ChanServ changed the topic of #linux-rockchip to: Rockchip development discussion | Wiki at http://linux-rockchip.info | Logs at http://irclog.whitequark.org/linux-rockchip
another_ has joined #linux-rockchip
another_ has quit [Quit: Page closed]
hipboi_ has quit [Read error: Connection reset by peer]
hipboi_ has joined #linux-rockchip
hipboi_ has quit [Ping timeout: 252 seconds]
hipboi_ has joined #linux-rockchip
<hipboi_> naobsd, how do you know emmc boot priority is higher than sd card
<naobsd> guessed. with bootable eMMC, bootable SD didn't work at all. I wiped eMMC, then same SD card worked
<naobsd> I just tried few times, something may be wrong/missing
<naobsd> I used different u-boot version to avoid confusion, so "eMMC is prior" should be correct...
<hipboi_> naobsd, checking rk3188 trm...
<naobsd> ah, good idea :)
<hipboi_> naobsd, it did not mention sd card
<hipboi_> only external nand, spi nor and emmc
<hipboi_> naobsd, rockchip should let sd boot priority higher
<naobsd> if eMMC is bootable but not functional
<hipboi_> naobsd, at the factory, they need to test the pcba, which runs the system on sd card
<naobsd> we need needle to short ;)
<naobsd> wipe eMMC, then it always boot from SD
<naobsd> flashing rom to eMMC will be last step to ship ;)
<naobsd> or, there might be a pin to change order
<naobsd> btw, first overview section mentioned about boot from SD
<naobsd> in TRM
<naobsd> oh wait, flow says "check NAND 1st, then eMMC"
<naobsd> it shouldn't be correct ;)
<naobsd> can anyone ask about boot media order to rockchip person? ;)
<naobsd> I have to leave here now
<naobsd> later
naobsd has quit [Quit: Page closed]
hramrach has quit [Ping timeout: 264 seconds]
hramrach has joined #linux-rockchip
aaaaaaaaaaa has joined #linux-rockchip
GriefNorth has joined #linux-rockchip
<aaaaaaaaaaa> Hi there, I am trying to develop additional net downloading function for a RK2818 based E-ink reader which named Boyue G10, the ROM is Linux but it is not open officially. I have tried several official Android unpack tools to unpack the official update img but all failed. So the question is, I have uncompressed that update img and get boot.img, kernel.img and so on, how to extract these img files? BTW it shows boot.img.
<aaaaaaaaaaa> any help is appeciate.
naobsd has joined #linux-rockchip
<naobsd> explaining detail is the best way to get answer
<naobsd> I decided to prepare new server for community
<naobsd> for git server
<naobsd> and for download files
<naobsd> probably I'll move wiki on it
<naobsd> linux-rockchip.org (web/wiki) dl.linux-rockchip.org git.linux-rockchip.org
<naobsd> hipboi: it seems I can make new clone from bitbucket/firefly, maybe there is short period bandwidth limit?
<naobsd> Repository size
<naobsd> Soft limit 1 GB: We will notify you by email and with a notification bar in Bitbucket.
<naobsd> Hard limit 2 GB: We will disable your ability to push to the repository. You will be notified by email and with a notification bar in Bitbucket.
<naobsd> Downloading archive .zip files
<naobsd> Must be no larger than 2 GB.
<naobsd> I see
<naobsd> I'll prepare server anyway
<aaaaaaaaaaa> Thanks @naobsd. G10 E-ink device has a RK2818 chip but it is not Android tablet, Linux ROM instead. Officially there is update img to update G10 ROM, so I choose the way that unpacking the update img to see what happens to the device. The update img is self uncompress package made by MakeFile actually. When I run the update img it uncompress several img files.
<aaaaaaaaaaa> And these files are named boot.img, kernel.img, oem ........ I thought these files are change files for Linux partation. But I was stopped here because no way to unpack these img files.
<aaaaaaaaaaa> I use command head -10 boot.img, there is a line shows boot.img.cpio
<aaaaaaaaaaa> But I cannot uncompress boot.img by using cpio -idv < boot.img
<aaaaaaaaaaa> So I thought it might not be a cpio package but no idea what is it and how to unpack boot.img or other img files.
<naobsd> ok, I understood your idea, but I still have no idea to get those files
<aaaaaaaaaaa> Anyway, thank you to check my question. @naobsd
<naobsd> tail -c +29 boot.img | gzip -d | cpio -t
<naobsd> I don't understand why you don't want to tell some basic info such as url
<naobsd> please don't ask how to repack/customize/etc. I don't know anything more than above commands
<aaaaaaaaaaa> @naobsd, WOW! it works and lists all package content. Great! Thank you so much. There is no URL because the official website delete all links of update images and no one explain why.
<aaaaaaaaaaa> Oh. it uncompressed!
<naobsd> you should have that file
<aaaaaaaaaaa> yes, I have this file , but it is over 100MB, and no idea where I can put it. btw, I am currently in China and dropbox doesnt work here.
<naobsd> *.img have header start with "KRNL", it should be removed by rkunpack and readded by rkcrc https://github.com/naobsd/rkutils
<naobsd> I have no idea about data (name and size) between RK header and gzip header
<naobsd> I already remove files, I really cannot do any, good luck
<aaaaaaaaaaa> That's already help, at least explain that is a real cpio package. Thank you very much! @naobsd
<naobsd> I remembered I want to get recent e-ink reader...
<naobsd> wanted
<aaaaaaaaaaa> LOL, dont buy a boyue device, it comes without any technical service and support and the system is very close to Internet.
<aaaaaaaaaaa> closed
<naobsd> ah, what I wanted is major one
<naobsd> and I'm sure about that kind of cheap device
<aaaaaaaaaaa> KPW2 is a good one. I choose Boyue G10 because it is cheapest 10 inch reader. 10 inch is what I must have to read scanned pdf book.
<aaaaaaaaaaa> And now I pay for the cheapest choice.
<aaaaaaaaaaa> I will be back later
<amstan> naobsd: why do you want a git server there? I really enjoy github
<naobsd> Rockchip's Android SDK is single git repo, too large for github/bitbucket
<naobsd> du -sh firefly-rk3288.git
<naobsd> 3.7Gfirefly-rk3288.git
<naobsd> du -sh radxa_rock_android4-4.git/
<naobsd> 3.6Gradxa_rock_android4-4.git/
<naobsd> I can split them as like as AOSP repositories
<naobsd> but handling updates from rockchip will be annoying...
<naobsd> new git server will be just for SDK distribution
<amstan> ah
<naobsd> I'm pushing kernel subtree in SDK to github
<naobsd> it also annoying a little ;)
hipboi_ has quit [Read error: Connection reset by peer]
ChiaSmurf has quit [Ping timeout: 268 seconds]
hipboi has joined #linux-rockchip
<rperier> hi all
dujw has joined #linux-rockchip
hipboi has quit [Ping timeout: 276 seconds]
<naobsd> hi
<naobsd> haha, "selinux" is buzzword in some community :)
<naobsd> when someone try to hack something (e.g. making custom rom)
<naobsd> something doesn't work well
<naobsd> "oh, selinux causes this issue!"
<rperier> I am no longer allowed to say "aaaaaaaaaaa" when I am surprise now. Otherwise I will ping aaaaaaaaaaa each time :P
<rperier> :D
<phh> well, it's an helper
<phh> just type aa<tab><backspace>
<naobsd> oh
<rperier> hehe, exactly
<naobsd> I just notice TAB can be used on webui
<naobsd> I cannot count how many a, but no need to count it now
hipboi has joined #linux-rockchip
<rperier> naobsd: webui ? from webchat.freenode ?
<naobsd> rperier: yes
<rperier> nice
hramrach has quit [Remote host closed the connection]
hramrach has joined #linux-rockchip
zombu2 has quit [Read error: Connection reset by peer]
wildea01 has joined #linux-rockchip
johan__ has joined #linux-rockchip
miro__ has joined #linux-rockchip
nighty^ has joined #linux-rockchip
johan__ has quit [Quit: Page closed]
hramrach has quit [Remote host closed the connection]
hramrach has joined #linux-rockchip
bengal has joined #linux-rockchip
<miro__> Hi guys. Where I can find the kernel source and .config file used to build "radxa_rock_debian_wheezy_140624" image ?
bengal has quit [Ping timeout: 268 seconds]
<naobsd> miro__: there are(were) some repos/branches, and probably tag was not added, finding it is difficult a little...
<naobsd> miro__: and here is not channel dedicated for radxa, #radxa is for radxa (and talk.radxa.com)
ChiaSmurf has joined #linux-rockchip
<miro__> ahh ok, I shall try there. Anyway is there any document that describes how RK83188 system boot works, and in private RK8133Loader. Where it must resides, is it configurable (load kernel from specific flash address)?
<naobsd> basically, no document for rockchip
<naobsd> RK3066 TRM may be useful
<naobsd> I will write wiki, but not available yet
<naobsd> address/size of logical partitions are defined in parameter (file/partition)
<naobsd> no document/source about physical NAND
<naobsd> you can disassemble RK loader
<naobsd> why do you want to know internals?
<miro__> I read TRM but it just say what mask loader do
<miro__> in general i want to build my own linux distro, and just starting to learn ho RK8133 works
<naobsd> to start, please type 3188 correctly...
<naobsd> and understand loader internal is not mandatory to "build my own distro"
FreezingCold has quit [Ping timeout: 268 seconds]
<karlp> the bootloader search order and requirements of the bootloader are though,
<miro__> It is as I have no idea on which address to load my kernel image
<naobsd> and "getting old kernel source which is used to build old image" is not mandatory too. latest one will be fine
<naobsd> miro__: did you read radxa's wiki?
<naobsd> miro__: did you already understand how to build/flash kernel etc described in that wiki?
<miro__> if you talk about Booting_Linux yes
<miro__> yes i do, I build my own kernel, I load it on address 0x0002000, and it boot actually, but was not able to load the ram fs.
<naobsd> then why you asked about "address of kernel"?
<miro__> But for example this RKAndroidTool has same default address for RK8133Loader and some "parameters" file. What is that "parameter", what is should contains? Can i tell that my boot.img start in different position then 4th MB ?
<miro__> I asked is this statr address is configurable?. I were not able to find any documentation about RK8133Loader
<naobsd> did you see parameter file and wiki page about parameter?
<naobsd> "wiki page about parameter" on wiki.radxa, that is linked from "build linux kernel" page
<naobsd> address of kernel partition on NAND can be changed, no need to understand RK loader internal
<miro__> wiki about "parameters" ?
<naobsd> as I already mentioned, parameter defines logical partitions
<naobsd> loader is stored outside of "logical partitions" area
<naobsd> it's not logical address "0x0"
<miro__> so the initial image code run boot image code (RK8133Loader) from NAND flash address 0x0. The RK8133Loader start OS (linux kernel) from address 0x00002000 (it is configurable). Is that correct ?
<naobsd> as I said RK loader is not stored at logical address 0x0, parameter is stored at logical address 0x0
<naobsd> address/size of logical partitions defined in parameter are configurable
<naobsd> RK loader load kernel and ramdisk from some logical partitions,
<naobsd> on normal boot: load both kernel&ramdisk from boot partition, or load kernel from kernel partition & ramdisk from boot partition
<naobsd> on recovery boot: load both kernel&ramdisk from recovery partition, or load kernel from kernel partition & ramdisk from recovery partition
<naobsd> mkbootimg makes kernel&ramdisk image for boot/recovery
<naobsd> rkcrc makes separated image for kernel&boot or kernel&recovery
GriefNorth has quit [Ping timeout: 246 seconds]
<naobsd> for RK3188/3288, there is another reason "no need to understand RK loader internal", there is u-boot...
<karlp> yes, if you already have uboot....
<naobsd> u-boot doesn't explain NAND internal, but loader behaviour can be changed
<naobsd> I think I posted here...
<karlp> are you going to put things back on linux-rockchip, the things about bootloaders and the good portions of the wiki? you seem to have stopped at just, "everything here is obsolete" ?
<naobsd> well
<naobsd> talking about _just now_, I'm stopping
<naobsd> I'm busy just now
<naobsd> I will refresh wiki, but I don't restrict others activity
<rperier> naobsd: does uboot work for rk3188 ? :/
<rperier> it did not few months ago
<naobsd> which u-boot source you used? latest one from radxa just work for NAND, and I fixed some for SD
<naobsd> but don't expect a lot, it's surely u-boot but only does same thing as RK loader
<rperier> does it support boot over the network ?
<naobsd> and size of RAM is hardcoded in u-boot, some kernel rely it :(
<rperier> indeed
<naobsd> rperier: of course no!
<naobsd> SD support is also special, no standard mmc interface in u-boot
<rperier> to have this feature, I need to load a zImage throught tftp from my stock initramfs and then kexec from it
<rperier> in this way I am able to "boot over the network". This is the only thing I found for the radxa
<naobsd> kexec works?
<rperier> yes
<naobsd> sounds good :)
<naobsd> better than u-boot :)
<rperier> the support is present for 3.17
<naobsd> btw barebox supports (only) network
<rperier> kexec only requires to be able to shutdown cores on the fly, support was added 2 months ago (something like that)
<naobsd> very nice
<rperier> these days I use kexec over network + yocto to develop on my board
<naobsd> my miniroot can act as bootloader which supports file system and network :)
<rperier> yeah I know. Few months ago I just created a bsp meta layer for yocto. Because I use yocto/openembedded for all of my boards, nothing more (and to be more familiar with it too)
<naobsd> btw I decided to prepare new server for SDK and other large file distribution, wiki will be moved to it
<rperier> there were not bsp meta layer for rockchip so... yeah I know miniroot too :)
<rperier> very nice btw
<naobsd> give me some time to refresh web things ;)
<rperier> in november I will make a presentation about the board (at toulouse) I will talk about miniroot, yocto mainline support , I think :)
<rperier> add a "," after yocto
<naobsd> interesting
<naobsd> ah
<naobsd> I remembered
FightingCat has joined #linux-rockchip
<naobsd> I didn't try fastboot in u-boot
<naobsd> it may be possible to do "fastboot boot xxx.img"
<naobsd> load/run kernel/ramdisk w/o flashing
<naobsd> via USB
<naobsd> it may be handy (if it really works)
<naobsd> if it doesn't work, we can fix it
<naobsd> mmm, it's late here, I have to return to home...
<naobsd> later
naobsd has quit [Quit: Page closed]
viric has quit [Read error: Connection reset by peer]
viric has joined #linux-rockchip
JohnDoe_71Rus has joined #linux-rockchip
jcarlos has quit [Ping timeout: 268 seconds]
jcarlos has joined #linux-rockchip
<miro__> why it is not possible to specify partition with zero offset in parameter file? If I add "0x00002000@0x00000000(parameter)" RKAndroidTool givs error.
FreezingCold has joined #linux-rockchip
<karlp> probably because the parameter itself is at zero?
<miro__> yes, I wanted to dump it from NAND actually. This is some .. "protection" ?
viric has quit [Ping timeout: 252 seconds]
viric has joined #linux-rockchip
jcarlos has quit [Ping timeout: 265 seconds]
jcarlos has joined #linux-rockchip
JohnDoe_71Rus has quit [Quit: KVIrc 4.0.4 Insomnia http://www.kvirc.net/]
viric has quit [Ping timeout: 245 seconds]
FightingCat has quit [Remote host closed the connection]
jcarlos has left #linux-rockchip ["Que os den ..."]
viric has joined #linux-rockchip
hramrach has quit [Ping timeout: 264 seconds]
hramrach has joined #linux-rockchip
dujw has quit [Read error: Connection reset by peer]
naobsd has joined #linux-rockchip
Bludot has joined #linux-rockchip
<naobsd> (old)RKAndroidTool doesn't have function to "dump NAND"
<naobsd> probably latest one can do it, but I prefer Linux
<naobsd> in recovery mode parameter at 0x0 can be accessed
<Bludot> cant you back up all the partitions?
<naobsd> all logical partition can be backup via USB
<naobsd> btw, I noticed
<naobsd> on rk3288
<naobsd> loader bin for emmc will not have functions for nand
<naobsd> many people believe "rockchip device cannot be bricked"
<naobsd> they will flash random rom image
Bludot has quit []
<naobsd> many bricks will be made ;)
<naobsd> (it can be unbricked by short NAND pins with needle method)
<miro__> which kernel tree I should use for radxa?
<naobsd> sleepy...
bengal has joined #linux-rockchip
<rperier> naobsd: are you able to mount the rootfs over nfs on the radxa ? did you try ? (with a mainline kernel)
<naobsd> rperier: not tried yet
<mrueg> is there any "working / not working" list for mainline kernel on radxa? :)
wildea01 has quit [Quit: leaving]
akaizen_ has joined #linux-rockchip
akaizen has quit [Ping timeout: 272 seconds]
<rperier> naobsd: found, I loose tcp packets with emac_rockchip :D
<rperier> and not with emac_arc... (the other platform driver on top of arc_emac, the core driver)...
<rperier> interesting...
<rperier> in the first case, I am unable to connect to a nfs server , I loose communication and packet... in the second case... everything works fine...
<rperier> the driver looses packets randomly... apparently.... (that's not deterministic because it works just fine most of the time)
<rperier> mhhhh.... fantastic ^^
akaizen_ has quit [Ping timeout: 260 seconds]
akaizen has joined #linux-rockchip
akaizen has quit [Remote host closed the connection]
akaizen has joined #linux-rockchip
<mmind00> mrueg: in https://github.com/mmind/linux-rockchip/blob/devel/workbench/arch/arm/boot/dts/rk3188-radxarock.dts you can see which devices get enabled ;-) ... which is largely the functionality 3.18 will provide
* karlp finally updates his devel/workbench again...
<karlp> time to try and merge all these dts cahnges :)
<karlp> do I want CONFIG_ROCKCHIP_IODOMAIN=y?
<karlp> the help just says, "if you want it" is there anyreason I wouldn't?
<karlp> naobsd: I'm trying to put together a miniroot for sdboot, like you describe http://androtab.info/radxa_rock/sdboot/
<karlp> hwoever, the zip just has boot.img, parameter.img, and sdboot_rk3188.img,
<karlp> I'm referring to "there is no "secret SD bootloader" in my sdboot_rk3188.img. no need to do decompiling, decrypting, etc. if you really want to use another bootloader, just replace FlashData and/or FlashBoot."
<karlp> how do I unpack/repack sdboot_rk3188.img to replace that section?
<karlp> you've listed them as offsets, but just say that you read them from "RK3188Loader(L)_V2.10.bin"
<karlp> also, I don't quite get the offsets you refer to, your dd commands use seek 0x2000, and 0x2000+0x4000 but your magic header andFlashData, flashBoot are both at 0x8xxx which overlaps?
<karlp> kernel.img should be the concated zImage+board.dtb, or does it need to be packed/crced/anything special?
<karlp> ok, checking my other notes, "rkcrc -k zImage+board.dtb kernel.img" so ok, that's that...
<karlp> I've got RK30xxLoader(L)_V1.22.bin from my factory firmware, so that's good.
viric has quit [Ping timeout: 268 seconds]
viric has joined #linux-rockchip
akaizen has quit [Read error: Connection reset by peer]
akaizen has joined #linux-rockchip
<karlp> http://pastebay.net/1492459 it looks _better_ now than it did in the past, but it still doesn't seem to see the sdcard at all,
<mmind00> karlp: try "rootwait" in your kernel-commandline
<mmind00> the kernel my still be occupied with finding the card when it's already missing its rootfs
<karlp> do I _need_ to put that in the kernel build config? or can I put that in the parameter file? (i'm having a little bit of a hard time working out which one is obeyed)
<karlp> haha! it's alive!
<karlp> is that rootwait something I can can work on with the dts or the board support?
<karlp> it found the sdcard at mmc1, not mmc0, I guess that's likely to be related?
<mmind00> mmc1 ... interesting but depends on how the board is wired up
<karlp> just got a new RFS on the sdcard, I think that nano thing I had was a little small,
<mmind00> rootwait simply tells the kernel to wait for the root-device ... so it will not fail if the mmc-subsystem is still looking for cards
<karlp> lsusb wasn't working or anything.
<karlp> how is naobsd reading the tsadc values? that's one of the new things I see in the updated .dts files.
<karlp> hrm, this is till not quite right though, I get "failed to execute /init" and then "The disk drive for / is not ready yet or not present.
<karlp> keys:Continue to wait, or Press S to skip mounting or M for manual recovery
<karlp> manual recovery shows that / is the right filesystem though,
<karlp> unless this is broken perms or something when I created the RFS?
<karlp> dmesg is flooded with thsi sort of thing: http://pastebay.net/1492482
<mmind00> doesn't look like errors, more like debug stuff
<karlp> I should be able to add emac stuff for wired ethernet now right?
<karlp> also, chosen->bootargs, is that _appended_ to the kernel cmdline or replaces the built in one?
bengal has quit [Ping timeout: 272 seconds]
<karlp> well, guess it wasn't used at all.
<karlp> this board has the lan8720a, like the radxa rock, how many of these pins do I need to find to make it work?
<karlp> ok, added the iio subsystem, and have /sys/bus/io/devices/iio:device[01], but nothing about saradc
<srao> you don't need CONFIG_ROCKCHIP_IODOMAIN=y right now, unless you know what you're doing
<srao> you almost certainly don't need it for SDcards since we don't yet have UHS support in dwmmc
<srao> we need it on rk3288 on our boards for wifi and audio
<karlp> was just something make oldconfig had turned up, help isn't super clear :)
<karlp> I still haven't got the factory kernel source for this, but if I can boot the factory firmwar,e can I work out the gpios for things like lan8720a->phy_int somehow online?
<srao> ah ok, it only responds to stuff in the device tree and i'm guessing it's not there for 3188 boards unless Heiko added it
<karlp> well, I'm on rk3066 actually
<srao> oh ok, then i think it really doesn't matter :-)
<karlp> yeah, I left it off.
<karlp> next steps are either ethernet or wifi, kinda thought ethernet would be easier, but really not sure which actually is :)
<srao> karlp: btw, your log from before looks like you have MMC_DEBUG defined
<karlp> huh, so I do, might have been somethign I added while I was trying to work out the bootup problems.
<karlp> any tips on whether sdio wifi (rtl8188es) is simpler to work out than ethernet?
<karlp> huh, host side usb wasn't even my .config yet!
<srao> how is ethernet connected?
<srao> built onto the chip?
<karlp> there's a SMSC LAN8720A phy, can't see anything else on the board that makes sense,
<karlp> CONFIG_USB_DWC2 is what I need right?
<karlp> or should I just forget about usb for the time being?
<srao> I know almost nothing about rk3066... so take my advice with a grain of salt. but if it's anything like 3288, yes you'd want CONFIG_USB_DWC2
<srao> generally it doesn't hurt to have extra drivers
<srao> they just won't be used
<karlp> genreally doesn't, it's nice when the kconfig help text says that though
<karlp> getting to a root shell is a major improvement at least :)
<naobsd> karlp:
<naobsd> well
<naobsd> sdboot page is not updated for u-boot yet
<naobsd> and that page just said
<naobsd> "loader code in sdboot.img is not special, it's came from RK Loader.bin"
<naobsd> it doesn't explain structure of img
<naobsd> because most user don't read such kind of info
<naobsd> complain from users are very quick, they don't read any detail, just say "give me working image and 1-step install procedure!"
<naobsd> then developers try to understand info, and say "hey, detail is not explained well, please write more" :)
<naobsd> then
<naobsd> structure of boot sd for rk3188 is
<naobsd> magic header is at 64th sector, it has info about size of following 2 area for code (and 65th sector, but 65th is only used by RK proprietary loader)
<naobsd> DDR init code is at 68th sector
<naobsd> DDR init code is at 68th sector, same data/format as FlashData in RK Loader (= encrypted)
<naobsd> loader code is at 92th sector, same data/format as FlashBoot in RK Loader (= encrypted)
<naobsd> well
<naobsd> DDR init code/loader code _start from_ 68th- / 92nd-
<naobsd> in my sdboot img, size of loader are is good for RK proprietary loaders, but too small for u-boot/barebox
<naobsd> size of whole header (include above 3(4) data and other) is 0x2000 sectors (4MiB)
<naobsd> 0x2000 sector = logical 0x0 = parameter is placed here
<naobsd> around end of 0x2000, there also be RSA key (probably for secureboot/DRM) and u-boot env(if loader is u-boot)
<naobsd> emmc structure is same, first raw 0x2000 sectors are for header
<naobsd> logical 0x0 starts at 0x2000-
<naobsd> there is a code in RK's linux kernel to make logical partitions defined in parameter as mmbclk0p?
<naobsd> RK3288 emmc/sdboot uses slightly different structure for 64/68/92th
<naobsd> ah
<naobsd> 0x2000 in dd line on my page is sector
<naobsd> 0x8000 in detail is byte
<naobsd> try hexdump -C sdboot.img, you'll understand soon
<naobsd> rperier: I didn't try emac_rockchip merged kernel yet, sorry
<naobsd> http://files.androtab.info/radxa/sdboot_rk3188.zip this is snapshot of my workdir. it should work but I cannot remember which version I tried last time ;)
<naobsd> it's free, but if you can please don't distribute it yet, people will complain "README.txt must be placed outside of zip" "wiki is best place" "write more" "easy 1step" "where is source code" etc etc etc
<naobsd> ah, it's not only for Radxa Rock
<naobsd> it should work on any RK3188 (I hope)
<naobsd> about encryption, see uboot/tools/rk_tools/, or https://gist.github.com/sarg/5028505
<naobsd> only openssl and split commands are needed
<naobsd> image format of logical partitions are same as NAND images
<naobsd> i.e. rkcrc -p for parameter, rkcrc -k for kernel/boot/recovery or mkbootimage for boot/recovery
<naobsd> oh
<naobsd> karlp: you tried sdboot on rk3066?
<karlp> man, taht list of sectors, could you jus tmake a page for that on linux-rockchip.info?
<karlp> I don't knwo what the magic header was, but I had my rk30 loader.bin for rk3066, just couldn't work out how to combine all the parts onto an sdboot.
<karlp> will try it later, but bedtime now. (will re-read your comments in the morning too)
<naobsd> I _always_ want to write wiki
<naobsd> but questions someone asked to me is too many
<naobsd> everyone say "write wiki"
<naobsd> but no one say "ok, I understood/confirmed what naobsd said, let's write wiki"
<naobsd> on IRC I post info at random
<naobsd> on wiki (I think) info should be arranged well