> I think it'd be fine if we used a piece of real RAM for it (hey, it's
> not that much. Besides it could be made a compile-time option or
> something like that if you wanna spare a few K's).
The recent traffic said that the screen gets scrolled by adjusting the
VGA's 'start of display memory' register, and that the kernel has no
clue as to what's scrollback and what's garbage. I'm not sure about the
"no clue" bit, as the original statement was broader but doesn't make
sense in retrospect. I see two solutions: one could split the vga
memory among the various VC's, or one could copy out and copy back in
the whole text memory section on VC switch, in addition to switching
which VC data structure the kernel looks at. The first would cut down
on the scrollback for individual VC's but would actually save memory, as
saving the current start-of-screen address takes much less space than
saving the screen itself. The second would be a serious memory pig, and
ought to be chucked into user space somehow. Alternately, we could
merge the two ideas, reserving tty1 for all the console spam and
splitting the rest of the memory among the rest of the VC's, but getting
the messages redirected to tty1, yet noticed (duplicated?) makes this
one even more ambitious than the other two.
> Comments? Is it possible?
Possible, yes. Possible to do what you want, maybe. Possible to do
what you want and make everyone happy, maybe. Now, go code it. :-)
Keith
-- "It moved faster. I swear, they are evolving right before my eyes. If you see something this big, with eight legs coming your way, let me know; I have to kill it before it develops language skills." --- Ambassador Londo Mollari, in 'Sic Transit Vir' (Babylon 5)