Elsa MicroLink 33.6TQV on HP-UX 10.20 with vgetty
Lutz Jaenicke (Lutz.Jaenicke@iee.TU-Berlin.DE)
Thu, 4 Dec 1997 19:08:30 +0100
On Wed, Dec 03, 1997 at 09:41:13PM +0100, Gert Doering wrote:
> > > I dimly remember that we have seen this on some other platform as well.
> > > It seems to be related to setting VTIME to something > 0.
> > Yes, when I changed this in init.c to unbuffered(?) c_cc[VTIME]=0 select
> > behaved as expected.
>
> This is what I remembered, yes. Unfortunately, vgetty depends on VTIME>0
> (at least that's what I remember -- don't forget that I didn't write a
> single line of vgetty code, and have not much idea about its inner
> workings...)
If I read the vgetty source carefully enough, it does use VTIME>0. Even then,
select() is not used so often: to be precise I only have the problem during
modem inactivity. After changing VTIME to zero in init.c the problem went
away. I hence had to leave the other VTIME settings (during voice play and
read) untouched!
> So, the question is, what does "ready to read" mean...
>
> a) "will not block"
> b) "there is data"
>
> ... I don't think it's "cast into stone" how select() should behave here.
I totally agree, obviously the HP man page is not clear enough!
> Well... I don't like HP/UX... but SGI with Irix seems to be even more
> vgetty-unfriendly ;-)
>
> [I'm not sure why I dislike HP/UX so much. I think it has something to do
> with very uncommon cursor key handling in "hpterm", with too many devices
> for a single serial port, and weird cabling schemes on some platforms...]
Hmm, I cannot comment on the cabling schems and the number of device files,
but I definitely can say, that with vi you don't even need cursor keys :-)
(And they rather work with tcsh, but don't ask about the page-up/-down keys :-)
And now back to the more serious facts:
By now I am also in control of the data flow problems (at least I hope so).
When looking at the log file I noticed that even though the Elsa ist set
to RTS/CTS handshaking with AT\Q3 and the HP is, too, the modem seems not
to use hardware handshaking. After switching to XON/XOFF with AT\Q1 and
AT\Q4 for recording and playing, respectively, it works flawlessly for
some 10 test calls. Yet, I still didn't try fax.
I am not sure whether or not my experience is really hard data, so I am not
sure that I should submit this to the vgetty author. After some more testing
I will submit the small adjustments for HP-UX 10.20 to you, Gert.
Many thanks for the support in solving my problem,
Lutz
--
Lutz Jaenicke Lutz.Jaenicke@iee.TU-Berlin.DE
TU Berlin http://www.iee.TU-Berlin.DE/personen/jaenicke/
Institut fuer Elektrische Energietechnik Tel. +49 30 314-24552
Einsteinufer 11, D-10587 Berlin Fax. +49 30 314-21133
.