[PATCH] D15504: Bug fix for shrink-wrap

[Quentin Colombet via llvm-commits]

> On Dec 14, 2015, at 12:09 PM, Quentin Colombet via llvm-commits <llvm-commits at lists.llvm.org> wrote:
> 
> 
>> On Dec 14, 2015, at 11:59 AM, Weiming Zhao <weimingz at codeaurora.org> wrote:
>> 
>> weimingz added a comment.
>> 
>> This patch is more of a workaround to fix a bug. It seems to me the conditions of the safe point is not strong enough.
> 
> I agree, but the fix seems too narrow, that’s why I would like a test case to make sure that whatever we come up with fixes it.

A proper workaround would be to push the restore block outside of the loop since post-dominance does not give us the proper guarantee for loops.

> 
> 
>> Since it's not SSA, the dominance relationship can't guarantee that the save point is before the def.

It’s not really about SSA by the way, it has more to do by the fact that post-dominance is counter intuitive in loops IMO :).

Q.

>> 
>> 
>> Repository:
>> rL LLVM
>> 
>> http://reviews.llvm.org/D15504
>> 
>> 
>> 
> 
> _______________________________________________
> llvm-commits mailing list
> llvm-commits at lists.llvm.org
> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-commits