]> granicus.if.org Git - llvm/commit
Merging r267916:
authorTom Stellard <thomas.stellard@amd.com>
Sat, 4 Jun 2016 03:43:04 +0000 (03:43 +0000)
committerTom Stellard <thomas.stellard@amd.com>
Sat, 4 Jun 2016 03:43:04 +0000 (03:43 +0000)
commit3dcb576961931fa9a653b42cc467f29044ec06d7
tree574163097a4f0837a7f7a3a6368bca8f4f9c7275
parentadf129408733e055c97b46ce0fa9b3f39a9b45ff
Merging r267916:

------------------------------------------------------------------------
r267916 | Matthew.Arsenault | 2016-04-28 11:38:48 -0700 (Thu, 28 Apr 2016) | 6 lines

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.

------------------------------------------------------------------------

git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_38@271768 91177308-0d34-0410-b5e6-96231b3b80d8
lib/Target/AMDGPU/AMDGPUPromoteAlloca.cpp
test/CodeGen/AMDGPU/array-ptr-calc-i32.ll
test/CodeGen/AMDGPU/indirect-private-64.ll
test/CodeGen/AMDGPU/promote-alloca-array-allocation.ll [new file with mode: 0644]