netdev - Re: [PATCH net-next 2/3] xfrm: clamp down spi range for IPComp when allocating spi

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

Date:	Mon, 9 Dec 2013 14:27:43 +0800
From:	Fan Du <fan.du@...driver.com>
To:	Steffen Klassert <steffen.klassert@...unet.com>
CC:	<davem@...emloft.net>, <netdev@...r.kernel.org>
Subject: Re: [PATCH net-next 2/3] xfrm: clamp down spi range for IPComp when
 allocating spi



On 2013年12月06日 19:42, Steffen Klassert wrote:
> On Thu, Nov 28, 2013 at 10:52:40AM +0800, Fan Du wrote:
>> otherwise xfrm state can not be found properly by peers.
>>
>> Signed-off-by: Fan Du<fan.du@...driver.com>
>> ---
>>   net/xfrm/xfrm_state.c |   13 +++++++++++++
>>   1 file changed, 13 insertions(+)
>>
>> diff --git a/net/xfrm/xfrm_state.c b/net/xfrm/xfrm_state.c
>> index 68c2f35..a6716d7 100644
>> --- a/net/xfrm/xfrm_state.c
>> +++ b/net/xfrm/xfrm_state.c
>> @@ -1506,6 +1506,19 @@ int xfrm_alloc_spi(struct xfrm_state *x, u32 low, u32 high)
>>   	__be32 maxspi = htonl(high);
>>   	u32 mark = x->mark.v&  x->mark.m;
>>
>> +	/* Compression Parameter Index(CPI) is 16bits wide
>> +	 * An 32 bits spi value will hash xfrm_state into wrong hash slot.
>> +	 * When the upper 16bits of spi values is used as CPI for the peer
>> +	 * to look up xfrm state, it would generate XfrmOutNoStates error,
>> +	 * as apparently we are looking for the wrong hash slot.
>> +	 *
>> +	 * So clamp down the spi range into only 16bits valid wide.
>> +	 */
>> +	if (x->id.proto == IPPROTO_COMP) {
>> +		minspi = htonl(0xc00);
>> +		maxspi = htonl(0xff00);
>> +	}
>
> This does not make sense. The spi is chosen based on minspi only
> if minspi is equal to maxspi. This will be never the case for
> IPPROTO_COMP with your patch.
>
> Also, the spi range is user defined, we should respect the
> users configuration if the range is valid.

Ok, then, speaking of respect user defined range, how about below informal
patch which only check the validity of the range? My original thoughts is CPI
is only 16bits wide, kernel itself can keep the CPI's validity. btw, v2 will
also fix patch1/3 align issue.

diff --git a/net/xfrm/xfrm_state.c b/net/xfrm/xfrm_state.c
index 6a9c402..2c6fb99 100644
--- a/net/xfrm/xfrm_state.c
+++ b/net/xfrm/xfrm_state.c
@@ -1507,6 +1507,9 @@ int xfrm_alloc_spi(struct xfrm_state *x, u32 low, u32 high)

         err = -ENOENT;

+       if ((x->id.proto == IPPROTO_COMP) && (high > 0xFFFF))
+               goto unlock;
+
         if (minspi == maxspi) {
                 x0 = xfrm_state_lookup(net, mark, &x->id.daddr, minspi, x->id.proto, x->props.family);
                 if (x0) {

-- 
浮沉随浪只记今朝笑

--fan
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html