[PATCH 3.2 027/107] net: Fix RCU splat in af_key

From: Ben Hutchings
Date: Thu Oct 08 2015 - 20:27:29 EST


3.2.72-rc1 review patch. If anyone has any objections, please let me know.

------------------

From: David Ahern <dsa@xxxxxxxxxxxxxxxxxxx>

commit ba51b6be38c122f7dab40965b4397aaf6188a464 upstream.

Hit the following splat testing VRF change for ipsec:

[ 113.475692] ===============================
[ 113.476194] [ INFO: suspicious RCU usage. ]
[ 113.476667] 4.2.0-rc6-1+deb7u2+clUNRELEASED #3.2.65-1+deb7u2+clUNRELEASED Not tainted
[ 113.477545] -------------------------------
[ 113.478013] /work/monster-14/dsa/kernel.git/include/linux/rcupdate.h:568 Illegal context switch in RCU read-side critical section!
[ 113.479288]
[ 113.479288] other info that might help us debug this:
[ 113.479288]
[ 113.480207]
[ 113.480207] rcu_scheduler_active = 1, debug_locks = 1
[ 113.480931] 2 locks held by setkey/6829:
[ 113.481371] #0: (&net->xfrm.xfrm_cfg_mutex){+.+.+.}, at: [<ffffffff814e9887>] pfkey_sendmsg+0xfb/0x213
[ 113.482509] #1: (rcu_read_lock){......}, at: [<ffffffff814e767f>] rcu_read_lock+0x0/0x6e
[ 113.483509]
[ 113.483509] stack backtrace:
[ 113.484041] CPU: 0 PID: 6829 Comm: setkey Not tainted 4.2.0-rc6-1+deb7u2+clUNRELEASED #3.2.65-1+deb7u2+clUNRELEASED
[ 113.485422] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.7.5.1-0-g8936dbb-20141113_115728-nilsson.home.kraxel.org 04/01/2014
[ 113.486845] 0000000000000001 ffff88001d4c7a98 ffffffff81518af2 ffffffff81086962
[ 113.487732] ffff88001d538480 ffff88001d4c7ac8 ffffffff8107ae75 ffffffff8180a154
[ 113.488628] 0000000000000b30 0000000000000000 00000000000000d0 ffff88001d4c7ad8
[ 113.489525] Call Trace:
[ 113.489813] [<ffffffff81518af2>] dump_stack+0x4c/0x65
[ 113.490389] [<ffffffff81086962>] ? console_unlock+0x3d6/0x405
[ 113.491039] [<ffffffff8107ae75>] lockdep_rcu_suspicious+0xfa/0x103
[ 113.491735] [<ffffffff81064032>] rcu_preempt_sleep_check+0x45/0x47
[ 113.492442] [<ffffffff8106404d>] ___might_sleep+0x19/0x1c8
[ 113.493077] [<ffffffff81064268>] __might_sleep+0x6c/0x82
[ 113.493681] [<ffffffff81133190>] cache_alloc_debugcheck_before.isra.50+0x1d/0x24
[ 113.494508] [<ffffffff81134876>] kmem_cache_alloc+0x31/0x18f
[ 113.495149] [<ffffffff814012b5>] skb_clone+0x64/0x80
[ 113.495712] [<ffffffff814e6f71>] pfkey_broadcast_one+0x3d/0xff
[ 113.496380] [<ffffffff814e7b84>] pfkey_broadcast+0xb5/0x11e
[ 113.497024] [<ffffffff814e82d1>] pfkey_register+0x191/0x1b1
[ 113.497653] [<ffffffff814e9770>] pfkey_process+0x162/0x17e
[ 113.498274] [<ffffffff814e9895>] pfkey_sendmsg+0x109/0x213

In pfkey_sendmsg the net mutex is taken and then pfkey_broadcast takes
the RCU lock.

Since pfkey_broadcast takes the RCU lock the allocation argument is
pointless since GFP_ATOMIC must be used between the rcu_read_{,un}lock.
The one call outside of rcu can be done with GFP_KERNEL.

Fixes: 7f6b9dbd5afbd ("af_key: locking change")
Signed-off-by: David Ahern <dsa@xxxxxxxxxxxxxxxxxxx>
Acked-by: Eric Dumazet <edumazet@xxxxxxxxxx>
Signed-off-by: David S. Miller <davem@xxxxxxxxxxxxx>
[bwh: Backported to 3.2: adjust context]
Signed-off-by: Ben Hutchings <ben@xxxxxxxxxxxxxxx>
---
net/key/af_key.c | 46 +++++++++++++++++++++++-----------------------
1 file changed, 23 insertions(+), 23 deletions(-)

