Re: [PATCH net] sctp: validate chunk length in the inqueue parser
From: Xin Long <lucien.xin@gmail.com>
Date: 2026-08-27 15:33:44
Also in:
linux-sctp, lkml
On Wed, Aug 26, 2026 at 9:01 PM Charles Vosburgh via B4 Relay [off-list ref] wrote:
From: Charles Vosburgh <redacted>
SCTP chunks always include a four-byte generic header, but
sctp_inq_pop() currently accepts shorter declared lengths. A zero-length
chunk leaves chunk_end at the current header.
When ASCONF is covered by the association's SCTP-AUTH policy,
sctp_assoc_bh_rcv() can continue before the state machine performs its
normal chunk-length check. sctp_inq_pop() then returns the same malformed
chunk repeatedly and the receive softirq can lock up.
A remote SCTP peer can trigger this after establishing an association on
a kernel built with CONFIG_IP_SCTP and configured with
net.sctp.addip_enable=1 and net.sctp.auth_enable=1. The reproducer did
not require application credentials, a shared SCTP AUTH key, or
net.sctp.addip_noauth_enable=1.
On commit f967455fb2a5 ("seg6: reset IP6CB after IPv6 decapsulation"),
one zero-length ASCONF caused repeated
watchdog soft-lockup reports in a two-vCPU KVM guest. All 3 pre-trigger
health probes succeeded, while 36 of 37 post-trigger probes failed. With
this change, all 37 post-trigger probes succeeded and no equivalent
soft-lockup signature appeared.
Reject chunks shorter than the generic SCTP header at the shared inqueue
parser boundary. Mark the packet for discard before either caller can
continue processing it, while preserving the four-byte generic minimum.
Declared-length 1 through 4 controls and kernel-generated ASCONF traffic
remained healthy. The patched sctp_hello selftest passed for IPv4 and
IPv6.
The complete private reproducer and validation evidence are available
directly to maintainers on request.Please share the PoC with maintainers.
quoted hunk ↗ jump to hunk
Fixes: bbd0d59809f9 ("[SCTP]: Implement the receive and verification of AUTH chunk") Cc: stable@vger.kernel.org Assisted-by: ChatGPT:GPT-5.6-Sol Assisted-by: Vantix:claude-opus-5 Assisted-by: Codex:GPT-5 Signed-off-by: Charles Vosburgh <redacted> --- net/sctp/inqueue.c | 5 +++++ 1 file changed, 5 insertions(+)diff --git a/net/sctp/inqueue.c b/net/sctp/inqueue.c index 5f988b3a8814..12a46868165a 100644 --- a/net/sctp/inqueue.c +++ b/net/sctp/inqueue.c@@ -212,6 +212,11 @@ struct sctp_chunk *sctp_inq_pop(struct sctp_inq *queue) chunk->chunk_end = ((__u8 *)ch) + SCTP_PAD4(ntohs(ch->length)); skb_pull(chunk->skb, sizeof(*ch)); chunk->subh.v = NULL; /* Subheader is no longer valid. */ + if (unlikely(ntohs(ch->length) < sizeof(*ch))) { + chunk->pdiscard = 1; + chunk->chunk_end = chunk->skb->data; + return chunk; + }
An early return of the chunk will skip the pr_debug(), which may not be what we want. Also, I don't think we have to update chunk->chunk_end.
if (chunk->chunk_end + sizeof(*ch) <= skb_tail_pointer(chunk->skb)) {
/* This is not a singleton */Maybe you can merge if (unlikely(ntohs(ch->length) < sizeof(*ch))) into the
check here, like:
if (unlikely(ntohs(ch->length) < sizeof(*ch))) {
chunk->pdiscard = 1;
} else if (chunk->chunk_end + sizeof(*ch) <=
skb_tail_pointer(chunk->skb)) {
...
Thanks.
--- base-commit: f967455fb2a5a2079b9eb5823e9ccf359174bf9f change-id: 20260826-sctp-zero-chunk-inqueue-b478dca552c6 Best regards, -- Charles Vosburgh [off-list ref]