Am Dienstag, 21. April 2015, 16:21:27 schrieb Masahiro Yamada:
> Initial commit for a new SoC family, UniPhier, developed by
> Socionext Inc. (formerly, System LSI Business Division of
> Panasonic Corporation).
>
> This commit includes a minimal set of components for booting the
> kernel, including SMP support.
>
> Signed-off-by: Masahiro Yamada <yamada.masahiro@xxxxxxxxxxxxx>
> ---
[...]
> +static int uniphier_boot_secondary(unsigned int cpu,
> + struct task_struct *idle)
> +{
> + struct regmap *sbcm_regmap;
> + int ret;
> +
> + sbcm_regmap = syscon_regmap_lookup_by_compatible(
> + "socionext,uniphier-system-bus-controller-misc");
> + if (IS_ERR(sbcm_regmap)) {
> + pr_err("failed to regmap system-bus-controller-misc\n");
> + return PTR_ERR(sbcm_regmap);
> + }
> +
> + ret = regmap_write(sbcm_regmap, 0x1208,
> + virt_to_phys(uniphier_secondary_startup));
> + if (!ret)
> + asm("sev"); /* wake up secondary CPU */
> +
> + return ret;
> +}
> +
> +struct smp_operations uniphier_smp_ops __initdata = {
> + .smp_prepare_cpus = uniphier_smp_prepare_cpus,
> + .smp_boot_secondary = uniphier_boot_secondary,
> +};
this is more of a drive-by comment, but you're doing the syscon lookup on
every boot of a core, which includes walking big parts of the devicetree every
time.
Is there anything speaking against doing this once in a .smp_prepare_cpus
callback?
Heiko
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
