AMDGPU/SI: Fix a GPU hang with POS_W_FLOAT enabled

[• tstellarAMD]

LGTM.

llvm/trunk/lib/Target/AMDGPU/SIISelLowering.cpp

SDValue SITargetLowering::LowerFormalArguments(
 
   // At least one interpolation mode must be enabled or else the GPU will hang.
   //
   // Check PSInputAddr instead of PSInputEna. The idea is that if the user set
   // PSInputAddr, the user wants to enable some bits after the compilation
   // based on run-time states. Since we can't know what the final PSInputEna
   // will look like, so we shouldn't do anything here and the user should take
   // responsibility for the correct programming.
+  //
+  // Otherwise, the following restrictions apply:
+  // - At least one of PERSP_* (0xF) or LINEAR_* (0x70) must be enabled.
+  // - If POS_W_FLOAT (11) is enabled, at least one of PERSP_* must be
+  //   enabled too.
   if (Info->getShaderType() == ShaderType::PIXEL &&
-      (Info->getPSInputAddr() & 0x7F) == 0) {
+      ((Info->getPSInputAddr() & 0x7F) == 0 ||
+       ((Info->getPSInputAddr() & 0xF) == 0 &&
+  Info->isPSInputAllocated(11)))) {
     CCInfo.AllocateReg(AMDGPU::VGPR0);
     CCInfo.AllocateReg(AMDGPU::VGPR1);
     Info->markPSInputAllocated(0);
     Info->PSInputEna |= 1;
   }
 
   if (Info->getShaderType() == ShaderType::COMPUTE) {
     getOriginalFunctionArgs(DAG, DAG.getMachineFunction().getFunction(), Ins,

llvm/trunk/test/CodeGen/AMDGPU/ret.ll

 ; GCN: v_mov_b32_e32 v0, 1.0
 ; GCN-NOT: s_endpgm
 define float @ps_input_ena_no_inputs([9 x <16 x i8>] addrspace(2)* byval, i32 inreg, i32 inreg, <2 x i32>, <2 x i32>, <2 x i32>, <3 x i32>, <2 x i32>, <2 x i32>, <2 x i32>, float, float, float, float, float, float, float, float, float) #1 {
   ret float 1.0
 }
 
 
 ; GCN: .long 165580
+; GCN-NEXT: .long 2081
+; GCN-NEXT: .long 165584
+; GCN-NEXT: .long 2081
+; GCN-LABEL: {{^}}ps_input_ena_pos_w:
+; GCN-DAG: v_mov_b32_e32 v0, v4
+; GCN-DAG: v_mov_b32_e32 v1, v2
+; GCN: v_mov_b32_e32 v2, v3
+; GCN-NOT: s_endpgm
+define {float, <2 x float>} @ps_input_ena_pos_w([9 x <16 x i8>] addrspace(2)* byval, i32 inreg, i32 inreg, <2 x i32>, <2 x i32>, <2 x i32>, <3 x i32>, <2 x i32>, <2 x i32>, <2 x i32>, float, float, float, float, float, float, float, float, float) #1 {
+  %f = bitcast <2 x i32> %8 to <2 x float>
+  %s = insertvalue {float, <2 x float>} undef, float %14, 0
+  %s1 = insertvalue {float, <2 x float>} %s, <2 x float> %f, 1
+  ret {float, <2 x float>} %s1
+}
+
+
+; GCN: .long 165580
 ; GCN-NEXT: .long 562
 ; GCN-NEXT: .long 165584
 ; GCN-NEXT: .long 563
 ; GCN-LABEL: {{^}}vgpr_ps_addr1:
 ; GCN-DAG: v_mov_b32_e32 v0, v2
 ; GCN-DAG: v_mov_b32_e32 v1, v3
 ; GCN: v_mov_b32_e32 v2, v4
 ; GCN-DAG: v_mov_b32_e32 v3, v6