RE: [v2,1/2] rtc/fsl: add FTM alarm driver as the wakeup source

[Biwen Li]

-----Original Message-----
From: Biwen Li <biwen.li@xxxxxxx> 
Sent: 2019年7月10日 19:04
To: a.zummo@xxxxxxxxxxxx; alexandre.belloni@xxxxxxxxxxx; Leo Li <leoyang.li@xxxxxxx>
Cc: linux-rtc@xxxxxxxxxxxxxxx; linux-kernel@xxxxxxxxxxxxxxx; Xiaobo Xie <xiaobo.xie@xxxxxxx>; Jiafei Pan <jiafei.pan@xxxxxxx>; Ran Wang <ran.wang_1@xxxxxxx>; Biwen Li <biwen.li@xxxxxxx>
Subject: [v2,1/2] rtc/fsl: add FTM alarm driver as the wakeup source

For the paltforms including LS1012A, LS1021A, LS1028A, LS1043A, LS1046A, LS1088A, LS208xA that has the FlexTimer module, implementing alarm functions within RTC subsystem to wakeup the system when system going to sleep (work with RCPM driver).

Signed-off-by: Biwen Li <biwen.li@xxxxxxx>
---

Change in v2:
    - remove code about setting rcpm

 drivers/rtc/Kconfig             |  14 ++
 drivers/rtc/Makefile            |   1 +
 drivers/rtc/rtc-fsl-ftm-alarm.c | 311 ++++++++++++++++++++++++++++++++++++++++
 3 files changed, 326 insertions(+)
 create mode 100644 drivers/rtc/rtc-fsl-ftm-alarm.c

diff --git a/drivers/rtc/Kconfig b/drivers/rtc/Kconfig index 03b60d5..0758a08 100644
--- a/drivers/rtc/Kconfig
+++ b/drivers/rtc/Kconfig
@@ -1313,6 +1313,20 @@ config RTC_DRV_IMXDI
 	   This driver can also be built as a module, if so, the module
 	   will be called "rtc-imxdi".
 
+config RTC_DRV_FSL_FTM_ALARM
+	tristate "Freescale FlexTimer alarm timer"
+	depends on ARCH_LAYERSCAPE
+	default y
+	help
+	   For the FlexTimer in LS1012A, LS1021A, LS1028A, LS1043A, LS1046A,
+	   LS1088A, LS208xA, we can use FTM as the wakeup source.
+
+	   Say y here to enable FTM alarm support. The FTM alarm provides
+	   alarm functions for wakeup system from deep sleep.
+
+	   This driver can also be built as a module, if so, the module
+	   will be called "rtc-fsl-ftm-alarm".
+
 config RTC_DRV_MESON
 	tristate "Amlogic Meson RTC"
 	depends on (ARM && ARCH_MESON) || COMPILE_TEST diff --git a/drivers/rtc/Makefile b/drivers/rtc/Makefile index 9d997fa..5cccb07 100644
--- a/drivers/rtc/Makefile
+++ b/drivers/rtc/Makefile
@@ -79,6 +79,7 @@ obj-$(CONFIG_RTC_DRV_HID_SENSOR_TIME) += rtc-hid-sensor-time.o
 obj-$(CONFIG_RTC_DRV_HYM8563)	+= rtc-hym8563.o
 obj-$(CONFIG_RTC_DRV_IMXDI)	+= rtc-imxdi.o
 obj-$(CONFIG_RTC_DRV_IMX_SC)	+= rtc-imx-sc.o
+obj-$(CONFIG_RTC_DRV_FSL_FTM_ALARM)	+= rtc-fsl-ftm-alarm.o
 obj-$(CONFIG_RTC_DRV_ISL12022)	+= rtc-isl12022.o
 obj-$(CONFIG_RTC_DRV_ISL12026)	+= rtc-isl12026.o
 obj-$(CONFIG_RTC_DRV_ISL1208)	+= rtc-isl1208.o
diff --git a/drivers/rtc/rtc-fsl-ftm-alarm.c b/drivers/rtc/rtc-fsl-ftm-alarm.c new file mode 100644 index 0000000..1836c2e
--- /dev/null
+++ b/drivers/rtc/rtc-fsl-ftm-alarm.c
@@ -0,0 +1,311 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Freescale FlexTimer Module (FTM) alarm device driver.
+ *
+ * Copyright 2014 Freescale Semiconductor, Inc.
+ * Copyright 2019 NXP
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2
+ * of the License, or (at your option) any later version.

