[PATCH] sock: fix /proc/net/sockstat underflow in sk_clone_lock()

sk_clone_lock() needs to call get_net() and sock_inuse_inc() together, or
socket_seq_show() will underflow when __sk_free() from sk_free() from
sk_free_unlock_clone() is called.

Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
---
 net/core/sock.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/net/core/sock.c b/net/core/sock.c
index 8f2b2f2c0e7b..41e91d0f7061 100644
--- a/net/core/sock.c
+++ b/net/core/sock.c

@@ -2124,8 +2124,10 @@ struct sock *sk_clone_lock(const struct sock *sk, const gfp_t priority)
 	newsk->sk_prot_creator = prot;
 
 	/* SANITY */
-	if (likely(newsk->sk_net_refcnt))
+	if (likely(newsk->sk_net_refcnt)) {
 		get_net(sock_net(newsk));
+		sock_inuse_add(sock_net(newsk), 1);
+	}
 	sk_node_init(&newsk->sk_node);
 	sock_lock_init(newsk);
 	bh_lock_sock(newsk);

@@ -2197,8 +2199,6 @@ struct sock *sk_clone_lock(const struct sock *sk, const gfp_t priority)
 	newsk->sk_err_soft = 0;
 	newsk->sk_priority = 0;
 	newsk->sk_incoming_cpu = raw_smp_processor_id();
-	if (likely(newsk->sk_net_refcnt))
-		sock_inuse_add(sock_net(newsk), 1);
 
 	/* Before updating sk_refcnt, we must commit prior changes to memory
 	 * (Documentation/RCU/rculist_nulls.rst for details)

-- 
2.32.0

On 11/13/21 10:02 PM, Tetsuo Handa wrote:

sk_clone_lock() needs to call get_net() and sock_inuse_inc() together, or
socket_seq_show() will underflow when __sk_free() from sk_free() from
sk_free_unlock_clone() is called.

IMO, a "sock_inuse_get() underflow" is a very different problem,
I suspect this should be fixed with the following patch.

diff --git a/net/core/sock.c b/net/core/sock.c
index c57d9883f62c75f522b7f6bc68451aaf8429dc83..bac8e2b62521301ce897728fff9622c4c05419a3 100644
--- a/net/core/sock.c
+++ b/net/core/sock.c

@@ -3573,7 +3573,7 @@ int sock_inuse_get(struct net *net)
        for_each_possible_cpu(cpu)
                res += *per_cpu_ptr(net->core.sock_inuse, cpu);
 
-       return res;
+       return max(res, 0);
 }
 
 EXPORT_SYMBOL_GPL(sock_inuse_get);

Bug added in commit 648845ab7e200993dccd3948c719c858368c91e7
Author: Tonghao Zhang [off-list ref]
Date:   Thu Dec 14 05:51:58 2017 -0800

    sock: Move the socket inuse to namespace.

On 2021/11/15 5:01, Eric Dumazet wrote:

On 11/13/21 10:02 PM, Tetsuo Handa wrote:

quoted

sk_clone_lock() needs to call get_net() and sock_inuse_inc() together, or

s/sock_inuse_inc/sock_inuse_add/

quoted

socket_seq_show() will underflow when __sk_free() from sk_free() from
sk_free_unlock_clone() is called.

IMO, a "sock_inuse_get() underflow" is a very different problem,

Yes, a different problem. I found this problem while trying to examine
https://syzkaller.appspot.com/bug?extid=694120e1002c117747ed where
somebody might be failing to call get_net() or we might be failing to
make sure that all timers are synchronously stopped before put_net().

I suspect this should be fixed with the following patch.

My patch addresses a permanent underflow problem which remains as long as
that namespace exists. Your patch addresses a temporal underflow problem
which happens due to calculating the sum without locks. Therefore, we can
apply both patches if we want.

On 11/14/21 2:43 PM, Tetsuo Handa wrote:

On 2021/11/15 5:01, Eric Dumazet wrote:

quoted

On 11/13/21 10:02 PM, Tetsuo Handa wrote:

quoted

sk_clone_lock() needs to call get_net() and sock_inuse_inc() together, or

s/sock_inuse_inc/sock_inuse_add/

quoted

quoted

socket_seq_show() will underflow when __sk_free() from sk_free() from
sk_free_unlock_clone() is called.

IMO, a "sock_inuse_get() underflow" is a very different problem,

Yes, a different problem. I found this problem while trying to examine
https://syzkaller.appspot.com/bug?extid=694120e1002c117747ed where
somebody might be failing to call get_net() or we might be failing to
make sure that all timers are synchronously stopped before put_net().

quoted

I suspect this should be fixed with the following patch.

My patch addresses a permanent underflow problem which remains as long as
that namespace exists. Your patch addresses a temporal underflow problem
which happens due to calculating the sum without locks. Therefore, we can
apply both patches if we want.

I think your changelog is a bit confusing.

It would really help if you add the Fixes: tag, because if you do

1) It makes the patch much easier to understand, by reading again the the old patch.

2) We can tell if we can merge both fixes in the same submission (if they share the same bug origin)

Thanks !

sk_clone_lock() needs to call sock_inuse_add(1) before entering the
sk_free_unlock_clone() error path, for __sk_free() from sk_free() from
sk_free_unlock_clone() calls sock_inuse_add(-1).

Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Fixes: 648845ab7e200993 ("sock: Move the socket inuse to namespace.")
---
Changes in v2:
  Rewrite patch description.

 net/core/sock.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/net/core/sock.c b/net/core/sock.c
index 8f2b2f2c0e7b..41e91d0f7061 100644
--- a/net/core/sock.c
+++ b/net/core/sock.c

@@ -2124,8 +2124,10 @@ struct sock *sk_clone_lock(const struct sock *sk, const gfp_t priority)
 	newsk->sk_prot_creator = prot;
 
 	/* SANITY */
-	if (likely(newsk->sk_net_refcnt))
+	if (likely(newsk->sk_net_refcnt)) {
 		get_net(sock_net(newsk));
+		sock_inuse_add(sock_net(newsk), 1);
+	}
 	sk_node_init(&newsk->sk_node);
 	sock_lock_init(newsk);
 	bh_lock_sock(newsk);

@@ -2197,8 +2199,6 @@ struct sock *sk_clone_lock(const struct sock *sk, const gfp_t priority)
 	newsk->sk_err_soft = 0;
 	newsk->sk_priority = 0;
 	newsk->sk_incoming_cpu = raw_smp_processor_id();
-	if (likely(newsk->sk_net_refcnt))
-		sock_inuse_add(sock_net(newsk), 1);
 
 	/* Before updating sk_refcnt, we must commit prior changes to memory
 	 * (Documentation/RCU/rculist_nulls.rst for details)

-- 
2.32.0

Hello:

This patch was applied to netdev/net.git (master)
by David S. Miller [off-list ref]:

On Mon, 15 Nov 2021 19:16:56 +0900 you wrote:

sk_clone_lock() needs to call sock_inuse_add(1) before entering the
sk_free_unlock_clone() error path, for __sk_free() from sk_free() from
sk_free_unlock_clone() calls sock_inuse_add(-1).

Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Fixes: 648845ab7e200993 ("sock: Move the socket inuse to namespace.")

[...]

Here is the summary with links:
  - [v2] sock: fix /proc/net/sockstat underflow in sk_clone_lock()
    https://git.kernel.org/netdev/net/c/938cca9e4109

You are awesome, thank you!
-- 
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html