Re: [PATCH v7 03/14] fs: provide accessors for ->i_state

[Dave Chinner <david@fromorbit.com>]

On Fri, Oct 10, 2025 at 04:44:19PM +0200, Jan Kara wrote:
> On Thu 09-10-25 09:59:17, Mateusz Guzik wrote:
> > +static inline void inode_state_set_raw(struct inode *inode,
> > +				       enum inode_state_flags_enum flags)
> > +{
> > +	WRITE_ONCE(inode->i_state, inode->i_state | flags);
> > +}
> 
> I think this shouldn't really exist as it is dangerous to use and if we
> deal with XFS, nobody will actually need this function.

XFS does it's own inode caching outside the VFS, so for the moment
it needs to have access to the same VFS inode initialisation APIs as
the core VFS inode cache instantiation functions to maintain the
same externally visible behaviours.

Yes, if we change how the VFS inode caches initialise inodes, we
have to update the XFS code, but that's always been the case. This
isn't very hard to do....

Keep in mind that XFS has been caching inodes outside the VFS and
doing external state initialisation since it was first ported to
Linux (i.e. ~25 years ago). It's kinda strange to suddenly hear
people claim that this sort of VFS inode state manipulation thing is
"too dangerous" to allow anyone to use given how long we've actually
been doing this....

-Dave.
-- 
Dave Chinner
david@fromorbit.com