<mripard>
hno: I finally took some time to dump the BROM on the A31, do you want me to put it somewhere?
<mnemoc>
dl.sunxi.org/users/mripard ?
<arokux1>
hi all. i'm back from vacation now, will continue my work on usb support
<mnemoc>
mripard: can you test sshing to sunxi.org and writting something in public_html/ ?
<mripard>
it seems to work fine
<mnemoc>
great :)
<mnemoc>
feel free to abuse it
dlan has quit [Ping timeout: 264 seconds]
dlan has joined #linux-sunxi
_enrico_ has joined #linux-sunxi
vicenteH has quit [Ping timeout: 260 seconds]
shaggy has quit [Quit: Page closed]
notmart has joined #linux-sunxi
notmart has joined #linux-sunxi
vicenteH has joined #linux-sunxi
tijs has joined #linux-sunxi
fldc has joined #linux-sunxi
gnufan has joined #linux-sunxi
MrBar has joined #linux-sunxi
<MrBar>
hi sunxi
<MrBar>
what is sunxi?
<mnemoc>
the name of the family of arm-based SoCs made by allwinner
<mnemoc>
sunXi with X between 3 and 7 currently
<MrBar>
but what is A10 A13 and like crap?
<mnemoc>
A10 is a SoC of the sun4i arch, A13 is a sun5i
<MrBar>
just alias?
<arokux1>
MrBar, yes
<mnemoc>
not really, a10s is also sun5i
<mnemoc>
also a12
<mnemoc>
and there are two known sun6i
<mnemoc>
a31 and a31s
<MrBar>
looks like chinese crap
<mnemoc>
they are pretty good SoCs, but you are free to label them as you like
<MrBar>
ok
<MrBar>
who know about axp209?
<mnemoc>
the PMU used with A10 and A13
<MrBar>
yes
<MrBar>
and it work like crap
<MrBar>
ok sorry for bad lexic
<MrBar>
maybe someone know how to
<MrBar>
how to avoid to boot in charge only mode? (androidboot.mode=charger) and how to complete remove android platform from uboot and linux?
<MrBar>
android just work like java crap
<tijs>
I dont think you can pull it apart that easy
<tijs>
you should rather pull out the uboot info, and drivers, and merge it into linux.
<MrBar>
may be there some #undef ANDROID_CRAP ?
<MrBar>
and remake all
<MrBar>
power_on button have random behavior while boot in debian, but android work stable
<tijs>
yeah just remake all, then you have also lots of other options for the linux kernel to use. However, it seems that linux drivers and android drivers are not the same. So if you want linux on it, you will have to rewrite some.
<MrBar>
i think someone use some GPIO to extracontrol AXP209 and change it from android
<MrBar>
android is just malware injected in linux kernel
<mnemoc>
there are defconfigs with android enabled and others with android disabled
<mnemoc>
pick what suits you
<MrBar>
heh nice joke
<Turl>
mnemoc: mali broken :( see ML
<tijs>
GPU's are always fun :P
<mnemoc>
Turl: so benn's update is a20-only?
<Turl>
well, not theorically, but seems so practically :p
<mnemoc>
:(
<hno>
MrBar, remoiving Android is not hard. Or you could just boot whatever you like from an SD card.
<stulluk>
Because I also need to use multiple UART interface
<stulluk>
Is that fixed?
<ssvb>
Turl: what kernel configuration are you using to get this "drivers/gpu/mali/mali/platform/ca7-virtex820-m400-2/mali_platform.c" file actually compiled?
<Turl>
TARGET_PLATFORM is "mali400-pmu" for all sunxi
<ssvb>
Turl: wait a second, do you mean that you need to patch this into stage/sunix-3.4 to get mali working on A20?
<Turl>
dunno
<ssvb>
Turl: what about the reports from the people who claimed that mali already works on A20 with stage/sunxi-3.4?
