Turl changed the topic of #linux-sunxi to: Allwinner/sunxi development discussion - Don't ask to ask. Just ask and wait! - See http://linux-sunxi.org | https://github.com/linux-sunxi/ | Logs at http://irclog.whitequark.org/linux-sunxi
akaizen has joined #linux-sunxi
Gerwin_J has quit [Quit: Gerwin_J]
wickwire has quit [Remote host closed the connection]
rz2k has joined #linux-sunxi
xavia has quit [Quit: Leaving.]
deffrag__ has joined #linux-sunxi
deffrag_ has quit [Ping timeout: 240 seconds]
xinj has joined #linux-sunxi
deasy has quit [Quit: Nom d'un quark, c'est Edmonton !]
egbert has quit [Ping timeout: 240 seconds]
egbert has joined #linux-sunxi
xavia has joined #linux-sunxi
srxa has joined #linux-sunxi
ecelis has joined #linux-sunxi
xavia has quit [Quit: Leaving.]
sap has joined #linux-sunxi
sappel has quit [Ping timeout: 268 seconds]
xinj has quit [Read error: Connection reset by peer]
xinj has joined #linux-sunxi
pseudomind_ has joined #linux-sunxi
<wens> ccaione: yeah sure
xinj has quit [Quit: xinj]
<libv> heh
<libv> ubuntu one is shutting down
<wens> I heard
<libv> so much for a lot of the cubie files stored there
setkeh has quit [Ping timeout: 240 seconds]
setkeh has joined #linux-sunxi
FreezingCold has joined #linux-sunxi
srxa has quit [Ping timeout: 268 seconds]
TheSeven has quit [Ping timeout: 264 seconds]
TheSeven has joined #linux-sunxi
<Turl> libv: oh right :|
rz2k has quit []
arky has joined #linux-sunxi
JohnDoe_71Rus has quit [Quit: KVIrc Insomnia 4.0.2, revision: 402, sources date: 20100627, built on: 2010-08-02 17:44:33 UTC 402 http://www.kvirc.net/]
akaizen has quit [Remote host closed the connection]
zumbi has quit [Ping timeout: 265 seconds]
zumbi has joined #linux-sunxi
* megal0maniac_afk gives lauri a cookie
pseudomind_ has quit [Quit: pseudomind_]
deffrag__ has quit [Remote host closed the connection]
deffrag__ has joined #linux-sunxi
akaizen has joined #linux-sunxi
akaizen has quit [Ping timeout: 240 seconds]
deffrag__ is now known as deffrag
bertrik has joined #linux-sunxi
Black_Horseman has quit [Quit: Zwi se logou mou!!!]
deffrag has quit []
deffrag has joined #linux-sunxi
sehraf has joined #linux-sunxi
shineworld has joined #linux-sunxi
Fusing has joined #linux-sunxi
arky has quit [Remote host closed the connection]
leviathanch2 has joined #linux-sunxi
sehraf has quit [Quit: www.miranda-fusion.de ... be part of it...]
sehraf has joined #linux-sunxi
ultraa has joined #linux-sunxi
bertrik has quit [Remote host closed the connection]
ultraa has quit [Remote host closed the connection]
JohnDoe_71Rus has joined #linux-sunxi
gonobolinka has joined #linux-sunxi
Fusing has quit [Read error: Connection reset by peer]
Fusing has joined #linux-sunxi
gonobolinka has quit [Ping timeout: 240 seconds]
nicksydney has quit [Remote host closed the connection]
nicksydney has joined #linux-sunxi
theskilledworker has joined #linux-sunxi
<lauri> Good morning guys
<lauri> Does it happen to you that occasionally HDMI link goes down for a second?
<lauri> My screen just randomly goes blank as if the monitor thinks the HDMI connection was disturbed or something
<Wizzup> anything in dmesg?
<lauri> nope
<lauri> I have this in uEnv: hdmi.audio=EDID:0 disp.screen0_output_mode=1920x1080p60
theskilledworker has quit [Quit: Leaving]
<ssvb> lauri: try 1920x1080p50 if it is supported by your monitor
<megal0maniac_afk> Interesting. I have the same issue
<ssvb> lauri: 1920x1080 screen resolution taxes memory bandwidth a bit more than we would like to - http://ssvb.github.io/2013/06/27/fullhd-x11-desktop-performance-of-the-allwinner-a10.html
srxa_ has joined #linux-sunxi
<lauri> ssvb: so the higher resolution on the screen, the less resources are available for the software?
<ssvb> lauri: insufficient memory bandwidth remaining for the display controller could be the source of hdmi signal getting occasionally disturbed
<ssvb> CPU and Mali are competing for the memory bandwidth with the display controller
<lauri> ssvb: alright, that sound like it makes sense but shouldn't display controller be prioritized?
theskilledworker has joined #linux-sunxi
theskilledworker has quit [Changing host]
theskilledworker has joined #linux-sunxi
<lauri> I'll try the 50Hz trick
<ssvb> there are hints that the memory controller can be tweaked to adjust priorities, but none of the old experiments with it helped
<ssvb> maybe this should be revisited
<megal0maniac_afk> Maybe allwinner should release sources :/
<lauri> Well let's say in future we'll have Wayland rendering stuff using GLES, is that going to change anything?
<lauri> The screen framebuffer still sits in RAM and display controller has to access it?
jemk has joined #linux-sunxi
<ssvb> old chat log about dram access priority - http://irclog.whitequark.org/linux-sunxi/2013-09-12#4938931;
kivutar has joined #linux-sunxi
<ssvb> lauri: wayland is not going to magically fix/improve the hardware
<ssvb> lauri: yes, the framebuffer is in ram, there is no dedicated video memory
<lauri> so everything is rendered to RAM
<lauri> okay
<lauri> and 4K is pretty much hopeless? :D
<lauri> I was thinking yesterday whether I should get a 4K screen..
<ssvb> 4K is not going to be pretty with the current hardware, assuming that it even works :)
<ccaione> wens: the paragraph "sleep and wakeup" for the axp209 doesn't make sense to me (how reg 31h can determine which power rail is gonna be disabled?)
<ccaione> is there anything more in the chinese version?
<lauri> ssvb: tinymembench debian packaging is missing, right?
<ssvb> lauri: yes, it's a tiny tool, which probably does not even need packaging
<lauri> ssvb: I think I don't even have gcc installed on my cubie
<lauri> ssvb: It's easier for me to run DIST=jessie ARCH=armhf pdebuild and pdebuild-upload :D
<ssvb> there should be a gcc package in debian :)
<megal0maniac_afk> Quick question, what does u-boot look for on NAND? The magical .bin file that partitions the NAND and installs whatever has the wrong mbr for A20. So now I need to do this manually
<ssvb> one more tool which might be useful to get packaged if you are up to it - https://github.com/maxnet/a10-meminfo
<lauri> ssvb: yeah I could do that if you're interested, that guy should have a release though
<lauri> ssvb: also, could you add 0.5.0 tag for fbturbo?
<ssvb> is 0.4.0 not good for you?
<lauri> ssvb: if you say changes are insignificant then yeah I guess it's good enoguh
<ssvb> there are too few changes in git since 0.4.0
bonbons has joined #linux-sunxi
srxa_ has quit [Ping timeout: 240 seconds]
* megal0maniac_afk idly cuts a SATA cable in half
leviathanch2 has quit [Ping timeout: 240 seconds]
shineworld has quit [Quit: Leaving]
leviathanch2 has joined #linux-sunxi
paulk-aldrin has joined #linux-sunxi
Black_Horseman has joined #linux-sunxi
nicksydney has quit [Remote host closed the connection]
nicksydney has joined #linux-sunxi
bbrezillon has joined #linux-sunxi
Black_Horseman has quit [Quit: Zwi se logou mou!!!]
<wens> ccaione: write 1 to reg31h [3] to enable wakeup, PMU keeps a copy of reg12h
<wens> then do whatever you want to regulators
<wens> AXP waits for wakeup event (by PEK or GPIO)
theskilledworker has quit [Quit: Leaving]
<wens> if wakeup event occurs, restore reg12h
<wens> voltages are restored to default
<wens> --
<wens> that's the translation for the flow chart
wingrime has joined #linux-sunxi
<wingrime> jemk: ping
<wingrime> jemk: I failed init ACE from uboot
<wingrime> jemk: on A20
<wingrime> jemk: maybe you can try it on A10
kivutar has quit [Quit: Ex-Chat]
<jemk> wingrime: cant try at the moment, but i think i already tried it some time ago and it worked
<wingrime> jemk: on a20
<wingrime> jemk: I read only R0 zeros
<wingrime> ?
<jemk> on a20 only zeros, on a10 i think it worked
<wingrime> jemk: thats strange, on a13 even ACE clocks are RO. but not on a20
wickwire has joined #linux-sunxi
<wingrime> jemk: than I should add to wiki
<jemk> wingrime: maybe they forgot removing the clock
<wingrime> jemk: also, ACE blobs are still in a23 sdk
leviathanch2 has quit [Ping timeout: 240 seconds]
srxa has joined #linux-sunxi
<libv> megal0maniac_afk: there's a reason why the howto talks about backing up nanda
<libv> megal0maniac_afk: you probably now have to flash a fresh image using livesuit
<megal0maniac_afk> libv: That's okay :)
setkeh has quit [Ping timeout: 240 seconds]
leviathanch2 has joined #linux-sunxi
setkeh has joined #linux-sunxi
bbrezillon has quit [Ping timeout: 240 seconds]
<srxa> is there any reason or limitations for mali 400 on a20 devices about clock settings, many devices are set to work around 300mhz with mali, and on wiki there is information that mali 400 core works at 500 mhz?
<libv> srxa: where does it state that?
<wingrime> srxa: techprocess/litography library limitations
<ssvb> srxa: where did you see the information about 500mhz? the a10 and a20 manuals specify the mali clock limit as 381mhz
<libv> srxa: i have mali 400 running at 800MHz
<libv> srxa: on _different_ hardware
<srxa> in the table for mali400 core clock rate on this page
<srxa> ok, i understand, thx.
<libv> srxa: when you say wiki, it means our wiki. use the word wikipedia if you mean wikipedia next time
<srxa> of course, sorry :)
<ssvb> srxa: btw, mali starts to misbehave if overclocked higher than 408mhz on a10 and a20 (on my devices at least)
<ssvb> srxa: bumping the mali clock speed a little bit in the linux-sunxi kernels may make sense
<srxa> my settings are for 320mhz now, i will try what is maximum for my device
<mru> a higher clock rate usually needs a higher voltage too
<mru> and too high voltage can damage the chip
jemk has quit [Read error: Operation timed out]
<ssvb> srxa: except that I would like to have a reliable temperature monitoring support first and hooks for throttling the cpu/mali clock speed on overheating
<srxa> i am aware of that, i will be gentle ;)
<wingrime> mru: can AXP control voltage for overclock?
<mru> I don't know how voltages are controlled on allwinner chips
<gzamboni> probably via I2C, you can use i2ctools to change the values
<gzamboni> if you want to do it in the userspace, otherwise you can change the fex / dts values
<ssvb> wingrime: fex files have voltage settings, which are used by the axp driver
<ssvb> wingrime: but reading the documentation and sources is needed for all the fine details, as usual
<gzamboni> question: for the fisrt use of a newly done board, does the AXP have the right voltages values to init the allwinner chip ?
<gzamboni> or do you have to use an external i2c master to set them up for the 1st usage
<wens> some of the defaults are usable, if you follow their guideline
deffrag has quit [Read error: Connection reset by peer]
<megal0maniac_afk> libv: LiveSuit just says "invalid image file" to everything
<megal0maniac_afk> They are specifically NAND .img files
<libv> megal0maniac_afk: talk to the itead guys
<libv> megal0maniac_afk: it's very much not our problem here, sorry, you wouldn't have been in this mess if you had followed our howto
deffrag has joined #linux-sunxi
<megal0maniac_afk> libv: Well nothing is "your problem" per se. But this is IRC, not a company support line
<megal0maniac_afk> If anyone can help with any problem, cool. If not, then that's fine too. It's a community
<ssvb> srxa: are you using a10 or a20?
<srxa> a20
<srxa> ssvb: olimex board
<ssvb> srxa: ok, just iirc mali is typically clocked at 312mhz on a20 and at 320mhz on a10
<srxa> ssvb: i change setting for pll4 to 960 and mali clock div to 3
<ssvb> srxa: ok, also there is a dedicated pll8 for mali on a20, which can be taken into use for more fine grained clock control
<srxa> oh, ok, i will prepare my hardware with some radiators and temperature monitoring and i will play with settings
<ssvb> :)
deasy has joined #linux-sunxi
xavia has joined #linux-sunxi
xavia has quit [Quit: Leaving.]
FR^2 has joined #linux-sunxi
jemk has joined #linux-sunxi
Seppoz has quit []
smotocel69 has joined #linux-sunxi
bbrezillon has joined #linux-sunxi
bbrezillon has quit [Ping timeout: 240 seconds]
bertrik has joined #linux-sunxi
<lauri> btw, is it possible to do something with cubietruck audio output? Once the board is turned off the analog output voltage level seems to be floating freely causing the 50Hz mains noise to be emitted from speakers..
<mru> stick a filter between the board and the amp
<lauri> mru: could you be more specific?
<libv> or get a better power supply
<libv> this board is provided with 5V DC
<megal0maniac_afk> Why are the Allwinner SoCs used in so many "development boards" if the support is so limited?
<mru> yeah, how is 50Hz getting onto the board?
