linux-toradex.git/drivers/media/platform/rockchip/rkisp1, branch master

media: rkisp1: Fix enum_framesizes accepting invalid pixel formats

2026-03-24T21:14:44+00:00

Reject unsupported pixel formats in rkisp1_enum_framesizes() to
fix v4l2-compliance failure.

v4l2-compliance test failure:

fail: ../utils/v4l2-compliance/v4l2-test-formats.cpp(403): Accepted framesize for invalid format
test VIDIOC_ENUM_FMT/FRAMESIZES/FRAMEINTERVALS: FAIL

Tested on: Debix i.MX8MP Model A
Kernel version: v6.17-rc3
v4l2-compliance: 1.31.0-5387

Signed-off-by: Tarang Raval 
Tested-by: Laurent Pinchart 
Reviewed-by: Laurent Pinchart 
Tested-by: Kieran Bingham 
Link: https://patch.msgid.link/20250829101425.95442-1-tarang.raval@siliconsignals.io
Signed-off-by: Laurent Pinchart 
Signed-off-by: Hans Verkuil

media: rkisp1: Fix filter mode register configuration

2026-01-21T07:25:45+00:00

The rkisp1_flt_config() function performs an initial direct write to
RKISP1_CIF_ISP_FILT_MODE without including the RKISP1_CIF_ISP_FLT_ENA
bit, which clears the filter enable bit in the hardware.

The subsequent read/modify/write sequence then reads back the register
with the enable bit already cleared and cannot restore it, resulting in
the filter being inadvertently disabled.

Remove the redundant direct write. The read/modify/write sequence alone
correctly preserves the existing enable bit state while updating the
DNR mode and filter configuration bits.

Signed-off-by: Rui Wang 
Reviewed-by: Stefan Klug 
Reviewed-by: Kieran Bingham 
Reviewed-by: Laurent Pinchart 
Link: https://patch.msgid.link/20260105171142.147792-2-rui.wang@ideasonboard.com
Signed-off-by: Laurent Pinchart 
Signed-off-by: Hans Verkuil

media: rkisp1: Discard pm_runtime_put() return value

2026-01-21T07:25:45+00:00

Printing error messages on pm_runtime_put() returning negative values
is not particularly useful.

Returning an error code from pm_runtime_put() merely means that it has
not queued up a work item to check whether or not the device can be
suspended and there are many perfectly valid situations in which that
can happen, like after writing "on" to the devices' runtime PM "control"
attribute in sysfs for one example.

Accordingly, update rkisp1_vb2_stop_streaming() to simply discard the
return value of pm_runtime_put().

This will facilitate a planned change of the pm_runtime_put() return
type to void in the future.

Signed-off-by: Rafael J. Wysocki 
Reviewed-by: Laurent Pinchart 
Link: https://patch.msgid.link/2356323.iZASKD2KPV@rafael.j.wysocki
Signed-off-by: Laurent Pinchart 
Signed-off-by: Hans Verkuil

media: v4l2-isp: Rename block_info to block_type_info

2025-11-14T14:48:49+00:00

The v4l2_isp_params_block_info structure contains validation information
that apply to a block -type- and not only to a specific ISP block
implementation.

Clarify this by renaming v4l2_isp_params_block_info in
v4l2_isp_params_block_type_info and update the documentation and the
users of v4l2-isp accordingly.

Signed-off-by: Jacopo Mondi 
Reviewed-by: Laurent Pinchart 
Reviewed-by: Sakari Ailus 
Signed-off-by: Hans Verkuil

media: rkisp1: Use v4l2-isp for validation

2025-11-14T14:48:48+00:00

Convert rkisp1-params.c to use the helpers defined in v4l2-isp.h
to perform validation of a ISP parameters buffer.

Reviewed-by: Daniel Scally 
Acked-by: Sakari Ailus 
Reviewed-by: Laurent Pinchart 
Tested-by: Lad Prabhakar 
Signed-off-by: Jacopo Mondi 
Signed-off-by: Hans Verkuil

media: rkisp1: Improve frame sequence correctness on stats and params buffers

2025-11-03T10:40:22+00:00

On the rkisp1 (in my case on a NXP i.MX8 M Plus) the ISP interrupt
handler is sometimes called with RKISP1_CIF_ISP_V_START (start of frame)
and RKISP1_CIF_ISP_FRAME (end of frame) being set at the same time. In
commit 8524fa22fd2f ("media: staging: rkisp1: isp: add a warning and
debugfs var for irq delay") a warning was added for that. There are two
cases where this condition can occur:

1. The v-sync and the frame-end belong to the same frame. This means,
   the irq was heavily delayed and the warning is likely appropriate.

2. The v-sync belongs to the next frame. This can happen if the vertical
   blanking between two frames is very short.

The current code always handles case 1 although case 2 is in my
experience the more common case and happens in regular usage. This leads
to incorrect sequence numbers on stats and params buffers which in turn
breaks the regulation in user space. Fix that by adding a frame_active
flag to distinguish between these cases and handle the start of frame
either at the beginning or at the end of the rkisp1_isp_isr().

Signed-off-by: Stefan Klug 
Reviewed-by: Laurent Pinchart 
Reviewed-by: Kieran Bingham 
Reviewed-by: Jacopo Mondi 
Link: https://patch.msgid.link/20250922182003.3712101-2-stefan.klug@ideasonboard.com
Signed-off-by: Laurent Pinchart 
Signed-off-by: Hans Verkuil

media: rkisp1: Use %pe format specifier

2025-10-17T09:31:16+00:00

The %pe format specifier is designed to print error pointers. It prints
a symbolic error name (eg. -EINVAL) and it makes the code simpler by
omitting PTR_ERR().

This patch fixes this cocci report:
./platform/rockchip/rkisp1/rkisp1-isp.c:940:3-10: WARNING: Consider using %pe to print PTR_ERR()
./platform/rockchip/rkisp1/rkisp1-csi.c:372:3-10: WARNING: Consider using %pe to print PTR_ERR()

Signed-off-by: Ricardo Ribalda 
Acked-by: Sakari Ailus 
Signed-off-by: Hans Verkuil

media: rkisp1: Add support for multiple power domains

2025-08-31T09:10:07+00:00

The ISP instances in the NXP i.MX8MP need two power domains. While
single power domains are managed automatically by the device core,
support for multiple power domains requires manually attaching to the
power domains. Do so based on platform data.

Link: https://lore.kernel.org/r/20250616011115.19515-6-laurent.pinchart@ideasonboard.com
Signed-off-by: Laurent Pinchart 
Reviewed-by: Jacopo Mondi 
Signed-off-by: Hans Verkuil

media: rkisp1: Acquire pclk clock on i.MX8MP

2025-08-31T09:10:07+00:00

The ISP instances in the NXP i.MX8MP need the input pixel clock to be
enabled in order to access the HDR stitching registers. The clock should
ideally be mandatory, but that would break backward compatibility with
old DT. Try to acquire it as an optional clock instead.

Link: https://lore.kernel.org/r/20250616011115.19515-5-laurent.pinchart@ideasonboard.com
Signed-off-by: Laurent Pinchart 
Reviewed-by: Jacopo Mondi 
Signed-off-by: Hans Verkuil

media: rkisp1: Refactor clocks initialization

2025-08-31T09:10:07+00:00

ISP instances in different SoCs differ in the number of clocks they use,
but not in the clock names. Refactor clocks initialization to avoid
duplicating the clock names per platform, and lower the total number of
clocks from 8 to 4 as no platform uses more than 4 clocks. Use a static
assert to ensure at build time that the size of the arrays match.

Link: https://lore.kernel.org/r/20250616011115.19515-4-laurent.pinchart@ideasonboard.com
Signed-off-by: Laurent Pinchart 
Reviewed-by: Jacopo Mondi 
Signed-off-by: Hans Verkuil