| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
The dss_mgr_*() functions take a channel argument to identify the
channel they operate on. This prevents the functions from accessing
driver data structures without resorting to global variables. In an
effort to remove global variables, pass the omap_dss_device pointer
associated with the channel instead. This will be used to look up the
omap_drm_private data structure to pass to the dss_mgr_ops.
Signed-off-by: Laurent Pinchart <[email protected]>
Reviewed-by: Sebastian Reichel <[email protected]>
|
|
The dsi_display_init_dispc() and dsi_display_uninit_dispc() functions
take a channel argument that is reduntant as it is always identical to
the dsi->output.dispc_channel. Remove the argument and use the field
directly in the functions to avoid misuse.
Signed-off-by: Laurent Pinchart <[email protected]>
Reviewed-by: Sebastian Reichel <[email protected]>
|
|
The dsi_data structure stores a pointer to a struct platform_device. The
driver only uses the dev member of the platform device structure. Store
the struct device pointer instead and use it directly.
Signed-off-by: Laurent Pinchart <[email protected]>
Reviewed-by: Sebastian Reichel <[email protected]>
|
|
The dsi_bind() function receives a pointer to a struct device that it
casts to a struct platform_device, only to use the platform device's dev
field through the code. Use the dev pointer directly.
While at it rename the struct platform_device pointer dsidev to pdev to
make it more explicit.
Signed-off-by: Laurent Pinchart <[email protected]>
Reviewed-by: Sebastian Reichel <[email protected]>
|
|
The dsi_get_dsidrv_data() and dsi_get_dsidev_from_dssdev() inline
functions convert a struct omap_dss_device pointer to the corresponding
struct platform_device, and a struct platform_device pointer to the
corresponding struct dsi_data. They are nearly always called together
without any use of the intermediate platform_device, so combine them
into a single function.
In the three locations where only dsi_get_dsidrv_data() is used, call
dev_get_drvdata() directly.
Signed-off-by: Laurent Pinchart <[email protected]>
Reviewed-by: Sebastian Reichel <[email protected]>
|
|
Internal dsi functions take a pointer to the DSI platform_device and
then cast it to a dsi_data pointer. That's pointless as the caller
already has the dsi_data pointer. Pass it directly instead of the
platform_device pointer.
Signed-off-by: Laurent Pinchart <[email protected]>
Reviewed-by: Sebastian Reichel <[email protected]>
|
|
The dss_unbind() function casts the struct device pointer to a struct
platform_device, only to later use the struct device pointer from
platform_device. Don't cast at all.
Signed-off-by: Laurent Pinchart <[email protected]>
Reviewed-by: Sebastian Reichel <[email protected]>
|
|
As part of an effort to remove the usage of global variables in the
driver, store the debugfs root directory in the dss_device structure
instead of a global variable.
Signed-off-by: Laurent Pinchart <[email protected]>
Reviewed-by: Sebastian Reichel <[email protected]>
|
|
As part of an effort to remove the usage of global variables in the
driver, store the registered plls array in the dss_device structure
instead of a global variable.
Signed-off-by: Laurent Pinchart <[email protected]>
Reviewed-by: Sebastian Reichel <[email protected]>
|
|
To simplify implementation of debugfs seq_file show handlers, the driver
passes the pointer to the show function through the debugfs_create_file
data pointer. This prevents using the pointer to pass driver private
data to the show handler, and requires all handlers to use global
variables to access private data.
To prepare for the removal of global private data in the driver, rework
the debugfs infrastructure to allow passing a private data pointer to
show handlers.
The price to pay is explicit removal of debugfs files to free the
internally allocated memory. This is desirable anyway as debugfs entries
should be removed when a component driver is unbound, otherwise crashes
will occur due to access to freed memory when the components will be
dynamically allocated instead of stored in global variables.
Signed-off-by: Laurent Pinchart <[email protected]>
Reviewed-by: Sebastian Reichel <[email protected]>
|
|
The DSS private data structure is currently stored as a global variable.
