<Montjoie>
ccaione, now it is time to use a conf engine for gain time, cfengine for example
<ccaione>
Montjoie: I'm a gentoo-noob, so yeah I have a lot to learn. Thank you for the advice :)
<Montjoie>
cfengine is for all OS, not only gentoo
MAT__ has joined #linux-sunxi
<ccaione>
didn't know
<MAT__>
Hi everyone, is there a way that I can compile the GPIO module for sun7i ?
MAT__ is now known as Guest79723
dlan_ has joined #linux-sunxi
mkutsevol has joined #linux-sunxi
mkutsevol has left #linux-sunxi [#linux-sunxi]
ysculo has joined #linux-sunxi
<ysculo>
hello
bbrezillon has joined #linux-sunxi
codekipper_ has joined #linux-sunxi
<ysculo>
nice talk at fosdem :) I realized at was not alone going mad with allwinner source code ...
NermaN has joined #linux-sunxi
<NermaN>
hi all =)
<NermaN>
how to set mac address for EMAC in device tree?
<nabblet>
I have a question about http://linux-sunxi.org/Linux_mainlining_effort The things you state under "work in progress": are those already implemented in your kernel and just need to get into main line - or are the work in progress as in "we are still trying to get it to work"
gerald_bostock is now known as lisca
<NermaN>
nabblet: some of things are waiting for mainline
<ysculo>
my boss heard about E200 processor and Malis 0S. He wants that because it's lowlowlowcost. And I'm freaking out. Does anyone knows about this OS ? I want to kill the idea before it spreads. Especially if there is no linux support on those processors as fallback.
mkutsevol has joined #linux-sunxi
<nabblet>
NermaN: so they are already working in the sunxi kernel?
<NermaN>
nabblet: SATA and NAND are working, and now as i know they waiting for mainline
<nabblet>
doh, stupid question
<nabblet>
NermaN: yeah, thank you :) Hand to read your first reply a second time XD
<alexvf>
ssvb: yesterday i checked the DEBE diagram in the a20 user manual
<alexvf>
ssvb: could i use the sprite controller to blend the UI over the video+subtitles?
<alexvf>
ssvb: i don't know what is the sprite controller or how it works
Guest79723 has quit [Quit: Page closed]
<nabblet>
NermaN: because i was wondering: olimex are basing their debian on your kernel. They offer sd cards with Debian pre installed for A20 and I wasn't sure if that will get all the hardware A20 has to run properly
<alexvf>
mmm, ssvb sleeps too :)
<alexvf>
anyone knows how can i blend 3 layers on the a20 display (UI + subtitles + video)?
<alexvf>
i learned that i can't blend 3 layers
<alexvf>
but there should be a way, maybe using the sprite controller?
notmart has joined #linux-sunxi
mkutsevol has left #linux-sunxi [#linux-sunxi]
FreezingCold has quit [Ping timeout: 260 seconds]
sehraf has joined #linux-sunxi
t3st3r has quit [Ping timeout: 240 seconds]
t3st3r has joined #linux-sunxi
sspiff has joined #linux-sunxi
sspiff has joined #linux-sunxi
NermaN has quit [Ping timeout: 245 seconds]
enrico_ has joined #linux-sunxi
arete74 has quit [Read error: Connection reset by peer]
geecko has joined #linux-sunxi
w00tc0d3 has joined #linux-sunxi
w00tc0d3 is now known as Guest13568
arete74 has joined #linux-sunxi
mkutsevol has joined #linux-sunxi
linkmauve1 has joined #linux-sunxi
fralle_ has quit [Remote host closed the connection]
Seppoz has joined #linux-sunxi
KEYTRONIC has joined #linux-sunxi
chkktri has joined #linux-sunxi
<chkktri>
hello again, just to make sure, are LDM volumes (windows dynamic disks) supported in 3.4 kernel?
