On Tue, Jan 21, 2020 at 05:47:52PM +0000, Qais Yousef wrote:
> On 01/21/20 17:03, Russell King - ARM Linux admin wrote:
> > On Mon, Nov 25, 2019 at 11:27:41AM +0000, Qais Yousef wrote:
> > > +void smp_shutdown_nonboot_cpus(unsigned int primary_cpu)
> > > +{
> > > + unsigned int cpu;
> > > +
> > > + if (!cpu_online(primary_cpu)) {
> > > + pr_info("Attempting to shutdodwn nonboot cpus while boot cpu is offline!\n");
> > > + cpu_online(primary_cpu);
>
> Eh, that should be cpu_up(primary_cpu)!
>
> Which I have to say I'm not if is the right thing to do.
> migrate_to_reboot_cpu() picks the first online cpu if reboot_cpu (assumed 0) is
> offline
>
> migrate_to_reboot_cpu():
> 225 /* Make certain the cpu I'm about to reboot on is online */
> 226 if (!cpu_online(cpu))
> 227 cpu = cpumask_first(cpu_online_mask);
>
> > > + }
> > > +
> > > + for_each_present_cpu(cpu) {
> > > + if (cpu == primary_cpu)
> > > + continue;
> > > + if (cpu_online(cpu))
> > > + cpu_down(cpu);
> > > + }
> >
> > How does this avoid racing with userspace attempting to restart CPUs
> > that have already been taken down by this function?
>
> This is meant to be called from machine_shutdown() only.
>
> But you've got a point.
>
> The previous logic that used disable_nonboot_cpus(), which in turn called
> freeze_secondary_cpus() didn't hold hotplug lock. So I assumed the higher level
> logic of machine_shutdown() ensures that hotplug lock is held to synchronize
> with potential other hotplug operations.
freeze_secondary_cpus() takes the CPU maps lock while it takes CPUs
down, and then disables cpu hotplug by incrementing
cpu_hotplug_disabled. Incrementing that prevents cpu_up() and
cpu_down() being used, thereby preventing userspace from changing the
online state of any CPU in the system.
> But I can see now that it doesn't.
>
> With this series that migrates users to use device_{online,offline}, holding
> the lock_device_hotplug() should protect against such races.
>
> Worth noting that this an existing problem in the code and not something
> I introduced, of course it makes sense to fix it properly as part of this
> series.
>
> I'm not sure how the other archs deal with this TBH.
>
> Thanks for having a look!
>
> Cheers
>
> --
> Qais Yousef
>
--
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/
FTTC broadband for 0.8mile line in suburbia: sync at 12.1Mbps down 622kbps up
According to speedtest.net: 11.9Mbps down 500kbps up
|