netdev - [PATCH net] net/sched: act_ct: fix wild memory access when clearing fragments

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: <a72b5a4239942c5cc0152e4efba17b544a8a19fb.1619450967.git.dcaratti@redhat.com>
Date:   Mon, 26 Apr 2021 17:45:51 +0200
From:   Davide Caratti <dcaratti@...hat.com>
To:     Jamal Hadi Salim <jhs@...atatu.com>,
        Cong Wang <xiyou.wangcong@...il.com>,
        Jiri Pirko <jiri@...nulli.us>,
        "David S. Miller" <davem@...emloft.net>,
        Jakub Kicinski <kuba@...nel.org>, wenxu <wenxu@...oud.cn>,
        Marcelo Ricardo Leitner <marcelo.leitner@...il.com>,
        Shuang Li <shuali@...hat.com>
Cc:     netdev@...r.kernel.org
Subject: [PATCH net] net/sched: act_ct: fix wild memory access when clearing fragments

while testing re-assembly/re-fragmentation using act_ct, it's possible to
observe a crash like the following one:

 KASAN: maybe wild-memory-access in range [0x0001000000000448-0x000100000000044f]
 CPU: 50 PID: 0 Comm: swapper/50 Tainted: G S                5.12.0-rc7+ #424
 Hardware name: Dell Inc. PowerEdge R730/072T6D, BIOS 2.4.3 01/17/2017
 RIP: 0010:inet_frag_rbtree_purge+0x50/0xc0
 Code: 00 fc ff df 48 89 c3 31 ed 48 89 df e8 a9 7a 38 ff 4c 89 fe 48 89 df 49 89 c6 e8 5b 3a 38 ff 48 8d 7b 40 48 89 f8 48 c1 e8 03 <42> 80 3c 20 00 75 59 48 8d bb d0 00 00 00 4c 8b 6b 40 48 89 f8 48
 RSP: 0018:ffff888c31449db8 EFLAGS: 00010203
 RAX: 0000200000000089 RBX: 000100000000040e RCX: ffffffff989eb960
 RDX: 0000000000000140 RSI: ffffffff97cfb977 RDI: 000100000000044e
 RBP: 0000000000000900 R08: 0000000000000000 R09: ffffed1186289350
 R10: 0000000000000003 R11: ffffed1186289350 R12: dffffc0000000000
 R13: 000100000000040e R14: 0000000000000000 R15: ffff888155e02160
 FS:  0000000000000000(0000) GS:ffff888c31440000(0000) knlGS:0000000000000000
 CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
 CR2: 00005600cb70a5b8 CR3: 0000000a2c014005 CR4: 00000000003706e0
 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
 DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
 Call Trace:
  <IRQ>
  inet_frag_destroy+0xa9/0x150
  call_timer_fn+0x2d/0x180
  run_timer_softirq+0x4fe/0xe70
  __do_softirq+0x197/0x5a0
  irq_exit_rcu+0x1de/0x200
  sysvec_apic_timer_interrupt+0x6b/0x80
  </IRQ>

when act_ct temporarily stores an IP fragment, restoring the skb qdisc cb
results in putting random data in FRAG_CB(), and this causes those "wild"
memory accesses later, when the rbtree is purged. Never overwrite the skb
cb in case tcf_ct_handle_fragments() returns -EINPROGRESS.

Fixes: ae372cb1750f ("net/sched: act_ct: fix restore the qdisc_skb_cb after defrag")
Fixes: 7baf2429a1a9 ("net/sched: cls_flower add CT_FLAGS_INVALID flag support")
Reported-by: Shuang Li <shuali@...hat.com>
Signed-off-by: Davide Caratti <dcaratti@...hat.com>
---
 net/sched/act_ct.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/net/sched/act_ct.c b/net/sched/act_ct.c
index 16e888a9601d..48fdf7293dea 100644
--- a/net/sched/act_ct.c
+++ b/net/sched/act_ct.c
@@ -732,7 +732,8 @@ static int tcf_ct_handle_fragments(struct net *net, struct sk_buff *skb,
 #endif
 	}
 
-	*qdisc_skb_cb(skb) = cb;
+	if (err != -EINPROGRESS)
+		*qdisc_skb_cb(skb) = cb;
 	skb_clear_hash(skb);
 	skb->ignore_df = 1;
 	return err;
@@ -967,7 +968,7 @@ static int tcf_ct_act(struct sk_buff *skb, const struct tc_action *a,
 	err = tcf_ct_handle_fragments(net, skb, family, p->zone, &defrag);
 	if (err == -EINPROGRESS) {
 		retval = TC_ACT_STOLEN;
-		goto out;
+		goto out_clear;
 	}
 	if (err)
 		goto drop;
@@ -1030,7 +1031,6 @@ static int tcf_ct_act(struct sk_buff *skb, const struct tc_action *a,
 out_push:
 	skb_push_rcsum(skb, nh_ofs);
 
-out:
 	qdisc_skb_cb(skb)->post_ct = true;
 out_clear:
 	tcf_action_update_bstats(&c->common, skb);
-- 
2.30.2