Ezequiel Garcia [Tue, 16 Nov 2021 14:38:34 +0000 (14:38 +0000)]
media: hantro: Add quirk for NV12/NV12_4L4 capture format
The G2 core decoder engine produces NV12_4L4 format,
which is a simple NV12 4x4 tiled format. The driver currently
hides this format by always enabling the post-processor engine,
and therefore offering NV12 directly.
This is done without using the logic in hantro_postproc.c
and therefore makes it difficult to add VP9 cleanly.
Since fixing this is not easy, add a small quirk to force
NV12 if HEVC was configured, but otherwise declare NV12_4L4
as the pixel format in imx8mq_vpu_g2_variant.dec_fmts.
This will be used by the VP9 decoder which will be added soon.
Signed-off-by: Ezequiel Garcia <ezequiel@collabora.com> Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@collabora.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Ezequiel Garcia [Tue, 16 Nov 2021 14:38:33 +0000 (14:38 +0000)]
media: hantro: Simplify postprocessor
Add a 'postprocessed' boolean property to struct hantro_fmt
to signal that a format is produced by the post-processor.
This will allow to introduce the G2 post-processor in a simple way.
Signed-off-by: Ezequiel Garcia <ezequiel@collabora.com> Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@collabora.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Turns out the post-processor block on the G2 core is substantially
different from the one on the G1 core. Introduce hantro_postproc_ops
with .enable and .disable methods, which will allow to support
the G2 post-processor cleanly.
Signed-off-by: Ezequiel Garcia <ezequiel@collabora.com> Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@collabora.com> Reviewed-by: Benjamin Gaignard <benjamin.gaignard@collabora.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Ezequiel Garcia [Tue, 16 Nov 2021 14:38:31 +0000 (14:38 +0000)]
media: hantro: postproc: Fix motion vector space size
When the post-processor hardware block is enabled, the driver
allocates an internal queue of buffers for the decoder enginer,
and uses the vb2 queue for the post-processor engine.
For instance, on a G1 core, the decoder engine produces NV12 buffers
and the post-processor engine can produce YUY2 buffers. The decoder
engine expects motion vectors to be appended to the NV12 buffers,
but this is only required for CODECs that need motion vectors,
such as H.264.
Fix the post-processor logic accordingly.
Signed-off-by: Ezequiel Garcia <ezequiel@collabora.com> Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@collabora.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
The pointer 'mem' was initialized to 'fw->data' but immediately after
that it was assigned 'fw->data + 10'. Lets remove the extra assignement
and initialize the pointer to the address its going to use.
media: media si2168: fully initialize si2168 on resume only when necessary
At connection time (or boot) in si2168_probe(), the firmware is not
loaded to the device and the device is not fully activated. It is
not useful or sensible to do this full initialization on resume in
case it has not been previously initialized and is expected to be
in this initialized state.
Calling si2168_init() and therefore reading the firmware file for
the first time during resume leads to problems and should be avoided.
It is however safe to read the firmware file once it has already been
read outside of a suspend/resume situation.
Add a staus flag 'initialized' to store whether si2168_init() has
successfully been called. If initialization fails (e.g. due to missing
firmware file), the flag is not set.
Register a separate si2168_resume callback which only calls
si2168_init() once the 'initialized' flag has been set and it is safe
to load the firmware at resume.
The first call to si2168_init() will now always happen when the device
is actually used for the first time and never during resume.
This avoids the unsafe firmware file reading and should also speed up
resume by skipping unnecessary device initialization.
media: si2168: drop support for old firmware file name for si2168 B40
The si2168 B40 firmware file name has been changed in or before 2014.
During initialization, the new file name is preferred, but the old file
name is used as a fallback when request_firmware with the new file name
fails. Once reading the old file name has been attempted, only this name
will be used on further firmware loading attempts. During resume,
firmware reading with the new file name can (and likely will) fail even
when it actually exists. So this permanent switch to the fallback
firmware name happens even when not desired.
Any system using a recent kernel version can be expected to have the
firmware under the new name. The major distributions are either using
the dvb firmware collection from LibreELEC, which has the new firmware
file name, or do not package the firmware file but have documentation
pointing towards a manual installation of the firmware file under the
new name. If the firmware is available under the old name, it is
severely outdated. If the switch to the old file name is performed,
further firmware loading will either permanently fail (if it is not
available) or an outdated firmware version will be used.
Drop support for the fallback firmware file name and fail directly if
the firmware is not available under its current name. On following
attempts, the firmware read will then be retried with the correct
current name instead of the old name.
As reasoned above, there should be no negative effects of this change,
while simplifying code (the B40 variant will be handled identical
compared to the other variants of the si2168) and at the same time
fixing possible problems if firmware loading fails on resume.
Michael Kuron [Sun, 26 Sep 2021 20:51:26 +0000 (21:51 +0100)]
media: dib0700: fix undefined behavior in tuner shutdown
This fixes a problem where closing the tuner would leave it in a state
where it would not tune to any channel when reopened. This problem was
discovered as part of https://github.com/hselasky/webcamd/issues/16.
Since adap->id is 0 or 1, this bit-shift overflows, which is undefined
behavior. The driver still worked in practice as the overflow would in
most environments result in 0, which rendered the line a no-op. When
running the driver as part of webcamd however, the overflow could lead
to 0xff due to optimizations by the compiler, which would, in the end,
improperly shut down the tuner.
The bug is a regression introduced in the commit referenced below. The
present patch causes identical behavior to before that commit for
adap->id equal to 0 or 1. The driver does not contain support for
dib0700 devices with more adapters, assuming such even exist.
Tests have been performed with the Xbox One Digital TV Tuner on amd64.
Not all dib0700 devices are expected to be affected by the regression;
this code path is only taken by those with incorrect endpoint numbers.
Scott K Logan [Sun, 3 Oct 2021 00:18:30 +0000 (01:18 +0100)]
media: s5h1411.c: Fix a typo in the VSB SNR table
This looks like a typo. By manipulating the antenna on a device while
monitoring the reported SNR, I was able to see the unexpected jump.
After applying this patch, the spike was no longer present.
GIT_AUTHOR_NAME=Colin King
GIT_AUTHOR_EMAIL=colin.king@canonical.com
Variable ADCClk_Hz is being initialised with a variable that is never read
and then re-assigned immediately afterwards. Clean up the code by removing
it and just returning the return value from the call to stv0367cab_get_mclk
Addresses-Coverity: ("Unused value")
Link: https://lore.kernel.org/linux-media/20211014153253.63527-1-colin.king@canonical.com Cc: kernel-janitors@vger.kernel.org, linux-kernel@vger.kernel.org Signed-off-by: Colin Ian King <colin.king@canonical.com> Reviewed-by: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
zhaoxiao [Wed, 17 Nov 2021 03:06:56 +0000 (03:06 +0000)]
media: dib9000: Use min() instead of doing it manually
Fix following coccicheck warning:
drivers/media/dvb-frontends/dib9000.c:261:10-11: WARNING opportunity for min()
drivers/media/dvb-frontends/dib9000.c:345:10-11: WARNING opportunity for min()
media: media dvb_frontend: add suspend and resume callbacks to dvb_frontend_ops
While dvb_tuner_ops already has dedicated suspend and resume callbacks,
dvb_frontend_ops currently does not have them. Add those callbacks and
use them for suspend and resume. If they are not set, the old behavior
of calling sleep or init is used.
This allows dvb_frontend drivers to handle resume differently from init,
and suspend differently from sleep. No change is required for drivers
not needing this functionality.
Zheyu Ma [Tue, 11 May 2021 09:00:03 +0000 (10:00 +0100)]
media: b2c2: Add missing check in flexcop_pci_isr:
A out-of-bounds bug can be triggered by an interrupt, the reason for
this bug is the lack of checking of register values.
In flexcop_pci_isr, the driver reads value from a register and uses it as
a dma address. Finally, this address will be passed to the count parameter
of find_next_packet. If this value is larger than the size of dma, the
index of buffer will be out-of-bounds.
Fix this by adding a check after reading the value of the register.
The following KASAN report reveals it:
BUG: KASAN: slab-out-of-bounds in find_next_packet
drivers/media/dvb-core/dvb_demux.c:528 [inline]
BUG: KASAN: slab-out-of-bounds in _dvb_dmx_swfilter
drivers/media/dvb-core/dvb_demux.c:572 [inline]
BUG: KASAN: slab-out-of-bounds in dvb_dmx_swfilter+0x3fa/0x420
drivers/media/dvb-core/dvb_demux.c:603
Read of size 1 at addr ffff8880608c00a0 by task swapper/2/0
Memory state around the buggy address: ffff8880608bff80: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ffff8880608c0000: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
>ffff8880608c0080: 00 00 00 00 fc fc fc fc fc fc fc fc 00 00 00 00
^ ffff8880608c0100: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ffff8880608c0180: fc fc fc fc fc fc fc fc 00 00 00 00 00 00 00 00
==================================================================
When dvb_register_device() in dvb_dmxdev_init() fails, dvb_dmxdev_init()
does not return a failure, and the memory pointed to by dvbdev or
dvr_dvbdev is invalid at this point. If they are used subsequently, it
will result in UFA or null-ptr-deref.
