Do you LOSE page table entries as time goes by, or are you just seeing an
extra PTE being used that you otherwise can’t account for? In either case,
you are properly undoing you call, right? Calling FreeAdapterChannel() or
whatever it is?
Yes, PTEs are consumed and never returned. In PerfMon I can see them drop
until the system starts squealing for resources.
I guess the next natural question is: If you turn off Verifier, do you still
see the extra PTE consumed?
Hmmm, gee hadn’t thought about that. I’ll test that tomorrow. That’s a good
call since 2000
Without the verifier active doesn’t show the symptom. I know there is one
level of IO checking
I can’t use. My code is based on Walter Oney’s and if I have it set for that
level of IO, I get the
Verifiers object instead of my own. That one is on a low priority burner for
fixing.
Maybe one of our Microsoft colleagues can tell us if Verifier can consume a
PTE during a call to AllocateAdapterChannel()? I won’t be able to check the
source code to confirm this until next week…
Peter
You are currently subscribed to ntdev as: xxxxx@delphieng.com
To unsubscribe send a blank email to leave-ntdev-$subst(‘Recip.MemberIDChar’)@lists.osr.com
You are currently subscribed to ntdev as: $subst(‘Recip.EmailAddr’)
To unsubscribe send a blank email to leave-ntdev-$subst(‘Recip.MemberIDChar’)@lists.osr.com