search for: d40375

...t hoisting in the presence of volatile loads? I was looking at this testcase: test/Transforms/LICM/volatile-alias.ll Context: MemorySSA treats volatile loads as defs. I'm looking to better understand expected behavior in the presence of volatile accesses. More context: https://reviews.llvm.org/D40375. Thanks in advance, Alina -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20171219/ebbf37b4/attachment.html>

...acker. > > > The conclusion I'm drawing is that right now AliasSetTracker and > MemorySSA have different behaviors and replacing one with the either > will naturally lead to different outcomes. > So, how can I make progress here? > > > I think it's reasonable in D40375 <https://reviews.llvm.org/D40375> to > have pointerInvalidatedByLoopWithMSSA only check if the defining > access is within the current loop or liveOnEntry, and rely on > MemorySSA to either consider a volatile load a clobbering access or > not. So, right now the LICM/volatile-...

On 12/20/2017 1:37 PM, Sanjoy Das wrote:> > Fwiw, I was under the impression that regular loads could *not* be > reordered with volatile loads since we could have e.g.: > > int *normal = &global_variable; > volatile int* ptr = 0; > int k = *ptr; // segfaults, and the signal handler writes to *normal > int value = *normal; > > and that we'd have

...gt;, D45301 <https://reviews.llvm.org/D45301>, D47022 <https://reviews.llvm.org/D47022>, D51718 <https://reviews.llvm.org/D51718>). LICM is the loop pass I've picked back up now, and the first that also uses MemorySSA vs just updating it. Patches in Phabricator are fairly old (D40375 <https://reviews.llvm.org/D40375>, D35741 <https://reviews.llvm.org/D35741>). I'm going to upload the rebased and updated versions once they're ready for review. The actual use/update is currently disabled by default (flag to enable: EnableMSSALoopDependency). I also recently b...

...? >> I was looking at this testcase: test/Transforms/LICM/volatile-alias.ll >> >> Context: MemorySSA treats volatile loads as defs. I'm looking to better >> understand expected behavior in the presence of volatile accesses. >> More context: https://reviews.llvm.org/D40375. >> >> Thanks in advance, >> Alina >> > > -- > Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, hosted > by The Linux Foundation > -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.llvm.org/piper...