[llvm] a073cbb - [X86] Fix misassemble due to not storing registers to state machine on RParen (#150252)

[via llvm-commits]

Author: circuit10
Date: 2025-07-24T10:38:16+01:00
New Revision: a073cbbb1aeaaeac01b12e818fe47e4c04080aac

URL: https://github.com/llvm/llvm-project/commit/a073cbbb1aeaaeac01b12e818fe47e4c04080aac
DIFF: https://github.com/llvm/llvm-project/commit/a073cbbb1aeaaeac01b12e818fe47e4c04080aac.diff

LOG: [X86] Fix misassemble due to not storing registers to state machine on RParen (#150252)

This fixes #116883.

The x86 parser saves any register it encounters to a TmpReg field in its
state machine, then on encountering the next valid token immediately
afterwards saves it to either BaseReg, or IndexReg if BaseReg was
already filled. However, this saving logic was missing on the RParen
token handler, causing the parser to "forget" the register immediately
beforehand. This also would prevent later validation logic from
detecting the addressing mode as invalid, leading to a silent
misassembly rather than an error.

Added: 
    llvm/test/MC/X86/intel-syntax-parentheses.s

Modified: 
    llvm/lib/Target/X86/AsmParser/X86AsmParser.cpp

Removed: 
    


################################################################################
diff  --git a/llvm/lib/Target/X86/AsmParser/X86AsmParser.cpp b/llvm/lib/Target/X86/AsmParser/X86AsmParser.cpp
index b642c1cfe383b..8213e512f45e1 100644
--- a/llvm/lib/Target/X86/AsmParser/X86AsmParser.cpp
+++ b/llvm/lib/Target/X86/AsmParser/X86AsmParser.cpp
@@ -1042,8 +1042,8 @@ class X86AsmParser : public MCTargetAsmParser {
       }
       PrevState = CurrState;
     }
-    void onRParen() {
-      PrevState = State;
+    bool onRParen(StringRef &ErrMsg) {
+      IntelExprState CurrState = State;
       switch (State) {
       default:
         State = IES_ERROR;
@@ -1054,9 +1054,27 @@ class X86AsmParser : public MCTargetAsmParser {
       case IES_RBRAC:
       case IES_RPAREN:
         State = IES_RPAREN;
+        // In the case of a multiply, onRegister has already set IndexReg
+        // directly, with appropriate scale.
+        // Otherwise if we just saw a register it has only been stored in
+        // TmpReg, so we need to store it into the state machine.
+        if (CurrState == IES_REGISTER && PrevState != IES_MULTIPLY) {
+          // If we already have a BaseReg, then assume this is the IndexReg with
+          // no explicit scale.
+          if (!BaseReg) {
+            BaseReg = TmpReg;
+          } else {
+            if (IndexReg)
+              return regsUseUpError(ErrMsg);
+            IndexReg = TmpReg;
+            Scale = 0;
+          }
+        }
         IC.pushOperator(IC_RPAREN);
         break;
       }
+      PrevState = CurrState;
+      return false;
     }
     bool onOffset(const MCExpr *Val, SMLoc OffsetLoc, StringRef ID,
                   const InlineAsmIdentifierInfo &IDInfo,
@@ -2172,7 +2190,11 @@ bool X86AsmParser::ParseIntelExpression(IntelExprStateMachine &SM, SMLoc &End) {
       }
       break;
     case AsmToken::LParen:  SM.onLParen(); break;
-    case AsmToken::RParen:  SM.onRParen(); break;
+    case AsmToken::RParen:
+      if (SM.onRParen(ErrMsg)) {
+        return Error(Tok.getLoc(), ErrMsg);
+      }
+      break;
     }
     if (SM.hadError())
       return Error(Tok.getLoc(), "unknown token in expression");

diff  --git a/llvm/test/MC/X86/intel-syntax-parentheses.s b/llvm/test/MC/X86/intel-syntax-parentheses.s
new file mode 100644
index 0000000000000..ae53f64089070
--- /dev/null
+++ b/llvm/test/MC/X86/intel-syntax-parentheses.s
@@ -0,0 +1,10 @@
+// RUN: not llvm-mc -triple x86_64-unknown-unknown %s 2>&1 | FileCheck %s
+
+.intel_syntax
+
+// CHECK: error: invalid base+index expression
+    lea     rdi, [(label + rsi) + rip]
+// CHECK: leaq    1(%rax,%rdi), %rdi
+    lea     rdi, [(rax + rdi) + 1]
+label:
+    .quad 42