James Carlson wrote:
It's wonderful to see this ... folks have been mis-attributing the
problem to PPP for a while now, and it's good news that the root cause
in USB has been found.
The problem seems located in usbser. But the root-cause is not so clear
yet. As I described in CR6840063, the usbser_wq_thread is supposed to
wake up, which is triggered by several conditions. However, it doesn't,
while there're a lot of messages queued. I have asked submitter to send
me a coredump with debug driver, so that I can figure out the root cause
from the history info. Unfortunately, I haven't successfully got such a
coredump for some reason. Chototsu's suggested fix seems too aggressive
and I'm afraid it may cause regression for other usb serial driver and
applications. I have another fix based on my guess of the root cause.
Somebody with such device has verified it. But Chototsu doesn't agree
it. See
http://www.opensolaris.org/jive/thread.jspa?threadID=70306&tstart=0 .
The RE on that bug is Raymond Chen. I've copied him on this reply,
and I'd hope that he'd follow up with the original poster.
Yes, I have some discussion with Chototsu. And I will contact submitter
of CR6840063 for further help. This CR is the first one with a coredump.
And thank you for the hint of STREAM analysis. Once the root cause is
known, I will putback my fix.
Thanks,
Lei Chen
_______________________________________________
opensolaris-discuss mailing list
opensolaris-discuss@opensolaris.org
