mnemoc changed the topic of #arm-netbook to: EOMA: Embedded Open Modular Architecture - Don't ask to ask. Just ask! - http://elinux.org/Embedded_Open_Modular_Architecture/EOMA-68 - ML arm-netbook@lists.phcomp.co.uk - Logs http://ibot.rikers.org/%23arm-netbook or http://irclog.whitequark.org/arm-netbook/ - http://rhombus-tech.net/
mysteryname has quit [Ping timeout: 240 seconds]
penguin42 has quit [Quit: Leaving.]
mysteryname has joined #arm-netbook
cncman has quit [Quit: BitchX-1.1-final -- just do it.]
Guest96895 is now known as _hp_
_hp_ is now known as hp__
tekzilla has quit [Ping timeout: 248 seconds]
tekzilla has joined #arm-netbook
QingPei has joined #arm-netbook
QingPei1 has joined #arm-netbook
QingPei1 has left #arm-netbook [#arm-netbook]
QingPei has quit [Ping timeout: 245 seconds]
tuliom has quit [Quit: Konversation terminated!]
Quarx has joined #arm-netbook
ceo16 has joined #arm-netbook
Marex has quit [Ping timeout: 260 seconds]
asus_ has joined #arm-netbook
ceo16 has quit [Ping timeout: 246 seconds]
asus__ has joined #arm-netbook
asus_ has quit [Ping timeout: 272 seconds]
asus__ has quit [Ping timeout: 252 seconds]
asus__ has joined #arm-netbook
tzafrir_laptop has quit [Ping timeout: 260 seconds]
asus__ has quit [Ping timeout: 248 seconds]
cat1 has quit [Ping timeout: 248 seconds]
asus__ has joined #arm-netbook
cheng has joined #arm-netbook
<cheng> anyone manage to get i2c work in meleA1000?
cat1 has joined #arm-netbook
<hno> cheng, yes
<cheng> hno: from rhombus-tech, i need to add R14/R15(jumper) to enable it?
<cheng> what i2c-line is that on the eeprom?
<cheng> as i saw it have i2c-0/1/2
<hno> no idea. have only tried talking to the pmu which is i2c connevted.
<cheng> that's the i2c device already on the melea1000?
<hno> and nowdays playing more with cubieboard & olinuxino where I/O functions are conveniently exposed on expansion pins.
<hno> cheng, yes.
avernos has quit [Ping timeout: 246 seconds]
<cheng> ok, thanks. i've work with what i have till i manage to get a cubieboard
<hno> If you don't need HDMI then http://linux-sunxi.org/A13-OLinuXino is very nice for hacking.
<cheng> too bad i need HDMI
<cheng> would it be useful if i put device that i found on meleA1000 on net for other reference?
<hno> cheng, please expand http://linux-sunxi.org/Mele_A1000 with your findings.
<hno> what did you find?
<cheng> i'm trying to scan any i2c device available on the board,
<cheng> so far seems like on i2c-0, there is a device 0x34 (not too sure)
<hno> 34 is the pmu
gimli has joined #arm-netbook
<cheng> but from the PMU(AXP209) datasheet, the device address is 0x68
<hno> That's because they include the read/write bit in the address there.
<hno> 0x68 >> 1 == 0x34.
<cheng> aiks...silly me..thanks..
<hno> Hm. why is so little info on A1000 on linux-sunxi? http://rhombus-tech.net/allwinner_a10/hacking_the_mele_a1000/ has plenty more info.
<hno> there. link added.
<cheng> cnx-software also have plenty of info on this device, i learn most stuff from there.
rellla has joined #arm-netbook
popolon has joined #arm-netbook
tzafrir_laptop has joined #arm-netbook
<discopig> sunxi is nice
MMlosh has quit [Quit: Bye...]
merbzt has joined #arm-netbook
avernos has joined #arm-netbook
avernos has quit [Changing host]
avernos has joined #arm-netbook
rellla1 has joined #arm-netbook
merbzt has quit [Ping timeout: 265 seconds]
cheng has quit [Quit: Leaving]
rellla1 has quit [Ping timeout: 255 seconds]
Marex has joined #arm-netbook
<oliv3r> anybody tried flashing anything with livesuit for linux? i keep getting the error 'invalid image'
<oliv3r> do I need to copy some 'key' file or something frmo the windos version from the oem?
<oliv3r> (the module is patched and loads fine as per patch i put on the ML)
rellla has quit [Quit: rellla]
ol1ver has quit [Quit: ol1ver]
oliv3r has quit [Quit: Leaving.]
rsalveti has quit [Ping timeout: 245 seconds]
rsalveti has joined #arm-netbook
oliv3r has joined #arm-netbook
ibot has quit [Ping timeout: 246 seconds]
ol1ver has joined #arm-netbook
avernos has quit [Remote host closed the connection]
eFfeM has joined #arm-netbook
penguin42 has joined #arm-netbook
eFfeM has quit [Quit: Leaving.]
ibot has joined #arm-netbook
dw4rf has joined #arm-netbook
Yakuzzi has joined #arm-netbook
<Yakuzzi> hey there, i just saw a good offer i think
<Yakuzzi> when i understand right there is a high propability that xbmc will run on it ?
<Yakuzzi> due to the amlogic chipset ?
QingPei has joined #arm-netbook
dw4rf has quit [Quit: This computer has gone to sleep]
dw4rf has joined #arm-netbook
<mnemoc> cat1: hi, does http://sprunge.us/hSDE let use rt8192cu on 3.4 on runtime?
avernos has joined #arm-netbook
avernos has quit [Changing host]
avernos has joined #arm-netbook
<techn> mnemoc: what you think if we merge disp branch.. And leave code which causes green screen to #if 0 ?
mano has joined #arm-netbook
Yakuzzi has quit [Remote host closed the connection]
Yakuzzi has joined #arm-netbook
asus__ has quit [Ping timeout: 244 seconds]
asus__ has joined #arm-netbook
mysteryname has quit [Ping timeout: 265 seconds]
<mnemoc> techn: if you have it narrowed, and works fine. sure
popolon has quit [Quit: Quitte]
dw4rf has quit [Quit: This computer has gone to sleep]
<oliv3r> anybody tried the linux livesuit yet? I kept getting an error about invalid image, where as it worked fine in the windos version
<mnemoc> it might rely in a newer packer tool. try to repack it in linux too
<oliv3r> i used the one supplied by the oem of the tablet
<oliv3r> is it 'safe' to unpack and repack?
<mnemoc> i suppose
popolon has joined #arm-netbook
avernos has quit [Quit: Leaving]
tuliom has joined #arm-netbook
tuliom has quit [Client Quit]
Yakuzzi has quit [Remote host closed the connection]
Yakuzzi has joined #arm-netbook
mikey_w has quit [Remote host closed the connection]
n6pfk has quit [Read error: Connection reset by peer]
mikey_w has joined #arm-netbook
n6pfk has joined #arm-netbook
QingPei has left #arm-netbook [#arm-netbook]
wingrime has joined #arm-netbook
<wingrime> mnemoc: You can see patches that make work ts on my tablet https://github.com/wingrime/linux-allwinner
<wingrime> but it need some Kconfig stuff...
<wingrime> also I now fixed driver re insmod
<wingrime> not make push yet
Yakuzzi has quit [Remote host closed the connection]
Yakuzzi has joined #arm-netbook
<mnemoc> wingrime: send them to the mailing list please
<wingrime> push requests is possible ?
<mnemoc> ML preferred
<mnemoc> so the rest of the devs can comment
Yakuzzi has quit [Remote host closed the connection]
<wingrime> give me link to you maillist
Yakuzzi has joined #arm-netbook
<wingrime> ok I make some cleanup and think about it
pwhalen has quit [Ping timeout: 245 seconds]
n6pfk has quit [Remote host closed the connection]
mikey_w has quit [Remote host closed the connection]
pwhalen has joined #arm-netbook
mikey_w has joined #arm-netbook
n6pfk has joined #arm-netbook
<techn> wingrime: have you tried latest codes.. there is already fix for some ts problem
<techn> atleast it helped drachensun
<cat1> mnemoc: on rt8192cu -- never tried, I am using rtlwifi.
rzk has joined #arm-netbook
<cat1> mnemoc: though some dirty hacking is required there as well..
<mnemoc> it seems the only problem is that sanity check over a list which doesn't exist in 3.4's struct wake_lock
<wingrime> techn:what commit?
rz2k has quit [Ping timeout: 245 seconds]
<wingrime> techn: one bug fixed in goodix_touch but not in ft5x_ts
<wingrime> +#define CTP_IRQ_NO (gpio_int_info[0].port_num)
<wingrime> techn: see my github, dirty fix for now, but it working
<mnemoc> err
<wingrime> I have ts on INT11
<mnemoc> it seems I screwed up the merge
<wingrime> but it not enough to make it working on my HW
<wingrime> now I will look for resume/suspend bug with ts
<mnemoc> wingrime: but please, to the ML
<techn> wingrime: could you try if #define CTP_IRQ_NO (gpio_int_info[0].port_num) helps you.. changing hardcoded 21 to 11 is just moving the bug :/
<wingrime> I know this
<mnemoc> is http://sprunge.us/JdCY?diff the only missing part?
