Skip to contentSkip to conversation

media: ov5640: Fix initial RESETB state and annotate timings

PATCH In review Marek Vasut 3 participants 3 messages Jul 24, 2023 – Aug 3, 2023

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

PATCH 1/1 media: ov5640: Fix initial RESETB state and annotate timings

The initial state of RESETB input signal of OV5640 should be asserted, i.e. the sensor should be in reset. This is not the case, make it so.

Since the subsequent assertion of RESETB signal is no longer necessary and the timing of the power sequencing could be slightly adjusted, add annotations to the delays which match OV5640 datasheet rev. 2.03, both:
figure 2-3 power up timing with internal DVDD figure 2-4 power up timing with external DVDD source

The 5..10ms delay between PWDN assertion and RESETB assertion is not even documented in the power sequencing diagram, and with this reset fix, it is no longer even necessary.

Fixes 19a81c1426c1 ("[media] add Omnivision OV5640 sensor driver")

---
Cc: Hans Verkuil <[email protected]>
Cc: Jacopo Mondi <[email protected]>
Cc: Mauro Carvalho Chehab <[email protected]>
Cc: Sakari Ailus <[email protected]>
Cc: Steve Longerbeam <[email protected]>
Cc: [email protected]

---
 drivers/media/i2c/ov5640.c | 11 ++++-------
 1 file changed, 4 insertions(+), 7 deletions(-)
drivers/media/i2c/ov5640.c +4 -7
index 7c065c39082dd..74b58380b5e69 100644
@@ -2452,16 +2452,13 @@ static void ov5640_power(struct ov5640_dev *sensor, bool enable)
 static void ov5640_powerup_sequence(struct ov5640_dev *sensor)
 {
 	if (sensor->pwdn_gpio) {
-		gpiod_set_value_cansleep(sensor->reset_gpio, 0);
+		gpiod_set_value_cansleep(sensor->reset_gpio, 1);
 
 		/* camera power cycle */
 		ov5640_power(sensor, false);
-		usleep_range(5000, 10000);
+		usleep_range(5000, 10000);	/* t2 */
 		ov5640_power(sensor, true);
-		usleep_range(5000, 10000);
-
-		gpiod_set_value_cansleep(sensor->reset_gpio, 1);
-		usleep_range(1000, 2000);
+		usleep_range(1000, 2000);	/* t3 */
 
 		gpiod_set_value_cansleep(sensor->reset_gpio, 0);
 	} else {
@@ -2469,7 +2466,7 @@ static void ov5640_powerup_sequence(struct ov5640_dev *sensor)
 		ov5640_write_reg(sensor, OV5640_REG_SYS_CTRL0,
 				 OV5640_REG_SYS_CTRL0_SW_RST);
 	}
-	usleep_range(20000, 25000);
+	usleep_range(20000, 25000);	/* t4 */
 
 	/*
 	 * software standby: allows registers programming;
Signature · 1 line
2.40.1
Reviewed-by on v1 1/1
#1 lore raw

Cc Jai which has recently changed this part to accomodate a design where RESETB and PWDN are not exposed as separate lines

Quoted 34 lines · Marek Vasut

The initial state of RESETB input signal of OV5640 should be asserted, i.e. the sensor should be in reset. This is not the case, make it so.

Since the subsequent assertion of RESETB signal is no longer necessary and the timing of the power sequencing could be slightly adjusted, add annotations to the delays which match OV5640 datasheet rev. 2.03, both:
figure 2-3 power up timing with internal DVDD figure 2-4 power up timing with external DVDD source

The 5..10ms delay between PWDN assertion and RESETB assertion is not even documented in the power sequencing diagram, and with this reset fix, it is no longer even necessary.

