From: Jiang Biao <[email protected]>
delta in update_stats_wait_end() might be negative, which would
make following statistics go wrong.
Add protection for delta of wait time, like what have been done in
update_stats_enqueue_sleeper() for deltas of sleep/block time.
Signed-off-by: Jiang Biao <[email protected]>
---
kernel/sched/fair.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index c0374c1152e0..ac950ac950bc 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -917,6 +917,9 @@ update_stats_wait_end(struct cfs_rq *cfs_rq, struct sched_entity *se)
delta = rq_clock(rq_of(cfs_rq)) - schedstat_val(se->statistics.wait_start);
+ if ((s64)delta < 0)
+ delta = 0;
+
if (entity_is_task(se)) {
p = task_of(se);
if (task_on_rq_migrating(p)) {
--
2.21.0
On Sun, 17 Jan 2021 at 13:31, Jiang Biao <[email protected]> wrote:
>
> From: Jiang Biao <[email protected]>
>
> delta in update_stats_wait_end() might be negative, which would
> make following statistics go wrong.
Could you describe the use case that generates a negative delta ?
rq_clock is always increasing so this should not lead to a negative
value even if update_stats_wait_end/start are not called in the right
order,
This situation could happen after a migration if we forgot to call
update_stats_wait_start
>
> Add protection for delta of wait time, like what have been done in
> update_stats_enqueue_sleeper() for deltas of sleep/block time.
>
> Signed-off-by: Jiang Biao <[email protected]>
> ---
> kernel/sched/fair.c | 3 +++
> 1 file changed, 3 insertions(+)
>
> diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
> index c0374c1152e0..ac950ac950bc 100644
> --- a/kernel/sched/fair.c
> +++ b/kernel/sched/fair.c
> @@ -917,6 +917,9 @@ update_stats_wait_end(struct cfs_rq *cfs_rq, struct sched_entity *se)
>
> delta = rq_clock(rq_of(cfs_rq)) - schedstat_val(se->statistics.wait_start);
>
> + if ((s64)delta < 0)
> + delta = 0;
> +
> if (entity_is_task(se)) {
> p = task_of(se);
> if (task_on_rq_migrating(p)) {
> --
> 2.21.0
>
Hi, Vincent
On Mon, 18 Jan 2021 at 15:56, Vincent Guittot
<[email protected]> wrote:
>
> On Sun, 17 Jan 2021 at 13:31, Jiang Biao <[email protected]> wrote:
> >
> > From: Jiang Biao <[email protected]>
> >
> > delta in update_stats_wait_end() might be negative, which would
> > make following statistics go wrong.
>
> Could you describe the use case that generates a negative delta ?
>
> rq_clock is always increasing so this should not lead to a negative
> value even if update_stats_wait_end/start are not called in the right
> order,
Yes, indeed.
> This situation could happen after a migration if we forgot to call
> update_stats_wait_start
The migration case was what I worried about, but no regular use case
comes into my mind. :)
As an extreme case, would it be a problem if we disable/re-enable
sched_schedstats during migration?
static inline void
update_stats_wait_start(struct cfs_rq *cfs_rq, struct sched_entity *se)
{
u64 wait_start, prev_wait_start;
if (!schedstat_enabled()) // disable during migration
return; // return here, and skip updating wait_start
...
}
static inline void
update_stats_wait_end(struct cfs_rq *cfs_rq, struct sched_entity *se)
{
struct task_struct *p;
u64 delta;
if (!schedstat_enabled()) // re-enable again
return;
/*
* When the sched_schedstat changes from 0 to 1, some sched se
* maybe already in the runqueue, the se->statistics.wait_start
* will be 0.So it will let the delta wrong. We need to avoid this
* scenario.
*/
if (unlikely(!schedstat_val(se->statistics.wait_start)))
return;
//stale wait_start which might be bigger than rq_clock would
be used. -)
delta = rq_clock(rq_of(cfs_rq)) -
schedstat_val(se->statistics.wait_start);
...
Thanks a lot.
Regards,
Jiang
}
>
> >
> > Add protection for delta of wait time, like what have been done in
> > update_stats_enqueue_sleeper() for deltas of sleep/block time.
> >
> > Signed-off-by: Jiang Biao <[email protected]>
> > ---
> > kernel/sched/fair.c | 3 +++
> > 1 file changed, 3 insertions(+)
> >
> > diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
> > index c0374c1152e0..ac950ac950bc 100644
> > --- a/kernel/sched/fair.c
> > +++ b/kernel/sched/fair.c
> > @@ -917,6 +917,9 @@ update_stats_wait_end(struct cfs_rq *cfs_rq, struct sched_entity *se)
> >
> > delta = rq_clock(rq_of(cfs_rq)) - schedstat_val(se->statistics.wait_start);
> >
> > + if ((s64)delta < 0)
> > + delta = 0;
> > +
> > if (entity_is_task(se)) {
> > p = task_of(se);
> > if (task_on_rq_migrating(p)) {
> > --
> > 2.21.0
> >
On Mon, 18 Jan 2021 at 15:11, Jiang Biao <[email protected]> wrote:
>
> Hi, Vincent
>
> On Mon, 18 Jan 2021 at 15:56, Vincent Guittot
> <[email protected]> wrote:
> >
> > On Sun, 17 Jan 2021 at 13:31, Jiang Biao <[email protected]> wrote:
> > >
> > > From: Jiang Biao <[email protected]>
> > >
> > > delta in update_stats_wait_end() might be negative, which would
> > > make following statistics go wrong.
