All of lore.kernel.org
 help / color / mirror / Atom feed
From: Sven Eckelmann <sven@narfation.org>
To: b.a.t.m.a.n@lists.open-mesh.org
Cc: Marek Lindner <mareklindner@neomailbox.ch>
Subject: Re: [B.A.T.M.A.N.] [PATCH maint] batman-adv: Reduce refcnt of removed router when updating route
Date: Sun, 20 Mar 2016 12:01:09 +0100	[thread overview]
Message-ID: <16765982.vnPfJFa6WA@sven-edge> (raw)
In-Reply-To: <9014983.RsdMSR9Elr@voltaire>

[-- Attachment #1: Type: text/plain, Size: 1671 bytes --]

On Sunday 20 March 2016 18:45:29 Marek Lindner wrote:
> On Saturday, March 05, 2016 15:53:47 Sven Eckelmann wrote:
> > --- a/net/batman-adv/routing.c
> > +++ b/net/batman-adv/routing.c
> > @@ -104,6 +104,8 @@ static void _batadv_update_route(struct batadv_priv
> > *bat_priv, neigh_node = NULL;
> > 
> >         spin_lock_bh(&orig_node->neigh_list_lock);
> > 
> > +       curr_router = rcu_dereference_protected(orig_ifinfo->router,
> > true);
> > +
> > 
> >         rcu_assign_pointer(orig_ifinfo->router, neigh_node);
> >         spin_unlock_bh(&orig_node->neigh_list_lock);
> >         batadv_orig_ifinfo_free_ref(orig_ifinfo);
> 
> Don't we also need to check for curr_router->refcount > 0 to mimic the check
> above ? Maybe a negative refcount does not hurt or is it unsigned ?

If this one gets negative then we would have a bug in a different place. The 
assignment only happens in this neigh_list_lock protected block. So the 
neigh_node behind orig_ifinfo->router must at least have a reference count of 
1 or there was no valid reference (as in reference counter) for the pointer.

The the kref_get_unless_zero before was only necessary because the curr_router 
was aquired inside a rcu_read_lock protected region which is not perfectly in 
sync with its writers. So it could happen that rcu_dereference returned a 
pointer to a neigh_node but this neigh_node will be free'd (reference counter 
== 0). And we cannot get a valid reference for an object which has refcount of 
0. This function avoids this problem by assuming that orig_ifinfo->router is 
NULL. This is not perfectly correct but better than having a pointer to free'd 
memory.

Kind regards,
	Sven

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 819 bytes --]

      reply	other threads:[~2016-03-20 11:01 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-03-05 14:53 [B.A.T.M.A.N.] [PATCH maint] batman-adv: Reduce refcnt of removed router when updating route Sven Eckelmann
2016-03-20 10:45 ` Marek Lindner
2016-03-20 11:01   ` Sven Eckelmann [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=16765982.vnPfJFa6WA@sven-edge \
    --to=sven@narfation.org \
    --cc=b.a.t.m.a.n@lists.open-mesh.org \
    --cc=mareklindner@neomailbox.ch \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.