+- PINE64 (https://forum.pine64.org)
+-- Forum: PINE A64(+) (https://forum.pine64.org/forumdisplay.php?fid=4)
+--- Forum: Linux on Pine A64(+) (https://forum.pine64.org/forumdisplay.php?fid=6)
+---- Forum: Ubuntu (https://forum.pine64.org/forumdisplay.php?fid=27)
+---- Thread: (ARCHIVED) Ubuntu Xenial Image (BSP Kernel) (/showthread.php?tid=376)
RE: Ubuntu Xenial Image (BSP Kernel) - Pander - 10-17-2016
Sticking for five years to LTS would not make sense as nobody wants to use old software. Sooner or later, all has to be upgraded. So use both ways, good to develop/support LTS but at the same time offer a build for Yakkety for those who want to test (and use) it.Even though lots is tied in with certain versions, not upgrading is a death for the user community and all the good work that has been done. So 'forward porting' I would recommend. See also https://en.m.wikipedia.org/wiki/Release_early,_release_often
RE: Ubuntu Xenial Image (BSP Kernel) - kflorek46 - 10-18-2016
Pander,
Ubuntu supports an LTS for 5 years, but it puts out a new LTS every 2 years. Longsleep does not supply or support anything but his kernel and scripts, with no desktop. One of his scripts installs the Mate desktop if you want. He has adapted his kernel to Ubuntu xenial.
I think a lot of people are under a wrong impression, as I was. The person who is doing the work (Longsleep) to create this image is not part of the Pine64 company and is doing this for his own reasons. It is by his generosity that there are any working versions of Linux for Pine64 (although he is using work from sunxi and others), but he is getting nothing for it. Pine64 is not funding any version of Linux, other than donating hardware, AFAIK. There are a few people, but very few, who are contributing any work or knowledge, and I am not one of them.
Some people who know all this seem to think that Raspberry PI is different, because of Raspbian (Debian for Raspberry PI). To the best of my knowledge Raspbian is done by a group not funded by Raspberry PI Foundation (a not-for-profit company) although PI does call Raspbian officially supported. (https://en.wikipedia.org/wiki/Raspbian) The user base for Raspberry PI has become huge over the many years since its first PI (2011), and so has the number of contributors, and therefore so has the amount of time and knowledge which is contributed. There is no reason to think the Pine64 can do without the same thing.
Open-source software has a cost, and it's cost is participation.
I think there is a fair chance that yakkety might work smoother than xenial on a Pine64. I know it does on my x86. For the heck of it, I might try the update I mentioned sometime in the next few years, or I might not, but if it doesn't work I would be totally lost trying to figure out why. If somebody tries it, I would appreciate hearing about it.
RE: Ubuntu Xenial Image (BSP Kernel) - Pander - 10-18-2016
Thanks for your reply. I am aware of all this already. I am offering my help in testing, as will probably others. Only I am not the one to add an (unstable yakkety) build to longsleep's PPA. so if it is there, I am there to test.
RE: Ubuntu Xenial Image (BSP Kernel) - Morp-heus - 10-28-2016
Hello I have a Pine64 2gb + Wifi module.
I have the xenial-pine64-bspkernel-20160716-1.img.xz 16-Jul-2016 15:41 171Mb
But unfortunately the wifi module is not recognized!
And how can I copy ubuntu server to sd carte?
Can someone please help me?
Thank you.
RE: Ubuntu Xenial Image (BSP Kernel) - pfeerick - 10-28-2016
(10-16-2016, 04:39 PM)kflorek46 Wrote: If I were going to try this, I would run the program that Ubuntu support recommends:
update-manager
kflorek46, that is only for the GUI/DE environment, isn't it? I'm running a headless build, so the appropriate command is do-release-upgrade (the ubuntu equivalent of apt dist-upgrade)
(10-18-2016, 03:36 AM)Pander Wrote: Thanks for your reply. I am aware of all this already. I am offering my help in testing, as will probably others. Only I am not the one to add an (unstable yakkety) build to longsleep's PPA. so if it is there, I am there to test.
I'm sure your offer will be much appreciated. I can confirm from a quick test that following three steps will successfully upgrade the xenial ubuntu image to yakkety. After the upgrade, I still have working ethernet, wifi and audio. No DE is installed, so I don't know if there are any issues there.
Code:
sudo apt-get install update-manager-core
#edit /etc/update-manager/release-upgrades so Prompt=normal (was Prompt=lts)
sudo do-release-upgradeCode:
Welcome to Ubuntu 16.10 (GNU/Linux 3.10.104-1-pine64-longsleep aarch64)
Linux pine64 3.10.104-1-pine64-longsleep #103 SMP PREEMPT Sun Oct 23 13:09:48 CEST 2016 aarch64 aarch64 aarch64 GNU/Linux
No LSB modules are available.
Distributor ID: Ubuntu
Description: Ubuntu 16.10
Release: 16.10
Codename: yakketyRE: Ubuntu Xenial Image (BSP Kernel) - longsleep - 10-28-2016
I will not support or recommend any non lts Ubuntu and neither should you. Especially 16.10 has almost no relevant changes. Any important major software upgrades will end up in the backports repo sooner or later. Ubuntu 16.04 is a great platform an there is no reason to use a newer one any time soon.
RE: Ubuntu Xenial Image (BSP Kernel) - kflorek46 - 10-28-2016
Pfeerick,
>That is only for the GUI/DE environment, isn't it?
Yes, update-manager is a GUI program. Also, upgrade-manager will not find an upgrade unless you have set synaptic to notify you of ALL upgrades. If it is set to only notify you of new LTS upgrades, it won't find yakkety, because it is not LTS (Long Time Support.) Other Ubuntu versions like yakkety are supported for 9 months (?), not 5 years.
I would expect what Pfeerick shows to do the same thing, including upgrading the desktop. Then Ubuntu should at least boot up either way. Since longsleep's repo is compiled for xenial, there is no assurance it will work with yakkety. But since the drivers only depend on the kernel, and the kernel does not change, the drivers really should work. One of the hitches is that update-manager will either disable the repo designations = xenial or update them to yakkety. I am too lazy right now to check how longsleep designates his repo, but if it needs to be set to back to xenial after an upgrade, you can do it manually (change text files in /etc/apt/sources.list or in subdirectory /etc/apt/sources.list.d/)
I'll repeat what I said before. If you are afraid of failing, don't try it.
If you are perfectly fine screwing up everything you have done so far, you are the right kind of person to try it. Don't demand any help or expect sympathy.
RE: Ubuntu Xenial Image (BSP Kernel) - pfeerick - 10-29-2016
(10-28-2016, 11:57 PM)kflorek46 Wrote: Pfeerick,
>That is only for the GUI/DE environment, isn't it?
Yes, update-manager is a GUI program. Also, upgrade-manager will not find an upgrade unless you have set synaptic to notify you of ALL upgrades. If it is set to only notify you of new LTS upgrades, it won't find yakkety, because it is not LTS (Long Time Support.) Other Ubuntu versions like yakkety are supported for 9 months (?), not 5 years.
I would expect what Pfeerick shows to do the same thing, including upgrading the desktop. Then Ubuntu should at least boot up either way. Since longsleep's repo is compiled for xenial, there is no assurance it will work with yakkety. But since the drivers only depend on the kernel, and the kernel does not change, the drivers really should work. One of the hitches is that update-manager will either disable the repo designations = xenial or update them to yakkety. I am too lazy right now to check how longsleep designates his repo, but if it needs to be set to back to xenial after an upgrade, you can do it manually (change text files in /etc/apt/sources.list or in subdirectory /etc/apt/sources.list.d/)
Yes, that's correct - if you're running a LTS build already, you also need to change the settings to find all updates, not just the next LTS build. And the sources file for the longsleep repo is in /etc/apt/sources.list.d
However, since longsleep has made it clear that he doesn't intend to support yakkety, I wouldn't recommend it (for that reason only). I have simply indicated that it is not only is it possible to upgrade as was asked, but still appears to be stable - in that the system doesn't immediately break. So for the adventurous, it is possible. However, since we don't have the benefits of the mainline kernel, not all of the improvements from newer versions of ubuntu make it to the pine64 - ie improvements to btrfs, docker, live kernel patching, etc. The reality of the situation is there should be no need to upgrade beyond a LTS build, unless you want some bleeding edge feature, and that is more relevant to a full blown desktop PC install anyway, not the pine64.
I am curious as to why Pander was so interested in the Yakkety build since it was only released... about 2 weeks ago? As longsleep pointed out, any fixes and software updates will make it through to xenial via the backports repository eventually. Ubuntu LTS builds are released every two years, are more thoroughly tested, and focus on consistency and stability, rather than introducing sudden feature and configuration changes which may arise in the normal releases. So why the rush? We're not talking about a version release with lots of wizz-bang features like an Android major version release...
Ubuntu Xenial Image (BSP Kernel) - RaspiBit - 11-02-2016
Does the wifi + BT module fully work on this image? I want to create an access point and link it to the ethernet interface.
RE: Ubuntu Xenial Image (BSP Kernel) - Pander - 11-03-2016
pfeerick, to also port the fixes/patches needed for LTS (xenial) to normal releases (yakkety and zesty) it will test if the fixes are forward compatible and also discover and confirm that certain fixes that have trickled upstream are no longer needed as fixes/patches in longsleep's build. Already managing PINE64 support at one level higher will serve as extra testing too.
There are a lot of PINE64 devices out there and user and developers will use them when also newer versions of distros run on them. Not for kernel parameters but for more up to date of other packages that don't need PINE64 specific support. So I am more talking about continuous integration. I know this is all done by volunteers, but I think many PINE64 owners are willing to provide support. Even though the 'promise' of the creators of PINE64 hasn't been all that. With a more structured infrastructure, I think we can lift it to a next level. Now all knowledge is distributed over IRC and a forum and some is on GitHub.
If there would be one overall project with a wiki, continuous integration and issue management, I am willing to put some financial support via https://www.bountysource.com on a lot of issues. I know this also takes time to setup, so that can have a bounty too in my opinion.