Turl changed the topic of #linux-sunxi to: Allwinner/sunxi development discussion - Don't ask to ask. Just ask and wait! - See http://linux-sunxi.org | https://github.com/linux-sunxi/ | Logs at http://irclog.whitequark.org/linux-sunxi
Turl has joined #linux-sunxi
zeRez has quit [Quit: zeRez]
torbenh3 has quit [Ping timeout: 272 seconds]
torbenh3 has joined #linux-sunxi
Gerwin_J has quit [Read error: Connection reset by peer]
montjoie[home] has quit [Ping timeout: 252 seconds]
montjoie[home] has joined #linux-sunxi
focus has quit [Ping timeout: 245 seconds]
wingrime has joined #linux-sunxi
focus has joined #linux-sunxi
popolon has quit [Quit: Quitte]
<wens> hramrach: dw u-boot confirmed broken @ Gbit
wingrime has quit [Ping timeout: 272 seconds]
<TheSeven> grr. both dragon and awimage are crashing in the same way
<TheSeven> trying to free an invalid malloc handle
<TheSeven> so either a double free, a bogus address free, or memory corruption of some kind
<TheSeven> wtf
tinti has quit [Quit: Leaving]
Turl has quit [Ping timeout: 252 seconds]
Turl has joined #linux-sunxi
printallthething has quit [Ping timeout: 246 seconds]
pfdm has quit [Ping timeout: 250 seconds]
egbert has quit [Disconnected by services]
egbert has joined #linux-sunxi
Turl has quit [Ping timeout: 252 seconds]
Turl has joined #linux-sunxi
hsildaerD has joined #linux-sunxi
captainigloo has quit [Ping timeout: 264 seconds]
captainigloo has joined #linux-sunxi
Dreadlish has quit [Ping timeout: 264 seconds]
<wens> so much for asking for gmac specs :(
<ganbold> wens: how is it going?
<wens> specs aren't released to the public
<ganbold> is it same to stmmac?
<ganbold> or totally different?
<arokux2> wens: you've got some reply from STMicroelectronics?
arokux2 has quit [Remote host closed the connection]
<wens> arokux: from allwinner. benn sent a mail yesterday
TehCaptain has joined #linux-sunxi
TheSeven has quit [Disconnected by services]
[7] has joined #linux-sunxi
prasannapete has joined #linux-sunxi
JohnDoe_71Rus has joined #linux-sunxi
eebrah has joined #linux-sunxi
eebrah has quit [Ping timeout: 272 seconds]
eebrah has joined #linux-sunxi
eebrah has quit [Ping timeout: 252 seconds]
eebrah has joined #linux-sunxi
hramrach has quit [Ping timeout: 240 seconds]
eebrah has quit [Read error: Connection reset by peer]
eebrah has joined #linux-sunxi
Montjoie has quit [Ping timeout: 252 seconds]
hramrach has joined #linux-sunxi
Montjoie has joined #linux-sunxi
eebrah has quit [Read error: Connection reset by peer]
deasy has joined #linux-sunxi
diego71 has quit [Ping timeout: 272 seconds]
diego71 has joined #linux-sunxi
Quarx has joined #linux-sunxi
Taduro_ has joined #linux-sunxi
tomboy64 has joined #linux-sunxi
BluesBoy has joined #linux-sunxi
wolfy has joined #linux-sunxi
wolfy has left #linux-sunxi [#linux-sunxi]
FR^2 has joined #linux-sunxi
<oliv3r> morni'n ya'll
<oliv3r> mnemoc: ping
<JohnDoe_71Rus> day
n01 has joined #linux-sunxi
Taduro_ has quit [Read error: Connection reset by peer]
rellla has joined #linux-sunxi
juanfont has left #linux-sunxi ["Saliendo"]
<mnemoc> oliv3r: pong. but tell me on /q to be able to read it in 1-2h
juanfont has joined #linux-sunxi
<wens> oliv3r: morning
wolfy has joined #linux-sunxi
prasannapete has quit [Quit: prasannapete]
wolfy has left #linux-sunxi [#linux-sunxi]
hramrach has quit [Ping timeout: 240 seconds]
hramrach has joined #linux-sunxi
tgaz has quit [Ping timeout: 272 seconds]
tgaz has joined #linux-sunxi
<hramrach> arokux: no, did not try cubietech kernel
notmart has joined #linux-sunxi
tzafrir has quit [Ping timeout: 272 seconds]
<focus> hi, parallel computers - anyone interested? i find a way to stack cubieboards to make full 3D computers and take care of the wiring and cooling no matter how deep :)
<focus> with the price of cubieboards going down to $59, i can afford to build an array
hsildaerD is now known as Dreadlish
<focus> highest priority is build a web server that allows a file to be uploaded to one server by ssh and then it gets read and all its component files like images get separated and put into different machines to speed up the entire delivery of a web page
<focus> with the price of 1TB 2.5" disks going down to about USD 70, i can give each cubie a 1TB disk
<focus> 6 cubies with hd arranged into cubic case
<rm> think on the software side of things
<focus> about usd 1000 per cube with 6TB storage
<rm> before wasing money on lots of boards and disks
AreaScout has joined #linux-sunxi
<rm> make a proof of concept with 2 boards and some leftover 160GB disks
<hramrach> how does that compare to a single high-end PC price and performance wise?
<rm> because I feel the software side is what will be the most problematic for you
<rm> not "take care of wiring and cooling"
<focus> rm: i work with 10 real desktop PCs and not impressed because of energy costs
<focus> hramrach: goes back to page turn speed of DDR - about 10MHz - so 6 x cubies will outperform 6 x PCs on a network when fully loaded
<hramrach> so the bottleneck is memory precharge times?
<focus> rm: i want to get into software as best as i can - the main problem holding it back is the availability of cheap enough hardware to spend time making software
<rm> as I said you need just two
<focus> hramrach: afraid so
<hramrach> still a PC has more modules so should be slightly better
Tsvetan has joined #linux-sunxi
<hramrach> but probably not as many times better as many times more power it draws, maybe
vicenteH has quit [Read error: Operation timed out]
<focus> hramrach: pc wins because it has lot deeper caching, but even caching can't help - i try with big data - no different between quad cores and dual cores - the speed of ssd is the limit
<hramrach> but there PC wind hands down - way faster buses
<hramrach> so you get more out of your ssd
<focus> hramrach: fast busses, deep caches, and GPUs
<hramrach> cpu sits there just to program dma for io intensive workloads
<focus> so my idea was to make one of the CPU responsible for rendering for example and let spread the load to other units - but rm says its software - it don't exist - but as I say - the hardware is now low enough in cost to start the experiments
<focus> i mean the serious experiments to break up processing and efficiently get the games rendering engines done up
<focus> and then move on to the database engines
<focus> a long waiting to be in such position :)
<focus> just notice the cubie price is $59 or UK pounds is 35 - very cheap
<focus> 1TB disk an overkill for experiments - settle for 32GB fast ssd
<focus> at $1000 per cube with 6 processors, don't give a hoot - just want the system :)
<hramrach> focus: you probably won't make use of fast SSD on cubieboard
<hramrach> but then again the cubiebaord sata may be faster than cheap spinning disks
<focus> i think the first experiment is build the cubic case with 6 cubies and 6 ssds, then write a gambas application to read an upload directory for html files, and then parse it, and then move component files to individual machines automatically, and then when the web page is requested, all 6 machines can respond at the same time
<focus> measure the response times to make sure the system works first
<hramrach> I guess you can measure some data that tells you if PCs or cubieboards have better perf/watt and perf/$ for your application
<hramrach> but hard to tell without really trying
<focus> the wattage i calculate as 40 Watts which easily defeats a PC
Xaros has joined #linux-sunxi
<wens> wasn't some vendor coming out with arm servers?
Black_Horseman has quit [Ping timeout: 245 seconds]
<focus> the performance i can bet easily defeats a web farm if set up correctly
<focus> wens: yes - they be 64 bit as well i believe
<focus> but until us 'poor' guys start doing the experiments and build computing with thundering ability available in bite size chunks, it always remains out of reach unobtainium
<wens> building clusters yourself, you'll reach a point where it's no longer easy to manage
<focus> at $1000 per 6 cubie solution, its still a stretch for many, but i'm betting that too will come down
<focus> wens: i built cluster before and you are totally write - even with 10 machines, there were drop outs and at times you could only count on 6
Taduro has joined #linux-sunxi
<focus> wens: i'm thinking software needs to be built that makes it easy to manage - don't want to spend time thinking how do i do xx
<focus> just point click and add more to the array :)
<focus> more realistic games rendering arrays, faster databases, and more AI capable machines
dlan has quit [Quit: Lost terminal]
<focus> even things like 3D printers now moving to 25 micron resolution has huge computing overheads and unless software learns to divide up tasks, these things are going to crawl
<focus> a lot of operations can't be done in 3D like 3D subtraction as soon as the data gets big
<focus> or if the details gets too much
<focus> most of this is the limitation of DDR page turn speed - 10MHz - no amount of caching will get over it once the data size exceeds a certain limit
<wens> you could get a decent PC or server for $1000 though
<hramrach> with web serving you get popular and less popular data
<hramrach> and the larger the cache on a single node the more data fits there
<hramrach> so even less popular data gets served fast
<hramrach> so my estimate is that latency on data with popularity slightly under the peak kills you for small nodes like cubies
<focus> with web serving i am confident i can defeat any pc with this array, but very unsure about 3D data and big data
Xaros has quit [Quit: Zwi se logou mou!!!]
Black_Horseman has joined #linux-sunxi
<hramrach> but depends on the data popularity distribution you use for your testing, of course
<focus> hramrach: the web page is content is divided into the 6 cpus by a parser - when the client PC makes a request, the web page contains references to 6 different IP addresses for images and various content - so 6 cubies respond simultaneously making it a lot quicker than a server PC
<focus> that be why i believe the web server i can win hands down
<focus> but not sure about winning with big data
<focus> unless a few steps are taken to link databases together - which is very hard
<wens> you could just use a load balancer up front
<focus> wens: i was going to do a little more intelligence like move content between servers on the fly as loads pick up and drop
<hramrach> more deterministic load balancing than just random could possibly hide the issue with small local cache
<hramrach> but then the software will be very important
<focus> i order 3 cubies later today and some more next month to see what can be done
<focus> hramrach: good to talk - will update on any progress :)
dlan has joined #linux-sunxi
<focus> thanks guys :)
prasannapete has joined #linux-sunxi
<libv> Tsvetan: are we happy with what itead provides?
wolfy has joined #linux-sunxi
<Tsvetan> libv I dont have cadence to verify what he uploaded (the files are with date today 10.30 so these were not on the web for sure)
<Tsvetan> but if these are the CAD files for the base board its OK
<libv> Tsvetan: what license do you use on your boards?
<libv> Tsvetan: perhaps you could help Alex stick a license on his files
<Tsvetan> we used to use CC
<libv> i will rework the iteaduino wiki page
<Tsvetan> but I think to move to Apache 2.0 in future
<libv> Tsvetan: help him out a bit, he seems lost with respect to licensing
<Tsvetan> Im lost myself also
<Tsvetan> I seriously though to use WTFPL
<Tsvetan> for LIME
<libv> at least get him to put a useful license on his stuff, he seems completely lost
<Tsvetan> but http://www.gnu.org/licenses/license-list.html#WTFPL says its better to use Apache 2.0
<libv> then tell him to put at least apache 2.0 on these files
<Tsvetan> as with WTFPL you are not protected from patent treachery
<libv> not that that is a worry for a chinese shop, but that is nice to have nonetheless
<libv> help him with that, i'll work the wiki
fredy has quit [Excess Flood]
<Tsvetan> ok
<oliv3r> if any of you have mod points on slashdot, i tried to post as clear as I could, but didn't have an account that I wanted to tie directly to my name handy (i do now) http://hardware.slashdot.org/comments.pl?sid=4485207&cid=45524191
<oliv3r> if i put anything wron gthere, feel free to shame me
<libv> oliv3r: pearls for swine :)
<oliv3r> focus: you should wait a little bit for the new 'rack-SoM' to come out ;)
fredy has joined #linux-sunxi
<oliv3r> libv: huh?
<libv> oliv3r: your /. post contains actual information and facts
<libv> no /.er will read stuff like that!
penghb has joined #linux-sunxi
rellla3 has joined #linux-sunxi
vicenteH has joined #linux-sunxi
HeHoPMaJIeH has joined #linux-sunxi
HeHoPMaJIeH has joined #linux-sunxi
HeHoPMaJIeH has quit [Changing host]
bamvor has quit [Ping timeout: 260 seconds]
bamvor has joined #linux-sunxi
<libv> mnemoc: would we want to host these files on sunxi.org once a license has been chosen?
<arokux> Tsvetan: when will you sell LIME?
<Tsvetan> I do not understand also this: this board is not OSHW if every bit if not open source too, in this case Arduino is also not open source as AVR microcontroller is without VHDL sources :)
<oliv3r> libv: i know, and i posted as AC, so i am downmodded to oblivion!
<Tsvetan> arokux the 50 developer edition are assembling now, the mass production mid December
<oliv3r> libv: but there was so much diss-information in such a short short thread, i just had to ;)
<oliv3r> if they don't have it on github; i think we should host them, having them only on their own account, while good, makes it failable to easily imo
<oliv3r> Tsvetan: yeah people are being stupid
<arokux> kriegerod: ping
<arokux> has somebody rebuilt cubie/sunxi-3.4?
tzafrir has joined #linux-sunxi
<wens> stopped reading /. a long time ago
<oliv3r> young readers tend to not know what they are talking about
<oliv3r> med-age readers tend to be windows-whores
<oliv3r> it's slowly saddening
<arokux> what are you talking about?
<wens> windows-whores?
<libv> hrm, i really wonder why benn huang is so hung up about an android image
<libv> yes, it is a competitor to cubieboard, but there are many competitors to the cubieboards
<libv> let's see what he has to say next time he jumps into the channel
<arokux> ugh... cubie/sunxi-3.4 (3.4.61) has problems with gcc-4.7
<arokux> I thought 4.7 is old enough.
<arokux> libv: maybe the other competitors do not make it that compatible to cubieboard? :)
<libv> arokux: quite a few boards will boot with images for other devices
<libv> arokux: but whether this is intended or not, the value is in the hw anyway
<libv> the cubie "brand" is pretty much established now, and cubietech will be our favourite as cubietech actively participates in sunxi.org
<arokux> libv: yes, this is true.
