device connection: Add support for device graphs
Subsystems (4)
Heikki Krogerus
maintainerv3 cover device connection: Add support for device graphs
Hi,
This is the third version of my proposal to add device graph parsing to the device connection API. There was only one problem reported in v2 by Jun - kernel-doc entry was missing for the new fwnode member in struct usb_role_switch_desc - and it's now fixed.
The second version of the series:
https://lkml.org/lkml/2019/1/30/622
The commit message from v1:
This series adds support for OF and ACPI device graph parsing to the device connection API.
Handling the graph is straightforward, but because I'm adding that fwnode member to struct device_connection, I had to make sure all the existing users consider it.
The plan is to only support matching with fwnode in the future, so no more device name matching. The software fwnodes that we now have in kernel should make that possible, once we add support for references to them.
The original RFC:
https://lkml.org/lkml/2018/10/24/619
thanks,
Heikki Krogerus (9):
platform/x86: intel_cht_int33fe: Prepare for better mux naming scheme
usb: typec: Rationalize the API for the muxes platform/x86: intel_cht_int33fe: Remove old style mux connections device connection: Add fwnode member to struct device_connection
usb: typec: mux: Find the muxes by also matching against the device
node
usb: roles: Find the muxes by also matching against the device node
usb: typec: Find the ports by also matching against the device node device connection: Prepare support for firmware described connections device connection: Find device connections also from device graphs
drivers/base/devcon.c | 62 ++++++++++++++- drivers/platform/x86/intel_cht_int33fe.c | 15 ++-- drivers/usb/roles/class.c | 21 +++++- drivers/usb/typec/class.c | 31 ++++++-- drivers/usb/typec/mux.c | 96 ++++++++++++++++++++---- include/linux/device.h | 6 ++ include/linux/usb/role.h | 2 + include/linux/usb/typec_mux.h | 3 +- 8 files changed, 196 insertions(+), 40 deletions(-)
Signature · 1 line
2.20.1
Heikki Krogerus
maintainerPATCH v3 1/9 platform/x86: intel_cht_int33fe: Prepare for better mux naming scheme
Adding new connections with for the muxes with new identifiers. The old connection are left in for now.
--- drivers/platform/x86/intel_cht_int33fe.c | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-)
drivers/platform/x86/intel_cht_int33fe.c +7 -1
@@ -32,7 +32,7 @@ struct cht_int33fe_data {struct i2c_client *fusb302;struct i2c_client *pi3usb30532;/* Contain a list-head must be per device */ -struct device_connection connections[5 ]; +struct device_connection connections[7 ]; };/*
@@ -184,6 +184,12 @@ static int cht_int33fe_probe(struct platform_device *pdev) data->connections[3 ].endpoint[0 ] ="i2c-fusb302" ; data->connections[3 ].endpoint[1 ] ="intel_xhci_usb_sw-role-switch" ; data->connections[3 ].id ="usb-role-switch" ; + data->connections[4 ].endpoint[0 ] ="port0" ; + data->connections[4 ].endpoint[1 ] ="i2c-pi3usb30532" ; + data->connections[4 ].id ="orientation-switch" ; + data->connections[5 ].endpoint[0 ] ="port0" ; + data->connections[5 ].endpoint[1 ] ="i2c-pi3usb30532" ; + data->connections[5 ].id ="mode-switch" ;device_connections_add (data->connections);
Re: [PATCH v3 1/9] platform/x86: intel_cht_int33fe: Prepare for better mux naming scheme
Hello!
Quoted 1 line · Heikki Krogerus
Adding new connections with for the muxes with new
^^^^^^^^
I can't parse this. :-)
Quoted 6 lines
identifiers. The old connection are left in for now.
Reviewed-by Hans de Goede <[email protected]>Reviewed-by Andy Shevchenko <[email protected]>Reviewed-by Jun Li <[email protected]>Signed-off-by Heikki Krogerus <[email protected]>
[...]
MBR, Sergei
Heikki Krogerus
maintainerPATCH v3 2/9 usb: typec: Rationalize the API for the muxes
In reply to Heikki KrogerusSince with accessory modes there is no need for additional identification when requesting a handle to the mux, we can replace the second parameter that is passed to the typec_mux_get() function with a pointer to alternate mode description structure, and simply passing NULL with accessory modes.
