RE: CE5 Debugger Disconnect

From: J (J_at_discussions.microsoft.com)
Date: 10/29/04 

Date: Fri, 29 Oct 2004 13:31:02 -0700

Update 2

It looks like PlatformBuilder is attempting to start another download, and
failing because the target device is already running (I can see packets
containing boot.bin.octet, which normally comes with the eboot protocol).
The download attempt eventually times out, and PlatformBuilder disconnects.
If I configure PB to not download always, the problem goes away.

This is different from 4.2, because 4.2 had the ability to initiate a
debugger connection separate from the download (the download could be
download only, and connect could be connect only). For some reason, these
two options were consolidated in 5.0 making it less flexible.

So, the question is this: What makes PB think that it should initiate a
download? Is it seeing packets on the network and interpreting them as
bootme packets? What is the recommended way of getting around this problem?

Any insight (preferably from MS) would be greatly appreciated...

"J" wrote:

> Update
>
> I've instrumented the code with lots of (serial) debug messages and noticed
> that the device stops sending data across the link after the debug break has
> been hit. It still receives data, but never makes any attempt to send
> anything. It's not as if the adapter is unable to transmit, the kernel isn't
> even trying.
>
> Obviously, debug messages are no longer flowing since I'm at a breakpoint,
> but I still think there is some kind of keep alive message that is probably
> not making it.
>
> Is there a change necessary in slightly higher level code that I missed on
> the port from CE42 to CE5?
>
> > I've got a problem where the CE5 debugger is disconnecting when it hits a
> > breakpoint. It appears to be some kind of timeout mechanism that wasn't
> > there in 4.2. This has only been observed with ethernet debugging
> > (LAN91C111), serial debugging is working fine.
> >
> > If you continue from the breakpoint within a couple of seconds, it will not
> > timeout. The ethernet debug code is the same as was used previously in 4.2,
> > where it was working fine. Not using the PQOAL libraries, although I've been
> > going through them to see if anything in supplied LAN91C111 libraries gives
> > me any ideas...
> >
> > What's different in CE5, and suggestions on where I should look to resolve
> > this problem?
> >
> >

Relevant Pages

  • Re: Strange PPPoe problem
    ... have PPPoe on my firewall. ... until it tries to download the mail (I see it login, ... sounds like icmp 3/4 packets are being dropped somewhere. ... firewall on the mail server is causing icmp 3/4s from reaching the mail ...
    (Debian-User)
  • Re: Data transfer over internet
    ... 1,28 to 1.29 KB/sec ... Interesting pattern with small gaps every 6 packets? ... I will log a QT download from APPLE to compare a SIMPLEX ... full and half duplex or do you really mean a true one-way transmission? ...
    (comp.sys.mac.system)
  • Re: OS image will not load driver.
    ... I put the CPU that I previously remove from the dual-processor. ... working so I though I could use sboot.bin to download the ... So I decided to copy nk.bin to the target device and see if I can see ...
    (microsoft.public.windowsce.platbuilder)
  • Re: Getting error 80070005, looks like a connect issue
    ... Although you are both getting the same code if you compare the traces ... that the download was going to be started and the message which indicates ... there are any packets sent then. ... start a netstat loop before allowing the install ...
    (microsoft.public.windowsupdate)
  • Re: Strange PPPoe problem
    ... The new service uses PPPoe - not a problem, ... thought - I have PPPoe on my firewall. ... until it tries to download the mail (I see it login, ... sounds like icmp 3/4 packets are being dropped somewhere. ...
    (Debian-User)