<bfree>
emdebian is for cross-compiling, looks like you are trying to use it natively on armhf. that page is basically about doing stuff from an amd64 computer
egbert has quit [Disconnected by services]
egbert_ has joined #linux-sunxi
<BJfreeman>
ah
<BJfreeman>
yes an A13
<lunra>
just install gcc, then it will work on ARM
hipboi|cubie has joined #linux-sunxi
_BJFreeman has joined #linux-sunxi
BJfreeman has quit [Ping timeout: 252 seconds]
_BJFreeman is now known as BJfreeman
tinti has joined #linux-sunxi
tinti has quit [Read error: Connection reset by peer]
bhoj has joined #linux-sunxi
[7] has quit [Disconnected by services]
TheSeven has joined #linux-sunxi
shineworld has joined #linux-sunxi
shineworld has quit [Ping timeout: 264 seconds]
rellla has joined #linux-sunxi
BJfreeman has quit [Quit: had a good time]
_BJFreeman has joined #linux-sunxi
_BJFreeman is now known as BJfreeman
rellla2 has joined #linux-sunxi
rellla has quit [Ping timeout: 268 seconds]
shineworld has joined #linux-sunxi
tzafrir has quit [Ping timeout: 264 seconds]
<shineworld>
hipboi|cubie, in your experience with AllWinner powered products and android (now 4.0.4) have you got some positive result using Android Open Accessory Mode (and so connecting android device, like cubie to external device like USB -> UART of FTDI) ?
n01|away has quit [Quit: leaving]
n01 has joined #linux-sunxi
n01 has quit [Client Quit]
n01_ has joined #linux-sunxi
n01_ has quit [Client Quit]
n01 has joined #linux-sunxi
Quarx has joined #linux-sunxi
awafaa is now known as FunkyPenguin
FunkyPenguin has quit [Changing host]
FunkyPenguin has joined #linux-sunxi
awafaa_ is now known as awafaa
san has joined #linux-sunxi
<shineworld>
I'm trying to connect a penmount capacitive touchscreen (PenMount-1201) and seem in penmount.c that there is already a kernel module.
<shineworld>
The module search for a /../input0
<shineworld>
what is input0 and where is connected on cubieboard ?
<shineworld>
till now I've used only sun4i-ts module which is simple to connect but support only resistive films
ijc has joined #linux-sunxi
rellla2 is now known as rellla
paulk-desktop has joined #linux-sunxi
<hramrach_>
ssvb: interesting thing is that when vlc crashes cedar seems to keep feeding garbage to disp and produces tons of kernel messages about something being wrong
<hramrach_>
since xbmc auto-restarts itself it will likely have less of this problem
<n01>
mripard_: I was implementing your solution for wdt. But in this way how can I get the wdt module (sunxi_wdt.ko)? I mean if sunxi_wdt.o is always compiled in with architecture how can we generate the module?
* slapin_n1
hates canned fish :(
shineworld has quit [Read error: Connection reset by peer]
<oliv3r>
Turl: crap, i did a save last friday on the RE page, that appearantly didn't go through (probably got a question before closing for going home) and continuing now brings up the latest saved change :( and I forgot what I allw rote :( i'll try to bring it all back from memory ;)
<n01>
here you have platform code to reboot _and_ wdt driver drivers/watchdog/s3c2410_wdt.c
<mripard_>
so that's your 80% ? come on :)
<mripard_>
and yes, it's a real burden to have to share dt-ids, register definitions, code logic
<mripard_>
etc
<mripard_>
wether you like it or not
<n01>
:D yeah it is not 80% but I haven't been able to find code using for reboot code already defined in wdt driver
<mripard_>
yes, because most of the SoCs don't have a proper wdt driver for it :)
<Turl>
oliv3r: why are you telling this to me? :P
<Turl>
hi mripard_
<n01>
uhm ... ok, I'll spend a few more hours trying to get something not ugly, otherwise I'll going to use only the code in wdt driver
<oliv3r>
Turl: you modified the page since friday, and fixed stuff i had fixed allready :)
<mripard_>
n01: anyway, do whatever you want
<oliv3r>
n01: poor carlo, what have you gotten yourself into :)
<Turl>
mripard_: I reworked the factor clocks a bit in preparation to support some more PLLs
<oliv3r>
n01: to be fair, it's an interesting research factor here :)
<mripard_>
whatever you feel is the cleanest
<n01>
lol :P I'm trying to learn :PPP
<mripard_>
and send it
<Turl>
mripard_: but I set it up to always be behind a composite clock, even when it has only one component
<Turl>
mripard_: would the overhead be acceptable?
<mripard_>
we will discuss the details afterwards
<n01>
no prob, thank you
<Turl>
oliv3r: I deleted an unneeded redirect as far as I recall
<mripard_>
Turl: hi
<oliv3r>
Turl: yeah it was broken, but had fixed it friday afternoon, i just noticed i missed a bunch of my last edit
<oliv3r>
but sometimes, the wiki asks you some spammer question, sometimes times out (when editing for an hour :p )
<mripard_>
Turl: hmmm, do you have a patch already ?
<oliv3r>
so lost all that
<mripard_>
Turl: also, mike has an a13-olinuxino now, so you'll probably not be alone anymore on the clocks side :)
<Turl>
mripard_: yeah, untested still
<Turl>
let me paste it so you can get an idea of what I'm saying
<mripard_>
Turl: yeah, I don't want to test it, just how it looks like :)
<bhoj>
wingrime, is there any such pin I need to look out for ?
<wingrime>
or try short nand flash
<bhoj>
what is short nand flash ?
<wingrime>
you can short some nand flash wire for make a31 make failsafe to fel
<wingrime>
I think thay keep that behavior
<wingrime>
simply say, If nand read fails on boot CPU must go in fel mode
<wingrime>
try short some nand information wires to GND
<wingrime>
bhoj: Trul: how idea?
<bhoj>
Turl, any idea ^^ ?
Yaku321 has joined #linux-sunxi
Yaku321 has quit [Disconnected by services]
Yaku-noob has joined #linux-sunxi
<Turl>
idea on what?
<bhoj>
Turl, how can I put allwinner a31 to fel mode.. Can shorting any pins help ? I tried fel-boot.c on sdcard but it did not work ..
<Turl>
bhoj: read the ML, if I recall correctly tom said that dding zeroes to nanda makes it boot to fel
<bhoj>
Turl, I did go through the ML discussion. The thing is I can't reach to the console. I have flashed a wrong image and hence the kernel boots and fails while mounting partition. I don't get to a console. adb shell doesn't work.
<oliv3r>
libvecore.{a,so} is seriously a hodgepodge of source :S
Tartarus has quit [Excess Flood]
Tartarus has joined #linux-sunxi
<wingrime>
oli3r: are you tryed make some stud library for libvecore
<wingrime>
only printf with parameters
<wingrime>
for real log from android for example
<wingrime>
or wraper
<wingrime>
printf and pass to real lib
<bhoj>
is there an spl support in u-boot ? I tried building u-boot but it did not generate an spl ..
<oliv3r>
wingrime: not yet, far beyond my knowledge :p
<oliv3r>
wingrime: but Its a horrible horrible mess. They used all sorts of random reference implementations they found on the internet
<oliv3r>
and kinda hacked it all together
<wingrime>
olb3r: 2 years ago I made stud library and remove 1Mb useless libs from phone
<oliv3r>
Right now, we'd ideally want to know what the core actually does/supports. someone mentioned a dedicated arm core even for that alone :) which I find somewhat hard to belive, where's the firmware :p
<oliv3r>
wingrime: REing cedar will be very time consuming job imo
<Turl>
bhoj: boot to android recovery?
paulk-desktop has quit [Quit: Ex-Chat]
<wingrime>
olv3r: It possible if someone steal from hex-rays arm decompiller for IDA
<wingrime>
olv3r: look at my lib, you can do same and get some logs according real calls
<wingrime>
next step make wrapper
<oliv3r>
wingrime: im' not smart/knowledgeable enough :p
<wingrime>
big work begin with single small step
<wingrime>
also you can add some kprint on cedar ioctrl
<oliv3r>
wingrime: ohh, a stub, not stud :D I know what you mean now, yeah we should start by having a wrapper around the entire goar filled thing, but on the other hand, don't we know how to call the library? so what will we find? we know the api allready, we'd want to know what happens on the register level; lucky it IS sniffable by modifying the driver/LD_PRELOAD the kernel?
<oliv3r>
wingrime: i think the best way, is to dump everything the library tries to write to the kernel, atleast we can get all register access dumped
<wingrime>
linux have API for dynamic lib loading
<oliv3r>
well fun sun7i it'll be easy, a simple IOCTL; for the sun4i, with the memory mapping, i don't know if it will be 'so' easy :(
<wingrime>
oliv3r: it not a problem "map" when lib have relocations
<wingrime>
olvi3r: there is way to do "map"
<wingrime>
HW breakpoint on address
<bhoj>
Turl, adb reboot recovery doesn't get me anywhere.
<wingrime>
or addresses
<wingrime>
olvi3r: kgdb can help with it
<oliv3r>
wingrime: see, that's a little above my knowledge, sun7i does it better, they have an IOCTL and do readl and writel when you want to read or write a register
<wingrime>
oliv3r: kernel can be debuged using UART with kgdb
<oliv3r>
wingrime: yeah but that's all way out of my league :p
<wingrime>
oliv3r: you can simply set debug breakping to address spaces and wait
vicenteH has quit [Ping timeout: 268 seconds]
<wingrime>
oliv3r: it realy does some COMMAND-WRITES to Cedar using mapped space?
<wingrime>
oliv3r: It must be only for buffer normaly
<shineworld>
hmm perhaps I've mistake everything about penmount touchscreen, so I haven't to use /driver/input/touchscreen/penmount.c but PenMount devices are already supported how HID components by 3.0 with USB
<shineworld>
I guess linux-sunxi 3.0 support HID devices
<oliv3r>
wingrime: i'm pretty sure that most/all decoding operations are handled via that mmap
<wingrime>
olv3r:You send some coded data and recive decoced data in that address space I guess commands must go using ioctrl
<oliv3r>
wingrime: libvecore.{a,so} is filled with register, sram references
<oliv3r>
wingrime: but i still have to setup my cubieboard SD card, and have a bunch of drivers to test/write
<oliv3r>
so for now, i'll only document my findings on the wiki
<oliv3r>
and hope someone smart will pick it up
<wingrime>
oliv3r: You always can make breakpint in userspace
<wingrime>
oliv3r: I realy sure that There is some internal function with command call
<oliv3r>
wingrime: well, it'll probably be easier on A20 hardware (with new blob) and hope A10 works the same way
<oliv3r>
sun7i has a simple IOCTL that does readl/writel
<wingrime>
oliv3r: Also there is some great trick for internal functions wrapper
<oliv3r>
wingrime: i'm sure there's many great hacks seasoned RE-ers know :p
<oliv3r>
wingrime: first more drivers for Mainline, while cedarX being close is shit, we want to wait for the new blob anyway :(
<wingrime>
oliv3r: on LD_PRELOAD 1) enable wirte to codepage 2) replace function start with long jump to our code 2) do our with 3) return old bytes to function 4) jump back to hijaked code
<wingrime>
olv3r: insiders always best way to get docs
bhoj2 has joined #linux-sunxi
<wingrime>
olv3r: why we just ask them for a cedarx docs
<oliv3r>
wingrime: we have some programming documentation
<oliv3r>
but that's -> libvecore.so
<oliv3r>
which isn't a problem
<wingrime>
?
<oliv3r>
but libvecore.so is a messy thing, it's a bunch of reference implementations, hacked together with some glue, that offload bits to the VPU
<oliv3r>
we know how to talk to libvecore.so, we want to know how and what we can offload to the VE
<wingrime>
olvi3r: have it some firmware?
bhoj has quit [Ping timeout: 272 seconds]
<oliv3r>
wingrime: we don't think so
<oliv3r>
but i don't think it's a super smart VPU either, I would be supprised if it does more then offload some very calculation intensive things
<oliv3r>
wingrime: your arm decompiler plan sounds a whole lot better ;)
Undertasker has joined #linux-sunxi
<oliv3r>
wingrime: and if i'm not mistaken, some debuging symbols are still left in the code!
<CountryGeek>
General Q - what's the process / timeline for upgrading the kernel to a more contemporary version? (is there a doc?)
<wingrime>
olvi3r: It must be easy and much simple: it guess : 1) send buffer 2) set data type flags and ask for decode 3) recive IRQ from cedar and get frame
<wingrime>
oliv3r: you can't simply buy IDA or ARM decompiller for hex-rays
<oliv3r>
wingrime: :(
<oliv3r>
CountryGeek: who's dwilkins?
* CountryGeek
is also known as dwilkins
<oliv3r>
CountryGeek: ah, ok, are you using gpio?
<oliv3r>
CountryGeek: sunxi does have a hardware PWM unit (2 actually)
<CountryGeek>
??? - nah - I interfaced with the h/w chip
<CountryGeek>
checked w/oscilloscope and everything
<oliv3r>
CountryGeek: well you name your sysfs entries pwm
<oliv3r>
erm you named your sysfs entries gpio
<wingrime>
oliv3r: we can only wait some one steal it (last time some russians steal IDA form NOD32)
<CountryGeek>
I tried to follow the lead of the sysfs gpio interface
<CountryGeek>
but named my class pwm-sunxi
<oliv3r>
CountryGeek: ah, ok
<CountryGeek>
/sys/class/pwm-sunxi/pwm0.gpio3/...
<oliv3r>
CountryGeek: i haven't gotten that far with mine :p
<oliv3r>
CountryGeek: do you use the PWM freamework at all?
<CountryGeek>
I tested on pcDuino
<CountryGeek>
no
<CountryGeek>
Kinda unwilling to add support for the *old* interface in my module
<oliv3r>
CountryGeek: in that case, keep this one for 3.0/3.4
<oliv3r>
CountryGeek: mine will be hopefully going into 3.10 or 3.11 using the new PWM kernel framework
<CountryGeek>
There's a patch that implemented a gpio style interface with the new framework
* CountryGeek
rummages around a bit
<oliv3r>
the 3.4 series and the 3.11 series won't easly work together with PWM anyway I don't think, the only user of the PWM hardware in 3.0/3.4 is the display engine, and that won't work with 3.11 :)
<CountryGeek>
Ah - that's what I'm trying to figure out - what's required for newer kernels
<CountryGeek>
Is there a branch or repo with 3.11 stuff in it?
<CountryGeek>
just linux-stable or something?
<oliv3r>
CountryGeek: look at linux-sunxi.org/Mainlining_Effort
<CountryGeek>
oliv3r: many thanks - heading there....
<oliv3r>
CountryGeek: so your going to rewrite/update your driver to match 3.11 and the PWM framework? or leave it as is?
<CountryGeek>
oliv3r: I'm willing to maintain a pwm driver
<CountryGeek>
Yeah - I'll update when I can get a kernel for 3.11 going
<oliv3r>
i'm aiming to get my security ID driver out tomorrow (got interrupted a lot of times) which you can compare against too, also carlos WDT driver is pretty clean allready
<CountryGeek>
oliv3r: The one I wrote is clumsy and oaf-ish I'm sure - I want to make sure it's kernel-quality
<CountryGeek>
oliv3r: Basically 3-4 days of work. I'll move on to an A<=>D driver next if the 3.11 kernel isn't at a good spot yet
* CountryGeek
needs pwm and A<=>D for his girl robot
<oliv3r>
lol
<oliv3r>
CountryGeek: if you want something usable now, stick with hat you have against 3.0/3.4
<oliv3r>
3.11 is far far from usable atm, it boots, but no i2c, no spi, no dma etc yet
<oliv3r>
no graphics, no usb (shall i continue? :)
<CountryGeek>
oliv3r: k - I'll be ready with a branch when 3.11 gets stable. Hopefully with A<=>D as well. I still need to make the code more friendly to compile while it's outside of the kernel
<rz2k>
x11 version cant fit in my 7" olinuxino lcd