>
> On Mon, May 22, 2000 at 08:19:28PM +0200, Ralf Baechle wrote:
>
> > since I merged the oss-cvs changes back into my own code by the end of
> > last week the IOC3 stopped generating interrupts on my machine. Got
> > any idea what's wrong? I suspect one of the recent interrupt
> > changes broke things.
>
> As a sideeffect of trying to solve this I now also found why the IOC3
> driver was performing less than great - it was always queueing at most
> a single packet for sending.
>
> Ralf
>
If at all possible, I would suggest working with top of trunk as
of now. Known problems: UP kernel does not do networking (but SMP
kernels do!). My work hours next few days to couple of weeks will
be choppy at best, but I will try to see what's up. Ralf, see if
you can pitch in, I think it has something to do with the way we
initialize stuff during bootup.
John, init 0 panicing the machine is a known problem. Takes a few
minutes to fix, I will try to see. Suggest you do a "reboot" in
the meantime.
Ralf, about sending results of our o2000 efforts outside - I was
thinking of sending a message to linux-kernel and linux-mm soon,
once we can get a kernel compile running "okay" on the 16 node,
32 p, 11 pci bus, 24 scsi controller, 13 disk, 3G memory machine.
(Multiple network interfaces wouldn't hurt, but initially a single
interface is fine, I guess). While most of this is tested piecemeal,
the combined thing has not been tested. Todo before that: get clearance
from Simon/Rich/Nancy about the (informal) announcement.
Kanoj
|