* [PATCH] clocksource: publish finished_booting with release semantics
@ 2026-09-22 1:22 Jaidev Shastri via B4 Relay
2026-09-29 11:53 ` Thomas Gleixner
0 siblings, 1 reply; 2+ messages in thread
From: Jaidev Shastri via B4 Relay @ 2026-09-22 1:22 UTC (permalink / raw)
To: John Stultz, Thomas Gleixner, Stephen Boyd, Miroslav Lichvar
Cc: linux-kernel, Jaidev Shastri
From: Jaidev Shastri <jaidevshastri@vt.edu>
clocksource_done_booting() sets curr_clocksource and then
finished_booting, both with plain stores under clocksource_mutex.
clocksource_find_best() tests finished_booting with a plain load before
it walks clocksource_list.
Set the flag with smp_store_release() and read it with
smp_load_acquire().
Found with MBCheck, a static herd7-based memory consistency checker.
Signed-off-by: Jaidev Shastri <jaidevshastri@vt.edu>
---
kernel/time/clocksource.c | 6 ++++--
1 file changed, 4 insertions(+), 2 deletions(-)
diff --git a/kernel/time/clocksource.c b/kernel/time/clocksource.c
index f1253f579..f683a787c 100644
--- a/kernel/time/clocksource.c
+++ b/kernel/time/clocksource.c
@@ -1086,7 +1086,8 @@ static struct clocksource *clocksource_find_best(bool oneshot, bool skipcur)
{
struct clocksource *cs;
- if (!finished_booting || list_empty(&clocksource_list))
+ /* Pairs with the smp_store_release() in clocksource_done_booting(). */
+ if (!smp_load_acquire(&finished_booting) || list_empty(&clocksource_list))
return NULL;
/*
@@ -1188,7 +1189,8 @@ static int __init clocksource_done_booting(void)
{
mutex_lock(&clocksource_mutex);
curr_clocksource = clocksource_default_clock();
- finished_booting = 1;
+ /* Pairs with the smp_load_acquire() in clocksource_find_best(). */
+ smp_store_release(&finished_booting, 1);
/*
* Run the watchdog first to eliminate unstable clock sources
*/
---
base-commit: 93f51579e7df248780214094418f205253383cc5
change-id: 20260921-mb-clocksource-49cd0b246a89
Best regards,
--
Jaidev Shastri <jaidevshastri@vt.edu>
^ permalink raw reply [flat|nested] 2+ messages in thread
* Re: [PATCH] clocksource: publish finished_booting with release semantics
2026-09-22 1:22 [PATCH] clocksource: publish finished_booting with release semantics Jaidev Shastri via B4 Relay
@ 2026-09-29 11:53 ` Thomas Gleixner
0 siblings, 0 replies; 2+ messages in thread
From: Thomas Gleixner @ 2026-09-29 11:53 UTC (permalink / raw)
To: Jaidev Shastri via B4 Relay, John Stultz, Stephen Boyd, Miroslav Lichvar
Cc: linux-kernel, Jaidev Shastri
On Mon, Sep 21 2026 at 21:22, Jaidev Shastri via wrote:
> From: Jaidev Shastri <jaidevshastri@vt.edu>
>
> clocksource_done_booting() sets curr_clocksource and then
> finished_booting, both with plain stores under clocksource_mutex.
> clocksource_find_best() tests finished_booting with a plain load before
> it walks clocksource_list.
>
> Set the flag with smp_store_release() and read it with
> smp_load_acquire().
>
> Found with MBCheck, a static herd7-based memory consistency checker.
Unfortunately your tool is missing the point. All callchains which end
up in clocksource_find_best() hold clocksource_mutex. So this
acquire/release is pointless.
Thanks,
tglx
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2026-09-29 11:53 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2026-09-22 1:22 [PATCH] clocksource: publish finished_booting with release semantics Jaidev Shastri via B4 Relay
2026-09-29 11:53 ` Thomas Gleixner
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®