| Message ID | 1401716463-23457-2-git-send-email-lee.jones@linaro.org |
|---|---|
| State | New |
| Headers | show |
> We have a problem. There are lots of I2C device ID tables scattered > around the kernel which are redundant in all Device Tree and/or ACPI > only supported device drivers. After recent discussions it has become > apparent that the only thing blocking the complete removal of these > tables is the continued support of 'register an I2C device via sysfs' > functionality. As the sysfs method doesn't know anything about Device > Tree or ACPI, we can not pass any nodes in. This patch searches all > known Device Tree nodes and attempts to acquire a match from the > device name provided via sysfs. It can not fail, but if found assigns > the matching of_node to i2c_board_info prior to registering. > > Signed-off-by: Lee Jones <lee.jones@linaro.org> > --- > drivers/i2c/i2c-core.c | 31 +++++++++++++++++++++++++++++++ > 1 file changed, 31 insertions(+) [...] > +static struct device_node *of_i2c_type_to_node(char *type) > +{ > + struct device_node *np; > + const char *compatible, *name; > + int len; > + > + if (!type) > + return NULL; > + > + for_each_of_allnodes(np) { > + compatible = of_get_property(np, "compatible", &len); > + if (!compatible) > + continue; > + > + name = strchr(compatible, ','); > + if (!name) > + name = compatible; > + else > + name++; > + > + if (!strncmp(name, type, len - (name - compatible))) > + return np; > + } > + > + return NULL; > +} Ah, this _might_ not work if there are more than one devices of the same type. I guess I need to check the 'reg' property too.
On Mon, 2 Jun 2014 14:41:01 +0100, Lee Jones <lee.jones@linaro.org> wrote: > We have a problem. There are lots of I2C device ID tables scattered > around the kernel which are redundant in all Device Tree and/or ACPI > only supported device drivers. After recent discussions it has become > apparent that the only thing blocking the complete removal of these > tables is the continued support of 'register an I2C device via sysfs' > functionality. As the sysfs method doesn't know anything about Device > Tree or ACPI, we can not pass any nodes in. This patch searches all > known Device Tree nodes and attempts to acquire a match from the > device name provided via sysfs. It can not fail, but if found assigns > the matching of_node to i2c_board_info prior to registering. Ummm... blech! :-) It shouldn't actually be necessary to have a node when instantiating an i2c_device... in fact, there /shouldn't/ be a node if the device is getting instantiated by had via sysfs. If there was a node, then I would expect a device to have been created at the same time. I think panto's overlays solve the problem where a node actually is required to instantiate the device by actually adding a node and letting the i2c subsystem create the associated device automatically. g. > > Signed-off-by: Lee Jones <lee.jones@linaro.org> > --- > drivers/i2c/i2c-core.c | 31 +++++++++++++++++++++++++++++++ > 1 file changed, 31 insertions(+) > > diff --git a/drivers/i2c/i2c-core.c b/drivers/i2c/i2c-core.c > index 7c7f4b8..2e47641 100644 > --- a/drivers/i2c/i2c-core.c > +++ b/drivers/i2c/i2c-core.c > @@ -64,6 +64,7 @@ static DEFINE_IDR(i2c_adapter_idr); > > static struct device_type i2c_client_type; > static int i2c_detect(struct i2c_adapter *adapter, struct i2c_driver *driver); > +static struct device_node *of_i2c_type_to_node(char *type); > > static struct static_key i2c_trace_msg = STATIC_KEY_INIT_FALSE; > > @@ -863,6 +864,8 @@ i2c_sysfs_new_device(struct device *dev, struct device_attribute *attr, > return -EINVAL; > } > > + info.of_node = of_i2c_type_to_node(info.type); > + > client = i2c_new_device(adap, &info); > if (!client) > return -EINVAL; > @@ -1088,8 +1091,36 @@ struct i2c_adapter *of_find_i2c_adapter_by_node(struct device_node *node) > return i2c_verify_adapter(dev); > } > EXPORT_SYMBOL(of_find_i2c_adapter_by_node); > + > +static struct device_node *of_i2c_type_to_node(char *type) > +{ > + struct device_node *np; > + const char *compatible, *name; > + int len; > + > + if (!type) > + return NULL; > + > + for_each_of_allnodes(np) { > + compatible = of_get_property(np, "compatible", &len); > + if (!compatible) > + continue; > + > + name = strchr(compatible, ','); > + if (!name) > + name = compatible; > + else > + name++; > + > + if (!strncmp(name, type, len - (name - compatible))) > + return np; > + } > + > + return NULL; > +} > #else > static void of_i2c_register_devices(struct i2c_adapter *adap) { } > +static struct device_node *of_i2c_type_to_node(char *type) { return NULL; } > #endif /* CONFIG_OF */ > > /* ACPI support code */ > -- > 1.8.3.2 > -- To unsubscribe from this list: send the line "unsubscribe linux-i2c" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
