Skip to contentSkip to conversation

at24: Support SMBus read/write of 16-bit devices

PATCH Superseded Aaron Sierra 2 participants 2 messages Nov 4, 2015 – Nov 5, 2015 1 message not in the archive

Subsystems (2)

v3 of 6 Superseded Latest is v6 →

lore
Share & tools
lore

A message that is not in the archive (sent off-list or not yet received): [email protected]

#1 lore raw

PATCH v3 1/1 at24: Support SMBus read/write of 16-bit devices

Previously, the at24 driver would bail out in the case of a 16-bit addressable EEPROM attached to an SMBus controller. This is because SMBus block reads and writes don't map to I2C multi-byte reads and writes when the offset portion is 2 bytes.

Instead of bailing out, this patch settles for functioning with single byte read SMBus cycles. Writes can be block or single-byte, depending on SMBus controller features.

Read access is not without some risk. Multiple SMBus cycles are required to read even one byte. If the SMBus has multiple masters and one accesses this EEPROM between the dummy address write and the subsequent current-address-read cycle(s), this driver will receive data from the wrong address.

Functionality has been tested with the following devices:

AT24CM01 attached to Intel ISCH SMBus
AT24C512 attached to Intel I801 SMBus

Read performance:
3.6 KB/s with 32-byte* access

*limited to 32-bytes by I2C_SMBUS_BLOCK_MAX.

Write performance:
248 B/s with 1-byte page (default)
3.9 KB/s with 128-byte* page (via platform data)

*limited to 31-bytes by I2C_SMBUS_BLOCK_MAX - 1.

---
v2 - Account for changes related to introduction of

      i2c_smbus_read_i2c_block_data_or_emulated()

v3 - Consolidate three patches into one
- Expand comments regarding SMBus multi-master read risks.
- Rely on current-address-read for improved read performance (i.e. one dummy address write followed by multiple individual byte reads). This improves performance from 1.4 KiB/s to 3.6 KiB/s.
- Use struct at24_data's writebuf instead of kzalloc-ing
- Only limit write_max by 1-byte when accessing a 16-bit device with block writes instead of attempting to preserve a power-of-two.
- Style fixes (indentation, parentheses, unnecessary masking, etc.)

 drivers/misc/eeprom/Kconfig |   4 +-
 drivers/misc/eeprom/at24.c  | 118 +++++++++++++++++++++++++++++++++++++++-----
 2 files changed, 109 insertions(+), 13 deletions(-)
drivers/misc/eeprom/Kconfig +3 -1
index 04f2e1f..bc79a44 100644
@@ -22,7 +22,9 @@ config EEPROM_AT24
 
 	  If you use this with an SMBus adapter instead of an I2C adapter,
 	  full functionality is not available.  Only smaller devices are
-	  supported (24c16 and below, max 4 kByte).
+	  supported via block reads (24c16 and below, max 4 kByte).
+	  Larger devices that use 16-bit addresses will only work with
+	  individual byte reads, which is very slow.
 
 	  This driver can also be built as a module.  If so, the module
 	  will be called at24.
drivers/misc/eeprom/at24.c +106 -12

162 lines not on this page · raw message

Signature · 1 line
1.9.1
#2 lore raw

Hi Aaron,

Quoted 55 lines · Aaron Sierra

Previously, the at24 driver would bail out in the case of a 16-bit addressable EEPROM attached to an SMBus controller. This is because SMBus block reads and writes don't map to I2C multi-byte reads and writes when the offset portion is 2 bytes.

Instead of bailing out, this patch settles for functioning with single byte read SMBus cycles. Writes can be block or single-byte, depending on SMBus controller features.

Read access is not without some risk. Multiple SMBus cycles are required to read even one byte. If the SMBus has multiple masters and one accesses this EEPROM between the dummy address write and the subsequent current-address-read cycle(s), this driver will receive data from the wrong address.

Functionality has been tested with the following devices:

AT24CM01 attached to Intel ISCH SMBus
AT24C512 attached to Intel I801 SMBus

Read performance:
3.6 KB/s with 32-byte* access

*limited to 32-bytes by I2C_SMBUS_BLOCK_MAX.

Write performance:
248 B/s with 1-byte page (default)
3.9 KB/s with 128-byte* page (via platform data)

*limited to 31-bytes by I2C_SMBUS_BLOCK_MAX - 1.

---
v2 - Account for changes related to introduction of

      i2c_smbus_read_i2c_block_data_or_emulated()

v3 - Consolidate three patches into one
- Expand comments regarding SMBus multi-master read risks.
- Rely on current-address-read for improved read performance (i.e. one dummy address write followed by multiple individual byte reads). This improves performance from 1.4 KiB/s to 3.6 KiB/s.
- Use struct at24_data's writebuf instead of kzalloc-ing
- Only limit write_max by 1-byte when accessing a 16-bit device with block writes instead of attempting to preserve a power-of-two.
- Style fixes (indentation, parentheses, unnecessary masking, etc.)

 drivers/misc/eeprom/Kconfig |   4 +-
 drivers/misc/eeprom/at24.c  | 118 +++++++++++++++++++++++++++++++++++++++-----
 2 files changed, 109 insertions(+), 13 deletions(-)
drivers/misc/eeprom/Kconfig +0 -0
index 04f2e1f..bc79a44 100644

drivers/misc/eeprom/Kconfig config EEPROM_AT24 · line 27

Quoted 7 lines
drivers/misc/eeprom/Kconfig +3 -1
@@ -22,7 +22,9 @@ config EEPROM_AT24
 
 	  If you use this with an SMBus adapter instead of an I2C adapter,
 	  full functionality is not available.  Only smaller devices are
-	  supported (24c16 and below, max 4 kByte).
+	  supported via block reads (24c16 and below, max 4 kByte).
+	  Larger devices that use 16-bit addresses will only work with
+	  individual byte reads, which is very slow.

... and unsafe on multi-master topologies.

Quoted 55 lines
drivers/misc/eeprom/Kconfig +0 -0
 	  This driver can also be built as a module.  If so, the module
 	  will be called at24.
drivers/misc/eeprom/at24.c +36 -1
index 5d7c090..35c5c83 100644
@@ -50,7 +50,7 @@
  * Other than binding model, current differences from "eeprom" driver are
  * that this one handles write access and isn't restricted to 24c02 devices.
  * It also handles larger devices (32 kbit and up) with two-byte addresses,
- * which won't work on pure SMBus systems.
+ * which don't work without risks on pure SMBus systems.
  */
 
 struct at24_data {
@@ -141,6 +141,86 @@ MODULE_DEVICE_TABLE(acpi, at24_acpi_ids);
 /*-------------------------------------------------------------------------*/
 
 /*
+ * Write a byte to an AT24 device using SMBus cycles.
+ */
+static inline s32 at24_smbus_write_byte_data(struct at24_data *at24,
+	struct i2c_client *client, u16 offset, u8 value)
+{
+	if (!(at24->chip.flags & AT24_FLAG_ADDR16))
+		return i2c_smbus_write_byte_data(client, offset, value);
+
+	/*
+	 * Emulate I2C multi-byte write by using SMBus "write word"
+	 * cycle.  We split up the 16-bit offset among the "command"
+	 * byte and the first data byte.
+	 */
+	return i2c_smbus_write_word_data(client,
+		offset >> 8, (value << 8) | (offset & 0xff));
+}
+
+/*
+ * Write block data to an AT24 device using SMBus cycles.
+ */
+static inline s32 at24_smbus_write_i2c_block_data(struct at24_data *at24,
+	const struct i2c_client *client, u16 off, u8 len, const u8 *vals)
+{
+	s32 res;
+
+	if (!(at24->chip.flags & AT24_FLAG_ADDR16))
+		return i2c_smbus_write_i2c_block_data(client, off, len, vals);
+
+	/* Insert extra address byte into data stream */
+	at24->writebuf[0] = off & 0xff;
+	memcpy(&at24->writebuf[1], vals, len);
+
+	res = i2c_smbus_write_i2c_block_data(client,
+		off >> 8, len + 1, at24->writebuf);
+

drivers/misc/eeprom/at24.c MODULE_DEVICE_TABLE(acpi, at24_acpi_ids); · line 188

Quoted 10 lines
drivers/misc/eeprom/at24.c +10 -0
+	return res;
+}
+
+/*
+ * Read block data from an AT24 device using SMBus cycles.
+ */
+static inline s32 at24_smbus_read_block_data(struct at24_data *at24,
+	const struct i2c_client *client, u16 off, u8 len, u8 *vals)
+{
+	int count = 0;

As a minor optimization, "count" could be initialized later, right before the do/while loop.

Quoted 20 lines
drivers/misc/eeprom/at24.c +20 -0
+	s32 res;
+
+	if (!(at24->chip.flags & AT24_FLAG_ADDR16))
+		return i2c_smbus_read_i2c_block_data_or_emulated(client,
+				off, len, vals);
+
+	/*
+	 * Emulate I2C multi-byte read by using SMBus "write byte" and
+	 * "receive byte".  This is slightly unsafe since there is an
+	 * additional STOP involved, which exposes the SMBus and (this
+	 * device!) to takeover by another bus master. However, it's the
+	 * only way to work on SMBus-only controllers when talking to
+	 * EEPROMs with multi-byte addresses.
+	 */
+
+	/* Address "dummy" write */
+	res = i2c_smbus_write_byte_data(client, off >> 8, off & 0xff);
+	if (res < 0)
+		return res;
+

drivers/misc/eeprom/at24.c MODULE_DEVICE_TABLE(acpi, at24_acpi_ids); · line 218

Quoted 10 lines
drivers/misc/eeprom/at24.c +10 -0
+	do {
+		/* Current Address Read */
+		res = i2c_smbus_read_byte(client);
+		if (res < 0)
+			break;
+
+		*(vals++) = res;
+		count++;
+		len--;
+	} while (len > 0);

You never checked that the adapter supports I2C_FUNC_SMBUS_WRITE_BYTE_DATA and I2C_FUNC_SMBUS_READ_BYTE.

drivers/misc/eeprom/at24.c static ssize_t at24_eeprom_read(struct at24_data *at24, char *buf,

Quoted 9 lines
drivers/misc/eeprom/at24.c +5 -0
+
+	return count;
+}
+
+/*
  * This routine supports chips which consume multiple I2C addresses. It
  * computes the addressing information to be used for a given r/w request.
  * Assumes that sanity checks for offset happened at sysfs-layer.
@@ -229,7 +309,7 @@ static ssize_t at24_eeprom_read(struct at24_data *at24, char *buf,

Earlier in this function is a test:

	if (at24->use_smbus) {
		/* Smaller eeproms can work given some SMBus extension calls */
		if (count > I2C_SMBUS_BLOCK_MAX)
			count = I2C_SMBUS_BLOCK_MAX;

Limiting count that way is not needed in the 16-bit addressing case. Skipping this will bring a small performance boost (4-5 %) as you will be able to read the whole EEPROM after setting the initial address pointer only once, rather than every 32 bytes. Can be done in a separate commit.

drivers/misc/eeprom/at24.c static ssize_t at24_eeprom_read(struct at24_data *at24, char *buf, · line 313

Quoted 6 lines
drivers/misc/eeprom/at24.c +1 -1
 	do {
 		read_time = jiffies;
 		if (at24->use_smbus) {
-			status = i2c_smbus_read_i2c_block_data_or_emulated(client, offset,
+			status = at24_smbus_read_block_data(at24, client, offset,
 									   count, buf);

Both lines are over 80 columns, and please preserve the alignment on the opening parenthesis.

Quoted 57 lines
drivers/misc/eeprom/at24.c +22 -9
 		} else {
 			status = i2c_transfer(client->adapter, msg, 2);
@@ -351,12 +431,12 @@ static ssize_t at24_eeprom_write(struct at24_data *at24, const char *buf,
 		if (at24->use_smbus_write) {
 			switch (at24->use_smbus_write) {
 			case I2C_SMBUS_I2C_BLOCK_DATA:
-				status = i2c_smbus_write_i2c_block_data(client,
-						offset, count, buf);
+				status = at24_smbus_write_i2c_block_data(at24,
+						client, offset, count, buf);
 				break;
 			case I2C_SMBUS_BYTE_DATA:
-				status = i2c_smbus_write_byte_data(client,
-						offset, buf[0]);
+				status = at24_smbus_write_byte_data(at24,
+						client, offset, buf[0]);
 				break;
 			}
 
@@ -527,10 +607,13 @@ static int at24_probe(struct i2c_client *client, const struct i2c_device_id *id)
 
 	/* Use I2C operations unless we're stuck with SMBus extensions. */
 	if (!i2c_check_functionality(client->adapter, I2C_FUNC_I2C)) {
-		if (chip.flags & AT24_FLAG_ADDR16)
-			return -EPFNOSUPPORT;
-
-		if (i2c_check_functionality(client->adapter,
+		if (chip.flags & AT24_FLAG_ADDR16) {
+			/*
+			 * This will be slow, but better than nothing
+			 * (e.g. read @ 3.6 KiB/s).
+			 */
+			use_smbus = I2C_SMBUS_BYTE_DATA;
+		} else if (i2c_check_functionality(client->adapter,
 				I2C_FUNC_SMBUS_READ_I2C_BLOCK)) {
 			use_smbus = I2C_SMBUS_I2C_BLOCK_DATA;
 		} else if (i2c_check_functionality(client->adapter,
@@ -549,7 +632,17 @@ static int at24_probe(struct i2c_client *client, const struct i2c_device_id *id)
 		if (i2c_check_functionality(client->adapter,
 				I2C_FUNC_SMBUS_WRITE_I2C_BLOCK)) {
 			use_smbus_write = I2C_SMBUS_I2C_BLOCK_DATA;
-		} else if (i2c_check_functionality(client->adapter,
+		} else if ((chip.flags & AT24_FLAG_ADDR16) &&
+		    i2c_check_functionality(client->adapter,
+				I2C_FUNC_SMBUS_WRITE_WORD_DATA)) {
+			/*
+			 * We need SMBUS_WRITE_WORD_DATA to implement
+			 * byte writes for 16-bit address devices.
+			 */
+			use_smbus_write = I2C_SMBUS_BYTE_DATA;
+			chip.page_size = 1;
+		} else if (!(chip.flags & AT24_FLAG_ADDR16) &&
+			i2c_check_functionality(client->adapter,
 				I2C_FUNC_SMBUS_WRITE_BYTE_DATA)) {
 			use_smbus_write = I2C_SMBUS_BYTE_DATA;
 			chip.page_size = 1;

drivers/misc/eeprom/at24.c static int at24_probe(struct i2c_client *client, const struct i2c_device_id *id) · line 696

Quoted 6 lines
drivers/misc/eeprom/at24.c +2 -1
@@ -599,7 +692,8 @@ static int at24_probe(struct i2c_client *client, const struct i2c_device_id *id)
 			if (write_max > io_limit)
 				write_max = io_limit;
 			if (use_smbus && write_max > I2C_SMBUS_BLOCK_MAX)
-				write_max = I2C_SMBUS_BLOCK_MAX;
+				write_max = I2C_SMBUS_BLOCK_MAX -
+					!!(chip.flags & AT24_FLAG_ADDR16);

This test still looks wrong to me when AT24_FLAG_ADDR16 is set. If write_max is 33 you'll properly clip it to 31. However if it's 32 you'll leave it at 32 and the subsequent block write attempts of 33 bytes will fail.

drivers/misc/eeprom/at24.c static int at24_probe(struct i2c_client *client, const struct i2c_device_id *id) · line 699

Quoted 3 lines
drivers/misc/eeprom/at24.c +0 -0
 			at24->write_max = write_max;
 
 			/* buffer (data + address at the beginning) */

Otherwise looks "good", assuming the user understands the risks and limitations.

Signature · 2 lines
Jean Delvare
SUSE L3 Support