bttv bug
Hi,
After upgrading to -test5, my bt878 card ceased to function properly. My kernel is tainted with the nvidia kernel module but I know for a fact that bttv worked just fine not too long ago and I have not upgraded my nvidia drivers since.
I have the following message in my dmesg:
kernel BUG at drivers/media/video/btcx-risc.c:66!
invalid operand: 0000 [#1] CPU: 0 EIP: 0060:[<e096b0a1>] Tainted: P EFLAGS: 00010286
EIP is at btcx_riscmem_alloc+0xa1/0xb0 [btcx_risc]
eax: 00000000 ebx: 000008b8 ecx: cdaed000 edx: ce59a000
esi: db5607dc edi: ce59a8b8 ebp: 00000280 esp: dbb95b54
ds: 007b es: 007b ss: 0068
Process motv (pid: 757, threadinfo=dbb94000 task=dc7c0140)
Stack: dfe8f054 000008b8 dbb95b64 00000020 0e59a000 00000280 00000000 db560740
e09b2a3e dfe8f000 db5607dc 000008b8 00000280 00012c00 db560740 00000001
e09b40f4 e09c1d00 db5607dc dd8cb400 00000000 00000280 00000000 000000f0
Call Trace:
[<e09b2a3e>] bttv_risc_packed+0x3e/0x150 [bttv]
[<e09b40f4>] bttv_buffer_risc+0x514/0x630 [bttv]
[<e0986846>] videobuf_waiton+0xb6/0xf0 [video_buf]
[<e09ab194>] bttv_prepare_buffer+0x104/0x1a0 [bttv]
[<e09ab2f2>] buffer_prepare+0x42/0x50 [bttv]
[<e0986dc2>] videobuf_qbuf+0xa2/0x110 [video_buf]
[<c044560f>] wait_hpet_tick+0xf/0x30
[<e09ad769>] bttv_do_ioctl+0x1109/0x1410 [bttv]
[<c013bb50>] __alloc_pages+0x90/0x300
[<c0262bbd>] do_rw_taskfile+0x1bd/0x2b0
[<c01456b2>] do_anonymous_page+0x122/0x230
[<c0145df9>] handle_mm_fault+0xf9/0x190
[<c011ab77>] try_to_wake_up+0xa7/0x150
[<e0a530fa>] __nvsym01580+0x3a/0x70 [nvidia]
[<e0a530fa>] __nvsym01580+0x3a/0x70 [nvidia]
[<e0a71ee4>] __nvsym01575+0x28/0x34 [nvidia]
[<e0a53179>] __nvsym01540+0x49/0x90 [nvidia]
[<c011ab77>] try_to_wake_up+0xa7/0x150
[<c011b66a>] default_wake_function+0x2a/0x30
[<c011b3d0>] schedule+0x1c0/0x3e0
[<c02c01f4>] kfree_skbmem+0x24/0x30
[<c02c0282>] __kfree_skb+0x82/0x100
[<c031b57d>] unix_stream_recvmsg+0x28d/0x580
[<c011b61a>] preempt_schedule+0x2a/0x50
[<c02bc798>] sock_aio_read+0xb8/0xd0
[<c01520bb>] do_sync_read+0x8b/0xc0
[<c044560f>] wait_hpet_tick+0xf/0x30
[<c0258588>] video_usercopy+0xc8/0x1c0
[<c044560f>] wait_hpet_tick+0xf/0x30
[<c01649b0>] __pollwait+0x0/0xd0
[<c0165094>] sys_select+0x234/0x4c0
[<e09adaae>] bttv_ioctl+0x3e/0x70 [bttv]
[<c044560f>] wait_hpet_tick+0xf/0x30
[<e09ac660>] bttv_do_ioctl+0x0/0x1410 [bttv]
[<c044560f>] wait_hpet_tick+0xf/0x30
[<c01641e3>] sys_ioctl+0xf3/0x290
[<c044560f>] wait_hpet_tick+0xf/0x30
[<c010915b>] syscall_call+0x7/0xb
[<c044560f>] wait_hpet_tick+0xf/0x30
Code: 0f 0b 42 00 e0 b5 96 e0 eb b7 90 8d 74 26 00 83 ec 18 89 7c
bttv0: skipped frame. no signal? high irq latency?Signature · 3 lines
Misha Nasledov [email protected] http://nasledov.com/misha/
Can you reproduce it without the nvidia module and produce a call trace that doesn't include nvidia symbols?
Quoted 76 lines · Misha Nasledov
Hi,
After upgrading to -test5, my bt878 card ceased to function properly. My kernel is tainted with the nvidia kernel module but I know for a fact that bttv worked just fine not too long ago and I have not upgraded my nvidia drivers since.
I have the following message in my dmesg:
kernel BUG at
drivers/media/video/btcx-risc.c:66!invalid operand: 0000 [#1] CPU: 0 EIP: 0060:[<e096b0a1>] Tainted: P EFLAGS: 00010286EIP is at
btcx_riscmem_alloc+0xa1/0xb0[btcx_risc]
eax: 00000000 ebx:000008b8ecx:cdaed000edx:ce59a000
esi:db5607dcedi:ce59a8b8ebp: 00000280 esp:dbb95b54
ds: 007b es: 007b ss: 0068
Process motv (pid: 757, threadinfo=dbb94000 task=dc7c0140)Stack: dfe8f054 000008b8 dbb95b64 00000020 0e59a000 00000280 00000000 db560740 e09b2a3e dfe8f000 db5607dc 000008b8 00000280 00012c00 db560740 00000001 e09b40f4 e09c1d00 db5607dc dd8cb400 00000000 00000280 00000000 000000f0 Call Trace: [<e09b2a3e>] bttv_risc_packed+0x3e/0x150 [bttv] [<e09b40f4>] bttv_buffer_risc+0x514/0x630 [bttv] [<e0986846>] videobuf_waiton+0xb6/0xf0 [video_buf] [<e09ab194>] bttv_prepare_buffer+0x104/0x1a0 [bttv] [<e09ab2f2>] buffer_prepare+0x42/0x50 [bttv] [<e0986dc2>] videobuf_qbuf+0xa2/0x110 [video_buf] [<c044560f>] wait_hpet_tick+0xf/0x30 [<e09ad769>] bttv_do_ioctl+0x1109/0x1410 [bttv] [<c013bb50>] __alloc_pages+0x90/0x300 [<c0262bbd>] do_rw_taskfile+0x1bd/0x2b0 [<c01456b2>] do_anonymous_page+0x122/0x230 [<c0145df9>] handle_mm_fault+0xf9/0x190 [<c011ab77>] try_to_wake_up+0xa7/0x150 [<e0a530fa>] __nvsym01580+0x3a/0x70 [nvidia] [<e0a530fa>] __nvsym01580+0x3a/0x70 [nvidia] [<e0a71ee4>] __nvsym01575+0x28/0x34 [nvidia] [<e0a53179>] __nvsym01540+0x49/0x90 [nvidia] [<c011ab77>] try_to_wake_up+0xa7/0x150 [<c011b66a>] default_wake_function+0x2a/0x30 [<c011b3d0>] schedule+0x1c0/0x3e0 [<c02c01f4>] kfree_skbmem+0x24/0x30 [<c02c0282>] __kfree_skb+0x82/0x100 [<c031b57d>] unix_stream_recvmsg+0x28d/0x580 [<c011b61a>] preempt_schedule+0x2a/0x50 [<c02bc798>] sock_aio_read+0xb8/0xd0 [<c01520bb>] do_sync_read+0x8b/0xc0 [<c044560f>] wait_hpet_tick+0xf/0x30 [<c0258588>] video_usercopy+0xc8/0x1c0 [<c044560f>] wait_hpet_tick+0xf/0x30 [<c01649b0>] __pollwait+0x0/0xd0 [<c0165094>] sys_select+0x234/0x4c0 [<e09adaae>] bttv_ioctl+0x3e/0x70 [bttv] [<c044560f>] wait_hpet_tick+0xf/0x30 [<e09ac660>] bttv_do_ioctl+0x0/0x1410 [bttv] [<c044560f>] wait_hpet_tick+0xf/0x30 [<c01641e3>] sys_ioctl+0xf3/0x290 [<c044560f>] wait_hpet_tick+0xf/0x30 [<c010915b>] syscall_call+0x7/0xb [<c044560f>] wait_hpet_tick+0xf/0x30 Code: 0f 0b 42 00 e0 b5 96 e0 eb b7 90 8d 74 26 00 83 ec 18 89 7c bttv0: skipped frame. no signal? high irq latency?Signature · 8 lines
Misha Nasledov [email protected] http://nasledov.com/misha/ - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Hi!
