BJfreeman has quit [Read error: Connection reset by peer]
_BJfreeman has joined #linux-sunxi
_BJfreeman is now known as BJfreeman
hurtigbuffer is now known as jelly-home
BJfreeman has quit [Quit: had a good time]
\\Mr_C\\ has quit [Read error: Connection reset by peer]
<hno>
oliv3r, no.
<hno>
and the olimex board works with both A10 and A20.
\\Mr_C\\ has joined #linux-sunxi
[7] has quit [Disconnected by services]
TheSeven has joined #linux-sunxi
_BJfreeman has joined #linux-sunxi
_BJfreeman is now known as BJfreeman
drachensun has quit [Quit: Leaving]
drachensun has joined #linux-sunxi
eebrah has quit [Ping timeout: 264 seconds]
BJfreeman has quit [Quit: had a good time]
Quarx has joined #linux-sunxi
piyushverma has quit [Ping timeout: 248 seconds]
Gorp has joined #linux-sunxi
<Gorp>
Anyone have any idea how to remove the "p2v8" at the end of the vermagic in a compiled kernel module?
<Gorp>
I'm trying to compile a kernel module, but the vermagic keeps showing the wrong string.
piyushverma has joined #linux-sunxi
<Gorp>
Anyone? .__.
ganbold__ has joined #linux-sunxi
dl9pf has joined #linux-sunxi
dl9pf has quit [Changing host]
dl9pf has joined #linux-sunxi
tkoskine_ has joined #linux-sunxi
grevaill1t has joined #linux-sunxi
focus_it has joined #linux-sunxi
ZetaNeta has joined #linux-sunxi
specing_ has joined #linux-sunxi
specing has quit [Ping timeout: 276 seconds]
LoCoZeNoz_ZUE has quit [Ping timeout: 276 seconds]
ganbold_ has quit [Ping timeout: 276 seconds]
focus has quit [Ping timeout: 276 seconds]
tkoskine has quit [Ping timeout: 276 seconds]
dl9pf_ has quit [Ping timeout: 276 seconds]
grevaillot has quit [Ping timeout: 276 seconds]
specing_ is now known as specing
tzafrir has quit [Ping timeout: 248 seconds]
aexl has joined #linux-sunxi
paulk-desktop has joined #linux-sunxi
aexl has quit [Quit: Leaving.]
aexl has joined #linux-sunxi
aexl has quit [Client Quit]
aexl has joined #linux-sunxi
<oliv3r>
hno: ping
aexl is now known as UMLAUTaxl
UMLAUTaxl is now known as aexl
_whitelogger has joined #linux-sunxi
_whitelogger has joined #linux-sunxi
_whitelogger__ has joined #linux-sunxi
_whitelogger__ has joined #linux-sunxi
_whitelogger__ has joined #linux-sunxi
_whitelogger___ has joined #linux-sunxi
_whitelogger___ has joined #linux-sunxi
_whitelogger___ has joined #linux-sunxi
_whitelogger___ has joined #linux-sunxi
_whitelogger__ has quit [Remote host closed the connection]
_whitelogger has quit [Remote host closed the connection]
_whitelogger___ has quit [Remote host closed the connection]
Gorp has quit [Quit: Page closed]
tzafrir has joined #linux-sunxi
sanka has joined #linux-sunxi
hamburger2000 has joined #linux-sunxi
piyushverma has quit [Remote host closed the connection]
piyushverma has joined #linux-sunxi
<oliv3r>
hmm, my olimex a20 refuses to boot from flash
<ZetaNeta>
yo, guys. Does anyone aware where android keeps the modules?
<oliv3r>
if I power it from usb-otg i it hangs at aw_clkevt_init: sun7i_timer0_clockevent mult 6442450, max_delta_ns 170000, min_delta_ns 1000, cpumask 0xc047e8c0, irq 54
<oliv3r>
ZetaNeta: /system/lib/modules?
<ZetaNeta>
ouch
<ZetaNeta>
i am wondering why "find / | grep SAMPLEMODULE" didnt work
<ZetaNeta>
but anyway, tnx
<oliv3r>
ZetaNeta: yw
<oliv3r>
if i power through a 5V powerbrick it hangs at: [ 45.114609] Freezing of user space aborted
<oliv3r>
before init devices.
<oliv3r>
before power init.
<oliv3r>
before dram enter selfresh.
<oliv3r>
before gating off dram c?
<oliv3r>
trying 12V brick now
<oliv3r>
same
<ZetaNeta>
and what exactly are you booting?
<oliv3r>
stock nand on the new olimex-a20-micro dev board
<ZetaNeta>
oh
<ZetaNeta>
i didnt have any of those devboards (through dreaming bout one). But a tablet with working android on it
<ZetaNeta>
So..... Try booting something else?
<ZetaNeta>
Maybe something else may work?
<ZetaNeta>
btw, what else you get before "[ 45.114609] Freezing of user space aborted"?
<oliv3r>
this is kinda new
<oliv3r>
we haven't support for anythign else; we need to load stock android once to get the memory timings
<ZetaNeta>
btw, does you wiki need a ALIP specific g_serial manual?
<oliv3r>
no clue
<ZetaNeta>
becouse now i got experience
<ZetaNeta>
in that
<ZetaNeta>
question
<oliv3r>
if you use the bsp to build the hwpack; all those will get installed automatically
<oliv3r>
so not quite sure what you are trying to say
<ZetaNeta>
i mean the g_serial module setup
<ZetaNeta>
i learned how to fully configure it on ALIP
<oliv3r>
ah ok
<oliv3r>
sure make a wiki page fro it
<aexl>
ZetaNeta: yes please. also you could maybe enhance /USB_Gadget please.
<ZetaNeta>
ok
<aexl>
ZetaNeta: you're LoCoZeNoz_Zue right? thx for the name change^^
<ZetaNeta>
?
<ZetaNeta>
I am?
<ZetaNeta>
Oh
<ZetaNeta>
Probably IRC took my secondary nick
<oliv3r>
that explians a lot :)
<aexl>
oliv3r: seems something is seriously broken. android kernel is normally more verbose.
<oliv3r>
*nods*
<oliv3r>
i'll ask tsvetsan later what is supposed to happen
<oliv3r>
it also looks like there's some LCD being initted or something
<aexl>
maybe they can send you a livesuit image. if they are available sundays ...
<oliv3r>
i'll ask tsvetsan for it prob
<oliv3r>
but who knows what they have :)
<oliv3r>
also, i'll have to find a box that runs livesuit :p
<aexl>
what's your distro?
<oliv3r>
gentoo
<oliv3r>
but i've heard bad things about the linux livesuit
<aexl>
i've got it running in squeeze. but it always says "invalid image".
<aexl>
or you try android/cwm on sdcard ...
<oliv3r>
yeah, but how do I boot the sdcard?
<oliv3r>
the problem is; we need to boot from nand (boot0/1) first to get the timings
<oliv3r>
i'm gonna try fel_boot next
<oliv3r>
since we boot from boot0/1 first; which sets up the clocks etc
<aexl>
doesn't it automatically when sdcard is inserted?
<oliv3r>
and if we can boot u-boot from there, we have all the timings setup right; but gotta ask hno if spl/fel wil change those timings again
<oliv3r>
aexl: but how do you configure your spl (dram timings etc) if you don't know these yet
<oliv3r>
normally, you build your spl, with timings you extract from a running stock android
<oliv3r>
because boot0/1 has setup the registers properly
<aexl>
i just mean when you have the android image but no livesuite ...
<oliv3r>
ah ok
<oliv3r>
put the stuff from the livesuit image on the sd card
<oliv3r>
yeah; hope it's on the wiki then :)
<oliv3r>
might be as easy as extract, dd; done
<aexl>
right. or you try out some other images right now.
