Skip to contentSkip to conversation

i2c: i801: Series with minor improvements

PATCH Superseded Heiner Kallweit 2 participants 26 messages Apr 15, 2022 – Dec 17, 2022

Subsystems (2)

v1 of 2 Superseded Latest is v2 →

lore
Share & tools
lore
#0 lore raw

cover i2c: i801: Series with minor improvements

This series includes a number of minor improvements, partially it's a re-send of patches submitted in December last year already.

Heiner Kallweit (8):
i2c: i801: improve interrupt handler
i2c: i801: make FEATURE_HOST_NOTIFY dependent on FEATURE_IRQ
i2c: i801: make FEATURE_BLOCK_PROC dependent on FEATURE_BLOCK_BUFFER
i2c: i801: enable FEATURE_IRQ and FEATURE_I2C_BLOCK_READ on all chip
versions
i2c: i801: add helper i801_set_hstadd
i2c: i801: add i801_single_transaction(), complementing

    i801_block_transaction()

i2c: i801: call i801_check_pre() from i801_access()
i2c: i801: call i801_check_post() from i801_access()

 drivers/i2c/busses/i2c-i801.c | 339 ++++++++++++++++++----------------
 1 file changed, 175 insertions(+), 164 deletions(-)
Signature · 1 line
2.35.3
#1 lore raw

PATCH 1/8 i2c: i801: improve interrupt handler

Not sure if it can happen, but better play safe: If SMBHSTSTS_BYTE_DONE and an error flag is set, then don't trust the result and skip calling i801_isr_byte_done(). In addition clear status bit SMBHSTSTS_BYTE_DONE in the main interrupt handler, this allows to simplify the code a little.

---
 drivers/i2c/busses/i2c-i801.c | 25 ++++++++-----------------
 1 file changed, 8 insertions(+), 17 deletions(-)
drivers/i2c/busses/i2c-i801.c +8 -17
index ff706349b..c481f121d 100644
@@ -556,9 +556,6 @@ static void i801_isr_byte_done(struct i801_priv *priv)
 		/* Write next byte, except for IRQ after last byte */
 		outb_p(priv->data[++priv->count], SMBBLKDAT(priv));
 	}
-
-	/* Clear BYTE_DONE to continue with next byte */
-	outb_p(SMBHSTSTS_BYTE_DONE, SMBHSTSTS(priv));
 }
 
 static irqreturn_t i801_host_notify_isr(struct i801_priv *priv)
@@ -588,7 +585,6 @@ static irqreturn_t i801_host_notify_isr(struct i801_priv *priv)
  *      BUS_ERR - SMI# transaction collision
  *      FAILED - transaction was canceled due to a KILL request
  *    When any of these occur, update ->status and signal completion.
- *    ->status must be cleared before kicking off the next transaction.
  *
  * 2) For byte-by-byte (I2C read/write) transactions, one BYTE_DONE interrupt
  *    occurs for each byte of a byte-by-byte to prepare the next byte.
@@ -613,23 +609,18 @@ static irqreturn_t i801_isr(int irq, void *dev_id)
 	}
 
 	status = inb_p(SMBHSTSTS(priv));
-	if (status & SMBHSTSTS_BYTE_DONE)
+	if ((status & SMBHSTSTS_BYTE_DONE) && !(status & STATUS_ERROR_FLAGS))
 		i801_isr_byte_done(priv);
 
 	/*
-	 * Clear remaining IRQ sources: Completion of last command, errors
-	 * and the SMB_ALERT signal. SMB_ALERT status is set after signal
-	 * assertion independently of the interrupt generation being blocked
-	 * or not so clear it always when the status is set.
-	 */
-	status &= SMBHSTSTS_INTR | STATUS_ERROR_FLAGS | SMBHSTSTS_SMBALERT_STS;
-	if (status)
-		outb_p(status, SMBHSTSTS(priv));
-	status &= ~SMBHSTSTS_SMBALERT_STS; /* SMB_ALERT not reported */
-	/*
-	 * Report transaction result.
-	 * ->status must be cleared before the next transaction is started.
+	 * Clear IRQ sources: SMB_ALERT status is set after signal assertion
+	 * independently of the interrupt generation being blocked or not
+	 * so clear it always when the status is set.
 	 */
+	status &= STATUS_FLAGS | SMBHSTSTS_SMBALERT_STS;
+	outb_p(status, SMBHSTSTS(priv));
+
+	status &= STATUS_ERROR_FLAGS | SMBHSTSTS_INTR;
 	if (status) {
 		priv->status = status;
 		complete(&priv->done);
Signature · 1 line
2.35.3
#2 lore raw

Re: [PATCH 1/8] i2c: i801: improve interrupt handler

Hi Heiner,

Quoted 34 lines · Heiner Kallweit

Not sure if it can happen, but better play safe: If SMBHSTSTS_BYTE_DONE and an error flag is set, then don't trust the result and skip calling i801_isr_byte_done(). In addition clear status bit SMBHSTSTS_BYTE_DONE in the main interrupt handler, this allows to simplify the code a little.

---
 drivers/i2c/busses/i2c-i801.c | 25 ++++++++-----------------
 1 file changed, 8 insertions(+), 17 deletions(-)
drivers/i2c/busses/i2c-i801.c +0 -4
index ff706349b..c481f121d 100644
@@ -556,9 +556,6 @@ static void i801_isr_byte_done(struct i801_priv *priv)
 		/* Write next byte, except for IRQ after last byte */
 		outb_p(priv->data[++priv->count], SMBBLKDAT(priv));
 	}
-
-	/* Clear BYTE_DONE to continue with next byte */
-	outb_p(SMBHSTSTS_BYTE_DONE, SMBHSTSTS(priv));
 }
 
 static irqreturn_t i801_host_notify_isr(struct i801_priv *priv)
@@ -588,7 +585,6 @@ static irqreturn_t i801_host_notify_isr(struct i801_priv *priv)
  *      BUS_ERR - SMI# transaction collision
  *      FAILED - transaction was canceled due to a KILL request
  *    When any of these occur, update ->status and signal completion.
- *    ->status must be cleared before kicking off the next transaction.
  *
  * 2) For byte-by-byte (I2C read/write) transactions, one BYTE_DONE interrupt
  *    occurs for each byte of a byte-by-byte to prepare the next byte.

drivers/i2c/busses/i2c-i801.c static irqreturn_t i801_isr(int irq, void *dev_id) · line 612

Quoted 5 lines
drivers/i2c/busses/i2c-i801.c +1 -1
@@ -613,23 +609,18 @@ static irqreturn_t i801_isr(int irq, void *dev_id)
 	}
 
 	status = inb_p(SMBHSTSTS(priv));
-	if (status & SMBHSTSTS_BYTE_DONE)
+	if ((status & SMBHSTSTS_BYTE_DONE) && !(status & STATUS_ERROR_FLAGS))

Isn't this better written

	if ((status & (SMBHSTSTS_BYTE_DONE | STATUS_ERROR_FLAGS)) == SMBHSTSTS_BYTE_DONE)

? At least my compiler generates smaller binary code.

Quoted 11 lines
drivers/i2c/busses/i2c-i801.c +0 -8
 		i801_isr_byte_done(priv);
 
 	/*
-	 * Clear remaining IRQ sources: Completion of last command, errors
-	 * and the SMB_ALERT signal. SMB_ALERT status is set after signal
-	 * assertion independently of the interrupt generation being blocked
-	 * or not so clear it always when the status is set.
-	 */
-	status &= SMBHSTSTS_INTR | STATUS_ERROR_FLAGS | SMBHSTSTS_SMBALERT_STS;
-	if (status)
-		outb_p(status, SMBHSTSTS(priv));

drivers/i2c/busses/i2c-i801.c static irqreturn_t i801_isr(int irq, void *dev_id) · line 621

Quoted 10 lines
drivers/i2c/busses/i2c-i801.c +5 -4
-	status &= ~SMBHSTSTS_SMBALERT_STS; /* SMB_ALERT not reported */
-	/*
-	 * Report transaction result.
-	 * ->status must be cleared before the next transaction is started.
+	 * Clear IRQ sources: SMB_ALERT status is set after signal assertion
+	 * independently of the interrupt generation being blocked or not
+	 * so clear it always when the status is set.
 	 */
+	status &= STATUS_FLAGS | SMBHSTSTS_SMBALERT_STS;
+	outb_p(status, SMBHSTSTS(priv));

You are making the call to outb_p() unconditional. Is this under the assumption that at least one of the bits must be set, so the condition was always met?

drivers/i2c/busses/i2c-i801.c static irqreturn_t i801_isr(int irq, void *dev_id) · line 626

Quoted 5 lines
drivers/i2c/busses/i2c-i801.c +2 -0
+
+	status &= STATUS_ERROR_FLAGS | SMBHSTSTS_INTR;
 	if (status) {
 		priv->status = status;
 		complete(&priv->done);

Tested OK on my system, looks good overall, nice simplification.

Signature · 2 lines
Jean Delvare
SUSE L3 Support
#3 lore raw

Re: [PATCH 1/8] i2c: i801: improve interrupt handler

Quoted 45 lines · Jean Delvare

Hi Heiner,

Quoted 34 lines · Heiner Kallweit

Not sure if it can happen, but better play safe: If SMBHSTSTS_BYTE_DONE and an error flag is set, then don't trust the result and skip calling i801_isr_byte_done(). In addition clear status bit SMBHSTSTS_BYTE_DONE in the main interrupt handler, this allows to simplify the code a little.

---
 drivers/i2c/busses/i2c-i801.c | 25 ++++++++-----------------
 1 file changed, 8 insertions(+), 17 deletions(-)
drivers/i2c/busses/i2c-i801.c +0 -4
index ff706349b..c481f121d 100644
@@ -556,9 +556,6 @@ static void i801_isr_byte_done(struct i801_priv *priv)
 		/* Write next byte, except for IRQ after last byte */
 		outb_p(priv->data[++priv->count], SMBBLKDAT(priv));
 	}
-
-	/* Clear BYTE_DONE to continue with next byte */
-	outb_p(SMBHSTSTS_BYTE_DONE, SMBHSTSTS(priv));
 }
 
 static irqreturn_t i801_host_notify_isr(struct i801_priv *priv)
@@ -588,7 +585,6 @@ static irqreturn_t i801_host_notify_isr(struct i801_priv *priv)
  *      BUS_ERR - SMI# transaction collision
  *      FAILED - transaction was canceled due to a KILL request
  *    When any of these occur, update ->status and signal completion.
- *    ->status must be cleared before kicking off the next transaction.
  *
  * 2) For byte-by-byte (I2C read/write) transactions, one BYTE_DONE interrupt
  *    occurs for each byte of a byte-by-byte to prepare the next byte.

drivers/i2c/busses/i2c-i801.c static irqreturn_t i801_isr(int irq, void *dev_id) · line 612

Quoted 5 lines
drivers/i2c/busses/i2c-i801.c +1 -1
@@ -613,23 +609,18 @@ static irqreturn_t i801_isr(int irq, void *dev_id)
 	}
 
 	status = inb_p(SMBHSTSTS(priv));
-	if (status & SMBHSTSTS_BYTE_DONE)
+	if ((status & SMBHSTSTS_BYTE_DONE) && !(status & STATUS_ERROR_FLAGS))

Isn't this better written

Quoted 4 lines
	if ((status & (SMBHSTSTS_BYTE_DONE | STATUS_ERROR_FLAGS)) == SMBHSTSTS_BYTE_DONE)

? At least my compiler generates smaller binary code.

OK. Meanwhile I prefer readability over saving few bytes, but here IMO the readability doesn't suffer.

Quoted 17 lines
Quoted 17 lines
drivers/i2c/busses/i2c-i801.c +2 -12
 		i801_isr_byte_done(priv);
 
 	/*
-	 * Clear remaining IRQ sources: Completion of last command, errors
-	 * and the SMB_ALERT signal. SMB_ALERT status is set after signal
-	 * assertion independently of the interrupt generation being blocked
-	 * or not so clear it always when the status is set.
-	 */
-	status &= SMBHSTSTS_INTR | STATUS_ERROR_FLAGS | SMBHSTSTS_SMBALERT_STS;
-	if (status)
-		outb_p(status, SMBHSTSTS(priv));
-	status &= ~SMBHSTSTS_SMBALERT_STS; /* SMB_ALERT not reported */
-	/*
-	 * Report transaction result.
-	 * ->status must be cleared before the next transaction is started.
+	 * Clear IRQ sources: SMB_ALERT status is set after signal assertion
+	 * independently of the interrupt generation being blocked or not
Quoted 9 lines

drivers/i2c/busses/i2c-i801.c static irqreturn_t i801_isr(int irq, void *dev_id) · line 621

Quoted 4 lines
drivers/i2c/busses/i2c-i801.c +3 -0
+	 * so clear it always when the status is set.
 	 */
+	status &= STATUS_FLAGS | SMBHSTSTS_SMBALERT_STS;
+	outb_p(status, SMBHSTSTS(priv));

You are making the call to outb_p() unconditional. Is this under the assumption that at least one of the bits must be set, so the condition was always met?

So far we exclude here the case where SMBHSTSTS is written in i801_isr_byte_done(). The patch allows to write SMBHSTSTS here only, so we don't need the condition any longer.

Quoted 8 lines

drivers/i2c/busses/i2c-i801.c static irqreturn_t i801_isr(int irq, void *dev_id) · line 626

Quoted 5 lines
drivers/i2c/busses/i2c-i801.c +2 -0
+
+	status &= STATUS_ERROR_FLAGS | SMBHSTSTS_INTR;
 	if (status) {
 		priv->status = status;
 		complete(&priv->done);

Tested OK on my system, looks good overall, nice simplification.

#4 lore raw

PATCH 2/8 i2c: i801: make FEATURE_HOST_NOTIFY dependent on FEATURE_IRQ

In reply to Heiner Kallweit

Host notification uses an interrupt, therefore it makes sense only if interrupts are enabled. Make this dependency explicit by removing FEATURE_HOST_NOTIFY if FEATURE_IRQ isn't set.

