Skip to contentSkip to conversation

i2c: i2c-ocores: clock improvements

PATCH Applied Max Filippov 2 participants 8 messages Feb 2, 2015 – Feb 5, 2015

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

v2 cover i2c: i2c-ocores: clock improvements

Peter, Wolfram,

this series fixes clock-frequency attribute usage by i2c-ocores driver and adds optional common clock framework support.

Changes v1->v2:
- add cleanup patch for clock-frequency attribute that implements options 2 and 3, as was suggested here: https://lkml.org/lkml/2015/1/22/527
- make CCF support optional

Max Filippov (2):
i2c: i2c-ocores: fix clock-frequency binding usage
i2c: i2c-ocores: add common clock support

 .../devicetree/bindings/i2c/i2c-ocores.txt         | 42 +++++++++-
 drivers/i2c/busses/i2c-ocores.c                    | 95 +++++++++++++++++++---
 2 files changed, 122 insertions(+), 15 deletions(-)
Signature · 1 line
1.8.1.4
#1 lore raw

PATCH v2 1/2 i2c: i2c-ocores: fix clock-frequency binding usage

clock-frequency property is meant to control the bus frequency for i2c bus drivers, but it was incorrectly used to specify i2c controller input clock frequency.
Introduce new attribute, opencores,ip-clock-frequency, that specifies i2c controller clock frequency and make clock-frequency attribute compatible with other i2c drivers. Maintain backwards compatibility in case opencores,ip-clock-frequency attribute is missing.

---
Changes v1->v2:
- new patch

 .../devicetree/bindings/i2c/i2c-ocores.txt         | 18 ++++++-
 drivers/i2c/busses/i2c-ocores.c                    | 55 +++++++++++++++++-----
 2 files changed, 58 insertions(+), 15 deletions(-)
Documentation/devicetree/bindings/i2c/i2c-ocores.txt +16 -2
index 1637c29..5bef3ad 100644
@@ -4,15 +4,29 @@ Required properties:
 - compatible      : "opencores,i2c-ocores" or "aeroflexgaisler,i2cmst"
 - reg             : bus address start and address range size of device
 - interrupts      : interrupt number
-- clock-frequency : frequency of bus clock in Hz
+- opencores,ip-clock-frequency: frequency of the controller clock in Hz;
+                    see the note below
 - #address-cells  : should be <1>
 - #size-cells     : should be <0>
 
 Optional properties:
+- clock-frequency : frequency of bus clock in Hz; see the note below.
+                    Defaults to 100 KHz when the property is not specified
 - reg-shift       : device register offsets are shifted by this value
 - reg-io-width    : io register width in bytes (1, 2 or 4)
 - regstep         : deprecated, use reg-shift above
 
+Note
+clock-frequency property is meant to control the bus frequency for i2c bus
+drivers, but it was incorrectly used to specify i2c controller input clock
+frequency. So the following rules are set to fix this situation:
+- if clock-frequency is present and opencores,ip-clock-frequency is not,
+  then clock-frequency specifies i2c controller clock frequency. This is
+  to keep backwards compatibility with setups using old DTB. i2c bus
+  frequency is fixed at 100 KHz.
+- if opencores,ip-clock-frequency is present it specifies i2c controller
+  clock frequency. clock-frequency property specifies i2c bus frequency.
+
 Example:
 
 	i2c0: ocores@a0000000 {
@@ -21,7 +35,7 @@ Example:
 		compatible = "opencores,i2c-ocores";
 		reg = <0xa0000000 0x8>;
 		interrupts = <10>;
-		clock-frequency = <20000000>;
+		opencores,ip-clock-frequency = <20000000>;
 
 		reg-shift = <0>;	/* 8 bit registers */
 		reg-io-width = <1>;	/* 8 bit read/write */
drivers/i2c/busses/i2c-ocores.c +42 -13

111 lines not on this page · raw message

Signature · 1 line
1.8.1.4
Applied
#2 lore raw

Re: [PATCH v2 1/2] i2c: i2c-ocores: fix clock-frequency binding usage

Quoted 9 lines · Max Filippov

clock-frequency property is meant to control the bus frequency for i2c bus drivers, but it was incorrectly used to specify i2c controller input clock frequency.
Introduce new attribute, opencores,ip-clock-frequency, that specifies i2c controller clock frequency and make clock-frequency attribute compatible with other i2c drivers. Maintain backwards compatibility in case opencores,ip-clock-frequency attribute is missing.

Exactly what I had in mind. Thank you for making the effort to fix this properly!

