[llvm-bugs] [Bug 32987] New: [inline asm] dot operator while using imm generates wrong ir + asm

[via llvm-bugs]

https://bugs.llvm.org/show_bug.cgi?id=32987

            Bug ID: 32987
           Summary: [inline asm] dot operator while using imm generates
                    wrong ir + asm
           Product: new-bugs
           Version: unspecified
          Hardware: PC
                OS: All
            Status: NEW
          Severity: enhancement
          Priority: P
         Component: new bugs
          Assignee: unassignedbugs at nondot.org
          Reporter: ziv.izhar at intel.com
                CC: llvm-bugs at lists.llvm.org

When using both imm and dot operator in an inline asm operand the generated ir
and asm are wrong.
Consider the code:

typedef struct foo {
   int x;
   int y;
} foo;
int main(){
  __asm {mov eax, 3[rbx]foo.y}
  return 0;
}

compiling with clang -cc1 -fasm-blocks -emit-llvm gives us:
call i32 asm sideeffect inteldialect "mov eax, $$3[ebx].11",
"={eax},~{eax},~{dirflag},~{fpsr},~{flags}"

the foo.y was rewrited as .11 (8+3 instead of 8) while the immediate is still
in the ir - $$3.

when compiling to asm, it's being resolved as 11+3 = 14, instead of 8+3=11.

-- 
You are receiving this mail because:
You are on the CC list for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-bugs/attachments/20170510/f6fa5790/attachment-0001.html>