[X86] Emit REX prefix immediately before the opcode

Fix prefix emission order to emit REX immediately before the opcode (SDM vol2, 2.1, Figure 2-1). According to SDM vol2 2.2.1, "Other placements are ignored". This fix has a side effect of outputting segment override prefix in a different order than previously (benign). Follow-up to https://reviews.llvm.org/D120592 Reviewed By: skan, craig.topper Differential Revision: https://reviews.llvm.org/D120871
llvm · Mar 16, 2022 · 2c4e38f · 2c4e38f
1 parent 1d37198
commit 2c4e38f
Show file tree

Hide file tree

Showing 11 changed files with 78 additions and 77 deletions.
diff --git a/llvm/lib/Target/X86/MCTargetDesc/X86MCCodeEmitter.cpp b/llvm/lib/Target/X86/MCTargetDesc/X86MCCodeEmitter.cpp
@@ -617,14 +617,6 @@ bool X86MCCodeEmitter::emitPrefixImpl(unsigned &CurOp, const MCInst &MI,
       Flags & X86::IP_HAS_AD_SIZE)
     emitByte(0x67, OS);
 
-  // Encoding type for this instruction.
-  uint64_t Encoding = TSFlags & X86II::EncodingMask;
-  bool HasREX = false;
-  if (Encoding)
-    emitVEXOpcodePrefix(MemoryOperand, MI, OS);
-  else
-    HasREX = emitOpcodePrefix(MemoryOperand, MI, STI, OS);
-
   uint64_t Form = TSFlags & X86II::FormMask;
   switch (Form) {
   default:
@@ -654,6 +646,15 @@ bool X86MCCodeEmitter::emitPrefixImpl(unsigned &CurOp, const MCInst &MI,
   }
   }
 
+  // REX prefix is optional, but if used must be immediately before the opcode
+  // Encoding type for this instruction.
+  uint64_t Encoding = TSFlags & X86II::EncodingMask;
+  bool HasREX = false;
+  if (Encoding)
+    emitVEXOpcodePrefix(MemoryOperand, MI, OS);
+  else
+    HasREX = emitOpcodePrefix(MemoryOperand, MI, STI, OS);
+
   return HasREX;
 }
 

diff --git a/llvm/test/MC/Disassembler/X86/moffs.txt b/llvm/test/MC/Disassembler/X86/moffs.txt
@@ -37,9 +37,9 @@
 # 64: movw 0x5a5a5a5a, %ax # encoding: [0x67,0x66,0xa1,0x5a,0x5a,0x5a,0x5a]
 0x67 0x66 0xa1 0x5a 0x5a 0x5a 0x5a 0x5a 0x5a 0x5a 0x5a
 
-# 16: movl %es:0x5a5a5a5a, %eax # encoding: [0x67,0x66,0x26,0xa1,0x5a,0x5a,0x5a,0x5a]
-# 32: movw %es:0x5a5a, %ax # encoding: [0x67,0x66,0x26,0xa1,0x5a,0x5a]
-# 64: movw %es:0x5a5a5a5a, %ax # encoding: [0x67,0x66,0x26,0xa1,0x5a,0x5a,0x5a,0x5a]
+# 16: movl %es:0x5a5a5a5a, %eax # encoding: [0x67,0x26,0x66,0xa1,0x5a,0x5a,0x5a,0x5a]
+# 32: movw %es:0x5a5a, %ax # encoding: [0x67,0x26,0x66,0xa1,0x5a,0x5a]
+# 64: movw %es:0x5a5a5a5a, %ax # encoding: [0x67,0x26,0x66,0xa1,0x5a,0x5a,0x5a,0x5a]
 0x67 0x26 0x66 0xa1 0x5a 0x5a 0x5a 0x5a 0x5a 0x5a 0x5a 0x5a # encoding: [0xa0,0x5a,0x5a]
 
 
@@ -79,8 +79,8 @@
 # 64: movw %ax, 0x5a5a5a5a # encoding: [0x67,0x66,0xa3,0x5a,0x5a,0x5a,0x5a]
 0x67 0x66 0xa3 0x5a 0x5a 0x5a 0x5a 0x5a 0x5a 0x5a 0x5a
 
