<!-- 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>

   benjiG: <u>hverkuil</u>: hello
   ***: hverkuil has left
   hverkuil: <u>benjiG</u>: pong
   benjiG: <u>hverkuil</u>: we (ST) have few patches (video decoder, hva debug) waiting for your remarks/ack, do we have to do something else ? or just wait ?
   hverkuil: Just wait, I'm afraid.
   <br> Busy at work/Christmas/moving to a new apartment all took a big bite out of my V4L2 time :-(
   benjiG: <u>hverkuil</u>: I understand :-) I was just afraid to have miss a blocking point
   hverkuil: I plan to go through some of those pending patch series today (no promises).
   benjiG: <u>hverkuil</u>: ok ping me if need
   ***: andrey_utkin has quit IRC (Read error: Connection reset by peer)
   sailus: <u>mchehab</u>: Bom dia!
   hverkuil: <u>sailus</u>: he's on vacation until the 21st.
   sailus: <u>hverkuil</u>: Oh.
   headless: note that when I used 1 desciriptor per skb on gen2, it hanged as well
   <br> oops
   <br> wrong chnanel
   <br> channel even
   pinchartl: <u>hverkuil</u>: that's -rc5 time :-/
   <br> <u>ttomov</u>: ping
   ttomov: <u>pinchartl</u>: pong
   pinchartl: how are you doing ? as we say here, onnellista uutta vuotta :-)
   ttomov: I'm starting today for 2017
   <br> and for the Finnish I'll have to check the dictionary :)
   pinchartl: :-D
   ttomov: Happy new year :)
   pinchartl: uusi means new and vuosi means year, I think you can guess the rest :-)
   <br> thank you
   <br> I was wondering whether you had worked on the msm8996
   <br> and whether you think it would be feasible to adapt the 8x16 driver to support msm8996
   <br> obviously the hardware is different
   <br> but I have msm8994 kernel code developed for Ara that I could use, which is hopefully close to the msm8996
   <br> (at least when it comes to the CSI receivers and RDI)
   ttomov: I haven't worked on 8996 so far
   pinchartl: should I ask Gjorgji ? he might be more familiar with the device
   ttomov: the camera support will reach DB820c some day
   <br> which means 8996
   <br> but it is not know when :)
   <br> but it is not known when :)
   pinchartl: that's my problem, I need it soon :-)
   <br> I can possibly work on that myself
   <br> but I don't have access to any documentation
   <br> other than the msm8x16 driver you posted, the msm8994 driver developed for Ara, and the msm8996 BSP code (which probably delegates register configuration to userspace)
   ttomov: I have checked 8996 code only briefly, for csiphy the register maps are different than 8916
   pinchartl: how about msm8994 vs. msm8996 ?
   ttomov: 96 supports c-phy in addition to d-phy. 16 and 94 only support d-phy
   <br> if you are able to compare the code for 94 and 96 and are able to make sense from the differences... it might work
   <br> although debugging any problem without documentation will be... challenging, or more :(
   pinchartl: I only need D-PHY, but I suppose I'll still need to set some registers
   <br> yes, that's what I was afraid of :-S
   ttomov: and what is your platform, if not secret?
   pinchartl: I'm trying to figure out :-)
   <br> my customer mentioned two different platforms
   <br> snapdragon 805 and snapdragon 820
   <br> (MSM8996 and APQ8084)
   <br> (the other way around)
   ttomov: ok
   pinchartl: http://www.inforcecomputing.com/products/single-board-computers-sbc/qualcomm-snapdragon-805-inforce-6540-sbc or http://www.inforcecomputing.com/products/system-on-modules-som/qualcomm-snapdragon-820-inforce-6601-micro-som
   <br> would the APQ8084 be easier ?
   ttomov: 84 is older than 94...
   <br> it's difficult to estimate which will be easier
   pinchartl: :-)
   <br> difficult and Qualcomm are often seen in the same sentence
   ttomov: but I would say that for 96 plans are that something will be available upstream... sometime :) so this is a big plus
   pinchartl: :-)
   <br> should I contact Nicolas about the "sometime" ?
   ttomov: you can always ask him, but I don't think that there is a very clear schedule for this at the moment
   pinchartl: thank you for the information
   hverkuil: <u>pinchartl</u>: any chance that you might be able to review koike's vimc driver this year? :-)
   pinchartl: <u>hverkuil</u>: I've booked time for reviews tomorrow morning
   hverkuil: great!
   ndufresne: <u>pinchartl</u>: 32 planes is currently a hard dep in most drivers, and user space, did something land with more then that ?
   hverkuil: <u>ndufresne</u>: no. There was a discussion about that 2 years ago or something like that, but nobody did anything about it. None of the current set of drivers need more than 32 buffers.
   <br> I'm sure this will become an issue again.
   pinchartl: <u>ndufresne</u>: I assume you mean 32 buffers, not 32 planes
   <br> patches are welcome :-)
   ndufresne: yeah, 32 buffers that is
   pinchartl: if we want to extend V4L2 beyond 32 buffers then we should do it properly
   <br> and implement a DELETE_BUFS ioctl
   <br> similar to CREATE_BUFS, but the other way around
   hverkuil: and that's exactly the reason this hasn't be done yet :-)
   ndufresne: definatly, though on my side, we don't shrink buffer pools, the maximum allocation may be unknown, but it is always capped
   <br> let's say we'd need a use case, and a driver
   pinchartl: even if you don't need to delete buffers for your use case, I believe increasing the number of buffers should be done by making the API fully dynamic, with create/delete support
   hverkuil: I agree.
   kbingham: <u>pinchartl</u>: hverkuil: if mchehab is on holiday - does that mean git://linuxtv.org/media_tree.git won't update anytime soon ?
   <br> media_tree.git/master is currently at v4.10-rc1, and has build failures for ARM :(
   <br> Works if I merge in v4.10-rc2 ...
   <br> but then I'm not on an upstream base ...
   <br> (meaning, I'm considering media_tree.git as the 'upstream' here ...)
   pinchartl: <u>kbingham</u>: I'm afraid so (unless Mauro works during his holidays)
   ***: awalls1 has left
   pinchartl: this calls for a V4L2 co-maintainer, doesn't it ? :-)
   kbingham: it certainly would be handy to have the linux-media/master fixed ...
   <br> <u>hverkuil</u>: Do you have commit rights?
   <br> <u>hverkuil</u>: Or can we get you commit rights ? :D
   hverkuil: <u>kbingham</u>: send an email to Mauro marked 'URGENT'. If all he has to do is to merged in rc2, then he might be willing to do that, even if he is on vacation.
   <br> I don't know how often he looks at such mails.
   ***: awalls has left