No need to add this with the SPDX tag.  And btw, the text here is
GPL-2.0+ instead of GPL-2.0.
[Biwen Li] ok,I will delete this and replace GPL-2.0 with GPL-2.0+ in v3.
+ */
+
+#include <linux/device.h>
+#include <linux/err.h>
+#include <linux/interrupt.h>
+#include <linux/io.h>
+#include <linux/of_address.h>
+#include <linux/of_irq.h>
+#include <linux/platform_device.h>
+#include <linux/of.h>
+#include <linux/of_device.h>
+#include <linux/module.h>
+#include <linux/fsl/ftm.h>
+#include <linux/rtc.h>
+#include <linux/time.h>
+
+#define FTM_SC_CLK(c)		((c) << FTM_SC_CLK_MASK_SHIFT)
+
+/*
+ * Select Fixed frequency clock as clock source
+ * of FlexTimer Module
+ */
+#define FTM_SC_CLKS_FIXED_FREQ	0x02
+
+#define FIXED_FREQ_CLK		32000
+#define MAX_FREQ_DIV		(1 << FTM_SC_PS_MASK)
+#define MAX_COUNT_VAL		0xffff
+
+struct ftm_rtc {
+	struct rtc_device *rtc_dev;
+	void __iomem *base;
+	bool endian;

Make it big_endian to be more clear.
[Biwen Li] yes, you are right.
+	u32 alarm_freq;
+};
+
+static inline u32 rtc_readl(struct ftm_rtc *dev, u32 reg) {
+	if (dev->endian)
+		return ioread32be(dev->base + reg); /*big endianness*/

No need to have the extra comment with new variable name.

And probably better to use "else" here to make clearer to
compiler/static analyser.
[Biwen Li] ok, I will remove the extra comment and add else branch in v3.

+
+	return ioread32(dev->base + reg);
+}
+
+static inline void rtc_writel(struct ftm_rtc *dev, u32 reg, u32 val) {
+	if (dev->endian)
+		iowrite32be(val, dev->base + reg);
+	else
+		iowrite32(val, dev->base + reg);
+}
+
+static inline void ftm_counter_enable(struct ftm_rtc *rtc) {
+	u32 val;
+
+	/* select and enable counter clock source */
+	val = rtc_readl(rtc, FTM_SC);
+	val &= ~(FTM_SC_PS_MASK | FTM_SC_CLK_MASK);
+	val |= (FTM_SC_PS_MASK | FTM_SC_CLK(FTM_SC_CLKS_FIXED_FREQ));
+	rtc_writel(rtc, FTM_SC, val);
+}
+
+static inline void ftm_counter_disable(struct ftm_rtc *rtc) {
+	u32 val;
+
+	/* disable counter clock source */
+	val = rtc_readl(rtc, FTM_SC);
+	val &= ~(FTM_SC_PS_MASK | FTM_SC_CLK_MASK);
+	rtc_writel(rtc, FTM_SC, val);
+}
+
+static inline void ftm_irq_acknowledge(struct ftm_rtc *rtc) {
+	unsigned int timeout = 100;
+
+	while ((FTM_SC_TOF & rtc_readl(rtc, FTM_SC)) && timeout--)
+		rtc_writel(rtc, FTM_SC, rtc_readl(rtc, FTM_SC) & (~FTM_SC_TOF)); 

Is there a reason that we need to loop here?
[Biwen Li] It is used to fix errata A-007728 for flextimer.i will add a reason for it in v3. 
[Biwen Li] If the FTM counter reaches the FTM_MOD value between the reading of the
[Biwen Li] TOF bit and the writing of 0 to the TOF bit, the process of clearing the
[Biwen Li] TOF bit does not work as expected when FTMx_CONF[NUMTOF] != 0 and the
[Biwen Li] current TOF count is less than FTMx_CONF[NUMTOF]. If the above condition
[Biwen Li] is met, the TOF bit remains set. If the TOF interrupt is enabled
[Biwen Li] (FTMx_SC[TOIE] = 1), the TOF interrupt also remains asserted.

[Biwen Li] Above is the errata discription

[Biwen Li] In one word: software clearing TOF bit not works when FTMx_CONF[NUMTOF]
[Biwen Li] was seted as nonzero and FTM counter reaches the FTM_MOD value.

[Biwen Li] The workaround is clearing TOF bit until it works(FTM counter doesn't
[Biwen Li] always reache the FTM_MOD anyway),which may cost some cycles.
 
