Arnd Bergmann [Fri, 8 Jan 2021 08:23:28 +0000 (09:23 +0100)]
media: atomisp: fix Wvisiblity warning
Some randconfig builds include ia_css_firmware.h without first
including linux/device.h:
In file included from atomisp/pci/mmu/sh_mmu_mrfld.c:23:
In file included from atomisp/pci/atomisp_compat.h:22:
In file included from atomisp/pci/atomisp_compat_css20.h:24:
In file included from atomisp/pci/ia_css.h:28:
In file included from atomisp/pci/ia_css_control.h:25:
drivers/staging/media/atomisp//pci/ia_css_firmware.h:52:29: error: declaration of 'struct device' will not be visible outside of this function [-Werror,-Wvisibility]
media: media/siano: Fix transmission parameters reporting for ISDB-T
The currently used functions for converting the transmission mode and guard
interval from the internal values to the DVBv5 API values do not return
correct values for ISDB broadcasts:
- The raw (debugfs) transmission mode is an integer whose values can be
1, 2 or 3, that corresponds to 2K, 4K and 8K FFT modes respectively.
However sms_to_mode() expects values 2, 4 or 8.
- Guard interval, as defined by smscoreapi.h returns "1 divided by value"
instead of 0, 1, 2, and 3 as defined in sms_to_guard_interval_table().
This commit implements ISDB-T specific methods for converting the internal
values for the aforementioned parameters to the DVBv5 API values. It also
adds support for reporting FEC and time interleaving values for each one
of the layers.
Peilin Ye [Fri, 11 Dec 2020 08:30:39 +0000 (09:30 +0100)]
media: dvbdev: Fix memory leak in dvb_media_device_free()
dvb_media_device_free() is leaking memory. Free `dvbdev->adapter->conn`
before setting it to NULL, as documented in include/media/media-device.h:
"The media_entity instance itself must be freed explicitly by the driver
if required."
Xu Wang [Fri, 25 Dec 2020 08:45:24 +0000 (09:45 +0100)]
media: atomisp: Remove redundant NULL check
./drivers/staging/media/atomisp/pci/isp/kernels/sdis/sdis_1.0/ia_css_sdis.host.c:390:2-8: WARNING: NULL check before some freeing functions is not needed.
Xu Wang [Fri, 25 Dec 2020 08:38:18 +0000 (09:38 +0100)]
media: atomisp: Remove redundant NULL check
Fix warnings reported by coccicheck:
./drivers/staging/media/atomisp/pci/runtime/isp_param/src/isp_param.c:159:4-10: WARNING: NULL check before some freeing functions is not needed.
media: atomisp: replace boolean comparison of values with bool variables
There are certain expressions in a condition in atomisp, where a boolean
variable is compared with true/false in forms such as (foo == true)
or (false != bar), which does not comply with the coding style rule by
checkpatch.pl (CHK: BOOL_COMPARISON), according to which the boolean
variables should be themselves used in the condition, rather than
comparing with true or false.
E.g. In drivers/staging/media/atomisp/pci/atomisp_compat_css20.c:
if (asd->stream_prepared == false) {
Can be replaced with:
if (!asd->stream_prepared) {
Replace such expressions with boolean variables appropriately.
media: atomisp: Write function decleration in one line
CHECK:OPEN_ENDED_LINE: Lines should not end with a '('
WARNING:LEADING_SPACE: please, no spaces at the start of a line
Avoid these errors by writing the function decleration in one line.
Write return_type, function_name and parameters in one line
because lines should not end with a '(' [OPEN_ENDED_LINE]
Write open brace ’{’ on the next line to fix OPEN_BRACE Error
media: atomisp: Convert comments to C99 initializers
The struct initalizers have been changed as recommended on
https://kernelnewbies.org/KernelJanitors/Todo
Also remove all the false, 0, and NULL members.
media: atomisp: do not select COMMON_CLK to fix builds
COMMON_CLK is a user-selectable option with its own dependencies. The
most important dependency is !HAVE_LEGACY_CLK. User-selectable drivers
should not select COMMON_CLK because they will create a dependency cycle
and build failures.
Kaixu Xia [Sat, 14 Nov 2020 13:55:50 +0000 (14:55 +0100)]
media: atomisp: make atomisp_css_continuous_set_num_raw_frames() void
The return value of atomisp_css_continuous_set_num_raw_frames() function
is always 0, so there's no reason for a return value. In addition, the
__enable_continuous_mode() checked the return value for possible error
which is unnecessary. Convert atomisp_css_continuous_set_num_raw_frames()
to a void function.
Kaixu Xia [Sat, 14 Nov 2020 13:19:36 +0000 (14:19 +0100)]
media: atomisp: make atomisp_css_stop() void
The return value of atomisp_css_stop() function is always 0 and there
are no callers check the return value, so there's no reason for a return
value. Convert atomisp_css_stop() to a void function.
Alex Dewar [Sat, 17 Oct 2020 14:28:00 +0000 (16:28 +0200)]
media: staging: media: atomisp: Remove unused function
The function ia_css_mipi_frame_specify() is not called from anywhere and
the comment above its declaration states that it should be removed when
there are no more users. So remove it.
Colin Ian King [Wed, 7 Oct 2020 12:16:28 +0000 (14:16 +0200)]
media: m88rs6000t: avoid potential out-of-bounds reads on arrays
There a 3 array for-loops that don't check the upper bounds of the
index into arrays and this may lead to potential out-of-bounds
reads. Fix this by adding array size upper bounds checks to be
full safe.
Leonid Kushnir [Tue, 6 Oct 2020 20:29:03 +0000 (22:29 +0200)]
media: staging: atomisp: Corrected error handling in function
This patch fixes check of a result of 'power_up()' function call in
function 'gc0310_s_power()' to do "error handling" instead of "success
handling" as Dan Carpenter noted in his comment on the previous patch.
Lines 'return gc0310_init(sd)' and 'return ret' are swapped, and direct
value of 'ret' is checked in IF statement now.
Leonid Kushnir [Tue, 6 Oct 2020 08:17:21 +0000 (10:17 +0200)]
media: staging: atomisp: Removed else branch in function
This patch fixes the checkpatch.pl warning :
WARNING: else is not generally useful after a break or return
Expressions under 'else' branch in function 'gc0310_s_power' are
executed whenever the exppression in 'if' is False. Otherwise, return
from function occurs. Therefore, there is no need in 'else', and it has
been removed.
Leonid Kushnir [Mon, 5 Oct 2020 09:41:07 +0000 (11:41 +0200)]
media: staging: atomisp: Fixed allignment to match open parenthesis
Moved the parameters of the function '__gc0310_write_reg_is_consecutive'
to the right under open parenthesis to fix warning message from
checkpatch.pl: 'CHECK: Alignment should match open parenthesis'.
Inside alloc_user_pages() based on flag value either pin_user_pages()
or get_user_pages_fast() will be called. However, these API might fail.
But free_user_pages() called in error handling path doesn't bother
about return value and will try to unpin bo->pgnr pages, which is
incorrect.
Fix this by passing the page_nr to free_user_pages(). If page_nr > 0
pages will be unpinned based on bo->mem_type. This will also take care
of non error handling path.
allocation")
Link: https://lore.kernel.org/linux-media/1601219284-13275-1-git-send-email-jrdr.linux@gmail.com Fixes: 14a638ab96c5 ("media: atomisp: use pin_user_pages() for memory Signed-off-by: Souptick Joarder <jrdr.linux@gmail.com> Cc: John Hubbard <jhubbard@nvidia.com> Cc: Ira Weiny <ira.weiny@intel.com> Reviewed-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
There is no Makefile rule to have drivers/media/pci/mantis/mantis_core.o
in build since the code overhaul with commit b3b961448f70
("V4L/DVB (13795): [Mantis/Hopper] Code overhaul, add Hopper devices into the PCI ID list").
It looks like drivers/media/pci/mantis/mantis_core.c is a leftover.
The wrappers in include/linux/pci-dma-compat.h should go away.
The patch has been generated with the coccinelle script below and has been
hand modified to replace GFP_ with a correct flag.
It has been compile tested.
When memory is allocated in 'mantis_alloc_buffers()' (mantis_dma.c)
GFP_KERNEL can be used because it is called from 'mantis_dma_init()' which
is only called from probe functions and no lock is taken in the between.
Alex Dewar [Thu, 3 Sep 2020 18:31:34 +0000 (20:31 +0200)]
media: staging: media: atomisp: Don't abort on error in module exit path
The function lm3554_remove() checks for the return code for
lm3554_gpio_uninit() even though this is on the exit path and exits the
function, leaving the variable flash unfreed. Print a warning instead
and free flash unconditionally.
media: dvb-frontends: lgdt3306a.c: remove dead code
Fix the following coccinelle report:
drivers/media/dvb-frontends/lgdt3306a.c:718:5-8:
Unneeded variable: "ret". Return "0" on line 744
Code was enclosed with #if 0 and thus this variable was not needed.
Remove that function altogether.
The lgdt3306a_spectral_inversion function was dropped completely
since the function doesn't do anything. A comment was left at its
calling site to let others know that spectral_inversion defaults
already set for VSB and QAM
Found using - Coccinelle (http://coccinelle.lip6.fr)
In the case where the call to lm3554_platform_data_func returns an
error there is a memory leak on the error return path of object
flash. Fix this by adding an error return path that will free
flash and rename labels fail2 to fail3 and fail1 to fail2.
media: dvb_frontend: warn if frontend driver has API issues
the kAPI for a frontend can use 3 different tuning methods:
1. The hardware tracks internally frequency shifts via its
own internal zigzag logic;
2. The hardware has a custom zigzag method, implemented via
fe search() ops;
3. The hardware doesn't have any internal zigzag logic. So,
the Kernel needs to implement it.
Drivers that use the in-kernel software zigzag are required to
provide some parameters for the zigzag code to work. Failing
to do that will just make the Kernel to tune several times
to the very same frequency, delaying the tuning time for
no good reason. This is actually a kAPI violation
(and an uAPI one, as the frequency shift is exported to the
uAPI).
Emit a warning on such case, as the driver needs to be fixed.
The wrappers in include/linux/pci-dma-compat.h should go away.
The patch has been generated with the coccinelle script below and has been
hand modified to replace GFP_ with a correct flag.
It has been compile tested.
When memory is allocated in 'cx25821_riscmem_alloc()' GFP_KERNEL can be
used because either this flag is already used in the call chain, or it is
called from a 'buf_prepare' function.
The call chains are:
vb2_ops.buf_prepare (in cx25821-video.c)
cx25821_buffer_prepare (in cx25821-video.c)
cx25821_risc_buffer
cx25821_riscmem_alloc
snd_cx25821_hw_params (in cx25821-alsa.c) <-- use GFP_KERNEL
cx25821_risc_databuffer_audio
cx25821_riscmem_alloc
Hans Verkuil [Thu, 18 Mar 2021 09:03:06 +0000 (10:03 +0100)]
media: vidioc-g-ext-ctrls.rst: reformat tables and clarify which vs ctrl_class
The VIDIOC_G/S/TRY_EXT_CTRLS documentation has large explanatory texts for
some of the fields in a table. This makes it hard to read. Move those text
to a new cell spanning the whole width of the table, similar to what was
done for struct v4l2_pix_format. This makes it much more readable.
Also move the 'ctrl_class' description to below the 'which' description and
just mention that it is deprecated and that 'which' should be used instead.
