This is preparation for support for different operational states
(defined by arche_platform_state) in the driver, to enable
user/developer to dynamically configure the state.
arche_platform_coldboot_seq() fn will be responsible for
rebooting SVC device.
Vaibhav Hiremath [Fri, 12 Feb 2016 20:34:02 +0000 (02:04 +0530)]
greybus: arche-platform: make apb_state common to both platform drivers
Make 'enum apb_state' common to both platform drivers, so that
both drivers can make use of same state and user will have
unified control configuration across devices (SVC, APB1 and APB2)
Testing Done: Tested on EVT1.2 and DB3.5 platform.
Michael Scott [Tue, 9 Feb 2016 01:08:46 +0000 (17:08 -0800)]
greybus: interface: clear upper 16-bits of version_id and product_id
Current userspace looks through the sysfs interface entries for
matching vendor_id and product_id any time an interface is opened
by module developers. The upper 16-bits of ES3 vendor_id and
product_id contain a reverse mask of the lower 16-bits. This
additional information is never used and should be removed so
that every consumer of these sysfs entries doesn't have to perform
the same bit clearing logic.
Signed-off-by: Michael Scott <michael.scott@linaro.org> Reviewed-by: Alex Elder <elder@linaro.org> Reviewed-by: Viresh Kumar <viresh.kumar@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
greybus: spi: add device_type field to device config
Add device_type field in device config operation to get the type of
device and try to expose less the kernel internal over greybus.
This include the spidev, spi-nor will fetch the correct nor id over
jede and a modalias that will have the previous behavior (name will set
the driver to be loaded).
As at it, fix a trivial error path and return immediately.
Tested: using gbsim and confirming that a spidev and mtd device were
created.
Signed-off-by: Rui Miguel Silva <rui.silva@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
When inserting and removing a module with a UART protocol defined a
double free of the tty_port would happen and that would generate a lot
of kernel oops in different places related to memory corruption.
Signed-off-by: Rui Miguel Silva <rui.silva@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
Rework the get property descriptors function to fix a memory handling
error for the response structure. This could corrupt the stack and
throw nonalignment PC or SP error:
Jacopo Mondi [Fri, 5 Feb 2016 08:35:32 +0000 (09:35 +0100)]
greybus: camera: Reset link speed on failed config
Improve the management of unipro power mode changes in response
to a configure_stream operation.
When sending a "test only" request to camera module, do not change power
mode to HS-G2 as no frame will be actually transmitted.
When receiveing an "adjusted" configuration response, reset power
mode to PWM-G1.
There's a definition missing in audio_manager causing the kernel build
to fail:
CC [M] ./greybus/audio_manager.o
./greybus/audio_manager.c:22:8: warning: type defaults to 'int' in declaration of 'DEFINE_IDA' [-Wimplicit-int]
error, forbidden warning: audio_manager.c:22
./kernel/scripts/Makefile.build:308: recipe for target './greybus/audio_manager.o' failed
Vaibhav Agarwal [Thu, 28 Jan 2016 15:45:40 +0000 (21:15 +0530)]
greybus: audio: use variable 'is_connected' to maintain module state
there is race condition between _disconnect() request &
stop_trigger() in case of abrupt module removal.
And sometimes this can lead to deadlock while acquiring
codec_info->lock.
To avoid such situation, atomic variable is used to maintain
codec connected state.
During dai operations (trigger, shutdown, etc.), 'is_connected'
variable is validated to avoid unnecessary lock acquire in
case module already removed.
Vaibhav Agarwal [Thu, 28 Jan 2016 15:45:39 +0000 (21:15 +0530)]
greybus: audio: codec driver cleanup
audio codec driver is now moved to bundle driver approach.
This resolved many race conditions related to audio mgmt &
data connection init/exit sequence.
Thus, a lot of helper functions can now be safely removed.
