Add hasSideEffects=0 to some forms of ROUND, RCP, and RSQRT.

[oota-llvm.git] / lib / Target / X86 / X86InstrInfo.td

diff --git a/lib/Target/X86/X86InstrInfo.td b/lib/Target/X86/X86InstrInfo.td
index 650fa95d7f23c779a0895b6c604914097a2f01f8..d0d236ab429e43ca7f3becf318a648d3862b550b 100644 (file)
--- a/lib/Target/X86/X86InstrInfo.td
+++ b/lib/Target/X86/X86InstrInfo.td
@@ -247,9 +247,9 @@ def X86and_flag  : SDNode<"X86ISD::AND",  SDTBinaryArithWithFlags,
                           [SDNPCommutative]>;
 def X86andn_flag : SDNode<"X86ISD::ANDN", SDTBinaryArithWithFlags>;
 
-def X86blsi_flag : SDNode<"X86ISD::BLSI",  SDTUnaryArithWithFlags>;
-def X86blsmsk_flag : SDNode<"X86ISD::BLSMSK",  SDTUnaryArithWithFlags>;
-def X86blsr_flag : SDNode<"X86ISD::BLSR",  SDTUnaryArithWithFlags>;
+def X86blsi   : SDNode<"X86ISD::BLSI",   SDTIntUnaryOp>;
+def X86blsmsk : SDNode<"X86ISD::BLSMSK", SDTIntUnaryOp>;
+def X86blsr   : SDNode<"X86ISD::BLSR",   SDTIntUnaryOp>;
 
 def X86mul_imm : SDNode<"X86ISD::MUL_IMM", SDTIntBinOp>;
 
@@ -594,9 +594,9 @@ def FPStackf32   : Predicate<"!Subtarget->hasSSE1()">;
 def FPStackf64   : Predicate<"!Subtarget->hasSSE2()">;
 def HasCmpxchg16b: Predicate<"Subtarget->hasCmpxchg16b()">;
 def In32BitMode  : Predicate<"!Subtarget->is64Bit()">,
-                             AssemblerPredicate<"!Mode64Bit">;
+                             AssemblerPredicate<"!Mode64Bit", "32-bit mode">;
 def In64BitMode  : Predicate<"Subtarget->is64Bit()">,
-                             AssemblerPredicate<"Mode64Bit">;
+                             AssemblerPredicate<"Mode64Bit", "64-bit mode">;
 def IsWin64      : Predicate<"Subtarget->isTargetWin64()">;
 def IsNaCl       : Predicate<"Subtarget->isTargetNaCl()">;
 def NotNaCl      : Predicate<"!Subtarget->isTargetNaCl()">;
@@ -1043,7 +1043,7 @@ def MOV64ao64 : RIi32<0xA3, RawFrm, (outs offset64:$dst), (ins),
 */
 
 
-let isCodeGenOnly = 1 in {
+let isCodeGenOnly = 1, hasSideEffects = 0 in {
 def MOV8rr_REV : I<0x8A, MRMSrcReg, (outs GR8:$dst), (ins GR8:$src),
                    "mov{b}\t{$src, $dst|$dst, $src}", [], IIC_MOV>;
 def MOV16rr_REV : I<0x8B, MRMSrcReg, (outs GR16:$dst), (ins GR16:$src),
@@ -1133,24 +1133,26 @@ def BT64rr : RI<0xA3, MRMDestReg, (outs), (ins GR64:$src1, GR64:$src2),
 // perspective, this is pretty bizarre. Make these instructions disassembly
 // only for now.
 
