Andrew Morton wrote:
>
> dart@iowaone.net wrote:
> >
> > Andrew Morton wrote:
> > >
> > > Jeff Garzik wrote:
> > > >
> > > > > Methinks that should be 'TX_RING_SIZE - 2'
> > > >
> > > > Good spotting, I'm not sure it should be 'TX_RING_SIZE -2' but it sure
> > > > looks like a bug. Donald, can you confirm? This is also in your latest
> > > > epic100.c...
> > >
> > > It was, effectively, 'TX_RING_SIZE - 2' in earlier versions. I'd expect this to fix dart@iowaone.net's problems, but it would be nice to get this sorted promptly: this
> > > problem also exists in latest 2.2 which is about to be released to manufacturing...
> > >
> > > --
> > > -akpm-
> > Jeff and I have been working on it...he killed a module
> > loading/unloading bug but I'm having trouble getting a decent oops on
> > the lockup itself.
>
> Did you try the TX_RING_SIZE fix? It's definitely a bug.
Yes, Jeffs current patch attached below..I've tried both his incarnation
with + 2 removed entirely and also - 2. This patch does fix the module
load/unload problem we found.
>
> > C-A-SysRQ/A-SysRQ is no good, can't get it to spit an oops to
> > console..etc...He mentioned increasing debug level to 7 but I don't know
> > the file to accomplish this in the kernel proper(already done the
> > driver)...clue please?
>
> Not sure what you mean...
>
> Jeff will be referring to the variable 'debug' in epic100.c
>
> If you're using it as a module, just use 'insmod epic100 debug=7'
>
> If you're linking the driver direct into the kernel, just change line 33
> to
>
> static int debug = 7;
I've already changed the driver to 7...I'm looking for a line that would
do the same for the kernel itself. Haven't had much success with the
console logging. I've redirected it to lp0 but the kernel complains that
it can't locate /dev/console...yes, it's there and I have serial console
debugging support enabled in the kernel. Booting with console=/dev/lp0.
This archive was generated by hypermail 2b29 : Wed May 31 2000 - 21:00:18 EST