On Tue, Mar 02, 2004 at 09:44:43AM +0000, Alan Hourihane wrote: >On Tue, Mar 02, 2004 at 12:53:22AM -0500, Mike A. Harris wrote: >> On Sun, 29 Feb 2004, Michal Kosmulski wrote: >> >> >I recently upgraded from linux 2.4.23 to 2.6.2 and this caused some >> >problems with my mouse in XFree86. I have XFree86 4.3.0 (official >> >slackware 9.1 build) and I use nvidia's binary video driver (version 5336 >> >at the moment). My mouse is a ps/2 logitech mouse with a mouse wheel >> >(s48). After I upgraded to linux 2.6.2, after starting up X, the mouse >> >cursor didn't react to mouse movement for about 2 seconds of moving the >> >mouse. After that, the mouse pointer did move, but the mouse wheel was not >> >working. At that time I had "Protocol" set to "Auto" for my mouse and with >> >kernel 2.4.23 the mouse was detected correctly and the mouse wheel worked. >> >After I manually changed the setting to "ImPS/2", the delay in mouse >> >motion stopped and the wheel works again. i didn't find anything >> >non-standard in my XFree86 logs, but there were some messages in the >> >syslog. The first two messages have disappeared after changing "Auto" to >> >"ImPS/2", the rest still appears whenever X is started. >> >I also get strange mouse behavior once in a while (once every 3 days or >> >so): suddenly the mouse starts moving all by itself - it seems to go to >> >one of the screens corners, but I can't really see where it goes. This >> >motion stops at the moment I press any key. I don't think this could be >> >attributed to dust in the mouse mechanism or anything similar - I believe >> >it is also a bug. >> >Excerpt from /var/log/syslog follows: >> > >> > >> >Feb 11 14:54:16 nowy kernel: psmouse.c: Wheel Mouse at >> >isa0060/serio1/input0 lost synchronization, throwing 3 bytes away. >> >Feb 11 15:46:17 nowy kernel: psmouse.c: Wheel Mouse at >> >isa0060/serio1/input0 lost synchronization, throwing 3 bytes away. >> >Feb 11 16:11:25 nowy kernel: atkbd.c: Unknown key released (translated set >> >2, code 0x7a on isa0060/serio0). >> >Feb 11 16:11:25 nowy kernel: atkbd.c: This is an XFree86 bug. It shouldn't >> >access hardware directly. >> >Feb 11 16:11:25 nowy kernel: atkbd.c: Unknown key released (translated set >> >2, code 0x7a on isa0060/serio0). >> >Feb 11 16:11:25 nowy kernel: atkbd.c: This is an XFree86 bug. It shouldn't >> >access hardware directly. >> >> This is indeed an XFree86 bug. A few weeks ago David posted a >> patch to attempt to fix it, however that patch didn't work. >> >> I added some debugging patches to the server and tracked the >> problem down and fixed it in the latest Fedora Core development >> XFree86 builds. >> >> You can grab the latest Red Hat XFree86 src.rpm from rawhide and >> extract the relevant patch if you like. >> >> Hope this helps. >> >> >> P.S. On a side note, before anyone asks ... I'd have submitted >> my fix in bugzilla, however nobody was motivated to respond to >> any of my emails on the subject of this bug over the last few >> weeks while I was trying to help find a solution, so I am not >> motivated to go out of my way to submit a patch either. Two way >> street. > >Mike, > >Sorry you felt that you needed a response. Reading your emails it >sounded like you were on top of things and would post a patch when >you were ready for people to test. > >That's certainly why I hadn't responded. Maybe others felt that way too. I'm not really sure what he's talking about. The patch I posted, plus the equivalent change for lnx_io.c, does fix the atkbd.c kernel warnings for me. I'll commit it. The mouse synchronisation is another matter. David _______________________________________________ XFree86 mailing list XFree86@xxxxxxxxxxx http://XFree86.Org/mailman/listinfo/xfree86
