Recent issues kernel-debugging over 1394 using Vista host
Ryan V. Bissell, Esquire
Although I'm not yet certain, it seems that some of the recent Windows
Updates for Vista kill the ability to debug in Vista-hosted WinDBG over
1394. Specifically, the "1394 Debugger Device" devices disappear from
Vista's Device Manager. (Manually recreating them with devcon.exe seemed to
resolve this for me the first time this happened to me, but not subsequent
times.)
This first happened to me last week. At the time I didn't mentally
corrolate the 1394 debugging issue with any Windows Update activity, and I
didn't have Restore Points enabled, so I eventually ended up re-installing
Vista Enterprise, and rebuilding my development system. All was good again
until this morning with I decided to catch up on all the Windows Updates
issued since Vista's launch. After reboot, 1394 debugging was again broken
exactly how it was last week.
This time, I had a Restore Point to rollback to, and rolling back to the
point immediately prior to my Windows Update activities this morning got me
debugging over 1394 again.
I can't yet conclusively claim that some particular Windows Update is the
culprit for this, but I intend to try when some of my schedule pressure
subsides.
Is this a known issue? I've checked recent posts to this newsgroup, and
also OSR's... googling yielded no help, either.
</ryan>
Ryan V. Bissell, Esquire
Courtesy follow-up: On this subject, I've been disabused of my Vista
Windows Updates theory, for two reasons:
* (1) Absolutely no one else has complained of this specific issue, nor
replied to this post.
* (2) Yesterday, while actively debugging from my (restored) Vista host over
1394, the target spontaneously stopped responding.[*] At that point in
time the 1394 kernel debugging pseudo-device was still present in the Vista
host's device manager. No other targets would connect on different channel
#s. Finally, I rebooted my Vista host, at which point the 1394 kernel
debugging pseudo-device went missing from the Vista host's device manager,
essentially reproducing my issue for the 3rd time. (Except, this time
there had been no intervening Windows Updates to cloud my perception of
cause.)
Since I've tried 3 different 1394 cards in this Vista host system, all with
the same result, I conclude that this is just somehow system-specific,
isolated to me, and now I'm simply going to regress to using XP for my
debugging host and hope that solves it. But I'm still confused as to why
this suddenly became an issue for me, after months of reliable use of
1394-kernel-WinDBG on this particular Vista host.
Hopefully this seemingly useless thread will be helpful to somebody,
someday.
</ryan>
[*] In case it matters to lurking WinDBG devs: This was WinDBG 6.7.5.1 (and
an older version that I don't recall.) The "active debugging" mentioned
above for this latest occurrence of 1394 debugging failure on my Vista host
was actually me just trying to repro a bug reported to occur after 100s of
reboots of the target system. Sometime after reboot 600, the target kept
rebooting as programmed but WinDBG on the Vista host was noted to just say
"Debuggee not connected." The last debug spew seen in the command window
was:
Shutdown occurred...unloading all symbol tables
Waiting to reconnect...
"1394 Windows Debug Driver(Kernel Mode)" was still in the Vista host's
device manager at that point-- listed as active and un-banged-- but
disappeared upon reboot of the Vista host.
"Ryan V. Bissell, Esquire" <ryan_i-wind...@bissell.org> wrote in
message news:325D29A9-8F6F-44EB...@microsoft.com...
Alberto
I have exactly the same problem as you. I was running an older version of
Windbg, I don't recall which, it had been working for months before on my
WinXP and it had worked before on a Beta version of Vista which I installed
on my system some time ago.
Well, I wiped out my old system and installed the Vista release fresh, and
guess what ? Windbg will not start. I installed 6.7.0005.1 hoping this was
something to do with the old version, same problem.
This is a solid, well tested Windbg install, which I have been running on XP
for several months without a glitch.
Alberto.
Alberto
I did some more investigation, and I could make it work. This is how:
The first time I open Windbg, I right click on the icon and left-click on
"Run as Administrator". I get yet another complaint about not being able to
install the driver, but now I can go on; and if I shutdown Windbg and just
run it again - without trying to run it as an Administrator, it works.
Problem is I was already an adminstrator on this machine, yet it didn't work
- go figure!
If I look now at my Device Manager, it says that the "1394 Debugger Device"
is now installed, and things are back to normal.
I hope this helps! I'm repeating this post on the other thread I posted my
question, hoping it will help the other people as well.
Alberto.
"Ryan V. Bissell, Esquire" wrote: