ChanServ changed the topic of #linux-sunxi to: Allwinner/sunxi development discussion - Don't ask to ask. Just ask! - See | | Logs at
<oliv3r> CountryGeek: pong
<oliv3r> hno: pwm is (should no longer) be all over th ekernel, we have a pwm framework and section in 3.7? or so kernels :)
ssvb has quit [*.net *.split]
awafaa_ has quit [*.net *.split]
Dreadlish has quit [*.net *.split]
mdfe has quit [Remote host closed the connection]
<hno> oliv3r, yes, noticed earlier.
<oliv3r> hno: i'm really slow with backreading :p
<oliv3r> and respond before backreading everything :(
<oliv3r> and now i go to bed ;p
Kadosch_ has quit [Remote host closed the connection]
WarheadsSE has joined #linux-sunxi
hipboi|cubie has quit [Ping timeout: 256 seconds]
cajg has quit [Ping timeout: 264 seconds]
zumbi has joined #linux-sunxi
zumbi is now known as Guest37274
FergusL_ has joined #linux-sunxi
awafaa_ has joined #linux-sunxi
Dreadlish has joined #linux-sunxi
ssvb has joined #linux-sunxi
tkoskine has joined #linux-sunxi
hno` has joined #linux-sunxi
_BJFreeman has joined #linux-sunxi
lunra has joined #linux-sunxi
hno has quit [Ping timeout: 245 seconds]
FergusL has quit [Ping timeout: 245 seconds]
Guest40180 has quit [Ping timeout: 245 seconds]
rm has quit [Ping timeout: 245 seconds]
tkoskine_ has quit [Ping timeout: 245 seconds]
lunra_ has quit [Ping timeout: 245 seconds]
_BJFreeman is now known as BJFreeman_
BJFreeman_ is now known as BJFreeman
rm has joined #linux-sunxi
CountryGeek has joined #linux-sunxi
cajg has joined #linux-sunxi
egbert has quit [Disconnected by services]
egbert has joined #linux-sunxi
hno` is now known as hno
hno has quit [Changing host]
hno has joined #linux-sunxi
ChanServ has quit [*.net *.split]
bfree has quit [*.net *.split]
drachensun__ has joined #linux-sunxi
drachensun_ has quit [*.net *.split]
TheSeven has quit [Ping timeout: 252 seconds]
bfree has joined #linux-sunxi
ChanServ has joined #linux-sunxi
TheSeven has joined #linux-sunxi
BJFreeman has quit [Quit: had a good time]
mturquette has joined #linux-sunxi
rz2k has joined #linux-sunxi
rellla has joined #linux-sunxi
Quarx has joined #linux-sunxi
eebrah|away is now known as eebrah
mdfe has joined #linux-sunxi
mdfe has quit [Remote host closed the connection]
mdfe has joined #linux-sunxi
mdfe_ has joined #linux-sunxi
mdfe has quit [Ping timeout: 246 seconds]
<oliv3r> mripard: Turl: could you do a really quick and dirty write up how to get from building binaries (initramfs, kernel image, (u-boot, but that's still the same I assume, unless you put the dtb there), dtb to a bootable SD card?
mdfe_ has quit [Read error: Connection reset by peer]
<oliv3r> CountryGeek: ping
<oliv3r> CountryGeek: i may of course be fully off, but a write mask of 0xffff0000 masks 16 bit does it not?
<oliv3r> if it would mask 8 bits to be written only, wouldn't it be 0xffffffff00
<oliv3r> my bitops mojo is weak however
<mripard> oliv3r: not today
<mripard> replace their u-boot by hno's one, and that applies.
lkcl_ is now known as lkcl
Kadosch has joined #linux-sunxi
sanka has joined #linux-sunxi
paulk-desktop has joined #linux-sunxi
<hno> oliv3r, Yes, 0xff or ~0xff depending on what way you want to mask.
<mnemoc> negation is much safer than counting fs...
<mnemoc> 'f's
<mnemoc> or i'm I the only who gets dizzy trying to count the 'f's of 0xffffffff00 ?
rellla has quit [Quit: Nettalk6 -]
<mnemoc> oliv3r: and indeed you have a pair of 'f's too much
fra79Wii has joined #linux-sunxi
<fra79Wii> I've seen a cunieboard with an A20 bga instead of an A10... So it could be done... In theory
<fra79Wii> Som of the pin are not perfectly the same..I'm concern about the RAM pins...
<mnemoc> fra79Wii: in olimex's blog you can see a detailed post about the diffs between A20 and A10
<mnemoc> iirc one specially important is that it works at 3.0V instead of 3.3V
Quarx has quit []
KamiKaze_Phoenix has joined #linux-sunxi
KamiKaze_Phoenix has quit [Client Quit]
<oliv3r> mripard: ok, i'll see if i can figure it out with those instructions ;)
<oliv3r> mnemoc: well CountryGeek said that 'it's only 8 bits, because they mask 8 bits. but they mask 16 bits, the docs say 16 bits, but his tests show it may only be 8 bits. So some more testing is needed to verify what it really is
<oliv3r> fra79Wii: they changed a view minor things around, GPS has been replaced with twi4; some hDMI pins have been changed I belive
<fra79Wii> It's howere a bit risky... I was planning to replace the bga on my tablet...
<fra79Wii> But it seems a pretty nifty job to do.... I'm not sure it's as easy as replacing a standard laptop bga.
<oliv3r> well first you'd have to desolder your SoC
<oliv3r> clean up the 'socket'
<oliv3r> new CPU will be pre-balled luckly
<oliv3r> but then allign it correctly and hot air melt it
<oliv3r> it IS possible
<oliv3r> if your really good at all that
<oliv3r> and appearantly change your voltage from 3.3 to 3.0 for some parts, though I think that should be done using the AXP
<fra79Wii> :-)
<oliv3r> it won't be easy, but if you manage to do it, yeah that'll deffinatly be interesting :)
<oliv3r> on the olimex blog you can also see what pins are of importantance to you, if they did actualyl change
<oliv3r> personally, i'd buy a new tablet :)
<oliv3r> as simply getting the chips will be hard enough
<oliv3r> time wasted in getting it done, will far exceed the 100 USD you can get a cheap tablet for
<fra79Wii> I saw the post on olimex. Yes i would wait that A20 would be out..and at least u-boot and kernel fully functional
<oliv3r> appearantly it uses a different interrupt controller
<oliv3r> it's pin compatible, but not 100% software compatible
<oliv3r> but it'll be close enough that support should be easy enough once we have hardware
<fra79Wii> Every day I have less motivation on do the hack...
<oliv3r> fra79Wii: it's very interesting, i just think the ROI is low :)
<oliv3r> like my tablet for example, it was about 100 USD (99 euro's locally, so cheaper when from china direct). It has crappy panel (not IPS), only 512 MiB ram
<oliv3r> If i'd get a 100 USD tabelt now, I'd get 1 G of ram, A20 and hopefully a better panel for roughly the same price
<fra79Wii> Yes..i have ips 1gb.. iI just wonted to recycle all the good HW...
<fra79Wii> But i might buy a crap tablet with A20 and swap the motherboard...
<oliv3r> lol it'll have to fit :(
<oliv3r> but yeah if you have a good tablet, that doens't use GPS, swapping out the CPU is an option
<oliv3r> very interesting hack at the very least
<fra79Wii> Yep.. :-)
<fra79Wii> At least knowing if it worth or not...
<fra79Wii> Maybe some service on ebay doing GPU reballing would do the job better...
<fra79Wii> Surely..
paulk_ has joined #linux-sunxi
<oliv3r> at a cost :S
paulk-desktop has quit [Ping timeout: 256 seconds]
<fra79Wii> Yes :-) ..but at least it wold not rack my PCB by myself.. Below the 70$ total would be worth the effort. Otherwise ...
paulk-desktop has joined #linux-sunxi
paulk_ has quit [Quit: Ex-Chat]
<mnemoc> fra79Wii: bdd, olimex and cubietech are working on A20 boards you can use as "core" of your own thing
de has joined #linux-sunxi
<mnemoc> don't know if luke plans to migrate his eoma68-a10 to a20
<lunra> whoever it was that kindly ported the 8188eu driver (was it you mnemoc? :) ) -- it seems to work! Connects to the access point and I can SSH to the cubieboard I tested. Sorry that it took so long for me to get to test. I'm...kind of new to raising 'issues' - what's the next step?
KamiKaze_Phoenix has joined #linux-sunxi
<KamiKaze_Phoenix> hello
<mnemoc> it was totally blind, and there is still a function with wrong prototype
<KamiKaze_Phoenix> i use 3.4.43 sunxi linux kernel, how i can have header files ?? for compile new module..
<KamiKaze_Phoenix> i have copy uImage and modules 3.4.43+
<KamiKaze_Phoenix> i cross compil
<KamiKaze_Phoenix> ok link in modules folder go to cross compile disk....