--- a/net/key/af_key.c
+++ b/net/key/af_key.c
@@ -220,7 +220,7 @@ static int pfkey_broadcast_one(struct sk
#define BROADCAST_ONE 1
#define BROADCAST_REGISTERED 2
#define BROADCAST_PROMISC_ONLY 4
-static int pfkey_broadcast(struct sk_buff *skb, gfp_t allocation,
+static int pfkey_broadcast(struct sk_buff *skb,
int broadcast_flags, struct sock *one_sk,
struct net *net)
{
@@ -246,7 +246,7 @@ static int pfkey_broadcast(struct sk_buf
* socket.
*/
if (pfk->promisc)
- pfkey_broadcast_one(skb, &skb2, allocation, sk);
+ pfkey_broadcast_one(skb, &skb2, GFP_ATOMIC, sk);

/* the exact target will be processed later */
if (sk == one_sk)
@@ -261,7 +261,7 @@ static int pfkey_broadcast(struct sk_buf
continue;
}

- err2 = pfkey_broadcast_one(skb, &skb2, allocation, sk);
+ err2 = pfkey_broadcast_one(skb, &skb2, GFP_ATOMIC, sk);

/* Error is cleare after succecful sending to at least one
* registered KM */
@@ -271,7 +271,7 @@ static int pfkey_broadcast(struct sk_buf
rcu_read_unlock();

if (one_sk != NULL)
- err = pfkey_broadcast_one(skb, &skb2, allocation, one_sk);
+ err = pfkey_broadcast_one(skb, &skb2, GFP_KERNEL, one_sk);

kfree_skb(skb2);
kfree_skb(skb);
@@ -294,7 +294,7 @@ static int pfkey_do_dump(struct pfkey_so
hdr = (struct sadb_msg *) pfk->dump.skb->data;
hdr->sadb_msg_seq = 0;
hdr->sadb_msg_errno = rc;
- pfkey_broadcast(pfk->dump.skb, GFP_ATOMIC, BROADCAST_ONE,
+ pfkey_broadcast(pfk->dump.skb, BROADCAST_ONE,
&pfk->sk, sock_net(&pfk->sk));
pfk->dump.skb = NULL;
}
@@ -335,7 +335,7 @@ static int pfkey_error(const struct sadb
hdr->sadb_msg_len = (sizeof(struct sadb_msg) /
sizeof(uint64_t));

- pfkey_broadcast(skb, GFP_KERNEL, BROADCAST_ONE, sk, sock_net(sk));
+ pfkey_broadcast(skb, BROADCAST_ONE, sk, sock_net(sk));

return 0;
}
@@ -1361,7 +1361,7 @@ static int pfkey_getspi(struct sock *sk,

xfrm_state_put(x);

- pfkey_broadcast(resp_skb, GFP_KERNEL, BROADCAST_ONE, sk, net);
+ pfkey_broadcast(resp_skb, BROADCAST_ONE, sk, net);

return 0;
}
@@ -1449,7 +1449,7 @@ static int key_notify_sa(struct xfrm_sta
hdr->sadb_msg_seq = c->seq;
hdr->sadb_msg_pid = c->pid;

- pfkey_broadcast(skb, GFP_ATOMIC, BROADCAST_ALL, NULL, xs_net(x));
+ pfkey_broadcast(skb, BROADCAST_ALL, NULL, xs_net(x));

return 0;
}
@@ -1566,7 +1566,7 @@ static int pfkey_get(struct sock *sk, st
out_hdr->sadb_msg_reserved = 0;
out_hdr->sadb_msg_seq = hdr->sadb_msg_seq;
out_hdr->sadb_msg_pid = hdr->sadb_msg_pid;
- pfkey_broadcast(out_skb, GFP_ATOMIC, BROADCAST_ONE, sk, sock_net(sk));
+ pfkey_broadcast(out_skb, BROADCAST_ONE, sk, sock_net(sk));

return 0;
}
@@ -1667,7 +1667,7 @@ static int pfkey_register(struct sock *s
return -ENOBUFS;
}

- pfkey_broadcast(supp_skb, GFP_KERNEL, BROADCAST_REGISTERED, sk, sock_net(sk));
+ pfkey_broadcast(supp_skb, BROADCAST_REGISTERED, sk, sock_net(sk));

return 0;
}
@@ -1686,7 +1686,7 @@ static int unicast_flush_resp(struct soc
hdr->sadb_msg_errno = (uint8_t) 0;
hdr->sadb_msg_len = (sizeof(struct sadb_msg) / sizeof(uint64_t));

- return pfkey_broadcast(skb, GFP_ATOMIC, BROADCAST_ONE, sk, sock_net(sk));
+ return pfkey_broadcast(skb, BROADCAST_ONE, sk, sock_net(sk));
}

static int key_notify_sa_flush(const struct km_event *c)
@@ -1707,7 +1707,7 @@ static int key_notify_sa_flush(const str
hdr->sadb_msg_len = (sizeof(struct sadb_msg) / sizeof(uint64_t));
hdr->sadb_msg_reserved = 0;

- pfkey_broadcast(skb, GFP_ATOMIC, BROADCAST_ALL, NULL, c->net);
+ pfkey_broadcast(skb, BROADCAST_ALL, NULL, c->net);

return 0;
}
@@ -1768,7 +1768,7 @@ static int dump_sa(struct xfrm_state *x,
out_hdr->sadb_msg_pid = pfk->dump.msg_pid;

if (pfk->dump.skb)
- pfkey_broadcast(pfk->dump.skb, GFP_ATOMIC, BROADCAST_ONE,
+ pfkey_broadcast(pfk->dump.skb, BROADCAST_ONE,
&pfk->sk, sock_net(&pfk->sk));
pfk->dump.skb = out_skb;

@@ -1829,7 +1829,7 @@ static int pfkey_promisc(struct sock *sk
new_hdr->sadb_msg_errno = 0;
}

- pfkey_broadcast(skb, GFP_KERNEL, BROADCAST_ALL, NULL, sock_net(sk));
+ pfkey_broadcast(skb, BROADCAST_ALL, NULL, sock_net(sk));
return 0;
}

@@ -2160,7 +2160,7 @@ static int key_notify_policy(struct xfrm
out_hdr->sadb_msg_errno = 0;
out_hdr->sadb_msg_seq = c->seq;
out_hdr->sadb_msg_pid = c->pid;
- pfkey_broadcast(out_skb, GFP_ATOMIC, BROADCAST_ALL, NULL, xp_net(xp));
+ pfkey_broadcast(out_skb, BROADCAST_ALL, NULL, xp_net(xp));
return 0;

}
@@ -2386,7 +2386,7 @@ static int key_pol_get_resp(struct sock
out_hdr->sadb_msg_errno = 0;
out_hdr->sadb_msg_seq = hdr->sadb_msg_seq;
out_hdr->sadb_msg_pid = hdr->sadb_msg_pid;
- pfkey_broadcast(out_skb, GFP_ATOMIC, BROADCAST_ONE, sk, xp_net(xp));
+ pfkey_broadcast(out_skb, BROADCAST_ONE, sk, xp_net(xp));
err = 0;

out:
@@ -2639,7 +2639,7 @@ static int dump_sp(struct xfrm_policy *x
out_hdr->sadb_msg_pid = pfk->dump.msg_pid;

if (pfk->dump.skb)
- pfkey_broadcast(pfk->dump.skb, GFP_ATOMIC, BROADCAST_ONE,
+ pfkey_broadcast(pfk->dump.skb, BROADCAST_ONE,
&pfk->sk, sock_net(&pfk->sk));
pfk->dump.skb = out_skb;

@@ -2690,7 +2690,7 @@ static int key_notify_policy_flush(const
hdr->sadb_msg_satype = SADB_SATYPE_UNSPEC;
hdr->sadb_msg_len = (sizeof(struct sadb_msg) / sizeof(uint64_t));
hdr->sadb_msg_reserved = 0;
- pfkey_broadcast(skb_out, GFP_ATOMIC, BROADCAST_ALL, NULL, c->net);
+ pfkey_broadcast(skb_out, BROADCAST_ALL, NULL, c->net);
return 0;

}
@@ -2756,7 +2756,7 @@ static int pfkey_process(struct sock *sk
void *ext_hdrs[SADB_EXT_MAX];
int err;

- pfkey_broadcast(skb_clone(skb, GFP_KERNEL), GFP_KERNEL,
+ pfkey_broadcast(skb_clone(skb, GFP_KERNEL),
BROADCAST_PROMISC_ONLY, NULL, sock_net(sk));

memset(ext_hdrs, 0, sizeof(ext_hdrs));
@@ -2962,7 +2962,7 @@ static int key_notify_sa_expire(struct x
out_hdr->sadb_msg_seq = 0;
out_hdr->sadb_msg_pid = 0;

- pfkey_broadcast(out_skb, GFP_ATOMIC, BROADCAST_REGISTERED, NULL, xs_net(x));
+ pfkey_broadcast(out_skb, BROADCAST_REGISTERED, NULL, xs_net(x));
return 0;
}

@@ -3134,7 +3134,7 @@ static int pfkey_send_acquire(struct xfr
xfrm_ctx->ctx_len);
}

- return pfkey_broadcast(skb, GFP_ATOMIC, BROADCAST_REGISTERED, NULL, xs_net(x));
+ return pfkey_broadcast(skb, BROADCAST_REGISTERED, NULL, xs_net(x));
}

static struct xfrm_policy *pfkey_compile_policy(struct sock *sk, int opt,
@@ -3332,7 +3332,7 @@ static int pfkey_send_new_mapping(struct
n_port->sadb_x_nat_t_port_port = sport;
n_port->sadb_x_nat_t_port_reserved = 0;

- return pfkey_broadcast(skb, GFP_ATOMIC, BROADCAST_REGISTERED, NULL, xs_net(x));
+ return pfkey_broadcast(skb, BROADCAST_REGISTERED, NULL, xs_net(x));
}

#ifdef CONFIG_NET_KEY_MIGRATE
@@ -3524,7 +3524,7 @@ static int pfkey_send_migrate(const stru
}

/* broadcast migrate message to sockets */
- pfkey_broadcast(skb, GFP_ATOMIC, BROADCAST_ALL, NULL, &init_net);
+ pfkey_broadcast(skb, BROADCAST_ALL, NULL, &init_net);

return 0;


--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/