From e258a2c3faec6aa2c29144e9cfef9e4edf4253da Mon Sep 17 00:00:00 2001 From: Lars Kellogg-Stedman Date: Wed, 22 May 2024 14:31:34 -0400 Subject: [PATCH] ax25: Fix refcount imbalance on inbound connections When releasing a socket in ax25_release(), we call netdev_put() to decrease the refcount on the associated ax.25 device. However, the execution path for accepting an incoming connection never calls netdev_hold(). This imbalance leads to refcount errors, and ultimately to kernel crashes. A typical call trace for the above situation looks like this: Call Trace: ? show_regs+0x64/0x70 ? __warn+0x83/0x120 ? refcount_warn_saturate+0xb2/0x100 ? report_bug+0x158/0x190 ? prb_read_valid+0x20/0x30 ? handle_bug+0x3e/0x70 ? exc_invalid_op+0x1c/0x70 ? asm_exc_invalid_op+0x1f/0x30 ? refcount_warn_saturate+0xb2/0x100 ? refcount_warn_saturate+0xb2/0x100 ax25_release+0x2ad/0x360 __sock_release+0x35/0xa0 sock_close+0x19/0x20 [...] On reboot (or any attempt to remove the interface), the kernel gets stuck in an infinite loop: unregister_netdevice: waiting for ax0 to become free. Usage count = 0 This patch corrects these issues by ensuring that we call netdev_hold() and ax25_dev_hold() for new connections in ax25_accept(). Fixes: 7d8a3a477b ("ax25: Fix ax25 session cleanup problems") Signed-off-by: Lars Kellogg-Stedman Signed-off-by: NipaLocal --- net/ax25/af_ax25.c | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/net/ax25/af_ax25.c b/net/ax25/af_ax25.c index 8077cf2ee44803..d6f9fae06a9d81 100644 --- a/net/ax25/af_ax25.c +++ b/net/ax25/af_ax25.c @@ -1378,8 +1378,10 @@ static int ax25_accept(struct socket *sock, struct socket *newsock, { struct sk_buff *skb; struct sock *newsk; + ax25_dev *ax25_dev; DEFINE_WAIT(wait); struct sock *sk; + ax25_cb *ax25; int err = 0; if (sock->state != SS_UNCONNECTED) @@ -1434,6 +1436,10 @@ static int ax25_accept(struct socket *sock, struct socket *newsock, kfree_skb(skb); sk_acceptq_removed(sk); newsock->state = SS_CONNECTED; + ax25 = sk_to_ax25(newsk); + ax25_dev = ax25->ax25_dev; + netdev_hold(ax25_dev->dev, &ax25->dev_tracker, GFP_ATOMIC); + ax25_dev_hold(ax25_dev); out: release_sock(sk);