[PATCH v5 37/42] xfs: support idmapped mounts

[Christoph Hellwig]

On Fri, Jan 15, 2021 at 07:51:54AM +1100, Dave Chinner wrote:
> > @@ -813,7 +818,7 @@ xfs_setattr_nonsize(
> >  	 * 	     Posix ACL code seems to care about this issue either.
> >  	 */
> >  	if (mask & ATTR_MODE) {
> > -		error = posix_acl_chmod(&init_user_ns, inode, inode->i_mode);
> > +		error = posix_acl_chmod(mnt_userns, inode, inode->i_mode);
> >  		if (error)
> >  			return error;
> >  	}
> > @@ -868,7 +873,7 @@ xfs_setattr_size(
> >  		 * Use the regular setattr path to update the timestamps.
> >  		 */
> >  		iattr->ia_valid &= ~ATTR_SIZE;
> > -		return xfs_setattr_nonsize(ip, iattr);
> > +		return xfs_setattr_nonsize(&init_user_ns, ip, iattr);
> 
> Shouldn't that be passing mnt_userns?

As Christian already explained we an't hit this with anything related
to uids/gids, the only thing that will be updated are the timestamps,
as also mentioned in the comment that only makes it partially into the
diff context.

> >  	trace_xfs_setattr(ip);
> >  
> > -	error = xfs_vn_change_ok(dentry, iattr);
> > +	error = xfs_vn_change_ok(mnt_userns, dentry, iattr);
> >  	if (error)
> >  		return error;
> >  	return xfs_setattr_size(ip, iattr);
> 
> And this passing mnt_userns down into xfs_setattr_size()?  Seems
> like a bit of a landmine...

That being said we could just pass down the argument, even if it doesn't
make much sense for the size update.