| Message ID | 1476234815-19084-1-git-send-email-gavin.guo@canonical.com |
|---|---|
| State | New |
| Headers | show |
On 10/11/2016 06:13 PM, Gavin Guo wrote: > BugLink: http://bugs.launchpad.net/bugs/1518457 > > On an AWS t2.micro instance (Xeon E5-2670, 991MiB of memory). > Occasionally (about once a day), kswapd0 falls into a busy loop and > spins on 100% CPU usage indefinitely. Reject to do the zone balance > when the memory is too small. > > Signed-off-by: Dan Streetman <dan.streetman@canonical.com> > Signed-off-by: Gavin Guo <gavin.guo@canonical.com> Acked-by: Leann Ogasawara <leann.ogasawara@canonical.com> > --- > mm/vmscan.c | 7 +++++++ > 1 file changed, 7 insertions(+) > > diff --git a/mm/vmscan.c b/mm/vmscan.c > index 0c114e2b01d3..21755dca16c2 100644 > --- a/mm/vmscan.c > +++ b/mm/vmscan.c > @@ -2959,6 +2959,13 @@ static void age_active_anon(struct zone *zone, struct scan_control *sc) > static bool zone_balanced(struct zone *zone, int order, > unsigned long balance_gap, int classzone_idx) > { > + /* > + * if zone is so small that watermarks are the same, don't bother trying > + * to balance; kswapd would just spin continuously trying to balance it. > + */ > + if (low_wmark_pages(zone) == high_wmark_pages(zone)) > + return true; > + > if (!zone_watermark_ok_safe(zone, order, high_wmark_pages(zone) + > balance_gap, classzone_idx)) > return false; >
On 10/11/2016 06:20 PM, Leann Ogasawara wrote: > On 10/11/2016 06:13 PM, Gavin Guo wrote: >> BugLink: http://bugs.launchpad.net/bugs/1518457 >> >> On an AWS t2.micro instance (Xeon E5-2670, 991MiB of memory). >> Occasionally (about once a day), kswapd0 falls into a busy loop and >> spins on 100% CPU usage indefinitely. Reject to do the zone balance >> when the memory is too small. >> >> Signed-off-by: Dan Streetman <dan.streetman@canonical.com> >> Signed-off-by: Gavin Guo <gavin.guo@canonical.com> > > Acked-by: Leann Ogasawara <leann.ogasawara@canonical.com> Jay (on CC) has also tested this so add his: Tested-by: Jay Vosburgh <jay.vosburgh@canonical.com> >> --- >> mm/vmscan.c | 7 +++++++ >> 1 file changed, 7 insertions(+) >> >> diff --git a/mm/vmscan.c b/mm/vmscan.c >> index 0c114e2b01d3..21755dca16c2 100644 >> --- a/mm/vmscan.c >> +++ b/mm/vmscan.c >> @@ -2959,6 +2959,13 @@ static void age_active_anon(struct zone *zone, struct scan_control *sc) >> static bool zone_balanced(struct zone *zone, int order, >> unsigned long balance_gap, int classzone_idx) >> { >> + /* >> + * if zone is so small that watermarks are the same, don't bother trying >> + * to balance; kswapd would just spin continuously trying to balance it. >> + */ >> + if (low_wmark_pages(zone) == high_wmark_pages(zone)) >> + return true; >> + >> if (!zone_watermark_ok_safe(zone, order, high_wmark_pages(zone) + >> balance_gap, classzone_idx)) >> return false; >>
On Wed, Oct 12, 2016 at 09:13:35AM +0800, Gavin Guo wrote: > BugLink: http://bugs.launchpad.net/bugs/1518457 > > On an AWS t2.micro instance (Xeon E5-2670, 991MiB of memory). > Occasionally (about once a day), kswapd0 falls into a busy loop and > spins on 100% CPU usage indefinitely. Reject to do the zone balance > when the memory is too small. > > Signed-off-by: Dan Streetman <dan.streetman@canonical.com> > Signed-off-by: Gavin Guo <gavin.guo@canonical.com> Applied to xenial and yakkety.
diff --git a/mm/vmscan.c b/mm/vmscan.c index 0c114e2b01d3..21755dca16c2 100644 --- a/mm/vmscan.c +++ b/mm/vmscan.c @@ -2959,6 +2959,13 @@ static void age_active_anon(struct zone *zone, struct scan_control *sc) static bool zone_balanced(struct zone *zone, int order, unsigned long balance_gap, int classzone_idx) { + /* + * if zone is so small that watermarks are the same, don't bother trying + * to balance; kswapd would just spin continuously trying to balance it. + */ + if (low_wmark_pages(zone) == high_wmark_pages(zone)) + return true; + if (!zone_watermark_ok_safe(zone, order, high_wmark_pages(zone) + balance_gap, classzone_idx)) return false;