> > We have a problem. There are lots of I2C device ID tables scattered > > around the kernel which are redundant in all Device Tree and/or ACPI > > only supported device drivers. After recent discussions it has become > > apparent that the only thing blocking the complete removal of these > > tables is the continued support of 'register an I2C device via sysfs' > > functionality. As the sysfs method doesn't know anything about Device > > Tree or ACPI, we can not pass any nodes in. This patch searches all > > known Device Tree nodes and attempts to acquire a match from the > > device name provided via sysfs. It can not fail, but if found assigns > > the matching of_node to i2c_board_info prior to registering. > > Ummm... blech! :-) > > It shouldn't actually be necessary to have a node when instantiating an > i2c_device... in fact, there /shouldn't/ be a node if the device is > getting instantiated by had via sysfs. If there was a node, then I would > expect a device to have been created at the same time. You're right. I realised that about 10 mins after I sent the patch! > I think panto's overlays solve the problem where a node actually is > required to instantiate the device by actually adding a node and letting > the i2c subsystem create the associated device automatically. This looks like it's exactly what I need. Pulling the code now. > > Signed-off-by: Lee Jones <lee.jones@linaro.org> > > --- > > drivers/i2c/i2c-core.c | 31 +++++++++++++++++++++++++++++++ > > 1 file changed, 31 insertions(+) > > > > diff --git a/drivers/i2c/i2c-core.c b/drivers/i2c/i2c-core.c > > index 7c7f4b8..2e47641 100644 > > --- a/drivers/i2c/i2c-core.c > > +++ b/drivers/i2c/i2c-core.c > > @@ -64,6 +64,7 @@ static DEFINE_IDR(i2c_adapter_idr); > > > > static struct device_type i2c_client_type; > > static int i2c_detect(struct i2c_adapter *adapter, struct i2c_driver *driver); > > +static struct device_node *of_i2c_type_to_node(char *type); > > > > static struct static_key i2c_trace_msg = STATIC_KEY_INIT_FALSE; > > > > @@ -863,6 +864,8 @@ i2c_sysfs_new_device(struct device *dev, struct device_attribute *attr, > > return -EINVAL; > > } > > > > + info.of_node = of_i2c_type_to_node(info.type); > > + > > client = i2c_new_device(adap, &info); > > if (!client) > > return -EINVAL; > > @@ -1088,8 +1091,36 @@ struct i2c_adapter *of_find_i2c_adapter_by_node(struct device_node *node) > > return i2c_verify_adapter(dev); > > } > > EXPORT_SYMBOL(of_find_i2c_adapter_by_node); > > + > > +static struct device_node *of_i2c_type_to_node(char *type) > > +{ > > + struct device_node *np; > > + const char *compatible, *name; > > + int len; > > + > > + if (!type) > > + return NULL; > > + > > + for_each_of_allnodes(np) { > > + compatible = of_get_property(np, "compatible", &len); > > + if (!compatible) > > + continue; > > + > > + name = strchr(compatible, ','); > > + if (!name) > > + name = compatible; > > + else > > + name++; > > + > > + if (!strncmp(name, type, len - (name - compatible))) > > + return np; > > + } > > + > > + return NULL; > > +} > > #else > > static void of_i2c_register_devices(struct i2c_adapter *adap) { } > > +static struct device_node *of_i2c_type_to_node(char *type) { return NULL; } > > #endif /* CONFIG_OF */ > > > > /* ACPI support code */ >
diff --git a/drivers/i2c/i2c-core.c b/drivers/i2c/i2c-core.c index 7c7f4b8..2e47641 100644 --- a/drivers/i2c/i2c-core.c +++ b/drivers/i2c/i2c-core.c @@ -64,6 +64,7 @@ static DEFINE_IDR(i2c_adapter_idr); static struct device_type i2c_client_type; static int i2c_detect(struct i2c_adapter *adapter, struct i2c_driver *driver); +static struct device_node *of_i2c_type_to_node(char *type); static struct static_key i2c_trace_msg = STATIC_KEY_INIT_FALSE; @@ -863,6 +864,8 @@ i2c_sysfs_new_device(struct device *dev, struct device_attribute *attr, return -EINVAL; } + info.of_node = of_i2c_type_to_node(info.type); + client = i2c_new_device(adap, &info); if (!client) return -EINVAL; @@ -1088,8 +1091,36 @@ struct i2c_adapter *of_find_i2c_adapter_by_node(struct device_node *node) return i2c_verify_adapter(dev); } EXPORT_SYMBOL(of_find_i2c_adapter_by_node); + +static struct device_node *of_i2c_type_to_node(char *type) +{ + struct device_node *np; + const char *compatible, *name; + int len; + + if (!type) + return NULL; + + for_each_of_allnodes(np) { + compatible = of_get_property(np, "compatible", &len); + if (!compatible) + continue; + + name = strchr(compatible, ','); + if (!name) + name = compatible; + else + name++; + + if (!strncmp(name, type, len - (name - compatible))) + return np; + } + + return NULL; +} #else static void of_i2c_register_devices(struct i2c_adapter *adap) { } +static struct device_node *of_i2c_type_to_node(char *type) { return NULL; } #endif /* CONFIG_OF */ /* ACPI support code */
We have a problem. There are lots of I2C device ID tables scattered around the kernel which are redundant in all Device Tree and/or ACPI only supported device drivers. After recent discussions it has become apparent that the only thing blocking the complete removal of these tables is the continued support of 'register an I2C device via sysfs' functionality. As the sysfs method doesn't know anything about Device Tree or ACPI, we can not pass any nodes in. This patch searches all known Device Tree nodes and attempts to acquire a match from the device name provided via sysfs. It can not fail, but if found assigns the matching of_node to i2c_board_info prior to registering. Signed-off-by: Lee Jones <lee.jones@linaro.org> --- drivers/i2c/i2c-core.c | 31 +++++++++++++++++++++++++++++++ 1 file changed, 31 insertions(+)