From: Zhang Changzhong zhangchangzhong@huawei.com

hulk inclusion category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I4S8PN CVE: NA

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

The sock_hold() in tcp_internal_pacing() is expected to pair with sock_put() in tcp_pace_kick(). But in some path tcp_internal_pacing() is called without checking if pacing timer is already armed, causing sock_hold() to be called one more time and tcp sock can't be released.

As Neal pointed out, this could happen from some of the retransmission code paths that don't use tcp_xmit_retransmit_queue(), such as tcp_retransmit_timer() and tcp_send_loss_probe().

The fix is provided by Eric, it extends the timer to cover all these points that Neal mentioned.

Following is the reproduce procedure provided by Jason: 0) cat /proc/slabinfo | grep TCP 1) switch net.ipv4.tcp_congestion_control to bbr 2) using wrk tool something like that to send packages 3) using tc to increase the delay and loss to simulate the RTO case. 4) cat /proc/slabinfo | grep TCP 5) kill the wrk command and observe the number of objects and slabs in TCP. 6) at last, you could notice that the number would not decrease.

Link: https://lore.kernel.org/all/CANn89i+7-wE4xr5D9DpH+N-xkL1SB8oVghCKgz+CT5eG1OD... Signed-off-by: Zhang Changzhong zhangchangzhong@huawei.com Reviewed-by: Wei Yongjun weiyongjun1@huawei.com Signed-off-by: Yang Yingliang yangyingliang@huawei.com --- net/ipv4/tcp_output.c | 31 +++++++++++++++++++++++-------- 1 file changed, 23 insertions(+), 8 deletions(-)

diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c index 6710056fd1b23..44250df02ee18 100644 --- a/net/ipv4/tcp_output.c +++ b/net/ipv4/tcp_output.c @@ -1021,6 +1021,8 @@ enum hrtimer_restart tcp_pace_kick(struct hrtimer *timer)

static void tcp_internal_pacing(struct sock *sk, const struct sk_buff *skb) { + struct tcp_sock *tp = tcp_sk(sk); + ktime_t expire, now; u64 len_ns; u32 rate;

@@ -1032,12 +1034,28 @@ static void tcp_internal_pacing(struct sock *sk, const struct sk_buff *skb)

len_ns = (u64)skb->len * NSEC_PER_SEC; do_div(len_ns, rate); - hrtimer_start(&tcp_sk(sk)->pacing_timer, - ktime_add_ns(ktime_get(), len_ns), + now = ktime_get(); + /* If hrtimer is already armed, then our caller has not + * used tcp_pacing_check(). + */ + if (unlikely(hrtimer_is_queued(&tp->pacing_timer))) { + expire = hrtimer_get_softexpires(&tp->pacing_timer); + if (ktime_after(expire, now)) + now = expire; + if (hrtimer_try_to_cancel(&tp->pacing_timer) == 1) + __sock_put(sk); + } + hrtimer_start(&tp->pacing_timer, ktime_add_ns(now, len_ns), HRTIMER_MODE_ABS_PINNED_SOFT); sock_hold(sk); }

+static bool tcp_pacing_check(const struct sock *sk) +{ + return tcp_needs_internal_pacing(sk) && + hrtimer_is_queued(&tcp_sk(sk)->pacing_timer); +} + static void tcp_update_skb_after_send(struct tcp_sock *tp, struct sk_buff *skb) { skb->skb_mstamp = tp->tcp_mstamp; @@ -2174,6 +2192,9 @@ static int tcp_mtu_probe(struct sock *sk) if (!tcp_can_coalesce_send_queue_head(sk, probe_size)) return -1;

+ if (tcp_pacing_check(sk)) + return -1; + /* We're allowed to probe. Build it now. */ nskb = sk_stream_alloc_skb(sk, probe_size, GFP_ATOMIC, false); if (!nskb) @@ -2247,12 +2268,6 @@ static int tcp_mtu_probe(struct sock *sk) return -1; }

-static bool tcp_pacing_check(const struct sock *sk) -{ - return tcp_needs_internal_pacing(sk) && - hrtimer_is_queued(&tcp_sk(sk)->pacing_timer); -} - /* TCP Small Queues : * Control number of packets in qdisc/devices to two packets / or ~1 ms. * (These limits are doubled for retransmits)