from Subversion.
This behavior is tentative and not discussed well. The point of
discussion will be just the length of commit hash, and I thought we
should include this kind of change in 2.7.0-preview1 release even before
the length is fixed yet.
Let's discuss that afterwards and fix it later as needed. Naruse
suggested that length=10 is very unlikely to cause conflict, and thus
it's used by email notification and rubyci now. This behavior is in
favor of that for now.
They seem to have never been used from the beginning (r45081).
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@67696 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
I believe no-one use this notation for years.
[Fix GH-2136]
From: Chikanaga Tomoyuki <t-chikanaga@groovenauts.jp>
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@67658 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
Reline is a readline stdlib compatible library. It also supports
multiline input. IRB is improved with Reline and supports multiline.
Besides, supports showing documents when completed.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@67645 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* tool/downloader.rb (Downloader::Unicode.download): enable cache
files when index.html has not been modified since the previous
download, even if beta version.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@67594 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
for ISeq including only leaf and no-handles_sp insns except leaf.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@67574 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
for opt_aref with inline cache to minimize the possibility of JIT cancel.
Also opt_aset and opt_mod are added for the targets.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@67547 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
I'm writing `//` comments in newer MJIT code after C99 enablement
(because I write 1-line comments more often than multi-line comments
and `//` requires fewer chars on 1-line) and then they are mixed
with `/* */` now.
For consistency and to avoid the conversion in future changes, let me
finish the rewrite in MJIT-related code.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@67533 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
Debugging output is no longer needed because the problem has been fixed with r67449.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@67451 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
Unicode file download doesn't work with Visual Studio, we need more debug output.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@67447 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* tool/downloader.rb (Downloader.save_cache): keep linked file
newer than cached file, so that GNU make triggers when the
content is updated. it uses the timestamp of symlink itself
instead of the target.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@67413 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
cfp->bp was (re-)introduced by Kokubun san, but VM doesn't use it
because I (ko1) want to remove it in a future. But using it make
leave instruction fast because of sp consisntency check.
So now VM uses cfp->bp.
To use cfp->bp, I checked the value and I found that it is not a
"initial value of sp" but a "initial value of ep". Fix this problem
and fix all bp references (this is why bp is renamed to bp_).
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@67342 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
Add counters to count ccf (call cache fastpath) usage.
These counters will help which kind of method dispatch
is important to optimize.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@67336 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
to avoid inlining a method call when it becomes argument_arity_error,
fixing a potential bug.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@67330 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
Remove empty UNICODE_FILES list to update Unicode files like as
r67025, for 2.6 branch.
[Bug #15593]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@67040 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
Moving public headers was 12-years ago, no depend files would
expect ruby.h in the top source directory now.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@67033 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
If a function has rb_warning() that is not a leaf because warning
ultimately is a method call of Warning#warn.
If a function has rb_name_error() that is not a leaf because
NameError is allocated, then initialized. This of course
involves calling NameError#initialize.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@66976 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
Now that comma at the end of enum is allowed, we can write this much
more straight-forward.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@66960 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
Also previous rb_vm_insn_len_info is now a function static variable.
It seems nobody else is using it.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@66958 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
C99 allows string literals of at least 4095 characters (cf: ISO/IEC
9899:1999 section 5.2.4.1) so the previous complex struct layout is no
longer necessary. Just dump the verbatim memory contents we want.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@66957 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
This changeset modifies the VM generator so that vm.inc is written in
C99. Also added some comments in _insn_entry.erb so that the
intention of each parts to be made more clear. I think this improves
overall readability of the generated VM.
Confirmed that the exact same binary is generated before/after this
changeset.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@66923 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
These changes break the behavior of default gems. Bug #13428 says
r58345 is reasonable because gemspec file is installed by `to_ruby_for_cache`
method. But I revert `to_ruby_for_cache` in rbinstall.rb at r58403.
There is no reason that we apply r58345 now.
But I'm not sure about gemspec of default gems affects standalone gems.
I'm going to investigate it on rubygems/rubygems.
[Bug #15500][ruby-core:90867]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@66867 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
and functions to clarify the intention and make sure it's not used in a
surprising way (like using 2, 3, ... other than 0, 1 even while it seems
to be a boolean).
This is a retry of r66775. It included some typos...
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@66778 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
In downloader.rb, add logic to handle file names of the form
UnicodeData-12.0.0d6.txt. To find the right file, we download
the index of the directory. Then we download the files by finding
the file names from the index. Files are always checked for changes,
because changes might be frequent during the beta period.
We also check whether any index.html files are left when we are not
in the beta period. This would indicate that we might have stale
data from the beta period rather than the actual release data.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@66737 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* Define major and minor version numbers only in the public
include/ruby/version.h header, as the API version numbers.
* Define only teeny version number in the private version.h
header.
* RUBY_VERSION moved to version.c.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@66599 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
These erb files are in fact erb comments + plain C. Adding #line
help us debug in case we have trouble there.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@66572 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
Just add more room for comments. This is a pure refactoring that does
not change anything but readability.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@66564 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* tool/rbinstall.rb (load_gemspec): do not hide syntax errors in
a gemspec file. check if the result instead.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@66460 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
The option --unicode-beta for tool/downloader.rb introduced in
r66448 must not be passed as an option to actual download machinery.
Thanks to MSP-Greg (Greg L.) for bug report and patch.
This closes issue #15434.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@66449 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
Unicode uses file names with explicit versions for beta publication.
This commit introduces a variable on the makefile level to distinguish
between beta and regular versions of file names.
common.mk: Define new variable UNICODE_BETA, usually set to NO,
but would be YES during tests with beta data files. Pass the value
of this variable to tool/downloader.rb with option --unicode-beta.
tool/downloader.rb: Receive and store value of --unicode-beta.
Raise an exception if value is YES, because we don't yet actually
deal with this case. Continue as usual if value is not YES.
This completes the changes needed in the makefile. Upcomming changes
will only affect tool/downloader.rb.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@66448 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* tool/vcs.rb (VCS.detect): limit level of travarsing parent
directories, 0 by the default. curretly always detecting at the
source directory itself.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@66305 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
Today I updated my ruby core clone, and I run the usual steps to compile
ruby. When running "make", I run into the following error:
```
(... stuff ...)
encdb.h unchanged
Traceback (most recent call last):
5: from ./tool/mkconfig.rb:19:in `<main>'
4: from ./tool/mkconfig.rb:19:in `require'
3: from /home/deivid/Code/ruby/lib/fileutils.rb:4:in `<top (required)>'
2: from /home/deivid/Code/ruby/lib/fileutils.rb:4:in `require'
1: from /home/deivid/Code/ruby/rbconfig.rb:11:in `<top (required)>'
/home/deivid/Code/ruby/rbconfig.rb:13:in `<module:RbConfig>': ruby lib version (2.5.0) doesn't match executable version (2.6.0) (RuntimeError)
uncommon.mk:780: recipe for target '.rbconfig.time' failed
make: *** [.rbconfig.time] Error 1
```
Apparently, the script that generates the root `rbconfig.rb` file requires
`fileutils`, which in turn requires 'rbconfig' (for mjit-headers it says in a
comment), which uses the `rbconfig.rb` config file in the root folder if it
exists. In my case, this file existed but had been generated on 2.5.0, thus
causing the error.
I think we can avoid this sort of circular dependency by not requiring
`fileutils`, since it does not seem to be used anywhere in the
`tool/mkconfig.rb` script since r55338.
[Fix GH-2045]
From: David Rodríguez <deivid.rodriguez@riseup.net>
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@66301 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* vm_trace.c (rb_tracepoint_enable_for_target): support targetting
TracePoint. [Feature #15289]
Tragetting TracePoint is only enabled on specified method, proc
and so on, example: `tp.enable(target: code)`.
`code` should be consisted of InstructionSeuqnece (iseq)
(RubyVM::InstructionSeuqnece.of(code) should not return nil)
If code is a tree of iseq, TracePoint is enabled on all of
iseqs in a tree.
Enabled tragetting TracePoints can not enabled again with
and without target.
* vm_core.h (rb_iseq_t): introduce `rb_iseq_t::local_hooks`
to store local hooks.
`rb_iseq_t::aux::trace_events` is renamed to
`global_trace_events` to contrast with `local_hooks`.
* vm_core.h (rb_hook_list_t): add `rb_hook_list_t::running`
to represent how many Threads/Fibers are used this list.
If this field is 0, nobody using this hooks and we can
delete it.
This is why we can remove code from cont.c.
* vm_core.h (rb_vm_t): because of above change, we can eliminate
`rb_vm_t::trace_running` field.
Also renamed from `rb_vm_t::event_hooks` to `global_hooks`.
* vm_core.h, vm.c (ruby_vm_event_enabled_global_flags): renamed
from `ruby_vm_event_enabled_flags.
