hno changed the topic of #linux-sunxi to: /Allwinner/sunxi development discussion - Don't ask to ask. Just ask and wait! - See | | Logs at
bsdfox has quit [Ping timeout: 264 seconds]
mcbrick has quit [Ping timeout: 264 seconds]
egbert has quit [Read error: Operation timed out]
egbert has joined #linux-sunxi
hipboi has joined #linux-sunxi
[7] has quit [Read error: Operation timed out]
TheSeven has joined #linux-sunxi
rings_IIV has quit []
Offshore has joined #linux-sunxi
rellla has joined #linux-sunxi
atiti has quit [Ping timeout: 264 seconds]
vi390 has joined #linux-sunxi
rellla2 has joined #linux-sunxi
rellla has quit [Ping timeout: 264 seconds]
rellla2 is now known as rellla
atiti has joined #linux-sunxi
rz2k has joined #linux-sunxi
<oliv3r> lkcl_: who dropped off?
<oliv3r> libv: yeah i wouldn't mind hacking on cederus, but time, plate; all that
<libv> oliv3r: :)
<oliv3r> libv: how's the heat in germany?
<oliv3r> i heard you guys had 40C the other week?
<libv> for nuernberg, the weatherfront arrived early enough on sunday to dampen that to less than the 36-37 we had the day before
<libv> now it's cool and rainy, which is much needed
<libv> and i currently have all windows open wide to try to cool the walls down. close the windows, and the temperature inside still jumps to 26°C
<libv> with the new temperature buildup starting today, and supposedly reaching another peak on sunday...
<libv> living without airconditioning
<libv> anyway, 38°C isn't the end of the world, it would just be a new record here
<oliv3r> yeah, our home was around 26-27 last week too
<oliv3r> but ehv is slightly more north so our max was around 35 i belive
<oliv3r> in the sun, the thermometer read 48C though ;)
<libv> yeah, it should be a few degrees less nearer to the coast indeed
<oliv3r> we'll be spending saturday on the entirity of the A3
<libv> ouch
<oliv3r> with Friday allready being prognosed as 33C
<libv> A3 has improved though
<oliv3r> i worry for saturday
<oliv3r> this year? or in recent years?
<libv> wuerzburg is not the big building site it was from 2007-2012
<oliv3r> been driving A3 for many years now :)
<oliv3r> oh god I was stuck im stau in wuerzburg in that timeframe
<oliv3r> one big ass construction site
<oliv3r> 1 bridge out of commission
<libv> and when one stretch was done, the next one was started
<libv> after wuerzburg you still only get dual lanes though
<oliv3r> can't wait to get past passau, then it'll be good :)
<libv> so if there is a lot of traffic you might get rather unlucky
<libv> saturday might be bad as school holidays start aug 1st
<oliv3r> yeah
<oliv3r> bayern has its last day friday
<oliv3r> luckly the rest of germany started last week and the weeks before that, bayern is the last bundesland to start
<libv> ah, ok
<oliv3r> so i'm planning to reach bayern around 15:00
<oliv3r> if i leave here really early, say 5 am, i'd be in bayern around 9-10
<oliv3r> which I dont' think will be the best time to be there ;)
<oliv3r> at 15:00 I might get lucky
<oliv3r> Once in austria it should be easy going
<libv> right
<oliv3r> big might ;)
<oliv3r> i'll wave when I pass nuernberg though ;)
<libv> my brother drove over from ghent 1.5 weeks ago on saturday... took him 9h
<libv> his drive back last sunday took him 7
<oliv3r> so 2 hours extra
<oliv3r> concidering the time of the year, that's not too horrible
<oliv3r> and 1.5 weeks ago, some other region's holliday started :S
<libv> yup :)
<oliv3r> but yeah, i've picked the worst time of the year to pass through DE
<libv> pretty much any weekend day in summer is bad
<libv> oliv3r: you're not taking a caravan, are you?
<oliv3r> HAhahaha
<rellla> oliv3r: the best comes last ;)
<oliv3r> no
<oliv3r> no caravan
<libv> oliv3r: well, 2ys ago i was visiting my family in belgium, and i took the latest train back
<libv> i had missed the last u-bahn, so i had to walk about 1.2km home, in nice august weather
<oliv3r> oh shit
<oliv3r> nice
<libv> i crossed the main road around the main station, and started walking
<libv> a car stops next to me
<libv> the window is wound down...
<libv> "entsschueldiggungg?"
<libv> in a real thick dutch accent
<oliv3r> lol
<libv> can you imagine
<oliv3r> chances of that happening
<oliv3r> did thy have a caravan?
<libv> you just drove 800kms from, i think amsterdam or something... you lost your way, and finally muster the courage to ask for directions
<libv> and, as a dutchman, you ask, in the middle of bavarian, the first person you see
<libv> and he turns out to be a flemish belgian :p
<oliv3r> alee!
<oliv3r> :
<oliv3r> ;)
<libv> he wanted to the way to his hotel
<oliv3r> did he offer you a ride?
<libv> +know
<libv> i answered "ge kunt me mij ook int nederlands klappen", in the flemish accent of my hometown (which i just practised all weekend)
<oliv3r> hahaha
<libv> that's when his brain shortcut
<oliv3r> lol
<libv> and he just repeated his entschueldigung and his question
<ganbold> where is the latest src of uboot for cubie2?
<oliv3r> in german?
<oliv3r> ganbold: github
<oliv3r> ganbold:
<ganbold> ok
<oliv3r> take the sunxi-current branch
<libv> so turned out that his hotel was 300m from my home, in a place next to where michael larabel ( stayed a month before
<libv> oliv3r: yes, german with a very thick dutch accent
<libv> so i told this guy and his buddy that if i had room in the back, i would show them where this street was
<libv> anyway... in the car, i got asked what i did in nue, i answered "moved for work, stayed."
<libv> i then asked the same from the dutchmen...
<oliv3r> libv: hahaha, the belgian scared him?
<libv> and apparently...
<libv> they were there to...
<libv> buy...
<libv> a...
<oliv3r> micheal was there for oktoberfest?
<oliv3r> lederhose?
<libv> no :)
<oliv3r> why would you go to nue to buy something
<libv> try again :)
<libv> what would a dutchman go halfway across the world for?
n01 has joined #linux-sunxi
<oliv3r> beer
<libv> a caravan!
<oliv3r> LOL
<oliv3r> bwahaha, ok that's funny
<libv> i nearly choked
<oliv3r> a german caravan
<libv> yeah, apparently... it's much cheaper down south
<libv> up north they are all really expensive, those secondhand caravans
<libv> so i kept my cool
<libv> guided them to their hotel
<libv> said "good luck"
<libv> then turned the corner
<libv> and laughed my head off.
<oliv3r> hahaha
<oliv3r> yeah i imgagine
<libv> it could not have been more stereotypical :)
<oliv3r> driving half way around the world, to get a better deal on a caravan ;)
<oliv3r> how's your german btw? Is it heavily bavarian influenced allready?
<libv> well, it is more franconian
<libv> pretty hairy in places, as i never had lessons, and i cannot be bothered too much with conjugations
<oliv3r> I always like when i start entereing bavarian territory
<oliv3r> they speak austrian-ish :p
<libv> that's proper bavaria
<libv> but as you might know, bavaria should've been 2 states
<libv> franconia and bavaria
<oliv3r> oh really? I do not know :)
<oliv3r> i know little of germany
<libv> but after wwii they did not want nuernberg to be the capital of a state anymore
<oliv3r> i did not know
<libv> my brothers gf stated as well that she found very little mention of the reichsparteitagesgelaende in touristic brochures
<libv> it should be swamped with tourists, but it sadly isn't. the most important historic site of europe, explaining quite a lot about the bad history of the last century, and nobody wants to talk about it
<libv> all the dutch come here for is to buy a caravan :p
<oliv3r> Hahaha
<oliv3r> my gf and I occasionally visit sites like those
<oliv3r> Kamp vught, Kamp Mauthausen
<oliv3r> she's been to auschwitz and Westerborg too, but that was before i met her ;)
<oliv3r> not sure if I want to visit auschwitz
rz2k has quit []
<oliv3r> libv: oh that does look interesting to go visit
<oliv3r> is it well preserved?
<oliv3r> mauthausen was extremly well preseved. was very impressive
<libv> i never visited a concentration camp yet, i really should
<libv> and what was started or halfway finished, is partly preserved
<oliv3r> kamp vught is nearly gone, they repurpoused the site over the years. of course now they regret it for its history value
<libv> the zeppelinfeld tribune is only a ghost of what it was, since the americans (rightfully) blew up the swastika (that's an image of wwii that everyone knows), and the germans blew up the columns in the 60s as they were not stable anymore
<libv> but the congresshall is a shell but pretty impressive, and it now has a good museum in there explaining the madness in great detail
<oliv3r> mauthausen I guess you could say is in pristine condition, but they swastika and eagle from the gates has been removed during the liberation
<libv> i wish miss merkel would visit it and then rethink her financial position towards countries like greece.
