tegra_cpu_die() dead code?

tegra_cpu_die() dead code?

[Stephen Warren]

Joseph,

In next-20121217, I see:

void __ref tegra_cpu_die(unsigned int cpu)
{
        cpu = cpu_logical_map(cpu);

        /* Flush the L1 data cache. */
        flush_cache_all();

        /* Shut down the current CPU. */
        tegra_hotplug_shutdown();

        /* Clock gate the CPU */
        tegra_wait_cpu_in_reset(cpu);
        tegra_disable_cpu_clock(cpu);
	
        /* Should never return here. */
        BUG();

Given that both implementations (Tegra20/30) of tegra_hotplug_shutdown
include:

        bl      tegra20_cpu_shutdown
        mov     pc, lr                  @ should never get here

why does tegra_cpu_die() include the calls to tegra_wait_cpu_in_reset()
and tegra_disable_cpu_clock() after calling tegra_hotplug_shutdown()?
(and how can a CPU wait for itself to be in reset anyway)?

Perhaps this is fixed by the various patches you've posted but are
awaiting being applied?

Re: tegra_cpu_die() dead code?

[Joseph Lo]

On Thu, 2012-12-20 at 02:11 +0800, Stephen Warren wrote:
> Joseph,
> 
> In next-20121217, I see:
> 
> void __ref tegra_cpu_die(unsigned int cpu)
> {
>         cpu = cpu_logical_map(cpu);
> 
>         /* Flush the L1 data cache. */
>         flush_cache_all();
> 
>         /* Shut down the current CPU. */
>         tegra_hotplug_shutdown();
> 
>         /* Clock gate the CPU */
>         tegra_wait_cpu_in_reset(cpu);
>         tegra_disable_cpu_clock(cpu);
> 	
>         /* Should never return here. */
>         BUG();
> 
> Given that both implementations (Tegra20/30) of tegra_hotplug_shutdown
> include:
> 
>         bl      tegra20_cpu_shutdown
>         mov     pc, lr                  @ should never get here
> 
> why does tegra_cpu_die() include the calls to tegra_wait_cpu_in_reset()
> and tegra_disable_cpu_clock() after calling tegra_hotplug_shutdown()?
> (and how can a CPU wait for itself to be in reset anyway)?
> 
Indeed, I had a patch to fix this. Will update to here today.

Thanks,
Joseph