Hi Greg,
In the 3.4 tree, can you consider a revert of commit
48f0f14ffb6ff4852922994d11fbda418d40100e?
It was a cherry pick of:
"sched: Add missing call to calc_load_exit_idle()"
commit 749c8814f08f12baa4a9c2812a7c6ede7d69507d upstream.
However Mark was reading the code and comparing to the original
and by chance noticed that it got mis-applied. It was meant to
patch tick_nohz_restart_sched_tick() but instead landed in
tick_nohz_update_jiffies().
I dug in a bit and found that this happens because 3.4 doesn't
have a tick_nohz_restart_sched_tick() yet. Instead, the code
in question lives in tick_nohz_idle_exit(), and the "missing"
call to calc_load_exit_idle() is in fact not missing in 3.4.x;
we've got it via our 3.4 cherry pick of commit
5167e8d5417bf5c322a703d2927daec727ea40dd
"sched/nohz: Rewrite and fix load-avg computation -- again"
So I think it makes sense to just drop it vs. looking for
a place where it should be applied to.
I checked the stable queue to see what other releases might
have the same issue and it happened in 3.5.x as well, but
that is EOL, so I guess it doesn't matter there.
Thanks,
Paul.
--
To unsubscribe from this list: send the line "unsubscribe stable" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html