| ↑back Search ←Prev date Next date→ Show only urls | (Click on time to select a line by its url) |
| Who | What | When |
|---|---|---|
| jhautbois | Hi there
I have an i.MX6 board here (a Digi ConnectCore 6) and I want to connect a FPD-LINK III module on it. It uses the DS90UB954 device from TI. AFAIK there is no support for it yet, but I would like to add it. As it is a deserializer with MIPI CSI-2 outputs, I think it is a good candidate to be a v4l2 device ? Connected to this deserializer I have its little brother :) a DS90UB953 and two OmniVision cameras but it could be pretty much any MIPI-CSI-2 camera | [07:56] |
| hverkuil | jhautbois: sounds like a v4l driver, yes. | [08:00] |
| jhautbois | hverkuil, according to you, which driver is a good candidate to have it as an example for this one ? | [08:05] |
| hverkuil | jhautbois: we don't have any CSI2 deserialization drivers yet to my knowledge. sailus: please correct me if I am wrong. | [08:09] |
| jhautbois: actually, I think this a deserializer: drivers/media/platform/cadence/cdns-csi2tx.c
this a -> this is a | [08:15] | |
| ................ (idle for 1h19mn) | ||
| kbingham | hverkuil, hautbois, The FPD-Link III is a similar system to GMSL from Maxim. We posted a MAX9286/RDACM20-MAX9271 implementation just before Japan OSS which might be useful reference.
oops - that should have been jhautbois :) ^ | [09:35] |
| hverkuil | kbingham: ah, that's not been merged yet AFAICS, so that's why I didn't find it. | [09:36] |
| jhautbois | kbingham, thx for the reference ! | [09:42] |
| kbingham | hverkuil, Correct - it's not there yet - but I think we're hoping it can go in soon. :D
jhautbois, No worries. Let me know if you want any help on the FPD-Link III ... I'm looking for an excuse to spread into the TI version :-) | [09:43] |
| ........ (idle for 39mn) | ||
| pinchartl | hverkuil: would you like to review those drivers ? :-) | [10:22] |
| hverkuil | pinchartl: no :-) | [10:28] |
| sailus | pinchartl: Which driver is it?
hverkuil: cdns-csi2tx.c is a driver for a CSI-2 TX IP block but I'm not sure it's part of an actual SoC at the moment. | [10:31] |
| pinchartl | sailus: max9286 and rdacm20 | [10:34] |
| kbingham | sailus, https://patchwork.linuxtv.org/patch/51483/ and https://patchwork.linuxtv.org/patch/51484/ | [10:38] |
| sailus | They're both assigned to me. Oh dear. | [10:39] |
| kbingham | (We have desires to further split the RDACM20 into it's components sometime too ... as we also have an RDACM21 which is the same max9271, but a different sensor) | [10:39] |
| sailus | Feel free to ping me if you think you're not receiving comments in a timely manner. | [10:40] |
| pinchartl | sailus: as Kieran said, please don't concentrate on the fact that the rdacm20 driver isn't split yet
it's longer term work, and planned, but not easy and in my opinion not a blocker | [10:40] |
| ........ (idle for 38mn) | ||
| mort | I have a camera and an encoder. I open the camera capture with mmap and export its buffers, then I open the encoder output with dmabuf, and I open the encoder capture with mmap. The first time I queue a buffer from the camera to the encoder output, I get an event on the encoder capture, and am able to dequeue that, and it looks perfect - but I neve
r get any more POLLIN or POLLRDNORM events on the encoder file descriptor. I do enqueue the buffer again after dequeuing it, and I get POLLIN events on the camera and then queue an available buffer to the encoder output, which the output happily accepts, and I get POLLOUT events on the encoder fd and successfully dequeue a buffer from it, everything seems to work perfectly, except that I never get any more POLLIN events f rom the encoder | [11:19] |
| anyone got any clue what I might be doing wrong? | [11:33] | |
| ............ (idle for 56mn) | ||
| pinchartl: do you have a second to help me with this a bit? I'm kind of stuck
I can't see how I'm doing anything differently from how I'm using a decoder device, other than using dma buffers instead of memory mapped buffers, and the decoder works | [12:29] | |
| svarbanov | mort: check does timestamps on encoder input are growing, i.e. the timestamps from capture device | [12:40] |
| mort | svarbanov: the camera's timestamp increases, but I'm not sending it in to the encoder's output | [12:46] |
| svarbanov | mort: I think you have to | [12:47] |
| mort | I didn't know that was necessary, and have never done it before, but I see in the documentation now that it's sometimes necessary
I mean, I have never done it before even when decoding or encoding without isues and indeed, adding a timestamp makes it work. Thanks :D | [12:47] |
| svarbanov | mort: you are welcome | [12:51] |
| jhautbois | kbingham, ok, I will ping you then ;) thx | [12:56] |
| ............................ (idle for 2h19mn) | ||
| *** | benjiG has left | [15:15] |
| ↑back Search ←Prev date Next date→ Show only urls | (Click on time to select a line by its url) |