5 Commits

Author SHA1 Message Date
Matt Arsenault
50caf6936b AMDGPU: Convert promote alloca tests to opaque pointers 2022-11-28 10:36:38 -05:00
Matt Arsenault
c1710e7779 AMDGPU: Use modern address spaces in some tests
This was way out of date, still using 4 for generic and 0 for private.
2022-11-28 10:05:06 -05:00
Matt Arsenault
1310aa1688 AMDGPU: Use -passes for amdgpu-promote-alloca tests 2022-11-16 17:14:48 -08:00
Matt Arsenault
3dbeefa978 AMDGPU: Mark all unspecified CC functions in tests as amdgpu_kernel
Currently the default C calling convention functions are treated
the same as compute kernels. Make this explicit so the default
calling convention can be changed to a non-kernel.

Converted with perl -pi -e 's/define void/define amdgpu_kernel void/'
on the relevant test directories (and undoing in one place that actually
wanted a non-kernel).

llvm-svn: 298444
2017-03-21 21:39:51 +00:00
Matt Arsenault
c5fce69031 AMDGPU: Fix mishandling array allocations when promoting alloca
The canonical form for allocas is a single allocation of the array type.
In case we see a non-canonical array alloca, make sure we aren't
replacing this with an array N times smaller.

llvm-svn: 267916
2016-04-28 18:38:48 +00:00