vsk (Vedant Kumar)
User

Projects

User Details

User Since
Jul 8 2015, 10:26 AM (167 w, 1 d)

Recent Activity

Yesterday

vsk updated the diff for D50478: Add support for artificial tail call frames.

Teach SBThread::StepOut and SBThread::ReturnFromFrame to behave as-if artificial frames were not present.

Thu, Sep 20, 6:42 PM
vsk planned changes to D50478: Add support for artificial tail call frames.

Can you add a test that makes sure that when you stop in a frame that has artificial frames above it, and then you do "finish", or "step out" past the end of frame 0, the presence of the artificial frame doesn't confuse us? I am pretty sure that will just work, but it would be good to make sure that it actually does.

Really "step out past the end of the frame" isn't really necessary, since that doesn't pay any attention to the frames above it, it just runs the code till the frame ID changes. But "finish" does look at the parent frame, so it might get confused.

Thu, Sep 20, 5:12 PM
vsk added inline comments to D52033: [GCOV] Add options to filter files which must be instrumented..
Thu, Sep 20, 4:34 PM
vsk added inline comments to D52034: [Clang] Add options -Xclang -coverage-filter and -Xclang -coverage-exclude to filter the files to instrument with gcov.
Thu, Sep 20, 4:34 PM
vsk added inline comments to D52034: [Clang] Add options -Xclang -coverage-filter and -Xclang -coverage-exclude to filter the files to instrument with gcov.
Thu, Sep 20, 4:21 PM
vsk updated the diff for D50478: Add support for artificial tail call frames.

I've added SB API support (SBFrame::IsArtificial), a SB API test, fleshed out the remaining tests, and rebased. PTAL, thanks!

Thu, Sep 20, 4:04 PM
vsk added a comment to D49915: [IR] Add a boolean field in DILocation to know if a line must covered or not.

This caused a backwards-compatibility issue which made llvm-dis crash when loading old bitcode (see https://reviews.llvm.org/rL342631#inline-2294). I've committed a fix in r342678 to get our bots back to a healthy state. I'm open to any post-commit review -- while I don't anticipate there being an issue, if there is one I'd rather iterate on a passing build.

Thu, Sep 20, 12:05 PM · debug-info
vsk added inline comments to rL342631: [IR] Add a boolean field in DILocation to know if a line must covered or not.
Thu, Sep 20, 11:39 AM

Wed, Sep 19

vsk added inline comments to D51813: [Util] Refer to [s|z]exts of args when converting dbg.declares (fix PR35400).
Wed, Sep 19, 11:20 AM · debug-info

Tue, Sep 18

vsk added a comment to D52236: [ADT][BitVector] Add push_back().

LGTM. As this is somewhat fundamental change another +1 would be nice.

Tue, Sep 18, 11:13 AM
vsk accepted D51976: [DebugInfo][Dexter] Speculated BB presents illegal variable value to debugger.

Thanks, LGTM. @aprantl?

Tue, Sep 18, 11:06 AM · debug-info

Mon, Sep 17

vsk created D52199: [profile] Install headers for custom runtime maintainers.
Mon, Sep 17, 5:28 PM
vsk updated subscribers of D46100: WIP: [IR/Verifier] Diagnose use-before-def scenarios for debug intrinsics.
Mon, Sep 17, 2:32 PM
vsk added a comment to D51976: [DebugInfo][Dexter] Speculated BB presents illegal variable value to debugger.

@aprantl and @probinson,

Based on your views and comments, I am happy to "change" the current fix towards implementing the extension of 'llvm.dbg.value', as it seems it would be more beneficial for a better debug information.

The added comments (https://bugs.llvm.org/show_bug.cgi?id=38763) by @aprantl, are a good starting point.

@vsk, how do you feel about moving the current proposed patch, to be in terms of extending the 'llvm.dbg.value'?

Mon, Sep 17, 2:06 PM · debug-info
vsk added a comment to D51813: [Util] Refer to [s|z]exts of args when converting dbg.declares (fix PR35400).

Ping.

Mon, Sep 17, 10:46 AM · debug-info

Fri, Sep 14

vsk updated the diff for D50751: Allow use of self.filecheck in LLDB tests (c.f self.expect).

Sorry for the delay, I was busy with other work. I think I've addressed the review feedback. PTAL.

Fri, Sep 14, 5:01 PM · Unknown Object (Project)
vsk added a comment to D52125: X86, AArch64, ARM: Do not attach debug location to spill/reload instructions.

As Matthias points out, a spill/reload can't unambiguously be associated with a specific instruction.

Fri, Sep 14, 3:31 PM
vsk added a comment to D51976: [DebugInfo][Dexter] Speculated BB presents illegal variable value to debugger.

Thanks, this is looking pretty close. A few comments inline.

Fri, Sep 14, 12:10 PM · debug-info
vsk accepted D52112: Fix debug info for SelectionDAG legalization of DAG nodes with two results..

LGTM, possibly with an added assert. The test looks great!

Fri, Sep 14, 11:50 AM · debug-info

Thu, Sep 13

vsk added inline comments to D52064: [Sema] Add a note pointing to the first use of an implicit capture.
Thu, Sep 13, 5:13 PM
vsk updated the diff for D52064: [Sema] Add a note pointing to the first use of an implicit capture.
Thu, Sep 13, 5:13 PM
vsk created D52064: [Sema] Add a note pointing to the first use of an implicit capture.
Thu, Sep 13, 4:38 PM
vsk added a comment to D52034: [Clang] Add options -Xclang -coverage-filter and -Xclang -coverage-exclude to filter the files to instrument with gcov.

Please document the filter behavior (see docs/UsersManual.rst).

Thu, Sep 13, 12:38 PM
vsk added inline comments to D52033: [GCOV] Add options to filter files which must be instrumented..
Thu, Sep 13, 12:28 PM
vsk accepted D51968: [deadargelim] Update dbg.value of 'unused' parameters.

Thanks, LGTM as well.

Thu, Sep 13, 11:46 AM · debug-info
vsk added a comment to D51520: Add libc++ data formatter for std::variant.

Please clang-format your diffs.

Thu, Sep 13, 11:29 AM

Wed, Sep 12

vsk accepted D51648: [UBSan] Partially fix `test/ubsan/TestCases/Misc/log-path_test.cc` so that it can run on devices..

I think what %device_rm is doing is reasonably clear. I'd rather have this available than rely on the device correctly deleting files as needed.

Wed, Sep 12, 12:15 PM
vsk added a comment to D51974: [GCOV] Remove useless loop when emiting edge counters.

It would help to have a regression test for this (somewhere in test/Transforms/GCOVProfiling). Something that exercises emitProfileArcs with 2 CU's would work.

Wed, Sep 12, 11:58 AM
vsk added a comment to D51968: [deadargelim] Update dbg.value of 'unused' parameters.

Thanks, this looks mostly good, minor comments for the testcase inline.

Why doesn't this do the right thing automatically though? What does the dbg.value point to after %k is dropped?

Wed, Sep 12, 11:26 AM · debug-info

Tue, Sep 11

vsk added a comment to D51619: [gcov] Fix branch counters with switch statements.

This needed a small follow-up fix: see r341985, http://lab.llvm.org:8011/builders/clang-cmake-x86_64-sde-avx512-linux/builds/12824/.

Tue, Sep 11, 12:45 PM
vsk added a comment to D51930: Add a basic test for 'memory region'.

This LGTM. Davide?

Tue, Sep 11, 11:49 AM
vsk added a comment to D51930: Add a basic test for 'memory region'.

I think it'd be useful to test the driver output specifically. The kind of testing lldb-test facilitates might not be a good fit here (too low-level).

Tue, Sep 11, 11:48 AM
vsk closed D51620: [gcov] Fix the switch tests.

Committed in r341977.

Tue, Sep 11, 11:41 AM
vsk accepted D51620: [gcov] Fix the switch tests.

Looks great, thanks.

Tue, Sep 11, 11:12 AM
vsk accepted D51619: [gcov] Fix branch counters with switch statements.

Thanks, LGTM.

Tue, Sep 11, 11:03 AM
vsk accepted D51881: [ADT] Made numerous methods of ImmutableList const.

Thanks, LGTM as well.

Tue, Sep 11, 10:46 AM

Mon, Sep 10

vsk added inline comments to D51520: Add libc++ data formatter for std::variant.
Mon, Sep 10, 4:56 PM
vsk added inline comments to D51619: [gcov] Fix branch counters with switch statements.
Mon, Sep 10, 4:31 PM
vsk added inline comments to D51881: [ADT] Made numerous methods of ImmutableList const.
Mon, Sep 10, 1:25 PM

Fri, Sep 7

vsk added inline comments to D51813: [Util] Refer to [s|z]exts of args when converting dbg.declares (fix PR35400).
Fri, Sep 7, 2:03 PM · debug-info
vsk created D51813: [Util] Refer to [s|z]exts of args when converting dbg.declares (fix PR35400).
Fri, Sep 7, 1:57 PM · debug-info

Tue, Aug 28

vsk updated the diff for D49887: [DebugInfo] Add support for DWARF5 call site-related attributes.

@JDevlieghere thanks for the feedback, and no worries!

Tue, Aug 28, 11:52 AM · debug-info
vsk updated the diff for D50927: [Sema] Remove location from implicit capture init expr.

Address the latest round of review feedback.

Tue, Aug 28, 11:39 AM
vsk added inline comments to D50927: [Sema] Remove location from implicit capture init expr.
Tue, Aug 28, 11:39 AM
vsk accepted D50887: [DWARF] Missing location debug information with -O2..

Thanks, LGTM. If you don't have commit access yet [1], I'd be happy to commit this on your behalf.

Tue, Aug 28, 10:49 AM

Mon, Aug 27

vsk accepted D51319: Use a RAII guard to control access to DisassemblerLLVMC..

Looks great, thanks!

Mon, Aug 27, 4:57 PM
vsk added inline comments to D50927: [Sema] Remove location from implicit capture init expr.
Mon, Aug 27, 3:36 PM
vsk updated the diff for D50927: [Sema] Remove location from implicit capture init expr.
  • Partially address some of @rsmith's feedback. Instead of using the capture default location, I used the beginning location of the capture list. This results in smoother single-stepping when those two locations are on different lines.
Mon, Aug 27, 3:36 PM
vsk added inline comments to D51319: Use a RAII guard to control access to DisassemblerLLVMC..
Mon, Aug 27, 1:51 PM
vsk added a comment to D50927: [Sema] Remove location from implicit capture init expr.

Ping.

Mon, Aug 27, 1:28 PM
vsk updated the diff for D51239: [ubsan] Enable -fsanitize=vptr on Apple devices and simulators.

Address some review feedback.

Mon, Aug 27, 1:14 PM
vsk accepted D51271: [UBSan] Add missing `%run` prefixes to Pointer tests..

Thanks!

Mon, Aug 27, 10:25 AM
vsk accepted D51273: [UBSan] Propagate `UBSAN_OPTIONS` environment variable when running ios simulator tests..

Thanks, LGTM.

Mon, Aug 27, 10:23 AM
vsk added inline comments to D51239: [ubsan] Enable -fsanitize=vptr on Apple devices and simulators.
Mon, Aug 27, 10:01 AM

Fri, Aug 24

vsk created D51239: [ubsan] Enable -fsanitize=vptr on Apple devices and simulators.
Fri, Aug 24, 3:56 PM
vsk updated the diff for D50751: Allow use of self.filecheck in LLDB tests (c.f self.expect).
Fri, Aug 24, 3:43 PM · Unknown Object (Project)
vsk updated the diff for D49887: [DebugInfo] Add support for DWARF5 call site-related attributes.
  • Per Adrian and Paul's feedback, remove the experimental "gating" cl::opt because the size overhead of call site info is low.
  • Rebase.
Fri, Aug 24, 11:25 AM · debug-info
vsk added inline comments to D50478: Add support for artificial tail call frames.
Fri, Aug 24, 10:39 AM
vsk updated the diff for D50478: Add support for artificial tail call frames.

Thanks for the feedback!

Fri, Aug 24, 10:39 AM
vsk added a comment to D50887: [DWARF] Missing location debug information with -O2..

