Vendor import of llvm RELEASE_34/dot1-final tag r208032 (effectively, 3.4.1 release):vendor/llvm/llvm-release_34-r208032

https://llvm.org/svn/llvm-project/llvm/tags/RELEASE_34/dot1-final@208032

12 files changed, 244 insertions, 16 deletions

diff --git a/test/CodeGen/R600/bfe_uint.ll b/test/CodeGen/R600/bfe_uint.ll
index 92570c315299..fe466e6ad5fd 100644
--- a/test/CodeGen/R600/bfe_uint.ll
+++ b/test/CodeGen/R600/bfe_uint.ll
@@ -1,5 +1,7 @@
 ; RUN: llc < %s -march=r600 -mcpu=redwood | FileCheck %s
 
+; XFAIL: *
+
 ; CHECK: @bfe_def
 ; CHECK: BFE_UINT
 define void @bfe_def(i32 addrspace(1)* %out, i32 %x) {
diff --git a/test/CodeGen/R600/fabs.ll b/test/CodeGen/R600/fabs.ll
index a5f5df96b5d9..2cd3a4f604f2 100644
--- a/test/CodeGen/R600/fabs.ll
+++ b/test/CodeGen/R600/fabs.ll
@@ -9,7 +9,7 @@
 ; R600-CHECK-NOT: AND
 ; R600-CHECK: |PV.{{[XYZW]}}|
 ; SI-CHECK-LABEL: @fabs_free
-; SI-CHECK: V_ADD_F32_e64 v{{[0-9]}}, s{{[0-9]}}, 0, 1, 0, 0, 0
+; SI-CHECK: V_AND_B32
 
 define void @fabs_free(float addrspace(1)* %out, i32 %in) {
 entry:
@@ -23,8 +23,8 @@ entry:
 ; R600-CHECK: |{{(PV|T[0-9])\.[XYZW]}}|
 ; R600-CHECK: |{{(PV|T[0-9])\.[XYZW]}}|
 ; SI-CHECK-LABEL: @fabs_v2
-; SI-CHECK: V_ADD_F32_e64 v{{[0-9]}}, s{{[0-9]}}, 0, 1, 0, 0, 0
-; SI-CHECK: V_ADD_F32_e64 v{{[0-9]}}, s{{[0-9]}}, 0, 1, 0, 0, 0
+; SI-CHECK: V_AND_B32
+; SI-CHECK: V_AND_B32
 define void @fabs_v2(<2 x float> addrspace(1)* %out, <2 x float> %in) {
 entry:
   %0 = call <2 x float> @llvm.fabs.v2f32(<2 x float> %in)
@@ -38,10 +38,10 @@ entry:
 ; R600-CHECK: |{{(PV|T[0-9])\.[XYZW]}}|
 ; R600-CHECK: |{{(PV|T[0-9])\.[XYZW]}}|
 ; SI-CHECK-LABEL: @fabs_v4
-; SI-CHECK: V_ADD_F32_e64 v{{[0-9]}}, s{{[0-9]}}, 0, 1, 0, 0, 0
-; SI-CHECK: V_ADD_F32_e64 v{{[0-9]}}, s{{[0-9]}}, 0, 1, 0, 0, 0
-; SI-CHECK: V_ADD_F32_e64 v{{[0-9]}}, s{{[0-9]}}, 0, 1, 0, 0, 0
-; SI-CHECK: V_ADD_F32_e64 v{{[0-9]}}, s{{[0-9]}}, 0, 1, 0, 0, 0
+; SI-CHECK: V_AND_B32
+; SI-CHECK: V_AND_B32
+; SI-CHECK: V_AND_B32
+; SI-CHECK: V_AND_B32
 define void @fabs_v4(<4 x float> addrspace(1)* %out, <4 x float> %in) {
 entry:
   %0 = call <4 x float> @llvm.fabs.v4f32(<4 x float> %in)
diff --git a/test/CodeGen/R600/fneg-fabs.ll b/test/CodeGen/R600/fneg-fabs.ll
new file mode 100644
