Hi
On 10.10.2014 21:47, Felipe Balbi wrote:
> Hi Mathias,
>
> I seem to be able to rather easily kill xHCI by just running
> test.sh/testusb with a USB2 device attached to it (I used my AM437x
> Starter Kit for the test).
>
> I noticed that after not too long, all tests started failing (see below)
> but using ehci, it works just fine (afaicr).
>
> Here's a snippet of the hang, but it has been running for a while and
> before testusb I ran over 24 hours of MSC testing, also passed full
> USB20CV and USB30CV with g_mass_storage and g_zero (and most other
> gadget drivers too):
>
Ouch, I don't have a proper device in my office to run testusb with, I probably should set up one here.
(If I understood correctly any device able to use g_zero should work?)
Could you try with xhci debugging enabled? (will probably produce a lot of output)
echo -n 'module xhci_hcd =p' > /sys/kernel/debug/dynamic_debug/control
> [874685.176903] usbtest 1-7:3.0: Linux gadget zero
> [874685.176906] usbtest 1-7:3.0: high-speed {control in/out bulk-in bulk-out} tests (+alt)
> [874703.140258] usbtest 1-7:3.0: TEST 9: ch9 (subset) control tests, 5000 times
> [874723.716857] usbtest 1-7:3.0: TEST 10: queue 32 control calls, 5000 times
> [874743.773236] usbtest 1-7:3.0: TEST 14: 15000 ep0out, 1..256 vary 1
> [874748.774291] usbtest 1-7:3.0: ctrl_out write failed, code -110, count 0
> [874754.297205] usbtest 1-7:3.0: set altsetting to 0 failed, -110
Timeout?
...
> [874790.680681] xhci_hcd 0000:00:14.0: URB transfer length is wrong, xHC issue? req. len = 0, act. len = 4294967288
Hmm, "act. len" value looks awfully close to full 32 bit, driver might be wrapping around something here.
-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
