Splitting the verifier into a separate diff: https://reviews.llvm.org/D68345
Oct 2 2019
It depends on the verifier function (CodeExtractor::verifyAssumptionCache) from D67941 to fail with the test case in: llvm/test/Transforms/HotColdSplit/assumption-cache-invalidation.ll
Relevant test case: llvm/test/Transforms/HotColdSplit/assumption-cache-invalidation.ll is already in trunk.
passed ninja check-all
ninja check passed
Oct 1 2019
Essentially did this:
diff --git a/llvm/lib/Transforms/Utils/CodeExtractor.cpp b/llvm/lib/Transforms/Utils/CodeExtractor.cpp index 952263c0b1f..49fa07a3648 100644 --- a/llvm/lib/Transforms/Utils/CodeExtractor.cpp +++ b/llvm/lib/Transforms/Utils/CodeExtractor.cpp @@ -1336,13 +1336,13 @@ void CodeExtractor::calculateNewCallTerminatorWeights(
Do we have further feedback for this patch?
LGTM, thanks for committing.
Fixed a bug in assumption cache where the AssumeVH wasn't getting cleared.
Sep 30 2019
added flag to the testcase.
Addressed comments from @vsk
- Move input parameter validation to isEligible().
This now requires recomputing inputs in extractCodeRegion because PHI nodes, and return blocks are split. The computation cost of input is same as before because previously the input parameter would be computed on the client side (caller of extractCodeRegion) and introduce extra overhead to be taken care of.
Comments from @compnerd
LGTM, makes it easier to read as well!
Sep 29 2019
fix clonetype tests
port to llvm bitcode version 8
after collecting numbers
Great, looking forward to it!
Can we add a test to verify cfi instructions are present without debug flag.
i think this will help quite a few optimizations get faster and enable others like instruction scheduling in SSA. We can push this patch and make incremental improvements as needed. Thanks for working on this!
LGTM, can we add test case from the PR?
closed in r345828, 314fbfa1c4c6665c54a220eefb10a6f23010a352
Can we get review on this patch.
it will be great to merge this patch.
Sep 28 2019
Is there any feedback for improvement, I'd like to push this if reasonable.
Sep 26 2019
@vsk I tried adding test to CodeExtractorTest (D68095), for some reason the verifier didn't fail. I may be missing something and can use some help.
We can follow up on the test in the new diff if that is reasonable.
Added AssumptionCache verifier to verify the original function.