-def BT16mr : I<0xA3, MRMDestMem, (outs), (ins i16mem:$src1, GR16:$src2),
-               "bt{w}\t{$src2, $src1|$src1, $src2}",
-//               [(X86bt (loadi16 addr:$src1), GR16:$src2),
-//                (implicit EFLAGS)]
-               [], IIC_BT_MR
-               >, OpSize, TB, Requires<[FastBTMem]>;
-def BT32mr : I<0xA3, MRMDestMem, (outs), (ins i32mem:$src1, GR32:$src2),
-               "bt{l}\t{$src2, $src1|$src1, $src2}",
-//               [(X86bt (loadi32 addr:$src1), GR32:$src2),
-//                (implicit EFLAGS)]
-               [], IIC_BT_MR
-               >, TB, Requires<[FastBTMem]>;
-def BT64mr : RI<0xA3, MRMDestMem, (outs), (ins i64mem:$src1, GR64:$src2),
-               "bt{q}\t{$src2, $src1|$src1, $src2}",
-//               [(X86bt (loadi64 addr:$src1), GR64:$src2),
-//                (implicit EFLAGS)]
-                [], IIC_BT_MR
-                >, TB;
+let mayLoad = 1, hasSideEffects = 0 in {
+  def BT16mr : I<0xA3, MRMDestMem, (outs), (ins i16mem:$src1, GR16:$src2),
+                 "bt{w}\t{$src2, $src1|$src1, $src2}",
+  //               [(X86bt (loadi16 addr:$src1), GR16:$src2),
+  //                (implicit EFLAGS)]
+                 [], IIC_BT_MR
+                 >, OpSize, TB, Requires<[FastBTMem]>;
+  def BT32mr : I<0xA3, MRMDestMem, (outs), (ins i32mem:$src1, GR32:$src2),
+                 "bt{l}\t{$src2, $src1|$src1, $src2}",
+  //               [(X86bt (loadi32 addr:$src1), GR32:$src2),
+  //                (implicit EFLAGS)]
+                 [], IIC_BT_MR
+                 >, TB, Requires<[FastBTMem]>;
+  def BT64mr : RI<0xA3, MRMDestMem, (outs), (ins i64mem:$src1, GR64:$src2),
+                 "bt{q}\t{$src2, $src1|$src1, $src2}",
+  //               [(X86bt (loadi64 addr:$src1), GR64:$src2),
+  //                (implicit EFLAGS)]
+                  [], IIC_BT_MR
+                  >, TB;
+}
 
 def BT16ri8 : Ii8<0xBA, MRM4r, (outs), (ins GR16:$src1, i16i8imm:$src2),
                 "bt{w}\t{$src2, $src1|$src1, $src2}",
@@ -1181,7 +1183,7 @@ def BT64mi8 : RIi8<0xBA, MRM4m, (outs), (ins i64mem:$src1, i64i8imm:$src2),
                 [(set EFLAGS, (X86bt (loadi64 addr:$src1),
                                      i64immSExt8:$src2))], IIC_BT_MI>, TB;
 
