Small question. I can get only half of my camera view. How can I get a full view? syoung: Can you look at these two v4l-utils warnings from the daily build? bpf_load.c:205:41: warning: pointer of type 'void *' used in arithmetic [-Wpointer-arith] bpf_load.c:322:54: warning: pointer of type 'void *' used in arithmetic [-Wpointer-arith] hverkuil, mchehab: Would it be possible to get https://patchwork.linuxtv.org/patch/50282/ collected sometime please? kbingham: for 4.19? hverkuil, Doesn't matter when :) hverkuil, in fact actaully I have a few trivial little patches which could be integrated sometime. Perhaps I should just batch them into a PR or somethign ? PR? <pull-request> ah hverkuil, Hrm ... the corresponding change for that binding update was picked up by Simon Horms ... So if that's landing in v4.19 then it might be good to make sure it matches - I'll ask him what it's target it. *is Anyway, this is something that will be picked up when processing patches. I probably do a run some time next week. If this is needed for 4.19, then let me know. Otherwise it will end up in 4.20. hverkuil, Ok sure - I was just looking at it to make sure it doesn't get missed or blocked due to me :) mchehab: ping pinchartl: pong kbingham, hi ! i have now 2 subdevices exposed (/dev/v4l-subdev0 and 1), how could i test them ? In particular should trigger the s_stream, can i do that by v4l2-ctl ? angelo_ts, Sounds promising - but do you have a corresponding /dev/videoX node too ? Those subdevs are probably objects to represent the sensor devices. but you'll still need a 'real' video node to be able to communicate with the pipeline. That should be provided by the ISP/IPU whatever component. kbingham, mm, i have a video0 and video1, but they seems to be mxc-jpeg decoder / encoder devices mchehab: have you noticed https://osseu18.sched.com/event/FwGc/complex-cameras-on-linux-mauro-carvalho-chehab-samsung and https://osseu18.sched.com/event/FwHe/why-embedded-cameras-are-difficult-and-how-to-make-them-easy-laurent-pinchart-ideas-on-board ? we both seem to talk about the same topic :-S didin't look at ELCE sched yet mchehab: I assume you'll talk about libcamera, right ? yes so will I... heh seems the program committee really screwed up angelo_ts, Next step then is to work out why there isn't a video node yet. It may have failed some other stage of the probe sequence, - Or, you mentioned that there are two CSI2 bus nodes ... but you are only working with one of them. Make sure that the driver is not waiting to register the device node because it's expecting more devices/subdevices to come on. mchehab: I think I'll contact the LF to ask why they have scheduled two talks about the same topic I don't see much value in presenting the same topic twice kbingham, mm ok. So the video0 and video1 i see (jpeg encoder / decoder) are not the ISP created video device ? angelo_ts, I don't know ? angelo_ts, They don't sound like it ... pinchartl: since my proposal was rejected, I'm happy to take over one of the slots :-) hverkuil: I've already had two similar requests :-) angelo_ts, I would expect a jpeg-encoder/decoder is separate from a camera device... kbingham, mm ok I have already material for a presentation with the current status I gave a presentation on an internal event on July, in KR can i check the type by v4l2-ctl ^? angelo_ts, v4l2-ctl --list-devices kbingham, many thanks. I found some missing devicetree parts kbingham, now i have https://pastebin.com/zKrcApVb that seems much better angelo_ts, Yes, that does indeed look like it might be better :) do you have any media controller devices ? (ls /dev/media*) oh yes, there is a media0 also now i see 0 to 5 subdevs Ok - so how about trying to capture something from /dev/video0? ok, what tool should i use ? angelo_ts, Depends on what you have available in your filesystem i have v4l2 utils now only I usually use yavta and guvcview as capture tools I believe v4l2-ctl can capture framees and write them out too. mm looks like the video0 is not properly linked to my subdevs has anybody ever used the adv7611 and similar in a 8-bit bus configuration? I'm wondering whether the MBUS_FMT is correct for those modes the datasheet is a bit unclear on this, but I think it should be MEDIA_BUS_FMT_VYUY8_2X8 instead of MEDIA_BUS_FMT_YVYU8_2X8 hverkuil: I'll fix those warnings in v4l-utils.