Quoted 2 lines · Aaron Lehmann
Can you reproduce it without the nvidia module and produce a call trace that doesn't include nvidia symbols?
Yes, and then try bttv patches from http://bytesex.org/bttv/
2.6.0's bttv driver is misbehaving in my setup for ages... Try bytesex's patches and report back, please.
Regards,
Nuno Silva
Quoted 10 lines · Misha Nasledov
Quoted 8 lines
Hi,
After upgrading to -test5, my bt878 card ceased to function properly. My
kernel is tainted with the nvidia kernel module but I know for a fact that bttv worked just fine not too long ago and I have not upgraded my nvidia drivers since.I have the following message in my dmesg:
Hi,
Quoted 3 lines · Nuno Silva
Quoted 2 lines · Aaron Lehmann
Can you reproduce it without the nvidia module and produce a call trace that doesn't include nvidia symbols?
It seems to work even with the nvidia module loaded, but once I start X, it stops working. I thought that perhaps the patch to the nvidia kernel module for -test5 was broken, so I used my older 2.6 nvidia kernel module source and replaced the call to kdev_val() with MINOR() but that still did not fix the problem. I can cat /dev/video0 before starting X, but afterwards it says "Device or resouce busy", even after I kill X.
Quoted 4 lines
Yes, and then try bttv patches from http://bytesex.org/bttv/
2.6.0's bttv driver is misbehaving in my setup for ages... Try bytesex's patches and report back, please.
I tried this set of patches and it didn't do anything for me either.
So am I pretty much on my own with this problem? Does anyone use a Bt878 card with an nvidia card under 2.6? It's strange that it used to work just fine, but after upgrading it stopped working. I might try downgrading later to -test4, but it would be unfortunate to have to run an earlier kernel without the VIA8237SATA chipset merged in =\
I read somewhere that Nvidia cards do not support the overlay extension, but I had xawtv working just fine if I disabled Xv, and it worked even better if I used grabdisplay.
Signature · 3 lines
Misha Nasledov [email protected] http://nasledov.com/misha/
Hi,
I'd also like to mention that the FM radio component still seems to work fine and that I get a strange error whenever I attempt to access /dev/video0:
(along with the Device or resource busy error)
tuner: TV freq (268435455.93) out of range (44-958)
Quoted 6 lines · Misha Nasledov
It seems to work even with the nvidia module loaded, but once I start X, it stops working. I thought that perhaps the patch to the nvidia kernel module for -test5 was broken, so I used my older 2.6 nvidia kernel module source and replaced the call to
kdev_val()withMINOR()but that still did not fix the problem. I can cat/dev/video0before starting X, but afterwards it says "Device or resouce busy", even after I kill X.
Signature · 3 lines
Misha Nasledov [email protected] http://nasledov.com/misha/
Hi!
Misha Nasledov wrote:
[..snip..]
Quoted 3 lines
So am I pretty much on my own with this problem? Does anyone use a Bt878 card with an nvidia card under 2.6? It's strange that it used to work just fine,
I have these:
# lspci|grep Brook
00:09.0 Multimedia video controller: Brooktree Corporation Bt878 Video Capture (rev 11) 00:09.1 Multimedia controller: Brooktree Corporation Bt878 Audio Capture (rev 11)
And this card (hauppagge wintv) works fine with 2.4 but doesn't work (at least for long...) in vanilla 2.6.0-test1 -> test3. I'll try a newer kernel shortly :-)
I also have a nvidia card but I don't have any nvidia drivers installed.
One more note: in my setup I must use "NoMTRR":
Section "Screen"
Identifier "Default Screen"
Device "nVidia Corporation [NV15]"
Monitor "Generic Monitor" Option "NoMTRR"
... etc...
Or I'll get screen corruption sooner or later.
Good luck,
Nuno Silva