With some excellent suggestions from this group, I was able to develop a root-enumerated NDIS virtual network adapter with custom KMDF lower edge that process network packets to/from user mode (actually comprised of basically netvmini sample as its upper filter; thread has title “NDIS Miniport with KMDF lower edge” for reference). This has worked well, as we have been able to use this for further development on our project over the last year and a half.
The time has now come to run this through HLK testing for MS driver certification (we are targeting Windows 10 only). I have been running the driver through some preliminary HLK testing, with some success, and a few failures (64/114 passed, 7 failed so far). I have looked at the recommended LAN topology for the 2-station tests, and so my question is this:
Since the driver package moves network frames to/from user mode, and never in/out of “real” ethernet port, how do I prepare for the NDIS 2-station tests?
- There might be similarities to OpenVPN, and based on some pages on their forum, it appears that they may have received a waiver for some of these tests. Might that be an option?
- My driver supports a “cross-over” mode that will talk to another instance of this device - albeit this requires a custom application to support the I/O Completion ports and IOCTLs to facilitate the user-mode communication (which could be started at system start-up, I suppose?) I also know that I am able to put an instance of my device into a bridge configuration along with a “real” network device. Do I need to use that sort of configuration to communicate with the SupportDevice?
Any insight here is appreciated.
TIA!