Page MenuHomePhabricator

ygao (Yunzhong Gao)
User

Projects

User does not belong to any projects.

User Details

User Since
May 9 2013, 11:09 AM (341 w, 10 h)

Recent Activity

Apr 19 2017

ygao added inline comments to D32162: Inline asm 0bH conflict.
Apr 19 2017, 11:22 AM

Oct 12 2016

ygao committed rL284080: [NFC] Fixing the description for _mm_store_ps and _mm_store_ps1..
[NFC] Fixing the description for _mm_store_ps and _mm_store_ps1.
Oct 12 2016, 4:54 PM
ygao committed rL284033: [NFC] Trial change to remove a redundant blank line..
[NFC] Trial change to remove a redundant blank line.
Oct 12 2016, 12:42 PM
ygao closed D1759: Fixing Intel format of the vshufpd instruction.
Oct 12 2016, 12:16 AM
ygao accepted D1759: Fixing Intel format of the vshufpd instruction.

Closed by rL191481

Oct 12 2016, 12:16 AM
ygao closed D1750: [1/2] Add intrinsics for TBM instruction set.
Oct 12 2016, 12:12 AM
ygao accepted D1750: [1/2] Add intrinsics for TBM instruction set.

Closed by rL191539.

Oct 12 2016, 12:11 AM

Oct 5 2016

ygao committed rL283374: Improve the debug-info test created in r274263..
Improve the debug-info test created in r274263.
Oct 5 2016, 1:35 PM
ygao closed D25290: Improve the debug-info test created in r274263 by committing rL283374: Improve the debug-info test created in r274263..
Oct 5 2016, 1:35 PM
ygao retitled D25290: Improve the debug-info test created in r274263 from to Improve the debug-info test created in r274263.
Oct 5 2016, 11:54 AM

Sep 2 2016

ygao committed rL280556: (clang part) Implement MASM-flavor intel syntax behavior for inline MS asm….
(clang part) Implement MASM-flavor intel syntax behavior for inline MS asm…
Sep 2 2016, 4:24 PM
ygao committed rL280555: (LLVM part) Implement MASM-flavor intel syntax behavior for inline MS asm block:.
(LLVM part) Implement MASM-flavor intel syntax behavior for inline MS asm block:
Sep 2 2016, 4:23 PM
ygao closed D22112: Disambiguate a constant with both 0B prefix and H suffix. by committing rL280555: (LLVM part) Implement MASM-flavor intel syntax behavior for inline MS asm block:.
Sep 2 2016, 4:23 PM

Sep 1 2016

ygao added inline comments to D22112: Disambiguate a constant with both 0B prefix and H suffix..
Sep 1 2016, 12:14 PM
ygao updated the diff for D22112: Disambiguate a constant with both 0B prefix and H suffix..
Sep 1 2016, 12:13 PM

Aug 31 2016

ygao added inline comments to D22112: Disambiguate a constant with both 0B prefix and H suffix..
Aug 31 2016, 6:20 PM
ygao updated the diff for D22112: Disambiguate a constant with both 0B prefix and H suffix..
Aug 31 2016, 6:19 PM

Aug 26 2016

ygao added inline comments to D22112: Disambiguate a constant with both 0B prefix and H suffix..
Aug 26 2016, 7:54 PM
ygao updated the diff for D22112: Disambiguate a constant with both 0B prefix and H suffix..
Aug 26 2016, 7:48 PM

Aug 23 2016

ygao added a comment to D22112: Disambiguate a constant with both 0B prefix and H suffix..

A gentle ping.

Aug 23 2016, 4:33 PM

Aug 21 2016

ygao added inline comments to D23761: clang-format: [JS] supports casts to types starting with punctuation ("{[(")..
Aug 21 2016, 4:11 PM

Aug 15 2016

ygao added inline comments to D22112: Disambiguate a constant with both 0B prefix and H suffix..
Aug 15 2016, 8:07 PM

Aug 8 2016

ygao added a comment to D22112: Disambiguate a constant with both 0B prefix and H suffix..

A gentle ping.

Aug 8 2016, 12:15 PM

Aug 2 2016

ygao updated the diff for D22112: Disambiguate a constant with both 0B prefix and H suffix..

Hi, I am updating the patch based on the latest feedback on PR27884.
This patch attempts to distinguish between the case of parsing MS inline asm blocks and that of parsing GNU inline asm. And in the former case, it implements the MASM-flavor intel-assembly parsing. I am adding a flag to the AsmLexer class.
There are a handful of MS inline asm test in clang that I need to modify for this patch. This review will be committed in one LLVM patch and one clang patch.

Aug 2 2016, 4:38 PM

Jul 29 2016

ygao committed rL277192: Improve documentation of the type safety attributes..
Improve documentation of the type safety attributes.
Jul 29 2016, 11:42 AM
ygao closed D22717: Improve documentation of the type safety attributes by committing rL277192: Improve documentation of the type safety attributes..
Jul 29 2016, 11:42 AM

Jul 22 2016

ygao retitled D22717: Improve documentation of the type safety attributes from to Improve documentation of the type safety attributes.
Jul 22 2016, 6:30 PM

Jul 19 2016

ygao committed rL276070: Fixing a few places in this doc which look like obvious typos..
Fixing a few places in this doc which look like obvious typos.
Jul 19 2016, 5:48 PM

Jul 18 2016

ygao committed rL275877: Support -masm= flag for x86 assembly targets..
Support -masm= flag for x86 assembly targets.
Jul 18 2016, 12:00 PM
ygao closed D22285: Support -masm= flag for x86 assembly targets.

Closed by rL275877.

Jul 18 2016, 11:57 AM
ygao added inline comments to D22285: Support -masm= flag for x86 assembly targets.
Jul 18 2016, 11:57 AM

Jul 12 2016

ygao retitled D22285: Support -masm= flag for x86 assembly targets from to Support -masm= flag for x86 assembly targets.
Jul 12 2016, 3:42 PM

Jul 11 2016

ygao added a comment to D22112: Disambiguate a constant with both 0B prefix and H suffix..

I think what you said makes a lot of sense to me... The practical difficulty
here is that if I actually disabled "0b" prefix under the Intel syntax, I would
be without a way to express binary numbers (sad face). Maybe I can implement
something? The Intel instruction manual talks of using [01]+[bB], but it looks
sufficiently similar to a backward label, and I could not figure out how the
Intel assembler actually tells them apart. For example,

Jul 11 2016, 11:25 AM

Jul 8 2016

ygao updated subscribers of D22112: Disambiguate a constant with both 0B prefix and H suffix..

Since the parsing style is specified on the command line at startup, can a
flag be attached to MCAsmInfo which specifies whether prefix-bool parsing is
activated? It looks like there are similar things already put there such as
CommentString and DollarIsPC

Jul 8 2016, 2:01 PM

Jul 7 2016

ygao retitled D22112: Disambiguate a constant with both 0B prefix and H suffix. from to Disambiguate a constant with both 0B prefix and H suffix..
Jul 7 2016, 2:29 PM

Jun 30 2016

ygao committed rL274263: Add an artificial line-0 debug location when the compiler emits a call to.
Add an artificial line-0 debug location when the compiler emits a call to
Jun 30 2016, 11:56 AM
ygao closed D21818: Add artificial debug information to avoid compiler crash by committing rL274263: Add an artificial line-0 debug location when the compiler emits a call to.
Jun 30 2016, 11:56 AM

Jun 28 2016

ygao added a comment to D21818: Add artificial debug information to avoid compiler crash.

Hi @aprantl,
I looked at r267904 per your suggestion. Thanks. It looks like it is a cfe patch, and ApplyDebugLocation::CreateArtificial() is defined in clang. Is it okay to use clang functions in the LLVM backend? This particular stackprotector pass is in the llvm backend, and I am under the impression that it is not allowed to use clang functions, but I could be wrong,

  • Gao
Jun 28 2016, 6:01 PM
ygao retitled D21818: Add artificial debug information to avoid compiler crash from to Add artificial debug information to avoid compiler crash.
Jun 28 2016, 3:28 PM

Jun 2 2016

ygao added a comment to D20549: [lit] Use os.devnull instead of named temp files.

Hi, sorry for the delay.
I tested your change on a 64-bit Windows 7 machine and it works fine. It seems a good idea to use os.devnull instead of a temporary file. Thanks.
I am not very familiar with the code myself to give a LGTM, so a second review would be appreciated.

Jun 2 2016, 11:51 AM

Apr 18 2016

ygao updated subscribers of D16427: FastISel should flush local value map.
Apr 18 2016, 12:23 PM

Apr 4 2016

ygao accepted D18708: Set C99 as default C Standard for PS4 target.

The test LGTM. Thanks!

Apr 4 2016, 2:42 PM
ygao committed rL265325: Add copyright notice to the modulemap file..
Add copyright notice to the modulemap file.
Apr 4 2016, 11:51 AM
ygao closed D18709: Add copyright notice to modulemap files by committing rL265325: Add copyright notice to the modulemap file..
Apr 4 2016, 11:51 AM

Apr 1 2016

ygao added inline comments to D18708: Set C99 as default C Standard for PS4 target.
Apr 1 2016, 5:41 PM
ygao retitled D18709: Add copyright notice to modulemap files from to Add copyright notice to modulemap files.
Apr 1 2016, 1:18 PM

Mar 29 2016

ygao added a comment to D18533: Fixing PR26558: the adc builtins do not require the adx target attribute.

Yep, only the ADCX instruction needs adx. This LGTM, but please just
remove "-target-feature +adx" from the run line instead of adding a
second run line.

Mar 29 2016, 4:05 PM
ygao committed rL264801: Fixing PR26558: remove the adx target attribute requirement from adc builtins..
Fixing PR26558: remove the adx target attribute requirement from adc builtins.
Mar 29 2016, 4:04 PM
ygao closed D18533: Fixing PR26558: the adc builtins do not require the adx target attribute by committing rL264801: Fixing PR26558: remove the adx target attribute requirement from adc builtins..
Mar 29 2016, 4:04 PM

Mar 28 2016

ygao retitled D18533: Fixing PR26558: the adc builtins do not require the adx target attribute from to Fixing PR26558: the adc builtins do not require the adx target attribute.
Mar 28 2016, 5:07 PM

Mar 7 2016

ygao added inline comments to D17066: libLTO: add a ThinLTOCodeGenerator on the model of LTOCodeGenerator..
Mar 7 2016, 11:30 AM
ygao updated subscribers of D17066: libLTO: add a ThinLTOCodeGenerator on the model of LTOCodeGenerator..
Mar 7 2016, 11:09 AM

Mar 1 2016

ygao added inline comments to D17682: [X86] AMD Bobcat CPU (btver1) doesn't support XSAVE .
Mar 1 2016, 12:36 PM

Feb 12 2016

ygao committed rL260764: Disable the vzeroupper insertion pass on PS4..
Disable the vzeroupper insertion pass on PS4.
Feb 12 2016, 3:42 PM
ygao closed D16837: Disable the vzeroupper insertion pass on PS4 by committing rL260764: Disable the vzeroupper insertion pass on PS4..
Feb 12 2016, 3:42 PM

Feb 3 2016

ygao added a comment to D16837: Disable the vzeroupper insertion pass on PS4.

I definitely remember there was some concern (or incident?) over correctness,
and it involves some library. Unfortunately I cannot recall the details.

Feb 3 2016, 11:16 AM

Feb 2 2016

ygao retitled D16837: Disable the vzeroupper insertion pass on PS4 from to Disable the vzeroupper insertion pass on PS4.
Feb 2 2016, 7:32 PM
ygao committed rL259615: Revert r259576: Disable the vzeroupper insertion pass on PS4..
Revert r259576: Disable the vzeroupper insertion pass on PS4.
Feb 2 2016, 5:29 PM
ygao committed rL259576: Disable the vzeroupper insertion pass on PS4..
Disable the vzeroupper insertion pass on PS4.
Feb 2 2016, 1:43 PM

Feb 1 2016

ygao added inline comments to D16588: [DebugInfo] Support zero-length CIE in the _eh_frame parser.
Feb 1 2016, 3:28 PM

Jan 29 2016

ygao closed D15705: Adding a scripted test for PR25717.

The test is commited in rL258898 and rL258902, and it is enabled on buildbot in rL259248.
Thanks!