This change means the naming of the mux device connections can be updated. Alternate and Accessory Modes will both be handled with muxes named "mode-switch", and the orientation switches will be named "orientation-switch".
Future identification of the alternate modes will be later done using device property "svid" of the mux.
--- drivers/usb/typec/class.c | 7 ++----- drivers/usb/typec/mux.c | 10 ++++++---- include/linux/usb/typec_mux.h | 3 ++- 3 files changed, 10 insertions(+), 10 deletions(-)
drivers/usb/typec/class.c +2 -5
@@ -1496,11 +1496,8 @@ typec_port_register_altmode(struct typec_port *port, {struct typec_altmode *adev;struct typec_mux *mux; -char id[10 ]; -sprintf (id,"id%04xm%02x" , desc->svid, desc->mode); - - mux =typec_mux_get (&port->dev, id); + mux =typec_mux_get (&port->dev, desc);if (IS_ERR (mux))return ERR_CAST (mux);
@@ -1593,7 +1590,7 @@ struct typec_port *typec_register_port(struct device *parent,return ERR_CAST (port->sw); } - port->mux =typec_mux_get (&port->dev,"typec-mux" ); + port->mux =typec_mux_get (&port->dev,NULL );if (IS_ERR (port->mux)) {put_device (&port->dev);return ERR_CAST (port->mux);
drivers/usb/typec/mux.c +6 -4
@@ -48,7 +48,7 @@ struct typec_switch *typec_switch_get(struct device *dev)struct typec_switch *sw;mutex_lock (&switch_lock); - sw =device_connection_find_match (dev,"typec-switch" ,NULL , + sw =device_connection_find_match (dev,"orientation-switch" ,NULL , typec_switch_match);if (!IS_ERR_OR_NULL (sw)) {WARN_ON (!try_module_get (sw->dev->driver->owner));
@@ -128,19 +128,21 @@ static void *typec_mux_match(struct device_connection *con, int ep, void *data)/** * typec_mux_get - Find USB Type-C Multiplexer * @dev: The caller device -* @name: Mux identifier +* @desc: Alt Mode description * * Finds a mux linked to the caller. This function is primarily meant for the * Type-C drivers. Returns a reference to the mux on success, NULL if no * matching connection was found, or ERR_PTR(-EPROBE_DEFER) when a connection * was found but the mux has not been enumerated yet. */ -struct typec_mux *typec_mux_get (struct device *dev,const char *name) +struct typec_mux *typec_mux_get (struct device *dev, +const struct typec_altmode_desc *desc) {struct typec_mux *mux;mutex_lock (&mux_lock); - mux =device_connection_find_match (dev, name,NULL , typec_mux_match); + mux =device_connection_find_match (dev,"mode-switch" , (void *)desc, + typec_mux_match);if (!IS_ERR_OR_NULL (mux)) {WARN_ON (!try_module_get (mux->dev->driver->owner));get_device (mux->dev);
include/linux/usb/typec_mux.h +2 -1
@@ -47,7 +47,8 @@ void typec_switch_put(struct typec_switch *sw);int typec_switch_register (struct typec_switch *sw);void typec_switch_unregister (struct typec_switch *sw); -struct typec_mux *typec_mux_get (struct device *dev,const char *name); +struct typec_mux * +typec_mux_get (struct device *dev,const struct typec_altmode_desc *desc);void typec_mux_put (struct typec_mux *mux);int typec_mux_register (struct typec_mux *mux);void typec_mux_unregister (struct typec_mux *mux);
Heikki Krogerus
maintainerPATCH v3 3/9 platform/x86: intel_cht_int33fe: Remove old style mux connections
In reply to Heikki KrogerusThe new mux connection naming scheme is now in use, so dropping the connections still using the old names. From now on the same connection description named "mode-switch" is used with both the port and the alternate modes, so on CHT the DP alt mode will use the same connection as the port to get a handle to the mux device.
