Re: [PATCH] clocksource/drivers/jcore: Use request_percpu_irq()

[Uros Bizjak]

On Mon, Sep 2, 2024 at 12:33 PM Daniel Lezcano
<daniel.lezcano@xxxxxxxxxx> wrote:
>
> On 02/09/2024 12:11, Uros Bizjak wrote:
> > On Mon, Sep 2, 2024 at 11:17 AM Daniel Lezcano
> > <daniel.lezcano@xxxxxxxxxx> wrote:
> >>
> >> On 30/07/2024 15:20, Uros Bizjak wrote:
> >>> Use request_percpu_irq() instead of request_irq() to solve
> >>> the following sparse warning:
> >>>
> >>> jcore-pit.c:173:40: warning: incorrect type in argument 5 (different address spaces)
> >>> jcore-pit.c:173:40:    expected void *dev
> >>> jcore-pit.c:173:40:    got struct jcore_pit [noderef] __percpu *static [assigned] [toplevel] jcore_pit_percpu
> >>>
> >>> Compile tested only.
> >>>
> >>> Signed-off-by: Uros Bizjak <ubizjak@xxxxxxxxx>
> >>> Cc: Daniel Lezcano <daniel.lezcano@xxxxxxxxxx>
> >>> Cc: Thomas Gleixner <tglx@xxxxxxxxxxxxx>
> >>> ---
> >>
> >> Added Rich Felker in Cc
> >>
> >> Applied, thanks
> >
> > I think we also need the following patch, since we changed request_irq
> > to request_percpu_irq:
>
> Hmm, I think you are right:
>
> I would say it is:
>
> static irqreturn_t jcore_timer_interrupt(int irq, void *dev_id)
> {
>          struct jcore_pit *pit = dev_id;
>
>         OR
>
>         struct jcore_pit *pit = this_cpu_ptr(jcore_pit_percpu);
>
>         [ ... ]
> }
>
> The former the better for the encapsulation.
>
> Do you mind to update the patch ?

Done, v2 with changed jcore_timer_interrupt() was just sent.

Sorry for the inconvenience.

Thanks,
Uros.