svn commit: r254986 - head/sys/ufs/ufs

[Davide Italiano]

On Wed, Aug 28, 2013 at 12:06 PM, Ivan Voras <ivoras at freebsd.org> wrote:
> Author: ivoras
> Date: Wed Aug 28 10:06:20 2013
> New Revision: 254986
> URL: http://svnweb.freebsd.org/changeset/base/254986
>
> Log:
>   Take a very small step toward the Century of the Anchovy by increasing the
>   time dirhash entries stay in memory before being considered for eviction to
>   1 minute.
>
> Modified:
>   head/sys/ufs/ufs/ufs_dirhash.c
>
> Modified: head/sys/ufs/ufs/ufs_dirhash.c
> ==============================================================================
> --- head/sys/ufs/ufs/ufs_dirhash.c      Wed Aug 28 07:48:44 2013        (r254985)
> +++ head/sys/ufs/ufs/ufs_dirhash.c      Wed Aug 28 10:06:20 2013        (r254986)
> @@ -85,7 +85,7 @@ SYSCTL_INT(_vfs_ufs, OID_AUTO, dirhash_d
>  static int ufs_dirhashlowmemcount = 0;
>  SYSCTL_INT(_vfs_ufs, OID_AUTO, dirhash_lowmemcount, CTLFLAG_RD,
>      &ufs_dirhashlowmemcount, 0, "number of times low memory hook called");
> -static int ufs_dirhashreclaimage = 5;
> +static int ufs_dirhashreclaimage = 60;
>  SYSCTL_INT(_vfs_ufs, OID_AUTO, dirhash_reclaimage, CTLFLAG_RW,
>      &ufs_dirhashreclaimage, 0,
>      "max time in seconds of hash inactivity before deletion in low VM events");

Hi,
do you have any evidence that this change impacts positively (or
negatively) performances for some workloads? If yes, can you share?
Also, why did you choose the '60' value (rather than something else)?
I don't see any 'Reviewed by:' line in your commit message neither I
remember a public discussion on -current or -arch or -fs about this.
OTOH I think such changes deserve a wider discussion.

Thanks,

-- 
Davide

"There are no solved problems; there are only problems that are more
or less solved" -- Henri Poincare