Messages in this thread Patch in this message | | | Date | Mon, 21 Jul 2014 21:44:08 -0400 | From | Steven Rostedt <> | Subject | [GIT PULL] tracing: Fix wraparound problems in "uptime" trace clock |
| |
Linus,
Tony Luck found that using the "uptime" trace clock that uses jiffies as a counter was converted to nanoseconds (silly), and after 1 hour 11 minutes and 34 seconds, this monotonic clock would wrap, causing havoc with the tracing system and making the clock useless.
He converted that clock to use jiffies_64 and made it into a counter instead of nanosecond conversions, and displayed the clock with the straight jiffy count, which works much better than it did in the past.
With your last email, it seems that you are a bit grumpy about fixes, I hope this is fine. You still scare me more than Greg does.
Please pull the latest trace-fixes-v3.16-rc6 tree, which can be found at:
git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace.git trace-fixes-v3.16-rc6
Tag SHA1: 86dfce3341c86da617cb3d03cad39ded60ed7543 Head SHA1: 58d4e21e50ff3cc57910a8abc20d7e14375d2f61
Tony Luck (1): tracing: Fix wraparound problems in "uptime" trace clock
---- kernel/trace/trace.c | 2 +- kernel/trace/trace_clock.c | 9 +++++---- 2 files changed, 6 insertions(+), 5 deletions(-) --------------------------- commit 58d4e21e50ff3cc57910a8abc20d7e14375d2f61 Author: Tony Luck <tony.luck@intel.com> Date: Fri Jul 18 11:43:01 2014 -0700
tracing: Fix wraparound problems in "uptime" trace clock The "uptime" trace clock added in: commit 8aacf017b065a805d27467843490c976835eb4a5 tracing: Add "uptime" trace clock that uses jiffies has wraparound problems when the system has been up more than 1 hour 11 minutes and 34 seconds. It converts jiffies to nanoseconds using: (u64)jiffies_to_usecs(jiffy) * 1000ULL but since jiffies_to_usecs() only returns a 32-bit value, it truncates at 2^32 microseconds. An additional problem on 32-bit systems is that the argument is "unsigned long", so fixing the return value only helps until 2^32 jiffies (49.7 days on a HZ=1000 system). Avoid these problems by using jiffies_64 as our basis, and not converting to nanoseconds (we do convert to clock_t because user facing API must not be dependent on internal kernel HZ values). Link: http://lkml.kernel.org/p/99d63c5bfe9b320a3b428d773825a37095bf6a51.1405708254.git.tony.luck@intel.com Cc: stable@vger.kernel.org # 3.10+ Fixes: 8aacf017b065 "tracing: Add "uptime" trace clock that uses jiffies" Signed-off-by: Tony Luck <tony.luck@intel.com> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index bda9621638cc..291397e66669 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -823,7 +823,7 @@ static struct { { trace_clock_local, "local", 1 }, { trace_clock_global, "global", 1 }, { trace_clock_counter, "counter", 0 }, - { trace_clock_jiffies, "uptime", 1 }, + { trace_clock_jiffies, "uptime", 0 }, { trace_clock, "perf", 1 }, ARCH_TRACE_CLOCKS }; diff --git a/kernel/trace/trace_clock.c b/kernel/trace/trace_clock.c index 26dc348332b7..57b67b1f24d1 100644 --- a/kernel/trace/trace_clock.c +++ b/kernel/trace/trace_clock.c @@ -59,13 +59,14 @@ u64 notrace trace_clock(void) /* * trace_jiffy_clock(): Simply use jiffies as a clock counter. + * Note that this use of jiffies_64 is not completely safe on + * 32-bit systems. But the window is tiny, and the effect if + * we are affected is that we will have an obviously bogus + * timestamp on a trace event - i.e. not life threatening. */ u64 notrace trace_clock_jiffies(void) { - u64 jiffy = jiffies - INITIAL_JIFFIES; - - /* Return nsecs */ - return (u64)jiffies_to_usecs(jiffy) * 1000ULL; + return jiffies_64_to_clock_t(jiffies_64 - INITIAL_JIFFIES); } /*
| |