Skip to contentSkip to conversation

ARM: omap: omap4-embt2ws: 32K clock for WLAN

PATCH Applied Andreas Kemnade 4 participants 14 messages Sep 16, 2023 – Nov 30, 2023

Subsystems (7)
lore
Share & tools
lore

Andreas Kemnade

author maintainer
#0 lore raw

v4 cover ARM: omap: omap4-embt2ws: 32K clock for WLAN

To have WLAN working properly, enable a 32K clock of the TWL6032. In earlier tests, it was still enabled from a previous boot into the vendor system.

Changes in V4:
- use dev_err_probe in clk probe()
- R-by

Changes in V3:
- maintainer change in binding doc
- fix references to binding doc
- additionalProperties: false
- remove subdevices also from examples until
subdevices are referenced/added

Changes in V2:
- no separate device node for the clock
- converted toplevel node of TWL

Andreas Kemnade (5):
dt-bindings: mfd: convert twl-family.txt to json-schema
dt-bindings: mfd: ti,twl: Add clock provider properties
mfd: twl-core: Add a clock subdevice for the TWL6032
clk: twl: add clock driver for TWL6032
ARM: dts: omap4-embt2ws: enable 32K clock on WLAN

 .../bindings/input/twl4030-pwrbutton.txt      |   2 +-
 .../devicetree/bindings/mfd/ti,twl.yaml       |  67 ++++++
 .../devicetree/bindings/mfd/twl-family.txt    |  46 ----
 .../boot/dts/ti/omap/omap4-epson-embt2ws.dts  |   8 +
 drivers/clk/Kconfig                           |   9 +
 drivers/clk/Makefile                          |   1 +
 drivers/clk/clk-twl.c                         | 197 ++++++++++++++++++
 drivers/mfd/twl-core.c                        |  16 ++
 8 files changed, 299 insertions(+), 47 deletions(-)
 create mode 100644 Documentation/devicetree/bindings/mfd/ti,twl.yaml
 delete mode 100644 Documentation/devicetree/bindings/mfd/twl-family.txt
 create mode 100644 drivers/clk/clk-twl.c
Signature · 1 line
2.39.2
author maintainer
#1 lore raw

PATCH v4 1/5 dt-bindings: mfd: convert twl-family.txt to json-schema

Convert the TWL[46]030 binding to DT schema format. To do it as a step by step work, do not include / handle nodes for subdevices yet, just convert things with minimal corrections. There are already some bindings for its subdevices in the tree.

---
 .../bindings/input/twl4030-pwrbutton.txt      |  2 +-
 .../devicetree/bindings/mfd/ti,twl.yaml       | 64 +++++++++++++++++++
 .../devicetree/bindings/mfd/twl-family.txt    | 46 -------------
 3 files changed, 65 insertions(+), 47 deletions(-)
 create mode 100644 Documentation/devicetree/bindings/mfd/ti,twl.yaml
 delete mode 100644 Documentation/devicetree/bindings/mfd/twl-family.txt
Documentation/devicetree/bindings/input/twl4030-pwrbutton.txt +1 -1
index f5021214edecb..6c201a2ba8acf 100644
@@ -1,7 +1,7 @@
 Texas Instruments TWL family (twl4030) pwrbutton module
 
 This module is part of the TWL4030. For more details about the whole
-chip see Documentation/devicetree/bindings/mfd/twl-family.txt.
+chip see Documentation/devicetree/bindings/mfd/ti,twl.yaml.
 
 This module provides a simple power button event via an Interrupt.
 
Documentation/devicetree/bindings/mfd/ti,twl.yaml new file +64 -0

64 lines not on this page · raw message

Documentation/devicetree/bindings/mfd/twl-family.txt deleted +0 -46

46 lines not on this page · raw message

Signature · 1 line
2.39.2
author maintainer
#2 lore raw

PATCH v4 2/5 dt-bindings: mfd: ti,twl: Add clock provider properties

In reply to Andreas Kemnade

Since these devices provide clock outputs, add the corresponding property.