While no platform with multiple DSS devices currently exists nor is
planned, this doesn't comply with the kernel device model and should
thus be fixed.
Allocate the DSS private data structure dynamically for each DSS
instance and remove the global variable. All code that need access to
the structure now retrieves it dynamically so we can remove the global
variable.
Signed-off-by: Laurent Pinchart <[email protected]>
Reviewed-by: Sebastian Reichel <[email protected]>
|
|
This removes the need to access the global DSS private data in those
functions (both for the current accesses and the future ones that will
be introduced when allocating the DSS device dynamically).
Signed-off-by: Laurent Pinchart <[email protected]>
Reviewed-by: Sebastian Reichel <[email protected]>
|
|
This removes the need to access the global DSS private data in those
functions (both for the current accesses and the future ones that will
be introduced when allocating the DSS device dynamically).
Signed-off-by: Laurent Pinchart <[email protected]>
Reviewed-by: Sebastian Reichel <[email protected]>
|
|
This removes the need to access the global DSS private data in those
functions (both for the current accesses and the future ones that will
be introduced when allocating the DSS device dynamically).
Signed-off-by: Laurent Pinchart <[email protected]>
Reviewed-by: Sebastian Reichel <[email protected]>
|
|
This removes the need to access the global DSS private data in those
functions (both for the current accesses and the future ones that will
be introduced when allocating the DSS device dynamically).
Signed-off-by: Laurent Pinchart <[email protected]>
Reviewed-by: Sebastian Reichel <[email protected]>
|
|
This removes the need to access the global DSS private data in those
functions (both for the current accesses and the future ones that will
be introduced when allocating the DSS device dynamically).
Signed-off-by: Laurent Pinchart <[email protected]>
Reviewed-by: Sebastian Reichel <[email protected]>
|
|
This will allow accessing the PLL data to get the DSS device pointer,
removing the need to access the global DSS private data.
Signed-off-by: Laurent Pinchart <[email protected]>
Reviewed-by: Sebastian Reichel <[email protected]>
|
|
To prepare for the removal of the global variable storing DSS private
data, pass its pointer to the dss_runtime_{get,put}() functions.
As this requires getting hold of the dss_device structure in the
callers, we add a new dss_get_device() function to retrieve it. The
function currently returns a pointer to the global data structure, and
will later be updated to get the pointer from device driver data when
the DSS private structure will be allocated dynamically.
Signed-off-by: Laurent Pinchart <[email protected]>
Reviewed-by: Sebastian Reichel <[email protected]>
|
|
The anoonymous dss structure in dss.c is the top-level component in the
omapdss driver. As such it should store all internal instance-specific
data that is currently stored in global variables. This however requires
both naming the structure to pass it around functions, and accessing it
from various locations in the omapdss driver. While we could implement
get and set functions for every field that needs to be accessed outside
of dss.c, that would introduce overhead and complexity that we could
avoid by exposing the structure to internal components of the omapdss
driver. Do so to prepare for removal of global variables.
Signed-off-by: Laurent Pinchart <[email protected]>
Reviewed-by: Sebastian Reichel <[email protected]>
|
|
When merging the omapdrm and omapdss drivers there will be not omapdrm
platform device anymore, and thus no associated probe and remove
functions. To prepare for that, split all the initialization code from
the probe function to make it usable without a platform device.
Similarly, split the cleanup code from the remove function.
Signed-off-by: Laurent Pinchart <[email protected]>
Reviewed-by: Sebastian Reichel <[email protected]>
|
|
The wait_for_bit_change() function returns the value of the bit it
polls. This requires the caller to compare the return value to the
expected bit value. As all the existing callers need is to check whether
the bit has reached the expected value, it's easier to return a boolean
status from the function.
Signed-off-by: Laurent Pinchart <[email protected]>
Reviewed-by: Sebastian Reichel <[email protected]>
Signed-off-by: Tomi Valkeinen <[email protected]>
|
|
The omap_dss_register_driver(), omap_dss_unregister_driver() and
dispc_enable_gamma_table() functions don't exist anymore, remove their
prototypes.
Signed-off-by: Laurent Pinchart <[email protected]>
Reviewed-by: Sebastian Reichel <[email protected]>
Signed-off-by: Tomi Valkeinen <[email protected]>
|
|
The function is unused, remove it.
Signed-off-by: Laurent Pinchart <[email protected]>
Reviewed-by: Sebastian Reichel <[email protected]>
Signed-off-by: Tomi Valkeinen <[email protected]>
|
|
There's no reason to delay initialization of most of the driver (such as
mapping memory I/O, getting clocks or enabling runtime PM) to the
component master bind handler.
This additionally fixes a real PM issue caused enabling runtime PM in
the bind handler.
The bind handler performs the following sequence of PM operations:
pm_runtime_enable(dev);
pm_runtime_get_sync(dev);
... (access the hardware to read the device revision) ...
pm_runtime_put_sync(dev);
If a failure occurs at this point, the error path calls
pm_runtime_disable() to balance the pm_runtime_enable() call.
To understand the problem, it should be noted that the bind handler is
called when one of the component registers itself, which happens in the
component's probe handler. Furthermore, as the components are children
of the DSS, the device core calls pm_runtime_get_sync() on the DSS
platform device before calling the component's probe handler. This
increases the DSS power usage count but doesn't runtime resume the
device, as runtime PM is disabled at that point.
The bind handler is thus called with runtime PM disabled, with the
device runtime suspended, but with the power usage count larger than 0.
The pm_runtime_get_sync() call will thus further increase the power
usage count and runtime resume the device. The pm_runtime_put_sync()
handler will decrease the power usage count to a non-zero value and will
thus not suspend the device. Finally, the pm_runtime_disable() call will
disable runtime PM, preventing the pm_runtime_put() call in the device
core from runtime suspending the device. The DSS device is thus left
powered on.
To fix this, move the initialization code from the bind handler to the
probe handler.
Signed-off-by: Laurent Pinchart <[email protected]>
Reviewed-by: Sebastian Reichel <[email protected]>
Signed-off-by: Tomi Valkeinen <[email protected]>
|
|
A few functions defined in omapdss-base are internal to the module.
Don't export them.
Signed-off-by: Laurent Pinchart <[email protected]>
Reviewed-by: Sebastian Reichel <[email protected]>
Signed-off-by: Tomi Valkeinen <[email protected]>
|
|
The function isn't used outside of its compilation unit, make it static.
Signed-off-by: Laurent Pinchart <[email protected]>
Reviewed-by: Sebastian Reichel <[email protected]>
Signed-off-by: Tomi Valkeinen <[email protected]>
|
|
The encoder drivers need a handle to the source they are connected to in
order to control the source.
All drivers get that handle at probe time, resulting in probe deferral
when the source hasn't been probed yet. However they don't need the
handle until their connect handler is called.
Move retrieval of the source handle to the connect handler to avoid
probe deferrals.
Signed-off-by: Laurent Pinchart <[email protected]>
Reviewed-by: Sebastian Reichel <[email protected]>
Signed-off-by: Tomi Valkeinen <[email protected]>
|
|
The connector drivers need a handle to the source they are connected to
in order to control the source.
All drivers get that handle at probe time, resulting in probe deferral
when the source hasn't been probed yet. However they don't need the
handle until their connect handler is called.
Move retrieval of the source handle to the connect handler to avoid
probe deferrals.
Signed-off-by: Laurent Pinchart <[email protected]>
Reviewed-by: Sebastian Reichel <[email protected]>
Signed-off-by: Tomi Valkeinen <[email protected]>
|
|
The connector drivers need a handle to the source they are connected to
in order to control the source.
All drivers get that handle at probe time, resulting in probe deferral
when the source hasn't been probed yet. However they don't need the
handle until their connect handler is called.
Move retrieval of the source handle to the connect handler to avoid
probe deferrals.
Signed-off-by: Laurent Pinchart <[email protected]>
Reviewed-by: Sebastian Reichel <[email protected]>
Signed-off-by: Tomi Valkeinen <[email protected]>
|
|
No panel is instantiated through platform data anymore, there is no
need to check for OF node presence.
Signed-off-by: Laurent Pinchart <[email protected]>
Reviewed-by: Sebastian Reichel <[email protected]>
Signed-off-by: Tomi Valkeinen <[email protected]>
|
|
No encoder is instantiated through platform data anymore, there is no
need to check for OF node presence.
Signed-off-by: Laurent Pinchart <[email protected]>
Reviewed-by: Sebastian Reichel <[email protected]>
Signed-off-by: Tomi Valkeinen <[email protected]>
|
|
No connector is instantiated through platform data anymore, there is no
need to check for OF node presence.
Signed-off-by: Laurent Pinchart <[email protected]>
Reviewed-by: Sebastian Reichel <[email protected]>
Signed-off-by: Tomi Valkeinen <[email protected]>
|
|
The tvc_of_match variable is never referenced before its definition.
Remove the forward declaration.
Signed-off-by: Laurent Pinchart <[email protected]>
Reviewed-by: Sebastian Reichel <[email protected]>
Signed-off-by: Tomi Valkeinen <[email protected]>
|
|
The kernel favours 'unsigned int' over plain 'unsigned'. Replace all
occurences of the latter by the former. This avoid lots of checkpatch
complaints in patches that touch lines where a plain 'unsigned' is used.
Signed-off-by: Laurent Pinchart <[email protected]>
Reviewed-by: Sebastian Reichel <[email protected]>
Signed-off-by: Tomi Valkeinen <[email protected]>
|
|
The standard kernel integer types are [us]{8,16,32}. Use them instead of
the u?int{8,16,32}_t types.
Signed-off-by: Laurent Pinchart <[email protected]>
Reviewed-by: Sebastian Reichel <[email protected]>
Signed-off-by: Tomi Valkeinen <[email protected]>
|
|
Replace of_find_backlight_by_node and of the code around it
with of_find_backlight helper to avoid repetition of code.
Reviewed-by: Noralf Trønnes <[email protected]>
Signed-off-by: Meghana Madhyastha <[email protected]>
Signed-off-by: Sean Paul <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/eb914313f9d0e58d81572ccd3c718a573a891bd7.1516810726.git.meghana.madhyastha@gmail.com
|
|
Use backlight_enable/disable helpers instead of changing
the property and calling backlight_update_status for cleaner
and simpler code and also to avoid repetitions.
Reviewed-by: Noralf Trønnes <[email protected]>
Reviewed-by: Sean Paul <[email protected]>
Signed-off-by: Meghana Madhyastha <[email protected]>
Signed-off-by: Sean Paul <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/30d5b26c3c7d75de3d2ab3cff9dee67084fc3caf.1516810725.git.meghana.madhyastha@gmail.com
|
|
Linux 4.15-rc8
Daniel requested this for so the intel CI won't fall over on drm-next
so often.
|
|
The omap4 CEC hardware cannot tell a Nack from a Low Drive from an
Arbitration Lost error, so just report a Nack, which is almost
certainly the reason for the error anyway.
This also simplifies the implementation. The only three interrupts
that need to be enabled are:
Transmit Buffer Full/Empty Change event: triggered when the
transmit finished successfully and cleared the buffer.
Receiver FIFO Not Empty event: triggered when a message was received.
Frame Retransmit Count Exceeded event: triggered when a transmit
failed repeatedly, usually due to the message being Nacked. Other
reasons are possible (Low Drive, Arbitration Lost) but there is no
way to know. If this happens the TX buffer needs to be cleared
manually.
While testing various error conditions I noticed that the hardware
can receive messages up to 18 bytes in total, which exceeds the legal
maximum of 16. This could cause a buffer overflow, so we check for
this and constrain the size to 16 bytes.
