- 22 Jan, 2016 10 commits
-
-
Johan Hovold authored
Return an ERR_PTR on errors when creating connections. This allows driver probe to fail with a more informative error message as not all connection creation errors are due to memory exhaustion. Reviewed-by:
Viresh Kumar <viresh.kumar@linaro.org> Signed-off-by:
Johan Hovold <johan@hovoldconsulting.com> Signed-off-by:
Greg Kroah-Hartman <gregkh@google.com>
-
Johan Hovold authored
Drop the _dynamic suffix from the exported interface that drivers use to create connections. Reviewed-by:
-
Johan Hovold authored
Simplify the exported generic interface for creating dynamic connections. Since all driver connections will have a bundle we can drop the redundant interface parameter. Reviewed-by:
-
Johan Hovold authored
The protocol-id parameter is only used by the legacy driver so remove it from the generic interface and move it to the legacy driver until it can be removed completely. Reviewed-by:
-
Johan Hovold authored
Add dedicated helper to create control connections. This will allow us to simplify the generic (dynamic) interface. Reviewed-by:
-
Johan Hovold authored
Drop legacy protocol parameter from static connection interface. Reviewed-by:
-
Johan Hovold authored
Now that connection creation has been separated from interface initialisation, we should explicitly check for duplicate CPort descriptors when parsing the manifest. Reviewed-by:
-
Johan Hovold authored
Defer connection creation to bundle driver probe instead of creating them when initialising the interface and parsing the manifest. Store copies of the CPorts descriptors in the bundle for the drivers to use, and update the legacy driver. This is needed for drivers that need more control over host-device resource management, for example, when a protocol needs to use a dedicated host CPort for traffic offloading (e.g. camera data). This also avoids allocating host CPorts for bundles that are not bound to a driver or for remote CPorts that a driver does not need. Reviewed-by:
-
Greg Kroah-Hartman authored
When bundles are added and then removed, we have a race where we go to read the sysfs file, but it is now for a different bundle than the uevent was originally for. So add the bundle class to the uevent so we "know" what the correct bundle class was. Signed-off-by:
Greg Kroah-Hartman <gregkh@linuxfoundation.org> Reviewed-by:
-
Greg Kroah-Hartman authored
This implements the SVC "ping" command. It's tiny and simple, but we need something like this in order for us to "know" if all is working well. Signed-off-by:
-
- 21 Jan, 2016 3 commits
-
-
Greg Kroah-Hartman authored
People are getting confused as to what the firmware file name is for their module, so log it to make it easier to find modules with "blank" vid/pid values and let people know what to name their firmware files. Because we are now logging this, turn the debug message after the firmware request into an error message if something fails. Signed-off-by:
Michael Scott <michael.scott@linaro.org> Signed-off-by:
-
Rui Miguel Silva authored
Add a new input device associated with the SVC to handle key events. This new events are transfer over a new greybus svc operation which is unidirectional. It was selected the KEY_A for representing the KEY_ARA_BUTTON key code. Signed-off-by:
Rui Miguel Silva <rui.silva@linaro.org> Reviewed-by:
-
Greg Kroah-Hartman authored
With the recent changes by Johan to the driver model of the greybus code, the audio_codec no longer should need to provide a "dummy" driver when registering the codec. In fact, having it there causes the greybus core to crash when inserting the module. Removing it all fixes the crash. Signed-off-by:
-
- 19 Jan, 2016 27 commits
-
-
Johan Hovold authored
Allocate private data structure at probe and release at disconnect. The private data will be used to track the connections managed by the driver, but for now only holds a count of the number of cports the bundle has. Reviewed-by:
-
Johan Hovold authored
Use the gb_connection_mutex when making sure a remote CPort is available. This is needed when moving to driver-managed connections that can be created and destroyed at any time. Also check for duplicate bundle-less connections. Reviewed-by:
-
Johan Hovold authored
Serialise connection creation against concurrent creation and destruction using a global mutex. This is needed to prevent two drivers from attempting to create a connection to the same interface CPort and to cope with a racing connection destroy when moving to driver managed connections. Note that the locking can not (easily) be made more fine-grained as not all connections have an interface, but these are not hot paths anyway. Reviewed-by:
-
Johan Hovold authored
Fix longstanding race that could lead to a connection being destroyed while processing an incoming message due to missing reference counting when looking up the connection. Add helpers to manipulate the connection kref. Reviewed-by:
-
Johan Hovold authored
Drop the useless connection_mutex that did not, and can not be used to prevent connection lookup races in atomic context and therefore serves no purpose. Reviewed-by:
-
Johan Hovold authored
Destroy the work queue when the connection is destroyed/deregistered instead of when the last reference is dropped. The work queue is not needed once the connection has been deregistered, and no operations will ever be added to it again (handled by checking connection->state) even if the connection object may not be deallocated until the final reference is dropped. The work-queue name is set based on the host-device name and host-device cport id, something which guarantees a unique name. This would no longer be true if the work queue was not destroyed at connection deregistration as a new connection could then be created for that very same host cport. This is not necessarily a problem unless some work queue features are used that require unique work-queue names, but let's try to be well behaved. Also update an obsolete comment and make explicit that a connection must be disabled before being destroyed. Reviewed-by:
-
Johan Hovold authored
Remove WARN_ON from connection destroy testing for a NULL-connection when destroying a connection. This will allow for simpler driver code when releasing driver-managed connections. Reviewed-by:
-
Johan Hovold authored
Add bundle-version operation to fetch the version of the bundle class. Retrieve the bundle version of all bundles when initialising the interface in case the control-protocol version is greater than 0.1. Reviewed-by:
-
Johan Hovold authored
Remove the broken legacy protocol-version handling from core and move it to the legacy driver instead. Note that version handling has always been broken as legacy protocols were looked up and bound to connections before the connections were enabled and version negotiation could take place (over that very connection). Document that in the legacy driver as well. Reviewed-by:
-
Johan Hovold authored
Store the "negotiated" protocol version to use in the svc state struct instead of the connection struct. The generic concept of a connection version is going away in favour of bundle-class versions. Reviewed-by:
-
Johan Hovold authored
Disable bundle connections in core before calling driver disconnect in case the interface is already gone. This avoids unnecessary timeouts on hot-unplug when a driver does I/O in its disconnect callback. Reviewed-by:
-
Johan Hovold authored
Use the protocol pointer to determine the legacy connection state. This is needed to allow core disable connections when an interface has been hot-unplugged without the legacy protocols leaking its resources. Reviewed-by:
-
Johan Hovold authored
Move legacy protocol and connection handling to the legacy driver. Rename the former global functions using a common legacy_ prefix. Note that all legacy protocols suffer from a connection initialisation race in that the protocol-specific initialisation, which includes allocation of protocol-specific state containers, can not happen until *after* the connection has been enabled. This is a major flaw in the original design that we can now finally address by converting the legacy protocol drivers into proper bundle (class) drivers. Reviewed-by:
-
Johan Hovold authored
Drop dependency on the legacy protocol abstraction. Remove the now unused and last legacy-protocol flag GB_PROTOCOL_SKIP_VERSION along with the protocol-flag feature. Reviewed-by:
-
Johan Hovold authored
Drop dependency on the legacy protocol abstraction. Instead implement the protocol-specific version request directly, and use the new interface for managing the control connection. Note that the version request is being removed from most protocols, but we need to keep the current request for the control protocol as-is indefinitely to maintain backwards compatibility (e.g. with the ES2/ES3 bootrom). Reviewed-by:
-
Johan Hovold authored
Remove control-protocol version from the exported protocol definitions as it is an implementation detail that makes no sense to export. Currently gbsim uses the kernel's control-protocol version definitions directly instead of reporting the version of the protocol it actually implements. Reviewed-by:
-
Johan Hovold authored
Remove the legacy protocol flags that were used to suppress the connected and disconnected events. Instead send the connected and disconnected event for all bundle connections and explicitly exclude static connections and control connections. Reviewed-by:
-
Johan Hovold authored
Remove the legacy protocol flag that was used to suppress sending the control disconnected operation. Instead rely on the fact that no control operations will be sent by core to an interface that has been removed (e.g. after having received a new hotplug event after the bootrom has jumped to the new image). Reviewed-by:
-
Johan Hovold authored
Bundle drivers *must* disable their connections in the disconnect callback, but add a defensive test and warn about buggy drivers nonetheless. Note that bundle drivers would generally release their state containers in disconnect so a failure stop I/O could potentially lead to use-after-free bugs in any late operation completion callbacks. Reviewed-by:
-
Johan Hovold authored
Disable and flush incoming operations before calling driver disconnect. Bundle drivers are still responsible for disabling their connections in their disconnect callback. Note that specifically the legacy protocols must have incoming operations disabled when their connection_exit callback is called as that is where their state is deallocated. Reviewed-by:
-
Johan Hovold authored
Add helper to disable and flush incoming operations. This is intended to be used by core to flush any incoming requests before calling driver disconnect, but could potentially later be exported for driver use as well. Note that we currently flush all incoming operation and allow the request handlers to run, but cancel any responses sent. This may need to be refined later. Reviewed-by:
-
Johan Hovold authored
Add a new connection state ENABLED_TX in which only outgoing operations are allowed. This allows drivers to query the device during probe before allocating their state containers without having to worry about racing incoming requests. Reviewed-by:
-
Johan Hovold authored
Add connection mutex to protect connection enabling and disabling. This is needed to eventually allow drivers to manage the state of their connections during operation (i.e. post probe and pre disconnect). Reviewed-by:
-
Johan Hovold authored
Make sure to cancel all (incoming) operations when failing to enable a connection. Reviewed-by:
-
Johan Hovold authored
Move helper to cancel active operations above gb_connection_enable and simplify locking. Reviewed-by:
-
Johan Hovold authored
Always cancel all operations on connection disable and remove the now unused DESTROYING state. Reviewed-by:
-
Johan Hovold authored
Add a connection request-handler field to struct gb_connection that is set when the connection is enabled. This is a step towards removing the legacy protocol abstraction from core, and will also be used to implement unidirectional connection states (e.g. only outgoing operations are allowed). Reviewed-by:
-
