On 12.08.2021 22:31, Thomas Gleixner wrote:
> clock_was_set() can be invoked from preemptible context. Use raw_cpu_ptr()
> to check whether high resolution mode is active or not. It does not matter
> whether the task migrates after acquiring the pointer.
>
> Fixes: e71a4153b7c2 ("hrtimer: Force clock_was_set() handling for the HIGHRES=n, NOHZ=y case")
> Reported-by: Mike Galbraith <efault@xxxxxx>
> Signed-off-by: Thomas Gleixner <tglx@xxxxxxxxxxxxx>
Tested-by: Marek Szyprowski <m.szyprowski@xxxxxxxxxxx>
This fixes the following issue:
BUG: using smp_processor_id() in preemptible [00000000] code: hwclock/227
> ---
> kernel/time/hrtimer.c | 5 +++--
> 1 file changed, 3 insertions(+), 2 deletions(-)
>
> --- a/kernel/time/hrtimer.c
> +++ b/kernel/time/hrtimer.c
> @@ -944,10 +944,11 @@ static bool update_needs_ipi(struct hrti
> */
> void clock_was_set(unsigned int bases)
> {
> + struct hrtimer_cpu_base *cpu_base = raw_cpu_ptr(&hrtimer_bases);
> cpumask_var_t mask;
> int cpu;
>
> - if (!hrtimer_hres_active() && !tick_nohz_active)
> + if (!__hrtimer_hres_active(cpu_base) && !tick_nohz_active)
> goto out_timerfd;
>
> if (!zalloc_cpumask_var(&mask, GFP_KERNEL)) {
> @@ -958,9 +959,9 @@ void clock_was_set(unsigned int bases)
> /* Avoid interrupting CPUs if possible */
> cpus_read_lock();
> for_each_online_cpu(cpu) {
> - struct hrtimer_cpu_base *cpu_base = &per_cpu(hrtimer_bases, cpu);
> unsigned long flags;
>
> + cpu_base = &per_cpu(hrtimer_bases, cpu);
> raw_spin_lock_irqsave(&cpu_base->lock, flags);
>
> if (update_needs_ipi(cpu_base, bases))
>
Best regards
--
Marek Szyprowski, PhD
Samsung R&D Institute Poland
|