Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

[AutoBump] Merge with 5901d400 (May 28) (54) #314

Merged
merged 23 commits into from
Sep 5, 2024
Merged

[AutoBump] Merge with 5901d400 (May 28) (54) #314

merged 23 commits into from
Sep 5, 2024

Conversation

mgehre-amd
Copy link
Collaborator

No description provided.

kazutakahirata and others added 23 commits May 28, 2024 09:03
@kazutakahirata
[clang] Fix a warning
fea7399 
This patch fixes:

  clang/unittests/Interpreter/IncrementalProcessingTest.cpp:39:13:
  error: unused function 'HostSupportsJit' [-Werror,-Wunused-function]
@AaronBallman
clang:: to llvm::; NFC
273777e 
These interfaces are LLVM interfaces, not Clang ones; but this worked
because of LLVM.h adding the interfaces to the clang namespace.
@a-tarasyuk
[Clang] Fix an assertion failure when checking invalid this (llvm#9…
259caad 
…3490)

Skip explicit this check in non-valid scopes due to `null` type in
lambdas with invalid captures or incomplete parameter lists during
parsing


Fixes llvm#91536
@fhahn
[LAA] Limit no-overlap check to at least one loop-invariant accesses.
234cc40 
Limit the logic added in llvm#9230
to cases where either sink or source are loop-invariant, to avoid
compile-time increases. This is not needed for correctness.

I am working on follow-up changes to reduce the compile-time impact in
general to allow us to enable this again for any source/sink.

This should fix the compile-time regression introduced by this change:

* compile-time improvement with this change:
  https://llvm-compile-time-tracker.com/compare.php?from=4351787fb650da6d1bfb8d6e58753c90dcd4c418&to=b89010a2eb5f98494787c1c3b77f25208c59090c&stat=instructions:u

* compile-time improvement with original patch reverted on top of this
  change:
  https://llvm-compile-time-tracker.com/compare.php?from=b89010a2eb5f98494787c1c3b77f25208c59090c&to=19a1103fe68115cfd7d6472c6961f4fabe81a593&stat=instructions:u
@AZero13
Revert "[Legalizer] Check full condition for UMIN and UMAX just like …
d582958 
…the code below does for SMIN and SMAX" (llvm#93573)

Reverts llvm#87932
@mbucko
Add SBAddressRange and SBAddressRangeList to SB API (llvm#92014)
42944e4 
This adds new SB API calls and classes to allow a user of the SB API to obtain an address ranges from SBFunction and SBBlock.
@llvmgnsyncbot
[gn build] Port 42944e4
79c7342
@artagnon
LoopIdiomRecognize: strip bad TODO (NFC) (llvm#92890)
7bea41e 
There are several reasons why handling powi in LoopIdiomRecognize is a
bad idea:

- powi corresponds to a GCC builtin that is only defined for C int
(which is i32 for most targets).
- powi isn't always lowered by targets correctly for non-i32 parameters.
Several targets fail to compile llvm.powi.f32.i16, for example.
- Unlike memcpy and memset, which tend to be important enough internal
intrinsics that you have to handle them correctly even in freestanding
modes, powi isn't.

Strip this bad TODO to avoid misleading contributors.
@arsenm
DAG: Use flags in isLegalToCombineMinNumMaxNum (llvm#93555)
16a5fd3
@alx32
[lld-macho][ObjC] Implement category merging into base class (llvm#92448
b963931 
)

Currently category merging only supports merging multiple categories
into one. With this commit we add the ability to fully merge categories
into the base class, if the base class is included in the current
module. This is the optimal approach for defined classes.
@JDevlieghere
[lldb] Remove lldbassert in AppleObjCTypeEncodingParser (llvm#93332)
d1d863c 
AppleObjCTypeEncodingParser::BuildObjCObjectPointerType currently
contains an lldbassert to detect situations where we have a forward
declaration without a definition. According to the accompanying comment,
its purpose is to catch "weird cases" during test suite runs.

However, because this is an lldbassert, we show a scary message to our
users who think this is a problem and report the issue to us.
Unfortunately those reports aren't very actionable without a way to know
the name of the type.

This patch changes the lldbassert to a regular assert and emits a log
message to the types log when this happens.

rdar://127439898
@JDevlieghere
[lldb] Add missing semicolon (NFC)
f69b6d2
@FLZ101 @jprotze
[OMPT] Set default values for tsan function pointers (llvm#93568)
c09787b 
Avoid calling NULL function pointers in cases where ompt_start_tool
succeeds but those tsan functions
do not really exist.

Fix llvm#93524

---------

Co-authored-by: Joachim <protze@rz.rwth-aachen.de>
@fhahn
[SCEV] Compute symbolic max backedge taken count in BTI directly. (NFC)
ef67f31 
Move symbolic max backedge taken count computation to BackedgeTakenInfo,
use existing ExitNotTaken info.

In preparation for llvm#93498.
@Sirraide
[Clang] [NFC] Remove debug printing
0b2094c
@JOE1994 @tbaederr @cpplearner
[clang][Sema] Don't emit 'declared here' note for builtin functions w…
f089996 
…ith no decl in source (llvm#93394)

Fixes llvm#93369

---------

Co-authored-by: Timm Baeder <tbaeder@redhat.com>
Co-authored-by: S. B. Tam <cpplearner@outlook.com>
@akshayrdeodhar @nikic
[Reassociate] Preserve NSW flags after expr tree rewriting (llvm#93105)
73e22ff 
We can guarantee NSW on all operands in a reassociated add expression
tree when:

- All adds in an add operator tree are NSW, AND either 
  - All add operands are guaranteed to be nonnegative, OR 
  - All adds are also NUW

- Alive2: 
- Nonnegative Operands
	- 3 operands: https://alive2.llvm.org/ce/z/G4XW6Q
	- 4 operands: https://alive2.llvm.org/ce/z/FWcZ6D
 - NUW NSW adds: https://alive2.llvm.org/ce/z/vRUxeC

---------

Co-authored-by: Nikita Popov <github@npopov.com>
[mlir][sparse] remove sparse encoding propagation pass. (llvm#93593)
9983592
@arsenm
DAG: Handle fminnum_ieee/fmaxnum_ieee in basic legalization
196a080 
Handle these in promote float and vector widening. Currently we happen
to avoid emitting these unless legal or custom. Avoids regression in
a future commit which wants to unconditionally emit these.
@aheejin
[WebAssembly] Add tests for EH/SjLj option errors (llvm#93583)
08de0b3 
This adds tests for EH/SjLj option errors and swaps the error checking
order for unimportant cosmetic reasons (I think checking EH/SjLj
conflicts is more important than the model checking)
@kartcq
[polly] Fix cppcheck SA comment reported in llvm#91235 (llvm#93505)
d33864d 
This patch moves the unreachable assert before return statement.
Fixes llvm#91235.
@AaronBallman
[C] Disallow declarations where a statement is required (llvm#92908)
5901d40 
This fixes a regression introduced in
8bd06d5 where Clang began to accept a
declaration where a statement is required. e.g.,
```
if (1)
  int x; // Previously accepted, now properly rejected
```

Fixes llvm#92775
@mgehre-amd
[AutoBump] Merge with 5901d400 (May 28)
9b0657c
@mgehre-amd mgehre-amd requested a review from cferry-AMD August 26, 2024 11:11
@mgehre-amd mgehre-amd mentioned this pull request Aug 28, 2024
Merged
Base automatically changed from bump_to_af22e274 to feature/fused-ops September 5, 2024 10:17
@jorickert jorickert merged commit 9d1002a into feature/fused-ops Sep 5, 2024
5 checks passed
@jorickert jorickert deleted the bump_to_5901d400 branch September 5, 2024 10:17
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@cferry-AMD cferry-AMD cferry-AMD approved these changes

Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.