Re: NanoX version 0.3 released

[Eric J. Korpela]

>       I think that's a good idea to potentially solve the lock problem,
> but of course lock code would have to be executed.  At this point, it's going
> to be *way* too slow for all the graphics primitives to go thru a kernel mode
> transition, however. 

I was thinking of only locking the initial open of the device driver,
and providing an ioctl that goes directly to "int 10" after it's open.
Given how badly video BIOS is generally written, it wouldn't add much overhead.
At any rate, once you have successfully opened the device, you should feel
free to do anything including writing directly to video memory.  If the open
didn't succeed, then you shouldn't do anything.

Eric