On Sunday 22 October 2006 23:28, Andrew McNabb wrote:
> Hi. I've been using ivtv for a long time with my PVR 500 (and with
> MythTV), and everything has worked great. However, this weekend I
> upgraded my kernel to 2.6.18 and switched to the 0.8.0 driver. I've
> been having some frightening problems. This message will contain a
> bunch of kernel logs, since I really don't know what's going on. In
> any case, I hope this information will be helpful in tracking down
> the problem, and I'd love to receive any recommendations. Thanks in
> advance.
>
> In a previous message to the list sent on 21 Oct 2006, I mentioned
> that I got a message about the following messages as soon as I
> upgraded to 0.8.0:
>
> ivtv0: Unknown card: vendor/device: 4444/0016
> ivtv0: subsystem vendor/device: 0070/f7f7
>
> Since the card still recorded just fine, I figured that this wasn't a
> serious problem.
It isn't a problem.
>
> However, later that night I got a kernel oops or something:
>
> Oct 21 01:15:50 mcbain kernel: ACPI: PCI interrupt for device
> 0000:03:09.0 disabled Oct 21 01:15:50 mcbain kernel: ivtv1: Removed
> WinTV PVR 500 (unit #1), card #1 Oct 21 01:15:50 mcbain kernel: ivtv0
> warning: ivtv_unreg_dev accessing uninitialied encode r PCM audio
> stream
> Oct 21 01:15:50 mcbain kernel: e8a594ed
> Oct 21 01:15:50 mcbain kernel: Modules linked in: snd_pcm_oss
> snd_mixer_oss snd_intel8x0 w m8775 cx25840 tuner ivtv firmware_class
> i2c_algo_bit cx2341x tveeprom videodev v4l1_compat v4l2_common nvidia
> i2c_core ppdev parport_pc lp parport raid1 raid456 xor md_mod
> dm_snaps hot dm_mirror 8250 serial_core usblp snd_ice1724
> snd_ice17xx_ak4xxx snd_ac97_codec snd_ac9 7_bus snd_ak4114
> snd_ak4xxx_adda snd_mpu401_uart snd_pcm snd_rawmidi snd_seq_device
> snd_ti mer snd ehci_hcd uhci_hcd e1000 snd_page_alloc usbcore
> sata_promise soundcore dm_mod Oct 21 01:15:50 mcbain kernel: EIP:
> 0060:[<e8a594ed>] Tainted: P VLI Oct 21 01:15:50 mcbain
> kernel: EFLAGS: 00210246 (2.6.18.1 #1) Oct 21 01:15:50 mcbain
> kernel: [<e8c2bfeb>] ivtv_remove+0x32a/0x3ba [ivtv] Oct 21 01:15:50
> mcbain kernel: [<c01b6285>] pci_device_remove+0x19/0x2c Oct 21
> 01:15:50 mcbain kernel: [<c01fbc92>]
> __device_release_driver+0x62/0x78 Oct 21 01:15:50 mcbain kernel:
> [<c01fbefb>] driver_detach+0x6a/0x9a Oct 21 01:15:50 mcbain kernel:
> [<c01fb40e>] bus_remove_driver+0x5e/0x83 Oct 21 01:15:50 mcbain
> kernel: [<c01fc012>] driver_unregister+0xb/0x15 Oct 21 01:15:50
> mcbain kernel: [<c01b63f8>] pci_unregister_driver+0x10/0x5d Oct 21
> 01:15:51 mcbain kernel: [<e8c2bb82>] module_cleanup+0x42/0x46 [ivtv]
> Oct 21 01:15:51 mcbain kernel: [<c012faf8>]
> sys_delete_module+0x1ba/0x1e2 Oct 21 01:15:51 mcbain kernel:
> [<c013fe33>] do_munmap+0x194/0x1b0 Oct 21 01:15:51 mcbain kernel:
> [<c01029db>] syscall_call+0x7/0xb
The driver is being unloaded and then an error happens. Are you sure
that the driver was loaded correctly? There is a bug in the driver
(fixed in subversion in fact) where, even though the initialization
failed, the card is still 'accepted' by the kernel. This can cause all
sorts of strange behavior, including a crash like this during unloading
of the driver.
>
> I ended up rebooting and things seemed fine for about 24 hours.
> Then, last night, I had some more problems that seem similar. I got
> messages in /var/log/messages about kernel oopses. These
> corresponded with periods of very high load where the machine froze
> for a few minutes, and then eventually recovered. After a few cycles
> of this, with a manual reboot somewhere in between, the machine froze
> again. This time it never seemed to come back. This afternoon I
> rebooted the machine and turned off MythTV to avoid using the card.
> Here are some more log excerpts:
>
>
> Oct 21 23:16:02 mcbain kernel: ivtv-enc: page allocation failure.
> order:4, mode:0xd0 Oct 21 23:16:02 mcbain kernel: [<c0137138>]
> __alloc_pages+0x246/0x25a Oct 21 23:16:02 mcbain kernel:
> [<c0148cad>] cache_alloc_refill+0x285/0x45d Oct 21 23:16:02 mcbain
> kernel: [<c0148ede>] __kmalloc+0x59/0x62 Oct 21 23:16:02 mcbain
> kernel: [<e88e22d2>] ivtv_init_buffer+0x51/0x180 [ivtv] Oct 21
> 23:16:02 mcbain kernel: [<e88e3852>]
> enc_gather_free_buffers+0x97/0x279 [ivtv] Oct 21 23:16:02 mcbain
> kernel: [<e88ecce2>] ivtv_sched_DMA+0x584/0x874 [ivtv] Oct 21
> 23:16:02 mcbain kernel: [<c011b959>] __do_softirq+0x59/0xbb Oct 21
> 23:16:02 mcbain kernel: [<c011212b>] default_wake_function+0x0/0x15
> Oct 21 23:16:02 mcbain kernel: [<e88f12d1>]
> ivtv_enc_thread+0x14a/0x1c3 [ivtv] Oct 21 23:16:02 mcbain kernel:
> [<c0127993>] autoremove_wake_function+0x0/0x3a Oct 21 23:16:02 mcbain
> kernel: [<e88f1187>] ivtv_enc_thread+0x0/0x1c3 [ivtv] Oct 21
> 23:16:02 mcbain kernel: [<c0100bfd>] kernel_thread_helper+0x5/0xb
1) Do you have the ivtv_dynbuf module option set? If so, remove it.
2) You may have to increase the memory set aside by the kernel for
drivers by adding this kernel option: vmalloc=192M
Hans
>
> This is the point where I restarted the machine and turned off
> MythTV. Anyway, I'm really stumped, and I'd love to hear any feedback
> on what I should do. Thanks for everything--I really love ivtv.
_______________________________________________
ivtv-devel mailing list
[email protected]
http://ivtvdriver.org/mailman/listinfo/ivtv-devel
