<amitk>
sjoerd: it is based on a branch khilman just created on top of linux-next and has changes to try to get mali working
<sjoerd>
oh extra nice
<sjoerd>
that's one of my things on my todo list as well so :)
<sjoerd>
(that and also have it workin on the cr2 with the same kernel)
<sjoerd>
amitk: will give it a try later on
<amitk>
sjoerd: at present the mali kernel driver fails to initialise due to lack of platform data
<amitk>
sjoerd: yes, we're targeting a tree that boots on both cb2 and odroid
<sjoerd>
great!
<sjoerd>
I'll keep you posted on my progress as well, hopefully you branch allows me to at least start userspace though.. that would be a good start :)
<zenmetsu>
would be nice to see mali added to a more recent kernel. kept running into ancient linaro kernels that lack DT support for the chromebook2 :/
<sjoerd>
zenmetsu: i've got a working 3.17 branch
<zenmetsu>
with mali?
<zenmetsu>
i wouldn't mind checking it out if you have a link to the repo. i'm looking into getting an XU3 as well
<sjoerd>
works fine on both peach pi (chromebook 2 13") and snow (original samsung arm chromebook)
<sjoerd>
khilman posted a picture on his G+ with the location of the pads
<zenmetsu>
sweet, i'll look into it
<zenmetsu>
surprisingly, i haven't even torn this thing open yet, and i usually hack up everything i get. any chance there is an i2c or spi bus hiding in there also, maybe even some gpio?
<sjoerd>
there is the debug header
<sjoerd>
I destroyed on cr2 with soldering on that one though, i suspect i managed to erase the EC flash but i'm not sure what happened
<sjoerd>
It did give me the EC serial though, which was cute while it lasted
<sjoerd>
Unfortunately if you buy a chreombook it doesn't have the debug connector on it, but the pads for it are exposed
<javier__>
zenmetsu: the RTC should work (both the S3C and the one in the max77802 PMIC)
<javier__>
zenmetsu: what kernel version are you using and what problem you have exactly? (e.g: RTC driver not probed, /sys/class/rtc/rtc0/time not incrementing, etc)
<javier__>
zenmetsu, amitk: I may have today to give wifi a try and see what's missing in the DTS
<javier__>
*time today
<zenmetsu>
sorry, was dealing with my puppy. I was getting rtc0 and rtc1 in /dev, rtc1 gave me an error when reading, and rtc0 takes forever to read and doesn't increment
<zenmetsu>
and i was using 3.18.0-rc6
<javier__>
zenmetsu: ok, I'll take a look
<javier__>
it was working last time I tried
<sjoerd>
javier__: see that's why you need to put some test into our lava for that :p
* sjoerd
runs
<zenmetsu>
it was khilman's branch though, pretty sure it was the wip/exynos/v3.19-integ
<javier__>
sjoerd: agreed
<javier__>
zenmetsu: thanks for the info, that branch is based on next-20141125 so I'll test it on linux-next then
<javier__>
zenmetsu: fortunately you haven't seen any soldering I did :)
* javier__
still has to solder serial on his peach pi
<zenmetsu>
there looks to be enough room in there to stick a level shifter. i'll probably give mine a 3.3V serial port broken out to the case. female headers are a great idea though, low profile and relatively high durability
<zenmetsu>
i'll probably opt for using some of my wire-wrap wire for this job, those pads look pretty small.
<javier__>
amitk: I confirmed that latest sin/ceti image does not bring display on peach pi, I'll take a look, thanks a lot for reporting!
<zenmetsu>
btw sjoerd, i ran into an issue compiling from that branch. getting cat:/drivers/gpu/midgard/platform//modules.order: input file is output file maybe my build environment is hosed o_O
<amitk>
javier__: thanks
amitk has quit [Ping timeout: 250 seconds]
amitk has joined #linux-exynos
amitk has quit [Ping timeout: 240 seconds]
amitk has joined #linux-exynos
amitk has quit [Ping timeout: 240 seconds]
amitk has joined #linux-exynos
<ukki>
amitk: any chance your branch would work with arndale octa?
<amitk>
ukki: it should, Kevin does some basic testing on arndale octa I believe
<amitk>
but I don't
<sjoerd>
oh that's curious
<sjoerd>
my kernel build does boot via tftp with a patched upstream u-boot
<ukki>
amitk: thanks
<ukki>
does Kevin use irc? would like a chat with him
<sjoerd>
ah that' sa fluke as newer u-boot does seem to load the initramfs correctly
<javier__>
ukki: yes, he is nickname is khilman and he is in #armlinux and #linaro-kernel at least
<ukki>
javier__: thanks
<zenmetsu>
sjoerd: disregard my previous comment. I double-noobed. First, i somehow configured the mali bits as built-in, not as modules. After resolving that, i ran into another error because i didn't have CONFIG_KDS set
<sjoerd>
yeah those patches relaly aren't pretty
<zenmetsu>
looks like it is compiling though, i'll know in a moment if it boots
<zenmetsu>
should teach me to not try to compile the kernel, dtbs, and modules on one shot; missed the fact that zImage failed to compile
<zenmetsu>
wondering if my .config is invalid
<javier__>
zenmetsu: can you post a complete log of your build failure and what config are you using?
<zenmetsu>
sure thing, i'm rebuilding after a make clean, and with -j1 this time. It was complaining about a truncated drivers/gpu/arm/midgard/built-in.o could be sd-card wonkiness, if it fails again, i'll post the code snippet online
<zenmetsu>
not getting much for build errors since it is failing with LD. didn't see any failures prior to that.
<zenmetsu>
going to have to build again and pipe the output to a build log to make certain
<javier__>
zenmetsu: are you using ccache? I've only seen similar errors when my ccache got confused and I had to flush the cache before building a new branch
<zenmetsu>
nah, i don't even have ccache installed on this box
<zenmetsu>
building natively though, and i did verify that i'm not using distcc either
amitk has quit [Ping timeout: 258 seconds]
amitk has joined #linux-exynos
<sjoerd>
mdrjr: If i want to send you an spl to sign to become bl2 what build target do you need ? u-boot-spl.bin or smdk5422-spl.bin
<amitk>
sjoerd: I just rebased and force-pushed on top of that branch, can you try again?
<amitk>
zenmetsu: are you saying that mali needs to always be compiled as a module (sorry, mali noobie here)
<sjoerd>
amitk: no my branch has some hacks to maek ti work one way or the other (i can't remember which way round)
<mdrjr>
sjoerd: smk5422-spl.bin :) please for reference use the odroid_defconfig on our u-boot
<sjoerd>
mdrjr: ok
<sjoerd>
comparing with the X2 u-boot looks like the u-boot size got load from 512K to 328K for the XU3
<sjoerd>
mdrjr: just wondering , was tehre a reason for that or just happened to be the size of your u-boot?
<mdrjr>
sjoerd: not my u-boot, I don't own hardkernel. And I have no idea why.. atm that XU3 was on bring up I was on another project.. so I didn't took part on that
<sjoerd>
Ok
<sjoerd>
mdrjr: Well you're the only one from hardkernle i know, so you get to be the target of my questions :)
<sjoerd>
mdrjr: hmm, silly question, looking at the u-boot code from hardkerenl, it looks like MOVI_UBOOT_POS only gets used when booting from SD card not when booting from EMMC
<mdrjr>
is CONFIG_SMC_CMD enabled? (I think it should be)
<sjoerd>
It is
<mdrjr>
in that case 128 will never the executed..
<sjoerd>
hmm?
<mdrjr>
the code clock on coldboot()
<mdrjr>
has a #ifdef
<sjoerd>
the L128 link was from the otehr branch
<mdrjr>
ahhh
<mdrjr>
lol.. my bad :(
<mdrjr>
I just saw the L128 and looked at line 128
<mdrjr>
:P
<sjoerd>
:)
afaerber__ has joined #linux-exynos
<sjoerd>
How is the SMC programmed? It looks like in case of EMMC boot both for odroid X2 and XU3 you only give it a size & destination but not the source of u-boot
<sjoerd>
ditto for TZSW
afaerber_ has quit [Ping timeout: 244 seconds]
<mdrjr>
that requires a quick check on the manual.. hehe
<sjoerd>
the manuals i have for exynos don't seem to document the smc
<sjoerd>
or maybe i'm just not looking at the right chapters
<mdrjr>
sjoerd: look at the trustzone mark
<sjoerd>
right but i'm failing to see any documentation abou tthe secure firmware that's running
<sjoerd>
maybe that's just not in my versions of the manuals though
<sjoerd>
mdrjr: Anyway, do the manuals you have explain the behaviour for loading u-boot from emmc on the XU3 ?
<mdrjr>
sort of.. hehe
<sjoerd>
do tell
<mdrjr>
sjoerd: gimme a second let me check on our internal svn :P
<sjoerd>
Thanks!
afaerber__ has quit [Quit: Verlassend]
afaerber has joined #linux-exynos
afaerber has quit [Quit: Verlassend]
liquidAcid has joined #linux-exynos
afaerber has joined #linux-exynos
amitk has quit [Quit: leaving]
liquidAcid has quit [Quit: Leaving]
<zenmetsu>
sjoerd, i threw the source at a different compiler, same result. i think my .config for my chromebook2 is the problem. compiles fine if i leave the mali bits out