netfilter-devel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Liping Zhang <zlpnobody@163.com>
To: pablo@netfilter.org
Cc: netfilter-devel@vger.kernel.org, Liping Zhang <zlpnobody@gmail.com>
Subject: [PATCH nf 4/5] netfilter: nfnl_cthelper: fix memory leak when do update
Date: Sun, 19 Mar 2017 22:36:01 +0800	[thread overview]
Message-ID: <1489934162-7415-5-git-send-email-zlpnobody@163.com> (raw)
In-Reply-To: <1489934162-7415-1-git-send-email-zlpnobody@163.com>

From: Liping Zhang <zlpnobody@gmail.com>

When invoke nfnl_cthelper_update, we will malloc a new expect_policy,
then only point the helper->expect_policy to the new one but ignore
the old one, so it will be leaked forever.

Another issue is that the user can modify the expect_class_max to a
new value, for example, decrease the expect_class_max from 3 to 0.
Then, exp->class created by ctnetlink_alloc_expect may become invalid,
and out of bound access will happen.

So keep it simple, when update the cthelper, reject to modify the
expect_class_max, and we can modify the struct nf_conntrack_expect_policy
directly instead of allocing a new object, then memory leak does not
exist anymore.

Signed-off-by: Liping Zhang <zlpnobody@gmail.com>
---
 net/netfilter/nfnetlink_cthelper.c | 35 +++++++++++++++++++++++------------
 1 file changed, 23 insertions(+), 12 deletions(-)

diff --git a/net/netfilter/nfnetlink_cthelper.c b/net/netfilter/nfnetlink_cthelper.c
index cc70dd5..fc4733f 100644
--- a/net/netfilter/nfnetlink_cthelper.c
+++ b/net/netfilter/nfnetlink_cthelper.c
@@ -156,7 +156,7 @@ nfnl_cthelper_expect_policy_set[NFCTH_POLICY_SET_MAX+1] = {
 
 static int
 nfnl_cthelper_parse_expect_policy(struct nf_conntrack_helper *helper,
-				  const struct nlattr *attr)
+				  const struct nlattr *attr, bool update)
 {
 	int i, ret;
 	struct nf_conntrack_expect_policy *expect_policy;
@@ -172,15 +172,23 @@ nfnl_cthelper_parse_expect_policy(struct nf_conntrack_helper *helper,
 		return -EINVAL;
 
 	class_max = ntohl(nla_get_be32(tb[NFCTH_POLICY_SET_NUM]));
-	if (class_max == 0)
-		return -EINVAL;
-	if (class_max > NF_CT_MAX_EXPECT_CLASSES)
-		return -EOVERFLOW;
 
-	expect_policy = kzalloc(sizeof(struct nf_conntrack_expect_policy) *
-				class_max, GFP_KERNEL);
-	if (expect_policy == NULL)
-		return -ENOMEM;
+	if (update) {
+		if (helper->expect_class_max + 1 != class_max)
+			return -EINVAL;
+
+		expect_policy = helper->expect_policy;
+	} else {
+		if (class_max == 0)
+			return -EINVAL;
+		if (class_max > NF_CT_MAX_EXPECT_CLASSES)
+			return -EOVERFLOW;
+
+		expect_policy = kcalloc(class_max, sizeof(*expect_policy),
+					GFP_KERNEL);
+		if (expect_policy == NULL)
+			return -ENOMEM;
+	}
 
 	for (i = 0; i < class_max; i++) {
 		if (!tb[NFCTH_POLICY_SET+i])
@@ -196,7 +204,8 @@ nfnl_cthelper_parse_expect_policy(struct nf_conntrack_helper *helper,
 	helper->expect_policy = expect_policy;
 	return 0;
 err:
-	kfree(expect_policy);
+	if (!update)
+		kfree(expect_policy);
 	return -EINVAL;
 }
 
@@ -214,7 +223,8 @@ nfnl_cthelper_create(const struct nlattr * const tb[],
 	if (helper == NULL)
 		return -ENOMEM;
 
-	ret = nfnl_cthelper_parse_expect_policy(helper, tb[NFCTH_POLICY]);
+	ret = nfnl_cthelper_parse_expect_policy(helper, tb[NFCTH_POLICY],
+						false);
 	if (ret < 0)
 		goto err1;
 
@@ -269,7 +279,8 @@ nfnl_cthelper_update(const struct nlattr * const tb[],
 
 	if (tb[NFCTH_POLICY]) {
 		ret = nfnl_cthelper_parse_expect_policy(helper,
-							tb[NFCTH_POLICY]);
+							tb[NFCTH_POLICY],
+							true);
 		if (ret < 0)
 			return ret;
 	}
-- 
2.5.5



  parent reply	other threads:[~2017-03-19 14:54 UTC|newest]

Thread overview: 21+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-03-19 14:35 [PATCH nf 0/5] netfilter: nfnl_cthelper: fix some bugs Liping Zhang
2017-03-19 14:35 ` [PATCH nf 1/5] netfilter: nfnl_cthelper: don't report error if NFCTH_PRIV_DATA_LEN is empty Liping Zhang
2017-03-21 10:18   ` Pablo Neira Ayuso
2017-03-21 14:26     ` Liping Zhang
2017-03-19 14:35 ` [PATCH nf 2/5] netfilter: nfnl_cthelper: fix incorrect helper->expect_class_max Liping Zhang
2017-03-21 10:27   ` Pablo Neira Ayuso
2017-03-21 14:35     ` Liping Zhang
2017-03-21 14:49       ` Pablo Neira Ayuso
2017-03-19 14:36 ` [PATCH nf 3/5] netfilter: drop const qualifier from struct nf_conntrack_expect_policy Liping Zhang
2017-03-21 10:34   ` Pablo Neira Ayuso
2017-03-21 15:00     ` Liping Zhang
2017-03-21 15:03       ` Pablo Neira Ayuso
2017-03-19 14:36 ` Liping Zhang [this message]
2017-03-21 10:32   ` [PATCH nf 4/5] netfilter: nfnl_cthelper: fix memory leak when do update Pablo Neira Ayuso
2017-03-21 13:23     ` Pablo Neira Ayuso
2017-03-19 14:36 ` [PATCH nf 5/5] netfilter: nfnl_cthelper: fix a race when walk the nf_ct_helper_hash table Liping Zhang
2017-03-21 10:33   ` Pablo Neira Ayuso
2017-03-21 14:48     ` Liping Zhang
2017-03-21 15:19       ` Liping Zhang
2017-03-21 15:26         ` Pablo Neira Ayuso
2017-03-22  0:06           ` Liping Zhang

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=1489934162-7415-5-git-send-email-zlpnobody@163.com \
    --to=zlpnobody@163.com \
    --cc=netfilter-devel@vger.kernel.org \
    --cc=pablo@netfilter.org \
    --cc=zlpnobody@gmail.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;
as well as URLs for NNTP newsgroup(s).