The old incorrect interrupt handler could cause the CEC framework to
enter into a bad state because it mis-detected the "Start Bit Irregularity
event" as an ARB_LOST transmit error when it actually is a receive error
which should be ignored.
Signed-off-by: Hans Verkuil <[email protected]>
Reported-by: Henrik Austad <[email protected]>
Tested-by: Henrik Austad <[email protected]>
Tested-by: Hans Verkuil <[email protected]>
Signed-off-by: Tomi Valkeinen <[email protected]>
|
|
Currently, calls into each file are used to register the various
platform drivers. Change this to a table of pointers to platform_driver
structs to allow using platform_register_drivers.
Signed-off-by: Andrew F. Davis <[email protected]>
|
|
Having the filename in the header serves little purpose and is
often wrong after renames as it is here in several places, just
drop it from all omapdrm files.
While we are here unify the copyright tags to the TI recommended style.
Signed-off-by: Andrew F. Davis <[email protected]>
|
|
Check the status of the DMM engine after it is reported that the
transaction was completed as in rare cases the engine might not reached a
working state.
The wait_status() will print information in case the DMM is not reached the
expected state and the dmm_txn_commit() will return with an error code to
make sure that we are not continuing with a broken setup.
Signed-off-by: Peter Ujfalusi <[email protected]>
Signed-off-by: Tomi Valkeinen <[email protected]>
|
|
To help diagnose DMM errors, print out information if any of the error bits
are set in the interrupt status register.
Signed-off-by: Peter Ujfalusi <[email protected]>
Signed-off-by: Tomi Valkeinen <[email protected]>
|
|
If the wait_status() fails either because of an error reported in the
STATUS register or because of a timeout waiting for the wait_mask, print
information which might help diagnose the reason.
Signed-off-by: Peter Ujfalusi <[email protected]>
Signed-off-by: Tomi Valkeinen <[email protected]>
|
|
The error bit definitions are typoed in DMM_IRQSTAT_ERR_MASK which went
unnoticed since the DMM_IRQSTAT_ERR_MASK was not used.
Change the bit definitions to the correct ones.
Signed-off-by: Peter Ujfalusi <[email protected]>
Signed-off-by: Tomi Valkeinen <[email protected]>
|
|
The number of function declarations in the omap_drv.h degrades
readability. To fix it, create new header files for each part of the
driver and move the related functions.
Signed-off-by: Laurent Pinchart <[email protected]>
Reviewed-by: Sebastian Reichel <[email protected]>
Signed-off-by: Tomi Valkeinen <[email protected]>
|
|
The venc_probe_of() function has an error cleanup path that returns
success instead of an error code. Fix it.
Signed-off-by: Laurent Pinchart <[email protected]>
Reviewed-by: Sebastian Reichel <[email protected]>
Signed-off-by: Tomi Valkeinen <[email protected]>
|
|
The dpi_data structure port_initialized field is used to check in the
cleanup path whether the DPI has been initialized. This can be performed
through the associated device_node data field instead. Remove the
port_initialized field.
Signed-off-by: Laurent Pinchart <[email protected]>
Reviewed-by: Sebastian Reichel <[email protected]>
Signed-off-by: Tomi Valkeinen <[email protected]>
|
|
The function isn't used outside of its compilation unit, make it static.
Signed-off-by: Laurent Pinchart <[email protected]>
Reviewed-by: Sebastian Reichel <[email protected]>
Signed-off-by: Tomi Valkeinen <[email protected]>
|
|
When merging the omapdrm and omapdss drivers the omapdrm virtual
platform device will disappear, and the omapdss platform device will be
used for DMA memory allocation. To prepare for that, set the DMA
coherent mask for the device.
Signed-off-by: Laurent Pinchart <[email protected]>
Reviewed-by: Sebastian Reichel <[email protected]>
Signed-off-by: Tomi Valkeinen <[email protected]>
|