linux.git
16 months agomedia: bttv: start_streaming should return a proper error code
Hans Verkuil [Thu, 30 Nov 2023 12:58:10 +0000 (13:58 +0100)]
media: bttv: start_streaming should return a proper error code

The start_streaming callback returned 0 or 1 instead of a
proper error code. Fix that.

Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Fixes: b7ec3212a73a ("media: bttv: convert to vb2")
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
16 months agomedia: ov9640: Don't set format in sub-device state
Sakari Ailus [Wed, 6 Dec 2023 17:25:07 +0000 (19:25 +0200)]
media: ov9640: Don't set format in sub-device state

For the purpose of setting old non-pad based sub-device try format as a
basis for VIDIOC_TRY_FMT implementation, there is no need to set the
format in the sub-device state. Drop the assignment to the state, which
would result in a NULL pointer dereference.

Fixes: fd17e3a9a788 ("media: i2c: Use accessors for pad config 'try_*' fields")
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: tw9910: Don't set format in sub-device state
Sakari Ailus [Wed, 6 Dec 2023 17:24:11 +0000 (19:24 +0200)]
media: tw9910: Don't set format in sub-device state

For the purpose of setting old non-pad based sub-device try format as a
basis for VIDIOC_TRY_FMT implementation, there is no need to set the
format in the sub-device state. Drop the assignment to the state, which
would result in a NULL pointer dereference.

Fixes: fd17e3a9a788 ("media: i2c: Use accessors for pad config 'try_*' fields")
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: rj54n1cb0c: Don't set format in sub-device state
Sakari Ailus [Wed, 6 Dec 2023 17:22:43 +0000 (19:22 +0200)]
media: rj54n1cb0c: Don't set format in sub-device state

For the purpose of setting old non-pad based sub-device try format as a
basis for VIDIOC_TRY_FMT implementation, there is no need to set the
format in the sub-device state. Drop the assignment to the state, which
would result in a NULL pointer dereference.

Fixes: fd17e3a9a788 ("media: i2c: Use accessors for pad config 'try_*' fields")
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: mt9t112: Don't set format in sub-device state
Sakari Ailus [Wed, 6 Dec 2023 17:22:03 +0000 (19:22 +0200)]
media: mt9t112: Don't set format in sub-device state

For the purpose of setting old non-pad based sub-device try format as a
basis for VIDIOC_TRY_FMT implementation, there is no need to set the
format in the sub-device state. Drop the assignment to the state, which
would result in a NULL pointer dereference.

Fixes: fd17e3a9a788 ("media: i2c: Use accessors for pad config 'try_*' fields")
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: adv7183: Don't set format in sub-device state
Sakari Ailus [Wed, 6 Dec 2023 17:20:19 +0000 (19:20 +0200)]
media: adv7183: Don't set format in sub-device state

For the purpose of setting old non-pad based sub-device try format as a
basis for VIDIOC_TRY_FMT implementation, there is no need to set the
format in the sub-device state. Drop the assignment to the state, which
would result in a NULL pointer dereference.

Fixes: fd17e3a9a788 ("media: i2c: Use accessors for pad config 'try_*' fields")
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: saa6752hs: Don't set format in sub-device state
Sakari Ailus [Wed, 6 Dec 2023 11:35:33 +0000 (13:35 +0200)]
media: saa6752hs: Don't set format in sub-device state

For the purpose of setting old non-pad based sub-device try format as a
basis for VIDIOC_TRY_FMT implementation, there is no need to set the
format in the sub-device state. Drop the assignment to the state, which
would result in a NULL pointer dereference.

Fixes: fd17e3a9a788 ("media: i2c: Use accessors for pad config 'try_*' fields")
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: ccs: Select V4L2_CCI_I2C
Sakari Ailus [Wed, 6 Dec 2023 10:12:13 +0000 (12:12 +0200)]
media: ccs: Select V4L2_CCI_I2C

Select V4L2_CCI_I2C Kconfig option which the CCS driver now depends on.

Fixes: 529322112a3b ("media: ccs: Use V4L2 CCI for accessing sensor registers")
Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/oe-kbuild-all/202312060941.CYiHppAp-lkp@intel.com/
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: rockchip: rga: add NV12M support
Michael Tretter [Fri, 13 Oct 2023 11:00:34 +0000 (13:00 +0200)]
media: rockchip: rga: add NV12M support

Add support for the multi-planar variants of NV12. The RGA is now able
to exchange DMABUFs with other devices that only support multi-planar
NV12, for example the Hantro JPEG encoder.

Signed-off-by: Michael Tretter <m.tretter@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: rockchip: rga: rework buffer handling for multi-planar formats
Michael Tretter [Fri, 13 Oct 2023 11:00:33 +0000 (13:00 +0200)]
media: rockchip: rga: rework buffer handling for multi-planar formats

Multi-planar formats may have multiple planes that must be handled and
correctly mapped into a continuous buffer for the RGA by using the DMA
descriptors.

The plane offsets in the continuous mapping may now start at page
boundaries and the previous calculation based on the frame sizes is only
valid for planar buffers in a single memory. Therefore, the offsets must
be detected and set while creating the mapping.

Signed-off-by: Michael Tretter <m.tretter@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: rockchip: rga: switch to multi-planar API
Michael Tretter [Fri, 13 Oct 2023 11:00:32 +0000 (13:00 +0200)]
media: rockchip: rga: switch to multi-planar API

Switch to the multi-planar API, which allows to handle buffers with
separate planes.

The RGA driver doesn't expose multi-planar formats, yet. The existing
contiguous planar formats can be used with the multi-planar API as well,
but the multi-planar API is required for multi-planar formats.

Signed-off-by: Michael Tretter <m.tretter@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: rockchip: rga: use macros for testing buffer type
Michael Tretter [Fri, 13 Oct 2023 11:00:31 +0000 (13:00 +0200)]
media: rockchip: rga: use macros for testing buffer type

Use the provided V4L2_TYPE_IS_{OUTPUT,CAPTURE} macros to check if the
buffer or queue is OUTPUT or CAPTURE. The macros work also work for the
_MPLANE buffer and queue types and make it easier to switch to the
multi-planar API.

Signed-off-by: Michael Tretter <m.tretter@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: rockchip: rga: add local variable for pix_format
Michael Tretter [Fri, 13 Oct 2023 11:00:30 +0000 (13:00 +0200)]
media: rockchip: rga: add local variable for pix_format

The local variable allows to simplify the accessed to the format and
makes it easier to change the type of the format.

Signed-off-by: Michael Tretter <m.tretter@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: rockchip: rga: use pixelformat to find format
Michael Tretter [Fri, 13 Oct 2023 11:00:29 +0000 (13:00 +0200)]
media: rockchip: rga: use pixelformat to find format

Use the pixelformat instead of the v4l2_format to find the rga_fmt. This
avoids knowing the structure and type of v4l2_format in rga_fmt_find and
simplifies the function.

Also cleanup the users of the function. In try_fmt always return the
found pixel format to make sure that the pixel format is always set.
Thus, we can be sure that we will find the rga_fmt in s_fmt and can drop
the check if a given format has been found.

Signed-off-by: Michael Tretter <m.tretter@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: rockchip: rga: use clamp() to clamp size to limits
Michael Tretter [Fri, 13 Oct 2023 11:00:28 +0000 (13:00 +0200)]
media: rockchip: rga: use clamp() to clamp size to limits

The try_fmt should limit the width and height to the know limits of the
RGA. Use the clamp() helper instead of open coding the clamping.

Signed-off-by: Michael Tretter <m.tretter@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: rockchip: rga: set dma mask to 32 bits
Michael Tretter [Fri, 13 Oct 2023 11:00:27 +0000 (13:00 +0200)]
media: rockchip: rga: set dma mask to 32 bits

The RGA DMA descriptor list contains only 32-bit addresses. Set the
dma_mask to only allocate memory that is addressable by the descriptors.

This prevents errors when preparing vb2 buffers that were allocated by
the RGA.

Signed-off-by: Michael Tretter <m.tretter@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: rockchip: rga: pre-calculate plane offsets
Michael Tretter [Fri, 13 Oct 2023 11:00:26 +0000 (13:00 +0200)]
media: rockchip: rga: pre-calculate plane offsets

Calculate the plane offsets and store them with the video buffer while
creating the buffer mapping.

This allows the driver to more freely handle the memory of the DMA
mapping as the offsets and the mapping can be kept in sync.

The driver still has to update the offsets to respect the configured
cropping and rotation, but this calculation is now separated from the
calculation of the plane offsets.

Signed-off-by: Michael Tretter <m.tretter@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: rockchip: rga: split src and dst buffer setup
Michael Tretter [Fri, 13 Oct 2023 11:00:25 +0000 (13:00 +0200)]
media: rockchip: rga: split src and dst buffer setup

Split the register setup for the source and destination video buffers
into separate functions.

