I think Definition of CRTSCTS in /sys/sys/termios.h is not right. It's now

#define CCTS_OFLOW      0x00010000      /* CTS flow control of output */
#define CRTSCTS         CCTS_OFLOW      /* ??? */
#define CRTS_IFLOW      0x00020000      /* RTS flow control of input */

But I think it should be

#define CCTS_OFLOW      0x00010000      /* CTS flow control of output */
#define CRTS_IFLOW      0x00020000      /* RTS flow control of input */
#define CRTSCTS         (CRTS_IFLOW|CCTS_OFLOW) /* RTS/CTS flow control */

or something. Am i wrong ? I enabled IER_EMSC in my patched driver ,
so Modem status interrupt can happen. I do 'stty crtsrts', then output
flow controlcan work(partly). And CRTS_IFLOW is used in /sys/kern/tty.c,
if Definition of CRTSCTS corrected and /bin/stty recompiled, input
flow control will work too. Why not? I hacked stty code and enabled
CRTS_IFLOW too when 'stty crtscts'. 

> BTW Is anyone working on a replacement driver - like one that is fast and
> allows bidir stuff ? Not just hack to com.c ?

 I have hacked com.c and got bidirectional com port. This is necessary
for me because My Dynabook SX001(386sx/16MHz) have only one com port.

 One more, I am japanese and using KANJI(requiers 8 bits through) from
com port. I set 'np' in /etc/gettytab, then getty set serial port to 
CS7 and AP. for this, I must login in 7 bits and even parity and then
must do stty pass8. It's not good. I changed getty to set CS8 if 'np'
set. But right way is to make getty to understand 'p8' flag. Is this
wrong? If this is right way, then I will change getty that way and
post patch. Any other ideas?
--
Koji Imada - je4owb/2
koji@math.nagoya-u.ac.jp

------------------------------------------------------------------------------