mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Kyuwon Cho <kyuwoncho18@gmail.com>
To: Xin Long <lucien.xin@gmail.com>
Cc: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>,
	"David S. Miller" <davem@davemloft.net>,
	Eric Dumazet <edumazet@kernel.org>,
	Jakub Kicinski <kuba@kernel.org>, Paolo Abeni <pabeni@redhat.com>,
	Simon Horman <horms@kernel.org>,
	Willem de Bruijn <willemb@google.com>,
	Jun Yang <junvyyang@tencent.com>,
	linux-sctp@vger.kernel.org, netdev@vger.kernel.org,
	linux-kernel@vger.kernel.org
Subject: [PATCH net v2] sctp: fail the association when the INIT-ACK cannot be processed
Date: Fri, 20 Nov 2026 13:55:05 +0900	[thread overview]
Message-ID: <179514851805.604157.8527055700821138420@gmail.com> (raw)
In-Reply-To: <CADvbK_citvtZ4eumLcx_szUvth7sQMP1oKu0mLbo0a3bD-eSxQ@mail.gmail.com>

sctp_sf_do_5_1C_ack() queues SCTP_CMD_PEER_INIT and SCTP_CMD_ASSOC_SHKEY.
If either fails to allocate memory, sctp_cmd_interpreter() stops there,
but the association stays: with the peer marked AUTH capable and
asoc->shkey still NULL, a chunk that the peer wants authenticated, such
as the ABORT built by close() with SO_LINGER {1, 0}, takes a reference
on the NULL key:

  BUG: KASAN: null-ptr-deref in sctp_auth_shkey_hold (net/sctp/auth.c:113)
  Write of size 4 at addr 0000000000000018
  Call Trace:
   sctp_auth_shkey_hold (net/sctp/auth.c:113)
   sctp_control_set_owner_w (net/sctp/sm_make_chunk.c:112)
   sctp_make_abort_user (net/sctp/sm_make_chunk.c:1016)
   sctp_close (net/sctp/socket.c:1564)

Process the INIT-ACK and set up the association key directly in
sctp_sf_do_5_1C_ack(), and fail the association if either step fails.
SCTP_CMD_PEER_INIT and SCTP_CMD_ASSOC_SHKEY have no other users, so
remove them.

Fixes: 1b1e0bc99474 ("sctp: add refcnt support for sh_key")
Suggested-by: Xin Long <lucien.xin@gmail.com>
Signed-off-by: Kyuwon Cho <kyuwoncho18@gmail.com>
---
v2:
- As Xin suggested, process the INIT-ACK and set up the association key
  directly in sctp_sf_do_5_1C_ack() and fail the association with
  SCTP_CMD_INIT_FAILED, instead of skipping AUTH in sctp_auth_send_cid().
  Remove SCTP_CMD_PEER_INIT and SCTP_CMD_ASSOC_SHKEY.
v1: https://lore.kernel.org/netdev/179505710266.3511678.14017674052424733262@gmail.com/

The patch was written with LLM assistance; we reviewed and tested it.
With it, the reproducer from the report no longer triggers the report,
also when the injected failure hits sctp_process_init() instead of the
association key.

 include/net/sctp/command.h |  4 ----
 net/sctp/sm_sideeffect.c   | 40 --------------------------------------
 net/sctp/sm_statefuns.c    | 15 +++++++++++---
 3 files changed, 12 insertions(+), 47 deletions(-)

diff --git a/include/net/sctp/command.h b/include/net/sctp/command.h
index 2058fabffbf6..8596e96c321a 100644
--- a/include/net/sctp/command.h
+++ b/include/net/sctp/command.h
@@ -34,7 +34,6 @@ enum sctp_verb {
 	SCTP_CMD_GEN_SACK,	/* Send a Selective ACK (maybe).  */
 	SCTP_CMD_PROCESS_SACK,	/* Process an inbound SACK.  */
 	SCTP_CMD_GEN_INIT_ACK,	/* Generate an INIT ACK chunk.  */
-	SCTP_CMD_PEER_INIT,	/* Process a INIT from the peer.  */
 	SCTP_CMD_GEN_COOKIE_ECHO, /* Generate a COOKIE ECHO chunk. */
 	SCTP_CMD_CHUNK_ULP,	/* Send a chunk to the sockets layer.  */
 	SCTP_CMD_EVENT_ULP,	/* Send a notification to the sockets layer. */
@@ -86,7 +85,6 @@ enum sctp_verb {
 	SCTP_CMD_ASSOC_CHANGE,	 /* generate and send assoc_change event */
 	SCTP_CMD_ADAPTATION_IND, /* generate and send adaptation event */
 	SCTP_CMD_PEER_NO_AUTH,   /* generate and send authentication event */
-	SCTP_CMD_ASSOC_SHKEY,    /* generate the association shared keys */
 	SCTP_CMD_T1_RETRAN,	 /* Mark for retransmission after T1 timeout  */
 	SCTP_CMD_UPDATE_INITTAG, /* Update peer inittag */
 	SCTP_CMD_SEND_MSG,	 /* Send the whole use message */
@@ -117,7 +115,6 @@ union sctp_arg {
 	struct sctp_association *asoc;
 	struct sctp_transport *transport;
 	struct sctp_bind_addr *bp;
-	struct sctp_init_chunk *init;
 	struct sctp_ulpevent *ulpevent;
 	struct sctp_packet *packet;
 	struct sctp_sackhdr *sackh;
@@ -158,7 +155,6 @@ SCTP_ARG_CONSTRUCTOR(CHUNK,	struct sctp_chunk *, chunk)
 SCTP_ARG_CONSTRUCTOR(ASOC,	struct sctp_association *, asoc)
 SCTP_ARG_CONSTRUCTOR(TRANSPORT,	struct sctp_transport *, transport)
 SCTP_ARG_CONSTRUCTOR(BA,	struct sctp_bind_addr *, bp)
-SCTP_ARG_CONSTRUCTOR(PEER_INIT,	struct sctp_init_chunk *, init)
 SCTP_ARG_CONSTRUCTOR(ULPEVENT,  struct sctp_ulpevent *, ulpevent)
 SCTP_ARG_CONSTRUCTOR(PACKET,	struct sctp_packet *, packet)
 SCTP_ARG_CONSTRUCTOR(SACKH,	struct sctp_sackhdr *, sackh)
diff --git a/net/sctp/sm_sideeffect.c b/net/sctp/sm_sideeffect.c
index 35f540fb15fc..a95bf10b852b 100644
--- a/net/sctp/sm_sideeffect.c
+++ b/net/sctp/sm_sideeffect.c
@@ -698,32 +698,6 @@ static void sctp_cmd_assoc_failed(struct sctp_cmd_seq *commands,
 	sctp_add_cmd_sf(commands, SCTP_CMD_DELETE_TCB, SCTP_NULL());
 }
 
-/* Process an init chunk (may be real INIT/INIT-ACK or an embedded INIT
- * inside the cookie.  In reality, this is only used for INIT-ACK processing
- * since all other cases use "temporary" associations and can do all
- * their work in statefuns directly.
- */
-static int sctp_cmd_process_init(struct sctp_cmd_seq *commands,
-				 struct sctp_association *asoc,
-				 struct sctp_chunk *chunk,
-				 struct sctp_init_chunk *peer_init,
-				 gfp_t gfp)
-{
-	int error;
-
-	/* We only process the init as a sideeffect in a single
-	 * case.   This is when we process the INIT-ACK.   If we
-	 * fail during INIT processing (due to malloc problems),
-	 * just return the error and stop processing the stack.
-	 */
-	if (!sctp_process_init(asoc, chunk, sctp_source(chunk), peer_init, gfp))
-		error = -ENOMEM;
-	else
-		error = 0;
-
-	return error;
-}
-
 /* Helper function to break out starting up of heartbeat timers.  */
 static void sctp_cmd_hb_timers_start(struct sctp_cmd_seq *cmds,
 				     struct sctp_association *asoc)
@@ -1397,16 +1371,6 @@ static int sctp_cmd_interpreter(enum sctp_event_type event_type,
 					SCTP_CHUNK(new_obj));
 			break;
 
