This series includes few logical sets that bring a support of non-ACPI
platforms for Intel Skylake.
First part is a refactoring of built-in device properties support:
- keep single value inside the structure
- provide helper macros to define built-in properties
- fall back to secondary fwnode if primary has no asked property
Second one is modifications to MFD code and intel-lpss.c driver in particular
to define and pass built-in properties to the individual drivers.
Last part is a fix for I2C bug found on Lenovo Yoga hardware and a first
converted user.
Built-in device properties is an alternative to platform data. It provides a
unified API that drivers can use to cover all cases at once: DT, ACPI, and
built-in properties.
With this series applied platform data can be considered obsolete. Moreover,
built-in device properties allows to adjust existing configuration, for
example, in cases when ACPI values are wrong on some platforms.
The series has been tested on available hardware and doesn't break current
behaviour. But we ask you, Kevin, to apply the series on your side and check
with Lenovo hardware.
Andy Shevchenko (8):
device property: always check for fwnode type
device property: rename helper functions
device property: refactor built-in properties support
device property: keep single value inplace
device property: improve readability of macros
device property: return -EINVAL when property isn't found in ACPI
device property: Fallback to secondary fwnode if primary misses the
property
mfd: core: propagate device properties to sub devices drivers
Heikki Krogerus (1):
device property: helper macros for property entry creation
Mika Westerberg (4):
mfd: intel-lpss: Add support for passing device properties
mfd: intel-lpss: Pass HSUART configuration via properties
mfd: intel-lpss: Pass SDA hold time to I2C host controller driver
i2c: designware: Convert to use unified device property API
drivers/acpi/property.c | 10 +-
drivers/base/property.c | 298 +++++++++++++++++++++-------
drivers/i2c/busses/i2c-designware-platdrv.c | 48 ++---
drivers/mfd/intel-lpss-acpi.c | 18 +-
drivers/mfd/intel-lpss-pci.c | 43 +++-
drivers/mfd/intel-lpss.c | 16 +-
drivers/mfd/intel-lpss.h | 2 +
drivers/mfd/mfd-core.c | 3 +
include/linux/mfd/core.h | 5 +
include/linux/property.h | 87 ++++++--
10 files changed, 394 insertions(+), 136 deletions(-)
--
2.6.2
Currently the property accessors unconditionally fall back to built-in property
set as a last resort. Make this strict and return an error in case the type of
fwnode is unknown.
This is actually a follow up to the commit 4fa7508e9f1c (device property:
Return -ENXIO if there is no suitable FW interface).
Signed-off-by: Andy Shevchenko <[email protected]>
---
drivers/base/property.c | 12 +++++++-----
1 file changed, 7 insertions(+), 5 deletions(-)
diff --git a/drivers/base/property.c b/drivers/base/property.c
index 1325ff2..09e488d 100644
--- a/drivers/base/property.c
+++ b/drivers/base/property.c
@@ -135,8 +135,9 @@ bool fwnode_property_present(struct fwnode_handle *fwnode, const char *propname)
return of_property_read_bool(to_of_node(fwnode), propname);
else if (is_acpi_node(fwnode))
return !acpi_node_prop_get(fwnode, propname, NULL);
-
- return !!pset_prop_get(to_pset(fwnode), propname);
+ else if (is_pset(fwnode))
+ return !!pset_prop_get(to_pset(fwnode), propname);
+ return false;
}
EXPORT_SYMBOL_GPL(fwnode_property_present);
@@ -494,9 +495,10 @@ int fwnode_property_read_string(struct fwnode_handle *fwnode,
else if (is_acpi_node(fwnode))
return acpi_node_prop_read(fwnode, propname, DEV_PROP_STRING,
val, 1);
-
- return pset_prop_read_array(to_pset(fwnode), propname,
- DEV_PROP_STRING, val, 1);
+ else if (is_pset(fwnode))
+ return pset_prop_read_array(to_pset(fwnode), propname,
+ DEV_PROP_STRING, val, 1);
+ return -ENXIO;
}
EXPORT_SYMBOL_GPL(fwnode_property_read_string);
--
2.6.2
To be in align with the rest of fwnode types we rename the built-in property
set ones, i.e.
is_pset() -> is_pset_node()
to_pset() -> to_pset_node()
There is no functional change.
Signed-off-by: Andy Shevchenko <[email protected]>
---
drivers/base/property.c | 22 +++++++++++-----------
1 file changed, 11 insertions(+), 11 deletions(-)
diff --git a/drivers/base/property.c b/drivers/base/property.c
index 09e488d..2e01f3f 100644
--- a/drivers/base/property.c
+++ b/drivers/base/property.c
@@ -37,14 +37,14 @@ void device_add_property_set(struct device *dev, struct property_set *pset)
}
EXPORT_SYMBOL_GPL(device_add_property_set);
-static inline bool is_pset(struct fwnode_handle *fwnode)
+static inline bool is_pset_node(struct fwnode_handle *fwnode)
{
return fwnode && fwnode->type == FWNODE_PDATA;
}
-static inline struct property_set *to_pset(struct fwnode_handle *fwnode)
+static inline struct property_set *to_pset_node(struct fwnode_handle *fwnode)
{
- return is_pset(fwnode) ?
+ return is_pset_node(fwnode) ?
container_of(fwnode, struct property_set, fwnode) : NULL;
}
@@ -135,8 +135,8 @@ bool fwnode_property_present(struct fwnode_handle *fwnode, const char *propname)
return of_property_read_bool(to_of_node(fwnode), propname);
else if (is_acpi_node(fwnode))
return !acpi_node_prop_get(fwnode, propname, NULL);
- else if (is_pset(fwnode))
- return !!pset_prop_get(to_pset(fwnode), propname);
+ else if (is_pset_node(fwnode))
+ return !!pset_prop_get(to_pset_node(fwnode), propname);
return false;
}
EXPORT_SYMBOL_GPL(fwnode_property_present);
@@ -323,8 +323,8 @@ EXPORT_SYMBOL_GPL(device_property_match_string);
else if (is_acpi_node(_fwnode_)) \
_ret_ = acpi_node_prop_read(_fwnode_, _propname_, _proptype_, \
_val_, _nval_); \
- else if (is_pset(_fwnode_)) \
- _ret_ = pset_prop_read_array(to_pset(_fwnode_), _propname_, \
+ else if (is_pset_node(_fwnode_)) \
+ _ret_ = pset_prop_read_array(to_pset_node(_fwnode_), _propname_, \
_proptype_, _val_, _nval_); \
else \
_ret_ = -ENXIO; \
@@ -465,8 +465,8 @@ int fwnode_property_read_string_array(struct fwnode_handle *fwnode,
else if (is_acpi_node(fwnode))
return acpi_node_prop_read(fwnode, propname, DEV_PROP_STRING,
val, nval);
- else if (is_pset(fwnode))
- return pset_prop_read_array(to_pset(fwnode), propname,
+ else if (is_pset_node(fwnode))
+ return pset_prop_read_array(to_pset_node(fwnode), propname,
DEV_PROP_STRING, val, nval);
return -ENXIO;
}
@@ -495,8 +495,8 @@ int fwnode_property_read_string(struct fwnode_handle *fwnode,
else if (is_acpi_node(fwnode))
return acpi_node_prop_read(fwnode, propname, DEV_PROP_STRING,
val, 1);
- else if (is_pset(fwnode))
- return pset_prop_read_array(to_pset(fwnode), propname,
+ else if (is_pset_node(fwnode))
+ return pset_prop_read_array(to_pset_node(fwnode), propname,
DEV_PROP_STRING, val, 1);
return -ENXIO;
}
--
2.6.2
Instead of using the type and nval fields we will use length (in bytes) of the
value. The sanity check is done in the accessors.
The built-in property accessors are split in the same way such as device tree.
