[PATCH] D119027: [AMDGPU][NFC]: Emit metadata for hidden_heap_v1 kernarg

Mon Feb 21 22:39:54 PST 2022

sameerds added inline comments.

================
Comment at: llvm/lib/Target/AMDGPU/AMDGPUAttributor.cpp:413
-    if (funcRetrievesHostcallPtr(A)) {
-      removeAssumedBits(IMPLICIT_ARG_PTR);
       removeAssumedBits(HOSTCALL_PTR);
----------------
cfang wrote:
> sameerds wrote:
> > cfang wrote:
> > > sameerds wrote:
> > > > Why was this removed? Is there some other place which guarantees the same relationship?
> > > I think the explicit call of amdgcn_implicitarg_ptr will guarantee IMPLICIT_ARG_PTR.
> > That sounds correct. But then this should be asserted somewhere ... probably in the code that manifests the attribute.
> Do you think the for loop a few lines ahead in the same function should and "MUST" catch amdgcn_implicitarg_ptr? So that we don't need an assert?
> 
> 386: for (Function *Callee : AAEdges.getOptimisticEdges()) {
Maybe it does. But that's not the point. This relationship between the two attributes is pretty important and we should assert. Even the coding standard says so [1] and it is pretty much a life-saver. Maybe I should have asserted myself, but that last change was kinda in a bit of a hurry!

[1] https://llvm.org/docs/CodingStandards.html#assert-liberally

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D119027/new/

https://reviews.llvm.org/D119027