diff options
277 files changed, 13381 insertions, 2290 deletions
diff --git a/bootstrap/bin/no_dot_erlang.boot b/bootstrap/bin/no_dot_erlang.boot Binary files differindex e259dc6ffd..a49298d260 100644 --- a/bootstrap/bin/no_dot_erlang.boot +++ b/bootstrap/bin/no_dot_erlang.boot diff --git a/bootstrap/bin/start.boot b/bootstrap/bin/start.boot Binary files differindex fc355b9c10..a49298d260 100644 --- a/bootstrap/bin/start.boot +++ b/bootstrap/bin/start.boot diff --git a/bootstrap/bin/start_clean.boot b/bootstrap/bin/start_clean.boot Binary files differindex fc355b9c10..a49298d260 100644 --- a/bootstrap/bin/start_clean.boot +++ b/bootstrap/bin/start_clean.boot diff --git a/bootstrap/lib/compiler/ebin/beam_listing.beam b/bootstrap/lib/compiler/ebin/beam_listing.beam Binary files differindex 74695349ba..281e9c2d03 100644 --- a/bootstrap/lib/compiler/ebin/beam_listing.beam +++ b/bootstrap/lib/compiler/ebin/beam_listing.beam diff --git a/bootstrap/lib/compiler/ebin/compile.beam b/bootstrap/lib/compiler/ebin/compile.beam Binary files differindex 9aea14162a..f6ad883ed8 100644 --- a/bootstrap/lib/compiler/ebin/compile.beam +++ b/bootstrap/lib/compiler/ebin/compile.beam diff --git a/bootstrap/lib/compiler/ebin/compiler.app b/bootstrap/lib/compiler/ebin/compiler.app index 24ff29cbf1..01e2c478d4 100644 --- a/bootstrap/lib/compiler/ebin/compiler.app +++ b/bootstrap/lib/compiler/ebin/compiler.app @@ -68,8 +68,7 @@ v3_codegen, v3_core, v3_kernel, - v3_kernel_pp, - v3_life + v3_kernel_pp ]}, {registered, []}, {applications, [kernel, stdlib]}, diff --git a/bootstrap/lib/compiler/ebin/v3_codegen.beam b/bootstrap/lib/compiler/ebin/v3_codegen.beam Binary files differindex e7cbc758f1..09ff7b253d 100644 --- a/bootstrap/lib/compiler/ebin/v3_codegen.beam +++ b/bootstrap/lib/compiler/ebin/v3_codegen.beam diff --git a/bootstrap/lib/compiler/ebin/v3_kernel.beam b/bootstrap/lib/compiler/ebin/v3_kernel.beam Binary files differindex 2286d9a646..56af498763 100644 --- a/bootstrap/lib/compiler/ebin/v3_kernel.beam +++ b/bootstrap/lib/compiler/ebin/v3_kernel.beam diff --git a/bootstrap/lib/compiler/ebin/v3_kernel_pp.beam b/bootstrap/lib/compiler/ebin/v3_kernel_pp.beam Binary files differindex e1b8c58c8f..b67d012f34 100644 --- a/bootstrap/lib/compiler/ebin/v3_kernel_pp.beam +++ b/bootstrap/lib/compiler/ebin/v3_kernel_pp.beam diff --git a/bootstrap/lib/compiler/ebin/v3_life.beam b/bootstrap/lib/compiler/ebin/v3_life.beam Binary files differdeleted file mode 100644 index c20070cf34..0000000000 --- a/bootstrap/lib/compiler/ebin/v3_life.beam +++ /dev/null diff --git a/bootstrap/lib/kernel/ebin/application.beam b/bootstrap/lib/kernel/ebin/application.beam Binary files differindex 05ecd20859..d242e045a1 100644 --- a/bootstrap/lib/kernel/ebin/application.beam +++ b/bootstrap/lib/kernel/ebin/application.beam diff --git a/bootstrap/lib/kernel/ebin/code_server.beam b/bootstrap/lib/kernel/ebin/code_server.beam Binary files differindex 536da5c692..f9ca469b0d 100644 --- a/bootstrap/lib/kernel/ebin/code_server.beam +++ b/bootstrap/lib/kernel/ebin/code_server.beam diff --git a/bootstrap/lib/kernel/ebin/disk_log.beam b/bootstrap/lib/kernel/ebin/disk_log.beam Binary files differindex e8a6acd235..fc71affe55 100644 --- a/bootstrap/lib/kernel/ebin/disk_log.beam +++ b/bootstrap/lib/kernel/ebin/disk_log.beam diff --git a/bootstrap/lib/kernel/ebin/disk_log_1.beam b/bootstrap/lib/kernel/ebin/disk_log_1.beam Binary files differindex 648c658f2e..a42a4c1b86 100644 --- a/bootstrap/lib/kernel/ebin/disk_log_1.beam +++ b/bootstrap/lib/kernel/ebin/disk_log_1.beam diff --git a/bootstrap/lib/kernel/ebin/global_group.beam b/bootstrap/lib/kernel/ebin/global_group.beam Binary files differindex ef0913f98d..1e496fdbbf 100644 --- a/bootstrap/lib/kernel/ebin/global_group.beam +++ b/bootstrap/lib/kernel/ebin/global_group.beam diff --git a/bootstrap/lib/kernel/ebin/group.beam b/bootstrap/lib/kernel/ebin/group.beam Binary files differindex 5c5f3bfa73..5eecaecec8 100644 --- a/bootstrap/lib/kernel/ebin/group.beam +++ b/bootstrap/lib/kernel/ebin/group.beam diff --git a/bootstrap/lib/kernel/ebin/group_history.beam b/bootstrap/lib/kernel/ebin/group_history.beam Binary files differindex 9328355025..80afde896f 100644 --- a/bootstrap/lib/kernel/ebin/group_history.beam +++ b/bootstrap/lib/kernel/ebin/group_history.beam diff --git a/bootstrap/lib/kernel/ebin/inet_db.beam b/bootstrap/lib/kernel/ebin/inet_db.beam Binary files differindex 4ce05435f7..0c77cee005 100644 --- a/bootstrap/lib/kernel/ebin/inet_db.beam +++ b/bootstrap/lib/kernel/ebin/inet_db.beam diff --git a/bootstrap/lib/kernel/ebin/inet_dns.beam b/bootstrap/lib/kernel/ebin/inet_dns.beam Binary files differindex ba611799a2..f36b12952e 100644 --- a/bootstrap/lib/kernel/ebin/inet_dns.beam +++ b/bootstrap/lib/kernel/ebin/inet_dns.beam diff --git a/bootstrap/lib/kernel/ebin/user.beam b/bootstrap/lib/kernel/ebin/user.beam Binary files differindex 947bc0f642..e8b49b2b33 100644 --- a/bootstrap/lib/kernel/ebin/user.beam +++ b/bootstrap/lib/kernel/ebin/user.beam diff --git a/bootstrap/lib/stdlib/ebin/base64.beam b/bootstrap/lib/stdlib/ebin/base64.beam Binary files differindex 00c908062a..4d8141a9c1 100644 --- a/bootstrap/lib/stdlib/ebin/base64.beam +++ b/bootstrap/lib/stdlib/ebin/base64.beam diff --git a/bootstrap/lib/stdlib/ebin/c.beam b/bootstrap/lib/stdlib/ebin/c.beam Binary files differindex 21ce30b8e7..3baf074b66 100644 --- a/bootstrap/lib/stdlib/ebin/c.beam +++ b/bootstrap/lib/stdlib/ebin/c.beam diff --git a/bootstrap/lib/stdlib/ebin/dets.beam b/bootstrap/lib/stdlib/ebin/dets.beam Binary files differindex 9e7801f637..e57da59d9d 100644 --- a/bootstrap/lib/stdlib/ebin/dets.beam +++ b/bootstrap/lib/stdlib/ebin/dets.beam diff --git a/bootstrap/lib/stdlib/ebin/edlin.beam b/bootstrap/lib/stdlib/ebin/edlin.beam Binary files differindex 1cabe095b3..06b7ba8b81 100644 --- a/bootstrap/lib/stdlib/ebin/edlin.beam +++ b/bootstrap/lib/stdlib/ebin/edlin.beam diff --git a/bootstrap/lib/stdlib/ebin/erl_compile.beam b/bootstrap/lib/stdlib/ebin/erl_compile.beam Binary files differindex d39b5a29b0..749c05c54f 100644 --- a/bootstrap/lib/stdlib/ebin/erl_compile.beam +++ b/bootstrap/lib/stdlib/ebin/erl_compile.beam diff --git a/bootstrap/lib/stdlib/ebin/erl_tar.beam b/bootstrap/lib/stdlib/ebin/erl_tar.beam Binary files differindex 6887230609..df0d6cf1b8 100644 --- a/bootstrap/lib/stdlib/ebin/erl_tar.beam +++ b/bootstrap/lib/stdlib/ebin/erl_tar.beam diff --git a/bootstrap/lib/stdlib/ebin/ets.beam b/bootstrap/lib/stdlib/ebin/ets.beam Binary files differindex 61ea75fb2b..95fbd243c3 100644 --- a/bootstrap/lib/stdlib/ebin/ets.beam +++ b/bootstrap/lib/stdlib/ebin/ets.beam diff --git a/bootstrap/lib/stdlib/ebin/filename.beam b/bootstrap/lib/stdlib/ebin/filename.beam Binary files differindex d48f6bc6e1..9ac0d3272c 100644 --- a/bootstrap/lib/stdlib/ebin/filename.beam +++ b/bootstrap/lib/stdlib/ebin/filename.beam diff --git a/bootstrap/lib/stdlib/ebin/io_lib.beam b/bootstrap/lib/stdlib/ebin/io_lib.beam Binary files differindex 82642fb353..3884967cba 100644 --- a/bootstrap/lib/stdlib/ebin/io_lib.beam +++ b/bootstrap/lib/stdlib/ebin/io_lib.beam diff --git a/bootstrap/lib/stdlib/ebin/io_lib_pretty.beam b/bootstrap/lib/stdlib/ebin/io_lib_pretty.beam Binary files differindex 6ed1e3b7cc..183e5b6278 100644 --- a/bootstrap/lib/stdlib/ebin/io_lib_pretty.beam +++ b/bootstrap/lib/stdlib/ebin/io_lib_pretty.beam diff --git a/bootstrap/lib/stdlib/ebin/rand.beam b/bootstrap/lib/stdlib/ebin/rand.beam Binary files differindex 887a09385c..2877b269c9 100644 --- a/bootstrap/lib/stdlib/ebin/rand.beam +++ b/bootstrap/lib/stdlib/ebin/rand.beam diff --git a/bootstrap/lib/stdlib/ebin/stdlib.app b/bootstrap/lib/stdlib/ebin/stdlib.app index 07309b88c4..a8abad00ea 100644 --- a/bootstrap/lib/stdlib/ebin/stdlib.app +++ b/bootstrap/lib/stdlib/ebin/stdlib.app @@ -101,6 +101,7 @@ timer, unicode, unicode_util, + uri_string, win32reg, zip]}, {registered,[timer_server,rsh_starter,take_over_monitor,pool_master, diff --git a/bootstrap/lib/stdlib/ebin/string.beam b/bootstrap/lib/stdlib/ebin/string.beam Binary files differindex eb00887566..406b2e19de 100644 --- a/bootstrap/lib/stdlib/ebin/string.beam +++ b/bootstrap/lib/stdlib/ebin/string.beam diff --git a/bootstrap/lib/stdlib/ebin/unicode_util.beam b/bootstrap/lib/stdlib/ebin/unicode_util.beam Binary files differindex 4fcf959fe6..1532f67ef4 100644 --- a/bootstrap/lib/stdlib/ebin/unicode_util.beam +++ b/bootstrap/lib/stdlib/ebin/unicode_util.beam diff --git a/bootstrap/lib/stdlib/ebin/uri_string.beam b/bootstrap/lib/stdlib/ebin/uri_string.beam Binary files differnew file mode 100644 index 0000000000..f60b12037a --- /dev/null +++ b/bootstrap/lib/stdlib/ebin/uri_string.beam diff --git a/bootstrap/lib/stdlib/ebin/zip.beam b/bootstrap/lib/stdlib/ebin/zip.beam Binary files differindex da5131d05f..1582220c5d 100644 --- a/bootstrap/lib/stdlib/ebin/zip.beam +++ b/bootstrap/lib/stdlib/ebin/zip.beam diff --git a/erts/configure.in b/erts/configure.in index 508e99a415..2f6043ee85 100644 --- a/erts/configure.in +++ b/erts/configure.in @@ -595,6 +595,8 @@ AS_HELP_STRING([--enable-pgo], esac ],enable_pgo=default) +LM_CHECK_ENABLE_CFLAG([-fprofile-use -fprofile-correction -Werror],[PROFILE_CORRECTION]) + USE_PGO=false AC_MSG_CHECKING([whether to do PGO of erts]) if test $enable_pgo = no; then @@ -607,22 +609,16 @@ elif test $CROSS_COMPILING = yes; then fi elif test "X$host" = "Xwin32"; then AC_MSG_RESULT([no, not supported in windows]) -elif test "X$PROFILE_GENERATE" = "Xtrue" -a "X$PROFILE_USE" = "Xtrue"; then +elif test "X$PROFILE_GENERATE" = "Xtrue" -a "X$PROFILE_USE" = "Xtrue" -a "X$PROFILE_CORRECTION" = "Xtrue"; then + ## We need -fprofile-generate and -fprofile-correction support to use PGO with + ## gcc as multiple threads run within the executed object files USE_PGO=true - AC_MSG_RESULT([yes, using -fprofile-generate]) PROFILE_COMPILER=gcc -# check if $CC accepts -fprofile-correction, if so we can use PGO on multi-threaded files. - LM_CHECK_ENABLE_CFLAG([-fprofile-use -fprofile-correction -Werror],[PROFILE_CORRECTION]) - if test "X$PROFILE_CORRECTION" = "Xtrue"; then - PROFILE_CORRECTION="-fprofile-correction" - else - PROFILE_CORRECTION="" - fi - AC_SUBST(PROFILE_CORRECTION) + AC_MSG_RESULT([yes, using -fprofile-generate -fprofile-correction]) elif test "X$PROFILE_INSTR_GENERATE" = "Xtrue" -a "X$PROFILE_INSTR_USE" = "Xtrue"; then USE_PGO=true - AC_MSG_RESULT([yes, using -fprofile-instr-generate]) PROFILE_COMPILER=clang + AC_MSG_RESULT([yes, using -fprofile-instr-generate]) else if $enable_pgo = yes; then AC_MSG_ERROR(cannot use PGO with this compiler) diff --git a/erts/doc/src/erlang.xml b/erts/doc/src/erlang.xml index d0588fe3c1..3b7b9d6a50 100644 --- a/erts/doc/src/erlang.xml +++ b/erts/doc/src/erlang.xml @@ -557,9 +557,7 @@ hello <name name="binary_to_term" arity="2"/> <fsummary>Decode an Erlang external term format binary.</fsummary> <desc> - <p>As <c>binary_to_term/1</c>, but takes options that affect decoding - of the binary.</p> - <p>Option:</p> + <p>As <c>binary_to_term/1</c>, but takes these options:</p> <taglist> <tag><c>safe</c></tag> <item> @@ -575,18 +573,31 @@ hello creation of new external function references. None of those resources are garbage collected, so unchecked creation of them can exhaust available memory.</p> - </item> - </taglist> - <p>Failure: <c>badarg</c> if <c>safe</c> is specified and unsafe - data is decoded.</p> <pre> -> <input>binary_to_term(<<131,100,0,5,104,101,108,108,111>>, [safe]).</input> +> <input>binary_to_term(<<131,100,0,5,"hello">>, [safe]).</input> ** exception error: bad argument > <input>hello.</input> hello -> <input>binary_to_term(<<131,100,0,5,104,101,108,108,111>>, [safe]).</input> +> <input>binary_to_term(<<131,100,0,5,"hello">>, [safe]).</input> hello </pre> + </item> + <tag><c>used</c></tag> + <item> + <p>Changes the return value to <c>{Term, Used}</c> where <c>Used</c> + is the number of bytes actually read from <c>Binary</c>.</p> + <pre> +> <input>Input = <<131,100,0,5,"hello","world">>.</input> +<<131,100,0,5,104,101,108,108,111,119,111,114,108,100>> +> <input>{Term, Used} = binary_to_term(Input, [used]).</input> +{hello, 9} +> <input>split_binary(Input, Used).</input> +{<<131,100,0,5,104,101,108,108,111>>, <<"world">>} +</pre> + </item> + </taglist> + <p>Failure: <c>badarg</c> if <c>safe</c> is specified and unsafe + data is decoded.</p> <p>See also <seealso marker="#term_to_binary/1"><c>term_to_binary/1</c></seealso>, <seealso marker="#binary_to_term/1"> @@ -4730,11 +4741,11 @@ RealSystem = system + MissedSystem</code> </func> <func> - <name name="process_flag" arity="2" clause_i="3"/> + <name name="process_flag" arity="2" clause_i="3" + anchor="process_flag_min_heap_size"/> <fsummary>Set process flag min_heap_size for the calling process. </fsummary> <desc> - <marker id="process_flag_min_heap_size"/> <p>Changes the minimum heap size for the calling process.</p> <p>Returns the old value of the flag.</p> </desc> @@ -4752,12 +4763,12 @@ RealSystem = system + MissedSystem</code> </func> <func> - <name name="process_flag" arity="2" clause_i="5"/> + <name name="process_flag" arity="2" clause_i="5" + anchor="process_flag_max_heap_size"/> <fsummary>Set process flag max_heap_size for the calling process. </fsummary> <type name="max_heap_size"/> <desc> - <marker id="process_flag_max_heap_size"/> <p>This flag sets the maximum heap size for the calling process. If <c><anno>MaxHeapSize</anno></c> is an integer, the system default values for <c>kill</c> and <c>error_logger</c> are used. @@ -4826,12 +4837,12 @@ RealSystem = system + MissedSystem</code> </func> <func> - <name name="process_flag" arity="2" clause_i="6"/> + <name name="process_flag" arity="2" clause_i="6" + anchor="process_flag_message_queue_data"/> <fsummary>Set process flag message_queue_data for the calling process. </fsummary> <type name="message_queue_data"/> <desc> - <marker id="process_flag_message_queue_data"/> <p>This flag determines how messages in the message queue are stored, as follows:</p> <taglist> @@ -4868,11 +4879,12 @@ RealSystem = system + MissedSystem</code> </func> <func> - <name name="process_flag" arity="2" clause_i="7"/> + <name name="process_flag" arity="2" clause_i="7" + anchor="process_flag_priority"/> <fsummary>Set process flag priority for the calling process.</fsummary> <type name="priority_level"/> <desc> - <p><marker id="process_flag_priority"></marker> + <p> Sets the process priority. <c><anno>Level</anno></c> is an atom. Four priority levels exist: <c>low</c>, <c>normal</c>, <c>high</c>, and <c>max</c>. Default @@ -6347,10 +6359,10 @@ true</pre> </func> <func> - <name name="statistics" arity="1" clause_i="1"/> + <name name="statistics" arity="1" clause_i="1" + anchor="statistics_active_tasks"/> <fsummary>Information about active processes and ports.</fsummary> <desc> - <marker id="statistics_active_tasks"></marker> <p>Returns the same as <seealso marker="#statistics_active_tasks_all"> <c>statistics(active_tasks_all)</c></seealso> @@ -6362,10 +6374,10 @@ true</pre> </func> <func> - <name name="statistics" arity="1" clause_i="2"/> + <name name="statistics" arity="1" clause_i="2" + anchor="statistics_active_tasks_all"/> <fsummary>Information about active processes and ports.</fsummary> <desc> - <marker id="statistics_active_tasks_all"></marker> <p>Returns a list where each element represents the amount of active processes and ports on each run queue and its associated schedulers. That is, the number of processes and @@ -6413,10 +6425,10 @@ true</pre> </func> <func> - <name name="statistics" arity="1" clause_i="4"/> + <name name="statistics" arity="1" clause_i="4" + anchor="statistics_exact_reductions"/> <fsummary>Information about exact reductions.</fsummary> <desc> - <marker id="statistics_exact_reductions"></marker> <p>Returns the number of exact reductions.</p> <note> <p><c>statistics(exact_reductions)</c> is @@ -6451,10 +6463,10 @@ true</pre> </func> <func> - <name name="statistics" arity="1" clause_i="7"/> + <name name="statistics" arity="1" clause_i="7" + anchor="statistics_microstate_accounting"/> <fsummary>Information about microstate accounting.</fsummary> <desc> - <marker id="statistics_microstate_accounting"></marker> <p>Microstate accounting can be used to measure how much time the Erlang runtime system spends doing various tasks. It is designed to be as lightweight as possible, but some overhead exists when this @@ -6594,10 +6606,10 @@ lists:map( </func> <func> - <name name="statistics" arity="1" clause_i="8"/> + <name name="statistics" arity="1" clause_i="8" + anchor="statistics_reductions"/> <fsummary>Information about reductions.</fsummary> <desc> - <marker id="statistics_reductions"></marker> <p>Returns information about reductions, for example:</p> <pre> > <input>statistics(reductions).</input> @@ -6613,9 +6625,10 @@ lists:map( </func> <func> - <name name="statistics" arity="1" clause_i="9"/> + <name name="statistics" arity="1" clause_i="9" + anchor="statistics_run_queue"/> <fsummary>Information about the run-queues.</fsummary> - <desc><marker id="statistics_run_queue"></marker> + <desc> <p>Returns the total length of all normal run-queues. That is, the number of processes and ports that are ready to run on all available normal run-queues. Dirty run queues are not part of the @@ -6629,9 +6642,10 @@ lists:map( </func> <func> - <name name="statistics" arity="1" clause_i="10"/> + <name name="statistics" arity="1" clause_i="10" + anchor="statistics_run_queue_lengths"/> <fsummary>Information about the run-queue lengths.</fsummary> - <desc><marker id="statistics_run_queue_lengths"></marker> + <desc> <p>Returns the same as <seealso marker="#statistics_run_queue_lengths_all"> <c>statistics(run_queue_lengths_all)</c></seealso> @@ -6643,9 +6657,10 @@ lists:map( </func> <func> - <name name="statistics" arity="1" clause_i="11"/> + <name name="statistics" arity="1" clause_i="11" + anchor="statistics_run_queue_lengths_all"/> <fsummary>Information about the run-queue lengths.</fsummary> - <desc><marker id="statistics_run_queue_lengths_all"></marker> + <desc> <p>Returns a list where each element represents the amount of processes and ports ready to run for each run queue. Values for normal run queues are located first in the @@ -6703,10 +6718,10 @@ lists:map( </func> <func> - <name name="statistics" arity="1" clause_i="13"/> + <name name="statistics" arity="1" clause_i="13" + anchor="statistics_scheduler_wall_time"/> <fsummary>Information about each schedulers work time.</fsummary> <desc> - <marker id="statistics_scheduler_wall_time"></marker> <p>Returns a list of tuples with <c>{<anno>SchedulerId</anno>, <anno>ActiveTime</anno>, <anno>TotalTime</anno>}</c>, where @@ -6824,10 +6839,10 @@ ok </func> <func> - <name name="statistics" arity="1" clause_i="14"/> + <name name="statistics" arity="1" clause_i="14" + anchor="statistics_scheduler_wall_time_all"/> <fsummary>Information about each schedulers work time.</fsummary> <desc> - <marker id="statistics_scheduler_wall_time_all"></marker> <p>The same as <seealso marker="#statistics_scheduler_wall_time"><c>statistics(scheduler_wall_time)</c></seealso>, except that it also include information about all dirty I/O @@ -6849,9 +6864,10 @@ ok </desc> </func> <func> - <name name="statistics" arity="1" clause_i="15"/> + <name name="statistics" arity="1" clause_i="15" + anchor="statistics_total_active_tasks"/> <fsummary>Information about active processes and ports.</fsummary> - <desc><marker id="statistics_total_active_tasks"></marker> + <desc> <p>The same as calling <c>lists:sum(</c><seealso marker="#statistics_active_tasks"><c>statistics(active_tasks)</c></seealso><c>)</c>, but more efficient.</p> @@ -6859,9 +6875,10 @@ ok </func> <func> - <name name="statistics" arity="1" clause_i="16"/> + <name name="statistics" arity="1" clause_i="16" + anchor="statistics_total_active_tasks_all"/> <fsummary>Information about active processes and ports.</fsummary> - <desc><marker id="statistics_total_active_tasks_all"></marker> + <desc> <p>The same as calling <c>lists:sum(</c><seealso marker="#statistics_active_tasks_all"><c>statistics(active_tasks_all)</c></seealso><c>)</c>, but more efficient.</p> @@ -6869,9 +6886,10 @@ ok </func> <func> - <name name="statistics" arity="1" clause_i="17"/> + <name name="statistics" arity="1" clause_i="17" + anchor="statistics_total_run_queue_lengths"/> <fsummary>Information about the run-queue lengths.</fsummary> - <desc><marker id="statistics_total_run_queue_lengths"></marker> + <desc> <p>The same as calling <c>lists:sum(</c><seealso marker="#statistics_run_queue_lengths"><c>statistics(run_queue_lengths)</c></seealso><c>)</c>, but more efficient.</p> @@ -6879,9 +6897,10 @@ ok </func> <func> - <name name="statistics" arity="1" clause_i="18"/> + <name name="statistics" arity="1" clause_i="18" + anchor="statistics_total_run_queue_lengths_all"/> <fsummary>Information about the run-queue lengths.</fsummary> - <desc><marker id="statistics_total_run_queue_lengths_all"></marker> + <desc> <p>The same as calling <c>lists:sum(</c><seealso marker="#statistics_run_queue_lengths_all"><c>statistics(run_queue_lengths_all)</c></seealso><c>)</c>, but more efficient.</p> @@ -7029,7 +7048,8 @@ ok </func> <func> - <name name="system_flag" arity="2" clause_i="2"/> + <name name="system_flag" arity="2" clause_i="2" + anchor="system_flag_cpu_topology"/> <fsummary>Set system flag <c>cpu_topology</c>.</fsummary> <type name="cpu_topology"/> <type name="level_entry"/> @@ -7038,7 +7058,7 @@ ok <type name="info_list"/> <desc> <warning> - <p><marker id="system_flag_cpu_topology"></marker> + <p> <em>This argument is deprecated.</em> Instead of using this argument, use command-line argument <seealso marker="erts:erl#+sct"><c>+sct</c></seealso> in @@ -7076,10 +7096,11 @@ ok </func> <func> - <name name="system_flag" arity="2" clause_i="3"/> + <name name="system_flag" arity="2" clause_i="3" + anchor="system_flag_dirty_cpu_schedulers_online"/> <fsummary>Set system_flag_dirty_cpu_schedulers_online.</fsummary> <desc> - <p><marker id="system_flag_dirty_cpu_schedulers_online"></marker> + <p> Sets the number of dirty CPU schedulers online. Range is <c><![CDATA[1 <= DirtyCPUSchedulersOnline <= N]]></c>, where <c>N</c> is the smallest of the return values of @@ -7123,10 +7144,11 @@ ok </func> <func> - <name name="system_flag" arity="2" clause_i="5"/> + <name name="system_flag" arity="2" clause_i="5" + anchor="system_flag_microstate_accounting"/> <fsummary>Set system flag microstate_accounting.</fsummary> <desc> - <p><marker id="system_flag_microstate_accounting"></marker> + <p> Turns on/off microstate accounting measurements. When passing reset, all counters are reset to 0.</p> <p>For more information see @@ -7168,28 +7190,29 @@ ok </func> <func> - <name name="system_flag" arity="2" clause_i="8"/> + <name name="system_flag" arity="2" clause_i="8" + anchor="system_flag_max_heap_size"/> <fsummary>Set system flag max_heap_size.</fsummary> <type name="max_heap_size"/> <desc> - <marker id="system_flag_max_heap_size"></marker> <p> Sets the default maximum heap size settings for processes. The size is specified in words. The new <c>max_heap_size</c> effects only processes spawned efter the change has been made. <c>max_heap_size</c> can be set for individual processes using <seealso marker="#spawn_opt/4"><c>spawn_opt/2,3,4</c></seealso> or - <seealso marker="#process_flag_message_queue_data"> + <seealso marker="#process_flag_max_heap_size"> <c>process_flag/2</c></seealso>.</p> <p>Returns the old value of the flag.</p> </desc> </func> <func> - <name name="system_flag" arity="2" clause_i="9"/> + <name name="system_flag" arity="2" clause_i="9" + anchor="system_flag_multi_scheduling"/> <fsummary>Set system flag multi_scheduling.</fsummary> <desc> - <p><marker id="system_flag_multi_scheduling"></marker> + <p> If multi-scheduling is enabled, more than one scheduler thread is used by the emulator. Multi-scheduling can be blocked in two different ways. Either all schedulers but @@ -7241,12 +7264,13 @@ ok </func> <func> - <name name="system_flag" arity="2" clause_i="10"/> + <name name="system_flag" arity="2" clause_i="10" + anchor="system_flag_scheduler_bind_type"/> <fsummary>Set system flag scheduler_bind_type.</fsummary> <type name="scheduler_bind_type"/> <desc> <warning> - <p><marker id="system_flag_scheduler_bind_type"></marker> + <p> <em>This argument is deprecated.</em> Instead of using this argument, use command-line argument <seealso marker="erts:erl#+sbt"><c>+sbt</c></seealso> in @@ -7367,10 +7391,11 @@ ok </func> <func> - <name name="system_flag" arity="2" clause_i="11"/> + <name name="system_flag" arity="2" clause_i="11" + anchor="system_flag_scheduler_wall_time"/> <fsummary>Set system flag scheduler_wall_time.</fsummary> <desc> - <p><marker id="system_flag_scheduler_wall_time"></marker> + <p> Turns on or off scheduler wall time measurements.</p> <p>For more information, see <seealso marker="#statistics_scheduler_wall_time"> @@ -7379,10 +7404,11 @@ ok </func> <func> - <name name="system_flag" arity="2" clause_i="12"/> + <name name="system_flag" arity="2" clause_i="12" + anchor="system_flag_schedulers_online"/> <fsummary>Set system flag schedulers_online.</fsummary> <desc> - <p><marker id="system_flag_schedulers_online"></marker> + <p> Sets the number of schedulers online. Range is <c><![CDATA[1 <= SchedulersOnline <= erlang:system_info(schedulers)]]></c>.</p> @@ -7421,10 +7447,11 @@ ok </func> <func> - <name name="system_flag" arity="2" clause_i="14"/> + <name name="system_flag" arity="2" clause_i="14" + anchor="system_flag_time_offset"/> <fsummary>Finalize the time offset.</fsummary> <desc> - <p><marker id="system_flag_time_offset"></marker> + <p> Finalizes the <seealso marker="#time_offset/0">time offset</seealso> when <seealso marker="time_correction#Single_Time_Warp_Mode">single time warp mode</seealso> is used. If another time warp mode @@ -7597,7 +7624,8 @@ ok </func> <func> - <name name="system_info" arity="1" clause_i="12"/> + <name name="system_info" arity="1" clause_i="12" + anchor="system_info_cpu_topology"/> <name name="system_info" arity="1" clause_i="13"/> <fsummary>Information about the CPU topology of the system.</fsummary> <type name="cpu_topology"/> @@ -7625,7 +7653,6 @@ ok </type_desc> <desc> <marker id="system_info_cpu_topology_tags"></marker> - <marker id="system_info_cpu_topology"></marker> <p>Returns various information about the CPU topology of the current system (emulator) as specified by <c><anno>Item</anno></c>:</p> diff --git a/erts/doc/src/notes.xml b/erts/doc/src/notes.xml index ff3c798179..bf41b61136 100644 --- a/erts/doc/src/notes.xml +++ b/erts/doc/src/notes.xml @@ -31,6 +31,72 @@ </header> <p>This document describes the changes made to the ERTS application.</p> +<section><title>Erts 9.1.4</title> + + <section><title>Fixed Bugs and Malfunctions</title> + <list> + <item> + <p>Microstate accounting sometimes produced incorrect + results for dirty schedulers.</p> + <p> + Own Id: OTP-14707</p> + </item> + <item> + <p>Fixed a regression in <c>zlib:gunzip/1</c> that + prevented it from working when the decompressed size was + a perfect multiple of 16384. This regression was + introduced in 20.1.1</p> + <p> + Own Id: OTP-14730 Aux Id: ERL-507 </p> + </item> + <item> + <p>Fixed a memory corruption bug in + <c>enif_inspect_iovec</c>; writable binaries stayed + writable after entering the iovec.</p> + <p> + Own Id: OTP-14745</p> + </item> + <item> + <p>Fixed a crash in <c>enif_inspect_iovec</c> on + encountering empty binaries.</p> + <p> + Own Id: OTP-14750</p> + </item> + <item> + <p><c>zlib:deflateParams/3</c> will no longer return + <c>buf_error</c> when called after <c>zlib:deflate/2</c> + with zlib <c>1.2.11</c>.</p> + <p> + Own Id: OTP-14751</p> + </item> + </list> + </section> + +</section> + +<section><title>Erts 9.1.3</title> + + <section><title>Fixed Bugs and Malfunctions</title> + <list> + <item> + <p>Added zlib:set_controlling_process/2 to move a + zstream() between processes.</p> + <p> + Own Id: OTP-14672 Aux Id: ERL-494 </p> + </item> + <item> + <p> + Fix so that schedulers are bound correctly when the first + available cpu is not the first detected cpu. e.g. when + using "taskset -c X..Y" when X is not equal to 0.</p> + <p> + Own Id: OTP-14694</p> + </item> + </list> + </section> + +</section> + <section><title>Erts 9.1.2</title> <section><title>Fixed Bugs and Malfunctions</title> diff --git a/erts/doc/src/zlib.xml b/erts/doc/src/zlib.xml index f5cc1b1e64..6f4c42da27 100644 --- a/erts/doc/src/zlib.xml +++ b/erts/doc/src/zlib.xml @@ -71,6 +71,11 @@ list_to_binary([Compressed|Last])</pre> called prior to a call to <seealso marker="#inflate/2"><c>inflate/2</c></seealso>. </item> + <tag><c>not_on_controlling_process</c></tag> + <item>The stream was used by a process that doesn't control it. Use + <seealso marker="#set_controlling_process/2"> + <c>set_controlling_process/2</c></seealso> if you need to transfer + a stream to a different process.</item> <tag><c>data_error</c></tag> <item>The data contains errors. </item> @@ -740,6 +745,15 @@ loop(Z, Handler, {finished, Output}) -> </func> <func> + <name name="set_controlling_process" arity="2"/> + <fsummary>Transfers ownership of a zlib stream.</fsummary> + <desc> + <p>Changes the controlling process of <c><anno>Z</anno></c> to + <c><anno>Pid</anno></c>, which must be a local process.</p> + </desc> + </func> + + <func> <name name="uncompress" arity="1"/> <fsummary>Uncompress data with standard zlib functionality.</fsummary> <desc> diff --git a/erts/emulator/Makefile.in b/erts/emulator/Makefile.in index 3117b24bcc..fdaf253fe5 100644 --- a/erts/emulator/Makefile.in +++ b/erts/emulator/Makefile.in @@ -74,9 +74,8 @@ endif endif ifeq ($(PROFILE_COMPILER), gcc) -PROFILE_CORRECTION=@PROFILE_CORRECTION@ PROFILE_GENERATE=-fprofile-generate -PROFILE_USE=-fprofile-use $(PROFILE_CORRECTION) +PROFILE_USE=-fprofile-use -fprofile-correction PROFILE_USE_DEPS=$(OBJDIR)/%_pu.gcda endif ifeq ($(PROFILE_COMPILER), clang) @@ -816,14 +815,7 @@ $(ERL_TOP)/lib/%.beam: INIT_OBJS = $(OBJDIR)/erl_main.o $(PRELOAD_OBJ) -# -fprofile-correction is needed in order to use PGO on erl_process -# as multiple threads execute in that file. -ifeq ($(PROFILE_CORRECTION),) -PROFILE_OBJS = $(OBJDIR)/beam_emu.o -RUN_OBJS = $(OBJDIR)/erl_process.o -else PROFILE_OBJS = $(OBJDIR)/beam_emu.o $(OBJDIR)/erl_process.o -endif EMU_OBJS = \ $(OBJDIR)/beam_opcodes.o \ diff --git a/erts/emulator/beam/beam_bif_load.c b/erts/emulator/beam/beam_bif_load.c index 00822d1415..87367b44ab 100644 --- a/erts/emulator/beam/beam_bif_load.c +++ b/erts/emulator/beam/beam_bif_load.c @@ -768,36 +768,45 @@ BIF_RETTYPE finish_after_on_load_2(BIF_ALIST_2) ErtsCodeIndex code_ix; Module* modp; + if (BIF_ARG_2 != am_false && BIF_ARG_2 != am_true) { + BIF_ERROR(BIF_P, BADARG); + } + if (!erts_try_seize_code_write_permission(BIF_P)) { ERTS_BIF_YIELD2(bif_export[BIF_finish_after_on_load_2], BIF_P, BIF_ARG_1, BIF_ARG_2); } - /* ToDo: Use code_ix staging instead of thread blocking */ - - erts_proc_unlock(BIF_P, ERTS_PROC_LOCK_MAIN); - erts_thr_progress_block(); - code_ix = erts_active_code_ix(); modp = erts_get_module(BIF_ARG_1, code_ix); - if (!modp || !modp->on_load || !modp->on_load->code_hdr) { - error: - erts_thr_progress_unblock(); - erts_proc_lock(BIF_P, ERTS_PROC_LOCK_MAIN); + if (!modp || !modp->on_load || !modp->on_load->code_hdr + || !modp->on_load->code_hdr->on_load_function_ptr) { + erts_release_code_write_permission(); BIF_ERROR(BIF_P, BADARG); } - if (modp->on_load->code_hdr->on_load_function_ptr == NULL) { - goto error; - } - if (BIF_ARG_2 != am_false && BIF_ARG_2 != am_true) { - goto error; - } if (BIF_ARG_2 == am_true) { + struct m mods[1]; + int is_blocking = 0; int i, num_exps; + erts_start_staging_code_ix(0); + code_ix = erts_staging_code_ix(); + modp = erts_get_module(BIF_ARG_1, code_ix); + + ASSERT(modp && modp->on_load && modp->on_load->code_hdr + && modp->on_load->code_hdr->on_load_function_ptr); + + if (erts_is_default_trace_enabled() + || IF_HIPE(hipe_need_blocking(modp))) { + + erts_proc_unlock(BIF_P, ERTS_PROC_LOCK_MAIN); + erts_thr_progress_block(); + is_blocking = 1; + } + /* * Make the code with the on_load function current. */ @@ -831,11 +840,13 @@ BIF_RETTYPE finish_after_on_load_2(BIF_ALIST_2) } } modp->curr.code_hdr->on_load_function_ptr = NULL; - set_default_trace_pattern(BIF_ARG_1); - #ifdef HIPE - hipe_redirect_to_module(modp); - #endif - } else if (BIF_ARG_2 == am_false) { + + mods[0].modp = modp; + mods[0].module = BIF_ARG_1; + mods[0].exception = THE_NON_VALUE; + return staging_epilogue(BIF_P, 1, am_true, is_blocking, mods, 1, 0); + } + else if (BIF_ARG_2 == am_false) { int i, num_exps; /* @@ -855,8 +866,6 @@ BIF_RETTYPE finish_after_on_load_2(BIF_ALIST_2) ep->beam[1] = 0; } } - erts_thr_progress_unblock(); - erts_proc_lock(BIF_P, ERTS_PROC_LOCK_MAIN); erts_release_code_write_permission(); BIF_RET(am_true); } diff --git a/erts/emulator/beam/beam_bp.c b/erts/emulator/beam/beam_bp.c index fe1e15701b..0832b3f374 100644 --- a/erts/emulator/beam/beam_bp.c +++ b/erts/emulator/beam/beam_bp.c @@ -998,7 +998,9 @@ do_call_trace(Process* c_p, ErtsCodeInfo* info, Eterm* reg, fixup_cp_before_trace(c_p, &return_to_trace); + ERTS_UNREQ_PROC_MAIN_LOCK(c_p); flags = erts_call_trace(c_p, info, ms, reg, local, &tracer); + ERTS_REQ_PROC_MAIN_LOCK(c_p); /* restore cp after potential fixup */ c_p->cp = cp_save; diff --git a/erts/emulator/beam/beam_debug.c b/erts/emulator/beam/beam_debug.c index 70078c8c59..509aa2a84f 100644 --- a/erts/emulator/beam/beam_debug.c +++ b/erts/emulator/beam/beam_debug.c @@ -408,7 +408,10 @@ print_op(fmtfn_t to, void *to_arg, int op, int size, BeamInstr* addr) addr++; ap = addr; } else { - BeamInstr instr_word = addr++[0]; +#if defined(ARCH_64) && defined(CODE_MODEL_SMALL) + BeamInstr instr_word = addr[0]; +#endif + addr++; /* * Copy all arguments to a local buffer for the unpacking. diff --git a/erts/emulator/beam/beam_emu.c b/erts/emulator/beam/beam_emu.c index aa94fbf536..bc95ccec52 100644 --- a/erts/emulator/beam/beam_emu.c +++ b/erts/emulator/beam/beam_emu.c @@ -400,12 +400,13 @@ static BeamInstr* apply_fun(Process* p, Eterm fun, Eterm args, Eterm* reg) NOINLINE; static Eterm new_fun(Process* p, Eterm* reg, ErlFunEntry* fe, int num_free) NOINLINE; -static Eterm new_map(Process* p, Eterm* reg, Uint live, Uint n, BeamInstr* ptr) NOINLINE; -static Eterm new_small_map_lit(Process* p, Eterm* reg, Eterm keys_literal, +static Eterm erts_gc_new_map(Process* p, Eterm* reg, Uint live, + Uint n, BeamInstr* ptr) NOINLINE; +static Eterm erts_gc_new_small_map_lit(Process* p, Eterm* reg, Eterm keys_literal, Uint live, BeamInstr* ptr) NOINLINE; -static Eterm update_map_assoc(Process* p, Eterm* reg, Uint live, +static Eterm erts_gc_update_map_assoc(Process* p, Eterm* reg, Uint live, Uint n, BeamInstr* new_p) NOINLINE; -static Eterm update_map_exact(Process* p, Eterm* reg, Uint live, +static Eterm erts_gc_update_map_exact(Process* p, Eterm* reg, Uint live, Uint n, Eterm* new_p) NOINLINE; static Eterm get_map_element(Eterm map, Eterm key); static Eterm get_map_element_hash(Eterm map, Eterm key, Uint32 hx); @@ -2755,7 +2756,7 @@ do { \ static Eterm -new_map(Process* p, Eterm* reg, Uint live, Uint n, BeamInstr* ptr) +erts_gc_new_map(Process* p, Eterm* reg, Uint live, Uint n, BeamInstr* ptr) { Uint i; Uint need = n + 1 /* hdr */ + 1 /*size*/ + 1 /* ptr */ + 1 /* arity */; @@ -2812,7 +2813,8 @@ new_map(Process* p, Eterm* reg, Uint live, Uint n, BeamInstr* ptr) } static Eterm -new_small_map_lit(Process* p, Eterm* reg, Eterm keys_literal, Uint live, BeamInstr* ptr) +erts_gc_new_small_map_lit(Process* p, Eterm* reg, Eterm keys_literal, + Uint live, BeamInstr* ptr) { Eterm* keys = tuple_val(keys_literal); Uint n = arityval(*keys); @@ -2846,7 +2848,8 @@ new_small_map_lit(Process* p, Eterm* reg, Eterm keys_literal, Uint live, BeamIns } static Eterm -update_map_assoc(Process* p, Eterm* reg, Uint live, Uint n, BeamInstr* new_p) +erts_gc_update_map_assoc(Process* p, Eterm* reg, Uint live, + Uint n, BeamInstr* new_p) { Uint num_old; Uint num_updates; @@ -2892,7 +2895,7 @@ update_map_assoc(Process* p, Eterm* reg, Uint live, Uint n, BeamInstr* new_p) */ if (num_old == 0) { - return new_map(p, reg, live, n, new_p); + return erts_gc_new_map(p, reg, live, n, new_p); } /* @@ -3048,7 +3051,7 @@ update_map_assoc(Process* p, Eterm* reg, Uint live, Uint n, BeamInstr* new_p) */ static Eterm -update_map_exact(Process* p, Eterm* reg, Uint live, Uint n, Eterm* new_p) +erts_gc_update_map_exact(Process* p, Eterm* reg, Uint live, Uint n, Eterm* new_p) { Uint i; Uint num_old; @@ -3199,13 +3202,16 @@ erts_is_builtin(Eterm Mod, Eterm Name, int arity) Export e; Export* ep; - if (Mod == am_erlang && Name == am_apply && arity == 3) { - /* - * Special case. apply/3 is built-in (implemented in C), - * but implemented in a different way than all other - * BIFs. - */ - return 1; + if (Mod == am_erlang) { + /* + * Special case for built-in functions that are implemented + * as instructions as opposed to SNIFs. + */ + if (Name == am_apply && (arity == 2 || arity == 3)) { + return 1; + } else if (Name == am_yield && arity == 0) { + return 1; + } } e.info.mfa.module = Mod; diff --git a/erts/emulator/beam/beam_load.c b/erts/emulator/beam/beam_load.c index 00dd28b26c..4fcfea527c 100644 --- a/erts/emulator/beam/beam_load.c +++ b/erts/emulator/beam/beam_load.c @@ -806,11 +806,6 @@ erts_finish_loading(Binary* magic, Process* c_p, struct erl_module_instance* inst_p; Uint size; - /* - * No other process may run since we will update the export - * table which is not protected by any locks. - */ - ERTS_LC_ASSERT(erts_initialized == 0 || erts_has_code_write_permission() || erts_thr_progress_is_blocking()); /* @@ -2389,8 +2384,18 @@ load_code(LoaderState* stp) code[ci++] = NIL; break; case TAG_q: - new_literal_patch(stp, ci); - code[ci++] = tmp_op->a[arg].val; + { + BeamInstr val = tmp_op->a[arg].val; + Eterm term = stp->literals[val].term; + new_literal_patch(stp, ci); + code[ci++] = val; + switch (loader_tag(term)) { + case LOADER_X_REG: + case LOADER_Y_REG: + LoadError1(stp, "the term '%T' would be confused " + "with a register", term); + } + } break; default: LoadError1(stp, "bad tag %d for general source", @@ -5027,7 +5032,7 @@ freeze_code(LoaderState* stp) */ codev[pos] = (BeamInstr) (codev + value); } else { -#ifdef DEBUG +#if defined(DEBUG) && defined(BEAM_WIDE_MASK) Uint w; #endif Sint32 rel = lp->offset + value; diff --git a/erts/emulator/beam/beam_ranges.c b/erts/emulator/beam/beam_ranges.c index 6e373a3480..01bda7f3c1 100644 --- a/erts/emulator/beam/beam_ranges.c +++ b/erts/emulator/beam/beam_ranges.c @@ -32,6 +32,15 @@ typedef struct { erts_atomic_t end; /* (BeamInstr*) Points one word beyond last function in module. */ } Range; +/* + * Used for crash dumping of literals. The size of erts_dump_lit_areas is + * always twice the number of active ranges (to allow for literals in both + * current and old code). + */ + +ErtsLiteralArea** erts_dump_lit_areas; +Uint erts_dump_num_lit_areas; + /* Range 'end' needs to be atomic as we purge module by setting end=start in active code_ix */ #define RANGE_END(R) ((BeamInstr*)erts_atomic_read_nob(&(R)->end)) @@ -97,6 +106,11 @@ erts_init_ranges(void) r[i].allocated = 0; erts_atomic_init_nob(&r[i].mid, 0); } + + erts_dump_num_lit_areas = 8; + erts_dump_lit_areas = (ErtsLiteralArea **) + erts_alloc(ERTS_ALC_T_CRASH_DUMP, + erts_dump_num_lit_areas * sizeof(ErtsLiteralArea*)); } void @@ -164,6 +178,14 @@ erts_end_staging_ranges(int commit) erts_atomic_set_nob(&r[dst].mid, (erts_aint_t) (r[dst].modules + r[dst].n / 2)); + + if (r[dst].allocated * 2 > erts_dump_num_lit_areas) { + erts_dump_num_lit_areas *= 2; + erts_dump_lit_areas = (ErtsLiteralArea **) + erts_realloc(ERTS_ALC_T_CRASH_DUMP, + (void *) erts_dump_lit_areas, + erts_dump_num_lit_areas * sizeof(ErtsLiteralArea*)); + } } } diff --git a/erts/emulator/beam/bif.c b/erts/emulator/beam/bif.c index ad555bb195..4b11884f38 100644 --- a/erts/emulator/beam/bif.c +++ b/erts/emulator/beam/bif.c @@ -4200,7 +4200,7 @@ BIF_RETTYPE list_to_port_1(BIF_ALIST_1) cp += 6; /* strlen("#Port<") */ - if (sscanf(cp, "%u.%u>", &n, &p) < 2) + if (sscanf(cp, "%u.%u>", (unsigned int*)&n, (unsigned int*)&p) < 2) goto bad; if (p > ERTS_MAX_PORT_NUMBER) diff --git a/erts/emulator/beam/break.c b/erts/emulator/beam/break.c index 35b2365655..d9ee940662 100644 --- a/erts/emulator/beam/break.c +++ b/erts/emulator/beam/break.c @@ -58,6 +58,8 @@ static void dump_attributes(fmtfn_t to, void *to_arg, byte* ptr, int size); extern char* erts_system_version[]; +#define WRITE_BUFFER_SIZE (64*1024) + static void port_info(fmtfn_t to, void *to_arg) { @@ -673,18 +675,28 @@ bin_check(void) static Sint64 crash_dump_limit = ERTS_SINT64_MAX; static Sint64 crash_dump_written = 0; -static int crash_dump_limited_writer(void* vfdp, char* buf, size_t len) +typedef struct LimitedWriterInfo_ { + fmtfn_t to; + void* to_arg; +} LimitedWriterInfo; + +static int +crash_dump_limited_writer(void* vfdp, char* buf, size_t len) { const char stop_msg[] = "\n=abort:CRASH DUMP SIZE LIMIT REACHED\n"; + LimitedWriterInfo* lwi = (LimitedWriterInfo *) vfdp; crash_dump_written += len; if (crash_dump_written <= crash_dump_limit) { - return erts_write_fd(vfdp, buf, len); + return lwi->to(lwi->to_arg, buf, len); } len -= (crash_dump_written - crash_dump_limit); - erts_write_fd(vfdp, buf, len); - erts_write_fd(vfdp, (char*)stop_msg, sizeof(stop_msg)-1); + lwi->to(lwi->to_arg, buf, len); + lwi->to(lwi->to_arg, (char*)stop_msg, sizeof(stop_msg)-1); + if (lwi->to == &erts_write_fp) { + fclose((FILE *) lwi->to_arg); + } /* We assume that crash dump was called from erts_exit_vv() */ erts_exit_epilogue(); @@ -707,6 +719,9 @@ erl_crash_dump_v(char *file, int line, char* fmt, va_list args) int i; fmtfn_t to = &erts_write_fd; void* to_arg; + FILE* fp = 0; + LimitedWriterInfo lwi; + static char* write_buffer; /* 'static' to avoid a leak warning in valgrind */ if (ERTS_SOMEONE_IS_CRASH_DUMPING) return; @@ -810,9 +825,30 @@ erl_crash_dump_v(char *file, int line, char* fmt, va_list args) fd = open(dumpname,O_WRONLY | O_CREAT | O_TRUNC,0640); if (fd < 0) return; /* Can't create the crash dump, skip it */ - to_arg = (void*)&fd; + + /* + * Wrap into a FILE* so that we can use buffered output. Set an + * explicit buffer to make sure the first write does not fail because + * of a failure to allocate a buffer. + */ + write_buffer = (char *) erts_alloc_fnf(ERTS_ALC_T_TMP, WRITE_BUFFER_SIZE); + if (write_buffer && (fp = fdopen(fd, "w")) != NULL) { + setvbuf(fp, write_buffer, _IOFBF, WRITE_BUFFER_SIZE); + lwi.to = &erts_write_fp; + lwi.to_arg = (void*)fp; + } else { + lwi.to = &erts_write_fd; + lwi.to_arg = (void*)&fd; + } + if (to == &crash_dump_limited_writer) { + to_arg = (void *) &lwi; + } else { + to = lwi.to; + to_arg = lwi.to_arg; + } + time(&now); - erts_cbprintf(to, to_arg, "=erl_crash_dump:0.3\n%s", ctime(&now)); + erts_cbprintf(to, to_arg, "=erl_crash_dump:0.4\n%s", ctime(&now)); if (file != NULL) erts_cbprintf(to, to_arg, "The error occurred in file %s, line %d\n", file, line); @@ -916,6 +952,9 @@ erl_crash_dump_v(char *file, int line, char* fmt, va_list args) } erts_cbprintf(to, to_arg, "=end\n"); + if (fp) { + fclose(fp); + } close(fd); erts_fprintf(stderr,"done\n"); } diff --git a/erts/emulator/beam/erl_alloc.types b/erts/emulator/beam/erl_alloc.types index 878b971b07..2960272eab 100644 --- a/erts/emulator/beam/erl_alloc.types +++ b/erts/emulator/beam/erl_alloc.types @@ -260,6 +260,7 @@ type MREF_TAB LONG_LIVED SYSTEM magic_ref_table type MINDIRECTION FIXED_SIZE SYSTEM magic_indirection type BINARY_FIND SHORT_LIVED PROCESSES binary_find type OPEN_PORT_ENV TEMPORARY SYSTEM open_port_env +type CRASH_DUMP STANDARD SYSTEM crash_dump type THR_Q_EL STANDARD SYSTEM thr_q_element type THR_Q_EL_SL FIXED_SIZE SYSTEM sl_thr_q_element diff --git a/erts/emulator/beam/erl_cpu_topology.c b/erts/emulator/beam/erl_cpu_topology.c index 49f9beb19f..6f8d2f8c35 100644 --- a/erts/emulator/beam/erl_cpu_topology.c +++ b/erts/emulator/beam/erl_cpu_topology.c @@ -602,7 +602,7 @@ write_schedulers_bind_change(erts_cpu_topology_t *cpudata, int size) cpu_bind_order_sort(cpudata, size, cpu_bind_order, 1); - for (cpu_ix = 0; cpu_ix < size && cpu_ix < erts_no_schedulers; cpu_ix++) + for (cpu_ix = 0; cpu_ix < size && s_ix <= erts_no_schedulers; cpu_ix++) if (erts_is_cpu_available(cpuinfo, cpudata[cpu_ix].logical)) scheduler2cpu_map[s_ix++].bind_id = cpudata[cpu_ix].logical; } diff --git a/erts/emulator/beam/erl_io_queue.c b/erts/emulator/beam/erl_io_queue.c index 190ba6bbb9..40d69ea6b0 100644 --- a/erts/emulator/beam/erl_io_queue.c +++ b/erts/emulator/beam/erl_io_queue.c @@ -658,7 +658,7 @@ io_list_vec_count(Eterm obj, Uint *v_size, int erts_ioq_iolist_vec_len(Eterm obj, int* vsize, Uint* csize, Uint* pvsize, Uint* pcsize, - Uint* total_size, Uint blimit) + size_t* total_size, Uint blimit) { DECLARE_ESTACK(s); Eterm* objp; @@ -669,7 +669,7 @@ erts_ioq_iolist_vec_len(Eterm obj, int* vsize, Uint* csize, Uint p_v_size = 0; Uint p_c_size = 0; Uint p_in_clist = 0; - Uint total; + size_t total; goto L_jump_start; /* avoid a push */ diff --git a/erts/emulator/beam/erl_io_queue.h b/erts/emulator/beam/erl_io_queue.h index 51abe99510..7d0fe6751c 100644 --- a/erts/emulator/beam/erl_io_queue.h +++ b/erts/emulator/beam/erl_io_queue.h @@ -103,7 +103,7 @@ Uint erts_ioq_sizeq(ErtsIOQueue *q); int erts_ioq_iolist_vec_len(Eterm obj, int* vsize, Uint* csize, Uint* pvsize, Uint* pcsize, - Uint* total_size, Uint blimit); + size_t* total_size, Uint blimit); int erts_ioq_iolist_to_vec(Eterm obj, SysIOVec* iov, ErtsIOQBinary** binv, ErtsIOQBinary* cbin, Uint bin_limit, int driver_binary); @@ -111,7 +111,7 @@ int erts_ioq_iolist_to_vec(Eterm obj, SysIOVec* iov, ERTS_GLB_INLINE int erts_ioq_iodata_vec_len(Eterm obj, int* vsize, Uint* csize, Uint* pvsize, Uint* pcsize, - Uint* total_size, Uint blimit); + size_t* total_size, Uint blimit); ERTS_GLB_INLINE int erts_ioq_iodata_to_vec(Eterm obj, SysIOVec* iov, ErtsIOQBinary** binv, ErtsIOQBinary* cbin, @@ -123,7 +123,7 @@ int erts_ioq_iodata_to_vec(Eterm obj, SysIOVec* iov, ERTS_GLB_INLINE int erts_ioq_iodata_vec_len(Eterm obj, int* vsize, Uint* csize, Uint* pvsize, Uint* pcsize, - Uint* total_size, Uint blimit) { + size_t* total_size, Uint blimit) { if (is_binary(obj)) { /* We optimize for when we get a procbin without a bit-offset * that fits in one iov slot diff --git a/erts/emulator/beam/erl_msacc.h b/erts/emulator/beam/erl_msacc.h index 2d4637f800..2588dec903 100644 --- a/erts/emulator/beam/erl_msacc.h +++ b/erts/emulator/beam/erl_msacc.h @@ -270,18 +270,32 @@ void erts_msacc_init_thread(char *type, int id, int liberty); #define ERTS_MSACC_PUSH_STATE_M() \ ERTS_MSACC_DECLARE_CACHE(); \ ERTS_MSACC_PUSH_STATE_CACHED_M() -#define ERTS_MSACC_PUSH_STATE_CACHED_M() \ - __erts_msacc_state = ERTS_MSACC_IS_ENABLED_CACHED() ? \ - erts_msacc_get_state_m__(__erts_msacc_cache) : ERTS_MSACC_STATE_OTHER +#define ERTS_MSACC_PUSH_STATE_CACHED_M() \ + do { \ + if (ERTS_MSACC_IS_ENABLED_CACHED()) { \ + ASSERT(!__erts_msacc_cache->unmanaged); \ + __erts_msacc_state = erts_msacc_get_state_m__(__erts_msacc_cache); \ + } else { \ + __erts_msacc_state = ERTS_MSACC_STATE_OTHER; \ + } \ + } while(0) #define ERTS_MSACC_SET_STATE_M(state) \ ERTS_MSACC_DECLARE_CACHE(); \ ERTS_MSACC_SET_STATE_CACHED_M(state) -#define ERTS_MSACC_SET_STATE_CACHED_M(state) \ - if (ERTS_MSACC_IS_ENABLED_CACHED()) \ - erts_msacc_set_state_m__(__erts_msacc_cache, state, 1) -#define ERTS_MSACC_POP_STATE_M() \ - if (ERTS_MSACC_IS_ENABLED_CACHED()) \ - erts_msacc_set_state_m__(__erts_msacc_cache, __erts_msacc_state, 0) +#define ERTS_MSACC_SET_STATE_CACHED_M(state) \ + do { \ + if (ERTS_MSACC_IS_ENABLED_CACHED()) { \ + ASSERT(!__erts_msacc_cache->unmanaged); \ + erts_msacc_set_state_m__(__erts_msacc_cache, state, 1); \ + } \ + } while(0) +#define ERTS_MSACC_POP_STATE_M() \ + do { \ + if (ERTS_MSACC_IS_ENABLED_CACHED()) { \ + ASSERT(!__erts_msacc_cache->unmanaged); \ + erts_msacc_set_state_m__(__erts_msacc_cache, __erts_msacc_state, 0); \ + } \ + } while(0) #define ERTS_MSACC_PUSH_AND_SET_STATE_M(state) \ ERTS_MSACC_PUSH_STATE_M(); ERTS_MSACC_SET_STATE_CACHED_M(state) diff --git a/erts/emulator/beam/erl_nif.c b/erts/emulator/beam/erl_nif.c index f7f12efe28..c21b139cfa 100644 --- a/erts/emulator/beam/erl_nif.c +++ b/erts/emulator/beam/erl_nif.c @@ -3323,36 +3323,38 @@ static int examine_iovec_term(Eterm list, UWord max_length, iovec_slice_t *resul size = binary_size(binary); binary_header = binary_val(binary); - /* If we're a sub-binary we'll need to check our underlying binary to - * determine whether we're on-heap or not. */ - if(thing_subtag(*binary_header) == SUB_BINARY_SUBTAG) { - ErlSubBin *sb = (ErlSubBin*)binary_header; - - /* Reject bitstrings */ - if((sb->bitoffs + sb->bitsize) > 0) { - return 0; + if (size > 0) { + /* If we're a sub-binary we'll need to check our underlying binary + * to determine whether we're on-heap or not. */ + if (thing_subtag(*binary_header) == SUB_BINARY_SUBTAG) { + ErlSubBin *sb = (ErlSubBin*)binary_header; + + /* Reject bitstrings */ + if((sb->bitoffs + sb->bitsize) > 0) { + return 0; + } + + ASSERT(size <= binary_size(sb->orig)); + binary_header = binary_val(sb->orig); } - ASSERT(size <= binary_size(sb->orig)); - binary_header = binary_val(sb->orig); - } - - if(thing_subtag(*binary_header) == HEAP_BINARY_SUBTAG) { - ASSERT(size <= ERL_ONHEAP_BIN_LIMIT); + if (thing_subtag(*binary_header) == HEAP_BINARY_SUBTAG) { + ASSERT(size <= ERL_ONHEAP_BIN_LIMIT); - result->iovec_len += 1; - result->onheap_size += size; - } else { - ASSERT(thing_subtag(*binary_header) == REFC_BINARY_SUBTAG); + result->iovec_len += 1; + result->onheap_size += size; + } else { + ASSERT(thing_subtag(*binary_header) == REFC_BINARY_SUBTAG); - result->iovec_len += 1 + size / MAX_SYSIOVEC_IOVLEN; - result->offheap_size += size; + result->iovec_len += 1 + size / MAX_SYSIOVEC_IOVLEN; + result->offheap_size += size; + } } result->sublist_length += 1; lookahead = CDR(cell); - if(result->sublist_length >= max_length) { + if (result->sublist_length >= max_length) { break; } } @@ -3385,6 +3387,10 @@ static void inspect_raw_binary_data(Eterm binary, ErlNifBinary *result) { if (thing_subtag(*parent_header) == REFC_BINARY_SUBTAG) { ProcBin *pb = (ProcBin*)parent_header; + if (pb->flags & (PB_IS_WRITABLE | PB_ACTIVE_WRITER)) { + erts_emasculate_writable_binary(pb); + } + ASSERT(pb->val != NULL); ASSERT(byte_offset < pb->size); ASSERT(&pb->bytes[byte_offset] >= (byte*)(pb->val)->orig_bytes); @@ -3428,7 +3434,7 @@ static int fill_iovec_with_slice(ErlNifEnv *env, /* If this isn't a refc binary, copy its contents to the onheap buffer * and reference that instead. */ - if (raw_data.ref_bin == NULL) { + if (raw_data.size > 0 && raw_data.ref_bin == NULL) { ASSERT(onheap_offset < onheap_data.size); ASSERT(slice->onheap_size > 0); @@ -3439,12 +3445,11 @@ static int fill_iovec_with_slice(ErlNifEnv *env, raw_data.ref_bin = onheap_data.ref_bin; } - ASSERT(raw_data.ref_bin != NULL); - while (raw_data.size > 0) { UWord chunk_len = MIN(raw_data.size, MAX_SYSIOVEC_IOVLEN); ASSERT(iovec_idx < iovec->iovcnt); + ASSERT(raw_data.ref_bin != NULL); iovec->iov[iovec_idx].iov_base = raw_data.data; iovec->iov[iovec_idx].iov_len = chunk_len; diff --git a/erts/emulator/beam/erl_node_tables.h b/erts/emulator/beam/erl_node_tables.h index 3bba673435..ee8277b5ea 100644 --- a/erts/emulator/beam/erl_node_tables.h +++ b/erts/emulator/beam/erl_node_tables.h @@ -271,5 +271,6 @@ erts_de_links_unlock(DistEntry *dep) #endif /* #if ERTS_GLB_INLINE_INCL_FUNC_DEF */ void erts_debug_test_node_tab_delayed_delete(Sint64 millisecs); +void erts_lcnt_update_distribution_locks(int enable); #endif diff --git a/erts/emulator/beam/erl_process.c b/erts/emulator/beam/erl_process.c index 3d12b3bc21..3c0a126fe2 100644 --- a/erts/emulator/beam/erl_process.c +++ b/erts/emulator/beam/erl_process.c @@ -3276,7 +3276,7 @@ scheduler_wait(int *fcalls, ErtsSchedulerData *esdp, ErtsRunQueue *rq) erts_aint32_t aux_work = 0; int thr_prgr_active = 1; erts_aint32_t flgs; - ERTS_MSACC_PUSH_STATE_M(); + ERTS_MSACC_PUSH_STATE(); ERTS_LC_ASSERT(erts_lc_runq_is_locked(rq)); @@ -3385,9 +3385,9 @@ scheduler_wait(int *fcalls, ErtsSchedulerData *esdp, ErtsRunQueue *rq) - 1) + 1; } else timeout = -1; - ERTS_MSACC_SET_STATE_CACHED_M(ERTS_MSACC_STATE_SLEEP); + ERTS_MSACC_SET_STATE_CACHED(ERTS_MSACC_STATE_SLEEP); res = erts_tse_twait(ssi->event, timeout); - ERTS_MSACC_POP_STATE_M(); + ERTS_MSACC_POP_STATE(); current_time = ERTS_SCHEDULER_IS_DIRTY(esdp) ? 0 : erts_get_monotonic_time(esdp); } while (res == EINTR); @@ -9701,7 +9701,7 @@ Process *erts_schedule(ErtsSchedulerData *esdp, Process *p, int calls) | ERTS_PROC_LOCK_STATUS | ERTS_PROC_LOCK_TRACE)); - ERTS_MSACC_SET_STATE_CACHED_M(ERTS_MSACC_STATE_OTHER); + ERTS_MSACC_SET_STATE_CACHED(ERTS_MSACC_STATE_OTHER); if (state & ERTS_PSFLG_FREE) { if (!is_normal_sched) { @@ -9916,7 +9916,7 @@ Process *erts_schedule(ErtsSchedulerData *esdp, Process *p, int calls) case 0: /* No process at all */ default: ASSERT(qmask == 0); - ERTS_MSACC_SET_STATE_CACHED_M(ERTS_MSACC_STATE_OTHER); + ERTS_MSACC_SET_STATE_CACHED(ERTS_MSACC_STATE_OTHER); goto check_activities_to_run; } @@ -10034,7 +10034,7 @@ Process *erts_schedule(ErtsSchedulerData *esdp, Process *p, int calls) } - ERTS_MSACC_SET_STATE_CACHED_M(ERTS_MSACC_STATE_EMULATOR); + ERTS_MSACC_SET_STATE_CACHED(ERTS_MSACC_STATE_EMULATOR); if (flags & ERTS_RUNQ_FLG_PROTECTED) @@ -10824,7 +10824,7 @@ request_system_task(Process *c_p, Eterm requester, Eterm target, goto badarg; req_type = tp[1]; req_id = tp[2]; - req_id_sz = is_immed(req_id) ? req_id : size_object(req_id); + req_id_sz = is_immed(req_id) ? 0 : size_object(req_id); tot_sz = req_id_sz; for (i = 0; i < ERTS_MAX_PROC_SYS_TASK_ARGS; i++) { int tix = 3 + i; diff --git a/erts/emulator/beam/erl_process_dump.c b/erts/emulator/beam/erl_process_dump.c index 5a2c262ff1..0b7f361622 100644 --- a/erts/emulator/beam/erl_process_dump.c +++ b/erts/emulator/beam/erl_process_dump.c @@ -32,6 +32,7 @@ #include "dist.h" #include "beam_catches.h" #include "erl_binary.h" +#include "erl_map.h" #define ERTS_WANT_EXTERNAL_TAGS #include "external.h" @@ -51,6 +52,11 @@ static void print_function_from_pc(fmtfn_t to, void *to_arg, BeamInstr* x); static void heap_dump(fmtfn_t to, void *to_arg, Eterm x); static void dump_binaries(fmtfn_t to, void *to_arg, Binary* root); static void dump_externally(fmtfn_t to, void *to_arg, Eterm term); +static void mark_literal(Eterm* ptr); +static void init_literal_areas(void); +static void dump_literals(fmtfn_t to, void *to_arg); +static void dump_module_literals(fmtfn_t to, void *to_arg, + ErtsLiteralArea* lit_area); static Binary* all_binaries; @@ -58,14 +64,14 @@ extern BeamInstr beam_apply[]; extern BeamInstr beam_exit[]; extern BeamInstr beam_continue_exit[]; - void erts_deep_process_dump(fmtfn_t to, void *to_arg) { int i, max = erts_ptab_max(&erts_proc); all_binaries = NULL; - + init_literal_areas(); + for (i = 0; i < max; i++) { Process *p = erts_pix2proc(i); if (p && p->i != ENULL) { @@ -75,6 +81,7 @@ erts_deep_process_dump(fmtfn_t to, void *to_arg) } } + dump_literals(to, to_arg); dump_binaries(to, to_arg, all_binaries); } @@ -373,7 +380,9 @@ heap_dump(fmtfn_t to, void *to_arg, Eterm x) next = (Eterm *) x; } else if (is_list(x)) { ptr = list_val(x); - if (ptr[0] != OUR_NIL) { + if (erts_is_literal(x, ptr)) { + mark_literal(ptr); + } else if (ptr[0] != OUR_NIL) { erts_print(to, to_arg, PTR_FMT ":l", ptr); dump_element(to, to_arg, ptr[0]); erts_putc(to, to_arg, '|'); @@ -392,7 +401,9 @@ heap_dump(fmtfn_t to, void *to_arg, Eterm x) ptr = boxed_val(x); hdr = *ptr; - if (hdr != OUR_NIL) { /* If not visited */ + if (erts_is_literal(x, ptr)) { + mark_literal(ptr); + } else if (hdr != OUR_NIL) { erts_print(to, to_arg, PTR_FMT ":", ptr); if (is_arity_value(hdr)) { Uint i; @@ -498,11 +509,77 @@ heap_dump(fmtfn_t to, void *to_arg, Eterm x) erts_print(to, to_arg, "p<%beu.%beu>\n", port_channel_no(x), port_number(x)); *ptr = OUR_NIL; + } else if (is_map_header(hdr)) { + if (is_flatmap_header(hdr)) { + flatmap_t* fmp = (flatmap_t *) flatmap_val(x); + Eterm* values = ptr + sizeof(flatmap_t) / sizeof(Eterm); + Uint map_size = fmp->size; + int i; + + erts_print(to, to_arg, "Mf" ETERM_FMT ":", map_size); + dump_element(to, to_arg, fmp->keys); + erts_putc(to, to_arg, ':'); + for (i = 0; i < map_size; i++) { + dump_element(to, to_arg, values[i]); + if (is_immed(values[i])) { + values[i] = make_small(0); + } + if (i < map_size-1) { + erts_putc(to, to_arg, ','); + } + } + erts_putc(to, to_arg, '\n'); + *ptr = OUR_NIL; + x = fmp->keys; + if (map_size) { + fmp->keys = (Eterm) next; + next = &values[map_size-1]; + } + continue; + } else { + Uint i; + Uint sz = 0; + Eterm* nodes = ptr + 1; + + switch (MAP_HEADER_TYPE(hdr)) { + case MAP_HEADER_TAG_HAMT_HEAD_ARRAY: + nodes++; + sz = 16; + erts_print(to, to_arg, "Mh" ETERM_FMT ":" ETERM_FMT ":", + hashmap_size(x), sz); + break; + case MAP_HEADER_TAG_HAMT_HEAD_BITMAP: + nodes++; + sz = hashmap_bitcount(MAP_HEADER_VAL(hdr)); + erts_print(to, to_arg, "Mh" ETERM_FMT ":" ETERM_FMT ":", + hashmap_size(x), sz); + break; + case MAP_HEADER_TAG_HAMT_NODE_BITMAP: + sz = hashmap_bitcount(MAP_HEADER_VAL(hdr)); + erts_print(to, to_arg, "Mn" ETERM_FMT ":", sz); + break; + } + *ptr = OUR_NIL; + for (i = 0; i < sz; i++) { + dump_element(to, to_arg, nodes[i]); + if (is_immed(nodes[i])) { + nodes[i] = make_small(0); + } + if (i < sz-1) { + erts_putc(to, to_arg, ','); + } + } + erts_putc(to, to_arg, '\n'); + x = nodes[0]; + nodes[0] = (Eterm) next; + next = &nodes[sz-1]; + continue; + } } else { /* * All other we dump in the external term format. */ - dump_externally(to, to_arg, x); + dump_externally(to, to_arg, x); erts_putc(to, to_arg, '\n'); *ptr = OUR_NIL; } @@ -564,11 +641,6 @@ dump_externally(fmtfn_t to, void *to_arg, Eterm term) } } - /* Do not handle maps */ - if (is_map(term)) { - term = am_undefined; - } - s = p = sbuf; erts_encode_ext(term, &p); erts_print(to, to_arg, "E%X:", p-s); @@ -577,6 +649,284 @@ dump_externally(fmtfn_t to, void *to_arg, Eterm term) } } +/* + * Handle dumping of literal areas. + */ + +static ErtsLiteralArea** lit_areas; +static Uint num_lit_areas; + +static int compare_areas(const void * a, const void * b) +{ + ErtsLiteralArea** a_p = (ErtsLiteralArea **) a; + ErtsLiteralArea** b_p = (ErtsLiteralArea **) b; + + if (*a_p < *b_p) { + return -1; + } else if (*b_p < *a_p) { + return 1; + } else { + return 0; + } +} + + +static void +init_literal_areas(void) +{ + int i; + Module* modp; + ErtsCodeIndex code_ix; + ErtsLiteralArea** area_p; + + code_ix = erts_active_code_ix(); + erts_rlock_old_code(code_ix); + + lit_areas = area_p = erts_dump_lit_areas; + num_lit_areas = 0; + for (i = 0; i < module_code_size(code_ix); i++) { + modp = module_code(i, code_ix); + if (modp == NULL) { + continue; + } + if (modp->curr.code_length > 0 && + modp->curr.code_hdr->literal_area) { + *area_p++ = modp->curr.code_hdr->literal_area; + } + if (modp->old.code_length > 0 && modp->old.code_hdr->literal_area) { + *area_p++ = modp->old.code_hdr->literal_area; + } + } + + num_lit_areas = area_p - lit_areas; + ASSERT(num_lit_areas <= erts_dump_num_lit_areas); + for (i = 0; i < num_lit_areas; i++) { + lit_areas[i]->off_heap = 0; + } + + qsort(lit_areas, num_lit_areas, sizeof(ErtsLiteralArea *), + compare_areas); + + erts_runlock_old_code(code_ix); +} + +static int search_areas(const void * a, const void * b) { + Eterm* key = (Eterm *) a; + ErtsLiteralArea** b_p = (ErtsLiteralArea **) b; + if (key < b_p[0]->start) { + return -1; + } else if (b_p[0]->end <= key) { + return 1; + } else { + return 0; + } +} + +static void mark_literal(Eterm* ptr) +{ + ErtsLiteralArea** ap; + + ap = bsearch(ptr, lit_areas, num_lit_areas, sizeof(ErtsLiteralArea*), + search_areas); + + /* + * If the literal was created by native code, this search will not + * find it and ap will be NULL. + */ + + if (ap) { + ap[0]->off_heap = (struct erl_off_heap_header *) 1; + } +} + + +static void +dump_literals(fmtfn_t to, void *to_arg) +{ + ErtsCodeIndex code_ix; + int i; + + code_ix = erts_active_code_ix(); + erts_rlock_old_code(code_ix); + + erts_print(to, to_arg, "=literals\n"); + for (i = 0; i < num_lit_areas; i++) { + if (lit_areas[i]->off_heap) { + dump_module_literals(to, to_arg, lit_areas[i]); + } + } + + erts_runlock_old_code(code_ix); +} + +static void +dump_module_literals(fmtfn_t to, void *to_arg, ErtsLiteralArea* lit_area) +{ + Eterm* htop; + Eterm* hend; + + htop = lit_area->start; + hend = lit_area->end; + while (htop < hend) { + Eterm w = *htop; + Eterm term; + Uint size; + + switch (primary_tag(w)) { + case TAG_PRIMARY_HEADER: + term = make_boxed(htop); + erts_print(to, to_arg, PTR_FMT ":", htop); + if (is_arity_value(w)) { + Uint i; + Uint arity = arityval(w); + + erts_print(to, to_arg, "t" ETERM_FMT ":", arity); + for (i = 1; i <= arity; i++) { + dump_element(to, to_arg, htop[i]); + if (i < arity) { + erts_putc(to, to_arg, ','); + } + } + erts_putc(to, to_arg, '\n'); + } else if (w == HEADER_FLONUM) { + FloatDef f; + char sbuf[31]; + int i; + + GET_DOUBLE_DATA((htop+1), f); + i = sys_double_to_chars(f.fd, sbuf, sizeof(sbuf)); + sys_memset(sbuf+i, 0, 31-i); + erts_print(to, to_arg, "F%X:%s\n", i, sbuf); + } else if (_is_bignum_header(w)) { + erts_print(to, to_arg, "B%T\n", term); + } else if (is_binary_header(w)) { + Uint tag = thing_subtag(w); + Uint size = binary_size(term); + Uint i; + + if (tag == HEAP_BINARY_SUBTAG) { + byte* p; + + erts_print(to, to_arg, "Yh%X:", size); + p = binary_bytes(term); + for (i = 0; i < size; i++) { + erts_print(to, to_arg, "%02X", p[i]); + } + } else if (tag == REFC_BINARY_SUBTAG) { + ProcBin* pb = (ProcBin *) binary_val(term); + Binary* val = pb->val; + + if (erts_atomic_xchg_nob(&val->intern.refc, 0) != 0) { + val->intern.flags = (UWord) all_binaries; + all_binaries = val; + } + erts_print(to, to_arg, + "Yc" PTR_FMT ":" PTR_FMT ":" PTR_FMT, + val, + pb->bytes - (byte *)val->orig_bytes, + size); + } else if (tag == SUB_BINARY_SUBTAG) { + ErlSubBin* Sb = (ErlSubBin *) binary_val(term); + Eterm* real_bin; + void* val; + + real_bin = boxed_val(Sb->orig); + if (thing_subtag(*real_bin) == REFC_BINARY_SUBTAG) { + /* + * Unvisited REFC_BINARY: Point directly to + * the binary. + */ + ProcBin* pb = (ProcBin *) real_bin; + val = pb->val; + } else { + /* + * Heap binary or visited REFC binary: Point + * to heap binary or ProcBin on the heap. + */ + val = real_bin; + } + erts_print(to, to_arg, + "Ys" PTR_FMT ":" PTR_FMT ":" PTR_FMT, + val, Sb->offs, size); + } + erts_putc(to, to_arg, '\n'); + } else if (is_map_header(w)) { + if (is_flatmap_header(w)) { + flatmap_t* fmp = (flatmap_t *) flatmap_val(term); + Eterm* values = htop + sizeof(flatmap_t) / sizeof(Eterm); + Uint map_size = fmp->size; + int i; + + erts_print(to, to_arg, "Mf" ETERM_FMT ":", map_size); + dump_element(to, to_arg, fmp->keys); + erts_putc(to, to_arg, ':'); + for (i = 0; i < map_size; i++) { + dump_element(to, to_arg, values[i]); + if (i < map_size-1) { + erts_putc(to, to_arg, ','); + } + } + erts_putc(to, to_arg, '\n'); + } else { + Uint i; + Uint sz = 0; + Eterm* nodes = htop + 1; + + switch (MAP_HEADER_TYPE(w)) { + case MAP_HEADER_TAG_HAMT_HEAD_ARRAY: + nodes++; + sz = 16; + erts_print(to, to_arg, "Mh" ETERM_FMT ":" ETERM_FMT ":", + hashmap_size(term), sz); + break; + case MAP_HEADER_TAG_HAMT_HEAD_BITMAP: + nodes++; + sz = hashmap_bitcount(MAP_HEADER_VAL(w)); + erts_print(to, to_arg, "Mh" ETERM_FMT ":" ETERM_FMT ":", + hashmap_size(term), sz); + break; + case MAP_HEADER_TAG_HAMT_NODE_BITMAP: + sz = hashmap_bitcount(MAP_HEADER_VAL(w)); + erts_print(to, to_arg, "Mn" ETERM_FMT ":", sz); + break; + } + for (i = 0; i < sz; i++) { + dump_element(to, to_arg, nodes[i]); + if (i < sz-1) { + erts_putc(to, to_arg, ','); + } + } + erts_putc(to, to_arg, '\n'); + } + } + size = 1 + header_arity(w); + switch (w & _HEADER_SUBTAG_MASK) { + case MAP_SUBTAG: + if (is_flatmap_header(w)) { + size += 1 + flatmap_get_size(htop); + } else { + size += hashmap_bitcount(MAP_HEADER_VAL(w)); + } + break; + case SUB_BINARY_SUBTAG: + size += 1; + break; + } + break; + default: + ASSERT(!is_header(htop[1])); + erts_print(to, to_arg, PTR_FMT ":l", htop); + dump_element(to, to_arg, htop[0]); + erts_putc(to, to_arg, '|'); + dump_element(to, to_arg, htop[1]); + erts_putc(to, to_arg, '\n'); + size = 2; + break; + } + htop += size; + } +} + void erts_dump_process_state(fmtfn_t to, void *to_arg, erts_aint32_t psflg) { char *s; diff --git a/erts/emulator/beam/erl_term.h b/erts/emulator/beam/erl_term.h index 6daf043117..7d42d1b396 100644 --- a/erts/emulator/beam/erl_term.h +++ b/erts/emulator/beam/erl_term.h @@ -55,9 +55,6 @@ struct erl_node_; /* Declared in erl_node_tables.h */ #if defined(ARCH_64) # define TAG_PTR_MASK__ 0x7 # if !defined(ERTS_HAVE_OS_PHYSICAL_MEMORY_RESERVATION) -# ifdef HIPE -# error Hipe on 64-bit needs a real mmap as it does not support the literal tag -# endif # define TAG_LITERAL_PTR 0x4 # else # undef TAG_LITERAL_PTR @@ -862,7 +859,7 @@ do { \ ((ErtsMRefThing *) (Hp))->mb = (Binp); \ ((ErtsMRefThing *) (Hp))->next = (Ohp)->first; \ (Ohp)->first = (struct erl_off_heap_header*) (Hp); \ - ASSERT(erts_is_ref_numbers_magic(&(Binp)->refn)); \ + ASSERT(erts_is_ref_numbers_magic((Binp)->refn)); \ } while (0) #endif /* ARCH_32 */ diff --git a/erts/emulator/beam/external.c b/erts/emulator/beam/external.c index 970158933f..6666c42778 100644 --- a/erts/emulator/beam/external.c +++ b/erts/emulator/beam/external.c @@ -122,8 +122,6 @@ static int encode_size_struct_int(struct TTBSizeContext_*, ErtsAtomCacheMap *acm static Export binary_to_term_trap_export; static BIF_RETTYPE binary_to_term_trap_1(BIF_ALIST_1); -static BIF_RETTYPE binary_to_term_int(Process* p, Uint32 flags, Eterm bin, Binary* context_b, - Export *bif, Eterm arg0, Eterm arg1); void erts_init_external(void) { erts_init_trap_export(&term_to_binary_trap_export, @@ -1220,7 +1218,8 @@ typedef struct B2TContext_t { ErtsBinary2TermState b2ts; Uint32 flags; SWord reds; - Eterm trap_bin; + Uint used_bytes; /* In: boolean, Out: bytes */ + Eterm trap_bin; /* THE_NON_VALUE if not exported */ Export *bif; Eterm arg[2]; enum B2TState state; @@ -1314,6 +1313,11 @@ binary2term_prepare(ErtsBinary2TermState *state, byte *data, Sint data_size, ctx->u.uc.dbytes = state->extp; ctx->u.uc.dleft = dest_len; + if (ctx->used_bytes) { + ASSERT(ctx->used_bytes == 1); + /* to be subtracted by stream.avail_in when done */ + ctx->used_bytes = data_size; + } ctx->state = B2TUncompressChunk; *ctxp = ctx; } @@ -1416,13 +1420,15 @@ static int b2t_context_destructor(Binary *context_bin) return 1; } +static BIF_RETTYPE binary_to_term_int(Process*, Eterm bin, B2TContext*); + + static BIF_RETTYPE binary_to_term_trap_1(BIF_ALIST_1) { Binary *context_bin = erts_magic_ref2bin(BIF_ARG_1); ASSERT(ERTS_MAGIC_BIN_DESTRUCTOR(context_bin) == b2t_context_destructor); - return binary_to_term_int(BIF_P, 0, THE_NON_VALUE, context_bin, NULL, - THE_NON_VALUE, THE_NON_VALUE); + return binary_to_term_int(BIF_P, THE_NON_VALUE, ERTS_MAGIC_BIN_DATA(context_bin)); } @@ -1448,6 +1454,8 @@ static B2TContext* b2t_export_context(Process* p, B2TContext* src) b2t_context_destructor); B2TContext* ctx = ERTS_MAGIC_BIN_DATA(context_b); Eterm* hp; + + ASSERT(is_non_value(src->trap_bin)); sys_memcpy(ctx, src, sizeof(B2TContext)); if (ctx->state >= B2TDecode && ctx->u.dc.next == &src->u.dc.res) { ctx->u.dc.next = &ctx->u.dc.res; @@ -1457,8 +1465,7 @@ static B2TContext* b2t_export_context(Process* p, B2TContext* src) return ctx; } -static BIF_RETTYPE binary_to_term_int(Process* p, Uint32 flags, Eterm bin, Binary* context_b, - Export *bif_init, Eterm arg0, Eterm arg1) +static BIF_RETTYPE binary_to_term_int(Process* p, Eterm bin, B2TContext *ctx) { BIF_RETTYPE ret_val; #ifdef EXTREME_B2T_TRAPPING @@ -1466,25 +1473,17 @@ static BIF_RETTYPE binary_to_term_int(Process* p, Uint32 flags, Eterm bin, Binar #else SWord initial_reds = (Uint)(ERTS_BIF_REDS_LEFT(p) * B2T_BYTES_PER_REDUCTION); #endif - B2TContext c_buff; - B2TContext *ctx; int is_first_call; - if (context_b == NULL) { + if (is_value(bin)) { /* Setup enough to get started */ is_first_call = 1; - ctx = &c_buff; ctx->state = B2TPrepare; ctx->aligned_alloc = NULL; - ctx->flags = flags; - ctx->bif = bif_init; - ctx->arg[0] = arg0; - ctx->arg[1] = arg1; - IF_DEBUG(ctx->trap_bin = THE_NON_VALUE;) } else { - is_first_call = 0; - ctx = ERTS_MAGIC_BIN_DATA(context_b); + ASSERT(is_value(ctx->trap_bin)); ASSERT(ctx->state != B2TPrepare); + is_first_call = 0; } ctx->reds = initial_reds; @@ -1528,6 +1527,10 @@ static BIF_RETTYPE binary_to_term_int(Process* p, Uint32 flags, Eterm bin, Binar && zret == Z_STREAM_END && ctx->u.uc.dleft == 0) { ctx->reds -= chunk; + if (ctx->used_bytes) { + ASSERT(ctx->used_bytes > 5 + ctx->u.uc.stream.avail_in); + ctx->used_bytes -= ctx->u.uc.stream.avail_in; + } ctx->state = B2TSizeInit; } else { @@ -1546,11 +1549,11 @@ static BIF_RETTYPE binary_to_term_int(Process* p, Uint32 flags, Eterm bin, Binar break; case B2TDecodeInit: - if (ctx == &c_buff && ctx->b2ts.extsize > ctx->reds) { + if (is_non_value(ctx->trap_bin) && ctx->b2ts.extsize > ctx->reds) { /* dec_term will maybe trap, allocate space for magic bin before result term to make it easy to trim with HRelease. */ - ctx = b2t_export_context(p, &c_buff); + ctx = b2t_export_context(p, ctx); } ctx->u.dc.ep = ctx->b2ts.extp; ctx->u.dc.res = (Eterm) (UWord) NULL; @@ -1593,6 +1596,25 @@ static BIF_RETTYPE binary_to_term_int(Process* p, Uint32 flags, Eterm bin, Binar return ret_val; case B2TDone: + if (ctx->used_bytes) { + Eterm *hp; + Eterm used; + if (!ctx->b2ts.exttmp) { + ASSERT(ctx->used_bytes == 1); + ctx->used_bytes = (ctx->u.dc.ep - ctx->b2ts.extp + +1); /* VERSION_MAGIC */ + } + if (IS_USMALL(0, ctx->used_bytes)) { + hp = erts_produce_heap(&ctx->u.dc.factory, 3, 0); + used = make_small(ctx->used_bytes); + } + else { + hp = erts_produce_heap(&ctx->u.dc.factory, 3+BIG_UINT_HEAP_SIZE, 0); + used = uint_to_big(ctx->used_bytes, hp); + hp += BIG_UINT_HEAP_SIZE; + } + ctx->u.dc.res = TUPLE2(hp, ctx->u.dc.res, used); + } b2t_destroy_context(ctx); if (ctx->u.dc.factory.hp > ctx->u.dc.factory.hp_end) { @@ -1613,11 +1635,10 @@ static BIF_RETTYPE binary_to_term_int(Process* p, Uint32 flags, Eterm bin, Binar } }while (ctx->reds > 0 || ctx->state >= B2TDone); - if (ctx == &c_buff) { - ASSERT(ctx->trap_bin == THE_NON_VALUE); - ctx = b2t_export_context(p, &c_buff); + if (is_non_value(ctx->trap_bin)) { + ctx = b2t_export_context(p, ctx); + ASSERT(is_value(ctx->trap_bin)); } - ASSERT(ctx->trap_bin != THE_NON_VALUE); if (is_first_call) { erts_set_gc_state(p, 0); @@ -1634,23 +1655,35 @@ HIPE_WRAPPER_BIF_DISABLE_GC(binary_to_term, 1) BIF_RETTYPE binary_to_term_1(BIF_ALIST_1) { - return binary_to_term_int(BIF_P, 0, BIF_ARG_1, NULL, bif_export[BIF_binary_to_term_1], - BIF_ARG_1, THE_NON_VALUE); + B2TContext ctx; + + ctx.flags = 0; + ctx.used_bytes = 0; + ctx.trap_bin = THE_NON_VALUE; + ctx.bif = bif_export[BIF_binary_to_term_1]; + ctx.arg[0] = BIF_ARG_1; + ctx.arg[1] = THE_NON_VALUE; + return binary_to_term_int(BIF_P, BIF_ARG_1, &ctx); } HIPE_WRAPPER_BIF_DISABLE_GC(binary_to_term, 2) BIF_RETTYPE binary_to_term_2(BIF_ALIST_2) { + B2TContext ctx; Eterm opts; Eterm opt; - Uint32 flags = 0; + ctx.flags = 0; + ctx.used_bytes = 0; opts = BIF_ARG_2; while (is_list(opts)) { opt = CAR(list_val(opts)); if (opt == am_safe) { - flags |= ERTS_DIST_EXT_BTT_SAFE; + ctx.flags |= ERTS_DIST_EXT_BTT_SAFE; + } + else if (opt == am_used) { + ctx.used_bytes = 1; } else { goto error; @@ -1661,8 +1694,11 @@ BIF_RETTYPE binary_to_term_2(BIF_ALIST_2) if (is_not_nil(opts)) goto error; - return binary_to_term_int(BIF_P, flags, BIF_ARG_1, NULL, bif_export[BIF_binary_to_term_2], - BIF_ARG_1, BIF_ARG_2); + ctx.trap_bin = THE_NON_VALUE; + ctx.bif = bif_export[BIF_binary_to_term_2]; + ctx.arg[0] = BIF_ARG_1; + ctx.arg[1] = BIF_ARG_2; + return binary_to_term_int(BIF_P, BIF_ARG_1, &ctx); error: BIF_ERROR(BIF_P, BADARG); @@ -4000,6 +4036,7 @@ dec_term_atom_common: if (ctx) { ctx->state = B2TDone; ctx->reds = reds; + ctx->u.dc.ep = ep; } return ep; diff --git a/erts/emulator/beam/global.h b/erts/emulator/beam/global.h index 09aeba00fa..3dd3a60939 100644 --- a/erts/emulator/beam/global.h +++ b/erts/emulator/beam/global.h @@ -947,6 +947,8 @@ void erts_update_ranges(BeamInstr* code, Uint size); void erts_remove_from_ranges(BeamInstr* code); UWord erts_ranges_sz(void); void erts_lookup_function_info(FunctionInfo* fi, BeamInstr* pc, int full_info); +ErtsLiteralArea** erts_dump_lit_areas; +Uint erts_dump_num_lit_areas; /* break.c */ void init_break_handler(void); diff --git a/erts/emulator/beam/instrs.tab b/erts/emulator/beam/instrs.tab index 20d356a81d..c17d1a8f69 100644 --- a/erts/emulator/beam/instrs.tab +++ b/erts/emulator/beam/instrs.tab @@ -788,7 +788,8 @@ is_eq_exact(Fail, X, Y) { } i_is_eq_exact_literal(Fail, Src, Literal) { - if (!eq($Src, $Literal)) { + Eterm src = $Src; + if (is_immed(src) || !eq(src, $Literal)) { $FAIL($Fail); } } @@ -800,7 +801,8 @@ is_ne_exact(Fail, X, Y) { } i_is_ne_exact_literal(Fail, Src, Literal) { - if (eq($Src, $Literal)) { + Eterm src = $Src; + if (!is_immed(src) && eq(src, $Literal)) { $FAIL($Fail); } } diff --git a/erts/emulator/beam/macros.tab b/erts/emulator/beam/macros.tab index e0b5f56b53..494fe8961e 100644 --- a/erts/emulator/beam/macros.tab +++ b/erts/emulator/beam/macros.tab @@ -20,13 +20,12 @@ // // -// Use if there is a garbage collection before storing to a -// general destination (either X or Y register). +// Define a regular expression that will match instructions that +// perform GC. That will allow beam_makeops to check for instructions +// that don't use $REFRESH_GEN_DEST() when they should. // -REFRESH_GEN_DEST() { - dst_ptr = REG_TARGET_PTR(dst); -} +GC_REGEXP=erts_garbage_collect|erts_gc|GcBifFunction; // $Offset is relative to the start of the instruction (not to the // location of the failure label reference). Since combined diff --git a/erts/emulator/beam/map_instrs.tab b/erts/emulator/beam/map_instrs.tab index bbb2f49b66..c594a87298 100644 --- a/erts/emulator/beam/map_instrs.tab +++ b/erts/emulator/beam/map_instrs.tab @@ -31,7 +31,7 @@ new_map(Dst, Live, N) { Eterm res; HEAVY_SWAPOUT; - res = new_map(c_p, reg, $Live, $N, $NEXT_INSTRUCTION); + res = erts_gc_new_map(c_p, reg, $Live, $N, $NEXT_INSTRUCTION); HEAVY_SWAPIN; $REFRESH_GEN_DEST(); $Dst = res; @@ -44,7 +44,7 @@ i_new_small_map_lit(Dst, Live, Keys) { Eterm keys = $Keys; HEAVY_SWAPOUT; - res = new_small_map_lit(c_p, reg, keys, $Live, $NEXT_INSTRUCTION); + res = erts_gc_new_small_map_lit(c_p, reg, keys, $Live, $NEXT_INSTRUCTION); HEAVY_SWAPIN; $REFRESH_GEN_DEST(); $Dst = res; @@ -133,7 +133,7 @@ update_map_assoc(Src, Dst, Live, N) { reg[live] = $Src; HEAVY_SWAPOUT; - res = update_map_assoc(c_p, reg, live, $N, $NEXT_INSTRUCTION); + res = erts_gc_update_map_assoc(c_p, reg, live, $N, $NEXT_INSTRUCTION); HEAVY_SWAPIN; ASSERT(is_value(res)); $REFRESH_GEN_DEST(); @@ -147,7 +147,7 @@ update_map_exact(Fail, Src, Dst, Live, N) { reg[live] = $Src; HEAVY_SWAPOUT; - res = update_map_exact(c_p, reg, live, $N, $NEXT_INSTRUCTION); + res = erts_gc_update_map_exact(c_p, reg, live, $N, $NEXT_INSTRUCTION); HEAVY_SWAPIN; if (is_value(res)) { $REFRESH_GEN_DEST(); diff --git a/erts/emulator/beam/module.c b/erts/emulator/beam/module.c index baeec115ea..1712dc803c 100644 --- a/erts/emulator/beam/module.c +++ b/erts/emulator/beam/module.c @@ -254,4 +254,3 @@ void module_end_staging(int commit) IF_DEBUG(dbg_load_code_ix = -1); } - diff --git a/erts/emulator/beam/module.h b/erts/emulator/beam/module.h index 9a81e6035b..a3f1ce1705 100644 --- a/erts/emulator/beam/module.h +++ b/erts/emulator/beam/module.h @@ -45,7 +45,7 @@ typedef struct erl_module { int seen; /* Used by finish_loading() */ struct erl_module_instance curr; - struct erl_module_instance old; /* protected by "old_code" rwlock */ + struct erl_module_instance old; /* active protected by "old_code" rwlock */ struct erl_module_instance* on_load; } Module; diff --git a/erts/emulator/beam/ops.tab b/erts/emulator/beam/ops.tab index 7a2c39b3a8..a560bde920 100644 --- a/erts/emulator/beam/ops.tab +++ b/erts/emulator/beam/ops.tab @@ -511,8 +511,6 @@ put_list y x x put_list y y x put_list x y x -put_list y x x - # put_list SrcReg Constant Dst put_list x c x diff --git a/erts/emulator/drivers/common/inet_drv.c b/erts/emulator/drivers/common/inet_drv.c index 7b1f4a0e9c..95d61fcc5d 100644 --- a/erts/emulator/drivers/common/inet_drv.c +++ b/erts/emulator/drivers/common/inet_drv.c @@ -1136,7 +1136,6 @@ static int packet_inet_init(void); static void packet_inet_stop(ErlDrvData); static void packet_inet_command(ErlDrvData, char*, ErlDrvSizeT); static void packet_inet_drv_input(ErlDrvData data, ErlDrvEvent event); -static void packet_inet_drv_output(ErlDrvData data, ErlDrvEvent event); static ErlDrvData udp_inet_start(ErlDrvPort, char* command); #ifdef HAVE_SCTP static ErlDrvData sctp_inet_start(ErlDrvPort, char* command); @@ -1161,7 +1160,7 @@ static struct erl_drv_entry udp_inet_driver_entry = NULL, #else packet_inet_drv_input, - packet_inet_drv_output, + NULL, #endif "udp_inet", NULL, @@ -1196,7 +1195,7 @@ static struct erl_drv_entry sctp_inet_driver_entry = NULL, #else packet_inet_drv_input, - packet_inet_drv_output, + NULL, #endif "sctp_inet", NULL, @@ -1262,7 +1261,6 @@ typedef struct { static int packet_inet_input(udp_descriptor* udesc, HANDLE event); -static int packet_inet_output(udp_descriptor* udesc, HANDLE event); #endif /* convert descriptor pointer to inet_descriptor pointer */ @@ -2199,13 +2197,16 @@ static int inet_reply_ok(inet_descriptor* desc) ErlDrvTermData caller = desc->caller; int i = 0; + desc->caller = 0; + if (is_not_internal_pid(caller)) + return 0; + i = LOAD_ATOM(spec, i, am_inet_reply); i = LOAD_PORT(spec, i, desc->dport); i = LOAD_ATOM(spec, i, am_ok); i = LOAD_TUPLE(spec, i, 3); ASSERT(i == sizeof(spec)/sizeof(*spec)); - desc->caller = 0; return erl_drv_send_term(desc->dport, caller, spec, i); } @@ -11297,24 +11298,20 @@ static ErlDrvSSizeT packet_inet_ctl(ErlDrvData e, unsigned int cmd, char* buf, (desc->sfamily, &remote, &buf, &len)) != NULL) return ctl_xerror(xerror, rbuf, rsize); - sock_select(desc, FD_CONNECT, 1); code = sock_connect(desc->s, &remote.sa, len); if (IS_SOCKET_ERROR(code) && (sock_errno() == EINPROGRESS)) { /* XXX: Unix only -- WinSock would have a different cond! */ - desc->state = INET_STATE_CONNECTING; if (timeout != INET_INFINITY) driver_set_timer(desc->port, timeout); enq_async(desc, tbuf, INET_REQ_CONNECT); + async_ok(desc); } else if (code == 0) { /* OK we are connected */ - sock_select(desc, FD_CONNECT, 0); - desc->state = INET_STATE_CONNECTED; enq_async(desc, tbuf, INET_REQ_CONNECT); async_ok(desc); } else { - sock_select(desc, FD_CONNECT, 0); return ctl_error(sock_errno(), rbuf, rsize); } return ctl_reply(INET_REP_OK, tbuf, 2, rbuf, rsize); @@ -11829,77 +11826,6 @@ static int packet_inet_input(udp_descriptor* udesc, HANDLE event) return count; } -static void packet_inet_drv_output(ErlDrvData e, ErlDrvEvent event) -{ - (void) packet_inet_output((udp_descriptor*)e, (HANDLE)event); -} - -/* UDP/SCTP socket ready for output: -** This is a Back-End for Non-Block SCTP Connect (INET_STATE_CONNECTING) -*/ -static int packet_inet_output(udp_descriptor* udesc, HANDLE event) -{ - inet_descriptor* desc = INETP(udesc); - int ret = 0; - ErlDrvPort ix = desc->port; - - DEBUGF(("packet_inet_output(%ld) {s=%d\r\n", - (long)desc->port, desc->s)); - - if (desc->state == INET_STATE_CONNECTING) { - sock_select(desc, FD_CONNECT, 0); - - driver_cancel_timer(ix); /* posssibly cancel a timer */ -#ifndef __WIN32__ - /* - * XXX This is strange. This *should* work on Windows NT too, - * but doesn't. An bug in Winsock 2.0 for Windows NT? - * - * See "Unix Netwok Programming", W.R.Stevens, p 412 for a - * discussion about Unix portability and non blocking connect. - */ - -#ifndef SO_ERROR - { - int sz = sizeof(desc->remote); - int code = sock_peer(desc->s, - (struct sockaddr*) &desc->remote, &sz); - - if (IS_SOCKET_ERROR(code)) { - desc->state = INET_STATE_OPEN; /* restore state */ - ret = async_error(desc, sock_errno()); - goto done; - } - } -#else - { - int error = 0; /* Has to be initiated, we check it */ - unsigned int sz = sizeof(error); /* even if we get -1 */ - int code = sock_getopt(desc->s, SOL_SOCKET, SO_ERROR, - (void *)&error, &sz); - - if ((code < 0) || error) { - desc->state = INET_STATE_OPEN; /* restore state */ - ret = async_error(desc, error); - goto done; - } - } -#endif /* SO_ERROR */ -#endif /* !__WIN32__ */ - - desc->state = INET_STATE_CONNECTED; - async_ok(desc); - } - else { - sock_select(desc,FD_CONNECT,0); - - DEBUGF(("packet_inet_output(%ld): bad state: %04x\r\n", - (long)desc->port, desc->state)); - } - done: - DEBUGF(("packet_inet_output(%ld) }\r\n", (long)desc->port)); - return ret; -} #endif /*---------------------------------------------------------------------------*/ diff --git a/erts/emulator/hipe/hipe_amd64_bifs.m4 b/erts/emulator/hipe/hipe_amd64_bifs.m4 index b3c9a460bb..cf4c59c9af 100644 --- a/erts/emulator/hipe/hipe_amd64_bifs.m4 +++ b/erts/emulator/hipe/hipe_amd64_bifs.m4 @@ -462,6 +462,7 @@ ASYM($1): TYPE_FUNCTION(ASYM($1)) #endif') + /* * noproc_primop_interface_0(nbif_name, cbif_name) * noproc_primop_interface_1(nbif_name, cbif_name) diff --git a/erts/emulator/hipe/hipe_bif0.tab b/erts/emulator/hipe/hipe_bif0.tab index 4038ca7ef8..0380e8c795 100644 --- a/erts/emulator/hipe/hipe_bif0.tab +++ b/erts/emulator/hipe/hipe_bif0.tab @@ -135,8 +135,8 @@ atom bs_utf16_size atom bs_put_utf16be atom bs_put_utf16le atom bs_get_utf16 -atom bs_validate_unicode atom bs_validate_unicode_retract atom emulate_fpe atom emasculate_binary atom is_divisible +atom is_unicode
\ No newline at end of file diff --git a/erts/emulator/hipe/hipe_bif_list.m4 b/erts/emulator/hipe/hipe_bif_list.m4 index bebe20a18e..625d8486fd 100644 --- a/erts/emulator/hipe/hipe_bif_list.m4 +++ b/erts/emulator/hipe/hipe_bif_list.m4 @@ -222,7 +222,8 @@ standard_bif_interface_2(nbif_rethrow, hipe_rethrow) standard_bif_interface_3(nbif_find_na_or_make_stub, hipe_find_na_or_make_stub) standard_bif_interface_2(nbif_nonclosure_address, hipe_nonclosure_address) nocons_nofail_primop_interface_0(nbif_fclearerror_error, hipe_fclearerror_error) -standard_bif_interface_2(nbif_is_divisible, hipe_is_divisible) +noproc_primop_interface_2(nbif_is_divisible, hipe_is_divisible) +noproc_primop_interface_1(nbif_is_unicode, hipe_is_unicode) /* * Mbox primops with implicit P parameter. @@ -247,7 +248,6 @@ nofail_primop_interface_3(nbif_bs_get_float_2, erts_bs_get_float_2) standard_bif_interface_3(nbif_bs_put_utf8, hipe_bs_put_utf8) standard_bif_interface_3(nbif_bs_put_utf16be, hipe_bs_put_utf16be) standard_bif_interface_3(nbif_bs_put_utf16le, hipe_bs_put_utf16le) -standard_bif_interface_1(nbif_bs_validate_unicode, hipe_bs_validate_unicode) /* * Bit-syntax primops without any P parameter. diff --git a/erts/emulator/hipe/hipe_debug.c b/erts/emulator/hipe/hipe_debug.c index 222a11db3d..cfe60b379e 100644 --- a/erts/emulator/hipe/hipe_debug.c +++ b/erts/emulator/hipe/hipe_debug.c @@ -135,7 +135,9 @@ static void print_heap(Eterm *pos, Eterm *end) printf("From: 0x%0*lx to 0x%0*lx\n\r", 2*(int)sizeof(long), (unsigned long)pos, 2*(int)sizeof(long), (unsigned long)end); - printf(" | H E A P |\r\n"); + printf(" | %*s H E A P %*s |\r\n", + 2*(int)sizeof(long)-1, "", + 2*(int)sizeof(long)-1, ""); printf(" | %*s | %*s |\r\n", 2+2*(int)sizeof(long), "Address", 2+2*(int)sizeof(long), "Contents"); @@ -158,8 +160,10 @@ static void print_heap(Eterm *pos, Eterm *end) ++pos; --ari; } - } else + } else { + fflush(stdout); erts_printf("%.30T", val); + } printf("\r\n"); } printf(" |%s|%s|\r\n", dashes, dashes); diff --git a/erts/emulator/hipe/hipe_mkliterals.c b/erts/emulator/hipe/hipe_mkliterals.c index 6ea120c65c..de00994d64 100644 --- a/erts/emulator/hipe/hipe_mkliterals.c +++ b/erts/emulator/hipe/hipe_mkliterals.c @@ -462,6 +462,15 @@ static const struct rts_param rts_params[] = { 0 #endif }, + /* This flag is always defined, but its value is configuration-dependent. */ + { 17, "ERTS_USE_LITERAL_TAG", + 1, +#if defined(TAG_LITERAL_PTR) + 1 +#else + 0 +#endif + }, /* This parameter is always defined, but its value depends on ERTS_SMP. */ { 19, "MSG_MESSAGE", 1, offsetof(struct erl_mesg, m[0]) diff --git a/erts/emulator/hipe/hipe_native_bif.c b/erts/emulator/hipe/hipe_native_bif.c index 23f64a6991..99c34532b9 100644 --- a/erts/emulator/hipe/hipe_native_bif.c +++ b/erts/emulator/hipe/hipe_native_bif.c @@ -482,15 +482,12 @@ static int validate_unicode(Eterm arg) return 1; } -BIF_RETTYPE nbif_impl_hipe_bs_validate_unicode(NBIF_ALIST_1) +Uint hipe_is_unicode(Eterm arg) { - Process *p = BIF_P; - Eterm arg = BIF_ARG_1; - if (!validate_unicode(arg)) - BIF_ERROR(p, BADARG); - return NIL; + return (Uint) validate_unicode(arg); } + int hipe_bs_validate_unicode_retract(ErlBinMatchBuffer* mb, Eterm arg) { if (!validate_unicode(arg)) { @@ -500,16 +497,12 @@ int hipe_bs_validate_unicode_retract(ErlBinMatchBuffer* mb, Eterm arg) return 1; } -/* Called via standard_bif_interface_2 */ -BIF_RETTYPE nbif_impl_hipe_is_divisible(NBIF_ALIST_2) +Uint hipe_is_divisible(Uint dividend, Uint divisor) { - /* Arguments are Eterm-sized unsigned integers */ - Uint dividend = BIF_ARG_1; - Uint divisor = BIF_ARG_2; if (dividend % divisor) { - BIF_ERROR(BIF_P, BADARG); + return 0; } else { - return NIL; + return 1; } } diff --git a/erts/emulator/hipe/hipe_native_bif.h b/erts/emulator/hipe/hipe_native_bif.h index cbc7ab8dc6..d5081b8438 100644 --- a/erts/emulator/hipe/hipe_native_bif.h +++ b/erts/emulator/hipe/hipe_native_bif.h @@ -66,9 +66,9 @@ AEXTERN(Eterm,nbif_bs_utf16_size,(Eterm)); AEXTERN(Eterm,nbif_bs_put_utf16be,(Process*,Eterm,byte*,unsigned int)); AEXTERN(Eterm,nbif_bs_put_utf16le,(Process*,Eterm,byte*,unsigned int)); AEXTERN(Eterm,nbif_bs_get_utf16,(void)); -AEXTERN(Eterm,nbif_bs_validate_unicode,(Process*,Eterm)); +AEXTERN(Uint,nbif_is_unicode,(Eterm)); AEXTERN(Eterm,nbif_bs_validate_unicode_retract,(void)); -AEXTERN(void,nbif_is_divisible,(Process*,Uint,Uint)); +AEXTERN(Uint,nbif_is_divisible,(Uint,Uint)); AEXTERN(void,nbif_select_msg,(Process*)); AEXTERN(Eterm,nbif_cmp_2,(void)); @@ -91,10 +91,10 @@ BIF_RETTYPE nbif_impl_hipe_bs_put_utf8(NBIF_ALIST_3); Eterm hipe_bs_utf16_size(Eterm); BIF_RETTYPE nbif_impl_hipe_bs_put_utf16be(NBIF_ALIST_3); BIF_RETTYPE nbif_impl_hipe_bs_put_utf16le(NBIF_ALIST_3); -BIF_RETTYPE nbif_impl_hipe_bs_validate_unicode(NBIF_ALIST_1); +Uint hipe_is_unicode(Eterm); struct erl_bin_match_buffer; int hipe_bs_validate_unicode_retract(struct erl_bin_match_buffer*, Eterm); -BIF_RETTYPE nbif_impl_hipe_is_divisible(NBIF_ALIST_2); +Uint hipe_is_divisible(Uint, Uint); #ifdef NO_FPE_SIGNALS AEXTERN(void,nbif_emulate_fpe,(Process*)); diff --git a/erts/emulator/hipe/hipe_primops.h b/erts/emulator/hipe/hipe_primops.h index 6aac5e6205..a6abd3e011 100644 --- a/erts/emulator/hipe/hipe_primops.h +++ b/erts/emulator/hipe/hipe_primops.h @@ -63,7 +63,7 @@ PRIMOP_LIST(am_bs_utf16_size, &nbif_bs_utf16_size) PRIMOP_LIST(am_bs_put_utf16be, &nbif_bs_put_utf16be) PRIMOP_LIST(am_bs_put_utf16le, &nbif_bs_put_utf16le) PRIMOP_LIST(am_bs_get_utf16, &nbif_bs_get_utf16) -PRIMOP_LIST(am_bs_validate_unicode, &nbif_bs_validate_unicode) +PRIMOP_LIST(am_is_unicode, &nbif_is_unicode) PRIMOP_LIST(am_bs_validate_unicode_retract, &nbif_bs_validate_unicode_retract) PRIMOP_LIST(am_is_divisible, &nbif_is_divisible) diff --git a/erts/emulator/internal_doc/GarbageCollection.md b/erts/emulator/internal_doc/GarbageCollection.md new file mode 100644 index 0000000000..1d9e3f4160 --- /dev/null +++ b/erts/emulator/internal_doc/GarbageCollection.md @@ -0,0 +1,187 @@ +# Erlang Garbage Collector + +Erlang manages dynamic memory with a [tracing garbage collector](https://en.wikipedia.org/wiki/Tracing_garbage_collection). More precisely a per process generational semi-space copying collector using [Cheney's](#cheney) copy collection algorithm together with a global large object space. + +## Overview + +Each Erlang process has its own stack and heap which are allocated in the same memory block and grow towards each other. When the stack and the heap [meet](https://github.com/erlang/otp/blob/OTP-18.0/erts/emulator/beam/beam_emu.c#L387), the garbage collector is triggered and memory is reclaimed. If not enough memory was reclaimed, the heap will grow. + +### Creating Data + +Terms are created on the heap by evaluating expressions. There are two major types of terms: [immediate terms](https://github.com/erlang/otp/blob/OTP-18.0/erts/emulator/beam/erl_term.h#L88-L97) which require no heap space (small integers, atoms, pids, port ids etc) and cons or [boxed terms](https://github.com/erlang/otp/blob/OTP-18.0/erts/emulator/beam/erl_term.h#L106-L120) (tuple, big num, binaries etc) that do require heap space. Immediate terms do not need any heap space because they are embedded into the containing structure. + +Let's look at an example that returns a tuple with the newly created data. + +```erlang +data(Foo) -> + Cons = [42|Foo], + Literal = {text, "hello world!"}, + {tag, Cons, Literal}. +``` + +In this example we first create a new cons cell with an integer and a tuple with some text. Then a tuple of size three wrapping the other values with an atom tag is created and returned. + +On the heap tuples require a word size for each of its elements as well as for the header. Cons cells always require two words. Adding these things together, we get seven words for the tuples and 26 words for the cons cells. The string `"hello world!"` is a list of cons cells and thus requires 24 words. The atom `tag` and the integer `42` do not require any additional heap memory since it is an *immediate*. Adding all the terms together, the heap space required in this example should be 33 words. + +Compiling this code to beam assembly (`erlc -S`) shows exactly what is happening. + +```erlang + ... + {test_heap,6,1}. + {put_list,{integer,42},{x,0},{x,1}}. + {put_tuple,3,{x,0}}. + {put,{atom,tag}}. + {put,{x,1}}. + {put,{literal,{text,"hello world!"}}}. + return. +``` + +Looking at the assembler code we can see three things; The heap requirement in this function turns out to be only six words, as seen by the `{test_heap,6,1}` instruction. All the allocations are combined to a single instruction. The bulk of the data `{text, "hello world!"}` is a *literal*. Literals, sometimes referred to as constants, are not allocated in the function since they are a part of the module and allocated at load time. + +If there is not enough space available on the heap to satisfy the `test_heap` instructions request for memory, then a garbage collection is initiated. It may happen immediately in the `test_heap` instruction, or it can be delayed until a later time depending on what state the process is in. If the garbage collection is delayed, any memory needed will be allocated in heap fragments. Heap fragments are extra memory blocks that are a part of the young heap, but are not allocated in the contigious area where terms normally reside. See [The young heap](#the-young-heap) for more details. + +### The collector + +Erlang has a copying semi-space garbage collector. This means that when doing a garbage collection, the terms are copied from one distinct area, called the *from space*, to a new clean area, called the *to space*. The collector starts by [scanning the root-set](https://github.com/erlang/otp/blob/OTP-18.0/erts/emulator/beam/erl_gc.c#L1980) (stack, registers, etc). + + + +It follows all the pointers from the root-set to the heap and copies each term word by word to the *to space*. + +After the header word has been copied a [*move marker*](https://github.com/erlang/otp/blob/OTP-18.0/erts/emulator/beam/erl_gc.h#L45-L46) is destructively placed in it pointing to the term in the *to space*. Any other term that points to the already moved term will [see this move marker](https://github.com/erlang/otp/blob/OTP-18.0/erts/emulator/beam/erl_gc.c#L1125) and copy the referring pointer instead. For example, if the have the following Erlang code: + +```erlang +foo(Arg) -> + T = {test, Arg}, + {wrapper, T, T, T}. +``` + +Only one copy of T exists on the heap and during the garbage collection only the first time T is encountered will it be copied. + + + +After [all terms](https://github.com/erlang/otp/blob/OTP-18.0/erts/emulator/beam/erl_gc.c#L1089) referenced by the root-set have been copied, the collector scans the *to space* and copies all terms that these terms reference. When scanning, the collector steps through each term on the *to space* and any term still referencing the *from space* is copied over to the *to space*. Some terms contain non-term data (the payload of a on heap binary for instance). When encountered by the collector, these values are simply skipped. + + + +Every term object we can reach is copied to the *to space* and stored on top off the *scan stop* line, and then the scan stop is moved to the end of the last object. + + + +When *scan stop* marker [catches up](https://github.com/erlang/otp/blob/OTP-18.0/erts/emulator/beam/erl_gc.c#L1103) to the *scan start* marker, the garbage collection is done. At this point we can [deallocate](https://github.com/erlang/otp/blob/OTP-18.0/erts/emulator/beam/erl_gc.c#L1206) the entire *from space* and therefore reclaim the entire young heap. + +## Generational Garbage Collection + +In addition to the collection algorithm described above, the Erlang garbage collector also provides generational garbage collection. An additional heap, called the old heap, is used where the long lived data is stored. The original heap is called the young heap, or sometimes the allocation heap. + +With this in mind we can look at the Erlang's garbage collection again. During the copy stage anything that should be copied to the young *to space* is instead copied to the old *to space* *if* it is [below the *high-watermark*](https://github.com/erlang/otp/blob/OTP-18.0/erts/emulator/beam/erl_gc.c#L1127). + + + +The [*high-watermark*](https://github.com/erlang/otp/blob/OTP-18.0/erts/emulator/beam/erl_process.h#L1021) is placed where the previous garbage collection (described in [Overview](#overview)) ended and we have introduced a new area called the old heap. When doing the normal garbage collection pass, any term that is located below the high-watermark is copied to the old *to space* instead of the young. + + + +In the next garbage collection, any pointers to the old heap will be ignored and not scanned. This way the garbage collector does not have to scan the long-lived terms. + +Generational garbage collection aims to increase performance at the expense of memory. This is achieved because only the young, smaller, heap is considered in most garbage collections. + +The generational [hypothesis](#ungar) predicts that most terms tend to die young, and for an immutable language such as Erlang, young terms die even faster than in other languages. So for most usage patterns the data in the new heap will die very soon after it is allocated. This is good because it limits the amount of data copied to the old heap and also because the garbage collection algorithm used is proportional to the amount of live data on the heap. + +One critical issue to note here is that any term on the young heap can reference terms on the old heap but *no* term on the old heap may refer to a term on the young heap. This is due to the nature of the copy algorithm. Anything referenced by an old heap term is not included in the reference tree, root-set and its followers, and hence is not copied. If it was, the data would be lost, fire and brimstone would rise to cover the earth. Fortunately, this comes naturally for Erlang because the terms are immutable and thus there can be no pointers modified on the old heap to point to the young heap. + +To reclaim data from the old heap, both young and old heaps are included during the collection and copied to a common *to space*. Both the *from space* of the young and old heap are then deallocated and the procedure will start over from the beginning. This type of garbage collection is called a full sweep and is triggered when the size of the area under the high-watermark is larger than the size of the free area of the old heap. It can also be triggered by doing a manual call to [erlang:garbage_collect()](http://erlang.org/doc/man/erlang.html#garbage_collect-0), or by running into the young garbage collection limit set by [spawn_opt(fun(),[{fullsweep_after, N}])](http://erlang.org/doc/man/erlang.html#spawn_opt-4) where N is the number of young garbage collections to do before forcing a garbage collection of both young and old heap. + +## The young heap + +The young heap, or the allocation heap, consists of the stack and heap as described in the Overview. However, it also includes any heap fragments that are attached to the heap. All of the heap fragments are considered to be above the high-watermark and part of the young generation. Heap fragments contain terms that either did not fit on the heap, or were created by another process and then attached to the heap. For instance if the bif binary_to_term created a term which does not fit on the current heap without doing a garbage collection, it will create a heap-fragment for the term and then schedule a garbage collection for later. Also if a message is sent to the process, the payload may be placed in a heap-fragment and that fragment is added to young heap when the message is matched in a receive clause. + +This procedure differs from how it worked prior to Erlang/OTP 19.0. Before 19.0, only a contiguous memory block where the young heap and stack resided was considered to be part of the young heap. Heap fragments and messages were immediately copied into the young heap before they could be inspected by the Erlang program. The behaviour introduced in 19.0 is superior in many ways - most significantly it reduces the number of necessary copy operations and the root set for garbage collection. + +## Sizing the heap + +As mentioned in the Overview the size of the heap [grows](https://github.com/erlang/otp/blob/OTP-18.0/erts/emulator/beam/erl_gc.c#L247) to accommodate more data. Heaps grow in two stages, first a [variation of the Fibonacci sequence](https://github.com/erlang/otp/blob/OTP-18.0/erts/emulator/beam/erl_gc.c#L199-L208) is used starting at 233 words. Then at about 1 mega words the heap only [grows in 20% increments](https://github.com/erlang/otp/blob/OTP-18.0/erts/emulator/beam/erl_gc.c#L215-L227). + +There are two occasions when the young heap grows: + +* if the total size of the heap + message and heap fragments exceeds the current heap size. +* if after a fullsweep, the total amount of live objects is greater than 75%. + +There are two occasions when the young heap is shrunk: + +* if after a young collection, the total amount of live objects is less than 25% of the heap and the young heap is "big" +* if after a fullsweep, the total amount of live objects is less than 25% of the heap. + +The old heap is always one step ahead in the heap growth stages than the young heap. + +## Literals + +When garbage collecting a heap (young or old) all literals are left in place and not copied. To figure out if a term should be copied or not when doing a garbage collection the following pseudo code is used: + +```c +if (erts_is_literal(ptr) || (on_old_heap(ptr) && !fullsweep)) { + /* literal or non fullsweep - do not copy */ +} else { + copy(ptr); +} +``` + +The [`erts_is_literal`](https://github.com/erlang/otp/blob/OTP-19.0/erts/emulator/beam/global.h#L1452-L1465) check works differently on different architectures and operating systems. + +On 64 bit systems that allow mapping of unreserved virtual memory areas (most operating systems except Windows), an area of size 1 GB (by default) is mapped and then all literals are placed within that area. Then all that has to be done to determine if something is a literal or not is [two quick pointer checks](https://github.com/erlang/otp/blob/OTP-19.0/erts/emulator/beam/erl_alloc.h#L322-L324). This system relies on the fact that a memory page that has not been touched yet does not take any actual space. So even if 1 GB of virtual memory is mapped, only the memory which is actually needed for literals is allocated in ram. The size of the literal area is configurable through the +MIscs erts_alloc option. + +On 32 bit systems, there is not enough virtual memory space to allocate 1 GB for just literals, so instead small 256 KB sized literal regions are created on demand and a card mark bit-array of the entire 32 bit memory space is then used to determine if a term is a literal or not. Since the total memory space is only 32 bits, the card mark bit-array is only 256 words large. On a 64 bit system the same bit-array would have to be 1 tera words large, so this technique is only viable on 32 bit systems. Doing [lookups in the array](https://github.com/erlang/otp/blob/OTP-19.0/erts/emulator/beam/erl_alloc.h#L316-L319) is a little more expensive then just doing the pointer checks that can be done in 64 bit systems, but not extremely so. + +On 64 bit windows, on which erts_alloc cannot do unreserved virtual memory mappings, a [special tag](https://github.com/erlang/otp/blob/OTP-19.0/erts/emulator/beam/erl_term.h#L59) within the Erlang term object is used to determine if something [is a literal or not](https://github.com/erlang/otp/blob/OTP-19.0/erts/emulator/beam/erl_term.h#L248-L252). This is very cheap, however, the tag is only available on 64 bit machines, and it is possible to do a great deal of other nice optimizations with this tag in the future (like for instance a more compact list implementation) so it is not used on operating systems where it is not needed. + +This behaviour is different from how it worked prior to Erlang/OTP 19.0. Before 19.0 the literal check was done by checking if the pointer pointed to the young or old heap block. If it did not, then it was considered a literal. This lead to considerable overhead and strange memory usage scenarios, so it was removed in 19.0. + +## Binary heap + +The binary heap works as a large object space for binary terms that are greater than 64 bytes (from now on called off-heap binaries). The binary heap is [reference counted](https://en.wikipedia.org/wiki/Reference_counting) and a pointer to the off-heap binary is stored on the process heap. To keep track of when to decrement the reference counter of the off-heap binary, a linked list (the MSO - mark and sweep object list) containing funs and externals as well as off-heap binaries is woven through the heap. After a garbage collection is done, the [MSO list is swept](https://github.com/erlang/otp/blob/OTP-18.0/erts/emulator/beam/erl_gc.c#L2299) and any off-heap binary that does not have a [move marker](https://github.com/erlang/otp/blob/OTP-18.0/erts/emulator/beam/erl_gc.c#L2325) written into the header words has its reference [decremented and is potentially freed](https://github.com/erlang/otp/blob/OTP-18.0/erts/emulator/beam/erl_gc.c#L2344-L2367). + +All items in the MSO list are ordered by the time they were added to the process heap, so when doing a minor garbage collection, the MSO sweeper only has to sweep until it [encounters an off-heap binary that is on the old heap](https://github.com/erlang/otp/blob/OTP-18.0/erts/emulator/beam/erl_gc.c#L2369). + +### Virtual Binary heap + +Each process has a virtual binary heap associated with it that has the size of all the current off-heap binaries that the process has references to. The virtual binary heap also has a limit and grows and shrinks depending on how off-heap binaries are used by the process. The same growth and shrink mechanisms are used for the binary heap and for the term heap, so first a Fibonacci like series and then 20% growth. + +The virtual binary heap exists in order to [trigger](https://github.com/erlang/otp/blob/OTP-18.0/erts/emulator/beam/beam_emu.c#L364) garbage collections earlier when potentially there is a very large amount of off-heap binary data that could be reclaimed. This approach does not catch all problems with binary memory not being released soon enough, but it does catch a lot of them. + +## Messages + +Messages can become a part of the process heap at different times. This depends on how the process is configured. +We can configure the behaviour of each process using `process_flag(message_queue_data, off_heap | on_heap)` or we can set a default for all processes at start using the option `+hmqd`. + +What do these different configurations do and when should we use them? +Let's start by going through what happens when one Erlang process sends a message to another. +The sending process needs to do a couple of things: + +1. calculate [how large](https://github.com/erlang/otp/blob/OTP-18.0/erts/emulator/beam/erl_message.c#L1031) the message to be sent is +2. [allocate enough space](https://github.com/erlang/otp/blob/OTP-18.0/erts/emulator/beam/erl_message.c#L1033) to fit the entire message +3. [copy](https://github.com/erlang/otp/blob/OTP-18.0/erts/emulator/beam/erl_message.c#L1040) the message payload +4. allocate a message container with some meta data +5. [insert](https://github.com/erlang/otp/blob/OTP-18.0/erts/emulator/beam/erl_message.c#L502) the message container in the receiver process' [message queue](https://github.com/erlang/otp/blob/OTP-18.0/erts/emulator/beam/erl_process.h#L1042) + +The process flag `message_queue_data`, of the receiver process, controls the message allocating strategy of the sender process in step 2 and also how the message data is treated by the garbage collector. + +The procedure above is different from how it worked prior to 19.0. Before 19.0 there was no configuration option, the behaviour was always very similar to how the `on_heap` option is in 19.0. + +### Message allocating strategies + +If set to `on_heap`, the sending process will first attempt to allocate the space for the message directly on the young heap block of the receiving process. +This is not always possible as it requires taking the *main lock* of the receiving process. The main lock is also held when the process is executing. The possibility for a lock conflict is thus likely in an intensely collaborating system. +If the sending process cannot acquire the main lock, a heap fragment is instead created for the message and the message payload is copied onto that. +With the `off_heap` option the sender process always creates heap fragments for messages sent to that process. + +There are a bunch of different tradeoffs that come into play when trying to figure out which of the strategies you want to use. + +Using `off_heap` may seem like a nice way to get a more scalable system as you get very little contention on the main locks, however, allocating a heap fragment is more expensive than allocating on the heap of the receiving process. So if it is very unlikely that contention will occur, it is more efficient to try to allocate the message directly on the receiving process' heap. + +Using `on_heap` will force all messages to be part of on the young heap which will increase the amount of data that the garbage collector has to move. So if a garbage collection is triggered while processing a large amount of messages, they will be copied to the young heap. This in turn will lead to that the messages will quickly be promoted to the old heap and thus increase its size. This may be good or bad depending on exactly what the process does. A large old heap means that the young heap will also be larger, which in turn means that less garbage collections will be triggered while processing the message queue. This will temporarly increase the throughput of the process at the cost of more memory usage. However, if after all the messages have been consumed the process enters a state where a lot less messages are being received. Then it may be a long time before the next fullsweep garbage collection happens and the messages that are on the old heap will be there until that happens. So while `on_heap` is potentially faster than the other modes, it uses more memory for a longer time. This mode is the legacy mode which is almost how the message queue was handled before Erlang/OTP 19.0. + +Which one of these strategies is best depends a lot on what the process is doing and how it interacts with other processes. So, as always, profile the application and see how it behaves with the different options. + + <a name="cheney">[1]</a>: C. J. Cheney. A nonrecursive list compacting algorithm. Commun. ACM, 13(11):677–678, Nov. 1970. + + <a name="ungar">[2]</a>: D. Ungar. Generation scavenging: A non-disruptive high performance storage reclamation algorithm. SIGSOFT Softw. Eng. Notes, 9(3):157–167, Apr. 1984. diff --git a/erts/emulator/internal_doc/beam_makeops.md b/erts/emulator/internal_doc/beam_makeops.md new file mode 100644 index 0000000000..1da8d2ab05 --- /dev/null +++ b/erts/emulator/internal_doc/beam_makeops.md @@ -0,0 +1,1846 @@ +The beam\_makeops script +======================= + +This document describes the **beam\_makeops** script. + +Introduction +------------ + +The **beam\_makeops** Perl script is used at build-time by both the +compiler and runtime system. Given a number of input files (all with +the extension `.tab`), it will generate source files used by the +Erlang compiler and by the runtime system to load and execute BEAM +instructions. + +Essentially those `.tab` files define: + +* External generic BEAM instructions. They are the instructions that +are known to both the compiler and the runtime system. Generic +instructions are stable between releases. New generic instructions +with high numbers than previous instructions can be added in major +releases. The OTP 20 release has 159 external generic instructions. + +* Internal generic instructions. They are known only to the runtime +system and can be changed at any time without compatibility issues. +They are created by transformation rules (described next). + +* Rules for transforming one or more generic instructions to other +generic instructions. The transformation rules allow combining, +splitting, and removal of instructions, as well as shuffling operands. +Because of the transformation rules, the runtime can have many +internal generic instructions that are only known to runtime system. + +* Specific BEAM instructions. The specific instructions are the +instructions that are actually executed by the runtime system. They +can be changed at any time without causing compatibility issues. +The loader translates generic instructions to specific instructions. +In general, for each generic instruction, there exists a family of +specific instructions. The OTP 20 release has 389 specific +instructions. + +* The implementation of specific instructions. + +Generic instructions have typed operands. Here are a few examples of +operands for `move/2`: + + {move,{atom,id},{x,5}}. + {move,{x,3},{x,0}}. + {move,{x,2},{y,1}}. + +When those instructions are loaded, the loader rewrites them +to specific instructions: + + move_cx id 5 + move_xx 3 0 + move_xy 2 1 + +Corresponding to each generic instruction, there is a family of +specific instructions. The types that an instance of a specific +instruction can handle are encoded in the instruction names. For +example, `move_xy` takes an X register number as the first operand and +a Y register number as the second operand. `move_cx` takes a tagged +Erlang term as the first operand and an X register number as the +second operand. + +An example: the move instruction +-------------------------------- + +Using the `move` instruction as an example, we will give a quick +tour to show the main features of **beam\_makeops**. + +In the `compiler` application, in the file `genop.tab`, there is the +following line: + + 64: move/2 + +This is a definition of an external generic BEAM instruction. Most +importantly it specifices that the opcode is 64. It also defines that +it has two operands. The BEAM assembler will use the opcode when +creating `.beam` files. The compiler does not really need the arity, +but it will use it as an internal sanity check when assembling the +BEAM code. + +Let's have a look at `ops.tab` in `erts/emulator/beam`, where the +specific `move` instructions are defined. Here are a few of them: + + move x x + move x y + move c x + +Each specific instructions is defined by following the name of the +instruction with the types for each operand. An operand type is a +single letter. For example, `x` means an X register, `y` +means a Y register, and `c` is a "constant" (a tagged term such as +an integer, an atom, or a literal). + +Now let's look at the implementation of the `move` instruction. There +are multiple files containing implementations of instructions in the +`erts/emulator/beam` directory. The `move` instruction is defined in +`instrs.tab`. It looks like this: + + move(Src, Dst) { + $Dst = $Src; + } + +The implementation for an instruction largely follows the C syntax, +except that the variables in the function head don't have any types. +The `$` before an identifier denotes a macro expansion. Thus, +`$Src` will expand to the code to pick up the source operand for +the instruction and `$Dst` to the code for the destination register. + +We will look at the code for each specific instruction in turn. To +make the code easier to understand, let's first look at the memory +layout for the instruction `{move,{atom,id},{x,5}}`: + + +--------------------+--------------------+ + I -> | 40 | &&lb_move_cx | + +--------------------+--------------------+ + | Tagged atom 'id' | + +--------------------+--------------------+ + +This example and all other examples in the document assumes a 64-bit +archictecture, and furthermore that pointers to C code fit in 32 bits. + +`I` in the BEAM virtual machine is the instruction pointer. When BEAM +executes an instruction, `I` points to the first word of the +instruction. + +`&&lb_move_cx` is the address to C code that implements `move_cx`. It +is stored in the lower 32 bits of the word. In the upper 32 bits is +the byte offset to the X register; the register number 5 has been +multiplied by the word size size 8. + +In the next word the tagged atom `id` is stored. + +With that background, we can look at the generated code for `move_cx` +in `beam_hot.h`: + + OpCase(move_cx): + { + BeamInstr next_pf = BeamCodeAddr(I[2]); + xb(BeamExtraData(I[0])) = I[1]; + I += 2; + ASSERT(VALID_INSTR(next_pf)); + GotoPF(next_pf); + } + +We will go through each line in turn. + +* `OpCase(move_cx):` defines a label for the instruction. The +`OpCase()` macro is defined in `beam_emu.c`. It will expand this line +to `lb_move_cx:`. + +* `BeamInstr next_pf = BeamCodeAddr(I[2]);` fetches the pointer to +code for the next instruction to be executed. The `BeamCodeAddr()` +macro extracts the pointer from the lower 32 bits of the instruction +word. + +* `xb(BeamExtraData(I[0])) = I[1];` is the expansion of `$Dst = $Src`. +`BeamExtraData()` is a macro that will extract the upper 32 bits from +the instruction word. In this example, it will return 40 which is the +byte offset for X register 5. The `xb()` macro will cast a byte +pointer to an `Eterm` pointer and dereference it. The `I[1]` on +the right side of the `=` fetches an Erlang term (the atom `id` in +this case). + +* `I += 2` advances the instruction pointer to the next +instruction. + +* In a debug-compiled emulator, `ASSERT(VALID_INSTR(next_pf));` makes +sure that `next_pf` is a valid instruction (that is, that it points +within the `process_main()` function in `beam_emu.c`). + +* `GotoPF(next_pf);` transfers control to the next instruction. + +Now let's look at the implementation of `move_xx`: + + OpCase(move_xx): + { + Eterm tmp_packed1 = BeamExtraData(I[0]); + BeamInstr next_pf = BeamCodeAddr(I[1]); + xb((tmp_packed1>>BEAM_TIGHT_SHIFT)) = xb(tmp_packed1&BEAM_TIGHT_MASK); + I += 1; + ASSERT(VALID_INSTR(next_pf)); + GotoPF(next_pf); + } + +We will go through the lines that are new or have changed compared to +`move_cx`. + +* `Eterm tmp_packed1 = BeamExtraData(I[0]);` picks up both X register +numbers packed into the upper 32 bits of the instruction word. + +* `BeamInstr next_pf = BeamCodeAddr(I[1]);` pre-fetches the address of +the next instruction. Note that because both X registers operands fits +into the instruction word, the next instruction is in the very next +word. + +* `xb((tmp_packed1>>BEAM_TIGHT_SHIFT)) = xb(tmp_packed1&BEAM_TIGHT_MASK);` +copies the source to the destination. (For a 64-bit architecture, +`BEAM_TIGHT_SHIFT` is 16 and `BEAM_TIGHT_MASK` is `0xFFFF`.) + +* `I += 1;` advances the instruction pointer to the next instruction. + +`move_xy` is almost identical to `move_xx`. The only difference is +the use of the `yb()` macro instead of `xb()` to reference the +destination register: + + OpCase(move_xy): + { + Eterm tmp_packed1 = BeamExtraData(I[0]); + BeamInstr next_pf = BeamCodeAddr(I[1]); + yb((tmp_packed1>>BEAM_TIGHT_SHIFT)) = xb(tmp_packed1&BEAM_TIGHT_MASK); + I += 1; + ASSERT(VALID_INSTR(next_pf)); + GotoPF(next_pf); + } + +### Transformation rules ### + +Next let's look at how we can do some optimizations using transformation +rules. For simple instructions such as `move/2`, the instruction dispatch +overhead can be substantial. A simple optimization is to combine common +instructions sequences to a single instruction. One such common sequence +is multiple `move` instructions moving X registers to Y registers. + +Using the following rule we can combine two `move` instructions +to a `move2` instruction: + + move X1=x Y1=y | move X2=x Y2=y => move2 X1 Y1 X2 Y2 + +The left side of the arrow (`=>`) is a pattern. If the pattern +matches, the matching instructions will be replaced by the +instructions on the right side. Variables in a pattern must start +with an uppercase letter just as in Erlang. A pattern variable may be +followed `=` and one or more type letters to constrain the match to +one of those types. The variables that are bound on the left side can +be used on the right side. + +We will also need to define a specific instruction and an implementation: + + # In ops.tab + move2 x y x y + + // In instrs.tab + move2(S1, D1, S2, D2) { + Eterm V1, V2; + V1 = $S1; + V2 = $S2; + $D1 = V1; + $D2 = V2; + } + +When the loader has found a match and replaced the matched instructions, +it will match the new instructions against the transformation rules. +Because of that, we can define the rule for a `move3/6` instruction +as follows: + + move2 X1=x Y1=y X2=x Y2=y | move X3=x Y3=y => \ + move3 X1 Y1 X2 Y2 X3 Y3 + +(A `\` before a newline can be used to break a long line for readability.) + +It would also be possible to define it like this: + + move X1=x Y1=y | move X2=x Y2=y | move X3=x Y3=y => \ + move3 X1 Y1 X2 Y2 X3 Y3 + +but in that case it must be defined before the rule for `move2/4` +because the first matching rule will be applied. + +One must be careful not to create infinite loops. For example, if we +for some reason would want to reverse the operand order for the `move` +instruction, we must not do like this: + + move Src Dst => move Dst Src + +The loader would swap the operands forever. To avoid the loop, we must +rename the instruction. For example: + + move Src Dst => assign Dst Src + +This concludes the quick tour of the features of **beam\_makeops**. + +Short overview of instruction loading +------------------------------------- + +To give some background to the rest of this document, here follows a +quick overview of how instructions are loaded. + +* The loader reads and decodes one instruction at a time from the BEAM +code and creates a generic instruction. Many transformation rules +must look at multiple instructions, so the loader will +keep multiple generic instructions in a linked list. + +* The loader tries to apply transformation rules against the +generic instructions in the linked list. If a rule matches, the +matched instructions will be removed and replaced with new +generic instructions constructed from the right side of the +transformation. + +* If a transformation rule matched, the loader applies the +transformation rules again. + +* If no transformation rule match, the loader will begin rewriting +the first of generic instructions to a specific instruction. + +* First the loader will search for a specific operation where the +types for all operands match the type for the generic instruction. +The first matching instruction will be selected. **beam\_makeops** +has ordered the specific instructions so that instructions with more +specific operands comes before instructions with less specific +operands. For example, `move_nx` is more specific than `move_cx`. If +the first operand is `[]` (NIL), `move_nx` will be selected. + +* Given the opcode for the selected specific instruction, the loader +looks up the pointer to the C code for the instruction and stores +in the code area for the module being loaded. + +* The loader translates each operand to a machine word and stores it +in the code area. The operand type for the selected specific +instruction guides the translation. For example, if the type is `e`, +the value of the operand is an index into an arry of external +functions and will be translated to a pointer to the export entry for +the function to call. If the type is `x`, the number of the X +register will be multiplied by the word size to produce a byte offset. + +* The loader runs the packing engine to pack multiple operands into a +single word. The packing engine is controlled by a small program, +which is a string where each character is an instruction. For +example, the code to pack the operands for `move_xy` is `"22#"` (on a +64-bit machine). That program will pack the byte offsets for both +registers into the same word as the pointer to C code. + +Running beam_makeops +-------------------- + +**beam\_makeops** is found in `$ERL_TOP/erts/emulator/utils`. Options +start with a hyphen (`-`). The options are followed by the name of +the input files. By convention, all input files have the extension +`.tab`, but is not enforced by **beam\_makeops**. + +### The -outdir option ### + +The option `-outdir Directory` specifies the output directory for +the generated files. Default is the current working directory. + +### Running beam_makeops for the compiler ### + +Give the option `-compiler` to produce output files for the compiler. +The following files will be written to the output directory: + +* `beam_opcodes.erl` - Used primarily by `beam_asm` and `beam_diasm`. + +* `beam_opcode.hrl` - Used by `beam_asm`. It contains tag definitions +used for encoding instruction operands. + +The input file should only contain the definition of BEAM_FORMAT_NUMBER +and external generic instructions. (Everything else would be ignored.) + +### Running beam_makeops for the emulator ### + +Give the option `-emulator` to produce output files for the emulator. +The following output files will be generated in the output directory. + +* `beam_hot.h`, `beam_warm.h`, `beam_cold.`h - Implementation of +instructions. Included inside the `process_main()` function in +`beam_emu.c`. + +* `beam_opcodes.c` - Defines static data used by the loader +(`beam_load.c`). Data about generic instructions, specific +instructions (including how to pack their operands), and +transformation rules are all part of this file. + +* `beam_opcodes.h` - Miscellanous preprocessor definitions, mainly +used by `beam_load.c` but also by `beam_{hot,warm,cold}.h`. + +* `beam_pred_funcs.h` - Included by `beam_load.c`. Contains defines +needed to call guard constraints in transformation rules. + +* `beam_tr_funcs.h` - Included by `beam_load.c`. Contains defines +needed to call a C function to the right of a transformation rule. + +The following options can be given: + +* `wordsize 32|64` - Defines the word size. Default is 32. + +* `code-model Model` - The code model as given to `-mcmodel` option +for GCC. Default is `unknown`. If the code model is `small` (and +the word size is 64 bits), **beam\_makeops** will pack operands +into the upper 32 bits of the instruction word. + +* `DSymbol=0|1` - Defines the value for a symbol. The symbol can be +used in `%if` and `%unless` directives. + +Syntax of .tab files +-------------------- + +### Comments ### + +Any line starting with `#` is a comment and is ignored. + +A line with `//` is also a comment. It is recommended to only +use this style of comments in files that define implementations of +instructions. + +A long line can be broken into shorter lines by a placing a`\` before +the newline. + +### Variable definitions ### + +A variable definition binds a variable to a Perl variable. It is only +meaningful to add a new definition if **beam\_makeops** is updated +at the same time to use the variable. A variable definition looks this: + +*name*=*value*[;] + +where *name* is the name of a Perl variable in **beam\_makeops**, +and *value* is the value to be given to the variable. The line +can optionally end with a `;` (to avoid messing up the +C indentation mode in Emacs). + +Here follows a description of the variables that are defined. + +#### BEAM\_FORMAT\_NUMBER #### + +`genop.tab` has the following definition: + + BEAM_FORMAT_NUMBER=0 + +It defines the version of the instruction set (which will be +included in the code header in the BEAM code). Theoretically, +the version could be bumped, and all instructions changed. +In practice, we would have two support two instruction sets +in the runtime system for at least two releases, so it will +probably never happen in practice. + +#### GC\_REGEXP #### + +In `macros.tab`, there is a definition of `GC_REGEXP`. +It will be described in [a later section](#the-gc_regexp-definition). + +### Directives ### + +There are directives to classify specific instructions depending +on how frequently used they are: + +* `%hot` - Implementation will be placed in `beam_hot.h`. Frequently +executed instructions. + +* `%warm` - Implementation will be placed in `beam_warm.h`. Binary +syntax instructions. + +* `%cold` - Implementation will be placed in `beam_cold.h`. Trace +instructions and infrequently used instructions. + +Default is `%hot`. The directives will be applied to declarations +of the specific instruction that follow. Here is an example: + + %cold + is_number f? xy + %hot + +#### Conditional compilation directives #### + +The `%if` directive includes a range of lines if a condition is +true. For example: + + %if ARCH_64 + i_bs_get_integer_32 x f? x + %endif + +The specific instruction `i_bs_get_integer_32` will only be defined +on a 64-bit machine. + +The condition can be inverted by using `%unless` instead of `%if`: + + %unless NO_FPE_SIGNALS + fcheckerror p => i_fcheckerror + i_fcheckerror + fclearerror + %endif + +It is also possible to add an `%else` clause: + + %if ARCH_64 + BS_SAFE_MUL(A, B, Fail, Dst) { + Uint64 res = ($A) * ($B); + if (res / $B != $A) { + $Fail; + } + $Dst = res; + } + %else + BS_SAFE_MUL(A, B, Fail, Dst) { + Uint64 res = (Uint64)($A) * (Uint64)($B); + if ((res >> (8*sizeof(Uint))) != 0) { + $Fail; + } + $Dst = res; + } + %endif + +#### Symbols that are defined in directives #### + +The following symbols are always defined. + +* `ARCH_64` - is 1 for a 64-bit machine, and 0 otherwise. +* `ARCH_32` - is 1 for 32-bit machine, and 1 otherwise. + +The `Makefile` for building the emulator currently defines the +following symbols by using the `-D` option on the command line for +**beam\_makeops**. + +* `NO_FPE_SIGNALS` - 1 if FPE signals are not enable in runtime system, +0 otherwise. +* `USE_VM_PROBES` - 1 if the runtime system is compiled to use VM probes (support for dtrace or systemtap), 0 otherwise. + +### Defining external generic instructions ### + +External generic BEAM instructions are known to both the compiler and +the runtime system. They remain stable between releases. A new major +release may add more external generic instructions, but must not change +the semantics for a previously defined instruction. + +The syntax for an external generic instruction is as follows: + +*opcode*: [-]*name*/*arity* + +*opcode* is an integer greater than or equal to 1. + +*name* is an identifier starting with a lowercase letter. *arity* is +an integer denoting the number of operands. + +*name* can optionally be preceded by `-` to indicate that it has been +obsoleted. The compiler is not allowed to generate BEAM files that +use obsolete instructions and the loader will refuse to load BEAM +files that use obsolete instructions. + +It only makes sense to define external generic instructions in the +file `genop.tab` in `lib/compiler/src`, because the compiler must +know about them in order to use them. + +New instructions must be added at the end of the file, with higher +numbers than the previous instructions. + +### Defining internal generic instructions ### + +Internal generic instructions are known only to the runtime +system and can be changed at any time without compatibility issues. + +There are two ways to define internal generic instructions: + +* Implicitly when a specific instruction is defined. This is by far +the most common way. Whenever a specific instruction is created, +**beam\_makeops** automatically creates an internal generic instruction +if it does not previously exist. + +* Explicitly. This is necessary only when a generic instruction does +not have any corresponding specific instruction. + +The syntax for an internal generic instruction is as follows: + +*name*/*arity* + +*name* is an identifier starting with a lowercase letter. *arity* is +an integer denoting the number of operands. + +### About generic instructions in general ### + +Each generic instruction has an opcode. The opcode is an integer, +greater than or equal to 1. For an external generic instruction, it +must be explicitly given `genop.tab`, while internal generic +instructions are automatically numbered by **beam\_makeops**. + +The identity of a generic instruction is its name combined with its +arity. That means that it is allowed to define two distinct generic +instructions having the same name but with different arities. For +example: + + move_window/5 + move_window/6 + +Each operand of a generic instruction is tagged with its type. A generic +instruction can have one of the following types: + +* `x` - X register. + +* `y` - Y register. + +* `l` - Floating point register number. + +* `i` - Tagged literal integer. + +* `a` - Tagged literal atom. + +* `n` - NIL (`[]`, the empty list). + +* `q` - Literal that don't fit in a word, that is an object stored on +the heap such as a list or tuple. Any heap object type is supported, +even types that don't have real literals such as external references. + +* `f` - Non-zero failure label. + +* `p` - Zero failure label. + +* `u` - Untagged integer that fits in a machine word. It is used for many +different purposes, such as the number of live registers in `test_heap/2`, +as a reference to the export for `call_ext/2`, and as the flags operand for +binary syntax instructions. When the generic instruction is translated to a +specific instruction, the type for the operand in the specific operation will +tell the loader how to treat the operand. + +* `o` - Overflow. If the value for an `u` operand does not fit in a machine +word, the type of the operand will be changed to `o` (with no associated +value). Currently only used internally in the loader in the guard constraint +function `binary_too_big()`. + +* `v` - Arity value. Only used internally in the loader. + + +### Defining specific instructions ### + +The specific instructions are known only to the runtime system and +are the instructions that are actually executed. They can be changed +at any time without causing compatibility issues. + +A specific instruction can have at most 6 operands. + +A specific instruction is defined by first giving its name followed by +the types for each operand. For example: + + move x y + +Internally, for example in the generated code and in the output from +the BEAM disassembler, the instruction `move x y` will be called `move_xy`. + +The name for a specific instruction is an identifier starting with a +lowercase letter. A type is an lowercase or uppercase letter. + +All specific instructions with a given name must have the same number +of operands. That is, the following is **not** allowed: + + move x x + move x y x y + +Here follows the type letters that more or less directly corresponds +to the types for generic instructions. + +* `x` - X register. Will be loaded as a byte offset to the X register +relative to the base of X register array. (Can be packed with other +operands.) + +* `y` - Y register. Will be loaded as a byte offset to the Y register +relative to the stack frame. (Can be packed with other operands.) + +* `r` - X register 0. An implicit operand that will not be stored in +the loaded code. + +* `l` - Floating point register number. (Can be packed with other +operands.) + +* `i` - Tagged literal integer (a SMALL that will fit in one word). + +* `a` - Tagged atom. + +* `n` - NIL or the empty list. (Will not be stored in the loaded code.) + +* `q` - Tagged CONS or BOXED pointer. That is, a term such as a list +or tuple. Any heap object type is supported, even types that don't +have real literals such as external references. + +* `f` - Failure label (non-zero). The target for a branch +or call instruction. + +* `p` - The 0 failure label, meaning that an exception should be raised +if the instruction fails. (Will not be stored in the loaded code.) + +* `c` - Any literal term; that is, immediate literals such as SMALL, +and CONS or BOXED pointers to literals. (Can be used where the +operand in the generic instruction has one of the types `i`, `a`, `n`, +or `q`.) + +The types that follow do a type test of the operand at runtime; thus, +they are generally more expensive in terms of runtime than the types +described earlier. However, those operand types are needed to avoid a +combinatorial explosion in the number of specific instructions and +overall code size of `process_main()`. + +* `s` - Tagged source: X register, Y register, or a literal term. The +tag will be tested at runtime to retrieve the value from an X +register, a Y register, or simply use the value as a tagged Erlang +term. (Implementation note: An X register is tagged as a pid, and a Y +register as a port. Therefore the literal term must not contain a +port or pid.) + +* `S` - Tagged source register (X or Y). The tag will be tested at +runtime to retrieve the value from an X register or a Y register. Slighly +cheaper than `s`. + +* `d` - Tagged destination register (X or Y). The tag will be tested +at runtime to set up a pointer to the destination register. If the +instrution performs a garbarge collection, it must use the +`$REFRESH_GEN_DEST()` macro to refresh the pointer before storing to +it (there are more details about that in a later section). + +* `j` - A failure label (combination of `f` and `p`). If the branch target 0, +an exception will be raised if instruction fails, otherwise control will be +transfered to the target address. + +The types that follows are all applied to an operand that has the `u` +type. + +* `t` - An untagged integer that will fit in 12 bits (0-4096). It can be +packed with other operands in a word. Most often used as the number +of live registers in instructions such as `test_heap`. + +* `I` - An untagged integer that will fit in 32 bits. It can be +packed with other operands in a word on a 64-bit system. + +* `W` - Untagged integer or pointer. Not possible to pack with other +operands. + +* `e` - Pointer to an export entry. Use by call instructions that call +other modules, such as `call_ext`. + +* `L` - A label. Only used by the `label/1` instruction. + +* `b` - Pointer to BIF. Used by instructions that BIFs, such as +`call_bif`. + +* `A` - A tagged arityvalue. Used in instructions that test the arity +of a tuple. + +* `P` - A byte offset into a tuple. + +* `Q` - A byte offset into the stack. Used for updating the frame +pointer register. Can be packed with other operands. + +When the loader translates a generic instruction a specific +instruction, it will choose the most specific instruction that will +fit the types. Consider the following two instructions: + + move c x + move n x + +The `c` operand can encode any literal value, including NIL. The +`n` operand only works for NIL. If we have the generic instruction +`{move,nil,{x,1}}`, the loader will translate it to `move_nx 1` +because `move n x` is more specific. `move_nx` could be slightly +faster or smaller (depending on the architecture), because the `[]` +is not stored explicitly as an operand. + +#### Syntactic sugar for specific instructions #### + +It is possible to specify more than one type letter for each operand. +Here is an example: + + move cxy xy + +This is syntactic sugar for: + + move c x + move c y + move x x + move x y + move y x + move y y + +Note the difference between `move c xy` and `move c d`. Note that `move c xy` +is equivalent to the following two definitions: + + move c x + move c y + +On the other hand, `move c d` is a single instruction. At runtime, +the `d` operand will be tested to see whether it refers to an X +register or a Y register, and a pointer to the register will be set +up. + +#### The '?' type modifier #### + +The character `?` can be added to the end of an operand to indicate +that the operand will not be used every time the instruction is executed. +For example: + + allocate_heap t I t? + is_eq_exact f? x xy + +In `allocate_heap`, the last operand is the number of live registers. +It will only be used if there is not enough heap space and a garbage +collection must be performed. + +In `is_eq_exact`, the failure address (the first operand) will only be +used if the two register operands are not equal. + +Knowing that an operand is not always used can improve how packing +is done for some instructions. + +For the `allocate_heap` instruction, without the `?` the packing would +be done like this: + + +--------------------+--------------------+ + I -> | Stack needed | &&lb_allocate_heap + + +--------------------+--------------------+ + | Heap needed | Live registers + + +--------------------+--------------------+ + +"Stack needed" and "Heap needed" are always used, but they are in +different words. Thus, at runtime the `allocate_heap` instruction +must read both words from memory even though it will not always use +"Live registers". + +With the `?`, the operands will be packed like this: + + +--------------------+--------------------+ + I -> | Live registers | &&lb_allocate_heap + + +--------------------+--------------------+ + | Heap needed | Stack needed + + +--------------------+--------------------+ + +Now "Stack needed" and "Heap needed" are in the same word. + +### Defining transformation rules ### + +Transformation rules are used to rewrite generic instructions to other +generic instructions. The transformations rules are applied +repeatedly until no rule match. At that point, the first instruction +in the resulting instruction sequence will be converted to a specific +instruction and added to the code for the module being loaded. Then +the transformation rules for the remaining instructions are run in the +same way. + +A rule is recognized by its right-pointer arrow: `=>`. To the left of +the arrow is one or more instruction patterns, separated by `|`. To +the right of the arrow is zero or more instructions, separated by `|`. +If the instructions from the BEAM code matches the instruction +patterns on the left side, they will be replaced with instructions on +the right side (or removed if there are no instructions on the right). + +#### Defining instruction patterns #### + +We will start looking at the patterns on the left side of the arrow. + +A pattern for an instruction consists of its name, followed by a pattern +for each of its operands. The operand patterns are separated by spaces. + +The simplest possible pattern is a variable. Just like in Erlang, +a variable must begin with an uppercase letter. If the same variable is +used in multiple operands, the pattern will only match if the operands +are equal. For example: + + move Same Same => + +This pattern will match if the operands for `move` are the same. If +the pattern match, the instruction will be removed. (That used to be an +actual rule a long time ago when the compiler would occasionally produce +instructions such as `{move,{x,2},{x,2}}`.) + +Variables that have been bound on the left side can be used on the +right side. For example, this rule will rewrite all `move` instructions +to `assign` instructions with the operands swapped: + + move Src Dst => assign Dst Src + +If we only want to match operands of a certain type, we can +use a type constraint. A type constraint consists of one or more +lowercase letters, each specifying a type. For example: + + is_integer Fail an => jump Fail + +The second operand pattern, `an`, will match if the second operand is +either an atom or NIL (the empty list). In case of a match, the +`is_integer/2` instruction will be replaced with a `jump/1` +instruction. + +An operand pattern can bind a variable and constrain the type at the +same time by following the variable with a `=` and the constraint. +For example: + + is_eq_exact Fail=f R=xy C=q => i_is_eq_exact_literal Fail R C + +Here the `is_eq_exact` instruction is replaced with a specialized instruction +that only compares literals, but only if the first operand is a register and +the second operand is a literal. + +#### Further constraining patterns #### + +In addition to specifying a type letter, the actual value for the type can +be specified. For example: + + move C=c x==1 => move_x1 C + +Here the second operand of `move` is constrained to be X register 1. + +When specifying an atom constraint, the atom is written as it would be +in the C source code. That is, it needs an `am_` prefix, and it must +be listed in `atom.names`. For example: + + is_boolean Fail=f a==am_true => + is_boolean Fail=f a==am_false => + +There are several constraints available for testing whether a call is to a BIF +or a function. + +The constraint `u$is_bif` will test whether the given operand refers to a BIF. +For example: + + call_ext u Bif=u$is_bif => call_bif Bif + call_ext u Func => i_call_ext Func + +The `call_ext` instruction can be used to call functions written in +Erlang as well as BIFs (or more properly called SNIFs). The +`u$is_bif` constraint will match if the operand refers to a BIF (that +is, if it is listed in the file `bif.tab`). Note that `u$is_bif` +should only be applied to operands that are known to contain an index +to the import table chunk in the BEAM file (such operands have the +type `b` or `e` in the corresponding specific instruction). If +applied to other `u` operands, it will at best return a nonsense +result. + +The `u$is_not_bif` constraint matches if the operand does not refer to +a BIF (not listed in `bif.tab`). For example: + + move S X0=x==0 | line Loc | call_ext_last Ar Func=u$is_not_bif D => \ + move S X0 | call_ext_last Ar Func D + +The `u$bif:Module:Name/Arity` constraint tests whether the given +operand refers to a specific BIF. Note that `Module:Name/Arity` +**must** be an existing BIF defined in `bif.tab`, or there will +be a compilation error. It is useful when a call to a specific BIF +should be replaced with an instruction as in this example: + + gc_bif2 Fail Live u$bif:erlang:splus/2 S1 S2 Dst => \ + gen_plus Fail Live S1 S2 Dst + +Here the call to the GC BIF `'+'/2` will be replaced with the instruction +`gen_plus/5`. Note that the same name as used in the C source code must be +used for the BIF, which in this case is `splus`. It is defined like this +in `bit.tab`: + + ubif erlang:'+'/2 splus_2 + +The `u$func:Module:Name/Arity` will test whether the given operand is a +a specific function. Here is an example: + + bif1 Fail u$func:erlang:is_constant/1 Src Dst => too_old_compiler + +`is_constant/1` used to be a BIF a long time ago. The transformation +replaces the call with the `too_old_compiler` instruction which will produce +a nicer error message than the default error would be for a missing guard BIF. + +#### Type constraints allowed in patterns #### + +Here are all type letters that are allowed on the left side of a transformation +rule. + +* `u` - An untagged integer that fits in a machine word. + +* `x` - X register. + +* `y` - Y register. + +* `l` - Floating point register number. + +* `i` - Tagged literal integer. + +* `a` - Tagged literal atom. + +* `n` - NIL (`[]`, the empty list). + +* `q` - Literals that don't fit in a word, such as list or tuples. + +* `f` - Non-zero failure label. + +* `p` - The zero failure label. + +* `j` - Any label. Equivalent to `fp`. + +* `c` - Any literal term. Equivalent to `ainq`. + +* `s` - X register, Y register, or any literal term. Equivalent to `xyc`. + +* `d` - X or Y register. Equivalent to `xy`. (In a pattern `d` will +match both source and destination registers. As an operand in a specific +instruction, it must only be used for a destination register.) + +* `o` - Overflow. An untagged integer that does not fit in a machine word. + +#### Guard constraints #### + +If the constraints described so far is not enough, additional +constraints can be written in C in `beam_load.c` and be called as a +guard function on the left side of the transformation. If the guard +function returns a non-zero value, the matching of the rule will +continue, otherwise the match will fail. For example: + + ensure_map Lit=q | literal_is_map(Lit) => + +The guard test `literal_is_map/1` tests whether the given literal is a map. +If the literal is a map, the instruction is unnecessary and can be removed. + +It is outside the scope for this document to describe in detail how such +guard functions are written, but for the curious here is the implementation +of `literal_is_map()`: + + static int + literal_is_map(LoaderState* stp, GenOpArg Lit) + { + Eterm term; + + ASSERT(Lit.type == TAG_q); + term = stp->literals[Lit.val].term; + return is_map(term); + } + +#### Handling instruction with variable number of operands #### + +Some instructions, such as `select_val/3`, essentially has a variable +number of operands. Such instructions have a `{list,[...]}` operand +as their last operand in the BEAM assembly code. For example: + + {select_val,{x,0}, + {f,1}, + {list,[{atom,b},{f,4},{atom,a},{f,5}]}}. + +The loader will convert a `{list,[...]}` operand to an `u` operand whose +value is the number of elements in the list, followed by each element in +the list. The instruction above would be translated to the following +generic instruction: + + {select_val,{x,0},{f,1},{u,4},{atom,b},{f,4},{atom,a},{f,5}} + +To match a variable number of arguments we need to use the special +operand type `*` like this: + + select_val Src=aiq Fail=f Size=u List=* => \ + i_const_select_val Src Fail Size List + +This transformation renames a `select_val/3` instruction +with a constant source operand to `i_const_select_val/3`. + +#### Constructing new instructions on the right side #### + +The most common operand on the right side is a variable that was bound while +matching the left side. For example: + + trim N Remaining => i_trim N + +An operand can also be a type letter to construct an operand of that type. +Each type has a default value. For example, the type `x` has the default +value 1023, which is the highest X register. That makes `x` on the right +side a convenient shortcut for a temporary X register. For example: + + is_number Fail Literal=q => move Literal x | is_number Fail x + +If the second operand for `is_number/2` is a literal, it will be moved to +X register 1023. Then `is_number/2` will test whether the value stored in +X register 1023 is a number. + +This kind of transformation is useful when it is rare that an operand can +be anything else but a register. In the case of `is_number/2`, the second +operand is always a register unless the compiler optimizations have been +disabled. + +If the default value is not suitable, the type letter can be followed +by `=` and a value. Most types take an integer value. The value for +an atom is written the same way as in the C source code. For example, +the atom `false` is written as `am_false`. The atom must be listed in +`atom.names`. + +Here is an example showing how values can be specified: + + bs_put_utf32 Fail=j Flags=u Src=s => \ + i_bs_validate_unicode Fail Src | \ + bs_put_integer Fail i=32 u=1 Flags Src + +#### Type letters on the right side #### + +Here follows all types that are allowed to be used in operands for +instructions being constructed on the right side of a transformation +rule. + +* `u` - Construct an untagged integer. The default value is 0. + +* `x` - X register. The default value is 1023. That makes `x` convenient to +use as a temporary X register. + +* `y` - Y register. The default value is 0. + +* `l` - Foating point register number. The default value is 0. + +* `i` - Tagged literal integer. The default value is 0. + +* `a` - Tagged atom. The default value is the empty atom (`am_Empty`). + +* `n` - NIL (`[]`, the empty list). + +#### Function call on the right side #### + +Transformations that are not possible to describe with the rule +language as described here can be written as a C function in +`beam_load.c` and called from the right side of a transformation. The +left side of the transformation will perform the match and bind +operands to variables. The variables can then be passed to a +generator function on the right side. For example: + + bif2 Fail=j u$bif:erlang:element/2 Index=s Tuple=xy Dst=d => \ + gen_element(Jump, Index, Tuple, Dst) + +This transformation rule matches a call to the BIF `element/2`. +The operands will be captured and the function `gen_element()` will +be called. + +`gen_element()` will produce one of two instructions depending +on `Index`. If `Index` is an integer in the range from 1 up to +the maximum tuple size, the instruction `i_fast_element/2` will +be produced, otherwise the instruction `i_element/4` will be +produced. The corresponding specific instructions are: + + i_fast_element xy j? I d + i_element xy j? s d + +The `i_fast_element/2` instruction is faster because the tuple is +already an untagged integer. It also knows that the index is at least +1, so it does not have to test for that. The `i_element/4` +instruction will have to fetch the index from a register, test that it +is an integer, and untag the integer. + +It is outside the scope of this document to describe in detail how +generator functions are written, but for the curious, here is the +implementation of `gen_element()`: + + static GenOp* + gen_element(LoaderState* stp, GenOpArg Fail, + GenOpArg Index, GenOpArg Tuple, GenOpArg Dst) + { + GenOp* op; + + NEW_GENOP(stp, op); + op->arity = 4; + op->next = NULL; + + if (Index.type == TAG_i && Index.val > 0 && + Index.val <= ERTS_MAX_TUPLE_SIZE && + (Tuple.type == TAG_x || Tuple.type == TAG_y)) { + op->op = genop_i_fast_element_4; + op->a[0] = Tuple; + op->a[1] = Fail; + op->a[2].type = TAG_u; + op->a[2].val = Index.val; + op->a[3] = Dst; + } else { + op->op = genop_i_element_4; + op->a[0] = Tuple; + op->a[1] = Fail; + op->a[2] = Index; + op->a[3] = Dst; + } + + return op; + } +} + +### Defining the implementation ### + +The actual implementation of instructions are also defined in `.tab` +files processed by **beam\_makeops**. For practical reasons, +instruction definitions are stored in several files, at the time of +writing in the following files: + + bif_instrs.tab + arith_instrs.tab + bs_instrs.tab + float_instrs.tab + instrs.tab + map_instrs.tab + msg_instrs.tab + select_instrs.tab + trace_instrs.tab + +There is also a file that only contains macro definitions: + + macros.tab + +The syntax of each file is similar to C code. In fact, most of +the contents *is* C code, interspersed with macro invocations. + +To allow Emacs to auto-indent the code, each file starts with the +following line: + + // -*- c -*- + +To avoid messing up the indentation, all comments are written +as C++ style comments (`//`) instead of `#`. Note that a comment +must start at the beginning of a line. + +The meat of an instruction definition file are macro definitions. +We have seen this macro definition before: + + move(Src, Dst) { + $Dst = $Src; + } + +A macro definitions must start at the beginning of the line (no spaces +allowed), the opening curly bracket must be on the same line, and the +finishing curly bracket must be at the beginning of a line. It is +recommended that the macro body is properly indented. + +As a convention, the macro arguments in the head all start with an +uppercase letter. In the body, the macro arguments can be expanded +by preceding them with `$`. + +A macro definition whose name and arity matches a family of +specific instructions is assumed to be the implementation of that +instruction. + +A macro can also be invoked from within another macro. For example, +`move_deallocate_return/2` avoids repeating code by invoking +`$deallocate_return()` as a macro: + + move_deallocate_return(Src, Deallocate) { + x(0) = $Src; + $deallocate_return($Deallocate); + } + +Here is the definition of `deallocate_return/1`: + + deallocate_return(Deallocate) { + //| -no_next + int words_to_pop = $Deallocate; + SET_I((BeamInstr *) cp_val(*E)); + E = ADD_BYTE_OFFSET(E, words_to_pop); + CHECK_TERM(x(0)); + DispatchReturn; + } + +The expanded code for `move_deallocate_return` will look this: + + OpCase(move_deallocate_return_cQ): + { + x(0) = I[1]; + do { + int words_to_pop = Qb(BeamExtraData(I[0])); + SET_I((BeamInstr *) cp_val(*E)); + E = ADD_BYTE_OFFSET(E, words_to_pop); + CHECK_TERM(x(0)); + DispatchReturn; + } while (0); + } + +When expanding macros, **beam\_makeops** wraps the expansion in a +`do`/`while` wrapper unless **beam\_makeops** can clearly see that no +wrapper is needed. In this case, the wrapper is needed. + +Note that arguments for macros cannot be complex expressions, because +the arguments are split on `,`. For example, the following would +not work because **beam\_makeops** would split the expression into +two arguments: + + $deallocate_return(get_deallocation(y, $Deallocate)); + +#### Code generation directives #### + +Within macro definitions, `//` comments are in general not treated +specially. They will be copied to the file with the generated code +along with the rest of code in the body. + +However, there is an exception. Within a macro definition, a line that +starts with whitespace followed by `//|` is treated specially. The +rest of the line is assumed to contain directives to control code +generation. + +Currently, two code generation directives are recognized: + +* `-no_prefetch` +* `-no_next` + +##### The -no_prefetch directive ##### + +To see what `-no_prefetch` does, let's first look at the default code +generation. Here is the code generated for `move_cx`: + + OpCase(move_cx): + { + BeamInstr next_pf = BeamCodeAddr(I[2]); + xb(BeamExtraData(I[0])) = I[1]; + I += 2; + ASSERT(VALID_INSTR(next_pf)); + GotoPF(next_pf); + } + +Note that the very first thing done is to fetch the address to the +next instruction. The reason is that it usually improves performance. + +Just as a demonstration, we can add a `-no_prefetch` directive to +the `move/2` instruction: + + move(Src, Dst) { + //| -no_prefetch + $Dst = $Src; + } + +We can see that the prefetch is no longer done: + + OpCase(move_cx): + { + xb(BeamExtraData(I[0])) = I[1]; + I += 2; + ASSERT(VALID_INSTR(*I)); + Goto(*I); + } + +When would we want to turn off the prefetch in practice? + +In instructions that will not always execute the next instruction. +For example: + + is_atom(Fail, Src) { + if (is_not_atom($Src)) { + $FAIL($Fail); + } + } + + // From macros.tab + FAIL(Fail) { + //| -no_prefetch + $SET_I_REL($Fail); + Goto(*I); + } + +`is_atom/2` may either execute the next instruction (if the second +operand is an atom) or branch to the failure label. + +The generated code looks like this: + + OpCase(is_atom_fx): + { + if (is_not_atom(xb(I[1]))) { + ASSERT(VALID_INSTR(*(I + (fb(BeamExtraData(I[0]))) + 0))); + I += fb(BeamExtraData(I[0])) + 0;; + Goto(*I);; + } + I += 2; + ASSERT(VALID_INSTR(*I)); + Goto(*I); + } + +##### The -no_next directive ##### + +Next we will look at when the `-no_next` directive can be used. Here +is the `jump/1` instruction: + + jump(Fail) { + $JUMP($Fail); + } + + // From macros.tab + JUMP(Fail) { + //| -no_next + $SET_I_REL($Fail); + Goto(*I); + } + +The generated code looks like this: + + OpCase(jump_f): + { + ASSERT(VALID_INSTR(*(I + (fb(BeamExtraData(I[0]))) + 0))); + I += fb(BeamExtraData(I[0])) + 0;; + Goto(*I);; + } + +If we remove the `-no_next` directive, the code would look like this: + + OpCase(jump_f): + { + BeamInstr next_pf = BeamCodeAddr(I[1]); + ASSERT(VALID_INSTR(*(I + (fb(BeamExtraData(I[0]))) + 0))); + I += fb(BeamExtraData(I[0])) + 0;; + Goto(*I);; + I += 1; + ASSERT(VALID_INSTR(next_pf)); + GotoPF(next_pf); + } + +In the end, the C compiler will probably optimize this code to the +same native code as the first version, but the first version is certainly +much easier to read for human readers. + +#### Macros in the macros.tab file #### + +The file `macros.tab` contains many useful macros. When implementing +new instructions it is good practice to look through `macros.tab` to +see if any of existing macros can be used rather than re-inventing +the wheel. + +We will describe a few of the most useful macros here. + +##### The GC_REGEXP definition ##### + +The following line defines a regular expression that will recognize +a call to a function that does a garbage collection: + + GC_REGEXP=erts_garbage_collect|erts_gc|GcBifFunction; + +The purpose is that **beam\_makeops** can verify that an instruction +that does a garbage collection and has an `d` operand uses the +`$REFRESH_GEN_DEST()` macro. + +If you need to define a new function that does garbage collection, +you should give it the prefix `erts_gc_`. If that is not possible +you should update the regular expression so that it will match your +new function. + +##### FAIL(Fail) ##### + +Branch to `$Fail`. Will suppress prefetch (`-no_prefetch`). Typical use: + + is_nonempty_list(Fail, Src) { + if (is_not_list($Src)) { + $FAIL($Fail); + } + } + +##### JUMP(Fail) ##### + +Branch to `$Fail`. Suppresses generation of dispatch of the next +instruction (`-no_next`). Typical use: + + jump(Fail) { + $JUMP($Fail); + } + +##### GC_TEST(NeedStack, NeedHeap, Live) ##### + +`$GC_TEST(NeedStack, NeedHeap, Live)` tests that given amount of +stack space and heap space is available. If not it will do a +garbage collection. Typical use: + + test_heap(Nh, Live) { + $GC_TEST(0, $Nh, $Live); + } + +##### AH(NeedStack, NeedHeap, Live) ##### + +`AH(NeedStack, NeedHeap, Live)` allocates a stack frame and +optionally additional heap space. + +#### Pre-defined macros and variables #### + +**beam\_makeops** defines several built-in macros and pre-bound variables. + +##### The NEXT_INSTRUCTION pre-bound variable ##### + +The NEXT_INSTRUCTION is a pre-bound variable that is available in +all instructions. It expands to the address of the next instruction. + +Here is an example: + + i_call(CallDest) { + SET_CP(c_p, $NEXT_INSTRUCTION); + $DISPATCH_REL($CallDest); + } + +When calling a function, the return address is first stored in `c_p->cp` +(using the `SET_CP()` macro defined in `beam_emu.c`), and then control is +transferred to the callee. Here is the generated code: + + OpCase(i_call_f): + { + SET_CP(c_p, I+1); + ASSERT(VALID_INSTR(*(I + (fb(BeamExtraData(I[0]))) + 0))); + I += fb(BeamExtraData(I[0])) + 0;; + DTRACE_LOCAL_CALL(c_p, erts_code_to_codemfa(I)); + Dispatch();; + } + +We can see that that `$NEXT_INSTRUCTION` has been expanded to `I+1`. +That makes sense since the size of the `i_call_f/1` instruction is +one word. + +##### The IP_ADJUSTMENT pre-bound variable ##### + +`$IP_ADJUSTMENT` is usually 0. In a few combined instructions +(described below) it can be non-zero. It is used like this +in `macros.tab`: + + SET_I_REL(Offset) { + ASSERT(VALID_INSTR(*(I + ($Offset) + $IP_ADJUSTMENT))); + I += $Offset + $IP_ADJUSTMENT; + } + +Avoid using `IP_ADJUSTMENT` directly. Use `SET_I_REL()` or +one of the macros that invoke such as `FAIL()` or `JUMP()` +defined in `macros.tab`. + +#### Pre-defined macro functions #### + +##### The IF() macro ##### + +`$IF(Expr, IfTrue, IfFalse)` evaluates `Expr`, which must be a valid +Perl expression (which for simple numeric expressions have the same +syntax as C). If `Expr` evaluates to 0, the entire `IF()` expression will be +replaced with `IfFalse`, otherwise it will be replaced with `IfTrue`. + +See the description of `OPERAND_POSITION()` for an example. + +##### The OPERAND\_POSITION() macro ##### + +`$OPERAND_POSITION(Expr)` returns the position for `Expr`, if +`Expr` is an operand that is not packed. The first operand is +at position 1. + +Returns 0 otherwise. + +This macro could be used like this in order to share code: + + FAIL(Fail) { + //| -no_prefetch + $IF($OPERAND_POSITION($Fail) == 1 && $IP_ADJUSTMENT == 0, + goto common_jump, + $DO_JUMP($Fail)); + } + + DO_JUMP(Fail) { + $SET_I_REL($Fail); + Goto(*I)); + } + + // In beam_emu.c: + common_jump: + I += I[1]; + Goto(*I)); + + +#### The $REFRESH\_GEN\_DEST() macro #### + +When a specific instruction has a `d` operand, early during execution +of the instruction, a pointer will be initialized to point to the X or +Y register in question. + +If there is a garbage collection before the result is stored, +the stack will move and if the `d` operand refered to a Y +register, the pointer will no longer be valid. (Y registers are +stored on the stack.) + +In those circumstances, `$REFRESH_GEN_DEST()` must be invoked +to set up the pointer again. **beam\_makeops** will notice +if there is a call to a function that does a garbage collection and +`$REFRESH_GEN_DEST()` is not called. + +Here is a complete example. The `new_map` instruction is defined +like this: + + new_map d t I + +It is implemented like this: + + new_map(Dst, Live, N) { + Eterm res; + + HEAVY_SWAPOUT; + res = erts_gc_new_map(c_p, reg, $Live, $N, $NEXT_INSTRUCTION); + HEAVY_SWAPIN; + $REFRESH_GEN_DEST(); + $Dst = res; + $NEXT($NEXT_INSTRUCTION+$N); + } + +If we have forgotten the `$REFRESH_GEN_DEST()` there would be a message +similar to this: + + pointer to destination register is invalid after GC -- use $REFRESH_GEN_DEST() + ... from the body of new_map at beam/map_instrs.tab(30) + +#### Combined instructions #### + +**Problem**: For frequently executed instructions we want to use +"fast" operands types such as `x` and `y`, as opposed to `s` or `S`. +To avoid an explosion in code size, we want to share most of the +implementation between the instructions. Here are the specific +instructions for `i_increment/5`: + + i_increment r W t d + i_increment x W t d + i_increment y W t d + +The `i_increment` instruction is implemented like this: + + i_increment(Source, IncrementVal, Live, Dst) { + Eterm increment_reg_source = $Source; + Eterm increment_val = $IncrementVal; + Uint live; + Eterm result; + + if (ERTS_LIKELY(is_small(increment_reg_val))) { + Sint i = signed_val(increment_reg_val) + increment_val; + if (ERTS_LIKELY(IS_SSMALL(i))) { + $Dst = make_small(i); + $NEXT0(); + } + } + live = $Live; + HEAVY_SWAPOUT; + reg[live] = increment_reg_val; + reg[live+1] = make_small(increment_val); + result = erts_gc_mixed_plus(c_p, reg, live); + HEAVY_SWAPIN; + ERTS_HOLE_CHECK(c_p); + if (ERTS_LIKELY(is_value(result))) { + $REFRESH_GEN_DEST(); + $Dst = result; + $NEXT0(); + } + ASSERT(c_p->freason != BADMATCH || is_value(c_p->fvalue)); + goto find_func_info; + } + +There will be three almost identical copies of the code. Given the +size of the code, that could be too high cost to pay. + +To avoid the three copies of the code, we could use only one specific +instruction: + + i_increment S W t d + +(The same implementation as above will work.) + +That reduces the code size, but is slower because `S` means that +there will be extra code to test whether the operand refers to an X +register or a Y register. + +**Solution**: We can use "combined instructions". Combined +instructions are combined from instruction fragments. The +bulk of the code can be shared. + +Here we will show how `i_increment` can be implemented as a combined +instruction. We will show each individual fragment first, and then +show how to connect them together. First we will need a variable that +we can store the value fetched from the register in: + + increment.head() { + Eterm increment_reg_val; + } + +The name `increment` is the name of the group that the fragment +belongs to. Note that it does not need to have the same +name as the instruction. The group name is followed by `.` and +the name of the fragment. The name `head` is pre-defined. +The code in it will be placed at the beginning of a block, so +that all fragments in the group can access it. + +Next we define the fragment that will pick up the value from the +register from the first operand: + + increment.fetch(Src) { + increment_reg_val = $Src; + } + +We call this fragment `fetch`. This fragment will be duplicated three +times, one for each value of the first operand (`r`, `x`, and `y`). + +Next we define the main part of the code that do the actual incrementing. + + increment.execute(IncrementVal, Live, Dst) { + Eterm increment_val = $IncrementVal; + Uint live; + Eterm result; + + if (ERTS_LIKELY(is_small(increment_reg_val))) { + Sint i = signed_val(increment_reg_val) + increment_val; + if (ERTS_LIKELY(IS_SSMALL(i))) { + $Dst = make_small(i); + $NEXT0(); + } + } + live = $Live; + HEAVY_SWAPOUT; + reg[live] = increment_reg_val; + reg[live+1] = make_small(increment_val); + result = erts_gc_mixed_plus(c_p, reg, live); + HEAVY_SWAPIN; + ERTS_HOLE_CHECK(c_p); + if (ERTS_LIKELY(is_value(result))) { + $REFRESH_GEN_DEST(); + $Dst = result; + $NEXT0(); + } + ASSERT(c_p->freason != BADMATCH || is_value(c_p->fvalue)); + goto find_func_info; + } + +We call this fragment `execute`. It will handle the three remaining +operands (`W t d`). There will only be one copy of this fragment. + +Now that we have defined the fragments, we need to inform +**beam\_makeops** how they should be connected: + + i_increment := increment.fetch.execute; + +To the left of the `:=` is the name of the specific instruction that +should be implemented by the fragments, in this case `i_increment`. +To the right of `:=` is the name of the group with the fragments, +followed by a `.`. Then the name of the fragments in the group are +listed in the order they should be executed. Note that the `head` +fragment is not listed. + +The line ends in `;` (to avoid messing up the indentation in Emacs). + +(Note that in practice the `:=` line is usually placed before the +fragments.) + +The generated code looks like this: + + { + Eterm increment_reg_val; + OpCase(i_increment_rWtd): + { + increment_reg_val = r(0); + } + goto increment__execute; + + OpCase(i_increment_xWtd): + { + increment_reg_val = xb(BeamExtraData(I[0])); + } + goto increment__execute; + + OpCase(i_increment_yWtd): + { + increment_reg_val = yb(BeamExtraData(I[0])); + } + goto increment__execute; + + increment__execute: + { + // Here follows the code from increment.execute() + . + . + . + } + +##### Some notes about combined instructions ##### + +The operands that are different must be at +the beginning of the instruction. All operands in the last +fragment must have the same operands in all variants of +the specific instruction. + +As an example, the following specific instructions cannot be +implemented as a combined instruction: + + i_times j? t x x d + i_times j? t x y d + i_times j? t s s d + +We would have to change the order of the operands so that the +two operands that are different are placed first: + + i_times x x j? t d + i_times x y j? t d + i_times s s j? t d + +We can then define: + + i_times := times.fetch.execute; + + times.head { + Eterm op1, op2; + } + + times.fetch(Src1, Src2) { + op1 = $Src1; + op2 = $Src2; + } + + times.execute(Fail, Live, Dst) { + // Multiply op1 and op2. + . + . + . + } + +Several instructions can share a group. As an example, the following +instructions have different names, but in the end they all create a +binary. The last two operands are common for all of them: + + i_bs_init_fail xy j? t? x + i_bs_init_fail_heap s I j? t? x + i_bs_init W t? x + i_bs_init_heap W I t? x + +The instructions are defined like this (formatted with extra +spaces for clarity): + + i_bs_init_fail_heap := bs_init . fail_heap . verify . execute; + i_bs_init_fail := bs_init . fail . verify . execute; + i_bs_init := bs_init . . plain . execute; + i_bs_init_heap := bs_init . heap . execute; + +Note that the first two instruction have three fragments, while the +other two only have two fragments. Here are the fragments: + + bs_init_bits.head() { + Eterm num_bits_term; + Uint num_bits; + Uint alloc; + } + + bs_init_bits.plain(NumBits) { + num_bits = $NumBits; + alloc = 0; + } + + bs_init_bits.heap(NumBits, Alloc) { + num_bits = $NumBits; + alloc = $Alloc; + } + + bs_init_bits.fail(NumBitsTerm) { + num_bits_term = $NumBitsTerm; + alloc = 0; + } + + bs_init_bits.fail_heap(NumBitsTerm, Alloc) { + num_bits_term = $NumBitsTerm; + alloc = $Alloc; + } + + bs_init_bits.verify(Fail) { + // Verify the num_bits_term, fail using $FAIL + // if there is a problem. + . + . + . + } + + bs_init_bits.execute(Live, Dst) { + // Long complicated code to a create a binary. + . + . + . + } + +The full definitions of those instructions can be found in `bs_instrs.tab`. +The generated code can be found in `beam_warm.h`. diff --git a/erts/emulator/internal_doc/figures/.gitignore b/erts/emulator/internal_doc/figures/.gitignore new file mode 100644 index 0000000000..c2813ac866 --- /dev/null +++ b/erts/emulator/internal_doc/figures/.gitignore @@ -0,0 +1 @@ +*.eps
\ No newline at end of file diff --git a/erts/emulator/internal_doc/figures/Makefile b/erts/emulator/internal_doc/figures/Makefile new file mode 100644 index 0000000000..111cb393fb --- /dev/null +++ b/erts/emulator/internal_doc/figures/Makefile @@ -0,0 +1,20 @@ +# In order to update the figures you have to have both dia +# and imagemagick installed. + +DIAGRAMS=$(wildcard *.dia) +EPS_DIAGRAMS=$(patsubst %.dia,%.eps,$(DIAGRAMS)) +PNG_DIAGRAMS=$(patsubst %.dia,%.png,$(DIAGRAMS)) + +diagrams: $(EPS_DIAGRAMS) + +png: $(PNG_DIAGRAMS) + +update_png: png + git add $(PNG_DIAGRAMS) + git commit -m "Update internal docs figures" + +%.eps: %.dia + dia --export=$@ $< + +%.png: %.eps + convert $< -resize 65% $@ diff --git a/erts/emulator/internal_doc/figures/gc-heap-scan1.dia b/erts/emulator/internal_doc/figures/gc-heap-scan1.dia Binary files differnew file mode 100644 index 0000000000..b8a32dc092 --- /dev/null +++ b/erts/emulator/internal_doc/figures/gc-heap-scan1.dia diff --git a/erts/emulator/internal_doc/figures/gc-heap-scan1.png b/erts/emulator/internal_doc/figures/gc-heap-scan1.png Binary files differnew file mode 100644 index 0000000000..9724fc8698 --- /dev/null +++ b/erts/emulator/internal_doc/figures/gc-heap-scan1.png diff --git a/erts/emulator/internal_doc/figures/gc-heap-stop.dia b/erts/emulator/internal_doc/figures/gc-heap-stop.dia Binary files differnew file mode 100644 index 0000000000..af219958c6 --- /dev/null +++ b/erts/emulator/internal_doc/figures/gc-heap-stop.dia diff --git a/erts/emulator/internal_doc/figures/gc-heap-stop.png b/erts/emulator/internal_doc/figures/gc-heap-stop.png Binary files differnew file mode 100644 index 0000000000..ec79790d36 --- /dev/null +++ b/erts/emulator/internal_doc/figures/gc-heap-stop.png diff --git a/erts/emulator/internal_doc/figures/gc-rootset-scan.dia b/erts/emulator/internal_doc/figures/gc-rootset-scan.dia Binary files differnew file mode 100644 index 0000000000..d6147740e5 --- /dev/null +++ b/erts/emulator/internal_doc/figures/gc-rootset-scan.dia diff --git a/erts/emulator/internal_doc/figures/gc-rootset-scan.png b/erts/emulator/internal_doc/figures/gc-rootset-scan.png Binary files differnew file mode 100644 index 0000000000..06509f83c3 --- /dev/null +++ b/erts/emulator/internal_doc/figures/gc-rootset-scan.png diff --git a/erts/emulator/internal_doc/figures/gc-start.dia b/erts/emulator/internal_doc/figures/gc-start.dia Binary files differnew file mode 100644 index 0000000000..41832d29b0 --- /dev/null +++ b/erts/emulator/internal_doc/figures/gc-start.dia diff --git a/erts/emulator/internal_doc/figures/gc-start.png b/erts/emulator/internal_doc/figures/gc-start.png Binary files differnew file mode 100644 index 0000000000..6c29e9adb0 --- /dev/null +++ b/erts/emulator/internal_doc/figures/gc-start.png diff --git a/erts/emulator/internal_doc/figures/gc-watermark-2.dia b/erts/emulator/internal_doc/figures/gc-watermark-2.dia Binary files differnew file mode 100644 index 0000000000..2ea3be3fb2 --- /dev/null +++ b/erts/emulator/internal_doc/figures/gc-watermark-2.dia diff --git a/erts/emulator/internal_doc/figures/gc-watermark-2.png b/erts/emulator/internal_doc/figures/gc-watermark-2.png Binary files differnew file mode 100644 index 0000000000..bca110282d --- /dev/null +++ b/erts/emulator/internal_doc/figures/gc-watermark-2.png diff --git a/erts/emulator/internal_doc/figures/gc-watermark.dia b/erts/emulator/internal_doc/figures/gc-watermark.dia Binary files differnew file mode 100644 index 0000000000..a5de7565dc --- /dev/null +++ b/erts/emulator/internal_doc/figures/gc-watermark.dia diff --git a/erts/emulator/internal_doc/figures/gc-watermark.png b/erts/emulator/internal_doc/figures/gc-watermark.png Binary files differnew file mode 100644 index 0000000000..b835c7694e --- /dev/null +++ b/erts/emulator/internal_doc/figures/gc-watermark.png diff --git a/erts/emulator/nifs/common/zlib_nif.c b/erts/emulator/nifs/common/zlib_nif.c index a9c5b05e47..b709ed5a6f 100644 --- a/erts/emulator/nifs/common/zlib_nif.c +++ b/erts/emulator/nifs/common/zlib_nif.c @@ -106,6 +106,7 @@ typedef struct { int inflateChunk_buffer_size; ErlNifPid controlling_process; + ErlNifMutex *controller_lock; ErlNifIOQueue *input_queue; @@ -117,6 +118,8 @@ typedef struct { static ERL_NIF_TERM zlib_open(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]); static ERL_NIF_TERM zlib_close(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]); +static ERL_NIF_TERM zlib_set_controller(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]); + static ERL_NIF_TERM zlib_deflateInit(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]); static ERL_NIF_TERM zlib_deflateSetDictionary(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]); static ERL_NIF_TERM zlib_deflateReset(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]); @@ -143,6 +146,11 @@ static ERL_NIF_TERM zlib_setBufSize(ErlNifEnv *env, int argc, const ERL_NIF_TERM static ERL_NIF_TERM zlib_enqueue_input(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]); static ErlNifFunc nif_funcs[] = { + {"close_nif", 1, zlib_close}, + {"open_nif", 0, zlib_open}, + + {"set_controller_nif", 2, zlib_set_controller}, + /* deflate */ {"deflateInit_nif", 6, zlib_deflateInit}, {"deflateSetDictionary_nif", 2, zlib_deflateSetDictionary}, @@ -162,10 +170,6 @@ static ErlNifFunc nif_funcs[] = { /* running checksum */ {"crc32_nif", 1, zlib_crc32}, - /* open & close */ - {"close_nif", 1, zlib_close}, - {"open_nif", 0, zlib_open}, - /* The stash keeps a single term alive across calls, and is used in * exception_on_need_dict/1 to retain the old error behavior, and for * saving data flushed through deflateParams/3. */ @@ -281,9 +285,7 @@ static ERL_NIF_TERM zlib_return(ErlNifEnv *env, int code) { return reason; } -static void gc_zlib(ErlNifEnv *env, void* data) { - zlib_data_t *d = (zlib_data_t*)data; - +static void zlib_internal_close(zlib_data_t *d) { if(d->state == ST_DEFLATE) { deflateEnd(&d->s); } else if(d->state == ST_INFLATE) { @@ -291,8 +293,6 @@ static void gc_zlib(ErlNifEnv *env, void* data) { } if(d->state != ST_CLOSED) { - enif_ioq_destroy(d->input_queue); - if(d->stash_env != NULL) { enif_free_env(d->stash_env); } @@ -301,17 +301,36 @@ static void gc_zlib(ErlNifEnv *env, void* data) { } } +static void gc_zlib(ErlNifEnv *env, void* data) { + zlib_data_t *d = (zlib_data_t*)data; + + enif_mutex_destroy(d->controller_lock); + enif_ioq_destroy(d->input_queue); + + zlib_internal_close(d); + + (void)env; +} + static int get_zlib_data(ErlNifEnv *env, ERL_NIF_TERM opaque, zlib_data_t **d) { return enif_get_resource(env, opaque, rtype_zlib, (void **)d); } static int zlib_process_check(ErlNifEnv *env, zlib_data_t *d) { + int is_controlling_process; ErlNifPid current_process; enif_self(env, ¤t_process); - return enif_is_identical(enif_make_pid(env, ¤t_process), + enif_mutex_lock(d->controller_lock); + + is_controlling_process = enif_is_identical( + enif_make_pid(env, ¤t_process), enif_make_pid(env, &d->controlling_process)); + + enif_mutex_unlock(d->controller_lock); + + return is_controlling_process; } static void zlib_reset_input(zlib_data_t *d) { @@ -516,6 +535,8 @@ static ERL_NIF_TERM zlib_open(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[ d->input_queue = enif_ioq_create(ERL_NIF_IOQ_NORMAL); + d->controller_lock = enif_mutex_create("zlib_controller_lock"); + d->s.zalloc = zlib_alloc; d->s.zfree = zlib_free; d->s.opaque = d; @@ -556,7 +577,28 @@ static ERL_NIF_TERM zlib_close(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv return enif_raise_exception(env, am_not_initialized); } - gc_zlib(env, d); + zlib_internal_close(d); + + return am_ok; +} + +static ERL_NIF_TERM zlib_set_controller(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]) { + zlib_data_t *d; + + ErlNifPid new_owner; + + if(argc != 2 || !get_zlib_data(env, argv[0], &d) + || !enif_get_local_pid(env, argv[1], &new_owner)) { + return enif_make_badarg(env); + } else if(!zlib_process_check(env, d)) { + return enif_raise_exception(env, am_not_on_controlling_process); + } + + enif_mutex_lock(d->controller_lock); + + d->controlling_process = new_owner; + + enif_mutex_unlock(d->controller_lock); return am_ok; } @@ -675,7 +717,9 @@ static ERL_NIF_TERM zlib_deflateEnd(ErlNifEnv *env, int argc, const ERL_NIF_TERM static ERL_NIF_TERM zlib_deflateParams(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]) { zlib_data_t *d; + int res, level, strategy; + Bytef dummy_buffer; if(argc != 3 || !get_zlib_data(env, argv[0], &d) || !enif_get_int(env, argv[1], &level) @@ -687,12 +731,27 @@ static ERL_NIF_TERM zlib_deflateParams(ErlNifEnv *env, int argc, const ERL_NIF_T return enif_raise_exception(env, am_not_initialized); } - /* deflateParams will flush everything currently in the stream, corrupting - * the heap unless it's empty. We therefore pretend to have a full output - * buffer, forcing a Z_BUF_ERROR if there's anything left to be flushed. */ - d->s.avail_out = 0; + /* This is a bit of a hack; deflateParams flushes with Z_BLOCK which won't + * stop at a byte boundary, so we can't split this operation up, and we + * can't allocate a buffer large enough to fit it in one go since we have + * to support zlib versions that lack deflatePending. + * + * We therefore flush everything prior to this call to ensure that we are + * stopped on a byte boundary and have no pending data. We then hand it a + * dummy buffer to detect when this assumption doesn't hold (Hopefully + * never), and to smooth over an issue with zlib 1.2.11 which always + * returns Z_BUF_ERROR when d->s.avail_out is 0, regardless of whether + * there's any pending data or not. */ + + d->s.next_out = &dummy_buffer; + d->s.avail_out = 1; + res = deflateParams(&d->s, level, strategy); + if(d->s.avail_out == 0) { + return zlib_return(env, Z_STREAM_ERROR); + } + return zlib_return(env, res); } @@ -887,7 +946,7 @@ static ERL_NIF_TERM zlib_inflate(ErlNifEnv *env, int argc, const ERL_NIF_TERM ar return enif_raise_exception(env, am_not_initialized); } - if(d->eos_seen) { + if(d->eos_seen && enif_ioq_size(d->input_queue) > 0) { int res; switch(d->eos_behavior) { @@ -901,11 +960,10 @@ static ERL_NIF_TERM zlib_inflate(ErlNifEnv *env, int argc, const ERL_NIF_TERM ar } d->eos_seen = 0; + break; case EOS_BEHAVIOR_CUT: zlib_reset_input(d); - - return enif_make_tuple2(env, am_finished, enif_make_list(env, 0)); } } diff --git a/erts/emulator/sys/common/erl_poll.c b/erts/emulator/sys/common/erl_poll.c index 30a595c17a..7aa53e8f36 100644 --- a/erts/emulator/sys/common/erl_poll.c +++ b/erts/emulator/sys/common/erl_poll.c @@ -2059,7 +2059,8 @@ uint32_t epoll_events(int kp_fd, int fd) int ev_fd; uint32_t events; uint64_t data; - if (fscanf(f,"tfd:%d events:%x data:%lx\n", &ev_fd, &events, &data) != 3) { + if (fscanf(f,"tfd:%d events:%x data:%llx\n", &ev_fd, &events, + (unsigned long long*)&data) != 3) { fprintf(stderr,"failed to parse file %s on line %d, errno = %d\n", fname, line, errno); @@ -2125,7 +2126,8 @@ ERTS_POLL_EXPORT(erts_poll_get_selected_events)(ErtsPollSet *ps, int fd; uint32_t events; uint64_t data; - if (fscanf(f,"tfd:%d events:%x data:%lx\n", &fd, &events, &data) != 3) { + if (fscanf(f,"tfd:%d events:%x data:%llx\n", &fd, &events, + (unsigned long long*)&data) != 3) { fprintf(stderr,"failed to parse file %s on line %d, errno = %d\n", fname, line, errno); ASSERT(0); diff --git a/erts/emulator/test/alloc_SUITE.erl b/erts/emulator/test/alloc_SUITE.erl index f0871ead7d..5e5cd6f578 100644 --- a/erts/emulator/test/alloc_SUITE.erl +++ b/erts/emulator/test/alloc_SUITE.erl @@ -89,10 +89,11 @@ mmsc_flags() -> mmsc_flags(Env) -> case os:getenv(Env) of false -> false; - V -> case string:str(V, "+MMsc") of - 0 -> false; - P -> Env ++ "=" ++ string:substr(V, P) - end + V -> + case string:find(V, "+MMsc") of + nomatch -> false; + SubStr -> Env ++ "=" ++ SubStr + end end. erts_mmap_do(Config, SCO, SCRPM, SCRFSD) -> diff --git a/erts/emulator/test/bif_SUITE.erl b/erts/emulator/test/bif_SUITE.erl index 04b7f2de15..e1b42e5d85 100644 --- a/erts/emulator/test/bif_SUITE.erl +++ b/erts/emulator/test/bif_SUITE.erl @@ -731,10 +731,12 @@ erlang_halt(Config) when is_list(Config) -> [broken_halt, "Validate correct crash dump"]), {ok,_} = wait_until_stable_size(CrashDump,-1), {ok, Bin} = file:read_file(CrashDump), - case {string:str(binary_to_list(Bin),"\n=end\n"), - string:str(binary_to_list(Bin),"\r\n=end\r\n")} of - {0,0} -> ct:fail("Could not find end marker in crash dump"); - _ -> ok + case {string:find(Bin, <<"\n=end\n">>), + string:find(Bin, <<"\r\n=end\r\n">>)} of + {nomatch,nomatch} -> + ct:fail("Could not find end marker in crash dump"); + {_,_} -> + ok end. wait_until_stable_size(_File,-10) -> @@ -779,14 +781,20 @@ is_builtin(_Config) -> {F,A} <- M:module_info(exports)], Exp = ordsets:from_list(Exp0), - %% erlang:apply/3 is considered to be built-in, but is not - %% implemented as other BIFs. + %% Built-ins implemented as special instructions. + Instructions = [{erlang,apply,2},{erlang,apply,3},{erlang,yield,0}], - Builtins0 = [{erlang,apply,3}|erlang:system_info(snifs)], + Builtins0 = Instructions ++ erlang:system_info(snifs), Builtins = ordsets:from_list(Builtins0), - NotBuiltin = ordsets:subtract(Exp, Builtins), - _ = [true = erlang:is_builtin(M, F, A) || {M,F,A} <- Builtins], - _ = [false = erlang:is_builtin(M, F, A) || {M,F,A} <- NotBuiltin], + + Fakes = [{M,F,42} || {M,F,_} <- Instructions], + All = ordsets:from_list(Fakes ++ Exp), + NotBuiltin = ordsets:subtract(All, Builtins), + + _ = [{true,_} = {erlang:is_builtin(M, F, A),MFA} || + {M,F,A}=MFA <- Builtins], + _ = [{false,_} = {erlang:is_builtin(M, F, A),MFA} || + {M,F,A}=MFA <- NotBuiltin], ok. diff --git a/erts/emulator/test/binary_SUITE.erl b/erts/emulator/test/binary_SUITE.erl index 61536bacd7..cbc2d8fae5 100644 --- a/erts/emulator/test/binary_SUITE.erl +++ b/erts/emulator/test/binary_SUITE.erl @@ -48,6 +48,7 @@ bad_list_to_binary/1, bad_binary_to_list/1, t_split_binary/1, bad_split/1, terms/1, terms_float/1, float_middle_endian/1, + b2t_used_big/1, external_size/1, t_iolist_size/1, t_hash/1, bad_size/1, @@ -72,6 +73,7 @@ all() -> t_split_binary, bad_split, bad_list_to_binary, bad_binary_to_list, terms, terms_float, float_middle_endian, external_size, t_iolist_size, + b2t_used_big, bad_binary_to_term_2, safe_binary_to_term2, bad_binary_to_term, bad_terms, t_hash, bad_size, bad_term_to_binary, more_bad_terms, otp_5484, otp_5933, @@ -425,40 +427,77 @@ bad_term_to_binary(Config) when is_list(Config) -> terms(Config) when is_list(Config) -> TestFun = fun(Term) -> - try - S = io_lib:format("~p", [Term]), - io:put_chars(S) - catch - error:badarg -> - io:put_chars("bit sized binary") - end, + S = io_lib:format("~p", [Term]), + io:put_chars(S), Bin = term_to_binary(Term), case erlang:external_size(Bin) of Sz when is_integer(Sz), size(Bin) =< Sz -> ok end, - Bin1 = term_to_binary(Term, [{minor_version, 1}]), - case erlang:external_size(Bin1, [{minor_version, 1}]) of - Sz1 when is_integer(Sz1), size(Bin1) =< Sz1 -> - ok - end, + Bin1 = term_to_binary(Term, [{minor_version, 1}]), + case erlang:external_size(Bin1, [{minor_version, 1}]) of + Sz1 when is_integer(Sz1), size(Bin1) =< Sz1 -> + ok + end, Term = binary_to_term_stress(Bin), Term = binary_to_term_stress(Bin, [safe]), - Unaligned = make_unaligned_sub_binary(Bin), - Term = binary_to_term_stress(Unaligned), - Term = binary_to_term_stress(Unaligned, []), - Term = binary_to_term_stress(Bin, [safe]), + Bin_sz = byte_size(Bin), + {Term,Bin_sz} = binary_to_term_stress(Bin, [used]), + + BinE = <<Bin/binary, 1, 2, 3>>, + {Term,Bin_sz} = binary_to_term_stress(BinE, [used]), + + BinU = make_unaligned_sub_binary(Bin), + Term = binary_to_term_stress(BinU), + Term = binary_to_term_stress(BinU, []), + Term = binary_to_term_stress(BinU, [safe]), + {Term,Bin_sz} = binary_to_term_stress(BinU, [used]), + + BinUE = make_unaligned_sub_binary(BinE), + {Term,Bin_sz} = binary_to_term_stress(BinUE, [used]), + BinC = erlang:term_to_binary(Term, [compressed]), + BinC_sz = byte_size(BinC), + true = BinC_sz =< size(Bin), Term = binary_to_term_stress(BinC), - true = size(BinC) =< size(Bin), + {Term, BinC_sz} = binary_to_term_stress(BinC, [used]), + Bin = term_to_binary(Term, [{compressed,0}]), terms_compression_levels(Term, size(Bin), 1), - UnalignedC = make_unaligned_sub_binary(BinC), - Term = binary_to_term_stress(UnalignedC) + + BinUC = make_unaligned_sub_binary(BinC), + Term = binary_to_term_stress(BinUC), + {Term,BinC_sz} = binary_to_term_stress(BinUC, [used]), + + BinCE = <<BinC/binary, 1, 2, 3>>, + {Term,BinC_sz} = binary_to_term_stress(BinCE, [used]), + + BinUCE = make_unaligned_sub_binary(BinCE), + Term = binary_to_term_stress(BinUCE), + {Term,BinC_sz} = binary_to_term_stress(BinUCE, [used]) end, test_terms(TestFun), ok. +%% Test binary_to_term(_, [used]) returning a big Used integer. +b2t_used_big(_Config) -> + case erlang:system_info(wordsize) of + 8 -> + {skipped, "This is not a 32-bit machine"}; + 4 -> + %% Use a long utf8 atom for large external format but compact on heap. + BigAtom = binary_to_atom(<< <<16#F0908D88:32>> || _ <- lists:seq(1,255) >>, + utf8), + Atoms = (1 bsl 17) + (1 bsl 9), + BigAtomList = lists:duplicate(Atoms, BigAtom), + BigBin = term_to_binary(BigAtomList), + {BigAtomList, Used} = binary_to_term(BigBin, [used]), + 2 = erts_debug:size(Used), + Used = byte_size(BigBin), + Used = 1 + 1 + 4 + Atoms*(1+2+4*255) + 1, + ok + end. + terms_compression_levels(Term, UncompressedSz, Level) when Level < 10 -> BinC = erlang:term_to_binary(Term, [{compressed,Level}]), Term = binary_to_term_stress(BinC), diff --git a/erts/emulator/test/decode_packet_SUITE.erl b/erts/emulator/test/decode_packet_SUITE.erl index 54ee4d5567..0ccdbd7ee8 100644 --- a/erts/emulator/test/decode_packet_SUITE.erl +++ b/erts/emulator/test/decode_packet_SUITE.erl @@ -239,7 +239,7 @@ packet_size(Config) when is_list(Config) -> %% Test OTP-9389, long HTTP header lines. Opts = [{packet_size, 128}], Pkt = list_to_binary(["GET / HTTP/1.1\r\nHost: localhost\r\nLink: /", - string:chars($Y, 64), "\r\n\r\n"]), + lists:duplicate(64, $Y), "\r\n\r\n"]), <<Pkt1:50/binary, Pkt2/binary>> = Pkt, {ok, {http_request,'GET',{abs_path,"/"},{1,1}}, Rest1} = erlang:decode_packet(http, Pkt1, Opts), @@ -250,7 +250,7 @@ packet_size(Config) when is_list(Config) -> erlang:decode_packet(httph, list_to_binary([Rest2, Pkt2]), Opts), Pkt3 = list_to_binary(["GET / HTTP/1.1\r\nHost: localhost\r\nLink: /", - string:chars($Y, 129), "\r\n\r\n"]), + lists:duplicate(129, $Y), "\r\n\r\n"]), {ok, {http_request,'GET',{abs_path,"/"},{1,1}}, Rest3} = erlang:decode_packet(http, Pkt3, Opts), {ok, {http_header,_,'Host',_,"localhost"}, Rest4} = @@ -509,9 +509,9 @@ decode_line(Bin,MaxLen) -> end. find_in_binary(Byte, Bin) -> - case string:chr(binary_to_list(Bin),Byte) of - 0 -> notfound; - P -> P + case string:find(Bin, [Byte]) of + nomatch -> notfound; + Suffix -> byte_size(Bin) - byte_size(Suffix) + 1 end. ssl(Config) when is_list(Config) -> @@ -562,7 +562,7 @@ decode_pkt(Type,Bin,Opts) -> otp_9389(Config) when is_list(Config) -> Opts = [{packet_size, 16384}, {line_length, 3000}], Pkt = list_to_binary(["GET / HTTP/1.1\r\nHost: localhost\r\nLink: /", - string:chars($X, 8192), + lists:duplicate(8192, $X), "\r\nContent-Length: 0\r\n\r\n"]), <<Pkt1:5000/binary, Pkt2/binary>> = Pkt, {ok, {http_request,'GET',{abs_path,"/"},{1,1}}, Rest1} = diff --git a/erts/emulator/test/distribution_SUITE.erl b/erts/emulator/test/distribution_SUITE.erl index 17cd1d1a3b..e2914cbc92 100644 --- a/erts/emulator/test/distribution_SUITE.erl +++ b/erts/emulator/test/distribution_SUITE.erl @@ -786,8 +786,8 @@ dist_auto_connect_once(Config) when is_list(Config) -> {ok, pong} = do_inet_rpc(Sock2,net_adm,ping,[NN]), {ok,[NN2]} = do_inet_rpc(Sock,erlang,nodes,[]), {ok,[NN]} = do_inet_rpc(Sock2,erlang,nodes,[]), - [_,HostPartPeer] = string:tokens(atom_to_list(NN),"@"), - [_,MyHostPart] = string:tokens(atom_to_list(node()),"@"), + [_,HostPartPeer] = string:lexemes(atom_to_list(NN),"@"), + [_,MyHostPart] = string:lexemes(atom_to_list(node()),"@"), % Give net_kernel a chance to change the state of the node to up to. receive after 1000 -> ok end, case HostPartPeer of @@ -1365,81 +1365,59 @@ bad_dist_structure(Config) when is_list(Config) -> start_monitor(Offender,P), P ! one, send_bad_structure(Offender, P,{?DOP_MONITOR_P_EXIT,'replace',P,normal},2), - pong = rpc:call(Victim, net_adm, ping, [Offender]), + start_monitor(Offender,P), send_bad_structure(Offender, P,{?DOP_MONITOR_P_EXIT,'replace',P,normal,normal},2), - pong = rpc:call(Victim, net_adm, ping, [Offender]), + start_link(Offender,P), send_bad_structure(Offender, P,{?DOP_LINK},0), - pong = rpc:call(Victim, net_adm, ping, [Offender]), + start_link(Offender,P), send_bad_structure(Offender, P,{?DOP_UNLINK,'replace'},2), - pong = rpc:call(Victim, net_adm, ping, [Offender]), + start_link(Offender,P), send_bad_structure(Offender, P,{?DOP_UNLINK,'replace',make_ref()},2), - pong = rpc:call(Victim, net_adm, ping, [Offender]), + start_link(Offender,P), send_bad_structure(Offender, P,{?DOP_UNLINK,make_ref(),P},0), - pong = rpc:call(Victim, net_adm, ping, [Offender]), + start_link(Offender,P), send_bad_structure(Offender, P,{?DOP_UNLINK,normal,normal},0), - pong = rpc:call(Victim, net_adm, ping, [Offender]), + start_monitor(Offender,P), send_bad_structure(Offender, P,{?DOP_MONITOR_P,'replace',P},2), - pong = rpc:call(Victim, net_adm, ping, [Offender]), + start_monitor(Offender,P), send_bad_structure(Offender, P,{?DOP_MONITOR_P,'replace',P,normal},2), - pong = rpc:call(Victim, net_adm, ping, [Offender]), + start_monitor(Offender,P), send_bad_structure(Offender, P,{?DOP_DEMONITOR_P,'replace',P},2), - pong = rpc:call(Victim, net_adm, ping, [Offender]), + start_monitor(Offender,P), send_bad_structure(Offender, P,{?DOP_DEMONITOR_P,'replace',P,normal},2), - pong = rpc:call(Victim, net_adm, ping, [Offender]), + send_bad_structure(Offender, P,{?DOP_EXIT,'replace',P},2), - pong = rpc:call(Victim, net_adm, ping, [Offender]), send_bad_structure(Offender, P,{?DOP_EXIT,make_ref(),normal,normal},0), - pong = rpc:call(Victim, net_adm, ping, [Offender]), send_bad_structure(Offender, P,{?DOP_EXIT_TT,'replace',token,P},2), - pong = rpc:call(Victim, net_adm, ping, [Offender]), send_bad_structure(Offender, P,{?DOP_EXIT_TT,make_ref(),token,normal,normal},0), - pong = rpc:call(Victim, net_adm, ping, [Offender]), send_bad_structure(Offender, P,{?DOP_EXIT2,'replace',P},2), - pong = rpc:call(Victim, net_adm, ping, [Offender]), send_bad_structure(Offender, P,{?DOP_EXIT2,make_ref(),normal,normal},0), - pong = rpc:call(Victim, net_adm, ping, [Offender]), send_bad_structure(Offender, P,{?DOP_EXIT2_TT,'replace',token,P},2), - pong = rpc:call(Victim, net_adm, ping, [Offender]), send_bad_structure(Offender, P,{?DOP_EXIT2_TT,make_ref(),token,normal,normal},0), - pong = rpc:call(Victim, net_adm, ping, [Offender]), send_bad_structure(Offender, P,{?DOP_GROUP_LEADER,'replace'},2), - pong = rpc:call(Victim, net_adm, ping, [Offender]), send_bad_structure(Offender, P,{?DOP_GROUP_LEADER,'replace','atomic'},2), - pong = rpc:call(Victim, net_adm, ping, [Offender]), send_bad_structure(Offender, P,{?DOP_GROUP_LEADER,'replace',P},0), - pong = rpc:call(Victim, net_adm, ping, [Offender]), send_bad_structure(Offender, P,{?DOP_REG_SEND_TT,'replace','',name},2,{message}), - pong = rpc:call(Victim, net_adm, ping, [Offender]), send_bad_structure(Offender, P,{?DOP_REG_SEND_TT,'replace','',name,token},0,{message}), - pong = rpc:call(Victim, net_adm, ping, [Offender]), send_bad_structure(Offender, P,{?DOP_REG_SEND,'replace',''},2,{message}), - pong = rpc:call(Victim, net_adm, ping, [Offender]), send_bad_structure(Offender, P,{?DOP_REG_SEND,'replace','',P},0,{message}), - pong = rpc:call(Victim, net_adm, ping, [Offender]), send_bad_structure(Offender, P,{?DOP_REG_SEND,'replace','',name},0,{message}), - pong = rpc:call(Victim, net_adm, ping, [Offender]), send_bad_structure(Offender, P,{?DOP_REG_SEND,'replace','',name,{token}},2,{message}), - pong = rpc:call(Victim, net_adm, ping, [Offender]), send_bad_structure(Offender, P,{?DOP_SEND_TT,'',P},0,{message}), - pong = rpc:call(Victim, net_adm, ping, [Offender]), send_bad_structure(Offender, P,{?DOP_SEND_TT,'',name,token},0,{message}), - pong = rpc:call(Victim, net_adm, ping, [Offender]), send_bad_structure(Offender, P,{?DOP_SEND,''},0,{message}), - pong = rpc:call(Victim, net_adm, ping, [Offender]), send_bad_structure(Offender, P,{?DOP_SEND,'',name},0,{message}), - pong = rpc:call(Victim, net_adm, ping, [Offender]), send_bad_structure(Offender, P,{?DOP_SEND,'',P,{token}},0,{message}), - pong = rpc:call(Victim, net_adm, ping, [Offender]), P ! two, P ! check_msgs, receive @@ -1685,13 +1663,16 @@ bad_dist_ext_size(Config) when is_list(Config) -> start_node_monitors([Offender,Victim]), Parent = self(), - P = spawn_link(Victim, + P = spawn_opt(Victim, fun () -> Parent ! {self(), started}, receive check_msgs -> ok end, %% DID CRASH HERE bad_dist_ext_check_msgs([one]), Parent ! {self(), messages_checked} - end), + end, + [link, + %% on_heap to force total_heap_size to inspect msg queue + {message_queue_data, on_heap}]), receive {P, started} -> ok end, P ! one, @@ -1714,6 +1695,7 @@ bad_dist_ext_size(Config) when is_list(Config) -> verify_still_up(Offender, Victim), + %% Let process_info(P, total_heap_size) find bad msg and disconnect rpc:call(Victim, erlang, process_info, [P, total_heap_size]), verify_down(Offender, connection_closed, Victim, killed), @@ -1795,10 +1777,11 @@ send_bad_structure(Offender,Victim,Bad,WhereToPutSelf) -> send_bad_structure(Offender,Victim,Bad,WhereToPutSelf,PayLoad) -> Parent = self(), Done = make_ref(), - spawn(Offender, + spawn_link(Offender, fun () -> Node = node(Victim), pong = net_adm:ping(Node), + erlang:monitor_node(Node, true), DCtrl = dctrl(Node), Bad1 = case WhereToPutSelf of 0 -> @@ -1812,7 +1795,16 @@ send_bad_structure(Offender,Victim,Bad,WhereToPutSelf,PayLoad) -> [] -> []; _Other -> [dmsg_ext(PayLoad)] end, + + receive {nodedown, Node} -> exit("premature nodedown") + after 10 -> ok + end, + dctrl_send(DCtrl, DData), + + receive {nodedown, Node} -> ok + after 5000 -> exit("missing nodedown") + end, Parent ! {DData,Done} end), receive @@ -2094,7 +2086,7 @@ start_relay_node(Node, Args) -> [{args, Args ++ " -setcookie "++Cookie++" -pa "++Pa++" "++ RunArg}]), - [N,H] = string:tokens(atom_to_list(NN),"@"), + [N,H] = string:lexemes(atom_to_list(NN),"@"), {ok, Sock} = gen_tcp:accept(LSock), pang = net_adm:ping(NN), {N,H,Sock}. diff --git a/erts/emulator/test/driver_SUITE.erl b/erts/emulator/test/driver_SUITE.erl index 33d0b708cf..475e03087a 100644 --- a/erts/emulator/test/driver_SUITE.erl +++ b/erts/emulator/test/driver_SUITE.erl @@ -1192,9 +1192,9 @@ check_driver_system_info_result(Result) -> io:format("All names: ~p~n", [?EXPECTED_SYSTEM_INFO_NAMES]), io:format("Result: ~p~n", [Result]), {[], Ns, DDVSN} = chk_sis(lists:map(fun (Str) -> - string:tokens(Str, "=") + string:lexemes(Str, "=") end, - string:tokens(Result, " ")), + string:lexemes(Result, " ")), ?EXPECTED_SYSTEM_INFO_NAMES), case {DDVSN, drv_vsn_str2tup(erlang:system_info(driver_version))} of @@ -2438,7 +2438,7 @@ wait_until(Fun) -> end. drv_vsn_str2tup(Str) -> - [Major, Minor] = string:tokens(Str, "."), + [Major, Minor] = string:lexemes(Str, "."), {list_to_integer(Major), list_to_integer(Minor)}. %% Build port data from a template. diff --git a/erts/emulator/test/num_bif_SUITE.erl b/erts/emulator/test/num_bif_SUITE.erl index 1c76eb8019..17555d63c6 100644 --- a/erts/emulator/test/num_bif_SUITE.erl +++ b/erts/emulator/test/num_bif_SUITE.erl @@ -145,7 +145,7 @@ t_float_to_string(Config) when is_list(Config) -> 123456789012345678.0, [{decimals, 237}])), {'EXIT', {badarg, _}} = (catch float_to_binary( 123456789012345678.0, [{decimals, 237}])), - test_fts("1." ++ string:copies("0", 249) ++ "e+00", + test_fts("1." ++ lists:duplicate(249, $0) ++ "e+00", 1.0, [{scientific, 249}, compact]), X1 = float_to_list(1.0), diff --git a/erts/emulator/test/port_SUITE.erl b/erts/emulator/test/port_SUITE.erl index 730a17d7e8..e46665a881 100644 --- a/erts/emulator/test/port_SUITE.erl +++ b/erts/emulator/test/port_SUITE.erl @@ -1035,6 +1035,9 @@ huge_env(Config) when is_list(Config) -> try erlang:open_port({spawn,Cmd},[exit_status, {env, Env}]) of P -> receive + {P, {exit_status,N}} = M when N > 127-> + %% If exit status is > 127 something went very wrong + ct:fail("Open port failed got ~p",[M]); {P, {exit_status,N}} = M -> %% We test that the exit status is an integer, this means %% that the child program has started. If we get an atom @@ -1662,13 +1665,7 @@ spawn_executable(Config) when is_list(Config) -> [ExactFile2,"hello world","dlrow olleh"] = run_echo_args_2(unicode:characters_to_binary("\""++ExactFile2++"\" "++"\"hello world\" \"dlrow olleh\"")), - ExeExt = - case string:to_lower(lists:last(string:tokens(ExactFile2,"."))) of - "exe" -> - ".exe"; - _ -> - "" - end, + ExeExt = filename:extension(ExactFile2), Executable2 = "spoky name"++ExeExt, file:copy(ExactFile1,filename:join([SpaceDir,Executable2])), ExactFile3 = filename:nativename(filename:join([SpaceDir,Executable2])), @@ -1836,7 +1833,7 @@ collect_data(Port) -> end. parse_echo_args_output(Data) -> - [lists:last(string:tokens(S,"|")) || S <- string:tokens(Data,"\r\n")]. + [lists:last(string:lexemes(S,"|")) || S <- string:lexemes(Data,["\r\n",$\n])]. %% Test that the emulator does not mix up ports when the port table wraps mix_up_ports(Config) when is_list(Config) -> @@ -1962,7 +1959,7 @@ max_ports() -> erlang:system_info(port_limit). port_ix(Port) when is_port(Port) -> - ["#Port",_,PortIxStr] = string:tokens(erlang:port_to_list(Port), + ["#Port",_,PortIxStr] = string:lexemes(erlang:port_to_list(Port), "<.>"), list_to_integer(PortIxStr). diff --git a/erts/emulator/test/process_SUITE.erl b/erts/emulator/test/process_SUITE.erl index a9f20f9928..a8bcfac84d 100644 --- a/erts/emulator/test/process_SUITE.erl +++ b/erts/emulator/test/process_SUITE.erl @@ -2532,8 +2532,13 @@ system_task_on_suspended(Config) when is_list(Config) -> end. gc_request_when_gc_disabled(Config) when is_list(Config) -> - Master = self(), AIS = erts_debug:set_internal_state(available_internal_state, true), + gc_request_when_gc_disabled_do(ref), + gc_request_when_gc_disabled_do(immed), + erts_debug:set_internal_state(available_internal_state, AIS). + +gc_request_when_gc_disabled_do(ReqIdType) -> + Master = self(), {P, M} = spawn_opt(fun () -> true = erts_debug:set_internal_state(gc_state, false), @@ -2545,7 +2550,10 @@ gc_request_when_gc_disabled(Config) when is_list(Config) -> receive after 100 -> ok end end, [monitor, link]), receive {P, gc_state, false} -> ok end, - ReqId = make_ref(), + ReqId = case ReqIdType of + ref -> make_ref(); + immed -> immed + end, async = garbage_collect(P, [{async, ReqId}]), receive {garbage_collect, ReqId, Result} -> @@ -2554,7 +2562,6 @@ gc_request_when_gc_disabled(Config) when is_list(Config) -> ok end, receive {garbage_collect, ReqId, true} -> ok end, - erts_debug:set_internal_state(available_internal_state, AIS), receive {'DOWN', M, process, P, _Reason} -> ok end, ok. diff --git a/erts/emulator/test/scheduler_SUITE.erl b/erts/emulator/test/scheduler_SUITE.erl index 7afb82a1b8..7eebbe8b19 100644 --- a/erts/emulator/test/scheduler_SUITE.erl +++ b/erts/emulator/test/scheduler_SUITE.erl @@ -894,11 +894,9 @@ adjust_schedulers_online() -> read_affinity(Data) -> Exp = "pid " ++ os:getpid() ++ "'s current affinity mask", - case string:tokens(Data, ":") of + case string:lexemes(Data, ":") of [Exp, DirtyAffinityStr] -> - AffinityStr = string:strip(string:strip(DirtyAffinityStr, - both, $ ), - both, $\n), + AffinityStr = string:trim(DirtyAffinityStr), case catch erlang:list_to_integer(AffinityStr, 16) of Affinity when is_integer(Affinity) -> Affinity; diff --git a/erts/emulator/test/system_info_SUITE.erl b/erts/emulator/test/system_info_SUITE.erl index 56522039da..fdf4aab24d 100644 --- a/erts/emulator/test/system_info_SUITE.erl +++ b/erts/emulator/test/system_info_SUITE.erl @@ -308,9 +308,9 @@ memory_test(_Config) -> mem_workers_call(MWs, fun () -> - list_to_atom("an ugly atom "++integer_to_list(erlang:system_info(scheduler_id))), - list_to_atom("another ugly atom "++integer_to_list(erlang:system_info(scheduler_id))), - list_to_atom("yet another ugly atom "++integer_to_list(erlang:system_info(scheduler_id))) + _ = list_to_atom("an ugly atom "++integer_to_list(erlang:system_info(scheduler_id))), + _ = list_to_atom("another ugly atom "++integer_to_list(erlang:system_info(scheduler_id))), + _ = list_to_atom("yet another ugly atom "++integer_to_list(erlang:system_info(scheduler_id))) end, []), cmp_memory(MWs, "new atoms"), diff --git a/erts/emulator/utils/beam_makeops b/erts/emulator/utils/beam_makeops index d7791d23fa..da994fae3e 100755 --- a/erts/emulator/utils/beam_makeops +++ b/erts/emulator/utils/beam_makeops @@ -19,7 +19,7 @@ # %CopyrightEnd% # use strict; -use vars qw($BEAM_FORMAT_NUMBER); +use vars qw($BEAM_FORMAT_NUMBER $GC_REGEXP); use constant COLD => 0; use constant WARM => 1; use constant HOT => 2; @@ -36,6 +36,7 @@ use constant PACK_CMD_LOOSE => '3'; use constant PACK_CMD_WIDE => '4'; $BEAM_FORMAT_NUMBER = undef; +$GC_REGEXP = undef; my $target = \&emulator_output; my $outdir = "."; # Directory for output files. @@ -77,6 +78,10 @@ my %num_specific; my %gen_to_spec; my %specific_op; +# The following hashes are used for error checking. +my %print_name; +my %specific_op_arity; + # Information about each specific operator. Key is the print name (e.g. get_list_xxy). # Value is a hash. my %spec_op_info; @@ -131,7 +136,10 @@ my $loader_types = "nprvlqo"; my $genop_types = $compiler_types . $loader_types; # -# Defines the argument types and their loaded size assuming no packing. +# Define the operand types and their loaded size assuming no packing. +# +# Those are the types that can be used in the definition of a specific +# instruction. # my %arg_size = ('r' => 0, # x(0) - x register zero 'x' => 1, # x(N), N > 0 - x register @@ -154,12 +162,35 @@ my %arg_size = ('r' => 0, # x(0) - x register zero 'A' => 1, # arity value 'P' => 1, # byte offset into tuple or stack 'Q' => 1, # like 'P', but packable - 'h' => 1, # character + 'h' => 1, # character (not used) 'l' => 1, # float reg 'q' => 1, # literal term ); # +# Define the types that may be used in a transformation rule. +# +# %pattern_type defines the types that may be used in a pattern +# on the left side. +# +# %construction_type defines the types that may be used when +# constructing a new instruction on the right side (a subset of +# the pattern types that are possible to construct). +# +my $pattern_types = "acdfjilnopqsuxy"; +my %pattern_type; +@pattern_type{split("", $pattern_types)} = (1) x length($pattern_types); + +my %construction_type; +foreach my $type (keys %pattern_type) { + $construction_type{$type} = 1 + if index($genop_types, $type) >= 0; +} +foreach my $makes_no_sense ('f', 'j', 'o', 'p', 'q') { + delete $construction_type{$makes_no_sense}; +} + +# # Generate bits. # my %type_bit; @@ -194,7 +225,8 @@ sub define_type_bit { define_type_bit('S', $type_bit{'d'}); define_type_bit('j', $type_bit{'f'} | $type_bit{'p'}); - # Aliases (for matching purposes). + # Aliases of 'u'. Those specify how to load the operand and + # what kind of packing can be done. define_type_bit('t', $type_bit{'u'}); define_type_bit('I', $type_bit{'u'}); define_type_bit('W', $type_bit{'u'}); @@ -279,9 +311,15 @@ if ($wordsize == 64) { # Add placeholders for built-in macros. # -$c_code{'IS_PACKED'} = ['$Expr',"built-in macro",('Expr')]; -$c_code{'ARG_POSITION'} = ['$Expr',"built-in macro",('Expr')]; -foreach my $name (keys %c_code) { +my %predef_macros = + (OPERAND_POSITION => ['Expr'], + IF => ['Expr','IfTrue','IfFalse'], + REFRESH_GEN_DEST => [], + ); +foreach my $name (keys %predef_macros) { + my @args = @{$predef_macros{$name}}; + my $body = join(':', map { '$' . $_ } @args); + $c_code{$name} = [$body,"built-in macro",@args], $c_code_used{$name} = 1; } @@ -359,8 +397,10 @@ while (<>) { # if (/^([\w_][\w\d_]+)=(.*)/) { no strict 'refs'; - my($name) = $1; - $$name = $2; + my $name = $1; + my $value = $2; + $value =~ s/;\s*$//; + $$name = $value; next; } @@ -1019,6 +1059,22 @@ sub parse_specific_op { my $key = "$name/$arity"; foreach my $args_ref (@res) { @args = @$args_ref; + my $arity = @args; + my $loc = "$ARGV($.)"; + if (defined $specific_op_arity{$name}) { + my($prev_arity,$loc) = @{$specific_op_arity{$name}}; + if ($arity != $prev_arity) { + error("$name defined with arity $arity, " . + "but previously defined with arity $prev_arity at $loc"); + } + } + $specific_op_arity{$name} = [$arity,$loc]; + my $print_name = print_name($name, @args); + if (defined $print_name{$print_name}) { + error("$name @args: already defined at " . + $print_name{$print_name}); + } + $print_name{$print_name} = $loc; push @{$specific_op{$key}}, [$name,$hotness,@args]; } @@ -1333,7 +1389,9 @@ sub cg_basic { # sub cg_combined_size { - my %params = (@_, pack_options => \@basic_pack_options); + my %params = (@_, + pack_options => \@basic_pack_options, + size_only => 1); $params{pack_options} = \@extended_pack_options if $params{first}; my($size) = code_gen(%params); @@ -1361,6 +1419,7 @@ sub code_gen { my %params = (extra_comments => '', offset => 0, inc => 0, + size_only => 0, @_); my $name = $params{name}; my $extra_comments = $params{extra_comments}; @@ -1393,6 +1452,7 @@ sub code_gen { my $need_block = 0; my $arg_offset = $offset; + my $has_gen_dest = 0; @args = map { s/[?]$//g; $_ } @args; foreach (@args) { my($this_size) = $arg_size{$_}; @@ -1403,6 +1463,7 @@ sub code_gen { "Eterm* dst_ptr = REG_TARGET_PTR(dst);\n"; push(@f, "*dst_ptr"); $this_size = $1; + $has_gen_dest = 1; last SWITCH; }; /^packed:[a-zA-z]:(\d):(.*)/ and do { @@ -1435,6 +1496,7 @@ sub code_gen { $var_decls .= "Eterm dst = " . arg_offset($arg_offset) . ";\n" . "Eterm* dst_ptr = REG_TARGET_PTR(dst);\n"; push(@f, "*dst_ptr"); + $has_gen_dest = 1; last SWITCH; }; defined $arg_size{$_} and do { @@ -1449,10 +1511,10 @@ sub code_gen { } # - # If the implementation is in beam_emu.c, there is nothing - # more to do. + # If the implementation is in beam_emu.c or if + # the caller only wants the size, we are done. # - unless (defined $c_code_ref) { + if (not defined $c_code_ref or $params{size_only}) { return ($size+1, undef, ''); } @@ -1517,9 +1579,36 @@ sub code_gen { "{", "$var_decls$body", "}", ""); + + # Make sure that $REFRESH_GEN_DEST() is used when a + # general destination ('d') may have been clobbered by + # a GC. + my $gc_error = verify_gc_code($code, $has_gen_dest); + if (defined $gc_error) { + warn $gc_error; + error("... from the body of $name at $where"); + } + + # Done. ($size+1, $code, $pack_spec); } +sub verify_gc_code { + my $code = shift; + my $has_gen_dest = shift; + + return unless $has_gen_dest; + + if ($code =~ /$GC_REGEXP/o) { + my $code_after_gc = substr($code, $+[0]); + unless ($code_after_gc =~ /dst_ptr = REG_TARGET_PTR/) { + return "pointer to destination register is invalid after GC -- " . + "use \$REFRESH_GEN_DEST()\n"; + } + } + return undef; +} + sub arg_offset { my $offset = shift; "I[" . ($offset+1) . "]"; @@ -1619,17 +1708,26 @@ sub expand_macro { } # Handle built-in macros. - if ($name eq 'ARG_POSITION') { + if ($name eq 'OPERAND_POSITION') { if ($body =~ /^I\[(\d+)\]$/) { $body = $1; } else { $body = 0; } - } elsif ($name eq 'IS_PACKED') { - $body = ($body =~ /^I\[\d+\]$/) ? 0 : 1; + } elsif ($name eq 'IF') { + my $expr = $new_bindings{Expr}; + my $bool = eval $expr; + if ($@ ne '') { + &error("bad expression '$expr' in \$IF()"); + } + my $part = $bool ? 'IfTrue' : 'IfFalse'; + $body = $new_bindings{$part}; + } elsif ($name eq 'REFRESH_GEN_DEST') { + $body = "dst_ptr = REG_TARGET_PTR(dst)"; } - # Wrap body if needed and return resul.t + + # Wrap body if needed and return result. $body = "do {\n$body\n} while (0)" if needs_do_wrapper($body); ($body,$rest); @@ -2156,12 +2254,19 @@ sub tr_parse_op { if (/^([a-z*]+)(.*)/) { $type = $1; $_ = $2; + error("$type: only a single type is allowed on right side of transformations") + if not $src and length($type) > 1; foreach (split('', $type)) { - error("bad type in $op") - unless defined $type_bit{$_} or $type eq '*'; - $_ eq 'r' and - error("$op: 'r' is not allowed in transformations") - } + next if $src and $type eq '*'; + error("$op: not a type") + unless defined $type_bit{$_}; + error("$op: the type '$_' is not allowed in transformations") + unless defined $pattern_type{$_}; + if (not $src) { + error("$op: type '$_' is not allowed on the right side of transformations") + unless defined $construction_type{$_}; + } + } } # Get an optional condition. (In source.) @@ -2194,10 +2299,18 @@ sub tr_parse_op { } # Get an optional value. (In destination.) - $type_val = $type eq 'x' ? 1023 : 0; + if ($type eq 'x') { + $type_val = 1023; + } elsif ($type eq 'a') { + $type_val = 'am_Empty'; + } else { + $type_val = 0; + } if (/^=(.*)/) { - error("value not allowed in source: $op") + error("$op: value not allowed in source") if $src; + error("$op: the type 'n' must not be given a value") + if $type eq 'n'; $type_val = $1; $_ = ''; } @@ -2207,13 +2320,16 @@ sub tr_parse_op { error("garbage '$_' after operand: $op") unless /^\s*$/; - # Test that destination has no conditions. + # Check the conditions. - unless ($src) { - error("condition not allowed in destination: $op") + if ($src) { + error("$op: the type '$type' is not allowed to be compared with a literal value") + if $cond and not $construction_type{$type}; + } else { + error("$op: condition not allowed in destination") if $cond; - error("variable name and type cannot be combined in destination: $op") - if $var && $type; + error("$op: variable name and type cannot be combined in destination") + if $var and $type; } ($var,$type,$type_val,$cond,$cond_val); diff --git a/erts/etc/common/erlexec.c b/erts/etc/common/erlexec.c index d61a3cbf95..08e326b33b 100644 --- a/erts/etc/common/erlexec.c +++ b/erts/etc/common/erlexec.c @@ -31,6 +31,7 @@ #ifdef __WIN32__ # include "erl_version.h" # include "init_file.h" +# include <Shlobj.h> #endif #define NO 0 @@ -1531,17 +1532,16 @@ static void get_parameters(int argc, char** argv) static void get_home(void) { - int len; - char tmpstr[MAX_PATH+1]; + wchar_t *profile; char* homedrive; char* homepath; homedrive = get_env("HOMEDRIVE"); homepath = get_env("HOMEPATH"); if (!homedrive || !homepath) { - if (len = GetWindowsDirectory(tmpstr,MAX_PATH)) { - home = emalloc(len+1); - strcpy(home,tmpstr); + if (SHGetKnownFolderPath(&FOLDERID_Profile, 0, NULL, &profile) == S_OK) { + home = utf16_to_utf8(profile); + /* CoTaskMemFree(profile); */ } else error("HOMEDRIVE or HOMEPATH is not set and GetWindowsDir failed"); } else { diff --git a/erts/include/internal/erl_printf.h b/erts/include/internal/erl_printf.h index f180a53f18..7e9807f6a8 100644 --- a/erts/include/internal/erl_printf.h +++ b/erts/include/internal/erl_printf.h @@ -44,6 +44,7 @@ struct erts_dsprintf_buf_t_ { typedef int (*fmtfn_t)(void*, char*, size_t); int erts_write_fd(void *vfdp, char* buf, size_t len); +int erts_write_fp(void *vfdp, char* buf, size_t len); int erts_write_ds(void *vdsbufp, char* buf, size_t len); int erts_printf(const char *, ...); diff --git a/erts/lib_src/common/erl_printf.c b/erts/lib_src/common/erl_printf.c index 9031a4c5b7..1de0f81e84 100644 --- a/erts/lib_src/common/erl_printf.c +++ b/erts/lib_src/common/erl_printf.c @@ -143,8 +143,8 @@ write_f_add_cr(void *vfp, char* buf, size_t len) return len; } -static int -write_f(void *vfp, char* buf, size_t len) +int +erts_write_fp(void *vfp, char* buf, size_t len) { ASSERT(vfp); #ifdef PUTC_ON_SMALL_WRITES @@ -253,7 +253,7 @@ erts_printf(const char *format, ...) FLOCKFILE(stdout); res = erts_printf_format(erts_printf_add_cr_to_stdout ? write_f_add_cr - : write_f, + : erts_write_fp, (void *) stdout, (char *) format, arglist); @@ -281,7 +281,7 @@ erts_fprintf(FILE *filep, const char *format, ...) else if (erts_printf_add_cr_to_stderr && filep == stderr) fmt_f = write_f_add_cr; else - fmt_f = write_f; + fmt_f = erts_write_fp; FLOCKFILE(filep); res = erts_printf_format(fmt_f,(void *)filep,(char *)format,arglist); FUNLOCKFILE(filep); @@ -386,7 +386,7 @@ erts_vprintf(const char *format, va_list arglist) errno = 0; res = erts_printf_format(erts_printf_add_cr_to_stdout ? write_f_add_cr - : write_f, + : erts_write_fp, (void *) stdout, (char *) format, arglist); @@ -410,7 +410,7 @@ erts_vfprintf(FILE *filep, const char *format, va_list arglist) else if (erts_printf_add_cr_to_stderr && filep == stderr) fmt_f = write_f_add_cr; else - fmt_f = write_f; + fmt_f = erts_write_fp; res = erts_printf_format(fmt_f,(void *)filep,(char *)format,arglist); } return res; diff --git a/erts/preloaded/ebin/zlib.beam b/erts/preloaded/ebin/zlib.beam Binary files differindex 5048bdb846..4ad5f37434 100644 --- a/erts/preloaded/ebin/zlib.beam +++ b/erts/preloaded/ebin/zlib.beam diff --git a/erts/preloaded/src/erlang.erl b/erts/preloaded/src/erlang.erl index f743b7d26b..80bceae506 100644 --- a/erts/preloaded/src/erlang.erl +++ b/erts/preloaded/src/erlang.erl @@ -427,9 +427,11 @@ binary_to_term(_Binary) -> erlang:nif_error(undefined). %% binary_to_term/2 --spec binary_to_term(Binary, Opts) -> term() when +-spec binary_to_term(Binary, Opts) -> term() | {term(), Used} when Binary :: ext_binary(), - Opts :: [safe]. + Opt :: safe | used, + Opts :: [Opt], + Used :: pos_integer(). binary_to_term(_Binary, _Opts) -> erlang:nif_error(undefined). diff --git a/erts/preloaded/src/zlib.erl b/erts/preloaded/src/zlib.erl index 3170ab6351..a4ef42204d 100644 --- a/erts/preloaded/src/zlib.erl +++ b/erts/preloaded/src/zlib.erl @@ -20,7 +20,8 @@ -module(zlib). --export([open/0,close/1,deflateInit/1,deflateInit/2,deflateInit/6, +-export([open/0,close/1,set_controlling_process/2, + deflateInit/1,deflateInit/2,deflateInit/6, deflateSetDictionary/2,deflateReset/1,deflateParams/3, deflate/2,deflate/3,deflateEnd/1, inflateInit/1,inflateInit/2,inflateInit/3, @@ -128,6 +129,14 @@ close(Z) -> close_nif(_Z) -> erlang:nif_error(undef). +-spec set_controlling_process(Z, Pid) -> 'ok' when + Z :: zstream(), + Pid :: pid(). +set_controlling_process(Z, Pid) -> + set_controller_nif(Z, Pid). +set_controller_nif(_Z, _Pid) -> + erlang:nif_error(undef). + -spec deflateInit(Z) -> 'ok' when Z :: zstream(). deflateInit(Z) -> @@ -179,14 +188,13 @@ deflateReset_nif(_Z) -> deflateParams(Z, Level0, Strategy0) -> Level = arg_level(Level0), Strategy = arg_strategy(Strategy0), + Progress = deflate(Z, <<>>, sync), case deflateParams_nif(Z, Level, Strategy) of - buf_error -> - %% We had data left in the pipe; flush everything and stash it away - %% for the next deflate call before trying again. - Output = deflate(Z, <<>>, full), - save_progress(Z, deflate, Output), - deflateParams_nif(Z, Level, Strategy); - Any -> Any + ok -> + save_progress(Z, deflate, Progress), + ok; + Other -> + Other end. deflateParams_nif(_Z, _Level, _Strategy) -> erlang:nif_error(undef). diff --git a/erts/start_scripts/Makefile b/erts/start_scripts/Makefile index e730422ed8..5a47ad9616 100644 --- a/erts/start_scripts/Makefile +++ b/erts/start_scripts/Makefile @@ -171,7 +171,7 @@ $(ERL_TOP)/bin/no_dot_erlang.script: $(ERLC) $(SCRIPT_PATH) +no_warn_sasl +otp_build +no_dot_erlang -o $@ $(SS_ROOT)/no_dot_erlang.rel ) ## Special target used from system/build/Makefile for source code release bootstrap. -bootstrap_scripts: $(SS_ROOT)/start_clean.rel +bootstrap_scripts: $(SS_ROOT)/start_clean.rel $(SS_ROOT)/no_dot_erlang.rel $(V_at)$(INSTALL_DIR) $(TESTROOT)/bin $(V_at)$(INSTALL_DIR) $(SS_TMP) $(V_at)( cd $(SS_TMP) && \ diff --git a/erts/vsn.mk b/erts/vsn.mk index 3d7ff2db66..220c3b5f6c 100644 --- a/erts/vsn.mk +++ b/erts/vsn.mk @@ -18,7 +18,7 @@ # %CopyrightEnd% # -VSN = 9.1.2 +VSN = 9.1.4 # Port number 4365 in 4.2 # Port number 4366 in 4.3 diff --git a/lib/common_test/src/cth_log_redirect.erl b/lib/common_test/src/cth_log_redirect.erl index 8b29d0f96d..1c55e17686 100644 --- a/lib/common_test/src/cth_log_redirect.erl +++ b/lib/common_test/src/cth_log_redirect.erl @@ -121,8 +121,8 @@ handle_event({_Type,GL,_Msg}, #eh_state{handle_remote_events = false} = State) when node(GL) /= node() -> {ok, State}; handle_event(Event, #eh_state{log_func = LogFunc} = State) -> - case lists:keyfind(sasl, 1, application:which_applications()) of - false -> + case whereis(sasl_sup) of + undefined -> sasl_not_started; _Else -> {ok, ErrLogType} = application:get_env(sasl, errlog_type), diff --git a/lib/compiler/internal_doc/cerl-notes.md b/lib/compiler/internal_doc/cerl-notes.md new file mode 100644 index 0000000000..705fe8f42d --- /dev/null +++ b/lib/compiler/internal_doc/cerl-notes.md @@ -0,0 +1,75 @@ +Some notes on the cerl modules +============================== + +Maps in cerl_clauses:match/3 +---------------------------- + +Not much optimization is done for maps in `cerl_clauses:match/3`. + +The reason is that the inliner (`cerl_inline`) was not designed for +data types that depend on variables bound in the enclosing environment +(for example, the keys for maps). If we attempt to extend the +optimizations for maps similar to the optimizations for the other data +types, the inliner will crash for certain code. Here is an example of +code that would crash the inliner: + + t() -> + f(key). + + f(K) -> + case #{K => value} of + #{key := V} -> V + end. + +The reason for the crash is that the inliner works in several +passes and calls `cerl_clauses:match/3` twice. The inliner will +assume that the same result will be returned both times, but +for this example different bindings will be returned. + +Here is the rough outline of what happens: + +* The first time `cerl_clauses:match/3` will be asked to match the +pattern `#{K := V}` against `#{key => value}`. It cannot say more +than that the pattern *may* match. + +* Now the inliner will add the bindings to body of the clause (which +is simply `V`). In this case, the bindings are still empty, so +nothing is added. + +* The inliner will now do some substitutions and renaming. The +variable `K` will be replaced with `key`. + +* The next time `cerl_clauses:match/3` is called, it will be asked to +match the pattern `#{key := V}` against `#{key => value#}`. In this +case, there will be a match and the bindings can be extended with +`{V,value}`. + +* The inliner will see that the entire case can be removed. Only +the body for the clause needs to be kept. + +Thus, after inlining the function `t/0` will look like this: + + t() -> + V. + +The problem here is that the inliner assumed that the bindings from +the first and second call to `cer_clauses:match/3` would be the same. +It used the empty bindings from the first call for the body. + +The correct way would be to add the bindings from the second call: + + t() -> + let V = value in V. + +### How to fix this ### + +The inliner will need to call `cerl_clauses:match/3` after doing +all substitutions and renaming. It is not clear to me how difficult +that would be. I assume that the inliner is written the way it is +for a good reason. That means that switching the order things are +done would lead to correctness and/or performance problems. + +### What must also be done to fix this ### + +`cerl_inline:make_template/3` must be extended to create a template +for maps. That is relatively straightforward. diff --git a/lib/compiler/src/Makefile b/lib/compiler/src/Makefile index 9b22e5197b..9e96147787 100644 --- a/lib/compiler/src/Makefile +++ b/lib/compiler/src/Makefile @@ -93,16 +93,14 @@ MODULES = \ v3_codegen \ v3_core \ v3_kernel \ - v3_kernel_pp \ - v3_life + v3_kernel_pp BEAM_H = $(wildcard ../priv/beam_h/*.h) HRL_FILES= \ beam_disasm.hrl \ core_parse.hrl \ - v3_kernel.hrl \ - v3_life.hrl + v3_kernel.hrl YRL_FILE = core_parse.yrl @@ -187,7 +185,7 @@ release_docs_spec: # ---------------------------------------------------- $(EBIN)/beam_disasm.beam: $(EGEN)/beam_opcodes.hrl beam_disasm.hrl -$(EBIN)/beam_listing.beam: v3_life.hrl +$(EBIN)/beam_listing.beam: core_parse.hrl v3_kernel.hrl $(EBIN)/beam_validator.beam: beam_disasm.hrl $(EBIN)/cerl.beam: core_parse.hrl $(EBIN)/compile.beam: core_parse.hrl ../../stdlib/include/erl_compile.hrl @@ -200,8 +198,7 @@ $(EBIN)/sys_core_dsetel.beam: core_parse.hrl $(EBIN)/sys_core_fold.beam: core_parse.hrl $(EBIN)/sys_core_fold_lists.beam: core_parse.hrl $(EBIN)/sys_core_inline.beam: core_parse.hrl -$(EBIN)/v3_codegen.beam: v3_life.hrl +$(EBIN)/v3_codegen.beam: v3_kernel.hrl $(EBIN)/v3_core.beam: core_parse.hrl $(EBIN)/v3_kernel.beam: core_parse.hrl v3_kernel.hrl $(EBIN)/v3_kernel_pp.beam: v3_kernel.hrl -$(EBIN)/v3_life.beam: v3_kernel.hrl v3_life.hrl diff --git a/lib/compiler/src/beam_listing.erl b/lib/compiler/src/beam_listing.erl index 94b47cf568..9422f9a78a 100644 --- a/lib/compiler/src/beam_listing.erl +++ b/lib/compiler/src/beam_listing.erl @@ -23,14 +23,12 @@ -include("core_parse.hrl"). -include("v3_kernel.hrl"). --include("v3_life.hrl"). -import(lists, [foreach/2]). -type code() :: cerl:c_module() | beam_utils:module_code() | #k_mdef{} - | {module(),_,_,_} %v3_life | [_]. %form-based format -spec module(file:io_device(), code()) -> 'ok'. @@ -42,13 +40,9 @@ module(File, #k_mdef{}=Kern) -> %% This is a kernel module. io:put_chars(File, v3_kernel_pp:format(Kern)); %%io:put_chars(File, io_lib:format("~p~n", [Kern])); -module(File, {Mod,Exp,Attr,Kern}) -> - %% This is output from beam_life (v3). - io:fwrite(File, "~w.~n~p.~n~p.~n", [Mod,Exp,Attr]), - foreach(fun (F) -> function(File, F) end, Kern); module(Stream, {Mod,Exp,Attr,Code,NumLabels}) -> - %% This is output from beam_codegen. - io:format(Stream, "{module, ~p}. %% version = ~w\n", + %% This is output from v3_codegen. + io:format(Stream, "{module, ~p}. %% version = ~w\n", [Mod, beam_opcodes:format_number()]), io:format(Stream, "\n{exports, ~p}.\n", [Exp]), io:format(Stream, "\n{attributes, ~p}.\n", [Attr]), @@ -68,60 +62,3 @@ format_asm([{label,L}|Is]) -> format_asm([I|Is]) -> [io_lib:format(" ~p", [I]),".\n"|format_asm(Is)]; format_asm([]) -> []. - -function(File, {function,Name,Arity,Args,Body,Vdb,_Anno}) -> - io:nl(File), - io:format(File, "function ~p/~p.\n", [Name,Arity]), - io:format(File, " ~p.\n", [Args]), - print_vdb(File, Vdb), - put(beam_listing_nl, false), - nl(File), - foreach(fun(F) -> format(File, F, []) end, Body), - nl(File), - erase(beam_listing_nl). - -format(File, #l{ke=Ke,i=I,vdb=Vdb}, Ind) -> - nl(File), - ind_format(File, Ind, "~p ", [I]), - print_vdb(File, Vdb), - nl(File), - format(File, Ke, Ind); -format(File, Tuple, Ind) when is_tuple(Tuple) -> - ind_format(File, Ind, "{", []), - format_list(File, tuple_to_list(Tuple), [$\s|Ind]), - ind_format(File, Ind, "}", []); -format(File, List, Ind) when is_list(List) -> - ind_format(File, Ind, "[", []), - format_list(File, List, [$\s|Ind]), - ind_format(File, Ind, "]", []); -format(File, F, Ind) -> - ind_format(File, Ind, "~p", [F]). - -format_list(File, [F], Ind) -> - format(File, F, Ind); -format_list(File, [F|Fs], Ind) -> - format(File, F, Ind), - ind_format(File, Ind, ",", []), - format_list(File, Fs, Ind); -format_list(_, [], _) -> ok. - - -print_vdb(File, [{Var,F,E}|Vs]) -> - io:format(File, "~p:~p..~p ", [Var,F,E]), - print_vdb(File, Vs); -print_vdb(_, []) -> ok. - -ind_format(File, Ind, Format, Args) -> - case get(beam_listing_nl) of - true -> - put(beam_listing_nl, false), - io:put_chars(File, Ind); - false -> ok - end, - io:format(File, Format, Args). - -nl(File) -> - case put(beam_listing_nl, true) of - true -> ok; - false -> io:nl(File) - end. diff --git a/lib/compiler/src/cerl_clauses.erl b/lib/compiler/src/cerl_clauses.erl index 7d6518c3c6..fa5104c01b 100644 --- a/lib/compiler/src/cerl_clauses.erl +++ b/lib/compiler/src/cerl_clauses.erl @@ -353,6 +353,8 @@ match(P, E, Bs) -> map -> %% The most we can do is to say "definitely no match" if a %% map pattern is matched against non-map data. + %% (Note: See the document internal_doc/cerl-notes.md for + %% information why we don't try to do more here.) case E of any -> {false, Bs}; diff --git a/lib/compiler/src/compile.erl b/lib/compiler/src/compile.erl index 1b359d1e59..b0d0da0f66 100644 --- a/lib/compiler/src/compile.erl +++ b/lib/compiler/src/compile.erl @@ -733,8 +733,6 @@ kernel_passes() -> ?pass(v3_kernel), {iff,dkern,{listing,"kernel"}}, {iff,'to_kernel',{done,"kernel"}}, - {pass,v3_life}, - {iff,dlife,{listing,"life"}}, {pass,v3_codegen}, {iff,dcg,{listing,"codegen"}} | asm_passes()]. @@ -1450,16 +1448,16 @@ beam_asm(Code0, #compile{ifile=File,extra_chunks=ExtraChunks,options=CompilerOpt {ok,DebugInfo,Opts0} -> Opts1 = [O || O <- Opts0, effects_code_generation(O)], Chunks = [{<<"Dbgi">>, DebugInfo} | ExtraChunks], - CompileInfo = compile_info(File, Opts1), + CompileInfo = compile_info(File, CompilerOpts, Opts1), {ok,Code} = beam_asm:module(Code0, Chunks, CompileInfo, CompilerOpts), {ok,Code,St#compile{abstract_code=[]}}; {error,Es} -> {error,St#compile{errors=St#compile.errors ++ [{File,Es}]}} end. -compile_info(File, Opts) -> - IsSlim = member(slim, Opts), - IsDeterministic = member(deterministic, Opts), +compile_info(File, CompilerOpts, Opts) -> + IsSlim = member(slim, CompilerOpts), + IsDeterministic = member(deterministic, CompilerOpts), Info0 = proplists:get_value(compile_info, Opts, []), Info1 = case paranoid_absname(File) of @@ -1947,7 +1945,6 @@ pre_load() -> sys_core_fold, v3_codegen, v3_core, - v3_kernel, - v3_life], + v3_kernel], _ = code:ensure_modules_loaded(L), ok. diff --git a/lib/compiler/src/compiler.app.src b/lib/compiler/src/compiler.app.src index 703cf1d1b8..cf32fd251c 100644 --- a/lib/compiler/src/compiler.app.src +++ b/lib/compiler/src/compiler.app.src @@ -68,8 +68,7 @@ v3_codegen, v3_core, v3_kernel, - v3_kernel_pp, - v3_life + v3_kernel_pp ]}, {registered, []}, {applications, [kernel, stdlib]}, diff --git a/lib/compiler/src/v3_codegen.erl b/lib/compiler/src/v3_codegen.erl index e705aefb96..006a6a82d2 100644 --- a/lib/compiler/src/v3_codegen.erl +++ b/lib/compiler/src/v3_codegen.erl @@ -19,25 +19,6 @@ %% %% Purpose : Code generator for Beam. -%% The following assumptions have been made: -%% -%% 1. Matches, i.e. things with {match,M,Ret} wrappers, only return -%% values; no variables are exported. If the match would have returned -%% extra variables then these have been transformed to multiple return -%% values. -%% -%% 2. All BIF's called in guards are gc-safe so there is no need to -%% put thing on the stack in the guard. While this would in principle -%% work it would be difficult to keep track of the stack depth when -%% trimming. -%% -%% The code generation uses variable lifetime information added by -%% the v3_life module to save variables, allocate registers and -%% move registers to the stack when necessary. -%% -%% We try to use a consistent variable name scheme throughout. The -%% StackReg record is always called Bef,Int<n>,Aft. - -module(v3_codegen). %% The main interface. @@ -45,12 +26,14 @@ -import(lists, [member/2,keymember/3,keysort/2,keydelete/3, append/1,flatmap/2,filter/2,foldl/3,foldr/3,mapfoldl/3, - sort/1,reverse/1,reverse/2]). --import(v3_life, [vdb_find/2]). + sort/1,reverse/1,reverse/2,map/2]). +-import(ordsets, [add_element/2,intersection/2,union/2]). -%%-compile([export_all]). +-include("v3_kernel.hrl"). --include("v3_life.hrl"). +%% These are not defined in v3_kernel.hrl. +get_kanno(Kthing) -> element(2, Kthing). +set_kanno(Kthing, Anno) -> setelement(2, Kthing, Anno). %% Main codegen structure. -record(cg, {lcount=1, %Label counter @@ -61,38 +44,273 @@ functable=#{}, %Map of local functions: {Name,Arity}=>Label in_catch=false, %Inside a catch or not. need_frame, %Need a stack frame. - ultimate_failure %Label for ultimate match failure. - }). + ultimate_failure, %Label for ultimate match failure. + ctx %Match context. + }). %% Stack/register state record. -record(sr, {reg=[], %Register table stk=[], %Stack table res=[]}). %Reserved regs: [{reserved,I,V}] --type life_module() :: {module(),_,_,[_]}. +%% Internal records. +-record(cg_need_heap, {anno=[] :: term(), + h=0 :: integer()}). +-record(cg_block, {anno=[] :: term(), + es=[] :: [term()]}). + +-type vdb_entry() :: {atom(),non_neg_integer(),non_neg_integer()}. + +-record(l, {i=0 :: non_neg_integer(), %Op number + vdb=[] :: [vdb_entry()], %Variable database + a=[] :: [term()]}). %Core annotation --spec module(life_module(), [compile:option()]) -> {'ok',beam_asm:module_code()}. +-spec module(#k_mdef{}, [compile:option()]) -> {'ok',beam_asm:module_code()}. -module({Mod,Exp,Attr,Forms}, _Options) -> - {Fs,St} = functions(Forms, {atom,Mod}), - {ok,{Mod,Exp,Attr,Fs,St#cg.lcount}}. +module(#k_mdef{name=Mod,exports=Es,attributes=Attr,body=Forms}, _Opts) -> + {Asm,St} = functions(Forms, {atom,Mod}), + {ok,{Mod,Es,Attr,Asm,St#cg.lcount}}. functions(Forms, AtomMod) -> mapfoldl(fun (F, St) -> function(F, AtomMod, St) end, #cg{lcount=1}, Forms). -function({function,Name,Arity,Asm0,Vb,Vdb,Anno}, AtomMod, St0) -> +function(#k_fdef{anno=#k{a=Anno},func=Name,arity=Arity, + vars=As,body=Kb}, AtomMod, St0) -> try - {Asm,EntryLabel,St} = cg_fun(Vb, Asm0, Vdb, AtomMod, - {Name,Arity}, Anno, St0), - Func = {function,Name,Arity,EntryLabel,Asm}, - {Func,St} + %% Annotate kernel records with variable usage. + #k_match{} = Kb, %Assertion. + Vdb0 = init_vars(As), + {Body,_,Vdb} = body(Kb, 1, Vdb0), + + %% Generate the BEAM assembly code. + {Asm,EntryLabel,St} = cg_fun(Body, As, Vdb, AtomMod, + {Name,Arity}, Anno, St0), + Func = {function,Name,Arity,EntryLabel,Asm}, + {Func,St} catch - Class:Error -> - Stack = erlang:get_stacktrace(), - io:fwrite("Function: ~w/~w\n", [Name,Arity]), - erlang:raise(Class, Error, Stack) + Class:Error -> + Stack = erlang:get_stacktrace(), + io:fwrite("Function: ~w/~w\n", [Name,Arity]), + erlang:raise(Class, Error, Stack) end. +%% This pass creates beam format annotated with variable lifetime +%% information. Each thing is given an index and for each variable we +%% store the first and last index for its occurrence. The variable +%% database, VDB, attached to each thing is only relevant internally +%% for that thing. +%% +%% For nested things like matches the numbering continues locally and +%% the VDB for that thing refers to the variable usage within that +%% thing. Variables which live through a such a thing are internally +%% given a very large last index. Internally the indexes continue +%% after the index of that thing. This creates no problems as the +%% internal variable info never escapes and externally we only see +%% variable which are alive both before or after. +%% +%% This means that variables never "escape" from a thing and the only +%% way to get values from a thing is to "return" them, with 'break' or +%% 'return'. Externally these values become the return values of the +%% thing. This is no real limitation as most nested things have +%% multiple threads so working out a common best variable usage is +%% difficult. + +%% body(Kbody, I, Vdb) -> {[Expr],MaxI,Vdb}. +%% Handle a body. + +body(#k_seq{arg=Ke,body=Kb}, I, Vdb0) -> + %%ok = io:fwrite("life ~w:~p~n", [?LINE,{Ke,I,Vdb0}]), + A = get_kanno(Ke), + Vdb1 = use_vars(union(A#k.us, A#k.ns), I, Vdb0), + {Es,MaxI,Vdb2} = body(Kb, I+1, Vdb1), + E = expr(Ke, I, Vdb2), + {[E|Es],MaxI,Vdb2}; +body(Ke, I, Vdb0) -> + %%ok = io:fwrite("life ~w:~p~n", [?LINE,{Ke,I,Vdb0}]), + A = get_kanno(Ke), + Vdb1 = use_vars(union(A#k.us, A#k.ns), I, Vdb0), + E = expr(Ke, I, Vdb1), + {[E],I,Vdb1}. + +%% expr(Kexpr, I, Vdb) -> Expr. + +expr(#k_test{anno=A}=Test, I, _Vdb) -> + Test#k_test{anno=#l{i=I,a=A#k.a}}; +expr(#k_call{anno=A}=Call, I, _Vdb) -> + Call#k_call{anno=#l{i=I,a=A#k.a}}; +expr(#k_enter{anno=A}=Enter, I, _Vdb) -> + Enter#k_enter{anno=#l{i=I,a=A#k.a}}; +expr(#k_bif{anno=A}=Bif, I, _Vdb) -> + Bif#k_bif{anno=#l{i=I,a=A#k.a}}; +expr(#k_match{anno=A,body=Kb,ret=Rs}, I, Vdb) -> + %% Work out imported variables which need to be locked. + Mdb = vdb_sub(I, I+1, Vdb), + M = match(Kb, A#k.us, I+1, Mdb), + L = #l{i=I,vdb=use_vars(A#k.us, I+1, Mdb),a=A#k.a}, + #k_match{anno=L,body=M,ret=Rs}; +expr(#k_guard_match{anno=A,body=Kb,ret=Rs}, I, Vdb) -> + %% Work out imported variables which need to be locked. + Mdb = vdb_sub(I, I+1, Vdb), + M = match(Kb, A#k.us, I+1, Mdb), + L = #l{i=I,vdb=use_vars(A#k.us, I+1, Mdb),a=A#k.a}, + #k_guard_match{anno=L,body=M,ret=Rs}; +expr(#k_protected{}=Protected, I, Vdb) -> + protected(Protected, I, Vdb); +expr(#k_try{anno=A,arg=Ka,vars=Vs,body=Kb,evars=Evs,handler=Kh}=Try, I, Vdb) -> + %% Lock variables that are alive before the catch and used afterwards. + %% Don't lock variables that are only used inside the try. + Tdb0 = vdb_sub(I, I+1, Vdb), + %% This is the tricky bit. Lock variables in Arg that are used in + %% the body and handler. Add try tag 'variable'. + Ab = get_kanno(Kb), + Ah = get_kanno(Kh), + Tdb1 = use_vars(union(Ab#k.us, Ah#k.us), I+3, Tdb0), + Tdb2 = vdb_sub(I, I+2, Tdb1), + Vnames = fun (Kvar) -> Kvar#k_var.name end, %Get the variable names + {Aes,_,Adb} = body(Ka, I+2, add_var({catch_tag,I+1}, I+1, locked, Tdb2)), + {Bes,_,Bdb} = body(Kb, I+4, new_vars(sort(map(Vnames, Vs)), I+3, Tdb2)), + {Hes,_,Hdb} = body(Kh, I+4, new_vars(sort(map(Vnames, Evs)), I+3, Tdb2)), + L = #l{i=I,vdb=Tdb1,a=A#k.a}, + Try#k_try{anno=L, + arg=#cg_block{es=Aes,anno=#l{i=I+1,vdb=Adb,a=[]}}, + vars=Vs,body=#cg_block{es=Bes,anno=#l{i=I+3,vdb=Bdb,a=[]}}, + evars=Evs,handler=#cg_block{es=Hes,anno=#l{i=I+3,vdb=Hdb,a=[]}}}; +expr(#k_try_enter{anno=A,arg=Ka,vars=Vs,body=Kb,evars=Evs,handler=Kh}, I, Vdb) -> + %% Lock variables that are alive before the catch and used afterwards. + %% Don't lock variables that are only used inside the try. + Tdb0 = vdb_sub(I, I+1, Vdb), + %% This is the tricky bit. Lock variables in Arg that are used in + %% the body and handler. Add try tag 'variable'. + Ab = get_kanno(Kb), + Ah = get_kanno(Kh), + Tdb1 = use_vars(union(Ab#k.us, Ah#k.us), I+3, Tdb0), + Tdb2 = vdb_sub(I, I+2, Tdb1), + Vnames = fun (Kvar) -> Kvar#k_var.name end, %Get the variable names + {Aes,_,Adb} = body(Ka, I+2, add_var({catch_tag,I+1}, I+1, 1000000, Tdb2)), + {Bes,_,Bdb} = body(Kb, I+4, new_vars(sort(map(Vnames, Vs)), I+3, Tdb2)), + {Hes,_,Hdb} = body(Kh, I+4, new_vars(sort(map(Vnames, Evs)), I+3, Tdb2)), + L = #l{i=I,vdb=Tdb1,a=A#k.a}, + #k_try_enter{anno=L, + arg=#cg_block{es=Aes,anno=#l{i=I+1,vdb=Adb,a=[]}}, + vars=Vs,body=#cg_block{es=Bes,anno=#l{i=I+3,vdb=Bdb,a=[]}}, + evars=Evs,handler=#cg_block{es=Hes,anno=#l{i=I+3,vdb=Hdb,a=[]}}}; +expr(#k_catch{anno=A,body=Kb}=Catch, I, Vdb) -> + %% Lock variables that are alive before the catch and used afterwards. + %% Don't lock variables that are only used inside the catch. + %% Add catch tag 'variable'. + Cdb0 = vdb_sub(I, I+1, Vdb), + {Es,_,Cdb1} = body(Kb, I+1, add_var({catch_tag,I}, I, locked, Cdb0)), + L = #l{i=I,vdb=Cdb1,a=A#k.a}, + Catch#k_catch{anno=L,body=#cg_block{es=Es}}; +expr(#k_receive{anno=A,var=V,body=Kb,action=Ka}=Recv, I, Vdb) -> + %% Work out imported variables which need to be locked. + Rdb = vdb_sub(I, I+1, Vdb), + M = match(Kb, add_element(V#k_var.name, A#k.us), I+1, + new_vars([V#k_var.name], I, Rdb)), + {Tes,_,Adb} = body(Ka, I+1, Rdb), + Le = #l{i=I,vdb=use_vars(A#k.us, I+1, Vdb),a=A#k.a}, + Recv#k_receive{anno=Le,body=M, + action=#cg_block{anno=#l{i=I+1,vdb=Adb,a=[]},es=Tes}}; +expr(#k_receive_accept{anno=A}, I, _Vdb) -> + #k_receive_accept{anno=#l{i=I,a=A#k.a}}; +expr(#k_receive_next{anno=A}, I, _Vdb) -> + #k_receive_next{anno=#l{i=I,a=A#k.a}}; +expr(#k_put{anno=A}=Put, I, _Vdb) -> + Put#k_put{anno=#l{i=I,a=A#k.a}}; +expr(#k_break{anno=A}=Break, I, _Vdb) -> + Break#k_break{anno=#l{i=I,a=A#k.a}}; +expr(#k_guard_break{anno=A}=Break, I, Vdb) -> + Locked = [V || {V,_,_} <- Vdb], + L = #l{i=I,a=A#k.a}, + Break#k_guard_break{anno=L,locked=Locked}; +expr(#k_return{anno=A}=Ret, I, _Vdb) -> + Ret#k_return{anno=#l{i=I,a=A#k.a}}. + +%% protected(Kprotected, I, Vdb) -> Protected. +%% Only used in guards. + +protected(#k_protected{anno=A,arg=Ts}=Prot, I, Vdb) -> + %% Lock variables that are alive before try and used afterwards. + %% Don't lock variables that are only used inside the protected + %% expression. + Pdb0 = vdb_sub(I, I+1, Vdb), + {T,MaxI,Pdb1} = body(Ts, I+1, Pdb0), + Pdb2 = use_vars(A#k.ns, MaxI+1, Pdb1), %Save "return" values + Prot#k_protected{arg=T,anno=#l{i=I,a=A#k.a,vdb=Pdb2}}. + +%% match(Kexpr, [LockVar], I, Vdb) -> Expr. +%% Convert match tree to old format. + +match(#k_alt{anno=A,first=Kf,then=Kt}, Ls, I, Vdb0) -> + Vdb1 = use_vars(union(A#k.us, Ls), I, Vdb0), + F = match(Kf, Ls, I+1, Vdb1), + T = match(Kt, Ls, I+1, Vdb1), + #k_alt{anno=[],first=F,then=T}; +match(#k_select{anno=A,var=V,types=Kts}=Select, Ls0, I, Vdb0) -> + Vanno = get_kanno(V), + Ls1 = case member(no_usage, Vanno) of + false -> add_element(V#k_var.name, Ls0); + true -> Ls0 + end, + Vdb1 = use_vars(union(A#k.us, Ls1), I, Vdb0), + Ts = [type_clause(Tc, Ls1, I+1, Vdb1) || Tc <- Kts], + Select#k_select{anno=[],types=Ts}; +match(#k_guard{anno=A,clauses=Kcs}, Ls, I, Vdb0) -> + Vdb1 = use_vars(union(A#k.us, Ls), I, Vdb0), + Cs = [guard_clause(G, Ls, I+1, Vdb1) || G <- Kcs], + #k_guard{anno=[],clauses=Cs}; +match(Other, Ls, I, Vdb0) -> + Vdb1 = use_vars(Ls, I, Vdb0), + {B,_,Vdb2} = body(Other, I+1, Vdb1), + Le = #l{i=I,vdb=Vdb2,a=[]}, + #cg_block{anno=Le,es=B}. + +type_clause(#k_type_clause{anno=A,type=T,values=Kvs}, Ls, I, Vdb0) -> + %%ok = io:format("life ~w: ~p~n", [?LINE,{T,Kvs}]), + Vdb1 = use_vars(union(A#k.us, Ls), I+1, Vdb0), + Vs = [val_clause(Vc, Ls, I+1, Vdb1) || Vc <- Kvs], + #k_type_clause{anno=[],type=T,values=Vs}. + +val_clause(#k_val_clause{anno=A,val=V,body=Kb}, Ls0, I, Vdb0) -> + New = (get_kanno(V))#k.ns, + Bus = (get_kanno(Kb))#k.us, + %%ok = io:format("Ls0 = ~p, Used=~p\n New=~p, Bus=~p\n", [Ls0,Used,New,Bus]), + Ls1 = union(intersection(New, Bus), Ls0), %Lock for safety + Vdb1 = use_vars(union(A#k.us, Ls1), I+1, new_vars(New, I, Vdb0)), + B = match(Kb, Ls1, I+1, Vdb1), + Le = #l{i=I,vdb=use_vars(Bus, I+1, Vdb1),a=A#k.a}, + #k_val_clause{anno=Le,val=V,body=B}. + +guard_clause(#k_guard_clause{anno=A,guard=Kg,body=Kb}, Ls, I, Vdb0) -> + Vdb1 = use_vars(union(A#k.us, Ls), I+2, Vdb0), + Gdb = vdb_sub(I+1, I+2, Vdb1), + G = protected(Kg, I+1, Gdb), + B = match(Kb, Ls, I+2, Vdb1), + Le = #l{i=I,vdb=use_vars((get_kanno(Kg))#k.us, I+2, Vdb1),a=A#k.a}, + #k_guard_clause{anno=Le,guard=G,body=B}. + + +%% Here follows the code generator pass. +%% +%% The following assumptions have been made: +%% +%% 1. Matches, i.e. things with {match,M,Ret} wrappers, only return +%% values; no variables are exported. If the match would have returned +%% extra variables then these have been transformed to multiple return +%% values. +%% +%% 2. All BIF's called in guards are gc-safe so there is no need to +%% put thing on the stack in the guard. While this would in principle +%% work it would be difficult to keep track of the stack depth when +%% trimming. +%% +%% The code generation uses variable lifetime information added by +%% the previous pass to save variables, allocate registers and +%% move registers to the stack when necessary. +%% +%% We try to use a consistent variable name scheme throughout. The +%% StackReg record is always called Bef,Int<n>,Aft. + %% cg_fun([Lkexpr], [HeadVar], Vdb, State) -> {[Ainstr],State} cg_fun(Les, Hvs, Vdb, AtomMod, NameArity, Anno, St0) -> @@ -114,14 +332,14 @@ cg_fun(Les, Hvs, Vdb, AtomMod, NameArity, Anno, St0) -> %% Note that and 'if_end' instruction does not need any %% live x registers, so it will always be safe to jump to %% it. (We never ever expect the jump to be taken, and in - %% must functions there will never be any references to + %% most functions there will never be any references to %% the label in the first place.) %% {UltimateMatchFail,St3} = new_label(St2), %% Create initial stack/register state, clear unused arguments. - Bef = clear_dead(#sr{reg=foldl(fun ({var,V}, Reg) -> + Bef = clear_dead(#sr{reg=foldl(fun (#k_var{name=V}, Reg) -> put_reg(V, Reg) end, [], Hvs), stk=[]}, 0, Vdb), @@ -136,45 +354,43 @@ cg_fun(Les, Hvs, Vdb, AtomMod, NameArity, Anno, St0) -> %% cg(Lkexpr, Vdb, StackReg, State) -> {[Ainstr],StackReg,State}. %% Generate code for a kexpr. -%% Split function into two steps for clarity, not efficiency. -cg(Le, Vdb, Bef, St) -> - cg(Le#l.ke, Le, Vdb, Bef, St). - -cg({block,Es}, Le, Vdb, Bef, St) -> +cg(#cg_block{anno=Le,es=Es}, Vdb, Bef, St) -> block_cg(Es, Le, Vdb, Bef, St); -cg({match,M,Rs}, Le, Vdb, Bef, St) -> +cg(#k_match{anno=Le,body=M,ret=Rs}, Vdb, Bef, St) -> match_cg(M, Rs, Le, Vdb, Bef, St); -cg({guard_match,M,Rs}, Le, Vdb, Bef, St) -> +cg(#k_guard_match{anno=Le,body=M,ret=Rs}, Vdb, Bef, St) -> guard_match_cg(M, Rs, Le, Vdb, Bef, St); -cg({call,Func,As,Rs}, Le, Vdb, Bef, St) -> +cg(#k_call{anno=Le,op=Func,args=As,ret=Rs}, Vdb, Bef, St) -> call_cg(Func, As, Rs, Le, Vdb, Bef, St); -cg({enter,Func,As}, Le, Vdb, Bef, St) -> +cg(#k_enter{anno=Le,op=Func,args=As}, Vdb, Bef, St) -> enter_cg(Func, As, Le, Vdb, Bef, St); -cg({bif,Bif,As,Rs}, Le, Vdb, Bef, St) -> - bif_cg(Bif, As, Rs, Le, Vdb, Bef, St); -cg({gc_bif,Bif,As,Rs}, Le, Vdb, Bef, St) -> - gc_bif_cg(Bif, As, Rs, Le, Vdb, Bef, St); -cg({internal,Bif,As,Rs}, Le, Vdb, Bef, St) -> - internal_cg(Bif, As, Rs, Le, Vdb, Bef, St); -cg({receive_loop,Te,Rvar,Rm,Tes,Rs}, Le, Vdb, Bef, St) -> +cg(#k_bif{anno=Le}=Bif, Vdb, Bef, St) -> + bif_cg(Bif, Le, Vdb, Bef, St); +cg(#k_receive{anno=Le,timeout=Te,var=Rvar,body=Rm,action=Tes,ret=Rs}, + Vdb, Bef, St) -> recv_loop_cg(Te, Rvar, Rm, Tes, Rs, Le, Vdb, Bef, St); -cg(receive_next, Le, Vdb, Bef, St) -> +cg(#k_receive_next{anno=Le}, Vdb, Bef, St) -> recv_next_cg(Le, Vdb, Bef, St); -cg(receive_accept, _Le, _Vdb, Bef, St) -> {[remove_message],Bef,St}; -cg({'try',Ta,Vs,Tb,Evs,Th,Rs}, Le, Vdb, Bef, St) -> +cg(#k_receive_accept{}, _Vdb, Bef, St) -> + {[remove_message],Bef,St}; +cg(#k_try{anno=Le,arg=Ta,vars=Vs,body=Tb,evars=Evs,handler=Th,ret=Rs}, + Vdb, Bef, St) -> try_cg(Ta, Vs, Tb, Evs, Th, Rs, Le, Vdb, Bef, St); -cg({try_enter,Ta,Vs,Tb,Evs,Th}, Le, Vdb, Bef, St) -> +cg(#k_try_enter{anno=Le,arg=Ta,vars=Vs,body=Tb,evars=Evs,handler=Th}, + Vdb, Bef, St) -> try_enter_cg(Ta, Vs, Tb, Evs, Th, Le, Vdb, Bef, St); -cg({'catch',Cb,R}, Le, Vdb, Bef, St) -> +cg(#k_catch{anno=Le,body=Cb,ret=[R]}, Vdb, Bef, St) -> catch_cg(Cb, R, Le, Vdb, Bef, St); -cg({set,Var,Con}, Le, Vdb, Bef, St) -> - set_cg(Var, Con, Le, Vdb, Bef, St); -cg({return,Rs}, Le, Vdb, Bef, St) -> return_cg(Rs, Le, Vdb, Bef, St); -cg({break,Bs}, Le, Vdb, Bef, St) -> break_cg(Bs, Le, Vdb, Bef, St); -cg({guard_break,Bs,N}, Le, Vdb, Bef, St) -> +cg(#k_put{anno=Le,arg=Con,ret=Var}, Vdb, Bef, St) -> + put_cg(Var, Con, Le, Vdb, Bef, St); +cg(#k_return{anno=Le,args=Rs}, Vdb, Bef, St) -> + return_cg(Rs, Le, Vdb, Bef, St); +cg(#k_break{anno=Le,args=Bs}, Vdb, Bef, St) -> + break_cg(Bs, Le, Vdb, Bef, St); +cg(#k_guard_break{anno=Le,args=Bs,locked=N}, Vdb, Bef, St) -> guard_break_cg(Bs, N, Le, Vdb, Bef, St); -cg({need_heap,H}, _Le, _Vdb, Bef, St) -> +cg(#cg_need_heap{h=H}, _Vdb, Bef, St) -> {[{test_heap,H,max_reg(Bef#sr.reg)}],Bef,St}. %% cg_list([Kexpr], FirstI, Vdb, StackReg, St) -> {[Ainstr],StackReg,St}. @@ -191,11 +407,11 @@ cg_list(Kes, I, Vdb, Bef, St0) -> %% Insert need_heap instructions in Kexpr list. Try to be smart and %% collect them together as much as possible. -need_heap(Kes0, I) -> +need_heap(Kes0, _I) -> {Kes,H} = need_heap_0(reverse(Kes0), 0, []), %% Prepend need_heap if necessary. - need_heap_need(I, H) ++ Kes. + need_heap_need(H) ++ Kes. need_heap_0([Ke|Kes], H0, Acc) -> {Ns,H} = need_heap_1(Ke, H0), @@ -203,27 +419,54 @@ need_heap_0([Ke|Kes], H0, Acc) -> need_heap_0([], H, Acc) -> {Acc,H}. -need_heap_1(#l{ke={set,_,{binary,_}},i=I}, H) -> - {need_heap_need(I, H),0}; -need_heap_1(#l{ke={set,_,{map,_,_,_}},i=I}, H) -> - {need_heap_need(I, H),0}; -need_heap_1(#l{ke={set,_,Val}}, H) -> +need_heap_1(#k_put{arg=#k_binary{}}, H) -> + {need_heap_need(H),0}; +need_heap_1(#k_put{arg=#k_map{}}, H) -> + {need_heap_need(H),0}; +need_heap_1(#k_put{arg=Val}, H) -> %% Just pass through adding to needed heap. {[],H + case Val of - {cons,_} -> 2; - {tuple,Es} -> 1 + length(Es); + #k_cons{} -> 2; + #k_tuple{es=Es} -> 1 + length(Es); _Other -> 0 end}; -need_heap_1(#l{ke={bif,_Bif,_As,_Rs}}, H) -> - {[],H}; -need_heap_1(#l{i=I}, H) -> +need_heap_1(#k_bif{}=Bif, H) -> + case is_gc_bif(Bif) of + false -> + {[],H}; + true -> + {need_heap_need(H),0} + end; +need_heap_1(_Ke, H) -> %% Call or call-like instruction such as set_tuple_element/3. - {need_heap_need(I, H),0}. - -need_heap_need(_I, 0) -> []; -need_heap_need(I, H) -> [#l{ke={need_heap,H},i=I}]. - -%% match_cg(Match, [Ret], Le, Vdb, StackReg, State) -> + {need_heap_need(H),0}. + +need_heap_need(0) -> []; +need_heap_need(H) -> [#cg_need_heap{h=H}]. + +%% is_gc_bif(#k_bif{}) -> true|false. +%% is_gc_bif(Name, Arity) -> true|false. +%% Determines whether the BIF Name/Arity might do a GC. + +is_gc_bif(#k_bif{op=#k_remote{name=#k_atom{val=Name}},args=Args}) -> + is_gc_bif(Name, length(Args)); +is_gc_bif(#k_bif{op=#k_internal{}}) -> + true. + +is_gc_bif(hd, 1) -> false; +is_gc_bif(tl, 1) -> false; +is_gc_bif(self, 0) -> false; +is_gc_bif(node, 0) -> false; +is_gc_bif(node, 1) -> false; +is_gc_bif(element, 2) -> false; +is_gc_bif(get, 1) -> false; +is_gc_bif(tuple_size, 1) -> false; +is_gc_bif(Bif, Arity) -> + not (erl_internal:bool_op(Bif, Arity) orelse + erl_internal:new_type_test(Bif, Arity) orelse + erl_internal:comp_op(Bif, Arity)). + +%% match_cg(Matc, [Ret], Le, Vdb, StackReg, State) -> %% {[Ainstr],StackReg,State}. %% Generate code for a match. First save all variables on the stack %% that are to survive after the match. We leave saved variables in @@ -252,7 +495,7 @@ guard_match_cg(M, Rs, Le, Vdb, Bef, St0) -> clear_dead(Aft#sr{reg=Reg}, I, Vdb), St2#cg{break=St1#cg.break}}. -guard_match_regs([{I,gbreakvar}|Rs], [{var,V}|Vs]) -> +guard_match_regs([{I,gbreakvar}|Rs], [#k_var{name=V}|Vs]) -> [{I,V}|guard_match_regs(Rs, Vs)]; guard_match_regs([R|Rs], Vs) -> [R|guard_match_regs(Rs, Vs)]; @@ -264,17 +507,14 @@ guard_match_regs([], []) -> []. %% down as each level which uses this takes its own internal Vdb not %% the outer one. -match_cg(Le, Fail, Bef, St) -> - match_cg(Le#l.ke, Le, Fail, Bef, St). - -match_cg({alt,F,S}, _Le, Fail, Bef, St0) -> +match_cg(#k_alt{first=F,then=S}, Fail, Bef, St0) -> {Tf,St1} = new_label(St0), {Fis,Faft,St2} = match_cg(F, Tf, Bef, St1), {Sis,Saft,St3} = match_cg(S, Fail, Bef, St2), Aft = sr_merge(Faft, Saft), {Fis ++ [{label,Tf}] ++ Sis,Aft,St3}; -match_cg({select,{var,Vname}=V,Scs0}, #l{a=Anno}, Fail, Bef, St) -> - ReuseForContext = member(reuse_for_context, Anno) andalso +match_cg(#k_select{var=#k_var{anno=Vanno,name=Vname}=V,types=Scs0}, Fail, Bef, St) -> + ReuseForContext = member(reuse_for_context, Vanno) andalso find_reg(Vname, Bef#sr.reg) =/= error, Scs = case ReuseForContext of false -> Scs0; @@ -283,10 +523,10 @@ match_cg({select,{var,Vname}=V,Scs0}, #l{a=Anno}, Fail, Bef, St) -> match_fmf(fun (S, F, Sta) -> select_cg(S, V, F, Fail, Bef, Sta) end, Fail, St, Scs); -match_cg({guard,Gcs}, _Le, Fail, Bef, St) -> +match_cg(#k_guard{clauses=Gcs}, Fail, Bef, St) -> match_fmf(fun (G, F, Sta) -> guard_clause_cg(G, F, Bef, Sta) end, Fail, St, Gcs); -match_cg({block,Es}, Le, _Fail, Bef, St) -> +match_cg(#cg_block{anno=Le,es=Es}, _Fail, Bef, St) -> %% Must clear registers and stack of dead variables. Int = clear_dead(Bef, Le#l.i, Le#l.vdb), block_cg(Es, Le, Int, St). @@ -294,8 +534,8 @@ match_cg({block,Es}, Le, _Fail, Bef, St) -> %% bsm_rename_ctx([Clause], Var) -> [Clause] %% We know from an annotation that the register for a binary can %% be reused for the match context because the two are not truly -%% alive at the same time (even though the conservative life time -%% information calculated by v3_life says so). +%% alive at the same time (even though the life time information +%% says so). %% %% The easiest way to have those variables share the same register is %% to rename the variable with the shortest life-span (the match @@ -306,12 +546,14 @@ match_cg({block,Es}, Le, _Fail, Bef, St) -> %% We must also remove all information about the match context %% variable from all life-time information databases (Vdb). -bsm_rename_ctx([#l{ke={type_clause,binary, - [#l{ke={val_clause,{binary,{var,Old}},Ke0}}=L2]}}=L1|Cs], New) -> +bsm_rename_ctx([#k_type_clause{type=k_binary,values=Vcs}=TC|Cs], New) -> + [#k_val_clause{val=#k_binary{segs=#k_var{name=Old}}=Bin, + body=Ke0}=VC0] = Vcs, Ke = bsm_rename_ctx(Ke0, Old, New, false), - [L1#l{ke={type_clause,binary, - [L2#l{ke={val_clause,{binary,{var,New}},Ke}}]}}|bsm_rename_ctx(Cs, New)]; -bsm_rename_ctx([C|Cs], New) -> + VC = VC0#k_val_clause{val=Bin#k_binary{segs=#k_var{name=New}}, + body=Ke}, + [TC#k_type_clause{values=[VC]}|bsm_rename_ctx(Cs, New)]; +bsm_rename_ctx([C|Cs], New) -> [C|bsm_rename_ctx(Cs, New)]; bsm_rename_ctx([], _) -> []. @@ -321,34 +563,24 @@ bsm_rename_ctx([], _) -> []. %% only complicatate things to recurse into blocks not in a protected %% (the match context variable is not live inside them). -bsm_rename_ctx(#l{ke={select,{var,V},Cs0}}=L, Old, New, InProt) -> +bsm_rename_ctx(#k_select{var=#k_var{name=V},types=Cs0}=Sel, + Old, New, InProt) -> Cs = bsm_rename_ctx_list(Cs0, Old, New, InProt), - L#l{ke={select,{var,bsm_rename_var(V, Old, New)},Cs}}; -bsm_rename_ctx(#l{ke={type_clause,Type,Cs0}}=L, Old, New, InProt) -> + Sel#k_select{var=#k_var{name=bsm_rename_var(V, Old, New)},types=Cs}; +bsm_rename_ctx(#k_type_clause{values=Cs0}=TC, Old, New, InProt) -> Cs = bsm_rename_ctx_list(Cs0, Old, New, InProt), - L#l{ke={type_clause,Type,Cs}}; -bsm_rename_ctx(#l{ke={val_clause,{bin_end,V},Ke0}}=L, Old, New, InProt) -> - Ke = bsm_rename_ctx(Ke0, Old, New, InProt), - L#l{ke={val_clause,{bin_end,bsm_rename_var(V, Old, New)},Ke}}; -bsm_rename_ctx(#l{ke={val_clause,{bin_seg,V,Sz,U,Type,Fl,Vs},Ke0}}=L, - Old, New, InProt) -> - Ke = bsm_rename_ctx(Ke0, Old, New, InProt), - L#l{ke={val_clause,{bin_seg,bsm_rename_var(V, Old, New),Sz,U,Type,Fl,Vs},Ke}}; -bsm_rename_ctx(#l{ke={val_clause,{bin_int,V,Sz,U,Fl,Val,Vs},Ke0}}=L, - Old, New, InProt) -> - Ke = bsm_rename_ctx(Ke0, Old, New, InProt), - L#l{ke={val_clause,{bin_int,bsm_rename_var(V, Old, New),Sz,U,Fl,Val,Vs},Ke}}; -bsm_rename_ctx(#l{ke={val_clause,Val,Ke0}}=L, Old, New, InProt) -> + TC#k_type_clause{values=Cs}; +bsm_rename_ctx(#k_val_clause{body=Ke0}=VC, Old, New, InProt) -> Ke = bsm_rename_ctx(Ke0, Old, New, InProt), - L#l{ke={val_clause,Val,Ke}}; -bsm_rename_ctx(#l{ke={alt,F0,S0}}=L, Old, New, InProt) -> + VC#k_val_clause{body=Ke}; +bsm_rename_ctx(#k_alt{first=F0,then=S0}=Alt, Old, New, InProt) -> F = bsm_rename_ctx(F0, Old, New, InProt), S = bsm_rename_ctx(S0, Old, New, InProt), - L#l{ke={alt,F,S}}; -bsm_rename_ctx(#l{ke={guard,Gcs0}}=L, Old, New, InProt) -> + Alt#k_alt{first=F,then=S}; +bsm_rename_ctx(#k_guard{clauses=Gcs0}=Guard, Old, New, InProt) -> Gcs = bsm_rename_ctx_list(Gcs0, Old, New, InProt), - L#l{ke={guard,Gcs}}; -bsm_rename_ctx(#l{ke={guard_clause,G0,B0}}=L, Old, New, InProt) -> + Guard#k_guard{clauses=Gcs}; +bsm_rename_ctx(#k_guard_clause{guard=G0,body=B0}=GC, Old, New, InProt) -> G = bsm_rename_ctx(G0, Old, New, InProt), B = bsm_rename_ctx(B0, Old, New, InProt), %% A guard clause may cause unsaved variables to be saved on the stack. @@ -356,49 +588,49 @@ bsm_rename_ctx(#l{ke={guard_clause,G0,B0}}=L, Old, New, InProt) -> %% same register), it is neither in the stack nor register descriptor %% lists and we would crash when we didn't find it unless we remove %% it from the database. - bsm_forget_var(L#l{ke={guard_clause,G,B}}, Old); -bsm_rename_ctx(#l{ke={protected,Ts0,Rs}}=L, Old, New, _InProt) -> + bsm_forget_var(GC#k_guard_clause{guard=G,body=B}, Old); +bsm_rename_ctx(#k_protected{arg=Ts0}=Prot, Old, New, _InProt) -> InProt = true, Ts = bsm_rename_ctx_list(Ts0, Old, New, InProt), - bsm_forget_var(L#l{ke={protected,Ts,Rs}}, Old); -bsm_rename_ctx(#l{ke={match,Ms0,Rs}}=L, Old, New, InProt) -> + bsm_forget_var(Prot#k_protected{arg=Ts}, Old); +bsm_rename_ctx(#k_match{body=Ms0}=Match, Old, New, InProt) -> Ms = bsm_rename_ctx(Ms0, Old, New, InProt), - L#l{ke={match,Ms,Rs}}; -bsm_rename_ctx(#l{ke={guard_match,Ms0,Rs}}=L, Old, New, InProt) -> + Match#k_match{body=Ms}; +bsm_rename_ctx(#k_guard_match{body=Ms0}=Match, Old, New, InProt) -> Ms = bsm_rename_ctx(Ms0, Old, New, InProt), - L#l{ke={guard_match,Ms,Rs}}; -bsm_rename_ctx(#l{ke={test,_,_,_}}=L, _, _, _) -> L; -bsm_rename_ctx(#l{ke={bif,_,_,_}}=L, _, _, _) -> L; -bsm_rename_ctx(#l{ke={gc_bif,_,_,_}}=L, _, _, _) -> L; -bsm_rename_ctx(#l{ke={set,_,_}}=L, _, _, _) -> L; -bsm_rename_ctx(#l{ke={call,_,_,_}}=L, _, _, _) -> L; -bsm_rename_ctx(#l{ke={block,_}}=L, Old, _, false) -> + Match#k_guard_match{body=Ms}; +bsm_rename_ctx(#k_test{}=Test, _, _, _) -> Test; +bsm_rename_ctx(#k_bif{}=Bif, _, _, _) -> Bif; +bsm_rename_ctx(#k_put{}=Put, _, _, _) -> Put; +bsm_rename_ctx(#k_call{}=Call, _, _, _) -> Call; +bsm_rename_ctx(#cg_block{}=Block, Old, _, false) -> %% This block is not inside a protected. The match context variable cannot %% possibly be live inside the block. - bsm_forget_var(L, Old); -bsm_rename_ctx(#l{ke={block,Bl0}}=L, Old, New, true) -> + bsm_forget_var(Block, Old); +bsm_rename_ctx(#cg_block{es=Es0}=Block, Old, New, true) -> %% A block in a protected. We must recursively rename the variable %% inside the block. - Bl = bsm_rename_ctx_list(Bl0, Old, New, true), - bsm_forget_var(L#l{ke={block,Bl}}, Old); -bsm_rename_ctx(#l{ke={guard_break,Bs,Locked0}}=L0, Old, _New, _InProt) -> + Es = bsm_rename_ctx_list(Es0, Old, New, true), + bsm_forget_var(Block#cg_block{es=Es}, Old); +bsm_rename_ctx(#k_guard_break{locked=Locked0}=Break, Old, _New, _InProt) -> Locked = Locked0 -- [Old], - L = L0#l{ke={guard_break,Bs,Locked}}, - bsm_forget_var(L, Old). + bsm_forget_var(Break#k_guard_break{locked=Locked}, Old). bsm_rename_ctx_list([C|Cs], Old, New, InProt) -> [bsm_rename_ctx(C, Old, New, InProt)| bsm_rename_ctx_list(Cs, Old, New, InProt)]; bsm_rename_ctx_list([], _, _, _) -> []. - + bsm_rename_var(Old, Old, New) -> New; bsm_rename_var(V, _, _) -> V. %% bsm_forget_var(#l{}, Variable) -> #l{} %% Remove a variable from the variable life-time database. -bsm_forget_var(#l{vdb=Vdb}=L, V) -> - L#l{vdb=keydelete(V, 1, Vdb)}. +bsm_forget_var(Ke, V) -> + #l{vdb=Vdb} = L0 = get_kanno(Ke), + L = L0#l{vdb=keydelete(V, 1, Vdb)}, + set_kanno(Ke, L). %% block_cg([Kexpr], Le, Vdb, StackReg, St) -> {[Ainstr],StackReg,St}. %% block_cg([Kexpr], Le, StackReg, St) -> {[Ainstr],StackReg,St}. @@ -421,7 +653,7 @@ cg_block(Kes0, I, Vdb, Bef, St0) -> case basic_block(Kes0) of {Kes1,LastI,Args,Rest} -> Ke = hd(Kes1), - Fb = Ke#l.i, + #l{i=Fb} = get_kanno(Ke), cg_basic_block(Kes1, Fb, LastI, Args, Vdb, Bef, St0); {Kes1,Rest} -> cg_list(Kes1, I, Vdb, Bef, St0) @@ -431,36 +663,46 @@ cg_block(Kes0, I, Vdb, Bef, St0) -> basic_block(Kes) -> basic_block(Kes, []). -basic_block([Le|Les], Acc) -> - case collect_block(Le#l.ke) of - include -> basic_block(Les, [Le|Acc]); +basic_block([Ke|Kes], Acc) -> + case collect_block(Ke) of + include -> basic_block(Kes, [Ke|Acc]); {block_end,As} -> case Acc of [] -> - %% If the basic block does not contain any set instructions, + %% If the basic block does not contain any #k_put{} instructions, %% it serves no useful purpose to do basic block optimizations. - {[Le],Les}; + {[Ke],Kes}; _ -> - {reverse(Acc, [Le]),Le#l.i,As,Les} + #l{i=I} = get_kanno(Ke), + {reverse(Acc, [Ke]),I,As,Kes} end; - no_block -> {reverse(Acc, [Le]),Les} + no_block -> {reverse(Acc, [Ke]),Kes} end. -%% sets that may garbage collect are not allowed in basic blocks. - -collect_block({set,_,{binary,_}}) -> no_block; -collect_block({set,_,{map,_,_,_}}) -> no_block; -collect_block({set,_,_}) -> include; -collect_block({call,{var,_}=Var,As,_Rs}) -> {block_end,As++[Var]}; -collect_block({call,Func,As,_Rs}) -> {block_end,As++func_vars(Func)}; -collect_block({enter,{var,_}=Var,As})-> {block_end,As++[Var]}; -collect_block({enter,Func,As}) -> {block_end,As++func_vars(Func)}; -collect_block({return,Rs}) -> {block_end,Rs}; -collect_block({break,Bs}) -> {block_end,Bs}; +%% #k_put{} instructions that may garbage collect are not allowed in basic blocks. + +collect_block(#k_put{arg=#k_binary{}}) -> + no_block; +collect_block(#k_put{arg=#k_map{}}) -> + no_block; +collect_block(#k_put{}) -> + include; +collect_block(#k_call{op=#k_var{}=Var,args=As}) -> + {block_end,As++[Var]}; +collect_block(#k_call{op=Func,args=As}) -> + {block_end,As++func_vars(Func)}; +collect_block(#k_enter{op=#k_var{}=Var,args=As}) -> + {block_end,As++[Var]}; +collect_block(#k_enter{op=Func,args=As}) -> + {block_end,As++func_vars(Func)}; +collect_block(#k_return{args=Rs}) -> + {block_end,Rs}; +collect_block(#k_break{args=Bs}) -> + {block_end,Bs}; collect_block(_) -> no_block. -func_vars({remote,M,F}) when element(1, M) =:= var; - element(1, F) =:= var -> +func_vars(#k_remote{mod=M,name=F}) + when is_record(M, k_var); is_record(F, k_var) -> [M,F]; func_vars(_) -> []. @@ -478,18 +720,19 @@ cg_basic_block(Kes, Fb, Lf, As, Vdb, Bef, St0) -> {Int0,X0_v0,St0}, need_heap(Kes, Fb)), {Keis,Aft,St1}. -cg_basic_block(#l{ke={need_heap,_}}=Ke, {Inta,X0v,Sta}, _Lf, Vdb) -> +cg_basic_block(#cg_need_heap{}=Ke, {Inta,X0v,Sta}, _Lf, Vdb) -> {Keis,Intb,Stb} = cg(Ke, Vdb, Inta, Sta), {Keis, {Intb,X0v,Stb}}; cg_basic_block(Ke, {Inta,X0_v1,Sta}, Lf, Vdb) -> - {Sis,Intb} = save_carefully(Inta, Ke#l.i, Lf+1, Vdb), - {X0_v2,Intc} = allocate_x0(X0_v1, Ke#l.i, Intb), + #l{i=I} = get_kanno(Ke), + {Sis,Intb} = save_carefully(Inta, I, Lf+1, Vdb), + {X0_v2,Intc} = allocate_x0(X0_v1, I, Intb), Intd = reserve(Intc), {Keis,Inte,Stb} = cg(Ke, Vdb, Intd, Sta), {Sis ++ Keis, {Inte,X0_v2,Stb}}. make_reservation([], _) -> []; -make_reservation([{var,V}|As], I) -> [{I,V}|make_reservation(As, I+1)]; +make_reservation([#k_var{name=V}|As], I) -> [{I,V}|make_reservation(As, I+1)]; make_reservation([A|As], I) -> [{I,A}|make_reservation(As, I+1)]. reserve(Sr) -> Sr#sr{reg=reserve(Sr#sr.res, Sr#sr.reg, Sr#sr.stk)}. @@ -538,7 +781,7 @@ save_carefully([V|Vs], Bef, Acc) -> end. x0_vars([], _Fb, _Lf, _Vdb) -> []; -x0_vars([{var,V}|_], Fb, _Lf, Vdb) -> +x0_vars([#k_var{name=V}|_], Fb, _Lf, Vdb) -> {V,F,_L} = VFL = vdb_find(V, Vdb), x0_vars1([VFL], Fb, F, Vdb); x0_vars([X0|_], Fb, Lf, Vdb) -> @@ -640,21 +883,27 @@ turn_yreg(Other, _MaxY) -> %% wrong. These are different as in the second case there is no need %% to try the next type, it will always fail. -select_cg(#l{ke={type_clause,cons,[S]}}, {var,V}, Tf, Vf, Bef, St) -> +select_cg(#k_type_clause{type=Type,values=Vs}, Var, Tf, Vf, Bef, St) -> + #k_var{name=V} = Var, + select_cg(Type, Vs, V, Tf, Vf, Bef, St). + +select_cg(k_cons, [S], V, Tf, Vf, Bef, St) -> select_cons(S, V, Tf, Vf, Bef, St); -select_cg(#l{ke={type_clause,nil,[S]}}, {var,V}, Tf, Vf, Bef, St) -> +select_cg(k_nil, [S], V, Tf, Vf, Bef, St) -> select_nil(S, V, Tf, Vf, Bef, St); -select_cg(#l{ke={type_clause,binary,[S]}}, {var,V}, Tf, Vf, Bef, St) -> +select_cg(k_binary, [S], V, Tf, Vf, Bef, St) -> select_binary(S, V, Tf, Vf, Bef, St); -select_cg(#l{ke={type_clause,bin_seg,S}}, {var,V}, Tf, _Vf, Bef, St) -> +select_cg(k_bin_seg, S, V, Tf, _Vf, Bef, St) -> select_bin_segs(S, V, Tf, Bef, St); -select_cg(#l{ke={type_clause,bin_int,S}}, {var,V}, Tf, _Vf, Bef, St) -> +select_cg(k_bin_int, S, V, Tf, _Vf, Bef, St) -> select_bin_segs(S, V, Tf, Bef, St); -select_cg(#l{ke={type_clause,bin_end,[S]}}, {var,V}, Tf, _Vf, Bef, St) -> +select_cg(k_bin_end, [S], V, Tf, _Vf, Bef, St) -> select_bin_end(S, V, Tf, Bef, St); -select_cg(#l{ke={type_clause,map,S}}, {var,V}, Tf, Vf, Bef, St) -> +select_cg(k_map, S, V, Tf, Vf, Bef, St) -> select_map(S, V, Tf, Vf, Bef, St); -select_cg(#l{ke={type_clause,Type,Scs}}, {var,V}, Tf, Vf, Bef, St0) -> +select_cg(k_literal, S, V, Tf, Vf, Bef, St) -> + select_literal(S, V, Tf, Vf, Bef, St); +select_cg(Type, Scs, V, Tf, Vf, Bef, St0) -> {Vis,{Aft,St1}} = mapfoldl(fun (S, {Int,Sta}) -> {Val,Is,Inta,Stb} = select_val(S, V, Vf, Bef, Sta), @@ -664,22 +913,29 @@ select_cg(#l{ke={type_clause,Type,Scs}}, {var,V}, Tf, Vf, Bef, St0) -> {Vls,Sis,St2} = select_labels(OptVls, St1, [], []), {select_val_cg(Type, fetch_var(V, Bef), Vls, Tf, Vf, Sis), Aft, St2}. -select_val_cg(tuple, R, [Arity,{f,Lbl}], Tf, Vf, [{label,Lbl}|Sis]) -> +select_val_cg(k_tuple, R, [Arity,{f,Lbl}], Tf, Vf, [{label,Lbl}|Sis]) -> [{test,is_tuple,{f,Tf},[R]},{test,test_arity,{f,Vf},[R,Arity]}|Sis]; -select_val_cg(tuple, R, Vls, Tf, Vf, Sis) -> +select_val_cg(k_tuple, R, Vls, Tf, Vf, Sis) -> [{test,is_tuple,{f,Tf},[R]},{select_tuple_arity,R,{f,Vf},{list,Vls}}|Sis]; select_val_cg(Type, R, [Val, {f,Lbl}], Fail, Fail, [{label,Lbl}|Sis]) -> - [{test,is_eq_exact,{f,Fail},[R,{Type,Val}]}|Sis]; + [{test,is_eq_exact,{f,Fail},[R,{type(Type),Val}]}|Sis]; select_val_cg(Type, R, [Val, {f,Lbl}], Tf, Vf, [{label,Lbl}|Sis]) -> [{test,select_type_test(Type),{f,Tf},[R]}, - {test,is_eq_exact,{f,Vf},[R,{Type,Val}]}|Sis]; + {test,is_eq_exact,{f,Vf},[R,{type(Type),Val}]}|Sis]; select_val_cg(Type, R, Vls0, Tf, Vf, Sis) -> - Vls1 = [case Value of {f,_Lbl} -> Value; _ -> {Type,Value} end || Value <- Vls0], + Vls1 = [case Value of + {f,_Lbl} -> Value; + _ -> {type(Type),Value} + end || Value <- Vls0], [{test,select_type_test(Type),{f,Tf},[R]}, {select_val,R,{f,Vf},{list,Vls1}}|Sis]. - -select_type_test(integer) -> is_integer; -select_type_test(atom) -> is_atom; -select_type_test(float) -> is_float. + +type(k_atom) -> atom; +type(k_float) -> float; +type(k_int) -> integer. + +select_type_test(k_int) -> is_integer; +select_type_test(k_atom) -> is_atom; +select_type_test(k_float) -> is_float. combine([{Is,Vs1}, {Is,Vs2}|Vis]) -> combine([{Is,Vs1 ++ Vs2}|Vis]); combine([V|Vis]) -> [V|combine(Vis)]; @@ -695,36 +951,49 @@ add_vls([V|Vs], Lbl, Acc) -> add_vls(Vs, Lbl, [V, {f,Lbl}|Acc]); add_vls([], _, Acc) -> Acc. -select_cons(#l{ke={val_clause,{cons,Es},B},i=I,vdb=Vdb}, V, Tf, Vf, Bef, St0) -> +select_literal(S, V, Tf, Vf, Bef, St) -> + Reg = fetch_var(V, Bef), + F = fun(ValClause, Fail, St0) -> + {Val,Is,Aft,St1} = select_val(ValClause, V, Vf, Bef, St0), + Test = {test,is_eq_exact,{f,Fail},[Reg,{literal,Val}]}, + {[Test|Is],Aft,St1} + end, + match_fmf(F, Tf, St, S). + +select_cons(#k_val_clause{val=#k_cons{hd=Hd,tl=Tl},body=B,anno=#l{i=I,vdb=Vdb}}, + V, Tf, Vf, Bef, St0) -> + Es = [Hd,Tl], {Eis,Int,St1} = select_extract_cons(V, Es, I, Vdb, Bef, St0), {Bis,Aft,St2} = match_cg(B, Vf, Int, St1), {[{test,is_nonempty_list,{f,Tf},[fetch_var(V, Bef)]}] ++ Eis ++ Bis,Aft,St2}. -select_nil(#l{ke={val_clause,nil,B}}, V, Tf, Vf, Bef, St0) -> +select_nil(#k_val_clause{val=#k_nil{},body=B}, V, Tf, Vf, Bef, St0) -> {Bis,Aft,St1} = match_cg(B, Vf, Bef, St0), {[{test,is_nil,{f,Tf},[fetch_var(V, Bef)]}] ++ Bis,Aft,St1}. -select_binary(#l{ke={val_clause,{binary,{var,V}},B},i=I,vdb=Vdb}, - V, Tf, Vf, Bef, St0) -> +select_binary(#k_val_clause{val=#k_binary{segs=#k_var{name=V}},body=B, + anno=#l{i=I,vdb=Vdb}}, V, Tf, Vf, Bef, St0) -> + #cg{ctx=OldCtx} = St0, Int0 = clear_dead(Bef#sr{reg=Bef#sr.reg}, I, Vdb), - {Bis0,Aft,St1} = match_cg(B, Vf, Int0, St0), + {Bis0,Aft,St1} = match_cg(B, Vf, Int0, St0#cg{ctx=V}), CtxReg = fetch_var(V, Int0), Live = max_reg(Bef#sr.reg), Bis1 = [{test,bs_start_match2,{f,Tf},Live,[CtxReg,V],CtxReg}, {bs_save2,CtxReg,{V,V}}|Bis0], Bis = finish_select_binary(Bis1), - {Bis,Aft,St1}; -select_binary(#l{ke={val_clause,{binary,{var,Ivar}},B},i=I,vdb=Vdb}, - V, Tf, Vf, Bef, St0) -> + {Bis,Aft,St1#cg{ctx=OldCtx}}; +select_binary(#k_val_clause{val=#k_binary{segs=#k_var{name=Ivar}},body=B, + anno=#l{i=I,vdb=Vdb}}, V, Tf, Vf, Bef, St0) -> + #cg{ctx=OldCtx} = St0, Regs = put_reg(Ivar, Bef#sr.reg), Int0 = clear_dead(Bef#sr{reg=Regs}, I, Vdb), - {Bis0,Aft,St1} = match_cg(B, Vf, Int0, St0), + {Bis0,Aft,St1} = match_cg(B, Vf, Int0, St0#cg{ctx=Ivar}), CtxReg = fetch_var(Ivar, Int0), Live = max_reg(Bef#sr.reg), Bis1 = [{test,bs_start_match2,{f,Tf},Live,[fetch_var(V, Bef),Ivar],CtxReg}, {bs_save2,CtxReg,{Ivar,Ivar}}|Bis0], Bis = finish_select_binary(Bis1), - {Bis,Aft,St1}. + {Bis,Aft,St1#cg{ctx=OldCtx}}. finish_select_binary([{bs_save2,R,Point}=I,{bs_restore2,R,Point}|Is]) -> [I|finish_select_binary(Is)]; @@ -746,9 +1015,16 @@ select_bin_segs(Scs, Ivar, Tf, Bef, St) -> select_bin_seg(S, Ivar, Fail, Bef, Sta) end, Tf, St, Scs). -select_bin_seg(#l{ke={val_clause,{bin_seg,Ctx,Size,U,T,Fs0,Es},B},i=I,vdb=Vdb,a=A}, - Ivar, Fail, Bef, St0) -> +select_bin_seg(#k_val_clause{val=#k_bin_seg{size=Size,unit=U,type=T, + seg=Seg,flags=Fs0,next=Next}, + body=B, + anno=#l{i=I,vdb=Vdb,a=A}}, Ivar, Fail, Bef, St0) -> + Ctx = St0#cg.ctx, Fs = [{anno,A}|Fs0], + Es = case Next of + [] -> [Seg]; + _ -> [Seg,Next] + end, {Mis,Int,St1} = select_extract_bin(Es, Size, U, T, Fs, Fail, I, Vdb, Bef, Ctx, B, St0), {Bis,Aft,St2} = match_cg(B, Fail, Int, St1), @@ -761,9 +1037,12 @@ select_bin_seg(#l{ke={val_clause,{bin_seg,Ctx,Size,U,T,Fs0,Es},B},i=I,vdb=Vdb,a= [{bs_restore2,CtxReg,{Ctx,Ivar}}|Mis++Bis] end, {Is,Aft,St2}; -select_bin_seg(#l{ke={val_clause,{bin_int,Ctx,Sz,U,Fs,Val,Es},B},i=I,vdb=Vdb}, - Ivar, Fail, Bef, St0) -> - {Mis,Int,St1} = select_extract_int(Es, Val, Sz, U, Fs, Fail, +select_bin_seg(#k_val_clause{val=#k_bin_int{size=Sz,unit=U,flags=Fs, + val=Val,next=Next}, + body=B, + anno=#l{i=I,vdb=Vdb}}, Ivar, Fail, Bef, St0) -> + Ctx = St0#cg.ctx, + {Mis,Int,St1} = select_extract_int(Next, Val, Sz, U, Fs, Fail, I, Vdb, Bef, Ctx, St0), {Bis,Aft,St2} = match_cg(B, Fail, Int, St1), CtxReg = fetch_var(Ctx, Bef), @@ -784,7 +1063,7 @@ select_bin_seg(#l{ke={val_clause,{bin_int,Ctx,Sz,U,Fs,Val,Es},B},i=I,vdb=Vdb}, end, {[{bs_restore2,CtxReg,{Ctx,Ivar}}|Is],Aft,St2}. -select_extract_int([{var,Tl}], Val, {integer,Sz}, U, Fs, Vf, +select_extract_int(#k_var{name=Tl}, Val, #k_int{val=Sz}, U, Fs, Vf, I, Vdb, Bef, Ctx, St) -> Bits = U*Sz, Bin = case member(big, Fs) of @@ -805,7 +1084,7 @@ select_extract_int([{var,Tl}], Val, {integer,Sz}, U, Fs, Vf, end, {Is,clear_dead(Bef, I, Vdb),St}. -select_extract_bin([{var,Hd},{var,Tl}], Size0, Unit, Type, Flags, Vf, +select_extract_bin([#k_var{name=Hd},#k_var{name=Tl}], Size0, Unit, Type, Flags, Vf, I, Vdb, Bef, Ctx, _Body, St) -> SizeReg = get_bin_size_reg(Size0, Bef), {Es,Aft} = @@ -828,11 +1107,11 @@ select_extract_bin([{var,Hd},{var,Tl}], Size0, Unit, Type, Flags, Vf, {bs_save2,CtxReg,{Ctx,Tl}}],Int1} end, {Es,clear_dead(Aft, I, Vdb),St}; -select_extract_bin([{var,Hd}], Size, Unit, binary, Flags, Vf, +select_extract_bin([#k_var{name=Hd}], Size, Unit, binary, Flags, Vf, I, Vdb, Bef, Ctx, Body, St) -> %% Match the last segment of a binary. We KNOW that the size %% must be 'all'. - Size = {atom,all}, %Assertion. + #k_atom{val=all} = Size, %Assertion. {Es,Aft} = case vdb_find(Hd, Vdb) of {_,_,Lhd} when Lhd =< I -> @@ -857,7 +1136,7 @@ select_extract_bin([{var,Hd}], Size, Unit, binary, Flags, Vf, Name = bs_get_binary2, Live = max_reg(Bef#sr.reg), {[{test,Name,{f,Vf},Live, - [CtxReg,Size,Unit,{field_flags,Flags}],Rhd}], + [CtxReg,atomic(Size),Unit,{field_flags,Flags}],Rhd}], Int1}; true -> %% Since the matching context will not be used again, @@ -872,43 +1151,42 @@ select_extract_bin([{var,Hd}], Size, Unit, binary, Flags, Vf, Name = bs_get_binary2, Live = max_reg(Int1#sr.reg), {[{test,Name,{f,Vf},Live, - [CtxReg,Size,Unit,{field_flags,Flags}],CtxReg}], + [CtxReg,atomic(Size),Unit,{field_flags,Flags}],CtxReg}], Int1} end end, {Es,clear_dead(Aft, I, Vdb),St}. %% is_context_unused(Ke) -> true | false -%% Simple heurististic to determine whether the code that follows will -%% use the current matching context again. (The information of liveness -%% calculcated by v3_life is too conservative to be useful for this purpose.) -%% 'true' means that the code that follows will definitely not use the context -%% again (because it is a block, not guard or matching code); 'false' that we -%% are not sure (there could be more matching). - -is_context_unused(#l{ke=Ke}) -> - is_context_unused(Ke); -is_context_unused({alt,_First,Then}) -> - %% {alt,First,Then} can be used for different purposes. If the Then part +%% Simple heurististic to determine whether the code that follows +%% will use the current matching context again. (The liveness +%% information is too conservative to be useful for this purpose.) +%% 'true' means that the code that follows will definitely not use +%% the context again (because it is a block, not guard or matching +%% code); 'false' that we are not sure (there could be more +%% matching). + +is_context_unused(#k_alt{then=Then}) -> + %% #k_alt{} can be used for different purposes. If the Then part %% is a block, it means that matching has finished and is used for a guard %% to choose between the matched clauses. is_context_unused(Then); -is_context_unused({block,_}) -> +is_context_unused(#cg_block{}) -> true; is_context_unused(_) -> false. -select_bin_end(#l{ke={val_clause,{bin_end,Ctx},B}}, - Ivar, Tf, Bef, St0) -> +select_bin_end(#k_val_clause{val=#k_bin_end{},body=B}, Ivar, Tf, Bef, St0) -> + Ctx = St0#cg.ctx, {Bis,Aft,St2} = match_cg(B, Tf, Bef, St0), CtxReg = fetch_var(Ctx, Bef), {[{bs_restore2,CtxReg,{Ctx,Ivar}}, {test,bs_test_tail2,{f,Tf},[CtxReg,0]}|Bis],Aft,St2}. -get_bin_size_reg({var,V}, Bef) -> +get_bin_size_reg(#k_var{name=V}, Bef) -> fetch_var(V, Bef); get_bin_size_reg(Literal, _Bef) -> - Literal. + atomic(Literal). build_bs_instr(Type, Vf, CtxReg, Live, SizeReg, Unit, Flags, Rhd) -> {Format,Name} = case Type of @@ -942,11 +1220,18 @@ build_skip_instr(Type, Vf, CtxReg, Live, SizeReg, Unit, Flags) -> {test,Name,{f,Vf},[CtxReg,Live,{field_flags,Flags}]} end. -select_val(#l{ke={val_clause,{tuple,Es},B},i=I,vdb=Vdb}, V, Vf, Bef, St0) -> +select_val(#k_val_clause{val=#k_tuple{es=Es},body=B,anno=#l{i=I,vdb=Vdb}}, + V, Vf, Bef, St0) -> {Eis,Int,St1} = select_extract_tuple(V, Es, I, Vdb, Bef, St0), {Bis,Aft,St2} = match_cg(B, Vf, Int, St1), {length(Es),Eis ++ Bis,Aft,St2}; -select_val(#l{ke={val_clause,{_,Val},B}}, _V, Vf, Bef, St0) -> +select_val(#k_val_clause{val=Val0,body=B}, _V, Vf, Bef, St0) -> + Val = case Val0 of + #k_atom{val=Lit} -> Lit; + #k_float{val=Lit} -> Lit; + #k_int{val=Lit} -> Lit; + #k_literal{val=Lit} -> Lit + end, {Bis,Aft,St1} = match_cg(B, Vf, Bef, St0), {Val,Bis,Aft,St1}. @@ -955,7 +1240,7 @@ select_val(#l{ke={val_clause,{_,Val},B}}, _V, Vf, Bef, St0) -> %% Extract tuple elements, but only if they do not immediately die. select_extract_tuple(Src, Vs, I, Vdb, Bef, St) -> - F = fun ({var,V}, {Int0,Elem}) -> + F = fun (#k_var{name=V}, {Int0,Elem}) -> case vdb_find(V, Vdb) of {V,_,L} when L =< I -> {[], {Int0,Elem+1}}; _Other -> @@ -972,9 +1257,10 @@ select_extract_tuple(Src, Vs, I, Vdb, Bef, St) -> select_map(Scs, V, Tf, Vf, Bef, St0) -> Reg = fetch_var(V, Bef), {Is,Aft,St1} = - match_fmf(fun(#l{ke={val_clause,{map,exact,_,Es},B},i=I,vdb=Vdb}, Fail, St1) -> - select_map_val(V, Es, B, Fail, I, Vdb, Bef, St1) - end, Vf, St0, Scs), + match_fmf(fun(#k_val_clause{val=#k_map{op=exact,es=Es}, + body=B,anno=#l{i=I,vdb=Vdb}}, Fail, St1) -> + select_map_val(V, Es, B, Fail, I, Vdb, Bef, St1) + end, Vf, St0, Scs), {[{test,is_map,{f,Tf},[Reg]}|Is],Aft,St1}. select_map_val(V, Es, B, Fail, I, Vdb, Bef, St0) -> @@ -991,29 +1277,32 @@ select_extract_map(Src, Vs, Fail, I, Vdb, Bef, St) -> %% Assume keys are term-sorted Rsrc = fetch_var(Src, Bef), - {{HasKs,GetVs,HasVarKs,GetVarVs},Aft} = lists:foldr(fun - ({map_pair,{var,K},{var,V}},{{HasKsi,GetVsi,HasVarVsi,GetVarVsi},Int0}) -> - case vdb_find(V, Vdb) of - {V,_,L} when L =< I -> - RK = fetch_var(K,Int0), - {{HasKsi,GetVsi,[RK|HasVarVsi],GetVarVsi},Int0}; - _Other -> - Reg1 = put_reg(V, Int0#sr.reg), - Int1 = Int0#sr{reg=Reg1}, - RK = fetch_var(K,Int0), - RV = fetch_reg(V,Reg1), - {{HasKsi,GetVsi,HasVarVsi,[[RK,RV]|GetVarVsi]},Int1} - end; - ({map_pair,Key,{var,V}},{{HasKsi,GetVsi,HasVarVsi,GetVarVsi},Int0}) -> - case vdb_find(V, Vdb) of - {V,_,L} when L =< I -> - {{[Key|HasKsi],GetVsi,HasVarVsi,GetVarVsi},Int0}; - _Other -> - Reg1 = put_reg(V, Int0#sr.reg), - Int1 = Int0#sr{reg=Reg1}, - {{HasKsi,[Key,fetch_reg(V, Reg1)|GetVsi],HasVarVsi,GetVarVsi},Int1} - end - end, {{[],[],[],[]},Bef}, Vs), + {{HasKs,GetVs,HasVarKs,GetVarVs},Aft} = + foldr(fun(#k_map_pair{key=#k_var{name=K},val=#k_var{name=V}}, + {{HasKsi,GetVsi,HasVarVsi,GetVarVsi},Int0}) -> + case vdb_find(V, Vdb) of + {V,_,L} when L =< I -> + RK = fetch_var(K,Int0), + {{HasKsi,GetVsi,[RK|HasVarVsi],GetVarVsi},Int0}; + _Other -> + Reg1 = put_reg(V, Int0#sr.reg), + Int1 = Int0#sr{reg=Reg1}, + RK = fetch_var(K,Int0), + RV = fetch_reg(V,Reg1), + {{HasKsi,GetVsi,HasVarVsi,[[RK,RV]|GetVarVsi]},Int1} + end; + (#k_map_pair{key=Key,val=#k_var{name=V}}, + {{HasKsi,GetVsi,HasVarVsi,GetVarVsi},Int0}) -> + case vdb_find(V, Vdb) of + {V,_,L} when L =< I -> + {{[atomic(Key)|HasKsi],GetVsi,HasVarVsi,GetVarVsi},Int0}; + _Other -> + Reg1 = put_reg(V, Int0#sr.reg), + Int1 = Int0#sr{reg=Reg1}, + {{HasKsi,[atomic(Key),fetch_reg(V, Reg1)|GetVsi], + HasVarVsi,GetVarVsi},Int1} + end + end, {{[],[],[],[]},Bef}, Vs), Code = [{test,has_map_fields,{f,Fail},Rsrc,{list,HasKs}} || HasKs =/= []] ++ [{test,has_map_fields,{f,Fail},Rsrc,{list,[K]}} || K <- HasVarKs] ++ @@ -1022,7 +1311,7 @@ select_extract_map(Src, Vs, Fail, I, Vdb, Bef, St) -> {Code, Aft, St}. -select_extract_cons(Src, [{var,Hd}, {var,Tl}], I, Vdb, Bef, St) -> +select_extract_cons(Src, [#k_var{name=Hd}, #k_var{name=Tl}], I, Vdb, Bef, St) -> {Es,Aft} = case {vdb_find(Hd, Vdb), vdb_find(Tl, Vdb)} of {{_,_,Lhd}, {_,_,Ltl}} when Lhd =< I, Ltl =< I -> %% Both head and tail are dead. No need to generate @@ -1045,7 +1334,7 @@ select_extract_cons(Src, [{var,Hd}, {var,Tl}], I, Vdb, Bef, St) -> {Es,Aft,St}. -guard_clause_cg(#l{ke={guard_clause,G,B},vdb=Vdb}, Fail, Bef, St0) -> +guard_clause_cg(#k_guard_clause{anno=#l{vdb=Vdb},guard=G,body=B}, Fail, Bef, St0) -> {Gis,Int,St1} = guard_cg(G, Fail, Vdb, Bef, St0), {Bis,Aft,St} = match_cg(B, Fail, Int, St1), {Gis ++ Bis,Aft,St}. @@ -1058,11 +1347,13 @@ guard_clause_cg(#l{ke={guard_clause,G,B},vdb=Vdb}, Fail, Bef, St0) -> %% the correct exit point. Primops and tests all go to the next %% instruction on success or jump to a failure label. -guard_cg(#l{ke={protected,Ts,Rs},i=I,vdb=Pdb}, Fail, _Vdb, Bef, St) -> +guard_cg(#k_protected{arg=Ts,ret=Rs,anno=#l{i=I,vdb=Pdb}}, Fail, _Vdb, Bef, St) -> protected_cg(Ts, Rs, Fail, I, Pdb, Bef, St); -guard_cg(#l{ke={block,Ts},i=I,vdb=Bdb}, Fail, _Vdb, Bef, St) -> +guard_cg(#cg_block{es=Ts,anno=#l{i=I,vdb=Bdb}}, Fail, _Vdb, Bef, St) -> guard_cg_list(Ts, Fail, I, Bdb, Bef, St); -guard_cg(#l{ke={test,Test,As,Inverted},i=I,vdb=_Tdb}, Fail, Vdb, Bef, St0) -> +guard_cg(#k_test{anno=#l{i=I},op=Test0,args=As,inverted=Inverted}, + Fail, Vdb, Bef, St0) -> + #k_remote{mod=#k_atom{val=erlang},name=#k_atom{val=Test}} = Test0, case Inverted of false -> test_cg(Test, As, Fail, I, Vdb, Bef, St0); @@ -1096,7 +1387,7 @@ protected_cg(Ts, Rs, _Fail, I, Vdb, Bef, St0) -> St2#cg{bfail=Pfail}), %%ok = io:fwrite("cg ~w: ~p~n", [?LINE,{Rs,I,Vdb,Aft}]), %% Set return values to false. - Mis = [{move,{atom,false},fetch_var(V,Aft)}||{var,V} <- Rs], + Mis = [{move,{atom,false},fetch_var(V,Aft)}||#k_var{name=V} <- Rs], {Tis ++ [{jump,{f,Psucc}}, {label,Pfail}] ++ Mis ++ [{label,Psucc}], Aft,St3#cg{bfail=St0#cg.bfail}}. @@ -1121,7 +1412,7 @@ test_cg(is_map, [A], Fail, I, Vdb, Bef, St) -> Arg = cg_reg_arg_prefer_y(A, Bef), Aft = clear_dead(Bef, I, Vdb), {[{test,is_map,{f,Fail},[Arg]}],Aft,St}; -test_cg(is_boolean, [{atom,Val}], Fail, I, Vdb, Bef, St) -> +test_cg(is_boolean, [#k_atom{val=Val}], Fail, I, Vdb, Bef, St) -> Aft = clear_dead(Bef, I, Vdb), Is = case is_boolean(Val) of true -> []; @@ -1167,7 +1458,7 @@ match_fmf(F, LastFail, St0, [H|T]) -> %% frame size. Finally the actual call is made. Call then needs the %% return values filled in. -call_cg({var,_V} = Var, As, Rs, Le, Vdb, Bef, St0) -> +call_cg(#k_var{}=Var, As, Rs, Le, Vdb, Bef, St0) -> {Sis,Int} = cg_setup_call(As++[Var], Bef, Le#l.i, Vdb), %% Put return values in registers. Reg = load_vars(Rs, clear_regs(Int#sr.reg)), @@ -1176,9 +1467,8 @@ call_cg({var,_V} = Var, As, Rs, Le, Vdb, Bef, St0) -> {Frees,Aft} = free_dead(clear_dead(Int#sr{reg=Reg}, Le#l.i, Vdb)), {Sis ++ Frees ++ [line(Le),{call_fun,Arity}],Aft, need_stack_frame(St0)}; -call_cg({remote,Mod,Name}, As, Rs, Le, Vdb, Bef, St0) - when element(1, Mod) =:= var; - element(1, Name) =:= var -> +call_cg(#k_remote{mod=Mod,name=Name}, As, Rs, Le, Vdb, Bef, St0) + when is_record(Mod, k_var); is_record(Name, k_var) -> {Sis,Int} = cg_setup_call(As++[Mod,Name], Bef, Le#l.i, Vdb), %% Put return values in registers. Reg = load_vars(Rs, clear_regs(Int#sr.reg)), @@ -1196,8 +1486,9 @@ call_cg(Func, As, Rs, Le, Vdb, Bef, St0) -> %% %% move {atom,ok} DestReg %% jump FailureLabel - {remote,{atom,erlang},{atom,error}} = Func, %Assertion. - [{var,DestVar}] = Rs, + #k_remote{mod=#k_atom{val=erlang}, + name=#k_atom{val=error}} = Func, %Assertion. + [#k_var{name=DestVar}] = Rs, Int0 = clear_dead(Bef, Le#l.i, Vdb), Reg = put_reg(DestVar, Int0#sr.reg), Int = Int0#sr{reg=Reg}, @@ -1216,11 +1507,11 @@ call_cg(Func, As, Rs, Le, Vdb, Bef, St0) -> {Sis ++ Frees ++ [line(Le)|Call],Aft,St1} end. -build_call({remote,{atom,erlang},{atom,'!'}}, 2, St0) -> +build_call(#k_remote{mod=#k_atom{val=erlang},name=#k_atom{val='!'}}, 2, St0) -> {[send],need_stack_frame(St0)}; -build_call({remote,{atom,Mod},{atom,Name}}, Arity, St0) -> +build_call(#k_remote{mod=#k_atom{val=Mod},name=#k_atom{val=Name}}, Arity, St0) -> {[{call_ext,Arity,{extfunc,Mod,Name,Arity}}],need_stack_frame(St0)}; -build_call(Name, Arity, St0) when is_atom(Name) -> +build_call(#k_local{name=Name}, Arity, St0) when is_atom(Name) -> {Lbl,St1} = local_func_label(Name, Arity, need_stack_frame(St0)), {[{call,Arity,{f,Lbl}}],St1}. @@ -1236,16 +1527,15 @@ free_dead([Any|Stk], Y, Instr, StkAcc) -> free_dead(Stk, Y+1, Instr, [Any|StkAcc]); free_dead([], _, Instr, StkAcc) -> {Instr,reverse(StkAcc)}. -enter_cg({var,_V} = Var, As, Le, Vdb, Bef, St0) -> +enter_cg(#k_var{} = Var, As, Le, Vdb, Bef, St0) -> {Sis,Int} = cg_setup_call(As++[Var], Bef, Le#l.i, Vdb), %% Build complete code and final stack/register state. Arity = length(As), {Sis ++ [line(Le),{call_fun,Arity},return], clear_dead(Int#sr{reg=clear_regs(Int#sr.reg)}, Le#l.i, Vdb), need_stack_frame(St0)}; -enter_cg({remote,Mod,Name}, As, Le, Vdb, Bef, St0) - when element(1, Mod) =:= var; - element(1, Name) =:= var -> +enter_cg(#k_remote{mod=Mod,name=Name}, As, Le, Vdb, Bef, St0) + when is_record(Mod, k_var); is_record(Name, k_var) -> {Sis,Int} = cg_setup_call(As++[Mod,Name], Bef, Le#l.i, Vdb), %% Build complete code and final stack/register state. Arity = length(As), @@ -1263,19 +1553,19 @@ enter_cg(Func, As, Le, Vdb, Bef, St0) -> clear_dead(Int#sr{reg=clear_regs(Int#sr.reg)}, Le#l.i, Vdb), St1}. -build_enter({remote,{atom,erlang},{atom,'!'}}, 2, St0) -> +build_enter(#k_remote{mod=#k_atom{val=erlang},name=#k_atom{val='!'}}, 2, St0) -> {[send,return],need_stack_frame(St0)}; -build_enter({remote,{atom,Mod},{atom,Name}}, Arity, St0) -> +build_enter(#k_remote{mod=#k_atom{val=Mod},name=#k_atom{val=Name}}, Arity, St0) -> St1 = case trap_bif(Mod, Name, Arity) of true -> need_stack_frame(St0); false -> St0 end, {[{call_ext_only,Arity,{extfunc,Mod,Name,Arity}}],St1}; -build_enter(Name, Arity, St0) when is_atom(Name) -> +build_enter(#k_local{name=Name}, Arity, St0) when is_atom(Name) -> {Lbl,St1} = local_func_label(Name, Arity, St0), {[{call_only,Arity,{f,Lbl}}],St1}. -enter_line({remote,{atom,Mod},{atom,Name}}, Arity, Le) -> +enter_line(#k_remote{mod=#k_atom{val=Mod},name=#k_atom{val=Name}}, Arity, Le) -> case erl_bifs:is_safe(Mod, Name, Arity) of false -> %% Tail-recursive call, possibly to a BIF. @@ -1323,6 +1613,22 @@ trap_bif(erlang, group_leader, 2) -> true; trap_bif(erlang, exit, 2) -> true; trap_bif(_, _, _) -> false. +%% bif_cg(#k_bif{}, Le, Vdb, StackReg, State) -> +%% {[Ainstr],StackReg,State}. +%% Generate code a BIF. + +bif_cg(#k_bif{op=#k_internal{name=Name},args=As,ret=Rs}, Le, Vdb, Bef, St) -> + internal_cg(Name, As, Rs, Le, Vdb, Bef, St); +bif_cg(#k_bif{op=#k_remote{mod=#k_atom{val=erlang},name=#k_atom{val=Name}}, + args=As,ret=Rs}, Le, Vdb, Bef, St) -> + Ar = length(As), + case is_gc_bif(Name, Ar) of + false -> + bif_cg(Name, As, Rs, Le, Vdb, Bef, St); + true -> + gc_bif_cg(Name, As, Rs, Le, Vdb, Bef, St) + end. + %% internal_cg(Bif, [Arg], [Ret], Le, Vdb, StackReg, State) -> %% {[Ainstr],StackReg,State}. @@ -1341,8 +1647,8 @@ internal_cg(dsetelement, [Index0,Tuple0,New0], _Rs, Le, Vdb, Bef, St0) -> clear_dead(Bef, Le#l.i, Vdb), St0}; internal_cg(make_fun, [Func0,Arity0|As], Rs, Le, Vdb, Bef, St0) -> %% This behaves more like a function call. - {atom,Func} = Func0, - {integer,Arity} = Arity0, + #k_atom{val=Func} = Func0, + #k_int{val=Arity} = Arity0, {Sis,Int} = cg_setup_call(As, Bef, Le#l.i, Vdb), Reg = load_vars(Rs, clear_regs(Int#sr.reg)), {FuncLbl,St1} = local_func_label(Func, Arity, St0), @@ -1362,7 +1668,7 @@ internal_cg(raise, As, Rs, Le, Vdb, Bef, St) -> %% bif_cg(Bif, [Arg], [Ret], Le, Vdb, StackReg, State) -> %% {[Ainstr],StackReg,State}. -bif_cg(Bif, As, [{var,V}], Le, Vdb, Bef, St0) -> +bif_cg(Bif, As, [#k_var{name=V}], Le, Vdb, Bef, St0) -> Ars = cg_reg_args(As, Bef), %% If we are inside a catch and in a body (not in guard) and the @@ -1400,7 +1706,7 @@ bif_cg(Bif, As, [{var,V}], Le, Vdb, Bef, St0) -> %% gc_bif_cg(Bif, [Arg], [Ret], Le, Vdb, StackReg, State) -> %% {[Ainstr],StackReg,State}. -gc_bif_cg(Bif, As, [{var,V}], Le, Vdb, Bef, St0) -> +gc_bif_cg(Bif, As, [#k_var{name=V}], Le, Vdb, Bef, St0) -> Ars = cg_reg_args(As, Bef), %% If we are inside a catch and in a body (not in guard) and the @@ -1446,7 +1752,7 @@ recv_loop_cg(Te, Rvar, Rm, Tes, Rs, Le, Vdb, Bef, St0) -> %% cg_recv_mesg( ) -> {[Ainstr],Aft,St}. -cg_recv_mesg({var,R}, Rm, Tl, Bef, St0) -> +cg_recv_mesg(#k_var{name=R}, Rm, Tl, Bef, St0) -> Int0 = Bef#sr{reg=put_reg(R, Bef#sr.reg)}, Ret = fetch_reg(R, Int0#sr.reg), %% Int1 = clear_dead(Int0, I, Rm#l.vdb), @@ -1456,22 +1762,22 @@ cg_recv_mesg({var,R}, Rm, Tl, Bef, St0) -> %% cg_recv_wait(Te, Tes, I, Vdb, Int2, St3) -> {[Ainstr],Aft,St}. -cg_recv_wait({atom,infinity}, Tes, I, Bef, St0) -> +cg_recv_wait(#k_atom{val=infinity}, #cg_block{anno=Le,es=Tes}, I, Bef, St0) -> %% We know that the 'after' body will never be executed. %% But to keep the stack and register information up to date, %% we will generate the code for the 'after' body, and then discard it. - Int1 = clear_dead(Bef, I, Tes#l.vdb), - {_,Int2,St1} = cg_block(Tes#l.ke, Tes#l.i, Tes#l.vdb, - Int1#sr{reg=clear_regs(Int1#sr.reg)}, St0), + Int1 = clear_dead(Bef, I, Le#l.vdb), + {_,Int2,St1} = cg_block(Tes, Le#l.i, Le#l.vdb, + Int1#sr{reg=clear_regs(Int1#sr.reg)}, St0), {[{wait,{f,St1#cg.recv}}],Int2,St1}; -cg_recv_wait({integer,0}, Tes, _I, Bef, St0) -> - {Tis,Int,St1} = cg_block(Tes#l.ke, Tes#l.i, Tes#l.vdb, Bef, St0), +cg_recv_wait(#k_int{val=0}, #cg_block{anno=Le,es=Tes}, _I, Bef, St0) -> + {Tis,Int,St1} = cg_block(Tes, Le#l.i, Le#l.vdb, Bef, St0), {[timeout|Tis],Int,St1}; -cg_recv_wait(Te, Tes, I, Bef, St0) -> +cg_recv_wait(Te, #cg_block{anno=Le,es=Tes}, I, Bef, St0) -> Reg = cg_reg_arg(Te, Bef), %% Must have empty registers here! Bug if anything in registers. - Int0 = clear_dead(Bef, I, Tes#l.vdb), - {Tis,Int,St1} = cg_block(Tes#l.ke, Tes#l.i, Tes#l.vdb, + Int0 = clear_dead(Bef, I, Le#l.vdb), + {Tis,Int,St1} = cg_block(Tes, Le#l.i, Le#l.vdb, Int0#sr{reg=clear_regs(Int0#sr.reg)}, St0), {[{wait_timeout,{f,St1#cg.recv},Reg},timeout] ++ Tis,Int,St1}. @@ -1489,7 +1795,7 @@ try_cg(Ta, Vs, Tb, Evs, Th, Rs, Le, Vdb, Bef, St0) -> {B,St1} = new_label(St0), %Body label {H,St2} = new_label(St1), %Handler label {E,St3} = new_label(St2), %End label - TryTag = Ta#l.i, + #l{i=TryTag} = get_kanno(Ta), Int1 = Bef#sr{stk=put_catch(TryTag, Bef#sr.stk)}, TryReg = fetch_stack({catch_tag,TryTag}, Int1#sr.stk), {Ais,Int2,St4} = cg(Ta, Vdb, Int1, St3#cg{break=B,in_catch=true}), @@ -1509,7 +1815,7 @@ try_cg(Ta, Vs, Tb, Evs, Th, Rs, Le, Vdb, Bef, St0) -> try_enter_cg(Ta, Vs, Tb, Evs, Th, Le, Vdb, Bef, St0) -> {B,St1} = new_label(St0), %Body label {H,St2} = new_label(St1), %Handler label - TryTag = Ta#l.i, + #l{i=TryTag} = get_kanno(Ta), Int1 = Bef#sr{stk=put_catch(TryTag, Bef#sr.stk)}, TryReg = fetch_stack({catch_tag,TryTag}, Int1#sr.stk), {Ais,Int2,St3} = cg(Ta, Vdb, Int1, St2#cg{break=B,in_catch=true}), @@ -1527,7 +1833,7 @@ try_enter_cg(Ta, Vs, Tb, Evs, Th, Le, Vdb, Bef, St0) -> %% catch_cg(CatchBlock, Ret, Le, Vdb, Bef, St) -> {[Ainstr],Aft,St}. -catch_cg(C, {var,R}, Le, Vdb, Bef, St0) -> +catch_cg(#cg_block{es=C}, #k_var{name=R}, Le, Vdb, Bef, St0) -> {B,St1} = new_label(St0), CatchTag = Le#l.i, Int1 = Bef#sr{stk=put_catch(CatchTag, Bef#sr.stk)}, @@ -1541,8 +1847,8 @@ catch_cg(C, {var,R}, Le, Vdb, Bef, St0) -> clear_dead(Aft, Le#l.i, Vdb), St2#cg{break=St1#cg.break,in_catch=St1#cg.in_catch}}. -%% set_cg([Var], Constr, Le, Vdb, Bef, St) -> {[Ainstr],Aft,St}. -%% We have to be careful how a 'set' works. First the structure is +%% put_cg([Var], Constr, Le, Vdb, Bef, St) -> {[Ainstr],Aft,St}. +%% We have to be careful how a 'put' works. First the structure is %% built, then it is filled and finally things can be cleared. The %% annotation must reflect this and make sure that the return %% variable is allocated first. @@ -1550,13 +1856,14 @@ catch_cg(C, {var,R}, Le, Vdb, Bef, St0) -> %% put_list and put_map are atomic instructions, both of %% which can safely resuse one of the source registers as target. -set_cg([{var,R}], {cons,Es}, Le, Vdb, Bef, St) -> - [S1,S2] = cg_reg_args(Es, Bef), +put_cg([#k_var{name=R}], #k_cons{hd=Hd,tl=Tl}, Le, Vdb, Bef, St) -> + [S1,S2] = cg_reg_args([Hd,Tl], Bef), Int0 = clear_dead(Bef, Le#l.i, Vdb), Int1 = Int0#sr{reg=put_reg(R, Int0#sr.reg)}, Ret = fetch_reg(R, Int1#sr.reg), {[{put_list,S1,S2,Ret}], Int1, St}; -set_cg([{var,R}], {binary,Segs}, Le, Vdb, Bef, #cg{bfail=Bfail}=St) -> +put_cg([#k_var{name=R}], #k_binary{segs=Segs}, Le, Vdb, Bef, + #cg{bfail=Bfail}=St) -> %% At run-time, binaries are constructed in three stages: %% 1) First the size of the binary is calculated. %% 2) Then the binary is allocated. @@ -1584,7 +1891,9 @@ set_cg([{var,R}], {binary,Segs}, Le, Vdb, Bef, #cg{bfail=Bfail}=St) -> {Sis++Code,Aft,St}; %% Map: single variable key. -set_cg([{var,R}], {map,Op,Map,[{map_pair,{var,_}=K,V}]}, Le, Vdb, Bef, St0) -> +put_cg([#k_var{name=R}], #k_map{op=Op,var=Map, + es=[#k_map_pair{key=#k_var{}=K,val=V}]}, + Le, Vdb, Bef, St0) -> {Sis,Int0} = maybe_adjust_stack(Bef, Le#l.i, Le#l.i+1, Vdb, St0), SrcReg = cg_reg_arg_prefer_y(Map, Int0), @@ -1599,22 +1908,23 @@ set_cg([{var,R}], {map,Op,Map,[{map_pair,{var,_}=K,V}]}, Le, Vdb, Bef, St0) -> Aft = Aft0#sr{reg=put_reg(R, Aft0#sr.reg)}, Target = fetch_reg(R, Aft#sr.reg), - {Is,St1} = set_cg_map(Line, Op, SrcReg, Target, Live, List, St0), + {Is,St1} = put_cg_map(Line, Op, SrcReg, Target, Live, List, St0), {Sis++Is,Aft,St1}; %% Map: (possibly) multiple literal keys. -set_cg([{var,R}], {map,Op,Map,Es}, Le, Vdb, Bef, St0) -> +put_cg([#k_var{name=R}], #k_map{op=Op,var=Map,es=Es}, Le, Vdb, Bef, St0) -> %% assert key literals - [] = [Var||{map_pair,{var,_}=Var,_} <- Es], + [] = [Var || #k_map_pair{key=#k_var{}=Var} <- Es], {Sis,Int0} = maybe_adjust_stack(Bef, Le#l.i, Le#l.i+1, Vdb, St0), SrcReg = cg_reg_arg_prefer_y(Map, Int0), Line = line(Le#l.a), %% fetch registers for values to be put into the map - Pairs = [{K,V} || {_,K,V} <- Es], - List = flatmap(fun({K,V}) -> [K,cg_reg_arg(V,Int0)] end, Pairs), + List = flatmap(fun(#k_map_pair{key=K,val=V}) -> + [atomic(K),cg_reg_arg(V, Int0)] + end, Es), Live = max_reg(Bef#sr.reg), @@ -1623,16 +1933,16 @@ set_cg([{var,R}], {map,Op,Map,Es}, Le, Vdb, Bef, St0) -> Aft = Aft0#sr{reg=put_reg(R, Aft0#sr.reg)}, Target = fetch_reg(R, Aft#sr.reg), - {Is,St1} = set_cg_map(Line, Op, SrcReg, Target, Live, List, St0), + {Is,St1} = put_cg_map(Line, Op, SrcReg, Target, Live, List, St0), {Sis++Is,Aft,St1}; %% Everything else. -set_cg([{var,R}], Con, Le, Vdb, Bef, St) -> +put_cg([#k_var{name=R}], Con, Le, Vdb, Bef, St) -> %% Find a place for the return register first. Int = Bef#sr{reg=put_reg(R, Bef#sr.reg)}, Ret = fetch_reg(R, Int#sr.reg), Ais = case Con of - {tuple,Es} -> + #k_tuple{es=Es} -> [{put_tuple,length(Es),Ret}] ++ cg_build_args(Es, Bef); Other -> [{move,cg_reg_arg(Other, Int),Ret}] @@ -1640,7 +1950,7 @@ set_cg([{var,R}], Con, Le, Vdb, Bef, St) -> {Ais,clear_dead(Int, Le#l.i, Vdb),St}. -set_cg_map(Line, Op0, SrcReg, Target, Live, List, St0) -> +put_cg_map(Line, Op0, SrcReg, Target, Live, List, St0) -> Bfail = St0#cg.bfail, Fail = {f,St0#cg.bfail}, Op = case Op0 of @@ -1854,7 +2164,8 @@ cg_bin_opt_1([I|Is]) -> cg_bin_opt_1([]) -> []. -cg_bin_put({bin_seg,[],S0,U,T,Fs,[E0,Next]}, Fail, Bef) -> +cg_bin_put(#k_bin_seg{size=S0,unit=U,type=T,flags=Fs,seg=E0,next=Next}, + Fail, Bef) -> S1 = cg_reg_arg(S0, Bef), E1 = cg_reg_arg(E0, Bef), {Format,Op} = case T of @@ -1871,7 +2182,7 @@ cg_bin_put({bin_seg,[],S0,U,T,Fs,[E0,Next]}, Fail, Bef) -> utf -> [{Op,Fail,{field_flags,Fs},E1}|cg_bin_put(Next, Fail, Bef)] end; -cg_bin_put({bin_end,[]}, _, _) -> []. +cg_bin_put(#k_bin_end{}, _, _) -> []. cg_build_args(As, Bef) -> [{put,cg_reg_arg(A, Bef)} || A <- As]. @@ -1925,11 +2236,11 @@ get_locked_regs([], _) -> []. cg_reg_args(As, Bef) -> [cg_reg_arg(A, Bef) || A <- As]. -cg_reg_arg({var,V}, Bef) -> fetch_var(V, Bef); -cg_reg_arg(Literal, _) -> Literal. +cg_reg_arg(#k_var{name=V}, Bef) -> fetch_var(V, Bef); +cg_reg_arg(Literal, _) -> atomic(Literal). -cg_reg_arg_prefer_y({var,V}, Bef) -> fetch_var_prefer_y(V, Bef); -cg_reg_arg_prefer_y(Literal, _) -> Literal. +cg_reg_arg_prefer_y(#k_var{name=V}, Bef) -> fetch_var_prefer_y(V, Bef); +cg_reg_arg_prefer_y(Literal, _) -> atomic(Literal). %% cg_setup_call([Arg], Bef, Cur, Vdb) -> {[Instr],Aft}. %% Do the complete setup for a call/enter. @@ -1967,9 +2278,9 @@ cg_call_args(As, Bef, I, Vdb) -> load_arg_regs(Regs, As) -> load_arg_regs(Regs, As, 0). -load_arg_regs([_|Rs], [{var,V}|As], I) -> [{I,V}|load_arg_regs(Rs, As, I+1)]; +load_arg_regs([_|Rs], [#k_var{name=V}|As], I) -> [{I,V}|load_arg_regs(Rs, As, I+1)]; load_arg_regs([_|Rs], [A|As], I) -> [{I,A}|load_arg_regs(Rs, As, I+1)]; -load_arg_regs([], [{var,V}|As], I) -> [{I,V}|load_arg_regs([], As, I+1)]; +load_arg_regs([], [#k_var{name=V}|As], I) -> [{I,V}|load_arg_regs([], As, I+1)]; load_arg_regs([], [A|As], I) -> [{I,A}|load_arg_regs([], As, I+1)]; load_arg_regs(Rs, [], _) -> Rs. @@ -2005,12 +2316,13 @@ move_unsaved([], _, Regs, Acc) -> {Acc,Regs}. gen_moves(As, Sr) -> gen_moves(As, Sr, 0, []). -gen_moves([{var,V}|As], Sr, I, Acc) -> +gen_moves([#k_var{name=V}|As], Sr, I, Acc) -> case fetch_var(V, Sr) of {x,I} -> gen_moves(As, Sr, I+1, Acc); Reg -> gen_moves(As, Sr, I+1, [{move,Reg,{x,I}}|Acc]) end; -gen_moves([A|As], Sr, I, Acc) -> +gen_moves([A0|As], Sr, I, Acc) -> + A = atomic(A0), gen_moves(As, Sr, I+1, [{move,A,{x,I}}|Acc]); gen_moves([], _, _, Acc) -> lists:keysort(3, Acc). @@ -2179,7 +2491,7 @@ fetch_var_prefer_y(V, #sr{reg=Reg,stk=Stk}) -> end. load_vars(Vs, Regs) -> - foldl(fun ({var,V}, Rs) -> put_reg(V, Rs) end, Regs, Vs). + foldl(fun (#k_var{name=V}, Rs) -> put_reg(V, Rs) end, Regs, Vs). %% put_reg(Val, Regs) -> Regs. %% find_reg(Val, Regs) -> {ok,r{R}} | error. @@ -2280,6 +2592,16 @@ put_catch(Tag, [Other|Stk], Acc) -> drop_catch(Tag, [{{catch_tag,Tag}}|Stk]) -> [free|Stk]; drop_catch(Tag, [Other|Stk]) -> [Other|drop_catch(Tag, Stk)]. +%% atomic(Klit) -> Lit. +%% atomic_list([Klit]) -> [Lit]. + +atomic(#k_literal{val=V}) -> {literal,V}; +atomic(#k_int{val=I}) -> {integer,I}; +atomic(#k_float{val=F}) -> {float,F}; +atomic(#k_atom{val=A}) -> {atom,A}; +%%atomic(#k_char{val=C}) -> {char,C}; +atomic(#k_nil{}) -> nil. + %% new_label(St) -> {L,St}. new_label(#cg{lcount=Next}=St) -> @@ -2322,3 +2644,86 @@ flatmapfoldl(F, Accu0, [Hd|Tail]) -> {Rs,Accu2} = flatmapfoldl(F, Accu1, Tail), {R++Rs,Accu2}; flatmapfoldl(_, Accu, []) -> {[],Accu}. + +%% Keep track of life time for variables. +%% +%% init_vars([{var,VarName}]) -> Vdb. +%% new_vars([VarName], I, Vdb) -> Vdb. +%% use_vars([VarName], I, Vdb) -> Vdb. +%% add_var(VarName, F, L, Vdb) -> Vdb. +%% +%% The list of variable names for new_vars/3 and use_vars/3 +%% must be sorted. + +init_vars(Vs) -> + vdb_new(Vs). + +new_vars([], _, Vdb) -> Vdb; +new_vars([V], I, Vdb) -> vdb_store_new(V, {V,I,I}, Vdb); +new_vars(Vs, I, Vdb) -> vdb_update_vars(Vs, Vdb, I). + +use_vars([], _, Vdb) -> + Vdb; +use_vars([V], I, Vdb) -> + case vdb_find(V, Vdb) of + {V,F,L} when I > L -> vdb_update(V, {V,F,I}, Vdb); + {V,_,_} -> Vdb; + error -> vdb_store_new(V, {V,I,I}, Vdb) + end; +use_vars(Vs, I, Vdb) -> vdb_update_vars(Vs, Vdb, I). + +add_var(V, F, L, Vdb) -> + vdb_store_new(V, {V,F,L}, Vdb). + +%% vdb + +vdb_new(Vs) -> + ordsets:from_list([{V,0,0} || #k_var{name=V} <- Vs]). + +-type var() :: atom(). + +-spec vdb_find(var(), [vdb_entry()]) -> 'error' | vdb_entry(). + +vdb_find(V, Vdb) -> + case lists:keyfind(V, 1, Vdb) of + false -> error; + Vd -> Vd + end. + +vdb_update(V, Update, [{V,_,_}|Vdb]) -> + [Update|Vdb]; +vdb_update(V, Update, [Vd|Vdb]) -> + [Vd|vdb_update(V, Update, Vdb)]. + +vdb_store_new(V, New, [{V1,_,_}=Vd|Vdb]) when V > V1 -> + [Vd|vdb_store_new(V, New, Vdb)]; +vdb_store_new(V, New, [{V1,_,_}|_]=Vdb) when V < V1 -> + [New|Vdb]; +vdb_store_new(_, New, []) -> [New]. + +vdb_update_vars([V|_]=Vs, [{V1,_,_}=Vd|Vdb], I) when V > V1 -> + [Vd|vdb_update_vars(Vs, Vdb, I)]; +vdb_update_vars([V|Vs], [{V1,_,_}|_]=Vdb, I) when V < V1 -> + %% New variable. + [{V,I,I}|vdb_update_vars(Vs, Vdb, I)]; +vdb_update_vars([V|Vs], [{_,F,L}=Vd|Vdb], I) -> + %% Existing variable. + if + I > L -> [{V,F,I}|vdb_update_vars(Vs, Vdb, I)]; + true -> [Vd|vdb_update_vars(Vs, Vdb, I)] + end; +vdb_update_vars([V|Vs], [], I) -> + %% New variable. + [{V,I,I}|vdb_update_vars(Vs, [], I)]; +vdb_update_vars([], Vdb, _) -> Vdb. + +%% vdb_sub(Min, Max, Vdb) -> Vdb. +%% Extract variables which are used before and after Min. Lock +%% variables alive after Max. + +vdb_sub(Min, Max, Vdb) -> + [ if L >= Max -> {V,F,locked}; + true -> Vd + end || {V,F,L}=Vd <- Vdb, + F < Min, + L >= Min ]. diff --git a/lib/compiler/src/v3_kernel.erl b/lib/compiler/src/v3_kernel.erl index 1fc05109c5..3eea058153 100644 --- a/lib/compiler/src/v3_kernel.erl +++ b/lib/compiler/src/v3_kernel.erl @@ -82,7 +82,8 @@ -export([module/2,format_error/1]). -import(lists, [map/2,foldl/3,foldr/3,mapfoldl/3,splitwith/2,member/2, - keymember/3,keyfind/3,partition/2,droplast/1,last/1,sort/1]). + keymember/3,keyfind/3,partition/2,droplast/1,last/1,sort/1, + reverse/1]). -import(ordsets, [add_element/2,del_element/2,union/2,union/1,subtract/2]). -import(cerl, [c_tuple/1]). @@ -159,8 +160,7 @@ function({#c_var{name={F,Arity}=FA},Body}, St0) -> {#ifun{anno=Ab,vars=Kvs,body=B0},[],St2} = expr(Body, new_sub(), St1), {B1,_,St3} = ubody(B0, return, St2), %%B1 = B0, St3 = St2, %Null second pass - {#k_fdef{anno=#k{us=[],ns=[],a=Ab}, - func=F,arity=Arity,vars=Kvs,body=B1},St3} + {make_fdef(#k{us=[],ns=[],a=Ab}, F, Arity, Kvs, B1),St3} catch Class:Error -> Stack = erlang:get_stacktrace(), @@ -1589,23 +1589,18 @@ match_var([U|Us], Cs0, Def, St) -> %% according to type, the order is really irrelevant but tries to be %% smart. -match_con(Us, [C], Def, St) -> - %% There is only one clause. We can keep literal tuples and - %% lists, but we must convert []/integer/float/atom literals - %% to the proper record (#k_nil{} and so on). - Cs = [expand_pat_lit_clause(C, false)], - match_con_1(Us, Cs, Def, St); match_con(Us, Cs0, Def, St) -> - %% More than one clause. Remove literals at the top level. - Cs = [expand_pat_lit_clause(C, true) || C <- Cs0], + %% Expand literals at the top level. + Cs = [expand_pat_lit_clause(C) || C <- Cs0], match_con_1(Us, Cs, Def, St). match_con_1([U|_Us] = L, Cs, Def, St0) -> %% Extract clauses for different constructors (types). %%ok = io:format("match_con ~p~n", [Cs]), - Ttcs = select_types([k_binary], Cs) ++ select_bin_con(Cs) ++ - select_types([k_cons,k_tuple,k_map,k_atom,k_float,k_int, - k_nil,k_literal], Cs), + Ttcs0 = select_types([k_binary], Cs) ++ select_bin_con(Cs) ++ + select_types([k_cons,k_tuple,k_map,k_atom,k_float, + k_int,k_nil], Cs), + Ttcs = opt_single_valued(Ttcs0), %%ok = io:format("ttcs = ~p~n", [Ttcs]), {Scs,St1} = mapfoldl(fun ({T,Tcs}, St) -> @@ -1618,28 +1613,14 @@ match_con_1([U|_Us] = L, Cs, Def, St0) -> select_types(Types, Cs) -> [{T,Tcs} || T <- Types, begin Tcs = select(T, Cs), Tcs =/= [] end]. - -expand_pat_lit_clause(#iclause{pats=[#ialias{pat=#k_literal{anno=A,val=Val}}=Alias|Ps]}=C, B) -> - P = case B of - true -> expand_pat_lit(Val, A); - false -> literal(Val, A) - end, + +expand_pat_lit_clause(#iclause{pats=[#ialias{pat=#k_literal{anno=A,val=Val}}=Alias|Ps]}=C) -> + P = expand_pat_lit(Val, A), C#iclause{pats=[Alias#ialias{pat=P}|Ps]}; -expand_pat_lit_clause(#iclause{pats=[#k_literal{anno=A,val=Val}|Ps]}=C, B) -> - P = case B of - true -> expand_pat_lit(Val, A); - false -> literal(Val, A) - end, +expand_pat_lit_clause(#iclause{pats=[#k_literal{anno=A,val=Val}|Ps]}=C) -> + P = expand_pat_lit(Val, A), C#iclause{pats=[P|Ps]}; -expand_pat_lit_clause(#iclause{pats=[#k_binary{anno=A,segs=#k_bin_end{}}|Ps]}=C, B) -> - case B of - true -> - C; - false -> - P = #k_literal{anno=A,val = <<>>}, - C#iclause{pats=[P|Ps]} - end; -expand_pat_lit_clause(C, _) -> C. +expand_pat_lit_clause(C) -> C. expand_pat_lit([H|T], A) -> #k_cons{anno=A,hd=literal(H, A),tl=literal(T, A)}; @@ -1659,6 +1640,107 @@ literal(Val, A) when is_atom(Val) -> literal(Val, A) when is_list(Val); is_tuple(Val) -> #k_literal{anno=A,val=Val}. +%% opt_singled_valued([{Type,Clauses}]) -> [{Type,Clauses}]. +%% If a type only has one clause and if the pattern is literal, +%% the matching can be done more efficiently by directly comparing +%% with the literal (that is especially true for binaries). + +opt_single_valued(Ttcs) -> + opt_single_valued(Ttcs, [], []). + +opt_single_valued([{_,[#iclause{pats=[P0|Ps]}=Tc]}=Ttc|Ttcs], TtcAcc, LitAcc) -> + try combine_lit_pat(P0) of + P -> + LitTtc = Tc#iclause{pats=[P|Ps]}, + opt_single_valued(Ttcs, TtcAcc, [LitTtc|LitAcc]) + catch + not_possible -> + opt_single_valued(Ttcs, [Ttc|TtcAcc], LitAcc) + end; +opt_single_valued([Ttc|Ttcs], TtcAcc, LitAcc) -> + opt_single_valued(Ttcs, [Ttc|TtcAcc], LitAcc); +opt_single_valued([], TtcAcc, []) -> + reverse(TtcAcc); +opt_single_valued([], TtcAcc, LitAcc) -> + Literals = {k_literal,reverse(LitAcc)}, + %% Test the literals as early as possible. + case reverse(TtcAcc) of + [{k_binary,_}=Bin|Ttcs] -> + %% The delayed creation of sub binaries requires + %% bs_start_match2 to be the first instruction in the + %% function. + [Bin,Literals|Ttcs]; + Ttcs -> + [Literals|Ttcs] + end. + +combine_lit_pat(#ialias{pat=Pat0}=Alias) -> + Pat = combine_lit_pat(Pat0), + Alias#ialias{pat=Pat}; +combine_lit_pat(Pat) -> + case do_combine_lit_pat(Pat) of + #k_literal{val=Val} when is_atom(Val) -> + throw(not_possible); + #k_literal{val=Val} when is_number(Val) -> + throw(not_possible); + #k_literal{val=[]} -> + throw(not_possible); + #k_literal{}=Lit -> + Lit + end. + +do_combine_lit_pat(#k_atom{anno=A,val=Val}) -> + #k_literal{anno=A,val=Val}; +do_combine_lit_pat(#k_float{anno=A,val=Val}) -> + #k_literal{anno=A,val=Val}; +do_combine_lit_pat(#k_int{anno=A,val=Val}) -> + #k_literal{anno=A,val=Val}; +do_combine_lit_pat(#k_nil{anno=A}) -> + #k_literal{anno=A,val=[]}; +do_combine_lit_pat(#k_binary{anno=A,segs=Segs}) -> + Bin = combine_bin_segs(Segs), + #k_literal{anno=A,val=Bin}; +do_combine_lit_pat(#k_cons{anno=A,hd=Hd0,tl=Tl0}) -> + #k_literal{val=Hd} = do_combine_lit_pat(Hd0), + #k_literal{val=Tl} = do_combine_lit_pat(Tl0), + #k_literal{anno=A,val=[Hd|Tl]}; +do_combine_lit_pat(#k_literal{}=Lit) -> + Lit; +do_combine_lit_pat(#k_tuple{anno=A,es=Es0}) -> + Es = [begin + #k_literal{val=Lit} = do_combine_lit_pat(El), + Lit + end || El <- Es0], + #k_literal{anno=A,val=list_to_tuple(Es)}; +do_combine_lit_pat(_) -> + throw(not_possible). + +combine_bin_segs(#k_bin_seg{size=Size0,unit=Unit,type=integer, + flags=[unsigned,big],seg=Seg,next=Next}) -> + #k_literal{val=Size1} = do_combine_lit_pat(Size0), + #k_literal{val=Int} = do_combine_lit_pat(Seg), + Size = Size1 * Unit, + if + 0 < Size, Size < 64 -> + Bin = <<Int:Size>>, + case Bin of + <<Int:Size>> -> + NextBin = combine_bin_segs(Next), + <<Bin/bits,NextBin/bits>>; + _ -> + %% The integer Int does not fit in the segment, + %% thus it will not match. + throw(not_possible) + end; + true -> + %% Avoid creating huge binary literals. + throw(not_possible) + end; +combine_bin_segs(#k_bin_end{}) -> + <<>>; +combine_bin_segs(_) -> + throw(not_possible). + %% select_bin_con([Clause]) -> [{Type,[Clause]}]. %% Extract clauses for the k_bin_seg constructor. As k_bin_seg %% matching can overlap, the k_bin_seg constructors cannot be @@ -2179,9 +2261,8 @@ iletrec_funs_gen(Fs, FreeVs, St) -> Arity0 = length(Vs), {Fb1,_,Lst1} = ubody(Fb0, return, Lst0#kern{ff={N,Arity0}}), Arity = Arity0 + length(FreeVs), - Fun = #k_fdef{anno=#k{us=[],ns=[],a=Fa}, - func=N,arity=Arity, - vars=Vs ++ FreeVs,body=Fb1}, + Fun = make_fdef(#k{us=[],ns=[],a=Fa}, N, Arity, + Vs++FreeVs, Fb1), Lst1#kern{funs=[Fun|Lst1#kern.funs]} end, St, Fs). @@ -2325,8 +2406,7 @@ uexpr(#ifun{anno=A,vars=Vs,body=B0}, {break,Rs}, St0) -> %% No id annotation. Must invent a fun name. new_fun_name(St1) end, - Fun = #k_fdef{anno=#k{us=[],ns=[],a=A},func=Fname,arity=Arity, - vars=Vs ++ Fvs,body=B1}, + Fun = make_fdef(#k{us=[],ns=[],a=A}, Fname, Arity, Vs++Fvs, B1), {#k_bif{anno=#k{us=Free,ns=lit_list_vars(Rs),a=A}, op=#k_internal{name=make_fun,arity=length(Free)+2}, args=[#k_atom{val=Fname},#k_int{val=Arity}|Fvs], @@ -2343,6 +2423,16 @@ uexpr(Lit, {break,Rs0}, St0) -> add_local_function(_, #kern{funs=ignore}=St) -> St; add_local_function(F, #kern{funs=Funs}=St) -> St#kern{funs=[F|Funs]}. +%% Make a #k_fdef{}, making sure that the body is always a #k_match{}. +make_fdef(Anno, Name, Arity, Vs, #k_match{}=Body) -> + #k_fdef{anno=Anno,func=Name,arity=Arity,vars=Vs,body=Body}; +make_fdef(Anno, Name, Arity, Vs, Body) -> + Ka = get_kanno(Body), + Match = #k_match{anno=#k{us=Ka#k.us,ns=[],a=Ka#k.a}, + vars=Vs,body=Body,ret=[]}, + #k_fdef{anno=Anno,func=Name,arity=Arity,vars=Vs,body=Match}. + + %% handle_reuse_annos([#k_var{}], State) -> State. %% In general, it is only safe to reuse a variable for a match context %% if the original value of the variable will no longer be needed. diff --git a/lib/compiler/src/v3_kernel.hrl b/lib/compiler/src/v3_kernel.hrl index 7cd30b25a8..87011b7680 100644 --- a/lib/compiler/src/v3_kernel.hrl +++ b/lib/compiler/src/v3_kernel.hrl @@ -79,7 +79,7 @@ -record(k_guard_clause, {anno=[],guard,body}). -record(k_break, {anno=[],args=[]}). --record(k_guard_break, {anno=[],args=[]}). +-record(k_guard_break, {anno=[],args=[],locked=[]}). -record(k_return, {anno=[],args=[]}). %%k_get_anno(Thing) -> element(2, Thing). diff --git a/lib/compiler/src/v3_life.erl b/lib/compiler/src/v3_life.erl deleted file mode 100644 index be3ade47ff..0000000000 --- a/lib/compiler/src/v3_life.erl +++ /dev/null @@ -1,468 +0,0 @@ -%% -%% %CopyrightBegin% -%% -%% Copyright Ericsson AB 1999-2016. All Rights Reserved. -%% -%% Licensed under the Apache License, Version 2.0 (the "License"); -%% you may not use this file except in compliance with the License. -%% You may obtain a copy of the License at -%% -%% http://www.apache.org/licenses/LICENSE-2.0 -%% -%% Unless required by applicable law or agreed to in writing, software -%% distributed under the License is distributed on an "AS IS" BASIS, -%% WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. -%% See the License for the specific language governing permissions and -%% limitations under the License. -%% -%% %CopyrightEnd% -%% -%% Purpose : Convert annotated kernel expressions to annotated beam format. - -%% This module creates beam format annotated with variable lifetime -%% information. Each thing is given an index and for each variable we -%% store the first and last index for its occurrence. The variable -%% database, VDB, attached to each thing is only relevant internally -%% for that thing. -%% -%% For nested things like matches the numbering continues locally and -%% the VDB for that thing refers to the variable usage within that -%% thing. Variables which live through a such a thing are internally -%% given a very large last index. Internally the indexes continue -%% after the index of that thing. This creates no problems as the -%% internal variable info never escapes and externally we only see -%% variable which are alive both before or after. -%% -%% This means that variables never "escape" from a thing and the only -%% way to get values from a thing is to "return" them, with 'break' or -%% 'return'. Externally these values become the return values of the -%% thing. This is no real limitation as most nested things have -%% multiple threads so working out a common best variable usage is -%% difficult. - --module(v3_life). - --export([module/2]). - --export([vdb_find/2]). - --import(lists, [member/2,map/2,reverse/1,sort/1]). --import(ordsets, [add_element/2,intersection/2,union/2]). - --include("v3_kernel.hrl"). --include("v3_life.hrl"). - --type fa() :: {atom(),arity()}. - -%% These are not defined in v3_kernel.hrl. -get_kanno(Kthing) -> element(2, Kthing). -%%set_kanno(Kthing, Anno) -> setelement(2, Kthing, Anno). - --spec module(#k_mdef{}, [compile:option()]) -> - {'ok',{module(),[fa()],[_],[_]}}. - -module(#k_mdef{name=M,exports=Es,attributes=As,body=Fs0}, _Opts) -> - Fs1 = functions(Fs0, []), - {ok,{M,Es,As,Fs1}}. - -functions([F|Fs], Acc) -> - functions(Fs, [function(F)|Acc]); -functions([], Acc) -> reverse(Acc). - -%% function(Kfunc) -> Func. - -function(#k_fdef{anno=#k{a=Anno},func=F,arity=Ar,vars=Vs,body=Kb}) -> - try - As = var_list(Vs), - Vdb0 = init_vars(As), - %% Force a top-level match! - B0 = case Kb of - #k_match{} -> Kb; - _ -> - Ka = get_kanno(Kb), - #k_match{anno=#k{us=Ka#k.us,ns=[],a=Ka#k.a}, - vars=Vs,body=Kb,ret=[]} - end, - {B1,_,Vdb1} = body(B0, 1, Vdb0), - {function,F,Ar,As,B1,Vdb1,Anno} - catch - Class:Error -> - Stack = erlang:get_stacktrace(), - io:fwrite("Function: ~w/~w\n", [F,Ar]), - erlang:raise(Class, Error, Stack) - end. - -%% body(Kbody, I, Vdb) -> {[Expr],MaxI,Vdb}. -%% Handle a body. - -body(#k_seq{arg=Ke,body=Kb}, I, Vdb0) -> - %%ok = io:fwrite("life ~w:~p~n", [?LINE,{Ke,I,Vdb0}]), - A = get_kanno(Ke), - Vdb1 = use_vars(union(A#k.us, A#k.ns), I, Vdb0), - {Es,MaxI,Vdb2} = body(Kb, I+1, Vdb1), - E = expr(Ke, I, Vdb2), - {[E|Es],MaxI,Vdb2}; -body(Ke, I, Vdb0) -> - %%ok = io:fwrite("life ~w:~p~n", [?LINE,{Ke,I,Vdb0}]), - A = get_kanno(Ke), - Vdb1 = use_vars(union(A#k.us, A#k.ns), I, Vdb0), - E = expr(Ke, I, Vdb1), - {[E],I,Vdb1}. - -%% protected(Kprotected, I, Vdb) -> Protected. -%% Only used in guards. - -protected(#k_protected{anno=A,arg=Ts,ret=Rs}, I, Vdb) -> - %% Lock variables that are alive before try and used afterwards. - %% Don't lock variables that are only used inside the protected - %% expression. - Pdb0 = vdb_sub(I, I+1, Vdb), - {T,MaxI,Pdb1} = body(Ts, I+1, Pdb0), - Pdb2 = use_vars(A#k.ns, MaxI+1, Pdb1), %Save "return" values - #l{ke={protected,T,var_list(Rs)},i=I,a=A#k.a,vdb=Pdb2}. - -%% expr(Kexpr, I, Vdb) -> Expr. - -expr(#k_test{anno=A,op=Op,args=As,inverted=Inverted}, I, _Vdb) -> - #l{ke={test,test_op(Op),atomic_list(As),Inverted},i=I,a=A#k.a}; -expr(#k_call{anno=A,op=Op,args=As,ret=Rs}, I, _Vdb) -> - #l{ke={call,call_op(Op),atomic_list(As),var_list(Rs)},i=I,a=A#k.a}; -expr(#k_enter{anno=A,op=Op,args=As}, I, _Vdb) -> - #l{ke={enter,call_op(Op),atomic_list(As)},i=I,a=A#k.a}; -expr(#k_bif{anno=A,op=Op,args=As,ret=Rs}, I, _Vdb) -> - Bif = k_bif(A, Op, As, Rs), - #l{ke=Bif,i=I,a=A#k.a}; -expr(#k_match{anno=A,body=Kb,ret=Rs}, I, Vdb) -> - %% Work out imported variables which need to be locked. - Mdb = vdb_sub(I, I+1, Vdb), - M = match(Kb, A#k.us, I+1, [], Mdb), - #l{ke={match,M,var_list(Rs)},i=I,vdb=use_vars(A#k.us, I+1, Mdb),a=A#k.a}; -expr(#k_guard_match{anno=A,body=Kb,ret=Rs}, I, Vdb) -> - %% Work out imported variables which need to be locked. - Mdb = vdb_sub(I, I+1, Vdb), - M = match(Kb, A#k.us, I+1, [], Mdb), - #l{ke={guard_match,M,var_list(Rs)},i=I,vdb=use_vars(A#k.us, I+1, Mdb),a=A#k.a}; -expr(#k_try{}=Try, I, Vdb) -> - body_try(Try, I, Vdb); -expr(#k_protected{}=Protected, I, Vdb) -> - protected(Protected, I, Vdb); -expr(#k_try_enter{anno=A,arg=Ka,vars=Vs,body=Kb,evars=Evs,handler=Kh}, I, Vdb) -> - %% Lock variables that are alive before the catch and used afterwards. - %% Don't lock variables that are only used inside the try. - Tdb0 = vdb_sub(I, I+1, Vdb), - %% This is the tricky bit. Lock variables in Arg that are used in - %% the body and handler. Add try tag 'variable'. - Ab = get_kanno(Kb), - Ah = get_kanno(Kh), - Tdb1 = use_vars(union(Ab#k.us, Ah#k.us), I+3, Tdb0), - Tdb2 = vdb_sub(I, I+2, Tdb1), - Vnames = fun (Kvar) -> Kvar#k_var.name end, %Get the variable names - {Aes,_,Adb} = body(Ka, I+2, add_var({catch_tag,I+1}, I+1, 1000000, Tdb2)), - {Bes,_,Bdb} = body(Kb, I+4, new_vars(sort(map(Vnames, Vs)), I+3, Tdb2)), - {Hes,_,Hdb} = body(Kh, I+4, new_vars(sort(map(Vnames, Evs)), I+3, Tdb2)), - #l{ke={try_enter,#l{ke={block,Aes},i=I+1,vdb=Adb,a=[]}, - var_list(Vs),#l{ke={block,Bes},i=I+3,vdb=Bdb,a=[]}, - var_list(Evs),#l{ke={block,Hes},i=I+3,vdb=Hdb,a=[]}}, - i=I,vdb=Tdb1,a=A#k.a}; -expr(#k_catch{anno=A,body=Kb,ret=[R]}, I, Vdb) -> - %% Lock variables that are alive before the catch and used afterwards. - %% Don't lock variables that are only used inside the catch. - %% Add catch tag 'variable'. - Cdb0 = vdb_sub(I, I+1, Vdb), - {Es,_,Cdb1} = body(Kb, I+1, add_var({catch_tag,I}, I, locked, Cdb0)), - #l{ke={'catch',Es,variable(R)},i=I,vdb=Cdb1,a=A#k.a}; -expr(#k_receive{anno=A,var=V,body=Kb,timeout=T,action=Ka,ret=Rs}, I, Vdb) -> - %% Work out imported variables which need to be locked. - Rdb = vdb_sub(I, I+1, Vdb), - M = match(Kb, add_element(V#k_var.name, A#k.us), I+1, [], - new_vars([V#k_var.name], I, Rdb)), - {Tes,_,Adb} = body(Ka, I+1, Rdb), - #l{ke={receive_loop,atomic(T),variable(V),M, - #l{ke=Tes,i=I+1,vdb=Adb,a=[]},var_list(Rs)}, - i=I,vdb=use_vars(A#k.us, I+1, Vdb),a=A#k.a}; -expr(#k_receive_accept{anno=A}, I, _Vdb) -> - #l{ke=receive_accept,i=I,a=A#k.a}; -expr(#k_receive_next{anno=A}, I, _Vdb) -> - #l{ke=receive_next,i=I,a=A#k.a}; -expr(#k_put{anno=A,arg=Arg,ret=Rs}, I, _Vdb) -> - #l{ke={set,var_list(Rs),literal(Arg, [])},i=I,a=A#k.a}; -expr(#k_break{anno=A,args=As}, I, _Vdb) -> - #l{ke={break,atomic_list(As)},i=I,a=A#k.a}; -expr(#k_guard_break{anno=A,args=As}, I, Vdb) -> - Locked = [V || {V,_,_} <- Vdb], - #l{ke={guard_break,atomic_list(As),Locked},i=I,a=A#k.a}; -expr(#k_return{anno=A,args=As}, I, _Vdb) -> - #l{ke={return,atomic_list(As)},i=I,a=A#k.a}. - -body_try(#k_try{anno=A,arg=Ka,vars=Vs,body=Kb,evars=Evs,handler=Kh,ret=Rs}, - I, Vdb) -> - %% Lock variables that are alive before the catch and used afterwards. - %% Don't lock variables that are only used inside the try. - Tdb0 = vdb_sub(I, I+1, Vdb), - %% This is the tricky bit. Lock variables in Arg that are used in - %% the body and handler. Add try tag 'variable'. - Ab = get_kanno(Kb), - Ah = get_kanno(Kh), - Tdb1 = use_vars(union(Ab#k.us, Ah#k.us), I+3, Tdb0), - Tdb2 = vdb_sub(I, I+2, Tdb1), - Vnames = fun (Kvar) -> Kvar#k_var.name end, %Get the variable names - {Aes,_,Adb} = body(Ka, I+2, add_var({catch_tag,I+1}, I+1, locked, Tdb2)), - {Bes,_,Bdb} = body(Kb, I+4, new_vars(sort(map(Vnames, Vs)), I+3, Tdb2)), - {Hes,_,Hdb} = body(Kh, I+4, new_vars(sort(map(Vnames, Evs)), I+3, Tdb2)), - #l{ke={'try',#l{ke={block,Aes},i=I+1,vdb=Adb,a=[]}, - var_list(Vs),#l{ke={block,Bes},i=I+3,vdb=Bdb,a=[]}, - var_list(Evs),#l{ke={block,Hes},i=I+3,vdb=Hdb,a=[]}, - var_list(Rs)}, - i=I,vdb=Tdb1,a=A#k.a}. - -%% call_op(Op) -> Op. -%% test_op(Op) -> Op. -%% Do any necessary name translations here to munge into beam format. - -call_op(#k_local{name=N}) -> N; -call_op(#k_remote{mod=M,name=N}) -> {remote,atomic(M),atomic(N)}; -call_op(Other) -> variable(Other). - -test_op(#k_remote{mod=#k_atom{val=erlang},name=#k_atom{val=N}}) -> N. - -%% k_bif(Anno, Op, [Arg], [Ret], Vdb) -> Expr. -%% Build bifs. - -k_bif(_A, #k_internal{name=Name}, As, Rs) -> - {internal,Name,atomic_list(As),var_list(Rs)}; -k_bif(_A, #k_remote{mod=#k_atom{val=erlang},name=#k_atom{val=Name}}, As, Rs) -> - Ar = length(As), - case is_gc_bif(Name, Ar) of - false -> - {bif,Name,atomic_list(As),var_list(Rs)}; - true -> - {gc_bif,Name,atomic_list(As),var_list(Rs)} - end. - -%% match(Kexpr, [LockVar], I, Vdb) -> Expr. -%% Convert match tree to old format. - -match(#k_alt{anno=A,first=Kf,then=Kt}, Ls, I, Ctxt, Vdb0) -> - Vdb1 = use_vars(union(A#k.us, Ls), I, Vdb0), - F = match(Kf, Ls, I+1, Ctxt, Vdb1), - T = match(Kt, Ls, I+1, Ctxt, Vdb1), - #l{ke={alt,F,T},i=I,vdb=Vdb1,a=A#k.a}; -match(#k_select{anno=A,var=V,types=Kts}, Ls0, I, Ctxt, Vdb0) -> - Vanno = get_kanno(V), - Ls1 = case member(no_usage, Vanno) of - false -> add_element(V#k_var.name, Ls0); - true -> Ls0 - end, - Anno = case member(reuse_for_context, Vanno) of - true -> [reuse_for_context|A#k.a]; - false -> A#k.a - end, - Vdb1 = use_vars(union(A#k.us, Ls1), I, Vdb0), - Ts = [type_clause(Tc, Ls1, I+1, Ctxt, Vdb1) || Tc <- Kts], - #l{ke={select,literal(V, Ctxt),Ts},i=I,vdb=Vdb1,a=Anno}; -match(#k_guard{anno=A,clauses=Kcs}, Ls, I, Ctxt, Vdb0) -> - Vdb1 = use_vars(union(A#k.us, Ls), I, Vdb0), - Cs = [guard_clause(G, Ls, I+1, Ctxt, Vdb1) || G <- Kcs], - #l{ke={guard,Cs},i=I,vdb=Vdb1,a=A#k.a}; -match(Other, Ls, I, _Ctxt, Vdb0) -> - Vdb1 = use_vars(Ls, I, Vdb0), - {B,_,Vdb2} = body(Other, I+1, Vdb1), - #l{ke={block,B},i=I,vdb=Vdb2,a=[]}. - -type_clause(#k_type_clause{anno=A,type=T,values=Kvs}, Ls, I, Ctxt, Vdb0) -> - %%ok = io:format("life ~w: ~p~n", [?LINE,{T,Kvs}]), - Vdb1 = use_vars(union(A#k.us, Ls), I+1, Vdb0), - Vs = [val_clause(Vc, Ls, I+1, Ctxt, Vdb1) || Vc <- Kvs], - #l{ke={type_clause,type(T),Vs},i=I,vdb=Vdb1,a=A#k.a}. - -val_clause(#k_val_clause{anno=A,val=V,body=Kb}, Ls0, I, Ctxt0, Vdb0) -> - New = (get_kanno(V))#k.ns, - Bus = (get_kanno(Kb))#k.us, - %%ok = io:format("Ls0 = ~p, Used=~p\n New=~p, Bus=~p\n", [Ls0,Used,New,Bus]), - Ls1 = union(intersection(New, Bus), Ls0), %Lock for safety - Vdb1 = use_vars(union(A#k.us, Ls1), I+1, new_vars(New, I, Vdb0)), - Ctxt = case V of - #k_binary{segs=#k_var{name=C0}} -> C0; - _ -> Ctxt0 - end, - B = match(Kb, Ls1, I+1, Ctxt, Vdb1), - #l{ke={val_clause,literal(V, Ctxt),B},i=I,vdb=use_vars(Bus, I+1, Vdb1),a=A#k.a}. - -guard_clause(#k_guard_clause{anno=A,guard=Kg,body=Kb}, Ls, I, Ctxt, Vdb0) -> - Vdb1 = use_vars(union(A#k.us, Ls), I+2, Vdb0), - Gdb = vdb_sub(I+1, I+2, Vdb1), - G = protected(Kg, I+1, Gdb), - B = match(Kb, Ls, I+2, Ctxt, Vdb1), - #l{ke={guard_clause,G,B}, - i=I,vdb=use_vars((get_kanno(Kg))#k.us, I+2, Vdb1), - a=A#k.a}. - -%% type(Ktype) -> Type. - -type(k_literal) -> literal; -type(k_int) -> integer; -%%type(k_char) -> integer; %Hhhmmm??? -type(k_float) -> float; -type(k_atom) -> atom; -type(k_nil) -> nil; -type(k_cons) -> cons; -type(k_tuple) -> tuple; -type(k_binary) -> binary; -type(k_bin_seg) -> bin_seg; -type(k_bin_int) -> bin_int; -type(k_bin_end) -> bin_end; -type(k_map) -> map. - -%% variable(Klit) -> Lit. -%% var_list([Klit]) -> [Lit]. - -variable(#k_var{name=N}) -> {var,N}. - -var_list(Ks) -> [variable(K) || K <- Ks]. - -%% atomic(Klit) -> Lit. -%% atomic_list([Klit]) -> [Lit]. - -atomic(#k_literal{val=V}) -> {literal,V}; -atomic(#k_var{name=N}) -> {var,N}; -atomic(#k_int{val=I}) -> {integer,I}; -atomic(#k_float{val=F}) -> {float,F}; -atomic(#k_atom{val=N}) -> {atom,N}; -%%atomic(#k_char{val=C}) -> {char,C}; -atomic(#k_nil{}) -> nil. - -atomic_list(Ks) -> [atomic(K) || K <- Ks]. - -%% literal(Klit) -> Lit. -%% literal_list([Klit]) -> [Lit]. - -literal(#k_var{name=N}, _) -> {var,N}; -literal(#k_literal{val=I}, _) -> {literal,I}; -literal(#k_int{val=I}, _) -> {integer,I}; -literal(#k_float{val=F}, _) -> {float,F}; -literal(#k_atom{val=N}, _) -> {atom,N}; -%%literal(#k_char{val=C}, _) -> {char,C}; -literal(#k_nil{}, _) -> nil; -literal(#k_cons{hd=H,tl=T}, Ctxt) -> - {cons,[literal(H, Ctxt),literal(T, Ctxt)]}; -literal(#k_binary{segs=V}, Ctxt) -> - {binary,literal(V, Ctxt)}; -literal(#k_bin_seg{size=S,unit=U,type=T,flags=Fs,seg=Seg,next=[]}, Ctxt) -> - %% Only occurs in patterns. - {bin_seg,Ctxt,literal(S, Ctxt),U,T,Fs,[literal(Seg, Ctxt)]}; -literal(#k_bin_seg{size=S,unit=U,type=T,flags=Fs,seg=Seg,next=N}, Ctxt) -> - {bin_seg,Ctxt,literal(S, Ctxt),U,T,Fs, - [literal(Seg, Ctxt),literal(N, Ctxt)]}; -literal(#k_bin_int{size=S,unit=U,flags=Fs,val=Int,next=N}, Ctxt) -> - %% Only occurs in patterns. - {bin_int,Ctxt,literal(S, Ctxt),U,Fs,Int, - [literal(N, Ctxt)]}; -literal(#k_bin_end{}, Ctxt) -> - {bin_end,Ctxt}; -literal(#k_tuple{es=Es}, Ctxt) -> - {tuple,literal_list(Es, Ctxt)}; -literal(#k_map{op=Op,var=Var,es=Es0}, Ctxt) -> - {map,Op,literal(Var, Ctxt),literal_list(Es0, Ctxt)}; -literal(#k_map_pair{key=K,val=V}, Ctxt) -> - {map_pair,literal(K, Ctxt),literal(V, Ctxt)}. - -literal_list(Ks, Ctxt) -> - [literal(K, Ctxt) || K <- Ks]. - - -%% is_gc_bif(Name, Arity) -> true|false -%% Determines whether the BIF Name/Arity might do a GC. - -is_gc_bif(hd, 1) -> false; -is_gc_bif(tl, 1) -> false; -is_gc_bif(self, 0) -> false; -is_gc_bif(node, 0) -> false; -is_gc_bif(node, 1) -> false; -is_gc_bif(element, 2) -> false; -is_gc_bif(get, 1) -> false; -is_gc_bif(tuple_size, 1) -> false; -is_gc_bif(Bif, Arity) -> - not (erl_internal:bool_op(Bif, Arity) orelse - erl_internal:new_type_test(Bif, Arity) orelse - erl_internal:comp_op(Bif, Arity)). - -%% Keep track of life time for variables. -%% -%% init_vars([{var,VarName}]) -> Vdb. -%% new_vars([VarName], I, Vdb) -> Vdb. -%% use_vars([VarName], I, Vdb) -> Vdb. -%% add_var(VarName, F, L, Vdb) -> Vdb. -%% -%% The list of variable names for new_vars/3 and use_vars/3 -%% must be sorted. - -init_vars(Vs) -> - vdb_new(Vs). - -new_vars([], _, Vdb) -> Vdb; -new_vars([V], I, Vdb) -> vdb_store_new(V, {V,I,I}, Vdb); -new_vars(Vs, I, Vdb) -> vdb_update_vars(Vs, Vdb, I). - -use_vars([], _, Vdb) -> - Vdb; -use_vars([V], I, Vdb) -> - case vdb_find(V, Vdb) of - {V,F,L} when I > L -> vdb_update(V, {V,F,I}, Vdb); - {V,_,_} -> Vdb; - error -> vdb_store_new(V, {V,I,I}, Vdb) - end; -use_vars(Vs, I, Vdb) -> vdb_update_vars(Vs, Vdb, I). - -add_var(V, F, L, Vdb) -> - vdb_store_new(V, {V,F,L}, Vdb). - -%% vdb - -vdb_new(Vs) -> - sort([{V,0,0} || {var,V} <- Vs]). - --type var() :: atom(). - --spec vdb_find(var(), [vdb_entry()]) -> 'error' | vdb_entry(). - -vdb_find(V, Vdb) -> - case lists:keyfind(V, 1, Vdb) of - false -> error; - Vd -> Vd - end. - -vdb_update(V, Update, [{V,_,_}|Vdb]) -> - [Update|Vdb]; -vdb_update(V, Update, [Vd|Vdb]) -> - [Vd|vdb_update(V, Update, Vdb)]. - -vdb_store_new(V, New, [{V1,_,_}=Vd|Vdb]) when V > V1 -> - [Vd|vdb_store_new(V, New, Vdb)]; -vdb_store_new(V, New, [{V1,_,_}|_]=Vdb) when V < V1 -> - [New|Vdb]; -vdb_store_new(_, New, []) -> [New]. - -vdb_update_vars([V|_]=Vs, [{V1,_,_}=Vd|Vdb], I) when V > V1 -> - [Vd|vdb_update_vars(Vs, Vdb, I)]; -vdb_update_vars([V|Vs], [{V1,_,_}|_]=Vdb, I) when V < V1 -> - %% New variable. - [{V,I,I}|vdb_update_vars(Vs, Vdb, I)]; -vdb_update_vars([V|Vs], [{_,F,L}=Vd|Vdb], I) -> - %% Existing variable. - if - I > L -> [{V,F,I}|vdb_update_vars(Vs, Vdb, I)]; - true -> [Vd|vdb_update_vars(Vs, Vdb, I)] - end; -vdb_update_vars([V|Vs], [], I) -> - %% New variable. - [{V,I,I}|vdb_update_vars(Vs, [], I)]; -vdb_update_vars([], Vdb, _) -> Vdb. - -%% vdb_sub(Min, Max, Vdb) -> Vdb. -%% Extract variables which are used before and after Min. Lock -%% variables alive after Max. - -vdb_sub(Min, Max, Vdb) -> - [ if L >= Max -> {V,F,locked}; - true -> Vd - end || {V,F,L}=Vd <- Vdb, F < Min, L >= Min ]. diff --git a/lib/compiler/test/compile_SUITE.erl b/lib/compiler/test/compile_SUITE.erl index 25983c6012..8fe2a93f95 100644 --- a/lib/compiler/test/compile_SUITE.erl +++ b/lib/compiler/test/compile_SUITE.erl @@ -120,9 +120,19 @@ file_1(Config) when is_list(Config) -> true = exists(Target), passed = run(Target, test, []), + %% Test option 'deterministic' as a compiler attribute. + Det = deterministic_module, + {DetPath, DetTarget} = get_files(Config, Det, "det_target"), + {ok,Det,DetCode} = compile:file(DetPath, [binary]), + {module,Det} = code:load_binary(Det, "", DetCode), + [{version,_}] = Det:module_info(compile), + true = code:delete(Det), + false = code:purge(Det), + %% Cleanup. ok = file:delete(Target), ok = file:del_dir(filename:dirname(Target)), + ok = file:del_dir(filename:dirname(DetTarget)), %% There should not be any messages in the messages. receive @@ -375,7 +385,6 @@ do_file_listings(DataDir, PrivDir, [File|Files]) -> do_listing(Simple, TargetDir, dcbsm, ".core_bsm"), do_listing(Simple, TargetDir, dsetel, ".dsetel"), do_listing(Simple, TargetDir, dkern, ".kernel"), - do_listing(Simple, TargetDir, dlife, ".life"), do_listing(Simple, TargetDir, dcg, ".codegen"), do_listing(Simple, TargetDir, dblk, ".block"), do_listing(Simple, TargetDir, dexcept, ".except"), @@ -1319,10 +1328,13 @@ do_warnings_2([], Next, F) -> %% pre-loads the modules that are used by a typical compilation. pre_load_check(Config) -> - case test_server:is_cover() of - true -> + case {test_server:is_cover(),code:module_info(native)} of + {true,_} -> {skip,"Cover is running"}; - false -> + {false,true} -> + %% Tracing won't work. + {skip,"'code' is native-compiled"}; + {false,false} -> try do_pre_load_check(Config) after diff --git a/lib/compiler/src/v3_life.hrl b/lib/compiler/test/compile_SUITE_data/deterministic_module.erl index 5c76312067..5e0e29c25e 100644 --- a/lib/compiler/src/v3_life.hrl +++ b/lib/compiler/test/compile_SUITE_data/deterministic_module.erl @@ -1,8 +1,8 @@ %% %% %CopyrightBegin% -%% -%% Copyright Ericsson AB 1999-2016. All Rights Reserved. -%% +%% +%% Copyright Ericsson AB 2017. All Rights Reserved. +%% %% Licensed under the Apache License, Version 2.0 (the "License"); %% you may not use this file except in compliance with the License. %% You may obtain a copy of the License at @@ -14,16 +14,8 @@ %% WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. %% See the License for the specific language governing permissions and %% limitations under the License. -%% +%% %% %CopyrightEnd% %% -%% This record contains variable life-time annotation for a -%% kernel expression. Added by v3_life, used by v3_codegen. - --type vdb_entry() :: {atom(),non_neg_integer(),non_neg_integer()}. - --record(l, {ke, %Kernel expression - i=0 :: non_neg_integer(), %Op number - vdb=[] :: [vdb_entry()], %Variable database - a=[] :: [term()]}). %Core annotation - +-module(deterministic_module). +-compile([deterministic]). diff --git a/lib/compiler/test/misc_SUITE.erl b/lib/compiler/test/misc_SUITE.erl index ea4aaf40a9..b12bcbeeab 100644 --- a/lib/compiler/test/misc_SUITE.erl +++ b/lib/compiler/test/misc_SUITE.erl @@ -171,7 +171,7 @@ silly_coverage(Config) when is_list(Config) -> expect_error(fun() -> sys_core_dsetel:module(BadCoreErlang, []) end), expect_error(fun() -> v3_kernel:module(BadCoreErlang, []) end), - %% v3_life + %% v3_codegen BadKernel = {k_mdef,[],?MODULE, [{foo,0}], [], @@ -179,11 +179,7 @@ silly_coverage(Config) when is_list(Config) -> {k,[],[],[]}, f,0,[], seriously_bad_body}]}, - expect_error(fun() -> v3_life:module(BadKernel, []) end), - - %% v3_codegen - CodegenInput = {?MODULE,[{foo,0}],[],[{function,foo,0,[a|b],a,b,[]}]}, - expect_error(fun() -> v3_codegen:module(CodegenInput, []) end), + expect_error(fun() -> v3_codegen:module(BadKernel, []) end), %% beam_a BeamAInput = {?MODULE,[{foo,0}],[], diff --git a/lib/crypto/c_src/Makefile.in b/lib/crypto/c_src/Makefile.in index af7c209c75..31124ba477 100644 --- a/lib/crypto/c_src/Makefile.in +++ b/lib/crypto/c_src/Makefile.in @@ -1,7 +1,7 @@ # # %CopyrightBegin% # -# Copyright Ericsson AB 1999-2016. All Rights Reserved. +# Copyright Ericsson AB 1999-2017. All Rights Reserved. # # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. @@ -78,12 +78,16 @@ CRYPTO_STATIC_OBJS = $(OBJDIR)/crypto_static$(TYPEMARKER).o\ NIF_ARCHIVE = $(LIBDIR)/crypto$(TYPEMARKER).a +TEST_ENGINE_OBJS = $(OBJDIR)/otp_test_engine$(TYPEMARKER).o + ifeq ($(findstring win32,$(TARGET)), win32) NIF_LIB = $(LIBDIR)/crypto$(TYPEMARKER).dll CALLBACK_LIB = $(LIBDIR)/crypto_callback$(TYPEMARKER).dll +TEST_ENGINE_LIB = $(LIBDIR)/otp_test_engine$(TYPEMARKER).dll else NIF_LIB = $(LIBDIR)/crypto$(TYPEMARKER).so CALLBACK_LIB = $(LIBDIR)/crypto_callback$(TYPEMARKER).so +TEST_ENGINE_LIB = $(LIBDIR)/otp_test_engine$(TYPEMARKER).so endif ifeq ($(HOST_OS),) @@ -129,10 +133,22 @@ ALL_STATIC_CFLAGS = $(DED_STATIC_CFLAGS) $(INCLUDES) _create_dirs := $(shell mkdir -p $(OBJDIR) $(LIBDIR)) -debug opt valgrind: $(NIF_LIB) $(CALLBACK_LIB) +debug opt valgrind: $(NIF_LIB) $(CALLBACK_LIB) $(TEST_ENGINE_LIB) static_lib: $(NIF_ARCHIVE) +$(OBJDIR)/otp_test_engine$(TYPEMARKER).o: otp_test_engine.c + $(V_at)$(INSTALL_DIR) $(OBJDIR) + $(V_CC) -c -o $@ $(ALL_CFLAGS) $< + +$(LIBDIR)/otp_test_engine$(TYPEMARKER).so: $(TEST_ENGINE_OBJS) + $(V_at)$(INSTALL_DIR) $(LIBDIR) + $(V_LD) $(LDFLAGS) -o $@ $^ $(LDLIBS) $(CRYPTO_LINK_LIB) + +$(LIBDIR)/otp_test_engine$(TYPEMARKER).dll: $(TEST_ENGINE_OBJS) + $(V_at)$(INSTALL_DIR) $(LIBDIR) + $(V_LD) $(LDFLAGS) -o $@ $(SSL_DED_LD_RUNTIME_LIBRARY_PATH) -L$(SSL_LIBDIR) $(TEST_ENGINE_OBJS) -l$(SSL_CRYPTO_LIBNAME) -l$(SSL_SSL_LIBNAME) + $(OBJDIR)/%$(TYPEMARKER).o: %.c $(V_at)$(INSTALL_DIR) $(OBJDIR) $(V_CC) -c -o $@ $(ALL_CFLAGS) $< @@ -170,6 +186,7 @@ ifeq ($(findstring win32,$(TARGET)), win32) rm -f $(LIBDIR)/crypto.debug.dll rm -f $(LIBDIR)/crypto_callback.dll rm -f $(LIBDIR)/crypto_callback.debug.dll + rm -f $(LIBDIR)/otp_test_engine.dll else rm -f $(LIBDIR)/crypto.so rm -f $(LIBDIR)/crypto.debug.so @@ -177,6 +194,7 @@ else rm -f $(LIBDIR)/crypto_callback.so rm -f $(LIBDIR)/crypto_callback.debug.so rm -f $(LIBDIR)/crypto_callback.valgrind.so + rm -f $(LIBDIR)/otp_test_engine.so endif rm -f $(OBJDIR)/crypto.o rm -f $(OBJDIR)/crypto_static.o @@ -187,6 +205,7 @@ endif rm -f $(OBJDIR)/crypto_callback.o rm -f $(OBJDIR)/crypto_callback.debug.o rm -f $(OBJDIR)/crypto_callback.valgrind.o + rm -f $(OBJDIR)/otp_test_engine.o rm -f core *~ docs: @@ -206,6 +225,8 @@ ifeq ($(DYNAMIC_CRYPTO_LIB),yes) $(INSTALL_PROGRAM) $(CALLBACK_OBJS) "$(RELSYSDIR)/priv/obj" $(INSTALL_PROGRAM) $(CALLBACK_LIB) "$(RELSYSDIR)/priv/lib" endif + $(INSTALL_PROGRAM) $(TEST_ENGINE_OBJS) "$(RELSYSDIR)/priv/obj" + $(INSTALL_PROGRAM) $(TEST_ENGINE_LIB) "$(RELSYSDIR)/priv/lib" release_docs_spec: diff --git a/lib/crypto/c_src/crypto.c b/lib/crypto/c_src/crypto.c index 53fe233790..b29c5082ba 100644 --- a/lib/crypto/c_src/crypto.c +++ b/lib/crypto/c_src/crypto.c @@ -1,4 +1,4 @@ -/* +/* * %CopyrightBegin% * * Copyright Ericsson AB 2010-2017. All Rights Reserved. @@ -19,8 +19,8 @@ */ /* - * Purpose: Dynamically loadable NIF library for cryptography. - * Based on OpenSSL. + * Purpose: Dynamically loadable NIF library for cryptography. + * Based on OpenSSL. */ #ifdef __WIN32__ @@ -60,6 +60,8 @@ #include <openssl/rand.h> #include <openssl/evp.h> #include <openssl/hmac.h> +#include <openssl/engine.h> +#include <openssl/err.h> /* Helper macro to construct a OPENSSL_VERSION_NUMBER. * See openssl/opensslv.h @@ -79,9 +81,9 @@ * * Therefor works tests like this as intendend: * OPENSSL_VERSION_NUMBER >= PACKED_OPENSSL_VERSION_PLAIN(1,0,0) - * (The test is for example "2.4.2" >= "1.0.0" although the test + * (The test is for example "2.4.2" >= "1.0.0" although the test * with the cloned OpenSSL test would be "1.0.1" >= "1.0.0") - * + * * But tests like this gives wrong result: * OPENSSL_VERSION_NUMBER < PACKED_OPENSSL_VERSION_PLAIN(1,1,0) * (The test is false since "2.4.2" < "1.1.0". It should have been @@ -119,6 +121,10 @@ #include <openssl/modes.h> #endif +#if OPENSSL_VERSION_NUMBER >= PACKED_OPENSSL_VERSION(0,9,8,'h') +#define HAS_ENGINE_SUPPORT +#endif + #include "crypto_callback.h" #if OPENSSL_VERSION_NUMBER >= PACKED_OPENSSL_VERSION_PLAIN(0,9,8) \ @@ -240,7 +246,7 @@ /* This shall correspond to the similar macro in crypto.erl */ /* Current value is: erlang:system_info(context_reductions) * 10 */ -#define MAX_BYTES_TO_NIF 20000 +#define MAX_BYTES_TO_NIF 20000 #define CONSUME_REDS(NifEnv, Ibin) \ do { \ @@ -277,7 +283,7 @@ static HMAC_CTX *HMAC_CTX_new() static void HMAC_CTX_free(HMAC_CTX *ctx) { HMAC_CTX_cleanup(ctx); - return CRYPTO_free(ctx); + CRYPTO_free(ctx); } #define EVP_MD_CTX_new() EVP_MD_CTX_create() @@ -342,6 +348,10 @@ static INLINE void RSA_get0_crt_params(const RSA *r, const BIGNUM **dmp1, const static INLINE int DSA_set0_key(DSA *d, BIGNUM *pub_key, BIGNUM *priv_key); static INLINE int DSA_set0_pqg(DSA *d, BIGNUM *p, BIGNUM *q, BIGNUM *g); +static INLINE void DSA_get0_pqg(const DSA *dsa, + const BIGNUM **p, const BIGNUM **q, const BIGNUM **g); +static INLINE void DSA_get0_key(const DSA *dsa, + const BIGNUM **pub_key, const BIGNUM **priv_key); static INLINE int DSA_set0_key(DSA *d, BIGNUM *pub_key, BIGNUM *priv_key) { @@ -358,6 +368,23 @@ static INLINE int DSA_set0_pqg(DSA *d, BIGNUM *p, BIGNUM *q, BIGNUM *g) return 1; } +static INLINE void +DSA_get0_pqg(const DSA *dsa, const BIGNUM **p, const BIGNUM **q, const BIGNUM **g) +{ + *p = dsa->p; + *q = dsa->q; + *g = dsa->g; +} + +static INLINE void +DSA_get0_key(const DSA *dsa, const BIGNUM **pub_key, const BIGNUM **priv_key) +{ + if (pub_key) *pub_key = dsa->pub_key; + if (priv_key) *priv_key = dsa->priv_key; +} + + + static INLINE int DH_set0_key(DH *dh, BIGNUM *pub_key, BIGNUM *priv_key); static INLINE int DH_set0_pqg(DH *dh, BIGNUM *p, BIGNUM *q, BIGNUM *g); static INLINE int DH_set_length(DH *dh, long length); @@ -387,6 +414,8 @@ static INLINE int DH_set_length(DH *dh, long length) return 1; } + + static INLINE void DH_get0_pqg(const DH *dh, const BIGNUM **p, const BIGNUM **q, const BIGNUM **g) { @@ -398,8 +427,8 @@ DH_get0_pqg(const DH *dh, const BIGNUM **p, const BIGNUM **q, const BIGNUM **g) static INLINE void DH_get0_key(const DH *dh, const BIGNUM **pub_key, const BIGNUM **priv_key) { - *pub_key = dh->pub_key; - *priv_key = dh->priv_key; + if (pub_key) *pub_key = dh->pub_key; + if (priv_key) *priv_key = dh->priv_key; } #else /* End of compatibility definitions. */ @@ -448,6 +477,7 @@ static ERL_NIF_TERM dh_generate_parameters_nif(ErlNifEnv* env, int argc, const E static ERL_NIF_TERM dh_check(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[]); static ERL_NIF_TERM dh_generate_key_nif(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[]); static ERL_NIF_TERM dh_compute_key_nif(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[]); +static ERL_NIF_TERM privkey_to_pubkey_nif(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[]); static ERL_NIF_TERM srp_value_B_nif(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[]); static ERL_NIF_TERM srp_user_secret_nif(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[]); static ERL_NIF_TERM srp_host_secret_nif(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[]); @@ -466,6 +496,22 @@ static ERL_NIF_TERM aes_gcm_decrypt_NO_EVP(ErlNifEnv* env, int argc, const ERL_N static ERL_NIF_TERM chacha20_poly1305_encrypt(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[]); static ERL_NIF_TERM chacha20_poly1305_decrypt(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[]); +static int get_engine_load_cmd_list(ErlNifEnv* env, const ERL_NIF_TERM term, char **cmds, int i); +static ERL_NIF_TERM engine_by_id_nif(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[]); +static ERL_NIF_TERM engine_init_nif(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[]); +static ERL_NIF_TERM engine_finish_nif(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[]); +static ERL_NIF_TERM engine_free_nif(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[]); +static ERL_NIF_TERM engine_load_dynamic_nif(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[]); +static ERL_NIF_TERM engine_ctrl_cmd_strings_nif(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[]); +static ERL_NIF_TERM engine_register_nif(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[]); +static ERL_NIF_TERM engine_unregister_nif(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[]); +static ERL_NIF_TERM engine_add_nif(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[]); +static ERL_NIF_TERM engine_remove_nif(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[]); +static ERL_NIF_TERM engine_get_first_nif(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[]); +static ERL_NIF_TERM engine_get_next_nif(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[]); +static ERL_NIF_TERM engine_get_id_nif(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[]); +static ERL_NIF_TERM engine_get_all_methods_nif(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[]); + /* helpers */ static void init_algorithms_types(ErlNifEnv*); static void init_digest_types(ErlNifEnv* env); @@ -477,6 +523,10 @@ static int term2point(ErlNifEnv* env, ERL_NIF_TERM term, #endif static ERL_NIF_TERM bin_from_bn(ErlNifEnv* env, const BIGNUM *bn); +#ifdef HAS_ENGINE_SUPPORT +static int zero_terminate(ErlNifBinary bin, char **buf); +#endif + static int library_refc = 0; /* number of users of this dynamic library */ static ErlNifFunc nif_funcs[] = { @@ -516,6 +566,7 @@ static ErlNifFunc nif_funcs[] = { {"dh_check", 1, dh_check}, {"dh_generate_key_nif", 4, dh_generate_key_nif}, {"dh_compute_key_nif", 3, dh_compute_key_nif}, + {"privkey_to_pubkey_nif", 2, privkey_to_pubkey_nif}, {"srp_value_B_nif", 5, srp_value_B_nif}, {"srp_user_secret_nif", 7, srp_user_secret_nif}, {"srp_host_secret_nif", 5, srp_host_secret_nif}, @@ -529,12 +580,27 @@ static ErlNifFunc nif_funcs[] = { {"aes_gcm_decrypt", 5, aes_gcm_decrypt}, {"chacha20_poly1305_encrypt", 4, chacha20_poly1305_encrypt}, - {"chacha20_poly1305_decrypt", 5, chacha20_poly1305_decrypt} + {"chacha20_poly1305_decrypt", 5, chacha20_poly1305_decrypt}, + + {"engine_by_id_nif", 1, engine_by_id_nif}, + {"engine_init_nif", 1, engine_init_nif}, + {"engine_finish_nif", 1, engine_finish_nif}, + {"engine_free_nif", 1, engine_free_nif}, + {"engine_load_dynamic_nif", 0, engine_load_dynamic_nif}, + {"engine_ctrl_cmd_strings_nif", 2, engine_ctrl_cmd_strings_nif}, + {"engine_register_nif", 2, engine_register_nif}, + {"engine_unregister_nif", 2, engine_unregister_nif}, + {"engine_add_nif", 1, engine_add_nif}, + {"engine_remove_nif", 1, engine_remove_nif}, + {"engine_get_first_nif", 0, engine_get_first_nif}, + {"engine_get_next_nif", 1, engine_get_next_nif}, + {"engine_get_id_nif", 1, engine_get_id_nif}, + {"engine_get_all_methods_nif", 0, engine_get_all_methods_nif} + }; ERL_NIF_INIT(crypto,nif_funcs,load,NULL,upgrade,unload) - #define MD5_CTX_LEN (sizeof(MD5_CTX)) #define MD4_CTX_LEN (sizeof(MD4_CTX)) #define RIPEMD160_CTX_LEN (sizeof(RIPEMD160_CTX)) @@ -603,7 +669,33 @@ static ERL_NIF_TERM atom_sha512; static ERL_NIF_TERM atom_md5; static ERL_NIF_TERM atom_ripemd160; - +#ifdef HAS_ENGINE_SUPPORT +static ERL_NIF_TERM atom_bad_engine_method; +static ERL_NIF_TERM atom_bad_engine_id; +static ERL_NIF_TERM atom_ctrl_cmd_failed; +static ERL_NIF_TERM atom_engine_init_failed; +static ERL_NIF_TERM atom_register_engine_failed; +static ERL_NIF_TERM atom_add_engine_failed; +static ERL_NIF_TERM atom_remove_engine_failed; +static ERL_NIF_TERM atom_engine_method_not_supported; + +static ERL_NIF_TERM atom_engine_method_rsa; +static ERL_NIF_TERM atom_engine_method_dsa; +static ERL_NIF_TERM atom_engine_method_dh; +static ERL_NIF_TERM atom_engine_method_rand; +static ERL_NIF_TERM atom_engine_method_ecdh; +static ERL_NIF_TERM atom_engine_method_ecdsa; +static ERL_NIF_TERM atom_engine_method_ciphers; +static ERL_NIF_TERM atom_engine_method_digests; +static ERL_NIF_TERM atom_engine_method_store; +static ERL_NIF_TERM atom_engine_method_pkey_meths; +static ERL_NIF_TERM atom_engine_method_pkey_asn1_meths; +static ERL_NIF_TERM atom_engine_method_ec; + +static ERL_NIF_TERM atom_engine; +static ERL_NIF_TERM atom_key_id; +static ERL_NIF_TERM atom_password; +#endif static ErlNifResourceType* hmac_context_rtype; struct hmac_context @@ -728,11 +820,13 @@ static struct cipher_type_t cipher_types[] = static struct cipher_type_t* get_cipher_type(ERL_NIF_TERM type, size_t key_len); + /* #define PRINTF_ERR0(FMT) enif_fprintf(stderr, FMT "\n") #define PRINTF_ERR1(FMT, A1) enif_fprintf(stderr, FMT "\n", A1) #define PRINTF_ERR2(FMT, A1, A2) enif_fprintf(stderr, FMT "\n", A1, A2) */ + #define PRINTF_ERR0(FMT) #define PRINTF_ERR1(FMT,A1) #define PRINTF_ERR2(FMT,A1,A2) @@ -758,6 +852,23 @@ static void evp_cipher_ctx_dtor(ErlNifEnv* env, struct evp_cipher_ctx* ctx) { } #endif +// Engine +#ifdef HAS_ENGINE_SUPPORT +static ErlNifResourceType* engine_ctx_rtype; +struct engine_ctx { + ENGINE *engine; + char *id; +}; +static void engine_ctx_dtor(ErlNifEnv* env, struct engine_ctx* ctx) { + PRINTF_ERR0("engine_ctx_dtor"); + if(ctx->id) { + PRINTF_ERR1(" non empty ctx->id=%s", ctx->id); + enif_free(ctx->id); + } else + PRINTF_ERR0(" empty ctx->id=NULL"); +} +#endif + static int verify_lib_version(void) { const unsigned long libv = SSLeay(); @@ -793,7 +904,7 @@ static char crypto_callback_name[] = "crypto_callback"; static int change_basename(ErlNifBinary* bin, char* buf, int bufsz, const char* newfile) { int i; - + for (i = bin->size; i > 0; i--) { if (bin->data[i-1] == '/') break; @@ -869,12 +980,23 @@ static int initialize(ErlNifEnv* env, ERL_NIF_TERM load_info) return __LINE__; } #endif +#ifdef HAS_ENGINE_SUPPORT + engine_ctx_rtype = enif_open_resource_type(env, NULL, "ENGINE_CTX", + (ErlNifResourceDtor*) engine_ctx_dtor, + ERL_NIF_RT_CREATE|ERL_NIF_RT_TAKEOVER, + NULL); + if (!engine_ctx_rtype) { + PRINTF_ERR0("CRYPTO: Could not open resource type 'ENGINE_CTX'"); + return __LINE__; + } + if (library_refc > 0) { /* Repeated loading of this library (module upgrade). * Atoms and callbacks are already set, we are done. */ return 0; } +#endif atom_true = enif_make_atom(env,"true"); atom_false = enif_make_atom(env,"false"); @@ -952,6 +1074,33 @@ static int initialize(ErlNifEnv* env, ERL_NIF_TERM load_info) atom_md5 = enif_make_atom(env,"md5"); atom_ripemd160 = enif_make_atom(env,"ripemd160"); +#ifdef HAS_ENGINE_SUPPORT + atom_bad_engine_method = enif_make_atom(env,"bad_engine_method"); + atom_bad_engine_id = enif_make_atom(env,"bad_engine_id"); + atom_ctrl_cmd_failed = enif_make_atom(env,"ctrl_cmd_failed"); + atom_engine_init_failed = enif_make_atom(env,"engine_init_failed"); + atom_engine_method_not_supported = enif_make_atom(env,"engine_method_not_supported"); + atom_add_engine_failed = enif_make_atom(env,"add_engine_failed"); + atom_remove_engine_failed = enif_make_atom(env,"remove_engine_failed"); + + atom_engine_method_rsa = enif_make_atom(env,"engine_method_rsa"); + atom_engine_method_dsa = enif_make_atom(env,"engine_method_dsa"); + atom_engine_method_dh = enif_make_atom(env,"engine_method_dh"); + atom_engine_method_rand = enif_make_atom(env,"engine_method_rand"); + atom_engine_method_ecdh = enif_make_atom(env,"engine_method_ecdh"); + atom_engine_method_ecdsa = enif_make_atom(env,"engine_method_ecdsa"); + atom_engine_method_store = enif_make_atom(env,"engine_method_store"); + atom_engine_method_ciphers = enif_make_atom(env,"engine_method_ciphers"); + atom_engine_method_digests = enif_make_atom(env,"engine_method_digests"); + atom_engine_method_pkey_meths = enif_make_atom(env,"engine_method_pkey_meths"); + atom_engine_method_pkey_asn1_meths = enif_make_atom(env,"engine_method_pkey_asn1_meths"); + atom_engine_method_ec = enif_make_atom(env,"engine_method_ec"); + + atom_engine = enif_make_atom(env,"engine"); + atom_key_id = enif_make_atom(env,"key_id"); + atom_password = enif_make_atom(env,"password"); +#endif + init_digest_types(env); init_cipher_types(env); init_algorithms_types(env); @@ -973,24 +1122,24 @@ static int initialize(ErlNifEnv* env, ERL_NIF_TERM load_info) #else /* !HAVE_DYNAMIC_CRYPTO_LIB */ funcp = &get_crypto_callbacks; #endif - + #ifdef OPENSSL_THREADS enif_system_info(&sys_info, sizeof(sys_info)); if (sys_info.scheduler_threads > 1) { - nlocks = CRYPTO_num_locks(); + nlocks = CRYPTO_num_locks(); } /* else no need for locks */ #endif - + ccb = (*funcp)(nlocks); - + if (!ccb || ccb->sizeof_me != sizeof(*ccb)) { PRINTF_ERR0("Invalid 'crypto_callbacks'"); return __LINE__; } - + CRYPTO_set_mem_functions(ccb->crypto_alloc, ccb->crypto_realloc, ccb->crypto_free); - + #ifdef OPENSSL_THREADS if (nlocks > 0) { CRYPTO_set_locking_callback(ccb->locking_function); @@ -1186,11 +1335,11 @@ static ERL_NIF_TERM info_lib(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[] * Version string is still from library though. */ - memcpy(enif_make_new_binary(env, name_sz, &name_term), libname, name_sz); + memcpy(enif_make_new_binary(env, name_sz, &name_term), libname, name_sz); memcpy(enif_make_new_binary(env, ver_sz, &ver_term), ver, ver_sz); return enif_make_list1(env, enif_make_tuple3(env, name_term, - enif_make_int(env, ver_num), + enif_make_int(env, ver_num), ver_term)); } @@ -1225,6 +1374,8 @@ static ERL_NIF_TERM enable_fips_mode(ErlNifEnv* env, int argc, const ERL_NIF_TER } } + +#if defined(HAVE_EC) static ERL_NIF_TERM make_badarg_maybe(ErlNifEnv* env) { ERL_NIF_TERM reason; @@ -1233,6 +1384,7 @@ static ERL_NIF_TERM make_badarg_maybe(ErlNifEnv* env) else return enif_make_badarg(env); } +#endif static ERL_NIF_TERM hash_nif(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[]) {/* (Type, Data) */ @@ -1668,7 +1820,7 @@ static ERL_NIF_TERM hmac_update_nif(ErlNifEnv* env, int argc, const ERL_NIF_TERM {/* (Context, Data) */ ErlNifBinary data; struct hmac_context* obj; - + if (!enif_get_resource(env, argv[0], hmac_context_rtype, (void**)&obj) || !enif_inspect_iolist_as_binary(env, argv[1], &data)) { return enif_make_badarg(env); @@ -1704,13 +1856,13 @@ static ERL_NIF_TERM hmac_final_nif(ErlNifEnv* env, int argc, const ERL_NIF_TERM enif_mutex_unlock(obj->mtx); return enif_make_badarg(env); } - + HMAC_Final(obj->ctx, mac_buf, &mac_len); HMAC_CTX_free(obj->ctx); obj->alive = 0; enif_mutex_unlock(obj->mtx); - if (argc == 2 && req_len < mac_len) { + if (argc == 2 && req_len < mac_len) { /* Only truncate to req_len bytes if asked. */ mac_len = req_len; } @@ -2021,7 +2173,7 @@ static ERL_NIF_TERM aes_ctr_stream_init(ErlNifEnv* env, int argc, const ERL_NIF_ } static ERL_NIF_TERM aes_ctr_stream_encrypt(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[]) -{/* ({Key, IVec, ECount, Num}, Data) */ +{/* ({Key, IVec, ECount, Num}, Data) */ ErlNifBinary key_bin, ivec_bin, text_bin, ecount_bin; AES_KEY aes_key; unsigned int num; @@ -2042,14 +2194,14 @@ static ERL_NIF_TERM aes_ctr_stream_encrypt(ErlNifEnv* env, int argc, const ERL_N return enif_make_badarg(env); } - ivec2_buf = enif_make_new_binary(env, ivec_bin.size, &ivec2_term); + ivec2_buf = enif_make_new_binary(env, ivec_bin.size, &ivec2_term); ecount2_buf = enif_make_new_binary(env, ecount_bin.size, &ecount2_term); - + memcpy(ivec2_buf, ivec_bin.data, 16); memcpy(ecount2_buf, ecount_bin.data, ecount_bin.size); AES_ctr128_encrypt((unsigned char *) text_bin.data, - enif_make_new_binary(env, text_bin.size, &cipher_term), + enif_make_new_binary(env, text_bin.size, &cipher_term), text_bin.size, &aes_key, ivec2_buf, ecount2_buf, &num); num2_term = enif_make_uint(env, num); @@ -2352,7 +2504,7 @@ out_err: } static ERL_NIF_TERM strong_rand_bytes_nif(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[]) -{/* (Bytes) */ +{/* (Bytes) */ unsigned bytes; unsigned char* data; ERL_NIF_TERM ret; @@ -2446,7 +2598,7 @@ static ERL_NIF_TERM rand_uniform_nif(ErlNifEnv* env, int argc, const ERL_NIF_TER bn_to = BN_new(); BN_sub(bn_to, bn_rand, bn_from); - BN_pseudo_rand_range(bn_rand, bn_to); + BN_pseudo_rand_range(bn_rand, bn_to); BN_add(bn_rand, bn_rand, bn_from); dlen = BN_num_bytes(bn_rand); data = enif_make_new_binary(env, dlen+4, &ret); @@ -2464,7 +2616,7 @@ static ERL_NIF_TERM mod_exp_nif(ErlNifEnv* env, int argc, const ERL_NIF_TERM arg BIGNUM *bn_base=NULL, *bn_exponent=NULL, *bn_modulo=NULL, *bn_result; BN_CTX *bn_ctx; unsigned char* ptr; - unsigned dlen; + unsigned dlen; unsigned bin_hdr; /* return type: 0=plain binary, 4: mpint */ unsigned extra_byte; ERL_NIF_TERM ret; @@ -2485,7 +2637,7 @@ static ERL_NIF_TERM mod_exp_nif(ErlNifEnv* env, int argc, const ERL_NIF_TERM arg dlen = BN_num_bytes(bn_result); extra_byte = bin_hdr && BN_is_bit_set(bn_result, dlen*8-1); ptr = enif_make_new_binary(env, bin_hdr+extra_byte+dlen, &ret); - if (bin_hdr) { + if (bin_hdr) { put_int32(ptr, extra_byte+dlen); ptr[4] = 0; /* extra zeroed byte to ensure a positive mpint */ ptr += bin_hdr + extra_byte; @@ -2545,6 +2697,7 @@ static struct cipher_type_t* get_cipher_type(ERL_NIF_TERM type, size_t key_len) return NULL; } + static ERL_NIF_TERM do_exor(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[]) {/* (Data1, Data2) */ ErlNifBinary d1, d2; @@ -2578,7 +2731,7 @@ static ERL_NIF_TERM rc4_set_key(ErlNifEnv* env, int argc, const ERL_NIF_TERM arg return enif_make_badarg(env); } RC4_set_key((RC4_KEY*)enif_make_new_binary(env, sizeof(RC4_KEY), &ret), - key.size, key.data); + key.size, key.data); return ret; #else return enif_raise_exception(env, atom_notsup); @@ -2846,7 +2999,7 @@ static ERL_NIF_TERM rsa_generate_key_nif(ErlNifEnv* env, int argc, const ERL_NIF static ERL_NIF_TERM dh_generate_parameters_nif(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[]) {/* (PrimeLen, Generator) */ int prime_len, generator; - DH* dh_params; + DH* dh_params = NULL; int p_len, g_len; unsigned char *p_ptr, *g_ptr; ERL_NIF_TERM ret_p, ret_g; @@ -2857,8 +3010,8 @@ static ERL_NIF_TERM dh_generate_parameters_nif(ErlNifEnv* env, int argc, const E return enif_make_badarg(env); } - dh_params = DH_generate_parameters(prime_len, generator, NULL, NULL); - if (dh_params == NULL) { + + if (DH_generate_parameters_ex(dh_params, prime_len, generator, NULL)) { return atom_error; } DH_get0_pqg(dh_params, &dh_p, &dh_q, &dh_g); @@ -2871,7 +3024,7 @@ static ERL_NIF_TERM dh_generate_parameters_nif(ErlNifEnv* env, int argc, const E BN_bn2bin(dh_g, g_ptr); ERL_VALGRIND_MAKE_MEM_DEFINED(p_ptr, p_len); ERL_VALGRIND_MAKE_MEM_DEFINED(g_ptr, g_len); - return enif_make_list2(env, ret_p, ret_g); + return enif_make_list2(env, ret_p, ret_g); } static ERL_NIF_TERM dh_check(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[]) @@ -2881,9 +3034,9 @@ static ERL_NIF_TERM dh_check(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[] ERL_NIF_TERM ret, head, tail; BIGNUM *dh_p, *dh_g; - if (!enif_get_list_cell(env, argv[0], &head, &tail) + if (!enif_get_list_cell(env, argv[0], &head, &tail) || !get_bn_from_bin(env, head, &dh_p) - || !enif_get_list_cell(env, tail, &head, &tail) + || !enif_get_list_cell(env, tail, &head, &tail) || !get_bn_from_bin(env, head, &dh_g) || !enif_is_empty_list(env,tail)) { @@ -2900,12 +3053,12 @@ static ERL_NIF_TERM dh_check(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[] else if (i & DH_NOT_SUITABLE_GENERATOR) ret = atom_not_suitable_generator; else ret = enif_make_tuple2(env, atom_unknown, enif_make_uint(env, i)); } - else { /* Check Failed */ + else { /* Check Failed */ ret = enif_make_tuple2(env, atom_error, atom_check_failed); } DH_free(dh_params); return ret; -} +} static ERL_NIF_TERM dh_generate_key_nif(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[]) {/* (PrivKey|undefined, DHParams=[P,G], Mpint, Len|0) */ @@ -3007,7 +3160,7 @@ static ERL_NIF_TERM dh_compute_key_nif(ErlNifEnv* env, int argc, const ERL_NIF_T i = DH_compute_key(ret_bin.data, other_pub_key, dh_params); if (i > 0) { if (i != ret_bin.size) { - enif_realloc_binary(&ret_bin, i); + enif_realloc_binary(&ret_bin, i); } ret = enif_make_binary(env, &ret_bin); } @@ -3803,9 +3956,69 @@ static int get_pkey_sign_options(ErlNifEnv *env, ERL_NIF_TERM algorithm, ERL_NIF } +#ifdef HAS_ENGINE_SUPPORT +static int get_engine_and_key_id(ErlNifEnv *env, ERL_NIF_TERM key, char ** id, ENGINE **e) +{ + ERL_NIF_TERM engine_res, key_id_term; + struct engine_ctx *ctx; + ErlNifBinary key_id_bin; + + if (!enif_get_map_value(env, key, atom_engine, &engine_res) || + !enif_get_resource(env, engine_res, engine_ctx_rtype, (void**)&ctx) || + !enif_get_map_value(env, key, atom_key_id, &key_id_term) || + !enif_inspect_binary(env, key_id_term, &key_id_bin)) { + return 0; + } + else { + *e = ctx->engine; + return zero_terminate(key_id_bin, id); + } +} + + +static char *get_key_password(ErlNifEnv *env, ERL_NIF_TERM key) { + ERL_NIF_TERM tmp_term; + ErlNifBinary pwd_bin; + char *pwd; + if (enif_get_map_value(env, key, atom_password, &tmp_term) && + enif_inspect_binary(env, tmp_term, &pwd_bin) && + zero_terminate(pwd_bin, &pwd) + ) return pwd; + + return NULL; +} + +static int zero_terminate(ErlNifBinary bin, char **buf) { + *buf = enif_alloc(bin.size+1); + if (!*buf) + return 0; + memcpy(*buf, bin.data, bin.size); + *(*buf+bin.size) = 0; + return 1; +} +#endif + static int get_pkey_private_key(ErlNifEnv *env, ERL_NIF_TERM algorithm, ERL_NIF_TERM key, EVP_PKEY **pkey) { - if (algorithm == atom_rsa) { + if (enif_is_map(env, key)) { +#ifdef HAS_ENGINE_SUPPORT + /* Use key stored in engine */ + ENGINE *e; + char *id; + char *password; + + if (!get_engine_and_key_id(env, key, &id, &e)) + return PKEY_BADARG; + password = get_key_password(env, key); + *pkey = ENGINE_load_private_key(e, id, NULL, password); + if (!pkey) + return PKEY_BADARG; + enif_free(id); +#else + return PKEY_BADARG; +#endif + } + else if (algorithm == atom_rsa) { RSA *rsa = RSA_new(); if (!get_rsa_private_key(env, key, rsa)) { @@ -3866,7 +4079,24 @@ static int get_pkey_private_key(ErlNifEnv *env, ERL_NIF_TERM algorithm, ERL_NIF_ static int get_pkey_public_key(ErlNifEnv *env, ERL_NIF_TERM algorithm, ERL_NIF_TERM key, EVP_PKEY **pkey) { - if (algorithm == atom_rsa) { + if (enif_is_map(env, key)) { +#ifdef HAS_ENGINE_SUPPORT + /* Use key stored in engine */ + ENGINE *e; + char *id; + char *password; + + if (!get_engine_and_key_id(env, key, &id, &e)) + return PKEY_BADARG; + password = get_key_password(env, key); + *pkey = ENGINE_load_public_key(e, id, NULL, password); + if (!pkey) + return PKEY_BADARG; + enif_free(id); +#else + return PKEY_BADARG; +#endif + } else if (algorithm == atom_rsa) { RSA *rsa = RSA_new(); if (!get_rsa_public_key(env, key, rsa)) { @@ -3924,7 +4154,7 @@ static int get_pkey_public_key(ErlNifEnv *env, ERL_NIF_TERM algorithm, ERL_NIF_T } static ERL_NIF_TERM pkey_sign_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]) -{/* (Algorithm, Type, Data|{digest,Digest}, Key, Options) */ +{/* (Algorithm, Type, Data|{digest,Digest}, Key|#{}, Options) */ int i; const EVP_MD *md = NULL; unsigned char md_value[EVP_MAX_MD_SIZE]; @@ -3944,6 +4174,13 @@ enif_get_atom(env,argv[0],buf,1024,ERL_NIF_LATIN1); printf("algo=%s ",buf); enif_get_atom(env,argv[1],buf,1024,ERL_NIF_LATIN1); printf("hash=%s ",buf); printf("\r\n"); */ + +#ifndef HAS_ENGINE_SUPPORT + if (enif_is_map(env, argv[3])) { + return atom_notsup; + } +#endif + i = get_pkey_sign_digest(env, argv[0], argv[1], argv[2], md_value, &md, &tbs, &tbslen); if (i != PKEY_OK) { if (i == PKEY_NOTSUP) @@ -3965,10 +4202,9 @@ printf("\r\n"); } #ifdef HAS_EVP_PKEY_CTX -/* printf("EVP interface\r\n"); - */ ctx = EVP_PKEY_CTX_new(pkey, NULL); if (!ctx) goto badarg; + if (EVP_PKEY_sign_init(ctx) <= 0) goto badarg; if (md != NULL && EVP_PKEY_CTX_set_signature_md(ctx, md) <= 0) goto badarg; @@ -4070,6 +4306,12 @@ static ERL_NIF_TERM pkey_verify_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM unsigned char *tbs; /* data to be signed */ size_t tbslen; +#ifndef HAS_ENGINE_SUPPORT + if (enif_is_map(env, argv[4])) { + return atom_notsup; + } +#endif + if (!enif_inspect_binary(env, argv[3], &sig_bin)) { return enif_make_badarg(env); } @@ -4095,7 +4337,7 @@ static ERL_NIF_TERM pkey_verify_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM } #ifdef HAS_EVP_PKEY_CTX -/* printf("EVP interface\r\n"); +/* printf("EVP interface\r\n"); */ ctx = EVP_PKEY_CTX_new(pkey, NULL); if (!ctx) goto badarg; @@ -4280,7 +4522,13 @@ static ERL_NIF_TERM pkey_crypt_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM int algo_init = 0; /* char algo[1024]; */ - + +#ifndef HAS_ENGINE_SUPPORT + if (enif_is_map(env, argv[2])) { + return atom_notsup; + } +#endif + if (!enif_inspect_binary(env, argv[1], &in_bin)) { return enif_make_badarg(env); } @@ -4542,6 +4790,83 @@ static ERL_NIF_TERM pkey_crypt_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM /*--------------------------------*/ +static ERL_NIF_TERM privkey_to_pubkey_nif(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[]) +{ /* (Algorithm, PrivKey | KeyMap) */ + EVP_PKEY *pkey; + ERL_NIF_TERM alg = argv[0]; + ERL_NIF_TERM result[8]; + + if (get_pkey_private_key(env, alg, argv[1], &pkey) != PKEY_OK) { + return enif_make_badarg(env); + } + + if (alg == atom_rsa) { + const BIGNUM *n = NULL, *e = NULL, *d = NULL; + RSA *rsa = EVP_PKEY_get1_RSA(pkey); + if (rsa) { + RSA_get0_key(rsa, &n, &e, &d); + result[0] = bin_from_bn(env, e); // Exponent E + result[1] = bin_from_bn(env, n); // Modulus N = p*q + EVP_PKEY_free(pkey); + return enif_make_list_from_array(env, result, 2); + } + + } else if (argv[0] == atom_dss) { + const BIGNUM *p = NULL, *q = NULL, *g = NULL, *pub_key = NULL; + DSA *dsa = EVP_PKEY_get1_DSA(pkey); + if (dsa) { + DSA_get0_pqg(dsa, &p, &q, &g); + DSA_get0_key(dsa, &pub_key, NULL); + result[0] = bin_from_bn(env, p); + result[1] = bin_from_bn(env, q); + result[2] = bin_from_bn(env, g); + result[3] = bin_from_bn(env, pub_key); + EVP_PKEY_free(pkey); + return enif_make_list_from_array(env, result, 4); + } + + } else if (argv[0] == atom_ecdsa) { +#if defined(HAVE_EC) + EC_KEY *ec = EVP_PKEY_get1_EC_KEY(pkey); + if (ec) { + /* Example of result: + { + Curve = {Field, Prime, Point, Order, CoFactor} = + { + Field = {prime_field,<<255,...,255>>}, + Prime = {<<255,...,252>>, + <<90,...,75>>, + <<196,...,144>> + }, + Point = <<4,...,245>>, + Order = <<255,...,81>>, + CoFactor = <<1>> + }, + Key = <<151,...,62>> + } + or + { + Curve = + {characteristic_two_field, + M, + Basis = {tpbasis, _} + | {ppbasis, k1, k2, k3} + }, + Key + } + */ + EVP_PKEY_free(pkey); + return atom_notsup; + } +#else + EVP_PKEY_free(pkey); + return atom_notsup; +#endif + } + + if (pkey) EVP_PKEY_free(pkey); + return enif_make_badarg(env); +} /*================================================================*/ @@ -4554,3 +4879,598 @@ static ERL_NIF_TERM rand_seed_nif(ErlNifEnv* env, int argc, const ERL_NIF_TERM a RAND_seed(seed_bin.data,seed_bin.size); return atom_ok; } + +/*================================================================*/ +/* Engine */ +/*================================================================*/ +static ERL_NIF_TERM engine_by_id_nif(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[]) +{/* (EngineId) */ +#ifdef HAS_ENGINE_SUPPORT + ERL_NIF_TERM ret; + ErlNifBinary engine_id_bin; + unsigned int engine_id_len = 0; + char *engine_id; + ENGINE *engine; + struct engine_ctx *ctx; + + // Get Engine Id + if(!enif_inspect_binary(env, argv[0], &engine_id_bin)) { + PRINTF_ERR0("engine_by_id_nif Leaved: badarg"); + return enif_make_badarg(env); + } else { + engine_id_len = engine_id_bin.size+1; + engine_id = enif_alloc(engine_id_len); + (void) memcpy(engine_id, engine_id_bin.data, engine_id_len); + engine_id[engine_id_len-1] = '\0'; + } + + engine = ENGINE_by_id(engine_id); + if(!engine) { + PRINTF_ERR0("engine_by_id_nif Leaved: {error, bad_engine_id}"); + return enif_make_tuple2(env, atom_error, atom_bad_engine_id); + } + + ctx = enif_alloc_resource(engine_ctx_rtype, sizeof(struct engine_ctx)); + ctx->engine = engine; + ctx->id = engine_id; + + ret = enif_make_resource(env, ctx); + enif_release_resource(ctx); + + return enif_make_tuple2(env, atom_ok, ret); +#else + return atom_notsup; +#endif +} + +static ERL_NIF_TERM engine_init_nif(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[]) +{/* (Engine) */ +#ifdef HAS_ENGINE_SUPPORT + ERL_NIF_TERM ret = atom_ok; + struct engine_ctx *ctx; + + // Get Engine + if (!enif_get_resource(env, argv[0], engine_ctx_rtype, (void**)&ctx)) { + PRINTF_ERR0("engine_init_nif Leaved: Parameter not an engine resource object"); + return enif_make_badarg(env); + } + if (!ENGINE_init(ctx->engine)) { + //ERR_print_errors_fp(stderr); + PRINTF_ERR0("engine_init_nif Leaved: {error, engine_init_failed}"); + return enif_make_tuple2(env, atom_error, atom_engine_init_failed); + } + + return ret; +#else + return atom_notsup; +#endif +} + +static ERL_NIF_TERM engine_free_nif(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[]) +{/* (Engine) */ +#ifdef HAS_ENGINE_SUPPORT + struct engine_ctx *ctx; + + // Get Engine + if (!enif_get_resource(env, argv[0], engine_ctx_rtype, (void**)&ctx)) { + PRINTF_ERR0("engine_free_nif Leaved: Parameter not an engine resource object"); + return enif_make_badarg(env); + } + + ENGINE_free(ctx->engine); + return atom_ok; +#else + return atom_notsup; +#endif +} + +static ERL_NIF_TERM engine_finish_nif(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[]) +{/* (Engine) */ +#ifdef HAS_ENGINE_SUPPORT + struct engine_ctx *ctx; + + // Get Engine + if (!enif_get_resource(env, argv[0], engine_ctx_rtype, (void**)&ctx)) { + PRINTF_ERR0("engine_finish_nif Leaved: Parameter not an engine resource object"); + return enif_make_badarg(env); + } + + ENGINE_finish(ctx->engine); + return atom_ok; +#else + return atom_notsup; +#endif +} + +static ERL_NIF_TERM engine_load_dynamic_nif(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[]) +{/* () */ +#ifdef HAS_ENGINE_SUPPORT + ENGINE_load_dynamic(); + return atom_ok; +#else + return atom_notsup; +#endif +} + +static ERL_NIF_TERM engine_ctrl_cmd_strings_nif(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[]) +{/* (Engine, Commands) */ +#ifdef HAS_ENGINE_SUPPORT + ERL_NIF_TERM ret = atom_ok; + unsigned int cmds_len = 0; + char **cmds = NULL; + struct engine_ctx *ctx; + int i; + + // Get Engine + if (!enif_get_resource(env, argv[0], engine_ctx_rtype, (void**)&ctx)) { + PRINTF_ERR0("engine_ctrl_cmd_strings_nif Leaved: Parameter not an engine resource object"); + return enif_make_badarg(env); + } + + PRINTF_ERR1("Engine Id: %s\r\n", ENGINE_get_id(ctx->engine)); + + // Get Command List + if(!enif_get_list_length(env, argv[1], &cmds_len)) { + PRINTF_ERR0("engine_ctrl_cmd_strings_nif Leaved: Bad Command List"); + return enif_make_badarg(env); + } else { + cmds_len *= 2; // Key-Value list from erlang + cmds = enif_alloc((cmds_len+1)*sizeof(char*)); + if(get_engine_load_cmd_list(env, argv[1], cmds, 0)) { + PRINTF_ERR0("engine_ctrl_cmd_strings_nif Leaved: Couldn't read Command List"); + ret = enif_make_badarg(env); + goto error; + } + } + + for(i = 0; i < cmds_len; i+=2) { + PRINTF_ERR2("Cmd: %s:%s\r\n", + cmds[i] ? cmds[i] : "(NULL)", + cmds[i+1] ? cmds[i+1] : "(NULL)"); + if(!ENGINE_ctrl_cmd_string(ctx->engine, cmds[i], cmds[i+1], 0)) { + PRINTF_ERR2("Command failed: %s:%s\r\n", + cmds[i] ? cmds[i] : "(NULL)", + cmds[i+1] ? cmds[i+1] : "(NULL)"); + //ENGINE_free(ctx->engine); + ret = enif_make_tuple2(env, atom_error, atom_ctrl_cmd_failed); + PRINTF_ERR0("engine_ctrl_cmd_strings_nif Leaved: {error, ctrl_cmd_failed}"); + goto error; + } +} + + error: + for(i = 0; cmds != NULL && cmds[i] != NULL; i++) + enif_free(cmds[i]); + return ret; +#else + return atom_notsup; +#endif +} + +static ERL_NIF_TERM engine_add_nif(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[]) +{/* (Engine) */ +#ifdef HAS_ENGINE_SUPPORT + struct engine_ctx *ctx; + + // Get Engine + if (!enif_get_resource(env, argv[0], engine_ctx_rtype, (void**)&ctx)) { + PRINTF_ERR0("engine_add_nif Leaved: Parameter not an engine resource object"); + return enif_make_badarg(env); + } + + if (!ENGINE_add(ctx->engine)) { + PRINTF_ERR0("engine_add_nif Leaved: {error, add_engine_failed}"); + return enif_make_tuple2(env, atom_error, atom_add_engine_failed); + } + return atom_ok; +#else + return atom_notsup; +#endif +} + +static ERL_NIF_TERM engine_remove_nif(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[]) +{/* (Engine) */ +#ifdef HAS_ENGINE_SUPPORT + struct engine_ctx *ctx; + + // Get Engine + if (!enif_get_resource(env, argv[0], engine_ctx_rtype, (void**)&ctx)) { + PRINTF_ERR0("engine_remove_nif Leaved: Parameter not an engine resource object"); + return enif_make_badarg(env); + } + + if (!ENGINE_remove(ctx->engine)) { + PRINTF_ERR0("engine_remove_nif Leaved: {error, remove_engine_failed}"); + return enif_make_tuple2(env, atom_error, atom_remove_engine_failed); + } + return atom_ok; +#else + return atom_notsup; +#endif +} + +static ERL_NIF_TERM engine_register_nif(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[]) +{/* (Engine, EngineMethod) */ +#ifdef HAS_ENGINE_SUPPORT + struct engine_ctx *ctx; + unsigned int method; + + // Get Engine + if (!enif_get_resource(env, argv[0], engine_ctx_rtype, (void**)&ctx)) { + PRINTF_ERR0("engine_register_nif Leaved: Parameter not an engine resource object"); + return enif_make_badarg(env); + } + // Get Method + if (!enif_get_uint(env, argv[1], &method)) { + PRINTF_ERR0("engine_register_nif Leaved: Parameter Method not an uint"); + return enif_make_badarg(env); + } + + switch(method) + { +#ifdef ENGINE_METHOD_RSA + case ENGINE_METHOD_RSA: + if (!ENGINE_register_RSA(ctx->engine)) + return enif_make_tuple2(env, atom_error, atom_register_engine_failed); + break; +#endif +#ifdef ENGINE_METHOD_DSA + case ENGINE_METHOD_DSA: + if (!ENGINE_register_DSA(ctx->engine)) + return enif_make_tuple2(env, atom_error, atom_register_engine_failed); + break; +#endif +#ifdef ENGINE_METHOD_DH + case ENGINE_METHOD_DH: + if (!ENGINE_register_DH(ctx->engine)) + return enif_make_tuple2(env, atom_error, atom_register_engine_failed); + break; +#endif +#ifdef ENGINE_METHOD_RAND + case ENGINE_METHOD_RAND: + if (!ENGINE_register_RAND(ctx->engine)) + return enif_make_tuple2(env, atom_error, atom_register_engine_failed); + break; +#endif +#ifdef ENGINE_METHOD_ECDH + case ENGINE_METHOD_ECDH: + if (!ENGINE_register_ECDH(ctx->engine)) + return enif_make_tuple2(env, atom_error, atom_register_engine_failed); + break; +#endif +#ifdef ENGINE_METHOD_ECDSA + case ENGINE_METHOD_ECDSA: + if (!ENGINE_register_ECDSA(ctx->engine)) + return enif_make_tuple2(env, atom_error, atom_register_engine_failed); + break; +#endif +#ifdef ENGINE_METHOD_STORE + case ENGINE_METHOD_STORE: + if (!ENGINE_register_STORE(ctx->engine)) + return enif_make_tuple2(env, atom_error, atom_register_engine_failed); + break; +#endif +#ifdef ENGINE_METHOD_CIPHERS + case ENGINE_METHOD_CIPHERS: + if (!ENGINE_register_ciphers(ctx->engine)) + return enif_make_tuple2(env, atom_error, atom_register_engine_failed); + break; +#endif +#ifdef ENGINE_METHOD_DIGESTS + case ENGINE_METHOD_DIGESTS: + if (!ENGINE_register_digests(ctx->engine)) + return enif_make_tuple2(env, atom_error, atom_register_engine_failed); + break; +#endif +#ifdef ENGINE_METHOD_PKEY_METHS + case ENGINE_METHOD_PKEY_METHS: + if (!ENGINE_register_pkey_meths(ctx->engine)) + return enif_make_tuple2(env, atom_error, atom_register_engine_failed); + break; +#endif +#ifdef ENGINE_METHOD_PKEY_ASN1_METHS + case ENGINE_METHOD_PKEY_ASN1_METHS: + if (!ENGINE_register_pkey_asn1_meths(ctx->engine)) + return enif_make_tuple2(env, atom_error, atom_register_engine_failed); + break; +#endif +#ifdef ENGINE_METHOD_EC + case ENGINE_METHOD_EC: + if (!ENGINE_register_EC(ctx->engine)) + return enif_make_tuple2(env, atom_error, atom_register_engine_failed); + break; +#endif + default: + return enif_make_tuple2(env, atom_error, atom_engine_method_not_supported); + break; + } + return atom_ok; +#else + return atom_notsup; +#endif +} + +static ERL_NIF_TERM engine_unregister_nif(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[]) +{/* (Engine, EngineMethod) */ +#ifdef HAS_ENGINE_SUPPORT + struct engine_ctx *ctx; + unsigned int method; + + // Get Engine + if (!enif_get_resource(env, argv[0], engine_ctx_rtype, (void**)&ctx)) { + PRINTF_ERR0("engine_unregister_nif Leaved: Parameter not an engine resource object"); + return enif_make_badarg(env); + } + // Get Method + if (!enif_get_uint(env, argv[1], &method)) { + PRINTF_ERR0("engine_unregister_nif Leaved: Parameter Method not an uint"); + return enif_make_badarg(env); + } + + switch(method) + { +#ifdef ENGINE_METHOD_RSA + case ENGINE_METHOD_RSA: + ENGINE_unregister_RSA(ctx->engine); + break; +#endif +#ifdef ENGINE_METHOD_DSA + case ENGINE_METHOD_DSA: + ENGINE_unregister_DSA(ctx->engine); + break; +#endif +#ifdef ENGINE_METHOD_DH + case ENGINE_METHOD_DH: + ENGINE_unregister_DH(ctx->engine); + break; +#endif +#ifdef ENGINE_METHOD_RAND + case ENGINE_METHOD_RAND: + ENGINE_unregister_RAND(ctx->engine); + break; +#endif +#ifdef ENGINE_METHOD_ECDH + case ENGINE_METHOD_ECDH: + ENGINE_unregister_ECDH(ctx->engine); + break; +#endif +#ifdef ENGINE_METHOD_ECDSA + case ENGINE_METHOD_ECDSA: + ENGINE_unregister_ECDSA(ctx->engine); + break; +#endif +#ifdef ENGINE_METHOD_STORE + case ENGINE_METHOD_STORE: + ENGINE_unregister_STORE(ctx->engine); + break; +#endif +#ifdef ENGINE_METHOD_CIPHERS + case ENGINE_METHOD_CIPHERS: + ENGINE_unregister_ciphers(ctx->engine); + break; +#endif +#ifdef ENGINE_METHOD_DIGESTS + case ENGINE_METHOD_DIGESTS: + ENGINE_unregister_digests(ctx->engine); + break; +#endif +#ifdef ENGINE_METHOD_PKEY_METHS + case ENGINE_METHOD_PKEY_METHS: + ENGINE_unregister_pkey_meths(ctx->engine); + break; +#endif +#ifdef ENGINE_METHOD_PKEY_ASN1_METHS + case ENGINE_METHOD_PKEY_ASN1_METHS: + ENGINE_unregister_pkey_asn1_meths(ctx->engine); + break; +#endif +#ifdef ENGINE_METHOD_EC + case ENGINE_METHOD_EC: + ENGINE_unregister_EC(ctx->engine); + break; +#endif + default: + break; + } + return atom_ok; +#else + return atom_notsup; +#endif +} + +static ERL_NIF_TERM engine_get_first_nif(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[]) +{/* (Engine) */ +#ifdef HAS_ENGINE_SUPPORT + ERL_NIF_TERM ret; + ENGINE *engine; + ErlNifBinary engine_bin; + struct engine_ctx *ctx; + + engine = ENGINE_get_first(); + if(!engine) { + enif_alloc_binary(0, &engine_bin); + engine_bin.size = 0; + return enif_make_tuple2(env, atom_ok, enif_make_binary(env, &engine_bin)); + } + + ctx = enif_alloc_resource(engine_ctx_rtype, sizeof(struct engine_ctx)); + ctx->engine = engine; + ctx->id = NULL; + + ret = enif_make_resource(env, ctx); + enif_release_resource(ctx); + + return enif_make_tuple2(env, atom_ok, ret); +#else + return atom_notsup; +#endif +} + +static ERL_NIF_TERM engine_get_next_nif(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[]) +{/* (Engine) */ +#ifdef HAS_ENGINE_SUPPORT + ERL_NIF_TERM ret; + ENGINE *engine; + ErlNifBinary engine_bin; + struct engine_ctx *ctx, *next_ctx; + + // Get Engine + if (!enif_get_resource(env, argv[0], engine_ctx_rtype, (void**)&ctx)) { + PRINTF_ERR0("engine_get_next_nif Leaved: Parameter not an engine resource object"); + return enif_make_badarg(env); + } + engine = ENGINE_get_next(ctx->engine); + if (!engine) { + enif_alloc_binary(0, &engine_bin); + engine_bin.size = 0; + return enif_make_tuple2(env, atom_ok, enif_make_binary(env, &engine_bin)); + } + + next_ctx = enif_alloc_resource(engine_ctx_rtype, sizeof(struct engine_ctx)); + next_ctx->engine = engine; + next_ctx->id = NULL; + + ret = enif_make_resource(env, next_ctx); + enif_release_resource(next_ctx); + + return enif_make_tuple2(env, atom_ok, ret); +#else + return atom_notsup; +#endif +} + +static ERL_NIF_TERM engine_get_id_nif(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[]) +{/* (Engine) */ +#ifdef HAS_ENGINE_SUPPORT + ErlNifBinary engine_id_bin; + const char *engine_id; + int size; + struct engine_ctx *ctx; + + // Get Engine + if (!enif_get_resource(env, argv[0], engine_ctx_rtype, (void**)&ctx)) { + PRINTF_ERR0("engine_get_id_nif Leaved: Parameter not an engine resource object"); + return enif_make_badarg(env); + } + + engine_id = ENGINE_get_id(ctx->engine); + if (!engine_id) { + enif_alloc_binary(0, &engine_id_bin); + engine_id_bin.size = 0; + return enif_make_tuple2(env, atom_ok, enif_make_binary(env, &engine_id_bin)); + } + + size = strlen(engine_id); + enif_alloc_binary(size, &engine_id_bin); + engine_id_bin.size = size; + memcpy(engine_id_bin.data, engine_id, size); + + return enif_make_tuple2(env, atom_ok, enif_make_binary(env, &engine_id_bin)); +#else + return atom_notsup; +#endif +} + +static int get_engine_load_cmd_list(ErlNifEnv* env, const ERL_NIF_TERM term, char **cmds, int i) +{ +#ifdef HAS_ENGINE_SUPPORT + ERL_NIF_TERM head, tail; + const ERL_NIF_TERM *tmp_tuple; + ErlNifBinary tmpbin; + int arity; + char* tmpstr; + int tmplen = 0; + + if(!enif_is_empty_list(env, term)) { + if(!enif_get_list_cell(env, term, &head, &tail)) { + cmds[i] = NULL; + return -1; + } else { + if(!enif_get_tuple(env, head, &arity, &tmp_tuple) || arity != 2) { + cmds[i] = NULL; + return -1; + } else { + if(!enif_inspect_binary(env, tmp_tuple[0], &tmpbin)) { + cmds[i] = NULL; + return -1; + } else { + tmplen = tmpbin.size+1; + tmpstr = enif_alloc(tmplen); + (void) memcpy(tmpstr, tmpbin.data, tmplen); + tmpstr[tmplen-1] = '\0'; + cmds[i++] = tmpstr; + } + if(!enif_inspect_binary(env, tmp_tuple[1], &tmpbin)) { + cmds[i] = NULL; + return -1; + } else { + if(tmpbin.size == 0) + cmds[i++] = NULL; + else { + tmplen = tmpbin.size+1; + tmpstr = enif_alloc(tmplen); + (void) memcpy(tmpstr, tmpbin.data, tmplen); + tmpstr[tmplen-1] = '\0'; + cmds[i++] = tmpstr; + } + } + return get_engine_load_cmd_list(env, tail, cmds, i); + } + } + } else { + cmds[i] = NULL; + return 0; + } +#else + return atom_notsup; +#endif +} + +static ERL_NIF_TERM engine_get_all_methods_nif(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[]) +{/* () */ +#ifdef HAS_ENGINE_SUPPORT + ERL_NIF_TERM method_array[12]; + int i = 0; + +#ifdef ENGINE_METHOD_RSA + method_array[i++] = atom_engine_method_rsa; +#endif +#ifdef ENGINE_METHOD_DSA + method_array[i++] = atom_engine_method_dsa; +#endif +#ifdef ENGINE_METHOD_DH + method_array[i++] = atom_engine_method_dh; +#endif +#ifdef ENGINE_METHOD_RAND + method_array[i++] = atom_engine_method_rand; +#endif +#ifdef ENGINE_METHOD_ECDH + method_array[i++] = atom_engine_method_ecdh; +#endif +#ifdef ENGINE_METHOD_ECDSA + method_array[i++] = atom_engine_method_ecdsa; +#endif +#ifdef ENGINE_METHOD_STORE + method_array[i++] = atom_engine_method_store; +#endif +#ifdef ENGINE_METHOD_CIPHERS + method_array[i++] = atom_engine_method_ciphers; +#endif +#ifdef ENGINE_METHOD_DIGESTS + method_array[i++] = atom_engine_method_digests; +#endif +#ifdef ENGINE_METHOD_PKEY_METHS + method_array[i++] = atom_engine_method_pkey_meths; +#endif +#ifdef ENGINE_METHOD_PKEY_ASN1_METHS + method_array[i++] = atom_engine_method_pkey_asn1_meths; +#endif +#ifdef ENGINE_METHOD_EC + method_array[i++] = atom_engine_method_ec; +#endif + + return enif_make_list_from_array(env, method_array, i); +#else + return atom_notsup; +#endif +} diff --git a/lib/crypto/c_src/otp_test_engine.c b/lib/crypto/c_src/otp_test_engine.c new file mode 100644 index 0000000000..a66bee2ddf --- /dev/null +++ b/lib/crypto/c_src/otp_test_engine.c @@ -0,0 +1,262 @@ +/* + * %CopyrightBegin% + * + * Copyright Ericsson AB 2017-2017. All Rights Reserved. + * + * Licensed under the Apache License, Version 2.0 (the "License"); + * you may not use this file except in compliance with the License. + * You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + * + * %CopyrightEnd% + */ + +#ifdef _WIN32 +#define OPENSSL_OPT_WINDLL +#endif +#include <stdio.h> +#include <string.h> + +#include <openssl/engine.h> +#include <openssl/md5.h> +#include <openssl/rsa.h> +#include <openssl/pem.h> + +#define PACKED_OPENSSL_VERSION(MAJ, MIN, FIX, P) \ + ((((((((MAJ << 8) | MIN) << 8 ) | FIX) << 8) | (P-'a'+1)) << 4) | 0xf) + +#define PACKED_OPENSSL_VERSION_PLAIN(MAJ, MIN, FIX) \ + PACKED_OPENSSL_VERSION(MAJ,MIN,FIX,('a'-1)) + +#if OPENSSL_VERSION_NUMBER < PACKED_OPENSSL_VERSION_PLAIN(1,1,0) \ + || defined(LIBRESSL_VERSION_NUMBER) +#define OLD +#endif + +static const char *test_engine_id = "MD5"; +static const char *test_engine_name = "MD5 test engine"; + +/* The callback that does the job of fetching keys on demand by the Engine */ +EVP_PKEY* test_key_load(ENGINE *er, const char *id, UI_METHOD *ui_method, void *callback_data); + + +static int test_init(ENGINE *e) { + printf("OTP Test Engine Initializatzion!\r\n"); + + /* Load all digest and cipher algorithms. Needed for password protected private keys */ + OpenSSL_add_all_algorithms(); + + return 111; +} + +static void add_test_data(unsigned char *md, unsigned int len) +{ + unsigned int i; + + for (i=0; i<len; i++) { + md[i] = (unsigned char)(i & 0xff); + } +} + +/* MD5 part */ +#undef data +#ifdef OLD +#define data(ctx) ((MD5_CTX *)ctx->md_data) +#endif + +static int test_engine_md5_init(EVP_MD_CTX *ctx) { + fprintf(stderr, "MD5 initialized\r\n"); +#ifdef OLD + return MD5_Init(data(ctx)); +#else + return 1; +#endif +} + +static int test_engine_md5_update(EVP_MD_CTX *ctx,const void *data, size_t count) +{ + fprintf(stderr, "MD5 update\r\n"); +#ifdef OLD + return MD5_Update(data(ctx), data, (size_t)count); +#else + return 1; +#endif +} + +static int test_engine_md5_final(EVP_MD_CTX *ctx,unsigned char *md) { +#ifdef OLD + int ret; + + fprintf(stderr, "MD5 final size of EVP_MD: %lu\r\n", sizeof(EVP_MD)); + ret = MD5_Final(md, data(ctx)); + + if (ret > 0) { + add_test_data(md, MD5_DIGEST_LENGTH); + } + return ret; +#else + fprintf(stderr, "MD5 final\r\n"); + add_test_data(md, MD5_DIGEST_LENGTH); + return 1; +#endif +} + +#ifdef OLD +static EVP_MD test_engine_md5_method= { + NID_md5, /* The name ID for MD5 */ + NID_undef, /* IGNORED: MD5 with private key encryption NID */ + MD5_DIGEST_LENGTH, /* Size of MD5 result, in bytes */ + 0, /* Flags */ + test_engine_md5_init, /* digest init */ + test_engine_md5_update, /* digest update */ + test_engine_md5_final, /* digest final */ + NULL, /* digest copy */ + NULL, /* digest cleanup */ + EVP_PKEY_NULL_method, /* IGNORED: pkey methods */ + MD5_CBLOCK, /* Internal blocksize, see rfc1321/md5.h */ + sizeof(EVP_MD *) + sizeof(MD5_CTX), + NULL, /* IGNORED: control function */ +}; +#endif + +static int test_digest_ids[] = {NID_md5}; + +static int test_engine_digest_selector(ENGINE *e, const EVP_MD **digest, + const int **nids, int nid) { + int ok = 1; + if (!digest) { + *nids = test_digest_ids; + fprintf(stderr, "Digest is empty! Nid:%d\r\n", nid); + return 2; + } + fprintf(stderr, "Digest no %d requested\r\n",nid); + if (nid == NID_md5) { +#ifdef OLD + *digest = &test_engine_md5_method; +#else + EVP_MD *md = EVP_MD_meth_new(NID_md5, NID_undef); + if (!md || + !EVP_MD_meth_set_result_size(md, MD5_DIGEST_LENGTH) || + !EVP_MD_meth_set_flags(md, 0) || + !EVP_MD_meth_set_init(md, test_engine_md5_init) || + !EVP_MD_meth_set_update(md, test_engine_md5_update) || + !EVP_MD_meth_set_final(md, test_engine_md5_final) || + !EVP_MD_meth_set_copy(md, NULL) || + !EVP_MD_meth_set_cleanup(md, NULL) || + !EVP_MD_meth_set_input_blocksize(md, MD5_CBLOCK) || + !EVP_MD_meth_set_app_datasize(md, sizeof(EVP_MD *) + sizeof(MD5_CTX)) || + !EVP_MD_meth_set_ctrl(md, NULL)) + { + ok = 0; + *digest = NULL; + } else + { + *digest = md; + } +#endif + } + else { + ok = 0; + *digest = NULL; + } + + return ok; +} + + +static int bind_helper(ENGINE * e, const char *id) +{ + if (!ENGINE_set_id(e, test_engine_id) || + !ENGINE_set_name(e, test_engine_name) || + !ENGINE_set_init_function(e, test_init) || + !ENGINE_set_digests(e, &test_engine_digest_selector) || + /* For testing of key storage in an Engine: */ + !ENGINE_set_load_privkey_function(e, &test_key_load) || + !ENGINE_set_load_pubkey_function(e, &test_key_load) + ) + return 0; + + return 1; +} + +IMPLEMENT_DYNAMIC_CHECK_FN(); + +IMPLEMENT_DYNAMIC_BIND_FN(bind_helper); + +/******************************************************** + * + * Engine storage simulation + * + */ +int pem_passwd_cb_fun(char *buf, int size, int rwflag, void *password); + +EVP_PKEY* test_key_load(ENGINE *er, const char *id, UI_METHOD *ui_method, void *callback_data) +{ + EVP_PKEY *pkey = NULL; + FILE *f = fopen(id, "r"); + + if (!f) { + fprintf(stderr, "%s:%d fopen(%s) failed\r\n", __FILE__,__LINE__,id); + return NULL; + } + + /* First try to read as a private key. If that fails, try to read as a public key: */ + pkey = PEM_read_PrivateKey(f, NULL, pem_passwd_cb_fun, callback_data); + if (!pkey) { + /* ERR_print_errors_fp (stderr); */ + fclose(f); + f = fopen(id, "r"); + pkey = PEM_read_PUBKEY(f, NULL, NULL, NULL); + } + fclose(f); + + if (!pkey) { + fprintf(stderr, "%s:%d Key read from file failed. ", __FILE__,__LINE__); + if (callback_data) + fprintf(stderr, "Pwd = \"%s\". ", (char *)callback_data); + fprintf(stderr, "Contents of file \"%s\":\r\n",id); + f = fopen(id, "r"); + { /* Print the contents of the key file */ + char c; + while (!feof(f)) { + switch (c=fgetc(f)) { + case '\n': + case '\r': putc('\r',stdout); putc('\n',stdout); break; + default: putc(c, stdout); + } + } + } + fclose(f); + } + + return pkey; +} + + +int pem_passwd_cb_fun(char *buf, int size, int rwflag, void *password) +{ + int i; + + fprintf(stderr, "In pem_passwd_cb_fun\r\n"); + if (!password) + return 0; + + i = strlen(password); + if (i < size) { + /* whole pwd (incl terminating 0) fits */ + fprintf(stderr, "Got FULL pwd %d(%d) chars\r\n", i, size); + memcpy(buf, (char*)password, i+1); + return i+1; + } else { + fprintf(stderr, "Got TO LONG pwd %d(%d) chars\r\n", i, size); + /* meaningless with a truncated password */ + return 0; + } +} diff --git a/lib/crypto/doc/src/Makefile b/lib/crypto/doc/src/Makefile index 9c503b8fe0..a902779383 100644 --- a/lib/crypto/doc/src/Makefile +++ b/lib/crypto/doc/src/Makefile @@ -9,11 +9,11 @@ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. -# +# # The Initial Developer of the Original Code is Ericsson Utvecklings AB. # Portions created by Ericsson are Copyright 1999, Ericsson Utvecklings # AB. All Rights Reserved.'' -# +# # $Id$ # include $(ERL_TOP)/make/target.mk @@ -38,13 +38,13 @@ XML_APPLICATION_FILES = ref_man.xml XML_REF3_FILES = crypto.xml XML_REF6_FILES = crypto_app.xml -XML_PART_FILES = release_notes.xml usersguide.xml -XML_CHAPTER_FILES = notes.xml licenses.xml fips.xml +XML_PART_FILES = usersguide.xml +XML_CHAPTER_FILES = notes.xml licenses.xml fips.xml engine_load.xml BOOK_FILES = book.xml XML_FILES = $(BOOK_FILES) $(XML_APPLICATION_FILES) $(XML_REF3_FILES) $(XML_REF6_FILES) \ - $(XML_PART_FILES) $(XML_CHAPTER_FILES) + $(XML_PART_FILES) $(XML_CHAPTER_FILES) GIF_FILES = @@ -63,9 +63,9 @@ HTML_REF_MAN_FILE = $(HTMLDIR)/index.html TOP_PDF_FILE = $(PDFDIR)/$(APPLICATION)-$(VSN).pdf # ---------------------------------------------------- -# FLAGS +# FLAGS # ---------------------------------------------------- -XML_FLAGS += +XML_FLAGS += # ---------------------------------------------------- # Targets @@ -73,7 +73,6 @@ XML_FLAGS += $(HTMLDIR)/%.gif: %.gif $(INSTALL_DATA) $< $@ - docs: pdf html man $(TOP_PDF_FILE): $(XML_FILES) @@ -86,7 +85,7 @@ man: $(MAN3_FILES) $(MAN6_FILES) gifs: $(GIF_FILES:%=$(HTMLDIR)/%) -debug opt valgrind: +debug opt valgrind: clean clean_docs clean_tex: rm -rf $(HTMLDIR)/* @@ -97,7 +96,7 @@ clean clean_docs clean_tex: # ---------------------------------------------------- # Release Target -# ---------------------------------------------------- +# ---------------------------------------------------- include $(ERL_TOP)/make/otp_release_targets.mk release_docs_spec: docs @@ -114,4 +113,3 @@ release_docs_spec: docs release_spec: - diff --git a/lib/crypto/doc/src/crypto.xml b/lib/crypto/doc/src/crypto.xml index 8c30d3d50c..3ab8f9c832 100644 --- a/lib/crypto/doc/src/crypto.xml +++ b/lib/crypto/doc/src/crypto.xml @@ -11,7 +11,7 @@ Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at - + http://www.apache.org/licenses/LICENSE-2.0 Unless required by applicable law or agreed to in writing, software @@ -19,7 +19,6 @@ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License. - </legalnotice> <title>crypto</title> @@ -68,11 +67,11 @@ <section> <title>DATA TYPES </title> - - <code>key_value() = integer() | binary() </code> + + <code>key_value() = integer() | binary() </code> <p>Always <c>binary()</c> when used as return value</p> - <code>rsa_public() = [key_value()] = [E, N] </code> + <code>rsa_public() = [key_value()] = [E, N] </code> <p> Where E is the public exponent and N is public modulus. </p> <code>rsa_private() = [key_value()] = [E, N, D] | [E, N, D, P1, P2, E1, E2, C] </code> @@ -85,7 +84,7 @@ <code>dss_public() = [key_value()] = [P, Q, G, Y] </code> <p>Where P, Q and G are the dss parameters and Y is the public key.</p> - <code>dss_private() = [key_value()] = [P, Q, G, X] </code> + <code>dss_private() = [key_value()] = [P, Q, G, X] </code> <p>Where P, Q and G are the dss parameters and X is the private key.</p> <code>srp_public() = key_value() </code> @@ -109,15 +108,16 @@ <code>ecdh_private() = key_value() </code> - <code>ecdh_params() = ec_named_curve() | ec_explicit_curve()</code> + <code>ecdh_params() = ec_named_curve() | ec_explicit_curve()</code> <code>ec_explicit_curve() = - {ec_field(), Prime :: key_value(), Point :: key_value(), Order :: integer(), CoFactor :: none | integer()} </code> + {ec_field(), Prime :: key_value(), Point :: key_value(), Order :: integer(), + CoFactor :: none | integer()} </code> <code>ec_field() = {prime_field, Prime :: integer()} | {characteristic_two_field, M :: integer(), Basis :: ec_basis()}</code> - <code>ec_basis() = {tpbasis, K :: non_neg_integer()} | + <code>ec_basis() = {tpbasis, K :: non_neg_integer()} | {ppbasis, K1 :: non_neg_integer(), K2 :: non_neg_integer(), K3 :: non_neg_integer()} | onbasis</code> @@ -136,16 +136,33 @@ See also <seealso marker="#supports-0">crypto:supports/0</seealso> </p> + <code>engine_key_ref() = #{engine := engine_ref(), + key_id := key_id(), + password => password()}</code> + + <code>engine_key_ref() = term()</code> + <p>The result of a call to <seealso marker="#engine_load-3">engine_load/3</seealso>. + </p> + + <code>key_id() = string() | binary()</code> + <p>Identifies the key to be used. The format depends on the loaded engine. It is passed to + the <c>ENGINE_load_(private|public)_key</c> functions in libcrypto. + </p> + + <code>password() = string() | binary()</code> + <p>The key's password + </p> + <code>stream_cipher() = rc4 | aes_ctr </code> - <code>block_cipher() = aes_cbc | aes_cfb8 | aes_cfb128 | aes_ige256 | blowfish_cbc | + <code>block_cipher() = aes_cbc | aes_cfb8 | aes_cfb128 | aes_ige256 | blowfish_cbc | blowfish_cfb64 | des_cbc | des_cfb | des3_cbc | des3_cfb | des_ede3 | rc2_cbc </code> - <code>aead_cipher() = aes_gcm | chacha20_poly1305 </code> + <code>aead_cipher() = aes_gcm | chacha20_poly1305 </code> - <code>stream_key() = aes_key() | rc4_key() </code> + <code>stream_key() = aes_key() | rc4_key() </code> - <code>block_key() = aes_key() | blowfish_key() | des_key()| des3_key() </code> + <code>block_key() = aes_key() | blowfish_key() | des_key()| des3_key() </code> <code>aes_key() = iodata() </code> <p>Key length is 128, 192 or 256 bits</p> @@ -174,13 +191,17 @@ Note that both md4 and md5 are recommended only for compatibility with existing applications. </p> <code> cipher_algorithms() = aes_cbc | aes_cfb8 | aes_cfb128 | aes_ctr | aes_gcm | - aes_ige256 | blowfish_cbc | blowfish_cfb64 | chacha20_poly1305 | des_cbc | des_cfb | - des3_cbc | des3_cfb | des_ede3 | rc2_cbc | rc4 </code> - <code> mac_algorithms() = hmac | cmac</code> - <code> public_key_algorithms() = rsa |dss | ecdsa | dh | ecdh | ec_gf2m</code> + aes_ige256 | blowfish_cbc | blowfish_cfb64 | chacha20_poly1305 | des_cbc | + des_cfb | des3_cbc | des3_cfb | des_ede3 | rc2_cbc | rc4 </code> + <code> mac_algorithms() = hmac | cmac</code> + <code> public_key_algorithms() = rsa |dss | ecdsa | dh | ecdh | ec_gf2m</code> <p>Note that ec_gf2m is not strictly a public key algorithm, but a restriction on what curves are supported with ecdsa and ecdh. </p> + <code>engine_method_type() = engine_method_rsa | engine_method_dsa | engine_method_dh | + engine_method_rand | engine_method_ecdh | engine_method_ecdsa | + engine_method_ciphers | engine_method_digests | engine_method_store | + engine_method_pkey_meths | engine_method_pkey_asn1_meths</code> </section> @@ -261,13 +282,13 @@ is not supported by the underlying OpenSSL implementation.</p> </desc> </func> - + <func> <name>bytes_to_integer(Bin) -> Integer </name> <fsummary>Convert binary representation, of an integer, to an Erlang integer.</fsummary> <type> <v>Bin = binary() - as returned by crypto functions</v> - + <v>Integer = integer() </v> </type> <desc> @@ -439,7 +460,7 @@ </type> <desc> <p>Updates the HMAC represented by <c>Context</c> using the given <c>Data</c>. <c>Context</c> - must have been generated using an HMAC init function (such as + must have been generated using an HMAC init function (such as <seealso marker="#hmac_init-2">hmac_init</seealso>). <c>Data</c> can be any length. <c>NewContext</c> must be passed into the next call to <c>hmac_update</c> or to one of the functions <seealso marker="#hmac_final-1">hmac_final</seealso> and @@ -580,7 +601,7 @@ <type> <v>Type = rsa</v> <v>CipherText = binary()</v> - <v>PrivateKey = rsa_private()</v> + <v>PrivateKey = rsa_private() | engine_key_ref()</v> <v>Padding = rsa_pkcs1_padding | rsa_pkcs1_oaep_padding | rsa_no_padding</v> <v>PlainText = binary()</v> </type> @@ -594,7 +615,22 @@ </p> </desc> </func> - + + <func> + <name>privkey_to_pubkey(Type, EnginePrivateKeyRef) -> PublicKey</name> + <fsummary>Fetches a public key from an Engine stored private key.</fsummary> + <type> + <v>Type = rsa | dss</v> + <v>EnginePrivateKeyRef = engine_key_ref()</v> + <v>PublicKey = rsa_public() | dss_public()</v> + </type> + <desc> + <p>Fetches the corresponding public key from a private key stored in an Engine. + The key must be of the type indicated by the Type parameter. + </p> + </desc> + </func> + <func> <name>private_encrypt(Type, PlainText, PrivateKey, Padding) -> CipherText</name> <fsummary>Encrypts PlainText using the private Key.</fsummary> @@ -605,7 +641,7 @@ than <c>byte_size(N)-11</c> if <c>rsa_pkcs1_padding</c> is used, and <c>byte_size(N)</c> if <c>rsa_no_padding</c> is used, where N is public modulus of the RSA key.</d> - <v>PrivateKey = rsa_private()</v> + <v>PrivateKey = rsa_private() | engine_key_ref()</v> <v>Padding = rsa_pkcs1_padding | rsa_no_padding</v> <v>CipherText = binary()</v> </type> @@ -624,7 +660,7 @@ <type> <v>Type = rsa</v> <v>CipherText = binary()</v> - <v>PublicKey = rsa_public() </v> + <v>PublicKey = rsa_public() | engine_key_ref()</v> <v>Padding = rsa_pkcs1_padding | rsa_no_padding</v> <v>PlainText = binary()</v> </type> @@ -649,7 +685,7 @@ than <c>byte_size(N)-11</c> if <c>rsa_pkcs1_padding</c> is used, and <c>byte_size(N)</c> if <c>rsa_no_padding</c> is used, where N is public modulus of the RSA key.</d> - <v>PublicKey = rsa_public()</v> + <v>PublicKey = rsa_public() | engine_key_ref()</v> <v>Padding = rsa_pkcs1_padding | rsa_pkcs1_oaep_padding | rsa_no_padding</v> <v>CipherText = binary()</v> </type> @@ -702,7 +738,7 @@ signed or it is the hashed value of "cleartext" i.e. the digest (plaintext).</d> <v>DigestType = rsa_digest_type() | dss_digest_type() | ecdsa_digest_type()</v> - <v>Key = rsa_private() | dss_private() | [ecdh_private(),ecdh_params()]</v> + <v>Key = rsa_private() | dss_private() | [ecdh_private(),ecdh_params()] | engine_key_ref()</v> <v>Options = sign_options()</v> </type> <desc> @@ -1014,7 +1050,7 @@ _FloatValue = rand:uniform(). % [0.0; 1.0[</pre> or it is the hashed value of "cleartext" i.e. the digest (plaintext).</d> <v>DigestType = rsa_digest_type() | dss_digest_type() | ecdsa_digest_type()</v> <v>Signature = binary()</v> - <v>Key = rsa_public() | dss_public() | [ecdh_public(),ecdh_params()]</v> + <v>Key = rsa_public() | dss_public() | [ecdh_public(),ecdh_params()] | engine_key_ref()</v> <v>Options = sign_options()</v> </type> <desc> @@ -1026,6 +1062,124 @@ _FloatValue = rand:uniform(). % [0.0; 1.0[</pre> </desc> </func> + <!-- Engine functions --> + <func> + <name>engine_get_all_methods() -> Result</name> + <fsummary>Return list of all possible engine methods</fsummary> + <type> + <v>Result = [EngineMethod::atom()]</v> + </type> + <desc> + <p> + Returns a list of all possible engine methods. + </p> + <p> + May throw exception notsup in case there is + no engine support in the underlying OpenSSL implementation. + </p> + <p> + See also the chapter <seealso marker="crypto:engine_load#engine_load">Engine Load</seealso> + in the User's Guide. + </p> + </desc> + </func> + + <func> + <name>engine_load(EngineId, PreCmds, PostCmds) -> Result</name> + <fsummary>Dynamical load an encryption engine</fsummary> + <type> + <v>EngineId = unicode:chardata()</v> + <v>PreCmds, PostCmds = [{unicode:chardata(), unicode:chardata()}]</v> + <v>Result = {ok, Engine::term()} | {error, Reason::term()}</v> + </type> + <desc> + <p> + Loads the OpenSSL engine given by <c>EngineId</c> if it is available and then returns ok and + an engine handle. This function is the same as calling <c>engine_load/4</c> with + <c>EngineMethods</c> set to a list of all the possible methods. An error tuple is + returned if the engine can't be loaded. + </p> + <p> + The function throws a badarg if the parameters are in wrong format. + It may also throw the exception notsup in case there is + no engine support in the underlying OpenSSL implementation. + </p> + <p> + See also the chapter <seealso marker="crypto:engine_load#engine_load">Engine Load</seealso> + in the User's Guide. + </p> + </desc> + </func> + + <func> + <name>engine_load(EngineId, PreCmds, PostCmds, EngineMethods) -> Result</name> + <fsummary>Dynamical load an encryption engine</fsummary> + <type> + <v>EngineId = unicode:chardata()</v> + <v>PreCmds, PostCmds = [{unicode:chardata(), unicode:chardata()}]</v> + <v>EngineMethods = [engine_method_type()]</v> + <v>Result = {ok, Engine::term()} | {error, Reason::term()}</v> + </type> + <desc> + <p> + Loads the OpenSSL engine given by <c>EngineId</c> if it is available and then returns ok and + an engine handle. An error tuple is returned if the engine can't be loaded. + </p> + <p> + The function throws a badarg if the parameters are in wrong format. + It may also throw the exception notsup in case there is + no engine support in the underlying OpenSSL implementation. + </p> + <p> + See also the chapter <seealso marker="crypto:engine_load#engine_load">Engine Load</seealso> + in the User's Guide. + </p> + </desc> + </func> + + <func> + <name>engine_unload(Engine) -> Result</name> + <fsummary>Dynamical load an encryption engine</fsummary> + <type> + <v>Engine = term()</v> + <v>Result = ok | {error, Reason::term()}</v> + </type> + <desc> + <p> + Unloads the OpenSSL engine given by <c>EngineId</c>. + An error tuple is returned if the engine can't be unloaded. + </p> + <p> + The function throws a badarg if the parameter is in wrong format. + It may also throw the exception notsup in case there is + no engine support in the underlying OpenSSL implementation. + </p> + <p> + See also the chapter <seealso marker="crypto:engine_load#engine_load">Engine Load</seealso> + in the User's Guide. + </p> + </desc> + </func> + + <func> + <name>engine_list() -> Result</name> + <fsummary>List the known engine ids</fsummary> + <type> + <v>Result = [EngineId::unicode:chardata()]</v> + </type> + <desc> + <p>List the id's of all engines in OpenSSL's internal list.</p> + <p> + It may also throw the exception notsup in case there is + no engine support in the underlying OpenSSL implementation. + </p> + <p> + See also the chapter <seealso marker="crypto:engine_load#engine_load">Engine Load</seealso> + in the User's Guide. + </p> + </desc> + </func> + </funcs> <!-- Maybe put this in the users guide --> @@ -1100,4 +1254,3 @@ _FloatValue = rand:uniform(). % [0.0; 1.0[</pre> <!-- </p> --> <!-- </section> --> </erlref> - diff --git a/lib/crypto/doc/src/engine_load.xml b/lib/crypto/doc/src/engine_load.xml new file mode 100644 index 0000000000..e5c3f5d561 --- /dev/null +++ b/lib/crypto/doc/src/engine_load.xml @@ -0,0 +1,110 @@ +<?xml version="1.0" encoding="utf-8" ?> +<!DOCTYPE chapter SYSTEM "chapter.dtd"> + +<chapter> + <header> + <copyright> + <year>2017</year><year>2017</year> + <holder>Ericsson AB. All Rights Reserved.</holder> + </copyright> + <legalnotice> + The contents of this file are subject to the Erlang Public License, + Version 1.1, (the "License"); you may not use this file except in + compliance with the License. You should have received a copy of the + Erlang Public License along with this software. If not, it can be + retrieved online at http://www.erlang.org/. + + Software distributed under the License is distributed on an "AS IS" + basis, WITHOUT WARRANTY OF ANY KIND, either express or implied. See + the License for the specific language governing rights and limitations + under the License. + </legalnotice> + <title>Engine Load</title> + <prepared>Lars Thorsén</prepared> + <date>2017-08-22</date> + <file>engine_load.xml</file> + </header> + <p> + <marker id="engine_load"></marker> + This chapter describes the support for loading encryption engines in the crypto application. + </p> + + <section> + <title>Background</title> + <p> + OpenSSL exposes an Engine API, which makes it possible to plug in alternative + implementations for some or all of the cryptographic operations implemented by OpenSSL. + When configured appropriately, OpenSSL calls the engine's implementation of these + operations instead of its own. + </p> + <p> + Typically, OpenSSL engines provide a hardware implementation of specific cryptographic + operations. The hardware implementation usually offers improved performance over its + software-based counterpart, which is known as cryptographic acceleration. + </p> + </section> + + <section> + <title>Use Cases</title> + <section> + <title>Dynamically load an engine from default directory</title> + <p> + If the engine is located in the OpenSSL/LibreSSL installation <c>engines</c> directory. + </p> + <code> +1> {ok, Engine} = crypto:engine_load(<<"otp_test_engine">>, [], []). + {ok, #Ref}</code> + <note> + <p>The file name requirement on the engine dynamic library can differ between SSL versions.</p> + </note> + </section> + + <section> + <title>Load an engine with the dynamic engine</title> + <p> + Load an engine with the help of the dynamic engine by giving the path to the library. + </p> + <code> + 2> {ok, Engine} = crypto:engine_load(<<"dynamic">>, + [{<<"SO_PATH">>, + <<"/some/path/otp_test_engine.so">>}, + {<<"ID">>, <<"MD5">>}, + <<"LOAD">>], + []). + {ok, #Ref}</code> + <note> + <p>The dynamic engine is not supported in LibreSSL from version 2.2.1</p> + </note> + </section> + + <section> + <title>Load an engine and replace some methods</title> + <p> + Load an engine with the help of the dynamic engine and just + replace some engine methods. + </p> + <code> + 3> Methods = crypto:engine_get_all_methods() -- [engine_method_dh,engine_method_rand, +engine_method_ciphers,engine_method_digests, engine_method_store, +engine_method_pkey_meths, engine_method_pkey_asn1_meths]. +[engine_method_rsa,engine_method_dsa, + engine_method_ecdh,engine_method_ecdsa] + 4> {ok, Engine} = crypto:engine_load(<<"dynamic">>, + [{<<"SO_PATH">>, + <<"/some/path/otp_test_engine.so">>}, + {<<"ID">>, <<"MD5">>}, + <<"LOAD">>], + [], + Methods). + {ok, #Ref}</code> + </section> + + <section> + <title>List all engines currently loaded</title> + <code> + 5> crypto:engine_list(). +[<<"dynamic">>, <<"MD5">>]</code> + </section> + + </section> +</chapter> diff --git a/lib/crypto/doc/src/usersguide.xml b/lib/crypto/doc/src/usersguide.xml index 7971aefff4..f637a1db79 100644 --- a/lib/crypto/doc/src/usersguide.xml +++ b/lib/crypto/doc/src/usersguide.xml @@ -11,7 +11,7 @@ Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at - + http://www.apache.org/licenses/LICENSE-2.0 Unless required by applicable law or agreed to in writing, software @@ -19,7 +19,7 @@ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License. - + </legalnotice> <title>Crypto User's Guide</title> @@ -48,5 +48,5 @@ </description> <xi:include href="licenses.xml"/> <xi:include href="fips.xml"/> + <xi:include href="engine_load.xml"/> </part> - diff --git a/lib/crypto/src/Makefile b/lib/crypto/src/Makefile index aea8a5a71c..edad0e6b61 100644 --- a/lib/crypto/src/Makefile +++ b/lib/crypto/src/Makefile @@ -39,8 +39,7 @@ MODULES= \ crypto \ crypto_ec_curves -HRL_FILES= - +HRL_FILES= ERL_FILES= $(MODULES:%=%.erl) TARGET_FILES= $(MODULES:%=$(EBIN)/%.$(EMULATOR)) @@ -56,16 +55,16 @@ APPUP_TARGET= $(EBIN)/$(APPUP_FILE) # ---------------------------------------------------- # FLAGS # ---------------------------------------------------- -ERL_COMPILE_FLAGS += -DCRYPTO_VSN=\"$(VSN)\" -Werror +ERL_COMPILE_FLAGS += -DCRYPTO_VSN=\"$(VSN)\" -Werror -I../include # ---------------------------------------------------- # Targets # ---------------------------------------------------- -debug opt valgrind: $(TARGET_FILES) $(APP_TARGET) $(APPUP_TARGET) +debug opt valgrind: $(TARGET_FILES) $(APP_TARGET) $(APPUP_TARGET) clean: - rm -f $(TARGET_FILES) $(APP_TARGET) $(APPUP_TARGET) + rm -f $(TARGET_FILES) $(APP_TARGET) $(APPUP_TARGET) rm -f errs core *~ $(APP_TARGET): $(APP_SRC) ../vsn.mk @@ -78,7 +77,7 @@ docs: # ---------------------------------------------------- # Release Target -# ---------------------------------------------------- +# ---------------------------------------------------- include $(ERL_TOP)/make/otp_release_targets.mk release_spec: opt @@ -89,10 +88,3 @@ release_spec: opt $(APPUP_TARGET) "$(RELSYSDIR)/ebin" release_docs_spec: - - - - - - - diff --git a/lib/crypto/src/crypto.erl b/lib/crypto/src/crypto.erl index 6b4f3a256d..9ba1623348 100644 --- a/lib/crypto/src/crypto.erl +++ b/lib/crypto/src/crypto.erl @@ -43,12 +43,26 @@ -export([public_encrypt/4, private_decrypt/4]). -export([private_encrypt/4, public_decrypt/4]). -export([dh_generate_parameters/2, dh_check/1]). %% Testing see +-export([privkey_to_pubkey/2]). -export([ec_curve/1, ec_curves/0]). -export([rand_seed/1]). +%% Engine +-export([ + engine_get_all_methods/0, + engine_load/3, + engine_load/4, + engine_unload/1, + engine_list/0 + ]). + +-export_type([engine_ref/0, + key_id/0, + password/0 + ]). -%% Private. For tests. --export([packed_openssl_version/4]). +%% Private. For tests. +-export([packed_openssl_version/4, engine_methods_convert_to_bitmask/2, get_test_engine/0]). -deprecated({rand_uniform, 2, next_major_release}). @@ -460,13 +474,24 @@ sign(Algorithm, Type, Data, Key, Options) -> end. + +-type key_id() :: string() | binary() . +-type password() :: string() | binary() . + +-type engine_key_ref() :: #{engine := engine_ref(), + key_id := key_id(), + password => password(), + term() => term() + }. + -type pk_algs() :: rsa | ecdsa | dss . --type pk_opt() :: list() | rsa_padding() . +-type pk_key() :: engine_key_ref() | [integer() | binary()] . +-type pk_opt() :: list() | rsa_padding() . --spec public_encrypt(pk_algs(), binary(), [binary()], pk_opt()) -> binary(). --spec public_decrypt(pk_algs(), binary(), [integer() | binary()], pk_opt()) -> binary(). --spec private_encrypt(pk_algs(), binary(), [integer() | binary()], pk_opt()) -> binary(). --spec private_decrypt(pk_algs(), binary(), [integer() | binary()], pk_opt()) -> binary(). +-spec public_encrypt(pk_algs(), binary(), pk_key(), pk_opt()) -> binary(). +-spec public_decrypt(pk_algs(), binary(), pk_key(), pk_opt()) -> binary(). +-spec private_encrypt(pk_algs(), binary(), pk_key(), pk_opt()) -> binary(). +-spec private_decrypt(pk_algs(), binary(), pk_key(), pk_opt()) -> binary(). public_encrypt(Algorithm, In, Key, Options) when is_list(Options) -> case pkey_crypt_nif(Algorithm, In, format_pkey(Algorithm, Key), Options, false, true) of @@ -607,10 +632,145 @@ compute_key(ecdh, Others, My, Curve) -> nif_curve_params(Curve), ensure_int_as_bin(My)). +%%====================================================================== +%% Engine functions +%%====================================================================== +%%---------------------------------------------------------------------- +%% Function: engine_get_all_methods/0 +%%---------------------------------------------------------------------- +-type engine_method_type() :: engine_method_rsa | engine_method_dsa | engine_method_dh | + engine_method_rand | engine_method_ecdh | engine_method_ecdsa | + engine_method_ciphers | engine_method_digests | engine_method_store | + engine_method_pkey_meths | engine_method_pkey_asn1_meths | + engine_method_ec. + +-type engine_ref() :: term(). + +-spec engine_get_all_methods() -> + [engine_method_type()]. +engine_get_all_methods() -> + notsup_to_error(engine_get_all_methods_nif()). + +%%---------------------------------------------------------------------- +%% Function: engine_load/3 +%%---------------------------------------------------------------------- +-spec engine_load(EngineId::unicode:chardata(), + PreCmds::[{unicode:chardata(), unicode:chardata()}], + PostCmds::[{unicode:chardata(), unicode:chardata()}]) -> + {ok, Engine::engine_ref()} | {error, Reason::term()}. +engine_load(EngineId, PreCmds, PostCmds) when is_list(PreCmds), is_list(PostCmds) -> + engine_load(EngineId, PreCmds, PostCmds, engine_get_all_methods()). + +%%---------------------------------------------------------------------- +%% Function: engine_load/4 +%%---------------------------------------------------------------------- +-spec engine_load(EngineId::unicode:chardata(), + PreCmds::[{unicode:chardata(), unicode:chardata()}], + PostCmds::[{unicode:chardata(), unicode:chardata()}], + EngineMethods::[engine_method_type()]) -> + {ok, Engine::term()} | {error, Reason::term()}. +engine_load(EngineId, PreCmds, PostCmds, EngineMethods) when is_list(PreCmds), + is_list(PostCmds) -> + try + ok = notsup_to_error(engine_load_dynamic_nif()), + case notsup_to_error(engine_by_id_nif(ensure_bin_chardata(EngineId))) of + {ok, Engine} -> + ok = engine_load_1(Engine, PreCmds, PostCmds, EngineMethods), + {ok, Engine}; + {error, Error1} -> + {error, Error1} + end + catch + throw:Error2 -> + Error2 + end. + +engine_load_1(Engine, PreCmds, PostCmds, EngineMethods) -> + try + ok = engine_nif_wrapper(engine_ctrl_cmd_strings_nif(Engine, ensure_bin_cmds(PreCmds))), + ok = engine_nif_wrapper(engine_add_nif(Engine)), + ok = engine_nif_wrapper(engine_init_nif(Engine)), + engine_load_2(Engine, PostCmds, EngineMethods), + ok + catch + throw:Error -> + %% The engine couldn't initialise, release the structural reference + ok = engine_free_nif(Engine), + throw(Error) + end. + +engine_load_2(Engine, PostCmds, EngineMethods) -> + try + ok = engine_nif_wrapper(engine_ctrl_cmd_strings_nif(Engine, ensure_bin_cmds(PostCmds))), + [ok = engine_nif_wrapper(engine_register_nif(Engine, engine_method_atom_to_int(Method))) || + Method <- EngineMethods], + ok + catch + throw:Error -> + %% The engine registration failed, release the functional reference + ok = engine_finish_nif(Engine), + throw(Error) + end. + +%%---------------------------------------------------------------------- +%% Function: engine_unload/1 +%%---------------------------------------------------------------------- +-spec engine_unload(Engine::term()) -> + ok | {error, Reason::term()}. +engine_unload(Engine) -> + engine_unload(Engine, engine_get_all_methods()). + +-spec engine_unload(Engine::term(), EngineMethods::[engine_method_type()]) -> + ok | {error, Reason::term()}. +engine_unload(Engine, EngineMethods) -> + try + [ok = engine_nif_wrapper(engine_unregister_nif(Engine, engine_method_atom_to_int(Method))) || + Method <- EngineMethods], + ok = engine_nif_wrapper(engine_remove_nif(Engine)), + %% Release the functional reference from engine_init_nif + ok = engine_nif_wrapper(engine_finish_nif(Engine)), + %% Release the structural reference from engine_by_id_nif + ok = engine_nif_wrapper(engine_free_nif(Engine)) + catch + throw:Error -> + Error + end. + +%%---------------------------------------------------------------------- +%% Function: engine_list/0 +%%---------------------------------------------------------------------- +-spec engine_list() -> + [EngineId::binary()]. +engine_list() -> + case notsup_to_error(engine_get_first_nif()) of + {ok, <<>>} -> + []; + {ok, Engine} -> + case notsup_to_error(engine_get_id_nif(Engine)) of + {ok, <<>>} -> + engine_list(Engine, []); + {ok, EngineId} -> + engine_list(Engine, [EngineId]) + end + end. + +engine_list(Engine0, IdList) -> + case notsup_to_error(engine_get_next_nif(Engine0)) of + {ok, <<>>} -> + lists:reverse(IdList); + {ok, Engine1} -> + case notsup_to_error(engine_get_id_nif(Engine1)) of + {ok, <<>>} -> + engine_list(Engine1, IdList); + {ok, EngineId} -> + engine_list(Engine1, [EngineId |IdList]) + end + end. + + %%-------------------------------------------------------------------- %%% On load %%-------------------------------------------------------------------- - on_load() -> LibBaseName = "crypto", PrivDir = code:priv_dir(crypto), @@ -670,12 +830,12 @@ path2bin(Path) when is_list(Path) -> end. %%-------------------------------------------------------------------- -%%% Internal functions +%%% Internal functions %%-------------------------------------------------------------------- max_bytes() -> ?MAX_BYTES_TO_NIF. -notsup_to_error(notsup) -> +notsup_to_error(notsup) -> erlang:error(notsup); notsup_to_error(Other) -> Other. @@ -799,7 +959,7 @@ do_stream_decrypt({rc4, State0}, Data) -> %% -%% AES - in counter mode (CTR) with state maintained for multi-call streaming +%% AES - in counter mode (CTR) with state maintained for multi-call streaming %% -type ctr_state() :: { iodata(), binary(), binary(), integer() } | binary(). @@ -808,11 +968,11 @@ do_stream_decrypt({rc4, State0}, Data) -> { ctr_state(), binary() }. -spec aes_ctr_stream_decrypt(ctr_state(), binary()) -> { ctr_state(), binary() }. - + aes_ctr_stream_init(_Key, _IVec) -> ?nif_stub. aes_ctr_stream_encrypt(_State, _Data) -> ?nif_stub. aes_ctr_stream_decrypt(_State, _Cipher) -> ?nif_stub. - + %% %% RC4 - symmetric stream cipher %% @@ -897,22 +1057,22 @@ pkey_verify_nif(_Algorithm, _Type, _Data, _Signature, _Key, _Options) -> ?nif_st rsa_generate_key_nif(_Bits, _Exp) -> ?nif_stub. %% DH Diffie-Hellman functions -%% +%% %% Generate (and check) Parameters is not documented because they are implemented %% for testing (and offline parameter generation) only. -%% From the openssl doc: +%% From the openssl doc: %% DH_generate_parameters() may run for several hours before finding a suitable prime. -%% Thus dh_generate_parameters may in this implementation block +%% Thus dh_generate_parameters may in this implementation block %% the emulator for several hours. %% -%% usage: dh_generate_parameters(1024, 2 or 5) -> +%% usage: dh_generate_parameters(1024, 2 or 5) -> %% [Prime=mpint(), SharedGenerator=mpint()] dh_generate_parameters(PrimeLen, Generator) -> case dh_generate_parameters_nif(PrimeLen, Generator) of error -> erlang:error(generation_failed, [PrimeLen,Generator]); Ret -> Ret - end. + end. dh_generate_parameters_nif(_PrimeLen, _Generator) -> ?nif_stub. @@ -938,6 +1098,16 @@ ec_curves() -> ec_curve(X) -> crypto_ec_curves:curve(X). + +privkey_to_pubkey(Alg, EngineMap) when Alg == rsa; Alg == dss; Alg == ecdsa -> + case privkey_to_pubkey_nif(Alg, format_pkey(Alg,EngineMap)) of + [_|_]=L -> map_ensure_bin_as_int(L); + X -> X + end. + +privkey_to_pubkey_nif(_Alg, _EngineMap) -> ?nif_stub. + + %% %% EC %% @@ -1005,6 +1175,19 @@ ensure_int_as_bin(Int) when is_integer(Int) -> ensure_int_as_bin(Bin) -> Bin. +map_ensure_bin_as_int(List) when is_list(List) -> + lists:map(fun ensure_bin_as_int/1, List). + +ensure_bin_as_int(Bin) when is_binary(Bin) -> + bin_to_int(Bin); +ensure_bin_as_int(E) -> + E. + +format_pkey(_Alg, #{engine:=_, key_id:=T}=M) when is_binary(T) -> format_pwd(M); +format_pkey(_Alg, #{engine:=_, key_id:=T}=M) when is_list(T) -> format_pwd(M#{key_id:=list_to_binary(T)}); +format_pkey(_Alg, #{engine:=_ }=M) -> error({bad_key_id, M}); +format_pkey(_Alg, #{}=M) -> error({bad_engine_map, M}); +%%% format_pkey(rsa, Key) -> map_ensure_int_as_bin(Key); format_pkey(ecdsa, [Key, Curve]) -> @@ -1014,6 +1197,9 @@ format_pkey(dss, Key) -> format_pkey(_, Key) -> Key. +format_pwd(#{password := Pwd}=M) when is_list(Pwd) -> M#{password := list_to_binary(Pwd)}; +format_pwd(M) -> M. + %%-------------------------------------------------------------------- %% -type rsa_padding() :: 'rsa_pkcs1_padding' | 'rsa_pkcs1_oaep_padding' | 'rsa_no_padding'. @@ -1024,7 +1210,7 @@ pkey_crypt_nif(_Algorithm, _In, _Key, _Options, _IsPrivate, _IsEncrypt) -> ?nif_ %% MP representaion (SSH2) mpint(X) when X < 0 -> mpint_neg(X); mpint(X) -> mpint_pos(X). - + -define(UINT32(X), X:32/unsigned-big-integer). @@ -1032,7 +1218,7 @@ mpint_neg(X) -> Bin = int_to_bin_neg(X, []), Sz = byte_size(Bin), <<?UINT32(Sz), Bin/binary>>. - + mpint_pos(X) -> Bin = int_to_bin_pos(X, []), <<MSB,_/binary>> = Bin, @@ -1054,7 +1240,6 @@ erlint(<<MPIntSize:32/integer,MPIntValue/binary>>) -> %% mod_exp_nif(_Base,_Exp,_Mod,_bin_hdr) -> ?nif_stub. - %%%---------------------------------------------------------------- %% 9470495 == V(0,9,8,zh). %% 268435615 == V(1,0,0,i). @@ -1065,3 +1250,92 @@ packed_openssl_version(MAJ, MIN, FIX, P0) -> P1 = atom_to_list(P0), P = lists:sum([C-$a||C<-P1]), ((((((((MAJ bsl 8) bor MIN) bsl 8 ) bor FIX) bsl 8) bor (P+1)) bsl 4) bor 16#f). + +%%-------------------------------------------------------------------- +%% Engine nifs +engine_by_id_nif(_EngineId) -> ?nif_stub. +engine_init_nif(_Engine) -> ?nif_stub. +engine_finish_nif(_Engine) -> ?nif_stub. +engine_free_nif(_Engine) -> ?nif_stub. +engine_load_dynamic_nif() -> ?nif_stub. +engine_ctrl_cmd_strings_nif(_Engine, _Cmds) -> ?nif_stub. +engine_add_nif(_Engine) -> ?nif_stub. +engine_remove_nif(_Engine) -> ?nif_stub. +engine_register_nif(_Engine, _EngineMethod) -> ?nif_stub. +engine_unregister_nif(_Engine, _EngineMethod) -> ?nif_stub. +engine_get_first_nif() -> ?nif_stub. +engine_get_next_nif(_Engine) -> ?nif_stub. +engine_get_id_nif(_Engine) -> ?nif_stub. +engine_get_all_methods_nif() -> ?nif_stub. + +%%-------------------------------------------------------------------- +%% Engine internals +engine_nif_wrapper(ok) -> + ok; +engine_nif_wrapper(notsup) -> + erlang:error(notsup); +engine_nif_wrapper({error, Error}) -> + throw({error, Error}). + +ensure_bin_chardata(CharData) when is_binary(CharData) -> + CharData; +ensure_bin_chardata(CharData) -> + unicode:characters_to_binary(CharData). + +ensure_bin_cmds(CMDs) -> + ensure_bin_cmds(CMDs, []). + +ensure_bin_cmds([], Acc) -> + lists:reverse(Acc); +ensure_bin_cmds([{Key, Value} |CMDs], Acc) -> + ensure_bin_cmds(CMDs, [{ensure_bin_chardata(Key), ensure_bin_chardata(Value)} | Acc]); +ensure_bin_cmds([Key | CMDs], Acc) -> + ensure_bin_cmds(CMDs, [{ensure_bin_chardata(Key), <<"">>} | Acc]). + +engine_methods_convert_to_bitmask([], BitMask) -> + BitMask; +engine_methods_convert_to_bitmask(engine_method_all, _BitMask) -> + 16#FFFF; +engine_methods_convert_to_bitmask(engine_method_none, _BitMask) -> + 16#0000; +engine_methods_convert_to_bitmask([M |Ms], BitMask) -> + engine_methods_convert_to_bitmask(Ms, BitMask bor engine_method_atom_to_int(M)). + +engine_method_atom_to_int(engine_method_rsa) -> 16#0001; +engine_method_atom_to_int(engine_method_dsa) -> 16#0002; +engine_method_atom_to_int(engine_method_dh) -> 16#0004; +engine_method_atom_to_int(engine_method_rand) -> 16#0008; +engine_method_atom_to_int(engine_method_ecdh) -> 16#0010; +engine_method_atom_to_int(engine_method_ecdsa) -> 16#0020; +engine_method_atom_to_int(engine_method_ciphers) -> 16#0040; +engine_method_atom_to_int(engine_method_digests) -> 16#0080; +engine_method_atom_to_int(engine_method_store) -> 16#0100; +engine_method_atom_to_int(engine_method_pkey_meths) -> 16#0200; +engine_method_atom_to_int(engine_method_pkey_asn1_meths) -> 16#0400; +engine_method_atom_to_int(engine_method_ec) -> 16#0800; +engine_method_atom_to_int(X) -> + erlang:error(badarg, [X]). + +get_test_engine() -> + Type = erlang:system_info(system_architecture), + LibDir = filename:join([code:priv_dir(crypto), "lib"]), + ArchDir = filename:join([LibDir, Type]), + case filelib:is_dir(ArchDir) of + true -> check_otp_test_engine(ArchDir); + false -> check_otp_test_engine(LibDir) + end. + +check_otp_test_engine(LibDir) -> + case filelib:wildcard("otp_test_engine*", LibDir) of + [] -> + {error, notexist}; + [LibName] -> + LibPath = filename:join(LibDir,LibName), + case filelib:is_file(LibPath) of + true -> + {ok, unicode:characters_to_binary(LibPath)}; + false -> + {error, notexist} + end + end. + diff --git a/lib/crypto/test/Makefile b/lib/crypto/test/Makefile index 138081d386..e046a25338 100644 --- a/lib/crypto/test/Makefile +++ b/lib/crypto/test/Makefile @@ -7,7 +7,8 @@ include $(ERL_TOP)/make/$(TARGET)/otp.mk MODULES = \ blowfish_SUITE \ - crypto_SUITE + crypto_SUITE \ + engine_SUITE ERL_FILES= $(MODULES:%=%.erl) @@ -27,7 +28,7 @@ RELSYSDIR = $(RELEASE_PATH)/crypto_test # FLAGS # ---------------------------------------------------- ERL_MAKE_FLAGS += -ERL_COMPILE_FLAGS += +ERL_COMPILE_FLAGS += +nowarn_export_all EBIN = . MAKE_EMAKE = $(wildcard $(ERL_TOP)/make/make_emakefile) @@ -77,7 +78,7 @@ release_spec: release_tests_spec: $(TEST_TARGET) $(INSTALL_DIR) "$(RELSYSDIR)" $(INSTALL_DATA) crypto.spec crypto.cover $(RELTEST_FILES) "$(RELSYSDIR)" - @tar cfh - crypto_SUITE_data | (cd "$(RELSYSDIR)"; tar xf -) + @tar cfh - *_SUITE_data | (cd "$(RELSYSDIR)"; tar xf -) chmod -R u+w "$(RELSYSDIR)" release_docs_spec: diff --git a/lib/crypto/test/crypto_SUITE.erl b/lib/crypto/test/crypto_SUITE.erl index 69f02d3da6..6dab459df6 100644 --- a/lib/crypto/test/crypto_SUITE.erl +++ b/lib/crypto/test/crypto_SUITE.erl @@ -198,7 +198,7 @@ init_per_suite(Config) -> %% This is NOT how you want to do seeding, it is just here %% to make the tests pass. Check your OS manual for how you %% really want to seed. - {H,M,L} = erlang:now(), + {H,M,L} = erlang:timestamp(), Bin = <<H:24,M:20,L:20>>, crypto:rand_seed(<< <<Bin/binary>> || _ <- lists:seq(1,16) >>), Config diff --git a/lib/crypto/test/engine_SUITE.erl b/lib/crypto/test/engine_SUITE.erl new file mode 100644 index 0000000000..72bd59f8ab --- /dev/null +++ b/lib/crypto/test/engine_SUITE.erl @@ -0,0 +1,513 @@ +%% +%% %CopyrightBegin% +%% +%% Copyright Ericsson AB 2017-2017. All Rights Reserved. +%% +%% Licensed under the Apache License, Version 2.0 (the "License"); +%% you may not use this file except in compliance with the License. +%% You may obtain a copy of the License at +%% +%% http://www.apache.org/licenses/LICENSE-2.0 +%% +%% Unless required by applicable law or agreed to in writing, software +%% distributed under the License is distributed on an "AS IS" BASIS, +%% WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +%% See the License for the specific language governing permissions and +%% limitations under the License. +%% +%% %CopyrightEnd% +%% +%% + +-module(engine_SUITE). + +-include_lib("common_test/include/ct.hrl"). + +%% Note: This directive should only be used in test suites. +-compile(export_all). + +%%-------------------------------------------------------------------- +%% Common Test interface functions ----------------------------------- +%%-------------------------------------------------------------------- +suite() -> + [{ct_hooks,[ts_install_cth]}, + {timetrap,{seconds, 10}} + ]. + +all() -> + [ + get_all_possible_methods, + engine_load_all_methods, + engine_load_some_methods, + bad_arguments, + unknown_engine, + pre_command_fail_bad_value, + pre_command_fail_bad_key, + failed_engine_init, + {group, engine_stored_key} + ]. + +groups() -> + [{engine_stored_key, [], + [sign_verify_rsa, + sign_verify_dsa, + sign_verify_ecdsa, + sign_verify_rsa_pwd, + priv_encrypt_pub_decrypt_rsa, + priv_encrypt_pub_decrypt_rsa_pwd, + pub_encrypt_priv_decrypt_rsa, + pub_encrypt_priv_decrypt_rsa_pwd, + get_pub_from_priv_key_dsa, + get_pub_from_priv_key_ecdsa + ]}]. + + +init_per_suite(Config) -> + try crypto:start() of + ok -> + Config; + {error,{already_started,crypto}} -> + Config + catch _:_ -> + {skip, "Crypto did not start"} + end. +end_per_suite(_Config) -> + ok. + +%%-------------------------------------------------------------------- +init_per_group(engine_stored_key, Config) -> + case load_storage_engine(Config) of + {ok, E} -> + KeyDir = key_dir(Config), + [{storage_engine,E}, {storage_dir,KeyDir} | Config]; + {error, notexist} -> + {skip, "OTP Test engine not found"}; + {error, notsup} -> + {skip, "Engine not supported on this OpenSSL version"}; + {error, bad_engine_id} -> + {skip, "Dynamic Engine not supported"}; + Other -> + ct:log("Engine load failed: ~p",[Other]), + {fail, "Engine load failed"} + end; +init_per_group(_Group, Config0) -> + Config0. + +end_per_group(engine_stored_key, Config) -> + case proplists:get_value(storage_engine, Config) of + undefined -> + ok; + E -> + ok = crypto:engine_unload(E) + end; +end_per_group(_, _) -> + ok. + +%%-------------------------------------------------------------------- +init_per_testcase(_Case, Config) -> + Config. +end_per_testcase(_Case, _Config) -> + ok. + +%%------------------------------------------------------------------------- +%% Test cases starts here. +%%------------------------------------------------------------------------- +get_all_possible_methods() -> + [{doc, "Just fetch all possible engine methods supported."}]. + +get_all_possible_methods(Config) when is_list(Config) -> + try + List = crypto:engine_get_all_methods(), + ct:log("crypto:engine_get_all_methods() -> ~p\n", [List]), + ok + catch + error:notsup -> + {skip, "Engine not supported on this OpenSSL version"} + end. + +engine_load_all_methods()-> + [{doc, "Use a dummy md5 engine that does not implement md5" + "but rather returns a static binary to test that crypto:engine_load " + "functions works."}]. + +engine_load_all_methods(Config) when is_list(Config) -> + case crypto:get_test_engine() of + {error, notexist} -> + {skip, "OTP Test engine not found"}; + {ok, Engine} -> + try + Md5Hash1 = <<106,30,3,246,166,222,229,158,244,217,241,179,50,232,107,109>>, + Md5Hash1 = crypto:hash(md5, "Don't panic"), + Md5Hash2 = <<0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15>>, + case crypto:engine_load(<<"dynamic">>, + [{<<"SO_PATH">>, Engine}, + {<<"ID">>, <<"MD5">>}, + <<"LOAD">>], + []) of + {ok, E} -> + case crypto:hash(md5, "Don't panic") of + Md5Hash1 -> + ct:fail(fail_to_load_still_original_engine); + Md5Hash2 -> + ok; + _ -> + ct:fail(fail_to_load_engine) + end, + ok = crypto:engine_unload(E), + case crypto:hash(md5, "Don't panic") of + Md5Hash2 -> + ct:fail(fail_to_unload_still_test_engine); + Md5Hash1 -> + ok; + _ -> + ct:fail(fail_to_unload_engine) + end; + {error, bad_engine_id} -> + {skip, "Dynamic Engine not supported"} + end + catch + error:notsup -> + {skip, "Engine not supported on this OpenSSL version"} + end + end. + +engine_load_some_methods()-> + [{doc, "Use a dummy md5 engine that does not implement md5" + "but rather returns a static binary to test that crypto:engine_load " + "functions works."}]. + +engine_load_some_methods(Config) when is_list(Config) -> + case crypto:get_test_engine() of + {error, notexist} -> + {skip, "OTP Test engine not found"}; + {ok, Engine} -> + try + Md5Hash1 = <<106,30,3,246,166,222,229,158,244,217,241,179,50,232,107,109>>, + Md5Hash1 = crypto:hash(md5, "Don't panic"), + Md5Hash2 = <<0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15>>, + EngineMethods = crypto:engine_get_all_methods() -- + [engine_method_dh,engine_method_rand, + engine_method_ciphers, engine_method_store, + engine_method_pkey_meths, engine_method_pkey_asn1_meths], + case crypto:engine_load(<<"dynamic">>, + [{<<"SO_PATH">>, Engine}, + {<<"ID">>, <<"MD5">>}, + <<"LOAD">>], + [], + EngineMethods) of + {ok, E} -> + case crypto:hash(md5, "Don't panic") of + Md5Hash1 -> + ct:fail(fail_to_load_engine_still_original); + Md5Hash2 -> + ok; + _ -> + ct:fail(fail_to_load_engine) + end, + ok = crypto:engine_unload(E), + case crypto:hash(md5, "Don't panic") of + Md5Hash2 -> + ct:fail(fail_to_unload_still_test_engine); + Md5Hash1 -> + ok; + _ -> + ct:fail(fail_to_unload_engine) + end; + {error, bad_engine_id} -> + {skip, "Dynamic Engine not supported"} + end + catch + error:notsup -> + {skip, "Engine not supported on this OpenSSL version"} + end + end. + +%%------------------------------------------------------------------------- +%% Error cases +bad_arguments()-> + [{doc, "Test different arguments in bad format."}]. + +bad_arguments(Config) when is_list(Config) -> + case crypto:get_test_engine() of + {error, notexist} -> + {skip, "OTP Test engine not found"}; + {ok, Engine} -> + try + try + crypto:engine_load(fail_engine, [], []) + catch + error:badarg -> + ok + end, + try + crypto:engine_load(<<"dynamic">>, + [{<<"SO_PATH">>, Engine}, + 1, + {<<"ID">>, <<"MD5">>}, + <<"LOAD">>], + []) + catch + error:badarg -> + ok + end, + try + crypto:engine_load(<<"dynamic">>, + [{<<"SO_PATH">>, Engine}, + {'ID', <<"MD5">>}, + <<"LOAD">>], + []) + catch + error:badarg -> + ok + end + catch + error:notsup -> + {skip, "Engine not supported on this OpenSSL version"} + end + end. + +unknown_engine() -> + [{doc, "Try to load a non existent engine."}]. + +unknown_engine(Config) when is_list(Config) -> + try + {error, bad_engine_id} = crypto:engine_load(<<"fail_engine">>, [], []), + ok + catch + error:notsup -> + {skip, "Engine not supported on this OpenSSL version"} + end. + +pre_command_fail_bad_value() -> + [{doc, "Test pre command due to bad value"}]. + +pre_command_fail_bad_value(Config) when is_list(Config) -> + DataDir = unicode:characters_to_binary(code:priv_dir(crypto)), + try + case crypto:engine_load(<<"dynamic">>, + [{<<"SO_PATH">>, + <<DataDir/binary, <<"/libfail_engine.so">>/binary >>}, + {<<"ID">>, <<"MD5">>}, + <<"LOAD">>], + []) of + {error, ctrl_cmd_failed} -> + ok; + {error, bad_engine_id} -> + {skip, "Dynamic Engine not supported"} + end + catch + error:notsup -> + {skip, "Engine not supported on this OpenSSL version"} + end. + +pre_command_fail_bad_key() -> + [{doc, "Test pre command due to bad key"}]. + +pre_command_fail_bad_key(Config) when is_list(Config) -> + try + case crypto:get_test_engine() of + {error, notexist} -> + {skip, "OTP Test engine not found"}; + {ok, Engine} -> + case crypto:engine_load(<<"dynamic">>, + [{<<"SO_WRONG_PATH">>, Engine}, + {<<"ID">>, <<"MD5">>}, + <<"LOAD">>], + []) of + {error, ctrl_cmd_failed} -> + ok; + {error, bad_engine_id} -> + {skip, "Dynamic Engine not supported"} + end + end + catch + error:notsup -> + {skip, "Engine not supported on this OpenSSL version"} + end. + +failed_engine_init()-> + [{doc, "Test failing engine init due to missed pre command"}]. + +failed_engine_init(Config) when is_list(Config) -> + try + case crypto:get_test_engine() of + {error, notexist} -> + {skip, "OTP Test engine not found"}; + {ok, Engine} -> + case crypto:engine_load(<<"dynamic">>, + [{<<"SO_PATH">>, Engine}, + {<<"ID">>, <<"MD5">>}], + []) of + {error, add_engine_failed} -> + ok; + {error, bad_engine_id} -> + {skip, "Dynamic Engine not supported"} + end + end + catch + error:notsup -> + {skip, "Engine not supported on this OpenSSL version"} + end. + +%%%---------------------------------------------------------------- +%%% Pub/priv key storage tests. Thoose are for testing the crypto.erl +%%% support for using priv/pub keys stored in an engine. + +sign_verify_rsa(Config) -> + Priv = #{engine => engine_ref(Config), + key_id => key_id(Config, "rsa_private_key.pem")}, + Pub = #{engine => engine_ref(Config), + key_id => key_id(Config, "rsa_public_key.pem")}, + sign_verify(rsa, sha, Priv, Pub). + +sign_verify_dsa(Config) -> + Priv = #{engine => engine_ref(Config), + key_id => key_id(Config, "dsa_private_key.pem")}, + Pub = #{engine => engine_ref(Config), + key_id => key_id(Config, "dsa_public_key.pem")}, + sign_verify(dss, sha, Priv, Pub). + +sign_verify_ecdsa(Config) -> + Priv = #{engine => engine_ref(Config), + key_id => key_id(Config, "ecdsa_private_key.pem")}, + Pub = #{engine => engine_ref(Config), + key_id => key_id(Config, "ecdsa_public_key.pem")}, + sign_verify(ecdsa, sha, Priv, Pub). + +sign_verify_rsa_pwd(Config) -> + Priv = #{engine => engine_ref(Config), + key_id => key_id(Config, "rsa_private_key_pwd.pem"), + password => "password"}, + Pub = #{engine => engine_ref(Config), + key_id => key_id(Config, "rsa_public_key_pwd.pem")}, + sign_verify(rsa, sha, Priv, Pub). + +priv_encrypt_pub_decrypt_rsa(Config) -> + Priv = #{engine => engine_ref(Config), + key_id => key_id(Config, "rsa_private_key.pem")}, + Pub = #{engine => engine_ref(Config), + key_id => key_id(Config, "rsa_public_key.pem")}, + priv_enc_pub_dec(rsa, Priv, Pub, rsa_pkcs1_padding). + +priv_encrypt_pub_decrypt_rsa_pwd(Config) -> + Priv = #{engine => engine_ref(Config), + key_id => key_id(Config, "rsa_private_key_pwd.pem"), + password => "password"}, + Pub = #{engine => engine_ref(Config), + key_id => key_id(Config, "rsa_public_key_pwd.pem")}, + priv_enc_pub_dec(rsa, Priv, Pub, rsa_pkcs1_padding). + +pub_encrypt_priv_decrypt_rsa(Config) -> + Priv = #{engine => engine_ref(Config), + key_id => key_id(Config, "rsa_private_key.pem")}, + Pub = #{engine => engine_ref(Config), + key_id => key_id(Config, "rsa_public_key.pem")}, + pub_enc_priv_dec(rsa, Pub, Priv, rsa_pkcs1_padding). + +pub_encrypt_priv_decrypt_rsa_pwd(Config) -> + Priv = #{engine => engine_ref(Config), + key_id => key_id(Config, "rsa_private_key.pem"), + password => "password"}, + Pub = #{engine => engine_ref(Config), + key_id => key_id(Config, "rsa_public_key.pem")}, + pub_enc_priv_dec(rsa, Pub, Priv, rsa_pkcs1_padding). + +get_pub_from_priv_key_rsa(Config) -> + Priv = #{engine => engine_ref(Config), + key_id => key_id(Config, "rsa_private_key.pem")}, + Pub = crypto:privkey_to_pubkey(rsa, Priv), + ct:log("rsa Pub = ~p",[Pub]), + sign_verify(rsa, sha, Priv, Pub). + +get_pub_from_priv_key_dsa(Config) -> + Priv = #{engine => engine_ref(Config), + key_id => key_id(Config, "dsa_private_key.pem")}, + Pub = crypto:privkey_to_pubkey(dss, Priv), + ct:log("dsa Pub = ~p",[Pub]), + sign_verify(dss, sha, Priv, Pub). + +get_pub_from_priv_key_ecdsa(Config) -> + Priv = #{engine => engine_ref(Config), + key_id => key_id(Config, "ecdsa_private_key.pem")}, + Pub = crypto:privkey_to_pubkey(ecdsa, Priv), + case Pub of + notsup -> {skip, "ECDSA not implemented"}; + _ -> + ct:log("ecdsa Pub = ~p",[Pub]), + sign_verify(ecdsa, sha, Priv, Pub) + end. + +%%%================================================================ +%%% Help for engine_stored_pub_priv_keys* test cases +%%% +load_storage_engine(_Config) -> + case crypto:get_test_engine() of + {ok, Engine} -> + try crypto:engine_load(<<"dynamic">>, + [{<<"SO_PATH">>, Engine}, + <<"LOAD">>], + []) + catch + error:notsup -> + {error, notsup} + end; + + {error, Error} -> + {error, Error} + end. + + +key_dir(Config) -> + DataDir = unicode:characters_to_binary(proplists:get_value(data_dir, Config)), + filename:join(DataDir, "pkcs8"). + + +engine_ref(Config) -> + proplists:get_value(storage_engine, Config). + +key_id(Config, File) -> + filename:join(proplists:get_value(storage_dir,Config), File). + +pubkey_alg_supported(Alg) -> + lists:member(Alg, + proplists:get_value(public_keys, crypto:supports())). + + +pub_enc_priv_dec(Alg, KeyEnc, KeyDec, Padding) -> + case pubkey_alg_supported(Alg) of + true -> + PlainText = <<"Hej på dig">>, + CryptoText = crypto:public_encrypt(Alg, PlainText, KeyEnc, Padding), + case crypto:private_decrypt(Alg, CryptoText, KeyDec, Padding) of + PlainText -> ok; + _ -> {fail, "Encrypt-decrypt error"} + end; + false -> + {skip, lists:concat([Alg," is not supported by cryptolib"])} + end. + +priv_enc_pub_dec(Alg, KeyEnc, KeyDec, Padding) -> + case pubkey_alg_supported(Alg) of + true -> + PlainText = <<"Hej på dig">>, + CryptoText = crypto:private_encrypt(Alg, PlainText, KeyEnc, Padding), + case crypto:public_decrypt(Alg, CryptoText, KeyDec, Padding) of + PlainText -> ok; + _ -> {fail, "Encrypt-decrypt error"} + end; + false -> + {skip, lists:concat([Alg," is not supported by cryptolib"])} + end. + +sign_verify(Alg, Sha, KeySign, KeyVerify) -> + case pubkey_alg_supported(Alg) of + true -> + PlainText = <<"Hej på dig">>, + Signature = crypto:sign(Alg, Sha, PlainText, KeySign), + case crypto:verify(Alg, Sha, PlainText, Signature, KeyVerify) of + true -> ok; + _ -> {fail, "Sign-verify error"} + end; + false -> + {skip, lists:concat([Alg," is not supported by cryptolib"])} + end. diff --git a/lib/crypto/test/engine_SUITE_data/pkcs8/dsa_private_key.pem b/lib/crypto/test/engine_SUITE_data/pkcs8/dsa_private_key.pem new file mode 100644 index 0000000000..778ffac675 --- /dev/null +++ b/lib/crypto/test/engine_SUITE_data/pkcs8/dsa_private_key.pem @@ -0,0 +1,9 @@ +-----BEGIN PRIVATE KEY----- +MIIBSwIBADCCASwGByqGSM44BAEwggEfAoGBAMyitTMR7vPbpqyAXJpqnB0AhFwQ +F87IE+JKFl5bD/MSkhhRV5sM73HUU1ooXY0FjhZ+cdLUCATuZR5ta4ydANqWIcAB +gX3IwF1B4zf5SXEKTWkUYneL9dOKtiZLtoG28swrk8xMxwX+0fLHkltCEj6FiTW9 +PFrv8GmIfV6DjcI9AhUAqXWbb3RtoN9Ld28fVMhGZrj3LJUCgYEAwnxGHGBMpJaF +2w7zAw3jHjL8PMYlV6vnufGHQlwF0ZUXJxRsvagMb/X1qACTu2VPYEVoLQGM3cfH +EhHoQmvSXGAyTfR7Bmn3gf1n/s/DcFbdZduUCZ/rAyIrfd0eSbc1I+kZk85UCsKK +w/IYdlqcuYa4Cgm2TapT5uEMqH4jhzEEFgIULh8swEUWmU8aJNWsrWl4eCiuUUg= +-----END PRIVATE KEY----- diff --git a/lib/crypto/test/engine_SUITE_data/pkcs8/dsa_public_key.pem b/lib/crypto/test/engine_SUITE_data/pkcs8/dsa_public_key.pem new file mode 100644 index 0000000000..0fa5428828 --- /dev/null +++ b/lib/crypto/test/engine_SUITE_data/pkcs8/dsa_public_key.pem @@ -0,0 +1,12 @@ +-----BEGIN PUBLIC KEY----- +MIIBtzCCASwGByqGSM44BAEwggEfAoGBAMyitTMR7vPbpqyAXJpqnB0AhFwQF87I +E+JKFl5bD/MSkhhRV5sM73HUU1ooXY0FjhZ+cdLUCATuZR5ta4ydANqWIcABgX3I +wF1B4zf5SXEKTWkUYneL9dOKtiZLtoG28swrk8xMxwX+0fLHkltCEj6FiTW9PFrv +8GmIfV6DjcI9AhUAqXWbb3RtoN9Ld28fVMhGZrj3LJUCgYEAwnxGHGBMpJaF2w7z +Aw3jHjL8PMYlV6vnufGHQlwF0ZUXJxRsvagMb/X1qACTu2VPYEVoLQGM3cfHEhHo +QmvSXGAyTfR7Bmn3gf1n/s/DcFbdZduUCZ/rAyIrfd0eSbc1I+kZk85UCsKKw/IY +dlqcuYa4Cgm2TapT5uEMqH4jhzEDgYQAAoGAXPygOFYdeKgfLmuIC303cESYXvic +e2GNJomv8vaWLZmbLVVDfwA1fNsuF1hZkWw8f7aYaN9iZ3yl9u4Yk4TbJKkqfJqd +dgVt288SUqvi+NMHODUzYi9KAOXxupXffZSvdu54gKRaDuFTZ5XNcRqIJWGYlJYg +NVHF5FPZ4Bk2FYA= +-----END PUBLIC KEY----- diff --git a/lib/crypto/test/engine_SUITE_data/pkcs8/ecdsa_private_key.pem b/lib/crypto/test/engine_SUITE_data/pkcs8/ecdsa_private_key.pem new file mode 100644 index 0000000000..a45522064f --- /dev/null +++ b/lib/crypto/test/engine_SUITE_data/pkcs8/ecdsa_private_key.pem @@ -0,0 +1,8 @@ +-----BEGIN PRIVATE KEY----- +MIHuAgEAMBAGByqGSM49AgEGBSuBBAAjBIHWMIHTAgEBBEIBparGjr0KcdNrVM2J +G0mW5ltP1QyvxDqBMyWLWo3fruRZv6Qoohl5skd1u4O+KJoM/UrrSTOXI/MDR7NN +i1yl7O+hgYkDgYYABAG8K2XVsK0ahG9+HIIPwCO0pJY8ulwSTXwIjkCGyB2lpglh +8qJmRzuyGcfRTslv8wfv0sPlT9H9PKDvgrTUL7rvQQDdOODNgVPXSecUoXoPn+X+ +eqxs77bjx+A5x0t/i3m5PfkaNPh5MZ1H/bWuOOdj2ZXZw0R4rlVc0zVrgnPU8L8S +BQ== +-----END PRIVATE KEY----- diff --git a/lib/crypto/test/engine_SUITE_data/pkcs8/ecdsa_public_key.pem b/lib/crypto/test/engine_SUITE_data/pkcs8/ecdsa_public_key.pem new file mode 100644 index 0000000000..6d22fe43fe --- /dev/null +++ b/lib/crypto/test/engine_SUITE_data/pkcs8/ecdsa_public_key.pem @@ -0,0 +1,6 @@ +-----BEGIN PUBLIC KEY----- +MIGbMBAGByqGSM49AgEGBSuBBAAjA4GGAAQBvCtl1bCtGoRvfhyCD8AjtKSWPLpc +Ek18CI5AhsgdpaYJYfKiZkc7shnH0U7Jb/MH79LD5U/R/Tyg74K01C+670EA3Tjg +zYFT10nnFKF6D5/l/nqsbO+248fgOcdLf4t5uT35GjT4eTGdR/21rjjnY9mV2cNE +eK5VXNM1a4Jz1PC/EgU= +-----END PUBLIC KEY----- diff --git a/lib/crypto/test/engine_SUITE_data/pkcs8/rsa_private_key.pem b/lib/crypto/test/engine_SUITE_data/pkcs8/rsa_private_key.pem new file mode 100644 index 0000000000..ea0e3d3958 --- /dev/null +++ b/lib/crypto/test/engine_SUITE_data/pkcs8/rsa_private_key.pem @@ -0,0 +1,28 @@ +-----BEGIN PRIVATE KEY----- +MIIEvwIBADANBgkqhkiG9w0BAQEFAASCBKkwggSlAgEAAoIBAQCwwb0/ddXGXTFK +4FLxXdV6a/WJMSoPPS55RvZIAHFsiTtvPLbJ8LxDsZ6wSVZLN0/UQ4wdWn9jftyj +U5/IxBVG8XOtKimTMvm3/ZOzVLueGHBbrLYscRv9oL85ulTKHWgrZDu0lBX5JJTI +v5UTCErzJRQbka9DG1GaBgDb1PlXfkzBWMwfsBZmwoC77KvCcIGCgbW/XCY03TP2 +3Tg8drvpByMStddP2FQ4fZ91qFUzPu8uhZEsqSQTFlmhgGEx7dLlky0xvu62RuAD +RTpINpcWZtWDHTdssOqu653LwwqBY8lBopCZ/4Af8QR3ZYkQhen1YLEbVheXRuzI +LSCZIiJNAgMBAAECggEBAJH4/fxpqQkvr2Shy33Pu1xlyhnpw01gfn/jrcKasxEq +aC4eWup86E2TY3U8q4pkfIXU3uLi+O9HNpmflwargNLc1mY8uqb44ygiv5bLNEKE +9k2PXcdoBfC4jxPyoNFl5cBn/7LK1TazEjiTl15na9ZPWcLG1pG5/vMPYCgsQ1sP +8J3c4E3aaXIj9QceYxBprl490OCzieGyZlRipncz3g4UShRc/b4cycvDZOJpmAy4 +zbWTcBcSMPVPi5coF0K8UcimiqZkotfb/2RLc433i34IdsIXMM+brdq+g8rmjg5a ++oQPy02M6tFApBruEhAz8DGgaLtDY6MLtyZAt3SjXnUCgYEA1zLgamdTHOqrrmIi +eIQBnAJiyIfcY8B9SX1OsLGYFCHiPVwgUY35B2c7MavMsGcExJhtE+uxU7o5djtM +R6r9cRHOXJ6EQwa8OwzzPqbM17/YqNDeK39bc9WOFUqRWrhDhVMPy6z8rmZr73mG +IUC7mBNx/1GBdVYXIlsXzC96dI8CgYEA0kUAhz6I5nyPa70NDEUYHLHf3IW1BCmE +UoVbraSePJtIEY/IqFx7oDuFo30d4n5z+8ICCtyid1h/Cp3mf3akOiqltYUfgV1G +JgcEjKKYWEnO7cfFyO7LB7Y3GYYDJNy6EzVWPiwTGk9ZTfFJEESmHC45Unxgd17m +Dx/R58rFgWMCgYBQXQWFdtSI5fH7C1bIHrPjKNju/h2FeurOuObcAVZDnmu4cmD3 +U8d9xkVKxVeJQM99A1coq0nrdI3k4zwXP3mp8fZYjDHkPe2pN6rW6L9yiohEcsuk +/siON1/5/4DMmidM8LnjW9R45HLGWWGHpX7oyco2iJ+Jy/6Tq+T1MX3PbQKBgQCm +hdsbQJ0u3CrBSmFQ/E9SOlRt0r4+45pVuCOY6yweF2QF9HcXTtbhWQJHLclDHJ5C +Ha18aKuKFN3XzKFFBPKe1jOSBDGlQ/dQGnKx5fr8wMdObM3oiaTlIJuWbRmEUgJT +QARjDIi8Z2b0YUhZx+Q9oSXoe3PyVYehJrQX+/BavQKBgQCIr7Zp0rQPbfqcTL+M +OYHUoNcb14f9f8hXeXHQOqVpsGwxGdRQAU9wbx/4+obKB5xIkzBsVNcJwavisNja +hegnGjTB/9Hc4m+5bMGwH0bhS2eQO4o+YYM2ypDmFQqDLRfFUlZ5PVHffm/aA9+g +GanNBCsmtoHtV6CJ1UZ7NmBuIA== +-----END PRIVATE KEY----- diff --git a/lib/crypto/test/engine_SUITE_data/pkcs8/rsa_private_key_pwd.pem b/lib/crypto/test/engine_SUITE_data/pkcs8/rsa_private_key_pwd.pem new file mode 100644 index 0000000000..501662fc35 --- /dev/null +++ b/lib/crypto/test/engine_SUITE_data/pkcs8/rsa_private_key_pwd.pem @@ -0,0 +1,30 @@ +-----BEGIN ENCRYPTED PRIVATE KEY----- +MIIFDjBABgkqhkiG9w0BBQ0wMzAbBgkqhkiG9w0BBQwwDgQIh888Iq6gxuMCAggA +MBQGCCqGSIb3DQMHBAic/11YZ8Nt5gSCBMjG/Jb4qiMoBS50iQvHXqcETPE+0NBr +jhsn9w94LkdRBstMPAsoKmY98Er96Rnde/NfmqlU9CupKTkd7Ce5poBf72Y6KMED +cPURyjbGRFsu6x9skXB2obhyKYEqAEF2oQAg4Qbe5v1qXBIgDuC/NgiJnM+w2zCZ +LkHSZB2/NmcnvDzcgPF7TM8pTO23xCJ33m37qjfWvHsgocVqZmL9wQ4+wr/NMYjJ +pJvX1OHW1vBsZsXh40WchalYRSB1VeO368QfsE8coRJztqbMzdce9EQdMB6Q6jlO +cetd3moLIoMP4I7HW0/SgokbycTbRiYSvRyU1TGc2WbW6BrFZV24IckcnnVUFatf +6HKUcaYLG68dJcRgs5QMGkcmgVvlddENHFmHZlo0eym/xSiUl/AT8/5odscm6ML8 +wW5sneax+TF4J2eYmiN7yjAUCodXVTNYNDVKo6uUhntlymbM0o4UitVIbPIfTDHl +sxJAEZ7vpuPqeNMxUk6G6zipuEjqsVbnuFSBSZmgKiGYcifRPUmqqINa3DdS4WVx +xaPWdHbHVRD//ze3h/FsA+1lIE5q2kUE0xXseJA1ISog++kJp14XeaaL2j/tx3Ob +OsbcaOAD/IUw/ItDt9kn0qzfnar7sS0Wov8AmJQxHmH7Lm93jHTLM05yE0AR/eBr +Mig2ZdC+9OqVC+GPuBkRjSs8NpltQIDroz6EV9IMwPwXm0szSYoyoPLmlHJUdnLs +ZUef+au6hYkEJBrvuisagnq5eT/fCV3hsjD7yODebNU2CmBTo6X2PRx/xsBHRMWl +QkoM9PBdSCnKv6HpHl4pchuoqU2NpFjN0BCaad6aHfZSTnqgzK4bEh1oO6dI8/rB +/eh71JyFFG5J4xbpaqz5Su01V1iwU5leK5bDwqals4M4+ZGHGciou7qnXUmX2fJl +r6DlMUa/xy+A2ZG0NuZR05yk2oB3+KVNMgp6zFty3XaxwoNtc8GTLtLnBnIh2rlP +mE1+I65LRWwrNQalPeOAUrYuEzhyp2Df7a8Ykas5PUH7MGR/S0Ge/dLxtE2bJuK4 +znbLAsGhvo/SbNxYqIp6D4iDtd3va6yUGncy41paA/vTKFVvXZDrXcwJQYYCVOGT +OwdzNuozU8Dc7oxsd8oakfC46kvmVaOrGvZbm56PFfprcaL/Hslska5xxEni/eZe +WRxZbCBhAVqS1pn5zkDQVUe9uFlR/x39Qi01HIlKLBsjpSs6qQsFArMe8hgXmXLG +xP+dyVuOE18NzSewdEjeqSRKIM7Qi8EOjZsI4HdSRBY7bh9VhmaVXDZiCSf33TTE +3y8nimzQAeuGoYg6WqHmWWC2Qnpki2HlaIH/ayXEyQWkP/qvg61e8ovdg9Fy8JOO +0AacXVt5zj0q00AW5bKx7usi4NIjZedi86hUm6H19aBm7r86BKjwYTEI/GOcdrbV +9HC/8ayOimgwiAG3gq+aLioWym+Z6KnsbVd7XReVbvM/InQx54WA2y5im0A+/c67 +oQFFPV84XGX9waeqv/K4Wzkm6HW+qVAEM67482VGOf0PVrlQMno6dOotT/Y7ljoZ +2iz0LmN9yylJnLPDrr1i6gzbs5OhhUgbF5LI2YP2wWdCZTl/DrKSIvQZWl8U+tw3 +ciA= +-----END ENCRYPTED PRIVATE KEY----- diff --git a/lib/crypto/test/engine_SUITE_data/pkcs8/rsa_public_key.pem b/lib/crypto/test/engine_SUITE_data/pkcs8/rsa_public_key.pem new file mode 100644 index 0000000000..d3fb5a2cc9 --- /dev/null +++ b/lib/crypto/test/engine_SUITE_data/pkcs8/rsa_public_key.pem @@ -0,0 +1,9 @@ +-----BEGIN PUBLIC KEY----- +MIIBIjANBgkqhkiG9w0BAQEFAAOCAQ8AMIIBCgKCAQEAsMG9P3XVxl0xSuBS8V3V +emv1iTEqDz0ueUb2SABxbIk7bzy2yfC8Q7GesElWSzdP1EOMHVp/Y37co1OfyMQV +RvFzrSopkzL5t/2Ts1S7nhhwW6y2LHEb/aC/ObpUyh1oK2Q7tJQV+SSUyL+VEwhK +8yUUG5GvQxtRmgYA29T5V35MwVjMH7AWZsKAu+yrwnCBgoG1v1wmNN0z9t04PHa7 +6QcjErXXT9hUOH2fdahVMz7vLoWRLKkkExZZoYBhMe3S5ZMtMb7utkbgA0U6SDaX +FmbVgx03bLDqruudy8MKgWPJQaKQmf+AH/EEd2WJEIXp9WCxG1YXl0bsyC0gmSIi +TQIDAQAB +-----END PUBLIC KEY----- diff --git a/lib/crypto/test/engine_SUITE_data/pkcs8/rsa_public_key_pwd.pem b/lib/crypto/test/engine_SUITE_data/pkcs8/rsa_public_key_pwd.pem new file mode 100644 index 0000000000..f74361cead --- /dev/null +++ b/lib/crypto/test/engine_SUITE_data/pkcs8/rsa_public_key_pwd.pem @@ -0,0 +1,9 @@ +-----BEGIN PUBLIC KEY----- +MIIBIjANBgkqhkiG9w0BAQEFAAOCAQ8AMIIBCgKCAQEAxquo1Na8C+kjeW0YESGm +vE1bgNW9xh+SQjU1fv/97ePK8mQW2zO1h/vUNz23pfZAKjQu3rlFW/VgGJQ0LgCs +8Gr/HbMwNcCJzuFMePUrnWn/qBeR7OKUZCJ3E1pp4kwsTdGDDO7jPtNzKf0bdKlg +G2GHfZWhUediRX8NsRg12X1odVPuRGVRsyJ952YODk9PFjK7pro7Ynf3Icx7di9d +PXL5vEcKSRdomXvt1rgM8XVHES94RQqoz60ZhfV2JnPfa9V8qu0KaGntpEr7p4rQ +5BSiLFPjPOArjsD5tKyo8ldKCdQjLfisEp7AetfMjLPVVPw9o/SmCjDxsYWTVRQ2 +tQIDAQAB +-----END PUBLIC KEY----- diff --git a/lib/dialyzer/src/dialyzer_analysis_callgraph.erl b/lib/dialyzer/src/dialyzer_analysis_callgraph.erl index a4b42c9367..9993c68fed 100644 --- a/lib/dialyzer/src/dialyzer_analysis_callgraph.erl +++ b/lib/dialyzer/src/dialyzer_analysis_callgraph.erl @@ -165,7 +165,11 @@ analysis_start(Parent, Analysis, LegalWarnings) -> remote_type_postprocessing(TmpCServer, Args) -> Fun = fun() -> - exit(remote_type_postproc(TmpCServer, Args)) + exit(try remote_type_postproc(TmpCServer, Args) of + R -> R + catch + throw:{error,_}=Error -> Error + end) end, {Pid, Ref} = erlang:spawn_monitor(Fun), dialyzer_codeserver:give_away(TmpCServer, Pid), diff --git a/lib/dialyzer/test/plt_SUITE.erl b/lib/dialyzer/test/plt_SUITE.erl index ebe79b2a6d..a8a9f176fc 100644 --- a/lib/dialyzer/test/plt_SUITE.erl +++ b/lib/dialyzer/test/plt_SUITE.erl @@ -9,14 +9,14 @@ -export([suite/0, all/0, build_plt/1, beam_tests/1, update_plt/1, local_fun_same_as_callback/1, remove_plt/1, run_plt_check/1, run_succ_typings/1, - bad_dialyzer_attr/1, merge_plts/1]). + bad_dialyzer_attr/1, merge_plts/1, bad_record_type/1]). suite() -> [{timetrap, ?plt_timeout}]. all() -> [build_plt, beam_tests, update_plt, run_plt_check, remove_plt, run_succ_typings, local_fun_same_as_callback, - bad_dialyzer_attr, merge_plts]. + bad_dialyzer_attr, merge_plts, bad_record_type]. build_plt(Config) -> OutDir = ?config(priv_dir, Config), @@ -369,6 +369,32 @@ create_plts(Mod1, Mod2, Config) -> %% End of merge_plts(). +bad_record_type(Config) -> + PrivDir = ?config(priv_dir, Config), + Source = lists:concat([bad_record_type, ".erl"]), + Filename = filename:join(PrivDir, Source), + PltFilename = dialyzer_common:plt_file(PrivDir), + + Opts = [{files, [Filename]}, + {check_plt, false}, + {from, src_code}, + {init_plt, PltFilename}], + + Prog = <<"-module(bad_record_type). + -export([r/0]). + -record(r, {f = 3 :: integer()}). + -spec r() -> #r{f :: atom()}. + r() -> + #r{}.">>, + ok = file:write_file(Filename, Prog), + {dialyzer_error, + "Analysis failed with error:\n" ++ Str} = + (catch dialyzer:run(Opts)), + P = string:str(Str, + "bad_record_type.erl:4: Illegal declaration of #r{f}"), + true = P > 0, + ok. + erlang_beam() -> case code:where_is_file("erlang.beam") of non_existing -> diff --git a/lib/diameter/doc/src/notes.xml b/lib/diameter/doc/src/notes.xml index 589e7d5145..eded788419 100644 --- a/lib/diameter/doc/src/notes.xml +++ b/lib/diameter/doc/src/notes.xml @@ -43,6 +43,23 @@ first.</p> <!-- ===================================================================== --> +<section><title>diameter 2.1.2</title> + + <section><title>Fixed Bugs and Malfunctions</title> + <list> + <item> + <p> + A fault introduced in diameter 2.1 could cause decode + errors to be ignored in AVPs following the header of a + Grouped AVP.</p> + <p> + Own Id: OTP-14684 Aux Id: ERIERL-85 </p> + </item> + </list> + </section> + +</section> + <section><title>diameter 2.1.1</title> <section><title>Fixed Bugs and Malfunctions</title> diff --git a/lib/diameter/src/base/diameter_gen.erl b/lib/diameter/src/base/diameter_gen.erl index d3b9f704fe..93ebe57685 100644 --- a/lib/diameter/src/base/diameter_gen.erl +++ b/lib/diameter/src/base/diameter_gen.erl @@ -320,8 +320,8 @@ decode(Bin, Code, Vid, DataLen, Pad, M, P, Name, Mod, Fmt, Strict, Opts0, index = Idx}, Dec = dec(Data, Name, NameT, Mod, Fmt, Opts, Avp), - Acc = decode(T, Name, Mod, Fmt, Strict, Opts, Idx+1, AM),%% recurse - acc(Acc, Dec, I, Field, Arity, Strict, Mod, Opts); + Acc = decode(T, Name, Mod, Fmt, Strict, Opts0, Idx+1, AM),%% recurse + acc(Acc, Dec, I, Field, Arity, Strict, Mod); _ -> {NameT, _Field, _Arity, {_, AM}} = incr(Name, Code, Vid, M, Mod, Strict, Opts0, AM0), @@ -574,15 +574,17 @@ dec_AVP(Dicts, Data, Name, Mod, Fmt, Opts, #diameter_avp{code = Code, %% Try to decode an AVP in the first alternate dictionary that defines %% it. -dec_AVP([Dict | Rest], Data, Name, Mod, Fmt, Opts, Code, Vid, Avp) -> +dec_AVP([Dict | Rest], Data, Name, Mod, Fmt, Opts0, Code, Vid, Avp) -> case Dict:avp_name(Code, Vid) of - {AvpName, Type} -> + {AvpName, Type} = NameT -> A = Avp#diameter_avp{name = AvpName, type = Type}, - #{failed_avp := Failed} = Opts, + #{failed_avp := Failed} + = Opts + = setopts(NameT, Name, Avp#diameter_avp.is_mandatory, Opts0), dec(Data, Name, AvpName, Type, Mod, Dict, Fmt, Failed, Opts, A); _ -> - dec_AVP(Rest, Data, Name, Mod, Fmt, Opts, Code, Vid, Avp) + dec_AVP(Rest, Data, Name, Mod, Fmt, Opts0, Code, Vid, Avp) end; dec_AVP([], _, _, _, _, _, _, _, Avp) -> @@ -680,30 +682,30 @@ set_failed('Failed-AVP', #{failed_avp := false} = Opts) -> set_failed(_, Opts) -> Opts. -%% acc/8 +%% acc/7 -acc([AM | Acc], As, I, Field, Arity, Strict, Mod, Opts) -> - [AM | acc1(Acc, As, I, Field, Arity, Strict, Mod, Opts)]. +acc([AM | Acc], As, I, Field, Arity, Strict, Mod) -> + [AM | acc1(Acc, As, I, Field, Arity, Strict, Mod)]. -%% acc1/8 +%% acc1/7 %% Faulty AVP, not grouped. -acc1(Acc, {_RC, Avp} = E, _, _, _, _, _, _) -> +acc1(Acc, {_RC, Avp} = E, _, _, _, _, _) -> [Avps, Failed | Rec] = Acc, [[Avp | Avps], [E | Failed] | Rec]; %% Faulty component in grouped AVP. -acc1(Acc, {RC, As, Avp}, _, _, _, _, _, _) -> +acc1(Acc, {RC, As, Avp}, _, _, _, _, _) -> [Avps, Failed | Rec] = Acc, [[As | Avps], [{RC, Avp} | Failed] | Rec]; %% Grouped AVP ... -acc1([Avps | Acc], [Avp|_] = As, I, Field, Arity, Strict, Mod, Opts) -> - [[As|Avps] | acc2(Acc, Avp, I, Field, Arity, Strict, Mod, Opts)]; +acc1([Avps | Acc], [Avp|_] = As, I, Field, Arity, Strict, Mod) -> + [[As|Avps] | acc2(Acc, Avp, I, Field, Arity, Strict, Mod)]; %% ... or not. -acc1([Avps | Acc], Avp, I, Field, Arity, Strict, Mod, Opts) -> - [[Avp|Avps] | acc2(Acc, Avp, I, Field, Arity, Strict, Mod, Opts)]. +acc1([Avps | Acc], Avp, I, Field, Arity, Strict, Mod) -> + [[Avp|Avps] | acc2(Acc, Avp, I, Field, Arity, Strict, Mod)]. %% The component list of a Grouped AVP is discarded when packing into %% the record (or equivalent): the values in an 'AVP' field are @@ -713,24 +715,24 @@ acc1([Avps | Acc], Avp, I, Field, Arity, Strict, Mod, Opts) -> %% retain the same structure as in diameter_packet.avps, but an 'AVP' %% list has always been flat. -%% acc2/8 +%% acc2/7 %% No errors, but nowhere to pack. -acc2(Acc, Avp, _, 'AVP', 0, _, _, _) -> +acc2(Acc, Avp, _, 'AVP', 0, _, _) -> [Failed | Rec] = Acc, [[{rc(Avp), Avp} | Failed] | Rec]; %% Relaxed arities. -acc2(Acc, Avp, _, Field, Arity, Strict, Mod, _) +acc2(Acc, Avp, _, Field, Arity, Strict, Mod) when Strict /= decode -> pack(Arity, Field, Avp, Mod, Acc); %% No maximum arity. -acc2(Acc, Avp, _, Field, {_,'*'} = Arity, _, Mod, _) -> +acc2(Acc, Avp, _, Field, {_,'*'} = Arity, _, Mod) -> pack(Arity, Field, Avp, Mod, Acc); %% Or check. -acc2(Acc, Avp, I, Field, Arity, _, Mod, _) -> +acc2(Acc, Avp, I, Field, Arity, _, Mod) -> Mx = max_arity(Arity), if Mx =< I -> [Failed | Rec] = Acc, diff --git a/lib/diameter/src/diameter.appup.src b/lib/diameter/src/diameter.appup.src index c2198de9ea..d0e58e8410 100644 --- a/lib/diameter/src/diameter.appup.src +++ b/lib/diameter/src/diameter.appup.src @@ -54,7 +54,9 @@ {"1.12.1", [{restart_application, diameter}]}, %% 19.1 {"1.12.2", [{restart_application, diameter}]}, %% 19.3 {"2.0", [{restart_application, diameter}]}, %% 20.0 - {"2.1", [{update, diameter_reg, {advanced, "2.1"}}]} %% 20.1 + {"2.1", [{load_module, diameter_gen}, %% 20.1 + {update, diameter_reg, {advanced, "2.1"}}]}, + {"2.1.1", [{load_module, diameter_gen}]} ], [ {"0.9", [{restart_application, diameter}]}, @@ -90,6 +92,7 @@ {"1.12.1", [{restart_application, diameter}]}, {"1.12.2", [{restart_application, diameter}]}, {"2.0", [{restart_application, diameter}]}, - {"2.1", [{restart_application, diameter}]} + {"2.1", [{restart_application, diameter}]}, + {"2.1.1", [{load_module, diameter_gen}]} ] }. diff --git a/lib/diameter/vsn.mk b/lib/diameter/vsn.mk index f73f68da0b..bfb260ed8f 100644 --- a/lib/diameter/vsn.mk +++ b/lib/diameter/vsn.mk @@ -17,5 +17,5 @@ # %CopyrightEnd% APPLICATION = diameter -DIAMETER_VSN = 2.1.1 +DIAMETER_VSN = 2.1.2 APP_VSN = $(APPLICATION)-$(DIAMETER_VSN)$(PRE_VSN) diff --git a/lib/erl_docgen/priv/dtd/erlref.dtd b/lib/erl_docgen/priv/dtd/erlref.dtd index 615b88b61a..78d6771f52 100644 --- a/lib/erl_docgen/priv/dtd/erlref.dtd +++ b/lib/erl_docgen/priv/dtd/erlref.dtd @@ -33,4 +33,5 @@ <!ATTLIST name name CDATA #IMPLIED arity CDATA #IMPLIED clause_i CDATA #IMPLIED + anchor CDATA #IMPLIED n_vars CDATA #IMPLIED> diff --git a/lib/erl_docgen/priv/xsl/db_html.xsl b/lib/erl_docgen/priv/xsl/db_html.xsl index 75614392fb..5b7eae4f73 100644 --- a/lib/erl_docgen/priv/xsl/db_html.xsl +++ b/lib/erl_docgen/priv/xsl/db_html.xsl @@ -172,6 +172,7 @@ <xsl:template name="spec_name"> <xsl:variable name="name" select="@name"/> <xsl:variable name="arity" select="@arity"/> + <xsl:variable name="anchor" select="@anchor"/> <xsl:variable name="spec0"> <xsl:call-template name="find_spec"/> </xsl:variable> @@ -198,6 +199,11 @@ </xsl:otherwise> </xsl:choose> + <!-- Insert an anchor for "anchor" attribute --> + <xsl:if test="string-length($anchor) > 0"> + <a name="{$anchor}"></a> + </xsl:if> + <xsl:variable name="global_types" select="ancestor::erlref/datatypes"/> <xsl:variable name="local_types" select="../type[string-length(@name) > 0]"/> diff --git a/lib/hipe/doc/src/hipe_app.xml b/lib/hipe/doc/src/hipe_app.xml index e489d155c3..9299c6d73f 100644 --- a/lib/hipe/doc/src/hipe_app.xml +++ b/lib/hipe/doc/src/hipe_app.xml @@ -47,6 +47,72 @@ Details on HiPE compiler options are given by <c>hipe:help_options()</c>.</p> </description> <section> + <title>Feature Limitations</title> + <p> + The HiPE compiler is in general compliant with the normal BEAM compiler, + with respect to semantic behavior. There are however features in the BEAM compiler + and the runtime system that have limited or no support for HiPE compiled modules. + </p> + <taglist> + <tag>Stack traces</tag> + <item><p>Stack traces returned from <seealso marker="erts:erlang#get_stacktrace/0"> + <c>erlang:get_stacktrace/0</c></seealso> or as part of <c>'EXIT'</c> terms + can look incomplete if HiPE compiled functions are involved. Typically a stack trace + will contain only BEAM compiled functions or only HiPE compiled functions, depending + on where the exception was raised.</p> + <p>Source code line numbers in stack traces are also not supported by HiPE compiled functions.</p> + </item> + + <tag>Tracing</tag> + <item><p>Erlang call trace is not supported by HiPE. Calling + <seealso marker="erts:erlang#trace_pattern/3"><c>erlang:trace_pattern({M,F,A}, ...)</c></seealso> + does not have any effect on HiPE compiled modules.</p> + </item> + + <tag>NIFs</tag> + <item><p>Modules compiled with HiPE can not call <seealso marker="erts:erlang#load_nif-2"> + <c>erlang:load_nif/2</c></seealso> to load NIFs.</p> + </item> + + <tag>-on_load</tag> + <item><p>Modules compiled with HiPE can not use + <seealso marker="doc/reference_manual:code_loading#on_load"><c>-on_load()</c></seealso> + directives.</p> + </item> + </taglist> + + </section> + <section> + <title>Performance Limitations</title> + <p> + The HiPE compiler does in general produce faster code than the + BEAM compiler. There are however some situation when HiPE + compiled code will perform worse than BEAM code. + </p> + <taglist> + <tag>Mode switches</tag> + <item><p>Every time a process changes from executing code in a + HiPE compiled module to a BEAM compiled module (or vice versa), + it will do a mode switch. This involves a certain amount of + CPU overhead which can have a negative net impact if the + process is switching back and forth without getting enough done in + each mode.</p> + </item> + + <tag>Optimization for <c>receive</c> with unique references</tag> + <item><p>The BEAM compiler can do an optimization when + a <c>receive</c> statement is <em>only</em> waiting for messages + containing a reference created before the receive. All messages + that existed in the queue when the reference was created will be + bypassed, as they cannot possibly contain the reference. HiPE + does not implement this optimization.</p> + <p>An example of this is when + <c>gen_server:call()</c> waits for the reply message.</p> + </item> + + </taglist> + </section> + <section> <title>SEE ALSO</title> <p> <seealso marker="stdlib:c">c(3)</seealso>, diff --git a/lib/hipe/icode/hipe_beam_to_icode.erl b/lib/hipe/icode/hipe_beam_to_icode.erl index 167f5c67bb..cb4c1cfbd3 100644 --- a/lib/hipe/icode/hipe_beam_to_icode.erl +++ b/lib/hipe/icode/hipe_beam_to_icode.erl @@ -794,7 +794,7 @@ trans_fun([{bs_append,{f,Lbl},Size,W,R,U,Binary,{field_flags,F},Dst}| SizeArg = trans_arg(Size), BinArg = trans_arg(Binary), IcodeDst = mk_var(Dst), - Offset = mk_var(reg), + Offset = mk_var(reg_gcsafe), Base = mk_var(reg), trans_bin_call({hipe_bs_primop,{bs_append,W,R,U,F}},Lbl,[SizeArg,BinArg], [IcodeDst,Base,Offset], @@ -805,7 +805,7 @@ trans_fun([{bs_private_append,{f,Lbl},Size,U,Binary,{field_flags,F},Dst}| SizeArg = trans_arg(Size), BinArg = trans_arg(Binary), IcodeDst = mk_var(Dst), - Offset = mk_var(reg), + Offset = mk_var(reg_gcsafe), Base = mk_var(reg), trans_bin_call({hipe_bs_primop,{bs_private_append,U,F}}, Lbl,[SizeArg,BinArg], @@ -844,7 +844,7 @@ trans_fun([{bs_init2,{f,Lbl},Size,_Words,_LiveRegs,{field_flags,Flags0},X}| Instructions], Env) -> Dst = mk_var(X), Flags = resolve_native_endianess(Flags0), - Offset = mk_var(reg), + Offset = mk_var(reg_gcsafe), Base = mk_var(reg), {Name, Args} = case Size of @@ -860,7 +860,7 @@ trans_fun([{bs_init_bits,{f,Lbl},Size,_Words,_LiveRegs,{field_flags,Flags0},X}| Instructions], Env) -> Dst = mk_var(X), Flags = resolve_native_endianess(Flags0), - Offset = mk_var(reg), + Offset = mk_var(reg_gcsafe), Base = mk_var(reg), {Name, Args} = case Size of @@ -1505,7 +1505,10 @@ clone_dst(Dest) -> New = case hipe_icode:is_reg(Dest) of true -> - mk_var(reg); + case hipe_icode:reg_is_gcsafe(Dest) of + true -> mk_var(reg_gcsafe); + false -> mk_var(reg) + end; false -> true = hipe_icode:is_var(Dest), mk_var(new) @@ -2126,7 +2129,12 @@ mk_var(reg) -> T = hipe_gensym:new_var(icode), V = (5*T)+4, hipe_gensym:update_vrange(icode,V), - hipe_icode:mk_reg(V). + hipe_icode:mk_reg(V); +mk_var(reg_gcsafe) -> + T = hipe_gensym:new_var(icode), + V = (5*T)+4, % same namespace as 'reg' + hipe_gensym:update_vrange(icode,V), + hipe_icode:mk_reg_gcsafe(V). %%----------------------------------------------------------------------- %% Make an icode label of proper type diff --git a/lib/hipe/icode/hipe_icode.erl b/lib/hipe/icode/hipe_icode.erl index 24b7ac4783..bc3403b0c5 100644 --- a/lib/hipe/icode/hipe_icode.erl +++ b/lib/hipe/icode/hipe_icode.erl @@ -515,10 +515,12 @@ annotate_variable/2, %% annotate_var_or_reg(VarOrReg, Type) unannotate_variable/1,%% unannotate_var_or_reg(VarOrReg) mk_reg/1, %% mk_reg(Id) + mk_reg_gcsafe/1, %% mk_reg_gcsafe(Id) mk_fvar/1, %% mk_fvar(Id) mk_new_var/0, %% mk_new_var() mk_new_fvar/0, %% mk_new_fvar() mk_new_reg/0, %% mk_new_reg() + mk_new_reg_gcsafe/0, %% mk_new_reg_gcsafe() mk_phi/1, %% mk_phi(Id) mk_phi/2 %% mk_phi(Id, ArgList) ]). @@ -1260,14 +1262,22 @@ is_var(_) -> false. -spec mk_reg(non_neg_integer()) -> #icode_variable{kind::'reg'}. mk_reg(V) -> #icode_variable{name=V, kind=reg}. --spec reg_name(#icode_variable{kind::'reg'}) -> non_neg_integer(). -reg_name(#icode_variable{name=Name, kind=reg}) -> Name. +-spec mk_reg_gcsafe(non_neg_integer()) -> #icode_variable{kind::'reg_gcsafe'}. +mk_reg_gcsafe(V) -> #icode_variable{name=V, kind=reg_gcsafe}. --spec reg_is_gcsafe(#icode_variable{kind::'reg'}) -> 'false'. -reg_is_gcsafe(#icode_variable{kind=reg}) -> false. % for now +-spec reg_name(#icode_variable{kind::'reg'|'reg_gcsafe'}) + -> non_neg_integer(). +reg_name(#icode_variable{name=Name, kind=reg}) -> Name; +reg_name(#icode_variable{name=Name, kind=reg_gcsafe}) -> Name. + +-spec reg_is_gcsafe(#icode_variable{kind::'reg'}) -> 'false'; + (#icode_variable{kind::'reg_gcsafe'}) -> 'true'. +reg_is_gcsafe(#icode_variable{kind=reg}) -> false; +reg_is_gcsafe(#icode_variable{kind=reg_gcsafe}) -> true. -spec is_reg(icode_argument()) -> boolean(). -is_reg(#icode_variable{kind=reg}) -> true; +is_reg(#icode_variable{kind=reg}) -> true; +is_reg(#icode_variable{kind=reg_gcsafe}) -> true; is_reg(_) -> false. -spec mk_fvar(non_neg_integer()) -> #icode_variable{kind::'fvar'}. @@ -1676,6 +1686,16 @@ mk_new_reg() -> mk_reg(hipe_gensym:get_next_var(icode)). %% +%% @doc Makes a new gcsafe register; that is, a register that is allowed to be +%% live over calls and other operations that might cause GCs and thus move heap +%% data around. +%% + +-spec mk_new_reg_gcsafe() -> icode_reg(). +mk_new_reg_gcsafe() -> + mk_reg_gcsafe(hipe_gensym:get_next_var(icode)). + +%% %% @doc Makes a new label. %% diff --git a/lib/hipe/icode/hipe_icode.hrl b/lib/hipe/icode/hipe_icode.hrl index 380ddd8371..7ed80a9ed4 100644 --- a/lib/hipe/icode/hipe_icode.hrl +++ b/lib/hipe/icode/hipe_icode.hrl @@ -41,9 +41,9 @@ -type variable_annotation() :: {atom(), any(), fun((any()) -> string())}. --record(icode_variable, {name :: non_neg_integer(), - kind :: 'var' | 'reg' | 'fvar', - annotation = [] :: [] | variable_annotation()}). +-record(icode_variable, {name :: non_neg_integer(), + kind :: 'var' | 'reg' | 'reg_gcsafe' | 'fvar', + annotation = [] :: [] | variable_annotation()}). %%--------------------------------------------------------------------- %% Type declarations for Icode instructions @@ -66,7 +66,7 @@ -type icode_funcall() :: mfa() | icode_primop(). -type icode_var() :: #icode_variable{kind::'var'}. --type icode_reg() :: #icode_variable{kind::'reg'}. +-type icode_reg() :: #icode_variable{kind::'reg'|'reg_gcsafe'}. -type icode_fvar() :: #icode_variable{kind::'fvar'}. -type icode_argument() :: #icode_const{} | #icode_variable{}. -type icode_term_arg() :: icode_var() | #icode_const{}. diff --git a/lib/hipe/icode/hipe_icode_liveness.erl b/lib/hipe/icode/hipe_icode_liveness.erl index 51e2855108..e61529a1bb 100644 --- a/lib/hipe/icode/hipe_icode_liveness.erl +++ b/lib/hipe/icode/hipe_icode_liveness.erl @@ -77,6 +77,7 @@ print_var(#icode_variable{name=V, kind=Kind, annotation=T}) -> case Kind of var -> io:format("v~p", [V]); reg -> io:format("r~p", [V]); + reg_gcsafe -> io:format("rs~p", [V]); fvar -> io:format("fv~p", [V]) end, case T of diff --git a/lib/hipe/icode/hipe_icode_pp.erl b/lib/hipe/icode/hipe_icode_pp.erl index 5b017dca32..33d1e62884 100644 --- a/lib/hipe/icode/hipe_icode_pp.erl +++ b/lib/hipe/icode/hipe_icode_pp.erl @@ -230,7 +230,10 @@ pp_arg(Dev, Arg) -> case hipe_icode:is_reg(Arg) of true -> N = hipe_icode:reg_name(Arg), - io:format(Dev, "r~p", [N]); + case hipe_icode:reg_is_gcsafe(Arg) of + true -> io:format(Dev, "rs~p", [N]); + false -> io:format(Dev, "r~p", [N]) + end; false -> N = hipe_icode:fvar_name(Arg), io:format(Dev, "fv~p", [N]) diff --git a/lib/hipe/main/hipe.app.src b/lib/hipe/main/hipe.app.src index 5b2280594f..66008a4178 100644 --- a/lib/hipe/main/hipe.app.src +++ b/lib/hipe/main/hipe.app.src @@ -179,6 +179,7 @@ hipe_rtl_to_sparc, hipe_rtl_to_x86, hipe_rtl_varmap, + hipe_rtl_verify_gcsafe, hipe_segment_trees, hipe_sdi, hipe_sparc, @@ -236,4 +237,4 @@ {applications, [kernel,stdlib]}, {env, []}, {runtime_dependencies, ["syntax_tools-1.6.14","stdlib-3.4","kernel-5.3", - "erts-9.0","compiler-5.0"]}]}. + "erts-9.2","compiler-5.0"]}]}. diff --git a/lib/hipe/main/hipe.erl b/lib/hipe/main/hipe.erl index f5f5bf5830..acb9b7b062 100644 --- a/lib/hipe/main/hipe.erl +++ b/lib/hipe/main/hipe.erl @@ -1414,6 +1414,7 @@ opt_keys() -> use_clusters, use_jumptable, verbose, + verify_gcsafe, %% verbose_spills, x87]. @@ -1510,7 +1511,8 @@ opt_negations() -> {no_use_callgraph, use_callgraph}, {no_use_clusters, use_clusters}, {no_use_inline_atom_search, use_inline_atom_search}, - {no_use_indexing, use_indexing}]. + {no_use_indexing, use_indexing}, + {no_verify_gcsafe, verify_gcsafe}]. %% Don't use negative forms in right-hand sides of aliases and expansions! %% We only expand negations once, before the other expansions are done. diff --git a/lib/hipe/main/hipe_main.erl b/lib/hipe/main/hipe_main.erl index dca6fddec3..4b5eb4c63e 100644 --- a/lib/hipe/main/hipe_main.erl +++ b/lib/hipe/main/hipe_main.erl @@ -410,6 +410,11 @@ icode_to_rtl(MFA, Icode, Options, Servers) -> hipe_llvm_liveness:analyze(RtlCfg4) end, pp(RtlCfg5, MFA, rtl, pp_rtl, Options, Servers), + case proplists:get_bool(verify_gcsafe, Options) of + false -> ok; + true -> + ok = hipe_rtl_verify_gcsafe:check(RtlCfg5) + end, LinearRTL1 = hipe_rtl_cfg:linearize(RtlCfg5), LinearRTL2 = hipe_rtl_cleanup_const:cleanup(LinearRTL1), %% hipe_rtl:pp(standard_io, LinearRTL2), diff --git a/lib/hipe/rtl/Makefile b/lib/hipe/rtl/Makefile index 5abc9ec049..becdd0b7d8 100644 --- a/lib/hipe/rtl/Makefile +++ b/lib/hipe/rtl/Makefile @@ -50,7 +50,7 @@ HIPE_MODULES = hipe_rtl hipe_rtl_cfg \ hipe_rtl_ssa hipe_rtl_ssa_const_prop \ hipe_rtl_cleanup_const hipe_rtl_symbolic hipe_rtl_lcm \ hipe_rtl_ssapre hipe_rtl_binary hipe_rtl_ssa_avail_expr \ - hipe_rtl_arch hipe_tagscheme + hipe_rtl_arch hipe_tagscheme hipe_rtl_verify_gcsafe else HIPE_MODULES = endif diff --git a/lib/hipe/rtl/hipe_rtl.erl b/lib/hipe/rtl/hipe_rtl.erl index 04c9728d5c..33027f3259 100644 --- a/lib/hipe/rtl/hipe_rtl.erl +++ b/lib/hipe/rtl/hipe_rtl.erl @@ -1740,7 +1740,10 @@ pp_reg(Dev, Arg) -> true -> pp_hard_reg(Dev, reg_index(Arg)); false -> - io:format(Dev, "r~w", [reg_index(Arg)]) + case reg_is_gcsafe(Arg) of + true -> io:format(Dev, "rs~w", [reg_index(Arg)]); + false -> io:format(Dev, "r~w", [reg_index(Arg)]) + end end. pp_var(Dev, Arg) -> diff --git a/lib/hipe/rtl/hipe_rtl_binary_construct.erl b/lib/hipe/rtl/hipe_rtl_binary_construct.erl index 52ea5db382..5b89d4946a 100644 --- a/lib/hipe/rtl/hipe_rtl_binary_construct.erl +++ b/lib/hipe/rtl/hipe_rtl_binary_construct.erl @@ -195,8 +195,13 @@ gen_rtl(BsOP, Dst, Args, TrueLblName, FalseLblName, SystemLimitLblName, ConstTab bs_validate_unicode -> [_Arg] = Args, - [hipe_rtl:mk_call([], bs_validate_unicode, Args, - TrueLblName, FalseLblName, not_remote)]; + [IsUnicode] = create_regs(1), + RetLbl = hipe_rtl:mk_new_label(), + [hipe_rtl:mk_call([IsUnicode], is_unicode, Args, + hipe_rtl:label_name(RetLbl), [], not_remote), + RetLbl, + hipe_rtl:mk_branch(IsUnicode, ne, hipe_rtl:mk_imm(0), + TrueLblName, FalseLblName, 0.99)]; bs_final -> Zero = hipe_rtl:mk_imm(0), @@ -354,7 +359,8 @@ not_writable_code(Bin, SizeReg, Dst, Base, Offset, Unit, allocate_writable(Dst, Base, UsedBytes, TotBytes, TotSize) -> Zero = hipe_rtl:mk_imm(0), [NextLbl] = create_lbls(1), - [EndSubSize, EndSubBitSize, ProcBin] = create_regs(3), + [EndSubSize, EndSubBitSize] = create_regs(2), + [ProcBin] = create_unsafe_regs(1), [hipe_rtl:mk_call([Base], bs_allocate, [UsedBytes], hipe_rtl:label_name(NextLbl), [], not_remote), NextLbl, @@ -581,12 +587,12 @@ const_init2(Size, Dst, Base, Offset, TrueLblName) -> false -> ByteSize = hipe_rtl:mk_new_reg(), [hipe_rtl:mk_gctest(?PROC_BIN_WORDSIZE+?SUB_BIN_WORDSIZE), - hipe_rtl:mk_move(Offset, hipe_rtl:mk_imm(0)), hipe_rtl:mk_move(ByteSize, hipe_rtl:mk_imm(Size)), hipe_rtl:mk_call([Base], bs_allocate, [ByteSize], hipe_rtl:label_name(NextLbl), [], not_remote), NextLbl, hipe_tagscheme:create_refc_binary(Base, ByteSize, Dst), + hipe_rtl:mk_move(Offset, hipe_rtl:mk_imm(0)), hipe_rtl:mk_goto(TrueLblName)] end. @@ -629,13 +635,12 @@ var_init2(Size, Dst, Base, Offset, TrueLblName, SystemLimitLblName, FalseLblName Log2WordSize = hipe_rtl_arch:log2_word_size(), WordSize = hipe_rtl_arch:word_size(), [ContLbl, HeapLbl, REFCLbl, NextLbl] = create_lbls(4), - [USize, Tmp] = create_unsafe_regs(2), + [USize, Tmp] = create_regs(2), [get_word_integer(Size, USize, SystemLimitLblName, FalseLblName), hipe_rtl:mk_branch(USize, leu, hipe_rtl:mk_imm(?MAX_BINSIZE), hipe_rtl:label_name(ContLbl), SystemLimitLblName), ContLbl, - hipe_rtl:mk_move(Offset, hipe_rtl:mk_imm(0)), hipe_rtl:mk_branch(USize, leu, hipe_rtl:mk_imm(?MAX_HEAP_BIN_SIZE), hipe_rtl:label_name(HeapLbl), hipe_rtl:label_name(REFCLbl)), @@ -645,6 +650,7 @@ var_init2(Size, Dst, Base, Offset, TrueLblName, SystemLimitLblName, FalseLblName hipe_rtl:mk_alu(Tmp, Tmp, add, hipe_rtl:mk_imm(?SUB_BIN_WORDSIZE)), hipe_rtl:mk_gctest(Tmp), hipe_tagscheme:create_heap_binary(Base, USize, Dst), + hipe_rtl:mk_move(Offset, hipe_rtl:mk_imm(0)), hipe_rtl:mk_goto(TrueLblName), REFCLbl, hipe_rtl:mk_gctest(?PROC_BIN_WORDSIZE+?SUB_BIN_WORDSIZE), @@ -652,6 +658,7 @@ var_init2(Size, Dst, Base, Offset, TrueLblName, SystemLimitLblName, FalseLblName hipe_rtl:label_name(NextLbl), [], not_remote), NextLbl, hipe_tagscheme:create_refc_binary(Base, USize, Dst), + hipe_rtl:mk_move(Offset, hipe_rtl:mk_imm(0)), hipe_rtl:mk_goto(TrueLblName)]. var_init_bits(Size, Dst, Base, Offset, TrueLblName, SystemLimitLblName, FalseLblName) -> @@ -858,7 +865,7 @@ get_base_offset_size(Binary, SrcBase, SrcOffset, SrcSize, FLName) -> JoinLbl, hipe_tagscheme:test_heap_binary(Orig, HeapLblName, REFCLblName), HeapLbl, - hipe_rtl:mk_alu(SrcBase, Orig, add, hipe_rtl:mk_imm(?HEAP_BIN_DATA-2)), + hipe_tagscheme:get_field_addr_from_term({heap_bin, {data, 0}}, Orig, SrcBase), hipe_rtl:mk_goto(EndLblName), REFCLbl, hipe_tagscheme:get_field_from_term({proc_bin,bytes}, Orig, SrcBase), @@ -1205,6 +1212,12 @@ is_divisible(Dividend, Divisor, SuccLbl, FailLbl) -> [hipe_rtl:mk_branch(Dividend, 'and', Mask, eq, SuccLbl, FailLbl, 0.99)]; false -> %% We need division, fall back to a primop - [hipe_rtl:mk_call([], is_divisible, [Dividend, hipe_rtl:mk_imm(Divisor)], - SuccLbl, FailLbl, not_remote)] + [Tmp] = create_regs(1), + RetLbl = hipe_rtl:mk_new_label(), + [hipe_rtl:mk_call([Tmp], is_divisible, + [Dividend, hipe_rtl:mk_imm(Divisor)], + hipe_rtl:label_name(RetLbl), [], not_remote), + RetLbl, + hipe_rtl:mk_branch(Tmp, ne, hipe_rtl:mk_imm(0), + SuccLbl, FailLbl, 0.99)] end. diff --git a/lib/hipe/rtl/hipe_rtl_binary_match.erl b/lib/hipe/rtl/hipe_rtl_binary_match.erl index 362a52f8fe..4575213838 100644 --- a/lib/hipe/rtl/hipe_rtl_binary_match.erl +++ b/lib/hipe/rtl/hipe_rtl_binary_match.erl @@ -730,7 +730,7 @@ get_base(Orig,Base) -> [hipe_tagscheme:test_heap_binary(Orig, hipe_rtl:label_name(HeapLbl), hipe_rtl:label_name(REFCLbl)), HeapLbl, - hipe_rtl:mk_alu(Base, Orig, 'add', hipe_rtl:mk_imm(?HEAP_BIN_DATA-2)), + hipe_tagscheme:get_field_addr_from_term({heap_bin, {data, 0}}, Orig, Base), hipe_rtl:mk_goto(hipe_rtl:label_name(EndLbl)), REFCLbl, get_field_from_term({proc_bin, flags}, Orig, Flags), @@ -740,7 +740,7 @@ get_base(Orig,Base) -> WritableLbl, hipe_rtl:mk_call([], emasculate_binary, [Orig], [], [], 'not_remote'), NotWritableLbl, - hipe_rtl:mk_load(Base, Orig, hipe_rtl:mk_imm(?PROC_BIN_BYTES-2)), + get_field_from_term({proc_bin, bytes}, Orig, Base), EndLbl]. extract_matchstate_var(binsize, Ms) -> @@ -842,12 +842,12 @@ make_dyn_prep(SizeReg, CCode) -> %%------------------------------------------------------------------------ get_unaligned_int(Dst1, Size, Base, Offset, Shiftr, Type, TrueLblName) -> - [Reg] = create_regs(1), + [Reg] = create_gcsafe_regs(1), [get_maybe_unaligned_int_to_reg(Reg, Size, Base, Offset, Shiftr, Type), do_bignum_code(Size, Type, Reg, Dst1, TrueLblName)]. get_maybe_unaligned_int_to_reg(Reg, Size, Base, Offset, Shiftr, Type) -> - [LowBits] = create_regs(1), + [LowBits] = create_gcsafe_regs(1), [AlignedLbl, UnAlignedLbl, EndLbl] = create_lbls(3), [hipe_rtl:mk_alub(LowBits, Offset, 'and', hipe_rtl:mk_imm(?LOW_BITS), eq, hipe_rtl:label_name(AlignedLbl), @@ -1001,7 +1001,7 @@ do_bignum_code(Size, {Signedness,_}, Src, Dst1, TrueLblName) end. signed_bignum(Dst1, Src, TrueLblName) -> - Tmp1 = hipe_rtl:mk_new_reg(), + Tmp1 = hipe_rtl:mk_new_reg_gcsafe(), BignumLabel = hipe_rtl:mk_new_label(), [hipe_tagscheme:realtag_fixnum(Dst1, Src), hipe_tagscheme:realuntag_fixnum(Tmp1, Dst1), diff --git a/lib/hipe/rtl/hipe_rtl_cleanup_const.erl b/lib/hipe/rtl/hipe_rtl_cleanup_const.erl index bfa6b9682e..00cc2bcb37 100644 --- a/lib/hipe/rtl/hipe_rtl_cleanup_const.erl +++ b/lib/hipe/rtl/hipe_rtl_cleanup_const.erl @@ -69,9 +69,9 @@ cleanup_instr([Const|Left], I, Acc) -> case I of X when is_record(X, fp_unop) orelse is_record(X, fp) -> Fdst = hipe_rtl:mk_new_fpreg(), - Fconv = hipe_tagscheme:unsafe_untag_float(Fdst, Dst), + Fconv = lists:flatten(hipe_tagscheme:unsafe_untag_float(Fdst, Dst)), NewI = hipe_rtl:subst_uses([{Const, Fdst}], I), - cleanup_instr(Left, NewI, Fconv ++ [Load|Acc]); + cleanup_instr(Left, NewI, lists:reverse(Fconv, [Load|Acc])); _ -> NewI = hipe_rtl:subst_uses([{Const, Dst}], I), cleanup_instr(Left, NewI, [Load|Acc]) diff --git a/lib/hipe/rtl/hipe_rtl_lcm.erl b/lib/hipe/rtl/hipe_rtl_lcm.erl index 9dcdd05fb1..af39c9a0a4 100644 --- a/lib/hipe/rtl/hipe_rtl_lcm.erl +++ b/lib/hipe/rtl/hipe_rtl_lcm.erl @@ -182,42 +182,41 @@ delete_exprs(Code, _, _, []) -> Code; delete_exprs(Code, ExprMap, IdMap, [ExprId|Exprs]) -> Expr = expr_id_map_get_expr(IdMap, ExprId), - %% Perform a foldl that goes through the code and deletes all - %% occurences of the expression. - NewCode = - lists:reverse - (lists:foldl(fun(CodeExpr, Acc) -> - case is_expr(CodeExpr) of - true -> - case expr_clear_dst(CodeExpr) =:= Expr of - true -> - pp_debug(" Deleting: ", []), - pp_debug_instr(CodeExpr), - %% Lookup expression entry. - Defines = - case expr_map_lookup(ExprMap, Expr) of - {value, {_, _, Defs}} -> - Defs; - none -> - exit({?MODULE, expr_map_lookup, - "expression missing"}) - end, - MoveCode = - mk_expr_move_instr(hipe_rtl:defines(CodeExpr), - Defines), - pp_debug(" Replacing with: ", []), - pp_debug_instr(MoveCode), - [MoveCode|Acc]; - false -> - [CodeExpr|Acc] - end; - false -> - [CodeExpr|Acc] - end - end, - [], Code)), + %% Lookup expression entry. + {value, {_, _, Defines}} = expr_map_lookup(ExprMap, Expr), + %% Go through the code and deletes all occurences of the expression. + NewCode = delete_expr(Code, Expr, Defines, []), delete_exprs(NewCode, ExprMap, IdMap, Exprs). +delete_expr([], _Expr, _Defines, Acc) -> lists:reverse(Acc); +delete_expr([CodeExpr|Code], Expr, Defines, Acc) -> + case exp_kill_expr(CodeExpr, [Expr]) of + [] -> % Expr was killed; deleting stops here + pp_debug(" Stopping before: ", []), + pp_debug_instr(CodeExpr), + lists:reverse(Acc, [CodeExpr|Code]); + [Expr] -> + NewCodeExpr = + case is_expr(CodeExpr) of + true -> + case expr_clear_dst(CodeExpr) =:= Expr of + true -> + pp_debug(" Deleting: ", []), + pp_debug_instr(CodeExpr), + MoveCode = mk_expr_move_instr(hipe_rtl:defines(CodeExpr), + Defines), + pp_debug(" Replacing with: ", []), + pp_debug_instr(MoveCode), + MoveCode; + false -> + CodeExpr + end; + false -> + CodeExpr + end, + delete_expr(Code, Expr, Defines, [NewCodeExpr|Acc]) + end. + %%============================================================================= %% Goes through the given list of expressions and inserts them at %% appropriate places in the code. @@ -226,13 +225,12 @@ insert_exprs(CFG, _, _, _, _, BetweenMap, []) -> insert_exprs(CFG, Pred, Succ, ExprMap, IdMap, BetweenMap, [ExprId|Exprs]) -> Expr = expr_id_map_get_expr(IdMap, ExprId), Instr = expr_map_get_instr(ExprMap, Expr), - case hipe_rtl_cfg:succ(CFG, Pred) of - [_] -> + case try_insert_expr_last(CFG, Pred, Instr) of + {ok, NewCFG} -> pp_debug(" Inserted last: ", []), pp_debug_instr(Instr), - NewCFG = insert_expr_last(CFG, Pred, Instr), insert_exprs(NewCFG, Pred, Succ, ExprMap, IdMap, BetweenMap, Exprs); - _ -> + not_safe -> case hipe_rtl_cfg:pred(CFG, Succ) of [_] -> pp_debug(" Inserted first: ", []), @@ -252,25 +250,31 @@ insert_exprs(CFG, Pred, Succ, ExprMap, IdMap, BetweenMap, [ExprId|Exprs]) -> %% Recursively goes through the code in a block and returns a new block %% with the new code inserted second to last (assuming the last expression %% is a branch operation). -insert_expr_last(CFG0, Label, Instr) -> - Code0 = hipe_bb:code(hipe_rtl_cfg:bb(CFG0, Label)), - %% FIXME: Use hipe_bb:butlast() instead? - Code1 = insert_expr_last_work(Label, Instr, Code0), - hipe_rtl_cfg:bb_add(CFG0, Label, hipe_bb:mk_bb(Code1)). +try_insert_expr_last(CFG0, Label, Instr) -> + case hipe_rtl_cfg:succ(CFG0, Label) of + [_] -> + Code0 = hipe_bb:code(hipe_rtl_cfg:bb(CFG0, Label)), + case insert_expr_last_work(Instr, Code0) of + not_safe -> not_safe; + Code1 -> + {ok, hipe_rtl_cfg:bb_add(CFG0, Label, hipe_bb:mk_bb(Code1))} + end; + _ -> not_safe + end. %%============================================================================= %% Recursively goes through the code in a block and returns a new block %% with the new code inserted second to last (assuming the last expression %% is a branch operation). -insert_expr_last_work(_, Instr, []) -> - %% This case should not happen since this means that block was completely - %% empty when the function was called. For compatibility we insert it last. - [Instr]; -insert_expr_last_work(_, Instr, [Code1]) -> +insert_expr_last_work(_Instr, [#call{}]) -> + %% Call instructions clobber all expressions; we musn't insert the expression + %% before it + not_safe; +insert_expr_last_work(Instr, [Code1]) -> %% We insert the code next to last. [Instr, Code1]; -insert_expr_last_work(Label, Instr, [Code|Codes]) -> - [Code|insert_expr_last_work(Label, Instr, Codes)]. +insert_expr_last_work(Instr, [Code|Codes]) -> + [Code|insert_expr_last_work(Instr, Codes)]. %%============================================================================= %% Inserts expression first in the block for the given label. @@ -305,7 +309,8 @@ insert_expr_between(CFG0, BetweenMap, Pred, Succ, Instr) -> {value, Label} -> pp_debug(" Using existing new bb for edge (~w,~w) with label ~w~n", [Pred, Succ, Label]), - {insert_expr_last(CFG0, Label, Instr), BetweenMap} + {ok, NewCfg} = try_insert_expr_last(CFG0, Label, Instr), + {NewCfg, BetweenMap} end. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% diff --git a/lib/hipe/rtl/hipe_rtl_varmap.erl b/lib/hipe/rtl/hipe_rtl_varmap.erl index 375a8f85c0..f34c66ab85 100644 --- a/lib/hipe/rtl/hipe_rtl_varmap.erl +++ b/lib/hipe/rtl/hipe_rtl_varmap.erl @@ -105,7 +105,7 @@ icode_var2rtl_var(Var, Map) -> {reg, IsGcSafe} -> NewVar = case IsGcSafe of - %% true -> hipe_rtl:mk_new_reg_gcsafe(); + true -> hipe_rtl:mk_new_reg_gcsafe(); false -> hipe_rtl:mk_new_reg() end, {NewVar, insert(Var, NewVar, Map)} diff --git a/lib/hipe/rtl/hipe_rtl_verify_gcsafe.erl b/lib/hipe/rtl/hipe_rtl_verify_gcsafe.erl new file mode 100644 index 0000000000..c3f20bfec1 --- /dev/null +++ b/lib/hipe/rtl/hipe_rtl_verify_gcsafe.erl @@ -0,0 +1,88 @@ +%% -*- mode: erlang; erlang-indent-level: 2 -*- +%% +%% Licensed under the Apache License, Version 2.0 (the "License"); +%% you may not use this file except in compliance with the License. +%% You may obtain a copy of the License at +%% +%% http://www.apache.org/licenses/LICENSE-2.0 +%% +%% Unless required by applicable law or agreed to in writing, software +%% distributed under the License is distributed on an "AS IS" BASIS, +%% WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +%% See the License for the specific language governing permissions and +%% limitations under the License. +%% +-module(hipe_rtl_verify_gcsafe). + +-export([check/1]). + +-include("../flow/cfg.hrl"). %% needed for the specs +-include("hipe_rtl.hrl"). + +check(CFG) -> + Liveness = hipe_rtl_liveness:analyze(CFG), + put({?MODULE, 'fun'}, CFG#cfg.info#cfg_info.'fun'), + lists:foreach( + fun(Lb) -> + put({?MODULE, label}, Lb), + Liveout = hipe_rtl_liveness:liveout(Liveness, Lb), + BB = hipe_rtl_cfg:bb(CFG, Lb), + check_instrs(lists:reverse(hipe_bb:code(BB)), Liveout) + end, hipe_rtl_cfg:labels(CFG)), + erase({?MODULE, 'fun'}), + erase({?MODULE, label}), + erase({?MODULE, instr}), + ok. + +check_instrs([], _Livein) -> ok; +check_instrs([I|Is], LiveOut) -> + Def = ordsets:from_list(hipe_rtl:defines(I)), + Use = ordsets:from_list(hipe_rtl:uses(I)), + LiveOver = ordsets:subtract(LiveOut, Def), + LiveIn = ordsets:union(LiveOver, Use), + case (hipe_rtl:is_call(I) + andalso not safe_primop(hipe_rtl:call_fun(I))) + orelse is_record(I, gctest) + of + false -> ok; + true -> + put({?MODULE, instr}, I), + lists:foreach(fun verify_live/1, LiveOver) + end, + check_instrs(Is, LiveIn). + +verify_live(T) -> + case hipe_rtl:is_reg(T) of + false -> ok; + true -> + case hipe_rtl:reg_is_gcsafe(T) of + true -> ok; + false -> + error({gcunsafe_live_over_call, + get({?MODULE, 'fun'}), + {label, get({?MODULE, label})}, + get({?MODULE, instr}), + T}) + end + end. + +%% Primops that can't gc +%% Note: This information is essentially duplicated from hipe_bif_list.m4 +safe_primop(is_divisible) -> true; +safe_primop(is_unicode) -> true; +safe_primop(cmp_2) -> true; +safe_primop(eq_2) -> true; +safe_primop(bs_allocate) -> true; +safe_primop(bs_reallocate) -> true; +safe_primop(bs_utf8_size) -> true; +safe_primop(bs_get_utf8) -> true; +safe_primop(bs_utf16_size) -> true; +safe_primop(bs_get_utf16) -> true; +safe_primop(bs_validate_unicode_retract) -> true; +safe_primop(bs_put_small_float) -> true; +safe_primop(bs_put_bits) -> true; +safe_primop(emasculate_binary) -> true; +safe_primop(atomic_inc) -> true; +%% Not noproc but manually verified +safe_primop(bs_put_big_integer) -> true; +safe_primop(_) -> false. diff --git a/lib/hipe/rtl/hipe_tagscheme.erl b/lib/hipe/rtl/hipe_tagscheme.erl index 68cbe75e85..737f0ec5e3 100644 --- a/lib/hipe/rtl/hipe_tagscheme.erl +++ b/lib/hipe/rtl/hipe_tagscheme.erl @@ -53,7 +53,8 @@ -export([test_subbinary/3, test_heap_binary/3]). -export([create_heap_binary/3, create_refc_binary/3, create_refc_binary/4]). -export([create_matchstate/6, convert_matchstate/1, compare_matchstate/4]). --export([get_field_from_term/3, get_field_from_pointer/3, +-export([get_field_addr_from_term/3, + get_field_from_term/3, get_field_from_pointer/3, set_field_from_term/3, set_field_from_pointer/3, extract_matchbuffer/2, extract_binary_bytes/2]). @@ -76,6 +77,10 @@ -define(TAG_PRIMARY_BOXED, 16#2). -define(TAG_PRIMARY_IMMED1, 16#3). +%% Only when ?ERTS_USE_LITERAL_TAG =:= 1 +-define(TAG_PTR_MASK__, 16#7). +-define(TAG_LITERAL_PTR, 16#4). + -define(TAG_IMMED1_SIZE, 4). -define(TAG_IMMED1_MASK, 16#F). -define(TAG_IMMED1_PID, ((16#0 bsl ?TAG_PRIMARY_SIZE) bor ?TAG_PRIMARY_IMMED1)). @@ -157,6 +162,38 @@ tag_cons(Res, X) -> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% +ptr_val(Res, X) -> + hipe_rtl:mk_alu(Res, X, 'and', hipe_rtl:mk_imm(bnot ?TAG_PTR_MASK__)). + +%% Returns {Base, Offset, Untag}. To be used like, for example: +%% {Base, Offset, Untag} = untag_ptr(X, ?TAG_PRIMARY_BOXED), +%% ... +%% [Untag, hipe_rtl:mk_load(Dst, Base, hipe_rtl:mk_imm(Offset))]. +%% +%% NB: Base might either be X or a new temp. It must thus not be modified. +untag_ptr(X, Tag) -> + case ?ERTS_USE_LITERAL_TAG of + 0 -> + {X, -Tag, []}; + 1 -> + Base = hipe_rtl:mk_new_reg(), + Untag = ptr_val(Base, X), + {Base, 0, Untag} + end. + +untag_ptr_nooffset(Dst, X, Tag) -> + %% We could just use ptr_val in all cases, but subtraction can use LEA on x86 + %% and can be inlined into effective address computations on several + %% architectures. + case ?ERTS_USE_LITERAL_TAG of + 0 -> + hipe_rtl:mk_alu(Dst, X, 'sub', hipe_rtl:mk_imm(Tag)); + 1 -> + ptr_val(Dst, X) + end. + +%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% + %%% Operations to test if an object has a known type T. test_nil(X, TrueLab, FalseLab, Pred) -> @@ -171,7 +208,8 @@ test_is_boxed(X, TrueLab, FalseLab, Pred) -> hipe_rtl:mk_branch(X, 'and', Mask, 'eq', TrueLab, FalseLab, Pred). get_header(Res, X) -> - hipe_rtl:mk_load(Res, X, hipe_rtl:mk_imm(-(?TAG_PRIMARY_BOXED))). + {Base, Offset, Untag} = untag_ptr(X, ?TAG_PRIMARY_BOXED), + [Untag, hipe_rtl:mk_load(Res, Base, hipe_rtl:mk_imm(Offset))]. mask_and_compare(X, Mask, Value, TrueLab, FalseLab, Pred) -> Tmp = hipe_rtl:mk_new_reg_gcsafe(), @@ -617,21 +655,25 @@ test_either_immed(Arg1, Arg2, TrueLab, FalseLab) -> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% unsafe_car(Dst, Arg) -> - hipe_rtl:mk_load(Dst, Arg, hipe_rtl:mk_imm(-(?TAG_PRIMARY_LIST))). + {Base, Offset, Untag} = untag_ptr(Arg, ?TAG_PRIMARY_LIST), + [Untag, hipe_rtl:mk_load(Dst, Base, hipe_rtl:mk_imm(Offset))]. unsafe_cdr(Dst, Arg) -> + {Base, Offset, Untag} = untag_ptr(Arg, ?TAG_PRIMARY_LIST), WordSize = hipe_rtl_arch:word_size(), - hipe_rtl:mk_load(Dst, Arg, hipe_rtl:mk_imm(-(?TAG_PRIMARY_LIST)+WordSize)). + [Untag, hipe_rtl:mk_load(Dst, Base, hipe_rtl:mk_imm(Offset+WordSize))]. unsafe_constant_element(Dst, Index, Tuple) -> % Index is an immediate WordSize = hipe_rtl_arch:word_size(), - Offset = -(?TAG_PRIMARY_BOXED) + WordSize * hipe_rtl:imm_value(Index), - hipe_rtl:mk_load(Dst, Tuple, hipe_rtl:mk_imm(Offset)). + {Base, Offset0, Untag} = untag_ptr(Tuple, ?TAG_PRIMARY_BOXED), + Offset = Offset0 + WordSize * hipe_rtl:imm_value(Index), + [Untag, hipe_rtl:mk_load(Dst, Base, hipe_rtl:mk_imm(Offset))]. unsafe_update_element(Tuple, Index, Value) -> % Index is an immediate WordSize = hipe_rtl_arch:word_size(), - Offset = -(?TAG_PRIMARY_BOXED) + WordSize * hipe_rtl:imm_value(Index), - hipe_rtl:mk_store(Tuple, hipe_rtl:mk_imm(Offset), Value). + {Base, Offset0, Untag} = untag_ptr(Tuple, ?TAG_PRIMARY_BOXED), + Offset = Offset0 + WordSize * hipe_rtl:imm_value(Index), + [Untag, hipe_rtl:mk_store(Base, hipe_rtl:mk_imm(Offset), Value)]. %%% wrong semantics %% unsafe_variable_element(Dst, Index, Tuple) -> % Index is an unknown fixnum @@ -644,10 +686,12 @@ unsafe_update_element(Tuple, Index, Value) -> % Index is an immediate %% Tmp1 = hipe_rtl:mk_new_reg_gcsafe(), %% Tmp2 = hipe_rtl:mk_new_reg_gcsafe(), %% Shift = ?TAG_IMMED1_SIZE - 2, -%% OffAdj = (?TAG_IMMED1_SMALL bsr Shift) + ?TAG_PRIMARY_BOXED, +%% {Base, Off0, Untag} = untag_ptr(Tuple, ?TAG_PRIMARY_BOXED), +%% OffAdj = (?TAG_IMMED1_SMALL bsr Shift) - Off0, %% [hipe_rtl:mk_alu(Tmp1, Index, 'srl', hipe_rtl:mk_imm(Shift)), %% hipe_rtl:mk_alu(Tmp2, Tmp1, 'sub', hipe_rtl:mk_imm(OffAdj)), -%% hipe_rtl:mk_load(Dst, Tuple, Tmp2)]. +%% Untag, +%% hipe_rtl:mk_load(Base, Tuple, Tmp2)]. element(Dst, Index, Tuple, FailLabName, {tuple, A}, IndexInfo) -> FixnumOkLab = hipe_rtl:mk_new_label(), @@ -660,7 +704,7 @@ element(Dst, Index, Tuple, FailLabName, {tuple, A}, IndexInfo) -> Offset = hipe_rtl:mk_new_reg_gcsafe(), Ptr = hipe_rtl:mk_new_reg(), % offset from Tuple [untag_fixnum(UIndex, Index), - hipe_rtl:mk_alu(Ptr, Tuple, 'sub', hipe_rtl:mk_imm(?TAG_PRIMARY_BOXED)), + untag_ptr_nooffset(Ptr, Tuple, ?TAG_PRIMARY_BOXED), hipe_rtl:mk_alu(Offset, UIndex, 'sll', hipe_rtl:mk_imm(hipe_rtl_arch:log2_word_size())), hipe_rtl:mk_load(Dst, Ptr, Offset)]; @@ -769,7 +813,7 @@ gen_element_tail(Dst, Tuple, Arity, UIndex, FailLabName, IndexOkLab) -> hipe_rtl:mk_branch(ZeroIndex, 'geu', Arity, FailLabName, hipe_rtl:label_name(IndexOkLab), 0.01), IndexOkLab, - hipe_rtl:mk_alu(Ptr, Tuple, 'sub', hipe_rtl:mk_imm(?TAG_PRIMARY_BOXED)), + untag_ptr_nooffset(Ptr, Tuple, ?TAG_PRIMARY_BOXED), hipe_rtl:mk_alu(Offset, UIndex, 'sll', hipe_rtl:mk_imm(hipe_rtl_arch:log2_word_size())), hipe_rtl:mk_load(Dst, Ptr, Offset)]. @@ -777,11 +821,13 @@ gen_element_tail(Dst, Tuple, Arity, UIndex, FailLabName, IndexOkLab) -> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% unsafe_closure_element(Dst, Index, Closure) -> % Index is an immediate - Offset = -(?TAG_PRIMARY_BOXED) %% Untag + %% XXX: Can there even be closure literals? + {Base, Offset0, Untag} = untag_ptr(Closure, ?TAG_PRIMARY_BOXED), + Offset = Offset0 %% Untag + ?EFT_ENV %% Field offset %% Index from 1 to N hence -1) + (hipe_rtl_arch:word_size() * (hipe_rtl:imm_value(Index)-1)), - hipe_rtl:mk_load(Dst, Closure, hipe_rtl:mk_imm(Offset)). + [Untag, hipe_rtl:mk_load(Dst, Base, hipe_rtl:mk_imm(Offset))]. mk_fun_header() -> hipe_rtl:mk_imm(?HEADER_FUN). @@ -790,7 +836,7 @@ tag_fun(Res, X) -> tag_boxed(Res, X). %% untag_fun(Res, X) -> -%% hipe_rtl:mk_alu(Res, X, 'sub', hipe_rtl:mk_imm(?TAG_PRIMARY_BOXED)). +%% untag_ptr_nooffset(Res, X, ?TAG_PRIMARY_BOXED). if_fun_get_arity_and_address(ArityReg, AddressReg, FunP, BadFunLab, Pred) -> %% EmuAddressPtrReg = hipe_rtl:mk_new_reg(), @@ -801,15 +847,15 @@ if_fun_get_arity_and_address(ArityReg, AddressReg, FunP, BadFunLab, Pred) -> TrueLab0 = hipe_rtl:mk_new_label(), %% TrueLab1 = hipe_rtl:mk_new_label(), IsFunCode = test_closure(FunP, hipe_rtl:label_name(TrueLab0), BadFunLab, Pred), + {Base, Offset, Untag} = untag_ptr(FunP, ?TAG_PRIMARY_BOXED), GetArityCode = [TrueLab0, %% Funp->arity contains the arity - hipe_rtl:mk_load(ArityReg, FunP, - hipe_rtl:mk_imm(-(?TAG_PRIMARY_BOXED)+ - ?EFT_ARITY)), - hipe_rtl:mk_load(FEPtrReg, FunP, - hipe_rtl:mk_imm(-(?TAG_PRIMARY_BOXED)+ - ?EFT_FE)), + Untag, + hipe_rtl:mk_load(ArityReg, Base, + hipe_rtl:mk_imm(Offset+?EFT_ARITY)), + hipe_rtl:mk_load(FEPtrReg, Base, + hipe_rtl:mk_imm(Offset+?EFT_FE)), hipe_rtl:mk_load(AddressReg, FEPtrReg, hipe_rtl:mk_imm(?EFE_NATIVE_ADDRESS))], IsFunCode ++ GetArityCode. @@ -927,20 +973,24 @@ test_subbinary(Binary, TrueLblName, FalseLblName) -> unsafe_load_float(DstLo, DstHi, Src) -> WordSize = hipe_rtl_arch:word_size(), - Offset1 = -(?TAG_PRIMARY_BOXED) + WordSize, + {Base, Offset0, Untag} = untag_ptr(Src, ?TAG_PRIMARY_BOXED), + Offset1 = Offset0 + WordSize, Offset2 = Offset1 + 4, %% This should really be 4 and not WordSize case hipe_rtl_arch:endianess() of little -> - [hipe_rtl:mk_load(DstLo, Src, hipe_rtl:mk_imm(Offset1), int32, unsigned), - hipe_rtl:mk_load(DstHi, Src, hipe_rtl:mk_imm(Offset2), int32, unsigned)]; + [Untag, + hipe_rtl:mk_load(DstLo, Base, hipe_rtl:mk_imm(Offset1), int32, unsigned), + hipe_rtl:mk_load(DstHi, Base, hipe_rtl:mk_imm(Offset2), int32, unsigned)]; big -> - [hipe_rtl:mk_load(DstHi, Src, hipe_rtl:mk_imm(Offset1), int32, unsigned), - hipe_rtl:mk_load(DstLo, Src, hipe_rtl:mk_imm(Offset2), int32, unsigned)] + [Untag, + hipe_rtl:mk_load(DstHi, Base, hipe_rtl:mk_imm(Offset1), int32, unsigned), + hipe_rtl:mk_load(DstLo, Base, hipe_rtl:mk_imm(Offset2), int32, unsigned)] end. unsafe_untag_float(Dst, Src) -> - Offset = -(?TAG_PRIMARY_BOXED) + hipe_rtl_arch:word_size(), - [hipe_rtl:mk_fload(Dst, Src, hipe_rtl:mk_imm(Offset))]. + {Base, Offset0, Untag} = untag_ptr(Src, ?TAG_PRIMARY_BOXED), + Offset = Offset0 + hipe_rtl_arch:word_size(), + [Untag, hipe_rtl:mk_fload(Dst, Base, hipe_rtl:mk_imm(Offset))]. unsafe_tag_float(Dst, Src) -> {GetHPInsn, HP, PutHPInsn} = hipe_rtl_arch:heap_pointer(), @@ -999,8 +1049,9 @@ get_one_word_pos_bignum(USize, Size, Fail) -> unsafe_get_one_word_pos_bignum(USize, Size) -> WordSize = hipe_rtl_arch:word_size(), - Imm = hipe_rtl:mk_imm(1*WordSize-?TAG_PRIMARY_BOXED), - [hipe_rtl:mk_load(USize, Size, Imm)]. + {Base, Offset, Untag} = untag_ptr(Size, ?TAG_PRIMARY_BOXED), + Imm = hipe_rtl:mk_imm(1*WordSize+Offset), + [Untag, hipe_rtl:mk_load(USize, Base, Imm)]. -spec bignum_sizeneed(non_neg_integer()) -> non_neg_integer(). @@ -1040,7 +1091,7 @@ create_matchstate(Max, BinSize, Base, Offset, Orig, Ms) -> SizeInWords = ((ByteSize div WordSize) - 1), Header = hipe_rtl:mk_imm(mk_header(SizeInWords, ?TAG_HEADER_BIN_MATCHSTATE)), [GetHPInsn, - hipe_rtl:mk_alu(Ms, HP, add, hipe_rtl:mk_imm(?TAG_PRIMARY_BOXED)), + tag_boxed(Ms, HP), set_field_from_term({matchstate,thing_word}, Ms, Header), set_field_from_term({matchstate,{matchbuffer,orig}}, Ms, Orig), set_field_from_term({matchstate,{matchbuffer,base}}, Ms, Base), @@ -1078,7 +1129,10 @@ convert_matchstate(Ms) -> size_from_header(SizeInWords, Header), hipe_rtl:mk_alu(Hole, SizeInWords, sub, hipe_rtl:mk_imm(?SUB_BIN_WORDSIZE)), mk_var_header(BigIntHeader, Hole, ?TAG_HEADER_POS_BIG), - hipe_rtl:mk_store(Ms, hipe_rtl:mk_imm(?SUB_BIN_WORDSIZE*WordSize-?TAG_PRIMARY_BOXED), + %% Matchstates can't be literals; so untagging with ?TAG_PRIMARY_BOXED is + %% fine here + hipe_rtl:mk_store(Ms, hipe_rtl:mk_imm(?SUB_BIN_WORDSIZE*WordSize + -?TAG_PRIMARY_BOXED), BigIntHeader)]. compare_matchstate(Max, Ms, LargeEnough, TooSmall) -> @@ -1087,8 +1141,10 @@ compare_matchstate(Max, Ms, LargeEnough, TooSmall) -> SizeInWords = ((ByteSize div WordSize) - 1), Header = hipe_rtl:mk_imm(mk_header(SizeInWords, ?TAG_HEADER_BIN_MATCHSTATE)), RealHeader = hipe_rtl:mk_new_reg_gcsafe(), - [hipe_rtl:mk_load(RealHeader, Ms, hipe_rtl:mk_imm(-?TAG_PRIMARY_BOXED)), - hipe_rtl:mk_branch(RealHeader, ge, Header, LargeEnough, TooSmall)]. + %% Matchstates can't be literals; so untagging with ?TAG_PRIMARY_BOXED is fine + %% here + [hipe_rtl:mk_load(RealHeader, Ms, hipe_rtl:mk_imm(-?TAG_PRIMARY_BOXED)), + hipe_rtl:mk_branch(RealHeader, ge, Header, LargeEnough, TooSmall)]. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% @@ -1207,15 +1263,22 @@ get_field_size1({matchbuffer, base}) -> get_field_size1({matchbuffer, binsize}) -> ?MB_SIZE_SIZE. +get_field_addr_from_term(Struct, Term, Dst) -> + {Base, Offset0, Untag} = untag_ptr(Term, ?TAG_PRIMARY_BOXED), + Offset = hipe_rtl:mk_imm(get_field_offset(Struct) + Offset0), + [Untag, hipe_rtl:mk_alu(Dst, Base, add, Offset)]. + get_field_from_term(Struct, Term, Dst) -> - Offset = hipe_rtl:mk_imm(get_field_offset(Struct) - ?TAG_PRIMARY_BOXED), + {Base, Offset0, Untag} = untag_ptr(Term, ?TAG_PRIMARY_BOXED), + Offset = hipe_rtl:mk_imm(get_field_offset(Struct) + Offset0), Size = get_field_size(Struct), - hipe_rtl:mk_load(Dst, Term, Offset, Size, unsigned). + [Untag, hipe_rtl:mk_load(Dst, Base, Offset, Size, unsigned)]. set_field_from_term(Struct, Term, Value) -> - Offset = hipe_rtl:mk_imm(get_field_offset(Struct) - ?TAG_PRIMARY_BOXED), + {Base, Offset0, Untag} = untag_ptr(Term, ?TAG_PRIMARY_BOXED), + Offset = hipe_rtl:mk_imm(get_field_offset(Struct) + Offset0), Size = get_field_size(Struct), - hipe_rtl:mk_store(Term, Offset, Value, Size). + [Untag, hipe_rtl:mk_store(Base, Offset, Value, Size)]. get_field_from_pointer(Struct, Term, Dst) -> Offset = hipe_rtl:mk_imm(get_field_offset(Struct)), @@ -1229,6 +1292,8 @@ set_field_from_pointer(Struct, Term, Value) -> extract_matchbuffer(Mb, Ms) -> What = {matchstate, matchbuffer}, + %% Matchstates can't be literals; so untagging with ?TAG_PRIMARY_BOXED is fine + %% here Offset = hipe_rtl:mk_imm(get_field_offset(What) - ?TAG_PRIMARY_BOXED), hipe_rtl:mk_alu(Mb, Ms, add, Offset). diff --git a/lib/hipe/ssa/hipe_ssa.inc b/lib/hipe/ssa/hipe_ssa.inc index c7c1a8e1d7..29e8b92266 100644 --- a/lib/hipe/ssa/hipe_ssa.inc +++ b/lib/hipe/ssa/hipe_ssa.inc @@ -463,20 +463,20 @@ updateStatementDefs([], Statement, Current, Acc) -> %%---------------------------------------------------------------------- updateIndices(Current, Variable) -> - case ?CODE:is_var(Variable) of - true -> - NewVar = ?CODE:mk_new_var(), - {NewVar,gb_trees:enter(Variable, NewVar, Current)}; - false -> - case is_fp_temp(Variable) of - true -> - NewFVar = mk_new_fp_temp(), - {NewFVar,gb_trees:enter(Variable, NewFVar, Current)}; - false -> - NewReg = ?CODE:mk_new_reg(), - {NewReg,gb_trees:enter(Variable, NewReg, Current)} - end - end. + New = + case ?CODE:is_var(Variable) of + true -> ?CODE:mk_new_var(); + false -> + case is_fp_temp(Variable) of + true -> mk_new_fp_temp(); + false -> + case ?CODE:reg_is_gcsafe(Variable) of + true -> ?CODE:mk_new_reg_gcsafe(); + false -> ?CODE:mk_new_reg() + end + end + end, + {New, gb_trees:enter(Variable, New, Current)}. %%---------------------------------------------------------------------- %% Procedure : updateSuccPhi/4 diff --git a/lib/hipe/test/hipe_testsuite_driver.erl b/lib/hipe/test/hipe_testsuite_driver.erl index ee9c57a908..8813af5dfc 100644 --- a/lib/hipe/test/hipe_testsuite_driver.erl +++ b/lib/hipe/test/hipe_testsuite_driver.erl @@ -161,7 +161,8 @@ run(TestCase, Dir, _OutDir) -> %% end, DataFiles), %% try ok = TestCase:test(), - HiPEOpts = try TestCase:hipe_options() catch error:undef -> [] end, + HiPEOpts0 = try TestCase:hipe_options() catch error:undef -> [] end, + HiPEOpts = HiPEOpts0 ++ hipe_options(), {ok, TestCase} = hipe:c(TestCase, HiPEOpts), ok = TestCase:test(), {ok, TestCase} = hipe:c(TestCase, [o1|HiPEOpts]), @@ -179,3 +180,6 @@ run(TestCase, Dir, _OutDir) -> %% lists:foreach(fun (DF) -> ok end, % = file:delete(DF) end, %% [filename:join(OutDir, D) || D <- DataFiles]) %% end. + +hipe_options() -> + [verify_gcsafe]. diff --git a/lib/inets/doc/src/notes.xml b/lib/inets/doc/src/notes.xml index 10ef84d7cf..07e29b5542 100644 --- a/lib/inets/doc/src/notes.xml +++ b/lib/inets/doc/src/notes.xml @@ -33,7 +33,58 @@ <file>notes.xml</file> </header> - <section><title>Inets 6.4.2</title> + <section><title>Inets 6.4.4</title> + + <section><title>Fixed Bugs and Malfunctions</title> + <list> + <item> + <p> + Correct the handling of location headers so that the + status code is not hard coded. This should have been + fixed by commit 2cc5ba70cbbc6b3ace81a2a0324417c3b65265bb + but unfortunately was broken during a code refactoring + and unnoticed due to a faulty placed test case.</p> + <p> + Own Id: OTP-14761</p> + </item> + </list> + </section> + +</section> + +<section><title>Inets 6.4.3</title> + + <section><title>Improvements and New Features</title> + <list> + <item> + <p> + Fix broken handling of POST requests</p> + <p> + New chunk mechanism of body data in POST requests added + in 5d01c70ca399edf28e99dc760506329689fab6ba broke + handling of POST body data not using the new mechanism.</p> + <p> + Own Id: OTP-14656</p> + </item> + <item> + <p> + Make sure ints:stop/2 of the service httpd is synchronous</p> + <p> + Own Id: OTP-14696</p> + </item> + <item> + <p> + Honor status code returned by ESI script and modernize + "location" header handling.</p> + <p> + Own Id: OTP-14716</p> + </item> + </list> + </section> + +</section> + +<section><title>Inets 6.4.2</title> <section><title>Fixed Bugs and Malfunctions</title> <list> diff --git a/lib/inets/src/http_client/httpc_handler.erl b/lib/inets/src/http_client/httpc_handler.erl index bd1d2e833a..6907bf5262 100644 --- a/lib/inets/src/http_client/httpc_handler.erl +++ b/lib/inets/src/http_client/httpc_handler.erl @@ -736,7 +736,7 @@ maybe_send_answer(Request, Answer, State) -> answer_request(Request, Answer, State). deliver_answer(#request{from = From} = Request) - when is_pid(From) -> + when From =/= answer_sent -> Response = httpc_response:error(Request, socket_closed_remotely), httpc_response:send(From, Response); deliver_answer(_Request) -> diff --git a/lib/inets/src/http_client/httpc_response.erl b/lib/inets/src/http_client/httpc_response.erl index b3b11b74ab..91638f5d2e 100644 --- a/lib/inets/src/http_client/httpc_response.erl +++ b/lib/inets/src/http_client/httpc_response.erl @@ -269,7 +269,7 @@ parse_headers(<<?LF,?LF,Body/binary>>, Header, Headers, MaxHeaderSize, Result, Relaxed); parse_headers(<<?CR,?LF,?CR,?LF,Body/binary>>, Header, Headers, - MaxHeaderSize, Result, _) -> + MaxHeaderSize, Result, Relaxed) -> HTTPHeaders = [lists:reverse(Header) | Headers], Length = lists:foldl(fun(H, Acc) -> length(H) + Acc end, 0, HTTPHeaders), @@ -277,8 +277,42 @@ parse_headers(<<?CR,?LF,?CR,?LF,Body/binary>>, Header, Headers, true -> ResponseHeaderRcord = http_response:headers(HTTPHeaders, #http_response_h{}), - {ok, list_to_tuple( - lists:reverse([Body, ResponseHeaderRcord | Result]))}; + + %% RFC7230, Section 3.3.3 + %% If a message is received with both a Transfer-Encoding and a + %% Content-Length header field, the Transfer-Encoding overrides the + %% Content-Length. Such a message might indicate an attempt to + %% perform request smuggling (Section 9.5) or response splitting + %% (Section 9.4) and ought to be handled as an error. A sender MUST + %% remove the received Content-Length field prior to forwarding such + %% a message downstream. + case ResponseHeaderRcord#http_response_h.'transfer-encoding' of + undefined -> + {ok, list_to_tuple( + lists:reverse([Body, ResponseHeaderRcord | Result]))}; + Value -> + TransferEncoding = string:lowercase(Value), + ContentLength = ResponseHeaderRcord#http_response_h.'content-length', + if + %% Respond without error but remove Content-Length field in relaxed mode + (Relaxed =:= true) + andalso (TransferEncoding =:= "chunked") + andalso (ContentLength =/= "-1") -> + ResponseHeaderRcordFixed = + ResponseHeaderRcord#http_response_h{'content-length' = "-1"}, + {ok, list_to_tuple( + lists:reverse([Body, ResponseHeaderRcordFixed | Result]))}; + %% Respond with error in default (not relaxed) mode + (Relaxed =:= false) + andalso (TransferEncoding =:= "chunked") + andalso (ContentLength =/= "-1") -> + throw({error, {headers_conflict, {'content-length', + 'transfer-encoding'}}}); + true -> + {ok, list_to_tuple( + lists:reverse([Body, ResponseHeaderRcord | Result]))} + end + end; false -> throw({error, {header_too_long, MaxHeaderSize, MaxHeaderSize-Length}}) diff --git a/lib/inets/src/http_server/httpd.erl b/lib/inets/src/http_server/httpd.erl index 0b632d24e3..540e68e749 100644 --- a/lib/inets/src/http_server/httpd.erl +++ b/lib/inets/src/http_server/httpd.erl @@ -99,7 +99,14 @@ start_service(Conf) -> stop_service({Address, Port}) -> stop_service({Address, Port, ?DEFAULT_PROFILE}); stop_service({Address, Port, Profile}) -> - httpd_sup:stop_child(Address, Port, Profile); + Name = httpd_util:make_name("httpd_instance_sup", Address, Port, Profile), + Pid = whereis(Name), + MonitorRef = erlang:monitor(process, Pid), + Result = httpd_sup:stop_child(Address, Port, Profile), + receive + {'DOWN', MonitorRef, _, _, _} -> + Result + end; stop_service(Pid) when is_pid(Pid) -> case service_info(Pid) of {ok, Info} -> diff --git a/lib/inets/src/http_server/httpd_esi.erl b/lib/inets/src/http_server/httpd_esi.erl index 9406b47802..f5493f6fad 100644 --- a/lib/inets/src/http_server/httpd_esi.erl +++ b/lib/inets/src/http_server/httpd_esi.erl @@ -66,7 +66,7 @@ handle_headers("") -> {ok, [], 200}; handle_headers(Headers) -> NewHeaders = string:tokens(Headers, ?CRLF), - handle_headers(NewHeaders, [], 200). + handle_headers(NewHeaders, [], 200, true). %%%======================================================================== %%% Internal functions @@ -80,21 +80,17 @@ parse_headers([?CR, ?LF, ?CR, ?LF | Rest], Acc) -> parse_headers([Char | Rest], Acc) -> parse_headers(Rest, [Char | Acc]). -handle_headers([], NewHeaders, StatusCode) -> +handle_headers([], NewHeaders, StatusCode, _) -> {ok, NewHeaders, StatusCode}; -handle_headers([Header | Headers], NewHeaders, StatusCode) -> +handle_headers([Header | Headers], NewHeaders, StatusCode, NoESIStatus) -> {FieldName, FieldValue} = httpd_response:split_header(Header, []), case FieldName of - "location" -> - case http_request:is_absolut_uri(FieldValue) of - true -> - handle_headers(Headers, - [{FieldName, FieldValue} | NewHeaders], - 302); - false -> - {proceed, FieldValue} - end; + "location" when NoESIStatus == true -> + handle_headers(Headers, + [{FieldName, FieldValue} | NewHeaders], + 302, NoESIStatus); + "status" -> NewStatusCode = case httpd_util:split(FieldValue," ",2) of @@ -103,8 +99,9 @@ handle_headers([Header | Headers], NewHeaders, StatusCode) -> _ -> 200 end, - handle_headers(Headers, NewHeaders, NewStatusCode); + handle_headers(Headers, NewHeaders, NewStatusCode, false); _ -> handle_headers(Headers, - [{FieldName, FieldValue}| NewHeaders], StatusCode) - end. + [{FieldName, FieldValue}| NewHeaders], StatusCode, + NoESIStatus) + end. diff --git a/lib/inets/src/http_server/httpd_example.erl b/lib/inets/src/http_server/httpd_example.erl index 45b6deba97..47a8c48d01 100644 --- a/lib/inets/src/http_server/httpd_example.erl +++ b/lib/inets/src/http_server/httpd_example.erl @@ -20,7 +20,7 @@ %% -module(httpd_example). -export([print/1]). --export([get/2, put/2, post/2, yahoo/2, test1/2, get_bin/2, peer/2]). +-export([get/2, put/2, post/2, yahoo/2, test1/2, get_bin/2, peer/2,new_status_and_location/2]). -export([newformat/3, post_chunked/3]). %% These are used by the inets test-suite @@ -90,6 +90,9 @@ post(Env,Input) -> yahoo(_Env,_Input) -> "Location: http://www.yahoo.com\r\n\r\n". +new_status_and_location(_Env,_Input) -> + "status:201 Created\r\n Location: http://www.yahoo.com\r\n\r\n". + default(Env,Input) -> [header(), top("Default Example"), diff --git a/lib/inets/src/http_server/httpd_request.erl b/lib/inets/src/http_server/httpd_request.erl index 0eaf073255..007d272323 100644 --- a/lib/inets/src/http_server/httpd_request.erl +++ b/lib/inets/src/http_server/httpd_request.erl @@ -306,10 +306,10 @@ add_chunk([<<>>, Body, Length, MaxChunk]) -> add_chunk([More, Body, Length, MaxChunk]) -> body_chunk(<<Body/binary, More/binary>>, Length, MaxChunk). -body_chunk(<<>> = Body, Length, MaxChunk) -> - {ok, {continue, ?MODULE, add_chunk, [Body, Length, MaxChunk]}}; body_chunk(Body, Length, nolimit) -> whole_body(Body, Length); +body_chunk(<<>> = Body, Length, MaxChunk) -> + {ok, {continue, ?MODULE, add_chunk, [Body, Length, MaxChunk]}}; body_chunk(Body, Length, MaxChunk) when Length > MaxChunk -> case size(Body) >= MaxChunk of diff --git a/lib/inets/src/http_server/httpd_request_handler.erl b/lib/inets/src/http_server/httpd_request_handler.erl index bd4fdd3832..d918f10424 100644 --- a/lib/inets/src/http_server/httpd_request_handler.erl +++ b/lib/inets/src/http_server/httpd_request_handler.erl @@ -516,6 +516,15 @@ handle_body(#state{headers = Headers, body = Body, case ((Length =< MaxBodySize) or (MaxBodySize == nolimit)) of true -> case httpd_request:body_chunk_first(Body, Length, MaxChunk) of + %% This is the case that the we need more data to complete + %% the body but chunking to the mod_esi user is not enabled. + {Module, add_chunk = Function, Args} -> + http_transport:setopts(ModData#mod.socket_type, + ModData#mod.socket, + [{active, once}]), + {noreply, State#state{mfa = + {Module, Function, Args}}}; + %% Chunking to mod_esi user is enabled {ok, {continue, Module, Function, Args}} -> http_transport:setopts(ModData#mod.socket_type, ModData#mod.socket, @@ -525,6 +534,8 @@ handle_body(#state{headers = Headers, body = Body, {ok, {{continue, Chunk}, Module, Function, Args}} -> handle_internal_chunk(State#state{chunk = chunk_start(MaxChunk), body = Chunk}, Module, Function, Args); + %% Whole body delivered, if chunking mechanism is enabled the whole + %% body fits in one chunk. {ok, NewBody} -> handle_response(State#state{chunk = chunk_finish(ChunkState, CbState, MaxChunk), diff --git a/lib/inets/src/http_server/mod_esi.erl b/lib/inets/src/http_server/mod_esi.erl index 3a589ca5f0..3206d957d9 100644 --- a/lib/inets/src/http_server/mod_esi.erl +++ b/lib/inets/src/http_server/mod_esi.erl @@ -339,26 +339,21 @@ erl_scheme_webpage_whole(Mod, Func, Env, Input, ModData) -> {Headers, Body} = httpd_esi:parse_headers(lists:flatten(Response)), Length = httpd_util:flatlength(Body), - case httpd_esi:handle_headers(Headers) of - {proceed, AbsPath} -> - {proceed, [{real_name, httpd_util:split_path(AbsPath)} - | ModData#mod.data]}; - {ok, NewHeaders, StatusCode} -> - send_headers(ModData, StatusCode, - [{"content-length", - integer_to_list(Length)}| NewHeaders]), - case ModData#mod.method of - "HEAD" -> - {proceed, [{response, {already_sent, 200, 0}} | - ModData#mod.data]}; - _ -> - httpd_response:send_body(ModData, - StatusCode, Body), - {proceed, [{response, {already_sent, 200, - Length}} | - ModData#mod.data]} - end - end + {ok, NewHeaders, StatusCode} = httpd_esi:handle_headers(Headers), + send_headers(ModData, StatusCode, + [{"content-length", + integer_to_list(Length)}| NewHeaders]), + case ModData#mod.method of + "HEAD" -> + {proceed, [{response, {already_sent, 200, 0}} | + ModData#mod.data]}; + _ -> + httpd_response:send_body(ModData, + StatusCode, Body), + {proceed, [{response, {already_sent, 200, + Length}} | + ModData#mod.data]} + end end. %% New API that allows the dynamic wepage to be sent back to the client @@ -398,29 +393,23 @@ deliver_webpage_chunk(#mod{config_db = Db} = ModData, Pid, Timeout) -> {continue, _} = Continue -> Continue; {Headers, Body} -> - case httpd_esi:handle_headers(Headers) of - {proceed, AbsPath} -> - {proceed, [{real_name, httpd_util:split_path(AbsPath)} - | ModData#mod.data]}; - {ok, NewHeaders, StatusCode} -> - IsDisableChunkedSend = - httpd_response:is_disable_chunked_send(Db), - case (ModData#mod.http_version =/= "HTTP/1.1") or - (IsDisableChunkedSend) of - true -> - send_headers(ModData, StatusCode, - [{"connection", "close"} | - NewHeaders]); - false -> - send_headers(ModData, StatusCode, - [{"transfer-encoding", - "chunked"} | NewHeaders]) - end, - handle_body(Pid, ModData, Body, Timeout, length(Body), - IsDisableChunkedSend) - end; - timeout -> - send_headers(ModData, 504, [{"connection", "close"}]), + {ok, NewHeaders, StatusCode} = httpd_esi:handle_headers(Headers), + IsDisableChunkedSend = httpd_response:is_disable_chunked_send(Db), + case (ModData#mod.http_version =/= "HTTP/1.1") or + (IsDisableChunkedSend) of + true -> + send_headers(ModData, StatusCode, + [{"connection", "close"} | + NewHeaders]); + false -> + send_headers(ModData, StatusCode, + [{"transfer-encoding", + "chunked"} | NewHeaders]) + end, + handle_body(Pid, ModData, Body, Timeout, length(Body), + IsDisableChunkedSend); + timeout -> + send_headers(ModData, 504, [{"connection", "close"}]), httpd_socket:close(ModData#mod.socket_type, ModData#mod.socket), {proceed,[{response, {already_sent, 200, 0}} | ModData#mod.data]} end. @@ -560,15 +549,10 @@ eval(#mod{method = Method} = ModData, ESIBody, Modules) {ok, Response} -> {Headers, _} = httpd_esi:parse_headers(lists:flatten(Response)), - case httpd_esi:handle_headers(Headers) of - {ok, _, StatusCode} -> - {proceed,[{response, {StatusCode, Response}} | - ModData#mod.data]}; - {proceed, AbsPath} -> - {proceed, [{real_name, AbsPath} | - ModData#mod.data]} - end - end; + {ok, _, StatusCode} =httpd_esi:handle_headers(Headers), + {proceed,[{response, {StatusCode, Response}} | + ModData#mod.data]} + end; false -> {proceed,[{status, {403, ModData#mod.request_uri, diff --git a/lib/inets/src/inets_app/inets.appup.src b/lib/inets/src/inets_app/inets.appup.src index a86413147c..fdf4cc6e07 100644 --- a/lib/inets/src/inets_app/inets.appup.src +++ b/lib/inets/src/inets_app/inets.appup.src @@ -18,10 +18,14 @@ %% %CopyrightEnd% {"%VSN%", [ + {<<"6.4.3">>, [{load_module, httpd_esi, + soft_purge, soft_purge, []}]}, {<<"6\\..*">>,[{restart_application, inets}]}, {<<"5\\..*">>,[{restart_application, inets}]} ], [ + {<<"6.4.3">>, [{load_module, httpd_esi, + soft_purge, soft_purge, []}]}, {<<"6\\..*">>,[{restart_application, inets}]}, {<<"5\\..*">>,[{restart_application, inets}]} ] diff --git a/lib/inets/test/http_format_SUITE.erl b/lib/inets/test/http_format_SUITE.erl index 4e10a97f58..647eff4f7c 100644 --- a/lib/inets/test/http_format_SUITE.erl +++ b/lib/inets/test/http_format_SUITE.erl @@ -535,8 +535,11 @@ esi_parse_headers(Config) when is_list(Config) -> {"location","http://foo.bar.se"}], 302} = httpd_esi:handle_headers(Headers2), - {proceed,"/foo/bar.html"} = - httpd_esi:handle_headers("location:/foo/bar.html\r\n"). + {ok,[{"location","/foo/bar.html"}], 302} = + httpd_esi:handle_headers("location:/foo/bar.html\r\n"), + + {ok,[{"location","http://foo/bar.html"}],201} = + httpd_esi:handle_headers("status:201 Created\r\nlocation:http://foo/bar.html\r\n"). %%-------------------------------------------------------------------- cgi_parse_headers() -> diff --git a/lib/inets/test/httpc_SUITE.erl b/lib/inets/test/httpc_SUITE.erl index 5dfb1474e5..cc166d522e 100644 --- a/lib/inets/test/httpc_SUITE.erl +++ b/lib/inets/test/httpc_SUITE.erl @@ -115,6 +115,7 @@ only_simulated() -> invalid_chunk_size, headers_dummy, headers_with_obs_fold, + headers_conflict_chunked_with_length, empty_response_header, remote_socket_close, remote_socket_close_async, @@ -130,7 +131,8 @@ only_simulated() -> port_in_host_header, redirect_port_in_host_header, relaxed, - multipart_chunks + multipart_chunks, + stream_fun_server_close ]. misc() -> @@ -745,7 +747,7 @@ empty_body() -> empty_body(Config) when is_list(Config) -> URL = url(group_name(Config), "/empty.html", Config), {ok, {{_,200,_}, [_ | _], []}} = - httpc:request(get, {URL, []}, [{timeout, 500}], []). + httpc:request(get, {URL, []}, [], []). %%------------------------------------------------------------------------- @@ -977,7 +979,6 @@ headers_dummy(Config) when is_list(Config) -> {"If-Range", "Sat, 29 Oct 1994 19:43:31 GMT"}, {"If-Match", "*"}, {"Content-Type", "text/plain"}, - {"Content-Encoding", "chunked"}, {"Content-Length", "6"}, {"Content-Language", "en"}, {"Content-Location", "http://www.foobar.se"}, @@ -1003,6 +1004,18 @@ headers_with_obs_fold(Config) when is_list(Config) -> %%------------------------------------------------------------------------- +headers_conflict_chunked_with_length(doc) -> + ["Test the code for handling headers with both Transfer-Encoding" + "and Content-Length which must receive error in default (not relaxed) mode" + "and must receive successful response in relaxed mode"]; +headers_conflict_chunked_with_length(Config) when is_list(Config) -> + Request = {url(group_name(Config), "/headers_conflict_chunked_with_length.html", Config), []}, + {error, {could_not_parse_as_http, _}} = httpc:request(get, Request, [{relaxed, false}], []), + {ok,{{_,200,_},_,_}} = httpc:request(get, Request, [{relaxed, true}], []), + ok. + +%%------------------------------------------------------------------------- + invalid_headers(Config) -> Request = {url(group_name(Config), "/dummy.html", Config), [{"cookie", undefined}]}, {error, _} = httpc:request(get, Request, [], []). @@ -1178,6 +1191,22 @@ wait_for_whole_response(Config) when is_list(Config) -> ReqSeqNumServer ! shutdown. %%-------------------------------------------------------------------- +stream_fun_server_close() -> + [{doc, "Test that an error msg is received when using a receiver fun as stream target"}]. +stream_fun_server_close(Config) when is_list(Config) -> + Request = {url(group_name(Config), "/delay_close.html", Config), []}, + Self = self(), + Fun = fun(X) -> Self ! X end, + {ok, RequestId} = httpc:request(get, Request, [], [{sync, false}, {receiver, Fun}]), + receive + {RequestId, {error, Reason}} -> + ct:pal("Close ~p", [Reason]), + ok + after 13000 -> + ct:fail(did_not_receive_close) + end. + +%%-------------------------------------------------------------------- %% Internal Functions ------------------------------------------------ %%-------------------------------------------------------------------- stream(ReceiverPid, Receiver, Config) -> @@ -1852,7 +1881,6 @@ handle_uri(_,"/dummy_headers.html",_,_,Socket,_) -> %% user to evaluate. This is not a valid response %% it only tests that the header handling code works. Head = "HTTP/1.1 200 ok\r\n" ++ - "Content-Length:32\r\n" ++ "Pragma:1#no-cache\r\n" ++ "Via:1.0 fred, 1.1 nowhere.com (Apache/1.1)\r\n" ++ "Warning:1#pseudonym foobar\r\n" ++ @@ -1882,6 +1910,15 @@ handle_uri(_,"/obs_folded_headers.html",_,_,_,_) -> " b\r\n\r\n" "Hello"; +handle_uri(_,"/headers_conflict_chunked_with_length.html",_,_,Socket,_) -> + Head = "HTTP/1.1 200 ok\r\n" + "Content-Length:32\r\n" + "Transfer-Encoding:Chunked\r\n\r\n", + send(Socket, Head), + send(Socket, http_chunk:encode("<HTML><BODY>fo")), + send(Socket, http_chunk:encode("obar</BODY></HTML>")), + http_chunk:encode_last(); + handle_uri(_,"/capital_transfer_encoding.html",_,_,Socket,_) -> Head = "HTTP/1.1 200 ok\r\n" ++ "Transfer-Encoding:Chunked\r\n\r\n", @@ -2029,6 +2066,9 @@ handle_uri(_,"/multipart_chunks.html",_,_,Socket,_) -> send(Socket, Head), send_multipart_chunks(Socket), http_chunk:encode_last(); +handle_uri(_,"/delay_close.html",_,_,Socket,_) -> + ct:sleep(10000), + close(Socket); handle_uri("HEAD",_,_,_,_,_) -> "HTTP/1.1 200 ok\r\n" ++ "Content-Length:0\r\n\r\n"; diff --git a/lib/inets/test/httpd_SUITE.erl b/lib/inets/test/httpd_SUITE.erl index 6c8728470b..9a85c51d24 100644 --- a/lib/inets/test/httpd_SUITE.erl +++ b/lib/inets/test/httpd_SUITE.erl @@ -129,7 +129,7 @@ groups() -> {http_1_1, [], [host, chunked, expect, cgi, cgi_chunked_encoding_test, trace, range, if_modified_since, mod_esi_chunk_timeout, - esi_put] ++ http_head() ++ http_get() ++ load()}, + esi_put, esi_post] ++ http_head() ++ http_get() ++ load()}, {http_1_0, [], [host, cgi, trace] ++ http_head() ++ http_get() ++ load()}, {http_0_9, [], http_head() ++ http_get() ++ load()} ]. @@ -923,8 +923,11 @@ esi(Config) when is_list(Config) -> {no_header, "cache-control"}]), ok = http_status("GET /cgi-bin/erl/httpd_example:peer ", Config, [{statuscode, 200}, - {header, "peer-cert-exist", peer(Config)}]). - + {header, "peer-cert-exist", peer(Config)}]), + ok = http_status("GET /cgi-bin/erl/httpd_example:new_status_and_location ", + Config, [{statuscode, 201}, + {header, "location"}]). + %%------------------------------------------------------------------------- esi_put() -> [{doc, "Test mod_esi PUT"}]. @@ -932,7 +935,20 @@ esi_put() -> esi_put(Config) when is_list(Config) -> ok = http_status("PUT /cgi-bin/erl/httpd_example/put/123342234123 ", Config, [{statuscode, 200}]). - +%%------------------------------------------------------------------------- +esi_post() -> + [{doc, "Test mod_esi POST"}]. + +esi_post(Config) when is_list(Config) -> + Chunk = "ZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZ", + Data = lists:duplicate(10000, Chunk), + Length = lists:flatlength(Data), + ok = http_status("POST /cgi-bin/erl/httpd_example/post ", + {"Content-Length:" ++ integer_to_list(Length) ++ "\r\n", + Data}, + [{http_version, "HTTP/1.1"} |Config], + [{statuscode, 200}]). + %%------------------------------------------------------------------------- mod_esi_chunk_timeout(Config) when is_list(Config) -> ok = httpd_1_1:mod_esi_chunk_timeout(proplists:get_value(type, Config), diff --git a/lib/inets/test/httpd_mod.erl b/lib/inets/test/httpd_mod.erl index d9118aa1a4..2035b50248 100644 --- a/lib/inets/test/httpd_mod.erl +++ b/lib/inets/test/httpd_mod.erl @@ -779,9 +779,14 @@ esi(Type, Port, Host, Node) -> [{statuscode, 200}, {no_header, "cache-control"}, {version, "HTTP/1.0"}]), + ok = httpd_test_lib:verify_request(Type, Host, Port, Node, + "GET /cgi-bin/erl/httpd_example:new_status_and_location" + " HTTP/1.1\r\n\r\n", + [{statuscode, 201}, + {header, "Location"}, + {version, "HTTP/1.1"}]), ok. - %%-------------------------------------------------------------------- get(Type, Port, Host, Node) -> ok = httpd_test_lib:verify_request(Type, Host, Port, Node, diff --git a/lib/inets/test/inets_SUITE.erl b/lib/inets/test/inets_SUITE.erl index 38b8229389..1abd96a228 100644 --- a/lib/inets/test/inets_SUITE.erl +++ b/lib/inets/test/inets_SUITE.erl @@ -213,7 +213,6 @@ start_httpd(Config) when is_list(Config) -> true = lists:member(Pid0, Pids0), [_|_] = inets:services_info(), inets:stop(httpd, Pid0), - ct:sleep(500), Pids1 = [ServicePid || {_, ServicePid} <- inets:services()], false = lists:member(Pid0, Pids1), {ok, Pid0b} = @@ -222,7 +221,6 @@ start_httpd(Config) when is_list(Config) -> true = lists:member(Pid0b, Pids0b), [_|_] = inets:services_info(), inets:stop(httpd, Pid0b), - ct:sleep(500), Pids1 = [ServicePid || {_, ServicePid} <- inets:services()], false = lists:member(Pid0b, Pids1), {ok, Pid1} = diff --git a/lib/inets/vsn.mk b/lib/inets/vsn.mk index 34b6902747..560d524bac 100644 --- a/lib/inets/vsn.mk +++ b/lib/inets/vsn.mk @@ -19,6 +19,6 @@ # %CopyrightEnd% APPLICATION = inets -INETS_VSN = 6.4.2 +INETS_VSN = 6.4.4 PRE_VSN = APP_VSN = "$(APPLICATION)-$(INETS_VSN)$(PRE_VSN)" diff --git a/lib/kernel/src/user.erl b/lib/kernel/src/user.erl index a5cc7b0ec1..872e63ab53 100644 --- a/lib/kernel/src/user.erl +++ b/lib/kernel/src/user.erl @@ -1,7 +1,7 @@ %% %% %CopyrightBegin% %% -%% Copyright Ericsson AB 1996-2016. All Rights Reserved. +%% Copyright Ericsson AB 1996-2017. All Rights Reserved. %% %% Licensed under the Apache License, Version 2.0 (the "License"); %% you may not use this file except in compliance with the License. @@ -398,7 +398,7 @@ get_line(Prompt, Port, Q, Acc, Enc) -> get_line_bytes(Prompt, Port, Q, Acc, Bytes, Enc); {Port, eof} -> put(eof, true), - {ok, eof, []}; + {ok, eof, queue:new()}; {io_request,From,ReplyAs,{get_geometry,_}=Req} when is_pid(From) -> do_io_request(Req, From, ReplyAs, Port, queue:new()), @@ -615,7 +615,7 @@ get_chars(Prompt, M, F, Xa, Port, Q, State, Enc) -> get_chars_bytes(State, M, F, Xa, Port, Q, Bytes, Enc); {Port, eof} -> put(eof, true), - {ok, eof, []}; + {ok, eof, queue:new()}; %%{io_request,From,ReplyAs,Request} when is_pid(From) -> %% get_chars_req(Prompt, M, F, Xa, Port, queue:new(), State, %% Request, From, ReplyAs); diff --git a/lib/kernel/test/erl_distribution_wb_SUITE.erl b/lib/kernel/test/erl_distribution_wb_SUITE.erl index 03aaee56b7..258ed4f88c 100644 --- a/lib/kernel/test/erl_distribution_wb_SUITE.erl +++ b/lib/kernel/test/erl_distribution_wb_SUITE.erl @@ -65,6 +65,7 @@ ?DFLAG_EXTENDED_PIDS_PORTS bor ?DFLAG_UTF8_ATOMS)). +-define(PASS_THROUGH, $p). -define(shutdown(X), exit(X)). -define(int16(X), [((X) bsr 8) band 16#ff, (X) band 16#ff]). @@ -676,10 +677,9 @@ recv_message(Socket) -> case gen_tcp:recv(Socket, 0) of {ok,Data} -> B0 = list_to_binary(Data), - {_,B1} = erlang:split_binary(B0,1), - Header = binary_to_term(B1), - Siz = byte_size(term_to_binary(Header)), - {_,B2} = erlang:split_binary(B1,Siz), + <<?PASS_THROUGH, B1/binary>> = B0, + {Header,Siz} = binary_to_term(B1,[used]), + <<_:Siz/binary,B2/binary>> = B1, Message = case (catch binary_to_term(B2)) of {'EXIT', _} -> could_not_digest_message; diff --git a/lib/kernel/test/gen_tcp_misc_SUITE.erl b/lib/kernel/test/gen_tcp_misc_SUITE.erl index 331864b5de..e47023d201 100644 --- a/lib/kernel/test/gen_tcp_misc_SUITE.erl +++ b/lib/kernel/test/gen_tcp_misc_SUITE.erl @@ -1572,52 +1572,56 @@ fill_sendq(Config) when is_list(Config) -> Master = self(), Server = spawn_link(fun () -> - {ok,L} = gen_tcp:listen - (0, [{active,false},binary, - {reuseaddr,true},{packet,0}]), + {ok,L} = gen_tcp:listen(0, [{active,false},binary, + {reuseaddr,true},{packet,0}]), {ok,Port} = inet:port(L), Master ! {self(),client, fill_sendq_client(Port, Master)}, fill_sendq_srv(L, Master) end), io:format("~p Server~n", [Server]), - receive {Server,client,Client} -> - io:format("~p Client~n", [Client]), - receive {Server,reader,Reader} -> - io:format("~p Reader~n", [Reader]), - fill_sendq_loop(Server, Client, Reader) + receive + {Server,client,Client} -> + io:format("~p Client~n", [Client]), + receive + {Server,reader,Reader} -> + io:format("~p Reader~n", [Reader]), + fill_sendq_loop(Server, Client, Reader) end end. fill_sendq_loop(Server, Client, Reader) -> %% Master %% - receive {Server,send} -> + receive + {Server,send} -> fill_sendq_loop(Server, Client, Reader) after 2000 -> %% Send queue full, sender blocked -> close client. io:format("Send timeout, closing Client...