I tried this, and it actually had zero effect. The REGINFO was still
happening very consistantly, correctly the first time and too late after
having restarted the device.

This actually puzzled me, so I did some more tracking to try to determine
where the REGINFO gets triggered. The first time through, when it works
right, the REGINFO was triggered sometime during the IoCallDriver that
passes the START_DEVICE down the stack. Subsequent times when it fails, the
REGINFO doesn't trigger until after my START_DEVICE is completely finished.

I am wondering now if the IRP stack is somehow getting corrupted. I say
this because the first time, a lower driver down the stack apparantly kicks
the REGINFO into gear. This may just be coincidental timing of multiple
threads, but I doubt it because it happens the EXACT same way every time...

Honestly, I don't even know if it's the WMI thing that's really giving me
trouble - that's just the only thing that I can see happening differently.

So I still have no idea...


Thanks,

Josh Stone


"Maxim S. Shatskih" <maxim@storagecraft.com> wrote in message
news:umNYtQWUDHA.1776@TK2MSFTNGP09.phx.gbl...

For me, it was always as the following on XP SP1 OS:
- for non-boot driver, IoWMIRegistrationControl send MN_REGINFO or
MN_REGINFO_EX inline. The call stack shows this.
- for boot driver, MN_REGINFO or MN_REGINFO_EX is sent much later, when the
System start drivers are loading.

Max