aboutsummaryrefslogtreecommitdiff
path: root/drivers/media/platform/rockchip
AgeCommit message (Collapse)AuthorFilesLines
2024-08-14media: rkisp1: Add support for the companding blockPaul Elder1-0/+168
Add support to the rkisp1 driver for the companding block that exists on the i.MX8MP version of the ISP. This requires usage of the new extensible parameters format, and showcases how the format allows for extensions without breaking backward compatibility. Signed-off-by: Paul Elder <[email protected]> Reviewed-by: Jacopo Mondi <[email protected]> Reviewed-by: Paul Elder <[email protected]> Signed-off-by: Jacopo Mondi <[email protected]> Tested-by: Kieran Bingham <[email protected]> Acked-by: Sakari Ailus <[email protected]> Signed-off-by: Laurent Pinchart <[email protected]>
2024-08-14media: rkisp1: Add feature flags for BLS and compandPaul Elder3-3/+17
Add feature flags for the dedicated black level subtraction hardware block and for the compand hardware block. The companding feature flag is added on its own (as opposed to "the absence of BLS") because we will need it later for when we add support for the companding block. Skip BLS configuration when the BLS feature flag is unset, as devices without the dedicated BLS block cannot configure a hardware block that doesn't exist. Signed-off-by: Paul Elder <[email protected]> Reviewed-by: Jacopo Mondi <[email protected]> Signed-off-by: Jacopo Mondi <[email protected]> Tested-by: Kieran Bingham <[email protected]> Acked-by: Sakari Ailus <[email protected]> Signed-off-by: Laurent Pinchart <[email protected]>
2024-08-14media: rkisp1: Add register definitions for the companding blockPaul Elder1-0/+23
To prepare for adding support for the companding block to the rkisp1 driver for the version of the ISP on the i.MX8MP, add the register definitions for it, including relevant register field values. Signed-off-by: Paul Elder <[email protected]> Reviewed-by: Jacopo Mondi <[email protected]> Signed-off-by: Jacopo Mondi <[email protected]> Tested-by: Kieran Bingham <[email protected]> Acked-by: Sakari Ailus <[email protected]> Signed-off-by: Laurent Pinchart <[email protected]>
2024-08-14media: rkisp1: Add features mask to extensible block handlersLaurent Pinchart1-1/+9
Future ISP parameter blocks for i.MX8MP-specific features will not support on Rockchip platforms as they lack the corresponding hardware. Introduce a features mask in the extensible block handlers to indicate which device features a block require, and ignore blocks that require unavailable features. Reviewed-by: Jacopo Mondi <[email protected]> Reviewed-by: Paul Elder <[email protected]> Signed-off-by: Jacopo Mondi <[email protected]> Tested-by: Kieran Bingham <[email protected]> Acked-by: Sakari Ailus <[email protected]> Signed-off-by: Laurent Pinchart <[email protected]>
2024-08-14media: rkisp1: Add helper function to swap colour channelsLaurent Pinchart4-85/+45
The BLS parameters passed by userspace are specified for named colour channels (R, Gr, Gb and B), while the hardware registers reference positions in the 2x2 CFA pattern (A, B, C and D). The BLS values are swapped based on the CFA pattern when writing to or reading from registers, using hand-roled switch statements. The logic is duplicated already, and new code will require similar processing. Move the swap logic to a shared function, using static data to control the channels order. Reviewed-by: Jacopo Mondi <[email protected]> Reviewed-by: Paul Elder <[email protected]> Signed-off-by: Jacopo Mondi <[email protected]> Tested-by: Kieran Bingham <[email protected]> Acked-by: Sakari Ailus <[email protected]> Signed-off-by: Laurent Pinchart <[email protected]>
2024-08-14media: rkisp1: Implement s_fmt/try_fmtJacopo Mondi1-8/+50
