Chris Johns wrote:
I am looking into the kernel oops I was getting. It is the same problem
I posted here:
http://www.linuxtv.org/pipermail/linux-dvb/2006-December/015009.html
Had the same problem with my Hauppauge Nova-T 500 Dual DVB-T...
Upgraded to the latest FC5 kernel and have switched to a single
processor kernel to simplify debugging.
I attach a cleaner couple of oops. It starts with the USB disconnects
then a couple of I2C writes which fail as they are not I2C but USB
writes then the oops. The first is in 'dvb_dmxdev_buffer_read'. The
second oops is in 'dvb_demux_release'. Looking in the specifics of the
code the mutex is claimed in release but no in the read. I attach a
patch that uncomments the mutex code in the read and also adds the same
to the poll function. With this change I have not seen any driver
problems for the last couple of days and the USB does not disconnect.
...and I were just about to test your patch when I saw that 2.6.20-rc4
were out and that they fixed a couple of things with usb and i2c so I
gave it a shot (without your patch) and the oops I had before seems to
be gone (and I really stressed the card, zapping like crazy and recorded
5 or 6 channels at the same time).
/Richard
_______________________________________________
linux-dvb mailing list
linux-dvb@xxxxxxxxxxx
http://www.linuxtv.org/cgi-bin/mailman/listinfo/linux-dvb
