Skip to contentSkip to conversation

i2c: imx_lpi2c: Fix a typo in bus_i2c_receive

PATCH Superseded [email protected] 2 participants 10 messages Jul 30, 2024 – Aug 7, 2024

Subsystems (2)

v2 of 3 Superseded Latest is v3 →

lore
Share & tools
lore
#0 lore raw

PATCH v2 1/3 i2c: imx_lpi2c: Fix a typo in bus_i2c_receive

From: Fedor Ross <[email protected]>

Fix a typo in a debug message. It should be 'for' not 'fot' .

---
Cc: Heiko Schocher <[email protected]>
Cc: Tom Rini <[email protected]>
Cc: Marek Vasut <[email protected]>
---

Changes in v2:
- Add additional '---' below SoB line to separate Cc list. This way it does not get included in the commit message when the patch is applied.

 drivers/i2c/imx_lpi2c.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/i2c/imx_lpi2c.c +1 -1
index a1be841b119..3c43f61d819 100644
@@ -130,7 +130,7 @@ static int bus_i2c_receive(struct udevice *bus, u8 *rxbuf, int len)
 
 	result = bus_i2c_wait_for_tx_ready(regs);
 	if (result) {
-		debug("i2c: receive wait fot tx ready: %d\n", result);
+		debug("i2c: receive wait for tx ready: %d\n", result);
 		return result;
 	}
 
Signature · 1 line
2.34.1
#1 lore raw

PATCH v2 2/3 i2c: imx_lpi2c: Replace hard-coded bus speed value with bus->speed_hz

From: Fedor Ross <[email protected]>

Instead of using the hard-coded bus speed value I2C_SPEED_STANDARD_RATE, use the actual configured bus speed. This way the bus speed doesn't change suddenly after calling the imx_lpi2c_probe_chip() function for example.

---
Cc: Heiko Schocher <[email protected]>
Cc: Tom Rini <[email protected]>
Cc: Marek Vasut <[email protected]>
---

Changes in v2:
- Add additional '---' below SoB line to separate Cc list. This way it does not get included in the commit message when the patch is applied.
- Drop the second parameter of bus_i2c_init().

 drivers/i2c/imx_lpi2c.c | 15 +++++++++------
 1 file changed, 9 insertions(+), 6 deletions(-)
drivers/i2c/imx_lpi2c.c +9 -6
index 3c43f61d819..54db6fa863f 100644
@@ -19,7 +19,7 @@
 #define LPI2C_NACK_TOUT_MS 1
 #define LPI2C_TIMEOUT_MS 100
 
-static int bus_i2c_init(struct udevice *bus, int speed);
+static int bus_i2c_init(struct udevice *bus);
 
 /* Weak linked function for overridden by some SoC power function */
 int __weak init_i2c_power(unsigned i2c_num)
@@ -172,7 +172,7 @@ static int bus_i2c_start(struct udevice *bus, u8 addr, u8 dir)
 		debug("i2c: start check busy bus: 0x%x\n", result);
 
 		/* Try to init the lpi2c then check the bus busy again */
-		bus_i2c_init(bus, I2C_SPEED_STANDARD_RATE);
+		bus_i2c_init(bus);
 		result = imx_lpci2c_check_busy_bus(regs);
 		if (result) {
 			printf("i2c: Error check busy bus: 0x%x\n", result);
@@ -344,11 +344,14 @@ static int bus_i2c_set_bus_speed(struct udevice *bus, int speed)
 	return 0;
 }
 
-static int bus_i2c_init(struct udevice *bus, int speed)
+static int bus_i2c_init(struct udevice *bus)
 {
 	u32 val;
 	int ret;
 
