[Simh] Problem using a real VT terminal with simh

[Peter Allan]

Thanks to everyone for their quick responses.

To my surprise, Xon/Xoff was turned off on the VAX, specifically, SHOW TERM
TTA0 showed No HostSync. However SET TERM/PERM/HOSTSYNC TTA0 made no
difference.

I have tried setting the line speed to 2400, but that only slows down the
output. It still contains the same characters, to the extent that I can
tell.

I will continue to experiment and I will try using a different computer so
that the serial port hardware is (presumably) different. If I can't solve
it, I will indeed create an issue on github as Mark suggests.

Cheers

Peter Allan


On 5 December 2017 at 19:30, Mark Pizzolato <Mark at infocomm.com> wrote:

> On Tuesday, December 5, 2017 at 10:08 AM, Paul Koning wrote:
> > Do you have flow control (Xon/Xoff) enabled both on the terminal and on
> the
> > VAX?
> >
> > You might try a slower speed to see if that makes a difference.  It
> shouldn't,
> > but it might give more information.
> >
> > Is autobaud enabled on the line?  You might turn that off and set the
> speed
> > and related parameters explicitly.
>
> These are all great suggestions.  Autobaud should definitely be turned off
> on
> the VMS side, but I don't think that is the problem since he's able to
> login
> and the commands he types produce output, which merely is garbled.
>
> A slower speed, which is explicitly stated on the ATTACH command might
> help.
>      sim> ATTACH DZ LINE=0,CONNECT=SER0;2400-8N1
>
> I think that what you're seeing is output from the simulator overrunning
> the host OS provided interface to the serial port.  This came up a while
> back.
> The simulator serial port interface delivers data to the host OS driver one
> character at a time as it is emitted by the simulator.  The simulator is
> notified of I/O completion when the host OS said the IO was complete.
> On some operating systems, this completion notification merely indicated
> that the data had made it to a kernel buffer and not actually out through
> the hardware.  This early completion indication then was passed back
> through the simulator and more data was then emitted.  Ultimately it seemed
> that there was not a reliable way to know that the data has actually, not
> only
> made it to the hardware, but actually been transmitted.  Some relatively
> common serial ports have an output buffer in the hardware that will take
> a number of characters to be transmitted as quickly as the host puts them
> into the hardware, and then deliver it outbound after the fact.  In any
> case
> due to the problem of not getting a precise indication from the OS driver
> of
> when the data actually made it out the serial port, an alternative approach
> was taken.  This approach attempts to time the delivery of character I/O
> data to the OS driver by explicitly scheduling the I/O completion time
> based
> on the port specified speed.  This approach presumes that the OS actually
> starts to send some of the data to the serial port when it is presented and
> doesn't hold off until some buffer full condition is achieved.
>
> If the above ATTACH suggestion doesn't work for you, please create an issue
> at https://github.com/simh/simh/issues and we can try to dig into this
> further.
>
> > > On Dec 5, 2017, at 12:56 PM, Peter Allan <petermallan at gmail.com>
> wrote:
> > >
> > > I am trying to get a real VT420 terminal working with a simulated
> micro VAX
> > 3900. However, I am having problems getting text displayed on the screen
> of
> > the VT420.
> > > ...
> > > Once the microVAX has booted, if I hit Enter on the VT420, then I can
> see
> > that a login process starts on the simulated microVAX, but all I see on
> the
> > screen of the VT420 is a few garbage characters. Despite that, when I
> type a
> > username and a password, I do get logged into the simulated microVAX. Any
> > commands that I type on the keyboard of the VT420 appears on the screen
> of
> > the VT420 (except the password, of course), but anything that the
> microVAX is
> > sending to the VT420 appears as garbage. There are recognisable
> characters,
> > but if I TYPE a file, even a small one, what appears seems to be every
> 7th
> > character from the file starting with the first character. The output of
> > commands such as DIR appear to have the same problem. ...
> > _______________________________________________
> > Simh mailing list
> > Simh at trailing-edge.com
> > http://mailman.trailing-edge.com/mailman/listinfo/simh
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.trailing-edge.com/pipermail/simh/attachments/20171205/2b264864/attachment-0001.html>