[Biwen Li] More info is here https://lore.kernel.org/patchwork/patch/782120/
+}
+
+static inline void ftm_irq_enable(struct ftm_rtc *rtc) {
+	u32 val;
+
+	val = rtc_readl(rtc, FTM_SC);
+	val |= FTM_SC_TOIE;
+	rtc_writel(rtc, FTM_SC, val);
+}
+
+static inline void ftm_irq_disable(struct ftm_rtc *rtc) {
+	u32 val;
+
+	val = rtc_readl(rtc, FTM_SC);
+	val &= ~FTM_SC_TOIE;
+	rtc_writel(rtc, FTM_SC, val);
+}
+
+static inline void ftm_reset_counter(struct ftm_rtc *rtc) {
+	/*
+	 * The CNT register contains the FTM counter value.
+	 * Reset clears the CNT register. Writing any value to COUNT
+	 * updates the counter with its initial value, CNTIN.
+	 */
+	rtc_writel(rtc, FTM_CNT, 0x00);
+}
+
+static void ftm_clean_alarm(struct ftm_rtc *rtc) {
+	ftm_counter_disable(rtc);
+
+	rtc_writel(rtc, FTM_CNTIN, 0x00);
+	rtc_writel(rtc, FTM_MOD, ~0U);
+
+	ftm_reset_counter(rtc);
+}
+
+static irqreturn_t ftm_rtc_alarm_interrupt(int irq, void *dev) {
+	struct ftm_rtc *rtc = dev;
+
+	ftm_irq_acknowledge(rtc);
+	ftm_irq_disable(rtc);
+	ftm_clean_alarm(rtc);
+
+	return IRQ_HANDLED;
+}
+
+static int ftm_rtc_alarm_irq_enable(struct device *dev,
+		unsigned int enabled)
+{
+	struct ftm_rtc *rtc = dev_get_drvdata(dev);
+
+	if (enabled)
+		ftm_irq_enable(rtc);
+	else
+		ftm_irq_disable(rtc);
+
+	return 0;
+}
+
+/*
+ * Note:
+ *	The function is not really getting time from the RTC
+ *	since FlexTimer is not a RTC device, but we need to
+ *	get time to setup alarm, so we are using system time
+ *	for now.
+ */
+static int ftm_rtc_read_time(struct device *dev, struct rtc_time *tm) {
+	struct timespec64 ts64;
+	unsigned long local_time;
+
+	ktime_get_real_ts64(&ts64);
+	local_time = (unsigned long)(ts64.tv_sec - (sys_tz.tz_minuteswest * 
+60));
+
+	rtc_time_to_tm(local_time, tm);
+
+	return 0;
+}
New line here.
[Biwen Li] yes. I will add new line here in v3. 
+static int ftm_rtc_read_alarm(struct device *dev, struct rtc_wkalrm 
+*alm) {
+	return 0;
Why we cannot read the alarm time?
[Biwen Li] The FlexTimer is not a real rtc device.it don't have alarm registers of rtc.
+}
+
+/*250Hz, 65536 / 250 = 262 second max*/ 
Can you make the comment more readable?
[Biwen Li] yes,I will add some comment in v3.
+static int ftm_rtc_set_alarm(struct device *dev, struct rtc_wkalrm *alm) {
+	struct rtc_time tm;
+	unsigned long now, alm_time, cycle;
+	struct ftm_rtc *rtc = dev_get_drvdata(dev);
+
+	ftm_rtc_read_time(dev, &tm);
+	rtc_tm_to_time(&tm, &now);
+	rtc_tm_to_time(&alm->time, &alm_time);
+
+	ftm_clean_alarm(rtc);
+	cycle = (alm_time - now) * rtc->alarm_freq;
+	if (cycle > MAX_COUNT_VAL) {
+		pr_err("Out of alarm range.\n");
Better to tell the users what the range is than letting them try out.
[Biwen Li] ok, I will add the range in v3.
+		return -EINVAL;
+	}
+
+	ftm_irq_disable(rtc);
+
+	/*
+	 * The counter increments until the value of MOD is reached,
+	 * at which point the counter is reloaded with the value of CNTIN.
+	 * The TOF (the overflow flag) bit is set when the FTM counter
+	 * changes from MOD to CNTIN. So we should using the cycle - 1.
+	 */
+	rtc_writel(rtc, FTM_MOD, cycle - 1);
+
+	ftm_counter_enable(rtc);
+	ftm_irq_enable(rtc);
+
+	return 0;
+
+}
+
+static const struct rtc_class_ops ftm_rtc_ops = {
+	.read_time		= ftm_rtc_read_time,
+	.read_alarm		= ftm_rtc_read_alarm,
