[DAG] Fold neg(splat(neg(x)) -> splat(x)

This add as a fold of sub(0, splat(sub(0, x))) -> splat(x). This can
come up in the lowering of right shifts under AArch64, where we generate
a shift left of a negated number.

Differential Revision: https://reviews.llvm.org/D103755

diff --git a/llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp b/llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp
index 02f3ea2..63c979c 100644 (file)
--- a/llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp
+++ b/llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp
@@ -3308,6 +3308,17 @@ SDValue DAGCombiner::visitSUB(SDNode *N) {
         !TLI.isOperationLegalOrCustom(ISD::ABS, VT) &&
         TLI.expandABS(N1.getNode(), Result, DAG, true))
       return Result;
+
+    // Fold neg(splat(neg(x)) -> splat(x)
+    if (VT.isVector()) {
+      SDValue N1S = DAG.getSplatValue(N1, true);
+      if (N1S && N1S.getOpcode() == ISD::SUB &&
+          isNullConstant(N1S.getOperand(0))) {
+        if (VT.isScalableVector())
+          return DAG.getSplatVector(VT, DL, N1S.getOperand(1));
+        return DAG.getSplatBuildVector(VT, DL, N1S.getOperand(1));
+      }
+    }
   }
 
   // Canonicalize (sub -1, x) -> ~x, i.e. (xor x, -1)