-		case SCTP_CMD_PEER_INIT:
-			/* Process a unified INIT from the peer.
-			 * Note: Only used during INIT-ACK processing.  If
-			 * there is an error just return to the outter
-			 * layer which will bail.
-			 */
-			error = sctp_cmd_process_init(commands, asoc, chunk,
-						      cmd->obj.init, gfp);
-			break;
-
 		case SCTP_CMD_GEN_COOKIE_ECHO:
 			/* Generate a COOKIE ECHO chunk.  */
 			new_obj = sctp_make_cookie_echo(asoc, chunk);
@@ -1772,10 +1736,6 @@ static int sctp_cmd_interpreter(enum sctp_event_type event_type,
 			sctp_cmd_peer_no_auth(commands, asoc);
 			break;
 
-		case SCTP_CMD_ASSOC_SHKEY:
-			error = sctp_auth_asoc_init_active_key(asoc,
-						GFP_ATOMIC);
-			break;
 		case SCTP_CMD_UPDATE_INITTAG:
 			asoc->peer.i.init_tag = cmd->obj.u32;
 			break;
diff --git a/net/sctp/sm_statefuns.c b/net/sctp/sm_statefuns.c
index 64519fd8f943..598eafdc5261 100644
--- a/net/sctp/sm_statefuns.c
+++ b/net/sctp/sm_statefuns.c
@@ -603,13 +603,22 @@ enum sctp_disposition sctp_sf_do_5_1C_ack(struct net *net,
 
 	initchunk = (struct sctp_init_chunk *)chunk->chunk_hdr;
 
-	sctp_add_cmd_sf(commands, SCTP_CMD_PEER_INIT,
-			SCTP_PEER_INIT(initchunk));
+	if (!sctp_process_init((struct sctp_association *)asoc, chunk,
+			       sctp_source(chunk), initchunk, GFP_ATOMIC)) {
+		sctp_add_cmd_sf(commands, SCTP_CMD_INIT_FAILED,
+				SCTP_U16(SCTP_ERROR_NO_RESOURCE));
+		return SCTP_DISPOSITION_CONSUME;
+	}
 
 	/* SCTP-AUTH: generate the association shared keys so that
 	 * we can potentially sign the COOKIE-ECHO.
 	 */
-	sctp_add_cmd_sf(commands, SCTP_CMD_ASSOC_SHKEY, SCTP_NULL());
+	if (sctp_auth_asoc_init_active_key((struct sctp_association *)asoc,
+					   GFP_ATOMIC)) {
+		sctp_add_cmd_sf(commands, SCTP_CMD_INIT_FAILED,
+				SCTP_U16(SCTP_ERROR_NO_RESOURCE));
+		return SCTP_DISPOSITION_CONSUME;
+	}
 
 	/* Reset init error count upon receipt of INIT-ACK.  */
 	sctp_add_cmd_sf(commands, SCTP_CMD_INIT_COUNTER_RESET, SCTP_NULL());
-- 
2.43.0


      reply	other threads:[~2026-11-20  4:55 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-11-19  3:04 [BUG] sctp: auth: NULL asoc->shkey after a failed SCTP_CMD_ASSOC_SHKEY makes close() oops Kyuwon Cho
2026-11-19 20:01 ` Xin Long
2026-11-20  4:55   ` Kyuwon Cho [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=179514851805.604157.8527055700821138420@gmail.com \
    --to=kyuwoncho18@gmail.com \
    --cc=davem@davemloft.net \
    --cc=edumazet@kernel.org \
    --cc=horms@kernel.org \
    --cc=junvyyang@tencent.com \
    --cc=kuba@kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-sctp@vger.kernel.org \
    --cc=lucien.xin@gmail.com \
    --cc=marcelo.leitner@gmail.com \
    --cc=netdev@vger.kernel.org \
    --cc=pabeni@redhat.com \
    --cc=willemb@google.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

all inboxes | Powered by JetHome®