<Turl>
well then you obviously don't need this :)
<Turl>
just wondered why it wasn't building and fixed it :P
hramrach has quit [Ping timeout: 240 seconds]
<ssvb>
mnemoc, Turl: my guess is that benn has probably just copied the redundant source file from 3.3, but never verified that it is really needed
<Turl>
ssvb: even if dropping this file though, it'll still leave sun45i broken
<Turl>
or maybe not, just saw the unknown symbol was unrelated to the ump and not the phys thing
<ssvb>
Turl: it works on sun4i if you remove the MEM_VALIDATION entry
<Turl>
I guess that's it then :)
<Turl>
I need to get going, see you later guys
<arokux1>
hno, ping
hipboi has quit [Quit: Leaving]
tijs has joined #linux-sunxi
tijs has left #linux-sunxi [#linux-sunxi]
<oliv3r>
mnemoc: have you started integrating sdk 2.0 into our lychee branches
<stulluk>
oliv3r: I figured out uboot tfp problem, because CONFIG_CMD_NET was undefined in u-boot source, so network scripts are not working..
<oliv3r>
gawd today is horrible ar $work
<stulluk>
Is there a way to use community u-boot to be compiled together with this SDK ?
<oliv3r>
working nonstop since 8:30 till now
<mnemoc>
oliv3r: not even extracted it yet. pretty busy with work and life issues :(
<oliv3r>
stulluk: you can't fix the multiple uart 'problem' as there is no problem. YOu as a developer decide which pin does what for you, and then create a fex file appropliatly for it. There is no '1 fex file to rule them all' for your board, because it is a dev board for one
<oliv3r>
stulluk: i dout none of us really use the SDK; we only extract kernel code from it at the most
<oliv3r>
mnemoc: yeah i can only imagine :( no pressure
<oliv3r>
mnemoc: just drawing you back into sunxi :)
<arokux1>
oliv3r, something interesting happened during two last weeks here?
<oliv3r>
mnemoc: i just came to realize today, that with our most recent effort, the sun457i might be the worlds most open 'PC'
<mnemoc>
do you know that in spain "Lumia" means slut?
sud0x3 has joined #linux-sunxi
<mnemoc>
nice marketing research they did
<MrBar>
heh
<leviathanch>
Turl: hi
<jlj>
I just learned a new word in Spanish!
<leviathanch>
Turl: is there DMA support for A20 within your branch already?
<jlj>
lurking irc sure is educational
<arokux1>
hno, is it possible to track down how GL850G usb hub is powered on mele? I think on-chip usb controller is working nice, but there is a failure during talking to the external usb hub.
tinti has joined #linux-sunxi
<MrBar>
no its not possible
<arokux1>
MrBar, do you own mele?
deasy has quit [Quit: Nom d'un quark, c'est Edmonton !]
<MrBar>
no
<MrBar>
but i have axp2009 based board it just pain
<MrBar>
*axp209
stulluk has quit [Ping timeout: 245 seconds]
stulluk has joined #linux-sunxi
<arokux1>
MrBar, what board exactly?
vicenteH has quit [Read error: Operation timed out]
<MrBar>
noname tablet
<MrBar>
A13 based
<MrBar>
there rtl8188E that crap too
<arokux1>
MrBar, I wonder why have you bought this tablet at all.
<MrBar>
heh
<MrBar>
really?
<arokux1>
MrBar, well, you say everything in it is crap..
<MrBar>
i get it witout money
<MrBar>
i wonder why realtek make crap too
<mripard>
because you're not working there I'm sure
hramrach has joined #linux-sunxi
<oliv3r>
mripard: lol
<oliv3r>
mripard: Hi!
<oliv3r>
mripard: if you are so unhappy with your crap; why not go out and buy something that's not crap :)
<oliv3r>
stulluk: does it tell you why it crashes?
<oliv3r>
i've bought so much crap; i doubt there is non-crap out there even :)
<MrBar>
where backtrace?
<stulluk>
I tried almost everything, when I enable UART6 or 7, it tell me like this: Unable to handle kernel NULL pointer dereference at virtual address 000000c4
<stulluk>
Let me give whole log by pastebin, wait
iamfrankenstein has quit [Quit: Nettalk6 - www.ntalk.de]
<mripard>
leviathanch: you should read the wiki before sending mails
<stulluk>
it will take some time, because I am building on Kubuntu1.02LTS, and writing the whole image to board by Livesuit... Unfortunately I couldn't figure out how to write only /env partition
<HeHoPMaJIeH>
stulluk, what is kernel version that you use ?
<stulluk>
3.3
<HeHoPMaJIeH>
try 3.4 i just test them
<mouchon>
stulluk i am on 3.4
iamfrankenstein has quit [Ping timeout: 264 seconds]
igor- has joined #linux-sunxi
<stulluk>
Ok, let me change back to 3.4 and let you know
<oliv3r>
stulluk: script.bin AND script.orig, configure the pins, it can work as it will reset to some default config of course, but a lot of functionality will be missing
<hramrach>
wingrime,jemk can't really play the sintel trailer either with the vdpau plugin. I don't see the screen but I get tons of errors about incompatbile output
<mouchon>
HeHoPMaJIeH: can you tell me wich boot param you use to have console on lcd ?
<oliv3r>
mouchon: not sure if that's possible :)
<oliv3r>
mouchon: console output only works on uart afaik
<HeHoPMaJIeH>
mouchon, you mean debug console or standart fbcon
<mouchon>
HeHoPMaJIeH: simple question, lcd_width/ heigh params are in which unit ?
<HeHoPMaJIeH>
in pixels
<mouchon>
what's the diffrence with lcd_x and lcd_y ?
<oliv3r>
X resolution and Y resolution? :)
<oliv3r>
Tsvetan: i'm just reading the forum and you wrote you have issue with TS right now?
<MrBar>
how to boot linux from nand?
<mouchon>
olivr3r: may be stupid question but does both are not same if you want use all the lcd surface ?
<Tsvetan>
oliv3r: old post?
<oliv3r>
Tsvetan: on the A20 forum :) you said 'there is issue with ts but we are investigating it
<Tsvetan>
this is written last Friday
<oliv3r>
Tsvetan: I don't have an LCD but was talking to HeHoPMaJIeH and I think he has the same issue (not sure since you didn't mention what issue) :)
<oliv3r>
Tsvetan: yeah i'm slow in reading posts :p
<Tsvetan>
today Dimitar solved the TS issue with you as far I understood
<oliv3r>
Tsvetan: ohh himitar and HeHoPMaJIeH are the same :D
<Tsvetan>
hahahah right
<oliv3r>
Tsvetan: but yeah, the TS driver wasn't ported yet which should be good
<oliv3r>
Tsvetan: now with dimitri's work :)
<Tsvetan>
and today UART6 and 7 are also working with correct multiplexing
<oliv3r>
Tsvetan: but i couldn't test any of his code as i don't have that particiular hardware, but i'll review his patch and possibly do a naming unification and get it merged
<Tsvetan>
so basically everything on A20 is working in 3.4
<oliv3r>
Tsvetan: i think the only ussue with uart6 and 7 was fex specific I don't think the driver needed any work (but i haven't checked my email yet :)
<Tsvetan>
Allwinner misleaded us when they said that if Ethernet works the UARTT6 and 7 should be forgotten
<mouchon>
Tsvetan: so a new image soon :-)
<Tsvetan>
they didnt made good explanation why they muxed them on PA and PI then, so today we found what they say is wrong
<stulluk>
mouchon, I checked your sys_config.fex, and you have disabled EMAC
<Tsvetan>
probably again lost in the translation
<oliv3r>
Tsvetan: hehe, yeah often happens; but also when answering these questions I just go by what I know, i never used uart6 and 7 :)
<Tsvetan>
today I reminded them we are waiting hardware encoding demo
<oliv3r>
Tsvetan: awesome
<oliv3r>
Tsvetan: i hope they respond timely
<stulluk>
mouchon: sorry, I confused... Let me try with your FEX file again
rellla2 has joined #linux-sunxi
<Tsvetan>
oliv3r usually they respond quickly when they have it, I think they do not have such demo and now somebody start doing it ... as we exchanged several e-mails and the ppl in CC list grow up to 7-8
<oliv3r>
Tsvetan: Hahaha, yeah i imagine
<oliv3r>
Tsvetan: they probably don't even know why you want it >: )
<stulluk>
mouchon: when I use your FEX file: port_to_index err: port 7, port_num 13, please check sys_config.fex
<stulluk>
and this one: script_init:mmc1_para->sdc_det gpio cfg invalid, please check sys_config.fex!
<stulluk>
So, now I will change back to 3.4 and try again
<Tsvetan>
oliv3r allwinner just benefit from being the first who will have completely open source drivers for everything
<oliv3r>
Tsvetan: and they don't even seem to care half the time :(
<hramrach>
they only ask how much money does it make them. in the short term there is definitely no benefit. in the long term they can probably sell the old chips that actually work due to having actual drivers
<hramrach>
but that remains to be seen if htere is profit to make there and if they are able to make it
<hramrach>
of course, working drivers are cool but the manufactureres usually don't care :/
<mouchon>
HeHoPMaJIeH: with you fex file fbcon seem working, need to adapt it for LCD7'' but something is displayed, thanks a lot.
<HeHoPMaJIeH>
mouchon, just change lcd settings , you can find fex files in olimex's github i think
BluesBoy has quit [Quit: HydraIRC -> http://www.hydrairc.com <- Nine out of ten l33t h4x0rz prefer it]
<mouchon>
<HeHoPMaJIeH>: have one from the r2 image tnks
<HeHoPMaJIeH>
mouchon, i am not sure
<HeHoPMaJIeH>
if it is done in r2 but this line in config may look like this
<HeHoPMaJIeH>
lcd_pwm_pol = 0
<stulluk>
HeHoPMaJIeH: Did you be able to compile new SDK for Android ?
<HeHoPMaJIeH>
stulluk, still not , i have a lot work today
<stulluk>
HeHoPMaJIeH: Let me make clear about one thing: If I use old SDK, UART6 -7 would still be working? Or I should switch to new SDK ?
<HeHoPMaJIeH>
stulluk, i don know you may test it yourself
<stulluk>
I am going really mad... now I tested UART5 ( not 6 or 7), and when I enable them in sys_config.fex file in "lichee/tools/pack/chips/sun7i/configs/android/wing-k70/" folder, just enable UART5, change PINs from PH06-PH07 to PI10-PI11, and kernel die again...
<stulluk>
HeHoPMaJIeH: The SDK2.0 file that you provided two days ago, it has lichee-2.0.tar.bz, and there is OSEPP folder... When I check OSEPP folder, everything is OLD. Only new part is lichee-2.0.tar.bz... IS there a Android.42.tar.bz seperately ?
boycottg00gle has joined #linux-sunxi
<hno>
arokux1, I bought my multimeter 25 years ago. haven't needed to buy another one since.
Soru has quit [Read error: Connection reset by peer]
<atsampson>
I used to think that way, and then CPC had a special offer on multimeters with RS232 output... since then I've found a number of uses for multiple multimeters ;)
<boycottg00gle>
is there a newer fedora image to test?
stulluk has quit [Remote host closed the connection]
xDR1TeK has joined #linux-sunxi
<xDR1TeK>
hi all
<arokux1>
I think I'll get EM 393
<arokux1>
Hama* EM393
<xDR1TeK>
i am changing the default touch screen driver of an android image v4 in init.sun5i.rc but everytime i reboot the change is lost.
<xDR1TeK>
please can someone explain why?
<xDR1TeK>
also i need to cat /dev/ttyS0 or any other one to get the touch feedback
<hno>
robertjw, that seems to be an A10 version with 1GB ram. But it does not say if it's the improved version with AXP209, or the earlier version with fixed DCDC converters (those had stability issues)
<jelly-home>
oh yeah, mine got rather warm, bordering on hot and stopped booting after wroking that way for 24-36 hours
<hramrach>
wingrime: pong
<robertjw>
any suggestions on selecting a device in the lower price range which includes the AXP209 power chip?
<hramrach>
that's cubieboard for me ;-)
<hramrach>
the HDMI sticks are cheaper but you get random insides with same case and name
panda84kde has quit [Quit: Konversation terminated!]
<jelly-home>
indeed
igor- has joined #linux-sunxi
<jelly-home>
also, pins exposed properly are a great boon
* jelly-home
still has to figure out where the hell is UART on Mele M5
Tsvetan2 has joined #linux-sunxi
Tsvetan has quit [Ping timeout: 240 seconds]
damed has joined #linux-sunxi
<damed>
is u-boot in linux-sunxi gitbub usable for cubieboard2 ?
rz2k has quit []
<wingrime>
hramrach: are you tested vdpau?
<wingrime>
mnemoc: mripard hno ping
<wingrime>
I find something interesting
<damed>
found ... works
<damed>
solved
<wingrime>
oliv3r: ping
<oliv3r>
wingrime: pong
<hramrach>
yeah, tested jemk's vdpau module
<oliv3r>
wingrime: what did you find
<hramrach>
it fails
<wingrime>
EXPORT_SYMBOL(sw_hcd_host_tx);
<wingrime>
* Host side RX (IN) using Mentor DMA works as follows:
<oliv3r>
is that a comment in the AW source?
<wingrime>
oliv3r: ^^^^^^^^^^^
<wingrime>
oliv3r: thats from our usb
<hramrach>
mplayer 1.1.1 crashes, maplyer svn fails to play anything
<wingrime>
hramrach: witch hw are you using?
<hramrach>
did you try the sintel trailer?
<wingrime>
hramrach: big bug bunny
<hramrach>
cubieboard
<wingrime>
hramrach: strange
<hramrach>
a10 one
<wingrime>
hramrach: also you change disp and cedar_dev to be 777
<wingrime>
?
<hramrach>
no. I run everything as root
damed has quit [Ping timeout: 250 seconds]
<hramrach>
other cedar players used to work, anyway
<oliv3r>
wingrime: well we know that the USB-OTG port is from mentor something
<hramrach>
I don't really care about 10 bit but every other tard knows how to frob the slider to encode 10bit and they do just because they can. And if you cannot decode that your codec is useless
<hramrach>
wingrime: it just sasy that but it's bogus
<wingrime>
hramrach: sintel not looks like 10b
<hramrach>
it inserts scale automagicall on its own and I don't have spp
<hramrach>
it does not play back anyway
<hramrach>
same error
<wingrime>
hramrach: have you any .mplayer/config
<hramrach>
empty
<wingrime>
hramrach: I will try myself with sintel
<hramrach>
oh, added vf=scale there to try out that awsome advice
<wingrime>
hramrach: and so?
<hramrach>
so what?
<hramrach>
it's just bogus
<wingrime>
hramrach: works?
<hramrach>
it's some generic advice which would work if the decoder worked
rellla has quit [Ping timeout: 260 seconds]
<oliv3r>
hramrach: remember, libsunxi is only testing/PoC code; it isn't to be used for anything else then to test and improve
<hramrach>
well, for PoC it should decode something at least
<hramrach>
of course, it may decode very narrow profile of media atm but then it has to be determined what profile it is so the results are reproducible at least
<oliv3r>
hramrach: did you try buck bunny?
<hramrach>
where do you get the exact buck bunny encode that works?
<hramrach>
that 10bit file does not play with vlc nor xbmc so maybe cedar really cannot do that
shineworld has quit [Remote host closed the connection]
<wingrime>
hramrach: I very doubt that 10b hw decoders exsist
<jukivili>
with host and peripheral modes (but without mode sensing / autodetection).. and appearently atleast peripheral does not work with sun5i
<hramrach>
well then hw decoders are useless
<wingrime>
hramrach: 10b not lead any standard
<hramrach>
any decoder is good only so long it can actually decode random videos
<hramrach>
if you need to recode every other video to get any playback at all
<hramrach>
you can as well live with a hand crafted NEON coded for one very specific format and recode *everything*
<oliv3r>
well then cedarX is simply crap
<oliv3r>
but do we know of any hardware (assisted) decoder that can do 10b?
<hramrach>
suposedly the Exynos decoder is just filter pack
<hramrach>
so basically some vector (or matrix if you wish) numeric coprocessor
<oliv3r>
heh which version? i know the S2 is having some issues decoding 1080p stuff
<oliv3r>
doubt its doing it in software, but might be
<hramrach>
so technically should be able to decode any format you can express in the arithmetic operations it supports
<oliv3r>
so it's hardware assisted
igor- has left #linux-sunxi ["Konversation terminated!"]
<hramrach>
it does not mean the performance is awesome or that it's used by every software or even that the drivers are good enough for it to be usabel at all
<hramrach>
see the cedar drivers which fail to work even for formats technically suported by hardware
<wingrime>
hramrach: you have bad carma
<wingrime>
hramrach: I just get 720p sintel
<oliv3r>
well hardware assist is much better, but problem with HA is that it costs a lot of memcopies
<hramrach>
yeah, bugs love me
<hramrach>
they throng around me :p
<hramrach>
oliv3r: depends on the hardware. if it can output to usable memory regions you do not have to copy more than with software
deasy has joined #linux-sunxi
<wingrime>
oliv3r: I get sintel 720 working
rings_IIV has joined #linux-sunxi
<hramrach>
I just got the big buck bunny. same as sintel
<wingrime>
hramrach: where was problem?
<hramrach>
I have no idea
<hramrach>
it just crashes or fails to decode
soul has joined #linux-sunxi
<hramrach>
well, adding the scale filter makes it fail and not crash so it definitely helps :p
Soru has quit [Ping timeout: 246 seconds]
Super-noob has quit [Ping timeout: 248 seconds]
Super-noob has joined #linux-sunxi
<hramrach>
if you have some idea where to add debug prints or some mplayer/vdpau-sunxi debug patches ..
<wingrime>
hramrach: not sure that bug in our code
<hramrach>
maybe I have some slightly different version of something that is not compatible
<hramrach>
well, mplayer works without that code
<hramrach>
but you cannot test mplayer using vdpau and not using your code so ..
<oliv3r>
swap some binaries around to test :)
<hramrach>
also do you depend on osme recent kernel changes?
<hramrach>
I have Linux A10 3.4.43+ #7 PREEMPT Wed Jul 24 07:47:24 CEST 2013 armv7l GNU/Linux
<hramrach>
I did swap some bianries but I am out of ideas now
<hramrach>
tried 3 different mplayes and upgraded libvdpau from .6.3 to .7
<wingrime>
hramrach: I use debian sid
<hramrach>
since each mplayer comes with its own version of most librraries there is not much to change
<hramrach>
I use debian wheezy
<wingrime>
hramrach: i can be helpfull if there somethig in our code crashes etc
<hramrach>
I can't really tell what crashes
<wingrime>
hramrach: but looks like our code not get control
<hramrach>
when you get the info you get a known tablet ;-)
geecko has quit [Quit: Quitte]
<ykchavan>
hramrach, I built img file for nand before and studied log. But I did not come accross 'sunxi-spl.bin' in that log. Is this file sdcard specific?
wingrime has quit [Ping timeout: 245 seconds]
<hramrach>
yes, that's for SD card
<ykchavan>
:-)
ganbold_ has quit [Remote host closed the connection]
<hramrach>
yeah, the page fault occurs after loading mali
rz2k has quit []
<hramrach>
yes, moving away the mali modules avoids the page fault
<hramrach>
that makes it a mali regression on sun4i I guess
<hramrach>
mnemoc:
uwe_ has joined #linux-sunxi
tinti has quit [Read error: Connection reset by peer]