Re: 2.6.17-mm4

["Jesse Brandeburg" <jesse.brandeburg@gmail.com>]

On 6/30/06, john stultz <johnstul@us.ibm.com> wrote:
> >  <IRQ> [<ffffffff8100d442>] main_timer_handler+0x1ed/0x3ad
> >  [<ffffffff8100d614>] timer_interrupt+0x12/0x27
> >  [<ffffffff8105076a>] handle_IRQ_event+0x29/0x5a
> >  [<ffffffff81050837>] __do_IRQ+0x9c/0xfd
> >  [<ffffffff8100bf27>] do_IRQ+0x63/0x71
> >  [<ffffffff810098b8>] ret_from_intr+0x0/0xa
> >  <EOI>
>
> Hmmm. From that trace I suspect something is enabling interrupts (likely
> in time_init) before timekeeping_init() has chosen the clocksource.
>
> Does the following workaround the issue?
>
> thanks
> -john
>
> diff --git a/init/main.c b/init/main.c
> index ae04eb7..41adc97 100644
> --- a/init/main.c
> +++ b/init/main.c
> @@ -497,8 +497,8 @@ asmlinkage void __init start_kernel(void
>         init_timers();
>         hrtimers_init();
>         softirq_init();
> -       time_init();
>         timekeeping_init();
> +       time_init();
>
>         /*
>          * HACK ALERT! This is early. We're enabling the console before
>

Yes it works, the previously failing bisect kernel boots with this
change. I'll take a look through andrew's suggestions next.