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/
L84Supper has quit [Ping timeout: 255 seconds]
L84Supper has joined #arm-netbook
rz2k has quit []
mikey_w has quit [Read error: No route to host]
n6pfk has joined #arm-netbook
NickNack has joined #arm-netbook
<NickNack> i have a question
<NickNack> im not sure if this is the right place to ask
n6pfk has quit [Quit: Leaving]
mikey_w has joined #arm-netbook
von_fritz has quit [Quit: vonfritz leaves, don't panic]
<mnemoc> probably the wrong place. there are no mind readers here.
NickNack has left #arm-netbook [#arm-netbook]
dyoung is now known as dyoung-away
<steev_> lol
<steev_> mnemoc: way to scare him off
ZaEarl has quit [Ping timeout: 244 seconds]
hp__ has joined #arm-netbook
ZaEarl has joined #arm-netbook
gimli has quit [Ping timeout: 250 seconds]
ZaEarl has quit [Ping timeout: 244 seconds]
voronaam has quit [Quit: Leaving.]
stefanro1 has joined #arm-netbook
stefanro has quit [Ping timeout: 244 seconds]
<countz> I need to make a modification to .config before the kernel is built, in sunxi-bsp
<countz> what's the clean way of doing this?
freakazoid0223 has joined #arm-netbook
dyoung-away is now known as dyoung
<servili007> countz: are you still around?
<countz> yes
<countz> I believe I am!
<countz> :-)
<servili007> haha, alright, here's the quick version of what you should do
<servili007> the first way is text based, not highly recommended. go into /arch/arm/configs, copy the defconfig you want to copy and make your edits, then target that instead of sun4i_defconfig or whatever you're using
<servili007> the other way is to use menuconfig in terminal to get a GUI based configuration tool
<countz> (yes, sun4i)
<countz> I like the first way actually
<countz> :-)
<servili007> If I just need to make a quick change, I do too
<servili007> Make sure you know if what you're editing is dual or tri state though
<servili007> (as in, some things can only be built as modules, no integrated completely)
popolon has quit [Quit: Quitte]
<servili007> if you don't know and want to cheat, just put that value into google and get the info page on it
<countz> I needed to modify a specific row, CONFIG_SUN4I_GPIO_UGLY=y
mikey_w has quit [Remote host closed the connection]
mikey_w has joined #arm-netbook
<countz> so you wouldn't make the modifications to the original file, you'd make a copy and change the target
<countz> I get that, it's cleaner
<servili007> you also don't want to modify .config because it's a temporary file
<servili007> and while you're making your copy, you can always name it as the device you're working on, this way everything feels more personalized ;)
<countz> right
<countz> I think that modification allows the a13 to shutdown properly
<servili007> does your device have an axp209 or similar?
<countz> it has some axp20 something
<countz> I remember that from the Xorg.0.log file
<countz> hold on i'll make sure
<countz> ok it's just axp20 :-)
<servili007> don't worry about it, as long as it's in the 20 series
<countz> by the way is this just some boot image I'm supposed to supply?
<countz> cp: cannot stat `build/boot.scr': No such file or directory
<countz> (while uboot loads the system?)
<servili007> you can add boot options to a boot.cmd and convert it to a boot.scr and stick it with your uImage and script
<countz> I see
<servili007> but if you're using u-boot, it'll auto supply some basic arguments to get you up and running
<servili007> you may need/want to add more though
<countz> must have forgotten to copy something... it panic'd just now over "init not found"...
<servili007> are you running from an SD card?
<countz> yes
<servili007> actually, give me some details on your setup, device, OS, what you're trying to do
<countz> is it possible sunxi-spl.bin and u-boot.bin are tailored for my kernel?
<countz> This is an A13, Olinuxino
<countz> with Debian Wheezy on an SD card
<countz> I previously extracted the 3.0.42+ kernel from an .img file I found, and used that
<countz> but it has the paranoid android kernel flag enabled, and I suspect it causes some issues for me with X11
<servili007> well
<servili007> on the default boot.scr
<servili007> the partition layout has to be similar to what is detailed at the bottom of - http://linux-sunxi.org/FirstSteps
<countz> this sd card is booting fine with 3.0.42+
<servili007> what branch is your linux-sunxi clone on?
<countz> how do I check?
<countz> git info
<countz> ?
<servili007> well
<servili007> when you cloned it what did you checkout
<countz> I'm on the master branch, it turns out.
<countz> (of sunxi-bsp)
<servili007> ah, okay, you're using sunxi-bsp
<countz> yes
<countz> I thought I mentioned it initially.
<countz> sorry.
<countz> oh wait, I did mention it. perhaps it wasn't very clear.
<servili007> no worries, unfortunately for you I'm not able to be much help remotely
<countz> That's something I can fix :-)
<servili007> yeah, I see it in your original message
<servili007> anyway
<countz> I can create a tunnel from this linux box to a public ip
<servili007> haha, sorry, I don't have enough free time for that at the moment, I gotta run soon
<countz> no worries I'll figure it out ;-)
<servili007> anyway, in your linux-sunxi submodule, it's a long shot, but try checking out sunxi-3.4 instead and remake your uImage
<countz> ok
mikey_w has quit [Remote host closed the connection]
<countz> how do I switch branches?
<servili007> within the linux-sunxi folder
<servili007> git checkout origin/sunxi-3.4
<countz> ok it's checking out
mikey_w has joined #arm-netbook
<servili007> good luck, I gotta run for now
<countz> thanks a lot!
