lkml.org 
[lkml]   [2005]   [Sep]   [20]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: [RFC PATCH 5/10] vfs: shared subtree aware bind mounts
On Fri, Sep 16, 2005 at 11:26:19AM -0700, Ram wrote:

This patch needs to be split *AND* accompanied by locking rules. It's
pretty much the core of the entire thing; if it's possible to offload
chunks elsewhere, life would become easier. Locking rules are badly
needed, along with the comments re "why can't that mntput()/dput()
block under a spinlock", etc.

BTW, how are you dealing with MS_MOVE?

> +void do_detach_prepare_mnt(struct vfsmount *mnt)
> +{
> + mnt->mnt_mountpoint->d_mounted--;
> + mntput(mnt->mnt_parent);
> + dput(mnt->mnt_mountpoint);
> + mnt->mnt_parent = mnt;
> +}

General note: mntput() should go _after_ dput() when we deal with pairs.
Doesn't cost anything, trivially safe.

> if (res) {
> spin_lock(&vfsmount_lock);
> + clean_propagation_reference(res);

Uh-oh... What makes that safe? We do mntput() here; are we guaranteed
that these pointers won't be the last references?

> + spin_lock(&vfspnode_lock);
> + propagate_abort_mount(m);

Calls do_detach_prepare() -> dput(), mntput(). At the very least such
cases need comments...

> +static void __do_make_private(struct vfsmount *mnt)
> +{
> + __do_make_slave(mnt);
> + list_del_init(&mnt->mnt_slave);
> + mnt->mnt_master = NULL;
> + set_mnt_private(mnt);
> +}
> +
> int do_make_private(struct vfsmount *mnt)
> {
> /*
> * a private mount is nothing but a
> * slave mount with no incoming
> * propagations.
> */
> spin_lock(&vfspnode_lock);
> - __do_make_slave(mnt);
> - list_del_init(&mnt->mnt_slave);
> + __do_make_private(mnt);
> spin_unlock(&vfspnode_lock);
> - mnt->mnt_master = NULL;
> - set_mnt_private(mnt);
> return 0;
> }

Why not do that from the very beginning, BTW?

> /*
> - * a unclonable mount is nothing but a
> + * a unclonable mount is a
> * private mount which is unclonnable.
> */
> spin_lock(&vfspnode_lock);
> - __do_make_slave(mnt);
> - list_del_init(&mnt->mnt_slave);
> + __do_make_private(mnt);
> spin_unlock(&vfspnode_lock);
> - mnt->mnt_master = NULL;
> set_mnt_unclonable(mnt);
> return 0;
> }
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

\
 
 \ /
  Last update: 2005-09-20 12:03    [W:0.050 / U:0.108 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site