> > > And I certainly agree that it ought to be replaced by will/wont pair to
> > > close the remount race. One that had been there all along. All fh_verify()
> > > callers of that kind need it - we want to pull mnt_{will,wont}_write()
> > > pair into callers *and* stretch to protect the entire relevant area.
> > >
> > > Which contains vfs_...() in case of nfsd_create, etc. See what I mean?
> > > That's exactly the thing I'd been talking about - the area we want to
> > > cover is _bigger_ than vfs_...() and contains nfsd-specific logic. IOW,
> > > doesn't get folded into any VFS-provided helper.
> >
> > I still don't get it why it needs to cover nfsd-specifi logic. What
> > does nfsd have to do with r/o mounts?
>
> Explain to me again, how fh_verify() manages to contain no nfsd-specific
> logics. Please.
I didn't say it doesn't contain nfsd specifics. What I said, that it
doesn't modify the filesystem. So there's no reason to cover it with
mnt_want_write()/drop_write().
> You are right - we do have races there. Always had.
> And nfsd_permission() is the next target for continuation of ro-bind
> series. Assuming that we don't simply make r/w export to hold will_write
> all along, in which case all these checks around calls of vfs_...() in
> there simply go away - that's also an arguable option.
Yes. And that _still_ doesn't make the path_*() interface wrong.
Miklos
--
To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
