| ↑back Search ←Prev date Next date→ Show only urls | (Click on time to select a line by its url) |
| Who | What | When |
|---|---|---|
| tfiga | hverkuil: isn't this a bug? https://git.linuxtv.org/v4l-utils.git/tree/utils/v4l2-compliance/v4l2-test-media.cpp#n450
how a link can be dynamic and immutable at the same time? | [06:11] |
| ............. (idle for 1h3mn) | ||
| hverkuil | tfiga: thanks for reporting this. I've pushed the fix for v4l2-compliance. | [07:15] |
| tfiga | hverkuil: thanks! (mediatek reported to me :)) | [07:29] |
| hverkuil | vimc doesn't have dynamic links, so I never noticed that this code was wrong. | [07:31] |
| tfiga | hverkuil: funny enough, I don't think we'll ultimately need them
for the driver it was found with it was basically a way to skip some DMA ports for a frame, without Request API with Request API, you can just omit buffers for matching queues | [07:43] |
| hverkuil | tfiga: I am almost ready with the compliance tests for the stateful encoder.
I replied to your proposed spec with several comments, and I am incorporating the suggestions I made in the compliance tests. It helps decide if my suggestions make sense. | [07:47] |
| tfiga | hverkuil: thanks! hopefully I can get to that soon... | [07:49] |
| hverkuil | The only remaining corner case is the STREAMOFF(OUTPUT) when draining. It is proving hard to test this.
I also think we will need to move some of this to v4l2-mem2mem.c as it is probably easier to handle there rather than in each driver. | [07:50] |
| tfiga: if you can look over my comments and let me know if you disagree with the proposed changes, then that would be helpful.
I'd like to get fixes in for vicodec so it complies to the future v4 of the stateful encoder spec. Next up will be tests for the stateful decoder. | [07:57] | |
| tfiga | okay, let me try to reserve time this week | [07:59] |
| hverkuil | Much appreciated! | [08:00] |
| ........................................................................................................... (idle for 8h51mn) | ||
| *** | benjiG has left | [16:51] |
| .......................................... (idle for 3h29mn) | ||
| lexano has quit IRC (Ping timeout: 245 seconds) | [20:20] | |
| ................. (idle for 1h24mn) | ||
| bparrot | hverkuil, i am using v4l2-ctl to test one of our m2m device. In one case where we do an NV12 to NM12 conversion where source and destination is to file. The capture file was correct upto until v4l-utils commit "8e695c1d3403 Hans Verkuil v4l2-ctl-streaming.cpp: use cv4l-helpers.h".
After this point it seems that the UV part of data is the same size as the Y part. But it should be only half the size. This mean that there about 1/4 size worth of "space" in between each frame in the file. Is this by design or is this truely a bug? | [21:44] |
| ↑back Search ←Prev date Next date→ Show only urls | (Click on time to select a line by its url) |