<deasy>
i have also see if sd plugged and power removed, then remove the sd with only uart plugged
<xtofury>
ok getting the same error
<deasy>
the power red led is on with the remove of the sd card
<hno>
deasy, just unplug the RX pin of the UART cable.
<deasy>
thank for the tips :)
<hno>
(RX pin on the board side)
<xtofury>
thanks so much for helping me get through this, I've been running around in circles for 2 weeks and have ignored this for another 2 weeks because my patience ran out...
<hno>
xtofury, open a new terminal window, then cd sunxi-bsp/u-boot-sunxi; make CROSS_COMPILE=$HOME/gcc-linaro-arm-linux-gnueabihf-4.7-2013.04-20130415_linux/bin/arm-linux-gnueabihf- Cubieboard
<deasy>
xtofury, you have probably do a mistake somewhere
<deasy>
i need to do it too
<woprr>
Cloning into u-boot-sunxi...
<woprr>
github.com[0: 204.232.175.90]: errno=Connection timed out
<deasy>
i want the last code :)
<woprr>
giving up
<navlrac>
github is overloaded? error: Failed connect to github.com:443; Operation now in progress while accessing
<navlrac>
hmm, working now
<deasy>
github is evil hehe
<hno>
navlrac, for u-boot use sunxi-current branch.
<deasy>
command not found is like you have not set well the path
<deasy>
i do it soon...if something in the wiki is wrong i correct it
<hno>
xtofury, ls -l $HOME/gcc-linaro-arm-linux-gnueabihf-4.7-2013.04-20130415_linux/bin/
<hno>
and also ls -l $HOME/gcc-linaro-arm-linux-gnueabihf-4.7-2013.04-20130415_linux/
<hno>
something very fundamentally wrong here...
<deasy>
it's an obligation to use ubuntu or it can works on others distro?
* hno
never use Ubuntu.
<deasy>
me too
<xtofury>
oh I'll try that one second
<hno>
Hmm... one thing.. the downloaded toolchain is 32-bit, maybe your ubuntu is 64-bit? It then need some additional packages to provide 32-bit compatibility.
<hno>
don't remember exacly what those packages are called on ubuntu.
<xtofury>
that could be it, the instructs said for me to grab amd 64
<xtofury>
I'd assumed since it said to grab that version of ubuntu, that the rest of the instructs were written for that.
<hno>
there is no mention of what OS you should use in FirstSteps. Doesn't need to, instructions is mostly generic (except a little apt part that have sneaked in in suxi-bsp part)
<xtofury>
now I can't even find where it said for me to grab amd 64 but I can see why that could be a problem... (it might have been in some other instructs, not on that first steps page, I am trying to find it where it said that)
<hno>
doesn't matter. If your Ubuntu is a 64-bit clean install then the errors makes a lot sense.
<hno>
xtofury, apt-get install ia32-libs
<hno>
might still work.
<hno>
(old method, but I think a compatibility meta package is still around to keep things simple)
<xtofury>
it looks like its downloading
<hno>
you'll get a bit more 32-bit libraries than you need, but it's a bit of a mess to find which ones you really need.
<hno>
with libc there is should start saying better error messages than "File not found" if ther eis more libraries needed.
<xtofury>
its installing a bunch of dependencies and structures
<hno>
all normal.
<xtofury>
ok, thats done installing
<xtofury>
error while loading shared object file: no such file or directory
<hno>
which one?
<xtofury>
libz.so.1
<hno>
I guess that becomes libz1:i386 in ubuntu speak.. but see page I linked before.
<hno>
which seems to say that it's libz1g:i386 have no idea what the g is about, completely senseless to me.
<jelly-home>
zlib1g
<deasy>
don't forget the i386
<hno>
right.. that lib is not renamed to lib... name syntax like all other libs.. and suddenly have a letter after the so version. Very consistent naming standard.
<jelly-home>
the package name is historical and there's little reason to rename things for consistency
<xtofury>
it seems to be working its flooding the terminal with a bunch of data
<deasy>
i do not love the naming convention on fedora, they have upper character
<deasy>
like in the original names of apps i think
<hno>
deasy, yes, Fedora uses original names where possible.
<deasy>
i prefer all lowercase
<jelly-home>
if one needs "libz.so.1", one uses apt-file to find it on Debian
<xtofury>
after using the
<xtofury>
make mele_a1000 CROSS_COMPILE=arm-linux-gnueabihf-
<hno>
which means some packages have as stupid names as the thing they package.
<deasy>
VirtualBox haha
<deasy>
i find it annoying
<hno>
xtofury, you don't really want a mele_a1000 u-boot.
egbert has joined #linux-sunxi
<xtofury>
why not?
<deasy>
you have not a cubieboard a10?
<xtofury>
yes
<xtofury>
cubieboard a10
<deasy>
so why compil with board mele ?
<deasy>
compile*
<hno>
deasy, VirtualBox is not in fedora. But sure we have some other packages with capital letters.
<xtofury>
just following instructions, what should I put in that line then?
<hno>
VirtualBox is not allowed in Fedora because it requires custom kernel modules which they have not cared to mainline.
<deasy>
hno, oh ok but the name of the package is VirtualBox, i need to change of distro because i really dislike it
<hno>
There is no VirtualBox package in Fedora.
<xtofury>
I'd assumed mele_a1000 was referencing a10 and mele_a2000 the a20... do I need to get rid of those extra 0's or something?
<hno>
xtofury, mele_a1000 is the Mele A1000 multimedia "PC".
<deasy>
a short time ago was the OSE version
<deasy>
but now it seems than only the original is available on rpmfusion
egbert_ has quit [Ping timeout: 264 seconds]
<hno>
deasy, what do you mean by OSE version?
<deasy>
Open Source Edition
<deasy>
without usb part
<woprr>
I've just stolen the uimage from the A20 archlinux image, was the smallest, but does not boot
<woprr>
Image Name: Linux-3.3.0-1-ARCH
<hno>
deasy, VirtualBox project have split the usb part from VirtualBox.
<xtofury>
ok so then what am I putting in there instead?
<woprr>
checking if github up again
<hno>
xtofury, yes.
<hno>
deasy, and VirtualBox name is really VirtualBox, not virtualbox. Trademark lawyers would frown upon you and throw you to jail :)
<deasy>
ow trademark :(
<deasy>
ok seriously? ok i understand
<deasy>
someone have say me than tvbgone is also a trademark...bad bad
<hno>
Thats surely why they call the source repository vbox
<deasy>
xtofury, you must use cubie or if the command can show you the list of boards
<deasy>
cubieboard*
<deasy>
xtofury, linux-sunxi is community website for allwinners cpu support
<deasy>
they have information about several boards allwinner based
<deasy>
the tutorial is not specific to cubieboard
rz2k has quit []
<deasy>
so there can be somes parameters to change
<deasy>
maybe YOURBOARD can be a better thing to show :)
<xtofury>
I don't understand why nobody has just uploaded a vbox image with this stuff already done :(
<hno>
oh, trademark wonders.. Oracle do not allow other distributions of VirtualBox than their own to go under the name VirtualBox. If this ever is going into Fedora it will needto use another name,,,
<xtofury>
aha cubieboard works
<deasy>
yes
<xtofury>
so I replaced melee_a1000 with "cubieboard"
<hno>
xtofury, because no one here likes huge binary images.
<deasy>
xtofury, if you get the experience you can do a proper one
<xtofury>
so not this: make cubieboard CROSS_COMPILE=arm-linux-gnueabihf-
<hno>
and it will be obsolete in 3 months.
<hno>
or maybe 6 if lucky
<deasy>
xtofury, not understand your last sentence
<xtofury>
the #boardmodel thing, is that referring to the cubieboard part in there or do I have to add it to the end, this is vague.... or is that just there to tell me that if I got a20 to do make cubieboard2 instead of make cubieboard
<xtofury>
ok yes, nvm
<xtofury>
I get it
<xtofury>
slow today
<arokux>
hno, any idea how pinctrl-0 works?
<hno>
arokux, what do you mean?
* hno
for sure knows how the PIO controller works...
* deasy
is registered on the wiki
<deasy>
wouhou
<arokux>
hno, in device node, you can assign a group of pins to it. you first define this group.
<hno>
arokux, DT syntax question?
<arokux>
hno, no.. "how it works"-question
<arokux>
hno, :)
<hno>
what works?
<hno>
need a little more detailed question.
<hno>
PIO as such do not care how you group pins.
<hno>
to PIO it is just a bunch of pins connections in the chip.
<arokux>
hno, so my question is: how those assigned pins are used in the code?
<hno>
deasy, you are only an "Autoconfirmed User" from what I can see. Is it maybe your IP that have been blocked?
<arokux>
hno, 2 hours of grepping and i cannot understand a *shit*
<deasy>
so i have not try to add about the device to change like it's write in the top
<deasy>
hno, maybe yes i think the message received from my contribution talk about ip, i can retry if you need
<hno>
deasy, seeing the actual error would help..
<hno>
arokux, pin names is in the code I think.
<hno>
the & in the first is a reference to the other, which lists which pinmuxer and pins to set up on it.
naobsd has quit [Quit: Page closed]
<arokux>
hno, yes, this is clear.
<arokux>
hno, but how to get to those pins in the code to poke at them?
<hno>
arokux, you don't poke at pinmux pins normally.
<xtofury>
ok almost at the sdcard setup, but now I am going to run into the raw access with vbox and I have no spare sdcard thing for use with usb to setup so I guess I do this next time...isn't there a way I can skip the sd card bit and just flash it onto my cubie like I do with livesuit?
<arokux>
hno, I need to enable the pin.
<hno>
which pin?
<deasy>
hno, We have doubts if you are a spambot or not, please <a href="/IRC">join us on IRC</a> to sort this issue out.
<hno>
Hm.. maybe you need to be approved to have write access.
<arokux>
deasy, have you answered some question? like if allwinners produces chips or smth else?
<arokux>
hno, the pin PH6 should be enabled in the ohci driver
<hno>
deasy, try now.
<hno>
arokux, why?
<arokux>
hno, because chinese code does so, VBUS power, remember?
<arokux>
hno, maybe it should be directly enabled in device tree?
<hno>
arokux, but that's not pinmuxing, thats gpio. You need it listed twice. Once for the pinmuxing setting, and then in another attribute definint it as the gpio needed for vbus control.
<hno>
or actually, listing it as the gpio might be sufficient, that should bring in the needed pinmuxing as well from gpio.
<hno>
but others know this much better than me. But I am sure that you DO NOT define that in the pinmuxing part.
<arokux>
hno, so you say I use pinctrl only if my pins are muxed for device?
<deasy>
arokux, yes it have ask
<deasy>
and i have answer CPUs like it have ask
<hno>
arete74, pinctrl is for pinmux control, configuring the pin to the right hardware function.
<arokux>
hno, hm... seems to be exactly what you are saying
<hno>
arokux, that makes sense to me. One pinmux section wiring up the pins, and one section defining their GPIO usage.
<hno>
just would think there is some smartness so you don't really need all that pinmux baggae when also referencing the same pin via a gpio controller, but maybe you do.
<xtofury>
how long does this take to compile?
<arokux>
xtofury, "this"?
<hno>
same gpio may have multiple different pin mappings, even if sunxi chips only have one possible physical pin per gpio pin.
<xtofury>
oh on this step : "make ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf- -j5 uImage modules"
<xtofury>
been about 10-15 minutes so far into it.
<deasy>
i have been marked as spammer? maybe it's my ip if i have dos a website of a scriptkiddies really annoying...
<arokux>
do you have 4 cores?
<hno>
xtofury, some minutes to an hour depending on what processing powers you have... but based on the fact tat you can run VirtualBox under 10 minutes.
<hno>
deasy, have you tried again?
<xtofury>
2 cores :(
<arokux>
then -j3
<deasy>
hno, nop
<arokux>
xtofury, not -j5
<hno>
deasy, I asked you to...
<xtofury>
ok what is this j5 thing for then?
<xtofury>
just remove it?
<arokux>
xtofury, it is trying to spawn 4 threads from compilation and one master thread for control
<hno>
I would use -j4
<arokux>
xtofury, do ^C and restart with proper -j
<xtofury>
ok restarting with -j4 option....
<deasy>
hno, works
<hno>
having slightly too large -j does not make much bad, provided there is sufficient RAM in the box.
<arokux>
xtofury, you are compiling s-3.4?
<hno>
deasy, approved you as a person.
<arokux>
hno, won't the OS keep scheduling threads imposing context-switch overhead?
<deasy>
xtofury, i have added the fact this page is focus on mele a1000 in stronger characters, hope it helps people to don't pass on it
<hno>
we also have a bureaucrat level, need to check what that means..
<arokux>
deasy, now you are a person :)
<hno>
arokux, Linux is quite good at identifying batch jobs such as compilers.
<xtofury>
when I am done I am going to pop cubie specific instructs on my blog.....
<arokux>
xtofury, and on our wiki too, plz
<hno>
to be exact I added you to the group "people".
<xtofury>
ok can do once I get this done... I figure this is a good way to n00b proof things since I'm starting out on this, was gunna use a pi but the specs are too good to pass up on the cubie. I'm actually getting interested in this cubietruck2 coming out but not sure if it's worth the wait for an A30 offering....
<arokux>
xtofury, what are you up to? how are you going to use the board?
<xtofury>
home automation controller with attached arduino and touchscreen -- specific use is for a greenhouse but I want to make it useable for multiple purposes.
<xtofury>
I wanted the arduino integration to provide expanded GPIO with 5v rather than 3.3v, there's more add-ons specific for arduino, plus it'll be easier to plop the RTC on the arduino since with cubie it is a definite issue.
<hno>
arokux, no, you need to check pio for what the pin definitions means.
<arokux>
xtofury, in my opinion RPi is nothing more than big advertisement... for almost the same money you could get much better perf
<arokux>
xtofury, have you checked olimex devices and their ext modules?
<xtofury>
no but now that you mention it I might give it a look... For now though I'm trying to get arduinocommander running on this with the screen in portrait mode, w/ RTC, w/resistive touch working (will have to find a capacitive one later, for this'll do, on a budget)...I'll be happy with it running in kiosk mode... still seems a ways away, gotta do something with USB to get arduino to work
<xtofury>
correctly for arduinocommander.
<xtofury>
ok all done :) thanks guys
<xtofury>
sdcard is going to have to wait till tommorow
<arokux>
hno, 7 stays for H :)
<arokux>
arokux, good night hno. thanks for help.
<arokux>
hno, tomorrow I'll enable this gpio..
<xtofury>
thanks everyone, night. help has been really appreciated
<arokux>
xtofury, night :)
xtofury has quit []
<hno>
arokux, yes, just figured the same.
<arokux>
hno, but this is the guess
<hno>
night, or actually morning..
<arokux>
hno, how to know which function gets called in pio, i.e. drivers/pinctrl/pinctrl-sunxi.c
<deasy>
nite as nightmare?
<deasy>
nice*
<deasy>
almost morning
<hno>
arokux, Documentation/gpio.txt
<arokux>
hno, and? not dev. tree info there.
<arokux>
gpios = <&pio 7 20 0>; /* LED2 */
<arokux>
pio is pinctrl-sunxi.c
<arokux>
hno, my question is where 7, 20 and 0 are passed to
ganbold_ has joined #linux-sunxi
TheSeven has quit [Disconnected by services]
[7] has joined #linux-sunxi
<woprr>
the archlinux A20 kernel wont boot on olimex micro with olimex or cubie uboot or I got a "disc geometry" problem with dd here...
<woprr>
to sdcard #1: 238+1 Datensätze aus, from sdcard #2 (olimex): 238+0 Datensätze aus with count=238 ???
<zerodamage>
(unrelated) resistive touchscreen works out of box?
<Seppoz>
there is no out of the box with touchscreens
<arokux>
deasy, yes
<Seppoz>
zerodamage: check if the driver is enabled and if it has edev support, if so you should be able to use it in x11 or whatever, if not look into tslib
eebrah|away is now known as eebrah
<Seppoz>
arokux: have you been using gpio before?
<Seppoz>
with export?
<arokux>
you mean out of sysfs?
<Seppoz>
yes
<Seppoz>
on the a10 i mean
<Seppoz>
because i dont see gpiochips
PaceyIV_ has joined #linux-sunxi
<Seppoz>
is that ok?
<arokux>
Seppoz, i do not know nothing about 3.4
<arokux>
sorry
<Seppoz>
how is it i mainline?
<Seppoz>
should be the same if there is sysfs gpio support aint it?
<zerodamage>
a10disp utility output: http://pastebin.com/Q3aDrgwL see what i mean with the crazy framebuffer size WTFF
<Seppoz>
whats the cat giving you?
<zerodamage>
sec
<arokux>
Seppoz, but i try to use gpios now in ml
<zerodamage>
U:65536x65536p-0
<zerodamage>
which as you guess, is not usable nor right :P
<Seppoz>
thats wired
<Seppoz>
with the default script.bin i get 800x480
<Seppoz>
did you just tried using the default cubieboard fex?
<zerodamage>
yes, but the timings are not right for my display unfortunatly, i can try it though to see if it give usable modes
<Seppoz>
even if the timings are incorrect you will see something on the screen
<Seppoz>
allso i think that this setting is for ttl not lvds
<PaceyIV_>
Hy! I've got a Olimex A20 board. I'm using kernel and boot scripts from official debian release 2. I create a custom uEnv.txt where I set root path to my sata hard disk.
<PaceyIV_>
When kernel try to boot from /dev/sda1 seams that sda doesn't exist. Without my uEnv.txt sda is recognized after init, when kernel populate /dev
<PaceyIV_>
How can i load rootfs from my sata hard disk, instead of micro SD?
<Seppoz>
PaceyIV_: mybe your eEnv is not loading script. bin and thus sata is not enabled ot so?
<Seppoz>
PaceyIV_: mybe your eEnv is not loading script. bin and thus sata is not enabled or so?
<zerodamage>
yeah now it just cycle-reboots, need to restore using other labtop, be back later
rellla has quit [Remote host closed the connection]
<hno>
Seppoz, I am not sure how much you can specify the i2c bus in script.bin.
<Seppoz>
also the ata one
<Seppoz>
its in device drivers Serial ANA and Parallel ATA drivers
<Seppoz>
hno: if i check the touchscreen stuff its more like support for some ready to go drivers
<PaceyIV_>
linux.sunxy doen't provide support for A20... olimex make the debian realease with allwinner sdk adding some missing support like sata and ethernet wired...
<PaceyIV_>
i wrong?
<ganbold_>
hno: thanks
<arokux>
hno, hi
<arokux>
hno, I need some help..
<arokux>
hno, do you have any idea how to test if the power is on after I write to gpio?
<ganbold_>
arokux: measure the voltage?
<arokux>
ganbold_, in software..
<mripard_>
arokux: yep, not for very long :)
<mripard_>
I'm in holiday for ~2 weeks, so I only have a limited internet access
<arokux>
mripard_, cool! plz, explain how to use gpio!
<mripard_>
but you shouldn't be able to export it at all, since your driver requested it.
<Seppoz>
that might also help a bit
<arokux>
mripard_, I freed it, so I could. did it for testing
<Seppoz>
its for s3c but probably the same for this
<arokux>
mripard_, so you think that I actually did what i wanted? turned it into OUT and make it output 1?
<Seppoz>
isnt there gpio_direction_output
<mripard_>
arokux: ah, right. I'm not exactly sure about the side-effects gpio_free has, but yeah, your code looks fine.
<mripard_>
the only way I'm thinking of to test that is actually to check it with a multimeter on the line connected to that gpio
<Seppoz>
why would you free it?
<arokux>
mripard_, :) ok. I'm adding usb host support. the hub appeared already. not I want to see something happens if i plug in flash drive
<arokux>
Seppoz, to be able to use it from sysfs
<Seppoz>
then export it right away
<Seppoz>
gpio_export
<mripard_>
arokux: using the musb driver ?
<arokux>
arokux, no, normal usb host, usb1 and usb2
<arokux>
mripard_, musb is usb0
<mripard_>
ok
<arokux>
mripard_, hm.. if i'll insert some device into the usb port, it should start to charge, if the power is there, or?
<mripard_>
I must admit I never looked too much at usb, the USB OHCI and EHCI specificatiosn also specify the register layout and all? or is it another shared IP ?
<hno>
arokux, connect your phone to the port, or anything else that indicates if there is power.
<arokux>
mripard_, yes, I hope so. we just need to get it up, the rest is done by drivers/usb/host/ohci-ucd.c
<hno>
arokux, using an USB lamp of fan if you have one is alsy very obvious indication of power.
<arokux>
hno, yes, also got this idea.
<arokux>
hno, no power...
<hno>
is the GPIO set?
<arokux>
hno, how do i know which port corresponds to which controller?
<hno>
arokux, by looking in schematics..
<hno>
it is possible none of them are directly connected to the controller. Might be a HUB inbetween.
<arokux>
hno, we do not have mele schematics.
<hno>
arokux, what board are you using=
<mripard_>
arokux: ok
<arokux>
hno, mele a1000
<ganbold_>
arokux: connect usb stick that has led inside
<arokux>
ganbold_, i'm poking with a phone already :)
<mripard_>
maybe you can enable the debug of the usb stack, to get what's going on
<hno>
The mele have all ports behind a hub I think. but not 100% sure. Maybe one of the host ports have direct connection.
<hno>
did map them out before, but have forgot.. might have put it in the wiki.
<arokux>
hno, if i poke that gpio from uboot, the power should be on? I don't need the usb setup
<hno>
Not sure what happens when there is a hub.
<arokux>
hno, oh, ok
<arokux>
hno, ok, no power from under uboot too.
<mripard_>
I guess the gpio is just to power up the PHY, maybe not the devices connected to it
<hno>
The GPIO is for powering VBUS. The PHY is internal.
<mripard_>
ok
<arokux>
vim drivers/usb/host/sw_hci_sunxi.c +449
<hno>
The mele have 5 USB ports, 3 external and 2 internal. The A10 have two USB controllers. What we know for sure is that the hub connects to one of the controllers and provides 4 ports.
<hno>
What we do not know is if the wifi is on the other controller or if it's connected on the hub.
<hno>
should be easy to look this up from a 3.4 kernel.
<hno>
or even the default android.
rellla has joined #linux-sunxi
PaceyIV_ has quit [Quit: Page closed]
<arokux>
hno, I think i need a diode.........
Black_Horseman has joined #linux-sunxi
<hno>
arokux, diode for?
<arokux>
hno, usb2serial
<hno>
right. haven't had that problem on mele. But haven't tried in a long while as my mele is also powered via the same usb connetion these days.
<hno>
so I naver have UART without power on the mele.
<woprr>
now, first e-cigarette then back to work... kernel, uboot, sata, etc
<woprr>
hi hn o
<hno>
h no would be more appropriate spelling.
<woprr>
dont't want to highlight you without important reason ...
<hno>
don't worry.
<arokux>
hno, which diode will you recommend
<arokux>
?
<ganbold_>
arokux: why diode?
<woprr>
I need a good toolchain or a unstable chroot for emdebian on oldstable :D
<ganbold_>
arokux: if you can have usb flash with led inside
<arokux>
ganbold_, usb2serial gives power to my mele a1000 board and this confuses it
<ganbold_>
arokux: then you can see whether power is coming or not
<arokux>
ganbold_, yep, no power, but this is different issue
<hno>
ganbold_, having the UART port connected with no power to the AXP controller makes most Allwinner boards go crazy.
<arokux>
mripard_, still there?
<hno>
quite likely also applies to HDMI port btw.
<woprr>
kernel on my oldstable is squeeze bpo / vanilla 3.2.x, so chroot may work
<woprr>
reading wiki first... first steps, etc
<arokux>
mripard_, why would I want to specify pinctrl-0 in the device none when after this i still need to say that the gpio shoud be out, when requesting it. I've thought pin function is specified in pinctrl-0 already?
