Re: nfs slowdown since 2.5.4

Trond Myklebust (trond.myklebust@fys.uio.no)
Wed, 5 Jun 2002 14:16:56 +0200


--------------Boundary-00=_8GF833WG5IQCPQQAZPU2
Content-Type: text/plain;
charset="iso-8859-1"
Content-Transfer-Encoding: 8bit

On Wednesday 05 June 2002 13:50, Matthias Welk wrote:

> Yes, the lookup rate increased 1/3 compared to the unpatched version.
> I have attached updated results of the build process, because there was a
> mistake for the 2.4.18 version.

I'm still confused about this... It suggests a sign error in a comparison
somewhere, but I cannot for the life of me see where...

However I do see some cases where we should update timestamps on dentries, but
are cleearly forgetting to do so. If this patch too leads to increased lookup
rates...

Cheers,
Trond

--------------Boundary-00=_8GF833WG5IQCPQQAZPU2
Content-Type: text/x-diff;
charset="iso-8859-1";
name="gnarf.dif"
Content-Transfer-Encoding: 7bit
Content-Disposition: attachment; filename="gnarf.dif"

--- fs/nfs/dir.c.orig Tue Mar 12 16:35:02 2002
+++ fs/nfs/dir.c Wed Jun 5 14:09:32 2002
@@ -554,6 +554,8 @@
}
if (is_bad_inode(inode))
force_delete(inode);
+ /* When creating a negative dentry, we want to renew d_time */
+ nfs_renew_times(dentry);
iput(inode);
}

@@ -1064,6 +1066,7 @@
d_rehash(rehash);
if (!error && !S_ISDIR(old_inode->i_mode))
d_move(old_dentry, new_dentry);
+ nfs_renew_times(new_dentry);

/* new dentry created? */
if (dentry)

--------------Boundary-00=_8GF833WG5IQCPQQAZPU2--
-
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/