-# 16: movl %eax, %es:0x5a5a5a5a # encoding: [0x67,0x66,0x26,0xa3,0x5a,0x5a,0x5a,0x5a]
-# 32: movw %ax, %es:0x5a5a # encoding: [0x67,0x66,0x26,0xa3,0x5a,0x5a]
-# 64: movw %ax, %es:0x5a5a5a5a # encoding: [0x67,0x66,0x26,0xa3,0x5a,0x5a,0x5a,0x5a]
+# 16: movl %eax, %es:0x5a5a5a5a # encoding: [0x67,0x26,0x66,0xa3,0x5a,0x5a,0x5a,0x5a]
+# 32: movw %ax, %es:0x5a5a # encoding: [0x67,0x26,0x66,0xa3,0x5a,0x5a]
+# 64: movw %ax, %es:0x5a5a5a5a # encoding: [0x67,0x26,0x66,0xa3,0x5a,0x5a,0x5a,0x5a]
 0x67 0x26 0x66 0xa3 0x5a 0x5a 0x5a 0x5a 0x5a 0x5a 0x5a 0x5a
 
diff --git a/llvm/test/MC/X86/I186-32.s b/llvm/test/MC/X86/I186-32.s
@@ -57,7 +57,7 @@ outsb %es:(%esi), %dx
 outsl %es:(%esi), %dx 
 
 // CHECK: outsw %es:(%esi), %dx 
-// CHECK: encoding: [0x66,0x26,0x6f]        
+// CHECK: encoding: [0x26,0x66,0x6f]        
 outsw %es:(%esi), %dx 
 
 // CHECK: popal 
@@ -269,7 +269,7 @@ repne outsb %es:(%esi), %dx
 repne outsl %es:(%esi), %dx 
 
 // CHECK: repne outsw %es:(%esi), %dx 
-// CHECK: encoding: [0xf2,0x66,0x26,0x6f]       
+// CHECK: encoding: [0xf2,0x26,0x66,0x6f]       
 repne outsw %es:(%esi), %dx 
 
 // CHECK: rep outsb %es:(%esi), %dx 
@@ -281,7 +281,7 @@ rep outsb %es:(%esi), %dx
 rep outsl %es:(%esi), %dx 
 
 // CHECK: rep outsw %es:(%esi), %dx 
-// CHECK: encoding: [0xf3,0x66,0x26,0x6f]       
+// CHECK: encoding: [0xf3,0x26,0x66,0x6f]       
 rep outsw %es:(%esi), %dx 
 
 // CHECK: rolb $0, -485498096(%edx,%eax,4) 

diff --git a/llvm/test/MC/X86/I186-64.s b/llvm/test/MC/X86/I186-64.s
@@ -33,7 +33,7 @@ outsb %gs:(%rsi), %dx
 outsl %gs:(%rsi), %dx 
 
 // CHECK: outsw %gs:(%rsi), %dx 
-// CHECK: encoding: [0x66,0x65,0x6f]        
+// CHECK: encoding: [0x65,0x66,0x6f]        
 outsw %gs:(%rsi), %dx 
 
 // CHECK: pushq $0 
@@ -289,7 +289,7 @@ repne outsb %gs:(%rsi), %dx
 repne outsl %gs:(%rsi), %dx 
 
 // CHECK: repne outsw %gs:(%rsi), %dx 
-// CHECK: encoding: [0xf2,0x66,0x65,0x6f]       
+// CHECK: encoding: [0xf2,0x65,0x66,0x6f]       
 repne outsw %gs:(%rsi), %dx 
 
 // CHECK: rep outsb %gs:(%rsi), %dx 
@@ -301,7 +301,7 @@ rep outsb %gs:(%rsi), %dx
 rep outsl %gs:(%rsi), %dx 
 
 // CHECK: rep outsw %gs:(%rsi), %dx 
-// CHECK: encoding: [0xf3,0x66,0x65,0x6f]       
+// CHECK: encoding: [0xf3,0x65,0x66,0x6f]       
 rep outsw %gs:(%rsi), %dx 
 
 // CHECK: rolb $0, 485498096 

diff --git a/llvm/test/MC/X86/I386-32.s b/llvm/test/MC/X86/I386-32.s
@@ -153,7 +153,7 @@ cmpsb %es:(%edi), %es:(%esi)
 cmpsl %es:(%edi), %es:(%esi) 
 
 // CHECK: cmpsw %es:(%edi), %es:(%esi) 
-// CHECK: encoding: [0x66,0x26,0xa7]        
+// CHECK: encoding: [0x26,0x66,0xa7]        
 cmpsw %es:(%edi), %es:(%esi) 
 
 // CHECK: insb %dx, %es:(%edi) 
@@ -193,7 +193,7 @@ movsb %es:(%esi), %es:(%edi)
 movsl %es:(%esi), %es:(%edi) 
 
 // CHECK: movsw %es:(%esi), %es:(%edi) 
-// CHECK: encoding: [0x66,0x26,0xa5]        
+// CHECK: encoding: [0x26,0x66,0xa5]        
 movsw %es:(%esi), %es:(%edi) 
 
 // CHECK: outsb %es:(%esi), %dx 
@@ -205,7 +205,7 @@ outsb %es:(%esi), %dx
 outsl %es:(%esi), %dx 
 
 // CHECK: outsw %es:(%esi), %dx 
-// CHECK: encoding: [0x66,0x26,0x6f]        
+// CHECK: encoding: [0x26,0x66,0x6f]        
 outsw %es:(%esi), %dx 
 
 // CHECK: popal 
@@ -249,7 +249,7 @@ rep cmpsb %es:(%edi), %es:(%esi)
 rep cmpsl %es:(%edi), %es:(%esi) 
 
 // CHECK: rep cmpsw %es:(%edi), %es:(%esi) 
-// CHECK: encoding: [0xf3,0x66,0x26,0xa7]       
+// CHECK: encoding: [0xf3,0x26,0x66,0xa7]       
 rep cmpsw %es:(%edi), %es:(%esi) 
 
 // CHECK: rep insb %dx, %es:(%edi) 
@@ -277,7 +277,7 @@ rep movsb %es:(%esi), %es:(%edi)
 rep movsl %es:(%esi), %es:(%edi) 
 
 // CHECK: rep movsw %es:(%esi), %es:(%edi) 
-// CHECK: encoding: [0xf3,0x66,0x26,0xa5]       
+// CHECK: encoding: [0xf3,0x26,0x66,0xa5]       
 rep movsw %es:(%esi), %es:(%edi) 
 
 // CHECK: repne cmpsb %es:(%edi), %es:(%esi) 
@@ -289,7 +289,7 @@ repne cmpsb %es:(%edi), %es:(%esi)
 repne cmpsl %es:(%edi), %es:(%esi) 
 
 // CHECK: repne cmpsw %es:(%edi), %es:(%esi) 
-// CHECK: encoding: [0xf2,0x66,0x26,0xa7]       
+// CHECK: encoding: [0xf2,0x26,0x66,0xa7]       
 repne cmpsw %es:(%edi), %es:(%esi) 
 
 // CHECK: repne insb %dx, %es:(%edi) 
@@ -317,7 +317,7 @@ repne movsb %es:(%esi), %es:(%edi)
 repne movsl %es:(%esi), %es:(%edi) 
 
 // CHECK: repne movsw %es:(%esi), %es:(%edi) 
-// CHECK: encoding: [0xf2,0x66,0x26,0xa5]       
+// CHECK: encoding: [0xf2,0x26,0x66,0xa5]       
 repne movsw %es:(%esi), %es:(%edi) 
 
 // CHECK: repne outsb %es:(%esi), %dx 
@@ -329,7 +329,7 @@ repne outsb %es:(%esi), %dx
 repne outsl %es:(%esi), %dx 
 
 // CHECK: repne outsw %es:(%esi), %dx 
-// CHECK: encoding: [0xf2,0x66,0x26,0x6f]       
+// CHECK: encoding: [0xf2,0x26,0x66,0x6f]       
 repne outsw %es:(%esi), %dx 
 
 // CHECK: repne scasl %es:(%edi), %eax 
@@ -349,7 +349,7 @@ rep outsb %es:(%esi), %dx
 rep outsl %es:(%esi), %dx 
 
 // CHECK: rep outsw %es:(%esi), %dx 
-// CHECK: encoding: [0xf3,0x66,0x26,0x6f]       
+// CHECK: encoding: [0xf3,0x26,0x66,0x6f]       
 rep outsw %es:(%esi), %dx 
 
 // CHECK: rep scasl %es:(%edi), %eax 

diff --git a/llvm/test/MC/X86/I386-64.s b/llvm/test/MC/X86/I386-64.s
@@ -226,11 +226,11 @@ cmpsb %es:(%rdi), %gs:(%rsi)
 cmpsl %es:(%rdi), %gs:(%rsi) 
 
 // CHECK: cmpsq %es:(%rdi), %gs:(%rsi) 
-// CHECK: encoding: [0x48,0x65,0xa7]        
+// CHECK: encoding: [0x65,0x48,0xa7]        
 cmpsq %es:(%rdi), %gs:(%rsi) 
 
 // CHECK: cmpsw %es:(%rdi), %gs:(%rsi) 
-// CHECK: encoding: [0x66,0x65,0xa7]        
+// CHECK: encoding: [0x65,0x66,0xa7]        
 cmpsw %es:(%rdi), %gs:(%rsi) 
 
 // CHECK: insb %dx, %es:(%rdi) 
@@ -306,11 +306,11 @@ movsbl (%rdx), %r13d
 movsl %gs:(%rsi), %es:(%rdi) 
 
 // CHECK: movsq %gs:(%rsi), %es:(%rdi) 
-// CHECK: encoding: [0x48,0x65,0xa5]        
+// CHECK: encoding: [0x65,0x48,0xa5]        
 movsq %gs:(%rsi), %es:(%rdi) 
 
 // CHECK: movsw %gs:(%rsi), %es:(%rdi) 
-// CHECK: encoding: [0x66,0x65,0xa5]        
+// CHECK: encoding: [0x65,0x66,0xa5]        
 movsw %gs:(%rsi), %es:(%rdi) 
 
 // CHECK: movswl 485498096, %r13d 
@@ -418,7 +418,7 @@ outsb %gs:(%rsi), %dx
 outsl %gs:(%rsi), %dx 
 
 // CHECK: outsw %gs:(%rsi), %dx 
-// CHECK: encoding: [0x66,0x65,0x6f]        
+// CHECK: encoding: [0x65,0x66,0x6f]        
 outsw %gs:(%rsi), %dx 
 
 // CHECK: rep cmpsb %es:(%rdi), %gs:(%rsi) 
@@ -430,11 +430,11 @@ rep cmpsb %es:(%rdi), %gs:(%rsi)
 rep cmpsl %es:(%rdi), %gs:(%rsi) 
 
 // CHECK: rep cmpsq %es:(%rdi), %gs:(%rsi) 
-// CHECK: encoding: [0xf3,0x48,0x65,0xa7]       
+// CHECK: encoding: [0xf3,0x65,0x48,0xa7]       
 rep cmpsq %es:(%rdi), %gs:(%rsi) 
 
 // CHECK: rep cmpsw %es:(%rdi), %gs:(%rsi) 
-// CHECK: encoding: [0xf3,0x66,0x65,0xa7]       
+// CHECK: encoding: [0xf3,0x65,0x66,0xa7]       
 rep cmpsw %es:(%rdi), %gs:(%rsi) 
 
 // CHECK: rep insb %dx, %es:(%rdi) 
@@ -462,11 +462,11 @@ rep movsb %gs:(%rsi), %es:(%rdi)
 rep movsl %gs:(%rsi), %es:(%rdi) 
 
 // CHECK: rep movsq %gs:(%rsi), %es:(%rdi) 
-// CHECK: encoding: [0xf3,0x48,0x65,0xa5]       
+// CHECK: encoding: [0xf3,0x65,0x48,0xa5]       
 rep movsq %gs:(%rsi), %es:(%rdi) 
 
 // CHECK: rep movsw %gs:(%rsi), %es:(%rdi) 
-// CHECK: encoding: [0xf3,0x66,0x65,0xa5]       
+// CHECK: encoding: [0xf3,0x65,0x66,0xa5]       
 rep movsw %gs:(%rsi), %es:(%rdi) 
 
 // CHECK: repne cmpsb %es:(%rdi), %gs:(%rsi) 
@@ -478,11 +478,11 @@ repne cmpsb %es:(%rdi), %gs:(%rsi)
 repne cmpsl %es:(%rdi), %gs:(%rsi) 
 
 // CHECK: repne cmpsq %es:(%rdi), %gs:(%rsi) 
-// CHECK: encoding: [0xf2,0x48,0x65,0xa7]       
+// CHECK: encoding: [0xf2,0x65,0x48,0xa7]       
 repne cmpsq %es:(%rdi), %gs:(%rsi) 
 
 // CHECK: repne cmpsw %es:(%rdi), %gs:(%rsi) 
-// CHECK: encoding: [0xf2,0x66,0x65,0xa7]       
+// CHECK: encoding: [0xf2,0x65,0x66,0xa7]       
 repne cmpsw %es:(%rdi), %gs:(%rsi) 
 
 // CHECK: repne insb %dx, %es:(%rdi) 
@@ -510,11 +510,11 @@ repne movsb %gs:(%rsi), %es:(%rdi)
 repne movsl %gs:(%rsi), %es:(%rdi) 
 
 // CHECK: repne movsq %gs:(%rsi), %es:(%rdi) 
-// CHECK: encoding: [0xf2,0x48,0x65,0xa5]       
+// CHECK: encoding: [0xf2,0x65,0x48,0xa5]       
 repne movsq %gs:(%rsi), %es:(%rdi) 
 
 // CHECK: repne movsw %gs:(%rsi), %es:(%rdi) 
-// CHECK: encoding: [0xf2,0x66,0x65,0xa5]       
+// CHECK: encoding: [0xf2,0x65,0x66,0xa5]       
 repne movsw %gs:(%rsi), %es:(%rdi) 
 
 // CHECK: repne outsb %gs:(%rsi), %dx 
@@ -526,7 +526,7 @@ repne outsb %gs:(%rsi), %dx
 repne outsl %gs:(%rsi), %dx 
 
 // CHECK: repne outsw %gs:(%rsi), %dx 
-// CHECK: encoding: [0xf2,0x66,0x65,0x6f]       
+// CHECK: encoding: [0xf2,0x65,0x66,0x6f]       
 repne outsw %gs:(%rsi), %dx 
 
 // CHECK: repne scasl %es:(%rdi), %eax 
@@ -546,7 +546,7 @@ rep outsb %gs:(%rsi), %dx
 rep outsl %gs:(%rsi), %dx 
 
 // CHECK: rep outsw %gs:(%rsi), %dx 
-// CHECK: encoding: [0xf3,0x66,0x65,0x6f]       
+// CHECK: encoding: [0xf3,0x65,0x66,0x6f]       
 rep outsw %gs:(%rsi), %dx 
 
 // CHECK: rep scasl %es:(%rdi), %eax 

diff --git a/llvm/test/MC/X86/I86-32.s b/llvm/test/MC/X86/I86-32.s
@@ -565,7 +565,7 @@ cmpsb %es:(%edi), %es:(%esi)
 cmpsl %es:(%edi), %es:(%esi) 
 
 // CHECK: cmpsw %es:(%edi), %es:(%esi) 
-// CHECK: encoding: [0x66,0x26,0xa7]        
+// CHECK: encoding: [0x26,0x66,0xa7]        
 cmpsw %es:(%edi), %es:(%esi) 
 