Finally remove 'note::' for the V4L2_CTRL_WHICH_DEF_VAL description. It
doesn't have to be marked as a note, it's just a simple paragraph.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
- @h264_dec was not documented, add that trivial change
- it is not possible to document a global variable in kernel-doc,
change to use 'DOC:' instead
- fix mismatched struct name (hantro_vp8d_hw_ctx -> hantro_vp8_dec_hw_ctx)
Hans Verkuil [Tue, 9 Mar 2021 13:03:36 +0000 (14:03 +0100)]
media: media/usb: fix kernel-doc header issues
Fix the following kernel-doc warnings:
drivers/media/usb/dvb-usb-v2/dvb_usb.h:109: warning: cannot understand function prototype: 'struct dvb_usb_driver_info '
drivers/media/usb/dvb-usb-v2/dvb_usb.h:126: warning: cannot understand function prototype: 'struct dvb_usb_rc '
drivers/media/usb/dvb-usb-v2/dvb_usb.h:143: warning: cannot understand function prototype: 'struct usb_data_stream_properties '
drivers/media/usb/dvb-usb-v2/dvb_usb.h:170: warning: expecting prototype for properties of dvb usb device adapter(). Prototype was for
MAX_NO_OF_FE_PER_ADAP() instead
drivers/media/usb/dvb-usb-v2/dvb_usb.h:221: error: Cannot parse struct or union!
drivers/media/usb/dvb-usb-v2/dvb_usb.h:277: warning: expecting prototype for generic object of an usb stream(). Prototype was for
MAX_NO_URBS_FOR_DATA_STREAM() instead
drivers/media/usb/dvb-usb-v2/dvb_usb.h:319: warning: cannot understand function prototype: 'struct dvb_usb_adapter '
drivers/media/usb/dvb-usb-v2/dvb_usb.h:363: warning: cannot understand function prototype: 'struct dvb_usb_device '
drivers/media/usb/dvb-usb/dvb-usb.h:104: warning: cannot understand function prototype: 'struct usb_data_stream_properties '
drivers/media/usb/dvb-usb/dvb-usb.h:163: warning: expecting prototype for struct dvb_usb_adapter_properties. Prototype was for struct
dvb_usb_adapter_fe_properties instead
drivers/media/usb/dvb-usb/dvb-usb.h:205: warning: cannot understand function prototype: 'struct dvb_rc '
drivers/media/usb/dvb-usb/dvb-usb.h:227: warning: Enum value 'DVB_RC_LEGACY' not described in enum 'dvb_usb_mode'
drivers/media/usb/dvb-usb/dvb-usb.h:227: warning: Enum value 'DVB_RC_CORE' not described in enum 'dvb_usb_mode'
drivers/media/usb/dvb-usb/dvb-usb.h:271: error: Cannot parse struct or union!
drivers/media/usb/dvb-usb/dvb-usb.h:325: error: Cannot parse struct or union!
drivers/media/usb/dvb-usb/dvb-usb.h:385: warning: expecting prototype for struct dvb_usb_adapter. Prototype was for struct
dvb_usb_fe_adapter instead
drivers/media/usb/dvb-usb/dvb-usb.h:476: warning: Function parameter or member 'num_adapters_initialized' not described in 'dvb_usb_device'
drivers/media/usb/dvb-usb/dvb-usb.h:476: warning: Function parameter or member 'adapter' not described in 'dvb_usb_device'
drivers/media/usb/dvb-usb/dvb-usb.h:476: warning: Function parameter or member 'rc_phys' not described in 'dvb_usb_device'
drivers/media/usb/em28xx/em28xx.h:377: warning: cannot understand function prototype: 'enum em28xx_amux '
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Robert Foss [Tue, 16 Mar 2021 17:19:18 +0000 (18:19 +0100)]
media: camss: Refactor CSID HW version support
In order to support Qualcomm ISP hardware architectures that diverge
from older architectures, the CSID subdevice drivers needs to be refactored
to better abstract the different ISP hardware architectures.
Signed-off-by: Robert Foss <robert.foss@linaro.org> Reviewed-by: Andrey Konovalov <andrey.konovalov@linaro.org> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Robert Foss [Tue, 16 Mar 2021 17:19:17 +0000 (18:19 +0100)]
media: camss: Add missing format identifiers
The CSI-2 spec defines format identifier for Data Type (DT),
and how the Decode Format (DF) & Encode Format (EF) are implemented.
The spec does however not define the DF, EF or Plain Format (PF)
identifiers, as those are vendor specific.
Plain formats describe the size of the pixels written by the RDI
units to memory. PLAIN8 for example has the size 8 bits, and
PLAIN32 32 bits. The appropriate Plain Format is determined by
the Decode Format used. The smallest Plain Format that is able
to contain a pixel of the used Decode Format is the appropriate
one to use.
As the vendor specific identifiers differ between hardware
generations, split them out into separate headers.
Signed-off-by: Robert Foss <robert.foss@linaro.org> Reviewed-by: Andrey Konovalov <andrey.konovalov@linaro.org> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Robert Foss [Tue, 16 Mar 2021 17:19:16 +0000 (18:19 +0100)]
media: camss: Add support for VFE hardware version Titan 170
Add register definitions for version 170 of the Titan architecture
and implement support for the RDI output mode.
The RDI mode as opposed to the PIX output mode for the VFE unit does
not support any ISP functionality. This means essentially only
supporting dumping the output of the whatever the CSI decoder receives
from the sensor.
For example will a sensor outputting YUV pixel format frames, only
allow the VFE to dump those frames as they are received by the ISP
to memory through the RDI interface.
Signed-off-by: Robert Foss <robert.foss@linaro.org> Reviewed-by: Andrey Konovalov <andrey.konovalov@linaro.org> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Robert Foss [Tue, 16 Mar 2021 17:19:15 +0000 (18:19 +0100)]
media: camss: Refactor VFE HW version support
In order to support Qualcomm ISP hardware architectures that diverge
from older architectures, the VFE subdevice driver needs to be refactored
to better abstract the different ISP architectures.
Gen1 represents the CAMSS ISP architecture. The ISP architecture developed
after CAMSS, Titan, will be referred to as Gen2.
Signed-off-by: Robert Foss <robert.foss@linaro.org> Reviewed-by: Andrey Konovalov <andrey.konovalov@linaro.org> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Robert Foss [Tue, 16 Mar 2021 17:19:14 +0000 (18:19 +0100)]
media: camss: Make ISPIF subdevice optional
This driver supports multiple architecture versions of the Qualcomm ISP.
The CAMSS architecure which this driver is name after, and with the
introduction of this series, the Titan architecture.
The ISPIF is an IP-block that is only present in the CAMSS generation of
the architecture. In order to support the Titan generation, make the
ISPIF an optional subdevice.
Signed-off-by: Robert Foss <robert.foss@linaro.org> Reviewed-by: Andrey Konovalov <andrey.konovalov@linaro.org> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Hans Verkuil [Fri, 12 Mar 2021 08:49:55 +0000 (09:49 +0100)]
media: v4l2-ctrls.c: initialize flags field of p_fwht_params
The V4L2_CID_STATELESS_FWHT_PARAMS compound control was missing a
proper initialization of the flags field, so after loading the vicodec
module for the first time, running v4l2-compliance for the stateless
decoder would fail on this control because the initial control value
was considered invalid by the vicodec driver.
Initializing the flags field to sane values fixes this.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Liu Ying [Thu, 11 Mar 2021 05:53:47 +0000 (06:53 +0100)]
media: docs: Fix data organization of MEDIA_BUS_FMT_RGB101010_1X30
The media bus bit width of MEDIA_BUS_FMT_RGB101010_1X30 is 30.
So, 'Bit31' and 'Bit30' cells for the 'MEDIA_BUS_FMT_RGB101010_1X30'
row should be spaces instead of '0's.
Fixes: 54f38fcae536 ("media: docs: move uAPI book to userspace-api/media") Signed-off-by: Liu Ying <victor.liu@nxp.com> Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Add a fix for the memory leak bugs that can occur when the
saa7164_encoder_register() function fails.
The function allocates memory without explicitly freeing
it when errors occur.
Add a better error handling that deallocate the unused buffers before the
function exits during a fail.
Signed-off-by: Daniel Niv <danielniv3@gmail.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>