Viresh Kumar [Thu, 28 Jan 2016 10:20:48 +0000 (15:50 +0530)]
greybus: connection: Fix sparse warnings around locking
The callers ensures that connection->lock is taken before calling few
routines, but that isn't enough for sparse as it sees an unexpected
unlock.
greybus/connection.c:380:29: warning: context imbalance in 'gb_connection_cancel_operations' - unexpected unlock
Fix that adding __must_lock() attribute to the function declaration.
This also adds the attribute for
gb_connection_flush_incoming_operations(), which isn't showing any
sparse warnings with the current state of code, but with minor
rearrangements of the code.
Viresh Kumar [Thu, 28 Jan 2016 10:20:47 +0000 (15:50 +0530)]
greybus: audio: Fix sparse warnings
greybus/audio_apbridgea.c:13:5: warning: symbol 'gb_audio_apbridgea_set_config' was not declared. Should it be static?
greybus/audio_apbridgea.c:30:5: warning: symbol 'gb_audio_apbridgea_register_cport' was not declared. Should it be static?
greybus/audio_apbridgea.c:44:5: warning: symbol 'gb_audio_apbridgea_unregister_cport' was not declared. Should it be static?
greybus/audio_apbridgea.c:58:5: warning: symbol 'gb_audio_apbridgea_set_tx_data_size' was not declared. Should it be static?
greybus/audio_apbridgea.c:72:5: warning: symbol 'gb_audio_apbridgea_get_tx_delay' was not declared. Should it be static?
greybus/audio_apbridgea.c:80:5: warning: symbol 'gb_audio_apbridgea_start_tx' was not declared. Should it be static?
greybus/audio_apbridgea.c:94:5: warning: symbol 'gb_audio_apbridgea_stop_tx' was not declared. Should it be static?
greybus/audio_apbridgea.c:106:5: warning: symbol 'gb_audio_apbridgea_set_rx_data_size' was not declared. Should it be static?
greybus/audio_apbridgea.c:120:5: warning: symbol 'gb_audio_apbridgea_get_rx_delay' was not declared. Should it be static?
greybus/audio_apbridgea.c:128:5: warning: symbol 'gb_audio_apbridgea_start_rx' was not declared. Should it be static?
greybus/audio_apbridgea.c:141:5: warning: symbol 'gb_audio_apbridgea_stop_rx' was not declared. Should it be static?
greybus/audio_gb.c:14:5: warning: symbol 'gb_audio_gb_get_topology' was not declared. Should it be static?
greybus/audio_gb.c:48:5: warning: symbol 'gb_audio_gb_get_control' was not declared. Should it be static?
greybus/audio_gb.c:70:5: warning: symbol 'gb_audio_gb_set_control' was not declared. Should it be static?
greybus/audio_gb.c:85:5: warning: symbol 'gb_audio_gb_enable_widget' was not declared. Should it be static?
greybus/audio_gb.c:97:5: warning: symbol 'gb_audio_gb_disable_widget' was not declared. Should it be static?
greybus/audio_gb.c:109:5: warning: symbol 'gb_audio_gb_get_pcm' was not declared. Should it be static?
greybus/audio_gb.c:133:5: warning: symbol 'gb_audio_gb_set_pcm' was not declared. Should it be static?
greybus/audio_gb.c:150:5: warning: symbol 'gb_audio_gb_set_tx_data_size' was not declared. Should it be static?
greybus/audio_gb.c:163:5: warning: symbol 'gb_audio_gb_get_tx_delay' was not declared. Should it be static?
greybus/audio_gb.c:183:5: warning: symbol 'gb_audio_gb_activate_tx' was not declared. Should it be static?
greybus/audio_gb.c:195:5: warning: symbol 'gb_audio_gb_deactivate_tx' was not declared. Should it be static?
greybus/audio_gb.c:207:5: warning: symbol 'gb_audio_gb_set_rx_data_size' was not declared. Should it be static?
greybus/audio_gb.c:220:5: warning: symbol 'gb_audio_gb_get_rx_delay' was not declared. Should it be static?
greybus/audio_gb.c:240:5: warning: symbol 'gb_audio_gb_activate_rx' was not declared. Should it be static?
greybus/audio_gb.c:252:5: warning: symbol 'gb_audio_gb_deactivate_rx' was not declared. Should it be static?
Fix them by including the header that declares the exported routines.
greybus: audio_manager: use an 'ida' for the module id
Every time we hotplug an audio module, we get a new audio module id. We
should recycle them instead of just constantly incrementing the number
so we don't see things like:
[178016.832580] Created audio module #6124
in the kernel logs.
Johan Hovold [Fri, 29 Jan 2016 14:42:31 +0000 (15:42 +0100)]
greybus: firmware: convert to bundle driver
Convert the legacy firmware protocol driver to a bundle driver.
This also fixes a potential crash should a (malicious) module have sent
an early request before the private data had been initialised.
Note that the firmware protocol needs to support the version request
indefinitely since it has been burnt into ROM.
In order to avoid having to update current module-loading scripts, keep
this driver internal to greybus core at least until modalias support is
added.
Note that there is no MODULE_DEVICE_TABLE defined for firmware as we
cannot have two greybus tables in one module on ancient 3.10 kernels and
that the legacy driver is currently also internal to core. This needs be
added once the driver can be built as a module.
Johan Hovold [Fri, 29 Jan 2016 14:42:30 +0000 (15:42 +0100)]
greybus: legacy: remove unimplemented classes from id table
Remove the unimplemented display, sensor, and svc classes from the
device-id table.
As Viresh noted the SVC protocol is special and having an SVC class
doesn't really make sense at all. Either way, the SVC protocol is
already implemented by core.
Michael Scott [Thu, 28 Jan 2016 00:41:01 +0000 (16:41 -0800)]
greybus: arche-platform: reduce wait between WAKE_OUT checks
SVC WAKE_OUT loop is estimated 400ms during which wake_detect line
is pulled low for AP detection. On AP side we have 500ms delay
between checks. To avoid timing issues, reduce delay between
checks and raise total # of checks so that overall time for sequence
is the same.
Testing Done:
- Used for DB3.5/EVT1.5 hardware during bringup
- Regression tested on DB3.1+ES2, DB3.1+ES3
Signed-off-by: Michael Scott <michael.scott@linaro.org> Reviewed-by: Vaibhav Hiremath <vaibhav.hiremath@linaro.org> Tested-by: Vaibhav Hiremath <vaibhav.hiremath@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
Michael Scott [Thu, 28 Jan 2016 00:41:00 +0000 (16:41 -0800)]
greybus: arche-platform: assert wake-detect to complete WAKE_OUT event
After SVC generates WAKE_OUT (pulling wake-detect pin low) and APB is
brought out of reset, we need to assert wake-detect again to complete
SVC WAKE_OUT logic.
Testing Done:
- Used for DB3.5/EVT1.5 hardware during bringup
- Regression tested on DB3.1+ES2, DB3.1+ES3
Signed-off-by: Michael Scott <michael.scott@linaro.org> Reviewed-by: Vaibhav Hiremath <vaibhav.hiremath@linaro.org> Tested-by: Vaibhav Hiremath <vaibhav.hiremath@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
Michael Scott [Thu, 28 Jan 2016 00:40:58 +0000 (16:40 -0800)]
greybus: arche-platform: ensure wake-detect pin is deasserted
During DB3.5 bringup, it was noted that wake_detect signal was not
properly generating SVC edge IRQ. To ensure signal goes from low
to high correctly, let's bring signal low (regardless of default
pin state).
Testing Done:
- Used for DB3.5/EVT1.5 hardware during bringup
- Regression tested on DB3.1+ES2, DB3.1+ES3
Signed-off-by: Michael Scott <michael.scott@linaro.org> Reviewed-by: Vaibhav Hiremath <vaibhav.hiremath@linaro.org> Tested-by: Vaibhav Hiremath <vaibhav.hiremath@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
Johan Hovold [Thu, 28 Jan 2016 11:43:29 +0000 (12:43 +0100)]
greybus: hd: fix host-device-removal race
Make sure to tear down the svc and flush any on-going hotplug processing
before removing the remaining interfaces.
This fixes crashes due to host-device removal racing with svc
hotplug/unplug processing (e.g. at "UniPro restart").
Testing Done:
Verified that this fixes crashes reproducible on SDB when unloading the
host-device driver module while generating hotplug/unplug events.
Signed-off-by: Johan Hovold <johan@hovoldconsulting.com> Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
Viresh Kumar [Wed, 27 Jan 2016 05:47:00 +0000 (11:17 +0530)]
greybus: audio: Rename Audio class and remove the unused one
There should be a single class macro for Audio and two protocol macros.
Rename class with value 0x12 as GREYBUS_CLASS_AUDIO and remove the other
unused class GREYBUS_CLASS_AUDIO_DATA.
Viresh Kumar [Wed, 27 Jan 2016 11:27:48 +0000 (16:57 +0530)]
greybus: audio_codec: convert to bundle driver
Convert the legacy audio management and data protocol drivers to a
bundle driver.
The Audio bundle driver can support a single management and any number
of data cports, and so we expect multiple data cports to be present for
the bundle during initialization.
When gb_protocol_get() fails in legacy_connection_create(), we end up
bailing-out before assigning lc->connection and lc->protocol. Calling
legacy_connection_destroy() in that case results in a null pointer
dereference.
Check if lc->connection is not null before freeing it.
greybus: svc watchdog: allow it to be enabled/disabled from userspace
Sometimes you want to disable the SVC watchdog without having to patch
your kernel, so provide a sysfs file to do this. This can let us do
power measurements, and let the firmware developers not go crazy
worrying that the kernel is going to reset their chips with no notice.
greybus: camera: only build module against msm kernel
The camera driver currently does not build against anything other than
the msm kernel, due to cross-dependancies, so enable that here so that
we can build against other kernels without failing the build.
Sometimes the ping response comes back _right_ after we timed out, as
the svc got its act together and squeaked out the ack, yet we miss it
and reset the whole bus. Double the delay to hopefully give the svc a
little more of a chance to fix itself. Odds are, it's still in trouble,
but we can just hold off resetting it for a bit more...
Vaibhav Hiremath [Tue, 26 Jan 2016 01:53:29 +0000 (07:23 +0530)]
greybus: arche-platform: Disable clock as part of driver remove
As part of driver remove (cleanup) function, disable the clock for both
SVC, APB1 & APB2.
Testing Done:
Tested on EVT1 platform with Connect=>disconnect=>connect
iteration, almost close to 100 iterations have passed (demo branch).
And also tested with kernel-only build.
Unipro network speed was increased at camera initialization time and
never slowed down.
This unnecessary drains power during the entire time camera module is
plugged in.
Increasing/decreasing unipro link speed before issuing stream
configuration request to camera module prevents this from happening.
greybus: svc: add a "watchdog" to check the network health
Now that we have a svc ping command, let's add a watchdog to call it
every so often (1 second at the moment.) If it finds something went
wrong, post a stern message to the kernel log and call:
start unipro_reset
to reset the whole greybus hardware subsystem.
Viresh Kumar [Fri, 22 Jan 2016 10:46:08 +0000 (16:16 +0530)]
greybus: svc: Expose and retain VID/PID received from bootrom for ES2
ES2 chips doesn't have efuses for storing module's vendor_id and
product_id and so we have hacked bootrom earlier using firmware
protocol, so that VID/PID can be received for fetching firmware
packages.
Another requirement is to expose them to sysfs, so that modules can be
identified properly.
That can be easily solved by updating interface's VID/PID, when fetched
using firmware protocol and later reusing them while the module switches
its identity from bootrom to firmware.
Do that only if the module is ES2 and the VID/PID sent during hotplug
are both 0.
Horrid hack, but a start at making it easier for people to see what is
currently connected without having to deal with poking around in sysfs
directly.
Vaibhav Agarwal [Thu, 21 Jan 2016 17:03:13 +0000 (22:33 +0530)]
greybus: audio_codec: update codec_name as per driver->name
Originally, driver->name was not poluated from GB and thus
manually set from audio_codec driver as a hack.
This is no more required.
Another patch already removes that hack.
Now, with new driver->name as "legacy.<id-bundle.interface>"
codec is registered with different name.
So, during DAI link registration as well it needs modification.
Johan Hovold [Thu, 21 Jan 2016 16:34:20 +0000 (17:34 +0100)]
greybus: legacy: look up protocol at connection creation
Look up legacy protocol at connection create instead of at init.
Note that we can now look up the protocol before even creating the
actual connection. This is needed to be able to set a connection request
handler when creating the connection rather than when enabling it.
Johan Hovold [Thu, 21 Jan 2016 16:34:13 +0000 (17:34 +0100)]
greybus: connection: drop the legacy protocol-id parameter
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.
Johan Hovold [Thu, 21 Jan 2016 16:34:10 +0000 (17:34 +0100)]
greybus: manifest: check for duplicate CPort descriptors when parsing
Now that connection creation has been separated from interface
initialisation, we should explicitly check for duplicate CPort
descriptors when parsing the manifest.
Johan Hovold [Thu, 21 Jan 2016 16:34:09 +0000 (17:34 +0100)]
greybus: core: defer connection creation to driver probe
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.
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.
greybus: firmware: log the name of the firmware being requested
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: Greg Kroah-Hartman <gregkh@google.com> Reviewed-by: Michael Scott <michael.scott@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
Rui Miguel Silva [Thu, 21 Jan 2016 01:42:17 +0000 (01:42 +0000)]
greybus: svc: add key event handling
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: Johan Hovold <johan@hovoldconsulting.com> Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
greybus: audio_codec: Don't be tricky with the driver model
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.
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.
Johan Hovold [Tue, 19 Jan 2016 11:51:25 +0000 (12:51 +0100)]
greybus: connection: fix lookup race
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.
Johan Hovold [Tue, 19 Jan 2016 11:51:24 +0000 (12:51 +0100)]
greybus: connection: drop the connection_mutex
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.
Johan Hovold [Tue, 19 Jan 2016 11:51:23 +0000 (12:51 +0100)]
greybus: connection: destroy workqueue at unregister
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.
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.
Johan Hovold [Tue, 19 Jan 2016 11:51:16 +0000 (12:51 +0100)]
greybus: connection: move legacy-protocol handling to legacy driver
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.
Johan Hovold [Tue, 19 Jan 2016 11:51:14 +0000 (12:51 +0100)]
greybus: control: drop legacy-protocol dependency
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).
Johan Hovold [Tue, 19 Jan 2016 11:51:13 +0000 (12:51 +0100)]
greybus: greybus_protocols: remove control-protocol version
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.
Johan Hovold [Tue, 19 Jan 2016 11:51:11 +0000 (12:51 +0100)]
greybus: firmware: remove skip-disconnected flag
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).
Johan Hovold [Tue, 19 Jan 2016 11:51:10 +0000 (12:51 +0100)]
greybus: core: add defensive connection disable post disconnect
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.
Johan Hovold [Tue, 19 Jan 2016 11:51:09 +0000 (12:51 +0100)]
greybus: core: disable incoming operations pre disconnect
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.
Johan Hovold [Tue, 19 Jan 2016 11:51:08 +0000 (12:51 +0100)]
greybus: connection: add helper to disable incoming operations
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.
projects / mirror_ubuntu-artful-kernel.git / log