Skip to contentSkip to conversation

fix iproc driver to handle master read request

PATCH Superseded Rayagonda Kokatanur 3 participants 19 messages Oct 11, 2020 – Oct 27, 2020 1 message not in the archive

Subsystems (2)

v1 of 3 Superseded Latest is v3 →

lore
Share & tools
lore
#0 lore raw

cover fix iproc driver to handle master read request

This series of patches adds the following,
- Handle master abort error
- Fix support for single/multi byte master read request with/without repeated start.
- Handle rx fifo full interrupt
- Fix typo

Rayagonda Kokatanur (6):
i2c: iproc: handle Master aborted error
i2c: iproc: handle only slave interrupts which are enabled
i2c: iproc: update slave isr mask (ISR_MASK_SLAVE)
i2c: iproc: fix typo in slave_isr function
i2c: iproc: handle master read request
i2c: iproc: handle rx fifo full interrupt

 drivers/i2c/busses/i2c-bcm-iproc.c | 254 +++++++++++++++++++++++------
 1 file changed, 200 insertions(+), 54 deletions(-)
Signature · 1 line
2.17.1
#1 lore raw

PATCH 1/6 i2c: iproc: handle Master aborted error

Handle Master aborted error by flushing tx and rx fifo and reinitializing the hw.

---
 drivers/i2c/busses/i2c-bcm-iproc.c | 8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)
drivers/i2c/busses/i2c-bcm-iproc.c +5 -3
index d8295b1c379d..834a98caeada 100644
@@ -93,6 +93,7 @@
 #define S_CMD_STATUS_MASK            0x07
 #define S_CMD_STATUS_SUCCESS         0x0
 #define S_CMD_STATUS_TIMEOUT         0x5
+#define S_CMD_STATUS_MASTER_ABORT    0x7
 
 #define IE_OFFSET                    0x38
 #define IE_M_RX_FIFO_FULL_SHIFT      31
@@ -311,9 +312,10 @@ static void bcm_iproc_i2c_check_slave_status(
 		return;
 
 	val = (val >> S_CMD_STATUS_SHIFT) & S_CMD_STATUS_MASK;
-	if (val == S_CMD_STATUS_TIMEOUT) {
-		dev_err(iproc_i2c->device, "slave random stretch time timeout\n");
-
+	if (val == S_CMD_STATUS_TIMEOUT || val == S_CMD_STATUS_MASTER_ABORT) {
+		dev_err(iproc_i2c->device, (val == S_CMD_STATUS_TIMEOUT) ?
+			"slave random stretch time timeout\n" :
+			"Master aborted read transaction\n");
 		/* re-initialize i2c for recovery */
 		bcm_iproc_i2c_enable_disable(iproc_i2c, false);
 		bcm_iproc_i2c_slave_init(iproc_i2c, true);
Signature · 1 line
2.17.1
Acked-by on v1 1/6
#2 lore raw

Re: [PATCH v1 1/6] i2c: iproc: handle Master aborted error

Quoted 25 lines · Rayagonda Kokatanur

Handle Master aborted error by flushing tx and rx fifo and reinitializing the hw.

---
 drivers/i2c/busses/i2c-bcm-iproc.c | 8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)
drivers/i2c/busses/i2c-bcm-iproc.c +1 -0
index d8295b1c379d..834a98caeada 100644
@@ -93,6 +93,7 @@
 #define S_CMD_STATUS_MASK            0x07
 #define S_CMD_STATUS_SUCCESS         0x0
 #define S_CMD_STATUS_TIMEOUT         0x5
+#define S_CMD_STATUS_MASTER_ABORT    0x7
 
 #define IE_OFFSET                    0x38
 #define IE_M_RX_FIFO_FULL_SHIFT      31
@@ -311,9 +312,10 @@ static void bcm_iproc_i2c_check_slave_status(
 		return;
 
 	val = (val >> S_CMD_STATUS_SHIFT) & S_CMD_STATUS_MASK;

drivers/i2c/busses/i2c-bcm-iproc.c static void bcm_iproc_i2c_check_slave_status( · line 321

Quoted 10 lines
drivers/i2c/busses/i2c-bcm-iproc.c +4 -3
-	if (val == S_CMD_STATUS_TIMEOUT) {
-		dev_err(iproc_i2c->device, "slave random stretch time timeout\n");
-
+	if (val == S_CMD_STATUS_TIMEOUT || val == S_CMD_STATUS_MASTER_ABORT) {
+		dev_err(iproc_i2c->device, (val == S_CMD_STATUS_TIMEOUT) ?
+			"slave random stretch time timeout\n" :
+			"Master aborted read transaction\n");
 		/* re-initialize i2c for recovery */
 		bcm_iproc_i2c_enable_disable(iproc_i2c, false);
 		bcm_iproc_i2c_slave_init(iproc_i2c, true);

This looks fine to me. Thanks.

#3 lore raw

PATCH 2/6 i2c: iproc: handle only slave interrupts which are enabled

In reply to Rayagonda Kokatanur

Handle only slave interrupts which are enabled.

The IS_OFFSET register contains the interrupt status bits which will be set regardless of the enabling of the corresponding interrupt condition. One must therefore look at both IS_OFFSET and IE_OFFSET to determine whether an interrupt condition is set and enabled.

Fixes c245d94ed106 ("i2c: iproc: Add multi byte read-write support for slave mode")
---
 drivers/i2c/busses/i2c-bcm-iproc.c | 13 +++++++++----
 1 file changed, 9 insertions(+), 4 deletions(-)
drivers/i2c/busses/i2c-bcm-iproc.c +9 -4
index 834a98caeada..b54f5130d246 100644
@@ -507,12 +507,17 @@ static void bcm_iproc_i2c_process_m_event(struct bcm_iproc_i2c_dev *iproc_i2c,
 static irqreturn_t bcm_iproc_i2c_isr(int irq, void *data)
 {
 	struct bcm_iproc_i2c_dev *iproc_i2c = data;
-	u32 status = iproc_i2c_rd_reg(iproc_i2c, IS_OFFSET);
+	u32 slave_status;
+	u32 status;
 	bool ret;
-	u32 sl_status = status & ISR_MASK_SLAVE;
 
-	if (sl_status) {
-		ret = bcm_iproc_i2c_slave_isr(iproc_i2c, sl_status);
+	status = iproc_i2c_rd_reg(iproc_i2c, IS_OFFSET);
+	/* process only slave interrupt which are enabled */
+	slave_status = status & iproc_i2c_rd_reg(iproc_i2c, IE_OFFSET) &
+		       ISR_MASK_SLAVE;
+
+	if (slave_status) {
+		ret = bcm_iproc_i2c_slave_isr(iproc_i2c, slave_status);
 		if (ret)
 			return IRQ_HANDLED;
 		else
Signature · 1 line
2.17.1
Acked-by on v1 2/6
#4 lore raw

Re: [PATCH v1 2/6] i2c: iproc: handle only slave interrupts which are enabled

Quoted 30 lines · Rayagonda Kokatanur

Handle only slave interrupts which are enabled.

The IS_OFFSET register contains the interrupt status bits which will be set regardless of the enabling of the corresponding interrupt condition. One must therefore look at both IS_OFFSET and IE_OFFSET to determine whether an interrupt condition is set and enabled.

