Mon, Sep 7
Thu, Sep 3
Wed, Sep 2
LGTM (looks like the other example plugins just use PRIVATE as well).
Jul 15 2020
Jul 14 2020
Jul 3 2020
This is causing libcxx test failures in AArch64 -O0. Reduced testcase (the inline asm ultimately comes from DoNotOptimize in libcxx/test/support/test_macros.h):
Jun 29 2020
Jun 26 2020
Don't print the diagnostic with -gen-reproducer.
Jun 25 2020
It looks like this patch is causing incorrect code generation for
Jun 23 2020
Jun 22 2020
Print the bug report message when forcing a crash, but also have the message available even with LLVM_ENABLE_BACKTRACES=OFF.
Jun 18 2020
Moved BugReportMsg printing into CrashRecoveryContext, and stopped printing of backtrace when forcing a crash.
Jun 16 2020
Added a couple of reviewers that have recently worked on CrashRecoveryContext.
Jun 11 2020
Changes look OK, though as I said before please add tests for this.
Jun 3 2020
It look like there have been recent changes to BranchProbabilityInfo that mean this patch needs to be rebased.
May 26 2020
May 22 2020
Detect more kinds of local class.
May 21 2020
Use the tag of the anonymous struct when emitting a diagnostic.
May 20 2020
Adjusted to not give multiple errors when more than one error criteria is met.
May 18 2020
Please add some tests for this to llvm/test/CodeGen/ARM/.
May 14 2020
May 13 2020
Could you add some tests (unless they're added in a later patch)?
May 12 2020
Apr 22 2020
Apr 21 2020
Adjust comments. Also, after double-checking the 8-A and 8-M ARMARMs it's OK if the base register being stored is the lowest one so adjusted to allow that.
Apr 20 2020
Apr 15 2020
With this patch the Threading.PhysicalConcurrency unit test fails when run with an affinity less than the number of physical cpus. I've raised https://bugs.llvm.org/show_bug.cgi?id=45556.
Apr 3 2020
Rebase and add adjust MVE masked load handling. Also make use of update_llc_checks - I tried it again and this time it did use a single CHECK when both little-endian and big-endian code generation are the same, so it looks like I was using it wrong somehow before (though I don't know in what way).
Mar 27 2020
Mar 26 2020
Do this in ARMTargetLowering::PerformDAGCombine
Release note added: https://reviews.llvm.org/rG0cff81cff05d
Mar 25 2020
Mar 24 2020
Would it make sense to do this as a DAG fold? That way we would get MVE for free too.
A few small changes:
- Don't start C++11 spelling with :: (changes to name normalisation to allow this are in D76704)
- Don't check that the Decl is null in diagAppertainsToDecl
- Return AttributeNotApplied on errors in handleDeclAttribute
Mar 23 2020
Apply clang-format changes.
Mar 20 2020
Mar 18 2020
Use an enum for the return value.
Updated to match the changes to D31342
Mar 17 2020
Mar 16 2020
Adjust GEP handling to reduce AllocSize by Offset instead of using the result size.
Mar 13 2020
Use the allocation size instead of the allocation type.
Mar 12 2020
Adjusted to avoid looking through pointer types by passing around the allocated type and checking it against the size of memory accesses. Also correctly handle GEPs where the addressed member is partially outside of the allocated object.
Mar 11 2020
Added bitcast handling.
Mar 10 2020
Removed loop tests.
Changes look good, but there's still the question about whether you really need the loop tests.
Mar 9 2020
Update based on review comments.
Mar 6 2020
Use getTypeAllocSize, apply clang-format.
Mar 5 2020
Mar 4 2020
Mar 3 2020
Abandoning this old patch (the underlying problem was solved by https://reviews.llvm.org/rL348496 which made GVN PRE not introduce the problematic phis in the first place).
Abandoning this old patch that I have no intention of doing any further work on.