Skip to contentSkip to conversation

drivers: tpm2: update reset gpio semantics

PATCH Superseded Jorge Ramirez-Ortiz 2 participants 5 messages Jun 1, 2021 – Jun 3, 2021

v2 of 5 Superseded Latest is v5 →

lore
Share & tools
lore
#0 lore raw

PATCH v2 1/1 drivers: tpm2: update reset gpio semantics

Use the more generic reset-gpios propery name.

---
v2: kept gpio-reset as legacy

 .../tpm2/tis-tpm2-spi.txt                     |  2 +-
 drivers/tpm/tpm2_tis_spi.c                    | 21 ++++++++++++-------
 2 files changed, 14 insertions(+), 9 deletions(-)
doc/device-tree-bindings/tpm2/tis-tpm2-spi.txt +1 -1
index 3a2ee4bd17..bbcd12950f 100644
@@ -6,7 +6,7 @@ Required properties:
 - reg			: SPI Chip select
 
 Optional properties:
-- gpio-reset		: Reset GPIO (if not connected to the SoC reset line)
+- reset-gpios		: Reset GPIO (if not connected to the SoC reset line)
 - spi-max-frequency	: See spi-bus.txt
 
 Example:
drivers/tpm/tpm2_tis_spi.c +13 -8
index 4b33ac8fd3..1f9f89f68f 100644
@@ -589,18 +589,23 @@ static int tpm_tis_spi_probe(struct udevice *dev)
 	if (CONFIG_IS_ENABLED(DM_GPIO)) {
 		struct gpio_desc reset_gpio;
 
-		ret = gpio_request_by_name(dev, "gpio-reset", 0,
+		ret = gpio_request_by_name(dev, "reset-gpios", 0,
 					   &reset_gpio, GPIOD_IS_OUT);
 		if (ret) {
-			log(LOGC_NONE, LOGL_NOTICE, "%s: missing reset GPIO\n",
-			    __func__);
-		} else {
-			dm_gpio_set_value(&reset_gpio, 1);
-			mdelay(1);
-			dm_gpio_set_value(&reset_gpio, 0);
+			/* legacy reset */
+			ret = gpio_request_by_name(dev, "gpio-reset", 0,
+						   &reset_gpio, GPIOD_IS_OUT);
+			if (ret) {
+				log(LOGC_NONE, LOGL_NOTICE,
+				    "%s: missing reset GPIO\n",  __func__);
+				goto init;
+			}
 		}
+		dm_gpio_set_value(&reset_gpio, 1);
+		mdelay(1);
+		dm_gpio_set_value(&reset_gpio, 0);
 	}
-
+init:
 	/* Ensure a minimum amount of time elapsed since reset of the TPM */
 	mdelay(drv_data->time_before_first_cmd_ms);
 
Signature · 1 line
2.31.1
#1 lore raw
Quoted 15 lines · Jorge Ramirez-Ortiz

Use the more generic reset-gpios propery name.

---
v2: kept gpio-reset as legacy

 .../tpm2/tis-tpm2-spi.txt                     |  2 +-
 drivers/tpm/tpm2_tis_spi.c                    | 21 ++++++++++++-------
 2 files changed, 14 insertions(+), 9 deletions(-)
doc/device-tree-bindings/tpm2/tis-tpm2-spi.txt +0 -0
index 3a2ee4bd17..bbcd12950f 100644

doc/device-tree-bindings/tpm2/tis-tpm2-spi.txt Required properties: · line 9

Quoted 4 lines
doc/device-tree-bindings/tpm2/tis-tpm2-spi.txt +0 -1
@@ -6,7 +6,7 @@ Required properties:
 - reg			: SPI Chip select
 
 Optional properties:
-- gpio-reset		: Reset GPIO (if not connected to the SoC reset line)

As I said you shouldn't remove this. Just extend description that it is deprecated and reset-gpios should be used instead.

Quoted 20 lines
doc/device-tree-bindings/tpm2/tis-tpm2-spi.txt +1 -0
+- reset-gpios		: Reset GPIO (if not connected to the SoC reset line)
 - spi-max-frequency	: See spi-bus.txt
 
 Example:
drivers/tpm/tpm2_tis_spi.c +1 -5
index 4b33ac8fd3..1f9f89f68f 100644
@@ -589,18 +589,23 @@ static int tpm_tis_spi_probe(struct udevice *dev)
 	if (CONFIG_IS_ENABLED(DM_GPIO)) {
 		struct gpio_desc reset_gpio;
 
-		ret = gpio_request_by_name(dev, "gpio-reset", 0,
+		ret = gpio_request_by_name(dev, "reset-gpios", 0,
 					   &reset_gpio, GPIOD_IS_OUT);
 		if (ret) {
-			log(LOGC_NONE, LOGL_NOTICE, "%s: missing reset GPIO\n",
-			    __func__);
-		} else {
-			dm_gpio_set_value(&reset_gpio, 1);

drivers/tpm/tpm2_tis_spi.c static int tpm_tis_spi_probe(struct udevice *dev) · line 602

Quoted 10 lines
drivers/tpm/tpm2_tis_spi.c +8 -2
-			mdelay(1);
-			dm_gpio_set_value(&reset_gpio, 0);
+			/* legacy reset */
+			ret = gpio_request_by_name(dev, "gpio-reset", 0,
+						   &reset_gpio, GPIOD_IS_OUT);
+			if (ret) {
+				log(LOGC_NONE, LOGL_NOTICE,
+				    "%s: missing reset GPIO\n",  __func__);
+				goto init;
+			}

And here it is clear that gpio-reset is used which should deprecated that's why you should print message about it here.

drivers/tpm/tpm2_tis_spi.c static int tpm_tis_spi_probe(struct udevice *dev) · line 607

Quoted 5 lines
drivers/tpm/tpm2_tis_spi.c +3 -0
 		}
+		dm_gpio_set_value(&reset_gpio, 1);
+		mdelay(1);
+		dm_gpio_set_value(&reset_gpio, 0);
 	}

What about this to remove that goto?

ret = gpio_request_by_name(dev, "reset-gpios", 0,

			   &reset_gpio, GPIOD_IS_OUT);
if (ret) {
	/* trying legacy reset */
	ret = gpio_request_by_name(dev, "gpio-reset", 0,
				   &reset_gpio, GPIOD_IS_OUT);
	if (!ret)
		log(LOGC_NONE, LOGL_NOTICE,
		    "%s: Using deprecated gpio-reset switch to reset-gpios\n",  __func__);
}
if (!ret) {
	dm_gpio_set_value(&reset_gpio, 1);
	mdelay(1);
	dm_gpio_set_value(&reset_gpio, 0);
}

drivers/tpm/tpm2_tis_spi.c static int tpm_tis_spi_probe(struct udevice *dev) · line 611

Quoted 6 lines
drivers/tpm/tpm2_tis_spi.c +1 -1
-
+init:
 	/* Ensure a minimum amount of time elapsed since reset of the TPM */
 	mdelay(drv_data->time_before_first_cmd_ms);
 

M

#2 lore raw
Quoted 20 lines · Michal Simek
Quoted 17 lines · Jorge Ramirez-Ortiz

Use the more generic reset-gpios propery name.

---
v2: kept gpio-reset as legacy

 .../tpm2/tis-tpm2-spi.txt                     |  2 +-
 drivers/tpm/tpm2_tis_spi.c                    | 21 ++++++++++++-------
 2 files changed, 14 insertions(+), 9 deletions(-)
doc/device-tree-bindings/tpm2/tis-tpm2-spi.txt +0 -0
index 3a2ee4bd17..bbcd12950f 100644
@@ -6,7 +6,7 @@ Required properties:
 - reg			: SPI Chip select
 
Quoted 5 lines

doc/device-tree-bindings/tpm2/tis-tpm2-spi.txt Required properties: · line 9

Quoted 2 lines
doc/device-tree-bindings/tpm2/tis-tpm2-spi.txt +0 -1
 Optional properties:
-- gpio-reset		: Reset GPIO (if not connected to the SoC reset line)

As I said you shouldn't remove this. Just extend description that it is deprecated and reset-gpios should be used instead.

I dont really agree with that. IMO we should remove the documentation since it is obsolete after this commit and anyone reading it should not care about the gpio-reset property.

Quoted 26 lines
Quoted 26 lines
doc/device-tree-bindings/tpm2/tis-tpm2-spi.txt +1 -0
+- reset-gpios		: Reset GPIO (if not connected to the SoC reset line)
 - spi-max-frequency	: See spi-bus.txt
 
 Example:
drivers/tpm/tpm2_tis_spi.c +4 -7
index 4b33ac8fd3..1f9f89f68f 100644
@@ -589,18 +589,23 @@ static int tpm_tis_spi_probe(struct udevice *dev)
 	if (CONFIG_IS_ENABLED(DM_GPIO)) {
 		struct gpio_desc reset_gpio;
 
-		ret = gpio_request_by_name(dev, "gpio-reset", 0,
+		ret = gpio_request_by_name(dev, "reset-gpios", 0,
 					   &reset_gpio, GPIOD_IS_OUT);
 		if (ret) {
-			log(LOGC_NONE, LOGL_NOTICE, "%s: missing reset GPIO\n",
-			    __func__);
-		} else {
-			dm_gpio_set_value(&reset_gpio, 1);
-			mdelay(1);
-			dm_gpio_set_value(&reset_gpio, 0);
+			/* legacy reset */
+			ret = gpio_request_by_name(dev, "gpio-reset", 0,
+						   &reset_gpio, GPIOD_IS_OUT);
Quoted 8 lines

drivers/tpm/tpm2_tis_spi.c static int tpm_tis_spi_probe(struct udevice *dev) · line 602

Quoted 5 lines
drivers/tpm/tpm2_tis_spi.c +5 -0
+			if (ret) {
+				log(LOGC_NONE, LOGL_NOTICE,
+				    "%s: missing reset GPIO\n",  __func__);
+				goto init;
+			}

And here it is clear that gpio-reset is used which should deprecated that's why you should print message about it here.

yes, I can do that. makes sense

Quoted 9 lines

drivers/tpm/tpm2_tis_spi.c static int tpm_tis_spi_probe(struct udevice *dev) · line 607

Quoted 7 lines
drivers/tpm/tpm2_tis_spi.c +3 -0
 		}
+		dm_gpio_set_value(&reset_gpio, 1);
+		mdelay(1);
+		dm_gpio_set_value(&reset_gpio, 0);
 	}

What about this to remove that goto?

um, what is the problem with the goto (IMO tidier than yet another conditional); it is not as if this goto is making the code obscure.

with the change below you just removed previous functionality (ie indicating that there is no GPIO reset provided, hence why at first sight might look cleaner than a goto)

Quoted 26 lines

ret = gpio_request_by_name(dev, "reset-gpios", 0,

			   &reset_gpio, GPIOD_IS_OUT);
if (ret) {
	/* trying legacy reset */
	ret = gpio_request_by_name(dev, "gpio-reset", 0,
				   &reset_gpio, GPIOD_IS_OUT);
	if (!ret)
		log(LOGC_NONE, LOGL_NOTICE,
		    "%s: Using deprecated gpio-reset switch to reset-gpios\n",  __func__);
}
if (!ret) {
	dm_gpio_set_value(&reset_gpio, 1);
	mdelay(1);
	dm_gpio_set_value(&reset_gpio, 0);
}

drivers/tpm/tpm2_tis_spi.c static int tpm_tis_spi_probe(struct udevice *dev) · line 611

Quoted 6 lines
drivers/tpm/tpm2_tis_spi.c +1 -1
-
+init:
 	/* Ensure a minimum amount of time elapsed since reset of the TPM */
 	mdelay(drv_data->time_before_first_cmd_ms);
 

M

#3 lore raw
Quoted 27 lines · Foundries
Quoted 20 lines · Michal Simek
Quoted 17 lines · Jorge Ramirez-Ortiz

Use the more generic reset-gpios propery name.

---
v2: kept gpio-reset as legacy

 .../tpm2/tis-tpm2-spi.txt                     |  2 +-
 drivers/tpm/tpm2_tis_spi.c                    | 21 ++++++++++++-------
 2 files changed, 14 insertions(+), 9 deletions(-)
doc/device-tree-bindings/tpm2/tis-tpm2-spi.txt +0 -0
index 3a2ee4bd17..bbcd12950f 100644
@@ -6,7 +6,7 @@ Required properties:
 - reg			: SPI Chip select
 
Quoted 5 lines

doc/device-tree-bindings/tpm2/tis-tpm2-spi.txt Required properties: · line 9

Quoted 2 lines
doc/device-tree-bindings/tpm2/tis-tpm2-spi.txt +0 -1
 Optional properties:
-- gpio-reset		: Reset GPIO (if not connected to the SoC reset line)

As I said you shouldn't remove this. Just extend description that it is deprecated and reset-gpios should be used instead.

Quoted 3 lines

I dont really agree with that. IMO we should remove the documentation since it is obsolete after this commit and anyone reading it should not care about the gpio-reset property.

Run this on linux kernel and you will see that normal style is to keep it there.
git grep deprecated Documentation/devicetree/bindings

Quoted 48 lines
Quoted 27 lines
Quoted 27 lines
doc/device-tree-bindings/tpm2/tis-tpm2-spi.txt +1 -0
+- reset-gpios		: Reset GPIO (if not connected to the SoC reset line)
 - spi-max-frequency	: See spi-bus.txt
 
 Example:
drivers/tpm/tpm2_tis_spi.c +4 -7
index 4b33ac8fd3..1f9f89f68f 100644
@@ -589,18 +589,23 @@ static int tpm_tis_spi_probe(struct udevice *dev)
 	if (CONFIG_IS_ENABLED(DM_GPIO)) {
 		struct gpio_desc reset_gpio;
 
-		ret = gpio_request_by_name(dev, "gpio-reset", 0,
+		ret = gpio_request_by_name(dev, "reset-gpios", 0,
 					   &reset_gpio, GPIOD_IS_OUT);
 		if (ret) {
-			log(LOGC_NONE, LOGL_NOTICE, "%s: missing reset GPIO\n",
-			    __func__);
-		} else {
-			dm_gpio_set_value(&reset_gpio, 1);
-			mdelay(1);
-			dm_gpio_set_value(&reset_gpio, 0);
+			/* legacy reset */
+			ret = gpio_request_by_name(dev, "gpio-reset", 0,
+						   &reset_gpio, GPIOD_IS_OUT);
Quoted 8 lines

drivers/tpm/tpm2_tis_spi.c static int tpm_tis_spi_probe(struct udevice *dev) · line 602

Quoted 5 lines
drivers/tpm/tpm2_tis_spi.c +5 -0
+			if (ret) {
+				log(LOGC_NONE, LOGL_NOTICE,
+				    "%s: missing reset GPIO\n",  __func__);
+				goto init;
+			}

And here it is clear that gpio-reset is used which should deprecated that's why you should print message about it here.

yes, I can do that. makes sense

Quoted 9 lines

drivers/tpm/tpm2_tis_spi.c static int tpm_tis_spi_probe(struct udevice *dev) · line 607

Quoted 7 lines
drivers/tpm/tpm2_tis_spi.c +3 -0
 		}
+		dm_gpio_set_value(&reset_gpio, 1);
+		mdelay(1);
+		dm_gpio_set_value(&reset_gpio, 0);
 	}

What about this to remove that goto?

Quoted 6 lines

um, what is the problem with the goto (IMO tidier than yet another conditional); it is not as if this goto is making the code obscure.

with the change below you just removed previous functionality (ie indicating that there is no GPIO reset provided, hence why at first sight might look cleaner than a goto)

I tend to use goto unless there is no way around. But up2you.

M

#4 lore raw
Quoted 30 lines · Michal Simek
Quoted 27 lines · Foundries
Quoted 25 lines · Michal Simek
Quoted 15 lines · Jorge Ramirez-Ortiz

Use the more generic reset-gpios propery name.

---
v2: kept gpio-reset as legacy

 .../tpm2/tis-tpm2-spi.txt                     |  2 +-
 drivers/tpm/tpm2_tis_spi.c                    | 21 ++++++++++++-------
 2 files changed, 14 insertions(+), 9 deletions(-)
doc/device-tree-bindings/tpm2/tis-tpm2-spi.txt +0 -0
index 3a2ee4bd17..bbcd12950f 100644

doc/device-tree-bindings/tpm2/tis-tpm2-spi.txt Required properties: · line 9

Quoted 4 lines
doc/device-tree-bindings/tpm2/tis-tpm2-spi.txt +0 -1
@@ -6,7 +6,7 @@ Required properties:
 - reg			: SPI Chip select
 
 Optional properties:
-- gpio-reset		: Reset GPIO (if not connected to the SoC reset line)

As I said you shouldn't remove this. Just extend description that it is deprecated and reset-gpios should be used instead.

Quoted 7 lines
Quoted 3 lines

I dont really agree with that. IMO we should remove the documentation since it is obsolete after this commit and anyone reading it should not care about the gpio-reset property.

Run this on linux kernel and you will see that normal style is to keep it there.
git grep deprecated Documentation/devicetree/bindings

um, ok then (although IMO it only makes sense when we have the tools to parse such documents and apply them at build time)...if you have to read the docs you might just as well change the property.

will post later today (I am also adding an additional test to tpm2 to read the rng)

Quoted 60 lines
Quoted 49 lines
Quoted 28 lines
Quoted 28 lines
doc/device-tree-bindings/tpm2/tis-tpm2-spi.txt +1 -0
+- reset-gpios		: Reset GPIO (if not connected to the SoC reset line)
 - spi-max-frequency	: See spi-bus.txt
 
 Example:
drivers/tpm/tpm2_tis_spi.c +4 -7
index 4b33ac8fd3..1f9f89f68f 100644
@@ -589,18 +589,23 @@ static int tpm_tis_spi_probe(struct udevice *dev)
 	if (CONFIG_IS_ENABLED(DM_GPIO)) {
 		struct gpio_desc reset_gpio;
 
-		ret = gpio_request_by_name(dev, "gpio-reset", 0,
+		ret = gpio_request_by_name(dev, "reset-gpios", 0,
 					   &reset_gpio, GPIOD_IS_OUT);
 		if (ret) {
-			log(LOGC_NONE, LOGL_NOTICE, "%s: missing reset GPIO\n",
-			    __func__);
-		} else {
-			dm_gpio_set_value(&reset_gpio, 1);
-			mdelay(1);
-			dm_gpio_set_value(&reset_gpio, 0);
+			/* legacy reset */
+			ret = gpio_request_by_name(dev, "gpio-reset", 0,
+						   &reset_gpio, GPIOD_IS_OUT);
Quoted 8 lines

drivers/tpm/tpm2_tis_spi.c static int tpm_tis_spi_probe(struct udevice *dev) · line 602

Quoted 5 lines
drivers/tpm/tpm2_tis_spi.c +5 -0
+			if (ret) {
+				log(LOGC_NONE, LOGL_NOTICE,
+				    "%s: missing reset GPIO\n",  __func__);
+				goto init;
+			}

And here it is clear that gpio-reset is used which should deprecated that's why you should print message about it here.

yes, I can do that. makes sense

Quoted 9 lines

drivers/tpm/tpm2_tis_spi.c static int tpm_tis_spi_probe(struct udevice *dev) · line 607

Quoted 7 lines
drivers/tpm/tpm2_tis_spi.c +3 -0
 		}
+		dm_gpio_set_value(&reset_gpio, 1);
+		mdelay(1);
+		dm_gpio_set_value(&reset_gpio, 0);
 	}

What about this to remove that goto?

Quoted 6 lines

um, what is the problem with the goto (IMO tidier than yet another conditional); it is not as if this goto is making the code obscure.

with the change below you just removed previous functionality (ie indicating that there is no GPIO reset provided, hence why at first sight might look cleaner than a goto)

I tend to use goto unless there is no way around. But up2you.

M