linux-rtc.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH v2 0/9] platform/chrome: rtc: Add support for Wilco EC
@ 2019-01-14 22:03 Nick Crews
  2019-01-14 22:03 ` [PATCH v2 5/9] platform/chrome: rtc: Add RTC driver " Nick Crews
  0 siblings, 1 reply; 5+ messages in thread
From: Nick Crews @ 2019-01-14 22:03 UTC (permalink / raw)
  To: linux-kernel
  Cc: groeck, sjg, dlaurie, Nick Crews, linux-rtc,
	Enric Balletbo i Serra, Alessandro Zummo, Benson Leung,
	Duncan Laurie, Alexandre Belloni


There is a new chromebook that contains a different Embedded Controller
(codename Wilco) than the rest of the chromebook series. Thus the kernel
requires a different driver than the already existing and generalized
cros_ec_* drivers. Specifically, this driver adds support for getting
and setting the RTC on the EC, adding a binary sysfs attribute
that receives ACPI events from the EC, adding a binary sysfs
attribute to request telemetry data from the EC (useful for enterprise
applications), and adding normal sysfs attributes to get/set various
other properties on the EC. The core of the communication with the EC
is implemented in wilco_ec/mailbox.c, using a simple byte-level protocol
with a checksum, transmitted over an eSPI bus. For debugging purposes,
a raw attribute is also provided which can write/read arbitrary
bytes to/from the eSPI bus.

We attempted to adhere to the sysfs principles of "one piece of data per
attribute" as much as possible, and mostly succeded. However, with the
wilco_ec/adv_power.h attributes, which deal with scheduling power usage,
we found it most elegant to bundle setting event times for an entire day
into a single attribute, so at most you are using attributes formatted
as "%d %d %d %d %d %d". With the telemetry attribute, we had to use a
binary attribute, instead of the preferable human-readable ascii, in
order to keep secure the information which is proprietary to the
enterprise service provider. This opaque binary data will be read and
sent using a proprietary daemon running on the OS. Finally, the
"version" attribute returns a formatted result that looks something
like:
> cat /sys/bus/platform/devices/GOOG000C\:00/version
Label        : 95.00.06
SVN Revision : 5960a.06
Model Number : 08;8
Build Date   : 11/29/18

The RTC driver is exposed as a standard RTC class driver with
read/write functionality.

For event notification, the Wilco EC can return extended events that
are not handled by standard ACPI objects. These events can
include hotkeys which map to standard functions like brightness
controls, or information about EC controlled features like the
charger or battery. These events are triggered with an
ACPI Notify(0x90) and the event data buffer is read through an ACPI
method provided by the BIOS which reads the event buffer from EC RAM.
These events are then processed, with hotkey events being sent
to the input subsystem and other events put into a queue which
can be read by a userspace daemon via a sysfs attribute.

The rest of the attributes are categorized as either "properties" or
"legacy". "legacy" implies that the attribute existed on the EC before it
was modified for ChromeOS, and "properties" implies that the attribute
exposes functionality that was added to the EC specifically for
ChromeOS. They are mostly boolean flags or percentages.

A full thread of the development of these patches can be found at
https://chromium-review.googlesource.com/c/1371034. This thread contains
comments and revisions that could be helpful in understanding how the
driver arrived at the state it is in now. The thread also contains some
ChromeOS specific patches that actually enable the driver. If you want
to test the patch yourself, you would have to install the ChromeOS SDK
and cherry pick in these patches.

I also wrote some integration tests using the Tast testing framework that
ChromeOS uses. It would require a full ChromeOS SDK to actually run the
tests, but the source of the tests, written in Go, are useful for
understanding what the desired behavior is. You can view the tests here:
https://chromium-review.googlesource.com/c/1372575

Thank you for your comments!

Changes in v2:
- Fixed kernel-doc comments
- Fixed include of linux/mfd/cros_ec_lpc_mec.h
- cros_ec_lpc_mec_in_range() returns -EINVAL on error
- Added parens around macro variables
- Removed COMPILE_TEST from Kconfig because inb()/outb()
won't work on anything but X86
- Moved everything to wilco_ec/ subdirectory
- Moved header file to include/platform_data/
so could be used by future sub-drivers
- Tweaked mailbox()
- Removed duplicate EC_MAILBOX_DATA_SIZE defs
- Removed some error messages / converted to dbg()
- Remove license boiler plate
- Remove "wilco_ec_sysfs -" docstring prefix
- Fix accidental Makefile deletion
- Add documentation for sysfs entries
- Change "enable ? 0 : 1" to "!enable"
- No longer override error code from sysfs_init()
- Put attributes in the legacy file to begin with, don't move later
- Remove duplicate error messages in sysfs_init() and its caller
- Add sysfs documentation
- rm duplicate EC_MAILBOX_DATA_SIZE defs
- Make docstrings follow kernel style
- Fix tags in commit msg
- Reading raw now includes ASCII translation
- rm license boiler plate
- rm "wilco_ec_rtc -" prefix in docstring
- Make rtc driver its own module within the drivers/rtc/ directory
- Register a rtc device from core.c that is picked up by this driver
- rm unused rtc_sync() function
- rm "wilco_ec_event -" prefix from docstring
- rm license boiler plate
- Add sysfs directory documentation
- Fix cosmetics
- Remove duplicate error messages
- rm license boiler plate
- rm "wilco_ec_properties -" prefix from docstring
- Add documentation
- rm license boiler plate
- rm "wilco_ec_adv_power - " prefix from docstring
- Add documentation
- Made format strings in store() and read() functions static
- rm "wilco_ec_telemetry - " prefix from docstring
- rm license boiler plate
- Fix commit msg tag

Duncan Laurie (6):
  platform/chrome: Remove cros_ec dependency in lpc_mec
  platform/chrome: Add new driver for Wilco EC
  platform/chrome: Add sysfs attributes
  platform/chrome: Add support for raw commands in sysfs
  platform/chrome: rtc: Add RTC driver for Wilco EC
  platform/chrome: Add event handling

Nick Crews (3):
  platform/chrome: Add EC properties
  platform/chrome: Add peakshift and adv_batt_charging
  platform/chrome: Add binary telemetry attributes

 .../ABI/testing/sysfs-platform-wilcoec        | 176 ++++++
 drivers/platform/chrome/Kconfig               |   4 +-
 drivers/platform/chrome/Makefile              |   2 +
 drivers/platform/chrome/cros_ec_lpc_mec.c     |  52 +-
 drivers/platform/chrome/cros_ec_lpc_mec.h     |  43 +-
 drivers/platform/chrome/cros_ec_lpc_reg.c     |  43 +-
 drivers/platform/chrome/wilco_ec/Kconfig      |  33 ++
 drivers/platform/chrome/wilco_ec/Makefile     |   6 +
 drivers/platform/chrome/wilco_ec/adv_power.c  | 544 ++++++++++++++++++
 drivers/platform/chrome/wilco_ec/adv_power.h  | 183 ++++++
 drivers/platform/chrome/wilco_ec/core.c       | 144 +++++
 drivers/platform/chrome/wilco_ec/event.c      | 347 +++++++++++
 drivers/platform/chrome/wilco_ec/legacy.c     | 190 ++++++
 drivers/platform/chrome/wilco_ec/legacy.h     |  93 +++
 drivers/platform/chrome/wilco_ec/mailbox.c    | 237 ++++++++
 drivers/platform/chrome/wilco_ec/properties.c | 344 +++++++++++
 drivers/platform/chrome/wilco_ec/properties.h | 179 ++++++
 drivers/platform/chrome/wilco_ec/sysfs.c      | 238 ++++++++
 drivers/platform/chrome/wilco_ec/telemetry.c  |  66 +++
 drivers/platform/chrome/wilco_ec/telemetry.h  |  41 ++
 drivers/platform/chrome/wilco_ec/util.h       |  38 ++
 drivers/rtc/Kconfig                           |  11 +
 drivers/rtc/Makefile                          |   1 +
 drivers/rtc/rtc-wilco-ec.c                    | 178 ++++++
 include/linux/platform_data/wilco-ec.h        | 186 ++++++
 25 files changed, 3319 insertions(+), 60 deletions(-)
 create mode 100644 Documentation/ABI/testing/sysfs-platform-wilcoec
 create mode 100644 drivers/platform/chrome/wilco_ec/Kconfig
 create mode 100644 drivers/platform/chrome/wilco_ec/Makefile
 create mode 100644 drivers/platform/chrome/wilco_ec/adv_power.c
 create mode 100644 drivers/platform/chrome/wilco_ec/adv_power.h
 create mode 100644 drivers/platform/chrome/wilco_ec/core.c
 create mode 100644 drivers/platform/chrome/wilco_ec/event.c
 create mode 100644 drivers/platform/chrome/wilco_ec/legacy.c
 create mode 100644 drivers/platform/chrome/wilco_ec/legacy.h
 create mode 100644 drivers/platform/chrome/wilco_ec/mailbox.c
 create mode 100644 drivers/platform/chrome/wilco_ec/properties.c
 create mode 100644 drivers/platform/chrome/wilco_ec/properties.h
 create mode 100644 drivers/platform/chrome/wilco_ec/sysfs.c
 create mode 100644 drivers/platform/chrome/wilco_ec/telemetry.c
 create mode 100644 drivers/platform/chrome/wilco_ec/telemetry.h
 create mode 100644 drivers/platform/chrome/wilco_ec/util.h
 create mode 100644 drivers/rtc/rtc-wilco-ec.c
 create mode 100644 include/linux/platform_data/wilco-ec.h

-- 
2.20.1.97.g81188d93c3-goog


^ permalink raw reply	[flat|nested] 5+ messages in thread

* [PATCH v2 5/9] platform/chrome: rtc: Add RTC driver for Wilco EC
  2019-01-14 22:03 [PATCH v2 0/9] platform/chrome: rtc: Add support for Wilco EC Nick Crews
@ 2019-01-14 22:03 ` Nick Crews
  2019-01-14 22:25   ` Alexandre Belloni
  0 siblings, 1 reply; 5+ messages in thread
From: Nick Crews @ 2019-01-14 22:03 UTC (permalink / raw)
  To: linux-kernel
  Cc: groeck, sjg, dlaurie, Duncan Laurie, Nick Crews, linux-rtc,
	Enric Balletbo i Serra, Alessandro Zummo, Benson Leung,
	Alexandre Belloni

From: Duncan Laurie <dlaurie@google.com>

This Embedded Controller has an internal RTC that is exposed
as a standard RTC class driver with read/write functionality.

The driver is added to the drivers/rtc/ so that the maintainer of that
directory will be able to comment on this change, as that maintainer is
the expert on this system. In addition, the driver code is called
indirectly after a corresponding device is registered from core.c,
as opposed to core.c registering the driver callbacks directly.

