[FFmpeg-devel] [PATCH] VC-1 MMX DSP functions
infernix
infernix
Sat Jul 7 16:05:58 CEST 2007
Christophe GISQUET wrote:
> int where long int should be used. This probably was in the first patch,
> but I forgot to reinsert a cast when refactoring vc1dsp_mmx.c.
>
> The patch attached should fix that.
Builds fine, but crashes on some vc1 samples
(http://dx.infernix.net/ffvc1/fixed/) where it does not crash without
this patch. VC1 clean up patches applied also.
CPU is an intel C2D E6600. gdb backtrace:
(gdb) run artifacts10.vc1
Starting program: /usr/bin/ffplay artifacts10.vc1
[Thread debugging using libthread_db enabled]
[New Thread 47471130704384 (LWP 19320)]
[New Thread 1082132832 (LWP 19323)]
[New Thread 1090525536 (LWP 19324)]
[New Thread 1098918240 (LWP 19325)]
Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 1098918240 (LWP 19325)]
0x00002b2cb88bedb8 in vc1_put_shift2_mmx (dst=0xb89047d641801ce8
<Address 0xb89047d641801ce8 out of bounds>,
dst_stride=8, src=0x2ab2aaefdfee <Address 0x2ab2aaefdfee out of
bounds>, src_stride=1920, h=10, rnd=7,
offset=1) at i386/vc1dsp_mmx.c:52
52 asm volatile(
(gdb) bt
#0 0x00002b2cb88bedb8 in vc1_put_shift2_mmx (dst=0xb89047d641801ce8
<Address 0xb89047d641801ce8 out of bounds>,
dst_stride=8, src=0x2ab2aaefdfee <Address 0x2ab2aaefdfee out of
bounds>, src_stride=1920, h=10, rnd=7,
offset=1) at i386/vc1dsp_mmx.c:52
#1 0x00002b2cb89047d6 in put_vc1_mspel_mc22_mmx (dst=0x2aaaab2900e0
'\200' <repeats 200 times>...,
src=<value optimized out>, stride=1, rnd=1) at i386/vc1dsp_mmx.c:191
#2 0x00002b2cb8a3fe95 in vc1_mc_1mv (v=0x563740, dir=<value optimized
out>) at vc1.c:559
#3 0x00002b2cb8a4609b in vc1_decode_p_blocks (v=0xb89047d641801ce8) at
vc1.c:3122
#4 0x00002b2cb8a4ef11 in vc1_decode_frame (avctx=0x53d7b0,
data=0x5d4d80, data_size=0x4180216c,
buf=0x1 <Address 0x1 out of bounds>, buf_size=211988) at vc1.c:3813
#5 0x00002b2cb87a32a1 in avcodec_decode_video (avctx=0x53d7b0,
picture=0x5d4d80, got_picture_ptr=0x4180216c,
buf=0x2aaaaada3010 "", buf_size=211988) at utils.c:920
#6 0x0000000000405409 in sdl_audio_callback ()
#7 0x00002b2cb908c477 in SDL_GetThreadID () from /usr/lib/libSDL-1.2.so.0
#8 0x00002b2cb90d1fa9 in SDL_ThreadID () from /usr/lib/libSDL-1.2.so.0
#9 0x00002b2cb923cf1a in start_thread () from /lib/libpthread.so.0
#10 0x00002b2cb9415602 in clone () from /lib/libc.so.6
#11 0x0000000000000000 in ?? ()
(gdb) quit
The program is running. Exit anyway? (y or n) n
Not confirmed.
(gdb) disass $pc-32 $pc+32
Dump of assembler code from 0x2b2cb88bed98 to 0x2b2cb88bedd8:
0x00002b2cb88bed98 <vc1_put_shift2_mmx+17>: or %al,(%rax)
0x00002b2cb88bed9a <vc1_put_shift2_mmx+19>: add %al,(%rax)
0x00002b2cb88bed9c <vc1_put_shift2_mmx+21>: sub %r9d,%eax
0x00002b2cb88bed9f <vc1_put_shift2_mmx+24>: mov
%eax,0xfffffffffffffff4(%rsp)
0x00002b2cb88beda3 <vc1_put_shift2_mmx+28>: lea (%rcx,%rcx,2),%rax
0x00002b2cb88beda7 <vc1_put_shift2_mmx+32>: movd
0xfffffffffffffff4(%rsp),%mm7
0x00002b2cb88bedac <vc1_put_shift2_mmx+37>: punpcklwd %mm7,%mm7
0x00002b2cb88bedaf <vc1_put_shift2_mmx+40>: punpckldq %mm7,%mm7
0x00002b2cb88bedb2 <vc1_put_shift2_mmx+43>: data16
0x00002b2cb88bedb3 <vc1_put_shift2_mmx+44>: data16
0x00002b2cb88bedb4 <vc1_put_shift2_mmx+45>: nop
0x00002b2cb88bedb5 <vc1_put_shift2_mmx+46>: data16
0x00002b2cb88bedb6 <vc1_put_shift2_mmx+47>: data16
0x00002b2cb88bedb7 <vc1_put_shift2_mmx+48>: nop
0x00002b2cb88bedb8 <vc1_put_shift2_mmx+49>: movd (%rdx,%rcx,1),%mm3
0x00002b2cb88bedbc <vc1_put_shift2_mmx+53>: movd 0x4(%rdx,%rcx,1),%mm4
0x00002b2cb88bedc1 <vc1_put_shift2_mmx+58>: movd (%rdx,%rcx,2),%mm1
0x00002b2cb88bedc5 <vc1_put_shift2_mmx+62>: movd 0x4(%rdx,%rcx,2),%mm2
0x00002b2cb88bedca <vc1_put_shift2_mmx+67>: punpcklbw %mm0,%mm3
0x00002b2cb88bedcd <vc1_put_shift2_mmx+70>: punpcklbw %mm0,%mm4
0x00002b2cb88bedd0 <vc1_put_shift2_mmx+73>: punpcklbw %mm0,%mm1
0x00002b2cb88bedd3 <vc1_put_shift2_mmx+76>: punpcklbw %mm0,%mm2
0x00002b2cb88bedd6 <vc1_put_shift2_mmx+79>: paddw %mm1,%mm3
End of assembler dump.
(gdb) info all-registers
rax 0x3 3
rbx 0x780 1920
rcx 0x1 1
rdx 0x2ab2aaefdfee 46946860392430
rsi 0x8 8
rdi 0xb89047d641801ce8 -5147535388536988440
rbp 0x41801ce0 0x41801ce0
rsp 0x41801cd0 0x41801cd0
r8 0xa 10
r9 0x1 1
r10 0x780 1920
r11 0x780 1920
r12 0x1 1
r13 0x2aaaab2900e0 46912504398048
r14 0x2aaaab0eba3f 46912502676031
r15 0xfffffe3e 4294966846
rip 0x2b2cb88bedb8 0x2b2cb88bedb8 <vc1_put_shift2_mmx+49>
eflags 0x10206 [ PF IF RF ]
cs 0x33 51
ss 0x2b 43
ds 0x0 0
es 0x0 0
fs 0x63 99
gs 0x0 0
st0 <invalid float value> (raw 0xffff0000000000000000)
st1 -nan(0x15001600160017) (raw 0xffff0015001600160017)
st2 -nan(0x15001500150015) (raw 0xffff0015001500150015)
st3 -nan(0x1515151516161717) (raw 0xffff1515151516161717)
st4 -nan(0x15001500150015) (raw 0xffff0015001500150015)
st5 -nan(0x35003500350035) (raw 0xffff0035003500350035)
st6 -nan(0x12001200120012) (raw 0xffff0012001200120012)
st7 -nan(0x7000700070007) (raw 0xffff0007000700070007)
fctrl 0x37f 895
fstat 0x0 0
ftag 0xaaaa 43690
fiseg 0x0 0
fioff 0x0 0
foseg 0x0 0
fooff 0x0 0
fop 0x0 0
xmm0 {v4_float = {0x0, 0x1a, 0x0, 0x0}, v2_double =
{0x468f9bf4, 0x0}, v16_int8 = {0x63, 0x9b, 0x33,
0xfd, 0xe6, 0xa3, 0xd1, 0x41, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0,
0x0}, v8_int16 = {0x9b63, 0xfd33, 0xa3e6,
0x41d1, 0x0, 0x0, 0x0, 0x0}, v4_int32 = {0xfd339b63, 0x41d1a3e6,
