[mythtv] Re: waiting for free video buffers timed out

[Cory Papenfuss]

 	Well, just to complete the ensemble, I figured I'd offer some 
debugging info that might be of help to those in the know.  I manually 
insmod'd the nvidia driver with X running and mythtv working.  Myth 
stopped working, and I got the following in dmesg when I attempted to 
watch live TV:

<snip>
nvidia: module license 'NVIDIA' taints kernel.
NVRM: loading NVIDIA Linux x86 NVIDIA Kernel Module  1.0-6111  Tue Jul 27 
07:55:38 PDT 2004
saa7115: decoder set picture bright=127 contrast=63 saturation=63 hue=0
saa7115: decoder set picture bright=127 contrast=63 saturation=63 hue=0
saa7115: decoder set picture bright=127 contrast=63 saturation=63 hue=0
saa7115: decoder set picture bright=127 contrast=63 saturation=63 hue=0
saa7115: decoder set size
saa7115: decoder disable output
saa7115: decoder enable output
eip: 022d5bf0
------------[ cut here ]------------
kernel BUG at include/asm/spinlock.h:152!
invalid operand: 0000 [#1]
SMP
Modules linked in: nvidia(U) snd_pcm_oss bttv video_buf(U) v4l2_common(U) 
btcx_risc lirc_i2c
(U) lirc_dev(U) floppy snd_mixer_oss snd_emu10k1 snd_rawmidi snd_pcm 
snd_timer snd_seq_devic
e snd_ac97_codec snd_page_alloc snd_util_mem snd_hwdep snd soundcore 
msp3400(U) saa7115(U) t
uner(U) tveeprom(U) ivtv(U) i2c_algo_bit i2c_core videodev md5 ipv6 
autofs4 sunrpc 3c59x mic
rocode joydev dm_mod uhci_hcd ext3 jbd
CPU:    0
EIP:    0060:[<022d5c42>]    Tainted: P
EFLAGS: 00010012   (2.6.8-1.521smp)
EIP is at __down+0x52/0x167
eax: 0000000e   ebx: 45740478   ecx: 022e5780   edx: 00000000
esi: 022d5bf0   edi: 00000246   ebp: 345168d0   esp: 33fbce94
ds: 007b   es: 007b   ss: 0068
Process mythfrontend (pid: 3129, threadinfo=33fbc000 task=345168d0)
Stack: 409b1200 00000001 00000000 00000000 345168d0 0211d07a 00000000 
00000000
        409b13c4 42823000 00000000 45740478 457403a0 00000000 24beace0 
022d5f04
        45740478 33fbc000 00000000 454887d7 00000000 454abc14 00000000 
02163f70
Call Trace:
  [<0211d07a>] default_wake_function+0x0/0xc
  [<022d5f04>] __down_failed+0x8/0xc
  [<454887d7>] .text.lock.nv+0x19/0x122 [nvidia]
  [<02163f70>] exact_match+0x0/0x3
  [<02163deb>] chrdev_open+0x21d/0x27e
  [<0215a5bc>] dentry_open+0xcd/0x1a5
  [<0215a4e9>] filp_open+0x36/0x3c
  [<0215a6e7>] get_unused_fd+0x53/0xf3
  [<0215a8c0>] sys_open+0x31/0x7d
Code: 0f 0b 98 00 3b 57 2e 02 f0 fe 4b 08 0f 88 0c 03 00 00 83 4c


Other relevant info:

[mythtv at localhost nuvexport]$ rpm -qa | grep kernel
kernel-utils-2.4-9.1.131
kernel-2.6.8-1.521
kernel-module-nvidia-graphics6106-2.6.8-1.521smp-1.0_6106-60.rhfc2.at
kernel-2.6.5-1.358
kernel-module-nvidia-graphics6111-2.6.8-1.521smp-1.0_6111-60.rhfc2.at
kernel-module-video4linux-2.6.8-1.521smp-20040910_121305-50.99.rhfc2.at
kernel-smp-2.6.8-1.541_10.rhfc2.at
kernel-module-lirc-2.6.8-1.521smp-0.7.0-38_pre7.rhfc2.at
kernel-module-ivtv-2.6.8-1.541_10.rhfc2.atsmp-0.1.10-47_pre2_ck100zz.rhfc2.at
kernel-smp-2.6.5-1.358
kernel-module-nvidia-graphics6111-2.6.8-1.541_10.rhfc2.atsmp-1.0_6111-60.rhfc2.at
kernel-module-video4linux-2.6.8-1.541_10.rhfc2.atsmp-20040910_121305-50.99.rhfc2.at
kernel-smp-2.6.8-1.521
kernel-sourcecode-2.6.8-1.521
kernel-module-ivtv-2.6.8-1.521smp-0.1.10-47_pre2_ck100zz.rhfc2.at

[mythtv at localhost nuvexport]$ uname -a
Linux localhost.localdomain 2.6.8-1.521smp #1 SMP Mon Aug 16 09:25:06 EDT 
2004 i686 i686 i386 GNU/Linux




*************************************************************************
* Cory Papenfuss							*
* Electrical Engineering candidate Ph.D. graduate student               *
* Virginia Polytechnic Institute and State University 			*
*************************************************************************


On Sat, 2 Oct 2004, Axel Thimm wrote:

> On Sat, Oct 02, 2004 at 12:15:27PM -0400, Doug Larrick wrote:
>> Cory Papenfuss wrote:
>>>    I guess it all makes sense now (except why X insisted on loading the
>>> nvidia module to begin with).  I still think it's mythtv-related (trying
>>> to open nvidia device if there?), since mplayer would play things fine.
>>
>> The nVidia driver was probably providing the OpenGL video sync
>> extension, but not returning from the sync call.  And I probably assumed
>> from your debug output that either (a) OpenGL video sync was working
>> with your ATI card, or (b) you had an nVidia card present.
>
> The mythtv rpms that were used during this testing were w/o
> opengl_vsync support (unless the used rpms were quite old, but from
> the thread it looks like both posters had fully updated systems).
>
> But perhaps the xvmc tried to check for nvidia?
> --
> Axel.Thimm at ATrpms.net
>