<mnemoc>
cat_x301: yes :( please disable that driver
<mnemoc>
cat_x301: if you have ethernet (mele or cubie) change the initial state of the usbc to enabled in script.bin
<cat_x301>
mnemoc: unf. nope, i only have mk802.
<cat_x301>
so perhaps i need this one
<mnemoc>
cool. then just disable that broken driver
<cat_x301>
mnemoc: but i wont have connectivity then
<mnemoc>
cat_x301: you have wifi
<mnemoc>
err
<mnemoc>
i got confused sorry
<mnemoc>
forget what I sad about ethernet
<mnemoc>
disable that driver and enable rtlwifi's
<cat_x301>
mnemoc: forgotten :)
<cat_x301>
mnemoc: ah, ok
<mnemoc>
then edit your script.bin to power that usbc since the begining
<mnemoc>
the allwinnerized driver adds poor-man power management by powering the usb controller on driver load and shut it down on unload
tender86 has joined #arm-netbook
<mnemoc>
but if you just set the usbc to powered in script.bin the mainline driver will deal with the wifi chip without troubles
<cat_x301>
mnemoc: thanks, will try
<mnemoc>
now time to brew some more coffee to wake up my brain...
* cat_x301
thinks of making dedicated defconfig for mk802.. now two things should be handled there..
<mnemoc>
this driver thing is generic, not mk802 specific
<cat_x301>
mnemoc: i had mine already: but brain is still kinda alseep..
<mnemoc>
:)
<lundman>
hmm
<cat_x301>
mnemoc: but i mean it was enabled by default in sun4i_defconfig
<cat_x301>
which does not work for mk802
<cat_x301>
another thing is that DVFS part.
<mnemoc>
ihmo the dvfs issue must be fixed in code, not adding a mostly redundant defconfig...
<cat_x301>
mnemoc: right. what about wifi stuff?
<mnemoc>
I think rtl8192cu should be removed... and plat-sunxi and usb-sunxi improved to deal with this config option behind the scene
<mnemoc>
scenes*
<cat_x301>
mnemoc: fair. anyway, will workaround by now. btw, what is the policy for sending patches should i come up with anything?
* cat_x301
apologizes for distracting mnemoc from coffee..
<mnemoc>
cat_x301: currently we use github's issue tracker... but it might be a good idea to create a real mailing list if the rest of the community approves
<cat_x301>
mnemoc: would be brilliant to have one.
arete74 has quit [Ping timeout: 244 seconds]
arete74 has joined #arm-netbook
<mnemoc>
sad groups.google.com doesn't let you create lists with custom domains in the free version :<
<mnemoc>
but I suppose we can just install mailman in the same server that is hosting the wiki
<mnemoc>
does vger.kernel.org accept not-mainlined projects?
<cat_x301>
mnemoc: no idea
<mnemoc>
cat_x301: as it was your idea, can you file a ticket requesting a mailing list and maybe a patch review thingy like gerrit? :)
<cat_x301>
mnemoc: need to learn something before doing this :)
<mnemoc>
how to register on github? :p
* cat_x301
is afraid of screwing things up publicly :)
<cat_x301>
ah, i was thinking you ask about vger.
* cat_x301
is looking for that ticket creator option on github..
<cat_x301>
mnemoc: any hints how do i create a ticket, pls :)
sspiff has quit [Remote host closed the connection]
<cat_x301>
mnemoc: ah, i thought of some github specific ticket system :) -- working for corporates eroded my brains..
<rz2k>
<hno> and how to make a Linux livesuit image is explained in the SDK documentation. -- what SDK you are talking about? where I can found it? I thought about using those windows unpack/pack and imgpacker tools for livesuit images. There is another way?
madmalkav has quit [Quit: Leaving]
sspiff has joined #arm-netbook
sspiff has quit [Changing host]
sspiff has joined #arm-netbook
Fremen2001 has joined #arm-netbook
tzafrir has joined #arm-netbook
<mnemoc>
rz2k: hno pushed a copy of the a10 sdk to his github long ago
<mnemoc>
rz2k: and olimex pushed the a13 sdk somewhere too
sspiff has quit [Ping timeout: 260 seconds]
<cat_x301>
mnemoc: stupid question i know, but how to reply to particular comment in github?
<mnemoc>
cat_x301: by mail... but they end up at the tail of the forum-ish list of comments anyway
<cat_x301>
Turl: it hits this in my case, drivers/video/sunxi/disp/dev_fb.c:
<cat_x301>
else if(need_open_hdmi == 0 && from == 1)
<cat_x301>
then it returns w/o initialization.
<cat_x301>
have to go now..
