[PATCH v2 19/25] timekeeping: Rework timekeeping_init() to use shadow_timekeeper

Anna-Maria Behnsen posted 25 patches 1 month, 2 weeks ago
[PATCH v2 19/25] timekeeping: Rework timekeeping_init() to use shadow_timekeeper
Posted by Anna-Maria Behnsen 1 month, 2 weeks ago
From: Anna-Maria Behnsen <anna-maria@linutronix.de>

For timekeeping_init() the sequence count write held time is not relevant
and it could keep working on the real timekeeper, but there is no reason to
make it different from other timekeeper updates.

Convert it to operate on the shadow timekeeper.

Signed-off-by: Anna-Maria Behnsen <anna-maria@linutronix.de>
---
 kernel/time/timekeeping.c | 8 +++-----
 1 file changed, 3 insertions(+), 5 deletions(-)

diff --git a/kernel/time/timekeeping.c b/kernel/time/timekeeping.c
index 6d21e9bf5c35..bb225534fee1 100644
--- a/kernel/time/timekeeping.c
+++ b/kernel/time/timekeeping.c
@@ -1778,7 +1778,7 @@ static bool persistent_clock_exists;
 void __init timekeeping_init(void)
 {
 	struct timespec64 wall_time, boot_offset, wall_to_mono;
-	struct timekeeper *tk = &tk_core.timekeeper;
+	struct timekeeper *tk = &tk_core.shadow_timekeeper;
 	struct clocksource *clock;
 
 	tkd_basic_setup(&tk_core);
@@ -1802,7 +1802,7 @@ void __init timekeeping_init(void)
 	wall_to_mono = timespec64_sub(boot_offset, wall_time);
 
 	guard(raw_spinlock_irqsave)(&tk_core.lock);
-	write_seqcount_begin(&tk_core.seq);
+
 	ntp_init();
 
 	clock = clocksource_default_clock();
@@ -1815,9 +1815,7 @@ void __init timekeeping_init(void)
 
 	tk_set_wall_to_mono(tk, wall_to_mono);
 
-	timekeeping_update(&tk_core, tk, TK_MIRROR | TK_CLOCK_WAS_SET);
-
-	write_seqcount_end(&tk_core.seq);
+	timekeeping_update_staged(&tk_core, TK_CLOCK_WAS_SET);
 }
 
 /* time in seconds when suspend began for persistent clock */

-- 
2.39.5
Re: [PATCH v2 19/25] timekeeping: Rework timekeeping_init() to use shadow_timekeeper
Posted by John Stultz 1 month ago
On Wed, Oct 9, 2024 at 1:29 AM Anna-Maria Behnsen
<anna-maria@linutronix.de> wrote:
>
> From: Anna-Maria Behnsen <anna-maria@linutronix.de>
>
> For timekeeping_init() the sequence count write held time is not relevant
> and it could keep working on the real timekeeper, but there is no reason to
> make it different from other timekeeper updates.
>
> Convert it to operate on the shadow timekeeper.
>
> Signed-off-by: Anna-Maria Behnsen <anna-maria@linutronix.de>
> ---
>  kernel/time/timekeeping.c | 8 +++-----
>  1 file changed, 3 insertions(+), 5 deletions(-)
>
> diff --git a/kernel/time/timekeeping.c b/kernel/time/timekeeping.c
> index 6d21e9bf5c35..bb225534fee1 100644
> --- a/kernel/time/timekeeping.c
> +++ b/kernel/time/timekeeping.c
> @@ -1778,7 +1778,7 @@ static bool persistent_clock_exists;
>  void __init timekeeping_init(void)
>  {
>         struct timespec64 wall_time, boot_offset, wall_to_mono;
> -       struct timekeeper *tk = &tk_core.timekeeper;
> +       struct timekeeper *tk = &tk_core.shadow_timekeeper;

Same nit as earlier, otherwise
Acked-by: John Stultz <jstultz@google.com>