FR^2 has joined #linux-sunxi
rellla2 has joined #linux-sunxi
rellla has quit [Ping timeout: 252 seconds]
gzamboni has quit [Remote host closed the connection]
JohnDoe71rus has joined #linux-sunxi
JohnDoe_71Rus has quit [Read error: Connection reset by peer]
gzamboni has joined #linux-sunxi
hramrach has quit [Remote host closed the connection]
hawi has joined #linux-sunxi
hramrach has joined #linux-sunxi
FunkyPenguin has quit [Remote host closed the connection]
<Nyuutwo>
but that crap sells to people as entry level linux board
<Turl>
libv: I was being ironic :p 56$
<libv>
ah, ok, kind of thought that it was awkward, but didn't think it through :)
<Turl>
libv: you can get a CB1 for that price :p
<libv>
people should get a lime instead
<Turl>
libv: if you're not in europe it costs a bit more though
<Turl>
but yeah, it's nicer if you can live with half the ram
<Turl>
eh
<Turl>
if you are on europe it costs even more heh
<Turl>
damn VAT :)
<libv>
i don't mind paying taxes, tbh. i like having roads and cyclelanes and a metro
<libv>
and in europe, the espionage and defense budgets are """reasonable""" ;)
<Turl>
libv: but don't you pay customs?
<Turl>
or is it free if you pay VAT?
<libv>
when importing from outside the european union, yes
<Turl>
ah
<Turl>
makes sense then
<libv>
again, also necessary, it's the missed taxes, and it allows a tiny bit of control, like the issue we now have with overly cheap solar panels from china
<libv>
(which killed several european companies already)
<speakman>
What's "surface size" when running the Mali test app?
<libv>
speakman: you mean, from the output on the console?
<speakman>
Surface size: 480x480
<speakman>
GL Vendor: "ARM"
<speakman>
GL Renderer: "Mali-400 MP"
<speakman>
GL Version: "OpenGL ES 2.0"
<libv>
width x height, in pixels?
<speakman>
Running the "test" binary included in sunxi-mali repo.
<speakman>
I'm having problems with Qt5.2 having problems with creating surface. A wild guess it may have to do with that specified surface size. The resolution on the LCD is 1024x768.
<libv>
no
<speakman>
EGL Error : Could not create the egl surface: error = 0x300b
<ssvb>
speakman: but I'm going to roll out a more foolproof 'solution' for the mali400 binary drivers soonish
<speakman>
ssvb: I've been trying that one. The patch might got lost while rebuilding. I'll have another look at it. Thanks!
<speakman>
ssvb: I've gone from Qt5.1 to Qt5.2, that's why I lost it.
<ssvb>
speakman: just been a bit busy during the daytime and lazy in the evenings/weekends lately
<speakman>
ssvb: Let me know if I can be to any help with the foolproof solution. :)
<chkktri>
Turl: sadly this didn't work for me, getting "Error scanning /dev/sda: Error reading from /dev/sda: Invalid argument"
<alexvf>
any piece of documentation/sample about how to use sprite controller in a20?
<speakman>
Huge improvements! But now this: "EGLFS: OpenGL windows cannot be mixed with others."
<speakman>
Seems to be an issue with Qt5 example I'm testing
<alexvf>
ssvb: do you know anything about sprite controller?
<ssvb>
alexvf: not really, haven't used them myself, but libv should know something because sprites had been mentioned in his FOSDEM talk
<alexvf>
ssvb: thanks, i'm searching for the talk in google :)
<libv>
alexvf: why would you need sprites?
<alexvf>
libv: it's just a crazy idea, maybe it has no sense
<libv>
alexvf: that sentence does not explain why you would need sprites
<alexvf>
libv: but, since i read that sprites are alpha blended over the result of video+osd
<alexvf>
libv: i'm trying to use them to draw an UI on top of all
<libv>
alexvf: we have 4 layers
<alexvf>
libv: what i want to do is to draw an UI over mplayer
<libv>
1 is used for the fb
<Turl>
chkktri: maybe something missing from kernel config?
<libv>
and they can be alpha blended two and two
<Turl>
chkktri: do you have support for GPT enabled?
<libv>
video doesn't often need to be alpha blended
<libv>
so you should get away with displaying that in the pair with the framebuffer
<libv>
then you have 2 further layers to play with, one of which you can alpha blend with the pair with the fb and the video
<alexvf>
libv: yes, but i also want subtitles over the video
<alexvf>
libv: i'm currently using the (non finished i think) osd branch of libvdpau-sunxi
<ssvb>
libv: libvdpau-sunxi uses up the layers but alexvf still wants to alpha blend something on top of it
<alexvf>
yep
<libv>
it uses them all up already?
<libv>
because it is using the second pair for video, perhaps?
<libv>
the first pair could be used for framebuffer and video
<alexvf>
libv: i don't quite understand how do you blend two by two
<libv>
then you have the second pair at your disposal
<alexvf>
libv: by software?
<libv>
alexvf: this is just how the hw works
<ssvb>
libv: libvdpau-sunxi uses up the "Alpha Blender 1"
<libv>
ssvb: alpha video?
<ssvb>
libv: osd and subtitles
<libv>
there are only limited uses for that
<libv>
well, alexvf, give it a go then
<libv>
but be careful, the sprites will be used as RGB cursors in kms
<libv>
as kms is a useful piece of shit badly reinvented twice by people who never did any display driver development before
<ssvb>
I'm afraid that kms might be a leaky abstraction, forcing us to forfeit some of the display controller features and go for the lowest common denominator
NermaN2718 has joined #linux-sunxi
<alexvf>
libv: do you think my idea can work?
<alexvf>
my problem now is the lack of documentation/samples on how to use the sprites
<libv>
alexvf: it's all there.
<alexvf>
i only have the disp ioctls, which have little documentation
<libv>
there are 4 of them, i do not recall any ordering
<libv>
each sprite can be set to a power of two width up to 1024, and power of two height up to 512
<libv>
with freely settable stride
<ssvb>
alexvf: the a20 user manual, "5.4.2 Display Engine Block Diagram", the registers description and also linux-sunxi sources should be enough
<libv>
and they support some basic colour formats
<ssvb>
libv: sounds like these sprites are a very powerful feature
<alexvf>
do you think i need to mess with registers?
<libv>
ssvb: yes, and no
<alexvf>
aren't ioctls enough?
<libv>
if they have ordering, then they could be
<libv>
but only when kms could expose them properly.
<libv>
which it can't
<alexvf>
i would prefer not to mess with registers as i have never worked so close to hardware
<libv>
kms planes are an afterthought, by yet another person who never did overlays or display development before
<libv>
and this afterthought only exists to make wayland more competitive with the android hwcomposer
<libv>
before that time, i, once again, was pretty much the only person who had done overlays for a decade
<ssvb>
alexvf: the sprites should be already exposed via disp ioctls
<alexvf>
ssvb: yes, they are but my first tests are not working
<libv>
alexvf: arbitrary stride means that you should be able to have 2 sprites next to eachother, showing different parts of the frame bit of framebuffer
<alexvf>
ssvb: just trying to request a sprite block
<libv>
or even get 2kx1k if you use all 4
<ssvb>
when I was implementing the hardware cursor feature for fbturbo, I did not know that the display controller documentation even existed :) so I just used the kernel sources to figure out everything important
<alexvf>
ssvb: so it could get reimplemented to take advantage of sprites at some moment?
<alexvf>
libv: i read on the a20 manual that up to 32 sprites could be used
<ssvb>
alexvf: yeah, the palette reduction for ARGB32 cursors might be unnecessary now
<alexvf>
libv: so you say only 4 are actually available?
<ssvb>
alexvf: just give it a try and report back ;)
<alexvf>
i'm going to experiment a little and hopefully see something on screen, which i am unable to see right now :)
<alexvf>
yes, i'll report any progress
<alexvf>
thank you ssvb and libv
<NermaN2718>
is there info about cpufreq on mainline? =)
Seppoz has quit [Remote host closed the connection]
Seppoz has joined #linux-sunxi
<speakman>
Does G2D has to be declared in script.bin or is it on by default when enabled in kernel?
