<hramrach_>
running the original android because I cannot replace it with *working* linux
<Dandel>
any ideas on when the image will be updated to kernel 3.14 ( or 3.16 )?
<Dandel>
*rockchip kernel images
<hramrach_>
when you update it
<hramrach_>
there is mainline suport in progress so you can flash mainline kernel on your device or put it on a SD card
<hramrach_>
problem is 1) you lose display support 2) you have to figure out how to configure dram and DT
<arokux>
very small problems :)
<Dandel>
DT ( as in display timings) ?
<hramrach_>
as in device tree
<hramrach_>
ther are no display timings with no display support
<hramrach_>
I have obtained some firmware images for my tablet so presumably first half of 2) can be achieved by unpacking those images
<hramrach_>
but I still need DT/board.c
<Dandel>
I see... so i haft to have a physical device to do the port, right?
<naobsd>
hramrach_: there is no code to use "how to configure dram" info
<Dandel>
the oem was kind enough to give the firmware image to the device, so it may be easier than most.
JeffyChen_ has joined #linux-rockchip
<Dandel>
also, I will probably be given more information in a week or two after the engineer responsible for the firmware gets back from the pr trip.
<hramrach_>
naobsd: presumably the firmware contains a dram blob so it can be flashed on empty nand so you can extract the blob and use it if rkunpack knows how to extract it
ganbold__ has quit [Ping timeout: 255 seconds]
<Dandel>
i also was wondering, has anyone else figured out what exactly is going on with the omniemu?
<hramrach_>
Dandel: to make your device easily supported you need the firmware source with board.c and hope they did not butcher the sources much so you can just drop the board.c in one of the linux-rockchip kernel branches and build it
<Dandel>
hramrach_, i already figure that, and i am also trying to get some user space gpl code identified.
<Dandel>
i sort of spotted six different sources of gpl code ( it appears they are using the jxk multi-emulator )
<hramrach_>
what do they use it for?
<Dandel>
old school game support ( namely nes/snes/etc )
<hramrach_>
probably the most useful code outside kernel to get would be video decode. afaik there is no opensource support for that
<naobsd>
hramrach_: please don't speak/write about something you presumed but not confirmed and/or not experienced _as if "the fact"_
<hramrach_>
or is there?
<hramrach_>
how can I confirm there is something?
<hramrach_>
oh, there is actually nice page about it
<Dandel>
ump.ko?
<hramrach_>
ump is some memory manager and there is source for it
<Dandel>
it appears that there is no kernel module for the media plugins
<naobsd>
hramrach_: why can you say something without confirmation, without experience, without knowledge?
<naobsd>
only "presume" is enough to answer question?
<JeffyChen_>
hramrach_: the pid in maskrom, is the chip id, which you can find in uboot/tools# grep -r NAME rk_tools/RKBOOT/*.ini, like for 3288:NAME=RK320A, then the pid is 320A
<Dandel>
not unless it is the ipp.ko module ( which "claims" to be gpl )
<hramrach_>
ipp.ko is supposed to be some 2d engine
<hramrach_>
or interface to it. it probably requires some userspace component as well
<Dandel>
the ipp.ko module falls under gpl ( at least for the copy i got )
<AstralixNB>
Dandel, RK just provided low level access code as gpl in source code. The real codecs are in the android interface libs. The reason is the needed licensing for codecs.
<JeffyChen_>
is that ./drivers/staging/rk29/ipp/rk29-ipp.c?
<Dandel>
JeffyChen_, I think so.
akaizen has quit [Remote host closed the connection]
akaizen has joined #linux-rockchip
<hramrach_>
ok, there is that libvpu which is very lowlevel but at least should give some idea what building blocks are available in the VE
<hramrach_>
so it should be workable but if you were to actually use that with a media player you will probably learn more about video decoding than you ever wanted
<AstralixNB>
hramrach_ this is pretty precise!
<arokux>
:)
<JeffyChen_>
there will be a v4l2 driver(support h264/vp8 en/decode)
<naobsd>
JeffyChen_: from RK?
<JeffyChen_>
yes, for 3.14 kernel
akaizen has quit [Ping timeout: 272 seconds]
<naobsd>
well, I saw some patches on chromium gerrit
<naobsd>
anyway, sounds very nice
<AstralixNB>
JeffyChen_ if anyone of us could support and help you to get this going, just call!
<JeffyChen_>
thanx~
<naobsd>
:)
<JeffyChen_>
we are trying to upstream a lot of stuff
<AstralixNB>
Yes, I know and we all really appreciate this
<naobsd>
JeffyChen_: I know, rockchip people are doing very nice work :)
<JeffyChen_>
:)
<AstralixNB>
I am still missing the RK3188 VPU documentation, so is the VPU unit in RK3188 and RK288 identical? Or is the code at least easy to backport to RK3188?
<JeffyChen_>
yes, i think so
<JeffyChen_>
i did not see much difference in libvpu, so ...
<arokux>
JeffyChen_: do you still work for RK?
<JeffyChen_>
yes, i used to do some android framework(java) stuff, but now working on chromeos's kernel...
<JeffyChen_>
i am very new to kernel...
<arokux>
JeffyChen_: I've been told Google is also working on the mainline/chromeos kernel
<Dandel>
JeffyChen_, how hard would it be to generate a mock up code if the kernel from a kernel image(s)?
<AstralixNB>
So you're probably at the right point here. Probably you should get good support from may guys in here
ganbold_ has joined #linux-rockchip
<JeffyChen_>
yes, we will upstream and get some reviews first, then backport to chromeos's kernel
<Dandel>
*if the kernel code is unavailable.
<JeffyChen_>
dandel: i do not quite understand
<arokux>
JeffyChen_: sounds fantastic as for a SoC vendor :)
<hramrach_>
naobsd: is it possible to read back data you write with the L l commands?
<Dandel>
JeffyChen_, i found a device oem that appears to not follow the gpl at all ><; ( sales never got asked about it either )
<naobsd>
Dandel: JeffyChen is not free/almighty answering machine...
<naobsd>
hramrach_: why you need to read back? you have data
<hramrach_>
naobsd: when you do not have data it's useful to get them
<Dandel>
naobsd, i know that, but the other idea would be to contact the company to remind them ( and offer help ) regarding gpl compliance.
<JeffyChen_>
dandel: i'd like to share my code(which would be very ugly...),but the company..
<hramrach_>
or to compare data from device with data extracted from image
<naobsd>
hramrach_: DDR init code is available on some places such as in firmware image, u-boot source tree. as far as I know, it's common code.
<naobsd>
Dandel: here is not the place for "fight for GPL!"
<hramrach_>
the code is common but it contains parameters specific to baord, right?
<AstralixNB>
Fortunately there are people around that have the license to test even closed source.
<JeffyChen_>
hramrach: we have a tool called upgrade_tool(which is close sourced), can dump data
<naobsd>
hramrach_: no
<AstralixNB>
This might speed up mainlining or verification and that will then result in code or blobs beeing available for all of us
<naobsd>
hramrach_: common code is common code
<hramrach_>
so every board uses same memory settings?
<JeffyChen_>
different ddr bins...
<JeffyChen_>
there also will be some ddr code in chromeos's firmware, i heard
<naobsd>
DDR init code is separeted by SoC type
<naobsd>
and there is some version
<AstralixNB>
Dandel, if you use DDR chips that are in the list of the Loader, they are supported without having the loader sources. So why do you need the Loader as source code?
<Dandel>
i think i got the ddr bin for the device :) ( easy part )
<naobsd>
but it's common code, same code is used by many boards
<AstralixNB>
exactly. The Loader supports a large list of DDR RAMs and NAND chips.
<hramrach_>
but it's still useful to extract it from firmware or board if firmware is not available because you cannot easily determine what loader your chips are compatible with
<AstralixNB>
You can ask RK for the pdf that lists all the chips. I am pretty sure that the document is not a secret.
<naobsd>
I have no objection that extract code from device when compat code is not available
<naobsd>
but I have objection that "DDR configuration _must_ be extracted for every device"
<AstralixNB>
And I am sure that you should have initialized DDR RAM interfaces once or twice before you understand what this code does and what is needed to sync SOC DDR/IF and DDR RAM.
<naobsd>
there is e.g. RK DDR Support List Ver2.07.pdf in RK3288 SDK
<hramrach_>
how do I tell if my configuration is compatible or if it should be extracted?
<AstralixNB>
hramrach_ you never need to extract something. Just get the DDR Support List Document and check if the chips are listed.
