We use dma_get_mask() of auxdev device for calculate iova pfn limit.
This is always 32 bit mask as we do not initialize the mask (and we can
not do so, since dev->dev_mask is NULL anyways for auxdev).
Since we need 31 bit mask for non-secure mode use mmu_info->aperture_end
which is properly initialized to correct mask for both modes.
Fixes: daabc5c647 ("media: ipu6: not override the dma_ops of device in driver")
Cc: stable@vger.kernel.org
Signed-off-by: Stanislaw Gruszka <stanislaw.gruszka@linux.intel.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
The clock frequency is often available via the clock itself and not read
by the driver from the "clock-frequency" property. Don't complain if the
property doesn't exist.
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
Generally any deviance from the standard is handled via CCS static data
nowadays and so not having quirks is expected. Don't warn about it.
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
The I²C retry hack has been there in order to address transient I²C
register write access issues on a few very old sensors and possibly it has
addressed also first I²C access problems (device not responding until a
certain amount of time has passed) but that is now separately handled. The
retry hack has a good potential for introducing hard to debug problems in
updating sensor settings while streaming. Remove it and instead pass those
rare errors to the user space -- which is also what virtually all other
drivers do.
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
Use read_poll_timeout() in polling the device after a reset, either hard
or soft. While at it, improve the related error message.
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
Some sensors take longer to respond after reset than the spec-required
time. Try up to 1 s for the sensor to become accessible.
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
Use pm_runtime_put_autosuspend() instead of pm_runtime_put()
to allow autosuspend. Set a 1000ms autosuspend delay in
imx219_probe() to improve power efficiency.
Signed-off-by: Tarang Raval <tarang.raval@siliconsignals.io>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
pcim_iomap_table() and pcim_iomap_regions() have been deprecated.
Furthermore, the "name" parameter in pcim_iomap_regions() and its
successor, pcim_iomap_region(), should always reflect the driver name,
whereas currently it is the device's name.
Replace the deprecated functions with pcim_iomap_region() and pass the
actual driver name.
Signed-off-by: Philipp Stanner <phasta@kernel.org>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
pcim_iomap_table() and pcim_iomap_regions() have been deprecated.
Replace them with pcim_iomap_region(), and pass the actual driver name
to that function.
Signed-off-by: Philipp Stanner <phasta@kernel.org>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
When shutting down capture extra care was needed to try and complete a
buffer that was involved in the emulated support for SEQ_{TB,BT}. This
was needed as a buffer might be queued once to the driver, but two times
to the hardware using different offsets.
As support for SEQ_{TB,BT} is now removed this shutdown process can be
greatly simplified. And in addition the state keeping of the VIN device
can be reduced to a single boolean value instead of keeping track of
this SEQ_{TB,BT} stopping dance.
Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Reviewed-by: Tomi Valkeinen <tomi.valkeinen+renesas@ideasonboard.com>
Tested-by: Tomi Valkeinen <tomi.valkeinen+renesas@ideasonboard.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
The STARTING state is superfluous and can be replaced with a check of
the sequence counter. The design idea is that the first buffer returned
from the driver have to come from the first hardware buffer slot.
Failing this the first 3 buffers queued to the device can be returned
out-of-order.
But it's much clearer to check the sequence counter to only return the
first buffer if it comes from hardware slot 0 then it is to carry around
an extra state just for this. Remove the unneeded state and replace it
with a simpler check.
Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Reviewed-by: Tomi Valkeinen <tomi.valkeinen+renesas@ideasonboard.com>
Tested-by: Tomi Valkeinen <tomi.valkeinen+renesas@ideasonboard.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
The VIN state of suspended is superfluous. The logic was that when the
device were suspended and in a RUNNING state the state was set to
SUSPENDED. And when resuming it checked if the state is SUSPENDED and if
so started the device and changed it to RUNNING.
This can be avoided by simply checking if the device is in a RUNNING
state at both suspend and resume callbacks. Remove the unneeded
complexity.
Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Reviewed-by: Tomi Valkeinen <tomi.valkeinen+renesas@ideasonboard.com>
Tested-by: Tomi Valkeinen <tomi.valkeinen+renesas@ideasonboard.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
When the driver was converted from soc_camera software support for
V4L2_FIELD_SEQ_TB and V4L2_FIELD_SEQ_BT were added. This was done by
capturing twice to the same VB2 buffer, but at different offsets.
This turned out to be a bad idea and it never really worked properly in
all situations. As the hardware can't support this mode natively remove
trying to emulate it in software. It's still possible to capture TOP or
BOTTOM fields separately or both ALTERNATING. If user-space wants the
same fields in the same buffer the same hack to capture twice to the
same buffer can be done.
Removing this error prone emulated support pave ways in future work to
simplify the internal buffer handling and making it less fragile, while
enabling adding support for other features the hardware actually
supports.
Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Reviewed-by: Tomi Valkeinen <tomi.valkeinen+renesas@ideasonboard.com>
Tested-by: Tomi Valkeinen <tomi.valkeinen+renesas@ideasonboard.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
Add multiplexed streams support. CAL has 8 DMA-engines and can capture 8
separate streams at the same time. The driver filters the streams based
on CSI-2 virtual channel number and datatype. CAL may have (depending on
the SoC) two CSI-2 RX blocks, which share the 8 DMA-engines, so the
number of capturable streams does not change even if there are two CSI-2
RX blocks.
Add 8 video device nodes, each representing a single DMA-engine, and set
the number of source pads on CSI-2 RX blocks to 8. Each video node can be
connected to any of the source pads on either of the CSI-2 RX instances
using media links. CSI-2 RX block's subdevice internal routing is used
to route the incoming CSI-2 streams to one of the 8 source pads.
Only video data streams are supported at the moment.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ideasonboard.com>
Reviewed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
When v4l2_subdev_init_finalize() fails no changes have been made to
the runtime-pm device state yet, so the probe_error_media_entity_cleanup
rollback path should not touch the runtime-pm device state.
Instead this should be done from the probe_error_v4l2_subdev_cleanup
rollback path. Note the pm_runtime_xxx() calls are put above
the v4l2_subdev_cleanup() call to have the reverse call order of probe().
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Bingbu Cao <bingbu.cao@intel.com>
Fixes: 289c25923e ("media: ov2740: Use sub-device active state")
Cc: stable@vger.kernel.org
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
Extract values needed by ipu6_queue_buf_ready() function from fw abi
structure. This will allow to reuse same buf ready code when fw abi
change.
Signed-off-by: Stanislaw Gruszka <stanislaw.gruszka@linux.intel.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
Document the CCS PLL flags with short comments. The CCS spec has more
information on them while the added documentation helps finding the
relevant information in the CCS spec.
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
Print the PLL calculator flags right away when the PLL calculator is
called. Previously this was done only in a successful case and that didn't
really help solving a problem when one happened.
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
The LINK_DECOUPLED flag isn't used by the PLL calculator other than
printing it. The number of OP/VT lanes are already printed in any case.
Thus drop the flag as it has no function.
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
In general the CCS PLL calculator prints debugging information on the
process to ease debugging. This case was not annotated, do that now.
Remove an extra multiplication while at it.
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
The PLL calculator does a search of the PLL configuration space for all
valid OP pre-PLL clock dividers. The maximum did not take into account the
CCS PLL flag CCS_PLL_FLAG_EXT_IP_PLL_DIVIDER in which case also odd PLL
dividers (other than 1) are valid. Do that now.
Fixes: 4e1e8d240d ("media: ccs-pll: Add support for extended input PLL clock divider")
Cc: stable@vger.kernel.org
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
The ccs_pll_calculate_vt_tree() function does a search over possible VT
PLL configurations to find the "best" one. If the sensor does not support
odd pre-PLL divisors and the minimum value (with constraints) isn't 1,
other odd values could be errorneously searched (and selected) for the
pre-PLL divisor. Fix this.
Fixes: 415ddd9939 ("media: ccs-pll: Split limits and PLL configuration into front and back parts")
Cc: stable@vger.kernel.org
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
The ccs_pll_calculate() function does a search over possible PLL
configurations to find the "best" one. If the sensor does not support odd
pre-PLL divisors and the minimum value (with constraints) isn't 1, other
odd values could be errorneously searched (and selected) for the pre-PLL
divisor. Fix this.
Fixes: 415ddd9939 ("media: ccs-pll: Split limits and PLL configuration into front and back parts")
Cc: stable@vger.kernel.org
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
