[00:55] *** awalls has left 
[08:36] <hverkuil> pinchartl: "implementing input support": I'm not sure what you mean with that.
[09:11] <hverkuil> mchehab: pong
[09:12] <mchehab> while looking at this patch: https://patchwork.linuxtv.org/patch/31435/
[09:13] <mchehab> I'm wandering why we're using the dvb old API inside ivtv
[09:13] <hverkuil> it predates VIDIOC_DQEVENT.
[09:14] <mchehab> I would very much prefer to just get rid of video.h and audio.h than to fix it for 2038-yr
[09:14] <hverkuil> I have considered removing it, but since it wasn't blocking anything I never bothered.
[09:15] <hverkuil> So you want to drop video.h and audio.h completely? Isn't it used by another driver as well?
[09:15] <mchehab> does it implement VIDIOC_DQEVENT? Are any application using the legacy DVB API to talk with ivtv?
[09:16] <hverkuil> Yes, it has the modern API as well.
[09:18] <hverkuil> As to applications using the old API: ivtv-ctl in v4l-utils still uses it (I never bothered updating it), but I suspect I can just drop that code from that utility since v4l2-ctl can do the same.
[09:19] <mchehab> does MythTV (or some other widely-used app) use it?
[09:19] <hverkuil> Whether there are other apps out there that use it, I have no idea.
[09:19] <mchehab> (and don't have a fall-back to VIDIOC_DQEVENT)
[09:19] <mchehab> Hmm.. let's trace a plan to remove it then...
[09:20] <mchehab> what about:
[09:20] <mchehab> 1) either drop ivtv-ctl or convert it to use VIDIOC_DQEVENT;
[09:20] <hverkuil> It would not surprise me if it is still in use: it was a very popular card in the past.
[09:20] <mchehab> 2) add a printk_once() msg telling that the API is obsolete;
[09:20] <hverkuil> 1: no problem
[09:20] <mchehab> 3) wait for a few Kernel cycles;
[09:21] <mchehab> 4) remove its support from ivtv
[09:21] <hverkuil> 2: ditto
[09:21] <mchehab> 5) drop support for it from Kernel, together with av7110 driver
[09:22] <hverkuil> do you have the mythtv source? If so, can you grep for VIDEO_GET_?
[09:22] <hverkuil> MythTV is the most likely app that support this.
[09:25] <mchehab`> no, I don't have mythtv source here
[09:25] <hverkuil> I just downloaded the latest.
[09:26] <mchehab`> https://github.com/MythTV
[09:26] <hverkuil> Looks like it isn't used anymore (if ever).
[09:26] <mchehab> getting it from git clone https://github.com/MythTV/mythtv
[09:28] <hverkuil> I'm OK with your plan.
[09:28] <mchehab> $ git grep dvb/video\\.h
[09:28] <mchehab> doesn't return anything
[09:29] <mchehab> neither:
[09:29] <mchehab> $ git grep VIDEO_GET_EVENT
[09:30] <mchehab> so, except if it hardcodes the IOCTL number, it should be safe to drop it
[09:39] <hverkuil> I'm removing the dvb ioctls from ivtv-ctl.
[09:39] <mchehab> OK
[09:39] <mchehab> thanks!
[09:44] <hverkuil> Done.
[09:46] <mchehab> thanks!
[09:48] <hverkuil> Can you make a patch for ivtv adding the printk? I'll review and ack it.
[09:48] <mchehab> sure
[09:48] <hverkuil> ivtv_default() in ivtv_ioctl.c
[09:49] <mchehab> only ivtv-alsa.h seems to be using pr_foo()
[09:50] <mchehab> I guess I'll convert it to pr_foo() first
[09:50] <mchehab> and use pr_err_once()
[09:51] <mchehab> ivtv-driver.h is included everywhere?
[09:53] <hverkuil> as far as I remember, yes.
[09:55] <mchehab> ok. I'll add pr_fmt() there and be sure that it is the first file included
[09:55] <mchehab> and use pr_foo() instead of printk()
[09:55] <mchehab> then, on a next patch, I'll add the pr_err_once()
[11:01] <mchehab> hverkuil: just sent the patches
[11:01] <mchehab> I opted to also add a kconfig option to disable the DVB API abuse
[15:31] *** prabhakarlad has left 
[15:49] <pinchartl> hverkuil: most drivers have a single input, and still have to implement useless code for g/s_input and enuminput
[15:50] <pinchartl> if we want to make those ioctls mandatory shouldn't we implement them in the core instead of duplicating useless code in drivers ?
[16:22] *** indy has quit IRC (Ping timeout: 244 seconds)
[17:02] *** ndufresne has left 
[17:18] *** benjiG has left 
[18:46] <joebobjoe> hverkuil: are you on?
[18:47] <joebobjoe> and can I ask you a question about proper HDMI quantization levels for RGB?
[18:47] <joebobjoe> I was told you might know
[19:08] <pinchartl> is there anyone with access to the Synopsys DWC HDMI TX documentation here ?
[19:46] <pinchartl> (scratch that Synopsys doc request, I've found the information I need)
[22:35] *** joebobjoe has left