<oliv3r> mnemoc: i wouldn't be suprised that luke would replace the a10 to a20 asap, since it is nearly pin compatible the effort is minimal
<oliv3r> anybody know if I can force hansg's fedora to stop rebooting? He says in his readme 'monitor must be on' etc, but I think the detection circuit puts it into standby really fast. I boot nandroid, power it off, put in SD card while it powers off and quickly turn it on again, appearantly that's not fast enough
<oliv3r> any tips on getting the cubieboard to boot with my hdmi monitor and actually 'stay up'? or atleast give some display out?
<de> hi, is there any way to test i2c interface used to connect touchscreen on a10 tablets? my tablet's touchscreen suddenly stopped working, i suspect erased firmware(tho i didnt flash any rom's, but this tablet has some issues, touchscreen sometimes doesnt work when tablet is charging). I have recompiled ft5x_ts module with debug and forced update, skipping ID check, but im getting write and possibly read errors all the time
<de> 358
<oliv3r> de: touchscreen not working when charging is a known problem, just forgot where I read about it
<de> yeah, but mine stopped working forever after one of those charging cycles..
fra79Wii has quit [Ping timeout: 245 seconds]
<oliv3r> ft5x_ts is just a simple microcontroller with firmware, it could be very well that the firmware is corrupted. I don't know of anybody extracting/writing firmware, let alone trying to fix it
<de> most of the cases i could find over the internets were with problems of ID read on step 3... but in my case it seems that ts is rather dead?
<oliv3r> always possible
<oliv3r> charger actually may have killed the chip too
<oliv3r> many things possible :) I think it was related either to refresh rate, or bad design of the hardware in itself
<oliv3r> i really don't rember, but I have read about it
<de> though... i have one of the stock ROMs for similar tablets, with a bit different ft5x_ts module, which shows during init, that it has really found touchscreen, and if i disconnect TS, it says to TS found... so i assume it is able to communicate over i2c with it..
<de> unfortunately that module was compiled without upgrade support :/
<de> but what im wondering, is there any way to determine physical corruption of TS?
<oliv3r> if the stock firmware detects the touschscreen, but still doesn't work
<oliv3r> chances are, your touchscreen simply is broken
<oliv3r> either some cables loose
<oliv3r> some components fried
<oliv3r> not really
tinti has joined #linux-sunxi
<oliv3r> we can't extract firmware (we dont' know afaik or don't have the tools (written))
<oliv3r> i do think we may have datasheet, but it still may not list a way to extract firmware
<oliv3r> and even then,w e don't have stock firmware to compare against
<de> but what about i2c interface itself? can it be tested?
<oliv3r> in general, of course, there's i2c tools, but what are you gonna test it against?
<oliv3r> you'd need a known good i2c device that connects on the i2c bus
<oliv3r> then, simply use the tools to probe it
<oliv3r> you can probe the i2c bus now, but probing the broken chip won't help you much :)
<oliv3r> well you can send/receive commands from the touchscreen to see what it does
tinti has quit [Read error: Connection reset by peer]
<oliv3r> personally, I don't think your firmware is corrupt
<oliv3r> i'd rather think, something blew up or some component isn't accurate anymore
<de> if the reason for not working TS during charge would be known, then could check components involved in that
<de> cause if its hardware problem, then, i think, it must be connected with the charger issue
<de> since tablet was left charging and next morning TS wasnt working anymore
<de> i could order new TS, but i agree that this can be hardware problem and not TS problem, but rather host one... so new TS will be useless :/
<oliv3r> of course :)
<oliv3r> if you are desperate to fix it :)
<oliv3r> you could get something like the bus-priate
<oliv3r> sounds quite plausible a component broke under stress
<oliv3r> bus-pirate*
<oliv3r> connect it to the touchscreen (I2C pins +power) and send commands
<oliv3r> you'd have to manually send all the init commands
<oliv3r> then start receiving commands and see if it responds to touch
<oliv3r> not impossible, probably an hour or two's worth of work, figuring out all the ts init commands (see source)
<oliv3r> in the end, just to test, you'd need atleast some device that does something with i2c
<oliv3r> be it a temperature sensor or anything
<oliv3r> connect it INSTEAD of the touchscreen, send i2c commands to probe the sensor, if it works, the bus is working right
<de> i have a host device (raspberry pi)
<oliv3r> i think the pi does proper i2c does it not?
<de> it does
<KamiKaze_Phoenix> use you TS in linux or android ?
<KamiKaze_Phoenix> ho linux, here....
<oliv3r> connect the TS to that :) you could in theory even port the driver over, remove the sunxi-ism, replace it with pi-ism
<de> android
<oliv3r> KamiKaze_Phoenix: i think a lot use regular linux here
<KamiKaze_Phoenix> search for gpm ts config, on X it work, on olimex A13
<oliv3r> de: make sure to double check the voltage of your TS
<oliv3r> KamiKaze_Phoenix: if the TS is broken/blown up; then that won't help much :)
<de> KamiKaze_Phoenix, so you say there is a tool for linux?
<KamiKaze_Phoenix> ts_xx ts_test ts_calibrate tslib
<KamiKaze_Phoenix> tslib
<KamiKaze_Phoenix> i have a TS 7" on debian on my olinuxino A13 wifi
fra79Wii has joined #linux-sunxi
<KamiKaze_Phoenix> work fine with ts_test etc and with X
<KamiKaze_Phoenix> but not with gpm.... console mode
<oliv3r> that'll only see if android is somehow broken
<KamiKaze_Phoenix> is a little gpm.conf problem....
<oliv3r> flashing stock rom excludes that
<KamiKaze_Phoenix> de: can you use linux distro on sd card on you re tablet ?
<de> thing is, that im not sure which of stock ROM's worked with my tablet. you know, those chinese clones of clones..
<de> KamiKaze_Phoenix, yes, i can
<KamiKaze_Phoenix> find a distro with TS support for test
<de> ive tried, TS doesnt work, but i wasnt aware of tslib-bin package then..
<KamiKaze_Phoenix> exist other soft for TS support in linux, not only tslib ?
<rz2k> xorg-input-mtouch
<rz2k> or something like that
<KamiKaze_Phoenix> now i try configure easycap usb for cam
<KamiKaze_Phoenix> linux header is the whole linux-sunxi folder, ok ?
<de> it is
<oliv3r> well linux-sunxi/include :p
<oliv3r> the rest is overkill
<mripard> oliv3r: linux/include/uapi actually.
<KamiKaze_Phoenix> ok not all, yes large tar.xz whole folder
<KamiKaze_Phoenix> strange now, when i reboot my arm, it s go down, but not run again.... must send a reset button
<oliv3r> mripard: what about linux/include/linux?
<oliv3r> or is that internal stuff? (uapi - user api? does make sense ;)
<oliv3r> mripard: does our sunxi uboot build device-tree support by default? I don't see anything device-tree specific in building linaro u-boot (in the cmd line)
<mripard> it depends on your board configuration iirc
<mripard> so you'll have to make sure that the configuration options are in your configuration
<mripard> and no, include/linux is mostly for internal stuff
<mripard> like, kernel frameworks, etc.
<oliv3r> yeah
<mripard> the userspace don't care about that.
<oliv3r> true true
<mripard> so it's not exported
<mripard> actually, uapi is quite recent
<oliv3r> makes perfect sense then
<mripard> like one or two versions old
<oliv3r> oh really
<oliv3r> didn't know that
<mripard> before, it was kind of merged with the headers
n01_ has joined #linux-sunxi
<mripard> and you had to make kernel_headers to export only the meaningful ones
<oliv3r> ah, never had to do that, so didn't know
<oliv3r> so if I have a sd card, with uBoot, uImage, dtb and rootfs.cpio on it, I should be able to pass allt hose from the uBoot console aswell, right?
<KamiKaze_Phoenix> make ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf- kernel_headers is a right cmd ?
<oliv3r> KamiKaze_Phoenix: appearantly, it used to be :)
<KamiKaze_Phoenix> after create link build to this folder ?
<KamiKaze_Phoenix> in modules folder
<mripard> oliv3r: yes
<oliv3r> mripard: i'll backread the log, i think Turl posted commands. i'm in u-boot shell so shouldn't be to hard
<n01_> re
<Turl> morning!
<hno> oliv3r, the 3.3V supply is usually done with a separate DCDC. The AXP do supply the core voltage however (around 1.2V) and a couple of other voltages.
<Turl> 70eba42..f722406 master -> torvalds/master
<Turl> * [new tag] v3.10-rc1 -> v3.10-rc1
<Turl> :)
<hno> oliv3r, all A10 tablets with GPS I have seen is using a separate GPS, not the built in one. Presumably because the discrete ones works very well and is easier to design with than the half-split one you need for the A10 (plus it's unmaintained driver)
<mripard> oliv3r: they are at the end of the linaro page
<CountryGeek> oliv3r: mnemoc the Allwinner docs say 8bit - only the register docs say 16 bit
<oliv3r> hno: well for a20, they've abandoned the gps pins, not even a mux left, it all converted straight to i2c
<oliv3r> from what i saw in the olimex files
<CountryGeek> oliv3r mnemoc - page 83 & 84
oliv3r has quit [Ping timeout: 246 seconds]
oliv3r has joined #linux-sunxi
mdfe has joined #linux-sunxi
shineworld has joined #linux-sunxi
shineworld has quit [Client Quit]
Undertasker has joined #linux-sunxi
fra79Wii has quit [Quit: AndroIRC - Android IRC Client ( )]
_BJFreeman has joined #linux-sunxi
_BJFreeman is now known as BJFreeman
ganbold__ has joined #linux-sunxi
ganbold___ has quit [Ping timeout: 246 seconds]
FergusL_ is now known as FergusL
<hno> oliv3r, the GPS have been gone since sun5i.
<hno> CountryGeek, what register are we talking about?
<hno> CountryGeek, that url do not open for me.
<hno> PWM period register?
<oliv3r> hno: well sun5i is a stripped sun4i :p anyhow, the same pins are now twi :)
<CountryGeek> hno yeah the period registers for ch0 & ch1
<oliv3r> hno: url opens fine, it'st he A10 user manual
heffer_ has joined #linux-sunxi
heffer_ has joined #linux-sunxi
<oliv3r> but opens on page TV Encoder :p
<CountryGeek> Google search for "A10 User Manual - v1.20 (" with quotes
<oliv3r> CountryGeek: that documents is from
<oliv3r> where do you think it came from :p
<CountryGeek> ha
<CountryGeek> Yeah - but not the same as on the wiki
<oliv3r> well i wrote the wiki page :)
<oliv3r> based on that document
<oliv3r> yeah, i wrote that
<CountryGeek> I saw "Oliver" as the author - thought it might be you
<oliv3r> :)
<CountryGeek> hard to cut / paste out of a pdf
<oliv3r> anyway, it's not PUREly based on the doc
<oliv3r> it's also based on the actual source
<oliv3r> source > doc
<CountryGeek> oliv3r: you have access to more than the docs? awesome
<CountryGeek> I have a constant (MAX_PWM_CYCLES) and more than 0x0ff will cause unpredictable results. I'll run my tests again though
spicerack has joined #linux-sunxi
hno has quit [Excess Flood]
hno has joined #linux-sunxi
de has quit [*.net *.split]
rz2k has quit [*.net *.split]
MadSpark has quit [*.net *.split]
heffer has quit [*.net *.split]
mnemoc has quit [*.net *.split]
<hno> Odd.. kicked for excessive flood while doing nothing.
zumbi has joined #linux-sunxi
jukivil1 has joined #linux-sunxi
oliv3r_ has joined #linux-sunxi
grevaill1t has joined #linux-sunxi
KamiKaze1Phoenix has joined #linux-sunxi
zumbi is now known as Guest64795
<oliv3r_> so your scope would be your backlight :D
<oliv3r_> hno major netsplit going on, maybe that's why
mnemoc has joined #linux-sunxi
oliv3r has quit [Write error: Broken pipe]
oliv3r_ is now known as oliv3r
MadSpark has joined #linux-sunxi
grevaillot has quit [Write error: Broken pipe]
vinifm has joined #linux-sunxi
<CountryGeek> oliv3r: testing now (with oscilloscope) - not seeing any strangeness yet - more cycles would be nice
n01_ has quit [Remote host closed the connection]
Guest37274 has quit [Remote host closed the connection]
<oliv3r> you have a cubieboard right?
* CountryGeek fixed a heinous bug yesterday
<CountryGeek> oliv3r: No - just the pcDuino
KamiKaze_Phoenix has quit [Remote host closed the connection]
jukivili has quit [Remote host closed the connection]
<oliv3r> ah no pcduino
<oliv3r> well it's 'the same'
<CountryGeek> Alas, no SATA connector
<CountryGeek> :-(
<oliv3r> well similar enough :p
<CountryGeek> IMHO, false advertising - I expected the "Arduino compatible" statement to mean 5v GPIO - it's 3v
n01_ has joined #linux-sunxi
<CountryGeek> AFAICT "Arduino Compatibility" means they copied all of the non-standard pin spacing from the Arduino and spent a lot of time working on an Arduino library
* CountryGeek <3's testing PWM with shell scripts
h36sa_ has joined #linux-sunxi
<oliv3r> neat little board
<oliv3r> did it come pre-installed with android?
