Skip to contentSkip to conversation

i2c: designware: fix master is holding SCL low while ENABLE bit is disabled

PATCH Superseded Kimriver Liu 4 participants 7 messages Sep 6, 2024 – Sep 8, 2024

Subsystems (3)

v1 of 11 Superseded Latest is v11 →

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 i2c: designware: fix master is holding SCL low while ENABLE bit is disabled

It was observed issuing ABORT bit(IC_ENABLE[1]) will not work when IC_ENABLE is already disabled.

Check if ENABLE bit(IC_ENABLE[0]) is disabled when the master is holding SCL low. If ENABLE bit is disabled, the software need enable it before trying to issue ABORT bit. otherwise, the controller ignores any write to ABORT bit.

---
V5->V6: restore i2c_dw_is_master_idling() function checking V4->V5: delete master idling checking
V3->V4:
1. update commit messages and add patch version and changelog
2. move print the error message in i2c_dw_xfer
V2->V3: change (!enable) to (!(enable & DW_IC_ENABLE_ENABLE)) V1->V2: used standard words in function names and addressed review comments

link to V1:
https://lore.kernel.org/lkml/[email protected]/

---
 drivers/i2c/busses/i2c-designware-common.c | 11 +++++++++++
 drivers/i2c/busses/i2c-designware-master.c | 22 ++++++++++++++++++++++
 2 files changed, 33 insertions(+)
drivers/i2c/busses/i2c-designware-common.c +11 -0
index e8a688d04aee..2b3398cd4382 100644
@@ -453,6 +453,17 @@ void __i2c_dw_disable(struct dw_i2c_dev *dev)
 
 	abort_needed = raw_intr_stats & DW_IC_INTR_MST_ON_HOLD;
 	if (abort_needed) {
+		if (!(enable & DW_IC_ENABLE_ENABLE)) {
+			regmap_write(dev->map, DW_IC_ENABLE, DW_IC_ENABLE_ENABLE);
+			enable |= DW_IC_ENABLE_ENABLE;
+			/*
+			 * Wait two ic_clk delay when enabling the I2C to ensure ENABLE bit
+			 * is already set by the driver (for 400KHz this is 25us)
+			 * as described in the DesignWare I2C databook.
+			 */
+			fsleep(25);
+		}
+
 		regmap_write(dev->map, DW_IC_ENABLE, enable | DW_IC_ENABLE_ABORT);
 		ret = regmap_read_poll_timeout(dev->map, DW_IC_ENABLE, enable,
 					       !(enable & DW_IC_ENABLE_ABORT), 10,
drivers/i2c/busses/i2c-designware-master.c +22 -0
index c7e56002809a..132b7237c004 100644
@@ -253,6 +253,19 @@ static void i2c_dw_xfer_init(struct dw_i2c_dev *dev)
 	__i2c_dw_write_intr_mask(dev, DW_IC_INTR_MASTER_MASK);
 }
 
+static bool i2c_dw_is_master_idling(struct dw_i2c_dev *dev)
+{
+	u32 status;
+
+	regmap_read(dev->map, DW_IC_STATUS, &status);
+	if (!(status & DW_IC_STATUS_MASTER_ACTIVITY))
+		return true;
+
+	return !regmap_read_poll_timeout(dev->map, DW_IC_STATUS, status,
+			!(status & DW_IC_STATUS_MASTER_ACTIVITY),
+			1100, 20000);
+}
+
 static int i2c_dw_check_stopbit(struct dw_i2c_dev *dev)
 {
 	u32 val;
@@ -788,6 +801,15 @@ i2c_dw_xfer(struct i2c_adapter *adap, struct i2c_msg msgs[], int num)
 		goto done;
 	}
 
+	/*
+	 * This happens rarely and is hard to reproduce. Debug trace
+	 * showed that IC_STATUS had value of 0x23 when STOP_DET occurred,
+	 * if disable IC_ENABLE.ENABLE immediately that can result in
+	 * IC_RAW_INTR_STAT.MASTER_ON_HOLD holding SCL low.
+	 */
+	if (!i2c_dw_is_master_idling(dev))
+		dev_err(dev->dev, "I2C master not idling\n");
+
 	/*
 	 * We must disable the adapter before returning and signaling the end
 	 * of the current transfer. Otherwise the hardware might continue
Signature · 1 line
2.17.1
maintainer Reviewed-by on v1 1/1
#1 lore raw

Hi,

Quoted 58 lines · Kimriver Liu

It was observed issuing ABORT bit(IC_ENABLE[1]) will not work when IC_ENABLE is already disabled.

Check if ENABLE bit(IC_ENABLE[0]) is disabled when the master is holding SCL low. If ENABLE bit is disabled, the software need enable it before trying to issue ABORT bit. otherwise, the controller ignores any write to ABORT bit.

---
V5->V6: restore i2c_dw_is_master_idling() function checking V4->V5: delete master idling checking
V3->V4:
1. update commit messages and add patch version and changelog
2. move print the error message in i2c_dw_xfer
V2->V3: change (!enable) to (!(enable & DW_IC_ENABLE_ENABLE)) V1->V2: used standard words in function names and addressed review comments

link to V1:
https://lore.kernel.org/lkml/[email protected]/

---
 drivers/i2c/busses/i2c-designware-common.c | 11 +++++++++++
 drivers/i2c/busses/i2c-designware-master.c | 22 ++++++++++++++++++++++
 2 files changed, 33 insertions(+)
drivers/i2c/busses/i2c-designware-common.c +11 -0
index e8a688d04aee..2b3398cd4382 100644
@@ -453,6 +453,17 @@ void __i2c_dw_disable(struct dw_i2c_dev *dev)
 
 	abort_needed = raw_intr_stats & DW_IC_INTR_MST_ON_HOLD;
 	if (abort_needed) {
+		if (!(enable & DW_IC_ENABLE_ENABLE)) {
+			regmap_write(dev->map, DW_IC_ENABLE, DW_IC_ENABLE_ENABLE);
+			enable |= DW_IC_ENABLE_ENABLE;
+			/*
+			 * Wait two ic_clk delay when enabling the I2C to ensure ENABLE bit
+			 * is already set by the driver (for 400KHz this is 25us)
+			 * as described in the DesignWare I2C databook.
+			 */
+			fsleep(25);
+		}
+
 		regmap_write(dev->map, DW_IC_ENABLE, enable | DW_IC_ENABLE_ABORT);
 		ret = regmap_read_poll_timeout(dev->map, DW_IC_ENABLE, enable,
 					       !(enable & DW_IC_ENABLE_ABORT), 10,
drivers/i2c/busses/i2c-designware-master.c +2 -0
index c7e56002809a..132b7237c004 100644
@@ -253,6 +253,19 @@ static void i2c_dw_xfer_init(struct dw_i2c_dev *dev)
 	__i2c_dw_write_intr_mask(dev, DW_IC_INTR_MASTER_MASK);
 }
 
+static bool i2c_dw_is_master_idling(struct dw_i2c_dev *dev)
+{

drivers/i2c/busses/i2c-designware-master.c static void i2c_dw_xfer_init(struct dw_i2c_dev *dev) · line 267

Quoted 10 lines
drivers/i2c/busses/i2c-designware-master.c +10 -0
+	u32 status;
+
+	regmap_read(dev->map, DW_IC_STATUS, &status);
+	if (!(status & DW_IC_STATUS_MASTER_ACTIVITY))
+		return true;
+
+	return !regmap_read_poll_timeout(dev->map, DW_IC_STATUS, status,
+			!(status & DW_IC_STATUS_MASTER_ACTIVITY),
+			1100, 20000);
+}

Yeah, I now realize that i2c_dw_wait_bus_not_busy() checks for DW_IC_STATUS_ACTIVITY not for DW_IC_STATUS_MASTER_ACTIVITY as I thought so consolidating them makes not that much sense.

This looks good to me,

#2 lore raw

Hi
I am sorry for not replying to questions in time on Friday.

-----邮件原件-----
发件人: Mika Westerberg <[email protected]>
发送时间: 2024年9月6日 16:08
收件人: Liu Kimriver/刘金河 <[email protected]>
抄送: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected] 主题: Re: [PATCH] i2c: designware: fix master is holding SCL low while ENABLE bit is disabled

Hi,

Quoted 73 lines · Kimriver Liu

It was observed issuing ABORT bit(IC_ENABLE[1]) will not work when IC_ENABLE is already disabled.

