01:12 <libv> magnificent. on the 3.8 kernel, using planes results in a nice kernel panic...

01:12 <libv> memcpy to fbdev it is :(

01:15 <libv> javier__ his tree would be pointless anyway as i first would need to get the mali kernel space going on it, and -ENOTIME for that before FOSDEM

01:26 <Tenkawa> brb

01:27 <R0b0t1> any idea what it would take to adapt hardkernel's u-boot to the mainline u-boot? I noticed there was some restructuring

01:28 <libv> R0b0t1: try it

01:29 <R0b0t1> well I mean this is me trying it

01:29 <R0b0t1> do what, just copy everything over and move the config?

01:29 <libv> wow, is that how you approach problems like this?

01:30 <R0b0t1> no, this is me gathering information

01:30 <Tenkawa> ugggh

01:30 <Tenkawa> I've got to be missing something simple

01:30 <libv> R0b0t1: they are both git trees with a communal basis, right?

01:31 <R0b0t1> I'm going to check the directory structure again, but I really don't think that fact helps me based on what I noticed...

01:31 <R0b0t1> (otherwise, yes, the solution would be rather braindead)

01:31 <Tenkawa> what would you all say is the most likely culprit if I dd a signed kernel to a partition and i get nothing but a black screen on boot (not even the white screen and boot)

01:32 <Tenkawa> goback to the internal one and no problem

01:32 <Tenkawa> I pulled down the tree. ran exynos_defconfig, make, vuilt.. dd to kern partition.. reboot and just black screen if I hit ctrl u

01:32 <Tenkawa> er built

01:33 <libv> R0b0t1: no, working off the git trees is also non-trivial, but it makes it a lot easier as it gives you direct points of comparison, so you can turn the odroid tree increasingly more in the u-boot tree

01:33 <R0b0t1> yes

01:33 <libv> what other pointers would you expect from an irc channel?

01:33 <R0b0t1> oh yeah whoops I'm not sure what I convinced myself of earlier

01:34 <libv> R0b0t1: did you expect someone else to have done this and explain in detail how to do this, without just putting the code online?

01:35 <R0b0t1> ... no, which is really what pissed me off about your reply. I never actually asked anyone to do it for me. I asked if anyone had knowledge of u-boot w.r.t. hardkernel's boards.

01:36 <libv> R0b0t1: to me, your question was quite wrong, as if you really expected people to have done what i just described

01:37 <libv> you should've asked "is anyone working on bringing the odroid uboot tree in line with mainline"

01:37 <Tenkawa> DOH

01:38 <Tenkawa> I think I migh know what I did

01:39 <R0b0t1> no, that is still the wrong question - most people consider it done with the inclusion of hardkernel's xu3 configuration. I looked, yes. I'm working with an older board which most people seem to be uninterested in. They seem uninterested because there's nothing about merging the xu configuration into mainline, just a bunch of orphaned repos.

01:40 <Tenkawa> I might have not copied the dtb to boot

01:41 <Tenkawa> does the cat the dtb to the end of uImage work?

01:42 <Tenkawa> lets see

01:42 <R0b0t1> there is an option for it, I believe, otherwise you specify it separately in the boot command.

01:43 <Tenkawa> brb.. i hope

01:55 <R0b0t1> must not have ended well

05:42 <R0b0t1> afaik doing everything properly, boot hangs at "Starting kernel..."

05:42 <R0b0t1> I mean. Not too properly. I expect it to fail when it can't find a root filesystem.

07:56 <sjoerd> There are loads of different odroid trees, really depends on what boards you're using for uboot/kernel

07:56 <sjoerd> libv: If you want the mali kernel space, we've got kernel trees with those as well

07:58 <sjoerd> R0b0t1: fwiw X2/U2/U3/XU3 work with upstream u-boot & kernel, not sure what happend with the XU, that board seems less popular

07:59 <R0b0t1> yeah might just ditch it

07:59 <R0b0t1> I have a frankenkernel and loader and all I've got to show for it is some instability

08:01 <sjoerd> I know some folks are booting that board with a mainline kernel using the exynos5410-smdk5410 dtb

08:01 <sjoerd> So if you care enough to do the work, taking that dts and correcting it for the XU shouldn't be that tricky

08:01 <R0b0t1> https://github.com/afaerber/u-boot/tree/odroid-hardkernel-v2012.07 - guy got ethernet in the bootloader working, and has a kernel fork and was/is working on mainlining some stuff? but I'm basically at the point where I don't know why the kernel isn't decompressing

08:02 <Wizzup> XU should do ok in 3.20 no?

08:02 <sjoerd> My XU3 dts got merged for 3.20, didn't think the XU was

08:03 <R0b0t1> sjoerd: yeah I don't mind doing the work but I think I effectively just did all there is to do.

08:03 <sjoerd> R0b0t1: typically because the dtb you added was wrong

08:04 <sjoerd> Or that's often what i have noobed if the kernel doesn't want to talk to me :)

08:04 <R0b0t1> I see... I used exynos5410-smdk5410.dtb

08:05 <R0b0t1> from odroid-13.y

08:05 <sjoerd> With what kernel tree?

08:07 <R0b0t1> w/ odroid-3.13.y, 3.17.7-hardened-r1, and https://github.com/afaerber/linux/tree/odroidxu

08:09 <R0b0t1> if you have a rough sketch of the steps I need to do, I will try it again in a bit

08:09 <sjoerd> heh, but afaebers tree has a dts for the odroid specifically

08:10 <sjoerd> mixing odroid-3.13.y and afaebers tree sounds like a pretty bad idea

08:10 <sjoerd> I don't know what your end-goal is :)

08:11 <sjoerd> but i'd start with just afaebers tree or with upstreamer master, pick up the dts from afeaber and slowly start adding on the things you need

08:11 <R0b0t1> yes, but I didn't have many options to start with

08:17 <sjoerd> fwiw in my experience, to get things rulling, it's best to get a dts upstream as soon as you have even the most basic support

08:17 <sjoerd> otherwise you'll keep the random trees issue which always leads to issues

08:44 <R0b0t1> yeah this wasn't intended to be long-term

08:45 <sjoerd> fwiw the hardkernel odroid kernels tend to be quite heavily hacked kernels based on some samsung vendor trees/bsp. So mixing those with anything mainline will quite likely result in total disaster

08:48 <R0b0t1> yeah wasn't my first option, but was the closest thing to a "known" "working" that I had

08:49 <R0b0t1> I wish the disasters were louder

11:51 <R0b0t1> sjoerd: thanks got it

19:25 <libv> sjoerd: i am going to make do with memcpy for now

19:25 <libv> but thanks