<mru> megal0maniac_afk: so people can "develop" support of course
nicksydney has quit [Remote host closed the connection]
<libv> if that is ripple free, then the board will not fluctuate like that
<libv> megal0maniac_afk: what do you mean with limited support?
<libv> megal0maniac_afk: we are about the most advanced out there, for this range of hardware
<mru> lauri: doesn't the cubietruck have a digital audio output?
<lauri> mru: I have analog speakers..
<mru> I'm pretty sure I saw an optical spdif output on it
<libv> mru: yeah
<megal0maniac_afk> Well, like we've discussed, the lack of support for mtd, cedarx etc. The manufacturer themselves aren't very helpful. I'm not saying that linux-sunxi hasn't come a long way, but it started at a serious disadvantage
<mru> I'd say it's doing remarkably well considering what it is
<libv> megal0maniac_afk: do you think this is any different on other hw?
<libv> megal0maniac_afk: you do know that it will soon be 11ys that i have been developing graphics drivers, right?
<mru> a lot has changed in that time
<libv> megal0maniac_afk: i pioneered modesetting and convinced half the planet that bios based modesetting was stupid and wrong, and could be fixed with proper code
<libv> megal0maniac_afk: i was, together with egbert and emmes, responsible for freeing ATI
<lauri> I think the problem is that speakers arent grounded..
<libv> megal0maniac_afk: i also did a lot of BIOS REing for flashrom, and added native VGA mode to VIA hw on coreboot
<megal0maniac_afk> mru: I agree, it is doing well. But am I wrong in saying that this is not made any easier by the manufacturer, who themselves don't even list linux as supported?
* mru remembers when ati mach64 was all the rage
<libv> megal0maniac_afk: and now i am on freeing ARM GPU drivers
<mru> in fact, they even branded some of the cards "rage $foo"
<libv> megal0maniac_afk: and i believe that this is about the free-est and best supported hw out there
<mru> megal0maniac_afk: what do they list as supported?
<megal0maniac_afk> mru: Android (yes I know it's Linux based, kind of)
<mru> android is linux
smotocel69 is now known as pirea
<lauri> Android is Linux but not GNU
<mru> close enough
<mru> none of the gnu tools need hardware-specific support
<lauri> Android has shitty BSD-licensed replacements for GNU userspace utils
<mru> so?
<mru> once you have a kernel you can run whatever userspace you like
<libv> lauri: usually, nothing stops you from installing busybox
<mru> you can run ubuntu on an android kernel if you want
<libv> (mru: apart from the userspace blobs :p)
<lauri> libv: Yeah I'm running Debian chroot on my Oppo Find 5
<mru> libv: nothing a little hacker can't fix
<mru> hackery
<megal0maniac_afk> libv: I fully respect and appreciate what you're doing, I'm just curious as to why this platform is so popular, being a WiP essentially
<lauri> but the whole Android in my opinion is a Franksenstein of operating systems
<libv> megal0maniac_afk: because the others are even bigger WIPs?
<mru> and it's cheap
<lauri> btw, enabling additional UART ports on Cubietruck is trivial as editing FEX files right?
<libv> megal0maniac_afk: at one poin tin 2012, for the first time for android hw, ever, some kernel and u-boot code fell out of device manufacturers
xinj has joined #linux-sunxi
<libv> megal0maniac_afk: and a lot of people went out and bought hw, and tom cubie started selling the mele a1000 over aliexpress so people in the west could get it too
Gerwin_J has joined #linux-sunxi
<libv> then the cubieboard was created, and then it really took off
<libv> then olimex joined the game, and now we have OSHW boards.
<libv> some of the code is still crapy chinese vendor code, but that's always how it is
<libv> only intel does better as intel employs a small army to write upstream code
<mru> TI at least has the decency to release full specs
<mru> (apart from GPU)
<libv> megal0maniac_afk: go out and buy another board
<libv> megal0maniac_afk: you will end up whining even more.
<mru> freescale is also pretty good with specs
<libv> and you will not have the benefit of a community like ours
<libv> and yes, this includes the rpi, as the rpi still is a closed system (even though they have no shown in a big way that they are working on that)
<mru> the beagleboard community is pretty good
<mru> the rpi is a piece of shit hardware with a piece of shit community
<libv> mru: that rules megal0maniac_afk out, as megal0maniac_afk didn't buy like a cubie or olimex board, he got the itead stuff
<mru> the rpi chip is actually just a validation chip for the gpu
<libv> mru: i am the guy who exposed that.
<mru> for some reason they made too many and found a clever way to offload them
<wens> mru: that sounds interesting
<mru> when it started selling really well they probably made a few more though
<mru> or someone is making them at least
<mru> I heard bcm gave the blueprints to someone else to manufacture at their leasure
<megal0maniac_afk> libv: This is all quite interesting. I thought the rpi was fairly open, in the sense that everything
kivutar has joined #linux-sunxi
<megal0maniac_afk> just "works" out of the box
<megal0maniac_afk> I still don't want one, but that's the impression I got
<rm> that's a novel definition of "open" :)
<libv> megal0maniac_afk: open != works out of the box
<megal0maniac_afk> libv: True
<libv> megal0maniac_afk: rpi is a single device, and every kid has written a howto
<libv> for their specific setup
<libv> plus, every kid has made a dd-able image.
<libv> megal0maniac_afk: this is probably what you are looking for in life
<libv> this is your misdefinition of open
<wens> the cubies work out of the box from nand as well, don't they?
FreezingCold has quit [Ping timeout: 240 seconds]
<libv> wens: the itead did as well, but then megal0maniac_afk went and used nand-part and such without reading our howto
<megal0maniac_afk> libv: I'm only here to learn. I don't claim that anything I know or do is correct
<libv> megal0maniac_afk: sure, but do talk to the itead guys to see if they can resolve this for you
<wens> hopefully they can give you a factory image, and tools to flash it
<megal0maniac_afk> I am, that's not what I'm talking about here though
<libv> megal0maniac_afk: ask them why your livesuit image cannot be flashed
netlynx has joined #linux-sunxi
<megal0maniac_afk> libv: How did the Dockstar (armv5te) come to have u-boot sources and proper mtd support? Through a community like this and reverse engineering, or manufacturers?
<libv> megal0maniac_afk: how old is the dockstar?
<libv> you know what, let's ask wikipedia.
<megal0maniac_afk> ~2009?
bbrezillon has joined #linux-sunxi
srxa has quit [Remote host closed the connection]
<libv> did seagate comply with the gpl straight away?
<megal0maniac_afk> Most likely not
<libv> 2009 is 5 years ago
<libv> linux-sunxi is not even 2 years old
<libv> linux-sunxi also deals with much more complicated hardware, and has a much wider range of devices out there
<megal0maniac_afk> libv: Understood
<libv> megal0maniac_afk: do you have a media decoding engine on your dockstar SoC?
<libv> the mtd thing, yes, i give you that, but that is due to what allwinner did
<megal0maniac_afk> libv: O
<libv> megal0maniac_afk: so you are comparing apples and oranges
<megal0maniac_afk> I'm not trying to swing a punch at linux-sunxi, I'm genuinely interested
<libv> and i believe we had this discussion several days before
<libv> where i said that there is probably only one dockstar with minor revisions
<libv> whereas sunxi...
<megal0maniac_afk> Well, no, there are a number of devices. The dockstar is one of them, but I see your point. They're also more mature, and headless
<libv> the advantage is that anyone who spends a minute on ebay, and spends 20EUR and upwards (that's for how much older allwinner tablets now go), and get an allwinner device
<megal0maniac_afk> Which I understand is one of the major challenges - media decoding and video output
<libv> there are not many many hundreds of fundamentally different dockstar designs
<libv> and we've only really been going for 2 years.
<libv> 2 years ago, i was just working up to demoing a textured cube on an amlogic based android tablet
<libv> i could only dream of a proper linux on a mali based device then
<libv> i got my mele in june 2012
<megal0maniac_afk> Okay, so to go back to my original question. Is the motivation for manufacturers to make and sell products based on Allwinner SoCs based on the price/performance, and active developer community?
<libv> no, the latter is mostly ignored.
<libv> megal0maniac_afk: how long have you been using open source software?
<libv> not long i think, as otherwise you would know that the latter has never played a measurable role
<megal0maniac_afk> libv: I agree, but in terms of "development platforms", it's kind of crucial to their success. Allwinner aren't doing anything about it that I can see - it's all here
<megal0maniac_afk> If this community hadn't been doing what you've been doing for the last 2 years, I doubt whether the "IBOX" or Olinuxino would even exist
<libv> megal0maniac_afk: nope
<libv> megal0maniac_afk: we kind of depend on cubie and olimex
<libv> megal0maniac_afk: and companies like itead are only in it to get a piece of the pie as well, hence their cubieboard compatibility
<megal0maniac_afk> Am I right in saying that Allwinner's target is Android-based TV boxes?
<libv> anything android
<megal0maniac_afk> Well yeah, tablets and such as well
<megal0maniac_afk> I am learning a lot here
<megal0maniac_afk> libv: I apologise if I frustrate you from time to time, but I'm understanding this better as I go :)
leviathanch2 has quit [Ping timeout: 240 seconds]
<megal0maniac_afk> With the Dockstar, I started out not knowing anything about mtd, FPUs, and ARM in general. I followed bits and pieces of guides with big warnings and outdated instructions, eventually figuring out what was applicable, what wasn't and what was safe to try out. I documented everything and now I can recreate a week's worth of work in 20 minutes. And I'm not following instructions blindly. I expect my experiences will be similar with
leviathanch2 has joined #linux-sunxi
<megal0maniac_afk> My premature frustration is mainly due to the way in which this ibox was marketed. It is not a polished product, it wasn't meant to be. Yet they advertise it as such
<megal0maniac_afk> So my expectations were inflated
theskilledworker has joined #linux-sunxi
deasy has quit [Quit: Nom d'un quark, c'est Edmonton !]
Net147 has joined #linux-sunxi
libv has quit [Ping timeout: 252 seconds]
leviathanch2 has quit [Ping timeout: 240 seconds]
libv has joined #linux-sunxi
Net147 has quit [Quit: HydraIRC -> http://www.hydrairc.com <- *I* use it, so it must be good!]
leviathanch2 has joined #linux-sunxi
<libv> megal0maniac_afk: sorry, was off replacing the grips on my better halfs bicycle.
<libv> Let's see whether rajesh does NDH this time round.
leviathanch2 has quit [Ping timeout: 240 seconds]
<megal0maniac_afk> libv: It's all good. As long as we're on the same page
<libv> ah, he was told on the 24th already
<ssvb> libv: hmm, is srxa the same person as rajesh?
<libv> is he, good, then i can complain to him directly when he comes back
<ssvb> well, I don't know, but 4K display resolution and 500mhz mali are mentioned at the link given by rajesh in the ml
<ssvb> http://en.sztena.com/pro/&productId=dcc094f2-b5a0-45a1-94cb-8575b68336bc&comp_stats=comp-FrontProducts_list01-004.html
<libv> yes, amazingly
<ssvb> or maybe not, because srxa mentioned olimex a20 board here
<ssvb> just a strange coincidence :)
xinj has quit [Ping timeout: 240 seconds]
<lauri> So 1920x1080 at 50Hz still gives me glitches..
<libv> he still has no git account though
xavia has joined #linux-sunxi
<lauri> ehm?
<libv> err, no wiki account
<libv> Rajesh
<ssvb> lauri: you can also enable 'scaler mode' in fex (the pixels are fetched by FE instead of BE from the framebuffer) and check the dram clock speed
<lauri> ssvb: could you ellaborate what's the difference?
theskilledworker has quit [Ping timeout: 240 seconds]
<ssvb> lauri: the scaler mode is more 'gentle' on the memory controller, and the difference is observable in benchmarks
<ssvb> lauri: the down side is that the 'scaler mode' wastes one hardware scaler, and this is a precious shared resource (used for scaled layers)
<ssvb> lauri: as for the dram speed, you have a cubietruck, right?
<ssvb> lauri: so normally you would have 432mhz dram clock speed, but you can check it with https://github.com/maxnet/a10-meminfo anyway to be sure
<ssvb> lauri: still the first thing to try would be to go down to 1280x720 resolution or to 16bpp color depth and check if these glitches disappear
nove has joined #linux-sunxi
JohnDoe_71Rus has quit [Read error: Connection reset by peer]
JohnDoe_71Rus has joined #linux-sunxi
akaizen_ has joined #linux-sunxi
<nove> another one fruit, http://www.lemaker.org/
<libv> banana pi?
<libv> now it's getting really ridiculous
<libv> is it me, or is the layout similar to rpi?
<megal0maniac_afk> Almost identical
<megal0maniac_afk> "Fully compatible with Raspberry Pi"
<megal0maniac_afk> Ooh! Gmac
<libv> rpi is arm11
<lauri> ssvb: scaler mode was already enabled for fb0
<lauri> ssvb: for fb1 it was disabled, btw what is up with fb1?
<rm> nice idea to reuse R Pi cases and GPIO expansion boards
<lauri> ssvb: you can have dual head if output type is set to non-zero value?
<rm> but they have an 1.2 MB JPEG on their front pae
<rm> ge
<rm> makes me wonder if they actually have a clue about anything
<libv> rm: true, it is nice to say that the formfactor is compatible
<libv> but "fully compatible" is a whole different thing
<ssvb> lauri: yes, you can have dual head (vga and hdmi monitors connected simultaneously)
<lauri> alrighty
<lauri> anywho fb0 was already in scaler mode
<ssvb> lauri: this means that without the scaler mode the glitches would be likely even worse
<libv> yay, rajesh found the example page :)
<wens> doesn't say how you can get one
<ssvb> lauri: have you tried reducing the screen resolution to confirm that it really only happens with 1920x1080?
