* Re: [PATCH v2 3/4] Regulators: TPS65218: Add Regulator driver for TPS65218 PMIC
From: Mark Brown @ 2014-02-14 20:21 UTC (permalink / raw)
To: Keerthy
Cc: rob.herring, pawel.moll, mark.rutland, swarren, ijc+devicetree,
rob, sameo, lee.jones, grant.likely, lgirdwood, devicetree,
linux-doc, linux-kernel, linux-omap
In-Reply-To: <1391665814-18814-4-git-send-email-j-keerthy@ti.com>
[-- Attachment #1: Type: text/plain, Size: 367 bytes --]
On Thu, Feb 06, 2014 at 11:20:13AM +0530, Keerthy wrote:
> This patch adds support for TPS65218 PMIC regulators.
>
> The regulators set consists of 6 DCDCs and 1 LDO. The output
> voltages are configurable and are meant to supply power to the
> main processor and other components.
Applied, thanks. Please use subject lines consistent with the
subsystem.
[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 836 bytes --]
^ permalink raw reply
* Re: [RFC 1/6] mailbox: add core framework
From: Courtney Cavin @ 2014-02-14 20:16 UTC (permalink / raw)
To: Arnd Bergmann
Cc: Rob Herring, Josh Cartwright, s-anna@ti.com, Rob Herring,
Wysocki, Rafael J, Mark Langsdorf, Tony Lindgren,
omar.ramirez@copitl.com, Greg Kroah-Hartman, Pawel Moll,
Mark Rutland, Ian Campbell, Kumar Gala, Rob Landley,
linux-doc@vger.kernel.org, devicetree@vger.kernel.org,
linux-kernel@vger.kernel.org, Linus Walleij, Linus Walleij,
Jassi Brar
In-Reply-To: <201402142048.25456.arnd@arndb.de>
On Fri, Feb 14, 2014 at 08:48:25PM +0100, Arnd Bergmann wrote:
> On Wednesday 12 February 2014, Courtney Cavin wrote:
> > On Tue, Feb 11, 2014 at 09:35:01AM +0100, Arnd Bergmann wrote:
> > > On Monday 10 February 2014 16:23:48 Courtney Cavin wrote:
>
> > Then again, I think that the context management stuff is the exception as well,
> > and I think that can/should also be handled in a higher level. Regardless, I
> > went ahead and drafted the async flags idea out anyway, so here's some
> > pseudo-code. I also tried to shoe-horn in 'peek', and you can see how that
> > turns out. Let me know if this is something like what you had in mind.
>
> The async implementation looks good to me, assuming we actually need both
> sync and async operations, which I can't tell for sure.
Yea, I would like some further input on that specifically. I have added
Linus Walleij and Jassi Brar, who have had good input on mailboxes in
the past, and somehow I missed in this series.
> For the peek operation, it wouldn't work for the ethernet case, which
> has to call it from atomic context in net_rx_action.
It wouldn't work if the mbox is not requested with MBOX_ASYNC, but
otherwise that should be fine, as it would just peek into the kfifo.
That doesn't seem like a desirable method for ethernet use-case though,
as it ends up being two extra copies.
> > /**
> > * so this is where this lock makes things difficult, as this function
> > * might_sleep(), but only really because of the lock. Either we can
> > * remove the lock and force the adapter to do its own locking
> > * spinlock-style, or we can accept the sleep here, which seems a bit
> > * stupid in a peek function. Neither option is good. Additionally,
> > * there's no guarantee that the adapter doesn't operate over a bus
> > * which itself might_sleep(), exacerbating the problem.
> > */
> > mutex_lock(&mbox->adapter->lock);
> > rc = mbox->adapter->ops->peek_message(mbox->adapter, mbox->chan, msg);
> > mutex_lock(&mbox->adapter->lock);
>
> If we decide that peek() must not sleep, any driver that operates on a
> slow bus could just always report "no data" here.
Yes indeed, or it could just not implement peek, which seems reasonable.
> Moving the locking into the mbox driver here sounds appropriate.
I don't really like doing that for the entirety of the mbox core, as it
makes the simple adapters harder to write properly. Since peek is not
a typical use-case, perhaps we could remove the locking for just peek,
and have a Big Fat Warning in the description of how to properly
implement it?
> Arnd
Thanks for the input!
-Courtney
^ permalink raw reply
* Re: How to select between different display timings?
From: Anatolij Gustschin @ 2014-02-14 20:06 UTC (permalink / raw)
To: Dirk Behme
Cc: Lothar Waßmann, Shawn Guo, Dirk Behme,
devicetree-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, Rob Herring,
linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r@public.gmane.org
In-Reply-To: <52FE53DD.7040102-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
On Fri, 14 Feb 2014 18:35:25 +0100
Dirk Behme <dirk.behme-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
...
> > You need to set the property 'native-mode' to the phandle of the
> > display-timings entry like marked above.
>
> Hmm, let's see if we talk about the same thing ;) Let me rephrase my
> questions:
>
> Using above example and booting a system with the resulting .dtb,
> timing5 is selected at boot time, correct?
yes.
> Assuming I want to use timing1 instead: who would set when the
> property 'native-mode' to &timing1?
>
> Can the switch to
>
> native-mode = <&timing1>;
>
> only be done at compile time of the dts -> dtb? Or can this be done at
> boot/kernel run time, too?
it could be done in U-Boot before booting the kernel. You can load
the dtb and patch the property as needed, i.e.:
U-Boot > tftp 107f0000 ${fdtfile}
U-Boot > fdt addr 107f0000
U-Boot > fdt pri /soc/aips-bus@02000000/ldb@020e0008/lvds-channel@0/display-timings
display-timings {
native-mode = <0x0000000d>;
hsd100pxn1 {
clock-frequency = <0x03dfd240>;
hactive = <0x00000400>;
vactive = <0x00000300>;
hback-porch = <0x000000dc>;
hfront-porch = <0x00000028>;
vback-porch = <0x00000015>;
vfront-porch = <0x00000007>;
hsync-len = <0x0000003c>;
vsync-len = <0x0000000a>;
linux,phandle = <0x0000000d>;
phandle = <0x0000000d>;
};
wvga {
clock-frequency = <0x019bfcc0>;
hactive = <0x00000320>;
vactive = <0x000001e0>;
hback-porch = <0x00000028>;
hfront-porch = <0x0000003c>;
vback-porch = <0x0000000a>;
vfront-porch = <0x0000000a>;
hsync-len = <0x00000014>;
vsync-len = <0x0000000a>;
hsync-active = <0x00000000>;
vsync-active = <0x00000000>;
de-active = <0x00000001>;
pixelclk-active = <0x00000000>;
linux,phandle = <0x0000000e>;
phandle = <0x0000000e>;
};
};
U-Boot > fdt set /soc/aips-bus@02000000/ldb@020e0008/lvds-channel@0/display-timings native-mode <0x0000000e>
U-Boot > fdt list /soc/aips-bus@02000000/ldb@020e0008/lvds-channel@0/display-timings/
display-timings {
native-mode = <0x0000000e>;
hsd100pxn1 {
};
wvga {
};
};
native-mode is referencing second timing node now.
But note that dtc could eliminate the phandles in the nodes if
these are not referenced in dts. To avoid this, the nodes could
contain a self-reference:
timing1: wvga {
...
linux,phandle = <&timing1>;
};
> In the end, I'm looking for a way to have several timings like above
> in the device tree, and select one at kernel boot time based on the
> display detected. Would this be possible?
You could add a function in U-Boot to detect the display and
update the dtb before booting. This function can be called
before booting if you define CONFIG_OF_BOARD_SETUP in the
U-Boot config file for your board and provide ft_board_setup()
in your board code.
HTH,
Anatolij
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply
* Re: [RFC 1/6] mailbox: add core framework
From: Arnd Bergmann @ 2014-02-14 19:48 UTC (permalink / raw)
To: Courtney Cavin
Cc: Rob Herring, Josh Cartwright, s-anna@ti.com, Rob Herring,
Wysocki, Rafael J, Mark Langsdorf, Tony Lindgren,
omar.ramirez@copitl.com, Greg Kroah-Hartman, Pawel Moll,
Mark Rutland, Ian Campbell, Kumar Gala, Rob Landley,
linux-doc@vger.kernel.org, devicetree@vger.kernel.org,
linux-kernel@vger.kernel.org
In-Reply-To: <20140212183143.GD1706@sonymobile.com>
On Wednesday 12 February 2014, Courtney Cavin wrote:
> On Tue, Feb 11, 2014 at 09:35:01AM +0100, Arnd Bergmann wrote:
> > On Monday 10 February 2014 16:23:48 Courtney Cavin wrote:
> Then again, I think that the context management stuff is the exception as well,
> and I think that can/should also be handled in a higher level. Regardless, I
> went ahead and drafted the async flags idea out anyway, so here's some
> pseudo-code. I also tried to shoe-horn in 'peek', and you can see how that
> turns out. Let me know if this is something like what you had in mind.
The async implementation looks good to me, assuming we actually need both
sync and async operations, which I can't tell for sure.
For the peek operation, it wouldn't work for the ethernet case, which
has to call it from atomic context in net_rx_action.
> /**
> * so this is where this lock makes things difficult, as this function
> * might_sleep(), but only really because of the lock. Either we can
> * remove the lock and force the adapter to do its own locking
> * spinlock-style, or we can accept the sleep here, which seems a bit
> * stupid in a peek function. Neither option is good. Additionally,
> * there's no guarantee that the adapter doesn't operate over a bus
> * which itself might_sleep(), exacerbating the problem.
> */
> mutex_lock(&mbox->adapter->lock);
> rc = mbox->adapter->ops->peek_message(mbox->adapter, mbox->chan, msg);
> mutex_lock(&mbox->adapter->lock);
If we decide that peek() must not sleep, any driver that operates on a
slow bus could just always report "no data" here.
Moving the locking into the mbox driver here sounds appropriate.
Arnd
^ permalink raw reply
* Re: [PATCH v3 1/3] ARM: EXYNOS: initial board support for exynos5260 SoC
From: Tomasz Figa @ 2014-02-14 18:58 UTC (permalink / raw)
To: Rahul Sharma, linux-samsung-soc
Cc: devicetree, linux-arm-kernel, kgene.kim, joshi, r.sh.open,
Pankaj Dubey
In-Reply-To: <1392385032-22015-2-git-send-email-rahul.sharma@samsung.com>
Hi,
On 14.02.2014 14:37, Rahul Sharma wrote:
> From: Pankaj Dubey <pankaj.dubey@samsung.com>
>
> This patch add basic arch side support for exynos5260 SoC.
>
> Signed-off-by: Pankaj Dubey <pankaj.dubey@samsung.com>
> Signed-off-by: Arun Kumar K <arun.kk@samsung.com>
> Signed-off-by: Rahul Sharma <rahul.sharma@samsung.com>
> ---
> arch/arm/mach-exynos/Kconfig | 9 +++++++++
> arch/arm/mach-exynos/common.c | 11 +++++++++++
> arch/arm/mach-exynos/include/mach/map.h | 1 +
> arch/arm/mach-exynos/mach-exynos5-dt.c | 1 +
> arch/arm/plat-samsung/include/plat/cpu.h | 8 ++++++++
> 5 files changed, 30 insertions(+)
Reviewed-by: Tomasz Figa <t.figa@samsung.com>
Best regards,
Tomasz
^ permalink raw reply
* Re: [PATCH v3 3/3] ARM: dts: add dts files for xyref5260 board
From: Tomasz Figa @ 2014-02-14 18:57 UTC (permalink / raw)
To: Rahul Sharma, linux-samsung-soc
Cc: devicetree, linux-arm-kernel, kgene.kim, joshi, r.sh.open
In-Reply-To: <1392385032-22015-4-git-send-email-rahul.sharma@samsung.com>
Hi Rahul,
The patch looks good, except one issue pointed inline.
On 14.02.2014 14:37, Rahul Sharma wrote:
> diff --git a/arch/arm/boot/dts/exynos5260-xyref5260-evt0.dts b/arch/arm/boot/dts/exynos5260-xyref5260-evt0.dts
> new file mode 100644
> index 0000000..c4efc1e
> --- /dev/null
> +++ b/arch/arm/boot/dts/exynos5260-xyref5260-evt0.dts
[snip]
> +&mmc_0 {
> + status = "okay";
> + num-slots = <1>;
> + broken-cd;
> + bypass-smu;
> + supports-highspeed;
> + supports-hs200-mode; /* 200 Mhz */
> + fifo-depth = <0x40>;
This is a SoC-level property and it is already specified in your
exynos5260.dtsi.
> + card-detect-delay = <200>;
> + samsung,dw-mshc-ciu-div = <3>;
> + samsung,dw-mshc-sdr-timing = <0 4>;
> + samsung,dw-mshc-ddr-timing = <0 2>;
> + pinctrl-names = "default";
> + pinctrl-0 = <&sd0_rdqs &sd0_clk &sd0_cmd &sd0_bus1 &sd0_bus4 &sd0_bus8>;
> +
> + slot@0 {
> + reg = <0>;
> + bus-width = <8>;
> + };
> +};
> +
> +&mmc_2 {
> + status = "okay";
> + num-slots = <1>;
> + supports-highspeed;
> + fifo-depth = <0x40>;
Ditto.
Best regards,
Tomasz
^ permalink raw reply
* [RFCv1 4/4] mfd: twl4030-madc: Move driver to drivers/iio/adc
From: Sebastian Reichel @ 2014-02-14 18:46 UTC (permalink / raw)
To: Sebastian Reichel, Marek Belisko
Cc: Jonathan Cameron, Lee Jones, Samuel Ortiz, Lars-Peter Clausen,
Rob Herring, Pawel Moll, Mark Rutland, Ian Campbell, Kumar Gala,
Grant Likely, linux-kernel, devicetree, linux-iio,
Sebastian Reichel
In-Reply-To: <1392403586-30540-1-git-send-email-sre@debian.org>
This is a driver for an A/D converter, which belongs into
drivers/iio/adc.
Signed-off-by: Sebastian Reichel <sre@debian.org>
---
drivers/iio/adc/Kconfig | 10 +
drivers/iio/adc/Makefile | 1 +
drivers/iio/adc/twl4030-madc.c | 922 +++++++++++++++++++++++++++++++++++++++++
drivers/mfd/Kconfig | 10 -
drivers/mfd/Makefile | 1 -
drivers/mfd/twl4030-madc.c | 922 -----------------------------------------
6 files changed, 933 insertions(+), 933 deletions(-)
create mode 100644 drivers/iio/adc/twl4030-madc.c
delete mode 100644 drivers/mfd/twl4030-madc.c
diff --git a/drivers/iio/adc/Kconfig b/drivers/iio/adc/Kconfig
index 2209f28..427f75c 100644
--- a/drivers/iio/adc/Kconfig
+++ b/drivers/iio/adc/Kconfig
@@ -183,6 +183,16 @@ config TI_AM335X_ADC
Say yes here to build support for Texas Instruments ADC
driver which is also a MFD client.
+config TWL4030_MADC
+ tristate "TWL4030 MADC (Monitoring A/D Converter)"
+ depends on TWL4030_CORE
+ help
+ This driver provides support for Triton TWL4030-MADC. The
+ driver supports both RT and SW conversion methods.
+
+ This driver can also be built as a module. If so, the module will be
+ called twl4030-madc.
+
config TWL6030_GPADC
tristate "TWL6030 GPADC (General Purpose A/D Converter) Support"
depends on TWL4030_CORE
diff --git a/drivers/iio/adc/Makefile b/drivers/iio/adc/Makefile
index ba9a10a..9acf2df 100644
--- a/drivers/iio/adc/Makefile
+++ b/drivers/iio/adc/Makefile
@@ -20,5 +20,6 @@ obj-$(CONFIG_MCP3422) += mcp3422.o
obj-$(CONFIG_NAU7802) += nau7802.o
obj-$(CONFIG_TI_ADC081C) += ti-adc081c.o
obj-$(CONFIG_TI_AM335X_ADC) += ti_am335x_adc.o
+obj-$(CONFIG_TWL4030_MADC) += twl4030-madc.o
obj-$(CONFIG_TWL6030_GPADC) += twl6030-gpadc.o
obj-$(CONFIG_VIPERBOARD_ADC) += viperboard_adc.o
diff --git a/drivers/iio/adc/twl4030-madc.c b/drivers/iio/adc/twl4030-madc.c
new file mode 100644
index 0000000..4da61c4
--- /dev/null
+++ b/drivers/iio/adc/twl4030-madc.c
@@ -0,0 +1,922 @@
+/*
+ *
+ * TWL4030 MADC module driver-This driver monitors the real time
+ * conversion of analog signals like battery temperature,
+ * battery type, battery level etc.
+ *
+ * Copyright (C) 2011 Texas Instruments Incorporated - http://www.ti.com/
+ * J Keerthy <j-keerthy@ti.com>
+ *
+ * Based on twl4030-madc.c
+ * Copyright (C) 2008 Nokia Corporation
+ * Mikko Ylinen <mikko.k.ylinen@nokia.com>
+ *
+ * Amit Kucheria <amit.kucheria@canonical.com>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * version 2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA
+ * 02110-1301 USA
+ *
+ */
+
+#include <linux/init.h>
+#include <linux/device.h>
+#include <linux/interrupt.h>
+#include <linux/kernel.h>
+#include <linux/delay.h>
+#include <linux/platform_device.h>
+#include <linux/slab.h>
+#include <linux/i2c/twl.h>
+#include <linux/i2c/twl4030-madc.h>
+#include <linux/module.h>
+#include <linux/stddef.h>
+#include <linux/mutex.h>
+#include <linux/bitops.h>
+#include <linux/jiffies.h>
+#include <linux/types.h>
+#include <linux/gfp.h>
+#include <linux/err.h>
+
+#include <linux/iio/iio.h>
+#include <linux/iio/machine.h>
+#include <linux/iio/driver.h>
+
+/*
+ * struct twl4030_madc_data - a container for madc info
+ * @dev - pointer to device structure for madc
+ * @lock - mutex protecting this data structure
+ * @requests - Array of request struct corresponding to SW1, SW2 and RT
+ * @imr - Interrupt mask register of MADC
+ * @isr - Interrupt status register of MADC
+ */
+struct twl4030_madc_data {
+ struct device *dev;
+ struct mutex lock; /* mutex protecting this data structure */
+ struct twl4030_madc_request requests[TWL4030_MADC_NUM_METHODS];
+ bool use_second_irq;
+ int imr;
+ int isr;
+};
+
+static int twl4030_madc_read(struct iio_dev *iio_dev,
+ const struct iio_chan_spec *chan,
+ int *val, int *val2, long mask)
+{
+ struct twl4030_madc_data *madc = iio_priv(iio_dev);
+ struct twl4030_madc_request req;
+ int channel = chan->channel;
+ int ret;
+
+ req.method = madc->use_second_irq ? TWL4030_MADC_SW2 : TWL4030_MADC_SW1;
+
+ req.channels = BIT(channel);
+ req.active = 0;
+ req.func_cb = NULL;
+ req.raw = (mask & IIO_CHAN_INFO_PROCESSED) ? false : true;
+ req.do_avg = (mask & IIO_CHAN_INFO_AVERAGE_RAW) ? true : false;
+
+ ret = twl4030_madc_conversion(&req);
+ if (ret < 0)
+ return ret;
+
+ *val = req.rbuf[channel];
+
+ return IIO_VAL_INT;
+}
+
+static const struct iio_info twl4030_madc_iio_info = {
+ .read_raw = &twl4030_madc_read,
+ .driver_module = THIS_MODULE,
+};
+
+#define ADC_CHANNEL(_channel, _type, _name, _mask) { \
+ .type = _type, \
+ .scan_type = IIO_ST('u', 10, 16, 0), \
+ .channel = _channel, \
+ .info_mask_separate = _mask, \
+ .datasheet_name = _name, \
+ .indexed = 1, \
+}
+
+static const struct iio_chan_spec twl4030_madc_iio_channels[] = {
+ ADC_CHANNEL(0, IIO_VOLTAGE, "ADCIN0", BIT(IIO_CHAN_INFO_AVERAGE_RAW)),
+ ADC_CHANNEL(1, IIO_TEMP, "ADCIN1", BIT(IIO_CHAN_INFO_PROCESSED) |
+ BIT(IIO_CHAN_INFO_AVERAGE_RAW)),
+ ADC_CHANNEL(2, IIO_VOLTAGE, "ADCIN2", BIT(IIO_CHAN_INFO_AVERAGE_RAW)),
+ ADC_CHANNEL(3, IIO_VOLTAGE, "ADCIN3", BIT(IIO_CHAN_INFO_AVERAGE_RAW)),
+ ADC_CHANNEL(4, IIO_VOLTAGE, "ADCIN4", BIT(IIO_CHAN_INFO_AVERAGE_RAW)),
+ ADC_CHANNEL(5, IIO_VOLTAGE, "ADCIN5", BIT(IIO_CHAN_INFO_AVERAGE_RAW)),
+ ADC_CHANNEL(6, IIO_VOLTAGE, "ADCIN6", BIT(IIO_CHAN_INFO_AVERAGE_RAW)),
+ ADC_CHANNEL(7, IIO_VOLTAGE, "ADCIN7", BIT(IIO_CHAN_INFO_AVERAGE_RAW)),
+ ADC_CHANNEL(8, IIO_VOLTAGE, "ADCIN8", BIT(IIO_CHAN_INFO_AVERAGE_RAW)),
+ ADC_CHANNEL(9, IIO_VOLTAGE, "ADCIN9", BIT(IIO_CHAN_INFO_AVERAGE_RAW)),
+ ADC_CHANNEL(10, IIO_CURRENT, "ADCIN10", BIT(IIO_CHAN_INFO_PROCESSED) |
+ BIT(IIO_CHAN_INFO_AVERAGE_RAW)),
+ ADC_CHANNEL(11, IIO_VOLTAGE, "ADCIN11", BIT(IIO_CHAN_INFO_AVERAGE_RAW)),
+ ADC_CHANNEL(12, IIO_VOLTAGE, "ADCIN12", BIT(IIO_CHAN_INFO_AVERAGE_RAW)),
+ ADC_CHANNEL(13, IIO_VOLTAGE, "ADCIN13", BIT(IIO_CHAN_INFO_AVERAGE_RAW)),
+ ADC_CHANNEL(14, IIO_VOLTAGE, "ADCIN14", BIT(IIO_CHAN_INFO_AVERAGE_RAW)),
+ ADC_CHANNEL(15, IIO_VOLTAGE, "ADCIN15", BIT(IIO_CHAN_INFO_AVERAGE_RAW)),
+};
+
+static struct twl4030_madc_data *twl4030_madc;
+
+struct twl4030_prescale_divider_ratios {
+ s16 numerator;
+ s16 denominator;
+};
+
+static const struct twl4030_prescale_divider_ratios
+twl4030_divider_ratios[16] = {
+ {1, 1}, /* CHANNEL 0 No Prescaler */
+ {1, 1}, /* CHANNEL 1 No Prescaler */
+ {6, 10}, /* CHANNEL 2 */
+ {6, 10}, /* CHANNEL 3 */
+ {6, 10}, /* CHANNEL 4 */
+ {6, 10}, /* CHANNEL 5 */
+ {6, 10}, /* CHANNEL 6 */
+ {6, 10}, /* CHANNEL 7 */
+ {3, 14}, /* CHANNEL 8 */
+ {1, 3}, /* CHANNEL 9 */
+ {1, 1}, /* CHANNEL 10 No Prescaler */
+ {15, 100}, /* CHANNEL 11 */
+ {1, 4}, /* CHANNEL 12 */
+ {1, 1}, /* CHANNEL 13 Reserved channels */
+ {1, 1}, /* CHANNEL 14 Reseved channels */
+ {5, 11}, /* CHANNEL 15 */
+};
+
+
+/*
+ * Conversion table from -3 to 55 degree Celcius
+ */
+static int therm_tbl[] = {
+30800, 29500, 28300, 27100,
+26000, 24900, 23900, 22900, 22000, 21100, 20300, 19400, 18700, 17900,
+17200, 16500, 15900, 15300, 14700, 14100, 13600, 13100, 12600, 12100,
+11600, 11200, 10800, 10400, 10000, 9630, 9280, 8950, 8620, 8310,
+8020, 7730, 7460, 7200, 6950, 6710, 6470, 6250, 6040, 5830,
+5640, 5450, 5260, 5090, 4920, 4760, 4600, 4450, 4310, 4170,
+4040, 3910, 3790, 3670, 3550
+};
+
+/*
+ * Structure containing the registers
+ * of different conversion methods supported by MADC.
+ * Hardware or RT real time conversion request initiated by external host
+ * processor for RT Signal conversions.
+ * External host processors can also request for non RT conversions
+ * SW1 and SW2 software conversions also called asynchronous or GPC request.
+ */
+static
+const struct twl4030_madc_conversion_method twl4030_conversion_methods[] = {
+ [TWL4030_MADC_RT] = {
+ .sel = TWL4030_MADC_RTSELECT_LSB,
+ .avg = TWL4030_MADC_RTAVERAGE_LSB,
+ .rbase = TWL4030_MADC_RTCH0_LSB,
+ },
+ [TWL4030_MADC_SW1] = {
+ .sel = TWL4030_MADC_SW1SELECT_LSB,
+ .avg = TWL4030_MADC_SW1AVERAGE_LSB,
+ .rbase = TWL4030_MADC_GPCH0_LSB,
+ .ctrl = TWL4030_MADC_CTRL_SW1,
+ },
+ [TWL4030_MADC_SW2] = {
+ .sel = TWL4030_MADC_SW2SELECT_LSB,
+ .avg = TWL4030_MADC_SW2AVERAGE_LSB,
+ .rbase = TWL4030_MADC_GPCH0_LSB,
+ .ctrl = TWL4030_MADC_CTRL_SW2,
+ },
+};
+
+/*
+ * Function to read a particular channel value.
+ * @madc - pointer to struct twl4030_madc_data
+ * @reg - lsb of ADC Channel
+ * If the i2c read fails it returns an error else returns 0.
+ */
+static int twl4030_madc_channel_raw_read(struct twl4030_madc_data *madc, u8 reg)
+{
+ u8 msb, lsb;
+ int ret;
+ /*
+ * For each ADC channel, we have MSB and LSB register pair. MSB address
+ * is always LSB address+1. reg parameter is the address of LSB register
+ */
+ ret = twl_i2c_read_u8(TWL4030_MODULE_MADC, &msb, reg + 1);
+ if (ret) {
+ dev_err(madc->dev, "unable to read MSB register 0x%X\n",
+ reg + 1);
+ return ret;
+ }
+ ret = twl_i2c_read_u8(TWL4030_MODULE_MADC, &lsb, reg);
+ if (ret) {
+ dev_err(madc->dev, "unable to read LSB register 0x%X\n", reg);
+ return ret;
+ }
+
+ return (int)(((msb << 8) | lsb) >> 6);
+}
+
+/*
+ * Return battery temperature
+ * Or < 0 on failure.
+ */
+static int twl4030battery_temperature(int raw_volt)
+{
+ u8 val;
+ int temp, curr, volt, res, ret;
+
+ volt = (raw_volt * TEMP_STEP_SIZE) / TEMP_PSR_R;
+ /* Getting and calculating the supply current in micro ampers */
+ ret = twl_i2c_read_u8(TWL_MODULE_MAIN_CHARGE, &val,
+ REG_BCICTL2);
+ if (ret < 0)
+ return ret;
+ curr = ((val & TWL4030_BCI_ITHEN) + 1) * 10;
+ /* Getting and calculating the thermistor resistance in ohms */
+ res = volt * 1000 / curr;
+ /* calculating temperature */
+ for (temp = 58; temp >= 0; temp--) {
+ int actual = therm_tbl[temp];
+
+ if ((actual - res) >= 0)
+ break;
+ }
+
+ return temp + 1;
+}
+
+static int twl4030battery_current(int raw_volt)
+{
+ int ret;
+ u8 val;
+
+ ret = twl_i2c_read_u8(TWL_MODULE_MAIN_CHARGE, &val,
+ TWL4030_BCI_BCICTL1);
+ if (ret)
+ return ret;
+ if (val & TWL4030_BCI_CGAIN) /* slope of 0.44 mV/mA */
+ return (raw_volt * CURR_STEP_SIZE) / CURR_PSR_R1;
+ else /* slope of 0.88 mV/mA */
+ return (raw_volt * CURR_STEP_SIZE) / CURR_PSR_R2;
+}
+/*
+ * Function to read channel values
+ * @madc - pointer to twl4030_madc_data struct
+ * @reg_base - Base address of the first channel
+ * @Channels - 16 bit bitmap. If the bit is set, channel value is read
+ * @buf - The channel values are stored here. if read fails error
+ * @raw - Return raw values without conversion
+ * value is stored
+ * Returns the number of successfully read channels.
+ */
+static int twl4030_madc_read_channels(struct twl4030_madc_data *madc,
+ u8 reg_base, unsigned
+ long channels, int *buf,
+ bool raw)
+{
+ int count = 0, count_req = 0, i;
+ u8 reg;
+
+ for_each_set_bit(i, &channels, TWL4030_MADC_MAX_CHANNELS) {
+ reg = reg_base + 2 * i;
+ buf[i] = twl4030_madc_channel_raw_read(madc, reg);
+ if (buf[i] < 0) {
+ dev_err(madc->dev,
+ "Unable to read register 0x%X\n", reg);
+ count_req++;
+ continue;
+ }
+ if (raw) {
+ count++;
+ continue;
+ }
+ switch (i) {
+ case 10:
+ buf[i] = twl4030battery_current(buf[i]);
+ if (buf[i] < 0) {
+ dev_err(madc->dev, "err reading current\n");
+ count_req++;
+ } else {
+ count++;
+ buf[i] = buf[i] - 750;
+ }
+ break;
+ case 1:
+ buf[i] = twl4030battery_temperature(buf[i]);
+ if (buf[i] < 0) {
+ dev_err(madc->dev, "err reading temperature\n");
+ count_req++;
+ } else {
+ buf[i] -= 3;
+ count++;
+ }
+ break;
+ default:
+ count++;
+ /* Analog Input (V) = conv_result * step_size / R
+ * conv_result = decimal value of 10-bit conversion
+ * result
+ * step size = 1.5 / (2 ^ 10 -1)
+ * R = Prescaler ratio for input channels.
+ * Result given in mV hence multiplied by 1000.
+ */
+ buf[i] = (buf[i] * 3 * 1000 *
+ twl4030_divider_ratios[i].denominator)
+ / (2 * 1023 *
+ twl4030_divider_ratios[i].numerator);
+ }
+ }
+ if (count_req)
+ dev_err(madc->dev, "%d channel conversion failed\n", count_req);
+
+ return count;
+}
+
+/*
+ * Enables irq.
+ * @madc - pointer to twl4030_madc_data struct
+ * @id - irq number to be enabled
+ * can take one of TWL4030_MADC_RT, TWL4030_MADC_SW1, TWL4030_MADC_SW2
+ * corresponding to RT, SW1, SW2 conversion requests.
+ * If the i2c read fails it returns an error else returns 0.
+ */
+static int twl4030_madc_enable_irq(struct twl4030_madc_data *madc, u8 id)
+{
+ u8 val;
+ int ret;
+
+ ret = twl_i2c_read_u8(TWL4030_MODULE_MADC, &val, madc->imr);
+ if (ret) {
+ dev_err(madc->dev, "unable to read imr register 0x%X\n",
+ madc->imr);
+ return ret;
+ }
+ val &= ~(1 << id);
+ ret = twl_i2c_write_u8(TWL4030_MODULE_MADC, val, madc->imr);
+ if (ret) {
+ dev_err(madc->dev,
+ "unable to write imr register 0x%X\n", madc->imr);
+ return ret;
+
+ }
+
+ return 0;
+}
+
+/*
+ * Disables irq.
+ * @madc - pointer to twl4030_madc_data struct
+ * @id - irq number to be disabled
+ * can take one of TWL4030_MADC_RT, TWL4030_MADC_SW1, TWL4030_MADC_SW2
+ * corresponding to RT, SW1, SW2 conversion requests.
+ * Returns error if i2c read/write fails.
+ */
+static int twl4030_madc_disable_irq(struct twl4030_madc_data *madc, u8 id)
+{
+ u8 val;
+ int ret;
+
+ ret = twl_i2c_read_u8(TWL4030_MODULE_MADC, &val, madc->imr);
+ if (ret) {
+ dev_err(madc->dev, "unable to read imr register 0x%X\n",
+ madc->imr);
+ return ret;
+ }
+ val |= (1 << id);
+ ret = twl_i2c_write_u8(TWL4030_MODULE_MADC, val, madc->imr);
+ if (ret) {
+ dev_err(madc->dev,
+ "unable to write imr register 0x%X\n", madc->imr);
+ return ret;
+ }
+
+ return 0;
+}
+
+static irqreturn_t twl4030_madc_threaded_irq_handler(int irq, void *_madc)
+{
+ struct twl4030_madc_data *madc = _madc;
+ const struct twl4030_madc_conversion_method *method;
+ u8 isr_val, imr_val;
+ int i, len, ret;
+ struct twl4030_madc_request *r;
+
+ mutex_lock(&madc->lock);
+ ret = twl_i2c_read_u8(TWL4030_MODULE_MADC, &isr_val, madc->isr);
+ if (ret) {
+ dev_err(madc->dev, "unable to read isr register 0x%X\n",
+ madc->isr);
+ goto err_i2c;
+ }
+ ret = twl_i2c_read_u8(TWL4030_MODULE_MADC, &imr_val, madc->imr);
+ if (ret) {
+ dev_err(madc->dev, "unable to read imr register 0x%X\n",
+ madc->imr);
+ goto err_i2c;
+ }
+ isr_val &= ~imr_val;
+ for (i = 0; i < TWL4030_MADC_NUM_METHODS; i++) {
+ if (!(isr_val & (1 << i)))
+ continue;
+ ret = twl4030_madc_disable_irq(madc, i);
+ if (ret < 0)
+ dev_dbg(madc->dev, "Disable interrupt failed%d\n", i);
+ madc->requests[i].result_pending = 1;
+ }
+ for (i = 0; i < TWL4030_MADC_NUM_METHODS; i++) {
+ r = &madc->requests[i];
+ /* No pending results for this method, move to next one */
+ if (!r->result_pending)
+ continue;
+ method = &twl4030_conversion_methods[r->method];
+ /* Read results */
+ len = twl4030_madc_read_channels(madc, method->rbase,
+ r->channels, r->rbuf, r->raw);
+ /* Return results to caller */
+ if (r->func_cb != NULL) {
+ r->func_cb(len, r->channels, r->rbuf);
+ r->func_cb = NULL;
+ }
+ /* Free request */
+ r->result_pending = 0;
+ r->active = 0;
+ }
+ mutex_unlock(&madc->lock);
+
+ return IRQ_HANDLED;
+
+err_i2c:
+ /*
+ * In case of error check whichever request is active
+ * and service the same.
+ */
+ for (i = 0; i < TWL4030_MADC_NUM_METHODS; i++) {
+ r = &madc->requests[i];
+ if (r->active == 0)
+ continue;
+ method = &twl4030_conversion_methods[r->method];
+ /* Read results */
+ len = twl4030_madc_read_channels(madc, method->rbase,
+ r->channels, r->rbuf, r->raw);
+ /* Return results to caller */
+ if (r->func_cb != NULL) {
+ r->func_cb(len, r->channels, r->rbuf);
+ r->func_cb = NULL;
+ }
+ /* Free request */
+ r->result_pending = 0;
+ r->active = 0;
+ }
+ mutex_unlock(&madc->lock);
+
+ return IRQ_HANDLED;
+}
+
+static int twl4030_madc_set_irq(struct twl4030_madc_data *madc,
+ struct twl4030_madc_request *req)
+{
+ struct twl4030_madc_request *p;
+ int ret;
+
+ p = &madc->requests[req->method];
+ memcpy(p, req, sizeof(*req));
+ ret = twl4030_madc_enable_irq(madc, req->method);
+ if (ret < 0) {
+ dev_err(madc->dev, "enable irq failed!!\n");
+ return ret;
+ }
+
+ return 0;
+}
+
+/*
+ * Function which enables the madc conversion
+ * by writing to the control register.
+ * @madc - pointer to twl4030_madc_data struct
+ * @conv_method - can be TWL4030_MADC_RT, TWL4030_MADC_SW2, TWL4030_MADC_SW1
+ * corresponding to RT SW1 or SW2 conversion methods.
+ * Returns 0 if succeeds else a negative error value
+ */
+static int twl4030_madc_start_conversion(struct twl4030_madc_data *madc,
+ int conv_method)
+{
+ const struct twl4030_madc_conversion_method *method;
+ int ret = 0;
+ method = &twl4030_conversion_methods[conv_method];
+ switch (conv_method) {
+ case TWL4030_MADC_SW1:
+ case TWL4030_MADC_SW2:
+ ret = twl_i2c_write_u8(TWL4030_MODULE_MADC,
+ TWL4030_MADC_SW_START, method->ctrl);
+ if (ret) {
+ dev_err(madc->dev,
+ "unable to write ctrl register 0x%X\n",
+ method->ctrl);
+ return ret;
+ }
+ break;
+ default:
+ break;
+ }
+
+ return 0;
+}
+
+/*
+ * Function that waits for conversion to be ready
+ * @madc - pointer to twl4030_madc_data struct
+ * @timeout_ms - timeout value in milliseconds
+ * @status_reg - ctrl register
+ * returns 0 if succeeds else a negative error value
+ */
+static int twl4030_madc_wait_conversion_ready(struct twl4030_madc_data *madc,
+ unsigned int timeout_ms,
+ u8 status_reg)
+{
+ unsigned long timeout;
+ int ret;
+
+ timeout = jiffies + msecs_to_jiffies(timeout_ms);
+ do {
+ u8 reg;
+
+ ret = twl_i2c_read_u8(TWL4030_MODULE_MADC, ®, status_reg);
+ if (ret) {
+ dev_err(madc->dev,
+ "unable to read status register 0x%X\n",
+ status_reg);
+ return ret;
+ }
+ if (!(reg & TWL4030_MADC_BUSY) && (reg & TWL4030_MADC_EOC_SW))
+ return 0;
+ usleep_range(500, 2000);
+ } while (!time_after(jiffies, timeout));
+ dev_err(madc->dev, "conversion timeout!\n");
+
+ return -EAGAIN;
+}
+
+/*
+ * An exported function which can be called from other kernel drivers.
+ * @req twl4030_madc_request structure
+ * req->rbuf will be filled with read values of channels based on the
+ * channel index. If a particular channel reading fails there will
+ * be a negative error value in the corresponding array element.
+ * returns 0 if succeeds else error value
+ */
+int twl4030_madc_conversion(struct twl4030_madc_request *req)
+{
+ const struct twl4030_madc_conversion_method *method;
+ u8 ch_msb, ch_lsb;
+ int ret;
+
+ if (!req || !twl4030_madc)
+ return -EINVAL;
+
+ mutex_lock(&twl4030_madc->lock);
+ if (req->method < TWL4030_MADC_RT || req->method > TWL4030_MADC_SW2) {
+ ret = -EINVAL;
+ goto out;
+ }
+ /* Do we have a conversion request ongoing */
+ if (twl4030_madc->requests[req->method].active) {
+ ret = -EBUSY;
+ goto out;
+ }
+ ch_msb = (req->channels >> 8) & 0xff;
+ ch_lsb = req->channels & 0xff;
+ method = &twl4030_conversion_methods[req->method];
+ /* Select channels to be converted */
+ ret = twl_i2c_write_u8(TWL4030_MODULE_MADC, ch_msb, method->sel + 1);
+ if (ret) {
+ dev_err(twl4030_madc->dev,
+ "unable to write sel register 0x%X\n", method->sel + 1);
+ goto out;
+ }
+ ret = twl_i2c_write_u8(TWL4030_MODULE_MADC, ch_lsb, method->sel);
+ if (ret) {
+ dev_err(twl4030_madc->dev,
+ "unable to write sel register 0x%X\n", method->sel + 1);
+ goto out;
+ }
+ /* Select averaging for all channels if do_avg is set */
+ if (req->do_avg) {
+ ret = twl_i2c_write_u8(TWL4030_MODULE_MADC,
+ ch_msb, method->avg + 1);
+ if (ret) {
+ dev_err(twl4030_madc->dev,
+ "unable to write avg register 0x%X\n",
+ method->avg + 1);
+ goto out;
+ }
+ ret = twl_i2c_write_u8(TWL4030_MODULE_MADC,
+ ch_lsb, method->avg);
+ if (ret) {
+ dev_err(twl4030_madc->dev,
+ "unable to write sel reg 0x%X\n",
+ method->sel + 1);
+ goto out;
+ }
+ }
+ if (req->type == TWL4030_MADC_IRQ_ONESHOT && req->func_cb != NULL) {
+ ret = twl4030_madc_set_irq(twl4030_madc, req);
+ if (ret < 0)
+ goto out;
+ ret = twl4030_madc_start_conversion(twl4030_madc, req->method);
+ if (ret < 0)
+ goto out;
+ twl4030_madc->requests[req->method].active = 1;
+ ret = 0;
+ goto out;
+ }
+ /* With RT method we should not be here anymore */
+ if (req->method == TWL4030_MADC_RT) {
+ ret = -EINVAL;
+ goto out;
+ }
+ ret = twl4030_madc_start_conversion(twl4030_madc, req->method);
+ if (ret < 0)
+ goto out;
+ twl4030_madc->requests[req->method].active = 1;
+ /* Wait until conversion is ready (ctrl register returns EOC) */
+ ret = twl4030_madc_wait_conversion_ready(twl4030_madc, 5, method->ctrl);
+ if (ret) {
+ twl4030_madc->requests[req->method].active = 0;
+ goto out;
+ }
+ ret = twl4030_madc_read_channels(twl4030_madc, method->rbase,
+ req->channels, req->rbuf, req->raw);
+ twl4030_madc->requests[req->method].active = 0;
+
+out:
+ mutex_unlock(&twl4030_madc->lock);
+
+ return ret;
+}
+EXPORT_SYMBOL_GPL(twl4030_madc_conversion);
+
+/*
+ * Return channel value
+ * Or < 0 on failure.
+ */
+int twl4030_get_madc_conversion(int channel_no)
+{
+ struct twl4030_madc_request req;
+ int temp = 0;
+ int ret;
+
+ req.channels = (1 << channel_no);
+ req.method = TWL4030_MADC_SW2;
+ req.active = 0;
+ req.func_cb = NULL;
+ ret = twl4030_madc_conversion(&req);
+ if (ret < 0)
+ return ret;
+ if (req.rbuf[channel_no] > 0)
+ temp = req.rbuf[channel_no];
+
+ return temp;
+}
+EXPORT_SYMBOL_GPL(twl4030_get_madc_conversion);
+
+/*
+ * Function to enable or disable bias current for
+ * main battery type reading or temperature sensing
+ * @madc - pointer to twl4030_madc_data struct
+ * @chan - can be one of the two values
+ * TWL4030_BCI_ITHEN - Enables bias current for main battery type reading
+ * TWL4030_BCI_TYPEN - Enables bias current for main battery temperature
+ * sensing
+ * @on - enable or disable chan.
+ */
+static int twl4030_madc_set_current_generator(struct twl4030_madc_data *madc,
+ int chan, int on)
+{
+ int ret;
+ u8 regval;
+
+ ret = twl_i2c_read_u8(TWL_MODULE_MAIN_CHARGE,
+ ®val, TWL4030_BCI_BCICTL1);
+ if (ret) {
+ dev_err(madc->dev, "unable to read BCICTL1 reg 0x%X",
+ TWL4030_BCI_BCICTL1);
+ return ret;
+ }
+ if (on)
+ regval |= chan ? TWL4030_BCI_ITHEN : TWL4030_BCI_TYPEN;
+ else
+ regval &= chan ? ~TWL4030_BCI_ITHEN : ~TWL4030_BCI_TYPEN;
+ ret = twl_i2c_write_u8(TWL_MODULE_MAIN_CHARGE,
+ regval, TWL4030_BCI_BCICTL1);
+ if (ret) {
+ dev_err(madc->dev, "unable to write BCICTL1 reg 0x%X\n",
+ TWL4030_BCI_BCICTL1);
+ return ret;
+ }
+
+ return 0;
+}
+
+/*
+ * Function that sets MADC software power on bit to enable MADC
+ * @madc - pointer to twl4030_madc_data struct
+ * @on - Enable or disable MADC software powen on bit.
+ * returns error if i2c read/write fails else 0
+ */
+static int twl4030_madc_set_power(struct twl4030_madc_data *madc, int on)
+{
+ u8 regval;
+ int ret;
+
+ ret = twl_i2c_read_u8(TWL_MODULE_MAIN_CHARGE,
+ ®val, TWL4030_MADC_CTRL1);
+ if (ret) {
+ dev_err(madc->dev, "unable to read madc ctrl1 reg 0x%X\n",
+ TWL4030_MADC_CTRL1);
+ return ret;
+ }
+ if (on)
+ regval |= TWL4030_MADC_MADCON;
+ else
+ regval &= ~TWL4030_MADC_MADCON;
+ ret = twl_i2c_write_u8(TWL4030_MODULE_MADC, regval, TWL4030_MADC_CTRL1);
+ if (ret) {
+ dev_err(madc->dev, "unable to write madc ctrl1 reg 0x%X\n",
+ TWL4030_MADC_CTRL1);
+ return ret;
+ }
+
+ return 0;
+}
+
+/*
+ * Initialize MADC and request for threaded irq
+ */
+static int twl4030_madc_probe(struct platform_device *pdev)
+{
+ struct twl4030_madc_data *madc;
+ struct twl4030_madc_platform_data *pdata = dev_get_platdata(&pdev->dev);
+ struct device_node *np = pdev->dev.of_node;
+ int irq, ret;
+ u8 regval;
+ struct iio_dev *iio_dev = NULL;
+
+ if (!pdata && !np) {
+ dev_err(&pdev->dev, "platform_data not available\n");
+ return -EINVAL;
+ }
+
+ iio_dev = devm_iio_device_alloc(&pdev->dev,
+ sizeof(struct twl4030_madc_data));
+ if (!iio_dev) {
+ dev_err(&pdev->dev, "failed allocating iio device\n");
+ return -ENOMEM;
+ }
+
+ madc = iio_priv(iio_dev);
+ madc->dev = &pdev->dev;
+
+ iio_dev->name = dev_name(&pdev->dev);
+ iio_dev->dev.parent = &pdev->dev;
+ iio_dev->dev.of_node = pdev->dev.of_node;
+ iio_dev->info = &twl4030_madc_iio_info;
+ iio_dev->modes = INDIO_DIRECT_MODE;
+ iio_dev->channels = twl4030_madc_iio_channels;
+ iio_dev->num_channels = 16;
+
+ /*
+ * Phoenix provides 2 interrupt lines. The first one is connected to
+ * the OMAP. The other one can be connected to the other processor such
+ * as modem. Hence two separate ISR and IMR registers.
+ */
+ if (pdata)
+ madc->use_second_irq = pdata->irq_line != 1;
+ else
+ madc->use_second_irq = false;
+
+ madc->imr = (madc->use_second_irq == 1) ?
+ TWL4030_MADC_IMR1 : TWL4030_MADC_IMR2;
+ madc->isr = (madc->use_second_irq == 1) ?
+ TWL4030_MADC_ISR1 : TWL4030_MADC_ISR2;
+
+ ret = twl4030_madc_set_power(madc, 1);
+ if (ret < 0)
+ return ret;
+ ret = twl4030_madc_set_current_generator(madc, 0, 1);
+ if (ret < 0)
+ goto err_current_generator;
+
+ ret = twl_i2c_read_u8(TWL_MODULE_MAIN_CHARGE,
+ ®val, TWL4030_BCI_BCICTL1);
+ if (ret) {
+ dev_err(&pdev->dev, "unable to read reg BCI CTL1 0x%X\n",
+ TWL4030_BCI_BCICTL1);
+ goto err_i2c;
+ }
+ regval |= TWL4030_BCI_MESBAT;
+ ret = twl_i2c_write_u8(TWL_MODULE_MAIN_CHARGE,
+ regval, TWL4030_BCI_BCICTL1);
+ if (ret) {
+ dev_err(&pdev->dev, "unable to write reg BCI Ctl1 0x%X\n",
+ TWL4030_BCI_BCICTL1);
+ goto err_i2c;
+ }
+
+ /* Check that MADC clock is on */
+ ret = twl_i2c_read_u8(TWL4030_MODULE_INTBR, ®val, TWL4030_REG_GPBR1);
+ if (ret) {
+ dev_err(&pdev->dev, "unable to read reg GPBR1 0x%X\n",
+ TWL4030_REG_GPBR1);
+ goto err_i2c;
+ }
+
+ /* If MADC clk is not on, turn it on */
+ if (!(regval & TWL4030_GPBR1_MADC_HFCLK_EN)) {
+ dev_info(&pdev->dev, "clk disabled, enabling\n");
+ regval |= TWL4030_GPBR1_MADC_HFCLK_EN;
+ ret = twl_i2c_write_u8(TWL4030_MODULE_INTBR, regval,
+ TWL4030_REG_GPBR1);
+ if (ret) {
+ dev_err(&pdev->dev, "unable to write reg GPBR1 0x%X\n",
+ TWL4030_REG_GPBR1);
+ goto err_i2c;
+ }
+ }
+
+ platform_set_drvdata(pdev, iio_dev);
+ mutex_init(&madc->lock);
+
+ irq = platform_get_irq(pdev, 0);
+ ret = devm_request_threaded_irq(&pdev->dev, irq, NULL,
+ twl4030_madc_threaded_irq_handler,
+ IRQF_TRIGGER_RISING, "twl4030_madc", madc);
+ if (ret) {
+ dev_dbg(&pdev->dev, "could not request irq\n");
+ goto err_i2c;
+ }
+ twl4030_madc = madc;
+
+ ret = iio_device_register(iio_dev);
+ if (ret) {
+ dev_dbg(&pdev->dev, "could not register iio device\n");
+ goto err_i2c;
+ }
+
+ return 0;
+
+err_i2c:
+ twl4030_madc_set_current_generator(madc, 0, 0);
+err_current_generator:
+ twl4030_madc_set_power(madc, 0);
+ return ret;
+}
+
+static int twl4030_madc_remove(struct platform_device *pdev)
+{
+ struct iio_dev *iio_dev = platform_get_drvdata(pdev);
+ struct twl4030_madc_data *madc = iio_priv(iio_dev);
+
+ twl4030_madc_set_current_generator(madc, 0, 0);
+ twl4030_madc_set_power(madc, 0);
+
+ iio_device_unregister(iio_dev);
+
+ return 0;
+}
+
+#ifdef CONFIG_OF
+static const struct of_device_id twl_madc_of_match[] = {
+ {.compatible = "ti,twl4030-madc", },
+ { },
+};
+MODULE_DEVICE_TABLE(of, twl_madc_of_match);
+#endif
+
+static struct platform_driver twl4030_madc_driver = {
+ .probe = twl4030_madc_probe,
+ .remove = twl4030_madc_remove,
+ .driver = {
+ .name = "twl4030_madc",
+ .owner = THIS_MODULE,
+ .of_match_table = of_match_ptr(twl_madc_of_match),
+ },
+};
+
+module_platform_driver(twl4030_madc_driver);
+
+MODULE_DESCRIPTION("TWL4030 ADC driver");
+MODULE_LICENSE("GPL");
+MODULE_AUTHOR("J Keerthy");
+MODULE_ALIAS("platform:twl4030_madc");
diff --git a/drivers/mfd/Kconfig b/drivers/mfd/Kconfig
index 49bb445..23a8a51 100644
--- a/drivers/mfd/Kconfig
+++ b/drivers/mfd/Kconfig
@@ -935,16 +935,6 @@ config TWL4030_CORE
high speed USB OTG transceiver, an audio codec (on most
versions) and many other features.
-config TWL4030_MADC
- tristate "TI TWL4030 MADC"
- depends on TWL4030_CORE
- help
- This driver provides support for triton TWL4030-MADC. The
- driver supports both RT and SW conversion methods.
-
- This driver can be built as a module. If so it will be
- named twl4030-madc
-
config TWL4030_POWER
bool "TI TWL4030 power resources"
depends on TWL4030_CORE && ARM
diff --git a/drivers/mfd/Makefile b/drivers/mfd/Makefile
index 5aea5ef..c8eb0bc 100644
--- a/drivers/mfd/Makefile
+++ b/drivers/mfd/Makefile
@@ -71,7 +71,6 @@ obj-$(CONFIG_MFD_TPS80031) += tps80031.o
obj-$(CONFIG_MENELAUS) += menelaus.o
obj-$(CONFIG_TWL4030_CORE) += twl-core.o twl4030-irq.o twl6030-irq.o
-obj-$(CONFIG_TWL4030_MADC) += twl4030-madc.o
obj-$(CONFIG_TWL4030_POWER) += twl4030-power.o
obj-$(CONFIG_MFD_TWL4030_AUDIO) += twl4030-audio.o
obj-$(CONFIG_TWL6040_CORE) += twl6040.o
diff --git a/drivers/mfd/twl4030-madc.c b/drivers/mfd/twl4030-madc.c
deleted file mode 100644
index 4da61c4..0000000
--- a/drivers/mfd/twl4030-madc.c
+++ /dev/null
@@ -1,922 +0,0 @@
-/*
- *
- * TWL4030 MADC module driver-This driver monitors the real time
- * conversion of analog signals like battery temperature,
- * battery type, battery level etc.
- *
- * Copyright (C) 2011 Texas Instruments Incorporated - http://www.ti.com/
- * J Keerthy <j-keerthy@ti.com>
- *
- * Based on twl4030-madc.c
- * Copyright (C) 2008 Nokia Corporation
- * Mikko Ylinen <mikko.k.ylinen@nokia.com>
- *
- * Amit Kucheria <amit.kucheria@canonical.com>
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License
- * version 2 as published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it will be useful, but
- * WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
- * General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write to the Free Software
- * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA
- * 02110-1301 USA
- *
- */
-
-#include <linux/init.h>
-#include <linux/device.h>
-#include <linux/interrupt.h>
-#include <linux/kernel.h>
-#include <linux/delay.h>
-#include <linux/platform_device.h>
-#include <linux/slab.h>
-#include <linux/i2c/twl.h>
-#include <linux/i2c/twl4030-madc.h>
-#include <linux/module.h>
-#include <linux/stddef.h>
-#include <linux/mutex.h>
-#include <linux/bitops.h>
-#include <linux/jiffies.h>
-#include <linux/types.h>
-#include <linux/gfp.h>
-#include <linux/err.h>
-
-#include <linux/iio/iio.h>
-#include <linux/iio/machine.h>
-#include <linux/iio/driver.h>
-
-/*
- * struct twl4030_madc_data - a container for madc info
- * @dev - pointer to device structure for madc
- * @lock - mutex protecting this data structure
- * @requests - Array of request struct corresponding to SW1, SW2 and RT
- * @imr - Interrupt mask register of MADC
- * @isr - Interrupt status register of MADC
- */
-struct twl4030_madc_data {
- struct device *dev;
- struct mutex lock; /* mutex protecting this data structure */
- struct twl4030_madc_request requests[TWL4030_MADC_NUM_METHODS];
- bool use_second_irq;
- int imr;
- int isr;
-};
-
-static int twl4030_madc_read(struct iio_dev *iio_dev,
- const struct iio_chan_spec *chan,
- int *val, int *val2, long mask)
-{
- struct twl4030_madc_data *madc = iio_priv(iio_dev);
- struct twl4030_madc_request req;
- int channel = chan->channel;
- int ret;
-
- req.method = madc->use_second_irq ? TWL4030_MADC_SW2 : TWL4030_MADC_SW1;
-
- req.channels = BIT(channel);
- req.active = 0;
- req.func_cb = NULL;
- req.raw = (mask & IIO_CHAN_INFO_PROCESSED) ? false : true;
- req.do_avg = (mask & IIO_CHAN_INFO_AVERAGE_RAW) ? true : false;
-
- ret = twl4030_madc_conversion(&req);
- if (ret < 0)
- return ret;
-
- *val = req.rbuf[channel];
-
- return IIO_VAL_INT;
-}
-
-static const struct iio_info twl4030_madc_iio_info = {
- .read_raw = &twl4030_madc_read,
- .driver_module = THIS_MODULE,
-};
-
-#define ADC_CHANNEL(_channel, _type, _name, _mask) { \
- .type = _type, \
- .scan_type = IIO_ST('u', 10, 16, 0), \
- .channel = _channel, \
- .info_mask_separate = _mask, \
- .datasheet_name = _name, \
- .indexed = 1, \
-}
-
-static const struct iio_chan_spec twl4030_madc_iio_channels[] = {
- ADC_CHANNEL(0, IIO_VOLTAGE, "ADCIN0", BIT(IIO_CHAN_INFO_AVERAGE_RAW)),
- ADC_CHANNEL(1, IIO_TEMP, "ADCIN1", BIT(IIO_CHAN_INFO_PROCESSED) |
- BIT(IIO_CHAN_INFO_AVERAGE_RAW)),
- ADC_CHANNEL(2, IIO_VOLTAGE, "ADCIN2", BIT(IIO_CHAN_INFO_AVERAGE_RAW)),
- ADC_CHANNEL(3, IIO_VOLTAGE, "ADCIN3", BIT(IIO_CHAN_INFO_AVERAGE_RAW)),
- ADC_CHANNEL(4, IIO_VOLTAGE, "ADCIN4", BIT(IIO_CHAN_INFO_AVERAGE_RAW)),
- ADC_CHANNEL(5, IIO_VOLTAGE, "ADCIN5", BIT(IIO_CHAN_INFO_AVERAGE_RAW)),
- ADC_CHANNEL(6, IIO_VOLTAGE, "ADCIN6", BIT(IIO_CHAN_INFO_AVERAGE_RAW)),
- ADC_CHANNEL(7, IIO_VOLTAGE, "ADCIN7", BIT(IIO_CHAN_INFO_AVERAGE_RAW)),
- ADC_CHANNEL(8, IIO_VOLTAGE, "ADCIN8", BIT(IIO_CHAN_INFO_AVERAGE_RAW)),
- ADC_CHANNEL(9, IIO_VOLTAGE, "ADCIN9", BIT(IIO_CHAN_INFO_AVERAGE_RAW)),
- ADC_CHANNEL(10, IIO_CURRENT, "ADCIN10", BIT(IIO_CHAN_INFO_PROCESSED) |
- BIT(IIO_CHAN_INFO_AVERAGE_RAW)),
- ADC_CHANNEL(11, IIO_VOLTAGE, "ADCIN11", BIT(IIO_CHAN_INFO_AVERAGE_RAW)),
- ADC_CHANNEL(12, IIO_VOLTAGE, "ADCIN12", BIT(IIO_CHAN_INFO_AVERAGE_RAW)),
- ADC_CHANNEL(13, IIO_VOLTAGE, "ADCIN13", BIT(IIO_CHAN_INFO_AVERAGE_RAW)),
- ADC_CHANNEL(14, IIO_VOLTAGE, "ADCIN14", BIT(IIO_CHAN_INFO_AVERAGE_RAW)),
- ADC_CHANNEL(15, IIO_VOLTAGE, "ADCIN15", BIT(IIO_CHAN_INFO_AVERAGE_RAW)),
-};
-
-static struct twl4030_madc_data *twl4030_madc;
-
-struct twl4030_prescale_divider_ratios {
- s16 numerator;
- s16 denominator;
-};
-
-static const struct twl4030_prescale_divider_ratios
-twl4030_divider_ratios[16] = {
- {1, 1}, /* CHANNEL 0 No Prescaler */
- {1, 1}, /* CHANNEL 1 No Prescaler */
- {6, 10}, /* CHANNEL 2 */
- {6, 10}, /* CHANNEL 3 */
- {6, 10}, /* CHANNEL 4 */
- {6, 10}, /* CHANNEL 5 */
- {6, 10}, /* CHANNEL 6 */
- {6, 10}, /* CHANNEL 7 */
- {3, 14}, /* CHANNEL 8 */
- {1, 3}, /* CHANNEL 9 */
- {1, 1}, /* CHANNEL 10 No Prescaler */
- {15, 100}, /* CHANNEL 11 */
- {1, 4}, /* CHANNEL 12 */
- {1, 1}, /* CHANNEL 13 Reserved channels */
- {1, 1}, /* CHANNEL 14 Reseved channels */
- {5, 11}, /* CHANNEL 15 */
-};
-
-
-/*
- * Conversion table from -3 to 55 degree Celcius
- */
-static int therm_tbl[] = {
-30800, 29500, 28300, 27100,
-26000, 24900, 23900, 22900, 22000, 21100, 20300, 19400, 18700, 17900,
-17200, 16500, 15900, 15300, 14700, 14100, 13600, 13100, 12600, 12100,
-11600, 11200, 10800, 10400, 10000, 9630, 9280, 8950, 8620, 8310,
-8020, 7730, 7460, 7200, 6950, 6710, 6470, 6250, 6040, 5830,
-5640, 5450, 5260, 5090, 4920, 4760, 4600, 4450, 4310, 4170,
-4040, 3910, 3790, 3670, 3550
-};
-
-/*
- * Structure containing the registers
- * of different conversion methods supported by MADC.
- * Hardware or RT real time conversion request initiated by external host
- * processor for RT Signal conversions.
- * External host processors can also request for non RT conversions
- * SW1 and SW2 software conversions also called asynchronous or GPC request.
- */
-static
-const struct twl4030_madc_conversion_method twl4030_conversion_methods[] = {
- [TWL4030_MADC_RT] = {
- .sel = TWL4030_MADC_RTSELECT_LSB,
- .avg = TWL4030_MADC_RTAVERAGE_LSB,
- .rbase = TWL4030_MADC_RTCH0_LSB,
- },
- [TWL4030_MADC_SW1] = {
- .sel = TWL4030_MADC_SW1SELECT_LSB,
- .avg = TWL4030_MADC_SW1AVERAGE_LSB,
- .rbase = TWL4030_MADC_GPCH0_LSB,
- .ctrl = TWL4030_MADC_CTRL_SW1,
- },
- [TWL4030_MADC_SW2] = {
- .sel = TWL4030_MADC_SW2SELECT_LSB,
- .avg = TWL4030_MADC_SW2AVERAGE_LSB,
- .rbase = TWL4030_MADC_GPCH0_LSB,
- .ctrl = TWL4030_MADC_CTRL_SW2,
- },
-};
-
-/*
- * Function to read a particular channel value.
- * @madc - pointer to struct twl4030_madc_data
- * @reg - lsb of ADC Channel
- * If the i2c read fails it returns an error else returns 0.