Fixes 19a81c1426c1 ("[media] add Omnivision OV5640 sensor driver")

---
Cc: Hans Verkuil <[email protected]>
Cc: Jacopo Mondi <[email protected]>
Cc: Mauro Carvalho Chehab <[email protected]>
Cc: Sakari Ailus <[email protected]>
Cc: Steve Longerbeam <[email protected]>
Cc: [email protected]

---
 drivers/media/i2c/ov5640.c | 11 ++++-------
 1 file changed, 4 insertions(+), 7 deletions(-)
drivers/media/i2c/ov5640.c +0 -0
index 7c065c39082dd..74b58380b5e69 100644

drivers/media/i2c/ov5640.c static void ov5640_power(struct ov5640_dev *sensor, bool enable) · line 2458

Quoted 8 lines
drivers/media/i2c/ov5640.c +1 -1
@@ -2452,16 +2452,13 @@ static void ov5640_power(struct ov5640_dev *sensor, bool enable)
 static void ov5640_powerup_sequence(struct ov5640_dev *sensor)
 {
 	if (sensor->pwdn_gpio) {
-		gpiod_set_value_cansleep(sensor->reset_gpio, 0);
+		gpiod_set_value_cansleep(sensor->reset_gpio, 1);
 
 		/* camera power cycle */
 		ov5640_power(sensor, false);

This is also probably a NOP for most designs but the one Jai had, as in his case a single "powerdown" line controls both PWDN and RESETB with some internal circuitry that handles signals inversions and delays. So I guess it is fine to have it here.

Quoted 12 lines
drivers/media/i2c/ov5640.c +2 -5
-		usleep_range(5000, 10000);
+		usleep_range(5000, 10000);	/* t2 */
 		ov5640_power(sensor, true);
-		usleep_range(5000, 10000);
-
-		gpiod_set_value_cansleep(sensor->reset_gpio, 1);
-		usleep_range(1000, 2000);
+		usleep_range(1000, 2000);	/* t3 */
 
 		gpiod_set_value_cansleep(sensor->reset_gpio, 0);
 	} else {

drivers/media/i2c/ov5640.c static void ov5640_powerup_sequence(struct ov5640_dev *sensor) · line 2469

Quoted 5 lines
drivers/media/i2c/ov5640.c +1 -1
@@ -2469,7 +2466,7 @@ static void ov5640_powerup_sequence(struct ov5640_dev *sensor)
 		ov5640_write_reg(sensor, OV5640_REG_SYS_CTRL0,
 				 OV5640_REG_SYS_CTRL0_SW_RST);
 	}
-	usleep_range(20000, 25000);
+	usleep_range(20000, 25000);	/* t4 */

This all looks good to me

Jai could you give this a run to make sure this works on your setup as well ?

Thanks
j

Quoted 6 lines
drivers/media/i2c/ov5640.c +0 -1
 	/*
 	 * software standby: allows registers programming;
--

2.40.1

Tested-by
#2 lore raw

Hi Marek, Jacopo,

Quoted 44 lines · Jacopo Mondi

Cc Jai which has recently changed this part to accomodate a design where RESETB and PWDN are not exposed as separate lines

Quoted 40 lines · Marek Vasut

The initial state of RESETB input signal of OV5640 should be asserted, i.e. the sensor should be in reset. This is not the case, make it so.

Since the subsequent assertion of RESETB signal is no longer necessary and the timing of the power sequencing could be slightly adjusted, add annotations to the delays which match OV5640 datasheet rev. 2.03, both:
figure 2-3 power up timing with internal DVDD figure 2-4 power up timing with external DVDD source

The 5..10ms delay between PWDN assertion and RESETB assertion is not even documented in the power sequencing diagram, and with this reset fix, it is no longer even necessary.

Fixes 19a81c1426c1 ("[media] add Omnivision OV5640 sensor driver")

---
Cc: Hans Verkuil <[email protected]>
Cc: Jacopo Mondi <[email protected]>
Cc: Mauro Carvalho Chehab <[email protected]>
Cc: Sakari Ailus <[email protected]>
Cc: Steve Longerbeam <[email protected]>
Cc: [email protected]

---
 drivers/media/i2c/ov5640.c | 11 ++++-------
 1 file changed, 4 insertions(+), 7 deletions(-)
drivers/media/i2c/ov5640.c +1 -1
index 7c065c39082dd..74b58380b5e69 100644
@@ -2452,16 +2452,13 @@ static void ov5640_power(struct ov5640_dev *sensor, bool enable)
 static void ov5640_powerup_sequence(struct ov5640_dev *sensor)
 {
 	if (sensor->pwdn_gpio) {
-		gpiod_set_value_cansleep(sensor->reset_gpio, 0);
+		gpiod_set_value_cansleep(sensor->reset_gpio, 1);
 
Quoted 8 lines

drivers/media/i2c/ov5640.c static void ov5640_power(struct ov5640_dev *sensor, bool enable) · line 2458

Quoted 2 lines
drivers/media/i2c/ov5640.c +0 -0
 		/* camera power cycle */
 		ov5640_power(sensor, false);

This is also probably a NOP for most designs but the one Jai had, as in his case a single "powerdown" line controls both PWDN and RESETB with some internal circuitry that handles signals inversions and delays. So I guess it is fine to have it here.

Yes the modules in question are -
- Digilent PCam 5C
- ALINX AN5641
- TEVI-OV5640-*-RPI

All of them have a 15-pin FFC (RPi) compatible connector.

Quoted 4 lines

drivers/media/i2c/ov5640.c static void ov5640_power(struct ov5640_dev *sensor, bool enable) · line 2461

Quoted 4 lines
drivers/media/i2c/ov5640.c +1 -2
-		usleep_range(5000, 10000);
+		usleep_range(5000, 10000);	/* t2 */
 		ov5640_power(sensor, true);
-		usleep_range(5000, 10000);

I remember noticing this before, but chose to keep it around as the above mentioned modules have a 5.5ms + 1ms delay mentioned in the schematics [1] between toggling the weird "PWUP" gpio to the RESET actually getting toggled.

But in my tests your patch is not breaking any of those modules, and this delay does look very redundant from driver p.o.v. so I am happy with removing it here.

In future if I face any issues I will try to come up with a better approach that does not add redundant delays for modules that don't perform magical delays on-board like those three special ones :)

[1]
https://digilent.com/reference/_media/reference/add-ons/pcam-5c/pcam_5c_sch.pdf

Quoted 16 lines

drivers/media/i2c/ov5640.c static void ov5640_powerup_sequence(struct ov5640_dev *sensor) · line 2469

Quoted 13 lines
drivers/media/i2c/ov5640.c +2 -4
-
-		gpiod_set_value_cansleep(sensor->reset_gpio, 1);
-		usleep_range(1000, 2000);
+		usleep_range(1000, 2000);	/* t3 */
 
 		gpiod_set_value_cansleep(sensor->reset_gpio, 0);
 	} else {
@@ -2469,7 +2466,7 @@ static void ov5640_powerup_sequence(struct ov5640_dev *sensor)
 		ov5640_write_reg(sensor, OV5640_REG_SYS_CTRL0,
 				 OV5640_REG_SYS_CTRL0_SW_RST);
 	}
-	usleep_range(20000, 25000);
+	usleep_range(20000, 25000);	/* t4 */

This all looks good to me

Quoted 5 lines

Jai could you give this a run to make sure this works on your setup as well ?

Works on BeaglePlay and SK-AM62 (with my CSI v8 series applied on next)

Quoted 10 lines

Thanks
j

Quoted 6 lines
drivers/media/i2c/ov5640.c +0 -1
 	/*
 	 * software standby: allows registers programming;
--

2.40.1

Signature · 4 lines
Thanks,
Jai

GPG Fingerprint: 4DE0 D818 E5D5 75E8 D45A AFC5 43DE 91F9 249A 7145