0x0, 0x0}, v2_int64 = {0x41d1a3e6fd339b63,
0x0}, uint128 = 0x000000000000000041d1a3e6fd339b63}
xmm1 {v4_float = {0x0, 0x2, 0x0, 0x0}, v2_double = {0x2, 0x0},
v16_int8 = {0x0, 0x0, 0x0, 0x0, 0x0,
---Type <return> to continue, or q <return> to quit---
0x0, 0x0, 0x40, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0}, v8_int16 =
{0x0, 0x0, 0x0, 0x4000, 0x0, 0x0, 0x0,
0x0}, v4_int32 = {0x0, 0x40000000, 0x0, 0x0}, v2_int64 =
{0x4000000000000000, 0x0},
uint128 = 0x00000000000000004000000000000000}
xmm2 {v4_float = {0x0, 0x1, 0x0, 0x0}, v2_double = {0x1, 0x0},
v16_int8 = {0x0, 0x0, 0x0, 0x0, 0x0,
0x0, 0xf0, 0x3f, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0}, v8_int16
= {0x0, 0x0, 0x0, 0x3ff0, 0x0, 0x0, 0x0,
0x0}, v4_int32 = {0x0, 0x3ff00000, 0x0, 0x0}, v2_int64 =
{0x3ff0000000000000, 0x0},
uint128 = 0x00000000000000003ff0000000000000}
xmm3 {v4_float = {0x0, 0x1, 0x0, 0x0}, v2_double = {0x1, 0x0},
v16_int8 = {0x0, 0x0, 0x0, 0x0, 0x0,
0x0, 0xf0, 0x3f, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0}, v8_int16
= {0x0, 0x0, 0x0, 0x3ff0, 0x0, 0x0, 0x0,
0x0}, v4_int32 = {0x0, 0x3ff00000, 0x0, 0x0}, v2_int64 =
{0x3ff0000000000000, 0x0},
uint128 = 0x00000000000000003ff0000000000000}
xmm4 {v4_float = {0x0, 0xfffffffd, 0x0, 0x0}, v2_double =
{0xffffffffffffffd2, 0x0}, v16_int8 = {0x5b,
0xaa, 0xa2, 0x2a, 0x9e, 0x6, 0x47, 0xc0, 0x0, 0x0, 0x0, 0x0, 0x0,
0x0, 0x0, 0x0}, v8_int16 = {0xaa5b,
0x2aa2, 0x69e, 0xc047, 0x0, 0x0, 0x0, 0x0}, v4_int32 = {0x2aa2aa5b,
0xc047069e, 0x0, 0x0}, v2_int64 = {
0xc047069e2aa2aa5b, 0x0}, uint128 = 0x0000000000000000c047069e2aa2aa5b}
xmm5 {v4_float = {0x0, 0xffffffff, 0x0, 0x0}, v2_double =
{0x0, 0x0}, v16_int8 = {0xf6, 0x8f, 0xee,
0x21, 0xa8, 0x74, 0xd3, 0xbf, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0,
0x0}, v8_int16 = {0x8ff6, 0x21ee, 0x74a8,
0xbfd3, 0x0, 0x0, 0x0, 0x0}, v4_int32 = {0x21ee8ff6, 0xbfd374a8,
0x0, 0x0}, v2_int64 = {0xbfd374a821ee8ff6,
0x0}, uint128 = 0x0000000000000000bfd374a821ee8ff6}
xmm6 {v4_float = {0x0, 0x0, 0x0, 0x0}, v2_double = {0x0, 0x0},
v16_int8 = {0x0, 0x0, 0x0, 0x0, 0x0,
0x80, 0x59, 0xbc, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0}, v8_int16
= {0x0, 0x0, 0x8000, 0xbc59, 0x0, 0x0,
0x0, 0x0}, v4_int32 = {0x0, 0xbc598000, 0x0, 0x0}, v2_int64 =
{0xbc59800000000000, 0x0},
uint128 = 0x0000000000000000bc59800000000000}
xmm7 {v4_float = {0x0, 0xffffffff, 0x0, 0x0}, v2_double =
{0x0, 0x0}, v16_int8 = {0x17, 0x11, 0xbe,
0x76, 0x67, 0x6b, 0xd3, 0xbf, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0,
0x0}, v8_int16 = {0x1117, 0x76be, 0x6b67,
0xbfd3, 0x0, 0x0, 0x0, 0x0}, v4_int32 = {0x76be1117, 0xbfd36b67,
0x0, 0x0}, v2_int64 = {0xbfd36b6776be1117,
0x0}, uint128 = 0x0000000000000000bfd36b6776be1117}
xmm8 {v4_float = {0x0, 0x0, 0x0, 0x0}, v2_double = {0x0, 0x0},
v16_int8 = {0x54, 0xec, 0x35, 0x16,
0xb3, 0xe9, 0x8f, 0xbd, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0},