<hramrach_>
I have no idea what chips I have
<JeffyChen_>
32_LPDDR2_300MHz_LPDDR3_300MHz_DDR3_300MHz_20140630.bin means it's for rk32xx, and is for lpddr2/lpddr3/ddr3 and ddr freq will be 300mhz after init by this ddr bin
<AstralixNB>
lol, you're kidding us?
<naobsd>
if you can get firmware for your device, DDR init code is included in it
<JeffyChen_>
also, you can find out in console log
<naobsd>
if you cannot get firmware and you cannot identify RAM chip, throw away it from window
<AstralixNB>
hramrach_ the loader only starts up the DDR RAM slowly (300MHz) and the kernel then uses it's ddr driver to speed up or slow down the DDR again. So if you like to have a fast device or a cool power-management, you need to dive into the ddr.c code of the kernel. Not the loader
<hramrach_>
which kind of explains why it does not open easily
<AstralixNB>
JeffyChen_ did you get my PN?
<JeffyChen_>
PN?
<AstralixNB>
I wrote a private message to you check your IRC program if there something popped up :)
akaizen has joined #linux-rockchip
<hramrach_>
naobsd: I browsed some Chinese forums and downloaded what seems to be firmware for this device. however, I don't know if the firmware actually works so I would prefer to get as much as possible from the known working firmware in the device
akaizen has quit [Read error: No route to host]
akaizen has joined #linux-rockchip
<arokux>
hramrach_: how much did you pay for your tablet?
<hramrach_>
it was like $130 when I got it I think
<AstralixNB>
hramrach_ from my own experience it is very hard to build your own working kernel into a device that you cannot open.
<AstralixNB>
The only way would be to get a working kernel from your tablet manufacturer
<arokux>
hramrach_: is it FNF ifive X2?
<hramrach_>
yes, it is
<AstralixNB>
The other way is to have good assembler knowledge and some good tools to disassemble a kernel
<AstralixNB>
But then you need the real kernel working in the tablet as a binary
<Dandel>
JeffyChen_, have ya ever heard of piglit by any chance?
<AstralixNB>
Hmmm... put a cooler ontop of the SOC first
<Dandel>
hramrach_, also does egl/opengl es 1.x/2.x/3.x
<arokux>
hramrach_: and where did you get it for $130? or has it got more expensive lately?
<Dandel>
and it does not crash most modern implementations of the catalyst driver on linux :)
<hramrach_>
yes, it does ES as well but the testsuite for non-desktop GL is limited
<naobsd>
I don't recommend cheap device for people who worry about support quality...
<hramrach_>
arokux: it was some time ago
<hramrach_>
so its price probably dropped since then
<arokux>
naobsd: which one would you recommend?
<naobsd>
arokux: anything off-topic here :)
<arokux>
hramrach_: hm... I couldn't find it cheaper than 150..
<Dandel>
arokux, i would look at the hisense devices... reasonably good experiences with their devices ( especially the sero 7 pro )
<arokux>
alright :)
akaizen has quit [Ping timeout: 260 seconds]
<hramrach_>
the hardware is good but the support is unfortunately Chinese
<naobsd>
I think, good device for here is,
<Dandel>
hramrach_, yes, it is chinese support, but it is gpl compliant :D
<naobsd>
the device which is opened just after arrived & before powered-on ;)
<arokux>
naobsd: true
<Dandel>
so that means you get good cfw and in my experience upwards of a week of battery life ( reading books). if i use it heavily it is still a few days though.
<hramrach_>
Dandel: did you get any sources from them?
<Dandel>
hramrach_, yes
<hramrach_>
they may well provide some but I cannot navigate the Chinese pages easily
<arokux>
last off-topic msg: I prefer to buy use nexus devices. rk is just to learn how things work
<Dandel>
just search for the model ( Sero 7 Pro or Sero 7 LT ) and it will give you the kernel code.
<Dandel>
the Sero 7 pro is an nvidia soc,, but the sero 7 lt is a Rockchip rk3066 device.
<hramrach_>
neither of which helps me
<JeffyChen_>
many of my friends like tesco 7' tablet(3188)
<naobsd>
JeffyChen_: btw, upgrade_tool support raw(physical sector) NAND access, it includes OOB data such as ECC, right? I'm not sure how -decode works
<naobsd>
I snooped USB communication some time ago, I can understand protocol header, but I don't know proper payload layout, ECC calculation etc
<naobsd>
this is the reason that I cannot implement raw NAND access to rkflashtool yet
<naobsd>
I know I can see flashed data on device, but raw dump is useless
<JeffyChen_>
i've no idea about ECC/OOB... i only know LBA and PBA
<naobsd>
ah, what I asked is PBA access
<JeffyChen_>
i thought pba is the first 4M of the storage
<Dandel>
and what naobsd means by "ECC calculation" is stuff like image verification using crc.
<hramrach_>
I have multiple A13 7" tablets which I took apart before running them because it was easy with them
<hramrach_>
the last one is still fully working :)
<naobsd>
when I snooped USB communication, I can see OOB data(ECC etc) between actual DDR init/loader code
bengal has joined #linux-rockchip
<naobsd>
every 1k or 2k or 4k or... I forgot it
<hramrach_>
but I wanted something with usable screen which is why I got the iFive tablet
<naobsd>
page size (OOB position and size) must depend on NAND type
<naobsd>
a lot of code is needed to handle NAND physical access
<naobsd>
so only LBA access is supported by rkflashtool ;)
<JeffyChen_>
i could not access the nand source code, i just use the nand api to read/write pba
<AstralixNB>
There are only patented FTL systems available. So there is no chance to get open source code legally.
<naobsd>
JeffyChen_: sorry, I talked about USB protocol, it seems rockusb PBA access want some work on host side
<arokux>
what is OOB?
<AstralixNB>
Even the mechanisms available in the kernel as open source are under software patent
<naobsd>
sorry, I'm not talking about source code openness
<naobsd>
what I talked is USB protocol payload format for PBA read/write
<naobsd>
arokux: out-of-band area on NAND
<AstralixNB>
No, but if you want to preset the NAND completely you need to replicate the FTL and ECC in the usb tool, so you basically need the license
<naobsd>
arokux: which stores ECC info etc
<naobsd>
AstralixNB: actual access to/from NAND is done by loader
<naobsd>
AstralixNB: USB protocol just passes data, but it includes OOB area, I'm not sure license is needed to generate info for OOB (I guess: it's needed)
<naobsd>
I'm not sure about patent
<JeffyChen_>
when flash loader though fastboot, it just pass the whole loader though usb, then the fastboot will unpack it and deal with pba. but when flash loader though rockusb, it's the host tool side to deal with the unpack/P_RC4/idblock struct
<naobsd>
but no need to worry about it, I cannot write NAND support code with current my knowledge ;)
<AstralixNB>
Naobsd, that is what I thought, but I wanted to go the more open way and check if we can just use one of the already working implementations in the kernel.
<naobsd>
oh I didn't check fastboot flash bootloader
<JeffyChen_>
fastboot flash loader loader-file
<naobsd>
JeffyChen_: but it requires working loader on device
<naobsd>
if we can run u-boot(fastboot) instead usbplug loader, it's possible
<AstralixNB>
you can do that with the upgrade-tool too. you can upload a loader to dram without flashing it
<naobsd>
AstralixNB: what we're talking is our open source tool ;)
<naobsd>
I know upgrade_tool can flash bootloader
<AstralixNB>
yes, but to verify the mechanisms you can inspect the tricks of fastboot and upgrade_tool
<AstralixNB>
And upgrade_tool can upload and run loader _without_ flashing it...
<naobsd>
about "upload loader to dram w/o flashing it" case, only usbplug loader is working well
<naobsd>
AstralixNB: I implemented it to rkflashtool to
<AstralixNB>
But what is your primary goal? Make your flash tool capable of flashing bootloader?
<hramrach_>
also regarding the dram code
<hramrach_>
is the dram code same for nand and mmc or does the dram code also include the code to load the next stage?
<naobsd>
AstralixNB: I don't want, people want
<JeffyChen_>
sorry guys, i must leave for my english classes now, bye :>
<naobsd>
JeffyChen_: no problem, really thanks
<AstralixNB>
JeffyChen, your english is pretty good already :)
<naobsd>
I think so too :)
<hramrach_>
JeffyChen_: thanks and good luck with your classes :)
<AstralixNB>
naobsd, btw. did you add my parameter read automatism to your tool?
<naobsd>
AstralixNB: function which uses partition name, not offset/size?
<AstralixNB>
Yes, I wrote that long long times ago
<naobsd>
AstralixNB: it's already added before forking
<AstralixNB>
I forked long times after getting the first code of it
<AstralixNB>
but probably I wasn't the only one who thought this might be useful
<AstralixNB>
So there may be multiple implementations of that thing out there
<AstralixNB>
My was dual-headed. I used a local file and if not present, I read it from the device.
<naobsd>
AstralixNB: it seems julien implement it
<naobsd>
about whole rkflashtool, I like stdin/out thing but it's really terrible interface!
<naobsd>
and single letter command is really really terrible!
<hramrach_>
single letter is passable
<hramrach_>
compared to overhaul of whole tool it's better to keep compatibility
<hramrach_>
stdin/out should probably go away once the rkflash does not depend on openssl to do encryption
<naobsd>
I have a right to blame that interface design
<naobsd>
I designed it.
<hramrach_>
there are programs out there that use the same interface for ages. Mostly archivers with DOS origin ;-)
bengal has quit [Quit: Leaving]
<naobsd>
I don't think compatibility to DOS archives is important
<naobsd>
archivers
<hramrach_>
well, they used that on DOS and the unix version is compatible with that to this day
<naobsd>
rkflashtool has no relation to them
<hramrach_>
except the old school single letter interface
cristian_c has quit [Quit: Sto andando via]
akaizen_ has quit [Remote host closed the connection]
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
akaizen has joined #linux-rockchip
akaizen has quit [Ping timeout: 272 seconds]
naobsd has quit [Quit: Page closed]
benja` is now known as benja
nighty^ has joined #linux-rockchip
FreezingCold has quit [Ping timeout: 272 seconds]
GriefNorth has quit [Ping timeout: 272 seconds]
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
FreezingCold has joined #linux-rockchip
FreezingCold has quit [Read error: Connection reset by peer]
FreezingCold has joined #linux-rockchip
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
<hramrach_>
hmm, so I upgraded to android 4.4.2 and ht tablet still works
<hramrach_>
so I can presumably extract anything from the firmware image
naobsd has joined #linux-rockchip
FreezingCold has quit [Ping timeout: 255 seconds]
FreezingCold has joined #linux-rockchip
Soopaman has joined #linux-rockchip
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
FreezingCold has quit [Ping timeout: 240 seconds]
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
FreezingCold has joined #linux-rockchip
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
AstralixNB has quit [Ping timeout: 250 seconds]
FreezingCold has quit [Read error: Connection reset by peer]
FreezingCold has joined #linux-rockchip
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
<hramrach_>
this whole android thing is fucking annoying. there is an annoying lock screen background in the 4.4 image and there is no way to change it
Soopaman has quit [Ping timeout: 244 seconds]
JeffyChen_ has quit [Ping timeout: 246 seconds]
Soopaman has joined #linux-rockchip
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
FreezingCold has quit [Read error: Connection reset by peer]
FreezingCold has joined #linux-rockchip
mrcan__ has joined #linux-rockchip
mrcan_ has quit [Ping timeout: 258 seconds]
akaizen has joined #linux-rockchip
FreezingCold has quit [Read error: Connection reset by peer]
FreezingCold has joined #linux-rockchip
gb_master has joined #linux-rockchip
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
Astralix has joined #linux-rockchip
bengal has joined #linux-rockchip
Dandel has quit [Remote host closed the connection]
gb_master has quit [Remote host closed the connection]
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
Soopaman has quit [Quit: Leaving.]
Soopaman has joined #linux-rockchip
akaizen has quit [Remote host closed the connection]
akaizen has joined #linux-rockchip
Soopaman has quit [Quit: Leaving.]
akaizen has quit [Ping timeout: 250 seconds]
FreezingCold has quit [Read error: Connection reset by peer]
FreezingCold has joined #linux-rockchip
akaizen has joined #linux-rockchip
akaizen has quit [Ping timeout: 260 seconds]
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
akaizen has joined #linux-rockchip
akaizen_ has joined #linux-rockchip
akaizen has quit [Ping timeout: 272 seconds]
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
arokux2 has joined #linux-rockchip
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
nighty^ has quit [Quit: Disappears in a puff of smoke]
hramrach_ has quit [Remote host closed the connection]
hramrach_ has joined #linux-rockchip
bengal has quit [Quit: Leaving]
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
benja is now known as benja`
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
naobsd has quit [Quit: Page closed]
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
hipboi has quit [Ping timeout: 246 seconds]
arokux2 has quit [Remote host closed the connection]