[llvm-commits] PATCH: Followup to r152556: Use more powerful instsimplify when actually performing the inline of a function

[Chris Lattner]

On Mar 21, 2012, at 4:56 AM, Chandler Carruth wrote:

> On Wed, Mar 21, 2012 at 4:15 AM, Chandler Carruth <chandlerc at gmail.com> wrote:
> There are a couple of things I want to do as follow-ups: simplify PHINodes to catch cases where all phi operands end up equal. This will in-turn want to recursively apply that simplification to users of the PHINodes. This has to happen in the lazy PHINode resolution pass, so I wanted to do it in a follow-up patch.
> 
> FYI, here is the patch for this. It's still a bit untested, but wanted to go ahead and get it out in front of folks for review. It at least didn't do anything horrible to a bootstrap.

This sounds great.  Just to verify, this does catch opportunities exposed by "pruning" the CFG (removing predecessors of phis), right?

-Chris

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20120321/b0e76a4c/attachment.html>