<libv> mk802ii is solid, only 50 of 84 left to go.
<ssvb> libv: there is (or will be) also HummingBoard, which is i.MX6 in the Raspberry Pi formfactor
<libv> 4 of those i own, so that should become 46/84, and hopefully Rajesh can be helped enough to take that device off the list as well
<libv> ssvb: it's a good idea, but... who still uses cvbs these days...
<libv> we already wrote that off when doing radeonhd back in 2007
<lauri> ssvb: not yet @ reducing resolution
<lauri> could anyone paste me a working OTG host config for CT
Gerwin_J has quit [Quit: Gerwin_J]
<lauri> never mind
<lauri> it works now
<lauri> I also enabled UART4, any clue where I could find the pinout of CT?
<libv> lauri: why do you need 5 uarts?
<lauri> I am experimenting here, I've got XBee modules, NFC cardreader etc I would like to try hooking onto Cubietruck
<lauri> login, screen lock & unlock would be pretty neat with NFC :D
<wens> lauri: get the schematics from cubietech
<lauri> right..
<lauri> under heavy load I am experiencing errors from SATA controller. Is this most probably issue with power supply?
<lauri> failed command is READ FPDMA QUEUED
<Turl> jemk: nice work on the dram stuff :D
<jemk> Turl: thx, I hope it was worth the effort
<jemk> but it still didn't crash
rz2k has joined #linux-sunxi
<ssvb> nove: wat? one more banana?
<megal0maniac_afk> Now, if they could do an x86 one of those...
* megal0maniac_afk runs
<rz2k> nove: anything that has "via" label is bad. period.
<nove> ssvb, appear to be where name come from, or aleast the older reference to the name
<rz2k> megal0maniac_afk: they have x86 something, official linux support: 2.6.2y...
leviathanch has quit [Remote host closed the connection]
<lauri> Turl: thanks, exactly what I was looking for :P
<nove> hint, the legend of the first image
leviathanch has joined #linux-sunxi
<lauri> I am a little bit confused about the virtual ports thing.. In the wiki I can see UART4 is connected to CN9 pins 13 and 14
<lauri> which correspond to port G pins 10 and 11?
<lauri> in my FEX file I currently have PH04 and PH05
<lauri> is this legit or I should change PH04 and PH05 to PG10 and PG11?
<megal0maniac_afk> lauri: My schematic says it's on PG10/11
<jemk> lauri: ph04 and ph05 is another pimux for uart4, not available at cubietruck. you have to change it to pg10 and pg11
<megal0maniac_afk> Yeah, just found it at PH04/05. So it's multiplexed
<megal0maniac_afk> How is multiplexing managed? :/
<lauri> could you elaborate @ multiplexing?
<lauri> I think there is same issue with Bluetooth module? The firmware upload utility seems to get no response from the module..
<megal0maniac_afk> Well a peripheral can be made available at various locations by means of multiplexing. It's usually managed by registers on the CPU, but I imagine that would be abstracted. I just don't know how
ZetaNeta has joined #linux-sunxi
<ZetaNeta> I know this is sunxi-unrelated, but is there any camera i can use from GPIO?
<megal0maniac_afk> So stick it on PG10/11 and see what happens. Maybe it will magically work :) Or PG10/11 is the default
<Turl> megal0maniac_afk: the PHNN<N><N><N>... stuff you write on fex
<Turl> indicates the pin, the function (ie, what's multiplexed to there) and a bunch of other things
<Turl> each pin has a handful of possible modes (ie, stuff that can "output" on it)
<megal0maniac_afk> Turl: Ah, so the fex file deals with mapping as well as multiplexing?
<lauri> ehm.. tell me in the dumbes way what should I do? :P
<megal0maniac_afk> Makes sense
akaizen_ has quit [Remote host closed the connection]
<megal0maniac_afk> lauri: Make it PG10/11
<Turl> megal0maniac_afk: well, the mapping *is* the multiplexing
<lauri> what about the other flags?
<lauri> port:PH04<4><1><default><default> --> port:PG10<4><1><default><default> is good enough or I should change the other flags aswell?
<Turl> the other flags are drive strength, pull up/down and don't remember
<lauri> okay
<lauri> what about the bluetooth?
<Turl> what about it?
<lauri> is internally connected via UART2 right?
<lauri> and not exported to the header
mkutsevol has joined #linux-sunxi
<Turl> it's on a uart, yes
<Turl> dunno which one
<Turl> wens probably knows :)
<lauri> /dev/ttyS1
<lauri> writing
<lauri> 01 03 0c 00
<ZetaNeta> so....
<ZetaNeta> anyone know anything?
<lauri> this is what the broadcom bluetooth firmware upload utility keeps on outputting
<lauri> I assume this means bluetooth module does not respond
<Turl> lauri: brcmpatchplus or whatever it's called?
<lauri> sudo brcm_patchram_plus -d --patchram /lib/firmware/ap6210/bcm20710a1.hcd --enable_hci --bd_addr 11:22:33:44:55:66 --no2bytes --tosleep 1000 /dev/ttyS1
<Turl> did you use the cubietech driver for that thing? the linux sunxi one doesn't work as far as I'm aware
<lauri> by driver what do you mean exactly?
<lauri> so far it hasn't got much to do with the driver yet?
<lauri> since the firmware is to be uploaded via UART
<lauri> but I haven't got past uploading the firmware
mkutsevol has left #linux-sunxi [#linux-sunxi]
mkutsevol has joined #linux-sunxi
<Turl> lauri: well yeah, but the chip needs to be blocked as well, fed with power and other things
<Turl> clocked*, not blocked
<lauri> well the Wifi works
<lauri> bcmdhd is loaded
<lauri> I am running 3.4.79-sun7i+
<Turl> lauri: are you using the firmware from cubietech or did you get it somewhere else?
<Turl> libv: http://dl.cubieboard.org/software/a23/tools/ btw, just seen it :)
Gerwin_J has joined #linux-sunxi
<lauri> Turl: it's from patwood's page I think
peterbjornx has joined #linux-sunxi
<Turl> libv: also very cool goodies around that :)
<libv> nice :)
<lauri> I am not expert of course but I assume since bcmdhd is loaded properly that the chip is up and running waiting for bluetooth firmware to be loaded but I think my FEX config is messed up and that's why I can't upload the firmware..
<wens> lauri: BT needs external clock and gpio enabled
<lauri> wens: could you elaborate :P
<wens> I'm not familiar with 3.4 kernel
setkeh has quit [Ping timeout: 240 seconds]
<Turl> libv: like pdfs :p
<wens> so I can't say if the kernel has proper support for the needed settings
setkeh has joined #linux-sunxi
<nove> why this?
