> > Although from the looks of the rest of the thread it appear too bad. > A small ROM table is easily loaded. > I'm assuming from context that that is "... doesn't appear too bad".
Yes sorry. I did proof read but obviously not good enough.
> All this is just the "way it is" in video land right now.. > So, from a LinuxBIOS developers point of view, what would you like the card to provide? Would it be sufficient to have a raw frame buffer and a set of library routines that blt'd text onto it? This is your opportunity to perhaps change "the way it is" in video land. I for one would like to here the requirements from the LinuxBIOS team for their "dream" video card. No guarantees, but input is always good.
Well there are really 2 senarios here. The first is just having the necessary docs to turn the card on without having to resort to using a 16-bit emulator. As Tim says that should be no problem as init example code will be easy to get and open docs are, of course, the whole purpose of the project. That code can run after RAM is enabled and be somewhat complex although simpler is always better. The other is a forward looking type thing where we are trying to figure out how we are going to port to the next generation of motherboards that don't have a serial port. At POR we don't have any [1] RAM available and bringing up a RAM controller is hard work. Serial port debug messages are crucial. So we need some sort of dirt simple TTY output mode where all thats necessary to produce characters on the screen is OUT calls to an IO address. The idea of using PCI serial cards has also been proposed and we are looking into that. (Got to find a good source for some with linux drivers) [1] On some mainboards we have a really slick trick where we can use the onboard cache of the chip as early boot RAM. But it does not work on all motherboards. Good news is that we think it will work on all newer motherboards. -- Richard A. Smith _______________________________________________ Open-graphics mailing list [email protected] http://lists.duskglow.com/mailman/listinfo/open-graphics List service provided by Duskglow Consulting, LLC (www.duskglow.com)