~n", []), Client ! {self(),close}, - receive {Server,[{error,closed}]} -> - io:format("Got server closed.~n"), - receive {Reader,[{error,closed}]} -> - io:format - ("Got reader closed.~n"), - ok - after 3000 -> - ct:fail({timeout,{closed,reader}}) - end; - {Reader,[{error,closed}]} -> - io:format("Got reader closed.~n"), - receive {Server,[{error,closed}]} -> - io:format("Got server closed~n"), - ok - after 3000 -> - ct:fail({timeout,{closed,server}}) - end - after 3000 -> - ct:fail({timeout,{closed,[server,reader]}}) - end + receive + {Server,[{error,closed}]} -> + io:format("Got server closed.~n"), + receive + {Reader,[{error,closed}]} -> + io:format("Got reader closed.~n"), + ok + after 3000 -> + ct:fail({timeout,{closed,reader}}) + end; + {Reader,[{error,closed}]} -> + io:format("Got reader closed.~n"), + receive + {Server,[{error,closed}]} -> + io:format("Got server closed~n"), + ok + after 3000 -> + ct:fail({timeout,{closed,server}}) + end + after 3000 -> + ct:fail({timeout,{closed,[server,reader]}}) + end end. fill_sendq_srv(L, Master) -> diff --git a/lib/kernel/test/zlib_SUITE.erl b/lib/kernel/test/zlib_SUITE.erl index 131a0685cd..f203ef878f 100644 --- a/lib/kernel/test/zlib_SUITE.erl +++ b/lib/kernel/test/zlib_SUITE.erl @@ -166,7 +166,7 @@ api_deflateInit(Config) when is_list(Config) -> ?m(ok, zlib:deflateInit(Z12,default,deflated,-Wbits,8,default)), ?m(ok,zlib:close(Z11)), ?m(ok,zlib:close(Z12)) - end, lists:seq(8, 15)), + end, lists:seq(9, 15)), lists:foreach(fun(MemLevel) -> Z = zlib:open(), @@ -213,12 +213,46 @@ api_deflateReset(Config) when is_list(Config) -> %% Test deflateParams. api_deflateParams(Config) when is_list(Config) -> + Levels = [none, default, best_speed, best_compression] ++ lists:seq(0, 9), + Strategies = [filtered, huffman_only, rle, default], + Z1 = zlib:open(), ?m(ok, zlib:deflateInit(Z1, default)), - ?m(L when is_list(L), zlib:deflate(Z1, <<1,1,1,1,1,1,1,1,1>>, none)), - ?m(ok, zlib:deflateParams(Z1, best_compression, huffman_only)), - ?m(L when is_list(L), zlib:deflate(Z1, <<1,1,1,1,1,1,1,1,1>>, sync)), - ?m(ok, zlib:close(Z1)). + + ApiTest = + fun(Level, Strategy) -> + ?m(ok, zlib:deflateParams(Z1, Level, Strategy)), + ?m(ok, zlib:deflateReset(Z1)) + end, + + [ ApiTest(Level, Strategy) || Level <- Levels, Strategy <- Strategies ], + + ?m(ok, zlib:close(Z1)), + + FlushTest = + fun FlushTest(Size, Level, Strategy) -> + Z = zlib:open(), + ok = zlib:deflateInit(Z, default), + Data = gen_determ_rand_bytes(Size), + case zlib:deflate(Z, Data, none) of + [<<120, 156>>] -> + %% All data is present in the internal zlib state, and will + %% be flushed on deflateParams. + + ok = zlib:deflateParams(Z, Level, Strategy), + Compressed = [<<120, 156>>, zlib:deflate(Z, <<>>, finish)], + Data = zlib:uncompress(Compressed), + zlib:close(Z), + + FlushTest(Size + (1 bsl 10), Level, Strategy); + _Other -> + ok + end + end, + + [ FlushTest(1, Level, Strategy) || Level <- Levels, Strategy <- Strategies ], + + ok. %% Test deflate. api_deflate(Config) when is_list(Config) -> @@ -652,6 +686,11 @@ api_g_un_zip(Config) when is_list(Config) -> Concatenated = <<Bin/binary, Bin/binary>>, ?m(Concatenated, zlib:gunzip([Comp, Comp])), + %% Don't explode if the uncompressed size is a perfect multiple of the + %% internal inflate chunk size. + ChunkSizedData = <<0:16384/unit:8>>, + ?m(ChunkSizedData, zlib:gunzip(zlib:gzip(ChunkSizedData))), + %% Bad CRC; bad length. BadCrc = bad_crc_data(), ?m(?EXIT(data_error),(catch zlib:gunzip(BadCrc))), @@ -762,13 +801,13 @@ zip_usage({run,ZIP,ORIG}) -> ?m(ok, zlib:deflateInit(Z, default, deflated, -15, 8, default)), C2 = zlib:deflate(Z, ORIG, finish), - ?m(true, C1 == list_to_binary(C2)), + ?m(ORIG, zlib:unzip(C2)), ?m(ok, zlib:deflateEnd(Z)), ?m(ok, zlib:deflateInit(Z, none, deflated, -15, 8, filtered)), ?m(ok, zlib:deflateParams(Z, default, default)), C3 = zlib:deflate(Z, ORIG, finish), - ?m(true, C1 == list_to_binary(C3)), + ?m(ORIG, zlib:unzip(C3)), ?m(ok, zlib:deflateEnd(Z)), ok = zlib:close(Z), @@ -978,23 +1017,37 @@ split_bin(Last,Acc) -> only_allow_owner(Config) when is_list(Config) -> Z = zlib:open(), + Owner = self(), ?m(ok, zlib:inflateInit(Z)), ?m(ok, zlib:inflateReset(Z)), {Pid, Ref} = spawn_monitor( fun() -> - ?m(?EXIT(not_on_controlling_process), zlib:inflateReset(Z)) + ?m(?EXIT(not_on_controlling_process), zlib:inflateReset(Z)), + Owner ! '$transfer_ownership', + receive + '$ownership_transferred' -> + ?m(ok, zlib:inflateReset(Z)) + after 200 -> + ct:fail("Never received transfer signal.") + end end), + ownership_transfer_check(Z, Pid, Ref). +ownership_transfer_check(Z, WorkerPid, Ref) -> receive - {'DOWN', Ref, process, Pid, _Reason} -> - ok + '$transfer_ownership' -> + zlib:set_controlling_process(Z, WorkerPid), + WorkerPid ! '$ownership_transferred', + ownership_transfer_check(Z, WorkerPid, Ref); + {'DOWN', Ref, process, WorkerPid, normal} -> + ok; + {'DOWN', Ref, process, WorkerPid, Reason} -> + ct:fail("Spawned worker crashed with reason ~p.", [Reason]) after 200 -> ct:fail("Spawned worker timed out.") - end, - - ?m(ok, zlib:inflateReset(Z)). + end. sub_heap_binaries(Config) when is_list(Config) -> Compressed = zlib:compress(<<"gurka">>), diff --git a/lib/mnesia/src/mnesia.erl b/lib/mnesia/src/mnesia.erl index b68b2de028..190fb2b56d 100644 --- a/lib/mnesia/src/mnesia.erl +++ b/lib/mnesia/src/mnesia.erl @@ -2681,7 +2681,7 @@ del_table_index(Tab, Ix) -> -spec transform_table(Tab::table(), Fun, [Attr]) -> t_result(ok) when Attr :: atom(), - Fun:: fun((Record::tuple()) -> Transformed::tuple()). + Fun:: fun((Record::tuple()) -> Transformed::tuple()) | ignore. transform_table(Tab, Fun, NewA) -> try val({Tab, record_name}) of OldRN -> mnesia_schema:transform_table(Tab, Fun, NewA, OldRN) @@ -2692,7 +2692,7 @@ transform_table(Tab, Fun, NewA) -> -spec transform_table(Tab::table(), Fun, [Attr], RecName) -> t_result(ok) when RecName :: atom(), Attr :: atom(), - Fun:: fun((Record::tuple()) -> Transformed::tuple()). + Fun:: fun((Record::tuple()) -> Transformed::tuple()) | ignore. transform_table(Tab, Fun, NewA, NewRN) -> mnesia_schema:transform_table(Tab, Fun, NewA, NewRN). diff --git a/lib/observer/src/cdv_bin_cb.erl b/lib/observer/src/cdv_bin_cb.erl index 5502869973..a4a542297c 100644 --- a/lib/observer/src/cdv_bin_cb.erl +++ b/lib/observer/src/cdv_bin_cb.erl @@ -71,6 +71,8 @@ hex_binary_fun(Bin) -> plain_html(io_lib:format("~s",[S])) end. +format_hex(<<>>,_) -> + []; format_hex(<<B1:4,B2:4>>,_) -> [integer_to_list(B1,16),integer_to_list(B2,16)]; format_hex(<<B1:4,B2:4,Bin/binary>>,0) -> diff --git a/lib/observer/src/crashdump_viewer.erl b/lib/observer/src/crashdump_viewer.erl index 51ff69fd69..50e9102ff0 100644 --- a/lib/observer/src/crashdump_viewer.erl +++ b/lib/observer/src/crashdump_viewer.erl @@ -26,10 +26,25 @@ %% Tables %% ------ %% cdv_dump_index_table: This table holds all tags read from the -%% crashdump. Each tag indicates where the information about a -%% specific item starts. The table entry for a tag includes the start -%% position for this item-information. In a crash dump file, all tags -%% start with a "=" at the beginning of a line. +%% crashdump, except the 'binary' tag. Each tag indicates where the +%% information about a specific item starts. The table entry for a +%% tag includes the start position for this item-information. In a +%% crash dump file, all tags start with a "=" at the beginning of a +%% line. +%% +%% cdv_binary_index_table: This table holds all 'binary' tags. The hex +%% address for each binary is converted to its integer value before +%% storing Address -> Start Position in this table. The hex value of +%% the address is never used for lookup. +%% +%% cdv_reg_proc_table: This table holds mappings between pid and +%% registered name. This is used for timers and monitors. +%% +%% cdv_heap_file_chars: For each 'proc_heap' and 'literals' tag, this +%% table contains the number of characters to read from the crash dump +%% file. This is used for giving an indication in percent of the +%% progress when parsing this data. +%% %% %% Process state %% ------------- @@ -73,6 +88,9 @@ -export([init/1, handle_call/3, handle_cast/2, handle_info/2, terminate/2, code_change/3]). +%% Test support +-export([get_dump_versions/0]). + %% Debug support -export([debug/1,stop_debug/0]). @@ -87,6 +105,7 @@ % line_head/1 function can return -define(not_available,"N/A"). -define(binary_size_progress_limit,10000). +-define(max_dump_version,[0,4]). %% All possible tags - use macros in order to avoid misspelling in the code @@ -104,6 +123,7 @@ -define(index_table,index_table). -define(instr_data,instr_data). -define(internal_ets,internal_ets). +-define(literals,literals). -define(loaded_modules,loaded_modules). -define(memory,memory). -define(memory_map,memory_map). @@ -293,6 +313,11 @@ port(Id) -> expand_binary(Pos) -> call({expand_binary,Pos}). +%%%----------------------------------------------------------------- +%%% For testing only - called from crashdump_viewer_SUITE +get_dump_versions() -> + call(get_dump_versions). + %%==================================================================== %% Server functions %%==================================================================== @@ -454,8 +479,9 @@ handle_call(index_tables,_From,State=#state{file=File}) -> handle_call(schedulers,_From,State=#state{file=File}) -> Schedulers=schedulers(File), TW = truncated_warning([?scheduler]), - {reply,{ok,Schedulers,TW},State}. - + {reply,{ok,Schedulers,TW},State}; +handle_call(get_dump_versions,_From,State=#state{dump_vsn=DumpVsn}) -> + {reply,{ok,{?max_dump_version,DumpVsn}},State}. %%-------------------------------------------------------------------- @@ -780,11 +806,12 @@ parse_vsn_str(Str,WS) -> %%%----------------------------------------------------------------- -%%% Traverse crash dump and insert index in table for each heading -%%% -%%% Progress is reported during the time and MUST be checked with -%%% crashdump_viewer:get_progress/0 until it returns {ok,done}. +%%% Traverse crash dump and insert index in table for each heading. +%%% Progress is reported during the time. do_read_file(File) -> + erase(?literals), %Clear literal cache. + put(truncated,false), %Not truncated (yet). + erase(truncated_reason), %Not truncated (yet). case file:read_file_info(File) of {ok,#file_info{type=regular, access=FileA, @@ -796,17 +823,21 @@ do_read_file(File) -> {Tag,Id,Rest,N1} = tag(Fd,TagAndRest,1), case Tag of ?erl_crash_dump -> - reset_tables(), - insert_index(Tag,Id,N1+1), - put_last_tag(Tag,""), - DumpVsn = [list_to_integer(L) || - L<-string:lexemes(Id,".")], - AddrAdj = get_bin_addr_adj(DumpVsn), - indexify(Fd,AddrAdj,Rest,N1), - end_progress(), - check_if_truncated(), - close(Fd), - {ok,DumpVsn}; + case check_dump_version(Id) of + {ok,DumpVsn} -> + reset_tables(), + insert_index(Tag,Id,N1+1), + put_last_tag(Tag,""), + AddrAdj = get_bin_addr_adj(DumpVsn), + indexify(Fd,AddrAdj,Rest,N1), + end_progress(), + check_if_truncated(), + close(Fd), + {ok,DumpVsn}; + Error -> + close(Fd), + Error + end; _Other -> R = io_lib:format( "~ts is not an Erlang crash dump~n", @@ -834,6 +865,18 @@ do_read_file(File) -> {error,R} end. +check_dump_version(Vsn) -> + DumpVsn = [list_to_integer(L) || L<-string:lexemes(Vsn,".")], + if DumpVsn > ?max_dump_version -> + Info = + "This Crashdump Viewer is too old for the given " + "Erlang crash dump. Please use a newer version of " + "Crashdump Viewer.", + {error,Info}; + true -> + {ok,DumpVsn} + end. + indexify(Fd,AddrAdj,Bin,N) -> case binary:match(Bin,<<"\n=">>) of {Start,Len} -> @@ -856,6 +899,19 @@ indexify(Fd,AddrAdj,Bin,N) -> {?proc_heap,LastId} -> [{_,LastPos}] = lookup_index(?proc_heap,LastId), ets:insert(cdv_heap_file_chars,{LastId,N+Start+1-LastPos}); + {?literals,[]} -> + case get(truncated_reason) of + undefined -> + [{_,LastPos}] = lookup_index(?literals,[]), + ets:insert(cdv_heap_file_chars, + {literals,N+Start+1-LastPos}); + _ -> + %% Literals are truncated. Make sure we never + %% attempt to read in the literals. (Heaps that + %% references literals will show markers for + %% incomplete heaps, but will otherwise work.) + delete_index(?literals, []) + end; _ -> ok end, indexify(Fd,AddrAdj,Rest,N1); @@ -908,6 +964,7 @@ check_if_truncated() -> find_truncated_proc({Tag,_Id}) when Tag==?atoms; Tag==?binary; Tag==?instr_data; + Tag==?literals; Tag==?memory_status; Tag==?memory_map -> put(truncated_proc,false); @@ -1065,7 +1122,7 @@ get_proc_details(File,Pid,WS,DumpVsn) -> {{Stack,MsgQ,Dict},TW} = case truncated_warning([{?proc,Pid}]) of [] -> - {expand_memory(Fd,Pid,DumpVsn),[]}; + expand_memory(Fd,Pid,DumpVsn); TW0 -> {{[],[],[]},TW0} end, @@ -1386,12 +1443,40 @@ maybe_other_node2(Channel) -> expand_memory(Fd,Pid,DumpVsn) -> BinAddrAdj = get_bin_addr_adj(DumpVsn), put(fd,Fd), - Dict = read_heap(Fd,Pid,BinAddrAdj,gb_trees:empty()), + Dict0 = case get(?literals) of + undefined -> + Literals = read_literals(Fd), + put(?literals,Literals), + put(fd,Fd), + Literals; + Literals -> + Literals + end, + Dict = read_heap(Fd,Pid,BinAddrAdj,Dict0), Expanded = {read_stack_dump(Fd,Pid,BinAddrAdj,Dict), read_messages(Fd,Pid,BinAddrAdj,Dict), read_dictionary(Fd,Pid,BinAddrAdj,Dict)}, erase(fd), - Expanded. + IncompleteWarning = + case erase(incomplete_heap) of + undefined -> + []; + true -> + ["WARNING: This process has an incomplete heap. " + "Some information might be missing."] + end, + {Expanded,IncompleteWarning}. + +read_literals(Fd) -> + case lookup_index(?literals,[]) of + [{_,Start}] -> + [{_,Chars}] = ets:lookup(cdv_heap_file_chars,literals), + init_progress("Reading literals",Chars), + pos_bof(Fd,Start), + read_heap(0,gb_trees:empty()); + [] -> + gb_trees:empty() + end. %%%----------------------------------------------------------------- %%% This is a workaround for a bug in dump versions prior to 0.3: @@ -2591,8 +2676,26 @@ parse_heap_term("Ys"++Line0, Addr, BinAddrAdj, D0) -> %Sub binary. end end, D = gb_trees:insert(Addr, Term, D0), - {Term,Line,D}. - + {Term,Line,D}; +parse_heap_term("Mf"++Line0, Addr, BinAddrAdj, D0) -> %Flatmap. + {Size,":"++Line1} = get_hex(Line0), + {Keys,":"++Line2,D1} = parse_term(Line1, BinAddrAdj, D0), + {Values,Line,D2} = parse_tuple(Size, Line2, Addr,BinAddrAdj, D1, []), + Pairs = zip_tuples(tuple_size(Keys), Keys, Values, []), + Map = maps:from_list(Pairs), + D = gb_trees:update(Addr, Map, D2), + {Map,Line,D}; +parse_heap_term("Mh"++Line0, Addr, BinAddrAdj, D0) -> %Head node in a hashmap. + {MapSize,":"++Line1} = get_hex(Line0), + {N,":"++Line2} = get_hex(Line1), + {Nodes,Line,D1} = parse_tuple(N, Line2, Addr, BinAddrAdj, D0, []), + Map = maps:from_list(flatten_hashmap_nodes(Nodes)), + MapSize = maps:size(Map), %Assertion. + D = gb_trees:update(Addr, Map, D1), + {Map,Line,D}; +parse_heap_term("Mn"++Line0, Addr, BinAddrAdj, D) -> %Interior node in a hashmap. + {N,":"++Line} = get_hex(Line0), + parse_tuple(N, Line, Addr, BinAddrAdj, D, []). parse_tuple(0, Line, Addr, _, D0, Acc) -> Tuple = list_to_tuple(lists:reverse(Acc)), @@ -2606,6 +2709,25 @@ parse_tuple(N, Line0, Addr, BinAddrAdj, D0, Acc) -> parse_tuple(N-1, Line, Addr, BinAddrAdj, D, [Term|Acc]) end. +zip_tuples(0, _T1, _T2, Acc) -> + Acc; +zip_tuples(N, T1, T2, Acc) when N =< tuple_size(T1) -> + zip_tuples(N-1, T1, T2, [{element(N, T1),element(N, T2)}|Acc]). + +flatten_hashmap_nodes(Tuple) -> + flatten_hashmap_nodes_1(tuple_size(Tuple), Tuple, []). + +flatten_hashmap_nodes_1(0, _Tuple, Acc) -> + Acc; +flatten_hashmap_nodes_1(N, Tuple0, Acc0) -> + case element(N, Tuple0) of + [K|V] -> + flatten_hashmap_nodes_1(N-1, Tuple0, [{K,V}|Acc0]); + Tuple when is_tuple(Tuple) -> + Acc = flatten_hashmap_nodes_1(N-1, Tuple0, Acc0), + flatten_hashmap_nodes_1(tuple_size(Tuple), Tuple, Acc) + end. + parse_term([$H|Line0], BinAddrAdj, D) -> %Pointer to heap term. {Ptr,Line} = get_hex(Line0), deref_ptr(Ptr, Line, BinAddrAdj, D); @@ -2672,6 +2794,7 @@ deref_ptr(Ptr, Line, BinAddrAdj, D0) -> none -> case get(fd) of end_of_heap -> + put(incomplete_heap,true), {['#CDVIncompleteHeap'],Line,D0}; Fd -> case bytes(Fd) of @@ -2794,6 +2917,10 @@ reset_tables() -> insert_index(Tag,Id,Pos) -> ets:insert(cdv_dump_index_table,{{Tag,Pos},Id}). +delete_index(Tag,Id) -> + Ms = [{{{Tag,'$1'},Id},[],[true]}], + ets:select_delete(cdv_dump_index_table, Ms). + lookup_index({Tag,Id}) -> lookup_index(Tag,Id); lookup_index(Tag) -> @@ -2810,6 +2937,7 @@ insert_binary_index(Addr,Pos) -> lookup_binary_index(Addr) -> ets:lookup(cdv_binary_index_table,Addr). + %%----------------------------------------------------------------- %% Convert tags read from crashdump to atoms used as first part of key %% in cdv_dump_index_table @@ -2827,6 +2955,7 @@ tag_to_atom("hidden_node") -> ?hidden_node; tag_to_atom("index_table") -> ?index_table; tag_to_atom("instr_data") -> ?instr_data; tag_to_atom("internal_ets") -> ?internal_ets; +tag_to_atom("literals") -> ?literals; tag_to_atom("loaded_modules") -> ?loaded_modules; tag_to_atom("memory") -> ?memory; tag_to_atom("mod") -> ?mod; @@ -2850,8 +2979,10 @@ tag_to_atom(UnknownTag) -> %%%----------------------------------------------------------------- %%% Store last tag for use when truncated, and reason if aborted put_last_tag(?abort,Reason) -> - %% Don't overwrite the real last tag - put(truncated_reason,Reason); + %% Don't overwrite the real last tag, and make sure to return + %% the previous last tag. + put(truncated_reason,Reason), + get(last_tag); put_last_tag(Tag,Id) -> put(last_tag,{Tag,Id}). diff --git a/lib/observer/src/observer_html_lib.erl b/lib/observer/src/observer_html_lib.erl index 68381bcc7b..8202f91030 100644 --- a/lib/observer/src/observer_html_lib.erl +++ b/lib/observer/src/observer_html_lib.erl @@ -355,11 +355,11 @@ href_proc_bin(From, T, Acc, LTB) -> PreviewStr end end; - [PreviewIntStr,SizeStr,Md5] when From =:= obs -> + [PreviewIntStr,PreviewBitSizeStr,SizeStr,Md5] when From =:= obs -> Size = list_to_integer(SizeStr), PreviewInt = list_to_integer(PreviewIntStr), - PrevSize = (trunc(math:log2(PreviewInt)/8)+1)*8, - PreviewStr = preview_string(Size,<<PreviewInt:PrevSize>>), + PreviewBitSize = list_to_integer(PreviewBitSizeStr), + PreviewStr = preview_string(Size,<<PreviewInt:PreviewBitSize>>), if LTB -> href("TARGET=\"expanded\"", ["#OBSBinary?key1="++PreviewIntStr++ diff --git a/lib/observer/src/observer_lib.erl b/lib/observer/src/observer_lib.erl index e5ffe61d25..52e6c3e52a 100644 --- a/lib/observer/src/observer_lib.erl +++ b/lib/observer/src/observer_lib.erl @@ -810,7 +810,7 @@ progress_dialog_destroy({Dialog,_,_}) -> make_obsbin(Bin,Tab) -> Size = byte_size(Bin), - Preview = + {Preview,PreviewBitSize} = try %% The binary might be a unicode string, in which case we %% don't want to split it in the middle of a grapheme @@ -819,14 +819,14 @@ make_obsbin(Bin,Tab) -> PB1 = string:slice(Bin,0,PL1), PS1 = byte_size(PB1) * 8, <<P1:PS1>> = PB1, - P1 + {P1,PS1} catch _:_ -> %% Probably not a string, so just split anywhere PS2 = min(Size, 10) * 8, <<P2:PS2, _/binary>> = Bin, - P2 + {P2,PS2} end, Hash = erlang:phash2(Bin), Key = {Preview, Size, Hash}, ets:insert(Tab, {Key,Bin}), - ['#OBSBin',Preview,Size,Hash]. + ['#OBSBin',Preview,PreviewBitSize,Size,Hash]. diff --git a/lib/observer/src/observer_trace_wx.erl b/lib/observer/src/observer_trace_wx.erl index 8127248262..2c3b46a3a1 100644 --- a/lib/observer/src/observer_trace_wx.erl +++ b/lib/observer/src/observer_trace_wx.erl @@ -1201,7 +1201,7 @@ make_ms(MS) -> make_ms(Name,Term,FunStr). make_ms(Name, Term, FunStr) -> - #match_spec{name=Name, term=Term, str=io_lib:format("~tw", Term), func = FunStr}. + #match_spec{name=Name, term=Term, str=io_lib:format("~tw", [Term]), func = FunStr}. parse_tp({tp, Mod, FAs}, State) -> Patterns = [#tpattern{m=Mod,fa={F,A}, ms=make_ms(List)} || diff --git a/lib/observer/test/crashdump_helper.erl b/lib/observer/test/crashdump_helper.erl index f37d9057cb..41041682c2 100644 --- a/lib/observer/test/crashdump_helper.erl +++ b/lib/observer/test/crashdump_helper.erl @@ -19,7 +19,9 @@ %% -module(crashdump_helper). --export([n1_proc/2,remote_proc/2]). +-export([n1_proc/2,remote_proc/2, + dump_maps/0,create_maps/0, + create_binaries/0]). -compile(r18). -include_lib("common_test/include/ct.hrl"). @@ -60,6 +62,7 @@ n1_proc(Creator,_N2,Pid2,Port2,_L) -> put(ref,Ref), put(pid,Pid), put(bin,Bin), + put(bins,create_binaries()), put(sub_bin,SubBin), put(bignum,83974938738373873), put(neg_bignum,-38748762783736367), @@ -92,3 +95,38 @@ remote_proc(P1,Creator) -> Creator ! {self(),done}, receive after infinity -> ok end end). + +create_binaries() -> + Sizes = lists:seq(60, 70) ++ lists:seq(120, 140), + [begin + <<H:16/unit:8>> = erlang:md5(<<Size:32>>), + Data = ((H bsl (8*150)) div (H+7919)), + <<Data:Size/unit:8>> + end || Size <- Sizes]. + +%%% +%%% Test dumping of maps. Dumping of maps only from OTP 20.2. +%%% + +dump_maps() -> + Parent = self(), + F = fun() -> + register(aaaaaaaa_maps, self()), + put(maps, create_maps()), + Parent ! {self(),done}, + receive _ -> ok end + end, + Pid = spawn_link(F), + receive + {Pid,done} -> + {ok,Pid} + end. + +create_maps() -> + Map0 = maps:from_list([{I,[I,I+1]} || I <- lists:seq(1, 40)]), + Map1 = maps:from_list([{I,{a,[I,I*I],{}}} || I <- lists:seq(1, 100)]), + Map2 = maps:from_list([{{I},(I*I) bsl 24} || I <- lists:seq(1, 10000)]), + Map3 = lists:foldl(fun(I, A) -> + A#{I=>I*I} + end, Map2, lists:seq(-10, 0)), + #{a=>Map0,b=>Map1,c=>Map2,d=>Map3,e=>#{}}. diff --git a/lib/observer/test/crashdump_viewer_SUITE.erl b/lib/observer/test/crashdump_viewer_SUITE.erl index f9ac884743..29b9e406ae 100644 --- a/lib/observer/test/crashdump_viewer_SUITE.erl +++ b/lib/observer/test/crashdump_viewer_SUITE.erl @@ -25,7 +25,7 @@ %% Test functions -export([all/0, suite/0,groups/0,init_per_group/2,end_per_group/2, start_stop/1,load_file/1,not_found_items/1, - non_existing/1,not_a_crashdump/1,old_crashdump/1]). + non_existing/1,not_a_crashdump/1,old_crashdump/1,new_crashdump/1]). -export([init_per_suite/1, end_per_suite/1]). -export([init_per_testcase/2, end_per_testcase/2]). @@ -83,6 +83,7 @@ all() -> non_existing, not_a_crashdump, old_crashdump, + new_crashdump, load_file, not_found_items ]. @@ -212,6 +213,25 @@ not_a_crashdump(Config) when is_list(Config) -> ok = crashdump_viewer:stop(). +%% Try to load a file with newer version than this crashdump viewer can handle +new_crashdump(Config) -> + Dump = hd(?config(dumps,Config)), + ok = start_backend(Dump), + {ok,{MaxVsn,CurrentVsn}} = crashdump_viewer:get_dump_versions(), + if MaxVsn =/= CurrentVsn -> + ct:fail("Current dump version is not equal to cdv's max version"); + true -> + ok + end, + ok = crashdump_viewer:stop(), + NewerVsn = lists:join($.,[integer_to_list(X+1) || X <- MaxVsn]), + PrivDir = ?config(priv_dir,Config), + NewDump = filename:join(PrivDir,"new_erl_crash.dump"), + ok = file:write_file(NewDump,"=erl_crash_dump:"++NewerVsn++"\n"), + {error, Reason} = start_backend(NewDump), + "This Crashdump Viewer is too old" ++_ = Reason, + ok = crashdump_viewer:stop(). + %% Load files into the tool and view all pages load_file(Config) when is_list(Config) -> case ?t:is_debug() of @@ -328,7 +348,7 @@ browse_file(File) -> io:format(" info read",[]), - lookat_all_pids(Procs), + lookat_all_pids(Procs,is_truncated(File),incomplete_allowed(File)), io:format(" pids ok",[]), lookat_all_ports(Ports), io:format(" ports ok",[]), @@ -339,6 +359,21 @@ browse_file(File) -> Procs. % used as second arg to special/2 +is_truncated(File) -> + case filename:extension(File) of + ".trunc"++_ -> + true; + _ -> + false + end. + +incomplete_allowed(File) -> + %% Incomplete heap is allowed for native libs, since some literals + %% are not dumped - and for pre OTP-20 (really pre 20.2) releases, + %% since literals were not dumped at all then. + Rel = get_rel_from_dump_name(File), + Rel < 20 orelse test_server:is_native(lists). + special(File,Procs) -> case filename:extension(File) of ".full_dist" -> @@ -364,6 +399,10 @@ special(File,Procs) -> crashdump_viewer:expand_binary({SOffset,SSize,SPos}), io:format(" expand binary ok",[]), + Binaries = crashdump_helper:create_binaries(), + verify_binaries(Binaries, proplists:get_value(bins,Dict)), + io:format(" binaries ok",[]), + #proc{last_calls=LastCalls} = ProcDetails, true = length(LastCalls) =< 4, @@ -513,20 +552,56 @@ special(File,Procs) -> io:format(" unicode table name ok",[]), ok; + ".maps" -> + %% I registered a process as aaaaaaaa_maps in the map dump + %% to make sure it will be the first in the list when sorted + %% on names. + [#proc{pid=Pid0,name=Name}|_Rest] = lists:keysort(#proc.name,Procs), + "aaaaaaaa_maps" = Name, + Pid = pid_to_list(Pid0), + {ok,ProcDetails=#proc{},[]} = crashdump_viewer:proc_details(Pid), + io:format(" process details ok",[]), + + #proc{dict=Dict} = ProcDetails, + %% io:format("~p\n", [Dict]), + Maps = crashdump_helper:create_maps(), + Maps = proplists:get_value(maps,Dict), + io:format(" maps ok",[]), + ok; _ -> ok end, ok. +verify_binaries([H|T1], [H|T2]) -> + %% Heap binary. + verify_binaries(T1, T2); +verify_binaries([Bin|T1], [['#CDVBin',Offset,Size,Pos]|T2]) -> + %% Refc binary. + {ok,<<Bin:Size/binary>>} = crashdump_viewer:expand_binary({Offset,Size,Pos}), + verify_binaries(T1, T2); +verify_binaries([], []) -> + ok. -lookat_all_pids([]) -> +lookat_all_pids([],_,_) -> ok; -lookat_all_pids([#proc{pid=Pid0}|Procs]) -> +lookat_all_pids([#proc{pid=Pid0}|Procs],TruncAllowed,IncompAllowed) -> Pid = pid_to_list(Pid0), - {ok,_ProcDetails=#proc{},_ProcTW} = crashdump_viewer:proc_details(Pid), - {ok,_Ets,_EtsTW} = crashdump_viewer:ets_tables(Pid), - {ok,_Timers,_TimersTW} = crashdump_viewer:timers(Pid), - lookat_all_pids(Procs). + {ok,_ProcDetails=#proc{},ProcTW} = crashdump_viewer:proc_details(Pid), + {ok,_Ets,EtsTW} = crashdump_viewer:ets_tables(Pid), + {ok,_Timers,TimersTW} = crashdump_viewer:timers(Pid), + case {ProcTW,EtsTW,TimersTW} of + {[],[],[]} -> + ok; + {["WARNING: This process has an incomplete heap."++_],[],[]} + when IncompAllowed -> + ok; % native libs, literals might not be included in dump + _ when TruncAllowed -> + ok; % truncated dump + TWs -> + ct:fail({unexpected_warning,TWs}) + end, + lookat_all_pids(Procs,TruncAllowed,IncompAllowed). lookat_all_ports([]) -> ok; @@ -574,16 +649,11 @@ do_create_dumps(DataDir,Rel) -> current -> CD3 = dump_with_args(DataDir,Rel,"instr","+Mim true"), CD4 = dump_with_strange_module_name(DataDir,Rel,"strangemodname"), - Tmp = dump_with_args(DataDir,Rel,"trunc_bytes",""), - {ok,#file_info{size=Max}} = file:read_file_info(Tmp), - ok = file:delete(Tmp), - Bytes = max(15,rand:uniform(Max)), - CD5 = dump_with_args(DataDir,Rel,"trunc_bytes", - "-env ERL_CRASH_DUMP_BYTES " ++ - integer_to_list(Bytes)), + CD5 = dump_with_size_limit_reached(DataDir,Rel,"trunc_bytes"), CD6 = dump_with_unicode_atoms(DataDir,Rel,"unicode"), + CD7 = dump_with_maps(DataDir,Rel,"maps"), TruncatedDumps = truncate_dump(CD1), - {[CD1,CD2,CD3,CD4,CD5,CD6|TruncatedDumps], DosDump}; + {[CD1,CD2,CD3,CD4,CD5,CD6,CD7|TruncatedDumps], DosDump}; _ -> {[CD1,CD2], DosDump} end. @@ -596,7 +666,10 @@ truncate_dump(File) -> {win32,_} -> <<"\r\n">>; _ -> <<"\n">> end, - [StartBin,AfterTag] = binary:split(Bin,BinTag), + %% Split after "our binary" created by crashdump_helper + %% (it may not be the first binary). + RE = <<"\n=binary:(?=[0-9A-Z]+",NewLine/binary,"FF:010203)">>, + [StartBin,AfterTag] = re:split(Bin,RE,[{parts,2}]), [AddrAndSize,BinaryAndRest] = binary:split(AfterTag,Colon), [Binary,_Rest] = binary:split(BinaryAndRest,NewLine), TruncSize = byte_size(Binary) - 2, @@ -689,6 +762,28 @@ dump_with_strange_module_name(DataDir,Rel,DumpName) -> ?t:stop_node(n1), CD. +dump_with_size_limit_reached(DataDir,Rel,DumpName) -> + Tmp = dump_with_args(DataDir,Rel,DumpName,""), + {ok,#file_info{size=Max}} = file:read_file_info(Tmp), + ok = file:delete(Tmp), + dump_with_size_limit_reached(DataDir,Rel,DumpName,Max). + +dump_with_size_limit_reached(DataDir,Rel,DumpName,Max) -> + Bytes = max(15,rand:uniform(Max)), + CD = dump_with_args(DataDir,Rel,DumpName, + "-env ERL_CRASH_DUMP_BYTES " ++ + integer_to_list(Bytes)), + {ok,#file_info{size=Size}} = file:read_file_info(CD), + if Size < Bytes -> + %% This means that the dump was actually smaller than the + %% randomly selected truncation size, so we'll just do it + %% again with a smaller numer + ok = file:delete(CD), + dump_with_size_limit_reached(DataDir,Rel,DumpName,Size-3); + true -> + CD + end. + dump_with_unicode_atoms(DataDir,Rel,DumpName) -> Opt = rel_opt(Rel), Pz = "-pz \"" ++ filename:dirname(code:which(?MODULE)) ++ "\"", @@ -699,6 +794,16 @@ dump_with_unicode_atoms(DataDir,Rel,DumpName) -> ?t:stop_node(n1), CD. +dump_with_maps(DataDir,Rel,DumpName) -> + Opt = rel_opt(Rel), + Pz = "-pz \"" ++ filename:dirname(code:which(?MODULE)) ++ "\"", + PzOpt = [{args,Pz}], + {ok,N1} = ?t:start_node(n1,peer,Opt ++ PzOpt), + {ok,_Pid} = rpc:call(N1,crashdump_helper,dump_maps,[]), + CD = dump(N1,DataDir,Rel,DumpName), + ?t:stop_node(n1), + CD. + dump(Node,DataDir,Rel,DumpName) -> Crashdump = filename:join(DataDir, dump_prefix(Rel)++DumpName), rpc:call(Node,os,putenv,["ERL_CRASH_DUMP",Crashdump]), @@ -751,6 +856,11 @@ dump_prefix(current) -> dump_prefix(Rel) -> lists:concat(["r",Rel,"_dump."]). +get_rel_from_dump_name(File) -> + Name = filename:basename(File), + ["r"++Rel|_] = string:split(Name,"_"), + list_to_integer(Rel). + compat_rel(current) -> ""; compat_rel(Rel) -> diff --git a/lib/parsetools/include/leexinc.hrl b/lib/parsetools/include/leexinc.hrl index b4449607cb..2a74c252ff 100644 --- a/lib/parsetools/include/leexinc.hrl +++ b/lib/parsetools/include/leexinc.hrl @@ -272,6 +272,8 @@ skip_cont(Rest, Line, {skip_token,Push}, Error) -> skip_cont(Rest, Line, {error,_S}, Error) -> skip_tokens(yystate(), Rest, Line, Rest, 0, Line, Error, reject, 0). +-compile({nowarn_unused_function, [yyrev/1, yyrev/2, yypre/2, yysuf/2]}). + yyrev(List) -> lists:reverse(List). yyrev(List, Tail) -> lists:reverse(List, Tail). yypre(List, N) -> lists:sublist(List, N). @@ -282,6 +284,8 @@ yysuf(List, N) -> lists:nthtail(N, List). %% Line has been updated with respect to newlines in the prefix of %% Chars consisting of (TokenLength - AcceptLength) characters. +-compile({nowarn_unused_function, adjust_line/4}). + adjust_line(N, N, _Cs, L) -> L; adjust_line(T, A, [$\n|Cs], L) -> adjust_line(T-1, A, Cs, L-1); diff --git a/lib/public_key/doc/src/public_key_records.xml b/lib/public_key/doc/src/public_key_records.xml index 739310c88b..9ebdbb244d 100644 --- a/lib/public_key/doc/src/public_key_records.xml +++ b/lib/public_key/doc/src/public_key_records.xml @@ -70,10 +70,10 @@ <p><c>| {dNSName, string()}</c></p> <p><c>| {x400Address, string()}</c></p> <p><c>| {directoryName, {rdnSequence, [#AttributeTypeAndValue'{}]}}</c></p> - <p><c>| {eidPartyName, special_string()}</c></p> - <p><c>| {eidPartyName, special_string(), special_string()}</c></p> + <p><c>| {ediPartyName, special_string()}</c></p> + <p><c>| {ediPartyName, special_string(), special_string()}</c></p> <p><c>| {uniformResourceIdentifier, string()}</c></p> - <p><c>| {ipAddress, string()}</c></p> + <p><c>| {iPAddress, string()}</c></p> <p><c>| {registeredId, oid()}</c></p> <p><c>| {otherName, term()}</c></p> </item> diff --git a/lib/public_key/src/pubkey_cert.erl b/lib/public_key/src/pubkey_cert.erl index 13833830a7..76fd0f8133 100644 --- a/lib/public_key/src/pubkey_cert.erl +++ b/lib/public_key/src/pubkey_cert.erl @@ -1144,7 +1144,7 @@ issuer(Contact, Role, Name) -> subject(Contact, Role ++ Name). subject(Contact, Name) -> - Opts = [{email, Contact ++ "@erlang.org"}, + Opts = [{email, Contact ++ "@example.org"}, {name, Name}, {city, "Stockholm"}, {country, "SE"}, @@ -1223,12 +1223,12 @@ cert_chain(Role, IssuerCert, IssuerKey, [PeerOpts], _, Acc) -> Key = gen_key(proplists:get_value(key, PeerOpts, default_key_gen())), Cert = cert(Role, public_key:pkix_decode_cert(IssuerCert, otp), IssuerKey, Key, "admin", " Peer cert", PeerOpts, peer), - [{Cert, Key}, {IssuerCert, IssuerKey} | Acc]; + [{Cert, encode_key(Key)}, {IssuerCert, encode_key(IssuerKey)} | Acc]; cert_chain(Role, IssuerCert, IssuerKey, [CAOpts | Rest], N, Acc) -> Key = gen_key(proplists:get_value(key, CAOpts, default_key_gen())), Cert = cert(Role, public_key:pkix_decode_cert(IssuerCert, otp), IssuerKey, Key, "webadmin", " Intermidiate CA " ++ integer_to_list(N), CAOpts, ca), - cert_chain(Role, Cert, Key, Rest, N+1, [{IssuerCert, IssuerKey} | Acc]). + cert_chain(Role, Cert, Key, Rest, N+1, [{IssuerCert, encode_key(IssuerKey)} | Acc]). cert(Role, #'OTPCertificate'{tbsCertificate = #'OTPTBSCertificate'{subject = Issuer}}, PrivKey, Key, Contact, Name, Opts, Type) -> @@ -1311,3 +1311,10 @@ add_default_extensions(Defaults0, Exts) -> end, Defaults0), Exts ++ Defaults. +encode_key(#'RSAPrivateKey'{} = Key) -> + {'RSAPrivateKey', public_key:der_encode('RSAPrivateKey', Key)}; +encode_key(#'ECPrivateKey'{} = Key) -> + {'ECPrivateKey', public_key:der_encode('ECPrivateKey', Key)}; +encode_key(#'DSAPrivateKey'{} = Key) -> + {'DSAPrivateKey', public_key:der_encode('DSAPrivateKey', Key)}. + diff --git a/lib/public_key/src/public_key.erl b/lib/public_key/src/public_key.erl index c6ab4d06ae..6788c1ee92 100644 --- a/lib/public_key/src/public_key.erl +++ b/lib/public_key/src/public_key.erl @@ -850,10 +850,10 @@ pkix_crls_validate(OtpCert, DPAndCRLs0, Options) -> %-------------------------------------------------------------------- -spec pkix_verify_hostname(Cert :: #'OTPCertificate'{} | binary(), - ReferenceIDs :: [{uri_id | dns_id | oid(), string()}]) -> boolean(). + ReferenceIDs :: [{uri_id | dns_id | ip | srv_id | oid(), string()}]) -> boolean(). -spec pkix_verify_hostname(Cert :: #'OTPCertificate'{} | binary(), - ReferenceIDs :: [{uri_id | dns_id | oid(), string()}], + ReferenceIDs :: [{uri_id | dns_id | ip | srv_id | oid(), string()}], Options :: proplists:proplist()) -> boolean(). %% Description: Validates a hostname to RFC 6125 @@ -1323,9 +1323,9 @@ ec_normalize_params(#'ECParameters'{} = ECParams) -> ec_normalize_params(Other) -> Other. -spec ec_curve_spec(ecpk_parameters_api()) -> term(). -ec_curve_spec( #'ECParameters'{fieldID = FieldId, curve = PCurve, base = Base, order = Order, cofactor = CoFactor }) -> - Field = {pubkey_cert_records:supportedCurvesTypes(FieldId#'FieldID'.fieldType), - FieldId#'FieldID'.parameters}, +ec_curve_spec( #'ECParameters'{fieldID = #'FieldID'{fieldType = Type, + parameters = Params}, curve = PCurve, base = Base, order = Order, cofactor = CoFactor }) -> + Field = format_field(pubkey_cert_records:supportedCurvesTypes(Type), Params), Curve = {PCurve#'Curve'.a, PCurve#'Curve'.b, none}, {Field, Curve, Base, Order, CoFactor}; ec_curve_spec({ecParameters, ECParams}) -> @@ -1335,6 +1335,26 @@ ec_curve_spec({namedCurve, OID}) when is_tuple(OID), is_integer(element(1,OID)) ec_curve_spec({namedCurve, Name}) when is_atom(Name) -> crypto:ec_curve(Name). +format_field(characteristic_two_field = Type, Params0) -> + #'Characteristic-two'{ + m = M, + basis = BasisOid, + parameters = Params} = der_decode('Characteristic-two', Params0), + {Type, M, field_param_decode(BasisOid, Params)}; +format_field(prime_field, Params0) -> + Prime = der_decode('Prime-p', Params0), + {prime_field, Prime}. + +field_param_decode(?'ppBasis', Params) -> + #'Pentanomial'{k1 = K1, k2 = K2, k3 = K3} = + der_decode('Pentanomial', Params), + {ppbasis, K1, K2, K3}; +field_param_decode(?'tpBasis', Params) -> + K = der_decode('Trinomial', Params), + {tpbasis, K}; +field_param_decode(?'gnBasis', _) -> + onbasis. + -spec ec_key({PubKey::term(), PrivateKey::term()}, Params::ecpk_parameters()) -> #'ECPrivateKey'{}. ec_key({PubKey, PrivateKey}, Params) -> #'ECPrivateKey'{version = 1, diff --git a/lib/public_key/test/public_key_SUITE.erl b/lib/public_key/test/public_key_SUITE.erl index 0100f0a912..38e8f30a25 100644 --- a/lib/public_key/test/public_key_SUITE.erl +++ b/lib/public_key/test/public_key_SUITE.erl @@ -64,7 +64,7 @@ all() -> groups() -> [{pem_decode_encode, [], [dsa_pem, rsa_pem, ec_pem, encrypted_pem, dh_pem, cert_pem, pkcs7_pem, pkcs10_pem, ec_pem2, - ec_pem_encode_generated]}, + ec_pem_encode_generated, gen_ec_param]}, {ssh_public_key_decode_encode, [], [ssh_rsa_public_key, ssh_dsa_public_key, ssh_ecdsa_public_key, ssh_rfc4716_rsa_comment, ssh_rfc4716_dsa_comment, @@ -102,8 +102,22 @@ init_per_testcase(pkix_test_data_all_default, Config) -> [] -> {skip, missing_ecc_support}; _ -> - init_common_per_testcase(Config) + init_common_per_testcase(Config) end; + +init_per_testcase(gen_ec_param, Config) -> + case crypto:ec_curves() of + [] -> + {skip, missing_ecc_support}; + Curves -> + case lists:member(secp521r1, Curves) of + true -> + init_common_per_testcase(Config); + false -> + {skip, missing_ecc_secp52r1_support} + end + end; + init_per_testcase(TestCase, Config) -> case TestCase of ssh_hostkey_fingerprint_md5_implicit -> init_fingerprint_testcase([md5], Config); @@ -1206,7 +1220,13 @@ short_crl_issuer_hash(Config) when is_list(Config) -> Issuer = public_key:pkix_crl_issuer(CrlDER), CrlIssuerHash = public_key:short_name_hash(Issuer). - +%%-------------------------------------------------------------------- +gen_ec_param() -> + [{doc, "Generate key with EC parameters"}]. +gen_ec_param(Config) when is_list(Config) -> + Datadir = proplists:get_value(data_dir, Config), + do_gen_ec_param(filename:join(Datadir, "ec_key_param0.pem")), + do_gen_ec_param(filename:join(Datadir, "ec_key_param1.pem")). %%-------------------------------------------------------------------- %% Internal functions ------------------------------------------------ @@ -1279,6 +1299,19 @@ strip_superfluous_newlines(Bin) -> Str = string:strip(binary_to_list(Bin), right, 10), re:replace(Str,"\n\n","\n", [{return,list}, global]). +do_gen_ec_param(File) -> + {ok, KeyPem} = file:read_file(File), + Entries = public_key:pem_decode(KeyPem), + [ParamInfo] = [Entry || Entry={'EcpkParameters', _, not_encrypted} <- Entries], + {ecParameters, Params} = public_key:pem_entry_decode(ParamInfo), + Key = public_key:generate_key(Params), + case check_entry_type(Key, 'ECPrivateKey') of + true -> + ok; + false -> + ct:fail({key_gen_fail, File}) + end. + incorrect_countryname_pkix_cert() -> <<48,130,5,186,48,130,4,162,160,3,2,1,2,2,7,7,250,61,63,6,140,137,48,13,6,9,42, 134,72,134,247,13,1,1,5,5,0,48,129,220,49,11,48,9,6,3,85,4,6,19,2,85,83,49, 16,48,14,6,3,85,4,8,19,7,65,114,105,122,111,110,97,49,19,48,17,6,3,85,4,7,19, 10,83,99,111,116,116,115,100,97,108,101,49,37,48,35,6,3,85,4,10,19,28,83,116, 97,114,102,105,101,108,100,32,84,101,99,104,110,111,108,111,103,105,101,115, 44,32,73,110,99,46,49,57,48,55,6,3,85,4,11,19,48,104,116,116,112,58,47,47,99, 101,114,116,105,102,105,99,97,116,101,115,46,115,116,97,114,102,105,101,108, 100,116,101,99,104,46,99,111,109,47,114,101,112,111,115,105,116,111,114,121, 49,49,48,47,6,3,85,4,3,19,40,83,116,97,114,102,105,101,108,100,32,83,101,99, 117,114,101,32,67,101,114,116,105,102,105,99,97,116,105,111,110,32,65,117, 116,104,111,114,105,116,121,49,17,48,15,6,3,85,4,5,19,8,49,48,54,56,56,52,51, 53,48,30,23,13,49,48,49,48,50,51,48,49,51,50,48,53,90,23,13,49,50,49,48,50, 51,48,49,51,50,48,53,90,48,122,49,11,48,9,6,3,85,4,6,12,2,85,83,49,11,48,9,6, 3,85,4,8,12,2,65,90,49,19,48,17,6,3,85,4,7,12,10,83,99,111,116,116,115,100, 97,108,101,49,38,48,36,6,3,85,4,10,12,29,83,112,101,99,105,97,108,32,68,111, 109,97,105,110,32,83,101,114,118,105,99,101,115,44,32,73,110,99,46,49,33,48, 31,6,3,85,4,3,12,24,42,46,108,111,103,105,110,46,115,101,99,117,114,101,115, 101,114,118,101,114,46,110,101,116,48,130,1,34,48,13,6,9,42,134,72,134,247, 13,1,1,1,5,0,3,130,1,15,0,48,130,1,10,2,130,1,1,0,185,136,240,80,141,36,124, 245,182,130,73,19,188,74,166,117,72,228,185,209,43,129,244,40,44,193,231,11, 209,12,234,88,43,142,1,162,48,122,17,95,230,105,171,131,12,147,46,204,36,80, 250,171,33,253,35,62,83,22,71,212,186,141,14,198,89,89,121,204,224,122,246, 127,110,188,229,162,67,95,6,74,231,127,99,131,7,240,85,102,203,251,50,58,58, 104,245,103,181,183,134,32,203,121,232,54,32,188,139,136,112,166,126,14,91, 223,153,172,164,14,61,38,163,208,215,186,210,136,213,143,70,147,173,109,217, 250,169,108,31,211,104,238,103,93,182,59,165,43,196,189,218,241,30,148,240, 109,90,69,176,194,52,116,173,151,135,239,10,209,179,129,192,102,75,11,25,168, 223,32,174,84,223,134,70,167,55,172,143,27,130,123,226,226,7,34,142,166,39, 48,246,96,231,150,84,220,106,133,193,55,95,159,227,24,249,64,36,1,142,171,16, 202,55,126,7,156,15,194,22,116,53,113,174,104,239,203,120,45,131,57,87,84, 163,184,27,83,57,199,91,200,34,43,98,61,180,144,76,65,170,177,2,3,1,0,1,163, 130,1,224,48,130,1,220,48,15,6,3,85,29,19,1,1,255,4,5,48,3,1,1,0,48,29,6,3, 85,29,37,4,22,48,20,6,8,43,6,1,5,5,7,3,1,6,8,43,6,1,5,5,7,3,2,48,14,6,3,85, 29,15,1,1,255,4,4,3,2,5,160,48,56,6,3,85,29,31,4,49,48,47,48,45,160,43,160, 41,134,39,104,116,116,112,58,47,47,99,114,108,46,115,116,97,114,102,105,101, 108,100,116,101,99,104,46,99,111,109,47,115,102,115,50,45,48,46,99,114,108, 48,83,6,3,85,29,32,4,76,48,74,48,72,6,11,96,134,72,1,134,253,110,1,7,23,2,48, 57,48,55,6,8,43,6,1,5,5,7,2,1,22,43,104,116,116,112,115,58,47,47,99,101,114, 116,115,46,115,116,97,114,102,105,101,108,100,116,101,99,104,46,99,111,109, 47,114,101,112,111,115,105,116,111,114,121,47,48,129,141,6,8,43,6,1,5,5,7,1, 1,4,129,128,48,126,48,42,6,8,43,6,1,5,5,7,48,1,134,30,104,116,116,112,58,47, 47,111,99,115,112,46,115,116,97,114,102,105,101,108,100,116,101,99,104,46,99, 111,109,47,48,80,6,8,43,6,1,5,5,7,48,2,134,68,104,116,116,112,58,47,47,99, 101,114,116,105,102,105,99,97,116,101,115,46,115,116,97,114,102,105,101,108, 100,116,101,99,104,46,99,111,109,47,114,101,112,111,115,105,116,111,114,121, 47,115,102,95,105,110,116,101,114,109,101,100,105,97,116,101,46,99,114,116, 48,31,6,3,85,29,35,4,24,48,22,128,20,73,75,82,39,209,27,188,242,161,33,106, 98,123,81,66,122,138,215,213,86,48,59,6,3,85,29,17,4,52,48,50,130,24,42,46, 108,111,103,105,110,46,115,101,99,117,114,101,115,101,114,118,101,114,46,110, 101,116,130,22,108,111,103,105,110,46,115,101,99,117,114,101,115,101,114,118, 101,114,46,110,101,116,48,29,6,3,85,29,14,4,22,4,20,138,233,191,208,157,203, 249,85,242,239,20,195,48,10,148,49,144,101,255,116,48,13,6,9,42,134,72,134, 247,13,1,1,5,5,0,3,130,1,1,0,82,31,121,162,49,50,143,26,167,202,143,61,71, 189,201,199,57,81,122,116,90,192,88,24,102,194,174,48,157,74,27,87,210,223, 253,93,3,91,150,109,120,1,110,27,11,200,198,141,222,246,14,200,71,105,41,138, 13,114,122,106,63,17,197,181,234,121,61,89,74,65,41,231,248,219,129,83,176, 219,55,107,55,211,112,98,38,49,69,77,96,221,108,123,152,12,210,159,157,141, 43,226,55,187,129,3,82,49,136,66,81,196,91,234,196,10,82,48,6,80,163,83,71, 127,102,177,93,209,129,26,104,2,84,24,255,248,161,3,244,169,234,92,122,110, 43,4,17,113,185,235,108,219,210,236,132,216,177,227,17,169,58,162,159,182, 162,93,160,229,200,9,163,229,110,121,240,168,232,14,91,214,188,196,109,210, 164,222,0,109,139,132,113,91,16,118,173,178,176,80,132,34,41,199,51,206,250, 224,132,60,115,192,94,107,163,219,212,226,225,65,169,148,108,213,46,174,173, 103,110,189,229,166,149,254,31,51,44,144,108,187,182,11,251,201,206,86,138, 208,59,51,86,132,235,81,225,88,34,190,8,184>>. diff --git a/lib/public_key/test/public_key_SUITE_data/ec_key_param0.pem b/lib/public_key/test/public_key_SUITE_data/ec_key_param0.pem new file mode 100644 index 0000000000..679b08f1a2 --- /dev/null +++ b/lib/public_key/test/public_key_SUITE_data/ec_key_param0.pem @@ -0,0 +1,28 @@ +-----BEGIN EC PARAMETERS----- +MIIBwgIBATBNBgcqhkjOPQEBAkIB//////////////////////////////////// +//////////////////////////////////////////////////8wgZ4EQgH///// +//////////////////////////////////////////////////////////////// +/////////////////ARBUZU+uWGOHJofkpohoLaFQO6i2nJbmbMV87i0iZGO8Qnh +Vhk5Uex+k3sWUsC9O7G/BzVz34g9LDTx70Uf1GtQPwADFQDQnogAKRy4U5bMZxc5 +MoSqoNpkugSBhQQAxoWOBrcEBOnNnj7LZiOVtEKcZIE5BT+1Ifgor2BrTT26oUte +d+/nWSj+HcEnov+o3jNIs8GFakKb+X5+McLlvWYBGDkpaniaO8AEXIpftCx9G9mY +9URJV5tEaBevvRcnPmYsl+5ymV70JkDFULkBP60HYTU8cIaicsJAiL6Udp/RZlAC +QgH///////////////////////////////////////////pRhoeDvy+Wa3/MAUj3 +CaXQO7XJuImcR667b7cekThkCQIBAQ== +-----END EC PARAMETERS----- +-----BEGIN EC PRIVATE KEY----- +MIICnAIBAQRBP6XqBV/b2Q49a8RUvrJBwotzY+IErK5FkzgkExqJuyzXXMM3jMtd +M1vlEF46OCjbldw7NaYITW1mpwz6z2xtqyagggHGMIIBwgIBATBNBgcqhkjOPQEB +AkIB//////////////////////////////////////////////////////////// +//////////////////////////8wgZ4EQgH///////////////////////////// +/////////////////////////////////////////////////////////ARBUZU+ +uWGOHJofkpohoLaFQO6i2nJbmbMV87i0iZGO8QnhVhk5Uex+k3sWUsC9O7G/BzVz +34g9LDTx70Uf1GtQPwADFQDQnogAKRy4U5bMZxc5MoSqoNpkugSBhQQAxoWOBrcE +BOnNnj7LZiOVtEKcZIE5BT+1Ifgor2BrTT26oUted+/nWSj+HcEnov+o3jNIs8GF +akKb+X5+McLlvWYBGDkpaniaO8AEXIpftCx9G9mY9URJV5tEaBevvRcnPmYsl+5y +mV70JkDFULkBP60HYTU8cIaicsJAiL6Udp/RZlACQgH///////////////////// +//////////////////////pRhoeDvy+Wa3/MAUj3CaXQO7XJuImcR667b7cekThk +CQIBAaGBiQOBhgAEAVtCjzs+HP67ZZheraLGJPY+iIJHwDYWeCyzn2J4/fOv5CTo +x0+1QcjrECh4V4F6jfqwW/oQaG9KKyEWhvvWQ0yVAeBfE+89DhYB3h2kyelqRcwZ +XUuB7n7TxijJiHJXq9b+u+sh0qH9ya6nbWjHk79V37pgQOegHjnRc1gIFz6IfRUa +-----END EC PRIVATE KEY----- diff --git a/lib/public_key/test/public_key_SUITE_data/ec_key_param1.pem b/lib/public_key/test/public_key_SUITE_data/ec_key_param1.pem new file mode 100644 index 0000000000..67095d8dc5 --- /dev/null +++ b/lib/public_key/test/public_key_SUITE_data/ec_key_param1.pem @@ -0,0 +1,25 @@ +-----BEGIN EC PARAMETERS----- +MIIBcQIBATAlBgcqhkjOPQECMBoCAgI7BgkqhkjOPQECAwMwCQIBAgIBBQIBCjBk +BAEBBEgC9A5+IiHyld4pcRe389YvXGqX/8uM7/HNa6jOSpoYrYT/q72O+lkzK+et +Z1ambilK/RhaeP8SqlIOTec5usoMf/7/fylVcnoDFQAqoFj3Og4zq0hrD2EEEMU6 +fxMjEASBkQQDAwAdNLhWKWwWwNQNPNd1CpPR0pVfqAql9A/I23sqvb3lOVD0wNKT +zdcRo1tn+xSZrmADhhTxOUq/o7TIUNkn4ed2nI7sLRkDe/JzQtpjm23M//63PWnX +jGwnpgCcu8oZgPhTOSHopoRCPkO6sIpXYpGvj0YbsqizUx0vBIXBmxbi8VFuI908 +GkgnrxuKwVsCSAP//////////////////////////////////////////////+Zh +zhj/VZhzCAWbGGgjhR7H3ZyhFh3pPVF01m6Dgum7L+hORwIBAg== +-----END EC PARAMETERS----- +-----BEGIN EC PRIVATE KEY----- +MIICXgIBAQRIAwaM3VblR1nZJseRVx4qJBQLLulB7uMF9KE+hMPlzZjj/f3KbcK4 +dfhxYFf4gqWNJW3fzFMEYF7JpzDjuFTmfX5jjO5f8cw9oIIBdTCCAXECAQEwJQYH +KoZIzj0BAjAaAgICOwYJKoZIzj0BAgMDMAkCAQICAQUCAQowZAQBAQRIAvQOfiIh +8pXeKXEXt/PWL1xql//LjO/xzWuozkqaGK2E/6u9jvpZMyvnrWdWpm4pSv0YWnj/ +EqpSDk3nObrKDH/+/38pVXJ6AxUAKqBY9zoOM6tIaw9hBBDFOn8TIxAEgZEEAwMA +HTS4VilsFsDUDTzXdQqT0dKVX6gKpfQPyNt7Kr295TlQ9MDSk83XEaNbZ/sUma5g +A4YU8TlKv6O0yFDZJ+HndpyO7C0ZA3vyc0LaY5ttzP/+tz1p14xsJ6YAnLvKGYD4 +Uzkh6KaEQj5DurCKV2KRr49GG7Kos1MdLwSFwZsW4vFRbiPdPBpIJ68bisFbAkgD +///////////////////////////////////////////////mYc4Y/1WYcwgFmxho +I4Uex92coRYd6T1RdNZug4Lpuy/oTkcCAQKhgZUDgZIABALOIIukF443IxnNZx1Q +4HKDifQ9Lj3Rh+inIKczXLLJNEGCJ0wwG3d4v/fOxzdepZcnYviFAijQQA/iYWsA +/Zet+5B4yxKISQcHei+PXkqwAAwEqq6D4hO2orlNOO430rgLoA0MNFc7I1THanOp +q4RhYp8qnCEjM7nfQ4R0F+hPgZK1VPojkrGolaY/cNT/oA== +-----END EC PRIVATE KEY----- diff --git a/lib/sasl/src/format_lib_supp.erl b/lib/sasl/src/format_lib_supp.erl index cfe2ec7668..2d37dfe117 100644 --- a/lib/sasl/src/format_lib_supp.erl +++ b/lib/sasl/src/format_lib_supp.erl @@ -86,8 +86,10 @@ print_data(Device, Line, [{Key, Value}|T]) -> print_data(Device, Line, [Value|T]) -> Modifier = misc_supp:modifier(Device), io:format(Device, "~"++Modifier++"p~n", [Value]), - print_data(Device, Line, T). - + print_data(Device, Line, T); +print_data(Device, _Line, Value) -> + Modifier = misc_supp:modifier(Device), + io:format(Device, "~"++Modifier++"p~n", [Value]). print_items(Device, Line, {Name, Items}) -> print_items(Device, Line, Name, Items). diff --git a/lib/sasl/test/rb_SUITE.erl b/lib/sasl/test/rb_SUITE.erl index 426dedbab5..4ba2540545 100644 --- a/lib/sasl/test/rb_SUITE.erl +++ b/lib/sasl/test/rb_SUITE.erl @@ -36,6 +36,7 @@ no_group_cases() -> groups() -> [{running_error_logger,[shuffle],[show, + show_other, list, rescan, start_stop_log, @@ -165,6 +166,23 @@ show(Config) -> ok. +show_other(Config) -> + PrivDir = ?config(priv_dir,Config), + OutFile = filename:join(PrivDir,"rb_SUITE_log.txt"), + + %% Insert some reports in the error log and start rb + error_logger:info_report([rb_test_term_in_list]), + error_logger:info_report(rb_test_term_no_list), + ok = start_rb(OutFile), + + %% Show by type and check content + [{_,I1},{_,I2}] = check_report(fun() -> rb:show(info_report) end,OutFile), + + true = contains(I1,"rb_test_term_no_list"), + true = contains(I2,"rb_test_term_in_list"), + + ok. + list(Config) -> PrivDir = ?config(priv_dir,Config), OutFile = filename:join(PrivDir,"rb_SUITE_log.txt"), diff --git a/lib/snmp/doc/src/notes.xml b/lib/snmp/doc/src/notes.xml index 4705804759..6bdcae5dd7 100644 --- a/lib/snmp/doc/src/notes.xml +++ b/lib/snmp/doc/src/notes.xml @@ -34,7 +34,23 @@ </header> - <section><title>SNMP 5.2.7</title> + <section><title>SNMP 5.2.8</title> + + <section><title>Fixed Bugs and Malfunctions</title> + <list> + <item> + <p> + The <c>recbuf</c> configuration option was not propagated + correctly to the socket for the SNMP Manager.</p> + <p> + Own Id: OTP-13372 Aux Id: ERIERL-73 </p> + </item> + </list> + </section> + +</section> + +<section><title>SNMP 5.2.7</title> <section><title>Fixed Bugs and Malfunctions</title> <list> diff --git a/lib/snmp/src/app/snmp.appup.src b/lib/snmp/src/app/snmp.appup.src index bde637744c..ca61782639 100644 --- a/lib/snmp/src/app/snmp.appup.src +++ b/lib/snmp/src/app/snmp.appup.src @@ -8,19 +8,6 @@ %% {update, snmpa_local_db, soft, soft_purge, soft_purge, []} %% {add_module, snmpm_net_if_mt} [ - {<<"5\\.2\\.6">>, - [{load_module, snmpc, soft_purge, soft_purge, []}, - {load_module, snmpc_lib, soft_purge, soft_purge, []}]}, - {<<"5\\.2\\.5">>, - [{load_module, snmpc, soft_purge, soft_purge, []}, - {load_module, snmpc_lib, soft_purge, soft_purge, []}, - {load_module, snmp_generic, soft_purge, soft_purge, []}]}, - {<<"5\\.2\\.4">>, - [{load_module, snmpc, soft_purge, soft_purge, []}, - {load_module, snmp_generic, soft_purge, soft_purge, []}, - {load_module, snmp, soft_purge, soft_purge, []}, - {load_module, snmpc_lib, soft_purge, soft_purge, []}, - {load_module, snmpc_mib_gram, soft_purge, soft_purge, []}]}, {<<"5\\..*">>, [{restart_application, snmp}]}, {<<"4\\..*">>, [{restart_application, snmp}]} ], @@ -30,19 +17,6 @@ %% {remove, {snmpm_net_if_mt, soft_purge, soft_purge}} [ - {<<"5\\.2\\.6">>, - [{load_module, snmpc, soft_purge, soft_purge, []}, - {load_module, snmpc_lib, soft_purge, soft_purge, []}]}, - {<<"5\\.2\\.5">>, - [{load_module, snmpc, soft_purge, soft_purge, []}, - {load_module, snmpc_lib, soft_purge, soft_purge, []}, - {load_module, snmp_generic, soft_purge, soft_purge, []}]}, - {<<"5\\.2\\.4">>, - [{load_module, snmpc, soft_purge, soft_purge, []}, - {load_module, snmp_generic, soft_purge, soft_purge, []}, - {load_module, snmp, soft_purge, soft_purge, []}, - {load_module, snmpc_lib, soft_purge, soft_purge, []}, - {load_module, snmpc_mib_gram, soft_purge, soft_purge, []}]}, {<<"5\\..*">>, [{restart_application, snmp}]}, {<<"4\\..*">>, [{restart_application, snmp}]} ] diff --git a/lib/snmp/src/manager/snmpm_net_if.erl b/lib/snmp/src/manager/snmpm_net_if.erl index 93c987eb0f..29216f9d6a 100644 --- a/lib/snmp/src/manager/snmpm_net_if.erl +++ b/lib/snmp/src/manager/snmpm_net_if.erl @@ -1,7 +1,7 @@ %% %% %CopyrightBegin% %% -%% Copyright Ericsson AB 2004-2015. All Rights Reserved. +%% Copyright Ericsson AB 2004-2017. All Rights Reserved. %% %% Licensed under the Apache License, Version 2.0 (the "License"); %% you may not use this file except in compliance with the License. @@ -367,7 +367,7 @@ common_socket_opts(Opts) -> default -> []; Sz -> - [{sndbuf, Sz}] + [{recbuf, Sz}] end ++ case get_opt(Opts, no_reuse, false) of false -> diff --git a/lib/snmp/test/snmp_agent_test.erl b/lib/snmp/test/snmp_agent_test.erl index 3c1a6f2afd..2ed2c4580c 100644 --- a/lib/snmp/test/snmp_agent_test.erl +++ b/lib/snmp/test/snmp_agent_test.erl @@ -605,7 +605,12 @@ init_per_group(multiple_reqs_3 = GroupName, Config) -> init_per_group(test_multi_threaded = GroupName, Config) -> init_mt(snmp_test_lib:init_group_top_dir(GroupName, Config)); init_per_group(test_v3 = GroupName, Config) -> - init_v3(snmp_test_lib:init_group_top_dir(GroupName, Config)); + case snmp_test_lib:crypto_start() of + ok -> + init_v3(snmp_test_lib:init_group_top_dir(GroupName, Config)); + _ -> + {skip, "Crypto did not start"} + end; init_per_group(test_v1_v2 = GroupName, Config) -> init_v1_v2(snmp_test_lib:init_group_top_dir(GroupName, Config)); init_per_group(test_v2 = GroupName, Config) -> @@ -631,11 +636,26 @@ init_per_group(mib_storage_varm_dets = GroupName, Config) -> init_varm_mib_storage_dets( snmp_test_lib:init_group_top_dir(GroupName, Config)); init_per_group(mib_storage_size_check_mnesia = GroupName, Config) -> - init_size_check_msm(snmp_test_lib:init_group_top_dir(GroupName, Config)); + case snmp_test_lib:crypto_start() of + ok -> + init_size_check_msm(snmp_test_lib:init_group_top_dir(GroupName, Config)); + _ -> + {skip, "Crypto did not start"} + end; init_per_group(mib_storage_size_check_dets = GroupName, Config) -> - init_size_check_msd(snmp_test_lib:init_group_top_dir(GroupName, Config)); + case snmp_test_lib:crypto_start() of + ok -> + init_size_check_msm(snmp_test_lib:init_group_top_dir(GroupName, Config)); + _ -> + {skip, "Crypto did not start"} + end; init_per_group(mib_storage_size_check_ets = GroupName, Config) -> - init_size_check_mse(snmp_test_lib:init_group_top_dir(GroupName, Config)); + case snmp_test_lib:crypto_start() of + ok -> + init_size_check_msm(snmp_test_lib:init_group_top_dir(GroupName, Config)); + _ -> + {skip, "Crypto did not start"} + end; init_per_group(mib_storage_mnesia = GroupName, Config) -> init_mib_storage_mnesia(snmp_test_lib:init_group_top_dir(GroupName, Config)); diff --git a/lib/snmp/test/snmp_manager_test.erl b/lib/snmp/test/snmp_manager_test.erl index 4bfeb0f8d1..6ced55f0cc 100644 --- a/lib/snmp/test/snmp_manager_test.erl +++ b/lib/snmp/test/snmp_manager_test.erl @@ -156,16 +156,25 @@ init_per_suite(Config0) when is_list(Config0) -> ?DBG("init_per_suite -> entry with" "~n Config0: ~p", [Config0]), - Config1 = snmp_test_lib:init_suite_top_dir(?MODULE, Config0), - Config2 = snmp_test_lib:fix_data_dir(Config1), - - %% Mib-dirs - %% data_dir is trashed by the test-server / common-test - %% so there is no point in fixing it... - MibDir = snmp_test_lib:lookup(data_dir, Config2), - StdMibDir = filename:join([code:priv_dir(snmp), "mibs"]), - - [{mib_dir, MibDir}, {std_mib_dir, StdMibDir} | Config2]. + %% Preferably this test SUITE should be divided into groups + %% so that if crypto does not work only v3 tests that + %% need crypto will be skipped, but as this is only a + %% problem with one legacy test machine, we will procrastinate + %% until we have a more important reason to fix this. + case snmp_test_lib:crypto_start() of + ok -> + Config1 = snmp_test_lib:init_suite_top_dir(?MODULE, Config0), + Config2 = snmp_test_lib:fix_data_dir(Config1), + %% Mib-dirs + %% data_dir is trashed by the test-server / common-test + %% so there is no point in fixing it... + MibDir = snmp_test_lib:lookup(data_dir, Config2), + StdMibDir = filename:join([code:priv_dir(snmp), "mibs"]), + + [{mib_dir, MibDir}, {std_mib_dir, StdMibDir} | Config2]; + _ -> + {skip, "Crypto did not start"} + end. end_per_suite(Config) when is_list(Config) -> diff --git a/lib/snmp/test/snmp_to_snmpnet_SUITE.erl b/lib/snmp/test/snmp_to_snmpnet_SUITE.erl index 24c14d86ea..6a3466b6e4 100644 --- a/lib/snmp/test/snmp_to_snmpnet_SUITE.erl +++ b/lib/snmp/test/snmp_to_snmpnet_SUITE.erl @@ -88,8 +88,17 @@ groups() -> ]. init_per_suite(Config) -> - [{agent_port, ?AGENT_PORT}, {manager_port, ?MANAGER_PORT} | Config]. - + case re:run(os:cmd("snmpd -v"),"NET-SNMP", [{capture, first}]) of + nomatch -> + {skip, "snmpd is NOT NET-SNMP"}; + {match, _} -> + case re:run(os:cmd("snmpd -v"),"5.4|5.6.2.1", [{capture, first}]) of + nomatch -> + [{agent_port, ?AGENT_PORT}, {manager_port, ?MANAGER_PORT} | Config]; + {match, _} -> + {skip, "buggy snmpd"} + end + end. end_per_suite(_Config) -> ok. @@ -322,7 +331,7 @@ snmpget(Oid, Transport, Config) -> Args = ["-c", "public", net_snmp_version(Versions), - "-m", "", + "-m", ":", "-Cf", net_snmp_addr_str(Transport), oid_str(Oid)], @@ -353,11 +362,13 @@ start_snmpd(Community, SysDescr, Config) -> ["--rocommunity"++domain_suffix(Domain)++"=" ++Community++" "++inet_parse:ntoa(Ip) || {Domain, {Ip, _}} <- Targets], + SnmpdArgs = - ["-f", "-r", %"-Dverbose", - "-c", filename:join(DataDir, "snmpd.conf"), - "-C", "-Lo", - "-m", "", + ["-f", "-r", %"-Dverbose", + "-c", filename:join(DataDir, "snmpd.conf"), + "-C", + "-Lo", + "-m", ":", "--sysDescr="++SysDescr, "--agentXSocket=tcp:localhost:"++integer_to_list(Port)] ++ CommunityArgs diff --git a/lib/snmp/vsn.mk b/lib/snmp/vsn.mk index 207f0084d8..ef48608bda 100644 --- a/lib/snmp/vsn.mk +++ b/lib/snmp/vsn.mk @@ -19,6 +19,6 @@ # %CopyrightEnd% APPLICATION = snmp -SNMP_VSN = 5.2.7 +SNMP_VSN = 5.2.8 PRE_VSN = APP_VSN = "$(APPLICATION)-$(SNMP_VSN)$(PRE_VSN)" diff --git a/lib/ssh/doc/src/ssh_client_key_api.xml b/lib/ssh/doc/src/ssh_client_key_api.xml index a1cd9d4b02..98a1676ca4 100644 --- a/lib/ssh/doc/src/ssh_client_key_api.xml +++ b/lib/ssh/doc/src/ssh_client_key_api.xml @@ -56,11 +56,17 @@ <tag><c>string() =</c></tag> <item><p><c>[byte()]</c></p></item> <tag><c>public_key() =</c></tag> - <item><p><c>#'RSAPublicKey'{}| {integer(), #'Dss-Parms'{}}| term()</c></p></item> + <item><p><c>#'RSAPublicKey'{} + | {integer(),#'Dss-Parms'{}} + | {#'ECPoint'{},{namedCurve,Curve::string()}}</c></p></item> <tag><c>private_key() =</c></tag> - <item><p><c>#'RSAPrivateKey'{} | #'DSAPrivateKey'{} | term()</c></p></item> + <item><p><c>#'RSAPrivateKey'{} + | #'DSAPrivateKey'{} + | #'ECPrivateKey'{}</c></p></item> <tag><c>public_key_algorithm() =</c></tag> - <item><p><c>'ssh-rsa'| 'ssh-dss' | atom()</c></p></item> + <item><p><c>'ssh-rsa' | 'ssh-dss' + | 'rsa-sha2-256' | 'rsa-sha2-384' | 'rsa-sha2-512' + | 'ecdsa-sha2-nistp256' | 'ecdsa-sha2-nistp384' | 'ecdsa-sha2-nistp521' </c></p></item> </taglist> </section> @@ -73,10 +79,11 @@ <d>Description of the host that owns the <c>PublicKey</c>.</d> <v>Key = public_key()</v> - <d>Normally an RSA or DSA public key, but handling of other public keys can be added.</d> + <d>Normally an RSA, DSA or ECDSA public key, but handling of other public keys can be added.</d> <v>ConnectOptions = proplists:proplist()</v> - <d>Options provided to <seealso marker="ssh#connect-3">ssh:connect/[3,4]</seealso></d> + <d>Options provided to <seealso marker="ssh#connect-3">ssh:connect/[3,4]</seealso>. The option list given in + the <c>key_cb</c> option is available with the key <c>key_cb_private</c>.</d> <v>Reason = term().</v> </type> <desc> @@ -89,17 +96,17 @@ <fsummary>Checks if a host key is trusted.</fsummary> <type> <v>Key = public_key() </v> - <d>Normally an RSA or DSA public key, but handling of other public keys can be added.</d> + <d>Normally an RSA, DSA or ECDSA public key, but handling of other public keys can be added.</d> <v>Host = string()</v> <d>Description of the host.</d> <v>Algorithm = public_key_algorithm()</v> - <d>Host key algorithm. Is to support <c>'ssh-rsa'| 'ssh-dss'</c>, but more algorithms - can be handled.</d> + <d>Host key algorithm.</d> <v>ConnectOptions = proplists:proplist() </v> - <d>Options provided to <seealso marker="ssh#connect-3">ssh:connect/[3,4]</seealso>.</d> + <d>Options provided to <seealso marker="ssh#connect-3">ssh:connect/[3,4]</seealso>. The option list given in + the <c>key_cb</c> option is available with the key <c>key_cb_private</c>.</d> <v>Result = boolean()</v> </type> @@ -110,15 +117,15 @@ <func> <name>Module:user_key(Algorithm, ConnectOptions) -> - {ok, PrivateKey} | {error, Reason}</name> + {ok, PrivateKey} | {error, Reason}</name> <fsummary>Fetches the users <em>public key</em> matching the <c>Algorithm</c>.</fsummary> <type> <v>Algorithm = public_key_algorithm()</v> - <d>Host key algorithm. Is to support <c>'ssh-rsa'| 'ssh-dss'</c> but more algorithms - can be handled.</d> + <d>Host key algorithm.</d> <v>ConnectOptions = proplists:proplist()</v> - <d>Options provided to <seealso marker="ssh#connect-3">ssh:connect/[3,4]</seealso></d> + <d>Options provided to <seealso marker="ssh#connect-3">ssh:connect/[3,4]</seealso>. The option list given in + the <c>key_cb</c> option is available with the key <c>key_cb_private</c>.</d> <v>PrivateKey = private_key()</v> <d>Private key of the user matching the <c>Algorithm</c>.</d> diff --git a/lib/ssh/doc/src/ssh_server_key_api.xml b/lib/ssh/doc/src/ssh_server_key_api.xml index a0694ca8d9..c6808b95d1 100644 --- a/lib/ssh/doc/src/ssh_server_key_api.xml +++ b/lib/ssh/doc/src/ssh_server_key_api.xml @@ -57,11 +57,17 @@ <tag><c>string() =</c></tag> <item><p><c>[byte()]</c></p></item> <tag><c>public_key() =</c></tag> - <item><p><c>#'RSAPublicKey'{}| {integer(), #'Dss-Parms'{}}| term()</c></p></item> + <item><p><c>#'RSAPublicKey'{} + | {integer(),#'Dss-Parms'{}} + | {#'ECPoint'{},{namedCurve,Curve::string()}}</c></p></item> <tag><c>private_key() =</c></tag> - <item><p><c>#'RSAPrivateKey'{} | #'DSAPrivateKey'{} | term()</c></p></item> + <item><p><c>#'RSAPrivateKey'{} + | #'DSAPrivateKey'{} + | #'ECPrivateKey'{}</c></p></item> <tag><c>public_key_algorithm() =</c></tag> - <item><p><c>'ssh-rsa'| 'ssh-dss' | atom()</c></p></item> + <item><p><c>'ssh-rsa' | 'ssh-dss' + | 'rsa-sha2-256' | 'rsa-sha2-384' | 'rsa-sha2-512' + | 'ecdsa-sha2-nistp256' | 'ecdsa-sha2-nistp384' | 'ecdsa-sha2-nistp521' </c></p></item> </taglist> </section> @@ -72,12 +78,13 @@ <fsummary>Fetches the host’s private key.</fsummary> <type> <v>Algorithm = public_key_algorithm()</v> - <d>Host key algorithm. Is to support <c>'ssh-rsa' | 'ssh-dss'</c>, but more algorithms - can be handled.</d> + <d>Host key algorithm.</d> <v>DaemonOptions = proplists:proplist()</v> - <d>Options provided to <seealso marker="ssh#daemon-2">ssh:daemon/[2,3]</seealso>.</d> - <v>Key = private_key()</v> - <d>Private key of the host matching the <c>Algorithm</c>.</d> + <d>Options provided to <seealso marker="ssh#daemon-2">ssh:daemon/[2,3]</seealso>. The option list given in + the <c>key_cb</c> option is available with the key <c>key_cb_private</c>.</d> + <v>Key = private_key() | crypto:engine_key_ref()</v> + <d>Private key of the host matching the <c>Algorithm</c>. + It may be a reference to a 'ssh-rsa', rsa-sha2-* or 'ssh-dss' (NOT ecdsa) key stored in a loaded Engine.</d> <v>Reason = term()</v> </type> <desc> @@ -90,11 +97,12 @@ <fsummary>Checks if the user key is authorized.</fsummary> <type> <v>Key = public_key()</v> - <d>Normally an RSA or DSA public key, but handling of other public keys can be added</d> + <d>Normally an RSA, DSA or ECDSA public key, but handling of other public keys can be added</d> <v>User = string()</v> <d>User owning the public key.</d> <v>DaemonOptions = proplists:proplist()</v> - <d>Options provided to <seealso marker="ssh#daemon-2">ssh:daemon/[2,3]</seealso>.</d> + <d>Options provided to <seealso marker="ssh#daemon-2">ssh:daemon/[2,3]</seealso>. The option list given in + the <c>key_cb</c> option is available with the key <c>key_cb_private</c>.</d> <v>Result = boolean()</v> </type> <desc> diff --git a/lib/ssh/src/ssh.erl b/lib/ssh/src/ssh.erl index 1a5d48baca..032d87bdad 100644 --- a/lib/ssh/src/ssh.erl +++ b/lib/ssh/src/ssh.erl @@ -188,6 +188,7 @@ daemon(Port) -> daemon(Socket, UserOptions) when is_port(Socket) -> try #{} = Options = ssh_options:handle_options(server, UserOptions), + case valid_socket_to_use(Socket, ?GET_OPT(transport,Options)) of ok -> {ok, {IP,Port}} = inet:sockname(Socket), @@ -461,6 +462,9 @@ open_listen_socket(_Host0, Port0, Options0) -> %%%---------------------------------------------------------------- finalize_start(Host, Port, Profile, Options0, F) -> try + %% throws error:Error if no usable hostkey is found + ssh_connection_handler:available_hkey_algorithms(server, Options0), + sshd_sup:start_child(Host, Port, Profile, Options0) of {error, {already_started, _}} -> @@ -470,6 +474,8 @@ finalize_start(Host, Port, Profile, Options0, F) -> Result = {ok,_} -> F(Options0, Result) catch + error:{shutdown,Err} -> + {error,Err}; exit:{noproc, _} -> {error, ssh_not_started} end. diff --git a/lib/ssh/src/ssh_auth.erl b/lib/ssh/src/ssh_auth.erl index ac64a7bf14..894877f8bf 100644 --- a/lib/ssh/src/ssh_auth.erl +++ b/lib/ssh/src/ssh_auth.erl @@ -145,14 +145,17 @@ get_public_key(SigAlg, #ssh{opts = Opts}) -> case KeyCb:user_key(KeyAlg, [{key_cb_private,KeyCbOpts}|UserOpts]) of {ok, PrivKey} -> try + %% Check the key - the KeyCb may be a buggy plugin + true = ssh_transport:valid_key_sha_alg(PrivKey, KeyAlg), Key = ssh_transport:extract_public_key(PrivKey), public_key:ssh_encode(Key, ssh2_pubkey) of PubKeyBlob -> {ok,{PrivKey,PubKeyBlob}} catch _:_ -> - not_ok + not_ok end; + _Error -> not_ok end. diff --git a/lib/ssh/src/ssh_connection_handler.erl b/lib/ssh/src/ssh_connection_handler.erl index 4158a52a27..802bf62570 100644 --- a/lib/ssh/src/ssh_connection_handler.erl +++ b/lib/ssh/src/ssh_connection_handler.erl @@ -46,6 +46,7 @@ %%% Internal application API -export([start_connection/4, + available_hkey_algorithms/2, open_channel/6, request/6, request/7, reply_request/3, @@ -432,13 +433,12 @@ init_ssh_record(Role, Socket, Opts) -> init_ssh_record(Role, Socket, PeerAddr, Opts). init_ssh_record(Role, _Socket, PeerAddr, Opts) -> - KeyCb = ?GET_OPT(key_cb, Opts), AuthMethods = ?GET_OPT(auth_methods, Opts), S0 = #ssh{role = Role, - key_cb = KeyCb, + key_cb = ?GET_OPT(key_cb, Opts), opts = Opts, userauth_supported_methods = AuthMethods, - available_host_keys = supported_host_keys(Role, KeyCb, Opts), + available_host_keys = available_hkey_algorithms(Role, Opts), random_length_padding = ?GET_OPT(max_random_length_padding, Opts) }, @@ -1544,44 +1544,42 @@ peer_role(client) -> server; peer_role(server) -> client. %%-------------------------------------------------------------------- -supported_host_keys(client, _, Options) -> - try - find_sup_hkeys(Options) - of - [] -> +available_hkey_algorithms(Role, Options) -> + KeyCb = ?GET_OPT(key_cb, Options), + case [A || A <- available_hkey_algos(Options), + (Role==client) orelse available_host_key(KeyCb, A, Options) + ] of + + [] when Role==client -> error({shutdown, "No public key algs"}); - Algs -> - [atom_to_list(A) || A<-Algs] - catch - exit:Reason -> - error({shutdown, Reason}) - end; -supported_host_keys(server, KeyCb, Options) -> - [atom_to_list(A) || A <- find_sup_hkeys(Options), - available_host_key(KeyCb, A, Options) - ]. + [] when Role==server -> + error({shutdown, "No host key available"}); -find_sup_hkeys(Options) -> - case proplists:get_value(public_key, - ?GET_OPT(preferred_algorithms,Options) - ) - of - undefined -> - ssh_transport:default_algorithms(public_key); - L -> - NonSupported = L--ssh_transport:supported_algorithms(public_key), - L -- NonSupported + Algs -> + [atom_to_list(A) || A<-Algs] end. +available_hkey_algos(Options) -> + SupAlgos = ssh_transport:supported_algorithms(public_key), + HKeys = proplists:get_value(public_key, + ?GET_OPT(preferred_algorithms,Options) + ), + NonSupported = HKeys -- SupAlgos, + AvailableAndSupported = HKeys -- NonSupported, + AvailableAndSupported. + %% Alg :: atom() available_host_key({KeyCb,KeyCbOpts}, Alg, Opts) -> UserOpts = ?GET_OPT(user_options, Opts), case KeyCb:host_key(Alg, [{key_cb_private,KeyCbOpts}|UserOpts]) of - {ok,_} -> true; - _ -> false + {ok,Key} -> + %% Check the key - the KeyCb may be a buggy plugin + ssh_transport:valid_key_sha_alg(Key, Alg); + _ -> + false end. diff --git a/lib/ssh/src/ssh_transport.erl b/lib/ssh/src/ssh_transport.erl index e92c727559..892db6b64f 100644 --- a/lib/ssh/src/ssh_transport.erl +++ b/lib/ssh/src/ssh_transport.erl @@ -795,8 +795,14 @@ get_host_key(SSH, SignAlg) -> #ssh{key_cb = {KeyCb,KeyCbOpts}, opts = Opts} = SSH, UserOpts = ?GET_OPT(user_options, Opts), case KeyCb:host_key(SignAlg, [{key_cb_private,KeyCbOpts}|UserOpts]) of - {ok, PrivHostKey} -> PrivHostKey; - Result -> exit({error, {Result, unsupported_key_type}}) + {ok, PrivHostKey} -> + %% Check the key - the KeyCb may be a buggy plugin + case valid_key_sha_alg(PrivHostKey, SignAlg) of + true -> PrivHostKey; + false -> exit({error, bad_hostkey}) + end; + Result -> + exit({error, {Result, unsupported_key_type}}) end. extract_public_key(#'RSAPrivateKey'{modulus = N, publicExponent = E}) -> @@ -805,7 +811,15 @@ extract_public_key(#'DSAPrivateKey'{y = Y, p = P, q = Q, g = G}) -> {Y, #'Dss-Parms'{p=P, q=Q, g=G}}; extract_public_key(#'ECPrivateKey'{parameters = {namedCurve,OID}, publicKey = Q}) -> - {#'ECPoint'{point=Q}, {namedCurve,OID}}. + {#'ECPoint'{point=Q}, {namedCurve,OID}}; +extract_public_key(#{engine:=_, key_id:=_, algorithm:=Alg} = M) -> + case {Alg, crypto:privkey_to_pubkey(Alg, M)} of + {rsa, [E,N]} -> + #'RSAPublicKey'{modulus = N, publicExponent = E}; + {dss, [P,Q,G,Y]} -> + {Y, #'Dss-Parms'{p=P, q=Q, g=G}} + end. + verify_host_key(#ssh{algorithms=Alg}=SSH, PublicKey, Digest, {AlgStr,Signature}) -> @@ -1255,10 +1269,12 @@ payload(<<PacketLen:32, PaddingLen:8, PayloadAndPadding/binary>>) -> <<Payload:PayloadLen/binary, _/binary>> = PayloadAndPadding, Payload. +sign(SigData, HashAlg, #{algorithm:=dss} = Key) -> + mk_dss_sig(crypto:sign(dss, HashAlg, SigData, Key)); +sign(SigData, HashAlg, #{algorithm:=SigAlg} = Key) -> + crypto:sign(SigAlg, HashAlg, SigData, Key); sign(SigData, HashAlg, #'DSAPrivateKey'{} = Key) -> - DerSignature = public_key:sign(SigData, HashAlg, Key), - #'Dss-Sig-Value'{r = R, s = S} = public_key:der_decode('Dss-Sig-Value', DerSignature), - <<R:160/big-unsigned-integer, S:160/big-unsigned-integer>>; + mk_dss_sig(public_key:sign(SigData, HashAlg, Key)); sign(SigData, HashAlg, Key = #'ECPrivateKey'{}) -> DerEncodedSign = public_key:sign(SigData, HashAlg, Key), #'ECDSA-Sig-Value'{r=R, s=S} = public_key:der_decode('ECDSA-Sig-Value', DerEncodedSign), @@ -1266,6 +1282,12 @@ sign(SigData, HashAlg, Key = #'ECPrivateKey'{}) -> sign(SigData, HashAlg, Key) -> public_key:sign(SigData, HashAlg, Key). + +mk_dss_sig(DerSignature) -> + #'Dss-Sig-Value'{r = R, s = S} = public_key:der_decode('Dss-Sig-Value', DerSignature), + <<R:160/big-unsigned-integer, S:160/big-unsigned-integer>>. + + verify(PlainText, HashAlg, Sig, {_, #'Dss-Parms'{}} = Key) -> case Sig of <<R:160/big-unsigned-integer, S:160/big-unsigned-integer>> -> @@ -1817,6 +1839,8 @@ kex_alg_dependent({Min, NBits, Max, Prime, Gen, E, F, K}) -> %%%---------------------------------------------------------------- +valid_key_sha_alg(#{engine:=_, key_id:=_}, _Alg) -> true; % Engine key + valid_key_sha_alg(#'RSAPublicKey'{}, 'rsa-sha2-512') -> true; valid_key_sha_alg(#'RSAPublicKey'{}, 'rsa-sha2-384') -> true; valid_key_sha_alg(#'RSAPublicKey'{}, 'rsa-sha2-256') -> true; @@ -1830,11 +1854,14 @@ valid_key_sha_alg(#'RSAPrivateKey'{}, 'ssh-rsa' ) -> true; valid_key_sha_alg({_, #'Dss-Parms'{}}, 'ssh-dss') -> true; valid_key_sha_alg(#'DSAPrivateKey'{}, 'ssh-dss') -> true; -valid_key_sha_alg({#'ECPoint'{},{namedCurve,OID}}, Alg) -> sha(OID) == sha(Alg); -valid_key_sha_alg(#'ECPrivateKey'{parameters = {namedCurve,OID}}, Alg) -> sha(OID) == sha(Alg); +valid_key_sha_alg({#'ECPoint'{},{namedCurve,OID}}, Alg) -> valid_key_sha_alg_ec(OID, Alg); +valid_key_sha_alg(#'ECPrivateKey'{parameters = {namedCurve,OID}}, Alg) -> valid_key_sha_alg_ec(OID, Alg); valid_key_sha_alg(_, _) -> false. - +valid_key_sha_alg_ec(OID, Alg) -> + Curve = public_key:oid2ssh_curvename(OID), + Alg == list_to_atom("ecdsa-sha2-" ++ binary_to_list(Curve)). + public_algo(#'RSAPublicKey'{}) -> 'ssh-rsa'; % FIXME: Not right with draft-curdle-rsa-sha2 public_algo({_, #'Dss-Parms'{}}) -> 'ssh-dss'; diff --git a/lib/ssh/test/Makefile b/lib/ssh/test/Makefile index 5ea048a352..a18383d148 100644 --- a/lib/ssh/test/Makefile +++ b/lib/ssh/test/Makefile @@ -38,6 +38,7 @@ MODULES= \ ssh_basic_SUITE \ ssh_bench_SUITE \ ssh_connection_SUITE \ + ssh_engine_SUITE \ ssh_protocol_SUITE \ ssh_property_test_SUITE \ ssh_sftp_SUITE \ @@ -49,6 +50,7 @@ MODULES= \ ssh_test_lib \ ssh_key_cb \ ssh_key_cb_options \ + ssh_key_cb_engine_keys \ ssh_trpt_test_lib \ ssh_echo_server \ ssh_bench_dev_null \ diff --git a/lib/ssh/test/property_test/ssh_eqc_client_info_timing.erl b/lib/ssh/test/property_test/ssh_eqc_client_info_timing.erl index c07140dc43..19e2754eba 100644 --- a/lib/ssh/test/property_test/ssh_eqc_client_info_timing.erl +++ b/lib/ssh/test/property_test/ssh_eqc_client_info_timing.erl @@ -57,9 +57,9 @@ %%% Properties: -prop_seq(_Config) -> +prop_seq(Config) -> {ok,Pid} = ssh_eqc_event_handler:add_report_handler(), - {_, _, Port} = init_daemon(), + {_, _, Port} = init_daemon(Config), numtests(1000, ?FORALL(Delay, choose(0,100),%% Micro seconds try @@ -86,7 +86,8 @@ any_relevant_error_report(Pid) -> end, Reports). %%%================================================================ -init_daemon() -> +init_daemon(Config) -> ok = begin ssh:stop(), ssh:start() end, - ssh_test_lib:daemon([]). + DataDir = proplists:get_value(data_dir, Config), + ssh_test_lib:daemon([{system_dir,DataDir}]). diff --git a/lib/ssh/test/ssh_basic_SUITE.erl b/lib/ssh/test/ssh_basic_SUITE.erl index db2ae241e5..202b0afe57 100644 --- a/lib/ssh/test/ssh_basic_SUITE.erl +++ b/lib/ssh/test/ssh_basic_SUITE.erl @@ -46,6 +46,7 @@ exec_key_differs2/1, exec_key_differs3/1, exec_key_differs_fail/1, + fail_daemon_start/1, idle_time_client/1, idle_time_server/1, inet6_option/1, @@ -105,6 +106,7 @@ all() -> {group, host_user_key_differs}, {group, key_cb}, {group, internal_error}, + {group, rsa_host_key_is_actualy_ecdsa}, daemon_already_started, double_close, daemon_opt_fd, @@ -121,6 +123,7 @@ groups() -> {ecdsa_sha2_nistp256_key, [], basic_tests()}, {ecdsa_sha2_nistp384_key, [], basic_tests()}, {ecdsa_sha2_nistp521_key, [], basic_tests()}, + {rsa_host_key_is_actualy_ecdsa, [], [fail_daemon_start]}, {host_user_key_differs, [], [exec_key_differs1, exec_key_differs2, exec_key_differs3, @@ -180,6 +183,31 @@ init_per_group(rsa_key, Config) -> false -> {skip, unsupported_pub_key} end; +init_per_group(rsa_host_key_is_actualy_ecdsa, Config) -> + case + lists:member('ssh-rsa', + ssh_transport:default_algorithms(public_key)) and + lists:member('ecdsa-sha2-nistp256', + ssh_transport:default_algorithms(public_key)) + of + true -> + DataDir = proplists:get_value(data_dir, Config), + PrivDir = proplists:get_value(priv_dir, Config), + ssh_test_lib:setup_ecdsa("256", DataDir, PrivDir), + %% The following sets up bad rsa keys: + begin + UserDir = PrivDir, + System = filename:join(UserDir, "system"), + file:copy(filename:join(DataDir, "id_rsa"), filename:join(UserDir, "id_rsa")), + file:rename(filename:join(System, "ssh_host_ecdsa_key"), filename:join(System, "ssh_host_rsa_key")), + file:rename(filename:join(System, "ssh_host_ecdsa_key.pub"), filename:join(System, "ssh_host_rsa_key.pub")), + ssh_test_lib:setup_rsa_known_host(DataDir, UserDir), + ssh_test_lib:setup_rsa_auth_keys(DataDir, UserDir) + end, + Config; + false -> + {skip, unsupported_pub_key} + end; init_per_group(ecdsa_sha2_nistp256_key, Config) -> case lists:member('ecdsa-sha2-nistp256', ssh_transport:default_algorithms(public_key)) of @@ -304,7 +332,8 @@ init_per_group(internal_error, Config) -> DataDir = proplists:get_value(data_dir, Config), PrivDir = proplists:get_value(priv_dir, Config), ssh_test_lib:setup_dsa(DataDir, PrivDir), - file:delete(filename:join(PrivDir, "system/ssh_host_dsa_key")), + %% In the test case the key will be deleted after the daemon start: + %% ... file:delete(filename:join(PrivDir, "system/ssh_host_dsa_key")), Config; init_per_group(dir_options, Config) -> PrivDir = proplists:get_value(priv_dir, Config), @@ -868,12 +897,17 @@ key_callback_options(Config) when is_list(Config) -> %%% Test that client does not hang if disconnects due to internal error internal_error(Config) when is_list(Config) -> process_flag(trap_exit, true), - SystemDir = filename:join(proplists:get_value(priv_dir, Config), system), + PrivDir = proplists:get_value(priv_dir, Config), UserDir = proplists:get_value(priv_dir, Config), + SystemDir = filename:join(PrivDir, system), {Pid, Host, Port} = ssh_test_lib:daemon([{system_dir, SystemDir}, {user_dir, UserDir}, {failfun, fun ssh_test_lib:failfun/2}]), + + %% Now provoke an error in the following connect: + file:delete(filename:join(PrivDir, "system/ssh_host_dsa_key")), + {error, Error} = ssh:connect(Host, Port, [{silently_accept_hosts, true}, {user_dir, UserDir}, @@ -902,6 +936,17 @@ send(Config) when is_list(Config) -> %%-------------------------------------------------------------------- +%%% +fail_daemon_start(Config) when is_list(Config) -> + process_flag(trap_exit, true), + SystemDir = filename:join(proplists:get_value(priv_dir, Config), system), + UserDir = proplists:get_value(priv_dir, Config), + + {error,_} = ssh_test_lib:daemon([{system_dir, SystemDir}, + {user_dir, UserDir}, + {failfun, fun ssh_test_lib:failfun/2}]). + +%%-------------------------------------------------------------------- %%% Test ssh:connection_info([peername, sockname]) peername_sockname(Config) when is_list(Config) -> process_flag(trap_exit, true), @@ -1300,14 +1345,11 @@ shell_exit_status(Config) when is_list(Config) -> %%-------------------------------------------------------------------- %% Due to timing the error message may or may not be delivered to %% the "tcp-application" before the socket closed message is recived -check_error("Invalid state") -> - ok; -check_error("Connection closed") -> - ok; -check_error("Selection of key exchange algorithm failed"++_) -> - ok; -check_error(Error) -> - ct:fail(Error). +check_error("Invalid state") -> ok; +check_error("Connection closed") -> ok; +check_error("Selection of key exchange algorithm failed"++_) -> ok; +check_error("No host key available") -> ok; +check_error(Error) -> ct:fail(Error). basic_test(Config) -> ClientOpts = proplists:get_value(client_opts, Config), diff --git a/lib/ssh/test/ssh_engine_SUITE.erl b/lib/ssh/test/ssh_engine_SUITE.erl new file mode 100644 index 0000000000..035446932b --- /dev/null +++ b/lib/ssh/test/ssh_engine_SUITE.erl @@ -0,0 +1,141 @@ +%% +%% %CopyrightBegin% +%% +%% Copyright Ericsson AB 2008-2017. All Rights Reserved. +%% +%% Licensed under the Apache License, Version 2.0 (the "License"); +%% you may not use this file except in compliance with the License. +%% You may obtain a copy of the License at +%% +%% http://www.apache.org/licenses/LICENSE-2.0 +%% +%% Unless required by applicable law or agreed to in writing, software +%% distributed under the License is distributed on an "AS IS" BASIS, +%% WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +%% See the License for the specific language governing permissions and +%% limitations under the License. +%% +%% %CopyrightEnd% +%% + +%% + +-module(ssh_engine_SUITE). + +-include_lib("common_test/include/ct.hrl"). +-include("ssh_test_lib.hrl"). + +%% Note: This directive should only be used in test suites. +-compile(export_all). + +%%-------------------------------------------------------------------- +%% Common Test interface functions ----------------------------------- +%%-------------------------------------------------------------------- + +suite() -> + [{ct_hooks,[ts_install_cth]}, + {timetrap,{seconds,40}}]. + +all() -> + [{group, dsa_key}, + {group, rsa_key} + ]. + +groups() -> + [{dsa_key, [], basic_tests()}, + {rsa_key, [], basic_tests()} + ]. + +basic_tests() -> + [simple_connect + ]. + + +%%-------------------------------------------------------------------- +init_per_suite(Config) -> + ssh:start(), + ?CHECK_CRYPTO( + case load_engine() of + {ok,E} -> + ssh_dbg:messages(fun ct:pal/2), + [{engine,E}|Config]; + {error, notsup} -> + {skip, "Engine not supported on this OpenSSL version"}; + {error, bad_engine_id} -> + {skip, "Dynamic Engine not supported"}; + Other -> + ct:log("Engine load failed: ~p",[Other]), + {fail, "Engine load failed"} + end + ). + +end_per_suite(Config) -> + catch crypto:engine_unload( proplists:get_value(engine,Config) ), + ssh:stop(). + +%%-------------------------------------------------------------------- +init_per_group(dsa_key, Config) -> + case lists:member('ssh-dss', + ssh_transport:default_algorithms(public_key)) of + true -> + start_daemon(Config, 'ssh-dss', "dsa_private_key.pem"); + false -> + {skip, unsupported_pub_key} + end; +init_per_group(rsa_key, Config) -> + case lists:member('ssh-rsa', + ssh_transport:default_algorithms(public_key)) of + true -> + start_daemon(Config, 'ssh-rsa', "rsa_private_key.pem"); + false -> + {skip, unsupported_pub_key} + end. + +start_daemon(Config, KeyType, KeyId) -> + SystemDir = proplists:get_value(data_dir, Config), + FullKeyId = filename:join(SystemDir, KeyId), + KeyCBOpts = [{engine, proplists:get_value(engine,Config)}, + {KeyType, FullKeyId} + ], + Opts = [{key_cb, {ssh_key_cb_engine_keys, KeyCBOpts}}], + {Pid, Host, Port} = ssh_test_lib:std_daemon(Config, Opts), + [{host_port,{Host,Port}}, {daemon_pid,Pid}| Config]. + + +end_per_group(_, Config) -> + catch ssh:stop_daemon(proplists:get_value(daemon_pid,Config)), + Config. + +%%-------------------------------------------------------------------- +%% Test Cases -------------------------------------------------------- +%%-------------------------------------------------------------------- + +%% A simple exec call +simple_connect(Config) -> + {Host,Port} = proplists:get_value(host_port, Config), + CRef = ssh_test_lib:std_connect(Config, Host, Port, []), + ssh:close(CRef). + +%%-------------------------------------------------------------------- +%%-------------------------------------------------------------------- +load_engine() -> + case crypto:get_test_engine() of + {ok, Engine} -> + try crypto:engine_load(<<"dynamic">>, + [{<<"SO_PATH">>, Engine}, + <<"LOAD">>], + []) + catch + error:notsup -> + {error, notsup} + end; + + {error, Error} -> + {error, Error} + end. + +start_std_daemon(Opts, Config) -> + ct:log("starting std_daemon",[]), + {Pid, Host, Port} = ssh_test_lib:std_daemon(Config, Opts), + ct:log("started ~p:~p ~p",[Host,Port,Opts]), + [{srvr_pid,Pid},{srvr_addr,{Host,Port}} | Config]. diff --git a/lib/ssh/test/ssh_engine_SUITE_data/dsa_private_key.pem b/lib/ssh/test/ssh_engine_SUITE_data/dsa_private_key.pem new file mode 100644 index 0000000000..778ffac675 --- /dev/null +++ b/lib/ssh/test/ssh_engine_SUITE_data/dsa_private_key.pem @@ -0,0 +1,9 @@ +-----BEGIN PRIVATE KEY----- +MIIBSwIBADCCASwGByqGSM44BAEwggEfAoGBAMyitTMR7vPbpqyAXJpqnB0AhFwQ +F87IE+JKFl5bD/MSkhhRV5sM73HUU1ooXY0FjhZ+cdLUCATuZR5ta4ydANqWIcAB +gX3IwF1B4zf5SXEKTWkUYneL9dOKtiZLtoG28swrk8xMxwX+0fLHkltCEj6FiTW9 +PFrv8GmIfV6DjcI9AhUAqXWbb3RtoN9Ld28fVMhGZrj3LJUCgYEAwnxGHGBMpJaF +2w7zAw3jHjL8PMYlV6vnufGHQlwF0ZUXJxRsvagMb/X1qACTu2VPYEVoLQGM3cfH +EhHoQmvSXGAyTfR7Bmn3gf1n/s/DcFbdZduUCZ/rAyIrfd0eSbc1I+kZk85UCsKK +w/IYdlqcuYa4Cgm2TapT5uEMqH4jhzEEFgIULh8swEUWmU8aJNWsrWl4eCiuUUg= +-----END PRIVATE KEY----- diff --git a/lib/ssh/test/ssh_engine_SUITE_data/ecdsa_private_key.pem b/lib/ssh/test/ssh_engine_SUITE_data/ecdsa_private_key.pem new file mode 100644 index 0000000000..a45522064f --- /dev/null +++ b/lib/ssh/test/ssh_engine_SUITE_data/ecdsa_private_key.pem @@ -0,0 +1,8 @@ +-----BEGIN PRIVATE KEY----- +MIHuAgEAMBAGByqGSM49AgEGBSuBBAAjBIHWMIHTAgEBBEIBparGjr0KcdNrVM2J +G0mW5ltP1QyvxDqBMyWLWo3fruRZv6Qoohl5skd1u4O+KJoM/UrrSTOXI/MDR7NN +i1yl7O+hgYkDgYYABAG8K2XVsK0ahG9+HIIPwCO0pJY8ulwSTXwIjkCGyB2lpglh +8qJmRzuyGcfRTslv8wfv0sPlT9H9PKDvgrTUL7rvQQDdOODNgVPXSecUoXoPn+X+ +eqxs77bjx+A5x0t/i3m5PfkaNPh5MZ1H/bWuOOdj2ZXZw0R4rlVc0zVrgnPU8L8S +BQ== +-----END PRIVATE KEY----- diff --git a/lib/ssh/test/ssh_engine_SUITE_data/rsa_private_key.pem b/lib/ssh/test/ssh_engine_SUITE_data/rsa_private_key.pem new file mode 100644 index 0000000000..ea0e3d3958 --- /dev/null +++ b/lib/ssh/test/ssh_engine_SUITE_data/rsa_private_key.pem @@ -0,0 +1,28 @@ +-----BEGIN PRIVATE KEY----- +MIIEvwIBADANBgkqhkiG9w0BAQEFAASCBKkwggSlAgEAAoIBAQCwwb0/ddXGXTFK +4FLxXdV6a/WJMSoPPS55RvZIAHFsiTtvPLbJ8LxDsZ6wSVZLN0/UQ4wdWn9jftyj +U5/IxBVG8XOtKimTMvm3/ZOzVLueGHBbrLYscRv9oL85ulTKHWgrZDu0lBX5JJTI +v5UTCErzJRQbka9DG1GaBgDb1PlXfkzBWMwfsBZmwoC77KvCcIGCgbW/XCY03TP2 +3Tg8drvpByMStddP2FQ4fZ91qFUzPu8uhZEsqSQTFlmhgGEx7dLlky0xvu62RuAD +RTpINpcWZtWDHTdssOqu653LwwqBY8lBopCZ/4Af8QR3ZYkQhen1YLEbVheXRuzI +LSCZIiJNAgMBAAECggEBAJH4/fxpqQkvr2Shy33Pu1xlyhnpw01gfn/jrcKasxEq +aC4eWup86E2TY3U8q4pkfIXU3uLi+O9HNpmflwargNLc1mY8uqb44ygiv5bLNEKE +9k2PXcdoBfC4jxPyoNFl5cBn/7LK1TazEjiTl15na9ZPWcLG1pG5/vMPYCgsQ1sP +8J3c4E3aaXIj9QceYxBprl490OCzieGyZlRipncz3g4UShRc/b4cycvDZOJpmAy4 +zbWTcBcSMPVPi5coF0K8UcimiqZkotfb/2RLc433i34IdsIXMM+brdq+g8rmjg5a ++oQPy02M6tFApBruEhAz8DGgaLtDY6MLtyZAt3SjXnUCgYEA1zLgamdTHOqrrmIi +eIQBnAJiyIfcY8B9SX1OsLGYFCHiPVwgUY35B2c7MavMsGcExJhtE+uxU7o5djtM +R6r9cRHOXJ6EQwa8OwzzPqbM17/YqNDeK39bc9WOFUqRWrhDhVMPy6z8rmZr73mG +IUC7mBNx/1GBdVYXIlsXzC96dI8CgYEA0kUAhz6I5nyPa70NDEUYHLHf3IW1BCmE +UoVbraSePJtIEY/IqFx7oDuFo30d4n5z+8ICCtyid1h/Cp3mf3akOiqltYUfgV1G +JgcEjKKYWEnO7cfFyO7LB7Y3GYYDJNy6EzVWPiwTGk9ZTfFJEESmHC45Unxgd17m +Dx/R58rFgWMCgYBQXQWFdtSI5fH7C1bIHrPjKNju/h2FeurOuObcAVZDnmu4cmD3 +U8d9xkVKxVeJQM99A1coq0nrdI3k4zwXP3mp8fZYjDHkPe2pN6rW6L9yiohEcsuk +/siON1/5/4DMmidM8LnjW9R45HLGWWGHpX7oyco2iJ+Jy/6Tq+T1MX3PbQKBgQCm +hdsbQJ0u3CrBSmFQ/E9SOlRt0r4+45pVuCOY6yweF2QF9HcXTtbhWQJHLclDHJ5C +Ha18aKuKFN3XzKFFBPKe1jOSBDGlQ/dQGnKx5fr8wMdObM3oiaTlIJuWbRmEUgJT +QARjDIi8Z2b0YUhZx+Q9oSXoe3PyVYehJrQX+/BavQKBgQCIr7Zp0rQPbfqcTL+M +OYHUoNcb14f9f8hXeXHQOqVpsGwxGdRQAU9wbx/4+obKB5xIkzBsVNcJwavisNja +hegnGjTB/9Hc4m+5bMGwH0bhS2eQO4o+YYM2ypDmFQqDLRfFUlZ5PVHffm/aA9+g +GanNBCsmtoHtV6CJ1UZ7NmBuIA== +-----END PRIVATE KEY----- diff --git a/lib/ssh/test/ssh_engine_SUITE_data/rsa_private_key_pwd.pem b/lib/ssh/test/ssh_engine_SUITE_data/rsa_private_key_pwd.pem new file mode 100644 index 0000000000..501662fc35 --- /dev/null +++ b/lib/ssh/test/ssh_engine_SUITE_data/rsa_private_key_pwd.pem @@ -0,0 +1,30 @@ +-----BEGIN ENCRYPTED PRIVATE KEY----- +MIIFDjBABgkqhkiG9w0BBQ0wMzAbBgkqhkiG9w0BBQwwDgQIh888Iq6gxuMCAggA +MBQGCCqGSIb3DQMHBAic/11YZ8Nt5gSCBMjG/Jb4qiMoBS50iQvHXqcETPE+0NBr +jhsn9w94LkdRBstMPAsoKmY98Er96Rnde/NfmqlU9CupKTkd7Ce5poBf72Y6KMED +cPURyjbGRFsu6x9skXB2obhyKYEqAEF2oQAg4Qbe5v1qXBIgDuC/NgiJnM+w2zCZ +LkHSZB2/NmcnvDzcgPF7TM8pTO23xCJ33m37qjfWvHsgocVqZmL9wQ4+wr/NMYjJ +pJvX1OHW1vBsZsXh40WchalYRSB1VeO368QfsE8coRJztqbMzdce9EQdMB6Q6jlO +cetd3moLIoMP4I7HW0/SgokbycTbRiYSvRyU1TGc2WbW6BrFZV24IckcnnVUFatf +6HKUcaYLG68dJcRgs5QMGkcmgVvlddENHFmHZlo0eym/xSiUl/AT8/5odscm6ML8 +wW5sneax+TF4J2eYmiN7yjAUCodXVTNYNDVKo6uUhntlymbM0o4UitVIbPIfTDHl +sxJAEZ7vpuPqeNMxUk6G6zipuEjqsVbnuFSBSZmgKiGYcifRPUmqqINa3DdS4WVx +xaPWdHbHVRD//ze3h/FsA+1lIE5q2kUE0xXseJA1ISog++kJp14XeaaL2j/tx3Ob +OsbcaOAD/IUw/ItDt9kn0qzfnar7sS0Wov8AmJQxHmH7Lm93jHTLM05yE0AR/eBr +Mig2ZdC+9OqVC+GPuBkRjSs8NpltQIDroz6EV9IMwPwXm0szSYoyoPLmlHJUdnLs +ZUef+au6hYkEJBrvuisagnq5eT/fCV3hsjD7yODebNU2CmBTo6X2PRx/xsBHRMWl +QkoM9PBdSCnKv6HpHl4pchuoqU2NpFjN0BCaad6aHfZSTnqgzK4bEh1oO6dI8/rB +/eh71JyFFG5J4xbpaqz5Su01V1iwU5leK5bDwqals4M4+ZGHGciou7qnXUmX2fJl +r6DlMUa/xy+A2ZG0NuZR05yk2oB3+KVNMgp6zFty3XaxwoNtc8GTLtLnBnIh2rlP +mE1+I65LRWwrNQalPeOAUrYuEzhyp2Df7a8Ykas5PUH7MGR/S0Ge/dLxtE2bJuK4 +znbLAsGhvo/SbNxYqIp6D4iDtd3va6yUGncy41paA/vTKFVvXZDrXcwJQYYCVOGT +OwdzNuozU8Dc7oxsd8oakfC46kvmVaOrGvZbm56PFfprcaL/Hslska5xxEni/eZe +WRxZbCBhAVqS1pn5zkDQVUe9uFlR/x39Qi01HIlKLBsjpSs6qQsFArMe8hgXmXLG +xP+dyVuOE18NzSewdEjeqSRKIM7Qi8EOjZsI4HdSRBY7bh9VhmaVXDZiCSf33TTE +3y8nimzQAeuGoYg6WqHmWWC2Qnpki2HlaIH/ayXEyQWkP/qvg61e8ovdg9Fy8JOO +0AacXVt5zj0q00AW5bKx7usi4NIjZedi86hUm6H19aBm7r86BKjwYTEI/GOcdrbV +9HC/8ayOimgwiAG3gq+aLioWym+Z6KnsbVd7XReVbvM/InQx54WA2y5im0A+/c67 +oQFFPV84XGX9waeqv/K4Wzkm6HW+qVAEM67482VGOf0PVrlQMno6dOotT/Y7ljoZ +2iz0LmN9yylJnLPDrr1i6gzbs5OhhUgbF5LI2YP2wWdCZTl/DrKSIvQZWl8U+tw3 +ciA= +-----END ENCRYPTED PRIVATE KEY----- diff --git a/lib/ssh/test/ssh_key_cb_engine_keys.erl b/lib/ssh/test/ssh_key_cb_engine_keys.erl new file mode 100644 index 0000000000..fc9cbfd49b --- /dev/null +++ b/lib/ssh/test/ssh_key_cb_engine_keys.erl @@ -0,0 +1,62 @@ +%% +%% %CopyrightBegin% +%% +%% Copyright Ericsson AB 2015-2017. All Rights Reserved. +%% +%% Licensed under the Apache License, Version 2.0 (the "License"); +%% you may not use this file except in compliance with the License. +%% You may obtain a copy of the License at +%% +%% http://www.apache.org/licenses/LICENSE-2.0 +%% +%% Unless required by applicable law or agreed to in writing, software +%% distributed under the License is distributed on an "AS IS" BASIS, +%% WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +%% See the License for the specific language governing permissions and +%% limitations under the License. +%% +%% %CopyrightEnd% +%% + +%% +%%---------------------------------------------------------------------- + +%% Note: This module is used by ssh_basic_SUITE + +-module(ssh_key_cb_engine_keys). +-behaviour(ssh_server_key_api). +-compile(export_all). + +host_key(SshAlg, Options) -> + KBopts = proplists:get_value(key_cb_private, Options, []), + Engine = proplists:get_value(engine, KBopts), + case proplists:get_value(SshAlg, KBopts) of + undefined -> + {error, {unknown_alg,SshAlg}}; + KeyId -> + case crypto_alg(SshAlg) of + undefined -> + {error, {unsupported_alg,SshAlg}}; + CryptoAlg -> + PrivKey = #{engine => Engine, + key_id => KeyId, + algorithm => CryptoAlg}, + %% Is there a key with this reference ? + case crypto:privkey_to_pubkey(CryptoAlg, PrivKey) of + [_|_] -> + {ok, PrivKey}; + _ -> + {error, {no_hostkey,SshAlg}} + end + end + end. + +is_auth_key(_PublicUserKey, _User, _Options) -> + false. + + + +crypto_alg('ssh-rsa') -> rsa; +crypto_alg('ssh-dss') -> dss; +crypto_alg(_) -> undefined. + diff --git a/lib/ssh/test/ssh_protocol_SUITE.erl b/lib/ssh/test/ssh_protocol_SUITE.erl index 74f802cf57..3e3e151781 100644 --- a/lib/ssh/test/ssh_protocol_SUITE.erl +++ b/lib/ssh/test/ssh_protocol_SUITE.erl @@ -630,11 +630,12 @@ client_handles_keyboard_interactive_0_pwds(Config) -> %%%-------------------------------------------------------------------- -client_info_line(_Config) -> +client_info_line(Config) -> %% A client must not send an info-line. If it does, the server should handle %% handle this gracefully {ok,Pid} = ssh_eqc_event_handler:add_report_handler(), - {_, _, Port} = ssh_test_lib:daemon([]), + DataDir = proplists:get_value(data_dir, Config), + {_, _, Port} = ssh_test_lib:daemon([{system_dir,DataDir}]), %% Fake client: {ok,S} = gen_tcp:connect("localhost",Port,[]), diff --git a/lib/ssl/doc/src/ssl.xml b/lib/ssl/doc/src/ssl.xml index ca2dcbb761..ac5a69c69b 100644 --- a/lib/ssl/doc/src/ssl.xml +++ b/lib/ssl/doc/src/ssl.xml @@ -69,7 +69,9 @@ <p><c>| {cert, public_key:der_encoded()}</c></p> <p><c>| {certfile, path()}</c></p> <p><c>| {key, {'RSAPrivateKey'| 'DSAPrivateKey' | 'ECPrivateKey' - | 'PrivateKeyInfo', public_key:der_encoded()}}</c></p> + | 'PrivateKeyInfo', public_key:der_encoded()} | + #{algorithm := rsa | dss | ecdsa, + engine := crypto:engine_ref(), key_id := crypto:key_id(), password => crypto:password()}</c></p> <p><c>| {keyfile, path()}</c></p> <p><c>| {password, string()}</c></p> <p><c>| {cacerts, [public_key:der_encoded()]}</c></p> @@ -202,8 +204,12 @@ <item><p>Path to a file containing the user certificate.</p></item> <tag><c>{key, {'RSAPrivateKey'| 'DSAPrivateKey' | 'ECPrivateKey' - |'PrivateKeyInfo', public_key:der_encoded()}}</c></tag> - <item><p>The DER-encoded user's private key. If this option + |'PrivateKeyInfo', public_key:der_encoded()} | #{algorithm := rsa | dss | ecdsa, + engine := crypto:engine_ref(), key_id := crypto:key_id(), password => crypto:password()}</c></tag> + <item><p>The DER-encoded user's private key or a map refering to a crypto + engine and its key reference that optionally can be password protected, + seealso <seealso marker="crypto:crypto#engine_load-4"> crypto:engine_load/4 + </seealso> and <seealso marker="crypto:engine_load"> Crypto's Users Guide</seealso>. If this option is supplied, it overrides option <c>keyfile</c>.</p></item> <tag><c>{keyfile, path()}</c></tag> @@ -589,22 +595,19 @@ fun(srp, Username :: string(), UserState :: term()) -> <tag><c>{server_name_indication, HostName :: hostname()}</c></tag> <item><p>Specify the hostname to be used in TLS Server Name Indication extension. - Is usefull when upgrading a TCP socket to a TLS socket or if the hostname can not be - derived from the Host argument to <seealso marker="ssl#connect-3">ssl:connect/3</seealso>. - Will also cause the client to preform host name verification of the peer certificate - <seealso marker="public_key:public_key#pkix_verify_hostname-2">public_key:pkix_verify_hostname(PeerCert, [{dns_id, HostName}])</seealso> - </p> during the x509-path validation. If the check fails the error {bad_cert, hostname_check_failiure} will be - propagated to the path validation fun <seealso marker="#verify_fun">verify_fun</seealso> - </item> - - <tag><c>{server_name_indication, disable}</c></tag> - <item> - <p>When starting a TLS connection without upgrade, the Server Name - Indication extension is sent if possible that is can be derived from the Host argument - to <seealso marker="ssl#connect-3">ssl:connect/3</seealso>. - This option can be used to disable that behavior.</p> - <note><p> Note that this also disables the default host name verification check of the peer certificate.</p></note> + If not specified it will default to the <c>Host</c> argument of <seealso marker="#connect-3">connect/[3,4]</seealso> + unless it is of type inet:ipaddress().</p> + <p> + The <c>HostName</c> will also be used in the hostname verification of the peer certificate using + <seealso marker="public_key:public_key#pkix_verify_hostname-2">public_key:pkix_verify_hostname/2</seealso>. + </p> </item> + <tag><c>{server_name_indication, disable}</c></tag> + <item> + <p> Prevents the Server Name Indication extension from being sent and + disables the hostname verification check + <seealso marker="public_key:public_key#pkix_verify_hostname-2">public_key:pkix_verify_hostname/2</seealso> </p> + </item> <tag><c>{fallback, boolean()}</c></tag> <item> <p> Send special cipher suite TLS_FALLBACK_SCSV to avoid undesired TLS version downgrade. @@ -881,6 +884,12 @@ fun(srp, Username :: string(), UserState :: term()) -> <desc><p>Upgrades a <c>gen_tcp</c>, or equivalent, connected socket to an SSL socket, that is, performs the client-side ssl handshake.</p> + + <note><p>If the option <c>verify</c> is set to <c>verify_peer</c> + the option <c>server_name_indication</c> shall also be specified, + if it is not no Server Name Indication extension will be sent, + and <seealso marker="public_key:public_key#pkix_verify_hostname-2">public_key:pkix_verify_hostname/2</seealso> + will be called with the IP-address of the connection as <c>ReferenceID</c>, which is proably not what you want.</p></note> </desc> </func> @@ -897,7 +906,24 @@ fun(srp, Username :: string(), UserState :: term()) -> <v>SslSocket = sslsocket()</v> <v>Reason = term()</v> </type> - <desc><p>Opens an SSL connection to <c>Host</c>, <c>Port</c>.</p></desc> + <desc><p>Opens an SSL connection to <c>Host</c>, <c>Port</c>.</p> + + <p> When the option <c>verify</c> is set to <c>verify_peer</c> the check + <seealso marker="public_key:public_key#pkix_verify_hostname-2">public_key:pkix_verify_hostname/2</seealso> + will be performed in addition to the usual x509-path validation checks. If the check fails the error {bad_cert, hostname_check_failed} will + be propagated to the path validation fun <seealso marker="#verify_fun">verify_fun</seealso>, where it is possible to do customized + checks by using the full possibilitis of the <seealso marker="public_key:public_key#pkix_verify_hostname-2">public_key:pkix_verify_hostname/2</seealso> API. + + When the option <c>server_name_indication</c> is provided, its value (the DNS name) will be used as <c>ReferenceID</c> + to <seealso marker="public_key:public_key#pkix_verify_hostname-2">public_key:pkix_verify_hostname/2</seealso>. + When no <c>server_name_indication</c> option is given, the <c>Host</c> argument will be used as + Server Name Indication extension. The <c>Host</c> argument will also be used for the + <seealso marker="public_key:public_key#pkix_verify_hostname-2">public_key:pkix_verify_hostname/2</seealso> check and if the <c>Host</c> + argument is an <c>inet:ip_address()</c> the <c>ReferenceID</c> used for the check will be <c>{ip, Host}</c> otherwise + <c>dns_id</c> will be assumed with a fallback to <c>ip</c> if that fails. </p> + <note><p>According to good practices certificates should not use IP-addresses as "server names". It would + be very surprising if this happen outside a closed network. </p></note> + </desc> </func> <func> diff --git a/lib/ssl/src/dtls_socket.erl b/lib/ssl/src/dtls_socket.erl index 5f854fbb4b..0e4ab089dc 100644 --- a/lib/ssl/src/dtls_socket.erl +++ b/lib/ssl/src/dtls_socket.erl @@ -24,7 +24,7 @@ -export([send/3, listen/3, accept/3, connect/4, socket/4, setopts/3, getopts/3, getstat/3, peername/2, sockname/2, port/2, close/2]). --export([emulated_options/0, internal_inet_values/0, default_inet_values/0, default_cb_info/0]). +-export([emulated_options/0, emulated_options/1, internal_inet_values/0, default_inet_values/0, default_cb_info/0]). send(Transport, {{IP,Port},Socket}, Data) -> Transport:send(Socket, IP, Port, Data). @@ -133,6 +133,9 @@ port(Transport, Socket) -> emulated_options() -> [mode, active, packet, packet_size]. +emulated_options(Opts) -> + emulated_options(Opts, internal_inet_values(), default_inet_values()). + internal_inet_values() -> [{active, false}, {mode,binary}]. @@ -158,3 +161,29 @@ emulated_socket_options(InetValues, #socket_options{ packet_size = proplists:get_value(packet_size, InetValues, PacketSize), active = proplists:get_value(active, InetValues, Active) }. + +emulated_options([{mode, Value} = Opt |Opts], Inet, Emulated) -> + validate_inet_option(mode, Value), + emulated_options(Opts, Inet, [Opt | proplists:delete(mode, Emulated)]); +emulated_options([{header, _} = Opt | _], _, _) -> + throw({error, {options, {not_supported, Opt}}}); +emulated_options([{active, Value} = Opt |Opts], Inet, Emulated) -> + validate_inet_option(active, Value), + emulated_options(Opts, Inet, [Opt | proplists:delete(active, Emulated)]); +emulated_options([{packet, _} = Opt | _], _, _) -> + throw({error, {options, {not_supported, Opt}}}); +emulated_options([{packet_size, _} = Opt | _], _, _) -> + throw({error, {options, {not_supported, Opt}}}); +emulated_options([Opt|Opts], Inet, Emulated) -> + emulated_options(Opts, [Opt|Inet], Emulated); +emulated_options([], Inet,Emulated) -> + {Inet, Emulated}. + +validate_inet_option(mode, Value) + when Value =/= list, Value =/= binary -> + throw({error, {options, {mode,Value}}}); +validate_inet_option(active, Value) + when Value =/= true, Value =/= false, Value =/= once -> + throw({error, {options, {active,Value}}}); +validate_inet_option(_, _) -> + ok. diff --git a/lib/ssl/src/inet_tls_dist.erl b/lib/ssl/src/inet_tls_dist.erl index d644cbe66a..96782dcfc0 100644 --- a/lib/ssl/src/inet_tls_dist.erl +++ b/lib/ssl/src/inet_tls_dist.erl @@ -360,7 +360,7 @@ do_setup(Driver, Kernel, Node, Type, MyNode, LongOrShortNames, SetupTime) -> Opts = trace(connect_options(get_ssl_options(client))), dist_util:reset_timer(Timer), case ssl:connect( - Ip, TcpPort, + Address, TcpPort, [binary, {active, false}, {packet, 4}, Driver:family(), nodelay()] ++ Opts, net_kernel:connecttime()) of diff --git a/lib/ssl/src/ssl.erl b/lib/ssl/src/ssl.erl index 62da498e76..4007e44a83 100644 --- a/lib/ssl/src/ssl.erl +++ b/lib/ssl/src/ssl.erl @@ -896,7 +896,8 @@ validate_option(key, {KeyType, Value}) when is_binary(Value), KeyType == 'ECPrivateKey'; KeyType == 'PrivateKeyInfo' -> {KeyType, Value}; - +validate_option(key, #{algorithm := _} = Value) -> + Value; validate_option(keyfile, undefined) -> <<>>; validate_option(keyfile, Value) when is_binary(Value) -> @@ -1003,7 +1004,7 @@ validate_option(server_name_indication = Opt, Value) when is_list(Value) -> validate_option(server_name_indication, undefined = Value) -> Value; validate_option(server_name_indication, disable) -> - undefined; + disable; validate_option(sni_hosts, []) -> []; @@ -1113,24 +1114,6 @@ dtls_validate_versions([Version | Rest], Versions) when Version == 'dtlsv1'; dtls_validate_versions([Ver| _], Versions) -> throw({error, {options, {Ver, {versions, Versions}}}}). -validate_inet_option(mode, Value) - when Value =/= list, Value =/= binary -> - throw({error, {options, {mode,Value}}}); -validate_inet_option(packet, Value) - when not (is_atom(Value) orelse is_integer(Value)) -> - throw({error, {options, {packet,Value}}}); -validate_inet_option(packet_size, Value) - when not is_integer(Value) -> - throw({error, {options, {packet_size,Value}}}); -validate_inet_option(header, Value) - when not is_integer(Value) -> - throw({error, {options, {header,Value}}}); -validate_inet_option(active, Value) - when Value =/= true, Value =/= false, Value =/= once -> - throw({error, {options, {active,Value}}}); -validate_inet_option(_, _) -> - ok. - %% The option cacerts overrides cacertsfile ca_cert_default(_,_, [_|_]) -> undefined; @@ -1145,31 +1128,11 @@ ca_cert_default(verify_peer, undefined, _) -> emulated_options(Protocol, Opts) -> case Protocol of tls -> - emulated_options(Opts, tls_socket:internal_inet_values(), tls_socket:default_inet_values()); + tls_socket:emulated_options(Opts); dtls -> - emulated_options(Opts, dtls_socket:internal_inet_values(), dtls_socket:default_inet_values()) + dtls_socket:emulated_options(Opts) end. -emulated_options([{mode, Value} = Opt |Opts], Inet, Emulated) -> - validate_inet_option(mode, Value), - emulated_options(Opts, Inet, [Opt | proplists:delete(mode, Emulated)]); -emulated_options([{header, Value} = Opt | Opts], Inet, Emulated) -> - validate_inet_option(header, Value), - emulated_options(Opts, Inet, [Opt | proplists:delete(header, Emulated)]); -emulated_options([{active, Value} = Opt |Opts], Inet, Emulated) -> - validate_inet_option(active, Value), - emulated_options(Opts, Inet, [Opt | proplists:delete(active, Emulated)]); -emulated_options([{packet, Value} = Opt |Opts], Inet, Emulated) -> - validate_inet_option(packet, Value), - emulated_options(Opts, Inet, [Opt | proplists:delete(packet, Emulated)]); -emulated_options([{packet_size, Value} = Opt | Opts], Inet, Emulated) -> - validate_inet_option(packet_size, Value), - emulated_options(Opts, Inet, [Opt | proplists:delete(packet_size, Emulated)]); -emulated_options([Opt|Opts], Inet, Emulated) -> - emulated_options(Opts, [Opt|Inet], Emulated); -emulated_options([], Inet,Emulated) -> - {Inet, Emulated}. - handle_cipher_option(Value, Version) when is_list(Value) -> try binary_cipher_suites(Version, Value) of Suites -> diff --git a/lib/ssl/src/ssl_certificate.erl b/lib/ssl/src/ssl_certificate.erl index 0dd5e5c5cf..a3333d35e9 100644 --- a/lib/ssl/src/ssl_certificate.erl +++ b/lib/ssl/src/ssl_certificate.erl @@ -138,13 +138,8 @@ validate(_, {bad_cert, _} = Reason, _) -> {fail, Reason}; validate(_, valid, UserState) -> {valid, UserState}; -validate(Cert, valid_peer, UserState = {client, _,_, Hostname, _, _}) when Hostname =/= undefined -> - case public_key:pkix_verify_hostname(Cert, [{dns_id, Hostname}]) of - true -> - {valid, UserState}; - false -> - {fail, {bad_cert, hostname_check_failed}} - end; +validate(Cert, valid_peer, UserState = {client, _,_, Hostname, _, _}) when Hostname =/= disable -> + verify_hostname(Hostname, Cert, UserState); validate(_, valid_peer, UserState) -> {valid, UserState}. @@ -337,3 +332,32 @@ new_trusteded_chain(DerCert, [_ | Rest]) -> new_trusteded_chain(DerCert, Rest); new_trusteded_chain(_, []) -> unknown_ca. + +verify_hostname({fallback, Hostname}, Cert, UserState) when is_list(Hostname) -> + case public_key:pkix_verify_hostname(Cert, [{dns_id, Hostname}]) of + true -> + {valid, UserState}; + false -> + case public_key:pkix_verify_hostname(Cert, [{ip, Hostname}]) of + true -> + {valid, UserState}; + false -> + {fail, {bad_cert, hostname_check_failed}} + end + end; + +verify_hostname({fallback, Hostname}, Cert, UserState) -> + case public_key:pkix_verify_hostname(Cert, [{ip, Hostname}]) of + true -> + {valid, UserState}; + false -> + {fail, {bad_cert, hostname_check_failed}} + end; + +verify_hostname(Hostname, Cert, UserState) -> + case public_key:pkix_verify_hostname(Cert, [{dns_id, Hostname}]) of + true -> + {valid, UserState}; + false -> + {fail, {bad_cert, hostname_check_failed}} + end. diff --git a/lib/ssl/src/ssl_config.erl b/lib/ssl/src/ssl_config.erl index e4611995ec..022fb7eac0 100644 --- a/lib/ssl/src/ssl_config.erl +++ b/lib/ssl/src/ssl_config.erl @@ -91,7 +91,15 @@ init_certificates(undefined, #{pem_cache := PemCache} = Config, CertFile, server end; init_certificates(Cert, Config, _, _) -> {ok, Config#{own_certificate => Cert}}. - +init_private_key(_, #{algorithm := Alg} = Key, <<>>, _Password, _Client) when Alg == ecdsa; + Alg == rsa; + Alg == dss -> + case maps:is_key(engine, Key) andalso maps:is_key(key_id, Key) of + true -> + Key; + false -> + throw({key, {invalid_key_id, Key}}) + end; init_private_key(_, undefined, <<>>, _Password, _Client) -> undefined; init_private_key(DbHandle, undefined, KeyFile, Password, _) -> diff --git a/lib/ssl/src/ssl_connection.erl b/lib/ssl/src/ssl_connection.erl index 56ba952e67..d83c9cb59f 100644 --- a/lib/ssl/src/ssl_connection.erl +++ b/lib/ssl/src/ssl_connection.erl @@ -511,12 +511,13 @@ certify(internal, #certificate{}, certify(internal, #certificate{} = Cert, #state{negotiated_version = Version, role = Role, + host = Host, cert_db = CertDbHandle, cert_db_ref = CertDbRef, crl_db = CRLDbInfo, ssl_options = Opts} = State, Connection) -> case ssl_handshake:certify(Cert, CertDbHandle, CertDbRef, - Opts, CRLDbInfo, Role) of + Opts, CRLDbInfo, Role, Host) of {PeerCert, PublicKeyInfo} -> handle_peer_cert(Role, PeerCert, PublicKeyInfo, State#state{client_certificate_requested = false}, Connection); @@ -725,6 +726,7 @@ cipher(internal, #finished{verify_data = Data} = Finished, expecting_finished = true, session = #session{master_secret = MasterSecret} = Session0, + ssl_options = SslOpts, connection_states = ConnectionStates0, tls_handshake_history = Handshake0} = State, Connection) -> case ssl_handshake:verify_connection(ssl:tls_version(Version), Finished, @@ -732,7 +734,7 @@ cipher(internal, #finished{verify_data = Data} = Finished, get_current_prf(ConnectionStates0, read), MasterSecret, Handshake0) of verified -> - Session = register_session(Role, Host, Port, Session0), + Session = register_session(Role, host_id(Role, Host, SslOpts), Port, Session0), cipher_role(Role, Data, Session, State#state{expecting_finished = false}, Connection); #alert{} = Alert -> @@ -2336,6 +2338,11 @@ register_session(server, _, Port, #session{is_resumable = new} = Session0) -> register_session(_, _, _, Session) -> Session. %% Already registered +host_id(client, _Host, #ssl_options{server_name_indication = Hostname}) when is_list(Hostname) -> + Hostname; +host_id(_, Host, _) -> + Host. + handle_new_session(NewId, CipherSuite, Compression, #state{session = Session0, protocol_cb = Connection} = State0) -> diff --git a/lib/ssl/src/ssl_handshake.erl b/lib/ssl/src/ssl_handshake.erl index fc4181a760..8681765284 100644 --- a/lib/ssl/src/ssl_handshake.erl +++ b/lib/ssl/src/ssl_handshake.erl @@ -50,7 +50,7 @@ finished/5, next_protocol/1]). %% Handle handshake messages --export([certify/6, client_certificate_verify/6, certificate_verify/6, verify_signature/5, +-export([certify/7, client_certificate_verify/6, certificate_verify/6, verify_signature/5, master_secret/4, server_key_exchange_hash/2, verify_connection/6, init_handshake_history/0, update_handshake_history/3, verify_server_key/5 ]). @@ -407,21 +407,21 @@ verify_signature(_, Hash, {HashAlgo, _SignAlg}, Signature, %%-------------------------------------------------------------------- -spec certify(#certificate{}, db_handle(), certdb_ref(), #ssl_options{}, term(), - client | server) -> {der_cert(), public_key_info()} | #alert{}. + client | server, inet:hostname() | inet:ip_address()) -> {der_cert(), public_key_info()} | #alert{}. %% %% Description: Handles a certificate handshake message %%-------------------------------------------------------------------- certify(#certificate{asn1_certificates = ASN1Certs}, CertDbHandle, CertDbRef, - Opts, CRLDbHandle, Role) -> + Opts, CRLDbHandle, Role, Host) -> + ServerName = server_name(Opts#ssl_options.server_name_indication, Host, Role), [PeerCert | _] = ASN1Certs, try {TrustedCert, CertPath} = ssl_certificate:trusted_cert_and_path(ASN1Certs, CertDbHandle, CertDbRef, Opts#ssl_options.partial_chain), ValidationFunAndState = validation_fun_and_state(Opts#ssl_options.verify_fun, Role, - CertDbHandle, CertDbRef, - Opts#ssl_options.server_name_indication, + CertDbHandle, CertDbRef, ServerName, Opts#ssl_options.crl_check, CRLDbHandle, CertPath), case public_key:pkix_path_validation(TrustedCert, CertPath, @@ -1584,6 +1584,8 @@ select_shared_curve([Curve | Rest], Curves) -> sni(undefined) -> undefined; +sni(disable) -> + undefined; sni(Hostname) -> #sni{hostname = Hostname}. @@ -1700,14 +1702,20 @@ digitally_signed(Version, Hashes, HashAlgo, PrivateKey) -> error:badkey-> throw(?ALERT_REC(?FATAL, ?HANDSHAKE_FAILURE, bad_key(PrivateKey))) end. - +do_digitally_signed({3, Minor}, Hash, HashAlgo, #{algorithm := Alg} = Engine) + when Minor >= 3 -> + crypto:sign(Alg, HashAlgo, {digest, Hash}, maps:remove(algorithm, Engine)); do_digitally_signed({3, Minor}, Hash, HashAlgo, Key) when Minor >= 3 -> public_key:sign({digest, Hash}, HashAlgo, Key); -do_digitally_signed(_Version, Hash, HashAlgo, #'DSAPrivateKey'{} = Key) -> - public_key:sign({digest, Hash}, HashAlgo, Key); do_digitally_signed(_Version, Hash, _HashAlgo, #'RSAPrivateKey'{} = Key) -> public_key:encrypt_private(Hash, Key, [{rsa_pad, rsa_pkcs1_padding}]); +do_digitally_signed({3, _}, Hash, _, + #{algorithm := rsa} = Engine) -> + crypto:private_encrypt(rsa, Hash, maps:remove(algorithm, Engine), + rsa_pkcs1_padding); +do_digitally_signed({3, _}, Hash, HashAlgo, #{algorithm := Alg} = Engine) -> + crypto:sign(Alg, HashAlgo, {digest, Hash}, maps:remove(algorithm, Engine)); do_digitally_signed(_Version, Hash, HashAlgo, Key) -> public_key:sign({digest, Hash}, HashAlgo, Key). @@ -2434,3 +2442,9 @@ available_signature_algs(#hash_sign_algos{hash_sign_algos = ClientHashSigns}, Su available_signature_algs(_, _, _, _) -> undefined. +server_name(_, _, server) -> + undefined; %% Not interesting to check your own name. +server_name(undefined, Host, client) -> + {fallback, Host}; %% Fallback to Host argument to connect +server_name(SNI, _, client) -> + SNI. %% If Server Name Indication is available diff --git a/lib/ssl/src/ssl_internal.hrl b/lib/ssl/src/ssl_internal.hrl index 24ac34653e..9bb1cbaeb0 100644 --- a/lib/ssl/src/ssl_internal.hrl +++ b/lib/ssl/src/ssl_internal.hrl @@ -95,7 +95,8 @@ certfile :: binary(), cert :: public_key:der_encoded() | secret_printout() | 'undefined', keyfile :: binary(), - key :: {'RSAPrivateKey' | 'DSAPrivateKey' | 'ECPrivateKey' | 'PrivateKeyInfo', public_key:der_encoded()} | secret_printout() | 'undefined', + key :: {'RSAPrivateKey' | 'DSAPrivateKey' | 'ECPrivateKey' | 'PrivateKeyInfo', + public_key:der_encoded()} | key_map() | secret_printout() | 'undefined', password :: string() | secret_printout() | 'undefined', cacerts :: [public_key:der_encoded()] | secret_printout() | 'undefined', cacertfile :: binary(), @@ -164,7 +165,15 @@ connection_cb }). - +-type key_map() :: #{algorithm := rsa | dss | ecdsa, + %% engine and key_id ought to + %% be :=, but putting it in + %% the spec gives dialyzer warning + %% of correct code! + engine => crypto:engine_ref(), + key_id => crypto:key_id(), + password => crypto:password() + }. -type state_name() :: hello | abbreviated | certify | cipher | connection. -type gen_fsm_state_return() :: {next_state, state_name(), term()} | {next_state, state_name(), term(), timeout()} | diff --git a/lib/ssl/src/tls_socket.erl b/lib/ssl/src/tls_socket.erl index e76d9c100a..453a908401 100644 --- a/lib/ssl/src/tls_socket.erl +++ b/lib/ssl/src/tls_socket.erl @@ -27,7 +27,7 @@ -export([send/3, listen/3, accept/3, socket/5, connect/4, upgrade/3, setopts/3, getopts/3, getstat/3, peername/2, sockname/2, port/2]). -export([split_options/1, get_socket_opts/3]). --export([emulated_options/0, internal_inet_values/0, default_inet_values/0, +-export([emulated_options/0, emulated_options/1, internal_inet_values/0, default_inet_values/0, init/1, start_link/3, terminate/2, inherit_tracker/3, emulated_socket_options/2, get_emulated_opts/1, set_emulated_opts/2, get_all_opts/1, handle_call/3, handle_cast/2, @@ -170,6 +170,9 @@ port(Transport, Socket) -> emulated_options() -> [mode, packet, active, header, packet_size]. +emulated_options(Opts) -> + emulated_options(Opts, internal_inet_values(), default_inet_values()). + internal_inet_values() -> [{packet_size,0}, {packet, 0}, {header, 0}, {active, false}, {mode,binary}]. @@ -328,3 +331,41 @@ emulated_socket_options(InetValues, #socket_options{ packet = proplists:get_value(packet, InetValues, Packet), packet_size = proplists:get_value(packet_size, InetValues, Size) }. + +emulated_options([{mode, Value} = Opt |Opts], Inet, Emulated) -> + validate_inet_option(mode, Value), + emulated_options(Opts, Inet, [Opt | proplists:delete(mode, Emulated)]); +emulated_options([{header, Value} = Opt | Opts], Inet, Emulated) -> + validate_inet_option(header, Value), + emulated_options(Opts, Inet, [Opt | proplists:delete(header, Emulated)]); +emulated_options([{active, Value} = Opt |Opts], Inet, Emulated) -> + validate_inet_option(active, Value), + emulated_options(Opts, Inet, [Opt | proplists:delete(active, Emulated)]); +emulated_options([{packet, Value} = Opt |Opts], Inet, Emulated) -> + validate_inet_option(packet, Value), + emulated_options(Opts, Inet, [Opt | proplists:delete(packet, Emulated)]); +emulated_options([{packet_size, Value} = Opt | Opts], Inet, Emulated) -> + validate_inet_option(packet_size, Value), + emulated_options(Opts, Inet, [Opt | proplists:delete(packet_size, Emulated)]); +emulated_options([Opt|Opts], Inet, Emulated) -> + emulated_options(Opts, [Opt|Inet], Emulated); +emulated_options([], Inet,Emulated) -> + {Inet, Emulated}. + +validate_inet_option(mode, Value) + when Value =/= list, Value =/= binary -> + throw({error, {options, {mode,Value}}}); +validate_inet_option(packet, Value) + when not (is_atom(Value) orelse is_integer(Value)) -> + throw({error, {options, {packet,Value}}}); +validate_inet_option(packet_size, Value) + when not is_integer(Value) -> + throw({error, {options, {packet_size,Value}}}); +validate_inet_option(header, Value) + when not is_integer(Value) -> + throw({error, {options, {header,Value}}}); +validate_inet_option(active, Value) + when Value =/= true, Value =/= false, Value =/= once -> + throw({error, {options, {active,Value}}}); +validate_inet_option(_, _) -> + ok. diff --git a/lib/ssl/test/Makefile b/lib/ssl/test/Makefile index c7e2f402af..aa01552c39 100644 --- a/lib/ssl/test/Makefile +++ b/lib/ssl/test/Makefile @@ -44,6 +44,7 @@ MODULES = \ ssl_certificate_verify_SUITE\ ssl_crl_SUITE\ ssl_dist_SUITE \ + ssl_engine_SUITE\ ssl_handshake_SUITE \ ssl_npn_hello_SUITE \ ssl_npn_handshake_SUITE \ diff --git a/lib/ssl/test/ssl_engine_SUITE.erl b/lib/ssl/test/ssl_engine_SUITE.erl new file mode 100644 index 0000000000..bc221d35fd --- /dev/null +++ b/lib/ssl/test/ssl_engine_SUITE.erl @@ -0,0 +1,142 @@ +%% +%% %CopyrightBegin% +%% +%% Copyright Ericsson AB 2017-2017. All Rights Reserved. +%% +%% Licensed under the Apache License, Version 2.0 (the "License"); +%% you may not use this file except in compliance with the License. +%% You may obtain a copy of the License at +%% +%% http://www.apache.org/licenses/LICENSE-2.0 +%% +%% Unless required by applicable law or agreed to in writing, software +%% distributed under the License is distributed on an "AS IS" BASIS, +%% WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +%% See the License for the specific language governing permissions and +%% limitations under the License. +%% +%% %CopyrightEnd% +%% + +%% +-module(ssl_engine_SUITE). + +%% Note: This directive should only be used in test suites. +-compile(export_all). + +-include_lib("common_test/include/ct.hrl"). +-include_lib("public_key/include/public_key.hrl"). + +%%-------------------------------------------------------------------- +%% Common Test interface functions ----------------------------------- +%%-------------------------------------------------------------------- +all() -> + [ + private_key + ]. + +init_per_suite(Config) -> + catch crypto:stop(), + try crypto:start() of + ok -> + ssl_test_lib:clean_start(), + case crypto:get_test_engine() of + {ok, EngineName} -> + try crypto:engine_load(<<"dynamic">>, + [{<<"SO_PATH">>, EngineName}, + <<"LOAD">>], + []) of + {ok, Engine} -> + [{engine, Engine} |Config]; + {error, Reason} -> + ct:pal("Reason ~p", [Reason]), + {skip, "No dynamic engine support"} + catch error:notsup -> + {skip, "No engine support in OpenSSL"} + end; + {error, notexist} -> + {skip, "Test engine not found"} + end + catch _:_ -> + {skip, "Crypto did not start"} + end. + +end_per_suite(Config) -> + Engine = proplists:get_value(engine, Config), + crypto:engine_unload(Engine), + ssl:stop(), + application:stop(crypto). + + +init_per_testcase(_TestCase, Config) -> + ssl:stop(), + ssl:start(), + ssl_test_lib:ct_log_supported_protocol_versions(Config), + ct:timetrap({seconds, 10}), + Config. + +end_per_testcase(_TestCase, Config) -> + Config. + +%%-------------------------------------------------------------------- +%% Test Cases -------------------------------------------------------- +%%-------------------------------------------------------------------- + +private_key(Config) when is_list(Config) -> + ClientFileBase = filename:join([proplists:get_value(priv_dir, Config), "client_engine"]), + ServerFileBase = filename:join([proplists:get_value(priv_dir, Config), "server_engine"]), + #{server_config := ServerConf, + client_config := ClientConf} = GenCertData = + public_key:pkix_test_data(#{server_chain => + #{root => [{key, ssl_test_lib:hardcode_rsa_key(1)}], + intermediates => [[{key, ssl_test_lib:hardcode_rsa_key(2)}]], + peer => [{key, ssl_test_lib:hardcode_rsa_key(3)} + ]}, + client_chain => + #{root => [{key, ssl_test_lib:hardcode_rsa_key(4)}], + intermediates => [[{key, ssl_test_lib:hardcode_rsa_key(5)}]], + peer => [{key, ssl_test_lib:hardcode_rsa_key(6)}]}}), + [{server_config, FileServerConf}, + {client_config, FileClientConf}] = + x509_test:gen_pem_config_files(GenCertData, ClientFileBase, ServerFileBase), + + Engine = proplists:get_value(engine, Config), + + ClientKey = engine_key(FileClientConf), + ServerKey = engine_key(FileServerConf), + + EngineClientConf = [{key, #{algorithm => rsa, + engine => Engine, + key_id => ClientKey}} | proplists:delete(key, ClientConf)], + + EngineServerConf = [{key, #{algorithm => rsa, + engine => Engine, + key_id => ServerKey}} | proplists:delete(key, ServerConf)], + %% Test with engine + test_tls_connection(EngineServerConf, EngineClientConf, Config), + %% Test that sofware fallback is available + test_tls_connection(ServerConf, [{reuse_sessions, false} |ClientConf], Config). + +engine_key(Conf) -> + FileStr = proplists:get_value(keyfile, Conf), + list_to_binary(FileStr). + + +test_tls_connection(ServerConf, ClientConf, Config) -> + {ClientNode, ServerNode, Hostname} = ssl_test_lib:run_where(Config), + + Server = ssl_test_lib:start_server([{node, ServerNode}, {port, 0}, + {from, self()}, + {mfa, {ssl_test_lib, send_recv_result_active, []}}, + {options, [{verify, verify_peer} + | ServerConf]}]), + Port = ssl_test_lib:inet_port(Server), + Client = ssl_test_lib:start_client([{node, ClientNode}, {port, Port}, + {host, Hostname}, + {from, self()}, + {mfa, {ssl_test_lib, send_recv_result_active, []}}, + {options, [{verify, verify_peer} | ClientConf]}]), + + ssl_test_lib:check_result(Server, ok, Client, ok), + ssl_test_lib:close(Server), + ssl_test_lib:close(Client). diff --git a/lib/ssl/test/ssl_packet_SUITE.erl b/lib/ssl/test/ssl_packet_SUITE.erl index 408d62ce9c..3261244ace 100644 --- a/lib/ssl/test/ssl_packet_SUITE.erl +++ b/lib/ssl/test/ssl_packet_SUITE.erl @@ -63,8 +63,10 @@ groups() -> {'tlsv1.1', [], socket_packet_tests() ++ protocol_packet_tests()}, {'tlsv1', [], socket_packet_tests() ++ protocol_packet_tests()}, {'sslv3', [], socket_packet_tests() ++ protocol_packet_tests()}, - {'dtlsv1.2', [], protocol_packet_tests()}, - {'dtlsv1', [], protocol_packet_tests()} + %% We will not support any packet types if the transport is + %% not reliable. We might support it for DTLS over SCTP in the future + {'dtlsv1.2', [], [reject_packet_opt]}, + {'dtlsv1', [], [reject_packet_opt]} ]. socket_packet_tests() -> @@ -1924,6 +1926,25 @@ header_decode_two_bytes_one_sent_passive(Config) when is_list(Config) -> ssl_test_lib:close(Client). %%-------------------------------------------------------------------- +reject_packet_opt() -> + [{doc,"Test packet option is rejected for DTLS over udp"}]. + +reject_packet_opt(Config) when is_list(Config) -> + + ServerOpts = ssl_test_lib:ssl_options(server_opts, Config), + + {error,{options,{not_supported,{packet,4}}}} = + ssl:listen(9999, [{packet, 4} | ServerOpts]), + {error,{options,{not_supported,{packet_size,1}}}} = + ssl:listen(9999, [{packet_size, 1} | ServerOpts]), + {error,{options,{not_supported,{header,1}}}} = + ssl:listen(9999, [{header, 1} | ServerOpts]), + + client_reject_packet_opt(Config, {packet,4}), + client_reject_packet_opt(Config, {packet_size, 1}), + client_reject_packet_opt(Config, {header, 1}). + +%%-------------------------------------------------------------------- %% Internal functions ------------------------------------------------ %%-------------------------------------------------------------------- @@ -2245,3 +2266,23 @@ add_tpkt_header(IOList) when is_list(IOList) -> Binary = list_to_binary(IOList), L = size(Binary) + 4, [3, 0, ((L) bsr 8) band 16#ff, (L) band 16#ff , Binary]. + + +client_reject_packet_opt(Config, PacketOpt) -> + ServerOpts = ssl_test_lib:ssl_options(server_opts, Config), + ClientOpts = ssl_test_lib:ssl_options(client_opts, Config), + {ClientNode, ServerNode, Hostname} = ssl_test_lib:run_where(Config), + + Server = ssl_test_lib:start_server([{node, ClientNode}, {port, 0}, + {from, self()}, + {mfa, {ssl_test_lib, no_result_msg ,[]}}, + {options, ServerOpts}]), + Port = ssl_test_lib:inet_port(Server), + Client = ssl_test_lib:start_client_error([{node, ServerNode}, {port, Port}, + {host, Hostname}, + {from, self()}, + {mfa, {ssl_test_lib, no_result_msg, []}}, + {options, [PacketOpt | + ClientOpts]}]), + + ssl_test_lib:check_result(Client, {error, {options, {not_supported, PacketOpt}}}). diff --git a/lib/ssl/test/ssl_sni_SUITE.erl b/lib/ssl/test/ssl_sni_SUITE.erl index 03676cb828..7e78c41444 100644 --- a/lib/ssl/test/ssl_sni_SUITE.erl +++ b/lib/ssl/test/ssl_sni_SUITE.erl @@ -25,6 +25,8 @@ -include_lib("common_test/include/ct.hrl"). -include_lib("public_key/include/public_key.hrl"). +-include_lib("kernel/include/inet.hrl"). + %%-------------------------------------------------------------------- %% Common Test interface functions ----------------------------------- @@ -55,7 +57,11 @@ sni_tests() -> sni_no_match, no_sni_header_fun, sni_match_fun, - sni_no_match_fun]. + sni_no_match_fun, + dns_name, + ip_fallback, + no_ip_fallback, + dns_name_reuse]. init_per_suite(Config0) -> catch crypto:stop(), @@ -82,6 +88,13 @@ end_per_suite(_) -> ssl:stop(), application:stop(crypto). +init_per_testcase(TestCase, Config) when TestCase == ip_fallback; + TestCase == no_ip_fallback; + TestCase == dns_name_reuse -> + ssl_test_lib:ct_log_supported_protocol_versions(Config), + ct:log("Ciphers: ~p~n ", [ ssl:cipher_suites()]), + ct:timetrap({seconds, 20}), + Config; init_per_testcase(_TestCase, Config) -> ssl_test_lib:ct_log_supported_protocol_versions(Config), ct:log("Ciphers: ~p~n ", [ ssl:cipher_suites()]), @@ -112,7 +125,119 @@ sni_no_match(Config) -> sni_no_match_fun(Config) -> run_sni_fun_handshake(Config, "c.server", undefined, "server Peer cert"). +dns_name(Config) -> + Hostname = "OTP.test.server", + #{server_config := ServerConf, + client_config := ClientConf} = public_key:pkix_test_data(#{server_chain => + #{root => [{key, ssl_test_lib:hardcode_rsa_key(1)}], + intermediates => [[{key, ssl_test_lib:hardcode_rsa_key(2)}]], + peer => [{extensions, [#'Extension'{extnID = + ?'id-ce-subjectAltName', + extnValue = [{dNSName, Hostname}], + critical = false}]}, + {key, ssl_test_lib:hardcode_rsa_key(3)}]}, + client_chain => + #{root => [{key, ssl_test_lib:hardcode_rsa_key(4)}], + intermediates => [[{key, ssl_test_lib:hardcode_rsa_key(5)}]], + peer => [{key, ssl_test_lib:hardcode_rsa_key(6)}]}}), + unsuccessfull_connect(ServerConf, [{verify, verify_peer} | ClientConf], undefined, Config), + successfull_connect(ServerConf, [{verify, verify_peer}, {server_name_indication, Hostname} | ClientConf], undefined, Config), + unsuccessfull_connect(ServerConf, [{verify, verify_peer}, {server_name_indication, "foo"} | ClientConf], undefined, Config), + successfull_connect(ServerConf, [{verify, verify_peer}, {server_name_indication, disable} | ClientConf], undefined, Config). + +ip_fallback(Config) -> + Hostname = net_adm:localhost(), + {ok, #hostent{h_addr_list = [IP |_]}} = inet:gethostbyname(net_adm:localhost()), + IPStr = tuple_to_list(IP), + #{server_config := ServerConf, + client_config := ClientConf} = public_key:pkix_test_data(#{server_chain => + #{root => [{key, ssl_test_lib:hardcode_rsa_key(1)}], + intermediates => [[{key, ssl_test_lib:hardcode_rsa_key(2)}]], + peer => [{extensions, [#'Extension'{extnID = + ?'id-ce-subjectAltName', + extnValue = [{dNSName, Hostname}, + {iPAddress, IPStr}], + critical = false}]}, + {key, ssl_test_lib:hardcode_rsa_key(3)}]}, + client_chain => + #{root => [{key, ssl_test_lib:hardcode_rsa_key(4)}], + intermediates => [[{key, ssl_test_lib:hardcode_rsa_key(5)}]], + peer => [{key, ssl_test_lib:hardcode_rsa_key(6)}]}}), + successfull_connect(ServerConf, [{verify, verify_peer} | ClientConf], Hostname, Config), + successfull_connect(ServerConf, [{verify, verify_peer} | ClientConf], IP, Config). + +no_ip_fallback(Config) -> + Hostname = net_adm:localhost(), + {ok, #hostent{h_addr_list = [IP |_]}} = inet:gethostbyname(net_adm:localhost()), + #{server_config := ServerConf, + client_config := ClientConf} = public_key:pkix_test_data(#{server_chain => + #{root => [{key, ssl_test_lib:hardcode_rsa_key(1)}], + intermediates => [[{key, ssl_test_lib:hardcode_rsa_key(2)}]], + peer => [{extensions, [#'Extension'{extnID = + ?'id-ce-subjectAltName', + extnValue = [{dNSName, Hostname}], + critical = false}]}, + {key, ssl_test_lib:hardcode_rsa_key(3)} + ]}, + client_chain => + #{root => [{key, ssl_test_lib:hardcode_rsa_key(4)}], + intermediates => [[{key, ssl_test_lib:hardcode_rsa_key(5)}]], + peer => [{key, ssl_test_lib:hardcode_rsa_key(6)}]}}), + successfull_connect(ServerConf, [{verify, verify_peer} | ClientConf], Hostname, Config), + unsuccessfull_connect(ServerConf, [{verify, verify_peer} | ClientConf], IP, Config). +dns_name_reuse(Config) -> + SNIHostname = "OTP.test.server", + #{server_config := ServerConf, + client_config := ClientConf} = public_key:pkix_test_data(#{server_chain => + #{root => [{key, ssl_test_lib:hardcode_rsa_key(1)}], + intermediates => [[{key, ssl_test_lib:hardcode_rsa_key(2)}]], + peer => [{extensions, [#'Extension'{extnID = + ?'id-ce-subjectAltName', + extnValue = [{dNSName, SNIHostname}], + critical = false} + ]}, + {key, ssl_test_lib:hardcode_rsa_key(3)} + ]}, + client_chain => + #{root => [{key, ssl_test_lib:hardcode_rsa_key(4)}], + intermediates => [[{key, ssl_test_lib:hardcode_rsa_key(5)}]], + peer => [{key, ssl_test_lib:hardcode_rsa_key(6)}]}}), + + {ClientNode, ServerNode, Hostname} = ssl_test_lib:run_where(Config), + + unsuccessfull_connect(ServerConf, [{verify, verify_peer} | ClientConf], undefined, Config), + + Server = + ssl_test_lib:start_server([{node, ServerNode}, {port, 0}, + {from, self()}, + {mfa, {ssl_test_lib, session_info_result, []}}, + {options, ServerConf}]), + Port = ssl_test_lib:inet_port(Server), + Client0 = + ssl_test_lib:start_client([{node, ClientNode}, + {port, Port}, {host, Hostname}, + {mfa, {ssl_test_lib, no_result, []}}, + {from, self()}, {options, [{verify, verify_peer}, + {server_name_indication, SNIHostname} | ClientConf]}]), + receive + {Server, _} -> + ok + end, + + Server ! {listen, {mfa, {ssl_test_lib, no_result, []}}}, + + %% Make sure session is registered + ct:sleep(1000), + + Client1 = + ssl_test_lib:start_client_error([{node, ClientNode}, + {port, Port}, {host, Hostname}, + {mfa, {ssl_test_lib, session_info_result, []}}, + {from, self()}, {options, [{verify, verify_peer} | ClientConf]}]), + + ssl_test_lib:check_result(Client1, {error, {tls_alert, "handshake failure"}}), + ssl_test_lib:close(Client0). %%-------------------------------------------------------------------- %% Internal Functions ------------------------------------------------ %%-------------------------------------------------------------------- @@ -217,3 +342,37 @@ run_handshake(Config, SNIHostname, ExpectedSNIHostname, ExpectedCN) -> ssl_test_lib:check_result(Server, ExpectedSNIHostname, Client, ExpectedCN), ssl_test_lib:close(Server), ssl_test_lib:close(Client). + +successfull_connect(ServerOptions, ClientOptions, Hostname0, Config) -> + {ClientNode, ServerNode, Hostname1} = ssl_test_lib:run_where(Config), + Hostname = host_name(Hostname0, Hostname1), + Server = ssl_test_lib:start_server([{node, ServerNode}, {port, 0}, + {from, self()}, {mfa, {ssl_test_lib, send_recv_result_active, []}}, + {options, ServerOptions}]), + Port = ssl_test_lib:inet_port(Server), + Client = ssl_test_lib:start_client([{node, ClientNode}, {port, Port}, + {host, Hostname}, {from, self()}, + {mfa, {ssl_test_lib, send_recv_result_active, []}}, + {options, ClientOptions}]), + ssl_test_lib:check_result(Server, ok, Client, ok), + ssl_test_lib:close(Server), + ssl_test_lib:close(Client). + +unsuccessfull_connect(ServerOptions, ClientOptions, Hostname0, Config) -> + {ClientNode, ServerNode, Hostname1} = ssl_test_lib:run_where(Config), + Hostname = host_name(Hostname0, Hostname1), + Server = ssl_test_lib:start_server_error([{node, ServerNode}, {port, 0}, + {from, self()}, + {options, ServerOptions}]), + Port = ssl_test_lib:inet_port(Server), + Client = ssl_test_lib:start_client_error([{node, ClientNode}, {port, Port}, + {host, Hostname}, + {from, self()}, + {options, ClientOptions}]), + + ssl_test_lib:check_result(Server, {error, {tls_alert, "handshake failure"}}, + Client, {error, {tls_alert, "handshake failure"}}). +host_name(undefined, Hostname) -> + Hostname; +host_name(Hostname, _) -> + Hostname. diff --git a/lib/ssl/test/x509_test.erl b/lib/ssl/test/x509_test.erl index 031fad1216..fea01efdaf 100644 --- a/lib/ssl/test/x509_test.erl +++ b/lib/ssl/test/x509_test.erl @@ -64,15 +64,12 @@ do_gen_pem_config_files(Config, CertFile, KeyFile, CAFile) -> cert_entry(Cert) -> {'Certificate', Cert, not_encrypted}. -key_entry(Key = #'RSAPrivateKey'{}) -> - Der = public_key:der_encode('RSAPrivateKey', Key), - {'RSAPrivateKey', Der, not_encrypted}; -key_entry(Key = #'DSAPrivateKey'{}) -> - Der = public_key:der_encode('DSAPrivateKey', Key), - {'DSAPrivateKey', Der, not_encrypted}; -key_entry(Key = #'ECPrivateKey'{}) -> - Der = public_key:der_encode('ECPrivateKey', Key), - {'ECPrivateKey', Der, not_encrypted}. +key_entry({'RSAPrivateKey', DERKey}) -> + {'RSAPrivateKey', DERKey, not_encrypted}; +key_entry({'DSAPrivateKey', DERKey}) -> + {'DSAPrivateKey', DERKey, not_encrypted}; +key_entry({'ECPrivateKey', DERKey}) -> + {'ECPrivateKey', DERKey, not_encrypted}. ca_entries(CAs) -> [{'Certificate', CACert, not_encrypted} || CACert <- CAs]. diff --git a/lib/stdlib/doc/src/Makefile b/lib/stdlib/doc/src/Makefile index 93eac8220d..aeed79408b 100644 --- a/lib/stdlib/doc/src/Makefile +++ b/lib/stdlib/doc/src/Makefile @@ -98,6 +98,7 @@ XML_REF3_FILES = \ sys.xml \ timer.xml \ unicode.xml \ + uri_string.xml \ win32reg.xml \ zip.xml diff --git a/lib/stdlib/doc/src/rand.xml b/lib/stdlib/doc/src/rand.xml index 89fb858823..21f680a0ee 100644 --- a/lib/stdlib/doc/src/rand.xml +++ b/lib/stdlib/doc/src/rand.xml @@ -133,8 +133,9 @@ variable <c>rand_seed</c> to remember the current state.</p> <p>If a process calls - <seealso marker="#uniform-0"><c>uniform/0</c></seealso> or - <seealso marker="#uniform-1"><c>uniform/1</c></seealso> without + <seealso marker="#uniform-0"><c>uniform/0</c></seealso>, + <seealso marker="#uniform-1"><c>uniform/1</c></seealso> or + <seealso marker="#uniform_real-0"><c>uniform_real/0</c></seealso> without setting a seed first, <seealso marker="#seed-1"><c>seed/1</c></seealso> is called automatically with the default algorithm and creates a non-constant seed.</p> @@ -168,10 +169,17 @@ R3 = rand:uniform(),</pre> S0 = rand:seed_s(exrop), {R4, S1} = rand:uniform_s(S0),</pre> + <p>Textbook basic form Box-Muller standard normal deviate</p> + + <pre> +R5 = rand:uniform_real(), +R6 = rand:uniform(), +SND0 = math:sqrt(-2 * math:log(R5)) * math:cos(math:pi() * R6)</pre> + <p>Create a standard normal deviate:</p> <pre> -{SND0, S2} = rand:normal_s(S1),</pre> +{SND1, S2} = rand:normal_s(S1),</pre> <p>Create a normal deviate with mean -3 and variance 0.5:</p> @@ -414,7 +422,8 @@ tests. We suggest to use a sign test to extract a random Boolean value.</pre> This function may return exactly <c>0.0</c> which can be fatal for certain applications. If that is undesired you can use <c>(1.0 - rand:uniform())</c> to get the - interval <c>0.0 < <anno>X</anno> =< 1.0</c>. + interval <c>0.0 < <anno>X</anno> =< 1.0</c>, or instead use + <seealso marker="#uniform_real-0"><c>uniform_real/0</c></seealso>. </p> <p> If neither endpoint is desired you can test and re-try @@ -432,6 +441,42 @@ end.</pre> </func> <func> + <name name="uniform_real" arity="0"/> + <fsummary>Return a random float.</fsummary> + <desc><marker id="uniform_real-0"/> + <p> + Returns a random float + uniformly distributed in the value range + <c>DBL_MIN =< <anno>X</anno> < 1.0</c> + and updates the state in the process dictionary. + </p> + <p> + Conceptually, a random real number <c>R</c> is generated + from the interval <c>0 =< R < 1</c> and then the + closest rounded down normalized number + in the IEEE 754 Double precision format + is returned. + </p> + <note> + <p> + The generated numbers from this function has got better + granularity for small numbers than the regular + <seealso marker="#uniform-0"><c>uniform/0</c></seealso> + because all bits in the mantissa are random. + This property, in combination with the fact that exactly zero + is never returned is useful for algoritms doing for example + <c>1.0 / <anno>X</anno></c> or <c>math:log(<anno>X</anno>)</c>. + </p> + </note> + <p> + See + <seealso marker="#uniform_real_s-1"><c>uniform_real_s/1</c></seealso> + for more explanation. + </p> + </desc> + </func> + + <func> <name name="uniform" arity="1"/> <fsummary>Return a random integer.</fsummary> <desc><marker id="uniform-1"/> @@ -460,7 +505,8 @@ end.</pre> This function may return exactly <c>0.0</c> which can be fatal for certain applications. If that is undesired you can use <c>(1.0 - rand:uniform(State))</c> to get the - interval <c>0.0 < <anno>X</anno> =< 1.0</c>. + interval <c>0.0 < <anno>X</anno> =< 1.0</c>, or instead use + <seealso marker="#uniform_real_s-1"><c>uniform_real_s/1</c></seealso>. </p> <p> If neither endpoint is desired you can test and re-try @@ -478,6 +524,68 @@ end.</pre> </func> <func> + <name name="uniform_real_s" arity="1"/> + <fsummary>Return a random float.</fsummary> + <desc> + <p> + Returns, for a specified state, a random float + uniformly distributed in the value range + <c>DBL_MIN =< <anno>X</anno> < 1.0</c> + and updates the state in the process dictionary. + </p> + <p> + Conceptually, a random real number <c>R</c> is generated + from the interval <c>0 =< R < 1</c> and then the + closest rounded down normalized number + in the IEEE 754 Double precision format + is returned. + </p> + <note> + <p> + The generated numbers from this function has got better + granularity for small numbers than the regular + <seealso marker="#uniform_s-1"><c>uniform_s/1</c></seealso> + because all bits in the mantissa are random. + This property, in combination with the fact that exactly zero + is never returned is useful for algoritms doing for example + <c>1.0 / <anno>X</anno></c> or <c>math:log(<anno>X</anno>)</c>. + </p> + </note> + <p> + The concept implicates that the probability to get + exactly zero is extremely low; so low that this function + is in fact guaranteed to never return zero. The smallest + number that it might return is <c>DBL_MIN</c>, which is + 2.0^(-1022). + </p> + <p> + The value range stated at the top of this function + description is technically correct, but + <c>0.0 =< <anno>X</anno> < 1.0</c> + is a better description of the generated numbers' + statistical distribution. Except that exactly 0.0 + is never returned, which is not possible to observe + statistically. + </p> + <p> + For example; for all sub ranges + <c>N*2.0^(-53) =< X < (N+1)*2.0^(-53)</c> + where + <c>0 =< integer(N) < 2.0^53</c> + the probability is the same. + Compare that with the form of the numbers generated by + <seealso marker="#uniform_s-1"><c>uniform_s/1</c></seealso>. + </p> + <p> + Having to generate extra random bits for + small numbers costs a little performance. + This function is about 20% slower than the regular + <seealso marker="#uniform_s-1"><c>uniform_s/1</c></seealso> + </p> + </desc> + </func> + + <func> <name name="uniform_s" arity="2"/> <fsummary>Return a random integer.</fsummary> <desc> diff --git a/lib/stdlib/doc/src/ref_man.xml b/lib/stdlib/doc/src/ref_man.xml index 878a3babc5..68bfddbc71 100644 --- a/lib/stdlib/doc/src/ref_man.xml +++ b/lib/stdlib/doc/src/ref_man.xml @@ -93,6 +93,7 @@ <xi:include href="sys.xml"/> <xi:include href="timer.xml"/> <xi:include href="unicode.xml"/> + <xi:include href="uri_string.xml"/> <xi:include href="win32reg.xml"/> <xi:include href="zip.xml"/> </application> diff --git a/lib/stdlib/doc/src/specs.xml b/lib/stdlib/doc/src/specs.xml index 45b207b13d..d559adf9b6 100644 --- a/lib/stdlib/doc/src/specs.xml +++ b/lib/stdlib/doc/src/specs.xml @@ -60,6 +60,7 @@ <xi:include href="../specs/specs_sys.xml"/> <xi:include href="../specs/specs_timer.xml"/> <xi:include href="../specs/specs_unicode.xml"/> + <xi:include href="../specs/specs_uri_string.xml"/> <xi:include href="../specs/specs_win32reg.xml"/> <xi:include href="../specs/specs_zip.xml"/> </specs> diff --git a/lib/stdlib/doc/src/uri_string.xml b/lib/stdlib/doc/src/uri_string.xml new file mode 100644 index 0000000000..9ace2b0a05 --- /dev/null +++ b/lib/stdlib/doc/src/uri_string.xml @@ -0,0 +1,230 @@ +<?xml version="1.0" encoding="utf-8" ?> +<!DOCTYPE erlref SYSTEM "erlref.dtd"> + +<erlref> + <header> + <copyright> + <year>2017</year><year>2017</year> + <holder>Ericsson AB. All Rights Reserved.</holder> + </copyright> + <legalnotice> + Licensed under the Apache License, Version 2.0 (the "License"); + you may not use this file except in compliance with the License. + You may obtain a copy of the License at + + http://www.apache.org/licenses/LICENSE-2.0 + + Unless required by applicable law or agreed to in writing, software + distributed under the License is distributed on an "AS IS" BASIS, + WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + See the License for the specific language governing permissions and + limitations under the License. + </legalnotice> + + <title>uri_string</title> + <prepared>Péter Dimitrov</prepared> + <docno>1</docno> + <date>2017-10-24</date> + <rev>A</rev> + </header> + <module>uri_string</module> + <modulesummary>URI processing functions.</modulesummary> + <description> + <p>This module contains functions for parsing and handling URIs + (<url href="https://www.ietf.org/rfc/rfc3986.txt">RFC 3986</url>). + </p> + <p>A URI is an identifier consisting of a sequence of characters matching the syntax + rule named <em>URI</em> in <url href="https://www.ietf.org/rfc/rfc3986.txt">RFC 3986</url>. + </p> + <p> The generic URI syntax consists of a hierarchical sequence of components referred + to as the scheme, authority, path, query, and fragment:</p> + <pre> + URI = scheme ":" hier-part [ "?" query ] [ "#" fragment ] + hier-part = "//" authority path-abempty + / path-absolute + / path-rootless + / path-empty + scheme = ALPHA *( ALPHA / DIGIT / "+" / "-" / "." ) + authority = [ userinfo "@" ] host [ ":" port ] + userinfo = *( unreserved / pct-encoded / sub-delims / ":" ) + + reserved = gen-delims / sub-delims + gen-delims = ":" / "/" / "?" / "#" / "[" / "]" / "@" + sub-delims = "!" / "$" / "&" / "'" / "(" / ")" + / "*" / "+" / "," / ";" / "=" + + unreserved = ALPHA / DIGIT / "-" / "." / "_" / "~" + </pre><br></br> + <p>The interpretation of a URI depends only on the characters used and not on how those + characters are represented in a network protocol.</p> + <p>The functions implemented by this module cover the following use cases:</p> + <list type="bulleted"> + <item>Parsing URIs into its components and returing a map<br></br> + <seealso marker="#parse/1"><c>parse/1</c></seealso> + </item> + <item>Recomposing a map of URI components into a URI string<br></br> + <seealso marker="#recompose/1"><c>recompose/1</c></seealso> + </item> + <item>Changing inbound binary and percent-encoding of URIs<br></br> + <seealso marker="#transcode/2"><c>transcode/2</c></seealso> + </item> + <item>Transforming URIs into a normalized form<br></br> + <seealso marker="#normalize/1"><c>normalize/1</c></seealso> + </item> + </list> + <p>There are four different encodings present during the handling of URIs:</p> + <list type="bulleted"> + <item>Inbound binary encoding in binaries</item> + <item>Inbound percent-encoding in lists and binaries</item> + <item>Outbound binary encoding in binaries</item> + <item>Outbound percent-encoding in lists and binaries</item> + </list> + <p>Functions with <c>uri_string()</c> argument accept lists, binaries and + mixed lists (lists with binary elements) as input type. All of the functions but + <c>transcode/2</c> expects input as lists of unicode codepoints, UTF-8 encoded binaries + and UTF-8 percent-encoded URI parts ("%C3%B6" corresponds to the unicode character "ö").</p> + <p>Unless otherwise specified the return value type and encoding are the same as the input + type and encoding. That is, binary input returns binary output, list input returns a list + output but mixed input returns list output.</p> + <p>In case of lists there is only percent-encoding. In binaries, however, both binary encoding + and percent-encoding shall be considered. <c>transcode/2</c> provides the means to convert + between the supported encodings, it takes a <c>uri_string()</c> and a list of options + specifying inbound and outbound encodings.</p> + <p><url href="https://www.ietf.org/rfc/rfc3986.txt">RFC 3986</url> does not mandate any specific + character encoding and it is usually defined by the protocol or surrounding text. This library + takes the same assumption, binary and percent-encoding are handled as one configuration unit, + they cannot be set to different values.</p> + </description> + + <datatypes> + <datatype> + <name name="error"/> + <desc> + <p>Error tuple indicating the type of error. Possible values of the second component:</p> + <list type="bulleted"> + <item><c>invalid_input</c></item> + <item><c>invalid_map</c></item> + <item><c>invalid_percent_encoding</c></item> + <item><c>invalid_scheme</c></item> + <item><c>invalid_uri</c></item> + <item><c>invalid_utf8</c></item> + </list> + <p>The third component is a term providing additional information about the + cause of the error.</p> + </desc> + </datatype> + <datatype> + <name name="uri_map"/> + <desc> + <p>Map holding the main components of a URI.</p> + </desc> + </datatype> + <datatype> + <name name="uri_string"/> + <desc> + <p>List of unicode codepoints, a UTF-8 encoded binary, or a mix of the two, + representing an <url href="https://www.ietf.org/rfc/rfc3986.txt">RFC 3986</url> + compliant URI (<em>percent-encoded form</em>). + A URI is a sequence of characters from a very limited set: the letters of + the basic Latin alphabet, digits, and a few special characters.</p> + </desc> + </datatype> + </datatypes> + + <funcs> + + <func> + <name name="normalize" arity="1"/> + <fsummary>Syntax-based normalization.</fsummary> + <desc> + <p>Transforms <c><anno>URIString</anno></c> into a normalized form + using Syntax-Based Normalization as defined by + <url href="https://www.ietf.org/rfc/rfc3986.txt">RFC 3986</url>.</p> + <p>This function implements case normalization, percent-encoding + normalization, path segment normalization and scheme based normalization + for HTTP(S) with basic support for FTP, SSH, SFTP and TFTP.</p> + <p><em>Example:</em></p> + <pre> +1> <input>uri_string:normalize("/a/b/c/./../../g").</input> +"/a/g" +2> <![CDATA[uri_string:normalize(<<"mid/content=5/../6">>).]]> +<![CDATA[<<"mid/6">>]]> +3> uri_string:normalize("http://localhost:80"). +"https://localhost/" + </pre> + </desc> + </func> + + <func> + <name name="parse" arity="1"/> + <fsummary>Parse URI into a map.</fsummary> + <desc> + <p>Parses an <url href="https://www.ietf.org/rfc/rfc3986.txt">RFC 3986</url> + compliant <c>uri_string()</c> into a <c>uri_map()</c>, that holds the parsed + components of the <c>URI</c>. + If parsing fails, an error tuple is returned.</p> + <p>See also the opposite operation <seealso marker="#recompose/1"> + <c>recompose/1</c></seealso>.</p> + <p><em>Example:</em></p> + <pre> +1> <input>uri_string:parse("foo://[email protected]:8042/over/there?name=ferret#nose").</input> +#{fragment => "nose",host => "example.com", + path => "/over/there",port => 8042,query => "name=ferret", + scheme => foo,userinfo => "user"} +2> <![CDATA[uri_string:parse(<<"foo://[email protected]:8042/over/there?name=ferret">>).]]> +<![CDATA[#{host => <<"example.com">>,path => <<"/over/there">>, + port => 8042,query => <<"name=ferret">>,scheme => <<"foo">>, + userinfo => <<"user">>}]]> + </pre> + </desc> + </func> + + <func> + <name name="recompose" arity="1"/> + <fsummary>Recompose URI.</fsummary> + <desc> + <p>Creates an <url href="https://www.ietf.org/rfc/rfc3986.txt">RFC 3986</url> compliant + <c><anno>URIString</anno></c> (percent-encoded), based on the components of + <c><anno>URIMap</anno></c>. + If the <c><anno>URIMap</anno></c> is invalid, an error tuple is returned.</p> + <p>See also the opposite operation <seealso marker="#parse/1"> + <c>parse/1</c></seealso>.</p> + <p><em>Example:</em></p> + <pre> +1> <input>URIMap = #{fragment => "nose", host => "example.com", path => "/over/there",</input> +1> port => 8042, query => "name=ferret", scheme => "foo", userinfo => "user"}. +#{fragment => "top",host => "example.com", + path => "/over/there",port => 8042,query => "?name=ferret", + scheme => foo,userinfo => "user"} + +2> <input>uri_string:recompose(URIMap).</input> +"foo://example.com:8042/over/there?name=ferret#nose"</pre> + </desc> + </func> + + <func> + <name name="transcode" arity="2"/> + <fsummary>Transcode URI.</fsummary> + <desc> + <p>Transcodes an <url href="https://www.ietf.org/rfc/rfc3986.txt">RFC 3986</url> + compliant <c><anno>URIString</anno></c>, + where <c><anno>Options</anno></c> is a list of tagged tuples, specifying the inbound + (<c>in_encoding</c>) and outbound (<c>out_encoding</c>) encodings. <c>in_encoding</c> + and <c>out_encoding</c> specifies both binary encoding and percent-encoding for the + input and output data. Mixed encoding, where binary encoding is not the same as + percent-encoding, is not supported. + If an argument is invalid, an error tuple is returned.</p> + <p><em>Example:</em></p> + <pre> +1> <input><![CDATA[uri_string:transcode(<<"foo%00%00%00%F6bar"/utf32>>,]]></input> +1> [{in_encoding, utf32},{out_encoding, utf8}]). +<![CDATA[<<"foo%C3%B6bar"/utf8>>]]> +2> uri_string:transcode("foo%F6bar", [{in_encoding, latin1}, +2> {out_encoding, utf8}]). +"foo%C3%B6bar" + </pre> + </desc> + </func> + + </funcs> +</erlref> diff --git a/lib/stdlib/src/Makefile b/lib/stdlib/src/Makefile index bf836203ec..8b156929d7 100644 --- a/lib/stdlib/src/Makefile +++ b/lib/stdlib/src/Makefile @@ -121,6 +121,7 @@ MODULES= \ timer \ unicode \ unicode_util \ + uri_string \ win32reg \ zip diff --git a/lib/stdlib/src/base64.erl b/lib/stdlib/src/base64.erl index 5885745fb1..c8cf6fdffe 100644 --- a/lib/stdlib/src/base64.erl +++ b/lib/stdlib/src/base64.erl @@ -113,9 +113,9 @@ encode_binary(Bin) -> Data :: ascii_binary(). decode(Bin) when is_binary(Bin) -> - decode_binary(<<>>, Bin); + decode_binary(Bin, <<>>); decode(List) when is_list(List) -> - list_to_binary(decode_l(List)). + decode_list(List, <<>>). -spec mime_decode(Base64) -> Data when Base64 :: ascii_string() | ascii_binary(), @@ -186,31 +186,41 @@ mime_decode_to_string(List) when is_list(List) -> bad,bad,bad,bad,bad,bad,bad,bad,bad,bad,bad,bad,bad,bad,bad,bad, bad,bad,bad,bad,bad,bad,bad,bad,bad,bad,bad,bad,bad,bad,bad,bad}). -decode_binary(Result0, <<C:8,T0/bits>>) -> - case element(C, ?DECODE_MAP) of - bad -> - erlang:error({badarg,C}); - ws -> - decode_binary(Result0, T0); - eq -> - case strip_ws(T0) of - <<$=:8,T/binary>> -> - <<>> = strip_ws(T), - Split = byte_size(Result0) - 1, - <<Result:Split/bytes,_:4>> = Result0, - Result; - T -> - <<>> = strip_ws(T), - Split = byte_size(Result0) - 1, - <<Result:Split/bytes,_:2>> = Result0, - Result - end; - Bits -> - decode_binary(<<Result0/bits,Bits:6>>, T0) +decode_binary(<<C1:8, Cs/bits>>, A) -> + case element(C1, ?DECODE_MAP) of + ws -> decode_binary(Cs, A); + B1 -> decode_binary(Cs, A, B1) end; -decode_binary(Result, <<>>) -> - true = is_binary(Result), - Result. +decode_binary(<<>>, A) -> + A. + +decode_binary(<<C2:8, Cs/bits>>, A, B1) -> + case element(C2, ?DECODE_MAP) of + ws -> decode_binary(Cs, A, B1); + B2 -> decode_binary(Cs, A, B1, B2) + end. + +decode_binary(<<C3:8, Cs/bits>>, A, B1, B2) -> + case element(C3, ?DECODE_MAP) of + ws -> decode_binary(Cs, A, B1, B2); + B3 -> decode_binary(Cs, A, B1, B2, B3) + end. + +decode_binary(<<C4:8, Cs/bits>>, A, B1, B2, B3) -> + case element(C4, ?DECODE_MAP) of + ws -> decode_binary(Cs, A, B1, B2, B3); + eq when B3 =:= eq -> only_ws_binary(Cs, <<A/binary,B1:6,(B2 bsr 4):2>>); + eq -> only_ws_binary(Cs, <<A/binary,B1:6,B2:6,(B3 bsr 2):4>>); + B4 -> decode_binary(Cs, <<A/binary,B1:6,B2:6,B3:6,B4:6>>) + end. + +only_ws_binary(<<>>, A) -> + A; +only_ws_binary(<<C:8, Cs/bits>>, A) -> + case element(C, ?DECODE_MAP) of + ws -> only_ws_binary(Cs, A); + _ -> erlang:error(function_clause) + end. %% Skipping pad character if not at end of string. Also liberal about %% excess padding and skipping of other illegal (non-base64 alphabet) @@ -262,6 +272,42 @@ mime_decode_binary_after_eq(Result0, <<>>, Eq) -> Result end. +decode_list([C1 | Cs], A) -> + case element(C1, ?DECODE_MAP) of + ws -> decode_list(Cs, A); + B1 -> decode_list(Cs, A, B1) + end; +decode_list([], A) -> + A. + +decode_list([C2 | Cs], A, B1) -> + case element(C2, ?DECODE_MAP) of + ws -> decode_list(Cs, A, B1); + B2 -> decode_list(Cs, A, B1, B2) + end. + +decode_list([C3 | Cs], A, B1, B2) -> + case element(C3, ?DECODE_MAP) of + ws -> decode_list(Cs, A, B1, B2); + B3 -> decode_list(Cs, A, B1, B2, B3) + end. + +decode_list([C4 | Cs], A, B1, B2, B3) -> + case element(C4, ?DECODE_MAP) of + ws -> decode_list(Cs, A, B1, B2, B3); + eq when B3 =:= eq -> only_ws(Cs, <<A/binary,B1:6,(B2 bsr 4):2>>); + eq -> only_ws(Cs, <<A/binary,B1:6,B2:6,(B3 bsr 2):4>>); + B4 -> decode_list(Cs, <<A/binary,B1:6,B2:6,B3:6,B4:6>>) + end. + +only_ws([], A) -> + A; +only_ws([C | Cs], A) -> + case element(C, ?DECODE_MAP) of + ws -> only_ws(Cs, A); + _ -> erlang:error(function_clause) + end. + decode([], A) -> A; decode([$=,$=,C2,C1|Cs], A) -> Bits2x6 = (b64d(C1) bsl 18) bor (b64d(C2) bsl 12), @@ -292,16 +338,6 @@ strip_spaces([$\r|Cs], A) -> strip_spaces(Cs, A); strip_spaces([$\n|Cs], A) -> strip_spaces(Cs, A); strip_spaces([C|Cs], A) -> strip_spaces(Cs, [C | A]). -strip_ws(<<$\t,T/binary>>) -> - strip_ws(T); -strip_ws(<<$\n,T/binary>>) -> - strip_ws(T); -strip_ws(<<$\r,T/binary>>) -> - strip_ws(T); -strip_ws(<<$\s,T/binary>>) -> - strip_ws(T); -strip_ws(T) -> T. - %% Skipping pad character if not at end of string. Also liberal about %% excess padding and skipping of other illegal (non-base64 alphabet) %% characters. See section 3.3 of RFC4648 diff --git a/lib/stdlib/src/ets.erl b/lib/stdlib/src/ets.erl index 4858c8d13c..b6548626f3 100644 --- a/lib/stdlib/src/ets.erl +++ b/lib/stdlib/src/ets.erl @@ -1700,6 +1700,8 @@ choice(Height, Width, P, Mode, Tab, Key, Turn, Opos) -> io:format("~ts\n", [ErrorString]), choice(Height, Width, P, Mode, Tab, Key, Turn, Opos) end; + eof -> + ok; _ -> choice(Height, Width, P, Mode, Tab, Key, Turn, Opos) end. diff --git a/lib/stdlib/src/filename.erl b/lib/stdlib/src/filename.erl index 919f8f20e6..a322bd002d 100644 --- a/lib/stdlib/src/filename.erl +++ b/lib/stdlib/src/filename.erl @@ -472,6 +472,10 @@ join(Name1, Name2) when is_atom(Name2) -> join1([UcLetter, $:|Rest], RelativeName, [], win32) when is_integer(UcLetter), UcLetter >= $A, UcLetter =< $Z -> join1(Rest, RelativeName, [$:, UcLetter+$a-$A], win32); +join1([$\\,$\\|Rest], RelativeName, [], win32) -> + join1([$/,$/|Rest], RelativeName, [], win32); +join1([$/,$/|Rest], RelativeName, [], win32) -> + join1(Rest, RelativeName, [$/,$/], win32); join1([$\\|Rest], RelativeName, Result, win32) -> join1([$/|Rest], RelativeName, Result, win32); join1([$/|Rest], RelativeName, [$., $/|Result], OsType) -> @@ -500,6 +504,10 @@ join1([Atom|Rest], RelativeName, Result, OsType) when is_atom(Atom) -> join1b(<<UcLetter, $:, Rest/binary>>, RelativeName, [], win32) when is_integer(UcLetter), UcLetter >= $A, UcLetter =< $Z -> join1b(Rest, RelativeName, [$:, UcLetter+$a-$A], win32); +join1b(<<$\\,$\\,Rest/binary>>, RelativeName, [], win32) -> + join1b(<<$/,$/,Rest/binary>>, RelativeName, [], win32); +join1b(<<$/,$/,Rest/binary>>, RelativeName, [], win32) -> + join1b(Rest, RelativeName, [$/,$/], win32); join1b(<<$\\,Rest/binary>>, RelativeName, Result, win32) -> join1b(<<$/,Rest/binary>>, RelativeName, Result, win32); join1b(<<$/,Rest/binary>>, RelativeName, [$., $/|Result], OsType) -> @@ -510,6 +518,8 @@ join1b(<<>>, <<>>, Result, OsType) -> list_to_binary(maybe_remove_dirsep(Result, OsType)); join1b(<<>>, RelativeName, [$:|Rest], win32) -> join1b(RelativeName, <<>>, [$:|Rest], win32); +join1b(<<>>, RelativeName, [$/,$/|Result], win32) -> + join1b(RelativeName, <<>>, [$/,$/|Result], win32); join1b(<<>>, RelativeName, [$/|Result], OsType) -> join1b(RelativeName, <<>>, [$/|Result], OsType); join1b(<<>>, RelativeName, [$., $/|Result], OsType) -> @@ -523,6 +533,8 @@ maybe_remove_dirsep([$/, $:, Letter], win32) -> [Letter, $:, $/]; maybe_remove_dirsep([$/], _) -> [$/]; +maybe_remove_dirsep([$/,$/], win32) -> + [$/,$/]; maybe_remove_dirsep([$/|Name], _) -> lists:reverse(Name); maybe_remove_dirsep(Name, _) -> @@ -712,6 +724,9 @@ win32_splitb(<<Letter0,$:,Rest/binary>>) when ?IS_DRIVELETTER(Letter0) -> Letter = fix_driveletter(Letter0), L = binary:split(Rest,[<<"/">>,<<"\\">>],[global]), [<<Letter,$:>> | [ X || X <- L, X =/= <<>> ]]; +win32_splitb(<<Slash,Slash,Rest/binary>>) when ((Slash =:= $\\) orelse (Slash =:= $/)) -> + L = binary:split(Rest,[<<"/">>,<<"\\">>],[global]), + [<<"//">> | [ X || X <- L, X =/= <<>> ]]; win32_splitb(<<Slash,Rest/binary>>) when ((Slash =:= $\\) orelse (Slash =:= $/)) -> L = binary:split(Rest,[<<"/">>,<<"\\">>],[global]), [<<$/>> | [ X || X <- L, X =/= <<>> ]]; @@ -723,6 +738,8 @@ win32_splitb(Name) -> unix_split(Name) -> split(Name, [], unix). +win32_split([Slash,Slash|Rest]) when ((Slash =:= $\\) orelse (Slash =:= $/)) -> + split(Rest, [[$/,$/]], win32); win32_split([$\\|Rest]) -> win32_split([$/|Rest]); win32_split([X, $\\|Rest]) when is_integer(X) -> diff --git a/lib/stdlib/src/rand.erl b/lib/stdlib/src/rand.erl index 7a8a5e6d4a..362e98006e 100644 --- a/lib/stdlib/src/rand.erl +++ b/lib/stdlib/src/rand.erl @@ -21,8 +21,8 @@ %% Multiple PRNG module for Erlang/OTP %% Copyright (c) 2015-2016 Kenji Rikitake %% -%% exrop (xoroshiro116+) added and statistical distribution -%% improvements by the Erlang/OTP team 2017 +%% exrop (xoroshiro116+) added, statistical distribution +%% improvements and uniform_real added by the Erlang/OTP team 2017 %% ===================================================================== -module(rand). @@ -30,10 +30,14 @@ -export([seed_s/1, seed_s/2, seed/1, seed/2, export_seed/0, export_seed_s/1, uniform/0, uniform/1, uniform_s/1, uniform_s/2, + uniform_real/0, uniform_real_s/1, jump/0, jump/1, normal/0, normal/2, normal_s/1, normal_s/3 ]). +%% Debug +-export([make_float/3, float2str/1, bc64/1]). + -compile({inline, [exs64_next/1, exsplus_next/1, exs1024_next/1, exs1024_calc/2, exrop_next/1, exrop_next_s/2, @@ -60,6 +64,10 @@ %% N i evaluated 3 times (?BSL((Bits), (X), (N)) bor ((X) bsr ((Bits)-(N))))). +-define( + BC(V, N), + bc((V), ?BIT((N) - 1), N)). + %%-define(TWO_POW_MINUS53, (math:pow(2, -53))). -define(TWO_POW_MINUS53, 1.11022302462515657e-16). @@ -84,14 +92,21 @@ %% The 'bits' field indicates how many bits the integer %% returned from 'next' has got, i.e 'next' shall return %% an random integer in the range 0..(2^Bits - 1). -%% At least 53 bits is required for the floating point -%% producing fallbacks. This field is only used when -%% the 'uniform' or 'uniform_n' fields are not defined. +%% At least 55 bits is required for the floating point +%% producing fallbacks, but 56 bits would be more future proof. %% %% The fields 'next', 'uniform' and 'uniform_n' -%% implement the algorithm. If 'uniform' or 'uinform_n' +%% implement the algorithm. If 'uniform' or 'uniform_n' %% is not present there is a fallback using 'next' and either -%% 'bits' or the deprecated 'max'. +%% 'bits' or the deprecated 'max'. The 'next' function +%% must generate a word with at least 56 good random bits. +%% +%% The 'weak_low_bits' field indicate how many bits are of +%% lesser quality and they will not be used by the floating point +%% producing functions, nor by the range producing functions +%% when more bits are needed, to avoid weak bits in the middle +%% of the generated bits. The lowest bits from the range +%% functions still have the generator's quality. %% -type alg_handler() :: #{type := alg(), @@ -148,11 +163,7 @@ %% For ranges larger than the algorithm bit size uniform_range(Range, #{next:=Next, bits:=Bits} = Alg, R, V) -> - WeakLowBits = - case Alg of - #{weak_low_bits:=WLB} -> WLB; - #{} -> 0 - end, + WeakLowBits = maps:get(weak_low_bits, Alg, 0), %% Maybe waste the lowest bit(s) when shifting in new bits Shift = Bits - WeakLowBits, ShiftMask = bnot ?MASK(WeakLowBits), @@ -297,7 +308,7 @@ uniform_s({#{bits:=Bits, next:=Next} = Alg, R0}) -> {(V bsr (Bits - 53)) * ?TWO_POW_MINUS53, {Alg, R1}}; uniform_s({#{max:=Max, next:=Next} = Alg, R0}) -> {V, R1} = Next(R0), - %% Old broken algorithm with non-uniform density + %% Old algorithm with non-uniform density {V / (Max + 1), {Alg, R1}}. @@ -317,7 +328,7 @@ uniform_s(N, {#{bits:=Bits, next:=Next} = Alg, R0}) ?uniform_range(N, Alg, R1, V, MaxMinusN, I); uniform_s(N, {#{max:=Max, next:=Next} = Alg, R0}) when is_integer(N), 1 =< N -> - %% Old broken algorithm with skewed probability + %% Old algorithm with skewed probability %% and gap in ranges > Max {V, R1} = Next(R0), if @@ -328,6 +339,189 @@ uniform_s(N, {#{max:=Max, next:=Next} = Alg, R0}) {trunc(F * N) + 1, {Alg, R1}} end. +%% uniform_real/0: returns a random float X where 0.0 < X =< 1.0, +%% updating the state in the process dictionary. + +-spec uniform_real() -> X :: float(). +uniform_real() -> + {X, Seed} = uniform_real_s(seed_get()), + _ = seed_put(Seed), + X. + +%% uniform_real_s/1: given a state, uniform_s/1 +%% returns a random float X where 0.0 < X =< 1.0, +%% and a new state. +%% +%% This function does not use the same form of uniformity +%% as the uniform_s/1 function. +%% +%% Instead, this function does not generate numbers with equal +%% distance in the interval, but rather tries to keep all mantissa +%% bits random also for small numbers, meaning that the distance +%% between possible numbers decreases when the numbers +%% approaches 0.0, as does the possibility for a particular +%% number. Hence uniformity is preserved. +%% +%% To generate 56 bits at the time instead of 53 is actually +%% a speed optimization since the probability to have to +%% generate a second word decreases by 1/2 for every extra bit. +%% +%% This function generates normalized numbers, so the smallest number +%% that can be generated is 2^-1022 with the distance 2^-1074 +%% to the next to smallest number, compared to 2^-53 for uniform_s/1. +%% +%% This concept of uniformity should work better for applications +%% where you need to calculate 1.0/X or math:log(X) since those +%% operations benefits from larger precision approaching 0.0, +%% and that this function does not return 0.0 nor denormalized +%% numbers very close to 0.0. The log() operation in The Box-Muller +%% transformation for normal distribution is an example of this. +%% +%%-define(TWO_POW_MINUS55, (math:pow(2, -55))). +%%-define(TWO_POW_MINUS110, (math:pow(2, -110))). +%%-define(TWO_POW_MINUS55, 2.7755575615628914e-17). +%%-define(TWO_POW_MINUS110, 7.7037197775489436e-34). +%% +-spec uniform_real_s(State :: state()) -> {X :: float(), NewState :: state()}. +uniform_real_s({#{bits:=Bits, next:=Next} = Alg, R0}) -> + %% Generate a 56 bit number without using the weak low bits. + %% + %% Be sure to use only 53 bits when multiplying with + %% math:pow(2.0, -N) to avoid rounding which would make + %% "even" floats more probable than "odd". + %% + {V1, R1} = Next(R0), + M1 = V1 bsr (Bits - 56), + if + ?BIT(55) =< M1 -> + %% We have 56 bits - waste 3 + {(M1 bsr 3) * math:pow(2.0, -53), {Alg, R1}}; + ?BIT(54) =< M1 -> + %% We have 55 bits - waste 2 + {(M1 bsr 2) * math:pow(2.0, -54), {Alg, R1}}; + ?BIT(53) =< M1 -> + %% We have 54 bits - waste 1 + {(M1 bsr 1) * math:pow(2.0, -55), {Alg, R1}}; + ?BIT(52) =< M1 -> + %% We have 53 bits - use all + {M1 * math:pow(2.0, -56), {Alg, R1}}; + true -> + %% Need more bits + {V2, R2} = Next(R1), + uniform_real_s(Alg, Next, M1, -56, R2, V2, Bits) + end; +uniform_real_s({#{max:=_, next:=Next} = Alg, R0}) -> + %% Generate a 56 bit number. + %% Ignore the weak low bits for these old algorithms, + %% just produce something reasonable. + %% + %% Be sure to use only 53 bits when multiplying with + %% math:pow(2.0, -N) to avoid rounding which would make + %% "even" floats more probable than "odd". + %% + {V1, R1} = Next(R0), + M1 = ?MASK(56, V1), + if + ?BIT(55) =< M1 -> + %% We have 56 bits - waste 3 + {(M1 bsr 3) * math:pow(2.0, -53), {Alg, R1}}; + ?BIT(54) =< M1 -> + %% We have 55 bits - waste 2 + {(M1 bsr 2) * math:pow(2.0, -54), {Alg, R1}}; + ?BIT(53) =< M1 -> + %% We have 54 bits - waste 1 + {(M1 bsr 1) * math:pow(2.0, -55), {Alg, R1}}; + ?BIT(52) =< M1 -> + %% We have 53 bits - use all + {M1 * math:pow(2.0, -56), {Alg, R1}}; + true -> + %% Need more bits + {V2, R2} = Next(R1), + uniform_real_s(Alg, Next, M1, -56, R2, V2, 56) + end. + +uniform_real_s(Alg, _Next, M0, -1064, R1, V1, Bits) -> % 19*56 + %% This is a very theoretical bottom case. + %% The odds of getting here is about 2^-1008, + %% through a white box test case, or thanks to + %% a malfunctioning PRNG producing 18 56-bit zeros in a row. + %% + %% Fill up to 53 bits, we have at most 52 + B0 = (53 - ?BC(M0, 52)), % Missing bits + {(((M0 bsl B0) bor (V1 bsr (Bits - B0))) * math:pow(2.0, -1064 - B0)), + {Alg, R1}}; +uniform_real_s(Alg, Next, M0, BitNo, R1, V1, Bits) -> + if + %% Optimize the most probable. + %% Fill up to 53 bits. + ?BIT(51) =< M0 -> + %% We have 52 bits in M0 - need 1 + {(((M0 bsl 1) bor (V1 bsr (Bits - 1))) + * math:pow(2.0, BitNo - 1)), + {Alg, R1}}; + ?BIT(50) =< M0 -> + %% We have 51 bits in M0 - need 2 + {(((M0 bsl 2) bor (V1 bsr (Bits - 2))) + * math:pow(2.0, BitNo - 2)), + {Alg, R1}}; + ?BIT(49) =< M0 -> + %% We have 50 bits in M0 - need 3 + {(((M0 bsl 3) bor (V1 bsr (Bits - 3))) + * math:pow(2.0, BitNo - 3)), + {Alg, R1}}; + M0 == 0 -> + M1 = V1 bsr (Bits - 56), + if + ?BIT(55) =< M1 -> + %% We have 56 bits - waste 3 + {(M1 bsr 3) * math:pow(2.0, BitNo - 53), {Alg, R1}}; + ?BIT(54) =< M1 -> + %% We have 55 bits - waste 2 + {(M1 bsr 2) * math:pow(2.0, BitNo - 54), {Alg, R1}}; + ?BIT(53) =< M1 -> + %% We have 54 bits - waste 1 + {(M1 bsr 1) * math:pow(2.0, BitNo - 55), {Alg, R1}}; + ?BIT(52) =< M1 -> + %% We have 53 bits - use all + {M1 * math:pow(2.0, BitNo - 56), {Alg, R1}}; + BitNo =:= -1008 -> + %% Endgame + %% For the last round we can not have 14 zeros or more + %% at the top of M1 because then we will underflow, + %% so we need at least 43 bits + if + ?BIT(42) =< M1 -> + %% We have 43 bits - get the last bits + uniform_real_s(Alg, Next, M1, BitNo - 56, R1); + true -> + %% Would underflow 2^-1022 - start all over + %% + %% We could just crash here since the odds for + %% the PRNG being broken is much higher than + %% for a good PRNG generating this many zeros + %% in a row. Maybe we should write an error + %% report or call this a system limit...? + uniform_real_s({Alg, R1}) + end; + true -> + %% Need more bits + uniform_real_s(Alg, Next, M1, BitNo - 56, R1) + end; + true -> + %% Fill up to 53 bits + B0 = 53 - ?BC(M0, 49), % Number of bits we need to append + {(((M0 bsl B0) bor (V1 bsr (Bits - B0))) + * math:pow(2.0, BitNo - B0)), + {Alg, R1}} + end. +%% +uniform_real_s(#{bits:=Bits} = Alg, Next, M0, BitNo, R0) -> + {V1, R1} = Next(R0), + uniform_real_s(Alg, Next, M0, BitNo, R1, V1, Bits); +uniform_real_s(#{max:=_} = Alg, Next, M0, BitNo, R0) -> + {V1, R1} = Next(R0), + uniform_real_s(Alg, Next, M0, BitNo, R1, ?MASK(56, V1), 56). + %% jump/1: given a state, jump/1 %% returns a new state which is equivalent to that %% after a large number of call defined for each algorithm. @@ -1025,3 +1219,42 @@ normal_fi(Indx) -> 1.0214971439701471e-02,8.6165827693987316e-03,7.0508754713732268e-03, 5.5224032992509968e-03,4.0379725933630305e-03,2.6090727461021627e-03, 1.2602859304985975e-03}). + +%%%bitcount64(0) -> 0; +%%%bitcount64(V) -> 1 + bitcount(V, 64). +%%% +%%%-define( +%%% BITCOUNT(V, N), +%%% bitcount(V, N) -> +%%% if +%%% (1 bsl ((N) bsr 1)) =< (V) -> +%%% ((N) bsr 1) + bitcount((V) bsr ((N) bsr 1), ((N) bsr 1)); +%%% true -> +%%% bitcount((V), ((N) bsr 1)) +%%% end). +%%%?BITCOUNT(V, 64); +%%%?BITCOUNT(V, 32); +%%%?BITCOUNT(V, 16); +%%%?BITCOUNT(V, 8); +%%%?BITCOUNT(V, 4); +%%%?BITCOUNT(V, 2); +%%%bitcount(_, 1) -> 0. + +bc64(V) -> ?BC(V, 64). + +%% Linear from high bit - higher probability first gives faster execution +bc(V, B, N) when B =< V -> N; +bc(V, B, N) -> bc(V, B bsr 1, N - 1). + +make_float(S, E, M) -> + <<F/float>> = <<S:1, E:11, M:52>>, + F. + +float2str(N) -> + <<S:1, E:11, M:52>> = <<(float(N))/float>>, + lists:flatten( + io_lib:format( + "~c~c.~13.16.0bE~b", + [case S of 1 -> $-; 0 -> $+ end, + case E of 0 -> $0; _ -> $1 end, + M, E - 16#3ff])). diff --git a/lib/stdlib/src/stdlib.app.src b/lib/stdlib/src/stdlib.app.src index ab0824ca17..5fb48acfab 100644 --- a/lib/stdlib/src/stdlib.app.src +++ b/lib/stdlib/src/stdlib.app.src @@ -101,6 +101,7 @@ timer, unicode, unicode_util, + uri_string, win32reg, zip]}, {registered,[timer_server,rsh_starter,take_over_monitor,pool_master, diff --git a/lib/stdlib/src/uri_string.erl b/lib/stdlib/src/uri_string.erl new file mode 100644 index 0000000000..22212da222 --- /dev/null +++ b/lib/stdlib/src/uri_string.erl @@ -0,0 +1,1842 @@ +%% +%% %CopyrightBegin% +%% +%% Copyright Ericsson AB 2017. All Rights Reserved. +%% +%% Licensed under the Apache License, Version 2.0 (the "License"); +%% you may not use this file except in compliance with the License. +%% You may obtain a copy of the License at +%% +%% http://www.apache.org/licenses/LICENSE-2.0 +%% +%% Unless required by applicable law or agreed to in writing, software +%% distributed under the License is distributed on an "AS IS" BASIS, +%% WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +%% See the License for the specific language governing permissions and +%% limitations under the License. +%% +%% %CopyrightEnd% +%% +%% +%% [RFC 3986, Chapter 2.2. Reserved Characters] +%% +%% reserved = gen-delims / sub-delims +%% +%% gen-delims = ":" / "/" / "?" / "#" / "[" / "]" / "@" +%% +%% sub-delims = "!" / "$" / "&" / "'" / "(" / ")" +%% / "*" / "+" / "," / ";" / "=" +%% +%% +%% [RFC 3986, Chapter 2.3. Unreserved Characters] +%% +%% unreserved = ALPHA / DIGIT / "-" / "." / "_" / "~" +%% +%% +%% [RFC 3986, Chapter 3. Syntax Components] +%% +%% The generic URI syntax consists of a hierarchical sequence of +%% components referred to as the scheme, authority, path, query, and +%% fragment. +%% +%% URI = scheme ":" hier-part [ "?" query ] [ "#" fragment ] +%% +%% hier-part = "//" authority path-abempty +%% / path-absolute +%% / path-rootless +%% / path-empty +%% +%% The scheme and path components are required, though the path may be +%% empty (no characters). When authority is present, the path must +%% either be empty or begin with a slash ("/") character. When +%% authority is not present, the path cannot begin with two slash +%% characters ("//"). These restrictions result in five different ABNF +%% rules for a path (Section 3.3), only one of which will match any +%% given URI reference. +%% +%% The following are two example URIs and their component parts: +%% +%% foo://example.com:8042/over/there?name=ferret#nose +%% \_/ \______________/\_________/ \_________/ \__/ +%% | | | | | +%% scheme authority path query fragment +%% | _____________________|__ +%% / \ / \ +%% urn:example:animal:ferret:nose +%% +%% +%% [RFC 3986, Chapter 3.1. Scheme] +%% +%% Each URI begins with a scheme name that refers to a specification for +%% assigning identifiers within that scheme. +%% +%% scheme = ALPHA *( ALPHA / DIGIT / "+" / "-" / "." ) +%% +%% +%% [RFC 3986, Chapter 3.2. Authority] +%% +%% Many URI schemes include a hierarchical element for a naming +%% authority so that governance of the name space defined by the +%% remainder of the URI is delegated to that authority (which may, in +%% turn, delegate it further). +%% +%% authority = [ userinfo "@" ] host [ ":" port ] +%% +%% +%% [RFC 3986, Chapter 3.2.1. User Information] +%% +%% The userinfo subcomponent may consist of a user name and, optionally, +%% scheme-specific information about how to gain authorization to access +%% the resource. The user information, if present, is followed by a +%% commercial at-sign ("@") that delimits it from the host. +%% +%% userinfo = *( unreserved / pct-encoded / sub-delims / ":" ) +%% +%% +%% [RFC 3986, Chapter 3.2.2. Host] +%% +%% The host subcomponent of authority is identified by an IP literal +%% encapsulated within square brackets, an IPv4 address in dotted- +%% decimal form, or a registered name. +%% +%% host = IP-literal / IPv4address / reg-name +%% +%% IP-literal = "[" ( IPv6address / IPvFuture ) "]" +%% +%% IPvFuture = "v" 1*HEXDIG "." 1*( unreserved / sub-delims / ":" ) +%% +%% IPv6address = 6( h16 ":" ) ls32 +%% / "::" 5( h16 ":" ) ls32 +%% / [ h16 ] "::" 4( h16 ":" ) ls32 +%% / [ *1( h16 ":" ) h16 ] "::" 3( h16 ":" ) ls32 +%% / [ *2( h16 ":" ) h16 ] "::" 2( h16 ":" ) ls32 +%% / [ *3( h16 ":" ) h16 ] "::" h16 ":" ls32 +%% / [ *4( h16 ":" ) h16 ] "::" ls32 +%% / [ *5( h16 ":" ) h16 ] "::" h16 +%% / [ *6( h16 ":" ) h16 ] "::" +%% +%% ls32 = ( h16 ":" h16 ) / IPv4address +%% ; least-significant 32 bits of address +%% +%% h16 = 1*4HEXDIG +%% ; 16 bits of address represented in hexadecimal +%% +%% IPv4address = dec-octet "." dec-octet "." dec-octet "." dec-octet +%% +%% dec-octet = DIGIT ; 0-9 +%% / %x31-39 DIGIT ; 10-99 +%% / "1" 2DIGIT ; 100-199 +%% / "2" %x30-34 DIGIT ; 200-249 +%% / "25" %x30-35 ; 250-255 +%% +%% reg-name = *( unreserved / pct-encoded / sub-delims ) +%% +%% +%% [RFC 3986, Chapter 3.2.2. Port] +%% +%% The port subcomponent of authority is designated by an optional port +%% number in decimal following the host and delimited from it by a +%% single colon (":") character. +%% +%% port = *DIGIT +%% +%% +%% [RFC 3986, Chapter 3.3. Path] +%% +%% The path component contains data, usually organized in hierarchical +%% form, that, along with data in the non-hierarchical query component +%% (Section 3.4), serves to identify a resource within the scope of the +%% URI's scheme and naming authority (if any). The path is terminated +%% by the first question mark ("?") or number sign ("#") character, or +%% by the end of the URI. +%% +%% path = path-abempty ; begins with "/" or is empty +%% / path-absolute ; begins with "/" but not "//" +%% / path-noscheme ; begins with a non-colon segment +%% / path-rootless ; begins with a segment +%% / path-empty ; zero characters +%% +%% path-abempty = *( "/" segment ) +%% path-absolute = "/" [ segment-nz *( "/" segment ) ] +%% path-noscheme = segment-nz-nc *( "/" segment ) +%% path-rootless = segment-nz *( "/" segment ) +%% path-empty = 0<pchar> +%% segment = *pchar +%% segment-nz = 1*pchar +%% segment-nz-nc = 1*( unreserved / pct-encoded / sub-delims / "@" ) +%% ; non-zero-length segment without any colon ":" +%% +%% pchar = unreserved / pct-encoded / sub-delims / ":" / "@" +%% +%% +%% [RFC 3986, Chapter 3.4. Query] +%% +%% The query component contains non-hierarchical data that, along with +%% data in the path component (Section 3.3), serves to identify a +%% resource within the scope of the URI's scheme and naming authority +%% (if any). The query component is indicated by the first question +%% mark ("?") character and terminated by a number sign ("#") character +%% or by the end of the URI. +%% +%% query = *( pchar / "/" / "?" ) +%% +%% +%% [RFC 3986, Chapter 3.5. Fragment] +%% +%% The fragment identifier component of a URI allows indirect +%% identification of a secondary resource by reference to a primary +%% resource and additional identifying information. +%% +%% fragment = *( pchar / "/" / "?" ) +%% +%% +%% [RFC 3986, Chapter 4.1. URI Reference] +%% +%% URI-reference is used to denote the most common usage of a resource +%% identifier. +%% +%% URI-reference = URI / relative-ref +%% +%% +%% [RFC 3986, Chapter 4.2. Relative Reference] +%% +%% A relative reference takes advantage of the hierarchical syntax +%% (Section 1.2.3) to express a URI reference relative to the name space +%% of another hierarchical URI. +%% +%% relative-ref = relative-part [ "?" query ] [ "#" fragment ] +%% +%% relative-part = "//" authority path-abempty +%% / path-absolute +%% / path-noscheme +%% / path-empty +%% +%% +%% [RFC 3986, Chapter 4.3. Absolute URI] +%% +%% Some protocol elements allow only the absolute form of a URI without +%% a fragment identifier. For example, defining a base URI for later +%% use by relative references calls for an absolute-URI syntax rule that +%% does not allow a fragment. +%% +%% absolute-URI = scheme ":" hier-part [ "?" query ] +%% +-module(uri_string). + +%%------------------------------------------------------------------------- +%% External API +%%------------------------------------------------------------------------- +-export([normalize/1, parse/1, + recompose/1, transcode/2]). +-export_type([error/0, uri_map/0, uri_string/0]). + + +%%------------------------------------------------------------------------- +%% Internal API +%%------------------------------------------------------------------------- +-export([is_host/1, is_path/1]). % suppress warnings + + +%%------------------------------------------------------------------------- +%% Macros +%%------------------------------------------------------------------------- +-define(CHAR(Char), <<Char/utf8>>). +-define(STRING_EMPTY, <<>>). +-define(STRING(MatchStr), <<MatchStr/binary>>). +-define(STRING_REST(MatchStr, Rest), <<MatchStr/utf8, Rest/binary>>). + +-define(DEC2HEX(X), + if ((X) >= 0) andalso ((X) =< 9) -> (X) + $0; + ((X) >= 10) andalso ((X) =< 15) -> (X) + $A - 10 + end). + +-define(HEX2DEC(X), + if ((X) >= $0) andalso ((X) =< $9) -> (X) - $0; + ((X) >= $A) andalso ((X) =< $F) -> (X) - $A + 10; + ((X) >= $a) andalso ((X) =< $f) -> (X) - $a + 10 + end). + + +%%%========================================================================= +%%% API +%%%========================================================================= + +%%------------------------------------------------------------------------- +%% URI compliant with RFC 3986 +%% ASCII %x21 - %x7A ("!" - "z") except +%% %x34 " double quote +%% %x60 < less than +%% %x62 > greater than +%% %x92 \ backslash +%% %x94 ^ caret / circumflex +%% %x96 ` grave / accent +%%------------------------------------------------------------------------- +-type uri_string() :: iodata(). +-type error() :: {error, atom(), term()}. + + +%%------------------------------------------------------------------------- +%% RFC 3986, Chapter 3. Syntax Components +%%------------------------------------------------------------------------- +-type uri_map() :: + #{fragment => unicode:chardata(), + host => unicode:chardata(), + path => unicode:chardata(), + port => non_neg_integer() | undefined, + query => unicode:chardata(), + scheme => unicode:chardata(), + userinfo => unicode:chardata()} | #{}. + + +%%------------------------------------------------------------------------- +%% Normalize URIs +%%------------------------------------------------------------------------- +-spec normalize(URIString) -> NormalizedURI when + URIString :: uri_string(), + NormalizedURI :: uri_string(). +normalize(URIString) -> + %% Percent-encoding normalization and case normalization for + %% percent-encoded triplets are achieved by running parse and + %% recompose on the input URI string. + recompose( + normalize_path_segment( + normalize_scheme_based( + normalize_case( + parse(URIString))))). + + +%%------------------------------------------------------------------------- +%% Parse URIs +%%------------------------------------------------------------------------- +-spec parse(URIString) -> URIMap when + URIString :: uri_string(), + URIMap :: uri_map() + | error(). +parse(URIString) when is_binary(URIString) -> + try parse_uri_reference(URIString, #{}) + catch + throw:{error, Atom, RestData} -> {error, Atom, RestData} + end; +parse(URIString) when is_list(URIString) -> + try + Binary = unicode:characters_to_binary(URIString), + Map = parse_uri_reference(Binary, #{}), + convert_mapfields_to_list(Map) + catch + throw:{error, Atom, RestData} -> {error, Atom, RestData} + end. + + +%%------------------------------------------------------------------------- +%% Recompose URIs +%%------------------------------------------------------------------------- +-spec recompose(URIMap) -> URIString when + URIMap :: uri_map(), + URIString :: uri_string() + | error(). +recompose(Map) -> + case is_valid_map(Map) of + false -> + {error, invalid_map, Map}; + true -> + try + T0 = update_scheme(Map, empty), + T1 = update_userinfo(Map, T0), + T2 = update_host(Map, T1), + T3 = update_port(Map, T2), + T4 = update_path(Map, T3), + T5 = update_query(Map, T4), + update_fragment(Map, T5) + catch + throw:{error, Atom, RestData} -> {error, Atom, RestData} + end + end. + + +%%------------------------------------------------------------------------- +%% Transcode URIs +%%------------------------------------------------------------------------- +-spec transcode(URIString, Options) -> Result when + URIString :: uri_string(), + Options :: [{in_encoding, unicode:encoding()}|{out_encoding, unicode:encoding()}], + Result :: uri_string() + | error(). +transcode(URIString, Options) when is_binary(URIString) -> + try + InEnc = proplists:get_value(in_encoding, Options, utf8), + OutEnc = proplists:get_value(out_encoding, Options, utf8), + List = convert_to_list(URIString, InEnc), + Output = transcode(List, [], InEnc, OutEnc), + convert_to_binary(Output, utf8, OutEnc) + catch + throw:{error, Atom, RestData} -> {error, Atom, RestData} + end; +transcode(URIString, Options) when is_list(URIString) -> + InEnc = proplists:get_value(in_encoding, Options, utf8), + OutEnc = proplists:get_value(out_encoding, Options, utf8), + Flattened = flatten_list(URIString, InEnc), + try transcode(Flattened, [], InEnc, OutEnc) + catch + throw:{error, Atom, RestData} -> {error, Atom, RestData} + end. + + +%%%======================================================================== +%%% Internal functions +%%%======================================================================== + +%%------------------------------------------------------------------------- +%% Converts Map fields to lists +%%------------------------------------------------------------------------- +convert_mapfields_to_list(Map) -> + Fun = fun (_, V) when is_binary(V) -> unicode:characters_to_list(V); + (_, V) -> V end, + maps:map(Fun, Map). + + +%%------------------------------------------------------------------------- +%% [RFC 3986, Chapter 4.1. URI Reference] +%% +%% URI-reference is used to denote the most common usage of a resource +%% identifier. +%% +%% URI-reference = URI / relative-ref +%%------------------------------------------------------------------------- +-spec parse_uri_reference(binary(), uri_map()) -> uri_map(). +parse_uri_reference(<<>>, _) -> #{path => <<>>}; +parse_uri_reference(URIString, URI) -> + try parse_scheme_start(URIString, URI) + catch + throw:{_,_,_} -> + parse_relative_part(URIString, URI) + end. + + +%%------------------------------------------------------------------------- +%% [RFC 3986, Chapter 4.2. Relative Reference] +%% +%% A relative reference takes advantage of the hierarchical syntax +%% (Section 1.2.3) to express a URI reference relative to the name space +%% of another hierarchical URI. +%% +%% relative-ref = relative-part [ "?" query ] [ "#" fragment ] +%% +%% relative-part = "//" authority path-abempty +%% / path-absolute +%% / path-noscheme +%% / path-empty +%%------------------------------------------------------------------------- +-spec parse_relative_part(binary(), uri_map()) -> uri_map(). +parse_relative_part(?STRING_REST("//", Rest), URI) -> + %% Parse userinfo - "//" is NOT part of authority + try parse_userinfo(Rest, URI) of + {T, URI1} -> + Userinfo = calculate_parsed_userinfo(Rest, T), + URI2 = maybe_add_path(URI1), + URI2#{userinfo => decode_userinfo(Userinfo)} + catch + throw:{_,_,_} -> + {T, URI1} = parse_host(Rest, URI), + Host = calculate_parsed_host_port(Rest, T), + URI2 = maybe_add_path(URI1), + URI2#{host => decode_host(remove_brackets(Host))} + end; +parse_relative_part(?STRING_REST($/, Rest), URI) -> + {T, URI1} = parse_segment(Rest, URI), % path-absolute + Path = calculate_parsed_part(Rest, T), + URI1#{path => decode_path(?STRING_REST($/, Path))}; +parse_relative_part(?STRING_REST($?, Rest), URI) -> + {T, URI1} = parse_query(Rest, URI), % path-empty ?query + Query = calculate_parsed_query_fragment(Rest, T), + URI2 = maybe_add_path(URI1), + URI2#{query => decode_query(Query)}; +parse_relative_part(?STRING_REST($#, Rest), URI) -> + {T, URI1} = parse_fragment(Rest, URI), % path-empty + Fragment = calculate_parsed_query_fragment(Rest, T), + URI2 = maybe_add_path(URI1), + URI2#{fragment => decode_fragment(Fragment)}; +parse_relative_part(?STRING_REST(Char, Rest), URI) -> + case is_segment_nz_nc(Char) of + true -> + {T, URI1} = parse_segment_nz_nc(Rest, URI), % path-noscheme + Path = calculate_parsed_part(Rest, T), + URI1#{path => decode_path(?STRING_REST(Char, Path))}; + false -> throw({error,invalid_uri,[Char]}) + end. + + +%%------------------------------------------------------------------------- +%% [RFC 3986, Chapter 3.3. Path] +%% +%% The path component contains data, usually organized in hierarchical +%% form, that, along with data in the non-hierarchical query component +%% (Section 3.4), serves to identify a resource within the scope of the +%% URI's scheme and naming authority (if any). The path is terminated +%% by the first question mark ("?") or number sign ("#") character, or +%% by the end of the URI. +%% +%% path = path-abempty ; begins with "/" or is empty +%% / path-absolute ; begins with "/" but not "//" +%% / path-noscheme ; begins with a non-colon segment +%% / path-rootless ; begins with a segment +%% / path-empty ; zero characters +%% +%% path-abempty = *( "/" segment ) +%% path-absolute = "/" [ segment-nz *( "/" segment ) ] +%% path-noscheme = segment-nz-nc *( "/" segment ) +%% path-rootless = segment-nz *( "/" segment ) +%% path-empty = 0<pchar> +%% segment = *pchar +%% segment-nz = 1*pchar +%% segment-nz-nc = 1*( unreserved / pct-encoded / sub-delims / "@" ) +%% ; non-zero-length segment without any colon ":" +%% +%% pchar = unreserved / pct-encoded / sub-delims / ":" / "@" +%%------------------------------------------------------------------------- + +%%------------------------------------------------------------------------- +%% path-abempty +%%------------------------------------------------------------------------- +-spec parse_segment(binary(), uri_map()) -> {binary(), uri_map()}. +parse_segment(?STRING_REST($/, Rest), URI) -> + parse_segment(Rest, URI); % segment +parse_segment(?STRING_REST($?, Rest), URI) -> + {T, URI1} = parse_query(Rest, URI), % ?query + Query = calculate_parsed_query_fragment(Rest, T), + {Rest, URI1#{query => decode_query(Query)}}; +parse_segment(?STRING_REST($#, Rest), URI) -> + {T, URI1} = parse_fragment(Rest, URI), + Fragment = calculate_parsed_query_fragment(Rest, T), + {Rest, URI1#{fragment => decode_fragment(Fragment)}}; +parse_segment(?STRING_REST(Char, Rest), URI) -> + case is_pchar(Char) of + true -> parse_segment(Rest, URI); + false -> throw({error,invalid_uri,[Char]}) + end; +parse_segment(?STRING_EMPTY, URI) -> + {?STRING_EMPTY, URI}. + + +%%------------------------------------------------------------------------- +%% path-noscheme +%%------------------------------------------------------------------------- +-spec parse_segment_nz_nc(binary(), uri_map()) -> {binary(), uri_map()}. +parse_segment_nz_nc(?STRING_REST($/, Rest), URI) -> + parse_segment(Rest, URI); % segment +parse_segment_nz_nc(?STRING_REST($?, Rest), URI) -> + {T, URI1} = parse_query(Rest, URI), % ?query + Query = calculate_parsed_query_fragment(Rest, T), + {Rest, URI1#{query => decode_query(Query)}}; +parse_segment_nz_nc(?STRING_REST($#, Rest), URI) -> + {T, URI1} = parse_fragment(Rest, URI), + Fragment = calculate_parsed_query_fragment(Rest, T), + {Rest, URI1#{fragment => decode_fragment(Fragment)}}; +parse_segment_nz_nc(?STRING_REST(Char, Rest), URI) -> + case is_segment_nz_nc(Char) of + true -> parse_segment_nz_nc(Rest, URI); + false -> throw({error,invalid_uri,[Char]}) + end; +parse_segment_nz_nc(?STRING_EMPTY, URI) -> + {?STRING_EMPTY, URI}. + + +%% Check if char is pchar. +-spec is_pchar(char()) -> boolean(). +is_pchar($%) -> true; % pct-encoded +is_pchar($:) -> true; +is_pchar($@) -> true; +is_pchar(Char) -> is_unreserved(Char) orelse is_sub_delim(Char). + +%% Check if char is segment_nz_nc. +-spec is_segment_nz_nc(char()) -> boolean(). +is_segment_nz_nc($%) -> true; % pct-encoded +is_segment_nz_nc($@) -> true; +is_segment_nz_nc(Char) -> is_unreserved(Char) orelse is_sub_delim(Char). + + +%%------------------------------------------------------------------------- +%% [RFC 3986, Chapter 3.1. Scheme] +%% +%% Each URI begins with a scheme name that refers to a specification for +%% assigning identifiers within that scheme. +%% +%% scheme = ALPHA *( ALPHA / DIGIT / "+" / "-" / "." ) +%%------------------------------------------------------------------------- +-spec parse_scheme_start(binary(), uri_map()) -> uri_map(). +parse_scheme_start(?STRING_REST(Char, Rest), URI) -> + case is_alpha(Char) of + true -> {T, URI1} = parse_scheme(Rest, URI), + Scheme = calculate_parsed_scheme(Rest, T), + URI2 = maybe_add_path(URI1), + URI2#{scheme => ?STRING_REST(Char, Scheme)}; + false -> throw({error,invalid_uri,[Char]}) + end. + +%% Add path component if it missing after parsing the URI. +%% According to the URI specification there is always a +%% path component in every URI-reference and it can be +%% empty. +maybe_add_path(Map) -> + case maps:is_key(path, Map) of + false -> + Map#{path => <<>>}; + _Else -> + Map + end. + + +-spec parse_scheme(binary(), uri_map()) -> {binary(), uri_map()}. +parse_scheme(?STRING_REST($:, Rest), URI) -> + {_, URI1} = parse_hier(Rest, URI), + {Rest, URI1}; +parse_scheme(?STRING_REST(Char, Rest), URI) -> + case is_scheme(Char) of + true -> parse_scheme(Rest, URI); + false -> throw({error,invalid_uri,[Char]}) + end; +parse_scheme(?STRING_EMPTY, _URI) -> + throw({error,invalid_uri,<<>>}). + + +%% Check if char is allowed in scheme +-spec is_scheme(char()) -> boolean(). +is_scheme($+) -> true; +is_scheme($-) -> true; +is_scheme($.) -> true; +is_scheme(Char) -> is_alpha(Char) orelse is_digit(Char). + + +%%------------------------------------------------------------------------- +%% hier-part = "//" authority path-abempty +%% / path-absolute +%% / path-rootless +%% / path-empty +%%------------------------------------------------------------------------- +-spec parse_hier(binary(), uri_map()) -> {binary(), uri_map()}. +parse_hier(?STRING_REST("//", Rest), URI) -> + % Parse userinfo - "//" is NOT part of authority + try parse_userinfo(Rest, URI) of + {T, URI1} -> + Userinfo = calculate_parsed_userinfo(Rest, T), + {Rest, URI1#{userinfo => decode_userinfo(Userinfo)}} + catch + throw:{_,_,_} -> + {T, URI1} = parse_host(Rest, URI), + Host = calculate_parsed_host_port(Rest, T), + {Rest, URI1#{host => decode_host(remove_brackets(Host))}} + end; +parse_hier(?STRING_REST($/, Rest), URI) -> + {T, URI1} = parse_segment(Rest, URI), % path-absolute + Path = calculate_parsed_part(Rest, T), + {Rest, URI1#{path => decode_path(?STRING_REST($/, Path))}}; +parse_hier(?STRING_REST($?, Rest), URI) -> + {T, URI1} = parse_query(Rest, URI), % path-empty ?query + Query = calculate_parsed_query_fragment(Rest, T), + {Rest, URI1#{query => decode_query(Query)}}; +parse_hier(?STRING_REST($#, Rest), URI) -> + {T, URI1} = parse_fragment(Rest, URI), % path-empty + Fragment = calculate_parsed_query_fragment(Rest, T), + {Rest, URI1#{fragment => decode_fragment(Fragment)}}; +parse_hier(?STRING_REST(Char, Rest), URI) -> % path-rootless + case is_pchar(Char) of + true -> % segment_nz + {T, URI1} = parse_segment(Rest, URI), + Path = calculate_parsed_part(Rest, T), + {Rest, URI1#{path => decode_path(?STRING_REST(Char, Path))}}; + false -> throw({error,invalid_uri,[Char]}) + end; +parse_hier(?STRING_EMPTY, URI) -> + {<<>>, URI}. + + +%%------------------------------------------------------------------------- +%% [RFC 3986, Chapter 3.2. Authority] +%% +%% Many URI schemes include a hierarchical element for a naming +%% authority so that governance of the name space defined by the +%% remainder of the URI is delegated to that authority (which may, in +%% turn, delegate it further). +%% +%% The authority component is preceded by a double slash ("//") and is +%% terminated by the next slash ("/"), question mark ("?"), or number +%% sign ("#") character, or by the end of the URI. +%% +%% authority = [ userinfo "@" ] host [ ":" port ] +%% +%% +%% [RFC 3986, Chapter 3.2.1. User Information] +%% +%% The userinfo subcomponent may consist of a user name and, optionally, +%% scheme-specific information about how to gain authorization to access +%% the resource. The user information, if present, is followed by a +%% commercial at-sign ("@") that delimits it from the host. +%% +%% userinfo = *( unreserved / pct-encoded / sub-delims / ":" ) +%%------------------------------------------------------------------------- +-spec parse_userinfo(binary(), uri_map()) -> {binary(), uri_map()}. +parse_userinfo(?CHAR($@), URI) -> + {?STRING_EMPTY, URI#{host => <<>>}}; +parse_userinfo(?STRING_REST($@, Rest), URI) -> + {T, URI1} = parse_host(Rest, URI), + Host = calculate_parsed_host_port(Rest, T), + {Rest, URI1#{host => decode_host(remove_brackets(Host))}}; +parse_userinfo(?STRING_REST(Char, Rest), URI) -> + case is_userinfo(Char) of + true -> parse_userinfo(Rest, URI); + false -> throw({error,invalid_uri,[Char]}) + end; +parse_userinfo(?STRING_EMPTY, _URI) -> + %% URI cannot end in userinfo state + throw({error,invalid_uri,<<>>}). + + +%% Check if char is allowed in userinfo +-spec is_userinfo(char()) -> boolean(). +is_userinfo($%) -> true; % pct-encoded +is_userinfo($:) -> true; +is_userinfo(Char) -> is_unreserved(Char) orelse is_sub_delim(Char). + + +%%------------------------------------------------------------------------- +%% [RFC 3986, Chapter 3.2.2. Host] +%% +%% The host subcomponent of authority is identified by an IP literal +%% encapsulated within square brackets, an IPv4 address in dotted- +%% decimal form, or a registered name. +%% +%% host = IP-literal / IPv4address / reg-name +%% +%% IP-literal = "[" ( IPv6address / IPvFuture ) "]" +%% +%% IPvFuture = "v" 1*HEXDIG "." 1*( unreserved / sub-delims / ":" ) +%% +%% IPv6address = 6( h16 ":" ) ls32 +%% / "::" 5( h16 ":" ) ls32 +%% / [ h16 ] "::" 4( h16 ":" ) ls32 +%% / [ *1( h16 ":" ) h16 ] "::" 3( h16 ":" ) ls32 +%% / [ *2( h16 ":" ) h16 ] "::" 2( h16 ":" ) ls32 +%% / [ *3( h16 ":" ) h16 ] "::" h16 ":" ls32 +%% / [ *4( h16 ":" ) h16 ] "::" ls32 +%% / [ *5( h16 ":" ) h16 ] "::" h16 +%% / [ *6( h16 ":" ) h16 ] "::" +%% +%% ls32 = ( h16 ":" h16 ) / IPv4address +%% ; least-significant 32 bits of address +%% +%% h16 = 1*4HEXDIG +%% ; 16 bits of address represented in hexadecimal +%% +%% IPv4address = dec-octet "." dec-octet "." dec-octet "." dec-octet +%% +%% dec-octet = DIGIT ; 0-9 +%% / %x31-39 DIGIT ; 10-99 +%% / "1" 2DIGIT ; 100-199 +%% / "2" %x30-34 DIGIT ; 200-249 +%% / "25" %x30-35 ; 250-255 +%% +%% reg-name = *( unreserved / pct-encoded / sub-delims ) +%%------------------------------------------------------------------------- +-spec parse_host(binary(), uri_map()) -> {binary(), uri_map()}. +parse_host(?STRING_REST($:, Rest), URI) -> + {T, URI1} = parse_port(Rest, URI), + H = calculate_parsed_host_port(Rest, T), + Port = get_port(H), + {Rest, URI1#{port => Port}}; +parse_host(?STRING_REST($/, Rest), URI) -> + {T, URI1} = parse_segment(Rest, URI), % path-abempty + Path = calculate_parsed_part(Rest, T), + {Rest, URI1#{path => decode_path(?STRING_REST($/, Path))}}; +parse_host(?STRING_REST($?, Rest), URI) -> + {T, URI1} = parse_query(Rest, URI), % path-empty ?query + Query = calculate_parsed_query_fragment(Rest, T), + {Rest, URI1#{query => decode_query(Query)}}; +parse_host(?STRING_REST($[, Rest), URI) -> + parse_ipv6_bin(Rest, [], URI); +parse_host(?STRING_REST($#, Rest), URI) -> + {T, URI1} = parse_fragment(Rest, URI), % path-empty + Fragment = calculate_parsed_query_fragment(Rest, T), + {Rest, URI1#{fragment => decode_fragment(Fragment)}}; +parse_host(?STRING_REST(Char, Rest), URI) -> + case is_digit(Char) of + true -> parse_ipv4_bin(Rest, [Char], URI); + false -> parse_reg_name(?STRING_REST(Char, Rest), URI) + end; +parse_host(?STRING_EMPTY, URI) -> + {?STRING_EMPTY, URI}. + + +-spec parse_reg_name(binary(), uri_map()) -> {binary(), uri_map()}. +parse_reg_name(?STRING_REST($:, Rest), URI) -> + {T, URI1} = parse_port(Rest, URI), + H = calculate_parsed_host_port(Rest, T), + Port = get_port(H), + {Rest, URI1#{port => Port}}; +parse_reg_name(?STRING_REST($/, Rest), URI) -> + {T, URI1} = parse_segment(Rest, URI), % path-abempty + Path = calculate_parsed_part(Rest, T), + {Rest, URI1#{path => decode_path(?STRING_REST($/, Path))}}; +parse_reg_name(?STRING_REST($?, Rest), URI) -> + {T, URI1} = parse_query(Rest, URI), % path-empty ?query + Query = calculate_parsed_query_fragment(Rest, T), + {Rest, URI1#{query => decode_query(Query)}}; +parse_reg_name(?STRING_REST($#, Rest), URI) -> + {T, URI1} = parse_fragment(Rest, URI), % path-empty + Fragment = calculate_parsed_query_fragment(Rest, T), + {Rest, URI1#{fragment => decode_fragment(Fragment)}}; +parse_reg_name(?STRING_REST(Char, Rest), URI) -> + case is_reg_name(Char) of + true -> parse_reg_name(Rest, URI); + false -> throw({error,invalid_uri,[Char]}) + end; +parse_reg_name(?STRING_EMPTY, URI) -> + {?STRING_EMPTY, URI}. + +%% Check if char is allowed in reg-name +-spec is_reg_name(char()) -> boolean(). +is_reg_name($%) -> true; +is_reg_name(Char) -> is_unreserved(Char) orelse is_sub_delim(Char). + + +-spec parse_ipv4_bin(binary(), list(), uri_map()) -> {binary(), uri_map()}. +parse_ipv4_bin(?STRING_REST($:, Rest), Acc, URI) -> + _ = validate_ipv4_address(lists:reverse(Acc)), + {T, URI1} = parse_port(Rest, URI), + H = calculate_parsed_host_port(Rest, T), + Port = get_port(H), + {Rest, URI1#{port => Port}}; +parse_ipv4_bin(?STRING_REST($/, Rest), Acc, URI) -> + _ = validate_ipv4_address(lists:reverse(Acc)), + {T, URI1} = parse_segment(Rest, URI), % path-abempty + Path = calculate_parsed_part(Rest, T), + {Rest, URI1#{path => decode_path(?STRING_REST($/, Path))}}; +parse_ipv4_bin(?STRING_REST($?, Rest), Acc, URI) -> + _ = validate_ipv4_address(lists:reverse(Acc)), + {T, URI1} = parse_query(Rest, URI), % path-empty ?query + Query = calculate_parsed_query_fragment(Rest, T), + {Rest, URI1#{query => decode_query(Query)}}; +parse_ipv4_bin(?STRING_REST($#, Rest), Acc, URI) -> + _ = validate_ipv4_address(lists:reverse(Acc)), + {T, URI1} = parse_fragment(Rest, URI), % path-empty + Fragment = calculate_parsed_query_fragment(Rest, T), + {Rest, URI1#{fragment => decode_fragment(Fragment)}}; +parse_ipv4_bin(?STRING_REST(Char, Rest), Acc, URI) -> + case is_ipv4(Char) of + true -> parse_ipv4_bin(Rest, [Char|Acc], URI); + false -> throw({error,invalid_uri,[Char]}) + end; +parse_ipv4_bin(?STRING_EMPTY, Acc, URI) -> + _ = validate_ipv4_address(lists:reverse(Acc)), + {?STRING_EMPTY, URI}. + + +%% Check if char is allowed in IPv4 addresses +-spec is_ipv4(char()) -> boolean(). +is_ipv4($.) -> true; +is_ipv4(Char) -> is_digit(Char). + +-spec validate_ipv4_address(list()) -> list(). +validate_ipv4_address(Addr) -> + case inet:parse_ipv4strict_address(Addr) of + {ok, _} -> Addr; + {error, _} -> throw({error,invalid_uri,Addr}) + end. + + +-spec parse_ipv6_bin(binary(), list(), uri_map()) -> {binary(), uri_map()}. +parse_ipv6_bin(?STRING_REST($], Rest), Acc, URI) -> + _ = validate_ipv6_address(lists:reverse(Acc)), + parse_ipv6_bin_end(Rest, URI); +parse_ipv6_bin(?STRING_REST(Char, Rest), Acc, URI) -> + case is_ipv6(Char) of + true -> parse_ipv6_bin(Rest, [Char|Acc], URI); + false -> throw({error,invalid_uri,[Char]}) + end; +parse_ipv6_bin(?STRING_EMPTY, _Acc, _URI) -> + throw({error,invalid_uri,<<>>}). + +%% Check if char is allowed in IPv6 addresses +-spec is_ipv6(char()) -> boolean(). +is_ipv6($:) -> true; +is_ipv6($.) -> true; +is_ipv6(Char) -> is_hex_digit(Char). + + +-spec parse_ipv6_bin_end(binary(), uri_map()) -> {binary(), uri_map()}. +parse_ipv6_bin_end(?STRING_REST($:, Rest), URI) -> + {T, URI1} = parse_port(Rest, URI), + H = calculate_parsed_host_port(Rest, T), + Port = get_port(H), + {Rest, URI1#{port => Port}}; +parse_ipv6_bin_end(?STRING_REST($/, Rest), URI) -> + {T, URI1} = parse_segment(Rest, URI), % path-abempty + Path = calculate_parsed_part(Rest, T), + {Rest, URI1#{path => decode_path(?STRING_REST($/, Path))}}; +parse_ipv6_bin_end(?STRING_REST($?, Rest), URI) -> + {T, URI1} = parse_query(Rest, URI), % path-empty ?query + Query = calculate_parsed_query_fragment(Rest, T), + {Rest, URI1#{query => decode_query(Query)}}; +parse_ipv6_bin_end(?STRING_REST($#, Rest), URI) -> + {T, URI1} = parse_fragment(Rest, URI), % path-empty + Fragment = calculate_parsed_query_fragment(Rest, T), + {Rest, URI1#{fragment => decode_fragment(Fragment)}}; +parse_ipv6_bin_end(?STRING_REST(Char, Rest), URI) -> + case is_ipv6(Char) of + true -> parse_ipv6_bin_end(Rest, URI); + false -> throw({error,invalid_uri,[Char]}) + end; +parse_ipv6_bin_end(?STRING_EMPTY, URI) -> + {?STRING_EMPTY, URI}. + +-spec validate_ipv6_address(list()) -> list(). +validate_ipv6_address(Addr) -> + case inet:parse_ipv6strict_address(Addr) of + {ok, _} -> Addr; + {error, _} -> throw({error,invalid_uri,Addr}) + end. + + +%%------------------------------------------------------------------------- +%% [RFC 3986, Chapter 3.2.2. Port] +%% +%% The port subcomponent of authority is designated by an optional port +%% number in decimal following the host and delimited from it by a +%% single colon (":") character. +%% +%% port = *DIGIT +%%------------------------------------------------------------------------- +-spec parse_port(binary(), uri_map()) -> {binary(), uri_map()}. +parse_port(?STRING_REST($/, Rest), URI) -> + {T, URI1} = parse_segment(Rest, URI), % path-abempty + Path = calculate_parsed_part(Rest, T), + {Rest, URI1#{path => decode_path(?STRING_REST($/, Path))}}; +parse_port(?STRING_REST($?, Rest), URI) -> + {T, URI1} = parse_query(Rest, URI), % path-empty ?query + Query = calculate_parsed_query_fragment(Rest, T), + {Rest, URI1#{query => decode_query(Query)}}; +parse_port(?STRING_REST($#, Rest), URI) -> + {T, URI1} = parse_fragment(Rest, URI), % path-empty + Fragment = calculate_parsed_query_fragment(Rest, T), + {Rest, URI1#{fragment => decode_fragment(Fragment)}}; +parse_port(?STRING_REST(Char, Rest), URI) -> + case is_digit(Char) of + true -> parse_port(Rest, URI); + false -> throw({error,invalid_uri,[Char]}) + end; +parse_port(?STRING_EMPTY, URI) -> + {?STRING_EMPTY, URI}. + + +%%------------------------------------------------------------------------- +%% [RFC 3986, Chapter 3.4. Query] +%% +%% The query component contains non-hierarchical data that, along with +%% data in the path component (Section 3.3), serves to identify a +%% resource within the scope of the URI's scheme and naming authority +%% (if any). The query component is indicated by the first question +%% mark ("?") character and terminated by a number sign ("#") character +%% or by the end of the URI. +%% +%% query = *( pchar / "/" / "?" ) +%%------------------------------------------------------------------------- +-spec parse_query(binary(), uri_map()) -> {binary(), uri_map()}. +parse_query(?STRING_REST($#, Rest), URI) -> + {T, URI1} = parse_fragment(Rest, URI), + Fragment = calculate_parsed_query_fragment(Rest, T), + {Rest, URI1#{fragment => decode_fragment(Fragment)}}; +parse_query(?STRING_REST(Char, Rest), URI) -> + case is_query(Char) of + true -> parse_query(Rest, URI); + false -> throw({error,invalid_uri,[Char]}) + end; +parse_query(?STRING_EMPTY, URI) -> + {?STRING_EMPTY, URI}. + + +%% Check if char is allowed in query +-spec is_query(char()) -> boolean(). +is_query($/) -> true; +is_query($?) -> true; +is_query(Char) -> is_pchar(Char). + + +%%------------------------------------------------------------------------- +%% [RFC 3986, Chapter 3.5. Fragment] +%% +%% The fragment identifier component of a URI allows indirect +%% identification of a secondary resource by reference to a primary +%% resource and additional identifying information. +%% +%% fragment = *( pchar / "/" / "?" ) +%%------------------------------------------------------------------------- +-spec parse_fragment(binary(), uri_map()) -> {binary(), uri_map()}. +parse_fragment(?STRING_REST(Char, Rest), URI) -> + case is_fragment(Char) of + true -> parse_fragment(Rest, URI); + false -> throw({error,invalid_uri,[Char]}) + end; +parse_fragment(?STRING_EMPTY, URI) -> + {?STRING_EMPTY, URI}. + + +%% Check if char is allowed in fragment +-spec is_fragment(char()) -> boolean(). +is_fragment($/) -> true; +is_fragment($?) -> true; +is_fragment(Char) -> is_pchar(Char). + + +%%------------------------------------------------------------------------- +%% [RFC 3986, Chapter 2.2. Reserved Characters] +%% +%% reserved = gen-delims / sub-delims +%% +%% gen-delims = ":" / "/" / "?" / "#" / "[" / "]" / "@" +%% +%% sub-delims = "!" / "$" / "&" / "'" / "(" / ")" +%% / "*" / "+" / "," / ";" / "=" +%% +%%------------------------------------------------------------------------- + +%% Check if char is sub-delim. +-spec is_sub_delim(char()) -> boolean(). +is_sub_delim($!) -> true; +is_sub_delim($$) -> true; +is_sub_delim($&) -> true; +is_sub_delim($') -> true; +is_sub_delim($() -> true; +is_sub_delim($)) -> true; + +is_sub_delim($*) -> true; +is_sub_delim($+) -> true; +is_sub_delim($,) -> true; +is_sub_delim($;) -> true; +is_sub_delim($=) -> true; +is_sub_delim(_) -> false. + + +%%------------------------------------------------------------------------- +%% [RFC 3986, Chapter 2.3. Unreserved Characters] +%% +%% unreserved = ALPHA / DIGIT / "-" / "." / "_" / "~" +%% +%%------------------------------------------------------------------------- +-spec is_unreserved(char()) -> boolean(). +is_unreserved($-) -> true; +is_unreserved($.) -> true; +is_unreserved($_) -> true; +is_unreserved($~) -> true; +is_unreserved(Char) -> is_alpha(Char) orelse is_digit(Char). + +-spec is_alpha(char()) -> boolean(). +is_alpha(C) + when $A =< C, C =< $Z; + $a =< C, C =< $z -> true; +is_alpha(_) -> false. + +-spec is_digit(char()) -> boolean(). +is_digit(C) + when $0 =< C, C =< $9 -> true; +is_digit(_) -> false. + +-spec is_hex_digit(char()) -> boolean(). +is_hex_digit(C) + when $0 =< C, C =< $9;$a =< C, C =< $f;$A =< C, C =< $F -> true; +is_hex_digit(_) -> false. + + +%% Remove enclosing brackets from binary +-spec remove_brackets(binary()) -> binary(). +remove_brackets(<<$[/utf8, Rest/binary>>) -> + {H,T} = split_binary(Rest, byte_size(Rest) - 1), + case T =:= <<$]/utf8>> of + true -> H; + false -> Rest + end; +remove_brackets(Addr) -> Addr. + + +%%------------------------------------------------------------------------- +%% Helper functions for calculating the parsed binary. +%%------------------------------------------------------------------------- +-spec calculate_parsed_scheme(binary(), binary()) -> binary(). +calculate_parsed_scheme(Input, <<>>) -> + strip_last_char(Input, [$:]); +calculate_parsed_scheme(Input, Unparsed) -> + get_parsed_binary(Input, Unparsed). + + +-spec calculate_parsed_part(binary(), binary()) -> binary(). +calculate_parsed_part(Input, <<>>) -> + strip_last_char(Input, [$?,$#]); +calculate_parsed_part(Input, Unparsed) -> + get_parsed_binary(Input, Unparsed). + + +-spec calculate_parsed_userinfo(binary(), binary()) -> binary(). +calculate_parsed_userinfo(Input, <<>>) -> + strip_last_char(Input, [$?,$#,$@]); +calculate_parsed_userinfo(Input, Unparsed) -> + get_parsed_binary(Input, Unparsed). + + +-spec calculate_parsed_host_port(binary(), binary()) -> binary(). +calculate_parsed_host_port(Input, <<>>) -> + strip_last_char(Input, [$:,$?,$#,$/]); +calculate_parsed_host_port(Input, Unparsed) -> + get_parsed_binary(Input, Unparsed). + + +calculate_parsed_query_fragment(Input, <<>>) -> + strip_last_char(Input, [$#]); +calculate_parsed_query_fragment(Input, Unparsed) -> + get_parsed_binary(Input, Unparsed). + + +get_port(<<>>) -> + undefined; +get_port(B) -> + try binary_to_integer(B) + catch + error:badarg -> + throw({error, invalid_uri, B}) + end. + + +%% Strip last char if it is in list +%% +%% This function is optimized for speed: parse/1 is about 10% faster than +%% with an alternative implementation based on lists and sets. +strip_last_char(<<>>, _) -> <<>>; +strip_last_char(Input, [C0]) -> + case binary:last(Input) of + C0 -> + init_binary(Input); + _Else -> + Input + end; +strip_last_char(Input, [C0,C1]) -> + case binary:last(Input) of + C0 -> + init_binary(Input); + C1 -> + init_binary(Input); + _Else -> + Input + end; +strip_last_char(Input, [C0,C1,C2]) -> + case binary:last(Input) of + C0 -> + init_binary(Input); + C1 -> + init_binary(Input); + C2 -> + init_binary(Input); + _Else -> + Input + end; +strip_last_char(Input, [C0,C1,C2,C3]) -> + case binary:last(Input) of + C0 -> + init_binary(Input); + C1 -> + init_binary(Input); + C2 -> + init_binary(Input); + C3 -> + init_binary(Input); + _Else -> + Input + end. + + +%% Get parsed binary +get_parsed_binary(Input, Unparsed) -> + {First, _} = split_binary(Input, byte_size(Input) - byte_size_exl_head(Unparsed)), + First. + + +%% Return all bytes of the binary except the last one. The binary must be non-empty. +init_binary(B) -> + {Init, _} = + split_binary(B, byte_size(B) - 1), + Init. + + +%% Returns the size of a binary exluding the first element. +%% Used in calls to split_binary(). +-spec byte_size_exl_head(binary()) -> number(). +byte_size_exl_head(<<>>) -> 0; +byte_size_exl_head(Binary) -> byte_size(Binary) + 1. + + +%%------------------------------------------------------------------------- +%% [RFC 3986, Chapter 2.1. Percent-Encoding] +%% +%% A percent-encoding mechanism is used to represent a data octet in a +%% component when that octet's corresponding character is outside the +%% allowed set or is being used as a delimiter of, or within, the +%% component. A percent-encoded octet is encoded as a character +%% triplet, consisting of the percent character "%" followed by the two +%% hexadecimal digits representing that octet's numeric value. For +%% example, "%20" is the percent-encoding for the binary octet +%% "00100000" (ABNF: %x20), which in US-ASCII corresponds to the space +%% character (SP). Section 2.4 describes when percent-encoding and +%% decoding is applied. +%% +%% pct-encoded = "%" HEXDIG HEXDIG +%%------------------------------------------------------------------------- +-spec decode_userinfo(binary()) -> binary(). +decode_userinfo(Cs) -> + check_utf8(decode(Cs, fun is_userinfo/1, <<>>)). + +-spec decode_host(binary()) -> binary(). +decode_host(Cs) -> + check_utf8(decode(Cs, fun is_host/1, <<>>)). + +-spec decode_path(binary()) -> binary(). +decode_path(Cs) -> + check_utf8(decode(Cs, fun is_path/1, <<>>)). + +-spec decode_query(binary()) -> binary(). +decode_query(Cs) -> + check_utf8(decode(Cs, fun is_query/1, <<>>)). + +-spec decode_fragment(binary()) -> binary(). +decode_fragment(Cs) -> + check_utf8(decode(Cs, fun is_fragment/1, <<>>)). + + +%% Returns Cs if it is utf8 encoded. +check_utf8(Cs) -> + case unicode:characters_to_list(Cs) of + {incomplete,_,_} -> + throw({error,invalid_utf8,Cs}); + {error,_,_} -> + throw({error,invalid_utf8,Cs}); + _ -> Cs + end. + +%%------------------------------------------------------------------------- +%% Percent-encode +%%------------------------------------------------------------------------- + +%% Only validates as scheme cannot have percent-encoded characters +-spec encode_scheme(list()|binary()) -> list() | binary(). +encode_scheme([]) -> + throw({error,invalid_scheme,""}); +encode_scheme(<<>>) -> + throw({error,invalid_scheme,<<>>}); +encode_scheme(Scheme) -> + case validate_scheme(Scheme) of + true -> Scheme; + false -> throw({error,invalid_scheme,Scheme}) + end. + +-spec encode_userinfo(list()|binary()) -> list() | binary(). +encode_userinfo(Cs) -> + encode(Cs, fun is_userinfo/1). + +-spec encode_host(list()|binary()) -> list() | binary(). +encode_host(Cs) -> + case classify_host(Cs) of + regname -> Cs; + ipv4 -> Cs; + ipv6 -> bracket_ipv6(Cs); + other -> encode(Cs, fun is_reg_name/1) + end. + +-spec encode_path(list()|binary()) -> list() | binary(). +encode_path(Cs) -> + encode(Cs, fun is_path/1). + +-spec encode_query(list()|binary()) -> list() | binary(). +encode_query(Cs) -> + encode(Cs, fun is_query/1). + +-spec encode_fragment(list()|binary()) -> list() | binary(). +encode_fragment(Cs) -> + encode(Cs, fun is_fragment/1). + +%%------------------------------------------------------------------------- +%% Helper funtions for percent-decode +%%------------------------------------------------------------------------- +decode(<<$%,C0,C1,Cs/binary>>, Fun, Acc) -> + case is_hex_digit(C0) andalso is_hex_digit(C1) of + true -> + B = ?HEX2DEC(C0)*16+?HEX2DEC(C1), + decode(Cs, Fun, <<Acc/binary, B>>); + false -> throw({error,invalid_percent_encoding,<<$%,C0,C1>>}) + end; +decode(<<C,Cs/binary>>, Fun, Acc) -> + case Fun(C) of + true -> decode(Cs, Fun, <<Acc/binary, C>>); + false -> throw({error,invalid_percent_encoding,<<C,Cs/binary>>}) + end; +decode(<<>>, _Fun, Acc) -> + Acc. + +%% Check if char is allowed in host +-spec is_host(char()) -> boolean(). +is_host($:) -> true; +is_host(Char) -> is_unreserved(Char) orelse is_sub_delim(Char). + +%% Check if char is allowed in path +-spec is_path(char()) -> boolean(). +is_path($/) -> true; +is_path(Char) -> is_pchar(Char). + + +%%------------------------------------------------------------------------- +%% Helper functions for percent-encode +%%------------------------------------------------------------------------- +-spec encode(list()|binary(), fun()) -> list() | binary(). +encode(Component, Fun) when is_list(Component) -> + B = unicode:characters_to_binary(Component), + unicode:characters_to_list(encode(B, Fun, <<>>)); +encode(Component, Fun) when is_binary(Component) -> + encode(Component, Fun, <<>>). +%% +encode(<<Char/utf8, Rest/binary>>, Fun, Acc) -> + C = encode_codepoint_binary(Char, Fun), + encode(Rest, Fun, <<Acc/binary,C/binary>>); +encode(<<Char, Rest/binary>>, _Fun, _Acc) -> + throw({error,invalid_input,<<Char,Rest/binary>>}); +encode(<<>>, _Fun, Acc) -> + Acc. + + +-spec encode_codepoint_binary(integer(), fun()) -> binary(). +encode_codepoint_binary(C, Fun) -> + case Fun(C) of + false -> percent_encode_binary(C); + true -> <<C>> + end. + + +-spec percent_encode_binary(integer()) -> binary(). +percent_encode_binary(Code) -> + percent_encode_binary(<<Code/utf8>>, <<>>). + + +percent_encode_binary(<<A:4,B:4,Rest/binary>>, Acc) -> + percent_encode_binary(Rest, <<Acc/binary,$%,(?DEC2HEX(A)),(?DEC2HEX(B))>>); +percent_encode_binary(<<>>, Acc) -> + Acc. + + +%%------------------------------------------------------------------------- +%%------------------------------------------------------------------------- +validate_scheme([]) -> true; +validate_scheme([H|T]) -> + case is_scheme(H) of + true -> validate_scheme(T); + false -> false + end; +validate_scheme(<<>>) -> true; +validate_scheme(<<H, Rest/binary>>) -> + case is_scheme(H) of + true -> validate_scheme(Rest); + false -> false + end. + + +%%------------------------------------------------------------------------- +%% Classifies hostname into the following categories: +%% regname, ipv4 - address does not contain reserved characters to be +%% percent-encoded +%% ipv6 - address does not contain reserved characters but it shall be +%% encolsed in brackets +%% other - address shall be percent-encoded +%%------------------------------------------------------------------------- +classify_host([]) -> other; +classify_host(Addr) when is_binary(Addr) -> + A = unicode:characters_to_list(Addr), + classify_host_ipv6(A); +classify_host(Addr) -> + classify_host_ipv6(Addr). + +classify_host_ipv6(Addr) -> + case is_ipv6_address(Addr) of + true -> ipv6; + false -> classify_host_ipv4(Addr) + end. + +classify_host_ipv4(Addr) -> + case is_ipv4_address(Addr) of + true -> ipv4; + false -> classify_host_regname(Addr) + end. + +classify_host_regname([]) -> regname; +classify_host_regname([H|T]) -> + case is_reg_name(H) of + true -> classify_host_regname(T); + false -> other + end. + +is_ipv4_address(Addr) -> + case inet:parse_ipv4strict_address(Addr) of + {ok, _} -> true; + {error, _} -> false + end. + +is_ipv6_address(Addr) -> + case inet:parse_ipv6strict_address(Addr) of + {ok, _} -> true; + {error, _} -> false + end. + +bracket_ipv6(Addr) when is_binary(Addr) -> + concat(<<$[,Addr/binary>>,<<$]>>); +bracket_ipv6(Addr) when is_list(Addr) -> + [$[|Addr] ++ "]". + + +%%------------------------------------------------------------------------- +%% Helper funtions for recompose +%%------------------------------------------------------------------------- + +%%------------------------------------------------------------------------- +%% Checks if input Map has valid combination of fields that can be +%% recomposed into a URI. +%% +%% The implementation is based on a decision tree that fulfills the +%% following rules: +%% - 'path' shall always be present in the input map +%% URI = scheme ":" hier-part [ "?" query ] [ "#" fragment ] +%% hier-part = "//" authority path-abempty +%% / path-absolute +%% / path-rootless +%% / path-empty +%% - 'host' shall be present in the input map when 'path' starts with +%% two slashes ("//") +%% path = path-abempty ; begins with "/" or is empty +%% / path-absolute ; begins with "/" but not "//" +%% / path-noscheme ; begins with a non-colon segment +%% / path-rootless ; begins with a segment +%% / path-empty ; zero characters +%% path-abempty = *( "/" segment ) +%% segment = *pchar +%% - 'host' shall be present if userinfo or port is present in input map +%% authority = [ userinfo "@" ] host [ ":" port ] +%% - All fields shall be valid (scheme, userinfo, host, port, path, query +%% or fragment). +%%------------------------------------------------------------------------- +is_valid_map(#{path := Path} = Map) -> + ((starts_with_two_slash(Path) andalso is_valid_map_host(Map)) + orelse + (maps:is_key(userinfo, Map) andalso is_valid_map_host(Map)) + orelse + (maps:is_key(port, Map) andalso is_valid_map_host(Map)) + orelse + all_fields_valid(Map)); +is_valid_map(#{}) -> + false. + + +is_valid_map_host(Map) -> + maps:is_key(host, Map) andalso all_fields_valid(Map). + + +all_fields_valid(Map) -> + Fun = fun(scheme, _, Acc) -> Acc; + (userinfo, _, Acc) -> Acc; + (host, _, Acc) -> Acc; + (port, _, Acc) -> Acc; + (path, _, Acc) -> Acc; + (query, _, Acc) -> Acc; + (fragment, _, Acc) -> Acc; + (_, _, _) -> false + end, + maps:fold(Fun, true, Map). + + +starts_with_two_slash([$/,$/|_]) -> + true; +starts_with_two_slash(?STRING_REST("//", _)) -> + true; +starts_with_two_slash(_) -> false. + + +update_scheme(#{scheme := Scheme}, _) -> + add_colon_postfix(encode_scheme(Scheme)); +update_scheme(#{}, _) -> + empty. + + +update_userinfo(#{userinfo := Userinfo}, empty) -> + add_auth_prefix(encode_userinfo(Userinfo)); +update_userinfo(#{userinfo := Userinfo}, URI) -> + concat(URI,add_auth_prefix(encode_userinfo(Userinfo))); +update_userinfo(#{}, empty) -> + empty; +update_userinfo(#{}, URI) -> + URI. + + +update_host(#{host := Host}, empty) -> + add_auth_prefix(encode_host(Host)); +update_host(#{host := Host} = Map, URI) -> + concat(URI,add_host_prefix(Map, encode_host(Host))); +update_host(#{}, empty) -> + empty; +update_host(#{}, URI) -> + URI. + + +%% URI cannot be empty for ports. E.g. ":8080" is not a valid URI +update_port(#{port := undefined}, URI) -> + concat(URI, <<":">>); +update_port(#{port := Port}, URI) -> + concat(URI,add_colon(encode_port(Port))); +update_port(#{}, URI) -> + URI. + + +update_path(#{path := Path}, empty) -> + encode_path(Path); +update_path(#{path := Path}, URI) -> + concat(URI,encode_path(Path)); +update_path(#{}, empty) -> + empty; +update_path(#{}, URI) -> + URI. + + +update_query(#{query := Query}, empty) -> + encode_query(Query); +update_query(#{query := Query}, URI) -> + concat(URI,add_question_mark(encode_query(Query))); +update_query(#{}, empty) -> + empty; +update_query(#{}, URI) -> + URI. + + +update_fragment(#{fragment := Fragment}, empty) -> + add_hashmark(encode_fragment(Fragment)); +update_fragment(#{fragment := Fragment}, URI) -> + concat(URI,add_hashmark(encode_fragment(Fragment))); +update_fragment(#{}, empty) -> + ""; +update_fragment(#{}, URI) -> + URI. + +%%------------------------------------------------------------------------- +%% Concatenates its arguments that can be lists and binaries. +%% The result is a list if at least one of its argument is a list and +%% binary otherwise. +%%------------------------------------------------------------------------- +concat(A, B) when is_binary(A), is_binary(B) -> + <<A/binary, B/binary>>; +concat(A, B) when is_binary(A), is_list(B) -> + unicode:characters_to_list(A) ++ B; +concat(A, B) when is_list(A) -> + A ++ maybe_to_list(B). + +add_hashmark(Comp) when is_binary(Comp) -> + <<$#, Comp/binary>>; +add_hashmark(Comp) when is_list(Comp) -> + [$#|Comp]. + +add_question_mark(Comp) when is_binary(Comp) -> + <<$?, Comp/binary>>; +add_question_mark(Comp) when is_list(Comp) -> + [$?|Comp]. + +add_colon(Comp) when is_binary(Comp) -> + <<$:, Comp/binary>>. + +add_colon_postfix(Comp) when is_binary(Comp) -> + <<Comp/binary,$:>>; +add_colon_postfix(Comp) when is_list(Comp) -> + Comp ++ ":". + +add_auth_prefix(Comp) when is_binary(Comp) -> + <<"//", Comp/binary>>; +add_auth_prefix(Comp) when is_list(Comp) -> + [$/,$/|Comp]. + +add_host_prefix(#{userinfo := _}, Host) when is_binary(Host) -> + <<$@,Host/binary>>; +add_host_prefix(#{}, Host) when is_binary(Host) -> + <<"//",Host/binary>>; +add_host_prefix(#{userinfo := _}, Host) when is_list(Host) -> + [$@|Host]; +add_host_prefix(#{}, Host) when is_list(Host) -> + [$/,$/|Host]. + +maybe_to_list(Comp) when is_binary(Comp) -> unicode:characters_to_list(Comp); +maybe_to_list(Comp) -> Comp. + +encode_port(Port) -> + integer_to_binary(Port). + +%%------------------------------------------------------------------------- +%% Helper functions for transcode +%%------------------------------------------------------------------------- + +%%------------------------------------------------------------------------- +%% uri_string:transcode(<<"x%00%00%00%F6"/utf32>>). +%% 1. Convert (transcode/2) input to list form (list of unicode codepoints) +%% "x%00%00%00%F6" +%% 2. Accumulate characters until percent-encoded segment (transcode/4). +%% Acc = "x" +%% 3. Convert percent-encoded triplets to binary form (transcode_pct/4) +%% <<0,0,0,246>> +%% 4. Transcode in-encoded binary to out-encoding (utf32 -> utf8): +%% <<195,182>> +%% 5. Percent-encode out-encoded binary: +%% <<"%C3%B6"/utf8>> = <<37,67,51,37,66,54>> +%% 6. Convert binary to list form, reverse it and append the accumulator +%% "6B%3C%" + "x" +%% 7. Reverse Acc and return it +%%------------------------------------------------------------------------- +transcode([$%,_C0,_C1|_Rest] = L, Acc, InEnc, OutEnc) -> + transcode_pct(L, Acc, <<>>, InEnc, OutEnc); +transcode([_C|_Rest] = L, Acc, InEnc, OutEnc) -> + transcode(L, Acc, [], InEnc, OutEnc). +%% +transcode([$%,_C0,_C1|_Rest] = L, Acc, List, InEncoding, OutEncoding) -> + transcode_pct(L, List ++ Acc, <<>>, InEncoding, OutEncoding); +transcode([C|Rest], Acc, List, InEncoding, OutEncoding) -> + transcode(Rest, Acc, [C|List], InEncoding, OutEncoding); +transcode([], Acc, List, _InEncoding, _OutEncoding) -> + lists:reverse(List ++ Acc). + + +%% Transcode percent-encoded segment +transcode_pct([$%,C0,C1|Rest] = L, Acc, B, InEncoding, OutEncoding) -> + case is_hex_digit(C0) andalso is_hex_digit(C1) of + true -> + Int = ?HEX2DEC(C0)*16+?HEX2DEC(C1), + transcode_pct(Rest, Acc, <<B/binary, Int>>, InEncoding, OutEncoding); + false -> throw({error, invalid_percent_encoding,L}) + end; +transcode_pct([_C|_Rest] = L, Acc, B, InEncoding, OutEncoding) -> + OutBinary = convert_to_binary(B, InEncoding, OutEncoding), + PctEncUtf8 = percent_encode_segment(OutBinary), + Out = lists:reverse(convert_to_list(PctEncUtf8, utf8)), + transcode(L, Out ++ Acc, [], InEncoding, OutEncoding); +transcode_pct([], Acc, B, InEncoding, OutEncoding) -> + OutBinary = convert_to_binary(B, InEncoding, OutEncoding), + PctEncUtf8 = percent_encode_segment(OutBinary), + Out = convert_to_list(PctEncUtf8, utf8), + lists:reverse(Acc) ++ Out. + + +%% Convert to binary +convert_to_binary(Binary, InEncoding, OutEncoding) -> + case unicode:characters_to_binary(Binary, InEncoding, OutEncoding) of + {error, _List, RestData} -> + throw({error, invalid_input, RestData}); + {incomplete, _List, RestData} -> + throw({error, invalid_input, RestData}); + Result -> + Result + end. + + +%% Convert to list +convert_to_list(Binary, InEncoding) -> + case unicode:characters_to_list(Binary, InEncoding) of + {error, _List, RestData} -> + throw({error, invalid_input, RestData}); + {incomplete, _List, RestData} -> + throw({error, invalid_input, RestData}); + Result -> + Result + end. + + +%% Flatten input list +flatten_list([], _) -> + []; +flatten_list(L, InEnc) -> + flatten_list(L, InEnc, []). +%% +flatten_list([H|T], InEnc, Acc) when is_binary(H) -> + L = convert_to_list(H, InEnc), + flatten_list(T, InEnc, lists:reverse(L) ++ Acc); +flatten_list([H|T], InEnc, Acc) when is_list(H) -> + flatten_list(H ++ T, InEnc, Acc); +flatten_list([H|T], InEnc, Acc) -> + flatten_list(T, InEnc, [H|Acc]); +flatten_list([], _InEnc, Acc) -> + lists:reverse(Acc); +flatten_list(Arg, _, _) -> + throw({error, invalid_input, Arg}). + + +percent_encode_segment(Segment) -> + percent_encode_binary(Segment, <<>>). + + +%%------------------------------------------------------------------------- +%% Helper functions for normalize +%%------------------------------------------------------------------------- + +%% 6.2.2.1. Case Normalization +normalize_case(#{scheme := Scheme, host := Host} = Map) -> + Map#{scheme => to_lower(Scheme), + host => to_lower(Host)}; +normalize_case(#{host := Host} = Map) -> + Map#{host => to_lower(Host)}; +normalize_case(#{scheme := Scheme} = Map) -> + Map#{scheme => to_lower(Scheme)}; +normalize_case(#{} = Map) -> + Map. + + +to_lower(Cs) when is_list(Cs) -> + B = convert_to_binary(Cs, utf8, utf8), + convert_to_list(to_lower(B), utf8); +to_lower(Cs) when is_binary(Cs) -> + to_lower(Cs, <<>>). +%% +to_lower(<<C,Cs/binary>>, Acc) when $A =< C, C =< $Z -> + to_lower(Cs, <<Acc/binary,(C + 32)>>); +to_lower(<<C,Cs/binary>>, Acc) -> + to_lower(Cs, <<Acc/binary,C>>); +to_lower(<<>>, Acc) -> + Acc. + + +%% 6.2.2.3. Path Segment Normalization +%% 5.2.4. Remove Dot Segments +normalize_path_segment(Map) -> + Path = maps:get(path, Map, undefined), + Map#{path => remove_dot_segments(Path)}. + + +remove_dot_segments(Path) when is_binary(Path) -> + remove_dot_segments(Path, <<>>); +remove_dot_segments(Path) when is_list(Path) -> + B = convert_to_binary(Path, utf8, utf8), + B1 = remove_dot_segments(B, <<>>), + convert_to_list(B1, utf8). +%% +remove_dot_segments(<<>>, Output) -> + Output; +remove_dot_segments(<<"../",T/binary>>, Output) -> + remove_dot_segments(T, Output); +remove_dot_segments(<<"./",T/binary>>, Output) -> + remove_dot_segments(T, Output); +remove_dot_segments(<<"/./",T/binary>>, Output) -> + remove_dot_segments(<<$/,T/binary>>, Output); +remove_dot_segments(<<"/.">>, Output) -> + remove_dot_segments(<<$/>>, Output); +remove_dot_segments(<<"/../",T/binary>>, Output) -> + Out1 = remove_last_segment(Output), + remove_dot_segments(<<$/,T/binary>>, Out1); +remove_dot_segments(<<"/..">>, Output) -> + Out1 = remove_last_segment(Output), + remove_dot_segments(<<$/>>, Out1); +remove_dot_segments(<<$.>>, Output) -> + remove_dot_segments(<<>>, Output); +remove_dot_segments(<<"..">>, Output) -> + remove_dot_segments(<<>>, Output); +remove_dot_segments(Input, Output) -> + {First, Rest} = first_path_segment(Input), + remove_dot_segments(Rest, <<Output/binary,First/binary>>). + + +first_path_segment(Input) -> + F = first_path_segment(Input, <<>>), + split_binary(Input, byte_size(F)). +%% +first_path_segment(<<$/,T/binary>>, Acc) -> + first_path_segment_end(<<T/binary>>, <<Acc/binary,$/>>); +first_path_segment(<<C,T/binary>>, Acc) -> + first_path_segment_end(<<T/binary>>, <<Acc/binary,C>>). + + +first_path_segment_end(<<>>, Acc) -> + Acc; +first_path_segment_end(<<$/,_/binary>>, Acc) -> + Acc; +first_path_segment_end(<<C,T/binary>>, Acc) -> + first_path_segment_end(<<T/binary>>, <<Acc/binary,C>>). + + +remove_last_segment(<<>>) -> + <<>>; +remove_last_segment(B) -> + {Init, Last} = split_binary(B, byte_size(B) - 1), + case Last of + <<$/>> -> + Init; + _Char -> + remove_last_segment(Init) + end. + + +%% RFC 3986, 6.2.3. Scheme-Based Normalization +normalize_scheme_based(Map) -> + Scheme = maps:get(scheme, Map, undefined), + Port = maps:get(port, Map, undefined), + Path= maps:get(path, Map, undefined), + normalize_scheme_based(Map, Scheme, Port, Path). +%% +normalize_scheme_based(Map, Scheme, Port, Path) + when Scheme =:= "http"; Scheme =:= <<"http">> -> + normalize_http(Map, Port, Path); +normalize_scheme_based(Map, Scheme, Port, Path) + when Scheme =:= "https"; Scheme =:= <<"https">> -> + normalize_https(Map, Port, Path); +normalize_scheme_based(Map, Scheme, Port, _Path) + when Scheme =:= "ftp"; Scheme =:= <<"ftp">> -> + normalize_ftp(Map, Port); +normalize_scheme_based(Map, Scheme, Port, _Path) + when Scheme =:= "ssh"; Scheme =:= <<"ssh">> -> + normalize_ssh_sftp(Map, Port); +normalize_scheme_based(Map, Scheme, Port, _Path) + when Scheme =:= "sftp"; Scheme =:= <<"sftp">> -> + normalize_ssh_sftp(Map, Port); +normalize_scheme_based(Map, Scheme, Port, _Path) + when Scheme =:= "tftp"; Scheme =:= <<"tftp">> -> + normalize_tftp(Map, Port); +normalize_scheme_based(Map, _, _, _) -> + Map. + + +normalize_http(Map, Port, Path) -> + M1 = normalize_port(Map, Port, 80), + normalize_http_path(M1, Path). + + +normalize_https(Map, Port, Path) -> + M1 = normalize_port(Map, Port, 443), + normalize_http_path(M1, Path). + + +normalize_ftp(Map, Port) -> + normalize_port(Map, Port, 21). + + +normalize_ssh_sftp(Map, Port) -> + normalize_port(Map, Port, 22). + + +normalize_tftp(Map, Port) -> + normalize_port(Map, Port, 69). + + +normalize_port(Map, Port, Default) -> + case Port of + Default -> + maps:remove(port, Map); + _Else -> + Map + end. + + +normalize_http_path(Map, Path) -> + case Path of + "" -> + Map#{path => "/"}; + <<>> -> + Map#{path => <<"/">>}; + _Else -> + Map + end. diff --git a/lib/stdlib/test/Makefile b/lib/stdlib/test/Makefile index 523cb95065..8490770f3d 100644 --- a/lib/stdlib/test/Makefile +++ b/lib/stdlib/test/Makefile @@ -87,6 +87,7 @@ MODULES= \ timer_simple_SUITE \ unicode_SUITE \ unicode_util_SUITE \ + uri_string_SUITE \ win32reg_SUITE \ y2k_SUITE \ select_SUITE \ diff --git a/lib/stdlib/test/ets_SUITE.erl b/lib/stdlib/test/ets_SUITE.erl index 5a5e282998..f329a07b7a 100644 --- a/lib/stdlib/test/ets_SUITE.erl +++ b/lib/stdlib/test/ets_SUITE.erl @@ -1682,7 +1682,7 @@ do_random_test() -> ets:delete(Set), verify_etsmem(EtsMem). -%% Ttest various variants of update_element. +%% Test various variants of update_element. update_element(Config) when is_list(Config) -> EtsMem = etsmem(), repeat_for_opts(fun update_element_opts/1), diff --git a/lib/stdlib/test/filename_SUITE.erl b/lib/stdlib/test/filename_SUITE.erl index fc77593bb8..f284eb1ed6 100644 --- a/lib/stdlib/test/filename_SUITE.erl +++ b/lib/stdlib/test/filename_SUITE.erl @@ -107,6 +107,17 @@ absname(Config) when is_list(Config) -> [Drive|":/erlang/src"] = filename:absname([Drive|":erlang/src"]), "a:/erlang" = filename:absname("a:erlang"), + "//foo" = filename:absname("//foo"), + "//foo/bar" = filename:absname("//foo/bar"), + "//foo/\bar" = filename:absname("//foo/\bar"), + "//foo/bar/baz" = filename:absname("//foo/bar\\baz"), + "//foo/bar/baz" = filename:absname("//foo\\bar/baz"), + "//foo" = filename:absname("\\\\foo"), + "//foo/bar" = filename:absname("\\\\foo/bar"), + "//foo/\bar" = filename:absname("\\\\foo/\bar"), + "//foo/bar/baz" = filename:absname("\\\\foo/bar\\baz"), + "//foo/bar/baz" = filename:absname("\\\\foo\\bar/baz"), + file:set_cwd(Cwd), ok; {unix, _} -> @@ -167,6 +178,23 @@ absname_2(Config) when is_list(Config) -> [Drive|":/"]), "a:/erlang" = filename:absname("a:erlang", [Drive|":/"]), + "//foo" = filename:absname("foo","//"), + "//foo/bar" = filename:absname("foo/bar", "//"), + "//foo/bar" = filename:absname("bar", "//foo"), + "//bar" = filename:absname("/bar", "//foo"), + "//foo/bar/baz" = filename:absname("bar/baz", "//foo"), + "//bar/baz" = filename:absname("//bar/baz", "//foo"), + "//\bar" = filename:absname("/\bar", "//foo"), + "//foo" = filename:absname("foo","\\\\"), + "//foo/bar" = filename:absname("foo/bar", "\\\\"), + "//foo/bar" = filename:absname("bar", "\\\\foo"), + "//bar" = filename:absname("/bar", "\\\\foo"), + "//foo/bar/baz" = filename:absname("bar/baz", "\\\\foo"), + "//bar/baz" = filename:absname("\\\\bar/baz", "\\\\foo"), + "//\bar" = filename:absname("/\bar", "\\\\foo"), + "//bar/baz" = filename:absname("\\\\bar/baz", "//foo"), + "//bar/baz" = filename:absname("//bar/baz", "\\\\foo"), + ok; _ -> "/usr/foo" = filename:absname(foo, "/usr"), @@ -244,6 +272,18 @@ dirname(Config) when is_list(Config) -> "A:usr" = filename:dirname("A:usr/foo.erl"), "/usr" = filename:dirname("\\usr\\foo.erl"), "/" = filename:dirname("\\usr"), + "//foo/bar" = filename:dirname("//foo/bar/baz.erl"), + "//foo/\bar" = filename:dirname("//foo/\bar/baz.erl"), + "//foo/bar" = filename:dirname("//foo\\bar/baz.erl"), + "//foo/bar" = filename:dirname("\\\\foo/bar/baz.erl"), + "//foo/\bar" = filename:dirname("\\\\foo/\bar/baz.erl"), + "//foo/bar" = filename:dirname("\\\\foo\\bar/baz.erl"), + "//foo" = filename:dirname("//foo/baz.erl"), + "//foo" = filename:dirname("//foo/\baz.erl"), + "//foo" = filename:dirname("//foo\\baz.erl"), + "//foo" = filename:dirname("\\\\foo/baz.erl"), + "//foo" = filename:dirname("\\\\foo/\baz.erl"), + "//foo" = filename:dirname("\\\\foo\\baz.erl"), "A:" = filename:dirname("A:"); _ -> true end, @@ -289,7 +329,6 @@ join(Config) when is_list(Config) -> %% join/1 and join/2 (OTP-12158) by using help function %% filename_join/2. "/" = filename:join(["/"]), - "/" = filename:join(["//"]), "usr/foo.erl" = filename_join("usr","foo.erl"), "/src/foo.erl" = filename_join(usr, "/src/foo.erl"), "/src/foo.erl" = filename_join("/src/",'foo.erl'), @@ -301,7 +340,6 @@ join(Config) when is_list(Config) -> "a/b/c/d/e/f/g" = filename_join("a//b/c/", "d//e/f/g"), "a/b/c/d/e/f/g" = filename_join("a//b/c", "d//e/f/g"), "/d/e/f/g" = filename_join("a//b/c", "/d//e/f/g"), - "/d/e/f/g" = filename:join("a//b/c", "//d//e/f/g"), "foo/bar" = filename_join([$f,$o,$o,$/,[]], "bar"), @@ -332,6 +370,7 @@ join(Config) when is_list(Config) -> case os:type() of {win32, _} -> + "//" = filename:join(["//"]), "d:/" = filename:join(["D:/"]), "d:/" = filename:join(["D:\\"]), "d:/abc" = filename_join("D:/", "abc"), @@ -345,8 +384,35 @@ join(Config) when is_list(Config) -> "c:/usr/foo.erl" = filename:join(["A:","C:/usr","foo.erl"]), "c:usr/foo.erl" = filename:join(["A:","C:usr","foo.erl"]), "d:/foo" = filename:join([$D, $:, $/, []], "foo"), + "//" = filename:join("\\\\", ""), + "//foo" = filename:join("\\\\", "foo"), + "//foo/bar" = filename:join("\\\\", "foo\\\\bar"), + "//foo/bar/baz" = filename:join("\\\\foo", "bar\\\\baz"), + "//foo/bar/baz" = filename:join("\\\\foo", "bar\\baz"), + "//foo/bar/baz" = filename:join("\\\\foo\\bar", baz), + "//foo/\bar/baz" = filename:join("\\\\foo/\bar", baz), + "//foo/bar/baz" = filename:join("\\\\foo/bar", baz), + "//bar/baz" = filename:join("\\\\foo", "\\\\bar\\baz"), + "//bar/baz" = filename:join("\\\\foo", "//bar\\baz"), + "//bar/baz" = filename:join("\\\\foo", "//bar/baz"), + "//bar/baz" = filename:join("\\\\foo", "\\\\bar/baz"), + "//d/e/f/g" = filename:join("a//b/c", "//d//e/f/g"), + "//" = filename:join("//", ""), + "//foo" = filename:join("//", "foo"), + "//foo/bar" = filename:join("//", "foo\\\\bar"), + "//foo/bar/baz" = filename:join("//foo", "bar\\\\baz"), + "//foo/bar/baz" = filename:join("//foo", "bar\\baz"), + "//foo/bar/baz" = filename:join("//foo\\bar", baz), + "//foo/\bar/baz" = filename:join("//foo/\bar", baz), + "//foo/bar/baz" = filename:join("//foo/bar", baz), + "//bar/baz" = filename:join("//foo", "\\\\bar\\baz"), + "//bar/baz" = filename:join("//foo", "//bar\\baz"), + "//bar/baz" = filename:join("//foo", "//bar/baz"), + "//bar/baz" = filename:join("//foo", "\\\\bar/baz"), ok; _ -> + "/" = filename:join(["//"]), + "/d/e/f/g" = filename:join("a//b/c", "//d//e/f/g"), ok end. @@ -402,6 +468,16 @@ split(Config) when is_list(Config) -> filename:split("a:\\msdev\\include"), ["a:","msdev","include"] = filename:split("a:msdev\\include"), + ["//","foo"] = + filename:split("\\\\foo"), + ["//","foo"] = + filename:split("//foo"), + ["//","foo","bar"] = + filename:split("\\\\foo\\\\bar"), + ["//","foo","baz"] = + filename:split("\\\\foo\\baz"), + ["//","foo","baz"] = + filename:split("//foo\\baz"), ok; _ -> ok @@ -630,7 +706,6 @@ extension_bin(Config) when is_list(Config) -> join_bin(Config) when is_list(Config) -> <<"/">> = filename:join([<<"/">>]), - <<"/">> = filename:join([<<"//">>]), <<"usr/foo.erl">> = filename:join(<<"usr">>,<<"foo.erl">>), <<"/src/foo.erl">> = filename:join(usr, <<"/src/foo.erl">>), <<"/src/foo.erl">> = filename:join([<<"/src/">>,'foo.erl']), @@ -642,7 +717,6 @@ join_bin(Config) when is_list(Config) -> <<"a/b/c/d/e/f/g">> = filename:join([<<"a//b/c/">>, <<"d//e/f/g">>]), <<"a/b/c/d/e/f/g">> = filename:join([<<"a//b/c">>, <<"d//e/f/g">>]), <<"/d/e/f/g">> = filename:join([<<"a//b/c">>, <<"/d//e/f/g">>]), - <<"/d/e/f/g">> = filename:join([<<"a//b/c">>, <<"//d//e/f/g">>]), <<"foo/bar">> = filename:join([$f,$o,$o,$/,[]], <<"bar">>), @@ -695,6 +769,7 @@ join_bin(Config) when is_list(Config) -> case os:type() of {win32, _} -> + <<"//">> = filename:join([<<"//">>]), <<"d:/">> = filename:join([<<"D:/">>]), <<"d:/">> = filename:join([<<"D:\\">>]), <<"d:/abc">> = filename:join([<<"D:/">>, <<"abc">>]), @@ -708,8 +783,35 @@ join_bin(Config) when is_list(Config) -> <<"c:/usr/foo.erl">> = filename:join([<<"A:">>,<<"C:/usr">>,<<"foo.erl">>]), <<"c:usr/foo.erl">> = filename:join([<<"A:">>,<<"C:usr">>,<<"foo.erl">>]), <<"d:/foo">> = filename:join([$D, $:, $/, []], <<"foo">>), + <<"//">> = filename:join(<<"\\\\">>, <<"">>), + <<"//foo">> = filename:join(<<"\\\\">>, <<"foo">>), + <<"//foo/bar">> = filename:join(<<"\\\\">>, <<"foo\\\\bar">>), + <<"//foo/bar/baz">> = filename:join(<<"\\\\foo">>, <<"bar\\\\baz">>), + <<"//bar/baz">> = filename:join(<<"\\\\foo">>, <<"\\\\bar\\baz">>), + <<"//foo/bar/baz">> = filename:join(<<"\\\\foo\\bar">>, baz), + <<"//foo/\bar/baz">> = filename:join(<<"\\\\foo/\bar">>, baz), + <<"//foo/bar/baz">> = filename:join(<<"\\\\foo/bar">>, baz), + <<"//bar/baz">> = filename:join(<<"\\\\foo">>, <<"\\\\bar\\baz">>), + <<"//bar/baz">> = filename:join(<<"\\\\foo">>, <<"//bar\\baz">>), + <<"//bar/baz">> = filename:join(<<"\\\\foo">>, <<"//bar/baz">>), + <<"//bar/baz">> = filename:join(<<"\\\\foo">>, <<"\\\\bar/baz">>), + <<"//d/e/f/g">> = filename:join([<<"a//b/c">>, <<"//d//e/f/g">>]), + <<"//">> = filename:join(<<"//">>, <<"">>), + <<"//foo">> = filename:join(<<"//">>, <<"foo">>), + <<"//foo/bar">> = filename:join(<<"//">>, <<"foo\\\\bar">>), + <<"//foo/bar/baz">> = filename:join(<<"//foo">>, <<"bar\\\\baz">>), + <<"//bar/baz">> = filename:join(<<"//foo">>, <<"\\\\bar\\baz">>), + <<"//foo/bar/baz">> = filename:join(<<"//foo\\bar">>, baz), + <<"//foo/\bar/baz">> = filename:join(<<"//foo/\bar">>, baz), + <<"//foo/bar/baz">> = filename:join(<<"//foo/bar">>, baz), + <<"//bar/baz">> = filename:join(<<"//foo">>, <<"\\\\bar\\baz">>), + <<"//bar/baz">> = filename:join(<<"//foo">>, <<"//bar\\baz">>), + <<"//bar/baz">> = filename:join(<<"//foo">>, <<"//bar/baz">>), + <<"//bar/baz">> = filename:join(<<"//foo">>, <<"\\\\bar/baz">>), ok; _ -> + <<"/">> = filename:join([<<"//">>]), + <<"/d/e/f/g">> = filename:join([<<"a//b/c">>, <<"//d//e/f/g">>]), ok end. @@ -756,6 +858,16 @@ split_bin(Config) when is_list(Config) -> filename:split(<<"a:\\msdev\\include">>), [<<"a:">>,<<"msdev">>,<<"include">>] = filename:split(<<"a:msdev\\include">>), + [<<"//">>,<<"foo">>] = + filename:split(<<"\\\\foo">>), + [<<"//">>,<<"foo">>] = + filename:split(<<"//foo">>), + [<<"//">>,<<"foo">>,<<"bar">>] = + filename:split(<<"\\\\foo\\\\bar">>), + [<<"//">>,<<"foo">>,<<"baz">>] = + filename:split(<<"\\\\foo\\baz">>), + [<<"//">>,<<"foo">>,<<"baz">>] = + filename:split(<<"//foo\\baz">>), ok; _ -> ok diff --git a/lib/stdlib/test/property_test/README b/lib/stdlib/test/property_test/README new file mode 100644 index 0000000000..57602bf719 --- /dev/null +++ b/lib/stdlib/test/property_test/README @@ -0,0 +1,12 @@ + +%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% +%%% %%% +%%% WARNING %%% +%%% %%% +%%% This is experimental code which may be changed or removed %%% +%%% anytime without any warning. %%% +%%% %%% +%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% + +The test in this directory are written assuming that the user has a QuickCheck license. They are to be run manually. Some may be possible to be run with other tools, e.g. PropEr. + diff --git a/lib/stdlib/test/property_test/uri_string_recompose.erl b/lib/stdlib/test/property_test/uri_string_recompose.erl new file mode 100644 index 0000000000..e51a671172 --- /dev/null +++ b/lib/stdlib/test/property_test/uri_string_recompose.erl @@ -0,0 +1,361 @@ +%% +%% %CopyrightBegin% +%% +%% Copyright Ericsson AB 2008-2017. All Rights Reserved. +%% +%% Licensed under the Apache License, Version 2.0 (the "License"); +%% you may not use this file except in compliance with the License. +%% You may obtain a copy of the License at +%% +%% http://www.apache.org/licenses/LICENSE-2.0 +%% +%% Unless required by applicable law or agreed to in writing, software +%% distributed under the License is distributed on an "AS IS" BASIS, +%% WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +%% See the License for the specific language governing permissions and +%% limitations under the License. +%% +%% %CopyrightEnd% +%% +-module(uri_string_recompose). + +-compile(export_all). + +-proptest(eqc). +-proptest([triq,proper]). + +-ifndef(EQC). +-ifndef(PROPER). +-ifndef(TRIQ). +-define(EQC,true). +-endif. +-endif. +-endif. + +-ifdef(EQC). +-include_lib("eqc/include/eqc.hrl"). +-define(MOD_eqc,eqc). + +-else. +-ifdef(PROPER). +-include_lib("proper/include/proper.hrl"). +-define(MOD_eqc,proper). + +-else. +-ifdef(TRIQ). +-define(MOD_eqc,triq). +-include_lib("triq/include/triq.hrl"). + +-endif. +-endif. +-endif. + + +-define(STRING_REST(MatchStr, Rest), <<MatchStr/utf8, Rest/binary>>). + +-define(SCHEME, {scheme, scheme()}). +-define(USER, {userinfo, unicode()}). +-define(HOST, {host, host_map()}). +-define(PORT, {port, port()}). +-define(PATH_ABE, {path, path_abempty_map()}). +-define(PATH_ABS, {path, path_absolute_map()}). +-define(PATH_NOS, {path, path_noscheme_map()}). +-define(PATH_ROO, {path, path_rootless_map()}). +-define(PATH_EMP, {path, path_empty_map()}). +-define(QUERY, {query, query_map()}). +-define(FRAGMENT, {fragment, fragment_map()}). + + +%%%======================================================================== +%%% Properties +%%%======================================================================== + +prop_recompose() -> + ?FORALL(Map, map(), + Map =:= uri_string:parse(uri_string:recompose(Map)) + ). + +%% Stats +prop_map_key_length_collect() -> + ?FORALL(List, map(), + collect(length(maps:keys(List)), true)). + +prop_map_collect() -> + ?FORALL(List, map(), + collect(lists:sort(maps:keys(List)), true)). + +prop_scheme_collect() -> + ?FORALL(List, scheme(), + collect(length(List), true)). + + +%%%======================================================================== +%%% Generators +%%%======================================================================== + +map() -> + ?LET(Gen, comp_proplist(), proplist_to_map(Gen)). + +comp_proplist() -> + frequency([ + {2, [?SCHEME,?PATH_ABS]}, + {2, [?SCHEME,?PATH_ROO]}, + {2, [?SCHEME,?PATH_EMP]}, + {2, [?SCHEME,?HOST,?PATH_ABE]}, + {2, [?SCHEME,?USER,?HOST,?PATH_ABE]}, + {2, [?SCHEME,?HOST,?PORT,?PATH_ABE]}, + {2, [?SCHEME,?USER,?HOST,?PORT,?PATH_ABE]}, + + {2, [?PATH_ABS]}, + {2, [?PATH_NOS]}, + {2, [?PATH_EMP]}, + {2, [?HOST,?PATH_ABE]}, + {2, [?USER,?HOST,?PATH_ABE]}, + {2, [?HOST,?PORT,?PATH_ABE]}, + {2, [?USER,?HOST,?PORT,?PATH_ABE]}, + + + {2, [?SCHEME,?PATH_ABS,?QUERY]}, + {2, [?SCHEME,?PATH_ROO,?QUERY]}, + {2, [?SCHEME,?PATH_EMP,?QUERY]}, + {2, [?SCHEME,?HOST,?PATH_ABE,?QUERY]}, + {2, [?SCHEME,?USER,?HOST,?PATH_ABE,?QUERY]}, + {2, [?SCHEME,?HOST,?PORT,?PATH_ABE,?QUERY]}, + {2, [?SCHEME,?USER,?HOST,?PORT,?PATH_ABE,?QUERY]}, + + {2, [?PATH_ABS,?QUERY]}, + {2, [?PATH_NOS,?QUERY]}, + {2, [?PATH_EMP,?QUERY]}, + {2, [?HOST,?PATH_ABE,?QUERY]}, + {2, [?USER,?HOST,?PATH_ABE,?QUERY]}, + {2, [?HOST,?PORT,?PATH_ABE,?QUERY]}, + {2, [?USER,?HOST,?PORT,?PATH_ABE,?QUERY]}, + + + {2, [?SCHEME,?PATH_ABS,?FRAGMENT]}, + {2, [?SCHEME,?PATH_ROO,?FRAGMENT]}, + {2, [?SCHEME,?PATH_EMP,?FRAGMENT]}, + {2, [?SCHEME,?HOST,?PATH_ABE,?FRAGMENT]}, + {2, [?SCHEME,?USER,?HOST,?PATH_ABE,?FRAGMENT]}, + {2, [?SCHEME,?HOST,?PORT,?PATH_ABE,?FRAGMENT]}, + {2, [?SCHEME,?USER,?HOST,?PORT,?PATH_ABE,?FRAGMENT]}, + + {2, [?PATH_ABS,?FRAGMENT]}, + {2, [?PATH_NOS,?FRAGMENT]}, + {2, [?PATH_EMP,?FRAGMENT]}, + {2, [?HOST,?PATH_ABE,?FRAGMENT]}, + {2, [?USER,?HOST,?PATH_ABE,?FRAGMENT]}, + {2, [?HOST,?PORT,?PATH_ABE,?FRAGMENT]}, + {2, [?USER,?HOST,?PORT,?PATH_ABE,?FRAGMENT]}, + + + {2, [?SCHEME,?PATH_ABS,?QUERY,?FRAGMENT]}, + {2, [?SCHEME,?PATH_ROO,?QUERY,?FRAGMENT]}, + {2, [?SCHEME,?PATH_EMP,?QUERY,?FRAGMENT]}, + {2, [?SCHEME,?HOST,?PATH_ABE,?QUERY,?FRAGMENT]}, + {2, [?SCHEME,?USER,?HOST,?PATH_ABE,?QUERY,?FRAGMENT]}, + {2, [?SCHEME,?HOST,?PORT,?PATH_ABE,?QUERY,?FRAGMENT]}, + {2, [?SCHEME,?USER,?HOST,?PORT,?PATH_ABE,?QUERY,?FRAGMENT]}, + + {2, [?PATH_ABS,?QUERY,?FRAGMENT]}, + {2, [?PATH_NOS,?QUERY,?FRAGMENT]}, + {2, [?PATH_EMP,?QUERY,?FRAGMENT]}, + {2, [?HOST,?PATH_ABE,?QUERY,?FRAGMENT]}, + {2, [?USER,?HOST,?PATH_ABE,?QUERY,?FRAGMENT]}, + {2, [?HOST,?PORT,?PATH_ABE,?QUERY,?FRAGMENT]}, + {2, [?USER,?HOST,?PORT,?PATH_ABE,?QUERY,?FRAGMENT]} + ]). + + +%%------------------------------------------------------------------------- +%% Path +%%------------------------------------------------------------------------- +path_abempty_map() -> + frequency([{90, path_abe_map()}, + {10, path_empty_map()}]). + +path_abe_map() -> + ?SIZED(Length, path_abe_map(Length, [])). +%% +path_abe_map(0, Segments) -> + ?LET(Gen, Segments, lists:append(Gen)); +path_abe_map(N, Segments) -> + path_abe_map(N-1, [slash(),segment()|Segments]). + + +path_absolute_map() -> + ?SIZED(Length, path_absolute_map(Length, [])). +%% +path_absolute_map(0, Segments) -> + ?LET(Gen, [slash(),segment_nz()|Segments], lists:append(Gen)); +path_absolute_map(N, Segments) -> + path_absolute_map(N-1, [slash(),segment()|Segments]). + + +path_noscheme_map() -> + ?SIZED(Length, path_noscheme_map(Length, [])). +%% +path_noscheme_map(0, Segments) -> + ?LET(Gen, [segment_nz_nc()|Segments], lists:append(Gen)); +path_noscheme_map(N, Segments) -> + path_noscheme_map(N-1, [slash(),segment()|Segments]). + +path_rootless_map() -> + ?SIZED(Length, path_rootless_map(Length, [])). +%% +path_rootless_map(0, Segments) -> + ?LET(Gen, [segment_nz()|Segments], lists:append(Gen)); +path_rootless_map(N, Segments) -> + path_rootless_map(N-1, [slash(),segment()|Segments]). + + +segment_nz() -> + non_empty(segment()). + +segment_nz_nc() -> + non_empty(list(frequency([{30, unreserved()}, + {10, sub_delims()}, + {10, unicode_char()}, + {5, oneof([$@])} + ]))). + + +segment() -> + list(frequency([{30, unreserved()}, + {10, sub_delims()}, + {10, unicode_char()}, + {5, oneof([$:, $@])} + ])). + +slash() -> + "/". + +path_empty_map() -> + "". + + +%%------------------------------------------------------------------------- +%% Path +%%------------------------------------------------------------------------- +host_map() -> + frequency([{30, reg_name()}, + {30, ip_address()} + ]). + + +reg_name() -> + list(frequency([{30, alpha()}, + {10, sub_delims()}, + {10, unicode_char()} + ])). + +ip_address() -> + oneof(["127.0.0.1", "::127.0.0.1", + "2001:0db8:0000:0000:0000:0000:1428:07ab", + "2001:0db8:0000:0000:0000::1428:07ab", + "2001:0db8:0:0:0:0:1428:07ab", + "2001:0db8:0::0:1428:07ab"]). + +%% Generating only reg-names +host_uri() -> + non_empty(list(frequency([{30, unreserved()}, + {10, sub_delims()}, + {10, pct_encoded()} + ]))). + +%%------------------------------------------------------------------------- +%% Port, Query, Fragment +%%------------------------------------------------------------------------- +port() -> + frequency([{10, undefined}, + {10, range(1,65535)} + ]). + +query_map() -> + unicode(). + + +query_uri() -> + [$?| non_empty(list(frequency([{20, pchar()}, + {5, oneof([$/, $?])} % punctuation + ])))]. + +fragment_map() -> + unicode(). + +fragment_uri() -> + [$?| non_empty(list(frequency([{20, pchar()}, + {5, oneof([$/, $?])} % punctuation + ])))]. + + +%%------------------------------------------------------------------------- +%% Scheme +%%------------------------------------------------------------------------- +scheme() -> + ?SIZED(Length, scheme_start(Length, [])). +%% +scheme_start(0, L) -> + ?LET(Gen, L, lists:reverse(Gen)); +scheme_start(N, L) -> + scheme(N-1,[alpha()|L]). + +scheme(0, L) -> + ?LET(Gen, L, lists:reverse(Gen)); +scheme(N, L) -> + scheme(N-1, [scheme_char()|L]). + + +%%------------------------------------------------------------------------- +%% Misc +%%------------------------------------------------------------------------- +unicode() -> + list(frequency([{20, alpha()}, % alpha + {10, digit()}, % digit + {10, unicode_char()} % unicode + ])). + +scheme_char() -> + frequency([{20, alpha()}, % alpha + {20, digit()}, % digit + {5, oneof([$+, $-, $.])} % punctuation + ]). + +sub_delims() -> + oneof([$!, $$, $&, $', $(, $), + $*, $+, $,,$;, $=]). + +pchar() -> + frequency([{20, unreserved()}, + {5, pct_encoded()}, + {5, sub_delims()}, + {1, oneof([$:, $@])} % punctuation + ]). + +unreserved() -> + frequency([{20, alpha()}, + {5, digit()}, + {1, oneof([$-, $., $_, $~])} % punctuation + ]). + +unicode_char() -> + range(913, 1023). + +alpha() -> + frequency([{20, range($a, $z)}, % letters + {20, range($A, $Z)}]). % letters + +digit() -> + range($0, $9). % numbers + +pct_encoded() -> + oneof(["%C3%A4", "%C3%A5", "%C3%B6"]). + + +%%%======================================================================== +%%% Helpers +%%%======================================================================== +proplist_to_map(L) -> + lists:foldl(fun({K,V},M) -> M#{K => V}; + (_,M) -> M + end, #{}, L). diff --git a/lib/stdlib/test/qlc_SUITE.erl b/lib/stdlib/test/qlc_SUITE.erl index 5e9e03e410..949142ec77 100644 --- a/lib/stdlib/test/qlc_SUITE.erl +++ b/lib/stdlib/test/qlc_SUITE.erl @@ -7871,7 +7871,7 @@ run_test(Config, Extra, {cres, Body, Opts, ExpectedCompileReturn}) -> {module, _} = code:load_abs(AbsFile, Mod), Ms0 = erlang:process_info(self(),messages), - Before = {{get(), lists:sort(ets:all()), Ms0}, pps()}, + Before = {{lget(), lists:sort(ets:all()), Ms0}, pps()}, %% Prepare the check that the qlc module does not call qlc_pt. _ = [unload_pt() || {file, Name} <- [code:is_loaded(qlc_pt)], @@ -7903,7 +7903,7 @@ run_test(Config, Extra, Body) -> wait_for_expected(R, {Strict0,PPS0}=Before, SourceFile, Wait) -> Ms = erlang:process_info(self(),messages), - After = {_,PPS1} = {{get(), lists:sort(ets:all()), Ms}, pps()}, + After = {_,PPS1} = {{lget(), lists:sort(ets:all()), Ms}, pps()}, case {R, After} of {ok, Before} -> ok; @@ -7931,6 +7931,18 @@ wait_for_expected(R, {Strict0,PPS0}=Before, SourceFile, Wait) -> expected({ok,Before}, {R,After}, SourceFile) end. +%% The qlc modules uses the process dictionary for storing names of files. +lget() -> + lists:sort([T || {K, _} = T <- get(), is_qlc_key(K)]). + +%% Copied from the qlc module. +-define(LCACHE_FILE(Ref), {Ref, '$_qlc_cache_tmpfiles_'}). +-define(MERGE_JOIN_FILE, '$_qlc_merge_join_tmpfiles_'). + +is_qlc_key(?LCACHE_FILE(_)) -> true; +is_qlc_key(?MERGE_JOIN_FILE) -> true; +is_qlc_key(_) -> false. + unload_pt() -> erlang:garbage_collect(), % get rid of references to qlc_pt... _ = code:purge(qlc_pt), diff --git a/lib/stdlib/test/rand_SUITE.erl b/lib/stdlib/test/rand_SUITE.erl index f69d42551e..ef4f9faad9 100644 --- a/lib/stdlib/test/rand_SUITE.erl +++ b/lib/stdlib/test/rand_SUITE.erl @@ -29,11 +29,14 @@ basic_stats_uniform_1/1, basic_stats_uniform_2/1, basic_stats_standard_normal/1, basic_stats_normal/1, + uniform_real_conv/1, plugin/1, measure/1, reference_jump_state/1, reference_jump_procdict/1]). -export([test/0, gen/1]). +-export([uniform_real_gen/1, uniform_gen/2]). + -include_lib("common_test/include/ct.hrl"). -define(LOOP, 1000000). @@ -46,7 +49,7 @@ all() -> [seed, interval_int, interval_float, api_eq, reference, - {group, basic_stats}, + {group, basic_stats}, uniform_real_conv, plugin, measure, {group, reference_jump} ]. @@ -101,7 +104,7 @@ seed_1(Alg) -> _ = rand:uniform(), S00 = get(rand_seed), erase(), - _ = rand:uniform(), + _ = rand:uniform_real(), false = S00 =:= get(rand_seed), %% hopefully %% Choosing algo and seed @@ -228,11 +231,13 @@ interval_float(Config) when is_list(Config) -> interval_float_1(0) -> ok; interval_float_1(N) -> X = rand:uniform(), + Y = rand:uniform_real(), if - 0.0 =< X, X < 1.0 -> + 0.0 =< X, X < 1.0, 0.0 < Y, Y < 1.0 -> ok; true -> - io:format("X=~p 0=<~p<1.0~n", [X,X]), + io:format("X=~p 0.0=<~p<1.0~n", [X,X]), + io:format("Y=~p 0.0<~p<1.0~n", [Y,Y]), exit({X, rand:export_seed()}) end, interval_float_1(N-1). @@ -334,7 +339,13 @@ basic_stats_normal(Config) when is_list(Config) -> IntendedMeanVariancePairs). basic_uniform_1(N, S0, Sum, A0) when N > 0 -> - {X,S} = rand:uniform_s(S0), + {X,S} = + case N band 1 of + 0 -> + rand:uniform_s(S0); + 1 -> + rand:uniform_real_s(S0) + end, I = trunc(X*100), A = array:set(I, 1+array:get(I,A0), A0), basic_uniform_1(N-1, S, Sum+X, A); @@ -400,6 +411,137 @@ normal_s(Mean, Variance, State0) -> rand:normal_s(Mean, Variance, State0). %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% +%% White box test of the conversion to float + +uniform_real_conv(Config) when is_list(Config) -> + [begin +%% ct:pal("~13.16.0bx~3.16.0b: ~p~n", [M,E,Gen]), + uniform_real_conv_check(M, E, Gen) + end || {M, E, Gen} <- uniform_real_conv_data()], + uniform_real_scan(0), + uniform_real_scan(3). + +uniform_real_conv_data() -> + [{16#fffffffffffff, -1, [16#3ffffffffffffff]}, + {16#fffffffffffff, -1, [16#3ffffffffffffe0]}, + {16#ffffffffffffe, -1, [16#3ffffffffffffdf]}, + %% + {16#0000000000000, -1, [16#200000000000000]}, + {16#fffffffffffff, -2, [16#1ffffffffffffff]}, + {16#fffffffffffff, -2, [16#1fffffffffffff0]}, + {16#ffffffffffffe, -2, [16#1ffffffffffffef]}, + %% + {16#0000000000000, -2, [16#100000000000000]}, + {16#fffffffffffff, -3, [16#0ffffffffffffff]}, + {16#fffffffffffff, -3, [16#0fffffffffffff8]}, + {16#ffffffffffffe, -3, [16#0fffffffffffff7]}, + %% + {16#0000000000000, -3, [16#080000000000000]}, + {16#fffffffffffff, -4, [16#07fffffffffffff]}, + {16#fffffffffffff, -4, [16#07ffffffffffffc]}, + {16#ffffffffffffe, -4, [16#07ffffffffffffb]}, + %% + {16#0000000000000, -4, [16#040000000000000]}, + {16#fffffffffffff, -5, [16#03fffffffffffff,16#3ffffffffffffff]}, + {16#fffffffffffff, -5, [16#03ffffffffffffe,16#200000000000000]}, + {16#ffffffffffffe, -5, [16#03fffffffffffff,16#1ffffffffffffff]}, + {16#ffffffffffffe, -5, [16#03fffffffffffff,16#100000000000000]}, + %% + {16#0000000000001, -56, [16#000000000000007,16#00000000000007f]}, + {16#0000000000001, -56, [16#000000000000004,16#000000000000040]}, + {16#0000000000000, -57, [16#000000000000003,16#20000000000001f]}, + {16#0000000000000, -57, [16#000000000000000,16#200000000000000]}, + {16#fffffffffffff, -58, [16#000000000000003,16#1ffffffffffffff]}, + {16#fffffffffffff, -58, [16#000000000000000,16#1fffffffffffff0]}, + {16#ffffffffffffe, -58, [16#000000000000000,16#1ffffffffffffef]}, + {16#ffffffffffffe, -58, [16#000000000000000,16#1ffffffffffffe0]}, + %% + {16#0000000000000, -58, [16#000000000000000,16#10000000000000f]}, + {16#0000000000000, -58, [16#000000000000000,16#100000000000000]}, + {2#11001100000000000000000000000000000000000011000000011, % 53 bits + -1022, + [0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, % 18 zeros + 2#1100110000000000000000000000000000000000001 bsl 2, % 43 bits + 2#1000000011 bsl (56-10+2)]}, % 10 bits + {0, -1, % 0.5 after retry + [0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, % 18 zeros + 2#111111111111111111111111111111111111111111 bsl 2, % 42 bits - retry + 16#200000000000003]}]. % 0.5 + +-define(UNIFORM_REAL_SCAN_PATTERN, (16#19000000000009)). % 53 bits +-define(UNIFORM_REAL_SCAN_NUMBER, (1021)). + +uniform_real_scan_template(K) -> + <<0:?UNIFORM_REAL_SCAN_NUMBER, + ?UNIFORM_REAL_SCAN_PATTERN:53,K:2,0:1>>. + +uniform_real_scan(K) -> + Templ = uniform_real_scan_template(K), + N = ?UNIFORM_REAL_SCAN_NUMBER, + uniform_real_scan(Templ, N, K). + +uniform_real_scan(Templ, N, K) when 0 =< N -> + <<_:N/bits,T/bits>> = Templ, + Data = uniform_real_scan_data(T, K), + uniform_real_conv_check( + ?UNIFORM_REAL_SCAN_PATTERN, N - 1 - ?UNIFORM_REAL_SCAN_NUMBER, Data), + uniform_real_scan(Templ, N - 1, K); +uniform_real_scan(_, _, _) -> + ok. + +uniform_real_scan_data(Templ, K) -> + case Templ of + <<X:56, T/bits>> -> + B = rand:bc64(X), + [(X bsl 2) bor K | + if + 53 =< B -> + []; + true -> + uniform_real_scan_data(T, K) + end]; + _ -> + <<X:56, _/bits>> = <<Templ/bits, 0:56>>, + [(X bsl 2) bor K] + end. + +uniform_real_conv_check(M, E, Gen) -> + <<F/float>> = <<0:1, (E + 16#3ff):11, M:52>>, + try uniform_real_gen(Gen) of + F -> F; + FF -> + ct:pal( + "~s =/= ~s: ~s~n", + [rand:float2str(FF), rand:float2str(F), + [["16#",integer_to_list(G,16),$\s]||G<-Gen]]), + ct:fail({neq, FF, F}) + catch + Error:Reason -> + ct:pal( + "~w:~p ~s: ~s~n", + [Error, Reason, rand:float2str(F), + [["16#",integer_to_list(G,16),$\s]||G<-Gen]]), + ct:fail({Error, Reason, F, erlang:get_stacktrace()}) + end. + + +uniform_real_gen(Gen) -> + State = rand_state(Gen), + {F, {#{type := rand_SUITE_list},[]}} = rand:uniform_real_s(State), + F. + +uniform_gen(Range, Gen) -> + State = rand_state(Gen), + {N, {#{type := rand_SUITE_list},[]}} = rand:uniform_s(Range, State), + N. + +%% Loaded dice for white box tests +rand_state(Gen) -> + {#{type => rand_SUITE_list, bits => 58, weak_low_bits => 1, + next => fun ([H|T]) -> {H, T} end}, + Gen}. + +%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Test that the user can write algorithms. plugin(Config) when is_list(Config) -> @@ -520,6 +662,21 @@ do_measure(_Config) -> end, Algs), %% + ct:pal("~nRNG uniform integer half range performance~n",[]), + _ = + measure_1( + fun (State) -> half_range(State) end, + fun (State, Range, Mod) -> + measure_loop( + fun (St0) -> + ?CHECK_UNIFORM_RANGE( + Mod:uniform_s(Range, St0), Range, + X, St1) + end, + State) + end, + Algs), + %% ct:pal("~nRNG uniform integer half range + 1 performance~n",[]), _ = measure_1( @@ -630,7 +787,8 @@ do_measure(_Config) -> Algs), %% ct:pal("~nRNG uniform float performance~n",[]), - _ = measure_1( + _ = + measure_1( fun (_) -> 0 end, fun (State, _, Mod) -> measure_loop( @@ -641,8 +799,22 @@ do_measure(_Config) -> end, Algs), %% + ct:pal("~nRNG uniform_real float performance~n",[]), + _ = + measure_1( + fun (_) -> 0 end, + fun (State, _, Mod) -> + measure_loop( + fun (St0) -> + ?CHECK_UNIFORM(Mod:uniform_real_s(St0), X, St) + end, + State) + end, + Algs), + %% ct:pal("~nRNG normal float performance~n",[]), - _ = measure_1( + [TMarkNormalFloat|_] = + measure_1( fun (_) -> 0 end, fun (State, _, Mod) -> measure_loop( @@ -652,10 +824,36 @@ do_measure(_Config) -> State) end, Algs), + %% Just for fun try an implementation of the Box-Muller + %% transformation for creating normal distribution floats + %% to compare with our Ziggurat implementation. + %% Generates two numbers per call that we add so they + %% will not be optimized away. Hence the benchmark time + %% is twice what it should be. + TwoPi = 2 * math:pi(), + _ = + measure_1( + fun (_) -> 0 end, + fun (State, _, Mod) -> + measure_loop( + fun (State0) -> + {U1, State1} = Mod:uniform_real_s(State0), + {U2, State2} = Mod:uniform_s(State1), + R = math:sqrt(-2.0 * math:log(U1)), + T = TwoPi * U2, + Z0 = R * math:cos(T), + Z1 = R * math:sin(T), + ?CHECK_NORMAL({Z0 + Z1, State2}, X, State3) + end, + State) + end, + exrop, TMarkNormalFloat), ok. +-define(LOOP_MEASURE, (?LOOP div 5)). + measure_loop(Fun, State) -> - measure_loop(Fun, State, ?LOOP). + measure_loop(Fun, State, ?LOOP_MEASURE). %% measure_loop(Fun, State, N) when 0 < N -> measure_loop(Fun, Fun(State), N-1); @@ -693,7 +891,8 @@ measure_1(RangeFun, Fun, Alg, TMark) -> end, io:format( "~.12w: ~p ns ~p% [16#~.16b]~n", - [Alg, (Time * 1000 + 500) div ?LOOP, Percent, Range]), + [Alg, (Time * 1000 + 500) div ?LOOP_MEASURE, + Percent, Range]), Parent ! {self(), Time}, normal end), diff --git a/lib/stdlib/test/uri_string_SUITE.erl b/lib/stdlib/test/uri_string_SUITE.erl new file mode 100644 index 0000000000..c625da56c6 --- /dev/null +++ b/lib/stdlib/test/uri_string_SUITE.erl @@ -0,0 +1,844 @@ +%% +%% %CopyrightBegin% +%% +%% Copyright Ericsson AB 2008-2017. All Rights Reserved. +%% +%% Licensed under the Apache License, Version 2.0 (the "License"); +%% you may not use this file except in compliance with the License. +%% You may obtain a copy of the License at +%% +%% http://www.apache.org/licenses/LICENSE-2.0 +%% +%% Unless required by applicable law or agreed to in writing, software +%% distributed under the License is distributed on an "AS IS" BASIS, +%% WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +%% See the License for the specific language governing permissions and +%% limitations under the License. +%% +%% %CopyrightEnd% +%% +-module(uri_string_SUITE). + +-include_lib("common_test/include/ct.hrl"). + +-export([all/0, suite/0,groups/0, + normalize/1, + parse_binary_fragment/1, parse_binary_host/1, parse_binary_host_ipv4/1, + parse_binary_host_ipv6/1, + parse_binary_path/1, parse_binary_pct_encoded_fragment/1, parse_binary_pct_encoded_query/1, + parse_binary_pct_encoded_userinfo/1, parse_binary_port/1, + parse_binary_query/1, parse_binary_scheme/1, parse_binary_userinfo/1, + parse_fragment/1, parse_host/1, parse_host_ipv4/1, parse_host_ipv6/1, + parse_path/1, parse_pct_encoded_fragment/1, parse_pct_encoded_query/1, + parse_pct_encoded_userinfo/1, parse_port/1, + parse_query/1, parse_scheme/1, parse_userinfo/1, + parse_list/1, parse_binary/1, parse_mixed/1, parse_relative/1, + parse_special/1, parse_special2/1, parse_negative/1, + recompose_fragment/1, recompose_parse_fragment/1, + recompose_query/1, recompose_parse_query/1, + recompose_path/1, recompose_parse_path/1, + recompose_autogen/1, parse_recompose_autogen/1, + transcode_basic/1, transcode_options/1, transcode_mixed/1, transcode_negative/1 + ]). + + +-define(SCHEME, "foo"). +-define(USERINFO, "åsa"). +-define(USERINFO_ENC, "%C3%A5sa"). +-define(HOST, "älvsjö"). +-define(HOST_ENC, "%C3%A4lvsj%C3%B6"). +-define(IPV6, "::127.0.0.1"). +-define(IPV6_ENC, "[::127.0.0.1]"). +-define(PORT, 8042). +-define(PORT_ENC, ":8042"). +-define(PATH, "/där"). +-define(PATH_ENC, "/d%C3%A4r"). +-define(QUERY, "name=örn"). +-define(QUERY_ENC, "?name=%C3%B6rn"). +-define(FRAGMENT, "näsa"). +-define(FRAGMENT_ENC, "#n%C3%A4sa"). + + +suite() -> + [{timetrap,{minutes,1}}]. + +all() -> + [ + normalize, + parse_binary_scheme, + parse_binary_userinfo, + parse_binary_pct_encoded_userinfo, + parse_binary_host, + parse_binary_host_ipv4, + parse_binary_host_ipv6, + parse_binary_port, + parse_binary_path, + parse_binary_query, + parse_binary_pct_encoded_query, + parse_binary_fragment, + parse_binary_pct_encoded_fragment, + parse_scheme, + parse_userinfo, + parse_pct_encoded_userinfo, + parse_host, + parse_host_ipv4, + parse_host_ipv6, + parse_port, + parse_path, + parse_query, + parse_pct_encoded_query, + parse_fragment, + parse_pct_encoded_fragment, + parse_list, + parse_binary, + parse_mixed, + parse_relative, + parse_special, + parse_special2, + parse_negative, + recompose_fragment, + recompose_parse_fragment, + recompose_query, + recompose_parse_query, + recompose_path, + recompose_parse_path, + recompose_autogen, + parse_recompose_autogen, + transcode_basic, + transcode_options, + transcode_mixed, + transcode_negative + ]. + +groups() -> + []. + + +%%------------------------------------------------------------------------- +%% Helper functions +%%------------------------------------------------------------------------- +uri_combinations() -> + [[Sch,Usr,Hst,Prt,Pat,Qry,Frg] || + Sch <- [fun update_scheme/1, fun update_scheme_binary/1, none], + Usr <- [fun update_userinfo/1, fun update_userinfo_binary/1, none], + Hst <- [fun update_host/1, fun update_host_binary/1, + fun update_ipv6/1, fun update_ipv6_binary/1, none], + Prt <- [fun update_port/1, none], + Pat <- [fun update_path/1, fun update_path_binary/1], + Qry <- [fun update_query/1,fun update_query_binary/1, none], + Frg <- [fun update_fragment/1, fun update_fragment_binary/1, none], + not (Usr =:= none andalso Hst =:= none andalso Prt =/= none), + not (Usr =/= none andalso Hst =:= none andalso Prt =:= none), + not (Usr =/= none andalso Hst =:= none andalso Prt =/= none)]. + + +generate_test_vector(Comb) -> + Fun = fun (F, {Map, URI}) when is_function(F) -> F({Map, URI}); + (_, Map) -> Map + end, + lists:foldl(Fun, {#{}, empty}, Comb). + +generate_test_vectors(L) -> + lists:map(fun generate_test_vector/1, L). + +update_fragment({In, empty}) -> + {In#{fragment => ?FRAGMENT}, ?FRAGMENT_ENC}; +update_fragment({In, Out}) when is_list(Out) -> + {In#{fragment => ?FRAGMENT}, Out ++ ?FRAGMENT_ENC}; +update_fragment({In, Out}) when is_binary(Out) -> + {In#{fragment => ?FRAGMENT}, binary_to_list(Out) ++ ?FRAGMENT_ENC}. + +update_fragment_binary({In, empty}) -> + {In#{fragment => <<?FRAGMENT/utf8>>}, <<?FRAGMENT_ENC>>}; +update_fragment_binary({In, Out}) when is_list(Out) -> + {In#{fragment => <<?FRAGMENT/utf8>>}, Out ++ ?FRAGMENT_ENC}; +update_fragment_binary({In, Out}) when is_binary(Out) -> + {In#{fragment => <<?FRAGMENT/utf8>>}, <<Out/binary,?FRAGMENT_ENC>>}. + + +update_query({In, empty}) -> + {In#{query => ?QUERY}, ?QUERY_ENC}; +update_query({In, Out}) when is_list(Out) -> + {In#{query => ?QUERY}, Out ++ ?QUERY_ENC}; +update_query({In, Out}) when is_binary(Out) -> + {In#{query => ?QUERY}, binary_to_list(Out) ++ ?QUERY_ENC}. + +update_query_binary({In, empty}) -> + {In#{query => <<?QUERY/utf8>>}, <<?QUERY_ENC>>}; +update_query_binary({In, Out}) when is_list(Out) -> + {In#{query => <<?QUERY/utf8>>}, Out ++ ?QUERY_ENC}; +update_query_binary({In, Out}) when is_binary(Out) -> + {In#{query => <<?QUERY/utf8>>}, <<Out/binary,?QUERY_ENC>>}. + +update_path({In, empty}) -> + {In#{path => ?PATH}, ?PATH_ENC}; +update_path({In, Out}) when is_list(Out) -> + {In#{path => ?PATH}, Out ++ ?PATH_ENC}; +update_path({In, Out}) when is_binary(Out) -> + {In#{path => ?PATH}, binary_to_list(Out) ++ ?PATH_ENC}. + +update_path_binary({In, empty}) -> + {In#{path => <<?PATH/utf8>>}, <<?PATH_ENC>>}; +update_path_binary({In, Out}) when is_list(Out) -> + {In#{path => <<?PATH/utf8>>}, Out ++ ?PATH_ENC}; +update_path_binary({In, Out}) when is_binary(Out) -> + {In#{path => <<?PATH/utf8>>}, <<Out/binary,?PATH_ENC>>}. + +update_port({In, Out}) when is_list(Out) -> + {In#{port => ?PORT}, Out ++ ?PORT_ENC}; +update_port({In, Out}) when is_binary(Out) -> + {In#{port => ?PORT}, <<Out/binary,?PORT_ENC>>}. + +update_host({In, empty}) -> + {In#{host => ?HOST}, "//" ++ ?HOST_ENC}; +update_host({In, Out}) when is_list(Out) -> + case maps:is_key(userinfo, In) of + true -> {In#{host => ?HOST}, Out ++ [$@|?HOST_ENC]}; + false -> {In#{host => ?HOST}, Out ++ [$/,$/|?HOST_ENC]} + end; +update_host({In, Out}) when is_binary(Out) -> + case maps:is_key(userinfo, In) of + true -> {In#{host => ?HOST}, binary_to_list(Out) ++ [$@|?HOST_ENC]}; + false -> {In#{host => ?HOST}, binary_to_list(Out) ++ [$/,$/|?HOST_ENC]} + end. + +update_host_binary({In, empty}) -> + {In#{host => <<?HOST/utf8>>}, <<"//",?HOST_ENC>>}; +update_host_binary({In, Out}) when is_list(Out) -> + case maps:is_key(userinfo, In) of + true -> {In#{host => <<?HOST/utf8>>}, Out ++ [$@|?HOST_ENC]}; + false -> {In#{host => <<?HOST/utf8>>}, Out ++ [$/,$/|?HOST_ENC]} + end; +update_host_binary({In, Out}) when is_binary(Out) -> + case maps:is_key(userinfo, In) of + true -> {In#{host => <<?HOST/utf8>>}, <<Out/binary,$@,?HOST_ENC>>}; + false-> {In#{host => <<?HOST/utf8>>}, <<Out/binary,"//",?HOST_ENC>>} + end. + +update_ipv6({In, empty}) -> + {In#{host => ?IPV6}, "//" ++ ?IPV6_ENC}; +update_ipv6({In, Out}) when is_list(Out) -> + case maps:is_key(userinfo, In) of + true -> {In#{host => ?IPV6}, Out ++ [$@|?IPV6_ENC]}; + false -> {In#{host => ?IPV6}, Out ++ [$/,$/|?IPV6_ENC]} + end; +update_ipv6({In, Out}) when is_binary(Out) -> + case maps:is_key(userinfo, In) of + true -> {In#{host => ?IPV6}, binary_to_list(Out) ++ [$@|?IPV6_ENC]}; + false -> {In#{host => ?IPV6}, binary_to_list(Out) ++ [$/,$/|?IPV6_ENC]} + end. + +update_ipv6_binary({In, empty}) -> + {In#{host => <<?IPV6/utf8>>}, <<"//",?IPV6_ENC>>}; +update_ipv6_binary({In, Out}) when is_list(Out) -> + case maps:is_key(userinfo, In) of + true -> {In#{host => <<?IPV6/utf8>>}, Out ++ [$@|?IPV6_ENC]}; + false -> {In#{host => <<?IPV6/utf8>>}, Out ++ [$/,$/|?IPV6_ENC]} + end; +update_ipv6_binary({In, Out}) when is_binary(Out) -> + case maps:is_key(userinfo, In) of + true -> {In#{host => <<?IPV6/utf8>>}, <<Out/binary,$@,?IPV6_ENC>>}; + false-> {In#{host => <<?IPV6/utf8>>}, <<Out/binary,"//",?IPV6_ENC>>} + end. + +update_userinfo({In, empty}) -> + {In#{userinfo => ?USERINFO}, "//" ++ ?USERINFO_ENC}; +update_userinfo({In, Out}) when is_list(Out) -> + {In#{userinfo => ?USERINFO}, Out ++ "//" ++ ?USERINFO_ENC}; +update_userinfo({In, Out}) when is_binary(Out) -> + {In#{userinfo => ?USERINFO}, binary_to_list(Out) ++ "//" ++ ?USERINFO_ENC}. + +update_userinfo_binary({In, empty}) -> + {In#{userinfo => <<?USERINFO/utf8>>}, <<"//",?USERINFO_ENC>>}; +update_userinfo_binary({In, Out}) when is_list(Out) -> + {In#{userinfo => <<?USERINFO/utf8>>}, Out ++ "//" ++ ?USERINFO_ENC}; +update_userinfo_binary({In, Out}) when is_binary(Out) -> + {In#{userinfo => <<?USERINFO/utf8>>}, <<Out/binary,"//",?USERINFO_ENC>>}. + +update_scheme({In, empty}) -> + {In#{scheme => ?SCHEME}, ?SCHEME ++ ":"}. + +update_scheme_binary({In, empty}) -> + {In#{scheme => <<?SCHEME/utf8>>}, <<?SCHEME,$:>>}. + + +%% Test recompose on a generated test vector +run_test_recompose({#{}, empty}) -> + try "" = uri_string:recompose(#{}) of + _ -> ok + catch + _:_ -> error({test_failed, #{}, ""}) + end; +run_test_recompose({Map, URI}) -> + try URI = uri_string:recompose(Map) of + URI -> ok + catch + _:_ -> error({test_failed, Map, URI}) + end. + +%% Test parse - recompose on a generated test vector +run_test_parse_recompose({#{}, empty}) -> + try "" = uri_string:recompose(uri_string:parse("")) of + _ -> ok + catch + _:_ -> error({test_failed, #{}, ""}) + end; +run_test_parse_recompose({Map, URI}) -> + try URI = uri_string:recompose(uri_string:parse(URI)) of + URI -> ok + catch + _:_ -> error({test_failed, Map, URI}) + end. + + +%%------------------------------------------------------------------------- +%% Parse tests +%%------------------------------------------------------------------------- + +parse_binary_scheme(_Config) -> + #{} = uri_string:parse(<<>>), + #{path := <<"foo">>} = uri_string:parse(<<"foo">>), + #{scheme := <<"foo">>} = uri_string:parse(<<"foo:">>), + #{scheme := <<"foo">>, path := <<"bar:nisse">>} = uri_string:parse(<<"foo:bar:nisse">>), + #{scheme := <<"foo">>, host := <<"">>} = uri_string:parse(<<"foo://">>), + #{scheme := <<"foo">>, host := <<"">>, path := <<"/">>} = uri_string:parse(<<"foo:///">>), + #{scheme := <<"foo">>, host := <<"">>, path := <<"//">>} = uri_string:parse(<<"foo:////">>), + + #{path := <<"/">>} = uri_string:parse(<<"/">>), + #{host := <<>>} = uri_string:parse(<<"//">>), + #{host := <<>>, path := <<"/">>} = uri_string:parse(<<"///">>). + +parse_binary_userinfo(_Config) -> + #{scheme := <<"user">>, path := <<"password@localhost">>} = + uri_string:parse(<<"user:password@localhost">>), + #{path := <<"user@">>} = uri_string:parse(<<"user@">>), + #{path := <<"/user@">>} = uri_string:parse(<<"/user@">>), + #{path := <<"user@localhost">>} = uri_string:parse(<<"user@localhost">>), + #{userinfo := <<"user">>, host := <<"localhost">>} = uri_string:parse(<<"//user@localhost">>), + #{userinfo := <<"user:password">>, host := <<"localhost">>} = + uri_string:parse(<<"//user:password@localhost">>), + #{scheme := <<"foo">>, path := <<"/user@">>} = + uri_string:parse(<<"foo:/user@">>), + #{scheme := <<"foo">>, userinfo := <<"user">>, host := <<"localhost">>} = + uri_string:parse(<<"foo://user@localhost">>), + #{scheme := <<"foo">>, userinfo := <<"user:password">>, host := <<"localhost">>} = + uri_string:parse(<<"foo://user:password@localhost">>). + +parse_binary_pct_encoded_userinfo(_Config) -> + #{scheme := <<"user">>, path := <<"合@気道"/utf8>>} = + uri_string:parse(<<"user:%E5%90%88@%E6%B0%97%E9%81%93">>), + #{path := <<"合気道@"/utf8>>} = uri_string:parse(<<"%E5%90%88%E6%B0%97%E9%81%93@">>), + #{path := <<"/合気道@"/utf8>>} = uri_string:parse(<<"/%E5%90%88%E6%B0%97%E9%81%93@">>), + #{path := <<"合@気道"/utf8>>} = uri_string:parse(<<"%E5%90%88@%E6%B0%97%E9%81%93">>), + #{userinfo := <<"合"/utf8>>, host := <<"気道"/utf8>>} = + uri_string:parse(<<"//%E5%90%88@%E6%B0%97%E9%81%93">>), + #{userinfo := <<"合:気"/utf8>>, host := <<"道"/utf8>>} = + uri_string:parse(<<"//%E5%90%88:%E6%B0%97@%E9%81%93">>), + #{scheme := <<"foo">>, path := <<"/合気道@"/utf8>>} = + uri_string:parse(<<"foo:/%E5%90%88%E6%B0%97%E9%81%93@">>), + #{scheme := <<"foo">>, userinfo := <<"合"/utf8>>, host := <<"気道"/utf8>>} = + uri_string:parse(<<"foo://%E5%90%88@%E6%B0%97%E9%81%93">>), + #{scheme := <<"foo">>, userinfo := <<"合:気"/utf8>>, host := <<"道"/utf8>>} = + uri_string:parse(<<"foo://%E5%90%88:%E6%B0%97@%E9%81%93">>), + {error,invalid_uri,"@"} = uri_string:parse(<<"//%E5%90%88@%E6%B0%97%E9%81%93@">>), + {error,invalid_uri,":"} = uri_string:parse(<<"foo://%E5%90%88@%E6%B0%97%E9%81%93@">>). + +parse_binary_host(_Config) -> + #{host := <<"hostname">>} = uri_string:parse(<<"//hostname">>), + #{host := <<"hostname">>,scheme := <<"foo">>} = uri_string:parse(<<"foo://hostname">>), + #{host := <<"hostname">>,scheme := <<"foo">>, userinfo := <<"user">>} = + uri_string:parse(<<"foo://user@hostname">>). + +parse_binary_host_ipv4(_Config) -> + #{host := <<"127.0.0.1">>} = uri_string:parse(<<"//127.0.0.1">>), + #{host := <<"127.0.0.1">>, path := <<"/over/there">>} = + uri_string:parse(<<"//127.0.0.1/over/there">>), + #{host := <<"127.0.0.1">>, query := <<"name=ferret">>} = + uri_string:parse(<<"//127.0.0.1?name=ferret">>), + #{host := <<"127.0.0.1">>, fragment := <<"nose">>} = uri_string:parse(<<"//127.0.0.1#nose">>), + {error,invalid_uri,"x"} = uri_string:parse(<<"//127.0.0.x">>), + {error,invalid_uri,"1227.0.0.1"} = uri_string:parse(<<"//1227.0.0.1">>). + +parse_binary_host_ipv6(_Config) -> + #{host := <<"::127.0.0.1">>} = uri_string:parse(<<"//[::127.0.0.1]">>), + #{host := <<"2001:0db8:0000:0000:0000:0000:1428:07ab">>} = + uri_string:parse(<<"//[2001:0db8:0000:0000:0000:0000:1428:07ab]">>), + #{host := <<"::127.0.0.1">>, path := <<"/over/there">>} = + uri_string:parse(<<"//[::127.0.0.1]/over/there">>), + #{host := <<"::127.0.0.1">>, query := <<"name=ferret">>} = + uri_string:parse(<<"//[::127.0.0.1]?name=ferret">>), + #{host := <<"::127.0.0.1">>, fragment := <<"nose">>} = + uri_string:parse(<<"//[::127.0.0.1]#nose">>), + {error,invalid_uri,"x"} = uri_string:parse(<<"//[::127.0.0.x]">>), + {error,invalid_uri,"::1227.0.0.1"} = uri_string:parse(<<"//[::1227.0.0.1]">>), + {error,invalid_uri,"G"} = uri_string:parse(<<"//[2001:0db8:0000:0000:0000:0000:1428:G7ab]">>). + +parse_binary_port(_Config) -> + #{path:= <<"/:8042">>} = + uri_string:parse(<<"/:8042">>), + #{host:= <<>>, port := 8042} = + uri_string:parse(<<"//:8042">>), + #{host := <<"example.com">>, port:= 8042} = + uri_string:parse(<<"//example.com:8042">>), + #{scheme := <<"foo">>, path := <<"/:8042">>} = + uri_string:parse(<<"foo:/:8042">>), + #{scheme := <<"foo">>, host := <<>>, port := 8042} = + uri_string:parse(<<"foo://:8042">>), + #{scheme := <<"foo">>, host := <<"example.com">>, port := 8042} = + uri_string:parse(<<"foo://example.com:8042">>), + {error,invalid_uri,":"} = uri_string:parse(":600"), + {error,invalid_uri,"x"} = uri_string:parse("//:8042x"). + +parse_binary_path(_Config) -> + #{path := <<"over/there">>} = uri_string:parse(<<"over/there">>), + #{path := <<"/over/there">>} = uri_string:parse(<<"/over/there">>), + #{scheme := <<"foo">>, path := <<"/over/there">>} = + uri_string:parse(<<"foo:/over/there">>), + #{scheme := <<"foo">>, host := <<"example.com">>, path := <<"/over/there">>} = + uri_string:parse(<<"foo://example.com/over/there">>), + #{scheme := <<"foo">>, host := <<"example.com">>, path := <<"/over/there">>, port := 8042} = + uri_string:parse(<<"foo://example.com:8042/over/there">>). + +parse_binary_query(_Config) -> + #{scheme := <<"foo">>, query := <<"name=ferret">>} = + uri_string:parse(<<"foo:?name=ferret">>), + #{scheme := <<"foo">>, path:= <<"over/there">>, query := <<"name=ferret">>} = + uri_string:parse(<<"foo:over/there?name=ferret">>), + #{scheme := <<"foo">>, path:= <<"/over/there">>, query := <<"name=ferret">>} = + uri_string:parse(<<"foo:/over/there?name=ferret">>), + #{scheme := <<"foo">>, host := <<"example.com">>, query := <<"name=ferret">>} = + uri_string:parse(<<"foo://example.com?name=ferret">>), + #{scheme := <<"foo">>, host := <<"example.com">>, path := <<"/">>, query := <<"name=ferret">>} = + uri_string:parse(<<"foo://example.com/?name=ferret">>), + + #{path := <<>>, query := <<"name=ferret">>} = + uri_string:parse(<<"?name=ferret">>), + #{path := <<"over/there">>, query := <<"name=ferret">>} = + uri_string:parse(<<"over/there?name=ferret">>), + #{path := <<"/">>, query := <<"name=ferret">>} = + uri_string:parse(<<"/?name=ferret">>), + #{path := <<"/over/there">>, query := <<"name=ferret">>} = + uri_string:parse(<<"/over/there?name=ferret">>), + #{host := <<"example.com">>, query := <<"name=ferret">>} = + uri_string:parse(<<"//example.com?name=ferret">>), + #{host := <<"example.com">>, path := <<"/">>, query := <<"name=ferret">>} = + uri_string:parse(<<"//example.com/?name=ferret">>). + +parse_binary_pct_encoded_query(_Config) -> + #{scheme := <<"foo">>, host := <<"example.com">>, path := <<"/">>, + query := <<"name=合気道"/utf8>>} = + uri_string:parse(<<"foo://example.com/?name=%E5%90%88%E6%B0%97%E9%81%93">>), + #{host := <<"example.com">>, path := <<"/">>, query := <<"name=合気道"/utf8>>} = + uri_string:parse(<<"//example.com/?name=%E5%90%88%E6%B0%97%E9%81%93">>). + +parse_binary_fragment(_Config) -> + #{scheme := <<"foo">>, fragment := <<"nose">>} = + uri_string:parse(<<"foo:#nose">>), + #{scheme := <<"foo">>, path:= <<"over/there">>, fragment := <<"nose">>} = + uri_string:parse(<<"foo:over/there#nose">>), + #{scheme := <<"foo">>, path:= <<"/over/there">>, fragment := <<"nose">>} = + uri_string:parse(<<"foo:/over/there#nose">>), + #{scheme := <<"foo">>, host := <<"example.com">>, fragment := <<"nose">>} = + uri_string:parse(<<"foo://example.com#nose">>), + #{scheme := <<"foo">>, host := <<"example.com">>, path := <<"/">>, fragment := <<"nose">>} = + uri_string:parse(<<"foo://example.com/#nose">>), + #{scheme := <<"foo">>, host := <<"example.com">>, fragment := <<"nose">>} = + uri_string:parse(<<"foo://example.com#nose">>), + + #{fragment := <<"nose">>} = + uri_string:parse(<<"#nose">>), + #{path := <<"over/there">>, fragment := <<"nose">>} = + uri_string:parse(<<"over/there#nose">>), + #{path := <<"/">>, fragment := <<"nose">>} = + uri_string:parse(<<"/#nose">>), + #{path := <<"/over/there">>, fragment := <<"nose">>} = + uri_string:parse(<<"/over/there#nose">>), + #{host := <<"example.com">>, fragment := <<"nose">>} = + uri_string:parse(<<"//example.com#nose">>), + #{host := <<"example.com">>, path := <<"/">>, fragment := <<"nose">>} = + uri_string:parse(<<"//example.com/#nose">>). + +parse_binary_pct_encoded_fragment(_Config) -> + #{scheme := <<"foo">>, host := <<"example.com">>, fragment := <<"合気道"/utf8>>} = + uri_string:parse(<<"foo://example.com#%E5%90%88%E6%B0%97%E9%81%93">>), + #{host := <<"example.com">>, path := <<"/">>, fragment := <<"合気道"/utf8>>} = + uri_string:parse(<<"//example.com/#%E5%90%88%E6%B0%97%E9%81%93">>). + +parse_scheme(_Config) -> + #{} = uri_string:parse(""), + #{path := "foo"} = uri_string:parse("foo"), + #{scheme := "foo"} = uri_string:parse("foo:"), + #{scheme := "foo", path := "bar:nisse"} = uri_string:parse("foo:bar:nisse"), + #{scheme := "foo", host := ""} = uri_string:parse("foo://"), + #{scheme := "foo", host := "", path := "/"} = uri_string:parse("foo:///"), + #{scheme := "foo", host := "", path := "//"} = uri_string:parse("foo:////"), + + #{path := "/"} = uri_string:parse("/"), + #{host := ""} = uri_string:parse("//"), + #{host := "", path := "/"} = uri_string:parse("///"). + +parse_userinfo(_Config) -> + #{scheme := "user", path := "password@localhost"} = uri_string:parse("user:password@localhost"), + #{path := "user@"} = uri_string:parse("user@"), + #{path := "/user@"} = uri_string:parse("/user@"), + #{path := "user@localhost"} = uri_string:parse("user@localhost"), + #{userinfo := "user", host := "localhost"} = uri_string:parse("//user@localhost"), + #{userinfo := "user:password", host := "localhost"} = + uri_string:parse("//user:password@localhost"), + #{scheme := "foo", path := "/user@"} = + uri_string:parse("foo:/user@"), + #{scheme := "foo", userinfo := "user", host := "localhost"} = + uri_string:parse("foo://user@localhost"), + #{scheme := "foo", userinfo := "user:password", host := "localhost"} = + uri_string:parse("foo://user:password@localhost"). + +parse_pct_encoded_userinfo(_Config) -> + #{scheme := "user", path := "合@気道"} = + uri_string:parse("user:%E5%90%88@%E6%B0%97%E9%81%93"), + #{path := "合気道@"} = uri_string:parse("%E5%90%88%E6%B0%97%E9%81%93@"), + #{path := "/合気道@"} = uri_string:parse("/%E5%90%88%E6%B0%97%E9%81%93@"), + #{path := "合@気道"} = uri_string:parse("%E5%90%88@%E6%B0%97%E9%81%93"), + #{userinfo := "合", host := "気道"} = + uri_string:parse("//%E5%90%88@%E6%B0%97%E9%81%93"), + #{userinfo := "合:気", host := "道"} = + uri_string:parse("//%E5%90%88:%E6%B0%97@%E9%81%93"), + #{scheme := "foo", path := "/合気道@"} = + uri_string:parse("foo:/%E5%90%88%E6%B0%97%E9%81%93@"), + #{scheme := "foo", userinfo := "合", host := "気道"} = + uri_string:parse("foo://%E5%90%88@%E6%B0%97%E9%81%93"), + #{scheme := "foo", userinfo := "合:気", host := "道"} = + uri_string:parse("foo://%E5%90%88:%E6%B0%97@%E9%81%93"), + {error,invalid_uri,"@"} = uri_string:parse("//%E5%90%88@%E6%B0%97%E9%81%93@"), + {error,invalid_uri,":"} = uri_string:parse("foo://%E5%90%88@%E6%B0%97%E9%81%93@"). + + +parse_host(_Config) -> + #{host := "hostname"} = uri_string:parse("//hostname"), + #{host := "hostname",scheme := "foo"} = uri_string:parse("foo://hostname"), + #{host := "hostname",scheme := "foo", userinfo := "user"} = + uri_string:parse("foo://user@hostname"). + +parse_host_ipv4(_Config) -> + #{host := "127.0.0.1"} = uri_string:parse("//127.0.0.1"), + #{host := "2001:0db8:0000:0000:0000:0000:1428:07ab"} = + uri_string:parse("//[2001:0db8:0000:0000:0000:0000:1428:07ab]"), + #{host := "127.0.0.1", path := "/over/there"} = uri_string:parse("//127.0.0.1/over/there"), + #{host := "127.0.0.1", query := "name=ferret"} = uri_string:parse("//127.0.0.1?name=ferret"), + #{host := "127.0.0.1", fragment := "nose"} = uri_string:parse("//127.0.0.1#nose"), + {error,invalid_uri,"x"} = uri_string:parse("//127.0.0.x"), + {error,invalid_uri,"1227.0.0.1"} = uri_string:parse("//1227.0.0.1"). + +parse_host_ipv6(_Config) -> + #{host := "::127.0.0.1"} = uri_string:parse("//[::127.0.0.1]"), + #{host := "::127.0.0.1", path := "/over/there"} = uri_string:parse("//[::127.0.0.1]/over/there"), + #{host := "::127.0.0.1", query := "name=ferret"} = + uri_string:parse("//[::127.0.0.1]?name=ferret"), + #{host := "::127.0.0.1", fragment := "nose"} = uri_string:parse("//[::127.0.0.1]#nose"), + {error,invalid_uri,"x"} = uri_string:parse("//[::127.0.0.x]"), + {error,invalid_uri,"::1227.0.0.1"} = uri_string:parse("//[::1227.0.0.1]"), + {error,invalid_uri,"G"} = uri_string:parse("//[2001:0db8:0000:0000:0000:0000:1428:G7ab]"). + +parse_port(_Config) -> + #{path:= "/:8042"} = + uri_string:parse("/:8042"), + #{host:= "", port := 8042} = + uri_string:parse("//:8042"), + #{host := "example.com", port:= 8042} = + uri_string:parse("//example.com:8042"), + #{scheme := "foo", path := "/:8042"} = + uri_string:parse("foo:/:8042"), + #{scheme := "foo", host := "", port := 8042} = + uri_string:parse("foo://:8042"), + #{scheme := "foo", host := "example.com", port := 8042} = + uri_string:parse("foo://example.com:8042"). + +parse_path(_Config) -> + #{path := "over/there"} = uri_string:parse("over/there"), + #{path := "/over/there"} = uri_string:parse("/over/there"), + #{scheme := "foo", path := "/over/there"} = + uri_string:parse("foo:/over/there"), + #{scheme := "foo", host := "example.com", path := "/over/there"} = + uri_string:parse("foo://example.com/over/there"), + #{scheme := "foo", host := "example.com", path := "/over/there", port := 8042} = + uri_string:parse("foo://example.com:8042/over/there"). + +parse_query(_Config) -> + #{scheme := "foo", query := "name=ferret"} = + uri_string:parse("foo:?name=ferret"), + #{scheme := "foo", path:= "over/there", query := "name=ferret"} = + uri_string:parse("foo:over/there?name=ferret"), + #{scheme := "foo", path:= "/over/there", query := "name=ferret"} = + uri_string:parse("foo:/over/there?name=ferret"), + #{scheme := "foo", host := "example.com", query := "name=ferret"} = + uri_string:parse("foo://example.com?name=ferret"), + #{scheme := "foo", host := "example.com", path := "/", query := "name=ferret"} = + uri_string:parse("foo://example.com/?name=ferret"), + + #{path := "", query := "name=ferret"} = + uri_string:parse("?name=ferret"), + #{path := "over/there", query := "name=ferret"} = + uri_string:parse("over/there?name=ferret"), + #{path := "/", query := "name=ferret"} = + uri_string:parse("/?name=ferret"), + #{path := "/over/there", query := "name=ferret"} = + uri_string:parse("/over/there?name=ferret"), + #{host := "example.com", query := "name=ferret"} = + uri_string:parse("//example.com?name=ferret"), + #{host := "example.com", path := "/", query := "name=ferret"} = + uri_string:parse("//example.com/?name=ferret"). + +parse_pct_encoded_query(_Config) -> + #{scheme := "foo", host := "example.com", path := "/", + query := "name=合気道"} = + uri_string:parse("foo://example.com/?name=%E5%90%88%E6%B0%97%E9%81%93"), + #{host := "example.com", path := "/", query := "name=合気道"} = + uri_string:parse("//example.com/?name=%E5%90%88%E6%B0%97%E9%81%93"). + +parse_fragment(_Config) -> + #{scheme := "foo", fragment := "nose"} = + uri_string:parse("foo:#nose"), + #{scheme := "foo", path:= "over/there", fragment := "nose"} = + uri_string:parse("foo:over/there#nose"), + #{scheme := "foo", path:= "/over/there", fragment := "nose"} = + uri_string:parse("foo:/over/there#nose"), + #{scheme := "foo", host := "example.com", fragment := "nose"} = + uri_string:parse("foo://example.com#nose"), + #{scheme := "foo", host := "example.com", path := "/", fragment := "nose"} = + uri_string:parse("foo://example.com/#nose"), + #{scheme := "foo", host := "example.com", fragment := "nose"} = + uri_string:parse("foo://example.com#nose"), + + #{fragment := "nose"} = + uri_string:parse("#nose"), + #{path := "over/there", fragment := "nose"} = + uri_string:parse("over/there#nose"), + #{path := "/", fragment := "nose"} = + uri_string:parse("/#nose"), + #{path := "/over/there", fragment := "nose"} = + uri_string:parse("/over/there#nose"), + #{host := "example.com", fragment := "nose"} = + uri_string:parse("//example.com#nose"), + #{host := "example.com", path := "/", fragment := "nose"} = + uri_string:parse("//example.com/#nose"). + +parse_pct_encoded_fragment(_Config) -> + #{scheme := "foo", host := "example.com", fragment := "合気道"} = + uri_string:parse("foo://example.com#%E5%90%88%E6%B0%97%E9%81%93"), + #{host := "example.com", path := "/", fragment := "合気道"} = + uri_string:parse("//example.com/#%E5%90%88%E6%B0%97%E9%81%93"). + +parse_list(_Config) -> + #{scheme := "foo", path := "bar:nisse"} = uri_string:parse("foo:bar:nisse"), + #{scheme := "foo", host := "example.com", port := 8042, + path := "/over/there", query := "name=ferret", fragment := "nose"} = + uri_string:parse("foo://example.com:8042/over/there?name=ferret#nose"), + #{scheme := "foo", userinfo := "admin:admin", host := "example.com", port := 8042, + path := "/over/there", query := "name=ferret", fragment := "nose"} = + uri_string:parse("foo://admin:[email protected]:8042/over/there?name=ferret#nose"). + +parse_binary(_Config) -> + #{scheme := <<"foo">>, path := <<"bar:nisse">>} = uri_string:parse(<<"foo:bar:nisse">>), + #{scheme := <<"foo">>, host := <<"example.com">>, port := 8042, + path := <<"/over/there">>, query := <<"name=ferret">>, fragment := <<"nose">>} = + uri_string:parse(<<"foo://example.com:8042/over/there?name=ferret#nose">>), + #{scheme := <<"foo">>, userinfo := <<"admin:admin">>, host := <<"example.com">>, port := 8042, + path := <<"/over/there">>, query := <<"name=ferret">>, fragment := <<"nose">>} = + uri_string:parse(<<"foo://admin:[email protected]:8042/over/there?name=ferret#nose">>). + + +parse_mixed(_Config) -> + #{scheme := "foo", path := "bar"} = + uri_string:parse(lists:append("fo",<<"o:bar">>)), + #{scheme := "foo", path := "bar"} = + uri_string:parse(lists:append("foo:b",<<"ar">>)), + #{scheme := "foo", path := "bar:bar"} = + uri_string:parse([[102],[111,111],<<":bar">>,58,98,97,114]). + +parse_relative(_Config) -> + #{path := "/path"} = + uri_string:parse(lists:append("/pa",<<"th">>)), + #{path := "foo"} = + uri_string:parse(lists:append("fo",<<"o">>)). + +parse_special(_Config) -> + #{host := [],query := []} = uri_string:parse("//?"), + #{fragment := [],host := []} = uri_string:parse("//#"), + #{host := [],query := [],scheme := "foo"} = uri_string:parse("foo://?"), + #{fragment := [],host := [],scheme := "foo"} = uri_string:parse("foo://#"), + #{host := <<>>, path := <<"/">>} = uri_string:parse(<<"///">>), + #{host := <<"hostname">>} = uri_string:parse(<<"//hostname">>), + #{host := <<>>, path := <<"/hostname">>} = uri_string:parse(<<"///hostname">>), + #{host := [],path := "/",query := []} = uri_string:parse("///?"), + #{fragment := [],host := [],path := "/"} = uri_string:parse("///#"), + #{host := "foo",query := []} = uri_string:parse("//foo?"), + #{fragment := [],host := "foo"} = uri_string:parse("//foo#"), + #{host := "foo",path := "/"} = uri_string:parse("//foo/"), + #{host := "foo",query := [],scheme := "http"} = uri_string:parse("http://foo?"), + #{fragment := [],host := "foo",scheme := "http"} = uri_string:parse("http://foo#"), + #{host := "foo",path := "/",scheme := "http"} = uri_string:parse("http://foo/"), + #{fragment := [],host := "host",port := 80,scheme := "http"} = uri_string:parse("http://host:80#"), + #{host := "host",port := 80,query := [],scheme := "http"} = uri_string:parse("http://host:80?"), + #{path := [],query := []} = uri_string:parse("?"), + #{path := [],query := "?"} = uri_string:parse("??"), + #{path := [],query := "??"} = uri_string:parse("???"). + +parse_special2(_Config) -> + #{host := [],path := "/",port := 1,scheme := "a"} = uri_string:parse("a://:1/"), + #{path := "/a/",scheme := "a"} = uri_string:parse("a:/a/"), + #{host := [],path := [],userinfo := []} = uri_string:parse("//@"), + #{host := [],path := [],scheme := "foo",userinfo := []} = uri_string:parse("foo://@"), + #{host := [],path := "/",userinfo := []} = uri_string:parse("//@/"), + #{host := [],path := "/",scheme := "foo",userinfo := []} = uri_string:parse("foo://@/"), + #{host := "localhost",path := "/",port := undefined} = uri_string:parse("//localhost:/"), + #{host := [],path := [],port := undefined} = uri_string:parse("//:"). + +parse_negative(_Config) -> + {error,invalid_uri,"å"} = uri_string:parse("å"), + {error,invalid_uri,"å"} = uri_string:parse("aå:/foo"), + {error,invalid_uri,":"} = uri_string:parse("foo://usär@host"), + {error,invalid_uri,"ö"} = uri_string:parse("//host/path?foö=bar"), + {error,invalid_uri,"ö"} = uri_string:parse("//host/path#foö"), + {error,invalid_uri,"127.256.0.1"} = uri_string:parse("//127.256.0.1"), + {error,invalid_uri,":::127.0.0.1"} = uri_string:parse("//[:::127.0.0.1]"), + {error,invalid_utf8,<<0,0,0,246>>} = uri_string:parse("//%00%00%00%F6"), + {error,invalid_uri,"A"} = uri_string:parse("//localhost:A8"). + + +%%------------------------------------------------------------------------- +%% Recompose tests +%%------------------------------------------------------------------------- +recompose_fragment(_Config) -> + <<?FRAGMENT_ENC>> = uri_string:recompose(#{fragment => <<?FRAGMENT/utf8>>, path => <<>>}), + ?FRAGMENT_ENC = uri_string:recompose(#{fragment => ?FRAGMENT, path => ""}). + +recompose_parse_fragment(_Config) -> + <<?FRAGMENT_ENC>> = uri_string:recompose(uri_string:parse(<<?FRAGMENT_ENC>>)), + ?FRAGMENT_ENC = uri_string:recompose(uri_string:parse(?FRAGMENT_ENC)). + +recompose_query(_Config) -> + <<?QUERY_ENC>> = + uri_string:recompose(#{query => <<?QUERY/utf8>>, path => <<>>}), + <<?QUERY_ENC?FRAGMENT_ENC>> = + uri_string:recompose(#{query => <<?QUERY/utf8>>, + fragment => <<?FRAGMENT/utf8>>, + path => <<>>}), + "?name=%C3%B6rn" = + uri_string:recompose(#{query => "name=örn", path => ""}), + "?name=%C3%B6rn#n%C3%A4sa" = + uri_string:recompose(#{query => "name=örn", + fragment => "näsa", + path => ""}). + +recompose_parse_query(_Config) -> + <<"?name=%C3%B6rn">> = uri_string:recompose(uri_string:parse(<<"?name=%C3%B6rn">>)), + <<"?name=%C3%B6rn#n%C3%A4sa">> = + uri_string:recompose(uri_string:parse(<<"?name=%C3%B6rn#n%C3%A4sa">>)), + "?name=%C3%B6rn" = uri_string:recompose(uri_string:parse("?name=%C3%B6rn")), + "?name=%C3%B6rn#n%C3%A4sa" = uri_string:recompose(uri_string:parse("?name=%C3%B6rn#n%C3%A4sa")). + +recompose_path(_Config) -> + <<"/d%C3%A4r">> = + uri_string:recompose(#{path => <<"/där"/utf8>>}), + <<"/d%C3%A4r#n%C3%A4sa">> = + uri_string:recompose(#{path => <<"/där"/utf8>>, + fragment => <<"näsa"/utf8>>}), + <<"/d%C3%A4r?name=%C3%B6rn">> = + uri_string:recompose(#{path => <<"/där"/utf8>>, + query => <<"name=örn"/utf8>>}), + <<"/d%C3%A4r?name=%C3%B6rn#n%C3%A4sa">> = + uri_string:recompose(#{path => <<"/där"/utf8>>, + query => <<"name=örn"/utf8>>, + fragment => <<"näsa"/utf8>>}), + + + "/d%C3%A4r" = + uri_string:recompose(#{path => "/där"}), + "/d%C3%A4r#n%C3%A4sa" = + uri_string:recompose(#{path => "/där", + fragment => "näsa"}), + "/d%C3%A4r?name=%C3%B6rn" = + uri_string:recompose(#{path => "/där", + query => "name=örn"}), + "/d%C3%A4r?name=%C3%B6rn#n%C3%A4sa" = + uri_string:recompose(#{path => "/där", + query => "name=örn", + fragment => "näsa"}). + + +recompose_parse_path(_Config) -> + <<"/d%C3%A4r">> = + uri_string:recompose(uri_string:parse(<<"/d%C3%A4r">>)), + <<"/d%C3%A4r#n%C3%A4sa">> = + uri_string:recompose(uri_string:parse(<<"/d%C3%A4r#n%C3%A4sa">>)), + <<"/d%C3%A4r?name=%C3%B6rn">> = + uri_string:recompose(uri_string:parse(<<"/d%C3%A4r?name=%C3%B6rn">>)), + + "/d%C3%A4r" = + uri_string:recompose(uri_string:parse("/d%C3%A4r")), + "/d%C3%A4r#n%C3%A4sa" = + uri_string:recompose(uri_string:parse("/d%C3%A4r#n%C3%A4sa")), + "/d%C3%A4r?name=%C3%B6rn" = + uri_string:recompose(uri_string:parse("/d%C3%A4r?name=%C3%B6rn")). + + +recompose_autogen(_Config) -> + Tests = generate_test_vectors(uri_combinations()), + lists:map(fun run_test_recompose/1, Tests). + +parse_recompose_autogen(_Config) -> + Tests = generate_test_vectors(uri_combinations()), + lists:map(fun run_test_parse_recompose/1, Tests). + +transcode_basic(_Config) -> + <<"foo%C3%B6bar"/utf8>> = + uri_string:transcode(<<"foo%00%00%00%F6bar"/utf32>>, [{in_encoding, utf32},{out_encoding, utf8}]), + "foo%C3%B6bar" = + uri_string:transcode("foo%00%00%00%F6bar", [{in_encoding, utf32},{out_encoding, utf8}]), + <<"foo%00%00%00%F6bar"/utf32>> = + uri_string:transcode(<<"foo%C3%B6bar"/utf8>>, [{in_encoding, utf8},{out_encoding, utf32}]), + "foo%00%00%00%F6bar" = + uri_string:transcode("foo%C3%B6bar", [{in_encoding, utf8},{out_encoding, utf32}]), + "foo%C3%B6bar" = + uri_string:transcode("foo%F6bar", [{in_encoding, latin1},{out_encoding, utf8}]). + +transcode_options(_Config) -> + <<"foo%C3%B6bar"/utf8>> = + uri_string:transcode(<<"foo%C3%B6bar"/utf8>>, []), + <<"foo%C3%B6bar"/utf8>> = + uri_string:transcode(<<"foo%00%00%00%F6bar"/utf32>>, [{in_encoding, utf32}]), + <<"foo%00%00%00%F6bar"/utf32>> = + uri_string:transcode(<<"foo%C3%B6bar"/utf8>>, [{out_encoding, utf32}]). + +transcode_mixed(_Config) -> + "foo%00%00%00%F6bar" = + uri_string:transcode(["foo",<<"%C3%B6"/utf8>>,<<"ba"/utf8>>,"r"], [{out_encoding, utf32}]), + "foo%00%00%00%F6bar" = + uri_string:transcode(["foo",<<"%C3%"/utf8>>,<<"B6ba"/utf8>>,"r"], [{out_encoding, utf32}]), + "foo%C3%B6bar" = + uri_string:transcode(["foo%00", <<"%00%0"/utf32>>,<<"0%F"/utf32>>,"6bar"], [{in_encoding, utf32},{out_encoding, utf8}]). + +transcode_negative(_Config) -> + {error,invalid_percent_encoding,"%BXbar"} = + uri_string:transcode(<<"foo%C3%BXbar"/utf8>>, [{in_encoding, utf8},{out_encoding, utf32}]), + {error,invalid_input,<<"ö">>} = + uri_string:transcode("foo%F6bar", [{in_encoding, utf8},{out_encoding, utf8}]). + +normalize(_Config) -> + "/a/g" = uri_string:normalize("/a/b/c/./../../g"), + <<"mid/6">> = uri_string:normalize(<<"mid/content=5/../6">>), + "http://localhost-%C3%B6rebro/a/g" = + uri_string:normalize("http://localhos%74-%c3%b6rebro:80/a/b/c/./../../g"), + <<"http://localhost-%C3%B6rebro/a/g">> = + uri_string:normalize(<<"http://localhos%74-%c3%b6rebro:80/a/b/c/./../../g">>), + <<"https://localhost/">> = + uri_string:normalize(<<"https://localhost:443">>), + <<"https://localhost:445/">> = + uri_string:normalize(<<"https://localhost:445">>), + <<"ftp://localhost">> = + uri_string:normalize(<<"ftp://localhost:21">>), + <<"ssh://localhost">> = + uri_string:normalize(<<"ssh://localhost:22">>), + <<"sftp://localhost">> = + uri_string:normalize(<<"sftp://localhost:22">>), + <<"tftp://localhost">> = + uri_string:normalize(<<"tftp://localhost:69">>). diff --git a/lib/stdlib/test/uri_string_property_test_SUITE.erl b/lib/stdlib/test/uri_string_property_test_SUITE.erl new file mode 100644 index 0000000000..ae2c61c7aa --- /dev/null +++ b/lib/stdlib/test/uri_string_property_test_SUITE.erl @@ -0,0 +1,39 @@ +%% +%% %CopyrightBegin% +%% +%% Copyright Ericsson AB 2008-2017. All Rights Reserved. +%% +%% Licensed under the Apache License, Version 2.0 (the "License"); +%% you may not use this file except in compliance with the License. +%% You may obtain a copy of the License at +%% +%% http://www.apache.org/licenses/LICENSE-2.0 +%% +%% Unless required by applicable law or agreed to in writing, software +%% distributed under the License is distributed on an "AS IS" BASIS, +%% WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +%% See the License for the specific language governing permissions and +%% limitations under the License. +%% +%% %CopyrightEnd% +%% +-module(uri_string_property_test_SUITE). + +-include_lib("common_test/include/ct.hrl"). +-compile(export_all). + +all() -> [recompose]. + +init_per_suite(Config) -> + ct_property_test:init_per_suite(Config). + +end_per_suite(Config) -> + Config. + +%%%======================================================================== +%%% Test suites +%%%======================================================================== +recompose(Config) -> + ct_property_test:quickcheck( + uri_string_recompose:prop_recompose(), + Config). diff --git a/lib/tools/test/xref_SUITE.erl b/lib/tools/test/xref_SUITE.erl index d651cbcfee..018f632948 100644 --- a/lib/tools/test/xref_SUITE.erl +++ b/lib/tools/test/xref_SUITE.erl @@ -1112,27 +1112,16 @@ read_expected(Version) -> {POS7+2,{FF,{erlang,spawn_opt,4}}}, {POS8+1,{FF,{hej,san,1}}}, {POS8+4,{FF,{a,b,1}}}, - {POS8+4,{FF,{erlang,apply,2}}}, - {POS8+5,{FF,{erlang,apply,2}}}, {POS8+6,{FF,{m,f,1}}}, {POS9+1,{FF,{read,bi,0}}}, {POS9+2,{FF,{a,b,1}}}, - {POS9+2,{FF,{erlang,apply,2}}}, - {POS9+3,{FF,{erlang,apply,2}}}, - {POS9+4,{FF,{erlang,apply,2}}}, {POS9+4,{FF,{erlang,not_a_function,1}}}, {POS9+5,{FF,{mod,func,2}}}, {POS9+6,{FF,{erlang,apply,1}}}, - {POS9+7,{FF,{erlang,apply,2}}}, {POS9+7,{FF,{math,add3,1}}}, {POS9+8,{FF,{q,f,1}}}, - {POS10+4,{FF,{erlang,apply,2}}}, {POS10+5,{FF,{mod1,fun1,1}}}, - {POS11+6,{FF,{erlang,apply,2}}}, - {POS12+1,{FF,{erlang,apply,2}}}, - {POS12+4,{FF,{erlang,apply,2}}}, {POS12+5,{FF,{m3,f3,2}}}, - {POS12+7,{FF,{erlang,apply,2}}}, {POS13+1,{FF,{dm,df,1}}}, {POS13+6,{{read,bi,0},{foo,module_info,0}}}, {POS13+7,{{read,bi,0},{read,module_info,0}}}, @@ -1162,15 +1151,26 @@ read_expected(Version) -> {POS3+3, {FF,{erlang,spawn_link,3}}}, {POS3+4, {FF,{erlang,spawn_link,3}}}, {POS6+4, {FF,{erlang,spawn,3}}}, + {POS8+4,{FF,{erlang,apply,2}}}, + {POS8+5,{FF,{erlang,apply,2}}}, {POS8+6,{FF,{erlang,apply,3}}}, {POS8+7,{FF,{erlang,apply,3}}}, {POS9+1,{FF,{erlang,apply,3}}}, + {POS9+2,{FF,{erlang,apply,2}}}, + {POS9+3,{FF,{erlang,apply,2}}}, + {POS9+4,{FF,{erlang,apply,2}}}, {POS9+5,{FF,{erlang,apply,3}}}, + {POS9+7,{FF,{erlang,apply,2}}}, + {POS10+4,{FF,{erlang,apply,2}}}, {POS11+1,{FF,{erlang,apply,3}}}, {POS11+2,{FF,{erlang,apply,3}}}, {POS11+3,{FF,{erlang,apply,3}}}, {POS11+4,{FF,{erlang,apply,3}}}, + {POS11+6,{FF,{erlang,apply,2}}}, + {POS12+1,{FF,{erlang,apply,2}}}, + {POS12+4,{FF,{erlang,apply,2}}}, {POS12+5,{FF,{erlang,apply,3}}}, + {POS12+7,{FF,{erlang,apply,2}}}, {POS12+8,{FF,{erlang,apply,3}}}, {POS13+5, {{read,bi,0},{erlang,length,1}}}, {POS14+3, {{read,bi,0},{erlang,length,1}}}], diff --git a/lib/wx/c_src/wxe_driver.c b/lib/wx/c_src/wxe_driver.c index 5da71818e5..26ae3564e7 100644 --- a/lib/wx/c_src/wxe_driver.c +++ b/lib/wx/c_src/wxe_driver.c @@ -33,6 +33,7 @@ #include <sys/types.h> #include <fcntl.h> +#include <assert.h> #include "wxe_driver.h" #define TEMP_BINARY_SIZE 512 @@ -234,13 +235,21 @@ standard_outputv(ErlDrvData drv_data, ErlIOVec* ev) sd->max_bins = max; } - if(ev->vsize == 2) { - binref->base = ev->iov[1].iov_base; - binref->size = ev->iov[1].iov_len; - binref->from = driver_caller(sd->port_handle); - bin = ev->binv[1]; - driver_binary_inc_refc(bin); /* Otherwise it could get deallocated */ - binref->bin = bin; + if(ev->size > 0) { + assert(ev->vsize == 2 && ev->iov[0].iov_len == 0 + && "erts changed how the ErlIOVec is structured for outputv"); + binref->from = driver_caller(sd->port_handle); + binref->size = ev->iov[1].iov_len; + if(ev->binv[1]) { + binref->base = ev->iov[1].iov_base; + bin = ev->binv[1]; + driver_binary_inc_refc(bin); /* Otherwise it could get deallocated */ + } else { + bin = driver_alloc_binary(ev->iov[1].iov_len); + memcpy(bin->orig_bytes, ev->iov[1].iov_base, ev->iov[1].iov_len); + binref->base = bin->orig_bytes; + } + binref->bin = bin; } else { /* Empty binary (becomes NULL) */ binref->base = NULL; binref->size = 0; diff --git a/otp_versions.table b/otp_versions.table index 46c160e63a..157ca8f0ba 100644 --- a/otp_versions.table +++ b/otp_versions.table @@ -1,3 +1,6 @@ +OTP-20.1.5 : erts-9.1.4 inets-6.4.4 # asn1-5.0.3 common_test-1.15.2 compiler-7.1.3 cosEvent-2.2.1 cosEventDomain-1.2.1 cosFileTransfer-1.2.1 cosNotification-1.2.2 cosProperty-1.2.2 cosTime-1.2.2 cosTransactions-1.3.2 crypto-4.1 debugger-4.2.3 dialyzer-3.2.2 diameter-2.1.2 edoc-0.9.1 eldap-1.2.2 erl_docgen-0.7.1 erl_interface-3.10 et-1.6.1 eunit-2.3.4 hipe-3.16.1 ic-4.4.2 jinterface-1.8 kernel-5.4 megaco-3.18.2 mnesia-4.15.1 observer-2.5 odbc-2.12 orber-3.8.3 os_mon-2.4.3 otp_mibs-1.1.1 parsetools-2.1.5 public_key-1.5 reltool-0.7.5 runtime_tools-1.12.2 sasl-3.1 snmp-5.2.8 ssh-4.6.1 ssl-8.2.1 stdlib-3.4.2 syntax_tools-2.1.3 tools-2.11 wx-1.8.2 xmerl-1.3.15 : +OTP-20.1.4 : inets-6.4.3 # asn1-5.0.3 common_test-1.15.2 compiler-7.1.3 cosEvent-2.2.1 cosEventDomain-1.2.1 cosFileTransfer-1.2.1 cosNotification-1.2.2 cosProperty-1.2.2 cosTime-1.2.2 cosTransactions-1.3.2 crypto-4.1 debugger-4.2.3 dialyzer-3.2.2 diameter-2.1.2 edoc-0.9.1 eldap-1.2.2 erl_docgen-0.7.1 erl_interface-3.10 erts-9.1.3 et-1.6.1 eunit-2.3.4 hipe-3.16.1 ic-4.4.2 jinterface-1.8 kernel-5.4 megaco-3.18.2 mnesia-4.15.1 observer-2.5 odbc-2.12 orber-3.8.3 os_mon-2.4.3 otp_mibs-1.1.1 parsetools-2.1.5 public_key-1.5 reltool-0.7.5 runtime_tools-1.12.2 sasl-3.1 snmp-5.2.8 ssh-4.6.1 ssl-8.2.1 stdlib-3.4.2 syntax_tools-2.1.3 tools-2.11 wx-1.8.2 xmerl-1.3.15 : +OTP-20.1.3 : diameter-2.1.2 erts-9.1.3 snmp-5.2.8 # asn1-5.0.3 common_test-1.15.2 compiler-7.1.3 cosEvent-2.2.1 cosEventDomain-1.2.1 cosFileTransfer-1.2.1 cosNotification-1.2.2 cosProperty-1.2.2 cosTime-1.2.2 cosTransactions-1.3.2 crypto-4.1 debugger-4.2.3 dialyzer-3.2.2 edoc-0.9.1 eldap-1.2.2 erl_docgen-0.7.1 erl_interface-3.10 et-1.6.1 eunit-2.3.4 hipe-3.16.1 ic-4.4.2 inets-6.4.2 jinterface-1.8 kernel-5.4 megaco-3.18.2 mnesia-4.15.1 observer-2.5 odbc-2.12 orber-3.8.3 os_mon-2.4.3 otp_mibs-1.1.1 parsetools-2.1.5 public_key-1.5 reltool-0.7.5 runtime_tools-1.12.2 sasl-3.1 ssh-4.6.1 ssl-8.2.1 stdlib-3.4.2 syntax_tools-2.1.3 tools-2.11 wx-1.8.2 xmerl-1.3.15 : OTP-20.1.2 : diameter-2.1.1 erts-9.1.2 # asn1-5.0.3 common_test-1.15.2 compiler-7.1.3 cosEvent-2.2.1 cosEventDomain-1.2.1 cosFileTransfer-1.2.1 cosNotification-1.2.2 cosProperty-1.2.2 cosTime-1.2.2 cosTransactions-1.3.2 crypto-4.1 debugger-4.2.3 dialyzer-3.2.2 edoc-0.9.1 eldap-1.2.2 erl_docgen-0.7.1 erl_interface-3.10 et-1.6.1 eunit-2.3.4 hipe-3.16.1 ic-4.4.2 inets-6.4.2 jinterface-1.8 kernel-5.4 megaco-3.18.2 mnesia-4.15.1 observer-2.5 odbc-2.12 orber-3.8.3 os_mon-2.4.3 otp_mibs-1.1.1 parsetools-2.1.5 public_key-1.5 reltool-0.7.5 runtime_tools-1.12.2 sasl-3.1 snmp-5.2.7 ssh-4.6.1 ssl-8.2.1 stdlib-3.4.2 syntax_tools-2.1.3 tools-2.11 wx-1.8.2 xmerl-1.3.15 : OTP-20.1.1 : compiler-7.1.3 erts-9.1.1 ssh-4.6.1 # asn1-5.0.3 common_test-1.15.2 cosEvent-2.2.1 cosEventDomain-1.2.1 cosFileTransfer-1.2.1 cosNotification-1.2.2 cosProperty-1.2.2 cosTime-1.2.2 cosTransactions-1.3.2 crypto-4.1 debugger-4.2.3 dialyzer-3.2.2 diameter-2.1 edoc-0.9.1 eldap-1.2.2 erl_docgen-0.7.1 erl_interface-3.10 et-1.6.1 eunit-2.3.4 hipe-3.16.1 ic-4.4.2 inets-6.4.2 jinterface-1.8 kernel-5.4 megaco-3.18.2 mnesia-4.15.1 observer-2.5 odbc-2.12 orber-3.8.3 os_mon-2.4.3 otp_mibs-1.1.1 parsetools-2.1.5 public_key-1.5 reltool-0.7.5 runtime_tools-1.12.2 sasl-3.1 snmp-5.2.7 ssl-8.2.1 stdlib-3.4.2 syntax_tools-2.1.3 tools-2.11 wx-1.8.2 xmerl-1.3.15 : OTP-20.1 : asn1-5.0.3 common_test-1.15.2 compiler-7.1.2 crypto-4.1 debugger-4.2.3 dialyzer-3.2.2 diameter-2.1 edoc-0.9.1 erl_docgen-0.7.1 erts-9.1 et-1.6.1 eunit-2.3.4 hipe-3.16.1 inets-6.4.2 kernel-5.4 mnesia-4.15.1 observer-2.5 os_mon-2.4.3 public_key-1.5 reltool-0.7.5 runtime_tools-1.12.2 sasl-3.1 snmp-5.2.7 ssh-4.6 ssl-8.2.1 stdlib-3.4.2 syntax_tools-2.1.3 tools-2.11 wx-1.8.2 # cosEvent-2.2.1 cosEventDomain-1.2.1 cosFileTransfer-1.2.1 cosNotification-1.2.2 cosProperty-1.2.2 cosTime-1.2.2 cosTransactions-1.3.2 eldap-1.2.2 erl_interface-3.10 ic-4.4.2 jinterface-1.8 megaco-3.18.2 odbc-2.12 orber-3.8.3 otp_mibs-1.1.1 parsetools-2.1.5 xmerl-1.3.15 : diff --git a/system/doc/design_principles/gen_server_concepts.xml b/system/doc/design_principles/gen_server_concepts.xml index c1b98189d5..3a1f876646 100644 --- a/system/doc/design_principles/gen_server_concepts.xml +++ b/system/doc/design_principles/gen_server_concepts.xml @@ -32,7 +32,7 @@ <marker id="gen_server"></marker> <p>This section is to be read with the <seealso marker="stdlib:gen_server">gen_server(3)</seealso> - manual page in <c>stdblib</c>, where all interface functions and + manual page in <c>stdlib</c>, where all interface functions and callback functions are described in detail.</p> <section> diff --git a/system/doc/oam/oam_intro.xml b/system/doc/oam/oam_intro.xml index d3867f03ca..ead8c026b9 100644 --- a/system/doc/oam/oam_intro.xml +++ b/system/doc/oam/oam_intro.xml @@ -211,7 +211,7 @@ snmp:c("MY-MIB", [{il, ["sasl/priv/mibs"]}]).</code> <p>The following MIBs are defined in the OTP system:</p> <list type="bulleted"> - <item><p><c>OTP-REG)</c> (in SASL) contains the top-level + <item><p><c>OTP-REG</c> (in SASL) contains the top-level OTP registration objects, used by all other MIBs.</p></item> <item><p><c>OTP-TC</c> (in SASL) contains the general Textual Conventions, which can be used by any other MIB.</p></item> |
