Re: Build device driver using Microsoft Visual C++ 2005
- From: mirage2k2 <mirage2k2@xxxxxxxxxxxxxxxxxxxxxxxxx>
- Date: Wed, 7 Jun 2006 08:15:02 -0700
Ok, I will be honest! I have never tried the WinDbg 2 machine thing ... just
read lots of reports about it being hard and slow, etc. I've also never
worked in the storage space!
Thanks for your advice ... I'll give it a go and see how I get on.
The only problems I've had with Softice is occasional incompatibilities with
keyboard, mouse or video drivers ... other than that none! For me Softice is
great because I implement some new functionality in my driver, test it, and
if something is wrong I just load the symbols, break into the code, step
through and find my bug! Then I break out, fix the bug, rebuild and then
restart my system and grab a coffee!
I find it very productive and very simple. Perhaps when I use WinDbg for a
while I'll realise that Softice is not that great ... or perhaps it is just a
case that one way of debugging suites my style more than the other.
Anyway, I was wrong for jumping to conclusions about 2 machine debugging
when I haven't actually tried it! Sorry! I use WinDbg for analyzing my
crash dumps and its great and ... I must admit ... the Softice command line
is horrible!
Mirage2k2
"Don Burn" wrote:
.
"mirage2k2" <mirage2k2@xxxxxxxxxxxxxxxxxxxxxxxxx> wrote in message
news:B4A6CF3D-D3C9-4801-ADEE-31DBB8BDE399@xxxxxxxxxxxxxxxx
So if I do not use SoftIce, how will I step through my code? I know that
you
can do this with WinDbg using 2 machines, but this is the hardest and
slowest
way of doing this! Our company simply cannot afford to waste so much time
using this ancient technique. Is there an alternative? Have MS updated
WinDbg so that it will do this? Do they have any plans to?
Why is this the hardest and slowest? The advantage of two machines is that
you are doing minimal disturbance of the test system with the debugger. The
other big advantage is that you are not risking corruption of the
development environment with the driver, this latter is critical if you work
in the storage space.
WinDBG on two machines was never slow compared to SoftIce, and nowadays with
1394 debugging it flies. I have never found 2 machine debugging harder, and
with the .kdfiles capability in WinDBG (where the debugger intercepts the
load of the driver, and installs one from the development machine on the
test machine) thing really fly. In a lot of cases, I will find a stupid bug
in WinDBG that caused a crash, fix them on source, and compile while the
test machine is rebooting. If you ever have one crash that requires a
reinstall, the cost of the second machine is probably covered in the wasted
time.
I really have to refute the slowest claim, I have been charging extra if a
customer forces me to use SoftIce since my productivity was so much worse.
--
Don Burn (MVP, Windows DDK)
Windows 2k/XP/2k3 Filesystem and Driver Consulting
http://www.windrvr.com
Remove StopSpam from the email to reply
- Follow-Ups:
- Re: Build device driver using Microsoft Visual C++ 2005
- From: Maxim S. Shatskih
- Re: Build device driver using Microsoft Visual C++ 2005
- References:
- Re: Build device driver using Microsoft Visual C++ 2005
- From: Gary G. Little
- Re: Build device driver using Microsoft Visual C++ 2005
- From: Gary G. Little
- Re: Build device driver using Microsoft Visual C++ 2005
- From: David J. Craig
- Re: Build device driver using Microsoft Visual C++ 2005
- From: mirage2k2
- Re: Build device driver using Microsoft Visual C++ 2005
- From: Don Burn
- Re: Build device driver using Microsoft Visual C++ 2005
- From: Don Burn
- Re: Build device driver using Microsoft Visual C++ 2005
- Prev by Date: Re: Parallel Port Driver
- Next by Date: Re: Parallel Port Driver
- Previous by thread: Re: Build device driver using Microsoft Visual C++ 2005
- Next by thread: Re: Build device driver using Microsoft Visual C++ 2005
- Index(es):
Relevant Pages
|