From: "Heiko Stübner" <heiko@sntech.de>
To: Xing Zheng <zhengxing@rock-chips.com>
Cc: linux@arm.linux.org.uk, linux-arm-kernel@lists.infradead.org,
linux-rockchip@lists.infradead.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH v2 8/9] ARM: rockchip: add support smp for rk3036
Date: Thu, 17 Sep 2015 22:15:46 +0200 [thread overview]
Message-ID: <22222181.r5MSqEbWpB@diego> (raw)
In-Reply-To: <1442486286-1873-1-git-send-email-zhengxing@rock-chips.com>
Am Donnerstag, 17. September 2015, 18:38:06 schrieb Xing Zheng:
> The rk3036 is dual-core soc, we can use this patch to enable cpu1
> enter boot secondary, and hotplug(online/offline).
>
> Signed-off-by: Xing Zheng <zhengxing@rock-chips.com>
> ---
>
> Changes in v2: None
>
> arch/arm/mach-rockchip/platsmp.c | 121
> ++++++++++++++++++++++++++++++++++++++ 1 file changed, 121 insertions(+)
>
> diff --git a/arch/arm/mach-rockchip/platsmp.c
> b/arch/arm/mach-rockchip/platsmp.c index 3e7a4b7..14218eb 100644
> --- a/arch/arm/mach-rockchip/platsmp.c
> +++ b/arch/arm/mach-rockchip/platsmp.c
> @@ -34,6 +34,8 @@
>
> static void __iomem *scu_base_addr;
> static void __iomem *sram_base_addr;
> +static void __iomem *cru_base_addr;
> +
> static int ncores;
>
> #define PMU_PWRDN_CON 0x08
> @@ -41,6 +43,8 @@ static int ncores;
>
> #define PMU_PWRDN_SCU 4
>
> +#define RK3036_SOFTRST_CON(x) ((x) * 0x4 + 0x110)
> +
> static struct regmap *pmu;
>
> static int pmu_power_domain_is_on(int pd)
> @@ -184,6 +188,7 @@ static int __init rockchip_smp_prepare_sram(struct
> device_node *node) }
>
> rsize = resource_size(&res);
> +
> if (rsize < trampoline_sz) {
> pr_err("%s: reserved block with size 0x%x is to small for trampoline
size
> 0x%x\n", __func__, rsize, trampoline_sz);
> @@ -350,3 +355,119 @@ static struct smp_operations rockchip_smp_ops
> __initdata = { };
>
> CPU_METHOD_OF_DECLARE(rk3066_smp, "rockchip,rk3066-smp",
> &rockchip_smp_ops); +
> +/* for rk3036 */
> +
> +static int rk3036_set_power_domain(int pd, bool on)
> +{
> + /* there are 2cpus on rk3036 soc, we just need to be care cpu1 */
> + if (pd != 1)
> + return 0;
> +
> + if (on)
> + writel_relaxed(0x20000, cru_base_addr + RK3036_SOFTRST_CON(0));
> + else
> + writel_relaxed(0x20002, cru_base_addr + RK3036_SOFTRST_CON(0));
this is definitly the wrong way. The reset handling is the same as on the
rk3288 - so you can use the regular reset mechanism like the other socs and
should definitly _never_ write to the cru directly.
The only visible difference is that you cannot control the power-domains of
the cpu cores. Its sibling the rk3128 seems to have controllable power domains
again.
I'm still pondering how to make this future proof...
Heiko
> +
> + dsb();
> +
> + return 0;
> +}
> +
> +static void __init rk3036_smp_prepare_cpus(unsigned int max_cpus)
> +{
> + struct device_node *node;
> + unsigned int l2ctlr;
> + unsigned int i, cpu;
> +
> + /* get cru_base_addr */
> + node = of_find_compatible_node(NULL, NULL, "rockchip,rk3036-cru");
> + if (!node) {
> + pr_err("%s: could not find cru dt node\n", __func__);
> + return;
> + }
> +
> + cru_base_addr = of_iomap(node, 0);
> + if (!cru_base_addr) {
> + pr_err("%s: could not map cru registers\n", __func__);
> + return;
> + }
> +
> + /* get sram_base_addr */
> + node = of_find_compatible_node(NULL, NULL, "rockchip,rk3036-smp-sram");
> + if (!node) {
> + pr_err("%s: could not find sram dt node\n", __func__);
> + return;
> + }
> +
> + sram_base_addr = of_iomap(node, 0);
> + if (!sram_base_addr) {
> + pr_err("%s: could not map sram registers\n", __func__);
> + return;
> + }
> +
> + /* get ncores */
> + asm ("mrc p15, 1, %0, c9, c0, 2\n" : "=r" (l2ctlr));
> + ncores = ((l2ctlr >> 24) & 0x3) + 1;
> + cpu = MPIDR_AFFINITY_LEVEL(read_cpuid_mpidr(), 0);
> +
> + /* Make sure that all cores except the first are really off */
> + for (i = 1; i < ncores; i++)
> + rk3036_set_power_domain(0 + i, false);
> +}
> +
> +static int rk3036_boot_secondary(unsigned int cpu, struct task_struct
> *idle) +{
> + if (cpu >= ncores) {
> + pr_err("%s: cpu %d outside maximum number of cpus %d\n",
> + __func__, cpu, ncores);
> + return -ENXIO;
> + }
> +
> + /* start the core */
> + rk3036_set_power_domain(0 + cpu, true);
> +
> + /*
> + * We need to wait a moment after soft reset CPUx on rk3036,
> + * otherwise, CPUx will startup failed.
> + */
> + udelay(10);
> + writel(virt_to_phys(secondary_startup), sram_base_addr + 8);
> + writel(0xDEADBEAF, sram_base_addr + 4);
> + dsb_sev();
> +
> + return 0;
> +}
> +
> +#ifdef CONFIG_HOTPLUG_CPU
> +static int rk3066_cpu_kill(unsigned int cpu)
> +{
> + /*
> + * We need a delay here to ensure that the dying CPU can finish
> + * executing v7_coherency_exit() and reach the WFI/WFE state
> + * prior to having the power domain disabled.
> + */
> + mdelay(1);
> +
> + rk3036_set_power_domain(0 + cpu, false);
> +
> + return 1;
> +}
> +
> +static void rk3066_cpu_die(unsigned int cpu)
> +{
> + v7_exit_coherency_flush(louis);
> + while (1)
> + cpu_do_idle();
> +}
> +#endif
> +
> +static struct smp_operations rk3036_smp_ops __initdata = {
> + .smp_prepare_cpus = rk3036_smp_prepare_cpus,
> + .smp_boot_secondary = rk3036_boot_secondary,
> +#ifdef CONFIG_HOTPLUG_CPU
> + .cpu_kill = rk3066_cpu_kill,
> + .cpu_die = rk3066_cpu_die,
> +#endif
> +};
> +CPU_METHOD_OF_DECLARE(rk3036_smp, "rockchip,rk3036-smp", &rk3036_smp_ops);
next prev parent reply other threads:[~2015-09-17 20:15 UTC|newest]
Thread overview: 37+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-09-17 8:28 [PATCH v2 0/9] Build and support rk3036 SoC platform Xing Zheng
2015-09-17 8:28 ` [PATCH v2 1/9] ARM: dts: rockchip: add core rk3036 dts Xing Zheng
2015-09-17 9:18 ` Heiko Stübner
2015-09-24 2:18 ` Xing Zheng
2015-09-17 8:28 ` [PATCH v2 2/9] clk: rockchip: add dt-binding header for rk3036 Xing Zheng
2015-09-17 9:25 ` Heiko Stübner
2015-09-24 2:17 ` Xing Zheng
2015-09-17 8:28 ` [PATCH v2 3/9] clk: rockchip: add clock controller " Xing Zheng
2015-09-17 9:47 ` Heiko Stübner
2015-09-24 3:04 ` Xing Zheng
2015-09-24 3:31 ` Xing Zheng
2015-10-07 10:24 ` Heiko Stuebner
2015-09-17 8:28 ` [PATCH v2 4/9] clk: rockchip: add new clock type and " Xing Zheng
2015-09-17 9:54 ` Heiko Stübner
2015-09-22 22:41 ` Stephen Boyd
2015-09-22 22:58 ` Heiko Stübner
2015-09-22 23:19 ` Stephen Boyd
2015-09-30 23:32 ` Heiko Stübner
2015-10-01 0:51 ` Stephen Boyd
2015-09-17 9:59 ` [PATCH v2 0/9] Build and support rk3036 SoC platform Heiko Stübner
2015-09-17 10:32 ` [PATCH v2 5/9] dt-bindings: add documentation of rk3036 clock controller Xing Zheng
2015-09-17 15:09 ` Heiko Stübner
2015-09-24 3:42 ` Xing Zheng
2015-09-17 10:34 ` [PATCH v2 6/9] pinctrl: rockchip: add support for the rk3036 Xing Zheng
2015-09-17 12:47 ` Heiko Stübner
2015-09-17 10:37 ` [PATCH v2 7/9] rockchip: make sure timer5 is enabled on rk3036 platforms Xing Zheng
2015-09-17 15:05 ` Heiko Stübner
2015-09-28 12:25 ` Xing Zheng
2015-09-28 12:44 ` Heiko Stübner
2015-09-28 12:53 ` Xing Zheng
2015-09-17 10:38 ` [PATCH v2 8/9] ARM: rockchip: add support smp for rk3036 Xing Zheng
2015-09-17 20:15 ` Heiko Stübner [this message]
2015-09-28 11:50 ` Xing Zheng
2015-09-17 10:39 ` [PATCH v2 9/9] rtc: hym8563: make sure hym8563 can be normal work Xing Zheng
2015-09-17 12:07 ` Heiko Stübner
2015-09-17 12:31 ` Alexandre Belloni
2015-09-17 12:44 ` Xing Zheng
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=22222181.r5MSqEbWpB@diego \
--to=heiko@sntech.de \
--cc=linux-arm-kernel@lists.infradead.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-rockchip@lists.infradead.org \
--cc=linux@arm.linux.org.uk \
--cc=zhengxing@rock-chips.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).