<oliv3r>
i think felboot might be the best direction now
<oliv3r>
i want to get into u-boot console asap
<oliv3r>
well one thing I can do
<oliv3r>
is boot 'anything' that gives me a console
<oliv3r>
nuke env partition
<oliv3r>
then, boot0/1 will work 'good enough'
<oliv3r>
e.g. i can access u-boot console
<aexl>
CWM_on_a_SDCard Boot_Android_from_Sdcard
<oliv3r>
because now (the main problem) is they've set the timeout to 0 seconds
<oliv3r>
aexl: you have to remember though; this image will have to support A20
<ZetaNeta>
I am fixing main page/ru alittle. Should i refer to a "ARM SoC family" as to a single thing, or as to several (разработаной, разработаных). It seems like to a single thing
<oliv3r>
erm, can't read russian :D
<ZetaNeta>
there is different words for developed for single and multiple things
<aexl>
ZetaNeta: at least write down the infos you have. nevermind the format.
<ZetaNeta>
hramrach_, Yet, i dont understand why you need a patch for configurations?
<ZetaNeta>
why cant you merge obvious things into the existing config?
<ZetaNeta>
why do you need patch?
<ZetaNeta>
nvm
<ZetaNeta>
il leave the config crazyness to you
<hramrach_>
how else do you merge something into a config other than by a patch?
<ZetaNeta>
?
<ZetaNeta>
on client side: make linux-config. On you side: just replace the old on on git?
<ZetaNeta>
i didnt understand what you mean with this patch
<ssvb_>
ZetaNeta: merge as in get these changes added to the linux-sunxi repository on github?
<hramrach_>
you replace old config with new config and git produces a path that updates the old config to new config
<ZetaNeta>
oh, i am just not so good into the git thingy
Yaku321 has joined #linux-sunxi
<ZetaNeta>
you sad g_serial is in somekind of patch
<ZetaNeta>
i mean to add g_serial to what people get when using the firststeps instructions
<ZetaNeta>
(i am really far away from git man)
<ssvb_>
ZetaNeta: the idea is that 1) a patch is created 2) then it is posted for review 3) then if nobody has objections, it gets added to github
<hramrach_>
the patch to which I posted link does add g_serial
<ZetaNeta>
ssvb_, You gave best answer
<hramrach_>
'but nobody has reviewed that so the config on github is still the old one with no drivers
<aexl>
ZetaNeta: it's also about who commited what and code history.
<ZetaNeta>
so, is it ok to make a folder /lib/modules/3.0.76/kernel/android and to put there all android modules?
<ZetaNeta>
my question, will it find them
<hramrach_>
not likely
<hramrach_>
the andriod kernel has different symbol version
<hramrach_>
but you can try
eebrah has joined #linux-sunxi
<oliv3r>
bah, can't write anything via fel; fel-read seems to work though
<oliv3r>
usb_bulk_recv error -7
<oliv3r>
well gotta go
damo22 has joined #linux-sunxi
sanka has quit [Remote host closed the connection]
<damo22>
i noticed that olimex has a product involving the A20 chip and 1xSATA... i cant find any gnu/linux stuff for the a20 chip, is this something very new?
<hramrach_>
it's very new
<hramrach_>
support is under development
<ZetaNeta>
i have copied modules in /lib/modules/3.0.76/kernel/android. Should i reboot to use them?
<hramrach_>
read it will crash a lot for some weeks or months to come ;-)
<hramrach_>
ZetaNeta: run depmod. then try to load the modules
<damo22>
hramrach_: i was following the A10 stuff a while ago from olimex they were going to release a board with A10 and 2x USB ports that could support 500mA each... but they stopped developing the board, do you know of a board that can support 2x full powered usb ports?
<hramrach_>
I have no problem with using USB prots on cubieboard
<hramrach_>
not sure what the exact power spec is, though
<hramrach_>
why would olinuxino not power the USB ports? or does it have only one?
<damo22>
hramrach_: i'll have another look
<damo22>
hramrach_: according to the spec for revision D of the A10s board, there is 1x usb + 1x usb-otg and they both provide full power requirements
<aexl>
maybe next time i'm confronted with it. but mnemoc told me around 3.0.42 that the problems should be fixed and that i should it try it. i'm still running 3.0.39 with my changes.^^
<ssvb_>
ZetaNeta: the "FBIOPUTCMAP: Invalid argument" can be workarounded in the kernel by not returning error when userland tries to use FBIOPUTCMAP incorrectly
<ssvb_>
ZetaNeta: that's how the x86 drivers (vesa and friends) usually behave
<ZetaNeta>
in what section may i find that option?
<ssvb_>
it's a single line kernel patch
<ZetaNeta>
and where may i find it?
<ZetaNeta>
or, if it single line. Paste it xD
<ssvb_>
I haven't sent it anywhere and lost it already :)
<ZetaNeta>
:(
<aexl>
^^
<ZetaNeta>
Did yeah know Allwinner infringes copyright? (sry, google translate)
<ZetaNeta>
Of america online?
BJfreeman has quit [Ping timeout: 252 seconds]
<ZetaNeta>
With the LiveSuit software
<ZetaNeta>
No?
_BJfreeman has joined #linux-sunxi
\\Mr_C\\ has quit []
_BJfreeman is now known as BJfreeman
<aexl>
ZetaNeta: ?
<ZetaNeta>
It uses AIM logo as exit button :D
<ZetaNeta>
AOL: *calls lawyers*
<ZetaNeta>
make: m2c: Command not found
<ZetaNeta>
What is m2c?
<ZetaNeta>
A random joke: What makes Binary and Koren similar?
<ZetaNeta>
not much people may answer this
<specing>
what is "Koren"?
<aexl>
korean?
<hno>
oliv3r, do you have adb root access? Then it's easy to nuke the env.
<ZetaNeta>
yeah
<ZetaNeta>
korean
<damo22>
binary and korean? they both use |o|o|Oo||
\\Mr_C\\ has joined #linux-sunxi
<specing>
One is spoken by computers and the other by people who build them?
BJfreeman has quit [Ping timeout: 240 seconds]
<aexl>
there are 10 kinds of koreans ...
<ZetaNeta>
damo22, You are correct
<ZetaNeta>
specing, A good answer, yet damo22 is right
_BJfreeman has joined #linux-sunxi
_BJfreeman is now known as BJfreeman\
BJfreeman\ is now known as BJfreeman
ZetaNeta has quit [Ping timeout: 264 seconds]
LoCoZeNoz_ZUE has joined #linux-sunxi
ganbold__ has quit [Remote host closed the connection]
<LoCoZeNoz_ZUE>
i have patched ft5x_ts.h in kernel sources, now how do i build it?
<LoCoZeNoz_ZUE>
What exactly do you mean under that
<mripard_>
You can't possibly compile every filesystem, every device driver into a static kernel image, it would make no sense
<LoCoZeNoz_ZUE>
why,
<paulk-desktop>
is there u-boot support for pcduino?
<mripard_>
aexl: yeah, from a architecture point of view, it's monolithic
<LoCoZeNoz_ZUE>
What exactly stops me from compiling all the FS into 1 kernel
<LoCoZeNoz_ZUE>
even through i dont need them all
<mripard_>
nothing
<aexl>
mripard_: okay. but i know now what you mean.
<mripard_>
except that it would be always loaded into RAM
<LoCoZeNoz_ZUE>
I dont think il ever use Amiga partition layout. Or some rare FS
<aexl>
LoCoZeNoz_ZUE: but even if you have 250M modules linux is still monolithic
<mripard_>
and you'll have it even if you don't use the feature
<LoCoZeNoz_ZUE>
aexl, Yet, how should we call the difference betwen "no-modules-linux" and the one who got em?
<LoCoZeNoz_ZUE>
if not monolithic?
<aexl>
modular and non-modular?
<LoCoZeNoz_ZUE>
GENIUS!
<mripard_>
hno: thanks for your work on A10s-olinuxino btw, it worked nicely here :)
<LoCoZeNoz_ZUE>
Now explain those 2 terms to entire internet and say that using "monolithic" is uncorrect. Go on
<mripard_>
hno: have you started working on A20/A31 yet?
<LoCoZeNoz_ZUE>
Its like explaining that 3D is term from graphics, and all 3d movies should be called "stereo movies"
<aexl>
modular means that you can load modules, non-modular means you cannot. monolithic is the kernel no matter how much is built-in or not ... i guess.
<mripard_>
aexl: yeah, strictly speaking, you're right. Yet the difference is blurry to a lot of people
<mripard_>
and so a lot of people use monolithic for !modular
<aexl>
but to say that i use a non-monolithic linux is wrong,
<aexl>
just saying.
<LoCoZeNoz_ZUE>
Thats the best example i could made aexl
<aexl>
i was just a little confused since i always wanted to try a microkernel like hurd (though i hear bad things) or minix.
<aexl>
especially to head something like this in a tech channel.
<hno>
mripard_, lkcl have done A20 support, pending for cleanup and integration. Available in lkcl-sunxi-a20 branch of my u-boot repository.
<LoCoZeNoz_ZUE>
aexl, Yet, people are not computers.
<hno>
Got my Olimex A20 board a couple of days ago so I have something to test on.
eebrah has joined #linux-sunxi
<aexl>
LoCoZeNoz_ZUE: yeah. this is why i asked.
<mripard_>
hno: yeah, I got it something like 20minutes ago, thus why I was asking
<LoCoZeNoz_ZUE>
i got no devboards, but a a10 tablet. Is A10 better than A31/A20?
<LoCoZeNoz_ZUE>
becouse they seems to be a "next generation" yet i see they (maybe something with my eyes) are slower than a10
<LoCoZeNoz_ZUE>
how many cores does you pc have? (a quick vote for entire channel)
<Turl>
LoCoZeNoz_ZUE: A10 is 1 Cortex A8
<LoCoZeNoz_ZUE>
no, i mean you desktops
<Turl>
A20 is 2 Cortex A7 and A31 is 4
<Turl>
LoCoZeNoz_ZUE: my desktop has 4 cores, 8 threads
<LoCoZeNoz_ZUE>
well, lets put it simplier. How many does /proc/cpuinfo show (including 0)
<Turl>
$ cat /proc/cpuinfo|grep 'model name' -c
<Turl>
8
<LoCoZeNoz_ZUE>
tnx
<aexl>
2 (it's a dell from 2007)
<LoCoZeNoz_ZUE>
anyone else wanna vote?
<LoCoZeNoz_ZUE>
oh, 2?
<LoCoZeNoz_ZUE>
well, ok
<LoCoZeNoz_ZUE>
Anyone else wanna vote?
<aexl>
will be 6 on 15th :D
<LoCoZeNoz_ZUE>
:D
<LoCoZeNoz_ZUE>
ok
<Turl>
linux-sunxi server votes 8
<LoCoZeNoz_ZUE>
i asked becouse the "FirstSteps" guide uses -j5 when doing make with kernel
<LoCoZeNoz_ZUE>
And as any gentoo use knows, it should be -j(cores+1)
<Turl>
I do -jcores*2
<LoCoZeNoz_ZUE>
Well, now i got another idea. To put a "hint from gentoo users" on bottom/top of the page
<LoCoZeNoz_ZUE>
gonna do that right now
<Turl>
:P
<aexl>
pf
<Turl>
just use -j :P
<aexl>
i can use -j 32 as well ...
<LoCoZeNoz_ZUE>
"i want people know how to compile faster and better" - average gentoo user
<LoCoZeNoz_ZUE>
s/know/"to know"
<aexl>
it's already in FirstSteps on also on rhombus-tech.net. this is no new gentoo information. ;)
<Turl>
LoCoZeNoz_ZUE: 1) buy a bigger box with an ssd 2) ... 3) profit! :)
<LoCoZeNoz_ZUE>
XD
<LoCoZeNoz_ZUE>
"for free!" - me
<jelly-home>
LoCoZeNoz_ZUE: it doesn't matter now many cores an intel or pc desktop has, when each single core can crunch integers 100x as fast, and access memory 20x as fast as your arm soc
<aexl>
i've got an ssd but i wouldn't use it for crunching ...
<LoCoZeNoz_ZUE>
yeah, yet. 5 min compilation on a machine that can do it in 2..... not gentoo way
<Turl>
aexl: code on ssd, object files on tmpfs :D
<aexl>
Turl: tmpfs is on ram. :D
<Turl>
aexl: yeah no ssd wear :)
<aexl>
optimizations. i love technology.
aexl is now known as aexl|away
<LoCoZeNoz_ZUE>
my mind stream: SSD/HDD > Cold Static Storage/Hot Dynamic Storage > Bitcoin cold wallet > omg, i forgot to install armory
<mripard_>
bfree: ping?
<LoCoZeNoz_ZUE>
guys, i have found a error at the wiki
<LoCoZeNoz_ZUE>
It says to add hf to all commands if you got a hardfloat toolchain
<LoCoZeNoz_ZUE>
yet, all commands are already with HF added :D
<LoCoZeNoz_ZUE>
so either 1 either another should be fixed
<LoCoZeNoz_ZUE>
i vote for removing the useless "add hf" thingy.
<LoCoZeNoz_ZUE>
oh.... my fail
<hno>
2
<LoCoZeNoz_ZUE>
failed to read properly XD
<LoCoZeNoz_ZUE>
btw, i predict that there will be no future where embedded systems will support bitcoins
<LoCoZeNoz_ZUE>
as to properly use bitcoins, you already need a 9gb file
<LoCoZeNoz_ZUE>
I think there is one thing bitcoin creator didnt think of :D
<hno>
LoCoZeNoz_ZUE, 2 of my Allwinner boards have SATA.
<LoCoZeNoz_ZUE>
oh... ouch
<LoCoZeNoz_ZUE>
yet, there are still embedded systems where they count storage in kb
<hno>
and trivial to attach any mount of USB storage to any board.
<hno>
Was long since modern ARMs counted in KB, if excluding the absolute bottom budget range.
<hno>
i.e.. Cortex M*.
<aexl|away>
there are even 512 GB flash sticks out there. and 64GB sd cards.
<aexl|away>
micro sd cards
<hramrach_>
does a10 support those 64gb cards?
<hramrach_>
it has xhci but spec lists 32gb max
<hramrach_>
not xhci
<hramrach_>
sdxc
<hramrach_>
but nobody tried the new cards afaik
<hno>
aexl|away, is there 64GB SD cards? Where?
<Turl>
hramrach_: the beauty of sd cards is that you can have 2x32GB and switch them :)
<hramrach_>
but they will come in a few months I guess
<aexl|away>
Turl: what's shipping to Germany from your location? :D
<hramrach_>
dx has free shipping ;-)
<Turl>
aexl|away: shipping it to you would probably cost more than buying it on dx :)
aexl|away is now known as aexl
<hramrach_>
wow, that looks ugly
<aexl>
okay.
<Turl>
hramrach_: it works :p
<Turl>
it's hardly top quality but it did get its job done
<hramrach_>
the reviewers say it is too loose
<hramrach_>
which is expected since they just stuff a flex cable with hardly any support into the slot
<hramrach_>
an that's the obviously ugly part
<Turl>
hramrach_: it has a hard part on the uSD end of the cable that clicks on the slot
<aexl>
ive got sparkfun breakout board. but i didnt get the job done. androids uboot only told me 3#3#3#3#3#3#3#3#3#3#3#3
<Turl>
aexl: wrong speed?
<aexl>
well it showed kernel booting.
<Turl>
aexl: you need to replace uboot with a modified one to output to uSD card too
<aexl>
yeah but i didn't know how to boot from sd card with the breakout board attached or how to replace androids uboot. maybe some day ...
<Turl>
aexl: you need to build with one of the lichee branches, don't remember which now
<Turl>
then replace the u-boot.bin on nanda with what you built :)
<aexl>
that's all (besides a modified script.bin)?
<Turl>
yes
<Turl>
the build needed the special branch and a little configuration, hno can probably clarify that part
<Turl>
but then it was justreplacing it and modifying script.bin
<hno>
hramrach_, the early A10 prototype and the A13 OLinuXIno micro boards only have micro-sd. A10s and A20 OLinuXIno MICRO both have micro-sd + full size sd.
<hno>
Turl, what?
<aexl>
hno: for using sdcard as uart - which lichee-branch do i need? and what do i need to modify in the code?
<Turl>
hno: what do you need to get uart on sd slot booting from nand? I recall you needed to use one of the lichee branches and touch something on the config
<aexl>
Turl: thx^
<aexl>
^^
<hno>
ah, console. With a modern lichee u-boot it should be sufficient to set the debug_uart to the SD slot in script.bin (and disable the same sd slot).
<hno>
earlier you needed to build sun4i_sdcon in the lichee-dev branch.
<aexl>
hno: ah yes i remember. and you said something about having console over sd and still be able to boot from sd.
<hno>
you can't. But you can in theory boot u-boot (and kernel) from SD and then switch to console for the kernel. But now with fel booting available it's not really needed.
<hno>
easier to just load u-boot & kernel (+ initrd if needed) over fel than switching between sd & uart adaper.
<aexl>
hno: hm. i thought you said something about pulling GND to devices or so.
<hno>
No, but I may have said something about shorting certain NAND lines to GND for disabling NAND so you can boot from MMC2 if you wire up those signals to an MMC like I have done on my Cubieboard
<hramrach_>
hno: nice. the full size slots come useful
<hno>
But I forgot to add that NAND disable wire before gluing on the MMC adapter.. but it's not a big deal, can skip the NAND boot step using JTAG to test BROM loading from MMC2, or use a screwdriver to short the right NAND lines at boot..
<hramrach_>
hw hackers ..
<aexl>
hno: but that includes opening my tablet right? :D
<hramrach_>
cubieboard comes open ;-)
<aexl>
hramrach_: one a10 device is enough for me.
<hno>
There is Samsung class 10 64GB micro-sd cards.
<aexl>
hno: thx for the info.
<hramrach_>
the mmc on a10 is also ~ class 10 so no need to look more ;-)
<hno>
And lexar SDXC 256GB UHS-1 rated.
<hno>
but costs more than all my Allwinner boards together.
<hramrach_>
lexar ..
<hno>
at ~2 EUR /GB.
<aexl>
Pretec SDXC 64GB, Class 16 :D
<hramrach_>
once I tried Lexar firewire CF reader. It did not work :/
<hno>
Class ratings varies greatly between vendors.
<hramrach_>
you could get the same functionality with IDE-CF and FW-IDE bridge but not in a usable case
<aexl>
oh the black thing right from the usb/ethernet socket right?
<aexl>
but the block diagram says "mmc0/1 8bit" and "mmc2 4bit"^^
Yaku321 has quit []
LoCoZeNoz_ZUE has quit [Ping timeout: 248 seconds]
LoCoZeNoz_ZUE has joined #linux-sunxi
<bfree>
mripard_: pong
aexl has left #linux-sunxi [#linux-sunxi]
LoCoZeNoz_ZUE has quit [Ping timeout: 276 seconds]
paulk-desktop has quit [Quit: Ex-Chat]
<mripard_>
bfree: see your inbox :)
LoCoZeNoz_ZUE has joined #linux-sunxi
<bfree>
mripard_: bah I'm blind, see it now (sorry guess the pastes died as didn't set them to very long). I'll give your suggested addresses a go now in a few minutes
<mripard_>
I really see only either the addresses or the configuration option regarding the compression that could cause this
<bfree>
just fyi (don't worry I'll check myself now) a proper debian linux-image kernel package should have the config in it ;)
fldc has quit [Ping timeout: 240 seconds]
<bfree>
it's not the config options anyway
fldc has joined #linux-sunxi
rz2k has joined #linux-sunxi
<bfree>
mripard_: http://paste.debian.net/9412/ :-( same error using the suggested addresses (well I'm presuming 0x4100000 for the kernel was a typo missing a 0 ... trying that it just fails at bootm in uboot with "Wrong Image Format for bootm command")
awafaa has quit [Ping timeout: 264 seconds]
<mripard_>
bfree: what's your config?
awafaa has joined #linux-sunxi
<bfree>
mripard_: the normal one is: http://paste.debian.net/9416/ ... will take me a minute to figure out where the one with the low level dubugging is, but it's basically the same
<mripard_>
I suggest you test my initrd and uimage first and then my initrd with your uimage and my uimage with your initramfs
<bfree>
mripard_: that seems revealing as I get the same with your uimage and uinitrd (could it be my dtb then or have I some sort of screwed up u-boot)!
<mripard_>
hmmm, odd
<mripard_>
I don't have a cubieboard here, so I never actually tested the cubie dt
<mripard_>
however, I know that Turl uses one to contribute
<mripard_>
I think it's working fine though
<mripard_>
I also use a way shorter cmdline
<mripard_>
try booting it with only the console set
<bfree>
yep, I tried that just in case
<mripard_>
ok
<mripard_>
with/without the NBD setup?
<bfree>
yes, just retried with only console, not even the panic/loglevel line
<mripard_>
ok, then I have the same thing here.
<mripard_>
Turl: could you try the above mentionned uImage/uInitrd on your cubie ? to double check
<bfree>
leaving devicetree, u-boot and hardware for me :-/
<mripard_>
well, if Turl is able to boot it, you would get a working dtb as well :)
<mripard_>
and presumably, he has a working u-boot as well
<mripard_>
but it looks too "clean" to be an hardware issue
<mripard_>
at least now we know that it's not a kernel/initramfs issue.
<bfree>
yep! boy do I hope I haven't had a stupid pebkac moment somewhere after dragging you (and now Turl hopefully) through this!
<mripard_>
don't worry, we all do some stupid mistakes :)
<Turl>
what's up :)
<Turl>
I don't have my cubie in here but I can give them a try tomorrow morning
<bfree>
Turl: so if you can boot mripard's uImage+uInitrd I'd love your dtb (and u-boot) to narrow down why I can't! and even if you can't it will help save my sanity ;)
<Turl>
the dtb I use is built from the linux tree
<Turl>
the uboot.. is something I built around nov/dec last year, burned to the card and forgot about :)
<Turl>
I don't even use the uboot console these days, I have it set up to autorun, pull kernel and dtb from tftp and boot
<hno>
bfree, which compression method are you using for the initrd? And how big is the uncompressed initrd? Have you tried moving it a bit later in memory?
<bfree>
Turl: so is my dtb ... I'd jumped up to sunxi-current u-boot for it's devicetree support ... I can go back and double check again with an older one (e.g. my ~March debian packaged one)
<hno>
also try using an uncompressed initramfs.
<bfree>
hno: gzip for the initramfs itself, none for mkimage. but I'm also failing with mripard's uImage+uInitrd
<mripard_>
hno: he enabled all the compression options in the kernel, and even the smaller dtb I provided doesn't work on his board, while it works fine here.
fldc has quit [Ping timeout: 245 seconds]
<hno>
bfree, and it wirks if the initramfs is embedded in the kernel?
<bfree>
well to be strictly accurate debian's kernel team enabled all the compression options ;) /me standing on the shoulders of giants
<bfree>
hno: with a small enough initramfs embedded yes, not with the full one though
<hno>
how big is the uInitrd file? And how are you loading it?
<bfree>
hno: sorry will double check the uncompressed size ... ~12M for mine. and yes I tried all sorts of moving around of memory addresses
<hno>
12 MB uncompressed is not big.
fldc has joined #linux-sunxi
rz2k has quit []
<bfree>
hno: might be simpler/safer to talk about mripard's uInitrd which is 4.1M. http://paste.debian.net/9423/ is how I tried to load it
<hno>
Nothing obviously wrong there.
fldc has quit [Ping timeout: 240 seconds]
n01 has quit [Ping timeout: 264 seconds]
<oliv3r>
hno: i cant use adb since android appears to be crashing, doubt there is some android recovery pre-installed?
<oliv3r>
does your olimex a20 boot android?
<oliv3r>
hno: anyway, i have started on the lkcl patches, but would first want to verify timings
<oliv3r>
hno: also what could cause fel-write to fail? read seems to work.error was -7
lioka has joined #linux-sunxi
<specing>
My first initrd was like 200 KB :D
<hno>
oliv3r, I have cleaned that pach up a bit in wip/a20 in my git.
<hno>
fel write generally fails if dram is not set right, or if you try writing outside ram regions.
<hno>
bfree, what dtb are you using?
<bfree>
I could cut mine back a fair bit through the initramfs-tools config (so it includes less modules) and dropping busybox, though the "big" one is what I would like to work. It's just a default debian unstable initramfs-tools+nbd(+busybox as nbd needs it)
<bfree>
hno: the cubieboard one built as part of the debian kernel package build
<bfree>
3.10+emac
<oliv3r>
hno: wel i noot using the boot1 + '2' trick