<oliv3r> it's one well worth seeing
<oliv3r> what is merkel's position atm?
<libv> well, just the german gouvernment is scared of spending, as the german population is scared of inflation (weimar republic)
<oliv3r> ah yes same as everywhere
<oliv3r> what really baffles me however
<libv> so they put the screws on greece and portugal and such
<oliv3r> in 2013, after we've seen wwi, wwii, god knows how many other wars
<oliv3r> people tend to forget that the greek, the spanish, you name it, are still human beings
<libv> which has the same effect, to a less extreme extent, as the weimar republic had, in the end
<oliv3r> and shouldn't be punished for the lies and bad choices their governement made
<libv> well, they should
<libv> but not this badly.
<libv> you just cannot justify these extreme measures
<libv> some things need to change, but it should not all be made right in such a short time
n01 has quit [Read error: Connection reset by peer]
<oliv3r> hopefully things will turn around for the better soon
<libv> wait and see, elections are coming up here
n01 has joined #linux-sunxi
<oliv3r> elections will only make things worse imo
<oliv3r> look at NL; some nutjob promisses the world (which he can't realize) and blames it on the foreigners
<oliv3r> 'hmm, i wonder where we heard that before'
<libv> hrm, to be honest, i now realize how cut off i am off the rest of europe
<libv> i cannot come up with the name of that dutch weirdly peroxide coloured hair anymore
<libv> he used to be in belgian news all the time
<libv> but he barely features in german news
<libv> this is perhaps a big problem in europe today
<libv> news from across europe is rare
<libv> and this means that european citizens don't feel connected
<oliv3r> Gert WIlders
<libv> ah, righto
<oliv3r> which is strange, since there's so little news nowadays, more eurpean news shouldn't be that hard
<libv> nobody talks about laws passed in the european parliament for instance
<libv> or very little
<libv> only about local laws
n01 has quit [Read error: Connection reset by peer]
<oliv3r> well a lot of countries do a lot of 'me me me' and dont' care about the rest
<oliv3r> france is really bad in that regard, and are well known for it :)
<libv> at least france is fully in the eu :)
<oliv3r> true, hollande isn't bad at all imo
n01 has joined #linux-sunxi
FR^2 has joined #linux-sunxi
eebrah|away is now known as eebrah
n01 has quit [Read error: Connection reset by peer]
n01 has joined #linux-sunxi
mcbrick has joined #linux-sunxi
eebrah is now known as eebrah|away
<ganbold> hm, what is this machid in uboot?
<oliv3r> should be setup correctly
<oliv3r> your kernel might have it's id set wrong?
<oliv3r> not sure
<oliv3r> check the irc backlog, machineID has been discussed before
<ganbold> well, what if different kernel?
<oliv3r> u-boot and kernel need to agree on the machine ID :)
<oliv3r> mnemoc: check your mail
<ganbold> does it matter if not linux kernel?
<oliv3r> what kernel are you booting?
<oliv3r> because of course it matters :)
<ganbold> I'm trying freebsd
<arokux1> :)
<oliv3r> ahh, well there are some people working on fbsd support, but don't know the A20 status, the fbsd kernel will of course also learn about the a20 machine ID
<oliv3r> so you have to add that to the kernel
<arokux1> oliv3r, does fbsd also support fdt?
<oliv3r> no clue
mcbrick has quit [Remote host closed the connection]
<hno> who pinge me?
mcbrick has joined #linux-sunxi
<mnemoc> oliv3r: what exactly on my mail?
<oliv3r> mnemoc: hansg, pull request
<mnemoc> ok. i'll try it
<oliv3r> appearantly things got fixed
<mnemoc> why the heck I *always* mistype hansg
<mnemoc> fatal: 'hasng' does not appear to be a git repository
<hno> mnemoc, at least the right characters, just ordering error. Typing too fast?
<mnemoc> :)
<oliv3r> using qwerty? Its intention has always been to slow you down :)
<oliv3r> try dvorak! (i am qwerty like everybody else)
mcbrick has quit [Remote host closed the connection]
mcbrick has joined #linux-sunxi
mcbrick has quit [Quit: Leaving]
<mnemoc> oliv3r: no improvement
<oliv3r> :(
<mnemoc> usb/gadget/android.c is totally f*ed up now
<oliv3r> git bisect; blame!
<oliv3r> blame and shame
fredy has quit [Excess Flood]
<arokux1> is there actually a software where one could (easily) construct a board computer with desired specs.
fredy has joined #linux-sunxi
<arokux1> and then order its printing
<arokux1> manufacturing*
<arokux1> "easily" like with lego :)
<oliv3r> with a10 preinstalled
<arokux1> oliv3r, not necessarily..
<oliv3r> google Co
<oliv3r> CoM
<arokux1> a SoC should be also an option there
<oliv3r> computer on module
<arokux1> yes, I know this, but not so many boards out there where you could actually plug CoM
rz2k_ has joined #linux-sunxi
<arokux1> oliv3r, plus the CoMs seem to be not compatible with each other
tzafrir has quit [Ping timeout: 268 seconds]
wingrime has quit [Ping timeout: 246 seconds]
rz2k_ has quit [Quit: leaving]
wingrime has joined #linux-sunxi
LooneyBin has joined #linux-sunxi
<LooneyBin> Hey all wondering if anyone knows the Kconfig necessary for a Mele A1000G (A10) for full memory reserves (don't need video it's a headless server). Any info is much appreciated. I'm about the compile sunxi3.4
<oliv3r> LooneyBin: check the wiki
<LooneyBin> Thanks oliv3r I checked the wiki and found info at Kernel_arguments. Last time I did this I remember the memory reserves had to be disabled during kernel compilation and kernel arguments would have no effect. Can I just use these arguments now?
hipboi has quit [Quit: Leaving]
<LooneyBin> Last time I had to disable things like CONFIG_MALI,
<LooneyBin> CONFIG_SUN4I_G2D. Is this still necessary? Sorry if this info is around but I can only seem to find outdated info
<oliv3r> well if you don't use it, it'll result in a smaller binary
<LooneyBin> Well I'd prefer a smaller binary if possible, but don't know the options. Do the above video CONFIGs still have the same effect?
<oliv3r> i belive so, but not sure
<oliv3r> only 1 way to find out ;)
<mripard_> n01: ping?
<LooneyBin> Great thank you I'll have another geez online.
eebrah|away is now known as eebrah
<n01> mripard_: pong
<mripard_> n01: hi :)
<n01> hi :)
<mripard_> will you have some time to address the few comments on your watchdog driver ?
<mripard_> we're pretty close to getting it merged it seems
<mripard_> so it would be great to have it in 3.12
<n01> yep, I'll submit it this evening/night
<mripard_> oh, great :)
<n01> sorry for the delay
<oliv3r> mripard_: have you heard anything from greg?
<mripard_> oliv3r: nope
<oliv3r> he said 'give me a few days'
<oliv3r> would it be rude if i'd mail for a status update so I can continue?
<mripard_> n01: no problem, actually, I was more concerned because I told you I could do the changes if you didn't have enough time
<mripard_> and I'm leaving at the end of the week for holidays, without much internet access
<oliv3r> i'm leaving saturday :D
<oliv3r> but only gone for a week, with excellent inernet access :p
<n01> those are really small changes, I can do it. thank you
<mripard_> so I just wanted to know if I needed to handle it this week, or if you could do it :)
<mripard_> oliv3r: where are you going ?
<oliv3r> austria :p
<oliv3r> visiting relatives :)
<oliv3r> you?
<mripard_> south-eastern france
<mripard_> visiting relatives (and the sea) :)
<ganbold> where is the definitions of arm CPUs in linux src?
<mripard_> what kind of definition are you looking for?
<ganbold> cpu id
<ganbold> cpu id and rev of A20
tzafrir has joined #linux-sunxi
arokux has joined #linux-sunxi
<oliv3r> vacation in own country, well france is big enough to do that well
Black_Horseman has quit [Remote host closed the connection]
<oliv3r> mripard_: he wants to use freebsd kernel :)
<mripard_> ganbold: hmmmm, presumably in head.S then
<ganbold> ok
<oliv3r> i heard of BSD work happening on A10
arokux1 has quit [Ping timeout: 264 seconds]
<oliv3r> so you can use that as your reference point
<ganbold> :) that is me
<mripard_> or in setup.c
<atiti> hm i gotta cedar again
marzubus has joined #linux-sunxi
<oliv3r> atiti: you got to cedar again?
marzubus_ has joined #linux-sunxi
marzubus has quit [Client Quit]
<atiti> yea, still havent got my code to decode h264 properly
<marzubus_> Hi Guys, im trying to compile a armhf kernel for a MK802 from ( tried 3.0 and 3.4 branch ) and I get error: arch/arm/kernel/head.S:90:2: error: #error "unsupport chip" #error "unsupport chip". Im using toolchain from
<marzubus_> using command "make ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf- uImage modules"
<marzubus_> have I missed something really stupid?
<zumbi> marzubus_: did you confiure the kernel? (make [..] menuconfig)
<marzubus_> zumi, I coped a .config from a working kernel, then ran menuconfig
<marzubus_> im retrying my copy, config and compile quick to be sure.
<oliv3r> make (linux-)config anyway
<marzubus_> @zumi, I seem to be able to get further now.
<oliv3r> you have to make some config after copying .cofnig
<oliv3r> but this is talking about the linux kerenl, you also talk about compiling u-boot
<marzubus_> im only doing the kernel build.
<marzubus_> Its building now at least. my next problem is the modules_install stage. I get a message saying DEPMOD 3.0.76 and a warning that I need to install modules-init-tools. I have that installed but as part of my OS packages. do I need a special ARM-ified versoin of that ?
<marzubus_> shows the exact output
<oliv3r> you might want to try the 'sunxi-bsp' from that same wiki page
<oliv3r> it does some checks, calls the proper commands etc
<marzubus_> ok.
<zumbi> marzubus_: modules-init-tools I think is harmless
<zumbi> marzubus_: you might need that on target rootfs
<marzubus_> baseically I have a working debian wheezy LXDE which I grabbed from: . I need to rebuild the kernel so I can have it talk to my arduino. hence this huge mission.
<marzubus_> but I need to add those modules to the rootfs. and since I cant capture them with modules_install, im a bit lots
<marzubus_> ill checkout the BSP way, but since I only need a kernel I thought id avoid re-doing the whole rootfs and u-boot setup.
<oliv3r> i find the sunxi-bsp the easiest solution to build u-boot and/or kernel
<marzubus_> ok. ill give it a bash
<oliv3r> it can make rootfs or reconfigure a rootfs (if it understands it) but that's a feature I don't use
<marzubus_> thanks.
<oliv3r> also, the depmod warning you can safely ignore ;)
<marzubus_> ah great!
Tsvetan has quit [Ping timeout: 268 seconds]
<bfree> marzubus_: what config options do you need for your arduino? just wondering if "my" deb kernel package config covers it, or if I should pencil in trying to enable it the next time I build new ones
<Turl> bfree: one of the USB serial things I think, the ones that produce ttyACM0 device nodes
<marzubus_> bfree: I need a kernel built with: ONFIG_USB_SERIAL_FTDI_SIO=y
deasy has joined #linux-sunxi
<bfree> Turl: marzubus_: thanks and ooops, sigh ... ... CONFIG_USB_SERIAL_* not set in my config (though CONFIG_USB_SERIAL=m) :-/ a little strange as the distro kernel config I used as my template to bloat it out has (most of) them
<marzubus_> btw, here is the project this is for: . Its a mini rover. Pololu Zumo chassis, arduino for running the motors. 3G modem and a MK802 for dealing with the video and 3G connection.
fredy has quit [Excess Flood]
fredy has joined #linux-sunxi
<oliv3r> marzubus_: very cool
<oliv3r> oh shit, temperature prognosis for friday evening: 34.7 degree's
<marzubus_> in UK?
<oliv3r> NL :)
<Turl> oliv3r: 10-16C here for friday :)
<oliv3r> are your winters warmer or colder then usual?
<mripard_> tssss, we had 39 last saturday :)
<marzubus_> 24C in Stockholm
<Turl> oliv3r: I dunno, I haven't been that many winters in here to tell :)
<oliv3r> mripard_: in south france? :p
<oliv3r> Turl: where are you originally from?
<mripard_> oliv3r: yep, in toulouse
<oliv3r> i think germany had 40C last week
<Turl> mripard_: actual 39C? or "feels as 39C"?
<marzubus_> I could live in toulouse. Im from RSA, sweden is cold!
<mripard_> Turl: actual 39C
<oliv3r> Turl: the main 24 Hz clk, is that something that is readable, queryable, 'get' able or we just assume it's 24 MHz?
<Turl> mripard_: I'm hoping your humidity is low then :P
<oliv3r> marzubus_: but sweden is also beaitful
<mripard_> Turl: yeah, it's around 50-60% these days
<Turl> oliv3r: it's an oscillator :P that means it outputs a fixed freq
<marzubus_> Sweden is great! winter is just too dark. like 5 hours of light a day! Summer is awesome though! 19 hours of day!
<Turl> oliv3r: and document and code mention it being 24MHz, so it's that then :)
<oliv3r> friday we'll have 65% humidity, today we have 86% ;)
<mripard_> marzubus_: haha, I trust you :)
<oliv3r> Turl: well you can always overclock it by putting a faster osc. in :p 25 MHz! :p
<Turl> lol
<mripard_> there's too much snow and not enough slope for me in sweden :)
<oliv3r> Turl: i'm just asking, because I need to make some calculations based on that 24 MHz clock, so I can make a nice define, or re-use an existing thing
<Turl> oliv3r: I think it's internal to the SoC though?
<oliv3r> mripard_: you enjoy sking?
<oliv3r> Turl: nope, external SMD crystal
<mripard_> oliv3r: yep
<Turl> oliv3r: I thought the external one was 32k
<Turl> oliv3r: anyway, if it's external why don't you just take a look at yours? :P
<oliv3r> i learned how to ski before I kenw how to ride a bike :p i'm not really amazingly good at it though :)
<mripard_> oliv3r: i've been skiing since i've been able to stand on my feet :)
<oliv3r> Turl: i want to le tthe PC do the math, always :p
<oliv3r> Turl: i KNow it's 24 MHz
<oliv3r> Turl: but why re-write #define CLK_SLK(24000000) if something central allready exists :p
<marzubus_> I tried snowboarding when I moved to Europe. broke ribs. :/
<oliv3r> mripard_: yep, born in the mountain area of austria
<Turl> I never went skiing :P I was on a snowy mountain once, but went on some kind of thingy where you sit and go down it
<Turl> oliv3r: clk_get(...) ? :P
<mripard_> oliv3r: ah, not me :)
<mripard_> but we went skiing with my parents every winter during christmas vacations
<mripard_> and now that the mountains are not that far away...
<mripard_> Turl: I actually considered coming to Argentina this summer to ski :)
<oliv3r> i tried snowboarding and got lessons, and while it's cool, I think i prefer sking
<wingrime> mripard_: I write dma-engine driver skeleton and have no idea how test driver (only want see printks)
<Turl> mripard_: to Las Leñas?
<oliv3r> Turl: i don't know, i never worked with the clocks, but i need to obtain the freq. not setup a clock or configure a clock, i don't think you can configure the 24MHz clock :)
<Turl> clk_get_rate(osc24M) :p
<Turl> or whatever clock you're interested in :)
<oliv3r> Turl: i think that's what i'm looking for
<mripard_> wingrime: you'd need a driver that use DMA
<oliv3r> but I think i might not use it afterall, i need to fill a table so want to do it with the preprocessor
<Turl> mripard_: there's a DMATEST driver
<mripard_> mdp's plan is to write an SPI PIO driver, and then use SPI with DMA
<mripard_> I believe it's a pretty good plan
<Turl> mripard_: emac can DMA too
<mripard_> I heard from him that we should expect the SPI PIO driver pretty soon now
<mripard_> Turl: yeah, but EMAC is much more complicated :)
<Turl> it just uses dma for tx from what I recall
<oliv3r> wouldn't it be usefull to use dma for rx aswell?
arokux1 has joined #linux-sunxi
<oliv3r> performance wise i mean?
<mripard_> oliv3r: I actually like snowboarding when doing off-track riding
<mripard_> for on-track, it's just meh.
<oliv3r> but as a computer geek; i do not excel at physical activies
<oliv3r> i cycle to work, 12 km each way on my tour-bike each day, and that's plenty for me :p
arokux has quit [Ping timeout: 264 seconds]
<marzubus_> so im trying to build the kernel via the sunxi-bsp "easy" way. but I get this error. whenever I "customize" the .config to include anything.
<oliv3r> lets see
<marzubus_> seems weird that it cant migrate my oldconfig
<oliv3r> it's probably best to save your .config somewher easily accessible
<oliv3r> then make linux-config CROSS_COMPILE=arm-pc-gnueabi-
<oliv3r> (you can modify the Makefile and adjust CROSS_COMPILE there to even ommit cross_compile
<Turl> oliv3r: arm-pc? o.O
<oliv3r> then load your saved config
<oliv3r> Turl: i'm just typing stuff :p i'm sure he gets what i mean
<oliv3r> Turl: mine is actually called arm-pc-linux-gnueabi-
<marzubus_> hehe .;)
<Turl> oliv3r: damn gentoo ;)
<oliv3r> Turl: hahah, you can fill in the name of your crosscompiler yourself when installing it :p
<oliv3r> marzubus_: anyhow, with the bsp, your config/output is seperated from your source tree
<oliv3r> marzubus_: your sourcetree will always be the same, no code/config is stored there, instead all output/config goes/comes from build/config_name/
<oliv3r> sun4i_defconfig-linux in your case
<marzubus_> aha
<oliv3r> so do cd sunxi-bsp/linux-sunxi; make mrproper
<Turl> oliv3r: was n01's watchdog merged?
<oliv3r> Turl: not yet, mripard_ asked him to fix the small comments left so that it could go into 3.12
<oliv3r> for SID, i'm waiting for greg to respond to mripard_ request/bug
<oliv3r> since i'm dependant on that fix
<oliv3r> if there's a bug in sysfs, sid should be done, if there's a change, i need to adapt
<n01> Turl: I didn't have time to work on it
<mripard_> oliv3r: you can ping gkh
<oliv3r> and since greg wants to merge the sysfs stuff in 3.11, i kinda have to wait for 3.12 so that every has it
<oliv3r> mripard_: ok will do, i asked you before, but we got distracted with frozen water ;)
<mripard_> did we ?
<mripard_> sorry :(
<oliv3r> mripard_: why, frozen water is awesome :)
<oliv3r> mripard_: on the upside, i should have pwm ready for first review and question session :p
<mripard_> great
<oliv3r> after that some tests and then the ML horror :p
<wingrime> mripard_: why not add sun7i to compatible and not use sun4i drivers in a20 dt....
<oliv3r> wingrime: because in dt, we use the first arch/plat where a device occured
<oliv3r> so if we have sunxi_dma.c, compatible should be allwinner,sun4i-dma
<mripard_> oliv3r: if the IP are the same
<wingrime> oliv3r: you not right here, any driver can have many compatible
<wingrime> mripard_: dma are same , but routes not
<wingrime> mripard_: end point address
<oliv3r> wingrime: well lets say PWM for example, the way I understand, sun4i and sun5i have the exact same PWM controller
<oliv3r> so you only need sun4i-pwm
<wingrime> oliv3r: no
<oliv3r> mripard_: ^
<wingrime> oliv3r: you can add not single compatible
<mripard_> you're both right.
<mripard_> if the IP is the same, you should use the same compatible
leowt has joined #linux-sunxi
<wingrime> you can add sun4i-dma and sun5i-dma string to match array
<mripard_> yet, if something differs, DMA routes, gpio pin sets, clock definitions, we should use different compatibles
<oliv3r> now sun5i is the SAME IP, but didn't route out all of the pins, so technically it's a different PWM (only 1 channel as opposed to 2)
<oliv3r> so then we have sun5i
<wingrime> than you can use in sun4i dt sun4i-dma version , and use sun5i-dma in sun5i dy
<wingrime> *dt
focus_well has quit [Quit: Ex-Chat]
<wingrime> oliv3r: driver can read own device name and make some diffences
<mripard_> wingrime: yes, but if there's no difference, there's no point in doing so
<wingrime> mripard_: estetic
<wingrime> mripard_: it looks stange in dmesg
focus has joined #linux-sunxi
<wingrime> mripard_: better use sunxi when no differences
<mripard_> no.
<mripard_> we got rid of sunxi, with reason
<mripard_> otherwise it's just a mess
<wingrime> mripard_: also, question , mainline wtd driver temorary stud?
<mripard_> ?
<wingrime> sun4i-wdt
<wingrime> not looks normal
<wingrime> mnemoc: ping
<mnemoc> wingrime:
<wingrime> mnemoc: are you readed backlog about cedarA
<wingrime> ?
<mnemoc> wingrime: no, sorry.... been mostly away from irc for over two months already :(
<mnemoc> wingrime: summary?
<wingrime> mnemoc: looks likes we have ac3 and dts audio decoders and jpeg encoder
<wingrime> mnemoc: also audio blob voliate gpl
<mnemoc> \o/
<wingrime> mnemoc: ACE driver in kerne
<oliv3r> it's from sun3i android drop mostly
<wingrime> a10 fast enought for audio
<oliv3r> dead silicon is a waste ;)
<Turl> mripard_: I'm updating sunxi-devel
<Turl> mripard_: I've got the clocksource, i2c on A10S, led names, wemac timeout & watchdog, anything else I'm missing?
<mripard_> sid?
<mripard_> plus a "few" patches you're going to see pretty soon
<mripard_> the A10s clocks
<wingrime> mripard_: also cau you add emac to sun7i dts
<mripard_> wingrime: I've not tested it yet, but that's the plan at some point yes.
<wingrime> mripard_: 90% it will work without any change
<oliv3r> wingrime: have you tested it?
<wingrime> oliv3r: no
<wingrime> oliv3r: guess
<wingrime> mripard_: also ac97 driver in a good shape for mainline
<mripard_> wingrime: i'm concerned about the 10% then
<wingrime> mripard_: if I clean it enought
<Turl> mripard_: but sid crashes, what's the point? :p
<mripard_> Turl: v4 worked
<Turl> oh, it did?
<mripard_> later versions didn't
<Turl> I thought it always crashed
<wingrime> mripard_: how about , I clean it and add dts support
\\Mr_C\\ has quit []
<wingrime> mripard_: also there nice soc-audio-dma-engine support
<mripard_> wingrime: so you'd work on audio instead of dma ?
<wingrime> mripard_: I need users for dma first
<wingrime> mripard_: also , ac97 driver lacks copyright
<mripard_> Turl: you should change your name :)
<mripard_> wingrime: and the ac97 codec works in PIO?
<wingrime> I think can
<Turl> mripard_: yeah :p
<mripard_> Turl: Emilio Lopez overflow :)
<Turl> mripard_: btw, about the A10S clocks, I had a look at both of them patches, and overall they looked good but I think we should start standarizing the gate names :)
<mripard_> why?
<mripard_> I tried to stick with the datasheet the more I could
<Turl> mripard_: for consistency mostly
<Turl> usbotg vs usb0
<mripard_> what inconsistencies did you find ?
<mripard_> ah
<mripard_> hmmm
<oliv3r> wingrime: i'm not so confinced it's in that good shape, it looked horrible last time i saw it
<oliv3r> wingrime: though hasng has been banging it in reasonable shape
<wingrime> oliv3r: not *whole* audio support
<oliv3r> sunxi_codec.c you speak of?
<wingrime> oliv3r: yes
<oliv3r> yep that didn't look so clean :p
<wingrime> also , no normal copyrights author anonimous
<wingrime> oliv3r: code in good shape for you?
eebrah is now known as eebrah|away
<oliv3r> i didn't think it looked too great
<wingrime> oliv3r: but if clean it
<oliv3r> a lot lot of work :)
<oliv3r> but sure, if you want to :)
<oliv3r> i'd pick a more simple driver initially
<oliv3r> but your time is yours to spend :)
<wingrime> oliv3r: only 12 registers
<oliv3r> hmm, it looked more complex then that :)
<oliv3r> 12 registers, ~20 settings each
<oliv3r> 240 options :p
<oliv3r> though you could ommit audio in at first, only worrya bout output
Superpelican has joined #linux-sunxi
\\Mr_C\\ has joined #linux-sunxi
<wingrime> oliv3r: but also can you find some USEFUL driver
<oliv3r> I find IR a simple usefull driver ;)
<wingrime> oliv3r: that realy important for manilne
<oliv3r> but lets look at the list
<wingrime> oliv3r: axp209 not so important
<wingrime> (only on tablets)
<oliv3r> mmc
<oliv3r> should work in PIO mode
<Turl> wingrime: axp is important for DVFS
<oliv3r> in u-boot it does pio mode only i think
<oliv3r> yeah but axp is quite complex
<Turl> I never said it was simple :p
<oliv3r> Turl: wingrime asked for a reasonable usefull driver
<oliv3r> i say, IR, MMC
<oliv3r> i2c = done; spi on its way
<oliv3r> lradc is usefull, but someone wanted to do that? (not on wiki is not claimed imo :p)
<oliv3r> also lrdac -> input connector is usefull
<oliv3r> for simple and usefull, that's it, MMC being the prime candiadte :)
<oliv3r> wingrime: MMC driver is very usefull :D
<mripard_> I plan on resuming my work on MMC as soon as I'm done with the A20/A31 bringup
<mripard_> LRADC, IR, would be indeed good candidates
<Turl> LRADC would be a good wakeup source :)
bsdfox has joined #linux-sunxi
rellla has quit [Ping timeout: 264 seconds]
BagOBones has joined #linux-sunxi
<BagOBones> Hey all, I'm having trouble using the 8192cu module, it loads but no wlan0 interface exists. Any ideas?
<BagOBones> I get only this "usbcore: registered new interface driver rtl8192cu"
<BagOBones> Is there a way to see what the module is doing?
<BagOBones> Also I compiled stage/sunxi-3.4 would this make a difference? Maybe 3.0 would work.
<rm> BagOBones, what does lsusb show?
<BagOBones> Nothing related, although it isn't a usb device
<BagOBones> Not sure how to see if it has even seen the device. dmesg shows nothing useful
<BagOBones> Is there something I need in the kernel config?
<Turl> BagOBones: 8192cu *is* an usb device
<mripard_> Turl: could you comment on the A10s stuff?
<mripard_> to point out the inconsistency you found?
<Turl> mripard_: sure, I'll do in a bit
<mripard_> Turl: patchbomb! :)
<BagOBones> I'm using a Mele a1000g is there another module I should be using?
<rm> BagOBones, most likely you are using a script.bin that does not enable the onboard USB port (and the WiFi in it) by default
<rm> try finding one for Mele A1000G, or decode yours using "fexc" and change in there somewhere usb_init_something to = 1
<rm> :)
<BagOBones> Ah okay so it is a usb device in side my apologies. Yeah you're right I forgot about the script.bin I'll check it out cheers.
<rm> another alternative is to use not "rtl8192cu" but "8192cu", the "Allwinnerized" driver that "manually" enables that port
<BagOBones> Yeah I'm using that module at the moment. Does these mean chancing script.bin won't fix the problem?
rellla has joined #linux-sunxi
<rm> BagOBones, by your dmesg you are using "usbcore: registered new interface driver rtl8192cu"
<rm> whatever
<rm> change dmesg anyway
<rm> er
<rm> change script.bin anyway
<rm> I find it easier to work with when you see the WiFi device in "lsusb" from the start
<rm> no matter if some module has loaded or not
naobsd has joined #linux-sunxi
BagOBones has quit [Ping timeout: 264 seconds]
LooneyBin has quit [Ping timeout: 264 seconds]
tinti has joined #linux-sunxi
LooneyBin has joined #linux-sunxi
BagOBones has joined #linux-sunxi
hansg has joined #linux-sunxi
hansg has quit [Client Quit]
leowt has quit [Quit: leowt]
tkoskine has quit [Quit: brb]
tkoskine has joined #linux-sunxi
atiti has quit [Ping timeout: 264 seconds]
wingrime has quit [Quit: Lost terminal]
<BagOBones> rm: Bingo thanks a bunch
marzubus_ has quit [Ping timeout: 250 seconds]
wingrime has joined #linux-sunxi
leowt has joined #linux-sunxi
<Turl> mripard_: I'm having a look at the patches
<Turl> mripard_: I'm not such a big fan of those expanded compatible declarations you made :p
<mnemoc> Turl: are you interested in android @ 3.4 at all?
bsdfox_ has joined #linux-sunxi
<Turl> mnemoc: I am, but I haven't touched android @ sunxi for ages :( mainline got all my time lately
<Turl> mnemoc: easier to test I must say :p
leowt_ has joined #linux-sunxi
<mnemoc> Turl: stage has broken usb/gadget/android.c on build time ... and hansg only cares about fedora's config obviusly
bsdfox has quit [Ping timeout: 268 seconds]
<wingrime> mripard_: wtf, lradc have no copyright too
<wingrime> MODULE_AUTHOR(" <@>");
<Turl> mnemoc: sigh :(
<Turl> mnemoc: what config are you using to test?
<mnemoc> defconfigs
leowt has quit [Ping timeout: 264 seconds]
leowt_ is now known as leowt
<Turl> mnemoc: yeah but which one :)
<mnemoc> all of them
<Turl> all of them are broken?
<wingrime> mnemoc: you can add usb-next branch ))
<mnemoc> wingrime: and reset the current stage/sunxi-3.4 ?
<wingrime> wingrime: cherry-pick
<mnemoc> getting it fixed sounded more appealing....
<wingrime> mnemoc: better you deal with ML for first
<Turl> mnemoc: if you let me know which tree/branch and config is broken I can try to come up with a fix
fredy has quit [Excess Flood]
<mnemoc> wingrime: problem is stage/sunxi-3.4 is already broken. i trusted too much I pushed the previous pull from hansg without testing
<mnemoc> Turl: stage/sunxi-3.4 or hansg's sunxi-3.4-for-amery and sun4i_crane_defconfig for example
<wingrime> mnemoc: do reset and apply working patches at least
<wingrime> mnemoc: than after you deal with mainline
dwilkins has quit [Ping timeout: 256 seconds]
rellla has quit [Quit: Nettalk6 -]
fredy has joined #linux-sunxi
FR^2 has quit [Quit: Connection reset by peer]
<wingrime> mnemoc:how plan?
<mnemoc> wingrime: i won't touch mainline. there is enough good people there. i care about sunxi unification/cleaning/improving 3.4 to be used by others to mainline sunxi
<wingrime> mnemoc: most important that usb merge patch not ONLY usb merge
<mnemoc> unfortunatelly I hardly find energy/time to do so atm.
<mnemoc> wingrime: let's stone techn_
<wingrime> mnemoc: actaly thats why we use MailList
<wingrime> ok
<wingrime> mnemoc: reset + deal with maillist
<mnemoc> yes, absolutely, but i'm over 2k mails behind there
n01 has quit [Ping timeout: 264 seconds]
<wingrime> mnemoc: after usb merge patch must be sended to ML
<mnemoc> but ok. if after finishing installing the new server there is no fix for gadget/android, i'll reset stage
<wingrime> mnemoc: usb patches must be in separated branch untill it mature
<wingrime> mnemoc: that other way fix it and keep stange working
<wingrime> *stage
<mnemoc> i'll wait for 2) until i finish the new server. if not, 1)
dwilkins has joined #linux-sunxi
<wingrime> mnemoc: webgit)
<wingrime> mnemoc: kernel reference)
<mnemoc> wingrime: new server will have cgit and lxr, yes
<wingrime> mnemoc: cool
<wingrime> oliv3r: fex key on cb are lradc?
woprr has joined #linux-sunxi
<woprr> lo
<woprr> Q1: (DVB) TS controller/GPIO, MPEG decoder Programming info for Allwinner A10/A20? Project ist direct attachment of dvb-receiver modules to Boards supporting the Transport Stream Controller's TSx inputs.
Superpelican has quit [Ping timeout: 264 seconds]
<woprr> (CSIx in GPIO multiplex mode "multi3", TSx in mode "multi2")
atiti has joined #linux-sunxi
<wingrime> woprr: we lack many docs in this area
<woprr> expected that, thx :) Reading in Basics from
<Turl> woprr: maybe of use,
<woprr> thx
<Turl> mnemoc: do you have a link to hansg's repo handy?
<Turl> nvm found it :)
<mnemoc> :)
<wingrime> woprr: also we have info for MPEG decoder in wiki
<woprr> nice, thx
<woprr> HDMI (A/V) status?
<wingrime> woprr: in mainline no anything around
<woprr> ok thx
leowt_ has joined #linux-sunxi
<Turl> mripard_: looks like a horrible mismerge
leowt has quit [Ping timeout: 260 seconds]
leowt_ is now known as leowt
<wingrime> woprr: ?
<woprr> I'm impressed... well, in the first stage I'll be satified with MPEG2 SDTV, got no dvb-s2 tuner yet
<wingrime> woprr: what you defenetly want to do?
<woprr> and alle HDTV is HD+ encrypted here, won't pay for
<woprr> direct attachment of dvb-receiver modules to Boards, decode MPEG2-TS + audio, output with HDMI to DD-receiver and TV, record with vdr
<woprr> Boards=Allwinner TSx inputs
BagOBones has quit [Quit: Leaving]
<wingrime> woprr: we have no info about TSx inputs
LooneyBin has quit [Quit: Leaving]
<wingrime> woprr: also. I have not seen any (!) driver of it writen by vendor
<woprr> switching between GPIO multiplex modes?
<woprr> driver for the TS controller?
<wingrime> woprr: TS
<woprr> ok
Yuuki has quit [Quit: Page closed]
<woprr> Will ask some maintainer of if (s)he can get the specs from Allwinner
<woprr> This could become a major use case, avoiding unneccessary "round trip" with dvb-usb devices
<woprr> SMT addon boards with dvb-receivers are cheap to build
<libv> ssvb: are you running your sunxifb on odroid btw?
<oliv3r> wingrime: fex key? fel key is dedicated pin
<wingrime> oliv3r: there anything connected to lradc on cb2
<oliv3r> mripard_: didn't know where working on mmc
<Turl> oliv3r: it's on the wiki isn't it?
<Turl> oliv3r: as "SDIO driver"
<libv> oh, just works, nice.
<libv> ssvb: my preliminary lima dri driver is now also clearing a buffer (which is easy on mali, you always have to attach a clear shader and a colour to every fragment job) on odroid
jemk has joined #linux-sunxi
<Turl> libv: dri, woooot :)
<mripard_> Turl: ?
<libv> very hackish, very early
<Turl> libv: and very cool :) keep up the great work
<Turl> mripard_: ? what did I miss? :)
<libv> anyway, it is a mess on the standard mali driver, but works on ssvbs sunxifb
<mripard_> Turl: I don't know, you tell me :)
<mripard_> 19:07:53 < Turl > mripard_: looks like a horrible mismerge
* libv can now try to proceed with plugging the binary compiler into mesa
<Turl> libv: the standard mali X driver was crap wasn't it?
<Turl> mripard_: ah, that was directed at mnemoc :) autocomplete derp
<ssvb> libv: kind of, but I probably broke it a bit with the latest changes (r3p0 window resize bug workaround)
<libv> well, it would be interesting to find out what went wrong there
<libv> ssvb: this seems to be a may 22nd checkout
<libv> ssvb: it works for now, and i do not want to ruin it for the time being :)
<ssvb> libv: well, then I guess it should be ok :)
<mripard_> Turl: ok :)
<Turl> mripard_: you never replied to my criticism of compatible tables though :P
<ssvb> libv: I'm currently hacking the driver to get proper double buffering with vsync for sunxi with r3p0 mali drivers
<ssvb> libv: it works fine in a quick and dirty prototype code, now I'm just trying to push a clean implementation
<libv> ok, so you are also poking at disp?
<libv> or is there a magic disp ioctl for vsync you are hooking off?
<mripard_> Turl: ? which one ?
<ssvb> libv: basically, this is a simple sunxi disp demo code -
<mnemoc> Turl: thank you! your fix fixed all the defconfigs
<mnemoc> pushing
<ssvb> libv: (using thin wrappers for ioctls, but the idea should be clear)
<ssvb> libv: any changes to disp layers take effect only at frame boundaries, and this makes everything quite simple
<woprr> wingrime, since Allwinner is an one chip media solution, there may be not much of a driver necessary, I assume setup of my "filter graph" from TS to HDMI above could be done by setting up a few registers only and the data flow is handled internally by A10/A20?
<Turl> mripard_: that I wasn't a fan of expanding the compatible tables to ocuppy 4 lines per compatible :p
<woprr> + the dvb frontend driver
<Turl> mnemoc: yw :)
<mripard_> Turl: I never received that mail :S
<ssvb> libv: that's why we also have proper vsync for XV -
<Turl> mripard_: that's because I said so over IRC :p
<libv> ssvb: ah, so you are using the disp layers for that
<mripard_> Turl: ah, this is why :)
<mripard_> this week-end?
<ssvb> libv: yes, the same also works for DRI
<Turl> mripard_: a couple of minutes/hours ago, after the patchbomb :)
arokux has joined #linux-sunxi
<oliv3r> mripard_: see greg's mail?
<ssvb> libv: actually we can even have triple buffering for DRI, but with certain quirks to take care of
<libv> oh, so you are using the FBIO ioctl
<mripard_> Turl: ah, yes, missed it indeed.
<mripard_> sorry
<arokux> mripard_, what is the best soc code in the mainline, to follow style etc?
<mripard_> arokux: freescale's or tegra, probably
<libv> i found that this delivered not too nice results when switching main fb offsets
<mripard_> oliv3r: yeah
<mripard_> oliv3r: I got a free beer \o/
<arokux> mripard_, thakns
<libv> it just moved the glitch to a fixed point on the screen
<mripard_> Turl: actually, I'm not a fan of one big wide line either
<wingrime> woprr: problem that we have NO any info about TS registers
<mripard_> and I'm not a fan of the macro stuff ;S
<ssvb> libv: was it when using scaler mode? and FBIOPAN ioctl is also taking effect only for the next frame
<woprr> wingrime, ACK. Will see how we get them
<libv> ssvb: i do not remember anymore, it was before fosdem
<Turl> mripard_: what about dropping the named members/tags then?
<libv> no, not scaling, but i did use fbiopan
<Turl> mripard_: { "sun4i,blah", data }
<mripard_> Turl: not great either, that would break if someone add some stuff to the structure
<oliv3r> wingrime: don't care about the MODULE_AUTHOR tag, MODULE_LICENSE is important and the top of the file, if it reads the GPL license, your in the clear
<wingrime> oliv3r: yes gpl there
<Turl> mripard_: if they're sensible and add it to the end it shouldn't break
<ssvb> libv: without scaler mode and doing WAITFORVSYNC immediately after FBIOPAN should work fine
<Turl> ssvb: do those ioctls actually work?
<ssvb> Turl: yes
<Turl> ssvb: last I played with them on android they returned immediately, they were no use
<Turl> I'll need to give it another try I suppose :)
<libv> ssvb: i will give it a go, but not right now :)
tzafrir has quit [Ping timeout: 264 seconds]
<oliv3r> woprr: we have docs for TS IP, just no sample code, no driver, nothing. So if someone wants to work on it, awesome, but it will have to be done from scratch
<oliv3r> Turl: sdio driver what? i missed my comment in reference to you
<woprr> oliv3r, ACK
<woprr> jemk, thx
<Turl> oliv3r: mripard_ is working on SDIO, check the wiki :)
<wingrime> woprr: you not simply connect TS and DISP and Cedar, there need normal diriver that use DMA
<wingrime> oliv3r: maybe ask lclk about some docs about hw without drivers?
<woprr> wingrime, ACK.
<wingrime> woprr: still have interst?
<woprr> yes :)
<wingrime> woprr: also we have CSI (for cameras)
<wingrime> woprr: better help with cedarx RE project
<woprr> yes in GPIO "multi2" mode , the TSx are CSIx ports
<wingrime> woprr: for CSI we have driver
<woprr> I need dvb, priority use case
<woprr> if the CSI controller can passthru the transport stream...maybe
<oliv3r> Turl: SDIO, IO over SD, didn't know mmc is part of that, well partially anyay
<wingrime> woprr: 1) we have no sample code for TS 2) our dma driver - crap 3) we have as result of RE - mjpeg mpeg12 code
<wingrime> woprr: with witch board you will play with it?
<oliv3r> wingrime: i think the doc has register info, let me reread
<oliv3r> that's the one
<oliv3r> yeah it has register info
<wingrime> oliv3r: not realy enought for work
<oliv3r> so we have all th einfo, but we dont' know how to connect hardware i belive
<oliv3r> it hsould be in theory
<oliv3r> let me read the docs again, its been awhile
<wingrime> oliv3r: may be ask lckl about driver less hw docs?
<oliv3r> i think they don't exist
<oliv3r> we could check sun3i code, but doubt it
<oliv3r> they wrote the register docs specifically for cubieboard actually
<wingrime> oliv3r: how much pressure will work
<wingrime> oliv3r: about GPL voliations
<oliv3r> there is no binary blob :p
<oliv3r> so it can't even violate anything :)
<oliv3r> i belive they added the hardware, but never used it
<oliv3r> they 'planned' on it maybe
<mnemoc> the original mele a1000 had AC3 and DTS
<mnemoc> then it was removed due to license issues
n01 has joined #linux-sunxi
<mnemoc> iirc licenses have to be paid by product seller, not chip vendor
<mnemoc> s/seller/manufacturer/
<oliv3r> we have found the ac3 and dts hardware decoder blobs
<oliv3r> and thanks mnemoc you just found out why there's an audio codec engine, but no drivers
<oliv3r> they stopped shipping it due to that then
<oliv3r> software can do it, so why bother with litigations etc
<oliv3r> ironically, at the least ac3 is violating the GPL, as it's ffmpeg based
<oliv3r> woprr: the ts unit , still requires a tuner, i _think_ its only a demodulator in dvb terms
<wingrime> mnemoc: can you find blob that come with mele
<mnemoc> even a single function from ffmpeg in the blob GPLizes the whole blob
<wingrime> oliv3r: I think TS only *split* MPEG-TS
<mnemoc> wingrime: i don't have my meles with me anymore :(
<mnemoc> only cb1 and olinuxinos
<mnemoc> hno probably has his ancient mele with original android on nand
<oliv3r> woprr: not sure how to hook up the descrambler, but i suppose with oscam that's not really an issue
<wingrime> oliv3r: find some chanel steam ant cut it
<oliv3r> wingrime: yeah, it's what linux-media/dvb calls the 'demodulator'
fredy has quit [Excess Flood]
<oliv3r> maybe even less, it's only a small part
<oliv3r> like 'pid filter'
<wingrime> mnemoc: can we get any pressure with this blob?
<oliv3r> so you have tuner, demodulator externally, then connect it to a10, an it will do hardware pid filter and deliver the ts to the system
<oliv3r> wingrime: lkcl_ is orking on it :) give it some time
<Turl> wingrime: I don't think they get any pressure with blob
fredy has joined #linux-sunxi
<oliv3r> Turl: why not?
<oliv3r> wingrime: actually it kinda sucks if it can only filter 1 pid (it's good if it can do 32 like the manual says)
<mnemoc> wingrime: I don't believe in gpl enforcement. if the violator doesn't want to give the code, no lawyer can get it
<oliv3r> USB sticks do pid filtering too, so you can do 1 stream via USB1.1 (or slow USB2 links)
<Turl> oliv3r: because they're a chinese SoC company, I dunno
<mnemoc> wingrime: you can only annoy them until they give up
<oliv3r> if you let the PC do pid filtering, you can tune to many channels easier
<mnemoc> or convince their CEO it's a good thing
bsdfox_ has quit [Ping timeout: 268 seconds]
<oliv3r> Turl: erm, why do you think lkcl got them to release u-boot and boot01 sources?
