OSDev.org

hello,
i just recently got a book out from my college's library on computer architecture and it talks about protected mode. I was wondering, what can't you do in pmode that you can in real? the book doesn't go too far into it.

And also, how do you implrement a scheduler? I was considering it and it shuold be that hard, in essence:
watch users commands, if the user types a runable command, then setup the scheduler to switch back and forth between the running apps.

keep a list of the apps somewhere and assign them prorites based on something, im thinking allow the user to assign it (it would be different heh but probabl least affective)

then highest prority 1st, etc etc

all of this in a loop

how far am i off? lol

Thanks in advance

let's see... you asked many things that deserver a whole thread for themselves

1) Actually, the question is what can you do in pmode that you can't do in rmode? There are plenty of texts out there, but if you're in rmode you are confined to segmends no longer than 64k, no more than 1 MEG memory and single tasking. however, there are some "patches", like unreal mode or A20 to get around things like these.

you can start by knowing rmode, you can go into pmode then. (pmode is excellent, but as always, compatibility seems to be a big deal for customers).

2) Scheduler: You have to define several things here. like are you a RTOS? are you multiuser? you can use the simplest algorithm: round robin. Then you can add several privileges and then round-robin them.

Also, is the scheduler inside or outside the kernel? is the scheduler itself a task?

3) under a multiuser environment is nice to have the user "renice" its process so he can drop some privileges, keep that door open

these are quite big topics to discuss

lol, well i always thought, when you think, think big!! ;D

i think it would be fun to make your own scheduler from scratch, with no examples, then see how far off (or on!) you are with really good ones, heh

i just get giddy thinking of it!