<libv>
speakman: why do you ask?
orly_owl has quit [Ping timeout: 265 seconds]
<Turl>
NermaN2718: there's no cpufreq support on mainline
<Turl>
(for sunxi that is)
orly_owl has joined #linux-sunxi
<Turl>
NermaN2718: I have a forward ported version of the one on 3.0/3.4 sitting on my git tree if you really need it
<Turl>
speakman, libv: there's a g2d_para section on script.bin with _used/_size
<speakman>
I've finally been able to test some Qt5 examples, but although it's all on accelerated GPU right now, it still seem anything but smooth. :/
<speakman>
I'm currently running "collidingmice" example, and they're all having "choppy" movements. Like it's 10 fps or something.
<Turl>
speakman: cpufreq?
<Turl>
have you adjusted it to performance?
<Turl>
speakman: echo performance|sudo tee /sys/devices/system/cpu/cpu0/cpufreq/scaling_governor
<speakman>
Oh! Didn't know about it. :D
<speakman>
it's at 50% CPU exactly while running collidingmice. I guess that's one core at 100% constantly.
<speakman>
Just cpu0? There is cpu1 as well.
<speakman>
HUGE difference!!!!
<Turl>
do it on both if they're there
<Turl>
they may be tied to each other already though
<wens>
Turl: did you see my gmac clock patch?
FR^2 has quit [Quit: Connection reset by peer]
<Turl>
wens: probably :)
<speakman>
Turl: yep, they're tied to eachother
<Turl>
wens: I still haven't got my network back in place so I cannot easily test stuff
<ssvb>
speakman: what kind of hardware are you using?
<speakman>
Anything else which might speed up the A20 a bit? Most demos are still slow. :/
<Turl>
wens: so I haven't started taking patches
<Turl>
wens: but I'll do next week at the latest
<Turl>
wens: the new gear is on the way :)
<Turl>
speakman: are you using fbturbo as X driver?
<Turl>
or are you running this with framebuffer mali?
<ssvb>
speakman: anyway, set the 'performance' cpufreq governor first, to get better performance and also more realistic cpu usage statistics
<wens>
Turl: I see. good luck :)
<speakman>
ssvb: That's what I've done.
<ssvb>
Turl: speakman is using the framebuffer mali driver, so no X11 there
<speakman>
Turl: nope, pure framebuffer/eglfs
<Turl>
oh ok then :)
<speakman>
Will there be even better performance with X11 and fbturbo?
<ssvb>
speakman: still 100% cpu usage on one core? 'htop' is a useful tool
* speakman
is also doing all this through Buildroot, which is a bit tricky compared to having a full fledged debian based system running.
<speakman>
ssvb: Yep, the collidingmice is still 50% constantly
<speakman>
Unfortenately I have to leave the office now. I'll scroll through the backlog tomorrow (UTC+2) if there is anything. Thanks A LOT so far guys!!
<ssvb>
speakman: no, with X11 there will be a few tens percents extra overhead, and that is assuming that everything is configured right
<ssvb>
speakman: try some other non-cpu hungry demo :)
<speakman>
ssvb: Will do first thing tomorrow :D
<speakman>
Btw, why isn't the Qt5.2 createNativeWindow patch posted upstreams?
<ssvb>
speakman: I guess that's because it is the mali400 framebuffer drivers bug/misfeature
<ssvb>
speakman: the other framebuffer egl/gles drivers allow to use EGL_DEFAULT_DISPLAY and NULL as the native window to get 'reasonable' behavior (fullscreen framebuffer use)
<ssvb>
speakman: the mali framebuffer drivers are not compatible with this assumption
kivutar has joined #linux-sunxi
<ccaione>
Turl: give him a call
<Turl>
ccaione: I don't have free calls to china numbers unfortunately
sehraf has quit [Ping timeout: 245 seconds]
shineworld has joined #linux-sunxi
fredy has quit [Excess Flood]
fredy has joined #linux-sunxi
rellla2 has quit [Remote host closed the connection]
hawi has joined #linux-sunxi
arete74 has quit [Ping timeout: 245 seconds]
sehraf has joined #linux-sunxi
<binaryferret>
Hi guys, I'm hoping someone can help me as I've been banging my head against a wall for days now. I'm trying to get a touchscreen working. The touch screen is gslx680. I'm using the driver source from https://gitorious.org/gslx680-for-sunxi . via dmesg etc it seems to insmod fine, and detects the chip - but I'm not getting any response from touching. I tried evtest and it detects the gslx680 event but again when I go to test it, it d
<binaryferret>
Here is the fex file http://pastebin.com/pf8CkG6u I think it has something to do with the wakeup and that the touchscreen is just not being made active. However, I'm still learning so request your help to debug the issue.
