[Bug c/60700] New: missing dependency between %ax and %eax when compiling 32bit on 64bit

[yzhou61 at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=60700

            Bug ID: 60700
           Summary: missing dependency between %ax and %eax when compiling
                    32bit on 64bit
           Product: gcc
           Version: 4.8.2
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c
          Assignee: unassigned at gcc dot gnu.org
          Reporter: yzhou61 at gmail dot com

When compiling with -m32 on a 64bit machine, gcc is generating wrong code for
the following snippet. The dependencies between %ax and %eax seems to have been
dropped, causing the memset to use the wrong value. Please fix. Thanks.

$ cat repro.c 
#include <stdlib.h>
#include <string.h>

extern int foo(void);
void *g = (void *)1;

struct st {
    char data[36]; // must be greater than 32
};

int repro(struct st **out)
{
    int status = 0;

    *out = NULL;

    status = foo();
    if (status != 0) {
        return status;
    }

    if (NULL == g) {
        status = 999;
        return status;
    }

    *out = (struct st *)malloc(sizeof(struct st));
    if (NULL == (*out)) {
        status = 42;
        return status;
    }

    memset(*out, 0, sizeof(struct st));

    return status;
}

$ gcc -c -o repro.o repro.c -m32 -march=i686 -O3 -I. -Wall -Wextra
-fno-strict-aliasing -fwrapv -fno-aggressive-loop-optimizations -save-temps
$ cat repro.s
        .file   "repro.c"
        .text
        .p2align 4,,15
        .globl  repro
        .type   repro, @function
repro:
.LFB19:
        .cfi_startproc
        pushl   %edi
        .cfi_def_cfa_offset 8
        .cfi_offset 7, -8
        pushl   %esi
        .cfi_def_cfa_offset 12
        .cfi_offset 6, -12
        pushl   %ebx
        .cfi_def_cfa_offset 16
        .cfi_offset 3, -16
        subl    $16, %esp
        .cfi_def_cfa_offset 32
        movl    32(%esp), %ebx
        movl    $0, (%ebx)
        call    foo
        testl   %eax, %eax
        jne     .L2
        movl    g, %edx
        movw    $999, %ax
        testl   %edx, %edx
        je      .L2
        movl    $36, (%esp)
        movl    %eax, %esi
        call    malloc
        movl    %eax, %edx
        testl   %edx, %edx
        movl    %eax, (%ebx)
        movl    $42, %eax
        je      .L2
        movl    %esi, %eax
        movl    $9, %ecx
        movl    %edx, %edi
        rep; stosl
        xorl    %eax, %eax
        .p2align 4,,7
        .p2align 3
.L2:
        addl    $16, %esp
        .cfi_def_cfa_offset 16
        popl    %ebx
        .cfi_restore 3
        .cfi_def_cfa_offset 12
        popl    %esi
        .cfi_restore 6
        .cfi_def_cfa_offset 8
        popl    %edi
        .cfi_restore 7
        .cfi_def_cfa_offset 4
        ret
        .cfi_endproc
.LFE19:
        .size   repro, .-repro
        .globl  g
        .data
        .align 4
        .type   g, @object
        .size   g, 4
g:
        .long   1
        .ident  "GCC: (GNU) 4.8.2"
        .section        .note.GNU-stack,"",@progbits

$ gcc -v                                                                        
Using built-in specs.
COLLECT_GCC=gcc
COLLECT_LTO_WRAPPER=/usr/local/libexec/gcc/x86_64-unknown-linux-gnu/4.8.2/lto-wrapper
Target: x86_64-unknown-linux-gnu
Configured with: ./configure
Thread model: posix
gcc version 4.8.2 (GCC)