Page MenuHomePhabricator
Feed All Stories

Mon, Jul 22

ibiryukov committed rL366674: [clangd] Log input code of failed highlighting tests. NFC.
[clangd] Log input code of failed highlighting tests. NFC
Mon, Jul 22, 3:07 AM
teemperor committed rGf94668e3360e: [lldb][NFC] Tablegenify breakpoint (authored by teemperor).
[lldb][NFC] Tablegenify breakpoint
Mon, Jul 22, 3:02 AM
grimar committed rL366668: [llvm-readobj] - Stop using precompiled objects in file-headers.test.
[llvm-readobj] - Stop using precompiled objects in file-headers.test
Mon, Jul 22, 3:02 AM
ilya-biryukov committed rG0a42fe70a566: [AST] Treat semantic form of InitListExpr as implicit code in traversals (authored by ilya-biryukov).
[AST] Treat semantic form of InitListExpr as implicit code in traversals
Mon, Jul 22, 3:02 AM
teemperor committed rL366673: [lldb][NFC] Tablegenify breakpoint.
[lldb][NFC] Tablegenify breakpoint
Mon, Jul 22, 3:02 AM
ibiryukov committed rL366672: [AST] Treat semantic form of InitListExpr as implicit code in traversals.
[AST] Treat semantic form of InitListExpr as implicit code in traversals
Mon, Jul 22, 3:02 AM
Diffusion closed D64762: [AST] Treat semantic form of InitListExpr as implicit code in traversals.
Mon, Jul 22, 3:02 AM · Restricted Project, Restricted Project
timcorringham updated the diff for D64557: Add llvm.loop.licm.disable metadata.

Changed the metadata name to llvm.licm.disable

Mon, Jul 22, 3:02 AM · Restricted Project
kren1 created D65070: [LLVM-C][OCaml] Add a fast linker binding.
Mon, Jul 22, 3:02 AM · Restricted Project
sgraenitz committed rG3a52e50d7379: Add location of SVN staging dir to git-llvm error output (authored by sgraenitz).
Add location of SVN staging dir to git-llvm error output
Mon, Jul 22, 2:50 AM
stefan.graenitz committed rL366671: Add location of SVN staging dir to git-llvm error output.
Add location of SVN staging dir to git-llvm error output
Mon, Jul 22, 2:50 AM
Diffusion closed D65038: Add location of SVN staging dir to git-llvm error output.
Mon, Jul 22, 2:50 AM · Restricted Project
mantognini closed D64569: [OpenCL] Improve destructor support in C++ for OpenCL.
Mon, Jul 22, 2:50 AM · Restricted Project, Restricted Project
samparker created D65069: [ARM][ParallelDSP] Improve smlad search.
Mon, Jul 22, 2:50 AM
ilya-biryukov added inline comments to D64762: [AST] Treat semantic form of InitListExpr as implicit code in traversals.
Mon, Jul 22, 2:45 AM · Restricted Project, Restricted Project
ilya-biryukov updated the diff for D64762: [AST] Treat semantic form of InitListExpr as implicit code in traversals.
  • Rewrite code as suggested in the review
Mon, Jul 22, 2:45 AM · Restricted Project, Restricted Project
martong added inline comments to D65064: [CrossTU] Add a function to retrieve original source location..
Mon, Jul 22, 2:45 AM · Restricted Project, Restricted Project
dnsampaio updated the diff for D65000: [ARM] Set default alignment to 64bits.
  • Set androideabi alignment to 128 bits
Mon, Jul 22, 2:45 AM · Restricted Project, Restricted Project
gchatelet updated the diff for D64790: [LLVM][NFC] Adding an Alignment type to LLVM.
  • Enable asserts since we will now fix the code while introducing the
Mon, Jul 22, 2:45 AM · Restricted Project
mantognini added an edge to rL366670: [OpenCL] Improve destructor support in C++ for OpenCL: D64569: [OpenCL] Improve destructor support in C++ for OpenCL.
Mon, Jul 22, 2:45 AM
mantognini added 3 commit(s) for D64569: [OpenCL] Improve destructor support in C++ for OpenCL: rG88559637641e: [OpenCL] Improve destructor support in C++ for OpenCL, rC366670: [OpenCL] Improve destructor support in C++ for OpenCL, rL366670: [OpenCL] Improve destructor support in C++ for OpenCL.
Mon, Jul 22, 2:45 AM · Restricted Project, Restricted Project
mantognini added an edge to rG88559637641e: [OpenCL] Improve destructor support in C++ for OpenCL: D64569: [OpenCL] Improve destructor support in C++ for OpenCL.
Mon, Jul 22, 2:45 AM
mantognini committed rG88559637641e: [OpenCL] Improve destructor support in C++ for OpenCL (authored by mantognini).
[OpenCL] Improve destructor support in C++ for OpenCL
Mon, Jul 22, 2:42 AM
mantognini committed rL366670: [OpenCL] Improve destructor support in C++ for OpenCL.
[OpenCL] Improve destructor support in C++ for OpenCL
Mon, Jul 22, 2:42 AM
grimar updated the summary of D65068: [Object/llvm-readobj] - Cleanup testing of the dynamic objects..
Mon, Jul 22, 2:41 AM · Restricted Project
martong accepted D65064: [CrossTU] Add a function to retrieve original source location..

