nieuwbie has quit [Quit: ERC Version 5.3 (IRC client for Emacs)]
pwhalen has quit [Ping timeout: 260 seconds]
newleaves has joined #linux-sunxi
pwhalen has joined #linux-sunxi
pwhalen has joined #linux-sunxi
pwhalen has quit [Changing host]
<wens>
hramrach: CT wifi coming, driver support is in 3.14
<libv>
ssvb: even swrast brings it down
jinzo has quit [Quit: Leaving]
egbert_ has quit [Ping timeout: 246 seconds]
egbert has joined #linux-sunxi
cajg has quit [Ping timeout: 260 seconds]
cajg has joined #linux-sunxi
FreezingCold has quit [Ping timeout: 246 seconds]
prasannapete has joined #linux-sunxi
pwhalen has quit [Ping timeout: 272 seconds]
pwhalen has joined #linux-sunxi
hunterhu has quit [Quit: This computer has gone to sleep]
JohnDoe_71Rus has joined #linux-sunxi
bamvor has quit [Ping timeout: 248 seconds]
[7] has joined #linux-sunxi
TheSeven has quit [Disconnected by services]
Vince has quit [Ping timeout: 246 seconds]
mouchon2 has quit [Ping timeout: 260 seconds]
deasy has quit [Quit: Nom d'un quark, c'est Edmonton !]
<ssvb>
libv: good
<libv>
?
arete74 has quit [Ping timeout: 252 seconds]
<libv>
ssvb: it seems to stay up if you don't use a dri based program
<ssvb>
libv: I mean that it must be something relatively simple
atiti has joined #linux-sunxi
<libv>
ssvb: well, gdb is not turning up anything, even with dumbSched
<libv>
so whatever it is, it will not be easy to find, no matter how simple you think it will be
<ssvb>
maybe Option "NoTrapSignals" is needed to make gdb more happy?
newleaves has quit [Quit: Leaving]
maz_ has joined #linux-sunxi
<libv>
heh, now it doesn't crash anymore. grand
<libv>
notrapsignals and dumbSched
prasannapete has quit [Quit: prasannapete]
<ssvb>
there is also valgrind, but it may be a bit slow
ssvb has quit [Remote host closed the connection]
maz_ has quit [Ping timeout: 260 seconds]
libv has quit [Ping timeout: 252 seconds]
libv has joined #linux-sunxi
Black_Horseman has quit [Ping timeout: 260 seconds]
Quarx has joined #linux-sunxi
<oliv3r>
mornin' ya'll
diego_r has joined #linux-sunxi
<wens>
morning
mouchon2 has joined #linux-sunxi
<oliv3r>
WarheadsSE: mplayer2 in ubuntu works; but it 'slows down' after about 30 minutes; very strange, i tried mpv under fedora, and ran that for hours
Maverick2k has quit [Quit: Konversation terminated!]
rellla has joined #linux-sunxi
shineworld has joined #linux-sunxi
shineworld has left #linux-sunxi [#linux-sunxi]
arete74 has joined #linux-sunxi
<gzamboni>
morning
HeHoPMaJIeH has joined #linux-sunxi
HeHoPMaJIeH has joined #linux-sunxi
HeHoPMaJIeH has quit [Changing host]
prasannapete has joined #linux-sunxi
<sspiff>
oliv3r: perhaps a memory leak and swapping kicking in?
* sspiff
is currently debugging a memory leak in ffmpeg, and his spider sense is tingling
<oliv3r>
sspiff: nah
<wens>
can't wait for rc1 to come out
<wens>
I spotted a bunch of changes to musb and otg
<oliv3r>
sweet
<oliv3r>
sspiff: if it where a memory leak in mplayer2; wouldn't it crash after swap runs out?
<oliv3r>
sspiff: or shouldn't I be able to see that in memory usage?
Vince has joined #linux-sunxi
<sspiff>
oliv3r: yeah
<sspiff>
It was just a guess based on my limited knowledge and assumptions :)
<sspiff>
other than that, it could be thermal throttling, but 30mins is a long time for that to kick in
<oliv3r>
and our kernel doesn't even throttle :p
<oliv3r>
i thought it was after dpms
<oliv3r>
but that's ocmpletly off
<oliv3r>
and i think it slows down, after a few itterations, not during playback
<oliv3r>
(mplayer -loop 0)
<oliv3r>
mpv didn't do this
AreaScout has joined #linux-sunxi
Vince has quit [Remote host closed the connection]
<oliv3r>
sspiff: no memory leaking;' so that's not it
<oliv3r>
not lagging yet either though :p
<oliv3r>
also, heh, there is no swap!
<sspiff>
oliv3r: oh well, sorry to have wasted your time :)
<oliv3r>
no time wasted; it was a fair guess
<oliv3r>
i'm waiting for it to stutter again
rz2k has joined #linux-sunxi
FreezingCold has joined #linux-sunxi
<gzamboni>
hey, im getting a probe error in my dma controller driver ( sunxi_dmac: probe of sunxi_dmac failed with error -22 ) any idea of how can i debug and find from where does it comes from ? i used somes printk but it seems to pass everything
notmart has joined #linux-sunxi
notmart has joined #linux-sunxi
<oliv3r>
Turl: melita keeps calling it FOSDEN
<Montjoie>
gzamboni, place a printk before each return -EINVAL ?
prasannapete has left #linux-sunxi [#linux-sunxi]
<wens>
it's probably passed from other functions, such as clk_get
<gzamboni>
Montjoie, i will dig deeper, shouldnt it break after a -EINVAL ? in almost all EINVAL i have a goto to unload stuff, but i think there are some without
<ccaione>
gzamboni: probably the errer is printed after the probe() returns
<ccaione>
probe() is supposed to return 0
<gzamboni>
i found something, i was using a bad ahb clk
<gzamboni>
well, im happy, im progressing :P
hansg has joined #linux-sunxi
rz2k has quit [Ping timeout: 252 seconds]
penghb1 has joined #linux-sunxi
penghb has quit [Ping timeout: 248 seconds]
<Montjoie>
gzamboni, you work on the DMA engine ?
hipboi has quit [Quit: Leaving]
<gzamboni>
Montjoie, yes, but i'm just playing around, i dont know if will be able to make it work
<Turl>
oliv3r: refresh my memory please :p
<Montjoie>
gzamboni, I have read that mripard_ would work on it
<wens>
Montjoie: mripard_ plan is to get SPI working first, then work on DMA with SPI as testbed
<gzamboni>
wens, he did sent a patch for spi without dma, but its only for A31
<Montjoie>
If he not work on it soon, I will try also for fun, I really need it for the Security System:)
<oliv3r>
Turl: FOSDEM2014, feb 1st 2nd
hansg has quit [Ping timeout: 272 seconds]
<gzamboni>
soon i will have to do my watchlist for Fosdem's
<gzamboni>
should i watch oliv3r's talk ? :P
<ccaione>
Turl: we will miss you at fosdem :)
<Turl>
oliv3r: the melita part :p
<Turl>
ccaione: I'll be watching the streams :)
<oliv3r>
Turl: from packt publishing, she keeps calling FOSDEM, FOSDEN
<oliv3r>
m vs n
<gzamboni>
how much does it costs for the air tickets ? 600 euros ? Argentine -Brussels
<Turl>
oliv3r: ahh now I remember :)
<Turl>
let me check
<gzamboni>
whos melita ?
kivutar has joined #linux-sunxi
chax has joined #linux-sunxi
<wens>
gzamboni: he'll probably get SPI merged first
<chax>
Could someone help me out with a UART question?
<Turl>
gzamboni: a book editor
<ccaione>
Turl has a private jet ;)
<Turl>
gzamboni: 2250$ from my local airport to BRU
<Turl>
(two way tickets)
<ccaione>
chax: simply ask
<chax>
Well, the documentation states that all UARTs on the A10 are able to support IRDA mode (SIR). I cannot find out if this is supported by the kernel or modifying a FEX file for his.
<Turl>
gzamboni: 1730$ from the capital city on air europe
<gzamboni>
olimex is fast, i order yesterday morning and here it is, im in a small city about 200km from Bordaux
rellla2 has joined #linux-sunxi
egbert has quit [Read error: Operation timed out]
egbert has joined #linux-sunxi
<gzamboni>
Turl its more expensive than from Brazil, its logical as its more far away
<gzamboni>
where are you in Argentina ?
rellla has quit [Ping timeout: 265 seconds]
<oliv3r>
Turl: LOCAL airport eh
<oliv3r>
Turl: you should do a maintrack talk, then fosdem will pay your tickets :p
FreezingCold has quit [Ping timeout: 272 seconds]
<oliv3r>
before i'll post on the ML: i'm looking for 1 more reviewer; free book!
<gzamboni>
do they do that ? ask for your gov ? are you attached to a university ?
popolon has joined #linux-sunxi
popolon has joined #linux-sunxi
<oliv3r>
chax: you sure it's the uart that handles that? the a10 has 2 IR ports capable of SIR i think
<wens>
gzamboni: FOSDEM pays, not the gov' :p
<gzamboni>
ccaione, are you going by plane to fosdem ? youre from spain right ? im going by car, i have still 3 free places
<chax>
Yes it has 2 dedicated IR ports capable of SIR/FIR and some other modes, but it also seems possible to modiufy a register to make a UART send IRDA
<ccaione>
gzamboni: from italy
<gzamboni>
in Brazil the gov do pay some brillant students to do the talks
<oliv3r>
chax: ah, i'm not aware of that :(
<chax>
Serial IRDA is just like normal UART except for the fact the signals are 3/16 of the length of normal UART.
<wens>
gzamboni: in Taiwan, too, but it has to be academic conferences I think
<gzamboni>
i lived in Sao Paulo - Brazil, damn how it was horrible to import stuff. i remember i had to wait for months and pay the double of the prices of taxes, sometimes packages just got stollen
<chax>
Section 18.3
<chax>
There are 8 UART controllers. UART1 has full modem control signals, including RTS, CTS, DTR,
<chax>
DSR, DCD and RING signal. UART2/3 has two data flow control singals, including RTS and CTS.
<chax>
Other UART controller has only two data signals, including DIN and DOUT. All UART
<chax>
controllers can be configured as Serial IrDA.
<chax>
You can configure the FEX file to support 2, 4 or 8 line serial mode
<Turl>
maybe if I ask the CS department I could get a grantlike thing
<chax>
But I cannot find out if there is a way to set the UART to IRDA mode
<Turl>
probably way easier to do for a local event when it costs cheap to go by bus (<200USD)
<chax>
It needs to be done in '18.4.9. UART Modem Control Register'
rz2k|2 has joined #linux-sunxi
<chax>
Bit 6: SIR Mode Enable
<chax>
0: IrDA SIR Mode disabled
<chax>
1: IrDA SIR Mode enabled
<gzamboni>
the school i studied was named Rosario, i went once to bariloche
<oliv3r>
chax: i don't think anybody done this :)
<gzamboni>
but for this Fosdem its too late
<chax>
oliv3r: probably not :). I could change the 8250 TTY serial driver in the kernel source tree but maybe there was another way
<oliv3r>
Turl: 200 USD from your uni to research something worthy of talking at fosdem
<chax>
I will go on and try this later ;)
<chax>
brb
hansg has joined #linux-sunxi
<wens>
chax: quick grep doesn't yield IRDA in 8250 driver
kivutar has quit [Read error: Connection reset by peer]
<oliv3r>
chax: so you might have to actually add this to the uart driver
kivutar has joined #linux-sunxi
enrico_ has joined #linux-sunxi
hansg has quit [Remote host closed the connection]
diego__r has joined #linux-sunxi
diego_r has quit [Disconnected by services]
diego_r has joined #linux-sunxi
diego__r has quit [Ping timeout: 248 seconds]
kivutar has quit [Read error: Connection reset by peer]
kivutar has joined #linux-sunxi
kivutar has quit [Read error: Connection reset by peer]
kivutar has joined #linux-sunxi
<chax>
I checked the file out and indeed there was no evidence about SIR mode. I think I will hack it in somehow (if I find out how this would work) and then see if the UART signal change in duration to IRDA
FreezingCold has joined #linux-sunxi
kivutar has quit [Read error: Connection reset by peer]
kivutar has joined #linux-sunxi
TomiK has joined #linux-sunxi
<TomiK>
hello :)
TomiK has quit [Read error: Connection reset by peer]
TomiK has joined #linux-sunxi
<TomiK>
:)
<oliv3r>
chax: why do you want to use uart so badly? what's wrong with the IR allready available? we do have drivers for that
kivutar has quit [Read error: Connection reset by peer]
<wens>
oliv3r: IR controller is for CIR, not IRDA
<chax>
Well, I only would like to use the UART in IR mode so without all ooverlay
<chax>
That too :)
<chax>
And as it merely is only a single register that needs to change
<chax>
That would probably be the easiest solution to use IRDA that way
<chax>
the rest of the system does not need to know about the fact that this is Serial IRDA, I can send out plain data to the UART where the 'other side' is actually not RS232 but infrared
<chax>
The actual IR interface supports CIR FIR and MIR but the driver only seems to do CIR
<chax>
That is not for Linux Sunxi I think? But I could use it as reference how to do it?
<Turl>
chax: it's from recent mainline linux, for another driver
<Turl>
but yeah, it could serve as inspiration :)
<oliv3r>
chax: check the mailing list, there's a newer driver that supports tons
<wens>
do all 8250 devices support IRDA? or even just all synopsys 8250 devices?
<chax>
I don't think all devices do
<chax>
But as only the UART signals are 3/16 of their original lenght, most UART peripherals support serial IRDA
<wens>
sounds like adding another quirk to 8250 driver
<chax>
The IRDA signals are 3/16 of the original length
<chax>
Yes probably
<chax>
just change the UART Modem Control Register bit 6 from 0 to 1
<chax>
and see what happens
<chax>
But I was wondering if there was a more elegant wat. Like changing the FEX file
<chax>
as the FEX also has settigns for 2, 4 or 8 wire serial interface
ganbold_ has joined #linux-sunxi
JohnDoe_71Rus has quit [Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org]
kivutar has joined #linux-sunxi
<oliv3r>
chax: changing the fex file is probably the least elegant :p
<oliv3r>
but quirck + dt/fex/sysfs toggle sounds nice
kivutar has quit [Read error: Connection reset by peer]
FreezingCold has quit [Ping timeout: 260 seconds]
<chax>
I am not that familiar with all files and the sunxi kernel
<chax>
What would that do?
<chax>
And what I was wondering, do we still need the sunxi kernel? Or can I get the latest kernel and compile that one too? Or is not everything in the mainline kernel yet?
<wens>
you can use sunxi-devel for mainline
<oliv3r>
but there are things missing (disp etc)
perr has joined #linux-sunxi
<perr>
hello
<perr>
anyone there
<perr>
what is the difference between TZIC & VIC
<Turl>
perr: background?
<oliv3r>
perr VIC being something interruptyness; and TZ being Trust Zone?
perr has quit [Ping timeout: 252 seconds]
perr has joined #linux-sunxi
<perr>
from the manual: The TZIC provides a software interface to control the generation of the nFIQ interrupt.
<perr>
In addition, it masks out the interrupt source(s) chosen for nFIQ from the interrupts that
<perr>
are passed onto a non-secure interrupt controller.
kivutar has quit [Read error: Connection reset by peer]
kivutar has joined #linux-sunxi
kivutar has quit [Read error: Connection reset by peer]
<perr>
in the VIC(PL192) manual, it does't mention the TZIC..and in the TZIC(SP890) manual, it does't talk much about the affect on VIC...is the TZIC just a mask for FIQ...not very clear..this is my question.
kivutar has joined #linux-sunxi
<oliv3r>
perr: someone on the mailining list implemented pcis
deasy has joined #linux-sunxi
<perr>
oliv3r► what is pcis
<perr>
?
<Turl>
perr: you may want to ask in #armlinux
<oliv3r>
i may have gotten that abbreviation wrong; but check the mailing list :)
<perr>
ok, i will try
<Turl>
perr: that's something mostly architectural and not really very sunxi specific
arete74 has quit [Quit: leaving]
<perr>
yes Turl
<oliv3r>
cubieboard is FCC tested?
<Turl>
oliv3r: I think it is in fact :p
<oliv3r>
yeah i'm reading the FCC pdf right now
<Turl>
I saw some certs on the site
<gzamboni>
is there a reason for a drive probe to run more than once ? for some reason i get errors when getting the dt data in the second time the driver probes
<oliv3r>
yeah
<oliv3r>
i dunno if they are fake or not; i mean it was done IN china :p
<Turl>
gzamboni: probe deferral
<oliv3r>
Tsvetan: where is your FCC report!
<Turl>
oliv3r: Tsvetan sells the stuff as dev boards so he doesn't need one :p
<oliv3r>
and cubietech doesn?T
<Turl>
oliv3r: it had pretty graphics and funny words, I think it's legit
<Turl>
oliv3r: they do too
<Turl>
oliv3r: but probably someone is shipping cubieboards as part of some end product
<Turl>
so they needed to fcc it or sth
TomiK has quit [Remote host closed the connection]
<gzamboni>
its dumb as certifications are company attached, its very probable that olimex clients does have ce and other certifications on olimex boards, but they are related to the company not to the product project.
<Turl>
gzamboni: probably you are returning EPROBE_DEFER on the error path of something (clock, regulator get most likely=
<gzamboni>
Turl, humm gotta check it, everything is new to me
<gzamboni>
from the supplier chain the certification pass to to buyer, but not the inverse
kivutar has quit [Read error: Connection reset by peer]
kivutar has joined #linux-sunxi
<Turl>
gzamboni: do you have the probe function somewhere? I can have a look if you want
<vector80>
Hi, when I checkout u-boot-sunxi, I can find my Interra-3 board, but when I checkout sunxi-bsp, I can't find it, why ?
<oliv3r>
vector80: because you forgot to cd sunxi-boards; git pull :)
<oliv3r>
sunxi-bsp determins the boards based on sunxi-boards
<vector80>
ok let me do it right now
<vector80>
today I studied git a little, I think I can create my 1st patch file and send you via email
<vector80>
oliv3r: Can I share my kernel .config via patch file also ?
<gzamboni>
Turl, this is my sun4 dtsi: http://pastebin.com/EQYurhy4 The driver is getting the info from the dtsi, but for some reason it does probe twice
FreezingCold has quit [Ping timeout: 252 seconds]
<Turl>
gzamboni: I see your code is also registered it via nonDT
kivutar has quit [Read error: Connection reset by peer]
<Turl>
gzamboni: that would be why it is first probed and fails to match device to of
kivutar has joined #linux-sunxi
<binaryferret>
I asked in olimex, but I thought I should also ask here. I have an a13 wifi olimex. I also have a 7inch lcd with ts. I'm running the image on an sd card provided on the olimex site. The issue I have is that once it boots into xfce and I open a menu, or bring something up on the screen it flickers madly. When I direct boot into android, or I'm in a terminal (eg tty4) screen the screen does not flicker.
<binaryferret>
I bought the power supply from the olimex site, and I've checked that it's working correctly (which it is). I'm stumpt.
<Turl>
libv: ^
<Turl>
binaryferret: are you using fbturbo?
<vector80>
binaryferret: I feel like it is a sys_config.fex issue
orly_owl has quit [Read error: Connection reset by peer]
orly_owl has joined #linux-sunxi
<binaryferret>
Turl: I have no idea, it's pretty out of the box image file put directly onto the sd without anything else put on it.
<libv>
binaryferret: sorry, use our stuff please
<Turl>
libv: olimex ships our stuff afaik
<libv>
binaryferret: the display doesn't flicker all the time
<libv>
Turl: i don't know :)
<binaryferret>
heh, that was my next port of call as I've had success with linux-sunix on a13 tablets. I was just wondering if it was something that had popped up before
<binaryferret>
libv: Nope, just when I click on a menu, or bring something up on the screen
<Turl>
libv: sounds like an fbturbo bug to me
<libv>
binaryferret: does the 7" lcd flicker, or does the graphics flicker?
<Turl>
binaryferret: you were using a weird res weren't you?
<binaryferret>
Turl: Nope, I ran the set display script and set it to LCD 800x480
<vector80>
Turl: oliv3r, thank you... I don't know how that happened, but when I branch to sunxi, now it is compiling...
kivutar has joined #linux-sunxi
<vector80>
Thats so strange, why the build system doesn't do these automatically ?
<oliv3r>
Turl: pwd turl
<oliv3r>
Turl: check your mail
<oliv3r>
Turl: slick!
<oliv3r>
i like that clock
<oliv3r>
499 USD
<libv>
pretty hard to read
<oliv3r>
yeah but it looks awesome
<oliv3r>
e-ink baseed
<Turl>
one cell batt for one year :)
<Turl>
libv: I've seen worse
perr has joined #linux-sunxi
<oliv3r>
our clock has been running on an old discarded battery for 3 years now
<oliv3r>
maybe 2 old discarded battereis; but still going
<oliv3r>
and it's a mechanical clock to boot
<Turl>
oliv3r: I put a new one on mine and it worked less than a week :p
<oliv3r>
Turl: short circuit :p
<oliv3r>
Turl: did you get my e-mail?
<oliv3r>
and did it work?
<Turl>
oliv3r: yep reading it now
<oliv3r>
using owncloud as your document storage backend is amazing; versioning!
<Turl>
oliv3r: chapter 8 o.O
<Turl>
oliv3r: pff can't beat git :)
kivutar has quit [Read error: Connection reset by peer]
<oliv3r>
to much work ;) and .odt is zip; so binary, and makes git cry
<Turl>
oliv3r: of course I meant git + tex + Make
<Turl>
:p
<oliv3r>
Turl: chapter 8 depends on how far we get :p it's 'optional'
kivutar has joined #linux-sunxi
<oliv3r>
Turl: but that should give you an estimate of the target users ;)
<Turl>
yeah, it's going to be a fine book :)
perr has left #linux-sunxi ["Leaving"]
vector80 has quit [Ping timeout: 248 seconds]
kivutar has quit [Read error: Connection reset by peer]
kivutar has joined #linux-sunxi
orly_owl has quit [Read error: Connection reset by peer]
FreezingCold has joined #linux-sunxi
TomiK has joined #linux-sunxi
orly_owl has joined #linux-sunxi
<Montjoie>
I want to try sunxi-devel but http://linux-sunxi.org/Mainline_Kernel_Howto said that I mus stop autoboot for doing some hack but I dont have any serial console, so mainline without serial is impossible?
FreezingCold has quit [Ping timeout: 252 seconds]
<hramrach>
you ca edit environment without serial
<hramrach>
the default copmiled-in env for one
<hramrach>
also uEnv.txt
<oliv3r>
are miniand's hackberries opensource anything?
orly_owl has quit [Read error: Connection reset by peer]
<Turl>
Montjoie: mainline doesn't have video out, how would you interact?
<hramrach>
but without console it's hard to see what's going on and mainline fails a lot
<Turl>
ssh would be the only option
rz2k|2 has quit []
<Montjoie>
I interact only with ssh
<Turl>
ok then :)
* Turl
should write a tool to easily make mainline-compatible cards
<hramrach>
like my tablet. I boot sunxi linux on it and nothing happens
<oliv3r>
hramrach: no LCD driver in mainline?
<Turl>
kind of like hans's fedora tool
<hramrach>
no disp driver
orly_owl has joined #linux-sunxi
<Montjoie>
does updating uboot is usefull, my uboot is the same since 4/5 month ?
<Turl>
oliv3r: mainline doesn't have disp nor dma nor audio
<Turl>
Montjoie: are you on a20?
<Montjoie>
yes
<Montjoie>
Cubieboard2
<hramrach>
you need current for a20
<Turl>
you may get some performance boost then
<hramrach>
for smp
<Turl>
(memory bw)
<oliv3r>
Turl: hansg's fedora images are so aweosme, because the setup makes it 'just owrk'
<oliv3r>
memory bw, gmac over emac
<hramrach>
it does not work for me :p
<Turl>
oliv3r: yeah the setup is nice :)
* hramrach
has some very uncooperative tablet
<Montjoie>
Ok I will update u-boot first
pasquale has quit [Quit: Ex-Chat]
<oliv3r>
hramrach: do you have serial?
kivutar has quit [Read error: Connection reset by peer]
kivutar has joined #linux-sunxi
mouchon2 has quit [Ping timeout: 264 seconds]
FreezingCold has joined #linux-sunxi
vector80 has joined #linux-sunxi
<vector80>
Hi, I have a question from my colleague. He is developing mostly on PIC microcontrollers, and he wants to understand how HIGH RESOLUTION TIMERS is working under linux.
<vector80>
For example, he wants to do a hard time critical job, say to enter one interrupt in every 100microseconds
<vector80>
As I know, A20 has 4 HRTs, can you give me some idea how to achieve this by code ?
<Turl>
oliv3r: I see the bsp has some provision for boot.cmd -> scr and embedding it into hwpacks, how does that work?
<Turl>
vector80: linux is not really a RTOS
<oliv3r>
Turl: magic :p
<oliv3r>
Turl: it uses mkimage from the makefile; bog standard stuff
<Turl>
oliv3r: I'm confused because of the relative paths
<Turl>
oliv3r: it checks if "boot.cmd" exists.. but where?
<vector80>
Turl: So, you mean, he can't do RealTime interrupts like on PIC ?
<Turl>
vector80: maybe processor is running with interrupts off for some period and it gets delayed
kivutar has quit [Read error: Connection reset by peer]
<Turl>
there is a series of patches to make linux better for rt if you are interested
<vector80>
yes, he just like to know, how to do such thing
<Turl>
vector80: tell him to read about linux-rt
kivutar has joined #linux-sunxi
<hramrach>
oliv3r: no serial so far. but it's a13 so probably doable one way or another
<oliv3r>
hramrach: yeah pins!
<vector80>
For example, on PIC, when CPU enter to that interrupt, it stops doing everything else, and only do that critical task, and when completes, return back to normal process
<oliv3r>
Turl: erm, i dunno; find . -name "boot.cmd" :)
<vector80>
Let me check linux-rt
<oliv3r>
vector80: the kernel uses one of the HR timers i think; i dunno if we have a driver for it
<oliv3r>
you could hack a custom thing to use the timer and trigger an interrupt from it
<vector80>
During kernel boot, it printk some very high res timing information, does it use HRT ?
<oliv3r>
pretty much like on the PIC; but don't know if it can be hard-reltime; as the interrupt service routine will still needed to be handled by the kernel as normal
<oliv3r>
vector80: yes
<oliv3r>
vector80: well, i would think so
<vector80>
Ok very clear now, thank you very much
<oliv3r>
vector80: but that's allt oo specific that i don't know
<oliv3r>
linux-rt are patches ontop of linux
<libv>
oliv3r: when did your fosdem interview get posted btw?
<oliv3r>
mripard added support for the HR timer as the main linux timer
<oliv3r>
libv: last week I think :)
<oliv3r>
libv: did I say something stupid?
<libv>
ok, you seem to not have made noise about it here, or i missed it
<oliv3r>
vector80: so if rt linux uses that timer, it should be fine, but you can still use one of the other times i guess
<oliv3r>
libv: you missed it; others made noise as they noticed it before I did :D
<libv>
oliv3r: don't be so insecure :p
<libv>
so no, nothing stupid, just a bit excited :)
<libv>
not that that is a bad thing :)
<oliv3r>
:D
<oliv3r>
i AM excited
<oliv3r>
i hope to start working on my slides tomorrow, have a book chapter deadline today :(
Coweater has quit [Ping timeout: 252 seconds]
alcides has joined #linux-sunxi
<Turl>
oliv3r: are they as strict as with the reviewers?
<Turl>
they told me they'd give me deadlines and stuff to comment on your chapters
<oliv3r>
is the imrpov opensource hardware?
<Turl>
is that the EOMA?
<oliv3r>
Turl: haha, yeah 'reminding' since yesterday, but i dont think they are strict at all
<oliv3r>
Turl: yeah
<Turl>
iirc it wasn't (pretty ironic I know)
<oliv3r>
well the 'motherboard' thing
<Turl>
lkcl: can you comment on that? ^^
<Turl>
lkcl: is improv OSHW?
<oliv3r>
they just like to push the authors/reviewers, as they know we are lazy slackers
<Turl>
oliv3r: they don't pay us much, so we should get that right :p
kivutar has quit [Read error: Connection reset by peer]
kivutar has joined #linux-sunxi
<Montjoie>
hramrach, so if I put all 4 lines in uEnv.txt, it will work ?
<vector80>
oliv3r: Thank you, we read many thing and discussed, but he is still not convinced about it if that one is really realy really real hard timer. Because he thinks, with this example: http://en.wikipedia.org/wiki/RTLinux, it is just setting higher priority to one task. So, it might delay in time, for example after 1 hour, would it be entering to interrupt in every 100ms ?
<Turl>
vector80: what do you mean 1h?
<vector80>
1 hour
<Turl>
you mean if it is going to take 1h instead of 100ms?
kivutar has quit [Read error: Connection reset by peer]
<Montjoie>
mainline kernel KO:(
<Turl>
Montjoie: what did you add on uEnv?
<Montjoie>
the 4 lines needed to type on u-boot
pwhalen has quit [Ping timeout: 260 seconds]
<Montjoie>
with the <board> changed by the filename
<Turl>
Montjoie: you may want to add root=... on bootargs too
nedko has joined #linux-sunxi
nedko has quit [Changing host]
nedko has joined #linux-sunxi
sspiff has quit [Quit: Konversation terminated!]
FreezingCold has quit [Ping timeout: 252 seconds]
kivutar has quit [Read error: Connection reset by peer]
kivutar has joined #linux-sunxi
paulk-collins has quit [Remote host closed the connection]
netlynx has joined #linux-sunxi
netlynx has quit [Changing host]
netlynx has joined #linux-sunxi
kivutar has quit [Read error: Connection reset by peer]
<lkcl>
Turl: que? let me read back....
Quarx has quit []
<lkcl>
oliv3r: oh! i just remembered: can you please make sure that you put "QiMod Ltd" prominently across any EOMA or EOMA68 material presented?
kivutar has joined #linux-sunxi
<lkcl>
Turl, oliv3r: right. Improv's engineering board comes with a promise that it will be released under an OSHW license once it's finalised.
<lkcl>
if anyone wants to do one themselves then i've already released the PCB and Schematics files for the MEBv1 (warts and all) in their original Altium files.
<Turl>
lkcl: promises, promises... so easy to break
<nedko>
lkcl: will the a20 cpu card be ever released as open source hardware?
atiti has quit [Ping timeout: 264 seconds]
Black_Horseman has joined #linux-sunxi
printallthething has joined #linux-sunxi
<lkcl>
nedko: possibly, once the investment - which includes my own personal financial wealth - is recouped.
<lkcl>
in other words, it would be financially irresponsible of me to place the CAD files into the public domain
<lkcl>
without anyone ever having given me any financial reward for having done the work
TomiK has quit [Read error: Connection reset by peer]
TomiK has joined #linux-sunxi
kivutar has quit [Read error: Connection reset by peer]
<nedko>
lkcl: i understand
<lkcl>
Turl: they're _going_ to release them... but *after* they've received the money from people.
<lkcl>
nedko: on the other hand... there is nothing to stop anyone from taking the GPL'd KiCAD schematics i developed, over two years ago, and completing them :)
<diego_r>
Turl: I think Aaron Seigo can be trusted regarding license promises.
<lkcl>
diego_r: yeah, exactly.
kivutar has joined #linux-sunxi
<lkcl>
this all ties in with the conversation yesterday.
<Turl>
eg what if they don't sell as many as they'd like?
<Turl>
lkcl: yeah, although I don't like it much :)
<lkcl>
Turl: well, he's the business entrepreneurial brains behind KDE. he's as well-known and as well-respected as other software libre project leaders
<nedko>
lkcl: what about the possibility of getting less money because the hardware is not open source?
<lkcl>
however, in his case, he also comes with a business brain attached :)
<lkcl>
nedko: you may not be aware of this but i have *REPEATEDLY* presented the open source community with opportunities to help move the EOMA project forward
<lkcl>
in the form of open hardware products.
<lkcl>
not a SINGLE person has committed to that.
<lkcl>
so i have been backed into a corner and have had to pay people to get the work done.
<lkcl>
simple as that.
<Turl>
lkcl: your description makes me think of De Icaza :)
<nedko>
lkcl: what commitment? money?
<lkcl>
if people truly _wanted_ open hardware, they would have committed to it
<lkcl>
nedko: money, time, skill, resources - anything that would create an open hardware EOMA68 CPU Card.
<lkcl>
there were no takers. look at the dates on the GPL-licensed git repo containing open hardware designs on git.rhombus-tech.net
FreezingCold has joined #linux-sunxi
bfree_ has joined #linux-sunxi
<Turl>
lkcl: you could've made a company and make it a product
<lkcl>
Turl: yeahhh, miguel has a different focus.
<lkcl>
Turl: i *have* made a company and made a product!! :)
specing has quit [Ping timeout: 260 seconds]
bfree has quit [Ping timeout: 260 seconds]
bamvor has quit [Ping timeout: 260 seconds]
<Turl>
lkcl: why is it not free hw then?!
bamvor_ has joined #linux-sunxi
<diego_r>
Turl: De Icaza used to spread FUD and copied Microsoft's .NET. Seigo did not.
bsdfox has quit [Ping timeout: 260 seconds]
ccaione has quit [Ping timeout: 260 seconds]
pasquale has quit [Quit: Ex-Chat]
bsdfox has joined #linux-sunxi
bsdfox has quit [Changing host]
bsdfox has joined #linux-sunxi
<lkcl>
Turl: because we need the guarantee that people will pay *us* for having provided them with value
<Turl>
lkcl: you give them hw they give you money :)
specing has joined #linux-sunxi
<lkcl>
it would be financially irresponsible to place the designs into the public domain such that *other people* make the hardware and receive the money instead!
<lkcl>
Turl: done!
<lkcl>
that's what's being done.
<Turl>
diego_r: "copying .NET" is pretty handy. Can't comment on the FUD, I've not followed him much
<Turl>
lkcl: why is not free then?
<Turl>
lkcl: if Olimex can, you can too :)
<lkcl>
Turl: define "free"
<Turl>
lkcl: OSHW
<Turl>
that's the whole point isn't it? :)
<lkcl>
and please don't mention tsevtan to me again.
<lkcl>
Turl: define OSHW
ccaione has joined #linux-sunxi
<nedko>
lkcl: public domain? i'm more worried about knowing what is in the hardware than being able to produce it myself.
<Turl>
lkcl: there's a pretty good definition..
<lkcl>
Turl: let me be explicit. for your money, do you want the schematics?
kivutar has quit [Read error: Connection reset by peer]
<Turl>
lkcl: and all the required files to modify and replicate such hw
<lkcl>
ok, not a problem. you can have the hardware.
<lkcl>
Turl: absolutely not. for that you're going to have to pay me sufficient money to recoup my investment.
<lkcl>
can you *guarantee* that?
kivutar has joined #linux-sunxi
<lkcl>
or you're going to have to come up with a scheme which *guarantees* that the investment of time and money is repaid
<Turl>
lkcl: then it's not oshw and you missed the whole point :)
<lkcl>
Turl: i have not.
<lkcl>
i *want* to release this as OSHW
<lkcl>
but i *HAVE* to guarantee success.
deasy has quit [Quit: Nom d'un quark, c'est Edmonton !]
<lkcl>
i CANNOT take the risk that some other fucker will come along - even before i've started - and rip the entire project off.
<Turl>
lkcl: if you can't guarantee success by the product alone you're pretty much doomed
<lkcl>
how do you suppose that can be guaranteed?
<lkcl>
Turl: remember that you're talking to someone who has been working with software libre for eighteen years
<lkcl>
so.
<lkcl>
simple question
<lkcl>
how do you guarantee that i will get my money back from the work i've put into this project so far?
<Turl>
by having a great product people will take off your hands
<nedko>
are "banks" supposed to give you credit for a business plan? so you basically outsource the risk of failure by selling part of your business to them?
<lkcl>
great!
<Turl>
(if you aren't aiming for that, why bother then?)
<Turl>
in any case, your business, your choices. I won't push you in either direction.
<lkcl>
good answer. so we have a great product. it's a fantastic idea, which, once the pressure is off, we will be free to explore the OSHW options, because that's *really* where i want to take this.
<lkcl>
so i have released ***SEVERAL*** OSHW I/O boards, Turl.
<lkcl>
did you know that?
<lkcl>
Turl: have you seen the -----*****>OSHW<****------ I/O boards that i've released under ****OSHW***** licenses?
<lkcl>
Turl: the router for example? the schematics and PCB files are released as ****OSHW****.
<Turl>
lkcl: but the card isn't? :)
<lkcl>
Turl: the *first* card - which i want my money and investment back on - is not. YET.
<lkcl>
i want the project to be financially stable first.
<lkcl>
*then* we can look at releasing CPU Cards under OSHW licenses.
Coweater has joined #linux-sunxi
<lkcl>
and... i have to say that if you look closely, you'll find that there are plenty of ***OSHW*** EOMA-68 CPU Cards designs that i've done.
<Turl>
lkcl: just so you know, I don't gain much by OSHW, other than the basic schematics (which nonOSHW also has available). I can't even solder a cable to a 3.5mm jack. I just ask for it because I know there's people who care and may find it useful
<lkcl>
including the jz4760 one.
<lkcl>
the schematics as a PDF i have absolutely no problem with providing that to people on request.
<lkcl>
if they need it.
<lkcl>
but, to be honest, EOMA68 is so simple that there are very few justifiable reasons why people need it.
kivutar has quit [Read error: Connection reset by peer]
<lkcl>
ok, listen: i gotta go, ok?
<Turl>
lkcl: sure, cya
kivutar has joined #linux-sunxi
<Turl>
I'm always idling here, you know you can ping me anytime if you want to bring this back
FR^2 has quit [Quit: Connection reset by peer]
<diego_r>
Turl: overall I understand the points made by lkcl. I think the problem might be that a dual license (GPL or commercial for business) like Qt did in the pre-Nokia era is not possible in the case of hardware.
<Turl>
diego_r: why would it not be possible?
* nedko
wonders the same
kivutar has quit [Read error: Connection reset by peer]
kivutar has joined #linux-sunxi
<diego_r>
Turl: nedko: there's nothing like "linking" in the hardware. The Qt dual license model worked because you could not release non-GPL software using Qt libraries. In the case of hardware you're not using a "library", you're using "the project of a board".
<diego_r>
Or if you want to see it in another way: the EOMA I/O board is the library, the EOMA CPU card is the program.
<Turl>
diego_r: but if you own the design you can license it as you see fit?
TomiK has joined #linux-sunxi
Fusing has joined #linux-sunxi
<nedko>
a license that allows doesnt allow other companies to make profit of the design makes sense to me
<diego_r>
Turl: read carefully the points 8 and 11 of the oshwa.org link you posted.
<diego_r>
8. For example, it must not restrict the hardware from being used in a business, or from being used in nuclear research.
<diego_r>
11. For example, the license must not insist that all other hardware sold with the licensed item be open source, nor that only open source software be used external to the device.
FreezingCold has quit [Ping timeout: 264 seconds]
<diego_r>
so: why would they license something that they can already do?
<nedko>
this sounds like mit/bsd
<Turl>
diego_r: so how does that block dual licensing?
<nedko>
i'm perfectly ok with gpl-like with some non-commercial clause
<diego_r>
Turl: dual licensing make sense when you want to have "free end users", but "paid commercial users / business"
<Turl>
diego_r: community version under gpl would require people who do modifications and distribute it to share their changes
<Turl>
diego_r: then a commercial version would lift hw vendors from that restriction
<hramrach>
in practice no people submitted modifications
<diego_r>
Turl: hardware is not "patchable"
<diego_r>
or to say it other works, "patches cost thousands of $"
kivutar has quit [Read error: Connection reset by peer]
<Turl>
diego_r: so it is patchable, just not as cheaply :)
<diego_r>
Turl: yeah
<montjoie[home]>
Turl thanks for the uEnv.txt but it is not better
<Turl>
montjoie[home]: what branch are you using?
kivutar has joined #linux-sunxi
<montjoie[home]>
mripard sunxi-next
<hramrach>
meaning that in practice only party that gains by releasing the sauce is a Chinese company that want s to make clones of your board
<Turl>
montjoie[home]: that doesn't have gmac support
<montjoie[home]>
but emac no ?
<Turl>
montjoie[home]: you have ct? or cb2?
<hramrach>
probably not
<hramrach>
try sunxi-devel
<diego_r>
or to say it better: what counts in hardware is the build, not the source. And OSHW is not giving restrictions on "builds", so if you give away your source, you more than likely give away your source of income (unlike dual licensing options that are used in software).
<hramrach>
gma should work on anything
<hramrach>
c
<montjoie[home]>
Turl cubieboard2
<Turl>
diego_r: for some reason I don't see olimex being broke, but w/e :)
<hramrach>
diego_r: on the other hand, olimex and dangerous prototypes make oshw and live
<Turl>
montjoie[home]: try sunxi-devel as hramrach said
<nedko>
i'd be very happy if olimex sold open source eoma68 stuff
<Turl>
diego_r: you win some income too with people requiring board modifications
<montjoie[home]>
ok I will try that thanks
<hramrach>
so the danger of Chinese clones is not as much of a problem as some people claim
<Turl>
hramrach: even arduino
<hramrach>
olimex does not sell eoma68 stuff because the do not want to and the standard is seriously outdated
<Turl>
hramrach: there the cloning is rampant
<Turl>
they still manage to be around
<hramrach>
because the clones are lower cost but lower quality
<nedko>
hramrach: what is outdated in eoma68?
<hramrach>
interfaces
<hramrach>
current SoCs have different interfaces from what eoma68 specifies
<hramrach>
and it's only getting worse over time
<nedko>
for example?
<hramrach>
does it have gbit ethernet, sata, hdmi, lvds, edp, usb3, pcie?
<nedko>
yes
<hramrach>
no
<Turl>
it only has rgb as display
<nedko>
well almost
<Turl>
and usb2
<Turl>
no pcie afaik
<nedko>
i read the spec last night, you are wrong :)
<hramrach>
less wires, cheaper displays, probably standardized connectors
<nedko>
Turl: pins 20,21,54 and 55?
<hramrach>
soo awesome
popolon has quit [Quit: Quitte]
<nedko>
you cannot implement eDP over existing interface?
<hramrach>
converters are expensive
<hramrach>
and it's pretty backwards when the SoC already ahs it
<Turl>
nedko: aha :) the first page really needs an update
enrico_ has quit [Quit: Bye]
<nedko>
hramrach: non-reusable interfaces are more expensive ;)
<hramrach>
non-reusable in what way?
diego_r has quit [Quit: Konversation terminated!]
<hramrach>
it's like you said SATA is more expensive than IDE
<hramrach>
in hte start the adaptors were quite expensive and you had to pay a premium for the disks
<nedko>
it is, if your current hardware is IDE only :)
<hramrach>
and where is ide now?
<hramrach>
but eoma is RGB-only artifically by poor design
<nedko>
for my project i don't need any display IO :)
<nedko>
i'd use another i2c bus though
<hramrach>
you can still get RGB on (some) current SoCs but it's the oldest and worst display interface the SoCs have
<hramrach>
there you go. most SoCs have like 4 i2c but only 1 on eoma (presumably for display ddc)
<nedko>
so do you think there can be long-term modular in any kind?
<HdkR>
eDP is awesome
<hramrach>
youo need to figure out a standard that is long-term sustainable even if the used interfaces change or live with different mutually incomaptible standards
<HdkR>
It's like everybody hates using DP, but eDP they are all over because LVDS sucks
<nedko>
hramrach: does a20 have eDP interface?
<hramrach>
it's not like you could not use the RGB pins for LVDS/eDP/HDMI but there is not even enough gpio to encode such switch
<hramrach>
nedko: a20 probably does not
<hramrach>
but not sure, anyhow
<nedko>
so eDP is still not here in my book :)
<hramrach>
it is on some SoCs, and it is the way for displays in new designs
pwhalen has joined #linux-sunxi
<nedko>
and as a user i don't care about it, even if i use eoma for a laptop
<HdkR>
Exynos 5 supports it :P
<HdkR>
Tegra supports it :D
<HdkR>
I don't think anyone cares if Qualcomm does or not
<hramrach>
and they both have PCIe so you can get an extra nVidia GPU with all sorts of connections ;-)
<HdkR>
Tegra 4 removed PCIe though
<HdkR>
Thought it was Tegra 3 and 5 that had it
<hramrach>
probably found nobody used it
<hramrach>
but it's just matter of time
<nedko>
i still think that opencompute approach with pcie connectors is nice
<hramrach>
the SoCs are fast enough to make use of it and it makes the designs so much more flexible
<HdkR>
Actually, I'd have to look again for T5, it may just be a internal PCIe connection to the kepler GPU
rz2k has joined #linux-sunxi
<mrnuke>
guys. nice-job with sunxi-devel. Keep up the good (mainlining) work.
<parabyte>
I got a interesting idea, I am not into hardware so why i ask, is it possible to attach a pci-e device regardless of bus width to any of the devices you guys play with?
ssvb has joined #linux-sunxi
<jelly-home>
ParaByte: and make it do something?
<parabyte>
yeah i just realised
<parabyte>
that Ati do not release there drivers open source
<parabyte>
which is a shame really
<nedko>
which drivers?
<nedko>
for amd video hardware?
<parabyte>
ati's closed source ones
<parabyte>
yeah
<nedko>
they are much better than nvidia
<nedko>
they release docs
<nedko>
and there is open source driver
<nedko>
good one
<parabyte>
I just went on ebay and got a nano x86 motherboard with pcie's
<parabyte>
and extender cables
<nedko>
it even works better than the closed source (for me, last time i tried)
<parabyte>
problem solved ;)
<nedko>
when i tried fglrx was not be unable to do KMS
<parabyte>
I using a old p4 for mining and thought be cool to utilize armel stuff, but i guess intel nano what ever its called will do
<parabyte>
i need a good opencl
<HdkR>
a machine good at bitmining you mean?
<HdkR>
I wonder how well a Tegra 5 will do with bitmining
<HdkR>
How efficient it is really
<jelly-home>
scrypt? don't expect too much for a chip optimized for low power
<HdkR>
Since bitmining is really just about efficiency, not being stupidly quick
<HdkR>
being stupidly quick just makes it better :P
<jelly-home>
HdkR: scrypt apparently relies on lots of memory bw, and that's one of the things that _will_ be crippled
<HdkR>
Any of these applications take advantage of OCL?
<HdkR>
on Android that is
<HdkR>
Considering Google broke OCL, meeh
<HdkR>
Q:Why is GPU mining not supported?
<HdkR>
A:Most manufacturers do not release OpenCL (GPU computation API) drivers for mobile devices. Only the Nexus 4 and Nexus 10 have support for it in Android 4.2, and, as the drivers are not included in Android 4.3, there will be no GPU mining on Android in the foreseeable future.
<HdkR>
blah
<HdkR>
Look forward to the future developers! D:<
<HdkR>
This is why Dolphin already supports OpenGL on Android
pwhalen has quit [Ping timeout: 248 seconds]
eebrah|mobile has joined #linux-sunxi
FR^2 has joined #linux-sunxi
TomiK has quit [Remote host closed the connection]
<oliv3r>
lkcl: qimod? not rhombus-tech?
TomiK has joined #linux-sunxi
deasy has joined #linux-sunxi
<parabyte>
yup scrypt
<parabyte>
I believe its Gpu Ram not system ram
<oliv3r>
vector80: no, first off; even linux software interupts; they will fire at a certain time. in soft-realtime; it may miss an interrupt, but will try for the next one to have it on time. there won't be any difference in the long run; hard real time systems 'guarantee' that the interrupts will be handled in time (up to a certain threshold); the HSR timer (and even the regular timer) will just trigger an intterupt at a fixed time (when the timer expires).
<parabyte>
i was getting very good results with a system with just 1 gig of ram and four 7950's
<oliv3r>
vector80: a pic does EXACTLY the same thing. As does AVR. the interrupt triggers; the 'os' will have to do something. however, and this is the huge difference, with avr/pic; if your program takes longer then there is time inbetween interrupts, you also will get jitter; but its easier to work things out, as you can count clock cycles etc
<oliv3r>
lkcl: please make sure, certailnly means 'would you be so kind'
<oliv3r>
lkcl: ok for now, i'll put 'opensource board' in my book, and will revise it in 2 months if nothing has changed
eebrah|mobile has quit [Read error: Connection reset by peer]
eebrah|mobile has joined #linux-sunxi
<oliv3r>
lkcl: what about if the project totally tanks; do you still promise to release it OSHW?
pwhalen has joined #linux-sunxi
pwhalen has quit [Changing host]
pwhalen has joined #linux-sunxi
<oliv3r>
lkcl: people want truely open hardware; but us RMS folk are few, and even less with EE backgrounds
eebrah|mobile has quit [Ping timeout: 246 seconds]
<mnemoc>
libv: pong
FreezingCold has joined #linux-sunxi
<mnemoc>
oliv3r: what's missing on the lime to be "truly" open?
<oliv3r>
nothing afiak
<oliv3r>
LIME is truely open
<oliv3r>
mnemoc: schematics, BOM, everything
<oliv3r>
i do understand what lkcl is saying; he wants make a living out of it; he wants to recoup all the time/money he has spent in the last 18 years
<mnemoc>
afaik the next generation of radxa products will be truly oshw too
<oliv3r>
but history has also tought us; those who say they will release, tend not to
<oliv3r>
and while i get that tsvetsan and lkcl don't seem to get along perfectly; olimex IS pulling it off
<mnemoc>
tsvetan will be giving a talk about oshw on fosdem iirc
<oliv3r>
i thought it was opensource 'pc'
<oliv3r>
like building a pc with lime
<mnemoc>
ah, ok
<oliv3r>
the OSHW license isn't viral is it
<oliv3r>
i cold be wrong about his talk; that's just what the description sounded like
<mnemoc>
i think it is viral
<mnemoc>
derived work needs to keep the license
<mnemoc>
but only 90% sure
pwhalen has quit [Ping timeout: 272 seconds]
TomiK has quit [Remote host closed the connection]
<oliv3r>
i'm no EE, but i do know that serial lines, like sata, need to be balanced, equal in length, in proximety etc
pwhalen has joined #linux-sunxi
Wizzup has quit [Quit: Change me]
<nedko>
lkcl: could you please comment on this? ^
<mnemoc>
isn it designed by wits?
<oliv3r>
the imrpov? nah
<oliv3r>
that's designed by make.play.live
<nedko>
both the cpu and the io boards?
* nedko
is lost (too manu "brands")
<nedko>
*many
FR^2 has quit [Quit: und weg...]
jinzo has joined #linux-sunxi
jinzo has joined #linux-sunxi
nieuwbie has joined #linux-sunxi
<oliv3r>
nedko: eoma board I think was designed by wits/lkcl; i remember him blogging about it being done in kicad, but the memory tracings being done in ... witscad as it has to be 'just right'
<plaes>
what is witscad?
<oliv3r>
wits-cad?
<oliv3r>
like ki-cad, or auto-cad, or altium-cad :)
<oliv3r>
wow, this shows you why the LIME is so small, and how tiney those pins really are
<nieuwbie>
hey guys.
<Turl>
oliv3r: what shows?
<oliv3r>
Turl: blog link
<nieuwbie>
I'm trying to write board configuration for inet_3f board and I have problem with option. I used INet97F-II as an example but I don't understand one thing here
<Nyuutwo>
oliv3r: these sata traces, they can just caccidentaly work
tgaz has quit [Ping timeout: 246 seconds]
Night-Shade has joined #linux-sunxi
<nieuwbie>
according to description found in boards.cfg option filed has following syntax
<Turl>
oliv3r: ah :)
<nieuwbie>
config_hearder:CPP_OPTION
<nieuwbie>
if no value is specified it gets one (right? Correct me if I'm wrong)
<Turl>
nieuwbie: it works like -D on gcc
<Turl>
-D OPTION defines the option so it will work for #ifdef and the like
<Turl>
if you -D OPTION=n apart from being defined it gets value n
<oliv3r>
nieuwbie: not sure if we use CPP_OPTION in boards.cfg at all
<oliv3r>
Nyuutwo: it's sloppy design io
<Turl>
on boards.cfg you use "blah:OPTION,SOMETHINGELSE=n
<nieuwbie>
ok I can specify sun4i:INet_3F but I bet I'll get an error cause I didn't define INET_3F anywhere else
<Turl>
oliv3r: we use it extensively <.<
<Turl>
oliv3r: it's like the last column on boards.cfg. the uberlong one :)
<oliv3r>
Turl: yeah, but i don't think we use CPP_OPTION; don't recall seeing that specific one
<oliv3r>
nieuwbie: try git log -p
<Turl>
oliv3r: that was just a generic specification :p
<Turl>
like saying "-D OPTION"
<oliv3r>
nieuwbie: it should give you very easy additions, where no dram*.c was added
<oliv3r>
Turl: i missed that
<Turl>
nieuwbie: did you need to add a dram file?
<Turl>
with it you should add an entry on board/sunxi/Makefile to include it
tgaz has joined #linux-sunxi
<nieuwbie>
Turl: I think I do. Not sure. but I don't know how.
<nieuwbie>
ok
<Turl>
saying CONFIG_BLAHBLAH
<Turl>
then on boards.cfg you put sun4i:BLAHBLAH
<Turl>
(and any other relevant options)
<oliv3r>
i don't think he needed to add a memory file
<Nyuutwo>
downloaded gerbers ... why connector on that side of board, and theese curves?
<oliv3r>
nieuwbie said he found one just like his :)
<Nyuutwo>
in tar.gz even more werid placement/traces
<libv>
nieuwbie: did you add your board to the boards/sunxi/Makefile ?
<gzamboni>
i just forgot to buy that lime pins adaptor =( they are really small, but it looks like an awesome board, havent tested yet, but the fact of having all the pins of the A10 is awesome
<nieuwbie>
oliv3r: that dram looks like COBY_MID7042 COBY_MID9742 and PENGPOD1000
<oliv3r>
Turl: can you check if that link in your mailbox still works? i accidentally deleted it; i replaced it with the original, but owncloud doesn't show it in my 'trashbin'
<oliv3r>
nieuwbie: so copy one of those lines, put INET_3F in the name bit, but leave the .c file identical in the Makefile
netlynx has quit [Quit: Leaving]
<oliv3r>
Nyuutwo: looks like crap to bef air
<nieuwbie>
I added something like this "obj-$(CONFIG_INET_3F) += dram_sun4i_408_1024_iow16.o" to board/sunxi/Makefile
<nieuwbie>
<nieuwbie>
Turl: Should I also add SPL to options in boards.cfg?
<Turl>
oliv3r: says it's been deleted
<Turl>
nieuwbie: yes
<nieuwbie>
Turl: If I understand correctly - it means that SPL has to be build as well. So some targets dont need SPL?
<Turl>
nieuwbie: all of them need spl if you want to boot from sd
Wizzup has joined #linux-sunxi
<nieuwbie>
Turl: ok. thanks.
<Nyuutwo>
more looking, and not so crappy, equalizing lengths done, (need to send LCD/TTL) but no ground plane, traces near each other beacuse of i2c power supply, and not crossed traces (even just like paraell) on sides of board
<Nyuutwo>
just ethernet/sata looks crappy
nove has quit [Quit: nove]
parabyte has quit [Quit: Leaving]
<oliv3r>
Turl: good good;
<oliv3r>
Nyuutwo: LCD looks shit imo; all weird
Nazcafan has joined #linux-sunxi
<Nyuutwo>
they try to equalize length of traces, but for small lcds not necessary
TomiK|exil has joined #linux-sunxi
TomiK has quit [Read error: Connection reset by peer]
<oliv3r>
they could have equalized them prettier :p they look all crooked; but yes, not SO bad, just visually appaling
<oliv3r>
sata/ethernet look horrible
<oliv3r>
i'm supprised it works
orly_owl has quit [Ping timeout: 264 seconds]
<Nazcafan>
Hi, I just acquired a cubietruck, and I figured I would install fedora 20 as a first experience, any feedback on it?
orly_owl has joined #linux-sunxi
popolon has joined #linux-sunxi
<Turl>
Nazcafan: it should be about the easiest to install
<Nazcafan>
Turl, it's quite an experience for me, the README.pdf is in Chinese (which is not my spoken language) and it seems that Livesuit.run refuses to install on my debian testing laptop
<Nazcafan>
Linux distribution Debian does not supported!!!
<Nazcafan>
I guess I'll nick someone else's windows machine unless there is a workaround
<Turl>
Nazcafan: didn't you say you were to install fedora? o.O
TomiK|exil has quit [Read error: Connection reset by peer]
<Nazcafan>
Turl, yes, but my laptop runs debian
TomiK has joined #linux-sunxi
<Nazcafan>
Turl, I am jumping ship because of the systemd VS upstart debacle :-p
<Nazcafan>
on the FAQ, they say that the revision of the A20 has an importance, but I already applied the heatsink on the chip, so I will assume I have revision B
<Turl>
Nazcafan: ahhh I think I got what you're trying to do
<Turl>
Nazcafan: you're trying to install the cubieboard provided fedora via livesuit
<Nazcafan>
Turl, is that not a recommended way to proceed?
<Turl>
Nazcafan: I (wrongly) assumed you were to use the sdcard version - that's just dd + running a script
<Turl>
well, if you want easy, I'd use the sd version :)
<Nazcafan>
Turl, I have spent too long wondering whether I should buy a 32 GB sdcard on amazon or a 16 GB on a highstreet shop, so for now, I figured I would resort to flashing the nand