--- drivers/platform/x86/intel_cht_int33fe.c | 21 ++++++--------------- 1 file changed, 6 insertions(+), 15 deletions(-)
drivers/platform/x86/intel_cht_int33fe.c +6 -15
@@ -32,7 +32,7 @@ struct cht_int33fe_data {struct i2c_client *fusb302;struct i2c_client *pi3usb30532;/* Contain a list-head must be per device */ -struct device_connection connections[7 ]; +struct device_connection connections[4 ]; };/*
@@ -174,22 +174,13 @@ static int cht_int33fe_probe(struct platform_device *pdev) data->connections[0 ].endpoint[0 ] ="port0" ; data->connections[0 ].endpoint[1 ] ="i2c-pi3usb30532" ; - data->connections[0 ].id ="typec-switch" ; + data->connections[0 ].id ="orientation-switch" ; data->connections[1 ].endpoint[0 ] ="port0" ; data->connections[1 ].endpoint[1 ] ="i2c-pi3usb30532" ; - data->connections[1 ].id ="typec-mux" ; - data->connections[2 ].endpoint[0 ] ="port0" ; - data->connections[2 ].endpoint[1 ] ="i2c-pi3usb30532" ; - data->connections[2 ].id ="idff01m01" ; - data->connections[3 ].endpoint[0 ] ="i2c-fusb302" ; - data->connections[3 ].endpoint[1 ] ="intel_xhci_usb_sw-role-switch" ; - data->connections[3 ].id ="usb-role-switch" ; - data->connections[4 ].endpoint[0 ] ="port0" ; - data->connections[4 ].endpoint[1 ] ="i2c-pi3usb30532" ; - data->connections[4 ].id ="orientation-switch" ; - data->connections[5 ].endpoint[0 ] ="port0" ; - data->connections[5 ].endpoint[1 ] ="i2c-pi3usb30532" ; - data->connections[5 ].id ="mode-switch" ; + data->connections[1 ].id ="mode-switch" ; + data->connections[2 ].endpoint[0 ] ="i2c-fusb302" ; + data->connections[2 ].endpoint[1 ] ="intel_xhci_usb_sw-role-switch" ; + data->connections[2 ].id ="usb-role-switch" ;device_connections_add (data->connections);
Heikki Krogerus
maintainerPATCH v3 4/9 device connection: Add fwnode member to struct device_connection
In reply to Heikki KrogerusThis will prepare the device connection API for connections described in firmware.
--- include/linux/device.h | 6 ++++++ 1 file changed, 6 insertions(+)
include/linux/device.h +6 -0
@@ -758,11 +758,17 @@ struct device_dma_parameters {/** * struct device_connection - Device Connection Descriptor +* @fwnode: The device node of the connected device * @endpoint: The names of the two devices connected together * @id: Unique identifier for the connection * @list: List head, private, for internal use only +* +* NOTE: @fwnode is not used together with @endpoint. @fwnode is used when +* platform firmware defines the connection. When the connection is registered +* with device_connection_add() @endpoint is used instead. */ struct device_connection { +struct fwnode_handle *fwnode;const char *endpoint[2 ];const char *id;struct list_head list;
Heikki Krogerus
maintainerPATCH v3 5/9 usb: typec: mux: Find the muxes by also matching against the device node
In reply to Heikki KrogerusWhen the connections are defined in firmware, struct device_connection will have the fwnode member pointing to the device node (struct fwnode_handle) of the requested device, and the endpoint will not be used at all in that case.
--- drivers/usb/typec/mux.c | 86 +++++++++++++++++++++++++++++++++++------ 1 file changed, 74 insertions(+), 12 deletions(-)
drivers/usb/typec/mux.c +74 -12
@@ -11,6 +11,8 @@#include <linux/list.h> #include <linux/module.h> #include <linux/mutex.h> +#include <linux/property.h> +#include <linux/slab.h> #include <linux/usb/typec_mux.h> static DEFINE_MUTEX (switch_lock);
@@ -23,15 +25,25 @@ static void *typec_switch_match(struct device_connection *con, int ep, {struct typec_switch *sw; -list_for_each_entry (sw, &switch_list, entry) -if (!strcmp (con->endpoint[ep],dev_name (sw->dev))) -return sw; +if (!con->fwnode) { +list_for_each_entry (sw, &switch_list, entry) +if (!strcmp (con->endpoint[ep],dev_name (sw->dev))) +return sw; +return ERR_PTR (-EPROBE_DEFER ); + }/* -* We only get called if a connection was found, tell the caller to -* wait for the switch to show up. +* With OF graph the mux node must have a boolean device property named +* "orientation-switch". */ -return ERR_PTR (-EPROBE_DEFER ); +if (con->id && !fwnode_property_present (con->fwnode, con->id)) +return NULL ; + +list_for_each_entry (sw, &switch_list, entry) +if (dev_fwnode (sw->dev) == con->fwnode) +return sw; + +return con->id ?ERR_PTR (-EPROBE_DEFER ) :NULL ; }/**
@@ -112,17 +124,67 @@ EXPORT_SYMBOL_GPL(typec_switch_unregister);static void *typec_mux_match (struct device_connection *con,int ep,void *data) { +const struct typec_altmode_desc *desc = data;struct typec_mux *mux; +size_t nval; +bool match; +u16 *val; +int i; -list_for_each_entry (mux, &mux_list, entry) -if (!strcmp (con->endpoint[ep],dev_name (mux->dev))) -return mux; +if (!con->fwnode) { +list_for_each_entry (mux, &mux_list, entry) +if (!strcmp (con->endpoint[ep],dev_name (mux->dev))) +return mux; +return ERR_PTR (-EPROBE_DEFER ); + }/* -* We only get called if a connection was found, tell the caller to -* wait for the switch to show up. +* Check has the identifier already been "consumed". If it +* has, no need to do any extra connection identification. */ -return ERR_PTR (-EPROBE_DEFER ); + match = !con->id; +if (match) + goto find_mux; + +/* Accessory Mode muxes */ +if (!desc) { + match =fwnode_property_present (con->fwnode,"accessory" ); +if (match) + goto find_mux; +return NULL ; + } + +/* Alternate Mode muxes */ + nval =fwnode_property_read_u16_array (con->fwnode,"svid" ,NULL ,0 ); +if (nval <=0 ) +return NULL ; + + val =kcalloc (nval,sizeof (*val),GFP_KERNEL ); +if (!val) +return ERR_PTR (-ENOMEM ); + + nval =fwnode_property_read_u16_array (con->fwnode,"svid" , val, nval); +if (nval <0 ) { +kfree (val); +return ERR_PTR (nval); + } + +for (i =0 ; i < nval; i++) { + match = val[i] == desc->svid; +if (match) { +kfree (val); + goto find_mux; + } + } +kfree (val); +return NULL ; + +find_mux: +list_for_each_entry (mux, &mux_list, entry) +if (dev_fwnode (mux->dev) == con->fwnode) +return mux; + +return match ?ERR_PTR (-EPROBE_DEFER ) :NULL ; }/**
Heikki Krogerus
maintainerPATCH v3 6/9 usb: roles: Find the muxes by also matching against the device node
In reply to Heikki KrogerusWhen the connections are defined in firmware, struct device_connection will have the fwnode member pointing to the device node (struct fwnode_handle) of the requested device, and the endpoint will not be used at all in that case.
--- drivers/usb/roles/class.c | 21 ++++++++++++++++++--- include/linux/usb/role.h | 2 ++ 2 files changed, 20 insertions(+), 3 deletions(-)
drivers/usb/roles/class.c +18 -3
@@ -8,6 +8,7 @@*/ #include <linux/usb/role.h> +#include <linux/property.h> #include <linux/device.h> #include <linux/module.h> #include <linux/mutex.h>
@@ -84,7 +85,12 @@ enum usb_role usb_role_switch_get_role(struct usb_role_switch *sw) }EXPORT_SYMBOL_GPL (usb_role_switch_get_role); -static int __switch_match (struct device *dev,const void *name) +static int switch_fwnode_match (struct device *dev,const void *fwnode) +{ +return dev_fwnode (dev) == fwnode; +} + +static int switch_name_match (struct device *dev,const void *name) {return !strcmp ((const char *)name,dev_name (dev)); }
@@ -94,8 +100,16 @@ static void *usb_role_switch_match(struct device_connection *con, int ep, {struct device *dev; - dev =class_find_device (role_class,NULL , con->endpoint[ep], - __switch_match); +if (con->fwnode) { +if (!fwnode_property_present (con->fwnode, con->id)) +return NULL ; + + dev =class_find_device (role_class,NULL , con->fwnode, + switch_fwnode_match); + }else { + dev =class_find_device (role_class,NULL , con->endpoint[ep], + switch_name_match); + }return dev ?to_role_switch (dev) :ERR_PTR (-EPROBE_DEFER ); }
@@ -266,6 +280,7 @@ usb_role_switch_register(struct device *parent, sw->get = desc->get; sw->dev.parent = parent; + sw->dev.fwnode = desc->fwnode; sw->dev.class = role_class; sw->dev.type = &usb_role_dev_type;dev_set_name (&sw->dev,"%s-role-switch" ,dev_name (parent));
include/linux/usb/role.h +2 -0
@@ -18,6 +18,7 @@ typedef enum usb_role (*usb_role_switch_get_t)(struct device *dev);/** * struct usb_role_switch_desc - USB Role Switch Descriptor +* @fwnode: The device node to be associated with the role switch * @usb2_port: Optional reference to the host controller port device (USB2) * @usb3_port: Optional reference to the host controller port device (USB3) * @udc: Optional reference to the peripheral controller device
@@ -32,6 +33,7 @@ typedef enum usb_role (*usb_role_switch_get_t)(struct device *dev);* usb_role_switch_register() before registering the switch. */ struct usb_role_switch_desc { +struct fwnode_handle *fwnode;struct device *usb2_port;struct device *usb3_port;struct device *udc;
Heikki Krogerus
maintainerPATCH v3 7/9 usb: typec: Find the ports by also matching against the device node
In reply to Heikki KrogerusWhen the connections are defined in firmware, struct device_connection will have the fwnode member pointing to the device node (struct fwnode_handle) of the requested device, and the endpoint will not be used at all in that case.
