[linux-dvb] Another Nova-T 500 Dual DVB-T Oops

[Matt Doran]

Matt Doran wrote:
>
> I've just upgraded to the ubuntu 2.6.20-5 kernel, which I believe is 
> based on 2.6.20-rc5.    I'll see if this reduces the occurrence of the 
> problem.
>
> I'm happy to help by running a different kernel or enabling 
> "USB-debugging".   Let me know what your after and how to do this and 
> I'll try to collect more info.
>
>
The 2.6.20-rc5 is definitely behaving better, but I just had my first 
oops since upgrading (about a week).  It was happening daily previously 
when running the Ubuntu 2.6.20-4-generic kernel which was based on the 
rc2 or rc3 kernel (not exactly sure).


The behaviour leading up to the oops was a little different this time.  
It started with the usually disconnect and reinitialize.... this 
happened at the start of a mythtv recording.

    usb 3-1: USB disconnect, address 2
    mt2060 I2C write failed
    mt2060 I2C write failed


The oops didn't occur until mythtv completed the recording 30 minutes 
later.   Myth must have thought the recording was working, but nothing 
was saved to file.

    BUG: unable to handle kernel paging request at virtual address 008c4530
     printing eip:
    c011c177
    *pde = 00000000
    Oops: 0000 [#1]
    SMP
    CPU:    1
    EIP:    0060:[__wake_up_common+23/96]    Tainted: PF     VLI
    EFLAGS: 00010013   (2.6.20-5-generic #2)
    EIP is at __wake_up_common+0x17/0x60
    eax: f8aea068   ebx: f8aea064   ecx: 00000001   edx: 008c4530
    esi: 00000000   edi: 00000001   ebp: ebc37f24   esp: ebc37f08
    ds: 007b   es: 007b   ss: 0068
    Process mythbackend (pid: 19502, ti=ebc36000 task=c54ed030
    task.ti=ebc36000)
    Stack: 00000000 00000001 00000003 00000282 f8aea064 00000000
    00000001 ebc37f48
           c011c738 00000000 00000000 00000003 00000292 f8aea000
    ebcbe888 00000000
           ebcbe8c8 f8a99f59 00000000 f8aea074 00000008 c1ad0540
    f49f9e7c f403b790
    Call Trace:
     [__wake_up+56/80] __wake_up+0x38/0x50
     [pg0+945962841/1069057024] dvb_demux_release+0xe9/0x120 [dvb_core]
     [__fput+167/400] __fput+0xa7/0x190
     [filp_close+71/128] filp_close+0x47/0x80
     [sys_close+107/208] sys_close+0x6b/0xd0
     [sysenter_past_esp+95/137] sysenter_past_esp+0x5f/0x89
     =======================



Do any of the developer have any idea what could be causing these?  I'm 
happy to enable whatever debugging is required to help track this down.

I've seen some comments (from Michael Krufky 
<http://bugzilla.kernel.org/show_bug.cgi?id=7819#c3> and others) that 
the reason for these types of oops' is that DVB core is not hot 
pluggable.  This card is a PCI card that contains a USB hub, so in 
theory will never be disconnected.  But when these disconnections occur 
(for whatever reason), the DVB core cannot handle this and causes the oops.

Are there any plans to make the dvb core support hot plugging?  This 
could help both the USB stick devices that are intended to be plugged in 
and removed from a running system ... but I assume it would also help 
when an internal USB card has a mysterious disconnection.

I'm wondering what path would be best to try to resolve the Nova 500 
crashes??

   1. Try to track down why the disconnections occur  (which will only
      benefit users of this card)
   2. Or implement hot plug support in DVB (probably a better solution,
      but probably a lot more work).  :)

I know very little about kernel/driver development, but would be happy 
to assist where possible.

Cheers,
Matt



-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.linuxtv.org/pipermail/linux-dvb/attachments/20070203/27114327/attachment-0001.htm