<oliv3r> mnemoc: and i belive that's hwt lkcl's idea is
<Turl> oliv3r: because EOMA will get them sales? and a happy lkcl means EOMA shipping
<oliv3r> Turl: if that gets us more stuff that we need, that's good
<oliv3r> and i think luke is after more then just more sales :p he actually genuinly cares about the GPL
<wingrime> mnemoc: gpl is aw advantage , rk and mtk have more power socs
<wingrime> mnemoc: if aw want live
<mnemoc> oliv3r: the best way is to convince AWs CEO that they are already violating the license, but that it will be great for their public image and sales to properly GPL cedarx libs
<wingrime> mnemoc: thay better not trow this adventage
<oliv3r> i agree kinda ith wingrime here, they are slowly going into obsolecance
<mnemoc> they don't target powerful, they target cheap
<mnemoc> and masssssssssssive cheap china devices
<wingrime> mnemoc: rk there too
<oliv3r> mnemoc: i am sure that's what lkcl is trying anyay
<mnemoc> ack
<oliv3r> yeah but rk is more powerfull and equal/cheaper price
<oliv3r> so why would they want to buy sunxi soc's anymore
<wingrime> mnemoc: I saw long time ago china iphone clones with TV and antenna
<wingrime> mnemoc: MTK based
<wingrime> mnemoc: and now LG makes phones with MTK
<wingrime> mnemoc: and more cool thing about MTK , old mtk have no second relatime core for gsm
<mnemoc> isn't google's stick also MTK based?
<mripard_> mnemoc: the chromecast ?
<wingrime> realy?
<rm> no
<rm> Marvel
<oliv3r> no
<mnemoc> i think i read so somewhere
<mnemoc> ok
<oliv3r> it's marvel indeed
<wingrime> Marvel have good docs but price
<mripard_> marvell actually
<oliv3r> yeah :p
<rm> what I don't get is the huge stupid AzureWave wifi
<mripard_> Marvel exists as well, but you know, they're in the superhero business :)
<rm> making dongle thicker at the back
<mnemoc> marvel have good super heroes
<oliv3r> ok i've build a kernel for the first time using 'the debian way' e.g fakeroot make-kpgk
<rm> and having no [good] Linux kernel drivers afaik
<oliv3r> all's fine and the .deb is ok, but what's up with this: -rw-r--r-- 1 root root 246M Jul 30 19:37 vmlinux.o
<arokux> mripard_, why do entries in dts end with _a, like emac_pins_a?
<mripard_> arokux: in case there's a _b ? :)
<mripard_> you can have several pins for the same function
<mripard_> so just emac_pins wouldn't fit
<Turl> emac can just be muxed in one way I think
<oliv3r> 250 MiB for vmlinuz.o? that can't be right. I don't see anything strange in my .config
<Turl> but stuff like uarts can have 3-5 different sets of pins
<oliv3r> my /usr (with 1 kernel) is 4G
<mripard_> Turl: some pins aren't
<wingrime> mnemoc: there any reason with google stick , google better deal with TV-set makers and add this internaly
<Turl> oliv3r: debug? :)
<mripard_> you have a few emac pins that can be muxed on several pins
<mripard_> but yeah, emac is probably a bad example in that case :)
<oliv3r> Turl: good point! let me double check that
<oliv3r> but 250mib?
<Turl> oliv3r: 63Mvmlinux
<Turl> 99Mvmlinux.o
<Turl> not that off :p
<mnemoc> i prefer to not have to replace the whole tv when a better SoC comes out
<oliv3r> far? mine's 2 1/2 yours :p
<oliv3r> eoma-68k based tv
<wingrime> mnemoc: I have mips in tv
<Turl> oliv3r: k for overclocking capability? :p
<oliv3r> Turl: lol
<Turl> wingrime: I dunno what mine has, but it runs linux with ffmpeg and friends, and takes like 5s to boot :)
<oliv3r> what tv do you have?
<Turl> pretty annoying for a computer monitor
<oliv3r> my tv takes 30 -45s to boot :)
<arokux> mripard_, so if i want to use some of the pins again in dts, then they need to be referenced by different name?
<Turl> oliv3r: syncmaster TA550
<oliv3r> well tv boots in < 1sec, PC that drives it takes the other 44s :p
<Turl> oliv3r: my PC boots faster than the monitors usually :p
<wingrime> mnemoc: in russian internet we call such devices like "bidlo devices" - it like game consoles , iphones, some devices that locked on vendor and can't do anothing but becomes money black hole
<Turl> wingrime: I think Stallman calls them "tivo" devices
<wingrime> Turl: also do surviliance
<Turl> and that was the main motivator for GPLv3
jemk_ has joined #linux-sunxi
jemk has quit [Disconnected by services]
jemk_ is now known as jemk
Soru has joined #linux-sunxi
<wingrime> Turl: and about locked functions
<wingrime> Turl: like modern intel cpu, you can bye unlock code
<oliv3r> Turl: i boot a reasonable standard debian, not sure why it's so slow, but this install is only a few days old, so haven't investiaged
<wingrime> Turl: I hope some one will fix that issue
<oliv3r> Turl: ho does your pc boot < 5sec? suspend?
<techn_> evening :)
<wingrime> techn_: !!
<Turl> wingrime: unlock code? :O
<oliv3r> techn_: uh oh
<Turl> oliv3r: try e4rat if you're using spinning rust
<wingrime> Turl: yes
<Turl> oliv3r: no, full boot
<oliv3r> Turl: my bios takes longer then that :p
<wingrime> Turl: unlock some cache , core in cpu
<Turl> oliv3r: it's a bit more than that when you consider slowBIOS but w/e
<oliv3r> even then, dhcp takes ages
<oliv3r> i have to clean some services probably
<wingrime> Turl: some instion set
<Turl> dhcp is instant? :P
<techn_> mnemoc: wingrime: usb problems solved?
<Turl> wingrime: I haven't seen any for mine at least :p how do you input them?
<arokux> mripard_, why there is a need to give pin names, when there already exists a mapping in pinctrl-sunxi-pins.h?
<mnemoc> techn_: at least build-time wise
<mnemoc> techn_: stage/sunxi-3.4 updated
<oliv3r> Turl: takes like 4 to 5 seconds here
<oliv3r> i'll install bootchart and see what's going on
<Turl> oliv3r: well, maybe it's done when I type my pw on gdm :P
<techn_> yeah.. hansg has two other build patches
<arokux> mripard_, hm.. maybe because of functions 0 and 1..
jlj has quit [Ping timeout: 268 seconds]
<mnemoc> techn_: I used the head of hansg's sunxi-3.4-for-amery
<techn_> mnemoc: he send new merge request today
<mnemoc> techn_: but usb/gadget/android.c was still broken. Turl fixed it
<techn_> ah..
<techn_> you merged it already
<mnemoc> techn_: yes, current HEAD of sunxi-3.4-for-amery
<Turl> wingrime: :O
<wingrime> Turl: it may be microcode upgrade that will burn some hw fuzes
<oliv3r> i'll stick with AMD for now ty
<wingrime> Turl: very possible that intel produce only one IP but test , sort it and flash on factory
<wingrime> Turl: so i3 i5 and i7 can be same chip
<oliv3r> Turl: looking at bootcharts of e4rat, that does look friggin amazing
<oliv3r> will test it now
<Turl> wingrime: chip vendors have always done that
<Turl> wingrime: never heard of "core unlocking" of AMD chips for example?
<oliv3r> Turl: yeah debugging as all enabled
<oliv3r> yeah binned cores that could be re-enabled
<mnemoc> afaik that's the result of the quality degradation when they build the.... plates of chips. the middle has the higher quality, the outer part the worst
<oliv3r> but wingrime, turl; this is a different thing, if you can 'buy' some unlock code
<oliv3r> so you buy a cpu, pay extra to unlock extra features
<oliv3r> mnemoc: wavers*
<mnemoc> that :)
<oliv3r> but yeah, center die's tend to be higher quality; they tend to make the server/mobile CPU's from those parts
<techn_> mnemoc: Turl: there is some stuf done twice now
jlj has joined #linux-sunxi
<oliv3r> lower power usage, higher frequencies, little more stable, for server parts taht's important
<mnemoc> techn_: uh
<mnemoc> techn_: :(
<wingrime> Turl: )
<Turl> techn_: ?
<techn_> Turl: check composite.c
<Turl> techn_: yeah it probably needs way more work :)
<Turl> techn_: but at least it builds now :)
<techn_> print_android_config is implemented twice.. luckily it's not called :p
<mnemoc> techn_: cleanup patches welcomed :)
<mnemoc> then we can merge to sunxi-3.4, tag, and jump forward
bsdfox_ has joined #linux-sunxi
rm has quit [Quit: ZNC -]
rm has joined #linux-sunxi
tzafrir has joined #linux-sunxi
fredy has quit [Excess Flood]
fredy has joined #linux-sunxi
leowt_ has joined #linux-sunxi
leowt has quit [Ping timeout: 256 seconds]
mcbrick has joined #linux-sunxi
shineworld has joined #linux-sunxi
eebrah|away is now known as eebrah
<arokux> hm.. there are ahb_usb0, ahb_ohci0, ahb_ohci1, ahb_echi0, ahb_ehci1 clocks, how do they relate to
<techn_> arokux: you working musb driver?
<arokux> no, ochi for now
<arokux> ohci*
<techn_> jukivil1 has done some driver prototype already
<arokux> techn_, i'm working on mainlining
<oliv3r> arokux: oh you found the host usb driver :) sw = softwinner :)
<oliv3r> probably will end up like sunxi_usb_hci i suppose?
<arokux> yes, those are usb1 and usb2
<oliv3r> you need all those clocks i'm affraid
<oliv3r> ahb_usb0 might be for otg actually
<oliv3r> ahb_ohci0 and ahb_ochi1 are the usb '1.1' clocks for 0 and 1 obviously
<oliv3r> ehci0 and ehci1 are the ehci controllers '2.0' clocks for 0 and 1
<oliv3r> so yep you'll need them
<arokux> for 1 and 2
<oliv3r> aye
<arokux> 0 is otg
<oliv3r> btw, you think you can make it work initially in PIO mode?
<oliv3r> since we don't have dma yet ;)
<arokux> oliv3r, i hardly understand things, so ...
<arokux> oliv3r, but it makes some sense.. :)
<oliv3r> then maybe USB is a bit to far fetched for the first one?
rellla has joined #linux-sunxi
<oliv3r> the IR driver is a nice one to start with too
<oliv3r> if you have a cubieboard, there's an IR pre-solderd
<arokux> oliv3r, i've spent some time already, so i'll try
<oliv3r> sure thing :) remember, 3.3 and 3.4 have buggy USB drivers to begin with
<arokux> yes.
<arokux> oliv3r, my hardware is mele a1000
<oliv3r> also has IR :D
<oliv3r> Turl: boot times haven't improved hugely, but xbmc starts significantly faster
fredy has quit [Excess Flood]
rellla has quit [Quit: Nettalk6 -]
fredy has joined #linux-sunxi
<wingrime> oliv3r: it not simple remote control
<wingrime> oliv3r: it can 4Mb irda
<oliv3r> n01: you should have added your e-mail to your copyright ;) so people can contact you
<oliv3r> wingrime: IR0 and IR1 can do both I belive, irda OR 36 khz remote control
<oliv3r> mripard_: actually even linus doesn't like the 80 collumns limit of checkpatch ;)
<woprr> re, sorry got a visitor, now answering
<oliv3r> (in certain situations of course)
<oliv3r> woprr: sure, go backread :)
<arokux> oliv3r, there are ahb_*hci* already, but where should usb_phy come frome?
<oliv3r> arokux: a little more specific? :)
<woprr> oliv3r, the TS interface is compatible with standard par or ser dvb (tin can) receivers (tuner+demod) like Philips CU1216
<woprr> dvb bus
<oliv3r> woprr: oh really nice
<arokux> oliv3r, they are supposed to go into dts, i suppose
<oliv3r> they are allready in dts :)
<oliv3r> Turl: ^
<oliv3r> arokux: there's 2 clocks to most IP cores
<Turl> oliv3r: 'sup
<oliv3r> 1 clock to communicate with the AHB bus (or the CPU side of the core)
<Turl> missing clocks? probably not implemented just yet :)
<oliv3r> 1 clock to make the IP core 'work'
<oliv3r> Turl: did you add usb clocks allready?
<woprr> oliv3r, descambler is no project target, no CA, 1. milestone is free tv dvb SDTV only, HD later, and we will surely get no specs for any media protection parts from Allwinner, nor CI+
<Turl> oliv3r: unless they're "mod0 style", no
<oliv3r> woprr: i think they probably just have some form to access the CI, but i find ci+ uninteresting, oscam handles CI nicely imo
<oliv3r> woprr: but yeah, dvb-t; dvb-c; dvb-s and dvb-s2 should all be supported
<oliv3r> Turl: they might not
Black_Horseman has joined #linux-sunxi
Black_Horseman has joined #linux-sunxi
Black_Horseman has quit [Changing host]
<oliv3r> Turl: rgr; thank you
<Turl> looks like a bunch of gates, shouldn't be hard to implement
<oliv3r> should be pretty trivial i suspect
<woprr> wingrime, board "to play with" (pardon?) is the Olimex A20 micro
<Turl> if you can make a proper list of all what's needed for USB I'll implement them
<wingrime> woprr: ack
<Turl> bbl :)
jemk has quit [Read error: Operation timed out]
<woprr> wingrime, 1) for what sample code? 2) we can utilize good DMA system professionals from 3) " mjpeg mpeg12 code" ? media?
<wingrime> woprr: ^
<wingrime> also mpeg there
<wingrime> woprr: for mpeg decoder
<woprr> oliv3r, CA: hard. CI+ with OSS? Mission impossible, security by obscurity restricted
<woprr> wingrime, OK
shineworld has quit [Quit: Leaving]
<wingrime> woprr: you can also see VE Register page that I do
<woprr> wingrime, nice, thx
<woprr> will look into tomorrow, thanks for all, Your nice, good night, it's 22:34 here :)
<woprr> cu
<oliv3r> woprr: we don't have DMA in mainline kernel at the moment, mdp is working on it i belive
<oliv3r> woprr: same here, nn )
<oliv3r> well same time :)
<woprr> oliv3r, ack
<oliv3r> 3.4 kernel is 'feature complete' hoewever
<woprr> ok
<oliv3r> but not mailnienable
<oliv3r> libv: you made rob go like crazy mad!
wingrime has quit [Ping timeout: 245 seconds]
wingrime has joined #linux-sunxi
<mnemoc> oliv3r: the purpose of 3.4 is to have a fully usable unified code base that can evolve into mainlineability while been broadly tested
<mnemoc> refactoring of legacy code vs. reimplementing
<oliv3r> not sure if refactoring is possible with a lot of the drivers we have
<oliv3r> they are messy and buggy :p
<mnemoc> not sure if reimplementing is possible with a lot of the drivers we have
<mnemoc> without documentation, new code is more likely to be more buggy
<mnemoc> even if more "elegant"
<oliv3r> well you can always use the 3.4 drivers as reference
<mnemoc> that means you can also refactor 3.4 drivers into decency
<oliv3r> in theory, of course
<oliv3r> but it's a lot of work, meaning the line between reimplementing them or refactoring them becomes very blurry
<mnemoc> leaving the script.bin vs dts issue aside
<oliv3r> so it's a mix of both if anything
<oliv3r> ;)
<oliv3r> mnemoc: btw, it's good to have you back
<oliv3r> i'm sure i speak for all that we've missed you and it's nice to see you being more active here again
<mnemoc> it feels good to be back
<mnemoc> but i still need to read more than 2k mails :(
<arokux> mripard_, there?
<oliv3r> yeah that's a lot of mail to process
<mnemoc> hopefully hansg's branches cover a big part of that
<oliv3r> they should i think
arokux has quit [Remote host closed the connection]
eebrah is now known as eebrah|away
Offshore has quit [Ping timeout: 240 seconds]
<wingrime> oliv3r: can you try unify sun4i sun3i sun5i IR drivers looks 100% same
<oliv3r> sure thing
<oliv3r> well i'll ingore sun3i; since i'm planning on removing it :p
<oliv3r> but sun457i should be identical
<libv> oliv3r: yeah, rob is a highly clued and highly motivated graphics driver developer
<oliv3r> libv: fun and exciting to read :D
<libv> oliv3r: he actually gets to work on freedreno on worktime
<oliv3r> oh wow
<oliv3r> well he works for RH now
<oliv3r> so figures
n01 has quit [Ping timeout: 246 seconds]
<hno> mnemoc, you are lucky. My inbox says 4090 unread 4677 total. Don't think I ever will catch up.
<oliv3r> hno well i bet 20 or so are from me that are dealt with ;)
<oliv3r> so could be deleted quickly
tvall has joined #linux-sunxi
Soru has quit [Ping timeout: 264 seconds]
wingrime has quit [Ping timeout: 268 seconds]
_whitelogger has joined #linux-sunxi
tinti has quit [Read error: Connection reset by peer]
<Turl> hno: I had something like that some time ago
<Turl> I just marked them all as read and started from scratch
<hno> Turl, mm.. I did that 4 months ago.
<Turl> hno: you need to keep up better then :p
<hno> Or find a drone that reads the mail for me...
<hno> Hmm.. confusing. have dealt with some hundred emails and now it says 4212 unread, 4704 total..
<Turl> hno: tagging can help a lot if you're not doing it already