lkml.org 
[lkml]   [2011]   [Jul]   [27]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
SubjectRe: [PATCH]vmscan: add block plug for page reclaim
From
Date
On Thu, 2011-07-28 at 09:15 +0800, Andrew Morton wrote:
> On Thu, 28 Jul 2011 09:04:20 +0800 Shaohua Li <shaohua.li@intel.com> wrote:
>
> > > Using an additional 44 bytes of stack on that path is also
> > > significant(ly bad). But we need to fix that problem anyway. One way
> > > we could improve things in mm/vmscan.c is to move the blk_plug into
> > > scan_control then get the scan_control off the stack in some manner.
> > > That's easy for kswapd: allocate one scan_control per kswapd at
> > > startup. Doing it for direct-reclaim would be a bit trickier...
> > unfortunately, the direct-reclaim case is what cares about stack.
> >
> > BTW, the scan_control can be dieted. may_unmap/may_swap/may_writepage
> > can be a bit. swappiness < 100, so can be a char. order <= 11, can be a
> > char. should I do it to cut the size?
>
> All five will fit in a 32-bit word, at some expense in code size.
oh, I missed the code size will increase, so it's not good then.

> But I think first it would be better to work on a way of getting it all
> off the stack, along with the blk_plug.
>
> Could be done with a per-cpu array and CPU pinning, but CPU pinning is
> a bit expensive nowadays. Could put a scan_control* into the
> tack_struct, but that's dopey.
looks it should be per task, as reclaim could sleep. either putting it
to task_struct or allocating it, both are not good.

Thanks,
Shaohua



\
 
 \ /
  Last update: 2011-07-28 03:37    [W:0.058 / U:0.732 seconds]
©2003-2018 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site