From: will.deacon@arm.com (Will Deacon) Date: Wed, 27 Feb 2013 17:14:10 +0000 Subject: [PATCH 0/4] Convert highbank to use PSCI calls In-Reply-To: <512E0F79.6080301@gmail.com> References: <1361936528-25998-1-git-send-email-robherring2@gmail.com> <201302270858.20864.arnd@arndb.de> <512E0F79.6080301@gmail.com> Message-ID: <20130227171410.GC22286@mudshark.cambridge.arm.com> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On Wed, Feb 27, 2013 at 01:51:53PM +0000, Rob Herring wrote: > On 02/27/2013 02:58 AM, Arnd Bergmann wrote: > >> This series converts highbank to use PSCI calls for secondary core boot, > >> hotplug, cpuidle, suspend/resume, reset, and poweroff. Doing this removes > >> about 200 lines of platform code (1/3 of it) and abstracts out the A9 and > >> A15 differences (some of which are not yet in place). > > > > This looks pretty cool. Can you or Will comment on what is still left > > for the SMP ops that is actually required as platform specific code? > > > > Specifically, I'm wondering if we could have a set of default smp_ops > > based on this that we could use with the default machine descriptor [1] > > and any platform that has SMP enabled but a NULL smp_ops pointer. > > My smp code is exactly the same as mach-virt with this, but adds cpu_die > for hotplug. smp_secondary_init will go away with Catalin's series to > use notifiers for the gic secondary init. This leaves smp_boot_secondary. > > We could move these remaining smp_ops to psci.c and create psci_smp_ops. > That probably needs to be after the gic notifier changes. > > Your patch, psci smp ops, and converting the arch timers to use > OF_CLKSRC_DECLARE will eliminate mach-virt unless Will has more things > planned for it. As long as we remain compatible with "linux,dummy-virt" then I'm not fussed what happens to the support code. I don't envisage adding anything else to it that can't live under drivers/ (e.g. new virtio devices). Will