Re: [PATCH] f2fs: introduce on-disk layout version checking functionality

From: Jaegeuk Kim
Date: Wed May 25 2016 - 13:12:56 EST


Hello,

On Tue, May 24, 2016 at 06:05:23PM -0700, Viacheslav Dubeyko wrote:
> Hi Jaegeuk,
>
> On Mon, 2016-05-23 at 14:13 -0700, Jaegeuk Kim wrote:
> > Hi Slava,
> >
> > On Thu, May 19, 2016 at 10:46:06AM -0700, Viacheslav Dubeyko wrote:
> > ...
> > >
> > > +#ifdef CONFIG_F2FS_16TB_VOLUME_SUPPORT
> > > +#define F2FS_MAX_SUPP_MAJOR_VERSION (2)
> > > +#define F2FS_MIN_16TB_VOLUME_SUPPORT_VERSION (2)
> > > +#else
> > > +#define F2FS_MAX_SUPP_MAJOR_VERSION (1)
> > > +#endif
> > > +
> > ...
> > >
> > > +static int f2fs_check_version_and_features(struct super_block *sb,
> > > + struct f2fs_super_block *raw_super)
> > > +{
> > > + u16 major_ver = le16_to_cpu(raw_super->major_ver);
> > > + u32 feature = le32_to_cpu(raw_super->feature);
> > > +
> > > + if (major_ver > F2FS_MAX_SUPP_MAJOR_VERSION) {
> >
> > This means, for example, f2fs driver in v4.8 will deny to mount a partition
> > formatted by mkfs.f2fs v3.x, which doesn't make sense, IIUC.
> >
>
> I didn't catch the point. Maybe, I've missed something but, as far as I
> can judge, f2fs driver v.4.8 will mount as old version of on-disk layout
> as the new one. But right now it doesn't make sense to discuss this
> topic because we haven't consent about ideology of this patch.
>
> > As Christoph mentioned, how about checking the feature only like this?
> >
> > 1. if the feature is ON,
> > - go 64 bits , when compiled w/ F2FS_MIN_16TB_VOLUME_SUPPORT
> > - fail to mount, when compiled w/o F2FS_MIN_16TB_VOLUME_SUPPORT
> >
> > 2. if the feature is OFF,
> > - fail to mount, when compiled w/ F2FS_MIN_16TB_VOLUME_SUPPORT
> > - go 32 bits , when compiled w/o F2FS_MIN_16TB_VOLUME_SUPPORT
> >
> > Thoughts?
> >
>
> So, my logic is simple. We are trying to modify the on-disk layout. As a
> result, we need to check the on-disk layout version, from my viewpoint.
> And this modification is not "feature" itself but simple bug fix. And I
> believe that "major_ver", "minor_ver" in F2FS superblock is the on-disk
> layout version.

Hmm, the versions are to indicate f2fs-tools, not on-disk layout something.
They are simply growing as whatever reasons such as bug fixes, new features,
and so on to provide debugging information.

Thanks,

>
> What do you think? Do you still believe that it should be a feature
> flag?
>
> Thanks,
> Vyacheslav Dubeyko.
>