Re: [RFC] mm: support MIGRATE_DISCARD

2012-09-05 Thread Mel Gorman
On Fri, Aug 24, 2012 at 01:25:30PM +0900, Minchan Kim wrote: > This patch introudes MIGRATE_DISCARD mode in migration. s/introudes/introduces/ It's also not a "mode" like the other MIGRATE_ modes but is instead a modifier. > It drops *unmapped clean cache pages* instead of migration so that > mi

Re: [RFC] mm: support MIGRATE_DISCARD

2012-08-24 Thread Michal Nazarewicz
Minchan Kim writes: > This patch introudes MIGRATE_DISCARD mode in migration. > It drops *unmapped clean cache pages* instead of migration so that > migration latency could be reduced by avoiding (memcpy + page remapping). > It's useful for CMA because latency of migration is very important rather

Re: [RFC] mm: support MIGRATE_DISCARD

2012-08-24 Thread Minchan Kim
Hi Rik, On Fri, Aug 24, 2012 at 11:04:16AM -0400, Rik van Riel wrote: > On 08/24/2012 12:25 AM, Minchan Kim wrote: > >This patch introudes MIGRATE_DISCARD mode in migration. > >It drops *unmapped clean cache pages* instead of migration so that > > Am I confused, or does the code not match the cha

Re: [RFC] mm: support MIGRATE_DISCARD

2012-08-24 Thread Rik van Riel
On 08/24/2012 12:25 AM, Minchan Kim wrote: This patch introudes MIGRATE_DISCARD mode in migration. It drops *unmapped clean cache pages* instead of migration so that Am I confused, or does the code not match the changelog? It looks like it is still trying to discard mapped page cache pages:

[RFC] mm: support MIGRATE_DISCARD

2012-08-23 Thread Minchan Kim
This patch introudes MIGRATE_DISCARD mode in migration. It drops *unmapped clean cache pages* instead of migration so that migration latency could be reduced by avoiding (memcpy + page remapping). It's useful for CMA because latency of migration is very important rather than eviction of background