Re: [PATCH v3 0/4] MR375: ntdll: Fixes for runtime 64-bit shift functions.

[Zebediah Figura]

On 7/7/22 06:58, Jinoh Kang (@iamahuman) wrote:
> Jinoh Kang (@iamahuman) commented about dlls/ntdll/tests/large_int.c:
> >   LONGLONG l;
> >
> > - static const BYTE call_shift_code[] =
> > - {
> > -
> >    0x55,                           /* pushl %ebp */
> > -
> >    0x89, 0xe5,                     /* movl %esp,%ebp */
> > -
> >    0x31, 0xc0,                     /* xorl %eax,%eax */
> > -
> >    0x31, 0xd2,                     /* xorl %edx,%edx */
> > -
> >    0x31, 0xc9,                     /* xorl %ecx,%ecx */
> > -
> >    0x87, 0x45, 0x0c,               /* xchgl 12(%ebp),%eax */
> > -
> >    0x87, 0x55, 0x10,               /* xchgl 16(%ebp),%edx */
> > -
> >    0x87, 0x4d, 0x14,               /* xchgl 20(%ebp),%ecx */
> > -
> >    0xff, 0x55, 0x08,               /* call *8(%ebp) */
> > -
> >    0x39, 0xe5,                     /* cmpl %esp,%ebp */
> > -
> >    0x75, 0x05,                     /* jne 1f */
> > -
> >    0xb9, 0xef, 0xbe, 0xad, 0xde,   /* movl $0xdeadbeef,%ecx */
>
> This won't affect the test outcome.  How about:
>          0x74, 0x05,                     /* je 1f */
>          0xb8, 0xef, 0xbe, 0xad, 0xde,   /* movl $0xdeadbeef,%eax */

Yes, that is what I meant to write; thanks for catching it. I've sent a 
new revision.