I boot tested 9.1.001 on Slice3 before posting the alpha images. At some point soon I will be in Europe and will post the box to HiassofT who is much better skilled at fixing the irregular device-tree breakage than me 
Posts by chewitt
-
-
VP9: "Currently software decode only. There is a hardware block that can do vp9 bitstream (cabac style) acceleration but the pixel operations need software decode which is the majority of the work." .. my 2c is that VP9 is never seen beyond YouTube which almost always has other (hardware decode supported) streams so it's probably not high on the list of things to spend time upon.
-
Incredibly exciting news. I wonder whether the device would support things such as Dolby Atmos and Dolby Vision.
I'm told Atmos should be okay. If you aren't using pass-through you'll only get the DTS core (5.1) anyway and if it is being passed through it's not being decoded and thus a license isn't needed. Or something like that .. IANAL
-
No need to build WeTek stuff for LE 9.2 as it doesn't exist. LE 9.0 is the last release with Amlogic_Legacy support. We might go back and push LE 9.0.3 so that people have K18.3 .. not decided yet (and no time to think about it right now).
-
I've been told that RPi4 will support open "static" metadata standards (HDR/HLG) not the commercial "dynamic" ones like Dolby Vision which would require a license fee to be paid on every board sold. Not sure about Atmos audio, but if it requires fees I'd expect the position to be the same. I've asked for a clarification.
-
Will LE ever be a stable image for odroid N2? If so, what do you think it takes? You're doing a great job! Thank you for
Yes, but I'm not expecting to add "stable" to my frequently used words list until LE 10.0 ships
-
Tthe mrfixit2001 codebase is using the "no public upstream Broadcom sources available" bcmdhd driver and I couldn't find the repo but would guess ayufan has the same. It's fine to have bcmdhd in distro releases that target a single piece of hardware. It's not okay to have that in the rest of LE which targets many things as proper support needs to be upstream.
-
Just as a general note for people. The team has made a decision not to release Amlogic (and Allwinner) images as part of LE 9.2 which is currently being built and will see an ALPHA1 release this week. Amlogic mainline has several outstanding technical decisions related to boot and Panfrost has some critical bugs that developers simply haven't had time to look at properly due to current work and family commitments. The summer vacation season is also about to start - so the last thing we need is a bump in user numbers while people resources are lighter. There are also large changes pending for Linux 5.3 which will be beneficial (G12 devices running at full clock speed will be great) and this also means a load of device-tree rework and testing effort. So instead of spending lots of time trying to support a Linux 5.1 based LE 9.2 release with a load of complex backporting work, the mainline codebase will make better progress with less effort if we concentrate upon the master branch where we can shift focus to Linux 5.3 where more changes are upstream and the amount of commit/patch juggling reduces.
As I ended up being release manager for LE 9.2 and I'm also one of the people with too many work commitments, the final plumbing changes needed to start master nightlies and push add-ons will have to wait until the dust settles on the first LE 9.2 alpha. NB: VIM3 users with a shiny new board will be catered for one way or another .. no need for anyone to panic about that.
-
The VIM3 devices being launched are all A311D chips. The VIM3L (notice the L) with S905D3 chips in the batch are developer samples - nothing that will be sold to users. Software support for these boards on the mainline kernel is trivial. The current device-tree that I made will need some rework due to changes coming in Linux 5.3 but that's no drama.
-
The initial production batch will be a small number of S905D3 (VIM3L) boards and the rest are split between VIM3 Basic/Pro using A311D chip - there are no more boards with S922X. Kodi runs fine on a board with 1GB ram and has no need for accelerometers, camera interfaces, or NPU functions. So it will be a nice fast board, but I think the VIM3L will be more aligned to LE user needs and budget. I also have an educated hunch we'll see more plot twists before Khadas settles on the final line-up for the VIM3 family.
-
Friends, we know that a lot of people are using dual boot android boxes.
I did not see whether someone has asked before but in the 9.0.2 version shutdown options there is no "Reboot To Android" ...
Does anyone have information?
Reboot to Android requires the GUI hack to write a value to the u-boot environment to toggle the boot device from SD card to eMMC. LE has never included that hack in any of our official images. NB: This thread is for mainline kernel images not 9.0.2 legacy image questions.
-
Bad SD card? LE does not need NOOBS to work. If you only want to install LE it will be easier to install it with our USB/SD Creator app.
-
It should work with the right combination of autoscript files on the USB, but I don't think the 8.2.x images have the right combination. Mainline kernel images should boot, but those won't be quite as functional right now (getting there, but still work in progress). The easiest option is probably to get another SD card.
-
-
Write the .img.gz file to an SD card using the LE USB/SD Creator App .. same as any other image that LE offers
-
The last image amlg12 ... X96 runs superbly on the bqeel y5
what device tree are you using? .. and what remote keymap (assuming it's the remote that ships with the box)?
-
"No rule to make target" means you're calling "make <something>" and <something> doesn't exist in the "Makefile" in the sources.
e.g. Makefile provides build targets for apples, pears and organges, but you're running "make apricots" which doesn't exist.
-
BT works fine with the commit I listed above - this chip is in a number of Amlogic boards that I'm poking right now. If you can point me to sources for those distros i'll have a look, but they're probably using the BCMDHD driver. This works, but there's no way we're switching from "supports everything except this one chip" well written/maintained upstream kernel code to "fugly out of tree Android BSP kernel cap" .. it would be retrograde step