+	.set_alarm		= ftm_rtc_set_alarm,
+	.alarm_irq_enable	= ftm_rtc_alarm_irq_enable,
+};
New line is needed.
[Biwen Li] ok, I will add new line here in v3.
+static int ftm_rtc_probe(struct platform_device *pdev) {
+	struct device_node *np = pdev->dev.of_node;
+	struct resource *r;
+	int irq;
+	int ret;
+	struct ftm_rtc *rtc;
+
+	rtc = devm_kzalloc(&pdev->dev, sizeof(*rtc), GFP_KERNEL);
+	if (unlikely(!rtc)) {
+		pr_err("ftm: cannot alloc memery for rtc\n");
+		return -ENOMEM;
+	}
+
+
No two new lines.
[Biwen Li] I will remove one line here in v3.
+	platform_set_drvdata(pdev, rtc);
+
+	r = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	if (!r) {
+		pr_err("ftm: cannot get resource for rtc\n");
+		return -ENODEV;
+	}
+
+	rtc->base = devm_ioremap_resource(&pdev->dev, r);
+	if (IS_ERR(rtc->base)) {
+		pr_err("ftm: cannot ioremap resource for rtc\n");
+		return PTR_ERR(rtc->base);
+	}
+
+	irq = irq_of_parse_and_map(np, 0);
+	if (irq <= 0) {
+		pr_err("ftm: unable to get IRQ from DT, %d\n", irq);
+		return -EINVAL;
+	}
+
+	rtc->endian = of_property_read_bool(np, "big-endian");
+
+	ret = devm_request_irq(&pdev->dev, irq, ftm_rtc_alarm_interrupt,
+			       IRQF_NO_SUSPEND, dev_name(&pdev->dev), rtc);
+	if (ret < 0) {
+		dev_err(&pdev->dev, "failed to request irq\n");
+		return ret;
+	}
+
+	rtc->alarm_freq = (u32)FIXED_FREQ_CLK / (u32)MAX_FREQ_DIV;
+
+	ftm_clean_alarm(rtc);
+
+	device_init_wakeup(&pdev->dev, true);
+	rtc->rtc_dev = devm_rtc_device_register(&pdev->dev, "ftm-alarm",
+							&ftm_rtc_ops,
+							THIS_MODULE);
+	if (IS_ERR(rtc->rtc_dev)) {
+		dev_err(&pdev->dev, "can't register rtc device\n");
+		return PTR_ERR(rtc->rtc_dev);
+	}
+	return ret;
+}
+
+static const struct of_device_id ftm_rtc_match[] = {
+	{ .compatible = "fsl,ftm-alarm", },
+	{ .compatible = "fsl,ls1012a-ftm-alarm", },
+	{ .compatible = "fsl,ls1021a-ftm-alarm", },
+	{ .compatible = "fsl,ls1043a-ftm-alarm", },
+	{ .compatible = "fsl,ls1046a-ftm-alarm", },
+	{ .compatible = "fsl,ls1088a-ftm-alarm", },
+	{ .compatible = "fsl,ls208xa-ftm-alarm", },
+	{ .compatible = "fsl,ls1028a-ftm-alarm", },
+	{ },
+};
+
+static struct platform_driver ftm_rtc_driver = {
+	.probe		= ftm_rtc_probe,
+	.driver		= {
+		.name	= "ftm-alarm",
+		.of_match_table = ftm_rtc_match,
+	},
+};
+
+static int __init ftm_alarm_init(void)
+{
+	return platform_driver_register(&ftm_rtc_driver);
+}
+
+/***************
This is not standard Linux multiline comment style.
[Biwen Li] ok, I will correct it in v3.
+ *Ensure that the driver is initialized after
*any real rtc driver
Can you explain why it has to be initialized later than other rtc drivers?
[Biwen Li] yes, I will add comment to explain this in v3.
[Biwen Li]		- The flextimer is not a real rtc device,it don't have time and date registers of rtc . 
[Biwen Li] 	- The flextimer rtc alarm driver gets time from wall time,
[Biwen Li]		 but the wall time is not ready.so the time from the driver is wrong.
[Biwen Li]		- If system regist it before any other real rtc device,it will be emulated as rtc0,
[Biwen Li]		 date command will read wrong time for user.
*/
+device_initcall_sync(ftm_alarm_init);
+
+MODULE_DESCRIPTION("NXP/Freescale FlexTimer alarm driver"); 
+MODULE_AUTHOR("Biwen Li <biwen.li@xxxxxxx>"); MODULE_LICENSE("GPL");
--
2.7.4