On Fri, 20 Oct 2023 15:45:30 +0100
"Russell King (Oracle)" <rmk+kernel@xxxxxxxxxxxxxxx> wrote:
> smp_prepare_boot_cpu() reads the cpuid of the first CPU, printing a
> message to state which processor booted, and setting it online and
> present.
>
> This cpuid is retrieved from per_cpu(cpu_data, 0).cpuid, which is
> initialised in arch/parisc/kernel/processor.c:processor_probe() thusly:
>
> p = &per_cpu(cpu_data, cpuid);
> ...
> p->cpuid = cpuid; /* save CPU id */
>
> Consequently, the cpuid retrieved seems to be guaranteed to also be
> zero, meaning that the message printed in this boils down to:
>
> pr_info("SMP: bootstrap CPU ID is 0\n");
>
> Moreover, since kernel/cpu.c::boot_cpu_init() already sets CPU 0 to
> be present and online, there is no need to do this again in
> smp_prepare_boot_cpu().
>
> Remove this code, and simplify the printk().
>
> Signed-off-by: Russell King (Oracle) <rmk+kernel@xxxxxxxxxxxxxxx>
LGTM, but trivial comment inline.
> ---
> arch/parisc/kernel/smp.c | 7 +------
> 1 file changed, 1 insertion(+), 6 deletions(-)
>
> diff --git a/arch/parisc/kernel/smp.c b/arch/parisc/kernel/smp.c
> index 2019c1f04bd0..989efc8c67a5 100644
> --- a/arch/parisc/kernel/smp.c
> +++ b/arch/parisc/kernel/smp.c
> @@ -404,13 +404,8 @@ static int smp_boot_one_cpu(int cpuid, struct task_struct *idle)
>
> void __init smp_prepare_boot_cpu(void)
> {
> - int bootstrap_processor = per_cpu(cpu_data, 0).cpuid;
> -
> /* Setup BSP mappings */
Given it's not doing any such thing, I think you should also drop the comment.
> - printk(KERN_INFO "SMP: bootstrap CPU ID is %d\n", bootstrap_processor);
> -
> - set_cpu_online(bootstrap_processor, true);
> - set_cpu_present(bootstrap_processor, true);
> + pr_info("SMP: bootstrap CPU ID is 0\n");
> }
>
>
