Re: GGI, EGCS/PGCC, Kernel source

Theodore Y. Ts'o (tytso@MIT.EDU)
Fri, 27 Feb 1998 13:48:43 -0500

From: (Alan Cox)
Date: Fri, 27 Feb 1998 10:45:38 +0000 (GMT)

> routing all tty IO through EvStack, ever. Only serial lines that are
> somehow (?) dedicated as keyboard input devices feed into EvStack and
> bypass the tty line
> disc. layer directly. Jason even gave a pseudo patch later:

I assume what you are actually doing is making your GGI-Keyboard a
line discipline in itself like PPP, SLIP, tty - which is nice and clean

A line discpline in itself is *not* clean, since it would require
replicating the N_TTY code in their line discpline.

Having a filter function that gets called from within the N_TTY line
discipline (assuming that you would never want to use SLIP or PPP line
discplines on the console) is a little better, but there's still a
fundamental mismatch in that by the time you get into a tty line
discpline, you *know* what tty the input is destined for.

That is, a tty line discpline is for a specific line discpline.
However, if you're sending raw keycodes to the line discpline, you
haven't yet done the dispatch of which console tty should be receiving the
input (and which keycodes means "switch to VC #3", etc.). Doing the
dispatch of which tty the keycodes should go to within the tty line
discpline is very unclean.

It's not at all clear people are thinking through all of the
implications here --- possibly, they're not explaining it clearly or I'm
not understanding what they're trying to do, but I see some real
problems here.

- Ted

To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to