PINE64

Full Version: Linux Kernal Development - BETA Boards
You're currently viewing a stripped down version of our content. View the full version with proper formatting.
Pages: 1 2 3 4 5 6 7 8
(01-14-2016, 10:29 AM)tllim Wrote: [ -> ]I will check out the broken SDK link at wiki page and correct in 24 hours.

Thx, but it's not a broken link. I tried to uncompress your lichee_linux_310.tar.gz (tar xzf) but get errors even in this early stage. Using shasum I get 0022016ef33fd520a1f48d266bb4f3011899a518 for lichee_linux_310.tar.gz when downloaded from Google drive. Can you please confirm this is correct?

(01-14-2016, 10:51 AM)paulieg Wrote: [ -> ]That particular issue with the arm boot protocol is just one of many that will prevent you from booting a mainline kernel.

Ah, I missed that they're talking about mainline kernel. I thought they try to combine the BSP's u-boot+kernel with a Linux userland...
(01-14-2016, 06:31 AM)paulieg Wrote: [ -> ]
(01-14-2016, 05:24 AM)Galileo Wrote: [ -> ]JFTR: Still nobody managed to get Linux running on A64:

http://linux-sunxi.org/Pine64#BSP
http://irclog.whitequark.org/linux-sunxi...4#15188515;

Considering very few devs actually have the boards yet, that's not surprising. I have no inside info, but it's probably just the 2 cards from an earlier batch pine64 promised sunxi devs. Mine is still sitting in customs.

-p

Should be more than 2, we have already delivered more than 50 developer boards.

(01-14-2016, 11:24 AM)Galileo Wrote: [ -> ]
(01-14-2016, 10:29 AM)tllim Wrote: [ -> ]I will check out the broken SDK link at wiki page and correct in 24 hours.

Thx, but it's not a broken link. I tried to uncompress your lichee_linux_310.tar.gz (tar xzf) but get errors even in this early stage. Using shasum I get 0022016ef33fd520a1f48d266bb4f3011899a518 for lichee_linux_310.tar.gz when downloaded from Google drive. Can you please confirm this is correct?

(01-14-2016, 10:51 AM)paulieg Wrote: [ -> ]That particular issue with the arm boot protocol is just one of many that will prevent you from booting a mainline kernel.

Ah, I missed that they're talking about mainline kernel. I thought they try to combine the BSP's u-boot+kernel with a Linux userland...
We have try and uncompressed correctly. However, we will check out again. Please make sure that you uncompress under Linux environment, not under Windows environment.
(01-14-2016, 11:24 AM)Galileo Wrote: [ -> ]Thx, but it's not a broken link. I tried to uncompress your lichee_linux_310.tar.gz (tar xzf) but get errors even in this early stage.

Are you talking about some git files being duplicates? That's harmless, they've managed to do something silly with links I'd imagine. In any case, that's likely only an issue under Windows.

-p
(01-14-2016, 12:02 PM)paulieg Wrote: [ -> ]
(01-14-2016, 11:24 AM)Galileo Wrote: [ -> ]Thx, but it's not a broken link. I tried to uncompress your lichee_linux_310.tar.gz (tar xzf) but get errors even in this early stage.

Are you talking about some git files being duplicates? That's harmless, they've managed to do something silly with links I'd imagine. In any case, that's likely only an issue under Windows.

-p
The duplicates happen when uncompress under Windows. When uncompress in Linux, should be OK.
(01-14-2016, 12:09 PM)tllim Wrote: [ -> ]
(01-14-2016, 12:02 PM)paulieg Wrote: [ -> ]
(01-14-2016, 11:24 AM)Galileo Wrote: [ -> ]Thx, but it's not a broken link. I tried to uncompress your lichee_linux_310.tar.gz (tar xzf) but get errors even in this early stage.

Are you talking about some git files being duplicates? That's harmless, they've managed to do something silly with links I'd imagine. In any case, that's likely only an issue under Windows.

-p
The duplicates happen when uncompress under Windows. When uncompress in Linux, should be OK.

Correct. In any case, they happen for files which are of no consequence and are not actually errors. This certainly does not equate to the BSP being 'broken'.

-p
I only tried it in Linux (Ubuntu 14.04 LTS, 64bit). Based on your statements it seems you're able to build the BSP? Contrary to http://linux-sunxi.org/Pine64#U-Boot -- Since you're convinced the BSP isn't broken I would assume you already built it?

And another one that puzzles me: When you start working on Linux for A64 where will development happen? Do you share your code/fixes on Github? Working together with or against the linux-sunxi folks?
(01-15-2016, 07:44 AM)Galileo Wrote: [ -> ]I only tried it in Linux (Ubuntu 14.04 LTS, 64bit). Based on your statements it seems you're able to build the BSP? Contrary to http://linux-sunxi.org/Pine64#U-Boot -- Since you're convinced the BSP isn't broken I would assume you already built it?

I said nothing of the sort and your induction is incorrect. All I said was that none of what you've cited indicates that it is broken. It may or may not be. At this point, it is orthogonal to the task of booting to a gnu/linux rootfs.

Quote:And another one that puzzles me: When you start working on Linux for A64 where will development happen? Do you share your code/fixes on Github? Working together with or against the linux-sunxi folks?

Since the first order of business is booting a minimal gnu/linux rootfs, the deliverable is likely to be an image + package repo and not actual code. My guess is that most, if not all, issues related to supporting a 3.0.x kernel can be resolved by rolling back to older versions that do not require kernel > 3.0.x, eg in the case of systemd. If code changes are needed for particular software, it might have to be forked since the chances of getting patches accepted to support an out of date kernel version are unlikely to be accepted, in which case it'll go up on github, yes.

My understanding of sunxi is that they work towards mainline support. While that is both admirable and desirable in the medium to long term, it does not gives us practical linux support in the short term. Since the latter is where my interests lie as of right now, that's where my efforts will be focused to the degree which other, more pressing time commitments allow. My guess is that all the different folks working on this sort of stuff will be posting on this forum and the pine64 wiki. If the community develops to a significant enough size that it makes collaboration and coordination necessary and practical, I'd like to see a mailing list as it's more congruent with the workflow most open source devs are familiar with.

I'm not sure what working 'against' linux-sunxi would even look like, never mind why it would be desirable. I'm just one hacker and don't speak for others or pine64, but I seriously doubt anyone would have an issue with anyone else working to improve linux support in whatever way they can along whichever vector suits them and their goals best.

-p
Ok, you neither looked into the BSP (kernel 3.10.65 not 3.0.x as you're always writing) nor linux-sunxi community (dealing also with 'legacy' kernels from Allwinner)
(01-15-2016, 08:02 AM)paulieg Wrote: [ -> ]I'm not sure what working 'against' linux-sunxi would even look like, never mind why it would be desirable. I'm just one hacker and don't speak for others or pine64, but I seriously doubt anyone would have an issue with anyone else working to improve linux support in whatever way they can along whichever vector suits them and their goals best.

-p

linux-sunxi is a diverse group of people. Those that are doing the actual work are not very vocal.
Those that are vocal, can be very rude. I really cannot understand why they stay if the dislike Allwinner so much.

The working Linux kernel is based on 3.10, not 3.0.x.
(01-15-2016, 10:09 AM)taros Wrote: [ -> ]
(01-15-2016, 08:02 AM)paulieg Wrote: [ -> ]I'm not sure what working 'against' linux-sunxi would even look like, never mind why it would be desirable. I'm just one hacker and don't speak for others or pine64, but I seriously doubt anyone would have an issue with anyone else working to improve linux support in whatever way they can along whichever vector suits them and their goals best.

-p

linux-sunxi is a diverse group of people. Those that are doing the actual work are not very vocal.
Those that are vocal, can be very rude.

I don't have experience with linux-sunxi in particular, but lkml is generally perceived as very rude by most people. I'm in Linus' camp on the issue and think it works well for a number of reasons and it doesn't bother me in the least. In any case, it's not going to change any time soon. I suspect linux-sunxi is just an extension of that particular culture.

Quote: I really cannot understand why they stay if the dislike Allwinner so much.

People who hack on open source licensed projects tend to dislike it when companies take their hard work, use it for commercial gain and don't cooperate with the community that did the work. I fully understand that sentiment. In fact, if people didn't scream, shout and complain about companies that do this, we'd be in a much worse position with vendor cooperation than we are now (generally speaking), so it's productive as well even though it may not appear to be in any given instance.

I can't speak for them of course, but I tend to get the angriest about companies and products that are OK-ish, but could be truly amazing if only they did the right thing. You first have to be passionate about the potential to get angry. If you're ambivalent, chances are you just won't be bothered enough to say anything.

Quote:The working Linux kernel is based on 3.10, not 3.0.x.

Must've been an update then since I looked at it or I was misremembering. Thanks for correcting me. In any case, that doesn't improve the situation wrt work needing to be done on the rootfs and package dependencies significantly. However, unlike other people, I'm not that hot under the collar about an older kernel.

-p
Pages: 1 2 3 4 5 6 7 8