Re: [PATCH] hugepage: allow parallelization of the hugepage fault path

2013-07-23 Thread Hush Bensen
On 07/18/2013 05:07 PM, Joonsoo Kim wrote: On Wed, Jul 17, 2013 at 12:50:25PM -0700, Davidlohr Bueso wrote: From: Davidlohr Bueso - Cleaned up and forward ported to Linus' latest. - Cache aligned mutexes. - Keep non SMP systems using a single mutex. It was found that this mutex can become qu

Re: [PATCH] hugepage: allow parallelization of the hugepage fault path

2013-07-22 Thread Hush Bensen
On 07/18/2013 03:50 AM, Davidlohr Bueso wrote: From: David Gibson At present, the page fault path for hugepages is serialized by a single mutex. This is used to avoid spurious out-of-memory conditions when the hugepage pool is fully utilized (two processes or threads can race to instantiate the

Re: [PATCH] hugepage: allow parallelization of the hugepage fault path

2013-07-21 Thread Joonsoo Kim
On Fri, Jul 19, 2013 at 02:24:15PM -0700, Davidlohr Bueso wrote: > On Fri, 2013-07-19 at 17:14 +1000, David Gibson wrote: > > On Thu, Jul 18, 2013 at 05:42:35PM +0900, Joonsoo Kim wrote: > > > On Wed, Jul 17, 2013 at 12:50:25PM -0700, Davidlohr Bueso wrote: > > > > From: David Gibson > > > > > >

Re: [PATCH] hugepage: allow parallelization of the hugepage fault path

2013-07-19 Thread Davidlohr Bueso
On Fri, 2013-07-19 at 17:14 +1000, David Gibson wrote: > On Thu, Jul 18, 2013 at 05:42:35PM +0900, Joonsoo Kim wrote: > > On Wed, Jul 17, 2013 at 12:50:25PM -0700, Davidlohr Bueso wrote: > > > From: David Gibson > > > > > > At present, the page fault path for hugepages is serialized by a > > > si

Re: [PATCH] hugepage: allow parallelization of the hugepage fault path

2013-07-19 Thread David Gibson
On Thu, Jul 18, 2013 at 05:42:35PM +0900, Joonsoo Kim wrote: > On Wed, Jul 17, 2013 at 12:50:25PM -0700, Davidlohr Bueso wrote: > > From: David Gibson > > > > At present, the page fault path for hugepages is serialized by a > > single mutex. This is used to avoid spurious out-of-memory conditions

Re: [PATCH] hugepage: allow parallelization of the hugepage fault path

2013-07-18 Thread Davidlohr Bueso
On Thu, 2013-07-18 at 17:19 -0700, Davidlohr Bueso wrote: > On Thu, 2013-07-18 at 18:07 +0900, Joonsoo Kim wrote: > > On Wed, Jul 17, 2013 at 12:50:25PM -0700, Davidlohr Bueso wrote: > > > > > From: Davidlohr Bueso > > > > > > - Cleaned up and forward ported to Linus' latest. > > > - Cache align

Re: [PATCH] hugepage: allow parallelization of the hugepage fault path

2013-07-18 Thread Davidlohr Bueso
On Thu, 2013-07-18 at 18:07 +0900, Joonsoo Kim wrote: > On Wed, Jul 17, 2013 at 12:50:25PM -0700, Davidlohr Bueso wrote: > > > From: Davidlohr Bueso > > > > - Cleaned up and forward ported to Linus' latest. > > - Cache aligned mutexes. > > - Keep non SMP systems using a single mutex. > > > > It

Re: [PATCH] hugepage: allow parallelization of the hugepage fault path

2013-07-18 Thread Joonsoo Kim
On Wed, Jul 17, 2013 at 12:50:25PM -0700, Davidlohr Bueso wrote: > From: Davidlohr Bueso > > - Cleaned up and forward ported to Linus' latest. > - Cache aligned mutexes. > - Keep non SMP systems using a single mutex. > > It was found that this mutex can become quite contended > during the early

Re: [PATCH] hugepage: allow parallelization of the hugepage fault path

2013-07-18 Thread Joonsoo Kim
On Wed, Jul 17, 2013 at 12:50:25PM -0700, Davidlohr Bueso wrote: > From: David Gibson > > At present, the page fault path for hugepages is serialized by a > single mutex. This is used to avoid spurious out-of-memory conditions > when the hugepage pool is fully utilized (two processes or threads c

[PATCH] hugepage: allow parallelization of the hugepage fault path

2013-07-17 Thread Davidlohr Bueso
From: David Gibson At present, the page fault path for hugepages is serialized by a single mutex. This is used to avoid spurious out-of-memory conditions when the hugepage pool is fully utilized (two processes or threads can race to instantiate the same mapping with the last hugepage from the poo