lkml.org 
[lkml]   [2004]   [Jul]   [27]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
SubjectRe: [PATCH] fix readahead breakage for sequential after random reads
From
Date
On Mon, 2004-07-26 at 17:08, Andrew Morton wrote:
> Ram Pai <linuxram@us.ibm.com> wrote:
> >
> > Andrew,
> > Yes the patch fixes a valid bug.
> >
>
> Please don't top-post :(
> > RP
> >
> > On Mon, 2004-07-26 at 16:29, Andrew Morton wrote:
> > > Miklos Szeredi <miklos@szeredi.hu> wrote:
> > > >
> > > > Current readahead logic is broken when a random read pattern is
> > > > followed by a long sequential read. The cause is that on a window
> > > > miss ra->next_size is set to ra->average, but ra->average is only
> > > > updated at the end of a sequence, so window size will remain 1 until
> > > > the end of the sequential read.
> > > >
> > > > This patch fixes this by taking the current sequence length into
> > > > account (code taken from towards end of page_cache_readahead()), and
> > > > also setting ra->average to a decent value in handle_ra_miss() when
> > > > sequential access is detected.
> > >
> > > Thanks. Do you have any performance testing results from this patch?
> > >
> > Ram Pai <linuxram@us.ibm.com> wrote:
> >
> > Andrew,
> > Yes the patch fixes a valid bug.
>
> Fine, but the readahead code is performance-sensitive, and it takes quite
> some time for any regressions to be discovered. So I'm going to need to
> either sit on this patch for a very long time, or extensively test it
> myself, or await convincing test results from someone else.
>
> Can you help with that?

yes I will run all my standard testsuites before we take this patch.
(DSS workload, iozone, sysbench). I will get back with some results
sooon. Probably by the end of this week.

Also I think the bug that Miklos, found is really hard to reproduce. Did
he find this bug by code inspection? Its really really hard to get into
a state where the current window is of size 1 page with zero pages in
the readahead window, and then the sequential read pattern to just right
then.

RP

>

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

\
 
 \ /
  Last update: 2005-03-22 14:04    [W:0.059 / U:0.704 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site