<mnemoc>
akaizen: but still an evil-rockchip device
Holo_ has quit [Ping timeout: 264 seconds]
<akaizen>
mnemoc: true, can you extract needed files from android ROM?
<rm>
this is a scam
<mnemoc>
akaizen: to build your own android rom, maybe
<mnemoc>
akaizen: to run linux, no
<rm>
the seller has no feedback
<akaizen>
rm: thats what a couple people said, but aliexpress / alibaba has a guarantee that you wont get scammed
<mnemoc>
they don't release the money to the seller until you are happy
<rm>
indeed, but maybe they are trying because they know of some way
<mysteryname>
They're way too cheap to be legitimate... akaizen: there is the escrow service, however it'll be a long wait till anything happens.
<akaizen>
mnemoc: hmmm i thought android 4.x + merged with mainline kernel, so necessary kernel modules cant be use in regular linux?
<mysteryname>
They could have a really short "acceptance" time and it might be automatically verified. I know Tom cubie had 15 days till shipping.
<mnemoc>
akaizen: you can't extract kernel sources from a blob ;-)
Thomas42 has joined #arm-netbook
<akaizen>
oh, you expect that its baked into the kernel blob
<akaizen>
some android (my HTC phone) has modules but its also very old (2+ yrs)
<akaizen>
guess ill have to get my hands on a rockchip image and see
<mnemoc>
rk is a wellknown gpl violator
<akaizen>
Or wait for Allwinner to release a 28nm Dual / Quad A9
<mnemoc>
from what i've read the next from allwinner will be a quad cortex-a7, not a9
<RaYmAn>
which will ne roughly just as bad wrt gpl violations
<RaYmAn>
:/
<akaizen>
All of China seems to be a GPL violator ;/
Holo_ has joined #arm-netbook
<RaYmAn>
Probably because it's unrealistic to pursue the GPL violations *sigh*
<RaYmAn>
I wonder if it would change anything if people went all-out on resellers wrt GPL
<mnemoc>
gpl is only honored by those who want to honor it
<akaizen>
And according to RMS: you either violate GPL or don't no 'degrees'
<akaizen>
Well GNU just needs a bunch of lawyers on retainer
<RaYmAn>
akaizen: it's still almost impossible to pursue it in china - from what I hear anyways
<akaizen>
can you image if FSF had as many lawyers as Apple lol
<akaizen>
imagine**
<mnemoc>
afaik tom shipped the rest of the cubieboards with a buildroot based system built from the open sources
<RaYmAn>
cool :)
<mnemoc>
except mine :<
<akaizen>
Tom used to work for Allwinner right?
<mnemoc>
akaizen: yes
<RaYmAn>
mnemoc: I'm sure you can get a copy though ;)
<akaizen>
So he re-implemented necessary parts for OSS version of system?
<akaizen>
mnemoc: Is he a hardware guy?
<mnemoc>
RaYmAn: I'm stuck on 2 options. 1) wait until hno gets uboot working there, 2) wait until tom revives and send me a dd-able phoenixcard installer
<mysteryname>
"used to work" so he left to make his own boards?
<RaYmAn>
mnemoc: u-boot fails? :S
<lundman>
hmm
<mnemoc>
akaizen: tom was working on the bootloader and then on the smp stuff for sun6i
<mnemoc>
RaYmAn: at least on the 1GB model it does
<mnemoc>
RaYmAn: gets hung after SPL
<RaYmAn>
meh
<RaYmAn>
did you dump boot0/boot1 to compare memory params?
<mnemoc>
mysteryname: he claimed family reasons. i believe he is moving to a different city
<mnemoc>
RaYmAn: if you tell me how to get boot0/boot1 using fel...
<RaYmAn>
mnemoc: oh, I thought you had jtag..
<mnemoc>
i have the adapter, but my flyswatter is dead
<RaYmAn>
ah :/
<mysteryname>
mnemoc: I see, so what is his current involvement in the A10 chips from the point he moves?
<RaYmAn>
presumably none since he quit his job? :P
<mnemoc>
iirc some of his former coworkers are also $involved$ in the cubieboard
<mnemoc>
but they don't speak english
<mysteryname>
presumably, he does seem very invested in the product with the cubie board.
<akaizen>
know of any info on how to get to kernel on all these ARM devices? from my understanding uboot -> sdcard -> kernel
<akaizen>
ah ok, so there were no oscilloscope, probes and logic tracing?
<mnemoc>
nope
<mnemoc>
but current u-boot work involves a lot of jtag
<mnemoc>
because tom's uboot-mmc had everything hardcoded and that limited it's usage across sunxi devices
<mnemoc>
akaizen: do you want to import that knowledge from the RT wiki into linux-sunxi's? :)
<akaizen>
Sure, I got the templates for boards and SoC working last night
eFfeM has quit [Client Quit]
<akaizen>
but then I decided to reorganize the info in a better way
<mnemoc>
:)
* akaizen
off to sleep
<akaizen>
more wiki and boot stuff later
<mnemoc>
:)
<mnemoc>
good night akaizen
Quinto has joined #arm-netbook
<Quinto>
hi i have imx53 QSB with Linux, i have modified my mx53_loco.c for set SD2 pin for to use a wireless module but i have some problem, who help me?
<zoobab_>
hi
<zoobab_>
I am trying to upgrade a MiniX with this:
<cat1>
mnemoc: do i read your last comment as this: "yes, we'd rather keep these as cmdline params"? sorry for being slow though..
<rm>
512 MB RAM :/
<mnemoc>
cat1: no problem. i expressed myself badly
<penguin42>
'Notes: our item(s) don't ship to italy.' what a bizarre restriction
<RaYmAn>
not really - a lot of ebay has that restriction because apparently italy customs are horrific
<cat1>
mnemoc: and the answer is?.. ;)
<mnemoc>
penguin42: unless you have lived in italy ;-)
<rm>
poor italians
<penguin42>
RaYmAn: Ah
<mnemoc>
penguin42: packages have an strong tendency to get stolen
<mnemoc>
penguin42: or simply trashed to avoid the hassle of delivering them
<mnemoc>
cat1: i rather keep those as cmdline params, BUT been careful about what the values they can take... which is basically default or 0
<mnemoc>
cat1: and I hate the prefix with all my guts :<
<cat1>
mnemoc: which one, sunxi?
<mnemoc>
yes
<cat1>
mnemoc: but how would you distinguish with other fbmem variants if any?
<mnemoc>
there is no way it could be confused with other driver as it's a platform thing
<cat1>
mnemoc: ok
<mnemoc>
i mean
<mnemoc>
in a perfect multiplatform world, yes, we need prefixing
<mnemoc>
but... i don't know
<mnemoc>
in that perfect multiplatform world the prefix would be the driver name
<cat1>
mnemoc: will keep them as you suggested for now.. just wonder if i need to come up with patch on top of patch or just patch, the latter will cost me force update on remote side :)
<mnemoc>
there is only one working copy involved
Gumboot_ has quit [Ping timeout: 240 seconds]
<mnemoc>
so I would prefer to get the fix squashed :p
<mnemoc>
also, you made this for your 3.6, while I'm going to merge them into 3.4
<cat1>
mnemoc: true
<Yakuzzi>
i wonder if the offers of that guy are for real, this looks too good to be true
<mnemoc>
because the datasheet is likely to have mistakes
<vinifr>
sorry, i mean the address map, like: "#define GPIOAOUT 0x1060"
<mnemoc>
afaik there is no such thing
<mnemoc>
GPIOs in allwinner's kernel are configured by script.bin,
<mnemoc>
using port notation, not a number
<vinifr>
so, could i use: "gpio_set_value(103, 1);"
<mnemoc>
you can't
<vinifr>
PB2103I/OPWM
<mnemoc>
and?
<vinifr>
so, could i use: "gpio_set_value(PB2, 1);"?
<mnemoc>
no, you can't
<mnemoc>
until the core of the allwinner kernel is entirely rewritten
von_fritz has joined #arm-netbook
<mnemoc>
what you can do is to enable the sun4i-gpio driver and pre-configure the pins in script.bin
<mnemoc>
[gpio_para] section
Quarx has quit []
<mnemoc>
then you'll get some basic control over them over /sys
<mnemoc>
but only those pin you set in the [gpio_para] section of script.bin
<mnemoc>
these are very flexible SoCs. but the kernel code is design for strictly preconfigured products
n6pfk has quit [Remote host closed the connection]
<mnemoc>
if you do C, and assuming the risk of been able to easily fry your board, it's compoennts and anything connected to it, you can take a look at pio.c in https://github.com/amery/sunxi-tools
<mnemoc>
and implement something to allow raw changes over /dev/mem
<RaYmAn>
the script.bin implementation does have on advantage - you can pretty much take any generic kernel with the right modules enabled and it'll boot, regardless of how things are connected
<RaYmAn>
that is a frigging pain on tegra...Have to hack up board files to get things even vaguely booting :P
<RaYmAn>
but I suppose device trees do the same, and in a much nicer way
<mnemoc>
DT + pinctrl allows the same
<RaYmAn>
yeah
<vinifr>
1 question, how do i edit the script.bin?
<vinifr>
:)
<mnemoc>
vinifr: same repo, bin2fex and fex2bin tools
<cat1>
guys, maybe this was answered already, but what one can do about this error:
<cat1>
[ 163.530] (EE) MALI(0): [maliModifyPixmapHeader:205] UMP failed to retrieve secure id
<RaYmAn>
what are you trying to do?
<cat1>
RaYmAn: just start xorg :)
<cat1>
nothing special though
<RaYmAn>
mali drivers aren't exactly working well yet
<RaYmAn>
as far as I can understand at least
<cat1>
so i may consider this error harmless?
* cat1
does not have access to hdmi capable screen and cannot see what is happening there :)
<RaYmAn>
I doubt it's harmless. It probably means a lot of stuff doesn't work - but as far as I know, a lot of stuff doesn't work :P as in, most stuff.
<cat1>
RaYmAn: actually xorg dies once in a while, and i believe this is the reason..
<cat1>
but not sure
gimli_ has quit [Quit: Verlassend]
<RaYmAn>
what drivers are you using?
Yakuzzi has quit [Remote host closed the connection]
Yakuzzi has joined #arm-netbook
<mnemoc>
cat1: poke techn or rz2k
<cat1>
kernel from next_mali and userspace r3p1.. well, it does not really match the kernel version but i think it does not really matter for this problem
<RaYmAn>
right, so known broken drivers :P
<RaYmAn>
but yeah, techn or rz2k
<RaYmAn>
:)
<cat1>
:D
<techn>
zup
<techn>
next_mali is for r3p0
<cat1>
techn: right
<cat1>
i just do not see that much difference, ioctl should be same
<techn>
yep.. they could fit together.. but anyway libs we got are not ment for x11
akaizen has quit [Ping timeout: 245 seconds]
<cat1>
btw, i had another problem with ump memory allocations, luckily fix was easy:
<cat1>
modprobe ump ump_memory_size=256000000
<cat1>
techn: actually what i want to achieve now is just to keep xorg alive
<cat1>
do not expect from it anything more atm :)
<Turl>
cat1: the mem allocation on the libs is hardcoded to 64M as far as I recall
<Turl>
do you really want to allocate half/a quarter of the ram to the gpu anyway? :P
<cat1>
Turl: i am talking about ump not mali
Yakuzzi has quit []
<techn>
cat1: but for that secureid(buf2) problem you could apply patch from my repo
<cat1>
Turl: is it just safety limit, will be tuned later
<Turl>
cat1: the userspace libs have the address and size hardcoded on them afaik
<cat1>
Turl: hmm..
<Turl>
I tried to move the allocation once and it complained because the address wasn't the one it had hardcoded on the libs
<captainigloo>
yep that on, but i tried with armel version
<captainigloo>
as i'm using a toolchain for armv7a softfp
<techn>
hmm.. you could check build command from binary
<techn>
if it's ment for x11
<captainigloo>
how to do that ?
<techn>
I used just notepad and it's there as plain text
<captainigloo>
i'm sorry what do you mean, which build command ?
<zoobab_>
@mnemoc how can I dump the NAND on the minix?
<zoobab_>
minix has some OTG with probably adb shell
<zoobab_>
don't know if I need a particular usb cable
<mnemoc>
zoobab_: please call it miniX, not minix. minix is an OS
<techn>
captainigloo: in that binary you can find build flags.. it's in plain text format
<captainigloo>
ahhh ok :)
<mnemoc>
zoobab_: i've never used that device, but I suppose it's just like any other A10 with android
<techn>
good night
<zoobab_>
miniX sorry
<mnemoc>
zoobab_: the stock gpl-violating kernel does let you see the whole logical nand, so you have to go partition my partition (/dev/block/nand[a-j]) dd'ing it to a file in /sdcard/
<zoobab_>
how can I access it if I do not have adb shell access?
<zoobab_>
if I install consolebot on it, with local shell I do not think I have enough rights to dump the nand
<captainigloo>
techn: yes you're right armel version is not meant for X11 !
<mnemoc>
zoobab_: all A10 devices I know come with `su`
<zoobab_>
trying
<mnemoc>
zoobab_: without root access you can't make a backup
<cat1>
techn: heh, was trying to mix r2p4 user with r3p0 kernel -- no luck! get this:
<cat1>
[nemo@localhost ~]$ glestest
<cat1>
UMP: ump_arch_open() failed to open UMP device driver
revident has quit [Quit: Combustible lemons? Bah, I bring you weaponized asparagus!]
<mnemoc>
cat1: that smells to permissions
<cat1>
mnemoc: good idea!
alcides has joined #arm-netbook
alcides has joined #arm-netbook
alcides has quit [Changing host]
<cat1>
now it confirms incompatibility:
<cat1>
Device driver API mismatch
<cat1>
Device driver API version: 14
vgrade has joined #arm-netbook
<zoobab_>
dumping the nand* partitions
<Turl>
cat1: yeah don't mix mali versions
<Turl>
that's asking for trouble :)
merbzt has joined #arm-netbook
<zoobab_>
is there a more recent rootfs then linaro-precise-ubuntu-desktop-20120524-177.tar.gz ?
<mnemoc>
there is a version from sept.
<zoobab_>
url?
<mnemoc>
on the phone, sorry :<
<zoobab_>
ok
ZaEarl has quit [Quit: Ex-Chat]
VarmVaffel has quit [Read error: Connection reset by peer]
vinifr has quit [Quit: Page closed]
sspiff has quit [Remote host closed the connection]
sspiff has joined #arm-netbook
sspiff has quit [Changing host]
sspiff has joined #arm-netbook
jlj has quit [Ping timeout: 265 seconds]
ZaEarl has joined #arm-netbook
mysteryname has joined #arm-netbook
<L84Supper>
http://tinyurl.com/8cbe42z A10 Mini PC's in a USB stick are down to $33, are they dumping the A10's in favor of something else already?
<traeak>
a13 ?
<Turl>
dual core chips? I saw some dual core "mini pc"s
<traeak>
rockchip maybe?
<Turl>
yeah possibly
<Turl>
if empat0 xmbc works as advertised though, these might make a sweet deal for some :)
<projectgus>
there could be other explanations, like cheaper ram prices or economies of scale
<traeak>
empat0?
<Turl>
I'm asking a random seller on alibaba now
<projectgus>
in my experience alibaba sellers will tell you anything non-commital that they think you want to hear
<L84Supper>
the order button returns with: Cubieboard is sold out now. Please add your email address to our list. We will inform you when cubieboard is available.
<L84Supper>
does they only have a limited supply of parts?
<penguin42>
L84Supper: I think they made a first batch of 100 or something, and that sold out
<penguin42>
L84Supper: It's always a good idea to do one batch first and see what the problems are
<L84Supper>
we usually make a batch to test in-house before unleashing them on the public
<penguin42>
L84Supper: I guess also it's difficult to guage how popular they would be
<L84Supper>
maybe
<zoobab_>
got the X running on the miniX
<zoobab_>
in 1280x800 though
<zoobab_>
any idea if it can go higher?
<zoobab_>
installing chromium on the miniX
<ZaEarl>
I don't see why not. All A10s should be able to do at least 1920x1080.
<lundman>
yeah but they cant :)
<lundman>
(right now)
<ZaEarl>
due to framebuffer driver limits?
von_fritz has quit [Quit: vonfritz leaves, don't panic]