<wingrime> not only for make it work for me
<mnemoc> can you compare it with the driver in the import-sunxi branch?
<wingrime> I now make git clone for new place of code
<mnemoc> also, make the commit mergable. not C++ comments, no "signatures", ...
<mnemoc> no hardcoded values
<mnemoc> but please compare it with the driver in import-sun5i branch to see what else I missed
<mnemoc> now I can't look deeper on that
<wingrime> ok, just wait
<wingrime> also, WRN:L2357(drivers/usb/sun4i_usb/udc/sw_udc.c):ERR : dev->driver=0xc0790e8c, dev->gadget.speed=0
<wingrime> have you this flood?
<wingrime> menmoc: you lost only interrupt http://pastebin.com/fRTcWwKG
<wingrime> please fix this in git and I make patch that fix second bug
<mnemoc> wingrime: ok. I'll commit the paste I showed before
toxicpsion has joined #arm-netbook
<wingrime> in my plaste reversive patch as you see
<wingrime> it need inverce
<wingrime> prefect
tzafrir_laptop has quit [Ping timeout: 246 seconds]
<mnemoc> wingrime: done
<wingrime> thanks
<mnemoc> thanks you for reporting it
<wingrime> send patch over ML or pool-req ?
<mnemoc> ML
<wingrime> ok
tzafrir_laptop has joined #arm-netbook
MMlosh has joined #arm-netbook
Yakuzza321 has joined #arm-netbook
Yakuzzi has quit [Ping timeout: 245 seconds]
Quarx has quit []
mano has quit [Quit: irc2go]
MMlosh has quit [Quit: Bye...]
<techn> mnemoc: I figured out how to fix that problem.. much better solution ;)
<mnemoc> \o/
<Yakuzza321> does anyone got experience how easy it is to use an external bluetooth dongle on a mele with stockrom ?
<Yakuzza321> * stockrom ics 1.2
rsalveti has quit [Read error: Connection reset by peer]
<mnemoc> techn: out_csc = 2; ????
<mnemoc> just that?
rsalveti has joined #arm-netbook
<techn> mnemoc: yes :)
<techn> :D
simosx has joined #arm-netbook
merbzt has joined #arm-netbook
dw4rf has joined #arm-netbook
wingrime has quit [Ping timeout: 255 seconds]
<hno> Yakuzza321, no idea if bluetooth is at all enabled in Mele stock Android rom.
lkcl has quit [Ping timeout: 245 seconds]
<Yakuzza321> which rom is recommendable at the moment, i have only one device and actually want to use it at an tv with cvbs input, so does there happen to be a rom where there might even be a uinput.ko for droidmote and bluetooth enabled ?
<specing> what is a rom?
<specing> Yakuzza321: ?
<Yakuzza321> the system image
<specing> no.
<specing> try again. What is a rom?
<Yakuzza321> you mean read only memory or what you mean ?
<specing> That would be the correct definition
<Yakuzza321> but i used it interchangebly with the image i´d flash on the device
<specing> And with this in mind, your question does not seem logical
<specing> Why can't you just ask "which system image ...", then?
<Yakuzza321> good point, is there a system image in which almost all output features of my mele are covered ?
<Yakuzza321> sata,vga,cvbs,ethernet
merbzt has quit [Ping timeout: 245 seconds]
dw4rf has quit [Quit: Leaving]
lkcl has joined #arm-netbook
<hno> Yakuzza321, outputting what?
rsalveti has quit [Read error: No route to host]
rsalveti has joined #arm-netbook
tzafrir_laptop has quit [Ping timeout: 252 seconds]
<toxicpsion> while on topic of mele... anyone figure out what unpopulated 13J1 (4pin by SATA) is connected to on a2000?
<toxicpsion> my guess is one of the a10 USB hosts (since all 3 external USB come from GL850G)... opinions?
<mnemoc> the otg header is behind the ethernet. and the other native usb host goes to the wifi
<mnemoc> and those are the 3 native usbs the a10 has
rsalveti_ has joined #arm-netbook
rsalveti has quit [Ping timeout: 248 seconds]
rsalveti_ is now known as rsalveti
<hno> toxicpsion, you mean the connector between sata and vga?
<hno> my guess is something audio related.
<hno> but have not traced it at all.
penguin42 has quit [Quit: Leaving.]
<hno> toxicpsion, the fourth USB port from the HUB is the 6 pin header between SD & USB on the side.
<hno> 12J4 is also unidentified.
<toxicpsion> yea... i forgot about the wifi using one....
<toxicpsion> where's 12j4? my eyes are missing it :|
<hno> 2 pin, next to SD
gimli has quit [Quit: Verlassend]
<toxicpsion> connected to 17R6(unpopulated)....
<hno> missing a resistor as well.'
<oliv3r> i'm comparing christian troy's script.bin (wherever he may got it from) to the one the oem supplied; found some differences in values that is not on the fex guide wiki. things plike pll4/6_freq, storage_type, [card_burn_para]
<toxicpsion> any idea what value that resistor should be? i'll throw a logic analyser on it.
<oliv3r> anybody got a more detailed list of explainations? (i haven't started going through the file further then the top two sections, was going to look why I can't get into recovery mode with christian troy's script.bin any more (which seems odd) (this all for the momo9 clone)
<hno> ppl frequencies are pretty obvious.
<hno> those are clock frequencies for different I/O functions. And there is several ways to wire those two.
Yakuzza321 has quit []
<hno> I think storage_type is telling livesuit if it should use NAND or SD. Not sure.
<oliv3r> I assume that since script.bin is generated by livesuit that one should be concidered 'best' for my board? anyhow, those where as an example, there's many more script.bin values that i don't know what should be used
<hno> script.bin is generated in the image build process. Don
<oliv3r> the entire [card_burn_para] section is 'new' with things like card_no, card_line, card_mode, sdc_d1 sdc_d0, sdc_clk, sdc_cmd, sdc_d?
<hno> Don't think livesuit changes it.
wingrime has joined #arm-netbook
<hno> Those are SD interface parameters for flashing to SD/eMMC.
<oliv3r> ah, i thought livesuit dynamically sets up some things (like ram timings? i forgot allready whattt exactly it did)
<oliv3r> 'card burn parameters' does make sense :) anyway, was just seeking an accurate list
<hno> It does, but not in script.bin. Those parameters are stored in boot0 & boot1.
<oliv3r> i'll add the valueees i find that aren't in the wiki to the wiki
<oliv3r> and someone more knowledgable hopefully can enter them :)
<hno> mnemoc, do you have the A13 fex guide?
<oliv3r> i'll probably use the OEM script.bin as that one atleast allows me to get back into recovery mode :)
<oliv3r> i was looking at the http://linux-sunxi.org/fex-guide
<oliv3r> s/fex-guide/Fex-Guide/
<ibot> oliv3r meant: i was looking at the http://linux-sunxi.org/Fex-Guide
<mnemoc> hno: the only difference i've noticed in A13 .fex files is the usage of port:power[0123] to refer to the AXP GPIOs
<mnemoc> we have a short datasheet, but no "guides"
<toxicpsion> hno: what's 11J1 (14pin @ front)?
<mnemoc> jtag?
<oliv3r> i'm actually just looking at [jtag_para] and if your really good, you could trace them back to PB14 PB15 PB16 and PB17 :p
<simosx> I want to get mkbootimg for Ubuntu. Is https://gist.github.com/1087757 the only way?
<mnemoc> simosx: that comes in the android sdk
<simosx> mnemoc, it's part of the utils like adb and so on? that would be nice.
<mnemoc> i get adb and fastboot from a ppa from linaro
<mnemoc> but it doesn't include mkbootimg
tzafrir_laptop has joined #arm-netbook
<hno> toxicpsion, JTAG.
<oliv3r> hmm, i don't know where this guy got his scripts.bin from, but it almost looks like he randomly enabled all sorts of thigns, or got it form a completly different device
<oliv3r> it's supposed to be for tablets, but things like sata ports, ir ports etc etc are all enabled
<hno> oliv3r, most script.bin:s are quite crappy.
<oliv3r> yeah, i guess the one that is most trustworthy would be the vendor supplied one, right?
<oliv3r> even though it's crappy :p
<mnemoc> they just tweak the template
<hno> they take the EVB one and modify it slightly for their device.
<oliv3r> how important is pmu_battery_rdc/cap?
<oliv3r> i think the android folks, 'chefs' they call them these days, just use a script.bin that comes from some device i guess
<oliv3r> anyway, those battery values, the resitance and capicitance should come from 'some' test. I guess I can't test it so i'll just have to hope the oem put some sane values there :)
<hno> The battery parameters should match the battery for charging and charge level to work right.
<oliv3r> thats what i mean, unless i have some verified sources, the vendor values are the most likly ones
<hno> yes
<oliv3r> hmm, why does the wiki say: 'please read the documentaiton of the cell parameter test instruction)
<hno> Because it's a copy-paste document.
<oliv3r> ohh
<hno> and annoyingly lacks reference to which chinese version the wiki one is based on.
<simosx> mnemoc, I could not find 'makebootimg' in the Android SDK. Continuing with source. I need makebootimg to create images for the linux-allwinner source.
<oliv3r> :S
<mnemoc> simosx: wasn't it `mkbootimg` and not `makebootimg` ?
<hno> there is a document on how to test the battery to get some of the parameters.
<akaizen> how did empat0 get all those cedarx docs?
<oliv3r> well, i went through script.fex but find nothing obvious as to why I can't enter recovery mode when flashing this stuff :S well, lets find out if the original script.bin works well :D
<mnemoc> hno: maybe it was there initially and got "lost" on wikification?
<mnemoc> in which case it would be in the history
wingrime has quit [Ping timeout: 272 seconds]
<simosx> mnemoc, sorry, you are right, it's 'mkbootimg' (not available in the Android SDK).
<hno> simosx, http://android-dls.com/wiki/index.php?title=HOWTO:_Unpack%2C_Edit%2C_and_Re-Pack_Boot_Images
<hno> another.
* hno feels yet another NIH syndrome in Android here.. Why couldn't they just use one of the existing multipart boot image formats?
<oliv3r> so what is repsonsible for entering recovery mode, that could be overwritten with some android rom (which shouldn't write more then /system, /bootloader (nande iirc)
<RaYmAn> There's also a 'standalone' variant of mkbootimg + a boot.img unpacker https://github.com/androidRoot/boottools
<hno> oliv3r, u-boot iirc. But not playing with Android.
<oliv3r> i just wanna have android 'properly' running on this tablet before playing with booting from uSD and more :)
<oliv3r> just something decent to fallback on, as the vendor firmware is ... well really crap; and i don't trust vendor binaries :)
<RaYmAn> oliv3r: on most devices, bootloader checks for key combo and for boot-recovery written to MISC partition (seperately)
<oliv3r> and the bootloader (u-boot I guess?) checks the bootloader partition for extra info? (/dev/block/nanda)
<oliv3r> so if the vendor firmware goes successfully into recovery mode, then flashing 'a' android rom 'that doesn't mess with /dev/nanda' should still mean that recovery mode can be entered normally?
<oliv3r> i really have to read up on the whole android thing. where the kernel resides, how partitions are laid out
<RaYmAn> the kernel+ramdisk resides on the two boot.img partitions
<RaYmAn> iirc nandc and nandg
<RaYmAn> nanda is iirc more for boot0/boot1/spl usage?
<oliv3r> i would assume so
<oliv3r> it contains script.bin among other thigns
<oliv3r> some drv_de.drv, font files, 'magic.bin'
<oliv3r> boot.axf and sprite.axf, boot.ini
<oliv3r> some dirs
<oliv3r> linux/u-boot.bin (THE uboot?)
<oliv3r> linux/some bitmaps, linux/linux.ini
<RaYmAn> yeah
<oliv3r> ok check this
<oliv3r> /dev/block/nanda/linux/linux.ini -> img_name = c:\linux\u-boot.bin
<oliv3r> also logo_name = c:\linux\linux.bmp
<oliv3r> how funny is that
<RaYmAn> yeah *rolls eyes*
<RaYmAn> allwinner-isms
<oliv3r> is this used by something else? or are they retranslating those \'s internally
<RaYmAn> it's boot0/boot1 that uses those inits afaik
<RaYmAn> It appears to just be symbolic naming, e.g. c: -> nanda
<oliv3r> so boot0/boot1 does some magic to translate c:\linux\u-boot.bin to nanda\linux\u-boot.bin?
<oliv3r> well i assume boot0/boot1 is some lowlvl C code, probably using very basic 'open()' or maybe fopen() functions.
<oliv3r> i haven't programmed for ages in win32, but i think even there you supply either both, or unix slashes?
<oliv3r> in any case, they have to actually code those translations?
ZaEarl_ has joined #arm-netbook
ZaEarl__ has quit [Read error: Connection reset by peer]
<oliv3r> also, and thid does actually quite bafle me; there's more interesting images in nanda\os_show. a 'melis[12].bmp' probably THE original template for the firmware? and wince[12].bmp!
<oliv3r> winCE out of all things!
<mnemoc> techn: can't c99-style struct initialization be used instead of each fb_para.foo one by one? { .foo = 1, ... } will zero anything else. with fb_para.foo you can always miss something
ZaEarl__ has joined #arm-netbook
ZaEarl_ has quit [Read error: Connection reset by peer]
<techn> mnemoc: true.
<mnemoc> that way you can also omit all those fields happy been 0
<mnemoc> techn: but I can merge the fixed version, and wait for the cleanup next
<techn> yep.. fix/cleanup will go to tomorrow
<mnemoc> i mean. do i merge what I've got now, or wait?
<techn> wait.. I'll fix that tomorrow :)
<mnemoc> ok
<oliv3r> assuming i'd use vanilla linux on the device (with sunxi support obviously) how 'free' am I in partitioning choice? I assume nanda is required (by uboot/boot[01]), but what about nandb? nandc? nandc is for the kernel yes? Does it always _have_ to live there? I want to write up a quick partitioning table on the wiki
<mnemoc> depends on the u-boot *file* in nanda
<mnemoc> by default it reads the env from a partition and the kernel from another
<mnemoc> but you can change it's env to load the kernel from nanda too
<mnemoc> for the env itself, not sure what relocation options you have
<oliv3r> so if i build my own u-boot, that directly loads my kernel from somewhere (must be a partition?) I could thus have 1 big partition for the rest?
<oliv3r> well still working on understanding the partitions i have here
<oliv3r> i see bootloader, env, boot, system, misc, recovery, cache, databk, UDISK atm for example
<mnemoc> I use the first block of http://linux-sunxi.org/User:Alejandro_Mery/Notes to make nanda's u-boot load my test kernel
<oliv3r> bootloader is for uboot/boot[01] i gather, env? boot probably holds the kernel and ramdisk; system obviously androids /system, misc, no idea, recovery is obviously CWM/recovery part, cache androi'ds cache, databk, not sure, i guess /data though; and UDISK would be the internal mass storage
<oliv3r> *reading*
<oliv3r> oh short note list :p
<mnemoc> boot0/boot1 live outside the logical nand space, you can't access them
<mnemoc> nanda is the vfat boot partition where u-boot lives in file form
<mnemoc> env is a raw partition dedicated to u-boot's env
<mnemoc> iirc the kernel is raw in nandd, but don't remmeber
<mnemoc> but that can be fixed changing some lines in u-boot's env
<mnemoc> like on that "note"
<Turl> misc is a partition to signal stuff to the boot chain
<Turl> eg 'boot in recovery mode' or 'boot into fastboot mode'
<oliv3r> yeah you told me a few times (sorry for asking so many questionns here!) but i ment boot0 data, e.g. battery1.bmp etc
<oliv3r> u-boot.bin IS u-boot right? basically, if BROM is the 'bios', nanda/linux/u-boot.bin is the MBR (to relate it to the PC way of thinking)
<hno> battery1.bmp is not used by boot0. It's used by boot.axf I think.
<oliv3r> actually, i'd think BROM is somethiiing before the bios, and boot0/1 is the bios
<hno> There is no BIOS as such.
<oliv3r> nono, i know, i'm just trying to relate it to my knowledge on how pc boots
<hno> BROM i the boot rom embedded in the CPU, which loads boot code from SD0/NAND/SD2/SPI and also does FEL USB mode.
<oliv3r> i guess any PC chip has some form of BROM that starts to execute the bios; so the BROM in the A10 executes boot0/1 which in turn run u-boot.bin from nand aa
<oliv3r> yeah :) I wrote the page! (on the wiki) :p
<mnemoc> here a "bios" is chomped in different apps
<mnemoc> brom, (boot0/boot1) or SPL, u-boot, and linux itself
<oliv3r> so boot0/boot1 reads the config file from nanda, which tells it where the bootloader lives (c:\linux\u-boot.bin) and takes it from there
<hno> BROM loads boot0 / u-boot-spl. boot0 / u-boot-spl configures DRAM etc and then loads boot1 / u-boot. In case of boot1 it then loads script.bin, configures more of the hardware and loads boot.axf. boot.axf then loads boot.cfg (or whatever it's name) which tells it to load u-boot.bin.
<oliv3r> then u-boot reads data from nandb (env)
<mnemoc> nanda's u-boot. yes
<oliv3r> so where recides uboot-spl?
<mnemoc> our u-boot reads that from before the first partiton
<oliv3r> hno: boot.ini probalby
<hno> u-boot-spl is the u-boot equivalence of boot0.
