Re: [RFC] Support volatile range for anon vma

2012-11-01 Thread Minchan Kim
Hi Christoph, On Thu, Nov 01, 2012 at 08:26:09PM +, Christoph Lameter wrote: > On Fri, 26 Oct 2012, Minchan Kim wrote: > > > I guess it would improve system performance very well. > > But as I wrote down in description, downside of the patch is that we have to > > age anon lru although we

Re: [RFC] Support volatile range for anon vma

2012-11-01 Thread Christoph Lameter
On Fri, 26 Oct 2012, Minchan Kim wrote: > I guess it would improve system performance very well. > But as I wrote down in description, downside of the patch is that we have to > age anon lru although we don't have swap. But gain via the patch is bigger > than > loss via aging of anon lru when

Re: [RFC] Support volatile range for anon vma

2012-11-01 Thread Christoph Lameter
On Fri, 26 Oct 2012, Minchan Kim wrote: I guess it would improve system performance very well. But as I wrote down in description, downside of the patch is that we have to age anon lru although we don't have swap. But gain via the patch is bigger than loss via aging of anon lru when memory

Re: [RFC] Support volatile range for anon vma

2012-11-01 Thread Minchan Kim
Hi Christoph, On Thu, Nov 01, 2012 at 08:26:09PM +, Christoph Lameter wrote: On Fri, 26 Oct 2012, Minchan Kim wrote: I guess it would improve system performance very well. But as I wrote down in description, downside of the patch is that we have to age anon lru although we don't have

Re: [RFC] Support volatile range for anon vma

2012-10-25 Thread Minchan Kim
Hi Christoph, On Thu, Oct 25, 2012 at 03:19:27PM +, Christoph Lameter wrote: > On Thu, 25 Oct 2012, Minchan Kim wrote: > > > #endif > > + /* > > +* True if page in this vma is reclaimed. > > What does that mean? All pages in the vma have been cleared out? It means at least, more

Re: [RFC] Support volatile range for anon vma

2012-10-25 Thread Christoph Lameter
On Thu, 25 Oct 2012, Minchan Kim wrote: > #endif > + /* > + * True if page in this vma is reclaimed. What does that mean? All pages in the vma have been cleared out? > + TTU_IGNORE_VOLATILE = (1 << 11),/* ignore volatile */ > }; > #define TTU_ACTION(x) ((x) & TTU_ACTION_MASK) >

Re: [RFC] Support volatile range for anon vma

2012-10-25 Thread Christoph Lameter
On Thu, 25 Oct 2012, Minchan Kim wrote: #endif + /* + * True if page in this vma is reclaimed. What does that mean? All pages in the vma have been cleared out? + TTU_IGNORE_VOLATILE = (1 11),/* ignore volatile */ }; #define TTU_ACTION(x) ((x) TTU_ACTION_MASK) int

Re: [RFC] Support volatile range for anon vma

2012-10-25 Thread Minchan Kim
Hi Christoph, On Thu, Oct 25, 2012 at 03:19:27PM +, Christoph Lameter wrote: On Thu, 25 Oct 2012, Minchan Kim wrote: #endif + /* +* True if page in this vma is reclaimed. What does that mean? All pages in the vma have been cleared out? It means at least, more than one is

[RFC] Support volatile range for anon vma

2012-10-24 Thread Minchan Kim
This patch introudce new madvise behavior MADV_VOLATILE and MADV_NOVOLATILE for anonymous pages. It's different with John Stultz's version which considers only tmpfs. This patch cannot cover John's one so if below idea is proved as reasonable maybe, we can unify both concepts and I hope interface

[RFC] Support volatile range for anon vma

2012-10-24 Thread Minchan Kim
This patch introudce new madvise behavior MADV_VOLATILE and MADV_NOVOLATILE for anonymous pages. It's different with John Stultz's version which considers only tmpfs. This patch cannot cover John's one so if below idea is proved as reasonable maybe, we can unify both concepts and I hope interface