LGTM! Thanks!

Mon, Jul 22, 2:38 AM · Restricted Project, Restricted Project
grimar updated the summary of D65068: [Object/llvm-readobj] - Cleanup testing of the dynamic objects..
Mon, Jul 22, 2:38 AM · Restricted Project
grimar created D65068: [Object/llvm-readobj] - Cleanup testing of the dynamic objects..
Mon, Jul 22, 2:38 AM · Restricted Project
lenary added inline comments to D54296: [RISCV] Lower inline asm constraint A for RISC-V.
Mon, Jul 22, 2:38 AM · Restricted Project
nhaehnle added inline comments to D64862: AMDGPU/GlobalISel: RegBankSelect interp intrinsics.
Mon, Jul 22, 2:38 AM
jhenderson added inline comments to D64999: [yaml2obj] - Change how we handle implicit sections..
Mon, Jul 22, 2:38 AM · Restricted Project
jhenderson added a comment to D64800: [llvm-readobj] - Stop using precompiled objects in file-headers.test.

Is there a danger that by moving in this direction we loose some test coverage since we are no longer verifying the libObject can read a know set of binaries, but only that libObject can read the output yaml2obj? i.e. we could evolve the object format in llvm such that these tests would continue pass but the objects would no longer be standards compliant?

I think the pre-canned binaries don't really help us any more than yaml2obj inputs - it would be just as straightforward to tune our object reading code to work with a very specific set of binaries (unintentionally or otherwise) as it would be to work with our yaml2obj output. It's worth noting that the yaml2obj writing code and llvm-readobj reading code are not the same set of code, so we are already testing two different implementations against each other. If we change one and not the other, a test should break, alerting us to a behaviour change. We would need a good reason to change both.

Mon, Jul 22, 2:38 AM
hokein created D65067: [clangd] Fix auto expand not work on "const au^to s = 123;".
Mon, Jul 22, 2:29 AM · Restricted Project
martong resigned from D65042: [Concept] Placeholder constraints and abbreviated templates.
Mon, Jul 22, 2:28 AM · Restricted Project
nhaehnle accepted D64901: [AMDGPU][NFC] Simplify test file for amdgcn intrinsics.

LGTM

Mon, Jul 22, 2:28 AM · Restricted Project
martong added a comment to D65042: [Concept] Placeholder constraints and abbreviated templates.

ASTImporter.cpp and ASTStructuralEquivalence.cpp looks good to me!

Mon, Jul 22, 2:26 AM · Restricted Project
nhaehnle accepted D64919: TableGen: Support physical register inputs > 255.

LGTM

Mon, Jul 22, 2:25 AM
jhenderson added a comment to D62462: [llvm-objdump] Add warning messages if disassembly + source for problematic inputs.

@jhenderson, can you think of an input that would cause Symbolizer->symbolizeCode(*Obj, Address) to return an Error?

