Hello Alexandre,

Le mar. 20 juil. 2021 à 22:01, Alexandre Belloni
[off-list ref] a écrit :

Hello Romain,

On 20/07/2021 19:22:50+0200, Romain Perier wrote:

quoted

From: Daniel Palmer <redacted>

This adds support for the RTC block on the Mstar MSC313e SoCs and newer.

Signed-off-by: Daniel Palmer <redacted>
Co-developed-by: Romain Perier <romain.perier@gmail.com>
Signed-off-by: Romain Perier <romain.perier@gmail.com>
---
 MAINTAINERS              |   1 +
 drivers/rtc/Kconfig      |  10 ++
 drivers/rtc/Makefile     |   1 +
 drivers/rtc/rtc-msc313.c | 246 +++++++++++++++++++++++++++++++++++++++
 4 files changed, 258 insertions(+)
 create mode 100644 drivers/rtc/rtc-msc313.c

diff --git a/MAINTAINERS b/MAINTAINERS
index 6c8be735cc91..7e8d1a375e0d 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS

@@ -2220,6 +2220,7 @@ F:      arch/arm/boot/dts/mstar-*
 F:   arch/arm/mach-mstar/
 F:   drivers/clk/mstar/
 F:   drivers/gpio/gpio-msc313.c
+F:   drivers/rtc/rtc-msc313.c
 F:   drivers/watchdog/msc313e_wdt.c
 F:   include/dt-bindings/clock/mstar-*
 F:   include/dt-bindings/gpio/msc313-gpio.h

diff --git a/drivers/rtc/Kconfig b/drivers/rtc/Kconfig
index 12153d5801ce..67870b422bc5 100644
--- a/drivers/rtc/Kconfig
+++ b/drivers/rtc/Kconfig

@@ -1925,4 +1925,14 @@ config RTC_DRV_WILCO_EC
        This can also be built as a module. If so, the module will
        be named "rtc_wilco_ec".

+config RTC_DRV_MSC313
+     tristate "MStar MSC313 RTC"
+        depends on ARCH_MSTARV7

|| COMPILE_TEST maybe ?

Arf, I always forget this option. Ack, I will add it in v2.

quoted

+     help
+       If you say yes here you get support for the Mstar MSC313e On-Chip
+       Real Time Clock.
+
+       This driver can also be built as a module, if so, the module
+       will be called "rtc-msc313".
+
 endif # RTC_CLASS

quoted

+static int msc313_rtc_read_time(struct device *dev, struct rtc_time *tm)
+{
+     struct msc313_rtc *priv = dev_get_drvdata(dev);
+     u32 seconds;
+     u16 reg;
+
+     reg = readw(priv->rtc_base + REG_RTC_CTRL);
+     writew(reg | READ_EN_BIT, priv->rtc_base + REG_RTC_CTRL);
+
+     /* Wait for HW latch done */
+     while (readw(priv->rtc_base + REG_RTC_CTRL) & READ_EN_BIT)
+             udelay(1);
+
+     seconds = readw(priv->rtc_base + REG_RTC_CNT_VAL_L)
+                     | (readw(priv->rtc_base + REG_RTC_CNT_VAL_H) << 16);
+
+     rtc_time64_to_tm(seconds, tm);
+
+     return rtc_valid_tm(tm);

This is not necessary, tm is valid at that point (and the core will
check anyway).

ack.

quoted

+}
+
+static int msc313_rtc_set_time(struct device *dev, struct rtc_time *tm)
+{
+     struct msc313_rtc *priv = dev_get_drvdata(dev);
+     unsigned long seconds;
+     u16 reg;
+
+     seconds = rtc_tm_to_time64(tm);
+     writew(seconds & 0xFFFF, priv->rtc_base + REG_RTC_LOAD_VAL_L);
+     writew((seconds >> 16) & 0xFFFF, priv->rtc_base + REG_RTC_LOAD_VAL_H);
+     reg = readw(priv->rtc_base + REG_RTC_CTRL);
+     writew(reg | LOAD_EN_BIT, priv->rtc_base + REG_RTC_CTRL);
+
+     /* need to check carefully if we want to clear REG_RTC_LOAD_VAL_H for customer*/
+     while (readw(priv->rtc_base + REG_RTC_CTRL) & LOAD_EN_BIT)
+             udelay(1);
+     writew(0, priv->rtc_base + REG_RTC_LOAD_VAL_H);

Why is that necessary? The comment is not super useful here.

quoted

+
+     return 0;
+}
+
+static const struct rtc_class_ops msc313_rtc_ops = {
+     .read_time = msc313_rtc_read_time,
+     .set_time = msc313_rtc_set_time,
+     .read_alarm = msc313_rtc_read_alarm,
+     .set_alarm = msc313_rtc_set_alarm,
+     .alarm_irq_enable = msc313_rtc_alarm_irq_enable,
+};
+
+static irqreturn_t msc313_rtc_interrupt(s32 irq, void *dev_id)
+{
+     struct msc313_rtc *priv = dev_get_drvdata(dev_id);
+     u16 reg;
+
+     reg = readw_relaxed(priv->rtc_base + REG_RTC_CTRL);
+     reg |= INT_CLEAR_BIT;
+     reg &= ~INT_FORCE_BIT;
+     writew_relaxed(reg, priv->rtc_base + REG_RTC_CTRL);
+

I'm not convinced the _relaxed functions are doing the right thing here.

Good catch for the _relaxed, I did not pay attention during refactoring.

Also, shouldn't you check the alarm actually fired?

You mean before set the flag "RTC_AF" ? This driver is mostly based on
reverse engineering,
even if we have some hardware info. I will try to find if it is
possible (I mean,  from the info
I have).

quoted

+     rtc_update_irq(priv->rtc_dev, 1, RTC_IRQF | RTC_AF);
+
+     return IRQ_HANDLED;
+}
+
+static int msc313_rtc_remove(struct platform_device *pdev)
+{
+     struct msc313_rtc *priv = platform_get_drvdata(pdev);
+
+     clk_disable_unprepare(priv->clk);

With a nice devm_add_action_or_reset() in the probe, you can remove the
need for msc313_rtc_remove().

devm_add_action_or_reset() you said, interesting I did not know this function.

quoted

+
+     return 0;
+}
+
+static int msc313_rtc_probe(struct platform_device *pdev)
+{
+     struct device *dev = &pdev->dev;
+     struct msc313_rtc *priv;
+     int ret;
+     int irq;
+     unsigned long rate;
+     u16 reg;
+
+     priv = devm_kzalloc(&pdev->dev, sizeof(struct msc313_rtc), GFP_KERNEL);
+     if (!priv)
+             return -ENOMEM;
+
+     priv->rtc_base = devm_platform_ioremap_resource(pdev, 0);
+     if (IS_ERR(priv->rtc_base))
+             return PTR_ERR(priv->rtc_base);
+
+     irq = platform_get_irq(pdev, 0);
+     if (irq < 0)
+             return -EINVAL;
+
+     ret = devm_request_irq(dev, irq, msc313_rtc_interrupt, IRQF_SHARED,
+                            dev_name(&pdev->dev), &pdev->dev);
+     if (ret) {
+             dev_err(dev, "Unable to request irq\n");
+             return ret;
+     }
+
+     priv->clk = devm_clk_get(dev, NULL);
+     if (IS_ERR(priv->clk)) {
+             dev_err(dev, "No input reference clock\n");
+             return PTR_ERR(priv->clk);
+     }
+
+     ret = clk_prepare_enable(priv->clk);
+     if (ret) {
+             dev_err(dev, "Failed to enable the reference clock, %d\n", ret);
+             return ret;
+     }
+
+     rate = clk_get_rate(priv->clk);
+
+     reg = readw(priv->rtc_base + REG_RTC_CTRL);
+     if (!(reg & SOFT_RSTZ_BIT)) {
+             reg |= SOFT_RSTZ_BIT;
+             writew(reg, priv->rtc_base + REG_RTC_CTRL);
+     }

What is the meaning of this bit? I would think it is better to use that
to know whether the RTC holds the correct time instead of killing the
info here.

quoted

+
+     writew(rate & 0xFFFF, priv->rtc_base + REG_RTC_FREQ_CW_L);
+     writew((rate >> 16) & 0xFFFF, priv->rtc_base + REG_RTC_FREQ_CW_H);
+
+     reg |= CNT_EN_BIT;
+     writew(reg, priv->rtc_base + REG_RTC_CTRL);
+
+     platform_set_drvdata(pdev, priv);
+
+     priv->rtc_dev = devm_rtc_device_register(dev, dev_name(dev), &msc313_rtc_ops, THIS_MODULE);
+     if (IS_ERR(priv->rtc_dev)) {
+             dev_err(dev, "Failed to register rtc device\n");
+             return PTR_ERR(priv->rtc_dev);
+     }

Please switch to devm_rtc_allocate_device and devm_rtc_register_device.
Also drop the error message, it is not necessary.

You must also set the RTC range. To help you, you can use:
https://git.kernel.org/pub/scm/linux/kernel/git/abelloni/rtc-tools.git/tree/rtc-range.c
then you must update the DT bindings as the RTC will support the
start-year property

You must also run rtctest and should include the results.

Ack, I will do this in v2.

Thanks!