On Tue, 16 Dec 2008, Mike Galbraith wrote:
> On Tue, 2008-12-16 at 01:49 +0100, Miklos Szeredi wrote:
> > On Sat, 13 Dec 2008, Rafael J. Wysocki wrote:
> > > Bug-Entry : http://bugzilla.kernel.org/show_bug.cgi?id=12208
> > > Subject : uml is very slow on 2.6.28 host
> > > Submitter : Miklos Szeredi <[email protected]>
> > > Date : 2008-12-12 9:35 (2 days old)
> > > References : http://marc.info/?l=linux-kernel&m=122907463518593&w=4
> >
> > I did a bisection, and this is the commit which is responsible:
> >
> > commit 464b75273f64be7c81fee975bd6ca9593df3427b
> > Author: Peter Zijlstra <[email protected]>
> > Date: Fri Oct 24 11:06:15 2008 +0200
> >
> > sched: re-instate vruntime based wakeup preemption
> >
> > The advantage is that vruntime based wakeup preemption has a better
> > conceptual model. Here wakeup_gran = 0 means: preempt when 'fair'.
> > Therefore wakeup_gran is the granularity of unfairness we allow in
> > order
> > to make progress.
> >
> > Signed-off-by: Peter Zijlstra <[email protected]>
> > Acked-by: Mike Galbraith <[email protected]>
> > Signed-off-by: Ingo Molnar <[email protected]>
>
> If that commit is responsible, then it should also be very slow in pre
> 28 kernels, where the same exists.
Everything prior to 2.6.28 was fine in this respect, so there must be
some subtle difference.
> Hm, there's another possibility.
> Can you try echo NO_LAST_BUDDY > /sys/kernel/debug/sched_features?
It didn't help, unfortunately.
Applying this patch on top of latest git (which essentially reverts
the above commit) fixes the slowness.
Thanks,
Miklos
---
kernel/sched_fair.c | 98 +++-------------------------------------------------
1 file changed, 6 insertions(+), 92 deletions(-)
Index: linux.git/kernel/sched_fair.c
===================================================================
--- linux.git.orig/kernel/sched_fair.c 2008-12-16 01:34:26.000000000 +0100
+++ linux.git/kernel/sched_fair.c 2008-12-16 01:34:54.000000000 +0100
@@ -143,49 +143,6 @@ static inline struct sched_entity *paren
return se->parent;
}
-/* return depth at which a sched entity is present in the hierarchy */
-static inline int depth_se(struct sched_entity *se)
-{
- int depth = 0;
-
- for_each_sched_entity(se)
- depth++;
-
- return depth;
-}
-
-static void
-find_matching_se(struct sched_entity **se, struct sched_entity **pse)
-{
- int se_depth, pse_depth;
-
- /*
- * preemption test can be made between sibling entities who are in the
- * same cfs_rq i.e who have a common parent. Walk up the hierarchy of
- * both tasks until we find their ancestors who are siblings of common
- * parent.
- */
-
- /* First walk up until both entities are at same depth */
- se_depth = depth_se(*se);
- pse_depth = depth_se(*pse);
-
- while (se_depth > pse_depth) {
- se_depth--;
- *se = parent_entity(*se);
- }
-
- while (pse_depth > se_depth) {
- pse_depth--;
- *pse = parent_entity(*pse);
- }
-
- while (!is_same_group(*se, *pse)) {
- *se = parent_entity(*se);
- *pse = parent_entity(*pse);
- }
-}
-
#else /* CONFIG_FAIR_GROUP_SCHED */
static inline struct rq *rq_of(struct cfs_rq *cfs_rq)
@@ -236,11 +193,6 @@ static inline struct sched_entity *paren
return NULL;
}
-static inline void
-find_matching_se(struct sched_entity **se, struct sched_entity **pse)
-{
-}
-
#endif /* CONFIG_FAIR_GROUP_SCHED */
@@ -1291,8 +1243,8 @@ static unsigned long wakeup_gran(struct
* More easily preempt - nice tasks, while not making it harder for
* + nice tasks.
*/
- if (!sched_feat(ASYM_GRAN) || se->load.weight > NICE_0_LOAD)
- gran = calc_delta_fair(sysctl_sched_wakeup_granularity, se);
+ if (sched_feat(ASYM_GRAN))
+ gran = calc_delta_mine(gran, NICE_0_LOAD, &se->load);
return gran;
}
@@ -1345,6 +1268,7 @@ static void check_preempt_wakeup(struct
{
struct task_struct *curr = rq->curr;
struct sched_entity *se = &curr->se, *pse = &p->se;
+ s64 delta_exec;
if (unlikely(rt_prio(p->prio))) {
struct cfs_rq *cfs_rq = task_cfs_rq(curr);
@@ -1398,19 +1322,9 @@ static void check_preempt_wakeup(struct
return;
}
- find_matching_se(&se, &pse);
-
- while (se) {
- BUG_ON(!pse);
-
- if (wakeup_preempt_entity(se, pse) == 1) {
- resched_task(curr);
- break;
- }
-
- se = parent_entity(se);
- pse = parent_entity(pse);
- }
+ delta_exec = se->sum_exec_runtime - se->prev_sum_exec_runtime;
+ if (delta_exec > wakeup_gran(pse))
+ resched_task(curr);
}
static struct task_struct *pick_next_task_fair(struct rq *rq)
--
To unsubscribe from this list: send the line "unsubscribe kernel-testers" in
the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