Jan 29 2016, 2:17 PM
ygao committed rL259248: [llvmlab] Enable clang tests that output to the console when we are on the….
[llvmlab] Enable clang tests that output to the console when we are on the…
Jan 29 2016, 2:08 PM
ygao closed D16694: [llvmlab] Enable clang tests that output to the console when we are on the buildbots by committing rL259248: [llvmlab] Enable clang tests that output to the console when we are on the….
Jan 29 2016, 2:07 PM

Jan 28 2016

ygao added a child revision for D16694: [llvmlab] Enable clang tests that output to the console when we are on the buildbots: D15705: Adding a scripted test for PR25717.
Jan 28 2016, 2:27 PM
ygao added a parent revision for D15705: Adding a scripted test for PR25717: D16694: [llvmlab] Enable clang tests that output to the console when we are on the buildbots.
Jan 28 2016, 2:27 PM
ygao retitled D16694: [llvmlab] Enable clang tests that output to the console when we are on the buildbots from to [llvmlab] Enable clang tests that output to the console when we are on the buildbots.
Jan 28 2016, 2:27 PM

Jan 26 2016

ygao committed rL258902: Adding a test for PR25717..
Adding a test for PR25717.
Jan 26 2016, 6:31 PM
ygao closed D16000: Add "/dev/tty" as a special file name to lit.

Committed rL258898. Thanks!

Jan 26 2016, 5:54 PM
ygao committed rL258898: Add "/dev/tty" as a special file name for lit tests..
Add "/dev/tty" as a special file name for lit tests.
Jan 26 2016, 5:52 PM
ygao retitled D16000: Add "/dev/tty" as a special file name to lit from Add "/dev/console" as a special file name to lit to Add "/dev/tty" as a special file name to lit.
Jan 26 2016, 5:50 PM
ygao committed rL258850: Do not define GXX_RTTI macro for C..
Do not define GXX_RTTI macro for C.
Jan 26 2016, 12:19 PM
ygao closed D16365: Do not define GXX_RTTI macro for C by committing rL258850: Do not define GXX_RTTI macro for C..
Jan 26 2016, 12:19 PM

Jan 25 2016

ygao added a comment to D16365: Do not define GXX_RTTI macro for C.

A gentle ping.

Jan 25 2016, 4:59 PM
ygao added a comment to D15705: Adding a scripted test for PR25717.

Thanks, Rafael!
I did verify that the test would fail without my fix. But what do I need to do to make buildbots turn on the new lit parameter?

Jan 25 2016, 11:03 AM

Jan 22 2016

ygao updated the diff for D16000: Add "/dev/tty" as a special file name to lit.

Thanks a lot! The lit parameter "-param enable_console=1" works for me and it is much cleaner than what I had earlier. I took out those changes.
The change to lit.cfg to use the lit parameter is relevant to the cfe project and I am updating the related diff D15705.

Jan 22 2016, 6:12 PM
ygao updated the diff for D15705: Adding a scripted test for PR25717.
Jan 22 2016, 6:05 PM
ygao updated the diff for D16000: Add "/dev/tty" as a special file name to lit.

I wonder if we could add a 'REQUIRES: console" and have that be enable
with a command line option. That way it can be off by default and
still enabled in a bot.

Jan 22 2016, 4:09 PM

Jan 21 2016

ygao added a comment to D16000: Add "/dev/tty" as a special file name to lit.

We should agree that we are fine with having a test in check-all that writes to
the console (or is the test-suite using lit these days?)

Jan 21 2016, 11:53 AM

Jan 20 2016

ygao updated the diff for D16000: Add "/dev/tty" as a special file name to lit.

You are right, and I like your idea of using a special device file on Windows for accessing the console. Thanks!

Jan 20 2016, 1:44 PM
ygao retitled D16365: Do not define GXX_RTTI macro for C from to Do not define GXX_RTTI macro for C.
Jan 20 2016, 11:15 AM

Jan 19 2016

ygao added a reviewer for D16177: Adding missing intrinsics _cvtsh_ss and _cvtss_sh: craig.topper.
Jan 19 2016, 2:54 PM

Jan 8 2016

ygao updated the diff for D15705: Adding a scripted test for PR25717.
Jan 8 2016, 11:45 AM
ygao added a parent revision for D15705: Adding a scripted test for PR25717: D16000: Add "/dev/tty" as a special file name to lit.
Jan 8 2016, 11:09 AM
ygao retitled D16000: Add "/dev/tty" as a special file name to lit from to Add "/dev/console" as a special file name to lit.
Jan 8 2016, 11:08 AM

Jan 5 2016

ygao added a comment to D15553: PR25717: fatal IO error writing large outputs to console on Windows.

