01:15
zombah has quit [Quit: Leaving]
03:52
dlan has quit [Ping timeout: 256 seconds]
03:52
dlan has joined #linux-exynos
05:16
amitk has joined #linux-exynos
08:48
Gottox has joined #linux-exynos
08:50
<
Gottox >
Hi! I'm currently trying to compile
08:50
<
Gottox >
u-boot for my chromebook
08:51
<
Gottox >
*samsung chromebook
08:51
<
Gottox >
is the vanilla u-boot enough for that?
08:54
<
Wizzup >
yes, recent u-boot should work fine
08:56
<
Gottox >
thx. trying to reproduce the nv-u-boot from source without the whole chromiumos tree.
09:14
<
Wizzup >
there's just a defconfig for that
09:14
<
Wizzup >
let me clone the repo
09:16
<
Wizzup >
what chromebook do you have btw
09:17
<
Wizzup >
./u-boot/configs/snow_defconfig ?
09:18
<
Gottox >
ah, I tried smdk5250 for now.
09:18
<
Wizzup >
make snow_config
09:20
<
Gottox >
thx, lets see.
09:24
* Wizzup
is going through old wiki pages
09:29
<
Gottox >
yea, saw that mainline was supported but overlooked the defconfig :D
09:29
<
Wizzup >
I just added that.
09:30
<
Wizzup >
(thanks to your feedback)
09:30
<
Wizzup >
we have generic (poor) guides for mainline kernel and u-boot building that need love
09:30
<
Gottox >
ah that's why I haven't foudn it :)
09:30
<
Wizzup >
if you want I can give you a wiki account
09:30
<
Gottox >
sure, but I can't promise anything.
09:31
<
Wizzup >
pm me username/email for the acc
09:48
<
Wizzup >
and then dd u-boot-dtb.bin to the first partition
09:49
<
Wizzup >
at least, I don't think vbutil is needed
09:49
<
sjoerd >
vbutil is needed
09:49
<
sjoerd >
you need to sign it just like you do a kernel fit image
09:52
<
Wizzup >
should reflect that in our u-boot guide
09:52
* sjoerd
still has a todo to add u-boot patches to do that when building for a chromebook
09:52
<
sjoerd >
so little time
09:53
<
Wizzup >
oh, it is in the normal one
09:53
<
Wizzup >
s/one/page/
09:53
<
Wizzup >
sjoerd: patches to what? u-boot?
09:54
<
sjoerd >
As doing this by hand is tedious, i'd prefer the build system to just give me a .kpart or whatever that i can put in place :)
09:54
<
sjoerd >
Especially when doing it for the nyans which need extra special care thusfar
09:58
<
Wizzup >
yeah, I also feel the mkimage step can be annoying
09:58
<
Wizzup >
since then you need to know what the loadaddr for the specific device is as well
09:58
<
Wizzup >
but perhaps that is already found somewhere in u-boot (doc?)
09:59
<
sjoerd >
They match the u-boot load addresses for the respective boards
09:59
<
sjoerd >
U-boot configuration tells u-boot where it expects to be loaded, so you need to have the mkimage match
10:01
<
Wizzup >
right, would be nice if the u-boot process would just do that ... :)
10:01
<
Wizzup >
the u-boot build process*
10:03
<
sjoerd >
it's jsut a matter of writing some code
10:04
<
javier__ >
sjoerd: that would be great indeed
11:02
zombah has joined #linux-exynos
11:05
jnettlet has joined #linux-exynos
11:13
<
Gottox >
how do I generate a key for u-boot?
11:14
<
Wizzup >
get vboot-utils
11:20
<
Gottox >
it's already there
11:21
<
Wizzup >
then you should have keys
11:22
<
Gottox >
unfortunately not, as I had to build vboot-utils by myself.
11:23
<
Wizzup >
see tests/devkeys/
14:05
si1v3r has quit [Ping timeout: 256 seconds]
14:54
si1v3r has joined #linux-exynos
15:03
<
si1v3r >
Wizzup: that page was probably a wiki editing mistake, if i remember right.
15:04
<
Wizzup >
perhaps we should hide/remove it
15:15
<
si1v3r >
its the arm template based one right?
15:22
<
Wizzup >
we need to fix the templates (as an aside)
16:03
<
si1v3r >
I think a 'state of the kernel' breakdown of features and when they started working would be very useful
16:04
<
Wizzup >
I have something like that per device
16:04
<
Wizzup >
well, yeah
16:04
<
Wizzup >
ok, I don't have that really
16:04
<
si1v3r >
It's all scattered around
16:11
<
Gottox >
yay, u-boot is booting! :D
16:13
<
Gottox >
I'd like to have some docs about how the signed uboot finds the right partition. Is it only done through gpt or can I use msdos too?
16:13
<
si1v3r >
It must be gpt
16:14
<
si1v3r >
it looks for a partition with a cros kernel part type, and a bootable flag.
16:14
<
si1v3r >
and a clear failure flag, I believe.
16:15
<
si1v3r >
I stick nv uboot in a partiton and give it those flags to get my boot process working.
16:16
<
Gottox >
but it can't fatload I guess.
16:19
<
si1v3r >
fatload works fine
16:19
<
si1v3r >
by default it picks up a file from mmc0:12/u-boot
16:19
<
si1v3r >
uboot being the folder the scripts live in.
16:19
<
Wizzup >
si1v3r: I do the same @ partition
16:30
amitk has quit [Quit: leaving]
16:41
<
si1v3r >
I set up my partitions with that script.
17:55
liquidAcid has joined #linux-exynos
18:57
<
javier__ >
Gottox: the RO u-boot can only boot a signed image from a gpt partition
18:58
<
javier__ >
that is a RAW image on a partition and the partition that chooses to boot depends on the GPT table priority (P) and successful (S) attributes
19:00
<
javier__ >
Gottox: it tries to boot the GTP partition that has the highest priority a number of times as specified in the GTP partition tries (T) attribute
19:01
<
javier__ >
but once the signed nv-u-boot image is loaded, then you can use fatload or whatever you want to load kernel, just like any other u-boot
19:07
<
javier__ >
si1v3r: the tries flag is only used if the successful boot flag is 0 so there is no need to set -T 5 in your script since you are already setting -S 1
19:14
<
si1v3r >
That's good to know, thanks.
19:58
<
Gottox >
javier__: Yea, I'm currently tinkering with u-boot to boot a kernel.
19:59
<
Wizzup >
shouldn't be too hard
19:59
<
Wizzup >
basically ext2load mmc 0:@ (or 1:2) ${loadaddr} /file
20:00
<
Wizzup >
basically ext2load mmc 0:2 (or 1:2) ${loadaddr} /file
20:00
<
si1v3r >
It's not compatible with the defaults but it works.
20:00
<
Wizzup >
si1v3r: that is not snow though
20:00
<
si1v3r >
They are the same
20:01
<
Wizzup >
so the loadaddr may be different?
20:01
<
si1v3r >
afaik, any loadaddr around 0x41000000 - 0x40008000 works
20:02
<
Wizzup >
yeah I use 0x400....
20:54
liquidAcid has quit [Quit: Leaving]