Re: [RFC PATCH v2] overlayfs: Provide mount options sync=off/fs to skip sync

[Vivek Goyal <vgoyal@redhat.com>]

On Wed, Jul 01, 2020 at 10:08:28PM +0300, Amir Goldstein wrote:
[..]
> > diff --git a/fs/overlayfs/readdir.c b/fs/overlayfs/readdir.c
> > index 6918b98faeb6..970319ca1623 100644
> > --- a/fs/overlayfs/readdir.c
> > +++ b/fs/overlayfs/readdir.c
> > @@ -858,11 +858,15 @@ static int ovl_dir_fsync(struct file *file, loff_t start, loff_t end,
> >         struct ovl_dir_file *od = file->private_data;
> >         struct dentry *dentry = file->f_path.dentry;
> >         struct file *realfile = od->realfile;
> > +       struct ovl_fs *ofs = OVL_FS(dentry->d_sb);
> >
> >         /* Nothing to sync for lower */
> >         if (!OVL_TYPE_UPPER(ovl_path_type(dentry)))
> >                 return 0;
> >
> > +       if (ofs->config.nosync || ofs->config.syncfs)
> > +               return 0;
> > +
> 
> Generally looks good, but those test conditions are quite weird IMO.
> I would go for either enum or flags, but not up to me to decide.
> But at the very least use inline helpers ovl_should_fsync(),
> ovl_should_syncfs(), because if we want to add new sync modes or
> whatever going over all those conditions and
> changing them would be sub-optimal.

Ok, I will add inline helpers.

I am not sure, what will I gain by switching to using enums or flags.
It will be just different style to keep same information.

Thanks
Vivek