diff options
Diffstat (limited to 'erts/doc')
| -rw-r--r-- | erts/doc/src/crash_dump.xml | 31 | ||||
| -rw-r--r-- | erts/doc/src/epmd.xml | 41 | ||||
| -rw-r--r-- | erts/doc/src/erl.xml | 20 | ||||
| -rw-r--r-- | erts/doc/src/erl_driver.xml | 5 | ||||
| -rw-r--r-- | erts/doc/src/erl_nif.xml | 158 | ||||
| -rw-r--r-- | erts/doc/src/erl_prim_loader.xml | 16 | ||||
| -rw-r--r-- | erts/doc/src/erlang.xml | 9 | ||||
| -rw-r--r-- | erts/doc/src/erts_alloc.xml | 14 | ||||
| -rw-r--r-- | erts/doc/src/notes.xml | 387 |
9 files changed, 534 insertions, 147 deletions
diff --git a/erts/doc/src/crash_dump.xml b/erts/doc/src/crash_dump.xml index c59741f250..d3de29b876 100644 --- a/erts/doc/src/crash_dump.xml +++ b/erts/doc/src/crash_dump.xml @@ -85,20 +85,22 @@ operating system.</p> <list type="bulleted"> <item>"<em><A></em>: Cannot allocate <em><N></em> - bytes of memory (of type "<em><T></em>")." - The system - has run out of memory. <A> is the allocator that failed - to allocate memory, <N> is the number of bytes that - <A> tried to allocate, and <T> is the memory block - type that the memory was needed for. The most common case is - that a process stores huge amounts of data. In this case - <T> is most often <c><![CDATA[heap]]></c>, <c><![CDATA[old_heap]]></c>, - <c><![CDATA[heap_frag]]></c>, or <c><![CDATA[binary]]></c>. For more information on - allocators see - <seealso marker="erts_alloc">erts_alloc(3)</seealso>.</item> + bytes of memory (of type "<em><T></em>", thread + <em><I></em>em>)." - The system has run out of memory. <A> + is the allocator that failed to allocate memory, <N> is the + number of bytes that <A> tried to allocate, <T> is the + memory block type that the memory was needed for, and <I> is the + thread identifier. The most common case is that a process stores huge + amounts of data. In this case <T> is most often + <c><![CDATA[heap]]></c>, <c><![CDATA[old_heap]]></c>, + <c><![CDATA[heap_frag]]></c>, or <c><![CDATA[binary]]></c>. + For more information on allocators see + <seealso marker="erts_alloc">erts_alloc(3)</seealso>.</item> <item>"<em><A></em>: Cannot reallocate <em><N></em> - bytes of memory (of type "<em><T></em>")." - Same as - above with the exception that memory was being reallocated - instead of being allocated when the system ran out of memory.</item> + bytes of memory (of type "<em><T></em>", thread + <em><I></em>em>)." - Same as above with the exception that memory + was being reallocated instead of being allocated when the system ran + out of memory.</item> <item>"Unexpected op code <em>N</em>" - Error in compiled code, <c><![CDATA[beam]]></c> file damaged or error in the compiler.</item> <item>"Module <em>Name</em> undefined" <c><![CDATA[|]]></c> "Function @@ -246,6 +248,9 @@ <tag><em>Last scheduled in for | Current call</em></tag> <item>The current function of the process. These fields will not always exist.</item> + <tag><em>Run queue</em></tag> + <item>The identifier of the scheduler run queue in which the process is + running.</item> <tag><em>Spawned by</em></tag> <item>The parent of the process, i.e. the process which executed <c><![CDATA[spawn]]></c> or <c><![CDATA[spawn_link]]></c>.</item> diff --git a/erts/doc/src/epmd.xml b/erts/doc/src/epmd.xml index 963d35c3c8..25f819ab50 100644 --- a/erts/doc/src/epmd.xml +++ b/erts/doc/src/epmd.xml @@ -58,12 +58,12 @@ of the IP address and a port number. The name of the node is an atom on the form of <c><![CDATA[Name@Node]]></c>. The job of the <c><![CDATA[epmd]]></c> daemon is to keep track of which - node name listens on which address. Hence, <c><![CDATA[epmd]]></c> map + node name listens on which address. Hence, <c><![CDATA[epmd]]></c> maps symbolic node names to machine addresses.</p> <p>The TCP/IP <c>epmd</c> daemon actually only keeps track of - the <c>Name</c> (first) part of an Erlang node name, the <c>Host</c> - part (whatever is after the <c><![CDATA[@]]></c> is implicit in the + the <c>Name</c> (first) part of an Erlang node name. The <c>Host</c> + part (whatever is after the <c><![CDATA[@]]></c>) is implicit in the node name where the <c>epmd</c> daemon was actually contacted, as is the IP address where the Erlang node can be reached. Consistent and correct TCP naming services are @@ -77,12 +77,12 @@ <p>The daemon is started automatically by the <c>erl</c> command if the node is to be distributed and there is no running instance present. If automatically launched, - environment variables has to be used to alter the behavior of + environment variables have to be used to alter the behavior of the daemon. See the <seealso marker="#environment_variables">Environment variables</seealso> section below.</p> - <p>If the -daemon argument is not given, the + <p>If the -daemon argument is not given, <c><![CDATA[epmd]]></c> runs as a normal program with the controlling terminal of the shell in which it is started. Normally, it should run as a daemon.</p> @@ -122,7 +122,7 @@ comma-separated list of IP addresses and on the loopback address (which is implicitly added to the list if it has not been specified). This can also be set using the - <c><![CDATA[ERL_EPMD_ADDRESS]]></c> environment variable, see the + <c><![CDATA[ERL_EPMD_ADDRESS]]></c> environment variable. See the section <seealso marker="#environment_variables">Environment variables</seealso> below.</p> </item> @@ -130,7 +130,7 @@ <item> <p>Let this instance of epmd listen to another TCP port than default 4369. This can also be set using the - <c><![CDATA[ERL_EPMD_PORT]]></c> environment variable, see the + <c><![CDATA[ERL_EPMD_PORT]]></c> environment variable. See the section <seealso marker="#environment_variables">Environment variables</seealso> below</p> </item> @@ -153,7 +153,7 @@ <p>With relaxed command checking, the <c>epmd</c> daemon can be killed from the localhost with i.e. <c>epmd -kill</c> even if there are active nodes registered. Normally only daemons with an empty node database can be killed with the <c>epmd -kill</c> command.