NDIS Driver debugging presentation at the MS Dev conference in November

List,
At the upcoming Driver Developer's Conference at
Microsoft's main campus, located in Redmond, WA (November
11- 14), the NDIS team will be presenting a session on
Debugging NDIS drivers.

This is your opportunity to tell us about anything
specific you would like to see demonstrated in this
presentation. Want to see how we track down pool
corruption? This is your chance to ask for it. How can you
determine if an NDIS Packet has been completed twice? We
can show you how. Need to track down that memory leak? We
have some of the best leak detector's in the business.

Email to me (jamesmat@microsoft.com) the issues you would
like to see debugged and explained, and we'll talk about
them at DevCon.

See you there!

Jim Mateer
Program Manager
Core Networking
Microsoft Corporation


This posting is provided "AS IS" with no warranties, and
confers no rights.

Duh
Sun Sep 14 11:34:18 CDT 2003

Jim Mateer [MS] wrote:
> List,
> At the upcoming Driver Developer's Conference at
> Microsoft's main campus, located in Redmond, WA (November
> 11- 14), the NDIS team will be presenting a session on
> Debugging NDIS drivers.
>
> This is your opportunity to tell us about anything
> specific you would like to see demonstrated in this
> presentation. Want to see how we track down pool

Want to see how to debug PNP/PM bugs at single machine environment.

Maxim
Sun Sep 14 20:23:39 CDT 2003

> Use softice. Better yet, get another machine, they are cheap.

A notebook is a good idea for PM debugging - they support S2 and S3, and maybe
even S1 (though I dunno how to employ S1 standby on the ones I saw).

--
Maxim Shatskih, Windows DDK MVP
StorageCraft Corporation
maxim@storagecraft.com
http://www.storagecraft.com

Duh
Sat Sep 20 11:40:55 CDT 2003

Mark Roddy wrote:
> On Sun, 14 Sep 2003 20:34:18 +0400, Duh Pohmelja
> <d2u7h1828p1828o45h90m45e23lj536a@people.ru> wrote:
>
>
>>Jim Mateer [MS] wrote:
>>
>>>List,
>>> At the upcoming Driver Developer's Conference at
>>>Microsoft's main campus, located in Redmond, WA (November
>>>11- 14), the NDIS team will be presenting a session on
>>>Debugging NDIS drivers.
>>>
>>>This is your opportunity to tell us about anything
>>>specific you would like to see demonstrated in this
>>>presentation. Want to see how we track down pool
>>
>>Want to see how to debug PNP/PM bugs at single machine environment.
>
>
> Use softice. Better yet, get another machine, they are cheap.
Nor SoftIce nor another machine won't help when something deadlock
inside PM manager - notebook simply hangs during awaking... neither it
won't accept network connections nor react keyboard/mice... just spins
HDD and blinks monitor (as it is in standby). I can see it starts up,
but then, in the middle of the show - PM manager deadlocks and no sense
where is it.

It is very appreciate to see how such errors can be debugged. May be
some "static testing" will help, because it is obviously a thread
synchronization problem (the behavior described earlier happens approx
10% times).

Would you try to answer once again? =)

Alexander
Sat Sep 20 15:38:03 CDT 2003

Use remote debugging with WinBug. Run with DriverVerifier. Put debug
messages into your driver to see how far it goes.

"Duh Pohmelja" <d2u7h1828p1828o45h90m45e23lj536a@people.ru> wrote in message
news:OlcTkZ5fDHA.908@tk2msftngp13.phx.gbl...
> Mark Roddy wrote:
> > On Sun, 14 Sep 2003 20:34:18 +0400, Duh Pohmelja
> > <d2u7h1828p1828o45h90m45e23lj536a@people.ru> wrote:
> >
> >
> >>Jim Mateer [MS] wrote:
> >>
> >>>List,
> >>> At the upcoming Driver Developer's Conference at
> >>>Microsoft's main campus, located in Redmond, WA (November
> >>>11- 14), the NDIS team will be presenting a session on
> >>>Debugging NDIS drivers.
> >>>
> >>>This is your opportunity to tell us about anything
> >>>specific you would like to see demonstrated in this
> >>>presentation. Want to see how we track down pool
> >>
> >>Want to see how to debug PNP/PM bugs at single machine environment.
> >
> >
> > Use softice. Better yet, get another machine, they are cheap.
> Nor SoftIce nor another machine won't help when something deadlock
> inside PM manager - notebook simply hangs during awaking... neither it
> won't accept network connections nor react keyboard/mice... just spins
> HDD and blinks monitor (as it is in standby). I can see it starts up,
> but then, in the middle of the show - PM manager deadlocks and no sense
> where is it.
>
> It is very appreciate to see how such errors can be debugged. May be
> some "static testing" will help, because it is obviously a thread
> synchronization problem (the behavior described earlier happens approx
> 10% times).
>
> Would you try to answer once again? =)
>
>
>
>
>

Duh
Mon Sep 22 04:13:52 CDT 2003

Alexander Grigoriev wrote:
> Use remote debugging with WinBug. Run with DriverVerifier. Put debug
> messages into your driver to see how far it goes.
>
Well, it goes as far as it falls asleep ok. Where it goes when it AWAKES
- that is a kind of information I want to get, but I can't see it with
any tool for now. It is a nice suggestion to put a lot of trace
messages, but if I can't observe any of them - how can it help? I'm
ready to proceed with the kind of debugging one of my friend had
performed when he was designing a kind of protected mode operational
system. As the only resource he could have after just switching into
protected mode was the i/o ports - he debugged his code with beeping the
PC's built in dynamics: "Aha, one beep instead of two, that means we
hang somewhere near that instruction!". =)

A good emulation for PM procedures (not requiring to really suspend the
whole machine, just single network card' stack) would be very grateful,
but it does not exist, AFAIK, don't it?