* [PATCH v4 1/2] at24: support eeproms that do not auto-rollover reads.
2017-12-05 15:23 [PATCH v4 0/2] at24: support eeproms that do not auto-rollover reads Sven Van Asbroeck
@ 2017-12-05 15:23 ` Sven Van Asbroeck
2017-12-05 15:23 ` [PATCH v4 2/2] dt-bindings: add eeprom "at24,no-read-rollover" property Sven Van Asbroeck
1 sibling, 0 replies; 6+ messages in thread
From: Sven Van Asbroeck @ 2017-12-05 15:23 UTC (permalink / raw)
To: svendev, robh+dt, mark.rutland, wsa, brgl, nsekhar, sakari.ailus,
david, javier, divagar.mohandass
Cc: devicetree, linux-kernel, linux-i2c
Some multi-address eeproms in the at24 family may not automatically
roll-over reads to the next slave address. On those eeproms, reads
that straddle slave boundaries will not work correctly.
Solution:
Mark such eeproms with a flag that prevents reads straddling
slave boundaries. Add the AT24_FLAG_NO_RDROL flag to the eeprom
entry in the device_id table, or add 'at24,no-read-rollover' to the
eeprom devicetree entry.
Note that I have not personally enountered an at24 chip that
does not support read rollovers. They may or may not exist.
However, my hardware requires this functionality because of
a quirk.
It's up to the Linux community to decide if this patch is useful/
general enough to warrant merging.
Signed-off-by: Sven Van Asbroeck <svendev@arcx.com>
---
drivers/misc/eeprom/at24.c | 37 +++++++++++++++++++++++++------------
include/linux/platform_data/at24.h | 2 ++
2 files changed, 27 insertions(+), 12 deletions(-)
diff --git a/drivers/misc/eeprom/at24.c b/drivers/misc/eeprom/at24.c
index 625b001..8c93ed0 100644
--- a/drivers/misc/eeprom/at24.c
+++ b/drivers/misc/eeprom/at24.c
@@ -251,15 +251,6 @@ struct at24_data {
* Slave address and byte offset derive from the offset. Always
* set the byte address; on a multi-master board, another master
* may have changed the chip's "current" address pointer.
- *
- * REVISIT some multi-address chips don't rollover page reads to
- * the next slave address, so we may need to truncate the count.
- * Those chips might need another quirk flag.
- *
- * If the real hardware used four adjacent 24c02 chips and that
- * were misconfigured as one 24c08, that would be a similar effect:
- * one "eeprom" file not four, but larger reads would fail when
- * they crossed certain pages.
*/
static struct at24_client *at24_translate_offset(struct at24_data *at24,
unsigned int *offset)
@@ -277,6 +268,28 @@ static struct at24_client *at24_translate_offset(struct at24_data *at24,
return &at24->client[i];
}
+static size_t at24_adjust_read_count(struct at24_data *at24,
+ unsigned int offset, size_t count)
+{
+ unsigned int bits;
+ size_t remainder;
+ /*
+ * In case of multi-address chips that don't rollover reads to
+ * the next slave address: truncate the count to the slave boundary,
+ * so that the read never straddles slaves.
+ */
+ if (at24->chip.flags & AT24_FLAG_NO_RDROL) {
+ bits = (at24->chip.flags & AT24_FLAG_ADDR16) ? 16 : 8;
+ remainder = BIT(bits) - offset;
+ if (count > remainder)
+ count = remainder;
+ }
+ if (count > io_limit)
+ count = io_limit;
+
+ return count;
+}
+
static ssize_t at24_regmap_read(struct at24_data *at24, char *buf,
unsigned int offset, size_t count)
{
@@ -289,9 +302,7 @@ static ssize_t at24_regmap_read(struct at24_data *at24, char *buf,
at24_client = at24_translate_offset(at24, &offset);
regmap = at24_client->regmap;
client = at24_client->client;
-
- if (count > io_limit)
- count = io_limit;
+ count = at24_adjust_read_count(at24, offset, count);
/* adjust offset for mac and serial read ops */
offset += at24->offset_adj;
@@ -457,6 +468,8 @@ static void at24_get_pdata(struct device *dev, struct at24_platform_data *chip)
if (device_property_present(dev, "read-only"))
chip->flags |= AT24_FLAG_READONLY;
+ if (device_property_present(dev, "at24,no-read-rollover"))
+ chip->flags |= AT24_FLAG_NO_RDROL;
err = device_property_read_u32(dev, "size", &val);
if (!err)
diff --git a/include/linux/platform_data/at24.h b/include/linux/platform_data/at24.h
index 271a4e2..841bb28 100644
--- a/include/linux/platform_data/at24.h
+++ b/include/linux/platform_data/at24.h
@@ -50,6 +50,8 @@ struct at24_platform_data {
#define AT24_FLAG_TAKE8ADDR BIT(4) /* take always 8 addresses (24c00) */
#define AT24_FLAG_SERIAL BIT(3) /* factory-programmed serial number */
#define AT24_FLAG_MAC BIT(2) /* factory-programmed mac address */
+#define AT24_FLAG_NO_RDROL BIT(1) /* does not auto-rollover reads to */
+ /* the next slave address */
void (*setup)(struct nvmem_device *nvmem, void *context);
void *context;
--
1.9.1
^ permalink raw reply related [flat|nested] 6+ messages in thread* [PATCH v4 2/2] dt-bindings: add eeprom "at24,no-read-rollover" property
2017-12-05 15:23 [PATCH v4 0/2] at24: support eeproms that do not auto-rollover reads Sven Van Asbroeck
2017-12-05 15:23 ` [PATCH v4 1/2] " Sven Van Asbroeck
@ 2017-12-05 15:23 ` Sven Van Asbroeck
2017-12-06 22:04 ` Rob Herring
1 sibling, 1 reply; 6+ messages in thread
From: Sven Van Asbroeck @ 2017-12-05 15:23 UTC (permalink / raw)
To: svendev, robh+dt, mark.rutland, wsa, brgl, nsekhar, sakari.ailus,
david, javier, divagar.mohandass
Cc: devicetree, linux-kernel, linux-i2c
Adds an optional property for at24 eeproms.
This parameterless property indicates that the multi-address eeprom
does not automatically roll over reads to the next slave address.
Signed-off-by: Sven Van Asbroeck <svendev@arcx.com>
---
Documentation/devicetree/bindings/eeprom/eeprom.txt | 5 +++++
1 file changed, 5 insertions(+)
diff --git a/Documentation/devicetree/bindings/eeprom/eeprom.txt b/Documentation/devicetree/bindings/eeprom/eeprom.txt
index 27f2bc1..5bfc0ac 100644
--- a/Documentation/devicetree/bindings/eeprom/eeprom.txt
+++ b/Documentation/devicetree/bindings/eeprom/eeprom.txt
@@ -38,6 +38,11 @@ Optional properties:
- size: total eeprom size in bytes
+ - at24,no-read-rollover:
+ This parameterless property indicates that the multi-address
+ eeprom does not automatically roll over reads to the next
+ slave address. Please consult the manual of your device.
+
Example:
eeprom@52 {
--
1.9.1
^ permalink raw reply related [flat|nested] 6+ messages in thread