On Fri, Sep 03, 2004 at 12:15:00PM -0300, Bruno Diniz wrote: > When the patch you mentioned was not applied, I could make it work > with any kernel version by simply adding a 'disable_APIC_timer' in the > beginning of acpi_bus_init, and a 'enable_APIC_timer' in the end of > this function. > > By the way, what is Local APIC for? What do I lose having Local APIC disabled?
Find it in kernel sources: A local APIC (Advanced Programmable Interrupt Controller) is an integrated interrupt controller in the CPU. If you have a single-CPU system which has a processor with a local APIC, you can say Y here to enable and use it. If you say Y here even though your machine doesn't have a local APIC, then the kernel will still run with no slowdown at all. The local APIC supports CPU-generated self-interrupts (timer, performance counters), and the NMI watchdog which detects hard lockups. The processor local APIC is of fundamental importance in SMP and RT environments to manage correctly I/O and and internal interrupts, timing and so. You can live without on uniprocessor systems. P5 did not have lapic indeed. If I remember correctly first multi-proc board had PPro which was the first with a lapic... Anyway tnis is the 'missing pointer' for the issue on LKML: http://www.ussg.iu.edu/hypermail/linux/kernel/0401.2/1525.html And it seems simply missing in debian sources. So the fix by Christopher was different (and unuseful for this issue). > > On Fri, 3 Sep 2004 09:44:07 +0200, Francesco P. Lovergine > <[EMAIL PROTECTED]> wrote: > > On Thu, Sep 02, 2004 at 11:00:50PM -0300, Bruno Diniz wrote: > > > Is there a problem with the idea of "turn off APIC timer before ACPI > > > initialization and then turn back on after ACPI initialization > > > finishes"? This is exactly what the tiny patch I've sent does. I would > > > agree that 'nolapic' boot param "solves" the problem, but IMHO the > > > idea above is clearer and works perfectly. > > > > > > > The idea is clear, but that patch does not work on the 2100 I tried. > > It was already implemented in current 2.6.8 AFAIK and it has exactly > > the same problem of hang up. > > > > > Thanks, > > > > > > Bruno. > > > > > > On Thu, 2 Sep 2004 12:52:35 +0200, Francesco P. Lovergine > > > <[EMAIL PROTECTED]> wrote: > > > > Ok I verified that: > > > > > > > > * This issue is yet valid for current 2.6.8 with acpi=on > > > > * It's valid for 2.4 series up to 2.4.27 with acpi=on > > > > * Patch proposed in 2.6.8 does not solve the issue, so it can be > > > > removed. > > > > > > > > The only workaround in all cases is the 'nolapic' boot param. > > > > > > > > -- > > > > Francesco P. Lovergine > > > > > > > > > > > > -- > > Francesco P. Lovergine > > > > > > -- > Bruno Diniz de Paula -- Francesco P. Lovergine
