> I have been talking on this subject privatly with Joe too, but > now in public again. that would be me :-) > I'm not going to dig into this theory, yet. First of all I > wrote > a patch to be able to switch a Edge interrupt to Level or to > switch of the use of IO-APIC for a certain interrupt number > using > kernel parameters. no problems last night when using this patch I will know more after a few days .. also no problems with using noapic at the command line, so if you are having problems with NFS and unexpected interrupts or the X bezerko mouse, you may try using noapic at the command line next time you boot and see if you still encounter problems > > Short explanation on what MAY be happening: > if a interrupt is not handled in the correct way (see > edge/level/xtpic > differences), this may cause nasty races and a rather large > flow > of interrupts, that are not caused by the mouse. > The network (NFS) may be affected by this problem too. one thing that I do want to add here is that this large flow of interrupts will be larger than normal. On my machine 1/3 of my interrupts are mouse interrupts on IRQ 3(com 2). This is because I make HEAVY use of the mouse, as I do almost everything in X these days. (cause I can!) I'd imagine that an NFS mount can also have lots of interrupts as well... depending on its traffic level... > What I will do: > If the switching of the interrupt handling does help, there > may > be a bug in the io-apic code, or another common encountered > BIOS > bug (not unlikely). In that case I'll take up with mingo. im my case I have a strange BIOS that allows me to switch my 430HX chipset to behave like the old chipset or the new chipset, which to me (imho) sounds buggy in itself Joe _________________________________________________________ Do You Yahoo!? Get your free @yahoo.com address at http://mail.yahoo.com - Linux SMP list: FIRST see FAQ at http://www.irisa.fr/prive/mentre/smp-faq/ To Unsubscribe: send "unsubscribe linux-smp" to [EMAIL PROTECTED]
