Sanjay Patel 4345e14522 [InstCombine] avoid infinite loops from min/max canonicalization ...

The intrinsics have an extra chunk of known bits logic
compared to the normal cmp+select idiom. That allows
folding the icmp in each case to something better, but
that then opposes the canonical form of min/max that
we try to form for a select.

I'm carving out a narrow exception to preserve all
existing regression tests while avoiding the inf-loop.
It seems unlikely that this is the only bug like this
left, but this should fix:
https://llvm.org/PR51419

(cherry picked from commit b267d3ce8defa092600bda717ff18440d002f316)

2021-08-16 11:35:38 -07:00

..

CMakeLists.txt

llvmbuildectomy - replace llvm-build by plain cmake

2020-11-13 10:35:24 +01:00

InstCombineAddSub.cpp

[InstCombine] fold reassociative FP add into start value of fadd reduction

2021-07-18 06:26:20 -04:00

InstCombineAndOrXor.cpp

[InstCombine] InstCombinerImpl::visitOr - enable bitreverse matching

2021-05-15 13:39:09 +01:00

InstCombineAtomicRMW.cpp

Add missing namespace closure comments. NFCI.

2020-09-23 16:19:25 +01:00

InstCombineCalls.cpp

[NFC][InstCombine] Fix typo

2021-07-27 11:33:10 +08:00

InstCombineCasts.cpp

[InstCombine] Eliminate casts to optimize ctlz operation

2021-06-23 11:19:12 -04:00

InstCombineCompares.cpp

[InstCombine] avoid infinite loops from min/max canonicalization

2021-08-16 11:35:38 -07:00

InstCombineInternal.h

[InstCombine] Add freezeAllUsesOfArgument to visitFreeze

2021-07-24 18:08:58 +09:00

InstCombineLoadStoreAlloca.cpp

Fixing an infinite loop problem in InstCombine

2021-08-04 16:51:40 -07:00

InstCombineMulDivRem.cpp

[InstCombine] Fold (sext bool X) * (sext bool X) to zext (and X, X)

2021-06-18 16:28:06 +07:00

InstCombineNegator.cpp

[Transforms] Use range-based for loops (NFC)

2021-02-08 22:33:53 -08:00

InstCombinePHI.cpp

[InstCombine] Don't combine PHI before catchswitch

2021-07-02 12:10:24 -07:00

InstCombineSelect.cpp

[InstCombine] Fixed select + masked load fold failure

2021-08-06 12:41:06 -07:00

InstCombineShifts.cpp

[InstCombine] Fold lshr/ashr(or(neg(x),x),bw-1) --> zext/sext(icmp_ne(x,0)) (PR50816)

2021-07-13 14:44:54 +01:00

InstCombineSimplifyDemanded.cpp

[InsCombine] Fix a few remaining vec transforms to use poison instead of undef

2021-05-31 18:47:09 +09:00

InstCombineVectorOps.cpp

[NFC] Fix a few whitespace issues and typos.

2021-07-04 11:49:58 +01:00

InstructionCombining.cpp

[LLVM IR] Allow volatile stores to trap.

2021-07-26 10:51:00 -07:00