yonghong-song added a comment.
I just tested the patch with kernel bpf selftests and it works fine. The patch
looks good to me too.
================
Comment at: llvm/lib/Target/BPF/BPFAbstractMemberAccess.cpp:303
CInfo.RecordAlignment =
DL->getABITypeAlign(CInfo.Base->getType()->getPointerElementType());
return true;
----------------
nikic wrote:
> I noticed too late that this intrinsic also uses the element type to get the
> alignment. I'll have to follow up with the same change for this intrinsic.
>
> Though I'm not completely sure if this is actually used, because no tests
> fail if I replace this with dummy values like 1 or 128.
This is for llvm.preserve.union.access.index. It is used only if the union
member type is a bitfield and we don't have any usecase and test case for that.
That is why we didn't hit it.
It is okay for now. I probably will change to use getBaseElementType(Call)
later by myself.
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D106184/new/
https://reviews.llvm.org/D106184
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
