Page MenuHomePhabricator
Feed All Stories

Fri, Jun 11

kda added inline comments to D104076: [clang-cl][sanitizer] Add -fsanitize-address-use-after-return to clang..
Fri, Jun 11, 11:06 AM · Restricted Project, Restricted Project
kda updated the diff for D104076: [clang-cl][sanitizer] Add -fsanitize-address-use-after-return to clang..
  • fixed up conflicting command line parameters.
  • only emit asan_stack_malloc calls when needed.
  • trimmed namespace ('llvm').
  • improved tests.
  • Changed default parameters from Never to Runtime.
Fri, Jun 11, 11:06 AM · Restricted Project, Restricted Project
JDevlieghere requested changes to D101340: Allows for dsymutil crashes to generate reproduceable information.
Fri, Jun 11, 11:06 AM · Restricted Project
vsavchenko requested review of D104135: [analyzer] Decouple NoteTag from its Factory.
Fri, Jun 11, 11:04 AM · Restricted Project
zoecarver committed rG7eba4856c702: [libcxx][ranges] Add class ref_view. (authored by zoecarver).
[libcxx][ranges] Add class ref_view.
Fri, Jun 11, 11:04 AM
zoecarver closed D102020: [libcxx][ranges] Add class ref_view..
Fri, Jun 11, 11:04 AM · Restricted Project
fhahn updated the diff for D102834: [SLPVectorizer] WIP Implement initial memory versioning (WIP!).

Finished off the main functionality, update includes

  • computing & comparing the cost of the versioned block to the original block,
  • more tests,
  • preserving LI & DT if changes to the CFG are made.
Fri, Jun 11, 11:02 AM · Restricted Project
klausler added inline comments to D104130: [flang] Handle multiple USE statements for the same module.
Fri, Jun 11, 11:02 AM · Restricted Project, Restricted Project
ThomasRaoux requested review of D104134: [mlir][VectorToGPU] Add conversion for scf::For op with Matrix operands.
Fri, Jun 11, 11:01 AM · Restricted Project
aartbik updated the diff for D104074: [mlir][sparse] allow all-dense annotated "sparse" tensor output.

removed code leftover

Fri, Jun 11, 11:01 AM · Restricted Project
CarlosAlbertoEnciso added inline comments to D103089: [Debug-Info][CodeView] Fix GUID string generation for MSVC generated objects..
Fri, Jun 11, 10:55 AM · Restricted Project, debug-info
CarlosAlbertoEnciso updated the diff for D103089: [Debug-Info][CodeView] Fix GUID string generation for MSVC generated objects..

Address comments from @amccarth:

Fri, Jun 11, 10:54 AM · Restricted Project, debug-info
aartbik updated the diff for D104074: [mlir][sparse] allow all-dense annotated "sparse" tensor output.

variadic operands (forward looking to sparse generalization)

Fri, Jun 11, 10:53 AM · Restricted Project
aeubanks added a reviewer for D104099: [NewPM] Remove SpeculateAroundPHIs pass: davidxl.
Fri, Jun 11, 10:53 AM · Restricted Project, Restricted Project
JDevlieghere accepted D103391: [lldb] Add missing reproducer instrumentation to some SB classes.

Thanks!

Fri, Jun 11, 10:49 AM · Restricted Project
JDevlieghere accepted D103575: Allow signposts to take advantage of deferred string substitution.

Looks good to me!

Fri, Jun 11, 10:48 AM · Restricted Project, Restricted Project
Quuxplusone updated the diff for D103753: [libc++] [P0619] Add _LIBCPP_ABI_NO_BINDER_BASES and remove binder typedefs in C++20..

Address review comments from @ldionne.

Fri, Jun 11, 10:46 AM · Restricted Project
arsenm committed rGa845dc1e562c: AMDGPU/GlobalISel: Remove leftover hack for argument memory sizes (authored by arsenm).
AMDGPU/GlobalISel: Remove leftover hack for argument memory sizes
Fri, Jun 11, 10:45 AM
arsenm committed rG6dd54dada3b1: AMDGPU/GlobalISel: Fix indentation (authored by arsenm).
AMDGPU/GlobalISel: Fix indentation
Fri, Jun 11, 10:45 AM
arsenm committed rG9d7299b6f0ab: GlobalISel: Reduce indentation and remove dead path (authored by arsenm).
GlobalISel: Reduce indentation and remove dead path
Fri, Jun 11, 10:45 AM
arsenm committed rG93f3c7cc3efc: CodeGen: Fix missing const (authored by arsenm).
CodeGen: Fix missing const
Fri, Jun 11, 10:45 AM
ayermolo added a comment to D104080: [LLD][LLVM] CG Graph profile using relocations.

The size went up from 107KB to 322KB, aggregate of all the input sections.

That sounds like a lot (3x increase). How does this impact link times?

Is this patch intended to be reviewed as-is, or is this still illustrative?

Fri, Jun 11, 10:42 AM · Restricted Project
reames added a comment to D97982: [MC] Introduce NeverAlign fragment type.

Drive by thought, not intended to be blocking.

Fri, Jun 11, 10:41 AM · Restricted Project
phosek added a comment to D101479: [Driver] Support libc++ in MSVC.

Couldn’t this commit have been kept in, and just reverting the one for using it in the fuchsia cmake cache? (I’m not using this particular commit myself, just observing.)

Fri, Jun 11, 10:41 AM · Restricted Project
MyDeveloperDay added a comment to D104044: [clang-format] Fix the issue that empty lines being removed at the beginning of namespace.

My point being there is inconsistency between how different types of blocks of code are handled, and rather than trying to fix another corner case maybe we should take a more holistic approach, all these KeepEmptyLines and EmptyLineAfterXXX options and what you'll need in order to fix this issue are all addressing what is effectively the same issue, and that is that the addition and/or removal of empty lines is a little hit or miss depending on your combination and permutation of settings and the type of block

Fri, Jun 11, 10:40 AM · Restricted Project, Restricted Project
klausler added a comment to D104100: [flang] Add clang-tidy check for braces around if.

This should probably not be applied to all of f18; the lowering code, which interacts heavily with mlir and llvm, respects their styles.

Fri, Jun 11, 10:39 AM · Restricted Project, Restricted Project
ThomasRaoux requested review of D104133: [mlir][VectorToGPU] Add conversion for splat constant to MMA const matrix.
Fri, Jun 11, 10:39 AM · Restricted Project
wenlei added a comment to D104131: [CSSPGO] Aggregation by the last K context frames for cold profiles.

Thanks for the patch. I think this can be useful for tuning. However as we discussed, it'd be nice if the level of merging (number of frames to keep after merging) can be adaptive based on hotness. But to do that, the preinliner in llvm-profgen is the way to go.

Fri, Jun 11, 10:38 AM · Restricted Project
klausler added a comment to D104019: [flang] Add initial implementation for CPU_TIME.

Please ensure that this code does not make the Fortran runtime library binary dependent on a C++ runtime library. We're trying to restrict dependences to C to make things easier for vendors and application writers to ship binaries without having to also ship one or more C++ environments alongside them.

Fri, Jun 11, 10:36 AM · Restricted Project, Restricted Project
lebedev.ri added a comment to D104099: [NewPM] Remove SpeculateAroundPHIs pass.

Which pass that comes after SpeculateAroundPHIs in the X86 pipeline (either in the optimization or codegen) would undo its effects?

Fri, Jun 11, 10:36 AM · Restricted Project, Restricted Project
melver abandoned D103958: [WIP] Support MustControl conditional control attribute.
Fri, Jun 11, 10:35 AM · Restricted Project, Restricted Project
chrish_ericsson_atx committed rGce44fe199bbf: [Sema] Address-space sensitive check for unbounded arrays (v2) (authored by chrish_ericsson_atx).
[Sema] Address-space sensitive check for unbounded arrays (v2)
Fri, Jun 11, 10:34 AM
chrish_ericsson_atx closed D88174: [Sema] Address-space sensitive check for unbounded arrays (v2).
Fri, Jun 11, 10:34 AM · Restricted Project
oontvoo added a comment to D103971: [lld-macho] Rework mergeFlag to closer mimic what LD64 does..

Could you add a test?

Fri, Jun 11, 10:33 AM · Restricted Project, Restricted Project
reames updated subscribers of D97982: [MC] Introduce NeverAlign fragment type.