<oliv3r> ah, there are android images
<CountryGeek> No - a nasty Lubuntu or something
<oliv3r> ah well ubuntu _should_ haev brightness control otherwise android should too
<CountryGeek> Large pink overscan, no modules (except for the pwm/adc) and no source for pwm/adc
<CountryGeek> Later version is much better - first one was 3.0 kernel, now it's 3.4.??
<oliv3r> ouch :p
<oliv3r> they use linux-sunxi kernel?
<CountryGeek> I think so
<oliv3r> don't recall any of the pcduino folks ever being in here or on ML
* CountryGeek isn't surprised
<CountryGeek> AFAICT it's a Strawberry something from Linksprite
<CountryGeek> Probably some Linksprite folks in here?
<oliv3r> i do appreciate the fact that they've build a A10->arduino wrapper
<CountryGeek> I got kind of angry on Friday -
hno has quit [Excess Flood]
hno has joined #linux-sunxi
<hno> I think my local network is going crazy.
<CountryGeek> oliv3r: The hardware isn't very useful if you're accustomed to Arduino.
<oliv3r> oh GPL violation
<oliv3r> always interesting
<oliv3r> hno: uh oh you spammer!
<CountryGeek> I guess they're keeping the pwm & adc source secret so that no-one can use their arduino wrapper
<CountryGeek> I'm gonna fork that next and make one that works with my driver
<oliv3r> CountryGeek: well they are obliged by the GPL to release (or point to) the full source
<oliv3r> i'm pretty sure that they are using sunxi linux code
<CountryGeek> As I stated in my post and my E-Mails to them and Linksprite
<oliv3r> then again, we don't have pwm.ko :)
<CountryGeek> So - maybe soon we can make the arduino environment work for Cubie/Mars/etc
<CountryGeek> Does Cubie have 2.54mm pin pitch headers or 2.0mm? I think that Mars has 2.0mm pitch:-(
<KamiKaze1Phoenix> is kernel 3.4 or 3.0 have probleme with RTL8188CUS, seem strange working
<KamiKaze1Phoenix> run, stop, run, stop etc....
<CountryGeek> Also - I've only written a driver for the hardware pwm - do you think it's useful to have an emulated software pwm for arbitrary gpio pins?
<KamiKaze1Phoenix> syslog say connected, disconnected.... is a bad kernel compil ?
<KamiKaze1Phoenix> drivers problem..
<vinifm> gpio-sunxi.c is for kernelspace too?
<vinifm> userspace and kernelspace
de has joined #linux-sunxi
<KamiKaze1Phoenix> vinifm: you use a rt kernel ?
de has quit [Ping timeout: 246 seconds]
deasy has joined #linux-sunxi
vinifm has quit [Ping timeout: 256 seconds]
rz2k has joined #linux-sunxi
sanka has quit [Quit: Leaving]
spicerack has quit [Quit: Page closed]
vinifm has joined #linux-sunxi
ganbold__ has quit [Remote host closed the connection]
<vinifm> KamiKaze1Phoenix, i use sunxi-3.0
Kadosch has quit [Read error: Connection reset by peer]
mturquette has quit [Changing host]
mturquette has joined #linux-sunxi
<KamiKaze1Phoenix> i search for solution for easycap on linux 3.4.43 sunxi kernel
<KamiKaze1Phoenix> olimex soft is only for 3.0 kernel i think.....
rz2k has quit []
rz2k has joined #linux-sunxi
rellla has joined #linux-sunxi
mdfe has quit [Disconnected by services]
mdfe has joined #linux-sunxi
<Dreadlish> eh
<Dreadlish> i dont know what was happened
<Dreadlish> but i tried to compile kernel and got this:
deasy has quit [Ping timeout: 264 seconds]
<oliv3r> CountryGeek: a software pwm driver could be interesting, but doesn't the kernel have that allready in the PWM framework? to use 'any' gpio? i would imagine something like that exists allready, but hey, who knows :)
deasy has joined #linux-sunxi
<KamiKaze1Phoenix> well i not found the good way for kernel headers....
<KamiKaze1Phoenix> with 3.4.43 linux sunxi
vinifm has quit [Quit: Saindo]
<Dreadlish> ok, fixed
<Dreadlish> just made a typo ;d
<oliv3r> KamiKaze1Phoenix: cp -ar linux-version/include/uapi/ /wherever
<oliv3r> ?
<KamiKaze1Phoenix> oliv3r: ok i try this
<KamiKaze1Phoenix> no i havent uapi folder....
<KamiKaze1Phoenix> but kernel work fine
<KamiKaze1Phoenix> i use this for kernel make ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf- uImage
rz2k has quit [Read error: Connection reset by peer]
koliqi has joined #linux-sunxi
koliqi has quit [Client Quit]
koliqi has joined #linux-sunxi
<oliv3r> Turl: ping
<Turl> oliv3r: pong
bsdfox\ has joined #linux-sunxi
<oliv3r> turl! hey :D
<oliv3r> can you spare 3 minutes to give me a hand?
bsdfox\ has quit [Quit: Leaving]
bsdfox has joined #linux-sunxi
bsdfox has joined #linux-sunxi
bsdfox has quit [Changing host]
<oliv3r> i've copied the uImage, dtb and cpio to my sd card, and booted u-boot from there
<oliv3r> i've intterupted the boot so got u-boot prompt
<oliv3r> how do i load these files and execute the kernel +cpio?
<Turl> oliv3r: I'd suggest you embed the cpio inside the kernel
<Turl> look for CONFIG_INITRAMFS_SOURCE on menuconfig
<oliv3r> yeah, but didn't wanna build an entire new kernel
<oliv3r> but i guess i have to :p
<oliv3r> just wanted to test something quickly :)
<Turl> then load them to memory normally and bootm 0xkernel - 0xdt where 0xkernel and dt are the addresses
<Turl> well you can load the kernel and dt without rootfs and watch it crash :)
<oliv3r> yeah, that's the bit i need help with :p
<oliv3r> never done anything with u-boot console
<Turl> look at the default env
<oliv3r> fatload?
<Turl> yeah
<oliv3r> fatload uImage 0x480000
<Turl> something like that I guess
<Turl> I think you need to tell it what partition too
<oliv3r> and device
<oliv3r> not sure how to passt hat along
<oliv3r> hence, my plead for help :p
<oliv3r> *reading*
<Turl> do the extloads or fatloads and then "bootm kerneladdr - dtaddr"
<oliv3r> so you pass the dtaddr as argument?
<Turl> yes, bootm with 3 arguments
<Turl> kernel, dash (meaning not used), dt
<oliv3r> dash by any chance stand for cpio location? :D
<Turl> supposedly it's ramdisk, but I never got it going :P
<Turl> I just embed it nowadays
<oliv3r> i'll rebuild my kernel then :p
<oliv3r> bah, per recommendation from the cubie wiki; i got one of those psp adapters to power the cubie
<oliv3r> i have a 500gb 2'5" disk i wanted to use with it
<oliv3r> sure enough, 2.5Amps isn't enough! (it doesn't deliver, is hould open it up and see what power converter they've used)
<Turl> make sure it's legit from sony and not a chinese clone :)
<oliv3r> yeah, i to bef air, this one was 2.12 USD
heffer_ is now known as heffer
<KamiKaze1Phoenix> bye
KamiKaze1Phoenix has left #linux-sunxi [#linux-sunxi]
<oliv3r> Turl: one more thing, I build as user, but the initramfs is root owned (the files in it0
<oliv3r> can I chown the entire thing to 1000:0 and will the kernel fix it? I would assume so
cajg has quit [Ping timeout: 246 seconds]
<Turl> oliv3r: the initramfs is a cpio file, I didn't understand your question
<Turl> if you unpack it, change stuff and repack, make sure you keep the permissions the same
vakkov has joined #linux-sunxi
<Turl> I noticed many people add stuff like "~~~" after sentences on emails, what does it stand for?
* Turl is curious :)
rellla has quit [Remote host closed the connection]
<rm> like "..." but more cutesy
<oliv3r> Turl: oh , since i used hansg's fedora18 sd card as base, some residual files (uenv etc) load the kernel automatically, of course, no console output etc :S
<oliv3r> gah, so much to setup to simply test some kernel stuff :p
<oliv3r> Turl: no clue what ~~~~ would stand for :S
<oliv3r> Turl: yeah I unpack it, it's quicker to modify and if you use initramfs_path (orwhatever it is called) it does the CPIO-ing for oyu. i think it even changes all permissions to 0
cajg has joined #linux-sunxi
<Turl> oliv3r: if you want to just test it, get a working image, then rebuild the kernel with ramdisk inside and with appended dtb and replace it on the image :P
<oliv3r> how do I append the dtb :)
<oliv3r> well 3.9+ kernels aren't usable yet are they
<oliv3r> so a 'all in kernel' would probably be best
<oliv3r> i take it you do your testing via uart1?
<oliv3r> uart0*
<hno> techn_, looking into your wip/usb branch I see removal of is_udc_enable in sw_udc_udc.c, but that is in both sun4i and sun5i Allwinner sources??
