Comsole hangs with ISC X r1.2 and fas2.07 controlled Logitech serial mouse

[Saul Lubkin]

I've had problems with fas2.07 and the ISC X v1.2 Xvga driver.  I emailed to
Uwe Doering, the developer and maintainer of fas, and he was kind enough to
write back promptly (twice).

To explain the problem, here's some of my last note, slightly edited:

>By puttering, I discoverd something else interesting:  If I tell X that the
>mouse port is ANY of the ports controlled by fas, I get the same result!  For
>example, fas2.07 controls COM2 (which has an old ASCII terminal attached), plus
>an old Stargate eight port dumb card.  The first port on the Stargate (like
>COM2) has full modem control; the other seven don't.  The first Stargate port
>is being used by a Hayes compatible 2400 Baud modem.  The other Starate ports
>have nothing attached.
>
>If I tell X that the mouse is on ANY of the fas2.07 controlled ports (including
>the ones with NOTHING attached (ttyF2 and above)-- or the modem port (ttyF1,
>the first Stargate port), or the terminal port (ttyF0, COM2 on a
>serial/parallel card) --- then the command "xinit&" results in an immortal
>Xvga process.  With kernels in which the mouse (on  COM1) is controlled by fas,
>the same thing happens if X is told (correctly) that the mouse is on COM1.
>
>Exactly what seems to be happening is:  
>
>xinit starts out.  It starts up Xvga.  Xvga tries to send, or recieve, on the
>serial device.  It waits forever for something.  In the course of doing this,
>it blanks the screen, and causes all keyboard or mouse input to be ineffective
>(including attempting to switch VT's).  I can shut the system down by logging
>in as root on an attached ASCII terminal, which is working fine.  On
>rebooting, an automatic fsck is done on the /usr filesystem -- apparently
>since the shutdown routine didn't succeed in unmounting /usr, undoubdetly
>because of the immortal Xvga.
>
>Again, this didn't happen with fas2.06.  In both cases , ISC X r1.2 was used.
>
>I noticed, in the release notes to fas2.07, that there have been some changes
>in initiallization proceedures.  Maybe something was inconsistent with some of
>the initialization code in the Xvga server in ISC 1.2?
>
>
>I appreciate your contacting me about the ISC X 1.2 Xvga server problem.  And
>I very much appreciate your writing this excellent driver, which has certainly
>become a permanent part of my system.


In his note, Uwe Doering had suggested trying a minor device number of zero
plus the port number -- same problem.

I should note that my modem traffic -- uucp, zmodem and pcomm -- have all
DRAMATICALLY speeded up since I've been using fas -- despite the fact that I am
using a mere Hayes compatible 2400 Baud modem, and AT class UART chips.  And
that now all fas serial ports are accessible to DOS under VP/ix (I've tried it;
works great.  VP/ix works on the ASCII terminal, and, on kernels with fas2.07
controlling my Logitech three button serial mouse, DOS Word Perfect 5.1 works
fine with the mouse).

My solution is to use fas2.07 to control all ports except for the mouse.
Asy has that port.

If anyone has any ideas, please email.  I'll report any further info.

					Sincerely yours,

					Saul Lubkin