@CarlosAlbertoEnciso thanks for investigating this and testing out a new way to collect debug values. Given the complexity of replacing the current algorithm, and that it's not strictly related to fixing MachineCSE, I'd recommend keeping the current algorithm (i.e not using MachineRegisterInfo for now). That will allow us to land a fix in tree soon and hopefully iterate on it.

Fri, Aug 24, 10:17 AM

Thu, Aug 23

vsk added a comment to D50478: Add support for artificial tail call frames.

Ping.

Thu, Aug 23, 4:21 PM
vsk accepted D51185: Reuse the SelectorTable from Clang's Preprocessor.

LGTM. This is NFC, it seems. There's a FIXME in Preprocessor.h about the lifetime of SelectorTable eventually not being tied to Preprocessor, but this is correct for now.

Thu, Aug 23, 3:24 PM · Restricted Project

Wed, Aug 22

vsk added a comment to D46100: WIP: [IR/Verifier] Diagnose use-before-def scenarios for debug intrinsics.

A problem with CodeGenPrepare::placeDbgValues is that it can move debug intrinsics more or less blindly, for example without taking other debug intrinsics refering to the same variable into account.
If there is a use-before-def scenario it often is correct to move the using dbg.value below the def, but that depends on which transform that resulted in such a situation (I guess that often it might be due to an opt-pass not caring about debug intrinsics).
Right now CodeGenPrepare::placeDbgValues also hoists dbg.value instructions when the use already is after the def. If I remember correctly it can even move a dbg.value from one basic block to another.

Wed, Aug 22, 5:50 PM
vsk accepted D50481: Fix broken builtin functions in the expression command.

Thanks, LGTM!

Wed, Aug 22, 1:39 PM · Restricted Project
vsk added inline comments to D50901: [clang][ubsan] Split Implicit Integer Truncation Sanitizer into unsigned and signed checks.
Wed, Aug 22, 11:49 AM · Restricted Project, Restricted Project
vsk added a comment to D46100: WIP: [IR/Verifier] Diagnose use-before-def scenarios for debug intrinsics.

Recently I addressed a dbg.value use-before-def bug in CodeGenPrepare (r340370). What I learned forced me to reconsider adding this verifier check.

Wed, Aug 22, 10:40 AM

Aug 21 2018

vsk added a comment to D50887: [DWARF] Missing location debug information with -O2..

This is looking really good. I think it's great that the test exercises more than just MachineCSE in isolation. Just two more minor comments (inline) --

Aug 21 2018, 10:26 AM

Aug 20 2018

vsk accepted D50997: Automatically set path to sanitizer runtime when running tests on macOS..

(LGTM with the second comment addressed.)

Aug 20 2018, 2:45 PM
vsk added a comment to D50997: Automatically set path to sanitizer runtime when running tests on macOS..

Thanks so much for doing this!

Aug 20 2018, 2:44 PM
vsk updated the diff for D50927: [Sema] Remove location from implicit capture init expr.
Aug 20 2018, 12:28 PM
vsk updated subscribers of D42043: c-index: CXString: fix MSAN read-past-end bug.

+ Jan and Volodymyr. This seemed to be in good shape the last time I looked at it. Not having touched libclang for a while I don't think I can give an official lgtm.

Aug 20 2018, 10:16 AM

Aug 17 2018

vsk added inline comments to D50864: Add libc++ data formatter for std::function.
Aug 17 2018, 5:13 PM
vsk created D50927: [Sema] Remove location from implicit capture init expr.
Aug 17 2018, 2:31 PM
vsk added inline comments to D50638: [AsmPrinter] Look inside bundles and rely on FrameDestroy in calculateDbgValueHistory.
Aug 17 2018, 11:40 AM
vsk updated subscribers of D50887: [DWARF] Missing location debug information with -O2..
Aug 17 2018, 10:42 AM
vsk added a comment to D50887: [DWARF] Missing location debug information with -O2..

Thanks for doing this!

Aug 17 2018, 10:41 AM

Aug 16 2018

vsk added inline comments to D50638: [AsmPrinter] Look inside bundles and rely on FrameDestroy in calculateDbgValueHistory.
Aug 16 2018, 3:23 PM
vsk added a comment to D50637: [CodeGen] Set FrameSetup/FrameDestroy on BUNDLE instructions.

Not having worked in this area I can't give an official lgtm, but this seems reasonable to me.

Aug 16 2018, 3:10 PM
vsk added inline comments to D50867: [InstrProf] Use atomic profile counter updates for TSan.
Aug 16 2018, 2:53 PM
vsk updated the diff for D50867: [InstrProf] Use atomic profile counter updates for TSan.
Aug 16 2018, 2:53 PM
vsk created D50867: [InstrProf] Use atomic profile counter updates for TSan.
Aug 16 2018, 2:40 PM

Aug 15 2018

vsk added inline comments to rL339716: Remove asseration from ConstString::GetConstCStringAndSetMangledCounterPart()….
Aug 15 2018, 11:59 AM

Aug 14 2018

vsk created D50751: Allow use of self.filecheck in LLDB tests (c.f self.expect).
Aug 14 2018, 5:28 PM · Unknown Object (Project)
vsk added a dependent revision for D49887: [DebugInfo] Add support for DWARF5 call site-related attributes: D50478: Add support for artificial tail call frames.
Aug 14 2018, 12:25 PM · debug-info
vsk updated the diff for D50478: Add support for artificial tail call frames.
Aug 14 2018, 12:25 PM

Aug 13 2018

vsk accepted D50620: Added test for Core/Range class..

Thanks, looks good with nitpicks.

Aug 13 2018, 11:27 AM
vsk added inline comments to D49887: [DebugInfo] Add support for DWARF5 call site-related attributes.
Aug 13 2018, 10:08 AM · debug-info

Aug 10 2018

vsk updated the diff for D50478: Add support for artificial tail call frames.

Rebase, and update the patch to use DW_AT_call_return_pc information.

Aug 10 2018, 7:00 PM
vsk accepted D50225: Use a DenseMap for looking up functions by UID in CompileUnit::FindFunctionByUID.

Thanks, LGTM!

Aug 10 2018, 5:08 PM
vsk added inline comments to D49887: [DebugInfo] Add support for DWARF5 call site-related attributes.
Aug 10 2018, 4:00 PM · debug-info
vsk updated the diff for D49887: [DebugInfo] Add support for DWARF5 call site-related attributes.

Thanks for your feedback @dblaikie! I've addressed some of it and responded to the rest inline.

Aug 10 2018, 4:00 PM · debug-info
vsk updated subscribers of D48994: Update DBG_VALUE register operand during LiveInterval operations.

@thegameg and @kparzysz, would either of you have some time to take a look?

Aug 10 2018, 9:59 AM · debug-info

Aug 9 2018

vsk planned changes to D49887: [DebugInfo] Add support for DWARF5 call site-related attributes.

Dsymutil will need to learn to resolve addresses in DW_AT_call_return_pc. Currently, I think it's skipping over them, which means that lldb can't use them.

Aug 9 2018, 6:54 PM · debug-info
vsk updated the diff for D49887: [DebugInfo] Add support for DWARF5 call site-related attributes.
  • Add verifier & bitcode tests, rebase.
  • Add complete DWARF generation support for call site entries (i.e, one call site entry per call, DW_AT_call_tail_call emission, and DW_AT_call_return_pc emission).
Aug 9 2018, 2:46 PM · debug-info

Aug 8 2018

vsk updated the summary of D50478: Add support for artificial tail call frames.
Aug 8 2018, 2:15 PM
vsk created D50478: Add support for artificial tail call frames.
Aug 8 2018, 2:14 PM

Aug 7 2018

vsk accepted D50340: [Local] Add dbg location on unreachable inst in changeToUnreachable.

LGTM, thanks!

Aug 7 2018, 12:31 PM
vsk added inline comments to D50340: [Local] Add dbg location on unreachable inst in changeToUnreachable.
Aug 7 2018, 12:01 PM

Aug 6 2018

vsk updated subscribers of D50271: [IRMemoryMap] Shrink Allocation make it move-only (NFC).
Aug 6 2018, 12:26 PM
vsk updated the diff for D50271: [IRMemoryMap] Shrink Allocation make it move-only (NFC).
  • Make Allocation move-only. This should prevent a std::vector copy in ::Malloc.
Aug 6 2018, 11:15 AM