speeding up development time with network drivers

NTDEV
1

I notice that each time I get a kernel fault, I am forced to go thru 2
reboots before I can get my changed kernel up and running. Is there any way
to reduce the same?. The 1st one is necessary… the machine comes up, I load
up my new driver thru the properties menu of a connection showed “Network
and dial-up connections”, and though it shows up that my driver is loaded, I
notice that the old driver is still in memory, *not* the new one, and so a
second reboot to bring the new one in. All this seems like an extermely long
cycle just to correct one single kernel fault. Any suggestions on cutting
the cycle time down?

Thanks
-Johnny

P.S. - Also if anyone needs clarification on my previous questions regarding
system call monitoring on Win2k (equiv of truss on solaris), please feel
free to email me.

Chat with friends online, try MSN Messenger: http://messenger.msn.com

2

Johnny

That’s the nice case. If the kernel fault is a boot up issue, you can
add a boot into Win’98 or safe boot just to remove your network driver
so that the normal boot can complete. With WDM, you have to let the PnP
manager do its thing. Oh for the good old days of NT4.

  1. Speed up your boots by loading as little as possible.
  2. Use a second machine to get something done while your sitting around
    waiting.

Michael S. Jackson
NetMotion Wireless, Inc.
xxxxx@nmwco.com

-----Original Message-----
From: Johnny D [mailto:xxxxx@hotmail.com]
Sent: Friday, March 29, 2002 12:48 PM
To: NT Developers Interest List
Subject: [ntdev] speeding up development time with network drivers

I notice that each time I get a kernel fault, I am forced to
go thru 2
reboots before I can get my changed kernel up and running. Is
there any way
to reduce the same?. The 1st one is necessary… the machine
comes up, I load
up my new driver thru the properties menu of a connection
showed “Network
and dial-up connections”, and though it shows up that my
driver is loaded, I
notice that the old driver is still in memory, *not* the new
one, and so a
second reboot to bring the new one in. All this seems like an
extermely long
cycle just to correct one single kernel fault. Any
suggestions on cutting
the cycle time down?

Thanks
-Johnny

P.S. - Also if anyone needs clarification on my previous
questions regarding
system call monitoring on Win2k (equiv of truss on solaris),
please feel
free to email me.

Chat with friends online, try MSN Messenger: http://messenger.msn.com

You are currently subscribed to ntdev as: xxxxx@nmwco.com
To unsubscribe send a blank email to %%email.unsub%%

3

Use Windbg with XP as Host and Target System.
Then make a Firewire Connection and use the Windbg feature called
‘Driver mapping’. This will inject a driver from your Host machine
into the target.
The cycle nows reduces to this:

  1. crash on Target. (Analysis with windbg
  2. changes in sources on Host,build new driver
  3. reboot target. (new driver form host will be loaded now).

| Norbert Kawulski | mailto:xxxxx@stollmann.de |
| Stollmann E+V GmbH, Development | http://www.stollmann.de |
–If it’s ISDN or Bluetooth, make sure it’s driven by Stollmann–

“No matter where you go, there you are”