Pull media fixes from Mauro Carvalho Chehab:

 - dvb-core fixes for vb2 check and device registration

 - v4l2-core: fix an issue with error handling for VIDIOC_G_CTRL

 - vb2 core: fix an issue with vb plane copy logic

 - videobuf2-core: copy vb planes unconditionally

 - vivid: fix buffer overwrite when using > 32 buffers

 - vivid: fix a potential division by zero due to an issue at v4l2-tpg

 - some spectre vulnerability fixes

 - several OOM access fixes

 - some buffer overflow fixes

* tag 'media/v6.12-2' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/linux-media:
  media: videobuf2-core: copy vb planes unconditionally
  media: dvbdev: fix the logic when DVB_DYNAMIC_MINORS is not set
  media: vivid: fix buffer overwrite when using > 32 buffers
  media: pulse8-cec: fix data timestamp at pulse8_setup()
  media: cec: extron-da-hd-4k-plus: don't use -1 as an error code
  media: stb0899_algo: initialize cfr before using it
  media: adv7604: prevent underflow condition when reporting colorspace
  media: cx24116: prevent overflows on SNR calculus
  media: ar0521: don't overflow when checking PLL values
  media: s5p-jpeg: prevent buffer overflows
  media: av7110: fix a spectre vulnerability
  media: mgb4: protect driver against spectre
  media: dvb_frontend: don't play tricks with underflow values
  media: dvbdev: prevent the risk of out of memory access
  media: v4l2-tpg: prevent the risk of a division by zero
  media: v4l2-ctrls-api: fix error handling for v4l2_g_ctrl()
  media: dvb-core: add missing buffer index check

Copy the relevant data from userspace to the vb->planes unconditionally
as it's possible some of the fields may have changed after the buffer
has been validated.

Keep the dma_buf_put(planes[plane].dbuf) calls in the first
`if (!reacquired)` case, in order to be close to the plane validation code
where the buffers were got in the first place.

Cc: stable@vger.kernel.org
Fixes: 95af7c00f35b ("media: videobuf2-core: release all planes first in __prepare_dmabuf()")
Signed-off-by: Tudor Ambarus <tudor.ambarus@linaro.org>
Tested-by: Will McVicker <willmcvicker@google.com>
Acked-by: Tomasz Figa <tfiga@chromium.org>
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>

When CONFIG_DVB_DYNAMIC_MINORS, ret is not initialized, and a
semaphore is left at the wrong state, in case of errors.

Make the code simpler and avoid mistakes by having just one error
check logic used weather DVB_DYNAMIC_MINORS is used or not.

Reported-by: kernel test robot <lkp@intel.com>
Reported-by: Dan Carpenter <dan.carpenter@linaro.org>
Closes: https://lore.kernel.org/r/202410201717.ULWWdJv8-lkp@intel.com/
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Link: https://lore.kernel.org/r/9e067488d8935b8cf00959764a1fa5de85d65725.1730926254.git.mchehab+huawei@kernel.org

The maximum number of buffers that can be requested was increased to
64 for the video capture queue. But video capture used a must_blank
array that was still sized for 32 (VIDEO_MAX_FRAME). This caused an
out-of-bounds write when using buffer indices >= 32.

Create a new define MAX_VID_CAP_BUFFERS that is used to access the
must_blank array and set max_num_buffers for the video capture queue.

This solves a crash reported by:

	https://bugzilla.kernel.org/show_bug.cgi?id=219258

Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
Fixes: cea70ed416b4 ("media: test-drivers: vivid: Increase max supported buffers for capture queues")
Cc: stable@vger.kernel.org

As pointed by Coverity, there is a hidden overflow condition there.
As date is signed and u8 is unsigned, doing:

	date = (data[0] << 24)

With a value bigger than 07f will make all upper bits of date
0xffffffff. This can be demonstrated with this small code:

<code>
typedef int64_t time64_t;
typedef uint8_t u8;

int main(void)
{
	u8 data[] = { 0xde ,0xad , 0xbe, 0xef };
	time64_t date;

	date = (data[0] << 24) | (data[1] << 16) | (data[2] << 8) | data[3];
	printf("Invalid data = 0x%08lx\n", date);

	date = ((unsigned)data[0] << 24) | (data[1] << 16) | (data[2] << 8) | data[3];
	printf("Expected data = 0x%08lx\n", date);

	return 0;
}
</code>

Fix it by converting the upper bit calculation to unsigned.

Fixes: cea28e7a55e7 ("media: pulse8-cec: reorganize function order")
Cc: stable@vger.kernel.org
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

The logic at get_edid_tag_location() returns either an offset
or an error condition. However, the error condition uses a
non-standard "-1" value. This hits a Coverity bug, as Coverity
assumes that positive values are underflow. While this is a
false positive, returning error codes as -1 is an issue.

So, instead, use -ENOENT to indicate that the tag was not found.

Fixes: 056f2821b631 ("media: cec: extron-da-hd-4k-plus: add the Extron DA HD 4K Plus CEC driver")
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

The loop at stb0899_search_carrier() starts with a random
value for cfr, as reported by Coverity.

Initialize it to zero, just like stb0899_dvbs_algo() to ensure
that carrier search won't bail out.

Fixes: 8bd135bab91f ("V4L/DVB (9375): Add STB0899 support")
Cc: stable@vger.kernel.org
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

Currently, adv76xx_log_status() reads some date using
io_read() which may return negative values. The current logic
doesn't check such errors, causing colorspace to be reported
on a wrong way at adv76xx_log_status(), as reported by Coverity.

If I/O error happens there, print a different message, instead
of reporting bogus messages to userspace.

Fixes: 54450f591c99 ("[media] adv7604: driver for the Analog Devices ADV7604 video decoder")
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Reviewed-by: Hans Verkuil <hverkuil@xs4all.nl>

as reported by Coverity, if reading SNR registers fail, a negative
number will be returned, causing an underflow when reading SNR
registers.

Prevent that.

Fixes: 8953db793d5b ("V4L/DVB (9178): cx24116: Add module parameter to return SNR as ESNO.")
Cc: stable@vger.kernel.org
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

The PLL checks are comparing 64 bit integers with 32 bit
ones, as reported by Coverity. Depending on the values of
the variables, this may underflow.

Fix it ensuring that both sides of the expression are u64.

Fixes: 852b50aeed15 ("media: On Semi AR0521 sensor driver")
Cc: stable@vger.kernel.org
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Acked-by: Sakari Ailus <sakari.ailus@linux.intel.com>