Applied to for-next!

#3 lore raw

Re: [PATCH v2 1/2] i2c: i2c-ocores: fix clock-frequency binding usage

Quoted 4 lines · Wolfram Sang

Exactly what I had in mind. Thank you for making the effort to fix this properly!

Applied to for-next!

Wolfram, is there any chance this series makes it into 3.20? (I need to update xtfpga platform, and the update depends on whether these changes are there or not).

Signature · 2 lines
Thanks.
-- Max
#4 lore raw

Re: [PATCH v2 1/2] i2c: i2c-ocores: fix clock-frequency binding usage

Quoted 9 lines · Max Filippov
Quoted 4 lines · Wolfram Sang

Exactly what I had in mind. Thank you for making the effort to fix this properly!

Applied to for-next!

Wolfram, is there any chance this series makes it into 3.20? (I need to update xtfpga platform, and the update depends on whether these changes are there or not).

Yes, it is planned for 3.20.

#5 lore raw

PATCH v2 2/2 i2c: i2c-ocores: add common clock support

In reply to Max Filippov

Allow bus clock specification as a common clock handle. This makes this controller easier to use in a setup based on common clock framework.

---
Changes v1->v2:
- fix binding documentation;
- make CCF support optional.

 .../devicetree/bindings/i2c/i2c-ocores.txt         | 32 +++++++++++++---
 drivers/i2c/busses/i2c-ocores.c                    | 44 +++++++++++++++++++++-
 2 files changed, 70 insertions(+), 6 deletions(-)
Documentation/devicetree/bindings/i2c/i2c-ocores.txt +27 -5
index 5bef3ad..17bef9a 100644
@@ -4,8 +4,10 @@ Required properties:
 - compatible      : "opencores,i2c-ocores" or "aeroflexgaisler,i2cmst"
 - reg             : bus address start and address range size of device
 - interrupts      : interrupt number
+- clocks          : handle to the controller clock; see the note below.
+                    Mutually exclusive with opencores,ip-clock-frequency
 - opencores,ip-clock-frequency: frequency of the controller clock in Hz;
-                    see the note below
+                    see the note below. Mutually exclusive with clocks
 - #address-cells  : should be <1>
 - #size-cells     : should be <0>
 
@@ -20,14 +22,16 @@ Note
 clock-frequency property is meant to control the bus frequency for i2c bus
 drivers, but it was incorrectly used to specify i2c controller input clock
 frequency. So the following rules are set to fix this situation:
-- if clock-frequency is present and opencores,ip-clock-frequency is not,
-  then clock-frequency specifies i2c controller clock frequency. This is
-  to keep backwards compatibility with setups using old DTB. i2c bus
+- if clock-frequency is present and neither opencores,ip-clock-frequency nor
+  clocks are, then clock-frequency specifies i2c controller clock frequency.
+  This is to keep backwards compatibility with setups using old DTB. i2c bus
   frequency is fixed at 100 KHz.
+- if clocks is present it specifies i2c controller clock. clock-frequency
+  property specifies i2c bus frequency.
 - if opencores,ip-clock-frequency is present it specifies i2c controller
   clock frequency. clock-frequency property specifies i2c bus frequency.
 
-Example:
+Examples:
 
 	i2c0: ocores@a0000000 {
 		#address-cells = <1>;
@@ -45,3 +49,21 @@ Example:
 			reg = <0x60>;
 		};
 	};
+or
+	i2c0: ocores@a0000000 {
+		#address-cells = <1>;
+		#size-cells = <0>;
+		compatible = "opencores,i2c-ocores";
+		reg = <0xa0000000 0x8>;
+		interrupts = <10>;
+		clocks = <&osc>;
+		clock-frequency = <400000>; /* i2c bus frequency 400 KHz */
+
+		reg-shift = <0>;	/* 8 bit registers */
+		reg-io-width = <1>;	/* 8 bit read/write */
+
+		dummy@60 {
+			compatible = "dummy";
+			reg = <0x60>;
+		};
+	};
drivers/i2c/busses/i2c-ocores.c +43 -1

80 lines not on this page · raw message

Signature · 1 line
1.8.1.4
#6 lore raw

Re: [PATCH v2 2/2] i2c: i2c-ocores: add common clock support

Quoted 9 lines
drivers/i2c/busses/i2c-ocores.c +8 -1
-	if (of_property_read_u32(np, "opencores,ip-clock-frequency", &val)) {
+#ifdef CONFIG_HAVE_CLK
+	i2c->clk = devm_clk_get(&pdev->dev, NULL);
+
+	if (!IS_ERR(i2c->clk)) {
+		int ret = clk_prepare_enable(i2c->clk);
+
+		if (ret) {
+			dev_err(&pdev->dev,

drivers/i2c/busses/i2c-ocores.c static int ocores_i2c_of_probe(struct platform_device *pdev, · line 363

