On 15.06.23 11:18, Sebastian Andrzej Siewior wrote:
> From: Thomas Gleixner <tglx@xxxxxxxxxxxxx>
>
> The tick period is aligned very early while the first clock_event_device
> is registered. The system runs in periodic mode and switches later to
> one-shot mode if possible.
>
> The next wake-up event is programmed based on aligned value
> (tick_next_period) but the delta value, that is used to program the
> clock_event_device, is computed based on ktime_get().
>
> With the subtracted offset, the devices fires in less than the exacted
> time frame. With a large enough offset the system programs the timer for
> the next wake-up and the remaining time left is too little to make any
> boot progress. The system hangs.
>
> Move the alignment later to the setup of tick_sched timer. At this point
> the system switches to oneshot mode and a highres clocksource is
> available. It safe to update tick_next_period ktime_get() will now
> return accurate (not jiffies based) time.
>
> [bigeasy: Patch description + testing].
>
> Reported-by: Mathias Krause <minipli@xxxxxxxxxxxxxx>
> Reported-by: "Bhatnagar, Rishabh" <risbhat@xxxxxxxxxx>
> Fixes: e9523a0d81899 ("tick/common: Align tick period with the HZ tick.")
Cc: stable, maybe? This commit already ended up in quite a few "stable"
kernels (v6.3.2, v6.2.15, v6.1.28, v5.15.111, v5.10.180 and v5.4.243)
and it might be better to list them explicitly to avoid one of them
getting missed.
> Signed-off-by: Thomas Gleixner <tglx@xxxxxxxxxxxxx>
> Signed-off-by: Sebastian Andrzej Siewior <bigeasy@xxxxxxxxxxxxx>
> Link: https://lore.kernel.org/5a56290d-806e-b9a5-f37c-f21958b5a8c0@xxxxxxxxxxxxxx
> Link: https://lore.kernel.org/12c6f9a3-d087-b824-0d05-0d18c9bc1bf3@xxxxxxxxxx
> ---
> kernel/time/tick-common.c | 11 +----------
> kernel/time/tick-sched.c | 13 ++++++++++++-
> 2 files changed, 13 insertions(+), 11 deletions(-)
>
> diff --git a/kernel/time/tick-common.c b/kernel/time/tick-common.c
> index 65b8658da829e..b85f2f9c32426 100644
> --- a/kernel/time/tick-common.c
> +++ b/kernel/time/tick-common.c
> @@ -218,19 +218,10 @@ static void tick_setup_device(struct tick_device *td,
> * this cpu:
> */
> if (tick_do_timer_cpu == TICK_DO_TIMER_BOOT) {
> - ktime_t next_p;
> - u32 rem;
>
> tick_do_timer_cpu = cpu;
>
> - next_p = ktime_get();
> - div_u64_rem(next_p, TICK_NSEC, &rem);
> - if (rem) {
> - next_p -= rem;
> - next_p += TICK_NSEC;
> - }
> -
> - tick_next_period = next_p;
> + tick_next_period = ktime_get();
> #ifdef CONFIG_NO_HZ_FULL
> /*
> * The boot CPU may be nohz_full, in which case set
> diff --git a/kernel/time/tick-sched.c b/kernel/time/tick-sched.c
> index 52254679ec489..42c0be3080bde 100644
> --- a/kernel/time/tick-sched.c
> +++ b/kernel/time/tick-sched.c
> @@ -161,8 +161,19 @@ static ktime_t tick_init_jiffy_update(void)
> raw_spin_lock(&jiffies_lock);
> write_seqcount_begin(&jiffies_seq);
> /* Did we start the jiffies update yet ? */
> - if (last_jiffies_update == 0)
> + if (last_jiffies_update == 0) {
> + u32 rem;
> +
> + /*
> + * Ensure that the tick is aligned to a multiple of
> + * TICK_NSEC.
> + */
> + div_u64_rem(tick_next_period, TICK_NSEC, &rem);
> + if (rem)
> + tick_next_period += TICK_NSEC - rem;
> +
> last_jiffies_update = tick_next_period;
> + }
> period = last_jiffies_update;
> write_seqcount_end(&jiffies_seq);
> raw_spin_unlock(&jiffies_lock);
Hah, nice spot. So you implemented what I suggested and it, indeed,
works as expected, thereby:
Tested-by: Mathias Krause <minipli@xxxxxxxxxxxxxx>
Thanks,
Mathias