<countz> good night :)
<servili007> night, and on a final note, make sure you don't mix and match gnueabihf parts to gnueabi ones
mentis has joined #arm-netbook
freakazoid0223 has quit [Quit: Leaving]
servili007 has quit [Quit: Leaving]
Turl_ has joined #arm-netbook
<Turl_> hey all :)
<Turl_> missed much these days?
mentis has quit [Quit: Konversation terminated!]
eebrah has quit [Ping timeout: 248 seconds]
aholler has joined #arm-netbook
<countz> probably not
aholler_ has quit [Ping timeout: 244 seconds]
eebrah has joined #arm-netbook
RaYmAn has quit [Read error: Operation timed out]
NAiL has quit [Read error: Operation timed out]
slapin_ has quit [Ping timeout: 246 seconds]
Thomas42_ has quit [Ping timeout: 260 seconds]
Holo_ has quit [Ping timeout: 252 seconds]
vgrade has quit [Ping timeout: 252 seconds]
uwe_mobile has quit [Ping timeout: 244 seconds]
slapin has joined #arm-netbook
Holo_ has joined #arm-netbook
Thomas42 has joined #arm-netbook
vgrade has joined #arm-netbook
jeremb__ has joined #arm-netbook
uwe_mobile has joined #arm-netbook
RaYmAn has joined #arm-netbook
Hexxeh_ has joined #arm-netbook
NAiL has joined #arm-netbook
steev__ has joined #arm-netbook
itamarjp has quit [Ping timeout: 252 seconds]
niftylettuce__ has joined #arm-netbook
hp__ has quit [Ping timeout: 276 seconds]
lerc_ has joined #arm-netbook
lerc has quit [Ping timeout: 265 seconds]
newbie has joined #arm-netbook
newbie is now known as hp__
dyoung is now known as dyoung-away
countz has quit [Quit: countz]
tallnerd1985 has joined #arm-netbook
countz has joined #arm-netbook
tallnerd1985 has quit [Quit: Colloquy for iPad - http://colloquy.mobi]
joekoocc1_ has joined #arm-netbook
gimli has joined #arm-netbook
antonl has joined #arm-netbook
mikey_w has quit [Remote host closed the connection]
antonl has quit [Ping timeout: 252 seconds]
mikey_w has joined #arm-netbook
Turl_ has quit [Quit: Page closed]
rellla has joined #arm-netbook
svp has joined #arm-netbook
L84Supper has quit [Ping timeout: 250 seconds]
joekoocc1_ has quit [Quit: Ex-Chat]
L84Supper has joined #arm-netbook
eebrah has quit [Remote host closed the connection]
niftylettuce__ has quit [*.net *.split]
Gumboot has quit [*.net *.split]
ssvb has quit [*.net *.split]
buZz has quit [*.net *.split]
niftylettuce__ has joined #arm-netbook
buZz has joined #arm-netbook
Gumboot has joined #arm-netbook
ssvb has joined #arm-netbook
RITRedbeard has quit [Ping timeout: 272 seconds]
Sv has joined #arm-netbook
RITRedbeard has joined #arm-netbook
Sv has quit [Client Quit]
Sv has joined #arm-netbook
Quarx has joined #arm-netbook
Sv has quit [Read error: Connection reset by peer]
Sv has joined #arm-netbook
Sv has quit [Client Quit]
Sv has joined #arm-netbook
Sv has quit [Client Quit]
Jef91 has quit [Remote host closed the connection]
Sv has joined #arm-netbook
<slapin> hno: interesting, there should be something to do about it... have you checked with LA what it does during such state?
<slapin> hno: also, please answer about bits
<hno> slapin, the LA does not say what the controller internals is doing.
<hno> which bits are unclear?
von_fritz has joined #arm-netbook
rsalveti_ has joined #arm-netbook
rsalveti has quit [Ping timeout: 272 seconds]
rsalveti_ is now known as rsalveti
von_fritz has quit [Quit: vonfritz leaves, don't panic]
steev_ has quit [Remote host closed the connection]
Hexxeh has quit [Remote host closed the connection]
jeremb_ has quit [Remote host closed the connection]
niftylettuce_ has quit [Write error: Broken pipe]
jeremb__ is now known as jeremb_
steev__ has quit [Remote host closed the connection]
jeremb_ has quit [Remote host closed the connection]
Hexxeh_ has quit [Remote host closed the connection]
niftylettuce__ has quit [Write error: Broken pipe]
<slapin> hno: SWAP
<slapin> hno: SWAP_DATA or something
<hno> Don't know exactly how that works yet.
Hexxeh_ has joined #arm-netbook
steev__ has joined #arm-netbook
jeremb_ has joined #arm-netbook
niftylettuce__ has joined #arm-netbook
popolon has joined #arm-netbook
Quarx has quit []
Quarx has joined #arm-netbook
<slapin> hno: as for LA, could you check what is on bus, when SRAM is inaccessible? also, odes IO_DATA is readable during such state?
<hno> slapin, what bus? The LA only have access to the NAND bus and the controller have finished the transfer there as per it's instructions.
<hno> the SRAM is internal in the SoC.
<slapin> hno: but since SRAM is not accessible, there might be something happenning.
<hno> The NAND bus is idle. Even CE is deasserted.
<slapin> hno: what makes me wonder, mw.l 01c03034 4a801209 then mw.l 01c03024 07ec0000 then mw.l 01c03024 16a00000 - at which stage SRAM is locked? does it generate DA when accessed?
<hno> what you mean by "generate DA"?
<slapin> hno: data abort
<hno> not sur. only get zeroes back when trying to read it from u-boot.'
<slapin> hno: ah, this is what happens when nand dump runs, hmmmm - locked down sram....
<hno> From what I remember it's locked after the command is sent. (0x24 register)
<slapin> hno: can you dump SRAM after each of mw.l 01c03034 ...? I don't have a board to test here at work :(
<hno> not now. have 6 kids to keep track of and no boart powered on.
