diff options
author | Tom Stellard <thomas.stellard@amd.com> | 2016-11-23 17:05:01 +0000 |
---|---|---|
committer | Tom Stellard <thomas.stellard@amd.com> | 2016-11-23 17:05:01 +0000 |
commit | 38f95bbf60fcfc13ad680a984811d821ea20cd40 (patch) | |
tree | da714ec18cbfece0b5e5698680cea6768f9114a2 | |
parent | c92639e0c5b9882c21135e71d41b6fe0600e042e (diff) | |
download | llvm-38f95bbf60fcfc13ad680a984811d821ea20cd40.zip llvm-38f95bbf60fcfc13ad680a984811d821ea20cd40.tar.gz llvm-38f95bbf60fcfc13ad680a984811d821ea20cd40.tar.bz2 |
Merging r280143:
------------------------------------------------------------------------
r280143 | dberlin | 2016-08-30 12:58:48 -0700 (Tue, 30 Aug 2016) | 2 lines
IntrArgMemOnly is only defined (and current AA machinery only sanely supports) pointer arguments, and these intrinsics have vector of pointer arguments. Remove ArgMemOnly until we either have the machinery, define a new attribute, or something similar
------------------------------------------------------------------------
llvm-svn: 287777
-rw-r--r-- | llvm/include/llvm/IR/Intrinsics.td | 5 | ||||
-rw-r--r-- | llvm/test/Transforms/GVN/2016-08-30-MaskedScatterGather.ll | 42 |
2 files changed, 44 insertions, 3 deletions
diff --git a/llvm/include/llvm/IR/Intrinsics.td b/llvm/include/llvm/IR/Intrinsics.td index 5ece731..e9264ec 100644 --- a/llvm/include/llvm/IR/Intrinsics.td +++ b/llvm/include/llvm/IR/Intrinsics.td @@ -668,13 +668,12 @@ def int_masked_gather: Intrinsic<[llvm_anyvector_ty], [LLVMVectorOfPointersToElt<0>, llvm_i32_ty, LLVMVectorSameWidth<0, llvm_i1_ty>, LLVMMatchType<0>], - [IntrReadMem, IntrArgMemOnly]>; + [IntrReadMem]>; def int_masked_scatter: Intrinsic<[], [llvm_anyvector_ty, LLVMVectorOfPointersToElt<0>, llvm_i32_ty, - LLVMVectorSameWidth<0, llvm_i1_ty>], - [IntrArgMemOnly]>; + LLVMVectorSameWidth<0, llvm_i1_ty>]>; // Test whether a pointer is associated with a type metadata identifier. def int_type_test : Intrinsic<[llvm_i1_ty], [llvm_ptr_ty, llvm_metadata_ty], diff --git a/llvm/test/Transforms/GVN/2016-08-30-MaskedScatterGather.ll b/llvm/test/Transforms/GVN/2016-08-30-MaskedScatterGather.ll new file mode 100644 index 0000000..3f8fdcc --- /dev/null +++ b/llvm/test/Transforms/GVN/2016-08-30-MaskedScatterGather.ll @@ -0,0 +1,42 @@ +; RUN: opt < %s -basicaa -gvn -S | FileCheck %s + +declare void @llvm.masked.scatter.v2i32(<2 x i32> , <2 x i32*> , i32 , <2 x i1> ) +declare <2 x i32> @llvm.masked.gather.v2i32(<2 x i32*>, i32, <2 x i1>, <2 x i32>) + +; This test ensures that masked scatter and gather operations, which take vectors of pointers, +; do not have pointer aliasing ignored when being processed. +; No scatter/gather calls should end up eliminated +; CHECK: llvm.masked.gather +; CHECK: llvm.masked.gather +; CHECK: llvm.masked.scatter +; CHECK: llvm.masked.gather +; CHECK: llvm.masked.scatter +; CHECK: llvm.masked.gather +define spir_kernel void @test(<2 x i32*> %in1, <2 x i32*> %in2, i32* %out) { +entry: + ; Just some temporary storage + %tmp.0 = alloca i32 + %tmp.1 = alloca i32 + %tmp.i = insertelement <2 x i32*> undef, i32* %tmp.0, i32 0 + %tmp = insertelement <2 x i32*> %tmp.i, i32* %tmp.1, i32 1 + ; Read from in1 and in2 + %in1.v = call <2 x i32> @llvm.masked.gather.v2i32(<2 x i32*> %in1, i32 1, <2 x i1> <i1 true, i1 true>, <2 x i32> undef) #1 + %in2.v = call <2 x i32> @llvm.masked.gather.v2i32(<2 x i32*> %in2, i32 1, <2 x i1> <i1 true, i1 true>, <2 x i32> undef) #1 + ; Store in1 to the allocas + call void @llvm.masked.scatter.v2i32(<2 x i32> %in1.v, <2 x i32*> %tmp, i32 1, <2 x i1> <i1 true, i1 true>); + ; Read in1 from the allocas + ; This gather should alias the scatter we just saw + %tmp.v.0 = call <2 x i32> @llvm.masked.gather.v2i32(<2 x i32*> %tmp, i32 1, <2 x i1> <i1 true, i1 true>, <2 x i32> undef) #1 + ; Store in2 to the allocas + call void @llvm.masked.scatter.v2i32(<2 x i32> %in2.v, <2 x i32*> %tmp, i32 1, <2 x i1> <i1 true, i1 true>); + ; Read in2 from the allocas + ; This gather should alias the scatter we just saw, and not be eliminated + %tmp.v.1 = call <2 x i32> @llvm.masked.gather.v2i32(<2 x i32*> %tmp, i32 1, <2 x i1> <i1 true, i1 true>, <2 x i32> undef) #1 + ; Store in2 to out for good measure + %tmp.v.1.0 = extractelement <2 x i32> %tmp.v.1, i32 0 + %tmp.v.1.1 = extractelement <2 x i32> %tmp.v.1, i32 1 + store i32 %tmp.v.1.0, i32* %out + %out.1 = getelementptr i32, i32* %out, i32 1 + store i32 %tmp.v.1.1, i32* %out.1 + ret void +} |