[PATCH] D89170: [AMDGPU] Select flat scratch instructions where available

Stanislav Mekhanoshin via Phabricator via llvm-commits llvm-commits at lists.llvm.org
Mon Oct 12 12:59:57 PDT 2020


rampitec added a comment.

In D89170#2325862 <https://reviews.llvm.org/D89170#2325862>, @t-tye wrote:

> In D89170#2325299 <https://reviews.llvm.org/D89170#2325299>, @scott.linder wrote:
>
>> I haven't done a meaningful review, but I wanted to note that this will require changes to the debug information (which isn't committed yet). I think this could be as simple as scaling the value of the SP/FP in the DWARF expression to recover the unswizzled scratch address.
>
> Is this change changing the call convention ABI? For example, making the SP be a swizzled address as opposed to a swizzled address? If so then AMDGPUUsage will also need updating.

Yes it does. However, it is a little premature to update the documentation. This is WIP, disabled by default and more or less not working at least until spilling is implemented. When it is at least working we can consider documenting it. Documenting it earlier just gives an impression there is an option to use it.


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D89170/new/

https://reviews.llvm.org/D89170



More information about the llvm-commits mailing list