---
 drivers/i2c/busses/i2c-i801.c | 3 +++
 1 file changed, 3 insertions(+)
drivers/i2c/busses/i2c-i801.c +3 -0
index c481f121d..eccdc7035 100644
@@ -1756,6 +1756,9 @@ static int i801_probe(struct pci_dev *dev, const struct pci_device_id *id)
 	dev_info(&dev->dev, "SMBus using %s\n",
 		 priv->features & FEATURE_IRQ ? "PCI interrupt" : "polling");
 
+	if (!(priv->features & FEATURE_IRQ))
+		priv->features &= ~FEATURE_HOST_NOTIFY;
+
 	i801_add_tco(priv);
 
 	snprintf(priv->adapter.name, sizeof(priv->adapter.name),
Signature · 1 line
2.35.3
#5 lore raw

Re: [PATCH 2/8] i2c: i801: make FEATURE_HOST_NOTIFY dependent on FEATURE_IRQ

Hi Heiner,

Quoted 2 lines · Heiner Kallweit

Host notification uses an interrupt, therefore it makes sense only if interrupts are enabled.

It would be nice to have this comment in the code itself.

Quoted 13 lines

Make this dependency explicit by removing
FEATURE_HOST_NOTIFY if FEATURE_IRQ isn't set.

---
 drivers/i2c/busses/i2c-i801.c | 3 +++
 1 file changed, 3 insertions(+)
drivers/i2c/busses/i2c-i801.c +0 -0
index c481f121d..eccdc7035 100644

drivers/i2c/busses/i2c-i801.c static int i801_probe(struct pci_dev *dev, const struct pci_device_id *id) · line 1760

Quoted 5 lines
drivers/i2c/busses/i2c-i801.c +2 -0
@@ -1756,6 +1756,9 @@ static int i801_probe(struct pci_dev *dev, const struct pci_device_id *id)
 	dev_info(&dev->dev, "SMBus using %s\n",
 		 priv->features & FEATURE_IRQ ? "PCI interrupt" : "polling");
 
+	if (!(priv->features & FEATURE_IRQ))
+		priv->features &= ~FEATURE_HOST_NOTIFY;

Earlier in this function, there's an action which depends on the FEATURE_HOST_NOTIFY flag being set. While this will only result in a useless action and wouldn't cause a bug as far as I can see, wouldn't it be cleaner to move that piece of code *after* the check you're adding?

drivers/i2c/busses/i2c-i801.c static int i801_probe(struct pci_dev *dev, const struct pci_device_id *id) · line 1764

Quoted 4 lines
drivers/i2c/busses/i2c-i801.c +1 -0
+
 	i801_add_tco(priv);
 
 	snprintf(priv->adapter.name, sizeof(priv->adapter.name),

Looks good, tested OK on my system (non-regression only, I'm not using the Host Notify feature).

Signature · 2 lines
Jean Delvare
SUSE L3 Support
#6 lore raw

Re: [PATCH 2/8] i2c: i801: make FEATURE_HOST_NOTIFY dependent on FEATURE_IRQ

Quoted 8 lines · Jean Delvare

Hi Heiner,

Quoted 2 lines · Heiner Kallweit

Host notification uses an interrupt, therefore it makes sense only if interrupts are enabled.

It would be nice to have this comment in the code itself.

OK

Quoted 16 lines
Quoted 16 lines

Make this dependency explicit by removing
FEATURE_HOST_NOTIFY if FEATURE_IRQ isn't set.

---
 drivers/i2c/busses/i2c-i801.c | 3 +++
 1 file changed, 3 insertions(+)
drivers/i2c/busses/i2c-i801.c +0 -0
index c481f121d..eccdc7035 100644
@@ -1756,6 +1756,9 @@ static int i801_probe(struct pci_dev *dev, const struct pci_device_id *id)
 	dev_info(&dev->dev, "SMBus using %s\n",
 		 priv->features & FEATURE_IRQ ? "PCI interrupt" : "polling");
 
Quoted 9 lines

drivers/i2c/busses/i2c-i801.c static int i801_probe(struct pci_dev *dev, const struct pci_device_id *id) · line 1760

Quoted 2 lines
drivers/i2c/busses/i2c-i801.c +2 -0
+	if (!(priv->features & FEATURE_IRQ))
+		priv->features &= ~FEATURE_HOST_NOTIFY;

Earlier in this function, there's an action which depends on the FEATURE_HOST_NOTIFY flag being set. While this will only result in a useless action and wouldn't cause a bug as far as I can see, wouldn't it be cleaner to move that piece of code *after* the check you're adding?

Yes, this would be better. Will be part of v2.

Quoted 8 lines

drivers/i2c/busses/i2c-i801.c static int i801_probe(struct pci_dev *dev, const struct pci_device_id *id) · line 1764

Quoted 4 lines
drivers/i2c/busses/i2c-i801.c +1 -0
+
 	i801_add_tco(priv);
 
 	snprintf(priv->adapter.name, sizeof(priv->adapter.name),

Looks good, tested OK on my system (non-regression only, I'm not using the Host Notify feature).

#7 lore raw

PATCH 3/8 i2c: i801: make FEATURE_BLOCK_PROC dependent on FEATURE_BLOCK_BUFFER

In reply to Heiner Kallweit

According to the datasheet the block process call requires block buffer mode. The user may disable block buffer mode by module parameter disable_features, in such a case we have to clear FEATURE_BLOCK_PROC.

---
 drivers/i2c/busses/i2c-i801.c | 3 +++
 1 file changed, 3 insertions(+)
drivers/i2c/busses/i2c-i801.c +3 -0
index eccdc7035..1d8182901 100644
@@ -1675,6 +1675,9 @@ static int i801_probe(struct pci_dev *dev, const struct pci_device_id *id)
 	}
 	priv->features &= ~disable_features;
 
+	if (!(priv->features & FEATURE_BLOCK_BUFFER))
+		priv->features &= ~FEATURE_BLOCK_PROC;
+
 	err = pcim_enable_device(dev);
 	if (err) {
 		dev_err(&dev->dev, "Failed to enable SMBus PCI device (%d)\n",
Signature · 1 line
2.35.3
#8 lore raw

Re: [PATCH 3/8] i2c: i801: make FEATURE_BLOCK_PROC dependent on FEATURE_BLOCK_BUFFER

Hi Heiner,

Quoted 4 lines · Heiner Kallweit

According to the datasheet the block process call requires block buffer mode. The user may disable block buffer mode by module parameter disable_features, in such a case we have to clear FEATURE_BLOCK_PROC.

In which datasheet are you seeing this? Can you point me to the specific section and/or quote the statement? I can't find it in the datasheet I'm looking at (ICH9, Intel document 316972-002) but it is huge and I may just be missing it.

Also, same request as previous patch, I'd like a comment in the code, so that developers don't have to read the git log to figure out why this piece of code is there.

Furthermore, as far as I can see, the FEATURE_BLOCK_PROC flag only affects the value returned by i801_func(). i801_access() does not verify whether this flag is set before processing a command where size == I2C_SMBUS_BLOCK_PROC_CALL. I think it should? Otherwise your fix is only partial (will work if the device driver calls .functionality as it is supposed to, will fail with - I suppose - unpredictable results if the device driver calls .smbus_xfer directly).

Quoted 10 lines
---
 drivers/i2c/busses/i2c-i801.c | 3 +++
 1 file changed, 3 insertions(+)
drivers/i2c/busses/i2c-i801.c +0 -0
index eccdc7035..1d8182901 100644

drivers/i2c/busses/i2c-i801.c static int i801_probe(struct pci_dev *dev, const struct pci_device_id *id) · line 1683

Quoted 9 lines
drivers/i2c/busses/i2c-i801.c +3 -0
@@ -1675,6 +1675,9 @@ static int i801_probe(struct pci_dev *dev, const struct pci_device_id *id)
 	}
 	priv->features &= ~disable_features;
 
+	if (!(priv->features & FEATURE_BLOCK_BUFFER))
+		priv->features &= ~FEATURE_BLOCK_PROC;
+
 	err = pcim_enable_device(dev);
 	if (err) {
 		dev_err(&dev->dev, "Failed to enable SMBus PCI device (%d)\n",

Thanks,

Signature · 2 lines
Jean Delvare
SUSE L3 Support
#9 lore raw

Re: [PATCH 3/8] i2c: i801: make FEATURE_BLOCK_PROC dependent on FEATURE_BLOCK_BUFFER

Quoted 8 lines · Jean Delvare

Hi Heiner,

Quoted 4 lines · Heiner Kallweit

According to the datasheet the block process call requires block buffer mode. The user may disable block buffer mode by module parameter disable_features, in such a case we have to clear FEATURE_BLOCK_PROC.

Quoted 5 lines

In which datasheet are you seeing this? Can you point me to the specific section and/or quote the statement? I can't find it in the datasheet I'm looking at (ICH9, Intel document 316972-002) but it is huge and I may just be missing it.

I used the following datasheet:
Intel 9 Series Chipset Family PCH
330550-002

On page 211 the block process call is described.
There's a note: E32B bit in the Auxiliary Control register must be set when using this protocol. The same note can be found on page 663.

Quoted 4 lines

Also, same request as previous patch, I'd like a comment in the code, so that developers don't have to read the git log to figure out why this piece of code is there.

OK

Quoted 8 lines

Furthermore, as far as I can see, the FEATURE_BLOCK_PROC flag only affects the value returned by i801_func(). i801_access() does not verify whether this flag is set before processing a command where size == I2C_SMBUS_BLOCK_PROC_CALL. I think it should? Otherwise your fix is only partial (will work if the device driver calls .functionality as it is supposed to, will fail with - I suppose - unpredictable results if the device driver calls .smbus_xfer directly).

If FEATURE_BLOCK_PROC isn't set then we would call i801_block_transaction_byte_by_byte() according to the following code in i801_block_transaction().

        if ((priv->features & FEATURE_BLOCK_BUFFER) &&
            command != I2C_SMBUS_I2C_BLOCK_DATA)
                result = i801_block_transaction_by_block(priv, data,
                                                         read_write,
                                                         command);
        else
                result = i801_block_transaction_byte_by_byte(priv, data,
                                                             read_write,
                                                             command);

And i801_block_transaction_byte_by_byte() immediately returns -EOPNOTSUPP in case of command == I2C_SMBUS_BLOCK_PROC_CALL.

Having said that the requested check is there, it's just executed a little bit later.

Quoted 21 lines
Quoted 10 lines
---
 drivers/i2c/busses/i2c-i801.c | 3 +++
 1 file changed, 3 insertions(+)
drivers/i2c/busses/i2c-i801.c +0 -0
index eccdc7035..1d8182901 100644

drivers/i2c/busses/i2c-i801.c static int i801_probe(struct pci_dev *dev, const struct pci_device_id *id) · line 1683

Quoted 9 lines
drivers/i2c/busses/i2c-i801.c +3 -0
@@ -1675,6 +1675,9 @@ static int i801_probe(struct pci_dev *dev, const struct pci_device_id *id)
 	}
 	priv->features &= ~disable_features;
 
+	if (!(priv->features & FEATURE_BLOCK_BUFFER))
+		priv->features &= ~FEATURE_BLOCK_PROC;
+
 	err = pcim_enable_device(dev);
 	if (err) {
 		dev_err(&dev->dev, "Failed to enable SMBus PCI device (%d)\n",

Thanks,

#10 lore raw

PATCH 4/8 i2c: i801: enable FEATURE_IRQ and FEATURE_I2C_BLOCK_READ on all chip versions

In reply to Heiner Kallweit

According to the datasheets interrupt mode and i2c block read are supported on all chip versions. Therefore set both feature flags for all chip versions.
Note: Don't remove the two feature flags as such (at least for now), so that in case of a problem users can use the disable_features module parameter to disable a problematic feature.

Patch is based solely on the datasheets. I don't have old enough test hw, therefore the patch is compile-tested only.

---
 drivers/i2c/busses/i2c-i801.c | 140 +++++++++++++++++-----------------
 1 file changed, 69 insertions(+), 71 deletions(-)
drivers/i2c/busses/i2c-i801.c +69 -71
index 1d8182901..a9737f14d 100644
@@ -12,70 +12,70 @@
 /*
  * Supports the following Intel I/O Controller Hubs (ICH):
  *
- *					I/O			Block	I2C
- *					region	SMBus	Block	proc.	block
- * Chip name			PCI ID	size	PEC	buffer	call	read
- * ---------------------------------------------------------------------------
- * 82801AA (ICH)		0x2413	16	no	no	no	no
- * 82801AB (ICH0)		0x2423	16	no	no	no	no
- * 82801BA (ICH2)		0x2443	16	no	no	no	no
- * 82801CA (ICH3)		0x2483	32	soft	no	no	no
- * 82801DB (ICH4)		0x24c3	32	hard	yes	no	no
- * 82801E (ICH5)		0x24d3	32	hard	yes	yes	yes
- * 6300ESB			0x25a4	32	hard	yes	yes	yes
- * 82801F (ICH6)		0x266a	32	hard	yes	yes	yes
- * 6310ESB/6320ESB		0x269b	32	hard	yes	yes	yes
- * 82801G (ICH7)		0x27da	32	hard	yes	yes	yes
- * 82801H (ICH8)		0x283e	32	hard	yes	yes	yes
- * 82801I (ICH9)		0x2930	32	hard	yes	yes	yes
- * EP80579 (Tolapai)		0x5032	32	hard	yes	yes	yes
- * ICH10			0x3a30	32	hard	yes	yes	yes
- * ICH10			0x3a60	32	hard	yes	yes	yes
- * 5/3400 Series (PCH)		0x3b30	32	hard	yes	yes	yes
- * 6 Series (PCH)		0x1c22	32	hard	yes	yes	yes
- * Patsburg (PCH)		0x1d22	32	hard	yes	yes	yes
- * Patsburg (PCH) IDF		0x1d70	32	hard	yes	yes	yes
- * Patsburg (PCH) IDF		0x1d71	32	hard	yes	yes	yes
- * Patsburg (PCH) IDF		0x1d72	32	hard	yes	yes	yes
- * DH89xxCC (PCH)		0x2330	32	hard	yes	yes	yes
- * Panther Point (PCH)		0x1e22	32	hard	yes	yes	yes
- * Lynx Point (PCH)		0x8c22	32	hard	yes	yes	yes
- * Lynx Point-LP (PCH)		0x9c22	32	hard	yes	yes	yes
- * Avoton (SOC)			0x1f3c	32	hard	yes	yes	yes
- * Wellsburg (PCH)		0x8d22	32	hard	yes	yes	yes
- * Wellsburg (PCH) MS		0x8d7d	32	hard	yes	yes	yes
- * Wellsburg (PCH) MS		0x8d7e	32	hard	yes	yes	yes
- * Wellsburg (PCH) MS		0x8d7f	32	hard	yes	yes	yes
- * Coleto Creek (PCH)		0x23b0	32	hard	yes	yes	yes
- * Wildcat Point (PCH)		0x8ca2	32	hard	yes	yes	yes
- * Wildcat Point-LP (PCH)	0x9ca2	32	hard	yes	yes	yes
- * BayTrail (SOC)		0x0f12	32	hard	yes	yes	yes
- * Braswell (SOC)		0x2292	32	hard	yes	yes	yes
- * Sunrise Point-H (PCH) 	0xa123  32	hard	yes	yes	yes
- * Sunrise Point-LP (PCH)	0x9d23	32	hard	yes	yes	yes
- * DNV (SOC)			0x19df	32	hard	yes	yes	yes
- * Emmitsburg (PCH)		0x1bc9	32	hard	yes	yes	yes
- * Broxton (SOC)		0x5ad4	32	hard	yes	yes	yes
- * Lewisburg (PCH)		0xa1a3	32	hard	yes	yes	yes
- * Lewisburg Supersku (PCH)	0xa223	32	hard	yes	yes	yes
- * Kaby Lake PCH-H (PCH)	0xa2a3	32	hard	yes	yes	yes
- * Gemini Lake (SOC)		0x31d4	32	hard	yes	yes	yes
- * Cannon Lake-H (PCH)		0xa323	32	hard	yes	yes	yes
- * Cannon Lake-LP (PCH)		0x9da3	32	hard	yes	yes	yes
- * Cedar Fork (PCH)		0x18df	32	hard	yes	yes	yes
- * Ice Lake-LP (PCH)		0x34a3	32	hard	yes	yes	yes
- * Ice Lake-N (PCH)		0x38a3	32	hard	yes	yes	yes
- * Comet Lake (PCH)		0x02a3	32	hard	yes	yes	yes
- * Comet Lake-H (PCH)		0x06a3	32	hard	yes	yes	yes
- * Elkhart Lake (PCH)		0x4b23	32	hard	yes	yes	yes
- * Tiger Lake-LP (PCH)		0xa0a3	32	hard	yes	yes	yes
- * Tiger Lake-H (PCH)		0x43a3	32	hard	yes	yes	yes
- * Jasper Lake (SOC)		0x4da3	32	hard	yes	yes	yes
- * Comet Lake-V (PCH)		0xa3a3	32	hard	yes	yes	yes
- * Alder Lake-S (PCH)		0x7aa3	32	hard	yes	yes	yes
- * Alder Lake-P (PCH)		0x51a3	32	hard	yes	yes	yes
- * Alder Lake-M (PCH)		0x54a3	32	hard	yes	yes	yes
- * Raptor Lake-S (PCH)		0x7a23	32	hard	yes	yes	yes
+ *					I/O			Block
+ *					region	SMBus	Block	proc.
+ * Chip name			PCI ID	size	PEC	buffer	call
+ * -------------------------------------------------------------------
+ * 82801AA (ICH)		0x2413	16	no	no	no
+ * 82801AB (ICH0)		0x2423	16	no	no	no
+ * 82801BA (ICH2)		0x2443	16	no	no	no
+ * 82801CA (ICH3)		0x2483	32	soft	no	no
+ * 82801DB (ICH4)		0x24c3	32	hard	yes	no
+ * 82801E (ICH5)		0x24d3	32	hard	yes	yes
+ * 6300ESB			0x25a4	32	hard	yes	yes
+ * 82801F (ICH6)		0x266a	32	hard	yes	yes
+ * 6310ESB/6320ESB		0x269b	32	hard	yes	yes
+ * 82801G (ICH7)		0x27da	32	hard	yes	yes
+ * 82801H (ICH8)		0x283e	32	hard	yes	yes
+ * 82801I (ICH9)		0x2930	32	hard	yes	yes
+ * EP80579 (Tolapai)		0x5032	32	hard	yes	yes
+ * ICH10			0x3a30	32	hard	yes	yes
+ * ICH10			0x3a60	32	hard	yes	yes
+ * 5/3400 Series (PCH)		0x3b30	32	hard	yes	yes
+ * 6 Series (PCH)		0x1c22	32	hard	yes	yes
+ * Patsburg (PCH)		0x1d22	32	hard	yes	yes
+ * Patsburg (PCH) IDF		0x1d70	32	hard	yes	yes
+ * Patsburg (PCH) IDF		0x1d71	32	hard	yes	yes
+ * Patsburg (PCH) IDF		0x1d72	32	hard	yes	yes
+ * DH89xxCC (PCH)		0x2330	32	hard	yes	yes
+ * Panther Point (PCH)		0x1e22	32	hard	yes	yes
+ * Lynx Point (PCH)		0x8c22	32	hard	yes	yes
+ * Lynx Point-LP (PCH)		0x9c22	32	hard	yes	yes
+ * Avoton (SOC)			0x1f3c	32	hard	yes	yes
+ * Wellsburg (PCH)		0x8d22	32	hard	yes	yes
+ * Wellsburg (PCH) MS		0x8d7d	32	hard	yes	yes
+ * Wellsburg (PCH) MS		0x8d7e	32	hard	yes	yes
+ * Wellsburg (PCH) MS		0x8d7f	32	hard	yes	yes
+ * Coleto Creek (PCH)		0x23b0	32	hard	yes	yes
+ * Wildcat Point (PCH)		0x8ca2	32	hard	yes	yes
+ * Wildcat Point-LP (PCH)	0x9ca2	32	hard	yes	yes
+ * BayTrail (SOC)		0x0f12	32	hard	yes	yes
+ * Braswell (SOC)		0x2292	32	hard	yes	yes
+ * Sunrise Point-H (PCH)	0xa123  32	hard	yes	yes
+ * Sunrise Point-LP (PCH)	0x9d23	32	hard	yes	yes
+ * DNV (SOC)			0x19df	32	hard	yes	yes
+ * Emmitsburg (PCH)		0x1bc9	32	hard	yes	yes
+ * Broxton (SOC)		0x5ad4	32	hard	yes	yes
+ * Lewisburg (PCH)		0xa1a3	32	hard	yes	yes
+ * Lewisburg Supersku (PCH)	0xa223	32	hard	yes	yes
+ * Kaby Lake PCH-H (PCH)	0xa2a3	32	hard	yes	yes
+ * Gemini Lake (SOC)		0x31d4	32	hard	yes	yes
+ * Cannon Lake-H (PCH)		0xa323	32	hard	yes	yes
+ * Cannon Lake-LP (PCH)		0x9da3	32	hard	yes	yes
+ * Cedar Fork (PCH)		0x18df	32	hard	yes	yes
+ * Ice Lake-LP (PCH)		0x34a3	32	hard	yes	yes
+ * Ice Lake-N (PCH)		0x38a3	32	hard	yes	yes
+ * Comet Lake (PCH)		0x02a3	32	hard	yes	yes
+ * Comet Lake-H (PCH)		0x06a3	32	hard	yes	yes
+ * Elkhart Lake (PCH)		0x4b23	32	hard	yes	yes
+ * Tiger Lake-LP (PCH)		0xa0a3	32	hard	yes	yes
+ * Tiger Lake-H (PCH)		0x43a3	32	hard	yes	yes
+ * Jasper Lake (SOC)		0x4da3	32	hard	yes	yes
+ * Comet Lake-V (PCH)		0xa3a3	32	hard	yes	yes
+ * Alder Lake-S (PCH)		0x7aa3	32	hard	yes	yes
+ * Alder Lake-P (PCH)		0x51a3	32	hard	yes	yes
+ * Alder Lake-M (PCH)		0x54a3	32	hard	yes	yes
+ * Raptor Lake-S (PCH)		0x7a23	32	hard	yes	yes
  *
  * Features supported by this driver:
  * Software PEC				no
@@ -168,7 +168,7 @@
 #define I801_WORD_DATA		0x0C
 #define I801_PROC_CALL		0x10
 #define I801_BLOCK_DATA		0x14
-#define I801_I2C_BLOCK_DATA	0x18	/* ICH5 and later */
+#define I801_I2C_BLOCK_DATA	0x18
 #define I801_BLOCK_PROC_CALL	0x1C
 
 /* I801 Host Control register bits */
@@ -973,11 +973,8 @@ static const struct i2c_algorithm smbus_algorithm = {
 	.functionality	= i801_func,
 };
 
-#define FEATURES_ICH5	(FEATURE_BLOCK_PROC | FEATURE_I2C_BLOCK_READ	| \
-			 FEATURE_IRQ | FEATURE_SMBUS_PEC		| \
-			 FEATURE_BLOCK_BUFFER | FEATURE_HOST_NOTIFY)
-#define FEATURES_ICH4	(FEATURE_SMBUS_PEC | FEATURE_BLOCK_BUFFER | \
-			 FEATURE_HOST_NOTIFY)
+#define FEATURES_ICH4	(FEATURE_SMBUS_PEC | FEATURE_BLOCK_BUFFER | FEATURE_HOST_NOTIFY)
+#define FEATURES_ICH5	(FEATURES_ICH4 | FEATURE_BLOCK_PROC)
 
 static const struct pci_device_id i801_ids[] = {
 	{ PCI_DEVICE_DATA(INTEL, 82801AA_3,		0)				 },
@@ -1665,7 +1662,8 @@ static int i801_probe(struct pci_dev *dev, const struct pci_device_id *id)
 	mutex_init(&priv->acpi_lock);
 
 	priv->pci_dev = dev;
-	priv->features = id->driver_data;
+	priv->features = FEATURE_IRQ | FEATURE_I2C_BLOCK_READ;
+	priv->features |= id->driver_data;
 
 	/* Disable features on user request */
 	for (i = 0; i < ARRAY_SIZE(i801_feature_names); i++) {
Signature · 1 line
2.35.3
#11 lore raw

Re: [PATCH 4/8] i2c: i801: enable FEATURE_IRQ and FEATURE_I2C_BLOCK_READ on all chip versions

Hi Heiner,

Quoted 3 lines · Heiner Kallweit

According to the datasheets interrupt mode and i2c block read are supported on all chip versions. Therefore set both feature flags for all chip versions.

While the datasheets do match your claims (I checked the 82801CAM aka ICH3-M datasheet), I have a hard time believing we would have made the feature device-dependent without a good reason (and I have vague memories that there was a problem, although I can't find any proof of that).

So I'll try to resurrect my old ICH3-M-based laptop and test these changes on it. If I manage to get a Linux distribution to install on that 20-year-old system...

Quoted 3 lines

Note: Don't remove the two feature flags as such (at least for now), so that in case of a problem users can use the disable_features module parameter to disable a problematic feature.

Agreed.

Signature · 2 lines
Jean Delvare
SUSE L3 Support
#12 lore raw

Re: [PATCH 4/8] i2c: i801: enable FEATURE_IRQ and FEATURE_I2C_BLOCK_READ on all chip versions

Hi Heiner,

I was able to resurrect my Sony Vaio GR214EP laptop to test this patch on an Intel 82801CAM (ICH3-M) chipset. And as I suspected, it does not work. I get the following error in the kernel log:

[13563.411101] i801_smbus 0000:00:1f.3: SMBus using PCI interrupt
[13565.434712] irq 9: nobody cared (try booting with the "irqpoll" option)
[13565.434728] Pid: 321, comm: udevd Tainted: G           O 3.4.63-2.44-default #1
[13565.434734] Call Trace:
[13565.434773]  [<c0205349>] try_stack_unwind+0x199/0x1b0
[13565.434793]  [<c02041c7>] dump_trace+0x47/0xf0
[13565.434808]  [<c02053ab>] show_trace_log_lvl+0x4b/0x60
[13565.434820]  [<c02053d8>] show_trace+0x18/0x20
[13565.434837]  [<c0682e81>] dump_stack+0x6d/0x72
[13565.434855]  [<c02adad1>] __report_bad_irq+0x21/0xc0
[13565.434870]  [<c02adef1>] note_interrupt+0x181/0x1d0
[13565.434887]  [<c02abe9e>] handle_irq_event_percpu+0x9e/0x1d0
[13565.434901]  [<c02abffe>] handle_irq_event+0x2e/0x50
[13565.434915]  [<c02ae3e6>] handle_level_irq+0x56/0x90
[13565.434928]  [<c0204168>] handle_irq+0x88/0xa0
[13565.434939] handlers:
[13565.434949] [<c04bbf6c>] acpi_irq
[13565.435027] [<e0baad00>] usb_hcd_irq
[13565.435054] [<e0baad00>] usb_hcd_irq
[13565.435079] [<e0baad00>] usb_hcd_irq
[13565.435193] [<e0d6fb80>] radeon_driver_irq_handler_kms
[13565.435206] [<e0af5c60>] yenta_interrupt
[13565.435214] [<e0af5c60>] yenta_interrupt
[13565.435227] [<e0b514c0>] irq_handler
[13565.435238] [<e0aeec90>] snd_intel8x0m_interrupt
[13565.435251] [<e0c514a0>] snd_intel8x0_interrupt
[13565.435264] [<e0ab4060>] e100_intr
[13565.435278] [<e09465d0>] i801_isr
[13565.435283] Disabling IRQ #9
[13565.437114] i801_smbus 0000:00:1f.3: Transaction timeout
[13565.437125] i801_smbus 0000:00:1f.3: Terminating the current operation
[13565.439189] i801_smbus 0000:00:1f.3: Failed terminating the transaction

If it matters, and as seen in the list of handlers above, IRQ9 is heavily shared on this laptop (ACPI, USB, graphics, PCMCIA, sound, network and SMBus).

Signature · 2 lines
Jean Delvare
SUSE L3 Support
#13 lore raw

Re: [PATCH 4/8] i2c: i801: enable FEATURE_IRQ and FEATURE_I2C_BLOCK_READ on all chip versions

Hi Heiner,

Quoted 6 lines · Jean Delvare

I was able to resurrect my Sony Vaio GR214EP laptop to test this patch on an Intel 82801CAM (ICH3-M) chipset. And as I suspected, it does not work. I get the following error in the kernel log:

[13563.411101] i801_smbus 0000:00:1f.3: SMBus using PCI interrupt
[13565.434712] irq 9: nobody cared (try booting with the "irqpoll" option)

And now it's all coming back to me. The reason why we did not enable interrupts on chipsets older than the ICH5 is because they lack bit INTS in PCI register PCISTS. When the IRQ line is shared, we don't know whether the interrupt was caused by our device or by another device. Specifically, the following piece of code fails (it returns IRQ_NONE unconditionally):

static irqreturn_t i801_isr(int irq, void *dev_id)

{
	(...)
	/* Confirm this is our interrupt */
	pci_read_config_word(priv->pci_dev, PCI_STATUS, &pcists);
	if (!(pcists & PCI_STATUS_INTERRUPT))
		return IRQ_NONE;

I tried replacing the code above by a check on SMBHSTSTS instead:

	status = inb_p(SMBHSTSTS(priv));
	if (!(status & STATUS_FLAGS))
		return IRQ_NONE;

It seems to work, however my testing is limited and I don't know how reliable that would be if the other devices sharing the interrupt line were used heavily at the same time (the laptop is idling in text mode at the moment so the fact that the interrupt line is heavily shared does not really get exercised).

Then I loaded the driver with interrupts disabled (disable_features=0x10) to see if I2C block read was working. It is NOT working, as can be seen from the following dumps from the memory SPD EEPROM:

linux-w6i1:/home/jdelvare/src/i2c-i801 # i2cdump -y 4 0x50 b

     0  1  2  3  4  5  6  7  8  9  a  b  c  d  e  f    0123456789abcdef
00: 80 08 04 0d 09 02 40 00 01 75 54 00 82 10 00 01    ??????@.?uT.??.?
10: 8f 04 06 01 01 00 0e a0 60 00 00 14 0f 14 2c 20    ?????.??`..???, 
20: 15 08 15 08 00 00 00 00 00 00 00 00 00 00 00 00    ????............
30: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 12 b9    ..............??
40: 7f da 00 00 00 00 00 00 53 53 50 31 33 33 2d 30    ??......SSP133-0
50: 36 34 33 32 33 4a 00 00 00 00 00 00 00 01 02 00    64323J.......??.
60: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00    ................
70: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 64 cd    ..............d?
80: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff    ................
90: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff    ................
a0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff    ................
b0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff    ................
c0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff    ................
d0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff    ................
e0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff    ................
f0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff    ................

linux-w6i1:/home/jdelvare/src/i2c-i801 # i2cdump -y 4 0x50 i

     0  1  2  3  4  5  6  7  8  9  a  b  c  d  e  f    0123456789abcdef
00: 08 04 0d 09 02 40 00 01 75 54 00 82 10 00 01 8f    ?????@.?uT.??.??
10: 04 06 01 01 00 0e a0 60 00 00 14 0f 14 2c 20 15    ????.??`..???, ?
20: 08 04 0d 09 02 40 00 01 75 54 00 82 10 00 01 8f    ?????@.?uT.??.??
30: 04 06 01 01 00 0e a0 60 00 00 14 0f 14 2c 20 15    ????.??`..???, ?
40: 08 04 0d 09 02 40 00 01 75 54 00 82 10 00 01 8f    ?????@.?uT.??.??
50: 04 06 01 01 00 0e a0 60 00 00 14 0f 14 2c 20 15    ????.??`..???, ?
60: 08 04 0d 09 02 40 00 01 75 54 00 82 10 00 01 8f    ?????@.?uT.??.??
70: 04 06 01 01 00 0e a0 60 00 00 14 0f 14 2c 20 15    ????.??`..???, ?
80: 08 04 0d 09 02 40 00 01 75 54 00 82 10 00 01 8f    ?????@.?uT.??.??
90: 04 06 01 01 00 0e a0 60 00 00 14 0f 14 2c 20 15    ????.??`..???, ?
a0: 08 04 0d 09 02 40 00 01 75 54 00 82 10 00 01 8f    ?????@.?uT.??.??
b0: 04 06 01 01 00 0e a0 60 00 00 14 0f 14 2c 20 15    ????.??`..???, ?
c0: 08 04 0d 09 02 40 00 01 75 54 00 82 10 00 01 8f    ?????@.?uT.??.??
d0: 04 06 01 01 00 0e a0 60 00 00 14 0f 14 2c 20 15    ????.??`..???, ?
e0: 08 04 0d 09 02 40 00 01 75 54 00 82 10 00 01 8f    ?????@.?uT.??.??
f0: 04 06 01 01 00 0e a0 60 00 00 14 0f 14 2c 20 15    ????.??`..???, ?

As you can see, the requested offset of the I2C block read (0x00, then 0x20, then 0x40 etc.) is being ignored, and instead every I2C block read starts at EEPROM offset 0x01.

If you compare the datasheets of the ICH5 (Intel doc 252516-001) and ICH3-M (Intel doc 290716-001), section "I2C Read", you will see that the description of the command is different. The format described for the ICH5 (table 114 in the datasheet) does match what the Linux i2c stack calls an I2C block read, while the format described for the ICH3-M (table 5-87 in the datasheet) does NOT. Apparently the original implementation was aimed at specific devices using 10-bit I2C addressing. As no other SMBus host device implemented that, we did not even allocate an SMBus command constant to it (and the fact that Intel changed the format in later hardware iterations proves that this was the right thing to do).

Looking at the ICH3-M implementation of the "I2C Block Read" transfer, I feel very lucky that I did not trash my memory SPD EEPROM while running the command. Because the format really starts with a WRITE of 2 bytes to the EEPROM before reading from it.

So at least the I2C block read part of the patch is never going to work. The interrupt part could work if we change the test as described above, however I would question the relevance of making that change considering that it is no longer the obvious clean-up you originally proposed, and would then impact recent hardware too. I would hate to introduce a regression for the sole purpose of enabling interrupts on 20-year old hardware which I doubt many people are still using.

I am available to perform any test you want me to on this old laptop. As long as it runs...

Signature · 2 lines
Jean Delvare
SUSE L3 Support
#14 lore raw

Re: [PATCH 4/8] i2c: i801: enable FEATURE_IRQ and FEATURE_I2C_BLOCK_READ on all chip versions

Quoted 109 lines · Jean Delvare

Hi Heiner,

Quoted 6 lines · Jean Delvare

I was able to resurrect my Sony Vaio GR214EP laptop to test this patch on an Intel 82801CAM (ICH3-M) chipset. And as I suspected, it does not work. I get the following error in the kernel log:

[13563.411101] i801_smbus 0000:00:1f.3: SMBus using PCI interrupt
[13565.434712] irq 9: nobody cared (try booting with the "irqpoll" option)

And now it's all coming back to me. The reason why we did not enable interrupts on chipsets older than the ICH5 is because they lack bit INTS in PCI register PCISTS. When the IRQ line is shared, we don't know whether the interrupt was caused by our device or by another device. Specifically, the following piece of code fails (it returns IRQ_NONE unconditionally):

static irqreturn_t i801_isr(int irq, void *dev_id)

{
	(...)
	/* Confirm this is our interrupt */
	pci_read_config_word(priv->pci_dev, PCI_STATUS, &pcists);
	if (!(pcists & PCI_STATUS_INTERRUPT))
		return IRQ_NONE;

I tried replacing the code above by a check on SMBHSTSTS instead:

	status = inb_p(SMBHSTSTS(priv));
	if (!(status & STATUS_FLAGS))
		return IRQ_NONE;

It seems to work, however my testing is limited and I don't know how reliable that would be if the other devices sharing the interrupt line were used heavily at the same time (the laptop is idling in text mode at the moment so the fact that the interrupt line is heavily shared does not really get exercised).

Then I loaded the driver with interrupts disabled (disable_features=0x10) to see if I2C block read was working. It is NOT working, as can be seen from the following dumps from the memory SPD EEPROM:

linux-w6i1:/home/jdelvare/src/i2c-i801 # i2cdump -y 4 0x50 b

     0  1  2  3  4  5  6  7  8  9  a  b  c  d  e  f    0123456789abcdef
00: 80 08 04 0d 09 02 40 00 01 75 54 00 82 10 00 01    ??????@.?uT.??.?
10: 8f 04 06 01 01 00 0e a0 60 00 00 14 0f 14 2c 20    ?????.??`..???, 
20: 15 08 15 08 00 00 00 00 00 00 00 00 00 00 00 00    ????............
30: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 12 b9    ..............??
40: 7f da 00 00 00 00 00 00 53 53 50 31 33 33 2d 30    ??......SSP133-0
50: 36 34 33 32 33 4a 00 00 00 00 00 00 00 01 02 00    64323J.......??.
60: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00    ................
70: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 64 cd    ..............d?
80: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff    ................
90: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff    ................
a0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff    ................
b0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff    ................
c0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff    ................
d0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff    ................
e0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff    ................
f0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff    ................

linux-w6i1:/home/jdelvare/src/i2c-i801 # i2cdump -y 4 0x50 i

     0  1  2  3  4  5  6  7  8  9  a  b  c  d  e  f    0123456789abcdef
00: 08 04 0d 09 02 40 00 01 75 54 00 82 10 00 01 8f    ?????@.?uT.??.??
10: 04 06 01 01 00 0e a0 60 00 00 14 0f 14 2c 20 15    ????.??`..???, ?
20: 08 04 0d 09 02 40 00 01 75 54 00 82 10 00 01 8f    ?????@.?uT.??.??
30: 04 06 01 01 00 0e a0 60 00 00 14 0f 14 2c 20 15    ????.??`..???, ?
40: 08 04 0d 09 02 40 00 01 75 54 00 82 10 00 01 8f    ?????@.?uT.??.??
50: 04 06 01 01 00 0e a0 60 00 00 14 0f 14 2c 20 15    ????.??`..???, ?
60: 08 04 0d 09 02 40 00 01 75 54 00 82 10 00 01 8f    ?????@.?uT.??.??
70: 04 06 01 01 00 0e a0 60 00 00 14 0f 14 2c 20 15    ????.??`..???, ?
80: 08 04 0d 09 02 40 00 01 75 54 00 82 10 00 01 8f    ?????@.?uT.??.??
90: 04 06 01 01 00 0e a0 60 00 00 14 0f 14 2c 20 15    ????.??`..???, ?
a0: 08 04 0d 09 02 40 00 01 75 54 00 82 10 00 01 8f    ?????@.?uT.??.??
b0: 04 06 01 01 00 0e a0 60 00 00 14 0f 14 2c 20 15    ????.??`..???, ?
c0: 08 04 0d 09 02 40 00 01 75 54 00 82 10 00 01 8f    ?????@.?uT.??.??
d0: 04 06 01 01 00 0e a0 60 00 00 14 0f 14 2c 20 15    ????.??`..???, ?
e0: 08 04 0d 09 02 40 00 01 75 54 00 82 10 00 01 8f    ?????@.?uT.??.??
f0: 04 06 01 01 00 0e a0 60 00 00 14 0f 14 2c 20 15    ????.??`..???, ?

As you can see, the requested offset of the I2C block read (0x00, then 0x20, then 0x40 etc.) is being ignored, and instead every I2C block read starts at EEPROM offset 0x01.

If you compare the datasheets of the ICH5 (Intel doc 252516-001) and ICH3-M (Intel doc 290716-001), section "I2C Read", you will see that the description of the command is different. The format described for the ICH5 (table 114 in the datasheet) does match what the Linux i2c stack calls an I2C block read, while the format described for the ICH3-M (table 5-87 in the datasheet) does NOT. Apparently the original implementation was aimed at specific devices using 10-bit I2C addressing. As no other SMBus host device implemented that, we did not even allocate an SMBus command constant to it (and the fact that Intel changed the format in later hardware iterations proves that this was the right thing to do).

Looking at the ICH3-M implementation of the "I2C Block Read" transfer, I feel very lucky that I did not trash my memory SPD EEPROM while running the command. Because the format really starts with a WRITE of 2 bytes to the EEPROM before reading from it.

So at least the I2C block read part of the patch is never going to work. The interrupt part could work if we change the test as described above, however I would question the relevance of making that change considering that it is no longer the obvious clean-up you originally proposed, and would then impact recent hardware too. I would hate to introduce a regression for the sole purpose of enabling interrupts on 20-year old hardware which I doubt many people are still using.

Quoted 3 lines

I am available to perform any test you want me to on this old laptop. As long as it runs...

Thanks for testing! So I'll drop this patch from the series.

#15 lore raw

PATCH 5/8 i2c: i801: add helper i801_set_hstadd()

In reply to Heiner Kallweit

Factor out setting SMBHSTADD to a helper. The current code makes the assumption that constant I2C_SMBUS_READ has bit 0 set, that's not ideal. Therefore let the new helper explicitly check for I2C_SMBUS_READ.

