diff options
| author | Peter Zijlstra <peterz@infradead.org> | 2017-01-19 14:36:33 +0100 |
|---|---|---|
| committer | Ingo Molnar <mingo@kernel.org> | 2017-01-20 02:38:46 +0100 |
| commit | acb04058de49458010c44bb35b849d45113fd668 (patch) | |
| tree | 6011ca77b47f21fcd5e65e5a2f02dddadd3f950c /kernel/sched/clock.c | |
| parent | 02cfdc95a0104fa5812d855d1e4ec687312aaa6f (diff) | |
sched/clock: Fix hotplug crash
Mike reported that he could trigger the WARN_ON_ONCE() in
set_sched_clock_stable() using hotplug.
This exposed a fundamental problem with the interface, we should never
mark the TSC stable if we ever find it to be unstable. Therefore
set_sched_clock_stable() is a broken interface.
The reason it existed is that not having it is a pain, it means all
relevant architecture code needs to call clear_sched_clock_stable()
where appropriate.
Of the three architectures that select HAVE_UNSTABLE_SCHED_CLOCK ia64
and parisc are trivial in that they never called
set_sched_clock_stable(), so add an unconditional call to
clear_sched_clock_stable() to them.
For x86 the story is a lot more involved, and what this patch tries to
do is ensure we preserve the status quo. So even is Cyrix or Transmeta
have usable TSC they never called set_sched_clock_stable() so they now
get an explicit mark unstable.
Reported-by: Mike Galbraith <efault@gmx.de>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Fixes: 9881b024b7d7 ("sched/clock: Delay switching sched_clock to stable")
Link: http://lkml.kernel.org/r/20170119133633.GB6536@twins.programming.kicks-ass.net
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Diffstat (limited to 'kernel/sched/clock.c')
| -rw-r--r-- | kernel/sched/clock.c | 29 |
1 files changed, 8 insertions, 21 deletions
diff --git a/kernel/sched/clock.c b/kernel/sched/clock.c index 7713b2b53f61..ad64efe41722 100644 --- a/kernel/sched/clock.c +++ b/kernel/sched/clock.c @@ -83,8 +83,15 @@ void sched_clock_init(void) } #ifdef CONFIG_HAVE_UNSTABLE_SCHED_CLOCK +/* + * We must start with !__sched_clock_stable because the unstable -> stable + * transition is accurate, while the stable -> unstable transition is not. + * + * Similarly we start with __sched_clock_stable_early, thereby assuming we + * will become stable, such that there's only a single 1 -> 0 transition. + */ static DEFINE_STATIC_KEY_FALSE(__sched_clock_stable); -static int __sched_clock_stable_early; +static int __sched_clock_stable_early = 1; /* * We want: ktime_get_ns() + gtod_offset == sched_clock() + raw_offset @@ -132,24 +139,6 @@ static void __set_sched_clock_stable(void) tick_dep_clear(TICK_DEP_BIT_CLOCK_UNSTABLE); } -void set_sched_clock_stable(void) -{ - __sched_clock_stable_early = 1; - - smp_mb(); /* matches sched_clock_init_late() */ - - /* - * This really should only be called early (before - * sched_clock_init_late()) when guestimating our sched_clock() is - * solid. - * - * After that we test stability and we can negate our guess using - * clear_sched_clock_stable, possibly from a watchdog. - */ - if (WARN_ON_ONCE(sched_clock_running == 2)) - __set_sched_clock_stable(); -} - static void __clear_sched_clock_stable(struct work_struct *work) { struct sched_clock_data *scd = this_scd(); @@ -199,8 +188,6 @@ void sched_clock_init_late(void) if (__sched_clock_stable_early) __set_sched_clock_stable(); - else - __clear_sched_clock_stable(NULL); } /* |