greybus: connection: fix up error patch logic in gb_connection_create()
Once you have called device_initialize() you have to call put_device()
on the structure to clean it up on an error path, otherwise you will
leak memory.
Alex Elder [Wed, 29 Oct 2014 00:36:00 +0000 (19:36 -0500)]
greybus: introduce protocol abstraction
Define a protocol structure that will allow protocols to be
registered dynamically. For now we just introduce a bookkeeping
data structure. Upcoming patches will move protocol-related methods
into the protocol structure, and will start registering protocol
handlers dynamically.
A list of connections using a given protocol is maintained so we can
tell when a protocol is no longer in use. This may not be necessary
(we could use a kref instead) but it may turn out to be a good way
to clean things up.
The interface is gb_protocol_get() and gb_protocol_put() for a
connection, allowing the protocol to be looked up and the connection
structure to be inserted into its list.
Signed-off-by: Alex Elder <elder@linaro.org> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
Alex Elder [Wed, 29 Oct 2014 00:35:59 +0000 (19:35 -0500)]
greybus: use protocol_id for numeric values
Switch to using "protocol_id" to refer to a byte-sized numeric
protocol number. A "protocol" will represent a protocol structure
(created in the next patch).
Signed-off-by: Alex Elder <elder@linaro.org> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
Alex Elder [Wed, 29 Oct 2014 00:35:58 +0000 (19:35 -0500)]
greybus: drop the cport id on error
In gb_connection_create(), if an error occurs adding a connection's
device, the cport id assigned to the AP end of the connection is not
getting freed. Fix that.
Signed-off-by: Alex Elder <elder@linaro.org> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
Matt Porter [Tue, 28 Oct 2014 14:08:13 +0000 (10:08 -0400)]
greybus: connection: call connection_init hook after setting the handler
In gb_connection_init() we set the connection_handler for each supported
protocol, but we never call the connection_init hook after doing so. This
results in a failure being returned so fix it by calling the connection_init
hook to get a good return and the associated driver initialized.
Signed-off-by: Matt Porter <mporter@linaro.org> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
Matt Porter [Tue, 28 Oct 2014 14:08:12 +0000 (10:08 -0400)]
greybus: module: set device_id when initializing an interface
gb_module_interface_init() looks for the interface corresponding
to the supplied interface_id, but fails to configure the
device_id that goes with it. This results in a set route command
being set with an uninitialized and bogus value. Fix it.
Signed-off-by: Matt Porter <mporter@linaro.org> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
We aren't using an id_table for "drivers" at this moment, as the whole
driver model interaction is under heavy rework. So remove these for now
to keep things easier to understand for future patches.
Alex Elder [Mon, 27 Oct 2014 11:04:30 +0000 (06:04 -0500)]
greybus: begin abstracting connection operations
This is part 1 of abstracting the connection operations into a set
of methods. This will avoid some big switch statements, but more
importantly this will be needed for supporting multiple versions of
each protocol.
For now only two methods are defined. The init method is used
to set up the device (or whatever the CPort represents) and the exit
method tears it down. There may need to be additional operations
added in the future, and once versioning is used we might stash
the version number in this structure as well.
The next patch adds dynamic registratration of these protocol
handlers, and will do away with the switch statement now found
in gb_connection_init().
Signed-off-by: Alex Elder <elder@linaro.org> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
greybus: module: remove unneeded XXX comment about module id values
We do properly check for duplicate module ids, as fixed in 008d85d90ae1ab31f1f7b80f245f6ee2eb5aed49 "module: don't create duplicate
module ids", so remove the XXX marker.
greybus: FIXME/XXX removals: We have proper reference counting now
Now that we have proper reference counting for modules, interfaces, and
connections, no need to worry about grabbing a pointer to your "parent"
structure, all is good.
The uart-gb code needs to init the tty core before it can add devices.
Previously we hard-coded this in the greybus core, move this to be
"dynamic" and self-contained within the uart-gb.c file.
greybus: operation: make the timeout a per-operation thing, not per-connection
An operation is what can timeout, not a connection itself. So notify
the operation timedout, and the connection can then do with it as it
sees fit, if necessary.
When registering a host controller, verify that all of the needed
callbacks are present, so we don't have to do the testing on any "hot"
paths when we want to send real data.
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com> Reviewed-by: Alex Elder <elder@linaro.org.
Hook up gbuf_kill_gbuf() by implementing yet-another-host-controller
callback and a gbuf-specific pointer to hold the tracking data the hcd
needs in order to be able to abort a transfer.
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com> Reviewed-by: Alex Elder <elder@linaro.org>
If we somehow get a hotplug event for a module id that we already have
created[1], don't try to create it again, or sysfs will complain loudly.
Instead, abort the creation properly.
[1] If, for example, you happened to run a script on a greybus emulator
twice in a row...
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com> Reviewed-by: Alex Elder <elder@linaro.org>
We were thinking that module attributes were known at the time the
device was created in the system, so we could query them to know if the
sysfs file was present or not. Unfortunatly that's not the case, we
create the device before we parse the values, so just always show the
sysfs attributes. If there is no such attribute, the sysfs file will be
empty (i.e. for the string attributes.)
This patch hooks up modules, interfaces, and connections to the driver
model. Now we have a correct hierarchy, and drivers can be correctly
bound to the proper portions in the future. Devices are correctly
reference counted and torn down in the proper order on removal of a
module.
Some basic sysfs attributes have been created for interfaces and
connections. Module attributes are not working properly, but that will
be fixed in future changes.
This has been tested on Alex's machine, with multiple hotplug and unplug
operations of a module working correctly.
Alex Elder [Wed, 22 Oct 2014 07:04:31 +0000 (02:04 -0500)]
greybus: make svc_set_route_send() public
Give svc_set_route_send() non-private scope so it can be used by a
function outside "ap.c" in the next patch. Change its type so it
can tell its caller if an error occurs.
Signed-off-by: Alex Elder <elder@linaro.org> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
Alex Elder [Wed, 22 Oct 2014 07:04:30 +0000 (02:04 -0500)]
greybus: define connection state
Define the state of a connection. A connection will not be
enabled until it has been successfully set up. Once it starts
getting torn down its state will move to "being destroyed".
Don't send any operation request messages unless the connection is
enabled. And drop any incoming messages if if the connection is
not enabled.
Signed-off-by: Alex Elder <elder@linaro.org> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
Alex Elder [Wed, 22 Oct 2014 07:04:29 +0000 (02:04 -0500)]
greybus: define operation_cancel()
Define a new function operation_cancel() that cancels an
outstanding operation. Use it to clear out any operations that
might be pending at the time a connection is torn down.
Note: This code isn't really functional yet, partially because
greybus_kill_gbuf() is not implemented.
Signed-off-by: Alex Elder <elder@linaro.org> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
Alex Elder [Wed, 22 Oct 2014 07:04:28 +0000 (02:04 -0500)]
greybus: time out operation requests
Arrange for operation requests that takke too long to time out.
At the moment, nothing happens when that occurs (other than a silly
message getting printed). When the connection and operation and
interface and module code are cleaned up properly, this event should
most likely cause the affected module to get torn down.
Signed-off-by: Alex Elder <elder@linaro.org> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
Matt Porter [Wed, 22 Oct 2014 06:06:09 +0000 (02:06 -0400)]
greybus: ap: add svc_set_route_send() command and use it on a link up event
When the AP receives a link up event, request that the SVC set a
route to the interface's device id (this device id has been
previously reported to the AP). In the future, we may not always
immediately set a route upon receiving a link up event but this
is sufficient for the known use cases at this time.
Signed-off-by: Matt Porter <mporter@linaro.org> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
Matt Porter [Wed, 22 Oct 2014 06:06:08 +0000 (02:06 -0400)]
greybus: svc: remove unneeded fields from the unipro set route message payload
CPort connections are being handled in the application layer connection
protocol and the layer 3 switch doesn't care about them. Also, the
switch doesn't care about a source device id when setting up the route
table. Reduce the message to just the necessary destination device ID.
As the SVC is aware of which switch port it found the module/interface
and assigned the device ID, we can simply tell the SVC to set a route
to the device ID it has reported to the AP as being active.
Signed-off-by: Matt Porter <mporter@linaro.org> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
Matt Porter [Wed, 22 Oct 2014 02:43:31 +0000 (22:43 -0400)]
greybus: ap: process the UniPro link up message
The link up message is the event that tells the AP what device ID
has been assigned to a particular interface on a module during
enumeration. The link up is sent *only* after the hotplug event
for a particular module has been sent to the AP.
The link up payload must carry the Module ID and Interface ID
to uniquely identify the struct gb_interface to which the
Device ID has been assigned.
After processing of the link up message, the interface's device_id
field will contain the assigned Device ID so that the AP has the
information necessary to issue network route commands.
Signed-off-by: Matt Porter <mporter@linaro.org> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
greybus: battery-gb: Allow kernel values to get out of sync with greybus spec
We can't know that the greybus values and the kernel values for a number
of battery enumerated types will remain in sync. And as theses are sent
by an external device from the kernel, we have to explicitly check these
values.
Matt Porter [Tue, 21 Oct 2014 05:52:27 +0000 (01:52 -0400)]
greybus: ap: add support for the AP Device ID unipro management function message
The AP needs to know its assigned Device ID in order to establish
Greybus connections between CPorts. We could have pulled the Device
ID from the controller hardware in a driver specific manner, but
instead we define one generic message from the SVC to let the
AP know this information. Add this additional unipro management
message and handle it by setting the supplied Device ID in the
struct greybus_host_device. The greybus core will use this to
populate the source Device ID when establishing a connection
between the AP and another module's CPort.
Signed-off-by: Matt Porter <mporter@linaro.org> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
Alex Elder [Tue, 21 Oct 2014 04:01:04 +0000 (23:01 -0500)]
greybus: improve module cleanup code
When a module gets destroyed all of its state and the state of its
interfaces and connections (etc.) need to be torn down. This is
not now being done properly. Add this teardown code.
Signed-off-by: Alex Elder <elder@linaro.org> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
Alex Elder [Mon, 20 Oct 2014 15:28:00 +0000 (10:28 -0500)]
greybus: leverage ap_disconnect() in ap_probe()
With a few minor changes, ap_disconnect() can correctly handle
cleaning up even a partially initialized USB interface. Make those
changes, and then use ap_disconnect() to simplify cleanup for all
the error paths in ap_probe(). Reset all fields as they're cleaned
up to facilitate debugging.
Alex Elder [Mon, 20 Oct 2014 15:27:59 +0000 (10:27 -0500)]
greybus: move ap_disconnect()
The next patch has ap_probe() reference ap_disconnect(). To prepare
for that, move ap_disconnect() up in the file.
This is done as a separate commit to make it easier to see this move
involves no other change to that function. This and the next commit
can be squashed if desired.
Signed-off-by: Alex Elder <elder@linaro.org> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
Alex Elder [Mon, 20 Oct 2014 15:27:58 +0000 (10:27 -0500)]
greybus: remove cports and strings from gb_module
We no longer keep copies of strings found in the manifuest in
a module's strings array, so we can get rid of the strings array.
Similarly, the new manifest parsing code sets up connections for
each cport id advertised for a module, so the cport array is
no longer needed either.
Signed-off-by: Alex Elder <elder@linaro.org> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
Alex Elder [Mon, 20 Oct 2014 15:27:57 +0000 (10:27 -0500)]
greybus: drop gb_* device fields from gb_module
A struct gb_module has a bunch of fields from the earlier skeleton
code, where a module was assumed to possibly have one of every
type of device available on the GP Bridge. The manifest parsing
code changed it so these things will be related to connection
endpoints, so these gb_module fields are no longer needed.
A few of these (battery and sdio) haven't been implemented the "new
way" yet, so just leave a bit of the code that was there commented
out for now.
Also, gb_tty seems to be partially implemented and I don't want to
remove that without knowing where it's headed, so that one stays.
Signed-off-by: Alex Elder <elder@linaro.org> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
Matt Porter [Mon, 20 Oct 2014 10:39:45 +0000 (06:39 -0400)]
greybus: gpio-gb: fix offset error checking and usage
Offset (or hwgpio num) is the offset within a gpiochip, not the
unique gpio namespace number. Adjust the error checking and use
of offset in our operation calls to fix this.
Signed-off-by: Matt Porter <mporter@linaro.org> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
greybus: gpio-gb: allow it to build properly for all current kernel versions.
GPIO remove changed the api for 3.17 to try to make up for some
previously foolish design decisions. Handle that in kernel_ver.h to
make the code simple.
Alex Elder [Fri, 17 Oct 2014 10:18:22 +0000 (05:18 -0500)]
greybus: update gbuf status for completion handlers
Currently, if a USB urb completes with an error, that error status
is not transferred back to the gbuf that it's associated with. For
inbound data there's not a lot we can do about an error, but for
outbound data, this means there is no notification to the submitter
that something went wrong.
For outbound data copy the urb status directly back to the gbuf as
its status. Follow USB's lead and set the status to -EINPROGRESS
while a gbuf is "in flight." Assign a gbuf an initial status value
of -EBADR to help identify use of never-set status values.
When an inbound urb fails (SVC or CPort), currently the urb is just
leaked, more or less (i.e., we lose an urb posted to receive
incoming data). Change that so such an error is reported, but
then re-submitted.
Signed-off-by: Alex Elder <elder@linaro.org> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
Alex Elder [Fri, 17 Oct 2014 10:09:21 +0000 (05:09 -0500)]
greybus: add write retry support for i2c
It is expected that i2c writes may fail, and in that case the driver
simply retries some number of times before actually treating it as a
failure. Define a GB_OP_RETRY status, which is interpreted by the
i2c driver as an indication a retry is in order. We just translate
that into an EAGAIN error passed back to the i2c core.
Signed-off-by: Alex Elder <elder@linaro.org> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
Alex Elder [Thu, 16 Oct 2014 11:35:38 +0000 (06:35 -0500)]
greybus: add i2c driver
This patch adds the i2c driver, based on the use of Greybus operations
over Greybus connections. It basically replaces almost all of what
was previously found in "i2c-gb.c".
When gb_connection_device_init(connection) is called, any connection
that talks the GREYBUS_PROTOCOL_I2C is passed to gb_i2c_device_init()
to be initialized.
Initialization involves verifying the code is able to support the
version of the protocol. For I2C, we then query the functionality
mask, and set the retry count and timeout to default values.
After that, we set up the i2c device and associate it with the
connection. The i2c_algorithm methods are then implemented
by translating them into Greybus operations.
Signed-off-by: Alex Elder <elder@linaro.org> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
Alex Elder [Thu, 16 Oct 2014 11:35:37 +0000 (06:35 -0500)]
greybus: kill old cport handlers
The original CPort message handlers are not needed. All incoming
data is passed to handlers based on the protocol used over the
connection over which the data was transferred. So get rid of the
old CPort handler code.
Signed-off-by: Alex Elder <elder@linaro.org> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
Alex Elder [Thu, 16 Oct 2014 11:35:33 +0000 (06:35 -0500)]
greybus: let operation layer examine incoming data
Give the operation layer a chance to examine incoming data so that
it can handle it appropriately.
Treat the data as an operation message header. If it's a response,
look up the operation it's associated with. If it's not, create a
new operation. Copy the incoming data into the request or response
buffer. The next patch adds a work queue to pick up handling
the request or response from there.
Get rid of gb_operation_submit(). Instead, we have two functions,
one for sending an operation's request message, the other for
sending an operation's response message.
Not fully functional yet, still just filling things in...
Signed-off-by: Alex Elder <elder@linaro.org> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
Alex Elder [Thu, 16 Oct 2014 11:35:32 +0000 (06:35 -0500)]
greybus: add gb_operation_find()
Add a red-black tree indexed by operation id to a connection to
allow pending operations (whose requests are in-flight) to be
found when their matching response is recieved.
Assign the id at the time an operation is inserted, and update
the operation's message header(s) to include it.
Rename gb_connection_op_id() to be more consistent with the
naming conventions being used elsewhere.
(Noting now that this may switch to a simple list implementation
based on Greg's assertion that lists are faster than red-black trees
for up to a few hundred entries.)
Signed-off-by: Alex Elder <elder@linaro.org> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
Alex Elder [Thu, 16 Oct 2014 11:35:31 +0000 (06:35 -0500)]
greybus: add response buffer to an operation
We need to track both request messages and response messages in
operations. So add another gbuf (and payload pointer) field to
the operation structure, and rename them to indicate which one
is which. Allow the creator specify the size of the response
buffer; just leave it a null pointer if the size is 0.
Define a new helper function gb_operation_gbuf_create() to
encapsulate creating either a request or a response buffer.
Any buffer associated with a connection will (eventually) have been
created as part of an operation. So stash the operation pointer in
the gbuf as the context pointer. Whether a buffer is for the
request or the response can be determined by pointer comparison.
Signed-off-by: Alex Elder <elder@linaro.org> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
Alex Elder [Thu, 16 Oct 2014 11:35:30 +0000 (06:35 -0500)]
greybus: stop init_subdevs stuff
Upcoming patches are going to set up devices based on what is
discovered in the module manifest. Get rid of the hard-coded
initialization done by gb_init_subdevs(), along with other related
code.
Signed-off-by: Alex Elder <elder@linaro.org> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
Alex Elder [Thu, 16 Oct 2014 11:35:29 +0000 (06:35 -0500)]
greybus: drop hd arg from submit_gbuf method
Every gbuf is associated with a connection when it is created. And
a connection contains a pointer to the host device that will carry
messages. So there's no need for the submit_gbuf() method to have
the host device pointer passed to it, the function can get it from
the gbuf's connection.
Signed-off-by: Alex Elder <elder@linaro.org> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
Alex Elder [Thu, 16 Oct 2014 11:35:27 +0000 (06:35 -0500)]
greybus: don't restrict input buffer size
Don't assume that input buffers have any particular content. The
only thing the gbuf layer needs to be concerned with is the presence
of the cport_id byte at the beginning of a transfer.
Signed-off-by: Alex Elder <elder@linaro.org> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
Alex Elder [Thu, 16 Oct 2014 11:35:25 +0000 (06:35 -0500)]
greybus: unlock gbuf mutex on free
To drop a reference on a gbuf, greybus_free_gbuf() is called. That
uses kref_put_mutex() to drop the refernce under protection of
gbuf_mutex. However the release routine, free_gbuf(), never
releases the mutex as it should. Fix that.
Signed-off-by: Alex Elder <elder@linaro.org> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
Matt Porter [Mon, 13 Oct 2014 07:00:53 +0000 (03:00 -0400)]
greybus: es1-ap-usb: handle -EPROTO in check_urb_status()
On a disconnect we can also have a status of -EPROTO. This results in
a flood of error messages due to the -EAGAIN handling of unsupported
status results. Fix this by also returning status when we have -EPROTO.
Signed-off-by: Matt Porter <mporter@linaro.org> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
Matt Porter [Mon, 6 Oct 2014 13:58:44 +0000 (09:58 -0400)]
greybus: fix manifest parsing problem with descriptor payload
The internal struct manifest_desc needs the data payload, rather
than the entire descriptor with header to be populated into the
data field. Also fix two places where the parser was trying to
extract the entire descriptor with header for the data payload
field.
Signed-off-by: Matt Porter <mporter@linaro.org> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
Alex Elder [Mon, 6 Oct 2014 11:53:13 +0000 (06:53 -0500)]
greybus: get rid of gbuf->hdpriv
Last time I tried to kill off gbuf->context my efforts were shot
down. Now that I've got the connection infrastructure in place,
maybe I'll have more luck getting rid of gbuf->hdpriv.
The only place it's used is to stash the es1_ap_dev structure
pointer in the buffer. But that information is now available
through the buffer's connection, so we don't need to use the
hdpriv field any more.
So get rid of it, and use hd_to_es1(gbuf->connection->hd) to
get at what we need.
Signed-off-by: Alex Elder <elder@linaro.org> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
projects / karo-tx-linux.git / log