[llvm] r309651 - [StackColoring] Update AliasAnalysis information in stack coloring pass

[Hans Wennborg via llvm-commits]

On Tue, Aug 1, 2017 at 10:51 AM, Hal Finkel <hfinkel at anl.gov> wrote:
>
> On 08/01/2017 12:38 PM, Matthias Braun wrote:
>
> This is going a little fast, given that Hal had raised concerns and the
> patch went in before he responded (at least it looks like that in phab). I'd
> like to hear his opinion first, given that my experience with alias analysis
> in LLVM is limited.
>
>
> I'd like to see this in 5.0, however, I just provided a post-commit review
> and it looks like it will need a small update before merging.

Sounds good. Thanks!



> On Aug 1, 2017, at 10:15 AM, Hans Wennborg via llvm-commits
> <llvm-commits at lists.llvm.org> wrote:
>
> Matthias, the PR for this was marked as a release blocker for 5.0.
> What do you think about merging this?
>
> On Mon, Jul 31, 2017 at 8:32 PM, Hiroshi Inoue via llvm-commits
> <llvm-commits at lists.llvm.org> wrote:
>
> Author: inouehrs
> Date: Mon Jul 31 20:32:15 2017
> New Revision: 309651
>
> URL: http://llvm.org/viewvc/llvm-project?rev=309651&view=rev
> Log:
> [StackColoring] Update AliasAnalysis information in stack coloring pass
>
> Stack coloring pass need to maintain AliasAnalysis information when merging
> stack slots of different types.
> Actually, there is a FIXME comment in StackColoring.cpp
>
> // FIXME: In order to enable the use of TBAA when using AA in CodeGen,
> // we'll also need to update the TBAA nodes in MMOs with values
> // derived from the merged allocas.
>
> But, TBAA has been already enabled in CodeGen without fixing this pass.
> The incorrect TBAA metadata results in recent failures in bootstrap test on
> ppc64le (PR33928) by allowing unsafe instruction scheduling.
> Although we observed the problem on ppc64le, this is a platform neutral
> issue.
>
> This patch makes the stack coloring pass maintains AliasAnalysis information
> when merging multiple stack slots.
>
>
> (This is https://reviews.llvm.org/D35907)
> _______________________________________________
> llvm-commits mailing list
> llvm-commits at lists.llvm.org
> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-commits
>
>
>
> --
> Hal Finkel
> Lead, Compiler Technology and Programming Languages
> Leadership Computing Facility
> Argonne National Laboratory