On Wed 16-08-17 11:33:37, Andreas Dilger wrote:
> On Aug 16, 2017, at 9:41 AM, Jan Kara <jack@xxxxxxx> wrote:
> >
> > Push down acquisition of dqio_sem into ->write_file_info() callback.
> > Mostly for consistency with other operations.
> >
> > Signed-off-by: Jan Kara <jack@xxxxxxx>
> > ---
> > fs/ocfs2/quota_global.c | 9 +++++++--
> > fs/quota/dquot.c | 10 +---------
> > fs/quota/quota_v1.c | 2 ++
> > fs/quota/quota_v2.c | 5 ++++-
> > 4 files changed, 14 insertions(+), 12 deletions(-)
> >
> > diff --git a/fs/ocfs2/quota_global.c b/fs/ocfs2/quota_global.c
> > index 4134d557a8e5..6f7c5bd3cbc6 100644
> > --- a/fs/ocfs2/quota_global.c
> > +++ b/fs/ocfs2/quota_global.c
> > @@ -443,13 +443,18 @@ static int __ocfs2_global_write_info(struct super_block *sb, int type)
> > int ocfs2_global_write_info(struct super_block *sb, int type)
>
> Would be nice if this was named ocfs2_global_write_file_info() so that
> it showed up in searches more easily, though not a huge deal. I guess
> the same for ocfs2_local_read_file_info() and ocfs2_local_free_file_info(),
> if we wanted to be consistent here, but that should go into a separate
> patch.
Well, rather than adding 'file' into the name I would add 'quota'
somewhere. But yes, the name seems too generic.
> > {
> > int err;
> > - struct ocfs2_mem_dqinfo *info = sb_dqinfo(sb, type)->dqi_priv;
> > + struct quota_info *dqopt = sb_dqopt(sb);
> > + struct ocfs2_mem_dqinfo *info = dqopt->info[type].dqi_priv;
> >
> > + down_write(&dqopt->dqio_sem);
> > err = ocfs2_qinfo_lock(info, 1);
> > - if (err < 0)
> > + if (err < 0) {
> > + up_write(&dqopt->dqio_sem);
> > return err;
> > + }
>
> goto out_sem?
>
> > err = __ocfs2_global_write_info(sb, type);
> > ocfs2_qinfo_unlock(info, 1);
>
> out_sem:
Thanks, done.
Honza
--
Jan Kara <jack@xxxxxxxx>
SUSE Labs, CR
