On 16/10/2017 14:16, Arnd Bergmann wrote:
On Mon, Oct 16, 2017 at 2:08 PM, Paolo Bonzini pbonzini@redhat.com wrote:
On 16/10/2017 10:11, Arnd Bergmann wrote:
Thanks!
Since you've looked at it overall, do you have an opinion on the question how to fix the PV interface to deal with the pvclock_wall_clock overflow?
It has to be done separately for each hypervisor.
In KVM, for example, it is probably best to abandon pvclock_read_wallclock altogether, and instead use the recently introduced KVM_HC_CLOCK_PAIRING hypercall. drivers/ptp/ptp_kvm.c is already using it and it's y2106 safe.
Right, makes sense. I see that this interface is currently implemented only for 64-bit x86 in kvm_emulate_hypercall(). Could this be extended to x86-32 and the non-x86 architectures as well?
Yes, it could be implemented for x86-32 too. The whole pvclock concept however is specific to x86.
Paolo