rommel@runx.UUCP (04/03/87)
After re-making System V.2 for the ICM3216 (with minor alterations to the 2681 device-driver) I noticed that GETTY would not notice a terminal connected to tty2 (the non-modem port nexto the console) unless it was there when the system was booted or when GETTY was respawned. This was not the case with the original binary, but is the case if I remake the system as supplied (without my device driver changes). Does anyone know the fix for this one? It is quite anoying since I only have one terminal (normally console), so when I get tied up there I just have tty0. The modem port tty0 exibits the same characteristics but it wakes up as soon as DCD is asserted. (It may have someting to do with the fact that asserting DCD interrupts the processor. This is not available on tty2 or console). PS shows the terminal for the getty as '?'. Once a terminal is connected to tty2 and getty killed, the new getty is quite happy. Michael Katzmann ATN channel 7 Sydney
joer@nscpdc.UUCP (04/06/87)
In article <750@runx.ips.oz>, rommel@runx.ips.oz (Stephen Joyce) writes: > > After re-making System V.2 for the ICM3216 (with minor alterations to the > 2681 device-driver) I noticed that GETTY would not notice a terminal > connected to tty2 (the non-modem port nexto the console) unless it was > there when the system was booted or when GETTY was respawned. This was > not the case with the original binary, but is the case if I remake the > system as supplied (without my device driver changes). > > Michael Katzmann > ATN channel 7 Sydney Greetings from National Semiconductor's Portland Development Center, Michael, To help you we need to know what version of the software you are running. Please call 503-629-4414 between 8AM and 5PM Pacific Coast Time. (Note to all ICM product users within U.S.A.; the number for product support is 800-222-2433). As an alternative, contact your local NSC sales office or field engineer who will forward your problem to us. regards, Joe Rawlings ICM-Product Support National Semiconductor Corp. Portland Development Center
rommel@runx.UUCP (04/07/87)
After more carefull examination of the 2681 device driver (/usr/src/buts/
ns32000/io/sig2681.c) I found the bug that caused the problem with tty2
on the ICM2681. In the procedure siginit() a line appears:
if ( dev > (sig_cnt-2) ) {
sig_telco[dev] = OFF;
sig_ds[PORT(dev)] = 0;
}
"sig_cnt" appears in the config file, and is set to 4 (4 ports), so that
the if statement is only true for tty3 (ie console). Thus tty2 is assumed
to have full modem control when it too is a "telco" connection.
Changing the ">" to ">=" fixes the bug.
It makes me wonder how many other bugs were patched in the sysV.2
binary and not corrected in the source and .o files supplied by nsc.
Michael Katzmann (VK2BEA / G4NYV)
ATN Channel 7 Sydney