amitk has quit [Read error: Connection reset by peer]
amitk has joined #linux-exynos
<James_T>
Does xen work on the Odroid-XU3's?
aballier has joined #linux-exynos
aballier has joined #linux-exynos
dlezcano has joined #linux-exynos
<lancelot>
DRM and FB are configured, I just checked what I did on pastebin. /dev/dri/card0 does not exist. [ 25.604] (II) No BusID or DriverName specified - opening /dev/dri/card0
<Wizzup>
By 'configured' you mean turned on in the config?
afaerber has quit [Quit: Verlassend]
<lancelot>
Correct, in the kernel .config
afaerber has joined #linux-exynos
dlezcano has quit [Ping timeout: 265 seconds]
dlezcano has joined #linux-exynos
afaerber has quit [Quit: Verlassend]
liquidAcid has joined #linux-exynos
amitk has quit [Quit: leaving]
zenmetsu has joined #linux-exynos
<zenmetsu>
Wizzup: are you around by any chance?
<Wizzup>
yes
<zenmetsu>
Wizzup: you're running Gentoo on a CB2 as well, right?
<Wizzup>
yep!
<zenmetsu>
any luck with getting sound working?
<Wizzup>
yep.
<Wizzup>
What does not work for you?
<zenmetsu>
i know wifi is still not working, but i have a nano N-dongle for that
<Wizzup>
wifi works.
<Wizzup>
I mean, depending on the kernel.
<Wizzup>
What kernel do you use? chromeos? mainline?
<zenmetsu>
god no
<Wizzup>
what?
<zenmetsu>
i'm trying to get a newer one working so i can have the 8-core global scheduler and mali... mali has been kinda hit and miss though with the 3.18+ kernel sources that i have tried thus far
<Wizzup>
so close to mainline kernel
<Wizzup>
then I am not sure about sound.
<Wizzup>
I run chromeos kernel on my cb2 atm
<Wizzup>
Because I need at least one work machine
<Wizzup>
I have mainline+patches on the cb1 atm
<zenmetsu>
gotcha. i'll have to ping someone that is maintaining a 3.18 repo... sjoerd or maybe javier
<zenmetsu>
i had a significant speedup when using all cores for compiling, enough that i almost switched over to my CB2 from my Macbook :)
<Wizzup>
what config option was that?
<Wizzup>
the one to turn on heterogenous mp
libv has quit [Ping timeout: 240 seconds]
<zenmetsu>
had to disable b.L switcher in the newer kernels, then it is automatic i believe
<sjoerd>
zenmetsu: You don't really get HMP scheduling on linux mainline yet though
libv has joined #linux-exynos
<si1v3r>
zenmetsu: blswitcher?
<si1v3r>
I've got a blank screen on mainline 3.19 boot, that sounds like it.
<sjoerd>
zenmetsu: It will just assume all cores are equal, so your performance depends in part on which core things happen to end up on
<zenmetsu>
sjoerd: really? well something bumped my compile time performance by about 12% :/
<Wizzup>
only 12%? with four more coures?
<Wizzup>
zenmetsu: ack @ switcher
<sjoerd>
zenmetsu: well if you do use the full 8 cores and make -j8 then things probably work out indeed
<sjoerd>
zenmetsu: See also Kevins main on the samsung kernel list
<zenmetsu>
sjoerd: yes, that is my understanding. chromeos kernel uses b.L switcher and it will end up running on all 4 big cores, but -j8 with the switcher disabled, as unintelligent as the scheduling may be in mainline at the moment, still ends up throwing jobs at all 8 cores so it helps
<si1v3r>
Ah, bigLITTLE. Thought you were talking about lcd backlighting. :/
<sjoerd>
zenmetsu: indeed
<zenmetsu>
sjoerd: i'll give kevin's a try. i am still using yours i believe, but didn't get mali working correctly and no sound, but other than that it has been pretty solid
afaerber has joined #linux-exynos
<sjoerd>
Not sure which you you're referring to as mine
<sjoerd>
Kevin pulled in more stuff for chromebooks
<sjoerd>
my tree was mostly geared towards XU3 bits we needed for a demo
<zenmetsu>
you linked me to a repo a while back
<sjoerd>
Ah right
<zenmetsu>
XU3 is pretty much CB2
<sjoerd>
It's the same SoC as the 13" CB2
<sjoerd>
so yeah pretty similar
<zenmetsu>
except the DT is different
<sjoerd>
Well it's a quite different board
<zenmetsu>
apparently the 3.10 kernels for the XU3 run pretty well, but i don't think i'll have any luck back-porting the peach-pi DTS :/
<sjoerd>
The kernels from hardkernel you mean?
<zenmetsu>
you mean, the Linaro kernels ;)
<sjoerd>
eh no
<zenmetsu>
i thought they just used the Linaro kernels
<sjoerd>
No
<sjoerd>
Looks like it's based of some samsung kernel tree
<zenmetsu>
ahh, then i was mistaken
<sjoerd>
and further hacked on by the guys from hardkernel
<zenmetsu>
hilman's 3.19-integ branch looks to have mali code in it
<zenmetsu>
i might give that a stab
<sjoerd>
I just care about mainline though and getting as much as possible there
<zenmetsu>
looks like he's got a bunch of additions to the DT also... so maybe a viable mainline kernel is just around the corner :)
<sjoerd>
Getting closer
<sjoerd>
there is quite a lot of outstanding stuff still
<sjoerd>
e.g. cpu frequency support, dvs, better support for the display panels, iommu support etc
libv_ has joined #linux-exynos
<zenmetsu>
oh, what is wrong with iommu? I thought that it was working... frequency support would be nice for on-battery, but not a gamebreaker for my intended use
<zenmetsu>
certain peripherals not accessible still?
libv has quit [Ping timeout: 240 seconds]
<sjoerd>
iommu doesn't properly work at all, not for the MFC nor for the DRM drivers
<sjoerd>
Marek has a ~20 patch branch outstanding to sort that out
<zenmetsu>
ah... i'll look. i'm still trying to get my hands dirty with this sort of thing. most of my work is in the ARM microcontroller realm, but i'm starting to get into the linux bits when i have free time
<sjoerd>
My branch does have some iommu work btw, but that's not the stuff that will be mainline
<zenmetsu>
sjoerd: i could try your branch again, maybe it has been updated since i last pulled down the source. which repo should i try for the CB2?