Signed-off-by: Andy Shevchenko <[email protected]>
---
drivers/base/property.c | 150 ++++++++++++++++++++++++++++++++++-------------
include/linux/property.h | 8 +--
2 files changed, 113 insertions(+), 45 deletions(-)
diff --git a/drivers/base/property.c b/drivers/base/property.c
index 2e01f3f..86834bd 100644
--- a/drivers/base/property.c
+++ b/drivers/base/property.c
@@ -63,45 +63,107 @@ static struct property_entry *pset_prop_get(struct property_set *pset,
return NULL;
}
-static int pset_prop_read_array(struct property_set *pset, const char *name,
- enum dev_prop_type type, void *val, size_t nval)
+static void *pset_prop_find(struct property_set *pset, const char *propname,
+ size_t length)
{
struct property_entry *prop;
- unsigned int item_size;
+ void *pointer;
- prop = pset_prop_get(pset, name);
+ prop = pset_prop_get(pset, propname);
+ if (!prop)
+ return ERR_PTR(-EINVAL);
+ pointer = prop->value.raw_data;
+ if (!pointer)
+ return ERR_PTR(-ENODATA);
+ if (length > prop->length)
+ return ERR_PTR(-EOVERFLOW);
+ return pointer;
+}
+
+static int pset_prop_read_u8_array(struct property_set *pset,
+ const char *propname,
+ u8 *values, size_t nval)
+{
+ void *pointer;
+ size_t length = nval * sizeof(*values);
+
+ pointer = pset_prop_find(pset, propname, length);
+ if (IS_ERR(pointer))
+ return PTR_ERR(pointer);
+
+ memcpy(values, pointer, length);
+ return 0;
+}
+
+static int pset_prop_read_u16_array(struct property_set *pset,
+ const char *propname,
+ u16 *values, size_t nval)
+{
+ void *pointer;
+ size_t length = nval * sizeof(*values);
+
+ pointer = pset_prop_find(pset, propname, length);
+ if (IS_ERR(pointer))
+ return PTR_ERR(pointer);
+
+ memcpy(values, pointer, length);
+ return 0;
+}
+
+static int pset_prop_read_u32_array(struct property_set *pset,
+ const char *propname,
+ u32 *values, size_t nval)
+{
+ void *pointer;
+ size_t length = nval * sizeof(*values);
+
+ pointer = pset_prop_find(pset, propname, length);
+ if (IS_ERR(pointer))
+ return PTR_ERR(pointer);
+
+ memcpy(values, pointer, length);
+ return 0;
+}
+
+static int pset_prop_read_u64_array(struct property_set *pset,
+ const char *propname,
+ u64 *values, size_t nval)
+{
+ void *pointer;
+ size_t length = nval * sizeof(*values);
+
+ pointer = pset_prop_find(pset, propname, length);
+ if (IS_ERR(pointer))
+ return PTR_ERR(pointer);
+
+ memcpy(values, pointer, length);
+ return 0;
+}
+
+static int pset_prop_count_elems_of_size(struct property_set *pset,
+ const char *propname, size_t length)
+{
+ struct property_entry *prop;
+
+ prop = pset_prop_get(pset, propname);
if (!prop)
- return -ENODATA;
-
- if (prop->type != type)
- return -EPROTO;
-
- if (!val)
- return prop->nval;
-
- if (prop->nval < nval)
- return -EOVERFLOW;
-
- switch (type) {
- case DEV_PROP_U8:
- item_size = sizeof(u8);
- break;
- case DEV_PROP_U16:
- item_size = sizeof(u16);
- break;
- case DEV_PROP_U32:
- item_size = sizeof(u32);
- break;
- case DEV_PROP_U64:
- item_size = sizeof(u64);
- break;
- case DEV_PROP_STRING:
- item_size = sizeof(const char *);
- break;
- default:
return -EINVAL;
- }
- memcpy(val, prop->value.raw_data, nval * item_size);
+
+ return prop->length / length;
+}
+
+static int pset_prop_read_string_array(struct property_set *pset,
+ const char *propname,
+ const char **strings, size_t nval)
+{
+ void *pointer;
+ size_t length = nval * sizeof(*strings);
+
+ pointer = pset_prop_find(pset, propname, length);
+ if (IS_ERR(pointer))
+ return PTR_ERR(pointer);
+
+ memcpy(strings, pointer, length);
return 0;
}
@@ -314,6 +376,10 @@ EXPORT_SYMBOL_GPL(device_property_match_string);
(val) ? of_property_read_##type##_array((node), (propname), (val), (nval)) \
: of_property_count_elems_of_size((node), (propname), sizeof(type))
+#define PSET_PROP_READ_ARRAY(node, propname, type, val, nval) \
+ (val) ? pset_prop_read_##type##_array((node), (propname), (val), (nval)) \
+ : pset_prop_count_elems_of_size((node), (propname), sizeof(type))
+
#define FWNODE_PROP_READ_ARRAY(_fwnode_, _propname_, _type_, _proptype_, _val_, _nval_) \
({ \
int _ret_; \
@@ -324,8 +390,8 @@ EXPORT_SYMBOL_GPL(device_property_match_string);
_ret_ = acpi_node_prop_read(_fwnode_, _propname_, _proptype_, \
_val_, _nval_); \
else if (is_pset_node(_fwnode_)) \
- _ret_ = pset_prop_read_array(to_pset_node(_fwnode_), _propname_, \
- _proptype_, _val_, _nval_); \
+ _ret_ = PSET_PROP_READ_ARRAY(to_pset_node(_fwnode_), _propname_, \
+ _type_, _val_, _nval_); \
else \
_ret_ = -ENXIO; \
_ret_; \
@@ -466,8 +532,12 @@ int fwnode_property_read_string_array(struct fwnode_handle *fwnode,
return acpi_node_prop_read(fwnode, propname, DEV_PROP_STRING,
val, nval);
else if (is_pset_node(fwnode))
- return pset_prop_read_array(to_pset_node(fwnode), propname,
- DEV_PROP_STRING, val, nval);
+ return val ?
+ pset_prop_read_string_array(to_pset_node(fwnode),
+ propname, val, nval) :
+ pset_prop_count_elems_of_size(to_pset_node(fwnode),
+ propname,
+ sizeof(const char *));
return -ENXIO;
}
EXPORT_SYMBOL_GPL(fwnode_property_read_string_array);
@@ -496,8 +566,8 @@ int fwnode_property_read_string(struct fwnode_handle *fwnode,
return acpi_node_prop_read(fwnode, propname, DEV_PROP_STRING,
val, 1);
else if (is_pset_node(fwnode))
- return pset_prop_read_array(to_pset_node(fwnode), propname,
- DEV_PROP_STRING, val, 1);
+ return pset_prop_read_string_array(to_pset_node(fwnode),
+ propname, val, 1);
return -ENXIO;
}
EXPORT_SYMBOL_GPL(fwnode_property_read_string);
diff --git a/include/linux/property.h b/include/linux/property.h
index 0a3705a..c29460a 100644
--- a/include/linux/property.h
+++ b/include/linux/property.h
@@ -144,14 +144,12 @@ static inline int fwnode_property_read_u64(struct fwnode_handle *fwnode,
/**
* struct property_entry - "Built-in" device property representation.
* @name: Name of the property.
- * @type: Type of the property.
- * @nval: Number of items of type @type making up the value.
- * @value: Value of the property (an array of @nval items of type @type).
+ * @length: Length of data making up the value.
+ * @value: Value of the property (an array of items of the given type).
*/
struct property_entry {
const char *name;
- enum dev_prop_type type;
- size_t nval;
+ size_t length;
union {
void *raw_data;
u8 *u8_data;
--
2.6.2
We may save a lot of lines of code and space by keeping single values inside
the struct property_entry. Refactor the implementation to do so.
Signed-off-by: Andy Shevchenko <[email protected]>
---
drivers/base/property.c | 31 ++++++++++++++++++++++++++++---
include/linux/property.h | 29 +++++++++++++++++++++--------
2 files changed, 49 insertions(+), 11 deletions(-)
diff --git a/drivers/base/property.c b/drivers/base/property.c
index 86834bd..3e603c0 100644
--- a/drivers/base/property.c
+++ b/drivers/base/property.c
@@ -72,7 +72,10 @@ static void *pset_prop_find(struct property_set *pset, const char *propname,
prop = pset_prop_get(pset, propname);
if (!prop)
return ERR_PTR(-EINVAL);
- pointer = prop->value.raw_data;
+ if (prop->is_array)
+ pointer = prop->pointer.raw_data;
+ else
+ pointer = &prop->value.raw_data;
if (!pointer)
return ERR_PTR(-ENODATA);
if (length > prop->length)
@@ -167,6 +170,29 @@ static int pset_prop_read_string_array(struct property_set *pset,
return 0;
}
+static int pset_prop_read_string(struct property_set *pset,
+ const char *propname, const char **strings)
+{
+ struct property_entry *prop;
+ const char **pointer;
+
+ prop = pset_prop_get(pset, propname);
+ if (!prop)
+ return -EINVAL;
+ if (prop->is_array) {
+ pointer = prop->pointer.str;
+ if (!pointer)
+ return -ENODATA;
+ } else {
+ pointer = &prop->value.str;
+ if (strnlen(*pointer, prop->length) >= prop->length)
+ return -EILSEQ;
+ }
+
+ *strings = *pointer;
+ return 0;
+}
+
static inline struct fwnode_handle *dev_fwnode(struct device *dev)
{
return IS_ENABLED(CONFIG_OF) && dev->of_node ?
@@ -566,8 +592,7 @@ int fwnode_property_read_string(struct fwnode_handle *fwnode,
return acpi_node_prop_read(fwnode, propname, DEV_PROP_STRING,
val, 1);
else if (is_pset_node(fwnode))
- return pset_prop_read_string_array(to_pset_node(fwnode),
- propname, val, 1);
+ return pset_prop_read_string(to_pset_node(fwnode), propname, val);
return -ENXIO;
}
EXPORT_SYMBOL_GPL(fwnode_property_read_string);
diff --git a/include/linux/property.h b/include/linux/property.h
index c29460a..5d0b9b6 100644
--- a/include/linux/property.h
+++ b/include/linux/property.h
@@ -145,19 +145,32 @@ static inline int fwnode_property_read_u64(struct fwnode_handle *fwnode,
* struct property_entry - "Built-in" device property representation.
* @name: Name of the property.
* @length: Length of data making up the value.
- * @value: Value of the property (an array of items of the given type).
+ * @is_array: True when the property is an array.
+ * @pointer: Pointer to the property (an array of items of the given type).
+ * @value: Value of the property (when it is a single item of the given type).
*/
struct property_entry {
const char *name;
size_t length;
+ bool is_array;
union {
- void *raw_data;
- u8 *u8_data;
- u16 *u16_data;
- u32 *u32_data;
- u64 *u64_data;
- const char **str;
- } value;
+ union {
+ void *raw_data;
+ u8 *u8_data;
+ u16 *u16_data;
+ u32 *u32_data;
+ u64 *u64_data;
+ const char **str;
+ } pointer;
+ union {
+ unsigned long long raw_data;
+ u8 u8_data;
+ u16 u16_data;
+ u32 u32_data;
+ u64 u64_data;
+ const char *str;
+ } value;
+ };
};
/**
--
2.6.2
From: Heikki Krogerus <[email protected]>
Marcos for easier creation of build-in property entries.
Signed-off-by: Heikki Krogerus <[email protected]>
Signed-off-by: Andy Shevchenko <[email protected]>
---
include/linux/property.h | 48 ++++++++++++++++++++++++++++++++++++++++++++++++
1 file changed, 48 insertions(+)
diff --git a/include/linux/property.h b/include/linux/property.h
index 5d0b9b6..2528a000 100644
--- a/include/linux/property.h
+++ b/include/linux/property.h
@@ -173,6 +173,54 @@ struct property_entry {
};
};
+#define PROPERTY_ENTRY_INTEGER_ARRAY(_name_, _type_, _val_) \
+{ \
+ .name = _name_, \
+ .length = ARRAY_SIZE(_val_) * sizeof(_type_), \
+ .is_array = true, \
+ .pointer._type_##_data = _val_, \
+}
+
+#define PROPERTY_ENTRY_U8_ARRAY(_name_, _val_) \
+ PROPERTY_ENTRY_INTEGER_ARRAY(_name_, u8, _val_)
+#define PROPERTY_ENTRY_U16_ARRAY(_name_, _val_) \
+ PROPERTY_ENTRY_INTEGER_ARRAY(_name_, u16, _val_)
+#define PROPERTY_ENTRY_U32_ARRAY(_name_, _val_) \
+ PROPERTY_ENTRY_INTEGER_ARRAY(_name_, u32, _val_)
+#define PROPERTY_ENTRY_U64_ARRAY(_name_, _val_) \
+ PROPERTY_ENTRY_INTEGER_ARRAY(_name_, u64, _val_)
+
+#define PROPERTY_ENTRY_STRING_ARRAY(_name_, _val_) \
+{ \
+ .name = _name_, \
+ .length = ARRAY_SIZE(_val_) * sizeof(const char *), \
+ .is_array = true, \
+ .pointer.str = _val_, \
+}
+
+#define PROPERTY_ENTRY_INTEGER(_name_, _type_, _val_) \
+{ \
+ .name = _name_, \
+ .length = sizeof(_type_), \
+ .value._type_##_data = _val_, \
+}
+
+#define PROPERTY_ENTRY_U8(_name_, _val_) \
+ PROPERTY_ENTRY_INTEGER(_name_, u8, _val_)
+#define PROPERTY_ENTRY_U16(_name_, _val_) \
+ PROPERTY_ENTRY_INTEGER(_name_, u16, _val_)
+#define PROPERTY_ENTRY_U32(_name_, _val_) \
+ PROPERTY_ENTRY_INTEGER(_name_, u32, _val_)
+#define PROPERTY_ENTRY_U64(_name_, _val_) \
+ PROPERTY_ENTRY_INTEGER(_name_, u64, _val_)
+
+#define PROPERTY_ENTRY_STRING(_name_, _val_) \
+{ \
+ .name = _name_, \
+ .length = sizeof(_val_), \
+ .value.str = _val_, \
+}
+
/**
* struct property_set - Collection of "built-in" device properties.
* @fwnode: Handle to be pointed to by the fwnode field of struct device.
--
2.6.2
There is no functional change.
Signed-off-by: Andy Shevchenko <[email protected]>
---
drivers/base/property.c | 28 ++++++++++++++--------------
include/linux/property.h | 4 ++--
2 files changed, 16 insertions(+), 16 deletions(-)
diff --git a/drivers/base/property.c b/drivers/base/property.c
index 3e603c0..c8e9722 100644
--- a/drivers/base/property.c
+++ b/drivers/base/property.c
@@ -398,29 +398,29 @@ int device_property_match_string(struct device *dev, const char *propname,
}
EXPORT_SYMBOL_GPL(device_property_match_string);
-#define OF_DEV_PROP_READ_ARRAY(node, propname, type, val, nval) \
- (val) ? of_property_read_##type##_array((node), (propname), (val), (nval)) \
+#define OF_DEV_PROP_READ_ARRAY(node, propname, type, val, nval) \
+ (val) ? of_property_read_##type##_array((node), (propname), (val), (nval)) \
: of_property_count_elems_of_size((node), (propname), sizeof(type))
#define PSET_PROP_READ_ARRAY(node, propname, type, val, nval) \
(val) ? pset_prop_read_##type##_array((node), (propname), (val), (nval)) \
: pset_prop_count_elems_of_size((node), (propname), sizeof(type))
-#define FWNODE_PROP_READ_ARRAY(_fwnode_, _propname_, _type_, _proptype_, _val_, _nval_) \
-({ \
- int _ret_; \
- if (is_of_node(_fwnode_)) \
- _ret_ = OF_DEV_PROP_READ_ARRAY(to_of_node(_fwnode_), _propname_, \
- _type_, _val_, _nval_); \
- else if (is_acpi_node(_fwnode_)) \
- _ret_ = acpi_node_prop_read(_fwnode_, _propname_, _proptype_, \
- _val_, _nval_); \
+#define FWNODE_PROP_READ_ARRAY(_fwnode_, _propname_, _type_, _proptype_, _val_, _nval_) \
+({ \
+ int _ret_; \
+ if (is_of_node(_fwnode_)) \
+ _ret_ = OF_DEV_PROP_READ_ARRAY(to_of_node(_fwnode_), _propname_, \
+ _type_, _val_, _nval_); \
+ else if (is_acpi_node(_fwnode_)) \
+ _ret_ = acpi_node_prop_read(_fwnode_, _propname_, _proptype_, \
+ _val_, _nval_); \
else if (is_pset_node(_fwnode_)) \
_ret_ = PSET_PROP_READ_ARRAY(to_pset_node(_fwnode_), _propname_, \
_type_, _val_, _nval_); \
- else \
- _ret_ = -ENXIO; \
- _ret_; \
+ else \
+ _ret_ = -ENXIO; \
+ _ret_; \
})
/**
diff --git a/include/linux/property.h b/include/linux/property.h
index 2528a000..e947403 100644
--- a/include/linux/property.h
+++ b/include/linux/property.h
@@ -73,8 +73,8 @@ int fwnode_property_match_string(struct fwnode_handle *fwnode,
struct fwnode_handle *device_get_next_child_node(struct device *dev,
struct fwnode_handle *child);
-#define device_for_each_child_node(dev, child) \
- for (child = device_get_next_child_node(dev, NULL); child; \
+#define device_for_each_child_node(dev, child) \
+ for (child = device_get_next_child_node(dev, NULL); child; \
child = device_get_next_child_node(dev, child))
void fwnode_handle_put(struct fwnode_handle *fwnode);
--
2.6.2
Change return code to be in align with OF and built-in device properties error
codes. In particular -EINVAL means property is not found.
Signed-off-by: Andy Shevchenko <[email protected]>
---
drivers/acpi/property.c | 10 +++++-----
1 file changed, 5 insertions(+), 5 deletions(-)
diff --git a/drivers/acpi/property.c b/drivers/acpi/property.c
index 88f4306..2aee416 100644
--- a/drivers/acpi/property.c
+++ b/drivers/acpi/property.c
@@ -346,7 +346,7 @@ void acpi_free_properties(struct acpi_device *adev)
*
* Return: %0 if property with @name has been found (success),
* %-EINVAL if the arguments are invalid,
- * %-ENODATA if the property doesn't exist,
+ * %-EINVAL if the property doesn't exist,
* %-EPROTO if the property value type doesn't match @type.
*/
static int acpi_data_get_property(struct acpi_device_data *data,
@@ -360,7 +360,7 @@ static int acpi_data_get_property(struct acpi_device_data *data,
return -EINVAL;
if (!data->pointer || !data->properties)
- return -ENODATA;
+ return -EINVAL;
properties = data->properties;
for (i = 0; i < properties->package.count; i++) {
@@ -375,13 +375,13 @@ static int acpi_data_get_property(struct acpi_device_data *data,
if (!strcmp(name, propname->string.pointer)) {
if (type != ACPI_TYPE_ANY && propvalue->type != type)
return -EPROTO;
- else if (obj)
+ if (obj)
*obj = propvalue;
return 0;
}
}
- return -ENODATA;
+ return -EINVAL;
}
/**
@@ -439,7 +439,7 @@ int acpi_node_prop_get(struct fwnode_handle *fwnode, const char *propname,
*
* Return: %0 if array property (package) with @name has been found (success),
* %-EINVAL if the arguments are invalid,
- * %-ENODATA if the property doesn't exist,
+ * %-EINVAL if the property doesn't exist,
* %-EPROTO if the property is not a package or the type of its elements
* doesn't match @type.
*/
--
2.6.2
The struct fwnode has notion of secondary fwnode. This is supposed to used
as fallback if the primary firmware interface (DT, ACPI) does not have the
property in question.
However, the current implementation never checks the secondary node which
prevents one to add default "built-in" properties to devices.
This patch adds fallback to the secondary fwnode if the primary fwnode
returns that the property does not exists.
Signed-off-by: Mika Westerberg <[email protected]>
Signed-off-by: Andy Shevchenko <[email protected]>
---
drivers/base/property.c | 109 ++++++++++++++++++++++++++++++++++--------------
1 file changed, 78 insertions(+), 31 deletions(-)
diff --git a/drivers/base/property.c b/drivers/base/property.c
index c8e9722..1768e61 100644
--- a/drivers/base/property.c
+++ b/drivers/base/property.c
@@ -212,12 +212,8 @@ bool device_property_present(struct device *dev, const char *propname)
}
EXPORT_SYMBOL_GPL(device_property_present);
-/**
- * fwnode_property_present - check if a property of a firmware node is present
- * @fwnode: Firmware node whose property to check
- * @propname: Name of the property
- */
-bool fwnode_property_present(struct fwnode_handle *fwnode, const char *propname)
+static bool __fwnode_property_present(struct fwnode_handle *fwnode,
+ const char *propname)
{
if (is_of_node(fwnode))
return of_property_read_bool(to_of_node(fwnode), propname);
@@ -227,6 +223,21 @@ bool fwnode_property_present(struct fwnode_handle *fwnode, const char *propname)
return !!pset_prop_get(to_pset_node(fwnode), propname);
return false;
}
+
+/**
+ * fwnode_property_present - check if a property of a firmware node is present
+ * @fwnode: Firmware node whose property to check
+ * @propname: Name of the property
+ */
+bool fwnode_property_present(struct fwnode_handle *fwnode, const char *propname)
+{
+ bool ret;
+
+ ret = __fwnode_property_present(fwnode, propname);
+ if (ret == false && fwnode->secondary)
+ ret = __fwnode_property_present(fwnode->secondary, propname);
+ return ret;
+}
EXPORT_SYMBOL_GPL(fwnode_property_present);
/**
@@ -406,7 +417,7 @@ EXPORT_SYMBOL_GPL(device_property_match_string);
(val) ? pset_prop_read_##type##_array((node), (propname), (val), (nval)) \
: pset_prop_count_elems_of_size((node), (propname), sizeof(type))
-#define FWNODE_PROP_READ_ARRAY(_fwnode_, _propname_, _type_, _proptype_, _val_, _nval_) \
+#define FWNODE_PROP_READ(_fwnode_, _propname_, _type_, _proptype_, _val_, _nval_) \
({ \
int _ret_; \
if (is_of_node(_fwnode_)) \
@@ -423,6 +434,17 @@ EXPORT_SYMBOL_GPL(device_property_match_string);
_ret_; \
})
+#define FWNODE_PROP_READ_ARRAY(_fwnode_, _propname_, _type_, _proptype_, _val_, _nval_) \
+({ \
+ int _ret_; \
+ _ret_ = FWNODE_PROP_READ(_fwnode_, _propname_, _type_, _proptype_, \
+ _val_, _nval_); \
+ if (_ret_ == -EINVAL && _fwnode_->secondary) \
+ _ret_ = FWNODE_PROP_READ(_fwnode_->secondary, _propname_, _type_, \
+ _proptype_, _val_, _nval_); \
+ _ret_; \
+})
+
/**
* fwnode_property_read_u8_array - return a u8 array property of firmware node
* @fwnode: Firmware node to get the property of
@@ -527,6 +549,41 @@ int fwnode_property_read_u64_array(struct fwnode_handle *fwnode,
}
EXPORT_SYMBOL_GPL(fwnode_property_read_u64_array);
+static int __fwnode_property_read_string_array(struct fwnode_handle *fwnode,
+ const char *propname,
+ const char **val, size_t nval)
+{
+ if (is_of_node(fwnode))
+ return val ?
+ of_property_read_string_array(to_of_node(fwnode),
+ propname, val, nval) :
+ of_property_count_strings(to_of_node(fwnode), propname);
+ else if (is_acpi_node(fwnode))
+ return acpi_node_prop_read(fwnode, propname, DEV_PROP_STRING,
+ val, nval);
+ else if (is_pset_node(fwnode))
+ return val ?
+ pset_prop_read_string_array(to_pset_node(fwnode),
+ propname, val, nval) :
+ pset_prop_count_elems_of_size(to_pset_node(fwnode),
+ propname,
+ sizeof(const char *));
+ return -ENXIO;
+}
+
+static int __fwnode_property_read_string(struct fwnode_handle *fwnode,
+ const char *propname, const char **val)
+{
+ if (is_of_node(fwnode))
+ return of_property_read_string(to_of_node(fwnode), propname, val);
+ else if (is_acpi_node(fwnode))
+ return acpi_node_prop_read(fwnode, propname, DEV_PROP_STRING,
+ val, 1);
+ else if (is_pset_node(fwnode))
+ return pset_prop_read_string(to_pset_node(fwnode), propname, val);
+ return -ENXIO;
+}
+
/**
* fwnode_property_read_string_array - return string array property of a node
* @fwnode: Firmware node to get the property of
@@ -549,22 +606,13 @@ int fwnode_property_read_string_array(struct fwnode_handle *fwnode,
const char *propname, const char **val,
size_t nval)
{
- if (is_of_node(fwnode))
- return val ?
- of_property_read_string_array(to_of_node(fwnode),
- propname, val, nval) :
- of_property_count_strings(to_of_node(fwnode), propname);
- else if (is_acpi_node(fwnode))
- return acpi_node_prop_read(fwnode, propname, DEV_PROP_STRING,
- val, nval);
- else if (is_pset_node(fwnode))
- return val ?
- pset_prop_read_string_array(to_pset_node(fwnode),
- propname, val, nval) :
- pset_prop_count_elems_of_size(to_pset_node(fwnode),
- propname,
- sizeof(const char *));
- return -ENXIO;
+ int ret;
+
+ ret = __fwnode_property_read_string_array(fwnode, propname, val, nval);
+ if (ret == -EINVAL && fwnode->secondary)
+ ret = __fwnode_property_read_string_array(fwnode->secondary,
+ propname, val, nval);
+ return ret;
}
EXPORT_SYMBOL_GPL(fwnode_property_read_string_array);
@@ -586,14 +634,13 @@ EXPORT_SYMBOL_GPL(fwnode_property_read_string_array);
int fwnode_property_read_string(struct fwnode_handle *fwnode,
const char *propname, const char **val)
{
- if (is_of_node(fwnode))
- return of_property_read_string(to_of_node(fwnode), propname, val);
- else if (is_acpi_node(fwnode))
- return acpi_node_prop_read(fwnode, propname, DEV_PROP_STRING,
- val, 1);
- else if (is_pset_node(fwnode))
- return pset_prop_read_string(to_pset_node(fwnode), propname, val);
- return -ENXIO;
+ int ret;
+
+ ret = __fwnode_property_read_string(fwnode, propname, val);
+ if (ret == -EINVAL && fwnode->secondary)
+ ret = __fwnode_property_read_string(fwnode->secondary,
+ propname, val);
+ return ret;
}
EXPORT_SYMBOL_GPL(fwnode_property_read_string);
--
2.6.2
In the similar way like we do for the platform data we propagate the device
properties. For example, in case of Intel LPSS drivers we may provide a
specific property to tell the actual device driver an additional information
such as platform name.
Signed-off-by: Andy Shevchenko <[email protected]>
---
drivers/mfd/mfd-core.c | 3 +++
include/linux/mfd/core.h | 5 +++++
2 files changed, 8 insertions(+)
diff --git a/drivers/mfd/mfd-core.c b/drivers/mfd/mfd-core.c
index 60b60dc..ec00f6f 100644
--- a/drivers/mfd/mfd-core.c
+++ b/drivers/mfd/mfd-core.c
@@ -14,6 +14,7 @@
#include <linux/kernel.h>
#include <linux/platform_device.h>
#include <linux/acpi.h>
+#include <linux/property.h>
#include <linux/mfd/core.h>
#include <linux/pm_runtime.h>
#include <linux/slab.h>
@@ -192,6 +193,8 @@ static int mfd_add_device(struct device *parent, int id,
goto fail_alias;
}
+ device_add_property_set(&pdev->dev, cell->pset);
+
ret = mfd_platform_add_cell(pdev, cell, usage_count);
if (ret)
goto fail_alias;
diff --git a/include/linux/mfd/core.h b/include/linux/mfd/core.h
index 27dac3f..0c62f17 100644
--- a/include/linux/mfd/core.h
+++ b/include/linux/mfd/core.h
@@ -15,6 +15,7 @@
#define MFD_CORE_H
#include <linux/platform_device.h>
+#include <linux/property.h>
struct irq_domain;
@@ -44,6 +45,10 @@ struct mfd_cell {
/* platform data passed to the sub devices drivers */
void *platform_data;
size_t pdata_size;
+
+ /* device properties passed to the sub devices drivers */
+ struct property_set *pset;
+
/*
* Device Tree compatible string
* See: Documentation/devicetree/usage-model.txt Chapter 2.2 for details
--
2.6.2
From: Mika Westerberg <[email protected]>
If the boot firmware does not support ACPI we need a way to pass device
configuration information to the drivers. The unified device properties API
already supports passing platform data via properties so let's take
advantage of that and allow probe drivers to pass set of properties to the
host controller driver.
In order to do that we need to be able to modify the MFD cell corresponding
the host controller, so make the core driver to take copy of the cell
instead of using it directly. Then we can assign info->pset to the
resulting copy of a cell and let the MFD core to assign that to the
resulting device.
Signed-off-by: Mika Westerberg <[email protected]>
Signed-off-by: Andy Shevchenko <[email protected]>
---
drivers/mfd/intel-lpss.c | 16 ++++++++++++----
drivers/mfd/intel-lpss.h | 2 ++
2 files changed, 14 insertions(+), 4 deletions(-)
diff --git a/drivers/mfd/intel-lpss.c b/drivers/mfd/intel-lpss.c
index 6255513..1743788 100644
--- a/drivers/mfd/intel-lpss.c
+++ b/drivers/mfd/intel-lpss.c
@@ -24,6 +24,7 @@
#include <linux/mfd/core.h>
#include <linux/pm_qos.h>
#include <linux/pm_runtime.h>
+#include <linux/property.h>
#include <linux/seq_file.h>
#include <linux/io-64-nonatomic-lo-hi.h>
@@ -72,7 +73,7 @@ struct intel_lpss {
enum intel_lpss_dev_type type;
struct clk *clk;
struct clk_lookup *clock;
- const struct mfd_cell *cell;
+ struct mfd_cell *cell;
struct device *dev;
void __iomem *priv;
int devid;
@@ -217,6 +218,7 @@ static void intel_lpss_ltr_hide(struct intel_lpss *lpss)
static int intel_lpss_assign_devs(struct intel_lpss *lpss)
{
+ const struct mfd_cell *cell;
unsigned int type;
type = lpss->caps & LPSS_PRIV_CAPS_TYPE_MASK;
@@ -224,18 +226,22 @@ static int intel_lpss_assign_devs(struct intel_lpss *lpss)
switch (type) {
case LPSS_DEV_I2C:
- lpss->cell = &intel_lpss_i2c_cell;
+ cell = &intel_lpss_i2c_cell;
break;
case LPSS_DEV_UART:
- lpss->cell = &intel_lpss_uart_cell;
+ cell = &intel_lpss_uart_cell;
break;
case LPSS_DEV_SPI:
- lpss->cell = &intel_lpss_spi_cell;
+ cell = &intel_lpss_spi_cell;
break;
default:
return -ENODEV;
}
+ lpss->cell = devm_kmemdup(lpss->dev, cell, sizeof(*cell), GFP_KERNEL);
+ if (!lpss->cell)
+ return -ENOMEM;
+
lpss->type = type;
return 0;
@@ -401,6 +407,8 @@ int intel_lpss_probe(struct device *dev,
if (ret)
return ret;
+ lpss->cell->pset = info->pset;
+
intel_lpss_init_dev(lpss);
lpss->devid = ida_simple_get(&intel_lpss_devid_ida, 0, 0, GFP_KERNEL);
diff --git a/drivers/mfd/intel-lpss.h b/drivers/mfd/intel-lpss.h
index 2c7f8d7..0dcea9e 100644
--- a/drivers/mfd/intel-lpss.h
+++ b/drivers/mfd/intel-lpss.h
@@ -16,12 +16,14 @@
struct device;
struct resource;
+struct property_set;
struct intel_lpss_platform_info {
struct resource *mem;
int irq;
unsigned long clk_rate;
const char *clk_con_id;
+ struct property_set *pset;
};
int intel_lpss_probe(struct device *dev,
--
2.6.2
From: Mika Westerberg <[email protected]>
The HS-UART host controller driver needs to know certain properties like
width of the register set if it cannot get that information from ACPI or
DT. In order to support non-ACPI systems we pass this information to the
driver via device properties.
Signed-off-by: Mika Westerberg <[email protected]>
Signed-off-by: Andy Shevchenko <[email protected]>
---
drivers/mfd/intel-lpss-pci.c | 13 +++++++++++++
1 file changed, 13 insertions(+)
diff --git a/drivers/mfd/intel-lpss-pci.c b/drivers/mfd/intel-lpss-pci.c
index 5bfdfcc..74f0d6d 100644
--- a/drivers/mfd/intel-lpss-pci.c
+++ b/drivers/mfd/intel-lpss-pci.c
@@ -65,9 +65,21 @@ static const struct intel_lpss_platform_info spt_info = {
.clk_rate = 120000000,
};
+static struct property_entry uart_properties[] = {
+ PROPERTY_ENTRY_U32("reg-io-width", 4),
+ PROPERTY_ENTRY_U32("reg-shift", 2),
+ PROPERTY_ENTRY_U8("snps,uart-16550-compatible", 1),
+ { },
+};
+
+static struct property_set uart_pset = {
+ .properties = uart_properties,
+};
+
static const struct intel_lpss_platform_info spt_uart_info = {
.clk_rate = 120000000,
.clk_con_id = "baudclk",
+ .pset = &uart_pset,
};
static const struct intel_lpss_platform_info bxt_info = {
@@ -77,6 +89,7 @@ static const struct intel_lpss_platform_info bxt_info = {
static const struct intel_lpss_platform_info bxt_uart_info = {
.clk_rate = 100000000,
.clk_con_id = "baudclk",
+ .pset = &uart_pset,
};
static const struct intel_lpss_platform_info bxt_i2c_info = {
--
2.6.2
From: Mika Westerberg <[email protected]>
Intel Skylake the LPSS I2C pad circuit has internal delays that require
programming non-zero SDA hold time for the I2C host controller. If this is
not done communication to slave devices may fail with arbitration lost
errors like the one seen below taken from Lenovo Yoga 900:
i2c_hid i2c-SYNA2B29:00: Fetching the HID descriptor
i2c_hid i2c-SYNA2B29:00: __i2c_hid_command: cmd=20 00
i2c_designware i2c_designware.1: i2c_dw_handle_tx_abort: lost arbitration
To fix this we follow what the Windows driver is doing and pass the default
SDA hold time of 230 ns to all Intel Skylake host controllers. This still
allows the platform to override these values by passing special ACPI
methods SSCN and FMCN.
Reported-by: Kevin Fenzi <[email protected]>
Signed-off-by: Mika Westerberg <[email protected]>
Signed-off-by: Andy Shevchenko <[email protected]>
---
drivers/mfd/intel-lpss-acpi.c | 18 ++++++++++++++++--
drivers/mfd/intel-lpss-pci.c | 30 ++++++++++++++++++++++--------
2 files changed, 38 insertions(+), 10 deletions(-)
diff --git a/drivers/mfd/intel-lpss-acpi.c b/drivers/mfd/intel-lpss-acpi.c
index b6fd904..6c5bbda 100644
--- a/drivers/mfd/intel-lpss-acpi.c
+++ b/drivers/mfd/intel-lpss-acpi.c
@@ -25,6 +25,20 @@ static const struct intel_lpss_platform_info spt_info = {
.clk_rate = 120000000,
};
+static struct property_entry spt_i2c_properties[] = {
+ PROPERTY_ENTRY_U32("i2c-sda-hold-time-ns", 230),
+ { },
+};
+
+static struct property_set spt_i2c_pset = {
+ .properties = spt_i2c_properties,
+};
+
+static const struct intel_lpss_platform_info spt_i2c_info = {
+ .clk_rate = 120000000,
+ .pset = &spt_i2c_pset,
+};
+
static const struct intel_lpss_platform_info bxt_info = {
.clk_rate = 100000000,
};
@@ -35,8 +49,8 @@ static const struct intel_lpss_platform_info bxt_i2c_info = {
static const struct acpi_device_id intel_lpss_acpi_ids[] = {
/* SPT */
- { "INT3446", (kernel_ulong_t)&spt_info },
- { "INT3447", (kernel_ulong_t)&spt_info },
+ { "INT3446", (kernel_ulong_t)&spt_i2c_info },
+ { "INT3447", (kernel_ulong_t)&spt_i2c_info },
/* BXT */
{ "80860AAC", (kernel_ulong_t)&bxt_i2c_info },
{ "80860ABC", (kernel_ulong_t)&bxt_info },
diff --git a/drivers/mfd/intel-lpss-pci.c b/drivers/mfd/intel-lpss-pci.c
index 74f0d6d..3c84ceb 100644
--- a/drivers/mfd/intel-lpss-pci.c
+++ b/drivers/mfd/intel-lpss-pci.c
@@ -65,6 +65,20 @@ static const struct intel_lpss_platform_info spt_info = {
.clk_rate = 120000000,
};
+static struct property_entry spt_i2c_properties[] = {
+ PROPERTY_ENTRY_U32("i2c-sda-hold-time-ns", 230),
+ { },
+};
+
+static struct property_set spt_i2c_pset = {
+ .properties = spt_i2c_properties,
+};
+
+static const struct intel_lpss_platform_info spt_i2c_info = {
+ .clk_rate = 120000000,
+ .pset = &spt_i2c_pset,
+};
+
static struct property_entry uart_properties[] = {
PROPERTY_ENTRY_U32("reg-io-width", 4),
PROPERTY_ENTRY_U32("reg-shift", 2),
@@ -134,20 +148,20 @@ static const struct pci_device_id intel_lpss_pci_ids[] = {
{ PCI_VDEVICE(INTEL, 0x9d28), (kernel_ulong_t)&spt_uart_info },
{ PCI_VDEVICE(INTEL, 0x9d29), (kernel_ulong_t)&spt_info },
{ PCI_VDEVICE(INTEL, 0x9d2a), (kernel_ulong_t)&spt_info },
- { PCI_VDEVICE(INTEL, 0x9d60), (kernel_ulong_t)&spt_info },
- { PCI_VDEVICE(INTEL, 0x9d61), (kernel_ulong_t)&spt_info },
- { PCI_VDEVICE(INTEL, 0x9d62), (kernel_ulong_t)&spt_info },
- { PCI_VDEVICE(INTEL, 0x9d63), (kernel_ulong_t)&spt_info },
- { PCI_VDEVICE(INTEL, 0x9d64), (kernel_ulong_t)&spt_info },
- { PCI_VDEVICE(INTEL, 0x9d65), (kernel_ulong_t)&spt_info },
+ { PCI_VDEVICE(INTEL, 0x9d60), (kernel_ulong_t)&spt_i2c_info },
+ { PCI_VDEVICE(INTEL, 0x9d61), (kernel_ulong_t)&spt_i2c_info },
+ { PCI_VDEVICE(INTEL, 0x9d62), (kernel_ulong_t)&spt_i2c_info },
+ { PCI_VDEVICE(INTEL, 0x9d63), (kernel_ulong_t)&spt_i2c_info },
+ { PCI_VDEVICE(INTEL, 0x9d64), (kernel_ulong_t)&spt_i2c_info },
+ { PCI_VDEVICE(INTEL, 0x9d65), (kernel_ulong_t)&spt_i2c_info },
{ PCI_VDEVICE(INTEL, 0x9d66), (kernel_ulong_t)&spt_uart_info },
/* SPT-H */
{ PCI_VDEVICE(INTEL, 0xa127), (kernel_ulong_t)&spt_uart_info },
{ PCI_VDEVICE(INTEL, 0xa128), (kernel_ulong_t)&spt_uart_info },
{ PCI_VDEVICE(INTEL, 0xa129), (kernel_ulong_t)&spt_info },
{ PCI_VDEVICE(INTEL, 0xa12a), (kernel_ulong_t)&spt_info },
- { PCI_VDEVICE(INTEL, 0xa160), (kernel_ulong_t)&spt_info },
- { PCI_VDEVICE(INTEL, 0xa161), (kernel_ulong_t)&spt_info },
+ { PCI_VDEVICE(INTEL, 0xa160), (kernel_ulong_t)&spt_i2c_info },
+ { PCI_VDEVICE(INTEL, 0xa161), (kernel_ulong_t)&spt_i2c_info },
{ PCI_VDEVICE(INTEL, 0xa166), (kernel_ulong_t)&spt_uart_info },
{ }
};
--
2.6.2
From: Mika Westerberg <[email protected]>
With ACPI _DSD (introduced in ACPI v5.1) it is now possible to pass device
configuration information from ACPI in addition to DT. In order to support
this, convert the driver to use the unified device property accessors
instead of DT specific.
Change to ordering a bit so that we first try platform data and if that's
not available look from device properties. ACPI *CNT methods are then used
as last resort to override everything else.
Signed-off-by: Mika Westerberg <[email protected]>
Signed-off-by: Andy Shevchenko <[email protected]>
---
drivers/i2c/busses/i2c-designware-platdrv.c | 48 +++++++++++++----------------
1 file changed, 22 insertions(+), 26 deletions(-)
diff --git a/drivers/i2c/busses/i2c-designware-platdrv.c b/drivers/i2c/busses/i2c-designware-platdrv.c
index 809579e..e9062be 100644
--- a/drivers/i2c/busses/i2c-designware-platdrv.c
+++ b/drivers/i2c/busses/i2c-designware-platdrv.c
@@ -36,6 +36,7 @@
#include <linux/platform_device.h>
#include <linux/pm.h>
#include <linux/pm_runtime.h>
+#include <linux/property.h>
#include <linux/io.h>
#include <linux/slab.h>
#include <linux/acpi.h>
@@ -156,33 +157,28 @@ static int dw_i2c_plat_probe(struct platform_device *pdev)
/* fast mode by default because of legacy reasons */
clk_freq = 400000;
- if (has_acpi_companion(&pdev->dev)) {
- dw_i2c_acpi_configure(pdev);
- } else if (pdev->dev.of_node) {
- of_property_read_u32(pdev->dev.of_node,
- "i2c-sda-hold-time-ns", &ht);
-
- of_property_read_u32(pdev->dev.of_node,
- "i2c-sda-falling-time-ns",
- &dev->sda_falling_time);
- of_property_read_u32(pdev->dev.of_node,
- "i2c-scl-falling-time-ns",
- &dev->scl_falling_time);
-
- of_property_read_u32(pdev->dev.of_node, "clock-frequency",
- &clk_freq);
-
- /* Only standard mode at 100kHz and fast mode at 400kHz
- * are supported.
- */
- if (clk_freq != 100000 && clk_freq != 400000) {
- dev_err(&pdev->dev, "Only 100kHz and 400kHz supported");
- return -EINVAL;
- }
+ if ((pdata = dev_get_platdata(&pdev->dev))) {
+ clk_freq = pdata->i2c_scl_freq;
} else {
- pdata = dev_get_platdata(&pdev->dev);
- if (pdata)
- clk_freq = pdata->i2c_scl_freq;
+ device_property_read_u32(&pdev->dev, "i2c-sda-hold-time-ns",
+ &ht);
+ device_property_read_u32(&pdev->dev, "i2c-sda-falling-time-ns",
+ &dev->sda_falling_time);
+ device_property_read_u32(&pdev->dev, "i2c-scl-falling-time-ns",
+ &dev->scl_falling_time);
+ device_property_read_u32(&pdev->dev, "clock-frequency",
+ &clk_freq);
+ }
+
+ if (has_acpi_companion(&pdev->dev))
+ dw_i2c_acpi_configure(pdev);
+
+ /* Only standard mode at 100kHz and fast mode at 400kHz
+ * are supported.
+ */
+ if (clk_freq != 100000 && clk_freq != 400000) {
+ dev_err(&pdev->dev, "Only 100kHz and 400kHz supported");
+ return -EINVAL;
}
r = i2c_dw_eval_lock_support(dev);
--
2.6.2
Hi Mika,
[auto build test ERROR on v4.4-rc2]
[also build test ERROR on next-20151124]
[cannot apply to ljones-mfd/for-mfd-next]
url: https://github.com/0day-ci/linux/commits/Andy-Shevchenko/device-property-always-check-for-fwnode-type/20151124-183221
config: x86_64-randconfig-x019-11241713 (attached as .config)
reproduce:
# save the attached .config to linux build tree
make ARCH=x86_64
All error/warnings (new ones prefixed by >>):
>> drivers/mfd/intel-lpss-pci.c:68:30: error: array type has incomplete element type 'struct property_entry'
static struct property_entry uart_properties[] = {
^
>> drivers/mfd/intel-lpss-pci.c:69:2: error: implicit declaration of function 'PROPERTY_ENTRY_U32' [-Werror=implicit-function-declaration]
PROPERTY_ENTRY_U32("reg-io-width", 4),
^
>> drivers/mfd/intel-lpss-pci.c:71:2: error: implicit declaration of function 'PROPERTY_ENTRY_U8' [-Werror=implicit-function-declaration]
PROPERTY_ENTRY_U8("snps,uart-16550-compatible", 1),
^
>> drivers/mfd/intel-lpss-pci.c:75:15: error: variable 'uart_pset' has initializer but incomplete type
static struct property_set uart_pset = {
^
>> drivers/mfd/intel-lpss-pci.c:76:2: error: unknown field 'properties' specified in initializer
.properties = uart_properties,
^
>> drivers/mfd/intel-lpss-pci.c:76:16: warning: excess elements in struct initializer
.properties = uart_properties,
^
drivers/mfd/intel-lpss-pci.c:76:16: note: (near initialization for 'uart_pset')
>> drivers/mfd/intel-lpss-pci.c:68:30: warning: 'uart_properties' defined but not used [-Wunused-variable]
static struct property_entry uart_properties[] = {
^
cc1: some warnings being treated as errors
vim +68 drivers/mfd/intel-lpss-pci.c
62 static INTEL_LPSS_PM_OPS(intel_lpss_pci_pm_ops);
63
64 static const struct intel_lpss_platform_info spt_info = {
65 .clk_rate = 120000000,
66 };
67
> 68 static struct property_entry uart_properties[] = {
> 69 PROPERTY_ENTRY_U32("reg-io-width", 4),
70 PROPERTY_ENTRY_U32("reg-shift", 2),
> 71 PROPERTY_ENTRY_U8("snps,uart-16550-compatible", 1),
72 { },
73 };
74
> 75 static struct property_set uart_pset = {
> 76 .properties = uart_properties,
77 };
78
79 static const struct intel_lpss_platform_info spt_uart_info = {
---
0-DAY kernel test infrastructure Open Source Technology Center
https://lists.01.org/pipermail/kbuild-all Intel Corporation
Hi Mika,
[auto build test ERROR on v4.4-rc2]
[also build test ERROR on next-20151124]
[cannot apply to ljones-mfd/for-mfd-next]
url: https://github.com/0day-ci/linux/commits/Andy-Shevchenko/device-property-always-check-for-fwnode-type/20151124-183221
config: x86_64-randconfig-x019-11241713 (attached as .config)
reproduce:
# save the attached .config to linux build tree
make ARCH=x86_64
All error/warnings (new ones prefixed by >>):
drivers/mfd/intel-lpss-pci.c:68:30: error: array type has incomplete element type 'struct property_entry'
static struct property_entry spt_i2c_properties[] = {
^
drivers/mfd/intel-lpss-pci.c:69:2: error: implicit declaration of function 'PROPERTY_ENTRY_U32' [-Werror=implicit-function-declaration]
PROPERTY_ENTRY_U32("i2c-sda-hold-time-ns", 230),
^
>> drivers/mfd/intel-lpss-pci.c:73:15: error: variable 'spt_i2c_pset' has initializer but incomplete type
static struct property_set spt_i2c_pset = {
^
drivers/mfd/intel-lpss-pci.c:74:2: error: unknown field 'properties' specified in initializer
.properties = spt_i2c_properties,
^
drivers/mfd/intel-lpss-pci.c:74:16: warning: excess elements in struct initializer
.properties = spt_i2c_properties,
^
drivers/mfd/intel-lpss-pci.c:74:16: note: (near initialization for 'spt_i2c_pset')
drivers/mfd/intel-lpss-pci.c:82:30: error: array type has incomplete element type 'struct property_entry'
static struct property_entry uart_properties[] = {
^
drivers/mfd/intel-lpss-pci.c:85:2: error: implicit declaration of function 'PROPERTY_ENTRY_U8' [-Werror=implicit-function-declaration]
PROPERTY_ENTRY_U8("snps,uart-16550-compatible", 1),
^
drivers/mfd/intel-lpss-pci.c:89:15: error: variable 'uart_pset' has initializer but incomplete type
static struct property_set uart_pset = {
^
drivers/mfd/intel-lpss-pci.c:90:2: error: unknown field 'properties' specified in initializer
.properties = uart_properties,
^
drivers/mfd/intel-lpss-pci.c:90:16: warning: excess elements in struct initializer
.properties = uart_properties,
^
drivers/mfd/intel-lpss-pci.c:90:16: note: (near initialization for 'uart_pset')
>> drivers/mfd/intel-lpss-pci.c:68:30: warning: 'spt_i2c_properties' defined but not used [-Wunused-variable]
static struct property_entry spt_i2c_properties[] = {
^
drivers/mfd/intel-lpss-pci.c:82:30: warning: 'uart_properties' defined but not used [-Wunused-variable]