This is a cleanup to make the code more readable by separating the
offset calculation for the different buffers and prepares the driver for
using pre-calculated offsets of planes.

Signed-off-by: Michael Tretter <m.tretter@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: rockchip: rga: allocate DMA descriptors per buffer
Michael Tretter [Fri, 13 Oct 2023 11:00:24 +0000 (13:00 +0200)]
media: rockchip: rga: allocate DMA descriptors per buffer

The RGA driver allocates two buffers for the DMA descriptors of the
input and output buffers. Whenever a new job is processed, the
descriptor list is updated for the current buffers.

By updating the descriptor list during buf_prepare, it is possible to
correctly fail DMABUF imports if the buffers that shall be imported are
not within the 32 bit address range that can be addressed by the RGA.

Managing the DMA descriptor list with the buffer also makes it easier to
track the buffer mapping and the plane offsets into this mapping.

The cost is that the driver now requires DMA coherent memory per buffer
for the descriptor list. However, the size scales with the size of the
video buffers and is not allocated if the RGA is not used.

While at it, use dma_alloc_coherent to allocate the descriptors and get
rid of the virt_to_phys calls to get the physical addresses.

Signed-off-by: Michael Tretter <m.tretter@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: rockchip: rga: extract helper to fill descriptors
Michael Tretter [Fri, 13 Oct 2023 11:00:23 +0000 (13:00 +0200)]
media: rockchip: rga: extract helper to fill descriptors

The IOMMU of the RGA is programmed with a list of DMA descriptors that
contain an 32 bit address per 4k page in the video buffers. The address
in the descriptor points to the start address of the page.

Introduce 'struct rga_dma_desc' to make the handling of the DMA
descriptors explicit instead of hiding them behind standard types.

While at it, use provided helpers for iterating the sg_table instead of
manually calculating the DMA addresses.

Signed-off-by: Michael Tretter <m.tretter@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: rockchip: rga: fix swizzling for RGB formats
Michael Tretter [Fri, 13 Oct 2023 11:00:22 +0000 (13:00 +0200)]
media: rockchip: rga: fix swizzling for RGB formats

When using 32 bit RGB formats, the RGA on the rk3568 produces wrong
colors as the wrong color channels are read or written.  The reason is
that the format description for the channel swizzeling is wrong and the
wrong bits are configured. For example, when converting ARGB32 to NV12,
the alpha channel is used as blue channel.. This doesn't happen if the
color format is the same on both sides.

Fix the color_swap settings of the formats to correctly handle 32 bit
RGB formats.

For RGA_COLOR_FMT_XBGR8888, the RGA_COLOR_ALPHA_SWAP bit doesn't have an
effect. Thus, it isn't possible to handle the V4L2_PIX_FMT_XRGB32. Thus,
it is removed from the list of supported formats.

Signed-off-by: Michael Tretter <m.tretter@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: qcom: camss: Add sm8250 named power-domain support
Bryan O'Donoghue [Thu, 23 Nov 2023 17:03:07 +0000 (17:03 +0000)]
media: qcom: camss: Add sm8250 named power-domain support

Declare power-domain names "top", "ife0" and "ife1" eponymously for the
power-domains TITAN_TOP_GDSC, IFE_0_GDSC and IFE_1_GDSC respectively.

Signed-off-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Reviewed-by: Konrad Dybcio <konrad.dybcio@linaro.org>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: qcom: camss: Flag CSID-lites to support more CSIDs
Matti Lehtimäki [Thu, 23 Nov 2023 17:03:06 +0000 (17:03 +0000)]
media: qcom: camss: Flag CSID-lites to support more CSIDs

Some platforms such as SC7280 have 3 CSIDs and 2 CSID-lites but current
code has hardcoded 2 as the maximum number of CSIDs. Remove the hardcoded
maximum number of VFEs to handle all possible combinations of CSIDs and
CSID-lites.

Signed-off-by: Matti Lehtimäki <matti.lehtimaki@gmail.com>
Signed-off-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Reviewed-by: Konrad Dybcio <konrad.dybcio@linaro.org>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: qcom: camss: Flag VFE-lites to support more VFEs
Matti Lehtimäki [Thu, 23 Nov 2023 17:03:05 +0000 (17:03 +0000)]
media: qcom: camss: Flag VFE-lites to support more VFEs

Some platforms such as SC7280 have three VFEs and two VFE-lites. Current
code has hard-coded two as the maximum number of VFEs. Remove the
hard-coded maximum number of VFEs to handle all possible combinations of
VFEs and VFE-lites.

Signed-off-by: Matti Lehtimäki <matti.lehtimaki@gmail.com>
Reviewed-by: Konrad Dybcio <konrad.dybcio@linaro.org>
Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/oe-kbuild-all/202311200405.h6G4L9oe-lkp@intel.com
Signed-off-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: qcom: camss: Add support for named power-domains
Bryan O'Donoghue [Wed, 6 Dec 2023 12:25:00 +0000 (13:25 +0100)]
media: qcom: camss: Add support for named power-domains

Right now we use fixed indexes to assign power-domains, with a
requirement for the TOP GDSC to come last in the list.

Adding support for named power-domains means the declaration in the dtsi
can come in any order.

After this change we continue to support the old indexing - if a SoC
resource declaration or the in-use dtb doesn't declare power-domain names
we fall back to the default legacy indexing.

From this point on though new SoC additions should contain named
power-domains, eventually we will drop support for legacy indexing.

Tested-by: Matti Lehtimäki <matti.lehtimaki@gmail.com>
Signed-off-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: qcom: camss: Move VFE power-domain specifics into vfe.c
Bryan O'Donoghue [Thu, 23 Nov 2023 17:03:03 +0000 (17:03 +0000)]
media: qcom: camss: Move VFE power-domain specifics into vfe.c

Moving the location of the hooks to VFE power domains has several
advantages.

1. Separation of concerns and functional decomposition.
   vfe.c should be responsible for and know best how manage
   power-domains for a VFE, excising from camss.c follows this
   principle.

2. Embedding a pointer to genpd in struct camss_vfe{} meas that we can
   dispense with a bunch of kmalloc array inside of camss.c.

3. Splitting up titan top gdsc from vfe/ife gdsc provides a base for
   breaking up magic indexes in dtsi.

Suggested-by: Matti Lehtimäki <matti.lehtimaki@gmail.com>
Tested-by: Matti Lehtimäki <matti.lehtimaki@gmail.com>
Signed-off-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Reviewed-by: Konrad Dybcio <konrad.dybcio@linaro.org>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: qcom: camss: Use common VFE pm_domain_on/pm_domain_off where applicable
Bryan O'Donoghue [Thu, 23 Nov 2023 17:03:02 +0000 (17:03 +0000)]
media: qcom: camss: Use common VFE pm_domain_on/pm_domain_off where applicable

For the various versions of VFE we have a boiler-plate
pm_domain_on/pm_domain_off callback pair of the general form.

- Error check.
  Not always done but applicable to all.

- device_link_add (DL_FLAG_STATELESS | DL_FLAG_PM_RUNTIME |
                   DL_FLAG_RPM_ACTIVE);

- Error check returning -EINVAL on error.

- Return 0

Reduce the pattern down to a common callback. VFE 4.1 is a special case
which to me also indicates that it is worthwhile maintaining an indirection
for the vfe_pm_domain_{on|off} for now.

Otherwise lets chuck out a bunch of needlessly replicated code.

Reviewed-by: Konrad Dybcio <konrad.dybcio@linaro.org>
Suggested-by: Matti Lehtimäki <matti.lehtimaki@gmail.com>
Tested-by: Matti Lehtimäki <matti.lehtimaki@gmail.com>
Signed-off-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: qcom: camss: Convert to per-VFE pointer for power-domain linkages
Bryan O'Donoghue [Thu, 23 Nov 2023 17:03:01 +0000 (17:03 +0000)]
media: qcom: camss: Convert to per-VFE pointer for power-domain linkages

Right now we use the top-level camss structure to provide pointers via
VFE id index back to genpd linkages.

In effect this hard-codes VFE indexes to power-domain indexes in the
dtsi and mandates a very particular ordering of power domains in the
dtsi, which bears no relationship to a real hardware dependency.

As a first step to rationalising the VFE power-domain code and breaking
the magic indexing in dtsi use per-VFE pointers to genpd linkages.

The top-level index in msm_vfe_subdev_init is still used to attain the
initial so no functional or logical change arises from this change.

Reviewed-by: Konrad Dybcio <konrad.dybcio@linaro.org>
Tested-by: Matti Lehtimäki <matti.lehtimaki@gmail.com>
Signed-off-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: qcom: camss: Flag which VFEs require a power-domain
Bryan O'Donoghue [Thu, 23 Nov 2023 17:03:00 +0000 (17:03 +0000)]
media: qcom: camss: Flag which VFEs require a power-domain

At the moment we have some complex code for determining if a VFE requires a
power-domain attachment. Particularly discordant in this scheme is the
subtle reliance on VFE and VFE Lite declaration ordering in our resources.

VFE id is used to determine if a VFE is lite or not and consequently if a
VFE requires power-domain attachment. VFE Lite though is not a correct
delineation between power-domain and non power-domain state since early
SoCs have neither VFE Lite nor power-domains attached to VFEs.

Introduce has_pd to the VFE resource structure to allow the CAMSS code to
understand if it needs to try to attach a power-domain for a given VFE.

As a side-effect from this we no longer need to care about VFE Lite or
non-Lite or the id number associated with either and which order the
VFE/VFE Lite was declared in.

Reviewed-by: Konrad Dybcio <konrad.dybcio@linaro.org>
Tested-by: Matti Lehtimäki <matti.lehtimaki@gmail.com>
Signed-off-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: stk1160: Fixed high volume of stk1160_dbg messages
Ghanshyam Agrawal [Sat, 25 Nov 2023 09:02:36 +0000 (14:32 +0530)]
media: stk1160: Fixed high volume of stk1160_dbg messages

The function stk1160_dbg gets called too many times, which causes
the output to get flooded with messages. Since stk1160_dbg uses
printk, it is now replaced with printk_ratelimited.

Suggested-by: Phillip Potter <phil@philpotter.co.uk>
Signed-off-by: Ghanshyam Agrawal <ghanshyam1898@gmail.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: cx231xx: fix a memleak in cx231xx_init_isoc
Zhipeng Lu [Fri, 1 Dec 2023 13:22:55 +0000 (21:22 +0800)]
media: cx231xx: fix a memleak in cx231xx_init_isoc

The dma_q->p_left_data alloced by kzalloc should be freed in all the
following error handling paths. However, it hasn't been freed in the
allocation error paths of dev->video_mode.isoc_ctl.urb and
dev->video_mode.isoc_ctl.transfer_buffer.

On the other hand, the dma_q->p_left_data did be freed in the
error-handling paths after that of dev->video_mode.isoc_ctl.urb and
dev->video_mode.isoc_ctl.transfer_buffer, by calling
cx231xx_uninit_isoc(dev). So the same free operation should be done in
error-handling paths of those two allocation.

Fixes: 64fbf4445526 ("[media] cx231xx: Added support for Carraera, Shelby, RDx_253S and VIDEO_GRABBER")
Signed-off-by: Zhipeng Lu <alexious@zju.edu.cn>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agostaging: media: VIDEO_STARFIVE_CAMSS should depend on ARCH_STARFIVE
Geert Uytterhoeven [Tue, 28 Nov 2023 18:31:53 +0000 (19:31 +0100)]
staging: media: VIDEO_STARFIVE_CAMSS should depend on ARCH_STARFIVE

The StarFive Camera Subsystem is only present on the StarFive JH7110
SoC.  Hence add a dependency on ARCH_STARFIVE, to prevent asking the
user about this driver when configuring a kernel without StarFive SoC
support.

Fixes: bba185d141b1 ("media: staging: media: starfive: camss: Add core driver")
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Reviewed-by: Changhuang Liang <changhuang.liang@starfivetech.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: chips-media: wave5: VIDEO_WAVE_VPU should depend on ARCH_K3
Geert Uytterhoeven [Tue, 28 Nov 2023 18:26:36 +0000 (19:26 +0100)]
media: chips-media: wave5: VIDEO_WAVE_VPU should depend on ARCH_K3

The Chips&Media Wave 5 Series multi-standard codec IP is currently only
supported on Texas Instruments K3 architecture.  Hence add a dependency on
ARCH_K3, to prevent asking the user about this driver when configuring a
kernel without Texas Instruments K3 Multicore SoC support.

Fixes: 9707a6254a8a ("media: chips-media: wave5: Add the v4l2 layer")
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Reviewed-by: Nishanth Menon <nm@ti.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
[hverkuil: small update in commit log, and reformatted the Fixes tag]

16 months agomedia: staging: starfive: camss: fix off by one in isp_enum_mbus_code()
Dan Carpenter [Tue, 28 Nov 2023 14:40:14 +0000 (17:40 +0300)]
media: staging: starfive: camss: fix off by one in isp_enum_mbus_code()

These > comparisons should be >=.  The formats->fmts[] array is either
a pointer to isp_formats_sink[] or isp_formats_source[] respectively.

Fixes: e57854628f58 ("media: staging: media: starfive: camss: Add ISP driver")
Signed-off-by: Dan Carpenter <dan.carpenter@linaro.org>
Reviewed-by: Changhuang Liang <changhuang.liang@starfivetech.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: videobuf2-dma-sg: fix vmap callback
Michael Grzeschik [Thu, 23 Nov 2023 22:32:05 +0000 (23:32 +0100)]
media: videobuf2-dma-sg: fix vmap callback

For dmabuf import users to be able to use the vaddr from another
videobuf2-dma-sg source, the exporter needs to set a proper vaddr on
vb2_dma_sg_dmabuf_ops_vmap callback. This patch adds vmap on map if
buf->vaddr was not set.

Cc: stable@kernel.org
Fixes: 7938f4218168 ("dma-buf-map: Rename to iosys-map")
Signed-off-by: Michael Grzeschik <m.grzeschik@pengutronix.de>
Acked-by: Tomasz Figa <tfiga@chromium.org>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agodoc: media: visl: Add AV1 support
Detlev Casanova [Thu, 23 Nov 2023 15:52:26 +0000 (10:52 -0500)]
doc: media: visl: Add AV1 support

Add AV1 information in visl documentation.

Reviewed-by: Daniel Almeida <daniel.almeida@collabora.com>
Signed-off-by: Detlev Casanova <detlev.casanova@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: visl: Add AV1 support
Detlev Casanova [Thu, 23 Nov 2023 15:52:25 +0000 (10:52 -0500)]
media: visl: Add AV1 support

Let the visl test driver accept the AV1 pixel format.

Reviewed-by: Daniel Almeida <daniel.almeida@collabora.com>
Tested-by: Daniel Almeida <daniel.almeida@collabora.com>
Signed-off-by: Detlev Casanova <detlev.casanova@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: exynos-gsc: remove unused improper CONFIG definition
Lukas Bulwahn [Thu, 23 Nov 2023 10:38:08 +0000 (11:38 +0100)]
media: exynos-gsc: remove unused improper CONFIG definition

Defines prefixed with "CONFIG" should be limited to proper Kconfig options,
that are introduced in a Kconfig file.

In the driver code, there is a define for CONFIG_VB2_GSC_DMA_CONTIG, but
this is not used anywhere in the code.

Just remove this unused definition.

No functional change.

Signed-off-by: Lukas Bulwahn <lukas.bulwahn@gmail.com>
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: platform: mtk-mdp3: Use devicetree phandle to retrieve SCP
AngeloGioacchino Del Regno [Thu, 5 Oct 2023 10:49:05 +0000 (12:49 +0200)]
media: platform: mtk-mdp3: Use devicetree phandle to retrieve SCP

Instead of walking the entire parent node for something that has the
right compatible, use the scp_get() function provided by the MediaTek
SCP remoteproc driver to retrieve a handle to mtk_scp through the
devicetree "mediatek,scp" (phandle) property.

In case of multi-core SCP, this also allows to select a specific core.

Reviewed-by: Chen-Yu Tsai <wenst@chromium.org>
Signed-off-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Reviewed-by: Alexandre Mergnat <amergnat@baylibre.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: dt-bindings: mediatek: Add phandle to mediatek,scp on MDP3 RDMA
AngeloGioacchino Del Regno [Thu, 5 Oct 2023 10:49:04 +0000 (12:49 +0200)]
media: dt-bindings: mediatek: Add phandle to mediatek,scp on MDP3 RDMA

The MDP3 RDMA needs to communicate with the SCP remote processor: allow
specifying a phandle to a SCP core.

Signed-off-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Reviewed-by: Conor Dooley <conor.dooley@microchip.com>
Reviewed-by: Alexandre Mergnat <amergnat@baylibre.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: v4l: async: Drop useless list move operation
Sebastian Reichel [Mon, 13 Nov 2023 22:57:24 +0000 (23:57 +0100)]
media: v4l: async: Drop useless list move operation

v4l2_async_unbind_subdev_one(), which is called in the line
following the list_move() operation contains list_move_tail()
for the same entry and overrides anything list_move() did.
Thus it can be removed.

Signed-off-by: Sebastian Reichel <sre@kernel.org>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: v4l: async: Fix duplicated list deletion
Sebastian Reichel [Mon, 13 Nov 2023 22:57:23 +0000 (23:57 +0100)]
media: v4l: async: Fix duplicated list deletion

The list deletion call dropped here is already called from the
helper function in the line before. Having a second list_del()
call results in either a warning (with CONFIG_DEBUG_LIST=y):

list_del corruption, c46c8198->next is LIST_POISON1 (00000100)

If CONFIG_DEBUG_LIST is disabled the operation results in a
kernel error due to NULL pointer dereference.

