drivers: video: CONFIG_VIDEO_SW_GENERATOR -> zephyr,video-sw-generator#88780
Conversation
f62e2d4 to
8bd2a1d
Compare
loicpoulain
left a comment
There was a problem hiding this comment.
LGTM, but not tested.
| child-binding: | ||
| child-binding: | ||
| include: video-interfaces.yaml |
There was a problem hiding this comment.
No need to add this because you don't have child node and use nothing from video-interfaces binding.
There was a problem hiding this comment.
Good point, no "sub-devices"... It is the "head" of the pipeline on its own.
There was a problem hiding this comment.
It seems you forgot to remove these ?
There was a problem hiding this comment.
Right thank you! I thought I needed to use it for the video_sw_pipeline but it ended-up not the case, I simply did not know how to best handle "m2m" devices. removed now!
|
It's fine to me to turn video sw generator into devicetree so that #76798 can get merged because it's been there for a long time. For a long-term solution, as discussed in #88182 and in #76798, I think
But we can do that later as a future extension (as currently, we are all based on zephyr,camera) to not postpone #76798. |
ec0c215 to
cb21eb8
Compare
|
force-push:
Thank you all for your patience! |
8ae099c to
e5127bc
Compare
|
Thanks for the review @avolmat-st! Though I just applied @JarmouniA diff saying |
e5127bc to
5dcb0e3
Compare
f57f701 to
0aca1b4
Compare
By having an "auto-fallback" mechanism, the VIDEO_SW_GENERATOR did shadow the fact that "platform:mimxrt1064_evk:SHIELD=dvp_fpc24_mt9m114" was missing the full specification, and therefore was not matched at all: the shield was not selected, but the CI still worked. Signed-off-by: Josuah Demangeon <me@josuah.net>
Make the video software generator a devicetree node, which allows enabling
several instances, and select it as chosen { zephyr,camera = &... }; node.
It can be enabled via a `video-sw-generator` snippet.
Signed-off-by: Josuah Demangeon <me@josuah.net>
Convert the log module declarationto the new syntax introduced in 5e34681 proposing a more compact syntax. This fixes the log level not being updated in the "capture" sample. Signed-off-by: Josuah Demangeon <me@josuah.net>
Since this sample's headers are now grouped it is possible to sort them alphabetically. Signed-off-by: Josuah Demangeon <me@josuah.net>
Variables that holds pointers to 'const struct device' typically end-up with '_dev'. The tcpserversink video sample did not have it. Add it. Signed-off-by: Josuah Demangeon <me@josuah.net>
The zephyr,camera was used in several samples, but was not documented. Add the missing documentation entry for it in the list. Signed-off-by: Josuah Demangeon <me@josuah.net>
0aca1b4 to
d845bd2
Compare
|
Force-push:
One more review round to go... Thanks all for your patience! |
|




Downstream:
Related:
As pointed out in the UVC PR, the typical way to configure hardware-like devices in Zephyr is via a chosen node. The other proposed approach for default video devices is #88182.
Because this is a purely virtual hardware, a snippet is provided to enable it in any example via
--snippet video-sw-generator. Document its use in the 3 capture samplesREADME.rst. Thezephyr,cameradoc was missing and is also added.If this is indeed the way to go, I will convert these other PRs to use the devicetree instead of Kconfig: