Re: Sparc release requalification

[Sébastien Bernard <seb@sfrdev.fr>]

David Miller a écrit :
> From: Sébastien Bernard <seb@sfrdev.fr>
> Date: Mon, 07 Sep 2009 18:53:15 +0200
>
>   
> > David Miller a écrit :
> >     
> > > [great explanation snipped]
> > > --- a/arch/sparc/kernel/nmi.c
> > > +++ b/arch/sparc/kernel/nmi.c
> > > @@ -103,7 +103,7 @@ notrace __kprobes void perfctr_irq(int irq, struct
> > > pt_regs *regs)
> > >  	}
> > >  	if (!touched && __get_cpu_var(last_irq_sum) == sum) {
> > >  		local_inc(&__get_cpu_var(alert_counter));
> > > -		if (local_read(&__get_cpu_var(alert_counter)) == 5 * nmi_hz)
> > > + if (local_read(&__get_cpu_var(alert_counter)) == 30 * nmi_hz)
> > >  			die_nmi("BUG: NMI Watchdog detected LOCKUP",
> > >  				regs, panic_on_timeout);
> > >  	} else {
> > >   
> > >       
> > Hum, I tested today, and no, it does not solve the problem. Kernel is
> > still hanging at the same place.
> > I'll get the initcall debug back when I'll have rebuild a kernel
> > withtou the config_prom_console.
> >     
>
> Then what bug are you talking about?
>
> You stated that disabling the NMI watchdog completely solves your
> problem right?  That's why I mentioned the above patch to you?
>   
My mistake, I didn't look the log hard enough.
I rebuild a new kernel (2.6.31-rc9) with no patch and there was no hang.
It passes the nmi_setup but crashes further when mounting the root 
partition.
I'll send the complete logs further.
The funny part is that booting with debug_initcalls=1 makes the kernel 
hangs.

Seb