[PATCH 2/5] ARM: add SMP support for Broadcom mobile SoCs

[elder@linaro.org (Alex Elder)]

On 04/04/2014 01:56 PM, Alex Elder wrote:
> On 04/04/2014 10:30 AM, Tim Kryger wrote:
>> On Thu, Apr 3, 2014 at 7:18 PM, Alex Elder <elder@linaro.org> wrote:
>>
>>> diff --git a/arch/arm/mach-bcm/platsmp.c b/arch/arm/mach-bcm/platsmp.c
>>> new file mode 100644
>>> index 0000000..46a64f2
>>> --- /dev/null
>>> +++ b/arch/arm/mach-bcm/platsmp.c
>>
>>> +/* Size of mapped Cortex A9 SCU address space */
>>> +#define SCU_SIZE               0x58
>>
>>> +/*
>>> + * Enable the Cortex A9 Snoop Control Unit
>>> + *
>>> + * By the time this is called we already know there are multiple
>>> + * cores present.  We assume we're running on a Cortex A9 processor,
>>> + * so any trouble getting the base address register or getting the
>>> + * SCU base is a problem.
>>> + *
>>> + * Return 0 if successful or an error code otherwise.
>>> + */
>>> +static int __init scu_a9_enable(void)
>>> +{
>>> +       unsigned long config_base;
>>> +       void __iomem *scu_base;
>>> +
>>> +       if (!scu_a9_has_base()) {
>>> +               pr_err("no configuration base address register!\n");
>>> +               return -ENXIO;
>>> +       }
>>> +
>>> +       /* Config base address register value is zero for uniprocessor */
>>> +       config_base = scu_a9_get_base();
>>> +       if (!config_base) {
>>> +               pr_err("hardware reports only one core; disabling SMP\n");
>>> +               return -ENOENT;
>>> +       }
>>> +
>>> +       scu_base = ioremap((phys_addr_t)config_base, SCU_SIZE);
>>> +       if (!scu_base) {
>>> +               pr_err("failed to remap config base (%lu/%u) for SCU\n",
>>> +                       config_base, SCU_SIZE);
>>> +               return -ENOMEM;
>>> +       }
>>> +
>>> +       scu_enable(scu_base);
>>> +
>>> +       iounmap(scu_base);      /* That's the last we'll need of this */
>>> +
>>> +       return 0;
>>> +}
>>
>> This function seems useful for Cortex A9 MPCore in general.
>>
>> While you gave it a generic name, you put it in a Broadcom file.
>>
>> Is there a better location for this code?
> 
> I think it belongs in arch/arm/kernel/smp_scu.c.  I was thinking
> it might be generally useful when I wrote it (hence the more
> complete header comment, for example).  And I'll gladly move
> it there, I just didn't want anybody to get hung up on that.

I'm going to re-submit this series this morning.  I looked at
where this function might be used, and it looks like only one
other platform could use it (in hi3xxx_smp_prepare_cpus()).
Man of the others define get their scu base address some
other way (using a fixed constant or using device tree).

For now I'm going to keep it where it is.  If you or someone else
reiterate the suggestion I'll move it to arch/arm/kernel/smp_scu.c.

					-Alex