Date: Sat, 31 Jan 2004 12:10:50 -0500 (EST) From: Robert Watson <rwatson@freebsd.org> To: Matt Douhan <matt@fruitsalad.org> Cc: current@freebsd.org Subject: Re: mouse problem Message-ID: <Pine.NEB.3.96L.1040131120604.9532B-100000@fledge.watson.org> In-Reply-To: <200401302201.13687.matt@fruitsalad.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Fri, 30 Jan 2004, Matt Douhan wrote: > With src from 30 jan 9 PM CET (central european time) I get terrible > mouse problems with SHED_ULE, the mouse lives itw own life, totally > unrelated to physical touch (it is a touchpad on my laptop) > > changing the scheduler to 4bsd allows me to use the mouse for about 2 > minutes after boot, then it behaves the same as per above. > > the mouse works under several other OS's so I doubt it could be hw > problems. > > any ideas highly appreciated What hardware are you using? My Dell C600 notebook has had this problem for the last six months, although it comes and goes some. All of the various changes to the psm driver to limit the effect of lost packets seem not to have helped, and it's Damn Annoying. Interestingly, in most situations, the psm driver does not report a sync error. The mouse zips to either the left or the right side of the screen, as though the mouse is being moved constantly to one side or the other. If I try to move the mouse, it zips back to the side. One really odd thing, though, is that if I "try hard enough", I can flip the mouse to the other side, where it then sticks, as though the constant "push" of the mouse has changed direction. I have trouble thinking of any explanation for this. There's been some suggestion that poor interrupt handling latency in -CURRENT is the blame, and that once we start merging in the various interrupt thread optimizations, etc, things will improve. However, I'm running with at least a couple of optimizations on my notebook, and haven't seen it improve yet. It's tempting to back out all the psm "improvements" to limit PS/2 problems and see if it fixes things, but I haven't found time yet. The change in scheduler changing the bad behavior of the mouse suggests that this could be related to kernel thread scheduling, however. Interestingly, I have the problem more often when running off battery than running off wall power, which also suggests it might be related to scheduling/response time, as my system runs at 1GHz when plugged in, and 700MHz off battery... Robert N M Watson FreeBSD Core Team, TrustedBSD Projects robert@fledge.watson.org Senior Research Scientist, McAfee Research
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.NEB.3.96L.1040131120604.9532B-100000>