* vm_core.h, vm.c (ruby_vm_event_local_num): added to count
enabled targetting TracePoints.
* vm_core.h, vm_trace.c (rb_exec_event_hooks): accepts
hook list.
* vm_core.h (rb_vm_global_hooks): added for convinience.
* method.h (rb_method_bmethod_t): added to maintain Proc
and `rb_hook_list_t` for bmethod (defined by define_method).
* prelude.rb (TracePoint#enable): extracet a keyword parameter
(because it is easy than writing in C).
It calls `TracePoint#__enable` internal method written in C.
* vm_insnhelper.c (vm_trace): check also iseq->local_hooks.
* vm.c (invoke_bmethod): check def->body.bmethod.hooks.
* vm.c (hook_before_rewind): check iseq->local_hooks
and def->body.bmethod.hooks before rewind by exception.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@66003 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* tool/vcs.rb (VCS::GIT#export_changelog): improve the
compatibility with svn-log.
remained differences are:
- in svn-log
- accented characters, left/right single quotation marks, and
non-break spaces are translated to ASCII characters
- other non-ASCII characters are excoded as `{U+XXXX}`
- in git-log
- tabs are expanded
- in git-log (intentional)
- lines looking too indented are unindented
- empty lines between headers and bodies are squeezed
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@65970 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
redirection to a tty, file or pipe is not a permanent status.
`rb_cv_` prefix means that it should be saved/restored across
re-configurations.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@65801 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
To reduce the amount of output, prefer --tty=no instead of
--color=never. This option not only disables color output but also
kill some tty-related features, like spinners. Travis limits its
output by the physical size of the log, not by the number of lines.
This change should make more room for new logs.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@65702 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* bin/*, lib/bundler/*, lib/bundler.rb, spec/bundler, man/*:
Merge from latest stable branch of bundler/bundler repository and
added workaround patches. I will backport them into upstream.
* common.mk, defs/gmake.mk: Added `test-bundler` task for test suite
of bundler.
* tool/sync_default_gems.rb: Added sync task for bundler.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@65509 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
because it's not supported by this file. Also, shared `def_iseq_ptr`
instead of copying the main definition of it.
vm_core.h: moved `def_iseq_ptr` to this place. added `inline` to avoid
compiler warnings since it's not used in some files including vm_core.h.
vm_insnhelper.c: moved `def_iseq_ptr` to vm_core.h.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@65440 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
The idea behind this commit is that handles_sp and leaf are two
concepts that are not mutually independent. By making one explicitly
depend another, we can reduces the number of lines of codes written,
thus making things concise.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@65426 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
This enhances stability of the generated source code (namely
insns_info.inc) across attribute insertion / deletion. It does
not change the compiled binary at all; just a bit of readability.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@65425 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
I assumed somehow this check was not needed, but it did need.
By canceling this instead of just warning here, we didn't lose the
current performance so much.
test_jit.rb: test the case that reproduces SEGV by that.
TestGemStreamUI.rb: delete. This test on --jit-wait is fixed.
=== Optcarrot Benchmark ===
$ benchmark-driver benchmark.yml --rbenv 'before::before --disable-gems --jit;after::after --disable-gems --jit' -v --repeat-count 24
before: ruby 2.6.0dev (2018-10-24 trunk 65355) +JIT [x86_64-linux]
after: ruby 2.6.0dev (2018-10-24 trunk 65355) +JIT [x86_64-linux]
last_commit=_mjit_compile_ivar.erb: cancel on undefined ivar
Calculating -------------------------------------
before after
Optcarrot Lan_Master.nes 85.344 84.849 fps
Comparison:
Optcarrot Lan_Master.nes
before: 85.3 fps
after: 84.8 fps - 1.01x slower
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@65356 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
to another file, because it's no longer shared. It was created when
attr_reader was inlined but it's no longer included.
common.mk: ditto
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@65276 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
on VM_CHECK_INTS. Letting MJIT worker directly see inline cache which
may be being updated could result in inconsistent IC index and serial.
mjit_worker.c: request the copy job after dequeue, and receive the
result synchronously.
tool/ruby_vm/views/_mjit_compile_ivar.erb: use the copied IC
mjit_compile.c: change the interface to pass is_entries
mjit.h: ditto
=== Optcarrot Benchmark ===
Thankfully this didn't have major performance regression.
$ benchmark-driver benchmark.yml --rbenv 'before::before --disable-gems --jit;after::after --disable-gems --jit' -v --repeat-count 24
before: ruby 2.6.0dev (2018-10-21 trunk 65263) +JIT [x86_64-linux]
after: ruby 2.6.0dev (2018-10-21 trunk 65263) +JIT [x86_64-linux]
last_commit=mjit.c: copy inline cache values to MJIT worker
Calculating -------------------------------------
before after
Optcarrot Lan_Master.nes 85.421 85.454 fps
Comparison:
Optcarrot Lan_Master.nes
after: 85.5 fps
before: 85.4 fps - 1.00x slower
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@65275 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* configure.ac: rename configure option `--disable-mjit` to
`--disable-mjit-support` because `--disable-mjit` is ambiguous that
runtime MJIT default enable option or supporting MJIT features.
`ENABLE_MJIT` is also renamed to `MJIT_SUPPORT`
* Makefile.in: catch up this fix.
* common.mk: ditto.
* test/ruby/test_jit.rb: ditto.
* win32/Makefile.sub: catch up this fix on mswin.
* tool/mkconfig.rb: fix to pass `MJIT_SUPPORT` key.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@65217 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
We have several options to ensure there's no race condition between main
thread and MJIT thead about IC reference:
1) Give up caching ivar for multiple classes (or multiple versions of the
same class) in the same getinstancevariable (This commit's approach)
2) Allocate new inline cache every time
Other ideas we could think of couldn't eliminate possibilities of race
condition.
In 2, it's memory allocation would be slow and it may trigger JIT
cancellation frequently. So 1 would be fast for both VM and JIT
situations.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@65213 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* tool/transform_mjit_header.rb: insert GCC system_header pragma
to get rid of errors caused by the contents included from system
headers.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@65169 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
`ic->ic_value.index` is size_t, not always unsigned long.
See also: https://travis-ci.org/ruby/ruby/jobs/443560810#L2263
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@65166 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
in the MJIT-header-specific path, not default path like vc140.pdb.
mjit_worker.c: specify the MJIT-header-specific pdb path.
tool/rbinstall.rb: install MJIT header pdb as well.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@65003 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
to prefix. This is a retry of r64947. So this doesn't still make mswin MJIT
on install directory succeed. One more step required.
tool/rbinstall.rb: This change is needed to install headers correctly since
the extensions are .obj and .pch, not .h
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@65000 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
Recent apple machines describe themselves being x86_64h. That
architecture is somehow supported by their C compiler and at least
by recent clang. However config.sub does not know that fact so
making universal binary targeting it is rejected by the program.
Why not skip the check by config.sub. [fix GH-1971]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@64905 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
`make install` has loaded forwardable.rb twice, from
forwardable.gemspec and prime.gemspec.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@64882 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
because r64849 seems to fix issues which we were confused about.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@64850 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
This reverts commit r64829. I'll prepare another temporary fix, but I'll
separately commit that to make it easier to revert that later.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@64838 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
not optimizing Array#& and Array#| because vm_insnhelper.c can't easily
inline it (large amount of array.c code would be needed in vm_insnhelper.c)
and the method body is a little complicated compared to Integer's ones.
So I thought only Integer#& and Integer#| have a significant impact,
and eliminating unnecessary branches would contribute to JIT's performance.
vm_insnhelper.c: ditto
tool/transform_mjit_header.rb: make sure these instructions are inlined
on JIT.
compile.c: compile vm_opt_and and vm_opt_or.
id.def: define id for them to be used in compile.c and vm*.c
vm.c: track redefinition of Integer#& and Integer#|
vm_core.h: allow detecting redefinition of & and |
test/ruby/test_jit.rb: test new insns
test/ruby/test_optimization.rb: ditto
* Optcarrot benchmark
This is a kind of experimental thing but I'm committing this since the
performance impact is significant especially on Optcarrot with JIT.