+	struct dm_i2c_bus *i2c = dev_get_uclass_priv(bus);
+	int speed = i2c->speed_hz;
+
 	struct imx_lpi2c_bus *i2c_bus = dev_get_priv(bus);
 	struct imx_lpi2c_reg *regs = (struct imx_lpi2c_reg *)(i2c_bus->base);
 	/* reset peripheral */
@@ -388,13 +391,13 @@ static int imx_lpi2c_probe_chip(struct udevice *bus, u32 chip,
 	result = bus_i2c_start(bus, chip, 0);
 	if (result) {
 		bus_i2c_stop(bus);
-		bus_i2c_init(bus, I2C_SPEED_STANDARD_RATE);
+		bus_i2c_init(bus);
 		return result;
 	}
 
 	result = bus_i2c_stop(bus);
 	if (result)
-		bus_i2c_init(bus, I2C_SPEED_STANDARD_RATE);
+		bus_i2c_init(bus);
 
 	return result;
 }
@@ -489,7 +492,7 @@ static int imx_lpi2c_probe(struct udevice *bus)
 			return ret;
 	}
 
-	ret = bus_i2c_init(bus, I2C_SPEED_STANDARD_RATE);
+	ret = bus_i2c_init(bus);
 	if (ret < 0)
 		return ret;
 
Signature · 1 line
2.34.1
Reviewed-by on v2 2/3
#2 lore raw

Re: [PATCHv2 2/3] i2c: imx_lpi2c: Replace hard-coded bus speed value with bus->speed_hz

Hello Fedor,

Quoted 14 lines · On 30.07.24 17:25, [email protected] wrote:

From: Fedor Ross <[email protected]>

Instead of using the hard-coded bus speed value I2C_SPEED_STANDARD_RATE, use the actual configured bus speed. This way the bus speed doesn't change suddenly after calling the imx_lpi2c_probe_chip() function for example.

---
Cc: Heiko Schocher <[email protected]>
Cc: Tom Rini <[email protected]>
Cc: Marek Vasut <[email protected]>
---

Quoted 7 lines

Changes in v2:
- Add additional '---' below SoB line to separate Cc list. This way it does not get included in the commit message when the patch is applied.
- Drop the second parameter of bus_i2c_init().

  drivers/i2c/imx_lpi2c.c | 15 +++++++++------
  1 file changed, 9 insertions(+), 6 deletions(-)

Thanks!

bye,
Heiko

Signature · 3 lines
DENX Software Engineering GmbH,      Managing Director: Erika Unter
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: +49-8142-66989-52   Fax: +49-8142-66989-80   Email: [email protected]
#3 lore raw

PATCH v2 3/3 i2c: imx_lpi2c: Support read transfers longer than 256 bytes

In reply to [email protected]

From: Fedor Ross <[email protected]>

The TXFIFO register of LPI2C only has one byte length, and if the length of the data that needs to be read exceeds 256 bytes, it needs to be written to TXFIFO multiple times.

---
Cc: Heiko Schocher <[email protected]>
Cc: Tom Rini <[email protected]>
Cc: Marek Vasut <[email protected]>
---

Changes in v2:
- Add additional '---' below SoB line to separate Cc list. This way it does not get included in the commit message when the patch is applied.
- Prefix 'CHUNK_DATA' macro with 'LPI2C_' to namespace it.
- Do bit manipulation in writel() function directly and drop the 'val = assignment' .

 drivers/i2c/imx_lpi2c.c | 71 ++++++++++++++++++++++++++---------------
 1 file changed, 46 insertions(+), 25 deletions(-)
drivers/i2c/imx_lpi2c.c +46 -25
index 54db6fa863f..a223d91bac3 100644
@@ -19,6 +19,8 @@
 #define LPI2C_NACK_TOUT_MS 1
 #define LPI2C_TIMEOUT_MS 100
 
+#define LPI2C_CHUNK_DATA	256
+
 static int bus_i2c_init(struct udevice *bus);
 
 /* Weak linked function for overridden by some SoC power function */
@@ -118,8 +120,10 @@ static int bus_i2c_send(struct udevice *bus, u8 *txbuf, int len)
 
 static int bus_i2c_receive(struct udevice *bus, u8 *rxbuf, int len)
 {
+	struct dm_i2c_bus *i2c = dev_get_uclass_priv(bus);
 	struct imx_lpi2c_bus *i2c_bus = dev_get_priv(bus);
 	struct imx_lpi2c_reg *regs = (struct imx_lpi2c_reg *)(i2c_bus->base);
+	unsigned int chunk_len, rx_remain, timeout;
 	lpi2c_status_t result = LPI2C_SUCESS;
 	u32 val;
 	ulong start_time = get_timer(0);
@@ -128,33 +132,50 @@ static int bus_i2c_receive(struct udevice *bus, u8 *rxbuf, int len)
 	if (!len)
 		return result;
 
-	result = bus_i2c_wait_for_tx_ready(regs);
-	if (result) {
-		debug("i2c: receive wait for tx ready: %d\n", result);
-		return result;
-	}
+	/*
+	 * Extend the timeout for a bulk read if needed.
+	 * The calculated timeout is the result of multiplying the
+	 * transfer length with 8 bit + ACK + one clock of extra time,
+	 * considering the I2C bus frequency.
+	 */
+	timeout = max(len * 10 * 1000 / i2c->speed_hz, LPI2C_TIMEOUT_MS);
 
-	/* clear all status flags */
-	writel(0x7f00, &regs->msr);
-	/* send receive command */
-	val = LPI2C_MTDR_CMD(0x1) | LPI2C_MTDR_DATA(len - 1);
-	writel(val, &regs->mtdr);
+	rx_remain = len;
+	while (rx_remain > 0) {
+		chunk_len = clamp(rx_remain, 1, LPI2C_CHUNK_DATA) - 1;
 
-	while (len--) {
-		do {
-			result = imx_lpci2c_check_clear_error(regs);
-			if (result) {
-				debug("i2c: receive check clear error: %d\n",
-				      result);
-				return result;
-			}
-			if (get_timer(start_time) > LPI2C_TIMEOUT_MS) {
-				debug("i2c: receive mrdr: timeout\n");
-				return -1;
-			}
-			val = readl(&regs->mrdr);
-		} while (val & LPI2C_MRDR_RXEMPTY_MASK);
-		*rxbuf++ = LPI2C_MRDR_DATA(val);
+		result = bus_i2c_wait_for_tx_ready(regs);
+		if (result) {
+			debug("i2c: receive wait for tx ready: %d\n", result);
+			return result;
+		}
+
+		/* clear all status flags */
+		writel(0x7f00, &regs->msr);
+		/* send receive command */
+		writel(LPI2C_MTDR_CMD(0x1) | LPI2C_MTDR_DATA(chunk_len), &regs->mtdr);
+		rx_remain = rx_remain - (chunk_len & 0xff) - 1;
+
+		while (len--) {
+			do {
+				result = imx_lpci2c_check_clear_error(regs);
+				if (result) {
+					debug("i2c: receive check clear error: %d\n",
+					      result);
+					return result;
+				}
+				if (get_timer(start_time) > timeout) {
+					debug("i2c: receive mrdr: timeout\n");
+					return -1;
+				}
+				val = readl(&regs->mrdr);
+			} while (val & LPI2C_MRDR_RXEMPTY_MASK);
+			*rxbuf++ = LPI2C_MRDR_DATA(val);
+
+			/* send next receive command before controller NACKs last byte */
+			if ((len - rx_remain) < 2 && rx_remain > 0)
+				break;
+		}
 	}
 
 	return result;
Signature · 1 line
2.34.1
Reviewed-by on v2 3/3
#4 lore raw

Re: [PATCHv2 3/3] i2c: imx_lpi2c: Support read transfers longer than 256 bytes

Hello Fedor,

Quoted 20 lines · On 30.07.24 17:25, [email protected] wrote:

From: Fedor Ross <[email protected]>

The TXFIFO register of LPI2C only has one byte length, and if the length of the data that needs to be read exceeds 256 bytes, it needs to be written to TXFIFO multiple times.

---
Cc: Heiko Schocher <[email protected]>
Cc: Tom Rini <[email protected]>
Cc: Marek Vasut <[email protected]>
---

Changes in v2:
- Add additional '---' below SoB line to separate Cc list. This way it does not get included in the commit message when the patch is applied.
- Prefix 'CHUNK_DATA' macro with 'LPI2C_' to namespace it.
- Do bit manipulation in writel() function directly and drop the 'val = assignment' .

Quoted 2 lines
  drivers/i2c/imx_lpi2c.c | 71 ++++++++++++++++++++++++++---------------
  1 file changed, 46 insertions(+), 25 deletions(-)

Thanks!

bye,
Heiko

Signature · 3 lines
DENX Software Engineering GmbH,      Managing Director: Erika Unter
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: +49-8142-66989-52   Fax: +49-8142-66989-80   Email: [email protected]
#5 lore raw

Re: [PATCHv2 3/3] i2c: imx_lpi2c: Support read transfers longer than 256 bytes

In reply to [email protected]

Hello Fedor,

Quoted 20 lines · On 30.07.24 17:25, [email protected] wrote:

From: Fedor Ross <[email protected]>

The TXFIFO register of LPI2C only has one byte length, and if the length of the data that needs to be read exceeds 256 bytes, it needs to be written to TXFIFO multiple times.

---
Cc: Heiko Schocher <[email protected]>
Cc: Tom Rini <[email protected]>
Cc: Marek Vasut <[email protected]>
---

Changes in v2:
- Add additional '---' below SoB line to separate Cc list. This way it does not get included in the commit message when the patch is applied.
- Prefix 'CHUNK_DATA' macro with 'LPI2C_' to namespace it.
- Do bit manipulation in writel() function directly and drop the 'val = assignment' .

Quoted 2 lines
  drivers/i2c/imx_lpi2c.c | 71 ++++++++++++++++++++++++++---------------
  1 file changed, 46 insertions(+), 25 deletions(-)

Patch breaks world build, see for example

https://dev.azure.com/hs0298/hs/_build/results?buildId=114&view=logs&j=9261107f-b16b-528d-034d-a5b366785bfc&t=aa057498-3634-5fad-2ef8-82ed58702aca&l=1197

Find the whole pipeline (with other patches too...)

https://dev.azure.com/hs0298/hs/_build/results?buildId=114&view=results

Could you take a look please?

Thanks!

bye,
Heiko

Signature · 3 lines
DENX Software Engineering GmbH,      Managing Director: Erika Unter
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: +49-8142-66989-52   Fax: +49-8142-66989-80   Email: [email protected]
author
#6 lore raw

Re: [PATCHv2 3/3] i2c: imx_lpi2c: Support read transfers longer than 256 bytes

On Wed, Aug 7, 2024 at 12:50 PM Heiko Schocher <[email protected]> wrote:

Hello Heiko,

Quoted 26 lines · On 30.07.24 17:25, [email protected] wrote:
Quoted 20 lines

From: Fedor Ross <[email protected]>

The TXFIFO register of LPI2C only has one byte length, and if the length of the data that needs to be read exceeds 256 bytes, it needs to be written to TXFIFO multiple times.

---
Cc: Heiko Schocher <[email protected]>
Cc: Tom Rini <[email protected]>
Cc: Marek Vasut <[email protected]>
---

Changes in v2:
- Add additional '---' below SoB line to separate Cc list. This way it does not get included in the commit message when the patch is applied.
- Prefix 'CHUNK_DATA' macro with 'LPI2C_' to namespace it.
- Do bit manipulation in writel() function directly and drop the 'val = assignment' .

Quoted 2 lines
  drivers/i2c/imx_lpi2c.c | 71 ++++++++++++++++++++++++++---------------
  1 file changed, 46 insertions(+), 25 deletions(-)

Patch breaks world build, see for example

Quoted 7 lines

https://dev.azure.com/hs0298/hs/_build/results?buildId=114&view=logs&j=9261107f-b16b-528d-034d-a5b366785bfc&t=aa057498-3634-5fad-2ef8-82ed58702aca&l=1197

Find the whole pipeline (with other patches too...)

https://dev.azure.com/hs0298/hs/_build/results?buildId=114&view=results

Could you take a look please?

