Scott Michel wrote:
> Matthew Dillon wrote:
>
>> It's really hard to track this stuff down unless we can reproduce it
>> on our own hardware. I don't have that laptop but I do have a
>> VAIO PCG-F560 that might be similar enough that the USB fails the same
>> way. What brand of USB ethernet is it? Is it internal or
>> external to
>> the laptop?
Looks like the actual bug is in bus/usb/usb_ethersubr.c:usbintr. No
packets get queued up to usbq_rx, consequently nothing gets dequeued.
Similarly, I suspect nothing gets queue on usbq_tx either. But the net
effect for *all* USB ethernet devices would be that nothing appears to
be received. This would also fix the earlier message about cue0 not
working, since it's also a USB Ethernet device that uses netisr_queue to
do its dirty work.
If I ifdef out the rx/tx queue looping code, which only appears to get
called when sent a message for a received packet, things start to appear
to work. Might have a patch after a bit more testing.