On 09/17/08 07:44, Li, Aubrey wrote: > Bill.Holler wrote: > > >> On 09/16/08 16:32, Li, Aubrey wrote: >> >>> Hi Bill, >>> >>> Bill.Holler wrote: >>> >>> >>> >>>> Hi Aubrey, >>>> >>>> My test machine has the same problem with the new push. >>>> >>>> The problem was a miss-merge; sorry it was probably my fault. :-( >>>> The promotion code can promote passed the end of the cstate list. >>>> >>>> The promotion code at the end of cpu_acpi_idle(): >>>> if ((cs_type < cpu_max_cstates) && (delta > >>>> cstate->promotion)) mcpu->mcpu_idle_type++; >>>> >>>> cs_type can be decremented allowing mcpu->mcpu_idle_type to >>>> overflow. /* * OSPM uses the BM_STS bit to determine the >>>> power state to enter >>>> * when considering a transition to or from the C2/C3 power >>>> state. >>>> * if C3 is determined, bus master activity demotes >>>> the power state >>>> * to C2. >>>> */ >>>> if ((cs_type >= CPU_ACPI_C3) && cpu_acpi_bm_sts()) >>>> cs_type = CPU_ACPI_C2; >>>> >>>> >>>> >>>> >>>> >>> It's not a miss-merge, It looks like a promotion bug. >>> Thanks to fix it. So, is your system working okay with C4 or the >>> second C3? >>> >>> >> Yes, my system is working great with the second C3. :-) >> >> >> My initial thoughts on duplicate C-States are: we should keep >> duplicate C-States. Does that sound reasonable? >> >> Is it possible to list the proper C-State for duplicate C-States? >> For example can a duplicate C3 be listed as C3 instead of C4? >> Will the current_cstate field of cpu_info kstat be able to >> distinguish duplicate C-States? >> >> > I failed to find the mail. But if I recall correctly, Mark mentioned > some > boxes with the buggy BIOS could have a few entries are the same. > Since we don't know how BIOS implementation the _CST object, > I think the current implentation may be more reasonable. > > Thanks, > -Aubrey > _______________________________________________ > tesla-dev mailing list > tesla-dev at opensolaris.org > http://mail.opensolaris.org/mailman/listinfo/tesla-dev >
I remember a bug Mark worked on with P-States. Buggy BIOS could create bad duplicate _PSS P-State ACPI object. The duplicate objects contained zeros? I agree we should ignore duplicate _CST objects until they prove to be stable. Thank you, Bill