$ benchmark-driver benchmark.yml --rbenv 'before::before --disable-gems;before+JIT::before --disable-gems --jit;after::after --disable-gems;after+JIT::after --disable-gems --jit' -v --repeat-count 24
before: ruby 2.6.0dev (2018-09-24 trunk 64821) [x86_64-linux]
before+JIT: ruby 2.6.0dev (2018-09-24 trunk 64821) +JIT [x86_64-linux]
after: ruby 2.6.0dev (2018-09-24 opt_and 64821) [x86_64-linux]
last_commit=opt_or
after+JIT: ruby 2.6.0dev (2018-09-24 opt_and 64821) +JIT [x86_64-linux]
last_commit=opt_or
Calculating -------------------------------------
before before+JIT after after+JIT
Optcarrot Lan_Master.nes 51.460 66.315 53.023 71.173 fps
Comparison:
Optcarrot Lan_Master.nes
after+JIT: 71.2 fps
before+JIT: 66.3 fps - 1.07x slower
after: 53.0 fps - 1.34x slower
before: 51.5 fps - 1.38x slower
[close https://github.com/ruby/ruby/pull/1963]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@64824 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
by sharing vm_call_iseq_setup_normal. This is a retry of r64280.
vm_insnhelper.c: Remove unused argument `ci` and pass `me` instead of
`cc` to share this with JIT. Declare this with ALWAYS_INLINE to make
sure this function is inlined in JIT.
tool/mk_call_iseq_optimized.rb: deal with the interface change of
vm_call_iseq_setup_normal.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@64820 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* tool/darwin-cc: ld in Apple's recent Xcode warns text-based stub
files, which are probably caused by Apple's broken package.
hide such (and architecture deprecation) warnings during
configuration to pass TRY_LDFLAGS.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@64802 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
This error does not happen right now so this typo is
not serious, unless you locally edit insns.def.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@64791 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
because it's actually setting fastpath to cc instead of ci since r51903.
vm_insnhelper.c: ditto
mjit_compile.c: ditto
tool/ruby_vm/views/_mjit_compile_send.erb: ditto
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@64772 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
tool/update-deps: tweak the comment to make sure it should be built in
the source directory, because building ruby outside source directory
failed on my trial.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@64764 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
Now that we can say for sure if an instruction calls a method or
not internally, it is now possible to reroute the bugs that
forced us to revert the "move PC around" optimization.
First try: r62051
Reverted: r63763
See also: r63999
----
trunk: ruby 2.6.0dev (2018-09-13 trunk 64736) [x86_64-darwin15]
ours: ruby 2.6.0dev (2018-09-13 trunk 64736) [x86_64-darwin15]
last_commit=move ADD_PC around (take 2)
Calculating -------------------------------------
trunk ours
so_ackermann 1.884 2.278 i/s - 1.000 times in 0.530926s 0.438935s
so_array 1.178 1.157 i/s - 1.000 times in 0.848786s 0.864467s
so_binary_trees 0.176 0.177 i/s - 1.000 times in 5.683895s 5.657707s
so_concatenate 0.220 0.221 i/s - 1.000 times in 4.546896s 4.518949s
so_count_words 6.729 6.470 i/s - 1.000 times in 0.148602s 0.154561s
so_exception 3.324 3.688 i/s - 1.000 times in 0.300872s 0.271147s
so_fannkuch 0.546 0.968 i/s - 1.000 times in 1.831328s 1.033376s
so_fasta 0.541 0.547 i/s - 1.000 times in 1.849923s 1.827091s
so_k_nucleotide 0.800 0.777 i/s - 1.000 times in 1.250635s 1.286295s
so_lists 2.101 1.848 i/s - 1.000 times in 0.475954s 0.541095s
so_mandelbrot 0.435 0.408 i/s - 1.000 times in 2.299328s 2.450535s
so_matrix 1.946 1.912 i/s - 1.000 times in 0.513872s 0.523076s
so_meteor_contest 0.311 0.317 i/s - 1.000 times in 3.219297s 3.152052s
so_nbody 0.746 0.703 i/s - 1.000 times in 1.339815s 1.423441s
so_nested_loop 0.899 0.901 i/s - 1.000 times in 1.111767s 1.109555s
so_nsieve 0.559 0.579 i/s - 1.000 times in 1.787763s 1.726552s
so_nsieve_bits 0.435 0.428 i/s - 1.000 times in 2.296282s 2.333852s
so_object 1.368 1.442 i/s - 1.000 times in 0.731237s 0.693684s
so_partial_sums 0.616 0.546 i/s - 1.000 times in 1.623592s 1.833097s
so_pidigits 0.831 0.832 i/s - 1.000 times in 1.203117s 1.202334s
so_random 2.934 2.724 i/s - 1.000 times in 0.340791s 0.367150s
so_reverse_complement 0.583 0.866 i/s - 1.000 times in 1.714144s 1.154615s
so_sieve 1.829 2.081 i/s - 1.000 times in 0.546607s 0.480562s
so_spectralnorm 0.524 0.558 i/s - 1.000 times in 1.908716s 1.792382s
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@64737 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
This reverts commit r64711, because EXEC_EC_CFP on JIT-ed code does not
call jit_func with the patch when catch_except_p is true. It wasn't intentional.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@64730 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
This is mostly cosmetic. Should generate a slightly readable
vm.inc output.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@64709 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
if and only if it's not moved yet, to avoid potential bugs in the future.
_mjit_compile_send.erb: ditto
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@64687 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
to raise descriptive KeyError instead of NoMethodError in case these
attrs are accidentally removed.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@64685 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
rather than `trace_enablable_insns` which is hard to maintain.
This would make performance bad because new branches for tracepoint will
be added.
Optcarrot before:
fps: 56.885371547337655
fps: 60.02493636060194
fps: 63.051028327122076
fps: 63.78463316242535
fps: 64.20391937940403
fps: 64.55990344731123
fps: 64.56771099162921
fps: 64.95991277629723
fps: 65.15120708973232
fps: 65.90558702393933
fps: 66.29579283026303
Optcarrot after:
fps: 52.7647027470875
fps: 53.67404855529564
fps: 58.40514319229468
fps: 60.90736996487708
fps: 62.83487236283472
fps: 63.01386139447994
fps: 63.42395443471596
fps: 63.78328559878602
fps: 64.58432081229746
fps: 64.78720429848532
fps: 65.48720618907552
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@64684 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
when catch_except_p is false and insn.always_leaf? is true (never makes
arbitrary method call in the insn).
On Optcarrot, unfortunately this didn't have measureable performance impact.
But still this is a good direction since it becomes much faster when
marking all insns as always leaf.
bare_instructions.rb: add `#always_leaf?` that indicates the insn can
always be considered as leaf. Using dynamic leaf for JIT would be hard
since it requires to discard outdated code somehow.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@64683 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
as well, to make CI succeed with VM_CHECK_MODE > 1.
vm_insnhelper.c: drop unnecessary MJIT_HEADER ifdef. This is intended to
be ignored by having `static inline`. Removing that by macro would be
helpful for minimizing compilation time, but the impact is not so big
and having many MJIT_HEADER check would be bad for maintainability.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@64682 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
An instruction is leaf if it has no rb_funcall inside. In order to
check this property, we introduce stack canary which is a random
number collected at runtime. Stack top is always filled with this
number and checked for stack smashing operations, when VM_CHECK_MODE.
[GH-1947]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@64677 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
[ruby-core:88699][Bug #15035]
This patch was provided by MSP-Greg.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@64585 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
GitHub download failed on 13:50:36
https://ci.appveyor.com/project/ruby/ruby/build/1.0.9221 and it also
failed on 13:51:35 (all builds between them failed too).
It means that we need to expect GitHub 502 that continues 1 minute.
So I configured 6 retries, that will sleep at most 91s in total.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@64317 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
This commit caused test-all failure with --jit-wait.
I don't know the reason yet, but let me revert it to normalize CI.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@64314 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
to fix the wrong line number on #caller_locations or rb_profile_frames.
Actually we would be able to move it only when method call (of
caller_locations) or C extension invocation (calling rb_profile_frames)
can happen.
This degrades performance. Optcarrot fps becomes...
before: 71.78976052783555
after: 67.65429356624131
I think I can lazily move it and fix the performance issue, even
improving the performance for the situation catch table exists.
But let me fix this bug first...
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@64283 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
_mjit_compile_send.erb: simplify code using the change
insns.def: adapt to the interface change
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@64281 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
implementation. This had no major performance impact by effort to keep
them inlined.
vm_insnhelper.c: ditto
mjit_compile.c: just update the comment about opt_pc=0 assumption
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@64280 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
between a type of return value and function identifier. Like:
```
rb_control_frame_t *
__attribute__ ((__fastcall__)) rb_vm_opt_struct_aref(rb_execution_context_t *ec, rb_control_frame_t *reg_cfp)
{
...
}
```
[Bug #14949]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@64138 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
because it's more suitable to describe the current behavior now.
tool/ruby_vm/models/bare_instructions.rb: ditto.
tool/ruby_vm/views/_insn_entry.erb: ditto.
tool/ruby_vm/views/_mjit_compile_insn_body.erb: ditto.
tool/ruby_vm/views/_mjit_compile_pc_and_sp.erb: ditto.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@64053 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
on r64031. Compiling vm_search_method_slowpath is very slow.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@64034 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* doc/*: Move entry to default gems category from standard library.
* lib/shell/*: Added `Shell::VERSION` and re-used it with @RELEASE_VERSION.
* test/shell/test_command_processor.rb: added missing require for test library.
* tool/sync_default_gems.rb: Support shell library. We need to ignore shellwords.rb when
syncing shell* files.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@64004 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
This optimization was reverted on r63863, but this commit resurrects the
optimization to skip some sp motions on JIT execution.
tool/ruby_vm/views/_mjit_compile_insn_body.erb: ditto
tool/ruby_vm/views/_mjit_compile_insn.erb: ditto
insns.def: resurrect handles_frame as handles_stack, which was deleted
on r63763.
tool/ruby_vm/models/bare_instructions.rb: ditto
vm_insnhelper.c: prevent moving sp outside insns.def to allow modifying
it by JIT.
* Optcarrot benchmark
$ benchmark-driver benchmark.yml --rbenv 'before --jit;after --jit' --repeat-count 12 -v
before --jit: ruby 2.6.0dev (2018-07-17 trunk 63987) +JIT [x86_64-linux]
after --jit: ruby 2.6.0dev (2018-07-17 local-stack 63987) +JIT [x86_64-linux]
last_commit=mjit_compile.c: resurrect local variable stack
Calculating -------------------------------------
before --jit after --jit
Optcarrot Lan_Master.nes 70.518 72.144 fps
Comparison:
Optcarrot Lan_Master.nes
after --jit: 72.1 fps
before --jit: 70.5 fps - 1.02x slower
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@63988 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* remove initial "./" from destdir to make downloading
./enc/unicode/data/11.0.0/ucd/auxiliary/GraphemeBreakProperty.txt
work properly
* tweak a comment
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@63978 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
This adds support for lldb as a debugger so that tests can be run under
lldb like this:
$ make test-all TESTS=test/some_test.rb RUNRUBYOPT=--debugger=lldb
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@63921 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
if the tag is already checked out, to optimize the execution time.
I'm going to prepare a task depending on this tool, and I want that to
finish fast and output nothing when it's already up-to-date.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@63883 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
r63655 was tightly coupled to handle_frames and some assumptions seems
to have been broken by r63763.
To partially resolve Bug#14892, this reverts the optimization for now. I
want to make MJIT CI happy first and then I'll probably retry r63655 by
partially reverting r63763 for sp changes.
The skipped test is not fixed yet.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@63863 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
I introduced this mechanism in r62051 to speed things up. Later it
was reported that the change causes problems. I searched for
workarounds but nothing seemed appropriate. I hereby officially
give it up. The idea to move ADD_PC around was a mistake.
Fixes [Bug #14809] and [Bug #14834].
Signed-off-by: Urabe, Shyouhei <shyouhei@ruby-lang.org>
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@63763 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
It never becomes `dispatched: true` with the current code.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@63751 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* configure.ac: removed DISABLE_RUBYGEMS macro from config.h, not
to rebuild everything when the flag changed.
* configure.ac, win32/configure.bat: make USE_RUBYGEMS lowercase.
* tool/mkconfig.rb: remove RUBYGEMS stuff from rbconfig.rb, not to
reconfigure and rebuild all extension libraries.
* Makefile.in (CPPFLAGS): enable/disable Rubygems by USE_RUBYGEMS.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@63718 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
This reverts commit 9d015aa91cc3ec45e41be58fd836fb7f6655a624.
Because tool/extlibs.rb runs with BASERUBY that may be ruby < 2.3
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@63617 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
This allows user to specify any name in `--with-so-name` that might
cause a name clash with LIBRUBY_ALIASES on the platform.
Without this, for example, configuring with `--with-soname=ruby
--enable-shared` on macOS would end up running `ln -sf libruby.dylib
libruby.dylib` only to fail with the following error in installation:
```
make[2]: stat: libruby.dylib: Too many levels of symbolic links
```
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@63544 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* tool/redmine-backporter.rb (StringScanner.readline): use `true`
to suppress a "literal in condition" warning.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@63458 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
for now, which was committed in r63333.
Currently trunk's JIT seems to have some wrong behaviors, and this is
the most suspicious culprit of them for now. In the future, I may have
a strict test environment to detect the cause, but there's no enough
time to test this until preview2. So let me revert this and see how it
goes after this.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@63455 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* tool/generic_erb.rb: get rid of FileUtils.touch, not to depend
on fileutils.rb which will depend on rbconfig.rb which does not
exist when creating encdb.h.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@63432 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* tool/transform_mjit_header.rb (MJITHeader.conflicting_types?):
Add workaround for Solaris 10 with old GCC (4.6.2), that is
essentially the same as for AIX (commit r62326), but probably
due to different GCC versions, different error message is shown.
