Russell King - ARM Linux Admin wrote:
> Hwa - do you have the SysRQ stuff enabled in your kernel? Does it still
> work after the crash?
My target machine only has serial console (no keyboard port). However
I also have a CATS machine, so I will try to see what happens on that
later on.
One thing I noticed in tulip.c driver -- it has a "while (1) ;" in a
couple
of places when transmit error happens. I think original writer of
this driver wanted to debug if the error happened. Unfortunately
when it does happen (and I can make it happen in my stress test
on my HW ), the driver busy loops forever in the kernel. I took
it out and the driver keeps on working (no more hangups!).
However there is still one more problem -- when sending a lot of data
out from ARM linux via tulip to outside machines, back-to-back,
flooding as fast as possible, the tulip transmitter seems to
halt the driver. Subsequent ifconfig down/up revives it. I am
trying to debug this now. This seems to be the only problem I can
think of, in terms of stability.
Oh... one more thing when I had IDE in PIO mode, when IDE and ethernet
are both busy it seems to behave badly. I ported latest PDC20246
(promise) driver code that does UDMA a little better and I am
exclusively running in UDMA mode with Quantum drives now. It works
well.
Other than that, 2.2.1 is quite stable, very good work all of you!
Thanks.
--
Hwa Jin Bae
PSO Systems Inc
mailto:[EMAIL PROTECTED]
http://www.pso.com
unsubscribe: body of `unsubscribe linux-arm' to [EMAIL PROTECTED]