static struct property_entry uart_properties[] = {
^
cc1: some warnings being treated as errors
vim +/spt_i2c_pset +73 drivers/mfd/intel-lpss-pci.c
62 static INTEL_LPSS_PM_OPS(intel_lpss_pci_pm_ops);
63
64 static const struct intel_lpss_platform_info spt_info = {
65 .clk_rate = 120000000,
66 };
67
> 68 static struct property_entry spt_i2c_properties[] = {
69 PROPERTY_ENTRY_U32("i2c-sda-hold-time-ns", 230),
70 { },
71 };
72
> 73 static struct property_set spt_i2c_pset = {
74 .properties = spt_i2c_properties,
75 };
76
77 static const struct intel_lpss_platform_info spt_i2c_info = {
78 .clk_rate = 120000000,
79 .pset = &spt_i2c_pset,
80 };
81
82 static struct property_entry uart_properties[] = {
83 PROPERTY_ENTRY_U32("reg-io-width", 4),
84 PROPERTY_ENTRY_U32("reg-shift", 2),
85 PROPERTY_ENTRY_U8("snps,uart-16550-compatible", 1),
86 { },
87 };
88
89 static struct property_set uart_pset = {
> 90 .properties = uart_properties,
91 };
92
93 static const struct intel_lpss_platform_info spt_uart_info = {
---
0-DAY kernel test infrastructure Open Source Technology Center
https://lists.01.org/pipermail/kbuild-all Intel Corporation
On 11/24/2015 12:22 PM, Andy Shevchenko wrote:
> From: Mika Westerberg <[email protected]>
>
> With ACPI _DSD (introduced in ACPI v5.1) it is now possible to pass device
> configuration information from ACPI in addition to DT. In order to support
> this, convert the driver to use the unified device property accessors
> instead of DT specific.
>
> Change to ordering a bit so that we first try platform data and if that's
> not available look from device properties. ACPI *CNT methods are then used
> as last resort to override everything else.
>
> Signed-off-by: Mika Westerberg <[email protected]>
> Signed-off-by: Andy Shevchenko <[email protected]>
> ---
> drivers/i2c/busses/i2c-designware-platdrv.c | 48 +++++++++++++----------------
> 1 file changed, 22 insertions(+), 26 deletions(-)
>
> diff --git a/drivers/i2c/busses/i2c-designware-platdrv.c b/drivers/i2c/busses/i2c-designware-platdrv.c
> index 809579e..e9062be 100644
> --- a/drivers/i2c/busses/i2c-designware-platdrv.c
> +++ b/drivers/i2c/busses/i2c-designware-platdrv.c
> @@ -36,6 +36,7 @@
> #include <linux/platform_device.h>
> #include <linux/pm.h>
> #include <linux/pm_runtime.h>
> +#include <linux/property.h>
> #include <linux/io.h>
> #include <linux/slab.h>
> #include <linux/acpi.h>
> @@ -156,33 +157,28 @@ static int dw_i2c_plat_probe(struct platform_device *pdev)
> /* fast mode by default because of legacy reasons */
> clk_freq = 400000;
>
> - if (has_acpi_companion(&pdev->dev)) {
> - dw_i2c_acpi_configure(pdev);
> - } else if (pdev->dev.of_node) {
> - of_property_read_u32(pdev->dev.of_node,
> - "i2c-sda-hold-time-ns", &ht);
> -
> - of_property_read_u32(pdev->dev.of_node,
> - "i2c-sda-falling-time-ns",
> - &dev->sda_falling_time);
> - of_property_read_u32(pdev->dev.of_node,
> - "i2c-scl-falling-time-ns",
> - &dev->scl_falling_time);
> -
> - of_property_read_u32(pdev->dev.of_node, "clock-frequency",
> - &clk_freq);
> -
> - /* Only standard mode at 100kHz and fast mode at 400kHz
> - * are supported.
> - */
> - if (clk_freq != 100000 && clk_freq != 400000) {
> - dev_err(&pdev->dev, "Only 100kHz and 400kHz supported");
> - return -EINVAL;
> - }
> + if ((pdata = dev_get_platdata(&pdev->dev))) {
> + clk_freq = pdata->i2c_scl_freq;
> } else {
> - pdata = dev_get_platdata(&pdev->dev);
> - if (pdata)
> - clk_freq = pdata->i2c_scl_freq;
> + device_property_read_u32(&pdev->dev, "i2c-sda-hold-time-ns",
> + &ht);
> + device_property_read_u32(&pdev->dev, "i2c-sda-falling-time-ns",
> + &dev->sda_falling_time);
> + device_property_read_u32(&pdev->dev, "i2c-scl-falling-time-ns",
> + &dev->scl_falling_time);
> + device_property_read_u32(&pdev->dev, "clock-frequency",
> + &clk_freq);
Mika, Andy: Was this one able to go separately? At least it builds
without rest of the set but is there anything that could break DT based
system if there are no patches 1-8/13?
Acked-by: Jarkko Nikula <[email protected]>
On Tue, Nov 24, 2015 at 12:53:06PM +0200, Jarkko Nikula wrote:
> Mika, Andy: Was this one able to go separately? At least it builds without
> rest of the set but is there anything that could break DT based system if
> there are no patches 1-8/13?
As far as I can tell this should not break existing DT systems but I
have not been able to test myself since I don't have any DT systems
here.
> Acked-by: Jarkko Nikula <[email protected]>
Thanks!
On Tue, Nov 24, 2015 at 06:44:53PM +0800, kbuild test robot wrote:
> Hi Mika,
>
> [auto build test ERROR on v4.4-rc2]
> [also build test ERROR on next-20151124]
> [cannot apply to ljones-mfd/for-mfd-next]
>
> url: https://github.com/0day-ci/linux/commits/Andy-Shevchenko/device-property-always-check-for-fwnode-type/20151124-183221
> config: x86_64-randconfig-x019-11241713 (attached as .config)
> reproduce:
> # save the attached .config to linux build tree
> make ARCH=x86_64
>
> All error/warnings (new ones prefixed by >>):
>
> >> drivers/mfd/intel-lpss-pci.c:68:30: error: array type has incomplete element type 'struct property_entry'
> static struct property_entry uart_properties[] = {
Thanks kbuild robot!
This patch misses include of <linux/property.h>. It seems that it got
added by a subsequent patch which was not sent out.
We will fix this.
On Tuesday, November 24, 2015 12:22:49 PM Andy Shevchenko wrote:
> Instead of using the type and nval fields we will use length (in bytes) of the
> value. The sanity check is done in the accessors.
>
> The built-in property accessors are split in the same way such as device tree.
Do I understand correctly that this is indended to make the built-in properties
follow the DT layout?
Thanks,
Rafael
On Tue, 2015-11-24 at 15:37 +0100, Rafael J. Wysocki wrote:
> On Tuesday, November 24, 2015 12:22:49 PM Andy Shevchenko wrote:
> > Instead of using the type and nval fields we will use length (in
> > bytes) of the
> > value. The sanity check is done in the accessors.
> >
> > The built-in property accessors are split in the same way such as
> > device tree.
>
> Do I understand correctly that this is indended to make the built-in
> properties
> follow the DT layout?
Correct.
>
> Thanks,
> Rafael
>
--
Andy Shevchenko <[email protected]>
Intel Finland Oy
On Tue, 2015-11-24 at 16:11 +0100, Rafael J. Wysocki wrote:
> On Tuesday, November 24, 2015 12:22:46 PM Andy Shevchenko wrote:
> > This series includes few logical sets that bring a support of non-
> > ACPI
> > platforms for Intel Skylake.
> >
> > First part is a refactoring of built-in device properties support:
> > - keep single value inside the structure
> > - provide helper macros to define built-in properties
> > - fall back to secondary fwnode if primary has no asked property
> >
> > Second one is modifications to MFD code and intel-lpss.c driver in
> > particular
> > to define and pass built-in properties to the individual drivers.
> >
> > Last part is a fix for I2C bug found on Lenovo Yoga hardware and a
> > first
> > converted user.
> >
> > Built-in device properties is an alternative to platform data. It
> > provides a
> > unified API that drivers can use to cover all cases at once: DT,
> > ACPI, and
> > built-in properties.
