Re: [PATCH v3 5/8] ARM: mediatek: add smp bringup code for MT6580

[Matthias Brugger]

On Tuesday, August 04, 2015 09:54:21 PM Scott Shu wrote:
> Add support for cpu enable-method "mediatek,mt6580-smp" for booting
> secondary CPUs on MT6580.
> 
> Signed-off-by: Scott Shu <scott.shu@xxxxxxxxxxxx>
> ---
>  arch/arm/mach-mediatek/platsmp.c |  137
> ++++++++++++++++++++++++++++++++++++++ 1 file changed, 137 insertions(+)
> 
> diff --git a/arch/arm/mach-mediatek/platsmp.c
> b/arch/arm/mach-mediatek/platsmp.c index a5bc108..94f7865 100644
> --- a/arch/arm/mach-mediatek/platsmp.c
> +++ b/arch/arm/mach-mediatek/platsmp.c
> @@ -21,10 +21,16 @@
>  #include <linux/of_address.h>
>  #include <linux/string.h>
>  #include <linux/threads.h>
> +#include <linux/delay.h>
> +#include <asm/cacheflush.h>
> +
> +#include <linux/soc/mediatek/scpsys.h>
> 
>  #define MTK_MAX_CPU		8
>  #define MTK_SMP_REG_SIZE	0x1000
> 
> +static DEFINE_SPINLOCK(boot_lock);
> +
>  struct mtk_smp_boot_info {
>  	unsigned long smp_base;
>  	unsigned int jump_reg;
> @@ -56,6 +62,126 @@ static const struct of_device_id mtk_smp_boot_infos[]
> __initconst = { static void __iomem *mtk_smp_base;
>  static const struct mtk_smp_boot_info *mtk_smp_info;
> 
> +#ifdef CONFIG_HOTPLUG_CPU
> +static int mt6580_cpu_kill(unsigned cpu)
> +{
> +	int ret;
> +
> +	ret = spm_cpu_mtcmos_off(cpu, 1);

is this function ever called with wfi == 0?

> +	if (ret < 0)
> +		return 0;
> +
> +	return 1;
> +}
> +
> +static void mt6580_cpu_die(unsigned int cpu)
> +{
> +	for (;;)
> +		cpu_do_idle();
> +}
> +#endif
> +
> +static void write_pen_release(int val)
> +{
> +	pen_release = val;
> +	/* Make sure this is visible to other CPUs */
> +	smp_wmb();
> +	sync_cache_w(&pen_release);
> +}
> +
> +/*
> + * Refer common "pen" secondary release method
> + */
> +static int mt6580_boot_secondary(unsigned int cpu, struct task_struct
> *idle) +{
> +	unsigned long timeout;
> +	int ret;
> +
> +	/*
> +	 * Set synchronisation state between this boot processor
> +	 * and the secondary one
> +	 */
> +	spin_lock(&boot_lock);
> +
> +	/*
> +	 * The secondary processor is waiting to be released from
> +	 * the holding pen - release it, then wait for it to flag
> +	 * that it has been released by resetting pen_release.
> +	 *
> +	 * Note that "pen_release" is the hardware CPU ID, whereas
> +	 * "cpu" is Linux's internal ID.
> +	 */
> +	write_pen_release(cpu);
> +
> +	/*
> +	 * CPU power on control by SPM
> +	 */
> +	ret = spm_cpu_mtcmos_on(cpu);
> +	if (ret < 0) {
> +		spin_unlock(&boot_lock);
> +		return -ENXIO;
> +	}
> +
> +	timeout = jiffies + (1 * HZ);
> +	while (time_before(jiffies, timeout)) {
> +		/* Read barrier */
> +		smp_rmb();
> +
> +		if (pen_release == -1)
> +			break;
> +
> +		udelay(10);
> +	}
> +
> +	/*
> +	 * Now the secondary core is starting up let it run its
> +	 * calibrations, then wait for it to finish
> +	 */
> +	spin_unlock(&boot_lock);
> +
> +	return (pen_release != -1 ? -ENXIO : 0);
> +}
> +
> +static void mt6580_secondary_init(unsigned int cpu)
> +{
> +	/*
> +	 * Let the primary processor know we're out of the
> +	 * pen, then head off into the C entry point
> +	 */
> +	write_pen_release(-1);
> +
> +	/*
> +	 * Synchronise with the boot thread.
> +	 */
> +	spin_lock(&boot_lock);
> +	spin_unlock(&boot_lock);
> +}
> +
> +#define MT6580_INFRACFG_AO	0x10001000
> +#define SW_ROM_PD		BIT(31)

Please put the defines on the beginning of the file.
SW_ROM_PD bit is the same on other SoCs? I wasn't able to reveal that from the 
data sheets. If this is MT6580 only, the define should be renamed.

> +
> +static void __init mt6580_smp_prepare_cpus(unsigned int max_cpus)
> +{
> +	static void __iomem *infracfg_ao_base;
> +
> +	infracfg_ao_base = ioremap(MT6580_INFRACFG_AO, 0x1000);
> +	if (!infracfg_ao_base) {
> +		pr_err("%s: Unable to map I/O memory\n", __func__);
> +		return;
> +	}
> +
> +	/* Enable bootrom power down mode */
> +	writel_relaxed(readl(infracfg_ao_base + 0x804) | SW_ROM_PD,
> +		       infracfg_ao_base + 0x804);

Please use a define for the offset.
I prefer to not cascade reads and writes but to use a temporary variable. But 
if this is fine with the kernel coding style (I doubt, but I'm not sure) then 
this is fine.

> +
> +	/* Write the address of slave startup into boot address
> +	   register for bootrom power down mode */
> +	writel_relaxed(virt_to_phys(secondary_startup_arm),
> +		       infracfg_ao_base + 0x800);
> +
> +	iounmap(infracfg_ao_base);
> +}
> +
>  static int mtk_boot_secondary(unsigned int cpu, struct task_struct *idle)
>  {
>  	if (!mtk_smp_base)
> @@ -142,3 +268,14 @@ static struct smp_operations mt6589_smp_ops __initdata
> = { .smp_boot_secondary = mtk_boot_secondary,
>  };
>  CPU_METHOD_OF_DECLARE(mt6589_smp, "mediatek,mt6589-smp", &mt6589_smp_ops);
> +
> +static struct smp_operations mt6580_smp_ops __initdata = {
> +	.smp_prepare_cpus = mt6580_smp_prepare_cpus,
> +	.smp_secondary_init = mt6580_secondary_init,
> +	.smp_boot_secondary = mt6580_boot_secondary,
> +#ifdef CONFIG_HOTPLUG_CPU
> +	.cpu_kill = mt6580_cpu_kill,
> +	.cpu_die = mt6580_cpu_die,
> +#endif
> +};
> +CPU_METHOD_OF_DECLARE(mt6580_smp, "mediatek,mt6580-smp", &mt6580_smp_ops);

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/