- */
-static int twl4030_madc_channel_raw_read(struct twl4030_madc_data *madc, u8 reg)
-{
- u8 msb, lsb;
- int ret;
- /*
- * For each ADC channel, we have MSB and LSB register pair. MSB address
- * is always LSB address+1. reg parameter is the address of LSB register
- */
- ret = twl_i2c_read_u8(TWL4030_MODULE_MADC, &msb, reg + 1);
- if (ret) {
- dev_err(madc->dev, "unable to read MSB register 0x%X\n",
- reg + 1);
- return ret;
- }
- ret = twl_i2c_read_u8(TWL4030_MODULE_MADC, &lsb, reg);
- if (ret) {
- dev_err(madc->dev, "unable to read LSB register 0x%X\n", reg);
- return ret;
- }
-
- return (int)(((msb << 8) | lsb) >> 6);
-}
-
-/*
- * Return battery temperature
- * Or < 0 on failure.
- */
-static int twl4030battery_temperature(int raw_volt)
-{
- u8 val;
- int temp, curr, volt, res, ret;
-
- volt = (raw_volt * TEMP_STEP_SIZE) / TEMP_PSR_R;
- /* Getting and calculating the supply current in micro ampers */
- ret = twl_i2c_read_u8(TWL_MODULE_MAIN_CHARGE, &val,
- REG_BCICTL2);
- if (ret < 0)
- return ret;
- curr = ((val & TWL4030_BCI_ITHEN) + 1) * 10;
- /* Getting and calculating the thermistor resistance in ohms */
- res = volt * 1000 / curr;
- /* calculating temperature */
- for (temp = 58; temp >= 0; temp--) {
- int actual = therm_tbl[temp];
-
- if ((actual - res) >= 0)
- break;
- }
-
- return temp + 1;
-}
-
-static int twl4030battery_current(int raw_volt)
-{
- int ret;
- u8 val;
-
- ret = twl_i2c_read_u8(TWL_MODULE_MAIN_CHARGE, &val,
- TWL4030_BCI_BCICTL1);
- if (ret)
- return ret;
- if (val & TWL4030_BCI_CGAIN) /* slope of 0.44 mV/mA */
- return (raw_volt * CURR_STEP_SIZE) / CURR_PSR_R1;
- else /* slope of 0.88 mV/mA */
- return (raw_volt * CURR_STEP_SIZE) / CURR_PSR_R2;
-}
-/*
- * Function to read channel values
- * @madc - pointer to twl4030_madc_data struct
- * @reg_base - Base address of the first channel
- * @Channels - 16 bit bitmap. If the bit is set, channel value is read
- * @buf - The channel values are stored here. if read fails error
- * @raw - Return raw values without conversion
- * value is stored
- * Returns the number of successfully read channels.
- */
-static int twl4030_madc_read_channels(struct twl4030_madc_data *madc,
- u8 reg_base, unsigned
- long channels, int *buf,
- bool raw)
-{
- int count = 0, count_req = 0, i;
- u8 reg;
-
- for_each_set_bit(i, &channels, TWL4030_MADC_MAX_CHANNELS) {
- reg = reg_base + 2 * i;
- buf[i] = twl4030_madc_channel_raw_read(madc, reg);
- if (buf[i] < 0) {
- dev_err(madc->dev,
- "Unable to read register 0x%X\n", reg);
- count_req++;
- continue;
- }
- if (raw) {
- count++;
- continue;
- }
- switch (i) {
- case 10:
- buf[i] = twl4030battery_current(buf[i]);
- if (buf[i] < 0) {
- dev_err(madc->dev, "err reading current\n");
- count_req++;
- } else {
- count++;
- buf[i] = buf[i] - 750;
- }
- break;
- case 1:
- buf[i] = twl4030battery_temperature(buf[i]);
- if (buf[i] < 0) {
- dev_err(madc->dev, "err reading temperature\n");
- count_req++;
- } else {
- buf[i] -= 3;
- count++;
- }
- break;
- default:
- count++;
- /* Analog Input (V) = conv_result * step_size / R
- * conv_result = decimal value of 10-bit conversion
- * result
- * step size = 1.5 / (2 ^ 10 -1)
- * R = Prescaler ratio for input channels.
- * Result given in mV hence multiplied by 1000.
- */
- buf[i] = (buf[i] * 3 * 1000 *
- twl4030_divider_ratios[i].denominator)
- / (2 * 1023 *
- twl4030_divider_ratios[i].numerator);
- }
- }
- if (count_req)
- dev_err(madc->dev, "%d channel conversion failed\n", count_req);
-
- return count;
-}
-
-/*
- * Enables irq.
- * @madc - pointer to twl4030_madc_data struct
- * @id - irq number to be enabled
- * can take one of TWL4030_MADC_RT, TWL4030_MADC_SW1, TWL4030_MADC_SW2
- * corresponding to RT, SW1, SW2 conversion requests.
- * If the i2c read fails it returns an error else returns 0.
- */
-static int twl4030_madc_enable_irq(struct twl4030_madc_data *madc, u8 id)
-{
- u8 val;
- int ret;
-
- ret = twl_i2c_read_u8(TWL4030_MODULE_MADC, &val, madc->imr);
- if (ret) {
- dev_err(madc->dev, "unable to read imr register 0x%X\n",
- madc->imr);
- return ret;
- }
- val &= ~(1 << id);
- ret = twl_i2c_write_u8(TWL4030_MODULE_MADC, val, madc->imr);
- if (ret) {
- dev_err(madc->dev,
- "unable to write imr register 0x%X\n", madc->imr);
- return ret;
-
- }
-
- return 0;
-}
-
-/*
- * Disables irq.
- * @madc - pointer to twl4030_madc_data struct
- * @id - irq number to be disabled
- * can take one of TWL4030_MADC_RT, TWL4030_MADC_SW1, TWL4030_MADC_SW2
- * corresponding to RT, SW1, SW2 conversion requests.
- * Returns error if i2c read/write fails.
- */
-static int twl4030_madc_disable_irq(struct twl4030_madc_data *madc, u8 id)
-{
- u8 val;
- int ret;
-
- ret = twl_i2c_read_u8(TWL4030_MODULE_MADC, &val, madc->imr);
- if (ret) {
- dev_err(madc->dev, "unable to read imr register 0x%X\n",
- madc->imr);
- return ret;
- }
- val |= (1 << id);
- ret = twl_i2c_write_u8(TWL4030_MODULE_MADC, val, madc->imr);
- if (ret) {
- dev_err(madc->dev,
- "unable to write imr register 0x%X\n", madc->imr);
- return ret;
- }
-
- return 0;
-}
-
-static irqreturn_t twl4030_madc_threaded_irq_handler(int irq, void *_madc)
-{
- struct twl4030_madc_data *madc = _madc;
- const struct twl4030_madc_conversion_method *method;
- u8 isr_val, imr_val;
- int i, len, ret;
- struct twl4030_madc_request *r;
-
- mutex_lock(&madc->lock);
- ret = twl_i2c_read_u8(TWL4030_MODULE_MADC, &isr_val, madc->isr);
- if (ret) {
- dev_err(madc->dev, "unable to read isr register 0x%X\n",
- madc->isr);
- goto err_i2c;
- }
- ret = twl_i2c_read_u8(TWL4030_MODULE_MADC, &imr_val, madc->imr);
- if (ret) {
- dev_err(madc->dev, "unable to read imr register 0x%X\n",
- madc->imr);
- goto err_i2c;
- }
- isr_val &= ~imr_val;
- for (i = 0; i < TWL4030_MADC_NUM_METHODS; i++) {
- if (!(isr_val & (1 << i)))
- continue;
- ret = twl4030_madc_disable_irq(madc, i);
- if (ret < 0)
- dev_dbg(madc->dev, "Disable interrupt failed%d\n", i);
- madc->requests[i].result_pending = 1;
- }
- for (i = 0; i < TWL4030_MADC_NUM_METHODS; i++) {
- r = &madc->requests[i];
- /* No pending results for this method, move to next one */
- if (!r->result_pending)
- continue;
- method = &twl4030_conversion_methods[r->method];
- /* Read results */
- len = twl4030_madc_read_channels(madc, method->rbase,
- r->channels, r->rbuf, r->raw);
- /* Return results to caller */
- if (r->func_cb != NULL) {
- r->func_cb(len, r->channels, r->rbuf);
- r->func_cb = NULL;
- }
- /* Free request */
- r->result_pending = 0;
- r->active = 0;
- }
- mutex_unlock(&madc->lock);
-
- return IRQ_HANDLED;
-
-err_i2c:
- /*
- * In case of error check whichever request is active
- * and service the same.
- */
- for (i = 0; i < TWL4030_MADC_NUM_METHODS; i++) {
- r = &madc->requests[i];
- if (r->active == 0)
- continue;
- method = &twl4030_conversion_methods[r->method];
- /* Read results */
- len = twl4030_madc_read_channels(madc, method->rbase,
- r->channels, r->rbuf, r->raw);
- /* Return results to caller */
- if (r->func_cb != NULL) {
- r->func_cb(len, r->channels, r->rbuf);
- r->func_cb = NULL;
- }
- /* Free request */
- r->result_pending = 0;
- r->active = 0;
- }
- mutex_unlock(&madc->lock);
-
- return IRQ_HANDLED;
-}
-
-static int twl4030_madc_set_irq(struct twl4030_madc_data *madc,
- struct twl4030_madc_request *req)
-{
- struct twl4030_madc_request *p;
- int ret;
-
- p = &madc->requests[req->method];
- memcpy(p, req, sizeof(*req));
- ret = twl4030_madc_enable_irq(madc, req->method);
- if (ret < 0) {
- dev_err(madc->dev, "enable irq failed!!\n");
- return ret;
- }
-
- return 0;
-}
-
-/*
- * Function which enables the madc conversion
- * by writing to the control register.
- * @madc - pointer to twl4030_madc_data struct
- * @conv_method - can be TWL4030_MADC_RT, TWL4030_MADC_SW2, TWL4030_MADC_SW1
- * corresponding to RT SW1 or SW2 conversion methods.
- * Returns 0 if succeeds else a negative error value
- */
-static int twl4030_madc_start_conversion(struct twl4030_madc_data *madc,
- int conv_method)
-{
- const struct twl4030_madc_conversion_method *method;
- int ret = 0;
- method = &twl4030_conversion_methods[conv_method];
- switch (conv_method) {
- case TWL4030_MADC_SW1:
- case TWL4030_MADC_SW2:
- ret = twl_i2c_write_u8(TWL4030_MODULE_MADC,
- TWL4030_MADC_SW_START, method->ctrl);
- if (ret) {
- dev_err(madc->dev,
- "unable to write ctrl register 0x%X\n",
- method->ctrl);
- return ret;
- }
- break;
- default:
- break;
- }
-
- return 0;
-}
-
-/*
- * Function that waits for conversion to be ready
- * @madc - pointer to twl4030_madc_data struct
- * @timeout_ms - timeout value in milliseconds
- * @status_reg - ctrl register
- * returns 0 if succeeds else a negative error value
- */
-static int twl4030_madc_wait_conversion_ready(struct twl4030_madc_data *madc,
- unsigned int timeout_ms,
- u8 status_reg)
-{
- unsigned long timeout;
- int ret;
-
- timeout = jiffies + msecs_to_jiffies(timeout_ms);
- do {
- u8 reg;
-
- ret = twl_i2c_read_u8(TWL4030_MODULE_MADC, ®, status_reg);
- if (ret) {
- dev_err(madc->dev,
- "unable to read status register 0x%X\n",
- status_reg);
- return ret;
- }
- if (!(reg & TWL4030_MADC_BUSY) && (reg & TWL4030_MADC_EOC_SW))
- return 0;
- usleep_range(500, 2000);
- } while (!time_after(jiffies, timeout));
- dev_err(madc->dev, "conversion timeout!\n");
-
- return -EAGAIN;
-}
-
-/*
- * An exported function which can be called from other kernel drivers.
- * @req twl4030_madc_request structure
- * req->rbuf will be filled with read values of channels based on the
- * channel index. If a particular channel reading fails there will
- * be a negative error value in the corresponding array element.
- * returns 0 if succeeds else error value
- */
-int twl4030_madc_conversion(struct twl4030_madc_request *req)
-{
- const struct twl4030_madc_conversion_method *method;
- u8 ch_msb, ch_lsb;
- int ret;
-
- if (!req || !twl4030_madc)
- return -EINVAL;
-
- mutex_lock(&twl4030_madc->lock);
- if (req->method < TWL4030_MADC_RT || req->method > TWL4030_MADC_SW2) {
- ret = -EINVAL;
- goto out;
- }
- /* Do we have a conversion request ongoing */
- if (twl4030_madc->requests[req->method].active) {
- ret = -EBUSY;
- goto out;
- }
- ch_msb = (req->channels >> 8) & 0xff;
- ch_lsb = req->channels & 0xff;
- method = &twl4030_conversion_methods[req->method];
- /* Select channels to be converted */
- ret = twl_i2c_write_u8(TWL4030_MODULE_MADC, ch_msb, method->sel + 1);
- if (ret) {
- dev_err(twl4030_madc->dev,
- "unable to write sel register 0x%X\n", method->sel + 1);
- goto out;
- }
- ret = twl_i2c_write_u8(TWL4030_MODULE_MADC, ch_lsb, method->sel);
- if (ret) {
- dev_err(twl4030_madc->dev,
- "unable to write sel register 0x%X\n", method->sel + 1);
- goto out;
- }
- /* Select averaging for all channels if do_avg is set */
- if (req->do_avg) {
- ret = twl_i2c_write_u8(TWL4030_MODULE_MADC,
- ch_msb, method->avg + 1);
- if (ret) {
- dev_err(twl4030_madc->dev,
- "unable to write avg register 0x%X\n",
- method->avg + 1);
- goto out;
- }
- ret = twl_i2c_write_u8(TWL4030_MODULE_MADC,
- ch_lsb, method->avg);
- if (ret) {
- dev_err(twl4030_madc->dev,
- "unable to write sel reg 0x%X\n",
- method->sel + 1);
- goto out;
- }
- }
- if (req->type == TWL4030_MADC_IRQ_ONESHOT && req->func_cb != NULL) {
- ret = twl4030_madc_set_irq(twl4030_madc, req);
- if (ret < 0)
- goto out;
- ret = twl4030_madc_start_conversion(twl4030_madc, req->method);
- if (ret < 0)
- goto out;
- twl4030_madc->requests[req->method].active = 1;
- ret = 0;
- goto out;
- }
- /* With RT method we should not be here anymore */
- if (req->method == TWL4030_MADC_RT) {
- ret = -EINVAL;
- goto out;
- }
- ret = twl4030_madc_start_conversion(twl4030_madc, req->method);
- if (ret < 0)
- goto out;
- twl4030_madc->requests[req->method].active = 1;
- /* Wait until conversion is ready (ctrl register returns EOC) */
- ret = twl4030_madc_wait_conversion_ready(twl4030_madc, 5, method->ctrl);
- if (ret) {
- twl4030_madc->requests[req->method].active = 0;
- goto out;
- }
- ret = twl4030_madc_read_channels(twl4030_madc, method->rbase,
- req->channels, req->rbuf, req->raw);
- twl4030_madc->requests[req->method].active = 0;
-
-out:
- mutex_unlock(&twl4030_madc->lock);
-
- return ret;
-}
-EXPORT_SYMBOL_GPL(twl4030_madc_conversion);
-
-/*
- * Return channel value
- * Or < 0 on failure.
- */
-int twl4030_get_madc_conversion(int channel_no)
-{
- struct twl4030_madc_request req;
- int temp = 0;
- int ret;
-
- req.channels = (1 << channel_no);
- req.method = TWL4030_MADC_SW2;
- req.active = 0;
- req.func_cb = NULL;
- ret = twl4030_madc_conversion(&req);
- if (ret < 0)
- return ret;
- if (req.rbuf[channel_no] > 0)
- temp = req.rbuf[channel_no];
-
- return temp;
-}
-EXPORT_SYMBOL_GPL(twl4030_get_madc_conversion);
-
-/*
- * Function to enable or disable bias current for
- * main battery type reading or temperature sensing
- * @madc - pointer to twl4030_madc_data struct
- * @chan - can be one of the two values
- * TWL4030_BCI_ITHEN - Enables bias current for main battery type reading
- * TWL4030_BCI_TYPEN - Enables bias current for main battery temperature
- * sensing
- * @on - enable or disable chan.
- */
-static int twl4030_madc_set_current_generator(struct twl4030_madc_data *madc,
- int chan, int on)
-{
- int ret;
- u8 regval;
-
- ret = twl_i2c_read_u8(TWL_MODULE_MAIN_CHARGE,
- ®val, TWL4030_BCI_BCICTL1);
- if (ret) {
- dev_err(madc->dev, "unable to read BCICTL1 reg 0x%X",
- TWL4030_BCI_BCICTL1);
- return ret;
- }
- if (on)
- regval |= chan ? TWL4030_BCI_ITHEN : TWL4030_BCI_TYPEN;
- else
- regval &= chan ? ~TWL4030_BCI_ITHEN : ~TWL4030_BCI_TYPEN;
- ret = twl_i2c_write_u8(TWL_MODULE_MAIN_CHARGE,
- regval, TWL4030_BCI_BCICTL1);
- if (ret) {
- dev_err(madc->dev, "unable to write BCICTL1 reg 0x%X\n",
- TWL4030_BCI_BCICTL1);
- return ret;
- }
-
- return 0;
-}
-
-/*
- * Function that sets MADC software power on bit to enable MADC
- * @madc - pointer to twl4030_madc_data struct
- * @on - Enable or disable MADC software powen on bit.
- * returns error if i2c read/write fails else 0
- */
-static int twl4030_madc_set_power(struct twl4030_madc_data *madc, int on)
-{
- u8 regval;
- int ret;
-
- ret = twl_i2c_read_u8(TWL_MODULE_MAIN_CHARGE,
- ®val, TWL4030_MADC_CTRL1);
- if (ret) {
- dev_err(madc->dev, "unable to read madc ctrl1 reg 0x%X\n",
- TWL4030_MADC_CTRL1);
- return ret;
- }
- if (on)
- regval |= TWL4030_MADC_MADCON;
- else
- regval &= ~TWL4030_MADC_MADCON;
- ret = twl_i2c_write_u8(TWL4030_MODULE_MADC, regval, TWL4030_MADC_CTRL1);
- if (ret) {
- dev_err(madc->dev, "unable to write madc ctrl1 reg 0x%X\n",
- TWL4030_MADC_CTRL1);
- return ret;
- }
-
- return 0;
-}
-
-/*
- * Initialize MADC and request for threaded irq
- */
-static int twl4030_madc_probe(struct platform_device *pdev)
-{
- struct twl4030_madc_data *madc;
- struct twl4030_madc_platform_data *pdata = dev_get_platdata(&pdev->dev);
- struct device_node *np = pdev->dev.of_node;
- int irq, ret;
- u8 regval;
- struct iio_dev *iio_dev = NULL;
-
- if (!pdata && !np) {
- dev_err(&pdev->dev, "platform_data not available\n");
- return -EINVAL;
- }
-
- iio_dev = devm_iio_device_alloc(&pdev->dev,
- sizeof(struct twl4030_madc_data));
- if (!iio_dev) {
- dev_err(&pdev->dev, "failed allocating iio device\n");
- return -ENOMEM;
- }
-
- madc = iio_priv(iio_dev);
- madc->dev = &pdev->dev;
-
- iio_dev->name = dev_name(&pdev->dev);
- iio_dev->dev.parent = &pdev->dev;
- iio_dev->dev.of_node = pdev->dev.of_node;
- iio_dev->info = &twl4030_madc_iio_info;
- iio_dev->modes = INDIO_DIRECT_MODE;
- iio_dev->channels = twl4030_madc_iio_channels;
- iio_dev->num_channels = 16;
-
- /*
- * Phoenix provides 2 interrupt lines. The first one is connected to
- * the OMAP. The other one can be connected to the other processor such
- * as modem. Hence two separate ISR and IMR registers.
- */
- if (pdata)
- madc->use_second_irq = pdata->irq_line != 1;
- else
- madc->use_second_irq = false;
-
- madc->imr = (madc->use_second_irq == 1) ?
- TWL4030_MADC_IMR1 : TWL4030_MADC_IMR2;
- madc->isr = (madc->use_second_irq == 1) ?
- TWL4030_MADC_ISR1 : TWL4030_MADC_ISR2;
-
- ret = twl4030_madc_set_power(madc, 1);
- if (ret < 0)
- return ret;
- ret = twl4030_madc_set_current_generator(madc, 0, 1);
- if (ret < 0)
- goto err_current_generator;
-
- ret = twl_i2c_read_u8(TWL_MODULE_MAIN_CHARGE,
- ®val, TWL4030_BCI_BCICTL1);
- if (ret) {
- dev_err(&pdev->dev, "unable to read reg BCI CTL1 0x%X\n",
- TWL4030_BCI_BCICTL1);
- goto err_i2c;
- }
- regval |= TWL4030_BCI_MESBAT;
- ret = twl_i2c_write_u8(TWL_MODULE_MAIN_CHARGE,
- regval, TWL4030_BCI_BCICTL1);
- if (ret) {
- dev_err(&pdev->dev, "unable to write reg BCI Ctl1 0x%X\n",
- TWL4030_BCI_BCICTL1);
- goto err_i2c;
- }
-
- /* Check that MADC clock is on */
- ret = twl_i2c_read_u8(TWL4030_MODULE_INTBR, ®val, TWL4030_REG_GPBR1);
- if (ret) {
- dev_err(&pdev->dev, "unable to read reg GPBR1 0x%X\n",
- TWL4030_REG_GPBR1);
- goto err_i2c;
- }
-
- /* If MADC clk is not on, turn it on */
- if (!(regval & TWL4030_GPBR1_MADC_HFCLK_EN)) {
- dev_info(&pdev->dev, "clk disabled, enabling\n");
- regval |= TWL4030_GPBR1_MADC_HFCLK_EN;
- ret = twl_i2c_write_u8(TWL4030_MODULE_INTBR, regval,
- TWL4030_REG_GPBR1);
- if (ret) {
- dev_err(&pdev->dev, "unable to write reg GPBR1 0x%X\n",
- TWL4030_REG_GPBR1);
- goto err_i2c;
- }
- }
-
- platform_set_drvdata(pdev, iio_dev);
- mutex_init(&madc->lock);
-
- irq = platform_get_irq(pdev, 0);
- ret = devm_request_threaded_irq(&pdev->dev, irq, NULL,
- twl4030_madc_threaded_irq_handler,
- IRQF_TRIGGER_RISING, "twl4030_madc", madc);
- if (ret) {
- dev_dbg(&pdev->dev, "could not request irq\n");
- goto err_i2c;
- }
- twl4030_madc = madc;
-
- ret = iio_device_register(iio_dev);
- if (ret) {
- dev_dbg(&pdev->dev, "could not register iio device\n");
- goto err_i2c;
- }
-
- return 0;
-
-err_i2c:
- twl4030_madc_set_current_generator(madc, 0, 0);
-err_current_generator:
- twl4030_madc_set_power(madc, 0);
- return ret;
-}
-
-static int twl4030_madc_remove(struct platform_device *pdev)
-{
- struct iio_dev *iio_dev = platform_get_drvdata(pdev);
- struct twl4030_madc_data *madc = iio_priv(iio_dev);
-
- twl4030_madc_set_current_generator(madc, 0, 0);
- twl4030_madc_set_power(madc, 0);
-
- iio_device_unregister(iio_dev);
-
- return 0;
-}
-
-#ifdef CONFIG_OF
-static const struct of_device_id twl_madc_of_match[] = {
- {.compatible = "ti,twl4030-madc", },
- { },
-};
-MODULE_DEVICE_TABLE(of, twl_madc_of_match);
-#endif
-
-static struct platform_driver twl4030_madc_driver = {
- .probe = twl4030_madc_probe,
- .remove = twl4030_madc_remove,
- .driver = {
- .name = "twl4030_madc",
- .owner = THIS_MODULE,
- .of_match_table = of_match_ptr(twl_madc_of_match),
- },
-};
-
-module_platform_driver(twl4030_madc_driver);
-
-MODULE_DESCRIPTION("TWL4030 ADC driver");
-MODULE_LICENSE("GPL");
-MODULE_AUTHOR("J Keerthy");
-MODULE_ALIAS("platform:twl4030_madc");
--
1.8.5.3
^ permalink raw reply related
* [RFCv1 3/4] Documentation: DT: Document twl4030-madc binding
From: Sebastian Reichel @ 2014-02-14 18:46 UTC (permalink / raw)
To: Sebastian Reichel, Marek Belisko
Cc: Jonathan Cameron, Lee Jones, Samuel Ortiz, Lars-Peter Clausen,
Rob Herring, Pawel Moll, Mark Rutland, Ian Campbell, Kumar Gala,
Grant Likely, linux-kernel, devicetree, linux-iio,
Sebastian Reichel
In-Reply-To: <1392403586-30540-1-git-send-email-sre@debian.org>
Add devicetree binding documentation for twl4030-madc
analog digital converter.
Signed-off-by: Sebastian Reichel <sre@debian.org>
---
.../devicetree/bindings/iio/adc/twl4030-madc.txt | 19 +++++++++++++++++++
1 file changed, 19 insertions(+)
create mode 100644 Documentation/devicetree/bindings/iio/adc/twl4030-madc.txt
diff --git a/Documentation/devicetree/bindings/iio/adc/twl4030-madc.txt b/Documentation/devicetree/bindings/iio/adc/twl4030-madc.txt
new file mode 100644
index 0000000..d074286
--- /dev/null
+++ b/Documentation/devicetree/bindings/iio/adc/twl4030-madc.txt
@@ -0,0 +1,19 @@
+* TWL4030 Monitoring Analog to Digital Converter (MADC)
+
+The MADC subsystem in the TWL4030 consists of a 10-bit ADC
+combined with a 16-input analog multiplexer.
+
+Required properties:
+ - compatible: Should contain "ti,twl4030-madc".
+ - interrupts: IRQ line for the MADC submodule.
+ - #io-channel-cells: Should be set to <1>.
+
+Example:
+
+&twl {
+ madc {
+ compatible = "ti,twl4030-madc";
+ interrupts = <3>;
+ #io-channel-cells = <1>;
+ };
+};
--
1.8.5.3
^ permalink raw reply related
* [RFCv1 2/4] mfd: twl4030-madc: Add DT support and convert to IIO framework
From: Sebastian Reichel @ 2014-02-14 18:46 UTC (permalink / raw)
To: Sebastian Reichel, Marek Belisko
Cc: Jonathan Cameron, Lee Jones, Samuel Ortiz, Lars-Peter Clausen,
Rob Herring, Pawel Moll, Mark Rutland, Ian Campbell, Kumar Gala,
Grant Likely, linux-kernel, devicetree, linux-iio,
Sebastian Reichel
In-Reply-To: <1392403586-30540-1-git-send-email-sre@debian.org>
This converts twl4030-madc module to use the Industrial IO ADC
framework and adds device tree support.
Signed-off-by: Sebastian Reichel <sre@debian.org>
---
drivers/mfd/twl4030-madc.c | 121 ++++++++++++++++++++++++++++++++++++++++++---
1 file changed, 114 insertions(+), 7 deletions(-)
diff --git a/drivers/mfd/twl4030-madc.c b/drivers/mfd/twl4030-madc.c
index 5458561..4da61c4 100644
--- a/drivers/mfd/twl4030-madc.c
+++ b/drivers/mfd/twl4030-madc.c
@@ -47,6 +47,10 @@
#include <linux/gfp.h>
#include <linux/err.h>
+#include <linux/iio/iio.h>
+#include <linux/iio/machine.h>
+#include <linux/iio/driver.h>
+
/*
* struct twl4030_madc_data - a container for madc info
* @dev - pointer to device structure for madc
@@ -59,10 +63,72 @@ struct twl4030_madc_data {
struct device *dev;
struct mutex lock; /* mutex protecting this data structure */
struct twl4030_madc_request requests[TWL4030_MADC_NUM_METHODS];
+ bool use_second_irq;
int imr;
int isr;
};
+static int twl4030_madc_read(struct iio_dev *iio_dev,
+ const struct iio_chan_spec *chan,
+ int *val, int *val2, long mask)
+{
+ struct twl4030_madc_data *madc = iio_priv(iio_dev);
+ struct twl4030_madc_request req;
+ int channel = chan->channel;
+ int ret;
+
+ req.method = madc->use_second_irq ? TWL4030_MADC_SW2 : TWL4030_MADC_SW1;
+
+ req.channels = BIT(channel);
+ req.active = 0;
+ req.func_cb = NULL;
+ req.raw = (mask & IIO_CHAN_INFO_PROCESSED) ? false : true;
+ req.do_avg = (mask & IIO_CHAN_INFO_AVERAGE_RAW) ? true : false;
+
+ ret = twl4030_madc_conversion(&req);
+ if (ret < 0)
+ return ret;
+
+ *val = req.rbuf[channel];
+
+ return IIO_VAL_INT;
+}
+
+static const struct iio_info twl4030_madc_iio_info = {
+ .read_raw = &twl4030_madc_read,
+ .driver_module = THIS_MODULE,
+};
+
+#define ADC_CHANNEL(_channel, _type, _name, _mask) { \
+ .type = _type, \
+ .scan_type = IIO_ST('u', 10, 16, 0), \
+ .channel = _channel, \
+ .info_mask_separate = _mask, \
+ .datasheet_name = _name, \
+ .indexed = 1, \
+}
+
+static const struct iio_chan_spec twl4030_madc_iio_channels[] = {
+ ADC_CHANNEL(0, IIO_VOLTAGE, "ADCIN0", BIT(IIO_CHAN_INFO_AVERAGE_RAW)),
+ ADC_CHANNEL(1, IIO_TEMP, "ADCIN1", BIT(IIO_CHAN_INFO_PROCESSED) |
+ BIT(IIO_CHAN_INFO_AVERAGE_RAW)),
+ ADC_CHANNEL(2, IIO_VOLTAGE, "ADCIN2", BIT(IIO_CHAN_INFO_AVERAGE_RAW)),
+ ADC_CHANNEL(3, IIO_VOLTAGE, "ADCIN3", BIT(IIO_CHAN_INFO_AVERAGE_RAW)),
+ ADC_CHANNEL(4, IIO_VOLTAGE, "ADCIN4", BIT(IIO_CHAN_INFO_AVERAGE_RAW)),
+ ADC_CHANNEL(5, IIO_VOLTAGE, "ADCIN5", BIT(IIO_CHAN_INFO_AVERAGE_RAW)),
+ ADC_CHANNEL(6, IIO_VOLTAGE, "ADCIN6", BIT(IIO_CHAN_INFO_AVERAGE_RAW)),
+ ADC_CHANNEL(7, IIO_VOLTAGE, "ADCIN7", BIT(IIO_CHAN_INFO_AVERAGE_RAW)),
+ ADC_CHANNEL(8, IIO_VOLTAGE, "ADCIN8", BIT(IIO_CHAN_INFO_AVERAGE_RAW)),
+ ADC_CHANNEL(9, IIO_VOLTAGE, "ADCIN9", BIT(IIO_CHAN_INFO_AVERAGE_RAW)),
+ ADC_CHANNEL(10, IIO_CURRENT, "ADCIN10", BIT(IIO_CHAN_INFO_PROCESSED) |
+ BIT(IIO_CHAN_INFO_AVERAGE_RAW)),
+ ADC_CHANNEL(11, IIO_VOLTAGE, "ADCIN11", BIT(IIO_CHAN_INFO_AVERAGE_RAW)),
+ ADC_CHANNEL(12, IIO_VOLTAGE, "ADCIN12", BIT(IIO_CHAN_INFO_AVERAGE_RAW)),
+ ADC_CHANNEL(13, IIO_VOLTAGE, "ADCIN13", BIT(IIO_CHAN_INFO_AVERAGE_RAW)),
+ ADC_CHANNEL(14, IIO_VOLTAGE, "ADCIN14", BIT(IIO_CHAN_INFO_AVERAGE_RAW)),
+ ADC_CHANNEL(15, IIO_VOLTAGE, "ADCIN15", BIT(IIO_CHAN_INFO_AVERAGE_RAW)),
+};
+
static struct twl4030_madc_data *twl4030_madc;
struct twl4030_prescale_divider_ratios {
@@ -702,28 +768,49 @@ static int twl4030_madc_probe(struct platform_device *pdev)
{
struct twl4030_madc_data *madc;
struct twl4030_madc_platform_data *pdata = dev_get_platdata(&pdev->dev);
+ struct device_node *np = pdev->dev.of_node;
int irq, ret;
u8 regval;
+ struct iio_dev *iio_dev = NULL;
- if (!pdata) {
+ if (!pdata && !np) {
dev_err(&pdev->dev, "platform_data not available\n");
return -EINVAL;
}
- madc = devm_kzalloc(&pdev->dev, sizeof(*madc), GFP_KERNEL);
- if (!madc)
+
+ iio_dev = devm_iio_device_alloc(&pdev->dev,
+ sizeof(struct twl4030_madc_data));
+ if (!iio_dev) {
+ dev_err(&pdev->dev, "failed allocating iio device\n");
return -ENOMEM;
+ }
+ madc = iio_priv(iio_dev);
madc->dev = &pdev->dev;
+ iio_dev->name = dev_name(&pdev->dev);
+ iio_dev->dev.parent = &pdev->dev;
+ iio_dev->dev.of_node = pdev->dev.of_node;
+ iio_dev->info = &twl4030_madc_iio_info;
+ iio_dev->modes = INDIO_DIRECT_MODE;
+ iio_dev->channels = twl4030_madc_iio_channels;
+ iio_dev->num_channels = 16;
+
/*
* Phoenix provides 2 interrupt lines. The first one is connected to
* the OMAP. The other one can be connected to the other processor such
* as modem. Hence two separate ISR and IMR registers.
