aboutsummaryrefslogtreecommitdiffstats
path: root/erts/emulator
AgeCommit message (Collapse)Author
2016-03-11Eliminate use of ?config() macroBjörn-Egil Dahlberg
2016-03-11Replace use of test_server:format/2 with io:format/2Björn-Egil Dahlberg
2016-03-11Eliminate use of test_server:fail/0,1Björn-Egil Dahlberg
2016-03-11Modernize use of timetrapsBjörn-Egil Dahlberg
2016-03-11Merge branch 'vinoski/ds-avoid-lock'Rickard Green
* vinoski/ds-avoid-lock: Skip run queue lock check for dirty schedulers
2016-03-11Merge branch 'rickard/ds-sched-suspend'Rickard Green
* rickard/ds-sched-suspend: Improved scheduler suspend functionality
2016-03-10Improved scheduler suspend functionalityRickard Green
- The calling process is now suspended while synchronizing scheduler suspends via erlang:system_flag(schedulers_online, _) and erlang:system_flag(multi_scheduling, _), instead of blocking the scheduler thread in the BIF call waiting for the operation to synchronize. Besides releasing the scheduler for other work (or immediate suspend) it also makes it possible to abort the operation by killing the process. - erlang:system_flag(schedulers_online, _) now only wait for normal schedulers to complete before it returns. This since it may take a very long time before all dirty schedulers suspends. - erlang:system_flag(multi_scheduling, block_normal|unblock_normal) which only operate on normal schedulers has been introduced. This since there are use cases where suspend of dirty schedulers are not of interest (hipe loader). - erlang:system_flag(multi_scheduling, block) still blocks all dirty schedulers as well as all normal schedulers except one since it is hard to redefine what multi scheduling block means. - The three operations: - changing amount of schedulers online - blocking/unblocking normal multi scheduling - blocking/unblocking full multi scheduling can now be done in parallel. This is important since otherwise a full multi scheduling block would potentially delay the other operations for a very long time.
2016-03-09Merge branch 'maint'Sverker Eriksson
2016-03-09erts: Fix wobbling tests in nif_SUITESverker Eriksson
No point in checking tmp_alloc instance 0 as any non-scheduler thread could race us.
2016-03-08Merge branch 'sverk/literal-alloc-polish'Sverker Eriksson
* sverk/literal-alloc-polish: erts: Add emulator flag +MIscs for literal super carrier size erts: Refactor init of erts_literal_mmapper erts: Make literal_alloc documented and configurable
2016-03-08Merge branch 'maint'Sverker Eriksson
2016-03-08erts: Fix alloc_SUITE:rbtree and migration for win64Sverker Eriksson
One little (unsigned long) left behind.
2016-03-08Merge branch 'maint'Sverker Eriksson
2016-03-08test: do not divide by zeroDan Gudmundsson
2016-03-01Merge branch 'sverk/master/halt-INT_MIN'Sverker Eriksson
* sverk/master/halt-INT_MIN: erts: Make erlang:halt() accept bignums as Status erts: Change erl_exit into erts_exit kernel: Remove calls to erl_exit
2016-03-01Merge branch 'sverk/halt-INT_MIN' into maintSverker Eriksson
OTP-13251 * sverk/halt-INT_MIN: erts: Make erlang:halt() accept bignums as Status erts: Change erl_exit into erts_exit kernel: Remove calls to erl_exit
2016-02-29Fix build wihtout thread supportRickard Green
2016-02-26Merge branch 'bjorn/multiple-load/OTP-13111'Björn Gustavsson
* bjorn/multiple-load/OTP-13111: code: Add functions that can load multiple modules Refactor post_beam_load handling Simplify and robustify code_server:all_loaded/1 Update preloaded modules Add erl_prim_loader:get_modules/3 Add has_prepared_code_on_load/1 BIF Allow erlang:finish_loading/1 to load more than one module beam_load.c: Add a function to check for an on_load function
2016-02-25Merge branch 'maint'Björn-Egil Dahlberg
Conflicts: erts/emulator/beam/erl_alloc.types erts/emulator/beam/erl_bif_info.c erts/emulator/beam/erl_process.c erts/preloaded/ebin/erts_internal.beam
2016-02-25Merge branch 'egil/heart-callback/OTP-13250' into maintBjörn-Egil Dahlberg
