Messages in this thread | | | Date | Sun, 08 Mar 2009 10:15:11 -0700 | From | Jeremy Fitzhardinge <> | Subject | Re: [PATCH 01/13] [VOYAGER] x86: add {safe,hard}_smp_processor_id to smp_ops |
| |
James Bottomley wrote: > Not having apics, Voyager can't use the default apic implementation of > these, it has to read from a special port in the VIC to get the > processor ID, so abstract these functions in smp_ops to allow voyager > to live simultaneously with the apic code. >
These aren't performance-sensitive at all, are they? smp_ops is not subject to patching/inlining optimisations happen to more hotpath pvops.
Is safe_smp_processor_id needed at all? It's only got two callers, and x86-64 just implements it as smp_processor_id().
> diff --git a/arch/x86/xen/smp.c b/arch/x86/xen/smp.c > index 035582a..0dfb8c0 100644 > --- a/arch/x86/xen/smp.c > +++ b/arch/x86/xen/smp.c > @@ -450,6 +450,11 @@ static irqreturn_t xen_call_function_single_interrupt(int irq, void *dev_id) > return IRQ_HANDLED; > } > > +static int xen_hard_smp_processor_id(void) > +{ > + return read_apic_id(); > +} > + > static const struct smp_ops xen_smp_ops __initdata = { > .smp_prepare_boot_cpu = xen_smp_prepare_boot_cpu, > .smp_prepare_cpus = xen_smp_prepare_cpus, > @@ -465,6 +470,8 @@ static const struct smp_ops xen_smp_ops __initdata = { > > .send_call_func_ipi = xen_smp_send_call_function_ipi, > .send_call_func_single_ipi = xen_smp_send_call_function_single_ipi, > + .hard_smp_processor_id = xen_hard_smp_processor_id, > + .safe_smp_processor_id = apic_safe_smp_processor_id, >
Hm, there's no meaningful apic-based implementation for these under Xen. DomU has no access to apics, and Dom0's vcpus don't have any fixed relationship to physical cpu apics. They should both just return smp_processor_id(), I guess.
J
| |