<binaryferret>
I should note, that it's for an Allwinner A13 tablet.
<binaryferret>
I'm running linux-sunxi kernel from the 3.4 branch (normal branch).
bsdfox has quit [Ping timeout: 245 seconds]
bertrik has joined #linux-sunxi
shineworld has left #linux-sunxi ["Leaving"]
arete74 has joined #linux-sunxi
setkeh has quit [Read error: Operation timed out]
setkeh has joined #linux-sunxi
popolon has quit [Quit: Quitte]
netlynx has joined #linux-sunxi
geecko has quit [Ping timeout: 245 seconds]
bsdfox has joined #linux-sunxi
bsdfox has quit [Ping timeout: 248 seconds]
<alexvf>
binaryferret: have you checked evtest output?
bbrezillon has quit [Quit: Ex-Chat]
<binaryferret>
alexvf: Yes alexvf once I select the input for evtest (glsx680) it displays osme information and then any trouches don't generate any events.
tomboy64 has quit [Read error: Connection reset by peer]
<alexvf>
ssvb: the sprite test almost worked
<alexvf>
ssvb: i could paint a semi-transparent rectangle over the video+subtitles thing
<alexvf>
libv: actually it seems that each sprite can be set to a power of two width up to 512, and power of two height up to 1024 (not the other way around)
<alexvf>
libv: and up to 32 sprites can be allocated, although i could not get more than one working
<alexvf>
libv: due to some memory bug that i have :)
gzamboni has quit [Ping timeout: 245 seconds]
<binaryferret>
alas no :(
gzamboni has joined #linux-sunxi
pwhalen has quit [Read error: Operation timed out]
<ssvb>
alexvf: very nice, it would be also interesting to check if multiple sprites can be overlapped and blended with each other
<alexvf>
yep, it is what i'm trying to do :)
enrico_ has quit [Quit: Bye]
_massi_ has quit [Quit: Leaving]
Nyuutwo has quit [Ping timeout: 245 seconds]
pwhalen has joined #linux-sunxi
pwhalen has quit [Changing host]
pwhalen has joined #linux-sunxi
diego_r has quit [Ping timeout: 248 seconds]
fredy has quit [Excess Flood]
fredy has joined #linux-sunxi
Nyuutwo has joined #linux-sunxi
rm has quit [Ping timeout: 264 seconds]
notmart has quit [Quit: notmart terminated!]
rm has joined #linux-sunxi
<binaryferret>
I'll forward the question on the mailing list.
nabblet has joined #linux-sunxi
hawi has joined #linux-sunxi
rellla has joined #linux-sunxi
fralle has joined #linux-sunxi
<fralle>
somebody know which keys or gpios witch will put the a10 into fel mode?
<fralle>
docs are talking about keys... guessing this are gpios? but which ones?
techn_ has joined #linux-sunxi
<oliv3r>
montjoie[home]: i prefer puppet over cfengine; but to be fair, i never used either!
<Turl>
binaryferret: check if you get interrupts when you touch in /proc/interrupts
<Turl>
fralle: usually it's a button dance you need to do
<Turl>
ie, some combination with some timing involved
<Turl>
the bootloader checks for it and puts the device in fel mode
<Turl>
fralle: the easiest way though would be to make a fel card
<fralle>
thnx turl
<fralle>
but im trying the contrary
<fralle>
my card stops in boot1
<fralle>
key
<fralle>
key found going into fel
<fralle>
same img into cubieboard runs
<Turl>
ah, so you want it not to go into fel?
<fralle>
yeah!
<Turl>
hmm
<fralle>
cubieboard boots
<fralle>
with same img
<Turl>
maybe check cubieboard schematics?
<Turl>
see what pin they wired
<fralle>
fel pin?
<fralle>
but thats not the one :)
<fralle>
seems it looks for more than that pin...
<fralle>
but how to configure which ones...
<codekipper>
exit
<Nyuutwo>
which pin is fel pin?