Quoted 10 lines
drivers/i2c/busses/i2c-ocores.c +10 -0
+				"clk_prepare_enable failed: %d\n", ret);
+			return ret;
+		}
+		i2c->ip_clock_khz = clk_get_rate(i2c->clk) / 1000;
+		if (clock_frequency_present)
+			i2c->bus_clock_khz = clock_frequency / 1000;
+	} else
+#endif
+		if (of_property_read_u32(np, "opencores,ip-clock-frequency",
+					 &val)) {

This is not very readable (wrong indentation) and fragile. I'd suggest to leave out #ifdeffery, the compiler will remove the code for !HAVE_CLK because of the empty dummy functions[1]. Then, you can write properly formatted code again.

[1] Well, it should if there wasn't a bug in that: https://lkml.org/lkml/2015/2/5/544

drivers/i2c/busses/i2c-ocores.c static int ocores_i2c_resume(struct device *dev) · line 540

Quoted 12 lines
drivers/i2c/busses/i2c-ocores.c +12 -0
+#ifdef CONFIG_HAVE_CLK
+	if (!IS_ERR(i2c->clk)) {
+		int ret = clk_prepare_enable(i2c->clk);
+
+		if (ret) {
+			dev_err(&pdev->dev,
+				"clk_prepare_enable failed: %d\n", ret);
+			return ret;
+		}
+		i2c->ip_clock_khz = clk_get_rate(i2c->clk) / 1000;
+	}
+#endif

Ahem, there is a build error in here :(

#7 lore raw

Re: [PATCH v2 2/2] i2c: i2c-ocores: add common clock support

Quoted 26 lines · Wolfram Sang
Quoted 10 lines
drivers/i2c/busses/i2c-ocores.c +8 -1
-     if (of_property_read_u32(np, "opencores,ip-clock-frequency", &val)) {
+#ifdef CONFIG_HAVE_CLK
+     i2c->clk = devm_clk_get(&pdev->dev, NULL);
+
+     if (!IS_ERR(i2c->clk)) {
+             int ret = clk_prepare_enable(i2c->clk);
+
+             if (ret) {
+                     dev_err(&pdev->dev,

drivers/i2c/busses/i2c-ocores.c static int ocores_i2c_of_probe(struct platform_device *pdev, · line 363

Quoted 10 lines
drivers/i2c/busses/i2c-ocores.c +10 -0
+                             "clk_prepare_enable failed: %d\n", ret);
+                     return ret;
+             }
+             i2c->ip_clock_khz = clk_get_rate(i2c->clk) / 1000;
+             if (clock_frequency_present)
+                     i2c->bus_clock_khz = clock_frequency / 1000;
+     } else
+#endif
+             if (of_property_read_u32(np, "opencores,ip-clock-frequency",
+                                      &val)) {

This is not very readable (wrong indentation) and fragile. I'd suggest to leave out #ifdeffery, the compiler will remove the code for !HAVE_CLK because of the empty dummy functions[1]. Then, you can write properly formatted code again.

Quoted 1 line

[1] Well, it should if there wasn't a bug in that: https://lkml.org/lkml/2015/2/5/544

Ok, will fix.

Quoted 6 lines
Quoted 6 lines
drivers/i2c/busses/i2c-ocores.c +6 -0
+#ifdef CONFIG_HAVE_CLK
+     if (!IS_ERR(i2c->clk)) {
+             int ret = clk_prepare_enable(i2c->clk);
+
+             if (ret) {
+                     dev_err(&pdev->dev,
Quoted 8 lines

drivers/i2c/busses/i2c-ocores.c static int ocores_i2c_resume(struct device *dev) · line 540

Quoted 6 lines
drivers/i2c/busses/i2c-ocores.c +6 -0
+                             "clk_prepare_enable failed: %d\n", ret);
+                     return ret;
+             }
+             i2c->ip_clock_khz = clk_get_rate(i2c->clk) / 1000;
+     }
+#endif

Ahem, there is a build error in here :(

Ooops, looks like I don't have CONFIG_PM_SLEEP in my test config. Will fix. Sorry about that.

Signature · 2 lines
Thanks.
-- Max