Messages in this thread | | | Date | Mon, 16 Feb 2015 15:29:48 +0000 | From | Mark Rutland <> | Subject | Re: [PATCH v2] ARM: smp: Only expose /sys/.../cpuX/online if hotpluggable |
| |
On Fri, Feb 13, 2015 at 11:01:37PM +0000, Russell King - ARM Linux wrote: > On Fri, Feb 13, 2015 at 09:44:50PM +0000, Magnus Damm wrote: > > Also, based on the comment in mcpm_cpu_can_disable() it looks like the > > PSCI hook may be executed once only with your change in place? > > Hopefully PSCI is OK not being invoked for every CPU shutdown. > > This is why I've said (in the parent thread) that I'm not happy to > apply this patch. Mark Rutland has indicated that he has MCPM cases > where the CPUs which can be disabled changes dynamically according > to the secure firmware requirements, and ripping out todays > infrastructure in light of that, only to have to add it back again > later makes no sense.
To clarify, PSCI and MCPM are unrelated. It was originally conceived that MCPM would use PSCI as a backend, but it turns out that they're effectively mutually exclusive, and are handled separately. I still want to add support for MIGRATE in the PSCI client code, but this is independent of MCPM.
In some cases it's possible to determine at boot time that a CPU cannot be hotplugged (e.g. in PSCI if the TOS is non-migrateable), so having separate hooks for determining that static and dynamic ability to hotplug a CPU sounds reasonable to me.
Thanks, Mark.
| |