nighty has quit [Quit: Disappears in a puff of smoke]
ganbold has quit [Quit: This computer has gone to sleep]
nighty has joined #linux-rockchip
nighty has quit [Quit: Disappears in a puff of smoke]
cnxsoft has joined #linux-rockchip
dlezcano has quit [Ping timeout: 255 seconds]
cnxsoft has quit [Remote host closed the connection]
cnxsoft has joined #linux-rockchip
ganbold has joined #linux-rockchip
cnxsoft has quit [Ping timeout: 265 seconds]
sunilmohan has quit [Quit: No Ping reply in 180 seconds.]
ganbold has quit [Quit: This computer has gone to sleep]
ganbold has joined #linux-rockchip
ganbold has quit [Client Quit]
sunilmohan has joined #linux-rockchip
ganbold has joined #linux-rockchip
cnxsoft has joined #linux-rockchip
dlezcano has joined #linux-rockchip
shahid__ has joined #linux-rockchip
wadim_ has joined #linux-rockchip
ganbold has quit [Remote host closed the connection]
ganbold has joined #linux-rockchip
zzarr has joined #linux-rockchip
zzarr has quit [Quit: Lämnar]
zzarr has joined #linux-rockchip
<zzarr>
hello! I have a Firefly RK3288 Plus and need some guidence
<zzarr>
I installed a bootloader (u-boot), installed the kernel.img (linux-3.14 I built specificaly for RK3288), the boot.img and the Ubuntu filesystem according to the guide
<stdint>
zzarr, why do you use 3.14 ?
dlezcano has quit [Ping timeout: 252 seconds]
<zzarr>
stdint, it's the kernel supplied by the wendor
<stdint>
zzarr, but if you are not intend to use android
<mmind00>
stdint: probably based on the chromeos kernel, which right now does provide the highest level of general support for the rk3288 (we're still missing your vpu-project as well as deep suspend, ddr frequency scaling and probably some other small things)
<zzarr>
I just want to try to get the board up and running first, then iron out what kernel and other stuff
<stdint>
then you could try the 4.4
<zzarr>
I think the Ubuntu graphics driver is using libhybris
<zzarr>
would 4.4.1 (which is an LTS if I'm not mistaken) work?
<stdint>
then you could try the 4.4, the mali library for graphics is designed for it
<jacob_>
Kernel gpu driver can get from 4.4 kernel in github
<zzarr>
okey, nice
<zzarr>
jacob_, do you know if it's possible to run Mir on it?
<libv>
jacob_: wikidot? really?
<libv>
ah, that's what rockchip did.
<libv>
that's a pretty useless and assholish move.
bbelos has quit [Ping timeout: 250 seconds]
<libv>
and rockchip should be seriously trashed for it.
<zzarr>
libv, what do you mean? (I would like to know what they have done)
<libv>
zzarr: created their own wiki
<libv>
instead of just creating accounts on the wiki mentioned in the topic here
<zzarr>
libv, okey, then I agree
<zzarr>
I flashed a kernel (4.4) to my Firefly RK3288 Plus board as well as a Ubuntu file system
<zzarr>
I did it on a remote connection from my work to my home computer and I forgot to change the network settings
<zzarr>
I rebooted the board, so I can't reach the board no more
<zzarr>
but I can see that it got an IP address and running nmap <IP> -Pn reveals that it is up
<zzarr>
does that mean that the kernel has booted?
<jacob_>
zzarr, i think can't support Mir
<jacob_>
haha, libv, don't be too strict with commercial company.
<jacob_>
They might just need a wiki for pushing document for their customer.
<libv>
jacob_: i have dealt with too many commercial companies in my lifetime
<libv>
google me.
<jacob_>
At leaset it's better than pdf which you can just get them with nda.
<libv>
now you definitely need to google me.
<jacob_>
cool, libv...... :-)
<jacob_>
zzarr, I think it has booted since it got ip adress.
<zzarr>
jacob_, that's nice, next time I have physical access to the board I will make sure it have a correct network interface settup
RayFlower has joined #linux-rockchip
<zzarr>
jacob_, is there a way to ask Rockchip if they will provide a Mir driver or make the user space open source?
<libv>
so kindly ask rockchip to make the mali userspace driver open source?
<libv>
<arrow-up><arrow-up><enter>
<zzarr>
libv, what?
<libv>
zzarr: i am the guy who started the lima driver back in 2011
<zzarr>
ohh.... I did not know that, cool
<zzarr>
how is that project comming along?
<libv>
i am at a big automotive supplier, doing useless and pointless work, so that i can have a roof over my head and finally do the things a 37y old should be doing
nighty has joined #linux-rockchip
<libv>
as i of course got myself blackballed with yet another rist of supposed linux companies.
<zzarr>
libv, sorry to hear that, I'm a developer
<zzarr>
I'm 33years old
<zzarr>
does the lima driver project have support for mali 764?
<zzarr>
(the one used by RK3288)
<libv>
i showed an early t604 based demo at fosdem 2015
<libv>
then i got crap for having done that
<libv>
now i am finally making some money and becoming a father, and thinking about normal human things like getting a car, etc
<libv>
at least the corporate fun and games i now have to deal with result in me not being dirtpoor.
<libv>
which is a massive improvement over me having fought bioses for display drivers (modesetting), freed ATI, and setting free ARM GPUs, which were all clearly such evil things to do that i should never be allowed to work in open source software.
<zzarr>
omg, I found your work on github and think it's amazing
<libv>
anyway, i also blogged about wayland vs mir in 2013
<zzarr>
don't be to hard on your self, people allways want more then they can get
<ssvb>
jacob_: http://rockchip.wikidot.com/xserver <- any idea why they are comparing "opaque stippled rectangle" vs. "tiled rectangle"? would it be way too difficult to benchmark exactly the same thing with and without glamor?
<libv>
and stated that, ignoring the useless games being played inside intel (a supposed hardware developer) , i preferred wayland over mir simply because canonical ran straight to libhybris
<zzarr>
indeed they did libv, but it was a way to make porting Ubuntu to devices quick and easy
<libv>
which took some pressure away from the need of having a proper open source gpu driver
<jacob_>
ssvb, i thought they are just lazy to benchmark with and without glamor. :->
<zzarr>
I would not have a Meizu MX4 Ubuntu Edition if Mediatek would have had to develop a Mir driver
<libv>
zzarr: canonical dug its own hole there
<libv>
iirc, there is a wayland binary available for mali t series
<ssvb>
jacob_: either way, that's only 140 MPix/s with glamor, which is actually not very impressive
<libv>
it's not """needed""" for mir.
<libv>
stellar long term thinking there
<zzarr>
I think there will be in the future (that's there goal)
<ssvb>
jacob_: consuming merely around ~560 MB/s of memory bandwidth, which is way below the memory controller capabilities
<zzarr>
in any way, I think companies will develop Mir drivers once Ubuntu have made the move to Mir completely
<libv>
zzarr: it won't happen
<libv>
or it would've happened years ago.
<zzarr>
libv, the future will tell, we can only speculate
<jacob_>
ssvb, they might test it in a old chip and low ddr rate.
<jacob_>
They just want show that compare to cpu, time-consuming didn‘t have linear growth when enable glamor.
<jacob_>
not the rate
<zzarr>
in any way, even if libhybris is the way for Ubuntu on phones/tablets running on ARM SoC's I don't think it's a big deal
<ssvb>
jacob_: with such unimpressive results, the cpu is going to slaughter glamor
<jacob_>
ssvb: uh...I will ask rockchip members to do a serious test with and without glamor.
cosm has quit [Remote host closed the connection]
cosm has joined #linux-rockchip
<ssvb>
jacob_: actually I stand corrected, on 1.4 GHz Exynos4 devboard I get "10000 trep @ 3.1027 msec ( 322.0/sec): 500x500 tiled rectangle (4x4 tile)" which is a bit less than the 556.0/sec result from the rockchip glamor
<ssvb>
jacob_: but perf shows that the CPU spends time in fbBlt and fbOddTile functions in the X server, rather than calling a SIMD optimized pixman function for this operation, looks like there is a lot of room for improvement
<ssvb>
jacob_: hmm, I also get the same result for the "opaque stippled rectangle" benchmark on exynos4 - "10000 trep @ 3.0808 msec ( 325.0/sec): 500x500 opaque stippled rectangle (8x8 stipple)"
<ssvb>
jacob_: and the CPU time is spent in fbBltOne and fbOddStipple now
<jacob_>
interesting... I will do this test on rockchip platform tomorrow
<jacob_>
and show what i found
<ssvb>
thanks, these benchmarks can be run as "DISPLAY=:0 x11perf -tilerect500" and "DISPLAY=:0 x11perf -osrect500"
<zzarr>
libv, with the help of a patch any Gallium based MESA driver could run Mir
nighty has quit [Quit: Disappears in a puff of smoke]
cnxsoft has quit [Quit: cnxsoft]
wens_ has joined #linux-rockchip
wens has quit [*.net *.split]
indy has quit [*.net *.split]
JohnDoe_71Rus has joined #linux-rockchip
chris2 has quit [Ping timeout: 265 seconds]
indy has joined #linux-rockchip
afaerber has quit [Quit: Ex-Chat]
vagrantc has joined #linux-rockchip
afaerber has joined #linux-rockchip
jacob_ has quit [Ping timeout: 264 seconds]
afaerber has quit [Ping timeout: 250 seconds]
afaerber has joined #linux-rockchip
wadim_ has quit [Read error: Connection reset by peer]
paulk-aldrin has joined #linux-rockchip
paulk-aldrin has quit [Remote host closed the connection]
paulk-aldrin has joined #linux-rockchip
JohnDoe8 has joined #linux-rockchip
JohnDoe_71Rus has quit [Ping timeout: 265 seconds]
<Fighter19>
Hey there. I'm wondering if somebody here knows anything about the kpanic partition on my RK31 device. I tried to create a custom rom (Kernel and Recovery definetly works), but when booting it gets stuck on the splash (I wonder is that one stored inside the kernel, because the splash is bigger than the stock kernel?). So my guess was that it couldn't mount the system partition.
erc has joined #linux-rockchip
paulk-aldrin has quit [Quit: Leaving]
<Fighter19>
I think I'm doing something wrong. I thought I could flash the recovery image to the boot partition. In the recovery partition it works fine, when loaded to boot, my device freezes.
<Fighter19>
the recovery image contains a kernel and a initramfs (both packed up with mkbootimg)
cbncvbc has joined #linux-rockchip
<cbncvbc>
test
<Fighter19>
hey there
cbncvbc has quit [Client Quit]
<Fighter19>
My partition strangely looks like this (stock): kernel (KRNL...) boot (ANDROID!)
<Fighter19>
*the recovery image contains a ramdisk.
<Fighter19>
unmkbootimg was also able to get a Kernel and a Ramdisk out of boot.
<Fighter19>
nvm I'm just stupid. (I have no clue, I just reflashed it and suddenly it worked)