--- drivers/usb/typec/class.c | 24 +++++++++++++++++++++--- 1 file changed, 21 insertions(+), 3 deletions(-)
drivers/usb/typec/class.c +21 -3
@@ -9,6 +9,7 @@#include <linux/device.h> #include <linux/module.h> #include <linux/mutex.h> +#include <linux/property.h> #include <linux/slab.h> #include "bus.h"
@@ -204,15 +205,32 @@ static void typec_altmode_put_partner(struct altmode *altmode)put_device (&adev->dev); } -static int __typec_port_match (struct device *dev,const void *name) +static int typec_port_fwnode_match (struct device *dev,const void *fwnode) +{ +return dev_fwnode (dev) == fwnode; +} + +static int typec_port_name_match (struct device *dev,const void *name) {return !strcmp ((const char *)name,dev_name (dev)); }static void *typec_port_match (struct device_connection *con,int ep,void *data) { -return class_find_device (typec_class,NULL , con->endpoint[ep], - __typec_port_match); +struct device *dev; + +/* +* FIXME: Check does the fwnode supports the requested SVID. If it does +* we need to return ERR_PTR(-PROBE_DEFER) when there is no device. +*/ +if (con->fwnode) +return class_find_device (typec_class,NULL , con->fwnode, + typec_port_fwnode_match); + + dev =class_find_device (typec_class,NULL , con->endpoint[ep], + typec_port_name_match); + +return dev ? dev :ERR_PTR (-EPROBE_DEFER ); } struct typec_altmode *
Re: [PATCH v3 7/9] usb: typec: Find the ports by also matching against the device node
Quoted 18 lines · Heikki Krogerus
When the connections are defined in firmware, struct
device_connectionwill have the fwnode member pointing to the device node (structfwnode_handle) of the requested device, and the endpoint will not be used at all in that case.Acked-by Hans de Goede <[email protected]>Reviewed-by Andy Shevchenko <[email protected]>Reviewed-by Jun Li <[email protected]>Signed-off-by Heikki Krogerus <[email protected]>--- drivers/usb/typec/class.c | 24 +++++++++++++++++++++--- 1 file changed, 21 insertions(+), 3 deletions(-)drivers/usb/typec/class.c +0
-0
[...]
drivers/usb/typec/class.c static void typec_altmode_put_partner(struct altmode *altmode) · line 223
Quoted 2 lines
drivers/usb/typec/class.c +2
-0+/* +* FIXME: Check does the fwnode supports the requested SVID. If it does
Perhaps "check if the fwnode supports"? Or at least s/supports/support/...
[...]
MBR, Sergei
Heikki Krogerus
maintainerPATCH v3 8/9 device connection: Prepare support for firmware described connections
In reply to Heikki KrogerusWhen the connections are defined in firmware, struct device_connection will have the fwnode member pointing to the device node (struct fwnode_handle) of the requested device. The endpoint member for the device names will not be used at all in that case.
