Re: [PATCH] mm: do not sleep in balance_pgdat if there's no i/o congestion

2012-12-29 Thread Zlatko Calusic
On 29.12.2012 08:25, Hillf Danton wrote: On Thu, Dec 27, 2012 at 11:42 PM, Zlatko Calusic wrote: On 21.12.2012 12:51, Hillf Danton wrote: On Thu, Dec 20, 2012 at 7:25 AM, Zlatko Calusic wrote: static unsigned long balance_pgdat(pg_data_t *pgdat, int order,

Re: [PATCH] mm: do not sleep in balance_pgdat if there's no i/o congestion

2012-12-28 Thread Hillf Danton
On Thu, Dec 27, 2012 at 11:42 PM, Zlatko Calusic wrote: > On 21.12.2012 12:51, Hillf Danton wrote: >> >> On Thu, Dec 20, 2012 at 7:25 AM, Zlatko Calusic >> wrote: >>> >>> static unsigned long balance_pgdat(pg_data_t *pgdat, int order, >>>

Re: [PATCH] mm: do not sleep in balance_pgdat if there's no i/o congestion

2012-12-27 Thread Zlatko Calusic
On 21.12.2012 12:51, Hillf Danton wrote: On Thu, Dec 20, 2012 at 7:25 AM, Zlatko Calusic wrote: static unsigned long balance_pgdat(pg_data_t *pgdat, int order, int *classzone_idx) { - int all_zones_ok; + struct zone *unbal

Re: [PATCH] mm: do not sleep in balance_pgdat if there's no i/o congestion

2012-12-21 Thread Hillf Danton
On Thu, Dec 20, 2012 at 7:25 AM, Zlatko Calusic wrote: > static unsigned long balance_pgdat(pg_data_t *pgdat, int order, > int *classzone_idx) > { > - int all_zones_ok; > + struct zone *unbalanced_zone; nit: less hunks if not e

Re: [PATCH] mm: do not sleep in balance_pgdat if there's no i/o congestion

2012-12-20 Thread Andrew Morton
On Thu, 20 Dec 2012 11:12:08 + Mel Gorman wrote: > On Thu, Dec 20, 2012 at 12:17:07AM +0100, Zlatko Calusic wrote: > > On a 4GB RAM machine, where Normal zone is much smaller than > > DMA32 zone, the Normal zone gets fragmented in time. This requires > > relatively more pressure in balance_pg

Re: [PATCH] mm: do not sleep in balance_pgdat if there's no i/o congestion

2012-12-20 Thread Mel Gorman
On Thu, Dec 20, 2012 at 12:17:07AM +0100, Zlatko Calusic wrote: > On a 4GB RAM machine, where Normal zone is much smaller than > DMA32 zone, the Normal zone gets fragmented in time. This requires > relatively more pressure in balance_pgdat to get the zone above the > required watermark. Unfortunate

Re: [PATCH] mm: do not sleep in balance_pgdat if there's no i/o congestion

2012-12-19 Thread Zlatko Calusic
On a 4GB RAM machine, where Normal zone is much smaller than DMA32 zone, the Normal zone gets fragmented in time. This requires relatively more pressure in balance_pgdat to get the zone above the required watermark. Unfortunately, the congestion_wait() call in there slows it down for a completely w

[PATCH] mm: do not sleep in balance_pgdat if there's no i/o congestion

2012-12-19 Thread Zlatko Calusic
On a 4GB RAM machine, where Normal zone is much smaller than DMA32 zone, the Normal zone gets fragmented in time. This requires relatively more pressure in balance_pgdat to get the zone above the required watermark. Unfortunately, the congestion_wait() call in there slows it down for a completely w