[PATCH v21 3/8] arm64: kdump: implement machine_crash_shutdown()

[AKASHI Takahiro]

On Wed, Jul 06, 2016 at 10:28:57AM -0700, Geoff Levand wrote:
> On Wed, 2016-07-06 at 16:52 +0900, AKASHI Takahiro wrote:
> > --- a/arch/arm64/include/asm/smp.h
> > +++ b/arch/arm64/include/asm/smp.h
> > @@ -136,6 +136,10 @@ static inline void cpu_panic_kernel(void)
> >   */
> >  bool cpus_are_stuck_in_kernel(void);
> >  
> > +#ifdef CONFIG_KEXEC_CORE
> > +extern void smp_send_crash_stop(void);
> > +#endif
> 
> It seems this could be simplified by removing these
> conditionals, then move the conditionals that are
> around the smp_send_crash_stop definition to inside
> that definition.

OK.

> > diff --git a/arch/arm64/kernel/smp.c b/arch/arm64/kernel/smp.c
> > index 62ff3c0..51e0b73 100644
> > --- a/arch/arm64/kernel/smp.c
> > +++ b/arch/arm64/kernel/smp.c
> > @@ -800,6 +803,28 @@ static void ipi_cpu_stop(unsigned int cpu)
> >  > 	> 	> cpu_relax();
> >  }
> >  
> > +#ifdef CONFIG_KEXEC_CORE
> > +static atomic_t waiting_for_crash_ipi;
> > +
> > +static void ipi_cpu_crash_stop(unsigned int cpu, struct pt_regs *regs)
> > +{
> > +> 	> crash_save_cpu(regs, cpu);
> > +
> > +> 	> atomic_dec(&waiting_for_crash_ipi);
> > +
> > +> 	> local_irq_disable();
> > +
> > +#ifdef CONFIG_HOTPLUG_CPU
> > +> 	> if (cpu_ops[cpu]->cpu_die)
> > +> 	> 	> cpu_ops[cpu]->cpu_die(cpu);
> > +#endif
> > +
> > +> 	> /* just in case */
> > +> 	> while (1)
> > +> 	> 	> wfi();
> > +}
> > +#endif
> 
>  /*
> >   * Main handler for inter-processor interrupts
> >   */
> > @@ -830,6 +855,14 @@ void handle_IPI(int ipinr, struct pt_regs *regs)
> >  > 	> 	> irq_exit();
> >  > 	> 	> break;
> >  
> > +#ifdef CONFIG_KEXEC_CORE
> > +> 	> case IPI_CPU_CRASH_STOP:
> > +> 	> 	> irq_enter();
> > +> 	> 	> ipi_cpu_crash_stop(cpu, regs);
> > +
> > +> 	> 	> unreachable();
> > +#endif
> 
> It seems like you could simplifiy with something like:
> 
> 	case IPI_CPU_CRASH_STOP:
> 		if (IS_ENABLED(CONFIG_KEXEC_CORE) {
> 			irq_enter();
> 			ipi_cpu_crash_stop(cpu, regs);
> 			unreachable();
> 		}

OK, I will take this.

> Then move the CONFIG_KEXEC_CORE conditional around
> ipi_cpu_crash_stop to just around crash_save_cpu().

Yeah, we can do that, but I added these conditionals just because
the kernel size can be reduced (a bit, obviously) if kexec is
not needed.
So instead of removing the conditional around ipi_cpu_crash_stop(),
an empty definition will be added for !CONFIG_KEXEC_CORE.

Thanks,
-Takahiro AKASHI

>  -Geoff