I went a few calls deep underneath Symbolizer->symbolizeCode(*Obj, Address) and it looks like LLVMSymbolizer and DWARFContext return a default-constructed DILineInfo when they fail to parse the debug info. Can we cause a failure that returns an Error instead? If so we can check whether the Error is actually an ErrorInfo subclass which is how custom errors are created according to error handling Detailed Description. A better implementation of SourcePrinter::printSourceLine() might call handleErrors() instead of consumeError(). Grepping the code for handleErrors and handleAllErrors provides some example uses. First I need some inputs to cause errors so I can see whether any ErrorInfo subclasses are returned and if they have useful context which we can incorporate into our warning. This is my first time wading into LLVM's error handling so please let me know if you have any reading suggestions. In addition to the doxygen docs for Error.h I also found a section on error handling in the programmer's manual.

Mon, Jul 22, 2:16 AM · Restricted Project
nhaehnle added a comment to D64726: AMDGPU/GlobalISel: Fix not constraining result reg of copies to VCC.

Okay, the possibility of an AssertZext is an interesting point. So let me try the other way around: What would the MIR at this stage look like to enforce an and?

Mon, Jul 22, 2:11 AM
peter.smith added reviewers for D65000: [ARM] Set default alignment to 64bits: srhines, danalbert.

I think that this may not apply for Android as AFAIK their ABI still requires 128-bit alignment in some cases. Adding some more reviewers from Android.

Mon, Jul 22, 2:08 AM · Restricted Project, Restricted Project
skatkov updated the diff for D64983: [Loop Peeling] Introduce an option for full peeling disabling.

the variable and option name changed.

Mon, Jul 22, 2:04 AM · Restricted Project
skatkov updated the diff for D64972: [Loop Peeling] Do not close further unroll/peel if profile based peeling was not used.

Test is updated to show what this patch changes.

Mon, Jul 22, 2:03 AM · Restricted Project
skatkov added a comment to D64972: [Loop Peeling] Do not close further unroll/peel if profile based peeling was not used.

I'm not quite sure this is the right macro approach.

The comment near the code your modifying makes me thing the whole reasoning behind disabling future peeling and unrolling after the first peel may be flawed. It doesn't make sense to "use up" profiling information. Assuming we correctly updated our profiling when doing the transform, the resulting profile for the loop should indicate that it's cold and thus not profitable to further peel/unroll. If that's not happening, maybe there's another issue in play? (I wonder if your other profile bug fix may help here?)

p.s. I'm opened to being convinced that this is a practical answer, even if not an ideal one. Just make the argument. :)

Mon, Jul 22, 2:02 AM · Restricted Project
ostannard committed rG6771a89fa01f: [IPRA][ARM] Make use of the "returned" parameter attribute (authored by ostannard).
[IPRA][ARM] Make use of the "returned" parameter attribute
Mon, Jul 22, 1:58 AM
rampitec added a comment to D64946: [AMDGPU] Fix trivial PHI into SI_END_CF..

How about the following simpler logic:

  • if the PHI is used by any basic-block prologue instruction (except other PHIs), then insert the COPY at the top of the basic block
  • otherwise, insert the COPY after the basic-block prologue

In this case COPY in the prologue also shall be marked as prologue instruction somehow.

You're right, my idea doesn't work.

Mon, Jul 22, 1:57 AM · Restricted Project
ostannard closed D64986: [IPRA][ARM] Make use of the "returned" parameter attribute.
Mon, Jul 22, 1:51 AM · Restricted Project
jhenderson added a comment to D65040: [llvm-objcopy] Add support for --add-section for COFF.

Please make sure to update the documentation at llvm/docs/CommandGuide/llvm-objcopy.rst to move this switch out of the ELF-specific directory.

Mon, Jul 22, 1:51 AM · Restricted Project
grimar added inline comments to D64999: [yaml2obj] - Change how we handle implicit sections..
Mon, Jul 22, 1:51 AM · Restricted Project
nhaehnle added a comment to D64935: [AMDGPU] Add llvm.amdgcn.softwqm intrinsic.

Okay thanks, I see the logic now.

Mon, Jul 22, 1:48 AM · Restricted Project
gchatelet added a comment to D64786: [Polly] Clang complains about missing virtual destructor.

A gentle ping. It's a one liner and should be reviewed easily :)

Mon, Jul 22, 1:48 AM · Restricted Project
ostannard added a comment to D63932: [GlobalDCE] Dead Virtual Function Elimination.

