Re: [LKP] Re: [clocksource] 6c52b5f3cf: stress-ng.opcode.ops_per_sec -14.4% regression
From: Paul E. McKenney
Date: Fri Apr 23 2021 - 19:40:03 EST
On Fri, Apr 23, 2021 at 02:14:44PM -0700, Paul E. McKenney wrote:
> On Fri, Apr 23, 2021 at 09:14:49PM +0200, Thomas Gleixner wrote:
> > On Thu, Apr 22 2021 at 07:24, Paul E. McKenney wrote:
> > > On Thu, Apr 22, 2021 at 03:41:26PM +0800, Feng Tang wrote:
> > > So what are our options?
> > >
> > > 1. Clear CLOCK_SOURCE_MUST_VERIFY from tsc-early.
> > >
> > > 2. #1, but add tsc-early into the watchdog list and set
> > > CLOCK_SOURCE_MUST_VERIFY once it is better calibrated.
> > >
> > > 3. Add a field to struct clocksource that, if non-zero, gives
> > > the maximum drift in nanoseconds per half second (AKA
> > > WATCHDOG_INTERVAL). If zero, the WATCHDOG_MAX_SKEW value
> > > is used. Set this to (say) 150,000ns for tsc-early.
> > >
> > > 4. As noted earlier, increase WATCHDOG_MAX_SKEW to 150 microseconds,
> > > which again is not a good approach given the real-world needs
> > > of real-world applications.
> > >
> > > 5. Your ideas here.
> >
> > #3 or add a flag to the clocksource which says 'frequency is guesswork' and
> > increase the threshold based on that.
> >
> > If that flag is still set max_drift is != 0 after 20 seconds yell.
>
> I made it 60 seconds based on recent experience with large systems,
> but sounds good!
>
> And the calls to clocksource_unregister(&clocksource_tsc_early) mean
> that it is not necessary to actually clear the .max_drift field, if
> I understand correctly.
>
> It looks to me that init_tsc_clocksource() is invoked at device_initcall()
> time, and that it either immediately unregisters clocksource_tsc_early
> (known TSC frequency), or schedules a delayed work to make
> tsc_refine_calibration_work() do the calibration and the unregister.
>
> Please see below for an untested patch.
And after a bit of testing, with two of the N bugs fixed...
Thanx, Paul
------------------------------------------------------------------------
commit e85f58eefb8e4803a3a6a0c68e5bdd82ccf2b039
Author: Paul E. McKenney <paulmck@xxxxxxxxxx>
Date: Fri Apr 23 15:53:55 2021 -0700
clocksource: Forgive tsc_early pre-calibration drift
Because the x86 tsc_early clocksource is given a quick and semi-accurate
calibration (by design!), it might have drift rates well in excess of
the 0.1% limit that is in the process of being adopted.
Therefore, add a max_drift field to the clocksource structure that, when
non-zero, specifies the maximum allowable drift rate in nanoseconds over
a half-second period. The tsc_early clocksource initializes this to five
miliseconds, which corresponds to the 1% drift rate limit suggested by
Xing Zhengjun. This max_drift field is intended only for early boot,
so clocksource_watchdog() splats if it encounters a non-zero value in
this field more than 60 seconds after boot, inspired by a suggestion by
Thomas Gleixner.
Cc: John Stultz <john.stultz@xxxxxxxxxx>
Cc: Thomas Gleixner <tglx@xxxxxxxxxxxxx>
Cc: Stephen Boyd <sboyd@xxxxxxxxxx>
Cc: Jonathan Corbet <corbet@xxxxxxx>
Cc: Mark Rutland <Mark.Rutland@xxxxxxx>
Cc: Marc Zyngier <maz@xxxxxxxxxx>
Cc: Andi Kleen <ak@xxxxxxxxxxxxxxx>
Cc: Xing Zhengjun <zhengjun.xing@xxxxxxxxxxxxxxx>
Cc: Feng Tang <feng.tang@xxxxxxxxx>
Signed-off-by: Paul E. McKenney <paulmck@xxxxxxxxxx>
diff --git a/arch/x86/kernel/tsc.c b/arch/x86/kernel/tsc.c
index 56289170753c..7192b8950322 100644
--- a/arch/x86/kernel/tsc.c
+++ b/arch/x86/kernel/tsc.c
@@ -1127,6 +1127,7 @@ static int tsc_cs_enable(struct clocksource *cs)
static struct clocksource clocksource_tsc_early = {
.name = "tsc-early",
.rating = 299,
+ .max_drift = 5 * NSEC_PER_MSEC,
.read = read_tsc,
.mask = CLOCKSOURCE_MASK(64),
.flags = CLOCK_SOURCE_IS_CONTINUOUS |
diff --git a/include/linux/clocksource.h b/include/linux/clocksource.h
index 83a3ebff7456..07d25a158d7a 100644
--- a/include/linux/clocksource.h
+++ b/include/linux/clocksource.h
@@ -42,6 +42,10 @@ struct module;
* @shift: Cycle to nanosecond divisor (power of two)
* @max_idle_ns: Maximum idle time permitted by the clocksource (nsecs)
* @maxadj: Maximum adjustment value to mult (~11%)
+ * @max_drift: Maximum drift rate in nanoseconds per half second.
+ * Zero says to use default WATCHDOG_THRESHOLD.
+ * A non-zero value for ->max_drift more than 60 seconds after boot
+ * will result in a splat.
* @archdata: Optional arch-specific data
* @max_cycles: Maximum safe cycle value which won't overflow on
* multiplication
@@ -93,6 +97,7 @@ struct clocksource {
u32 shift;
u64 max_idle_ns;
u32 maxadj;
+ u32 max_drift;
#ifdef CONFIG_ARCH_CLOCKSOURCE_DATA
struct arch_clocksource_data archdata;
#endif
diff --git a/kernel/time/clocksource.c b/kernel/time/clocksource.c
index a1f90e2b1039..001a9ed1edf3 100644
--- a/kernel/time/clocksource.c
+++ b/kernel/time/clocksource.c
@@ -126,6 +126,7 @@ static void __clocksource_change_rating(struct clocksource *cs, int rating);
*/
#define WATCHDOG_INTERVAL (HZ >> 1)
#define WATCHDOG_THRESHOLD (200 * NSEC_PER_USEC)
+#define WATCHDOG_SYNC_FORGIVENESS (HZ * 60UL)
/*
* Maximum permissible delay between two readouts of the watchdog
@@ -377,6 +378,7 @@ static void clocksource_watchdog(struct timer_list *unused)
int next_cpu, reset_pending;
int64_t wd_nsec, cs_nsec;
struct clocksource *cs;
+ u32 md;
spin_lock(&watchdog_lock);
if (!watchdog_running)
@@ -423,7 +425,13 @@ static void clocksource_watchdog(struct timer_list *unused)
continue;
/* Check the deviation from the watchdog clocksource. */
- if (abs(cs_nsec - wd_nsec) > WATCHDOG_THRESHOLD) {
+ if (!cs->max_drift) {
+ md = WATCHDOG_MAX_SKEW;
+ } else {
+ WARN_ON_ONCE(time_after(jiffies, WATCHDOG_SYNC_FORGIVENESS));
+ md = cs->max_drift;
+ }
+ if (abs(cs_nsec - wd_nsec) > md) {
pr_warn("timekeeping watchdog on CPU%d: Marking clocksource '%s' as unstable because the skew is too large:\n",
smp_processor_id(), cs->name);
pr_warn(" '%s' wd_now: %llx wd_last: %llx mask: %llx\n",