Message ID | 20230516095517.611711-1-sakari.ailus@linux.intel.com (mailing list archive) |
---|---|
Headers | show |
Series | Separate links and async sub-devices | expand |
Hi Sakari, thanks for the update. That's awesome, device setup on media-ctl works again on TQMa6x (i.MX6). For the record, below is what the media-ctl dot looks like. I was able to capture data on ipu1_csi0_capture device. Tested-by: Alexander Stein <alexander.stein@ew.tq-group.com> #imx6q Best regards, Alexander --8<-- digraph board { rankdir=TB n00000001 [label="{{<port0> 0} | ipu1_csi0\n/dev/v4l-subdev0 | {<port1> 1 | <port2> 2}}", shape=Mrecord, style=filled, fillcolor=green] n00000001:port2 -> n00000005 n00000001:port1 -> n0000000f:port0 [style=dashed] n00000001:port1 -> n0000000b:port0 [style=dashed] n00000005 [label="ipu1_csi0 capture\n/dev/video4", shape=box, style=filled, fillcolor=yellow] n0000000b [label="{{<port0> 0 | <port1> 1} | ipu1_vdic\n/dev/v4l- subdev1 | {<port2> 2}}", shape=Mrecord, style=filled, fillcolor=green] n0000000b:port2 -> n0000000f:port0 [style=dashed] n0000000f [label="{{<port0> 0} | ipu1_ic_prp\n/dev/v4l-subdev2 | {<port1> 1 | <port2> 2}}", shape=Mrecord, style=filled, fillcolor=green] n0000000f:port1 -> n00000013:port0 [style=dashed] n0000000f:port2 -> n0000001c:port0 [style=dashed] n00000013 [label="{{<port0> 0} | ipu1_ic_prpenc\n/dev/v4l-subdev3 | {<port1> 1}}", shape=Mrecord, style=filled, fillcolor=green] n00000013:port1 -> n00000016 [style=dashed] n00000016 [label="ipu1_ic_prpenc capture\n/dev/video5", shape=box, style=filled, fillcolor=yellow] n0000001c [label="{{<port0> 0} | ipu1_ic_prpvf\n/dev/v4l-subdev4 | {<port1> 1}}", shape=Mrecord, style=filled, fillcolor=green] n0000001c:port1 -> n0000001f [style=dashed] n0000001f [label="ipu1_ic_prpvf capture\n/dev/video6", shape=box, style=filled, fillcolor=yellow] n0000002f [label="{{<port0> 0 | <port1> 1} | ipu1_csi0_mux\n/dev/ v4l-subdev5 | {<port2> 2}}", shape=Mrecord, style=filled, fillcolor=green] n0000002f:port2 -> n00000001:port0 n00000035 [label="{{<port0> 0} | imx6-mipi-csi2\n/dev/v4l-subdev6 | {<port1> 1 | <port2> 2 | <port3> 3 | <port4> 4}}", shape=Mrecord, style=filled, fillcolor=green] n00000035:port1 -> n0000002f:port0 n00000035:port2 -> n00000041:port0 [style=dashed] n00000035:port3 -> n00000051:port0 [style=dashed] n00000035:port4 -> n0000008f:port0 [style=dashed] n0000003d [label="{{} | tc358743 0-000f\n/dev/v4l-subdev7 | {<port0> 0}}", shape=Mrecord, style=filled, fillcolor=green] n0000003d:port0 -> n00000035:port0 n00000041 [label="{{<port0> 0} | ipu1_csi1\n/dev/v4l-subdev8 | {<port1> 1 | <port2> 2}}", shape=Mrecord, style=filled, fillcolor=green] n00000041:port2 -> n00000045 [style=dashed] n00000041:port1 -> n0000000f:port0 [style=dashed] n00000041:port1 -> n0000000b:port0 [style=dashed] n00000045 [label="ipu1_csi1 capture\n/dev/video7", shape=box, style=filled, fillcolor=yellow] n00000051 [label="{{<port0> 0} | ipu2_csi0\n/dev/v4l-subdev9 | {<port1> 1 | <port2> 2}}", shape=Mrecord, style=filled, fillcolor=green] n00000051:port2 -> n00000055 [style=dashed] n00000051:port1 -> n0000005f:port0 [style=dashed] n00000051:port1 -> n0000005b:port0 [style=dashed] n00000055 [label="ipu2_csi0 capture\n/dev/video8", shape=box, style=filled, fillcolor=yellow] n0000005b [label="{{<port0> 0 | <port1> 1} | ipu2_vdic\n/dev/v4l- subdev10 | {<port2> 2}}", shape=Mrecord, style=filled, fillcolor=green] n0000005b:port2 -> n0000005f:port0 [style=dashed] n0000005f [label="{{<port0> 0} | ipu2_ic_prp\n/dev/v4l-subdev11 | {<port1> 1 | <port2> 2}}", shape=Mrecord, style=filled, fillcolor=green] n0000005f:port1 -> n00000063:port0 [style=dashed] n0000005f:port2 -> n0000006c:port0 [style=dashed] n00000063 [label="{{<port0> 0} | ipu2_ic_prpenc\n/dev/v4l-subdev12 | {<port1> 1}}", shape=Mrecord, style=filled, fillcolor=green] n00000063:port1 -> n00000066 [style=dashed] n00000066 [label="ipu2_ic_prpenc capture\n/dev/video9", shape=box, style=filled, fillcolor=yellow] n0000006c [label="{{<port0> 0} | ipu2_ic_prpvf\n/dev/v4l-subdev13 | {<port1> 1}}", shape=Mrecord, style=filled, fillcolor=green] n0000006c:port1 -> n0000006f [style=dashed] n0000006f [label="ipu2_ic_prpvf capture\n/dev/video10", shape=box, style=filled, fillcolor=yellow] n00000081 [label="{{<port0> 0} | ipu2_csi1\n/dev/v4l-subdev14 | {<port1> 1 | <port2> 2}}", shape=Mrecord, style=filled, fillcolor=green] n00000081:port2 -> n00000085 [style=dashed] n00000081:port1 -> n0000005f:port0 [style=dashed] n00000081:port1 -> n0000005b:port0 [style=dashed] n00000085 [label="ipu2_csi1 capture\n/dev/video11", shape=box, style=filled, fillcolor=yellow] n0000008f [label="{{<port0> 0 | <port1> 1} | ipu2_csi1_mux\n/dev/ v4l-subdev15 | {<port2> 2}}", shape=Mrecord, style=filled, fillcolor=green] n0000008f:port2 -> n00000081:port0 [style=dashed] } --8<-- Am Dienstag, 16. Mai 2023, 11:54:46 CEST schrieb Sakari Ailus: > Hi all, > > This set adds support for multiple downstream links in an async > sub-device, by separating the sub-device registration from the link > creation. > > A new concept, V4L2 async connection is added. A connection later on > translates to an MC ancillary or data link. Generally async notifiers have > a number of connections but at that level there is no knowledge of how > many sub-devices they will connect to. The bound and unbound callbacks now > work on connections. For the existing drivers there's only one connection > so I do not expect regressions because of that. > > Async sub-device fwnode matching will now take place between the device > (the dev field of struct v4l2_subdev) and a struct v4l2_async_connection > (an endpoint for devices that have endpoints or the device for those that > do not). This is because the graph data structure only describes > point-to-point connections so therefore defining one end of the connection > defines the entire connection. > > This set is unlikely to address all needs people have related to the async > framework but I think that beyond what it does, it paves some way for > addressing more of those additional needs. > > To be frank, I'd like to get rid of the entire V4L2 async framework, but > it would require allowing much more dynamic driver initialisation, > including sub-devices and device nodes popping up in the system in the > order and extent there is successfully probed hardware. Until that, and > this may well be the entire foreseeable future, we have at least some of > this complexity. > > since v1: > > - Fixed object relation issues. The set has now been tested on an > async sub-device with two pads connected by data links to sub-devices on > a notifier (struct) device. (Multiple notifiers should work, too, but > has not been tested.) > > - Add a function to obtain an async connection based on the sub-device. > This is useful for drivers for accessing their own link specific data. > > - Improved documentation. Include a patch documenting > v4l2_async_nf_add_fwnode(). > > - Return endpoint matching and address adv748x driver breakage in v1. It's > a special case so other drivers can remain simpler. > > - Swap notifier initialisation arguments, by making the notifier the first > argument. > > - Remove extra