public inbox for stable@vger.kernel.org
 help / color / mirror / Atom feed
From: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
To: linux-kernel@vger.kernel.org
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
	stable@vger.kernel.org, Tom Parkin <tparkin@katalix.com>,
	Eric Dumazet <edumazet@google.com>,
	"David S. Miller" <davem@davemloft.net>
Subject: [ 39/42] net: drop dst before queueing fragments
Date: Mon, 29 Apr 2013 12:02:21 -0700	[thread overview]
Message-ID: <20130429184756.815952144@linuxfoundation.org> (raw)
In-Reply-To: <20130429184752.435249613@linuxfoundation.org>

3.8-stable review patch.  If anyone has any objections, please let me know.

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


From: Eric Dumazet <edumazet@google.com>

[ Upstream commit 97599dc792b45b1669c3cdb9a4b365aad0232f65 ]

Commit 4a94445c9a5c (net: Use ip_route_input_noref() in input path)
added a bug in IP defragmentation handling, as non refcounted
dst could escape an RCU protected section.

Commit 64f3b9e203bd068 (net: ip_expire() must revalidate route) fixed
the case of timeouts, but not the general problem.

Tom Parkin noticed crashes in UDP stack and provided a patch,
but further analysis permitted us to pinpoint the root cause.

Before queueing a packet into a frag list, we must drop its dst,
as this dst has limited lifetime (RCU protected)

When/if a packet is finally reassembled, we use the dst of the very
last skb, still protected by RCU and valid, as the dst of the
reassembled packet.

Use same logic in IPv6, as there is no need to hold dst references.

Reported-by: Tom Parkin <tparkin@katalix.com>
Tested-by: Tom Parkin <tparkin@katalix.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
---
 net/ipv4/ip_fragment.c |   15 +++++++++++----
 net/ipv6/reassembly.c  |   13 +++++++++++--
 2 files changed, 22 insertions(+), 6 deletions(-)

--- a/net/ipv4/ip_fragment.c
+++ b/net/ipv4/ip_fragment.c
@@ -255,8 +255,7 @@ static void ip_expire(unsigned long arg)
 		if (!head->dev)
 			goto out_rcu_unlock;
 
-		/* skb dst is stale, drop it, and perform route lookup again */
-		skb_dst_drop(head);
+		/* skb has no dst, perform route lookup again */
 		iph = ip_hdr(head);
 		err = ip_route_input_noref(head, iph->daddr, iph->saddr,
 					   iph->tos, head->dev);
@@ -525,8 +524,16 @@ found:
 		qp->q.max_size = skb->len + ihl;
 
 	if (qp->q.last_in == (INET_FRAG_FIRST_IN | INET_FRAG_LAST_IN) &&
-	    qp->q.meat == qp->q.len)
-		return ip_frag_reasm(qp, prev, dev);
+	    qp->q.meat == qp->q.len) {
+		unsigned long orefdst = skb->_skb_refdst;
+
+		skb->_skb_refdst = 0UL;
+		err = ip_frag_reasm(qp, prev, dev);
+		skb->_skb_refdst = orefdst;
+		return err;
+	}
+
+	skb_dst_drop(skb);
 
 	write_lock(&ip4_frags.lock);
 	list_move_tail(&qp->q.lru_list, &qp->q.net->lru_list);
--- a/net/ipv6/reassembly.c
+++ b/net/ipv6/reassembly.c
@@ -342,8 +342,17 @@ found:
 	}
 
 	if (fq->q.last_in == (INET_FRAG_FIRST_IN | INET_FRAG_LAST_IN) &&
-	    fq->q.meat == fq->q.len)
-		return ip6_frag_reasm(fq, prev, dev);
+	    fq->q.meat == fq->q.len) {
+		int res;
+		unsigned long orefdst = skb->_skb_refdst;
+
+		skb->_skb_refdst = 0UL;
+		res = ip6_frag_reasm(fq, prev, dev);
+		skb->_skb_refdst = orefdst;
+		return res;
+	}
+
+	skb_dst_drop(skb);
 
 	write_lock(&ip6_frags.lock);
 	list_move_tail(&fq->q.lru_list, &fq->q.net->lru_list);



  parent reply	other threads:[~2013-04-29 19:02 UTC|newest]