@lebedev.ri @reames Could you weigh in from performance perspective?

JFYI, I no longer have access to the infrastructure used to test previous patches in this area. You could consider asking @skatkov if needed.

Fri, Jun 11, 10:33 AM · Restricted Project
wlei updated the summary of D104131: [CSSPGO] Aggregation by the last K context frames for cold profiles.
Fri, Jun 11, 10:31 AM · Restricted Project
reames planned changes to D103844: [SCEV] Cache wrap facts for positive IVs w/LT exits.

Change appears to be wrong as written, discussing why before adjusting.

Fri, Jun 11, 10:30 AM · Restricted Project
marbre requested review of D104132: [mlir][docs] Reorder PassWrapper arguments.
Fri, Jun 11, 10:29 AM · Restricted Project
ftynse committed rGfdc0d4360b4e: Introduce alloca_scope op (authored by shabalin).
Introduce alloca_scope op
Fri, Jun 11, 10:29 AM
thakis added a comment to D103184: [AArch64] handle -Wa,-march=.

FWIW the failure goes away locally if I revert this change here, so it's definitely due to this change here.

Things have been red for a while, probably good to revert while you investigate by this point.

Sorry for the delay but my macbook had some issues so I just finished building LLVM, but I can't seem to reproduce the test failure locally (neither by running the test directly with llvm-lit nor with ninja check-clang). In fact I don't quite understand why this test failed. Like @nickdesaulniers mentioned the build failure should have been fixed in https://reviews.llvm.org/rG76d9bc72784d88f4dd57b9939e52c73739438af5 for a different patch. Do you mind sharing the instructions you used to reproduce the test failure? Thanks.

Fri, Jun 11, 10:29 AM · Restricted Project
ftynse closed D97768: Introduce alloca_scope op.
Fri, Jun 11, 10:28 AM · Restricted Project
JDevlieghere committed rGfc71a5c6e8e8: [lldb] Support new objective-c hash table layout (authored by JDevlieghere).
[lldb] Support new objective-c hash table layout
Fri, Jun 11, 10:26 AM
aeubanks added a comment to D104099: [NewPM] Remove SpeculateAroundPHIs pass.

Which pass that comes after SpeculateAroundPHIs in the X86 pipeline (either in the optimization or codegen) would undo its effects?

Fri, Jun 11, 10:25 AM · Restricted Project, Restricted Project
JDevlieghere committed rGc7dee6aefd5b: [lldb] Enable TestRuntimeTypes on Apple Silicon (authored by JDevlieghere).
[lldb] Enable TestRuntimeTypes on Apple Silicon
Fri, Jun 11, 10:25 AM
jasonliu accepted D93659: [AIX][XCOFF] emit vector info of traceback table..

LGTM; Thanks!

Fri, Jun 11, 10:24 AM · Restricted Project
asm added a comment to D104054: [lldb] Enable Rust v0 symbol demangling.

And thank you for the quick review!

Fri, Jun 11, 10:24 AM · Restricted Project
chrish_ericsson_atx updated the diff for D88174: [Sema] Address-space sensitive check for unbounded arrays (v2).

Corrected APSInt.toString() usage to comply with intervening change 61cdaf66fe22be2b5

Fri, Jun 11, 10:24 AM · Restricted Project
Quuxplusone accepted D102028: [libcxx][ranges] Implement views::all..

Sure, I don't need to be red anymore.

Fri, Jun 11, 10:22 AM · Restricted Project
craig.topper added a comment to D104037: [X86] Check immediate before get it..

I don't what Intel's original failure looked like, but here's a test that should reproduce this with -run-pass=machinelicm https://reviews.llvm.org/P8267 needs more cleanup.

Fri, Jun 11, 10:21 AM · Restricted Project
mstorsjo added a comment to D101479: [Driver] Support libc++ in MSVC.

Couldn’t this commit have been kept in, and just reverting the one for using it in the fuchsia cmake cache? (I’m not using this particular commit myself, just observing.)

Fri, Jun 11, 10:21 AM · Restricted Project
nickdesaulniers added a comment to D103928: [IR] make -warn-frame-size into a module attr.

FWIW, it was pointed out to me that the commit message doesn't precisely match the actual attribute name. The actual attribute name is warn-stack-size, but the commit message uses -warn-frame-size. Oh well.

Fri, Jun 11, 10:21 AM · Restricted Project, Restricted Project
int3 added a comment to D104084: [lld-macho] Add options -(no_)keep_dwarf_unwind.

Yes, I will do it, but wanted this out quickly.

Fri, Jun 11, 10:21 AM · Restricted Project, Restricted Project
lebedev.ri added a comment to D104099: [NewPM] Remove SpeculateAroundPHIs pass.

Some backends don't run SimplifyCFG, e.g. X86. I believe the pass was originally created specifically for X86 (the header has some X86 examples) and may or may not extend to other targets (I'm not very familiar with the pass itself).

I'm not opposed to landing this and seeing who complains, but if somebody does, we can make this pass X86-specific by adding it to X86TargetMachine::registerPassBuilderCallbacks() (which doesn't exist yet).

Fri, Jun 11, 10:20 AM · Restricted Project, Restricted Project
hoy planned changes to D104129: [CSSPGO] Report zero-count probe in profile instead of dangling probes..

Will come back with more numbers.

Fri, Jun 11, 10:18 AM · Restricted Project
gkm added a comment to D104084: [lld-macho] Add options -(no_)keep_dwarf_unwind.

The intelligent approach isn't _that_ hard, is it? It's < 200 lines in ld64 I think. ld64-linked Chromium Framework leaves some __eh_frame around, so I think removing the whole section generally probably doesn't work.

Fri, Jun 11, 10:18 AM · Restricted Project, Restricted Project
lattner added a comment to D103811: [docs] Set Phabricator as the tool for pre-commit code reviews.

nice, thank you!

Fri, Jun 11, 10:17 AM · Restricted Project
efriedma added a comment to D104007: [BasicAA] Properly mark that coroutine suspension may modify parameters.

Do you know if there is a stable contract in IR to tell whether an argument is in fact a callee-owned memory?

Fri, Jun 11, 10:17 AM · Restricted Project
aeubanks added a comment to D104099: [NewPM] Remove SpeculateAroundPHIs pass.

Some backends don't run SimplifyCFG, e.g. X86. I believe the pass was originally created specifically for X86 (the header has some X86 examples) and may or may not extend to other targets (I'm not very familiar with the pass itself).

Fri, Jun 11, 10:16 AM · Restricted Project, Restricted Project
gkm updated the summary of D104084: [lld-macho] Add options -(no_)keep_dwarf_unwind.
Fri, Jun 11, 10:14 AM · Restricted Project, Restricted Project
wlei requested review of D104131: [CSSPGO] Aggregation by the last K context frames for cold profiles.
Fri, Jun 11, 10:13 AM · Restricted Project
reames added inline comments to D103991: [SCEV] Extend mustprogress reasoning to ne exit tests.
Fri, Jun 11, 10:12 AM · Restricted Project
baziotis added inline comments to D103861: [Attributor] Look through selects in genericValueTraversal.
Fri, Jun 11, 10:12 AM · Restricted Project
efriedma added inline comments to D104075: [ScalarEvolution] Merge howManyGreaterThans with howManyLessThans..
Fri, Jun 11, 10:11 AM · Restricted Project
sdesmalen added a comment to D103939: [SVE][LSR] Teach LSR to enable simple scaled-index addressing mode generation for SVE..

The change looks good to me, just have an open question about the test.

Fri, Jun 11, 10:10 AM · Restricted Project
teemperor added a comment to D103675: [LLDB/API] Expose args and env from SBProcessInfo..

Not sure why this failed. Is it perhaps a difference where some processes have args including the executable and some do not?

Fri, Jun 11, 10:10 AM · Restricted Project
ftynse accepted D104108: [mlir] Add ExpOp to Complex dialect..
Fri, Jun 11, 10:09 AM · Restricted Project
vdmitrie committed rG94a07c79cf10: [SLP][NFC] Fix condition that was supposed to save a bit of compile time. (authored by vdmitrie).
[SLP][NFC] Fix condition that was supposed to save a bit of compile time.
Fri, Jun 11, 10:09 AM
vdmitrie closed D104064: [SLP][NFC] Fix condition that was supposed to save a bit of compile time..
Fri, Jun 11, 10:09 AM · Restricted Project
ABataev updated the diff for D100486: [COST]Improve cost model for shuffles in SLP..

Rebase + fixed gathering cost calculation.

Fri, Jun 11, 10:06 AM · Restricted Project
LLVM GN Syncbot <llvmgnsyncbot@gmail.com> committed rGee98f600de9e: [gn build] Port c54d3050f7b9 (authored by LLVM GN Syncbot <llvmgnsyncbot@gmail.com>).
[gn build] Port c54d3050f7b9
Fri, Jun 11, 10:01 AM
lebedev.ri updated the summary of D104099: [NewPM] Remove SpeculateAroundPHIs pass.
Fri, Jun 11, 10:00 AM · Restricted Project, Restricted Project
aeubanks added reviewers for D104099: [NewPM] Remove SpeculateAroundPHIs pass: asbirlea, echristo.
Fri, Jun 11, 9:58 AM · Restricted Project, Restricted Project
ldionne committed rGc54d3050f7b9: [libc++] NFC: Move indirect_concepts.h to __iterator/concepts.h (authored by ldionne).
[libc++] NFC: Move indirect_concepts.h to __iterator/concepts.h
Fri, Jun 11, 9:57 AM
ldionne closed D104048: [libc++] NFC: Move indirect_concepts.h to __iterator/concepts.h.
Fri, Jun 11, 9:57 AM · Restricted Project
cjdb added a comment to D103478: [libc++][compare] Implement three_way_comparable[_with] concepts.

Ping. @rarutyun this is blocking other work now; will you have time to work on this next week, or would you prefer I take over?

Fri, Jun 11, 9:56 AM · Restricted Project
ldionne added a comment to D104048: [libc++] NFC: Move indirect_concepts.h to __iterator/concepts.h.

LGTM. But btw, what's going on with these GCC failures? I thought you (@ldionne) were going to fix them, but it's been like a month now. Should I make a PR for it? Some of these failures are just straight up typos in the test cases; others might require some digging, I don't know (which is why I was happy to let you deal with it, I thought).

Fri, Jun 11, 9:56 AM · Restricted Project
SouraVX added a comment to D103177: [flang][driver] Extend the `flang` bash script to act as a driver.

Overall I support & I'd like to have one driver i.e flang-new considering all the effort that has gone in it and the long term plan associated with the driver(that will cater to the entire flang project as a primary interface). :)

Fri, Jun 11, 9:56 AM · Restricted Project, Restricted Project
reames requested changes to D104075: [ScalarEvolution] Merge howManyGreaterThans with howManyLessThans..

Several level of comments.

Fri, Jun 11, 9:55 AM · Restricted Project
PeteSteinfeld updated the summary of D104130: [flang] Handle multiple USE statements for the same module.
Fri, Jun 11, 9:55 AM · Restricted Project, Restricted Project
darwin added a comment to D104044: [clang-format] Fix the issue that empty lines being removed at the beginning of namespace.

Devils advocate how is this any different from

class Foo {

class Bar {} ;
}

};

This would become

class Foo {
   class Bar {};
};

i.e. its going to remove the extra lines, just asking so we can understand if the removal of the line is the error or the fact it doesn't remove the line in the first place?

Fri, Jun 11, 9:54 AM · Restricted Project, Restricted Project
teemperor accepted D104054: [lldb] Enable Rust v0 symbol demangling.

LGTM, thanks! (And congrats on your first patch)

Fri, Jun 11, 9:53 AM · Restricted Project
cjdb abandoned D103330: [libcxx][nfc] splices non-modifying algorithms into their own headers.
Fri, Jun 11, 9:53 AM · Restricted Project
PeteSteinfeld requested review of D104130: [flang] Handle multiple USE statements for the same module.
Fri, Jun 11, 9:52 AM · Restricted Project, Restricted Project
AndrewLitteken updated the diff for D87302: [IRSim][IROutliner] Adding DebugInfo handling for IR outlined functions..

Remade tests with regexed metadata information.

Fri, Jun 11, 9:52 AM · debug-info, Restricted Project
RedDocMD updated the diff for D103750: [analyzer] Handle std::make_unique for SmartPtrModeling.

