[X86] Add isel patterns to match BMI/TBMI instructions when lowering has turned the root nodes into one of the flag producing binops.

This fixes the patterns that have or/and as a root. 'and' is handled differently since thy usually have a CMP wrapped around them.

I had to look for uses of the CF flag because all these nodes have non-standard CF flag behavior. A real or/xor would always clear CF. In practice we shouldn't be using the CF flag from these nodes as far as I know.

Differential Revision: https://reviews.llvm.org/D55813

llvm-svn: 349962

2 files changed, 62 insertions, 2 deletions

diff --git a/llvm/lib/Target/X86/X86ISelDAGToDAG.cpp b/llvm/lib/Target/X86/X86ISelDAGToDAG.cpp
index 8115753904f..0274867459d 100644
--- a/llvm/lib/Target/X86/X86ISelDAGToDAG.cpp
+++ b/llvm/lib/Target/X86/X86ISelDAGToDAG.cpp
@@ -472,6 +472,9 @@ namespace {
                                 SDValue &InFlag);
 
     bool tryOptimizeRem8Extend(SDNode *N);
+
+    bool hasNoSignFlagUses(SDValue Flags) const;
+    bool hasNoCarryFlagUses(SDValue Flags) const;
   };
 }
 
@@ -2225,7 +2228,7 @@ static X86::CondCode getCondFromOpc(unsigned Opc) {
 
 /// Test whether the given X86ISD::CMP node has any uses which require the SF
 /// flag to be accurate.
-static bool hasNoSignFlagUses(SDValue Flags) {
+bool X86DAGToDAGISel::hasNoSignFlagUses(SDValue Flags) const {
   // Examine each user of the node.
   for (SDNode::use_iterator UI = Flags->use_begin(), UE = Flags->use_end();
          UI != UE; ++UI) {
@@ -2265,7 +2268,7 @@ static bool hasNoSignFlagUses(SDValue Flags) {
 
 /// Test whether the given node which sets flags has any uses which require the
 /// CF flag to be accurate.
-static bool hasNoCarryFlagUses(SDValue Flags) {
+ bool X86DAGToDAGISel::hasNoCarryFlagUses(SDValue Flags) const {
   // Examine each user of the node.
   for (SDNode::use_iterator UI = Flags->use_begin(), UE = Flags->use_end();
          UI != UE; ++UI) {
diff --git a/llvm/lib/Target/X86/X86InstrInfo.td b/llvm/lib/Target/X86/X86InstrInfo.td
index fe9b530e8ea..7a401d4fc35 100644
--- a/llvm/lib/Target/X86/X86InstrInfo.td
+++ b/llvm/lib/Target/X86/X86InstrInfo.td
@@ -2387,6 +2387,16 @@ let Predicates = [HasBMI], Defs = [EFLAGS] in {
 // Pattern fragments to auto generate BMI instructions.
 //===----------------------------------------------------------------------===//
 
+def or_flag_nocf : PatFrag<(ops node:$lhs, node:$rhs),
+                           (X86or_flag node:$lhs, node:$rhs), [{
+  return hasNoCarryFlagUses(SDValue(N, 1));
+}]>;
+
+def xor_flag_nocf : PatFrag<(ops node:$lhs, node:$rhs),
+                            (X86xor_flag node:$lhs, node:$rhs), [{
+  return hasNoCarryFlagUses(SDValue(N, 1));
+}]>;
+
 let Predicates = [HasBMI] in {
   // FIXME: patterns for the load versions are not implemented
   def : Pat<(and GR32:$src, (add GR32:$src, -1)),
@@ -2403,6 +2413,14 @@ let Predicates = [HasBMI] in {
             (BLSI32rr GR32:$src)>;
   def : Pat<(and GR64:$src, (ineg GR64:$src)),
             (BLSI64rr GR64:$src)>;
+
+  // Versions to match flag producing ops.
+  // X86and_flag nodes are rarely created. Those should use CMP+AND. We do
+  // TESTrr matching in PostProcessISelDAG to allow BLSR/BLSI to be formed.
+  def : Pat<(xor_flag_nocf GR32:$src, (add GR32:$src, -1)),
+            (BLSMSK32rr GR32:$src)>;
+  def : Pat<(xor_flag_nocf GR64:$src, (add GR64:$src, -1)),
+            (BLSMSK64rr GR64:$src)>;
 }
 
 multiclass bmi_bextr<bits<8> opc, string mnemonic, RegisterClass RC,
@@ -2801,6 +2819,45 @@ let Predicates = [HasTBM] in {
             (TZMSK32rr GR32:$src)>;
   def : Pat<(and (not GR64:$src), (add GR64:$src, -1)),
             (TZMSK64rr GR64:$src)>;
+
+  // Patterns to match flag producing ops.
+  // X86and_flag nodes are rarely created. Those should use CMP+AND. We do
+  // TESTrr matching in PostProcessISelDAG to allow BLSR/BLSI to be formed.
+  def : Pat<(or_flag_nocf GR32:$src, (not (add GR32:$src, 1))),
+            (BLCI32rr GR32:$src)>;
+  def : Pat<(or_flag_nocf GR64:$src, (not (add GR64:$src, 1))),
+            (BLCI64rr GR64:$src)>;
+
+  // Extra patterns because opt can optimize the above patterns to this.
+  def : Pat<(or_flag_nocf GR32:$src, (sub -2, GR32:$src)),
+            (BLCI32rr GR32:$src)>;
+  def : Pat<(or_flag_nocf GR64:$src, (sub -2, GR64:$src)),
+            (BLCI64rr GR64:$src)>;
+
+  def : Pat<(xor_flag_nocf GR32:$src, (add GR32:$src, 1)),
+            (BLCMSK32rr GR32:$src)>;
+  def : Pat<(xor_flag_nocf GR64:$src, (add GR64:$src, 1)),
+            (BLCMSK64rr GR64:$src)>;
+
+  def : Pat<(or_flag_nocf GR32:$src, (add GR32:$src, 1)),
+            (BLCS32rr GR32:$src)>;
+  def : Pat<(or_flag_nocf GR64:$src, (add GR64:$src, 1)),
+            (BLCS64rr GR64:$src)>;
+
+  def : Pat<(or_flag_nocf GR32:$src, (add GR32:$src, -1)),
+            (BLSFILL32rr GR32:$src)>;
+  def : Pat<(or_flag_nocf GR64:$src, (add GR64:$src, -1)),
+            (BLSFILL64rr GR64:$src)>;
+
+  def : Pat<(or_flag_nocf (not GR32:$src), (add GR32:$src, -1)),
+            (BLSIC32rr GR32:$src)>;
+  def : Pat<(or_flag_nocf (not GR64:$src), (add GR64:$src, -1)),
+            (BLSIC64rr GR64:$src)>;
+
+  def : Pat<(or_flag_nocf (not GR32:$src), (add GR32:$src, 1)),
+            (T1MSKC32rr GR32:$src)>;
+  def : Pat<(or_flag_nocf (not GR64:$src), (add GR64:$src, 1)),
+            (T1MSKC64rr GR64:$src)>;
 } // HasTBM
 
 //===----------------------------------------------------------------------===//