---
 Documentation/devicetree/bindings/mfd/ti,twl.yaml | 3 +++
 1 file changed, 3 insertions(+)
Documentation/devicetree/bindings/mfd/ti,twl.yaml +3 -0
index f125b254a4b93..c04d57ba22b49 100644
@@ -37,6 +37,9 @@ properties:
   "#interrupt-cells":
     const: 1
 
+  "#clock-cells":
+    const: 1
+
 additionalProperties: false
 
 required:
Signature · 1 line
2.39.2
author maintainer
#4 lore raw

PATCH v4 3/5 mfd: twl-core: Add a clock subdevice for the TWL6032

In reply to Andreas Kemnade

Clock device needs no separate devicetree node, so add it as a platform device. Other devices in the family also have controllable clocks, but due to the lack of testing, just add it for the TWL6032 now.

---
 drivers/mfd/twl-core.c | 16 ++++++++++++++++
 1 file changed, 16 insertions(+)
drivers/mfd/twl-core.c +16 -0
index ce01a87f8dc39..234500b2e53fc 100644
@@ -31,6 +31,8 @@
 #include <linux/regulator/machine.h>
 
 #include <linux/i2c.h>
+
+#include <linux/mfd/core.h>
 #include <linux/mfd/twl.h>
 
 /* Register descriptions for audio */
@@ -690,6 +692,10 @@ static struct of_dev_auxdata twl_auxdata_lookup[] = {
 	{ /* sentinel */ },
 };
 
+static const struct mfd_cell twl6032_cells[] = {
+	{ .name = "twl6032-clk" },
+};
+
 /* NOTE: This driver only handles a single twl4030/tps659x0 chip */
 static int
 twl_probe(struct i2c_client *client)
@@ -836,6 +842,16 @@ twl_probe(struct i2c_client *client)
 				 TWL4030_DCDC_GLOBAL_CFG);
 	}
 
+	if (id->driver_data == (TWL6030_CLASS | TWL6032_SUBCLASS)) {
+		status = devm_mfd_add_devices(&client->dev,
+					      PLATFORM_DEVID_NONE,
+					      twl6032_cells,
+					      ARRAY_SIZE(twl6032_cells),
+					      NULL, 0, NULL);
+		if (status < 0)
+			goto free;
+	}
+
 	status = of_platform_populate(node, NULL, twl_auxdata_lookup,
 				      &client->dev);
 
Signature · 1 line
2.39.2
author maintainer
#5 lore raw

PATCH v4 4/5 clk: twl: add clock driver for TWL6032

In reply to Andreas Kemnade

The TWL6032 has some clock outputs which are controlled like fixed-voltage regulators, in some drivers for these chips found in the wild, just the regulator api is abused for controlling them, so simply use something similar to the regulator functions. Due to a lack of hardware available for testing, leave out the TWL6030-specific part of those functions.

---
 drivers/clk/Kconfig   |   9 ++
 drivers/clk/Makefile  |   1 +
 drivers/clk/clk-twl.c | 197 ++++++++++++++++++++++++++++++++++++++++++
 3 files changed, 207 insertions(+)
 create mode 100644 drivers/clk/clk-twl.c
drivers/clk/Kconfig +9 -0
index c30099866174d..3944f081ebade 100644
@@ -277,6 +277,15 @@ config COMMON_CLK_S2MPS11
 	  clock. These multi-function devices have two (S2MPS14) or three
 	  (S2MPS11, S5M8767) fixed-rate oscillators, clocked at 32KHz each.
 
