<naobsd>
I didn't & will not try to access rockchip's git repo
<naobsd>
fonix232: where is RKDeviceManager you mentioned? I didn't see everything yet
<naobsd>
I know "RKDeviceManage.exe" is already available for some HDMI stick
<naobsd>
but I'm not sure it's same as "RKDEviceManager" you mentioned
<naobsd>
well, I have to go
<naobsd>
bye
naobsd has quit [Quit: Page closed]
_whitelogger has joined #linux-rockchip
_whitelogger has joined #linux-rockchip
cnxsoft has joined #linux-rockchip
hipboi has joined #linux-rockchip
hipboi has quit [Ping timeout: 246 seconds]
hipboi has joined #linux-rockchip
akaizen has quit [Remote host closed the connection]
akaizen has joined #linux-rockchip
Astralix1 has joined #linux-rockchip
Astralix has quit [Ping timeout: 264 seconds]
Theueip has quit [Quit: Leaving...]
cnxsoft has quit [Ping timeout: 264 seconds]
hramrach has quit [Ping timeout: 240 seconds]
hramrach has joined #linux-rockchip
cnxsoft has joined #linux-rockchip
rellla has joined #linux-rockchip
Astralix has joined #linux-rockchip
Topgun100 has quit [Quit: Topgun100]
fonix232 has joined #linux-rockchip
<fonix232>
mornin'
hramrach has quit [Remote host closed the connection]
hramrach has joined #linux-rockchip
Super-noob has joined #linux-rockchip
<Super-noob>
good morning
<fonix232>
I'm pushing the SDK to github right now
<fonix232>
but I'd suggest we make a clone of the whole AOSP repo, try to diff it to the SDK, and add the most necessary changes to the proper projects
<fonix232>
a tedious and long work, especially on this crappy laptop, but would give us a better overview of the stuff Rockchip changed, what's required, what's not required, and clean it up
<hipboi>
fonix232, good
<fonix232>
anyone up for the challenge of finding the closest match of AOSP vs. SDK?
<hipboi>
fonix232, i can help with it
<fonix232>
okay
<fonix232>
just finished counting the object, ugh
<fonix232>
I think this is going to be a looooooong day
akaizen has quit [Read error: Connection reset by peer]
<hipboi>
i think rockchips development is based on AOSP tag 4.2.2_r1
akaizen has joined #linux-rockchip
<fonix232>
the tag is okay, but they might have pulled in some further updates
<Astralix>
like they did in the kernel. It is still called 3.0.36+ but it is based on 3.0.66
<fonix232>
well it IS 3.0.36+ :D
<fonix232>
though I suppose it is mostly for compatibility
<Astralix>
nope, it is cause the proprietary rknand_ko.ko driver checks the kernel version and fails if it doesn't match. So if they would change numbering too often, they get messed up the nand flash driver.
<fonix232>
See, compatibility :D
<fonix232>
damn if only once they would make the mistake of releasing rknand sources...
<fonix232>
the whole device tree is a mess, using 50% proprietary libraries that could be built, or are built during compile
<jeremb>
it seems to be an internal forum for rockchip customers
Astralix has joined #linux-rockchip
cnxsoft has joined #linux-rockchip
<fonix232>
nice find
Astralix has quit [Quit: Leaving.]
Astralix has joined #linux-rockchip
naobsd has joined #linux-rockchip
Super-noob has quit [Ping timeout: 268 seconds]
<fonix232>
hey naobsd
<naobsd>
hi
<fonix232>
I have some good news for you
<fonix232>
yesterdays hackery on the Rockchip Redmine has some awesome results
<naobsd>
you used account information described in readme.txt?
<fonix232>
yes, but did some other things too
<naobsd>
I'm sorry, probably you have good news, but
<fonix232>
the main thing being, sending in an issue report
<fonix232>
requesting for better work with the community, etc.
<fonix232>
received an e-mail not long ago from a Rockchip employee to detail how exactly would we like to work with them
<naobsd>
in my country, using others account is illegal even if it's disclosed by any reason :(
<fonix232>
oh, okay
<fonix232>
not here
<fonix232>
well, point is, someone who matters got our message
tonikasch has joined #linux-rockchip
<fonix232>
just answered their e-mail
<fonix232>
where I detailed that an option would be an "Ambassador" program
<fonix232>
ambassadors being community members who did something huge, like you, would get access to all the internals of Rockchip
<fonix232>
including proprietary stuff (rknand, and others)
<naobsd>
well, RK guy promised they gives everything to us?