Ping.

Mon, Jul 22, 1:48 AM · Restricted Project, Restricted Project
nhaehnle added a comment to D64946: [AMDGPU] Fix trivial PHI into SI_END_CF..

How about the following simpler logic:

  • if the PHI is used by any basic-block prologue instruction (except other PHIs), then insert the COPY at the top of the basic block
  • otherwise, insert the COPY after the basic-block prologue

In this case COPY in the prologue also shall be marked as prologue instruction somehow.

Mon, Jul 22, 1:48 AM · Restricted Project
nhaehnle added a comment to D64508: AMDGPU: Force s_waitcnt after GWS instructions.

My understanding is that this is mostly related to CWSR. The trap handler has to be able to "replay" the GWS instruction.

Mon, Jul 22, 1:40 AM
jhenderson added inline comments to D64999: [yaml2obj] - Change how we handle implicit sections..
Mon, Jul 22, 1:39 AM · Restricted Project
wuzish added a comment to D64195: [PowerPC] Replace float load/store pair with integer load/store pair when it's only used in load/store.

Gentle pin..

Mon, Jul 22, 1:39 AM · Restricted Project
nhaehnle added a comment to D63281: [TargetLowering] Add SimplifyMultipleUseDemandedBits.

The AMDGPU changes seem fine to me overall.

Mon, Jul 22, 1:39 AM · Restricted Project
avl added a comment to D64595: [Debuginfo][SROA] Need to handle dbg.value in SROA pass..

Converting a dbg.value into a dbg.declare implicitly extends the duration ("lifetime" in the docs [0]) of the variable location from "until the next dbg.value" to "the entire scope". Consider what happens if we make the modifications here [1] to your test case. Imagine an IR producer that wants to temporarily bind some variable to a field of 'result' (the dbg.value added with DW_OP_deref), and later assigns '1' to the variable with the second dbg.value, to cover the rest of the function.

Mon, Jul 22, 1:32 AM · Restricted Project, debug-info
nhaehnle added a comment to D64954: [IR][Verifier] Allow IntToPtrInst to be !dereferenceable.

Thanks! Could you please also add a test to Analysis/ValueTracking/memory-dereferenceable.ll?

Mon, Jul 22, 1:31 AM · Restricted Project
avl updated the diff for D64595: [Debuginfo][SROA] Need to handle dbg.value in SROA pass..

added check for aggregate dbg.value.

Mon, Jul 22, 1:27 AM · Restricted Project, debug-info
ostannard accepted D64406: [ARM][AArch64] Cortex-A65AE, Neoverse E1 and Neoverse N1 support.

LGTM

Mon, Jul 22, 1:22 AM · Restricted Project
grimar committed rG6522a7df544e: [llvm-readobj] - Stop using precompiled objects in file-headers.test (authored by grimar).
[llvm-readobj] - Stop using precompiled objects in file-headers.test
Mon, Jul 22, 1:20 AM
dmgreen added a parent revision for D65066: [ARM] Disable MVE fptosi and friends: D65062: [ARM] More MVE compare vector splat combines for AND's.
Mon, Jul 22, 1:17 AM · Restricted Project
grimar closed D64800: [llvm-readobj] - Stop using precompiled objects in file-headers.test.

r366668

Mon, Jul 22, 1:17 AM
dmgreen added a child revision for D65062: [ARM] More MVE compare vector splat combines for AND's: D65066: [ARM] Disable MVE fptosi and friends.
Mon, Jul 22, 1:17 AM · Restricted Project
dmgreen created D65066: [ARM] Disable MVE fptosi and friends.
Mon, Jul 22, 1:17 AM · Restricted Project
grimar added a comment to D64800: [llvm-readobj] - Stop using precompiled objects in file-headers.test.

I am going to commit this to free road for another patches. If during futher discussion(s)
will be decided to keep these or other binaries checked in for something, we can do it
in a follow-up independent commit(s) with appropriate test case(s) and comment(s) about
how exactly binaries were produced and for what we keep them.

