pass flags from device close to l_close

kernel reorg

new copyright notice

ttysleep to catch revoked terminals

set WOPEN on first open

check for revoked terminal (ttclosed)

new tsleep; unwind when signal interrupts; close routines must return values

merge in vnodes

speed table, initializations, t_param, other tty related changes

../machine => machine

from Keith Muller;
>>> fixed problem with using up all events. should check that this is a
>>> event, not just that one is pending. Cleanup up code that was using
>>> only partially cleaned events. P

from Keith Muller;
>>> fixed problem with using up all events. should check that this is a
>>> event, not just that one is pending. Cleanup up code that was using
>>> only partially cleaned events. Ports still wedge, but can now be
>>> cleaned without a reboot.
>>>
>>> close off a race condition in open when background jobs exit
>>>
>>> add flush code to mpstop. matches a change to mpccdl firmware.
>>>
>>> missing \n from a log call
>>>
>>> repairs races with IOCTL mp events caused by rapid ioctl calls on loaded
>>> mp controllers. Fixes problem with delays not being processed. Fixes
>>> problem that a stty 0 was setting the baud rate to 0 (a new form of
>>> screen lock????) rather than dropping the line as specified in tty (4).

show more ...

add CCI credit, rm unneeded wakup; how can 2 closes
be in progress at once on a line?

mpcc ports wedge; bug report 4.3BSD/sys/217

mpcc ports on CCI can wedge; bug report 4.3BSD/sys/216

install approved copyright notice

mpcc ports hang with TS_BUSY; bug report 4.3BSD-tahoe/sys/23

bring up to revision 7 for tahoe release

lint

don't know what reg points to, but it's not an mblok

From: muller%sdcc7@ucsd.edu (Keith Muller)
We had problems with many terminals that only do ^S/^Q flow control. They
were losing characters. The time to process a ^S was at times 240 characters
(24

From: muller%sdcc7@ucsd.edu (Keith Muller)
We had problems with many terminals that only do ^S/^Q flow control. They
were losing characters. The time to process a ^S was at times 240 characters
(240 characters were transmitted at 9600 baud after a ^S was sent). The
problem was that the mpcc firmware would support ^S/^Q flow control, but the
driver only used it in tandem mode.

show more ...

From: muller%sdcc7@ucsd.edu (Keith Muller)
Tracked down the bizzarre error we were getting on the console from the
mp driver. It was due to a missing case in a switch statement. Turns out
it was repo

From: muller%sdcc7@ucsd.edu (Keith Muller)
Tracked down the bizzarre error we were getting on the console from the
mp driver. It was due to a missing case in a switch statement. Turns out
it was reporting that a port saw the BREAK key pressed. It also had a
overzealous logging of frame and parity errors. In actual use with
lots of ports this a bit much. Since none of the other tty drivers log
every break, framing error or parity error I made the following change:

show more ...

need reset (maybe)

date and time created 87/11/17 20:26:20 by sam