buglwIP - A Lightweight TCP/IP stack - Bugs: bug #1713, Out-of-sequence TCP segments...

 
 

You are not allowed to post comments on this tracker with your current authentication level.

bug #1713: Out-of-sequence TCP segments should be timed out

Submitter:  Adam Dunkels <adamdunkels>
Submitted:  Fri 15 Nov 2002 09:37:25 AM UTC
   
 
Category:  TCP Severity:  1 - Wish
Item Group:  Faulty Behaviour Status:  None
Privacy:  Public Assigned to:  None
Open/Closed:  Closed Planned Release:  None
lwIP version:  None

Mon 26 Mar 2007 11:49:26 AM UTC, comment #2: 

Closing, thanks for spotting this fix.

Kieran Mansley <kieranm>
Group Member
Fri 23 Mar 2007 02:43:52 PM UTC, comment #1: 


Noticed that this bug is still in the list... but browsing the code, it appears that a patch has already been applied to address this (see below). Perhaps this bug just needs a status update and closed? (I didn't go back and browse the history to see when this was added... but it was in at least by 0.7.2 I think).

in tcp_slowtmr():
-----------------------------------------------------------

    /* If this PCB has queued out of sequence data, but has been
       inactive for too long, will drop the data (it will eventually
       be retransmitted). */
#if TCP_QUEUE_OOSEQ
    if (pcb->ooseq != NULL &&
       (u32_t)tcp_ticks - pcb->tmr >=
       pcb->rto * TCP_OOSEQ_TIMEOUT) {
      tcp_segs_free(pcb->ooseq);
      pcb->ooseq = NULL;
      LWIP_DEBUGF(TCP_CWND_DEBUG, ("tcp_slowtmr: dropping OOSEQ queued data\n"))
    }
#endif /* TCP_QUEUE_OOSEQ */

------------------------------------------------------------

Jim Pettinato <jim_pettinato>
Group Member
Fri 15 Nov 2002 09:37:25 AM UTC, original submission:  

When TCP receives out-of-sequence segments, those are queued on the tcp_pcb. Those segments use up pbufs from the pbuf pool until in-sequence segments arrives to fill the gap. The problem is that when all pbufs from the pool are tied up in out-of-sequence queues, no new segments can be received.

The best way to deal with this is probably to add a timeout in tcp_slowtmr() which kills all out-of-sequence segments after four round-trip times (tcp_pcb->rto). During this time, new segments should have arrived.

Nothing will break when killing out-of-sequence segments, since they will eventually be retransmitted.

Adam Dunkels <adamdunkels>
Group Member

 

(Note: upload size limit is set to 16384 kB, after insertion of the required escape characters.)

No files currently attached

 

Depends on the following items: None found

Items that depend on this one: None found

 

Carbon-Copy List
  • -email is unavailable- added by kieranm (Posted a comment)
  • -email is unavailable- added by jim_pettinato (Posted a comment)
  •  

    There are 0 votes so far. Votes easily highlight which items people would like to see resolved in priority, independently of the priority of the item set by tracker managers.

     

    Follow 3 latest changes.

    Date Changed by Updated Field Previous Value => Replaced by
    2007-03-26 kieranm Open/ClosedOpen Closed
    2003-05-07 likewise CategoryNone None
        Item GroupNone None

    Back to the top

    Powered by Savane 3.13-02a9.
    Corresponding source code