[finally answering the review emails now that I'm rebasing most of my
pending stuff for v9]
On 04/06/12 03:32, Rusty Russell wrote:
> On Thu, 31 May 2012 10:04:14 +0100, Marc Zyngier <marc.zyngier at arm.com> wrote:
>> +static bool write_dcsw(struct kvm_vcpu *vcpu,
>> + const struct coproc_params *p,
>> + unsigned long cp15_reg)
>> +{
>> + u32 val;
>> +
>> + val = *vcpu_reg(vcpu, p->Rt1);
>> +
>> + switch(p->CRm) {
>> + case 6: /* Upgrade DCISW to DCCISW, as per HCR.SWIO */
>> + case 14: /* DCCISW */
>> + asm volatile("mcr p15, 0, %0, c7, c14, 2" : : "r" (val));
>> + break;
>> +
>> + case 10: /* DCCSW */
>> + asm volatile("mcr p15, 0, %0, c7, c10, 2" : : "r" (val));
>> + break;
>> + }
>> +
>> + cpumask_setall(&vcpu->arch.require_dcache_flush);
>> + cpumask_clear_cpu(vcpu->cpu, &vcpu->arch.require_dcache_flush);
>
> Hmm, but vcpu->cpu is not necessarily smp_processor_id() here.
>
> I think you do need get_cpu() and put_cpu() around this:
> kvm_arch_vcpu_load() will do a full flush if vcpu->cpu !=
> smp_processor_id().
I think that now that we have both __kvm_vcpu_run() and handle_exit() as
part of a non-preemptible section, we're guaranteed that vcpu->cpu is
smp_processor_id().
M.
--
Jazz is not dead. It just smells funny...
