From mboxrd@z Thu Jan 1 00:00:00 1970 From: Vlad Yasevich Subject: Re: [PATCH net v2] sctp: asconf's process should verify address parameter is in the beginning Date: Tue, 25 Aug 2015 12:03:59 -0400 Message-ID: <55DC91EF.8030706@redhat.com> References: <3ffe8f6b86e33c016dddec672fab23a206c21acf.1440505764.git.lucien.xin@gmail.com> <20150825140113.GD1873@localhost.localdomain> Reply-To: vyasevic@redhat.com Mime-Version: 1.0 Content-Type: text/plain; charset=windows-1252 Content-Transfer-Encoding: 7bit Cc: Marcelo Ricardo Leitner , network dev , davem@davemloft.net To: Xin Long Return-path: Received: from mx1.redhat.com ([209.132.183.28]:46335 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754784AbbHYQED (ORCPT ); Tue, 25 Aug 2015 12:04:03 -0400 In-Reply-To: <20150825140113.GD1873@localhost.localdomain> Sender: netdev-owner@vger.kernel.org List-ID: On 08/25/2015 10:01 AM, Marcelo Ricardo Leitner wrote: > On Tue, Aug 25, 2015 at 08:29:24PM +0800, Xin Long wrote: >> in sctp_process_asconf(), we get address parameter from the beginning of >> the addip params. but we never check if it's really there. if the addr >> param is not there, it still can pass sctp_verify_asconf(), then to be >> handled by sctp_process_asconf(), it will not be safe. >> >> so add a code in sctp_verify_asconf() to check the address parameter is in >> the beginning, or return false to send abort. >> >> v1->v2: >> * put the check behind the params' length verify. >> >> Signed-off-by: Xin Long >> --- >> net/sctp/sm_make_chunk.c | 7 +++++++ >> 1 file changed, 7 insertions(+) >> >> diff --git a/net/sctp/sm_make_chunk.c b/net/sctp/sm_make_chunk.c >> index 06320c8..89a4d1c 100644 >> --- a/net/sctp/sm_make_chunk.c >> +++ b/net/sctp/sm_make_chunk.c >> @@ -3166,6 +3166,13 @@ bool sctp_verify_asconf(const struct sctp_association *asoc, >> return false; >> if (!addr_param_needed && addr_param_seen) >> return false; >> + if (addr_param_needed && addr_param_seen) { >> + /* Ensure the address parameter is in the beginning */ >> + param.v = chunk->skb->data + sizeof(sctp_addiphdr_t); > > Using param.v before the loop made sense but after the loop, it will > cause all packets that hits here to be reject due to the check below. > >> + if (param.p->type != SCTP_PARAM_IPV4_ADDRESS && >> + param.p->type != SCTP_PARAM_IPV6_ADDRESS) >> + return false; >> + } >> if (param.v != chunk->chunk_end) > this one -----^ > > Maybe it's easier if you put this check inside the loop for each ipv4/6, > and check if it is the first parameter or not by mimicing the way > sctp_walk_params() finds the first chunk, it's just a pointer > derreference and that was already checked and performed to reach there. > > (You can have some logic with addr_param_seen so you don't catch the > multiple parameters in there.) Exactly! something like this: SCTP_PARAM_IPV4_ADDRESS: if (param.v == addip->addip_hdr.params) addr_param_seen = true; Thus making sure that the parameter as seen only when it's at the beginning... Then later we can do things like: SCTP_PARAM_IPV4_ADDRESS: if (addr_param_seen) { /* peer placed multiple address parameters into the same * asconf. reject it. */ return false; } -vlad > > Marcelo > >> return false; >> >> -- >> 2.1.0 >>