[PATCH][AVX512] Fix TERNLOG when folding broadcast

[Cameron McInally via llvm-commits]

Hey guys,

Here's a little patch to fix ternlog instructions with a folded
broadcast. The broadcast decorator, e.g. {1toX}, was missing.

Thanks,
Cam
-------------- next part --------------
Index: lib/Target/X86/X86InstrAVX512.td
===================================================================
--- lib/Target/X86/X86InstrAVX512.td	(revision 315065)
+++ lib/Target/X86/X86InstrAVX512.td	(working copy)
@@ -9399,13 +9399,13 @@
                    (OpNode (X86VBroadcast (_.ScalarLdFrag addr:$src3)),
                     _.RC:$src2, _.RC:$src1, (i8 imm:$src4)),
                    _.RC:$src1)),
-            (!cast<Instruction>(NAME#_.ZSuffix#rmik) _.RC:$src1, _.KRCWM:$mask,
+            (!cast<Instruction>(NAME#_.ZSuffix#rmbik) _.RC:$src1, _.KRCWM:$mask,
              _.RC:$src2, addr:$src3, (VPTERNLOG321_imm8 imm:$src4))>;
   def : Pat<(_.VT (vselect _.KRCWM:$mask,
                    (OpNode _.RC:$src2, _.RC:$src1,
                     (X86VBroadcast (_.ScalarLdFrag addr:$src3)),
                     (i8 imm:$src4)), _.RC:$src1)),
-            (!cast<Instruction>(NAME#_.ZSuffix#rmik) _.RC:$src1, _.KRCWM:$mask,
+            (!cast<Instruction>(NAME#_.ZSuffix#rmbik) _.RC:$src1, _.KRCWM:$mask,
              _.RC:$src2, addr:$src3, (VPTERNLOG213_imm8 imm:$src4))>;
   def : Pat<(_.VT (vselect _.KRCWM:$mask,
                    (OpNode _.RC:$src2,
@@ -9412,13 +9412,13 @@
                     (X86VBroadcast (_.ScalarLdFrag addr:$src3)),
                     _.RC:$src1, (i8 imm:$src4)),
                    _.RC:$src1)),
-            (!cast<Instruction>(NAME#_.ZSuffix#rmik) _.RC:$src1, _.KRCWM:$mask,
+            (!cast<Instruction>(NAME#_.ZSuffix#rmbik) _.RC:$src1, _.KRCWM:$mask,
              _.RC:$src2, addr:$src3, (VPTERNLOG231_imm8 imm:$src4))>;
   def : Pat<(_.VT (vselect _.KRCWM:$mask,
                    (OpNode (X86VBroadcast (_.ScalarLdFrag addr:$src3)),
                     _.RC:$src1, _.RC:$src2, (i8 imm:$src4)),
                    _.RC:$src1)),
-            (!cast<Instruction>(NAME#_.ZSuffix#rmik) _.RC:$src1, _.KRCWM:$mask,
+            (!cast<Instruction>(NAME#_.ZSuffix#rmbik) _.RC:$src1, _.KRCWM:$mask,
              _.RC:$src2, addr:$src3, (VPTERNLOG312_imm8 imm:$src4))>;
 }
 
Index: test/CodeGen/X86/avx512-vpternlog-commute.ll
===================================================================
--- test/CodeGen/X86/avx512-vpternlog-commute.ll	(revision 315065)
+++ test/CodeGen/X86/avx512-vpternlog-commute.ll	(working copy)
@@ -1008,7 +1008,7 @@
 ; CHECK-LABEL: vpternlog_v16i32_012_broadcast0_mask1:
 ; CHECK:       ## BB#0:
 ; CHECK-NEXT:    kmovd %esi, %k1
-; CHECK-NEXT:    vpternlogd $92, (%rdi), %zmm1, %zmm0 {%k1}
+; CHECK-NEXT:    vpternlogd $92, (%rdi){1to16}, %zmm1, %zmm0 {%k1}
 ; CHECK-NEXT:    retq
   %x0scalar = load i32, i32* %x0ptr
   %vecinit.i = insertelement <16 x i32> undef, i32 %x0scalar, i32 0
@@ -1023,7 +1023,7 @@
 ; CHECK-LABEL: vpternlog_v16i32_012_broadcast0_mask2:
 ; CHECK:       ## BB#0:
 ; CHECK-NEXT:    kmovd %esi, %k1
-; CHECK-NEXT:    vpternlogd $58, (%rdi), %zmm0, %zmm1 {%k1}
+; CHECK-NEXT:    vpternlogd $58, (%rdi){1to16}, %zmm0, %zmm1 {%k1}
 ; CHECK-NEXT:    vmovdqa64 %zmm1, %zmm0
 ; CHECK-NEXT:    retq
   %x0scalar = load i32, i32* %x0ptr
@@ -1039,7 +1039,7 @@
 ; CHECK-LABEL: vpternlog_v16i32_012_broadcast1_mask2:
 ; CHECK:       ## BB#0:
 ; CHECK-NEXT:    kmovd %esi, %k1
-; CHECK-NEXT:    vpternlogd $46, (%rdi), %zmm0, %zmm1 {%k1}
+; CHECK-NEXT:    vpternlogd $46, (%rdi){1to16}, %zmm0, %zmm1 {%k1}
 ; CHECK-NEXT:    vmovdqa64 %zmm1, %zmm0
 ; CHECK-NEXT:    retq
   %x1scalar = load i32, i32* %x1ptr
@@ -1055,7 +1055,7 @@
 ; CHECK-LABEL: vpternlog_v16i32_012_broadcast2_mask1:
 ; CHECK:       ## BB#0:
 ; CHECK-NEXT:    kmovd %esi, %k1
-; CHECK-NEXT:    vpternlogd $78, (%rdi), %zmm0, %zmm1 {%k1}
+; CHECK-NEXT:    vpternlogd $78, (%rdi){1to16}, %zmm0, %zmm1 {%k1}
 ; CHECK-NEXT:    vmovdqa64 %zmm1, %zmm0
 ; CHECK-NEXT:    retq
   %x2scalar = load i32, i32* %x2ptr