[PATCH] D22879: [CUDA] Align kernel launch args correctly when the LLVM type's alignment is different from the clang type's alignment.

[Justin Lebar via cfe-commits]

jlebar added inline comments.

================
Comment at: test/CodeGenCUDA/kernel-args-alignment.cu:1-2
@@ +1,3 @@
+// REQUIRES: x86-registered-target
+// REQUIRES: nvptx-registered-target
+
----------------
rnk wrote:
> Typically clang doesn't need a registered backend for a target to generate IR for that target. It "knows" a whole bunch of stuff about all target calling conventions and data layout. Unless CUDA goes out of its way to query LLVM backend information, we shouldn't need these REQUIRES lines.
> 
> You should probably test this theory, though, by configuring an ARM-only clang and running the tests. :)
Yeah, I don't think we actually need this, as we have a bunch of other codegen tests that don't have these REQUIRES lines.


https://reviews.llvm.org/D22879