[PATCH] D14489: [AArch64] Applying load pair optimization for volatile load/store

[Chad Rosier via llvm-commits]

mcrosier added a comment.

In http://reviews.llvm.org/D14489#285326, @t.p.northover wrote:

> Stopping the scan at 1 seems a bit hacky though, and is there any particular reason you limit the transformation to loads?


I agree with Tim.  Setting the scan limit to 1 isn't the right solution.  You should be able to scan until you encounter the first memory operation.


http://reviews.llvm.org/D14489