If dvb_register_device() in dvb_dmxdev_init() fails, fix the bug by making
dvb_dmxdev_init() return an error as well.
Martin Weber [Tue, 9 Nov 2021 11:43:09 +0000 (11:43 +0000)]
media: coda: V4L2_PIX_FMT_GREY for coda960 JPEG Encoder
support greyscale pix fmt input for coda9_jpeg_encoder. The hardware
supports it, so allow V4L2 Mem2Mem JPEG Encoder use it as well. Tested
on an i.MX6QP.
[hverkuil: updated the Subject line as suggested by Philipp]
Signed-off-by: Martin Weber <martin.weber@br-automation.com> Reviewed-by: Philipp Zabel <p.zabel@pengutronix.de> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
media: rcar-vin: Free buffers with error if hardware stop fails
The driver already has logic to detect if it fails to stop properly and
report this error to the user. The driver however did not report the
unused buffers or buffers given to the hardware (if any) with an error,
the buffers where instead returned to user-space in the active state.
Build on the existing detection of the error condition and correctly
return the buffers with an error if it triggers.
Johan Hovold [Mon, 25 Oct 2021 12:16:40 +0000 (13:16 +0100)]
media: s2255: fix control-message timeouts
USB control-message timeouts are specified in milliseconds and should
specifically not vary with CONFIG_HZ.
Use the common control-message timeout define for the five-second
timeouts.
Fixes: 38f993ad8b1f ("V4L/DVB (8125): This driver adds support for the Sensoray 2255 devices.") Cc: stable@vger.kernel.org # 2.6.27 Signed-off-by: Johan Hovold <johan@kernel.org> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Dafna Hirschfeld [Fri, 22 Oct 2021 15:04:08 +0000 (16:04 +0100)]
media: mtk-vcodec: enc: add vp8 profile ctrl
In order for the encoder to work with gstreamer
it needs to have the V4L2_CID_MPEG_VIDEO_VP8_PROFILE
ctrl. This patch adds that ctrl with only profile 0
supported.
Signed-off-by: Dafna Hirschfeld <dafna.hirschfeld@collabora.com> Acked-by: Nicolas Dufresne <nicolas.dufresne@collabora.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
media: atomisp: pipe_binarydesc: drop logic incompatible with firmware
As we're using this firmware for ISP2401:
https://github.com/intel-aero/meta-intel-aero-base/blob/master/recipes-kernel/linux/linux-yocto/shisp_2401a0_v21.bin
Revert some changes that are not compatible with it, making
the code close to the Intel Aero one.
media: atomisp: binary.c: drop logic incompatible with firmware
As we're using this firmware for ISP2401:
https://github.com/intel-aero/meta-intel-aero-base/blob/master/recipes-kernel/linux/linux-yocto/shisp_2401a0_v21.bin
It makes no sense to try to select a non-existing binary
inside the firmware. So, revert it to reflect the Intel Aero
device driver.
media: atomisp: get rid of set pipe version custom ctrl
It doesn't make sense to have a control for that. Besides that,
the Intel Aero implementation doesn't have, which means that
even the custom control is not used in practice, at least
outside Android.
media: atomisp: sh_css_params: remove tests for ISP2401
Those tests are related to the input system, which is the same
for the chosen firmware, so both ISP2400 and ISP2401 will be
identical with that regards.
With the ISP2401 firmware we're using, the code differences
are not that much from ISP2400. Cleanup the code in order
to make it closer to Intel Aero driver.
There are several unused macros. Simplify the logic there, making
it closer to the Intel Aero driver and the corresponding firmware,
as this is what we have widely available for this device.
media: atomisp: drop check_pipe_resolutions() logic
Such function doesn't exist on Intel Aero driver. As we're using
its firmware, it may mean that this is not compatible with the
current file. So, drop it.
There's a commented dead code related to some future thing
to be implemented. As this won't happen, as it would require
a newer firmware, just drop the code.
There are two versions of those functions. It turns that the
choosen firmware use the old version. So, drop the unused
ones and ensure that all devices will use the right functions.
media: atomisp: get rid of phys event abstractions
ia_css_dequeue_event() is just an alias to ia_css_dequeue_psys_event(),
and atomisp_css_dequeue_event() do nothing but calling the event
dequeue logic.
Get rid of both abstractions, calling the function directly.
media: atomisp: handle errors at sh_css_create_isp_params()
The succ var tracks memory allocation erros on this function.
Fix it, in order to stop this W=1 Werror in clang:
drivers/staging/media/atomisp/pci/sh_css_params.c:2430:7: error: variable 'succ' set but not used [-Werror,-Wunused-but-set-variable]
bool succ = true;
^
Allow the sensor to report the per-resolution frame interval.
It should be noticed that, despite the ov2680 datasheet says
that, for resolutions <= 1280x720, the sensor supports 60fps,
the data at ov2680.h tells otherwise, limiting it to 30fps.
Not sure yet if this is a limitation of atomisp or if it is
just because the tables are incorrect.
media: atomisp-ov2680: uncomment other resolutions
There are several resolutions defined for this sensor, but just
two are currently uncommented. Change the driver to allow other
resolutions to also be used.
Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
drivers/staging/media/ipu3/ipu3-css-params.c:774:8: error: variable 'pin_scale' set but not used [-Werror,-Wunused-but-set-variable]
int pin_scale = 0;
^
Cai Huoqing [Tue, 9 Nov 2021 12:55:20 +0000 (12:55 +0000)]
media: rc: ir-hix5hd2: Add the dependency on HAS_IOMEM
The helper function devm_platform_ioremap_resource()
needs HAS_IOMEM enabled, so add the dependency on HAS_IOMEM.
Signed-off-by: Cai Huoqing <caihuoqing@baidu.com> Signed-off-by: Sean Young <sean@mess.org> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
>> drivers/media/platform/mtk-vcodec/mtk_vcodec_dec_stateless.c:133:18:
warning: Value stored to 'pfb' during its initialization is never
read [clang-analyzer-deadcode.DeadStores]
Jammy Huang [Tue, 9 Nov 2021 03:12:27 +0000 (03:12 +0000)]
media: aspeed: Update signal status immediately to ensure sane hw state
If res-chg, VE_INTERRUPT_MODE_DETECT_WD irq will be raised. But
v4l2_input_status won't be updated to no-signal immediately until
aspeed_video_get_resolution() in aspeed_video_resolution_work().
During the period of time, aspeed_video_start_frame() could be called
because it doesn't know signal becomes unstable now. If it goes with
aspeed_video_init_regs() of aspeed_video_irq_res_change()
simultaneously, it will mess up hw state.
To fix this problem, v4l2_input_status is updated to no-signal
immediately for VE_INTERRUPT_MODE_DETECT_WD irq.
Fixes: d2b4387f3bdf ("media: platform: Add Aspeed Video Engine driver") Signed-off-by: Jammy Huang <jammy_huang@aspeedtech.com> Acked-by: Paul Menzel <pmenzel@molgen.mpg.de> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Yunfei Dong [Thu, 4 Nov 2021 12:24:26 +0000 (12:24 +0000)]
media: mtk-vcodec: Align width and height to 64 bytes
Width and height need to 64 bytes aligned when setting the format.
Need to make sure all is 64 bytes align when use width and height to
calculate buffer size.
Signed-off-by: Yunfei Dong <yunfei.dong@mediatek.com> Acked-by: Nicolas Dufresne <nicolas.dufresne@collabora.com> Tested-by: Steve Cho <stevecho@chromium.org> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Hans Verkuil [Wed, 3 Nov 2021 12:28:31 +0000 (12:28 +0000)]
media: v4l2-ioctl.c: readbuffers depends on V4L2_CAP_READWRITE
If V4L2_CAP_READWRITE is not set, then readbuffers must be set to 0,
otherwise v4l2-compliance will complain.
A note on the Fixes tag below: this patch does not really fix that commit,
but it can be applied from that commit onwards. For older code there is no
guarantee that device_caps is set, so even though this patch would apply,
it will not work reliably.
projects / thirdparty / linux.git / log
