On Tue, 2008-10-14 at 11:04 -0400, Jeff Campbell wrote:
> > cat /dev/video1 > test1.mpg
> >
> > I get zero length file
> >
> > -rw-r--r-- 1 root root 0 Oct 13 17:46 test1.mpg
>
>
> At first glance, I'd say you're not getting interrupts for
> some reason.
> Here is /proc/interrupts
>
> enc239-009-060:~# cat /proc/interrupts
> CPU0
> 0: 699019 IO-APIC-edge timer
> 1: 2 IO-APIC-edge i8042
> 9: 0 IO-APIC-fasteoi acpi
> 12: 4 IO-APIC-edge i8042
> 16: 0 IO-APIC-fasteoi cx18-1
^ ^^^^^^
Zero interrupts means zero video buffers were transferred.
> 17: 63 IO-APIC-fasteoi cx18-0
> 18: 109280 IO-APIC-fasteoi eth0
> 19: 3 IO-APIC-fasteoi
> 21: 506 IO-APIC-fasteoi ehci_hcd:usb1, uhci_hcd:usb2,
> uhci_hcd:usb3, uhci_hcd:usb4, uhci_hcd:usb5
> 22: 0 IO-APIC-fasteoi VIA8237
> NMI: 0 Non-maskable interrupts
> LOC: 113 Local timer interrupts
> TRM: 0 Thermal event interrupts
> SPU: 0 Spurious interrupts
> ERR: 0
> MIS: 0
>
>
>
> after the failed attempt. You may also wish to load the cx18
> module
> with some of the debug flags enabled. Namely the flags for
> info, warn,
> mailbox, & irq. See the output of modinfo cx18 for details on
> how to
> set them.
>
>
> I tried modinfo cx18 and was told it couldn't find the module, which I
> expected, as I have it compiled in directly to the kernel.
Oops. I missed that in your original email. BTW you don't need the
ivtv driver at all for CX23418 based cards.
> Given that it is part of the kernel itself (and not a loadable
> module) is there a way to set the various debug options or do I need
> to spin a new kernel with cx18 as a loadable module so I can change
> flags on insert?
Probably for the cx18_debug option, since it's a global symbol. You
could find it's address from /proc/ksyms, and then muck with it's value
directly via /dev/kmem (I think), but it's hardly worth the effort.
Given the rate changes are applied to the cx18 driver in the v4l-dvb
repository (a change every week or two), you'll probably want to build
it as a module anyway. See below...
> Here is some interesting information. I can reproduce it again this
> morning on the same machine. I have a similar machine (same
> motherboard and memory and cards but a different PCI riser card) and
> it was able to pass the "cat" test no problem last night (identical
> kernel/os load (same svn)).
You should really download and test with the latest cx18 driver from the
v4l-dvb repository:
http://www.linuxtv.org/hg/v4l-dvb
Instructions are on this page:
http://www.ivtvdriver.org/index.php/Cx18
Not being able to start the card working and generating interrupts is an
indication that commands didn't get received by the CX23418 over the PCI
bus. cx18 driver v1.0.1 adds retries of PCI mmio read and write
accesses to improve communications with the CX23418.
You can load the v1.0.1 cx18 module with the debug parameter set to log
at least "info", if you want to see statistics on PCI mmio read and
write retries with v4l2-ctl --log-status.
> The host bridge latency timer of 8 seems odd - it's a little
> low.
>
> Is this something I can set myself or is that determined by the BIOS?
You can change it yourself. Values should be multiples of 8.
$ man setpci
Regards,
Andy
_______________________________________________
ivtv-devel mailing list
[email protected]
http://ivtvdriver.org/mailman/listinfo/ivtv-devel
