Due to man page of setsockopt, if optlen is not valid, kernel should return
-EINVAL. But a simple testcase as following, errno is 0, which means setsockopt
is successful.
addr.s_addr = inet_addr("192.1.2.3");
setsockopt(s, IPPROTO_IP, IP_MULTICAST_IF, &addr, 1);
printf("errno is %d\n", errno);
This patch fixes the optlen check part, with the patch, we got errno EINVAL.
Signed-off-by: Xiaotian Feng <[email protected]>
Cc: Patrick McHardy <[email protected]>
Cc: David S. Miller <[email protected]>
---
net/ipv4/ip_sockglue.c | 8 ++++----
1 files changed, 4 insertions(+), 4 deletions(-)
diff --git a/net/ipv4/ip_sockglue.c b/net/ipv4/ip_sockglue.c
index fc7993e..5a29dce 100644
--- a/net/ipv4/ip_sockglue.c
+++ b/net/ipv4/ip_sockglue.c
@@ -615,13 +615,13 @@ static int do_ip_setsockopt(struct sock *sk, int level,
if (optlen >= sizeof(struct ip_mreqn)) {
if (copy_from_user(&mreq, optval, sizeof(mreq)))
break;
- } else {
+ } else if (optlen >= sizeof(struct in_addr)) {
memset(&mreq, 0, sizeof(mreq));
- if (optlen >= sizeof(struct in_addr) &&
- copy_from_user(&mreq.imr_address, optval,
+ if (copy_from_user(&mreq.imr_address, optval,
sizeof(struct in_addr)))
break;
- }
+ } else /* Invalid optlen */
+ goto e_inval;
if (!mreq.imr_ifindex) {
if (mreq.imr_address.s_addr == htonl(INADDR_ANY)) {
--
1.6.2.5
Xiaotian Feng wrote, at 09/17/2009 01:19 PM:
> Due to man page of setsockopt, if optlen is not valid, kernel should return
> -EINVAL. But a simple testcase as following, errno is 0, which means setsockopt
> is successful.
>
> addr.s_addr = inet_addr("192.1.2.3");
> setsockopt(s, IPPROTO_IP, IP_MULTICAST_IF, &addr, 1);
> printf("errno is %d\n", errno);
>
> This patch fixes the optlen check part, with the patch, we got errno EINVAL.
>
I also think it's a bug, the freebsd also does the optlen check.
But the style should be coincident with other option: firstly check the
availability of optlen, then copy option value from user and deal with it.
How about this one:
diff --git a/net/ipv4/ip_sockglue.c b/net/ipv4/ip_sockglue.c
index fc7993e..5a06935 100644
--- a/net/ipv4/ip_sockglue.c
+++ b/net/ipv4/ip_sockglue.c
@@ -611,6 +611,9 @@ static int do_ip_setsockopt(struct sock *sk, int level,
* Check the arguments are allowable
*/
+ if (optlen < sizeof(struct in_addr))
+ goto e_inval;
+
err = -EFAULT;
if (optlen >= sizeof(struct ip_mreqn)) {
if (copy_from_user(&mreq, optval, sizeof(mreq)))
Best Regards
-----
Shan Wei
From: Shan Wei <[email protected]>
Date: Thu, 17 Sep 2009 17:15:22 +0800
> Xiaotian Feng wrote, at 09/17/2009 01:19 PM:
>> Due to man page of setsockopt, if optlen is not valid, kernel should return
>> -EINVAL. But a simple testcase as following, errno is 0, which means setsockopt
>> is successful.
>>
>> addr.s_addr = inet_addr("192.1.2.3");
>> setsockopt(s, IPPROTO_IP, IP_MULTICAST_IF, &addr, 1);
>> printf("errno is %d\n", errno);
>>
>> This patch fixes the optlen check part, with the patch, we got errno EINVAL.
>>
>
> I also think it's a bug, the freebsd also does the optlen check.
