[PATCH] [Power] Use AtomicExpandPass for fence insertion, and use lwsync where appropriate
Hal Finkel
hfinkel at anl.gov
Thu Sep 4 09:34:04 PDT 2014
Manuel,
I posted a review to this patch about an hour ago, but I don't think a corresponding e-mail made it to the list. Can you please check on it?
Thanks again,
Hal
----- Original Message -----
> From: "Robin Morisset" <morisset at google.com>
> To: morisset at google.com, jfb at chromium.org
> Cc: llvm-commits at cs.uiuc.edu
> Sent: Wednesday, September 3, 2014 6:23:37 PM
> Subject: [PATCH] [Power] Use AtomicExpandPass for fence insertion, and use lwsync where appropriate
>
> Hi jfb,
>
> This patch makes use of AtomicExpandPass in Power for inserting
> fences around
> atomic as part of an effort to remove fence insertion from
> SelectionDAGBuilder.
> As a big bonus, it lets us use sync 1 (lightweight sync, often used
> by the mnemonic
> lwsync) instead of sync 0 (heavyweight sync) in many cases.
>
> I also added a test, as there was no test for the barriers emitted by
> the Power
> backend for atomic loads and stores.
>
> http://reviews.llvm.org/D5180
>
> Files:
> include/llvm/IR/IntrinsicsPowerPC.td
> lib/Target/PowerPC/PPCISelLowering.cpp
> lib/Target/PowerPC/PPCISelLowering.h
> lib/Target/PowerPC/PPCInstrInfo.td
> lib/Target/PowerPC/PPCTargetMachine.cpp
> test/CodeGen/PowerPC/atomics.ll
>
> _______________________________________________
> llvm-commits mailing list
> llvm-commits at cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits
>
--
Hal Finkel
Assistant Computational Scientist
Leadership Computing Facility
Argonne National Laboratory
More information about the llvm-commits
mailing list