Check if ENABLE bit(IC_ENABLE[0]) is disabled when the master is holding SCL low. If ENABLE bit is disabled, the software need enable it before trying to issue ABORT bit. otherwise, the controller ignores any write to ABORT bit.

---
V5->V6: restore i2c_dw_is_master_idling() function checking V4->V5: delete master idling checking
V3->V4:
1. update commit messages and add patch version and changelog
2. move print the error message in i2c_dw_xfer
V2->V3: change (!enable) to (!(enable & DW_IC_ENABLE_ENABLE)) V1->V2: used standard words in function names and addressed review comments

link to V1:
https://lore.kernel.org/lkml/[email protected]/

---
 drivers/i2c/busses/i2c-designware-common.c | 11 +++++++++++
 drivers/i2c/busses/i2c-designware-master.c | 22 ++++++++++++++++++++++
 2 files changed, 33 insertions(+)
drivers/i2c/busses/i2c-designware-common.c +11 -0
index e8a688d04aee..2b3398cd4382 100644
@@ -453,6 +453,17 @@ void __i2c_dw_disable(struct dw_i2c_dev *dev)
 
 	abort_needed = raw_intr_stats & DW_IC_INTR_MST_ON_HOLD;
 	if (abort_needed) {
+		if (!(enable & DW_IC_ENABLE_ENABLE)) {
+			regmap_write(dev->map, DW_IC_ENABLE, DW_IC_ENABLE_ENABLE);
+			enable |= DW_IC_ENABLE_ENABLE;
+			/*
+			 * Wait two ic_clk delay when enabling the I2C to ensure ENABLE bit
+			 * is already set by the driver (for 400KHz this is 25us)
+			 * as described in the DesignWare I2C databook.
+			 */
+			fsleep(25);
+		}
+
 		regmap_write(dev->map, DW_IC_ENABLE, enable | DW_IC_ENABLE_ABORT);
 		ret = regmap_read_poll_timeout(dev->map, DW_IC_ENABLE, enable,
 					       !(enable & DW_IC_ENABLE_ABORT), 10,
drivers/i2c/busses/i2c-designware-master.c +12 -0
index c7e56002809a..132b7237c004 100644
@@ -253,6 +253,19 @@ static void i2c_dw_xfer_init(struct dw_i2c_dev *dev)
 	__i2c_dw_write_intr_mask(dev, DW_IC_INTR_MASTER_MASK);
 }
 
+static bool i2c_dw_is_master_idling(struct dw_i2c_dev *dev)
+{
+	u32 status;
+
+	regmap_read(dev->map, DW_IC_STATUS, &status);
+	if (!(status & DW_IC_STATUS_MASTER_ACTIVITY))
+		return true;
+
+	return !regmap_read_poll_timeout(dev->map, DW_IC_STATUS, status,
+			!(status & DW_IC_STATUS_MASTER_ACTIVITY),
+			1100, 20000);
+}
 

Yeah, I now realize that i2c_dw_wait_bus_not_busy() checks for DW_IC_STATUS_ACTIVITY not for DW_IC_STATUS_MASTER_ACTIVITY as I thought so consolidating them makes not that much sense.

Quoted 1 line

This looks good to me,

Thanks.
This case happens rarely and is hard to reproduce. We reproduce this issue in RTL simulation. It is necessary to add waiting DW_IC_STATUS_MASTER_ACTIVITY idling before disabling I2C when I2C transfer completed. as described in the DesignWare I2C databook(Flowchart for DW_apb_i2c Controller)

-----------------
Best Regards
Kimriver Liu

#3 lore raw
In reply to Kimriver Liu
Quoted 7 lines · Kimriver Liu

It was observed issuing ABORT bit(IC_ENABLE[1]) will not work when IC_ENABLE is already disabled.

Check if ENABLE bit(IC_ENABLE[0]) is disabled when the master is holding SCL low. If ENABLE bit is disabled, the software need enable it before trying to issue ABORT bit. otherwise, the controller ignores any write to ABORT bit.

This is *still* version 1!

