Page MenuHomePhabricator
Feed Advanced Search

Thu, Jul 22

madhur13490 added a comment to D106452: [AMDGPU] Fix running ResourceUsageAnalysis.

If you want to you can emit some debugging message after clearing the map and have a llc lit test with the debug flag to CHECK the debugging message.

Thu, Jul 22, 7:01 AM · Restricted Project

Mon, Jul 19

madhur13490 added inline comments to D105839: [AMDGPU] Improve register computation for indirect calls.
Mon, Jul 19, 6:22 AM · Restricted Project

Fri, Jul 16

madhur13490 committed rG1a7cddb0ea63: [NFC] Fix typo intrinisic (authored by madhur13490).
[NFC] Fix typo intrinisic
Fri, Jul 16, 9:15 AM
madhur13490 closed D106161: [NFC] Fix typo intrinisic.
Fri, Jul 16, 9:15 AM · Restricted Project
madhur13490 requested review of D106161: [NFC] Fix typo intrinisic.
Fri, Jul 16, 9:14 AM · Restricted Project

Sun, Jul 11

madhur13490 added inline comments to D105780: [Attributes] Determine attribute properties from TableGen data.
Sun, Jul 11, 9:27 AM · Restricted Project

Jun 30 2021

madhur13490 committed rGa7ed55f64c5f: [AMDGPU] Simplify getReservedNumSGPRs (authored by madhur13490).
[AMDGPU] Simplify getReservedNumSGPRs
Jun 30 2021, 3:50 AM
madhur13490 closed D104513: [AMDGPU] Simplify getReservedNumSGPRs.
Jun 30 2021, 3:49 AM · Restricted Project

Jun 24 2021

madhur13490 added a comment to D103694: [AMDGPU] Simplify handleAddressTakenFunctions. NFC..

Please remove NFC due to the new context. @JonChesterfield found an issue with recursions in the base patch. The following case depicts the scenario

Jun 24 2021, 12:21 AM · Restricted Project
madhur13490 added a reviewer for D103694: [AMDGPU] Simplify handleAddressTakenFunctions. NFC.: JonChesterfield.
Jun 24 2021, 12:19 AM · Restricted Project
madhur13490 added a comment to D103694: [AMDGPU] Simplify handleAddressTakenFunctions. NFC..

Commit 660cae84c3144a42272daa16415fc9a2532773c4 reverts the base patch due to an unsupported scenario but thankfully this patch fixes that issue. Please rebase and keep the support intact.

Jun 24 2021, 12:15 AM · Restricted Project
madhur13490 added inline comments to D103694: [AMDGPU] Simplify handleAddressTakenFunctions. NFC..
Jun 24 2021, 12:12 AM · Restricted Project

Jun 23 2021

madhur13490 added inline comments to D104513: [AMDGPU] Simplify getReservedNumSGPRs.
Jun 23 2021, 3:26 AM · Restricted Project
madhur13490 updated the diff for D104513: [AMDGPU] Simplify getReservedNumSGPRs.

optimize conditions

Jun 23 2021, 3:25 AM · Restricted Project

Jun 21 2021

madhur13490 added a comment to D104513: [AMDGPU] Simplify getReservedNumSGPRs.

Ping to @arsenm

Jun 21 2021, 11:36 PM · Restricted Project
madhur13490 updated the diff for D104513: [AMDGPU] Simplify getReservedNumSGPRs.

address review comments

Jun 21 2021, 11:35 PM · Restricted Project

Jun 18 2021

madhur13490 requested review of D104513: [AMDGPU] Simplify getReservedNumSGPRs.
Jun 18 2021, 2:45 AM · Restricted Project

Jun 11 2021

madhur13490 added inline comments to D103636: [AMDGPU][IndirectCalls] Fix register usage propagation for indirect/external calls.
Jun 11 2021, 11:30 PM · Restricted Project
madhur13490 committed rGc27e8141b3d1: [AMDGPU][IndirectCalls] Fix register usage propagation for indirect/external… (authored by madhur13490).
[AMDGPU][IndirectCalls] Fix register usage propagation for indirect/external…
Jun 11 2021, 11:30 PM
madhur13490 closed D103636: [AMDGPU][IndirectCalls] Fix register usage propagation for indirect/external calls.
Jun 11 2021, 11:29 PM · Restricted Project

Jun 9 2021

madhur13490 committed rG62bd7da889c1: [LangRef] Add link to opaque pointers (authored by madhur13490).
[LangRef] Add link to opaque pointers
Jun 9 2021, 11:41 AM
madhur13490 closed D103981: [LangRef] Add link to opaque pointers.
Jun 9 2021, 11:41 AM · Restricted Project
madhur13490 requested review of D103981: [LangRef] Add link to opaque pointers.
Jun 9 2021, 10:43 AM · Restricted Project
madhur13490 added inline comments to D103636: [AMDGPU][IndirectCalls] Fix register usage propagation for indirect/external calls.
Jun 9 2021, 2:40 AM · Restricted Project
madhur13490 updated the diff for D103636: [AMDGPU][IndirectCalls] Fix register usage propagation for indirect/external calls.

address review comments + rebase

Jun 9 2021, 2:39 AM · Restricted Project

Jun 8 2021

madhur13490 added a comment to D103636: [AMDGPU][IndirectCalls] Fix register usage propagation for indirect/external calls.

Ping to @arsenm and @rampitec!

Jun 8 2021, 8:53 AM · Restricted Project

Jun 4 2021

madhur13490 added inline comments to D103694: [AMDGPU] Simplify handleAddressTakenFunctions. NFC..
Jun 4 2021, 6:03 AM · Restricted Project
madhur13490 added inline comments to D103694: [AMDGPU] Simplify handleAddressTakenFunctions. NFC..
Jun 4 2021, 6:00 AM · Restricted Project

Jun 3 2021

madhur13490 committed rG6a3beb1f68d6: [AMDGPU] [IndirectCalls] Don't propagate attributes to address taken functions… (authored by madhur13490).
[AMDGPU] [IndirectCalls] Don't propagate attributes to address taken functions…
Jun 3 2021, 11:07 PM
madhur13490 closed D103138: [AMDGPU] [IndirectCalls] Don't propagate attributes to address taken functions and their callees.
Jun 3 2021, 11:07 PM · Restricted Project
madhur13490 added inline comments to D103636: [AMDGPU][IndirectCalls] Fix register usage propagation for indirect/external calls.
Jun 3 2021, 10:56 AM · Restricted Project
madhur13490 added reviewers for D103636: [AMDGPU][IndirectCalls] Fix register usage propagation for indirect/external calls: rampitec, arsenm, scchan.
Jun 3 2021, 10:54 AM · Restricted Project
madhur13490 requested review of D103636: [AMDGPU][IndirectCalls] Fix register usage propagation for indirect/external calls.
Jun 3 2021, 10:52 AM · Restricted Project

Jun 1 2021

madhur13490 committed rG3c874ce4271c: [AMDGPU][NFC] Remove author's name from codebase (authored by madhur13490).
[AMDGPU][NFC] Remove author's name from codebase
Jun 1 2021, 12:22 PM
madhur13490 closed D103484: [AMDGPU][NFC] Remove author's name from codebase.
Jun 1 2021, 12:22 PM · Restricted Project
madhur13490 retitled D103484: [AMDGPU][NFC] Remove author's name from codebase from [AMDGPU][NFC] Remove author's name from codebase This must have made to code by accident. to [AMDGPU][NFC] Remove author's name from codebase.
Jun 1 2021, 12:20 PM · Restricted Project
madhur13490 requested review of D103484: [AMDGPU][NFC] Remove author's name from codebase.
Jun 1 2021, 12:19 PM · Restricted Project
madhur13490 added inline comments to D103138: [AMDGPU] [IndirectCalls] Don't propagate attributes to address taken functions and their callees.
Jun 1 2021, 11:56 AM · Restricted Project
madhur13490 added a reviewer for D102740: Convert TableGen assert to error: Paul-C-Anagnostopoulos.
Jun 1 2021, 11:51 AM · Restricted Project
madhur13490 updated the diff for D103138: [AMDGPU] [IndirectCalls] Don't propagate attributes to address taken functions and their callees.

add comment about keeping indirect calls handling separate

Jun 1 2021, 11:41 AM · Restricted Project

May 28 2021

madhur13490 added a comment to D103138: [AMDGPU] [IndirectCalls] Don't propagate attributes to address taken functions and their callees.

Fixed one more test after rebase.

