[llvm-bugs] [Bug 46244] New: Clang does not propagate range when using uint8_t type while gcc does.

via llvm-bugs llvm-bugs at lists.llvm.org
Mon Jun 8 09:12:13 PDT 2020


https://bugs.llvm.org/show_bug.cgi?id=46244

            Bug ID: 46244
           Summary: Clang does not propagate range when using uint8_t type
                    while gcc does.
           Product: libraries
           Version: trunk
          Hardware: PC
                OS: All
            Status: NEW
          Severity: enhancement
          Priority: P
         Component: Scalar Optimizations
          Assignee: unassignedbugs at nondot.org
          Reporter: jmuizelaar at mozilla.com
                CC: llvm-bugs at lists.llvm.org

The following program compiles to:

#include <stdint.h>
#include <stdlib.h>
int foo(uint8_t *a , int q) {
    if (q == 5) {
        *a = 1;
    } else {
        *a = 2;
    };
    switch (*a) {
        case 1: return 1;
        case 2: return 2;
        default: exit(1);
    }
}

compiles to:
foo(unsigned char*, int):                              # @foo(unsigned char*,
int)
        pushq   %rax
        xorl    %ecx, %ecx
        cmpl    $5, %esi
        sete    %cl
        movb    $2, %dl
        subb    %cl, %dl
        movb    %dl, (%rdi)
        movl    $2, %eax
        subl    %ecx, %eax
        cmpb    $1, %dl
        je      .LBB0_3
        cmpl    $2, %eax
        jne     .LBB0_4
        movl    $2, %eax
.LBB0_3:
        popq    %rcx
        retq
.LBB0_4:
        movl    $1, %edi
        callq   exit

GCC compiles it to:
foo(unsigned char*, int):
        cmpl    $5, %esi
        setne   %dl
        xorl    %eax, %eax
        addl    $1, %edx
        cmpl    $5, %esi
        setne   %al
        movb    %dl, (%rdi)
        addl    $1, %eax
        ret

Changing the type of `a` from `uint8_t *` to `int *` fixes the problem.

-- 
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/20200608/5d88004f/attachment.html>


More information about the llvm-bugs mailing list