--- drivers/base/devcon.c | 24 ++++++++++++++++++++++++ 1 file changed, 24 insertions(+)
drivers/base/devcon.c +24 -0
@@ -75,12 +75,36 @@ static struct bus_type *generic_match_buses[] = { NULL, }; +static int device_fwnode_match (struct device *dev,void *fwnode) +{ +return dev_fwnode (dev) == fwnode; +} + +static void *device_connection_fwnode_match (struct device_connection *con) +{ +struct bus_type *bus; +struct device *dev; + +for (bus = generic_match_buses[0 ]; bus; bus++) { + dev =bus_find_device (bus,NULL , (void *)con->fwnode, + device_fwnode_match); +if (dev && !strncmp (dev_name (dev), con->id,strlen (con->id))) +return dev; + +put_device (dev); + } +return NULL ; +} +/* This tries to find the device from the most common bus types by name. */ static void *generic_match (struct device_connection *con,int ep,void *data) {struct bus_type *bus;struct device *dev; +if (con->fwnode) +return device_connection_fwnode_match (con); +for (bus = generic_match_buses[0 ]; bus; bus++) { dev =bus_find_device_by_name (bus,NULL , con->endpoint[ep]);if (dev)
Heikki Krogerus
maintainerPATCH v3 9/9 device connection: Find device connections also from device graphs
In reply to Heikki KrogerusIf connections between devices are described in OF graph or ACPI device graph, we can find them by using the fwnode_graph_*() functions.
--- drivers/base/devcon.c | 38 +++++++++++++++++++++++++++++++++++--- 1 file changed, 35 insertions(+), 3 deletions(-)
drivers/base/devcon.c +35 -3
@@ -7,10 +7,37 @@*/ #include <linux/device.h> +#include <linux/property.h> static DEFINE_MUTEX (devcon_lock);static LIST_HEAD (devcon_list); +typedef void *(*devcon_match_fn_t )(struct device_connection *con,int ep, +void *data); + +static void * +fwnode_graph_devcon_match (struct fwnode_handle *fwnode,const char *con_id, +void *data,devcon_match_fn_t match) +{ +struct device_connection con = { .id = con_id }; +struct fwnode_handle *ep; +void *ret; + +fwnode_graph_for_each_endpoint (fwnode, ep) { + con.fwnode =fwnode_graph_get_remote_port_parent (ep); +if (!fwnode_device_is_available (con.fwnode)) +continue ; + + ret =match (&con,-1 , data); +fwnode_handle_put (con.fwnode); +if (ret) { +fwnode_handle_put (ep); +return ret; + } + } +return NULL ; +} +/** * device_connection_find_match - Find physical connection to a device * @dev: Device with the connection
@@ -23,10 +50,9 @@ static LIST_HEAD(devcon_list);* caller is expecting to be returned. */ void *device_connection_find_match (struct device *dev,const char *con_id, -void *data, -void *(*match)(struct device_connection *con, -int ep,void *data)) +void *data,devcon_match_fn_t match) { +struct fwnode_handle *fwnode =dev_fwnode (dev);const char *devname =dev_name (dev);struct device_connection *con;void *ret =NULL ;
@@ -35,6 +61,12 @@ void *device_connection_find_match(struct device *dev, const char *con_id,if (!match)return NULL ; +if (fwnode) { + ret =fwnode_graph_devcon_match (fwnode, con_id, data, match); +if (ret) +return ret; + } +mutex_lock (&devcon_lock);list_for_each_entry (con, &devcon_list, list) {
Greg Kroah-Hartman
maintainerQuoted 19 lines · Heikki Krogerus
Hi,
This is the third version of my proposal to add device graph parsing to the device connection API. There was only one problem reported in v2 by Jun - kernel-doc entry was missing for the new fwnode member in struct
usb_role_switch_desc- and it's now fixed.The second version of the series:
https://lkml.org/lkml/2019/1/30/622The commit message from v1:
This series adds support for OF and ACPI device graph parsing to the device connection API.
Handling the graph is straightforward, but because I'm adding that fwnode member to struct
device_connection, I had to make sure all the existing users consider it.
Quoted 7 lines
The plan is to only support matching with fwnode in the future, so no more device name matching. The software fwnodes that we now have in kernel should make that possible, once we add support for references to them.