<nove> not even a email?
<peterbjornx> so, is there and chance of getting usb controller 1 running on sunxi-3.4
<Turl> peterbjornx: eh?
<jemk> nove: wow, that looks good.
<nove> dear cubietech, if you don't help us, we can't help you, is simples like that
bbrezillon has quit [Ping timeout: 255 seconds]
<jemk> nove: VENC_CODEC_TYPE: VENC_CODEC_VP8 !?!
<jemk> ok, only on aw1639, whatever this is
<nove> jemk: sorry, but is bad timing, i am a bit busy right now, and will not be able to work to make this usable, until next month
<jemk> nove: hehe, no problem, i should do other things too, but it looked good
<lauri> Okay UART4 works with PG10/PG11
<lauri> Going to attach NFC reader and let's see what we can get cookin' :D
<lauri> I bet most default UART configs are broken
paulk-aldrin has quit [Ping timeout: 268 seconds]
<Turl> lauri: serial nfc reader? cool
<nove> people, the behavior of cubietech of not tell us of new things is unacceptable, and we should let clear know cubietech of that
<peterbjornx> turl, its not implemented is it?
<peterbjornx> on sun5i
<peterbjornx> i only see code for hcd0
<nove> and is not the first time, last january there was a new android sdk for a20, with openmax, and we also weren't informed
akaizen has joined #linux-sunxi
* nove will add dl.cubieboard.org to the weekly must check
<Turl> peterbjornx: and A13 has just one controller, checking the user manual
<Turl> peterbjornx: so dunno what you're referring to :)
<peterbjornx> err, i got a tablet which has both a builtin wifi module and a otg port
<peterbjornx> wifi module is usb
<Turl> peterbjornx: are you sure of that?
<Turl> peterbjornx: some are SDIO
<Turl> peterbjornx: ah, but you said OTG
<peterbjornx> yes
<Turl> peterbjornx: yeah, maybe it's using the host one for wifi
<Turl> the OTG is a different controller
<peterbjornx> and linux only works with otg
<peterbjornx> i got my script set up properly
<peterbjornx> im using a modified defconfig but no usb related changes
<Turl> host should be working
<Turl> can you paste it?
<Turl> and your fex
<Turl> and a dmesg :)
<peterbjornx> dmesg is going to be hard cause i dont have access to ui or net atm
<peterbjornx> i do have serial but thats on a windows vm
<peterbjornx> script http://sprunge.us/YgbS
<peterbjornx> o btw, im not using wifi on that port anymore
<peterbjornx> i connected a hub
paulk-aldrin has joined #linux-sunxi
<peterbjornx> http://sprunge.us/JDER kconfig
<Turl> usb_drv_vbus_gpio = port:power203<1><0><default><0> <- that's interesting
<Turl> wens: ^
<peterbjornx> i know
<peterbjornx> but vbus is active
<lauri> Turl: yeah @ serial NFC
akaizen has quit [Remote host closed the connection]
akaizen has joined #linux-sunxi
jemk has quit [Quit: bye]
geecko has quit [Ping timeout: 268 seconds]
paulk-aldrin has quit [Remote host closed the connection]
paulk-aldrin has joined #linux-sunxi
Gerwin_J has quit [Quit: Gerwin_J]
bertrik has quit [Remote host closed the connection]
<nove> ah better, linux-codec-release-V1.2.rar has date of 28 of March
<peterbjornx> turl, any clue?
geecko has joined #linux-sunxi
<Turl> peterbjornx: hard to tell without looking at dmesg
setkeh has quit [Ping timeout: 240 seconds]
setkeh has joined #linux-sunxi
<peterbjornx> oops
<lauri> Interestingly 3.3V pins in CN9 don't output 3.3V :/
<peterbjornx> http://pastebin.com/bLHSTUS8 <- turl
<lauri> I noticed that HDMI issues are correlated with load
akaizen has quit [Remote host closed the connection]
<ssvb> lauri: more like with the heavy memory accesses done by the CPU, which are starving framebuffer scanout
<lauri> I guessed so
<lauri> btw, could 30Hz also work?
<Turl> rm: can you update your wiki (a10/debian)? :)
<Turl> rm: the uboot link and instructions are outdated
<ssvb> lauri: if your monitor supports it (or TV?), then you can give it a try
<rm> I can
<peterbjornx> turl, i posted console dump above
<rm> well I think those still work but for A10 only
<lauri> ssvb: btw, isn't it possible to read framebuffer only when some pixels have changed?
<Turl> rm: yeah, but there's like 3 devices there :)
<lauri> ssvb: that could potentially save a lot bandwidth
<Turl> peterbjornx: can you lsusb?
* nove was to send a email to list to make know about the new binaries, but after half way written, why bother
deasy has joined #linux-sunxi
<peterbjornx> not atm but previously it would list only otg root hub
<rm> Turl, if you can write up better instructions at linux-sunxi.org wiki, I can just link to that
<rm> or I will look into updating that myself
<rm> but not now, a bit later (monday+)
<ssvb> lauri: there are some advanced things like framebuffer compression (in other hardware)
<lauri> ssvb: I noticed Intel has something like that
<lauri> ssvb: anywho even if lower resolution or lower resolution resolves the problem then what?
wickwire has quit [Ping timeout: 240 seconds]
<ssvb> lauri: then you have confirmed that it's this particular issue
<lauri> I'll let you know next boot
<ssvb> lauri: if you really want 1920x1080 with full color depth, then you may consider clocking dram faster
<lauri> ssvb: how would you do that?
<lauri> btw, would 24-bit help or cause more issues?
jemk has joined #linux-sunxi
<lauri> ssvb: so I have to tune u-boot compilation, right?
<ssvb> lauri: yes, that's currently the only way to change the memory clock speed
<ssvb> lauri: 24-bit per pixel (3 bytes) would help, but will also have a performance cost
<lauri> I think it's at 480MHz
<lauri> what should I put there?
<lauri> sorry, I misread 480MHz was for cb2
<ssvb> lauri: really? did a10-meminfo report that?
<lauri> 432MHz is for ct
<lauri> I haven't run meminfo just yet, probably going to compile it soon
<lauri> so it should be switched to 480MHz aswell?
<ssvb> lauri: you can increase it in 24mhz steps, so that would be 456 or 480
<ssvb> lauri: not all cubietrucks can run memory at 480mhz reliably, but you can try
<lauri> why not?
Wizzup has quit [Quit: Change me]
Gerwin_J has joined #linux-sunxi
<ssvb> lauri: a different PCB, longer tracks to DDR chips, more DDR3 chips than in Cubieboard, ...