---
 drivers/i2c/busses/i2c-i801.c | 41 ++++++++++++++++++-----------------
 1 file changed, 21 insertions(+), 20 deletions(-)
drivers/i2c/busses/i2c-i801.c +21 -20
index a9737f14d..bf77f8640 100644
@@ -771,6 +771,14 @@ static int i801_block_transaction(struct i801_priv *priv, union i2c_smbus_data *
 	return result;
 }
 
+static void i801_set_hstadd(struct i801_priv *priv, u8 addr, char read_write)
+{
+	addr <<= 1;
+	if (read_write == I2C_SMBUS_READ)
+		addr |= 0x01;
+	outb_p(addr, SMBHSTADD(priv));
+}
+
 /* Return negative errno on error. */
 static s32 i801_access(struct i2c_adapter *adap, u16 addr,
 		       unsigned short flags, char read_write, u8 command,
@@ -795,28 +803,24 @@ static s32 i801_access(struct i2c_adapter *adap, u16 addr,
 
 	switch (size) {
 	case I2C_SMBUS_QUICK:
-		outb_p(((addr & 0x7f) << 1) | (read_write & 0x01),
-		       SMBHSTADD(priv));
+		i801_set_hstadd(priv, addr, read_write);
 		xact = I801_QUICK;
 		break;
 	case I2C_SMBUS_BYTE:
-		outb_p(((addr & 0x7f) << 1) | (read_write & 0x01),
-		       SMBHSTADD(priv));
+		i801_set_hstadd(priv, addr, read_write);
 		if (read_write == I2C_SMBUS_WRITE)
 			outb_p(command, SMBHSTCMD(priv));
 		xact = I801_BYTE;
 		break;
 	case I2C_SMBUS_BYTE_DATA:
-		outb_p(((addr & 0x7f) << 1) | (read_write & 0x01),
-		       SMBHSTADD(priv));
+		i801_set_hstadd(priv, addr, read_write);
 		outb_p(command, SMBHSTCMD(priv));
 		if (read_write == I2C_SMBUS_WRITE)
 			outb_p(data->byte, SMBHSTDAT0(priv));
 		xact = I801_BYTE_DATA;
 		break;
 	case I2C_SMBUS_WORD_DATA:
-		outb_p(((addr & 0x7f) << 1) | (read_write & 0x01),
-		       SMBHSTADD(priv));
+		i801_set_hstadd(priv, addr, read_write);
 		outb_p(command, SMBHSTCMD(priv));
 		if (read_write == I2C_SMBUS_WRITE) {
 			outb_p(data->word & 0xff, SMBHSTDAT0(priv));
@@ -825,7 +829,7 @@ static s32 i801_access(struct i2c_adapter *adap, u16 addr,
 		xact = I801_WORD_DATA;
 		break;
 	case I2C_SMBUS_PROC_CALL:
-		outb_p((addr & 0x7f) << 1, SMBHSTADD(priv));
+		i801_set_hstadd(priv, addr, I2C_SMBUS_WRITE);
 		outb_p(command, SMBHSTCMD(priv));
 		outb_p(data->word & 0xff, SMBHSTDAT0(priv));
 		outb_p((data->word & 0xff00) >> 8, SMBHSTDAT1(priv));
@@ -833,8 +837,7 @@ static s32 i801_access(struct i2c_adapter *adap, u16 addr,
 		read_write = I2C_SMBUS_READ;
 		break;
 	case I2C_SMBUS_BLOCK_DATA:
-		outb_p(((addr & 0x7f) << 1) | (read_write & 0x01),
-		       SMBHSTADD(priv));
+		i801_set_hstadd(priv, addr, read_write);
 		outb_p(command, SMBHSTCMD(priv));
 		block = 1;
 		break;
@@ -845,10 +848,11 @@ static s32 i801_access(struct i2c_adapter *adap, u16 addr,
 		 * However if SPD Write Disable is set (Lynx Point and later),
 		 * the read will fail if we don't set the R/#W bit.
 		 */
-		outb_p(((addr & 0x7f) << 1) |
-		       ((priv->original_hstcfg & SMBHSTCFG_SPD_WD) ?
-			(read_write & 0x01) : 0),
-		       SMBHSTADD(priv));
+		if (priv->original_hstcfg & SMBHSTCFG_SPD_WD)
+			i801_set_hstadd(priv, addr, read_write);
+		else
+			i801_set_hstadd(priv, addr, I2C_SMBUS_WRITE);
+
 		if (read_write == I2C_SMBUS_READ) {
 			/* NB: page 240 of ICH5 datasheet also shows
 			 * that DATA1 is the cmd field when reading */
@@ -858,11 +862,8 @@ static s32 i801_access(struct i2c_adapter *adap, u16 addr,
 		block = 1;
 		break;
 	case I2C_SMBUS_BLOCK_PROC_CALL:
-		/*
-		 * Bit 0 of the slave address register always indicate a write
-		 * command.
-		 */
-		outb_p((addr & 0x7f) << 1, SMBHSTADD(priv));
+		/* Needs to be flagged as write transaction */
+		i801_set_hstadd(priv, addr, I2C_SMBUS_WRITE);
 		outb_p(command, SMBHSTCMD(priv));
 		block = 1;
 		break;
Signature · 1 line
2.35.3
#16 lore raw

Re: [PATCH 5/8] i2c: i801: add helper i801_set_hstadd()

Hi Heiner,

Quoted 2 lines · Heiner Kallweit

Factor out setting SMBHSTADD to a helper. The current code makes the assumption that constant I2C_SMBUS_READ has bit 0 set, that's not ideal.

This isn't an "assumption". The values of I2C_SMBUS_WRITE and I2C_SMBUS_READ were chosen to match the bit position and values in the I2C protocol. Maybe it should have been made clearer by defining them as hexadecimal values instead of decimal values. Nevertheless, I find it unfortunate to not use this fact to optimize the code, see below.

Quoted 12 lines

Therefore let the new helper explicitly check for I2C_SMBUS_READ.

---
 drivers/i2c/busses/i2c-i801.c | 41 ++++++++++++++++++-----------------
 1 file changed, 21 insertions(+), 20 deletions(-)
drivers/i2c/busses/i2c-i801.c +0 -0
index a9737f14d..bf77f8640 100644

drivers/i2c/busses/i2c-i801.c static int i801_block_transaction(struct i801_priv *priv, union i2c_smbus_data * · line 779

Quoted 9 lines
drivers/i2c/busses/i2c-i801.c +6 -0
@@ -771,6 +771,14 @@ static int i801_block_transaction(struct i801_priv *priv, union i2c_smbus_data *
 	return result;
 }
 
+static void i801_set_hstadd(struct i801_priv *priv, u8 addr, char read_write)
+{
+	addr <<= 1;
+	if (read_write == I2C_SMBUS_READ)
+		addr |= 0x01;
+	outb_p(addr, SMBHSTADD(priv));

This can be written:

	outb_p((addr << 1) | read_write, SMBHSTADD(priv));

Net result -48 bytes of (x86_64) binary code. That's basically what the original code was doing, minus the useless masking.

Quoted 59 lines
drivers/i2c/busses/i2c-i801.c +8 -11
+}
+
 /* Return negative errno on error. */
 static s32 i801_access(struct i2c_adapter *adap, u16 addr,
 		       unsigned short flags, char read_write, u8 command,
@@ -795,28 +803,24 @@ static s32 i801_access(struct i2c_adapter *adap, u16 addr,
 
 	switch (size) {
 	case I2C_SMBUS_QUICK:
-		outb_p(((addr & 0x7f) << 1) | (read_write & 0x01),
-		       SMBHSTADD(priv));
+		i801_set_hstadd(priv, addr, read_write);
 		xact = I801_QUICK;
 		break;
 	case I2C_SMBUS_BYTE:
-		outb_p(((addr & 0x7f) << 1) | (read_write & 0x01),
-		       SMBHSTADD(priv));
+		i801_set_hstadd(priv, addr, read_write);
 		if (read_write == I2C_SMBUS_WRITE)
 			outb_p(command, SMBHSTCMD(priv));
 		xact = I801_BYTE;
 		break;
 	case I2C_SMBUS_BYTE_DATA:
-		outb_p(((addr & 0x7f) << 1) | (read_write & 0x01),
-		       SMBHSTADD(priv));
+		i801_set_hstadd(priv, addr, read_write);
 		outb_p(command, SMBHSTCMD(priv));
 		if (read_write == I2C_SMBUS_WRITE)
 			outb_p(data->byte, SMBHSTDAT0(priv));
 		xact = I801_BYTE_DATA;
 		break;
 	case I2C_SMBUS_WORD_DATA:
-		outb_p(((addr & 0x7f) << 1) | (read_write & 0x01),
-		       SMBHSTADD(priv));
+		i801_set_hstadd(priv, addr, read_write);
 		outb_p(command, SMBHSTCMD(priv));
 		if (read_write == I2C_SMBUS_WRITE) {
 			outb_p(data->word & 0xff, SMBHSTDAT0(priv));
@@ -825,7 +829,7 @@ static s32 i801_access(struct i2c_adapter *adap, u16 addr,
 		xact = I801_WORD_DATA;
 		break;
 	case I2C_SMBUS_PROC_CALL:
-		outb_p((addr & 0x7f) << 1, SMBHSTADD(priv));
+		i801_set_hstadd(priv, addr, I2C_SMBUS_WRITE);
 		outb_p(command, SMBHSTCMD(priv));
 		outb_p(data->word & 0xff, SMBHSTDAT0(priv));
 		outb_p((data->word & 0xff00) >> 8, SMBHSTDAT1(priv));
@@ -833,8 +837,7 @@ static s32 i801_access(struct i2c_adapter *adap, u16 addr,
 		read_write = I2C_SMBUS_READ;
 		break;
 	case I2C_SMBUS_BLOCK_DATA:
-		outb_p(((addr & 0x7f) << 1) | (read_write & 0x01),
-		       SMBHSTADD(priv));
+		i801_set_hstadd(priv, addr, read_write);
 		outb_p(command, SMBHSTCMD(priv));
 		block = 1;
 		break;
@@ -845,10 +848,11 @@ static s32 i801_access(struct i2c_adapter *adap, u16 addr,
 		 * However if SPD Write Disable is set (Lynx Point and later),

drivers/i2c/busses/i2c-i801.c static s32 i801_access(struct i2c_adapter *adap, u16 addr, · line 854

Quoted 10 lines
drivers/i2c/busses/i2c-i801.c +4 -4
 		 * the read will fail if we don't set the R/#W bit.
 		 */
-		outb_p(((addr & 0x7f) << 1) |
-		       ((priv->original_hstcfg & SMBHSTCFG_SPD_WD) ?
-			(read_write & 0x01) : 0),
-		       SMBHSTADD(priv));
+		if (priv->original_hstcfg & SMBHSTCFG_SPD_WD)
+			i801_set_hstadd(priv, addr, read_write);
+		else
+			i801_set_hstadd(priv, addr, I2C_SMBUS_WRITE);

Preserving the use of the ternary operator makes the generated binary smaller once again:

		i801_set_hstadd(priv, addr,
				(priv->original_hstcfg & SMBHSTCFG_SPD_WD) ?
				read_write : I2C_SMBUS_WRITE);

Net result -11 bytes of (x86_64) binary code.

Quoted 18 lines
drivers/i2c/busses/i2c-i801.c +3 -5
+
 		if (read_write == I2C_SMBUS_READ) {
 			/* NB: page 240 of ICH5 datasheet also shows
 			 * that DATA1 is the cmd field when reading */
@@ -858,11 +862,8 @@ static s32 i801_access(struct i2c_adapter *adap, u16 addr,
 		block = 1;
 		break;
 	case I2C_SMBUS_BLOCK_PROC_CALL:
-		/*
-		 * Bit 0 of the slave address register always indicate a write
-		 * command.
-		 */
-		outb_p((addr & 0x7f) << 1, SMBHSTADD(priv));
+		/* Needs to be flagged as write transaction */
+		i801_set_hstadd(priv, addr, I2C_SMBUS_WRITE);
 		outb_p(command, SMBHSTCMD(priv));
 		block = 1;
 		break;
Signature · 2 lines
Jean Delvare
SUSE L3 Support
#17 lore raw

Re: [PATCH 5/8] i2c: i801: add helper i801_set_hstadd()

Quoted 36 lines · Jean Delvare

Hi Heiner,

Quoted 2 lines · Heiner Kallweit

Factor out setting SMBHSTADD to a helper. The current code makes the assumption that constant I2C_SMBUS_READ has bit 0 set, that's not ideal.

This isn't an "assumption". The values of I2C_SMBUS_WRITE and I2C_SMBUS_READ were chosen to match the bit position and values in the I2C protocol. Maybe it should have been made clearer by defining them as hexadecimal values instead of decimal values. Nevertheless, I find it unfortunate to not use this fact to optimize the code, see below.

Quoted 12 lines

Therefore let the new helper explicitly check for I2C_SMBUS_READ.

---
 drivers/i2c/busses/i2c-i801.c | 41 ++++++++++++++++++-----------------
 1 file changed, 21 insertions(+), 20 deletions(-)
drivers/i2c/busses/i2c-i801.c +0 -0
index a9737f14d..bf77f8640 100644

drivers/i2c/busses/i2c-i801.c static int i801_block_transaction(struct i801_priv *priv, union i2c_smbus_data * · line 779

Quoted 9 lines
drivers/i2c/busses/i2c-i801.c +6 -0
@@ -771,6 +771,14 @@ static int i801_block_transaction(struct i801_priv *priv, union i2c_smbus_data *
 	return result;
 }
 
+static void i801_set_hstadd(struct i801_priv *priv, u8 addr, char read_write)
+{
+	addr <<= 1;
+	if (read_write == I2C_SMBUS_READ)
+		addr |= 0x01;
+	outb_p(addr, SMBHSTADD(priv));

This can be written:

Quoted 5 lines
	outb_p((addr << 1) | read_write, SMBHSTADD(priv));

Net result -48 bytes of (x86_64) binary code. That's basically what the original code was doing, minus the useless masking.

OK

Quoted 73 lines
Quoted 59 lines
drivers/i2c/busses/i2c-i801.c +8 -11
+}
+
 /* Return negative errno on error. */
 static s32 i801_access(struct i2c_adapter *adap, u16 addr,
 		       unsigned short flags, char read_write, u8 command,
@@ -795,28 +803,24 @@ static s32 i801_access(struct i2c_adapter *adap, u16 addr,
 
 	switch (size) {
 	case I2C_SMBUS_QUICK:
-		outb_p(((addr & 0x7f) << 1) | (read_write & 0x01),
-		       SMBHSTADD(priv));
+		i801_set_hstadd(priv, addr, read_write);
 		xact = I801_QUICK;
 		break;
 	case I2C_SMBUS_BYTE:
-		outb_p(((addr & 0x7f) << 1) | (read_write & 0x01),
-		       SMBHSTADD(priv));
+		i801_set_hstadd(priv, addr, read_write);
 		if (read_write == I2C_SMBUS_WRITE)
 			outb_p(command, SMBHSTCMD(priv));
 		xact = I801_BYTE;
 		break;
 	case I2C_SMBUS_BYTE_DATA:
-		outb_p(((addr & 0x7f) << 1) | (read_write & 0x01),
-		       SMBHSTADD(priv));
+		i801_set_hstadd(priv, addr, read_write);
 		outb_p(command, SMBHSTCMD(priv));
 		if (read_write == I2C_SMBUS_WRITE)
 			outb_p(data->byte, SMBHSTDAT0(priv));
 		xact = I801_BYTE_DATA;
 		break;
 	case I2C_SMBUS_WORD_DATA:
-		outb_p(((addr & 0x7f) << 1) | (read_write & 0x01),
-		       SMBHSTADD(priv));
+		i801_set_hstadd(priv, addr, read_write);
 		outb_p(command, SMBHSTCMD(priv));
 		if (read_write == I2C_SMBUS_WRITE) {
 			outb_p(data->word & 0xff, SMBHSTDAT0(priv));
@@ -825,7 +829,7 @@ static s32 i801_access(struct i2c_adapter *adap, u16 addr,
 		xact = I801_WORD_DATA;
 		break;
 	case I2C_SMBUS_PROC_CALL:
-		outb_p((addr & 0x7f) << 1, SMBHSTADD(priv));
+		i801_set_hstadd(priv, addr, I2C_SMBUS_WRITE);
 		outb_p(command, SMBHSTCMD(priv));
 		outb_p(data->word & 0xff, SMBHSTDAT0(priv));
 		outb_p((data->word & 0xff00) >> 8, SMBHSTDAT1(priv));
@@ -833,8 +837,7 @@ static s32 i801_access(struct i2c_adapter *adap, u16 addr,
 		read_write = I2C_SMBUS_READ;
 		break;
 	case I2C_SMBUS_BLOCK_DATA:
-		outb_p(((addr & 0x7f) << 1) | (read_write & 0x01),
-		       SMBHSTADD(priv));
+		i801_set_hstadd(priv, addr, read_write);
 		outb_p(command, SMBHSTCMD(priv));
 		block = 1;
 		break;
@@ -845,10 +848,11 @@ static s32 i801_access(struct i2c_adapter *adap, u16 addr,
 		 * However if SPD Write Disable is set (Lynx Point and later),

drivers/i2c/busses/i2c-i801.c static s32 i801_access(struct i2c_adapter *adap, u16 addr, · line 854

Quoted 10 lines
drivers/i2c/busses/i2c-i801.c +4 -4
 		 * the read will fail if we don't set the R/#W bit.
 		 */
-		outb_p(((addr & 0x7f) << 1) |
-		       ((priv->original_hstcfg & SMBHSTCFG_SPD_WD) ?
-			(read_write & 0x01) : 0),
-		       SMBHSTADD(priv));
+		if (priv->original_hstcfg & SMBHSTCFG_SPD_WD)
+			i801_set_hstadd(priv, addr, read_write);
+		else
+			i801_set_hstadd(priv, addr, I2C_SMBUS_WRITE);

Preserving the use of the ternary operator makes the generated binary smaller once again:

Quoted 6 lines
		i801_set_hstadd(priv, addr,
				(priv->original_hstcfg & SMBHSTCFG_SPD_WD) ?
				read_write : I2C_SMBUS_WRITE);

Net result -11 bytes of (x86_64) binary code.

OK

Quoted 20 lines
Quoted 18 lines
drivers/i2c/busses/i2c-i801.c +3 -5
+
 		if (read_write == I2C_SMBUS_READ) {
 			/* NB: page 240 of ICH5 datasheet also shows
 			 * that DATA1 is the cmd field when reading */
@@ -858,11 +862,8 @@ static s32 i801_access(struct i2c_adapter *adap, u16 addr,
 		block = 1;
 		break;
 	case I2C_SMBUS_BLOCK_PROC_CALL:
-		/*
-		 * Bit 0 of the slave address register always indicate a write
-		 * command.
-		 */
-		outb_p((addr & 0x7f) << 1, SMBHSTADD(priv));
+		/* Needs to be flagged as write transaction */
+		i801_set_hstadd(priv, addr, I2C_SMBUS_WRITE);
 		outb_p(command, SMBHSTCMD(priv));
 		block = 1;
 		break;
#18 lore raw

PATCH 6/8 i2c: i801: add i801_single_transaction(), complementing i801_block_transaction()

In reply to Heiner Kallweit

This patch factors out non-block pre/post processing to a new function i801_single_transaction(), complementing existing function i801_block_transaction(). This makes i801_access() better readable.

---
 drivers/i2c/busses/i2c-i801.c | 95 +++++++++++++++++++++--------------
 1 file changed, 58 insertions(+), 37 deletions(-)
drivers/i2c/busses/i2c-i801.c +58 -37
index bf77f8640..8c2245f38 100644
@@ -771,6 +771,62 @@ static int i801_block_transaction(struct i801_priv *priv, union i2c_smbus_data *
 	return result;
 }
 
+/* Single transaction function */
+static int i801_single_transaction(struct i801_priv *priv, union i2c_smbus_data *data,
+				   char read_write, int command)
+{
+	int xact, ret;
+
+	switch (command) {
+	case I2C_SMBUS_QUICK:
+		xact = I801_QUICK;
+		break;
+	case I2C_SMBUS_BYTE:
+                xact = I801_BYTE;
+                break;
+	case I2C_SMBUS_BYTE_DATA:
+		if (read_write == I2C_SMBUS_WRITE)
+			outb_p(data->byte, SMBHSTDAT0(priv));
+		xact = I801_BYTE_DATA;
+		break;
+	case I2C_SMBUS_WORD_DATA:
+		if (read_write == I2C_SMBUS_WRITE) {
+			outb_p(data->word & 0xff, SMBHSTDAT0(priv));
+			outb_p((data->word & 0xff00) >> 8, SMBHSTDAT1(priv));
+		}
+		xact = I801_WORD_DATA;
+		break;
+	case I2C_SMBUS_PROC_CALL:
+		outb_p(data->word & 0xff, SMBHSTDAT0(priv));
+		outb_p((data->word & 0xff00) >> 8, SMBHSTDAT1(priv));
+		xact = I801_PROC_CALL;
+		break;
+	default:
+		return -EOPNOTSUPP;
+	}
+
+	ret = i801_transaction(priv, xact);
+
+	if (ret || read_write == I2C_SMBUS_WRITE)
+		return ret;
+
+	switch (command) {
+	case I2C_SMBUS_BYTE:
+	case I2C_SMBUS_BYTE_DATA:
+		data->byte = inb_p(SMBHSTDAT0(priv));
+		break;
+	case I2C_SMBUS_WORD_DATA:
+	case I2C_SMBUS_PROC_CALL:
+		data->word = inb_p(SMBHSTDAT0(priv)) +
+			     (inb_p(SMBHSTDAT1(priv)) << 8);
+		break;
+	default:
+		break;
+	}
+
+	return 0;
+}
+
 static void i801_set_hstadd(struct i801_priv *priv, u8 addr, char read_write)
 {
 	addr <<= 1;
@@ -784,9 +840,7 @@ static s32 i801_access(struct i2c_adapter *adap, u16 addr,
 		       unsigned short flags, char read_write, u8 command,
 		       int size, union i2c_smbus_data *data)
 {
-	int hwpec;
-	int block = 0;
-	int ret, xact;
+	int hwpec, ret, block = 0;
 	struct i801_priv *priv = i2c_get_adapdata(adap);
 
 	mutex_lock(&priv->acpi_lock);
@@ -804,36 +858,23 @@ static s32 i801_access(struct i2c_adapter *adap, u16 addr,
 	switch (size) {
 	case I2C_SMBUS_QUICK:
 		i801_set_hstadd(priv, addr, read_write);
-		xact = I801_QUICK;
 		break;
 	case I2C_SMBUS_BYTE:
 		i801_set_hstadd(priv, addr, read_write);
 		if (read_write == I2C_SMBUS_WRITE)
 			outb_p(command, SMBHSTCMD(priv));
-		xact = I801_BYTE;
 		break;
 	case I2C_SMBUS_BYTE_DATA:
 		i801_set_hstadd(priv, addr, read_write);
 		outb_p(command, SMBHSTCMD(priv));
-		if (read_write == I2C_SMBUS_WRITE)
-			outb_p(data->byte, SMBHSTDAT0(priv));
-		xact = I801_BYTE_DATA;
 		break;
 	case I2C_SMBUS_WORD_DATA:
 		i801_set_hstadd(priv, addr, read_write);
 		outb_p(command, SMBHSTCMD(priv));
-		if (read_write == I2C_SMBUS_WRITE) {
-			outb_p(data->word & 0xff, SMBHSTDAT0(priv));
-			outb_p((data->word & 0xff00) >> 8, SMBHSTDAT1(priv));
-		}
-		xact = I801_WORD_DATA;
 		break;
 	case I2C_SMBUS_PROC_CALL:
 		i801_set_hstadd(priv, addr, I2C_SMBUS_WRITE);
 		outb_p(command, SMBHSTCMD(priv));
-		outb_p(data->word & 0xff, SMBHSTDAT0(priv));
-		outb_p((data->word & 0xff00) >> 8, SMBHSTDAT1(priv));
-		xact = I801_PROC_CALL;
 		read_write = I2C_SMBUS_READ;
 		break;
 	case I2C_SMBUS_BLOCK_DATA:
@@ -883,7 +924,7 @@ static s32 i801_access(struct i2c_adapter *adap, u16 addr,
 	if (block)
 		ret = i801_block_transaction(priv, data, read_write, size);
 	else
-		ret = i801_transaction(priv, xact);
+		ret = i801_single_transaction(priv, data, read_write, size);
 
 	/* Some BIOSes don't like it when PEC is enabled at reboot or resume
 	   time, so we forcibly disable it after every transaction. Turn off
@@ -891,26 +932,6 @@ static s32 i801_access(struct i2c_adapter *adap, u16 addr,
 	if (hwpec || block)
 		outb_p(inb_p(SMBAUXCTL(priv)) &
 		       ~(SMBAUXCTL_CRC | SMBAUXCTL_E32B), SMBAUXCTL(priv));
-
-	if (block)
-		goto out;
-	if (ret)
-		goto out;
-	if ((read_write == I2C_SMBUS_WRITE) || (xact == I801_QUICK))
-		goto out;
-
-	switch (xact) {
-	case I801_BYTE:	/* Result put in SMBHSTDAT0 */
-	case I801_BYTE_DATA:
-		data->byte = inb_p(SMBHSTDAT0(priv));
-		break;
-	case I801_WORD_DATA:
-	case I801_PROC_CALL:
-		data->word = inb_p(SMBHSTDAT0(priv)) +
-			     (inb_p(SMBHSTDAT1(priv)) << 8);
-		break;
-	}
-
 out:
 	/*
 	 * Unlock the SMBus device for use by BIOS/ACPI,
Signature · 1 line
2.35.3
#19 lore raw

Re: [PATCH 6/8] i2c: i801: add i801_single_transaction(), complementing i801_block_transaction()

Hi Heiner,

Quoted 3 lines · Heiner Kallweit

This patch factors out non-block pre/post processing to a new function i801_single_transaction(), complementing existing function i801_block_transaction(). This makes i801_access() better readable.

I like the idea, but I have objections about some implementation details, see below.

Quoted 10 lines
---
 drivers/i2c/busses/i2c-i801.c | 95 +++++++++++++++++++++--------------
 1 file changed, 58 insertions(+), 37 deletions(-)
drivers/i2c/busses/i2c-i801.c +0 -0
index bf77f8640..8c2245f38 100644

drivers/i2c/busses/i2c-i801.c static int i801_block_transaction(struct i801_priv *priv, union i2c_smbus_data * · line 774

Quoted 4 lines
drivers/i2c/busses/i2c-i801.c +1 -0
@@ -771,6 +771,62 @@ static int i801_block_transaction(struct i801_priv *priv, union i2c_smbus_data *
 	return result;
 }
 
+/* Single transaction function */

The term "single transaction" is a bit misleading. Block transactions are also single transactions, in the sense that there's one start condition at the beginning and one stop condition at the end. I'd rather call non-block transactions "single value transactions" or "simple transactions".

drivers/i2c/busses/i2c-i801.c static int i801_block_transaction(struct i801_priv *priv, union i2c_smbus_data * · line 786

Quoted 12 lines
drivers/i2c/busses/i2c-i801.c +12 -0
+static int i801_single_transaction(struct i801_priv *priv, union i2c_smbus_data *data,
+				   char read_write, int command)
+{
+	int xact, ret;
+
+	switch (command) {
+	case I2C_SMBUS_QUICK:
+		xact = I801_QUICK;
+		break;
+	case I2C_SMBUS_BYTE:
+                xact = I801_BYTE;
+                break;

Previous 2 lines are indented with spaces instead of tabs.

Quoted 9 lines
drivers/i2c/busses/i2c-i801.c +9 -0
+	case I2C_SMBUS_BYTE_DATA:
+		if (read_write == I2C_SMBUS_WRITE)
+			outb_p(data->byte, SMBHSTDAT0(priv));
+		xact = I801_BYTE_DATA;
+		break;
+	case I2C_SMBUS_WORD_DATA:
+		if (read_write == I2C_SMBUS_WRITE) {
+			outb_p(data->word & 0xff, SMBHSTDAT0(priv));
+			outb_p((data->word & 0xff00) >> 8, SMBHSTDAT1(priv));

drivers/i2c/busses/i2c-i801.c static int i801_block_transaction(struct i801_priv *priv, union i2c_smbus_data * · line 805

Quoted 10 lines
drivers/i2c/busses/i2c-i801.c +10 -0
+		}
+		xact = I801_WORD_DATA;
+		break;
+	case I2C_SMBUS_PROC_CALL:
+		outb_p(data->word & 0xff, SMBHSTDAT0(priv));
+		outb_p((data->word & 0xff00) >> 8, SMBHSTDAT1(priv));
+		xact = I801_PROC_CALL;
+		break;
+	default:
+		return -EOPNOTSUPP;

That's never going to happen.

Generally speaking, I'm worried about having the same switch/case construct here that we already have in i801_access. Looks to me like we are doing half of the work here and the other half there and I fail to see the rationale for splitting the work like that. I mean, I see how it solves the asymmetry between the block and non-block code paths, but the result doesn't look appealing. From a performance perspective it's questionable too.

What prevents us from doing all the work on either side? Maybe we should move more code into i801_single_transaction (possibly in a subsequent patch)?

drivers/i2c/busses/i2c-i801.c static int i801_block_transaction(struct i801_priv *priv, union i2c_smbus_data * · line 809

Quoted 4 lines
drivers/i2c/busses/i2c-i801.c +4 -0
+	}
+
+	ret = i801_transaction(priv, xact);
+

Traditionally no blank line here.

Quoted 5 lines
drivers/i2c/busses/i2c-i801.c +5 -0
+	if (ret || read_write == I2C_SMBUS_WRITE)
+		return ret;
+
+	switch (command) {
+	case I2C_SMBUS_BYTE:

drivers/i2c/busses/i2c-i801.c static int i801_block_transaction(struct i801_priv *priv, union i2c_smbus_data * · line 824

Quoted 10 lines
drivers/i2c/busses/i2c-i801.c +10 -0
+	case I2C_SMBUS_BYTE_DATA:
+		data->byte = inb_p(SMBHSTDAT0(priv));
+		break;
+	case I2C_SMBUS_WORD_DATA:
+	case I2C_SMBUS_PROC_CALL:
+		data->word = inb_p(SMBHSTDAT0(priv)) +
+			     (inb_p(SMBHSTDAT1(priv)) << 8);
+		break;
+	default:
+		break;

Default case is not needed.

Quoted 92 lines
drivers/i2c/busses/i2c-i801.c +7 -37
+	}
+
+	return 0;
+}
+
 static void i801_set_hstadd(struct i801_priv *priv, u8 addr, char read_write)
 {
 	addr <<= 1;
@@ -784,9 +840,7 @@ static s32 i801_access(struct i2c_adapter *adap, u16 addr,
 		       unsigned short flags, char read_write, u8 command,
 		       int size, union i2c_smbus_data *data)
 {
-	int hwpec;
-	int block = 0;
-	int ret, xact;
+	int hwpec, ret, block = 0;
 	struct i801_priv *priv = i2c_get_adapdata(adap);
 
 	mutex_lock(&priv->acpi_lock);
@@ -804,36 +858,23 @@ static s32 i801_access(struct i2c_adapter *adap, u16 addr,
 	switch (size) {
 	case I2C_SMBUS_QUICK:
 		i801_set_hstadd(priv, addr, read_write);
-		xact = I801_QUICK;
 		break;
 	case I2C_SMBUS_BYTE:
 		i801_set_hstadd(priv, addr, read_write);
 		if (read_write == I2C_SMBUS_WRITE)
 			outb_p(command, SMBHSTCMD(priv));
-		xact = I801_BYTE;
 		break;
 	case I2C_SMBUS_BYTE_DATA:
 		i801_set_hstadd(priv, addr, read_write);
 		outb_p(command, SMBHSTCMD(priv));
-		if (read_write == I2C_SMBUS_WRITE)
-			outb_p(data->byte, SMBHSTDAT0(priv));
-		xact = I801_BYTE_DATA;
 		break;
 	case I2C_SMBUS_WORD_DATA:
 		i801_set_hstadd(priv, addr, read_write);
 		outb_p(command, SMBHSTCMD(priv));
-		if (read_write == I2C_SMBUS_WRITE) {
-			outb_p(data->word & 0xff, SMBHSTDAT0(priv));
-			outb_p((data->word & 0xff00) >> 8, SMBHSTDAT1(priv));
-		}
-		xact = I801_WORD_DATA;
 		break;
 	case I2C_SMBUS_PROC_CALL:
 		i801_set_hstadd(priv, addr, I2C_SMBUS_WRITE);
 		outb_p(command, SMBHSTCMD(priv));
-		outb_p(data->word & 0xff, SMBHSTDAT0(priv));
-		outb_p((data->word & 0xff00) >> 8, SMBHSTDAT1(priv));
-		xact = I801_PROC_CALL;
 		read_write = I2C_SMBUS_READ;
 		break;
 	case I2C_SMBUS_BLOCK_DATA:
@@ -883,7 +924,7 @@ static s32 i801_access(struct i2c_adapter *adap, u16 addr,
 	if (block)
 		ret = i801_block_transaction(priv, data, read_write, size);
 	else
-		ret = i801_transaction(priv, xact);
+		ret = i801_single_transaction(priv, data, read_write, size);
 
 	/* Some BIOSes don't like it when PEC is enabled at reboot or resume
 	   time, so we forcibly disable it after every transaction. Turn off
@@ -891,26 +932,6 @@ static s32 i801_access(struct i2c_adapter *adap, u16 addr,
 	if (hwpec || block)
 		outb_p(inb_p(SMBAUXCTL(priv)) &
 		       ~(SMBAUXCTL_CRC | SMBAUXCTL_E32B), SMBAUXCTL(priv));
-
-	if (block)
-		goto out;
-	if (ret)
-		goto out;
-	if ((read_write == I2C_SMBUS_WRITE) || (xact == I801_QUICK))
-		goto out;
-
-	switch (xact) {
-	case I801_BYTE:	/* Result put in SMBHSTDAT0 */
-	case I801_BYTE_DATA:
-		data->byte = inb_p(SMBHSTDAT0(priv));
-		break;
-	case I801_WORD_DATA:
-	case I801_PROC_CALL:
-		data->word = inb_p(SMBHSTDAT0(priv)) +
-			     (inb_p(SMBHSTDAT1(priv)) << 8);
-		break;
-	}
-
 out:
 	/*
 	 * Unlock the SMBus device for use by BIOS/ACPI,
Signature · 2 lines
Jean Delvare
SUSE L3 Support
#20 lore raw

Re: [PATCH 6/8] i2c: i801: add i801_single_transaction(), complementing i801_block_transaction()

Quoted 23 lines · Jean Delvare

Hi Heiner,

Quoted 3 lines · Heiner Kallweit

This patch factors out non-block pre/post processing to a new function i801_single_transaction(), complementing existing function i801_block_transaction(). This makes i801_access() better readable.

I like the idea, but I have objections about some implementation details, see below.

Quoted 13 lines
---
 drivers/i2c/busses/i2c-i801.c | 95 +++++++++++++++++++++--------------
 1 file changed, 58 insertions(+), 37 deletions(-)
drivers/i2c/busses/i2c-i801.c +0 -0
index bf77f8640..8c2245f38 100644
@@ -771,6 +771,62 @@ static int i801_block_transaction(struct i801_priv *priv, union i2c_smbus_data *
 	return result;
 }
 
Quoted 8 lines

drivers/i2c/busses/i2c-i801.c static int i801_block_transaction(struct i801_priv *priv, union i2c_smbus_data * · line 774

Quoted 1 line
drivers/i2c/busses/i2c-i801.c +1 -0
+/* Single transaction function */

The term "single transaction" is a bit misleading. Block transactions are also single transactions, in the sense that there's one start condition at the beginning and one stop condition at the end. I'd rather call non-block transactions "single value transactions" or "simple transactions".

OK

Quoted 6 lines
Quoted 6 lines
drivers/i2c/busses/i2c-i801.c +6 -0
+static int i801_single_transaction(struct i801_priv *priv, union i2c_smbus_data *data,
+				   char read_write, int command)
+{
+	int xact, ret;
+
+	switch (command) {
Quoted 9 lines

drivers/i2c/busses/i2c-i801.c static int i801_block_transaction(struct i801_priv *priv, union i2c_smbus_data * · line 786

Quoted 6 lines
drivers/i2c/busses/i2c-i801.c +6 -0
+	case I2C_SMBUS_QUICK:
+		xact = I801_QUICK;
+		break;
+	case I2C_SMBUS_BYTE:
+                xact = I801_BYTE;
+                break;

Previous 2 lines are indented with spaces instead of tabs.

OK

Quoted 30 lines
Quoted 9 lines
drivers/i2c/busses/i2c-i801.c +9 -0
+	case I2C_SMBUS_BYTE_DATA:
+		if (read_write == I2C_SMBUS_WRITE)
+			outb_p(data->byte, SMBHSTDAT0(priv));
+		xact = I801_BYTE_DATA;
+		break;
+	case I2C_SMBUS_WORD_DATA:
+		if (read_write == I2C_SMBUS_WRITE) {
+			outb_p(data->word & 0xff, SMBHSTDAT0(priv));
+			outb_p((data->word & 0xff00) >> 8, SMBHSTDAT1(priv));

drivers/i2c/busses/i2c-i801.c static int i801_block_transaction(struct i801_priv *priv, union i2c_smbus_data * · line 805

Quoted 10 lines
drivers/i2c/busses/i2c-i801.c +10 -0
+		}
+		xact = I801_WORD_DATA;
+		break;
+	case I2C_SMBUS_PROC_CALL:
+		outb_p(data->word & 0xff, SMBHSTDAT0(priv));
+		outb_p((data->word & 0xff00) >> 8, SMBHSTDAT1(priv));
+		xact = I801_PROC_CALL;
+		break;
+	default:
+		return -EOPNOTSUPP;

That's never going to happen.

Generally speaking, I'm worried about having the same switch/case construct here that we already have in i801_access. Looks to me like we are doing half of the work here and the other half there and I fail to see the rationale for splitting the work like that. I mean, I see how it solves the asymmetry between the block and non-block code paths, but the result doesn't look appealing. From a performance perspective it's questionable too.

Quoted 4 lines

What prevents us from doing all the work on either side? Maybe we should move more code into i801_single_transaction (possibly in a subsequent patch)?

Makes sense. Ill add this in v2.

Quoted 7 lines

drivers/i2c/busses/i2c-i801.c static int i801_block_transaction(struct i801_priv *priv, union i2c_smbus_data * · line 809

Quoted 4 lines
drivers/i2c/busses/i2c-i801.c +4 -0
+	}
+
+	ret = i801_transaction(priv, xact);
+

Traditionally no blank line here.

OK

Quoted 9 lines
Quoted 9 lines
drivers/i2c/busses/i2c-i801.c +9 -0
+	if (ret || read_write == I2C_SMBUS_WRITE)
+		return ret;
+
+	switch (command) {
+	case I2C_SMBUS_BYTE:
+	case I2C_SMBUS_BYTE_DATA:
+		data->byte = inb_p(SMBHSTDAT0(priv));
+		break;
+	case I2C_SMBUS_WORD_DATA:
Quoted 9 lines

drivers/i2c/busses/i2c-i801.c static int i801_block_transaction(struct i801_priv *priv, union i2c_smbus_data * · line 824

Quoted 6 lines
drivers/i2c/busses/i2c-i801.c +6 -0
+	case I2C_SMBUS_PROC_CALL:
+		data->word = inb_p(SMBHSTDAT0(priv)) +
+			     (inb_p(SMBHSTDAT1(priv)) << 8);
+		break;
+	default:
+		break;

Default case is not needed.

OK

Quoted 94 lines
Quoted 92 lines
drivers/i2c/busses/i2c-i801.c +7 -37
+	}
+
+	return 0;
+}
+
 static void i801_set_hstadd(struct i801_priv *priv, u8 addr, char read_write)
 {
 	addr <<= 1;
@@ -784,9 +840,7 @@ static s32 i801_access(struct i2c_adapter *adap, u16 addr,
 		       unsigned short flags, char read_write, u8 command,
 		       int size, union i2c_smbus_data *data)
 {
-	int hwpec;
-	int block = 0;
-	int ret, xact;
+	int hwpec, ret, block = 0;
 	struct i801_priv *priv = i2c_get_adapdata(adap);
 
 	mutex_lock(&priv->acpi_lock);
@@ -804,36 +858,23 @@ static s32 i801_access(struct i2c_adapter *adap, u16 addr,
 	switch (size) {
 	case I2C_SMBUS_QUICK:
 		i801_set_hstadd(priv, addr, read_write);
-		xact = I801_QUICK;
 		break;
 	case I2C_SMBUS_BYTE:
 		i801_set_hstadd(priv, addr, read_write);
 		if (read_write == I2C_SMBUS_WRITE)
 			outb_p(command, SMBHSTCMD(priv));
-		xact = I801_BYTE;
 		break;
 	case I2C_SMBUS_BYTE_DATA:
 		i801_set_hstadd(priv, addr, read_write);
 		outb_p(command, SMBHSTCMD(priv));
-		if (read_write == I2C_SMBUS_WRITE)
-			outb_p(data->byte, SMBHSTDAT0(priv));
-		xact = I801_BYTE_DATA;
 		break;
 	case I2C_SMBUS_WORD_DATA:
 		i801_set_hstadd(priv, addr, read_write);
 		outb_p(command, SMBHSTCMD(priv));
-		if (read_write == I2C_SMBUS_WRITE) {
-			outb_p(data->word & 0xff, SMBHSTDAT0(priv));
-			outb_p((data->word & 0xff00) >> 8, SMBHSTDAT1(priv));
-		}
-		xact = I801_WORD_DATA;
 		break;
 	case I2C_SMBUS_PROC_CALL:
 		i801_set_hstadd(priv, addr, I2C_SMBUS_WRITE);
 		outb_p(command, SMBHSTCMD(priv));
-		outb_p(data->word & 0xff, SMBHSTDAT0(priv));
-		outb_p((data->word & 0xff00) >> 8, SMBHSTDAT1(priv));
-		xact = I801_PROC_CALL;
 		read_write = I2C_SMBUS_READ;
 		break;
 	case I2C_SMBUS_BLOCK_DATA:
@@ -883,7 +924,7 @@ static s32 i801_access(struct i2c_adapter *adap, u16 addr,
 	if (block)
 		ret = i801_block_transaction(priv, data, read_write, size);
 	else
-		ret = i801_transaction(priv, xact);
+		ret = i801_single_transaction(priv, data, read_write, size);
 
 	/* Some BIOSes don't like it when PEC is enabled at reboot or resume
 	   time, so we forcibly disable it after every transaction. Turn off
@@ -891,26 +932,6 @@ static s32 i801_access(struct i2c_adapter *adap, u16 addr,
 	if (hwpec || block)
 		outb_p(inb_p(SMBAUXCTL(priv)) &
 		       ~(SMBAUXCTL_CRC | SMBAUXCTL_E32B), SMBAUXCTL(priv));
-
-	if (block)
-		goto out;
-	if (ret)
-		goto out;
-	if ((read_write == I2C_SMBUS_WRITE) || (xact == I801_QUICK))
-		goto out;
-
-	switch (xact) {
-	case I801_BYTE:	/* Result put in SMBHSTDAT0 */
-	case I801_BYTE_DATA:
-		data->byte = inb_p(SMBHSTDAT0(priv));
-		break;
-	case I801_WORD_DATA:
-	case I801_PROC_CALL:
-		data->word = inb_p(SMBHSTDAT0(priv)) +
-			     (inb_p(SMBHSTDAT1(priv)) << 8);
-		break;
-	}
-
 out:
 	/*
 	 * Unlock the SMBus device for use by BIOS/ACPI,
#21 lore raw

PATCH 7/8 i2c: i801: call i801_check_pre() from i801_access()

In reply to Heiner Kallweit

This avoids code duplication, in a next step we'll call i801_check_post() from i801_access() as well.

---
 drivers/i2c/busses/i2c-i801.c | 12 ++++--------
 1 file changed, 4 insertions(+), 8 deletions(-)
drivers/i2c/busses/i2c-i801.c +4 -8
index 8c2245f38..9061333f2 100644
@@ -460,10 +460,6 @@ static int i801_transaction(struct i801_priv *priv, int xact)
 	unsigned long result;
 	const struct i2c_adapter *adap = &priv->adapter;
 
-	status = i801_check_pre(priv);
-	if (status < 0)
-		return status;
-
 	if (priv->features & FEATURE_IRQ) {
 		reinit_completion(&priv->done);
 		outb_p(xact | SMBHSTCNT_INTREN | SMBHSTCNT_START,
@@ -647,10 +643,6 @@ static int i801_block_transaction_byte_by_byte(struct i801_priv *priv,
 	if (command == I2C_SMBUS_BLOCK_PROC_CALL)
 		return -EOPNOTSUPP;
 
-	status = i801_check_pre(priv);
-	if (status < 0)
-		return status;
-
 	len = data->block[0];
 
 	if (read_write == I2C_SMBUS_WRITE) {
@@ -851,6 +843,10 @@ static s32 i801_access(struct i2c_adapter *adap, u16 addr,
 
 	pm_runtime_get_sync(&priv->pci_dev->dev);
 
+	ret = i801_check_pre(priv);
+	if (ret)
+		goto out;
+
 	hwpec = (priv->features & FEATURE_SMBUS_PEC) && (flags & I2C_CLIENT_PEC)
 		&& size != I2C_SMBUS_QUICK
 		&& size != I2C_SMBUS_I2C_BLOCK_DATA;
Signature · 1 line
2.35.3
Reviewed-by on v1 7/8
#22 lore raw

Re: [PATCH 7/8] i2c: i801: call i801_check_pre() from i801_access()

Hi Heiner,

Quoted 35 lines · Heiner Kallweit

This avoids code duplication, in a next step we'll call i801_check_post() from i801_access() as well.

---
 drivers/i2c/busses/i2c-i801.c | 12 ++++--------
 1 file changed, 4 insertions(+), 8 deletions(-)
drivers/i2c/busses/i2c-i801.c +0 -8
index 8c2245f38..9061333f2 100644
@@ -460,10 +460,6 @@ static int i801_transaction(struct i801_priv *priv, int xact)
 	unsigned long result;
 	const struct i2c_adapter *adap = &priv->adapter;
 
-	status = i801_check_pre(priv);
-	if (status < 0)
-		return status;
-
 	if (priv->features & FEATURE_IRQ) {
 		reinit_completion(&priv->done);
 		outb_p(xact | SMBHSTCNT_INTREN | SMBHSTCNT_START,
@@ -647,10 +643,6 @@ static int i801_block_transaction_byte_by_byte(struct i801_priv *priv,
 	if (command == I2C_SMBUS_BLOCK_PROC_CALL)
 		return -EOPNOTSUPP;
 
-	status = i801_check_pre(priv);
-	if (status < 0)
-		return status;
-
 	len = data->block[0];
 
 	if (read_write == I2C_SMBUS_WRITE) {

drivers/i2c/busses/i2c-i801.c static s32 i801_access(struct i2c_adapter *adap, u16 addr, · line 852

Quoted 10 lines
drivers/i2c/busses/i2c-i801.c +4 -0
@@ -851,6 +843,10 @@ static s32 i801_access(struct i2c_adapter *adap, u16 addr,
 
 	pm_runtime_get_sync(&priv->pci_dev->dev);
 
+	ret = i801_check_pre(priv);
+	if (ret)
+		goto out;
+
 	hwpec = (priv->features & FEATURE_SMBUS_PEC) && (flags & I2C_CLIENT_PEC)
 		&& size != I2C_SMBUS_QUICK
 		&& size != I2C_SMBUS_I2C_BLOCK_DATA;

Makes sense, thanks for the clean-up.

Signature · 2 lines
Jean Delvare
SUSE L3 Support
#23 lore raw

PATCH 8/8 i2c: i801: call i801_check_post() from i801_access()

In reply to Heiner Kallweit

Avoid code duplication by calling i801_check_post() from i801_access().

---
 drivers/i2c/busses/i2c-i801.c | 20 +++++++++-----------
 1 file changed, 9 insertions(+), 11 deletions(-)
drivers/i2c/busses/i2c-i801.c +9 -11
index 9061333f2..ecec7a3a8 100644
@@ -432,7 +432,7 @@ static int i801_wait_intr(struct i801_priv *priv)
 		busy = status & SMBHSTSTS_HOST_BUSY;
 		status &= STATUS_ERROR_FLAGS | SMBHSTSTS_INTR;
 		if (!busy && status)
-			return status;
+			return status & STATUS_ERROR_FLAGS;
 	} while (time_is_after_eq_jiffies(timeout));
 
 	return -ETIMEDOUT;
@@ -456,7 +456,6 @@ static int i801_wait_byte_done(struct i801_priv *priv)
 
 static int i801_transaction(struct i801_priv *priv, int xact)
 {
-	int status;
 	unsigned long result;
 	const struct i2c_adapter *adap = &priv->adapter;
 
@@ -465,13 +464,12 @@ static int i801_transaction(struct i801_priv *priv, int xact)
 		outb_p(xact | SMBHSTCNT_INTREN | SMBHSTCNT_START,
 		       SMBHSTCNT(priv));
 		result = wait_for_completion_timeout(&priv->done, adap->timeout);
-		return i801_check_post(priv, result ? priv->status : -ETIMEDOUT);
+		return result ? priv->status : -ETIMEDOUT;
 	}
 
 	outb_p(xact | SMBHSTCNT_START, SMBHSTCNT(priv));
 
-	status = i801_wait_intr(priv);
-	return i801_check_post(priv, status);
+	return i801_wait_intr(priv);
 }
 
 static int i801_block_transaction_by_block(struct i801_priv *priv,
@@ -618,7 +616,7 @@ static irqreturn_t i801_isr(int irq, void *dev_id)
 
 	status &= STATUS_ERROR_FLAGS | SMBHSTSTS_INTR;
 	if (status) {
-		priv->status = status;
+		priv->status = status & STATUS_ERROR_FLAGS;
 		complete(&priv->done);
 	}
 
@@ -668,7 +666,7 @@ static int i801_block_transaction_byte_by_byte(struct i801_priv *priv,
 		reinit_completion(&priv->done);
 		outb_p(priv->cmd | SMBHSTCNT_START, SMBHSTCNT(priv));
 		result = wait_for_completion_timeout(&priv->done, adap->timeout);
-		return i801_check_post(priv, result ? priv->status : -ETIMEDOUT);
+		return result ? priv->status : -ETIMEDOUT;
 	}
 
 	for (i = 1; i <= len; i++) {
@@ -682,7 +680,7 @@ static int i801_block_transaction_byte_by_byte(struct i801_priv *priv,
 
 		status = i801_wait_byte_done(priv);
 		if (status)
-			goto exit;
+			return status;
 
 		if (i == 1 && read_write == I2C_SMBUS_READ
 		 && command != I2C_SMBUS_I2C_BLOCK_DATA) {
@@ -712,9 +710,7 @@ static int i801_block_transaction_byte_by_byte(struct i801_priv *priv,
 		outb_p(SMBHSTSTS_BYTE_DONE, SMBHSTSTS(priv));
 	}
 
-	status = i801_wait_intr(priv);
-exit:
-	return i801_check_post(priv, status);
+	return i801_wait_intr(priv);
 }
 
 /* Block transaction function */
@@ -922,6 +918,8 @@ static s32 i801_access(struct i2c_adapter *adap, u16 addr,
 	else
 		ret = i801_single_transaction(priv, data, read_write, size);
 
+	ret = i801_check_post(priv, ret);
+
 	/* Some BIOSes don't like it when PEC is enabled at reboot or resume
 	   time, so we forcibly disable it after every transaction. Turn off
 	   E32B for the same reason. */
Signature · 1 line
2.35.3
#24 lore raw

Re: [PATCH 8/8] i2c: i801: call i801_check_post() from i801_access()

Hi Heiner,

Quoted 6 lines · Heiner Kallweit

Avoid code duplication by calling i801_check_post() from i801_access().

---
 drivers/i2c/busses/i2c-i801.c | 20 +++++++++-----------
 1 file changed, 9 insertions(+), 11 deletions(-)

Overall I like the idea. I only have one question to make sure I'm not missing something.

drivers/i2c/busses/i2c-i801.c static int i801_wait_intr(struct i801_priv *priv) · line 436

Quoted 11 lines
drivers/i2c/busses/i2c-i801.c +1 -1
index 9061333f2..ecec7a3a8 100644
@@ -432,7 +432,7 @@ static int i801_wait_intr(struct i801_priv *priv)
 		busy = status & SMBHSTSTS_HOST_BUSY;
 		status &= STATUS_ERROR_FLAGS | SMBHSTSTS_INTR;
 		if (!busy && status)
-			return status;
+			return status & STATUS_ERROR_FLAGS;
 	} while (time_is_after_eq_jiffies(timeout));

Do I understand correctly that this change isn't really related to the rest of the patch, and could have been done independently?

You are filtering out SMBHSTSTS_INTR simply because i801_check_post() will never check it anyway, right? If so, I wonder if that's really something we want to do, as ultimately this adds code with no functional benefit just to be "cleaner". But please correct me if I'm wrong.

Signature · 2 lines
Jean Delvare
SUSE L3 Support
#25 lore raw

Re: [PATCH 8/8] i2c: i801: call i801_check_post() from i801_access()

Quoted 28 lines · Jean Delvare

Hi Heiner,

Quoted 6 lines · Heiner Kallweit

Avoid code duplication by calling i801_check_post() from i801_access().

---
 drivers/i2c/busses/i2c-i801.c | 20 +++++++++-----------
 1 file changed, 9 insertions(+), 11 deletions(-)

Overall I like the idea. I only have one question to make sure I'm not missing something.

drivers/i2c/busses/i2c-i801.c static int i801_wait_intr(struct i801_priv *priv) · line 436

Quoted 11 lines
drivers/i2c/busses/i2c-i801.c +1 -1
index 9061333f2..ecec7a3a8 100644
@@ -432,7 +432,7 @@ static int i801_wait_intr(struct i801_priv *priv)
 		busy = status & SMBHSTSTS_HOST_BUSY;
 		status &= STATUS_ERROR_FLAGS | SMBHSTSTS_INTR;
 		if (!busy && status)
-			return status;
+			return status & STATUS_ERROR_FLAGS;
 	} while (time_is_after_eq_jiffies(timeout));

Do I understand correctly that this change isn't really related to the rest of the patch, and could have been done independently?

Quoted 6 lines

You are filtering out SMBHSTSTS_INTR simply because i801_check_post() will never check it anyway, right? If so, I wonder if that's really something we want to do, as ultimately this adds code with no functional benefit just to be "cleaner". But please correct me if I'm wrong.

Reason is that in few places we check whether return value of i801_wait_intr() is zero, this would fail if not filtering out SMBHSTSTS_INTR. Example:
i801_transaction() returns the return value of i801_wait_intr() now. And in i801_block_transaction_by_block() we check whether return value of i801_transaction() is zero.