<!-- Some styling for better description lists --><style type='text/css'>dt { font-weight: bold;float: left;display:inline;margin-right: 1em} dd { display:block; margin-left: 2em}</style>

   hverkuil: <u>mchehab</u>: I would like to know this week (preferably no later than Thursday) whether the meeting in Helsinki will be held and when. If I book travel less than two weeks before the meeting getting approval may be more difficult for me.
   <br> just FYI
   mchehab: <u>hverkuil</u>: I hope to have this defined this week too
   <br> need to know if shuah will be there or not
   <br> if we move to 29-31
   <br> Aug, 1 doesn't work for sailus
   <br> <u>shuah</u>: Aug, 1 won't work for sailus
   <br> it seems that the only period that fits for all is Jul, 29-31
   <br> would that work for you?
   shuah: I don't think I can get in until mid-day or so on Aug 29th. So I can be there for Jul 30th and 31st
   <br> LPC won't work for all I am assuming
   mchehab: sailus can't be there
   <br> I would prefer to do it at LPC
   shuah: ah ok
   mchehab: it would be easier to get it approved
   shuah: right
   mchehab: but I think we should have sailus there too
   <br> so, let's do what we can
   shuah: of course
   hverkuil: pinchartl also can't easily (and/or cheaply) change his flight to the LPC to make room for a brainstorm meeting.
   mchehab: pinchartl, hverkuil, sailus, shuah: my understanding is that Jul, 29-31 is the best option
   <br> and that works for everybody
   <br> (except that shuah will join us a little late)
   <br> so, I'll run the approval process for that schedule and cross my fingers
   shuah: right he mentioned that - I did a quick check for flights looks like the earliest I can get there is just before noon on Aug 29th - So realistically I can join July 30th and 31st
   hverkuil: I think 31 was problematic for sailus, but it's not clear if it is a big problem. But I think if sailus can't be there on the 31st we can live with that.
   <br> we just have to accept that it is very difficult to get 5 busy people in the same place for three days :-)
   mchehab: <u>hverkuil</u>: I checked with sailus last week, he said 31 would work for him as well
   <br> ok, I submitted the request
   <br> I expect I'll have a feedback in the next couple days
   sailus: <u>mchehab</u>: Ack. We have the meeting room every day that week.
   <br> I need to leave new, back late this evening.
   ezequielg: I've been testing the recently submitted tw686x driver. The driver doesn't support V4L2_FIELD_INTERLACED, just V4L2_FIELD_TOP, BOTTOM and SEQ. Since there's no support for ALTERNATE or INTERLACED, how is userspace expected to compose frames?
   <br> The driver manages to deliver frames without CPU intervention, and it would be interesting to keep it that way
   <br> here's the driver btw: https://patchwork.linuxtv.org/patch/30448/
   hverkuil: <u>ezequielg</u>: qv4l2 can handle SEQ frames using the GPU.
   ezequielg: <u>hverkuil</u>: so, if the platform has GPU in can offload the frame composition, and otherwise it must do that using the CPU?
   hverkuil: Right. Which shouldn't be a big deal for modern CPUs.
   <br> it's strange that the driver doesn't support FIELD_ALTERNATE though.
   ezequielg: <u>hverkuil</u>: but is that something that must be done on userspace, or can drivers do that theirselves.
   hverkuil: userspace
   ezequielg: ack
   hverkuil: thanks for reviewing this driver BTW.
   ezequielg: no problem. the device internal DMA engine works in terms of P-B buffers, but the specs doens't give much info, and I'm yet to figure out what that means.
   <br> ..and there are some other DMA modes
   hverkuil: I gather you have one of these cards. Where did you buy it?
   ezequielg: <u>hverkuil</u>: customer sent them
   <br> it was rather amusing, I was just about to write a driver when I found Krzysztof's
   <br> so it was really a joy to see that :)
   hverkuil: Did you compare it with this driver: https://github.com/igorizyumin/tw6869/tree/master/src
   <br> It claims to do FIELD_INTERLACED.
   ezequielg: yeah
   <br> that driver is the one I was using
   <br> it has a few crashes, not so hard to reproduce
   <br> and it uses old videobuf
   <br> and it's a mess in general
   hverkuil: that driver is in bad shape, but it might give clues to the DMA handling and interlaced support.
   ezequielg: regarding the INTERLACE, I believe it's copying the fields into a buffer