The gpio-siox driver uses handle_nested_irq() to implement its
interrupt support. This is only capable of handling threaded irq
actions. For a hardirq action it triggers a NULL pointer oops.
(It calls action->thread_fn which is NULL then.)
Prevent registration of a hardirq action by setting
gpio_irq_chip::threaded to true.
Cc: [email protected]
Signed-off-by: Ahmad Fatoum <[email protected]>
---
v1 -> v2:
reworded commit message (Uwe)
---
drivers/gpio/gpio-siox.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/drivers/gpio/gpio-siox.c b/drivers/gpio/gpio-siox.c
index 26e1fe092304..f8c5e9fc4bac 100644
--- a/drivers/gpio/gpio-siox.c
+++ b/drivers/gpio/gpio-siox.c
@@ -245,6 +245,7 @@ static int gpio_siox_probe(struct siox_device *sdevice)
girq->chip = &ddata->ichip;
girq->default_type = IRQ_TYPE_NONE;
girq->handler = handle_level_irq;
+ girq->threaded = true;
ret = devm_gpiochip_add_data(dev, &ddata->gchip, NULL);
if (ret)
--
2.28.0
Hello Ahmad,
On Mon, Sep 07, 2020 at 05:31:35PM +0200, Ahmad Fatoum wrote:
> The gpio-siox driver uses handle_nested_irq() to implement its
> interrupt support. This is only capable of handling threaded irq
> actions. For a hardirq action it triggers a NULL pointer oops.
> (It calls action->thread_fn which is NULL then.)
>
> Prevent registration of a hardirq action by setting
> gpio_irq_chip::threaded to true.
cool commit log :-)
> Cc: [email protected]
> Signed-off-by: Ahmad Fatoum <[email protected]>
Acked-by: Uwe Kleine-K?nig <[email protected]>
Thanks for picking this up.
Best regards
Uwe
--
Pengutronix e.K. | Uwe Kleine-K?nig |
Industrial Linux Solutions | https://www.pengutronix.de/ |
On Mon, Sep 7, 2020 at 5:32 PM Ahmad Fatoum <[email protected]> wrote:
>
> The gpio-siox driver uses handle_nested_irq() to implement its
> interrupt support. This is only capable of handling threaded irq
> actions. For a hardirq action it triggers a NULL pointer oops.
> (It calls action->thread_fn which is NULL then.)
>
> Prevent registration of a hardirq action by setting
> gpio_irq_chip::threaded to true.
>
> Cc: [email protected]
> Signed-off-by: Ahmad Fatoum <[email protected]>
> ---
Could you add a Fixes tag? This looks like stable material.
Bart
On Wed, Sep 09, 2020 at 11:43:24AM +0200, Bartosz Golaszewski wrote:
> On Mon, Sep 7, 2020 at 5:32 PM Ahmad Fatoum <[email protected]> wrote:
> >
> > The gpio-siox driver uses handle_nested_irq() to implement its
> > interrupt support. This is only capable of handling threaded irq
> > actions. For a hardirq action it triggers a NULL pointer oops.
> > (It calls action->thread_fn which is NULL then.)
> >
> > Prevent registration of a hardirq action by setting
> > gpio_irq_chip::threaded to true.
> >
> > Cc: [email protected]
> > Signed-off-by: Ahmad Fatoum <[email protected]>
> > ---
>
> Could you add a Fixes tag? This looks like stable material.
Ah, we talked about this in the v1 thread---tglx and I agreed we want
it. That's:
Fixes: be8c8facc707 ("gpio: new driver to work with a 8x12 siox")
Best regards
Uwe
--
Pengutronix e.K. | Uwe Kleine-K?nig |
Industrial Linux Solutions | https://www.pengutronix.de/ |
On Wed, Sep 9, 2020 at 12:30 PM Uwe Kleine-König
<[email protected]> wrote:
>
> On Wed, Sep 09, 2020 at 11:43:24AM +0200, Bartosz Golaszewski wrote:
> > On Mon, Sep 7, 2020 at 5:32 PM Ahmad Fatoum <[email protected]> wrote:
> > >
> > > The gpio-siox driver uses handle_nested_irq() to implement its
> > > interrupt support. This is only capable of handling threaded irq
> > > actions. For a hardirq action it triggers a NULL pointer oops.
> > > (It calls action->thread_fn which is NULL then.)
> > >
> > > Prevent registration of a hardirq action by setting
> > > gpio_irq_chip::threaded to true.
> > >
> > > Cc: [email protected]
> > > Signed-off-by: Ahmad Fatoum <[email protected]>
> > > ---
> >
> > Could you add a Fixes tag? This looks like stable material.
>
> Ah, we talked about this in the v1 thread---tglx and I agreed we want
> it. That's:
>
> Fixes: be8c8facc707 ("gpio: new driver to work with a 8x12 siox")
>
Thanks, now queued for fixes.
Bartosz