Ronald G Minnich <[EMAIL PROTECTED]> writes:

> On Monday 19 November 2001 15:18, Eric W. Biederman wrote:
> > Ronald G Minnich <[EMAIL PROTECTED]> writes:
> > > On Monday 19 November 2001 13:29, Eric W. Biederman wrote:
> > > > Cool.  No /proc/cpuinfo?
> > >
> > > Don't know, it is still not up (but has not crashed). Serial port echos
> > > take a really long time (i.e. hit return and wait ...)
> >
> > Ah, o.k.  I wonder what broke serial interrupts.
> > At least that is my guess if it starts happening as soon as init starts...
> >
> 
> yes but init is the first process. When we saw this before it was because 
> init got scheduled onto an SMP node with cache off. 

O.k.  I haven't seen that but I have seen a similiar problem with
serial interrupts.
  
> But I think you're right it is something else. 
> 
> It's right here,and has been for an hour or two.

Ouch.  That is slower then I estimate serial interrupt problem should
be.  That is about 30s for 1/2 line.  That is slow.  But it is the
go to lunch and you are in kind of slow.

You might try adding init=/bin/bash.
And skip most of the slow kernel setup.

I guess telnet/ssh and ping are also interesting things to try.

Eric

Reply via email to