<tgall_foo>
Wizzup, so like I was saying I've got the r3p0 kernel and user space mali driver running on octa
<tgall_foo>
Wizzup, performance is less than great tho as compared to the arndale
<Wizzup>
hiya:)
<Wizzup>
the octa as in chromebook, right?
<tgall_foo>
Wizzup, yeah the chromebook2 essentially I believe
<tgall_foo>
the arndale octa has an earlier exynos 5 tho as compared to the chromebook2
<tgall_foo>
both have mali 628 MP6 tho
<Wizzup>
Awesome.
<tgall_foo>
Wizzup, FWIW I'm just getting my gentoo dev credentials back and intended on pushing all this into gentoo
<Wizzup>
So you have some u-boot and some kernel working on the chromebook2?
<Wizzup>
If you could send me some doc I'll try to do the same and document it on the wiki :)
<Wizzup>
yay - gentoo dev, even more awesome
<tfiga>
I assume it's all about vendor kernel?
<tgall_foo>
yeah i haven't done much with the chromebook2 yet. I put together a linux install on an SD card but haven't been able to get a single message out of it so not sure what I've messed up
<tgall_foo>
tfiga, in an ideal universe I'd very much like to see the upstream kernels get into shape
<tfiga>
tgall_foo: well, the biggest problem would be probably getting blob drivers running with upstream kernels
<tfiga>
although on x86 example we can see that it can be done
<tgall_foo>
tfiga, it can be done. yeah we've gotten on arndale things to work well with 3.10 stable series via the LSK IE the linaro stable kernel
* WarheadsSE
listens in the background
<tgall_foo>
tfiga, still the mali driver really needs work to get it into shape
<tfiga>
hmm, most of platform-independent code should just work
<tfiga>
the biggest issue IMHO would be the platform-specific glue
<tgall_foo>
"should"
<tfiga>
although with most platforms already using generic frameworks in mainline it might be even possible to create a generic one
<tfiga>
I mean DVFS and so
<tfiga>
we have something like this for Exynos 4 and Mali T400
<Wizzup>
tgall_foo: wrt "Haven't been able to get a single message out of it", I assume you mean that you didn't get the chromebook2 to boot well yet?
<Wizzup>
(that's where I am stuck as well at the moment)
<tfiga>
device instantiation from DT, power control using runtime PM, clock scaling using clock API and voltage scaling using regulator API
<tgall_foo>
Wizzup, exactly .. I've only spent like an afternoon on it. have had some other projects demanding my time
<tfiga>
that should work for most platforms
<Wizzup>
Wrt upstream kernels, I believe exynos already has a lot of drivers in place, for example for the samsung chromebook 1, there's a lot of drivers but we didn't get the DT to work that well ye
<Wizzup>
yet*
<Wizzup>
sound only worked on and off, same for wifi, probably had something to do with the power regulator missing
* Wizzup
pokes swabbles
<tfiga>
Wizzup: the problem of wifi in upstream is that there is no good way to power it up, so that it can be detected by the SDIO controller
<tfiga>
I assume chromebooks use the same design of semi-enumerable WLAN chips
<Wizzup>
tgall_foo: I see, I thought that may have been solved recently because there were some power regulators added recently
<Wizzup>
but I .. mostly don't know I think :)
<tfiga>
there is some work going on to create some framework for this, but in our internal tree we ended up with just another platform device controlling the power
* tgall_foo
catches up
<tgall_foo>
Wizzup, where you getting further with your kernel ?
<Wizzup>
tgall_foo: no, not yet, right now I'm not even in my homeland, but I will be tomorrow evening
<tgall_foo>
ah ok
<Wizzup>
tgall_foo: so what did you mean with 'I have a r3p0 kernel running on the octa core' -> what octa core were you talking about then?
<Wizzup>
Or did you mean you had it working on chromeos?
<tgall_foo>
Wizzup, r3p0 == Mali device driver version
<tgall_foo>
for Octa that's the Arndale Octa board