Skip to contentSkip to conversation

FAILED: patch "[PATCH] mptcp: pm: Fix uaf in __timer_delete_sync" failed to apply to 5.10-stable tree

PATCH Applied [email protected] 2 participants 9 messages Sep 13, 2024 – Sep 27, 2024

Subsystems (3)
lore
Share & tools
lore

[email protected]

author maintainer
#0 lore raw

The patch below does not apply to the 5.10-stable tree.
If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <[email protected]>.

To reproduce the conflict and resubmit, you may use the following commands:

git fetch https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/ linux-5.10.y
git checkout FETCH_HEAD
git cherry-pick -x b4cd80b0338945a94972ac3ed54f8338d2da2076
# <resolve conflicts, build, test, etc.>
git commit -s
git send-email --to '<[email protected]>' --in-reply-to '2024091330-reps-craftsman-ab67@gregkh' --subject-prefix 'PATCH 5.10.y' HEAD^..

Possible dependencies:

b4cd80b03389 ("mptcp: pm: Fix uaf in __timer_delete_sync") d58300c3185b ("mptcp: validate 'id' when stopping the ADD_ADDR retransmit timer") f7dafee18538 ("mptcp: use mptcp_addr_info in mptcp_options_received") dc87efdb1a5c ("mptcp: add mptcp reset option support")
557963c383e8 ("mptcp: move to next addr when subflow creation fail") d88c476f4a7d ("mptcp: export lookup_anno_list_by_saddr") efd13b71a3fa ("Merge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net")

thanks,

greg k-h

------------------ original commit in Linus's tree ------------------

From b4cd80b0338945a94972ac3ed54f8338d2da2076 Mon Sep 17 00:00:00 2001
From: Edward Adam Davis <[email protected]>
Date: Tue, 10 Sep 2024 17:58:56 +0800
Subject: [PATCH] mptcp: pm: Fix uaf in __timer_delete_sync

There are two paths to access mptcp_pm_del_add_timer, result in a race condition:

     CPU1				CPU2
     ====                               ====
     net_rx_action
     napi_poll                          netlink_sendmsg
     __napi_poll                        netlink_unicast
     process_backlog                    netlink_unicast_kernel
     __netif_receive_skb                genl_rcv
     __netif_receive_skb_one_core       netlink_rcv_skb
     NF_HOOK                            genl_rcv_msg
     ip_local_deliver_finish            genl_family_rcv_msg
     ip_protocol_deliver_rcu            genl_family_rcv_msg_doit
     tcp_v4_rcv                         mptcp_pm_nl_flush_addrs_doit
     tcp_v4_do_rcv                      mptcp_nl_remove_addrs_list
     tcp_rcv_established                mptcp_pm_remove_addrs_and_subflows
     tcp_data_queue                     remove_anno_list_by_saddr
     mptcp_incoming_options             mptcp_pm_del_add_timer
     mptcp_pm_del_add_timer             kfree(entry)

In remove_anno_list_by_saddr(running on CPU2), after leaving the critical zone protected by "pm.lock", the entry will be released, which leads to the occurrence of uaf in the mptcp_pm_del_add_timer(running on CPU1).

Keeping a reference to add_timer inside the lock, and calling sk_stop_timer_sync() with this reference, instead of "entry->add_timer".

Move list_del(&entry->list) to mptcp_pm_del_add_timer and inside the pm lock, do not directly access any members of the entry outside the pm lock, which can avoid similar "entry->x" uaf.

net/mptcp/pm_netlink.c +9 -4
index f891bc714668..ad935d34c973 100644
@@ -334,15 +334,21 @@ mptcp_pm_del_add_timer(struct mptcp_sock *msk,
 {
 	struct mptcp_pm_add_entry *entry;
 	struct sock *sk = (struct sock *)msk;
+	struct timer_list *add_timer = NULL;
 
 	spin_lock_bh(&msk->pm.lock);
 	entry = mptcp_lookup_anno_list_by_saddr(msk, addr);
-	if (entry && (!check_id || entry->addr.id == addr->id))
+	if (entry && (!check_id || entry->addr.id == addr->id)) {
 		entry->retrans_times = ADD_ADDR_RETRANS_MAX;
+		add_timer = &entry->add_timer;
+	}
+	if (!check_id && entry)
+		list_del(&entry->list);
 	spin_unlock_bh(&msk->pm.lock);
 
-	if (entry && (!check_id || entry->addr.id == addr->id))
-		sk_stop_timer_sync(sk, &entry->add_timer);
+	/* no lock, because sk_stop_timer_sync() is calling del_timer_sync() */
+	if (add_timer)
+		sk_stop_timer_sync(sk, add_timer);
 
 	return entry;
 }
@@ -1462,7 +1468,6 @@ static bool remove_anno_list_by_saddr(struct mptcp_sock *msk,
 
 	entry = mptcp_pm_del_add_timer(msk, addr, false);
 	if (entry) {
-		list_del(&entry->list);
 		kfree(entry);
 		return true;
 	}
#1 lore raw

cover Backport of "mptcp: pm: Fix uaf in __timer_delete_sync"

To ease the inclusion of commit b4cd80b03389 ("mptcp: pm: Fix uaf in __timer_delete_sync"), commit d88c476f4a7d ("mptcp: export lookup_anno_list_by_saddr") and commit d58300c3185b ("mptcp: validate 'id' when stopping the ADD_ADDR retransmit timer") look safe to be backported as well, even if there was one small conflict, but easy to resolve. The first one is a refactoring, while the second one is a fix that is helping to prevent the new issues being fixed here.

Davide Caratti (1):
mptcp: validate 'id' when stopping the ADD_ADDR retransmit timer

Edward Adam Davis (1):
mptcp: pm: Fix uaf in __timer_delete_sync

Geliang Tang (1):
mptcp: export lookup_anno_list_by_saddr

 net/mptcp/options.c    |  2 +-
 net/mptcp/pm_netlink.c | 27 ++++++++++++++++-----------
 net/mptcp/protocol.h   |  5 ++++-
 3 files changed, 21 insertions(+), 13 deletions(-)
Signature · 1 line
2.45.2
author maintainer
#2 lore raw

Re: [PATCH 5.10.y 0/3] Backport of "mptcp: pm: Fix uaf in __timer_delete_sync"

Quoted 14 lines · On Tue, Sep 17, 2024 at 09:26:08AM +0200, Matthieu Baerts (NGI0) wrote:

To ease the inclusion of commit b4cd80b03389 ("mptcp: pm: Fix uaf in __timer_delete_sync"), commit d88c476f4a7d ("mptcp: export lookup_anno_list_by_saddr") and commit d58300c3185b ("mptcp: validate 'id' when stopping the ADD_ADDR retransmit timer") look safe to be backported as well, even if there was one small conflict, but easy to resolve. The first one is a refactoring, while the second one is a fix that is helping to prevent the new issues being fixed here.

Davide Caratti (1):
mptcp: validate 'id' when stopping the ADD_ADDR retransmit timer

Edward Adam Davis (1):
mptcp: pm: Fix uaf in __timer_delete_sync

Quoted 12 lines

Geliang Tang (1):
mptcp: export lookup_anno_list_by_saddr

 net/mptcp/options.c    |  2 +-
 net/mptcp/pm_netlink.c | 27 ++++++++++++++++-----------
 net/mptcp/protocol.h   |  5 ++++-
 3 files changed, 21 insertions(+), 13 deletions(-)
Signature · 1 line
2.45.2

ALl now queued up, thanks!

greg k-h

#3 lore raw

PATCH 1/3 mptcp: export lookup_anno_list_by_saddr

In reply to [email protected]

From: Geliang Tang <[email protected]>

commit d88c476f4a7dd69a2588470f6c4f8b663efa16c6 upstream.

This patch exported the static function lookup_anno_list_by_saddr, and renamed it to mptcp_lookup_anno_list_by_saddr.

Stable-dep-of: b4cd80b03389 ("mptcp: pm: Fix uaf in __timer_delete_sync")

---
 net/mptcp/pm_netlink.c | 10 +++++-----
 net/mptcp/protocol.h   |  3 +++
 2 files changed, 8 insertions(+), 5 deletions(-)
net/mptcp/pm_netlink.c +5 -5
index f115c92c45d4..0b566678cc96 100644
@@ -194,9 +194,9 @@ static void check_work_pending(struct mptcp_sock *msk)
 		WRITE_ONCE(msk->pm.work_pending, false);
 }
 
-static struct mptcp_pm_add_entry *
-lookup_anno_list_by_saddr(struct mptcp_sock *msk,
-			  struct mptcp_addr_info *addr)
+struct mptcp_pm_add_entry *
+mptcp_lookup_anno_list_by_saddr(struct mptcp_sock *msk,
+				struct mptcp_addr_info *addr)
 {
 	struct mptcp_pm_add_entry *entry;
 
@@ -255,7 +255,7 @@ mptcp_pm_del_add_timer(struct mptcp_sock *msk,
 	struct sock *sk = (struct sock *)msk;
 
 	spin_lock_bh(&msk->pm.lock);
-	entry = lookup_anno_list_by_saddr(msk, addr);
+	entry = mptcp_lookup_anno_list_by_saddr(msk, addr);
 	if (entry)
 		entry->retrans_times = ADD_ADDR_RETRANS_MAX;
 	spin_unlock_bh(&msk->pm.lock);
@@ -272,7 +272,7 @@ static bool mptcp_pm_alloc_anno_list(struct mptcp_sock *msk,
 	struct mptcp_pm_add_entry *add_entry = NULL;
 	struct sock *sk = (struct sock *)msk;
 
-	if (lookup_anno_list_by_saddr(msk, &entry->addr))
+	if (mptcp_lookup_anno_list_by_saddr(msk, &entry->addr))
 		return false;
 
 	add_entry = kmalloc(sizeof(*add_entry), GFP_ATOMIC);
net/mptcp/protocol.h +3 -0
index b8351b671c2f..eaaff2cee4d5 100644
@@ -451,6 +451,9 @@ void mptcp_pm_free_anno_list(struct mptcp_sock *msk);
 struct mptcp_pm_add_entry *
 mptcp_pm_del_add_timer(struct mptcp_sock *msk,
 		       struct mptcp_addr_info *addr);
+struct mptcp_pm_add_entry *
+mptcp_lookup_anno_list_by_saddr(struct mptcp_sock *msk,
+				struct mptcp_addr_info *addr);
 
 int mptcp_pm_announce_addr(struct mptcp_sock *msk,
 			   const struct mptcp_addr_info *addr,
Signature · 1 line
2.45.2
author maintainer Applied
#4 lore raw

Patch "mptcp: export lookup_anno_list_by_saddr" has been added to the 5.10-stable tree

This is a note to let you know that I've just added the patch titled

mptcp: export lookup_anno_list_by_saddr

to the 5.10-stable tree which can be found at:
http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=summary

The filename of the patch is:
mptcp-export-lookup_anno_list_by_saddr.patch
and it can be found in the queue-5.10 subdirectory.

If you, or anyone else, feels it should not be added to the stable tree, please let <[email protected]> know about it.

From [email protected] Tue Sep 17 09:26:33 2024
From: "Matthieu Baerts (NGI0)" <[email protected]>
Date: Tue, 17 Sep 2024 09:26:09 +0200
Subject: mptcp: export lookup_anno_list_by_saddr
To: [email protected], [email protected], [email protected]
Cc: Geliang Tang <[email protected]>, Mat Martineau <[email protected]>, "David S . Miller" <[email protected]>, Matthieu Baerts <[email protected]>
Message-ID: <[email protected]>

From: Geliang Tang <[email protected]>

commit d88c476f4a7dd69a2588470f6c4f8b663efa16c6 upstream.

This patch exported the static function lookup_anno_list_by_saddr, and renamed it to mptcp_lookup_anno_list_by_saddr.

Stable-dep-of: b4cd80b03389 ("mptcp: pm: Fix uaf in __timer_delete_sync")

---
 net/mptcp/pm_netlink.c |   10 +++++-----
 net/mptcp/protocol.h   |    3 +++
 2 files changed, 8 insertions(+), 5 deletions(-)
net/mptcp/pm_netlink.c +5 -5
@@ -194,9 +194,9 @@ static void check_work_pending(struct mp
 		WRITE_ONCE(msk->pm.work_pending, false);
 }
 
-static struct mptcp_pm_add_entry *
-lookup_anno_list_by_saddr(struct mptcp_sock *msk,
-			  struct mptcp_addr_info *addr)
+struct mptcp_pm_add_entry *
+mptcp_lookup_anno_list_by_saddr(struct mptcp_sock *msk,
+				struct mptcp_addr_info *addr)
 {
 	struct mptcp_pm_add_entry *entry;
 
@@ -255,7 +255,7 @@ mptcp_pm_del_add_timer(struct mptcp_sock
 	struct sock *sk = (struct sock *)msk;
 
 	spin_lock_bh(&msk->pm.lock);
-	entry = lookup_anno_list_by_saddr(msk, addr);
+	entry = mptcp_lookup_anno_list_by_saddr(msk, addr);
 	if (entry)
 		entry->retrans_times = ADD_ADDR_RETRANS_MAX;
 	spin_unlock_bh(&msk->pm.lock);
@@ -272,7 +272,7 @@ static bool mptcp_pm_alloc_anno_list(str
 	struct mptcp_pm_add_entry *add_entry = NULL;
 	struct sock *sk = (struct sock *)msk;
 
