<ZaEarl>
well, it got further, but now a error without details:
<ZaEarl>
Fetching projects: 100% (319/319), done.
<ZaEarl>
including device/allwinner/zatab/vendorsetup.sh
<ZaEarl>
including device/generic/armv7-a/vendorsetup.sh
<ZaEarl>
including device/ti/panda/vendorsetup.sh
<ZaEarl>
including device/generic/armv7-a-neon/vendorsetup.sh
<ZaEarl>
including vendor/cm/vendorsetup.sh
<ZaEarl>
including sdk/bash_completion/adb.bash
<ZaEarl>
including vendor/cm/vendorsetup.sh
<ZaEarl>
make: *** [android-build] Error 1
<lundman>
disk full!
<ZaEarl>
argh, what good is error 1 without a line number?
<Turl>
ZaEarl: prepend echo to the last line of the script and see if it runs
<lundman>
also, shell "-x" option is handy, like bash -x ./some_cript
<lundman>
or simply add -x to the shellbang
<ZaEarl>
-x nice
t0dbld|work has quit [Ping timeout: 255 seconds]
drachensun has quit [Quit: Leaving]
<ZaEarl>
+ unset f
<ZaEarl>
+ '[' 1 -eq 0 ']'
<ZaEarl>
+ grep -q -
<ZaEarl>
+ echo zzatab
<ZaEarl>
make: *** [android-build] Error 1
tekzilla has quit [Ping timeout: 260 seconds]
<Turl>
o.O
tekzilla has joined #arm-netbook
<lundman>
the grep -q - liine probably uses a $ENVNAME or somekind
<lundman>
that is empty
Maqs has quit [Ping timeout: 246 seconds]
Maqs has joined #arm-netbook
slash_random has quit [Ping timeout: 256 seconds]
<Turl>
ZaEarl: there, it should work now
<Turl>
update the bsp and try again :)
<ZaEarl>
Turl, make[1]: *** No rule to make target `vendor/cm/proprietary/Term.apk', needed by `/home/malmrose/bsp/android/out/target/product/zatab/system/app/Term.apk'. Stop.
<ZaEarl>
am I going to need a zatab on usb for pulling files off it?
<Turl>
no
<Turl>
I made a repo to put the props in
<ZaEarl>
ok, looks like android is compiling now
<Turl>
and included it on the manifest
<ZaEarl>
I've got all 4 cores pegged. Does it default to something like -j 8?
<Turl>
it defaults to -j `cat /proc/cpuinfo | grep "^processor" | wc -l`
<ZaEarl>
that should be ok
<Turl>
it's also run as "schedtool -B -n 1 -e ionice -n 1 make ..." so you should still be able to use the machine
avernos has quit [Quit: Leaving]
avernos has joined #arm-netbook
avernos has quit [Changing host]
avernos has joined #arm-netbook
<Turl>
WarheadsSE: any ETA on an A10-compatible XBMC arch package? :P
gimli has joined #arm-netbook
n6pfk has quit [Remote host closed the connection]
<ZaEarl>
building "external/llvm/lib/Target/X86/...". Is that really to cross compile x86 on arm? lol
<ZaEarl>
time to setup a beowolf cluster of cubieboards
Sv has quit [Read error: Connection reset by peer]
Sv has joined #arm-netbook
Sv has joined #arm-netbook
Sv has quit [Changing host]
eebrah has quit [Remote host closed the connection]
mikey_w has quit [Remote host closed the connection]
mikey_w has joined #arm-netbook
n6pfk has joined #arm-netbook
specing has quit [Ping timeout: 252 seconds]
<ZaEarl>
Turl, almost 2 hrs to build on my laptop. .zip created.
specing has joined #arm-netbook
specing has quit [Changing host]
specing has joined #arm-netbook
AndChat138129 has joined #arm-netbook
gimli has quit [Quit: Verlassend]
AndChat|138129 has joined #arm-netbook
AndChat138129 has quit [Read error: Connection reset by peer]
rellla has joined #arm-netbook
orly_owl has joined #arm-netbook
Quarx has joined #arm-netbook
rellla has quit [Ping timeout: 255 seconds]
rellla has joined #arm-netbook
freakazoid0223 has quit [Ping timeout: 240 seconds]
<lundman>
500 units seems a bit stiff
pawel5870 has joined #arm-netbook
<stefanro>
hno: i just noticed that you added the sunxi board so the MAINTAINERS file - good
<stefanro>
hno: nitpick: you should use the board target name here (meaning CUBIEBOARD -> cubieboard)
kaspter has joined #arm-netbook
mSquare has quit [Ping timeout: 260 seconds]
mSquare has joined #arm-netbook
sspiff has joined #arm-netbook
cat_x301 has quit [Ping timeout: 256 seconds]
cat_x301 has joined #arm-netbook
tzafrir_laptop has quit [Ping timeout: 244 seconds]
Thomas42 has quit [Read error: Operation timed out]
avernos has quit [Ping timeout: 246 seconds]
avernos has joined #arm-netbook
avernos has quit [Changing host]
avernos has joined #arm-netbook
Thomas42 has joined #arm-netbook
avernos has quit [Ping timeout: 246 seconds]
avernos has joined #arm-netbook
avernos has quit [Changing host]
avernos has joined #arm-netbook
mysteryname has joined #arm-netbook
<oliv3r>
Have the 4k 'notify me' been notified allready?