<fralle>
let me see
codekipper has quit [Quit: leaving]
<Turl>
fralle: do you have LRADC keys on board?
<Turl>
bbl
<fralle>
nop
<fralle>
:)
<fralle>
UBOOT_SEL is fel pin
<Turl>
fralle: usually on tablet you need to press a key on lradc to enter fel, maybe if you have it floating it is reading a key?
<Turl>
I dunno on cubieboard
<Nyuutwo>
on cubie there is button on other side of board than USB otg connector
FreezingCold has joined #linux-sunxi
<Nyuutwo>
is there a pio connected to this pin, or just tabletmakers connect it to other pio used for button?
<fralle>
ok ok
<fralle>
ill check
<fralle>
can u tell boot1 through fex file to skip LRADC?
<binaryferret>
Turl: Sorry for the wait on a reply was away from pooter. I'll check out interrupts.
<binaryferret>
Cheers.
<binaryferret>
Not sure what I'm looking at, so I might go away and have a gander and see what the output says but here is the output http://pastebin.com/d52ptba9
<binaryferret>
what the output means*
AreaScout has quit []
tomboy65 has quit [Ping timeout: 252 seconds]
<fralle>
Turl, the lradc is it supposed to be pulled up?
geecko has quit [Ping timeout: 264 seconds]
tomboy65 has joined #linux-sunxi
<Turl>
fralle: I don't know really
<Turl>
fralle: it just occured to me that it may be that
<Turl>
Tsvetan: maybe oliv3r or Tsvetan know
<Turl>
err fralle ^
<Turl>
binaryferret: and I bet the gslx680 number remains static if you don't do anything and grows when you touch?
<Turl>
binaryferret: seems to be fine interrupt wise
<binaryferret>
Turl: I'll just do some more tests to confirm that.
<fralle>
i can see its pulled up in the cubieboard
<binaryferret>
Turl: Weirdly, gslx680 interrupts are growing when I keep running cat /proc/interrupts
<fralle>
mine is down...
<Turl>
binaryferret: anything touching the screen or near it?
<Turl>
it's odd that it interrupts "just because"
<binaryferret>
Without touching anything.
<Turl>
binaryferret: are you sure the fex config is correct?
<binaryferret>
Turl: It's the fex file that came with the device and I took it from Android nanda. Now, with that said the ctp_name was different, so I edited the fex file. The fex file referred to ekt4ft (or something like that). When insmod the module in android it doesn't seem to use the script.bin
jemk has joined #linux-sunxi
<binaryferret>
so I imagine that the fex file could actually be incorrect.
<Turl>
binaryferret: do you have a copy of the module handy?
<Turl>
the original android one
<binaryferret>
Which one, the one from android or the one I built?
<Turl>
^
<binaryferret>
I can do in a moment if you don't mind waiting 5 mins.
<Turl>
I'm in no hurry :)
<jemk>
Turl: the dram gate names are in the wrong order, they need to be the other way around. But otherwise it works :)
<libv>
Turl: done :)
<Turl>
jemk: I listed them the wrong way? oops :(
<Turl>
jemk: thanks for the heads up, I'll swap them around
<Turl>
jemk: just the names right? the mask is fine isn't it?
<libv>
and these devices are ones that customers sent back
<libv>
under german law, you get like a month where you can send your online or mail bought stuff back and get your money back, without having to give any reason
<binaryferret>
Turl: Not sure if it matters, but this http://binaryferret.com/tmp/gslx680.fw is the firmware file that is taken from the original module, and used with the built module.
<Turl>
binaryferret: try removing lines 163-166 and ctp_io_port from your fex
<libv>
oliv3r: true
<binaryferret>
Ok will do that now.
