On Tue, Apr 17, 2012 at 2:43 AM, Sascha Hauer s.hauer@pengutronix.de wrote:
On Mon, Apr 16, 2012 at 06:50:12PM -0500, Robert Lee wrote:
Add common cpuidle init functionality that can be used by various imx platforms.
Signed-off-by: Robert Lee rob.lee@linaro.org
diff --git a/arch/arm/plat-mxc/Makefile b/arch/arm/plat-mxc/Makefile index e81290c..7c9e05f 100644 --- a/arch/arm/plat-mxc/Makefile +++ b/arch/arm/plat-mxc/Makefile @@ -16,6 +16,8 @@ obj-$(CONFIG_MXC_ULPI) += ulpi.o obj-$(CONFIG_MXC_USE_EPIT) += epit.o obj-$(CONFIG_MXC_DEBUG_BOARD) += 3ds_debugboard.o obj-$(CONFIG_CPU_FREQ_IMX) += cpufreq.o +obj-$(CONFIG_CPU_IDLE) += cpuidle.o
ifdef CONFIG_SND_IMX_SOC obj-y += ssi-fiq.o obj-y += ssi-fiq-ksym.o diff --git a/arch/arm/plat-mxc/cpuidle.c b/arch/arm/plat-mxc/cpuidle.c new file mode 100644 index 0000000..d1c9301 --- /dev/null +++ b/arch/arm/plat-mxc/cpuidle.c @@ -0,0 +1,89 @@ +/*
- Copyright 2012 Freescale Semiconductor, Inc.
- Copyright 2012 Linaro Ltd.
- The code contained herein is licensed under the GNU General Public
- License. You may obtain a copy of the GNU General Public License
- Version 2 or later at the following locations:
- */
+#include <linux/kernel.h> +#include <linux/io.h> +#include <linux/cpuidle.h> +#include <linux/hrtimer.h> +#include <linux/err.h> +#include <linux/slab.h>
+static struct cpuidle_device __percpu * imx_cpuidle_devices; +static struct cpuidle_driver *drv;
+void __init imx_cpuidle_set_driver(struct cpuidle_driver *p) +{
- drv = p;
+}
You like it complicated, eh? Why do you introduce a function which sets a variable...
This complication is used to deal with the timing of various levels of init calls. More explanation below.
+void imx_cpuidle_devices_uninit(void) +{
- int cpu_id;
- struct cpuidle_device *dev;
- for_each_possible_cpu(cpu_id) {
- dev = per_cpu_ptr(imx_cpuidle_devices, cpu_id);
- cpuidle_unregister_device(dev);
- }
- free_percpu(imx_cpuidle_devices);
+}
+static int __init imx_cpuidle_init(void)
... instead of passing it directly to the function which uses it?
If I called imx_cpuidle_init directly from imx5 or imx6q init routines, it would be getting called before the coreinit_call of core cpuidle causing a failure. There were various other directions to take and all seemed less desirable than this one.
One alternative would be to add a function to return the pointer to the cpuidle driver object based on the machine type. Functionality exists to identify imx5 as a machine type but not imx6q, so I couldn't use that machine based method without adding that extra code.
Another alternative would be to add a general platform lateinit_call function to each platforms that support cpuidle.
+{
- struct cpuidle_device *dev;
- int cpu_id, ret;
- if (!drv || drv->state_count > CPUIDLE_STATE_MAX) {
- pr_err("%s: Invalid Input\n", __func__);
You introduce a pointless error message on SoCs not setting a cpuidle driver. When will people learn that initcalls do not only run on machines they have on their desk?
Will fix in next version.
- return -EINVAL;
- }
- ret = cpuidle_register_driver(drv);
- if (ret) {
- pr_err("%s: Failed to register cpuidle driver\n", __func__);
It's always nice to add the return value to the error message to get a clue *what* went wrong. The function name though is rather uninteresting.
Will add the return value in the next version.
Sascha
-- Pengutronix e.K. | | Industrial Linux Solutions | http://www.pengutronix.de/ | Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0 | Amtsgericht Hildesheim, HRA 2686 | Fax: +49-5121-206917-5555 |