> >
> > With this series applied platform data can be considered obsolete.
> > Moreover,
> > built-in device properties allows to adjust existing configuration,
> > for
> > example, in cases when ACPI values are wrong on some platforms.
> >
> > The series has been tested on available hardware and doesn't break
> > current
> > behaviour. But we ask you, Kevin, to apply the series on your side
> > and check
> > with Lenovo hardware.
> >
> > Andy Shevchenko (8):
> > device property: always check for fwnode type
> > device property: rename helper functions
> > device property: refactor built-in properties support
> > device property: keep single value inplace
> > device property: improve readability of macros
> > device property: return -EINVAL when property isn't found in ACPI
> > device property: Fallback to secondary fwnode if primary misses
> > the
> > property
> > mfd: core: propagate device properties to sub devices drivers
> >
> > Heikki Krogerus (1):
> > device property: helper macros for property entry creation
> >
> > Mika Westerberg (4):
> > mfd: intel-lpss: Add support for passing device properties
> > mfd: intel-lpss: Pass HSUART configuration via properties
> > mfd: intel-lpss: Pass SDA hold time to I2C host controller driver
> > i2c: designware: Convert to use unified device property API
> >
> > drivers/acpi/property.c | 10 +-
> > drivers/base/property.c | 298
> > +++++++++++++++++++++-------
> > drivers/i2c/busses/i2c-designware-platdrv.c | 48 ++---
> > drivers/mfd/intel-lpss-acpi.c | 18 +-
> > drivers/mfd/intel-lpss-pci.c | 43 +++-
> > drivers/mfd/intel-lpss.c | 16 +-
> > drivers/mfd/intel-lpss.h | 2 +
> > drivers/mfd/mfd-core.c | 3 +
> > include/linux/mfd/core.h | 5 +
> > include/linux/property.h | 87 ++++++--
> > 10 files changed, 394 insertions(+), 136 deletions(-)
>
> All patches in this series look good to me overall, but please fix
> the build
> problems reported by 0-day and resubmit.
Thanks for fast response. We will do that later this week while
gathering comments from others.
> Thanks,
> Rafael
>
--
Andy Shevchenko <[email protected]>
Intel Finland Oy
On Tuesday, November 24, 2015 12:22:46 PM Andy Shevchenko wrote:
> This series includes few logical sets that bring a support of non-ACPI
> platforms for Intel Skylake.
>
> First part is a refactoring of built-in device properties support:
> - keep single value inside the structure
> - provide helper macros to define built-in properties
> - fall back to secondary fwnode if primary has no asked property
>
> Second one is modifications to MFD code and intel-lpss.c driver in particular
> to define and pass built-in properties to the individual drivers.
>
> Last part is a fix for I2C bug found on Lenovo Yoga hardware and a first
> converted user.
>
> Built-in device properties is an alternative to platform data. It provides a
> unified API that drivers can use to cover all cases at once: DT, ACPI, and
> built-in properties.
>
> With this series applied platform data can be considered obsolete. Moreover,
> built-in device properties allows to adjust existing configuration, for
> example, in cases when ACPI values are wrong on some platforms.
>
> The series has been tested on available hardware and doesn't break current
> behaviour. But we ask you, Kevin, to apply the series on your side and check
> with Lenovo hardware.
>
> Andy Shevchenko (8):
> device property: always check for fwnode type
> device property: rename helper functions
> device property: refactor built-in properties support
> device property: keep single value inplace
> device property: improve readability of macros
> device property: return -EINVAL when property isn't found in ACPI
> device property: Fallback to secondary fwnode if primary misses the
> property
> mfd: core: propagate device properties to sub devices drivers
>
> Heikki Krogerus (1):
> device property: helper macros for property entry creation
>
> Mika Westerberg (4):
> mfd: intel-lpss: Add support for passing device properties
> mfd: intel-lpss: Pass HSUART configuration via properties
> mfd: intel-lpss: Pass SDA hold time to I2C host controller driver
> i2c: designware: Convert to use unified device property API
>
> drivers/acpi/property.c | 10 +-
> drivers/base/property.c | 298 +++++++++++++++++++++-------
> drivers/i2c/busses/i2c-designware-platdrv.c | 48 ++---
> drivers/mfd/intel-lpss-acpi.c | 18 +-
> drivers/mfd/intel-lpss-pci.c | 43 +++-
> drivers/mfd/intel-lpss.c | 16 +-
> drivers/mfd/intel-lpss.h | 2 +
> drivers/mfd/mfd-core.c | 3 +
> include/linux/mfd/core.h | 5 +
> include/linux/property.h | 87 ++++++--
> 10 files changed, 394 insertions(+), 136 deletions(-)
All patches in this series look good to me overall, but please fix the build
problems reported by 0-day and resubmit.
Thanks,
Rafael
On Tuesday 24 November 2015 12:22:57 Andy Shevchenko wrote:
> +static struct property_entry uart_properties[] = {
> + PROPERTY_ENTRY_U32("reg-io-width", 4),
> + PROPERTY_ENTRY_U32("reg-shift", 2),
> + PROPERTY_ENTRY_U8("snps,uart-16550-compatible", 1),
> + { },
>
If I read the binding correctly, the "snps,uart-16550-compatible" property
is meant to be boolean, meaning true if present and zero-length or false
if absent. Using a u8 propert instead feels wrong.
Maybe we can have a PROPERTY_ENTRY_BOOL() for that?
Arnd
On Tuesday 24 November 2015 12:22:46 Andy Shevchenko wrote:
> This series includes few logical sets that bring a support of non-ACPI
> platforms for Intel Skylake.
>
> First part is a refactoring of built-in device properties support:
> - keep single value inside the structure
> - provide helper macros to define built-in properties
> - fall back to secondary fwnode if primary has no asked property
>
> Second one is modifications to MFD code and intel-lpss.c driver in particular
> to define and pass built-in properties to the individual drivers.
>
> Last part is a fix for I2C bug found on Lenovo Yoga hardware and a first
> converted user.
>
> Built-in device properties is an alternative to platform data. It provides a
> unified API that drivers can use to cover all cases at once: DT, ACPI, and
> built-in properties.
>
> With this series applied platform data can be considered obsolete. Moreover,
> built-in device properties allows to adjust existing configuration, for
> example, in cases when ACPI values are wrong on some platforms.
>
> The series has been tested on available hardware and doesn't break current
> behaviour. But we ask you, Kevin, to apply the series on your side and check
> with Lenovo hardware.
I agree with Rafael, this looks really nice. I found one small thing that
could be improved, see the comment on patch 11.
Aside from that, I think we should have a nicer way to pass a property
list through platform_device_info when calling
platform_device_register_full(). You don't do that here because the drivers
you change are based on MFD cells rather than direct platform devices,
but it would fit in the series and should be easy enough to do.
I don't know why Rafael didn't do that for the initial series already, maybe
he had a good reason.
Arnd
On Tue, Nov 24, 2015 at 9:00 PM, Arnd Bergmann <[email protected]> wrote:
> On Tuesday 24 November 2015 12:22:46 Andy Shevchenko wrote:
>> This series includes few logical sets that bring a support of non-ACPI
>> platforms for Intel Skylake.
>>
>> First part is a refactoring of built-in device properties support:
>> - keep single value inside the structure
>> - provide helper macros to define built-in properties
>> - fall back to secondary fwnode if primary has no asked property
>>
>> Second one is modifications to MFD code and intel-lpss.c driver in particular
>> to define and pass built-in properties to the individual drivers.
>>
>> Last part is a fix for I2C bug found on Lenovo Yoga hardware and a first
>> converted user.
>>
>> Built-in device properties is an alternative to platform data. It provides a
>> unified API that drivers can use to cover all cases at once: DT, ACPI, and
>> built-in properties.
>>
>> With this series applied platform data can be considered obsolete. Moreover,
>> built-in device properties allows to adjust existing configuration, for
>> example, in cases when ACPI values are wrong on some platforms.
>>
>> The series has been tested on available hardware and doesn't break current
>> behaviour. But we ask you, Kevin, to apply the series on your side and check
>> with Lenovo hardware.
>
> I agree with Rafael, this looks really nice. I found one small thing that
> could be improved, see the comment on patch 11.
>
> Aside from that, I think we should have a nicer way to pass a property
> list through platform_device_info when calling
> platform_device_register_full(). You don't do that here because the drivers
> you change are based on MFD cells rather than direct platform devices,
> but it would fit in the series and should be easy enough to do.
>
> I don't know why Rafael didn't do that for the initial series already, maybe
> he had a good reason.
I was focusing on the generics rather than on any particular
individual bus type.
Thanks,
Rafael
On Tue, Nov 24, 2015 at 08:53:04PM +0100, Arnd Bergmann wrote:
> On Tuesday 24 November 2015 12:22:57 Andy Shevchenko wrote:
> > +static struct property_entry uart_properties[] = {
> > + PROPERTY_ENTRY_U32("reg-io-width", 4),
> > + PROPERTY_ENTRY_U32("reg-shift", 2),
> > + PROPERTY_ENTRY_U8("snps,uart-16550-compatible", 1),
> > + { },
> >
>
> If I read the binding correctly, the "snps,uart-16550-compatible" property
> is meant to be boolean, meaning true if present and zero-length or false
> if absent. Using a u8 propert instead feels wrong.
>
> Maybe we can have a PROPERTY_ENTRY_BOOL() for that?
That's a good idea. We'll add it to the next revision of the series.