index 000000000000..d95e1311bc10
--- /dev/null
+++ b/test/CodeGen/R600/fneg-fabs.ll
@@ -0,0 +1,55 @@
+; RUN: llc < %s -march=r600 -mcpu=redwood | FileCheck %s --check-prefix=R600-CHECK
+; RUN: llc < %s -march=r600 -mcpu=SI -verify-machineinstrs | FileCheck %s --check-prefix=SI-CHECK
+
+; DAGCombiner will transform:
+; (fabs (f32 bitcast (i32 a))) => (f32 bitcast (and (i32 a), 0x7FFFFFFF))
+; unless isFabsFree returns true
+
+; R600-CHECK-LABEL: @fneg_fabs_free
+; R600-CHECK-NOT: AND
+; R600-CHECK: |PV.{{[XYZW]}}|
+; R600-CHECK: -PV
+; SI-CHECK-LABEL: @fneg_fabs_free
+; SI-CHECK: V_OR_B32
+
+define void @fneg_fabs_free(float addrspace(1)* %out, i32 %in) {
+entry:
+  %0 = bitcast i32 %in to float
+  %1 = call float @fabs(float %0)
+  %2 = fsub float -0.000000e+00, %1
+  store float %2, float addrspace(1)* %out
+  ret void
+}
+
+; R600-CHECK-LABEL: @fneg_fabs_v2
+; R600-CHECK: |{{(PV|T[0-9])\.[XYZW]}}|
+; R600-CHECK: -PV
+; R600-CHECK: |{{(PV|T[0-9])\.[XYZW]}}|
+; R600-CHECK: -PV
+; SI-CHECK-LABEL: @fneg_fabs_v2
+; SI-CHECK: V_OR_B32
+; SI-CHECK: V_OR_B32
+define void @fneg_fabs_v2(<2 x float> addrspace(1)* %out, <2 x float> %in) {
+entry:
+  %0 = call <2 x float> @llvm.fabs.v2f32(<2 x float> %in)
+  %1 = fsub <2 x float> <float -0.000000e+00, float -0.000000e+00>, %0
+  store <2 x float> %1, <2 x float> addrspace(1)* %out
+  ret void
+}
+
+; SI-CHECK-LABEL: @fneg_fabs_v4
+; SI-CHECK: V_OR_B32
+; SI-CHECK: V_OR_B32
+; SI-CHECK: V_OR_B32
+; SI-CHECK: V_OR_B32
+define void @fneg_fabs_v4(<4 x float> addrspace(1)* %out, <4 x float> %in) {
+entry:
+  %0 = call <4 x float> @llvm.fabs.v4f32(<4 x float> %in)
+  %1 = fsub <4 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %0
+  store <4 x float> %1, <4 x float> addrspace(1)* %out
+  ret void
+}
+
+declare float @fabs(float ) readnone
+declare <2 x float> @llvm.fabs.v2f32(<2 x float> ) readnone
+declare <4 x float> @llvm.fabs.v4f32(<4 x float> ) readnone
diff --git a/test/CodeGen/R600/fneg.ll b/test/CodeGen/R600/fneg.ll
index 9446aa8ea9c3..f4e6be62467a 100644
--- a/test/CodeGen/R600/fneg.ll
+++ b/test/CodeGen/R600/fneg.ll
@@ -4,7 +4,7 @@
 ; R600-CHECK-LABEL: @fneg
 ; R600-CHECK: -PV
 ; SI-CHECK-LABEL: @fneg
-; SI-CHECK: V_ADD_F32_e64 v{{[0-9]}}, s{{[0-9]}}, 0, 0, 0, 0, 1
+; SI-CHECK: V_XOR_B32
 define void @fneg(float addrspace(1)* %out, float %in) {
 entry:
   %0 = fsub float -0.000000e+00, %in
@@ -16,8 +16,8 @@ entry:
 ; R600-CHECK: -PV
 ; R600-CHECK: -PV
 ; SI-CHECK-LABEL: @fneg_v2
-; SI-CHECK: V_ADD_F32_e64 v{{[0-9]}}, s{{[0-9]}}, 0, 0, 0, 0, 1
-; SI-CHECK: V_ADD_F32_e64 v{{[0-9]}}, s{{[0-9]}}, 0, 0, 0, 0, 1
+; SI-CHECK: V_XOR_B32
+; SI-CHECK: V_XOR_B32
 define void @fneg_v2(<2 x float> addrspace(1)* nocapture %out, <2 x float> %in) {
 entry:
   %0 = fsub <2 x float> <float -0.000000e+00, float -0.000000e+00>, %in
@@ -31,10 +31,10 @@ entry:
 ; R600-CHECK: -PV
 ; R600-CHECK: -PV
 ; SI-CHECK-LABEL: @fneg_v4
-; SI-CHECK: V_ADD_F32_e64 v{{[0-9]}}, s{{[0-9]}}, 0, 0, 0, 0, 1
-; SI-CHECK: V_ADD_F32_e64 v{{[0-9]}}, s{{[0-9]}}, 0, 0, 0, 0, 1
-; SI-CHECK: V_ADD_F32_e64 v{{[0-9]}}, s{{[0-9]}}, 0, 0, 0, 0, 1
-; SI-CHECK: V_ADD_F32_e64 v{{[0-9]}}, s{{[0-9]}}, 0, 0, 0, 0, 1
+; SI-CHECK: V_XOR_B32
+; SI-CHECK: V_XOR_B32
+; SI-CHECK: V_XOR_B32
+; SI-CHECK: V_XOR_B32
 define void @fneg_v4(<4 x float> addrspace(1)* nocapture %out, <4 x float> %in) {
 entry:
   %0 = fsub <4 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %in
diff --git a/test/CodeGen/R600/lds-oqap-crash.ll b/test/CodeGen/R600/lds-oqap-crash.ll
new file mode 100644
index 000000000000..79591506e8cb
--- /dev/null
+++ b/test/CodeGen/R600/lds-oqap-crash.ll
@@ -0,0 +1,28 @@
+; RUN: llc < %s -march=r600 -mcpu=redwood -verify-machineinstrs | FileCheck %s
+
+; The test is for a bug in R600EmitClauseMarkers.cpp where this pass
+; was searching for a use of the OQAP register in order to determine
+; if an LDS instruction could fit in the current clause, but never finding
+; one.  This created an infinite loop and hung the compiler.
+;
+; The LDS instruction should not have been defining OQAP in the first place,
+; because the LDS instructions are pseudo instructions and the OQAP
+; reads and writes are bundled together in the same instruction.
+
+; CHECK: @lds_crash
+define void @lds_crash(i32 addrspace(1)* %out, i32 addrspace(3)* %in, i32 %a, i32 %b, i32 %c) {
+entry:
+  %0 = load i32 addrspace(3)* %in
+  ; This block needs to be > 115 ISA instructions to hit the bug,
+  ; so we'll use udiv instructions.
