Re: System Freeze (update)
Am Donnerstag, 23. November 2006 04:59 schrieb Richard L. Mace:
> I recently posted a similar freeze issue. In my case I had made many
> changes in a short space of time (xorg 7.1, kernel 2.6.18.x, from
> linuxant's driverloader to ndiswrapper).
> After weeks of testing I have confirmed that, for me, the source of my
> freezes is also wireless related. It is ndiswrapper 1.28 (with a Broadcom
> BCM4318 and 64 bit windows NDIS drivers). So, sadly, I'm back to using
> Linuxant's driverloader (which is rock-solid stable) and my freezes are
> BTW I had at least one response to my previous post where the poster
> confirmed the source of his freezes was the bcm43xx kernel module. This
> seems to agree with your findings. Hope this helps.
well, I would agree to this, if other people without the broadcom driver and
card would not have the freezes, too !
I myself use the native drier without linuxant or ndiswrapper.
So you can exclude them both, too.
If you will collect all messages about freezes and exclude all the things,
which all people with freezes NOT have, then there will be not much left,
which could cause this.
> On Wednesday 22 November 2006 23:48, Hans-J. Ullrich wrote:
> > Hi all,
> > as my computer suddenly freezes (this problem was alsready often
> > discussed), here are my newest watchings:
> > 1. Xorg (and X) and the closed source drivers for X by nvidia or ATI can
> > be wiped out, as it freezes also, when they are NOT loaded !
> > 2. The kernel can also be wiped out, as it freezes also at older kernel
> > (2.6.17) like other people reported
> > 3. The broadcom-module (wireless) in the kernel is still in responsible
> > in my mind, as it makes the last kernel-message before crash. (network
> > dies, the module loads automatically new)
> > 4. When I was listening to sound during a crash (i.e. to a mp3-file), the
> > file goes into a loop. This means, the last 2-3 seconds are played again
> > and again. (race condition with atiixp and atiixp-modem ???)
> > 5. There are some errors with APIC in dmesg before crash. Please take a
> > look: Nov 22 20:12:11 protheus2 kernel: Bluetooth: RFCOMM socket layer
> > initialized Nov 22 20:12:11 protheus2 kernel: Bluetooth: RFCOMM TTY layer
> > initialized Nov 22 20:12:11 protheus2 kernel: Bluetooth: RFCOMM ver 1.8
> > Nov 22 20:40:13 protheus2 kernel: APIC error on CPU0: 00(40)
> > Nov 22 20:45:37 protheus2 kernel: APIC error on CPU0: 40(40)
> > Nov 22 21:11:53 protheus2 kernel: APIC error on CPU0: 40(40)
> > Nov 22 21:32:44 protheus2 kernel: APIC error on CPU0: 40(40)
> > Nov 22 21:33:37 protheus2 kernel: APIC error on CPU0: 40(40)
> > Nov 22 22:12:34 protheus2 kernel: APIC error on CPU0: 40(40)
> > Do you see, what I mean ? This was just a few seconds before the crash !
> > Maybe someone can confirm this ! I remember something about starting the
> > kernel with "noapic" function, dunno if that is correct. Did someone test
> > this ?
> > Is the reason for the freeze found by now ???
> > Best regards
> > Hans