<libv> the download link states that this is the cubieboard image
<arokux> what is this compiler thing, how can one know what is broken once the kernel crashes with a new compiler? :(
<arokux> wens: ping
<arokux> focus: have you seen LIME? the price is even cheaper.
<oliv3r> i love lime and don't even have it yet
<arokux> oliv3r: it is only A10/512MB though :(
<arokux> oliv3r: but nice case :((
<oliv3r> same as pi :p
<oliv3r> but xbmc box, limebmc or olimexbmc it'll be enough :)
<libv> but much more open than pi
<oliv3r> i forsee in a year or so, when prices have dropped, a 1 GiB a20 version will be possible
<oliv3r> i do hope he will do an option, A10 and A20 versions
<oliv3r> in a year from now, hopefully it will be 100% open
<arokux> oliv3r: in a year from now there will be new hardware...
Black_Horseman has quit [Ping timeout: 264 seconds]
<oliv3r> arokux: i'm not holding my breath for that
<oliv3r> we will have what, a23? which will be like a13, so a20 may be more interesting for certain loads
<oliv3r> q3/q4 of 2014 we may see a60
<oliv3r> so boards at the end of next year
<oliv3r> by then, I hope sunxi to be close to being done with the RE effords
hansg has joined #linux-sunxi
<oliv3r> and if that's the case, chances are, A60 will be very similar to A20 but with big.LITTLE, the IP will mostly be the same, there's no point for them to swap working things out
<oliv3r> so we probably can use 3.4 at the getgo
<oliv3r> meaning we don't get a huge setback
<arokux> oliv3r: why do you think A60 won't be based on A31 which is different from A20 and for which we do not have support?
tomboy64 has quit [Ping timeout: 240 seconds]
tomboy64 has joined #linux-sunxi
boycottg00gle has joined #linux-sunxi
<oliv3r> arokux: because A80 will be beased on A31, we know that pretty much allready
<wolfy> speaking about hardware, are there any boards ( or systems) quad-core with decent linux support? by decent I mean that there is at least an image which can be downloaded and booted.
<oliv3r> a80 will be dual channel memory controller; powerVR
boycottg00gle has left #linux-sunxi [#linux-sunxi]
<oliv3r> wolfy: radxa rock
<oliv3r> wolfy: though bootloader is closed source
<oliv3r> or rather, GPL infringing u-boot
<arokux> oliv3r: I see, two lines A20 ---> A60, A31 ---> A80
<wolfy> as long as it boots, I do not care. I was looking at cubox-i but linux support is .. suffering
<wolfy> Subscribe to be notified when Radxa Rock is available.""
<arokux> wolfy: you won't get any support for radxa too, only community.
<arokux> wolfy: it is currently in production.
<wolfy> arokux: I am involved in MANY communities. codounder 14 years ago of the first LUG in the country, in fedora since 2005, in centos.. and others
<wolfy> i know the drill
<arokux> wolfy: nice :) I'm just saying do not expect support from Rockchip...
<wolfy> arokux: The company I work for is a TI contractor. And I do not expect support for BBB from TI :)
<arokux> BBB?
<wolfy> beaglebone black
<wolfy> I tested rpi and bbb before purchasing a cubie2. but i need more juice or decent hardware accelerated decoding
<wolfy> and since the decoder in cubie is not yet properly supported...
<arokux> wolfy: what SoC has a better supported decoder?
<wolfy> arokux: no idea, that's why I am lurking here
<oliv3r> arokux: yep, A10 -> A20; A13 -> A23; though those are pretty much identical, A13 is a slightly upgraded subset of A10; A31 is the ugly stepchild :)
<oliv3r> wolfy: decoder is slowly improving :)
<oliv3r> but i don't think many other SoC's are better
<oliv3r> AM Logic has a 'ok' supported soc with decoder
<oliv3r> if you want a consumer product that 'just work'
<oliv3r> but you may aswell get an apple-tv then
<wolfy> oliv3r: right. but i still need a functional device. and the alternatives are NUC or pc-fit. and rather than paying 5* the price, I'd rather use a quad-core
<arokux> wolfy: have you seen Utilite? (i'm not sure about multimedia)
<wolfy> arokux: yes. they are in pair with cubox-i . and quite pricey so I am still window shopping :)
boycottg00gle has joined #linux-sunxi
boycottg00gle has quit [Remote host closed the connection]
<kriegerod> arokux: pong
<arokux> kriegerod: i've written an e-mail to you
abcd593 has joined #linux-sunxi
<abcd593> Hi, is there any cheap, but reliable Chinese tablet or netbook that runs Linux (preferably Debian or Ubuntu)?
rellla3 has quit [Remote host closed the connection]
<kriegerod> arokux: i used gcc 4.7.3 armhf, prepared on x86_64 box using crossdev
zeRez has joined #linux-sunxi
rz2k has joined #linux-sunxi
Guest34299 has joined #linux-sunxi
Guest34299 is now known as eebrah_
<Tsvetan> guys do you know http://openinventionnetwork.com/about.php ?
<focus> arokux: Olimex LIME - EU30+tax+shipping > USD59+cn shipping :(
<zeRez> hi
adb_ has joined #linux-sunxi
eebrah_ has quit [Read error: Connection reset by peer]
<Tsvetan> focus compared to ?
<wens> arokux: pong
eebrah_ has joined #linux-sunxi
<focus> Tsevetan: cubieboard2
<Tsvetan> how much is cubie2?
<focus> Tsevetan: i find way to make 3D array of computers to make parallel computers
<Tsvetan> and how $39 is more than $59 :)
<Tsvetan> now I got what you wrote above
<focus> cubie2 is $59 each + $18 postage for 3 x cubie
eebrah_ has quit [Read error: Connection reset by peer]
<Tsvetan> focus where you are from?
<focus> uk
eebrah__ has joined #linux-sunxi
<focus> i pay 20% tax
<Tsvetan> then you pay tax on all imports
<Tsvetan> 3 LIMEs to UK are probably EUR 5.50
<focus> Tsevtan: China deal means no tax below USD199
<Tsvetan> there is no such deal unless you cheat your government
<Tsvetan> here everything above EUR 5 is taxed
<focus> Tsvetan: genuine
<Tsvetan> no matter where it comes from
<Tsvetan> VAT rules are same in EU
<Tsvetan> there is no exemption
<focus> Tsvetan: HK was UK territory
<abcd593> uk is not quite EU :P
<Tsvetan> I see :)
<Tsvetan> so now China is part of UK
<Tsvetan> :)))
<wens> haha
<focus> Tsevetan: Also Korea has deal with UK - a $50k laser costs me no tax i was told
<abcd593> anyway: is there any cheap, but reliable Chinese tablet or netbook that runs Linux (preferably Debian or Ubuntu)?
<libv> abcd593: linux-sunxi.org
<torbenh3> thats why all stuff from china is sent into the e.u. via uk...
<Tsvetan> I guess your covernmet put a big hole in the tax budged with this $200 deal :)
<wens> but there's still VAT, right?
<focus> Tsevetan: no real problems - because if it turns into trade, or you sell it on, then gov collect tax - everyone wins :)
<focus> (trade volume i.e. > USD 200)
<wolfy> wens: it depends. i ordered (in personal name) a lot of stuff from dx.com and did not pay any tax at all. I ordered a cubie2 from eleduino for the company I worlk for and paid 24% (VAT)
geecko has joined #linux-sunxi
<wens> wolfy: i think companies always have to pay taxes, because they need the invoices for expensing and stuff
<wolfy> i wanted to buy from miniand but their policy is stupid. they add 6$ shipping per item. so added a Wi-Fi stick for cubie=> +6 USD. added a case => + another 6
<wens> wolfy: do they actually ship each item seperately?