> >
> > Could you describe the use case that generates a negative delta ?
> >
> > rq_clock is always increasing so this should not lead to a negative
> > value even if update_stats_wait_end/start are not called in the right
> > order,
> Yes, indeed.
>
> > This situation could happen after a migration if we forgot to call
> > update_stats_wait_start
> The migration case was what I worried about, but no regular use case
> comes into my mind. :)
IIUC, you haven't faced the problem and it's only based on studying the code.
> As an extreme case, would it be a problem if we disable/re-enable
> sched_schedstats during migration?
>
> static inline void
> update_stats_wait_start(struct cfs_rq *cfs_rq, struct sched_entity *se)
> {
> u64 wait_start, prev_wait_start;
>
> if (!schedstat_enabled()) // disable during migration
> return; // return here, and skip updating wait_start
> ...
> }
>
> static inline void
> update_stats_wait_end(struct cfs_rq *cfs_rq, struct sched_entity *se)
> {
> struct task_struct *p;
> u64 delta;
>
> if (!schedstat_enabled()) // re-enable again
> return;
>
> /*
> * When the sched_schedstat changes from 0 to 1, some sched se
> * maybe already in the runqueue, the se->statistics.wait_start
> * will be 0.So it will let the delta wrong. We need to avoid this
> * scenario.
> */
> if (unlikely(!schedstat_val(se->statistics.wait_start)))
> return;
> //stale wait_start which might be bigger than rq_clock would
> be used. -)
> delta = rq_clock(rq_of(cfs_rq)) -
> schedstat_val(se->statistics.wait_start);
> ...
>
> Thanks a lot.
> Regards,
> Jiang
>
> }
> >
> > >
> > > Add protection for delta of wait time, like what have been done in
> > > update_stats_enqueue_sleeper() for deltas of sleep/block time.
> > >
> > > Signed-off-by: Jiang Biao <[email protected]>
> > > ---
> > > kernel/sched/fair.c | 3 +++
> > > 1 file changed, 3 insertions(+)
> > >
> > > diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
> > > index c0374c1152e0..ac950ac950bc 100644
> > > --- a/kernel/sched/fair.c
> > > +++ b/kernel/sched/fair.c
> > > @@ -917,6 +917,9 @@ update_stats_wait_end(struct cfs_rq *cfs_rq, struct sched_entity *se)
> > >
> > > delta = rq_clock(rq_of(cfs_rq)) - schedstat_val(se->statistics.wait_start);
> > >
> > > + if ((s64)delta < 0)
> > > + delta = 0;
> > > +
> > > if (entity_is_task(se)) {
> > > p = task_of(se);
> > > if (task_on_rq_migrating(p)) {
> > > --
> > > 2.21.0
> > >
Hi, Vincent
On Mon, 18 Jan 2021 at 23:32, Vincent Guittot
<[email protected]> wrote:
>
> On Mon, 18 Jan 2021 at 15:11, Jiang Biao <[email protected]> wrote:
> >
> > Hi, Vincent
> >
> > On Mon, 18 Jan 2021 at 15:56, Vincent Guittot
> > <[email protected]> wrote:
> > >
> > > On Sun, 17 Jan 2021 at 13:31, Jiang Biao <[email protected]> wrote:
> > > >
> > > > From: Jiang Biao <[email protected]>
> > > >
> > > > delta in update_stats_wait_end() might be negative, which would
> > > > make following statistics go wrong.
> > >
> > > Could you describe the use case that generates a negative delta ?
> > >
> > > rq_clock is always increasing so this should not lead to a negative
> > > value even if update_stats_wait_end/start are not called in the right
> > > order,
> > Yes, indeed.
> >
> > > This situation could happen after a migration if we forgot to call
> > > update_stats_wait_start
> > The migration case was what I worried about, but no regular use case
> > comes into my mind. :)
>
> IIUC, you haven't faced the problem and it's only based on studying the code.
Not yet. :). Just found there are protections for
sleep_time/block_time, but no protection
for wait_time.
Think more later, the sleep_time/block_time do need to be protected
for the migration case,
because update_stats_enqueue_sleeper could be called right after
migration with src cpu's
sleep_start/block_start. But wait_time is not the case.
The following case might be too extreme to happen. :)
Thanks a lot for your patience.
Regards,
Jiang
>
> > As an extreme case, would it be a problem if we disable/re-enable
> > sched_schedstats during migration?
> >
> > static inline void
> > update_stats_wait_start(struct cfs_rq *cfs_rq, struct sched_entity *se)
> > {
> > u64 wait_start, prev_wait_start;
> >
> > if (!schedstat_enabled()) // disable during migration
> > return; // return here, and skip updating wait_start
> > ...
> > }
> >
> > static inline void
> > update_stats_wait_end(struct cfs_rq *cfs_rq, struct sched_entity *se)
> > {
> > struct task_struct *p;
> > u64 delta;
> >
> > if (!schedstat_enabled()) // re-enable again
> > return;
> >
> > /*
> > * When the sched_schedstat changes from 0 to 1, some sched se
> > * maybe already in the runqueue, the se->statistics.wait_start
> > * will be 0.So it will let the delta wrong. We need to avoid this
> > * scenario.
> > */
> > if (unlikely(!schedstat_val(se->statistics.wait_start)))
> > return;
> > //stale wait_start which might be bigger than rq_clock would
> > be used. -)
> > delta = rq_clock(rq_of(cfs_rq)) -
> > schedstat_val(se->statistics.wait_start);
> > ...
> >
> > Thanks a lot.
> > Regards,
> > Jiang