-
+let hasSideEffects = 0 in {
 def BTC16rr : I<0xBB, MRMDestReg, (outs), (ins GR16:$src1, GR16:$src2),
                 "btc{w}\t{$src2, $src1|$src1, $src2}", [], IIC_BTX_RR>,
                 OpSize, TB;
@@ -1189,6 +1191,8 @@ def BTC32rr : I<0xBB, MRMDestReg, (outs), (ins GR32:$src1, GR32:$src2),
                 "btc{l}\t{$src2, $src1|$src1, $src2}", [], IIC_BTX_RR>, TB;
 def BTC64rr : RI<0xBB, MRMDestReg, (outs), (ins GR64:$src1, GR64:$src2),
                  "btc{q}\t{$src2, $src1|$src1, $src2}", [], IIC_BTX_RR>, TB;
+
+let mayLoad = 1, mayStore = 1 in {
 def BTC16mr : I<0xBB, MRMDestMem, (outs), (ins i16mem:$src1, GR16:$src2),
                 "btc{w}\t{$src2, $src1|$src1, $src2}", [], IIC_BTX_MR>,
                 OpSize, TB;
@@ -1196,6 +1200,8 @@ def BTC32mr : I<0xBB, MRMDestMem, (outs), (ins i32mem:$src1, GR32:$src2),
                 "btc{l}\t{$src2, $src1|$src1, $src2}", [], IIC_BTX_MR>, TB;
 def BTC64mr : RI<0xBB, MRMDestMem, (outs), (ins i64mem:$src1, GR64:$src2),
                  "btc{q}\t{$src2, $src1|$src1, $src2}", [], IIC_BTX_MR>, TB;
+}
+
 def BTC16ri8 : Ii8<0xBA, MRM7r, (outs), (ins GR16:$src1, i16i8imm:$src2),
                     "btc{w}\t{$src2, $src1|$src1, $src2}", [], IIC_BTX_RI>,
                     OpSize, TB;
@@ -1203,6 +1209,8 @@ def BTC32ri8 : Ii8<0xBA, MRM7r, (outs), (ins GR32:$src1, i32i8imm:$src2),
                     "btc{l}\t{$src2, $src1|$src1, $src2}", [], IIC_BTX_RI>, TB;
 def BTC64ri8 : RIi8<0xBA, MRM7r, (outs), (ins GR64:$src1, i64i8imm:$src2),
                     "btc{q}\t{$src2, $src1|$src1, $src2}", [], IIC_BTX_RI>, TB;
+
+let mayLoad = 1, mayStore = 1 in {
 def BTC16mi8 : Ii8<0xBA, MRM7m, (outs), (ins i16mem:$src1, i16i8imm:$src2),
                     "btc{w}\t{$src2, $src1|$src1, $src2}", [], IIC_BTX_MI>,
                     OpSize, TB;
@@ -1210,6 +1218,7 @@ def BTC32mi8 : Ii8<0xBA, MRM7m, (outs), (ins i32mem:$src1, i32i8imm:$src2),
                     "btc{l}\t{$src2, $src1|$src1, $src2}", [], IIC_BTX_MI>, TB;
 def BTC64mi8 : RIi8<0xBA, MRM7m, (outs), (ins i64mem:$src1, i64i8imm:$src2),
                     "btc{q}\t{$src2, $src1|$src1, $src2}", [], IIC_BTX_MI>, TB;
+}
 
 def BTR16rr : I<0xB3, MRMDestReg, (outs), (ins GR16:$src1, GR16:$src2),
                 "btr{w}\t{$src2, $src1|$src1, $src2}", [], IIC_BTX_RR>,
@@ -1218,6 +1227,8 @@ def BTR32rr : I<0xB3, MRMDestReg, (outs), (ins GR32:$src1, GR32:$src2),
                 "btr{l}\t{$src2, $src1|$src1, $src2}", [], IIC_BTX_RR>, TB;
 def BTR64rr : RI<0xB3, MRMDestReg, (outs), (ins GR64:$src1, GR64:$src2),
                  "btr{q}\t{$src2, $src1|$src1, $src2}", []>, TB;
+
+let mayLoad = 1, mayStore = 1 in {
 def BTR16mr : I<0xB3, MRMDestMem, (outs), (ins i16mem:$src1, GR16:$src2),
                 "btr{w}\t{$src2, $src1|$src1, $src2}", [], IIC_BTX_MR>,
                 OpSize, TB;
@@ -1225,6 +1236,8 @@ def BTR32mr : I<0xB3, MRMDestMem, (outs), (ins i32mem:$src1, GR32:$src2),
                 "btr{l}\t{$src2, $src1|$src1, $src2}", [], IIC_BTX_MR>, TB;
 def BTR64mr : RI<0xB3, MRMDestMem, (outs), (ins i64mem:$src1, GR64:$src2),
                  "btr{q}\t{$src2, $src1|$src1, $src2}", [], IIC_BTX_MR>, TB;
+}
+
 def BTR16ri8 : Ii8<0xBA, MRM6r, (outs), (ins GR16:$src1, i16i8imm:$src2),
                     "btr{w}\t{$src2, $src1|$src1, $src2}", [], IIC_BTX_RI>,
                     OpSize, TB;
@@ -1232,6 +1245,8 @@ def BTR32ri8 : Ii8<0xBA, MRM6r, (outs), (ins GR32:$src1, i32i8imm:$src2),
                     "btr{l}\t{$src2, $src1|$src1, $src2}", [], IIC_BTX_RI>, TB;
 def BTR64ri8 : RIi8<0xBA, MRM6r, (outs), (ins GR64:$src1, i64i8imm:$src2),
                     "btr{q}\t{$src2, $src1|$src1, $src2}", [], IIC_BTX_RI>, TB;
+
+let mayLoad = 1, mayStore = 1 in {
 def BTR16mi8 : Ii8<0xBA, MRM6m, (outs), (ins i16mem:$src1, i16i8imm:$src2),
                     "btr{w}\t{$src2, $src1|$src1, $src2}", [], IIC_BTX_MI>,
                     OpSize, TB;
@@ -1239,6 +1254,7 @@ def BTR32mi8 : Ii8<0xBA, MRM6m, (outs), (ins i32mem:$src1, i32i8imm:$src2),
                     "btr{l}\t{$src2, $src1|$src1, $src2}", [], IIC_BTX_MI>, TB;
 def BTR64mi8 : RIi8<0xBA, MRM6m, (outs), (ins i64mem:$src1, i64i8imm:$src2),
                     "btr{q}\t{$src2, $src1|$src1, $src2}", [], IIC_BTX_MI>, TB;
+}
 
 def BTS16rr : I<0xAB, MRMDestReg, (outs), (ins GR16:$src1, GR16:$src2),
                 "bts{w}\t{$src2, $src1|$src1, $src2}", [], IIC_BTX_RR>,
@@ -1247,6 +1263,8 @@ def BTS32rr : I<0xAB, MRMDestReg, (outs), (ins GR32:$src1, GR32:$src2),
                 "bts{l}\t{$src2, $src1|$src1, $src2}", [], IIC_BTX_RR>, TB;
 def BTS64rr : RI<0xAB, MRMDestReg, (outs), (ins GR64:$src1, GR64:$src2),
                  "bts{q}\t{$src2, $src1|$src1, $src2}", [], IIC_BTX_RR>, TB;
+
+let mayLoad = 1, mayStore = 1 in {
 def BTS16mr : I<0xAB, MRMDestMem, (outs), (ins i16mem:$src1, GR16:$src2),
                 "bts{w}\t{$src2, $src1|$src1, $src2}", [], IIC_BTX_MR>,
                 OpSize, TB;
@@ -1254,6 +1272,8 @@ def BTS32mr : I<0xAB, MRMDestMem, (outs), (ins i32mem:$src1, GR32:$src2),
                 "bts{l}\t{$src2, $src1|$src1, $src2}", [], IIC_BTX_MR>, TB;
 def BTS64mr : RI<0xAB, MRMDestMem, (outs), (ins i64mem:$src1, GR64:$src2),
                  "bts{q}\t{$src2, $src1|$src1, $src2}", [], IIC_BTX_MR>, TB;
+}
+
 def BTS16ri8 : Ii8<0xBA, MRM5r, (outs), (ins GR16:$src1, i16i8imm:$src2),
                     "bts{w}\t{$src2, $src1|$src1, $src2}", [], IIC_BTX_RI>,
                     OpSize, TB;
@@ -1261,6 +1281,8 @@ def BTS32ri8 : Ii8<0xBA, MRM5r, (outs), (ins GR32:$src1, i32i8imm:$src2),
                     "bts{l}\t{$src2, $src1|$src1, $src2}", [], IIC_BTX_RI>, TB;
 def BTS64ri8 : RIi8<0xBA, MRM5r, (outs), (ins GR64:$src1, i64i8imm:$src2),
                     "bts{q}\t{$src2, $src1|$src1, $src2}", [], IIC_BTX_RI>, TB;
+
+let mayLoad = 1, mayStore = 1 in {
 def BTS16mi8 : Ii8<0xBA, MRM5m, (outs), (ins i16mem:$src1, i16i8imm:$src2),
                     "bts{w}\t{$src2, $src1|$src1, $src2}", [], IIC_BTX_MI>,
                     OpSize, TB;
@@ -1268,6 +1290,8 @@ def BTS32mi8 : Ii8<0xBA, MRM5m, (outs), (ins i32mem:$src1, i32i8imm:$src2),
                     "bts{l}\t{$src2, $src1|$src1, $src2}", [], IIC_BTX_MI>, TB;
 def BTS64mi8 : RIi8<0xBA, MRM5m, (outs), (ins i64mem:$src1, i64i8imm:$src2),
                     "bts{q}\t{$src2, $src1|$src1, $src2}", [], IIC_BTX_MI>, TB;
+}
+} // hasSideEffects = 0
 } // Defs = [EFLAGS]
 
