win2k verifier causing system problems?

NTDEV
1

I’m currently testing a checked version of a PCI driver that only runs on
win2k and has never been run with driver verifier. When I setup driver
verifier to verify this (and only this) driver, everything gets wonky. Let
me be more specific: after the requisite reboot, it takes several login
attempts to get past the login prompt. By that I mean that I enter my UN/PW
and get the windows startup sound, but am immediately presented with the
dialogue that says “press ctrl+alt+del to login” again. No “bad password”
messages, I promise! This happens 3 - 12 times after which I am presented
with a partially loaded desktop (generally I have no icons or task bar, but
sometimes one or the other or both) and a bevy of “could not load” errors
mostly related to explorer.exe, and our anti-virus software, etc, etc, etc.
At this point, I usually cannot run any applications although sometimes
taskmanager will come up… This happened on both a free and
partially-checked version of win2k.

Disabling verifier.exe, deleting the driver.sys and associated .infs and
OEM#.infs in safe-mode did not restore functionality. I eventually had our
systems group copy a backup of the system disk from before using
verifier.exe onto this machine to get us back up and running.

My question is: have any of you seen this before? I think this a big hint
about a major problem in the driver that I just don’t understand. Everything
I’ve seen about verifier.exe is perfume and roses, so if you know of any
system integrity problems like this, I’d love to hear about it. We’re using
the 2k version of verifier, but 2k3 DDK for driver builds.

2

I would guess your test driver is a disk driver, and when you installed it
one of the drive letter assignments changed. For instance, C: might have
become D:. At this point Windows can’t find about 60% of the stuff it needs
to run, so very strange things happen, or mostly don’t happen.

Loren

3

Please do not include “Low resources simulation” in DV for all drivers :slight_smile:

This option causes each 1 of 5 (or so) ExAllocateXxx to fail :slight_smile: the only
exception is first several seconds after boot. It is only designed to test the
driver’s error handling path - so it will not BSOD on allocation failure.

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

----- Original Message -----
From: “Carl Denslow”
Newsgroups: ntdev
To: “Windows System Software Devs Interest List”
Sent: Monday, January 23, 2006 7:06 AM
Subject: [ntdev] win2k verifier causing system problems?

> I’m currently testing a checked version of a PCI driver that only runs on
> win2k and has never been run with driver verifier. When I setup driver
> verifier to verify this (and only this) driver, everything gets wonky. Let
> me be more specific: after the requisite reboot, it takes several login
> attempts to get past the login prompt. By that I mean that I enter my UN/PW
> and get the windows startup sound, but am immediately presented with the
> dialogue that says “press ctrl+alt+del to login” again. No “bad password”
> messages, I promise! This happens 3 - 12 times after which I am presented
> with a partially loaded desktop (generally I have no icons or task bar, but
> sometimes one or the other or both) and a bevy of “could not load” errors
> mostly related to explorer.exe, and our anti-virus software, etc, etc, etc.
> At this point, I usually cannot run any applications although sometimes
> taskmanager will come up… This happened on both a free and
> partially-checked version of win2k.
>
> Disabling verifier.exe, deleting the driver.sys and associated .infs and
> OEM#.infs in safe-mode did not restore functionality. I eventually had our
> systems group copy a backup of the system disk from before using
> verifier.exe onto this machine to get us back up and running.
>
> My question is: have any of you seen this before? I think this a big hint
> about a major problem in the driver that I just don’t understand. Everything
> I’ve seen about verifier.exe is perfume and roses, so if you know of any
> system integrity problems like this, I’d love to hear about it. We’re using
> the 2k version of verifier, but 2k3 DDK for driver builds.
>
>
>
> —
> Questions? First check the Kernel Driver FAQ at
http://www.osronline.com/article.cfm?id=256
>
> You are currently subscribed to ntdev as: xxxxx@storagecraft.com
> To unsubscribe send a blank email to xxxxx@lists.osr.com

4

> I would guess your test driver is a disk driver, and when you installed it

one of the drive letter assignments changed. For instance, C: might have
become D:. At this point Windows can’t find about 60% of the stuff it needs
to run, so very strange things happen, or mostly don’t happen.

In this case, you will need to boot another copy of Windows from another disk
on this machine, and manually patch the MountedDevices registry in the ill copy
of Windows.

BTW - w2k never allowed be even to logon to desktop if this problem occured.

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

5

Is there any debug spew into WinDbg, or have you used WinDbg?

Gary G. Little

-----Original Message-----
From: xxxxx@lists.osr.com
[mailto:xxxxx@lists.osr.com] On Behalf Of Carl Denslow
Sent: Sunday, January 22, 2006 10:06 PM
To: Windows System Software Devs Interest List
Subject: [ntdev] win2k verifier causing system problems?

I’m currently testing a checked version of a PCI driver that only runs on
win2k and has never been run with driver verifier. When I setup driver
verifier to verify this (and only this) driver, everything gets wonky.
Let
me be more specific: after the requisite reboot, it takes several login
attempts to get past the login prompt. By that I mean that I enter my
UN/PW
and get the windows startup sound, but am immediately presented with the
dialogue that says “press ctrl+alt+del to login” again. No “bad password”

messages, I promise! This happens 3 - 12 times after which I am presented

with a partially loaded desktop (generally I have no icons or task bar,
but
sometimes one or the other or both) and a bevy of “could not load” errors
mostly related to explorer.exe, and our anti-virus software, etc, etc,
etc.
At this point, I usually cannot run any applications although sometimes
taskmanager will come up… This happened on both a free and
partially-checked version of win2k.

Disabling verifier.exe, deleting the driver.sys and associated .infs and
OEM#.infs in safe-mode did not restore functionality. I eventually had
our
systems group copy a backup of the system disk from before using
verifier.exe onto this machine to get us back up and running.

My question is: have any of you seen this before? I think this a big hint

about a major problem in the driver that I just don’t understand.
Everything
I’ve seen about verifier.exe is perfume and roses, so if you know of any
system integrity problems like this, I’d love to hear about it. We’re
using
the 2k version of verifier, but 2k3 DDK for driver builds.

Questions? First check the Kernel Driver FAQ at
http://www.osronline.com/article.cfm?id=256

You are currently subscribed to ntdev as: xxxxx@seagate.com
To unsubscribe send a blank email to xxxxx@lists.osr.com

6

Actually, this is not a disk driver. It is a custom PCI card where all
functions and registers are accessible through plain memory reads and
writes.

“Loren Wilton” wrote in message news:xxxxx@ntdev…
>I would guess your test driver is a disk driver, and when you installed it
> one of the drive letter assignments changed. For instance, C: might have
> become D:. At this point Windows can’t find about 60% of the stuff it
> needs
> to run, so very strange things happen, or mostly don’t happen.
>
> Loren
>
>
>

7

I agree. I am not using the low resources option.

“Maxim S. Shatskih” wrote in message
news:xxxxx@ntdev…
> Please do not include “Low resources simulation” in DV for all drivers
> :slight_smile:
>
> This option causes each 1 of 5 (or so) ExAllocateXxx to fail :slight_smile: the
> only
> exception is first several seconds after boot. It is only designed to test
> the
> driver’s error handling path - so it will not BSOD on allocation failure.
>
> Maxim Shatskih, Windows DDK MVP
> StorageCraft Corporation
> xxxxx@storagecraft.com
> http://www.storagecraft.com
>
> ----- Original Message -----
> From: “Carl Denslow”
> Newsgroups: ntdev
> To: “Windows System Software Devs Interest List”
> Sent: Monday, January 23, 2006 7:06 AM
> Subject: [ntdev] win2k verifier causing system problems?
>
>
>> I’m currently testing a checked version of a PCI driver that only runs on
>> win2k and has never been run with driver verifier. When I setup driver
>> verifier to verify this (and only this) driver, everything gets wonky.
>> Let
>> me be more specific: after the requisite reboot, it takes several login
>> attempts to get past the login prompt. By that I mean that I enter my
>> UN/PW
>> and get the windows startup sound, but am immediately presented with the
>> dialogue that says “press ctrl+alt+del to login” again. No “bad
>> password”
>> messages, I promise! This happens 3 - 12 times after which I am
>> presented
>> with a partially loaded desktop (generally I have no icons or task bar,
>> but
>> sometimes one or the other or both) and a bevy of “could not load” errors
>> mostly related to explorer.exe, and our anti-virus software, etc, etc,
>> etc.
>> At this point, I usually cannot run any applications although sometimes
>> taskmanager will come up… This happened on both a free and
>> partially-checked version of win2k.
>>
>> Disabling verifier.exe, deleting the driver.sys and associated .infs and
>> OEM#.infs in safe-mode did not restore functionality. I eventually had
>> our
>> systems group copy a backup of the system disk from before using
>> verifier.exe onto this machine to get us back up and running.
>>
>> My question is: have any of you seen this before? I think this a big
>> hint
>> about a major problem in the driver that I just don’t understand.
>> Everything
>> I’ve seen about verifier.exe is perfume and roses, so if you know of any
>> system integrity problems like this, I’d love to hear about it. We’re
>> using
>> the 2k version of verifier, but 2k3 DDK for driver builds.
>>
>>
>>
>> —
>> Questions? First check the Kernel Driver FAQ at
> http://www.osronline.com/article.cfm?id=256
>>
>> You are currently subscribed to ntdev as: xxxxx@storagecraft.com
>> To unsubscribe send a blank email to xxxxx@lists.osr.com
>
>
>

8

Excellent question. I am testing on a free system without a debugger
connected. I was told by the developer (when he saw the same symptoms on
the checked+debugger rig) that the only difference in windbg output was that
there was much less, owing to the fact that he was not able to exercise the
driver because he could not run any apps. Throughout the loading and
initialization, it was the same output.

The thing that bothers me the most about this problem is that after booting
in safe mode to disable verifer and remove the driver (.sys, .inf and
OEM#.inf) and rebooting, the system problems persisted. This says “memory
corruption” to me. I have a hard time with both of my two obvious choices:

  1. The driver is walking on system memory without windbg catching it. 2)
    Verifer (win2k) is walking on system memory without anyone having posted
    information about this previously.

wrote in message news:xxxxx@ntdev…
> Is there any debug spew into WinDbg, or have you used WinDbg?
>
> Gary G. Little
>
> -----Original Message-----
> From: xxxxx@lists.osr.com
> [mailto:xxxxx@lists.osr.com] On Behalf Of Carl Denslow
> Sent: Sunday, January 22, 2006 10:06 PM
> To: Windows System Software Devs Interest List
> Subject: [ntdev] win2k verifier causing system problems?
>
> I’m currently testing a checked version of a PCI driver that only runs on
> win2k and has never been run with driver verifier. When I setup driver
> verifier to verify this (and only this) driver, everything gets wonky.
> Let
> me be more specific: after the requisite reboot, it takes several login
> attempts to get past the login prompt. By that I mean that I enter my
> UN/PW
> and get the windows startup sound, but am immediately presented with the
> dialogue that says “press ctrl+alt+del to login” again. No “bad password”
>
> messages, I promise! This happens 3 - 12 times after which I am presented
>
> with a partially loaded desktop (generally I have no icons or task bar,
> but
> sometimes one or the other or both) and a bevy of “could not load” errors
> mostly related to explorer.exe, and our anti-virus software, etc, etc,
> etc.
> At this point, I usually cannot run any applications although sometimes
> taskmanager will come up… This happened on both a free and
> partially-checked version of win2k.
>
> Disabling verifier.exe, deleting the driver.sys and associated .infs and
> OEM#.infs in safe-mode did not restore functionality. I eventually had
> our
> systems group copy a backup of the system disk from before using
> verifier.exe onto this machine to get us back up and running.
>
> My question is: have any of you seen this before? I think this a big hint
>
> about a major problem in the driver that I just don’t understand.
> Everything
> I’ve seen about verifier.exe is perfume and roses, so if you know of any
> system integrity problems like this, I’d love to hear about it. We’re
> using
> the 2k version of verifier, but 2k3 DDK for driver builds.
>
>
>
> —
> Questions? First check the Kernel Driver FAQ at
> http://www.osronline.com/article.cfm?id=256
>
> You are currently subscribed to ntdev as: xxxxx@seagate.com
> To unsubscribe send a blank email to xxxxx@lists.osr.com
>
>

9

While not impossible, I doubt verifier is affecting the system after
disabling an rebooting. I assume that during the removal of the driver files
that you also pulled the card, or did you leave it in the peripheral slot?

The personal opinion of
Gary G. Little

-----Original Message-----
From: xxxxx@lists.osr.com
[mailto:xxxxx@lists.osr.com] On Behalf Of Carl Denslow
Sent: Monday, January 23, 2006 7:52 PM
To: Windows System Software Devs Interest List
Subject: Re:[ntdev] win2k verifier causing system problems?

Excellent question. I am testing on a free system without a debugger
connected. I was told by the developer (when he saw the same symptoms on
the checked+debugger rig) that the only difference in windbg output was that
there was much less, owing to the fact that he was not able to exercise the
driver because he could not run any apps. Throughout the loading and
initialization, it was the same output.

The thing that bothers me the most about this problem is that after booting
in safe mode to disable verifer and remove the driver (.sys, .inf and
OEM#.inf) and rebooting, the system problems persisted. This says “memory
corruption” to me. I have a hard time with both of my two obvious choices:

  1. The driver is walking on system memory without windbg catching it. 2)
    Verifer (win2k) is walking on system memory without anyone having posted
    information about this previously.

wrote in message news:xxxxx@ntdev…
> Is there any debug spew into WinDbg, or have you used WinDbg?
>
> Gary G. Little
>
> -----Original Message-----
> From: xxxxx@lists.osr.com
> [mailto:xxxxx@lists.osr.com] On Behalf Of Carl Denslow
> Sent: Sunday, January 22, 2006 10:06 PM
> To: Windows System Software Devs Interest List
> Subject: [ntdev] win2k verifier causing system problems?
>
> I’m currently testing a checked version of a PCI driver that only runs
> on win2k and has never been run with driver verifier. When I setup
> driver verifier to verify this (and only this) driver, everything gets
wonky.
> Let
> me be more specific: after the requisite reboot, it takes several
> login attempts to get past the login prompt. By that I mean that I
> enter my UN/PW and get the windows startup sound, but am immediately
> presented with the dialogue that says “press ctrl+alt+del to login”
> again. No “bad password”
>
> messages, I promise! This happens 3 - 12 times after which I am
> presented
>
> with a partially loaded desktop (generally I have no icons or task
> bar, but sometimes one or the other or both) and a bevy of “could not
> load” errors mostly related to explorer.exe, and our anti-virus
> software, etc, etc, etc.
> At this point, I usually cannot run any applications although
> sometimes taskmanager will come up… This happened on both a free
> and partially-checked version of win2k.
>
> Disabling verifier.exe, deleting the driver.sys and associated .infs
> and OEM#.infs in safe-mode did not restore functionality. I
> eventually had our systems group copy a backup of the system disk from
> before using verifier.exe onto this machine to get us back up and
> running.
>
> My question is: have any of you seen this before? I think this a big
> hint
>
> about a major problem in the driver that I just don’t understand.
> Everything
> I’ve seen about verifier.exe is perfume and roses, so if you know of
> any system integrity problems like this, I’d love to hear about it.
> We’re using the 2k version of verifier, but 2k3 DDK for driver builds.
>
>
>
> —
> Questions? First check the Kernel Driver FAQ at
> http://www.osronline.com/article.cfm?id=256
>
> You are currently subscribed to ntdev as: xxxxx@seagate.com To
> unsubscribe send a blank email to xxxxx@lists.osr.com
>
>

—
Questions? First check the Kernel Driver FAQ at
http://www.osronline.com/article.cfm?id=256

You are currently subscribed to ntdev as: glittle@mn.rr.com To unsubscribe
send a blank email to xxxxx@lists.osr.com

NOD32 1.1376 (20060123) Information

This message was checked by NOD32 antivirus system.
http://www.eset.com

10

>The thing that bothers me the most about this problem is that after booting

in safe mode to disable verifer and remove the driver (.sys, .inf and
OEM#.inf) and rebooting, the system problems persisted. This says “memory
corruption” to me. I have a hard time with both of my two obvious choices:

There is a third choice: something in installing the driver changed something about the system configuration. I really can’t offhand think what it would be, and trying to diff the registry before and after the install would be an act of futile insanity if you didn’t have a real good idea where to look to start with.

I’d still almost be inclined to the concept of a drive letter change. But as Maxim points out, that generally means you can *never* log onto the partition again, not just “mostly can’t”. So it doesn’t seem to be quite that drastic. Perhaps a directory changed name? That seems rather unlikely; I’d mostly expect a crash or outright boot failure from that. Perhaps somehow a whole scad of inf files disappeared? Again, I’d expect that would be consistent.

You’re absolutely certain that your driver is getting removed? Are you absolutely certain (I bet not) that all traces of it are being removed from the registry? Perhaps there is something there that is causing other driver loads to fail, even if your driver isn’t loaded. (And no, I have no idea what might do that. Load order group changes maybe?)

Loren

11

I removed the card, then removed the files.

After looking it over, my first post was a bit misleading. Let me clarify:

> Disabling verifier.exe, deleting the driver.sys and associated .infs
> and OEM#.infs in safe-mode did not restore functionality.
After removing the device and removing all of the driver files in safemode I
WAS able to log in to non-safemode properly. After I reinstalled the
driver, but BEFORE re-attaching verifier, all of the system problems came
back. So there was a step in the middle where the machine seemed to behave
properly, but that was without the test driver (or device) installed at all.

“Gary G. Little” wrote in message news:xxxxx@ntdev…
> While not impossible, I doubt verifier is affecting the system after
> disabling an rebooting. I assume that during the removal of the driver
> files
> that you also pulled the card, or did you leave it in the peripheral slot?
>
> The personal opinion of
> Gary G. Little
>
> -----Original Message-----
> From: xxxxx@lists.osr.com
> [mailto:xxxxx@lists.osr.com] On Behalf Of Carl Denslow
> Sent: Monday, January 23, 2006 7:52 PM
> To: Windows System Software Devs Interest List
> Subject: Re:[ntdev] win2k verifier causing system problems?
>
> Excellent question. I am testing on a free system without a debugger
> connected. I was told by the developer (when he saw the same symptoms on
> the checked+debugger rig) that the only difference in windbg output was
> that
> there was much less, owing to the fact that he was not able to exercise
> the
> driver because he could not run any apps. Throughout the loading and
> initialization, it was the same output.
>
> The thing that bothers me the most about this problem is that after
> booting
> in safe mode to disable verifer and remove the driver (.sys, .inf and
> OEM#.inf) and rebooting, the system problems persisted. This says “memory
> corruption” to me. I have a hard time with both of my two obvious
> choices:
> 1) The driver is walking on system memory without windbg catching it. 2)
> Verifer (win2k) is walking on system memory without anyone having posted
> information about this previously.
>
>
> wrote in message news:xxxxx@ntdev…
>> Is there any debug spew into WinDbg, or have you used WinDbg?
>>
>> Gary G. Little
>>
>> -----Original Message-----
>> From: xxxxx@lists.osr.com
>> [mailto:xxxxx@lists.osr.com] On Behalf Of Carl Denslow
>> Sent: Sunday, January 22, 2006 10:06 PM
>> To: Windows System Software Devs Interest List
>> Subject: [ntdev] win2k verifier causing system problems?
>>
>> I’m currently testing a checked version of a PCI driver that only runs
>> on win2k and has never been run with driver verifier. When I setup
>> driver verifier to verify this (and only this) driver, everything gets
> wonky.
>> Let
>> me be more specific: after the requisite reboot, it takes several
>> login attempts to get past the login prompt. By that I mean that I
>> enter my UN/PW and get the windows startup sound, but am immediately
>> presented with the dialogue that says “press ctrl+alt+del to login”
>> again. No “bad password”
>>
>> messages, I promise! This happens 3 - 12 times after which I am
>> presented
>>
>> with a partially loaded desktop (generally I have no icons or task
>> bar, but sometimes one or the other or both) and a bevy of “could not
>> load” errors mostly related to explorer.exe, and our anti-virus
>> software, etc, etc, etc.
>> At this point, I usually cannot run any applications although
>> sometimes taskmanager will come up… This happened on both a free
>> and partially-checked version of win2k.
>>
>> Disabling verifier.exe, deleting the driver.sys and associated .infs
>> and OEM#.infs in safe-mode did not restore functionality. I
>> eventually had our systems group copy a backup of the system disk from
>> before using verifier.exe onto this machine to get us back up and
>> running.
>>
>> My question is: have any of you seen this before? I think this a big
>> hint
>>
>> about a major problem in the driver that I just don’t understand.
>> Everything
>> I’ve seen about verifier.exe is perfume and roses, so if you know of
>> any system integrity problems like this, I’d love to hear about it.
>> We’re using the 2k version of verifier, but 2k3 DDK for driver builds.
>>
>>
>>
>> —
>> Questions? First check the Kernel Driver FAQ at
>> http://www.osronline.com/article.cfm?id=256
>>
>> You are currently subscribed to ntdev as: xxxxx@seagate.com To
>> unsubscribe send a blank email to xxxxx@lists.osr.com
>>
>>
>
>
>
> —
> Questions? First check the Kernel Driver FAQ at
> http://www.osronline.com/article.cfm?id=256
>
> You are currently subscribed to ntdev as: glittle@mn.rr.com To unsubscribe
> send a blank email to xxxxx@lists.osr.com
>
> NOD32 1.1376 (20060123) Information
>
> This message was checked by NOD32 antivirus system.
> http://www.eset.com
>
>
>
>

12

When we’re talking about the magical forest of the registry, I am not ever
certain!

Clearly, all traces of the driver are not being removed since I can see that
problems persist over uninstall, reboot, reinstall cycles. (Also, please
see my clarification in response to G.G.Little)

I kind of like the idea that the combination of ill-behaved test driver +
verifier is breaking something outside of the test driver. That seems like
a reasonable explanation. I’m not opposed to throwing windiff at the
registry and seeing what it comes up with. Any suggestions what kinds of
mischief the test driver could do that windbg / verifier would not overtly
complain about? Also, perhaps I’m debugging the wrong driver!

“Loren Wilton” wrote in message news:xxxxx@ntdev…
> >The thing that bothers me the most about this problem is that after
> >booting
>>in safe mode to disable verifer and remove the driver (.sys, .inf and
>>OEM#.inf) and rebooting, the system problems persisted. This says “memory
>>corruption” to me. I have a hard time with both of my two obvious
>>choices:
>
> There is a third choice: something in installing the driver changed
> something about the system configuration. I really can’t offhand think
> what it would be, and trying to diff the registry before and after the
> install would be an act of futile insanity if you didn’t have a real good
> idea where to look to start with.
>
> I’d still almost be inclined to the concept of a drive letter change. But
> as Maxim points out, that generally means you can never log onto the
> partition again, not just “mostly can’t”. So it doesn’t seem to be quite
> that drastic. Perhaps a directory changed name? That seems rather
> unlikely; I’d mostly expect a crash or outright boot failure from that.
> Perhaps somehow a whole scad of inf files disappeared? Again, I’d expect
> that would be consistent.
>
> You’re absolutely certain that your driver is getting removed? Are you
> absolutely certain (I bet not) that all traces of it are being removed
> from the registry? Perhaps there is something there that is causing other
> driver loads to fail, even if your driver isn’t loaded. (And no, I have
> no idea what might do that. Load order group changes maybe?)
>
> Loren
>
>
>

13

When inserting your card, did you move the disk controller’s card to
another slot?

d

-----Original Message-----
From: xxxxx@lists.osr.com
[mailto:xxxxx@lists.osr.com] On Behalf Of Carl Denslow
Sent: Monday, January 23, 2006 5:52 PM
To: Windows System Software Devs Interest List
Subject: Re:[ntdev] win2k verifier causing system problems?

Excellent question. I am testing on a free system without a debugger
connected. I was told by the developer (when he saw the same symptoms
on
the checked+debugger rig) that the only difference in windbg output was
that
there was much less, owing to the fact that he was not able to exercise
the
driver because he could not run any apps. Throughout the loading and
initialization, it was the same output.

The thing that bothers me the most about this problem is that after
booting
in safe mode to disable verifer and remove the driver (.sys, .inf and
OEM#.inf) and rebooting, the system problems persisted. This says
“memory
corruption” to me. I have a hard time with both of my two obvious
choices:

  1. The driver is walking on system memory without windbg catching it.

Verifer (win2k) is walking on system memory without anyone having posted

information about this previously.

wrote in message news:xxxxx@ntdev…
> Is there any debug spew into WinDbg, or have you used WinDbg?
>
> Gary G. Little
>
> -----Original Message-----
> From: xxxxx@lists.osr.com
> [mailto:xxxxx@lists.osr.com] On Behalf Of Carl Denslow
> Sent: Sunday, January 22, 2006 10:06 PM
> To: Windows System Software Devs Interest List
> Subject: [ntdev] win2k verifier causing system problems?
>
> I’m currently testing a checked version of a PCI driver that only runs
on
> win2k and has never been run with driver verifier. When I setup
driver
> verifier to verify this (and only this) driver, everything gets wonky.
> Let
> me be more specific: after the requisite reboot, it takes several
login
> attempts to get past the login prompt. By that I mean that I enter my
> UN/PW
> and get the windows startup sound, but am immediately presented with
the
> dialogue that says “press ctrl+alt+del to login” again. No “bad
password”
>
> messages, I promise! This happens 3 - 12 times after which I am
presented
>
> with a partially loaded desktop (generally I have no icons or task
bar,
> but
> sometimes one or the other or both) and a bevy of “could not load”
errors
> mostly related to explorer.exe, and our anti-virus software, etc, etc,
> etc.
> At this point, I usually cannot run any applications although
sometimes
> taskmanager will come up… This happened on both a free and
> partially-checked version of win2k.
>
> Disabling verifier.exe, deleting the driver.sys and associated .infs
and
> OEM#.infs in safe-mode did not restore functionality. I eventually
had
> our
> systems group copy a backup of the system disk from before using
> verifier.exe onto this machine to get us back up and running.
>
> My question is: have any of you seen this before? I think this a big
hint
>
> about a major problem in the driver that I just don’t understand.
> Everything
> I’ve seen about verifier.exe is perfume and roses, so if you know of
any
> system integrity problems like this, I’d love to hear about it. We’re
> using
> the 2k version of verifier, but 2k3 DDK for driver builds.
>
>
>
> —
> Questions? First check the Kernel Driver FAQ at
> http://www.osronline.com/article.cfm?id=256
>
> You are currently subscribed to ntdev as: xxxxx@seagate.com
> To unsubscribe send a blank email to xxxxx@lists.osr.com
>
>

—
Questions? First check the Kernel Driver FAQ at
http://www.osronline.com/article.cfm?id=256

You are currently subscribed to ntdev as: xxxxx@microsoft.com
To unsubscribe send a blank email to xxxxx@lists.osr.com