Implement in the rkisp1 driver support for the s_fmt and try_fmt operation to allow userspace to select between the extensible and the fixed parameters formats. Implement enum_mbus_code to enumerate the fixed and the extensible formats and disallow changing the data format while the queue is busy. Signed-off-by: Jacopo Mondi <[email protected]> Reviewed-by: Laurent Pinchart <[email protected]> Reviewed-by: Paul Elder <[email protected]> Tested-by: Kieran Bingham <[email protected]> Acked-by: Sakari Ailus <[email protected]> Signed-off-by: Laurent Pinchart <[email protected]>
2024-08-14media: rkisp1: Implement extensible params supportJacopo Mondi2-11/+622
Implement support in rkisp1-params for the extensible configuration parameters format. Create a list of handlers for each ISP block that wraps the existing configuration functions and handles the ISP block enablement. Parse the configuration parameters buffer in rkisp1_ext_params_config and filter the enable blocks by group, to allow setting the 'other' groups separately from the 'lsc' group to support the pre/post-configure operations. Implement parameter buffer validation for the extensible format at .buf_prepare() time. Signed-off-by: Jacopo Mondi <[email protected]> Reviewed-by: Laurent Pinchart <[email protected]> Reviewed-by: Jacopo Mondi <[email protected]> Reviewed-by: Paul Elder <[email protected]> Tested-by: Kieran Bingham <[email protected]> Acked-by: Sakari Ailus <[email protected]> Signed-off-by: Laurent Pinchart <[email protected]>
2024-08-12media: rkisp1: Cache the currently active formatJacopo Mondi2-22/+38
The rkisp1-params driver assumes the data buffer format is the only currently supported "fixed" one. The usage of the "fixed" format is assumed when allocating memory for the scratch buffers and when initializing the vb2 queue. In order to prepare to support the "extensible" format beside the existing "fixed" one, add support in the driver for both formats by caching a pointer to the active one in the driver structure and use it in the vb2 queue operations and subdev pad operations implementations. Do not yet allow userspace to select between the two formats as the support for the "extensible" format parsing will be introduced in a later patch in the series. While at it, document the un-documented ycbcr_encoding field of struct rkisp1_params_ops. Signed-off-by: Jacopo Mondi <[email protected]> Reviewed-by: Laurent Pinchart <[email protected]> Reviewed-by: Paul Elder <[email protected]> Tested-by: Kieran Bingham <[email protected]> Acked-by: Sakari Ailus <[email protected]> Signed-off-by: Laurent Pinchart <[email protected]>
2024-08-12media: rkisp1: Copy the parameters bufferJacopo Mondi2-31/+60
The ISP parameters buffers are queued by userspace to the params video device and appended by the driver to the list of available buffers for later consumption. As the parameters buffer is mapped in the userspace process memory, applications have access to the buffer content after the buffer has been queued. To prevent userspace from modifying the contents of the parameters buffer after it has been queued to the video device, add to 'struct rkisp1_params_buffer' a scratch buffer where to copy the parameters. Allocate the scratch buffer in the vb2 buf_init() operation and copy the buffer content in the buf_prepare() operation. Free the scratch buffer in the newly introduced buf_cleanup() operation handler. Modify the ISP configuration function to access the ISP configuration from the cached copy of the parameters buffer instead of using the userspace-mapped one. Signed-off-by: Jacopo Mondi <[email protected]> Reviewed-by: Laurent Pinchart <[email protected]> Reviewed-by: Paul Elder <[email protected]> Tested-by: Kieran Bingham <[email protected]> Acked-by: Sakari Ailus <[email protected]> Signed-off-by: Laurent Pinchart <[email protected]>
2024-08-12media: rkisp1: Add struct rkisp1_params_bufferJacopo Mondi2-11/+24
Create the 'struct rkisp1_params_buffer' type that wraps a vb2_v4l2_buffer to prepare to hold a copy of the parameters buffer that will be used to cache the user-provided configuration buffer in the following patches. Replace usage of 'struct rkisp1_buffer' with 'struct rkisp1_params_buffer' in rkisp1-params.c to prepare for that. Signed-off-by: Jacopo Mondi <[email protected]> Reviewed-by: Daniel Scally <[email protected]> Reviewed-by: Laurent Pinchart <[email protected]> Reviewed-by: Paul Elder <[email protected]> Tested-by: Kieran Bingham <[email protected]> Acked-by: Sakari Ailus <[email protected]> Signed-off-by: Laurent Pinchart <[email protected]>
2024-08-12media: rkisp1: Adapt to different SoCs having different size limitsOndrej Jirman5-11/+24
- RK3399 has input/output limit of main path 4416 x 3312 - PX30 has input/output limit of main path 3264 x 2448 - i.MX8MP has input/output limit of main path 4096 x 3072 Use rkisp1_info struct to encode the limits. Signed-off-by: Ondrej Jirman <[email protected]> Reviewed-by: Laurent Pinchart <[email protected]> Reviewed-by: Paul Elder <[email protected]> Signed-off-by: Laurent Pinchart <[email protected]>
2024-06-04media: rkisp1: Mark subdev state pointers as constLaurent Pinchart2-8/+8
Several subdev state pointers are used to access the state read-only. Mark them as const. Signed-off-by: Laurent Pinchart <[email protected]> Reviewed-by: Tomi Valkeinen <[email protected]> Reviewed-by: Nicolas Dufresne <[email protected]> Signed-off-by: Sakari Ailus <[email protected]> Signed-off-by: Hans Verkuil <[email protected]>
2024-05-31media: rockchip: rga: fix sequence number handlingMichael Tretter3-0/+12
The RGA driver didn't set the sequence numbers of the returned buffers. Keep track of the CAPTURE and OUTPUT sequence numbers, and set the sequence numbers in the source and destination buffers. Signed-off-by: Michael Tretter <[email protected]> Signed-off-by: Hans Verkuil <[email protected]>
2024-03-15Merge tag 'media/v6.9-1' of ↵Linus Torvalds6-65/+443
git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/linux-media Pull media updates from Mauro Carvalho Chehab: - DVB budget legacy API was finally documented. It took only 20+ years to get some documentation about it... - hantro driver has gained support for STM32MP25 VDEC/VENC - rkisp1 has gained support for i.MX8MP - atomisp got rid of two items from its todo list. Still 5 items pending for moving it out of staging - lots of driver fixes, cleanups and improvements * tag 'media/v6.9-1' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/linux-media: (252 commits) media: rcar-isp: Disallow unbind of devices media: usbtv: Remove useless locks in usbtv_video_free() media: mediatek: vcodec: avoid -Wcast-function-type-strict warning media: ttpci: fix two memleaks in budget_av_attach media: go7007: fix a memleak in go7007_load_encoder media: dvb-frontends: avoid stack overflow warnings with clang media: pvrusb2: fix uaf in pvr2_context_set_notify media: usb: s2255: Refactor s2255_get_fx2fw media: ti: j721e-csi2rx: Convert to platform remove callback returning void media: stm32-dcmipp: Convert to platform remove callback returning void media: nxp: imx8-isi: Convert to platform remove callback returning void media: nuvoton: Convert to platform remove callback returning void media: chips-media: wave5: Convert to platform remove callback returning void media: chips-media: wave5: Remove unnecessary semicolons media: i2c: imx290: Fix IMX920 typo media: platform: replace of_graph_get_next_endpoint() media: i2c: replace of_graph_get_next_endpoint() media: ivsc: csi: Make use of sub-device state media: ivsc: csi: Swap SINK and SOURCE pads media: ipu-bridge: Serialise calls to IPU bridge init ...
2024-02-23media: rkisp1: Add match data for i.MX8MP ISPPaul Elder1-0/+24
Add match data to the rkisp1 driver to match the i.MX8MP ISP. Signed-off-by: Paul Elder <[email protected]> Reviewed-by: Rob Herring <[email protected]> Tested-by: Alexander Stein <[email protected]> Tested-by: Adam Ford <[email protected]> Reviewed-by: Tomi Valkeinen <[email protected]> Reviewed-by: Laurent Pinchart <[email protected]> Signed-off-by: Laurent Pinchart <[email protected]>
2024-02-23media: rkisp1: Fix endianness on raw streams on i.MX8MPPaul Elder2-12/+85
The i.MX8MP has extra register fields in the memory interface control register for setting the output format, which work with the output alignment format register for byte-swapping and LSB/MSB alignment. With processed and 8-bit raw streams, it doesn't cause any problems to not set these, but with raw streams of higher bit depth the endianness is swapped and the data is not aligned properly. Add support for setting these registers and plumb them in to fix this. While at it, reflow a comment related to the forced configuration update. Signed-off-by: Paul Elder <[email protected]> Signed-off-by: Laurent Pinchart <[email protected]> Tested-by: Alexander Stein <[email protected]>
2024-02-23media: rkisp1: Add UYVY as an output formatPaul Elder1-0/+41
Add support for UYVY as an output format. The uv_swap bit in the MI_XTD_FORMAT_CTRL register that is used for the NV formats does not work for packed YUV formats. Thus, UYVY support is implemented via byte-swapping. This method clearly does not work for implementing support for YVYU and VYUY. Signed-off-by: Paul Elder <[email protected]> Reviewed-by: Laurent Pinchart <[email protected]> Signed-off-by: Laurent Pinchart <[email protected]> Tested-by: Alexander Stein <[email protected]> Tested-by: Adam Ford <[email protected]>
2024-02-23media: rkisp1: Add YC swap capabilityPaul Elder1-7/+19
The ISP version in the i.MX8MP has an MI_OUTPUT_ALIGN_FORMAT register that the rk3399 does not have. This register allows swapping bytes, which can be used to implement UYVY from YUYV. Add a flag to the format info in the list of formats supported by the capture v4l2 devices, and update enum_fmt and s_fmt to take it into account. To signify the presence of this feature, reuse the MAIN_STRIDE feature flag, as it is very likely that any ISP version that supports one of these two features will also support the other. Signed-off-by: Paul Elder <[email protected]> Reviewed-by: Laurent Pinchart <[email protected]> Signed-off-by: Laurent Pinchart <[email protected]> Tested-by: Alexander Stein <[email protected]> Tested-by: Adam Ford <[email protected]>
2024-02-23media: rkisp1: Support i.MX8MP's 34-bit DMAPaul Elder3-10/+22
On the ISP that is integrated in the i.MX8MP, DMA addresses have been extended to 34 bits, with the 32 MSBs stored in the DMA address registers and the 2 LSBs set to 0. To support this: - Shift the addresses to the right by 2 when writing to registers - Set the dma mask to 34 bits - Use dma_addr_t instead of u32 when storing the addresses Signed-off-by: Paul Elder <[email protected]> Reviewed-by: Laurent Pinchart <[email protected]> Signed-off-by: Tomi Valkeinen <[email protected]> Tested-by: Alexander Stein <[email protected]> Tested-by: Adam Ford <[email protected]> Reviewed-by: Tomi Valkeinen <[email protected]> Signed-off-by: Laurent Pinchart <[email protected]>
2024-02-23media: rkisp1: Configure gasket on i.MX8MPLaurent Pinchart3-3/+149
The i.MX8MP has a gasket between the CSI-2 receiver and the ISP. Configure and enable it when starting the ISP, and disable it when stopping. Signed-off-by: Laurent Pinchart <[email protected]> Signed-off-by: Paul Elder <[email protected]> Tested-by: Alexander Stein <[email protected]> Tested-by: Adam Ford <[email protected]> Reviewed-by: Tomi Valkeinen <[email protected]>
2024-02-23media: rkisp1: Support devices lacking dual cropPaul Elder3-7/+14
Some versions of the ISP supported by the rkisp1 driver, such as the ISP in the i.MX8MP, lack the dual crop registers and don't support cropping at the resizer input. They instead rely on cropping in the Image Stabilization module, at the output of the ISP, to modify the resizer input size and implement digital zoom. Add a dual crop feature flag to distinguish the versions of the ISP that support dual crop from those that don't, and make sure that the sink crop is set to the sink format when dual crop is not supported. Signed-off-by: Paul Elder <[email protected]> Tested-by: Alexander Stein <[email protected]> Tested-by: Adam Ford <[email protected]> Reviewed-by: Tomi Valkeinen <[email protected]> Reviewed-by: Laurent Pinchart <[email protected]> Signed-off-by: Laurent Pinchart <[email protected]>
2024-02-23media: rkisp1: Support devices lacking self pathPaul Elder4-8/+31
Some versions of the ISP supported by the rkisp1 driver, such as the ISP in the i.MX8MP, lack the self path. Support those ISP versions by adding a self path feature flag, and massage the rest of the driver to support the lack of a self path. Signed-off-by: Paul Elder <[email protected]> Reviewed-by: Laurent Pinchart <[email protected]> Tested-by: Alexander Stein <[email protected]> Tested-by: Adam Ford <[email protected]> Reviewed-by: Tomi Valkeinen <[email protected]> Signed-off-by: Laurent Pinchart <[email protected]>
2024-02-23media: rkisp1: Support setting memory stride for main pathPaul Elder3-15/+52
Some versions of the ISP supported by the rkisp1 driver, such as the ISP in the i.MX8MP, implement configurable memory stride for the main path the same way as already implemented by the driver for the self path. Support this feature by adding a main stride feature flag and program the corresponding registers accordingly. Signed-off-by: Laurent Pinchart <[email protected]> Signed-off-by: Paul Elder <[email protected]> Reviewed-by: Laurent Pinchart <[email protected]> Tested-by: Alexander Stein <[email protected]> Tested-by: Adam Ford <[email protected]> Reviewed-by: Tomi Valkeinen <[email protected]>
2024-02-23media: rkisp1: Add and use rkisp1_has_feature() macroLaurent Pinchart2-6/+9
Simplify feature tests with a macro that shortens lines. Signed-off-by: Laurent Pinchart <[email protected]> Reviewed-by: Paul Elder <[email protected]> Reviewed-by: Kieran Bingham <[email protected]> Signed-off-by: Paul Elder <[email protected]> Tested-by: Alexander Stein <[email protected]> Tested-by: Adam Ford <[email protected]> Reviewed-by: Tomi Valkeinen <[email protected]>
2024-02-01media: rkisp1: Fix IRQ handling due to shared interruptsTomi Valkeinen5-0/+33
The driver requests the interrupts as IRQF_SHARED, so the interrupt handlers can be called at any time. If such a call happens while the ISP is powered down, the SoC will hang as the driver tries to access the ISP registers. This can be reproduced even without the platform sharing the IRQ line: Enable CONFIG_DEBUG_SHIRQ and unload the driver, and the board will hang. Fix this by adding a new field, 'irqs_enabled', which is used to bail out from the interrupt handler when the ISP is not operational. Link: https://lore.kernel.org/r/[email protected] Signed-off-by: Tomi Valkeinen <[email protected]> Signed-off-by: Laurent Pinchart <[email protected]> Signed-off-by: Mauro Carvalho Chehab <[email protected]>
2024-02-01media: Revert "media: rkisp1: Drop IRQF_SHARED"Tomi Valkeinen1-1/+1
This reverts commit 85d2a31fe4d9be1555f621ead7a520d8791e0f74. The rkisp1 does share interrupt lines on some platforms, after all. Thus we need to revert this, and implement a fix for the rkisp1 shared irq handling in a follow-up patch. Closes: https://lore.kernel.org/all/[email protected]/ Link: https://lore.kernel.org/r/[email protected] Reported-by: Mikhail Rudenko <[email protected]> Signed-off-by: Tomi Valkeinen <[email protected]> Signed-off-by: Laurent Pinchart <[email protected]> Signed-off-by: Mauro Carvalho Chehab <[email protected]>
2023-12-13media: videobuf2: core: Rename min_buffers_needed field in vb2_queueBenjamin Gaignard1-1/+1
Rename min_buffers_needed into min_queued_buffers and update the documentation about it. Signed-off-by: Benjamin Gaignard <[email protected]> Signed-off-by: Hans Verkuil <[email protected]> [hverkuil: Drop the change where min_queued_buffers + 1 buffers would be] [hverkuil: allocated. Now this patch only renames this field instead of making] [hverkuil: a functional change as well.] [hverkuil: Renamed 3 remaining min_buffers_needed occurrences.]
2023-12-09media: rkisp1: resizer: Stop manual allocation of v4l2_subdev_stateLaurent Pinchart1-14/+24
Supported media bus codes on the resizer sink pad are identical to the ISP source pad. The .enum_mbus_code() handler thus delegates the enumeration to the ISP's operation. This is problematic for two reasons: - Format enumeration on the ISP source pad is dependent on the format configured on the ISP sink pad for the same subdev state (TRY or ACTIVE), while format enumeration on the resizer sink pad should return all formats supported by the resizer subdev, regardless of the ISP configuration. - Delegating the operation involves creating a fake v4l2_subdev_state on the stack to pass to the ISP .enum_mbus_code() handler. This gets in the way of evolution of both the ISP enumeration handler and, more generally, the V4L2 subdev state infrastructure. Fix those two issues by implementing format enumeration manually for the resizer. Link: https://lore.kernel.org/r/[email protected] Reviewed-by: Paul Elder <[email protected]> Reviewed-by: Tomi Valkeinen <[email protected]> Signed-off-by: Laurent Pinchart <[email protected]> Signed-off-by: Mauro Carvalho Chehab <[email protected]>
2023-12-09media: rkisp1: debug: Count completed frame interruptsPaul Elder3-0/+5
Add a counter to debugfs to count the number of frame-end interrupts. Link: https://lore.kernel.org/r/[email protected] Signed-off-by: Paul Elder <[email protected]> Reviewed-by: Alexander Stein <[email protected]> Reviewed-by: Kieran Bingham <[email protected]> Reviewed-by: Tomi Valkeinen <[email protected]> Signed-off-by: Laurent Pinchart <[email protected]> Signed-off-by: Mauro Carvalho Chehab <[email protected]>
2023-12-09media: rkisp1: debug: Add register dump for ISPaul Elder1-0/+4
Add register dump for the ISP image stabilizer module to debugfs. This helps debugging issues related to digital zoom. Link: https://lore.kernel.org/r/[email protected] Signed-off-by: Paul Elder <[email protected]> Tested-by: Alexander Stein <[email protected]> Reviewed-by: Kieran Bingham <[email protected]> Reviewed-by: Tomi Valkeinen <[email protected]> Signed-off-by: Laurent Pinchart <[email protected]> Signed-off-by: Mauro Carvalho Chehab <[email protected]>
2023-12-09media: rkisp1: regs: Consolidate MI interrupt wrap fieldsPaul Elder1-6/+3
Consolidate the wraparound fields in the memory interface interrupt status registers, so that it can be more succinctly expressed by taking the stream ID (main or self) as a parameter. Link: https://lore.kernel.org/r/[email protected] Signed-off-by: Paul Elder <[email protected]> Reviewed-by: Kieran Bingham <[email protected]> Reviewed-by: Tomi Valkeinen <[email protected]> Signed-off-by: Laurent Pinchart <[email protected]> Signed-off-by: Mauro Carvalho Chehab <[email protected]>
2023-12-09media: rkisp1: Fix IRQ disable race issueTomi Valkeinen2-4/+30
In rkisp1_isp_stop() and rkisp1_csi_disable() the driver masks the interrupts and then apparently assumes that the interrupt handler won't be running, and proceeds in the stop procedure. This is not the case, as the interrupt handler can already be running, which would lead to the ISP being disabled while the interrupt handler handling a captured frame. This brings up two issues: 1) the ISP could be powered off while the interrupt handler is still running and accessing registers, leading to board lockup, and 2) the interrupt handler code and the code that disables the streaming might do things that conflict. It is not clear to me if 2) causes a real issue, but 1) can be seen with a suitable delay (or printk in my case) in the interrupt handler, leading to board lockup. Link: https://lore.kernel.org/r/[email protected] Tested-by: Adam Ford <[email protected]> #imx8mp-beacon Signed-off-by: Tomi Valkeinen <[email protected]> Signed-off-by: Laurent Pinchart <[email protected]> Signed-off-by: Mauro Carvalho Chehab <[email protected]>
2023-12-09media: rkisp1: Store IRQ linesTomi Valkeinen2-5/+23
Store the IRQ lines used by the driver for easy access. These are needed in future patches which fix IRQ race issues. Link: https://lore.kernel.org/r/[email protected] Tested-by: Adam Ford <[email protected]> #imx8mp-beacon Signed-off-by: Tomi Valkeinen <[email protected]> Signed-off-by: Laurent Pinchart <[email protected]> Signed-off-by: Mauro Carvalho Chehab <[email protected]>
2023-12-09media: rkisp1: Fix IRQ handler return valuesTomi Valkeinen1-4/+12
The IRQ handler rkisp1_isr() calls sub-handlers, all of which returns an irqreturn_t value, but rkisp1_isr() ignores those values and always returns IRQ_HANDLED. Fix this by collecting the return values, and returning IRQ_HANDLED or IRQ_NONE as appropriate. Link: https://lore.kernel.org/r/[email protected] Tested-by: Adam Ford <[email protected]> #imx8mp-beacon Signed-off-by: Tomi Valkeinen <[email protected]> Signed-off-by: Laurent Pinchart <[email protected]> Signed-off-by: Mauro Carvalho Chehab <[email protected]>
2023-12-09media: rkisp1: Drop IRQF_SHAREDTomi Valkeinen1-1/+1
In all known platforms the ISP has dedicated IRQ lines, but for some reason the driver uses IRQF_SHARED. Supporting IRQF_SHARED properly requires handling interrupts even when our device is disabled, and the driver does not handle this. To avoid adding such code, and to be sure the driver won't accidentally be used in a platform with shared interrupts, let's drop the IRQF_SHARED flag. Link: https://lore.kernel.org/r/[email protected] Tested-by: Adam Ford <[email protected]> #imx8mp-beacon Signed-off-by: Tomi Valkeinen <[email protected]> Signed-off-by: Laurent Pinchart <[email protected]> Signed-off-by: Mauro Carvalho Chehab <[email protected]>
2023-12-09media: rkisp1: Fix memory leaks in rkisp1_isp_unregister()Tomi Valkeinen1-0/+1
Add missing call to v4l2_subdev_cleanup() to fix memory leak. Link: https://lore.kernel.org/r/[email protected] Fixes: 2cce0a369dbd ("media: rkisp1: isp: Use V4L2 subdev active state") Reviewed-by: Tommaso Merciai <[email protected]> Signed-off-by: Tomi Valkeinen <[email protected]> Signed-off-by: Laurent Pinchart <[email protected]> Signed-off-by: Mauro Carvalho Chehab <[email protected]>
2023-12-09media: rkisp1: Fix media device memory leakTomi Valkeinen1-1/+5
Add missing calls to media_device_cleanup() to fix memory leak. Link: https://lore.kernel.org/r/[email protected] Fixes: d65dd85281fb ("media: staging: rkisp1: add Rockchip ISP1 base driver") Reviewed-by: Tommaso Merciai <[email protected]> Signed-off-by: Tomi Valkeinen <[email protected]> Signed-off-by: Laurent Pinchart <[email protected]> Signed-off-by: Mauro Carvalho Chehab <[email protected]>
2023-12-07media: rockchip: rga: add NV12M supportMichael Tretter1-0/+9
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 <[email protected]> Signed-off-by: Hans Verkuil <[email protected]>
2023-12-07media: rockchip: rga: rework buffer handling for multi-planar formatsMichael Tretter3-11/+58
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 <[email protected]> Signed-off-by: Hans Verkuil <[email protected]>
2023-12-07media: rockchip: rga: switch to multi-planar APIMichael Tretter1-31/+29
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 <[email protected]> Signed-off-by: Hans Verkuil <[email protected]>
2023-12-07media: rockchip: rga: use macros for testing buffer typeMichael Tretter1-12/+9
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 <[email protected]> Signed-off-by: Hans Verkuil <[email protected]>
2023-12-07media: rockchip: rga: add local variable for pix_formatMichael Tretter1-23/+32
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 <[email protected]> Signed-off-by: Hans Verkuil <[email protected]>
2023-12-07media: rockchip: rga: use pixelformat to find formatMichael Tretter1-11/+7
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 <[email protected]> Signed-off-by: Hans Verkuil <[email protected]>
2023-12-07media: rockchip: rga: use clamp() to clamp size to limitsMichael Tretter1-9/+4
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 <[email protected]> Signed-off-by: Hans Verkuil <[email protected]>
2023-12-07media: rockchip: rga: set dma mask to 32 bitsMichael Tretter1-0/+8
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 <[email protected]> Signed-off-by: Hans Verkuil <[email protected]>
2023-12-07media: rockchip: rga: pre-calculate plane offsetsMichael Tretter3-21/+40
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 <[email protected]> Signed-off-by: Hans Verkuil <[email protected]>
2023-12-07media: rockchip: rga: split src and dst buffer setupMichael Tretter1-26/+68
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 <[email protected]> Signed-off-by: Hans Verkuil <[email protected]>
2023-12-07media: rockchip: rga: allocate DMA descriptors per bufferMichael Tretter4-66/+74
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 <[email protected]> Signed-off-by: Hans Verkuil <[email protected]>
2023-12-07media: rockchip: rga: extract helper to fill descriptorsMichael Tretter3-25/+30
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 <[email protected]> Signed-off-by: Hans Verkuil <[email protected]>
2023-12-07media: rockchip: rga: fix swizzling for RGB formatsMichael Tretter1-12/+3
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 <[email protected]> Signed-off-by: Hans Verkuil <[email protected]>
Powered by cgit, Gruvboxed by Blaster4385.