<fonix232>
no promises yet
<naobsd>
internal information is very nice
<fonix232>
but a select few, after signing an NDA, could work on currently safely locked down codes
<fonix232>
and many of the currently private code would go public, to, with the help of the community, be restructured, reorganized, optimized, and brought on par with AOSP standards
<fonix232>
making both hardware manufacturers and community developers work easier
<naobsd>
my small pieces of source code for RK were written by myself. progress was slow because of no information.
<naobsd>
but it's free, no NDA, no illegal material
<naobsd>
it's important for me
<naobsd>
I'm not an active developer
<fonix232>
yes, I know this
<naobsd>
my output is almost nothing
<naobsd>
maybe some other person will help you
<fonix232>
oh, I wouldn't say so, all those CM10 device trees
<fonix232>
generic, yes, but useful? definitely
<fonix232>
any way, I'm just saying that soon there might be more work on RK based devices than before
<naobsd>
it will be nice
<naobsd>
I hope so too.
<naobsd>
if NDA-ed information will be disclosed in this community
<naobsd>
I have to public my work-in-progress small codes related to my rkutils
<naobsd>
s/public/publish/
<naobsd>
(or doing my work outside of here)
<naobsd>
(it's very small code, probably no one care about it...)
<tonikasch>
I do care :)
<naobsd>
maybe it's good time to try mainline kernel :)
<tonikasch>
mainline kernel? Has rk3188 support been introduced yet there?
<naobsd>
tonikasch: some amount of code were committed, what I said is I want to try it :)
<tonikasch>
oh, ok :)
<tonikasch>
I though only bare support was present, I'm not following it live
<naobsd>
I already soldered wires to UART on my RK3188/RK3066 sticks :)
<Astralix>
It sounds nice, but we already learned that it isn't that easy with rockchip-
<tonikasch>
naobsd, do you know rk3188 diagram to uart pins?
<tonikasch>
well, lunch time for me, my family waits for me :P
<Astralix>
fonix232: if you could get the abassador state for some of us, it would be pretty nice and I surely will support
<Astralix>
tonikasch: we are working on mainline
<naobsd>
tonikasch: sorry, I only know about hotack's HTC-T010-V1/V2 boards
<naobsd>
it has 3 TPs with R T G print
<Astralix>
started with 3.10-rc7 and still proceeding. Just some slow progress due to vacation times
<fonix232>
guys, guys guys
<fonix232>
first of all
<fonix232>
NDAd information WON'T be published
<fonix232>
well, most of it
<Astralix>
sure
<fonix232>
but, some of us will have access to NDAd source code
<fonix232>
to develop it further
<naobsd>
yes please don't
<Astralix>
fonix, that will not work
<fonix232>
by my idea, it will be a simple git repo that you cannot release to the public
<Astralix>
cause if you do it like that it results in some people doing hard work for Rockchip without beeing payed
<fonix232>
but compiled binaries can be released
<fonix232>
Astralix, right now the same is going on
<Astralix>
so you develop closed source for closed circle of people
<fonix232>
except that the people who actually do the hard work will get credit
<Astralix>
no it isn't
<Astralix>
as I reverse engineer everything I can publish my code
<Astralix>
it was based on things I evaluated and I never signed any nda for
<fonix232>
yes, of course
<fonix232>
but a great example is how many manufacturers simply go on github, grab the patches, and claim the fixes to themselves
<Astralix>
so reverse engineering is going slowe and hard work but then it is fully open to all
<fonix232>
I am working with a bunch of asshats right now
<Astralix>
yes, I fully agree
<fonix232>
I do not want to push them too much initially
<fonix232>
for now, full access to their git repo is more than we can wager for
<Astralix>
but if I code in NDA code where I cannot make any firther use of than limited to the NDA, I would like to get 140€/h
<fonix232>
hmm, okay
<Astralix>
because I do it in my free time and RK is making money of it
<fonix232>
yes, of course
<fonix232>
I think we can discuss that parts of your own code can be published, but the whole thing, especially the rockchip-made code, should not be
rellla has quit [Ping timeout: 260 seconds]
<Astralix>
So they are grabbing our results, give us all information we need to be even more effective and then we may not puiblish it but they sell it in SDK 2.x
<fonix232>
yes, basically, that's the offer
<fonix232>
they give us access, properly, and we work with them to better the platform
<Astralix>
guess it would be much more easy if some of us geeks sit together and throw the bootloader out of the flash. Then port redboot or uboot
<fonix232>
hmm
<Astralix>
after that no rknand is needed anymore as there are good NAND chip drivers in the mainline kernel available
<fonix232>
might be
<fonix232>
but as far as I know, the current bootloader is a heavily modified u-boot
<naobsd>
oh
<Astralix>
not sure
<Astralix>
but we could check. I decoded it and I could disassemble it. So I could give it a try
<Astralix>
AFAIK booting from NAND or (e)MMC/SD should be no problem as aside from NAND-IP the eMMC and SD IPs in the SOC are standard IPs well known to the marked
<naobsd>
I'm not active, I don't know many things you may know, it will be better not to count me as "some of us" ;)
<Astralix>
just spreading my thoughts with the whole world.