The original RFC:
https://lkml.org/lkml/2018/10/24/619
All now merged, thanks.
greg k-h
Heikki Krogerus
maintainerHi Greg,
Quoted 25 lines · Greg Kroah-Hartman
Quoted 24 lines · Heikki Krogerus
Hi,
This is the third version of my proposal to add device graph parsing to the device connection API. There was only one problem reported in v2 by Jun - kernel-doc entry was missing for the new fwnode member in struct
usb_role_switch_desc- and it's now fixed.The second version of the series:
https://lkml.org/lkml/2019/1/30/622The commit message from v1:
This series adds support for OF and ACPI device graph parsing to the device connection API.
Handling the graph is straightforward, but because I'm adding that fwnode member to struct
device_connection, I had to make sure all the existing users consider it.The plan is to only support matching with fwnode in the future, so no more device name matching. The software fwnodes that we now have in kernel should make that possible, once we add support for references to them.
Quoted 4 lines
Quoted 2 lines
The original RFC:
https://lkml.org/lkml/2018/10/24/619All now merged, thanks.
It looks like you have not followed the order of the patches in this series.
You applied at least the patch 4/9 ("device connection: Add fwnode member to struct device_connection") as the last patch to your usb-next branch. The rest of the series, patches starting from 5/9, depend on that patch.
thanks,
Signature · 1 line
heikki
Greg Kroah-Hartman
maintainerQuoted 33 lines · Heikki Krogerus
Hi Greg,
Quoted 25 lines · Greg Kroah-Hartman
Quoted 24 lines · Heikki Krogerus
Hi,
This is the third version of my proposal to add device graph parsing to the device connection API. There was only one problem reported in v2 by Jun - kernel-doc entry was missing for the new fwnode member in struct
usb_role_switch_desc- and it's now fixed.The second version of the series:
https://lkml.org/lkml/2019/1/30/622The commit message from v1:
This series adds support for OF and ACPI device graph parsing to the device connection API.
Handling the graph is straightforward, but because I'm adding that fwnode member to struct
device_connection, I had to make sure all the existing users consider it.The plan is to only support matching with fwnode in the future, so no more device name matching. The software fwnodes that we now have in kernel should make that possible, once we add support for references to them.
Quoted 4 lines
Quoted 2 lines
The original RFC:
https://lkml.org/lkml/2018/10/24/619All now merged, thanks.
Quoted 7 lines
It looks like you have not followed the order of the patches in this series.
You applied at least the patch 4/9 ("device connection: Add fwnode member to struct
device_connection") as the last patch to your usb-next branch. The rest of the series, patches starting from 5/9, depend on that patch.
Ugh, did I just apply them out of order? I can't rewrite my tree now, sorry about that, odd that my sorting didn't work.
I think all of the patches are now in the tree, so I didn't miss anything, is there anything I can do now? Want me to revert them and then add them back in the correct order?
thanks,
greg k-h
Heikki Krogerus
maintainerQuoted 42 lines · Greg Kroah-Hartman
Quoted 33 lines · Heikki Krogerus
Hi Greg,
Quoted 25 lines · Greg Kroah-Hartman
Quoted 24 lines · Heikki Krogerus
Hi,
This is the third version of my proposal to add device graph parsing to the device connection API. There was only one problem reported in v2 by Jun - kernel-doc entry was missing for the new fwnode member in struct
usb_role_switch_desc- and it's now fixed.The second version of the series:
https://lkml.org/lkml/2019/1/30/622The commit message from v1:
This series adds support for OF and ACPI device graph parsing to the device connection API.
Handling the graph is straightforward, but because I'm adding that fwnode member to struct
device_connection, I had to make sure all the existing users consider it.The plan is to only support matching with fwnode in the future, so no more device name matching. The software fwnodes that we now have in kernel should make that possible, once we add support for references to them.
Quoted 4 lines
Quoted 2 lines
The original RFC:
https://lkml.org/lkml/2018/10/24/619All now merged, thanks.
Quoted 7 lines
It looks like you have not followed the order of the patches in this series.
You applied at least the patch 4/9 ("device connection: Add fwnode member to struct
device_connection") as the last patch to your usb-next branch. The rest of the series, patches starting from 5/9, depend on that patch.
Quoted 6 lines
Ugh, did I just apply them out of order? I can't rewrite my tree now, sorry about that, odd that my sorting didn't work.
I think all of the patches are now in the tree, so I didn't miss anything, is there anything I can do now? Want me to revert them and then add them back in the correct order?
The problem is with bisecting, so I'm not sure if reverting helps. I don't think there's anything else that can be done about this now.
thanks,
Signature · 1 line
heikki