[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <a9ccd0cd-82ee-4e31-9815-f8f3871b30c8@linaro.org>
Date: Fri, 16 Aug 2024 13:40:11 +0200
From: Philippe Mathieu-Daudé <philmd@...aro.org>
To: Jiaxun Yang <jiaxun.yang@...goat.com>,
Thomas Bogendoerfer <tsbogend@...ha.franken.de>
Cc: linux-mips@...r.kernel.org, linux-kernel@...r.kernel.org,
Serge Semin <fancer.lancer@...il.com>,
Frederic Weisbecker <frederic@...nel.org>,
Thomas Gleixner <tglx@...utronix.de>, Ingo Molnar <mingo@...nel.org>,
Daniel Lezcano <daniel.lezcano@...aro.org>
Subject: Re: [PATCH] MIPS: cevt-r4k: Don't call get_c0_compare_int if timer
irq is installed
Hi Jiaxun,
+Daniel/Thomas for CLOCKEVENTS
On 13/8/24 11:59, Jiaxun Yang wrote:
> This avoids warning:
>
> [ 0.118053] BUG: sleeping function called from invalid context at kernel/locking/mutex.c:283
>
> Caused by get_c0_compare_int on secondary CPU.
>
> We also skipped saving IRQ number to struct clock_event_device *cd as
> it's never used by clockevent core, as per comments it's only meant
> for "non CPU local devices".
>
> Reported-by: Serge Semin <fancer.lancer@...il.com>
> Closes: https://lore.kernel.org/linux-mips/6szkkqxpsw26zajwysdrwplpjvhl5abpnmxgu2xuj3dkzjnvsf@4daqrz4mf44k/
> Signed-off-by: Jiaxun Yang <jiaxun.yang@...goat.com>
> ---
> arch/mips/kernel/cevt-r4k.c | 15 +++++++--------
> 1 file changed, 7 insertions(+), 8 deletions(-)
>
> diff --git a/arch/mips/kernel/cevt-r4k.c b/arch/mips/kernel/cevt-r4k.c
> index 368e8475870f..5f6e9e2ebbdb 100644
> --- a/arch/mips/kernel/cevt-r4k.c
> +++ b/arch/mips/kernel/cevt-r4k.c
> @@ -303,13 +303,6 @@ int r4k_clockevent_init(void)
> if (!c0_compare_int_usable())
> return -ENXIO;
>
> - /*
> - * With vectored interrupts things are getting platform specific.
> - * get_c0_compare_int is a hook to allow a platform to return the
> - * interrupt number of its liking.
> - */
> - irq = get_c0_compare_int();
> -
> cd = &per_cpu(mips_clockevent_device, cpu);
>
> cd->name = "MIPS";
> @@ -320,7 +313,6 @@ int r4k_clockevent_init(void)
> min_delta = calculate_min_delta();
>
> cd->rating = 300;
> - cd->irq = irq;
What should be the unset value is not clear, clock_event_device::irq
is an int, described in <linux/clockchips.h> as:
@irq: IRQ number (only for non CPU local devices)
Most arch clocks set irq > 0 or don't set it.
Core code / drivers check for irq != -1, irq != 0 or irq >= 0:
kernel/time/tick-common.c-295- if (newdev->irq >= 0 &&
!irq_can_set_affinity(newdev->irq))
kernel/time/tick-common.c-296- return false;
drivers/clocksource/exynos_mct.c-479- if (evt->irq == -1)
drivers/clocksource/exynos_mct.c-480- return -EIO;
drivers/clocksource/timer-ti-dm-systimer.c-562- if (!dev->irq)
drivers/clocksource/timer-ti-dm-systimer.c-563- return -ENXIO;
Using -1 for unset value seems reasonable, but since most don't set
this default value, I guess that's OK to do the same.
Reviewed-by: Philippe Mathieu-Daudé <philmd@...aro.org>
> cd->cpumask = cpumask_of(cpu);
> cd->set_next_event = mips_next_event;
> cd->event_handler = mips_event_handler;
> @@ -332,6 +324,13 @@ int r4k_clockevent_init(void)
>
> cp0_timer_irq_installed = 1;
>
> + /*
> + * With vectored interrupts things are getting platform specific.
> + * get_c0_compare_int is a hook to allow a platform to return the
> + * interrupt number of its liking.
> + */
> + irq = get_c0_compare_int();
> +
> if (request_irq(irq, c0_compare_interrupt, flags, "timer",
> c0_compare_interrupt))
> pr_err("Failed to request irq %d (timer)\n", irq);
>
> ---
> base-commit: 9e6869691724b12e1f43655eeedc35fade38120c
> change-id: 20240812-get_c0_compare_int-66935853a308
>
> Best regards,
Powered by blists - more mailing lists