Re: [PATCH RFC 1/1] arm64: Use PSCI calls for CPU stop when hotplug is supported

[Sudeep Holla]

On Fri, Jan 25, 2019 at 12:33:56PM +0530, Pramod Kumar wrote:

[...]

>
> This is an GPIO interrupt. This can not be marked secure as for that
> we need to mark whole GPIO controller as secure which is not possible
> as GPIO controller is meant for non-secure world having more than 100
> lines connected.
>
> I agree we have work around where we invoke handler in Linux and
> switch to ATF via SMC and from ATF we need bring all secondary CPU to
> ATF via sending SGI and and then respective core flushes the L1/L2 and
> bring himself out of coherency domain and cluster and MCU shutdowns
> the CPU subsystem gracefully. This could work for our requirement.
> Need to check ATF support for that.
>

Yes platform specific requirement and platform specific solution,
happy ending :).

> But What about generic system? This patch address the generic
> multi-master system's requirement.

Why do we need to address this in generic system ? And what exactly
is this multi-master system's requirement ? Linux runs on one or more
masters and will own it completely. Shutdown must involve everything
it owns if it needs to be graceful.

> Consider system where shutting down the linux does not mean shutting
> down the complete system.

Why is that the case ? Is it forceful shutdown ? Does Linux just own
CPUs and don't care about other blocks in the system ? Irrespective
of what it owns, system shutdown will take care.

> Lets take an example of smartnic case Where NIC master and CPUs access
> cachable DDR. In smarnic its quite common to bring CPUs on demand means
> when needed via MCU help.

Yes you are talking about CPU hotplug or system here ? The above
indicates, it's just CPU hotplug and a solution already exists for that.

> Now in full-fledged system. if CPU subsystem is shutdown via poweroff
> command which does not bring secondary CPUs out of coherency domain,
> it will bring the complete system unstable when NIC master tries to
> access DDR and snoop is send to CPUs as well which is not available.
> Fabric/System hangs...
>

That's because your custom solution just sends ipi to stop CPUs. If you
shutdown the system, all the required information is save to non-volatile
memory and system is powered off gracefully.

> I feel While shutting down the CPUs subsystem or powering off, All
> secondary CPUs must be shutdown properly by bring-out of coherency
> domain to remain rest of subsystem usable. I agree that introducing
> PSCI call introduce delay for shutdown/reboot case but stability
> matter than little delay.

IPI_STOP is not designed to do a graceful shutdown of CPU subsystem.
Use CPU hotplug. You are trying to make you custom requirement a generic
one. We have CPU hotplug framework to do what you want, you just have to
use it. Infact you are doing almost the same with you patch, I don't see
any point as why CPU hotplug can't be used.

--
Regards,
Sudeep