[linux-dvb] cx88 pci_abort messages

Scott list-dvb at bluecamel.eml.cc
Thu Sep 21 05:45:55 CEST 2006


On Sep 20, 2006, at 11:05 PM, Lincoln Dale wrote:

>> I'm continuing to troubleshoot the above problem. Since I was seeing
>> this under 2.6.18-r7 and 2.6.17-r8-gentoo (uses a slightly older
>> version of linux-dvb and cx88 driver) I've moved back to using the
>> 2.6.17-r8-gentoo kernel. That of course didn't make to much of a
>> difference, I didn't expect it to.
>
> i'd suggest you try latest 2.6.18 & perhaps combine it with dvb-v4l
> tip-of-tree.

2.6.18-rc7 was the most recent available snapshot as of 48hrs ago and  
I didn't see any differences. I didn't try to use linux-dvb tip  
though. I did look at the code differences between linux-dvb in  
2.6.18-rc7 and dvb-tip. It looked like it was mostly refactoring and  
I saw nothing to indicate new support. I could be wrong.

> most motherboard bioses allow you to configure explicit IRQs for  
> different
> PCI slots.  you should be able to find a combination where at least  
> cx88
> doesn't share its IRQ with any other device.

I'm fairly sure that the current Asus P5B bios does not allow for  
this. Of course, I'll double check it, and,I could always achieve the  
same result by disabling onboard audio in the BIOS and testing the  
card on it's dedicated irq. It's something new to try but I don't  
have much hope for it.

I don't see how it's current shared IRQ with a _very_ lightly loaded  
USB hub would be causing the problems I'm seeing. Prior to the  
current slot it was sharing an IRQ with the onboard audio. If I'm not  
playing any sound, don't have the audio portion of the dvb card  
loaded, and I'm capturing an ATSC stream which has the audio embedded  
right into it, I don't see how the onboard audio could be stealing  
the IRQ.

> from personal experience, i've found in the past that i got cx88_abort
> errors where something had (incorrectly) put an incorrect card=XX into
> /etc/modprobe.conf - so the card wasn't being autodetected but  
> instead was
> being forced to something it wasn't.

This is a good suggestion. From dmesg:

cx2388x dvb driver version 0.0.5 loaded
CORE cx88[0]: subsystem: 18ac:d500, board: DViCO FusionHDTV 5 Gold  
[card=31,autodetected]
TV tuner 64 at 0x1fe, Radio tuner -1 at 0x1fe
ACPI: PCI Interrupt 0000:04:02.2[A] -> Link [LNKH] -> GSI 7 (level,  
low) -> IRQ 7
cx88[0]/2: found at 0000:04:02.2, rev: 5, irq: 7, latency: 32, mmio:  
0xfd000000
cx88[0]/2: cx2388x based dvb card
DVB: registering new adapter (cx88[0]).
DVB: registering frontend 0 (LG Electronics LGDT3303 VSB/QAM  
Frontend)...

This seems to match my actual device which is the "FusionHDTV 5 RT  
Gold". I had read that the "FusionHDTV5 RT Gold" and the "FusionHDTV5  
Gold Plus" were actually identical boards. The chips seemed to match  
up with the modules and the dvb-v4l tables I've seen.

04:02.2 Multimedia controller: Conexant CX23880/1/2/3 PCI Video and  
Audio Decoder [MPEG Port] (rev 05)
         Subsystem: DViCO Corporation DViCO FusionHDTV5 Gold
         Flags: bus master, medium devsel, latency 32, IRQ 7
         Memory at fd000000 (32-bit, non-prefetchable) [size=16M]


Thanks for the help!

--
Scott



More information about the linux-dvb mailing list