forked from OSchip/llvm-project
d57e9aca01
Summary: As [[ https://bugs.llvm.org/show_bug.cgi?id=45360 | PR45360 ]] reports, with new cost-model we can sometimes end up being able to expand `udiv`/`urem` instructions. And that exposes at least one instance of when we do that regardless of whether or not it is safe to do. In this particular case, it's `SimplifyIndvar::replaceIVUserWithLoopInvariant()`. It seems to me, we simply need to check with `isSafeToExpandAt()` first. The test isn't great. I'm not sure how to make it only run `-indvars`. Fixes [[ https://bugs.llvm.org/show_bug.cgi?id=45360 | PR45360 ]]. Reviewers: mkazantsev, reames, helloqirun Reviewed By: mkazantsev Subscribers: hiraditya, llvm-commits Tags: #llvm Differential Revision: https://reviews.llvm.org/D82108 |
||
|---|---|---|
| .. | ||
| 2009-04-14-shorten_iv_vars.ll | ||
| 2009-04-15-shorten-iv-vars-2.ll | ||
| 2011-11-15-multiexit.ll | ||
| deterministic-scev-verify.ll | ||
| eliminate-trunc.ll | ||
| huge_muls.ll | ||
| indvar-debug-value.ll | ||
| indvar-debug-value2.ll | ||
| inner-loop-by-latch-cond.ll | ||
| inner-loop.ll | ||
| iv-widen.ll | ||
| lit.local.cfg | ||
| loop-invariant-conditions.ll | ||
| loop_evaluate10.ll | ||
| overflow-intrinsics.ll | ||
| polynomial-expand.ll | ||
| pr24356.ll | ||
| pr24804.ll | ||
| pr24956.ll | ||
| pr25047.ll | ||
| pr25051.ll | ||
| pr25360.ll | ||
| pr25421.ll | ||
| pr25576.ll | ||
| pr26973.ll | ||
| pr27133.ll | ||
| pr28935.ll | ||
| pr35406.ll | ||
| pr45360.ll | ||
| replace-iv-with-loop-invariant.ll | ||
| scev-phi-debug-info.ll | ||
| uglygep.ll | ||
| variable-stride-ivs-1.ll | ||
| verify-scev.ll | ||
| widen-nsw.ll | ||