by Jeff Layton

[permalink] [raw]

Subject: Re: [PATCH 09/12] locks: helper functions for delegation breaking

On Tue, 9 Jul 2013 15:38:36 -0400
"J. Bruce Fields" <[email protected]> wrote:

> On Tue, Jul 09, 2013 at 09:23:10AM -0400, Jeff Layton wrote:
> > On Wed, 3 Jul 2013 16:12:33 -0400
> > "J. Bruce Fields" <[email protected]> wrote:
> >
> > > From: "J. Bruce Fields" <[email protected]>
> > >
> > > We'll need the same logic for rename and link.
> > >
> > > Signed-off-by: J. Bruce Fields <[email protected]>
> > > ---
> > > fs/namei.c | 13 +++----------
> > > include/linux/fs.h | 33 +++++++++++++++++++++++++++++++--
> > > 2 files changed, 34 insertions(+), 12 deletions(-)
> > >
> > > diff --git a/fs/namei.c b/fs/namei.c
> > > index cba3db1..a9d4031 100644
> > > --- a/fs/namei.c
> > > +++ b/fs/namei.c
> > > @@ -3401,14 +3401,9 @@ int vfs_unlink(struct inode *dir, struct dentry *dentry, struct inode **delegate
> > > else {
> > > error = security_inode_unlink(dir, dentry);
> > > if (!error) {
> > > - error = break_deleg(target, O_WRONLY|O_NONBLOCK);
> > > - if (error) {
> > > - if (error == -EWOULDBLOCK && delegated_inode) {
> > > - *delegated_inode = target;
> > > - ihold(target);
> > > - }
> > > + error = try_break_deleg(target, delegated_inode);
> > > + if (error)
> > > goto out;
> > > - }
> > > error = dir->i_op->unlink(dir, dentry);
> > > if (!error)
> > > dont_mount(dentry);
> > > @@ -3478,9 +3473,7 @@ exit2:
> > > iput(inode); /* truncate the inode here */
> > > inode = NULL;
> > > if (delegated_inode) {
> > > - error = break_deleg(delegated_inode, O_WRONLY);
> > > - iput(delegated_inode);
> > > - delegated_inode = NULL;
> > > + error = break_deleg_wait(&delegated_inode);
> > > if (!error)
> > > goto retry_deleg;
> > > }
> > > diff --git a/include/linux/fs.h b/include/linux/fs.h
> > > index f951588..c37e463 100644
> > > --- a/include/linux/fs.h
> > > +++ b/include/linux/fs.h
> > > @@ -1894,6 +1894,9 @@ extern bool our_mnt(struct vfsmount *mnt);
> > >
> > > extern int current_umask(void);
> > >
> > > +extern void ihold(struct inode * inode);
> > > +extern void iput(struct inode *);
> > > +
> > > /* /sys/fs */
> > > extern struct kobject *fs_kobj;
> > >
> > > @@ -1962,6 +1965,28 @@ static inline int break_deleg(struct inode *inode, unsigned int mode)
> > > return 0;
> > > }
> > >
> > > +static inline int try_break_deleg(struct inode *inode, struct inode **delegated_inode)
> > > +{
> > > + int ret;
> > > +
> > > + ret = break_deleg(inode, O_WRONLY|O_NONBLOCK);
> > > + if (ret == -EWOULDBLOCK && delegated_inode) {
> > > + *delegated_inode = inode;
> > > + ihold(inode);
> > > + }
> > > + return ret;
> > > +}
> > > +
> >
> > Actually, now that I look...
> >
> > Suppose a vfs_unlink caller passes in a NULL delegated_inode pointer.
> > He'll get back a -EWOULDBLOCK here if there's a delegation on it.
> > Presumably he'll just treat that as a hard error and the delegation
> > would never get broken. Is that expected?
>
> Yes.
>
> The callers that pass in NULL are callers who should really never
> encounter a write delegation. E.g. it might happen if you exported a
> filesystem that you're also stacking ecryptfs on top of (assuming that's
> possible). Maybe I'm missing something, but that looks pointless.
> Erroring out seems a safe way to handle it.
>
> It wouldn't be a problem to convert more of those callers to do the same
> retry loop, it just seems better to avoid the extra spaghetti to support
> what looks like a nutty use case.
>
> --b.

Ok, thanks for clarifying that. I agree that those others are nutty
use-cases, at least for now. If someone were to ever allow setting of
FL_DELEGs from userland though they'll probably need to contend with
them.

--
Jeff Layton <[email protected]>