On Fri, 26 Aug 2011, Liming Wang wrote:
> The preempt_rt kernel forces all irq interrupts to be threaded,
> but special interrupts can be excluded from this conversion.
> The cascade interrupt should be part of these exceptions.
>
> In this case, irq2 is initialized before "kthreadd" task, which
> converts irq interrupt to threaded.
>
> If this irq is threaded, the kernel calls "try_to_wake_up" function
> to wake up "kthreadd" task, but at that moment, "kthreadd" task
> has not been initialized and try_to_wake_up wakes up a NULL task.
That's not a reason to mark an interrupt NO_THREAD. The interrupt
handler could be non hardirq safe on RT. You'd run into other
problems.
In the case at hand it's the no_action handler which is hardirq safe
on RT and being no_action which means it simply returns IRQ_NONE.
I'll queue it for the next rt release and fixup the changelog myself.
Thanks,
tglx
> Signed-off-by: Liming Wang <liming.wang@xxxxxxxxxxxxx>
> Signed-off-by: Bruce Ashfield <bruce.ashfield@xxxxxxxxxxxxx>
> ---
> arch/mips/kernel/i8259.c | 1 +
> 1 files changed, 1 insertions(+), 0 deletions(-)
>
> diff --git a/arch/mips/kernel/i8259.c b/arch/mips/kernel/i8259.c
> index 5c74eb7..fb338db 100644
> --- a/arch/mips/kernel/i8259.c
> +++ b/arch/mips/kernel/i8259.c
> @@ -295,6 +295,7 @@ static void init_8259A(int auto_eoi)
> static struct irqaction irq2 = {
> .handler = no_action,
> .name = "cascade",
> + .flags = IRQF_NO_THREAD,
> };
>
> static struct resource pic1_io_resource = {
> --
> 1.7.0.4
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-rt-users" in
> the body of a message to majordomo@xxxxxxxxxxxxxxx
> More majordomo info at http://vger.kernel.org/majordomo-info.html
>
--
To unsubscribe from this list: send the line "unsubscribe linux-rt-users" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