2016-02-25Add has_prepared_code_on_load/1 BIFBjörn Gustavsson
2016-02-25Allow erlang:finish_loading/1 to load more than one moduleBjörn Gustavsson
The BIFs prepare_loading/2 and finish_loading/1 have been designed to allow fast loading in parallel of many modules. Because of the complications with on_load functions, the initial implementation of finish_loading/1 only allowed a single element in the list of prepared modules. finish_loading/1 does not suspend other processes, but it must wait for all schedulers to pass a write barrier ("thread progress"). The time for all schedulers to pass the write barrier is highly variable, depending on what kind of code they are executing. Therefore, allowing finish_loading/1 to finish the loading for more than one module before passing the write barrier could potentially be much faster than calling finish_loading/1 multiple times. The test case many/1 run on my computer shows that with "heavy load", finish loading of 100 modules in parallel is almost 50 times faster than loading them sequentially. With "light load", the gain is still almost 10 times. Here follows an actual sample of the output from the test case on my computer (an 2012 iMac): Light load ========== Sequential: 22361 µs Parallel: 2586 µs Ratio: 9 Heavy load ========== Sequential: 254512 µs Parallel: 5246 µs Ratio: 49
2016-02-24Merge branch 'master' into sverk/master/halt-INT_MINSverker Eriksson
2016-02-24erts: Make erlang:halt() accept bignums as StatusSverker Eriksson
Just mask away the high bits to get a more tolerant erlang:halt that behaves the same on 32 and 64 bit architectures.
2016-02-24erts: Change erl_exit into erts_exitSverker Eriksson
This is mostly a pure refactoring. Except for the buggy cases when calling erlang:halt() with a positive integer in the range -(INT_MIN+2) to -INT_MIN that got confused with ERTS_ABORT_EXIT, ERTS_DUMP_EXIT and ERTS_INTR_EXIT. Outcome OLD erl_exit(n, ) NEW erts_exit(n, ) ------- ------------------- ------------------------------------------- exit(Status) n = -Status <= 0 n = Status >= 0 crashdump+abort n > 0, ignore n n = ERTS_ERROR_EXIT < 0 The outcome of the old ERTS_ABORT_EXIT, ERTS_INTR_EXIT and ERTS_DUMP_EXIT are the same as before (even though their values have changed).
2016-02-24Merge branch 'maint'Rickard Green
* maint: Do not wait for main lock when looking up process not running
2016-02-23Do not wait for main lock when looking up process not runningRickard Green
2016-02-23erts: Add emulator flag +MIscs for literal super carrier sizeSverker Eriksson
2016-02-23erts: Refactor init of erts_literal_mmapperSverker Eriksson
2016-02-23erts: Make literal_alloc documented and configurableSverker Eriksson
Except it cannot be disabled and cannot be multi-threaded. The bit-vector 'erts_literal_vspace_map' on 32-bit is currently only protected by the literal allocator mutex. We could allow multiple instances on 64-bit (I think), but what would be the point?
2016-02-22Merge branch 'maint'Sverker Eriksson
2016-02-22Merge branch 'sverk/proc-lock-check-fix' into maintSverker Eriksson
* sverk/proc-lock-check-fix: erts: Fix lock checker for process locks
2016-02-21Skip run queue lock check for dirty schedulersSteve Vinoski
The wake_scheduler function asserts that the run queue is not locked, but this assertion sometimes fails for dirty schedulers (in January 2016 a user in erlang-questions reported a dirty schedulers problem related to this). After discussing it with Rickard, we decided modifying the assertion was the most practical way to address the problem.
2016-02-20beam_load.c: Add a function to check for an on_load functionBjörn Gustavsson
We will need a way to check whether an prepared BEAM modules has an on_load function.
2016-02-19Merge branch 'rickard/rq-state-bug/OTP-13298' into maintRickard Green
* rickard/rq-state-bug/OTP-13298: Fix bug causing run-queue mask to become inconsistent
2016-02-18erts: Fix lock checker for process locksSverker Eriksson
Do lock order check *before* trying to seize lock... duh!