> But the style should be coincident with other option: firstly check the
> availability of optlen, then copy option value from user and deal with it.
>
> How about this one:
This definitely is better and cleaner, but please don't post such
things without proper signoffs and commit messages because now
I have to ask you to do that instead of me just applying your
patch :-/
David Miller wrote, at 09/23/2009 04:38 AM:
> From: Shan Wei <[email protected]>
> Date: Thu, 17 Sep 2009 17:15:22 +0800
>
>> Xiaotian Feng wrote, at 09/17/2009 01:19 PM:
>>> Due to man page of setsockopt, if optlen is not valid, kernel should return
>>> -EINVAL. But a simple testcase as following, errno is 0, which means setsockopt
>>> is successful.
>>>
>>> addr.s_addr = inet_addr("192.1.2.3");
>>> setsockopt(s, IPPROTO_IP, IP_MULTICAST_IF, &addr, 1);
>>> printf("errno is %d\n", errno);
>>>
>>> This patch fixes the optlen check part, with the patch, we got errno EINVAL.
>>>
>> I also think it's a bug, the freebsd also does the optlen check.
>> But the style should be coincident with other option: firstly check the
>> availability of optlen, then copy option value from user and deal with it.
>>
>> How about this one:
>
> This definitely is better and cleaner, but please don't post such
> things without proper signoffs and commit messages because now
> I have to ask you to do that instead of me just applying your
> patch :-/
>
I'm so sorry about that. The whole patch is below.
[PATCH BUGFIX] ipv4: check optlen for IP_MULTICAST_IF option
Due to man page of setsockopt, if optlen is not valid, kernel should return
-EINVAL. But a simple testcase as following, errno is 0, which means setsockopt
is successful.
addr.s_addr = inet_addr("192.1.2.3");
setsockopt(s, IPPROTO_IP, IP_MULTICAST_IF, &addr, 1);
printf("errno is %d\n", errno);
Xiaotian Feng([email protected]) caught the bug. We fix it firstly checking
the availability of optlen and then dealing with the logic like other options.
Reported-by: Xiaotian Feng <[email protected]>
Signed-off-by: Shan Wei <[email protected]>
Acked-by: Alexey Kuznetsov <[email protected]>
---
net/ipv4/ip_sockglue.c | 3 +++
1 files changed, 3 insertions(+), 0 deletions(-)
diff --git a/net/ipv4/ip_sockglue.c b/net/ipv4/ip_sockglue.c
index fc7993e..5a06935 100644
--- a/net/ipv4/ip_sockglue.c
+++ b/net/ipv4/ip_sockglue.c
@@ -611,6 +611,9 @@ static int do_ip_setsockopt(struct sock *sk, int level,
* Check the arguments are allowable
*/
+ if (optlen < sizeof(struct in_addr))
+ goto e_inval;
+
err = -EFAULT;
if (optlen >= sizeof(struct ip_mreqn)) {
if (copy_from_user(&mreq, optval, sizeof(mreq)))
--
1.6.0.4
From: Shan Wei <[email protected]>
Date: Wed, 23 Sep 2009 09:41:10 +0800
> [PATCH BUGFIX] ipv4: check optlen for IP_MULTICAST_IF option
>
> Due to man page of setsockopt, if optlen is not valid, kernel should return
> -EINVAL. But a simple testcase as following, errno is 0, which means setsockopt
> is successful.
> addr.s_addr = inet_addr("192.1.2.3");
> setsockopt(s, IPPROTO_IP, IP_MULTICAST_IF, &addr, 1);
> printf("errno is %d\n", errno);
>
> Xiaotian Feng([email protected]) caught the bug. We fix it firstly checking
> the availability of optlen and then dealing with the logic like other options.
>
> Reported-by: Xiaotian Feng <[email protected]>
> Signed-off-by: Shan Wei <[email protected]>
> Acked-by: Alexey Kuznetsov <[email protected]>
Applied.