Eirik Oeverby wrote: > Eirik Oeverby wrote: > >> Just for those interested: >> I do *not* get any messages at all from the kernel (or elsewhere) when >> my mouse goes haywire. And it's an absolute truth (just tested back >> and forth 8 times) that it *only* happens with SCHED_ULE and *only* >> with old versions (~1.50) and the very latest ones (1.75 as I'm >> currently running). 1.69 for instance did *not* show any such problems. >> >> I will, however, update my kernel again now, to get the latest >> sched_ule.c (if any changes have been made since 1.75) and to test >> with the new interrupt handler. I have a suspicion it might be a >> combination of SCHED_ULE and some signal/message/interrupt handling >> causing messages to get lost along the way. Because that's exactly how >> it feels... > > > Whee. Either the bump from sched_ule.c 1.75 to 1.77 changed something > back to the old status, or the new interrupt handling has had some major > influence. > All I can say is - wow. My system is now more responsive than ever, I > cannot (so far) reproduce any mouse jerkiness or bogus input or > anything, and things seem smoother. > > As always I cannot guarantee that this report is not influenced by the > placebo effect, but I do feel that it's a very real improvement. The > fact that I can start VMWare, Firebird, Thunderbird, Gaim and gkrellm at > the same time without having *one* mouse hickup speaks for itself. I > couldn't even do that with ULE. > > So Jeff or whoever did the interrupt stuff - what did you do? Oh and just to add to the goods/bads: A make -j 16 buildworld still makes the box sluggish when it gets to the crypto stuff, but not anywhere close to what it was like before. I'd say it probably matches or beats ULE now. And one *major* thing: I can now play sound again! Without clicks or pops like I've had since 5.1-RELEASE .. I can play sound for real! *meep* *meep* what a relief! This upped my QOL (quality-of-life) quite significantly, given that I haven't been able to play music (wihout being annoyed beyond what is good for me or anyone else near) since, what, spring? *phew* Thanks, to whomever of you guys made this possible... /Eirik > > /Eirik > >> >> Greetings, >> /Eirik >> >> Morten Johansen wrote: >> >>> On Tue, 4 Nov 2003, Sheldon Hearn wrote: >>> >>>> On (2003/11/04 09:29), Eirik Oeverby wrote: >>>> >>>> > The problem is two parts: The mouse tends to 'lock up' for brief >>>> moments >>>> > when the system is under load, in particular during heavy UI >>>> operations >>>> > or when doing compile jobs and such. >>>> > The second part of the problem is related, and is manifested by the >>>> > mouse actually making movements I never asked it to make. >>>> >>>> Wow, I just assumed it was a local problem. I'm also seeing >>>> unrequested >>>> mouse movement, as if the signals from movements are repeated or >>>> amplified. >>>> >>>> The thing is, I'm using 4BSD, not ULE, so I wouldn't trouble Jeff to >>>> look for a cause for that specific problem in ULE. >>> >>> >>> >>> >>> >>> Me too. Have had this problem since I got a "Intellimouse" PS/2 >>> wheel-mouse. (It worked fine with previous mice (no wheel)). >>> With any scheduler in 5-CURRENT and even more frequent in 4-STABLE, >>> IIRC. Using moused or not doesn't make a difference. >>> Get these messages on console: "psmintr: out of sync", and the mouse >>> freezes then goes wild for a few seconds. >>> Can happen under load and sometimes when closing Mozilla (not often). >>> It could be related to the psm-driver. Or maybe I have a bad mouse, I >>> don't know. >>> I will try another mouse, but it does work perfectly in Linux and >>> Windogs... >>> >>> mj >>> >>> >>> >>> _______________________________________________ >>> freebsd-current_at_freebsd.org mailing list >>> http://lists.freebsd.org/mailman/listinfo/freebsd-current >>> To unsubscribe, send any mail to >>> "freebsd-current-unsubscribe_at_freebsd.org" >> >> >> >> >> _______________________________________________ >> freebsd-current_at_freebsd.org mailing list >> http://lists.freebsd.org/mailman/listinfo/freebsd-current >> To unsubscribe, send any mail to >> "freebsd-current-unsubscribe_at_freebsd.org" > > > > _______________________________________________ > freebsd-current_at_freebsd.org mailing list > http://lists.freebsd.org/mailman/listinfo/freebsd-current > To unsubscribe, send any mail to "freebsd-current-unsubscribe_at_freebsd.org"Received on Tue Nov 04 2003 - 15:55:20 UTC
This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:37:27 UTC