: Fix invalid return values from sctp_new
PATCH 1/1 : Fix invalid return values from sctp_new
Hi Kiran,
there are multiple places in sctp_new where it returns -1, but it looks like it really wants to return "invalid". The call to protocol->new in init_conntrack looks like this:
if (!protocol->new(conntrack, skb)) {
kmem_cache_free(ip_conntrack_cachep, conntrack);
return NULL;
}so it has to return 0 to say "invalid". Please check if the attached patch which fixes these places is correct.
Regards
Patrick
# This is a BitKeeper generated diff -Nru style patch. #
# ChangeSet
# 2004/09/20 10:06:40+02:00 [email protected]
# [NETFILTER]: Fix invalid return values from sctp_new
#
# Signed-off-by: Patrick McHardy <[email protected]>
#
# net/ipv4/netfilter/ip_conntrack_proto_sctp.c # 2004/09/20 10:06:16+02:00 [email protected] +5 -5 # [NETFILTER]: Fix invalid return values from sctp_new # # Signed-off-by: Patrick McHardy <[email protected]> # diff -Nru a/net/ipv4/netfilter/ip_conntrack_proto_sctp.c b/net/ipv4/netfilter/ip_conntrack_proto_sctp.c --- a/net/ipv4/netfilter/ip_conntrack_proto_sctp.c 2004-09-20 10:26:10 +02:00 +++ b/net/ipv4/netfilter/ip_conntrack_proto_sctp.c 2004-09-20 10:26:10 +02:00 @@ -430,16 +430,16 @@ DEBUGP("\n");
if (skb_copy_bits(skb, skb->nh.iph->ihl * 4, &sctph, sizeof(sctph)) != 0)
- return -1;
+ return 0;
if (do_basic_checks(conntrack, skb, map) != 0)
- return -1;
+ return 0;
/* If an OOTB packet has any of these chunks discard (Sec 8.4) */ if ((test_bit (SCTP_CID_ABORT, (void *)map))
|| (test_bit (SCTP_CID_SHUTDOWN_COMPLETE, (void *)map))
|| (test_bit (SCTP_CID_COOKIE_ACK, (void *)map))) {- return -1;
+ return 0;
}
newconntrack = SCTP_CONNTRACK_MAX;
@@ -461,7 +461,7 @@
if (skb_copy_bits(skb, offset + sizeof (sctp_chunkhdr_t),
&inithdr, sizeof(inithdr)) != 0) {
- return -1;
+ return 0;
}
DEBUGP("Setting vtag %x for new conn\n",
@@ -471,7 +471,7 @@
inithdr.init_tag;
} else {
/* Sec 8.5.1 (A) */
- return -1;
+ return 0;
}
}
/* If it is a shutdown ack OOTB packet, we expect a return
Quoted 11 lines · Patrick McHardy
there are multiple places in
sctp_newwhere it returns -1, but it looks like it really wants to return "invalid". The call toprotocol->newininit_conntracklooks like this:if (!protocol->new(conntrack, skb)) { kmem_cache_free(ip_conntrack_cachep, conntrack); return NULL; }so it has to return 0 to say "invalid". Please check if the attached patch which fixes these places is correct.
This seems fair. But the place I copied this from
[ ip_conntrack_proto_tcp.c:tcp_new() ] also needs to be fixed I guess..
@@ -237,7 +237,7 @@
struct tcphdr tcph;
if (skb_copy_bits(skb, skb->nh.iph->ihl * 4, &tcph, sizeof(tcph)) != 0)
- return -1;
+ return 0;
Signature · 2 lines
Regards, Kiran Kumar Immidi
Quoted 4 lines · Kiran Kumar Immidi
This seems fair. But the place I copied this from
[ ip_conntrack_proto_tcp.c:tcp_new() ] also needs to be fixed I guess..
Thanks. The others are fine, and I can't find an old version of ip_conntrack_proto_tcp where it returned -1.
Regards
Patrick
Quoted 2 lines · Patrick McHardy
Thanks. The others are fine, and I can't find an old version of
ip_conntrack_proto_tcpwhere it returned -1.
http://lxr.linux.no/source/net/ipv4/netfilter/ip_conntrack_proto_tcp.c?v=2.6.8.1#L240
I still find it in 2.6.8.1
Signature · 2 lines
Regards, Kiran Kumar Immidi
Quoted 5 lines · Kiran Kumar Immidi
http://lxr.linux.no/source/net/ipv4/netfilter/ip_conntrack_proto_tcp.c?v=2.6.8.1#L240
I still find it in 2.6.8.1
You're right. Seems like it was silently fixed during the skb_header_pointer conversion.
Regards
Patrick