</p> </item> <item> - <p>The <c>epmd -stop</c> command (and the corresponding messages to epmd, as can be given using <c>erl_interface/ei</c>) is normally always ignored, as it opens up for strange situation when two nodes of the same name can be alive at the same time. A node unregisters itself by just closing the connection to epmd, why the <c>stop</c> command was only intended for use in debugging situations.</p> + <p>The <c>epmd -stop</c> command (and the corresponding messages to epmd, as can be given using <c>erl_interface/ei</c>) is normally always ignored, as it opens up the possibility of a strange situation where two nodes of the same name can be alive at the same time. A node unregisters itself by just closing the connection to epmd, which is why the <c>stop</c> command was only intended for use in debugging situations.</p> <p>With relaxed command checking enabled, you can forcibly unregister live nodes.</p> </item> </list> @@ -166,7 +166,7 @@ <section> <marker id="debug_flags"></marker> <title>DbgExtra options</title> - <p>These options are purely for debugging and testing epmd clients, they should not be used in normal operation.</p> + <p>These options are purely for debugging and testing epmd clients. They should not be used in normal operation.</p> <taglist> <tag><c><![CDATA[-packet_timeout Seconds]]></c></tag> @@ -177,9 +177,9 @@ </item> <tag><c><![CDATA[-delay_accept Seconds]]></c></tag> <item> - <p>To simulate a busy server you can insert a delay between epmd - gets notified about that a new connection is requested and - when the connections gets accepted.</p> + <p>To simulate a busy server you can insert a delay between when epmd + gets notified that a new connection is requested and + when the connection gets accepted.</p> </item> <tag><c><![CDATA[-delay_write Seconds]]></c></tag> <item> @@ -191,15 +191,15 @@ <section> <marker id="interactive_flags"></marker> <title>Interactive options</title> - <p>These options make <c>epmd</c> run as an interactive command displaying the results of sending queries ta an already running instance of <c>epmd</c>. The epmd contacted is always on the local node, but the <c>-port</c> option can be used to select between instances if several are running using different port on the host.</p> + <p>These options make <c>epmd</c> run as an interactive command, displaying the results of sending queries to an already running instance of <c>epmd</c>. The epmd contacted is always on the local node, but the <c>-port</c> option can be used to select between instances if several are running using different ports on the host.</p> <taglist> <tag><c><![CDATA[-port No]]></c></tag> <item> <p>Contacts the <c>epmd</c> listening on the given TCP port number (default 4369). This can also be set using the - <c><![CDATA[ERL_EPMD_PORT]]></c> environment variable, see the + <c><![CDATA[ERL_EPMD_PORT]]></c> environment variable. See the section <seealso marker="#environment_variables">Environment - variables</seealso> below</p> + variables</seealso> below.</p> </item> <tag><c><![CDATA[-names]]></c></tag> <item> @@ -210,7 +210,7 @@ <p>Kill the currently running <c>epmd</c>.</p> <p>Killing the running <c>epmd</c> is only allowed if <c>epmd - -names</c> show an empty database or + -names</c> shows an empty database or <c>-relaxed_command_check</c> was given when the running instance of <c>epmd</c> was started. Note that <c>-relaxed_command_check</c> is given when starting the @@ -228,7 +228,7 @@ <p>This command can only be used when contacting <c>epmd</c> instances started with the <c>-relaxed_command_check</c> flag. Note that relaxed command checking has to be enabled for - the <c>epmd</c> daemon contacted, When running epmd + the <c>epmd</c> daemon contacted. When running epmd interactively, <c>-relaxed_command_check</c> has no effect.</p> </item> @@ -259,7 +259,7 @@ <item> <p>If set prior to start, the <c>epmd</c> daemon will behave as if the <c>-relaxed_command_check</c> option was given at - start-up. If consequently setting this option before starting + start-up. Consequently, if this option is set before starting the Erlang virtual machine, the automatically started <c>epmd</c> will accept the <c>-kill</c> and <c>-stop</c> commands without restrictions.</p> @@ -287,8 +287,8 @@ remote hosts. However, only the query commands are answered (and acted upon) if the query comes from a remote host. It is always an error to try to register a nodename if the client is not a process - located on the same host as the <c>epmd</c> instance is running on, - why such requests are considered hostile and the connection is + located on the same host as the <c>epmd</c> instance is running on- + such requests are considered hostile and the connection is immediately closed.</p> <p>The queries accepted from remote nodes are:</p> @@ -307,3 +307,4 @@ </comref> + diff --git a/erts/doc/src/erl.xml b/erts/doc/src/erl.xml index 9724a1345a..f856b9ab86 100644 --- a/erts/doc/src/erl.xml +++ b/erts/doc/src/erl.xml @@ -495,7 +495,7 @@ <c><![CDATA[werl]]></c>, not <c><![CDATA[erl]]></c> (<c><![CDATA[oldshell]]></c>). Note also that <c><![CDATA[Ctrl-Break]]></c> is used instead of <c><![CDATA[Ctrl-C]]></c> on Windows.</p> </item> - <tag><c><![CDATA[+c]]></c></tag> + <tag><marker id="+c"><c><![CDATA[+c]]></c></marker></tag> <item> <p>Disable compensation for sudden changes of system time.</p> <p>Normally, <c><![CDATA[erlang:now/0]]></c> will not immediately reflect @@ -510,6 +510,9 @@ reflect the current system time. Note that timers are based on <c><![CDATA[erlang:now/0]]></c>. If the system time jumps, timers then time out at the wrong time.</p> + <p><em>NOTE</em>: You can check whether the adjustment is enabled or + disabled by calling + <seealso marker="erlang#system_info_tolerant_timeofday">erlang:system_info(tolerant_timeofday)</seealso>.