Ok, I had to modify the fix somewhat to make a mingw bot happy. The bot is
i686-mingw32-RA-on-linux.

Jan 5 2016, 7:21 PM
ygao committed rL256904: Do not define NOGDI. Mingw defines LOGFONTW type in wingdi.h and the mingw.
Do not define NOGDI. Mingw defines LOGFONTW type in wingdi.h and the mingw
Jan 5 2016, 7:04 PM
ygao committed rL256902: Another attempt at fixing the i686-mingw32-RA-on-linux buildbot. I am getting.
Another attempt at fixing the i686-mingw32-RA-on-linux buildbot. I am getting
Jan 5 2016, 6:52 PM
ygao committed rL256901: Another attempt at fixing the i686-mingw32-RA-on-linux buildbot..
Another attempt at fixing the i686-mingw32-RA-on-linux buildbot.
Jan 5 2016, 6:36 PM
ygao committed rL256896: Hopefully fix a mingw32 buildbot (i686-mingw32-RA-on-linux) which does not have.
Hopefully fix a mingw32 buildbot (i686-mingw32-RA-on-linux) which does not have
Jan 5 2016, 5:40 PM
ygao committed rL256892: Fixing PR25717: fatal IO error writing large outputs to console on Windows..
Fixing PR25717: fatal IO error writing large outputs to console on Windows.
Jan 5 2016, 4:53 PM
ygao closed D15553: PR25717: fatal IO error writing large outputs to console on Windows by committing rL256892: Fixing PR25717: fatal IO error writing large outputs to console on Windows..
Jan 5 2016, 4:53 PM
ygao added a comment to D15705: Adding a scripted test for PR25717.

In llvm/utils/lit/lit/TestRunner.py line#202 inside function executeShCmd(), the test is spawned by
subprocess.Popen(command, ..., stdout = subprocess.PIPE, stderr = subprocess.PIPE, ...)
And the test passes. I can verify with a small python script doing just a subprocess.Popen.

Jan 5 2016, 2:39 PM

Jan 4 2016

ygao added a comment to D15705: Adding a scripted test for PR25717.

What I would suggest is:

Check that this test fails at least on windows with your patch

reverted. If so, commit it. BTW, don't you need to drop the "|
FileCheck" to cause a crash?

Jan 4 2016, 5:00 PM

Dec 23 2015

ygao added a comment to D15705: Adding a scripted test for PR25717.

If you take a look at

void PrintPreprocessedAction::ExecuteAction() {
      ...
      while (next < end) {
      if (*cur == 0x0D) {  // CR
        if (*next == 0x0A)  // CRLF
          BinaryMode = false;
      ...
Dec 23 2015, 2:43 PM
ygao updated the diff for D15705: Adding a scripted test for PR25717.

I think what I can do is to run
svn propset svn:eol-style 'LF'
on the new file, and hopefully the subversion or git client will get the hint and not auto-translate the line endings.

Dec 23 2015, 12:22 PM
ygao added a comment to D15705: Adding a scripted test for PR25717.

Hi Rafael,
Yes I can do that, and it works for me. Do you know whether there is a way to keep the UNIX line ending on the test file upon checkout rather than having svn or git auto-translate the line endings? The UNIX line ending is necessary to trigger the original bug,

  • Gao
Dec 23 2015, 11:44 AM

Dec 21 2015

ygao added a parent revision for D15705: Adding a scripted test for PR25717: D15553: PR25717: fatal IO error writing large outputs to console on Windows.
Dec 21 2015, 5:12 PM
ygao added a child revision for D15553: PR25717: fatal IO error writing large outputs to console on Windows: D15705: Adding a scripted test for PR25717.
Dec 21 2015, 5:12 PM
ygao added inline comments to D15705: Adding a scripted test for PR25717.
Dec 21 2015, 5:10 PM
ygao retitled D15705: Adding a scripted test for PR25717 from to Adding a scripted test for PR25717.
Dec 21 2015, 5:07 PM