<oliv3r>
still 30E is a nice price
<libv>
oliv3r: but tbh, you must have a few remotes lying about already
<libv>
oliv3r: surely you have like a universal remote somewhere
<kivutar>
anybody here with a cubieboard2 and a serial to usb willing to boot an image and tell me the kernel error? I'm trying to port openelec
<libv>
somewhere in some drawer, with batteries which are flowing out already
<Turl>
libv: and if he doesn't, they're cheap
<Turl>
:p
<Turl>
oliv3r: ouch oliv3r :(
<Nyuutwo>
kivutar: I have on desk
<Turl>
odd that they run their forum on :81 btw
Seppoz has quit [Remote host closed the connection]
<libv>
kivutar: please spend 5$currency on a uart ;p
<libv>
well, uart-to-usb thingie
<binaryferret>
Turl: Just to check when you say remove lined, are you talking about the original one from android, or the script.fex file I posted to pastebin?
Seppoz has joined #linux-sunxi
<Turl>
binaryferret: the original one from android doesn't have it
<Nyuutwo>
kivutar: like, when you recompile it doesn't show you log what was done
<Nyuutwo>
but you can force to do it
<Nyuutwo>
or nonremoving sources of compiled packages which just eat space by default
<nove>
Turl, Enrico announced here, is the right path but i think is a bit too early, better to wait for when the driver starts to get in shape
<kivutar>
wow
<Turl>
nove: oh, good then :)
<kivutar>
what's that bitbake thing
<Turl>
isn't that what yocto et al use?
<kivutar>
derived from portage.. ok
<kivutar>
is it C?
<Nyuutwo>
kivutar: python/bash heap of kludgecoding to leash crosscompiling of packages which developers didn't thought of doing that
<Nyuutwo>
kivutar: more like emerge
<Turl>
kivutar: python acording to wiki
<Nyuutwo>
bb files are like portage
<kivutar>
wow python...
<binaryferret>
Turl: I'm going to have to rebuild the kernel before hand, as I've just realised that the tab I'm using I built at work. I'm at home so when I build the driver and now try and insmod it, it moans about versions.
<kivutar>
its even worse than cmake
<nove>
Nyuutwo: the hardware appears to be capable of high profile encoding, appears because the result wasn't tested to see if it really uses high profile features
<Turl>
binaryferret: check uname, you can probably find the hash used to build there and check that out with git
<Nyuutwo>
kivutar: cmake doing rootfs would be much worse than bitbake
<Nyuutwo>
i once tried use it on htc wizard but it was so slow
<Nyuutwo>
and some android ui
<Nyuutwo>
on tablet i plan use archlinux
<binaryferret>
Hehe alas not. Rebuild it is.
paulk-collins has quit [Quit: Ex-Chat]
<Nyuutwo>
but I need axp support to powerdown it
<Nyuutwo>
and some suspend would be nice
Guest76017 is now known as Wizzup
geecko has joined #linux-sunxi
FreezingCold has quit [Ping timeout: 272 seconds]
<kivutar>
oe looks tentaculous
<Nyuutwo>
kivutar: ??
<kivutar>
Nyuutwo, i'm giving a try to yocto
<kivutar>
it is almost the same right?
<Nyuutwo>
kivutar: not so same
<kivutar>
oh
<Nyuutwo>
oe is a system of building and it provides some basic packages
<Nyuutwo>
yocto (pocky) is a distro based on oe
<Nyuutwo>
afair it is reference distro
<kivutar>
mmmh
<Nyuutwo>
but yocto says about itself it is not a distro but tool to create your custom one
<Nyuutwo>
and oe/yocto ecosystem just develops wild
<Nyuutwo>
trying to upgrade from one release to another it is like to bump version of kernel to your kernel modules
<Nyuutwo>
it just sometime breaks
<kivutar>
well it is the same with openelec
bertrik has quit [Remote host closed the connection]
kuldeepdhaka has quit [Ping timeout: 252 seconds]
FreezingCold has joined #linux-sunxi
<binaryferret>
Turl: Ok, rebuilt (issue with 8188eu wifi driver now but will look into that later) and I get this output http://pastebin.com/ZP6QFrEd
hawi has quit [Quit: Leaving]
<Turl>
binaryferret: does it work?
nove has quit [Quit: nove]
cubear has quit [Quit: Leaving]
<binaryferret>
Alas no. Still no events on evtest :(
<Turl>
binaryferret: what about the interrupts now?
<binaryferret>
Turl: Still growing when not touching screen.
<Turl>
binaryferret: hrm
<Turl>
binaryferret: do you have/can you install i2cdetect?