<hno>
woprr, the linaro 4.7 toolchain work on any distribution if you want an easy path.
<woprr>
k, thx
<hno>
woprr, just keep in mind that it's 32-bit binaries so you need 32-bit libraries installed (libc + libstdc++ + zlib)
<arokux>
hno, connected withOUT or with power makes them crazy?
<woprr>
easy not necessary, but low on HD space on my "compile engine" /dev/sda4 45G 20G 23G 47% /mnt/data
<hno>
arokux, without.
<woprr>
ok
<arokux>
hno, but my problem is because usb2serial provides power, or?
<hno>
arokux, without any power to the AXP controller makes the AXP controller go crazy due to the current leakage from the UART.
<woprr>
arokux, a self powered usbserial cable? for what is that for?
<arokux>
woprr, usb2serial is connected to the PC, and it powers it, from what i understand
<woprr>
all my are bus powered and some are crap, not all RS232 pins are used, only one works with my UPS
<hno>
woprr, no we are talking about hardware design issues in 99.9% of all Allwinner devices if you have an UART cable connected.
<woprr>
ah ok
<arokux>
hno, ok.. so how can I avoid this problem (without soldering)
<woprr>
arokux, that's what lsusb -v says "bus powered"
naobsd has joined #linux-sunxi
<hno>
arokux, provide power to the board, the nconnect UART cable.
<arokux>
hno, i'm confused. my problem there is current leakage from uart and it is cured by providing even more current?
* woprr
uart cable is handmade, max3232 and 5 caps, a sub D plug and 4 PIN bar plug, thats it, 15min work
<arokux>
hno, sorry..
<arokux>
woprr, you mean I can check if the bus is powered by lsusb -v?
<woprr>
ask hn o ^^
<arokux>
woprr, that was your message!
<woprr>
no you can see how a device is powered
<woprr>
usually on x86 but may not be implemented here?
<Seppoz>
lsusb shows you how devices are connected and whats on the bus
<hno>
lsusb shows how a device claims it is powered. Says nothing about how it is actually powered.
<hno>
and is irrelevant to todays topic.
<Seppoz>
what are you guys trying to find out anways
<woprr>
sorry
<arokux>
hno, you said i need a diode to prevent current leakage?
<hno>
arokux, yes, if having the UART cable connected before you connect the power cable is causing problems for you.
<Seppoz>
just drop GND of the uart and you should be fine aswell
<mnemoc>
o_O
<Seppoz>
if we talking about a ttl levelshifter conntected to tx tx and gnd of an uart
<arokux>
mnemoc, hi :)
<mnemoc>
arokux: hi :)
<Seppoz>
arokux: what are you trying to do?
<arokux>
mnemoc, i'm adding usb host (1 and 2) but as i'm a noob i need to test every step, are you into usb stuff?
<woprr>
gnueabihf-4.7 is minimum required toolchain version?
<Seppoz>
woprr: just use the arm-linux-gnueabi-gcc from universe
<mnemoc>
arokux: need to urgently finish a $milestone$, sorry :(
<woprr>
Seppoz, not on ubuntu , debian oldstable here
<Seppoz>
oh well just update those packages then?
<Seppoz>
or just use the one you asked
<arokux>
Seppoz, if i have usb2serial connected to my board it confuses it. maybe it could influence the usb stuff, that i'm working on..
<Seppoz>
oh i see
<Seppoz>
maybe use ssh instead?
_hipboi_ has joined #linux-sunxi
wickwire has joined #linux-sunxi
hipboi_ has quit [Ping timeout: 245 seconds]
<wickwire>
Hi guys, I'd like to use a USB webcam with A10 (sunxi kernel), any special requirements? The webcam has MJPEG, will I be able to take advange of it?
<arokux>
wickwire, have u tried it?
<wickwire>
I'm at it right now
<wickwire>
I plugged the camera in, I'm getting lsbusb results I think
<wickwire>
I'm following a guide originally for the Raspberry Pi, which worked for this camera as well
<woprr>
eh.. wildcards in PATH do not work in debian oldstable bash http://linux-sunxi.org/FirstSteps#For_developers export PATH="$PATH":/home/user/folder/gcc-linaro-arm-linux-gnueabihf-*_linux/bin/
<woprr>
no olimex A20 defconfig in # On branch stage/sunxi-3.4, use sun7i_defconfig ?
<woprr>
Already up-to-date.
<Seppoz>
A20 is sun7i
<Seppoz>
so sun7i defconfig should be fine yes
<woprr>
ack
<woprr>
olimex supplied debian sdcard uboot identifies with machid(?) EOMA86-A20
<woprr>
near match omap1,p2 ...
<woprr>
using sun7i
<Seppoz>
i think the sunxi current branch has support for the olimex boards
<Seppoz>
the uboot
<woprr>
yes , make a20-olinuxino_micro CROSS_COMPILE=arm-linux-gnueabihf- succeeded
<Seppoz>
ok
<Seppoz>
and you know how to flash it to the card right?
<woprr>
first build kernel
<Seppoz>
ok
<woprr>
if the wiki is up to date?
<woprr>
topics, first steps,debian, building on debiab
<woprr>
more?
<Seppoz>
?
<Seppoz>
sorr yi dont understand the question
<arokux>
hno ?
vinifr has quit [Quit: Saindo]
<woprr>
never mind
<woprr>
why is FPU emulation needed, A20 has got an FPU
<woprr>
in sun7i_defconfig ?
<hno>
arokux, what?
<arokux>
hno, nothing already. I decided to clean up old 3.4 tree and gradually move usb host to mainline
<hno>
ok
<arokux>
hno, hope dropping of unused code is welcome
<woprr>
enable the ugly gpio driver as M?
<hno>
arokux, sure. Killing unused code. refactoring used code so it makes sense, all allowed.
<woprr>
A20 needs CONFIG_SATA_AHCI_PLATFORM=Y or is CONFIG_SW_SATA_AHCI_PLATFORM=Y sufficent for use with SATA?
<woprr>
do not know "softwinner"
dwilkins has joined #linux-sunxi
<ssvb>
woprr: there are some legacy vfp instructions not supported by vfpv4 anymore, and it may be reasonable to keep the vfp emulation enabled in the kernel just in case
<woprr>
OK
<woprr>
enabling both sata drivers, not mutually exclusive
<mripard_>
wingrime: everything I had for the A20 is ready to push mainline, and actually, it's the cubieboard2 dt has been sent today.
<wingrime>
mripard_: nice
<wingrime>
mripard_: about IR0 clk....
<mripard_>
if I didn't submit everything yet, it's just to avoid a maintainer overflow :)
<wingrime>
mripard_: driver is ready but it use hack
<mripard_>
about the clock, hmm, no idea what it's about, let me look at it
<wingrime>
mripard_: I pushed to linux-sunxi for RFC
<wingrime>
Turl: ^
<Turl>
arokux: yes, you can poke directly, but there should be no need. make sure you've got the latest patches from maxime, the pinctrl bug had quite a few bugs that might be getting to you
<arokux>
Turl, oh..
<arokux>
Turl, I just want to poke directly so that I know this is the problem. how do I do this?
<mripard_>
wingrime: it looks like it the exact same clock than on the A10
<mripard_>
so if Turl come up with a clock for it, you can use it on the A20 as well.
<Turl>
arokux: read the user manual and look for the pin bank stuff
* hramrach
emerges from the woods
<Turl>
mripard_: I just need to write bindings, user manual in hand :)
<wingrime>
Turl: mripard_ I want get rid hacks soon as possible
<mripard_>
the only difference seems to be the clock sources it uses, but that's not a big deal, just a few tweaks in the DT
<Turl>
mripard_: wingrime discovered the SATA IP is from synopsis :) should be good to go with a bit of glue he says
<mripard_>
wingrime: then, either do it yourself or see if Turl can do/did it
<Turl>
mripard_: I'll do it before I send v2 of my series
<Turl>
need to do so for sun5i too
<wingrime>
Turl: aw sata driver are 500 lines
<mripard_>
because I won't have time to hack this kind of stuff for something like two weeks
<mripard_>
Turl: yep, he told me, that's great :)
<wingrime>
Turl: I need only normaly port it to DT
<wingrime>
Turl: sata use own dma ,
<arokux>
mripard_, another stupid question. I set interrupts = <N> in device tree. then I do irq = platform_get_irq(pdev, 0);, but I get different number from this function. it's ok?
<Turl>
wingrime: you got good dma consumer to test now
<wingrime>
Turl: no
<wingrime>
Turl: it have OWN dma
<Seppoz>
arokux: there is a chip called SY6280
<Turl>
wingrime: :/
<Turl>
two dma?
<arokux>
Seppoz, and?
<Seppoz>
arokux: this chip is able to interrupt the 5V power supply and does the current detection
<wingrime>
Turl: we have bunch dma
<wingrime>
Turl: mmc have
<Seppoz>
the USBDRV is enabling or disabling this chip
<wingrime>
Turl: cedar have
<mripard_>
arokux: yes, the number you set in the DT is the hardware number (ie, the one given in the datasheet)
<Seppoz>
when its disabled the 5V is gone
<mripard_>
the one you get by platform_get_irq is virtual
<Seppoz>
if you dont meassure 5V its most likely that this pin is off or there is too much power taken
<arokux>
mripard_, and the one I get is some kind of handle?
<arokux>
mripard_, ok, thanks
servili007 has joined #linux-sunxi
vinifr has joined #linux-sunxi
<wingrime>
Turl: for sata I need PLL6 support
<Seppoz>
arokux: can you pastbin me the driver please? with coloring
<Turl>
wingrime: PLL6 will be there, don't worry
<arokux>
Seppoz, it won't help. it's to messy atm
<Seppoz>
just set all 3 pins manually to 1 on probing the driver
<arokux>
mripard_, does the settings here: http://pastebin.com/yvz2xibr help in anyway? if cat /sys/class/gpio/gpio230/direction it give me "in", shouldn't it be out?
<vinifr>
dwilkins, hi
<Seppoz>
IMO the driver should do that allready
<wingrime>
mripard_: I can't find yours a20 patches in gmane
<mripard_>
arokux: default value is in
<wingrime>
mripard_: only some pinctl fixes
<Seppoz>
arete74: maybe there is something that suspends the USB port?
<Seppoz>
arokux: ^^^^
<arokux>
mripard_, so the setting function=gpio_out doesn't do anything?
<wingrime>
Turl: mripard_ I claim sata
<arokux>
Seppoz, maybe-maybe, who knows.
<wingrime>
arokux: also, check cb scheme , axp can be related with usb
<wingrime>
take multimiter and measure power on (17) pin AXP
<wingrime>
with linux-sunxi and mainline
<arokux>
wingrime, no multimeter here.....
<arokux>
wingrime, maybe you have mele a1000?
<wingrime>
arokux: no
<arokux>
hno can u measure please? -----^
<mripard_>
arokux: doing kernel dev on a board without schematics and without proper tools, especially on something as tricky as USB, isn't probably a good idea :S
<mripard_>
maybe if you are interested, you can do easier stuff to begin with.
<arokux>
mripard_, ok.. didn't know...
<wingrime>
arokux: do you have buttons on mele?
<arokux>
yes 1
<mripard_>
maybe working on the keypad or the ADC driver depending where it's wired to would be a good idea then
<mripard_>
the IPs look simple enough
<arokux>
mripard_, that seemed to be easy. a bunch of clocks and pokes at different registers
<wingrime>
arokux: only powerbutton
<wingrime>
>
<wingrime>
?
<arokux>
wingrime, yes
<wingrime>
arokux: bad,
<wingrime>
arokux: if there some other buttons like volume
<arokux>
wingrime, no
<wingrime>
arokux: you can possible debug lradc
<arokux>
wingrime, what's that?
<wingrime>
arokux: low resoluton alnalog digital converter
<wingrime>
arokux: used for buttons on tables
<arokux>
mripard_, if I do keypad or ADC, how do I test them? :)
<wingrime>
arokux: vol+ vol-
<arokux>
wingrime, ok, no buttons here, as I said
<wingrime>
arokux: keypad=lrdac
<arokux>
wingrime, there is remote control only :)
<mripard_>
arokux: I thought you had :)
<wingrime>
arokux: you can test my driver
<wingrime>
arokux: for IR
<mripard_>
it will probably be trickier without any button connected to it :)
<arokux>
wingrime, with pleasure. tell me what to do.
<mripard_>
yeah, testing wingrime's driver looks like a good idea
<wingrime>
arokux: wait when trul push all stuff
<mripard_>
anyway, off for tonight
<wingrime>
arokux: hacks still there
<arokux>
wingrime, ok
<arokux>
mripard_, thanks a lot
<mripard_>
have a good morning/afternoon/evening/night :)
<wingrime>
arokux: also LRADC should be not difficult
<arokux>
wingrime, but I won't be able to test it
<wingrime>
arokux: more difficult , but not requires any HW - cryptoengine
<wingrime>
arokux: you can try implement hardware Random generator
<arokux>
wingrime, to tell the truth I want usb.. and wifi.. :(
<wingrime>
arokux: there is no way, but try and try again blind
<wingrime>
arokux: thats way also works
<wingrime>
arokux: but need more time
<wingrime>
arokux: first things that you can try
<wingrime>
arokux: check GATE and CLOCKS set right way
<wingrime>
arokux: reading usb regs
<wingrime>
arokux: if you setuped all things correct
<wingrime>
arokux: you must read registers non zeroed
<arokux>
wingrime, ok
<wingrime>
arokux: otherwice you will get read only zeros
<wingrime>
arokux: and second check, write 0xffffffff to it and readback
<Seppoz>
are we armel or armhf?
<wingrime>
arokux: this will help figure Is ip configured right way
<wingrime>
arokux: also a20 manual now have registers for usb
<arokux>
wingrime, this info is not needed as ohci and ehci are standardized. chinese just dumped it info I do not know for which reason.
<arokux>
wingrime, in the code they actually use kernel stuff. everything they do is just poking at clocks and PHY to turn it ON
<wingrime>
arokux: then check it
<wingrime>
arokux: registers sould act
<wingrime>
*should
<arokux>
wingrime, so that's why I thought I should be easy thing, once I've understood it is like I said
<Turl>
Seppoz: both, but you'll probably prefer armhf for perf reasons
<arokux>
wingrime, yes, I need to triple check each step.
<wingrime>
arokux: check registers, if it reads/writes normaly than gate and clock correct
<wingrime>
arokux: and if it right we will look for other things
<arokux>
wingrime, but now i'm asking myself how I can use the working 3.4 kernel to see if AXP is working correctly
<wingrime>
arokux: just boot
<arokux>
wingrime, booted it. USBs working fine.
<wingrime>
arokux: you can try use aw axp sysfs for see
<woprr>
HAVE_LATENCYTOP_SUPPORT [=n] ?
<wingrime>
arokux: also do you try set PB09 to "1"
<arokux>
wingrime, no, what is that for?
<wingrime>
usb_drv_vbus_gpio
<woprr>
LATENCYTOP [=n] cause Depends on: HAVE_LATENCYTOP_SUPPORT [=n] && DEBUG_KERNEL [=y] && STACKTRACE_SUPPORT [=y] && PROC_FS [=y]
<woprr>
no support On branch stage/sunxi-3.4 sun7i_defconfig ?
<wingrime>
mripard_: I just thinked
<wingrime>
mripard_: most tables have 3.7 voltage
<wingrime>
mripard_: usb have 5v
<wingrime>
mripard_: so, DC-DC is required to make 5V from 3.7
<wingrime>
Turl: ^
<wingrime>
arokux: ^
<wingrime>
arokux: also I checked cb2 scheme, looks like gpio used onlt for OTG
<arokux>
wingrime, DC-DC needs a driver too?
<wingrime>
arokux: AXP have DC-DCs
<wingrime>
and MOSFET too internaly
<wingrime>
hno: ping
<arokux>
wingrime, (sorry, I know nothing about electronics)
<wingrime>
arokux: simply say AXP required, for make 5V from 3.3-3.7 for usb
<arokux>
wingrime, yes, this one I've got.
<wingrime>
arokux: also you must know, that battery usualy discharge, and voltage is decreacing
<wingrime>
arokux: but for CPU and USB you must have stable voltage
<arokux>
wingrime, mele is not powered by battery
<wingrime>
arokux: AXP up volate to noraml level
<arokux>
wingrime, ok, but CPU works, so AXP is doing something...
<arokux>
wingrime, maybe not everything :)
<wingrime>
arokux: axp have default config
<wingrime>
arokux: cpu must be on for boot
<wingrime>
arokux: and turn on voltage for usb
<arokux>
wingrime, from where do you know all this?
<wingrime>
arokux: logic
<arokux>
wingrime, for logic you need to know the principles, where is this from?
<wingrime>
arokux: I have RF engineering barcelor degree
<Turl>
arokux: basically wingrime is saying you need axp driver (or uboot hack) to get vbus on usb
<arokux>
Turl, yep
<Seppoz>
is there a howto on how to make any debian/ubuntu from cratch?
<wingrime>
Seppoz: see wiki
<wingrime>
Seppoz: all there
<Seppoz>
if i debootstrap debian on armel or armhf i allways get stucked
<Seppoz>
or only basic devices with address and no configuration
<wingrime>
Seppoz: you fistly need add some DT support
<wingrime>
to it
<Seppoz>
oh i see
<Seppoz>
mach files for the win!!!!
<Seppoz>
:P
<wingrime>
Seppoz: configuration also can be readed
<Seppoz>
i like the idea of dt files
<hno>
wingrime, pong?
<wingrime>
hno: do you know anuthing about usb power and axp
<Seppoz>
but only if you have some kind of editor and support for everything
<Seppoz>
what does apx have to do with usb power?