Mon, Jul 22, 1:17 AM
bruntib created D65065: [clang-tidy] Possibility of displaying duplicate warnings.
Mon, Jul 22, 1:08 AM · Restricted Project
SjoerdMeijer accepted D65062: [ARM] More MVE compare vector splat combines for AND's.

looks okay to me.

Mon, Jul 22, 1:04 AM · Restricted Project
SjoerdMeijer accepted D65061: [ARM] MVE compare vector splat combine.

LGTM

Mon, Jul 22, 1:04 AM · Restricted Project
SjoerdMeijer added inline comments to D65059: [ARM] Better OR's for MVE compares.
Mon, Jul 22, 12:58 AM · Restricted Project
sdesmalen added inline comments to D64739: [SVE][Inline-Asm] Add support to specify SVE registers in the clobber list.
Mon, Jul 22, 12:58 AM · Restricted Project, Restricted Project
grimar added inline comments to D64999: [yaml2obj] - Change how we handle implicit sections..
Mon, Jul 22, 12:50 AM · Restricted Project
SjoerdMeijer accepted D65058: [ARM] Better AND's for MVE compares.

Looks like a good optimisation to me.

Mon, Jul 22, 12:50 AM · Restricted Project
grimar updated the diff for D64999: [yaml2obj] - Change how we handle implicit sections..
  • Addressed review comments.
Mon, Jul 22, 12:50 AM · Restricted Project
SjoerdMeijer accepted D65053: [ARM] Basic And/Or/Xor handling for MVE predicates.

looks reasonable

Mon, Jul 22, 12:50 AM · Restricted Project
balazske added a comment to D64554: [CrossTU] Add a function to retrieve original source location..

New patch:
https://reviews.llvm.org/D65064

Mon, Jul 22, 12:50 AM · Restricted Project, Restricted Project

Jul 22 2019

balazske added a child revision for D65064: [CrossTU] Add a function to retrieve original source location.: D64638: [CrossTU] Fix plist macro expansion if macro in other file..
Jul 22 2019, 12:44 AM · Restricted Project, Restricted Project
balazske added a parent revision for D64638: [CrossTU] Fix plist macro expansion if macro in other file.: D65064: [CrossTU] Add a function to retrieve original source location..
Jul 22 2019, 12:43 AM · Restricted Project, Restricted Project
SjoerdMeijer accepted D65049: [ARM] Fix for MVE VPT block pass.

Thanks for fixing this.

Jul 22 2019, 12:39 AM · Restricted Project
Herald updated subscribers of D65064: [CrossTU] Add a function to retrieve original source location..
Jul 22 2019, 12:39 AM · Restricted Project, Restricted Project
balazske created D65064: [CrossTU] Add a function to retrieve original source location..
Jul 22 2019, 12:38 AM · Restricted Project, Restricted Project
foad committed rG298500ae3319: [AMDGPU] Save some work when an atomic op has no uses (authored by foad).
[AMDGPU] Save some work when an atomic op has no uses
Jul 22 2019, 12:24 AM
foad committed rL366667: [AMDGPU] Save some work when an atomic op has no uses.
[AMDGPU] Save some work when an atomic op has no uses
Jul 22 2019, 12:19 AM
foad closed D64981: [AMDGPU] Save some work when an atomic op has no uses.
Jul 22 2019, 12:19 AM · Restricted Project
foad added a comment to rL366585: AMDGPU/GlobalISel: Selection for fminnum/fmaxnum.

The added tests are failing for me, both on NetBSD and on one of my Linux systems:
http://lab.llvm.org:8011/builders/netbsd-amd64/builds/21070/steps/run%20unit%20tests

Jul 22 2019, 12:15 AM
MaskRay added a comment to D64968: [Object] Create MutableObject class for doing Object Mutations [Part 2].

If you have branches a, b, and c (a depends on master, b depends on a, c depends on b):

Jul 22 2019, 12:15 AM · Restricted Project

Jul 21 2019

grimar added inline comments to D64913: [yaml2obj] - Add a support for defining null sections in YAMLs..
Jul 21 2019, 11:47 PM · Restricted Project
MaskRay added inline comments to D64999: [yaml2obj] - Change how we handle implicit sections..
Jul 21 2019, 11:22 PM · Restricted Project
seiya added a comment to D64969: [llvm-objdump][NFC] Make the PrettyPrinter::printInst() output buffered.