2016-02-18Merge branch 'sverk/erts-hipe_x86_signal-cleanups'Sverker Eriksson
OTP-13341 * sverk/erts-hipe_x86_signal-cleanups: hipe_sigaltstack: correct initialization of ss.ss_flags hipe_x86_signal: cleanups hipe_x86_signal: cleanups hipe_x86_signal: cleanups hipe_x86_signal: cleanups hipe_x86_signal: cleanups hipe_x86_signal: cleanups
2016-02-18Merge branch 'sverk/fix-list-length-int/OTP-13288'Sverker Eriksson
* sverk/fix-list-length-int/OTP-13288: erts: Fix error cases in enif_get_list_length erts: Use Sint instead of int for list lengths
2016-02-18Merge branch 'bjorn/remove-test_server/OTP-12705'Björn Gustavsson
* bjorn/remove-test_server/OTP-12705: Remove test_server as a standalone application Erlang mode for Emacs: Include ct.hrl instead test_server.hrl Remove out-commented references to the test_server applications Makefiles: Remove test_server from include path and code path Eliminate use of test_server.hrl and test_server_line.hrl
2016-02-18Merge branch 'maint'Sverker Eriksson
2016-02-17Merge branch 'jv/erts/optimize-cmp'Björn Gustavsson
* jv/erts/optimize-cmp: Unify comparison macros in erl_utils.h Avoid erts_cmp jump in atom, int and float comparisons
2016-02-17Makefiles: Remove test_server from include path and code pathBjörn Gustavsson
Since no test suites includede test_server.hrl, there is no need to have test_server in the include path or code path.
2016-02-17Eliminate use of test_server.hrl and test_server_line.hrlBjörn Gustavsson
As a first step to removing the test_server application as as its own separate application, change the inclusion of test_server.hrl to an inclusion of ct.hrl and remove the inclusion of test_server_line.hrl.
2016-02-16erts: Add BIF erts_internal:system_check/1Björn-Egil Dahlberg
This commit implements erts_internal:system_check(schedulers) with the intent of a basic responsiveness test check of the schedulers.
2016-02-15Unify comparison macros in erl_utils.hJosé Valim
This removes the duplication in having both `cmp_eq` and `CMP_EQ` and normalizes their name to uppercase.
2016-02-15Avoid erts_cmp jump in atom, int and float comparisonsJosé Valim
Given the function definition below: check(X) when X >= 0, X <= 20 -> true. @nox has originally noticed that perfoming lt and ge guard tests were performing slower than they should be. Further investigation revealed that most of the cost was in jumping to the erts_cmp function. This patch brings the operations already inlined in erts_cmp into the emulator, removing the jump cost. After applying these changes, invoking the check/1 function defined above 30000 times with different values from 0 to 20 has fallen from 367us to 213us (measured as average of 3 runs). This is a considerably improvement over Erlang 18 which takes 556us on average. Floats have also dropped their time from 1126us (on Erlang 18) to 613us.
2016-02-15Merge branch 'rickard/ds-fixes' into rickard/ds-fixes-masterRickard Green
* rickard/ds-fixes: Fix unique_SUITE for dirty schedulers Add dirty scheduler process termination test Ensure that work is done on the correct type of schedulers Conflicts: erts/emulator/beam/erl_process.c erts/emulator/beam/erl_process.h erts/emulator/beam/erl_process_dump.c
2016-02-15Fix unique_SUITE for dirty schedulersRickard Green
2016-02-15Add dirty scheduler process termination testSteve Vinoski
In scheduler_SUITE add a new test that runs a single dirty I/O scheduler and launches a number of dirty I/O NIF calls that each sleep for 3 seconds. Given the single scheduler, the first of these will run while the rest queue up. Then start killing these processes, and verify they call exit correctly.
2016-02-15Ensure that work is done on the correct type of schedulersRickard Green
Only the actual call to the dirty nif is allowed to execute on dirty schedulers. The dirty nif is not allowed to execute on normal schedulers if dirty schedulers are available. Arrival of exit signals and system tasks, while a process was scheduled for execution on a dirty scheduler, could mess up the process internal state. Preparation for dirty system task has been made, but is currently unused.