10 files changed, 145 insertions, 105 deletions

diff --git a/contrib/llvm/lib/Target/X86/AsmParser/X86AsmParser.cpp b/contrib/llvm/lib/Target/X86/AsmParser/X86AsmParser.cpp
index 899b50d0f78f..81391b96d126 100644
--- a/contrib/llvm/lib/Target/X86/AsmParser/X86AsmParser.cpp
+++ b/contrib/llvm/lib/Target/X86/AsmParser/X86AsmParser.cpp
@@ -1115,8 +1115,7 @@ bool X86AsmParser::ParseRegister(unsigned &RegNo,
   }
 
   // Parse "%st" as "%st(0)" and "%st(1)", which is multiple tokens.
-  if (RegNo == 0 && (Tok.getString() == "st" || Tok.getString() == "ST")) {
-    RegNo = X86::ST0;
+  if (RegNo == X86::ST0) {
     Parser.Lex(); // Eat 'st'
 
     // Check to see if we have '(4)' after %st.
diff --git a/contrib/llvm/lib/Target/X86/InstPrinter/X86ATTInstPrinter.cpp b/contrib/llvm/lib/Target/X86/InstPrinter/X86ATTInstPrinter.cpp
index 0e861d5ddbc9..3a074818c762 100644
--- a/contrib/llvm/lib/Target/X86/InstPrinter/X86ATTInstPrinter.cpp
+++ b/contrib/llvm/lib/Target/X86/InstPrinter/X86ATTInstPrinter.cpp
@@ -200,3 +200,14 @@ void X86ATTInstPrinter::printU8Imm(const MCInst *MI, unsigned Op,
   O << markup("<imm:") << '$' << formatImm(MI->getOperand(Op).getImm() & 0xff)
     << markup(">");
 }
+
+void X86ATTInstPrinter::printSTiRegOperand(const MCInst *MI, unsigned OpNo,
+                                           raw_ostream &OS) {
+  const MCOperand &Op = MI->getOperand(OpNo);
+  unsigned Reg = Op.getReg();
+  // Override the default printing to print st(0) instead st.
+  if (Reg == X86::ST0)
+    OS << markup("<reg:") << "%st(0)" << markup(">");
+  else
+    printRegName(OS, Reg);
+}
diff --git a/contrib/llvm/lib/Target/X86/InstPrinter/X86ATTInstPrinter.h b/contrib/llvm/lib/Target/X86/InstPrinter/X86ATTInstPrinter.h
index 57422bc9a0b2..584dc9c286e6 100644
--- a/contrib/llvm/lib/Target/X86/InstPrinter/X86ATTInstPrinter.h
+++ b/contrib/llvm/lib/Target/X86/InstPrinter/X86ATTInstPrinter.h
@@ -44,6 +44,7 @@ public:
   void printSrcIdx(const MCInst *MI, unsigned Op, raw_ostream &O);
   void printDstIdx(const MCInst *MI, unsigned Op, raw_ostream &O);
   void printU8Imm(const MCInst *MI, unsigned Op, raw_ostream &OS);
+  void printSTiRegOperand(const MCInst *MI, unsigned OpNo, raw_ostream &OS);
 
   void printanymem(const MCInst *MI, unsigned OpNo, raw_ostream &O) {
     printMemReference(MI, OpNo, O);
diff --git a/contrib/llvm/lib/Target/X86/InstPrinter/X86IntelInstPrinter.cpp b/contrib/llvm/lib/Target/X86/InstPrinter/X86IntelInstPrinter.cpp
index 044b71564152..b31f8ab80838 100644
--- a/contrib/llvm/lib/Target/X86/InstPrinter/X86IntelInstPrinter.cpp
+++ b/contrib/llvm/lib/Target/X86/InstPrinter/X86IntelInstPrinter.cpp
@@ -160,3 +160,14 @@ void X86IntelInstPrinter::printU8Imm(const MCInst *MI, unsigned Op,
 
   O << formatImm(MI->getOperand(Op).getImm() & 0xff);
 }
+
+void X86IntelInstPrinter::printSTiRegOperand(const MCInst *MI, unsigned OpNo,
+                                            raw_ostream &OS) {
+  const MCOperand &Op = MI->getOperand(OpNo);
+  unsigned Reg = Op.getReg();
+  // Override the default printing to print st(0) instead st.
+  if (Reg == X86::ST0)
+    OS << "st(0)";
+  else
+    printRegName(OS, Reg);
+}
diff --git a/contrib/llvm/lib/Target/X86/InstPrinter/X86IntelInstPrinter.h b/contrib/llvm/lib/Target/X86/InstPrinter/X86IntelInstPrinter.h
index 3b34a8052bec..fe52bd482a26 100644
--- a/contrib/llvm/lib/Target/X86/InstPrinter/X86IntelInstPrinter.h
+++ b/contrib/llvm/lib/Target/X86/InstPrinter/X86IntelInstPrinter.h
@@ -39,6 +39,7 @@ public:
   void printSrcIdx(const MCInst *MI, unsigned OpNo, raw_ostream &O);
   void printDstIdx(const MCInst *MI, unsigned OpNo, raw_ostream &O);
   void printU8Imm(const MCInst *MI, unsigned Op, raw_ostream &O);
+  void printSTiRegOperand(const MCInst *MI, unsigned OpNo, raw_ostream &OS);
 
   void printanymem(const MCInst *MI, unsigned OpNo, raw_ostream &O) {
     printMemReference(MI, OpNo, O);
diff --git a/contrib/llvm/lib/Target/X86/X86ISelLowering.cpp b/contrib/llvm/lib/Target/X86/X86ISelLowering.cpp
index 3637562c8ec3..f4f37a894620 100644
--- a/contrib/llvm/lib/Target/X86/X86ISelLowering.cpp
+++ b/contrib/llvm/lib/Target/X86/X86ISelLowering.cpp
@@ -42507,6 +42507,14 @@ X86TargetLowering::getRegForInlineAsmConstraint(const TargetRegisterInfo *TRI,
     if (StringRef("{flags}").equals_lower(Constraint))
       return std::make_pair(X86::EFLAGS, &X86::CCRRegClass);
 
+    // dirflag -> DF
+    if (StringRef("{dirflag}").equals_lower(Constraint))
+      return std::make_pair(X86::DF, &X86::DFCCRRegClass);
+
+    // fpsr -> FPSW
+    if (StringRef("{fpsr}").equals_lower(Constraint))
+      return std::make_pair(X86::FPSW, &X86::FPCCRRegClass);
+
     // 'A' means [ER]AX + [ER]DX.
     if (Constraint == "A") {
       if (Subtarget.is64Bit())
diff --git a/contrib/llvm/lib/Target/X86/X86InstrFPStack.td b/contrib/llvm/lib/Target/X86/X86InstrFPStack.td
index 5912a3199613..8e12efff77ea 100644
--- a/contrib/llvm/lib/Target/X86/X86InstrFPStack.td
+++ b/contrib/llvm/lib/Target/X86/X86InstrFPStack.td
@@ -230,7 +230,7 @@ def _FI32m  : FPI<0xDA, fp, (outs), (ins i32mem:$src),
 } // mayLoad = 1, hasSideEffects = 1
 }
 
-let Defs = [FPSW] in {
+let Defs = [FPSW], Uses = [FPCW] in {
 // FPBinary_rr just defines pseudo-instructions, no need to set a scheduling
 // resources.
 let hasNoSchedulingInfo = 1 in {
@@ -258,42 +258,42 @@ defm DIVR: FPBinary<fdiv, MRM7m, "divr", 0>;
 } // Defs = [FPSW]
 
 class FPST0rInst<Format fp, string asm>
-  : FPI<0xD8, fp, (outs), (ins RST:$op), asm>;
+  : FPI<0xD8, fp, (outs), (ins RSTi:$op), asm>;
 class FPrST0Inst<Format fp, string asm>
-  : FPI<0xDC, fp, (outs), (ins RST:$op), asm>;
+  : FPI<0xDC, fp, (outs), (ins RSTi:$op), asm>;
 class FPrST0PInst<Format fp, string asm>
-  : FPI<0xDE, fp, (outs), (ins RST:$op), asm>;
+  : FPI<0xDE, fp, (outs), (ins RSTi:$op), asm>;
 