<Astralix>
so if someone knows what I am talking about, just jump in and leave us a note
<naobsd>
(sorry, I talked about relationship with RK)
<Astralix>
I do not have any relationship with rk
<naobsd>
yes I know
<Astralix>
I had some emails to guys woking there but conversations always end when it starts to get interesting
<Astralix>
I guess the real geeks already have met. Even it looks like Omegamoon or crewrktablet are different crews working on different things, but in reality we are all sitting 'together' bundling forces and noone is hiding anything from the others. There is no competition, just co-working
<naobsd>
sounds good :)
<Astralix>
As a result some nicer 4.2.2 images could be ready in a few days, and the WiFi problems of picuntu are now under close inspection
<naobsd>
really good
<tonikasch>
I was talking to Omegamoon a while ago :)
<vinifr>
seems that Rockchip is like Allwinner
<tonikasch>
we'll see if Rockchip finally opens up like has done Allwinner...
<naobsd>
I'm not sure about Allwinner. all I know is hipboi is nice guy :)
<fonix232>
Allwinner really opened up, it was a nice move from their part
<naobsd>
I thought VPU is not open
<fonix232>
not everything is open
<fonix232>
but for example, their bootloader is, completely
<fonix232>
at least on A10
<fonix232>
haven't checked the further platforms
<fonix232>
also the whole sunxi plat and mach was merged into the mainstream kernel AFAIK
<vinifr>
but still lack good documentation :0
<fonix232>
good documentation <> chinese
<naobsd>
I think Allwinner mainlining is community's effort
<naobsd>
as like as what Astralix is doing...
tinti_ has joined #linux-rockchip
<naobsd>
(I have to take dinner, sorry)
Small-noob has joined #linux-rockchip
Small-noob has quit [Client Quit]
<fonix232>
no problem, I'm cooking right now :D
Super-noob has joined #linux-rockchip
<vinifr>
is there a 'Linux mainlining effort' for Rockchip?
<Super-noob>
i know it´s a bit off topic but, can the mk908 still only stereo with stockrom ?
<Super-noob>
was almost tempted to replace my 808b
<tonikasch>
I think some German guys are working on it
<tonikasch>
dunno, sorry
<tonikasch>
damn, I cannot access any post on that bbs.rockchip.com.cn site, insufficient rights or smoething alike
<Astralix>
we had to wait a bit for some other modules of the official kernel have riped and where accepted by linus. That happened now and we can continue.#
<Astralix>
I have a booting kernel 3.10-rc7 on RK3188 and mind00 has one on RK3066
rellla has joined #linux-rockchip
<tonikasch>
woooow, Astralix :)
<Astralix>
But that is still far away from solving all of our problems.
<Astralix>
For getting a stick or a tablet running with this kernel, you still need drivers for touches, sensors, battery and HDMI
<tonikasch>
oh, yes, of course...
<Astralix>
and all those drivers working and available as source have to be rewritten mostly. otherwise they will be thrown out by luns' team
<Astralix>
cause they where written the same way the ARM and core related code was. And we filtered and rewrote 100% of it
<Astralix>
(almost)
<tonikasch>
aps... :S
<Astralix>
The truth is, that the cpus are made like any other ARM too: Buy the core and buy peripherals for it, put all to a die and sell it.
<tonikasch>
ahm
Theueip has joined #linux-rockchip
<Astralix>
So most components are already available and code is available. Like SD, USB, DMA, SRAM and many many more. But you have to reverse engineer the code to find out registers and addresses. Then you have to look which other existing IP matches best and then you adopt the IP
<Astralix>
Unfortunately the important things for us are not that easy:
<Astralix>
LCDC is an open yard of work inside the kernel team. I cannot tell if it makes sense to add RK LCDC at the moment
<tonikasch>
mmmm
<Astralix>
HDMI of 3066 is a thing that incorporates encryption things. HDMI for 3188 requires external chips and therefore additional drivers
<Astralix>
And then the speed optimization
<Astralix>
we have only bad quality sources and no document about the CRU unit
<Astralix>
So setting up the system timers is difficult without a witch and a glass bowl
<Astralix>
actually our kernels run from the timers as they are set up by the bootloader
<Astralix>
So even if you are a lucky guy and have a copy of 3188 or 3066 TRM, it helps you almost nothing, as Clocks, Rests, Power Manager and many other components of the die are andled in separate documents which I never did see out in the wild
<tonikasch>
so when the bootloader will be replaced there will be new problems to solve?
<tonikasch>
ahms
<Astralix>
for replacing the bootloader you need to know how to initialize the chip first.
<fonix232>
Astralix, exactly what documents do you need?
naobsd has quit [Ping timeout: 250 seconds]
<Astralix>
I have headhunter list :)
<fonix232>
Redmine had a lot of technical data on it
<Astralix>
was just kidding. I know how fast cn domains are
<fonix232>
here are some screenshots and the zip containing documents
<fonix232>
actually, the redmine site is quite fast
<fonix232>
peak speed was around 1MB/s
<fonix232>
btw yes, toni's link is right, omegamoon found most of the RK30 related technical data alreadt
<fonix232>
*already
<fonix232>
well, most of the available ones
<fonix232>
dammit!
<fonix232>
this repo sucks
<fonix232>
cannot push it to github
<fonix232>
"one of the refs exceeds the limit"
<Astralix>
lol
<Astralix>
fonix did you get my private chat?
<jeremb>
actually, I have a full copy of redmine as of yesterday evening if some of you are interested
<tonikasch>
:D I am
fonix232 has quit [Ping timeout: 250 seconds]
<jeremb>
just kidding haha :)
<tonikasch>
oh xD
<tonikasch>
I have a full copy of that ftp but I think that has been already uploaded
fonix232 has joined #linux-rockchip
<fonix232>
okay, I'm back
<fonix232>
what was my last message?
<fonix232>
my internet went down quite unexpected
<tonikasch>
<fonix232> cannot push it to github
<tonikasch>
<fonix232> "one of the refs exceeds the limit"
<fonix232>
hmm good
<fonix232>
thanks
<tonikasch>
:)
<tonikasch>
so...?
<fonix232>
well pushing the whole repo failed
<fonix232>
probably related to the loss of internet
<tonikasch>
:(
<tonikasch>
well, have to leave, see you later
tonikasch has quit [Quit: Saliendo]
hramrach has joined #linux-rockchip
cnxsoft has quit [Remote host closed the connection]
hramrach has quit [Ping timeout: 240 seconds]
hramrach has joined #linux-rockchip
<PiyushVerma>
VPU & GPU Not open
rellla has quit [Ping timeout: 264 seconds]
<Astralix>
VPU should be enough open to get it running and GPU is MALI400 so we cannot really blame RK for it, it is ARM.
<Astralix>
But ARM offerend partly open source code and closed source modules. So we should try to combine that into a working AOSP image
<PiyushVerma>
VPU is unique for each SOC
tonikasch has joined #linux-rockchip
<Astralix>
I know that these graphics parts are a pain and the most protected IP on each chip. But it really would help if they would make am open source interface for a closed source driver available.
<Astralix>
I can live with that. They must keep their IP to stay alive on the marked and not beeing killed by beeing copied.
<PiyushVerma>
Astralix: There is many like we use vaapi for AMD VPU
<Astralix>
yes
<PiyushVerma>
I am download android leak need to check what level is open there
<Astralix>
And the MALI SDK was done by ARM with the same intention. But actually we need some specialists that put these parts together
<PiyushVerma>
usualy there is binay build for specific platform which is provided by soc maker
<Astralix>
yes, let's see if we can get that one for RK chips
<Astralix>
but I have to leave now. See you all later in here.
<PiyushVerma>
right as some dicussion going directly with RK may be we can ask them
<PiyushVerma>
or we can work around with libhybris
<PiyushVerma>
on android binary
<fonix232>
I will definitely ask them about these, as soon as they reply
<fonix232>
collect your questions somewhere
<Astralix>
I would really appreciate to work with them together than to work on leaked unapproved illegal stuff
<PiyushVerma>
yes but they must should release GPS what ever they release
<Astralix>
GPL?
<Astralix>
Because GPS doesn't work I think
<PiyushVerma>
GNU PUBLIC LICENCE
<PiyushVerma>
sorry
<Astralix>
(sorry was just kidding)
<PiyushVerma>
:)
<Astralix>
so, have to leave now
<Astralix>
family is waiting
<fonix232>
well we can try to make GPS work
<PiyushVerma>
Astralix: see u
<tonikasch>
which GPS? ah, that on tablets, I guess...
<tonikasch>
bye, Astralix
<PiyushVerma>
fonix232: gps wont be dificult right it just use UART
<Astralix>
fonix, guess there is a bug in the SOC
<Astralix>
not that one, the one that uses the SOC high speed counter
<Astralix>
GPS I am trying to repair on MTK6628
<PiyushVerma>
tonikasch: sorry GPS -> GPL
<tonikasch>
oh, yes, I knew, PiyushVerma ;)
<tonikasch>
but by the way I continued talking about gps :)
<PiyushVerma>
aah ok so it's gps or agps
<tonikasch>
don't know, actually
<tonikasch>
I only own a mk808
<PiyushVerma>
I think gps would be easy it's just about reading RS232 right ?
<tonikasch>
fonix232 said so
hramrach has quit [Ping timeout: 240 seconds]
<PiyushVerma>
right
<PiyushVerma>
What suggestion for Mali should we wait for binary or bring up lima or try libhybris
<tonikasch>
I tried lima about a month ago but support was still on the way...
<tonikasch>
I didn't know about libhybris, what's it about?
<PiyushVerma>
it's a way to call android lib by daynamicaly loading android lib
<tonikasch>
ahm
<tonikasch>
Perhaps best option for now would be trying libhybris
<tonikasch>
I'm taking a look, but I'm on holidays now and on 3g, only 100 MB left...
<fonix232>
I was just kidding about the GPS thing :D
<fonix232>
so, you plan to use hybris to use the existing GPU libraries on a pure glibc Linux build?
<PiyushVerma>
Just discussion if we expecting binary soon then no use
<PiyushVerma>
else start wokring on it
<PiyushVerma>
one more option try some prebuilt lib
<PiyushVerma>
today I setup RS232 on my device tomorrow will try to build kernel and setup
<PiyushVerma>
it's 12am here now
<PiyushVerma>
So I have to leave
<tonikasch>
ok, PiyushVerma, see you
<PiyushVerma>
see you
<fonix232>
you don't really need RS232 to flash a kernel
<fonix232>
and UART would be more useful for debugging
<PiyushVerma>
Yes any we we need that to debug
hramrach has joined #linux-rockchip
<tonikasch>
bye!
hramrach has quit [Ping timeout: 240 seconds]
tonikasch has quit [Quit: Itaka, berriro...]
<Tsvetan>
fonix232 is the login info in the readme.txt left intentional :)
<fonix232>
it probably was :D
hramrach has joined #linux-rockchip
hramrach has quit [Ping timeout: 240 seconds]
hramrach has joined #linux-rockchip
Astralix has quit [Ping timeout: 264 seconds]
fonix232 has quit [Quit: Page closed]
hramrach has quit [Ping timeout: 240 seconds]
hramrach has joined #linux-rockchip
<Astralix1>
You could just laugh if you think twice about the idea that RK left the login info int the txt file by intention...
<Astralix1>
We should find the bugs they are not able to solve and do the job for them.... Auuu!
<Astralix1>
Just kiddi'n
<Astralix1>
dont take the serious :)
mcbrick has joined #linux-rockchip
Topgun100 has joined #linux-rockchip
<Astralix1>
Topgun100... where are Maveric and Goose?
<Astralix1>
SCNR
<Topgun100>
hehe
<Astralix1>
Hi
mcbrick has quit [Quit: Leaving]
<Topgun100>
Hi
<Topgun100>
Anyone here using rock chip in an embedded solution? What is the stability like with the the cheap sticks?
mcbrick has joined #linux-rockchip
<Astralix1>
I am not sure how stable the RK3188 is, but I stressed the NAND they bundle with it very much and it survived
<Topgun100>
Nice. I think rk3188 is overkill for most embedded scenarios. I just need something that is stable and runs linux. One of the most important things I need to work on is a build with a read only root partition. Ubuntu writes a lot to /var/ so if there is a power cut, the OS will get trashed.
Theueip has quit [Quit: Leaving...]
<Astralix1>
I guess for more industrial than commercial use, you should look at TI sitara or Freescale i.MX series
<Topgun100>
The cost of rk3066 is very appealing though.
<Topgun100>
The biggest issue is going to be quality control.
<Astralix1>
yes but most tablets or sticks made from it don't survive a year or two.
<Astralix1>
But that is cause they get replaced cause of old Software or dead batteries
<Topgun100>
Oh dear. I think that answers my question
<Astralix1>
so even the RK3xxx is still alive and kicking, these cheap sticks get replaced for lots of other reasons that do not have anything to do with quality
<Topgun100>
the MT Box, or whatever you want to call it, looks at least superficially to be pretty well made.
<Topgun100>
the Kaiboer C9 also looks very good and only £32 from china (ebay).
<Astralix1>
Yes there are several intersting boxes. I didn't want to say any bad about the boxes at all, but I bet there are no people that can tell you that their system is running well for 5 years now.
<Topgun100>
that's cool. It is not going to be used for controlling life support systems or anything like that. :)
hramrach has quit [Ping timeout: 240 seconds]
thefinn93 has quit [Ping timeout: 240 seconds]
hramrach has joined #linux-rockchip
thefinn93 has joined #linux-rockchip
akaizen has quit [Ping timeout: 276 seconds]
akaizen has joined #linux-rockchip
JochenKauz has joined #linux-rockchip
tonikasch has joined #linux-rockchip
tonikasch has quit [Changing host]
tonikasch has joined #linux-rockchip
<tonikasch>
hi
thefinn93 has quit [Ping timeout: 240 seconds]
<JochenKauz>
hi tonikasch, wiki working?
<tonikasch>
mk808b?
<tonikasch>
I don't have one now, I donated mine
<tonikasch>
But I hope that kernel patch will work
<tonikasch>
Omegamoon has not had yet time to test it and mawkish has not replied, perhaps he/she didn't succeed in compiling kernel, I still don't know
<tonikasch>
ouch
<tonikasch>
sorry
<tonikasch>
I read wifi
<tonikasch>
xD
<tonikasch>
JochenKauz, yeah, I began to add some models
<tonikasch>
thanks ;)
thefinn93 has joined #linux-rockchip
<tonikasch>
... :)
JochenKauz1 has joined #linux-rockchip
<tonikasch>
JochenKauz1, wiki ok, adding some models in the free time
JochenKauz has quit [Ping timeout: 264 seconds]
<JochenKauz1>
fine, glad to see that
<tonikasch>
:)
<JochenKauz1>
kernels are sometimes a very tricky thing
<JochenKauz1>
one really bad thing are the toolchains and that andorid has linked modules against the kerne
<tonikasch>
well, until now I was using the gnueabihf toolchain, but I read somewhere that it is not good for development
<JochenKauz1>
yes... its an hf, it could work with newer designs based on A9. I never used it, caused I come from RK2918 kernel development, and the A8 is not compatible with hf
<tonikasch>
aha
<tonikasch>
well, I have to leave now, cu :)
<JochenKauz1>
I use on A8 an 4.6.2 gcc, but A9 does not work with that toolchain. highest version for me was 4.4