<mnemoc> u-boot spl at 8*1024 of the mmc
<oliv3r> ah duh, of course, i should have known that, partition-table, uboot-spl, partition0 with uboot in it
<hno> we have two different boot loader chains depending on which bootloader you use.
<oliv3r> i'll let that all sink in a little more, and maybe write a page on the wiki about the boot chains
<hno> allwinner boot0/boot1 or u-boot-spl.
<oliv3r> and right now, we can only boot u-boot-spl via SD
<hno> yes.
<oliv3r> got it
<hno> and u-boot-spl also do not know as much about the PMU and possibly other parts as boot1, but is all free software with full source available.
<oliv3r> yeah, mnemoc told me a while ago, uboot-spl isn't as far as boot0, but should get there in time
<hno> it should be pretty up to spee with boot0, but not boot1.
<mnemoc> it's very close these days
<hno> have not cared about batteries yet.
<hno> none of my devices have a battery.
<oliv3r> so i get nanda (uboot.bin and datafiles for boot*etc), nandb (env for uboot), nandc (boot, holds the kernel + ramdisk) then, nandd (system which is 'android' or whatever you configure it as, followed by nande (misc, which turl explained above, not sure why there's a partition required for this and how this works at all yet) and then the rest
<oliv3r> i'll wikify that info as best as i can
<hno> Plesae do.
<hno> Another note about A10 documentation. It is imho not OK to copy material from the A10 User Guide. It's a restricted document.
<oliv3r> one more thing, i have a boot.img to be written directly into nandc, it starts however with ANDROID! as it's magic signature
<hno> boot.img is an android thing.
<hno> u-boot uses other image formats.
<oliv3r> in this case, it's what gets dumped to nandc
<hno> In an Android setup yes.
<oliv3r> ah ok
<hno> You can repartition the NAND as you please.
<hno> and also load kernel + initrd as u-boot images from files.
<oliv3r> well a lot of A10 hardware usualy ocmes pre-installed with android, so some notes on default partition layout can be usefull
<hno> just remember to keep "boot" partition reasonably intact if using boot0/boot1.
<hno> and a smallish "env" partition if using allwinner u-boot. Beyond that partition as you please.
<oliv3r> makes sense
<oliv3r> i'll write a minimal partitioning chapter with that info sometime tomorrow
<oliv3r> hmm, somehow a letter got lost in the alphabet. i'll have to dig deeper into that :S
<mnemoc> Android/Partitions ?
<oliv3r> i know for a fact that nandg is the recovery partition on my device. i've written CWM to it several times. however dmesg shows 'The 5 disk name = recovery' which is nandf
<oliv3r> ah, better idea
<hno> idea?
<markvandenborre> oliv3r: you may also want to talk to Dimitar Gmishev
<markvandenborre> he has managed to get Debian running from nand on an a13 based board
<oliv3r> Android/partitions for the wiki location :p
<simosx> mnemoc, hno: I eventually used a standalone repo (the one from github given above) to compile mkbootimg. Thanks to both.
<mnemoc> if you are going to talk about android/sunxi partiotions, yes
<mnemoc> simosx: quarkx's?
<oliv3r> mnemoc: did and done :)
<mnemoc> :)
tuliom has joined #arm-netbook
<mnemoc> ok
<oliv3r> great scots! i figured out why CWM and CM10 won't boot. I think somewhere, someone somehow swapped recovery and /data partitons, so my /data is only 15MiB! atleast i know what to fix now :D
<hno> oliv3r, what does u-boot say? Mine says http://fpaste.org/3sSA/
<oliv3r> but taht's not from u-boot
<oliv3r> that's from dmesg/kernel
<hno> this one is from u-boot.
<hno> but probably the same in dmesg. NAND driver is mostly the same in u-boot and kernel.
<mnemoc> the filesystem can be smaller
<hno> mnemoc, what?
<mnemoc> in the android images they generate filesystem images of a given size
<mnemoc> and then dump it
<hno> yes.
<mnemoc> ignoring the real partition size
<mnemoc> as a result you can have a 16MB filesystem in a 2GB partition...
<hno> not sure the data partition is included in images. But maybe it is.
<Turl> oliv3r: did you device come with gingerbread on it?
<Turl> oliv3r: I recall reading GB A10 devices had a different partition layout
<oliv3r> it came with android 4.0.3
<oliv3r> and they have an update for android 4.0.4 (both livesuit images)
<oliv3r> but they changed something with the partitions, not only size, but also order
<Turl> most A10 devices have that layout on the link there ^
<oliv3r> i'm installing 4.0.3 to make a quick dmesg dump of partitinos, then install 4.0.4 stock to do the same to compare
<oliv3r> gimme 10 minutes and i'll give you a paste of both what i have :)
<oliv3r> but they deffo changed the sizes
bsdfox\ has quit [Ping timeout: 240 seconds]
<oliv3r> i'll fpaste it tomorrow :p
<oliv3r> bed time now :)
<mnemoc> hno: btw, my cubie still halts on spl on reboot. yours doesn't?
<hno> Don't boot the cubie that often. Usually test stuff on the A13 via JTAG.
<hno> does it also halt on reboot from u-boot? Or only from kernel?
<hno> oliv3r, remember to select to format the flash. If not existing partitioning layout is kept which may give unexpected results.
<hno> mnemoc, does it print the SPL header and then silently dies?
<mnemoc> it used to die on MMC:
<mnemoc> now it's after the greeting
<mnemoc> i suppose still on the same place
<mnemoc> just that boot device detection code changed
<hno> Many small things changed. Now the greeting is very early, even before DRAM.
<hno> How reproducible is it?
<mnemoc> 100%
<mnemoc> in the mele it's like 50%
<hno> every time you reboot?
<mnemoc> yes
<hno> Same SD card in both?
<mnemoc> no, one SD per device, one ttl/usb dongle per device
<hno> Hm. You don't get MMC: now?
<Turl> fuu, forgot to merge next_mali when building with cpufreq for zatab :P
<mnemoc> hno: only: U-Boot SPL 2012.10-rc3-04244-gfa4a874-dirty (Oct 13 2012 - 16:47:06)
<mnemoc> Turl: does next_mali work well in android?
<Turl> mnemoc: yeah, I've been using it since I created the branch :x
<mnemoc> :)
<Turl> that's like 2 months ago *hint* :P
<mnemoc> Turl: not my call
<Turl> techn: what's holding next_mali from merging?
<hno> Hm... odd, I don't get MMC: either.
<Turl> hm, wtf
<Turl> it still doesn't work and on boot the screen does all weird glitches
<Turl> with the uboot(?) splash
<hno> mnemoc, so yes, it's most likely dying at the same place as before. Just different output.
<hno> hadn't noticed the MMC: output is removed in rc3.
<mnemoc> good thing is that same happens (not 100% but often) on the mele too. so jtag could be used ;-)
<Turl> nevermind, I suck
<Turl> I pushed the kernel and not the modules ~.~
<mnemoc> Turl: :)
<mnemoc> make an script to do so for you. write once, works always
<Turl> there, it works
<Turl> it still does the fugly glitches though
<Turl> wonder what's causing those
<Turl> they weren't there before
<Turl> hno: the cpufreq stuff is on my github if you want to test/review before I send it to the ML
<Turl> I need to clean up the patches and organize it better, and I noticed I commited a defconfig by mistake too together with some other stuff
<hno> mnemoc, try enabling SUNXI_MMCDBG in drivers/mmc/sunxi_mmc.
<hno> .c
<mnemoc> http://dpaste.com/813681/ <--- first boot
<hno> mnemoc, isn't that from u-boot and not spl?
<mnemoc> oops
<hno> Should be mostly the same, but...
<mnemoc> http://dpaste.com/813683/ <--- spl, first boot
mysteryname has joined #arm-netbook
ZaEarl_ has joined #arm-netbook
ZaEarl__ has quit [Read error: Connection reset by peer]
ibrah has joined #arm-netbook
<mnemoc> hno: and early death on reboot
<mnemoc> U-Boot SPL 2012.10-rc3-04245-g77f7322-dirty (Oct 15 2012 - 00:36:44)
<mnemoc> [mmc]: init mmc 0 resource
<mnemoc> but i have to go to take a nap know. good night
ZaEarl_ has quit [Ping timeout: 256 seconds]
ZaEarl_ has joined #arm-netbook
<hno> mnemoc, that's odd. There is no I/O until after "init mmc 0 clock and io". Only preparation of data structures.
popolon has quit [Remote host closed the connection]
<hno> But hmm... that data is in DRAM not SRAM. Wonder if that's a good idea.
itamarjp has joined #arm-netbook
itamarjp has joined #arm-netbook
itamarjp has quit [Changing host]
mysteryname has quit [Ping timeout: 245 seconds]
<Turl> https://plus.google.com/u/0/+Linux/posts/iBvjvyovDCA arm multiplatform on rc1 :)
bsdfox has joined #arm-netbook
bsdfox has quit [Changing host]
bsdfox has joined #arm-netbook