linux-kernel - [PATCH net] net/sched: flower: Fix wrong handle assignment during filter change

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [thread-next>] [day] [month] [year] [list]

Message-Id: <20230425140604.169881-1-ivecera@redhat.com>
Date:   Tue, 25 Apr 2023 16:06:04 +0200
From:   Ivan Vecera <ivecera@...hat.com>
To:     netdev@...r.kernel.org
Cc:     Jamal Hadi Salim <jhs@...atatu.com>,
        Cong Wang <xiyou.wangcong@...il.com>,
        Jiri Pirko <jiri@...nulli.us>,
        "David S. Miller" <davem@...emloft.net>,
        Eric Dumazet <edumazet@...gle.com>,
        Jakub Kicinski <kuba@...nel.org>,
        Paolo Abeni <pabeni@...hat.com>,
        Simon Horman <simon.horman@...igine.com>,
        Marcelo Ricardo Leitner <marcelo.leitner@...il.com>,
        Paul Blakey <paulb@...dia.com>,
        linux-kernel@...r.kernel.org (open list)
Subject: [PATCH net] net/sched: flower: Fix wrong handle assignment during filter change

Commit 08a0063df3ae ("net/sched: flower: Move filter handle initialization
earlier") moved filter handle initialization but an assignment of
the handle to fnew->handle is done regardless of fold value. This is wrong
because if fold != NULL (so fold->handle == handle) no new handle is
allocated and passed handle is assigned to fnew->handle. Then if any
subsequent action in fl_change() fails then the handle value is
removed from IDR that is incorrect as we will have still valid old filter
instance with handle that is not present in IDR.
Fix this issue by moving the assignment so it is done only when passed
fold == NULL.

Prior the patch:
[root@...hine tc-testing]# ./tdc.py -d enp1s0f0np0 -e 14be
Test 14be: Concurrently replace same range of 100k flower filters from 10 tc instances
exit: 123
exit: 0
RTNETLINK answers: Invalid argument
We have an error talking to the kernel
Command failed tmp/replace_6:1885


All test results:

1..1
not ok 1 14be - Concurrently replace same range of 100k flower filters from 10 tc instances
        Command exited with 123, expected 0
RTNETLINK answers: Invalid argument
We have an error talking to the kernel
Command failed tmp/replace_6:1885

After the patch:
[root@...hine tc-testing]# ./tdc.py -d enp1s0f0np0 -e 14be
Test 14be: Concurrently replace same range of 100k flower filters from 10 tc instances

All test results:

1..1
ok 1 14be - Concurrently replace same range of 100k flower filters from 10 tc instances

Fixes: 08a0063df3ae ("net/sched: flower: Move filter handle initialization earlier")
Signed-off-by: Ivan Vecera <ivecera@...hat.com>
---
 net/sched/cls_flower.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/net/sched/cls_flower.c b/net/sched/cls_flower.c
index 475fe222a855..fa6c2bb0b626 100644
--- a/net/sched/cls_flower.c
+++ b/net/sched/cls_flower.c
@@ -2231,8 +2231,8 @@ static int fl_change(struct net *net, struct sk_buff *in_skb,
 			kfree(fnew);
 			goto errout_tb;
 		}
+		fnew->handle = handle;
 	}
-	fnew->handle = handle;
 
 	err = tcf_exts_init_ex(&fnew->exts, net, TCA_FLOWER_ACT, 0, tp, handle,
 			       !tc_skip_hw(fnew->flags));
-- 
2.39.1