<Turl> oliv3r: yes, uart0
<Turl> oliv3r: take zImage, cat zImage stuff.dtb > zImage-dtb then make uImage with zImage-dtb
<Turl> you need to enable appended dtb when configuring kernel too
<vakkov> Is there somebody who has the schematics of the q88 nuclear a13 mid tablet?
<vakkov> this one -
<oliv3r> Turl: ah, ok it's quite litteral then :p
<oliv3r> i will do that then
<oliv3r> what will the kernel do without the dtb passed to it? I assume not even uart will work?
<rm> "nuclear"?
<oliv3r> can I use mm to actually modify registers? I can use md 0x1c0000 for example to see sram register status in u-boot
<oliv3r> I assume the same can be used to modfy it?
<oliv3r> I also assume, its quite safe to say, that during u-boot memory display, those should be conciderd the absolute 'default' values?
<vakkov> rm: its device name is nuclear..
<vgrade> vakkov: tried to send this in PM but you closed the session,
drachensun__ has quit [Ping timeout: 246 seconds]
<hno> oliv3r, u-boot can modify any register you ask for. mw.l ... ...
ssvb has quit [Quit: Leaving]
<techn_> hno: ? It should be there?
ssvb has joined #linux-sunxi
<hno> Hmm.. let me check again. Was rebasing the changes to test in 3.4.
<hno> Odd. It's not there in my tree.
<techn_> 3.4 has it removed :/
<techn_> ok.. seems that 3.4 is missing commits :/
<oliv3r> hno: awesome, i assumed as much
<hno> techn_, indeed. and what a mixed commit that is compared to the commit message. Some of it is not relevant to 3.4 as well.
eebrah is now known as eebrah|away
<oliv3r> hno: is it normal that md 'wraps around'? I was examining sram registers, which run from 0x01c00000 to 0x01c00fff according to the documetnation. I md the first 0x100 bytes, all's well. But if i probe 0x100 - 0x200 I get the same first 0x100 bytes again
vakkov has quit [Ping timeout: 255 seconds]
<hno> oliv3r, that is pretty commmon in I/O space. Most I/O modules only partially decode the address lines.
<hno> techn_, I do not think the is_udc_enable part is reletant to 3.4, but hard to say what it's really about.
<hno> the udc driver uses the newer framework for this in 3.4 (framework also available in 3.0 I thnk)
Undertasker has quit [Quit: Leaving.]
drachensun has joined #linux-sunxi
paulk-desktop has quit [Quit: Ex-Chat]
<oliv3r> hno so is it fair to say that thus sram adres space runs in that limited space?
<hno> oliv3r, that address is I/O, not SRAM. It's the SRAM controller.
drachensun has quit [Ping timeout: 245 seconds]
<hno> There is some references to SRAM also being at that address but I it's confusing.
<hno> techn_, do drivers/usb/sunxi_usb/hcd/hcd0/sw_hcd0.c compile for you? I need for it to compile.
<oliv3r> sorry, thats whay i ment. touchtyping on tablet in ned atm
<oliv3r> what refrence do you mean
<hno> I don't remember exacly where it was now. Collected what I could find about the memory map at
<techn_> hno: it compiles fine. make system changed between 3.0 <-> 3.4?
<hno> techn_, not that I am aware of but maybe it has.
<oliv3r> then im not sure what you ref. too.
<techn_> hno: I tried adding ../.. to 3.0, that worked too
<hno> techn_, either way that include should move to include if it's needed by both.
<techn_> so your path is most likely correct one.. 3.4 is just pickyer
drachensun has joined #linux-sunxi
<hno> techn_, which platforms have wip/usb been tested on?
<techn_> cubie and a13_mid
<techn_> linux and android
<techn_> gotta sleep. good night :)
<oliv3r> nn
<hno> Good night. Should do the same.
<oliv3r> me 3
<oliv3r> before you, what did you mean with sram?
<hno> Static RAM embedded into the CPU. There is quite plenty of it in the A10.
<oliv3r> yeah
<oliv3r> i was speaking of the sram controler registers
<oliv3r> was poking those
mdfe has quit [Disconnected by services]
mdfe has joined #linux-sunxi
mdfe has quit [Disconnected by services]
mdfe_ has joined #linux-sunxi
mdfe_ has quit [Remote host closed the connection]
deasy has quit [Quit: *freesoftware song* :p]
da-dave has joined #linux-sunxi
da-dave has quit [Client Quit]
da-dave has joined #linux-sunxi
da-dave has left #linux-sunxi [#linux-sunxi]