From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from smtp.kernel.org (aws-us-west-2-korg-mail-1.web.codeaurora.org [10.30.226.201]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id BF8AE17AA6 for ; Wed, 9 Aug 2023 11:26:16 +0000 (UTC) Received: by smtp.kernel.org (Postfix) with ESMTPSA id 40F13C433C7; Wed, 9 Aug 2023 11:26:16 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=linuxfoundation.org; s=korg; t=1691580376; bh=TK3ZmN3Im5PTHrMOUPItLn64AudpaPQVsNNTMjzB7/E=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=Cq3jP8pij0WK3zbIuWSUjeM5cVciiA4UJSiBUlIl6qHxTecXDrkObZ+2qjEiPYh+l Z0vZDMci6pmdtKWSoKimuOpsUzjgKSCzFIZcxw/vw27ueY1E4KC071dPvphnrF5pQ2 sCKPjqpI966pT/NEuhkvnnNmph8ZWiW0yBe5C37c= From: Greg Kroah-Hartman To: stable@vger.kernel.org Cc: Greg Kroah-Hartman , patches@lists.linux.dev, Eric Dumazet , Yuchung Cheng , Kuniyuki Iwashima , Jakub Kicinski , Sasha Levin Subject: [PATCH 4.19 304/323] tcp_metrics: fix data-race in tcpm_suck_dst() vs fastopen Date: Wed, 9 Aug 2023 12:42:22 +0200 Message-ID: <20230809103711.949463886@linuxfoundation.org> X-Mailer: git-send-email 2.41.0 In-Reply-To: <20230809103658.104386911@linuxfoundation.org> References: <20230809103658.104386911@linuxfoundation.org> User-Agent: quilt/0.67 Precedence: bulk X-Mailing-List: patches@lists.linux.dev List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit From: Eric Dumazet [ Upstream commit ddf251fa2bc1d3699eec0bae6ed0bc373b8fda79 ] Whenever tcpm_new() reclaims an old entry, tcpm_suck_dst() would overwrite data that could be read from tcp_fastopen_cache_get() or tcp_metrics_fill_info(). We need to acquire fastopen_seqlock to maintain consistency. For newly allocated objects, tcpm_new() can switch to kzalloc() to avoid an extra fastopen_seqlock acquisition. Fixes: 1fe4c481ba63 ("net-tcp: Fast Open client - cookie cache") Signed-off-by: Eric Dumazet Cc: Yuchung Cheng Reviewed-by: Kuniyuki Iwashima Link: https://lore.kernel.org/r/20230802131500.1478140-7-edumazet@google.com Signed-off-by: Jakub Kicinski Signed-off-by: Sasha Levin --- net/ipv4/tcp_metrics.c | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-) diff --git a/net/ipv4/tcp_metrics.c b/net/ipv4/tcp_metrics.c index dfd224979cf65..7bbd9125b5000 100644 --- a/net/ipv4/tcp_metrics.c +++ b/net/ipv4/tcp_metrics.c @@ -93,6 +93,7 @@ static struct tcpm_hash_bucket *tcp_metrics_hash __read_mostly; static unsigned int tcp_metrics_hash_log __read_mostly; static DEFINE_SPINLOCK(tcp_metrics_lock); +static DEFINE_SEQLOCK(fastopen_seqlock); static void tcpm_suck_dst(struct tcp_metrics_block *tm, const struct dst_entry *dst, @@ -129,11 +130,13 @@ static void tcpm_suck_dst(struct tcp_metrics_block *tm, tcp_metric_set(tm, TCP_METRIC_REORDERING, dst_metric_raw(dst, RTAX_REORDERING)); if (fastopen_clear) { + write_seqlock(&fastopen_seqlock); tm->tcpm_fastopen.mss = 0; tm->tcpm_fastopen.syn_loss = 0; tm->tcpm_fastopen.try_exp = 0; tm->tcpm_fastopen.cookie.exp = false; tm->tcpm_fastopen.cookie.len = 0; + write_sequnlock(&fastopen_seqlock); } } @@ -194,7 +197,7 @@ static struct tcp_metrics_block *tcpm_new(struct dst_entry *dst, } tm = oldest; } else { - tm = kmalloc(sizeof(*tm), GFP_ATOMIC); + tm = kzalloc(sizeof(*tm), GFP_ATOMIC); if (!tm) goto out_unlock; } @@ -204,7 +207,7 @@ static struct tcp_metrics_block *tcpm_new(struct dst_entry *dst, tm->tcpm_saddr = *saddr; tm->tcpm_daddr = *daddr; - tcpm_suck_dst(tm, dst, true); + tcpm_suck_dst(tm, dst, reclaim); if (likely(!reclaim)) { tm->tcpm_next = tcp_metrics_hash[hash].chain; @@ -561,8 +564,6 @@ bool tcp_peer_is_proven(struct request_sock *req, struct dst_entry *dst) return ret; } -static DEFINE_SEQLOCK(fastopen_seqlock); - void tcp_fastopen_cache_get(struct sock *sk, u16 *mss, struct tcp_fastopen_cookie *cookie) { -- 2.40.1