[clang] [clang][x86] Ensure we use the shifted value bit width to check for out of bounds per-element shift amounts (PR #156019)

[Simon Pilgrim via cfe-commits]

================
@@ -3256,8 +3256,8 @@ bool InterpretBuiltin(InterpState &S, CodePtr OpPC, const 
CallExpr *Call,
   case clang::X86::BI__builtin_ia32_psllv8si:
----------------
RKSimon wrote:

@Arghnews with this change you should be able to add the slli cases and reuse 
this code (same for the srli/srai cases as well below):
```
  case clang::X86::BI__builtin_ia32_psllwi128:
  case clang::X86::BI__builtin_ia32_psllwi256:
  case clang::X86::BI__builtin_ia32_psllwi512:
  case clang::X86::BI__builtin_ia32_pslldi128:
  case clang::X86::BI__builtin_ia32_pslldi256:
  case clang::X86::BI__builtin_ia32_pslldi512:
  case clang::X86::BI__builtin_ia32_psllqi128:
  case clang::X86::BI__builtin_ia32_psllqi256:
  case clang::X86::BI__builtin_ia32_psllqi512:
```

https://github.com/llvm/llvm-project/pull/156019
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits