Re: Start and stop via SCM not multiprocessor-safe?
From: Alexander Grigoriev (alegr_at_earthlink.net)
Date: 11/15/04
- Next message: Martin Borve [MSFT]: "Re: USB Video Class, How to Make 3 configuration"
- Previous message: Alexander Grigoriev: "Re: IoStatus.Status for IRP originator"
- In reply to: Calvin Guan: "Re: Start and stop via SCM not multiprocessor-safe?"
- Next in thread: Alexander Grigoriev: "Re: Start and stop via SCM not multiprocessor-safe?"
- Messages sorted by: [ date ] [ thread ]
Date: Mon, 15 Nov 2004 14:20:11 -0800
"Calvin Guan" <cguan@pleasenospams.ati.com> wrote in message
news:u18xhAtyEHA.2196@TK2MSFTNGP14.phx.gbl...
> Do you still see your driver object at the time it bugchecked? (kd>!object
> yourdriverobject) if yes, what was the pointer count and handle count of
> it?
> If you still see the object and the pointer is >0 (but not huge number
> such
> as 0xfffexxxx), then the driver shouldn't have been unloaded and the stack
> trace is probably bogus. If you don't see the object, then continue
> reading...
>
>> Still, I ask myself how the driver can be unloaded while I'm still
>> processing inside of DriverEntry(). :-(
>
> It should not. I was suspecting that the driverobject somehow was being
> accedentally dereferenced. I would do a BA on the DriverObject's pointer
> count right after DriverEntry was entered, and see if there's a bite.
>
> Good luck and happy hacking...
> --
> Calvin Guan Software Engineer/Radeon NT Drivers
> ATI Technologies Inc. Markham ON, Canada
> www.ati.com
>
>
- Next message: Martin Borve [MSFT]: "Re: USB Video Class, How to Make 3 configuration"
- Previous message: Alexander Grigoriev: "Re: IoStatus.Status for IRP originator"
- In reply to: Calvin Guan: "Re: Start and stop via SCM not multiprocessor-safe?"
- Next in thread: Alexander Grigoriev: "Re: Start and stop via SCM not multiprocessor-safe?"
- Messages sorted by: [ date ] [ thread ]
Relevant Pages
|