hyper-v timer support

The hyper-v spec on MSDN states:

“If a significant number of timer signals have been missed, the
hypervisor might be unable to compensate by using period modulation.
In this situation, some timer expiration signals might be skipped
completely.”

It isn’t clear if this sentence is referring to the lazy timer
mechanism introduced in the following paragraph, or if this sentence
is describing a situation that can occur while using periodic
modulation to catch up a virtual processor. If it is the latter, how
can that possibly work (correctly) in a VM that depends on periodic
interrupts?

See: http://msdn.microsoft.com/en-us/library/ff542561(v=VS.85).aspx

Mark Roddy

“Hypervisor timers” aren’t used for the periodic interrupt. The kernel in
the VM uses what it thinks is PC timer hardware. The section that you’re
reading refers to something entirely different.

Our hypervisor spec documents a lot of things that aren’t used in guest VMs,
only in the parent partition. I’ve argued that there’s no point in this,
and we may remove documentation of this kind of thing in the future.

Jake Oshins
Hyper-V I/O Architect
Windows Kernel Group

This post implies no warranties and confers no rights.


“Mark Roddy” wrote in message news:xxxxx@ntdev…

The hyper-v spec on MSDN states:

“If a significant number of timer signals have been missed, the
hypervisor might be unable to compensate by using period modulation.
In this situation, some timer expiration signals might be skipped
completely.”

It isn’t clear if this sentence is referring to the lazy timer
mechanism introduced in the following paragraph, or if this sentence
is describing a situation that can occur while using periodic
modulation to catch up a virtual processor. If it is the latter, how
can that possibly work (correctly) in a VM that depends on periodic
interrupts?

See: http://msdn.microsoft.com/en-us/library/ff542561(v=VS.85).aspx

Mark Roddy