github
/
ruby
зеркало из https://github.com/github/ruby.git
1
0
Форкнуть 0
Код Задачи Пакеты Проекты Релизы Вики Активность
83 303 коммитов 219 Ветки 1 255 Теги 478 MiB
Граф коммитов

30 Коммитов

Автор SHA1 Сообщение Дата
Takashi Kokubun da02d08f27 RJIT: Rewind $! after compilation 2023-12-22 22:18:14 -08:00
Takashi Kokubun 2e0277b133 RJIT: Make --rjit-trace actually work 2023-12-21 21:46:19 -08:00
Takashi Kokubun 64c52cd1c2 RJIT: Add --rjit-trace to allow TracePoint during JIT 2023-12-21 21:05:13 -08:00
Takashi Kokubun 009968a700 RJIT: Avoid incrementing vm_insns_count 
during compilation
 2023-12-21 17:12:40 -08:00
Takashi Kokubun c73959cff4 RJIT: Rename pause/resume to disable/enable 
like YJIT. They don't work in the same way yet, but it's nice to make
the naming consistent first so that we will not need to rename them
later.
 2023-12-21 14:25:41 -08:00
Takashi Kokubun eb872d1752 RJIT: Share rb_vm_insns_count for vm_insns_count 2023-12-18 23:55:40 -08:00
Koichi Sasada c4c39082af add `flags` to `rb_postponed_job_preregister` 
for future extensions.
 2023-12-10 15:39:06 +09:00
KJ Tsanaktsidis f8effa209a Change the semantics of rb_postponed_job_register 
Our current implementation of rb_postponed_job_register suffers from
some safety issues that can lead to interpreter crashes (see bug #1991).
Essentially, the issue is that jobs can be called with the wrong
arguments.

We made two attempts to fix this whilst keeping the promised semantics,
but:
  * The first one involved masking/unmasking when flushing jobs, which
    was believed to be too expensive
  * The second one involved a lock-free, multi-producer, single-consumer
    ringbuffer, which was too complex

The critical insight behind this third solution is that essentially the
only user of these APIs are a) internal, or b) profiling gems.

For a), none of the usages actually require variable data; they will
work just fine with the preregistration interface.

For b), generally profiling gems only call a single callback with a
single piece of data (which is actually usually just zero) for the life
of the program. The ringbuffer is complex because it needs to support
multi-word inserts of job & data (which can't be atomic); but nobody
actually even needs that functionality, really.

So, this comit:
  * Introduces a pre-registration API for jobs, with a GVL-requiring
    rb_postponed_job_prereigster, which returns a handle which can be
    used with an async-signal-safe rb_postponed_job_trigger.
  * Deprecates rb_postponed_job_register (and re-implements it on top of
    the preregister function for compatability)
  * Moves all the internal usages of postponed job register
    pre-registration
 2023-12-10 15:00:37 +09:00
Takashi Kokubun 417eb83b48 RJIT: Tweak the default call threshold 
This number works better on yjit-bench (ruby-lsp).
 2023-04-26 18:19:48 -07:00
Takashi Kokubun 19506650ef RJIT: Add --rjit-verify-ctx option 2023-04-04 00:35:29 -07:00
Takashi Kokubun 6002b12611 RJIT: Support entry with different PCs 2023-04-02 15:27:40 -07:00
Takashi Kokubun d189f8d870 RJIT: Prefix rjit_options with rb_ 2023-03-18 21:28:55 -07:00
Takashi Kokubun 644c998525 RJIT: Support --rjit-stats on release build as well 2023-03-17 22:31:41 -07:00
Takashi Kokubun 9cd5441d28 RJIT: Implement --rjit-trace-exits 2023-03-12 15:15:08 -07:00
Takashi Kokubun c364e0745d RJIT: Introduce --rjit-exec-mem-size 2023-03-10 13:04:45 -08:00
Takashi Kokubun 3c35c13aaa RJIT: Resurrect --rjit-pause and RJIT.resume 2023-03-09 22:43:53 -08:00
Takashi Kokubun 1a0d3ec4b9 RJIT: Make functions in rjit_c.c static 
They don't need to be global.
 2023-03-08 23:38:02 -08:00
Takashi Kokubun 4be224eaf5 RJIT: Reorganize rjit.c 
Reordering functions for readability, adding stats whenever possible.
 2023-03-08 23:31:06 -08:00
Takashi Kokubun f5909ac6d9 RJIT: Stop allowing leaked globals rjit_* 2023-03-08 23:24:38 -08:00
Takashi Kokubun a0918a4a80 RJIT: Get rid of verbose logging 2023-03-08 23:12:51 -08:00
Takashi Kokubun 7d7b67a472 RJIT: Clean up the declaration mess 2023-03-08 23:07:30 -08:00
Takashi Kokubun 9ad19069f9 Remove obsoleted functions in rjit.c 2023-03-07 23:59:50 -08:00
Takashi Kokubun 4bf037bebd Update documentation about RJIT 2023-03-07 23:31:39 -08:00
Takashi Kokubun 43de2365a5 Fix a typo 2023-03-07 23:06:01 -08:00
Takashi Kokubun 6d91df08b5
Allow enabling YJIT and RJIT independently (#7474) 
We used to require MJIT is supported when YJIT is supported. However,
now that RJIT dropped some platforms that YJIT supports, it no longer
makes sense. We should be able to enable only YJIT, and vice versa.
 2023-03-07 22:43:37 -08:00
Takashi Kokubun 0bf4cd8e1c Actually stop supporting those options 2023-03-07 21:21:19 -08:00
Takashi Kokubun 1d39d2d334 Update options available in RJIT 2023-03-07 21:19:04 -08:00
Takashi Kokubun 23ec248e48 s/mjit/rjit/ 2023-03-06 23:44:01 -08:00
Takashi Kokubun 2e875549a9 s/MJIT/RJIT/ 2023-03-06 23:44:01 -08:00
Takashi Kokubun eaccdc1941 Rename MJIT filenames to RJIT 2023-03-06 23:44:01 -08:00