fwnode_handle_put() in max9286_v4l2_unregister(). > > - Make struct device available before notifier initialisation for > consistent debug messages. > > - Simplify notifier and async sub-device linked lists. Consistent list > head and list entry naming. > > - Drop leftovers from an early experimenation work in rkisp1 and omap3isp > drivers. > > - Simplify xilinx-vipp sub-device binding. > > - Use if()s in notifier_dev() of v4l2-async.c. > > - Improved debug messages in v4l2-async.c, use v4l2-async prefix and > generally with notifier device. > > - Call match types with macros V4L2_ASYNC_MATCH_TYPE_* (was > V4L2_ASYNC_MATCH_*). > > - Create ancillary links only when the sub-device is registered, not when > a connection is bound (which can take place more than once for a > sub-device). > > - Rename struct v4l2_async_match as v4l2_async_match_desc. > > - Perform list initialisation in notifier init rather than registration. > > - Get rid of the "readd" parameter for v4l2_async_nf_unbind_all_subdevs(). > > - Check async sub-device validity on a notifier only when the notifier is > registered. This removes extra list traversal and simplifies the code. > > - Remove extra list initialisation in v4l2_async_register_subdev(). > > - Drop v4l2_async_cleanup(). It was no longer useful, called from a single > place. > > - Lots of kerneldoc fixes (mostly changed argument names). > > since RFC v1: > > - Address missing API usage changes in a lot of drivers. > > - Fix compilation problems in intermediate patches. > > - Move V4L2 device registration earlier or move notifier initialisation > and fwnode endpoint parsing past the current V4L2 device registration > (patches 11--16). > > Jacopo Mondi (1): > media: v4l: async: Drop v4l2_async_nf_parse_fwnode_endpoints() > > Sakari Ailus (30): > media: Documentation: v4l: Document missing async subdev function > media: xilinx-vipp: Clean up bound async notifier callback > media: v4l: async: Add some debug prints > media: v4l: async: Clean testing for duplicated async subdevs > media: v4l: async: Drop unneeded list entry initialisation > media: v4l: async: Don't check whether asd is NULL in validity check > media: v4l: async: Make V4L2 async match information a struct > media: v4l: async: Rename V4L2_ASYNC_MATCH_ macros, add TYPE_ > media: v4l: async: Only pass match information for async subdev > validation > media: v4l: async: Clean up list heads and entries > media: v4l: async: Simplify async sub-device fwnode matching > media: v4l: async: Rename v4l2_async_subdev as v4l2_async_connection > media: v4l: async: Clean up error handling in v4l2_async_match_notify > media: v4l: async: Drop duplicate handling when adding connections > media: v4l: async: Rework internal lists > media: v4l: async: Obtain async connection based on sub-device > media: v4l: async: Differentiate connecting and creating sub-devices > media: v4l: async: Try more connections > media: v4l: async: Support fwnode endpoint list matching for subdevs > media: adv748x: Return to endpoint matching > media: pxa_camera: Fix probe error handling > media: pxa_camera: Register V4L2 device early, fix probe error > handling > media: marvell: cafe: Register V4L2 device earlier > media: am437x-vpfe: Register V4L2 device early > media: omap3isp: Initialise V4L2 async notifier later > media: xilinx-vipp: Init async notifier after registering V4L2 device > media: davinci: Init async notifier after registering V4L2 device > media: qcom: Initialise V4L2 async notifier later > media: v4l: async: Set v4l2_device in async notifier init > media: Documentation: v4l: Document sub-device notifiers > > .../driver-api/media/v4l2-subdev.rst | 28 +- > drivers/media/i2c/adv748x/adv748x-csi2.c | 13 +- > drivers/media/i2c/max9286.c | 27 +- > drivers/media/i2c/rdacm20.c | 16 +- > drivers/media/i2c/rdacm21.c | 15 +- > drivers/media/i2c/st-mipid02.c | 12 +- > drivers/media/i2c/tc358746.c | 15 +- > drivers/media/pci/intel/ipu3/ipu3-cio2-main.c | 14 +- > drivers/media/platform/atmel/atmel-isi.c | 12 +- > drivers/media/platform/atmel/atmel-isi.h | 2 +- > drivers/media/platform/cadence/cdns-csi2rx.c | 10 +- > drivers/media/platform/intel/pxa_camera.c | 77 +- > drivers/media/platform/marvell/cafe-driver.c | 18 +- > drivers/media/platform/marvell/mcam-core.c | 12 +- > drivers/media/platform/marvell/mmp-driver.c | 6 +- > .../platform/microchip/microchip-csi2dc.c | 11 +- > .../platform/microchip/microchip-isc-base.c | 4 +- > .../media/platform/microchip/microchip-isc.h | 2 +- > .../microchip/microchip-sama5d2-isc.c | 9 +- > .../microchip/microchip-sama7g5-isc.c | 9 +- > drivers/media/platform/nxp/imx-mipi-csis.c | 10 +- > drivers/media/platform/nxp/imx7-media-csi.c | 10 +- > .../platform/nxp/imx8-isi/imx8-isi-core.c | 12 +- > drivers/media/platform/qcom/camss/camss.c | 26 +- > drivers/media/platform/qcom/camss/camss.h | 2 +- > drivers/media/platform/renesas/rcar-isp.c | 12 +- > .../platform/renesas/rcar-vin/rcar-core.c | 26 +- > .../platform/renesas/rcar-vin/rcar-csi2.c | 12 +- > .../platform/renesas/rcar-vin/rcar-vin.h | 4 +- > drivers/media/platform/renesas/rcar_drif.c | 12 +- > drivers/media/platform/renesas/renesas-ceu.c | 10 +- > .../platform/renesas/rzg2l-cru/rzg2l-core.c | 14 +- > .../platform/renesas/rzg2l-cru/rzg2l-cru.h | 2 +- > .../platform/renesas/rzg2l-cru/rzg2l-csi2.c | 12 +- > .../platform/rockchip/rkisp1/rkisp1-common.h | 2 +- > .../platform/rockchip/rkisp1/rkisp1-csi.c | 7 +- > .../platform/rockchip/rkisp1/rkisp1-dev.c | 12 +- > .../platform/rockchip/rkisp1/rkisp1-isp.c | 8 +- > .../platform/samsung/exynos4-is/media-dev.c | 11 +- > .../platform/samsung/exynos4-is/media-dev.h | 2 +- > drivers/media/platform/st/stm32/stm32-dcmi.c | 12 +- > .../platform/sunxi/sun4i-csi/sun4i_csi.c | 10 +- > .../sunxi/sun6i-csi/sun6i_csi_bridge.c | 8 +- > .../sunxi/sun6i-csi/sun6i_csi_bridge.h | 2 +- > .../sunxi/sun6i-mipi-csi2/sun6i_mipi_csi2.c | 10 +- > .../sun8i_a83t_mipi_csi2.c | 10 +- > .../media/platform/ti/am437x/am437x-vpfe.c | 36 +- > .../media/platform/ti/am437x/am437x-vpfe.h | 2 +- > drivers/media/platform/ti/cal/cal.c | 10 +- > .../media/platform/ti/davinci/vpif_capture.c | 33 +- > drivers/media/platform/ti/omap3isp/isp.c | 28 +- > drivers/media/platform/ti/omap3isp/isp.h | 15 +- > drivers/media/platform/ti/omap3isp/ispccdc.c | 13 +- > drivers/media/platform/ti/omap3isp/ispccp2.c | 2 + > drivers/media/platform/ti/omap3isp/ispcsi2.c | 2 + > .../media/platform/ti/omap3isp/ispcsiphy.c | 15 +- > drivers/media/platform/video-mux.c | 10 +- > drivers/media/platform/xilinx/xilinx-vipp.c | 55 +- > drivers/media/v4l2-core/v4l2-async.c | 781 +++++++++++------- > drivers/media/v4l2-core/v4l2-fwnode.c | 109 +-- > drivers/media/v4l2-core/v4l2-subdev.c | 13 + > .../media/deprecated/atmel/atmel-isc-base.c | 4 +- > .../media/deprecated/atmel/atmel-isc.h | 2 +- > .../deprecated/atmel/atmel-sama5d2-isc.c | 9 +- > .../deprecated/atmel/atmel-sama7g5-isc.c | 4 +- > drivers/staging/media/imx/imx-media-csi.c | 10 +- > .../staging/media/imx/imx-media-dev-common.c | 6 +- > drivers/staging/media/imx/imx-media-dev.c | 2 +- > drivers/staging/media/imx/imx-media-of.c | 4 +- > drivers/staging/media/imx/imx6-mipi-csi2.c | 12 +- > drivers/staging/media/imx/imx8mq-mipi-csi2.c | 10 +- > .../media/sunxi/sun6i-isp/sun6i_isp_proc.c | 6 +- > .../media/sunxi/sun6i-isp/sun6i_isp_proc.h | 2 +- > drivers/staging/media/tegra-video/vi.c | 18 +- > drivers/staging/media/tegra-video/vi.h | 2 +- > include/media/davinci/vpif_types.h | 2 +- > include/media/v4l2-async.h | 236 +++--- > include/media/v4l2-fwnode.h | 68 +- > include/media/v4l2-subdev.h | 12 +- > 79 files changed, 1088 insertions(+), 1015 deletions(-)
Hi Alexander, On Wed, May 17, 2023 at 09:57:59AM +0200, Alexander Stein wrote: > Hi Sakari, > > thanks for the update. That's awesome, device setup on media-ctl works again > on TQMa6x (i.MX6). > For the record, below is what the media-ctl dot looks like. I was able to > capture data on ipu1_csi0_capture device. > > Tested-by: Alexander Stein <alexander.stein@ew.tq-group.com> #imx6q Thanks for testing! I'll wait for Niklas to test them, too. V1 inadvertly broke the adv748x driver, that should be now fixed. (But please avoid top posting.)
Hi Alexander, (Resending, fixed Niklas's e-mail.) On Wed, May 17, 2023 at 09:57:59AM +0200, Alexander Stein wrote: > Hi Sakari, > > thanks for the update. That's awesome, device setup on media-ctl works again > on TQMa6x (i.MX6). > For the record, below is what the media-ctl dot looks like. I was able to > capture data on ipu1_csi0_capture device. > > Tested-by: Alexander Stein <alexander.stein@ew.tq-group.com> #imx6q Thanks for testing! I'll wait for Niklas to test them, too. V1 inadvertly broke the adv748x driver, that should be now fixed. (But please avoid top posting.)
Hi Sakari, On Di, 2023-05-16 at 12:54 +0300, Sakari Ailus wrote: > Hi all, > > This set adds support for multiple downstream links in an async > sub-device, by separating the sub-device registration from the link > creation. thank you, capture works again on i.MX6QP (tested with TC358764). Tested-by: Philipp Zabel <p.zabel@pengutronix.de> # imx6qp regards Philipp
Hi Philipp On Wed, 2023-05-17 at 11:43 +0200, Philipp Zabel wrote: > Hi Sakari, > > On Di, 2023-05-16 at 12:54 +0300, Sakari Ailus wrote: > > Hi all, > > > > This set adds support for multiple downstream links in an async > > sub-device, by separating the sub-device registration from the link > > creation. > > thank you, capture works again on i.MX6QP (tested with TC358764). You probably meant TC358743 being HDMI/CSI as TC358764 is DSI/LVDS, right? > Tested-by: Philipp Zabel <p.zabel@pengutronix.de> # imx6qp Thanks! > regards > Philipp Cheers Marcel
Hi Marcel, On Mi, 2023-05-17 at 13:24 +0200, Marcel Ziswiler wrote: > Hi Philipp > > On Wed, 2023-05-17 at 11:43 +0200, Philipp Zabel wrote: > > Hi Sakari, > > > > On Di, 2023-05-16 at 12:54 +0300, Sakari Ailus wrote: > > > Hi all, > > > > > > This set adds support for multiple downstream links in an async > > > sub-device, by separating the sub-device registration from the link > > > creation. > > > > thank you, capture works again on i.MX6QP (tested with TC358764). > > You probably meant TC358743 being HDMI/CSI as TC358764 is DSI/LVDS, right? Yes, muscle memory auto-completion failure. Thanks! regards Philipp
On Wed, May 17, 2023 at 11:43:52AM +0200, Philipp Zabel wrote: > Hi Sakari, > > On Di, 2023-05-16 at 12:54 +0300, Sakari Ailus wrote: > > Hi all, > > > > This set adds support for multiple downstream links in an async > > sub-device, by separating the sub-device registration from the link > > creation. > > thank you, capture works again on i.MX6QP (tested with TC358764). > > Tested-by: Philipp Zabel <p.zabel@pengutronix.de> # imx6qp Thanks for testing this!