[PATCH] D52504: [DAGCombiner] Div/rem folds

[Dávid Bolvanský via Phabricator via llvm-commits]

xbolva00 added inline comments.


================
Comment at: test/CodeGen/X86/pr38539.ll:6
 ; This test is targeted at 64-bit mode. It used to crash due to the creation of an EXTRACT_SUBREG after the peephole pass had ran.
 define void @f() {
 ; X64-LABEL: f:
----------------
RKSimon wrote:
> xbolva00 wrote:
> > RKSimon wrote:
> > > These test cases need rebuilding so that they still check for PR38539 
> > So change %B8 = srem i66 0, %B20 to e.g.
> > %B8 = srem i66 3 (any value which does not fold), %B20 should be okay?
> @craig.topper will be able to confirm but it might be that you need to (locally) revert the fix for PR38539, apply your DAGCombiner.cpp patch and then run bugpoint to reduce the original fuzz code again: https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=8882
> 
I reverted https://reviews.llvm.org/rL339945 and I ran llvm-lit on pr38539.ll - test passed... (flags-copy-lowering.mir failed). Something else between that commit and top of trunk must fix it too.

So I think we should either remove test file or modify it with no worries.


https://reviews.llvm.org/D52504