mgetty on null modem

• Messages sorted by:[ date ][ thread ][ subject ][ author ]

Hi,

On Sun, Nov 08, 1998 at 04:25:03PM +0100, etrius@hem2.passagen.se wrote:
> >> Unfortunately I don't have a place where I can try that...
> > No friends around, no "distribution kernel" to try (vs. the one you
> > compiled yourself)?
> No right now

Hmmm, bad.

> > Can you send us the output of "gcc -v"?
> Reading specs from /usr/lib/gcc-lib/i386-redhat-linux/2.7.2.3/specs
> gcc version 2.7.2.3

OK, that one is definitely working.

> > If serial ports stop working "suddenly", especially after a recompile of
> > the kernel, it's usually a very good sign that something in the kernel
> > broke.  Somebody else reported a very similar log file to the list
> > yesterday, where all his modems stopped working right after a kernel
> > recompilation - and his log looked exactly like yours.
> 
> This is kind of strange. The modem worked fine on ttyS1, and when I run
> the kernel that I used then (just did a recompile to put some thernet
> stuff in) it still doesn't work, but the errors are a bit different:
> 
> 11/08 15:28:27 yS1  mgetty: experimental test release 1.1.14-Apr02
[..]
> 11/08 15:28:27 yS1   do_makelock: lock='/var/lock/LCK..ttyS1'
> 11/08 15:28:27 yS1   lock made
> 11/08 15:28:28 yS1   tio_get_rs232_lines: status: RTS DTR

No error here, this is good.

> 11/08 15:28:28 yS1  WARNING: DSR is off - modem turned off or bad cable?

That one could hint at "no modem there" (or the modem isn't set to
hardwire DSR to on, AFAIR AT&S0 or &S1).

> 11/08 15:28:28 yS1   tss: set speed to 38400 (017)
> 11/08 15:28:28 yS1   tio_set_flow_control( HARD )
> 11/08 15:28:28 yS1   waiting for line to clear (VTIME), read:
> 11/08 15:28:28 yS1   waiting for line to clear (VTIME), read:
> 11/08 15:28:28 yS1   removing lock file
> 11/08 15:28:28 yS1  waiting...
> 11/08 15:31:39 ##### failed dev=ttyS1, pid=227, got signal 15, exiting

This looks good: mgetty initialized and ran fine until killed (low debug
level and "-r" set, thus no more stuff logged).

I think with this kernel, things should work fine.  Try it :-)

> > Try to start mgetty on the port that your mouse is currently connected to
> > - if that doesn't give errors, it might be the serial port.   But I still
> > suspect the kernel.
> 
> The mouse IS a serial mouse, but mgetty won't start on that port anyway.
> It prints the same things as ttyS1, but the mouse works fine both when
> I start gpm and within X (of course I killed all mouse related programs
> before trying).

I assume that mgetty triggers something in the kernel that gets
miscompiled, while the "other" things (mouse drivers) don't do that.  I
do not know enough about the way a serial mouse operates to pinpoint it,
though.

> It is quite an old computer 486/66, and all cards are the same age
> (around 1994).

Shouldn't be a problem.  One of my Linux machines is a 386/33.

gert
-- 
USENET is *not* the non-clickable part of WWW!
                                                           //www.muc.de/~gert/
Gert Doering - Munich, Germany                             gert@greenie.muc.de
fax: +49-89-35655025                        gert.doering@physik.tu-muenchen.de