[PATCH] D79052: [clang codegen] Fix alignment of "Address" for incomplete array pointer.

[John McCall via Phabricator via cfe-commits]

rjmccall added inline comments.


================
Comment at: clang/lib/CodeGen/CodeGenFunction.cpp:176
   CharUnits Alignment;
-  if (T->isIncompleteType()) {
+  if (T->getBaseElementTypeUnsafe()->isIncompleteType()) {
     Alignment = CharUnits::One(); // Shouldn't be used, but pessimistic is best.
----------------
efriedma wrote:
> rsmith wrote:
> > I don't know if it matters in practice, but this is still wrong. An incomplete type can have a known alignment, for a case like `struct alignas(32) S;`. Perhaps we should remove this test entirely and call `getTypeAlignIfKnown` instead of `getTypeAlign[InChars]` below.
> I can't think of any way to observe the alignment computed by getNaturalTypeAlignment for an incomplete class. We usually only use the alignment computed by getNaturalTypeAlignment() to set the alignment of memory operations, and you can't do any memory operations with an incomplete class.
> 
> But the result might be easier to read, in any case.
If there's a `getTypeAlignIfKnown()`, it would be better to use it.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D79052/new/

https://reviews.llvm.org/D79052