Re: Oops in 2.5.38-mm2

[Andrew Morton]

Jens Axboe wrote:
> 
> On Mon, Sep 23 2002, Andrew Morton wrote:
> > Grega Fajdiga wrote:
> > >
> > > Good day,
> > >
> > > I get this oops at startup in 2.5.38-mm2. The oops
> >
> > It's not an oops - it's just a warning.
> >
> > > ..
> > > Trace; c0117826 <__might_sleep+56/5d>
> > > Trace; c0134386 <kmalloc+66/1f0>
> > > Trace; c01d2e60 <ide_intr+0/1d0>
> >
> > ide_intr() is calling sleeping functions inside ide_lock.
> 
> this is ludicris, why on earth would ide_intr() call kmalloc() from its
> isr?! the trace is obviously bogus.
> 

Argh, sorry - brain is mush.  It's init_irq() which is calling
sleeping functions inside ide_lock.  The ide_intr is just stack gunk.