AndChat|138129 has quit [Ping timeout: 252 seconds]
deepblue has joined #arm-netbook
<mnemoc>
oliv3r: Tom is still offline :|
<RaYmAn>
33k :)
<RaYmAn>
Does seem like it'll reach it just fine :P
<oliv3r>
it's 'rude' for those 4k
<RaYmAn>
It's kind of an issue with paypal though. (also, I'm sure a good deal of those 4k has already seen the project on indiegogo)
AndChat138129 has joined #arm-netbook
<oliv3r>
also true
<oliv3r>
it just reflects as messy and bad to those interested
freakazoid0223 has joined #arm-netbook
<mnemoc>
the problem with those 33k is that Tom won't receive them until the end of december
AndChat138129 has quit [Ping timeout: 252 seconds]
AndChat138129 has joined #arm-netbook
<Turl>
ZaEarl: good, thanks
AndChat138129 has quit [Ping timeout: 252 seconds]
anton_____ has joined #arm-netbook
<ZaEarl>
that bsp is a really nice setup. so easy to do.
<mnemoc>
:)
<ZaEarl>
does it generate a livesuit img?
<oliv3r>
you mean when those 38 days are over?
<mnemoc>
oliv3r: + 5 days for indiegogo to process, + a week for the wire transfer
<mnemoc>
ZaEarl: not yet
<oliv3r>
ah, yeah december at the earliest
<sspiff>
is there any documentation on how to build CM/AOSP roms for the A10, that you can flash using that Chinese flash tool they use? (I forgot the name)
AndChat138129 has joined #arm-netbook
<mnemoc>
our CM currently creates standard android images only, to be flashed from CWM
<sspiff>
then is there an easy way to install CWM recovery to a stock A10 device?
<mnemoc>
Turl: ---^
<Turl>
sspiff: just dd the CWM image to your device's recovery partition :)
<Turl>
(or use cat image > /dev/block/.. if you don't have dd)
<ZaEarl>
but you need a CWM built for your device
tzafrir_laptop has joined #arm-netbook
Quarx has quit [Read error: Connection reset by peer]
Quarx has joined #arm-netbook
cheng has joined #arm-netbook
AndChat|138129 has joined #arm-netbook
<oliv3r>
gonna sell my pivos
<oliv3r>
was a gift but i'm not gonna do much with it, i'm to invested in a10 stuff :)
<oliv3r>
anybody want int? jinzo? :p
<lundman>
huh, I've all but abandonned a10
<lundman>
i was going to get a pivos, but they wouldnt sell overseas :)
<oliv3r>
i'm in NL, bought one before getting into the A10
<oliv3r>
so now i have a pivos, but not doing anything with it
<oliv3r>
checking the pivos github page, hardly any forks, very little development, community around the amlogic seems non-existant
<lundman>
yeah but a10 is a deadend :)
<oliv3r>
so why bother?
<oliv3r>
lundman: how so?
<oliv3r>
linux-sunxi.org/XBMC gives hope?
<mysteryname>
yes.. Hope :D
<oliv3r>
mysteryname: do i smell some sarcasm there? :)
<mysteryname>
yes a little however I'm rather invested in the a10s
AndChat138129 has quit [Ping timeout: 252 seconds]
<mysteryname>
I have two sitting on my table and all of this is over my head so I'm following what I can
<oliv3r>
so how is hope bad?
AndChat|138129 has quit [Ping timeout: 252 seconds]
<mysteryname>
sorry, hope is not bad. I fear it's not enough :P
<lundman>
yeah I think it could just-about-almost be amedialpayer
<oliv3r>
mysteryname: what else would be needed you think?
<lundman>
but it will never be enough
<oliv3r>
lundman: why not?
<lundman>
so, personally, I'm looking at teh duals or quads, not that they are out
<mysteryname>
oliver3: I think if it makes a breakthrough for the hardware acceleration, it will ignite.
<oliv3r>
so your complaint is libcedarx driver support
<mysteryname>
lundman: I think it can do it and it can when using android. but I want linux so I can have more power over it.
<oliv3r>
lundman: well lets assume the a40 will be available in an mele a4000 in a year or so
<oliv3r>
a lot of work done here will still be relevant
<lundman>
oli: yes, definitely. I said I gave up on A10, not allwinner
<mysteryname>
oliv3r: agreed.
<lundman>
for me, to be a media player, it needs to be 1080p, and do more than stereo. but a10, everything is geared towards ICS
<lundman>
mystery: I'm hoping for android, in that, maaayybee google can do a good itunes replacement, ie, get tv and movies from play
<lundman>
then to have a legal solution finally would be nice
<mysteryname>
It can do that in ICS right?
<Turl>
lundman: google already does that with play :P you can buy movies and music and books
<mnemoc>
google play rents (overpriced) movies and mosuic in mose countries
<lundman>
turl: alas, only for some areas
<mysteryname>
indeed.
<oliv3r>
lundman: stereo is fine for me
<lundman>
it shouldnt matter one bit where I live, if I am willing to pay
<Turl>
lundman: go tell that to copyright owners
<oliv3r>
1080p it should do more then easily (it does quad HD iirc) so 1080p stereo = fine :)
<oliv3r>
doesn't it do dolby passthrough however?
<Turl>
lundman: you alternatively can use a VPN to USA and buy still
<lundman>
oliver: then if quality doesnt matter, you might be happy with A10. half the framerate, steam poewr
<mysteryname>
lundman: for sure.
<lundman>
turl: its an exercise in getting shit legally, so to illegally gain access to the shop would defeat the purpose
<mnemoc>
lundman: it's mostly due to the contracts each copyright owner has with distributors on each country
<lundman>
I am already the biggest pirate in asia
<mysteryname>
Turl: sometimes credit card numbers can be country orientated, I tried to buy in the US once and It knew I had a AUS card.
* lundman
lies
<lundman>
mnemoc: oh i know why it is, it is almost very out dates and stupid
<lundman>
and I understand trhey dont want to change what they are already making millions on
<mysteryname>
lundman: if too many companies resist change it'll bit them hard.
<mysteryname>
lundman: look at dymocks (I don't know how it;s spelt) they resisted change and went from the biggest book store to bankrupt.