Removed un-necessary check

Fri, Jun 11, 9:51 AM · Restricted Project
MyDeveloperDay awarded D104096: [Clang-Format] Add ReferenceAlignment directive a Like token.
Fri, Jun 11, 9:50 AM · Restricted Project
Carrot added a comment to D101970: [X86FixupLEAs] Transform the sequence LEA/SUB to SUB/SUB.

The fix has been committed as f35bcea1d4748889b8240defdf00cb7a71cbe070.

Fri, Jun 11, 9:50 AM · Restricted Project
reames added a comment to D104066: [SCEV] Use knowledge of stride to prove loops finite for LT exit count computation.

I agree with the request to hold until latent miscompile in underlying code is addressed. I will hold this patch until that happens, and then land without further review. Max, good suggestion on condition ordering, will address in landed patch.

Fri, Jun 11, 9:49 AM · Restricted Project
Carrot committed rGf35bcea1d474: [X86FixupLEAs] Sub register usage of LEA dest should block LEA/SUB optimization (authored by Carrot).
[X86FixupLEAs] Sub register usage of LEA dest should block LEA/SUB optimization
Fri, Jun 11, 9:46 AM
Carrot closed D103922: [X86FixupLEAs] Sub register usage of LEA dest should block LEA/SUB optimization.
Fri, Jun 11, 9:46 AM · Restricted Project
MyDeveloperDay added a comment to D104096: [Clang-Format] Add ReferenceAlignment directive.

A new nits but I'd say it was pretty good.

Fri, Jun 11, 9:46 AM · Restricted Project
ldionne committed rGf84dbd2f2bef: [libc++] Enable the synchronization library on Apple platforms (authored by ldionne).
[libc++] Enable the synchronization library on Apple platforms
Fri, Jun 11, 9:45 AM
ldionne closed D96790: [libc++] Enable the synchronization library on Apple platforms.
Fri, Jun 11, 9:45 AM · Restricted Project
asm updated the diff for D104054: [lldb] Enable Rust v0 symbol demangling.

code style fixes, use modern log

Fri, Jun 11, 9:44 AM · Restricted Project
hoy added reviewers for D104129: [CSSPGO] Report zero-count probe in profile instead of dangling probes.: wenlei, wlei, wmi.
Fri, Jun 11, 9:44 AM · Restricted Project
Meinersbur added a comment to D104126: [MicroBenchmarks] Add initial SLP vectorization benchmarks..

Does this expect on loop unrolling to happened? Did you consider adding #pragma clang loop unroll(full) to make sure of it?

Fri, Jun 11, 9:42 AM
hoy requested review of D104129: [CSSPGO] Report zero-count probe in profile instead of dangling probes..
Fri, Jun 11, 9:42 AM · Restricted Project
LLVM GN Syncbot <llvmgnsyncbot@gmail.com> committed rG2244a0f5fe62: [gn build] Port 9106047ee3dd (authored by LLVM GN Syncbot <llvmgnsyncbot@gmail.com>).
[gn build] Port 9106047ee3dd
Fri, Jun 11, 9:38 AM
dmgreen added a comment to D104042: [AArch64] Improve SAD pattern.

It is good point! I have tried below pattern following your suggestion. It seems to work. If you are ok, let me add below pattern in this patch.

let AddedComplexity = 10 in { 
def : Pat<(i32 (extractelt
                 (v4i32 (AArch64uaddv (v4i32 (AArch64uaddlp (v8i16 V128:$op))))),
                 (i64 0))),
          (UADDLVv8i16v V128:$op)>;
}

Do you mind doing this as a new patch? As it does feel logically separable. If we can test them, it would be good to add the various other sizes too. And, I'm not sure about this, but maybe it doesn't need to start from the extract, and can produce a INSERT_SUBREG like some of the other patterns do (like the ones from SIMDAcrossLanesIntrinsic). That might remove the need for the added complexity, and the INSERT_SUBREG / EXRACT_SUBREG should all get cleared up later in the pipeline.

Fri, Jun 11, 9:38 AM · Restricted Project
zoecarver added a comment to D102028: [libcxx][ranges] Implement views::all..

@Quuxplusone friendly ping. I'm going to rebase this shortly. After the tests pass I'm planning to land it.

Fri, Jun 11, 9:36 AM · Restricted Project