On 16.09.2015 00:06, Alexey Brodkin wrote:
Hi Mathias,
On Tue, 2015-09-15 at 16:18 +0300, Mathias Nyman wrote:
On 10.09.2015 14:54, Alexey Brodkin wrote:
I'm not sure what happens here, have you tried the 4.2 kernel?
I've just tried vanilla 4.3-rc1 and see exactly the same picture.
If I send additional xhci debugging patches can you apply and test them?
Sure!
An additional usbmon trace showing which URBs are actually sent could help.
Below are simple "usbmon" outputs. If there's a need for additional
options for usbmon let me know which ones I need to pass.
Thanks,
Looks like a we have at least one bug in how we handle short packet transfers for bulk transfer
we expect to be large.
The driver ask for 65664 bytes from the device, but the device only sends 73 bytes (see log below)
09c4b540 0.374261 C Bo:1:002:1 0 5 >
09c4b540 0.374312 S Bi:1:002:2 - 65664 <
09c4b540 0.374368 C Bi:1:002:2 0 73 =
00000000 00000001 08200716 00000000 00000000 00100400 00000000 00000000
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
If xhci gets less data than requested it will issue a short transfer event. But as the
difference is bigger than 64k we will get this event before the last TRB of that TD, and then
we will get an additional event for the actual last TRB.
This doesn't yet explain the 5 seconds timeout, but it should solve the
"WARN Event TRB for slot 8 ep 4 with no TDs queued?" warning.
I'll scribble down a patch for this shortly
-Mathias
--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