Done.
Sorry for the inconvenience.

[...]

Best regards,
Fedor

#7 lore raw

Re: [PATCHv2 3/3] i2c: imx_lpi2c: Support read transfers longer than 256 bytes

Hello Fedor,

Quoted 32 lines · Fedor Ross

On Wed, Aug 7, 2024 at 12:50 PM Heiko Schocher <[email protected]> wrote:

Hello Heiko,

Quoted 28 lines · On 30.07.24 17:25, [email protected] wrote:
Quoted 20 lines

From: Fedor Ross <[email protected]>

The TXFIFO register of LPI2C only has one byte length, and if the length of the data that needs to be read exceeds 256 bytes, it needs to be written to TXFIFO multiple times.

---
Cc: Heiko Schocher <[email protected]>
Cc: Tom Rini <[email protected]>
Cc: Marek Vasut <[email protected]>
---

Changes in v2:
- Add additional '---' below SoB line to separate Cc list. This way it does not get included in the commit message when the patch is applied.
- Prefix 'CHUNK_DATA' macro with 'LPI2C_' to namespace it.
- Do bit manipulation in writel() function directly and drop the 'val = assignment' .

Quoted 2 lines
   drivers/i2c/imx_lpi2c.c | 71 ++++++++++++++++++++++++++---------------
   1 file changed, 46 insertions(+), 25 deletions(-)

Patch breaks world build, see for example

https://dev.azure.com/hs0298/hs/_build/results?buildId=114&view=logs&j=9261107f-b16b-528d-034d-a5b366785bfc&t=aa057498-3634-5fad-2ef8-82ed58702aca&l=1197

Quoted 7 lines
Quoted 5 lines

Find the whole pipeline (with other patches too...)

https://dev.azure.com/hs0298/hs/_build/results?buildId=114&view=results

Could you take a look please?

Done.

Thanks!

Quoted 1 line

Sorry for the inconvenience.

No problem.

bye,
Heiko

Signature · 3 lines
DENX Software Engineering GmbH,      Managing Director: Erika Unter
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: +49-8142-66989-52   Fax: +49-8142-66989-80   Email: [email protected]
Reviewed-by on v2 1/3
#8 lore raw
In reply to [email protected]

Hello Fedor,

Quoted 11 lines · On 30.07.24 17:25, [email protected] wrote:

From: Fedor Ross <[email protected]>

Fix a typo in a debug message. It should be 'for' not 'fot' .

---
Cc: Heiko Schocher <[email protected]>
Cc: Tom Rini <[email protected]>
Cc: Marek Vasut <[email protected]>
---

Quoted 3 lines

Changes in v2:
- Add additional '---' below SoB line to separate Cc list. This way it does not get included in the commit message when the patch is applied.

You may also take a look at patman tool in u-boot:/tools/patman

Thanks!

bye,
Heiko

Signature · 4 lines
-- 
DENX Software Engineering GmbH,      Managing Director: Erika Unter
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: +49-8142-66989-52   Fax: +49-8142-66989-80   Email: [email protected]
author
#9 lore raw
Quoted 2 lines · Heiko Schocher

Hello Fedor,

Hello Heiko,

Quoted 12 lines · On 30.07.24 17:25, [email protected] wrote:
Quoted 11 lines

From: Fedor Ross <[email protected]>

Fix a typo in a debug message. It should be 'for' not 'fot' .

---
Cc: Heiko Schocher <[email protected]>
Cc: Tom Rini <[email protected]>
Cc: Marek Vasut <[email protected]>
---

Quoted 5 lines
Quoted 3 lines

Changes in v2:
- Add additional '---' below SoB line to separate Cc list. This way it does not get included in the commit message when the patch is applied.

You may also take a look at patman tool in u-boot:/tools/patman

I'll take a look at that. Thanks!

Quoted 1 line

[...]

Best regards,
Fedor