[OLPC-devel] error report on usb8xxx with new kernel and firmware

[supat at supat.eu.org]

Thank you.
That is very helpful/useful mechanism for me to follow.
Now I have work to continue after one day stuck :(

regards,
supat

On Mon, 11 Sep 2006, Dan Williams wrote:

> On Mon, 2006-09-11 at 12:42 +0700, supat at supat.eu.org wrote:
>> Yes. I know what is build 81 is already.
>> And my original report is correct. usb8xxx cause errors on slackware that
>> using build 81 under new Linux BIOS.
>>
>> My suspect is some fire-up mechanism of slackware are different from OLPC
>> used. I notice it has /etc/module.d/blacklist under fedora to protect
>> kernel to probe the wrong VGA. Slackware do not have that mechanism.
>>
>> If you gave me exactly mechanism to fire-up usb8xxx then I may be able to
>> solve or find out the problem.
>
> When you say "fire-up", what do you mean?
>
> When the kernel probes USB devices, it sees that the marvell device has
> certain USB product & vendor IDs.  It then looks for modules that
> support that specific USB P/V ID.  The usb8xxx driver does support the
> USB P/V ID for the marvell device, and the kernel loads the usb8xxx
> module.
>
> The usb8xxx module determines that yes, it actually does support the
> device in question.  Then, it calls request_firmware(), which is a
> kernel function, to retrieve the firmware data from the filesystem.
> request_firmware() generates a hotplug event, which runs udevsend in
> userspace, which handles the firmware request.  The hotplug event is
> handled by whatever executable the kernel is told to run for hotplug
> requests, which is specified by telling /proc/sys/kernel/hotplug which
> thing to run.  The firmware is then written back to a file in sysfs,
> and, back in kernel space, that data is given back to the usb8xxx driver
> on return from request_firmware().
>
> Check you udev rules files (likely in /etc/udev/rules.d) and see how the
> slackware udev handles firmware requests.  Ensure that your hotplug
> setup is working correctly to call whatever slackware hotplug event
> handling scripts there are for firmware.
>
> Dan
>
>> Conclusion: Yes. errors report earlier are reproducible under build 81.
>> And I did use build 81 since at the beginning.
>>
>> regards,
>> supat
>>
>> On Fri, 8 Sep 2006, James Cameron wrote:
>>
>>> Have you been able to reproduce this on build 81 binaries?
>>>
>>> --
>>> James Cameron    mailto:quozl at us.netrek.org     http://quozl.netrek.org/
>>> _______________________________________________
>>> Devel mailing list
>>> Devel at laptop.org
>>> http://mailman.laptop.org/mailman/listinfo/devel
>>>
>> _______________________________________________
>> Devel mailing list
>> Devel at laptop.org
>> http://mailman.laptop.org/mailman/listinfo/devel
>