Fixes: 28a1295795d8 ("media: v4l: async: Allow multiple connections between entities")
Signed-off-by: Sebastian Reichel <sre@kernel.org>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: v4l2-subdev: Fix indentation in v4l2-subdev.h
Laurent Pinchart [Mon, 23 Oct 2023 18:19:22 +0000 (21:19 +0300)]
media: v4l2-subdev: Fix indentation in v4l2-subdev.h

Fix a simple indentation issue in the v4l2-subdev.h header.

Fixes: f57fa2959244 ("media: v4l2-subdev: Add new ioctl for client capabilities")
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: ccs: Use V4L2 CCI for accessing sensor registers
Sakari Ailus [Tue, 7 Nov 2023 10:46:42 +0000 (12:46 +0200)]
media: ccs: Use V4L2 CCI for accessing sensor registers

Use V4L2 CCI for accessing device's registers. The 8-bit compatibility
read option is removed but this is supported by regmap through other
means.

Also the CCS register definitions are re-generated with V4L2 CCI
definitions. The older SMIA++ register definitions have been manually
converted.

Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: ccs: Better separate CCS static data access
Sakari Ailus [Tue, 7 Nov 2023 09:37:21 +0000 (11:37 +0200)]
media: ccs: Better separate CCS static data access

Separate CCS static data read-only register access in ccs-reg-access.c by
naming them differently.

The code in this file generally deals with reading and writing registers
where as static data (when it comes to ccs_static_data_read_ro_reg())
contains the read-only register values but no hardware registers are
accessed in that case.

Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: ccs: Generate V4L2 CCI compliant register definitions
Sakari Ailus [Tue, 7 Nov 2023 09:32:46 +0000 (11:32 +0200)]
media: ccs: Generate V4L2 CCI compliant register definitions

Generate register definitions that are fit for use with V4L2 CCI.

Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: v4l: cci: Add macros to obtain register width and address
Sakari Ailus [Tue, 7 Nov 2023 15:42:40 +0000 (17:42 +0200)]
media: v4l: cci: Add macros to obtain register width and address

Add CCI_REG_WIDTH() macro to obtain register width in bits and similarly,
CCI_REG_WIDTH_BYTES() to obtain it in bytes.

Also add CCI_REG_ADDR() macro to obtain the address of a register.

Use both macros in v4l2-cci.c, too.

Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: v4l: cci: Add driver-private bit definitions
Sakari Ailus [Tue, 7 Nov 2023 09:31:13 +0000 (11:31 +0200)]
media: v4l: cci: Add driver-private bit definitions

Provide a few bits for drivers to store private information on register
definitions.

Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: v4l: cci: Include linux/bits.h
Sakari Ailus [Tue, 7 Nov 2023 08:45:30 +0000 (10:45 +0200)]
media: v4l: cci: Include linux/bits.h

linux/bits.h is needed for GENMASK(). Include it.

Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: microchip-isc: Remove dead code in pipeline validation
Laurent Pinchart [Fri, 27 Oct 2023 09:16:43 +0000 (12:16 +0300)]
media: microchip-isc: Remove dead code in pipeline validation

The isc_try_fse() function, called from isc_validate(), takes two
parameters, an isc_device pointer, and a v4l2_subdev_state pointer. The
isc_device is accessed but not modified by the function. The state is
modified, including the struct v4l2_subdev_pad_config array it points
to, but they are then never used by the caller. Furthermore, the V4L2
subdev operation called by isc_try_fse() doesn't modify the subdev it is
called on. The isc_try_fse() function has thus no effect, and can just
be dropped.

Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Eugen Hristev <eugen.hristev@collabora.com>
Tested-by: Eugen Hristev <eugen.hristev@collabora.com>
[Sakari Ailus: Resolve conflicts due to API changes.]
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: Documentation: Initialisation finishes before subdev registration
Sakari Ailus [Wed, 1 Nov 2023 08:00:05 +0000 (10:00 +0200)]
media: Documentation: Initialisation finishes before subdev registration

Document that sub-device initialisation needs to complete before the async
sub-device is registered as there is no further driver action needed
before the sensor becomes accessible via the UAPI.

Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: imx319: Enable runtime PM before registering async sub-device
Sakari Ailus [Wed, 1 Nov 2023 08:22:20 +0000 (10:22 +0200)]
media: imx319: Enable runtime PM before registering async sub-device

As the sensor may be accessible right after its async sub-device is
registered, enable runtime PM before doing so.

Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: ccs: Print ireal and float limits converted to integers
Sakari Ailus [Fri, 27 Oct 2023 08:44:05 +0000 (11:44 +0300)]
media: ccs: Print ireal and float limits converted to integers

A number of CCS register value limits are in ireal or float format.
Also convert them to integers for easier interpretation.

Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: ov9734: Enable runtime PM before registering async sub-device
Bingbu Cao [Wed, 22 Nov 2023 09:46:09 +0000 (17:46 +0800)]
media: ov9734: Enable runtime PM before registering async sub-device

As the sensor device maybe accessible right after its async sub-device is
registered, such as ipu-bridge will try to power up sensor by sensor's
client device's runtime PM from the async notifier callback, if runtime PM
is not enabled, it will fail.

So runtime PM should be ready before its async sub-device is registered
and accessible by others.

Fixes: d3f863a63fe4 ("media: i2c: Add ov9734 image sensor driver")
Cc: stable@vger.kernel.org
Signed-off-by: Bingbu Cao <bingbu.cao@intel.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: ov13b10: Enable runtime PM before registering async sub-device
Bingbu Cao [Wed, 22 Nov 2023 09:46:08 +0000 (17:46 +0800)]
media: ov13b10: Enable runtime PM before registering async sub-device

As the sensor device maybe accessible right after its async sub-device is
registered, such as ipu-bridge will try to power up sensor by sensor's
client device's runtime PM from the async notifier callback, if runtime PM
is not enabled, it will fail.

So runtime PM should be ready before its async sub-device is registered
and accessible by others.

Fixes: 7ee850546822 ("media: Add sensor driver support for the ov13b10 camera.")
Cc: stable@vger.kernel.org
Signed-off-by: Bingbu Cao <bingbu.cao@intel.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: ov01a10: Enable runtime PM before registering async sub-device
Bingbu Cao [Wed, 22 Nov 2023 09:46:07 +0000 (17:46 +0800)]
media: ov01a10: Enable runtime PM before registering async sub-device

As the sensor device maybe accessible right after its async sub-device is
registered, such as ipu-bridge will try to power up sensor by sensor's
client device's runtime PM from the async notifier callback, if runtime PM
is not enabled, it will fail.

So runtime PM should be ready before its async sub-device is registered
and accessible by others.

It also sets the runtime PM status to active as the sensor was turned
on by i2c-core.

Fixes: 0827b58dabff ("media: i2c: add ov01a10 image sensor driver")
Cc: stable@vger.kernel.org
Signed-off-by: Bingbu Cao <bingbu.cao@intel.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: imx355: Enable runtime PM before registering async sub-device
Bingbu Cao [Wed, 22 Nov 2023 09:46:06 +0000 (17:46 +0800)]
media: imx355: Enable runtime PM before registering async sub-device

As the sensor device maybe accessible right after its async sub-device is
registered, such as ipu-bridge will try to power up sensor by sensor's
client device's runtime PM from the async notifier callback, if runtime PM
is not enabled, it will fail.

So runtime PM should be ready before its async sub-device is registered
and accessible by others.

Fixes: df0b5c4a7ddd ("media: add imx355 camera sensor driver")
Cc: stable@vger.kernel.org
Signed-off-by: Bingbu Cao <bingbu.cao@intel.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: stm32-dcmipp: STM32 DCMIPP camera interface driver
Hugues Fruchet [Mon, 27 Nov 2023 17:08:17 +0000 (18:08 +0100)]
media: stm32-dcmipp: STM32 DCMIPP camera interface driver

This V4L2 subdev driver enables Digital Camera Memory Interface
Pixel Processor(DCMIPP) of STMicroelectronics STM32 SoC series.

Signed-off-by: Hugues Fruchet <hugues.fruchet@foss.st.com>
Signed-off-by: Alain Volmat <alain.volmat@foss.st.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
[hverkuil: remove empty line at end of source]

16 months agomedia: MAINTAINERS: add entry for STM32 DCMIPP driver
Alain Volmat [Mon, 27 Nov 2023 17:08:16 +0000 (18:08 +0100)]
media: MAINTAINERS: add entry for STM32 DCMIPP driver

Add the entry related to the STM32 MEDIA DCMIPP driver within the
MAINTAINERS file.
Add myself as maintainer of the DCMI driver as well.

