| ↑back Search ←Prev date Next date→ Show only urls | (Click on time to select a line by its url) |
| Who | What | When |
|---|---|---|
| *** | ChanServ sets mode: +v hverkuil | [08:40] |
| ................................... (idle for 2h53mn) | ||
| syoung | commit 0b41a64fe4f03460b16e2d3d973017cf9cc2a056 only changes the code path on rc_unregister_device() so that does not explain why IR is not working or those uevent errors
I must admit I am a bit lost on this. anyway I think 0b41a64fe4f03460b16e2d3d973017cf9cc2a056 should be merged | [11:33] |
| ......................... (idle for 2h3mn) | ||
| hverkuil | sailus: patch "ad608fb media: v4l: event: Prevent freeing event subscriptions while accessed" broke event handling (vivid and others now fail). Please take a look asap.
This is being backported as well to older kernels, so that's not good at all. *Always* test such core changes with vivid and v4l2-compliance in the future. We really need to get the regressions tests up and running soon, that would have prevented this. | [13:37] |
| ................. (idle for 1h22mn) | ||
| sailus | hverkuil: I agree.
The sooner we have automated testing, the better. It'd be pretty much impossible to miss such things. Albeit this could have been catched in review, too. Thorough testing is always better though. | [15:01] |
| hverkuil | I missed it as well when I reviewed. I should have asked if it was tested with v4l2-compliance :-( We should make this a requirement for core-code changes as well. | [15:07] |
| sailus | I do agree we need more testing, but putting all the burden on developers is not the right way to do it.
What I'm saying is that this need to be made automatic. | [15:10] |
| pinchartl | sailus: we should give developers means to catch issues by offloading tests
so the responsibility of testing would still be on developers (with the added bonus that we would get more visibility on it), but with less burden | [15:11] |
| sailus | Offloading tests... from where? | [15:12] |
| pinchartl | from developer's workstations to a centralized server | [15:12] |
| sailus | https://pastebin.com/P7kyqt0Q
So I won't be testing with vivid today. :-( Oh, wait... Perhaps it's my media device refcounting patches that need more work after all... :-P | [15:14] |
| pinchartl | :-)
pinchartl is gone, I'll be back later | [15:17] |
| hverkuil | ANN: I've updated v4l-utils with the latest media_tree master. Only the patch adding tests for the Request API is missing, that can be expected later this week. I need to think about that a little bit more. | [15:26] |
| sailus | https://pastebin.com/TF2kjv2r
I don't believe it's my event patch. :-) | [15:29] |
| hverkuil: Can you probe vivid successfully? | [15:35] | |
| hverkuil | On my PC, yes. | [15:35] |
| sailus | See the paste... that's on ARM though. But I believe there's something wrong, possibly with vivid. | [15:37] |
| hverkuil | I know it runs fine on a tegra, never seen any problems with vivid. It seems to die in drivers/base/dd.c, __device_attach_driver. | [15:43] |
| sailus | For some reason other drivers probe nicely.
Sent. I'll be offline for some time, back later for a while. | [15:44] |
| *** | ChanServ sets mode: +v mchehab` | [15:51] |
| hverkuil | Traveling to Oslo this evening. I'll be back tomorrow. | [15:54] |
| .... (idle for 16mn) | ||
| *** | ChanServ sets mode: +v snawrocki | [16:10] |
| snawrocki | hverkuil: just tested your crop2sel patch series on one of Exynos reference boards, FWIW I could do it only now after returning from a 4 months leave | [16:16] |
| ↑back Search ←Prev date Next date→ Show only urls | (Click on time to select a line by its url) |