I have need for a virtual keyboard driver and set upon a mission to create one. I have started with a model from GITHUB koharubiyori (user) (virtualinput) is there project. and have been using that to a) understand the code as much as possible b) learn WINDBG. The thing is I cannot get the client app + driver to actually show typing by the OS. I have spent hours going through the code and everything I read and check suggests it should work - but it isn't.
But...when I am stepping through the code in WINDBG to see the queues and IRP etc. then return to the VM. And the VM has gone to the WINLOGON screen, I can see the key-presses (typing) is showing up in the WINLOGON username input field.
I am completely stuck!!! Any advice would be greatly appreciated.
Immediately after riting this I worked it out.
Turned off enhanced mode in hyper-v.
Tried again...it worked...
I have spent hours working out the code to understand it to debug it.
And in the end it was hyper v
I could cry.
Are you trying to inject input BEFORE logging in? Client apps don't run before that point. Or are you saying you were in an up and running system, but then it logged out? The login screen runs in a separate windows "session". Perhaps you should describe the sequence of events in a bit more detail.
In most cases, it's much easier to use the SendInputAPI to do simulated keyboard and mouse input instead of going to the trouble of installing a kernel driver.
Thank you for your response. On whim, after posting the message, I cut and pasted the question into Perplexity (no affiliation) and it pointed me towards my - debug - VM running in enhanced mode. Once I turned enhanced mode off, things started working as I wanted.
Although I spent hours debugging code that was actually working, the good news is that, rather than obtaining code from GITHUB and running it without understanding it, I feel I more fully understand the code, and and have learned lots about using WINDBG in the process.
Thanks again for your response. R