May 28 2021, 12:28 PM · Restricted Project
madhur13490 updated the diff for D103138: [AMDGPU] [IndirectCalls] Don't propagate attributes to address taken functions and their callees.

use assert

May 28 2021, 12:27 PM · Restricted Project
madhur13490 added a comment to D103138: [AMDGPU] [IndirectCalls] Don't propagate attributes to address taken functions and their callees.

You cannot recover after an "unreachable" or a failed assertion.

C [...] returns gracefully returns in release builds.

No it doesn't. In a release build llvm_unreachable will usually expand to __builtin_unreachable() or __assume(false) and the optimising compiler will remove any code after that point because it is allowed to assume that it is not reachable.

May 28 2021, 8:29 AM · Restricted Project
madhur13490 added inline comments to D103138: [AMDGPU] [IndirectCalls] Don't propagate attributes to address taken functions and their callees.
May 28 2021, 7:40 AM · Restricted Project
madhur13490 added inline comments to D103138: [AMDGPU] [IndirectCalls] Don't propagate attributes to address taken functions and their callees.
May 28 2021, 7:38 AM · Restricted Project

May 27 2021

madhur13490 updated the diff for D103138: [AMDGPU] [IndirectCalls] Don't propagate attributes to address taken functions and their callees.

use NSDMI for ModuleCG

May 27 2021, 3:10 AM · Restricted Project
madhur13490 added inline comments to D103138: [AMDGPU] [IndirectCalls] Don't propagate attributes to address taken functions and their callees.
May 27 2021, 2:28 AM · Restricted Project

May 26 2021

madhur13490 added inline comments to D103138: [AMDGPU] [IndirectCalls] Don't propagate attributes to address taken functions and their callees.
May 26 2021, 11:17 PM · Restricted Project
madhur13490 updated the diff for D103138: [AMDGPU] [IndirectCalls] Don't propagate attributes to address taken functions and their callees.

address review comments

May 26 2021, 11:13 PM · Restricted Project

May 25 2021

madhur13490 requested review of D103138: [AMDGPU] [IndirectCalls] Don't propagate attributes to address taken functions and their callees.
May 25 2021, 10:50 PM · Restricted Project

May 17 2021

madhur13490 accepted D101847: [AMDGPU] Fix function pointer argument bug in AMDGPU Propagate Attributes pass..

LGTM

May 17 2021, 8:33 AM · Restricted Project, Restricted Project

May 13 2021

madhur13490 added inline comments to D101847: [AMDGPU] Fix function pointer argument bug in AMDGPU Propagate Attributes pass..
May 13 2021, 11:20 AM · Restricted Project, Restricted Project
madhur13490 added a comment to D101847: [AMDGPU] Fix function pointer argument bug in AMDGPU Propagate Attributes pass..

I think at this point I'd prefer to just not worry about the cast stripping case and go back to just handling the direct callee.

This pass should not be required for correctness, and is only because opencl is still not linking the device libraries correctly such that the linked in functions have the correct subtarget/subtarget features. Long term I would like to move this pass into a form that does conservative merges of attributes, which can safely give up on anything indirect looking

May 13 2021, 11:17 AM · Restricted Project, Restricted Project

May 11 2021

madhur13490 added a comment to D102177: [AMDGPU][RFC] Improve sgpr function arguments.

We have another proposal we were working on to rearrange these a bit differently. We need to account for a few more inputs in the layout

As long as this remains in GFX land, we should be fine with it because our new proposal is for compute only (as of now).

I would like to keep the same calling convention in compute and graphics. At least regarding the stack pointer and others, because I don’t see a compelling reason to diverge even more. Actually, I’d like it if they were more common than they are now, because we implement some things twice at the moment.
The compute proposal should work just fine; if we move the stack and frame pointer, we end up with the same benefits as in this patch. I commented on the internal proposal for this (I hope I found the right one?).

May 11 2021, 6:03 AM · Restricted Project
madhur13490 added a comment to D102177: [AMDGPU][RFC] Improve sgpr function arguments.