> hwclock --show --rtc /dev/rtc1
2007-12-31 16:01:20.460959-08:00
> hwclock --systohc --rtc /dev/rtc1
> hwclock --show --rtc /dev/rtc1
2018-11-29 17:08:00.780793-08:00

Signed-off-by: Duncan Laurie <dlaurie@google.com>
Signed-off-by: Nick Crews <ncrews@google.com>
---

Changes in v2:
- rm license boiler plate
- rm "wilco_ec_rtc -" prefix in docstring
- Make rtc driver its own module within the drivers/rtc/ directory
- Register a rtc device from core.c that is picked up by this driver
- rm unused rtc_sync() function

 drivers/platform/chrome/wilco_ec/core.c |  18 +++
 drivers/rtc/Kconfig                     |  11 ++
 drivers/rtc/Makefile                    |   1 +
 drivers/rtc/rtc-wilco-ec.c              | 178 ++++++++++++++++++++++++
 4 files changed, 208 insertions(+)
 create mode 100644 drivers/rtc/rtc-wilco-ec.c

diff --git a/drivers/platform/chrome/wilco_ec/core.c b/drivers/platform/chrome/wilco_ec/core.c
index 3193e8b2ec91..61dacd582582 100644
--- a/drivers/platform/chrome/wilco_ec/core.c
+++ b/drivers/platform/chrome/wilco_ec/core.c
@@ -37,6 +37,7 @@ static int wilco_ec_probe(struct platform_device *pdev)
 {
 	struct device *dev = &pdev->dev;
 	struct wilco_ec_device *ec;
+	struct platform_device *child_pdev;
 	int ret;
 
 	ec = devm_kzalloc(dev, sizeof(*ec), GFP_KERNEL);
@@ -70,8 +71,25 @@ static int wilco_ec_probe(struct platform_device *pdev)
 		dev_err(dev, "Failed to create sysfs attributes\n");
 		goto destroy_lpc_mec;
 	}
+
+	/*
+	 * Register a RTC platform device that will get picked up by the RTC
+	 * subsystem. This platform device will get freed when its parent device
+	 * dev is unregistered.
+	 */
+	child_pdev = platform_device_register_data(dev, "rtc-wilco-ec",
+						   PLATFORM_DEVID_AUTO,
+						   NULL, 0);
+	if (IS_ERR(child_pdev)) {
+		dev_err(dev, "Failed to create RTC platform device\n");
+		ret = PTR_ERR(child_pdev);
+		goto remove_sysfs;
+	}
+
 	return 0;
 
+remove_sysfs:
+	wilco_ec_sysfs_remove(ec);
 destroy_lpc_mec:
 	cros_ec_lpc_mec_destroy();
 	return ret;
diff --git a/drivers/rtc/Kconfig b/drivers/rtc/Kconfig
index 225b0b8516f3..d5063c791515 100644
--- a/drivers/rtc/Kconfig
+++ b/drivers/rtc/Kconfig
@@ -1814,4 +1814,15 @@ config RTC_DRV_GOLDFISH
 	  Goldfish is a code name for the virtual platform developed by Google
 	  for Android emulation.
 
+config RTC_DRV_WILCO_EC
+	tristate "Wilco EC RTC"
+	depends on WILCO_EC
+	default m
+	help
+	  If you say yes here, you get read/write support for the Real Time
+	  Clock on the Wilco Embedded Controller (Wilco is a kind of Chromebook)
+
+	  This can also be built as a module. If so, the module will
+	  be named "rtc_wilco_ec".
+
 endif # RTC_CLASS
diff --git a/drivers/rtc/Makefile b/drivers/rtc/Makefile
index df022d820bee..6255ea78da25 100644
--- a/drivers/rtc/Makefile
+++ b/drivers/rtc/Makefile
@@ -172,6 +172,7 @@ obj-$(CONFIG_RTC_DRV_V3020)	+= rtc-v3020.o
 obj-$(CONFIG_RTC_DRV_VR41XX)	+= rtc-vr41xx.o
 obj-$(CONFIG_RTC_DRV_VRTC)	+= rtc-mrst.o
 obj-$(CONFIG_RTC_DRV_VT8500)	+= rtc-vt8500.o
+obj-$(CONFIG_RTC_DRV_WILCO_EC)	+= rtc-wilco-ec.o
 obj-$(CONFIG_RTC_DRV_WM831X)	+= rtc-wm831x.o
 obj-$(CONFIG_RTC_DRV_WM8350)	+= rtc-wm8350.o
 obj-$(CONFIG_RTC_DRV_X1205)	+= rtc-x1205.o
diff --git a/drivers/rtc/rtc-wilco-ec.c b/drivers/rtc/rtc-wilco-ec.c
new file mode 100644
index 000000000000..d8ed791da82d
--- /dev/null
+++ b/drivers/rtc/rtc-wilco-ec.c
@@ -0,0 +1,178 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * RTC interface for Wilco Embedded Controller with R/W abilities
+ *
+ * Copyright 2018 Google LLC
+ *
+ * The corresponding platform device is typically registered in
+ * drivers/platform/chrome/wilco_ec/core.c
+ */
+
+#include <linux/bcd.h>
+#include <linux/err.h>
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/platform_device.h>
+#include <linux/platform_data/wilco-ec.h>
+#include <linux/rtc.h>
+#include <linux/timekeeping.h>
+
+#define DRV_NAME "rtc-wilco-ec"
+
+#define EC_COMMAND_CMOS			0x7c
+#define EC_CMOS_TOD_WRITE		0x02
+#define EC_CMOS_TOD_READ		0x08
+
+/**
+ * struct ec_rtc_read - Format of RTC returned by EC.
+ * @second: Second value (0..59)
+ * @minute: Minute value (0..59)
+ * @hour: Hour value (0..23)
+ * @day: Day value (1..31)
+ * @month: Month value (1..12)
+ * @year: Year value (full year % 100)
+ * @century: Century value (full year / 100)
+ *
+ * All values are presented in binary (not BCD).
+ */
+struct ec_rtc_read {
+	u8 second;
+	u8 minute;
+	u8 hour;
+	u8 day;
+	u8 month;
+	u8 year;
+	u8 century;
+} __packed;
+
+/**
+ * struct ec_rtc_write - Format of RTC sent to the EC.
+ * @param: EC_CMOS_TOD_WRITE
+ * @century: Century value (full year / 100)
+ * @year: Year value (full year % 100)
+ * @month: Month value (1..12)
+ * @day: Day value (1..31)
+ * @hour: Hour value (0..23)
+ * @minute: Minute value (0..59)
+ * @second: Second value (0..59)
+ * @weekday: Day of the week (0=Saturday)
+ *
+ * All values are presented in BCD.
+ */
+struct ec_rtc_write {
+	u8 param;
+	u8 century;
+	u8 year;
+	u8 month;
+	u8 day;
+	u8 hour;
+	u8 minute;
+	u8 second;
+	u8 weekday;
+} __packed;
+
+int wilco_ec_rtc_read(struct device *dev, struct rtc_time *tm)
+{
+	struct wilco_ec_device *ec = dev_get_drvdata(dev->parent);
+	u8 param = EC_CMOS_TOD_READ;
+	struct ec_rtc_read rtc;
+	struct wilco_ec_message msg = {
+		.type = WILCO_EC_MSG_LEGACY,
+		.flags = WILCO_EC_FLAG_RAW_RESPONSE,
+		.command = EC_COMMAND_CMOS,
+		.request_data = &param,
+		.request_size = sizeof(param),
+		.response_data = &rtc,
+		.response_size = sizeof(rtc),
+	};
+	struct rtc_time calc_tm;
+	unsigned long time;
+	int ret;
+
+	ret = wilco_ec_mailbox(ec, &msg);
+	if (ret < 0)
+		return ret;
+
+	tm->tm_sec	= rtc.second;
+	tm->tm_min	= rtc.minute;
+	tm->tm_hour	= rtc.hour;
+	tm->tm_mday	= rtc.day;
+	/*
+	 * The RTC stores the month value as 1-12 but the kernel expects 0-11,
+	 * so ensure invalid/zero month value from RTC is not converted to -1.
+	 */
+	tm->tm_mon	= rtc.month ? rtc.month - 1 : 0;
+	tm->tm_year	= rtc.year + (rtc.century * 100) - 1900;
+	tm->tm_yday	= rtc_year_days(tm->tm_mday, tm->tm_mon, tm->tm_year);
+
+	/* Compute day of week */
+	rtc_tm_to_time(tm, &time);
+	rtc_time_to_tm(time, &calc_tm);
+	tm->tm_wday = calc_tm.tm_wday;
+
+	return 0;
+}
+
+int wilco_ec_rtc_write(struct device *dev, struct rtc_time *tm)
+{
+	struct wilco_ec_device *ec = dev_get_drvdata(dev->parent);
+	struct ec_rtc_write rtc;
+	struct wilco_ec_message msg = {
+		.type = WILCO_EC_MSG_LEGACY,
+		.flags = WILCO_EC_FLAG_RAW_RESPONSE,
+		.command = EC_COMMAND_CMOS,
+		.request_data = &rtc,
+		.request_size = sizeof(rtc),
+	};
+	int year = tm->tm_year + 1900;
+	/* Convert from 0=Sunday to 0=Saturday for the EC */
+	int wday = tm->tm_wday == 6 ? 0 : tm->tm_wday + 1;
+	int ret;
+
+	rtc.param	= EC_CMOS_TOD_WRITE;
+	rtc.century	= bin2bcd(year / 100);
+	rtc.year	= bin2bcd(year % 100);
+	rtc.month	= bin2bcd(tm->tm_mon + 1);
+	rtc.day		= bin2bcd(tm->tm_mday);
+	rtc.hour	= bin2bcd(tm->tm_hour);
+	rtc.minute	= bin2bcd(tm->tm_min);
+	rtc.second	= bin2bcd(tm->tm_sec);
+	rtc.weekday	= bin2bcd(wday);
+
+	ret = wilco_ec_mailbox(ec, &msg);
+	if (ret < 0)
+		return ret;
+
+	return 0;
+}
+
+static const struct rtc_class_ops wilco_ec_rtc_ops = {
+	.read_time = wilco_ec_rtc_read,
+	.set_time = wilco_ec_rtc_write,
+};
+
+static int wilco_ec_rtc_probe(struct platform_device *pdev)
+{
+	struct rtc_device *rtc = devm_rtc_device_register(&pdev->dev,
+							  DRV_NAME,
+							  &wilco_ec_rtc_ops,
+							  THIS_MODULE);
+	if (IS_ERR(rtc))
+		return PTR_ERR(rtc);
+
+	return 0;
+}
+
+static struct platform_driver wilco_ec_rtc_driver = {
+	.driver = {
+		.name = DRV_NAME,
+	},
+	.probe = wilco_ec_rtc_probe,
+};
+
+module_platform_driver(wilco_ec_rtc_driver);
+
+MODULE_ALIAS("platform:" DRV_NAME);
+MODULE_AUTHOR("Nick Crews <ncrews@google.com>");
+MODULE_LICENSE("GPL v2");
+MODULE_DESCRIPTION("Wilco EC RTC driver");
-- 
2.20.1.97.g81188d93c3-goog


^ permalink raw reply related	[flat|nested] 5+ messages in thread

* Re: [PATCH v2 5/9] platform/chrome: rtc: Add RTC driver for Wilco EC
  2019-01-14 22:03 ` [PATCH v2 5/9] platform/chrome: rtc: Add RTC driver " Nick Crews
@ 2019-01-14 22:25   ` Alexandre Belloni
  2019-01-15  0:30     ` Nick Crews
  2019-01-18 21:24     ` Nick Crews
  0 siblings, 2 replies; 5+ messages in thread
From: Alexandre Belloni @ 2019-01-14 22:25 UTC (permalink / raw)
  To: Nick Crews
  Cc: linux-kernel, groeck, sjg, dlaurie, Duncan Laurie, linux-rtc,
	Enric Balletbo i Serra, Alessandro Zummo, Benson Leung

Hello,

On 14/01/2019 15:03:52-0700, Nick Crews wrote:
> diff --git a/drivers/rtc/rtc-wilco-ec.c b/drivers/rtc/rtc-wilco-ec.c
> new file mode 100644
> index 000000000000..d8ed791da82d
> --- /dev/null
> +++ b/drivers/rtc/rtc-wilco-ec.c
> @@ -0,0 +1,178 @@
> +// SPDX-License-Identifier: GPL-2.0
> +/*
> + * RTC interface for Wilco Embedded Controller with R/W abilities
> + *
> + * Copyright 2018 Google LLC
> + *
> + * The corresponding platform device is typically registered in
> + * drivers/platform/chrome/wilco_ec/core.c
> + */
> +
> +#include <linux/bcd.h>
> +#include <linux/err.h>
> +#include <linux/kernel.h>
> +#include <linux/module.h>
> +#include <linux/platform_device.h>
> +#include <linux/platform_data/wilco-ec.h>
> +#include <linux/rtc.h>
> +#include <linux/timekeeping.h>
> +
> +#define DRV_NAME "rtc-wilco-ec"

Please get rid of that define.

> +int wilco_ec_rtc_read(struct device *dev, struct rtc_time *tm)
> +{
> +	struct wilco_ec_device *ec = dev_get_drvdata(dev->parent);
> +	u8 param = EC_CMOS_TOD_READ;
> +	struct ec_rtc_read rtc;
> +	struct wilco_ec_message msg = {
> +		.type = WILCO_EC_MSG_LEGACY,
> +		.flags = WILCO_EC_FLAG_RAW_RESPONSE,
> +		.command = EC_COMMAND_CMOS,
> +		.request_data = &param,
> +		.request_size = sizeof(param),
> +		.response_data = &rtc,
> +		.response_size = sizeof(rtc),
> +	};
> +	struct rtc_time calc_tm;
> +	unsigned long time;
> +	int ret;
> +
> +	ret = wilco_ec_mailbox(ec, &msg);
> +	if (ret < 0)
> +		return ret;
> +
> +	tm->tm_sec	= rtc.second;
> +	tm->tm_min	= rtc.minute;
> +	tm->tm_hour	= rtc.hour;
> +	tm->tm_mday	= rtc.day;
> +	/*
> +	 * The RTC stores the month value as 1-12 but the kernel expects 0-11,
> +	 * so ensure invalid/zero month value from RTC is not converted to -1.

I'm pretty sure this is not necessary and will certainly make the core
think the date is valid when it has obviously been corrupted.

> +	 */
> +	tm->tm_mon	= rtc.month ? rtc.month - 1 : 0;
> +	tm->tm_year	= rtc.year + (rtc.century * 100) - 1900;
> +	tm->tm_yday	= rtc_year_days(tm->tm_mday, tm->tm_mon, tm->tm_year);
> +
> +	/* Compute day of week */
> +	rtc_tm_to_time(tm, &time);
> +	rtc_time_to_tm(time, &calc_tm);
> +	tm->tm_wday = calc_tm.tm_wday;

Do you really need an accurate wday? There are no userspace applications
that I know of that care and converting back and forth is quite
expensive.

> +
> +	return 0;
> +}
> +
> +int wilco_ec_rtc_write(struct device *dev, struct rtc_time *tm)
> +{
> +	struct wilco_ec_device *ec = dev_get_drvdata(dev->parent);
> +	struct ec_rtc_write rtc;
> +	struct wilco_ec_message msg = {
> +		.type = WILCO_EC_MSG_LEGACY,
> +		.flags = WILCO_EC_FLAG_RAW_RESPONSE,
> +		.command = EC_COMMAND_CMOS,
> +		.request_data = &rtc,
> +		.request_size = sizeof(rtc),
> +	};
> +	int year = tm->tm_year + 1900;
> +	/* Convert from 0=Sunday to 0=Saturday for the EC */
> +	int wday = tm->tm_wday == 6 ? 0 : tm->tm_wday + 1;

I'm not sure how useful it is to set an accurate wday when the RTC is
not able to give it back.

> +	int ret;
> +
> +	rtc.param	= EC_CMOS_TOD_WRITE;
> +	rtc.century	= bin2bcd(year / 100);
> +	rtc.year	= bin2bcd(year % 100);
> +	rtc.month	= bin2bcd(tm->tm_mon + 1);
> +	rtc.day		= bin2bcd(tm->tm_mday);
> +	rtc.hour	= bin2bcd(tm->tm_hour);
> +	rtc.minute	= bin2bcd(tm->tm_min);
> +	rtc.second	= bin2bcd(tm->tm_sec);
> +	rtc.weekday	= bin2bcd(wday);
> +
> +	ret = wilco_ec_mailbox(ec, &msg);
> +	if (ret < 0)
> +		return ret;
> +
> +	return 0;
> +}
> +
> +static const struct rtc_class_ops wilco_ec_rtc_ops = {
> +	.read_time = wilco_ec_rtc_read,
> +	.set_time = wilco_ec_rtc_write,
> +};
> +
> +static int wilco_ec_rtc_probe(struct platform_device *pdev)
> +{
> +	struct rtc_device *rtc = devm_rtc_device_register(&pdev->dev,
> +							  DRV_NAME,
> +							  &wilco_ec_rtc_ops,
> +							  THIS_MODULE);

Please use devm_rtc_allocate_device() and set rtc->range_min and
rtc->range_max before registering with rtc_register_device().

> +	if (IS_ERR(rtc))
> +		return PTR_ERR(rtc);
> +
> +	return 0;
> +}
> +
> +static struct platform_driver wilco_ec_rtc_driver = {
> +	.driver = {
> +		.name = DRV_NAME,
> +	},
> +	.probe = wilco_ec_rtc_probe,
> +};
> +
> +module_platform_driver(wilco_ec_rtc_driver);
> +
> +MODULE_ALIAS("platform:" DRV_NAME);
> +MODULE_AUTHOR("Nick Crews <ncrews@google.com>");
> +MODULE_LICENSE("GPL v2");
> +MODULE_DESCRIPTION("Wilco EC RTC driver");
> -- 
> 2.20.1.97.g81188d93c3-goog
> 

-- 
Alexandre Belloni, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [PATCH v2 5/9] platform/chrome: rtc: Add RTC driver for Wilco EC
  2019-01-14 22:25   ` Alexandre Belloni
@ 2019-01-15  0:30     ` Nick Crews
  2019-01-18 21:24     ` Nick Crews
  1 sibling, 0 replies; 5+ messages in thread
From: Nick Crews @ 2019-01-15  0:30 UTC (permalink / raw)
  To: Alexandre Belloni
  Cc: linux-kernel, groeck, sjg, dlaurie, Duncan Laurie, linux-rtc,
	Enric Balletbo i Serra, Alessandro Zummo, Benson Leung

Thanks for the comments Alexandre, I've responded to your comments
inline. I'll send out a new version of the patch in a bit

On Mon, Jan 14, 2019 at 3:26 PM -700 Alexandre Belloni wrote:
>
> Hello,
>
> On 14/01/2019 15:03:52-0700, Nick Crews wrote:
> > diff --git a/drivers/rtc/rtc-wilco-ec.c b/drivers/rtc/rtc-wilco-ec.c
> > new file mode 100644
> > index 000000000000..d8ed791da82d
> > --- /dev/null
> > +++ b/drivers/rtc/rtc-wilco-ec.c
> > @@ -0,0 +1,178 @@
> > +// SPDX-License-Identifier: GPL-2.0
> > +/*
> > + * RTC interface for Wilco Embedded Controller with R/W abilities
> > + *
> > + * Copyright 2018 Google LLC
> > + *
> > + * The corresponding platform device is typically registered in
> > + * drivers/platform/chrome/wilco_ec/core.c
> > + */
> > +
> > +#include <linux/bcd.h>
> > +#include <linux/err.h>
> > +#include <linux/kernel.h>
> > +#include <linux/module.h>
> > +#include <linux/platform_device.h>
> > +#include <linux/platform_data/wilco-ec.h>
> > +#include <linux/rtc.h>
> > +#include <linux/timekeeping.h>
> > +
> > +#define DRV_NAME "rtc-wilco-ec"
>
> Please get rid of that define.

Done in the new patch.

>
> > +int wilco_ec_rtc_read(struct device *dev, struct rtc_time *tm)
> > +{
> > +     struct wilco_ec_device *ec = dev_get_drvdata(dev->parent);
> > +     u8 param = EC_CMOS_TOD_READ;
> > +     struct ec_rtc_read rtc;
> > +     struct wilco_ec_message msg = {
> > +             .type = WILCO_EC_MSG_LEGACY,
> > +             .flags = WILCO_EC_FLAG_RAW_RESPONSE,
> > +             .command = EC_COMMAND_CMOS,
> > +             .request_data = &param,
> > +             .request_size = sizeof(param),
> > +             .response_data = &rtc,
> > +             .response_size = sizeof(rtc),
> > +     };
> > +     struct rtc_time calc_tm;
> > +     unsigned long time;
> > +     int ret;
> > +
> > +     ret = wilco_ec_mailbox(ec, &msg);
> > +     if (ret < 0)
> > +             return ret;
> > +
> > +     tm->tm_sec      = rtc.second;
> > +     tm->tm_min      = rtc.minute;
> > +     tm->tm_hour     = rtc.hour;
> > +     tm->tm_mday     = rtc.day;
> > +     /*
> > +      * The RTC stores the month value as 1-12 but the kernel expects 0-11,
> > +      * so ensure invalid/zero month value from RTC is not converted to -1.
>
> I'm pretty sure this is not necessary and will certainly make the core
> think the date is valid when it has obviously been corrupted.

OK, I'll get rid of the safety check and just do rtc.month -1.

>
> > +      */
> > +     tm->tm_mon      = rtc.month ? rtc.month - 1 : 0;
> > +     tm->tm_year     = rtc.year + (rtc.century * 100) - 1900;
> > +     tm->tm_yday     = rtc_year_days(tm->tm_mday, tm->tm_mon, tm->tm_year);
> > +
> > +     /* Compute day of week */
> > +     rtc_tm_to_time(tm, &time);
> > +     rtc_time_to_tm(time, &calc_tm);
> > +     tm->tm_wday = calc_tm.tm_wday;
>
> Do you really need an accurate wday? There are no userspace applications
> that I know of that care and converting back and forth is quite
> expensive.
>

We probably don't, this RTC should only get used for very specific purposes,
and any userspace programs should be using a different RTC.

> > +
> > +     return 0;
> > +}
> > +
> > +int wilco_ec_rtc_write(struct device *dev, struct rtc_time *tm)
> > +{
> > +     struct wilco_ec_device *ec = dev_get_drvdata(dev->parent);
> > +     struct ec_rtc_write rtc;
> > +     struct wilco_ec_message msg = {
> > +             .type = WILCO_EC_MSG_LEGACY,
> > +             .flags = WILCO_EC_FLAG_RAW_RESPONSE,
> > +             .command = EC_COMMAND_CMOS,
> > +             .request_data = &rtc,
> > +             .request_size = sizeof(rtc),
> > +     };
> > +     int year = tm->tm_year + 1900;
> > +     /* Convert from 0=Sunday to 0=Saturday for the EC */
> > +     int wday = tm->tm_wday == 6 ? 0 : tm->tm_wday + 1;
>
> I'm not sure how useful it is to set an accurate wday when the RTC is
> not able to give it back.
>

This is actually needed. The RTC on the EC controls battery charging
schedules and thus needs to know the day of the week, and this is the
way to do this.

