Add needed logic into dvb_input_attach(), ddb_port_probe() and
ddb_ports_init() to initialize and support these new cards.
Picked up from the upstream dddvb-0.9.33 release.
Signed-off-by: Daniel Scheller <d.scheller@gmx.net>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Add PCI IDs and ddb_info for the new MCI-based MaxSX8 cards. Also add
needed defines so the cards can be hooked up into ddbridge's probe and
attach handling.
Picked up from the upstream dddvb-0.9.33 release.
Signed-off-by: Daniel Scheller <d.scheller@gmx.net>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Implement frontend attachment as ddb_fe_attach_mci() into the
ddbridge-max module. The MaxSX8 MCI cards are part of the Max card series
and make use of the LNB controller driven by the already existing lnb
functionality, so here's where this code belongs to.
Picked up from the upstream dddvb-0.9.33 release.
Signed-off-by: Daniel Scheller <d.scheller@gmx.net>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
This adds initial support for the new MCI-based (micro-code interface)
DD cards, with the first one being the MaxSX8 eight-tuner DVB-S/S2/S2X
PCIe card. The MCI is basically a generalized interface implemented in
the card's FPGA firmware and usable for all kind of cards, without the
need to implement any demod/tuner drivers as this interface "hides" any
I2C interface to the actual ICs, in other words any required driver is
implemented in the card firmware.
At this stage, the MCI interface is quite rudimentary with things like
signal statistics reporting missing, but is already working to serve
DVB streams to DVB applications. Missing functionality will be enabled
over time.
This implements only the ddbridge-mci sub-object and hooks it up to the
Makefile so the object gets build. The upcoming commits hook this module
into all other ddbridge parts where required, including device IDs etc.
Picked up from the upstream dddvb-0.9.33 release.
Signed-off-by: Daniel Scheller <d.scheller@gmx.net>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
The Octopus V3 and Octopus Mini devices support set up of a dummy tuner
mode on port 0 that will deliver a continuous data stream of 125MBytes
per second while raising IRQs and filling the DMA buffers, which comes
handy for some stress, PCIe link and IRQ handling testing. The dummy
frontend is registered using dvb_dummy_fe's QAM dummy frontend. Set
ddbridge.dummy_tuner to 1 to enable this on the supported cards.
Picked up from the upstream dddvb-0.9.33 release.
Signed-off-by: Daniel Scheller <d.scheller@gmx.net>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Make the number of DMA buffers and their size configurable using module
parameters. Being able to set these to a higher number might help on
busy systems when handling overall high data rates without having to
edit the driver sources and recompile things.
Picked up from the upstream dddvb-0.9.33 release.
Signed-off-by: Daniel Scheller <d.scheller@gmx.net>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Currently, /sys/class/ddbridgeX/devid always reports 0 due to devid not
being set at all. Set the devid field alongside while storing all other
hardware ID data.
Picked up from the upstream dddvb-0.9.33 release.
Signed-off-by: Daniel Scheller <d.scheller@gmx.net>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
A 188 byte gap has to be left between the writer and the consumer. This
requires 2*188 bytes available to be able to write to the output buffers.
So, change ddb_output_free() to report free bytes according to this rule.
Picked up from the upstream dddvb-0.9.33 release.
Signed-off-by: Daniel Scheller <d.scheller@gmx.net>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Make sure to save IRQ states before taking the dma lock, as already done
in it's input_work() counterpart.
Picked up from the upstream dddvb-0.9.33 release.
Signed-off-by: Daniel Scheller <d.scheller@gmx.net>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Improve IRQ handling in the separated MSG/I2C and IO/TSDATA handlers by
applying a mask for recognized bits immediately upon reading the IRQ mask
from the hardware, so only the bits/IRQs that actually were set will be
acked.
Picked up from the upstream dddvb-0.9.33 release.
Signed-off-by: Daniel Scheller <d.scheller@gmx.net>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Currently, each IRQ requires one IRQ_HANDLE() line to call each IRQ
handler that was set up. Add a IRQ_HANDLE_NIBBLE() and IRQ_HANDLE_BYTE()
macro to call all handlers in blocks of four (_NIBBLE) or eight (_BYTE)
handlers at a time, to make this construct more compact.
Picked up from the upstream dddvb-0.9.33 release.
Signed-off-by: Daniel Scheller <d.scheller@gmx.net>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Introduce the ddb_irq_set() helper function (along with a matching
prototype in ddbridge.h) to improve the set up of the IRQ handlers
and handler_data, and rework storing this data into the ddb_link
using a new ddb_irq struct. This also does the necessary rework
of affected variables. And while at it, always do queue_work in
input_handler() as there's not much of a difference to directly
calling input_work if there's no ptr at input->redi, or queueing
this call.
Picked up from the upstream dddvb-0.9.33 release.
Signed-off-by: Daniel Scheller <d.scheller@gmx.net>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Instead of trying to manage IRQ numbers on itself, utilise the
pci_irq_vector() function to do this, which will take care of correct IRQ
numbering for MSI and non-MSI IRQs. While at it, request and enable MSI-X
interrupts for hardware (boards and cards) that support this.
Picked up from the upstream dddvb-0.9.33 release.
Signed-off-by: Daniel Scheller <d.scheller@gmx.net>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Introduce the ddb_msi_exit() helper to be used for cleaning up previously
allocated MSI IRQ vectors. Deduplicates code and makes things look
cleaner as for all cleanup work the CONFIG_PCI_MSI ifdeffery is only
needed in the helper now. Also, replace the call to the deprecated
pci_disable_msi() function with pci_free_irq_vectors().
Picked up from the upstream dddvb-0.9.33 release.
Signed-off-by: Daniel Scheller <d.scheller@gmx.net>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Move the ddbridge module initialisation and cleanup code to ddbridge-core
and set up the ddb_wq workqueue there, and create and destroy the ddb
device class there aswell. Due to this, the prototypes for ddb_wq,
ddb_class_create() and ddb_class_destroy() aren't required in ddbridge.h
anymore, so remove them. Also, declare ddb_wq and the ddb_class_*()
functions static.
Picked up from the upstream dddvb-0.9.33 release.
Signed-off-by: Daniel Scheller <d.scheller@gmx.net>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Besides the 'msi' module option, all options are used from within
ddbridge-core only, so move them over from ddbridge-main, and declare the
associated variables static. Since the prototypes in ddbridge.h aren't
necessary anymore now, remove them. As a side effect, this has the benefit
of aligning things more with the dddvb upstream.
Signed-off-by: Daniel Scheller <d.scheller@gmx.net>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
The CNR value determined in read_snr() is reported via the wrong variable.
It uses FE_SCALE_DECIBEL, which implies the value to be reported in svalue
instead of uvalue. Fix this accordingly.
Picked up from the upstream dddvb-0.9.33 release.
Signed-off-by: Daniel Scheller <d.scheller@gmx.net>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
The TSINSDEL registers were lacking initialisation in the stv0910 demod
driver. Initialise them (both demods) in the probe() function.
Picked up from the upstream dddvb-0.9.33 release.
Signed-off-by: Daniel Scheller <d.scheller@gmx.net>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
After adding this driver, gcc complains with:
drivers/media/i2c/video-i2c.c:55:1: warning: 'static' is not at beginning of declaration [-Wold-style-declaration]
const static struct v4l2_fmtdesc amg88xx_format = {
^~~~~
drivers/media/i2c/video-i2c.c:59:1: warning: 'static' is not at beginning of declaration [-Wold-style-declaration]
const static struct v4l2_frmsize_discrete amg88xx_size = {
^~~~~
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Now that FB_OMAP has stubs, the omap2 media drivers can be
built on ARM with COMPILE_TEST && DRM_OMAP.
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Despite depending on ACPI, this driver builds fine on non-x86
archtecture with COMPILE_TEST, as it doesn't depend on
ACPI-specific functions/structs.
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Acked-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
The pnp header already provide enough stub to build those
drivers with COMPILE_TEST on non-x86 archs.
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Acked-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
This driver doesn't use any weird API. So, allow building it
with COMPILE_TEST.
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Several radio devices only build on i386, because they depend
on ISA. Allow them to build on other archs by adding a
COMPILE_TEST check.
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Coverity complains about werid stuff at the debug logic:
CID 113542 (#1 of 1): Out-of-bounds access (ARRAY_VS_SINGLETON)10.
callee_ptr_arith: Passing buf to function flexcop_i2c_write4
which uses it as an array. This might corrupt or misinterpret
adjacent memory locations.
Instead of directly addressing the issue there, let's rework at
the logic there.
On newer kernels, KERN_CONT does nothing, as the previous message
won't wait for a continuation. Also, both flexcop_i2c_read4() and
flexcop_i2c_write4(), called by it, will print stuff if (debug &4).
So, the way it is is too buggy.
There are two kinds of debug stuff there: deb_i2c() and a code hidden
under #ifdef DUMP_I2C_MESSAGES, with can't be selected without touching
the source code.
Also, if both debug & 0x4 and DUMP_I2C_MESSAGES, flexcop_i2c_request()
will emit two debug messages per call with different data,
with sounds messy.
Simplify it by getting rid of DUMP_I2C_MESSAGES and adding a new
flag to debug (0x40), and making the debug logic there more
consistent.
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
As warned by Coverity:
CID 1415211 (#1 of 1): Unused value (UNUSED_VALUE)assigned_value:
Assigning value -22 to ret here, but that stored value is
overwritten before it can be used.
On all cases where the there's a goto 'unlock_out' or 'streamof',
ret was filled with a non-sero value. It toesn't make sense to override
such error code with a videobuf_streamoff() error.
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Acked-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Building this driver on arm64 gives this warning:
drivers/media/platform/s5p-jpeg/jpeg-hw-exynos3250.c:430:16: error: return expression in void function
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Reviewed-by: Jacek Anaszewski <jacek.anaszewski@gmail.com>
Acked-by: Andrzej Pietrasiewicz <andrzej.p@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Right now, at siano driver, all places where devpath is
defined has sizeof(devpath) == 32. So, there's no practical
risc of going past devpath array anywhere.
Still, code changes might cause troubles. It also confuses
Coverity:
CID 139059 (#1 of 1): Copy into fixed size buffer (STRING_OVERFLOW)
9. fixed_size_dest: You might overrun the 32-character
fixed-size string entry->devpath by copying devpath
without checking the length.
10. parameter_as_source: Note: This defect has an
elevated risk because the source argument
is a parameter of the current function.
So, explicitly limit strcmp() and strcpy() to ensure that the
devpath size (32) will be respected.
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Both Hauppauge WinTV 44981 (bt878) and the HVR 1110 (saa7134) have a Zilog
Z8F0811. The transmitter was not probed. Receive and transmit tested on
both cards.
Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
The iMON PAD controller has a analog stick, which can be switched to
keyboard mode (cursor keys) or work as a crappy mouse.
Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
The raw_register function exists to create input devices associated with
that IR protocol.
If the mce_kbd module is loaded, then every rc device will have mce_kbd
input devices, even if the protocol is not enabled. Change this to call
the register function to when the protocol is enabled.
Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Some VSP instances have two blending units named BRU (Blend/ROP Unit)
and BRS (Blend/ROP Sub unit). The BRS is a smaller version of the BRU
with only two inputs, but otherwise offers similar features and offers
the same register interface. The BRU and BRS can be used exchangeably in
VSP pipelines (provided no more than two inputs are needed).
Due to historical reasons, the VSP1 driver implements support for both
the BRU and BRS through objects named vsp1_bru. The code uses the name
BRU to refer to either the BRU or the BRS, except in a few places where
noted explicitly. This creates confusion.
In an effort to avoid confusion, rename the vsp1_bru object and the
corresponding API to vsp1_brx, and use BRx to refer to blend unit
instances regardless of their type. The names BRU and BRS are retained
where reference to a particular blend unit type is needed, as well as in
hardware registers to stay close to the datasheet.
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Acked-by: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Dynamic assignment of the BRU and BRS to pipelines is prone to
regressions, add messages to make debugging easier. Keep it as a
separate commit to ease removal of those messages once the code will
deem to be completely stable.
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Acked-by: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
The VSPDL variant drives two DU channels through two LIF and two
blenders, BRU and BRS. The DU channels thus share the five available
VSPDL inputs and expose them as five KMS planes.
The current implementation assigns the BRS to the second LIF and thus
artificially limits the number of planes for the second display channel
to two at most.
Lift this artificial limitation by assigning the BRU and BRS to the
display pipelines on demand based on the number of planes used by each
pipeline. When a display pipeline needs more than two inputs and the BRU
is already in use by the other pipeline, this requires reconfiguring the
other pipeline to free the BRU before processing, which can result in
frame drop on both pipelines.