-	if (lookup_anno_list_by_saddr(msk, &entry->addr))
+	if (mptcp_lookup_anno_list_by_saddr(msk, &entry->addr))
 		return false;
 
 	add_entry = kmalloc(sizeof(*add_entry), GFP_ATOMIC);
net/mptcp/protocol.h +3 -0
@@ -451,6 +451,9 @@ void mptcp_pm_free_anno_list(struct mptc
 struct mptcp_pm_add_entry *
 mptcp_pm_del_add_timer(struct mptcp_sock *msk,
 		       struct mptcp_addr_info *addr);
+struct mptcp_pm_add_entry *
+mptcp_lookup_anno_list_by_saddr(struct mptcp_sock *msk,
+				struct mptcp_addr_info *addr);
 
 int mptcp_pm_announce_addr(struct mptcp_sock *msk,
 			   const struct mptcp_addr_info *addr,

Patches currently in stable-queue which might be from [email protected] are

queue-5.10/mptcp-pm-fix-uaf-in-__timer_delete_sync.patch queue-5.10/mptcp-validate-id-when-stopping-the-add_addr-retransmit-timer.patch queue-5.10/mptcp-export-lookup_anno_list_by_saddr.patch

#5 lore raw

PATCH 2/3 mptcp: validate 'id' when stopping the ADD_ADDR retransmit timer

In reply to [email protected]

From: Davide Caratti <[email protected]>

commit d58300c3185b78ab910092488126b97f0abe3ae2 upstream.

when Linux receives an echo-ed ADD_ADDR, it checks the IP address against the list of "announced" addresses. In case of a positive match, the timer that handles retransmissions is stopped regardless of the 'Address Id' in the received packet: this behaviour does not comply with RFC8684 3.4.1.

Fix it by validating the 'Address Id' in received echo-ed ADD_ADDRs. Tested using packetdrill, with the following captured output:

unpatched kernel:

Out <...> Flags [.], ack 1, win 256, options [mptcp add-addr v1 id 1 198.51.100.2 hmac 0xfd2e62517888fe29,mptcp dss ack 3007449509], length 0 In <...> Flags [.], ack 1, win 257, options [mptcp add-addr v1-echo id 1 1.2.3.4,mptcp dss ack 3013740213], length 0 Out <...> Flags [.], ack 1, win 256, options [mptcp add-addr v1 id 1 198.51.100.2 hmac 0xfd2e62517888fe29,mptcp dss ack 3007449509], length 0 In <...> Flags [.], ack 1, win 257, options [mptcp add-addr v1-echo id 90 198.51.100.2,mptcp dss ack 3013740213], length 0
^^^ retransmission is stopped here, but 'Address Id' is 90

patched kernel:

Out <...> Flags [.], ack 1, win 256, options [mptcp add-addr v1 id 1 198.51.100.2 hmac 0x1cf372d59e05f4b8,mptcp dss ack 3007449509], length 0 In <...> Flags [.], ack 1, win 257, options [mptcp add-addr v1-echo id 1 1.2.3.4,mptcp dss ack 1672384568], length 0 Out <...> Flags [.], ack 1, win 256, options [mptcp add-addr v1 id 1 198.51.100.2 hmac 0x1cf372d59e05f4b8,mptcp dss ack 3007449509], length 0 In <...> Flags [.], ack 1, win 257, options [mptcp add-addr v1-echo id 90 198.51.100.2,mptcp dss ack 1672384568], length 0 Out <...> Flags [.], ack 1, win 256, options [mptcp add-addr v1 id 1 198.51.100.2 hmac 0x1cf372d59e05f4b8,mptcp dss ack 3007449509], length 0 In <...> Flags [.], ack 1, win 257, options [mptcp add-addr v1-echo id 1 198.51.100.2,mptcp dss ack 1672384568], length 0
^^^ retransmission is stopped here, only when both 'Address Id' and 'IP Address' match

Fixes 00cfd77b9063 ("mptcp: retransmit ADD_ADDR when timeout")

Stable-dep-of: b4cd80b03389 ("mptcp: pm: Fix uaf in __timer_delete_sync") [ Conflicts in options.c, because some features are missing in this
version, e.g. commit 557963c383e8 ("mptcp: move to next addr when subflow creation fail") and commit f7dafee18538 ("mptcp: use mptcp_addr_info in mptcp_options_received"). ]

---
 net/mptcp/options.c    | 2 +-
 net/mptcp/pm_netlink.c | 8 ++++----
 net/mptcp/protocol.h   | 2 +-
 3 files changed, 6 insertions(+), 6 deletions(-)
net/mptcp/options.c +1 -1
index 9b11396552df..8bc8812f7526 100644
@@ -915,7 +915,7 @@ void mptcp_incoming_options(struct sock *sk, struct sk_buff *skb)
 			mptcp_pm_add_addr_received(msk, &addr);
 			MPTCP_INC_STATS(sock_net(sk), MPTCP_MIB_ADDADDR);
 		} else {
-			mptcp_pm_del_add_timer(msk, &addr);
+			mptcp_pm_del_add_timer(msk, &addr, true);
 			MPTCP_INC_STATS(sock_net(sk), MPTCP_MIB_ECHOADD);
 		}
 		mp_opt.add_addr = 0;
net/mptcp/pm_netlink.c +4 -4
index 0b566678cc96..f4f5cc76870a 100644
@@ -249,18 +249,18 @@ static void mptcp_pm_add_timer(struct timer_list *timer)
 
 struct mptcp_pm_add_entry *
 mptcp_pm_del_add_timer(struct mptcp_sock *msk,
-		       struct mptcp_addr_info *addr)
+		       struct mptcp_addr_info *addr, bool check_id)
 {
 	struct mptcp_pm_add_entry *entry;
 	struct sock *sk = (struct sock *)msk;
 
 	spin_lock_bh(&msk->pm.lock);
 	entry = mptcp_lookup_anno_list_by_saddr(msk, addr);
-	if (entry)
+	if (entry && (!check_id || entry->addr.id == addr->id))
 		entry->retrans_times = ADD_ADDR_RETRANS_MAX;
 	spin_unlock_bh(&msk->pm.lock);
 
-	if (entry)
+	if (entry && (!check_id || entry->addr.id == addr->id))
 		sk_stop_timer_sync(sk, &entry->add_timer);
 
 	return entry;
@@ -764,7 +764,7 @@ static bool remove_anno_list_by_saddr(struct mptcp_sock *msk,
 {
 	struct mptcp_pm_add_entry *entry;
 
-	entry = mptcp_pm_del_add_timer(msk, addr);
+	entry = mptcp_pm_del_add_timer(msk, addr, false);
 	if (entry) {
 		list_del(&entry->list);
 		kfree(entry);
net/mptcp/protocol.h +1 -1
index eaaff2cee4d5..44944e8f73c5 100644
@@ -450,7 +450,7 @@ void mptcp_pm_rm_addr_received(struct mptcp_sock *msk, u8 rm_id);
 void mptcp_pm_free_anno_list(struct mptcp_sock *msk);
 struct mptcp_pm_add_entry *
 mptcp_pm_del_add_timer(struct mptcp_sock *msk,
-		       struct mptcp_addr_info *addr);
+		       struct mptcp_addr_info *addr, bool check_id);
 struct mptcp_pm_add_entry *
 mptcp_lookup_anno_list_by_saddr(struct mptcp_sock *msk,
 				struct mptcp_addr_info *addr);
Signature · 1 line
2.45.2
author maintainer Applied
#6 lore raw

Patch "mptcp: validate 'id' when stopping the ADD_ADDR retransmit timer" has been added to the 5.10-stable tree

This is a note to let you know that I've just added the patch titled

mptcp: validate 'id' when stopping the ADD_ADDR retransmit timer

to the 5.10-stable tree which can be found at:
http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=summary

The filename of the patch is:
mptcp-validate-id-when-stopping-the-add_addr-retransmit-timer.patch
and it can be found in the queue-5.10 subdirectory.

If you, or anyone else, feels it should not be added to the stable tree, please let <[email protected]> know about it.

From [email protected] Tue Sep 17 09:26:36 2024
From: "Matthieu Baerts (NGI0)" <[email protected]>
Date: Tue, 17 Sep 2024 09:26:10 +0200
Subject: mptcp: validate 'id' when stopping the ADD_ADDR retransmit timer
To: [email protected], [email protected], [email protected]
Cc: Davide Caratti <[email protected]>, Mat Martineau <[email protected]>, "David S . Miller" <[email protected]>, Matthieu Baerts <[email protected]>
Message-ID: <[email protected]>

From: Davide Caratti <[email protected]>

commit d58300c3185b78ab910092488126b97f0abe3ae2 upstream.

when Linux receives an echo-ed ADD_ADDR, it checks the IP address against the list of "announced" addresses. In case of a positive match, the timer that handles retransmissions is stopped regardless of the 'Address Id' in the received packet: this behaviour does not comply with RFC8684 3.4.1.

Fix it by validating the 'Address Id' in received echo-ed ADD_ADDRs. Tested using packetdrill, with the following captured output:

unpatched kernel:

Out <...> Flags [.], ack 1, win 256, options [mptcp add-addr v1 id 1 198.51.100.2 hmac 0xfd2e62517888fe29,mptcp dss ack 3007449509], length 0 In <...> Flags [.], ack 1, win 257, options [mptcp add-addr v1-echo id 1 1.2.3.4,mptcp dss ack 3013740213], length 0 Out <...> Flags [.], ack 1, win 256, options [mptcp add-addr v1 id 1 198.51.100.2 hmac 0xfd2e62517888fe29,mptcp dss ack 3007449509], length 0 In <...> Flags [.], ack 1, win 257, options [mptcp add-addr v1-echo id 90 198.51.100.2,mptcp dss ack 3013740213], length 0
^^^ retransmission is stopped here, but 'Address Id' is 90

patched kernel:

Out <...> Flags [.], ack 1, win 256, options [mptcp add-addr v1 id 1 198.51.100.2 hmac 0x1cf372d59e05f4b8,mptcp dss ack 3007449509], length 0 In <...> Flags [.], ack 1, win 257, options [mptcp add-addr v1-echo id 1 1.2.3.4,mptcp dss ack 1672384568], length 0 Out <...> Flags [.], ack 1, win 256, options [mptcp add-addr v1 id 1 198.51.100.2 hmac 0x1cf372d59e05f4b8,mptcp dss ack 3007449509], length 0 In <...> Flags [.], ack 1, win 257, options [mptcp add-addr v1-echo id 90 198.51.100.2,mptcp dss ack 1672384568], length 0 Out <...> Flags [.], ack 1, win 256, options [mptcp add-addr v1 id 1 198.51.100.2 hmac 0x1cf372d59e05f4b8,mptcp dss ack 3007449509], length 0 In <...> Flags [.], ack 1, win 257, options [mptcp add-addr v1-echo id 1 198.51.100.2,mptcp dss ack 1672384568], length 0
^^^ retransmission is stopped here, only when both 'Address Id' and 'IP Address' match

Fixes 00cfd77b9063 ("mptcp: retransmit ADD_ADDR when timeout")

Stable-dep-of: b4cd80b03389 ("mptcp: pm: Fix uaf in __timer_delete_sync") [ Conflicts in options.c, because some features are missing in this
version, e.g. commit 557963c383e8 ("mptcp: move to next addr when subflow creation fail") and commit f7dafee18538 ("mptcp: use mptcp_addr_info in mptcp_options_received"). ]

---
 net/mptcp/options.c    |    2 +-
 net/mptcp/pm_netlink.c |    8 ++++----
 net/mptcp/protocol.h   |    2 +-
 3 files changed, 6 insertions(+), 6 deletions(-)
net/mptcp/options.c +1 -1
@@ -915,7 +915,7 @@ void mptcp_incoming_options(struct sock
 			mptcp_pm_add_addr_received(msk, &addr);
 			MPTCP_INC_STATS(sock_net(sk), MPTCP_MIB_ADDADDR);
 		} else {
-			mptcp_pm_del_add_timer(msk, &addr);
+			mptcp_pm_del_add_timer(msk, &addr, true);
 			MPTCP_INC_STATS(sock_net(sk), MPTCP_MIB_ECHOADD);
 		}
 		mp_opt.add_addr = 0;
net/mptcp/pm_netlink.c +4 -4
@@ -249,18 +249,18 @@ out:
 
 struct mptcp_pm_add_entry *
 mptcp_pm_del_add_timer(struct mptcp_sock *msk,
-		       struct mptcp_addr_info *addr)
+		       struct mptcp_addr_info *addr, bool check_id)
 {
 	struct mptcp_pm_add_entry *entry;
 	struct sock *sk = (struct sock *)msk;
 
 	spin_lock_bh(&msk->pm.lock);
 	entry = mptcp_lookup_anno_list_by_saddr(msk, addr);
-	if (entry)
+	if (entry && (!check_id || entry->addr.id == addr->id))
 		entry->retrans_times = ADD_ADDR_RETRANS_MAX;
 	spin_unlock_bh(&msk->pm.lock);
 
-	if (entry)
+	if (entry && (!check_id || entry->addr.id == addr->id))
 		sk_stop_timer_sync(sk, &entry->add_timer);
 
 	return entry;
@@ -764,7 +764,7 @@ static bool remove_anno_list_by_saddr(st
 {
 	struct mptcp_pm_add_entry *entry;
 
-	entry = mptcp_pm_del_add_timer(msk, addr);
+	entry = mptcp_pm_del_add_timer(msk, addr, false);
 	if (entry) {
 		list_del(&entry->list);
 		kfree(entry);
net/mptcp/protocol.h +1 -1
@@ -450,7 +450,7 @@ void mptcp_pm_rm_addr_received(struct mp
 void mptcp_pm_free_anno_list(struct mptcp_sock *msk);
 struct mptcp_pm_add_entry *
 mptcp_pm_del_add_timer(struct mptcp_sock *msk,
-		       struct mptcp_addr_info *addr);
+		       struct mptcp_addr_info *addr, bool check_id);
 struct mptcp_pm_add_entry *
 mptcp_lookup_anno_list_by_saddr(struct mptcp_sock *msk,
 				struct mptcp_addr_info *addr);

Patches currently in stable-queue which might be from [email protected] are

queue-5.10/mptcp-pm-fix-uaf-in-__timer_delete_sync.patch queue-5.10/mptcp-validate-id-when-stopping-the-add_addr-retransmit-timer.patch queue-5.10/mptcp-export-lookup_anno_list_by_saddr.patch

#7 lore raw

PATCH 3/3 mptcp: pm: Fix uaf in __timer_delete_sync

In reply to [email protected]

From: Edward Adam Davis <[email protected]>

There are two paths to access mptcp_pm_del_add_timer, result in a race condition:

     CPU1				CPU2
     ====                               ====
     net_rx_action
     napi_poll                          netlink_sendmsg
     __napi_poll                        netlink_unicast
     process_backlog                    netlink_unicast_kernel
     __netif_receive_skb                genl_rcv
     __netif_receive_skb_one_core       netlink_rcv_skb
     NF_HOOK                            genl_rcv_msg
     ip_local_deliver_finish            genl_family_rcv_msg
     ip_protocol_deliver_rcu            genl_family_rcv_msg_doit
     tcp_v4_rcv                         mptcp_pm_nl_flush_addrs_doit
     tcp_v4_do_rcv                      mptcp_nl_remove_addrs_list
     tcp_rcv_established                mptcp_pm_remove_addrs_and_subflows
     tcp_data_queue                     remove_anno_list_by_saddr
     mptcp_incoming_options             mptcp_pm_del_add_timer
     mptcp_pm_del_add_timer             kfree(entry)

In remove_anno_list_by_saddr(running on CPU2), after leaving the critical zone protected by "pm.lock", the entry will be released, which leads to the occurrence of uaf in the mptcp_pm_del_add_timer(running on CPU1).

Keeping a reference to add_timer inside the lock, and calling sk_stop_timer_sync() with this reference, instead of "entry->add_timer".

Move list_del(&entry->list) to mptcp_pm_del_add_timer and inside the pm lock, do not directly access any members of the entry outside the pm lock, which can avoid similar "entry->x" uaf.

Fixes 00cfd77b9063 ("mptcp: retransmit ADD_ADDR when timeout")
Reported-and-tested-by [email protected]
(cherry picked from commit b4cd80b0338945a94972ac3ed54f8338d2da2076)
---
 net/mptcp/pm_netlink.c | 13 +++++++++----
 1 file changed, 9 insertions(+), 4 deletions(-)
net/mptcp/pm_netlink.c +9 -4
index f4f5cc76870a..bd03fb6df729 100644
@@ -253,15 +253,21 @@ mptcp_pm_del_add_timer(struct mptcp_sock *msk,
 {
 	struct mptcp_pm_add_entry *entry;
 	struct sock *sk = (struct sock *)msk;
+	struct timer_list *add_timer = NULL;
 
 	spin_lock_bh(&msk->pm.lock);
 	entry = mptcp_lookup_anno_list_by_saddr(msk, addr);
-	if (entry && (!check_id || entry->addr.id == addr->id))
+	if (entry && (!check_id || entry->addr.id == addr->id)) {
 		entry->retrans_times = ADD_ADDR_RETRANS_MAX;
+		add_timer = &entry->add_timer;
+	}
+	if (!check_id && entry)
+		list_del(&entry->list);
 	spin_unlock_bh(&msk->pm.lock);
 
-	if (entry && (!check_id || entry->addr.id == addr->id))
-		sk_stop_timer_sync(sk, &entry->add_timer);
+	/* no lock, because sk_stop_timer_sync() is calling del_timer_sync() */
+	if (add_timer)
+		sk_stop_timer_sync(sk, add_timer);
 
 	return entry;
 }
@@ -766,7 +772,6 @@ static bool remove_anno_list_by_saddr(struct mptcp_sock *msk,
 
 	entry = mptcp_pm_del_add_timer(msk, addr, false);
 	if (entry) {
-		list_del(&entry->list);
 		kfree(entry);
 		return true;
 	}
Signature · 1 line
2.45.2
author maintainer Applied
#8 lore raw

Patch "mptcp: pm: Fix uaf in __timer_delete_sync" has been added to the 5.10-stable tree

This is a note to let you know that I've just added the patch titled

mptcp: pm: Fix uaf in __timer_delete_sync

to the 5.10-stable tree which can be found at:
http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=summary

The filename of the patch is:
mptcp-pm-fix-uaf-in-__timer_delete_sync.patch
and it can be found in the queue-5.10 subdirectory.

If you, or anyone else, feels it should not be added to the stable tree, please let <[email protected]> know about it.

From [email protected] Tue Sep 17 09:26:39 2024
From: "Matthieu Baerts (NGI0)" <[email protected]>
Date: Tue, 17 Sep 2024 09:26:11 +0200
Subject: mptcp: pm: Fix uaf in __timer_delete_sync
To: [email protected], [email protected], [email protected]
Cc: Edward Adam Davis <[email protected]>, [email protected], Matthieu Baerts <[email protected]>, Paolo Abeni <[email protected]>, Jakub Kicinski <[email protected]>
Message-ID: <[email protected]>

From: Edward Adam Davis <[email protected]>

There are two paths to access mptcp_pm_del_add_timer, result in a race condition:

     CPU1				CPU2
     ====                               ====
     net_rx_action
     napi_poll                          netlink_sendmsg
     __napi_poll                        netlink_unicast
     process_backlog                    netlink_unicast_kernel
     __netif_receive_skb                genl_rcv
     __netif_receive_skb_one_core       netlink_rcv_skb
     NF_HOOK                            genl_rcv_msg
     ip_local_deliver_finish            genl_family_rcv_msg
     ip_protocol_deliver_rcu            genl_family_rcv_msg_doit
     tcp_v4_rcv                         mptcp_pm_nl_flush_addrs_doit
     tcp_v4_do_rcv                      mptcp_nl_remove_addrs_list
     tcp_rcv_established                mptcp_pm_remove_addrs_and_subflows
     tcp_data_queue                     remove_anno_list_by_saddr
     mptcp_incoming_options             mptcp_pm_del_add_timer
     mptcp_pm_del_add_timer             kfree(entry)

In remove_anno_list_by_saddr(running on CPU2), after leaving the critical zone protected by "pm.lock", the entry will be released, which leads to the occurrence of uaf in the mptcp_pm_del_add_timer(running on CPU1).

Keeping a reference to add_timer inside the lock, and calling sk_stop_timer_sync() with this reference, instead of "entry->add_timer".

Move list_del(&entry->list) to mptcp_pm_del_add_timer and inside the pm lock, do not directly access any members of the entry outside the pm lock, which can avoid similar "entry->x" uaf.

---
 net/mptcp/pm_netlink.c |   13 +++++++++----
 1 file changed, 9 insertions(+), 4 deletions(-)
net/mptcp/pm_netlink.c +9 -4
@@ -253,15 +253,21 @@ mptcp_pm_del_add_timer(struct mptcp_sock
 {
 	struct mptcp_pm_add_entry *entry;
 	struct sock *sk = (struct sock *)msk;
+	struct timer_list *add_timer = NULL;
 
 	spin_lock_bh(&msk->pm.lock);
 	entry = mptcp_lookup_anno_list_by_saddr(msk, addr);
-	if (entry && (!check_id || entry->addr.id == addr->id))
+	if (entry && (!check_id || entry->addr.id == addr->id)) {
 		entry->retrans_times = ADD_ADDR_RETRANS_MAX;
+		add_timer = &entry->add_timer;
+	}
+	if (!check_id && entry)
+		list_del(&entry->list);
 	spin_unlock_bh(&msk->pm.lock);
 
-	if (entry && (!check_id || entry->addr.id == addr->id))
-		sk_stop_timer_sync(sk, &entry->add_timer);
+	/* no lock, because sk_stop_timer_sync() is calling del_timer_sync() */
+	if (add_timer)
+		sk_stop_timer_sync(sk, add_timer);
 
 	return entry;
 }
@@ -766,7 +772,6 @@ static bool remove_anno_list_by_saddr(st
 
 	entry = mptcp_pm_del_add_timer(msk, addr, false);
 	if (entry) {
-		list_del(&entry->list);
 		kfree(entry);
 		return true;
 	}

Patches currently in stable-queue which might be from [email protected] are

queue-5.10/mptcp-pm-fix-uaf-in-__timer_delete_sync.patch queue-5.10/mptcp-validate-id-when-stopping-the-add_addr-retransmit-timer.patch queue-5.10/mptcp-export-lookup_anno_list_by_saddr.patch