Hi, Michael
Since only virtual ioapic uses the lock-less logic, so it shouldn't introduce the issues. Anyway, it should be safer to add the lock considering its semantic and future usage models. Would you like to prepare a patch to enhance it ? Thanks!
Xiantao
Michael S. Tsirkin wrote:
> Hi,
> arch/ia64/kvm/kvm-ia64.c does this:
>
> mmio_dev = vcpu_find_mmio_dev(vcpu, p->addr, p->size, !p->dir);
> if (mmio_dev) {
> if (!p->dir)
> kvm_iodevice_write(mmio_dev, p->addr, p->size,
> &p->data);
> else
> kvm_iodevice_read(mmio_dev, p->addr, p->size,
> &p->data);
> without holding kvm lock.
>
> I know that find/read/write on the bus require locking on x86 -
> are they safe to do without locking on ia64?
--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
