From owner-freebsd-hackers Tue Sep 9 12:00:11 1997 Return-Path: Received: (from root@localhost) by hub.freebsd.org (8.8.7/8.8.7) id MAA27562 for hackers-outgoing; Tue, 9 Sep 1997 12:00:11 -0700 (PDT) Received: from counterintelligence.ml.org (mdean.vip.best.com [206.86.94.101]) by hub.freebsd.org (8.8.7/8.8.7) with ESMTP id MAA27555 for ; Tue, 9 Sep 1997 12:00:06 -0700 (PDT) Received: from localhost (jamil@localhost) by counterintelligence.ml.org (8.8.7/8.8.5) with SMTP id LAA05275; Tue, 9 Sep 1997 11:46:11 -0700 (PDT) Date: Tue, 9 Sep 1997 11:46:11 -0700 (PDT) From: "Jamil J. Weatherbee" To: Andrew Atrens cc: freebsd-hackers@freebsd.org Subject: re:Help! keyboard lockups - could be kernel bug ? In-Reply-To: <199709090549.WAA05083@hub.freebsd.org> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-hackers@freebsd.org X-Loop: FreeBSD.org Precedence: bulk This has happened to me also quite a few times, at first I suspected netscape but then discovered that when the machine is under heavy load the CTRL-ALT-F? sequences don't work correctly to change from X to the Virtual consoles. On 9 Sep 1997, Andrew Atrens wrote: > > Hi again, > > Last week I wrote about bizarre problem I was experiencing with keyboard > lockups under X. Well a week has passed and the situation has not improved > despite my best efforts. Actually to be honest I'm not sure what to try > next. Any pointers would be appreciated... > > Actually, one further clue: if I'm pounding my IDE disk ( while running xdm > of course ) the keyboard will also occasionally lock up. The lock up is > fairly gradual, usually starts with a single key press being missed, then > a few, and then eventually ( after, say 30 seconds ) I can play `chopsticks' > without effect ;) . Its almost as though the activity on the PCI bus is > somehow gating the keyboard i/o... unfortunately for the keyboard its a > one way trip - once its locked up its dead. > > One last strange observation is that, under xdm if I do a - to > get back to the text-console, the background color of the screen changes > from black to red. No such behaviour with good ol' startx ! > > Cheers, > > Andrew. ( opinions are mine, not Nortel's ) > > -- > > In message "Help! keyboard lockups - could be kernel bug ?", I wrote: > > > > > Hi Folks, > > > > Last week I was happily running FreeBSD-2.2.2 on my Pentium system with an > > ASUS T2P4 motherboard ( has PCI 430HX chipset ). But unfortunately I got > > the upgrade itch, and moved up (?) to an ASUS TX97 ( has PCI 430TX chipset ). > > > > To make a long story short, 2.2.2 consistently locked up on boot - a real > > showstopper. After much trial and error I discovered the culprit to be > > AUTO_EOI_X kernel option(s). When I built a kernel without these options it > > booted and appeared (initially at least) to function properly. > > > > Then, occasionally I noticed that my xdm login window was dropping characters > > and locking up. Enough of that, so I downloaded the sources and built/installed > > the August 08 3.0-SNAP. > > > > Xdm login was still dropping characters, so I went out and got Xfree86-3.3.1 > > sources and rebuilt X. Xdm login was *still* dropping characters and > > occasionally locking up the keyboard, so I began to look at this symptom more > > closely. > > > > While the xdm login banner is displayed, I run an xlock -inroot -mode random > > process in the background. Depending on how graphics/cpu intensive the > > particular screen saver is, the lock up behaviour worsens. For example, the > > lock ups are worse when running xlock in `galaxy' mode. > > > > Furthermore, if I get past the xdm login window, xterms and rxvts in my > > session cause keyboard lockup when xlock is running `galaxy' in the root > > window. > > > > So here's the strange thing: IF I DON'T RUN XDM, and instead just use startx > > then the system behaves normally - no keyboard lockups regardless of how > > hard I pound X. > > > > So what the heck is going on? -- What do I try next ? > > > > Any help would be appreciated :) > > > > Andrew > > ( opinions are mine, not Nortels ) > > > > > > >