Signed-off-by: Alain Volmat <alain.volmat@foss.st.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
[Sakari Ailus: Arrange files alphabetically.]
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agodt-bindings: media: add bindings for stm32 dcmipp
Alain Volmat [Mon, 27 Nov 2023 17:08:15 +0000 (18:08 +0100)]
dt-bindings: media: add bindings for stm32 dcmipp

Add the yaml binding for the ST Microelectronics STM32 DCMIPP
(Digital Camera Memory Interface Pixel Processor)

Signed-off-by: Alain Volmat <alain.volmat@foss.st.com>
Reviewed-by: Conor Dooley <conor.dooley@microchip.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: i2c: st-vgxy61: add v4l2_fwnode ctrls parse and addition
Alain Volmat [Mon, 27 Nov 2023 13:51:12 +0000 (14:51 +0100)]
media: i2c: st-vgxy61: add v4l2_fwnode ctrls parse and addition

Allow parsing of the v4l2_fwnode properties from the DT and addition
of those properties (such as orientation, rotation).

Signed-off-by: Alain Volmat <alain.volmat@foss.st.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: i2c: st-vgxy61: Add V4L2_SUBDEV_FL_HAS_EVENTS and subscribe hooks
Alain Volmat [Mon, 27 Nov 2023 13:51:11 +0000 (14:51 +0100)]
media: i2c: st-vgxy61: Add V4L2_SUBDEV_FL_HAS_EVENTS and subscribe hooks

Any V4L2 subdevice that implements controls and declares
V4L2_SUBDEV_FL_HAS_DEVNODE should also declare V4L2_SUBDEV_FL_HAS_EVENTS
and implement subscribe_event and unsubscribe_event hooks.

With that done, v4l2-compliance testing is now ok.

Signed-off-by: Alain Volmat <alain.volmat@foss.st.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: i2c: st-mipid02: add Y8 format support
Alain Volmat [Sat, 25 Nov 2023 18:20:55 +0000 (19:20 +0100)]
media: i2c: st-mipid02: add Y8 format support

Add support of MEDIA_BUS_FMT_Y8_1X8.

Signed-off-by: Alain Volmat <alain.volmat@foss.st.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: i2c: st-mipid02: removal of unused link_frequency variable
Alain Volmat [Sat, 25 Nov 2023 18:20:54 +0000 (19:20 +0100)]
media: i2c: st-mipid02: removal of unused link_frequency variable

link_frequency variable within struct mipid02_dev seems to have never
been used hence remove it.

Signed-off-by: Alain Volmat <alain.volmat@foss.st.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: i2c: st-mipid02: use mipi-csi macro for data-type
Alain Volmat [Sat, 25 Nov 2023 18:20:53 +0000 (19:20 +0100)]
media: i2c: st-mipid02: use mipi-csi macro for data-type

Use MIPI data-type macros.

Signed-off-by: Alain Volmat <alain.volmat@foss.st.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: i2c: st-mipid02: use active state to store pad formats
Alain Volmat [Sat, 25 Nov 2023 18:20:52 +0000 (19:20 +0100)]
media: i2c: st-mipid02: use active state to store pad formats

Store formats information within pad allowing to simplify further more
the driver (mutex / format store within the driver structure no more
necessary).

Signed-off-by: Alain Volmat <alain.volmat@foss.st.com>
[Sakari Ailus: Address init_cfg -> init_state API change.]
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: i2c: st-mipid02: use cci_* helpers for register access.
Alain Volmat [Sat, 25 Nov 2023 18:20:51 +0000 (19:20 +0100)]
media: i2c: st-mipid02: use cci_* helpers for register access.

Use cci_read & cci_write functions for accessing registers.

Signed-off-by: Alain Volmat <alain.volmat@foss.st.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: i2c: st-mipid02: don't keep track of streaming status
Alain Volmat [Sat, 25 Nov 2023 18:20:50 +0000 (19:20 +0100)]
media: i2c: st-mipid02: don't keep track of streaming status

As explained in the following series, subdev do not have to keep
track of their streaming status:

https://lore.kernel.org/linux-media/20230914181704.4811-1-laurent.pinchart@ideasonboard.com/

Signed-off-by: Alain Volmat <alain.volmat@foss.st.com>
[Sakari Ailus: Remove redundant local variable in mipid02_set_fmt.]
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: i2c: st-mipid02: add usage of v4l2_get_link_freq
Alain Volmat [Sat, 25 Nov 2023 18:20:49 +0000 (19:20 +0100)]
media: i2c: st-mipid02: add usage of v4l2_get_link_freq

Use the helper v4l2_get_link_freq instead of performing manually
check of the LINK_FREQ or PIXELRATE ctrls.

Signed-off-by: Alain Volmat <alain.volmat@foss.st.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: i2c: st-mipid02: correct format propagation
Alain Volmat [Mon, 13 Nov 2023 14:57:30 +0000 (15:57 +0100)]
media: i2c: st-mipid02: correct format propagation

Use a copy of the struct v4l2_subdev_format when propagating
format from the sink to source pad in order to avoid impacting the
sink format returned to the application.

Thanks to Jacopo Mondi for pointing the issue.

Fixes: 6c01e6f3f27b ("media: st-mipid02: Propagate format from sink to source pad")
Signed-off-by: Alain Volmat <alain.volmat@foss.st.com>
Cc: stable@vger.kernel.org
Reviewed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Reviewed-by: Daniel Scally <dan.scally@ideasonboard.com>
Reviewed-by: Benjamin Mugnier <benjamin.mugnier@foss.st.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: v4l: fwnode: Parse MIPI DisCo for Imaging properties
Sakari Ailus [Tue, 31 Oct 2023 18:22:37 +0000 (20:22 +0200)]
media: v4l: fwnode: Parse MIPI DisCo for Imaging properties

Parse MIPI DisCo for Imaging properties "mipi-img-lens-focus" and
"mipi-img-flash-leds" for VCMs and flash LEDs.

Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: v4l2-subdev: Rename .init_cfg() operation to .init_state()
Laurent Pinchart [Mon, 27 Nov 2023 09:07:44 +0000 (11:07 +0200)]
media: v4l2-subdev: Rename .init_cfg() operation to .init_state()

The subdev .init_cfg() operation is affected by two issues:

- It has long been extended to initialize a whole v4l2_subdev_state
  instead of just a v4l2_subdev_pad_config, but its name has stuck
  around.

- Despite operating on a whole subdev state and not being directly
  exposed to the subdev users (either in-kernel or through the userspace
  API), .init_cfg() is categorized as a subdev pad operation.

This participates in making the subdev API confusing for new developers.
Fix it by renaming the operation to .init_state(), and make it a subdev
internal operation.

Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: Michael Riesch <michael.riesch@wolfvision.net> # for imx415
Acked-by: Shuah Khan <skhan@linuxfoundation.org> # for vimc
Reviewed-by: Philipp Zabel <p.zabel@pengutronix.de>
Reviewed-by: Tomi Valkeinen <tomi.valkeinen@ideasonboard.com>
[Sakari Ailus: Resolved a conflict in Renesas vsp1 driver.]
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: i2c: Add driver for THine THP7312
Paul Elder [Fri, 24 Nov 2023 14:26:25 +0000 (16:26 +0200)]
media: i2c: Add driver for THine THP7312

The THP7312 is an external camera ISP from THine. Add a V4L2 subdev
driver for it.

Signed-off-by: Paul Elder <paul.elder@ideasonboard.com>
Co-developed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
[Sakari Ailus: squash a patch to fix missing mutex_unlock by Laurent.]
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: uapi: Add controls for the THP7312 ISP
Laurent Pinchart [Fri, 24 Nov 2023 14:26:24 +0000 (16:26 +0200)]
media: uapi: Add controls for the THP7312 ISP

The THP7312 is an external ISP from THine. As such, it implements a
large number of parameters to control all aspects of the image
processing. Many of those controls are already standard in V4L2, but
some are fairly device-specific.

Reserve a range of 32 controls for the device. The driver will implement
4 device-specific controls to start with, define and document them. 28
additional device-specific controls should be enough for future
development.

Co-developed-by: Paul Elder <paul.elder@ideasonboard.com>
Signed-off-by: Paul Elder <paul.elder@ideasonboard.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agodt-bindings: media: Add bindings for THine THP7312 ISP
Paul Elder [Fri, 24 Nov 2023 14:26:23 +0000 (16:26 +0200)]
dt-bindings: media: Add bindings for THine THP7312 ISP

The THP7312 is an external ISP from THine. Add DT bindings for it.

Signed-off-by: Paul Elder <paul.elder@ideasonboard.com>
Co-developed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agodevice property: Add fwnode_name_eq()
Sakari Ailus [Mon, 6 Nov 2023 07:31:41 +0000 (09:31 +0200)]
device property: Add fwnode_name_eq()

Add fwnode_name_eq() to implement the functionality of of_node_name_eq()
on fwnode property API. The same convention of ending the comparison at
'@' (besides NUL) is applied on also both ACPI and swnode. The function
is intended for comparing unit address-less node names on DT and firmware
or swnodes compliant with DT bindings.

Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Tested-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: Rafael J. Wysocki <rafael@kernel.org>
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: i2c: gc2145: Galaxy Core GC2145 sensor support
Alain Volmat [Wed, 22 Nov 2023 07:51:49 +0000 (08:51 +0100)]
media: i2c: gc2145: Galaxy Core GC2145 sensor support

Addition of support for the Galaxy Core GC2145 XVGA sensor.
The sensor supports both DVP and CSI-2 interfaces however for
the time being only CSI-2 is implemented.

Configurations are currently based on initialization scripts
coming from Galaxy Core and so for that purpose only 3 static
resolutions are supported:
 - 640x480
 - 1280x720
 - 1600x1200

Signed-off-by: Alain Volmat <alain.volmat@foss.st.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agodt-bindings: media: i2c: add galaxycore,gc2145 dt-bindings
Alain Volmat [Wed, 22 Nov 2023 07:51:48 +0000 (08:51 +0100)]
dt-bindings: media: i2c: add galaxycore,gc2145 dt-bindings

Introduction of the Galaxy Core GC2145 XVGA CMOS camera sensor.

Signed-off-by: Alain Volmat <alain.volmat@foss.st.com>
Reviewed-by: Conor Dooley <conor.dooley@microchip.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agodt-bindings: vendor-prefixes: Add prefix for GalaxyCore Inc.
Alain Volmat [Wed, 22 Nov 2023 07:51:47 +0000 (08:51 +0100)]
dt-bindings: vendor-prefixes: Add prefix for GalaxyCore Inc.

Add a vendor prefix entry for galaxycore (https://www.gcoreinc.com)

Signed-off-by: Alain Volmat <alain.volmat@foss.st.com>
Acked-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Reviewed-by: Pavel Machek <pavel@ucw.cz>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agomedia: renesas: vsp1: Fix references to pad config
Laurent Pinchart [Sun, 26 Nov 2023 01:37:15 +0000 (03:37 +0200)]
media: renesas: vsp1: Fix references to pad config

V4L2 subdev operations have moved from operating on a
v4l2_subdev_pad_config to a v4l2_subdev_state a long time ago. Fix
remaining incorrect references to pad config in function and variable
names.

Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agoMerge tag 'v6.7-rc4' into media_stage
Hans Verkuil [Sun, 3 Dec 2023 15:35:17 +0000 (16:35 +0100)]
Merge tag 'v6.7-rc4' into media_stage

Linux 6.7-rc4

This is needed for a vsp1 fix that upcoming media patches depend on.

Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
16 months agoLinux 6.7-rc4
Linus Torvalds [Sun, 3 Dec 2023 09:52:56 +0000 (18:52 +0900)]
Linux 6.7-rc4

16 months agoMerge tag 'v6.7-rc3-smb3-client-fixes' of git://git.samba.org/sfrench/cifs-2.6
Linus Torvalds [Sun, 3 Dec 2023 00:08:26 +0000 (09:08 +0900)]
Merge tag 'v6.7-rc3-smb3-client-fixes' of git://git.samba.org/sfrench/cifs-2.6

Pull smb client fixes from Steve French:

 - Two fallocate fixes

 - Fix warnings from new gcc

 - Two symlink fixes

* tag 'v6.7-rc3-smb3-client-fixes' of git://git.samba.org/sfrench/cifs-2.6:
  smb: client, common: fix fortify warnings
  cifs: Fix FALLOC_FL_INSERT_RANGE by setting i_size after EOF moved
  cifs: Fix FALLOC_FL_ZERO_RANGE by setting i_size if EOF moved
  smb: client: report correct st_size for SMB and NFS symlinks
  smb: client: fix missing mode bits for SMB symlinks

16 months agoMerge tag 'firewire-fixes-6.7-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Sun, 3 Dec 2023 00:03:07 +0000 (09:03 +0900)]
Merge tag 'firewire-fixes-6.7-rc4' of git://git./linux/kernel/git/ieee1394/linux1394

Pull firewire fix from Takashi Sakamoto:
 "A single patch to fix long-standing issue of memory leak at failure of
  device registration for fw_unit. We rarely encounter the issue, but it
  should be applied to stable releases, since it fixes inappropriate API
  usage"

* tag 'firewire-fixes-6.7-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/ieee1394/linux1394:
  firewire: core: fix possible memory leak in create_units()

16 months agoMerge tag 'powerpc-6.7-3' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc...
Linus Torvalds [Sat, 2 Dec 2023 23:43:35 +0000 (08:43 +0900)]
Merge tag 'powerpc-6.7-3' of git://git./linux/kernel/git/powerpc/linux

Pull powerpc fixes from Michael Ellerman:

 - Fix corruption of f0/vs0 during FP/Vector save, seen as userspace
   crashes when using io-uring workers (in particular with MariaDB)

 - Fix KVM_RUN potentially clobbering all host userspace FP/Vector
   registers

Thanks to Timothy Pearson, Jens Axboe, and Nicholas Piggin.

* tag 'powerpc-6.7-3' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux:
  KVM: PPC: Book3S HV: Fix KVM_RUN clobbering FP/VEC user registers
  powerpc: Don't clobber f0/vs0 during fp|altivec register save

16 months agoMerge tag 'vfio-v6.7-rc4' of https://github.com/awilliam/linux-vfio
Linus Torvalds [Sat, 2 Dec 2023 23:37:39 +0000 (08:37 +0900)]
Merge tag 'vfio-v6.7-rc4' of https://github.com/awilliam/linux-vfio

Pull vfio fixes from Alex Williamson:

 - Fix the lifecycle of a mutex in the pds variant driver such that a
   reset prior to opening the device won't find it uninitialized.
   Implement the release path to symmetrically destroy the mutex. Also
   switch a different lock from spinlock to mutex as the code path has
   the potential to sleep and doesn't need the spinlock context
   otherwise (Brett Creeley)

 - Fix an issue detected via randconfig where KVM tries to symbol_get an
   undeclared function. The symbol is temporarily declared
   unconditionally here, which resolves the problem and avoids churn
   relative to a series pending for the next merge window which resolves
   some of this symbol ugliness, but also fixes Kconfig dependencies
   (Sean Christopherson)

* tag 'vfio-v6.7-rc4' of https://github.com/awilliam/linux-vfio:
  vfio: Drop vfio_file_iommu_group() stub to fudge around a KVM wart
  vfio/pds: Fix possible sleep while in atomic context
  vfio/pds: Fix mutex lock->magic != lock warning

16 months agoMerge tag 'for-linus-6.7a-rc4-tag' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Sat, 2 Dec 2023 23:31:53 +0000 (08:31 +0900)]
Merge tag 'for-linus-6.7a-rc4-tag' of git://git./linux/kernel/git/xen/tip

Pull xen fixes from Juergen Gross:

 - A fix for the Xen event driver setting the correct return value when
   experiencing an allocation failure

 - A fix for allocating space for a struct in the percpu area to not
   cross page boundaries (this one is for x86, a similar one for Arm was
   already in the pull request for rc3)

* tag 'for-linus-6.7a-rc4-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip:
  xen/events: fix error code in xen_bind_pirq_msi_to_irq()
  x86/xen: fix percpu vcpu_info allocation

16 months agoMerge tag 'probes-fixes-v6.7-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Sat, 2 Dec 2023 23:02:49 +0000 (08:02 +0900)]
Merge tag 'probes-fixes-v6.7-rc3' of git://git./linux/kernel/git/trace/linux-trace

Pull probes fixes from Masami Hiramatsu:

 - objpool: Fix objpool overrun case on memory/cache access delay
   especially on the big.LITTLE SoC. The objpool uses a copy of object
   slot index internal loop, but the slot index can be changed on
   another processor in parallel. In that case, the difference of 'head'
   local copy and the 'slot->last' index will be bigger than local slot
   size. In that case, we need to re-read the slot::head to update it.

 - kretprobe: Fix to use appropriate rcu API for kretprobe holder. Since
   kretprobe_holder::rp is RCU managed, it should use
   rcu_assign_pointer() and rcu_dereference_check() correctly. Also
   adding __rcu tag for finding wrong usage by sparse.

 - rethook: Fix to use appropriate rcu API for rethook::handler. The
   same as kretprobe, rethook::handler is RCU managed and it should use
   rcu_assign_pointer() and rcu_dereference_check(). This also adds
   __rcu tag for finding wrong usage by sparse.

* tag 'probes-fixes-v6.7-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace:
  rethook: Use __rcu pointer for rethook::handler
  kprobes: consistent rcu api usage for kretprobe holder
  lib: objpool: fix head overrun on RK3588 SBC

16 months agoMerge tag 'pm-6.7-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm
Linus Torvalds [Sat, 2 Dec 2023 00:01:00 +0000 (09:01 +0900)]
Merge tag 'pm-6.7-rc4' of git://git./linux/kernel/git/rafael/linux-pm

Pull power management fixes from Rafael Wysocki:
 "These fix issues in two cpufreq drivers, in the AMD P-state driver and
  in the power-capping DTPM framework.

  Specifics:

   - Fix the AMD P-state driver's EPP sysfs interface in the cases when
     the performance governor is in use (Ayush Jain)

   - Make the ->fast_switch() callback in the AMD P-state driver return
     the target frequency as expected (Gautham R. Shenoy)

   - Allow user space to control the range of frequencies to use via
     scaling_min_freq and scaling_max_freq when AMD P-state driver is in
     use (Wyes Karny)

   - Prevent power domains needed for wakeup signaling from being turned
     off during system suspend on Qualcomm systems and prevent
     performance states votes from runtime-suspended devices from being
     lost across a system suspend-resume cycle in qcom-cpufreq-nvmem
     (Stephan Gerhold)

   - Fix disabling the 792 Mhz OPP in the imx6q cpufreq driver for the
     i.MX6ULL types that can run at that frequency (Christoph
     Niedermaier)

   - Eliminate unnecessary and harmful conversions to uW from the DTPM
     (dynamic thermal and power management) framework (Lukasz Luba)"

* tag 'pm-6.7-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
  cpufreq/amd-pstate: Only print supported EPP values for performance governor
  cpufreq/amd-pstate: Fix scaling_min_freq and scaling_max_freq update
  powercap: DTPM: Fix unneeded conversions to micro-Watts
  cpufreq/amd-pstate: Fix the return value of amd_pstate_fast_switch()
  pmdomain: qcom: rpmpd: Set GENPD_FLAG_ACTIVE_WAKEUP
  cpufreq: qcom-nvmem: Preserve PM domain votes in system suspend
  cpufreq: qcom-nvmem: Enable virtual power domain devices
  cpufreq: imx6q: Don't disable 792 Mhz OPP unnecessarily

16 months agoMerge tag 'acpi-6.7-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael...
Linus Torvalds [Fri, 1 Dec 2023 23:52:20 +0000 (08:52 +0900)]
Merge tag 'acpi-6.7-rc4' of git://git./linux/kernel/git/rafael/linux-pm

Pull ACPI fixes from Rafael Wysocki:
 "This fixes a recently introduced build issue on ARM32 and a NULL
  pointer dereference in the ACPI backlight driver due to a design issue
  exposed by a recent change in the ACPI bus type code.

  Specifics:

   - Fix a recently introduced build issue on ARM32 platforms caused by
     an inadvertent header file breakage (Dave Jiang)

   - Eliminate questionable usage of acpi_driver_data() in the ACPI
     backlight cooling device code that leads to NULL pointer
     dereferences after recent ACPI core changes (Hans de Goede)"

* tag 'acpi-6.7-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
  ACPI: video: Use acpi_video_device for cooling-dev driver data
  ACPI: Fix ARM32 platforms compile issue introduced by fw_table changes

16 months agoMerge tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux
Linus Torvalds [Fri, 1 Dec 2023 23:48:59 +0000 (08:48 +0900)]
Merge tag 'arm64-fixes' of git://git./linux/kernel/git/arm64/linux

Pull arm64 fix from Catalin Marinas:
 "Fix a regression where the arm64 KPTI ends up enabled even on systems
  that don't need it"

* tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux:
  arm64: Avoid enabling KPTI unnecessarily

16 months agoMerge tag 'iommu-fixes-v6.7-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Fri, 1 Dec 2023 23:42:39 +0000 (08:42 +0900)]
Merge tag 'iommu-fixes-v6.7-rc3' of git://git./linux/kernel/git/joro/iommu

Pull iommu fixes from Joerg Roedel:

 - Fix race conditions in device probe path

 - Handle ERR_PTR() returns in __iommu_domain_alloc() path

 - Update MAINTAINERS entry for Qualcom IOMMUs

 - Printk argument fix in device tree specific code

 - Several Intel VT-d fixes from Lu Baolu:
     - Do not support enforcing cache coherency for non-empty domains
     - Avoid devTLB invalidation if iommu is off
     - Disable PCI ATS in legacy passthrough mode
     - Support non-PCI devices when clearing context
     - Fix incorrect cache invalidation for mm notification
     - Add MTL to quirk list to skip TE disabling
     - Set variable intel_dirty_ops to static

* tag 'iommu-fixes-v6.7-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/joro/iommu:
  iommu: Fix printk arg in of_iommu_get_resv_regions()
  iommu/vt-d: Set variable intel_dirty_ops to static
  iommu/vt-d: Fix incorrect cache invalidation for mm notification
  iommu/vt-d: Add MTL to quirk list to skip TE disabling
  iommu/vt-d: Make context clearing consistent with context mapping
  iommu/vt-d: Disable PCI ATS in legacy passthrough mode
  iommu/vt-d: Omit devTLB invalidation requests when TES=0
  iommu/vt-d: Support enforce_cache_coherency only for empty domains
  iommu: Avoid more races around device probe
  MAINTAINERS: list all Qualcomm IOMMU drivers in the QUALCOMM IOMMU entry
  iommu: Flow ERR_PTR out from __iommu_domain_alloc()

16 months agoMerge tag 'sound-6.7-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai...
Linus Torvalds [Fri, 1 Dec 2023 23:33:29 +0000 (08:33 +0900)]
Merge tag 'sound-6.7-rc4' of git://git./linux/kernel/git/tiwai/sound

Pull sound fixes from Takashi Iwai:
 "No surprise here, including only a collection of HD-audio
  device-specific small fixes"

* tag 'sound-6.7-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound:
  ALSA: hda: Disable power-save on KONTRON SinglePC
  ALSA: hda/realtek: Add supported ALC257 for ChromeOS
  ALSA: hda/realtek: Headset Mic VREF to 100%
  ALSA: hda: intel-nhlt: Ignore vbps when looking for DMIC 32 bps format
  ALSA: hda: cs35l56: Enable low-power hibernation mode on SPI
  ALSA: cs35l41: Fix for old systems which do not support command
  ALSA: hda: cs35l41: Remove unnecessary boolean state variable firmware_running
  ALSA: hda - Fix speaker and headset mic pin config for CHUWI CoreBook XPro

16 months agoMerge tag 'drm-fixes-2023-12-01' of git://anongit.freedesktop.org/drm/drm
Linus Torvalds [Fri, 1 Dec 2023 23:18:59 +0000 (08:18 +0900)]
Merge tag 'drm-fixes-2023-12-01' of git://anongit.freedesktop.org/drm/drm

Pull drm fixes from Dave Airlie:
 "Weekly fixes, mostly amdgpu fixes with a scattering of nouveau, i915,
  and a couple of reverts. Hopefully it will quieten down in coming
  weeks.

  drm:
   - Revert unexport of prime helpers for fd/handle conversion

  dma_resv:
   - Do not double add fences in dma_resv_add_fence.

  gpuvm:
   - Fix GPUVM license identifier.

  i915:
   - Mark internal GSC engine with reserved uabi class
   - Take VGA converters into account in eDP probe
   - Fix intel_pre_plane_updates() call to ensure workarounds get applied

  panel:
   - Revert panel fixes as they require exporting device_is_dependent.

  nouveau:
   - fix oversized allocations in new vm path
   - fix zero-length array
   - remove a stray lock

  nt36523:
   - Fix error check for nt36523.

  amdgpu:
   - DMUB fix
   - DCN 3.5 fixes
   - XGMI fix
   - DCN 3.2 fixes
   - Vangogh suspend fix
   - NBIO 7.9 fix
   - GFX11 golden register fix
   - Backlight fix
   - NBIO 7.11 fix
   - IB test overflow fix
   - DCN 3.1.4 fixes
   - fix a runtime pm ref count
   - Retimer fix
   - ABM fix
   - DCN 3.1.5 fix
   - Fix AGP addressing
   - Fix possible memory leak in SMU error path
   - Make sure PME is enabled in D3
   - Fix possible NULL pointer dereference in debugfs
   - EEPROM fix
   - GC 9.4.3 fix

  amdkfd:
   - IP version check fix
   - Fix memory leak in pqm_uninit()"

* tag 'drm-fixes-2023-12-01' of git://anongit.freedesktop.org/drm/drm: (53 commits)
  Revert "drm/prime: Unexport helpers for fd/handle conversion"
  drm/amdgpu: Use another offset for GC 9.4.3 remap
  drm/amd/display: Fix some HostVM parameters in DML
  drm/amdkfd: Free gang_ctx_bo and wptr_bo in pqm_uninit
  drm/amdgpu: Update EEPROM I2C address for smu v13_0_0
  drm/amd/display: Allow DTBCLK disable for DCN35
  drm/amdgpu: Fix cat debugfs amdgpu_regs_didt causes kernel null pointer
  drm/amd: Enable PCIe PME from D3
  drm/amd/pm: fix a memleak in aldebaran_tables_init
  drm/amdgpu: fix AGP addressing when GART is not at 0
  drm/amd/display: update dcn315 lpddr pstate latency
  drm/amd/display: fix ABM disablement
  drm/amd/display: Fix black screen on video playback with embedded panel
  drm/amd/display: Fix conversions between bytes and KB
  drm/amdkfd: Use common function for IP version check
  drm/amd/display: Remove config update
  drm/amd/display: Update DCN35 clock table policy
  drm/amd/display: force toggle rate wa for first link training for a retimer
  drm/amdgpu: correct the amdgpu runtime dereference usage count
  drm/amd/display: Update min Z8 residency time to 2100 for DCN314
  ...