We have another proposal we were working on to rearrange these a bit differently. We need to account for a few more inputs in the layout

May 11 2021, 2:56 AM · Restricted Project

May 5 2021

madhur13490 added inline comments to D101847: [AMDGPU] Fix function pointer argument bug in AMDGPU Propagate Attributes pass..
May 5 2021, 2:28 AM · Restricted Project, Restricted Project

Apr 30 2021

madhur13490 accepted D101534: [AMDGPU] Do not annotate features for graphics.
Apr 30 2021, 5:35 PM · Restricted Project

Apr 29 2021

madhur13490 added inline comments to D101534: [AMDGPU] Do not annotate features for graphics.
Apr 29 2021, 9:25 AM · Restricted Project

Apr 22 2021

madhur13490 added inline comments to D98608: [Attributor] Make sure Heap2Stack works properly on a GPU target.
Apr 22 2021, 10:36 AM · Restricted Project
madhur13490 added a comment to D100904: [AMDGPU] Lower regbanks reassign threshold to 15000.

I feel we should probably discuss more about the optimal solution than the current one. I have a gut feeling that the problem pass is trying to solve is a flavor of graph coloring problem where we have to color the interference graph with K number of colors where K is the number of register banks. We may need to redo the interference graph in this pass and then attempt the coloring.

Apr 22 2021, 2:12 AM · Restricted Project

Apr 21 2021

madhur13490 added inline comments to D100877: AMDGPU: Fix indirect tail calls.
Apr 21 2021, 7:42 AM · Restricted Project

Apr 20 2021

madhur13490 added inline comments to D100633: [AMDGPU] Remove error check for indirect calls and add missing queue-ptr.
Apr 20 2021, 12:13 AM · Restricted Project

Apr 19 2021

madhur13490 committed rG6a4d9cb7e04d: [AMDGPU] Remove error check for indirect calls and add missing queue-ptr (authored by madhur13490).
[AMDGPU] Remove error check for indirect calls and add missing queue-ptr
Apr 19 2021, 12:05 PM
madhur13490 closed D100633: [AMDGPU] Remove error check for indirect calls and add missing queue-ptr.
Apr 19 2021, 12:05 PM · Restricted Project
madhur13490 added inline comments to D100633: [AMDGPU] Remove error check for indirect calls and add missing queue-ptr.
Apr 19 2021, 12:04 PM · Restricted Project
madhur13490 updated the diff for D100633: [AMDGPU] Remove error check for indirect calls and add missing queue-ptr.

remove new test and merge in existing one

Apr 19 2021, 11:56 AM · Restricted Project

Apr 16 2021

madhur13490 updated the diff for D100633: [AMDGPU] Remove error check for indirect calls and add missing queue-ptr.

keep tests alone as discussed offline and add explicit subtarget to test

Apr 16 2021, 11:03 AM · Restricted Project
madhur13490 added inline comments to D100633: [AMDGPU] Remove error check for indirect calls and add missing queue-ptr.
Apr 16 2021, 3:28 AM · Restricted Project
madhur13490 added a reviewer for D100633: [AMDGPU] Remove error check for indirect calls and add missing queue-ptr: rampitec.
Apr 16 2021, 3:27 AM · Restricted Project
madhur13490 requested review of D100633: [AMDGPU] Remove error check for indirect calls and add missing queue-ptr.
Apr 16 2021, 3:26 AM · Restricted Project

Apr 15 2021

madhur13490 added a comment to D96179: Ignore assume like calls by default in hasAddressTaken().

I hope this doesn't run into the issues I observed in D98884. Doing bootstrap builds locally would strengthen testing.

That's the point. It is effectively default on for a long time already.

Apr 15 2021, 10:28 PM · Restricted Project
madhur13490 added a comment to D96179: Ignore assume like calls by default in hasAddressTaken().

I hope this doesn't run into the issues I observed in D98884. Doing bootstrap builds locally would strengthen testing.

Apr 15 2021, 10:06 PM · Restricted Project

Apr 13 2021

madhur13490 added inline comments to D100339: [Attributor] Run lightweight version of the Attributor by default..
Apr 13 2021, 6:25 AM · Restricted Project
madhur13490 committed rG5682ae2fc6fe: [AMDGPU] Set implicit arg attributes for indirect calls (authored by madhur13490).
[AMDGPU] Set implicit arg attributes for indirect calls
Apr 13 2021, 6:15 AM
madhur13490 closed D99347: [AMDGPU] Set implicit arg attributes for indirect calls.
Apr 13 2021, 6:15 AM · Restricted Project
madhur13490 added inline comments to D100339: [Attributor] Run lightweight version of the Attributor by default..
Apr 13 2021, 5:36 AM · Restricted Project

Apr 12 2021

madhur13490 updated the diff for D99347: [AMDGPU] Set implicit arg attributes for indirect calls.

change tests at HEAD

Apr 12 2021, 12:01 PM · Restricted Project
madhur13490 added a comment to D98884: [IR] Ignore bitcasts of function pointers which are only used as callees in callbase instruction.

Are we sure this is not a single bad cast that evolves into something different if assertions are disabled. Is the cast problem located, that should be easy after all given the minimal reproducer in PR49861?

Apr 12 2021, 1:07 AM · Restricted Project

Apr 10 2021

madhur13490 added a comment to D98884: [IR] Ignore bitcasts of function pointers which are only used as callees in callbase instruction.

This change has wider impact than was expected as it affects several components and backends. Bootstrap builds fail with this change and as the buildbots show, this change generates a buggy Tablegen binary in release mode.

Debugging:

  1. Disabling the newly added code in hasAddressTaken(), we get sane Tablegen and builds pass. To note, debug builds pass with the change. This clearly shows that codegen is functionally different in release and debug mode for Tablegen.
  2. Doing “ninja check-all” on stage1 build does not help although we have ~90000 tests.
  3. Running valgrind also did not reveal any new things.
  4. Doing ASAN build of the compiler led to several errors and drawing any conclusion from it is close to impossible for this patch.

However, if I disable IPSCCP pass on top of this patch then everything turns green. I tried to fix trivial issues in the pass but it seems the pass needs deeper and major fixes to make this patch land.

Till we fix issues in IPSCCP, this patch stands reverted. The whole intention of this patch was to handle a subset of cases for AMDGPU backend, specifically, needed for https://reviews.llvm.org/D99347.

This doesn't necessarily mean there's a problem with IPSCCP. What is the failing IR?

Apr 10 2021, 9:27 AM · Restricted Project

Apr 8 2021

madhur13490 added a comment to D98884: [IR] Ignore bitcasts of function pointers which are only used as callees in callbase instruction.

This change has wider impact than was expected as it affects several components and backends. Bootstrap builds fail with this change and as the buildbots show, this change generates a buggy Tablegen binary in release mode.

Apr 8 2021, 10:17 PM · Restricted Project
madhur13490 added a comment to D98884: [IR] Ignore bitcasts of function pointers which are only used as callees in callbase instruction.

We've also seen failures on all of our 2 stage bots e.g. http://lab.llvm.org:8011/#/builders/7/builds/2348. Which are fixed now this is reverted.

[494/5842] Building Attributes.inc...
FAILED: include/llvm/IR/Attributes.inc 
cd /home/tcwg-buildslave/worker/clang-cmake-aarch64-full/stage2 && /home/tcwg-buildslave/worker/clang-cmake-aarch64-full/stage2/bin/llvm-tblgen -gen-attrs -I /home/tcwg-buildslave/worker/clang-cmake-aarch64-full/llvm/llvm/include/llvm/IR -I/home/tcwg-buildslave/worker/clang-cmake-aarch64-full/stage2/include -I/home/tcwg-buildslave/worker/clang-cmake-aarch64-full/llvm/llvm/include /home/tcwg-buildslave/worker/clang-cmake-aarch64-full/llvm/llvm/include/llvm/IR/Attributes.td --write-if-changed -o include/llvm/IR/Attributes.inc -d include/llvm/IR/Attributes.inc.d
pure virtual method called
terminate called without an active exception

Bisecting this I have seen clang-tblgen crash with a few different asserts. Bad alloc, one in erase() in StringMap, illegal instruction. Our bot doesn't have sanitizers on so it could be the same finicky memory corruption.

This PPC one does though https://lab.llvm.org/buildbot/#/builders/18/builds/1140 and saw similar failures. No immediate explanation but some configs worth checking locally before the reland.

Apr 8 2021, 2:09 AM · Restricted Project

Apr 6 2021

madhur13490 added a comment to D98884: [IR] Ignore bitcasts of function pointers which are only used as callees in callbase instruction.

These failures were not caught by LLVM bots which run on a patch. Can I do a trial run of the bots and run all tests that would run when the patch lands?

Apr 6 2021, 11:54 AM · Restricted Project
madhur13490 committed rG167ea67d76c2: [IR] Ignore bitcasts of function pointers which are only used as callees in… (authored by madhur13490).
[IR] Ignore bitcasts of function pointers which are only used as callees in…
Apr 6 2021, 2:24 AM
madhur13490 closed D98884: [IR] Ignore bitcasts of function pointers which are only used as callees in callbase instruction.
Apr 6 2021, 2:24 AM · Restricted Project
madhur13490 updated the diff for D98884: [IR] Ignore bitcasts of function pointers which are only used as callees in callbase instruction.

make positive checks

Apr 6 2021, 2:16 AM · Restricted Project

Apr 3 2021

madhur13490 added a comment to D98884: [IR] Ignore bitcasts of function pointers which are only used as callees in callbase instruction.

Ping!

Apr 3 2021, 8:14 PM · Restricted Project

Mar 30 2021

madhur13490 added inline comments to D98884: [IR] Ignore bitcasts of function pointers which are only used as callees in callbase instruction.
Mar 30 2021, 7:25 AM · Restricted Project
madhur13490 updated the diff for D98884: [IR] Ignore bitcasts of function pointers which are only used as callees in callbase instruction.

Addressed Stas' comments

Mar 30 2021, 7:24 AM · Restricted Project
madhur13490 added inline comments to D98884: [IR] Ignore bitcasts of function pointers which are only used as callees in callbase instruction.
Mar 30 2021, 3:30 AM · Restricted Project

Mar 27 2021

madhur13490 added inline comments to D98884: [IR] Ignore bitcasts of function pointers which are only used as callees in callbase instruction.
Mar 27 2021, 6:32 AM · Restricted Project
madhur13490 updated the diff for D98884: [IR] Ignore bitcasts of function pointers which are only used as callees in callbase instruction.

add new test

Mar 27 2021, 6:28 AM · Restricted Project

Mar 26 2021

madhur13490 added a comment to D98884: [IR] Ignore bitcasts of function pointers which are only used as callees in callbase instruction.

This should work I guess. It needs tests for non-callee use and non call base uses. I would also prefer to GlobalOpt change to be extracted into a separate patch.

non call base users - What all such cases possible?We can use invoke but that is also CallBase.

That is more than possible. Take a pointer and store it, compare to another pointer, bitcast and store or compare.

Mar 26 2021, 11:02 AM · Restricted Project
madhur13490 added a comment to D98884: [IR] Ignore bitcasts of function pointers which are only used as callees in callbase instruction.

This should work I guess. It needs tests for non-callee use and non call base uses. I would also prefer to GlobalOpt change to be extracted into a separate patch.

Mar 26 2021, 10:50 AM · Restricted Project
madhur13490 added inline comments to D98884: [IR] Ignore bitcasts of function pointers which are only used as callees in callbase instruction.
Mar 26 2021, 10:26 AM · Restricted Project
madhur13490 updated the diff for D98884: [IR] Ignore bitcasts of function pointers which are only used as callees in callbase instruction.

specifically check for callee

Mar 26 2021, 10:22 AM · Restricted Project

Mar 25 2021

madhur13490 added inline comments to D99347: [AMDGPU] Set implicit arg attributes for indirect calls.
Mar 25 2021, 11:10 PM · Restricted Project
madhur13490 added inline comments to D98884: [IR] Ignore bitcasts of function pointers which are only used as callees in callbase instruction.
Mar 25 2021, 9:41 PM · Restricted Project
madhur13490 updated the diff for D98884: [IR] Ignore bitcasts of function pointers which are only used as callees in callbase instruction.

go back to dyn_cast to handle all users

Mar 25 2021, 9:40 PM · Restricted Project
madhur13490 added a comment to D98884: [IR] Ignore bitcasts of function pointers which are only used as callees in callbase instruction.

The bot shows failure of bad-signature.c at https://reviews.llvm.org/B95560. When ran locally it does report the correct values and thus pass. I am not sure how to fix the test failure.
Failure in tile-and-distribute.mlir seems unrelated to this change.

@jdoerfert/ @arsenm / @rampitec , please advise.

The bots are imperfect, I would say it's unrelated

Mar 25 2021, 12:29 PM · Restricted Project
madhur13490 added a comment to D98884: [IR] Ignore bitcasts of function pointers which are only used as callees in callbase instruction.

The bot shows failure of bad-signature.c at https://reviews.llvm.org/B95560. When ran locally it does report the correct values and thus pass. I am not sure how to fix the test failure.
Failure in tile-and-distribute.mlir seems unrelated to this change.

Mar 25 2021, 12:11 PM · Restricted Project