[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <47EAFB12.9040800@cn.fujitsu.com>
Date: Thu, 27 Mar 2008 09:40:34 +0800
From: Gui Jianfeng <guijianfeng@...fujitsu.com>
To: Vlad Yasevich <vladislav.yasevich@...com>
CC: Wei Yongjun <yjwei@...fujitsu.com>,
netdev <netdev@...r.kernel.org>,
lksctp-dev <lksctp-developers@...ts.sourceforge.net>,
David Miller <davem@...emloft.net>
Subject: Re: [PATCH] SCTP: Fix Protocol violation when receiving a error length
INIT ACK
Vlad Yasevich wrote:
> There is a side-effect to this patch that now we will completely ignore
> the verification
> tag in the INIT-ACK regardless of the violation.
>
> In particular, if the INIT-ACK contains all the fixed parameters but
> violates structure
> in some variable parameters, we'll currently use the Initiate Tag from
> the INIT-ACK in
> the ABORT. We should not be changing this behavior.
>
> The simple hack is to add some conditional code into the the different
> violation functions, but
> I'd like to see if there is a cleaner way to solve this.
Vlad,
The problem you said has been addressed, this is a new patch. Just treat
an INIT-ACK received during COOKIE-WAIT as a special case. this patch will
not break the normal behavior.
Signed-off-by: Gui Jianfeng <guijianfeng@...fujitsu.com>
---
include/net/sctp/command.h | 1 +
net/sctp/outqueue.c | 3 +++
net/sctp/sm_sideeffect.c | 5 ++++-
net/sctp/sm_statefuns.c | 18 ++++++++++++++++++
4 files changed, 26 insertions(+), 1 deletions(-)
diff --git a/include/net/sctp/command.h b/include/net/sctp/command.h
index 10ae2da..35b1e83 100644
--- a/include/net/sctp/command.h
+++ b/include/net/sctp/command.h
@@ -104,6 +104,7 @@ typedef enum {
SCTP_CMD_ADAPTATION_IND, /* generate and send adaptation 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_LAST
} sctp_verb_t;
diff --git a/net/sctp/outqueue.c b/net/sctp/outqueue.c
index 1bb3c5c..c071446 100644
--- a/net/sctp/outqueue.c
+++ b/net/sctp/outqueue.c
@@ -793,6 +793,9 @@ int sctp_outq_flush(struct sctp_outq *q, int rtx_timeout)
break;
case SCTP_CID_ABORT:
+ if (sctp_test_T_bit(chunk)) {
+ packet->vtag = asoc->c.my_vtag;
+ }
case SCTP_CID_SACK:
case SCTP_CID_HEARTBEAT:
case SCTP_CID_HEARTBEAT_ACK:
diff --git a/net/sctp/sm_sideeffect.c b/net/sctp/sm_sideeffect.c
index 28eb38e..2dbc7bd 100644
--- a/net/sctp/sm_sideeffect.c
+++ b/net/sctp/sm_sideeffect.c
@@ -1536,7 +1536,10 @@ static int sctp_cmd_interpreter(sctp_event_t event_type,
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;
+
default:
printk(KERN_WARNING "Impossible command: %u, %p\n",
cmd->verb, cmd->obj.ptr);
diff --git a/net/sctp/sm_statefuns.c b/net/sctp/sm_statefuns.c
index f2ed647..b3ed52c 100644
--- a/net/sctp/sm_statefuns.c
+++ b/net/sctp/sm_statefuns.c
@@ -4144,6 +4144,24 @@ static sctp_disposition_t sctp_sf_abort_violation(
goto nomem;
if (asoc) {
+ /* Treat INIT-ACK as a special case during COOKIE-WAIT. */
+ if (chunk->chunk_hdr->type == SCTP_CID_INIT_ACK
+ && !asoc->peer.i.init_tag) {
+ sctp_initack_chunk_t *initack;
+
+ initack = (sctp_initack_chunk_t *)chunk->chunk_hdr;
+ if (!sctp_chunk_length_valid(chunk,
+ sizeof(sctp_initack_chunk_t)))
+ abort->chunk_hdr->flags |= SCTP_CHUNK_FLAG_T;
+ else {
+ unsigned int inittag;
+
+ inittag = ntohl(initack->init_hdr.init_tag);
+ sctp_add_cmd_sf(commands, SCTP_CMD_UPDATE_INITTAG,
+ SCTP_U32(inittag));
+ }
+ }
+
sctp_add_cmd_sf(commands, SCTP_CMD_REPLY, SCTP_CHUNK(abort));
SCTP_INC_STATS(SCTP_MIB_OUTCTRLCHUNKS);
--
1.5.3
--
Regards
Gui Jianfeng
--
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
Powered by blists - more mailing lists