Re: [lwip-users] TCP causing out of mem pool [RAW]

[Kieran Mansley]

On Tue, 2009-07-28 at 06:34 -0700, Chris_S wrote:
> IE6.0.2800
> 
> But since you asked, I just tried FF as well.  It causes lwIP to crash
> quicker.
> By crash I mean stuck in DAbort.  On the 2nd page refresh.

Not sure what DAbort is, or what sort of things might trigger it, but
I'm guessing it's terminal.

> It appears FF is trying to load the "favicon.ico".
> My HTTPd returns the 404 correctly, 1st page load is fine.
> But on the 2nd page load, locks up.
> 
> Really don't know why.  Another wierdness yet to trace.
> 
> Here's a trace below.  I put a ping set in between 1st/2nd page refresh
> loads.

Hard to read through wireshark traces inline in an email (a pcap file is
best, with details about which address is running lwIP) but in this case
it looks unlikely that it's the same as bug24830.  That would show up as
connections with a zero window and lots of retransmissions or
zero-window probes.  In this trace we see 192.168.2.10 closing
connections but 192.168.2.1 not doing likewise.

If we're to make any progress in fixing this you're going to have to
look in more detail to see what is going wrong.  I think it would be
best to isolate each of the problems you're seeing and try to tackle
them one at a time.  At the moment you're describing lots of different
behaviours:
 - lwIP out of PCBs: this is probably OK as it's expected and should
recycle from the TIME_WAIT state.
 - lwIP stops responding to new TCP connections: this is not OK and
needs more detail.
 - lwIP crashes and goes to DAbort: this is not OK and needs more
detail.

Please also be aware that there are lots of interacting components here:
the network driver, lwIP itself, and the application/OS.  A bug in any
one of these could be causing the problems you're seeing.

Kieran