Re: [PATCH 2/3] stop_machine: simplify

["Akinobu Mita" <akinobu.mita@gmail.com>]

2008/7/8 Rusty Russell <rusty@rustcorp.com.au>:
> On Tuesday 08 July 2008 21:44:40 Akinobu Mita wrote:
>> I found a small possible cleanup in this patch.
>
> Well spotted!  I think this cleanup is actually orthogonal to my patch,
> so best served as a separate patch, how's this?

Actually the cpu_online() check was necessary before appling this
stop_machine: simplify patch.

With old __stop_machine_run(), __stop_machine_run() could succeed
(return !IS_ERR(p) value) even if take_cpu_down() returned non-zero value.
The return value of take_cpu_down() was obtained through kthread_stop().

> ===
> Hotplug CPU: don't check cpu_online after take_cpu_down

So it seems that folding this patch into stop_machine: simplify patch is
more appropriate.

> Akinobu points out that if take_cpu_down() succeeds, the cpu must be offline
> (otherwise we're in deep trouble anyway.
>
> Remove the cpu_online() check, and put a BUG_ON().
>
> Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
> Cc: "Akinobu Mita" <akinobu.mita@gmail.com>
>
> diff -r 805a2e5e68dd kernel/cpu.c
> --- a/kernel/cpu.c      Tue Jul 08 23:04:48 2008 +1000
> +++ b/kernel/cpu.c      Tue Jul 08 23:07:43 2008 +1000
> @@ -226,8 +226,7 @@ static int __ref _cpu_down(unsigned int
>        set_cpus_allowed_ptr(current, &tmp);
>
>        err = __stop_machine_run(take_cpu_down, &tcd_param, cpu);
> -
> -       if (err || cpu_online(cpu)) {
> +       if (err) {
>                /* CPU didn't die: tell everyone.  Can't complain. */
>                if (raw_notifier_call_chain(&cpu_chain, CPU_DOWN_FAILED | mod,
>                                            hcpu) == NOTIFY_BAD)
> @@ -235,6 +234,7 @@ static int __ref _cpu_down(unsigned int
>
>                goto out_allowed;
>        }
> +       BUG_ON(cpu_online(cpu));
>
>        /* Wait for it to sleep (leaving idle task). */
>        while (!idle_cpu(cpu))
>