http://defect.opensolaris.org/bz/show_bug.cgi?id=5444
Bill Holler <bill.holler at sun.com> changed:
What |Removed |Added
----------------------------------------------------------------------------
Priority|P4 |P2
Severity|minor |normal
AssignedTo|aubrey.li at intel.com |bill.holler at sun.com
--- Comment #7 from Bill Holler <bill.holler at sun.com> 2009-02-01 21:20:00
---
The following webrev is the result of extensive testing on a kernel
instrumented to record events and state at all CPU power domain levels.
This fix has 3 mechanisms for detecting workloads where there will be a loss
of performance if the CPU too aggresively enters deeper C-states.
This version improves SPECweb slightly on our 2 socket system,
it allows the system to save much more power at idle and low utilizations,
and it allows very good performance on benchmarks.
http://cr.opensolaris.org/~bholler/5444/webrev/
I would like to push this ASAP to allow further testing.
I am not sure what the locking requirements are in
cpupm_alloc_ms_cstate() and cpupm_free_ms_cstate()?
These functions can be called during boot, during _CST change event
notifications, and when C-states are disabled/enabled.
--
Configure bugmail: http://defect.opensolaris.org/bz/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