<Turl>
(I meant the mali build issue)
<cat_x301>
Turl: ah. ok :)
yaayaa has joined #arm-netbook
cat_x301 is now known as cat_away
Simon_G has left #arm-netbook [#arm-netbook]
cat_n9 has quit [Ping timeout: 255 seconds]
Fremen2001 has joined #arm-netbook
yaayaa has quit [Quit: WeeChat 0.3.8]
jlj has joined #arm-netbook
Almamuetya has joined #arm-netbook
cat_n9 has joined #arm-netbook
grevaillot has joined #arm-netbook
cat_away is now known as cat_x301
ssalevan has joined #arm-netbook
cat_n9 has quit [Ping timeout: 264 seconds]
ssalevan has quit [Quit: Leaving.]
kaspter has quit [Ping timeout: 240 seconds]
cat_n9 has joined #arm-netbook
Fremen2001 has quit [Remote host closed the connection]
hipboi has joined #arm-netbook
<hipboi>
hno: ping
<hipboi>
sun5i u-boot-mmc
<hipboi>
> (u-boot boots ok from sdcard, but kernel dies at any printk)
<hipboi>
is this the clock related?
itamarjp has joined #arm-netbook
itamarjp has quit [Changing host]
itamarjp has joined #arm-netbook
revident has joined #arm-netbook
Fremen2001 has joined #arm-netbook
hno has quit [Excess Flood]
hno has joined #arm-netbook
kaspter has joined #arm-netbook
rellla has quit [Quit: rellla]
ssalevan has joined #arm-netbook
<oliv3r>
lkcl: I think we missunderstood eachother, I wasn't speaking of the mali GPU, but specifically the CedarX VPU. Is there any talks/plans in getting proper open source support for this bit aswell? I know there's some reverse engineering going on and in time, that could be fruitfull, but until then, anything in the open pipes?
polto2010 has joined #arm-netbook
RITRedbeard has joined #arm-netbook
<merbanan>
oliv3r: I looked though the libve code and it is based on alot of reference software, I don't see it likely that it will be released
<merbanan>
what would be possible is reimplementation of it
sspiff has quit [Remote host closed the connection]
<merbanan>
IIRC there was code for mjpeg
<merbanan>
reimplementing that looks like a doable task
<merbanan>
but for vc1 and h264 there was quite a lot of code
<merbanan>
if I understand correctly the biggest complaint from gimli was that the time for decoding some frames where to slow
<merbanan>
and I can understand the reason for the slowness if it is as a lot of reference code that parses the codec data before passing it of to the hardware for decode
<merbanan>
reference code is nearly never optimized
<andoma>
merbanan: you're lurking here as well? :)
polto2010 has quit [Quit: Computer has gone to sleep.]
<lkcl>
oliv3r: no, i wasn't misunderstanding at all. i will only speak to allwinner once there is a financial incentive - a contract that is significant to them that would be lost if they *didn't* provide the full source code of libcedarx.
<lkcl>
in the mean time i hope that someone continues with the reverse-engineering to the point where that provides an alternative avenue and additional leverage
<merbanan>
lkcl: is libcedarx != libve ?
<traeak>
lkcl: so when's the mips multicore coming out?
<traeak>
heh
<lkcl>
merbanan: no idea!
<lkcl>
1sec...
<lkcl>
traeak: MIPS doesn't do actual CPUs, they just do designs (like ARM). are you referring to ICubeCorp?
<lkcl>
their CPU design is completely different
<traeak>
lkcl: yes assuing they are using a mips core?
<traeak>
s/assuing/assuming/
<ibot>
traeak meant: lkcl: yes assuming they are using a mips core?
<lkcl>
traeak: absolutely not. it's a completely independent architecture, waaayy more advanced than anything that MIPS have ever dreamed of.
<andoma>
merbanan: mk802
<lkcl>
it's a VLIW architecture for a start.
<hno>
hipboi, sun5i branch should boot fine.
<lkcl>
merbanan: i'm not able to find anything about what libve is - do you have any links/references?
<lkcl>
where you share a portion of a wafer's cost with other people. you usually have to wait for a 3-month run
<lkcl>
the costs of doing so however are far far less than if you get a fab to run an entire wafer just for you.
<merbanan>
andoma: we need working hw decode
_jm has joined #arm-netbook
<lkcl>
so the last time ICubeCorp did this, they were able to get a number of working 65nm 700mhz Dual-Core processors, codenamed "IC1".
<merbanan>
andoma: I need to come by with your stuff some day also
<lkcl>
700mhz really isn't fast enough, even though the VLIW and the parallelism ramps the actual performance up considerably, it still isn't enough in today's market.
<andoma>
merbanan: yes.. i'm in NYC now though
<andoma>
back late next week
<merbanan>
ok
<lkcl>
so, it'll be at least 18 months before they come out with something (40nm, 28nm)
<traeak>
vliw brings up the itanium
<hno>
had that printk issue during early development.
<traeak>
itanium pretty expensive, and isn't the code density sort of poor on vliw ?
<specing>
vliw brings up the 100 core tilera
<specing>
vapourware.
<traeak>
there's a korean company with an interesting instruction set that shows exceptionally good density and good extendability
<traeak>
but no clue...anyways all vapor and speculation
<specing>
yep
* specing
goes out to the store to buy another x86 :(
<andoma>
merbanan: but i would like to run it without X so i guess i'll have to wait/work with the lima driver
<hipboi>
hno: i beleive it's the freqence of the cpu at u-boot stage
<traeak>
lkcl: so 18 months before something comes out, there needs to be stuff like compilers, etc written in addition to OS stuff as well. fun fun, so 18 months maybe to even start anything
_jm has quit [Quit: Loqui powered]
rellla has joined #arm-netbook
kaspter has quit [Ping timeout: 276 seconds]
<hno>
hipboi, yes, quite likely due to the core voltage issue / frequency.
<hno>
hipboi, On my A13 board with the core running under voltage the CPU hang when trying to jump to printk, sometimes randomly earlier.