vmm: Lookup vcpu pointers in vmmdev_ioctl.

[jhb]

Summary

Centralize mapping vCPU IDs to struct vcpu objects in vmmdev_ioctl and
pass vcpu pointers to the routines in vmm.c. For operations that want
to perform an action on all vCPUs or on a single vCPU, pass pointers
to both the VM and the vCPU using a NULL vCPU pointer to request
global actions.

[corvink]

sys/amd64/vmm/vmm_dev.c

445

Nit

541–548

555

vcpu could be left uninitialized. Might be a good idea to always initialize it. Not sure if the compiler will complain about it when enabling more warnings.

[jhb]

sys/amd64/vmm/vmm_dev.c

435

vcpu is initialized to NULL here btw.

[jhb]

Review feedback

[markj]

sys/amd64/vmm/vmm_lapic.c

80

This is a semantic change, right? Before, if vm_active_cpus() is empty, this function would do nothing and return 0 instead of ENOENT.

[jhb]

sys/amd64/vmm/vmm_lapic.c

80

Hmm, it is. Probably I shouldn't change the semantics in this commit (enough moving parts).

[corvink]

sys/amd64/vmm/vmm_lapic.c

80

Btw: why did you changed the layout of the function? The old layout avoids some nesting. This commit could change it to something like:

if (vcpu == NULL)
  dmask = vm_active_cpus(vm);
else
  CPU_SETOF(vcpu_vcpuid(vcpu), &dmask);
error = 0;
CPU_FOREACH_ISSET(cpuid, &dmask)
...

[jhb]

sys/amd64/vmm/vmm_lapic.c

80

I think it was just to avoid doing a round trip of vcpu = vm_vcpu(vm, vm_vcpuid(vcpu)).

[corvink]

sys/amd64/vmm/vmm_lapic.c

80

Yeah, this round trip looks unnecessary. The old layout avoids some code duplication for vlapic = vm_lapic(vcpu); error = vlapic_trigger_lvt(vlapic, vector);.