[PATCH] D20268: [wip] Resolution-based LTO API.

Teresa Johnson via llvm-commits llvm-commits at lists.llvm.org
Thu May 19 19:43:35 PDT 2016


tejohnson added a comment.

In http://reviews.llvm.org/D20268#435172, @pcc wrote:

> > That was my recommendation in the reviews on the earlier version of this patch, but a comment on the LTOBackend class made me think pcc planned to use this for the distributed backends themselves too (see the comment I noted below).
>
>
> Maybe I misunderstood you then. My understanding from this:
>
> > > Now that I think about it, there's probably no need to use the LTO class itself in the distributed backends.
>
> > 
>
> > 
>
> > Yes and I think it shouldn't be in the same library, but rather in Transforms/IPO.
>
>
> was that we should move the code that implements the distributed backend to IPO. That's what LTOBackend is.


Ok, I missed the fact that the LTOBackend code is in Transforms/IPO, not also in LTO (like the LTO class derived from it). So then this does seem to address my original concern.

Somehow EmitAssembly or CreatePasses will need to invoke this instead of some portion of the work they currently do to setup and run the pass pipeline.


http://reviews.llvm.org/D20268





More information about the llvm-commits mailing list