Nyuutwo has quit [Ping timeout: 245 seconds]
<binaryferret>
Yep alreayd installed
<binaryferret>
on bus i2c-1 I get a 27 at 0x27, UU at 0x40 and a UU at 0x51
<Turl>
binaryferret: what's UU?
<binaryferret>
Not sure. It's just the charactrers that are displayed. I'll google and find out
<Turl>
ah, in use by driver
<binaryferret>
Yep that's the one.
<Turl>
0x51 is your rtc
Nyuutwo has joined #linux-sunxi
<binaryferret>
Ok.
<Turl>
binaryferret: iirc there was a flag to force the scan, can you see if there's anything on 40 really?
<Turl>
0x40 that is
<binaryferret>
Okdokey. So uninstall the mod and then do i2cdetect?
<Turl>
that should work too
<binaryferret>
ok on i2c 1 it at 0x40 it shows 40
Nyuutwo has quit [Quit: No Ping reply in 180 seconds.]
Nyuutwo has joined #linux-sunxi
<Turl>
binaryferret: if you run ./pio -m 'port:PB03<1><default><default><1>'
<Turl>
does it disappear?
<binaryferret>
I don't have pio. I'll have to get that.
<binaryferret>
Well I have it, but not on the device.
<binaryferret>
Turl: Sorry to sound like a newb, but I get a usage message (what commands to input) when I run pio. I can't seem to correct the above.
<Turl>
binaryferret: did you pass -m?
<Turl>
binaryferret: maybe it doesn't take the port:, I don't have the tool here to double chek
<Turl>
check*
<binaryferret>
I typed ./pio -m 'port:PB03<1><default><default><1>'
<Turl>
try without port:
<binaryferret>
Ok I ran it without the word port: and it didn't complain. I then did i2cdetect 1 and on address 0x40 I got 40.
<Turl>
any new numbers?
<binaryferret>
Nope.
<Turl>
binaryferret: now rerun pio, but change the first 1 to a 0
<Turl>
and try agai
<Turl>
n
<binaryferret>
Okdokey. Still the same numbers.
pwhalen has quit [Ping timeout: 252 seconds]
<Turl>
ahh sorry
<Turl>
it's the last 1, not the first that you need to change
FreezingCold has quit [Ping timeout: 248 seconds]
<Turl>
keep the first 1 as 1
<binaryferret>
Ok, so we sent 0 as the data? part of the pin config?
<binaryferret>
0x40 no longer contains anything.
rah has quit [Ping timeout: 272 seconds]
rah has joined #linux-sunxi
<Turl>
ok, so 0 shuts down the chip
<Turl>
1 wakes it up
<binaryferret>
Ahhhhh I see.
<Turl>
binaryferret: now load your driver and do ./pio -m PB03
<Turl>
make sure it's at 1
<binaryferret>
ok so wake up chip, then install mod.
<Turl>
the module should wake up the chip itself
<Turl>
install it and confirm it's awake using pio
<Turl>
./pio -m PB03
<Turl>
should print the pin config
<binaryferret>
Okdokey PB3<1><0><1><1>
<Turl>
that's fine
<Turl>
:/
<Turl>
dunno then
<Turl>
have you visually confirmed that's the chip on the tablet?
<binaryferret>
I haven't visually. I've just manually uninstalled and installed the android driver while android is running, and read that the gslx680 will be on i2c 1 address 0x40.
<binaryferret>
However I'll open the backup and get out my loop.
pwhalen has joined #linux-sunxi
<Turl>
binaryferret: usually that testing is enough
<Turl>
I don't see anything going wrong though, sorry :(
<Turl>
the continuous interrupts are a bit odd however
<Turl>
maybe the driver author has some other idea
FreezingCold has joined #linux-sunxi
<binaryferret>
No worries matey. Thank you for all the assistance. At least I've learnt some new ways to debug.
<Turl>
although a userspace driver is kind of weird :)
<binaryferret>
That was my next port of call. However I wanted to exhaust as many avenues in this before I went skipping off to try other things. I tend to sometimes get excited and start jumping around areas a little too often, so I'm being good and ticking of things as I go along :D