Skip to contentSkip to conversation

timekeeping: Copy the shadow-timekeeper over the real timekeeper last

PATCH New John Stultz 2 participants 2 messages Jun 17, 2015 – Jun 18, 2015 Timekeeping, clocksource core, NTP, alarmtimer

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 timekeeping: Copy the shadow-timekeeper over the real timekeeper last

The fix in d151832650ed9 (time: Move clock_was_set_seq update before updating shadow-timekeeper) was unfortunately incomplete.

The main gist of that change was to do the shadow-copy update last, so that any state changes were properly duplicated, and we wouldn't accidentally have stale data in the shadow.

Unfortunately in the main update_wall_time() logic, we update use the shadow-timekeeper to calculate the next update values, then while holding the lock, copy the shadow-timekeeper over, then call timekeeping_update() to do some additional bookkeeping, (skipping the shadow mirror). The bug with this is the additional bookkeeping isn't all read-only, and some changes timkeeper state. Thus we might then overwrite this state change on the next update.

To avoid this problem, do the timekeeping_update() on the shadow-timekeeper prior to copying the full state over to the real-timekeeper.

This avoids problems with both the clock_was_set_seq and next_leap_ktime being overwritten and possibly the fast-timekeepers as well.

Many thanks to Prarit for his rigorous testing, which discovered this problem, along with Prarit and Daniel's work validating this fix.

Cc Prarit Bhargava <[email protected]>
Cc Daniel Bristot de Oliveira <[email protected]>
Cc Richard Cochran <[email protected]>
Cc Jan Kara <[email protected]>
Cc Jiri Bohac <[email protected]>
Cc Ingo Molnar <[email protected]>
Cc Thomas Gleixner <[email protected]>

---
(This patch applies to tip/timers/core)

 kernel/time/timekeeping.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
kernel/time/timekeeping.c +2 -1
index 5d67ffb..30b7a40 100644
@@ -1853,8 +1853,9 @@ void update_wall_time(void)
 	 * memcpy under the tk_core.seq against one before we start
 	 * updating.
 	 */
+	timekeeping_update(tk, clock_set);
 	memcpy(real_tk, tk, sizeof(*tk));
-	timekeeping_update(real_tk, clock_set);
+	/* The memcpy must come last. Do not put anything here! */
 	write_seqcount_end(&tk_core.seq);
 out:
 	raw_spin_unlock_irqrestore(&timekeeper_lock, flags);
Signature · 1 line
1.9.1
#1 lore raw
timekeeping: Copy the shadow-timekeeper over the real timekeeper last
Commit-ID:  906c55579a6360dd9ef5a3101bb2e3ae396dfb97
Gitweb:     http://git.kernel.org/tip/906c55579a6360dd9ef5a3101bb2e3ae396dfb97
Author:     John Stultz <[email protected]>
AuthorDate: Wed, 17 Jun 2015 10:05:53 -0700
Committer:  Thomas Gleixner <[email protected]>
CommitDate: Thu, 18 Jun 2015 09:27:02 +0200

timekeeping: Copy the shadow-timekeeper over the real timekeeper last

The fix in d151832650ed9 (time: Move clock_was_set_seq update before updating shadow-timekeeper) was unfortunately incomplete.

The main gist of that change was to do the shadow-copy update last, so that any state changes were properly duplicated, and we wouldn't accidentally have stale data in the shadow.

Unfortunately in the main update_wall_time() logic, we update use the shadow-timekeeper to calculate the next update values, then while holding the lock, copy the shadow-timekeeper over, then call timekeeping_update() to do some additional bookkeeping, (skipping the shadow mirror). The bug with this is the additional bookkeeping isn't all read-only, and some changes timkeeper state. Thus we might then overwrite this state change on the next update.

To avoid this problem, do the timekeeping_update() on the shadow-timekeeper prior to copying the full state over to the real-timekeeper.

This avoids problems with both the clock_was_set_seq and next_leap_ktime being overwritten and possibly the fast-timekeepers as well.

Many thanks to Prarit for his rigorous testing, which discovered this problem, along with Prarit and Daniel's work validating this fix.

---
 kernel/time/timekeeping.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
kernel/time/timekeeping.c +2 -1
index 5d67ffb..30b7a40 100644
@@ -1853,8 +1853,9 @@ void update_wall_time(void)
 	 * memcpy under the tk_core.seq against one before we start
 	 * updating.
 	 */
+	timekeeping_update(tk, clock_set);
 	memcpy(real_tk, tk, sizeof(*tk));
-	timekeeping_update(real_tk, clock_set);
+	/* The memcpy must come last. Do not put anything here! */
 	write_seqcount_end(&tk_core.seq);
 out:
 	raw_spin_unlock_irqrestore(&timekeeper_lock, flags);