16 months agoMerge tag 'io_uring-6.7-2023-11-30' of git://git.kernel.dk/linux
Linus Torvalds [Fri, 1 Dec 2023 21:47:32 +0000 (06:47 +0900)]
Merge tag 'io_uring-6.7-2023-11-30' of git://git.kernel.dk/linux

Pull io_uring fixes from Jens Axboe:

 - Fix an issue with discontig page checking for IORING_SETUP_NO_MMAP

 - Fix an issue with not allowing IORING_SETUP_NO_MMAP also disallowing
   mmap'ed buffer rings

 - Fix an issue with deferred release of memory mapped pages

 - Fix a lockdep issue with IORING_SETUP_NO_MMAP

 - Use fget/fput consistently, even from our sync system calls. No real
   issue here, but if we were ever to allow closing io_uring descriptors
   it would be required. Let's play it safe and just use the full ref
   counted versions upfront. Most uses of io_uring are threaded anyway,
   and hence already doing the full version underneath.

* tag 'io_uring-6.7-2023-11-30' of git://git.kernel.dk/linux:
  io_uring: use fget/fput consistently
  io_uring: free io_buffer_list entries via RCU
  io_uring/kbuf: prune deferred locked cache when tearing down
  io_uring/kbuf: recycle freed mapped buffer ring entries
  io_uring/kbuf: defer release of mapped buffer rings
  io_uring: enable io_mem_alloc/free to be used in other parts
  io_uring: don't guard IORING_OFF_PBUF_RING with SETUP_NO_MMAP
  io_uring: don't allow discontig pages for IORING_SETUP_NO_MMAP

16 months agoMerge tag 'block-6.7-2023-12-01' of git://git.kernel.dk/linux
Linus Torvalds [Fri, 1 Dec 2023 21:39:30 +0000 (06:39 +0900)]
Merge tag 'block-6.7-2023-12-01' of git://git.kernel.dk/linux

Pull block fixes from Jens Axboe:

 - NVMe pull request via Keith:
     - Invalid namespace identification error handling (Marizio Ewan,
       Keith)
     - Fabrics keep-alive tuning (Mark)

 - Fix for a bad error check regression in bcache (Markus)

 - Fix for a performance regression with O_DIRECT (Ming)

 - Fix for a flush related deadlock (Ming)

 - Make the read-only warn on per-partition (Yu)

* tag 'block-6.7-2023-12-01' of git://git.kernel.dk/linux:
  nvme-core: check for too small lba shift
  blk-mq: don't count completed flush data request as inflight in case of quiesce
  block: Document the role of the two attribute groups
  block: warn once for each partition in bio_check_ro()
  block: move .bd_inode into 1st cacheline of block_device
  nvme: check for valid nvme_identify_ns() before using it
  nvme-core: fix a memory leak in nvme_ns_info_from_identify()
  nvme: fine-tune sending of first keep-alive
  bcache: revert replacing IS_ERR_OR_NULL with IS_ERR

16 months agoMerge tag 'dm-6.7/dm-fixes-2' of git://git.kernel.org/pub/scm/linux/kernel/git/device...
Linus Torvalds [Fri, 1 Dec 2023 21:32:29 +0000 (06:32 +0900)]
Merge tag 'dm-6.7/dm-fixes-2' of git://git./linux/kernel/git/device-mapper/linux-dm

Pull device mapper fixes from Mike Snitzer:

 - Fix DM verity target's FEC support to always initialize IO before it
   frees it. Also fix alignment of struct dm_verity_fec_io within the
   per-bio-data

 - Fix DM verity target to not FEC failed readahead IO

 - Update DM flakey target to use MAX_ORDER rather than MAX_ORDER - 1

* tag 'dm-6.7/dm-fixes-2' of git://git.kernel.org/pub/scm/linux/kernel/git/device-mapper/linux-dm:
  dm-flakey: start allocating with MAX_ORDER
  dm-verity: align struct dm_verity_fec_io properly
  dm verity: don't perform FEC for failed readahead IO
  dm verity: initialize fec io before freeing it

16 months agoMerge tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi
Linus Torvalds [Fri, 1 Dec 2023 21:27:20 +0000 (06:27 +0900)]
Merge tag 'scsi-fixes' of git://git./linux/kernel/git/jejb/scsi

Pull SCSI fixes from James Bottomley:
 "Three small fixes, one in drivers.

  The core changes are to the internal representation of flags in
  scsi_devices which removes space wasting bools in favour of single bit
  flags and to add a flag to force a runtime resume which is used by ATA
  devices"

* tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi:
  scsi: sd: Fix system start for ATA devices
  scsi: Change SCSI device boolean fields to single bit flags
  scsi: ufs: core: Clear cmd if abort succeeds in MCQ mode

16 months agoMerge tag 'fs_for_v6.7-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/jack...
Linus Torvalds [Fri, 1 Dec 2023 21:19:27 +0000 (06:19 +0900)]
Merge tag 'fs_for_v6.7-rc4' of git://git./linux/kernel/git/jack/linux-fs

Pull ext2 fix from Jan Kara:
 "Fix an ext2 bug introduced by changes in ext2 & iomap stepping on each
  other toes (apparently ext2 driver does not get much testing in
  linux-next)"

* tag 'fs_for_v6.7-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs:
  ext2: Fix ki_pos update for DIO buffered-io fallback case

16 months agoMerge tag 'bcachefs-2023-11-29' of https://evilpiepirate.org/git/bcachefs
Linus Torvalds [Fri, 1 Dec 2023 21:02:16 +0000 (06:02 +0900)]
Merge tag 'bcachefs-2023-11-29' of https://evilpiepirate.org/git/bcachefs

Pull more bcachefs bugfixes from Kent Overstreet:

 - bcache & bcachefs were broken with CFI enabled; patch for closures to
   fix type punning

 - mark erasure coding as extra-experimental; there are incompatible
   disk space accounting changes coming for erasure coding, and I'm
   still seeing checksum errors in some tests

 - several fixes for durability-related issues (durability is a device
   specific setting where we can tell bcachefs that data on a given
   device should be counted as replicated x times)

 - a fix for a rare livelock when a btree node merge then updates a
   parent node that is almost full

 - fix a race in the device removal path, where dropping a pointer in a
   btree node to a device would be clobbered by an in flight btree write
   updating the btree node key on completion

 - fix one SRCU lock hold time warning in the btree gc code - ther's
   still a bunch more of these to fix

 - fix a rare race where we'd start copygc before initializing the "are
   we rw" percpu refcount; copygc would think we were already ro and die
   immediately

* tag 'bcachefs-2023-11-29' of https://evilpiepirate.org/git/bcachefs: (23 commits)
  bcachefs: Extra kthread_should_stop() calls for copygc
  bcachefs: Convert gc_alloc_start() to for_each_btree_key2()
  bcachefs: Fix race between btree writes and metadata drop
  bcachefs: move journal seq assertion
  bcachefs: -EROFS doesn't count as move_extent_start_fail
  bcachefs: trace_move_extent_start_fail() now includes errcode
  bcachefs: Fix split_race livelock
  bcachefs: Fix bucket data type for stripe buckets
  bcachefs: Add missing validation for jset_entry_data_usage
  bcachefs: Fix zstd compress workspace size
  bcachefs: bpos is misaligned on big endian
  bcachefs: Fix ec + durability calculation
  bcachefs: Data update path won't accidentaly grow replicas
  bcachefs: deallocate_extra_replicas()
  bcachefs: Proper refcounting for journal_keys
  bcachefs: preserve device path as device name
  bcachefs: Fix an endianness conversion
  bcachefs: Start gc, copygc, rebalance threads after initing writes ref
  bcachefs: Don't stop copygc thread on device resize
  bcachefs: Make sure bch2_move_ratelimit() also waits for move_ops
  ...