+config CLK_TWL
+	tristate "Clock driver for the TWL PMIC family"
+	depends on TWL4030_CORE
+	help
+	  Enable support for controlling the clock resources on TWL family
+	  PMICs. These devices have some 32K clock outputs which can be
+	  controlled by software. For now, only the TWL6032 clocks are
+	  supported.
+
 config CLK_TWL6040
 	tristate "External McPDM functional clock from twl6040"
 	depends on TWL6040_CORE
drivers/clk/Makefile +1 -0
index 18969cbd4bb1e..86e46adcb619c 100644
@@ -72,6 +72,7 @@ obj-$(CONFIG_COMMON_CLK_STM32H7)	+= clk-stm32h7.o
 obj-$(CONFIG_COMMON_CLK_STM32MP157)	+= clk-stm32mp1.o
 obj-$(CONFIG_COMMON_CLK_TPS68470)      += clk-tps68470.o
 obj-$(CONFIG_CLK_TWL6040)		+= clk-twl6040.o
+obj-$(CONFIG_CLK_TWL)			+= clk-twl.o
 obj-$(CONFIG_ARCH_VT8500)		+= clk-vt8500.o
 obj-$(CONFIG_COMMON_CLK_RS9_PCIE)	+= clk-renesas-pcie.o
 obj-$(CONFIG_COMMON_CLK_SI521XX)	+= clk-si521xx.o
drivers/clk/clk-twl.c new file +197 -0

197 lines not on this page · raw message

Signature · 1 line
2.39.2
author maintainer
#6 lore raw

Re: [PATCH v4 4/5] clk: twl: add clock driver for TWL6032

Quoted 9 lines · On Sat, 16 Sep 2023 12:05:14 +0200 Andreas Kemnade <[email protected]> wrote:

The TWL6032 has some clock outputs which are controlled like fixed-voltage regulators, in some drivers for these chips found in the wild, just the regulator api is abused for controlling them, so simply use something similar to the regulator functions. Due to a lack of hardware available for testing, leave out the TWL6030-specific part of those functions.

---

ping...

anything left to do here?

Regards,
Andreas

maintainer Acked-by on v4 4/5
#7 lore raw

Re: [PATCH v4 4/5] clk: twl: add clock driver for TWL6032

In reply to Andreas Kemnade

Quoting Andreas Kemnade (2023-09-16 03:05:14)

Quoted 9 lines

The TWL6032 has some clock outputs which are controlled like fixed-voltage regulators, in some drivers for these chips found in the wild, just the regulator api is abused for controlling them, so simply use something similar to the regulator functions. Due to a lack of hardware available for testing, leave out the TWL6030-specific part of those functions.

---

Did you want me to pick this up in clk tree?

author maintainer
#8 lore raw

Re: [PATCH v4 4/5] clk: twl: add clock driver for TWL6032

Am Wed, 18 Oct 2023 18:20:22 -0700
schrieb Stephen Boyd <[email protected]>:

Quoted 8 lines

Quoting Andreas Kemnade (2023-09-16 03:05:14)

Quoted 7 lines

The TWL6032 has some clock outputs which are controlled like fixed-voltage regulators, in some drivers for these chips found in the wild, just the regulator api is abused for controlling them, so simply use something similar to the regulator functions. Due to a lack of hardware available for testing, leave out the TWL6030-specific part of those functions.

Quoted 7 lines
Quoted 2 lines

---

Did you want me to pick this up in clk tree?

Well, Lee has picked up 1-3 of this series, so I think it is a good idea if you take it. I do not think it will make things any easier if it goes through Tony's tree because part 5 is a bit tricky.
If picked up independently (of the whole set and not only part 4), there will be a range of commits were WLAN will not work at all instead of basically working randomly by the clock is enabled for some reason (can happen if warm-booted from the vendor kernel).

So IMHO for part 5 we have to decide wether to decide if above is considered to break bisectability and take the patch for 6.8 or ignore that and take it for 6.7.

But no matter how we decide, that should not affect the route for this part 4. So it is a good idea to have it through the clk tree.

Regrads,
Andreas

maintainer Applied
#9 lore raw

Re: [PATCH v4 4/5] clk: twl: add clock driver for TWL6032

In reply to Andreas Kemnade

Quoting Andreas Kemnade (2023-09-16 03:05:14)

Quoted 9 lines

The TWL6032 has some clock outputs which are controlled like fixed-voltage regulators, in some drivers for these chips found in the wild, just the regulator api is abused for controlling them, so simply use something similar to the regulator functions. Due to a lack of hardware available for testing, leave out the TWL6030-specific part of those functions.

