ipv6: Fix "waiting for %s to become free" hang

From: Krishna Kumar <krkumar2@in.ibm.com>

(From code-walkthrough, and untested)

Fix "waiting for %s to become free" hang. dev has an
extra hold (by ip6mr_reg_vif) which was not dropped
before calling unregister_netdevice.

Signed-off-by: Krishna Kumar <krkumar2@in.ibm.com>
---
 net/ipv6/ip6mr.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Krishna Kumar a écrit :
> From: Krishna Kumar <krkumar2@in.ibm.com>
> 
> (From code-walkthrough, and untested)
> 
> Fix "waiting for %s to become free" hang. dev has an
> extra hold (by ip6mr_reg_vif) which was not dropped
> before calling unregister_netdevice.
> 
> Signed-off-by: Krishna Kumar <krkumar2@in.ibm.com>
> ---
>  net/ipv6/ip6mr.c |    2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff -ruNp org/net/ipv6/ip6mr.c new/net/ipv6/ip6mr.c
> --- org/net/ipv6/ip6mr.c	2009-12-04 16:08:07.000000000 +0530
> +++ new/net/ipv6/ip6mr.c	2009-12-04 16:09:56.000000000 +0530
> @@ -648,8 +648,8 @@ static int mif6_add(struct net *net, str
>  			return -ENOBUFS;
>  		err = dev_set_allmulti(dev, 1);
>  		if (err) {
> -			unregister_netdevice(dev);
>  			dev_put(dev);
> +			unregister_netdevice(dev);
>  			return err;
>  		}
>  		break;

This changes nothing, since we hold rtnl at this point.

unregister_netdevice() only queues the device (net_set_todo()) in a list.

And we process this list later, when renl_unlock() is finally called.

--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

> Eric Dumazet <eric.dumazet@gmail.com>
>
> Re: [PATCH] ipv6: Fix "waiting for %s to become free" hang
>
> Krishna Kumar a écrit :
> > From: Krishna Kumar <krkumar2@in.ibm.com>
> >
> > (From code-walkthrough, and untested)
> >
> > Fix "waiting for %s to become free" hang. dev has an
> > extra hold (by ip6mr_reg_vif) which was not dropped
> > before calling unregister_netdevice.
> >
> > Signed-off-by: Krishna Kumar <krkumar2@in.ibm.com>
> > ---
> >  net/ipv6/ip6mr.c |    2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff -ruNp org/net/ipv6/ip6mr.c new/net/ipv6/ip6mr.c
> > --- org/net/ipv6/ip6mr.c   2009-12-04 16:08:07.000000000 +0530
> > +++ new/net/ipv6/ip6mr.c   2009-12-04 16:09:56.000000000 +0530
> > @@ -648,8 +648,8 @@ static int mif6_add(struct net *net, str
> >           return -ENOBUFS;
> >        err = dev_set_allmulti(dev, 1);
> >        if (err) {
> > -         unregister_netdevice(dev);
> >           dev_put(dev);
> > +         unregister_netdevice(dev);
> >           return err;
> >        }
> >        break;
>
> This changes nothing, since we hold rtnl at this point.
>
> unregister_netdevice() only queues the device (net_set_todo()) in a list.
>
> And we process this list later, when renl_unlock() is finally called.

Yes, you are right. I sent the patch on my memory of unregister_netdev,
which gets
the rtnl and drops it, and that would have caused this problem. Please drop
the patch.

thanks,

- KK

--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html