<!-- 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>
tfiga: <u>stark3y</u>: ndufresne: would the fence timestamp be helpful for that problem?
<br> AFAIR one can query the fence for signal timestamp
gnurou: <u>hverkuil_</u>: not much, if any. Don't think my SoB is necessary for this one.
<br> feel free to add if you think it should be though
hverkuil_: <u>gnurou</u>: tfiga: should I remove the Google copyright from patch 4 as well?
gnurou: <u>hverkuil_</u>: I don't see a Google copyright in patch 4
<br> looking at v14
hverkuil_: Sorry, patch 2. (It's patch 4 in the upcoming v15 :-) )
tfiga: I think anything that says "Chromium OS authors" should be removed
hverkuil_: There is nothing like that at all. The only thing is a "Copyright (C) 2018 Google, Inc." in patch 2 (v14)
tfiga: Okay, that sounds like correct copyright for our code, but I'll leave it to Alex to judge how much of our code is there :)
gnurou: if you decide to keep the Google copyright then there probably should also be one for Ideas on Board
<br> I don't feel like either is particularly necessary though
hverkuil_: I don't mind either way, but if there is a Google copyright, then I also need your SoB line.
<br> <u>pinchartl</u>: do you want to add your copyright + SoB to patch 2 of the request API? (v14)
pinchartl: <u>hverkuil_</u>: good point. let me check it
<br> I'll try to read and reply to your e-mails today
hverkuil_: <u>svarbanov</u>: what is the status of the venus patch series?
tfiga: <u>hverkuil_</u>: I guess it's blocked on my review being slow...
<br> I reviewed 12 patches out of v2 series until yesterday
<br> if you think it makes sense for svarbanov to respin it, I'm okay with it
hverkuil_: Today is the last chance for this to be added to 4.18. So if he can make a v3 with only minor changes from v2, then I can make a pull request. Otherwise it will slip to 4.19.
tfiga: <u>hverkuil_</u>: I'd wait with this then, I'd say some of my comments already go beyond "minor changes"
hverkuil_: OK, good to know.
stark3y: <u>ezequielg</u>: ping?
<br> <u>tfiga</u>: I think the relevant fence would be lost by the time it reached the consumer (might have been through several devices, with/without their own fences)
<br> You'd need some metadata which accompanies the content through multiple devices (and even multiple buffers)
***: prabhakarlad has left
javier__: <u>mchehab</u>: have you tested the tvp5150 + omap3isp on your igepv2 board lately?
<br> I wanted to finally re-spin the tvp5150 input connectors series this afternoon but noticed that it doesn't work on v4.17-rc6
<br> last time I tested was on v4.9, VIDIOC_STREAMON returns -EPIPE that's because cdc_link_validate() fails due ISP CCDC src and sink pads fmt not being the same
<br> https://elixir.bootlin.com/linux/v4.17-rc6/source/drivers/media/platform/omap3isp/ispccdc.c#L2410
<br> the culprit is commit 0866df8dffd ("[media] tvp5150: fix pad format frame height")
<br> <u>mchehab</u>: anyways, just because you mentioned the igepv3 and omap3isp the other day in your e-mail about MC-based devices the other day
mchehab: <u>javier__</u>: no, I didn't
<br> have been busy those days, and was without space to test boards
<br> this week I added an extra "layer" of computers on the top of my table, where I placed the usual stuff...
<br> that should be giving me space for testing other boards
javier__: <u>mchehab</u>: I see, so it seems that's not the only regression. I reverted that commit to test and now it hangs on DQBUF
mchehab: if needed, I can do some tests during the merge window
<br> :-(
javier__: same user-space and MC pipeline but with v4.9 works
mchehab: there weren't many changes on omapisp3 since 4.9, as far as I remember
pinchartl: no, not many changes
<br> but code around the driver has changed too
<br> such as vb2
<br> omap3isp still receives testing, but quite infrequently
mchehab: could it be a performance issue?
pinchartl: likely not
mchehab: seek for this thread: dvb usb issues since kernel 4.9
javier__: <u>mchehab</u>: no, it hangs forever. And if you send a signal to the process the driver says:
pinchartl: usb isn't involved here
javier__: [ 63.518585] omap3isp 480bc000.isp: CCDC stop timeout!
<br> [ 63.525604] omap3isp 480bc000.isp: Unable to stop OMAP3 ISP CCDC
pinchartl: <u>javier__</u>: I suspect a change in the source configuration
mchehab: the problem is not at USB layer
pinchartl: in the tvp driver in this case
mchehab: https://patchwork.kernel.org/patch/10150455/
<br> try Linus patches and see if it changes something
<br> (no fix related to it was applied upstream, afaikt)
pinchartl: it's likely not an irq latency issue, no
javier__: <u>mchehab</u>: I really don't think it's related, but I can give it a try
pinchartl: <u>javier__</u>: I'd recommend looking at the changes in the tvp driver
javier__: <u>pinchartl</u>: yes, I'll dig on this
<br> <u>mchehab</u>: btw, I think we should just merge the revert I posted for the connectors
<br> the DT binging has been reverted years ago anyway
<br> I wanted to post the latest version today but it seems that won't be trivial for me to test with these issues
<br> and I can only do this on my free time
<br> *binding, it seems I can't today
pH5: I guess 0866df8dffd was wrong after all? At the time it was incorrectly(?) understood that v4l2_mbus_framefmt height referred to the full frame height (https://patchwork.linuxtv.org/patch/40513/), but the spec has been clarified in the other direction since: 0018147c964e ("media: v4l: doc: Clarify v4l2_mbus_fmt height definition").
javier__: <u>pH5</u>: this is what I use to test fyi https://hastebin.com/ucosoviyot.hs
<br> <u>mchehab</u>: this is the revert patch https://patchwork.kernel.org/patch/10094209/, it still applies cleanly on top of media_tree master branch
pH5: That 720x240 interlaced-tb on the CCDC output pad doesn't make any sense to me.
javier__: <u>pH5</u>: right, that may be wrong. It's set to 720x480 though
<br> this is the MC graph https://hastebin.com/oferujezah.hs
<br> media-ctl -v --set-format '"OMAP3 ISP CCDC":1 [UYVY2X8 720x480 field:interlaced-tb]' has the same effect
***: Guest11917 has quit IRC (Ping timeout: 276 seconds)