---

Applied to clk-next

author maintainer
#10 lore raw

PATCH v4 5/5 ARM: dts: omap4-embt2ws: enable 32K clock on WLAN

In reply to Andreas Kemnade

WLAN did only work if clock was left enabled by the original system, so make it fully enable the needed resources itself.

---
 arch/arm/boot/dts/ti/omap/omap4-epson-embt2ws.dts | 8 ++++++++
 1 file changed, 8 insertions(+)
arch/arm/boot/dts/ti/omap/omap4-epson-embt2ws.dts +8 -0
index ee86981b2e448..9d2f2d8639496 100644
@@ -69,6 +69,12 @@ unknown_supply: unknown-supply {
 		regulator-name = "unknown";
 	};
 
+	wl12xx_pwrseq: wl12xx-pwrseq {
+		compatible = "mmc-pwrseq-simple";
+		clocks = <&twl 1>;
+		clock-names = "ext_clock";
+	};
+
        /* regulator for wl12xx on sdio2 */
 	wl12xx_vmmc: wl12xx-vmmc {
 		pinctrl-names = "default";
@@ -92,6 +98,7 @@ &i2c1 {
 	twl: pmic@48 {
 		compatible = "ti,twl6032";
 		reg = <0x48>;
+		#clock-cells = <1>;
 		/* IRQ# = 7 */
 		interrupts = <GIC_SPI 7 IRQ_TYPE_LEVEL_HIGH>; /* IRQ_SYS_1N cascaded to gic */
 		interrupt-controller;
@@ -316,6 +323,7 @@ &mmc3 {
 	pinctrl-names = "default";
 	pinctrl-0 = <&wl12xx_pins>;
 	vmmc-supply = <&wl12xx_vmmc>;
+	mmc-pwrseq = <&wl12xx_pwrseq>;
 	interrupts-extended = <&wakeupgen GIC_SPI 94 IRQ_TYPE_LEVEL_HIGH
 			       &omap4_pmx_core 0x12e>;
 	non-removable;
Signature · 1 line
2.39.2
maintainer
#11 lore raw

Re: [PATCH v4 5/5] ARM: dts: omap4-embt2ws: enable 32K clock on WLAN

Quoted 2 lines · Andreas Kemnade

WLAN did only work if clock was left enabled by the original system, so make it fully enable the needed resources itself.

Seems applying this dts change before the clock patch is applied would break wlan so please let me know when this is safe to apply.

Regards,

Tony

maintainer
#12 lore raw

Re: [PATCH v4 5/5] ARM: dts: omap4-embt2ws: enable 32K clock on WLAN

Quoted 6 lines · Tony Lindgren
Quoted 2 lines · Andreas Kemnade

WLAN did only work if clock was left enabled by the original system, so make it fully enable the needed resources itself.

Seems applying this dts change before the clock patch is applied would break wlan so please let me know when this is safe to apply.

Applying into omap-for-v6.8/dt thanks.

Tony

maintainer Applied
#13 lore raw

Re: (subset) [PATCH v4 0/5] ARM: omap: omap4-embt2ws: 32K clock for WLAN

In reply to Andreas Kemnade
Quoted 9 lines · Andreas Kemnade

To have WLAN working properly, enable a 32K clock of the TWL6032. In earlier tests, it was still enabled from a previous boot into the vendor system.

Changes in V4:
- use dev_err_probe in clk probe()
- R-by

[...]

Applied, thanks!

[1/5] dt-bindings: mfd: convert twl-family.txt to json-schema

      commit: 098e2d6fd72fc99097af33e6e8cb4cd0921a26ac
[2/5] dt-bindings: mfd: ti,twl: Add clock provider properties
      commit: eb9fba08e386ff98818de12b40a5ee1443129229
[3/5] mfd: twl-core: Add a clock subdevice for the TWL6032
      commit: 3e359099b245227c6729dd5f81f2757dea5d4656

--
Lee Jones [李琼斯]