v8_int16 = {0xec54, 0x1635, 0xe9b3, 0xbd8f,
0x0, 0x0, 0x0, 0x0}, v4_int32 = {0x1635ec54, 0xbd8fe9b3, 0x0, 0x0},
v2_int64 = {0xbd8fe9b31635ec54, 0x0},
uint128 = 0x0000000000000000bd8fe9b31635ec54}
xmm9 {v4_float = {0x0, 0x1, 0x0, 0x0}, v2_double = {0x1, 0x0},
v16_int8 = {0x6d, 0x7d, 0xbf, 0xbb,
0x27, 0xaf, 0xf5, 0x3f, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0},
v8_int16 = {0x7d6d, 0xbbbf, 0xaf27, 0x3ff5,
0x0, 0x0, 0x0, 0x0}, v4_int32 = {0xbbbf7d6d, 0x3ff5af27, 0x0, 0x0},
v2_int64 = {0x3ff5af27bbbf7d6d, 0x0},
uint128 = 0x00000000000000003ff5af27bbbf7d6d}
xmm10 {v4_float = {0x0, 0xfffffffd, 0x0, 0x0}, v2_double =
{0xffffffffffffffd2, 0x0}, v16_int8 = {0xe0,
0xe6, 0x35, 0x67, 0x9e, 0x6, 0x47, 0xc0, 0x0, 0x0, 0x0, 0x0, 0x0,
0x0, 0x0, 0x0}, v8_int16 = {0xe6e0,
0x6735, 0x69e, 0xc047, 0x0, 0x0, 0x0, 0x0}, v4_int32 = {0x6735e6e0,
0xc047069e, 0x0, 0x0}, v2_int64 = {
0xc047069e6735e6e0, 0x0}, uint128 = 0x0000000000000000c047069e6735e6e0}
xmm11 {v4_float = {0x0, 0x0, 0x0, 0x0}, v2_double = {0x0, 0x0},
v16_int8 = {0x29, 0xf2, 0x88, 0x6c,
0xa6, 0x49, 0xde, 0x3e, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0},
v8_int16 = {0xf229, 0x6c88, 0x49a6, 0x3ede,
0x0, 0x0, 0x0, 0x0}, v4_int32 = {0x6c88f229, 0x3ede49a6, 0x0, 0x0},
v2_int64 = {0x3ede49a66c88f229, 0x0},
uint128 = 0x00000000000000003ede49a66c88f229}
xmm12 {v4_float = {0x9689a800, 0x0, 0x0, 0x0}, v2_double =
{0x0, 0x0}, v16_int8 = {0x6a, 0xa2, 0x65,
0x50, 0xf2, 0xea, 0x8f, 0xbd, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0,
0x0}, v8_int16 = {0xa26a, 0x5065, 0xeaf2,
---Type <return> to continue, or q <return> to quit---
0xbd8f, 0x0, 0x0, 0x0, 0x0}, v4_int32 = {0x5065a26a, 0xbd8feaf2,
0x0, 0x0}, v2_int64 = {0xbd8feaf25065a26a,
0x0}, uint128 = 0x0000000000000000bd8feaf25065a26a}
xmm13 {v4_float = {0x0, 0x0, 0x0, 0x0}, v2_double = {0x0, 0x0},
v16_int8 = {0x0 <repeats 16 times>},
v8_int16 = {0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0}, v4_int32 = {0x0,
0x0, 0x0, 0x0}, v2_int64 = {0x0, 0x0},
uint128 = 0x00000000000000000000000000000000}
xmm14 {v4_float = {0x0, 0x0, 0x0, 0x0}, v2_double = {0x0, 0x0},
v16_int8 = {0x0, 0x0, 0x46, 0x84, 0x24,
0x59, 0xd6, 0x3e, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0}, v8_int16
= {0x0, 0x8446, 0x5924, 0x3ed6, 0x0,
0x0, 0x0, 0x0}, v4_int32 = {0x84460000, 0x3ed65924, 0x0, 0x0},
v2_int64 = {0x3ed6592484460000, 0x0},
uint128 = 0x00000000000000003ed6592484460000}
xmm15 {v4_float = {0x0, 0x1, 0x0, 0x0}, v2_double = {0x1, 0x0},
v16_int8 = {0x0, 0x0, 0x0, 0x0, 0x0,
0x0, 0xf0, 0x3f, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0}, v8_int16
= {0x0, 0x0, 0x0, 0x3ff0, 0x0, 0x0, 0x0,
0x0}, v4_int32 = {0x0, 0x3ff00000, 0x0, 0x0}, v2_int64 =
{0x3ff0000000000000, 0x0},
uint128 = 0x00000000000000003ff0000000000000}
mxcsr 0x1fa0 [ PE IM DM ZM OM UM PM ]
More information about the ffmpeg-devel
mailing list