Page MenuHomePhabricator
Feed All Stories

Today

zequanwu retitled D85390: [Clang] Add note for bad conversion error when expression is of forward-declared type from [Clang] Add note for bad conversion when expression is of forward-declared type to [Clang] Add note for bad conversion error when expression is of forward-declared type.
Wed, Aug 5, 6:48 PM · Restricted Project
zequanwu requested review of D85390: [Clang] Add note for bad conversion error when expression is of forward-declared type.
Wed, Aug 5, 6:48 PM · Restricted Project
ruiling committed rG5ddc8b49babb: [AMDGPU] add buffer_atomic_swap for float (authored by ruiling).
[AMDGPU] add buffer_atomic_swap for float
Wed, Aug 5, 6:46 PM
ruiling closed D85187: [AMDGPU] add buffer_atomic_swap for float.
Wed, Aug 5, 6:46 PM · Restricted Project
delcypher requested review of D85389: Fix bug where we request a shadow memory one page larger than necessary..
Wed, Aug 5, 6:43 PM · Restricted Project
JDevlieghere added inline comments to D85388: [lldb] Fix bug in skipIfRosetta decorator.
Wed, Aug 5, 6:39 PM · Restricted Project
smeenai added inline comments to D85255: [lld-macho] Generate ObjC symbols from .tbd files.
Wed, Aug 5, 6:37 PM · Restricted Project
JDevlieghere requested review of D85388: [lldb] Fix bug in skipIfRosetta decorator.
Wed, Aug 5, 6:37 PM · Restricted Project
JDevlieghere committed rG3a538de65360: [lldb] Make UBSan tests remote ready (authored by JDevlieghere).
[lldb] Make UBSan tests remote ready
Wed, Aug 5, 6:32 PM
hiraditya added a comment to D85368: [llvm][CodeGen] Machine Function Splitter.

Please share performance numbers for publicly available workload(s).

Wed, Aug 5, 6:27 PM · Restricted Project
aqjune added a comment to D84948: [InstCombine] Fold select/and/or with freeze(undef) that is used in the op only.

To expand on this, if we want to handle multi-use freeze undef in the future, there's generally two ways that can work:

  1. Individual folds for certain operators (as done here), but replacing the freeze itself with an advantageous value rather than just the operator.
  2. Walking the uses of the freeze to find the best replacement value and replace all uses. (E.g. by counting for how many uses a certain value is favorable.)

The current patch would be the starting point for 1, my suggestion would be the starting point for 2.

I should say though that I'm also ok with your current patch.

Wed, Aug 5, 6:24 PM · Restricted Project
smeenai accepted D85153: [lld-macho] Add .tbd support for frameworks.

LGTM

Wed, Aug 5, 6:22 PM · Restricted Project
nemanjai updated the diff for D84360: [LLD][PowerPC] Implement GOT to PC-Rel relaxation.
  • Move implementation completely into LLD
  • Add option to disable the optimization
  • Handle computation of the total offset when the access instruction has a non-zero displacement
  • Add tests with the new option as well as a test with non-zero displacement on the access instructions
Wed, Aug 5, 6:18 PM · Restricted Project, Restricted Project
delcypher requested review of D85387: [NFC] Make argument name passed to `FindAvailableMemoryRange()` consistent with parameter name..
Wed, Aug 5, 6:16 PM · Restricted Project
cgyurgyik updated the diff for D85386: [libc] Add strpbrk implementation..
Wed, Aug 5, 6:12 PM · Restricted Project
NoQ accepted D85311: [analyzer][tests] Understand when diagnostics change between builds.

This is one of the scripts that really deserves tests outside of buildbots. Like, it's not uncommon to use it "by hand".

Wed, Aug 5, 6:09 PM · Restricted Project
tskeith committed rG6d2d73059fda: [flang] Fix compilation warning in check-directive-structure.h (authored by tskeith).
[flang] Fix compilation warning in check-directive-structure.h
Wed, Aug 5, 6:04 PM
tskeith closed D85383: [flang] Fix compilation warning in check-directive-structure.h.
Wed, Aug 5, 6:03 PM · Restricted Project, Restricted Project
davide added inline comments to D85376: Fix how ValueObjectChild handles bit-fields stored in a Scalar in UpdateValue() .
Wed, Aug 5, 6:00 PM
davide requested changes to D85376: Fix how ValueObjectChild handles bit-fields stored in a Scalar in UpdateValue() .
Wed, Aug 5, 5:58 PM
smeenai accepted D85362: [CMake] Print the autodetected host linker version.

LGTM

Wed, Aug 5, 5:57 PM · Restricted Project
NoQ added inline comments to D67422: [analyzer] NFC: Move path diagnostic consumer implementations to libAnalysis..
Wed, Aug 5, 5:55 PM · Restricted Project
NoQ updated the diff for D67422: [analyzer] NFC: Move path diagnostic consumer implementations to libAnalysis..

Fxd.

Wed, Aug 5, 5:55 PM · Restricted Project
cgyurgyik updated the diff for D85386: [libc] Add strpbrk implementation..
Wed, Aug 5, 5:54 PM · Restricted Project
smeenai accepted D85080: [lld-macho] Support static linking of thread-locals.

LGTM

Wed, Aug 5, 5:54 PM · Restricted Project
cgyurgyik requested review of D85386: [libc] Add strpbrk implementation..
Wed, Aug 5, 5:52 PM · Restricted Project
aqjune committed rG9f717d7b941f: [JumpThreading] Allow duplicating a basic block into preds when its branch… (authored by aqjune).
[JumpThreading] Allow duplicating a basic block into preds when its branch…
Wed, Aug 5, 5:52 PM
aqjune committed rGfd86d67b8283: [JumpThreading] Add a test that duplicates insts of a basic block with cond br… (authored by aqjune).
[JumpThreading] Add a test that duplicates insts of a basic block with cond br…
Wed, Aug 5, 5:51 PM
aqjune closed D85029: [JumpThreading] Allow duplicating a basic block into preds when its branch condition is freeze(phi).
Wed, Aug 5, 5:51 PM · Restricted Project
aqjune updated the summary of D85029: [JumpThreading] Allow duplicating a basic block into preds when its branch condition is freeze(phi).
Wed, Aug 5, 5:51 PM · Restricted Project
pengfei requested review of D85385: [X86][FPEnv] Teach X86 mask compare intrinsics to respect strict FP semantics..
Wed, Aug 5, 5:50 PM · Restricted Project, Restricted Project
craig.topper requested review of D85384: [X86] Add basic support for -mtune command line option in clang.
Wed, Aug 5, 5:45 PM
efriedma accepted D85327: [SVE] Implement lowering for fixed length vector multiplication..

I've updated the patch to only affect 64/128-bit vector mul operations when wide vectors are enabled.

Wed, Aug 5, 5:44 PM · Restricted Project
klausler accepted D85383: [flang] Fix compilation warning in check-directive-structure.h.
Wed, Aug 5, 5:42 PM · Restricted Project, Restricted Project
ychen accepted D85381: [NewPM][opt] Add more codegen passes.

LGTM. Thanks.

Wed, Aug 5, 5:41 PM · Restricted Project
tskeith requested review of D85383: [flang] Fix compilation warning in check-directive-structure.h.
Wed, Aug 5, 5:37 PM · Restricted Project, Restricted Project
friss added inline comments to D85376: Fix how ValueObjectChild handles bit-fields stored in a Scalar in UpdateValue() .
Wed, Aug 5, 5:37 PM
aprantl added inline comments to D85049: Unify the code that updates the ArchSpec after finding a fat binary with how it is done for a lean binary.
Wed, Aug 5, 5:36 PM
delcypher added a comment to D85378: Fix off-by-one error in size of the required shadow memory passed to `MapDynamicShadow`..

@tejohnson Looks like this bug is present on Linux too

Wed, Aug 5, 5:35 PM · Restricted Project
tmsriram accepted D85380: [NFC] Rename BBSectionsPrepare -> BasicBlockSections..

IMO, I think this naming is very clear as this pass exists to carve basic block sections.

Wed, Aug 5, 5:31 PM · Restricted Project
timshen requested review of D85382: [MLIR] Add getSizeInBits() for tensor of complex.
Wed, Aug 5, 5:28 PM · Restricted Project
tmsriram added a comment to D85085: Fix debug_loc offset difference with basic block sections.

Trying to reproduce this (by taking the .ll test case and compiling it with clang -c) I don't get any debug_loc section or DW_AT_locations. Is there something I'm missing about how to reproduce this?

(trying to see what the non-bb-sections location description looks like - if it depends on basic block ordering, it's probably incorrect anyway, and might be worth fixing so it doesn't depend on that non-guaranteed property & then there'll be less divergence between bb-sections and not)

I see why debug_loc is not produced in the default case. This is what happens:

+ For the default case (no bbsections), DebugLoc in https://github.com/llvm/llvm-project/blob/master/llvm/lib/CodeGen/AsmPrinter/DwarfDebug.cpp#L1744 , it has exactly one value that contains the whole function scope as the start label is lfunc_begin and the end label is the end of the function.
+ Such location lists are handled specially here when the scope is the whole function : https://github.com/llvm/llvm-project/blob/master/llvm/lib/CodeGen/AsmPrinter/DwarfDebug.cpp#L1841 and the debug_loc is not generated.

OK ^ and with bb sections do we see the same behavior? If not, why not?

Wed, Aug 5, 5:27 PM · Restricted Project
davidxl added inline comments to D85368: [llvm][CodeGen] Machine Function Splitter.
Wed, Aug 5, 5:23 PM · Restricted Project
jdenny updated the diff for D85369: [OpenMP] Fix ref count dec for implicit map of partial data.

Don't examine LIBOMPTARGET_DEBUG=1 output in the test. It doesn't work in release builds.

Wed, Aug 5, 5:20 PM · Restricted Project
aeubanks requested review of D85381: [NewPM][opt] Add more codegen passes.
Wed, Aug 5, 5:19 PM · Restricted Project
rriddle updated the diff for D82597: [mlir] Add a utility class, ThreadLocalCache, for storing non static thread local objects..

Switch to a combo of weak_ptr/shared_ptr.

Wed, Aug 5, 5:18 PM · Restricted Project
snehasish requested review of D85380: [NFC] Rename BBSectionsPrepare -> BasicBlockSections..
Wed, Aug 5, 5:16 PM · Restricted Project
ldionne updated subscribers of D82627: Fix CFI issues in <future>.

You’re good to go. Thanks!

Wed, Aug 5, 5:16 PM · Restricted Project
vabridgers added a comment to D85097: [Sema] add warning for comparisons like 'x<=y<=z'.

I added prototype fixits per request by Roman, updated the LIT test, and added an additional RUN line (one for -Wparentheses and one for -Wcompare-op-parentheses). Also filed https://bugs.llvm.org/show_bug.cgi?id=47010 to follow up on the FIXME cases at the bottom of the LIT since they are out of scope for this change. Thanks for the feedback and comments so far, I look forward to driving this change to completion.

Wed, Aug 5, 5:12 PM · Restricted Project
stephan.yichao.zhao requested review of D85379: Improve dropTriviallyDeadConstantArrays time cost ratio from 17% to 4%.
Wed, Aug 5, 5:11 PM · Restricted Project
vabridgers updated the diff for D85097: [Sema] add warning for comparisons like 'x<=y<=z'.

added prototype fixits for review.
added additional RUN test case.
filed https://bugs.llvm.org/show_bug.cgi?id=47010 for other
warnings improvement post landing of this patch, after lgtm

Wed, Aug 5, 5:09 PM · Restricted Project
efriedma added a comment to D85364: [SVE][WIP] Implement lowering for fixed width select.

For load <8 x i1> specifically, the code is terrible because we're using the generic target-independent expansion, which goes element by element. If we cared, we could custom-lower it to something more reasonable. Nobody has looked into it because there isn't any way to generate that operation from C code.

That's interesting. If we could load the vXi1 and then vector extend it, it might be more palatable. I haven't checked if there are instructions to support that though. And I wonder if it will get weird with a vector-of-i1s smaller than a byte...

Wed, Aug 5, 5:07 PM · Restricted Project
delcypher requested review of D85378: Fix off-by-one error in size of the required shadow memory passed to `MapDynamicShadow`..
Wed, Aug 5, 5:06 PM · Restricted Project
stefanp added a reviewer for D85377: [PowerPC] Add option to control PCRel GOT indirect linker optimization: Restricted Project.
Wed, Aug 5, 5:03 PM · Restricted Project
shafik added a comment to D85376: Fix how ValueObjectChild handles bit-fields stored in a Scalar in UpdateValue() .

Note: for the test I did not want to rely on clang choosing to pass the union by register, so I used assembly which ensures I will obtain the behavior I am looking to capture for the test.

Wed, Aug 5, 5:02 PM
stefanp requested review of D85377: [PowerPC] Add option to control PCRel GOT indirect linker optimization.
Wed, Aug 5, 5:01 PM · Restricted Project
MaskRay committed rG3af4039876a1: [ELF][test] Add another -triple=x86_64 to linkorder-mixed.s (authored by MaskRay).
[ELF][test] Add another -triple=x86_64 to linkorder-mixed.s
Wed, Aug 5, 5:00 PM
smeenai committed rG8b9e86fea66f: [lld-macho] Fix errant search and replace. NFC (authored by smeenai).
[lld-macho] Fix errant search and replace. NFC
Wed, Aug 5, 5:00 PM
shafik requested review of D85376: Fix how ValueObjectChild handles bit-fields stored in a Scalar in UpdateValue() .
Wed, Aug 5, 4:59 PM
davidxl added a comment to D85368: [llvm][CodeGen] Machine Function Splitter.

Regarding reroller -- compiler with PGO will adjust the agressiveness of the unroller based on instruction workset size estimation. Doing this in later pass or in Propeller can help catch cases that are mis-handled.

Wed, Aug 5, 4:59 PM · Restricted Project
jcai19 added a comment to D69411: [MC] Parse .if conditions with symbols in consecutive MCDataFragements.

@yknight while working on a different issue, I happened to take a look at the implementation of subsection (MCSection::getSubsectionInsertionPoint) and from what I understood, fragments in subsections were always placed after fragments not in any subsection., and fragments in the same subsection (or not in any subsection) were always inserted in order and separate from fragments in other subsections. Is it safe to assume what between two fragments either not in any subsections or in the same subsection will be remain unchanged once they are inserted to the fragment list of a section? If so, can we resolve differences of two labels in fragments with such restriction?

Wed, Aug 5, 4:58 PM · Restricted Project
paquette requested review of D85375: [GlobalISel] Fix computing known bits for loads with range metadata.
Wed, Aug 5, 4:57 PM · Restricted Project
rampitec accepted D85374: AMDGPU: Remove ATOMIC_PK_FADD.

LGTM, thanks!

Wed, Aug 5, 4:49 PM · Restricted Project
hubert.reinterpretcast added a comment to D85329: [AIX] Don't try to use llvm tools while building runtimes.

@daltenty, if these tools happened to be on the user's $PATH, then would they get picked up? Is that the desired behaviour?

Wed, Aug 5, 4:48 PM · Restricted Project
steven_wu updated the diff for D84803: [ThinLTO][MachO] Preserve both possible GUIDs from exported list.

Change how the GUID is computed in ThinLTOCodeGenerator.

Wed, Aug 5, 4:47 PM · Restricted Project
asbirlea committed rGbeb9993d960b: [MSSA] Update test with more detailed and resilient checks. [NFC] (authored by asbirlea).
[MSSA] Update test with more detailed and resilient checks. [NFC]
Wed, Aug 5, 4:47 PM
MaskRay added a comment to D84610: [ELF] --icf: don't fold text sections with LSDA.

This probably breaks the build on systems where uint64_t and size_t are different types, such as apparently on macOS:

http://45.33.8.238/mac/18492/step_4.txt

I'm only seeing this on some of my Macs though, so maybe it's dependent on sdk or host clang version. If the error makes sense to you, do something about it, else I'll try installing a newer host clang on that bot later.

Already fixed by 279e4cf78262dfef560631d643b85aa2032ac566

Wed, Aug 5, 4:47 PM · Restricted Project
aartbik committed rG39379916a7f0: [mlir] [VectorOps] Add masked load/store operations to Vector dialect (authored by aartbik).
[mlir] [VectorOps] Add masked load/store operations to Vector dialect
Wed, Aug 5, 4:46 PM
aartbik closed D85357: [mlir] [VectorOps] Add masked load/store operations to Vector dialect.
Wed, Aug 5, 4:46 PM · Restricted Project, Restricted Project
MaskRay added a comment to D84610: [ELF] --icf: don't fold text sections with LSDA.

This probably breaks the build on systems where uint64_t and size_t are different types, such as apparently on macOS:

http://45.33.8.238/mac/18492/step_4.txt

I'm only seeing this on some of my Macs though, so maybe it's dependent on sdk or host clang version. If the error makes sense to you, do something about it, else I'll try installing a newer host clang on that bot later.

Wed, Aug 5, 4:45 PM · Restricted Project
thakis added a comment to D84610: [ELF] --icf: don't fold text sections with LSDA.

This probably breaks the build on systems where uint64_t and size_t are different types, such as apparently on macOS:

Wed, Aug 5, 4:45 PM · Restricted Project
aartbik added a comment to D85357: [mlir] [VectorOps] Add masked load/store operations to Vector dialect.

Thanks!

Wed, Aug 5, 4:44 PM · Restricted Project, Restricted Project
arsenm requested review of D85374: AMDGPU: Remove ATOMIC_PK_FADD.
Wed, Aug 5, 4:42 PM · Restricted Project
hubert.reinterpretcast retitled D85315: [AIX][Clang][Driver] Generate reference to the C++ library on the link step from [AIX][Clang] Add C++ linker option to the driver toolchain to [AIX][Clang][Driver] Generate reference to the C++ library on the link step.
Wed, Aug 5, 4:42 PM · Restricted Project, Restricted Project
tskeith accepted D84334: [flang] Version information in flang/f18.

Thanks, looks good to me.

Wed, Aug 5, 4:41 PM · Restricted Project
LLVM GN Syncbot <llvmgnsyncbot@gmail.com> committed rGdd892a33e1ec: [gn build] Port 820e8d8656e (authored by LLVM GN Syncbot <llvmgnsyncbot@gmail.com>).
[gn build] Port 820e8d8656e
Wed, Aug 5, 4:40 PM
MaskRay committed rG279e4cf78262: [ELF] Fix type of ciesWithLSDA after D84610 (authored by MaskRay).
[ELF] Fix type of ciesWithLSDA after D84610
Wed, Aug 5, 4:34 PM
MaskRay added a comment to D72904: [ELF] Allow SHF_LINK_ORDER sections to have sh_link=0.

This breaks tests on windows: http://45.33.8.238/win/21396/step_10.txt

(Probably just needs a triple)

Wed, Aug 5, 4:29 PM · Restricted Project
MaskRay committed rGeb45b978b7d6: [ELF][test] Add REQUIES: x86 and -triple=x86_64 to linkorder-mixed.s (authored by MaskRay).
[ELF][test] Add REQUIES: x86 and -triple=x86_64 to linkorder-mixed.s
Wed, Aug 5, 4:29 PM
cameron.mcinally added a comment to D85364: [SVE][WIP] Implement lowering for fixed width select.

For load <8 x i1> specifically, the code is terrible because we're using the generic target-independent expansion, which goes element by element. If we cared, we could custom-lower it to something more reasonable. Nobody has looked into it because there isn't any way to generate that operation from C code.

Wed, Aug 5, 4:29 PM · Restricted Project
thakis added a comment to D72904: [ELF] Allow SHF_LINK_ORDER sections to have sh_link=0.

This breaks tests on windows: http://45.33.8.238/win/21396/step_10.txt

Wed, Aug 5, 4:26 PM · Restricted Project
jkorous committed rG820e8d8656ec: [Analyzer][WebKit] UncountedLambdaCaptureChecker (authored by jkorous).
[Analyzer][WebKit] UncountedLambdaCaptureChecker
Wed, Aug 5, 4:24 PM
jkorous closed D82837: [Analyzer][WebKit] UncountedLambdaCaptureChecker.
Wed, Aug 5, 4:24 PM · Restricted Project
nickdesaulniers updated subscribers of rGc41a18cf6179: [CMake] Default ENABLE_X86_RELAX_RELOCATIONS to ON.

Cross referencing the boot failure report from @nathanchance : https://github.com/ClangBuiltLinux/linux/issues/1121

Wed, Aug 5, 4:23 PM
coti updated the diff for D84334: [flang] Version information in flang/f18.

Thank you @tskeith for the tip on the tests!

Wed, Aug 5, 4:22 PM · Restricted Project
mtrofin requested review of D85373: [llvm][MLInliner] Don't log 'mandatory' events.
Wed, Aug 5, 4:18 PM · Restricted Project
MaskRay committed rGb216c80cc249: [ELF] Allow SHF_LINK_ORDER sections to have sh_link=0 (authored by MaskRay).
[ELF] Allow SHF_LINK_ORDER sections to have sh_link=0
Wed, Aug 5, 4:18 PM
MaskRay closed D72904: [ELF] Allow SHF_LINK_ORDER sections to have sh_link=0.
Wed, Aug 5, 4:18 PM · Restricted Project
dcaballe requested changes to D84998: [MLIR] Add affine.parallel canonicalizations and folding.

Thanks Frank! It looks good to me. Just a few minor comments

Wed, Aug 5, 4:12 PM · Restricted Project
eugenis added a comment to D82627: Fix CFI issues in <future>.

I can push this tomorrow unless anyone objects before then.

Wed, Aug 5, 4:12 PM · Restricted Project
paulwalker-arm updated the diff for D85327: [SVE] Implement lowering for fixed length vector multiplication..

Call me a scaredy cat but I'll take the conservative option :)
I've updated the patch to only affect 64/128-bit vector mul operations when wide vectors are enabled.

Wed, Aug 5, 4:11 PM · Restricted Project
MaskRay updated the summary of D72904: [ELF] Allow SHF_LINK_ORDER sections to have sh_link=0.
Wed, Aug 5, 4:09 PM · Restricted Project
xbolva00 added a comment to D85368: [llvm][CodeGen] Machine Function Splitter.

I am wondering what is is your opinion on machine unroller/reroller? Aggressive loop unrolling may destroy code cache too.

Wed, Aug 5, 4:09 PM · Restricted Project
davide added inline comments to D85049: Unify the code that updates the ArchSpec after finding a fat binary with how it is done for a lean binary.
Wed, Aug 5, 4:08 PM
phosek committed rG1adc494bce44: [CMake] Simplify CMake handling for zlib (authored by phosek).
[CMake] Simplify CMake handling for zlib
Wed, Aug 5, 4:07 PM
phosek closed D79219: [CMake] Simplify CMake handling for zlib.
Wed, Aug 5, 4:07 PM · Restricted Project, Restricted Project, Restricted Project, Restricted Project, Restricted Project
efriedma added a comment to D85368: [llvm][CodeGen] Machine Function Splitter.

We probably need to discuss how to make basic-block-section stuff work on non-X86 targets at some point, but I guess we don't have to do it in this patch if it's off by default.

Wed, Aug 5, 4:06 PM · Restricted Project
xbolva00 added reviewers for D62990: [SCEV]When safe, compute MinStart as unsigned_min(Start - Stride) + Stride in computeMaxBECountForLT: fhahn, reames.
Wed, Aug 5, 4:05 PM · Restricted Project
rriddle accepted D85372: [mlir] take MLIRContext instead of LLVMDialect in getters of LLVMType's.
Wed, Aug 5, 4:05 PM · Restricted Project
pcc accepted D72904: [ELF] Allow SHF_LINK_ORDER sections to have sh_link=0.

LGTM

Wed, Aug 5, 4:05 PM · Restricted Project