*/
- madc->imr = (pdata->irq_line == 1) ?
+ if (pdata)
+ madc->use_second_irq = pdata->irq_line != 1;
+ else
+ madc->use_second_irq = false;
+
+ madc->imr = (madc->use_second_irq == 1) ?
TWL4030_MADC_IMR1 : TWL4030_MADC_IMR2;
- madc->isr = (pdata->irq_line == 1) ?
+ madc->isr = (madc->use_second_irq == 1) ?
TWL4030_MADC_ISR1 : TWL4030_MADC_ISR2;
+
ret = twl4030_madc_set_power(madc, 1);
if (ret < 0)
return ret;
@@ -768,7 +855,7 @@ static int twl4030_madc_probe(struct platform_device *pdev)
}
}
- platform_set_drvdata(pdev, madc);
+ platform_set_drvdata(pdev, iio_dev);
mutex_init(&madc->lock);
irq = platform_get_irq(pdev, 0);
@@ -780,7 +867,15 @@ static int twl4030_madc_probe(struct platform_device *pdev)
goto err_i2c;
}
twl4030_madc = madc;
+
+ ret = iio_device_register(iio_dev);
+ if (ret) {
+ dev_dbg(&pdev->dev, "could not register iio device\n");
+ goto err_i2c;
+ }
+
return 0;
+
err_i2c:
twl4030_madc_set_current_generator(madc, 0, 0);
err_current_generator:
@@ -790,20 +885,32 @@ err_current_generator:
static int twl4030_madc_remove(struct platform_device *pdev)
{
- struct twl4030_madc_data *madc = platform_get_drvdata(pdev);
+ struct iio_dev *iio_dev = platform_get_drvdata(pdev);
+ struct twl4030_madc_data *madc = iio_priv(iio_dev);
twl4030_madc_set_current_generator(madc, 0, 0);
twl4030_madc_set_power(madc, 0);
+ iio_device_unregister(iio_dev);
+
return 0;
}
+#ifdef CONFIG_OF
+static const struct of_device_id twl_madc_of_match[] = {
+ {.compatible = "ti,twl4030-madc", },
+ { },
+};
+MODULE_DEVICE_TABLE(of, twl_madc_of_match);
+#endif
+
static struct platform_driver twl4030_madc_driver = {
.probe = twl4030_madc_probe,
.remove = twl4030_madc_remove,
.driver = {
.name = "twl4030_madc",
.owner = THIS_MODULE,
+ .of_match_table = of_match_ptr(twl_madc_of_match),
},
};
--
1.8.5.3
^ permalink raw reply related
* [RFCv1 1/4] mfd: twl4030-madc: Use managed resources
From: Sebastian Reichel @ 2014-02-14 18:46 UTC (permalink / raw)
To: Sebastian Reichel, Marek Belisko
Cc: Jonathan Cameron, Lee Jones, Samuel Ortiz, Lars-Peter Clausen,
Rob Herring, Pawel Moll, Mark Rutland, Ian Campbell, Kumar Gala,
Grant Likely, linux-kernel, devicetree, linux-iio,
Sebastian Reichel
In-Reply-To: <20140214174040.GB26969@earth.universe>
Update twl4030-madc driver to use managed resources.
Signed-off-by: Sebastian Reichel <sre@debian.org>
---
drivers/mfd/twl4030-madc.c | 15 ++++++---------
1 file changed, 6 insertions(+), 9 deletions(-)
diff --git a/drivers/mfd/twl4030-madc.c b/drivers/mfd/twl4030-madc.c
index 4c583e4..5458561 100644
--- a/drivers/mfd/twl4030-madc.c
+++ b/drivers/mfd/twl4030-madc.c
@@ -702,14 +702,14 @@ static int twl4030_madc_probe(struct platform_device *pdev)
{
struct twl4030_madc_data *madc;
struct twl4030_madc_platform_data *pdata = dev_get_platdata(&pdev->dev);
- int ret;
+ int irq, ret;
u8 regval;
if (!pdata) {
dev_err(&pdev->dev, "platform_data not available\n");
return -EINVAL;
}
- madc = kzalloc(sizeof(*madc), GFP_KERNEL);
+ madc = devm_kzalloc(&pdev->dev, sizeof(*madc), GFP_KERNEL);
if (!madc)
return -ENOMEM;
@@ -726,7 +726,7 @@ static int twl4030_madc_probe(struct platform_device *pdev)
TWL4030_MADC_ISR1 : TWL4030_MADC_ISR2;
ret = twl4030_madc_set_power(madc, 1);
if (ret < 0)
- goto err_power;
+ return ret;
ret = twl4030_madc_set_current_generator(madc, 0, 1);
if (ret < 0)
goto err_current_generator;
@@ -770,7 +770,9 @@ static int twl4030_madc_probe(struct platform_device *pdev)
platform_set_drvdata(pdev, madc);
mutex_init(&madc->lock);
- ret = request_threaded_irq(platform_get_irq(pdev, 0), NULL,
+
+ irq = platform_get_irq(pdev, 0);
+ ret = devm_request_threaded_irq(&pdev->dev, irq, NULL,
twl4030_madc_threaded_irq_handler,
IRQF_TRIGGER_RISING, "twl4030_madc", madc);
if (ret) {
@@ -783,9 +785,6 @@ err_i2c:
twl4030_madc_set_current_generator(madc, 0, 0);
err_current_generator:
twl4030_madc_set_power(madc, 0);
-err_power:
- kfree(madc);
-
return ret;
}
@@ -793,10 +792,8 @@ static int twl4030_madc_remove(struct platform_device *pdev)
{
struct twl4030_madc_data *madc = platform_get_drvdata(pdev);
- free_irq(platform_get_irq(pdev, 0), madc);
twl4030_madc_set_current_generator(madc, 0, 0);
twl4030_madc_set_power(madc, 0);
- kfree(madc);
return 0;
}
--
1.8.5.3
^ permalink raw reply related
* Re: [PATCH v3 0/3] dp83640: Get pin and master/slave configuration from DT
From: David Miller @ 2014-02-14 18:36 UTC (permalink / raw)
To: richardcochran-Re5JQEeQqe8AvxtiuMwx3w
Cc: stefan.sorensen-usnHOLptxrsHrNJx0XZkJA,
grant.likely-QSEj5FYQhm4dnm+yROfE0A,
robh+dt-DgEjT+Ai2ygdnm+yROfE0A, mark.rutland-5wv7dgnIgG8,
devicetree-u79uwXL29TY76Z2rM5mHXA,
linux-kernel-u79uwXL29TY76Z2rM5mHXA,
netdev-u79uwXL29TY76Z2rM5mHXA
In-Reply-To: <20140214090601.GA3906@netboy>
From: Richard Cochran <richardcochran-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
Date: Fri, 14 Feb 2014 10:06:01 +0100
> People want to be able to configure the auxiliary functions on the
> pins of their PTP devices. My preference for supporting this is:
>
> 1. additional ioctl on the PTP character device
> 2. ethtool ioctl
> 3. DT and/or ACPI
>
> The first option seems best because that is how you activate the
> auxiliary functions. I am working on something in this direction.
Probably since PTP already uses ioctls, and that is the activation
point, yes that seems the best approach.
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply
* [PATCH v2 3/3] ARM: dts: bcm281xx: define real clocks
From: Alex Elder @ 2014-02-14 18:29 UTC (permalink / raw)
To: mturquette, mporter, bcm
Cc: Alex Elder, Rob Herring, Pawel Moll, Mark Rutland, Ian Campbell,
Kumar Gala, Russell King, bcm-kernel-feedback-list, devicetree,
linux-arm-kernel, linux-kernel
In-Reply-To: <1392402560-12929-1-git-send-email-elder@linaro.org>
From: Alex Elder <alex.elder@linaro.org>
Replace the "fake" clocks defined in the "bcm11351.dtsi" device tree
file with real definitions backed by the new BCM281xx clock driver.
Signed-off-by: Alex Elder <elder@linaro.org>
Reviewed-by: Matt Porter <matt.porter@linaro.org>
Reviewed-by: Tim Kryger <tim.kryger@linaro.org>
---
arch/arm/boot/dts/bcm11351.dtsi | 192 ++++++++++++++++++++++++++++-----------
1 file changed, 140 insertions(+), 52 deletions(-)
diff --git a/arch/arm/boot/dts/bcm11351.dtsi b/arch/arm/boot/dts/bcm11351.dtsi
index e491b82..94b36f6 100644
--- a/arch/arm/boot/dts/bcm11351.dtsi
+++ b/arch/arm/boot/dts/bcm11351.dtsi
@@ -14,6 +14,8 @@
#include <dt-bindings/interrupt-controller/arm-gic.h>
#include <dt-bindings/interrupt-controller/irq.h>
+#include "dt-bindings/clock/bcm281xx.h"
+
#include "skeleton.dtsi"
/ {
@@ -43,7 +45,7 @@
compatible = "brcm,bcm11351-dw-apb-uart", "snps,dw-apb-uart";
status = "disabled";
reg = <0x3e000000 0x1000>;
- clocks = <&uartb_clk>;
+ clocks = <&slave_ccu BCM281XX_SLAVE_CCU_UARTB>;
interrupts = <GIC_SPI 67 IRQ_TYPE_LEVEL_HIGH>;
reg-shift = <2>;
reg-io-width = <4>;
@@ -53,7 +55,7 @@
compatible = "brcm,bcm11351-dw-apb-uart", "snps,dw-apb-uart";
status = "disabled";
reg = <0x3e001000 0x1000>;
- clocks = <&uartb2_clk>;
+ clocks = <&slave_ccu BCM281XX_SLAVE_CCU_UARTB2>;
interrupts = <GIC_SPI 66 IRQ_TYPE_LEVEL_HIGH>;
reg-shift = <2>;
reg-io-width = <4>;
@@ -63,7 +65,7 @@
compatible = "brcm,bcm11351-dw-apb-uart", "snps,dw-apb-uart";
status = "disabled";
reg = <0x3e002000 0x1000>;
- clocks = <&uartb3_clk>;
+ clocks = <&slave_ccu BCM281XX_SLAVE_CCU_UARTB3>;
interrupts = <GIC_SPI 65 IRQ_TYPE_LEVEL_HIGH>;
reg-shift = <2>;
reg-io-width = <4>;
@@ -73,7 +75,7 @@
compatible = "brcm,bcm11351-dw-apb-uart", "snps,dw-apb-uart";
status = "disabled";
reg = <0x3e003000 0x1000>;
- clocks = <&uartb4_clk>;
+ clocks = <&slave_ccu BCM281XX_SLAVE_CCU_UARTB4>;
interrupts = <GIC_SPI 64 IRQ_TYPE_LEVEL_HIGH>;
reg-shift = <2>;
reg-io-width = <4>;
@@ -95,7 +97,7 @@
compatible = "brcm,kona-timer";
reg = <0x35006000 0x1000>;
interrupts = <GIC_SPI 7 IRQ_TYPE_LEVEL_HIGH>;
- clocks = <&hub_timer_clk>;
+ clocks = <&aon_ccu BCM281XX_AON_CCU_HUB_TIMER>;
};
gpio: gpio@35003000 {
@@ -118,7 +120,7 @@
compatible = "brcm,kona-sdhci";
reg = <0x3f180000 0x10000>;
interrupts = <GIC_SPI 77 IRQ_TYPE_LEVEL_HIGH>;
- clocks = <&sdio1_clk>;
+ clocks = <&master_ccu BCM281XX_MASTER_CCU_SDIO1>;
status = "disabled";
};
@@ -126,7 +128,7 @@
compatible = "brcm,kona-sdhci";
reg = <0x3f190000 0x10000>;
interrupts = <GIC_SPI 76 IRQ_TYPE_LEVEL_HIGH>;
- clocks = <&sdio2_clk>;
+ clocks = <&master_ccu BCM281XX_MASTER_CCU_SDIO2>;
status = "disabled";
};
@@ -134,7 +136,7 @@
compatible = "brcm,kona-sdhci";
reg = <0x3f1a0000 0x10000>;
interrupts = <GIC_SPI 74 IRQ_TYPE_LEVEL_HIGH>;
- clocks = <&sdio3_clk>;
+ clocks = <&master_ccu BCM281XX_MASTER_CCU_SDIO3>;
status = "disabled";
};
@@ -142,7 +144,7 @@
compatible = "brcm,kona-sdhci";
reg = <0x3f1b0000 0x10000>;
interrupts = <GIC_SPI 73 IRQ_TYPE_LEVEL_HIGH>;
- clocks = <&sdio4_clk>;
+ clocks = <&master_ccu BCM281XX_MASTER_CCU_SDIO4>;
status = "disabled";
};
@@ -157,7 +159,7 @@
interrupts = <GIC_SPI 103 IRQ_TYPE_LEVEL_HIGH>;
#address-cells = <1>;
#size-cells = <0>;
- clocks = <&bsc1_clk>;
+ clocks = <&slave_ccu BCM281XX_SLAVE_CCU_BSC1>;
status = "disabled";
};
@@ -167,7 +169,7 @@
interrupts = <GIC_SPI 102 IRQ_TYPE_LEVEL_HIGH>;
#address-cells = <1>;
#size-cells = <0>;
- clocks = <&bsc2_clk>;
+ clocks = <&slave_ccu BCM281XX_SLAVE_CCU_BSC2>;
status = "disabled";
};
@@ -177,7 +179,7 @@
interrupts = <GIC_SPI 169 IRQ_TYPE_LEVEL_HIGH>;
#address-cells = <1>;
#size-cells = <0>;
- clocks = <&bsc3_clk>;
+ clocks = <&slave_ccu BCM281XX_SLAVE_CCU_BSC3>;
status = "disabled";
};
@@ -187,105 +189,191 @@
interrupts = <GIC_SPI 14 IRQ_TYPE_LEVEL_HIGH>;
#address-cells = <1>;
#size-cells = <0>;
- clocks = <&pmu_bsc_clk>;
+ clocks = <&aon_ccu BCM281XX_AON_CCU_PMU_BSC>;
status = "disabled";
};
clocks {
- bsc1_clk: bsc1 {
- compatible = "fixed-clock";
- clock-frequency = <13000000>;
+ #address-cells = <1>;
+ #size-cells = <1>;
+ ranges;
+
+ root_ccu: root_ccu {
+ compatible = "brcm,bcm11351-root-ccu";
+ reg = <0x35001000 0x0f00>;
+ #clock-cells = <1>;
+ clock-output-names = "frac_1m";
+ };
+
+ hub_ccu: hub_ccu {
+ compatible = "brcm,bcm11351-hub-ccu";
+ reg = <0x34000000 0x0f00>;
+ #clock-cells = <1>;
+ clock-output-names = "tmon_1m";
+ };
+
+ aon_ccu: aon_ccu {
+ compatible = "brcm,bcm11351-aon-ccu";
+ reg = <0x35002000 0x0f00>;
+ #clock-cells = <1>;
+ clock-output-names = "hub_timer",
+ "pmu_bsc",
+ "pmu_bsc_var";
+ };
+
+ master_ccu: master_ccu {
+ compatible = "brcm,bcm11351-master-ccu";
+ reg = <0x3f001000 0x0f00>;
+ #clock-cells = <1>;
+ clock-output-names = "sdio1",
+ "sdio2",
+ "sdio3",
+ "sdio4",
+ "usb_ic",
+ "hsic2_48m",
+ "hsic2_12m";
+ };
+
+ slave_ccu: slave_ccu {
+ compatible = "brcm,bcm11351-slave-ccu";
+ reg = <0x3e011000 0x0f00>;
+ #clock-cells = <1>;
+ clock-output-names = "uartb",
+ "uartb2",
+ "uartb3",
+ "uartb4",
+ "ssp0",
+ "ssp2",
+ "bsc1",
+ "bsc2",
+ "bsc3",
+ "pwm";
+ };
+
+ ref_1m_clk: ref_1m {
#clock-cells = <0>;
+ compatible = "fixed-clock";
+ clock-frequency = <1000000>;
};
- bsc2_clk: bsc2 {
+ ref_32k_clk: ref_32k {
+ #clock-cells = <0>;
compatible = "fixed-clock";
- clock-frequency = <13000000>;
+ clock-frequency = <32768>;
+ };
+
+ bbl_32k_clk: bbl_32k {
#clock-cells = <0>;
+ compatible = "fixed-clock";
+ clock-frequency = <32768>;
};
- bsc3_clk: bsc3 {
+ ref_13m_clk: ref_13m {
+ #clock-cells = <0>;
compatible = "fixed-clock";
clock-frequency = <13000000>;
- #clock-cells = <0>;
};
- pmu_bsc_clk: pmu_bsc {
+ var_13m_clk: var_13m {
+ #clock-cells = <0>;
compatible = "fixed-clock";
clock-frequency = <13000000>;
- #clock-cells = <0>;
};
- hub_timer_clk: hub_timer {
- compatible = "fixed-clock";
- clock-frequency = <32768>;
+ dft_19_5m_clk: dft_19_5m {
#clock-cells = <0>;
+ compatible = "fixed-clock";
+ clock-frequency = <19500000>;
};
- pwm_clk: pwm {
+ ref_crystal_clk: ref_crystal {
+ #clock-cells = <0>;
compatible = "fixed-clock";
clock-frequency = <26000000>;
- #clock-cells = <0>;
};
- sdio1_clk: sdio1 {
- compatible = "fixed-clock";
- clock-frequency = <48000000>;
+ ref_cx40_clk: ref_cx40 {
#clock-cells = <0>;
+ compatible = "fixed-clock";
+ clock-frequency = <40000000>;
};
- sdio2_clk: sdio2 {
- compatible = "fixed-clock";
- clock-frequency = <48000000>;
+ ref_52m_clk: ref_52m {
#clock-cells = <0>;
+ compatible = "fixed-clock";
+ clock-frequency = <52000000>;
};
- sdio3_clk: sdio3 {
- compatible = "fixed-clock";
- clock-frequency = <48000000>;
+ var_52m_clk: var_52m {
#clock-cells = <0>;
+ compatible = "fixed-clock";
+ clock-frequency = <52000000>;
};
- sdio4_clk: sdio4 {
+ usb_otg_ahb_clk: usb_otg_ahb {
compatible = "fixed-clock";
- clock-frequency = <48000000>;
+ clock-frequency = <52000000>;
#clock-cells = <0>;
};
- tmon_1m_clk: tmon_1m {
- compatible = "fixed-clock";
- clock-frequency = <1000000>;
+ ref_96m_clk: ref_96m {
#clock-cells = <0>;
+ compatible = "fixed-clock";
+ clock-frequency = <96000000>;
};
- uartb_clk: uartb {
- compatible = "fixed-clock";
- clock-frequency = <13000000>;
+ var_96m_clk: var_96m {
#clock-cells = <0>;
+ compatible = "fixed-clock";
+ clock-frequency = <96000000>;
};
- uartb2_clk: uartb2 {
+ ref_104m_clk: ref_104m {
+ #clock-cells = <0>;
compatible = "fixed-clock";
- clock-frequency = <13000000>;
+ clock-frequency = <104000000>;
+ };
+
+ var_104m_clk: var_104m {
#clock-cells = <0>;
+ compatible = "fixed-clock";
+ clock-frequency = <104000000>;
};
- uartb3_clk: uartb3 {
+ ref_156m_clk: ref_156m {
+ #clock-cells = <0>;
compatible = "fixed-clock";
- clock-frequency = <13000000>;
+ clock-frequency = <156000000>;
+ };
+
+ var_156m_clk: var_156m {
#clock-cells = <0>;
+ compatible = "fixed-clock";
+ clock-frequency = <156000000>;
};
- uartb4_clk: uartb4 {
+ ref_208m_clk: ref_208m {
+ #clock-cells = <0>;
compatible = "fixed-clock";
- clock-frequency = <13000000>;
+ clock-frequency = <208000000>;
+ };
+
+ var_208m_clk: var_208m {
#clock-cells = <0>;
+ compatible = "fixed-clock";
+ clock-frequency = <208000000>;
};
- usb_otg_ahb_clk: usb_otg_ahb {
+ ref_312m_clk: ref_312m {
+ #clock-cells = <0>;
compatible = "fixed-clock";
- clock-frequency = <52000000>;
+ clock-frequency = <312000000>;
+ };
+
+ var_312m_clk: var_312m {
#clock-cells = <0>;
+ compatible = "fixed-clock";
+ clock-frequency = <312000000>;
};
};
--
1.7.9.5
^ permalink raw reply related
* [PATCH v2 1/3] clk: bcm281xx: add initial clock framework support
From: Alex Elder @ 2014-02-14 18:29 UTC (permalink / raw)
To: mturquette-QSEj5FYQhm4dnm+yROfE0A, mporter-QSEj5FYQhm4dnm+yROfE0A,
bcm-xK7y4jjYLqYh9ZMKESR00Q
Cc: Alex Elder, Rob Herring, Pawel Moll, Mark Rutland, Ian Campbell,
Kumar Gala, Tim Kryger, Matt Porter,
linux-kernel-u79uwXL29TY76Z2rM5mHXA,
linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r,
devicetree-u79uwXL29TY76Z2rM5mHXA
In-Reply-To: <1392402560-12929-1-git-send-email-elder-QSEj5FYQhm4dnm+yROfE0A@public.gmane.org>
Add code for device tree support of clocks in the BCM281xx family of
SoCs. Machines in this family use peripheral clocks implemented by
"Kona" clock control units (CCUs). (Other Broadcom SoC families use
Kona style CCUs as well, but support for them is not yet upstream.)
A BCM281xx SoC has multiple CCUs, each of which manages a set of
clocks on the SoC. A Kona peripheral clock is composite clock that
may include a gate, a parent clock multiplexor, and zero, one
or two dividers. There is a variety of gate types, and many gates
implement hardware-managed gating (often called "auto-gating").
Most dividers divide their input clock signal by an integer value
(one or more). There are also "fractional" dividers which allow
division by non-integer values. To accomodate such dividers,
clock rates and dividers are generally maintained by the code in
"scaled" form, which allows integer and fractional dividers to
be handled in a uniform way.
If present, the gate for a Kona peripheral clock must be enabled
when a change is made to its multiplexor or one of its dividers.
Additionally, dividers and multiplexors have trigger registers which
must be used whenever the divider value or selected parent clock is
changed. The same trigger is often used for a divider and
multiplexor, and a BCM281xx peripheral clock occasionally has two
triggers.
The gate, dividers, and parent clock selector are treated in this
code as "components" of a peripheral clock. Their functionality is
implemented directly--e.g. the common clock framework gate
implementation is not used for a Kona peripheral clock gate. (This
has being considered though, and the intention is to evolve this
code to leverage common code as much as possible.)
The source code is divided into three general portions:
drivers/clk/bcm/clk-kona.h
drivers/clk/bcm/clk-kona.c
These implement the basic Kona clock functionality,
including the clk_ops methods and various routines to
manipulate registers and interpret their values. This
includes some functions used to set clocks to a desired
initial state (though this feature is only partially
implemented here).
drivers/clk/bcm/clk-kona-setup.c
This contains generic run-time initialization code for
data structures representing Kona CCUs and clocks. This
encapsulates the clock structure initialization that can't
be done statically. Note that there is a great deal of
validity-checking code here, making explicit certain
assumptions in the code. This is mostly useful for adding
new clock definitions and could possibly be disabled for
production use.
drivers/clk/bcm/clk-bcm281xx.c
This file defines the specific CCUs used by BCM281XX family
SoCs, as well as the specific clocks implemented by each.
It declares a device tree clock match entry for each CCU
defined.
include/dt-bindings/clock/bcm281xx.h
This file defines the selector (index) values used to
identify a particular clock provided by a CCU. It consists
entirely of C preprocessor constants, to be used by both the
C source and device tree source files.
Signed-off-by: Alex Elder <elder-QSEj5FYQhm4dnm+yROfE0A@public.gmane.org>
Reviewed-by: Matt Porter <matt.porter-QSEj5FYQhm4dnm+yROfE0A@public.gmane.org>
Reviewed-by: Tim Kryger <tim.kryger-QSEj5FYQhm4dnm+yROfE0A@public.gmane.org>
Acked-by: Mike Turquette <mturquette-QSEj5FYQhm4dnm+yROfE0A@public.gmane.org>
---
drivers/clk/Kconfig | 1 +
drivers/clk/Makefile | 1 +
drivers/clk/bcm/Kconfig | 9 +
drivers/clk/bcm/Makefile | 3 +
drivers/clk/bcm/clk-bcm281xx.c | 416 ++++++++++++++
drivers/clk/bcm/clk-kona-setup.c | 769 +++++++++++++++++++++++++
drivers/clk/bcm/clk-kona.c | 1033 ++++++++++++++++++++++++++++++++++
drivers/clk/bcm/clk-kona.h | 410 ++++++++++++++
include/dt-bindings/clock/bcm281xx.h | 65 +++
9 files changed, 2707 insertions(+)
create mode 100644 drivers/clk/bcm/Kconfig
create mode 100644 drivers/clk/bcm/Makefile
create mode 100644 drivers/clk/bcm/clk-bcm281xx.c
create mode 100644 drivers/clk/bcm/clk-kona-setup.c
create mode 100644 drivers/clk/bcm/clk-kona.c
create mode 100644 drivers/clk/bcm/clk-kona.h
create mode 100644 include/dt-bindings/clock/bcm281xx.h
diff --git a/drivers/clk/Kconfig b/drivers/clk/Kconfig
index 7641965..f9f6056 100644
--- a/drivers/clk/Kconfig
+++ b/drivers/clk/Kconfig
@@ -111,4 +111,5 @@ source "drivers/clk/qcom/Kconfig"
endmenu
+source "drivers/clk/bcm/Kconfig"
source "drivers/clk/mvebu/Kconfig"
diff --git a/drivers/clk/Makefile b/drivers/clk/Makefile
index a367a98..88af4a3 100644
--- a/drivers/clk/Makefile
+++ b/drivers/clk/Makefile
@@ -29,6 +29,7 @@ obj-$(CONFIG_ARCH_VT8500) += clk-vt8500.o
obj-$(CONFIG_COMMON_CLK_WM831X) += clk-wm831x.o
obj-$(CONFIG_COMMON_CLK_XGENE) += clk-xgene.o
obj-$(CONFIG_COMMON_CLK_AT91) += at91/
+obj-$(CONFIG_ARCH_BCM_MOBILE) += bcm/
obj-$(CONFIG_ARCH_HI3xxx) += hisilicon/
obj-$(CONFIG_COMMON_CLK_KEYSTONE) += keystone/
ifeq ($(CONFIG_COMMON_CLK), y)
diff --git a/drivers/clk/bcm/Kconfig b/drivers/clk/bcm/Kconfig
new file mode 100644
index 0000000..a7262fb
--- /dev/null
+++ b/drivers/clk/bcm/Kconfig
@@ -0,0 +1,9 @@
+config CLK_BCM_KONA
+ bool "Broadcom Kona CCU clock support"
+ depends on ARCH_BCM_MOBILE
+ depends on COMMON_CLK
+ default y
+ help
+ Enable common clock framework support for Broadcom SoCs
+ using "Kona" style clock control units, including those
+ in the BCM281xx family.
diff --git a/drivers/clk/bcm/Makefile b/drivers/clk/bcm/Makefile
new file mode 100644
index 0000000..cf93359
--- /dev/null
+++ b/drivers/clk/bcm/Makefile
@@ -0,0 +1,3 @@
+obj-$(CONFIG_CLK_BCM_KONA) += clk-kona.o
+obj-$(CONFIG_CLK_BCM_KONA) += clk-kona-setup.o
+obj-$(CONFIG_CLK_BCM_KONA) += clk-bcm281xx.o
diff --git a/drivers/clk/bcm/clk-bcm281xx.c b/drivers/clk/bcm/clk-bcm281xx.c
new file mode 100644
index 0000000..3c66de6
--- /dev/null
+++ b/drivers/clk/bcm/clk-bcm281xx.c
@@ -0,0 +1,416 @@
+/*
+ * Copyright (C) 2013 Broadcom Corporation
+ * Copyright 2013 Linaro Limited
+ *
+ * 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 version 2.
+ *
+ * This program is distributed "as is" WITHOUT ANY WARRANTY of any
+ * kind, whether express or implied; without even the implied warranty
+ * of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ */
+
+#include "clk-kona.h"
+#include "dt-bindings/clock/bcm281xx.h"
+
+/* bcm11351 CCU device tree "compatible" strings */
+#define BCM11351_DT_ROOT_CCU_COMPAT "brcm,bcm11351-root-ccu"
+#define BCM11351_DT_AON_CCU_COMPAT "brcm,bcm11351-aon-ccu"
+#define BCM11351_DT_HUB_CCU_COMPAT "brcm,bcm11351-hub-ccu"
+#define BCM11351_DT_MASTER_CCU_COMPAT "brcm,bcm11351-master-ccu"
+#define BCM11351_DT_SLAVE_CCU_COMPAT "brcm,bcm11351-slave-ccu"
+
+/* Root CCU clocks */
+
+static struct peri_clk_data frac_1m_data = {
+ .gate = HW_SW_GATE(0x214, 16, 0, 1),
+ .trig = TRIGGER(0x0e04, 0),
+ .div = FRAC_DIVIDER(0x0e00, 0, 22, 16),
+ .clocks = CLOCKS("ref_crystal"),
+};
+
+/* AON CCU clocks */
+
+static struct peri_clk_data hub_timer_data = {
+ .gate = HW_SW_GATE(0x0414, 16, 0, 1),
+ .clocks = CLOCKS("bbl_32k",
+ "frac_1m",
+ "dft_19_5m"),
+ .sel = SELECTOR(0x0a10, 0, 2),
+ .trig = TRIGGER(0x0a40, 4),
+};
+
+static struct peri_clk_data pmu_bsc_data = {
+ .gate = HW_SW_GATE(0x0418, 16, 0, 1),
+ .clocks = CLOCKS("ref_crystal",
+ "pmu_bsc_var",
+ "bbl_32k"),
+ .sel = SELECTOR(0x0a04, 0, 2),
+ .div = DIVIDER(0x0a04, 3, 4),
+ .trig = TRIGGER(0x0a40, 0),
+};
+
+static struct peri_clk_data pmu_bsc_var_data = {
+ .clocks = CLOCKS("var_312m",
+ "ref_312m"),
+ .sel = SELECTOR(0x0a00, 0, 2),
+ .div = DIVIDER(0x0a00, 4, 5),
+ .trig = TRIGGER(0x0a40, 2),
+};
+
+/* Hub CCU clocks */
+
+static struct peri_clk_data tmon_1m_data = {
+ .gate = HW_SW_GATE(0x04a4, 18, 2, 3),
+ .clocks = CLOCKS("ref_crystal",
+ "frac_1m"),
+ .sel = SELECTOR(0x0e74, 0, 2),
+ .trig = TRIGGER(0x0e84, 1),
+};
+
+/* Master CCU clocks */
+
+static struct peri_clk_data sdio1_data = {
+ .gate = HW_SW_GATE(0x0358, 18, 2, 3),
+ .clocks = CLOCKS("ref_crystal",
+ "var_52m",
+ "ref_52m",
+ "var_96m",
+ "ref_96m"),
+ .sel = SELECTOR(0x0a28, 0, 3),
+ .div = DIVIDER(0x0a28, 4, 14),
+ .trig = TRIGGER(0x0afc, 9),
+};
+
+static struct peri_clk_data sdio2_data = {
+ .gate = HW_SW_GATE(0x035c, 18, 2, 3),
+ .clocks = CLOCKS("ref_crystal",
+ "var_52m",
+ "ref_52m",
+ "var_96m",
+ "ref_96m"),
+ .sel = SELECTOR(0x0a2c, 0, 3),
+ .div = DIVIDER(0x0a2c, 4, 14),
+ .trig = TRIGGER(0x0afc, 10),
+};
+
+static struct peri_clk_data sdio3_data = {
+ .gate = HW_SW_GATE(0x0364, 18, 2, 3),
+ .clocks = CLOCKS("ref_crystal",
+ "var_52m",
+ "ref_52m",
+ "var_96m",
+ "ref_96m"),
+ .sel = SELECTOR(0x0a34, 0, 3),
+ .div = DIVIDER(0x0a34, 4, 14),
+ .trig = TRIGGER(0x0afc, 12),
+};
+
+static struct peri_clk_data sdio4_data = {
+ .gate = HW_SW_GATE(0x0360, 18, 2, 3),
+ .clocks = CLOCKS("ref_crystal",
+ "var_52m",
+ "ref_52m",
+ "var_96m",
+ "ref_96m"),
+ .sel = SELECTOR(0x0a30, 0, 3),
+ .div = DIVIDER(0x0a30, 4, 14),
+ .trig = TRIGGER(0x0afc, 11),
+};
+
+static struct peri_clk_data usb_ic_data = {
+ .gate = HW_SW_GATE(0x0354, 18, 2, 3),
+ .clocks = CLOCKS("ref_crystal",
+ "var_96m",
+ "ref_96m"),
+ .div = FIXED_DIVIDER(2),
+ .sel = SELECTOR(0x0a24, 0, 2),
+ .trig = TRIGGER(0x0afc, 7),
+};
+
+/* also called usbh_48m */
+static struct peri_clk_data hsic2_48m_data = {
+ .gate = HW_SW_GATE(0x0370, 18, 2, 3),
+ .clocks = CLOCKS("ref_crystal",
+ "var_96m",
+ "ref_96m"),
+ .sel = SELECTOR(0x0a38, 0, 2),
+ .div = FIXED_DIVIDER(2),
+ .trig = TRIGGER(0x0afc, 5),
+};
+
+/* also called usbh_12m */
+static struct peri_clk_data hsic2_12m_data = {
+ .gate = HW_SW_GATE(0x0370, 20, 4, 5),
+ .div = DIVIDER(0x0a38, 12, 2),
+ .clocks = CLOCKS("ref_crystal",
+ "var_96m",
+ "ref_96m"),
+ .pre_div = FIXED_DIVIDER(2),
+ .sel = SELECTOR(0x0a38, 0, 2),
+ .trig = TRIGGER(0x0afc, 5),
+};
+
+/* Slave CCU clocks */
+
+static struct peri_clk_data uartb_data = {
+ .gate = HW_SW_GATE(0x0400, 18, 2, 3),
+ .clocks = CLOCKS("ref_crystal",
+ "var_156m",
+ "ref_156m"),
+ .sel = SELECTOR(0x0a10, 0, 2),
+ .div = FRAC_DIVIDER(0x0a10, 4, 12, 8),
+ .trig = TRIGGER(0x0afc, 2),
+};
+
+static struct peri_clk_data uartb2_data = {
+ .gate = HW_SW_GATE(0x0404, 18, 2, 3),
+ .clocks = CLOCKS("ref_crystal",
+ "var_156m",
+ "ref_156m"),
+ .sel = SELECTOR(0x0a14, 0, 2),
+ .div = FRAC_DIVIDER(0x0a14, 4, 12, 8),
+ .trig = TRIGGER(0x0afc, 3),
+};
+
+static struct peri_clk_data uartb3_data = {
+ .gate = HW_SW_GATE(0x0408, 18, 2, 3),
+ .clocks = CLOCKS("ref_crystal",
+ "var_156m",
+ "ref_156m"),
+ .sel = SELECTOR(0x0a18, 0, 2),
+ .div = FRAC_DIVIDER(0x0a18, 4, 12, 8),
+ .trig = TRIGGER(0x0afc, 4),
+};
+
+static struct peri_clk_data uartb4_data = {
+ .gate = HW_SW_GATE(0x0408, 18, 2, 3),
+ .clocks = CLOCKS("ref_crystal",
+ "var_156m",
+ "ref_156m"),
+ .sel = SELECTOR(0x0a1c, 0, 2),
+ .div = FRAC_DIVIDER(0x0a1c, 4, 12, 8),
+ .trig = TRIGGER(0x0afc, 5),
+};
+
+static struct peri_clk_data ssp0_data = {
+ .gate = HW_SW_GATE(0x0410, 18, 2, 3),
+ .clocks = CLOCKS("ref_crystal",
+ "var_104m",
+ "ref_104m",
+ "var_96m",
+ "ref_96m"),
+ .sel = SELECTOR(0x0a20, 0, 3),
+ .div = DIVIDER(0x0a20, 4, 14),
+ .trig = TRIGGER(0x0afc, 6),
+};
+
+static struct peri_clk_data ssp2_data = {
+ .gate = HW_SW_GATE(0x0418, 18, 2, 3),
+ .clocks = CLOCKS("ref_crystal",
+ "var_104m",
+ "ref_104m",
+ "var_96m",
+ "ref_96m"),
+ .sel = SELECTOR(0x0a28, 0, 3),
+ .div = DIVIDER(0x0a28, 4, 14),
+ .trig = TRIGGER(0x0afc, 8),
+};
+
+static struct peri_clk_data bsc1_data = {
+ .gate = HW_SW_GATE(0x0458, 18, 2, 3),
+ .clocks = CLOCKS("ref_crystal",
+ "var_104m",
+ "ref_104m",
+ "var_13m",
+ "ref_13m"),
+ .sel = SELECTOR(0x0a64, 0, 3),
+ .trig = TRIGGER(0x0afc, 23),
+};
+
+static struct peri_clk_data bsc2_data = {
+ .gate = HW_SW_GATE(0x045c, 18, 2, 3),
+ .clocks = CLOCKS("ref_crystal",
+ "var_104m",
+ "ref_104m",
+ "var_13m",
+ "ref_13m"),
+ .sel = SELECTOR(0x0a68, 0, 3),
+ .trig = TRIGGER(0x0afc, 24),
+};
+
+static struct peri_clk_data bsc3_data = {
+ .gate = HW_SW_GATE(0x0484, 18, 2, 3),
+ .clocks = CLOCKS("ref_crystal",
+ "var_104m",
+ "ref_104m",
+ "var_13m",
+ "ref_13m"),
+ .sel = SELECTOR(0x0a84, 0, 3),
+ .trig = TRIGGER(0x0b00, 2),
+};
+
+static struct peri_clk_data pwm_data = {
+ .gate = HW_SW_GATE(0x0468, 18, 2, 3),
+ .clocks = CLOCKS("ref_crystal",
+ "var_104m"),
+ .sel = SELECTOR(0x0a70, 0, 2),
+ .div = DIVIDER(0x0a70, 4, 3),
+ .trig = TRIGGER(0x0afc, 15),
+};
+
+/*
+ * CCU setup routines
+ *
+ * These are called from kona_dt_ccu_setup() to initialize the array
+ * of clocks provided by the CCU. Once allocated, the entries in
+ * the array are initialized by calling kona_clk_setup() with the
+ * initialization data for each clock. They return 0 if successful
+ * or an error code otherwise.
+ */
+static int __init bcm281xx_root_ccu_clks_setup(struct ccu_data *ccu)
+{
+ struct clk **clks;
+ size_t count = BCM281XX_ROOT_CCU_CLOCK_COUNT;
+
+ clks = kzalloc(count * sizeof(*clks), GFP_KERNEL);
+ if (!clks) {
+ pr_err("%s: failed to allocate root clocks\n", __func__);
+ return -ENOMEM;
+ }
+ ccu->data.clks = clks;
+ ccu->data.clk_num = count;
+
+ PERI_CLK_SETUP(clks, ccu, BCM281XX_ROOT_CCU_FRAC_1M, frac_1m);
+
+ return 0;
+}
+
+static int __init bcm281xx_aon_ccu_clks_setup(struct ccu_data *ccu)
+{
+ struct clk **clks;
+ size_t count = BCM281XX_AON_CCU_CLOCK_COUNT;
+
+ clks = kzalloc(count * sizeof(*clks), GFP_KERNEL);
+ if (!clks) {
+ pr_err("%s: failed to allocate aon clocks\n", __func__);
+ return -ENOMEM;
+ }
+ ccu->data.clks = clks;
+ ccu->data.clk_num = count;
+
+ PERI_CLK_SETUP(clks, ccu, BCM281XX_AON_CCU_HUB_TIMER, hub_timer);
+ PERI_CLK_SETUP(clks, ccu, BCM281XX_AON_CCU_PMU_BSC, pmu_bsc);
+ PERI_CLK_SETUP(clks, ccu, BCM281XX_AON_CCU_PMU_BSC_VAR, pmu_bsc_var);
+
+ return 0;
+}
+
+static int __init bcm281xx_hub_ccu_clks_setup(struct ccu_data *ccu)
+{
+ struct clk **clks;
+ size_t count = BCM281XX_HUB_CCU_CLOCK_COUNT;
+
+ clks = kzalloc(count * sizeof(*clks), GFP_KERNEL);
+ if (!clks) {
+ pr_err("%s: failed to allocate hub clocks\n", __func__);
+ return -ENOMEM;
+ }
+ ccu->data.clks = clks;
+ ccu->data.clk_num = count;
+
+ PERI_CLK_SETUP(clks, ccu, BCM281XX_HUB_CCU_TMON_1M, tmon_1m);
+
+ return 0;
+}
+
+static int __init bcm281xx_master_ccu_clks_setup(struct ccu_data *ccu)
+{
+ struct clk **clks;
+ size_t count = BCM281XX_MASTER_CCU_CLOCK_COUNT;
+
+ clks = kzalloc(count * sizeof(*clks), GFP_KERNEL);
+ if (!clks) {
+ pr_err("%s: failed to allocate master clocks\n", __func__);
+ return -ENOMEM;
+ }
+ ccu->data.clks = clks;
+ ccu->data.clk_num = count;
+
+ PERI_CLK_SETUP(clks, ccu, BCM281XX_MASTER_CCU_SDIO1, sdio1);
+ PERI_CLK_SETUP(clks, ccu, BCM281XX_MASTER_CCU_SDIO2, sdio2);
+ PERI_CLK_SETUP(clks, ccu, BCM281XX_MASTER_CCU_SDIO3, sdio3);
+ PERI_CLK_SETUP(clks, ccu, BCM281XX_MASTER_CCU_SDIO4, sdio4);
+ PERI_CLK_SETUP(clks, ccu, BCM281XX_MASTER_CCU_USB_IC, usb_ic);
+ PERI_CLK_SETUP(clks, ccu, BCM281XX_MASTER_CCU_HSIC2_48M, hsic2_48m);
+ PERI_CLK_SETUP(clks, ccu, BCM281XX_MASTER_CCU_HSIC2_12M, hsic2_12m);
+
+ return 0;
+}
+
+static int __init bcm281xx_slave_ccu_clks_setup(struct ccu_data *ccu)
+{
+ struct clk **clks;
+ size_t count = BCM281XX_SLAVE_CCU_CLOCK_COUNT;
+
+ clks = kzalloc(count * sizeof(*clks), GFP_KERNEL);
+ if (!clks) {
+ pr_err("%s: failed to allocate slave clocks\n", __func__);
+ return -ENOMEM;
+ }
+ ccu->data.clks = clks;
+ ccu->data.clk_num = count;
+
+ PERI_CLK_SETUP(clks, ccu, BCM281XX_SLAVE_CCU_UARTB, uartb);
+ PERI_CLK_SETUP(clks, ccu, BCM281XX_SLAVE_CCU_UARTB2, uartb2);
+ PERI_CLK_SETUP(clks, ccu, BCM281XX_SLAVE_CCU_UARTB3, uartb3);
+ PERI_CLK_SETUP(clks, ccu, BCM281XX_SLAVE_CCU_UARTB4, uartb4);
+ PERI_CLK_SETUP(clks, ccu, BCM281XX_SLAVE_CCU_SSP0, ssp0);
+ PERI_CLK_SETUP(clks, ccu, BCM281XX_SLAVE_CCU_SSP2, ssp2);
+ PERI_CLK_SETUP(clks, ccu, BCM281XX_SLAVE_CCU_BSC1, bsc1);
+ PERI_CLK_SETUP(clks, ccu, BCM281XX_SLAVE_CCU_BSC2, bsc2);
+ PERI_CLK_SETUP(clks, ccu, BCM281XX_SLAVE_CCU_BSC3, bsc3);
+ PERI_CLK_SETUP(clks, ccu, BCM281XX_SLAVE_CCU_PWM, pwm);
+
+ return 0;
+}
+
+/* Device tree match table callback functions */
+
+static void __init kona_dt_root_ccu_setup(struct device_node *node)
+{
+ kona_dt_ccu_setup(node, bcm281xx_root_ccu_clks_setup);
+}
+
+static void __init kona_dt_aon_ccu_setup(struct device_node *node)
+{
+ kona_dt_ccu_setup(node, bcm281xx_aon_ccu_clks_setup);
+}
+
+static void __init kona_dt_hub_ccu_setup(struct device_node *node)
+{
+ kona_dt_ccu_setup(node, bcm281xx_hub_ccu_clks_setup);
+}
+
+static void __init kona_dt_master_ccu_setup(struct device_node *node)
+{
+ kona_dt_ccu_setup(node, bcm281xx_master_ccu_clks_setup);
+}
+
+static void __init kona_dt_slave_ccu_setup(struct device_node *node)
+{
+ kona_dt_ccu_setup(node, bcm281xx_slave_ccu_clks_setup);
+}
+
+CLK_OF_DECLARE(bcm11351_root_ccu, BCM11351_DT_ROOT_CCU_COMPAT,
+ kona_dt_root_ccu_setup);
+CLK_OF_DECLARE(bcm11351_aon_ccu, BCM11351_DT_AON_CCU_COMPAT,
+ kona_dt_aon_ccu_setup);
+CLK_OF_DECLARE(bcm11351_hub_ccu, BCM11351_DT_HUB_CCU_COMPAT,
+ kona_dt_hub_ccu_setup);
+CLK_OF_DECLARE(bcm11351_master_ccu, BCM11351_DT_MASTER_CCU_COMPAT,
+ kona_dt_master_ccu_setup);
+CLK_OF_DECLARE(bcm11351_slave_ccu, BCM11351_DT_SLAVE_CCU_COMPAT,
+ kona_dt_slave_ccu_setup);
diff --git a/drivers/clk/bcm/clk-kona-setup.c b/drivers/clk/bcm/clk-kona-setup.c
new file mode 100644
index 0000000..f1e88fe
--- /dev/null
+++ b/drivers/clk/bcm/clk-kona-setup.c
@@ -0,0 +1,769 @@
+/*
+ * Copyright (C) 2013 Broadcom Corporation
+ * Copyright 2013 Linaro Limited
+ *
+ * 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 version 2.
+ *
+ * This program is distributed "as is" WITHOUT ANY WARRANTY of any
+ * kind, whether express or implied; without even the implied warranty
+ * of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ */
+
+#include <linux/io.h>
+#include <linux/of_address.h>
+
+#include "clk-kona.h"
+
+/* These are used when a selector or trigger is found to be unneeded */
+#define selector_clear_exists(sel) ((sel)->width = 0)
+#define trigger_clear_exists(trig) FLAG_CLEAR(trig, TRIG, EXISTS)
+
+LIST_HEAD(ccu_list); /* The list of set up CCUs */
+
+/* Validity checking */
+
+static bool clk_requires_trigger(struct kona_clk *bcm_clk)
+{
+ struct peri_clk_data *peri = bcm_clk->peri;
+ struct bcm_clk_sel *sel;
+ struct bcm_clk_div *div;
+
+ if (bcm_clk->type != bcm_clk_peri)
+ return false;
+
+ sel = &peri->sel;
+ if (sel->parent_count && selector_exists(sel))
+ return true;
+
+ div = &peri->div;
+ if (!divider_exists(div))
+ return false;
+
+ /* Fixed dividers don't need triggers */
+ if (!divider_is_fixed(div))
+ return true;
+
+ div = &peri->pre_div;
+
+ return divider_exists(div) && !divider_is_fixed(div);
+}
+
+static bool peri_clk_data_offsets_valid(struct kona_clk *bcm_clk)
+{
+ struct peri_clk_data *peri;
+ struct bcm_clk_gate *gate;
+ struct bcm_clk_div *div;
+ struct bcm_clk_sel *sel;
+ struct bcm_clk_trig *trig;
+ const char *name;
+ u32 range;
+ u32 limit;
+
+ BUG_ON(bcm_clk->type != bcm_clk_peri);
+ peri = bcm_clk->peri;
+ name = bcm_clk->name;
+ range = bcm_clk->ccu->range;
+
+ limit = range - sizeof(u32);
+ limit = round_down(limit, sizeof(u32));
+
+ gate = &peri->gate;
+ if (gate_exists(gate)) {
+ if (gate->offset > limit) {
+ pr_err("%s: bad gate offset for %s (%u > %u)\n",
+ __func__, name, gate->offset, limit);
+ return false;
+ }
+ }
+
+ div = &peri->div;
+ if (divider_exists(div)) {
+ if (div->offset > limit) {
+ pr_err("%s: bad divider offset for %s (%u > %u)\n",
+ __func__, name, div->offset, limit);
+ return false;
+ }
+ }
+
+ div = &peri->pre_div;
+ if (divider_exists(div)) {
+ if (div->offset > limit) {
+ pr_err("%s: bad pre-divider offset for %s "
+ "(%u > %u)\n",
+ __func__, name, div->offset, limit);
+ return false;
+ }
+ }
+
+ sel = &peri->sel;
+ if (selector_exists(sel)) {
+ if (sel->offset > limit) {
+ pr_err("%s: bad selector offset for %s (%u > %u)\n",
+ __func__, name, sel->offset, limit);
+ return false;
+ }
+ }
+
+ trig = &peri->trig;
+ if (trigger_exists(trig)) {
+ if (trig->offset > limit) {
+ pr_err("%s: bad trigger offset for %s (%u > %u)\n",
+ __func__, name, trig->offset, limit);
+ return false;
+ }
+ }
+
+ trig = &peri->pre_trig;
+ if (trigger_exists(trig)) {
+ if (trig->offset > limit) {
+ pr_err("%s: bad pre-trigger offset for %s (%u > %u)\n",
+ __func__, name, trig->offset, limit);
+ return false;
+ }
+ }
+
+ return true;
+}
+
+/* A bit position must be less than the number of bits in a 32-bit register. */
+static bool bit_posn_valid(u32 bit_posn, const char *field_name,
+ const char *clock_name)
+{
+ u32 limit = BITS_PER_BYTE * sizeof(u32) - 1;
+
+ if (bit_posn > limit) {
+ pr_err("%s: bad %s bit for %s (%u > %u)\n", __func__,
+ field_name, clock_name, bit_posn, limit);
+ return false;
+ }
+ return true;
+}
+
+/*
+ * A bitfield must be at least 1 bit wide. Both the low-order and
+ * high-order bits must lie within a 32-bit register. We require
+ * fields to be less than 32 bits wide, mainly because we use
+ * shifting to produce field masks, and shifting a full word width
+ * is not well-defined by the C standard.
+ */
+static bool bitfield_valid(u32 shift, u32 width, const char *field_name,
+ const char *clock_name)
+{
+ u32 limit = BITS_PER_BYTE * sizeof(u32);
+
+ if (!width) {
+ pr_err("%s: bad %s field width 0 for %s\n", __func__,
+ field_name, clock_name);
+ return false;
+ }
+ if (shift + width > limit) {
+ pr_err("%s: bad %s for %s (%u + %u > %u)\n", __func__,
+ field_name, clock_name, shift, width, limit);
+ return false;
+ }
+ return true;
+}
+
+/*
+ * All gates, if defined, have a status bit, and for hardware-only
+ * gates, that's it. Gates that can be software controlled also
+ * have an enable bit. And a gate that can be hardware or software
+ * controlled will have a hardware/software select bit.
+ */
+static bool gate_valid(struct bcm_clk_gate *gate, const char *field_name,
+ const char *clock_name)
+{
+ if (!bit_posn_valid(gate->status_bit, "gate status", clock_name))
+ return false;
+
+ if (gate_is_sw_controllable(gate)) {
+ if (!bit_posn_valid(gate->en_bit, "gate enable", clock_name))
+ return false;
+
+ if (gate_is_hw_controllable(gate)) {
+ if (!bit_posn_valid(gate->hw_sw_sel_bit,
+ "gate hw/sw select",
+ clock_name))
+ return false;
+ }
+ } else {
+ BUG_ON(!gate_is_hw_controllable(gate));
+ }
+
+ return true;
+}
+
+/*
+ * A selector bitfield must be valid. Its parent_sel array must
+ * also be reasonable for the field.
+ */
+static bool sel_valid(struct bcm_clk_sel *sel, const char *field_name,
+ const char *clock_name)
+{
+ if (!bitfield_valid(sel->shift, sel->width, field_name, clock_name))
+ return false;
+
+ if (sel->parent_count) {
+ u32 max_sel;
+ u32 limit;
+
+ /*
+ * Make sure the selector field can hold all the
+ * selector values we expect to be able to use. A
+ * clock only needs to have a selector defined if it
+ * has more than one parent. And in that case the
+ * highest selector value will be in the last entry
+ * in the array.
+ */
+ max_sel = sel->parent_sel[sel->parent_count - 1];
+ limit = (1 << sel->width) - 1;
+ if (max_sel > limit) {
+ pr_err("%s: bad selector for %s "
+ "(%u needs > %u bits)\n",
+ __func__, clock_name, max_sel,
+ sel->width);
+ return false;
+ }
+ } else {
+ pr_warn("%s: ignoring selector for %s (no parents)\n",
+ __func__, clock_name);
+ selector_clear_exists(sel);
+ kfree(sel->parent_sel);
+ sel->parent_sel = NULL;
+ }
+
+ return true;
+}
+
+/*
+ * A fixed divider just needs to be non-zero. A variable divider
+ * has to have a valid divider bitfield, and if it has a fraction,
+ * the width of the fraction must not be no more than the width of
+ * the divider as a whole.
+ */
+static bool div_valid(struct bcm_clk_div *div, const char *field_name,
+ const char *clock_name)
+{
+ if (divider_is_fixed(div)) {
+ /* Any fixed divider value but 0 is OK */
+ if (div->fixed == 0) {
+ pr_err("%s: bad %s fixed value 0 for %s\n", __func__,
+ field_name, clock_name);
+ return false;
+ }
+ return true;
+ }
+ if (!bitfield_valid(div->shift, div->width, field_name, clock_name))
+ return false;
+
+ if (divider_has_fraction(div))
+ if (div->frac_width > div->width) {
+ pr_warn("%s: bad %s fraction width for %s (%u > %u)\n",
+ __func__, field_name, clock_name,
+ div->frac_width, div->width);
+ return false;
+ }
+
+ return true;
+}
+
+/*
+ * If a clock has two dividers, the combined number of fractional
+ * bits must be representable in a 32-bit unsigned value. This
+ * is because we scale up a dividend using both dividers before
+ * dividing to improve accuracy, and we need to avoid overflow.
+ */
+static bool kona_dividers_valid(struct kona_clk *bcm_clk)
+{
+ struct peri_clk_data *peri = bcm_clk->peri;
+ struct bcm_clk_div *div;
+ struct bcm_clk_div *pre_div;
+ u32 limit;
+
+ BUG_ON(bcm_clk->type != bcm_clk_peri);
+
+ if (!divider_exists(&peri->div) || !divider_exists(&peri->pre_div))
+ return true;
+
+ div = &peri->div;
+ pre_div = &peri->pre_div;
+ if (divider_is_fixed(div) || divider_is_fixed(pre_div))
+ return true;
+
+ limit = BITS_PER_BYTE * sizeof(u32);
+
+ return div->frac_width + pre_div->frac_width <= limit;
+}
+
+
+/* A trigger just needs to represent a valid bit position */
+static bool trig_valid(struct bcm_clk_trig *trig, const char *field_name,
+ const char *clock_name)
+{
+ return bit_posn_valid(trig->bit, field_name, clock_name);
+}
+
+/* Determine whether the set of peripheral clock registers are valid. */
+static bool
+peri_clk_data_valid(struct kona_clk *bcm_clk)
+{
+ struct peri_clk_data *peri;
+ struct bcm_clk_gate *gate;
+ struct bcm_clk_sel *sel;
+ struct bcm_clk_div *div;
+ struct bcm_clk_div *pre_div;
+ struct bcm_clk_trig *trig;
+ const char *name;
+
+ BUG_ON(bcm_clk->type != bcm_clk_peri);
+
+ /*
+ * First validate register offsets. This is the only place
+ * where we need something from the ccu, so we do these
+ * together.
+ */
+ if (!peri_clk_data_offsets_valid(bcm_clk))
+ return false;
+
+ peri = bcm_clk->peri;
+ name = bcm_clk->name;
+ gate = &peri->gate;
+ if (gate_exists(gate) && !gate_valid(gate, "gate", name))
+ return false;
+
+ sel = &peri->sel;
+ if (selector_exists(sel)) {
+ if (!sel_valid(sel, "selector", name))
+ return false;
+
+ } else if (sel->parent_count > 1) {
+ pr_err("%s: multiple parents but no selector for %s\n",
+ __func__, name);
+
+ return false;
+ }
+
+ div = &peri->div;
+ pre_div = &peri->pre_div;
+ if (divider_exists(div)) {
+ if (!div_valid(div, "divider", name))
+ return false;
+
+ if (divider_exists(pre_div))
+ if (!div_valid(pre_div, "pre-divider", name))
+ return false;
+ } else if (divider_exists(pre_div)) {
+ pr_err("%s: pre-divider but no divider for %s\n", __func__,
+ name);
+ return false;
+ }
+
+ trig = &peri->trig;
+ if (trigger_exists(trig)) {
+ if (!trig_valid(trig, "trigger", name))
+ return false;
+
+ if (trigger_exists(&peri->pre_trig)) {
+ if (!trig_valid(trig, "pre-trigger", name)) {
+ return false;
+ }
+ }
+ if (!clk_requires_trigger(bcm_clk)) {
+ pr_warn("%s: ignoring trigger for %s (not needed)\n",
+ __func__, name);
+ trigger_clear_exists(trig);
+ }
+ } else if (trigger_exists(&peri->pre_trig)) {
+ pr_err("%s: pre-trigger but no trigger for %s\n", __func__,
+ name);
+ return false;
+ } else if (clk_requires_trigger(bcm_clk)) {
+ pr_err("%s: required trigger missing for %s\n", __func__,
+ name);
+ return false;
+ }
+
+ return kona_dividers_valid(bcm_clk);
+}
+
+static bool kona_clk_valid(struct kona_clk *bcm_clk)
+{
+ switch (bcm_clk->type) {
+ case bcm_clk_peri:
+ if (!peri_clk_data_valid(bcm_clk))
+ return false;
+ break;
+ default:
+ pr_err("%s: unrecognized clock type (%d)\n", __func__,
+ (int)bcm_clk->type);
+ return false;
+ }
+ return true;
+}
+
+/*
+ * Scan an array of parent clock names to determine whether there
+ * are any entries containing BAD_CLK_NAME. Such entries are
+ * placeholders for non-supported clocks. Keep track of the
+ * position of each clock name in the original array.
+ *
+ * Allocates an array of pointers to to hold the names of all
+ * non-null entries in the original array, and returns a pointer to
+ * that array in *names. This will be used for registering the
+ * clock with the common clock code. On successful return,
+ * *count indicates how many entries are in that names array.
+ *
+ * If there is more than one entry in the resulting names array,
+ * another array is allocated to record the parent selector value
+ * for each (defined) parent clock. This is the value that
+ * represents this parent clock in the clock's source selector
+ * register. The position of the clock in the original parent array
+ * defines that selector value. The number of entries in this array
+ * is the same as the number of entries in the parent names array.
+ *
+ * The array of selector values is returned. If the clock has no
+ * parents, no selector is required and a null pointer is returned.
+ *
+ * Returns a null pointer if the clock names array supplied was
+ * null. (This is not an error.)
+ *
+ * Returns a pointer-coded error if an error occurs.
+ */
+static u32 *parent_process(const char *clocks[],
+ u32 *count, const char ***names)
+{
+ static const char **parent_names;
+ static u32 *parent_sel;
+ const char **clock;
+ u32 parent_count;
+ u32 bad_count = 0;
+ u32 orig_count;
+ u32 i;
+ u32 j;
+
+ *count = 0; /* In case of early return */
+ *names = NULL;
+ if (!clocks)
+ return NULL;
+
+ /*
+ * Count the number of names in the null-terminated array,
+ * and find out how many of those are actually clock names.
+ */
+ for (clock = clocks; *clock; clock++)
+ if (*clock == BAD_CLK_NAME)
+ bad_count++;
+ orig_count = (u32)(clock - clocks);
+ parent_count = orig_count - bad_count;
+
+ /* If all clocks are unsupported, we treat it as no clock */
+ if (!parent_count)
+ return NULL;
+
+ /* Avoid exceeding our parent clock limit */
+ if (parent_count > PARENT_COUNT_MAX) {
+ pr_err("%s: too many parents (%u > %u)\n", __func__,
+ parent_count, PARENT_COUNT_MAX);
+ return ERR_PTR(-EINVAL);
+ }
+
+ /*
+ * There is one parent name for each defined parent clock.
+ * We also maintain an array containing the selector value
+ * for each defined clock. If there's only one clock, the
+ * selector is not required, but we allocate space for the
+ * array anyway to keep things simple.
+ */
+ parent_names = kmalloc(parent_count * sizeof(parent_names), GFP_KERNEL);
+ if (!parent_names) {
+ pr_err("%s: error allocating %u parent names\n", __func__,
+ parent_count);
+ return ERR_PTR(-ENOMEM);
+ }
+
+ /* There is at least one parent, so allocate a selector array */
+
+ parent_sel = kmalloc(parent_count * sizeof(*parent_sel), GFP_KERNEL);
+ if (!parent_sel) {
+ pr_err("%s: error allocating %u parent selectors\n", __func__,
+ parent_count);
+ kfree(parent_names);
+
+ return ERR_PTR(-ENOMEM);
+ }
+
+ /* Now fill in the parent names and selector arrays */
+ for (i = 0, j = 0; i < orig_count; i++) {
+ if (clocks[i] != BAD_CLK_NAME) {
+ parent_names[j] = clocks[i];
+ parent_sel[j] = i;
+ j++;
+ }
+ }
+ *names = parent_names;
+ *count = parent_count;
+
+ return parent_sel;
+}
+
+static int
+clk_sel_setup(const char **clocks, struct bcm_clk_sel *sel,
+ struct clk_init_data *init_data)
+{
+ const char **parent_names = NULL;
+ u32 parent_count = 0;
+ u32 *parent_sel;
+
+ /*
+ * If a peripheral clock has multiple parents, the value
+ * used by the hardware to select that parent is represented
+ * by the parent clock's position in the "clocks" list. Some
+ * values don't have defined or supported clocks; these will
+ * have BAD_CLK_NAME entries in the parents[] array. The
+ * list is terminated by a NULL entry.
+ *
+ * We need to supply (only) the names of defined parent
+ * clocks when registering a clock though, so we use an
+ * array of parent selector values to map between the
+ * indexes the common clock code uses and the selector
+ * values we need.
+ */
+ parent_sel = parent_process(clocks, &parent_count, &parent_names);
+ if (IS_ERR(parent_sel)) {
+ int ret = PTR_ERR(parent_sel);
+
+ pr_err("%s: error processing parent clocks for %s (%d)\n",
+ __func__, init_data->name, ret);
+
+ return ret;
+ }
+
+ init_data->parent_names = parent_names;
+ init_data->num_parents = parent_count;
+
+ sel->parent_count = parent_count;
+ sel->parent_sel = parent_sel;
+
+ return 0;
+}
+
+static void clk_sel_teardown(struct bcm_clk_sel *sel,
+ struct clk_init_data *init_data)
+{
+ kfree(sel->parent_sel);
+ sel->parent_sel = NULL;
+ sel->parent_count = 0;
+
+ init_data->num_parents = 0;
+ kfree(init_data->parent_names);
+ init_data->parent_names = NULL;
+}
+
+static void peri_clk_teardown(struct peri_clk_data *data,
+ struct clk_init_data *init_data)
+{
+ clk_sel_teardown(&data->sel, init_data);
+ init_data->ops = NULL;
+}
+
+/*
+ * Caller is responsible for freeing the parent_names[] and
+ * parent_sel[] arrays in the peripheral clock's "data" structure
+ * that can be assigned if the clock has one or more parent clocks
+ * associated with it.
+ */
+static int peri_clk_setup(struct ccu_data *ccu, struct peri_clk_data *data,
+ struct clk_init_data *init_data)
+{
+ init_data->ops = &kona_peri_clk_ops;
+ init_data->flags = 0;
+
+ return clk_sel_setup(data->clocks, &data->sel, init_data);
+}
+
+static void bcm_clk_teardown(struct kona_clk *bcm_clk)
+{
+ switch (bcm_clk->type) {
+ case bcm_clk_peri:
+ peri_clk_teardown(bcm_clk->data, &bcm_clk->init_data);
+ break;
+ default:
+ break;
+ }
+ bcm_clk->data = NULL;
+ bcm_clk->type = bcm_clk_none;
+}
+
+static void kona_clk_teardown(struct clk *clk)
+{
+ struct clk_hw *hw;
+ struct kona_clk *bcm_clk;
+
+ if (!clk)
+ return;
+
+ hw = __clk_get_hw(clk);
+ if (!hw) {
+ pr_err("%s: clk %p has null hw pointer\n", __func__, clk);
+ return;
+ }
+ clk_unregister(clk);
+
+ bcm_clk = to_kona_clk(hw);
+ bcm_clk_teardown(bcm_clk);
+}
+
+struct clk *kona_clk_setup(struct ccu_data *ccu, const char *name,
+ enum bcm_clk_type type, void *data)
+{
+ struct kona_clk *bcm_clk;
+ struct clk_init_data *init_data;
+ struct clk *clk = NULL;
+
+ bcm_clk = kzalloc(sizeof(*bcm_clk), GFP_KERNEL);
+ if (!bcm_clk) {
+ pr_err("%s: failed to allocate bcm_clk for %s\n", __func__,
+ name);
+ return NULL;
+ }
+ bcm_clk->ccu = ccu;
+ bcm_clk->name = name;
+
+ init_data = &bcm_clk->init_data;
+ init_data->name = name;
+ switch (type) {
+ case bcm_clk_peri:
+ if (peri_clk_setup(ccu, data, init_data))
+ goto out_free;
+ break;
+ default:
+ data = NULL;
+ break;
+ }
+ bcm_clk->type = type;
+ bcm_clk->data = data;
+
+ /* Make sure everything makes sense before we set it up */
+ if (!kona_clk_valid(bcm_clk)) {
+ pr_err("%s: clock data invalid for %s\n", __func__, name);
+ goto out_teardown;
+ }
+
+ bcm_clk->hw.init = init_data;
+ clk = clk_register(NULL, &bcm_clk->hw);
+ if (IS_ERR(clk)) {
+ pr_err("%s: error registering clock %s (%ld)\n", __func__,
+ name, PTR_ERR(clk));
+ goto out_teardown;
+ }
+ BUG_ON(!clk);
+
+ return clk;
+out_teardown:
+ bcm_clk_teardown(bcm_clk);
+out_free:
+ kfree(bcm_clk);
+
+ return NULL;
+}
+
+static void ccu_clks_teardown(struct ccu_data *ccu)
+{
+ u32 i;
+
+ for (i = 0; i < ccu->data.clk_num; i++)
+ kona_clk_teardown(ccu->data.clks[i]);
+ kfree(ccu->data.clks);
+}
+
+static void kona_ccu_teardown(struct ccu_data *ccu)
+{
+ if (!ccu)
+ return;
+
+ if (!ccu->base)
+ goto done;
+
+ of_clk_del_provider(ccu->node); /* safe if never added */
+ ccu_clks_teardown(ccu);
+ list_del(&ccu->links);
+ of_node_put(ccu->node);
+ iounmap(ccu->base);
+done:
+ kfree(ccu->name);
+ kfree(ccu);
+}
+
+/*
+ * Set up a CCU. Call the provided ccu_clks_setup callback to
+ * initialize the array of clocks provided by the CCU.
+ */
+void __init kona_dt_ccu_setup(struct device_node *node,
+ int (*ccu_clks_setup)(struct ccu_data *))
+{
+ struct ccu_data *ccu;
+ struct resource res = { 0 };
+ resource_size_t range;
+ int ret;
+
+ ccu = kzalloc(sizeof(*ccu), GFP_KERNEL);
+ if (ccu)
+ ccu->name = kstrdup(node->name, GFP_KERNEL);
+ if (!ccu || !ccu->name) {
+ pr_err("%s: unable to allocate CCU struct for %s\n",
+ __func__, node->name);
+ kfree(ccu);
+
+ return;
+ }
+
+ ret = of_address_to_resource(node, 0, &res);
+ if (ret) {
+ pr_err("%s: no valid CCU registers found for %s\n", __func__,
+ node->name);
+ goto out_err;
+ }
+
+ range = resource_size(&res);
+ if (range > (resource_size_t)U32_MAX) {
+ pr_err("%s: address range too large for %s\n", __func__,
+ node->name);
+ goto out_err;
+ }
+
+ ccu->range = (u32)range;
+ ccu->base = ioremap(res.start, ccu->range);
+ if (!ccu->base) {
+ pr_err("%s: unable to map CCU registers for %s\n", __func__,
+ node->name);
+ goto out_err;
+ }
+
+ spin_lock_init(&ccu->lock);
+ INIT_LIST_HEAD(&ccu->links);
+ ccu->node = of_node_get(node);
+
+ list_add_tail(&ccu->links, &ccu_list);
+
+ /* Set up clocks array (in ccu->data) */
+ if (ccu_clks_setup(ccu))
+ goto out_err;
+
+ ret = of_clk_add_provider(node, of_clk_src_onecell_get, &ccu->data);
+ if (ret) {
+ pr_err("%s: error adding ccu %s as provider (%d)\n", __func__,
+ node->name, ret);
+ goto out_err;
+ }
+
+ if (!kona_ccu_init(ccu))
+ pr_err("Broadcom %s initialization had errors\n", node->name);
+
+ return;
+out_err:
+ kona_ccu_teardown(ccu);
+ pr_err("Broadcom %s setup aborted\n", node->name);
+}
diff --git a/drivers/clk/bcm/clk-kona.c b/drivers/clk/bcm/clk-kona.c
new file mode 100644
index 0000000..e3d339e
--- /dev/null
+++ b/drivers/clk/bcm/clk-kona.c
@@ -0,0 +1,1033 @@
+/*
+ * Copyright (C) 2013 Broadcom Corporation
+ * Copyright 2013 Linaro Limited
+ *
+ * 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 version 2.
+ *
+ * This program is distributed "as is" WITHOUT ANY WARRANTY of any
+ * kind, whether express or implied; without even the implied warranty
+ * of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ */
+
+#include "clk-kona.h"
+
+#include <linux/delay.h>
+
+#define CCU_ACCESS_PASSWORD 0xA5A500
+#define CLK_GATE_DELAY_LOOP 2000
+
+/* Bitfield operations */
+
+/* Produces a mask of set bits covering a range of a 32-bit value */
+static inline u32 bitfield_mask(u32 shift, u32 width)
+{
+ return ((1 << width) - 1) << shift;
+}
+
+/* Extract the value of a bitfield found within a given register value */
+static inline u32 bitfield_extract(u32 reg_val, u32 shift, u32 width)
+{
+ return (reg_val & bitfield_mask(shift, width)) >> shift;
+}
+
+/* Replace the value of a bitfield found within a given register value */
+static inline u32 bitfield_replace(u32 reg_val, u32 shift, u32 width, u32 val)
+{
+ u32 mask = bitfield_mask(shift, width);
+
+ return (reg_val & ~mask) | (val << shift);
+}
+
+/* Divider and scaling helpers */
+
+/*
+ * Implement DIV_ROUND_CLOSEST() for 64-bit dividend and both values
+ * unsigned. Note that unlike do_div(), the remainder is discarded
+ * and the return value is the quotient (not the remainder).
+ */
+u64 do_div_round_closest(u64 dividend, unsigned long divisor)
+{
+ u64 result;
+
+ result = dividend + ((u64)divisor >> 1);
+ (void)do_div(result, divisor);
+
+ return result;
+}
+
+/* Convert a divider into the scaled divisor value it represents. */
+static inline u64 scaled_div_value(struct bcm_clk_div *div, u32 reg_div)
+{
+ return (u64)reg_div + ((u64)1 << div->frac_width);
+}
+
+/*
+ * Build a scaled divider value as close as possible to the
+ * given whole part (div_value) and fractional part (expressed
+ * in billionths).
+ */
+u64 scaled_div_build(struct bcm_clk_div *div, u32 div_value, u32 billionths)
+{
+ u64 combined;
+
+ BUG_ON(!div_value);
+ BUG_ON(billionths >= BILLION);
+
+ combined = (u64)div_value * BILLION + billionths;
+ combined <<= div->frac_width;
+
+ return do_div_round_closest(combined, BILLION);
+}
+
+/* The scaled minimum divisor representable by a divider */
+static inline u64
+scaled_div_min(struct bcm_clk_div *div)
+{
+ if (divider_is_fixed(div))
+ return (u64)div->fixed;
+
+ return scaled_div_value(div, 0);
+}
+
+/* The scaled maximum divisor representable by a divider */
+u64 scaled_div_max(struct bcm_clk_div *div)
+{
+ u32 reg_div;
+
+ if (divider_is_fixed(div))
+ return (u64)div->fixed;
+
+ reg_div = ((u32)1 << div->width) - 1;
+
+ return scaled_div_value(div, reg_div);
+}
+
+/*
+ * Convert a scaled divisor into its divider representation as
+ * stored in a divider register field.
+ */
+static inline u32
+divider(struct bcm_clk_div *div, u64 scaled_div)
+{
+ BUG_ON(scaled_div < scaled_div_min(div));
+ BUG_ON(scaled_div > scaled_div_max(div));
+
+ return (u32)(scaled_div - ((u64)1 << div->frac_width));
+}
+
+/* Return a rate scaled for use when dividing by a scaled divisor. */
+static inline u64
+scale_rate(struct bcm_clk_div *div, u32 rate)
+{
+ if (divider_is_fixed(div))
+ return (u64)rate;
+
+ return (u64)rate << div->frac_width;
+}
+
+/* CCU access */
+
+/* Read a 32-bit register value from a CCU's address space. */
+static inline u32 __ccu_read(struct ccu_data *ccu, u32 reg_offset)
+{
+ return readl(ccu->base + reg_offset);
+}
+
+/* Write a 32-bit register value into a CCU's address space. */
+static inline void
+__ccu_write(struct ccu_data *ccu, u32 reg_offset, u32 reg_val)
+{
+ writel(reg_val, ccu->base + reg_offset);
+}
+
+static inline unsigned long ccu_lock(struct ccu_data *ccu)
+{
+ unsigned long flags;
+
+ spin_lock_irqsave(&ccu->lock, flags);
+
+ return flags;
+}
+static inline void ccu_unlock(struct ccu_data *ccu, unsigned long flags)
+{
+ spin_unlock_irqrestore(&ccu->lock, flags);
+}
+
+/*
+ * Enable/disable write access to CCU protected registers. The
+ * WR_ACCESS register for all CCUs is at offset 0.
+ */
+static inline void __ccu_write_enable(struct ccu_data *ccu)
+{
+ if (ccu->write_enabled) {
+ pr_err("%s: access already enabled for %s\n", __func__,
+ ccu->name);
+ return;
+ }
+ ccu->write_enabled = true;
+ __ccu_write(ccu, 0, CCU_ACCESS_PASSWORD | 1);
+}
+
+static inline void __ccu_write_disable(struct ccu_data *ccu)
+{
+ if (!ccu->write_enabled) {
+ pr_err("%s: access wasn't enabled for %s\n", __func__,
+ ccu->name);
+ return;
+ }
+
+ __ccu_write(ccu, 0, CCU_ACCESS_PASSWORD);
+ ccu->write_enabled = false;
+}
+
+/*
+ * Poll a register in a CCU's address space, returning when the
+ * specified bit in that register's value is set (or clear). Delay
+ * a microsecond after each read of the register. Returns true if
+ * successful, or false if we gave up trying.
+ *
+ * Caller must ensure the CCU lock is held.
+ */
+static inline bool
+__ccu_wait_bit(struct ccu_data *ccu, u32 reg_offset, u32 bit, bool want)
+{
+ unsigned int tries;
+ u32 bit_mask = 1 << bit;
+
+ for (tries = 0; tries < CLK_GATE_DELAY_LOOP; tries++) {
+ u32 val;
+ bool bit_val;
+
+ val = __ccu_read(ccu, reg_offset);
+ bit_val = (val & bit_mask) != 0;
+ if (bit_val == want)
+ return true;
+ udelay(1);
+ }
+ return false;
+}
+
+/* Gate operations */
+
+/* Determine whether a clock is gated. CCU lock must be held. */
+static bool
+__is_clk_gate_enabled(struct ccu_data *ccu, struct bcm_clk_gate *gate)
+{
+ u32 bit_mask;
+ u32 reg_val;
+
+ /* If there is no gate we can assume it's enabled. */
+ if (!gate_exists(gate))
+ return true;
+
+ bit_mask = 1 << gate->status_bit;
+ reg_val = __ccu_read(ccu, gate->offset);
+
+ return (reg_val & bit_mask) != 0;
+}
+
+/* Determine whether a clock is gated. */
+static bool
+is_clk_gate_enabled(struct ccu_data *ccu, struct bcm_clk_gate *gate)
+{
+ long flags;
+ bool ret;
+
+ /* Avoid taking the lock if we can */
+ if (!gate_exists(gate))
+ return true;
+
+ flags = ccu_lock(ccu);
+ ret = __is_clk_gate_enabled(ccu, gate);
+ ccu_unlock(ccu, flags);
+
+ return ret;
+}
+
+/*
+ * Commit our desired gate state to the hardware.
+ * Returns true if successful, false otherwise.
+ */
+static bool
+__gate_commit(struct ccu_data *ccu, struct bcm_clk_gate *gate)
+{
+ u32 reg_val;
+ u32 mask;
+ bool enabled = false;
+
+ BUG_ON(!gate_exists(gate));
+ if (!gate_is_sw_controllable(gate))
+ return true; /* Nothing we can change */
+
+ reg_val = __ccu_read(ccu, gate->offset);
+
+ /* For a hardware/software gate, set which is in control */
+ if (gate_is_hw_controllable(gate)) {
+ mask = (u32)1 << gate->hw_sw_sel_bit;
+ if (gate_is_sw_managed(gate))
+ reg_val |= mask;
+ else
+ reg_val &= ~mask;
+ }
+
+ /*
+ * If software is in control, enable or disable the gate.
+ * If hardware is, clear the enabled bit for good measure.
+ * If a software controlled gate can't be disabled, we're
+ * required to write a 0 into the enable bit (but the gate
+ * will be enabled).
+ */
+ mask = (u32)1 << gate->en_bit;
+ if (gate_is_sw_managed(gate) && (enabled = gate_is_enabled(gate)) &&
+ !gate_is_no_disable(gate))
+ reg_val |= mask;
+ else
+ reg_val &= ~mask;
+
+ __ccu_write(ccu, gate->offset, reg_val);
+
+ /* For a hardware controlled gate, we're done */
+ if (!gate_is_sw_managed(gate))
+ return true;
+
+ /* Otherwise wait for the gate to be in desired state */
+ return __ccu_wait_bit(ccu, gate->offset, gate->status_bit, enabled);
+}
+
+/*
+ * Initialize a gate. Our desired state (hardware/software select,
+ * and if software, its enable state) is committed to hardware
+ * without the usual checks to see if it's already set up that way.
+ * Returns true if successful, false otherwise.
+ */
+static bool gate_init(struct ccu_data *ccu, struct bcm_clk_gate *gate)
+{
+ if (!gate_exists(gate))
+ return true;
+ return __gate_commit(ccu, gate);
+}
+
+/*
+ * Set a gate to enabled or disabled state. Does nothing if the
+ * gate is not currently under software control, or if it is already
+ * in the requested state. Returns true if successful, false
+ * otherwise. CCU lock must be held.
+ */
+static bool
+__clk_gate(struct ccu_data *ccu, struct bcm_clk_gate *gate, bool enable)
+{
+ bool ret;
+
+ if (!gate_exists(gate) || !gate_is_sw_managed(gate))
+ return true; /* Nothing to do */
+
+ if (!enable && gate_is_no_disable(gate)) {
+ pr_warn("%s: invalid gate disable request (ignoring)\n",
+ __func__);
+ return true;
+ }
+
+ if (enable == gate_is_enabled(gate))
+ return true; /* No change */
+
+ gate_flip_enabled(gate);
+ ret = __gate_commit(ccu, gate);
+ if (!ret)
+ gate_flip_enabled(gate); /* Revert the change */
+
+ return ret;
+}
+
+/* Enable or disable a gate. Returns 0 if successful, -EIO otherwise */
+static int clk_gate(struct ccu_data *ccu, const char *name,
+ struct bcm_clk_gate *gate, bool enable)
+{
+ unsigned long flags;
+ bool success;
+
+ /*
+ * Avoid taking the lock if we can. We quietly ignore
+ * requests to change state that don't make sense.
+ */
+ if (!gate_exists(gate) || !gate_is_sw_managed(gate))
+ return 0;
+ if (!enable && gate_is_no_disable(gate))
+ return 0;
+
+ flags = ccu_lock(ccu);
+ __ccu_write_enable(ccu);
+
+ success = __clk_gate(ccu, gate, enable);
+
+ __ccu_write_disable(ccu);
+ ccu_unlock(ccu, flags);
+
+ if (success)
+ return 0;
+
+ pr_err("%s: failed to %s gate for %s\n", __func__,
+ enable ? "enable" : "disable", name);
+
+ return -EIO;
+}
+
+/* Trigger operations */
+
+/*
+ * Caller must ensure CCU lock is held and access is enabled.
+ * Returns true if successful, false otherwise.
+ */
+static bool __clk_trigger(struct ccu_data *ccu, struct bcm_clk_trig *trig)
+{
+ /* Trigger the clock and wait for it to finish */
+ __ccu_write(ccu, trig->offset, 1 << trig->bit);
+
+ return __ccu_wait_bit(ccu, trig->offset, trig->bit, false);
+}
+
+/* Divider operations */
+
+/* Read a divider value and return the scaled divisor it represents. */
+static u64 divider_read_scaled(struct ccu_data *ccu, struct bcm_clk_div *div)
+{
+ unsigned long flags;
+ u32 reg_val;
+ u32 reg_div;
+
+ if (divider_is_fixed(div))
+ return (u64)div->fixed;
+
+ flags = ccu_lock(ccu);
+ reg_val = __ccu_read(ccu, div->offset);
+ ccu_unlock(ccu, flags);
+
+ /* Extract the full divider field from the register value */
+ reg_div = bitfield_extract(reg_val, div->shift, div->width);
+
+ /* Return the scaled divisor value it represents */
+ return scaled_div_value(div, reg_div);
+}
+
+/*
+ * Convert a divider's scaled divisor value into its recorded form
+ * and commit it into the hardware divider register.
+ *
+ * Returns 0 on success. Returns -EINVAL for invalid arguments.
+ * Returns -ENXIO if gating failed, and -EIO if a trigger failed.
+ */
+static int __div_commit(struct ccu_data *ccu, struct bcm_clk_gate *gate,
+ struct bcm_clk_div *div, struct bcm_clk_trig *trig)
+{
+ bool enabled;
+ u32 reg_div;
+ u32 reg_val;
+ int ret = 0;
+
+ BUG_ON(divider_is_fixed(div));
+
+ /*
+ * If we're just initializing the divider, and no initial
+ * state was defined in the device tree, we just find out
+ * what its current value is rather than updating it.
+ */
+ if (div->scaled_div == BAD_SCALED_DIV_VALUE) {
+ reg_val = __ccu_read(ccu, div->offset);
+ reg_div = bitfield_extract(reg_val, div->shift, div->width);
+ div->scaled_div = scaled_div_value(div, reg_div);
+
+ return 0;
+ }
+
+ /* Convert the scaled divisor to the value we need to record */
+ reg_div = divider(div, div->scaled_div);
+
+ /* Clock needs to be enabled before changing the rate */
+ enabled = __is_clk_gate_enabled(ccu, gate);
+ if (!enabled && !__clk_gate(ccu, gate, true)) {
+ ret = -ENXIO;
+ goto out;
+ }
+
+ /* Replace the divider value and record the result */
+ reg_val = __ccu_read(ccu, div->offset);
+ reg_val = bitfield_replace(reg_val, div->shift, div->width, reg_div);
+ __ccu_write(ccu, div->offset, reg_val);
+
+ /* If the trigger fails we still want to disable the gate */
+ if (!__clk_trigger(ccu, trig))
+ ret = -EIO;
+
+ /* Disable the clock again if it was disabled to begin with */
+ if (!enabled && !__clk_gate(ccu, gate, false))
+ ret = ret ? ret : -ENXIO; /* return first error */
+out:
+ return ret;
+}
+
+/*
+ * Initialize a divider by committing our desired state to hardware
+ * without the usual checks to see if it's already set up that way.
+ * Returns true if successful, false otherwise.
+ */
+static bool div_init(struct ccu_data *ccu, struct bcm_clk_gate *gate,
+ struct bcm_clk_div *div, struct bcm_clk_trig *trig)
+{
+ if (!divider_exists(div) || divider_is_fixed(div))
+ return true;
+ return !__div_commit(ccu, gate, div, trig);
+}
+
+static int divider_write(struct ccu_data *ccu, struct bcm_clk_gate *gate,
+ struct bcm_clk_div *div, struct bcm_clk_trig *trig,
+ u64 scaled_div)
+{
+ unsigned long flags;
+ u64 previous;
+ int ret;
+
+ BUG_ON(divider_is_fixed(div));
+
+ previous = div->scaled_div;
+ if (previous == scaled_div)
+ return 0; /* No change */
+
+ div->scaled_div = scaled_div;
+
+ flags = ccu_lock(ccu);
+ __ccu_write_enable(ccu);
+
+ ret = __div_commit(ccu, gate, div, trig);
+
+ __ccu_write_disable(ccu);
+ ccu_unlock(ccu, flags);
+
+ if (ret)
+ div->scaled_div = previous; /* Revert the change */
+
+ return ret;
+
+}
+
+/* Common clock rate helpers */
+
+/*
+ * Implement the common clock framework recalc_rate method, taking
+ * into account a divider and an optional pre-divider. The
+ * pre-divider register pointer may be NULL.
+ */
+static unsigned long clk_recalc_rate(struct ccu_data *ccu,
+ struct bcm_clk_div *div, struct bcm_clk_div *pre_div,
+ unsigned long parent_rate)
+{
+ u64 scaled_parent_rate;
+ u64 scaled_div;
+ u64 result;
+
+ if (!divider_exists(div))
+ return parent_rate;
+
+ if (parent_rate > (unsigned long)LONG_MAX)
+ return 0; /* actually this would be a caller bug */
+
+ /*
+ * If there is a pre-divider, divide the scaled parent rate
+ * by the pre-divider value first. In this case--to improve
+ * accuracy--scale the parent rate by *both* the pre-divider
+ * value and the divider before actually computing the
+ * result of the pre-divider.
+ *
+ * If there's only one divider, just scale the parent rate.
+ */
+ if (pre_div && divider_exists(pre_div)) {
+ u64 scaled_rate;
+
+ scaled_rate = scale_rate(pre_div, parent_rate);
+ scaled_rate = scale_rate(div, scaled_rate);
+ scaled_div = divider_read_scaled(ccu, pre_div);
+ scaled_parent_rate = do_div_round_closest(scaled_rate,
+ scaled_div);
+ } else {
+ scaled_parent_rate = scale_rate(div, parent_rate);
+ }
+
+ /*
+ * Get the scaled divisor value, and divide the scaled
+ * parent rate by that to determine this clock's resulting
+ * rate.
+ */
+ scaled_div = divider_read_scaled(ccu, div);
+ result = do_div_round_closest(scaled_parent_rate, scaled_div);
+
+ return (unsigned long)result;
+}
+
+/*
+ * Compute the output rate produced when a given parent rate is fed
+ * into two dividers. The pre-divider can be NULL, and even if it's
+ * non-null it may be nonexistent. It's also OK for the divider to
+ * be nonexistent, and in that case the pre-divider is also ignored.
+ *
+ * If scaled_div is non-null, it is used to return the scaled divisor
+ * value used by the (downstream) divider to produce that rate.
+ */
+static long round_rate(struct ccu_data *ccu, struct bcm_clk_div *div,
+ struct bcm_clk_div *pre_div,
+ unsigned long rate, unsigned long parent_rate,
+ u64 *scaled_div)
+{
+ u64 scaled_parent_rate;
+ u64 min_scaled_div;
+ u64 max_scaled_div;
+ u64 best_scaled_div;
+ u64 result;
+
+ BUG_ON(!divider_exists(div));
+ BUG_ON(!rate);
+ BUG_ON(parent_rate > (u64)LONG_MAX);
+
+ /*
+ * If there is a pre-divider, divide the scaled parent rate
+ * by the pre-divider value first. In this case--to improve
+ * accuracy--scale the parent rate by *both* the pre-divider
+ * value and the divider before actually computing the
+ * result of the pre-divider.
+ *
+ * If there's only one divider, just scale the parent rate.
+ *
+ * For simplicity we treat the pre-divider as fixed (for now).
+ */
+ if (divider_exists(pre_div)) {
+ u64 scaled_rate;
+ u64 scaled_pre_div;
+
+ scaled_rate = scale_rate(pre_div, parent_rate);
+ scaled_rate = scale_rate(div, scaled_rate);
+ scaled_pre_div = divider_read_scaled(ccu, pre_div);
+ scaled_parent_rate = do_div_round_closest(scaled_rate,
+ scaled_pre_div);
+ } else {
+ scaled_parent_rate = scale_rate(div, parent_rate);
+ }
+
+ /*
+ * Compute the best possible divider and ensure it is in
+ * range. A fixed divider can't be changed, so just report
+ * the best we can do.
+ */
+ if (!divider_is_fixed(div)) {
+ best_scaled_div = do_div_round_closest(scaled_parent_rate,
+ rate);
+ min_scaled_div = scaled_div_min(div);
+ max_scaled_div = scaled_div_max(div);
+ if (best_scaled_div > max_scaled_div)
+ best_scaled_div = max_scaled_div;
+ else if (best_scaled_div < min_scaled_div)
+ best_scaled_div = min_scaled_div;
+ } else {
+ best_scaled_div = divider_read_scaled(ccu, div);
+ }
+
+ /* OK, figure out the resulting rate */
+ result = do_div_round_closest(scaled_parent_rate, best_scaled_div);
+
+ if (scaled_div)
+ *scaled_div = best_scaled_div;
+
+ return (long)result;
+}
+
+/* Common clock parent helpers */
+
+/*
+ * For a given parent selector (register field) value, find the
+ * index into a selector's parent_sel array that contains it.
+ * Returns the index, or BAD_CLK_INDEX if it's not found.
+ */
+static u8 parent_index(struct bcm_clk_sel *sel, u8 parent_sel)
+{
+ u8 i;
+
+ BUG_ON(sel->parent_count > (u32)U8_MAX);
+ for (i = 0; i < sel->parent_count; i++)
+ if (sel->parent_sel[i] == parent_sel)
+ return i;
+ return BAD_CLK_INDEX;
+}
+
+/*
+ * Fetch the current value of the selector, and translate that into
+ * its corresponding index in the parent array we registered with
+ * the clock framework.
+ *
+ * Returns parent array index that corresponds with the value found,
+ * or BAD_CLK_INDEX if the found value is out of range.
+ */
+static u8 selector_read_index(struct ccu_data *ccu, struct bcm_clk_sel *sel)
+{
+ unsigned long flags;
+ u32 reg_val;
+ u32 parent_sel;
+ u8 index;
+
+ /* If there's no selector, there's only one parent */
+ if (!selector_exists(sel))
+ return 0;
+
+ /* Get the value in the selector register */
+ flags = ccu_lock(ccu);
+ reg_val = __ccu_read(ccu, sel->offset);
+ ccu_unlock(ccu, flags);
+
+ parent_sel = bitfield_extract(reg_val, sel->shift, sel->width);
+
+ /* Look up that selector's parent array index and return it */
+ index = parent_index(sel, parent_sel);
+ if (index == BAD_CLK_INDEX)
+ pr_err("%s: out-of-range parent selector %u (%s 0x%04x)\n",
+ __func__, parent_sel, ccu->name, sel->offset);
+
+ return index;
+}
+
+/*
+ * Commit our desired selector value to the hardware.
+ *
+ * Returns 0 on success. Returns -EINVAL for invalid arguments.
+ * Returns -ENXIO if gating failed, and -EIO if a trigger failed.
+ */
+static int
+__sel_commit(struct ccu_data *ccu, struct bcm_clk_gate *gate,
+ struct bcm_clk_sel *sel, struct bcm_clk_trig *trig)
+{
+ u32 parent_sel;
+ u32 reg_val;
+ bool enabled;
+ int ret = 0;
+
+ BUG_ON(!selector_exists(sel));
+
+ /*
+ * If we're just initializing the selector, and no initial
+ * state was defined in the device tree, we just find out
+ * what its current value is rather than updating it.
+ */
+ if (sel->clk_index == BAD_CLK_INDEX) {
+ u8 index;
+
+ reg_val = __ccu_read(ccu, sel->offset);
+ parent_sel = bitfield_extract(reg_val, sel->shift, sel->width);
+ index = parent_index(sel, parent_sel);
+ if (index == BAD_CLK_INDEX)
+ return -EINVAL;
+ sel->clk_index = index;
+
+ return 0;
+ }
+
+ BUG_ON((u32)sel->clk_index >= sel->parent_count);
+ parent_sel = sel->parent_sel[sel->clk_index];
+
+ /* Clock needs to be enabled before changing the parent */
+ enabled = __is_clk_gate_enabled(ccu, gate);
+ if (!enabled && !__clk_gate(ccu, gate, true))
+ return -ENXIO;
+
+ /* Replace the selector value and record the result */
+ reg_val = __ccu_read(ccu, sel->offset);
+ reg_val = bitfield_replace(reg_val, sel->shift, sel->width, parent_sel);
+ __ccu_write(ccu, sel->offset, reg_val);
+
+ /* If the trigger fails we still want to disable the gate */
+ if (!__clk_trigger(ccu, trig))
+ ret = -EIO;
+
+ /* Disable the clock again if it was disabled to begin with */
+ if (!enabled && !__clk_gate(ccu, gate, false))
+ ret = ret ? ret : -ENXIO; /* return first error */
+
+ return ret;
+}
+
+/*
+ * Initialize a selector by committing our desired state to hardware
+ * without the usual checks to see if it's already set up that way.
+ * Returns true if successful, false otherwise.
+ */
+static bool sel_init(struct ccu_data *ccu, struct bcm_clk_gate *gate,
+ struct bcm_clk_sel *sel, struct bcm_clk_trig *trig)
+{
+ if (!selector_exists(sel))
+ return true;
+ return !__sel_commit(ccu, gate, sel, trig);
+}
+
+/*
+ * Write a new value into a selector register to switch to a
+ * different parent clock. Returns 0 on success, or an error code
+ * (from __sel_commit()) otherwise.
+ */
+static int selector_write(struct ccu_data *ccu, struct bcm_clk_gate *gate,
+ struct bcm_clk_sel *sel, struct bcm_clk_trig *trig,
+ u8 index)
+{
+ unsigned long flags;
+ u8 previous;
+ int ret;
+
+ previous = sel->clk_index;
+ if (previous == index)
+ return 0; /* No change */
+
+ sel->clk_index = index;
+
+ flags = ccu_lock(ccu);
+ __ccu_write_enable(ccu);
+
+ ret = __sel_commit(ccu, gate, sel, trig);
+
+ __ccu_write_disable(ccu);
+ ccu_unlock(ccu, flags);
+
+ if (ret)
+ sel->clk_index = previous; /* Revert the change */
+
+ return ret;
+}
+
+/* Clock operations */
+
+static int kona_peri_clk_enable(struct clk_hw *hw)
+{
+ struct kona_clk *bcm_clk = to_kona_clk(hw);
+ struct bcm_clk_gate *gate = &bcm_clk->peri->gate;
+
+ return clk_gate(bcm_clk->ccu, bcm_clk->name, gate, true);
+}
+
+static void kona_peri_clk_disable(struct clk_hw *hw)
+{
+ struct kona_clk *bcm_clk = to_kona_clk(hw);
+ struct bcm_clk_gate *gate = &bcm_clk->peri->gate;
+
+ (void)clk_gate(bcm_clk->ccu, bcm_clk->name, gate, false);
+}
+
+static int kona_peri_clk_is_enabled(struct clk_hw *hw)
+{
+ struct kona_clk *bcm_clk = to_kona_clk(hw);
+ struct bcm_clk_gate *gate = &bcm_clk->peri->gate;
+
+ return is_clk_gate_enabled(bcm_clk->ccu, gate) ? 1 : 0;
+}
+
+static unsigned long kona_peri_clk_recalc_rate(struct clk_hw *hw,
+ unsigned long parent_rate)
+{
+ struct kona_clk *bcm_clk = to_kona_clk(hw);
+ struct peri_clk_data *data = bcm_clk->peri;
+
+ return clk_recalc_rate(bcm_clk->ccu, &data->div, &data->pre_div,
+ parent_rate);
+}
+
+static long kona_peri_clk_round_rate(struct clk_hw *hw, unsigned long rate,
+ unsigned long *parent_rate)
+{
+ struct kona_clk *bcm_clk = to_kona_clk(hw);
+ struct bcm_clk_div *div = &bcm_clk->peri->div;
+
+ if (!divider_exists(div))
+ return __clk_get_rate(hw->clk);
+
+ /* Quietly avoid a zero rate */
+ return round_rate(bcm_clk->ccu, div, &bcm_clk->peri->pre_div,
+ rate ? rate : 1, *parent_rate, NULL);
+}
+
+static int kona_peri_clk_set_parent(struct clk_hw *hw, u8 index)
+{
+ struct kona_clk *bcm_clk = to_kona_clk(hw);
+ struct peri_clk_data *data = bcm_clk->peri;
+ struct bcm_clk_sel *sel = &data->sel;
+ struct bcm_clk_trig *trig;
+ int ret;
+
+ BUG_ON(index >= sel->parent_count);
+
+ /* If there's only one parent we don't require a selector */
+ if (!selector_exists(sel))
+ return 0;
+
+ /*
+ * The regular trigger is used by default, but if there's a
+ * pre-trigger we want to use that instead.
+ */
+ trig = trigger_exists(&data->pre_trig) ? &data->pre_trig
+ : &data->trig;
+
+ ret = selector_write(bcm_clk->ccu, &data->gate, sel, trig, index);
+ if (ret == -ENXIO) {
+ pr_err("%s: gating failure for %s\n", __func__, bcm_clk->name);
+ ret = -EIO; /* Don't proliferate weird errors */
+ } else if (ret == -EIO) {
+ pr_err("%s: %strigger failed for %s\n", __func__,
+ trig == &data->pre_trig ? "pre-" : "",
+ bcm_clk->name);
+ }
+
+ return ret;
+}
+
+static u8 kona_peri_clk_get_parent(struct clk_hw *hw)
+{
+ struct kona_clk *bcm_clk = to_kona_clk(hw);
+ struct peri_clk_data *data = bcm_clk->peri;
+ u8 index;
+
+ index = selector_read_index(bcm_clk->ccu, &data->sel);
+
+ /* Not all callers would handle an out-of-range value gracefully */
+ return index == BAD_CLK_INDEX ? 0 : index;
+}
+
+static int kona_peri_clk_set_rate(struct clk_hw *hw, unsigned long rate,
+ unsigned long parent_rate)
+{
+ struct kona_clk *bcm_clk = to_kona_clk(hw);
+ struct peri_clk_data *data = bcm_clk->peri;
+ struct bcm_clk_div *div = &data->div;
+ u64 scaled_div = 0;
+ int ret;
+
+ if (parent_rate > (unsigned long)LONG_MAX)
+ return -EINVAL;
+
+ if (rate == __clk_get_rate(hw->clk))
+ return 0;
+
+ if (!divider_exists(div))
+ return rate == parent_rate ? 0 : -EINVAL;
+
+ /*
+ * A fixed divider can't be changed. (Nor can a fixed
+ * pre-divider be, but for now we never actually try to
+ * change that.) Tolerate a request for a no-op change.
+ */
+ if (divider_is_fixed(&data->div))
+ return rate == parent_rate ? 0 : -EINVAL;
+
+ /*
+ * Get the scaled divisor value needed to achieve a clock
+ * rate as close as possible to what was requested, given
+ * the parent clock rate supplied.
+ */
+ (void)round_rate(bcm_clk->ccu, div, &data->pre_div,
+ rate ? rate : 1, parent_rate, &scaled_div);
+
+ /*
+ * We aren't updating any pre-divider at this point, so
+ * we'll use the regular trigger.
+ */
+ ret = divider_write(bcm_clk->ccu, &data->gate, &data->div,
+ &data->trig, scaled_div);
+ if (ret == -ENXIO) {
+ pr_err("%s: gating failure for %s\n", __func__, bcm_clk->name);
+ ret = -EIO; /* Don't proliferate weird errors */
+ } else if (ret == -EIO) {
+ pr_err("%s: trigger failed for %s\n", __func__, bcm_clk->name);
+ }
+
+ return ret;
+}
+
+struct clk_ops kona_peri_clk_ops = {
+ .enable = kona_peri_clk_enable,
+ .disable = kona_peri_clk_disable,
+ .is_enabled = kona_peri_clk_is_enabled,
+ .recalc_rate = kona_peri_clk_recalc_rate,
+ .round_rate = kona_peri_clk_round_rate,
+ .set_parent = kona_peri_clk_set_parent,
+ .get_parent = kona_peri_clk_get_parent,
+ .set_rate = kona_peri_clk_set_rate,
+};
+
+/* Put a peripheral clock into its initial state */
+static bool __peri_clk_init(struct kona_clk *bcm_clk)
+{
+ struct ccu_data *ccu = bcm_clk->ccu;
+ struct peri_clk_data *peri = bcm_clk->peri;
+ const char *name = bcm_clk->name;
+ struct bcm_clk_trig *trig;
+
+ BUG_ON(bcm_clk->type != bcm_clk_peri);
+
+ if (!gate_init(ccu, &peri->gate)) {
+ pr_err("%s: error initializing gate for %s\n", __func__, name);
+ return false;
+ }
+ if (!div_init(ccu, &peri->gate, &peri->div, &peri->trig)) {
+ pr_err("%s: error initializing divider for %s\n", __func__,
+ name);
+ return false;
+ }
+
+ /*
+ * For the pre-divider and selector, the pre-trigger is used
+ * if it's present, otherwise we just use the regular trigger.
+ */
+ trig = trigger_exists(&peri->pre_trig) ? &peri->pre_trig
+ : &peri->trig;
+
+ if (!div_init(ccu, &peri->gate, &peri->pre_div, trig)) {
+ pr_err("%s: error initializing pre-divider for %s\n", __func__,
+ name);
+ return false;
+ }
+
+ if (!sel_init(ccu, &peri->gate, &peri->sel, trig)) {
+ pr_err("%s: error initializing selector for %s\n", __func__,
+ name);
+ return false;
+ }
+
+ return true;
+}
+
+static bool __kona_clk_init(struct kona_clk *bcm_clk)
+{
+ switch (bcm_clk->type) {
+ case bcm_clk_peri:
+ return __peri_clk_init(bcm_clk);
+ default:
+ BUG();
+ }
+ return -EINVAL;
+}
+
+/* Set a CCU and all its clocks into their desired initial state */
+bool __init kona_ccu_init(struct ccu_data *ccu)
+{
+ unsigned long flags;
+ unsigned int which;
+ struct clk **clks = ccu->data.clks;
+ bool success = true;
+
+ flags = ccu_lock(ccu);
+ __ccu_write_enable(ccu);
+
+ for (which = 0; which < ccu->data.clk_num; which++) {
+ struct kona_clk *bcm_clk;
+
+ if (!clks[which])
+ continue;
+ bcm_clk = to_kona_clk(__clk_get_hw(clks[which]));
+ success &= __kona_clk_init(bcm_clk);
+ }
+
+ __ccu_write_disable(ccu);
+ ccu_unlock(ccu, flags);
+ return success;
+}
diff --git a/drivers/clk/bcm/clk-kona.h b/drivers/clk/bcm/clk-kona.h
new file mode 100644
index 0000000..5e139ad
--- /dev/null
+++ b/drivers/clk/bcm/clk-kona.h
@@ -0,0 +1,410 @@
+/*
+ * Copyright (C) 2013 Broadcom Corporation
+ * Copyright 2013 Linaro Limited
+ *
+ * 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 version 2.
+ *
+ * This program is distributed "as is" WITHOUT ANY WARRANTY of any
+ * kind, whether express or implied; without even the implied warranty
+ * of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ */
+
+#ifndef _CLK_KONA_H
+#define _CLK_KONA_H
+
+#include <linux/kernel.h>
+#include <linux/list.h>
+#include <linux/spinlock.h>
+#include <linux/slab.h>
+#include <linux/device.h>
+#include <linux/of.h>
+#include <linux/clk-provider.h>
+
+#define BILLION 1000000000
+
+/* The common clock framework uses u8 to represent a parent index */
+#define PARENT_COUNT_MAX ((u32)U8_MAX)
+
+#define BAD_CLK_INDEX U8_MAX /* Can't ever be valid */
+#define BAD_CLK_NAME ((const char *)-1)
+
+#define BAD_SCALED_DIV_VALUE U64_MAX
+
+/*
+ * Utility macros for object flag management. If possible, flags
+ * should be defined such that 0 is the desired default value.
+ */
+#define FLAG(type, flag) BCM_CLK_ ## type ## _FLAGS_ ## flag
+#define FLAG_SET(obj, type, flag) ((obj)->flags |= FLAG(type, flag))
+#define FLAG_CLEAR(obj, type, flag) ((obj)->flags &= ~(FLAG(type, flag)))
+#define FLAG_FLIP(obj, type, flag) ((obj)->flags ^= FLAG(type, flag))
+#define FLAG_TEST(obj, type, flag) (!!((obj)->flags & FLAG(type, flag)))
+
+/* Clock field state tests */
+
+#define gate_exists(gate) FLAG_TEST(gate, GATE, EXISTS)
+#define gate_is_enabled(gate) FLAG_TEST(gate, GATE, ENABLED)
+#define gate_is_hw_controllable(gate) FLAG_TEST(gate, GATE, HW)
+#define gate_is_sw_controllable(gate) FLAG_TEST(gate, GATE, SW)
+#define gate_is_sw_managed(gate) FLAG_TEST(gate, GATE, SW_MANAGED)
+#define gate_is_no_disable(gate) FLAG_TEST(gate, GATE, NO_DISABLE)
+
+#define gate_flip_enabled(gate) FLAG_FLIP(gate, GATE, ENABLED)
+
+#define divider_exists(div) FLAG_TEST(div, DIV, EXISTS)
+#define divider_is_fixed(div) FLAG_TEST(div, DIV, FIXED)
+#define divider_has_fraction(div) (!divider_is_fixed(div) && \
+ (div)->frac_width > 0)
+
+#define selector_exists(sel) ((sel)->width != 0)
+#define trigger_exists(trig) FLAG_TEST(trig, TRIG, EXISTS)
+
+/* Clock type, used to tell common block what it's part of */
+enum bcm_clk_type {
+ bcm_clk_none, /* undefined clock type */
+ bcm_clk_bus,
+ bcm_clk_core,
+ bcm_clk_peri
+};
+
+/*
+ * Each CCU defines a mapped area of memory containing registers
+ * used to manage clocks implemented by the CCU. Access to memory
+ * within the CCU's space is serialized by a spinlock. Before any
+ * (other) address can be written, a special access "password" value
+ * must be written to its WR_ACCESS register (located at the base
+ * address of the range). We keep track of the name of each CCU as
+ * it is set up, and maintain them in a list.
+ */
+struct ccu_data {
+ void __iomem *base; /* base of mapped address space */
+ spinlock_t lock; /* serialization lock */
+ bool write_enabled; /* write access is currently enabled */
+ struct list_head links; /* for ccu_list */
+ struct device_node *node;
+ struct clk_onecell_data data;
+ const char *name;
+ u32 range; /* byte range of address space */
+};
+
+/*
+ * Gating control and status is managed by a 32-bit gate register.
+ *
+ * There are several types of gating available:
+ * - (no gate)
+ * A clock with no gate is assumed to be always enabled.
+ * - hardware-only gating (auto-gating)
+ * Enabling or disabling clocks with this type of gate is
+ * managed automatically by the hardware. Such clocks can be
+ * considered by the software to be enabled. The current status
+ * of auto-gated clocks can be read from the gate status bit.
+ * - software-only gating
+ * Auto-gating is not available for this type of clock.
+ * Instead, software manages whether it's enabled by setting or
+ * clearing the enable bit. The current gate status of a gate
+ * under software control can be read from the gate status bit.
+ * To ensure a change to the gating status is complete, the
+ * status bit can be polled to verify that the gate has entered
+ * the desired state.
+ * - selectable hardware or software gating
+ * Gating for this type of clock can be configured to be either
+ * under software or hardware control. Which type is in use is
+ * determined by the hw_sw_sel bit of the gate register.
+ */
+struct bcm_clk_gate {
+ u32 offset; /* gate register offset */
+ u32 status_bit; /* 0: gate is disabled; 0: gatge is enabled */
+ u32 en_bit; /* 0: disable; 1: enable */
+ u32 hw_sw_sel_bit; /* 0: hardware gating; 1: software gating */
+ u32 flags; /* BCM_CLK_GATE_FLAGS_* below */
+};
+
+/*
+ * Gate flags:
+ * HW means this gate can be auto-gated
+ * SW means the state of this gate can be software controlled
+ * NO_DISABLE means this gate is (only) enabled if under software control
+ * SW_MANAGED means the status of this gate is under software control
+ * ENABLED means this software-managed gate is *supposed* to be enabled
+ */
+#define BCM_CLK_GATE_FLAGS_EXISTS ((u32)1 << 0) /* Gate is valid */
+#define BCM_CLK_GATE_FLAGS_HW ((u32)1 << 1) /* Can auto-gate */
+#define BCM_CLK_GATE_FLAGS_SW ((u32)1 << 2) /* Software control */
+#define BCM_CLK_GATE_FLAGS_NO_DISABLE ((u32)1 << 3) /* HW or enabled */
+#define BCM_CLK_GATE_FLAGS_SW_MANAGED ((u32)1 << 4) /* SW now in control */
+#define BCM_CLK_GATE_FLAGS_ENABLED ((u32)1 << 5) /* If SW_MANAGED */
+
+/*
+ * Gate initialization macros.
+ *
+ * Any gate initially under software control will be enabled.
+ */
+
+/* A hardware/software gate initially under software control */
+#define HW_SW_GATE(_offset, _status_bit, _en_bit, _hw_sw_sel_bit) \
+ { \
+ .offset = (_offset), \
+ .status_bit = (_status_bit), \
+ .en_bit = (_en_bit), \
+ .hw_sw_sel_bit = (_hw_sw_sel_bit), \
+ .flags = FLAG(GATE, HW)|FLAG(GATE, SW)| \
+ FLAG(GATE, SW_MANAGED)|FLAG(GATE, ENABLED)| \
+ FLAG(GATE, EXISTS), \
+ }
+
+/* A hardware/software gate initially under hardware control */
+#define HW_SW_GATE_AUTO(_offset, _status_bit, _en_bit, _hw_sw_sel_bit) \
+ { \
+ .offset = (_offset), \
+ .status_bit = (_status_bit), \
+ .en_bit = (_en_bit), \
+ .hw_sw_sel_bit = (_hw_sw_sel_bit), \
+ .flags = FLAG(GATE, HW)|FLAG(GATE, SW)| \
+ FLAG(GATE, EXISTS), \
+ }
+
+/* A hardware-or-enabled gate (enabled if not under hardware control) */
+#define HW_ENABLE_GATE(_offset, _status_bit, _en_bit, _hw_sw_sel_bit) \
+ { \
+ .offset = (_offset), \
+ .status_bit = (_status_bit), \
+ .en_bit = (_en_bit), \
+ .hw_sw_sel_bit = (_hw_sw_sel_bit), \
+ .flags = FLAG(GATE, HW)|FLAG(GATE, SW)| \
+ FLAG(GATE, NO_DISABLE)|FLAG(GATE, EXISTS), \
+ }
+
+/* A software-only gate */
+#define SW_ONLY_GATE(_offset, _status_bit, _en_bit) \
+ { \
+ .offset = (_offset), \
+ .status_bit = (_status_bit), \
+ .en_bit = (_en_bit), \
+ .flags = FLAG(GATE, SW)|FLAG(GATE, SW_MANAGED)| \
+ FLAG(GATE, ENABLED)|FLAG(GATE, EXISTS), \
+ }
+
+/* A hardware-only gate */
+#define HW_ONLY_GATE(_offset, _status_bit) \
+ { \
+ .offset = (_offset), \
+ .status_bit = (_status_bit), \
+ .flags = FLAG(GATE, HW)|FLAG(GATE, EXISTS), \
+ }
+
+/*
+ * Each clock can have zero, one, or two dividers which change the
+ * output rate of the clock. Each divider can be either fixed or
+ * variable. If there are two dividers, they are the "pre-divider"
+ * and the "regular" or "downstream" divider. If there is only one,
+ * there is no pre-divider.
+ *
+ * A fixed divider is any non-zero (positive) value, and it
+ * indicates how the input rate is affected by the divider.
+ *
+ * The value of a variable divider is maintained in a sub-field of a
+ * 32-bit divider register. The position of the field in the
+ * register is defined by its offset and width. The value recorded
+ * in this field is always 1 less than the value it represents.
+ *
+ * In addition, a variable divider can indicate that some subset
+ * of its bits represent a "fractional" part of the divider. Such
+ * bits comprise the low-order portion of the divider field, and can
+ * be viewed as representing the portion of the divider that lies to
+ * the right of the decimal point. Most variable dividers have zero
+ * fractional bits. Variable dividers with non-zero fraction width
+ * still record a value 1 less than the value they represent; the
+ * added 1 does *not* affect the low-order bit in this case, it
+ * affects the bits above the fractional part only. (Often in this
+ * code a divider field value is distinguished from the value it
+ * represents by referring to the latter as a "divisor".)
+ *
+ * In order to avoid dealing with fractions, divider arithmetic is
+ * performed using "scaled" values. A scaled value is one that's
+ * been left-shifted by the fractional width of a divider. Dividing
+ * a scaled value by a scaled divisor produces the desired quotient
+ * without loss of precision and without any other special handling
+ * for fractions.
+ *
+ * The recorded value of a variable divider can be modified. To
+ * modify either divider (or both), a clock must be enabled (i.e.,
+ * using its gate). In addition, a trigger register (described
+ * below) must be used to commit the change, and polled to verify
+ * the change is complete.
+ */
+struct bcm_clk_div {
+ union {
+ struct { /* variable divider */
+ u32 offset; /* divider register offset */
+ u32 shift; /* field shift */
+ u32 width; /* field width */
+ u32 frac_width; /* field fraction width */
+
+ u64 scaled_div; /* scaled divider value */
+ };
+ u32 fixed; /* non-zero fixed divider value */
+ };
+ u32 flags; /* BCM_CLK_DIV_FLAGS_* below */
+};
+
+/*
+ * Divider flags:
+ * EXISTS means this divider exists
+ * FIXED means it is a fixed-rate divider
+ */
+#define BCM_CLK_DIV_FLAGS_EXISTS ((u32)1 << 0) /* Divider is valid */
+#define BCM_CLK_DIV_FLAGS_FIXED ((u32)1 << 1) /* Fixed-value */
+
+/* Divider initialization macros */
+
+/* A fixed (non-zero) divider */
+#define FIXED_DIVIDER(_value) \
+ { \
+ .fixed = (_value), \
+ .flags = FLAG(DIV, EXISTS)|FLAG(DIV, FIXED), \
+ }
+
+/* A divider with an integral divisor */
+#define DIVIDER(_offset, _shift, _width) \
+ { \
+ .offset = (_offset), \
+ .shift = (_shift), \
+ .width = (_width), \
+ .scaled_div = BAD_SCALED_DIV_VALUE, \
+ .flags = FLAG(DIV, EXISTS), \
+ }
+
+/* A divider whose divisor has an integer and fractional part */
+#define FRAC_DIVIDER(_offset, _shift, _width, _frac_width) \
+ { \
+ .offset = (_offset), \
+ .shift = (_shift), \
+ .width = (_width), \
+ .frac_width = (_frac_width), \
+ .scaled_div = BAD_SCALED_DIV_VALUE, \
+ .flags = FLAG(DIV, EXISTS), \
+ }
+
+/*
+ * Clocks may have multiple "parent" clocks. If there is more than
+ * one, a selector must be specified to define which of the parent
+ * clocks is currently in use. The selected clock is indicated in a
+ * sub-field of a 32-bit selector register. The range of
+ * representable selector values typically exceeds the number of
+ * available parent clocks. Occasionally the reset value of a
+ * selector field is explicitly set to a (specific) value that does
+ * not correspond to a defined input clock.
+ *
+ * We register all known parent clocks with the common clock code
+ * using a packed array (i.e., no empty slots) of (parent) clock
+ * names, and refer to them later using indexes into that array.
+ * We maintain an array of selector values indexed by common clock
+ * index values in order to map between these common clock indexes
+ * and the selector values used by the hardware.
+ *
+ * Like dividers, a selector can be modified, but to do so a clock
+ * must be enabled, and a trigger must be used to commit the change.
+ */
+struct bcm_clk_sel {
+ u32 offset; /* selector register offset */
+ u32 shift; /* field shift */
+ u32 width; /* field width */
+
+ u32 parent_count; /* number of entries in parent_sel[] */
+ u32 *parent_sel; /* array of parent selector values */
+ u8 clk_index; /* current selected index in parent_sel[] */
+};
+
+/* Selector initialization macro */
+#define SELECTOR(_offset, _shift, _width) \
+ { \
+ .offset = (_offset), \
+ .shift = (_shift), \
+ .width = (_width), \
+ .clk_index = BAD_CLK_INDEX, \
+ }
+
+/*
+ * Making changes to a variable divider or a selector for a clock
+ * requires the use of a trigger. A trigger is defined by a single
+ * bit within a register. To signal a change, a 1 is written into
+ * that bit. To determine when the change has been completed, that
+ * trigger bit is polled; the read value will be 1 while the change
+ * is in progress, and 0 when it is complete.
+ *
+ * Occasionally a clock will have more than one trigger. In this
+ * case, the "pre-trigger" will be used when changing a clock's
+ * selector and/or its pre-divider.
+ */
+struct bcm_clk_trig {
+ u32 offset; /* trigger register offset */
+ u32 bit; /* trigger bit */
+ u32 flags; /* BCM_CLK_TRIG_FLAGS_* below */
+};
+
+/*
+ * Trigger flags:
+ * EXISTS means this trigger exists
+ */
+#define BCM_CLK_TRIG_FLAGS_EXISTS ((u32)1 << 0) /* Trigger is valid */
+
+/* Trigger initialization macro */
+#define TRIGGER(_offset, _bit) \
+ { \
+ .offset = (_offset), \
+ .bit = (_bit), \
+ .flags = FLAG(TRIG, EXISTS), \
+ }
+
+struct peri_clk_data {
+ struct bcm_clk_gate gate;
+ struct bcm_clk_trig pre_trig;
+ struct bcm_clk_div pre_div;
+ struct bcm_clk_trig trig;
+ struct bcm_clk_div div;
+ struct bcm_clk_sel sel;
+ const char *clocks[]; /* must be last; use CLOCKS() to declare */
+};
+#define CLOCKS(...) { __VA_ARGS__, NULL, }
+#define NO_CLOCKS { NULL, } /* Must use of no parent clocks */
+
+struct kona_clk {
+ struct clk_hw hw;
+ struct clk_init_data init_data;
+ const char *name; /* name of this clock */
+ struct ccu_data *ccu; /* ccu this clock is associated with */
+ enum bcm_clk_type type;
+ union {
+ void *data;
+ struct peri_clk_data *peri;
+ };
+};
+#define to_kona_clk(_hw) \
+ container_of(_hw, struct kona_clk, hw)
+
+/* Exported globals */
+
+extern struct clk_ops kona_peri_clk_ops;
+
+/* Help functions */
+
+#define PERI_CLK_SETUP(clks, ccu, id, name) \
+ clks[id] = kona_clk_setup(ccu, #name, bcm_clk_peri, &name ## _data)
+
+/* Externally visible functions */
+
+extern u64 do_div_round_closest(u64 dividend, unsigned long divisor);
+extern u64 scaled_div_max(struct bcm_clk_div *div);
+extern u64 scaled_div_build(struct bcm_clk_div *div, u32 div_value,
+ u32 billionths);
+
+extern struct clk *kona_clk_setup(struct ccu_data *ccu, const char *name,
+ enum bcm_clk_type type, void *data);
+extern void __init kona_dt_ccu_setup(struct device_node *node,
+ int (*ccu_clks_setup)(struct ccu_data *));
+extern bool __init kona_ccu_init(struct ccu_data *ccu);
+
+#endif /* _CLK_KONA_H */
diff --git a/include/dt-bindings/clock/bcm281xx.h b/include/dt-bindings/clock/bcm281xx.h
new file mode 100644
index 0000000..e009694
--- /dev/null
+++ b/include/dt-bindings/clock/bcm281xx.h
@@ -0,0 +1,65 @@
+/*
+ * Copyright (C) 2013 Broadcom Corporation
+ * Copyright 2013 Linaro Limited
+ *
+ * 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 version 2.
+ *
+ * This program is distributed "as is" WITHOUT ANY WARRANTY of any
+ * kind, whether express or implied; without even the implied warranty
+ * of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ */
+
+#ifndef _CLOCK_BCM281XX_H
+#define _CLOCK_BCM281XX_H
+
+/*
+ * This file defines the values used to specify clocks provided by
+ * the clock control units (CCUs) on Broadcom BCM281XX family SoCs.
