> Hi, 
>
> I will be following up this project. So to recapitulate, the aim is to take
> down a processor based on a policy implemented in the kernel, and to
> afterwards re-enable it.

The mechanisms are in the kernel but it would be better to keep the policy
in userland.  We have three userland policy mechanisms for CPUs running
in userland today:

One is the Fault Management architecture, running inside of fmd(1M), which
knows how to diagnose cpu failures based on error telemetry and disable
bad CPUs.  The other is the interrupt distribution daemon, intrd(1M),
which tries to keep interrupt mappings distributed appropriately.
And the third is powerd(1M), which does E-* compliance power management of
many things, including disks and CPUs.

The most recent work here to look at is PSARC 2005/067, which added
features for x86 CPUs.  Remember that x86 CPUs have multiple power states
before you shut them off (S0, S1, etc.)  It would be nice to see this
enhanced for things other than E-*, i.e. let an administrator tell the
system to "optimize for performance" or "optimize for least power consumed"

-Mike

-- 
Mike Shapiro, Solaris Kernel Development. blogs.sun.com/mws/
_______________________________________________
opensolaris-code mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/opensolaris-code

Reply via email to