Drew Northup wrote:
> My "solution:
> is this (not that I have the faintest idea about how to pull it off): Pass
> some sort of ACPI or APM compliant clock data to the guest OS from the
> monitor derived from the reality that the host OS so whishes to hide for the
> most part. This allows the guest to, for the most part, "not care" what
> happens while it is not in control of the clock cycles on the CPU. This
> would allow mail, schedule, and calendar applications to function properly
> in most cases.
Most OSes I know of do not use the ACPI or APM clock... usually,
the software (OS) time is synchronised with the clock hardware
on startup, after which the software clock is updated independently
by the OS on every timer interrupt. Every version of Linux I have
read the code of works this way, anyway, though it may be that the
newer versions do use ACPI (but probably not --- accessing something
like the APM BIOS is rather costly, compared to keeping up a cheap
timer counter.)
Obviously, the RTC, etc. should be synchronised to the host clock.
-- Ramon