Signature · 2 lines
With Best Regards,
Andy Shevchenko
maintainer
#4 lore raw
In reply to Kimriver Liu

Hi Kimriver,

Quoted 19 lines · Kimriver Liu

It was observed issuing ABORT bit(IC_ENABLE[1]) will not work when IC_ENABLE is already disabled.

Check if ENABLE bit(IC_ENABLE[0]) is disabled when the master is holding SCL low. If ENABLE bit is disabled, the software need enable it before trying to issue ABORT bit. otherwise, the controller ignores any write to ABORT bit.

---
V5->V6: restore i2c_dw_is_master_idling() function checking V4->V5: delete master idling checking
V3->V4:
1. update commit messages and add patch version and changelog
2. move print the error message in i2c_dw_xfer
V2->V3: change (!enable) to (!(enable & DW_IC_ENABLE_ENABLE)) V1->V2: used standard words in function names and addressed review comments

Quoted 2 lines

link to V1:
https://lore.kernel.org/lkml/[email protected]/

You've ignored all of Andy's reviews, and this is not acceptable.

You've submitted this patch three times today in a short period, and Andy has responded to each one with the same comment, which you continue to disregard.

You must follow up on the reviews. I will accept this patch for now, but if you keep ignoring feedback, I will reject future patches until you address all reviews from those taking their time to review your changes.

Andi

maintainer
#5 lore raw
In reply to Kimriver Liu

Hi Kimriver,

...

drivers/i2c/busses/i2c-designware-common.c void __i2c_dw_disable(struct dw_i2c_dev *dev) · line 457

Quoted 10 lines
drivers/i2c/busses/i2c-designware-common.c +2 -0
index e8a688d04aee..2b3398cd4382 100644
@@ -453,6 +453,17 @@ void __i2c_dw_disable(struct dw_i2c_dev *dev)
 
 	abort_needed = raw_intr_stats & DW_IC_INTR_MST_ON_HOLD;
 	if (abort_needed) {
+		if (!(enable & DW_IC_ENABLE_ENABLE)) {
+			regmap_write(dev->map, DW_IC_ENABLE, DW_IC_ENABLE_ENABLE);

BTW, your patch doesn't compile. Please make sure that you have everything in place and please resend.

This time I expect you to follow Andy's suggestion.

Thanks,
Andi

#6 lore raw

Hi Andi
I am sorry for not replying to questions in time, when I left the office early on Friday. I sincerely apologize to you again.

I will improve my response progress and follow Andy's suggestion. Thanks. And will resend new patch V7 on Moday:
[PATCH V7] i2c: designware: fix master is holding SCL low while ENABLE bit is disabled

-----邮件原件-----
发件人: Andi Shyti <[email protected]>
发送时间: 2024年9月7日 5:35
收件人: Liu Kimriver/刘金河 <[email protected]>
抄送: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected] 主题: Re: [PATCH] i2c: designware: fix master is holding SCL low while ENABLE bit is disabled

Hi Kimriver,

...

Quoted 14 lines
drivers/i2c/busses/i2c-designware-common.c +2 -0
index e8a688d04aee..2b3398cd4382 100644
@@ -453,6 +453,17 @@ void __i2c_dw_disable(struct dw_i2c_dev *dev)
 
 	abort_needed = raw_intr_stats & DW_IC_INTR_MST_ON_HOLD;
 	if (abort_needed) {
+		if (!(enable & DW_IC_ENABLE_ENABLE)) {
+			regmap_write(dev->map, DW_IC_ENABLE, DW_IC_ENABLE_ENABLE);
 

BTW, your patch doesn't compile. Please make sure that you have everything in place and please resend.

Quoted 1 line

This time I expect you to follow Andy's suggestion.

Sorry, I forget to merge DW_IC_ENABLE_ENABLE patch to here. I will update V7 patch on Monday, [PATCH V7] i2c: designware: fix master is holding SCL low while ENABLE bit is disabled I will update V7 patch on Monday which includes the following issues:
-------
V6->V7:
1. add Subject versioning
2. change fsleep(25) to usleep_range(25, 250)
3. Add macro definition DW_iC_ENABLE_ENABLE for compiling error
------------------
Best Regards
Kimriver Liu