diff --git a/llvm/test/CodeGen/AArch64/neon-shift-neg.ll b/llvm/test/CodeGen/AArch64/neon-shift-neg.ll
index ecfa430..9ea0e96 100644 (file)
--- a/llvm/test/CodeGen/AArch64/neon-shift-neg.ll
+++ b/llvm/test/CodeGen/AArch64/neon-shift-neg.ll
@@ -4,9 +4,7 @@
 define <2 x i64> @shr64x2(<2 x i64> %a, i64 %b) {
 ; CHECK-LABEL: shr64x2:
 ; CHECK:       // %bb.0: // %entry
-; CHECK-NEXT:    neg x8, x0
-; CHECK-NEXT:    dup v1.2d, x8
-; CHECK-NEXT:    neg v1.2d, v1.2d
+; CHECK-NEXT:    dup v1.2d, x0
 ; CHECK-NEXT:    sshl v0.2d, v0.2d, v1.2d
 ; CHECK-NEXT:    ret
 entry:
@@ -20,9 +18,7 @@ entry:
 define <4 x i32> @shr32x4(<4 x i32> %a, i32 %b) {
 ; CHECK-LABEL: shr32x4:
 ; CHECK:       // %bb.0: // %entry
-; CHECK-NEXT:    neg w8, w0
-; CHECK-NEXT:    dup v1.4s, w8
-; CHECK-NEXT:    neg v1.4s, v1.4s
+; CHECK-NEXT:    dup v1.4s, w0
 ; CHECK-NEXT:    sshl v0.4s, v0.4s, v1.4s
 ; CHECK-NEXT:    ret
 entry:
@@ -36,9 +32,7 @@ entry:
 define <4 x i32> @shr32x4undef(<4 x i32> %a, i32 %b) {
 ; CHECK-LABEL: shr32x4undef:
 ; CHECK:       // %bb.0: // %entry
-; CHECK-NEXT:    neg w8, w0
-; CHECK-NEXT:    dup v1.4s, w8
-; CHECK-NEXT:    neg v1.4s, v1.4s
+; CHECK-NEXT:    dup v1.4s, w0
 ; CHECK-NEXT:    sshl v0.4s, v0.4s, v1.4s
 ; CHECK-NEXT:    ret
 entry:
@@ -52,9 +46,7 @@ entry:
 define <8 x i16> @shr16x8(<8 x i16> %a, i16 %b) {
 ; CHECK-LABEL: shr16x8:
 ; CHECK:       // %bb.0: // %entry
-; CHECK-NEXT:    neg w8, w0
-; CHECK-NEXT:    dup v1.8h, w8
-; CHECK-NEXT:    neg v1.8h, v1.8h
+; CHECK-NEXT:    dup v1.8h, w0
 ; CHECK-NEXT:    sshl v0.8h, v0.8h, v1.8h
 ; CHECK-NEXT:    ret
 entry:
@@ -68,9 +60,7 @@ entry:
 define <16 x i8> @shr8x16(<16 x i8> %a, i8 %b) {
 ; CHECK-LABEL: shr8x16:
 ; CHECK:       // %bb.0: // %entry
-; CHECK-NEXT:    neg w8, w0
-; CHECK-NEXT:    dup v1.16b, w8
-; CHECK-NEXT:    neg v1.16b, v1.16b
+; CHECK-NEXT:    dup v1.16b, w0
 ; CHECK-NEXT:    sshl v0.16b, v0.16b, v1.16b
 ; CHECK-NEXT:    ret
 entry:
@@ -84,9 +74,7 @@ entry:
 define <1 x i64> @shr64x1(<1 x i64> %a, i64 %b) {
 ; CHECK-LABEL: shr64x1:
 ; CHECK:       // %bb.0: // %entry
-; CHECK-NEXT:    neg x8, x0
-; CHECK-NEXT:    fmov d1, x8
-; CHECK-NEXT:    neg d1, d1
+; CHECK-NEXT:    fmov d1, x0
 ; CHECK-NEXT:    sshl d0, d0, d1
 ; CHECK-NEXT:    ret
 entry:
@@ -99,9 +87,7 @@ entry:
 define <2 x i32> @shr32x2(<2 x i32> %a, i32 %b) {
 ; CHECK-LABEL: shr32x2:
 ; CHECK:       // %bb.0: // %entry
-; CHECK-NEXT:    neg w8, w0
-; CHECK-NEXT:    dup v1.2s, w8
-; CHECK-NEXT:    neg v1.2s, v1.2s
+; CHECK-NEXT:    dup v1.2s, w0
 ; CHECK-NEXT:    sshl v0.2s, v0.2s, v1.2s
 ; CHECK-NEXT:    ret
 entry:
@@ -115,9 +101,7 @@ entry:
 define <4 x i16> @shr16x4(<4 x i16> %a, i16 %b) {
 ; CHECK-LABEL: shr16x4:
 ; CHECK:       // %bb.0: // %entry
-; CHECK-NEXT:    neg w8, w0
-; CHECK-NEXT:    dup v1.4h, w8
-; CHECK-NEXT:    neg v1.4h, v1.4h
+; CHECK-NEXT:    dup v1.4h, w0
 ; CHECK-NEXT:    sshl v0.4h, v0.4h, v1.4h
 ; CHECK-NEXT:    ret
 entry:
@@ -131,9 +115,7 @@ entry:
 define <8 x i8> @shr8x8(<8 x i8> %a, i8 %b) {
 ; CHECK-LABEL: shr8x8:
 ; CHECK:       // %bb.0: // %entry
-; CHECK-NEXT:    neg w8, w0
-; CHECK-NEXT:    dup v1.8b, w8
-; CHECK-NEXT:    neg v1.8b, v1.8b
+; CHECK-NEXT:    dup v1.8b, w0
 ; CHECK-NEXT:    sshl v0.8b, v0.8b, v1.8b
 ; CHECK-NEXT:    ret
 entry:
@@ -147,9 +129,7 @@ entry:
 define <2 x i64> @lshr64x2(<2 x i64> %a, i64 %b) {
 ; CHECK-LABEL: lshr64x2:
 ; CHECK:       // %bb.0: // %entry
-; CHECK-NEXT:    neg x8, x0
-; CHECK-NEXT:    dup v1.2d, x8
-; CHECK-NEXT:    neg v1.2d, v1.2d
+; CHECK-NEXT:    dup v1.2d, x0
 ; CHECK-NEXT:    ushl v0.2d, v0.2d, v1.2d
 ; CHECK-NEXT:    ret
 entry:
@@ -163,9 +143,7 @@ entry:
 define <4 x i32> @lshr32x4(<4 x i32> %a, i32 %b) {
 ; CHECK-LABEL: lshr32x4:
 ; CHECK:       // %bb.0: // %entry
-; CHECK-NEXT:    neg w8, w0
-; CHECK-NEXT:    dup v1.4s, w8
-; CHECK-NEXT:    neg v1.4s, v1.4s
+; CHECK-NEXT:    dup v1.4s, w0
 ; CHECK-NEXT:    ushl v0.4s, v0.4s, v1.4s
 ; CHECK-NEXT:    ret
 entry:
@@ -179,9 +157,7 @@ entry:
 define <4 x i32> @lshr32x4undef(<4 x i32> %a, i32 %b) {
 ; CHECK-LABEL: lshr32x4undef:
 ; CHECK:       // %bb.0: // %entry
-; CHECK-NEXT:    neg w8, w0
-; CHECK-NEXT:    dup v1.4s, w8
-; CHECK-NEXT:    neg v1.4s, v1.4s
+; CHECK-NEXT:    dup v1.4s, w0
 ; CHECK-NEXT:    ushl v0.4s, v0.4s, v1.4s
 ; CHECK-NEXT:    ret
 entry:
@@ -195,9 +171,7 @@ entry:
 define <8 x i16> @lshr16x8(<8 x i16> %a, i16 %b) {
 ; CHECK-LABEL: lshr16x8:
 ; CHECK:       // %bb.0: // %entry
-; CHECK-NEXT:    neg w8, w0
-; CHECK-NEXT:    dup v1.8h, w8
-; CHECK-NEXT:    neg v1.8h, v1.8h
+; CHECK-NEXT:    dup v1.8h, w0
 ; CHECK-NEXT:    ushl v0.8h, v0.8h, v1.8h
 ; CHECK-NEXT:    ret
 entry:
@@ -211,9 +185,7 @@ entry:
 define <16 x i8> @lshr8x16(<16 x i8> %a, i8 %b) {
 ; CHECK-LABEL: lshr8x16:
 ; CHECK:       // %bb.0: // %entry
-; CHECK-NEXT:    neg w8, w0
-; CHECK-NEXT:    dup v1.16b, w8
-; CHECK-NEXT:    neg v1.16b, v1.16b
+; CHECK-NEXT:    dup v1.16b, w0
 ; CHECK-NEXT:    ushl v0.16b, v0.16b, v1.16b
 ; CHECK-NEXT:    ret
 entry:
@@ -227,9 +199,7 @@ entry:
 define <1 x i64> @lshr64x1(<1 x i64> %a, i64 %b) {
 ; CHECK-LABEL: lshr64x1:
 ; CHECK:       // %bb.0: // %entry
-; CHECK-NEXT:    neg x8, x0
-; CHECK-NEXT:    fmov d1, x8
-; CHECK-NEXT:    neg d1, d1
+; CHECK-NEXT:    fmov d1, x0
 ; CHECK-NEXT:    ushl d0, d0, d1
 ; CHECK-NEXT:    ret
 entry:
@@ -242,9 +212,7 @@ entry:
 define <2 x i32> @lshr32x2(<2 x i32> %a, i32 %b) {
 ; CHECK-LABEL: lshr32x2:
 ; CHECK:       // %bb.0: // %entry
-; CHECK-NEXT:    neg w8, w0
-; CHECK-NEXT:    dup v1.2s, w8
-; CHECK-NEXT:    neg v1.2s, v1.2s
+; CHECK-NEXT:    dup v1.2s, w0
 ; CHECK-NEXT:    ushl v0.2s, v0.2s, v1.2s
 ; CHECK-NEXT:    ret
 entry:
@@ -258,9 +226,7 @@ entry:
 define <4 x i16> @lshr16x4(<4 x i16> %a, i16 %b) {
 ; CHECK-LABEL: lshr16x4:
 ; CHECK:       // %bb.0: // %entry
-; CHECK-NEXT:    neg w8, w0
-; CHECK-NEXT:    dup v1.4h, w8
-; CHECK-NEXT:    neg v1.4h, v1.4h
+; CHECK-NEXT:    dup v1.4h, w0
 ; CHECK-NEXT:    ushl v0.4h, v0.4h, v1.4h
 ; CHECK-NEXT:    ret
 entry:
@@ -274,9 +240,7 @@ entry:
 define <8 x i8> @lshr8x8(<8 x i8> %a, i8 %b) {
 ; CHECK-LABEL: lshr8x8:
 ; CHECK:       // %bb.0: // %entry
-; CHECK-NEXT:    neg w8, w0
-; CHECK-NEXT:    dup v1.8b, w8
-; CHECK-NEXT:    neg v1.8b, v1.8b
+; CHECK-NEXT:    dup v1.8b, w0
 ; CHECK-NEXT:    ushl v0.8b, v0.8b, v1.8b
 ; CHECK-NEXT:    ret
 entry:
@@ -603,9 +567,7 @@ entry:
 define <vscale x 16 x i8> @subsub(<vscale x 16 x i8> %a, i8 %b) {
 ; CHECK-LABEL: subsub:
 ; CHECK:       // %bb.0: // %entry
-; CHECK-NEXT:    neg w8, w0
-; CHECK-NEXT:    mov z0.b, w8
-; CHECK-NEXT:    subr z0.b, z0.b, #0 // =0x0
+; CHECK-NEXT:    mov z0.b, w0
 ; CHECK-NEXT:    ret
 entry:
   %sub = sub i8 0, %b

diff --git a/llvm/test/CodeGen/AArch64/sub-splat-sub.ll b/llvm/test/CodeGen/AArch64/sub-splat-sub.ll
index 8a5c616..5402259 100644 (file)
--- a/llvm/test/CodeGen/AArch64/sub-splat-sub.ll
+++ b/llvm/test/CodeGen/AArch64/sub-splat-sub.ll
@@ -4,9 +4,7 @@
 define <16 x i8> @subsubii8(<16 x i8> %a, i8 %b) {
 ; CHECK-LABEL: subsubii8:
 ; CHECK:       // %bb.0: // %entry
-; CHECK-NEXT:    neg w8, w0
-; CHECK-NEXT:    dup v0.16b, w8
-; CHECK-NEXT:    neg v0.16b, v0.16b
+; CHECK-NEXT:    dup v0.16b, w0
 ; CHECK-NEXT:    ret
 entry:
   %sub = sub i8 0, %b
@@ -19,9 +17,7 @@ entry:
 define <vscale x 16 x i8> @subsubni8(<vscale x 16 x i8> %a, i8 %b) {
 ; CHECK-LABEL: subsubni8:
 ; CHECK:       // %bb.0: // %entry
-; CHECK-NEXT:    neg w8, w0
-; CHECK-NEXT:    mov z0.b, w8
-; CHECK-NEXT:    subr z0.b, z0.b, #0 // =0x0
+; CHECK-NEXT:    mov z0.b, w0
 ; CHECK-NEXT:    ret
 entry:
   %sub = sub i8 0, %b