[PATCH] D69498: IR: Invert convergent attribute handling

Mehdi AMINI via Phabricator via llvm-commits llvm-commits at lists.llvm.org
Sat Nov 2 10:57:27 PDT 2019


mehdi_amini added a comment.

In D69498#1727650 <https://reviews.llvm.org/D69498#1727650>, @dexonsmith wrote:

> In D69498#1723606 <https://reviews.llvm.org/D69498#1723606>, @rjmccall wrote:
>
> > Perhaps there should be a global metadata, or something in the increasingly-misnamed "data layout" string, which says that convergence is meaningful, and we should only add the attribute in appropriately-annotated modules?
>
>
> Just wanted to resurface these alternatives from John.  Given that some targets want a fundamentally different default from what most frontends expect, I think we ought to find some way to encode the difference.


Just thought about a slight variation on this: what about adding a flag on the datalayout (or a module flag) but not use it in the transformations/analyses, instead use it only when creating Function by always setting the `convergent` attribute when the flag is present? This would require to always have a Module passed to the Function constructor though (the C API already does, we would just need to update the C++ users).

So creating a Function in a Module with this flag would have the convergent attribute set on creation (can be unset explicitly).


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D69498/new/

https://reviews.llvm.org/D69498





More information about the llvm-commits mailing list