 // NOTE: GAS and apparently all other AT&T style assemblers have a broken notion
 // of some of the 'reverse' forms of the fsub and fdiv instructions.  As such,
 // we have to put some 'r's in and take them out of weird places.
-let SchedRW = [WriteFAdd] in {
-def ADD_FST0r   : FPST0rInst <MRM0r, "fadd\t$op">;
-def ADD_FrST0   : FPrST0Inst <MRM0r, "fadd\t{%st(0), $op|$op, st(0)}">;
-def ADD_FPrST0  : FPrST0PInst<MRM0r, "faddp\t$op">;
-def SUBR_FST0r  : FPST0rInst <MRM5r, "fsubr\t$op">;
-def SUB_FrST0   : FPrST0Inst <MRM5r, "fsub{r}\t{%st(0), $op|$op, st(0)}">;
-def SUB_FPrST0  : FPrST0PInst<MRM5r, "fsub{r}p\t$op">;
-def SUB_FST0r   : FPST0rInst <MRM4r, "fsub\t$op">;
-def SUBR_FrST0  : FPrST0Inst <MRM4r, "fsub{|r}\t{%st(0), $op|$op, st(0)}">;
-def SUBR_FPrST0 : FPrST0PInst<MRM4r, "fsub{|r}p\t$op">;
+let SchedRW = [WriteFAdd], Defs = [FPSW], Uses = [FPCW] in {
+def ADD_FST0r   : FPST0rInst <MRM0r, "fadd\t{$op, %st|st, $op}">;
+def ADD_FrST0   : FPrST0Inst <MRM0r, "fadd\t{%st, $op|$op, st}">;
+def ADD_FPrST0  : FPrST0PInst<MRM0r, "faddp\t{%st, $op|$op, st}">;
+def SUBR_FST0r  : FPST0rInst <MRM5r, "fsubr\t{$op, %st|st, $op}">;
+def SUB_FrST0   : FPrST0Inst <MRM5r, "fsub{r}\t{%st, $op|$op, st}">;
+def SUB_FPrST0  : FPrST0PInst<MRM5r, "fsub{r}p\t{%st, $op|$op, st}">;
+def SUB_FST0r   : FPST0rInst <MRM4r, "fsub\t{$op, %st|st, $op}">;
+def SUBR_FrST0  : FPrST0Inst <MRM4r, "fsub{|r}\t{%st, $op|$op, st}">;
+def SUBR_FPrST0 : FPrST0PInst<MRM4r, "fsub{|r}p\t{%st, $op|$op, st}">;
 } // SchedRW
-let SchedRW = [WriteFCom] in {
+let SchedRW = [WriteFCom], Defs = [FPSW], Uses = [FPCW] in {
 def COM_FST0r   : FPST0rInst <MRM2r, "fcom\t$op">;
 def COMP_FST0r  : FPST0rInst <MRM3r, "fcomp\t$op">;
 } // SchedRW
-let SchedRW = [WriteFMul] in {
-def MUL_FST0r   : FPST0rInst <MRM1r, "fmul\t$op">;
-def MUL_FrST0   : FPrST0Inst <MRM1r, "fmul\t{%st(0), $op|$op, st(0)}">;
-def MUL_FPrST0  : FPrST0PInst<MRM1r, "fmulp\t$op">;
+let SchedRW = [WriteFMul], Defs = [FPSW], Uses = [FPCW] in {
+def MUL_FST0r   : FPST0rInst <MRM1r, "fmul\t{$op, %st|st, $op}">;
+def MUL_FrST0   : FPrST0Inst <MRM1r, "fmul\t{%st, $op|$op, st}">;
+def MUL_FPrST0  : FPrST0PInst<MRM1r, "fmulp\t{%st, $op|$op, st}">;
 } // SchedRW
-let SchedRW = [WriteFDiv] in {
-def DIVR_FST0r  : FPST0rInst <MRM7r, "fdivr\t$op">;
-def DIV_FrST0   : FPrST0Inst <MRM7r, "fdiv{r}\t{%st(0), $op|$op, st(0)}">;
-def DIV_FPrST0  : FPrST0PInst<MRM7r, "fdiv{r}p\t$op">;
-def DIV_FST0r   : FPST0rInst <MRM6r, "fdiv\t$op">;
-def DIVR_FrST0  : FPrST0Inst <MRM6r, "fdiv{|r}\t{%st(0), $op|$op, st(0)}">;
-def DIVR_FPrST0 : FPrST0PInst<MRM6r, "fdiv{|r}p\t$op">;
+let SchedRW = [WriteFDiv], Defs = [FPSW], Uses = [FPCW] in {
+def DIVR_FST0r  : FPST0rInst <MRM7r, "fdivr\t{$op, %st|st, $op}">;
+def DIV_FrST0   : FPrST0Inst <MRM7r, "fdiv{r}\t{%st, $op|$op, st}">;
+def DIV_FPrST0  : FPrST0PInst<MRM7r, "fdiv{r}p\t{%st, $op|$op, st}">;
+def DIV_FST0r   : FPST0rInst <MRM6r, "fdiv\t{$op, %st|st, $op}">;
+def DIVR_FrST0  : FPrST0Inst <MRM6r, "fdiv{|r}\t{%st, $op|$op, st}">;
+def DIVR_FPrST0 : FPrST0PInst<MRM6r, "fdiv{|r}p\t{%st, $op|$op, st}">;
 } // SchedRW
 
 // Unary operations.
@@ -307,7 +307,7 @@ def _Fp80  : FpI_<(outs RFP80:$dst), (ins RFP80:$src), OneArgFPRW,
 def _F     : FPI<0xD9, fp, (outs), (ins), asmstring>;
 }
 
-let Defs = [FPSW] in {
+let Defs = [FPSW], Uses = [FPCW] in {
 
 let SchedRW = [WriteFSign] in {
 defm CHS : FPUnary<fneg, MRM_E0, "fchs">;
@@ -335,7 +335,7 @@ def TST_F  : FPI<0xD9, MRM_E4, (outs), (ins), "ftst">;
 
 // Versions of FP instructions that take a single memory operand.  Added for the
 //   disassembler; remove as they are included with patterns elsewhere.
-let SchedRW = [WriteFComLd] in {
+let SchedRW = [WriteFComLd], Defs = [FPSW], Uses = [FPCW] in {
 def FCOM32m  : FPI<0xD8, MRM2m, (outs), (ins f32mem:$src), "fcom{s}\t$src">;
 def FCOMP32m : FPI<0xD8, MRM3m, (outs), (ins f32mem:$src), "fcomp{s}\t$src">;
 
@@ -398,22 +398,22 @@ defm CMOVNP : FPCMov<X86_COND_NP>;
 
 let Predicates = [HasCMov] in {
 // These are not factored because there's no clean way to pass DA/DB.
-def CMOVB_F  : FPI<0xDA, MRM0r, (outs), (ins RST:$op),
-                  "fcmovb\t{$op, %st(0)|st(0), $op}">;
-def CMOVBE_F : FPI<0xDA, MRM2r, (outs), (ins RST:$op),
-                  "fcmovbe\t{$op, %st(0)|st(0), $op}">;
-def CMOVE_F  : FPI<0xDA, MRM1r, (outs), (ins RST:$op),
-                  "fcmove\t{$op, %st(0)|st(0), $op}">;
-def CMOVP_F  : FPI<0xDA, MRM3r, (outs), (ins RST:$op),
-                  "fcmovu\t{$op, %st(0)|st(0), $op}">;
-def CMOVNB_F : FPI<0xDB, MRM0r, (outs), (ins RST:$op),
-                  "fcmovnb\t{$op, %st(0)|st(0), $op}">;
-def CMOVNBE_F: FPI<0xDB, MRM2r, (outs), (ins RST:$op),
-                  "fcmovnbe\t{$op, %st(0)|st(0), $op}">;
-def CMOVNE_F : FPI<0xDB, MRM1r, (outs), (ins RST:$op),
-                  "fcmovne\t{$op, %st(0)|st(0), $op}">;
-def CMOVNP_F : FPI<0xDB, MRM3r, (outs), (ins RST:$op),
-                  "fcmovnu\t{$op, %st(0)|st(0), $op}">;
+def CMOVB_F  : FPI<0xDA, MRM0r, (outs), (ins RSTi:$op),
+                  "fcmovb\t{$op, %st|st, $op}">;
+def CMOVBE_F : FPI<0xDA, MRM2r, (outs), (ins RSTi:$op),
+                  "fcmovbe\t{$op, %st|st, $op}">;
+def CMOVE_F  : FPI<0xDA, MRM1r, (outs), (ins RSTi:$op),
+                  "fcmove\t{$op, %st|st, $op}">;
+def CMOVP_F  : FPI<0xDA, MRM3r, (outs), (ins RSTi:$op),
+                  "fcmovu\t{$op, %st|st, $op}">;
+def CMOVNB_F : FPI<0xDB, MRM0r, (outs), (ins RSTi:$op),
+                  "fcmovnb\t{$op, %st|st, $op}">;
+def CMOVNBE_F: FPI<0xDB, MRM2r, (outs), (ins RSTi:$op),
+                  "fcmovnbe\t{$op, %st|st, $op}">;
+def CMOVNE_F : FPI<0xDB, MRM1r, (outs), (ins RSTi:$op),
+                  "fcmovne\t{$op, %st|st, $op}">;
+def CMOVNP_F : FPI<0xDB, MRM3r, (outs), (ins RSTi:$op),
+                  "fcmovnu\t{$op, %st|st, $op}">;
 } // Predicates = [HasCMov]
 } // SchedRW
 
@@ -454,7 +454,7 @@ def ILD_Fp64m80: FpI_<(outs RFP80:$dst), (ins i64mem:$src), ZeroArgFP,
                   [(set RFP80:$dst, (X86fild addr:$src, i64))]>;
 } // SchedRW
 
-let SchedRW = [WriteStore] in {
+let SchedRW = [WriteStore], Uses = [FPCW] in {
 def ST_Fp32m   : FpIf32<(outs), (ins f32mem:$op, RFP32:$src), OneArgFP,
                   [(store RFP32:$src, addr:$op)]>;
 def ST_Fp64m32 : FpIf64<(outs), (ins f32mem:$op, RFP64:$src), OneArgFP,
@@ -489,7 +489,7 @@ def IST_Fp16m80  : FpI_<(outs), (ins i16mem:$op, RFP80:$src), OneArgFP, []>;
 def IST_Fp32m80  : FpI_<(outs), (ins i32mem:$op, RFP80:$src), OneArgFP, []>;
 def IST_Fp64m80  : FpI_<(outs), (ins i64mem:$op, RFP80:$src), OneArgFP, []>;
 } // mayStore
-} // SchedRW
+} // SchedRW, Uses = [FPCW]
 
 let mayLoad = 1, SchedRW = [WriteLoad] in {
 def LD_F32m   : FPI<0xD9, MRM0m, (outs), (ins f32mem:$src), "fld{s}\t$src">;
@@ -499,7 +499,7 @@ def ILD_F16m  : FPI<0xDF, MRM0m, (outs), (ins i16mem:$src), "fild{s}\t$src">;
 def ILD_F32m  : FPI<0xDB, MRM0m, (outs), (ins i32mem:$src), "fild{l}\t$src">;
 def ILD_F64m  : FPI<0xDF, MRM5m, (outs), (ins i64mem:$src), "fild{ll}\t$src">;
 }
-let mayStore = 1, SchedRW = [WriteStore] in {
+let mayStore = 1, SchedRW = [WriteStore], Uses = [FPCW] in {
 def ST_F32m   : FPI<0xD9, MRM2m, (outs), (ins f32mem:$dst), "fst{s}\t$dst">;
 def ST_F64m   : FPI<0xDD, MRM2m, (outs), (ins f64mem:$dst), "fst{l}\t$dst">;
 def ST_FP32m  : FPI<0xD9, MRM3m, (outs), (ins f32mem:$dst), "fstp{s}\t$dst">;
@@ -513,7 +513,7 @@ def IST_FP64m : FPI<0xDF, MRM7m, (outs), (ins i64mem:$dst), "fistp{ll}\t$dst">;
 }
 
 // FISTTP requires SSE3 even though it's a FPStack op.
-let Predicates = [HasSSE3], SchedRW = [WriteStore] in {
+let Predicates = [HasSSE3], SchedRW = [WriteStore], Uses = [FPCW] in {
 def ISTT_Fp16m32 : FpI_<(outs), (ins i16mem:$op, RFP32:$src), OneArgFP,
                     [(X86fp_to_i16mem RFP32:$src, addr:$op)]>;
 def ISTT_Fp32m32 : FpI_<(outs), (ins i32mem:$op, RFP32:$src), OneArgFP,
@@ -534,7 +534,7 @@ def ISTT_Fp64m80 : FpI_<(outs), (ins i64mem:$op, RFP80:$src), OneArgFP,
                     [(X86fp_to_i64mem RFP80:$src, addr:$op)]>;
 } // Predicates = [HasSSE3]
 
-let mayStore = 1, SchedRW = [WriteStore] in {
+let mayStore = 1, SchedRW = [WriteStore], Uses = [FPCW] in {
 def ISTT_FP16m : FPI<0xDF, MRM1m, (outs), (ins i16mem:$dst), "fisttp{s}\t$dst">;
 def ISTT_FP32m : FPI<0xDB, MRM1m, (outs), (ins i32mem:$dst), "fisttp{l}\t$dst">;
 def ISTT_FP64m : FPI<0xDD, MRM1m, (outs), (ins i64mem:$dst), "fisttp{ll}\t$dst">;
@@ -542,10 +542,10 @@ def ISTT_FP64m : FPI<0xDD, MRM1m, (outs), (ins i64mem:$dst), "fisttp{ll}\t$dst">
 
 // FP Stack manipulation instructions.
 let SchedRW = [WriteMove] in {
-def LD_Frr   : FPI<0xD9, MRM0r, (outs), (ins RST:$op), "fld\t$op">;
-def ST_Frr   : FPI<0xDD, MRM2r, (outs), (ins RST:$op), "fst\t$op">;
-def ST_FPrr  : FPI<0xDD, MRM3r, (outs), (ins RST:$op), "fstp\t$op">;
-def XCH_F    : FPI<0xD9, MRM1r, (outs), (ins RST:$op), "fxch\t$op">;
+def LD_Frr   : FPI<0xD9, MRM0r, (outs), (ins RSTi:$op), "fld\t$op">;
+def ST_Frr   : FPI<0xDD, MRM2r, (outs), (ins RSTi:$op), "fst\t$op">;
+def ST_FPrr  : FPI<0xDD, MRM3r, (outs), (ins RSTi:$op), "fstp\t$op">;
+def XCH_F    : FPI<0xD9, MRM1r, (outs), (ins RSTi:$op), "fxch\t$op">;
 }
 
 // Floating point constant loads.
@@ -570,7 +570,7 @@ def LD_F0 : FPI<0xD9, MRM_EE, (outs), (ins), "fldz">;
 let SchedRW = [WriteFLD1] in
 def LD_F1 : FPI<0xD9, MRM_E8, (outs), (ins), "fld1">;
 
-let SchedRW = [WriteFLDC], Defs = [FPSW] in {
+let SchedRW = [WriteFLDC] in {
 def FLDL2T : I<0xD9, MRM_E9, (outs), (ins), "fldl2t", []>;
 def FLDL2E : I<0xD9, MRM_EA, (outs), (ins), "fldl2e", []>;
 def FLDPI : I<0xD9, MRM_EB, (outs), (ins), "fldpi", []>;
@@ -579,7 +579,7 @@ def FLDLN2 : I<0xD9, MRM_ED, (outs), (ins), "fldln2", []>;
 } // SchedRW
 
 // Floating point compares.
-let SchedRW = [WriteFCom] in {
+let SchedRW = [WriteFCom], Uses = [FPCW] in {
 def UCOM_Fpr32 : FpIf32<(outs), (ins RFP32:$lhs, RFP32:$rhs), CompareFP,
                         [(set FPSW, (trunc (X86cmp RFP32:$lhs, RFP32:$rhs)))]>;
 def UCOM_Fpr64 : FpIf64<(outs), (ins RFP64:$lhs, RFP64:$rhs), CompareFP,
@@ -591,37 +591,37 @@ def UCOM_Fpr80 : FpI_  <(outs), (ins RFP80:$lhs, RFP80:$rhs), CompareFP,
 
 let SchedRW = [WriteFCom] in {
 // CC = ST(0) cmp ST(i)
-let Defs = [EFLAGS, FPSW] in {
-let Predicates = [FPStackf32, HasCMov] in
-def UCOM_FpIr32: FpIf32<(outs), (ins RFP32:$lhs, RFP32:$rhs), CompareFP,
-                  [(set EFLAGS, (X86cmp RFP32:$lhs, RFP32:$rhs))]>;
-let Predicates = [FPStackf64, HasCMov] in
-def UCOM_FpIr64: FpIf64<(outs), (ins RFP64:$lhs, RFP64:$rhs), CompareFP,
-                  [(set EFLAGS, (X86cmp RFP64:$lhs, RFP64:$rhs))]>;
-let Predicates = [HasCMov] in
+let Defs = [EFLAGS, FPSW], Uses = [FPCW] in {
+def UCOM_FpIr32: FpI_<(outs), (ins RFP32:$lhs, RFP32:$rhs), CompareFP,
+                  [(set EFLAGS, (X86cmp RFP32:$lhs, RFP32:$rhs))]>,
+                  Requires<[FPStackf32, HasCMov]>;
+def UCOM_FpIr64: FpI_<(outs), (ins RFP64:$lhs, RFP64:$rhs), CompareFP,
+                  [(set EFLAGS, (X86cmp RFP64:$lhs, RFP64:$rhs))]>,
+                  Requires<[FPStackf64, HasCMov]>;
 def UCOM_FpIr80: FpI_<(outs), (ins RFP80:$lhs, RFP80:$rhs), CompareFP,
-                  [(set EFLAGS, (X86cmp RFP80:$lhs, RFP80:$rhs))]>;
+                  [(set EFLAGS, (X86cmp RFP80:$lhs, RFP80:$rhs))]>,
+                  Requires<[HasCMov]>;
 }
 
-let Defs = [FPSW], Uses = [ST0] in {
+let Defs = [FPSW], Uses = [ST0, FPCW] in {
 def UCOM_Fr    : FPI<0xDD, MRM4r,    // FPSW = cmp ST(0) with ST(i)
-                    (outs), (ins RST:$reg), "fucom\t$reg">;
+                    (outs), (ins RSTi:$reg), "fucom\t$reg">;
 def UCOM_FPr   : FPI<0xDD, MRM5r,    // FPSW = cmp ST(0) with ST(i), pop
-                    (outs), (ins RST:$reg), "fucomp\t$reg">;
+                    (outs), (ins RSTi:$reg), "fucomp\t$reg">;
 def UCOM_FPPr  : FPI<0xDA, MRM_E9,       // cmp ST(0) with ST(1), pop, pop
                     (outs), (ins), "fucompp">;
 }
 
-let Defs = [EFLAGS, FPSW], Uses = [ST0] in {
+let Defs = [EFLAGS, FPSW], Uses = [ST0, FPCW] in {
 def UCOM_FIr   : FPI<0xDB, MRM5r,     // CC = cmp ST(0) with ST(i)
-                    (outs), (ins RST:$reg), "fucomi\t$reg">;
+                    (outs), (ins RSTi:$reg), "fucomi\t{$reg, %st|st, $reg}">;
 def UCOM_FIPr  : FPI<0xDF, MRM5r,     // CC = cmp ST(0) with ST(i), pop
-                    (outs), (ins RST:$reg), "fucompi\t$reg">;
-}
+                    (outs), (ins RSTi:$reg), "fucompi\t{$reg, %st|st, $reg}">;
 
-let Defs = [EFLAGS, FPSW] in {
-def COM_FIr : FPI<0xDB, MRM6r, (outs), (ins RST:$reg), "fcomi\t$reg">;
-def COM_FIPr : FPI<0xDF, MRM6r, (outs), (ins RST:$reg), "fcompi\t$reg">;
+def COM_FIr : FPI<0xDB, MRM6r, (outs), (ins RSTi:$reg),
+                  "fcomi\t{$reg, %st|st, $reg}">;
+def COM_FIPr : FPI<0xDF, MRM6r, (outs), (ins RSTi:$reg),
+                   "fcompi\t{$reg, %st|st, $reg}">;
 }
 } // SchedRW
 
@@ -631,12 +631,12 @@ let Defs = [AX], Uses = [FPSW] in
 def FNSTSW16r : I<0xDF, MRM_E0,                  // AX = fp flags
                   (outs), (ins), "fnstsw\t{%ax|ax}",
                   [(set AX, (X86fp_stsw FPSW))]>;
-let Defs = [FPSW] in
+let Defs = [FPSW], Uses = [FPCW] in
 def FNSTCW16m : I<0xD9, MRM7m,                   // [mem16] = X87 control world
                   (outs), (ins i16mem:$dst), "fnstcw\t$dst",
                   [(X86fp_cwd_get16 addr:$dst)]>;
 } // SchedRW
-let Defs = [FPSW], mayLoad = 1 in
+let Defs = [FPSW,FPCW], mayLoad = 1 in
 def FLDCW16m  : I<0xD9, MRM5m,                   // X87 control world = [mem16]
                   (outs), (ins i16mem:$dst), "fldcw\t$dst", []>,
                 Sched<[WriteLoad]>;
@@ -645,8 +645,8 @@ def FLDCW16m  : I<0xD9, MRM5m,                   // X87 control world = [mem16]
 let SchedRW = [WriteMicrocoded] in {
 let Defs = [FPSW] in {
 def FNINIT : I<0xDB, MRM_E3, (outs), (ins), "fninit", []>;
-def FFREE : FPI<0xDD, MRM0r, (outs), (ins RST:$reg), "ffree\t$reg">;
-def FFREEP : FPI<0xDF, MRM0r, (outs), (ins RST:$reg), "ffreep\t$reg">;
+def FFREE : FPI<0xDD, MRM0r, (outs), (ins RSTi:$reg), "ffree\t$reg">;
+def FFREEP : FPI<0xDF, MRM0r, (outs), (ins RSTi:$reg), "ffreep\t$reg">;
 
 // Clear exceptions
 def FNCLEX : I<0xDB, MRM_E2, (outs), (ins), "fnclex", []>;
diff --git a/contrib/llvm/lib/Target/X86/X86InstrInfo.td b/contrib/llvm/lib/Target/X86/X86InstrInfo.td
index e53f83baa3c6..4ec4d566ca99 100644
--- a/contrib/llvm/lib/Target/X86/X86InstrInfo.td
+++ b/contrib/llvm/lib/Target/X86/X86InstrInfo.td
@@ -3231,39 +3231,39 @@ def : InstAlias<"fucompi",      (UCOM_FIPr   ST1), 0>;
 // instructions like "fadd %st(0), %st(0)" as "fadd %st(0)" for consistency with
 // gas.
 multiclass FpUnaryAlias<string Mnemonic, Instruction Inst, bit EmitAlias = 1> {
- def : InstAlias<!strconcat(Mnemonic, "\t{$op, %st(0)|st(0), $op}"),
-                 (Inst RST:$op), EmitAlias>;
- def : InstAlias<!strconcat(Mnemonic, "\t{%st(0), %st(0)|st(0), st(0)}"),
+ def : InstAlias<!strconcat(Mnemonic, "\t$op"),
+                 (Inst RSTi:$op), EmitAlias>;
+ def : InstAlias<!strconcat(Mnemonic, "\t{%st, %st|st, st}"),
                  (Inst ST0), EmitAlias>;
 }
 
-defm : FpUnaryAlias<"fadd",   ADD_FST0r>;
+defm : FpUnaryAlias<"fadd",   ADD_FST0r, 0>;
 defm : FpUnaryAlias<"faddp",  ADD_FPrST0, 0>;
-defm : FpUnaryAlias<"fsub",   SUB_FST0r>;
-defm : FpUnaryAlias<"fsub{|r}p",  SUBR_FPrST0>;
-defm : FpUnaryAlias<"fsubr",  SUBR_FST0r>;
-defm : FpUnaryAlias<"fsub{r|}p", SUB_FPrST0>;
-defm : FpUnaryAlias<"fmul",   MUL_FST0r>;
-defm : FpUnaryAlias<"fmulp",  MUL_FPrST0>;
-defm : FpUnaryAlias<"fdiv",   DIV_FST0r>;
-defm : FpUnaryAlias<"fdiv{|r}p",  DIVR_FPrST0>;
-defm : FpUnaryAlias<"fdivr",  DIVR_FST0r>;
-defm : FpUnaryAlias<"fdiv{r|}p", DIV_FPrST0>;
+defm : FpUnaryAlias<"fsub",   SUB_FST0r, 0>;
+defm : FpUnaryAlias<"fsub{|r}p",  SUBR_FPrST0, 0>;
+defm : FpUnaryAlias<"fsubr",  SUBR_FST0r, 0>;
+defm : FpUnaryAlias<"fsub{r|}p", SUB_FPrST0, 0>;
+defm : FpUnaryAlias<"fmul",   MUL_FST0r, 0>;
+defm : FpUnaryAlias<"fmulp",  MUL_FPrST0, 0>;
+defm : FpUnaryAlias<"fdiv",   DIV_FST0r, 0>;
+defm : FpUnaryAlias<"fdiv{|r}p",  DIVR_FPrST0, 0>;
+defm : FpUnaryAlias<"fdivr",  DIVR_FST0r, 0>;
+defm : FpUnaryAlias<"fdiv{r|}p", DIV_FPrST0, 0>;
 defm : FpUnaryAlias<"fcomi",   COM_FIr, 0>;
 defm : FpUnaryAlias<"fucomi",  UCOM_FIr, 0>;
-defm : FpUnaryAlias<"fcompi",   COM_FIPr>;
-defm : FpUnaryAlias<"fucompi",  UCOM_FIPr>;
+defm : FpUnaryAlias<"fcompi",   COM_FIPr, 0>;
+defm : FpUnaryAlias<"fucompi",  UCOM_FIPr, 0>;
 
 
-// Handle "f{mulp,addp} st(0), $op" the same as "f{mulp,addp} $op", since they
+// Handle "f{mulp,addp} $op, %st(0)" the same as "f{mulp,addp} $op", since they
 // commute.  We also allow fdiv[r]p/fsubrp even though they don't commute,
 // solely because gas supports it.
-def : InstAlias<"faddp\t{%st(0), $op|$op, st(0)}", (ADD_FPrST0 RST:$op), 0>;
-def : InstAlias<"fmulp\t{%st(0), $op|$op, st(0)}", (MUL_FPrST0 RST:$op)>;
-def : InstAlias<"fsub{|r}p\t{%st(0), $op|$op, st(0)}", (SUBR_FPrST0 RST:$op)>;
-def : InstAlias<"fsub{r|}p\t{%st(0), $op|$op, st(0)}", (SUB_FPrST0 RST:$op)>;
-def : InstAlias<"fdiv{|r}p\t{%st(0), $op|$op, st(0)}", (DIVR_FPrST0 RST:$op)>;
-def : InstAlias<"fdiv{r|}p\t{%st(0), $op|$op, st(0)}", (DIV_FPrST0 RST:$op)>;
+def : InstAlias<"faddp\t{$op, %st|st, $op}", (ADD_FPrST0 RSTi:$op), 0>;
+def : InstAlias<"fmulp\t{$op, %st|st, $op}", (MUL_FPrST0 RSTi:$op), 0>;
+def : InstAlias<"fsub{|r}p\t{$op, %st|st, $op}", (SUBR_FPrST0 RSTi:$op), 0>;
+def : InstAlias<"fsub{r|}p\t{$op, %st|st, $op}", (SUB_FPrST0 RSTi:$op), 0>;
+def : InstAlias<"fdiv{|r}p\t{$op, %st|st, $op}", (DIVR_FPrST0 RSTi:$op), 0>;
+def : InstAlias<"fdiv{r|}p\t{$op, %st|st, $op}", (DIV_FPrST0 RSTi:$op), 0>;
 
 def : InstAlias<"fnstsw"     , (FNSTSW16r), 0>;
 
diff --git a/contrib/llvm/lib/Target/X86/X86RegisterInfo.cpp b/contrib/llvm/lib/Target/X86/X86RegisterInfo.cpp
index 55842a4a2091..bc39cee34c4a 100644
--- a/contrib/llvm/lib/Target/X86/X86RegisterInfo.cpp
+++ b/contrib/llvm/lib/Target/X86/X86RegisterInfo.cpp
@@ -497,6 +497,9 @@ BitVector X86RegisterInfo::getReservedRegs(const MachineFunction &MF) const {
   BitVector Reserved(getNumRegs());
   const X86FrameLowering *TFI = getFrameLowering(MF);
 
+  // Set the floating point control register as reserved.
+  Reserved.set(X86::FPCW);
+
   // Set the stack-pointer register and its aliases as reserved.
   for (MCSubRegIterator I(X86::RSP, this, /*IncludeSelf=*/true); I.isValid();
        ++I)
diff --git a/contrib/llvm/lib/Target/X86/X86RegisterInfo.td b/contrib/llvm/lib/Target/X86/X86RegisterInfo.td
index aa20273f89ab..6a0538138528 100644
--- a/contrib/llvm/lib/Target/X86/X86RegisterInfo.td
+++ b/contrib/llvm/lib/Target/X86/X86RegisterInfo.td
@@ -278,7 +278,7 @@ def K7 : X86Reg<"k7", 7>, DwarfRegNum<[125, 100, 100]>;
 // pseudo registers, but we still mark them as aliasing FP registers. That
 // way both kinds can be live without exceeding the stack depth. ST registers
 // are only live around inline assembly.
-def ST0 : X86Reg<"st(0)", 0>, DwarfRegNum<[33, 12, 11]>;
+def ST0 : X86Reg<"st", 0>, DwarfRegNum<[33, 12, 11]>;
 def ST1 : X86Reg<"st(1)", 1>, DwarfRegNum<[34, 13, 12]>;
 def ST2 : X86Reg<"st(2)", 2>, DwarfRegNum<[35, 14, 13]>;
 def ST3 : X86Reg<"st(3)", 3>, DwarfRegNum<[36, 15, 14]>;
@@ -288,7 +288,10 @@ def ST6 : X86Reg<"st(6)", 6>, DwarfRegNum<[39, 18, 17]>;
 def ST7 : X86Reg<"st(7)", 7>, DwarfRegNum<[40, 19, 18]>;
 
 // Floating-point status word
-def FPSW : X86Reg<"fpsw", 0>;
+def FPSW : X86Reg<"fpsr", 0>;
+
+// Floating-point control word
+def FPCW : X86Reg<"fpcr", 0>;
 
 // Status flags register.
 //
@@ -539,6 +542,9 @@ def RST : RegisterClass<"X86", [f80, f64, f32], 32, (sequence "ST%u", 0, 7)> {
   let isAllocatable = 0;
 }
 
+// Helper to allow %st to print as %st(0) when its encoded in the instruction.
+def RSTi : RegisterOperand<RST, "printSTiRegOperand">;
+
 // Generic vector registers: VR64 and VR128.
 // Ensure that float types are declared first - only float is legal on SSE1.
 def VR64: RegisterClass<"X86", [x86mmx], 64, (sequence "MM%u", 0, 7)>;