I have upgraded the system concerned to include sarge/security updates
and this bug still appears to be present. I appreciate that the 2.6.8
kernel may not be upgraded/updated to include a patch of this kind, but
obviously I would like to see this problem fixed - and I've heard
something like it may still be present in 2.6.12-something, so may still
be relevant in any case.
I am not sure what else to provide that may solve the problem, but I
have run lsusb, lspci and dmesg straight after booting the 2.6.7 and
2.6.8 kernels and they are currently stored at:
http://www.kepiernet.ukfsn.org/debian/
My comments: (via tkdiff analysis)
* The lspci outputs are the same.
* From the lsusb logs, bus 5 is not getting included at all in 2.6.8,
and the mouse device number is one lower too (off-by-one bug?)
* The dmesg output is fairly similar, the key differences (from someone
not versed in the kernel):
- ACPI: IRQ0 used by override.
ACPI: IRQ2 used by override.
ACPI: IRQ9 used by override.
...in 2.6.8 and not 2.6.7
- the 'IRQ redirection table' is different (vector-based indexing?),
as are the 'IRQ to pin mappings'
- irq differences are spread throughout, presumably resulting from
this difference
- about line 284 of dmesg_2.6.8 is I think where USB bus 5 fails:
--8<--
ehci_hcd 0000:00:1d.7: BIOS handoff failed (104, 1010001)
ehci_hcd 0000:00:1d.7: can't reset
ehci_hcd 0000:00:1d.7: init 0000:00:1d.7 fail, -95
ehci_hcd: probe of 0000:00:1d.7 failed with error -95
------
Possibly of use: googling the last 'probe' error gave me:
http://www.ussg.iu.edu/hypermail/linux/kernel/0407.2/0298.html
which if you follow the thread forwards says:
--8<--
"This was previously discussed as the subject "proper bios handoff in
ehci-hcd". Basically it's a broken bios, and David Brownell has a fix
for it that hasn't made it into the kernel yet.
Changing the "return 1" at line 308 in drivers/usb/host/ehci-hcd.c" to a
"return 0" works for me until the patch makes it into the tree.
------
I've looked at a few online linux source browsers and this workaround
seem to be in:
* 2.6.11: http://lxr.linux.no/source/drivers/usb/host/ehci-hcd.c#L284
* 2.6.9 :
http://fxr.watson.org/fxr/source/drivers/usb/host/ehci-hcd.c?v=linux-2.6.9#L289
I've not yet found (or downloaded) 2.6.8 to see if this workaround is in
there - but this thread implies that this workaround was the 'final'
patch and its just a buggy bios :(
http://groups.google.com/group/linux.kernel/browse_frm/thread/9aff9b4c47439e30/092fc21b96cd7c0e#092fc21b96cd7c0e
Please let me know if there is anything else that I could try to debug
or fix the problem - I will try the kernel from etch when I have time.
Time to try another usb port/controller/whatever and see if that works :|
--
Neil
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