Fixes c245d94ed106 ("i2c: iproc: Add multi byte read-write support for slave mode")
---
 drivers/i2c/busses/i2c-bcm-iproc.c | 13 +++++++++----
 1 file changed, 9 insertions(+), 4 deletions(-)
drivers/i2c/busses/i2c-bcm-iproc.c +2 -4
index 834a98caeada..b54f5130d246 100644
@@ -507,12 +507,17 @@ static void bcm_iproc_i2c_process_m_event(struct bcm_iproc_i2c_dev *iproc_i2c,
 static irqreturn_t bcm_iproc_i2c_isr(int irq, void *data)
 {
 	struct bcm_iproc_i2c_dev *iproc_i2c = data;
-	u32 status = iproc_i2c_rd_reg(iproc_i2c, IS_OFFSET);
+	u32 slave_status;
+	u32 status;
 	bool ret;
-	u32 sl_status = status & ISR_MASK_SLAVE;
 
-	if (sl_status) {
-		ret = bcm_iproc_i2c_slave_isr(iproc_i2c, sl_status);

drivers/i2c/busses/i2c-bcm-iproc.c static void bcm_iproc_i2c_process_m_event(struct bcm_iproc_i2c_dev *iproc_i2c, · line 523

Quoted 10 lines
drivers/i2c/busses/i2c-bcm-iproc.c +7 -0
+	status = iproc_i2c_rd_reg(iproc_i2c, IS_OFFSET);
+	/* process only slave interrupt which are enabled */
+	slave_status = status & iproc_i2c_rd_reg(iproc_i2c, IE_OFFSET) &
+		       ISR_MASK_SLAVE;
+
+	if (slave_status) {
+		ret = bcm_iproc_i2c_slave_isr(iproc_i2c, slave_status);
 		if (ret)
 			return IRQ_HANDLED;
 		else
#5 lore raw

PATCH 3/6 i2c: iproc: update slave isr mask (ISR_MASK_SLAVE)

In reply to Rayagonda Kokatanur

Update slave isr mask (ISR_MASK_SLAVE) to include remaining two slave interrupts.

Fixes c245d94ed106 ("i2c: iproc: Add multi byte read-write support for slave mode")
---
 drivers/i2c/busses/i2c-bcm-iproc.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/i2c/busses/i2c-bcm-iproc.c +2 -1
index b54f5130d246..cd687696bf0b 100644
@@ -216,7 +216,8 @@ struct bcm_iproc_i2c_dev {
 
 #define ISR_MASK_SLAVE (BIT(IS_S_START_BUSY_SHIFT)\
 		| BIT(IS_S_RX_EVENT_SHIFT) | BIT(IS_S_RD_EVENT_SHIFT)\
-		| BIT(IS_S_TX_UNDERRUN_SHIFT))
+		| BIT(IS_S_TX_UNDERRUN_SHIFT) | BIT(IS_S_RX_FIFO_FULL_SHIFT)\
+		| BIT(IS_S_RX_THLD_SHIFT))
 
 static int bcm_iproc_i2c_reg_slave(struct i2c_client *slave);
 static int bcm_iproc_i2c_unreg_slave(struct i2c_client *slave);
Signature · 1 line
2.17.1
Acked-by on v1 3/6
#6 lore raw

Re: [PATCH v1 3/6] i2c: iproc: update slave isr mask (ISR_MASK_SLAVE)

Quoted 15 lines · Rayagonda Kokatanur

Update slave isr mask (ISR_MASK_SLAVE) to include remaining two slave interrupts.

Fixes c245d94ed106 ("i2c: iproc: Add multi byte read-write support for slave mode")
---
 drivers/i2c/busses/i2c-bcm-iproc.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/i2c/busses/i2c-bcm-iproc.c +0 -0
index b54f5130d246..cd687696bf0b 100644

drivers/i2c/busses/i2c-bcm-iproc.c struct bcm_iproc_i2c_dev { · line 223

Quoted 9 lines
drivers/i2c/busses/i2c-bcm-iproc.c +2 -1
@@ -216,7 +216,8 @@ struct bcm_iproc_i2c_dev {
 
 #define ISR_MASK_SLAVE (BIT(IS_S_START_BUSY_SHIFT)\
 		| BIT(IS_S_RX_EVENT_SHIFT) | BIT(IS_S_RD_EVENT_SHIFT)\
-		| BIT(IS_S_TX_UNDERRUN_SHIFT))
+		| BIT(IS_S_TX_UNDERRUN_SHIFT) | BIT(IS_S_RX_FIFO_FULL_SHIFT)\
+		| BIT(IS_S_RX_THLD_SHIFT))
 
 static int bcm_iproc_i2c_reg_slave(struct i2c_client *slave);
 static int bcm_iproc_i2c_unreg_slave(struct i2c_client *slave);
#7 lore raw

PATCH 4/6 i2c: iproc: fix typo in slave_isr function

In reply to Rayagonda Kokatanur

Fix typo in bcm_iproc_i2c_slave_isr().

Fixes c245d94ed106 ("i2c: iproc: Add multi byte read-write support for slave mode")
---
 drivers/i2c/busses/i2c-bcm-iproc.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/i2c/busses/i2c-bcm-iproc.c +1 -1
index cd687696bf0b..7a235f9f5884 100644
@@ -382,7 +382,7 @@ static bool bcm_iproc_i2c_slave_isr(struct bcm_iproc_i2c_dev *iproc_i2c,
 	if (status & BIT(IS_S_START_BUSY_SHIFT)) {
 		i2c_slave_event(iproc_i2c->slave, I2C_SLAVE_STOP, &value);
 		/*
-		 * Enable interrupt for TX FIFO becomes empty and
+		 * Disable interrupt for TX FIFO becomes empty and
 		 * less than PKT_LENGTH bytes were output on the SMBUS
 		 */
 		val = iproc_i2c_rd_reg(iproc_i2c, IE_OFFSET);
Signature · 1 line
2.17.1
#8 lore raw

Re: [PATCH v1 4/6] i2c: iproc: fix typo in slave_isr function

Quoted 3 lines · Rayagonda Kokatanur

Fix typo in bcm_iproc_i2c_slave_isr().

Fixes c245d94ed106 ("i2c: iproc: Add multi byte read-write support for slave mode")

This is merely a fix of typo in code comment and there's no functional impact. Why do we need a Fixes tag on this (which indicates the fix needs to be backported to LTS kernels)?

Quoted 19 lines
---
 drivers/i2c/busses/i2c-bcm-iproc.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/i2c/busses/i2c-bcm-iproc.c +1 -1
index cd687696bf0b..7a235f9f5884 100644
@@ -382,7 +382,7 @@ static bool bcm_iproc_i2c_slave_isr(struct bcm_iproc_i2c_dev *iproc_i2c,
 	if (status & BIT(IS_S_START_BUSY_SHIFT)) {
 		i2c_slave_event(iproc_i2c->slave, I2C_SLAVE_STOP, &value);
 		/*
-		 * Enable interrupt for TX FIFO becomes empty and
+		 * Disable interrupt for TX FIFO becomes empty and
 		 * less than PKT_LENGTH bytes were output on the SMBUS
 		 */
 		val = iproc_i2c_rd_reg(iproc_i2c, IE_OFFSET);
#9 lore raw

Re: [PATCH v1 4/6] i2c: iproc: fix typo in slave_isr function

Quoted 11 lines · Ray Jui
Quoted 3 lines · Rayagonda Kokatanur

Fix typo in bcm_iproc_i2c_slave_isr().

Fixes c245d94ed106 ("i2c: iproc: Add multi byte read-write support for slave mode")

This is merely a fix of typo in code comment and there's no functional impact. Why do we need a Fixes tag on this (which indicates the fix needs to be backported to LTS kernels)?

Thank you Ray.
I will remove Fixes tag and push patch v2.

Best regards,
Rayagonda

Quoted 20 lines
Quoted 20 lines
---
 drivers/i2c/busses/i2c-bcm-iproc.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/i2c/busses/i2c-bcm-iproc.c +1 -1
index cd687696bf0b..7a235f9f5884 100644
@@ -382,7 +382,7 @@ static bool bcm_iproc_i2c_slave_isr(struct bcm_iproc_i2c_dev *iproc_i2c,
      if (status & BIT(IS_S_START_BUSY_SHIFT)) {
              i2c_slave_event(iproc_i2c->slave, I2C_SLAVE_STOP, &value);
              /*
-              * Enable interrupt for TX FIFO becomes empty and
+              * Disable interrupt for TX FIFO becomes empty and
               * less than PKT_LENGTH bytes were output on the SMBUS
               */
              val = iproc_i2c_rd_reg(iproc_i2c, IE_OFFSET);
#10 lore raw

PATCH 5/6 i2c: iproc: handle master read request

In reply to Rayagonda Kokatanur

Handle single or multi byte master read request with or without repeated start.

Fixes c245d94ed106 ("i2c: iproc: Add multi byte read-write support for slave mode")
---
 drivers/i2c/busses/i2c-bcm-iproc.c | 215 +++++++++++++++++++++++------
 1 file changed, 170 insertions(+), 45 deletions(-)
drivers/i2c/busses/i2c-bcm-iproc.c +170 -45
index 7a235f9f5884..22e04055b447 100644
@@ -160,6 +160,11 @@
 
 #define IE_S_ALL_INTERRUPT_SHIFT     21
 #define IE_S_ALL_INTERRUPT_MASK      0x3f
+/*
+ * It takes ~18us to reading 10bytes of data, hence to keep tasklet
+ * running for less time, max slave read per tasklet is set to 10 bytes.
+ */
+#define MAX_SLAVE_RX_PER_INT         10
 
 enum i2c_slave_read_status {
 	I2C_SLAVE_RX_FIFO_EMPTY = 0,
@@ -206,8 +211,18 @@ struct bcm_iproc_i2c_dev {
 	/* bytes that have been read */
 	unsigned int rx_bytes;
 	unsigned int thld_bytes;
+
+	bool slave_rx_only;
+	bool rx_start_rcvd;
+	bool slave_read_complete;
+	u32 tx_underrun;
+	u32 slave_int_mask;
+	struct tasklet_struct slave_rx_tasklet;
 };
 
+/* tasklet to process slave rx data */
+static void slave_rx_tasklet_fn(unsigned long);
+
 /*
  * Can be expanded in the future if more interrupt status bits are utilized
  */
@@ -261,6 +276,7 @@ static void bcm_iproc_i2c_slave_init(
 {
 	u32 val;
 
+	iproc_i2c->tx_underrun = 0;
 	if (need_reset) {
 		/* put controller in reset */
 		val = iproc_i2c_rd_reg(iproc_i2c, CFG_OFFSET);
@@ -297,8 +313,11 @@ static void bcm_iproc_i2c_slave_init(
 
 	/* Enable interrupt register to indicate a valid byte in receive fifo */
 	val = BIT(IE_S_RX_EVENT_SHIFT);
+	/* Enable interrupt register to indicate a Master read transaction */
+	val |= BIT(IE_S_RD_EVENT_SHIFT);
 	/* Enable interrupt register for the Slave BUSY command */
 	val |= BIT(IE_S_START_BUSY_SHIFT);
+	iproc_i2c->slave_int_mask = val;
 	iproc_i2c_wr_reg(iproc_i2c, IE_OFFSET, val);
 }
 
@@ -324,76 +343,176 @@ static void bcm_iproc_i2c_check_slave_status(
 	}
 }
 
-static bool bcm_iproc_i2c_slave_isr(struct bcm_iproc_i2c_dev *iproc_i2c,
-				    u32 status)
+static void bcm_iproc_i2c_slave_read(struct bcm_iproc_i2c_dev *iproc_i2c)
 {
+	u8 rx_data, rx_status;
+	u32 rx_bytes = 0;
 	u32 val;
-	u8 value, rx_status;
 
-	/* Slave RX byte receive */
-	if (status & BIT(IS_S_RX_EVENT_SHIFT)) {
+	while (rx_bytes < MAX_SLAVE_RX_PER_INT) {
 		val = iproc_i2c_rd_reg(iproc_i2c, S_RX_OFFSET);
 		rx_status = (val >> S_RX_STATUS_SHIFT) & S_RX_STATUS_MASK;
-		if (rx_status == I2C_SLAVE_RX_START) {
-			/* Start of SMBUS for Master write */
-			i2c_slave_event(iproc_i2c->slave,
-					I2C_SLAVE_WRITE_REQUESTED, &value);
+		rx_data = ((val >> S_RX_DATA_SHIFT) & S_RX_DATA_MASK);
 
-			val = iproc_i2c_rd_reg(iproc_i2c, S_RX_OFFSET);
-			value = (u8)((val >> S_RX_DATA_SHIFT) & S_RX_DATA_MASK);
+		if (rx_status == I2C_SLAVE_RX_START) {
+			/* Start of SMBUS Master write */
 			i2c_slave_event(iproc_i2c->slave,
-					I2C_SLAVE_WRITE_RECEIVED, &value);
-		} else if (status & BIT(IS_S_RD_EVENT_SHIFT)) {
-			/* Start of SMBUS for Master Read */
+					I2C_SLAVE_WRITE_REQUESTED, &rx_data);
+			iproc_i2c->rx_start_rcvd = true;
+			iproc_i2c->slave_read_complete = false;
+		} else if (rx_status == I2C_SLAVE_RX_DATA &&
+			   iproc_i2c->rx_start_rcvd) {
+			/* Middle of SMBUS Master write */
 			i2c_slave_event(iproc_i2c->slave,
-					I2C_SLAVE_READ_REQUESTED, &value);
-			iproc_i2c_wr_reg(iproc_i2c, S_TX_OFFSET, value);
+					I2C_SLAVE_WRITE_RECEIVED, &rx_data);
+		} else if (rx_status == I2C_SLAVE_RX_END &&
+			   iproc_i2c->rx_start_rcvd) {
+			/* End of SMBUS Master write */
+			if (iproc_i2c->slave_rx_only)
+				i2c_slave_event(iproc_i2c->slave,
+						I2C_SLAVE_WRITE_RECEIVED,
+						&rx_data);
+
+			i2c_slave_event(iproc_i2c->slave, I2C_SLAVE_STOP,
+					&rx_data);
+		} else if (rx_status == I2C_SLAVE_RX_FIFO_EMPTY) {
+			iproc_i2c->rx_start_rcvd = false;
+			iproc_i2c->slave_read_complete = true;
+			break;
+		}
 
-			val = BIT(S_CMD_START_BUSY_SHIFT);
-			iproc_i2c_wr_reg(iproc_i2c, S_CMD_OFFSET, val);
+		rx_bytes++;
+	}
+}
 
