[gpm]Re: gpm-root

[Alessandro Rubini]

Hello.

I Cc: the gpm mailing list, as I'm not the maintainer any more. The
new maintainer, Nico Schottelius <nicos@pcsystems.de>, is subscribed
to the list, as other interested people is (and I am as well).
I include most of your message so the list can see it.

> I have installed gpm on my (2.2.19) linux box without any problem
> and wanted to thank you for this work and also ... for your nice
> book about the linux device drivers. This opportunity is given to me
> because gpm seems working as expected (fits the operation section in
> "man") but gpm-root has no effect at all.

I can't tell right away what may be going wrong, but you should find
all error messages in the system logs. What file they fall in depend
on your /etc/syslog.conf; a portable way to find it is using "ls -lrt
/var/log" right after the error, so you see what files are most
recent.  Similarly, "tail `ls -rt /var/log | tail -1`" may do the job
for you.

> All right with your book in hand I should be able to find out the
> trouble.

Not really. The book is about device drivers, not daemon
programs. They are completely different things.

> In matter of fact I have installed gpm for having a go to
> frame buffering and like to try something with Dinx, and Dinx
> suppose gpm is running. So time being always too short and having
> already spent much of it torturing "aty_fb" [x/yres parameters
> passed to kernel by lilo, ou-la-la], well, I'd like to have the
> solution at once for this time, if possible.

I don't know what Dinx is. Actually, I don't even understand what is
the relation of gpm with frame buffers. They are completely orthogonal
issue, because the part of the kernel related to cutting/copying text
and to drawing highlighted area is independent of whether the console
is VGA or frame-buffer (the only difference is the shape and color of
the cursor, and I find the white block quite annoying, personally --
but I just switched to fb, I still have to dig (I now have a ppc and
fb is mandatory :).

> I see in the doc that gpm is no longer maintained, specially for a
> lack of interest in it. By the way, don't you think that
> frame-buffering and Dinx, and what is going to come this way could
> open an interesting route for itself ?

Maybe, but I don't know what Dinx is and base my life on email-only
for a variety of reasons (no surfing, thus).

> I put some trace in attachment. Thank you anyway.

I remove stuff from this reply to not flood the list. The ps listing
shows no gpm-root, so only this line is interesting:

>   400 ?        S      0:02 gpm -m /dev/psaux -t imps2 -R raw

The trace of gpm-root is not useful, as it shows that the process
forks and exits. You'd better "strace -fF" instead, so the child
process is traced as well. But please look in the logs before that.
If the logs don't help, please post the new trace to the list.

Thanks for your message, it's appreciated. We just need a little more info.
/alessandro