[01:42] *** syoung has quit IRC (Ping timeout: 240 seconds) [10:31] <sailus> I'll be late from today's meeting. [10:31] <sailus> Expect ½ an hour or so. [10:38] <hverkuil> let's start half an our later than usual. That works better for me as well. [10:38] <hverkuil> our -> hour :-) [11:09] <sailus> Heippa! [11:09] <sailus> I'm actually free now. [11:10] <sailus> Or... am I actually 50 minutes early?? [11:12] <hverkuil> yes, you're early :-) [12:04] <sailus> Am I still early? [12:06] <hverkuil> no, you are on time, but my meeting will likely run for another 25 minutes. Possibly longer. [12:06] <sailus> Ack. [12:30] <hverkuil> meeting over, let me get some coffee and we can start. [12:32] <syoung> Hello [12:32] <pinchartl> hello [12:33] <hverkuil> OK, I'm ready. [12:34] <hverkuil> Not that I have much to tell. [12:35] <hverkuil> I would like some feedback for https://patchwork.linuxtv.org/patch/55458/ [12:36] <hverkuil> I've also reviewed Boris Brezillon's series for the new format/streaming ioctls. [12:36] <hverkuil> We need to think about how to represent buffers (single planar, multi planar, offsets). [12:36] <pinchartl> hverkuil: the change seems ok, although my preferred fix would be removal of the userptr support :-) [12:36] <sailus> hverkuil: I'll take a look at that. [12:37] <sailus> V4L2 has too many ways for supporting formats with multiple planes. [12:37] <hverkuil> I'll see if I can make a proposal for the formats and people can than shoot at it. [12:38] <pinchartl> sailus: let's create a new one that will work for everybody [12:38] <pinchartl> (https://xkcd.com/927/) [12:38] <hverkuil> pinchartl: that's how we keep in work! [12:38] <sailus> pinchartl: Do we need yet one more? [12:39] <hverkuil> Anything else I need to look at? [12:40] <sailus> Well, I'll send soon a pull request on my V4L2 corner case fixes. [12:40] <sailus> No-one commented for a while, I guess that's mostly settled, there were just some details mchehab wanted to have a say on... [12:41] <hverkuil> sailus: please run the test-media script with a kernel with those patches. [12:41] <hverkuil> Just to make sure nothing regressed. [12:42] <hverkuil> That should be a standard check for any patches touching on the v4l2 core. [12:42] <sailus> hverkuil: Yeah, I can do that. [12:43] <hverkuil> There should be no errors and 12 warnings (hopefully a colleague will post a patch series fixing those remaining vivid warnings soon) [12:44] <pinchartl> sailus: would you also be able to add tests that get trigerred by the bugs you're fixing ? [12:46] <hverkuil> Note that I posted a patch deleting the zoran driver from staging. Are there other deprecated drivers that should be deleted for 5.2? [12:47] <pinchartl> a bit sad to see that driver go, but I think it makes sense [12:48] <sailus> pinchartl: It requires driver "support" as well. It would not work on all drivers, but with those that do not limit the buffer size, it's possible. [12:49] <sailus> But would you want such a test? Most likely drivers should still carry some kind of sanity checks. [12:49] <pinchartl> sailus: it's a case by case decision, but bug fixes are supposed to fix bugs, so a way to trigger those bugs is useful to check for regressions later [12:50] <sailus> Yeah, true. [12:52] <hverkuil> Anything else? [12:52] <sailus> I guess we're done then? [12:53] <hverkuil> I'd say so. [12:54] <hverkuil> I'm not joining the meeting next week (public holiday), but I'll read any comment made here later that day. [13:12] <sailus> Ack. [13:26] *** ChanServ sets mode: +v mchehab [13:43] *** ChanServ sets mode: +v mchehab [13:52] *** ChanServ sets mode: +v hverkuil