 // CHECK: cmpw $0, -485498096(%edx,%eax,4) 
@@ -1197,7 +1197,7 @@ lock xchgl %eax, (%edx)
 lodsb %es:(%esi), %al 
 
 // CHECK: lodsw %es:(%esi), %ax 
-// CHECK: encoding: [0x66,0x26,0xad]        
+// CHECK: encoding: [0x26,0x66,0xad]        
 lodsw %es:(%esi), %ax 
 
 // CHECK: loop 64 
@@ -1345,7 +1345,7 @@ movsb %es:(%esi), %es:(%edi)
 movsl %es:(%esi), %es:(%edi) 
 
 // CHECK: movsw %es:(%esi), %es:(%edi) 
-// CHECK: encoding: [0x66,0x26,0xa5]        
+// CHECK: encoding: [0x26,0x66,0xa5]        
 movsw %es:(%esi), %es:(%edi) 
 
 // CHECK: movw $0, -485498096(%edx,%eax,4) 
@@ -2221,15 +2221,15 @@ rep cmpsb %es:(%edi), %es:(%esi)
 rep cmpsl %es:(%edi), %es:(%esi) 
 
 // CHECK: rep cmpsw %es:(%edi), %es:(%esi) 
-// CHECK: encoding: [0xf3,0x66,0x26,0xa7]       
+// CHECK: encoding: [0xf3,0x26,0x66,0xa7]       
 rep cmpsw %es:(%edi), %es:(%esi) 
 
 // CHECK: rep lodsb %es:(%esi), %al 
 // CHECK: encoding: [0xf3,0x26,0xac]       
 rep lodsb %es:(%esi), %al 
 
 // CHECK: rep lodsw %es:(%esi), %ax 
-// CHECK: encoding: [0xf3,0x66,0x26,0xad]       
+// CHECK: encoding: [0xf3,0x26,0x66,0xad]       
 rep lodsw %es:(%esi), %ax 
 
 // CHECK: rep movsb %es:(%esi), %es:(%edi) 
@@ -2241,7 +2241,7 @@ rep movsb %es:(%esi), %es:(%edi)
 rep movsl %es:(%esi), %es:(%edi) 
 
 // CHECK: rep movsw %es:(%esi), %es:(%edi) 
-// CHECK: encoding: [0xf3,0x66,0x26,0xa5]       
+// CHECK: encoding: [0xf3,0x26,0x66,0xa5]       
 rep movsw %es:(%esi), %es:(%edi) 
 
 // CHECK: repne cmpsb %es:(%edi), %es:(%esi) 
@@ -2253,15 +2253,15 @@ repne cmpsb %es:(%edi), %es:(%esi)
 repne cmpsl %es:(%edi), %es:(%esi) 
 
 // CHECK: repne cmpsw %es:(%edi), %es:(%esi) 
-// CHECK: encoding: [0xf2,0x66,0x26,0xa7]       
+// CHECK: encoding: [0xf2,0x26,0x66,0xa7]       
 repne cmpsw %es:(%edi), %es:(%esi) 
 
 // CHECK: repne lodsb %es:(%esi), %al 
 // CHECK: encoding: [0xf2,0x26,0xac]       
 repne lodsb %es:(%esi), %al 
 
 // CHECK: repne lodsw %es:(%esi), %ax 
-// CHECK: encoding: [0xf2,0x66,0x26,0xad]       
+// CHECK: encoding: [0xf2,0x26,0x66,0xad]       
 repne lodsw %es:(%esi), %ax 
 
 // CHECK: repne movsb %es:(%esi), %es:(%edi) 
@@ -2273,7 +2273,7 @@ repne movsb %es:(%esi), %es:(%edi)
 repne movsl %es:(%esi), %es:(%edi) 
 
 // CHECK: repne movsw %es:(%esi), %es:(%edi) 
-// CHECK: encoding: [0xf2,0x66,0x26,0xa5]       
+// CHECK: encoding: [0xf2,0x26,0x66,0xa5]       
 repne movsw %es:(%esi), %es:(%edi) 
 
 // CHECK: repne scasb %es:(%edi), %al