[Bug #14751] [ruby-dev:50541]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@63400 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
reverting r63379 (revert of r63360).
The cause of error seems to be r63350. See r63382.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@63384 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
Revert "_mjit_compile_send.erb: inline recursive call"
I reverted r63360 in r63379, but the errors were reproductive from
r63350. So I need to revert this.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@63382 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
Revert "_mjit_compile_pc_and_sp.erb: make sure no uninitialized"
This triggered some `NoMethodError`s which seem to be caused by the
commit like: https://travis-ci.org/k0kubun/mjit-test/builds/376416934
I'll add tests and fix it later...
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@63379 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
area on VM stack to prevent SEGV on GC. GC may mark every value in VM
stack.
Unfortunately I couldn't write a test for it... So let me explain the
situation.
SEGV example:
https://gist.github.com/k0kubun/c7cea2b5761ffdff29ec79ea1a8f7f91
```
$ ruby --dump=insns -e 'def oct(num, len); "%0#{len}o" % num; end'
== disasm: #<ISeq:oct@-e:1 (1,0)-(1,41)> (catch: FALSE)
local table (size: 2, argc: 2 [opts: 0, rest: -1, post: 0, block: -1, kw: -1@-1, kwrest: -1])
[ 2] num@0<Arg> [ 1] len@1<Arg>
0000 putobject "%0" ( 1)[LiCa]
0002 getlocal_WC_0 len@1
0004 dup
0005 checktype T_STRING
0007 branchif 14
0009 dup
0010 opt_send_without_block <callinfo!mid:to_s, argc:0, FCALL|ARGS_SIMPLE>, <callcache>
0013 tostring
0014 putobject "o"
0016 concatstrings 3
0018 getlocal_WC_0 num@0
0020 opt_mod <callinfo!mid:%, argc:1, ARGS_SIMPLE>, <callcache>
0023 leave [Re]
```
Prior to this commit, after arguments are pushed on 0010, stacks were:
VM stack: [uninitialized, uninitialized, len]
JIT stack: ["%0", len, len]
And then, when GC is invoked on 0016, VM stack will be [uninitialized, uninitialized]
and those uninitialized values will be marked by GC.
With this commit, after arguments are pushed on 0010, stacks will be:
VM stack: [len]
JIT stack: ["%0", len, len]
And VM stack will be [] on 0016.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@63360 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
This reverts r63249 (revert r63212) and fixes a bug in it. The test to
prevent the bug is added as well.
vm_insnhelper.c: add `index` argument to vm_getivar. The argument is
created so that MJIT can pass the value of `cc->aux.index` on compilation
time. The cache invalidation in _mjit_compile_send_guard.erb is only
working for the cache value on compilation time.
Note: As `index` is always passed as constant and it's force-inlined,
the performance of `vm_getivar` won't be degraded in VM.
_mjit_compile_send_guard.erb: New. Used to invalidate inlined values of cc.
common.mk: update dependencies for _mjit_compile_send_guard.erb
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@63333 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
between branches.
mjit_compile.inc.erb: move the compiled_for_pos reference to
mjit_compile.c
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@63331 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
on JIT compilation. r63092 was risky without this check.
mjit_compile.c: update comment about stack consistency check
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@63329 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* tool/rbinstall.rb ($script_installer.stub): read stub file on
demand. as `$cmdtype` is set to "exe" in parse_args, it is not
set yet when `$script_installer` is defined.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@63268 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
except test_jit.rb.
In some situations, this generates a wrong code. I'll add a test for it
later but let me revert this to make it work for now.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@63249 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
to introduce additional optimization for another `cc->me->def->type`
later. I carved out the `cc->me->def->type == VM_METHOD_TYPE_ISEQ`
part because I wanted to check other types as well.
mjit_compile.c: drop get_iseq_if_available and define simplified version
of it, has_valid_method_type.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@63199 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
Takashi Kokubun
Nobuyoshi Nakada
aycabta
Urabe, Shyouhei
Kazuhiro NISHIYAMA
k0kubun
aycabta
hsbt
nobu
duerst
svn
ko1
nagachika
kazu
shyouhei
naruse
usa
normal
tenderlove
knu
ngoto
mame
stomar