Skip to contentSkip to conversation

mptcp: A few fixes

PATCH Applied Mat Martineau 2 participants 5 messages Jan 21, 2022

Subsystems (4)
lore
Share & tools
lore
#0 lore raw

cover mptcp: A few fixes

Here are some fixes from the mptcp tree:

Patch 1 fixes a RCU locking issue when processing a netlink command that updates endpoint flags in the in-kernel MPTCP path manager.

Patch 2 fixes a typo affecting available endpoint id tracking.

Patch 3 fixes IPv6 routing in the MPTCP self tests.

Geliang Tang (1):
mptcp: fix removing ids bitmap setting

Paolo Abeni (2):
mptcp: fix msk traversal in mptcp_nl_cmd_set_flags()
selftests: mptcp: fix ipv6 routing setup

 net/mptcp/pm_netlink.c                        | 39 +++++++++++++------
 .../testing/selftests/net/mptcp/mptcp_join.sh |  5 ++-
 2 files changed, 31 insertions(+), 13 deletions(-)

base-commit: 57afdc0aab094b4c811b3fe030b2567812a495f3

Signature · 1 line
2.34.1
#1 lore raw

PATCH 1/3 mptcp: fix msk traversal in mptcp_nl_cmd_set_flags()

From: Paolo Abeni <[email protected]>

The MPTCP endpoint list is under RCU protection, guarded by the pernet spinlock. mptcp_nl_cmd_set_flags() traverses the list without acquiring the spin-lock nor under the RCU critical section.

This change addresses the issue performing the lookup and the endpoint update under the pernet spinlock.

Fixes 0f9f696a502e ("mptcp: add set_flags command in PM netlink")
---
 net/mptcp/pm_netlink.c | 37 +++++++++++++++++++++++++++----------
 1 file changed, 27 insertions(+), 10 deletions(-)
net/mptcp/pm_netlink.c +27 -10
index 75af1f701e1d..f17a09f7fbf9 100644
@@ -478,6 +478,20 @@ __lookup_addr_by_id(struct pm_nl_pernet *pernet, unsigned int id)
 	return NULL;
 }
 
+static struct mptcp_pm_addr_entry *
+__lookup_addr(struct pm_nl_pernet *pernet, const struct mptcp_addr_info *info,
+	      bool lookup_by_id)
+{
+	struct mptcp_pm_addr_entry *entry;
+
+	list_for_each_entry(entry, &pernet->local_addr_list, list) {
+		if ((!lookup_by_id && addresses_equal(&entry->addr, info, true)) ||
+		    (lookup_by_id && entry->addr.id == info->id))
+			return entry;
+	}
+	return NULL;
+}
+
 static int
 lookup_id_by_addr(struct pm_nl_pernet *pernet, const struct mptcp_addr_info *addr)
 {
@@ -1763,18 +1777,21 @@ static int mptcp_nl_cmd_set_flags(struct sk_buff *skb, struct genl_info *info)
 			return -EOPNOTSUPP;
 	}
 
-	list_for_each_entry(entry, &pernet->local_addr_list, list) {
-		if ((!lookup_by_id && addresses_equal(&entry->addr, &addr.addr, true)) ||
-		    (lookup_by_id && entry->addr.id == addr.addr.id)) {
-			mptcp_nl_addr_backup(net, &entry->addr, bkup);
-
-			if (bkup)
-				entry->flags |= MPTCP_PM_ADDR_FLAG_BACKUP;
-			else
-				entry->flags &= ~MPTCP_PM_ADDR_FLAG_BACKUP;
-		}
+	spin_lock_bh(&pernet->lock);
+	entry = __lookup_addr(pernet, &addr.addr, lookup_by_id);
+	if (!entry) {
+		spin_unlock_bh(&pernet->lock);
+		return -EINVAL;
 	}
 
+	if (bkup)
+		entry->flags |= MPTCP_PM_ADDR_FLAG_BACKUP;
+	else
+		entry->flags &= ~MPTCP_PM_ADDR_FLAG_BACKUP;
+	addr = *entry;
+	spin_unlock_bh(&pernet->lock);
+
+	mptcp_nl_addr_backup(net, &addr.addr, bkup);
 	return 0;
 }
 
Signature · 1 line
2.34.1
#2 lore raw

PATCH 2/3 mptcp: fix removing ids bitmap setting

In reply to Mat Martineau

From: Geliang Tang <[email protected]>

In mptcp_pm_nl_rm_addr_or_subflow(), the bit of rm_list->ids[i] in the id_avail_bitmap should be set, not rm_list->ids[1]. This patch fixed it.

