[PATCH] AMDGPU/SI: Fix the V_FRACT_F64 SI bug workaround

[Marek Olšák]

From: Marek Olšák <marek.olsak at amd.com>

This is a candidate for 3.7.
---
 lib/Target/AMDGPU/SIInstructions.td          |  4 ++--
 test/CodeGen/AMDGPU/llvm.AMDGPU.fract.f64.ll | 12 ++++++------
 2 files changed, 8 insertions(+), 8 deletions(-)

diff --git a/lib/Target/AMDGPU/SIInstructions.td b/lib/Target/AMDGPU/SIInstructions.td
index de8675e..b2edc03 100644
--- a/lib/Target/AMDGPU/SIInstructions.td
+++ b/lib/Target/AMDGPU/SIInstructions.td
@@ -3273,13 +3273,13 @@ def : Pat <
   (f64 (fadd (f64 (VOP3Mods f64:$x, i32:$mods)),
              (f64 (fneg (f64 (ffloor (f64 (VOP3Mods f64:$x, i32:$mods)))))))),
   (V_CNDMASK_B64_PSEUDO
-      $x,
       (V_MIN_F64
           SRCMODS.NONE,
           (V_FRACT_F64_e64 $mods, $x, DSTCLAMP.NONE, DSTOMOD.NONE),
           SRCMODS.NONE,
           (V_MOV_B64_PSEUDO 0x3fefffffffffffff),
           DSTCLAMP.NONE, DSTOMOD.NONE),
+      $x,
       (V_CMP_CLASS_F64_e64 SRCMODS.NONE, $x, 3/*NaN*/))
 >;
 
@@ -3291,13 +3291,13 @@ def : Pat <
       $x,
       SRCMODS.NEG,
       (V_CNDMASK_B64_PSEUDO
-         $x,
          (V_MIN_F64
              SRCMODS.NONE,
              (V_FRACT_F64_e64 $mods, $x, DSTCLAMP.NONE, DSTOMOD.NONE),
              SRCMODS.NONE,
              (V_MOV_B64_PSEUDO 0x3fefffffffffffff),
              DSTCLAMP.NONE, DSTOMOD.NONE),
+         $x,
          (V_CMP_CLASS_F64_e64 SRCMODS.NONE, $x, 3/*NaN*/)),
       DSTCLAMP.NONE, DSTOMOD.NONE)
 >;
diff --git a/test/CodeGen/AMDGPU/llvm.AMDGPU.fract.f64.ll b/test/CodeGen/AMDGPU/llvm.AMDGPU.fract.f64.ll
index e098dd3..6049dca 100644
--- a/test/CodeGen/AMDGPU/llvm.AMDGPU.fract.f64.ll
+++ b/test/CodeGen/AMDGPU/llvm.AMDGPU.fract.f64.ll
@@ -11,8 +11,8 @@ declare double @llvm.AMDGPU.fract.f64(double) nounwind readnone
 ; SI: v_mov_b32_e32 v[[UPHI:[0-9]+]], 0x3fefffff
 ; SI: v_min_f64 v{{\[}}[[MINLO:[0-9]+]]:[[MINHI:[0-9]+]]], v{{\[}}[[UPLO]]:[[UPHI]]], [[FRC]]
 ; SI: v_cmp_class_f64_e64 [[COND:s\[[0-9]+:[0-9]+\]]], v{{\[}}[[LO]]:[[HI]]], 3
-; SI: v_cndmask_b32_e64 v[[RESLO:[0-9]+]], v[[LO]], v[[MINLO]], [[COND]]
-; SI: v_cndmask_b32_e64 v[[RESHI:[0-9]+]], v[[HI]], v[[MINHI]], [[COND]]
+; SI: v_cndmask_b32_e64 v[[RESLO:[0-9]+]], v[[MINLO]], v[[LO]], [[COND]]
+; SI: v_cndmask_b32_e64 v[[RESHI:[0-9]+]], v[[MINHI]], v[[HI]], [[COND]]
 ; SI: buffer_store_dwordx2 v{{\[}}[[RESLO]]:[[RESHI]]]
 ; CI: buffer_store_dwordx2 [[FRC]]
 define void @fract_f64(double addrspace(1)* %out, double addrspace(1)* %src) nounwind {
@@ -28,8 +28,8 @@ define void @fract_f64(double addrspace(1)* %out, double addrspace(1)* %src) nou
 ; SI: v_mov_b32_e32 v[[UPHI:[0-9]+]], 0x3fefffff
 ; SI: v_min_f64 v{{\[}}[[MINLO:[0-9]+]]:[[MINHI:[0-9]+]]], v{{\[}}[[UPLO]]:[[UPHI]]], [[FRC]]
 ; SI: v_cmp_class_f64_e64 [[COND:s\[[0-9]+:[0-9]+\]]], v{{\[}}[[LO]]:[[HI]]], 3
-; SI: v_cndmask_b32_e64 v[[RESLO:[0-9]+]], v[[LO]], v[[MINLO]], [[COND]]
-; SI: v_cndmask_b32_e64 v[[RESHI:[0-9]+]], v[[HI]], v[[MINHI]], [[COND]]
+; SI: v_cndmask_b32_e64 v[[RESLO:[0-9]+]], v[[MINLO]], v[[LO]], [[COND]]
+; SI: v_cndmask_b32_e64 v[[RESHI:[0-9]+]], v[[MINHI]], v[[HI]], [[COND]]
 ; SI: buffer_store_dwordx2 v{{\[}}[[RESLO]]:[[RESHI]]]
 ; CI: buffer_store_dwordx2 [[FRC]]
 define void @fract_f64_neg(double addrspace(1)* %out, double addrspace(1)* %src) nounwind {
@@ -46,8 +46,8 @@ define void @fract_f64_neg(double addrspace(1)* %out, double addrspace(1)* %src)
 ; SI: v_mov_b32_e32 v[[UPHI:[0-9]+]], 0x3fefffff
 ; SI: v_min_f64 v{{\[}}[[MINLO:[0-9]+]]:[[MINHI:[0-9]+]]], v{{\[}}[[UPLO]]:[[UPHI]]], [[FRC]]
 ; SI: v_cmp_class_f64_e64 [[COND:s\[[0-9]+:[0-9]+\]]], v{{\[}}[[LO]]:[[HI]]], 3
-; SI: v_cndmask_b32_e64 v[[RESLO:[0-9]+]], v[[LO]], v[[MINLO]], [[COND]]
-; SI: v_cndmask_b32_e64 v[[RESHI:[0-9]+]], v[[HI]], v[[MINHI]], [[COND]]
+; SI: v_cndmask_b32_e64 v[[RESLO:[0-9]+]], v[[MINLO]], v[[LO]], [[COND]]
+; SI: v_cndmask_b32_e64 v[[RESHI:[0-9]+]], v[[MINHI]], v[[HI]], [[COND]]
 ; SI: buffer_store_dwordx2 v{{\[}}[[RESLO]]:[[RESHI]]]
 ; CI: buffer_store_dwordx2 [[FRC]]
 define void @fract_f64_neg_abs(double addrspace(1)* %out, double addrspace(1)* %src) nounwind {
-- 
2.1.4