<lauri> alrighty
<ssvb> lauri: if it was stable on every cubietruck, then u-boot would have 480mhz for it by default
<lauri> I recompiled it and overwrote if using dd
<lauri> lemme see if it boots
<ssvb> lauri: however proper dram configuration is currently being investigated - https://www.mail-archive.com/linux-sunxi@googlegroups.com/msg03239.html
<lauri> well it booted
wickwire has joined #linux-sunxi
<mnemoc> Turl: uh
<lauri> blerrgh still HDMI issues..
<ssvb> lauri: check with a10-meminfo if the changes have actually taken effect
<lauri> yep
<Turl> lauri: do you have some other equipment working where you are?
<Turl> lauri: like a fridge?
wingrime has quit [Quit: Lost terminal]
<lauri> Turl: ehm, what? :D
<Turl> lauri: yeah, I know it sounds weird
<lauri> You mean to put it in the fridge to see whether heat might be the issue?
<Turl> lauri: but that same thing (losing signal) happens on my x86 PC with a HDMI monitor
<Turl> when my fridge kicks in/stops
<Turl> no, no
<lauri> aa
<Turl> try unplugging the fridge a bit :p
Gerwin_J has quit [Quit: Gerwin_J]
<lauri> yeah there is fridge
<lauri> dram_clk = 480
<lauri> so it's at 480MHz alright
<ssvb> lauri: but yes, higher memory clock speed, 50hz refresh and "scaler mode" help to reduce this effect, but still don't eliminate it completely
<ssvb> lauri: is it really "that" bad?
<lauri> it could be that the power supply grounding is issue
<ssvb> lauri: I mean how easy is it for you to trigger it?
<lauri> the monitor is grounded but cubietruck powersupply is not
<lauri> okay now I am starting to think the problem is most probably grounding
<lauri> it's under full load and no glitches
<ssvb> lauri: that's why you could try lower resolution or 16bpp color depth
<ssvb> lauri: if the glitches are still there, then it's likely your fridge, grounding or whatever
akaizen has joined #linux-sunxi
<lauri> anywho overclocking memory should increate the performance generally aswell?
<lauri> *increase
<ssvb> yes
<Turl> as long as you're memory bound :)
<lauri> 504MHz will most probably work aswell?
geecko has quit [Ping timeout: 255 seconds]
<lauri> overclocking DRAM and CPU do not affect voltage right? And therefore lifetime is not reduced?
Gerwin_J has joined #linux-sunxi
akaizen has quit [Remote host closed the connection]
Gerwin_J has quit [Client Quit]
Gerwin_J has joined #linux-sunxi
Gerwin_J has quit [Client Quit]
nove has quit [Quit: nove]
xavia has left #linux-sunxi ["PART #interlock :PART #arduino :PART #wnydev :PART #buffalolab :PART #docker :PART #symfony :PART #virtuvia :QUIT :Leaving."]
<mnemoc> Turl: fixed
<libv> rajesh apparently fried his "ttl to usb" adapter
<libv> i think it might be his uart, but i now told him to come join us here
<libv> so we can find out what he did there, and perhaps help him with usb as well
akaizen has joined #linux-sunxi
mallah has joined #linux-sunxi
<mallah> USB ports work with android with kernel 3.3.0+ with same script.bin sunxi-3.4 panics in EHCI
<libv> mallah: this ttl to usb adapter of yours...
<libv> mallah: was this a uart/serial adapter?
<mallah> it was the one that cubie sells!
<mallah> which got fried.
<libv> you wrote in your wiki page that this had something to do with power?
<libv> mallah: you disconnected power, but you had vcc connected on the uart?
<libv> mallah: was this on a cubieboard or was this on this device?
<mallah> yes i had powered the device from other computers' usb port becoz i had forgotten to carry the AC adaptor to office.
<mallah> it was on this device.
<mallah> cubieboard warns was ON connecting the VCC !
<libv> mallah: how did you power it, just used the +5V line on the adapter and tied it to the power plug?
<libv> mallah: usb can only deliver .5A from a host device
<mallah> the device was safe but the pl2303 of the cubie cable got fried.
<libv> how did you wire up things? Vcc on the usb module to vcc of the uart of the device?
vincenzo has joined #linux-sunxi
<mallah> true but i guess the 2.5W are sufficient unless u ar doing some heavy wt.
<mallah> i had connected both , powered the device from the DC port ie 5V port from the host port of one computer and ALSO connected the UART to yet another computer ie my laptop.
<mallah> i think it was risky
vincenzo has quit [Client Quit]
<mallah> and my fault.
<libv> i am still not entirely clear on what you did.
<mallah> probably we can remove that note.
<libv> mallah: you used 2 power connections now?
<mallah> thats rite.
<libv> mallah: one just doing host usb -> the device DC in
<mallah> not now but on the day my pl2303 got fried.
<libv> mallah: and then you also connected VCC from the usb-uart to VCC on the device's uart port?
<mallah> rite now i am using a DC adator that came with the device. and its working well.
<libv> or did the pl2303 get fried without that VCC connected?
<mallah> yes thats rite :)
<mallah> i am not sure here.
<libv> hrm. 5V from my cp2102 module seems to go straight to the usb port
* libv gets a multimeter out
<mallah> hey libv i think it was my folly to get things fried leave it.
<libv> mallah: we all are quite interested in this, as we warn against connecting VCC, but we have no confirmed cases of this damaging anything
<mallah> ok i will try to replicate the damage ....
<libv> no, don't
<mallah> when i catch hold of a pl2303 based adaptor
<mallah> ok
<libv> no need to burn yet another one
<libv> so when did it break?
<mallah> hmmmm 3 days back
<libv> when you tried to power up the board, or did it break when you disconnected the +5V from the other usb port?
<libv> cp2102 has +5V wired directly to the usb connector
<libv> 3.3V comes from the cp2102 though
<mallah> i cannot recall the exact sequence but for sure the cable was fried. as syslog stopped reporting the pl2303 on jack in.
Fusing2 has joined #linux-sunxi
<libv> a host computer reboot didn't change anything? (my usb stack sometimes plays up here)
Fusing has quit [Ping timeout: 255 seconds]
<libv> or did you fry the usb port on the host?
<mallah> sorry for late i was putting the kernel on the card.
<mallah> hmmm noooo
<mallah> that port is the one which is feeding me the ooops.
<libv> ok
<mallah> CFLAGS_ehci-sunxi.o := -DDEBUG
<mallah> drivers/usb/host/Makefile
<libv> anyway, interesting, i will try to remember to document that on our UART page
<mallah> is the above correcct way to make the kernel spit pr_debug s?
<libv> we had someone ask yesterday btw
<mallah> ok
akaizen has quit [Remote host closed the connection]
<libv> hrm, something already assigned these io pins?
<libv> arokux2 might be to blame :p
akaizen has joined #linux-sunxi
sehraf has quit [Ping timeout: 245 seconds]
<mallah> who is hrm?
sehraf has joined #linux-sunxi
<libv> it's not a name or a word, it's a sound one tends to make
<libv> nobody else seems to be using these io pins from simply looking at the .fex
<libv> PD02 and PD03...
* libv checks the datasheet
<mallah> just now i saw that the GPIO support in the kernel was disabled.
<libv> mallah: ?
<mallah> i enabled it NOW and recompiling
<mallah> yes
<libv> mallah: i do not know what you mean with gpio support?
<mallah> wait
<libv> is that sunxi specific?
petr has quit [Quit: http://quassel-irc.org - Chat comfortably. Anywhere.]
<libv> anyway, i was going to suggest to disable usbc0 and usbc2
<libv> from the fex
<mallah> Symbol: GPIO_SUNXI
petr has joined #linux-sunxi
<libv> mallah: that's probably a driver for playing with gpios from userspace
<mallah> ok
setkeh has quit [Ping timeout: 240 seconds]
<mallah> lemme disable usbc0 and usbc2 as u suggest
<libv> anyway, disable both, and see if it boots then, then re-enable usbc0 again and see if it boots then
<libv> disable means, set usb_used = 0
<mallah> shall i just flip usb_used = 1 or totally delete the section
<mallah> ok
setkeh has joined #linux-sunxi
<mallah> u just answered
<libv> ah!
<libv> lcd0_para
<libv> set lcd_used = 0
<mallah> what
<mallah> ok
<libv> but try the previous two as well
petr has quit [Remote host closed the connection]
<libv> pd02/pd02 are usually data lines for the first directly attached lcd
<mallah> sorry it still paniced after disabling the usbc0 and usbc2
<libv> right
<libv> now disable the lcd
petr has joined #linux-sunxi
<mallah> k
<libv> hrm, strange, disabling both should mean that you wouldn't request a gpio line anymore
* libv reads deeper
<mallah> sorry
<mallah> LR is at sw_ehci_hcd_probe+0xe8/0x2e0
<mallah> there was still a kernel oops.
<mallah> shall i also disable usbc1
<mallah> when i swap the usbc0,usbc1 and usbc2 of this fex with the corresponding section of olimex-a20-micro fex kernel oops disappear but the ports does not work either :(
<libv> aha, could it be that our code ignores the used flag?
<mallah> :)
<mallah> shall i make the sections disappear ?
<mallah> and try again.
akaizen has quit [Remote host closed the connection]
<libv> no, that will not change a thing
<libv> or...
<mallah> no it did not
Black_Horseman has joined #linux-sunxi
Black_Horseman has quit [Changing host]
Black_Horseman has joined #linux-sunxi
bertrik has joined #linux-sunxi
leviathanch2 has joined #linux-sunxi
<mallah> with above it will boot.
<libv> right
<libv> interesting
<bertrik> I'm planning to work on a A31 boot0 replacement
leviathanch2 has quit [Ping timeout: 240 seconds]
<mallah> EUREKA!!
<mallah> i do see power in one of the usb ports
<mallah> my mouse is glowing
<libv> mallah: what did you change?
<mallah> i changed usbc0 , c1 and c2
<mallah> the oops was gone
<mallah> dmesg | grep -i -E 'ehci|ohci|usb'
<libv> how did you change them?
<mallah> i just lifted from the olimex-A20-micro
<mallah> i donot understand much
<libv> so the changes that you pasted above?
<mallah> how does the messages look
<mallah> yep http://pastebin.com/raw.php?i=AG5mk9DG is the live config now
<libv> that's pretty worrying
<libv> how did it work before?
* libv doesn't know enough though
<mallah> it never worked in linux
<mallah> it worked only on android.
<libv> but i do know that this code now seems to ignore the usb_used flag
<mallah> i mean still works on android
<libv> mallah: very strange
<mallah> ok
<libv> one of the guys with more usb knowledge really should look into this
<mallah> do u want me to bisect
<libv> no, this smells like a job for a multimeter
<mallah> the real change that is making or breaking
<mallah> :)
<libv> a multimeter and a tool for setting gpios individually
<libv> so that you can determine exactly which gpios are wired to what
<mallah> ok
<mallah> u mean which gpio pins of the A20 feeds the USB port ?
<libv> yes
<libv> so that you can hand tune your config there.
<libv> because this wholescale copying is kind of scary stuff
<mallah> is the possible to RE from the android running instance?
<libv> mallah: you do not need to
geecko has joined #linux-sunxi
<mallah> how does the messages looks like on http://pastebin.com/raw.php?i=QCjMLV3Q
<libv> this would need some peering into the a20-olinuxino schematic as well
<mallah> I also have the correspoinding messages of the android kernel
<libv> mallah: can you load the wifi module?
<libv> because that usually has a usbc of its own as well
<mallah> lemme chekc,
paulk-aldrin has quit [Quit: Ex-Chat]
<libv> mallah: when you have the wifi working, line by line, comment out (with ;) the usbc gpio lines, and see what the basic configuration that you need is
deffrag_ has joined #linux-sunxi
deffrag has quit [Ping timeout: 255 seconds]
tomboy65 has joined #linux-sunxi
tomboy64 has quit [Ping timeout: 245 seconds]
<mallah> i had builtin the RTL8188EU i see no sign of it in dmesg
akaizen has joined #linux-sunxi
FreezingCold has joined #linux-sunxi
<mallah> now building icplus phy
<mallah> libv: i will take leave now
<mallah> its 3:00am morning here
<libv> goodnight
<mallah> which part of world r u in?
jemk has quit [Remote host closed the connection]
Xaros has joined #linux-sunxi
Black_Horseman has quit [Ping timeout: 268 seconds]
<mallah> Bye
mallah has left #linux-sunxi [#linux-sunxi]
geecko has quit [Ping timeout: 255 seconds]
TheSeven has quit [Ping timeout: 255 seconds]
TheSeven has joined #linux-sunxi
geecko has joined #linux-sunxi
ZetaNeta has quit [Quit: Leaving]
xavia has joined #linux-sunxi
kivutar has quit [Ping timeout: 240 seconds]
Wizzup has joined #linux-sunxi
sehraf has quit [Quit: www.miranda-fusion.de ... be part of it...]
xavia has quit [Quit: Leaving.]
akaizen has quit [Remote host closed the connection]
wickwire has quit [Remote host closed the connection]
kivutar has joined #linux-sunxi
bertrik has quit [Quit: Lost terminal]
akaizen has joined #linux-sunxi
netlynx has quit [Quit: Leaving]
peterbjornx has quit [Quit: Lost terminal]
setkeh has quit [Remote host closed the connection]
Fusing has joined #linux-sunxi
setkeh has joined #linux-sunxi
Fusing2 has quit [Ping timeout: 268 seconds]
Fusing has quit [Ping timeout: 252 seconds]
setkeh1 has joined #linux-sunxi
setkeh has quit [Ping timeout: 240 seconds]
setkeh has joined #linux-sunxi
kivutar has quit [Quit: Ex-Chat]
deffrag_ has quit [Ping timeout: 268 seconds]