<slapin> hno: kids' party? :)
* slapin needs to organize one this weekend
<hno> slapin, no, it's only my kids.
* stefanro1 wonders how one could handle 6 kids!!! - its hard enough with one - fun though :)
<mnemoc> with the help of an outstanding woman ;-)
<stefanro1> of course
tinti has joined #arm-netbook
<hno> slapin, telnet home.hno.se 7777
<slapin> hno: lucky you are
<slapin> hno: cool, what is this little program you use?
<slapin> hno: also, I never seen sun5i u-boot yet
<slapin> hno: where will it save its environment?
eebrah has joined #arm-netbook
merbanan has joined #arm-netbook
<slapin> hno: do you remember an address of SRAM?
<hno> #define NFC_RAM0_BASE 0x0400 1KB SRAM buffer
<hno> #define NFC_RAM1_BASE 0x0800 1KB SRAM buffer
<hno> offsets from nand controller base address.
<slapin> hno: from your script at https://github.com/hno/Allwinner-Info/blob/master/NAND/boot0.ush I get only zeroes, which is strange...
<hno> Hm.. probably the clock is not enabled right now.
<hno> nand controller is now enabled. try again
<hno> slapin^
gimli has quit [Ping timeout: 248 seconds]
<hno> slapin, sun4i & sun5i u-boot is the same. only SPL differ a little bit.
gimli has joined #arm-netbook
Turl has quit [Ping timeout: 264 seconds]
<slapin> hno: now works, but no reset is needed
<slapin> hno: also, last command leads to garbage in SRAM
<slapin> previous one returns info which looks valid for boot block
Turl has joined #arm-netbook
<slapin> repetition works too
<slapin> so, no reset is needed.
<slapin> cool
Quarx has quit []
<jinzo> I don't know if it was allready pasted - but could be usefull: http://xda-university.com/
Quarx has joined #arm-netbook
<slapin> hno:
Quarx has quit [Client Quit]
<slapin> hno: see backlog
itamarjp has joined #arm-netbook
rellla has quit [Quit: Nettalk6 - www.ntalk.de]
* slapin asked wife to setup board at home, so I managed to reproduce the same thing at my board, so works like charm
<slapin> hno:
<slapin> hno: at which address does bootrom loads data from NAND to execute?
Quarx has joined #arm-netbook
lerc has joined #arm-netbook
lerc_ has quit [Ping timeout: 265 seconds]
<slapin> hno: ping
<slapin> hno: drivers/mmc/mmc.c:1146: undefined reference to `init_part'
hp__ has quit [Ping timeout: 276 seconds]
<hno> slapin, the bootrom loads to the CPU SRAM at address 0-
<hno> slapin, I have seen that when switching between branches sometimes. Try a "make distclean".
<slapin> hno: is it possible with our u-boot to loady it then run?
<hno> boot0?
<slapin> hno: no, another u-boot, as I can;t switch cards at the moment
<slapin> can't
<hno> yes ofcourse. u-boot should be loaded to 4a000000
<hno> do that often.
eebrah has quit [Remote host closed the connection]
<hno> slapin, are you saying that you suceeded in doing an ECC+RAND read without locking the SRAM?
<slapin> hno: per your instruction, yes, dunno if ecc, but rand 100%
<slapin> hno: I try to implement this in code now
<slapin> hno: will need to add some thought to it
Turl_ has joined #arm-netbook
<hno> You olny used a CMD type 0, right? That does not cut it for ECC.
<slapin> hno: details?
<slapin> hno: I want to make experimental base first with all settings right.
<hno> There is four command types. 0 = simple CPU command. 1 = ECC command. 2 = page command (this locks SRAM)
<slapin> hno: interleaved read, I know
<slapin> type 1 just reports ECC without reading
<slapin> at least can't find anything
<hno> page commands do the whole sequence, for as many sectors instructed.
<slapin> hno: there is no store in SRAM for the whole page
<hno> type 1 reads the ECC data. Ignoring size always based on the ECC size/type configured.
<hno> I know a whole page do not fit in SRAM.
<slapin> hno: full read is possible with smple command
<hno> of first sector yes.
<slapin> hno: any page
<slapin> hno: any parto of page
<hno> not with randomizer enabled.
<slapin> hno: why?
<hno> I only get garbage then.
hipboi has joined #arm-netbook
<hno> and also have not figured out how to use type 1 commands proper.
pwhalen has quit [Quit: Leaving]
<hno> or to be precise, how to trigger ECC & Randomizer processing correctly when combining type 0 & trype 1 commands.
<slapin> hno: so I think we need to setup experimental base, with ability to turn various things on and off and experiment
itamarjp has quit [Read error: Connection reset by peer]
<slapin> hno: will work on it then
<slapin> will add commands for all things
rz2k has joined #arm-netbook
svp has left #arm-netbook [#arm-netbook]
<hno> slapin, ok.
<hno> there is also a number of unidentified bits.
<hno> mainly in CTL. Bit 15-21 is R/W, but never mentioned in driver sources.
hp__ has joined #arm-netbook
SouL_ has joined #arm-netbook
SouL_ has left #arm-netbook [#arm-netbook]
<hno> Whoa! Tried again and second sector read just fine!
<slapin> hno: so I think it is time to celebrate!
<slapin> hno could you remove your function which returns seed value and put static array to global space? current implementation looks extremely ugly
<hno> Feel free to rewrite.
<slapin> hno: ok
rz2k has quit []
<hno> Wonder why it did not work last time.
<slapin> mnemoc: did a story with mediawiki templates have any conclusion?
<mnemoc> slapin: no idea, oliv3r was on that
<slapin> oliv3r: I've got special request for you re mediawiki templates
hg_5 has joined #arm-netbook
<hno> slapin, now it even worked fine seeking directly to sector 2.
<hno> seems the random seed is the same for all sectors in a page.
<slapin> hno: that's great
Jef91 has joined #arm-netbook
hipboi has quit [Quit: Leaving]
Dessimat0r has quit [Ping timeout: 264 seconds]
<hno> Hm... odd. Can't seem to seec to page 2.
<hno> Ah, yes I can. User error.
Dessimat0r has joined #arm-netbook
<hno> slapin, boot0.ush updated a bit.
<hno> Wonder how BROM figures out the page size. Or maybe it simply assumes that page size is at least 8K + ECC.
<slapin> hno: it can ask chip about that
<slapin> hno: or read until page ends
<slapin> hno: what if page size is 2K?
Skaag has joined #arm-netbook
<drachensun> Man so I got the latest kernel commits this morning, my full size usb is working again but my OTG usb isn't now
<drachensun> I'm getting "parent hub has no TT"
<drachensun> any one else seeing that?
<mnemoc> drachensun: please test the stage branch
focus_well has quit [Ping timeout: 245 seconds]
<drachensun> ok, can do
ZaEarl has joined #arm-netbook
focus_well has joined #arm-netbook
Quarx has quit [Read error: Connection reset by peer]
Quarx has joined #arm-netbook
<slapin> damn
<mnemoc> killed your nand?
<slapin> my nand tells it has 640 bytes OOB
<hno> slapin, it never asks the chip anything. See boot0/A13-brom file. That's all the nand actions BROM takes during loading boot0 on A13.
<slapin> but attempt to read such oob kills u-boot
<hno> heh. u-boot also did not like mine.
<slapin> if I force it to 512 bytes, or less, it works...
<hno> Don't remember what I concluded that my nand have in OOB area, but was certainly more than MTD expected.
<slapin> probably this is some max value in u-boot or something...
<slapin> let me check...
<slapin> heheheh
<hno> ?
<slapin> so we need to allocate our buffers depending on chip size(?) or just 1K, and use NAND_OWN_BUFFERS chip option
<slapin> as it allocates too small chunk
<hno> sizeof(*chip->buffers));
<hno> chip->buffers = memalign(ARCH_DMA_MINALIGN,
<hno> ?
<slapin> ARCH_DMA_MINALIGN = 64 bytes it seems...
<slapin> hno: omap_gpmc seems to use NAND_OWN_BUFFERS
<hno> #define NAND_MAX_OOBSIZE 576
<hno> #define NAND_MAX_PAGESIZE 8192
<hno> include/linux/mtd/nand.h
<hno> there is also nand_buffers which is the struct allocated here.
rellla has joined #arm-netbook
servili007 has joined #arm-netbook
<hno> Header says that one should adjust MAX_OOBSIZE if adding a chip with larger oob size.
<slapin> hno: lets do that then
<slapin> hno: well, I really hope u-boot guys will update mtd
<hno> Well, I am a bit disturbed that MTD is this behind both in u-boot and kernel. THese nand chips are not new.
focus_well has quit [Ping timeout: 248 seconds]
<hno> MLC nand needs larger ECC codes + randomizer for reliable operation.
<slapin> hno: nobody cares now with MLC NANDs, everybody use SLC in their boards (who cares to commit to mainline)
<drachensun> mnemoc: Staging branch fixes the OTG port for me, accidently sent that in the other channel
<slapin> hno: actually kernel mtd should work properly
<slapin> hno: as it has wider chips support
<servili007> drachensun, on what board?
<drachensun> its not on a cubieboard, its on a tablet
<hno> 3.0.53 have the same limits.
<slapin> hno: check the recent one
<slapin> hno: I fixed mine though.
* slapin definitely will make randomizer work in code tonight
<hno> slapin, yes. 3.7 have 640.
<hno> But I wonder what OOB size my chip actually have. Not even 3.7 have an OOB size definition for this chip.
arokux1 has quit [Remote host closed the connection]
<slapin> hno: probably 640 too
<hno> Will probe later when we venture into trying write commands.
<slapin> hno: http://fpaste.org/D5dJ/ what is wrong here?
<slapin> hno: if not obvious, I will check a bit later myself
<hno> slapin, what should I look for?
<slapin> hno: when nandrand on, reads garbage of one type, when off -another
<slapin> hno: I try to read with randomizer from u-boot, I miss something, so I ask if you see if something wrong
<slapin> hno: otherwise I will check it setting-by-setting myself
<hno> you are trying to read the boot block?
<hno> or some logical sector?
<mnemoc> drachensun: noticed any side effect?
<slapin> hno: first page
<hno> slapin, seed is not right for boot blocks
<slapin> hno: weird, why they define 2b75 as seed for first page, while they never use it?
<slapin> will replace for 4a80 and see
<hno> and you are not reading the ECC data.
focus_well has joined #arm-netbook
<slapin> what randomizer does? randomizes locations of bytes on page or adds some random xor to data?
steev__ is now known as steev
<hno> probably xor with a stream of random numbers.
<hno> based on the seed.
<slapin> hno: is ecc required for randomizer?
<slapin> then I don't quite see a reason for randomizer...
<slapin> hno: cool, randomizer works.
<slapin> then lets play with ecc...
<slapin> hno: ok, so I need to implement interkeaved read, how you did yours?
<slapin> s/interkeaved/interleaved/
<ibot> slapin meant: hno: ok, so I need to implement interleaved read, how you did yours?
<drachensun> mnemoc: Nothing yet, everything seems to be fine
<mnemoc> drachensun: great. I'll merge the stage branch once someone confirms it works fine on a13 too
gzamboni has quit [Ping timeout: 265 seconds]
gzamboni has joined #arm-netbook
<hno> slapin, see boot0.ush
<ZaEarl> I have an a13 kernel. Would it be useful to compare to what we have in github?
<hno> ZaEarl, A13 kernel from where?
<ZaEarl> a tablet manufacturer
<hno> do you have arch/arm/mach-sun5i/ sources?
<ZaEarl> I think so. let me check.
<hno> In all A13 SDKs I have seen there is only a binary object there and headers.
steev has quit [Excess Flood]
<ZaEarl> I see a lot of .c files there
<hno> Interesting.
steev__ has joined #arm-netbook
<ZaEarl> File : cpu-freq.c
<ZaEarl> Descript: cpufreq driver on Reuuimlla chips;
<Turl_> sure that's an A13 kernel?
<ZaEarl> it's in a mach-sun5i dir
ganbold_ has joined #arm-netbook
<mnemoc> ZaEarl: can you paste a find -type f arch/arm/mach-sun5i/ ?
<mnemoc> err
<mnemoc> find arch/arm/mach-sun5i/ -type f
<Turl_> new *winner name? :P
<ZaEarl> It's got all my .o's too. http://paste.ubuntu.com/1455216/
<mnemoc> ZaEarl: \o/
<mnemoc> ZaEarl: and to increase a bit the greediness... does `git log` work? :p
<ZaEarl> no, I don't see any .gits
<Turl_> the new name is even more unpronunceable than the previous
<arete74>
gimli has quit [Ping timeout: 248 seconds]
Quarx has quit []
<ZaEarl> reuuimllatech.com isn't even registered
<steev__> it is now!
<ZaEarl> Tom Cubie <tangliang@reuuimllatech.com>
<mnemoc> * by tom cubie <tangliang@reuuimllatech.com>
<ZaEarl> weird
<steev__> oh i was being sarcastic
<jinzo> mali drivers in there too
<jinzo> but nothing of worth probably
<mnemoc> jinzo: mali kernel drivers are gpl. problem is on userspace
<jinzo> oh.
<jinzo> still, some interesting sources
<mnemoc> a13_source_1.5.tar.gz is 14G!
<Turl_> heh
<ZaEarl> my a13 source is about 1g
<mnemoc> ZaEarl: url? :)
<Turl_> I see an android repo in there
<Turl_> and buildroot
<Turl_> and uboot
<ZaEarl> mnemoc, where do you want it? :)
<Turl_> mnemoc: I see an a10 sdk there too
<Turl_> and an "a09" o.O
<jelly-home> good heavens I've barely gotten my hands on an A10
<arete74> a09 is an ARM Cortex A9 dual core processor with up to 1.5 GHz, as well as 1 GB DDR III RAM
<Turl_> a GPL compliant vendor, nice
drachensun has quit [Quit: Leaving]
<mnemoc> Turl_: up to know the problem was that the vendors weren't getting complete sources
<mnemoc> s/know/now/
<ibot> mnemoc meant: Turl_: up to now the problem was that the vendors weren't getting complete sources
popolon has quit [Quit: Quitte]
drachensun has joined #arm-netbook
<ZaEarl> Few companies selling tablets can actually build the source. They usually get it from another company,.
<ZaEarl> lol. .tar.bz2 is 1.7G, .tar.bz2.bz2 is .9G
<drachensun> is everybody else downloaded it all too?
<Turl_> nope
<drachensun> s/downloaded/downloading/
<Turl_> run it though codeen if you intend to though
<ibot> drachensun meant: is everybody else downloading it all too?
eFfeM has joined #arm-netbook
<hno> Turl, that name looks like someone just tried to remove Allwinner name from it. I do not think it's a real name.
<hno> mnemoc, there is nothing in that .git. Only symlinks to a non-existing .repo directory as in all other "proper" SDK releases.
<ZaEarl> lol, it looks like "allwinner" backwards and upsidedown
<hno> ZaEarl, good catch. It is.
voronaam has joined #arm-netbook
<Turl_> ZaEarl: ha
<Turl_> wonder why they'd change their name that way though
<Turl_> I guess tom would probably know
<mnemoc> short after he GPLed our sources he mentioned his boss asked him to change the headers in their internal trees to some random company
<mnemoc> on the mailing list
<mnemoc> i guess is for "protection"
<mnemoc> s/is/it's/
<ibot> mnemoc meant: i guess it's for "protection"
<Turl_> protection of what?
<mnemoc> reduce traces, make gpl enforcing more difficult or something weird
<Turl_> traces of what?
* Turl_ doesn't get it
<hno> Turl, don't try.
<hno> you only turn mad
<mnemoc> :p
<Turl_> haha
<Turl_> I guess working on clocks will be time better used than trying to figure that out
<hno> ZaEarl, would be interesting to have those sources.
<mnemoc> they are 100% aware they are violating the licenses of linux, u-boot and ffmpeg
<Turl_> the new clock framework implementation is working nicely :)
<mnemoc> probably also the licenses of the decoders supported by cedarx
<hno> clocks... right.. need to figure out what's wrong with the bus clocks on A13.
* Turl_ recalls mele killing support for one of the codecs on an update
<hno> AACS?
<Turl_> DTS I think
<hno> Right.. AACS is something else entirely.
* Turl_ now needs to move the clock initialization to device tree
<ZaEarl> hno, i'll send mnemoc a copy, he can play with it.
<hno> great
<hno> Turl, it's the responsibility of each manufacturer to license HDMI, MPEG, DTS, Ethernet MACs, etc etc...
<hno> Allwinner can not do that.
<hno> Allwinner also need a license in many cases.
<mnemoc> i assumed they need to pay to get decoders in the SoC
<slapin> hno: how long is your boot0?
<slapin> hno: I read first page and it happens to be half-empty, it is either a bug in code or something strange somewhere
<slapin> hno: not half - only 0x400+4 bytes are filled
<hno> slapin, i get data in all sectors nicely, over several pages.
<slapin> hno: I see, so it is bug.
<hno> tried boot0.ush sequence as reference?
<slapin> hno: I can't make random read work - probably positioning is broken somewhere
<slapin> hno: yes, but mtd wants to run RNDOUT occasionally, so it should work
<slapin> hno: you state it works for you but is somehow doesn't work for me
<slapin> will check in a hour or two
* slapin goes home
<hno> slapin, seeking using RNDOUT works, but is different than any other nand access due to ECC being interleaved
<slapin> hno: but it should not fill data with ffs
<slapin> hno: could you try my work branch with LA?
freakazoid0223 has joined #arm-netbook
<hno> yes when I get computer time again. Still running around trying to watch 5 kids. (1 to bed already)
eFfeM has quit [Quit: Leaving.]
gsilvis has quit [Read error: Connection reset by peer]
gsilvis has joined #arm-netbook
gimli has joined #arm-netbook
mikey_w has quit [Remote host closed the connection]
hp__ has quit [Read error: Connection reset by peer]
Hexxeh_ is now known as Hexxeh
jukivili has joined #arm-netbook
<Skaag> what's build.sh for in sunxi-bsp?
<Skaag> is it an old remnant?
<mnemoc> $ git ls | grep build | wc -l
<mnemoc> 0
<jinzo> that's from the "new" sources?
<mnemoc> lichee has a build.sh, sunxi-bsp doesn't
tinti_ has joined #arm-netbook
<jinzo> also I'm wondering what makes up the 10G difference between a10 and a13 sources
<mnemoc> android?
<jinzo> ah didn't notice that one on the first try. Blind as it looks like :D
<mnemoc> I downloaded i-onik.de's a10 tarball at 8MB/s ... now the A13's is downloading at mere 300k :<
<jinzo> :/
<jinzo> I'm conteptating a download too, but on the other hand I have way too much work on my hands allready :D
<techn_> mnemoc: is there something new?
<mnemoc> techn_: haven't checked. just downloading while $working$ :)
<mnemoc> techn_: when you have some spare time, can you test stage/sunxi-3.0 on your A13? otg 1.1/2.0 support and partial driver unification
eebrah has joined #arm-netbook
<techn_> mnemoc: I havent got usb working on that tablet.. but I'll try if it works now
<servili007> what's the proper way to add more default bootargs to u-boot?
<techn_> uEnv.txt or boot.scr
<servili007> I'll be working in nand, everything still applies?
<techn_> hmm.. I have forgot to document how to use boot.scr in sunxi-bsp :/
<servili007> hm, how can I modify environment for accomplishing the same thing on the original u-boot?
gzamboni has quit [Ping timeout: 252 seconds]
<hno> servili007, you need uart console to modify env of original uart, or a suitably modified fwenv binary capable of modifying the env in the right nand partition.
<hno> slapin, it seems ECC processing actually works.
SouL_ has joined #arm-netbook
SouL_ has left #arm-netbook [#arm-netbook]
<servili007> hno, if I have an existing environment dump, do we have a way to look into it?
<hno> fwenv modified to work on a file will do.
<slapin> hno: how it is?
<ZaEarl> sunxi-bsp can do an android build?
<hno> slapin, what?
<slapin> hno: have you got computer time?
<hno> some.
tinti_ has quit [Quit: Leaving]
<slapin> hno: can you try my branch?
<mnemoc> ZaEarl: poorly. it still requires manual integration of each board
<ZaEarl> but once the board is in, then it's just builds?
<slapin> hno: I can build you a binary if you wish
<mnemoc> haven't tried it myself... Turl did the integration
<ZaEarl> I want to get a fresh a13 build running on my tablet
<techn_> ZaEarl: I tried with my a13 tablet.. and I can't get recovery working
<ZaEarl> oh, good point, I need to get recovery on
<servili007> ZaEarl: zatab should work
<servili007> also, making your own configuration isn't too difficult, and I've gotten recovery working on a10
<techn_> ZaEarl: is there some known to be working recovery for a13 chips
<ZaEarl> I guess that's my first thing to figure out
<techn_> ZaEarl: sunxi-bsp "make android" seems to generate recovery.img
<techn_> but for me it doesn't boot.. dunno whats problem since I dont have serial soldered on that device
<servili007> techn_: the issue was probably in the kernel config if I had to guess.
<hno> slapin: Lots of warnings in sunxi_nand.c
<techn_> servili007: yes.. it doesn't even compile as stock
<techn_> someone know does android kernel run under linaro rootfs?
<servili007> techn_: I was just talking to Turl about this
<slapin> hno: my compiler doesn't report these
<servili007> the one built in sunxi-bsp for android is just sunxi-3.0 with some zatab patches and a couple of other things
<slapin> hno: anyway, use nandrand on then nand read.raw 46000000 0 1
<servili007> since the mk802ii doesn't boot the sunxi-3.0 kernel without some boot.scr additions, I either need to sub in 3.4 or modify the env on the nand to add it
<slapin> hno: look for rndout stuff with LA
SouL_ has joined #arm-netbook
SouL_ has left #arm-netbook [#arm-netbook]
<mnemoc> techn_: beside the annoyance of paranoid network, the android kernel behaves like a linux kernel
<hno> slapin, What u-boot command should I used?
<slapin> hno: ^
<techn_> servili007: what addions you need to boot.scr?
<servili007> techn_: I haven't actually dug into which ones it is specifically, but hold on, let me paste the one I know to work
<slapin> hno: nandrand on;nand read.raw 46000000 0 1
<techn_> mnemoc: I could then verify a13_nuclear_defconfig under of known to be working linux rootfs
<hno> slapin, your brand do not like my nand..
<slapin> hno: strange, it was rebased from your branch
<hno> you redid nand_base.c some time ago.
<servili007> techn_: http://pastebin.com/kbNWmtPH
<hno> moving chip detection to sunxi_nand.c.
<slapin> hno: yes
<hno> that broke..
<slapin> hno: how?
<hno> NAND device: Manufacturer ID: 0xad, Chip ID: 0xd7 (Hynix NAND 4GiB 3,3V 8-bit)
<hno> NAND bus width 8 instead 16 bit
<hno> No NAND device found!!!
<hno> nand_select_chip(-1)
<hno> NAND device: Manufacturer ID: 0xad, Chip ID: 0xd7 (Hynix NAND 4GiB 3,3V 8-bit)
<hno> NAND bus width 8 instead 16 bit
<hno> No NAND device found!!!
<servili007> techn_: since what it fixes for me is display detection on the hdmi port, I'd have to guess that it's line 6
<hno> nand_select_chip(-1)
<techn_> servili007: strange
<slapin> hno: awww, weird
<servili007> techn_: but again, that's 3.0 specific, 3.4 works sans boot.scr
<slapin> NAND bus width 8 instead 16 bit means no hynox code path activated in detection routine
<slapin> hmmmm
<techn_> servili007: I have no real need for boot.scr.. I only define bootdelay '0' there
<servili007> techn_: right, since the default values passed by u-boot are supposed to handle everything
<servili007> seems like an mk802ii specific issue
rellla has quit [Quit: rellla]
lerc has quit [Ping timeout: 246 seconds]
<hno> slapin, works fine without that part.
<slapin> hno: init_size should be activated regardless, and hynix code path should be executed without problems...
<slapin> hno: how come init_size is not set in your case?
<hno> I have no idea, and not going to look into that tonight. I did a checkout of your branch with no modifications.
<hno> now cherrypicked the new changes to my mtd branch.
Turl_ has quit [Ping timeout: 245 seconds]
<slapin> nand->init_size = sunxi_nand_init_size; is always set
<techn_> anyone know if disp module supports dual display?
<ZaEarl> cwm from http://www.mediafire.com/?t3ctcths6r2229q basically works on a13.
<slapin> hynix path is always running :(
<hno> but not working.
<techn_> ZaEarl: where you found that link?
<servili007> ZaEarl: ask the maker for their kernel config
<hno> slapin, better to update mtd I think.
<slapin> hno: you can revert that patch for now if you feel the problem is there then well test it later
<hno> slapin, that's what I have done already, except that I have not merged it at all.
<ZaEarl> lack of a working 'back' button is a pain, but usable
<slapin> hno: it is better, but we'll drown in mess if we'll depend too much on local patches to mtd
<hno> slapin, well, update and submit naturally.
<servili007> ZaEarl: It's pretty easy to reassign the keybindings when building
<ZaEarl> sure, once I locate the source for it :)
<hno> slapin, this can not work. It's not reading the ECC, and it's seeking to the wrong offset for second sector.
<slapin> hno: u-boot IIRC mtd update process is from kernel only
<hno> slapin, yes.
<hno> I said update, not patch.
rz2k has joined #arm-netbook
<hno> Linux-3.7 code have a good chance of working with my chip.
<hno> slapin, what is this read.raw supposed to read here?
<slapin> hno: page including oob
<hno> slapin, remember that the layout of these boot block pages is very odd.
<slapin> hno: regardless of whatever I don't see how wrong offsets could lead to ffs
<slapin> hno: different from all the remaining pages?
<hno> slapin, the bootblock is different from any other nand usage.
<hno> bootblock: <data> <ecc> <data> <ecc> <data> <ecc> <data> <ecc> [8 times]
<hno> normal block: <data> <data> [8 times] <ecc> <ecc> [8 times]
<hno> and different ECC mode as well.
<slapin> hno: well, but it should then read something, not ffs...
<hno> I got random garbage, not FFs
<techn_> thanks ZaEarl I got recovery :)
Jef91 is now known as Jef91|Away
<hno> slapin, I see 00 <00 00 00 00 00> 30 [1024 random bytes] 05 <00 04> E0 [1024 random bytes] ...05 <00 08> E0 [1024 random bytes]
<slapin> hno: I get only ffs and first 512-byte + ecc allright
<hno> sector size is 1K.
<slapin> hno: have you did nandrand on?
<slapin> actually first 0x474 nytes allright
<slapin> s/nytes/bytes/
<ibot> slapin meant: actually first 0x474 bytes allright
<hno> nandrand is on.
<hno> I get first 0x400 bytes fine at 46000000. The rest is only random garbage as expected from reading at wrong offsets.
<hno> no blocks of FFs.
<slapin> hno: so rndout works for you but not for me which is straaaaange
<hno> Different NAND chips and different detection code.
<hno> I have pushed my mtd branch if you want to verify what I am running.
<hno> but could not see any changes to your branch other than the chip detection.
<hno> now I need to do some other stuff. Talk more tomorrow.
<hno> and plan on making a small fel tool for dumping boot0 + boot1.
<hno> slapin, do the sequence in boot0.ush work for you?
<hno> remember to update. there was some fixes and clarifications today.
<slapin> hno: for reading boot blocks, yes
<hno> good.
<hno> reading normal blocks is slightly different. There another ECC mode is used, and all ECC data is in the OOB/spare region in the NAND page.
<slapin> hno how did you come to idea to run commands just to read data without actual commands?
<hno> Because it's just a sequential read of the nand page.
<hno> data, ecc, data, ecc, data, ecc, data, ecc...
<hno> nand do not want any commands for sequential reads.
<hno> using 05 <xxxx> e0 in each step also works fine. See last example.
<slapin> hno: which ECC mode does normal blocks use?
<hno> as long as you remember to seek to the right offset (0x42e, not 0x400)
<Skaag> I configured sunxi-bsp to a13_olinuxino, how do I know if linux-sunxi is built with sun4i_defconfig or sun5i_defconfig ?
<hno> slapin, Which ECC mode is used depends on oob size.
<slapin> hno: 42e for boot blocks
<hno> slapin, for my chip normal pages use ECC mode 3.
<hno> and boot blocks always use ECC mode 1.
<techn_> Skaag: build. it should show on prints
<slapin> ok, so we need to write proper ecc functions and do all this magic stuff from them
<Skaag> techn_: in builds all I see is a13_olinuxino
<hno> slapin, or use hardware ecc.
<Skaag> techn_: I just finished building it
<hno> it works.
<mnemoc> ZaEarl: the 1.7G file called .tar.bz2 is indeed just .tar misnamed
<slapin> hno: I mean write appropriate callbacks for chip->ecc.* functionslike other drivers do
<ZaEarl> haha
<Skaag> ok it built it with a13 instead of sun4i
<Skaag> I guess the A13 Olinuxino is Sun5i?
<Skaag> since in the a13_defconfig I see this: CONFIG_ARCH_SUN5I=y
<jinzo> all A13s are sun5i, and there's a lot of info in the wiki
<voronaam> oops
<voronaam> sorry, wrong window
<Skaag> jinzo: It's confusing sometimes because some modules are sun4i despite this, like the touchscreen for example, it's sun4i-ts
<techn_> Skaag: that's true.. thats why we are trying to get sunxi
<Skaag> that's cool, and anything I can do to help that effort, I will do
<Skaag> I'm trying to find what happened with the sun4i-gpio module
<Skaag> and how to get it compiled as a module
<Skaag> what's the recommended branch in linux-sunxi?
<Skaag> The default one that comes with sunxi-bsp makes kernel 3.0.52
<mnemoc> sunxi-3.0 is still the recommended branch
<Skaag> is it possible sun4i-gpio changed its name to axp20-gpio?
<mnemoc> but in this particular instance you might be better trying stage/sunxi-3.0 due to a fix to OTG
<mnemoc> Skaag: different things
<Skaag> OTG = ?
<mnemoc> host/slave usb port
<Skaag> thanks
<Skaag> I tried enabling CONFIG_SUN4I_GPIO_UGLY=m thinking it would produce sun4i-gpio.o
<Skaag> but it doesn't seem to do it
<techn_> Skaag: it could have some dependencies
<Skaag> how do I find out?
<Skaag> do I need to enable general purpose GPIO support?
<mnemoc> make menuconfig
<techn_> check kconfig where SUN4O_GPIO_UGLY is defined
<mnemoc> only if you want to play with the pins from userspace
<Skaag> when I make menuconfig, there's no .config
<Skaag> should I copy the a13_defconfig from the arch/arm/configs dir?
<techn_> Skaag: no.. I suggest to modify that Makefile so that it will call make menuconfig with correct params
<techn_> it needs ie. ARCH=arm
<techn_> also correct o= param
<Skaag> this is the linux-sunxi makefile...
<techn_> one moment
voronaam has quit [Quit: Leaving.]
<Skaag> also there's a build.sh in there, which references more specific scripts (per platform) under scripts/
<techn_> Skaag: try latest sunxi-bsp and "make linux-config"
<Skaag> I just cloned this one: git://github.com/linux-sunxi/sunxi-bsp.git
<Skaag> this would be the latest, right?
<Skaag> there's no rule for linux-config in the linux-sunxi that this brings
drachensun has quit [Quit: Leaving]
<techn_> Skaag: I just updated repo
<Skaag> cool, running make update
<techn_> hmm.. :)
<techn_> use git pull :)
<Skaag> that's ok right? I mean running "make update" under sunxi-bsp/
<Skaag> ok
<Skaag> git pull tells me i'm not currently on a branch
<Skaag> (if I run it under linux-sunxi)
<techn_> I mean git pull under sunxi-bsp
<Skaag> that worked
<Skaag> and now I have linux-config as a make target
freakazoid0223 has quit [Read error: Connection reset by peer]
<Skaag> ok this feels much more organized
<Skaag> :)
<Skaag> so make linux-config will edit the correct .config file
<techn_> Skaag: yes
<Skaag> yet!!! it worked. now in the list of modules I see sun4i-gpio!
<Skaag> Halleluja!
<Skaag> s/yet/yes/g
<Skaag> ok this is was a fantastic update to sunxi-bsp in my opinion
<techn_> thanks :)
mikey_w has joined #arm-netbook
gimli has quit [Ping timeout: 264 seconds]
<techn_> Skaag: I improved sunxi-bsp little more
<techn_> now make update updates also sunxi-bsp ;)
mikey_w has quit [Remote host closed the connection]
drachensun has joined #arm-netbook
Jef91|Away is now known as Jef91
hg_5 has quit [Remote host closed the connection]
popolon has joined #arm-netbook
rz2k has quit []
<servili007> techn_: getting an error with the new make update 'unknown option for 'stash save': --include-untracked' at cedarx-libs
<techn_> servili007: git version?
<servili007> 1.7.5.4
<techn_> git version 1.7.10.4
<servili007> I believe my version is just whatever ubuntu pulls down by default
<servili007> yeah, it's still the latest version in standard repos
<techn_> I have 12.10 ubuntu
<servili007> alright, I'm still at 11.10
Jef91 has quit [Quit: Instantbird 1.3 -- http://www.instantbird.com]
countz has quit [Read error: Connection reset by peer]
Skaag has quit [Read error: Connection reset by peer]
countz has joined #arm-netbook
Skaag has joined #arm-netbook