[PATCH] D46191: [LV] Preserve inbounds on created GEPs

[Daniel Neilson via llvm-commits]

Yes, this was the cause of the break. Put an X86 test in the general-arch directory.

Fixed by: https://reviews.llvm.org/rL331281


On May 1, 2018, at 12:02 PM, Ben Hamilton via Phabricator <reviews at reviews.llvm.org<mailto:reviews at reviews.llvm.org>> wrote:

benhamilton added a comment.

I believe this broke at least one test:

http://lab.llvm.org:8011/builders/clang-cmake-armv7-quick/builds/1233/steps/ninja%20check%201/logs/FAIL%3A%20LLVM%3A%3Apr23997.ll

 ******************** TEST 'LLVM :: Transforms/LoopVectorize/pr23997.ll' FAILED ********************
 Script:
 --
 /home/buildslave/buildslave/clang-cmake-armv7-quick/stage1/bin/opt -S -loop-vectorize -dce -instcombine < /home/buildslave/buildslave/clang-cmake-armv7-quick/llvm/test/Transforms/LoopVectorize/pr23997.ll | /home/buildslave/buildslave/clang-cmake-armv7-quick/stage1/bin/FileCheck /home/buildslave/buildslave/clang-cmake-armv7-quick/llvm/test/Transforms/LoopVectorize/pr23997.ll
 --
 Exit Code: 1

 Command Output (stderr):
 --
 /home/buildslave/buildslave/clang-cmake-armv7-quick/llvm/test/Transforms/LoopVectorize/pr23997.ll:17:15: error: expected string not found in input
 ; CHECK-NEXT: [[TMP3:%.*]] = icmp ugt i64 [[TMP2:%.*]], 1
               ^
 <stdin>:16:2: note: scanning from here
  br label %loop
  ^
 <stdin>:25:2: note: possible intended match here
  %.21 = icmp ult i64 %indvars.iv.next4, %2
  ^

 --

 ********************

Can you take a look, please?


Repository:
 rL LLVM

https://reviews.llvm.org/D46191



-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20180501/c7ff8926/attachment.html>