Page MenuHomePhabricator
Feed Advanced Search

Sat, Aug 17

rnk accepted D66373: Fixed placement of llvm.global_dtors on Windows..

lgtm, thanks!

Sat, Aug 17, 4:49 PM · Restricted Project
rnk accepted D65964: [lld-link] implement -lto-obj-path.

lgtm

Sat, Aug 17, 4:48 PM · Restricted Project

Fri, Aug 16

rnk added inline comments to D66361: Improve behavior in the case of stack exhaustion..
Fri, Aug 16, 2:18 PM · Restricted Project
rnk added inline comments to D66355: [COFF] Add libcall symbols to the link when LTO is being used..
Fri, Aug 16, 2:03 PM · Restricted Project
rnk added inline comments to D66344: Filesystem/Windows: fix inconsistency in readNativeFileSlice API.
Fri, Aug 16, 11:47 AM · Restricted Project
rnk added inline comments to D66355: [COFF] Add libcall symbols to the link when LTO is being used..
Fri, Aug 16, 11:33 AM · Restricted Project

Thu, Aug 15

rnk committed rGed399a69e2a0: [Sema] Implement DR2386 for C++17 structured binding (authored by rnk).
[Sema] Implement DR2386 for C++17 structured binding
Thu, Aug 15, 12:47 PM
rnk committed rL369043: [Sema] Implement DR2386 for C++17 structured binding.
[Sema] Implement DR2386 for C++17 structured binding
Thu, Aug 15, 12:44 PM
rnk closed D66040: [Sema] Implement DR2386 for C++17 structured binding.
Thu, Aug 15, 12:44 PM · Restricted Project, Restricted Project
rnk updated the diff for D66040: [Sema] Implement DR2386 for C++17 structured binding.
  • add DR test case
Thu, Aug 15, 10:52 AM · Restricted Project, Restricted Project
rnk accepted D65859: [LLVM][Alignment] Fix MSVC warning.

I prefer push/pop generally.

Thu, Aug 15, 9:17 AM · Restricted Project
rnk added a comment to D66040: [Sema] Implement DR2386 for C++17 structured binding.

rnk, what's the status here?

Thu, Aug 15, 9:10 AM · Restricted Project, Restricted Project

Tue, Aug 13

rnk accepted D66065: [MinGW] Restructure Options.td to use multiclass where sensible. NFC..

lgtm

Tue, Aug 13, 12:19 PM · Restricted Project
rnk accepted D66066: [LLD] [MinGW] Correct handling of a few options .

lgtm

Tue, Aug 13, 12:19 PM · Restricted Project

Mon, Aug 12

rnk updated subscribers of D65964: [lld-link] implement -lto-obj-path.

-lto-merged-obj sounds good to me, unless anyone objects. I see it was added to the gold plugin back in rL125663, and @rdhindsa ported it to LLD LTO in rL331817.

Mon, Aug 12, 4:36 PM · Restricted Project
rnk committed rGe9865b9b31bb: [WinEH] Fix catch block parent frame pointer offset (authored by rnk).
[WinEH] Fix catch block parent frame pointer offset
Mon, Aug 12, 4:02 PM
rnk committed rL368631: [WinEH] Fix catch block parent frame pointer offset.
[WinEH] Fix catch block parent frame pointer offset
Mon, Aug 12, 4:01 PM

Fri, Aug 9

rnk created D66040: [Sema] Implement DR2386 for C++17 structured binding.
Fri, Aug 9, 4:55 PM · Restricted Project, Restricted Project
rnk accepted D65923: [X86] Fix stack probe issue on windows32..

lgtm, thanks!

Fri, Aug 9, 1:10 PM · Restricted Project
rnk accepted D66023: Warning disable for mod-by-zero and union 'uninitialized' use.

lgtm

Fri, Aug 9, 12:57 PM · Restricted Project, Restricted Project
rnk committed rG50fcf7285eeb: Don't diagnose errors when a file matches an include component (authored by rnk).
Don't diagnose errors when a file matches an include component
Fri, Aug 9, 12:52 PM
rnk committed rL368475: Don't diagnose errors when a file matches an include component.
Don't diagnose errors when a file matches an include component
Fri, Aug 9, 12:48 PM

Thu, Aug 8

rnk added a comment to D65964: [lld-link] implement -lto-obj-path.

I think the command line flag needs a better name to indicate that it's the destination for the merged, whole-program bitcode file. In ELF, the config name is ltoObjPath, and the flag is -plugin-opt=obj-path=. I think we can be more descriptive. How about: -split-bc-path:, -merged-bc-path:, -lto-bc-path, or something else descriptive?

Thu, Aug 8, 3:24 PM · Restricted Project
rnk accepted D65970: [GlobalOpt] prevent crashing on wide integer types (PR42932).

lgtm

Thu, Aug 8, 3:11 PM · Restricted Project
rnk committed rG7859fed6eaff: Mark clang-scan-deps test as requiring thread support (authored by rnk).
Mark clang-scan-deps test as requiring thread support
Thu, Aug 8, 2:46 PM
rnk committed rL368354: Mark clang-scan-deps test as requiring thread support.
Mark clang-scan-deps test as requiring thread support
Thu, Aug 8, 2:45 PM
rnk committed rG1d63b02f5305: Fix up fd limit diagnosis code (authored by rnk).
Fix up fd limit diagnosis code
Thu, Aug 8, 2:35 PM
rnk committed rL368348: Fix up fd limit diagnosis code.
Fix up fd limit diagnosis code
Thu, Aug 8, 2:34 PM
rnk added inline comments to D64931: Change X86 datalayout for three address spaces that specify pointer sizes..
Thu, Aug 8, 11:35 AM · Restricted Project, Restricted Project
rnk added a comment to D65923: [X86] Fix stack probe issue on windows32..

Seems reasonable, but I have some suggestions.

Thu, Aug 8, 10:13 AM · Restricted Project
rnk accepted D65956: clang: Diag running out of file handles while looking for files.

lgtm

Thu, Aug 8, 9:54 AM · Restricted Project

Wed, Aug 7

rnk accepted D65023: [MS] Pass S_HEAPALLOCSITE metadata through SelectionDAG.

lgtm, let's try again...

Wed, Aug 7, 3:45 PM · Restricted Project
rnk accepted D65838: [Driver] Use enumeration for quoting mode. NFC.

lgtm+2

Wed, Aug 7, 2:35 PM · Restricted Project
rnk accepted D65890: Support: Remove needless allocation when getMainExecutable() calls readlink().

Huh, code seems to date from 2012: rL164676.

Wed, Aug 7, 9:50 AM · Restricted Project

Tue, Aug 6

rnk added a comment to D65826: Add support for deterministically linked binaries on macOS to lldb..

Code seems fine, I think this is probably the right fix, but this probably deserves a test along the lines of what's in the commit message. I'm not familiar with LLDB's test suite, so I can't give much guidance there.

Tue, Aug 6, 1:48 PM · Restricted Project, Restricted Project
rnk removed a reviewer for D64884: [PHINode] Preserve use-list order when removing incoming values.: rnk.
Tue, Aug 6, 1:43 PM · Restricted Project
rnk added a comment to D65380: [InstCombine] Shift amount reassociation: shl-trunc-shl pattern.

The assertion and stack trace:
https://reviews.llvm.org/P8158
clang: /usr/local/google/home/rnk/llvm-project/llvm/lib/IR/Constants.cpp:1864: static llvm::Constant *llvm::ConstantExpr::get(unsigned int, llvm::Constant *, llvm::Constant *
, unsigned int, llvm::Type *): Assertion `C1->getType() == C2->getType() && "Operand types in binary constant expression should match"' failed.
#9 0x0000000003ca39ae llvm::ConstantExpr::get(unsigned int, llvm::Constant*, llvm::Constant*, unsigned int, llvm::Type*) /usr/local/google/home/rnk/llvm-project/llvm/lib/IR/
Constants.cpp:1896:5
#10 0x00000000036e4fec llvm::ConstantFoldBinaryOpOperands(unsigned int, llvm::Constant*, llvm::Constant*, llvm::DataLayout const&) /usr/local/google/home/rnk/llvm-project/llv
m/lib/Analysis/ConstantFolding.cpp:1286:10

Tue, Aug 6, 1:34 PM · Restricted Project
rnk committed rGe4bd38478b38: Revert [InstCombine] Shift amount reassociation: shl-trunc-shl pattern (authored by rnk).
Revert [InstCombine] Shift amount reassociation: shl-trunc-shl pattern
Tue, Aug 6, 1:32 PM
rnk committed rL368084: Revert [InstCombine] Shift amount reassociation: shl-trunc-shl pattern.
Revert [InstCombine] Shift amount reassociation: shl-trunc-shl pattern
Tue, Aug 6, 1:31 PM
rnk added a comment to D65380: [InstCombine] Shift amount reassociation: shl-trunc-shl pattern.

This broke the Windows bot:
http://lab.llvm.org:8011/builders/clang-x64-windows-msvc/builds/9418
The failure mode doesn't manifest until the next build due to unrelated issues:
http://lab.llvm.org:8011/builders/clang-x64-windows-msvc/builds/9419

Tue, Aug 6, 1:29 PM · Restricted Project
rnk added inline comments to D65023: [MS] Pass S_HEAPALLOCSITE metadata through SelectionDAG.
Tue, Aug 6, 12:06 PM · Restricted Project
rnk accepted D65686: IR: Disable verifier check for GlobalValues with private linkage named after a comdat for non-COFF..

In the past, Rafael advocated for having more verifier check more things that are unrepresentable at an object file level. That's how we ended up with most of the checks in Verifier::visitGlobalAlias. I think this is consistent with that, even if we have to check which object file invariant we're enforcing.

Tue, Aug 6, 11:45 AM · Restricted Project
rnk accepted D65736: Handle /align option..

lgtm with question, no need to re-review

Tue, Aug 6, 10:51 AM · Restricted Project

Mon, Aug 5

rnk added a comment to D65686: IR: Disable verifier check for GlobalValues with private linkage named after a comdat for non-COFF..

I haven't had time to really think about this, but I believe @majnemer felt it was important to add verifier checks like these. There's no reason we can't make this triple dependent, right? If there's no triple on the module, we can skip the check.

Mon, Aug 5, 11:57 AM · Restricted Project
rnk committed rGd67c90a8c4a6: Try to fix failing AMDGPU disasm test, both Lin/Win agree this is 0 not 0x0 (authored by rnk).
Try to fix failing AMDGPU disasm test, both Lin/Win agree this is 0 not 0x0
Mon, Aug 5, 11:50 AM
rnk committed rL367907: Try to fix failing AMDGPU disasm test, both Lin/Win agree this is 0 not 0x0.
Try to fix failing AMDGPU disasm test, both Lin/Win agree this is 0 not 0x0
Mon, Aug 5, 11:49 AM
rnk added a comment to D65620: [AMDGPU] Use S_DENORM_MODE for gfx10.
Mon, Aug 5, 11:49 AM · Restricted Project

Fri, Aug 2

rnk committed rGe6ef4f98d289: The MinGW linker supports response files (authored by rnk).
The MinGW linker supports response files
Fri, Aug 2, 4:00 PM
rnk committed rL367733: The MinGW linker supports response files.
The MinGW linker supports response files
Fri, Aug 2, 3:54 PM
rnk accepted D64793: [Driver] Properly use values-X[ca].o, values-xpg[46].o on Solaris.

If I interpret @jyknight correctly, having failed to dissuade you, he doesn't feel strongly enough about this to block it, he just wants to reduce legacy when feasible.

Fri, Aug 2, 3:46 PM · Restricted Project, Restricted Project
rnk accepted D65662: Remove support for unsupported MSVC versions.

lgtm

Fri, Aug 2, 2:44 PM · Restricted Project
rnk accepted D65562: Move LangStandard*, InputKind::Language to Basic.

Looks good either way.

Fri, Aug 2, 2:29 PM · Restricted Project, Restricted Project
rnk accepted D65582: IR: accept and print numbered %N names for function args.

lgtm

Fri, Aug 2, 2:22 PM · Restricted Project, Restricted Project
rnk accepted D65624: [lit] Print internal env commands.

lgtm, thanks!

Fri, Aug 2, 2:11 PM · Restricted Project
rnk added a comment to D65624: [lit] Print internal env commands.

Suggestion, but this looks good too.

Fri, Aug 2, 11:34 AM · Restricted Project
rnk accepted D65668: [Driver][test] Avoid undefined grep in darwin-ld.c.

Probably like most programmers, I am not familiar with section 9.3.2 of posix. I assume this is affecting a real platform that you care about, but I'm curious what it is.

Fri, Aug 2, 11:29 AM · Restricted Project, Restricted Project
rnk accepted D65179: Changing representation of .cv_def_range directives in Codeview debug info assembly format for better readability.

lgtm

Fri, Aug 2, 11:29 AM · Restricted Project

Thu, Aug 1

rnk added a comment to D65619: [lit] use built-in clear_environment instead of external env command.

I think this ended up being reverted in rL367627, see the discussion on the review:
https://reviews.llvm.org/D65156#1610296

Thu, Aug 1, 4:36 PM · Restricted Project
rnk accepted D61027: Fix crash on switch conditions of non-integer types in templates.

Looks good to me. I was hoping Richard would take a look, but I reproduced the crash, and I'd rather see the fix land sooner than later. Thanks!

Thu, Aug 1, 3:41 PM · Restricted Project
rnk added a comment to D65562: Move LangStandard*, InputKind::Language to Basic.

Thanks!

Thu, Aug 1, 11:24 AM · Restricted Project, Restricted Project
rnk added a comment to D65543: Use library basenames when autolinking on Windows.

I'll look into that. I also noticed that check-ubsan fails. I think we should also change clang's driver to add this libpath when it invokes the linker, so that this works transparently when using the GCC-style driver.

Thu, Aug 1, 11:12 AM · Restricted Project
rnk accepted D65579: Don't try emitting dllexported explicitly defaulted non-trivial ctors twice during explicit template instantiation definition (PR42857).

lgtm

Thu, Aug 1, 10:59 AM · Restricted Project
rnk accepted D65429: Improving CodeView debug info type record's inline comments.

lgtm, thanks! I think that's everything we need to remove the type dump comment block.

Thu, Aug 1, 10:50 AM · Restricted Project
rnk added a comment to D65582: IR: accept and print numbered %N names for function args.

This is awesome! Thanks for doing the test updates. As I read it, LLVM will still read old-style IR just fine, which seems nice.

Thu, Aug 1, 10:44 AM · Restricted Project, Restricted Project
rnk created Image Macro "celebration_balloons".
Thu, Aug 1, 10:38 AM

Wed, Jul 31

rnk created D65543: Use library basenames when autolinking on Windows.
Wed, Jul 31, 3:00 PM · Restricted Project
rnk accepted D65511: Delay emitting dllexport explicitly defaulted members until the class is fully parsed (PR40006).

lgtm

Wed, Jul 31, 10:48 AM · Restricted Project
rnk added inline comments to D65511: Delay emitting dllexport explicitly defaulted members until the class is fully parsed (PR40006).
Wed, Jul 31, 9:55 AM · Restricted Project

Tue, Jul 30

rnk added inline comments to D65179: Changing representation of .cv_def_range directives in Codeview debug info assembly format for better readability.
Tue, Jul 30, 7:36 PM · Restricted Project
rnk committed rGc41b58fd409a: Remove cmake checks for MSVC 1900 / VS 2013 (authored by rnk).
Remove cmake checks for MSVC 1900 / VS 2013
Tue, Jul 30, 3:50 PM
rnk committed rL367362: Remove cmake checks for MSVC 1900 / VS 2013.
Remove cmake checks for MSVC 1900 / VS 2013
Tue, Jul 30, 3:48 PM
rnk added a comment to D64483: [CMake] Define _FILE_OFFSET_BITS=64 on Solaris.
In D64483#1591428, @ro wrote:
In D64483#1590120, @rnk wrote:

I would place this macro setting with the other code that sets this macro, in cmake/modules/HandleLLVMOptions.cmake.

This doesn't work unfortunately: I still get the same link failure for Sanitizer-i386-Test with gcc 9. For the original patch I'd followed how things are done
for AIX (_XOPEN_SOURCE=700, _LARGE_FILE_API) which is similar.

Tue, Jul 30, 1:10 PM · Restricted Project
rnk accepted D61105: [MS] Emit S_HEAPALLOCSITE debug info in SelectionDAG.

One issue, but otherwise this looks good to me.

Tue, Jul 30, 12:58 PM · Restricted Project
rnk added inline comments to D65179: Changing representation of .cv_def_range directives in Codeview debug info assembly format for better readability.
Tue, Jul 30, 11:53 AM · Restricted Project
rnk added inline comments to D65429: Improving CodeView debug info type record's inline comments.
Tue, Jul 30, 11:45 AM · Restricted Project
rnk committed rGc0e275df3d5d: Remove cache for macro arg stringization (authored by rnk).
Remove cache for macro arg stringization
Tue, Jul 30, 10:59 AM
rnk committed rL367337: Remove cache for macro arg stringization.
Remove cache for macro arg stringization
Tue, Jul 30, 10:59 AM
rnk closed D65428: Remove cache for macro arg stringization.
Tue, Jul 30, 10:59 AM · Restricted Project, Restricted Project
rnk added a comment to D65428: Remove cache for macro arg stringization.

Thanks for looking into it as well, I would not have seen the issue without reviewing the test case again. :)

Tue, Jul 30, 10:59 AM · Restricted Project, Restricted Project
rnk accepted D65458: [NFC] Remove LLVM_ALIGNAS.

lgtm

Tue, Jul 30, 10:41 AM · Restricted Project, Restricted Project

Mon, Jul 29

rnk added a comment to D65426: [Coverage] Hide coverage for regions with incorrect end locations (PR39942).

PTAL: https://reviews.llvm.org/D65428

Mon, Jul 29, 6:31 PM
rnk created D65428: Remove cache for macro arg stringization.
Mon, Jul 29, 6:31 PM · Restricted Project, Restricted Project
rnk added a comment to D65426: [Coverage] Hide coverage for regions with incorrect end locations (PR39942).

I did some digging and I figured out where things go wrong. The issue is the repeated string-izing of the macro argument, the repetition of #func here. This hacky patch seems to fix the issue:

diff --git a/clang/lib/Lex/MacroArgs.cpp b/clang/lib/Lex/MacroArgs.cpp
index 5aa4679fad4..29fd25a43bb 100644
--- a/clang/lib/Lex/MacroArgs.cpp
+++ b/clang/lib/Lex/MacroArgs.cpp
@@ -318,7 +318,7 @@ const Token &MacroArgs::getStringifiedArgument(unsigned ArgNo,
   if (StringifiedArgs.empty())
     StringifiedArgs.resize(getNumMacroArguments(), {});
Mon, Jul 29, 6:01 PM
rnk accepted D65249: [NFC] use C++11 in AlignOf.h, remove AlignedCharArray.

I have concerns that some of the patches that you landed prior to this will cause issues with old versions of MSVC, but in isolation, this is fine, and if anyone complains, we can address it on a case-by-case basis. lgtm

Mon, Jul 29, 5:39 PM · Restricted Project, Restricted Project, Restricted Project
rnk added a comment to D63048: Update __VERSION__ to remove the hardcoded 4.2.1 version.

Should we bump GNUC, GNUC_MINOR too?

Mon, Jul 29, 1:55 PM · Restricted Project, Restricted Project
rnk committed rG2336c1b872a6: Fix taint-generic.c on Windows, handle case in OS error (authored by rnk).
Fix taint-generic.c on Windows, handle case in OS error
Mon, Jul 29, 11:52 AM
rnk committed rL367249: Fix taint-generic.c on Windows, handle case in OS error.
Fix taint-generic.c on Windows, handle case in OS error
Mon, Jul 29, 11:52 AM
rnk added a comment to D64793: [Driver] Properly use values-X[ca].o, values-xpg[46].o on Solaris.
In D64793#1597765, @ro wrote:

How about instead just adding "values-xpg6.o" unconditionally, alongside the current unconditional "values-Xa.o", and just forget about the xpg4 and Xc modes?

If all else fails, that would have to be the last fallback. I'd rather do things correctly, though.

Mon, Jul 29, 11:07 AM · Restricted Project, Restricted Project

Fri, Jul 26

rnk committed rG95ed4c383535: Fix remarks unit test on Windows (authored by rnk).
Fix remarks unit test on Windows
Fri, Jul 26, 3:11 PM
rnk added inline comments to rL367155: Reland: [Remarks] Support parsing remark metadata in the YAML remark parser.
Fri, Jul 26, 3:11 PM
rnk committed rL367159: Fix remarks unit test on Windows.
Fix remarks unit test on Windows
Fri, Jul 26, 3:10 PM
rnk added inline comments to rL367155: Reland: [Remarks] Support parsing remark metadata in the YAML remark parser.
Fri, Jul 26, 3:02 PM
rnk added a comment to D65249: [NFC] use C++11 in AlignOf.h, remove AlignedCharArray.

I still think this concern applies:
https://reviews.llvm.org/D64417#1576675

Fri, Jul 26, 1:41 PM · Restricted Project, Restricted Project, Restricted Project
rnk accepted D65233: driver: Don't warn about assembler flags being unused when not assembling; different approach.

lgtm

Fri, Jul 26, 12:44 PM · Restricted Project

Mon, Jul 22

rnk accepted D65018: [LLD] [COFF] Unbreak sorting of mingw comdat .tls sections after SVN r363457.

lgtm, thanks!

Mon, Jul 22, 3:14 PM · Restricted Project
rnk added inline comments to D65018: [LLD] [COFF] Unbreak sorting of mingw comdat .tls sections after SVN r363457.
Mon, Jul 22, 1:46 PM · Restricted Project
rnk added a comment to D65023: [MS] Pass S_HEAPALLOCSITE metadata through SelectionDAG.

We discussed this offline, and after looking at rL364516, it seems like the side table might be a better way to go.

Mon, Jul 22, 1:11 PM · Restricted Project
rnk added inline comments to D65108: Reland "driver: Don't warn about assembler flags being unused when not assembling".
Mon, Jul 22, 1:06 PM
rnk added a comment to D64595: [Debuginfo][SROA] Need to handle dbg.value in SROA pass..

There's a lot of text here, I'm sorry I haven't read it all, but I think we should try to avoid getting in this situation where dbg.value points to an alloca in the first place. In Chromium, we use -instcombine-lower-dbg-declare to prevent instcombine from conservatively demoting from the "high availability" dbg.declare to the "always accurate but often missing" dbg.value. I don't really have time to work on this or participate in the discussion, so I'll just drop my two cents in and run away... =S

Mon, Jul 22, 11:39 AM · Restricted Project, debug-info

Jul 19 2019

rnk accepted D65016: Changes to emit CodeView debug info nested type records properly using MCStreamer directives.

Looks good, but please run git-clang-format before committing. Thanks!

Jul 19 2019, 3:59 PM · Restricted Project