> > +     int ret;
> > +
> > +     rtc.param       = EC_CMOS_TOD_WRITE;
> > +     rtc.century     = bin2bcd(year / 100);
> > +     rtc.year        = bin2bcd(year % 100);
> > +     rtc.month       = bin2bcd(tm->tm_mon + 1);
> > +     rtc.day         = bin2bcd(tm->tm_mday);
> > +     rtc.hour        = bin2bcd(tm->tm_hour);
> > +     rtc.minute      = bin2bcd(tm->tm_min);
> > +     rtc.second      = bin2bcd(tm->tm_sec);
> > +     rtc.weekday     = bin2bcd(wday);
> > +
> > +     ret = wilco_ec_mailbox(ec, &msg);
> > +     if (ret < 0)
> > +             return ret;
> > +
> > +     return 0;
> > +}
> > +
> > +static const struct rtc_class_ops wilco_ec_rtc_ops = {
> > +     .read_time = wilco_ec_rtc_read,
> > +     .set_time = wilco_ec_rtc_write,
> > +};
> > +
> > +static int wilco_ec_rtc_probe(struct platform_device *pdev)
> > +{
> > +     struct rtc_device *rtc = devm_rtc_device_register(&pdev->dev,
> > +                                                       DRV_NAME,
> > +                                                       &wilco_ec_rtc_ops,
> > +                                                       THIS_MODULE);
>
> Please use devm_rtc_allocate_device() and set rtc->range_min and
> rtc->range_max before registering with rtc_register_device().
>

Thanks, will do this.

> > +     if (IS_ERR(rtc))
> > +             return PTR_ERR(rtc);
> > +
> > +     return 0;
> > +}
> > +
> > +static struct platform_driver wilco_ec_rtc_driver = {
> > +     .driver = {
> > +             .name = DRV_NAME,
> > +     },
> > +     .probe = wilco_ec_rtc_probe,
> > +};
> > +
> > +module_platform_driver(wilco_ec_rtc_driver);
> > +
> > +MODULE_ALIAS("platform:" DRV_NAME);
> > +MODULE_AUTHOR("Nick Crews <ncrews@google.com>");
> > +MODULE_LICENSE("GPL v2");
> > +MODULE_DESCRIPTION("Wilco EC RTC driver");
> > --
> > 2.20.1.97.g81188d93c3-goog
> >
>
> --
> Alexandre Belloni, Bootlin
> Embedded Linux and Kernel engineering
> https://bootlin.com

Nick Crews
ncrews@google.com

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [PATCH v2 5/9] platform/chrome: rtc: Add RTC driver for Wilco EC
  2019-01-14 22:25   ` Alexandre Belloni
  2019-01-15  0:30     ` Nick Crews
@ 2019-01-18 21:24     ` Nick Crews
  1 sibling, 0 replies; 5+ messages in thread
From: Nick Crews @ 2019-01-18 21:24 UTC (permalink / raw)
  To: Alexandre Belloni
  Cc: linux-kernel, groeck, sjg, dlaurie, Duncan Laurie, linux-rtc,
	Enric Balletbo i Serra, Alessandro Zummo, Benson Leung

Hi Alexandre, thanks for taking the time to review this. I've responded to your
comments inline below. I'll send out a new version of this patch soon.

On Mon, Jan 14, 2019 at 3:26 PM Alexandre Belloni
<alexandre.belloni@bootlin.com> wrote:
>
> Hello,
>
> On 14/01/2019 15:03:52-0700, Nick Crews wrote:
> > diff --git a/drivers/rtc/rtc-wilco-ec.c b/drivers/rtc/rtc-wilco-ec.c
> > new file mode 100644
> > index 000000000000..d8ed791da82d
> > --- /dev/null
> > +++ b/drivers/rtc/rtc-wilco-ec.c
> > @@ -0,0 +1,178 @@
> > +// SPDX-License-Identifier: GPL-2.0
> > +/*
> > + * RTC interface for Wilco Embedded Controller with R/W abilities
> > + *
> > + * Copyright 2018 Google LLC
> > + *
> > + * The corresponding platform device is typically registered in
> > + * drivers/platform/chrome/wilco_ec/core.c
> > + */
> > +
> > +#include <linux/bcd.h>
> > +#include <linux/err.h>
> > +#include <linux/kernel.h>
> > +#include <linux/module.h>
> > +#include <linux/platform_device.h>
> > +#include <linux/platform_data/wilco-ec.h>
> > +#include <linux/rtc.h>
> > +#include <linux/timekeeping.h>
> > +
> > +#define DRV_NAME "rtc-wilco-ec"
>
> Please get rid of that define.
>

Done.

> > +int wilco_ec_rtc_read(struct device *dev, struct rtc_time *tm)
> > +{
> > +     struct wilco_ec_device *ec = dev_get_drvdata(dev->parent);
> > +     u8 param = EC_CMOS_TOD_READ;
> > +     struct ec_rtc_read rtc;
> > +     struct wilco_ec_message msg = {
> > +             .type = WILCO_EC_MSG_LEGACY,
> > +             .flags = WILCO_EC_FLAG_RAW_RESPONSE,
> > +             .command = EC_COMMAND_CMOS,
> > +             .request_data = &param,
> > +             .request_size = sizeof(param),
> > +             .response_data = &rtc,
> > +             .response_size = sizeof(rtc),
> > +     };
> > +     struct rtc_time calc_tm;
> > +     unsigned long time;
> > +     int ret;
> > +
> > +     ret = wilco_ec_mailbox(ec, &msg);
> > +     if (ret < 0)
> > +             return ret;
> > +
> > +     tm->tm_sec      = rtc.second;
> > +     tm->tm_min      = rtc.minute;
> > +     tm->tm_hour     = rtc.hour;
> > +     tm->tm_mday     = rtc.day;
> > +     /*
> > +      * The RTC stores the month value as 1-12 but the kernel expects 0-11,
> > +      * so ensure invalid/zero month value from RTC is not converted to -1.
>
> I'm pretty sure this is not necessary and will certainly make the core
> think the date is valid when it has obviously been corrupted.

Done, I'll just use rtc.month - 1 always.

>
> > +      */
> > +     tm->tm_mon      = rtc.month ? rtc.month - 1 : 0;
> > +     tm->tm_year     = rtc.year + (rtc.century * 100) - 1900;
> > +     tm->tm_yday     = rtc_year_days(tm->tm_mday, tm->tm_mon, tm->tm_year);
> > +
> > +     /* Compute day of week */
> > +     rtc_tm_to_time(tm, &time);
> > +     rtc_time_to_tm(time, &calc_tm);
> > +     tm->tm_wday = calc_tm.tm_wday;
>
> Do you really need an accurate wday? There are no userspace applications
> that I know of that care and converting back and forth is quite
> expensive.

No, I'm not sure we do need this. Removing it.

>
> > +
> > +     return 0;
> > +}
> > +
> > +int wilco_ec_rtc_write(struct device *dev, struct rtc_time *tm)
> > +{
> > +     struct wilco_ec_device *ec = dev_get_drvdata(dev->parent);
> > +     struct ec_rtc_write rtc;
> > +     struct wilco_ec_message msg = {
> > +             .type = WILCO_EC_MSG_LEGACY,
> > +             .flags = WILCO_EC_FLAG_RAW_RESPONSE,
> > +             .command = EC_COMMAND_CMOS,
> > +             .request_data = &rtc,
> > +             .request_size = sizeof(rtc),
> > +     };
> > +     int year = tm->tm_year + 1900;
> > +     /* Convert from 0=Sunday to 0=Saturday for the EC */
> > +     int wday = tm->tm_wday == 6 ? 0 : tm->tm_wday + 1;
>
> I'm not sure how useful it is to set an accurate wday when the RTC is
> not able to give it back.

This is actually needed. The RTC controls battery charging schedule,
which varies by the day of the week, and it requires this to be set.

>
> > +     int ret;
> > +
> > +     rtc.param       = EC_CMOS_TOD_WRITE;
> > +     rtc.century     = bin2bcd(year / 100);
> > +     rtc.year        = bin2bcd(year % 100);
> > +     rtc.month       = bin2bcd(tm->tm_mon + 1);
> > +     rtc.day         = bin2bcd(tm->tm_mday);
> > +     rtc.hour        = bin2bcd(tm->tm_hour);
> > +     rtc.minute      = bin2bcd(tm->tm_min);
> > +     rtc.second      = bin2bcd(tm->tm_sec);
> > +     rtc.weekday     = bin2bcd(wday);
> > +
> > +     ret = wilco_ec_mailbox(ec, &msg);
> > +     if (ret < 0)
> > +             return ret;
> > +
> > +     return 0;
> > +}
> > +
> > +static const struct rtc_class_ops wilco_ec_rtc_ops = {
> > +     .read_time = wilco_ec_rtc_read,
> > +     .set_time = wilco_ec_rtc_write,
> > +};
> > +
> > +static int wilco_ec_rtc_probe(struct platform_device *pdev)
> > +{
> > +     struct rtc_device *rtc = devm_rtc_device_register(&pdev->dev,
> > +                                                       DRV_NAME,
> > +                                                       &wilco_ec_rtc_ops,
> > +                                                       THIS_MODULE);
>
> Please use devm_rtc_allocate_device() and set rtc->range_min and
> rtc->range_max before registering with rtc_register_device().

Done.

>
> > +     if (IS_ERR(rtc))
> > +             return PTR_ERR(rtc);
> > +
> > +     return 0;
> > +}
> > +
> > +static struct platform_driver wilco_ec_rtc_driver = {
> > +     .driver = {
> > +             .name = DRV_NAME,
> > +     },
> > +     .probe = wilco_ec_rtc_probe,
> > +};
> > +
> > +module_platform_driver(wilco_ec_rtc_driver);
> > +
> > +MODULE_ALIAS("platform:" DRV_NAME);
> > +MODULE_AUTHOR("Nick Crews <ncrews@google.com>");
> > +MODULE_LICENSE("GPL v2");
> > +MODULE_DESCRIPTION("Wilco EC RTC driver");
> > --
> > 2.20.1.97.g81188d93c3-goog
> >
>
> --
> Alexandre Belloni, Bootlin
> Embedded Linux and Kernel engineering
> https://bootlin.com

Cheers,
Nick Crews

^ permalink raw reply	[flat|nested] 5+ messages in thread

end of thread, other threads:[~2019-01-18 21:25 UTC | newest]

Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2019-01-14 22:03 [PATCH v2 0/9] platform/chrome: rtc: Add support for Wilco EC Nick Crews
2019-01-14 22:03 ` [PATCH v2 5/9] platform/chrome: rtc: Add RTC driver " Nick Crews
2019-01-14 22:25   ` Alexandre Belloni
2019-01-15  0:30     ` Nick Crews
2019-01-18 21:24     ` Nick Crews

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).