</p> </item> <tag><c><![CDATA[+d]]></c></tag> <item> @@ -848,6 +851,19 @@ </p> </item> <tag><marker id="+SDio"><c><![CDATA[+SDio IOSchedulers]]></c></marker></tag> + <item> + <p>Sets the number of dirty I/O scheduler threads to create when threading + support has been enabled. The valid range is 0-1024. By default, the number + of dirty I/O scheduler threads created is 10, same as the default number of + threads in the <seealso marker="#async_thread_pool_size">async thread pool + </seealso>. + </p> + <p>This option is ignored if the emulator doesn't have threading support + enabled. Currently, <em>this option is experimental</em> and is supported only + if the emulator was configured and built with support for dirty schedulers + enabled (it's disabled by default). + </p> + </item> <tag><c><![CDATA[+sFlag Value]]></c></tag> <item> <p>Scheduling specific flags.</p> @@ -1170,7 +1186,7 @@ utilization. </p> </item> - <tag><marker id="+swct"><c>+sws very_eager|eager|medium|lazy|very_lazy</c></marker></tag> + <tag><marker id="+swct"><c>+swct very_eager|eager|medium|lazy|very_lazy</c></marker></tag> <item> <p> Set scheduler wake cleanup threshold. Default is <c>medium</c>. diff --git a/erts/doc/src/erl_driver.xml b/erts/doc/src/erl_driver.xml index ad37813ac0..4a1aab75c7 100644 --- a/erts/doc/src/erl_driver.xml +++ b/erts/doc/src/erl_driver.xml @@ -546,6 +546,7 @@ typedef struct ErlDrvSysInfo { int scheduler_threads; int nif_major_version; int nif_minor_version; + int dirty_scheduler_support; } ErlDrvSysInfo; </code> @@ -610,6 +611,10 @@ typedef struct ErlDrvSysInfo { <tag><c>nif_minor_version</c></tag> <item>The value of <c>ERL_NIF_MINOR_VERSION</c> when the runtime system was compiled. </item> + <tag><c>dirty_scheduler_support</c></tag> + <item>A value <c>!= 0</c> if the runtime system has support for dirty scheduler threads; + otherwise <c>0</c>. + </item> </taglist> </item> <tag><marker id="ErlDrvBinary"/>ErlDrvBinary</tag> diff --git a/erts/doc/src/erl_nif.xml b/erts/doc/src/erl_nif.xml index 6b1f4cccf8..3de94be9ff 100644 --- a/erts/doc/src/erl_nif.xml +++ b/erts/doc/src/erl_nif.xml @@ -168,16 +168,18 @@ ok <p><marker id="lengthy_work"/> As mentioned in the <seealso marker="#WARNING">warning</seealso> text at the beginning of this document it is of vital importance that a native function - does return relatively fast. It is hard to give an exact maximum amount + return relatively quickly. It is hard to give an exact maximum amount of time that a native function is allowed to work, but as a rule of thumb - a well behaving native function should return to its caller before a + a well-behaving native function should return to its caller before a millisecond has passed. This can be achieved using different approaches. - If you have full control over the code that are to execute in the native + If you have full control over the code to execute in the native function, the best approach is to divide the work into multiple chunks of - work and call the native function multiple times. Function + work and call the native function multiple times, either directly from Erlang code + or by having a native function schedule a future NIF call via the + <seealso marker="#enif_schedule_nif"> enif_schedule_nif</seealso> function. Function <seealso marker="#enif_consume_timeslice">enif_consume_timeslice</seealso> can be - used this facilitate such work division. In some cases, however, this might not - be possible, e.g. when calling third party libraries. Then you typically want + used to help with such work division. In some cases, however, this might not + be possible, e.g. when calling third-party libraries. Then you typically want to dispatch the work to another thread, return from the native function, and wait for the result. The thread can send the result back to the calling thread using message passing. Information @@ -342,37 +344,38 @@ ok libraries might however fail if deprecated features are used. </p></item> - <tag>Dirty NIFs</tag> - <item><p><marker id="dirty_nifs"/><em>Note that the dirty NIF functionality - is experimental</em> and that you have to enable support for dirty - schedulers when building OTP in order to try the functionality out. Native functions + <tag>Long-running NIFs</tag> + <item><p><marker id="dirty_nifs"/>Native functions <seealso marker="#lengthy_work"> must normally run quickly</seealso>, as explained earlier in this document. They generally should execute for no more than a millisecond. But not all native functions can execute so quickly; for example, functions that encrypt large blocks of data or perform lengthy file system operations can often run for tens of seconds or more.</p> - <p>A NIF that cannot execute in a millisecond or less is called a "dirty NIF" since - it performs work that the Erlang runtime cannot handle cleanly. Applications - that make use of such functions must indicate to the runtime that the functions are + <p>If the functionality of a long-running NIF can be split so that its work can be + achieved through a series of shorter NIF calls, the application can either make that series + of NIF calls from the Erlang level, or it can call a NIF that first performs a chunk of the + work, then invokes the <seealso marker="#enif_schedule_nif">enif_schedule_nif</seealso> + function to schedule another NIF call to perform the next chunk. The final call scheduled + in this manner can then return the overall result. Breaking up a long-running function in + this manner enables the VM to regain control between calls to the NIFs, thereby avoiding + degraded responsiveness, scheduler load balancing problems, and other strange behaviours.