Re: Alphastation 600 5/333, sk0 works

On Wednesday 05 January 2005 18.19, you wrote:
> that looks like the famous alpha bug. i. e. unrelated.

Right.
I may be way off now but after having had a lot of these alpha bugs during 
make build I for some reason opened the box and noticed it didn't have very 
adequate cooling.
So I threw in a couple of fans (most notably on top of the CPU heatsink which 
was getting awfully warm) and tried a make build again.
A couple of hours ago it did finnish make build for the first time.

So I guess this box will be put to proper use in some dmz or build farm now.


Regards
Johan

> * Johan M:son Lindman <[email protected]> [2005-01-05 18:07]:
> > On Wednesday 05 January 2005 18.35, you wrote:
> > > "Johan M:son Lindman" <[email protected]> wrote:
> > > > On Wednesday 05 January 2005 14.57, you wrote:
> > > > > sk works maybe because of *current* ? :)
> > > >
> > > > In case you actually do read alpha.html you'll notice the below
> > > > quoted. Emphasize on the "but any feedback is appreciated." part.
> > > > Hence, here you've got it, it works.
> > >
> > > Did you make build and/or make release? Did you kick sk(4) really hard?
> > > (Like flood pinging and ftping huge files, jumbo frames and all the fun
> > > ...)
> >
> > Yeah, the box has been making build all day, until now that it paniced
> > when I stress tested the sk0 (ping -fs 10000 $IP concurrently from two
> > different gigE PCs).
> >
> > Trace spits out an endless amount of
> > "cpu_switch(?, ?, fffffc00006d23ea,1404, 0, fffffc00007ecbb0) at
> > cpu_switch+0xf"
> >
> > I'll get back with ps as soon as I have the serial console cable crimped
> > and connected.
> >
> >
> > Regards
> > Johan