AMDGPU: Don't look at chain users when adjusting writemask

[mareko]

LGTM.

[arsenm]

r295878

lib/Target/AMDGPU/SIISelLowering.cpp

void SITargetLowering::adjustWritemask(MachineSDNode *&Node,
   unsigned DmaskIdx = (Node->getNumOperands() - Node->getNumValues() == 9) ? 2 : 3;
   unsigned OldDmask = Node->getConstantOperandVal(DmaskIdx);
   unsigned NewDmask = 0;
 
   // Try to figure out the used register components
   for (SDNode::use_iterator I = Node->use_begin(), E = Node->use_end();
        I != E; ++I) {
 
+    // Don't look at users of the chain.
+    if (I.getUse().getResNo() != 0)
+      continue;
+
     // Abort if we can't understand the usage
     if (!I->isMachineOpcode() ||
         I->getMachineOpcode() != TargetOpcode::EXTRACT_SUBREG)
       return;
 
     // Lane means which subreg of %VGPRa_VGPRb_VGPRc_VGPRd is used.
     // Note that subregs are packed, i.e. Lane==0 is the first bit set
     // in OldDmask, so it can be any of X,Y,Z,W; Lane==1 is the second bit

test/CodeGen/AMDGPU/llvm.amdgcn.image.sample.ll

 ; GCN: image_sample {{v\[[0-9]+:[0-9]+\]}}, {{v\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}} dmask:0x3
 define void @sample_v2f32(<2 x float> addrspace(1)* %out) {
 main_body:
   %r = call <2 x float> @llvm.amdgcn.image.sample.v2f32.v4f32.v8i32(<4 x float> undef, <8 x i32> undef, <4 x i32> undef, i32 3, i1 0, i1 0, i1 0, i1 0, i1 0)
   store <2 x float> %r, <2 x float> addrspace(1)* %out
   ret void
 }
 
+; GCN-LABEL: {{^}}adjust_writemask_sample_0:
+; GCN: image_sample v{{[0-9]+}}, v{{\[[0-9]+:[0-9]+\]}}, s{{\[[0-9]+:[0-9]+\]}}, s{{\[[0-9]+:[0-9]+\]}} dmask:0x1{{$}}
+define void @adjust_writemask_sample_0(float addrspace(1)* %out) {
+main_body:
+  %r = call <4 x float> @llvm.amdgcn.image.sample.v4f32.v4f32.v8i32(<4 x float> undef, <8 x i32> undef, <4 x i32> undef, i32 15, i1 0, i1 0, i1 0, i1 0, i1 0)
+  %elt0 = extractelement <4 x float> %r, i32 0
+  store float %elt0, float addrspace(1)* %out
+  ret void
+}
+
+; GCN-LABEL: {{^}}adjust_writemask_sample_01:
+; GCN: image_sample v{{\[[0-9]+:[0-9]+\]}}, v{{\[[0-9]+:[0-9]+\]}}, s{{\[[0-9]+:[0-9]+\]}}, s{{\[[0-9]+:[0-9]+\]}} dmask:0x3{{$}}
+define void @adjust_writemask_sample_01(float addrspace(1)* %out) {
+main_body:
+  %r = call <4 x float> @llvm.amdgcn.image.sample.v4f32.v4f32.v8i32(<4 x float> undef, <8 x i32> undef, <4 x i32> undef, i32 15, i1 0, i1 0, i1 0, i1 0, i1 0)
+  %elt0 = extractelement <4 x float> %r, i32 0
+  %elt1 = extractelement <4 x float> %r, i32 1
+  store volatile float %elt0, float addrspace(1)* %out
+  store volatile float %elt1, float addrspace(1)* %out
+  ret void
+}
+
+; GCN-LABEL: {{^}}adjust_writemask_sample_012:
+; GCN: image_sample v{{\[[0-9]+:[0-9]+\]}}, v{{\[[0-9]+:[0-9]+\]}}, s{{\[[0-9]+:[0-9]+\]}}, s{{\[[0-9]+:[0-9]+\]}} dmask:0x7{{$}}
+define void @adjust_writemask_sample_012(float addrspace(1)* %out) {
+main_body:
+  %r = call <4 x float> @llvm.amdgcn.image.sample.v4f32.v4f32.v8i32(<4 x float> undef, <8 x i32> undef, <4 x i32> undef, i32 15, i1 0, i1 0, i1 0, i1 0, i1 0)
+  %elt0 = extractelement <4 x float> %r, i32 0
+  %elt1 = extractelement <4 x float> %r, i32 1
+  %elt2 = extractelement <4 x float> %r, i32 2
+  store volatile float %elt0, float addrspace(1)* %out
+  store volatile float %elt1, float addrspace(1)* %out
+  store volatile float %elt2, float addrspace(1)* %out
+  ret void
+}
+
+; GCN-LABEL: {{^}}adjust_writemask_sample_12:
+; GCN: image_sample v{{\[[0-9]+:[0-9]+\]}}, v{{\[[0-9]+:[0-9]+\]}}, s{{\[[0-9]+:[0-9]+\]}}, s{{\[[0-9]+:[0-9]+\]}} dmask:0x6{{$}}
+define void @adjust_writemask_sample_12(float addrspace(1)* %out) {
+main_body:
+  %r = call <4 x float> @llvm.amdgcn.image.sample.v4f32.v4f32.v8i32(<4 x float> undef, <8 x i32> undef, <4 x i32> undef, i32 15, i1 0, i1 0, i1 0, i1 0, i1 0)
+  %elt1 = extractelement <4 x float> %r, i32 1
+  %elt2 = extractelement <4 x float> %r, i32 2
+  store volatile float %elt1, float addrspace(1)* %out
+  store volatile float %elt2, float addrspace(1)* %out
+  ret void
+}
+
+; GCN-LABEL: {{^}}adjust_writemask_sample_03:
+; GCN: image_sample v{{\[[0-9]+:[0-9]+\]}}, v{{\[[0-9]+:[0-9]+\]}}, s{{\[[0-9]+:[0-9]+\]}}, s{{\[[0-9]+:[0-9]+\]}} dmask:0x9{{$}}
+define void @adjust_writemask_sample_03(float addrspace(1)* %out) {
+main_body:
+  %r = call <4 x float> @llvm.amdgcn.image.sample.v4f32.v4f32.v8i32(<4 x float> undef, <8 x i32> undef, <4 x i32> undef, i32 15, i1 0, i1 0, i1 0, i1 0, i1 0)
+  %elt0 = extractelement <4 x float> %r, i32 0
+  %elt3 = extractelement <4 x float> %r, i32 3
+  store volatile float %elt0, float addrspace(1)* %out
+  store volatile float %elt3, float addrspace(1)* %out
+  ret void
+}
+
+; GCN-LABEL: {{^}}adjust_writemask_sample_13:
+; GCN: image_sample v{{\[[0-9]+:[0-9]+\]}}, v{{\[[0-9]+:[0-9]+\]}}, s{{\[[0-9]+:[0-9]+\]}}, s{{\[[0-9]+:[0-9]+\]}} dmask:0xa{{$}}
+define void @adjust_writemask_sample_13(float addrspace(1)* %out) {
+main_body:
+  %r = call <4 x float> @llvm.amdgcn.image.sample.v4f32.v4f32.v8i32(<4 x float> undef, <8 x i32> undef, <4 x i32> undef, i32 15, i1 0, i1 0, i1 0, i1 0, i1 0)
+  %elt1 = extractelement <4 x float> %r, i32 1
+  %elt3 = extractelement <4 x float> %r, i32 3
+  store volatile float %elt1, float addrspace(1)* %out
+  store volatile float %elt3, float addrspace(1)* %out
+  ret void
+}
+
+; GCN-LABEL: {{^}}adjust_writemask_sample_123:
+; GCN: image_sample v{{\[[0-9]+:[0-9]+\]}}, v{{\[[0-9]+:[0-9]+\]}}, s{{\[[0-9]+:[0-9]+\]}}, s{{\[[0-9]+:[0-9]+\]}} dmask:0xe{{$}}
+define void @adjust_writemask_sample_123(float addrspace(1)* %out) {
+main_body:
+  %r = call <4 x float> @llvm.amdgcn.image.sample.v4f32.v4f32.v8i32(<4 x float> undef, <8 x i32> undef, <4 x i32> undef, i32 15, i1 0, i1 0, i1 0, i1 0, i1 0)
+  %elt1 = extractelement <4 x float> %r, i32 1
+  %elt2 = extractelement <4 x float> %r, i32 2
+  %elt3 = extractelement <4 x float> %r, i32 3
+  store volatile float %elt1, float addrspace(1)* %out
+  store volatile float %elt2, float addrspace(1)* %out
+  store volatile float %elt3, float addrspace(1)* %out
+  ret void
+}
+
 declare <4 x float> @llvm.amdgcn.image.sample.v4f32.v4f32.v8i32(<4 x float>, <8 x i32>, <4 x i32>, i32, i1, i1, i1, i1, i1) #0
 declare <4 x float> @llvm.amdgcn.image.sample.cl.v4f32.v4f32.v8i32(<4 x float>, <8 x i32>, <4 x i32>, i32, i1, i1, i1, i1, i1) #0
 declare <4 x float> @llvm.amdgcn.image.sample.d.v4f32.v4f32.v8i32(<4 x float>, <8 x i32>, <4 x i32>, i32, i1, i1, i1, i1, i1) #0
 declare <4 x float> @llvm.amdgcn.image.sample.d.cl.v4f32.v4f32.v8i32(<4 x float>, <8 x i32>, <4 x i32>, i32, i1, i1, i1, i1, i1) #0
 declare <4 x float> @llvm.amdgcn.image.sample.l.v4f32.v4f32.v8i32(<4 x float>, <8 x i32>, <4 x i32>, i32, i1, i1, i1, i1, i1) #0
 declare <4 x float> @llvm.amdgcn.image.sample.b.v4f32.v4f32.v8i32(<4 x float>, <8 x i32>, <4 x i32>, i32, i1, i1, i1, i1, i1) #0
 declare <4 x float> @llvm.amdgcn.image.sample.b.cl.v4f32.v4f32.v8i32(<4 x float>, <8 x i32>, <4 x i32>, i32, i1, i1, i1, i1, i1) #0
 declare <4 x float> @llvm.amdgcn.image.sample.lz.v4f32.v4f32.v8i32(<4 x float>, <8 x i32>, <4 x i32>, i32, i1, i1, i1, i1, i1) #0