SUMMARY -- screenblank refuses to acknowledge mouse/keyboard activity

From: Alireza Vali (reza@voyager.me.utexas.edu)
Date: Wed Feb 26 1992 - 10:13:20 CST


A few days ago, I posted a question regarding problems with the screenblank
utility on a Sun Sparc 2 running SunOS 4.1.1. I received many helpful
comments and would like to post a summary for anyone who's interested. I
did not include the names of the persons who responded because I wasn't sure
whether it was customary to do so, but I thanked those individuals through
e-mail instead. The text of the original post follows:

>Hello. We're having some problems with the screenblank utility on a Sparc 2.
>For some strange reason, when the screen blanker kicks in, it refuses to
>acknowledge any keystrokes or mouse clicks to restore the screen. The only
>way to restore the screen is to either kill screenblank or to get on another
>terminal and write or talk to a a window on the console. The rc.local entry
>for it is "/usr/bin/screenblank -d 600 -e 0.20" and it has always been that
>since day one. This problem started happening only last week. I have tried
>all the obvious remedies such as killing and restarting the process, rebooting,
>and varying the -d and -e numbers, but nothing seems to work. Can anybody
>clue me in as to what's going on and/or suggest solutions and/or recommend
>alternate public domain screen savers? Please e-mail responses as I am unable
>to keep up with the newsgroup(s) on a daily basis, and thank you in advance.

The following are the suggestions that actually solved the problem:

-------------------------------------------------------------------------
What version of the OS? We had this problem crop up last week and
we had just upgraded our hosts to 4.1.2 within a couple of weeks
of that. The *really* strange part was the _all_ of our hosts did
it at the same time. We walked in one morning and they were all
"hung". The only thing that I had done "weird" was the night before
I had reset the time on the server (back 10 minutes). All the other
machines (BTW, the server was not affected) reset their times via
rdate every night at midnight. Our screenblank is set at 300 seconds
which is less than the time adjustment...I'm sort of wondering about
some interaction between rdate and setting the time back too much...
you know,

(current time) - (time of last keypress) < 0

being given to an unsigned int kind of thing...

Keep this one alive

-------------------------------------------------------------------------

One common cause of the screenblank problems you describe is that
the access times of /dev/kbd, /dev/mouse, or /dev/console are somehow
set in the future. This causes screenblank to assume there is nothing
changed. You can verify this by typing:

        # ls -lu /dev/kbd /dev/mouse /dev/console

and comparing the times printed to the current date.

A simple fix is to simply remove and recreate the offending party,
        # rm /dev/kbd; mknod /dev/kbd c 29 0; chmod 666 /dev/kbd
        # rm /dev/mouse; mknod /dev/mouse c 13 0; chmod 666 /dev/mouse

This is a SunOS standard bug in screenblank, which has been fixed in
current Solbourne releases.

-------------------------------------------------------------------------

That is all. Again, thanks to the folks who responded.



This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:06:36 CDT