-			/*
-			 * Enable interrupt for TX FIFO becomes empty and
-			 * less than PKT_LENGTH bytes were output on the SMBUS
-			 */
-			val = iproc_i2c_rd_reg(iproc_i2c, IE_OFFSET);
-			val |= BIT(IE_S_TX_UNDERRUN_SHIFT);
-			iproc_i2c_wr_reg(iproc_i2c, IE_OFFSET, val);
-		} else {
-			/* Master write other than start */
-			value = (u8)((val >> S_RX_DATA_SHIFT) & S_RX_DATA_MASK);
+static void slave_rx_tasklet_fn(unsigned long data)
+{
+	struct bcm_iproc_i2c_dev *iproc_i2c = (struct bcm_iproc_i2c_dev *)data;
+	u32 int_clr;
+
+	bcm_iproc_i2c_slave_read(iproc_i2c);
+
+	/* clear pending IS_S_RX_EVENT_SHIFT interrupt */
+	int_clr = BIT(IS_S_RX_EVENT_SHIFT);
+
+	if (!iproc_i2c->slave_rx_only && iproc_i2c->slave_read_complete) {
+		/*
+		 * In case of single byte master-read request,
+		 * IS_S_TX_UNDERRUN_SHIFT event is generated before
+		 * IS_S_START_BUSY_SHIFT event. Hence start slave data send
+		 * from first IS_S_TX_UNDERRUN_SHIFT event.
+		 *
+		 * This means don't send any data from slave when
+		 * IS_S_RD_EVENT_SHIFT event is generated else it will increment
+		 * eeprom or other backend slave driver read pointer twice.
+		 */
+		iproc_i2c->tx_underrun = 0;
+		iproc_i2c->slave_int_mask |= BIT(IE_S_TX_UNDERRUN_SHIFT);
+
+		/* clear IS_S_RD_EVENT_SHIFT interrupt */
+		int_clr |= BIT(IS_S_RD_EVENT_SHIFT);
+	}
+
+	/* clear slave interrupt */
+	iproc_i2c_wr_reg(iproc_i2c, IS_OFFSET, int_clr);
+	/* enable slave interrupts */
+	iproc_i2c_wr_reg(iproc_i2c, IE_OFFSET, iproc_i2c->slave_int_mask);
+}
+
+static bool bcm_iproc_i2c_slave_isr(struct bcm_iproc_i2c_dev *iproc_i2c,
+				    u32 status)
+{
+	u32 val;
+	u8 value;
+
+	/*
+	 * Slave events in case of master-write, master-write-read and,
+	 * master-read
+	 *
+	 * Master-write     : only IS_S_RX_EVENT_SHIFT event
+	 * Master-write-read: both IS_S_RX_EVENT_SHIFT and IS_S_RD_EVENT_SHIFT
+	 *                    events
+	 * Master-read      : both IS_S_RX_EVENT_SHIFT and IS_S_RD_EVENT_SHIFT
+	 *                    events or only IS_S_RD_EVENT_SHIFT
+	 */
+	if (status & BIT(IS_S_RX_EVENT_SHIFT) ||
+	    status & BIT(IS_S_RD_EVENT_SHIFT)) {
+		/* disable slave interrupts */
+		val = iproc_i2c_rd_reg(iproc_i2c, IE_OFFSET);
+		val &= ~iproc_i2c->slave_int_mask;
+		iproc_i2c_wr_reg(iproc_i2c, IE_OFFSET, val);
+
+		if (status & BIT(IS_S_RD_EVENT_SHIFT))
+			/* Master-write-read request */
+			iproc_i2c->slave_rx_only = false;
+		else
+			/* Master-write request only */
+			iproc_i2c->slave_rx_only = true;
+
+		/* schedule tasklet to read data later */
+		tasklet_schedule(&iproc_i2c->slave_rx_tasklet);
+
+		/* clear only IS_S_RX_EVENT_SHIFT interrupt */
+		iproc_i2c_wr_reg(iproc_i2c, IS_OFFSET,
+				 BIT(IS_S_RX_EVENT_SHIFT));
+	}
+
+	if (status & BIT(IS_S_TX_UNDERRUN_SHIFT)) {
+		iproc_i2c->tx_underrun++;
+		if (iproc_i2c->tx_underrun == 1)
+			/* Start of SMBUS for Master Read */
 			i2c_slave_event(iproc_i2c->slave,
-					I2C_SLAVE_WRITE_RECEIVED, &value);
-			if (rx_status == I2C_SLAVE_RX_END)
-				i2c_slave_event(iproc_i2c->slave,
-						I2C_SLAVE_STOP, &value);
-		}
-	} else if (status & BIT(IS_S_TX_UNDERRUN_SHIFT)) {
-		/* Master read other than start */
-		i2c_slave_event(iproc_i2c->slave,
-				I2C_SLAVE_READ_PROCESSED, &value);
+					I2C_SLAVE_READ_REQUESTED,
+					&value);
+		else
+			/* Master read other than start */
+			i2c_slave_event(iproc_i2c->slave,
+					I2C_SLAVE_READ_PROCESSED,
+					&value);
 
 		iproc_i2c_wr_reg(iproc_i2c, S_TX_OFFSET, value);
+		/* start transfer */
 		val = BIT(S_CMD_START_BUSY_SHIFT);
 		iproc_i2c_wr_reg(iproc_i2c, S_CMD_OFFSET, val);
+
+		/* clear interrupt */
+		iproc_i2c_wr_reg(iproc_i2c, IS_OFFSET,
+				 BIT(IS_S_TX_UNDERRUN_SHIFT));
 	}
 
-	/* Stop */
+	/* Stop received from master in case of master read transaction */
 	if (status & BIT(IS_S_START_BUSY_SHIFT)) {
-		i2c_slave_event(iproc_i2c->slave, I2C_SLAVE_STOP, &value);
 		/*
 		 * Disable interrupt for TX FIFO becomes empty and
 		 * less than PKT_LENGTH bytes were output on the SMBUS
 		 */
-		val = iproc_i2c_rd_reg(iproc_i2c, IE_OFFSET);
-		val &= ~BIT(IE_S_TX_UNDERRUN_SHIFT);
-		iproc_i2c_wr_reg(iproc_i2c, IE_OFFSET, val);
+		iproc_i2c->slave_int_mask &= ~BIT(IE_S_TX_UNDERRUN_SHIFT);
+		iproc_i2c_wr_reg(iproc_i2c, IE_OFFSET,
+				 iproc_i2c->slave_int_mask);
+
+		/* End of SMBUS for Master Read */
+		val = BIT(S_TX_WR_STATUS_SHIFT);
+		iproc_i2c_wr_reg(iproc_i2c, S_TX_OFFSET, val);
+
+		val = BIT(S_CMD_START_BUSY_SHIFT);
+		iproc_i2c_wr_reg(iproc_i2c, S_CMD_OFFSET, val);
+
+		/* flush TX FIFOs */
+		val = iproc_i2c_rd_reg(iproc_i2c, S_FIFO_CTRL_OFFSET);
+		val |= (BIT(S_FIFO_TX_FLUSH_SHIFT));
+		iproc_i2c_wr_reg(iproc_i2c, S_FIFO_CTRL_OFFSET, val);
+
+		i2c_slave_event(iproc_i2c->slave, I2C_SLAVE_STOP, &value);
+
+		/* clear interrupt */
+		iproc_i2c_wr_reg(iproc_i2c, IS_OFFSET,
+				 BIT(IS_S_START_BUSY_SHIFT));
 	}
 
-	/* clear interrupt status */
-	iproc_i2c_wr_reg(iproc_i2c, IS_OFFSET, status);
+	/* check slave transmit status only if slave is transmitting */
+	if (!iproc_i2c->slave_rx_only)
+		bcm_iproc_i2c_check_slave_status(iproc_i2c);
 
-	bcm_iproc_i2c_check_slave_status(iproc_i2c);
 	return true;
 }
 
@@ -1074,6 +1193,10 @@ static int bcm_iproc_i2c_reg_slave(struct i2c_client *slave)
 		return -EAFNOSUPPORT;
 
 	iproc_i2c->slave = slave;
+
+	tasklet_init(&iproc_i2c->slave_rx_tasklet, slave_rx_tasklet_fn,
+		     (unsigned long)iproc_i2c);
+
 	bcm_iproc_i2c_slave_init(iproc_i2c, false);
 	return 0;
 }
@@ -1094,6 +1217,8 @@ static int bcm_iproc_i2c_unreg_slave(struct i2c_client *slave)
 			IE_S_ALL_INTERRUPT_SHIFT);
 	iproc_i2c_wr_reg(iproc_i2c, IE_OFFSET, tmp);
 
+	tasklet_kill(&iproc_i2c->slave_rx_tasklet);
+
 	/* Erase the slave address programmed */
 	tmp = iproc_i2c_rd_reg(iproc_i2c, S_CFG_SMBUS_ADDR_OFFSET);
 	tmp &= ~BIT(S_CFG_EN_NIC_SMB_ADDR3_SHIFT);
Signature · 1 line
2.17.1
#11 lore raw

PATCH 5/6 i2c: iproc: handle master read request

Quoted 24 lines · Rayagonda Kokatanur
drivers/i2c/busses/i2c-bcm-iproc.c +16 -4
 
-		} else if (status & BIT(IS_S_RD_EVENT_SHIFT)) {
-			/* Start of SMBUS for Master Read */
+					I2C_SLAVE_WRITE_REQUESTED, &rx_data);
+			iproc_i2c->rx_start_rcvd = true;
+			iproc_i2c->slave_read_complete = false;
+		} else if (rx_status == I2C_SLAVE_RX_DATA &&
+			   iproc_i2c->rx_start_rcvd) {
+			/* Middle of SMBUS Master write */
 			i2c_slave_event(iproc_i2c->slave,
-					I2C_SLAVE_READ_REQUESTED, &value);
-			iproc_i2c_wr_reg(iproc_i2c, S_TX_OFFSET, value);
+					I2C_SLAVE_WRITE_RECEIVED, &rx_data);
+		} else if (rx_status == I2C_SLAVE_RX_END &&
+			   iproc_i2c->rx_start_rcvd) {
+			/* End of SMBUS Master write */
+			if (iproc_i2c->slave_rx_only)
+				i2c_slave_event(iproc_i2c->slave,
+						I2C_SLAVE_WRITE_RECEIVED,
+						&rx_data);
+
+			i2c_slave_event(iproc_i2c->slave, I2C_SLAVE_STOP,

drivers/i2c/busses/i2c-bcm-iproc.c static void bcm_iproc_i2c_check_slave_status( · line 384

Quoted 10 lines
drivers/i2c/busses/i2c-bcm-iproc.c +7 -2
+					&rx_data);
+		} else if (rx_status == I2C_SLAVE_RX_FIFO_EMPTY) {
+			iproc_i2c->rx_start_rcvd = false;
+			iproc_i2c->slave_read_complete = true;
+			break;
+		}
 
-			val = BIT(S_CMD_START_BUSY_SHIFT);
-			iproc_i2c_wr_reg(iproc_i2c, S_CMD_OFFSET, val);
+		rx_bytes++;

rx_bytes should be incremented only along with I2C_SLAVE_WRITE_RECEIVED event?

Quoted 28 lines
drivers/i2c/busses/i2c-bcm-iproc.c +26 -0
+static bool bcm_iproc_i2c_slave_isr(struct bcm_iproc_i2c_dev *iproc_i2c,
+				    u32 status)
+{
+	u32 val;
+	u8 value;
+
+	/*
+	 * Slave events in case of master-write, master-write-read and,
+	 * master-read
+	 *
+	 * Master-write     : only IS_S_RX_EVENT_SHIFT event
+	 * Master-write-read: both IS_S_RX_EVENT_SHIFT and IS_S_RD_EVENT_SHIFT
+	 *                    events
+	 * Master-read      : both IS_S_RX_EVENT_SHIFT and IS_S_RD_EVENT_SHIFT
+	 *                    events or only IS_S_RD_EVENT_SHIFT
+	 */
+	if (status & BIT(IS_S_RX_EVENT_SHIFT) ||
+	    status & BIT(IS_S_RD_EVENT_SHIFT)) {
+		/* disable slave interrupts */
+		val = iproc_i2c_rd_reg(iproc_i2c, IE_OFFSET);
+		val &= ~iproc_i2c->slave_int_mask;
+		iproc_i2c_wr_reg(iproc_i2c, IE_OFFSET, val);
+
+		if (status & BIT(IS_S_RD_EVENT_SHIFT))
+			/* Master-write-read request */
+			iproc_i2c->slave_rx_only = false;

drivers/i2c/busses/i2c-bcm-iproc.c static void bcm_iproc_i2c_check_slave_status( · line 457

Quoted 10 lines
drivers/i2c/busses/i2c-bcm-iproc.c +10 -0
+		else
+			/* Master-write request only */
+			iproc_i2c->slave_rx_only = true;
+
+		/* schedule tasklet to read data later */
+		tasklet_schedule(&iproc_i2c->slave_rx_tasklet);
+
+		/* clear only IS_S_RX_EVENT_SHIFT interrupt */
+		iproc_i2c_wr_reg(iproc_i2c, IS_OFFSET,
+				 BIT(IS_S_RX_EVENT_SHIFT));

Both tasklet and isr are writing to status (IS_OFFSET) reg.

The tasklet seems to be batching up rx fifo reads because of time-sensitive Master-write-read transaction? Linux I2C framework is byte interface anyway. Can the need to batch reads be avoided by setting slave rx threshold for interrupt (S_FIFO_RX_THLD) to 1-byte?

Also, wouldn't tasklets be susceptible to other interrupts? If fifo reads have to be batched up, can it be changed to threaded irq?

#12 lore raw

Re: [PATCH v1 5/6] i2c: iproc: handle master read request

Quoted 33 lines · Dhananjay Phadke
Quoted 31 lines · Rayagonda Kokatanur
drivers/i2c/busses/i2c-bcm-iproc.c +22 -4
 
-             } else if (status & BIT(IS_S_RD_EVENT_SHIFT)) {
-                     /* Start of SMBUS for Master Read */
+                                     I2C_SLAVE_WRITE_REQUESTED, &rx_data);
+                     iproc_i2c->rx_start_rcvd = true;
+                     iproc_i2c->slave_read_complete = false;
+             } else if (rx_status == I2C_SLAVE_RX_DATA &&
+                        iproc_i2c->rx_start_rcvd) {
+                     /* Middle of SMBUS Master write */
                      i2c_slave_event(iproc_i2c->slave,
-                                     I2C_SLAVE_READ_REQUESTED, &value);
-                     iproc_i2c_wr_reg(iproc_i2c, S_TX_OFFSET, value);
+                                     I2C_SLAVE_WRITE_RECEIVED, &rx_data);
+             } else if (rx_status == I2C_SLAVE_RX_END &&
+                        iproc_i2c->rx_start_rcvd) {
+                     /* End of SMBUS Master write */
+                     if (iproc_i2c->slave_rx_only)
+                             i2c_slave_event(iproc_i2c->slave,
+                                             I2C_SLAVE_WRITE_RECEIVED,
+                                             &rx_data);
+
+                     i2c_slave_event(iproc_i2c->slave, I2C_SLAVE_STOP,
+                                     &rx_data);
+             } else if (rx_status == I2C_SLAVE_RX_FIFO_EMPTY) {
+                     iproc_i2c->rx_start_rcvd = false;
+                     iproc_i2c->slave_read_complete = true;
+                     break;
+             }
Quoted 5 lines

drivers/i2c/busses/i2c-bcm-iproc.c static void bcm_iproc_i2c_check_slave_status( · line 384

Quoted 3 lines
drivers/i2c/busses/i2c-bcm-iproc.c +1 -2
-                     val = BIT(S_CMD_START_BUSY_SHIFT);
-                     iproc_i2c_wr_reg(iproc_i2c, S_CMD_OFFSET, val);
+             rx_bytes++;

rx_bytes should be incremented only along with I2C_SLAVE_WRITE_RECEIVED event?

It should be incremented in both I2C_SLAVE_WRITE_REQUESTED and I2C_SLAVE_WRITE_RECEIVED cases because in both cases it is reading valid bytes from rx fifo.

Quoted 42 lines
Quoted 29 lines
drivers/i2c/busses/i2c-bcm-iproc.c +26 -0
+static bool bcm_iproc_i2c_slave_isr(struct bcm_iproc_i2c_dev *iproc_i2c,
+                                 u32 status)
+{
+     u32 val;
+     u8 value;
+
+     /*
+      * Slave events in case of master-write, master-write-read and,
+      * master-read
+      *
+      * Master-write     : only IS_S_RX_EVENT_SHIFT event
+      * Master-write-read: both IS_S_RX_EVENT_SHIFT and IS_S_RD_EVENT_SHIFT
+      *                    events
+      * Master-read      : both IS_S_RX_EVENT_SHIFT and IS_S_RD_EVENT_SHIFT
+      *                    events or only IS_S_RD_EVENT_SHIFT
+      */
+     if (status & BIT(IS_S_RX_EVENT_SHIFT) ||
+         status & BIT(IS_S_RD_EVENT_SHIFT)) {
+             /* disable slave interrupts */
+             val = iproc_i2c_rd_reg(iproc_i2c, IE_OFFSET);
+             val &= ~iproc_i2c->slave_int_mask;
+             iproc_i2c_wr_reg(iproc_i2c, IE_OFFSET, val);
+
+             if (status & BIT(IS_S_RD_EVENT_SHIFT))
+                     /* Master-write-read request */
+                     iproc_i2c->slave_rx_only = false;

drivers/i2c/busses/i2c-bcm-iproc.c static void bcm_iproc_i2c_check_slave_status( · line 457

Quoted 10 lines
drivers/i2c/busses/i2c-bcm-iproc.c +10 -0
+             else
+                     /* Master-write request only */
+                     iproc_i2c->slave_rx_only = true;
+
+             /* schedule tasklet to read data later */
+             tasklet_schedule(&iproc_i2c->slave_rx_tasklet);
+
+             /* clear only IS_S_RX_EVENT_SHIFT interrupt */
+             iproc_i2c_wr_reg(iproc_i2c, IS_OFFSET,
+                              BIT(IS_S_RX_EVENT_SHIFT));

Both tasklet and isr are writing to status (IS_OFFSET) reg.

Quoted 4 lines

The tasklet seems to be batching up rx fifo reads because of time-sensitive Master-write-read transaction? Linux I2C framework is byte interface anyway. Can the need to batch reads be avoided by setting slave rx threshold for interrupt (S_FIFO_RX_THLD) to 1-byte?

To process more data with a single interrupt we are batching up rx fifo reads. This will reduce the number of interrupts.

Also to avoid tasklet running more time (20us) we have a threshold of 10 bytes for batching read.
This is a better/optimised approach than reading single byte data per interrupt.

Quoted 3 lines

Also, wouldn't tasklets be susceptible to other interrupts? If fifo reads have to be batched up, can it be changed to threaded irq?

tasklets have higher priority than threaded irq, since i2c is time sensitive so using a tasklet is preferred over threaded irq.

Best regards,
Rayagonda

Quoted 2 lines

A message that is not in the archive (sent off-list or not yet received): CAHO=5PEtoJrFEPin0hH19Ubs9Zmhxiay4jSGAhXBFE=ft=+CYg@mail.gmail.com

#14 lore raw

Re: [PATCH v1 5/6] i2c: iproc: handle master read request

Quoted 100 lines · Rayagonda Kokatanur

On Wed, Oct 14, 2020 at 8:50 AM Dhananjay Phadke <[email protected] <mailto:[email protected]>> wrote:

    On Sun, 11 Oct 2020 23:52:53 +0530, Rayagonda Kokatanur wrote:
    > --- a/drivers/i2c/busses/i2c-bcm-iproc.c
    > +++ b/drivers/i2c/busses/i2c-bcm-iproc.c
    >
    > -             } else if (status & BIT(IS_S_RD_EVENT_SHIFT)) {
    > -                     /* Start of SMBUS for Master Read */
    > +                                     I2C_SLAVE_WRITE_REQUESTED,
    &rx_data);
    > +                     iproc_i2c->rx_start_rcvd = true;
    > +                     iproc_i2c->slave_read_complete = false;
    > +             } else if (rx_status == I2C_SLAVE_RX_DATA &&
    > +                        iproc_i2c->rx_start_rcvd) {
    > +                     /* Middle of SMBUS Master write */
    >                       i2c_slave_event(iproc_i2c->slave,
    > -                                     I2C_SLAVE_READ_REQUESTED,
    &value);
    > -                     iproc_i2c_wr_reg(iproc_i2c, S_TX_OFFSET, value);
    > +                                     I2C_SLAVE_WRITE_RECEIVED,
    &rx_data);
    > +             } else if (rx_status == I2C_SLAVE_RX_END &&
    > +                        iproc_i2c->rx_start_rcvd) {
    > +                     /* End of SMBUS Master write */
    > +                     if (iproc_i2c->slave_rx_only)
    > +                             i2c_slave_event(iproc_i2c->slave,
    > +                                           
     I2C_SLAVE_WRITE_RECEIVED,
    > +                                             &rx_data);
    > +
    > +                     i2c_slave_event(iproc_i2c->slave,
    I2C_SLAVE_STOP,
    > +                                     &rx_data);
    > +             } else if (rx_status == I2C_SLAVE_RX_FIFO_EMPTY) {
    > +                     iproc_i2c->rx_start_rcvd = false;
    > +                     iproc_i2c->slave_read_complete = true;
    > +                     break;
    > +             }
    > 
    > -                     val = BIT(S_CMD_START_BUSY_SHIFT);
    > -                     iproc_i2c_wr_reg(iproc_i2c, S_CMD_OFFSET, val);
    > +             rx_bytes++;

    rx_bytes should be incremented only along with
    I2C_SLAVE_WRITE_RECEIVED event?

It should be incremented in both I2C_SLAVE_WRITE_REQUESTED and I2C_SLAVE_WRITE_RECEIVED cases because in both case it is reading valid bytes from rx fifo.

Quoted 2 lines

+static bool bcm_iproc_i2c_slave_isr(struct bcm_iproc_i2c_dev

    *iproc_i2c,
    > +                                 u32 status)
    > +{
    > +     u32 val;
    > +     u8 value;
    > +
    > +     /*
    > +      * Slave events in case of master-write, master-write-read and,
    > +      * master-read
    > +      *
    > +      * Master-write     : only IS_S_RX_EVENT_SHIFT event
    > +      * Master-write-read: both IS_S_RX_EVENT_SHIFT and
    IS_S_RD_EVENT_SHIFT
    > +      *                    events
    > +      * Master-read      : both IS_S_RX_EVENT_SHIFT and
    IS_S_RD_EVENT_SHIFT
    > +      *                    events or only IS_S_RD_EVENT_SHIFT
    > +      */
    > +     if (status & BIT(IS_S_RX_EVENT_SHIFT) ||
    > +         status & BIT(IS_S_RD_EVENT_SHIFT)) {
    > +             /* disable slave interrupts */
    > +             val = iproc_i2c_rd_reg(iproc_i2c, IE_OFFSET);
    > +             val &= ~iproc_i2c->slave_int_mask;
    > +             iproc_i2c_wr_reg(iproc_i2c, IE_OFFSET, val);
    > +
    > +             if (status & BIT(IS_S_RD_EVENT_SHIFT))
    > +                     /* Master-write-read request */
    > +                     iproc_i2c->slave_rx_only = false;
    > +             else
    > +                     /* Master-write request only */
    > +                     iproc_i2c->slave_rx_only = true;
    > +
    > +             /* schedule tasklet to read data later */
    > +             tasklet_schedule(&iproc_i2c->slave_rx_tasklet);
    > +
    > +             /* clear only IS_S_RX_EVENT_SHIFT interrupt */
    > +             iproc_i2c_wr_reg(iproc_i2c, IS_OFFSET,
    > +                              BIT(IS_S_RX_EVENT_SHIFT));
    >

    Both tasklet and isr are writing to status (IS_OFFSET) reg.
Quoted 6 lines

Yes this is required.

For ex, If IS_S_RD_EVENT_SHIFT interrupt, this should be cleared once the driver completes reading all data from rx fifo.
After this the driver can start sending data to master.

If both tasklet and isr are accessing the IS_OFFSET register, don't you need lock protection against race condition? That is, ISR can interrupt tasklet.

Quoted 28 lines

The tasklet seems to be batching up rx fifo reads because of time-sensitive
Master-write-read transaction? Linux I2C framework is byte interface anyway.
Can the need to batch reads be avoided by setting slave rx threshold for interrupt (S_FIFO_RX_THLD) to 1-byte?

To process more data with a single interrupt we are batching up rx fifo reads.
This will reduce the number of interrupts.

Also to avoid tasklet running more time (20us) we have a threshold of 10 bytes for batching read.
This is a better/optimised approach than reading single byte data per interrupt.

Also, wouldn't tasklets be susceptible to other interrupts? If fifo reads
have to be batched up, can it be changed to threaded irq?

tasklets have higher priority than threaded irq, since i2c is time sensitive so using a tasklet is preferred over threaded irq.

#15 lore raw

Re: [PATCH v1 5/6] i2c: iproc: handle master read request

Quoted 106 lines · Ray Jui
Quoted 102 lines · Rayagonda Kokatanur

On Wed, Oct 14, 2020 at 8:50 AM Dhananjay Phadke <[email protected] <mailto:[email protected]>> wrote:

    On Sun, 11 Oct 2020 23:52:53 +0530, Rayagonda Kokatanur wrote:
    > --- a/drivers/i2c/busses/i2c-bcm-iproc.c
    > +++ b/drivers/i2c/busses/i2c-bcm-iproc.c
    >
    > -             } else if (status & BIT(IS_S_RD_EVENT_SHIFT)) {
    > -                     /* Start of SMBUS for Master Read */
    > +                                     I2C_SLAVE_WRITE_REQUESTED,
    &rx_data);
    > +                     iproc_i2c->rx_start_rcvd = true;
    > +                     iproc_i2c->slave_read_complete = false;
    > +             } else if (rx_status == I2C_SLAVE_RX_DATA &&
    > +                        iproc_i2c->rx_start_rcvd) {
    > +                     /* Middle of SMBUS Master write */
    >                       i2c_slave_event(iproc_i2c->slave,
    > -                                     I2C_SLAVE_READ_REQUESTED,
    &value);
    > -                     iproc_i2c_wr_reg(iproc_i2c, S_TX_OFFSET, value);
    > +                                     I2C_SLAVE_WRITE_RECEIVED,
    &rx_data);
    > +             } else if (rx_status == I2C_SLAVE_RX_END &&
    > +                        iproc_i2c->rx_start_rcvd) {
    > +                     /* End of SMBUS Master write */
    > +                     if (iproc_i2c->slave_rx_only)
    > +                             i2c_slave_event(iproc_i2c->slave,
    > +
     I2C_SLAVE_WRITE_RECEIVED,
    > +                                             &rx_data);
    > +
    > +                     i2c_slave_event(iproc_i2c->slave,
    I2C_SLAVE_STOP,
    > +                                     &rx_data);
    > +             } else if (rx_status == I2C_SLAVE_RX_FIFO_EMPTY) {
    > +                     iproc_i2c->rx_start_rcvd = false;
    > +                     iproc_i2c->slave_read_complete = true;
    > +                     break;
    > +             }
    >
    > -                     val = BIT(S_CMD_START_BUSY_SHIFT);
    > -                     iproc_i2c_wr_reg(iproc_i2c, S_CMD_OFFSET, val);
    > +             rx_bytes++;

    rx_bytes should be incremented only along with
    I2C_SLAVE_WRITE_RECEIVED event?

It should be incremented in both I2C_SLAVE_WRITE_REQUESTED and I2C_SLAVE_WRITE_RECEIVED cases because in both case it is reading valid bytes from rx fifo.

Quoted 2 lines

+static bool bcm_iproc_i2c_slave_isr(struct bcm_iproc_i2c_dev

    *iproc_i2c,
    > +                                 u32 status)
    > +{
    > +     u32 val;
    > +     u8 value;
    > +
    > +     /*
    > +      * Slave events in case of master-write, master-write-read and,
    > +      * master-read
    > +      *
    > +      * Master-write     : only IS_S_RX_EVENT_SHIFT event
    > +      * Master-write-read: both IS_S_RX_EVENT_SHIFT and
    IS_S_RD_EVENT_SHIFT
    > +      *                    events
    > +      * Master-read      : both IS_S_RX_EVENT_SHIFT and
    IS_S_RD_EVENT_SHIFT
    > +      *                    events or only IS_S_RD_EVENT_SHIFT
    > +      */
    > +     if (status & BIT(IS_S_RX_EVENT_SHIFT) ||
    > +         status & BIT(IS_S_RD_EVENT_SHIFT)) {
    > +             /* disable slave interrupts */
    > +             val = iproc_i2c_rd_reg(iproc_i2c, IE_OFFSET);
    > +             val &= ~iproc_i2c->slave_int_mask;
    > +             iproc_i2c_wr_reg(iproc_i2c, IE_OFFSET, val);
    > +
    > +             if (status & BIT(IS_S_RD_EVENT_SHIFT))
    > +                     /* Master-write-read request */
    > +                     iproc_i2c->slave_rx_only = false;
    > +             else
    > +                     /* Master-write request only */
    > +                     iproc_i2c->slave_rx_only = true;
    > +
    > +             /* schedule tasklet to read data later */
    > +             tasklet_schedule(&iproc_i2c->slave_rx_tasklet);
    > +
    > +             /* clear only IS_S_RX_EVENT_SHIFT interrupt */
    > +             iproc_i2c_wr_reg(iproc_i2c, IS_OFFSET,
    > +                              BIT(IS_S_RX_EVENT_SHIFT));
    >

    Both tasklet and isr are writing to status (IS_OFFSET) reg.

Yes this is required.

Quoted 8 lines
Quoted 4 lines

For ex, If IS_S_RD_EVENT_SHIFT interrupt, this should be cleared once the driver completes reading all data from rx fifo.
After this the driver can start sending data to master.

If both tasklet and isr are accessing the IS_OFFSET register, don't you need lock protection against race condition? That is, ISR can interrupt tasklet.

All interrupts are disbaled when the tasklet is running. Interrupts are re-enabled at the end of the tasklet. So no race condition between tasklet and isr.

Best regards,
Rayagonda

Quoted 29 lines
Quoted 29 lines

The tasklet seems to be batching up rx fifo reads because of time-sensitive
Master-write-read transaction? Linux I2C framework is byte interface anyway.
Can the need to batch reads be avoided by setting slave rx threshold for interrupt (S_FIFO_RX_THLD) to 1-byte?

To process more data with a single interrupt we are batching up rx fifo reads.
This will reduce the number of interrupts.

Also to avoid tasklet running more time (20us) we have a threshold of 10 bytes for batching read.
This is a better/optimised approach than reading single byte data per interrupt.

Also, wouldn't tasklets be susceptible to other interrupts? If fifo reads
have to be batched up, can it be changed to threaded irq?

tasklets have higher priority than threaded irq, since i2c is time sensitive so using a tasklet is preferred over threaded irq.

#16 lore raw

PATCH 6/6 i2c: iproc: handle rx fifo full interrupt

In reply to Rayagonda Kokatanur

Add code to handle IS_S_RX_FIFO_FULL_SHIFT interrupt to support master write request with >= 64 bytes.

Iproc has a slave rx fifo size of 64 bytes.
Rx fifo full interrupt (IS_S_RX_FIFO_FULL_SHIFT) will be generated when RX fifo becomes full. This can happen if master issues write request of more than 64 bytes.

---
 drivers/i2c/busses/i2c-bcm-iproc.c | 21 +++++++++++++++++----
 1 file changed, 17 insertions(+), 4 deletions(-)
drivers/i2c/busses/i2c-bcm-iproc.c +17 -4
index 22e04055b447..cceaf69279a9 100644
@@ -313,6 +313,8 @@ static void bcm_iproc_i2c_slave_init(
 
 	/* Enable interrupt register to indicate a valid byte in receive fifo */
 	val = BIT(IE_S_RX_EVENT_SHIFT);
+	/* Enable interrupt register to indicate Slave Rx FIFO Full */
+	val |= BIT(IE_S_RX_FIFO_FULL_SHIFT);
 	/* Enable interrupt register to indicate a Master read transaction */
 	val |= BIT(IE_S_RD_EVENT_SHIFT);
 	/* Enable interrupt register for the Slave BUSY command */
@@ -434,9 +436,15 @@ static bool bcm_iproc_i2c_slave_isr(struct bcm_iproc_i2c_dev *iproc_i2c,
 	 *                    events
 	 * Master-read      : both IS_S_RX_EVENT_SHIFT and IS_S_RD_EVENT_SHIFT
 	 *                    events or only IS_S_RD_EVENT_SHIFT
+	 *
+	 * iproc has a slave rx fifo size of 64 bytes. Rx fifo full interrupt
+	 * (IS_S_RX_FIFO_FULL_SHIFT) will be generated when RX fifo becomes
+	 * full. This can happen if Master issues write requests of more than
+	 * 64 bytes.
 	 */
 	if (status & BIT(IS_S_RX_EVENT_SHIFT) ||
-	    status & BIT(IS_S_RD_EVENT_SHIFT)) {
+	    status & BIT(IS_S_RD_EVENT_SHIFT) ||
+	    status & BIT(IS_S_RX_FIFO_FULL_SHIFT)) {
 		/* disable slave interrupts */
 		val = iproc_i2c_rd_reg(iproc_i2c, IE_OFFSET);
 		val &= ~iproc_i2c->slave_int_mask;
@@ -452,9 +460,14 @@ static bool bcm_iproc_i2c_slave_isr(struct bcm_iproc_i2c_dev *iproc_i2c,
 		/* schedule tasklet to read data later */
 		tasklet_schedule(&iproc_i2c->slave_rx_tasklet);
 
-		/* clear only IS_S_RX_EVENT_SHIFT interrupt */
-		iproc_i2c_wr_reg(iproc_i2c, IS_OFFSET,
-				 BIT(IS_S_RX_EVENT_SHIFT));
+		/*
+		 * clear only IS_S_RX_EVENT_SHIFT and
+		 * IS_S_RX_FIFO_FULL_SHIFT interrupt.
+		 */
+		val = BIT(IS_S_RX_EVENT_SHIFT);
+		if (status & BIT(IS_S_RX_FIFO_FULL_SHIFT))
+			val |= BIT(IS_S_RX_FIFO_FULL_SHIFT);
+		iproc_i2c_wr_reg(iproc_i2c, IS_OFFSET, val);
 	}
 
 	if (status & BIT(IS_S_TX_UNDERRUN_SHIFT)) {
Signature · 1 line
2.17.1
#17 lore raw

Re: [PATCH v1 6/6] i2c: iproc: handle rx fifo full interrupt

Quoted 12 lines · Dhananjay Phadke

From: Rayagonda Kokatanur <[email protected]>

Quoted 8 lines · Rayagonda Kokatanur

Add code to handle IS_S_RX_FIFO_FULL_SHIFT interrupt to support master write request with >= 64 bytes.

Iproc has a slave rx fifo size of 64 bytes.
Rx fifo full interrupt (IS_S_RX_FIFO_FULL_SHIFT) will be generated when RX fifo becomes full. This can happen if master issues write request of more than 64 bytes.

Quoted 3 lines

ARM cores run much faster than I2C bus, why would rx fifo go full when rx interrupt is enabled and bytes are read out by bus driver isr? Isn't fifo read pointer updated on these byte reads?

Hi Rayagonda,

Could you please reply on this question? For transactions > 64 bytes, do we batch until RX FIFO is full before we read out the data?

Thanks,

Ray

Quoted 3 lines

Does controller stretch clock when rx fifo is full (e.g. kernel has crashed, bus driver isn't draining fifo)?

#18 lore raw

Re: [PATCH v1 6/6] i2c: iproc: handle rx fifo full interrupt

Hi Dhanajay,

Quoted 16 lines · Ray Jui
Quoted 12 lines · Dhananjay Phadke

From: Rayagonda Kokatanur <[email protected]>

Quoted 8 lines · Rayagonda Kokatanur

Add code to handle IS_S_RX_FIFO_FULL_SHIFT interrupt to support master write request with >= 64 bytes.

Iproc has a slave rx fifo size of 64 bytes.
Rx fifo full interrupt (IS_S_RX_FIFO_FULL_SHIFT) will be generated when RX fifo becomes full. This can happen if master issues write request of more than 64 bytes.

Quoted 8 lines
Quoted 3 lines

ARM cores run much faster than I2C bus, why would rx fifo go full when rx interrupt is enabled and bytes are read out by bus driver isr? Isn't fifo read pointer updated on these byte reads?

Hi Rayagonda,

Could you please reply on this question? For transactions > 64 bytes, do we batch until RX FIFO is full before we read out the data?

Sorry I missed this question.
Yes with current design we are batching 64 bytes for translation > 64 bytes.

Best regards,
Rayagonda

Quoted 8 lines

Thanks,

Ray

Quoted 3 lines

Does controller stretch clock when rx fifo is full (e.g. kernel has crashed, bus driver isn't draining fifo)?

#19 lore raw

Re: [PATCH v1 6/6] i2c: iproc: handle rx fifo full interrupt

Quoted 23 lines · Rayagonda Kokatanur

Hi Dhanajay,

Quoted 20 lines · Ray Jui
Quoted 15 lines · Dhananjay Phadke

From: Rayagonda Kokatanur <[email protected]>

Quoted 8 lines · Rayagonda Kokatanur

Add code to handle IS_S_RX_FIFO_FULL_SHIFT interrupt to support master write request with >= 64 bytes.

Iproc has a slave rx fifo size of 64 bytes.
Rx fifo full interrupt (IS_S_RX_FIFO_FULL_SHIFT) will be generated when RX fifo becomes full. This can happen if master issues write request of more than 64 bytes.

ARM cores run much faster than I2C bus, why would rx fifo go full when rx interrupt is enabled and bytes are read out by bus driver isr? Isn't fifo read pointer updated on these byte reads?

Quoted 8 lines
Quoted 4 lines

Hi Rayagonda,

Could you please reply on this question? For transactions > 64 bytes, do we batch until RX FIFO is full before we read out the data?

Sorry I missed this question.
Yes with current design we are batching 64 bytes for translation > 64 bytes.

So we do batch the transfer and read them in one shot, and that's how the FIFO full interrupt is being utilized for. That sounds okay to me.

Thanks,

Ray

Quoted 12 lines

Best regards,
Rayagonda

Quoted 8 lines

Thanks,

Ray

Quoted 3 lines

Does controller stretch clock when rx fifo is full (e.g. kernel has crashed, bus driver isn't draining fifo)?

_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel