08:10 <brads> Hi Guys, on the odroid c2 there was a usb issue in 4.18/4.19 which caused issues initialising the 4 port usb hub and devices not appearing when plugged in after boot. Looks like there are some dwc2 drivers work applied to 4.20 which might solve the problem (similar to the fix for dwc2 rockchip driver which has just been added) - https://git.kernel.org/pub/scm/linux/kernel/git/torvald

08:10 <brads> s/linux.git/log/?h=v4.20-rc6&qt=grep&q=dwc2

08:20 <narmstrong> It does and xdarklight pushes the fix yesterday

08:22 <brads> Thanks!

08:57 <xdarklight> brads: I'd appreciate a Tested-by on the patch as I only tested this on an older SoC (older than GXBB / S905): https://lore.kernel.org/linux-amlogic/20181209190129.11870-1-martin.blumenstingl@googlemail.com/T/#t

09:25 <brads> xdarklight: I'll try it out

09:40 <Darkmatter66> Hi all, any idea on mainlining vdec support in ffmpeg,mpv and kodi? I haven't seen the patches by lukas in the mailing list since august and they where not merged AFAIK!

09:40 <Darkmatter66> mjourdan: I think you mentioned that you are working on fixing userspace support now that the driver mainlining is blocked by licence issues, any news on that ?!

10:07 <mjourdan> Darkmatter66: this work resulted in off-tree patches for kodi, ffmpeg and chromium. Most of those live in LibreElec and an ubuntu image builder. ffmpeg's v4l2m2m decoder needs an overhaul and no one's really sure about the way to do that yet.

10:10 <Darkmatter66> mjourdan yes I am already following chewitt LibreElec tree

10:10 <Darkmatter66> I was hoping that the patches reach the mainline tree

12:24 <brads> xdarklight: usb now works fine with the patch applied, Nice! ill remove it to confirm the reverse and give a Tested-by

13:30 <sheepman> ccaione: the patch you suggested works a treat! Ignore me if you're busy but mind if I ask how you go about diagnosing such an issue?

13:33 <ccaione> sheepman: that's a good news. Briefly: we had a way to reproduce the issue. If you look at the GMAC registers you can see that the IRQs are not ACKed anymore. Manually acking the registers make the problem going away. The ACK is done in the ISR so the problem was that the ISR was not triggered anymore. From there the idea to change the trigger type

13:35 <sheepman> I vaguely get what you're saying but GMAC? and ISR?

13:36 <sheepman> interrupt service routine presumably

13:37 <sheepman> Gigabit Media Access Controller?

13:37 <sheepman> if so i'll start reading :)

13:38 <ccaione> yup

13:40 <sheepman> ok cool thanks again :)

13:42 <sheepman> I rarely delve into hardware at this level so it'll be interesting that's for sure.

18:02 <warpme_> fantastic! LIMA guys are doing fantastic job. I have initial working mythtv on s905w with GLES. https://pasteboard.co/HR7b6UD.jpg good!