On Tue, May 17, 2011 at 8:50 AM, Minchan Kim <minchan.kim@xxxxxxxxx> wrote:
> On Mon, May 16, 2011 at 7:27 PM, Mel Gorman <mgorman@xxxxxxx> wrote:
>> On Mon, May 16, 2011 at 05:58:59PM +0900, Minchan Kim wrote:
>>> On Mon, May 16, 2011 at 5:45 PM, Mel Gorman <mgorman@xxxxxxx> wrote:
>>> > On Mon, May 16, 2011 at 02:04:00PM +0900, Minchan Kim wrote:
>>> >> On Mon, May 16, 2011 at 1:21 PM, James Bottomley
>>> >> <James.Bottomley@xxxxxxxxxxxxxxxxxxxxx> wrote:
>>> >> > On Sun, 2011-05-15 at 19:27 +0900, KOSAKI Motohiro wrote:
>>> >> >> (2011/05/13 23:03), Mel Gorman wrote:
>>> >> >> > Under constant allocation pressure, kswapd can be in the situation where
>>> >> >> > sleeping_prematurely() will always return true even if kswapd has been
>>> >> >> > running a long time. Check if kswapd needs to be scheduled.
>>> >> >> >
>>> >> >> > Signed-off-by: Mel Gorman<mgorman@xxxxxxx>
>>> >> >> > ---
>>> >> >> > Â mm/vmscan.c | Â Â4 ++++
>>> >> >> > Â 1 files changed, 4 insertions(+), 0 deletions(-)
>>> >> >> >
>>> >> >> > diff --git a/mm/vmscan.c b/mm/vmscan.c
>>> >> >> > index af24d1e..4d24828 100644
>>> >> >> > --- a/mm/vmscan.c
>>> >> >> > +++ b/mm/vmscan.c
>>> >> >> > @@ -2251,6 +2251,10 @@ static bool sleeping_prematurely(pg_data_t *pgdat, int order, long remaining,
>>> >> >> > Â Â unsigned long balanced = 0;
>>> >> >> > Â Â bool all_zones_ok = true;
>>> >> >> >
>>> >> >> > + Â /* If kswapd has been running too long, just sleep */
>>> >> >> > + Â if (need_resched())
>>> >> >> > + Â Â Â Â Â return false;
>>> >> >> > +
>>> >> >>
>>> >> >> Hmm... I don't like this patch so much. because this code does
>>> >> >>
>>> >> >> - don't sleep if kswapd got context switch at shrink_inactive_list
>>> >> >
>>> >> > This isn't entirely true: Âneed_resched() will be false, so we'll follow
>>> >> > the normal path for determining whether to sleep or not, in effect
>>> >> > leaving the current behaviour unchanged.
>>> >> >
>>> >> >> - sleep if kswapd didn't
>>> >> >
>>> >> > This also isn't entirely true: whether need_resched() is true at this
>>> >> > point depends on a whole lot more that whether we did a context switch
>>> >> > in shrink_inactive. It mostly depends on how long we've been running
>>> >> > without giving up the CPU. ÂGenerally that will mean we've been round
>>> >> > the shrinker loop hundreds to thousands of times without sleeping.
>>> >> >
>>> >> >> It seems to be semi random behavior.
>>> >> >
>>> >> > Well, we have to do something. ÂChris Mason first suspected the hang was
>>> >> > a kswapd rescheduling problem a while ago. ÂWe tried putting
>>> >> > cond_rescheds() in several places in the vmscan code, but to no avail.
>>> >>
>>> >> Is it a result of Âtest with patch of Hannes(ie, !pgdat_balanced)?
>>> >>
>>> >> If it isn't, it would be nop regardless of putting cond_reshed at vmscan.c.
>>> >> Because, although we complete zone balancing, kswapd doesn't sleep as
>>> >> pgdat_balance returns wrong result. And at last VM calls
>>> >> balance_pgdat. In this case, balance_pgdat returns without any work as
>>> >> kswap couldn't find zones which have not enough free pages and goto
>>> >> out. kswapd could repeat this work infinitely. So you don't have a
>>> >> chance to call cond_resched.
>>> >>
>>> >> But if your test was with Hanne's patch, I am very curious how come
>>> >> kswapd consumes CPU a lot.
>>> >>
>>> >> > The need_resched() in sleeping_prematurely() seems to be about the best
>>> >> > option. ÂThe other option might be just to put a cond_resched() in
>>> >> > kswapd_try_to_sleep(), but that will really have about the same effect.
>>> >>
>>> >> I don't oppose it but before that, I think we have to know why kswapd
>>> >> consumes CPU a lot although we applied Hannes' patch.
>>> >>
>>> >
>>> > Because it's still possible for processes to allocate pages at the same
>>> > rate kswapd is freeing them leading to a situation where kswapd does not
>>> > consider the zone balanced for prolonged periods of time.
>>>
>>> We have cond_resched in shrink_page_list, shrink_slab and balance_pgdat.
>>> So I think kswapd can be scheduled out although it's scheduled in
>>> after a short time as task scheduled also need page reclaim. Although
>>> all task in system need reclaim, kswapd cpu 99% consumption is a
>>> natural result, I think.
>>> Do I miss something?
>>>
>>
>> Lets see;
>>
>> shrink_page_list() only applies if inactive pages were isolated
>> Â Â Â Âwhich in turn may not happen if all_unreclaimable is set in
>> Â Â Â Âshrink_zones(). If for whatver reason, all_unreclaimable is
>> Â Â Â Âset on all zones, we can miss calling cond_resched().
>>
>> shrink_slab only applies if we are reclaiming slab pages. If the first
>> Â Â Â Âshrinker returns -1, we do not call cond_resched(). If that
>> Â Â Â Âfirst shrinker is dcache and __GFP_FS is not set, direct
>> Â Â Â Âreclaimers will not shrink at all. However, if there are
>> Â Â Â Âenough of them running or if one of the other shrinkers
>> Â Â Â Âis running for a very long time, kswapd could be starved
>> Â Â Â Âacquiring the shrinker_rwsem and never reaching the
>> Â Â Â Âcond_resched().
>
> Don't we have to move cond_resched?
>
> diff --git a/mm/vmscan.c b/mm/vmscan.c
> index 292582c..633e761 100644
> --- a/mm/vmscan.c
> +++ b/mm/vmscan.c
> @@ -231,8 +231,10 @@ unsigned long shrink_slab(struct shrink_control *shrink,
> Â Â Â Âif (scanned == 0)
> Â Â Â Â Â Â Â Âscanned = SWAP_CLUSTER_MAX;
>
> - Â Â Â if (!down_read_trylock(&shrinker_rwsem))
> - Â Â Â Â Â Â Â return 1; Â Â Â /* Assume we'll be able to shrink next time */
> + Â Â Â if (!down_read_trylock(&shrinker_rwsem)) {
> + Â Â Â Â Â Â Â ret = 1;
> + Â Â Â Â Â Â Â goto out; /* Assume we'll be able to shrink next time */
> + Â Â Â }
>
> Â Â Â Âlist_for_each_entry(shrinker, &shrinker_list, list) {
> Â Â Â Â Â Â Â Âunsigned long long delta;
> @@ -280,12 +282,14 @@ unsigned long shrink_slab(struct shrink_control *shrink,
> Â Â Â Â Â Â Â Â Â Â Â Âcount_vm_events(SLABS_SCANNED, this_scan);
> Â Â Â Â Â Â Â Â Â Â Â Âtotal_scan -= this_scan;
>
> - Â Â Â Â Â Â Â Â Â Â Â cond_resched();
> Â Â Â Â Â Â Â Â}
>
> Â Â Â Â Â Â Â Âshrinker->nr += total_scan;
> + Â Â Â Â Â Â Â cond_resched();
> Â Â Â Â}
> Â Â Â Âup_read(&shrinker_rwsem);
> +out:
> + Â Â Â cond_resched();
> Â Â Â Âreturn ret;
> Â}
>
>
>>
>> balance_pgdat() only calls cond_resched if the zones are not
>> Â Â Â Âbalanced. For a high-order allocation that is balanced, it
>> Â Â Â Âchecks order-0 again. During that window, order-0 might have
>> Â Â Â Âbecome unbalanced so it loops again for order-0 and returns
>> Â Â Â Âthat was reclaiming for order-0 to kswapd(). It can then find
>> Â Â Â Âthat a caller has rewoken kswapd for a high-order and re-enters
>> Â Â Â Âbalance_pgdat() without ever have called cond_resched().
>
> If kswapd reclaims order-o followed by high order, it would have a
> chance to call cond_resched in shrink_page_list. But if all zones are
> all_unreclaimable is set, balance_pgdat could return any work.
Typo : without any work.
--
Kind regards,
Minchan Kim
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@xxxxxxxxxx For more info on Linux MM,
see: http://www.linux-mm.org/ .
Fight unfair telecom internet charges in Canada: sign http://stopthemeter.ca/
Don't email: <a href