+  %div0 = udiv i32 %0, %b
+  %div1 = udiv i32 %div0, %a
+  %div2 = udiv i32 %div1, 11
+  %div3 = udiv i32 %div2, %a
+  %div4 = udiv i32 %div3, %b
+  %div5 = udiv i32 %div4, %c
+  %div6 = udiv i32 %div5, %div0
+  %div7 = udiv i32 %div6, %div1
+  store i32 %div7, i32 addrspace(1)* %out
+  ret void
+}
diff --git a/test/CodeGen/R600/llvm.AMDGPU.kill.ll b/test/CodeGen/R600/llvm.AMDGPU.kill.ll
new file mode 100644
index 000000000000..bec5cdf65f1b
--- /dev/null
+++ b/test/CodeGen/R600/llvm.AMDGPU.kill.ll
@@ -0,0 +1,18 @@
+; RUN: llc < %s -march=r600 -mcpu=verde -verify-machineinstrs | FileCheck --check-prefix=SI %s
+
+; SI-LABEL: @kill_gs
+; SI: V_CMPX_LE_F32
+
+define void @kill_gs() #0 {
+main_body:
+  %0 = icmp ule i32 0, 3
+  %1 = select i1 %0, float 1.000000e+00, float -1.000000e+00
+  call void @llvm.AMDGPU.kill(float %1)
+  ret void
+}
+
+declare void @llvm.AMDGPU.kill(float)
+
+attributes #0 = { "ShaderType"="2" }
+
+!0 = metadata !{metadata !"const", null, i32 1}
diff --git a/test/CodeGen/R600/llvm.SI.load.dword.ll b/test/CodeGen/R600/llvm.SI.load.dword.ll
new file mode 100644
index 000000000000..a6227755b72e
--- /dev/null
+++ b/test/CodeGen/R600/llvm.SI.load.dword.ll
@@ -0,0 +1,40 @@
+;RUN: llc < %s -march=r600 -mcpu=verde -verify-machineinstrs | FileCheck %s
+
+; Example of a simple geometry shader loading vertex attributes from the
+; ESGS ring buffer
+
+; CHECK-LABEL: @main
+; CHECK: BUFFER_LOAD_DWORD
+; CHECK: BUFFER_LOAD_DWORD
+; CHECK: BUFFER_LOAD_DWORD
+; CHECK: BUFFER_LOAD_DWORD
+
+define void @main([17 x <16 x i8>] addrspace(2)* byval, [32 x <16 x i8>] addrspace(2)* byval, [16 x <32 x i8>] addrspace(2)* byval, [2 x <16 x i8>] addrspace(2)* byval, [17 x <16 x i8>] addrspace(2)* inreg, [17 x <16 x i8>] addrspace(2)* inreg, i32, i32, i32, i32) #0 {
+main_body:
+  %10 = getelementptr [2 x <16 x i8>] addrspace(2)* %3, i64 0, i32 1
+  %11 = load <16 x i8> addrspace(2)* %10, !tbaa !0
+  %12 = shl i32 %6, 2
+  %13 = call i32 @llvm.SI.buffer.load.dword.i32.i32(<16 x i8> %11, i32 0, i32 0, i32 0, i32 0, i32 0, i32 1, i32 1, i32 0)
+  %14 = bitcast i32 %13 to float
+  %15 = call i32 @llvm.SI.buffer.load.dword.i32.i32(<16 x i8> %11, i32 %12, i32 0, i32 0, i32 1, i32 0, i32 1, i32 1, i32 0)
+  %16 = bitcast i32 %15 to float
+  %17 = call i32 @llvm.SI.buffer.load.dword.i32.i32(<16 x i8> %11, i32 %12, i32 0, i32 0, i32 0, i32 1, i32 1, i32 1, i32 0)
+  %18 = bitcast i32 %17 to float
+  %19 = call i32 @llvm.SI.buffer.load.dword.i32.v2i32(<16 x i8> %11, <2 x i32> <i32 0, i32 0>, i32 0, i32 0, i32 1, i32 1, i32 1, i32 1, i32 0)
+  %20 = bitcast i32 %19 to float