 
@@ -1486,10 +1510,10 @@ def BOUNDS32rm : I<0x62, MRMSrcMem, (outs GR32:$dst), (ins i32mem:$src),
                    Requires<[In32BitMode]>;
 
 // Adjust RPL Field of Segment Selector
-def ARPL16rr : I<0x63, MRMDestReg, (outs GR16:$src), (ins GR16:$dst),
+def ARPL16rr : I<0x63, MRMDestReg, (outs GR16:$dst), (ins GR16:$src),
                  "arpl\t{$src, $dst|$dst, $src}", [], IIC_ARPL_REG>,
                  Requires<[In32BitMode]>;
-def ARPL16mr : I<0x63, MRMSrcMem, (outs GR16:$src), (ins i16mem:$dst),
+def ARPL16mr : I<0x63, MRMDestMem, (outs), (ins i16mem:$dst, GR16:$src),
                  "arpl\t{$src, $dst|$dst, $src}", [], IIC_ARPL_MEM>,
                  Requires<[In32BitMode]>;
 
@@ -1605,26 +1629,26 @@ multiclass bmi_bls<string mnemonic, Format RegMRM, Format MemMRM,
                   PatFrag ld_frag> {
   def rr : I<0xF3, RegMRM, (outs RC:$dst), (ins RC:$src),
              !strconcat(mnemonic, "\t{$src, $dst|$dst, $src}"),
-             [(set RC:$dst, EFLAGS, (OpNode RC:$src))]>, T8, VEX_4V;
+             [(set RC:$dst, (OpNode RC:$src)), (implicit EFLAGS)]>, T8, VEX_4V;
   def rm : I<0xF3, MemMRM, (outs RC:$dst), (ins x86memop:$src),
              !strconcat(mnemonic, "\t{$src, $dst|$dst, $src}"),
-             [(set RC:$dst, EFLAGS, (OpNode (ld_frag addr:$src)))]>,
+             [(set RC:$dst, (OpNode (ld_frag addr:$src))), (implicit EFLAGS)]>,
              T8, VEX_4V;
 }
 
 let Predicates = [HasBMI], Defs = [EFLAGS] in {
   defm BLSR32 : bmi_bls<"blsr{l}", MRM1r, MRM1m, GR32, i32mem,
-                        X86blsr_flag, loadi32>;
+                        X86blsr, loadi32>;
   defm BLSR64 : bmi_bls<"blsr{q}", MRM1r, MRM1m, GR64, i64mem,
-                        X86blsr_flag, loadi64>, VEX_W;
+                        X86blsr, loadi64>, VEX_W;
   defm BLSMSK32 : bmi_bls<"blsmsk{l}", MRM2r, MRM2m, GR32, i32mem,
-                          X86blsmsk_flag, loadi32>;
+                          X86blsmsk, loadi32>;
   defm BLSMSK64 : bmi_bls<"blsmsk{q}", MRM2r, MRM2m, GR64, i64mem,
-                          X86blsmsk_flag, loadi64>, VEX_W;
+                          X86blsmsk, loadi64>, VEX_W;
   defm BLSI32 : bmi_bls<"blsi{l}", MRM3r, MRM3m, GR32, i32mem,
-                        X86blsi_flag, loadi32>;
+                        X86blsi, loadi32>;
   defm BLSI64 : bmi_bls<"blsi{q}", MRM3r, MRM3m, GR64, i64mem,
-                        X86blsi_flag, loadi64>, VEX_W;
+                        X86blsi, loadi64>, VEX_W;
 }
 
 multiclass bmi_bextr_bzhi<bits<8> opc, string mnemonic, RegisterClass RC,