+ */
+
+/* root CCU clock ids */
+
+#define BCM281XX_ROOT_CCU_FRAC_1M 0
+#define BCM281XX_ROOT_CCU_CLOCK_COUNT 1
+
+/* aon CCU clock ids */
+
+#define BCM281XX_AON_CCU_HUB_TIMER 0
+#define BCM281XX_AON_CCU_PMU_BSC 1
+#define BCM281XX_AON_CCU_PMU_BSC_VAR 2
+#define BCM281XX_AON_CCU_CLOCK_COUNT 3
+
+/* hub CCU clock ids */
+
+#define BCM281XX_HUB_CCU_TMON_1M 0
+#define BCM281XX_HUB_CCU_CLOCK_COUNT 1
+
+/* master CCU clock ids */
+
+#define BCM281XX_MASTER_CCU_SDIO1 0
+#define BCM281XX_MASTER_CCU_SDIO2 1
+#define BCM281XX_MASTER_CCU_SDIO3 2
+#define BCM281XX_MASTER_CCU_SDIO4 3
+#define BCM281XX_MASTER_CCU_USB_IC 4
+#define BCM281XX_MASTER_CCU_HSIC2_48M 5
+#define BCM281XX_MASTER_CCU_HSIC2_12M 6
+#define BCM281XX_MASTER_CCU_CLOCK_COUNT 7
+
+/* slave CCU clock ids */
+
+#define BCM281XX_SLAVE_CCU_UARTB 0
+#define BCM281XX_SLAVE_CCU_UARTB2 1
+#define BCM281XX_SLAVE_CCU_UARTB3 2
+#define BCM281XX_SLAVE_CCU_UARTB4 3
+#define BCM281XX_SLAVE_CCU_SSP0 4
+#define BCM281XX_SLAVE_CCU_SSP2 5
+#define BCM281XX_SLAVE_CCU_BSC1 6
+#define BCM281XX_SLAVE_CCU_BSC2 7
+#define BCM281XX_SLAVE_CCU_BSC3 8
+#define BCM281XX_SLAVE_CCU_PWM 9
+#define BCM281XX_SLAVE_CCU_CLOCK_COUNT 10
+
+#endif /* _CLOCK_BCM281XX_H */
--
1.7.9.5
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply related
* Re: [PATCH 1/2] mfd: twl4030-madc: Add devicetree support.
From: Sebastian Reichel @ 2014-02-14 17:40 UTC (permalink / raw)
To: Marek Belisko
Cc: robh+dt, pawel.moll, mark.rutland, ijc+devicetree, galak, rob,
linux, grant.likely, devicetree, linux-doc, linux-kernel,
linux-arm-kernel, hns
In-Reply-To: <1392383861-14169-1-git-send-email-marek@goldelico.com>
[-- Attachment #1: Type: text/plain, Size: 245 bytes --]
Hi Marek,
I have prepared a patchset, which adds DT bindings for twl4030-madc
using the standard IIO DT API. I have not yet send the patchset,
since I have not yet found the time to test the patchset. I will
send them as RFC now.
-- Sebastian
[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 819 bytes --]
^ permalink raw reply
* Re: [PATCH net-next v5 09/10] Documentation: add Device tree bindings for Broadcom GENET
From: Florian Fainelli @ 2014-02-14 17:36 UTC (permalink / raw)
To: Mark Rutland
Cc: netdev@vger.kernel.org, davem@davemloft.net,
devicetree@vger.kernel.org, cernekee@gmail.com,
romieu@fr.zoreil.com
In-Reply-To: <20140214101942.GA9907@e106331-lin.cambridge.arm.com>
2014-02-14 2:19 GMT-08:00 Mark Rutland <mark.rutland@arm.com>:
> On Fri, Feb 14, 2014 at 12:08:50AM +0000, Florian Fainelli wrote:
>> This patch adds the Device Tree bindings for the Broadcom GENET Gigabit
>> Ethernet controller. A bunch of examples are provided to illustrate the
>> versatile aspect of the hardare.
>>
>> Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
>> ---
>> Changes in v5:
>> - respin due to the CONFIG_OF dependency fix
>>
>> Changes in v4:
>> - respin due to the Kconfig dependency
>>
>> Changes in v3
>> - improve compatible property description to specify all compatible strings
>> supported
>> - move MDIO bus node to a separate section, to separate from the properties
>> - detail which exact phy-mode strings are supported and that this refers to
>> the ePAPR 'phy-connection-type' modes
>> - fixed MDIO bus node compatible string to match what is really provided by
>> the bootloader and use the same details as for the GENET compatible string
>> - add address-cells and size-cells required properties for the GENET device
>> tree node
>> - add a better description for the 'fixed-link' property with reference to
>> existing binding documentation
>> - add a reference to the Ethernet PHY device tree binding
>> - add a description for the interrupts properties
>> - add a documentation for the optional clocks properties
>> - removed unused and deprecated device_type properties
>>
>> Changes in v2:
>> - rebased against net-next/master
>>
>> .../devicetree/bindings/net/broadcom-bcmgenet.txt | 121 +++++++++++++++++++++
>> 1 file changed, 121 insertions(+)
>> create mode 100644 Documentation/devicetree/bindings/net/broadcom-bcmgenet.txt
>>
>> diff --git a/Documentation/devicetree/bindings/net/broadcom-bcmgenet.txt b/Documentation/devicetree/bindings/net/broadcom-bcmgenet.txt
>> new file mode 100644
>> index 0000000..afd31f9
>> --- /dev/null
>> +++ b/Documentation/devicetree/bindings/net/broadcom-bcmgenet.txt
>> @@ -0,0 +1,121 @@
>> +* Broadcom BCM7xxx Ethernet Controller (GENET)
>> +
>> +Required properties:
>> +- compatible: should contain one of "brcm,genet-v1", "brcm,genet-v2",
>> + "brcm,genet-v3", "brcm,genet-v4".
>> +- reg: address and length of the register set for the device
>> +- interrupts: must be two cells, the first cell is the general purpose
>> + interrupt line, while the second cell is the interrupt for the ring
>> + RX and TX queues operating in ring mode
>> +- phy-mode: String, operation mode of the PHY interface. Supported values are
>> + "mii", "rgmii", "rgmii-txid", "rev-mii", "moca". Analogous to ePAPR
>> + "phy-connection-type" values
>> +- address-cells: should be 1
>> +- size-cells: should be 1
>> +
>> +Optional properties:
>> +- clocks: When provided, must be two cells, first one is the main GENET clock
>> + and the second cell is the Genet Wake-on-LAN clock.
>
> Clocks aren't just cells. They are referred to with phandle +
> clock-specificer pairs.
>
> This doesn't document the names that the driver requests the clocks by.
> Just listing the clocks isn't sufficient, as any dts will have to add a
> clock-names property to get those clocks probed.
I will follow up with a fix for this, thanks!
--
Florian
^ permalink raw reply
* Re: How to select between different display timings?
From: Dirk Behme @ 2014-02-14 17:35 UTC (permalink / raw)
To: Lothar Waßmann, Shawn Guo
Cc: Dirk Behme, devicetree-u79uwXL29TY76Z2rM5mHXA@public.gmane.org,
Rob Herring,
linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r@public.gmane.org
In-Reply-To: <20140214150208.12417b11-VjFSrY7JcPWvSplVBqRQBQ@public.gmane.org>
Am 14.02.2014 15:02, schrieb Lothar Waßmann:
> Hi,
>
> Dirk Behme wrote:
>> Hi Lothar and Shawn,
>>
>> On 08.08.2013 14:51, Lothar Waßmann wrote:
>>> - add Copyright header
>>> - use label references for better readability
>>> - sort the entries alphabetically
>>> - add some aliases used by U-Boot to edit the DT data
>>>
>>> Signed-off-by: Lothar Waßmann <LW-bxm8fMRDkQLDiMYJYoSAnRvVK+yQ3ZXh@public.gmane.org>
>>> ---
>>> arch/arm/boot/dts/imx28-tx28.dts | 693 +++++++++++++++++++++++++++++++++-----
>>> 1 files changed, 611 insertions(+), 82 deletions(-)
>>>
>>> diff --git a/arch/arm/boot/dts/imx28-tx28.dts b/arch/arm/boot/dts/imx28-tx28.dts
>>> index 37be532..866af60 100644
>>> --- a/arch/arm/boot/dts/imx28-tx28.dts
>>> +++ b/arch/arm/boot/dts/imx28-tx28.dts
>> ...
>>> +&lcdif {
>>> + pinctrl-names = "default";
>>> + pinctrl-0 = <&lcdif_24bit_pins_a &lcdif_sync_pins_a &lcdif_ctrl_pins_a>;
>>> + lcd-supply = <®_lcd>;
>>> + display = <&display>;
>>> + status = "okay";
>>> +
>>> + display: display@0 {
>>> + bits-per-pixel = <32>;
>>> + bus-width = <24>;
>>> +
>>> + display-timings {
>>> + native-mode = <&timing5>;
> ^^^^^^^^^^^^^^^^^^^^^^^^^
>>> + timing0: timing0 {
> [...]
>>> + };
>>> + };
>>> + };
>>> +};
>>
>> Being no graphics expert, looking at above device tree change, I'd like
>> to understand how this can be used to switch between different display
>> timings?
>>
>> In the kernel, I've found the code which selects the default/native-mode
>> (of_display_timing.c). But, as here, if there is a native/default mode,
>> there are alternative modes. I haven't understood how and who to select
>> these other modes? In this case, how could the alternative modes
>> timing0/1/2/3/4 be selected in the kernel?
>>
>> Do you have any pointers regarding this?
>>
> You need to set the property 'native-mode' to the phandle of the
> display-timings entry like marked above.
Hmm, let's see if we talk about the same thing ;) Let me rephrase my
questions:
Using above example and booting a system with the resulting .dtb,
timing5 is selected at boot time, correct?
Assuming I want to use timing1 instead: who would set when the
property 'native-mode' to &timing1?
Can the switch to
native-mode = <&timing1>;
only be done at compile time of the dts -> dtb? Or can this be done at
boot/kernel run time, too?
In the end, I'm looking for a way to have several timings like above
in the device tree, and select one at kernel boot time based on the
display detected. Would this be possible?
Many thanks and best regards
Dirk
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply
* Re: [PATCH 1/2] clk: fixed-rate: use full DT node name
From: Stephen Warren @ 2014-02-14 16:43 UTC (permalink / raw)
To: Mark Rutland
Cc: Mike Turquette,
linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r@public.gmane.org,
devicetree-u79uwXL29TY76Z2rM5mHXA@public.gmane.org,
linux-rpi-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r@public.gmane.org
In-Reply-To: <20140214103556.GC9907-NuALmloUBlrZROr8t4l/smS4ubULX0JqMm0uRHvK7Nw@public.gmane.org>
On 02/14/2014 03:35 AM, Mark Rutland wrote:
> On Fri, Feb 14, 2014 at 06:16:52AM +0000, Stephen Warren wrote:
>> clk-fixed-rate currently names clocks according to a node's name without
>> the unit address. When faced with the legal and technically correct DT
>> structure below, this causes rgistration attempts for 3 clocks with the
>> same name, 2 of which fail.
>>
>> clocks {
>> compatible = "simple-bus";
>> #address-cells = <1>;
>> #size-cells = <0>;
>>
>> clk_mmc: clock@0 {
>> compatible = "fixed-clock";
>> reg = <0>;
>> ...
>> clk_i2c: clock@1 {
>> compatible = "fixed-clock";
>> reg = <1>;
>> ...
>> clk_spi: clock@2 {
>> compatible = "fixed-clock";
>> reg = <2>;
>> ...
>
> I'd argue that this case isn't valid.
>
> The fixed-clock binding doesn't define a reg, yet simple bus binding
> implies that the reg property of child nodes should be interpretted as
> the same address space as their parent (MMIO in this case?). The
> fixed-clock nodes reg proeprties clearly aren't MMIO addresses.
>
> Additionally, the _requred_ ranges property is missing.
Oh, IIRC that was deliberate to indicate that the child address space
was disjoint from the parent address space.
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply
* Re: [PATCH 1/2] clk: fixed-rate: use full DT node name
From: Stephen Warren @ 2014-02-14 16:43 UTC (permalink / raw)
To: Mark Rutland
Cc: Mike Turquette,
linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r@public.gmane.org,
devicetree-u79uwXL29TY76Z2rM5mHXA@public.gmane.org,
linux-rpi-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r@public.gmane.org
In-Reply-To: <20140214103556.GC9907-NuALmloUBlrZROr8t4l/smS4ubULX0JqMm0uRHvK7Nw@public.gmane.org>
On 02/14/2014 03:35 AM, Mark Rutland wrote:
> On Fri, Feb 14, 2014 at 06:16:52AM +0000, Stephen Warren wrote:
>> clk-fixed-rate currently names clocks according to a node's name without
>> the unit address. When faced with the legal and technically correct DT
>> structure below, this causes rgistration attempts for 3 clocks with the
>> same name, 2 of which fail.
>>
>> clocks {
>> compatible = "simple-bus";
>> #address-cells = <1>;
>> #size-cells = <0>;
>>
>> clk_mmc: clock@0 {
>> compatible = "fixed-clock";
>> reg = <0>;
>> ...
>> clk_i2c: clock@1 {
>> compatible = "fixed-clock";
>> reg = <1>;
>> ...
>> clk_spi: clock@2 {
>> compatible = "fixed-clock";
>> reg = <2>;
>> ...
>
> I'd argue that this case isn't valid.
Well, it's very widely used, and was the result of numerous discussions
of how this kind of thing should be represented:-/
> The fixed-clock binding doesn't define a reg, yet simple bus binding
> implies that the reg property of child nodes should be interpretted as
> the same address space as their parent (MMIO in this case?). The
> fixed-clock nodes reg proeprties clearly aren't MMIO addresses.
>
> Additionally, the _requred_ ranges property is missing.
Perhaps we need to invent a simple-container instead then?
> It's just nonsensical; rename them to clock_{0,1,..} instead and get rid
> of the reg properties. Then they're named uniquely.
That's not legal either. DT node names are supposed to represent the
type of device/object (i.e. just "clock"), not the identity of the
device/object (i.e. not include IDs etc.). Hence, the node name needs to
be "clock" for all of them, and the unit address must be used to
differentiate them.
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply
* Re: [PATCH V2 1/2] mmc: omap_hsmmc: Add support for quirky omap3 hsmmc controller
From: Tony Lindgren @ 2014-02-14 16:31 UTC (permalink / raw)
To: Nishanth Menon
Cc: Balaji T K, Chris Ball, linux-mmc, linux-omap, linux-kernel,
linux-doc, linux-arm-kernel, devicetree, Felipe Balbi
In-Reply-To: <1392356749-32091-2-git-send-email-nm@ti.com>
* Nishanth Menon <nm@ti.com> [140213 21:48]:
> When device is booted using devicetree, platforms impacted by Erratum
> 2.1.1.128 is not detected easily in the mmc driver. This erratum
> indicates that the module cannot do multi-block transfers. Platforms
> such as LDP which use OMAP3 ES revision prior to ES3.0 are impacted by
> this.
>
> Provide a new compatible property "ti,omap3-pre-es3-hsmmc" to allow
> driver to determine if driver needs to implement quirks associated
> with the specific module version (primarily because the IP revision
> information is not sufficient for the same).
>
> Signed-off-by: Nishanth Menon <nm@ti.com>
Looks good to me, would be good to get this into the -rc cycle
as LDP is one of the automated boot testing boards for us.
If you guys want me to pick up both of the patches let me know,
otherwise feel free to add my ack.
Regards,
Tony
^ permalink raw reply
* Re: [PATCH 2/2] of: search the best compatible match first in __of_match_node()
From: Kumar Gala @ 2014-02-14 16:23 UTC (permalink / raw)
To: Rob Herring
Cc: Kevin Hao, devicetree-u79uwXL29TY76Z2rM5mHXA@public.gmane.org,
Stephen N Chivers, Rob Herring, Grant Likely, linuxppc-dev,
Sebastian Hesselbarth
In-Reply-To: <CAL_JsqKq4_1K8EF+PZoP=0=H6tiRxbgdzs9UHHVdbHS014n74Q-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
On Feb 14, 2014, at 9:53 AM, Rob Herring <robherring2-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
> On Thu, Feb 13, 2014 at 11:22 PM, Kevin Hao <haokexin-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
>> Currently, of_match_node compares each given match against all node's
>> compatible strings with of_device_is_compatible.
>>
>> To achieve multiple compatible strings per node with ordering from
>> specific to generic, this requires given matches to be ordered from
>> specific to generic. For most of the drivers this is not true and also
>> an alphabetical ordering is more sane there.
>>
>> Therefore, this patch introduces a function to match each of the node's
>> compatible strings against all given compatible matches without type and
>> name first, before checking the next compatible string. This implies
>> that node's compatibles are ordered from specific to generic while
>> given matches can be in any order. If we fail to find such a match
>> entry, then fall-back to the old method in order to keep compatibility.
>>
>> Cc: Sebastian Hesselbarth <sebastian.hesselbarth-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
>> Signed-off-by: Kevin Hao <haokexin-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
>
> Looks good to me. I'll put this in next for a few days. I'd really
> like to see some acks and tested-by's before sending to Linus.
>
> We could be a bit more strict here and fallback to the old matching if
> the match table has any entries with name or type. I don't think that
> should be necessary though.
>
> Rob
>
Can you push the revert to Linus sooner, since currently a ton of boards wouldn’t be working on the PPC side, so at least -rc3 has the possibility of working for them.
- k
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply
* Re: [PATCH 2/2] of: search the best compatible match first in __of_match_node()
From: Kumar Gala @ 2014-02-14 16:19 UTC (permalink / raw)
To: Rob Herring
Cc: Kevin Hao, devicetree-u79uwXL29TY76Z2rM5mHXA@public.gmane.org,
linuxppc-dev, Sebastian Hesselbarth, Stephen N Chivers,
Grant Likely, Rob Herring
In-Reply-To: <CAL_JsqKq4_1K8EF+PZoP=0=H6tiRxbgdzs9UHHVdbHS014n74Q-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
On Feb 14, 2014, at 9:53 AM, Rob Herring <robherring2-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
> On Thu, Feb 13, 2014 at 11:22 PM, Kevin Hao <haokexin-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
>> Currently, of_match_node compares each given match against all node's
>> compatible strings with of_device_is_compatible.
>>
>> To achieve multiple compatible strings per node with ordering from
>> specific to generic, this requires given matches to be ordered from
>> specific to generic. For most of the drivers this is not true and also
>> an alphabetical ordering is more sane there.
>>
>> Therefore, this patch introduces a function to match each of the node's
>> compatible strings against all given compatible matches without type and
>> name first, before checking the next compatible string. This implies
>> that node's compatibles are ordered from specific to generic while
>> given matches can be in any order. If we fail to find such a match
>> entry, then fall-back to the old method in order to keep compatibility.
>>
>> Cc: Sebastian Hesselbarth <sebastian.hesselbarth-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
>> Signed-off-by: Kevin Hao <haokexin-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
>
> Looks good to me. I'll put this in next for a few days. I'd really
> like to see some acks and tested-by's before sending to Linus.
>
> We could be a bit more strict here and fallback to the old matching if
> the match table has any entries with name or type. I don't think that
> should be necessary though.
>
> Rob
Can you push the revert to Linus sooner, since currently a ton of boards wouldn’t be working on the PPC side, so at least -rc3 has the possibility of working for them.
- k
>
>> ---
>> drivers/of/base.c | 43 ++++++++++++++++++++++++++++++++++++++++++-
>> 1 file changed, 42 insertions(+), 1 deletion(-)
>>
>> diff --git a/drivers/of/base.c b/drivers/of/base.c
>> index ba195fbce4c6..10b51106c854 100644
>> --- a/drivers/of/base.c
>> +++ b/drivers/of/base.c
>> @@ -730,13 +730,49 @@ out:
>> }
>> EXPORT_SYMBOL(of_find_node_with_property);
>>
>> +static const struct of_device_id *
>> +of_match_compatible(const struct of_device_id *matches,
>> + const struct device_node *node)
>> +{
>> + const char *cp;
>> + int cplen, l;
>> + const struct of_device_id *m;
>> +
>> + cp = __of_get_property(node, "compatible", &cplen);
>> + while (cp && (cplen > 0)) {
>> + m = matches;
>> + while (m->name[0] || m->type[0] || m->compatible[0]) {
>> + /* Only match for the entries without type and name */
>> + if (m->name[0] || m->type[0] ||
>> + of_compat_cmp(m->compatible, cp,
>> + strlen(m->compatible)))
>> + m++;
>> + else
>> + return m;
>> + }
>> +
>> + /* Get node's next compatible string */
>> + l = strlen(cp) + 1;
>> + cp += l;
>> + cplen -= l;
>> + }
>> +
>> + return NULL;
>> +}
>> +
>> static
>> const struct of_device_id *__of_match_node(const struct of_device_id *matches,
>> const struct device_node *node)
>> {
>> + const struct of_device_id *m;
>> +
>> if (!matches)
>> return NULL;
>>
>> + m = of_match_compatible(matches, node);
>> + if (m)
>> + return m;
>> +
>> while (matches->name[0] || matches->type[0] || matches->compatible[0]) {
>> int match = 1;
>> if (matches->name[0])
>> @@ -760,7 +796,12 @@ const struct of_device_id *__of_match_node(const struct of_device_id *matches,
>> * @matches: array of of device match structures to search in
>> * @node: the of device structure to match against
>> *
>> - * Low level utility function used by device matching.
>> + * Low level utility function used by device matching. We have two ways
>> + * of matching:
>> + * - Try to find the best compatible match by comparing each compatible
>> + * string of device node with all the given matches respectively.
>> + * - If the above method failed, then try to match the compatible by using
>> + * __of_device_is_compatible() besides the match in type and name.
>> */
>> const struct of_device_id *of_match_node(const struct of_device_id *matches,
>> const struct device_node *node)
>> --
>> 1.8.5.3
--
Employee of Qualcomm Innovation Center, Inc.
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, hosted by The Linux Foundation
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply
* Re: [PATCH 2/2] of: search the best compatible match first in __of_match_node()
From: Rob Herring @ 2014-02-14 15:53 UTC (permalink / raw)
To: Kevin Hao
Cc: devicetree-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, linuxppc-dev,
Sebastian Hesselbarth, Stephen N Chivers, Grant Likely,
Rob Herring, Kumar Gala
In-Reply-To: <1392355366-1445-3-git-send-email-haokexin-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
On Thu, Feb 13, 2014 at 11:22 PM, Kevin Hao <haokexin-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
> Currently, of_match_node compares each given match against all node's
> compatible strings with of_device_is_compatible.
>
> To achieve multiple compatible strings per node with ordering from
> specific to generic, this requires given matches to be ordered from
> specific to generic. For most of the drivers this is not true and also
> an alphabetical ordering is more sane there.
>
> Therefore, this patch introduces a function to match each of the node's
> compatible strings against all given compatible matches without type and
> name first, before checking the next compatible string. This implies
> that node's compatibles are ordered from specific to generic while
> given matches can be in any order. If we fail to find such a match
> entry, then fall-back to the old method in order to keep compatibility.
>
> Cc: Sebastian Hesselbarth <sebastian.hesselbarth-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
> Signed-off-by: Kevin Hao <haokexin-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
Looks good to me. I'll put this in next for a few days. I'd really
like to see some acks and tested-by's before sending to Linus.
We could be a bit more strict here and fallback to the old matching if
the match table has any entries with name or type. I don't think that
should be necessary though.
Rob
> ---
> drivers/of/base.c | 43 ++++++++++++++++++++++++++++++++++++++++++-
> 1 file changed, 42 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/of/base.c b/drivers/of/base.c
> index ba195fbce4c6..10b51106c854 100644
> --- a/drivers/of/base.c
> +++ b/drivers/of/base.c
> @@ -730,13 +730,49 @@ out:
> }
> EXPORT_SYMBOL(of_find_node_with_property);
>
> +static const struct of_device_id *
> +of_match_compatible(const struct of_device_id *matches,
> + const struct device_node *node)
> +{
> + const char *cp;
> + int cplen, l;
> + const struct of_device_id *m;
> +
> + cp = __of_get_property(node, "compatible", &cplen);
> + while (cp && (cplen > 0)) {
> + m = matches;
> + while (m->name[0] || m->type[0] || m->compatible[0]) {
> + /* Only match for the entries without type and name */
> + if (m->name[0] || m->type[0] ||
> + of_compat_cmp(m->compatible, cp,
> + strlen(m->compatible)))
> + m++;
> + else
> + return m;
> + }
> +
> + /* Get node's next compatible string */
> + l = strlen(cp) + 1;
> + cp += l;
> + cplen -= l;
> + }
> +
> + return NULL;
> +}
> +
> static
> const struct of_device_id *__of_match_node(const struct of_device_id *matches,
> const struct device_node *node)
> {
> + const struct of_device_id *m;
> +
> if (!matches)
> return NULL;
>
> + m = of_match_compatible(matches, node);
> + if (m)
> + return m;
> +
> while (matches->name[0] || matches->type[0] || matches->compatible[0]) {
> int match = 1;
> if (matches->name[0])
> @@ -760,7 +796,12 @@ const struct of_device_id *__of_match_node(const struct of_device_id *matches,
> * @matches: array of of device match structures to search in
> * @node: the of device structure to match against
> *
> - * Low level utility function used by device matching.
> + * Low level utility function used by device matching. We have two ways
> + * of matching:
> + * - Try to find the best compatible match by comparing each compatible
> + * string of device node with all the given matches respectively.
> + * - If the above method failed, then try to match the compatible by using
> + * __of_device_is_compatible() besides the match in type and name.
> */
> const struct of_device_id *of_match_node(const struct of_device_id *matches,
> const struct device_node *node)
> --
> 1.8.5.3
>
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply
* Re: [PATCH v3] phy: Add new Exynos5 USB 3.0 PHY driver
From: Tomasz Figa @ 2014-02-14 15:47 UTC (permalink / raw)
To: Vivek Gautam
Cc: Vivek Gautam, Linux USB Mailing List,
linux-samsung-soc-u79uwXL29TY76Z2rM5mHXA@public.gmane.org,
linux-kernel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org,
devicetree-u79uwXL29TY76Z2rM5mHXA@public.gmane.org,
linux-doc-u79uwXL29TY76Z2rM5mHXA, Greg KH, Kukjin Kim,
Felipe Balbi, kishon, Kamil Debski, Sylwester Nawrocki,
Julius Werner, Jingoo Han
In-Reply-To: <CAFp+6iFYoN1w_BX7mMEa-Xx5ADaPTTKeDzjsSbQmwVKdSEsmOw-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
Hi Vivek,
On 14.02.2014 14:53, Vivek Gautam wrote:
>>> Changes from v2:
>>> 1) Added support for multiple PHYs (UTMI+ and PIPE3) and
>>> related changes in the driver structuring.
>>
>>
>> I'm a bit skeptical about this separation. Can the PHY operate with just the
>> UTMI+ or PIPE3 part enabled alone without the other? Can any PHY consumer
>> operate this way?
>
> Yes :-)
> As also pointed by Kishon the PHY consumer (which is DWC3 in case of
> Exynos5 SoC series)
> should theoretically be able use either UTMI+ phy for High speed
> operations or both (UTMI+ and PIPE3)
> for Super Speed operations.
OK, that's fine then. This is the explanation I needed, thanks.
>>
>> I believe the right thing to do here is to do all the initialization in
>> .power_on() and let the driver simply call phy_power_on() when it needs the
>> PHY and phy_power_off() otherwise.
>
> If this is what we should be doing then what will be the purpose of
> two separate APIs :
> phy_power_on() and phy_init().
> Am i missing while understanding the things.
>
I don't understand this separation as well. Operations that should be
done together shouldn't be separated. Is there any case when you can
call one of phy_power_on() and phy_init() without calling another one
right before/after it?
Best regards,
Tomasz
--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply
* Re: [PATCH 1/2] mfd: twl4030-madc: Add devicetree support.
From: Lee Jones @ 2014-02-14 15:28 UTC (permalink / raw)
To: Belisko Marek
Cc: Rob Herring, Pawel Moll, Mark Rutland,
ijc+devicetree-KcIKpvwj1kUDXYZnReoRVg@public.gmane.org,
Kumar Gala, Rob Landley, Russell King - ARM Linux, Grant Likely,
devicetree-u79uwXL29TY76Z2rM5mHXA@public.gmane.org,
linux-doc-u79uwXL29TY76Z2rM5mHXA@public.gmane.org,
Dr. H. Nikolaus Schaller, LKML, linux-arm-kernel
In-Reply-To: <CAAfyv36FH364OPnktx_Qh6sUfwYe7hynqrc3Xs_tY4gPXqQWvA-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
> >> Signed-off-by: Marek Belisko <marek-xXXSsgcRVICgSpxsJD1C4w@public.gmane.org>
> >> ---
> >> .../devicetree/bindings/mfd/twl4030-madc.txt | 18 +++++++++++++
> >> drivers/mfd/twl4030-madc.c | 31
> >> ++++++++++++++++++++--
> >
<snip>
> >> +static struct twl4030_madc_platform_data *
> >> + twl4030_madc_of_parse(struct platform_device *pdev)
> >> +{
> >> + struct twl4030_madc_platform_data *pdata;
> >> +
> >> + pdata = devm_kzalloc(&pdev->dev,
> >> + sizeof(struct twl4030_madc_platform_data), GFP_KERNEL);
> >
> > s/struct twl4030_madc_platform_data/*pdata/
> Right typo.
Sorry, my comment of ambiguous.
I mean do this:
sizeof(*pdata)
... instead of this:
sizeof(struct twl4030_madc_platform_data)
> >> + if (!pdata)
> >> + return ERR_PTR(-ENOMEM);
> >> +
> >> + pdata->irq_line = platform_get_irq(pdev, 0);
> >
> > Why weren't 'resources' used in the original implementation?
> Not sure I'm not an author :). It's passed in platform data.
Yes, I saw that. It should be changed.
--
Lee Jones
Linaro STMicroelectronics Landing Team Lead
Linaro.org │ Open source software for ARM SoCs
Follow Linaro: Facebook | Twitter | Blog
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply
* Re: [PATCH v10 0/4] ata: Add APM X-Gene SoC SATA host controller support
From: Tejun Heo @ 2014-02-14 15:03 UTC (permalink / raw)
To: Loc Ho
Cc: Olof Johansson, Arnd Bergmann, Linux SCSI List,
linux-ide@vger.kernel.org, devicetree@vger.kernel.org,
linux-arm-kernel@lists.infradead.org, David Milburn, Jon Masters,
patches@apm.com, Tuan Phan, Suman Tripathi
In-Reply-To: <CAPw-ZTn5c=v+7pbetEjXWpw9i=cG+zdeHPxO1QD0e42uFrKnCg@mail.gmail.com>
Hello, Loc.
On Thu, Feb 13, 2014 at 03:28:01PM -0800, Loc Ho wrote:
> 1. There are a number of errata that require workaround. Some can be
> fixed by adding broken flags while others are better to just wrap
> around the existent libahci library routines and not overly polluting
> the libahci routines.
> 2. There are additional controller programming sequences to configure.
> 2a. By default, RAM are powered down and require brought out of shutdown.
> 2b. The controller has an additional corresponding PHY part that needs
> to be programmed after PHY configuration.
Have you looked at the latest patchset Hans posted? He added multiple
PHY support and split up init to three steps so that each platform
driver can mix and match as they see fit. Looking at xgene driver,
sure there are things specific to the driver but there also are
non-insignificant amount of boilerplate code and that's what I'm
primarily concerned about. It may be okay when you have two or three
drivers duplicating some code but it looks like we could have many
more and I *really* want to avoid the situation where the same piece
of code is copied over N times. In addition, frankly, not many people
except yourself would care about these drivers once they're merged and
many of these are gonna be painful to test making later refactoring a
lot harder.
> 2c. The controller requires extra programming sequence for the
> hardreset due to errata.
> 2d. For the IO flush, it requires additional memory resources.
Sure, you'll need to override good parts of the driver. What I'm
saying is please try to reuse whatever you can. If that takes
refactoring and librarize ahci_platform, please do so and I do see
healthy chunk of duplicated code in the init path. Please take a look
at Hans' patches and if necessary work with him so that your driver
can be part of the refactoring.
Thanks.
--
tejun
^ permalink raw reply
page: next (older) | prev (newer) | latest
- recent:[subjects (threaded)|topics (new)|topics (active)]
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox