According to runtime pm architecture, the kernel side driver should be
as smart as needed to know when the module is idle or active, so that it can
issue the suspend/resume operations to the firmware side at the right time.
To add logics prevents AP from issuing the suspend request to the firmware
when a channel turning to active state, and put it to suspend if the state
is going to inactive with still holding a reference.

Testing Done: Compiled and verified on EVT2 and gpbridge-test module
              with device class daughter board.
Signed-off-by: NKris Huang <huang_kris@projectara.com>
Reviewed-by: NRui Miguel Silva <rui.silva@linaro.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@google.com>

Fix incorrect attribute size when the channel supports fader, as well as
fixing the issue that the attribute list is not null terminated.

Testing Done:
 - Verified by setting brightness and color on red/green/blue leds
   of the device class test board.
Signed-off-by: NDavid Lin <dtwlin@google.com>
Reviewed-by: NRui Miguel Silva <rui.silva@linaro.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@google.com>

The direction value passed to gb_audio_apbridgea_register_cport()
in the gbaudio_module_enable_rx() routine is TX and not RX like
it should be so fix it.

Testing Done: Recorded microphone data from a headset.
Signed-off-by: NMark Greer <mgreer@animalcreek.com>
Fixes: c80e7c6fafa5 ("audio: Split helper APIs based on stream direction")
Tested-by: NVaibhav Agarwal <vaibhav.agarwal@linaro.org>
Reviewd-by: NVaibhav Agarwal <vaibhav.agarwal@linaro.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@google.com>

The svc functions are only supposed to be called by core and should not
be exported.

Most of these functions should never have been exported in the first
place, while a few no longer needs to be (e.g. since core gained support
for offloaded connections).

The only remaining exception is gb_svc_intf_set_power_mode() which is
needed by the camera driver until proper link management is in place.
Signed-off-by: NJohan Hovold <johan@hovoldconsulting.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@google.com>

This reverts commit a0e559d03bbb3f9774a091e31c0f77d98db1f60d.

It breaks the camera driver :(
Signed-off-by: NGreg Kroah-Hartman <gregkh@google.com>

The svc functions are only supposed to be called by core and should not
be exported.

Most of these functions should never have been exported in the first
place, while a few no longer needs to be (e.g. since core gained support
for offloaded connections).
Signed-off-by: NJohan Hovold <johan@hovoldconsulting.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@google.com>

When we receive a update request we shall not trust the cache mechanism
and for that we need a way to invalidate the cache. Add a field that
will control the cache status and refactor the code to check if cache is
valid in a helper function.

This will fix the scenario where an update request is received within
the cache expiration time after a previous update as happened and would
be ignored.
Signed-off-by: NRui Miguel Silva <rui.silva@linaro.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@google.com>

We use the update interval to control the remove path and we set it to
zero when we do not want to have more updates in transit. That means
that the check in the request handler needs to be for interval update
zero to discard the newly received request and not the other way around
like it is.

This will fix the issue that all incoming requests were being discard.
Signed-off-by: NRui Miguel Silva <rui.silva@linaro.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@google.com>

When checking for property changes we may need to act upon that change
besides reporting it using power_supply_changed. So, add a function that
will be call if the specific property changed.

As at it, adjust some indentation of the psy_props_changes array.
Signed-off-by: NRui Miguel Silva <rui.silva@linaro.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@google.com>

While releasing memory during error path exit, invalid memory pointer
was used for dapm_routes. Use a valid one.
Signed-off-by: NVaibhav Agarwal <vaibhav.agarwal@linaro.org>
Reviewed-by: NMark Greer <mgreer@animalcreek.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@google.com>

While mapping control id to define DAPM routes, invalid control index
may cause kernel oops. Add extra check to validate index while mapping
names to control_id.
Signed-off-by: NVaibhav Agarwal <vaibhav.agarwal@linaro.org>
Reviewed-by: NMark Greer <mgreer@animalcreek.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@google.com>

In the beginning, module side can control the vibrator timeout value,
it can disable vibrator until timeout. But after Runtime PM control added
in, AP side didn't know when module can be suspended, the vibrator
task will be interrupted by suspending event.

Because of this problem, the module can not be in charge of counting down
the timeout value, it is now up to the AP to manage this. So add workqueue
to handle the vibrator timeout.
Signed-off-by: NAnn Chen <chen_ann@projectara.com>
Reviewed-by: NViresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@google.com>

In case a bulk-in transfer-buffer allocation failed during probe, we'd
currently leak the corresponding URB.
Signed-off-by: NJohan Hovold <johan@hovoldconsulting.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@google.com>

Remove support for module implementing legacy version of camera bandwidth
requirements specifications, now that all available camera modules have
been updated to use the new version
Signed-off-by: NJacopo Mondi <jacopo.mondi@linaro.org>
Reviewed-by: NLaurent Pinchart <laurent.pinchart@linaro.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@google.com>

The firmware file name's maximum length is set to 56 right now, but it
isn't very much readable how we got to that value (unless someone reads
that line).

Update the comment to show calculations.
Reported-by: NAlex Elder <alex.elder@linaro.org>
Signed-off-by: NViresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@google.com>

Replace 'should' with 'must' to clear the expectation a bit more.
Reported-by: NAlex Elder <alex.elder@linaro.org>
Signed-off-by: NViresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@google.com>

Alex Elder pointed out that the macros also count the trailing NULL
('\0') character and so it should be using SIZE instead of LEN.

This patch makes that change.
Reported-by: NAlex Elder <alex.elder@linaro.org>
Signed-off-by: NViresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@google.com>

Remove the unnecessary parenthesis.
Reported-by: NAlex Elder <alex.elder@linaro.org>
Signed-off-by: NViresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@google.com>

s/shall be used the user/shall be used by the user/
Signed-off-by: NViresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@google.com>

We already have another direct pointer for this, use that instead.

Testing Done: Compiled.
Signed-off-by: NViresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@google.com>

Make the control-protocol disconnecting operation a "core" operation.

This is needed to be able to use the operation in the new connection
tear-down sequence of control connections, which moves disconnecting
after the flush barrier.
Signed-off-by: NJohan Hovold <johan@hovoldconsulting.com>
Reviewed-by: NViresh Kumar <viresh.kumar@linaro.org>
Acked-by: NSandeep Patil <sspatil@google.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@google.com>

Add a host-device cport_shutdown callback which will be called as part
of the new connection tear-down sequence for offloaded connection in
order to do a cport_shutdown operation on behalf of the AP.
Signed-off-by: NJohan Hovold <johan@hovoldconsulting.com>
Reviewed-by: NViresh Kumar <viresh.kumar@linaro.org>
Acked-by: NSandeep Patil <sspatil@google.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@google.com>

Add a host-device cport_clear callback, which will be called as part
of the new connection tear-down sequence to reset the CPort state.
Signed-off-by: NJohan Hovold <johan@hovoldconsulting.com>
Reviewed-by: NViresh Kumar <viresh.kumar@linaro.org>
Acked-by: NSandeep Patil <sspatil@google.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@google.com>

Add a host-device cport_quiesce callback, which will be called as part
of the new connection tear-down sequence to disable flow control after
first making sure that enough peer buffer space is available for the
next messages about to be sent.
Signed-off-by: NJohan Hovold <johan@hovoldconsulting.com>
Reviewed-by: NViresh Kumar <viresh.kumar@linaro.org>
Acked-by: NSandeep Patil <sspatil@google.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@google.com>

Add a host-device cport_connected callback, which will be called after a
connection has been created and that can be used by the host-device
driver to make sure its internal state is updated to match the CPort
attributes set by the SVC.

This callback will eventually replace the cport_features_enable
callback.
Signed-off-by: NJohan Hovold <johan@hovoldconsulting.com>
Reviewed-by: NViresh Kumar <viresh.kumar@linaro.org>
Acked-by: NSandeep Patil <sspatil@google.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@google.com>

I got them wrong, Johan was right, my fault.
Signed-off-by: NGreg Kroah-Hartman <gregkh@google.com>

A backport (commit 79c4de08c0e5a26b04a4ac9e6543dad6379f0b40) was
applied in kernel which adding attribute-group support in led-class.
Remove the LED_HAVE_GROUPS flag check entirely that allow led
drivers to create custom attributes like color/fade_in/fade_out.

Testing Done: Compiled and verified on EVT2 and gpbridge-test module
              with device class daughter board.
Signed-off-by: NKris Huang <huang_kris@projectara.com>
Reviewed-by: NRui Miguel Silva <rui.silva@linaro.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@google.com>

Add intf_oops operation to SVC Protocol. This
operation will notify the AP about a fatal error
in a module. The request has two arguments:
    -u8 intf - the interface in question
    -u8 reason - reason of the error
The response has no payload.
Upon receiving the Request, the driver disables
the Interface.
Signed-off-by: NGeorgi Dobrev <gdobrev@mm-sol.com>
Reviewed-by: NViresh Kumar <viresh.kumar@linaro.org>
Reviewed-by: NAshwin Chaugule <ashwinch@google.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@google.com>

Remove last occurrence of "ara" term in camera driver.
Replace reference to "ara subdevice" with "gmp subdevice"
Signed-off-by: NJacopo Mondi <jacopo.mondi@linaro.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@google.com>

Move the ARPC definitions to their own header.

ARPC is not part of greybus, but is rather an implementation-specific
means of communicating with a certain class of host-device hardware.

Note that the same is true for the APBA USB vendor requests, but we keep
them in the greybus header for the time being.
Signed-off-by: NJohan Hovold <johan@hovoldconsulting.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@google.com>

Make sure to use the common GB_APB prefix for all APBA USB vendor
requests.
Signed-off-by: NJohan Hovold <johan@hovoldconsulting.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@google.com>

Fix copy-paste error in an APBA USB vendor-request comment.
Signed-off-by: NJohan Hovold <johan@hovoldconsulting.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@google.com>

Adjust maximum segment size of the sg list to meet the maximum request
size, this will allow to have less segments which increase the
performance of data movement during transfer operations.

Test Done: using mmc_test with best-case read/write performance and see
that only one segment is used and that the results are better (288KiB vs
300KiB in read operation).
Signed-off-by: NRui Miguel Silva <rui.silva@linaro.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@google.com>

Order of the field of the sdio get caps operation response were wrong,
that influence later the frequencies used by core, during normal
operation.

Tested: verified that the values inserted by the fw are the correct ones
for the field.
Suggested-by: NJackson Chang <jacksonc@bsquare.com>
Signed-off-by: NRui Miguel Silva <rui.silva@linaro.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@google.com>

When enabling an interface, control device should be registered after
having successfully added the timesync. Similarly for the interface
disable path, control device should be removed first before removing
timesync.

Testing Done:
 - Enable and disable the interface
Reviewed-by: NViresh Kumar <viresh.kumar@linaro.org>
Reviewed-by: NJohan Hovold <johan@hovoldconsulting.com>
Suggested-by: NJohan Hovold <johan@hovoldconsulting.com>
Signed-off-by: NDavid Lin <dtwlin@google.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@google.com>

For GB module with jack slot supported, headset/headphone can still
be inserted at the time of module removal. In this case, above layer is
unaware about jack removal event which happened due to module removal.
This may lead to inconsistent state in above HAL layer.  Fix this by
reporting jack removal event explicitly.
Signed-off-by: NVaibhav Agarwal <vaibhav.agarwal@linaro.org>
Reviewed-by: NMark Greer <mgreer@animalcreek.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@google.com>

Modify Lights greybus driver to support runtime PM framework.
The suspend and resume function have been tested with gpbridge-test
image by sysfs. Lights functions work well on suspend/resume.

Testing Done: Compiled and verified on EVT2 and gpbridge-test module
              with device class daughter board.
Signed-off-by: NKris Huang <huang_kris@projectara.com>
Reviewed-by: NViresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@google.com>

Modify SDIO greybus driver to support runtime PM framework.
To enable SDIO runtime PM, it needs to remove MMC_CAP_NEEDS_POLL
and add MMC_CAP2_CORE_RUNTIME_PM in set_host_caps().
The suspend function and resume function have been tested
with micron-sdio image by sysfs. SDIO functions work well
on suspend/resume.

Testing Done: Compiled and verified on EVT2.0 + Micron ARA
              SD module with USB connector
Signed-off-by: NJackson Chang <chang_jackson@projectara.com>
Reviewed-by: NRui Miguel Silva <rui.silva@linaro.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@google.com>

For SPK module, each data connection corresponds to codec DAI. Now
stream state is maintained for each DAI. So, need to maintain stream
state for each DAI/data connection for individual module as well.
Signed-off-by: NVaibhav Agarwal <vaibhav.agarwal@linaro.org>
Reviewed-by: NMark Greer <mgreer@animalcreek.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@google.com>

Runtime streams are required while configuring GB module plugged-in
during active stream. Currently, it is maintained for single stream.
However, this should be maintained for a stream corresponding to each
DAI. Fix this!
Signed-off-by: NVaibhav Agarwal <vaibhav.agarwal@linaro.org>
Reviewed-by: NMark Greer <mgreer@animalcreek.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@google.com>