Nice finding!

[1]: https://github.com/llvm/llvm-project/blob/2946cd701067404b99c39fb29dc9c74bd7193eb3/llvm/include/llvm/Support/FormattedStream.h#L83

I think the 4x (line buffered vs unbuffered) slowdown can be explained by your first link. The formatted_raw_ostream ctor resets the underlying stream (outs()) to unbuffered. There are several solutions to the problem. Given formatted_raw_ostream FOS(OS); is only used once in PrettyPrinter and can be easily replaced, I suggest deleting it and tracking the column manually.

Jul 21 2019, 11:13 PM · Restricted Project
MaskRay added a comment to D64969: [llvm-objdump][NFC] Make the PrettyPrinter::printInst() output buffered.

Nice finding!

Jul 21 2019, 11:07 PM · Restricted Project
MaskRay updated the diff for D64906: [ELF][PPC] Allow PT_LOAD to have overlapping p_offset ranges.

Fix gcc 8 -Wparentheses

Jul 21 2019, 10:52 PM · Restricted Project
seiya added a comment to D64969: [llvm-objdump][NFC] Make the PrettyPrinter::printInst() output buffered.

I've added a few more reviewers for visibility.

My initial thought was "why not just delete the "flush"" from printInst, but I see that it's there for a "good" reason, looking at PrettyPrinter. I can't help but feel that a better solution would be to fix formatted_raw_ostream so that the indentation only gets added after any already-written content. Assuming that's not viable (e.g. it's intentional behaviour), I think what you have written works. @seiya, what are your thoughts on the formatted_raw_ostream implementation? Can it be sensibly changed?

First I thought that it is not straightforward to update the position (what FOS.getColumn() returns) in formatted_raw_stream without flushing because it's done in its write_impl.

However, now I realized that it's trivial: add new method formatted_raw_ostream::updatePositionByBufferedString() (I couldn't come up with a good name) method, which reads the buffered string and updates the position by doing ComputePosition(getBufferStart(), GetNumBytesInBuffer()); .

I haven't tried this, but with this method, we no longer need to add a string buffer like this current patch. What do you think on this approach?

It's hard to envisage this without seeing a patch with it in. Could you create a separate patch to this one, with your alternative proposal in, so that I and others can compare and contrast the two options, please?

Jul 21 2019, 10:46 PM · Restricted Project
MaskRay added inline comments to D64913: [yaml2obj] - Add a support for defining null sections in YAMLs..
Jul 21 2019, 10:44 PM · Restricted Project
lkail committed rG3d72a58981e9: [PowerPC][NFC] Precommit a test case where ppc-mi-peepholes miscompiles extswsli (authored by lkail).
[PowerPC][NFC] Precommit a test case where ppc-mi-peepholes miscompiles extswsli
Jul 21 2019, 10:33 PM
lkail committed rL366666: [PowerPC][NFC] Precommit a test case where ppc-mi-peepholes miscompiles extswsli.
[PowerPC][NFC] Precommit a test case where ppc-mi-peepholes miscompiles extswsli
Jul 21 2019, 10:32 PM
MaskRay added a comment to D64913: [yaml2obj] - Add a support for defining null sections in YAMLs..

It says (https://docs.oracle.com/cd/E19683-01/817-3677/6mj8mbtc9/index.html):

Jul 21 2019, 10:29 PM · Restricted Project
craig.topper added inline comments to D63847: [MC] Add MCInstrAnalysis::evaluateMemoryOperandAddress.
Jul 21 2019, 10:29 PM · Restricted Project
wuzish created D65063: [PowerPC] Eliminate loads/swap feeding swap/store for vector type by using big-endian load/store.
Jul 21 2019, 10:26 PM · Restricted Project
skatkov committed rGc6c31da86773: [Loop Peeling] Fix the handling of branch weights of peeled off branches. (authored by skatkov).
[Loop Peeling] Fix the handling of branch weights of peeled off branches.
Jul 21 2019, 10:17 PM
skatkov committed rL366665: [Loop Peeling] Fix the handling of branch weights of peeled off branches..
[Loop Peeling] Fix the handling of branch weights of peeled off branches.
Jul 21 2019, 10:17 PM