Thread overview: 54+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2013-04-29 19:01 [ 00/42] 3.8.11-stable review Greg Kroah-Hartman
2013-04-29 19:01 ` [ 01/42] aio: fix possible invalid memory access when DEBUG is enabled Greg Kroah-Hartman
2013-04-29 19:01 ` [ 02/42] TTY: do not update atime/mtime on read/write Greg Kroah-Hartman
2013-04-30  0:14   ` Simon Kirby
2013-04-30  0:21     ` Greg Kroah-Hartman
2013-04-30  0:36       ` Simon Kirby
2013-04-30  1:37         ` Greg Kroah-Hartman
2013-04-30 23:50           ` Simon Kirby
2013-05-01  0:57           ` Linus Torvalds
2013-05-01  1:41             ` Linus Torvalds
2013-05-01  5:23               ` Jiri Slaby
2013-05-02 16:11               ` Simon Kirby
2013-04-29 19:01 ` [ 03/42] TTY: fix atime/mtime regression Greg Kroah-Hartman
2013-04-29 19:01 ` [ 04/42] sparc64: Fix race in TLB batch processing Greg Kroah-Hartman
2013-04-29 19:01 ` [ 05/42] atm: update msg_namelen in vcc_recvmsg() Greg Kroah-Hartman
2013-04-29 19:01 ` [ 06/42] ax25: fix info leak via msg_name in ax25_recvmsg() Greg Kroah-Hartman
2013-04-29 19:01 ` [ 07/42] Bluetooth: fix possible info leak in bt_sock_recvmsg() Greg Kroah-Hartman
2013-04-29 19:01 ` [ 08/42] Bluetooth: RFCOMM - Fix missing msg_namelen update in rfcomm_sock_recvmsg() Greg Kroah-Hartman
2013-04-29 19:01 ` [ 09/42] Bluetooth: SCO - Fix missing msg_namelen update in sco_sock_recvmsg() Greg Kroah-Hartman
2013-04-29 19:01 ` [ 10/42] caif: Fix missing msg_namelen update in caif_seqpkt_recvmsg() Greg Kroah-Hartman
2013-04-29 19:01 ` [ 11/42] irda: Fix missing msg_namelen update in irda_recvmsg_dgram() Greg Kroah-Hartman
2013-04-29 19:01 ` [ 12/42] iucv: Fix missing msg_namelen update in iucv_sock_recvmsg() Greg Kroah-Hartman
2013-04-29 19:01 ` [ 13/42] l2tp: fix info leak in l2tp_ip6_recvmsg() Greg Kroah-Hartman
2013-04-29 19:01 ` [ 14/42] llc: Fix missing msg_namelen update in llc_ui_recvmsg() Greg Kroah-Hartman
2013-04-29 19:01 ` [ 15/42] netrom: fix info leak via msg_name in nr_recvmsg() Greg Kroah-Hartman
2013-04-29 19:01 ` [ 16/42] NFC: llcp: fix info leaks via msg_name in llcp_sock_recvmsg() Greg Kroah-Hartman
2013-04-29 19:01 ` [ 17/42] rose: fix info leak via msg_name in rose_recvmsg() Greg Kroah-Hartman
2013-04-29 19:02 ` [ 18/42] tipc: fix info leaks via msg_name in recv_msg/recv_stream Greg Kroah-Hartman
2013-04-29 19:02 ` [ 19/42] cbq: incorrect processing of high limits Greg Kroah-Hartman
2013-04-29 19:02 ` [ 20/42] net IPv6 : Fix broken IPv6 routing table after loopback down-up Greg Kroah-Hartman
2013-04-29 19:02 ` [ 21/42] net: count hw_addr syncs so that unsync works properly Greg Kroah-Hartman
2013-04-29 19:02 ` [ 22/42] atl1e: limit gso segment size to prevent generation of wrong ip length fields Greg Kroah-Hartman
2013-04-29 19:02 ` [ 23/42] bonding: fix bonding_masters race condition in bond unloading Greg Kroah-Hartman
2013-04-29 19:02 ` [ 24/42] bonding: IFF_BONDING is not stripped on enslave failure Greg Kroah-Hartman
2013-04-29 19:02 ` [ 25/42] bonding: fix l23 and l34 load balancing in forwarding path Greg Kroah-Hartman
2013-04-29 19:02 ` [ 26/42] af_unix: If we dont care about credentials coallesce all messages Greg Kroah-Hartman
2013-04-29 19:02 ` [ 27/42] netfilter: dont reset nf_trace in nf_reset() Greg Kroah-Hartman
2013-04-29 19:02 ` [ 28/42] ipv6/tcp: Stop processing ICMPv6 redirect messages Greg Kroah-Hartman
2013-04-29 19:02 ` [ 29/42] rtnetlink: Call nlmsg_parse() with correct header length Greg Kroah-Hartman
2013-04-29 19:02 ` [ 30/42] tcp: incoming connections might use wrong route under synflood Greg Kroah-Hartman
2013-04-29 19:02 ` [ 31/42] tcp: Reallocate headroom if it would overflow csum_start Greg Kroah-Hartman
2013-04-29 19:02 ` [ 32/42] net: mvmdio: add select PHYLIB Greg Kroah-Hartman
2013-04-29 19:02 ` [ 33/42] esp4: fix error return code in esp_output() Greg Kroah-Hartman
2013-04-29 19:02 ` [ 34/42] net: mvneta: fix improper tx queue usage in mvneta_tx() Greg Kroah-Hartman
2013-04-29 19:02 ` [ 35/42] net: cdc_mbim: remove bogus sizeof() Greg Kroah-Hartman
2013-04-29 19:02 ` [ 36/42] tcp: call tcp_replace_ts_recent() from tcp_ack() Greg Kroah-Hartman
2013-04-29 19:02 ` [ 37/42] net: rate-limit warn-bad-offload splats Greg Kroah-Hartman
2013-04-29 19:02 ` [ 38/42] net: fix incorrect credentials passing Greg Kroah-Hartman
2013-04-29 19:02 ` Greg Kroah-Hartman [this message]
2013-04-29 19:02 ` [ 40/42] tracing: Fix selftest function recursion accounting Greg Kroah-Hartman
2013-04-29 19:02 ` [ 41/42] ARM: 7699/1: sched_clock: Add more notrace to prevent recursion Greg Kroah-Hartman
2013-04-29 19:02 ` [ 42/42] ARM: 7692/1: iop3xx: move IOP3XX_PERIPHERAL_VIRT_BASE Greg Kroah-Hartman
2013-04-30  1:54 ` [ 00/42] 3.8.11-stable review Shuah Khan
2013-04-30  2:02   ` Greg Kroah-Hartman

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20130429184756.815952144@linuxfoundation.org \
    --to=gregkh@linuxfoundation.org \
    --cc=davem@davemloft.net \
    --cc=edumazet@google.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=stable@vger.kernel.org \
    --cc=tparkin@katalix.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox