[PATCH 4/7] cputime: Correctly handle task guest time on housekeepers

From: Frederic Weisbecker
Date: Thu Nov 19 2015 - 10:49:17 EST

When a task runs on a housekeeper (a CPU running with the periodic tick
with neighbours running tickless), it doesn't account cputime using vtime
but relies on the tick. Such a task has its vtime_snap_whence value set

Readers won't handle that correctly though. As long as vtime is running
on some CPU, readers incorretly assume that vtime runs on all CPUs and
always compute the tickless cputime delta, which is only junk on

So lets fix this with checking that the target runs on a vtime CPU through
the appropriate state check before computing the tickless delta.

Cc: Christoph Lameter <cl@xxxxxxxxx>
Cc: Chris Metcalf <cmetcalf@xxxxxxxxxx>
Cc: Ingo Molnar <mingo@xxxxxxxxxx>
Cc: Luiz Capitulino <lcapitulino@xxxxxxxxxx>
Cc: Peter Zijlstra <peterz@xxxxxxxxxxxxx>
Cc: Rik van Riel <riel@xxxxxxxxxx>
Cc: Thomas Gleixner <tglx@xxxxxxxxxxxxx>
Cc: Paul E. McKenney <paulmck@xxxxxxxxxxxxxxxxxx>
Cc: Hiroshi Shimamoto <h-shimamoto@xxxxxxxxxxxxx>
Signed-off-by: Frederic Weisbecker <fweisbec@xxxxxxxxx>
kernel/sched/cputime.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/kernel/sched/cputime.c b/kernel/sched/cputime.c
index 4a18a6e..5cf24e7 100644
--- a/kernel/sched/cputime.c
+++ b/kernel/sched/cputime.c
@@ -795,7 +795,7 @@ cputime_t task_gtime(struct task_struct *t)
seq = read_seqbegin(&t->vtime_seqlock);

gtime = t->gtime;
- if (t->flags & PF_VCPU)
+ if (t->vtime_snap_whence == VTIME_SYS && t->flags & PF_VCPU)
gtime += vtime_delta(t);

} while (read_seqretry(&t->vtime_seqlock, seq));

To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/