[11:20] <hverkuil> Do we need a meeting this week and next week? Nothing much is happening during this merge window. I'm happy to cancel these two meetings.
[11:43] <mchehab> hverkuil: good question. I have just one thing for today's meeting
[11:44] <mchehab> Subject: [PATCH v2 00/17] follow_pfn and other iomap races
[11:44] <mchehab> someone has to test it, and probably port it also to VB1, due to the two DMA contig drivers that still use VB1
[11:46] <mchehab> (davinci and viu)
[11:46] <mchehab> or, as an alternative, port both to VB2
[11:48] <hverkuil> I can't test either. Lad Prabhakar might be able to test davinci.
[11:53] <hverkuil> I'm inclined to mark fsl-viu and davinci/vpfe_capture as deprecated, move them to staging and kill them if nobody takes any action after 3-4 kernel cycles.
[11:54] <mchehab> So, I guess we should talk with Lad and see if he could port the pin_user_pages_fast() code from patch 05/17 to VB1 and test on davinci
[11:54] <hverkuil> Or ideally, port that driver to vb2. All other davinci drivers are all vb2, except that one.
[11:55] <mchehab> yeah, porting them to VB2 would solve
[11:56] <mchehab> after having both drivers addressed, we can get rid of DMA contig code at VB1
[11:56] <mchehab> with would be a good thing
[11:56] <mchehab> unfortunately, currently I can't help with those... most of my cards are too far away from me
[11:58] <mchehab> (well, I don't have davinci/viu boards. I could perhaps with tm6000 and bttv if I had the needed hardware here with me, instead of having them in Brazil)
[12:01] <mchehab> anyway, that's the only subject I had for today
[12:01] <mchehab> I guess we can skip having a meeting next week
[12:01] <hverkuil> you can buy bttv cards very cheaply on ebay. That might be the easiest way to get hold of it.
[12:01] <mchehab> yes, but I would also need a computer with PCI slots
[12:02] <hverkuil> I use PCIe-to-PCI riser cards. Quite cheap and they do the job.
[12:02] <mchehab> I don't have PCIe here either
[12:02] <mchehab> I'm using just NUCs
[12:02] <hverkuil> Ah, then you have a problem :-)
[12:02] <mchehab> yep
[12:03] <hverkuil> I'd be happy to test for you, if needed.
[12:04] <mchehab> ok, thanks! I'll try to find some time for trying to port bttv to VB2 during 5.11 cycle
[12:06] <mchehab> (I have a few other things on my todo list for 5.11, but I guess I can find some time for bttv)
[12:07] <hverkuil> I have bttv cards here in Norway, but it looks like my tm6000 might be in the Netherlands.
[12:08] <mchehab> I was thinking that I had at least some of my tm6000 cards here, but it seems I didn't take any
[12:10] <mchehab> I'll double-check it later
[12:43] <pinchartl> a bit delayed, but I also think that porting davinci to VB2 is better than porting the fixes from VB2 to VB1
[12:44] <pinchartl> I expect that to actually be easier
[12:44] <pinchartl> as fixing VB1 is hindered by VB1 being a mess, and the fix needing good knowledge of memory management
[12:44] <pinchartl> while porting davinci to VB2 is a more mechenical conversion
[12:59] *** hverkuil has quit IRC (Ping timeout: 272 seconds)
[13:09] *** ChanServ sets mode: +v hverkuil
[15:32] <mchehab> pinchartl: in ths specific case, I suspect that adding VB1 support shouldn't be hard, as they probably don't need to
[15:32] <mchehab> ... support the hack added by Ricardo at VB2
[15:33] <mchehab> but yeah, porting to VB2 is better
[15:35] <mchehab> btw, maybe VB1 could just call frame_vector API instead of having to port all the changes into it
[19:59] *** [LOGGER] has quit IRC (Ping timeout: 272 seconds)