Fixes 86e39e04482b ("mptcp: keep track of local endpoint still available for each msk")
---
 net/mptcp/pm_netlink.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
net/mptcp/pm_netlink.c +1 -1
index f17a09f7fbf9..782b1d452269 100644
@@ -791,7 +791,7 @@ static void mptcp_pm_nl_rm_addr_or_subflow(struct mptcp_sock *msk,
 			removed = true;
 			__MPTCP_INC_STATS(sock_net(sk), rm_type);
 		}
-		__set_bit(rm_list->ids[1], msk->pm.id_avail_bitmap);
+		__set_bit(rm_list->ids[i], msk->pm.id_avail_bitmap);
 		if (!removed)
 			continue;
 
Signature · 1 line
2.34.1
#3 lore raw

PATCH 3/3 selftests: mptcp: fix ipv6 routing setup

In reply to Mat Martineau

From: Paolo Abeni <[email protected]>

MPJ ipv6 selftests currently lack per link route to the server net. Additionally, ipv6 subflows endpoints are created without any interface specified. The end-result is that in ipv6 self-tests subflows are created all on the same link, leading to expected delays and sporadic self-tests failures.

Fix the issue by adding the missing setup bits.

Fixes 523514ed0a99 ("selftests: mptcp: add ADD_ADDR IPv6 test cases")
Reported-and-tested-by Geliang Tang <[email protected]>
---
 tools/testing/selftests/net/mptcp/mptcp_join.sh | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)
tools/testing/selftests/net/mptcp/mptcp_join.sh +3 -2
index 27d0eb9afdca..b8bdbec0cf69 100755
@@ -75,6 +75,7 @@ init()
 
 		# let $ns2 reach any $ns1 address from any interface
 		ip -net "$ns2" route add default via 10.0.$i.1 dev ns2eth$i metric 10$i
+		ip -net "$ns2" route add default via dead:beef:$i::1 dev ns2eth$i metric 10$i
 	done
 }
 
@@ -1476,7 +1477,7 @@ ipv6_tests()
 	reset
 	ip netns exec $ns1 ./pm_nl_ctl limits 0 1
 	ip netns exec $ns2 ./pm_nl_ctl limits 0 1
-	ip netns exec $ns2 ./pm_nl_ctl add dead:beef:3::2 flags subflow
+	ip netns exec $ns2 ./pm_nl_ctl add dead:beef:3::2 dev ns2eth3 flags subflow
 	run_tests $ns1 $ns2 dead:beef:1::1 0 0 0 slow
 	chk_join_nr "single subflow IPv6" 1 1 1
 
@@ -1511,7 +1512,7 @@ ipv6_tests()
 	ip netns exec $ns1 ./pm_nl_ctl limits 0 2
 	ip netns exec $ns1 ./pm_nl_ctl add dead:beef:2::1 flags signal
 	ip netns exec $ns2 ./pm_nl_ctl limits 1 2
-	ip netns exec $ns2 ./pm_nl_ctl add dead:beef:3::2 flags subflow
+	ip netns exec $ns2 ./pm_nl_ctl add dead:beef:3::2 dev ns2eth3 flags subflow
 	run_tests $ns1 $ns2 dead:beef:1::1 0 -1 -1 slow
 	chk_join_nr "remove subflow and signal IPv6" 2 2 2
 	chk_add_nr 1 1
Signature · 1 line
2.34.1
#4 lore raw
In reply to Mat Martineau
This series was applied to netdev/net.git (master) by Jakub Kicinski <[email protected]>:

Hello:

This series was applied to netdev/net.git (master) by Jakub Kicinski <[email protected]>:

Quoted 8 lines · On Thu, 20 Jan 2022 16:35:26 -0800 you wrote:

Here are some fixes from the mptcp tree:

Patch 1 fixes a RCU locking issue when processing a netlink command that updates endpoint flags in the in-kernel MPTCP path manager.

Patch 2 fixes a typo affecting available endpoint id tracking.

[...]

Here is the summary with links:
- [net,1/3] mptcp: fix msk traversal in mptcp_nl_cmd_set_flags()

    https://git.kernel.org/netdev/net/c/8e9eacad7ec7
  - [net,2/3] mptcp: fix removing ids bitmap setting
    https://git.kernel.org/netdev/net/c/a4c0214fbee9
  - [net,3/3] selftests: mptcp: fix ipv6 routing setup
    https://git.kernel.org/netdev/net/c/9846921dba49

You are awesome, thank you!

Signature · 2 lines
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html