Add crc32 instruction and intrinsics. Add a new class of prefix

bytes for F2 0F 38 and propagate. Add a FIXME for a set
of possibilities which correspond to intrinsics already used.

New test.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78508 91177308-0d34-0410-b5e6-96231b3b80d8

6 files changed, 131 insertions, 0 deletions

diff --git a/lib/Target/X86/X86CodeEmitter.cpp b/lib/Target/X86/X86CodeEmitter.cpp
index 03d871a..3dd3c6b 100644
--- a/lib/Target/X86/X86CodeEmitter.cpp
+++ b/lib/Target/X86/X86CodeEmitter.cpp
@@ -517,6 +517,10 @@ void Emitter<CodeEmitter>::emitInstruction(
   case X86II::TA:  // 0F 3A
     Need0FPrefix = true;
     break;
+  case X86II::TF: // F2 0F 38
+    MCE.emitByte(0xF2);
+    Need0FPrefix = true;
+    break;
   case X86II::REP: break; // already handled.
   case X86II::XS:   // F3 0F
     MCE.emitByte(0xF3);
@@ -548,6 +552,7 @@ void Emitter<CodeEmitter>::emitInstruction(
     MCE.emitByte(0x0F);
 
   switch (Desc->TSFlags & X86II::Op0Mask) {
+  case X86II::TF:  // F2 0F 38
   case X86II::T8:  // 0F 38
     MCE.emitByte(0x38);
     break;
diff --git a/lib/Target/X86/X86InstrFormats.td b/lib/Target/X86/X86InstrFormats.td
index eeed5bd..6f5941c 100644
--- a/lib/Target/X86/X86InstrFormats.td
+++ b/lib/Target/X86/X86InstrFormats.td
@@ -79,6 +79,7 @@ class XD     { bits<4> Prefix = 11; }
 class XS     { bits<4> Prefix = 12; }
 class T8     { bits<4> Prefix = 13; }
 class TA     { bits<4> Prefix = 14; }
+class TF     { bits<4> Prefix = 15; }
 
 class X86Inst<bits<8> opcod, Format f, ImmType i, dag outs, dag ins,
               string AsmStr>
@@ -229,6 +230,11 @@ class SS428I<bits<8> o, Format F, dag outs, dag ins, string asm,
              list<dag> pattern>
       : I<o, F, outs, ins, asm, pattern>, T8, Requires<[HasSSE42]>;
 
+//   SS42FI - SSE 4.2 instructions with TF prefix.
+class SS42FI<bits<8> o, Format F, dag outs, dag ins, string asm,
+              list<dag> pattern>
+      : I<o, F, outs, ins, asm, pattern>, TF, Requires<[HasSSE42]>;
+      
 // X86-64 Instruction templates...
 //
 
diff --git a/lib/Target/X86/X86InstrInfo.cpp b/lib/Target/X86/X86InstrInfo.cpp
index 393bd5c..b59f8e8 100644
--- a/lib/Target/X86/X86InstrInfo.cpp
+++ b/lib/Target/X86/X86InstrInfo.cpp
@@ -2931,6 +2931,10 @@ static unsigned GetInstSizeWithDesc(const MachineInstr &MI,
   case X86II::TA:  // 0F 3A
     Need0FPrefix = true;
     break;
+  case X86II::TF: // F2 0F 38
+    ++FinalSize;
+    Need0FPrefix = true;
+    break;
   case X86II::REP: break; // already handled.
   case X86II::XS:   // F3 0F
     ++FinalSize;
@@ -2966,6 +2970,9 @@ static unsigned GetInstSizeWithDesc(const MachineInstr &MI,
   case X86II::TA:  // 0F 3A
     ++FinalSize;
     break;
+  case X86II::TF: // F2 0F 38
+    ++FinalSize;
+    break;
   }
 
   // If this is a two-address instruction, skip one of the register operands.
diff --git a/lib/Target/X86/X86InstrInfo.h b/lib/Target/X86/X86InstrInfo.h
index 0fb2052..2e0235a 100644
--- a/lib/Target/X86/X86InstrInfo.h
+++ b/lib/Target/X86/X86InstrInfo.h
@@ -321,6 +321,9 @@ namespace X86II {
 
     // T8, TA - Prefix after the 0x0F prefix.
     T8 = 13 << Op0Shift,  TA = 14 << Op0Shift,
+    
+    // TF - Prefix before and after 0x0F
+    TF = 15 << Op0Shift,
 
     //===------------------------------------------------------------------===//
     // REX_W - REX prefixes are instruction prefixes used in 64-bit mode.
diff --git a/lib/Target/X86/X86InstrSSE.td b/lib/Target/X86/X86InstrSSE.td
index 890ebae..fefdba0 100644
--- a/lib/Target/X86/X86InstrSSE.td
+++ b/lib/Target/X86/X86InstrSSE.td
@@ -3678,3 +3678,75 @@ def : Pat<(v2i64 (X86pcmpgtq VR128:$src1, VR128:$src2)),
           (PCMPGTQrr VR128:$src1, VR128:$src2)>;
 def : Pat<(v2i64 (X86pcmpgtq VR128:$src1, (memop addr:$src2))),
           (PCMPGTQrm VR128:$src1, addr:$src2)>;
+
+// crc intrinsic instruction
+// This set of instructions are only rm, the only difference is the size
+// of r and m.
+let Constraints = "$src1 = $dst" in {
+  def CRC32m8  : SS42FI<0xF0, MRMSrcMem, (outs GR32:$dst), 
+                      (ins GR32:$src1, i8mem:$src2),
+                      "crc32 \t{$src2, $src1|$src1, $src2}",
+                       [(set GR32:$dst,
+                         (int_x86_sse42_crc32_8 GR32:$src1,
+                         (load addr:$src2)))]>, OpSize;
+  def CRC32r8  : SS42FI<0xF0, MRMSrcReg, (outs GR32:$dst), 
+                      (ins GR32:$src1, GR8:$src2),
+                      "crc32 \t{$src2, $src1|$src1, $src2}",
+                       [(set GR32:$dst,
+                         (int_x86_sse42_crc32_8 GR32:$src1, GR8:$src2))]>, 
+                         OpSize;
+  def CRC32m16  : SS42FI<0xF1, MRMSrcMem, (outs GR32:$dst), 
+                      (ins GR32:$src1, i16mem:$src2),
+                      "crc32 \t{$src2, $src1|$src1, $src2}",
+                       [(set GR32:$dst,
+                         (int_x86_sse42_crc32_16 GR32:$src1,
+                         (load addr:$src2)))]>,
+                         OpSize;
+  def CRC32r16  : SS42FI<0xF1, MRMSrcReg, (outs GR32:$dst), 
+                      (ins GR32:$src1, GR16:$src2),
+                      "crc32 \t{$src2, $src1|$src1, $src2}",
+                       [(set GR32:$dst,
+                         (int_x86_sse42_crc32_16 GR32:$src1, GR16:$src2))]>, 
+                         OpSize;
+  def CRC32m32  : SS42FI<0xF1, MRMSrcMem, (outs GR32:$dst), 
+                      (ins GR32:$src1, i32mem:$src2),
+                      "crc32 \t{$src2, $src1|$src1, $src2}",
+                       [(set GR32:$dst,
+                         (int_x86_sse42_crc32_32 GR32:$src1,
+                         (load addr:$src2)))]>, OpSize;
+  def CRC32r32  : SS42FI<0xF1, MRMSrcReg, (outs GR32:$dst), 
+                      (ins GR32:$src1, GR32:$src2),
+                      "crc32 \t{$src2, $src1|$src1, $src2}",
+                       [(set GR32:$dst,
+                         (int_x86_sse42_crc32_32 GR32:$src1, GR32:$src2))]>, 
+                         OpSize;
+  def CRC64m64  : SS42FI<0xF0, MRMSrcMem, (outs GR64:$dst), 
+                      (ins GR64:$src1, i64mem:$src2),
+                      "crc32 \t{$src2, $src1|$src1, $src2}",
+                       [(set GR64:$dst,
+                         (int_x86_sse42_crc32_64 GR64:$src1,
+                         (load addr:$src2)))]>,
+                         OpSize, REX_W;
+  def CRC64r64  : SS42FI<0xF0, MRMSrcReg, (outs GR64:$dst), 
+                      (ins GR64:$src1, GR64:$src2),
+                      "crc32 \t{$src2, $src1|$src1, $src2}",
+                       [(set GR64:$dst,
+                         (int_x86_sse42_crc32_64 GR64:$src1, GR64:$src2))]>, 
+                         OpSize, REX_W;
+                         
+  // TODO: These correspond to int_x86_sse42_crc32_8 but with a 64-bit src
+  // and dest, figure it out.
+  //def CRC64m8  : SS42FI<0xF1, MRMSrcMem, (outs GR64:$dst), 
+  //                    (ins GR32:$src1, i8mem:$src2),
+  //                    "crc32 \t{$src2, $src1|$src1, $src2}",
+  //                     [(set GR64:$dst,
+  //                       (int_x86_sse42_crc32_8 GR64:$src1,
+  //                       (load addr:$src2)))]>, 
+  //                       OpSize, REX_W;
+  //def CRC64r8  : SS42FI<0xF1, MRMSrcReg, (outs GR64:$dst), 
+  //                    (ins GR64:$src1, GR8:$src2),
+  //                    "crc32 \t{$src2, $src1|$src1, $src2}",
+  //                     [(set GR64:$dst,
+  //                       (int_x86_sse42_crc32_8 GR32:$src1, GR8:$src2))]>, 
+  //                       OpSize, REX_W;
+}
diff --git a/test/CodeGen/X86/sse42.ll b/test/CodeGen/X86/sse42.ll
new file mode 100644
index 0000000..1652294
--- /dev/null
+++ b/test/CodeGen/X86/sse42.ll
@@ -0,0 +1,38 @@
+; RUN: llvm-as < %s | llc -mtriple=i686-apple-darwin9 -mattr=sse42 | FileCheck %s -check-prefix=X32
+; RUN: llvm-as < %s | llc -mtriple=x86_64-apple-darwin9 -mattr=sse42 | FileCheck %s -check-prefix=X64
+
+declare i32 @llvm.x86.sse42.crc32.8(i32, i8) nounwind
+declare i32 @llvm.x86.sse42.crc32.16(i32, i16) nounwind
+declare i32 @llvm.x86.sse42.crc32.32(i32, i32) nounwind
+
+define i32 @crc32_8(i32 %a, i8 %b) nounwind {
+  %tmp = call i32 @llvm.x86.sse42.crc32.8(i32 %a, i8 %b)
+  ret i32 %tmp
+; X32: _crc32_8:
+; X32:     crc32   8(%esp), %eax
+
+; X64: _crc32_8:
+; X64:     crc32   %sil, %eax
+}
+
+
+define i32 @crc32_16(i32 %a, i16 %b) nounwind {
+  %tmp = call i32 @llvm.x86.sse42.crc32.16(i32 %a, i16 %b)
+  ret i32 %tmp
+; X32: _crc32_16:
+; X32:     crc32   8(%esp), %eax
+
+; X64: _crc32_16:
+; X64:     crc32   %si, %eax
+}
+
+
+define i32 @crc32_32(i32 %a, i32 %b) nounwind {
+  %tmp = call i32 @llvm.x86.sse42.crc32.32(i32 %a, i32 %b)
+  ret i32 %tmp
+; X32: _crc32_32:
+; X32:     crc32   8(%esp), %eax
+
+; X64: _crc32_32:
+; X64:     crc32   %esi, %eax
+}