[MPlayer-users] Re: intel os x crash using window codecs
Martin Simmons
vyslnqaaxytp at spammotel.com
Sun Dec 3 00:32:47 CET 2006
>>>>> On Fri, 1 Dec 2006 10:58:40 -0500, Dave Chand <devchan1 at bellsouth.net> said:
>
> On Nov 30, 2006, at 10:27 PM, Zuxy Meng wrote:
>
> > Could u do a disassem in gdb to see what the bad instruction is?
>
> (gdb) disass $pc-32 $pc+32
> Dump of assembler code from 0x900373ed to 0x9003742d:
> 0x900373ed <nanosleep+157>: punpckldq %xmm0,%xmm2
> 0x900373f1 <nanosleep+161>: punpckldq 1258178(%ebx),%xmm2
> 0x900373f9 <nanosleep+169>: movapd 1258194(%ebx),%xmm3
> 0x90037401 <nanosleep+177>: subpd %xmm3,%xmm2
> 0x90037405 <nanosleep+181>: movapd %xmm2,%xmm0
> 0x90037409 <nanosleep+185>: haddpd %xmm2,%xmm0
> 0x9003740d <nanosleep+189>: movapd %xmm0,-72(%ebp)
> 0x90037412 <nanosleep+194>: movsd -72(%ebp),%xmm2
> 0x90037417 <nanosleep+199>: movd -28(%ebp),%xmm1
> 0x9003741c <nanosleep+204>: movdqa %xmm4,%xmm0
> 0x90037420 <nanosleep+208>: punpckldq %xmm1,%xmm0
> 0x90037424 <nanosleep+212>: punpckldq 1258178(%ebx),%xmm0
> 0x9003742c <nanosleep+220>: subpd %xmm3,%xmm0
> End of assembler dump.
> (gdb) info all-registers
> eax 0x0 0
> ecx 0x0 0
> edx 0x0 0
> ebx 0x9003735e -1878822050
> esp 0xbfffd1fc 0xbfffd1fc
> ebp 0xbfffd284 0xbfffd284
> esi 0xbfffd2a4 -1073753436
> edi 0xbfffd34c -1073753268
> eip 0x9003740d 0x9003740d <nanosleep+189>
> eflags 0x210246 2163270
> cs 0x17 23
> ss 0x1f 31
> ds 0x1f 31
> es 0x1f 31
> fs 0x8f 143
> gs 0x37 55
Looks like a stack alignment problem to me -- the Intel Mac calling convention
requires 16 byte alignment of the stack before the call, which won't be
happening when a Win32 DLL makes calls to functions in loader/win32.c.
__Martin
More information about the MPlayer-users
mailing list