+  call void @llvm.SI.export(i32 15, i32 0, i32 1, i32 12, i32 0, float %14, float %16, float %18, float %20)
+  ret void
+}
+
+; Function Attrs: nounwind readonly
+declare i32 @llvm.SI.buffer.load.dword.i32.i32(<16 x i8>, i32, i32, i32, i32, i32, i32, i32, i32) #1
+
+; Function Attrs: nounwind readonly
+declare i32 @llvm.SI.buffer.load.dword.i32.v2i32(<16 x i8>, <2 x i32>, i32, i32, i32, i32, i32, i32, i32) #1
+
+declare void @llvm.SI.export(i32, i32, i32, i32, i32, float, float, float, float)
+
+attributes #0 = { "ShaderType"="1" }
+attributes #1 = { nounwind readonly }
+
+!0 = metadata !{metadata !"const", null, i32 1}
diff --git a/test/CodeGen/R600/llvm.SI.sendmsg.ll b/test/CodeGen/R600/llvm.SI.sendmsg.ll
new file mode 100644
index 000000000000..cfcc7c4e40ee
--- /dev/null
+++ b/test/CodeGen/R600/llvm.SI.sendmsg.ll
@@ -0,0 +1,21 @@
+;RUN: llc < %s -march=r600 -mcpu=verde -verify-machineinstrs | FileCheck %s
+
+; CHECK-LABEL: @main
+; CHECK: S_SENDMSG 34
+; CHECK: S_SENDMSG 274
+; CHECK: S_SENDMSG 562
+; CHECK: S_SENDMSG 3
+
+define void @main() {
+main_body:
+  call void @llvm.SI.sendmsg(i32 34, i32 0);
+  call void @llvm.SI.sendmsg(i32 274, i32 0);
+  call void @llvm.SI.sendmsg(i32 562, i32 0);
+  call void @llvm.SI.sendmsg(i32 3, i32 0);
+  ret void
+}
+
+; Function Attrs: nounwind
+declare void @llvm.SI.sendmsg(i32, i32) #0
+
+attributes #0 = { nounwind }
diff --git a/test/CodeGen/R600/load.ll b/test/CodeGen/R600/load.ll
index e4492d7d6e7b..0153524d136c 100644
--- a/test/CodeGen/R600/load.ll
+++ b/test/CodeGen/R600/load.ll
@@ -445,6 +445,7 @@ define void @load_const_addrspace_f32(float addrspace(1)* %out, float addrspace(
 ; R600-CHECK: LDS_UBYTE_READ_RET
 ; SI-CHECK-LABEL: @load_i8_local
 ; SI-CHECK-NOT: S_WQM_B64
+; SI-CHECK: S_MOV_B32 m0
 ; SI-CHECK: DS_READ_U8
 define void @load_i8_local(i32 addrspace(1)* %out, i8 addrspace(3)* %in) {
   %1 = load i8 addrspace(3)* %in
@@ -458,6 +459,7 @@ define void @load_i8_local(i32 addrspace(1)* %out, i8 addrspace(3)* %in) {
 ; R600-CHECK: ASHR
 ; SI-CHECK-LABEL: @load_i8_sext_local
 ; SI-CHECK-NOT: S_WQM_B64
+; SI-CHECK: S_MOV_B32 m0
 ; SI-CHECK: DS_READ_I8
 define void @load_i8_sext_local(i32 addrspace(1)* %out, i8 addrspace(3)* %in) {
 entry:
@@ -472,6 +474,7 @@ entry:
 ; R600-CHECK: LDS_UBYTE_READ_RET
 ; SI-CHECK-LABEL: @load_v2i8_local
 ; SI-CHECK-NOT: S_WQM_B64
+; SI-CHECK: S_MOV_B32 m0
 ; SI-CHECK: DS_READ_U8
 ; SI-CHECK: DS_READ_U8
 define void @load_v2i8_local(<2 x i32> addrspace(1)* %out, <2 x i8> addrspace(3)* %in) {
@@ -489,6 +492,7 @@ entry:
 ; R600-CHECK-DAG: ASHR
 ; SI-CHECK-LABEL: @load_v2i8_sext_local
 ; SI-CHECK-NOT: S_WQM_B64
+; SI-CHECK: S_MOV_B32 m0
 ; SI-CHECK: DS_READ_I8
 ; SI-CHECK: DS_READ_I8
 define void @load_v2i8_sext_local(<2 x i32> addrspace(1)* %out, <2 x i8> addrspace(3)* %in) {
@@ -506,6 +510,7 @@ entry:
 ; R600-CHECK: LDS_UBYTE_READ_RET
 ; SI-CHECK-LABEL: @load_v4i8_local
 ; SI-CHECK-NOT: S_WQM_B64
+; SI-CHECK: S_MOV_B32 m0
 ; SI-CHECK: DS_READ_U8
 ; SI-CHECK: DS_READ_U8
 ; SI-CHECK: DS_READ_U8
@@ -529,6 +534,7 @@ entry:
 ; R600-CHECK-DAG: ASHR
 ; SI-CHECK-LABEL: @load_v4i8_sext_local
 ; SI-CHECK-NOT: S_WQM_B64
+; SI-CHECK: S_MOV_B32 m0
 ; SI-CHECK: DS_READ_I8
 ; SI-CHECK: DS_READ_I8
 ; SI-CHECK: DS_READ_I8
@@ -546,6 +552,7 @@ entry:
 ; R600-CHECK: LDS_USHORT_READ_RET
 ; SI-CHECK-LABEL: @load_i16_local
 ; SI-CHECK-NOT: S_WQM_B64
+; SI-CHECK: S_MOV_B32 m0
 ; SI-CHECK: DS_READ_U16
 define void @load_i16_local(i32 addrspace(1)* %out, i16 addrspace(3)* %in) {
 entry:
@@ -560,6 +567,7 @@ entry:
 ; R600-CHECK: ASHR
 ; SI-CHECK-LABEL: @load_i16_sext_local
 ; SI-CHECK-NOT: S_WQM_B64
+; SI-CHECK: S_MOV_B32 m0
 ; SI-CHECK: DS_READ_I16
 define void @load_i16_sext_local(i32 addrspace(1)* %out, i16 addrspace(3)* %in) {
 entry:
@@ -574,6 +582,7 @@ entry:
 ; R600-CHECK: LDS_USHORT_READ_RET
 ; SI-CHECK-LABEL: @load_v2i16_local
 ; SI-CHECK-NOT: S_WQM_B64
+; SI-CHECK: S_MOV_B32 m0
 ; SI-CHECK: DS_READ_U16
 ; SI-CHECK: DS_READ_U16
 define void @load_v2i16_local(<2 x i32> addrspace(1)* %out, <2 x i16> addrspace(3)* %in) {
@@ -591,6 +600,7 @@ entry:
 ; R600-CHECK-DAG: ASHR
 ; SI-CHECK-LABEL: @load_v2i16_sext_local
 ; SI-CHECK-NOT: S_WQM_B64
+; SI-CHECK: S_MOV_B32 m0
 ; SI-CHECK: DS_READ_I16
 ; SI-CHECK: DS_READ_I16
 define void @load_v2i16_sext_local(<2 x i32> addrspace(1)* %out, <2 x i16> addrspace(3)* %in) {
@@ -608,6 +618,7 @@ entry:
 ; R600-CHECK: LDS_USHORT_READ_RET
 ; SI-CHECK-LABEL: @load_v4i16_local
 ; SI-CHECK-NOT: S_WQM_B64
+; SI-CHECK: S_MOV_B32 m0
 ; SI-CHECK: DS_READ_U16
 ; SI-CHECK: DS_READ_U16
 ; SI-CHECK: DS_READ_U16
@@ -631,6 +642,7 @@ entry:
 ; R600-CHECK-DAG: ASHR
 ; SI-CHECK-LABEL: @load_v4i16_sext_local
 ; SI-CHECK-NOT: S_WQM_B64
+; SI-CHECK: S_MOV_B32 m0
 ; SI-CHECK: DS_READ_I16
 ; SI-CHECK: DS_READ_I16
 ; SI-CHECK: DS_READ_I16
@@ -643,11 +655,12 @@ entry:
   ret void
 }
 
-; load an i32 value from the glocal address space.
+; load an i32 value from the local address space.
 ; R600-CHECK-LABEL: @load_i32_local
 ; R600-CHECK: LDS_READ_RET
 ; SI-CHECK-LABEL: @load_i32_local
 ; SI-CHECK-NOT: S_WQM_B64
+; SI-CHECK: S_MOV_B32 m0
 ; SI-CHECK: DS_READ_B32
 define void @load_i32_local(i32 addrspace(1)* %out, i32 addrspace(3)* %in) {
 entry:
@@ -656,10 +669,11 @@ entry:
   ret void
 }
 
-; load a f32 value from the global address space.
+; load a f32 value from the local address space.
 ; R600-CHECK-LABEL: @load_f32_local
 ; R600-CHECK: LDS_READ_RET
 ; SI-CHECK-LABEL: @load_f32_local
+; SI-CHECK: S_MOV_B32 m0
 ; SI-CHECK: DS_READ_B32
 define void @load_f32_local(float addrspace(1)* %out, float addrspace(3)* %in) {
 entry:
@@ -673,6 +687,7 @@ entry:
 ; R600-CHECK: LDS_READ_RET
 ; R600-CHECK: LDS_READ_RET
 ; SI-CHECK-LABEL: @load_v2f32_local
+; SI-CHECK: S_MOV_B32 m0
 ; SI-CHECK: DS_READ_B32
 ; SI-CHECK: DS_READ_B32
 define void @load_v2f32_local(<2 x float> addrspace(1)* %out, <2 x float> addrspace(3)* %in) {
diff --git a/test/CodeGen/R600/trunc.ll b/test/CodeGen/R600/trunc.ll
index 0bd320ad9ceb..6bbd7f7b510e 100644
--- a/test/CodeGen/R600/trunc.ll
+++ b/test/CodeGen/R600/trunc.ll
@@ -28,3 +28,13 @@ define void @trunc_shl_i64(i32 addrspace(1)* %out, i64 %a) {
   store i32 %result, i32 addrspace(1)* %out, align 4
   ret void
 }
+
+; SI-LABEL: @trunc_i32_to_i1:
+; SI: V_AND_B32
+; SI: V_CMP_EQ_I32
+define void @trunc_i32_to_i1(i32 addrspace(1)* %out, i32 %a) {
+  %trunc = trunc i32 %a to i1
+  %result = select i1 %trunc, i32 1, i32 0
+  store i32 %result, i32 addrspace(1)* %out, align 4
+  ret void
+}
diff --git a/test/CodeGen/R600/vtx-fetch-branch.ll b/test/CodeGen/R600/vtx-fetch-branch.ll
new file mode 100644
index 000000000000..0fc99dee0dbe
--- /dev/null
+++ b/test/CodeGen/R600/vtx-fetch-branch.ll
@@ -0,0 +1,29 @@
+; RUN: llc -march=r600 -mcpu=redwood %s -o - | FileCheck %s
+
+; This tests for a bug where vertex fetch clauses right before an ENDIF
+; instruction where being emitted after the ENDIF.  We were using ALU_POP_AFTER
+; for the ALU clause before the vetex fetch instead of emitting a POP instruction
+; after the fetch clause.
+
+
+; CHECK-LABEL: @test
+; CHECK-NOT: ALU_POP_AFTER
+; CHECK: TEX
+; CHECK-NEXT: POP
+define void @test(i32 addrspace(1)* %out, i32 addrspace(1)* %in, i32 %cond) {
+entry:
+  %0 = icmp eq i32 %cond, 0
+  br i1 %0, label %endif, label %if
+
+if:
+  %1 = load i32 addrspace(1)* %in
+  br label %endif
+
+endif:
+  %x = phi i32 [ %1, %if], [ 0, %entry]
+  store i32 %x, i32 addrspace(1)* %out
+  br label %done
+
+done:
+  ret void
+}
diff --git a/test/CodeGen/R600/zero_extend.ll b/test/CodeGen/R600/zero_extend.ll
index 481b3b328259..a114bfc4a02b 100644
--- a/test/CodeGen/R600/zero_extend.ll
+++ b/test/CodeGen/R600/zero_extend.ll
@@ -16,3 +16,13 @@ entry:
   store i64 %2, i64 addrspace(1)* %out
   ret void
 }
+
+; SI-CHECK-LABEL: @testi1toi32
+; SI-CHECK: V_CNDMASK_B32
+define void @testi1toi32(i32 addrspace(1)* %out, i32 %a, i32 %b) {
+entry:
+  %0 = icmp eq i32 %a, %b
+  %1 = zext i1 %0 to i32
+  store i32 %1, i32 addrspace(1)* %out
+  ret void
+}