On Mon, 2012-05-14 at 17:33 -0600, Andreas Dilger wrote:
> Even better than just blindly bumping the i_version on every change,
> it would be better to have users of i_version (i.e. knfsd) flag the
> inode with "needs i_version update" then read the version. When the
> filesystem/VFS bumps i_version the next time it can clear this flag
> and not update i_version again until after the next time i_version
> is actually used.
That doubly penalises actual users of i_version: not only does the
filesystem have to bump it on a file change, but it now has to set and
write this flag on the next read of i_version.
If those reads of i_version were really few and far between, then I
agree that it might be a solution, but most NFS clients will ask for the
updated i_version after every change that they make to the file because
they need to track that value for cache consistency management purposes.
--
Trond Myklebust
Linux NFS client maintainer
NetApp
Trond.Myklebust@xxxxxxxxxx
www.netapp.com
��.n��������+%������w��{.n�����{�����ܨ}���Ơz�j:+v�����w����ޙ��&�)ߡ�a����z�ޗ���ݢj��w�f
