On 01/08/09 14:21, Dan Mick wrote: > where do you all stand with HPET usage? How are you using it, and when >
The Deep C-state project will allocate an interrupt for the HPET's timer during pcplusmp initialization. The cpu module plumbs the I/O APIC for the HPET's interrupt when it detects the processors and platform support deep C-states. I do not remember if we start the HPET's main counter in the cpu-specific code or in pcplusmp initialization (either will work). Anyone is free to read the HPET's main counter at any time. We will require that no one stops the HPET's main counter. We expect the HPET's interrupt to only be used to suppliment the Local APIC Timer when it stops in deep C-state. We have only seen 1 user available HPET timer on our test systems. This timer will be exclusively dedicated to LAPIC Timer support. There are 2 "legacy" HPET timers available for any other project. In theory someone can use these to replace the PIT and RTC. Some AMD processors stop the Local APIC Timer when all cores enter C1e. The HPET's timer will also be used as a LAPIC Timer proxy on these AMD system. AMD C1e support is future work. > do you expect the code that uses it to end up in Solaris? > As soon as you code review it. :-b We expect to send out the code review request for deep C-state/PAD any day now. :-) snv_108 is our target. Thank you, Bill > _______________________________________________ > tesla-dev mailing list > tesla-dev at opensolaris.org > http://mail.opensolaris.org/mailman/listinfo/tesla-dev >
