Extended TTY Modes, and User Interface


To allow greater flexibility on TTY lines (e.g. the ability to handle raw,
8-bit data), the existing stty()/gtty() interface must be extended, as it has
no spare bits for other modes. (The approach historically taken at MIT, of
adding a new system call, ttymod(), is being avoided.)

The extension uses over-laid semantics on the existing stty()/gtty()
interface; the use of the printing characters 'S' and 'C' in the 'erase
character' field of the stty() call. Since one would never normally use these
characters as the erase character, over-laying changed semantics on their use
seems acceptable.

When a 'S' is seen, that means this command is setting an extended TTY mode;
and when a 'C' is seen, that means it is clearing an extended mode. The name
of the extended mode bit/byte in question is a two-character constant carried
in the 'mode' field of the stty() call - this will allow up to 2^16 different
mode bit/bytes.

If the affected extended mode being set is a byte, the value is carried in
the 'kill character' field. 'Clearing' an extended mode byte is usually
undefined, but in some cases will return it to its default value.

Currently assigned values:

LI	Literal (i.e. 8-bit) input
LO	Literal (i.e. 8-bit) output
BR	Send break