<Seppoz>
i mean its either direcltly connnected to the 5V regulator or indirectly using this chip
<wingrime>
hno: arokux trying get usb
<wingrime>
hno: but no power
<Seppoz>
arokux: did you meassure that there is no 5V?
<arokux>
Seppoz, nothing to measure with. if i connect my phone it won't charge.
<Seppoz>
want me to meassure it for you quickly?
<wingrime>
arokux: whait, change can be related with datalines
<Seppoz>
told him allready :P
<wingrime>
arokux: better connect something more simpler
<Seppoz>
ye like usb fan xD
<Seppoz>
or your thoung
<Seppoz>
lol
<hno>
wingrime, AXP209 and AXP202 seems virtually identical to me. Did translate most of 209 datasheet and always end up in same description (taking auto translation in account) as the 202 datasheet.
<arokux>
nothing simple here
<Seppoz>
arokux: want me to start your uImage and meassure?
<arokux>
Seppoz, not now, will test everything else works.
<Seppoz>
ok :) if you want me to just tell me
<arokux>
Seppoz, ok, thanks!
<hno>
AXP don't do much for the power of USB HOST ports.
<hno>
It monitors the VBUS of the OTG port and draws power from there, but that's about it I think,
<hno>
5V for HOST VBUS is taken directly from the power source on 5V powered devices such as Mele, Cubieboard etc.
<hno>
i don't know how they make the 5V in battery powerd devices, but doubt it's done by the AXP. Probably separate DCDC.The AXP is more used for voltage sensitive uses such as CPU core power,
<arokux>
hno, do you know why kernel log get's duplicated in 3.4? same message with time stamp and same without
heffer has quit [Quit: No Ping reply in 180 seconds.]
<hno>
yes, it's an allwinner hack.
heffer has joined #linux-sunxi
heffer has quit [Changing host]
heffer has joined #linux-sunxi
<Seppoz>
yea i wanna get rid of it too
<hno>
old leftover and can be torn out.
<hno>
just don't remember where it is...
<arokux>
hno, let's do it!
<arokux>
:)
<Seppoz>
i thought its because of the debuglevel in uboot
<Seppoz>
that is passed to the kernel
<hno>
no.
<arokux>
Seppoz, those are kernel msgs
<hno>
arokux, "gpio set ph6" from u-boot powers on the side USB port on my a1000.
<hno>
gpio set ph9
<hno>
err, wrong window.
<hno>
no power from u-boot on the back USB ports.
<hno>
guess those are controlled by the hub chip
<arokux>
gpio set ph6 <--- does nothing to me
<hno>
for me it turns on the usb led light I connected in the port.
<arokux>
hno, how do you check it gets power?
<arokux>
hno, hm, i've pluged in the card reader
<arokux>
hno, and it's led flushed
<arokux>
hno, it also flushes when i clear/set
<arokux>
hno, but the phone won't charge
<hno>
right.. phone is probably too smart.
<hno>
it won't charge unless it can negotiate sufficient power. a smart device is only allowed to draw 100mA without negotiation.
<hno>
which isn't sufficient for charging.
<arokux>
hno, btw, my cardreader's LED flushes for the ports at the back too
<hno>
arokux, when you do what?
<arokux>
hno, when my *new* driver starts :)
<Seppoz>
arokux, hno: this provves that the a1000 has the sxxx chip
<Seppoz>
because if it wouldnt have it the led would be on all the time
<hno>
led light indicates side USB port is on USBC1
<Seppoz>
SY6280
<hno>
Can be some other as well, but yes.
<arokux>
hno, why? as I said LED is flushing for all the ports for me
<Seppoz>
IMO the drriver need to handle this pin anyways
<hno>
arokux, maybe everything is working?
<arokux>
hno, maybe... I just need to find a way to understand it.
<arokux>
hno, I thought working should mean "registering of the usb stick" if i plug it in
<arokux>
hno, or charging of the phone.
<hno>
arokux, do you see the GL850G USB hub chip on the bus?
<arokux>
hno, yes, hub is up
deasy has quit [Remote host closed the connection]
<woprr>
whew! report sata working [ < 2.405598] ata1.00: ATA-8: WDC WD800BEVE-00A0HT0, 11.01A11, max UDMA/100
<woprr>
but lost sound & lan on olimex A20 micro
<woprr>
[ 1.692388] emac driver is disabled
<hno>
woprr, that means you have emac_used=0 in your script.bin.
<woprr>
[ 5.814594] ALSA device list:<6> No soundcards found.
<woprr>
should I use wemac, but menuconfig says its for older < sunxi
<hno>
arokux, can't find it now.
<arokux>
hno, ok
drachensun has quit [Ping timeout: 248 seconds]
* woprr
is checking schematics for realtek phy <-> Ports PA...
<woprr>
wemac from the olimex debian card worked with the olimex kernel
<woprr>
no wemac in a20-olinuxino_micro.fex
<arokux>
hno, good night
<hno>
odd..
<hno>
arokux, goot night.
<hno>
woprr, please extract script.bin from that debian image and compare with the sunxi-boards one to find if anything else is missing.. and submit a patch fixing.
<woprr>
ok, tomorrow
<hno>
machine is obviously completely wrong as well.
<hno>
makes me wonder if it's at all the right fex.
<woprr>
may not fit, I've got a rev.C board , other reported D, so we need revision suffixes?
navlrac has quit [Ping timeout: 264 seconds]
<hno>
Hmm.. was me that added that fex..
<woprr>
olimex uboot said EOMA86-A20
<woprr>
make help has a list
<woprr>
but don't know the matching name
<hno>
The A20 OLinuXino fex is sys_config/a20/a20-olinuxino_micro.fex
<woprr>
A10 rev.b sch confirms ethernet connected to PAx ports
<woprr>
yes, used that
<hno>
where is the debian image?
<woprr>
olimex said A10 = A20 board
<hno>
yes it is the same PCB.
wickwire has quit [Remote host closed the connection]
<woprr>
don't know, all links a dead, some torrent, lol, android guys...
<hno>
only differen CPU soldered, and some other small component changes.