<lundman>
we've gone through a decade of "piracy is the root of all evil" but now we are finally changing, itunes, amazon video, yahoo video, google video.. just needs to get ridof region blocking :)
<Turl>
and DRM
<mysteryname>
agreed. Mega upload had something ground breaking going till the FBI rocked up and stole everything
<lundman>
the worst is itunes japan "oh hi, we have jpop!" "what about tv?" "jpop!" "ok, movies?" "sure, dubbed!"
<mnemoc>
google play only has dubbed movies here in spain :<
<oliv3r>
half the framerate? you say it DOES do quadHD video decoding, but only at 15FPS? then full hd should be fine @ 30fps. if it does 15fps at fullHD, what are those 4k HD claims then? 5FPS?
<lundman>
o9li: last test I did, on ahh beta3, it could not play bluray TS at more than 12fps, it could do "scene standard 1080p" most of the time, but had frame drops
<lundman>
1080p 4frames and less was ok, 720p was ok
<lundman>
very far from acceptable, if you actually care about entertainment
<oliv3r>
lundman: but passthrough surround? does it do that?
<lundman>
ah hmm, never tested, since xbmc did exist backthen
<lundman>
assuming that the allwinner kernel sources are up to it, it should be able to passthrough the normal stuff
<oliv3r>
LOL
<lundman>
but when I checked kernel sources, they didnt talk about the high bitrate, DTS-HD and Master
<oliv3r>
you _really_ should. the whole A10 world is changing extremly fast
<lundman>
yeah I'm waiting for someone to do a binary armhf for me
<RaYmAn>
mnemoc: technically, the only thing that gives is precompiled binaries and a fancy graphical installer ;) ALl their wine changes are (obviously) on their public source repos :P
<lundman>
i have a upload area and everything!
<mnemoc>
RaYmAn: doh
<RaYmAn>
the wonders of GPL :)
AndChat138129 has joined #arm-netbook
<lundman>
but the a10 would be great for the boys room
<Turl>
today's the 'download crossover for free' day right?
<sspiff>
Turl, ZaEarl: any idea where I can get such an image? I've got an MK802 and an Ainol Novo Aurora - are the CWM images different per device or only per SoC?
<sspiff>
also, is JB so fundamentally different from ICS that it won't work (well) on A10 hardware?
<Turl>
usually per device, as keyboard mapping varies
<Turl>
JB works fine on A10 hardware
popolon has joined #arm-netbook
<deepblue>
@Turl - question on JB. will it work with 512MB RAM? I have a LY/F1 clone
<Turl>
512M is a bit too tight in my opinion
<oliv3r>
sspiff: i run JB on my A10 tablet and little/no issues
<oliv3r>
(it sometimes doesn't sleep when i turn the display off, thus draining the battery to fast, could be an app causing that of course)
<sspiff>
oliv3r: hmmm, interesting, are there any play by play instructions to get that running?
<oliv3r>
and it sometimes doesn't wake (battery usage just says from full to null, could also be battery related)
<oliv3r>
for my specific tablet? yes :)
<sspiff>
Turl: how do I make or get such an image?
<oliv3r>
following that post/links there's an entire list of 'compatibility patches' (e.g. driver zips)
AndChat|138129 has joined #arm-netbook
<sspiff>
oliv3r: seems like my tablet is also listed, neat
AndChat138129 has quit [Ping timeout: 252 seconds]
AndChat|138129 has quit [Ping timeout: 252 seconds]
furan has quit [Ping timeout: 252 seconds]
ajmitch has quit [Ping timeout: 245 seconds]
ajmitch has joined #arm-netbook
ajmitch has quit [Changing host]
ajmitch has joined #arm-netbook
[trip] has joined #arm-netbook
furan has joined #arm-netbook
furan is now known as Guest65804
AndChat138129 has joined #arm-netbook
Almamuetya10 has joined #arm-netbook
jeremb has quit [Remote host closed the connection]
ZaEarl has quit [Ping timeout: 255 seconds]
jeremb has joined #arm-netbook
<mnemoc>
hno: just tried `sunxi-patchqueue` ... my cubie halts on spl. right after `Board: sunxi` .... sunxi? shouldn't it be cubieboard? .oO
AndChat|138129 has joined #arm-netbook
AndChat138129 has quit [Read error: Connection reset by peer]
slash_random has joined #arm-netbook
AndChat138129 has joined #arm-netbook
AndChat|138129 has quit [Ping timeout: 255 seconds]
<[trip]>
mnemoc: hows the bsp project going?
AndChat138129 has quit [Ping timeout: 255 seconds]
<mnemoc>
[trip]: still not ready to replace cnxsoft's, but works
<[trip]>
mnemoc: very cool. i just got my mele a1000 in, and i wanted to start helping dev on it.
<[trip]>
it's got a great set of hardware, it just needs the proper software to go with it now =)
<mnemoc>
:)
<[trip]>
are the current builds using the lima drivers or the cedarx binaries?
AndChat138129 has joined #arm-netbook
<[trip]>
I've read everything I could find on the state of the graphics drivers, but I wasn't able to pin anything solid down on where everything is at.
<mnemoc>
lima is alternative to mali closed userspace, and not ready to replace it yet
<mnemoc>
cedarx is video decoding
<[trip]>
I do know that my mele cuts off part of the screen when in VGA mode. It's off center.
<[trip]>
That makes more sense.
<[trip]>
CederX is a closed blob too I take it then?
AndChat138129 has quit [Ping timeout: 255 seconds]
pwhalen has quit [Read error: Connection reset by peer]
<[trip]>
mnemoc: what are you using as your build platform?
<mnemoc>
ubuntu 12.04/64
<[trip]>
I've got mint on my current build machine, but ubuntu on my server. That should work.
<[trip]>
mnemoc: are there still HDMI builds for the Mele?
<mnemoc>
the neat thing with ubuntu 12.04 is that sf and hf toolchains are just an apt-get install away
<mnemoc>
[trip]: eh?
<jinzo>
mnemoc, the linaro ones I presume?
<[trip]>
mnemoc: Sorry, I meant VGA builds for the Mele.
<mnemoc>
the default output is set on script.bin
<mnemoc>
jinzo: yes
<[trip]>
mnemoc: gotcha.
<[trip]>
I'm sure I'll be pestering the # with more questions soon, so I'll apologize in advance.
<mnemoc>
:)
<[trip]>
Once I get everything setup, I'll hopefully get some time in the evenings to hack on the core, and hopefully some of the drivers.
<mnemoc>
nice
<mnemoc>
help is always welcomed
<[trip]>
I know nothing about the Mali architecture, but there's always room to learn something new.
<stefanro>
mnemoc: regarding the "sunxi" board name output in u-boot
<stefanro>
mnemoc: this is because of the board code consolidation i started a few days ago (never tested on real hw)
<[trip]>
And just randomly, have there been any theories on why the VGA output is off center on the Mele?
<stefanro>
mnemoc: and hno consolidated a bit more (most likely a bit too much ;))
<mnemoc>
stefanro: :)
<[trip]>
mnemoc: are the documents on cnx still the best set of instructions for getting a working sd card with ubuntu working on it?
<mnemoc>
[trip]: i had no idea mele's vga output was off center
<libv>
once that heap of ... roses. is clean, it will do free modes
<[trip]>
Maybe it's just mine. At 720p there's about a half inch of screen missing to the left, at 1024x768 there's about an inch, and at 800x600 it's actually off center to the right about an inch. Go figure.
pwhalen has joined #arm-netbook
<libv>
[trip]: that will different on many monitors, but with free modes you can adjust any way you like
AndChat138129 has joined #arm-netbook
<mnemoc>
[trip]: as you are here, i would suggest sunxi-bsp instead :)
<[trip]>
libv: interesting. i've only got it running android 4.0 right now, so it's a slight annoyance. I'm going to test it on my projector, an old monitor, and an old TV over the component out to see if I can get a stable resolution on one of them.
<[trip]>
mnemoc: just clone it, build, and then follow the old instructions to setup an SD card?
<mnemoc>
[trip]: android 4.0 with open or gpl-violating kernel?
<mnemoc>
[trip]: yes
<[trip]>
mnemoc: not sure, my mele was flashed with 4.0 on it already. it's got a bunch of chinese stuff that I can't read, so I'd go with GPL violating (sadly).
<stefanro>
mnemoc: let me know if this fixes your problem (i don't have the linaro toolchain installed)
<libv>
wow, someone really finds it important that copyright is with a capital C?
* libv
is amazed
* stefanro
was wondering about that too...
<mnemoc>
stefanro: /srv/build/amery/allwinner/sunxi-bsp-3.0/u-boot-sunxi/fs/zfs/zfs.c:2161: undefined reference to `__aeabi_uldivmod'
<stefanro>
mnemoc: then the u-boot zfs port needs fixing to not use FP i assume
<stefanro>
mnemoc: the other errors are gone now?
<mnemoc>
stefanro: let me disable zfs. 1m
<stefanro>
yes, i was just going to propose this
<mnemoc>
stefanro: USE_PRIVATE_LIBGCC=yes + disabling ZFS worked :)
<stefanro>
ufff ;)
<mnemoc>
libv: I doubt he really finds it important. it smells more like a "getting involved" trick :)
<oliv3r>
if its the first commit out of many, all you can say is 'welcome' :)
<libv>
he's a bit late to the disp party for beauty cleanups
<libv>
and i do not understand why he undoes the separate copyright and date for the danliang
<RaYmAn>
mnemoc: chromebook arrived yet? =P
<mnemoc>
what makes me ask.... how long before we merge stage/sunxi ?
<mnemoc>
RaYmAn: it's resting in TN :<
<RaYmAn>
aww
<RaYmAn>
did your co-worker receive it yet?
<mnemoc>
yes
<mnemoc>
on monday
<RaYmAn>
cool
<mnemoc>
:)
<RaYmAn>
mine is shipping around 10th, heh.
<mnemoc>
RaYmAn: your fault :p
<RaYmAn>
yes :/
<xenoxaos>
my chromebook.... "Weather conditions might delay delivery of this shipment. This is the most current information available. Please track again later for status updates."
* xenoxaos
grumbles
<RaYmAn>
aww
<xenoxaos>
FU Sandy
<RaYmAn>
hmm, that u-boot hardfp gerrit actually reveals one quite cool thing
<mnemoc>
slapin: are you still interested in doing the mtd nand driver for sunxi?
<oliv3r>
amlogic? oss in one sentence?
<slapin>
mnemoc: yeah, sure
<wingrime>
drachensun:I have power port field in my FEX that constain GPIO pin that enable power to TS
<slapin>
oliv3r: they released kernel and u-boot as allwinner and rockchip
<oliv3r>
slapin: but no VPU details?
<wingrime>
drachensunif GPIO state logical "0" ts powered off
<oliv3r>
slapin: amlogic doesn't have a usefull gitrepo, and their only link to the outside world, pivos, doesn't have much changes/community support imo
<slapin>
oliv3r: I don't know any details yet, they've just arrived, were ordered by accident.
<mnemoc>
slapin: the thing is we need to be able to extract the first raw blocks for the nand where boot0/boot1 lives
<wingrime>
drachensun:without this TS5x not answered over I2C
<slapin>
oliv3r: openlinux.amlogic.com, no?
avernos has quit [Ping timeout: 246 seconds]
<oliv3r>
mnemoc: if anything documentation wise would help in that regard, point me to the next section to do after the CCM :)
<mnemoc>
slapin: and maybe you could do your experiments to understand the nand controller with a baremetal tool + fel
<slapin>
mnemoc: I think so
<oliv3r>
slapin: last time i checked, openlinux wasn't that usefull
<oliv3r>
doesn't compare to linux-sunxi :)
<drachensun>
wingrime: I see, makes sense I guess, now mine is showing an error that it is not responding. Our devices are some how opposites
<mnemoc>
oliv3r: the only documentation we have about the nand controller is in the code
<oliv3r>
mnemoc: bleh
<drachensun>
wingrime: Of course it might just be mine needed a different pin and now its lost power
<mnemoc>
slapin: so.... can you help us with that? :)
<wingrime>
drachensun: best way to figure out what wrong - take a look with IDa
<wingrime>
drachensun: are you worked with IDA ?
<mnemoc>
slapin: sunxi-tools has a couple of other fel-able apps
<drachensun>
wingrime: I haven't, but I figured since mine did work before I just have to figure out what changed and then once I am sure which change was critical I will try to find a way to make both our devices happy
avernos has joined #arm-netbook
avernos has quit [Changing host]
avernos has joined #arm-netbook
<slapin>
mnemoc: of course, but I still don't understand what I have to do. I have u-boot source I play with, with some weird stuff
<slapin>
mnemoc: ?
<slapin>
mnemoc: any details?
<wingrime>
drachensun: I can take a quck look for you problem but I need ft5x_ts.so - working binary
<mnemoc>
slapin: initialize the dram controller, and the copy the first blocks of the nand to a memory address from where we can read it :)
<slapin>
mnemoc: what is fel, and what baremetal tool?
<mnemoc>
slapin: get the sunxi-tools repo
<mnemoc>
slapin: fel is a usb recovery mode in a1x's brom
<drachensun>
wingrime: Are all your changes committed? it looks like most of the changes likely to cause my problem are actually related to the A13-SDK commit
<wingrime>
drachecnsun:I can compare linux-sunxi driver code with assebler equvalent with vendor changes
<mnemoc>
slapin: you can write memory, read memory, and call memory
<mnemoc>
slapin: it's used by their flashing too
<drachensun>
wingrime: Oh, I get you, yeah let me post the driver somewhere
<mnemoc>
slapin: this runs very early, only sram and usb is initialized at that point
<wingrime>
drachensun: mostly you not find solution in commits, it can be some vendor driver fix
<slapin>
mnemoc: so fel can read NAND in raw?
<mnemoc>
slapin: no. fel lets you upload a baremetal app
<slapin>
mnemoc: doesn't u-boot environment is more friendly with this?
<slapin>
mnemoc: which baremetal app?
<mnemoc>
slapin: the one you are writting to experiment with the nand controller :)
<WarheadsSE>
Turl: I might be able to to get to that friday/monday\
<mnemoc>
slapin: we need a fel-based app so users can extract boot0/boot1 from their devices without a serial console
<mnemoc>
slapin: and later to be able to replace boot0/boot1 with our spl
<mnemoc>
slapin: but the boot1 is the only reliable source for dram initialization data
<wingrime>
mnemoc: I too intresed in dump/write tool like livesuit
<oliv3r>
PLL5 up on the wiki :)
<oliv3r>
hometime
<mnemoc>
oliv3r: thanks!
<mnemoc>
wingrime: livesuit pushes pre-made apps to do the job
<slapin>
mnemoc: I use u-boot in my work. I usually hate to write big ARM CPU init stuff, especially without documentation :-E~~~
<wingrime>
mnemoc: can it tools be pushen manualy from your fel tools?
<slapin>
mnemoc: serial console is easier :)
<mnemoc>
slapin: most devices and most people can't get a serial console
<RaYmAn>
mnemoc: the fun part will be that if you do it through fel, you need to either fit the driver in SRAM or initialize ram :P
<libv>
slapin: i just got a hyundai
<slapin>
and still, I don't quite understanf NFC inner workings to meka bare-bone app to read NAND.
<mnemoc>
RaYmAn: it should fit in sram....
<libv>
slapin: so i will get the lcd working
<mnemoc>
slapin: that's the fun part ;-) while you get this tool working you learn about the nand controller ;-)
<slapin>
serial console is just to solder 3 wires, and you think writing such an application is easier?
<mnemoc>
RaYmAn: "driver" wise, we only need raw access at this point
<drachensun>
wingrime: I sent you a pm with the link
<RaYmAn>
mnemoc: sure, it's just important to be aware of :)
<mnemoc>
slapin: soldering 3 wires on every allwinner-based device we want to set free is... hard
<slapin>
mnemoc: I think direct approach might get better experience; u-boot is scriptable and can be loaded with AllWinner's own strange boot loader, and scripts can dump whatever you want to uSD. and no bare-bone apps...
t0dbld has joined #arm-netbook
<wingrime>
slapin:UART much easier to debug than lcd
<slapin>
libv: it is so great!
<wingrime>
slapin:it easiest interface ever and universal over different tablets lcd
<slapin>
wingrime: somebody with serial console can script everything for poor guys without serial
* slapin
discovered loots of junk food and harvesting
<mnemoc>
slapin: making a card with a uboot that boots from mmc and can also read raw nand will work too, obviusly
<wingrime>
slapin:is it big deal to use mmc pin?
<RaYmAn>
mnemoc: not really
<RaYmAn>
unless it's in the u-boot SPL
<RaYmAn>
:P
<mnemoc>
RaYmAn: very good point!
<mnemoc>
slapin: -----&
<RaYmAn>
and we really can't expect people to just randomly swap out files on nanda to get it dumped :/
<mnemoc>
slapin: -----^
<RaYmAn>
(because then they have to livesuite or *chicken-egg problem* boot from sd to fix it)
<slapin>
AW A10 does have quitel large SRAM, doesn't it?
<mnemoc>
32k
<RaYmAn>
if we only need to dump boot0, we only need to read one block, don't we?
<mnemoc>
with need both
fredy has quit [Excess Flood]
<mnemoc>
but boot0 is a good start
<mnemoc>
fel fel fel fel
<mnemoc>
:)
<RaYmAn>
I wonder if there is a way to initialize the ram at some safe frequency and safe settings such that they'll at least work?
<slapin>
so current SPL will do as barebone app, doesn't it?
<mnemoc>
yes
fredy has joined #arm-netbook
<mnemoc>
but it needs to know how to initialize the dram, and we only know for certain devices
<mnemoc>
but you can steal the arm initialization stuff from there
<slapin>
mnemoc: and will very slow default work?
<mnemoc>
sure
<mnemoc>
it's a very small piece of nand :)
<slapin>
and what fel will do then?
<mnemoc>
fel is a protocol
<mnemoc>
to transfer stuff via usb
<slapin>
so how rill you use fel then then?
<slapin>
s/rill/will/
<ibot>
slapin meant: so how will you use fel then then?
<mnemoc>
you press some keys when powering up the device
<mnemoc>
connect the otg to your computer, and you'll see a usb device without name
<slapin>
mnemoc: this is not what I ask
<mnemoc>
slapin: then, please rephrase :)
<slapin>
mnemoc: You got hypothetical bare-bones app running, what then?
<slapin>
mnemoc: how to extract acquired data?
<mnemoc>
using fel again, from sram
<mnemoc>
then gets parsed and turned into something u-boot can use (binary header or something)
<slapin>
mnemoc: so bare-bone app needs to put these data into SRAM, then somehow give back control to fel?
<mnemoc>
afaik it doesn't need to "give back control"... simply die
<wingrime>
drachensun:are you still here?
<drachensun>
back
akaizen has quit []
<slapin>
mnemoc: ok, will see what happens...
<mnemoc>
\o/
<mnemoc>
in the worst case you should be able to start fel again, calling 0xffff0020
<RaYmAn>
mnemoc: no, it has to do "regular c thing" to return control. (which means branch to return address). If you do it as a single c file with a function that just ends, you're good. :)
<RaYmAn>
fel also sets up a c stack for you, so :)
<mnemoc>
how kind of it :)
<slapin>
well, can't start fel on my a10 stick, what exactly should be done to do so?
<mnemoc>
you need to start your device in fel mode, just like for using livesuit
<mnemoc>
more than one device tree repo per device?
tzafrir_laptop has quit [Ping timeout: 260 seconds]
<Turl>
yeah
<mnemoc>
:<
<Turl>
device trees are pretty much make once, use wherever you feel like, these compatibility "hwpacks" have random binaries and kernel modules you need to build from somewhere (ie you'd still need the device tree)
<Turl>
and you'd need to remake all the compatibility patches every time there's a kernel update or abi change
<Turl>
and besides you cannot change android overlayable configs with these (including random jars and apks would be even more of a maint. nightmare)
<libv>
hno: i could halt the linux easily, but not the mele itself
<libv>
hno: and a reboot also gets stuck
<libv>
so i would halt after a panic, and then press power until shutdown and power up again
Mazon has quit [Quit: meh]
<libv>
not having to reboot everytime for testing changes to disp is a big improvement :)
<mnemoc>
libv: the watchdog is supposed to reboot it
<libv>
does this correctly bring the device back up?
<Turl>
yep
<libv>
ok
<Turl>
I reboot my mele all the time, it works fine :<
<mnemoc>
i wouldn't be so sure...
<Turl>
libv: you can 'kexec' too
<mnemoc>
my cubie with current u-boot halts on spl when rebooting
<Turl>
I use ancient uboot
<techn>
Turl: so.. only good solution for android multidevice is to have it's own build system. nd use ?
<techn>
-nd use
<techn>
and that generates only full livesuite images per device?
<Turl>
it generates flashable zips
<Turl>
making a device tree is not hard at all, I don't get why is everyone so scared of them in here :<
<mnemoc>
Turl: I can do the bsp generate an initial device tree of each device, no problem with that
<mnemoc>
but need a template, and need to get `make android` use our kernel, our script.bin, ...
ZaEarl has joined #arm-netbook
<mnemoc>
but i'm really more inclined to have a single repo for all of them
freakazoid0223 has quit [Read error: Connection reset by peer]
<mnemoc>
content will be almost the same so maintaining them all in a single repo will be simpler
<Turl>
mnemoc: script.bin is on nanda, untouched
<mnemoc>
ok
<mnemoc>
android_device_allwinner ? :)
<techn>
Turl: so android can't be installed to sdcard?
<Turl>
it can, but it'd be wasting all that NAND space on android devices :)
<mnemoc>
one of the goals of the bsp is to separate building from image making
<mnemoc>
so we can cross the different build options with the different image options
<Turl>
mnemoc: the zip is pretty much location-agnostic
<libv>
slapin: how long does it take to charge your device?
<Turl>
mnemoc: you could make a sdcard with CWM on it that flashed zips to the sdcard :)
<libv>
here it seems to take about 5h already and it still isn't full
<Turl>
mnemoc: you just need to edit the recovery fstab
<mnemoc>
500mA?
<libv>
should be 2A from the psu
<Turl>
libv: zatab takes like a full 12h or so to charge
<libv>
wow
<mnemoc>
ouch
<libv>
well, i will soon find out just how big the battery is :)
<Turl>
well maybe a bit less too, but it's still more than an afternoon :P
<libv>
but it cannot be much, this thing seems much smaller than a current ainol
Mazon has joined #arm-netbook
<hno>
libv, the watchdog resets the CPU, so as long as the PMU haven't got completely messed up settings it will reset.
WarheadsSE has quit [Quit: WeeChat 0.3.9]
<libv>
is it only th mele that is suffering from failing poweroff or failing reboot?
<hno>
but it's maximum timeout is ~30 seconds and if you are doing kernel debugging with gdb then it's easily triggered.
<hno>
libv, no it's not only the mele. mnemoc also have problems rebooting his cubieboard, where it gets stuck in u-boot SPL after reboot.
<libv>
hno: are you reproducing this for yourself?
<techn>
has anyone got hno's u-boot working perfectly on tabled?
pawel5870 has quit [Ping timeout: 276 seconds]
<techn>
*tablet
lkcl has quit [Ping timeout: 255 seconds]
<mnemoc>
techn: it doesn't do battery initialization stuff yet
ZaEarl has quit [Read error: Connection reset by peer]
ZaEarl has joined #arm-netbook
<mnemoc>
techn: that's one of the reason why I'm bugging slapin about the tool to extract nand blocks using fel :)
<techn>
mnemoc: is that hard to do.. so it needs something from nand?
<mnemoc>
we need reliable data, and that only comes from boot1
<mnemoc>
not script.bin
<mnemoc>
shouldn't be that hard, having the time to dive into the nand controller code
<mnemoc>
but slapin seems like the perfect candidate for that task :)
<mnemoc>
not that I want to poke slapin :p
<Turl>
two 3.6.x on a week, Greg is on a roll :)
<mnemoc>
and... 3.0.50 >> 3.0.42
<mnemoc>
I'll merge stage this friday, tag, and then jump forward
<libv>
slapin: do you know the pinout of the serial?
<libv>
mnemoc: not for long, i am also taking pictures
<mnemoc>
:)
<libv>
slapin: my feeling is that it is the two middle ones that are tx/rx, and the right one on your picture (http://ossfans.org/a7hd_serial.jpg) is gnd
<mnemoc>
nice little board
WarheadsSE has joined #arm-netbook
pawel5870 has joined #arm-netbook
<drachensun>
techn: define perfect, I've got the newer u-boot running on my tablet
<techn>
drachensun: mine boot's but not every time.. It could be becouse that battery.. or not supporting different boot modes
<drachensun>
its actually the only way I can compile xbmc, with the older one I was getting seg faults randomly while compiling, but I'm not 100% sure of the cause of that though
<techn>
I just checked from source that lichee u-boot supports recovery and fastboot.. so it could be the cause
<drachensun>
I've pretty much been running on on the charger so I'm actually not sure about the battery during boot
<drachensun>
what kind of touch screen does yours have? does it work?
<drachensun>
techn: did you pull the dram settings from boot1 for your tablet?
<techn>
drachensun: no
<techn>
from fex
<drachensun>
heh only 1 of the 6 devices I have pulled the fex on had the ram settings, pretty lucky
<techn>
drachensun: Last week we got gt811 ts driver .. it works perfectly :)
<drachensun>
I can't get mine to work with multitouch and I can't figure out why
<drachensun>
i'
<drachensun>
I'm starting to wonder if I am missing some user space piece
Sternennebel has joined #arm-netbook
<libv>
"turbo switch", what a perfect name for a set of serial wires sticking out of a tablet case
<hno>
guessing that the issue is caused by 7383d6175e515db1192115d661e351a2736da00a or 6262be6104ccfba448b10295b1c26ba9750851a5
* mnemoc
slaps ibot
rellla has joined #arm-netbook
<hno>
ibot: It''s not your fault. Blame IRC.
<hno>
libv, wireless noisoiness?
hp__ has quit [Read error: Connection reset by peer]
hp__ has joined #arm-netbook
<libv>
hno: from the android kernel driver, yes
<hno>
Ah, you mean kernel spews lots of messages about the wireless controller?
<hno>
libv, thought you had noticeable noise in the serial signal.
<hno>
mnemoc, any success/failure?
<libv>
nope, it's really rather dandy.
<libv>
is an android kernel bootlog, plus uboot going to help anyone?
<mnemoc>
hno: sorry, got distracted. 5m
freakazoid0223 has joined #arm-netbook
<drachensun>
rellla: Updated the wiki with my cross compile notes
slash_random has quit [Ping timeout: 256 seconds]
<mnemoc>
drachensun: haven't looked at your change, but please try to keep not-xbmc-specific stuff out of xbmc's wiki page
<drachensun>
?
<drachensun>
its how to cross compile
<drachensun>
you think the build stuff should be a on a new page?
<mnemoc>
if it can be used to cross compile something else too, yes
<mnemoc>
tutorials have the tendency to duplicate everything
<mnemoc>
and in a short time that makes a bunch of incomplete/inconsistent/outdated info
<mnemoc>
also, if something needs to be fixed, instead of giving instructions, provide a patch, and try to "upstream" it :)
<drachensun>
I got ya
<mnemoc>
:)
<drachensun>
I didn't think the fix stuff was always a problem, I just figured they we the result of the cross compile setup, but yeah they real are generic "how to build" tips
<mnemoc>
hno: 7c3e3191f0bc0fc905511294830611dcf06d5b1b boots fine
slash_random has joined #arm-netbook
<hno>
can you try 7383d6175e515db1192115d661e351a2736da00a and 6262be6104ccfba448b10295b1c26ba9750851a5 as well?
<mnemoc>
hno: ok
<Marex>
hno: hey, any news on the u-boot stuff ?
<hno>
Marex, yes, me and Stefan is busy cleaning things up and breaking it a bit.
<Marex>
mmmm, sounds good :)
<techn>
mnemoc: should we make sunxi-bsp avare of changed kernel config?
<rellla>
drachensun: looks good. like mnemoc mentioned, xbmc-wiki-site perhaps could be refactored ... imho it's not that overwhelmed with xbmc-unrelated stuff. maybe an extra page howto setup a functional (cross)build environment would be useful.
<mnemoc>
hno: that .git is to trigger a `git clone`
<rellla>
drachensun: btw, did you test empat0's latest changes and pvr?
<mnemoc>
techn: not really convinced about that.... we already do an oldconfig, and the user might want to tweak his .config
<techn>
ok.. I'll leave that out
<techn>
mnemoc: pushed
<mnemoc>
techn: btw, the media create script now needs the hwpack with full path because of the `cd` before extracting
<techn>
mnemoc: I patched that already
<techn>
:p
<mnemoc>
:)
Sternennebel1 has joined #arm-netbook
Sternennebel has quit [Ping timeout: 244 seconds]
<hno>
Who is Андрей Панов (apxii)?
<mnemoc>
i remember to have merged a commit from him. not i believe he is not here
gimli has quit [Remote host closed the connection]
<mnemoc>
yw
<mnemoc>
techn: any reason to use 7z? is that the "standard" for hwpacks?
<Turl>
techn: any special reason why you switched to rsync?
<Marex>
Turl: xz and 7z differ
<Marex>
7z is lzma1 and it's some windows crap
<Marex>
xz is lzma2 and it becoming new standard on linux systems (not unix though)
<Turl>
"other unix" still use gzip right?
<Marex>
(yes, xz has better compression ratio than 7z)
<Marex>
Turl: yea, eg. OpenBSD comes only with gzip and compress I think
<Marex>
but you can install bzip2
<mnemoc>
techn: can we switch to xz?
<mnemoc>
do they still use cvs?
<techn>
mnemoc: sure.. if that works
<Marex>
mnemoc: who? OpenBSD still uses cvs ;-)
<Marex>
maybe they already finished their own OpenCVS
<Marex>
(yay)
<mnemoc>
doh
<Turl>
next in line is an svn clone right? :P
<techn>
mnemoc: we need to add support to media-create.sh
<mnemoc>
sure
<hno>
mnemoc, if you are still awake, test current HEAD of sunxi-current.
<hno>
backed out the problematic change.
<mnemoc>
hno: 1m
<Turl>
techn: (19:28:03) Turl: techn: any special reason why you switched to rsync? <- just asking because cp is on any half decent distro, rsync isn't so widely available
<mnemoc>
cp -r doesn't support overlapping well
<Turl>
rm + cp?
<hno>
rsync is available on any half-decent distro, but perhaps not installed by default.
<mnemoc>
tar -c | tar -C ... -x does the same
<Turl>
yeah, it's not much use if it's on a remote server on a package :)
<mnemoc>
is tar installed by default?
<Turl>
yeah
<mnemoc>
then we can use tar | tar
<mnemoc>
too fancy?
<hno>
what is overlapping?
<Marex>
ugh
<Marex>
what are you trying to do ?
<hno>
?
<mnemoc>
a recursive cp to a dir where the wanted subdir already exists
<hno>
that works.
<mnemoc>
to install overlays
<mnemoc>
if `cp -r` does the job, let's use that then
<mnemoc>
i had bad memories about it, but might be something else
<Marex>
cp -Rp you mean ?
<Marex>
or cp -Rfp
<Marex>
(preserve permissions)
<Marex>
ev. force
<mnemoc>
the original script used `cp -rf`
<hno>
the shell often have aliases cp to add -i flag by default which confuses a bit. But not in scripts. -p should be there.
<Marex>
ugh ... alias for cp -i ?
<Marex>
that's so wrong
<mnemoc>
luser protection
<mnemoc>
pretty stupid
<hno>
same often done for rm
rellla has quit [Quit: rellla]
<Marex>
hno: yea ... it's pretty sad :-(
<mnemoc>
i've seen rm aliased to mv :<
<Marex>
hno: rm -rf / was always such a fun
<hno>
Still is
<Marex>
mnemoc: I've seen (done) all commands aliased to rm -rf / :D
<Marex>
hno: nay ... booboontu won't let you anymore
<mnemoc>
:)
<hno>
only needs to be done from a script now.
<mnemoc>
hno: hno/sunxi-current boots :)
<hno>
and the -f overrides -i
<mnemoc>
so `tar | tar` is considered evil, right?
<hno>
unneccesary obfuiscation of what is done.
<mnemoc>
`ssh tar | tar` is pretty handy :p
<mnemoc>
and faster than rsync
<hno>
scp -r?
<mnemoc>
works backward?
<hno>
what is backward?
<mnemoc>
from the server to you
<hno>
yes.
<mnemoc>
nice
<lundman>
rsync
<mnemoc>
rsync checksums every single file....
<lundman>
if you want it to
<lundman>
but yes, lots and lots of small files, tar will be faster
<hno>
unless you already have most of the files.
<lundman>
also true
<mnemoc>
sure
<lundman>
and use --inplace
<mnemoc>
there are uses for each tool :)
<mnemoc>
hno: btw, I only tested on the cubie.... but I guess it's enough in this case
<hno>
it's good enough for tonight at least. Need to runn a full test after that change have been fixed.
<hno>
Marex, you are right. "rm -rf /" is blocked these days, but it's not related to -i, only -r
<hno>
bin/rm: it is dangerous to operate recursively on `/'
<hno>
bin/rm: use --no-preserve-root to override this failsafe
<Marex>
lundman: use netcat if you want to milk the most from the transfer :p
<Marex>
hno: it's STUPID :-E~
<Marex>
bloody
<hno>
How should new sysadmins learn anything now?
<Marex>
exactly :-C
<Marex>
and now that forkbomb is also easily limited ... :(){:|:&};: ... :-C
orly_owl has joined #arm-netbook
slapin_nb has joined #arm-netbook
hp__ has quit [Read error: Connection reset by peer]