Skip to contentSkip to conversation

time: Revert to calling clock_was_set_delayed() while in irq context

PATCH In review John Stultz 3 participants 3 messages Mar 27, 2014 – Mar 28, 2014 Timekeeping, clocksource core, NTP, alarmtimer

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 time: Revert to calling clock_was_set_delayed() while in irq context

Ingo,

After testing a patch queued for 3.15, I noticed there was a WARNON message that I realized was triggered by one of my patches that came in the 3.14 merge window. I'm not very happy to find this now, but sadly here it is. The fix is luckily fairly simple and just falls back to previous behavior.

For tip/timers/urgent.

thanks
-john

In commit 47a1b796306356f35 (tick/timekeeping: Call update_wall_time outside the jiffies lock), we moved to calling clock_was_set() due to the fact that we were no longer holding the timekeeping or jiffies lock.

However, there is still the problem that clock_was_set() triggers an IPI, which cannot be done from the timer's hard irq context, and will generate WARN_ON warnings.

Apparently in my earlier testing, I'm guessing I didn't bump the dmesg log level, so I somehow missed the WARN_ONs.

Thus we need to revert back to calling clock_was_set_delayed().

Cc Thomas Gleixner <[email protected]>
Cc Ingo Molnar <[email protected]>
Cc Linus Torvalds <[email protected]>
---
 kernel/time/timekeeping.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
kernel/time/timekeeping.c +2 -1
index 0aa4ce8..5b40279 100644
@@ -1435,7 +1435,8 @@ void update_wall_time(void)
 out:
 	raw_spin_unlock_irqrestore(&timekeeper_lock, flags);
 	if (clock_set)
-		clock_was_set();
+		/* Have to call _delayed version, since in irq context*/
+		clock_was_set_delayed();
 }
 
 /**
Signature · 1 line
1.8.3.2
#1 lore raw
Quoted 9 lines · John Stultz

Ingo,

After testing a patch queued for 3.15, I noticed there was a WARNON message that I realized was triggered by one of my patches that came in the 3.14 merge window. I'm not very happy to find this now, but sadly here it is. The fix is luckily fairly simple and just falls back to previous behavior.

For tip/timers/urgent.

No problem - I've applied your fix to timers/urgent and will get it to Linus later today, after a bit of testing.

Thanks,

Ingo

#2 lore raw

[tip:timers/urgent] time: Revert to calling clock_was_set_delayed () while in irq context

In reply to John Stultz
time: Revert to calling clock_was_set_delayed() while in irq context
Commit-ID:  cab5e127eef040399902caa8e1510795583fa03a
Gitweb:     http://git.kernel.org/tip/cab5e127eef040399902caa8e1510795583fa03a
Author:     John Stultz <[email protected]>
AuthorDate: Thu, 27 Mar 2014 16:30:49 -0700
Committer:  Ingo Molnar <[email protected]>
CommitDate: Fri, 28 Mar 2014 08:07:07 +0100

time: Revert to calling clock_was_set_delayed() while in irq context

In commit 47a1b796306356f35 ("tick/timekeeping: Call update_wall_time outside the jiffies lock"), we moved to calling clock_was_set() due to the fact that we were no longer holding the timekeeping or jiffies lock.

However, there is still the problem that clock_was_set() triggers an IPI, which cannot be done from the timer's hard irq context, and will generate WARN_ON warnings.

Apparently in my earlier testing, I'm guessing I didn't bump the dmesg log level, so I somehow missed the WARN_ONs.

Thus we need to revert back to calling clock_was_set_delayed().

---
 kernel/time/timekeeping.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
kernel/time/timekeeping.c +2 -1
index 0aa4ce8..5b40279 100644
@@ -1435,7 +1435,8 @@ void update_wall_time(void)
 out:
 	raw_spin_unlock_irqrestore(&timekeeper_lock, flags);
 	if (clock_set)
-		clock_was_set();
+		/* Have to call _delayed version, since in irq context*/
+		clock_was_set_delayed();
 }
 
 /**