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
When disabling a DRM plane, the corresponding RPF is only marked as
removed from the pipeline in the atomic update handler, with the actual
removal happening when configuring the pipeline at atomic commit time.
This is required as the RPF has to be disabled in the hardware, which
can't be done from the atomic update handler.
The current implementation is RPF-specific. Make it independent of the
entity type by using the entity's pipe pointer to mark removal from the
pipeline. This will allow using the mechanism to remove BRU instances.
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Display list completion is already reported to the frame end handler,
but that mechanism is global to all display lists. In order to implement
BRU and BRS reassignment in DRM pipelines we will need to commit a
display list and wait for its completion internally, without reporting
it to the DRM driver. Extend the display list API to support such an
internal use of the display list.
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
We will soon need to return more than a boolean completion status from
the vsp1_dlm_irq_frame_end() IRQ handler. Turn the return value into a
bitfield to prepare for that. No functional change is introduced here.
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
In order to make the vsp1_du_setup_lif() easier to read, and for
symmetry with the DRM pipeline input setup, move the pipeline output
setup code to a separate function.
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
The vsp1_du_setup_lif() function sets up the DRM pipeline input
manually. This duplicates the code from the
vsp1_du_pipeline_setup_inputs() function. Replace the manual
implementation by a call to the function.
As the pipeline has no enabled input in vsp1_du_setup_lif(), the
vsp1_du_pipeline_setup_inputs() function will not setup any RPF, and
will thus not setup formats on the BRU sink pads. This isn't a problem
as all inputs are disabled, and the BRU sink pads will be reconfigured
from the atomic commit handler when inputs will be enabled.
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
To implement fully dynamic plane assignment to pipelines, we need to
reassign the BRU and BRS to the DRM pipelines in the atomic commit
handler. In preparation for this setup factor out the BRU source pad
code and call it both at LIF setup and atomic commit time.
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
The DRM pipeline setup code used at atomic commit time is similar to the
setup code used when enabling the pipeline. Move it to a separate
function in order to share it.
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Move the duplicated DRM pipeline configuration code to a function and
call it from vsp1_du_setup_lif() and vsp1_du_atomic_flush().
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
The DRM pipeline handling code uses the entity's pipe list head to check
whether the entity is already included in a pipeline. This method is a
bit fragile in the sense that it uses list_empty() on a list_head that
is a list member. Replace it by a simpler check for the entity pipe
pointer.
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Various types of objects subclassing vsp1_entity currently store a
pointer to the pipeline. Move the pointer to vsp1_entity to simplify the
code and avoid storing the pipeline in more entity subclasses later.
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
The vsp1_drm_pipeline enabled field is set but never used. Remove it.
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
The DRM support code manages a pipeline of VSP entities, each backed by
a media entity. When starting or stopping the pipeline, it starts and
stops the media pipeline through the media API in order to store the
pipeline pointer in every entity.
The driver doesn't use the pipe pointer in media entities, neither does
it rely on the other effects of the media_pipeline_start() and
media_pipeline_stop() functions. Furthermore, as the media links for the
DRM pipeline are never set up correctly, and as the pipeline can be
modified dynamically when enabling or disabling planes, the current
implementation is not correct. Remove the incorrect and unneeded code.
While at it remove the outdated comment that states that entities are
not started when the LIF is setup, as they now are.
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Use the USBTV_TV_STD define instead of repeating ourselves.
Signed-off-by: Hugo Grostabussiat <bonstra@bonstra.fr.eu.org>
Signed-off-by: Hans Verkuil <hansverk@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Depending on the chosen standard, configure the decoder to use the
appropriate color encoding standard (PAL-like, NTSC-like or SECAM).
Until now, the decoder was not configured for a specific color standard,
making it autodetect the color encoding.
While this may sound fine, it potentially causes the wrong image tuning
parameters to be applied (e.g. tuning parameters for NTSC are applied to
a PAL source), and may confuse users about what the actual standard is
in use.
This commit explicitly configures the color standard the decoder will
use, making it visually obvious if a wrong standard was chosen.
Signed-off-by: Hugo Grostabussiat <bonstra@bonstra.fr.eu.org>
Signed-off-by: Hans Verkuil <hansverk@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