</p> + <p>A NIF that cannot be split and cannot execute in a millisecond or less is called a "dirty NIF" + because it performs work that the Erlang runtime cannot handle cleanly. + <em>Note that the dirty NIF functionality described here is experimental</em> and that you have to + enable support for dirty schedulers when building OTP in order to try the functionality out. + Applications that make use of such functions must indicate to the runtime that the functions are dirty so they can be handled specially. To schedule a dirty NIF for execution, the - application calls <seealso marker="#enif_schedule_dirty_nif">enif_schedule_dirty_nif</seealso>, - passing to it a pointer to the dirty NIF to be executed and indicating with a flag + appropriate flags value can be set for the NIF in its <seealso marker="#ErlNifFunc">ErlNifFunc</seealso> + entry, or the application can call <seealso marker="#enif_schedule_nif">enif_schedule_nif</seealso>, + passing to it a pointer to the dirty NIF to be executed and indicating with the <c>flags</c> argument whether it expects the operation to be CPU-bound or I/O-bound.</p> - <p>All dirty NIFs must ultimately invoke the <seealso marker="#enif_schedule_dirty_nif_finalizer"> - enif_schedule_dirty_nif_finalizer</seealso> as their final action, passing to it the - result they wish to return to the original caller. A finalizer function can either - receive the result and return it directly, or it can return a different value instead. - For convenience, the NIF API provides the <seealso marker="#enif_dirty_nif_finalizer"> - enif_dirty_nif_finalizer</seealso> function that applications can use as a finalizer; - it simply returns its result argument.</p> <note><p>Dirty NIF support is available only when the emulator is configured with dirty schedulers enabled. This feature is currently disabled by default. To determine whether the dirty NIF API is available, native code can check to see if the C preprocessor macro <c>ERL_NIF_DIRTY_SCHEDULER_SUPPORT</c> is defined. Also, if the Erlang runtime was built without threading support, dirty schedulers are disabled. To check at runtime for the presence - of dirty scheduler threads, code can call the <seealso marker="#enif_have_dirty_schedulers"><c> - enif_have_dirty_schedulers()</c></seealso> API function, which returns true if dirty - scheduler threads are present, false otherwise.</p></note> + of dirty scheduler threads, code can use the <seealso marker="#enif_system_info"><c> + enif_system_info()</c></seealso> API function.</p></note> </item> </taglist> </section> @@ -498,6 +501,7 @@ typedef struct { const char* <em>name</em>; unsigned <em>arity</em>; ERL_NIF_TERM (*<em>fptr</em>)(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[]); + unsigned flags; } ErlNifFunc; </code> <p>Describes a NIF by its name, arity and implementation. @@ -508,7 +512,17 @@ typedef struct { will thus denote the Nth argument to the NIF. Note that the <c>argc</c> argument allows for the same C function to implement several Erlang functions with different arity (but - same name probably).</p> + same name probably). For a regular NIF, <c>flags</c> is 0 (and + so its value can be omitted for statically initialized <c>ErlNifFunc</c> + instances), or it can be used to indicate that the NIF is a <seealso + marker="#dirty_nifs">dirty NIF</seealso> that should be executed + on a dirty scheduler thread (<em>note that the dirty NIF functionality + described here is experimental</em> and that you have to enable + support for dirty schedulers when building OTP in order to try the + functionality out). If the dirty NIF is expected to be + CPU-bound, its <c>flags</c> field should be set to + <c>ERL_NIF_DIRTY_JOB_CPU_BOUND</c>, or for I/O-bound jobs, + <c>ERL_NIF_DIRTY_JOB_IO_BOUND</c>.</p> </item> <tag><marker id="ErlNifBinary"/>ErlNifBinary</tag> <item> @@ -672,18 +686,6 @@ typedef enum { See also the <seealso marker="#WARNING">warning</seealso> text at the beginning of this document.</p> </desc> </func> - <func><name><ret>ERL_NIF_TERM</ret><nametext>enif_dirty_nif_finalizer(ErlNifEnv* env, ERL_NIF_TERM result)</nametext></name> - <fsummary>Simple dirty NIF result finalizer</fsummary> - <desc> - <p>A convenience function that a dirty NIF can use as a finalizer that simply - return its <c>result</c> argument as its return value. This function is provided - for dirty NIFs with results that should be returned directly to the original caller.</p> - <note><p>This function is available only when the emulator is configured with dirty - schedulers enabled. This feature is currently disabled by default. To determine whether - the dirty NIF API is available, native code can check to see if the C preprocessor macro - <c>ERL_NIF_DIRTY_SCHEDULER_SUPPORT</c> is defined.</p></note> - </desc> - </func> <func><name><ret>int</ret><nametext>enif_equal_tids(ErlNifTid tid1, ErlNifTid tid2)</nametext></name> <fsummary></fsummary> <desc><p>Same as <seealso marker="erl_driver#erl_drv_equal_tids">erl_drv_equal_tids</seealso>. @@ -804,22 +806,6 @@ typedef enum { and return true, or return false if <c>term</c> is not an unsigned integer or is outside the bounds of type <c>unsigned long</c>.</p></desc> </func> - <func><name><ret>int</ret><nametext>enif_have_dirty_schedulers()</nametext></name> - <fsummary>Runtime check for the presence of dirty scheduler threads</fsummary> - <desc> - <p>Check at runtime for the presence of dirty scheduler threads. If the emulator is - built with threading support, dirty scheduler threads are available and - <c>enif_have_dirty_schedulers()</c> returns true. If the emulator was built without - threading support, <c>enif_have_dirty_schedulers()</c> returns false.</p> - <p>If dirty scheduler threads are not available in the emulator, calls to - <c>enif_schedule_dirty_nif</c> and <c>enif_schedule_dirty_nif_finalizer</c> result in - the NIF and finalizer functions being called directly within the calling thread.</p> - <note><p>This function is available only when the emulator is configured with dirty - schedulers enabled. This feature is currently disabled by default. To determine whether - the dirty NIF API is available, native code can check to see if the C preprocessor macro - <c>ERL_NIF_DIRTY_SCHEDULER_SUPPORT</c> is defined.</p></note> - </desc> - </func> <func><name><ret>int</ret><nametext>enif_inspect_binary(ErlNifEnv* env, ERL_NIF_TERM bin_term, ErlNifBinary* bin)</nametext></name> <fsummary>Inspect the content of a binary</fsummary> <desc><p>Initialize the structure pointed to by <c>bin</c> with @@ -873,8 +859,8 @@ typedef enum { <p>Check to see if the current NIF is executing on a dirty scheduler thread. If the emulator is built with threading support, calling <c>enif_is_on_dirty_scheduler</c> from within a dirty NIF returns true. It returns false when the calling NIF is a regular - NIF or a NIF finalizer, both of which run on normal scheduler threads, or when the emulator - is built without threading support.</p> + NIF running on a normal scheduler thread, or when the emulator is built without threading + support.</p> <note><p>This function is available only when the emulator is configured with dirty schedulers enabled. This feature is currently disabled by default. To determine whether the dirty NIF API is available, native code can check to see if the C preprocessor macro @@ -1245,46 +1231,29 @@ typedef enum { <desc><p>Same as <seealso marker="erl_driver#erl_drv_rwlock_tryrwlock">erl_drv_rwlock_tryrwlock</seealso>. </p></desc> </func> - <func><name><ret>ERL_NIF_TERM</ret><nametext>enif_schedule_dirty_nif(ErlNifEnv* env, int flags, ERL_NIF_TERM (*fp)(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[]), int argc, const ERL_NIF_TERM argv[])</nametext></name> - <fsummary>Schedule a dirty NIF for execution</fsummary> + <func><name><ret>ERL_NIF_TERM</ret><nametext>enif_schedule_nif(ErlNifEnv* env, const char* fun_name, int flags, ERL_NIF_TERM (*fp)(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[]), int argc, const ERL_NIF_TERM argv[])</nametext></name> + <fsummary>Schedule a NIF for execution</fsummary> <desc> - <p>Schedule dirty NIF <c>fp</c> to execute a long-running operation. The <c>flags</c> - argument must be set to either <c>ERL_NIF_DIRTY_JOB_CPU_BOUND</c> if the job is expected to - be primarily CPU-bound, or <c>ERL_NIF_DIRTY_JOB_IO_BOUND</c> for jobs that will be - I/O-bound. The <c>argc</c> and <c>argv</c> arguments can either be the originals passed - into the calling NIF, or they can be values created by the calling NIF. The calling - NIF must use the return value of <c>enif_schedule_dirty_nif</c> as its own return value.</p> - <p>Be aware that <c>enif_schedule_dirty_nif</c>, as its name implies, only schedules the - dirty NIF for future execution. The calling NIF does not block waiting for the dirty NIF to - execute and return, which means that the calling NIF can't expect to receive the dirty NIF + <p>Schedule NIF <c>fp</c> to execute. This function allows an application to break up long-running + work into multiple regular NIF calls or to schedule a <seealso marker="#dirty_nifs">dirty NIF</seealso> + to execute on a dirty scheduler thread (<em>note that the dirty NIF functionality described here is + experimental</em> and that you have to enable support for dirty schedulers when building OTP in + order to try the functionality out).</p> + <p>The <c>fun_name</c> argument provides a name for the NIF being scheduled for execution. If it cannot + be converted to an atom, <c>enif_schedule_nif</c> returns a <c>badarg</c> exception.</p> + <p>The <c>flags</c> argument must be set to 0 for a regular NIF, or if the emulator was built the + experimental dirty scheduler support enabled, <c>flags</c> can be set to either <c>ERL_NIF_DIRTY_JOB_CPU_BOUND</c> + if the job is expected to be primarily CPU-bound, or <c>ERL_NIF_DIRTY_JOB_IO_BOUND</c> for jobs that will + be I/O-bound. If dirty scheduler threads are not available in the emulator, a try to schedule such a job + will result in a <c>badarg</c> exception.</p> + + <p>The <c>argc</c> and <c>argv</c> arguments can either be the originals passed into the calling NIF, or + they can be values created by the calling NIF.</p> + <p>The calling NIF must use the return value of <c>enif_schedule_nif</c> as its own return value.</p> + <p>Be aware that <c>enif_schedule_nif</c>, as its name implies, only schedules the + NIF for future execution. The calling NIF does not block waiting for the scheduled NIF to + execute and return, which means that the calling NIF can't expect to receive the scheduled NIF return value and use it for further operations.</p> - <p>A dirty NIF may not invoke the <seealso marker="#enif_make_badarg">enif_make_badarg</seealso> - to raise an exception. If it wishes to return an exception, the dirty NIF should pass a - regular result indicating the exception details to its finalizer, and allow the finalizer - to raise the exception on its behalf.</p> - <note><p>This function is available only when the emulator is configured with dirty schedulers - enabled. This feature is currently disabled by default. To determine whether the dirty NIF API - is available, native code can check to see if the C preprocessor macro - <c>ERL_NIF_DIRTY_SCHEDULER_SUPPORT</c> is defined.</p></note> - </desc> - </func> - <func><name><ret>ERL_NIF_TERM</ret><nametext>enif_schedule_dirty_nif_finalizer(ErlNifEnv* env, ERL_NIF_TERM result, ERL_NIF_TERM (*fp)(ErlNifEnv* env, ERL_NIF_TERM result))</nametext></name> - <fsummary>Schedule a dirty NIF finalizer</fsummary> - <desc> - <p>When a dirty NIF finishes executing, it must schedule a finalizer function to return - its result to the original NIF caller. The dirty NIF passes <c>result</c> as the value it - wants the finalizer to use as the return value. The <c>fp</c> argument is a pointer to the - finalizer function. The NIF API provides the <seealso marker="#enif_dirty_nif_finalizer"> - enif_dirty_nif_finalizer</seealso> function that can be used as a finalizer that simply - returns its <c>result</c> argument. You are also free to write your own custom finalizer - that uses <c>result</c> to derive a different return value, or ignores <c>result</c> - entirely and returns a completely different value.</p> - <p>Without exception, all dirty NIFs must invoke <c>enif_schedule_dirty_nif_finalizer</c> - to complete their execution.</p> - <note><p>This function is available only when the emulator is configured with dirty - schedulers enabled. This feature is currently disabled by default. To determine whether - the dirty NIF API is available, native code can check to see if the C preprocessor macro - <c>ERL_NIF_DIRTY_SCHEDULER_SUPPORT</c> is defined.</p></note> </desc> </func> <func><name><ret>ErlNifPid *</ret><nametext>enif_self(ErlNifEnv* caller_env, ErlNifPid* pid)</nametext></name> @@ -1384,4 +1353,3 @@ typedef enum { <p><seealso marker="erlang#load_nif-2">erlang:load_nif/2</seealso></p> </section> </cref> - diff --git a/erts/doc/src/erl_prim_loader.xml b/erts/doc/src/erl_prim_loader.xml index 6751deda4d..171f84decc 100644 --- a/erts/doc/src/erl_prim_loader.xml +++ b/erts/doc/src/erl_prim_loader.xml @@ -148,6 +148,22 @@ </desc> </func> <func> + <name name="read_link_info" arity="1"/> + <fsummary>Get information about a link or file</fsummary> + <desc> + <p>This function works like + <seealso marker="#read_file_info/1">read_file_info/1</seealso> + except that if <c><anno>Filename</anno></c> is a symbolic link, + information about the link will be returned in the <c>file_info</c> + record and the <c>type</c> field of the record will be set to + <c>symlink</c>.</p> + <p>If <c><anno>Filename</anno></c> is not a symbolic link, this function + returns exactly the same result as <c>read_file_info/1</c>. + On platforms that do not support symbolic links, this function + is always equivalent to <c>read_file_info/1</c>.</p> + </desc> + </func> + <func> <name name="set_path" arity="1"/> <fsummary>Set the path of the loader</fsummary> <desc> diff --git a/erts/doc/src/erlang.xml b/erts/doc/src/erlang.xml index 0f4dfc0f98..84168397f6 100644 --- a/erts/doc/src/erlang.xml +++ b/erts/doc/src/erlang.xml @@ -4968,7 +4968,7 @@ true</pre> <desc> <p>Note that the run-time is the sum of the run-time for all threads in the Erlang run-time system and may therefore be greater - than the wall-clock time.</p> + than the wall-clock time. The time is returned in milliseconds.</p> <pre> > <input>statistics(runtime).</input> {1690,1620} @@ -6295,6 +6295,13 @@ ok (<seealso marker="erts:erl_driver#driver_async">driver_async()</seealso>) as an integer.</p> </item> + <tag><marker id="system_info_tolerant_timeofday"><c>tolerant_timeofday</c></marker></tag> + <item> + <p>Returns whether compensation for sudden changes of system + time is <c>enabled</c> or <c>disabled</c>.</p> + <p>See also <seealso marker="erts:erl#+c">+c</seealso> + command line flag.</p> + </item> <tag><c>trace_control_word</c></tag> <item> <p>Returns the value of the node's trace control word. diff --git a/erts/doc/src/erts_alloc.xml b/erts/doc/src/erts_alloc.xml index c9eca39a99..1ade41f1aa 100644 --- a/erts/doc/src/erts_alloc.xml +++ b/erts/doc/src/erts_alloc.xml @@ -4,7 +4,7 @@ <cref> <header> <copyright> - <year>2002</year><year>2013</year> + <year>2002</year><year>2014</year> <holder>Ericsson AB. All Rights Reserved.</holder> </copyright> <legalnotice> @@ -531,15 +531,9 @@ <p>Multiple, thread specific instances of the allocator. This option will only have any effect on the runtime system with SMP support. Default behaviour on the runtime system with - SMP support:</p> - <taglist> - <tag><c>ll_alloc</c></tag> - <item><c>1</c> instance.</item> - <tag>Other allocators</tag> - <item><c>NoSchedulers+1</c> instances. Each scheduler will use - a lock-free instance of its own and other threads will use - a common instance.</item> - </taglist> + SMP support is <c>NoSchedulers+1</c> instances. Each scheduler will use + a lock-free instance of its own and other threads will use + a common instance.</p> <p>It was previously (before ERTS version 5.9) possible to configure a smaller amount of thread specific instances than schedulers. This is, however, not possible any more.</p> diff --git a/erts/doc/src/notes.xml b/erts/doc/src/notes.xml index 68feaa027a..743369951f 100644 --- a/erts/doc/src/notes.xml +++ b/erts/doc/src/notes.xml @@ -30,6 +30,360 @@ </header> <p>This document describes the changes made to the ERTS application.</p> +<section><title>Erts 6.2</title> + + <section><title>Fixed Bugs and Malfunctions</title> + <list> + <item> + <p> + General documentation updates.</p> + <p> + Own Id: OTP-12052</p> + </item> + <item> + <p>A bug in the VM code implementing sending of signals + to ports could cause the receiving port queue to remain + in a busy state forever. When this state had been + reached, processes sending command signals to the port + either got suspended forever, or, if the <c>nosuspend</c> + feature was used, always failed to send to the port. This + bug was introduced in ERTS version 5.10.</p> + <p>In order for this bug to be triggered on a port, one + had to at least once utilize the <c>nosuspend</c> + functionality when passing a signal to the port. This by + either calling</p> <list> <item> <seealso + marker="erlang#port_command/3"><c>port_command(Port, + Data, [nosuspend | Options])</c></seealso>, </item> + <item> <seealso + marker="erlang#send/3"><c>erlang:send(Port, {PortOwner, + {command, Data}}, [nosuspend | Options])</c></seealso>, + </item> <item> <seealso + marker="erlang#send_nosuspend/2"><c>erlang:send_nosuspend(Port, + {PortOwner, {command, Data}})</c></seealso>, or </item> + <item> <seealso + marker="erlang#send_nosuspend/3"><c>erlang:send_nosuspend(Port, + {PortOwner, {command, Data}}, Options)</c></seealso>. + </item> </list> + <p>Thanks Vasily Demidenok for reporting the issue, and + Sergey Kudryashov for providing a testcase.</p> + <p> + Own Id: OTP-12082 Aux Id: OTP-10336 </p> + </item> + <item> + <p> + Fix size overflow bug at memory allocation. A memory + allocation call, with an insane size close to the entire + address space, could return successfully as if it had + allocated just a few bytes. (Thanks to Don A. Bailey for + reporting)</p> + <p> + Own Id: OTP-12091</p> + </item> + <item> + <p> + Fix various issues where negating a signed integer would + trigger undefined behaviour. This fixes issues in the + enif_make_int64 interface and some edge cases inside the + erlang runtime system.</p> + <p> + Own Id: OTP-12097</p> + </item> + <item> + <p> + The documentation erroneously listed the <seealso + marker="erl#+swct"><c>+swct</c></seealso> command line + argument under <c>+sws</c>.</p> + <p> + Own Id: OTP-12102 Aux Id: OTP-10994 </p> + </item> + <item> + <p> + Profiling messages could be delivered out of order when + profiling on <c>runnable_procs</c> and/or + <c>runnable_ports</c> using <seealso + marker="erlang#system_profile/2"><c>erlang:system_profile/2</c></seealso>. + This bug was introduced in ERTS version 5.10.</p> + <p> + Own Id: OTP-12105 Aux Id: OTP-10336 </p> + </item> + <item> + <p> + Various logging fixes, including: Add run queue index to + the process dump in crash dumps.<br/> Add thread index to + enomem slogan when crashing.<br/> Remove error logger + message for sending messages to old instances of the same + node.</p> + <p> + Own Id: OTP-12115</p> + </item> + <item> + <p> + Fix compiler warnings reported by LLVM</p> + <p> + Own Id: OTP-12138</p> + </item> + <item> + <p> + Correct conversion of <c>MIN_SMALL</c> by + <c>list_to_integer/1</c> and <c>binary_to_integer/1</c>. + The bug produced an unnormalized bignum which can cause + strange behavior such as comparing different to a correct + <c>MIN_SMALL</c> integer. The value <c>MIN_SMALL</c> is + <c>-(1 bsl 27) = -134217728</c> on a 32-bit VM and <c>-(1 + bsl 59) = -576460752303423488</c> on a 64-bit VM. (Thanks + to Jesper Louis Andersen, Mikael Pettersson and Anthony + Ramine for report, patch and optimization suggestion)</p> + <p> + Own Id: OTP-12140</p> + </item> + <item> + <p> + Fix bug in <c>term_to_binary</c> that reallocates binary + with inconsistent size information. Bug has never been + confirmed to be the cause of any faulty behavior.</p> + <p> + Own Id: OTP-12141</p> + </item> + <item> + <p> + Real_path method used while prim loading archive files + was not taking into account the fact that windows + directory symlinks can be across different drives.</p> + <p> + Own Id: OTP-12155</p> + </item> + </list> + </section> + + + <section><title>Improvements and New Features</title> + <list> + <item> + <p> + Add log2 histogram to lcnt for lock wait time</p> + <p> + Own Id: OTP-12059</p> + </item> + <item> + <p> + Introduced <seealso + marker="erl_nif#enif_schedule_nif"><c>enif_schedule_nif()</c></seealso> + to the NIF API.</p> + <p> + The <c>enif_schedule_nif()</c> function allows a + long-running NIF to be broken into separate NIF + invocations without the help of a wrapper function + written in Erlang. The NIF first executes part of the + long-running task, then calls <c>enif_schedule_nif()</c> + to schedule a NIF for later execution to continue the + task. Any number of NIFs can be scheduled in this manner, + one after another. Since the emulator regains control + between invocations, this helps avoid problems caused by + native code tying up scheduler threads for too long.</p> + <p> + The <c>enif_schedule_nif()</c> function also replaces the + <c>enif_schedule_dirty_nif()</c> in the experimental + dirty NIF API. Note that the only incompatible changes + made are in the experimental dirty NIF API.</p> + <p> + See the <seealso marker="erl_nif">NIF + documentation</seealso> for more information.</p> + <p> + Thanks to Steve Vinoski.</p> + <p> + *** POTENTIAL INCOMPATIBILITY ***</p> + <p> + Own Id: OTP-12128</p> + </item> + </list> + </section> + +</section> + +<section><title>Erts 6.1.2</title> + + <section><title>Fixed Bugs and Malfunctions</title> + <list> + <item> + <p> + OTP-11850 fixed filelib:wildcard/1 to work with broken + symlinks. This correction, however, introduced problems + since symlinks were no longer followed for functions like + filelib:ensure_dir/1, filelib:is_dir/1, + filelib:file_size/1, etc. This is now corrected.</p> + <p> + Own Id: OTP-12054 Aux Id: seq12660 </p> + </item> + </list> + </section> + +</section> + +<section><title>Erts 6.1.1</title> + + <section><title>Fixed Bugs and Malfunctions</title> + <list> + <item> + <p> + Fixed ETHR_FORCE_INLINE which caused the build to break + on some platforms without adequate thread support + (VxWorks).</p> + <p> + Own Id: OTP-12010</p> + </item> + </list> + </section> + +</section> + +<section><title>Erts 6.1</title> + + <section><title>Fixed Bugs and Malfunctions</title> + <list> + <item> + <p>The documentation for <c>spawn_opt/5</c> now has a + note mentioning that the <c>monitor</c> option is not + supported.</p> + <p> + Own Id: OTP-11849</p> + </item> + <item> + <p> + Fix broken system monitoring of <c>large_heap</c> for + non-smp VM. No message for <c>large_heap</c> was ever + sent on non-smp VM. Bug exist since R16B.</p> + <p> + Own Id: OTP-11852</p> + </item> + <item> + <p> + The emulator without SMP support crashed when passing a + message to a process without enough heap space for the + message. This bug was introduced in <c>erts-6.0</c>.</p> + <p> + Own Id: OTP-11887 Aux Id: OTP-11388 </p> + </item> + <item> + <p> + Fix race between ETS table deletion and unfixation that + could cause VM crash. The race could happen between a + terminating process that does not own the table but has a + fixation on it and another process that deletes the table + (maybe the owner terminating) at the same time. Bug + existed since R15B02.</p> + <p> + Own Id: OTP-11892</p> + </item> + <item> + <p>The string following the <c>-eval</c> option when + invoking <c>erl</c> would not be properly translated from + UTF-8 to a list of Unicode characters (as would the + arguments for <c>-run</c>).</p> + <p>That bug would cause the build of Erlang/OTP to fail + when building in a directory whose pathname contained + non-US ASCII characters encoded in UTF-8. (Thanks to Eric + Pailleau for reporting this bug.)</p> + <p> + Own Id: OTP-11916</p> + </item> + <item> + <p> + Fix erts_debug:size/1 to handle Map sizes</p> + <p> + Own Id: OTP-11923</p> + </item> + <item> + <p> + Removed <c>erlang:bitstr_to_list/1</c> and + <c>erlang:list_to_bitstr/1</c>. They were added by + mistake, and have always raised an <c>undefined</c> + exception when called.</p> + <p> + Own Id: OTP-11942</p> + </item> + <item> + <p> + Fixed compilation using mingw-w64 on Windows.</p> + <p> + Thanks to Jani Hakala.</p> + <p> + Own Id: OTP-11945</p> + </item> + <item> + <p> + The git sha is no longer printed in the shell start + header when erlang is built from a tagged git release.</p> + <p> + Own Id: OTP-11961</p> + </item> + <item> + <p> + Fixed a bug where <c>send</c> trace events were + erroneously dropped when the send was done to a + registered process. This bug was introduced in R16B.</p> + <p> + Own Id: OTP-11968</p> + </item> + </list> + </section> + + + <section><title>Improvements and New Features</title> + <list> + <item> + <p>The following native functions now bump an appropriate + amount of reductions and yield when out of + reductions:</p> <list> + <item><c>erlang:binary_to_list/1</c></item> + <item><c>erlang:binary_to_list/3</c></item> + <item><c>erlang:bitstring_to_list/1</c></item> + <item><c>erlang:list_to_binary/1</c></item> + <item><c>erlang:iolist_to_binary/1</c></item> + <item><c>erlang:list_to_bitstring/1</c></item> + <item><c>binary:list_to_bin/1</c></item> </list> + <p>Characteristics impact:</p> <taglist> + <tag>Performance</tag> <item>The functions converting + from lists got a performance loss for very small lists, + and a performance gain for very large lists.</item> + <tag>Priority</tag> <item>Previously a process executing + one of these functions effectively got an unfair priority + boost. This priority boost depended on the input size. + The larger the input was, the larger the priority boost + got. This unfair priority boost is now lost. </item> + </taglist> + <p> + Own Id: OTP-11888</p> + </item> + <item> + <p> + The systemd features of epmd have been removed from epmd + by default. To enable them you have to build erlang with + the configure option --enable-systemd.</p> + <p> + Own Id: OTP-11921</p> + </item> + <item> + <p> + Removed Erlang wrapper code used when calling + <c>binary_to_term/1</c>, and <c>binary_to_term/2</c>. + This improves the performance of these BIFs especially + when they are called with small binaries as input.</p> + <p> + Own Id: OTP-11931</p> + </item> + <item> + <p> + Add erlang:system_info(tolerant_timeofday), an API to + check whether compensation for sudden changes of system + time is enabled or not.</p> + <p> + Own Id: OTP-11970</p> + </item> + </list> + </section> + +</section> + <section><title>Erts 6.0.1</title> <section><title>Fixed Bugs and Malfunctions</title> @@ -595,9 +949,9 @@ "hi" := V1, a := V2, b := V3} = M2. % match keys with values</c></item> </taglist></p> <p> - For information on how to use Maps please see the - <seealso marker="doc/reference_manual:maps">Reference - Manual</seealso>.</p> + For information on how to use Maps please see Map Expressions in the + <seealso marker="doc/reference_manual:expressions#map_expressions"> + Reference Manual</seealso>.</p> <p> The current implementation is without the following features: <taglist> <item>No variable keys</item> @@ -756,6 +1110,27 @@ Thanks to Matwey V. Kornilov</p> <p> Own Id: OTP-11829</p> + </item> + </list> + </section> + +</section> + +<section><title>Erts 5.10.4.1</title> + + <section><title>Known Bugs and Problems</title> + <list> + <item> + <p> + When using gen_tcp:connect and the <c>fd</c> option with + <c>port</c> and/or <c>ip</c>, the <c>port</c> and + <c>ip</c> options were ignored. This has been fixed so + that if <c>port</c> and/or <c>ip</c> is specified + together with <c>fd</c> a bind is requested for that + <c>fd</c>. If <c>port</c> and/or <c>ip</c> is not + specified bind will not be called.</p> + <p> + Own Id: OTP-12061</p> </item> </list> </section> @@ -4854,7 +5229,7 @@ <item> <p> The <c>configure</c> command line argument <seealso - marker="doc/installation_guide:INSTALL#How-to-Build-and-Install-ErlangOTP_A-Closer-Look-at-the-individual-Steps_Configuring">--enable-ethread-pre-pentium4-compatibility</seealso> + marker="doc/installation_guide:INSTALL#Advanced-configuration-and-build-of-ErlangOTP">--enable-ethread-pre-pentium4-compatibility</seealso> had no effect. This option is now also automatically enabled if required on the build machine.</p> <p> @@ -5433,7 +5808,7 @@ platforms than before. If <c>configure</c> warns about no atomic implementation available, try using the <c>libatomic_ops</c> library. Use the <seealso - marker="doc/installation_guide:INSTALL#How-to-Build-and-Install-ErlangOTP_A-Closer-Look-at-the-individual-Steps_Configuring">--with-libatomic_ops=PATH</seealso> + marker="doc/installation_guide:INSTALL#Advanced-configuration-and-build-of-ErlangOTP">--with-libatomic_ops=PATH</seealso> <c>configure</c> command line argument when specifying where the <c>libatomic_ops</c> installation is located. The <c>libatomic_ops</c> library can be downloaded from: @@ -5451,7 +5826,7 @@ the pentium 4 processor. If you want the runtime system to be compatible with older processors (back to 486) you need to pass the <seealso - marker="doc/installation_guide:INSTALL#How-to-Build-and-Install-ErlangOTP_A-Closer-Look-at-the-individual-Steps_Configuring">--enable-ethread-pre-pentium4-compatibility</seealso> + marker="doc/installation_guide:INSTALL#Advanced-configuration-and-build-of-ErlangOTP">--enable-ethread-pre-pentium4-compatibility</seealso> <c>configure</c> command line argument when configuring the system.</p> <p> |