[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <1474990799.28155.54.camel@edumazet-glaptop3.roam.corp.google.com>
Date: Tue, 27 Sep 2016 08:39:59 -0700
From: Eric Dumazet <eric.dumazet@...il.com>
To: Cyrill Gorcunov <gorcunov@...il.com>
Cc: David Ahern <dsa@...ulusnetworks.com>, netdev@...r.kernel.org,
linux-kernel@...r.kernel.org, David Miller <davem@...emloft.net>,
kuznet@....inr.ac.ru, jmorris@...ei.org, yoshfuji@...ux-ipv6.org,
kaber@...sh.net, avagin@...nvz.org, stephen@...workplumber.org
Subject: Re: [PATCH v4] net: ip, diag -- Add diag interface for raw sockets
On Tue, 2016-09-27 at 18:08 +0300, Cyrill Gorcunov wrote:
> +static struct sock *raw_sock_get(struct net *net, const struct inet_diag_req_v2 *r)
> +{
> + struct raw_hashinfo *hashinfo = raw_get_hashinfo(r);
> + struct sock *sk = NULL, *s;
> + int slot;
> +
> + if (IS_ERR(hashinfo))
> + return ERR_CAST(hashinfo);
> +
> + read_lock(&hashinfo->lock);
> + for (slot = 0; slot < RAW_HTABLE_SIZE; slot++) {
> + sk_for_each(s, &hashinfo->ht[slot]) {
> + sk = raw_lookup(net, s, r);
> + if (sk)
> + break;
> + }
> + }
> + if (sk && !atomic_inc_not_zero(&sk->sk_refcnt))
> + sk = NULL;
Minor detail, but note that raw sockets do not use rcu (yet)
Since you have read_lock(&hashinfo->lock), no writer can suddenly change
sk->sk_refcnt to a zero value.
Therefore, a mere "sock_hold(sk)" should be enough to get a reference on
the socket.
Using atomic_inc_not_zero() also works, but might distract/confuse the
next guy trying to understand this code ;)
> + read_unlock(&hashinfo->lock);
> +
> + return sk ? sk : ERR_PTR(-ENOENT);
> +}
> +
Powered by blists - more mailing lists