Hi!
I’m trying to debug an IFS driver that unfortunately keeps wanting to do a paging write after MJ_CLOSE. I can see that there are several (many!) IRP_MJ_CREATE calls that are all handled correctly, at least inasmuch as I can tell. So for one of the create calls that succeeded I can see the MJ_CLEANUP call, but before the lazy writer sends an IRP_MJ_WRITE I see the IRP_MJ_CLOSE happen. I don’t see myself going through the purge cache path in the cleanup routine, by the way. I’m thinking this might be a reference counting bug. Any ideas are most welcome.
Thanks!
Steve