<wolfy> i doubt that
<focus> wolfy: depends on store - don't buy from stores that do 1:1 costing
<wolfy> i even mailed them. received no reply
<rm> Tsvetan, A13-OLinuXino-MICRO 35.00 EUR Shipping 16.97 EUR Total 51.97 EUR = 70.37 US Dollar
<focus> wolfy: ah!
<rm> MINI PC cubieboard A20 ARM Cortex A7 Dual-Core US $66.31 / piece Free Shipping to Russian Federation via China Post Air Mail
<Tsvetan> rm if you use courier yes, but you can buy from our Russian distributor too locally
<abcd593> libv: thanks
abcd593 has left #linux-sunxi [#linux-sunxi]
<focus> rm: thanks for those numbers - save me from Tsvetan wrath!
<focus> :)
eebrah__ has quit [Read error: Connection reset by peer]
<Tsvetan> rm so you compare China post with UPS/DHL courier :)))
<rm> Tsvetan, I do not think I can, I am 2000km from Moscow
hansg has quit [Quit: Leaving]
<rm> and shipping within russia is pretty expensive for some reason
ganbold_ has joined #linux-sunxi
<Tsvetan> post shipping to Russia is EUR 5.50
<Tsvetan> but obviously for some reason you do not want to compare apples with apples
<rm> oh right
<rm> it's down at the very bottom
<rm> I can select "AIRMAIL ROW"
<rm> and it changes to Shipping 8.50 EUR
<focus> Tsevetan: I buy the A20 olxino from Mouser one with and one without flash - cost me total USD 214 :(
<rm> and shows "Insurance: No"
<wens> i remember seeing an article about some postal service, where the package was all busted on arrival
<JohnDoe_71Rus> focus: you want to build some like this http://farm4.static.flickr.com/3605/3401919852_1e842623e5_o.jpg ? google your competitor
<rm> which likely means no tracking number
<Tsvetan> rm same like China post
<rm> whereas stuff from Aliexpress always shipped with one
<rm> always
rellla has quit [Quit: Nettalk6 - www.ntalk.de]
<rm> even if you buy an $2 gadget, they send registered and give you a tracking number
<wens> rm: Aliexpress is a huge company :D
<libv> .us postage somehow always seem insanely overpriced and very intransparent
<rm> it's not Aliexpress that actually ships, it's individual sellers (the site is just kind of like eBay)
<focus> JohnDoe_71Rus: kind of - practical 3D stackable computers with cables and cooling issues removed - i find way to do it :)
eebrah_ has joined #linux-sunxi
<rm> why they use tracking, is because their payment Escrow system depends on it
<Tsvetan> libv they are state owned here and although charge more they are still not profitable
<libv> you get 50 options for any us package, but you only get the premium insured ones, which tend to cost more than the object is worth
<focus> JohnDoe_71Rus: don't want to build racks - want to build genuine 3D stackable - with deep stacking like 4 x 4 x 4 arrays and not have any problems
<libv> then you get to pay customs when it arrives, on both the object _and_ the postage
<rm> Tsvetan, relying on distributors is a very poor choice in reaching customers in Russia
<rm> choice for*
<wens> rm: more like a site wide policy then?
<rm> yes
<Tsvetan> rm Russia is not normal country (sorry no offence) to give as example
<Tsvetan> your customs is under heavy corruption
<rm> reasons: they all tend to be in Moscow only, tend to be very greedy (+30...50...100% markup)
<rm> also: post while slow, actually works
<Tsvetan> our Russian distributors do not import directly but through 3rd party KGB company which only knows how to officially smuggle goods
<rm> everyone used to ordering stuff abroad
<rm> because NO TAX FOR UP TO 1000 EUR and up to 31KG of goods per month per person
<JohnDoe_71Rus> Tsvetan: I ordered by mail
<rm> so my suggestion is, check why the site shows 8.50 EUR and not 5.50 EUR like you said to me
<rm> and
<rm> check if you can send registered, is it really that much more expensive
<Tsvetan> rm as you said Russia is big country
<Tsvetan> part of it is in EU where tax is 5.50 part is close to Japan where tax is 8.50
<JohnDoe_71Rus> 69$ with shiped
<kriegerod> part of Russia in EU?
eebrah_ has quit [Read error: Connection reset by peer]
<Tsvetan> rm I do not use Bulgarian post anymore, but DHL global mail
<Tsvetan> kriegord - yes the part till Ural mountain is still geographically in EU
<kriegerod> in Europe, but not Europe Union
<rm> the website doesn't seem to
<Tsvetan> EU = Europe
<rm> update itself if I type different cities
<rm> just select "Russia", bam, 8.50 EUR
<rm> no matter which city
<arokux> wens: where does allwinner know that dw in u-boot is broken, do you believe them?
<kriegerod> Tsvetan: and taxes rely on Ural boundary, not on countries borders?
<focus> best way to get rid of corruption is to have a government no tax policy for anything below USD500 - everyone will complain bitterly and try to reduce package sizes when they encounter corruption putting pressure all of the time :)
<Tsvetan> focus you never lived in ex-communist country if you believe this
<focus> the buyer can make money and use it to fight corruption
<Tsvetan> now in Bulgaria the government is dominated by the ex-communists
<focus> Tsvetan: a friend of a friend got killed
<Tsvetan> people are on the streets 180+ days protesting against
<Tsvetan> and they do not care
<arokux> Tsvetan: wow, didn't know something like this happens in Bulgaria
<arokux> Tsvetan: and what about elections?
<Tsvetan> they do not resign
<Tsvetan> nor they care that thousands people every day go and protest against them
<Tsvetan> focus so far about the pressure against the corruption :)
<arokux> this is happening in Kiev now, btw
<JohnDoe_71Rus> Russian mail. Experiment sent the parcel from Ukraine to Moscow. In Russian, but the pictures speak for themselves. http://habrahabr.ru/post/185510/
eebrah_ has joined #linux-sunxi
<focus> Tsevtan: I met some Russian buyers for a project i was trying to do - they drink too much + don't want to do straight buy sell - its always like come to russian, the grass is greener etc :)
<focus> I gave up
<oliv3r> wolfy: personally, i'm holding my breath for cubietruck/lime :)
<arokux> focus: +1
<arokux> oliv3r: you've got you ct already!!
<oliv3r> Tsvetan: yeah, wasn't OIN something in combination with IBM trying to protect Linux (with patents?)
<TomiK> hello everybody :)
<focus> lime i will buy x 2 when its out
<libv> oliv3r: lime should sell real well, as it is a proper rpi competitor (and a properly open one at that)
<TomiK> how could I specify compiler that my sun3i is an arm7tdmi ?
<TomiK> in the u-boot + spl file, of course
<libv> and cubietruck covers the top end of allwinner boards.
<libv> if anyone cannot find what he needs between the olimex and cubie devices, then he has pretty wrong expectations :)
<oliv3r> arokux: i ment as a mediacenter PC :)
<arokux> have you actually understood http://habrahabr.ru/post/185510/
<oliv3r> libv: it will depend on proper distribution though, e.g. 'cheap' on amazon, 'cheap' on mouser etc.
<arokux> the parcel was sent across the Russian for customs control and then back!
<arokux> :DDD
<oliv3r> if it's 30 USD without nand (before taxes, I know i know) so 36 after, on amazon/mouser etc; yeah it will annhilate the pi
<oliv3r> but with the example calculations I saw here, noway
<JohnDoe_71Rus> arokux: they send mobile phone and trase it
<arokux> ( JohnDoe_71Rus: I speak russian)
boycottg00gle has joined #linux-sunxi
<JohnDoe_71Rus> arokux: right?
<arokux> JohnDoe_71Rus: правда-правда)
<JohnDoe_71Rus> :)
<oliv3r> TomiK: oh good issue; not sure
<TomiK> oliv3r> :)
<TomiK> my dummy bootloader is compiled with -mcpu=arm7tdmi option, but how to tell it u-boot ? :)
<wens> arokux: do they know?
<arokux> wens: what? you told aw told you dw is broken in u-boot. where do they know?
<wens> arokux: i said dw @ Gbit is broken/not working in u-boot, based on our tests
<wens> arokux: it works fine @ 100Mbps
<arokux> wens: you also told about some e-mail from aw?
<wens> arokux: benn sent mail to shuge @ allwinner asking about gmac specifics, and cc'ed me
<arokux> wens: did he reply?
<wens> arokux: reply from allwinner was, specifics on gmac are not released to public.
<wens> most likely NDA issues :(
eebrah_ has quit [Read error: Connection reset by peer]
<arokux> wens: ...
<arokux> wens: you should ask him about some concrete details in private e-mail, maybe he'll reply
<oliv3r> yay, my post got +4ed ;)
<oliv3r> TomiK: i don't know, sun4+i are all armv7 based; you basically god armv5 it hink that is?
<libv> oliv3r: what? intellegent lifeforms at /.?
<oliv3r> libv: Informative!
<oliv3r> so atleast i finally got a /. account sorted to my name
<TomiK> oliv3r> i'll try
<oliv3r> TomiK: so you may have to define a whole new platform sun3i
<TomiK> ok and create an armv5 or arm7tdmi directory in arch/arm/cpu ?
eebrah_ has joined #linux-sunxi
tomboy64 has quit [Remote host closed the connection]
eebrah_ has quit [Read error: Connection reset by peer]
JohnDoe_71Rus has quit [Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org]
naobsd has joined #linux-sunxi
<oliv3r> TomiK: follow it as the rest goes I suppose; i do'nt know; but i do know there's armv5/arm7 support there allready isn't there
<TomiK> oliv3r> ok :)
eebrah_ has joined #linux-sunxi
eebrah_ has quit [Read error: Connection reset by peer]
enrico_ has joined #linux-sunxi
eebrah_ has joined #linux-sunxi
Black_Horseman has joined #linux-sunxi
popolon has joined #linux-sunxi
prasannapete has quit [Quit: prasannapete]
Black_Horseman has quit [Quit: Zwi se logou mou!!!]
rwmjones has joined #linux-sunxi
<rwmjones> has anyone managed to boot any upstream or sunxi-next kernel on the cubietruck?
<rwmjones> I can't work out what *load commands to use (the ones on the wiki don't work ..)
<rwmjones> it usually hangs very early, after:
<rwmjones> CPU1: failed to boot: -38
<rwmjones> bounce pool size: 64 pages
<rwmjones> Could not switch to high resolution mode on CPU 0
<rwmjones> (and that's it)
<mripard> rwmjones: would you have the full bootlogs ?
<rwmjones> that's about it, but I'm going to post everything to the linux-sunxi mailing list in a min
<mripard> could you put me in CC ?
fredy has quit [Excess Flood]
fredy has joined #linux-sunxi
<rwmjones> sure, I will do
<mripard> thanks
Taduro has quit [Ping timeout: 272 seconds]
<rwmjones> great, sent, thanks
vicenteH has quit [Ping timeout: 245 seconds]
<oliv3r> rwmjones: cpu1 failed to boot isn't serious :)
<rwmjones> yup, that's not what I'm worried about :-)
<rwmjones> I also tried the non-SMP kernel, but they're all failing really early
<rwmjones> what I suspect is I just have the addresses wrong or something really simple/stupid
<rwmjones> a good question might be, which upstream kernel *should* I be using, given I just want something to work
<oliv3r> afaik i haven't played with allwinner hardware in a few weeks; but last i checked, mainline booted fine, as I submitted the dtb myself ;)
<rwmjones> I tried Linus's git, but no dice, hanging very early
<rz2k> rwmjones: do you have latest sunxi u-boot?
<rwmjones> I'm using hansg's uboot on his Fedora AllWinner images, they are the sunxi uboot, and relatively recent, but perhaps not the absolute latest
<oliv3r> rwmjones: you need u-boot from git from linux-sunxi, and make sure you get a dtb from linux-sunxi
<oliv3r> i don't think those patches have been merged upstream yet
<oliv3r> try sunxi-next from our git, it should have all our submitted work
<rwmjones> ok .. how does uboot affect the kernel?
<libv> hrm. upstream fetish again.
<rz2k> rwmjones: we have 2gb enabled u-boot available, maybe the one from Fedora is too old
NotRelevant has joined #linux-sunxi
<rwmjones> ok, but this uboot boots the 3.4.61 kernel fine
<rwmjones> with 2GB of ram
<oliv3r> rwmjones: i don't htink hansg's supports cubietruck
<rz2k> rwmjones: checkout https://github.com/linux-sunxi/u-boot-sunxi branch:sunxi and try again.
<rwmjones> it does
<oliv3r> unless hansg has added new flavors, i don't think it does
<rwmjones> it really does, I'm using it right this minute
<rwmjones> Linux cube.home.annexia.org 3.4.61.sun7i+ #212 SMP Sun Oct 13 12:11:09 CEST 2013 armv7l armv7l armv7l GNU/Linux
<oliv3r> oh shit nightlies are broken!
<mripard> libv: it would be great if you could just like... ignore us.
<rwmjones> the problem is, I want to make a KVM development platform
<oliv3r> mnemoc: nightlies are broken! :p
<mripard> it would give us a very appreciated break
<mripard> and you would even have more time to hack on mali
<mripard> wonderful, isn't it?
<libv> if i'd be following that fetish, i'd never get to mali.
<oliv3r> i can understand where both of you are comming from :)
<oliv3r> libv basically says, random users who wants to run the latest and greatest shouldn't be messing with that unless they know what they are doing
<oliv3r> right, time to fix u-boot nightlies
<oliv3r> why oh why did it break
<mripard> libv: feel free to follow your own fetish. Just don't blame/mock/whatever us for having different one.
<mripard> rwmjones: somehow, it looks like you mixed a device tree from one branch with a kernel of another.
<mripard> sunxi-next-* are just development branches, so things might be quite rough around the edges
<rwmjones> so the first question: given uboot seems to work, do I need to compile another one?
<mripard> you should probably try a fresh v3.13-rc1
<mripard> and start from there
<rwmjones> or can uboot somehow affect the kernel after the kernel start?
<rwmjones> mripard: so you mean, an upstream linus tree kernel, is that right?
<mripard> nah, it looks fine for me, but I don't own a CT myself
<mripard> rwmjones: yep
<rwmjones> ok, I'll try compile it again and make sure I have the right dtb
<rwmjones> do the load addresses look OK?
<rwmjones> and in fact, what do all the addresses _mean_?
<mripard> it's the address at which you load the kernel and DT to
<rwmjones> right, I get that, but how do you know what the memory layout is?
<rwmjones> and how big the uncompressed kernel is
<oliv3r> hno: you committed something 13 hours ago and now the nightlies don't build anymore? I'll try and see if i can find what's up
<rwmjones> and where it moves to (assuming it moves itself, as on x86)
<rwmjones> and the compiled in LOADADDR has no resemblence to any of the other addresses
<oliv3r> mripard: didn't you get one?
<mripard> for the memory layout, it's usually in the SoC datasheet.
rz2k has quit [Ping timeout: 245 seconds]
<mripard> how big the uncompressed kernel depends on your configuration
<mripard> but it should be less than 10MB
<mripard> it moves itself to the RAM base + 0x8000 iirc
<mripard> (RAM base in our case being 0x40000000)
<mripard> and you could use other RAM adresses to load the kernel to
<rwmjones> does it uncompress itself to 0x40008000?
<rwmjones> or does it move there, then uncompress itself somewhere else?
<mripard> I'm not really sure about the exact details, but I'd go for the former option
wolfy has quit [Quit: Primul barbat care s-a insurat poate fi iertat. Ca nu stia ce-l asteapta. Ceilalti n-au nici o scuza. (Sacha Guiry)]
JohnDoe_71Rus has joined #linux-sunxi
panda84kde has joined #linux-sunxi
eebrah_ has quit [Ping timeout: 248 seconds]
<oliv3r> mnemoc: looks like the build only fails on the buildserver, building locally with gcc-4.7.3 works fine
eebrah_ has joined #linux-sunxi
<mripard> rwmjones: also, please try to stick with tags when using linus' branch
<mripard> oliv3r: have you tested the 3.13-rc1 on the CT ?
<mripard> rwmjones: also, what configuration are you using ?
<mripard> oliv3r: oh, btw, I just took my plane tickes for Brussels.
<rwmjones> ok; no however I will try 3.13-rc1
<oliv3r> mripard: nope; only whatever i had a while ago
<oliv3r> mripard: i really hope and pray to the linux gods that I have a bit more time soon
Seppoz has quit []
<mripard> oliv3r: ok
<mripard> I tested it yesterday on my A20-olinuxino, and it works, so I guess it's either something in the u-boot configuration/load process, in the cubietruck DT, or the configuration
<oliv3r> mripard: i will try really hard to check it wednesday
TomiK has quit [Quit: KVirc a enfin son Script! [AndarisScript 4.3 sur www.andaris.net]]
<mripard> oliv3r: and to answer your question, I actually had one, but a colleague of mine wanted to start working on the Allwinner SoCs, so I gave it to him insteda
<mripard> and am still waiting for his patches ...
<oliv3r> mripard: you told me and I forgot
TomiK has joined #linux-sunxi
<oliv3r> mripard: make him work harder! :p
<oliv3r> like I should
<mripard> oliv3r: he's too far away for me to whip him unfortunately ...
<oliv3r> digitally whip him!
eebrah_ has quit [Read error: Connection reset by peer]
eebrah_ has joined #linux-sunxi
wingrime has joined #linux-sunxi
NotRelevant has left #linux-sunxi ["."]
eebrah_ has quit [Read error: Connection reset by peer]
rellla has joined #linux-sunxi
<oliv3r> rwmjones: I didn't see env set fdt_high ffffffff in your mailo; are you using that?
<rwmjones> not in the a20-smp test, but yes in the upstream test
<oliv3r> rwmjones: and double check your DEBUG_LL and console output; the second mail looks like it runs fine, but isn't printing the console
<rwmjones> I'm just rebuilding 3.13-rc1 right now, and checking I got the right dtb
<rwmjones> what is DEBUG_LL?
<oliv3r> Lowlevel Debug
<rwmjones> and how do I set/check it?
eebrah_ has joined #linux-sunxi
<mripard> oliv3r: I don't think it's debug_ll related, you wouldn't see anything
<mripard> what's odd, is that, yesterday when I booted it, I had much more verbose logs
<oliv3r> mripard: oh ian got ahci fixed; it's turls' fault!
<oliv3r> i have seen that or similar output in the past
<oliv3r> i probably complained about it here, but have completly forgotten
boycottg00gle has quit [Remote host closed the connection]
<mripard> oliv3r: yeah, I saw that :)
<oliv3r> ok now i'm really exicted about working on it again :)
<oliv3r> mripard: so ian is basically saying we're missing some sata clk enable stuff?
eebrah_ has quit [Read error: Connection reset by peer]
<mripard> oliv3r: that's what I understood :)
<oliv3r> yeah it sounded really trivial
<oliv3r> but i don't know clock stuff :)
<oliv3r> Turl: ^^
eebrah_ has joined #linux-sunxi
Turl has quit [Ping timeout: 252 seconds]
Turl has joined #linux-sunxi
eebrah__ has joined #linux-sunxi
eebrah_ has quit [Ping timeout: 245 seconds]
<Turl> oliv3r: he also mentioned some typo on a shift on the ahci code
drachensun has joined #linux-sunxi
eebrah__ has quit [Read error: Connection reset by peer]
eebrah_ has joined #linux-sunxi
zeRez has quit [Quit: zeRez]
eebrah__ has joined #linux-sunxi
eebrah_ has quit [Read error: Connection reset by peer]
<binaryferret> Apart from Fedora are there any other linux-sunxi images?
<buZz> millions
<binaryferret> Huzzah
HeHoPMaJIeH has quit [Quit: Leaving]
<arokux> binaryferret: Bootable OS Images at wiki
<arokux> +1
<buZz> +100
<binaryferret> Yeah, I stupidly wrote and pushed enter without checking the site. Sorry
<buZz> :)
<buZz> binaryferret: thats ok, you get three strikes
<buZz> >:)
<binaryferret> I think that's my second then.
<arokux> :) binaryferret, I do not understand what is stopping you from grabbing _any_ rootfs
<buZz> +5
<arokux> -1
<binaryferret> Nothing, bar lazyiness due to never putting my own rootfs together
<rm> a bootable image "for board XX" at least somewhat guaranteed to boot
<binaryferret> I do plan to move to doing my own rootfs, but I need to figure out why this one A13 tablet I have won't play nice with the fedora image that works on my other two.
<arokux> rm: the kernel guarantees that, hm.. but true, some times kernel should have some features on for userspace to work.
<rm> useful to quickly get your board booting and working, and think about further experiments after that
<arokux> Charbax behaviour is so ugly
<arokux> "wha are you doing?!?"
<arokux> lkcl: started allwinner revolution, and now there are dozen of boards
<arokux> is*
vicenteH has joined #linux-sunxi
<buZz> cool board
<slapin> if only it had full eSATA...
eebrah__ has quit [Read error: Connection reset by peer]
<Turl> rwmjones: ping
<rwmjones> pong
<rwmjones> is env set a real command?
<Turl> rwmjones: your bootm is missing two arguments
<rwmjones> uboot seems to give an error
<rwmjones> ok, what should it be?
<Turl> rwmjones: bootm 0xkernel 0xramdisk 0xdt
<Turl> if you don't have a ramdisk, you can just put a dash in place (-)
eebrah_ has joined #linux-sunxi
woprr has quit [Read error: Operation timed out]
<rwmjones> I'm using:
<rwmjones> bootm 0x46000000 - 0x49000000
<rwmjones> which is ok right?
<Turl> rwmjones: yeah, but your uEnv.txt isn't using that
<rwmjones> Turl: I don't understand .. uEnv.txt doesn't mention bootm
<Turl> rwmjones: exactly
<rwmjones> so I still don't understand
<Turl> bootcmd=if run loadbootenv; then echo Loaded environment from ${bootenv}; env import -t ${scriptaddr} ${filesize}; fi; if test -n ${uenvcmd}; then echo Running uenvcmd ...; run uenvcmd; fi; if run loadbootscr; then echo Jumping to ${bootscr}; source ${scriptaddr}; fi; run setargs boot_mmc;
<Turl> that's the boot command embedded on uboot
<Turl> bootenv is uEnv.txt
<rwmjones> I still don't understand
Taduro_ has joined #linux-sunxi
<rwmjones> should I just forget about uEnv.txt?
<rwmjones> it's the exact same one which is in the hansg Fedora image
<arokux> rwmjones: sunxi-3.4 (kernel in fedora image) and mainline are two very different kernels and need to be booted differently.
<Turl> rwmjones: yeah but hans' image uses 3.4 and can boot with the one argument incantation (default)
<rwmjones> is this documented anywhere?
<arokux> rwmjones: of course!!!
<arokux> sunxi.org
<rwmjones> imagine I have no knowledge whatsoever about uboot, and that would be approximately correct; but I'm willing to read documentation if it's available
<arokux> rwmjones: and what is your ultimate goal?
<rwmjones> provide a KVM development platform
<rwmjones> for a bunch of KVM developers
<rwmjones> (ARM-based, obviously)
<arokux> rwmjones: is the connected to Xen? there is something about it at our wiki.
<rwmjones> nope, KVM
<focus> arokux: merri sounds like witstech or witstech engineers in a new shell company - the shipping price is more than board price :)
<arokux> :D
<focus> i ran away screaming
<arokux> focus: then buy from there house and let the other ppl ship!
<rwmjones> let me try those uboot commands from the prompt and see if it makes any difference
Taduro_ has quit [Ping timeout: 245 seconds]
<focus> paid witstech $1000 x 2 for linux systems and they 'promise' Linux support - when it came to the crunch they say only android and when i say android runs on Linux they hardly understand and don't offer a full supported linux
<focus> this get one of my engineers sacked - very annoyed - not his fault!!
<arokux> focus: and you understood - your only hope is sunxi.org? :)
<rwmjones> Turl: how do I specify the linux kernel command line using uboot commands?
<focus> months later I found sunxi through tom's aliexpress store and a link there for support
BluesBoy has quit [Quit: Want to be different? Try HydraIRC -> http://www.hydrairc.com <-]
<Turl> rwmjones: stuff them into the bootargs variable
<Turl> setenv bootargs blahblahblah
<rwmjones> and console should be tty0 or ttyS0 or something else?
<rwmjones> (I mean console=...)
<Turl> ttyS0,115200 for the first uart
Taduro has joined #linux-sunxi
<focus> arokux: only only only hope for allwinner is sunxi
<focus> :)
<arokux> focus: for allwinner? don't thinks so. for this dozen of small companies doing dev boards - yes.
<rwmjones> nah, still hanging at
<rwmjones> [ 0.252994] drivers/rtc/hctosys.c: unable to open rtc device (rtc0)
<Turl> rwmjones: did you expect much more without a root filesystem? :p
<Turl> you should be able to see the panic though
<focus> arokux: they all get their linux experience from sunxi?
<Turl> unless you passed rootwait, then it's going to just wait
<arokux> focus: yes. but allwinner simply doesn't care. this is my understanding at least.
<rwmjones> I'd expect a bunch more messages .. not really expecting a full boot at the moment, just something that panics trying to mount root would be a start here
woprr has joined #linux-sunxi
eebrah_ is now known as eebrah
<Turl> rwmjones: you can try enabling DEBUG_LL and EARLY_PRINTK and passing earlyprintk on kernel command line
<rwmjones> these are kernel CONFIG_* options right?
<rwmjones> let have a look ..
<torbenh3> passing in "initcall_debug logleval=9" is also pretty helpful
<Turl> rwmjones: yes
<focus> arokux: may be growing fat on android and ignoring the hand that feeds them which happens to be linux at the core of android
tomee^ has joined #linux-sunxi
<arokux> focus: they focus on things that bring them money, not on philosophy :)
<Turl> brb
<arokux> focus: but at least they can release docs, but even this doesn't happen.
<focus> arokux: shame really - some day a true Linux company will come along with everything open and they and fellow non sharing companies all go down with one flush :)
<torbenh3> i highly doubt that :S
eebrah has quit [Read error: Connection reset by peer]
<arokux> focus: and you do believe in this?
<torbenh3> the chips would be too expensive.
<focus> arokux: if I make money i do it :) - i design digital version of quantum computer - at least 10 to 100x faster with today's chip tech with about 5% or thereabout extra silicon
<focus> programming is 4 x complicated
<focus> can't run Linux directly, but then i show everyone how to write proper software :)
<focus> need a mil order to make it real
eebrah has joined #linux-sunxi
eebrah is now known as Guest76827
<focus> <deep snort!> :)
<Turl> rwmjones: any luck?
<rwmjones> compiling still
<Turl> :p
<rwmjones> I should cross-compile it really, but it's nearly done ..
<Turl> you're compiling on target? ouch :)
Guest76827 has quit [Ping timeout: 246 seconds]
Fusing has joined #linux-sunxi
eebrah_ has joined #linux-sunxi
woprr has quit [Ping timeout: 245 seconds]
<buZz> focus: you need to write qinux for it
<buZz> quantum linux
<buZz> of course
<buZz> with inf timestamps on all the files
<buZz> etc
* slapin is tired of quantum things over there so much
woprr has joined #linux-sunxi
rellla has quit [Quit: Nettalk6 - www.ntalk.de]
fredy has quit [Excess Flood]
fredy has joined #linux-sunxi
Quarx has quit []
rz2k has joined #linux-sunxi
Fusing has quit [Ping timeout: 240 seconds]
FR^2 has quit [Quit: Connection reset by peer]
_massi_ has quit [Remote host closed the connection]
paulk-collins has joined #linux-sunxi
popolon has quit [Quit: Quitte]
<hno> oliv3r, yes.. something... what is the error message?
<hno> Merge branch 'master' of git://git.denx.de/u-boot into sunxi
n01 has quit [Quit: leaving]
geecko has quit [Ping timeout: 272 seconds]
deasy has quit [Ping timeout: 272 seconds]
eebrah_ has quit [Read error: Connection reset by peer]
<wingrime> more one dev board
<wingrime> heh
<wingrime> they are like mushrooms
<wingrime> after rain
netlynx has joined #linux-sunxi
panda84kde has quit [Quit: Konversation terminated!]
eebrah has joined #linux-sunxi
eebrah is now known as Guest41005
Guest41005 is now known as eebrah_
wingrime has quit [Ping timeout: 272 seconds]
eebrah__ has joined #linux-sunxi
eebrah_ has quit [Ping timeout: 246 seconds]
n01 has joined #linux-sunxi
n01_ has joined #linux-sunxi
rz2k has quit []
eebrah_ has joined #linux-sunxi
eebrah__ has quit [Read error: Connection reset by peer]
tzafrir has quit [Ping timeout: 246 seconds]
<mnemoc> arokux: should I stop tracking that? refs/remotes/origin/sunxi-next-usb does not point to a valid object!
<mnemoc> aaaaarg...
<mnemoc> $ git fetch origin
<mnemoc> error: refs/remotes/origin/sunxi-next-usb does not point to a valid object!
JohnDoe_71Rus has quit [Quit: KVIrc 4.0.4 Insomnia http://www.kvirc.net/]
geecko has joined #linux-sunxi
LoCoZeNoz_ZUE has quit [Quit: Leaving]
eebrah_ has quit [Quit: leaving]
<oliv3r> hno: i only looked at the .txt files of the dl.linux-sunxi.org nightlies, they seem to build fine here
<oliv3r> mnemoc: stuff. broken. boom!
rellla3 has joined #linux-sunxi
wingrime has joined #linux-sunxi
<mnemoc> oliv3r: give me one error link please
<mnemoc> Fetching origin
<mnemoc> error: refs/remotes/origin/sunxi-next-usb does not point to a valid object!
<mnemoc> ^--- still trying to fix those.... git fsck didn't solve it
<mnemoc> for x in */.git/..; do cd "$x"; pwd; git fsck; cd - > /dev/null; done
<mnemoc> doing it again
tinti has joined #linux-sunxi
<hno> mnemoc, where are build logs for the nighly u-boot builds?
<oliv3r> every single one of them
<oliv3r> hno: ^ :p
<oliv3r> make[1]: *** No rule to make target `/srv/build/amery/allwinner/sunxi-nightly/u-boot-sunxi/include/linux/config.h', needed by `/srv/build/amery/allwinner/sunxi-nightly/build_u-boot/build_cubietruck/board/sunxi/board.o'. Stop.
<hno> Right. just found them. Should have looked first..
<hno> Ok, that's already fixed I think.
<hno> so just wait for next nightly build.
<oliv3r> ah ok nvm then
<hno> then look again.
<oliv3r> actually
<oliv3r> doesn't the nightly build every hour after each cvommit?
Nyuutwo has quit [Ping timeout: 240 seconds]
<mnemoc> I froze it to fix the freaking refs/remotes/origin/sunxi-next-usb problem
<mnemoc> but I'll disable linux for now and restore the loop
<hno> I don't know. Would be nice if it mentioned the git rev it's building.
<mnemoc> it does. -latest is a symlink
<hno> ah.
Nyuutwo has joined #linux-sunxi
<mnemoc> per-board -latest doesn't really scale well :(
<hno> Hm.. that is the latest.
<mnemoc> so I'm doing it per commit alone... sadly that keeps broken builds in the -latest
<mnemoc> with the kernel I can keep -latest pointing to the last that built correctly
<hno> Hm.. why are the nightly build failing but not mine or oliv3r..
<mnemoc> I'll force a fresh rebuild
<hno> Is the nightly build in-tree or out-of-tree?
<oliv3r> i only built cubietruck
<oliv3r> i've built using the bsp though
<mnemoc> hno: O=
<hno> so out-of-tree?
<hno> same as me then.
netlynx has quit [Quit: Leaving]
<hno> mnemoc, ok. I do suspect this is a stale dependencies issue and will be fixed by a fresh build.
<hno> Date: Mon Oct 7 16:04:18 2013 +0900
<hno> include: delete include/linux/config.h
<mnemoc> $ rm -rf build_u-boot/
<mnemoc> <3 O=foo/
<mnemoc> sad xen doesn't have such thing
<hno> what?
<mnemoc> O=
<hno> I don't follow
<mnemoc> on u-boot or linux you can build on a dedicated dir away from the sources
<mnemoc> xen can't
<hno> and you can't when building xen??
<mnemoc> you do?
<hno> I don't use xen, even less build it.. prefer kvm.
<hno> but now I must build it... must see this for myself :)
<mnemoc> distclean on each iteration :\
n01 has quit [Read error: Connection reset by peer]
n01_ has quit [Read error: Connection reset by peer]
n01 has joined #linux-sunxi
n01_ has joined #linux-sunxi
shineworld has joined #linux-sunxi
pfdm has joined #linux-sunxi
tzafrir has joined #linux-sunxi
<mnemoc> good
<mnemoc> rm -rf build_u-boot/ fixed it
<mnemoc> and `rm -v linux-sunxi*/.git/refs/remotes/origin/sunxi-next-usb` fixed the other problem :p
<oliv3r> +5 points
<hno> gah.. xen is an odd project. Claims using autoconf only to masquerade to be a good citisen, but dont actually use it at all other than to specify prefix I think. And indeed, no out-of-tree support anywhere in sight.
shineworld has quit [Quit: Leaving]
<[7]> any idea what could cause this?
<[7]> Load /media/theseven/f40355b3-9e2c-45a4-b228-5a20f9e538df/tmp/cubie/LiveSuit/bin/tools.scj failed
<[7]> ./image/ASuitImage.cpp, LoadToolsPlug failed 206
<[7]> official images work, just not my homebrew ones
eebrah has joined #linux-sunxi
<[7]> I tried both dragon and awimage for packing, same result
<hno> oliv3r, have you tried runnin the new u-boot?
<oliv3r> hno: not yet
AreaScout has quit []
n01 has quit [Ping timeout: 246 seconds]
n01_ has quit [Ping timeout: 272 seconds]
rellla3 has quit [Quit: Nettalk6 - www.ntalk.de]
zeRez has joined #linux-sunxi
<techn__> hno: fel boot seems to atleast start
<techn__> I have troubles with gmac though
<techn__> .. still.. this is not related to new u-boot
Black_Horseman has joined #linux-sunxi
Black_Horseman has joined #linux-sunxi
Black_Horseman has quit [Changing host]
eebrah has quit [Ping timeout: 272 seconds]
wingrime has quit [Ping timeout: 248 seconds]
paulk-collins has quit [Remote host closed the connection]
hglm has joined #linux-sunxi
eebrah has joined #linux-sunxi
<hglm> I have A20 device that didn't run stable with 432 MHz DRAM clock (only 408) and default MBUS (memory controller) clock but it is perfectly stable after bumping up MBUS speed to the DRAM speed (using the patch on the mailing list)...
<hglm> ..and increasing MBUS speed to DRAM speed gives a nice performance boost, as does raising DRAM clock from 408 to 432 MHz. I guess it could help to to make MBUS more configurable in u-boot.
naobsd has quit [Quit: Page closed]
techn_ has joined #linux-sunxi
<Turl> hglm: higher mbus caused problems with the mmc controller during ssvb's tests though
<Turl> hglm: we got confirmation from aw that 400MHz should be stable though with a small voltage bump
<hglm> Turl: Yeah but tht was at 480 Mhz which is high...I guess MBUS need to be clocked lower.
eebrah has quit [Ping timeout: 246 seconds]
<hglm> Turl: voltage bump of CPU?
techn__ has quit [Ping timeout: 240 seconds]
eebrah has joined #linux-sunxi
<Turl> hglm: dcdc3, 1.25v->1.3v
<hglm> Turl, I guess that voltage bump is not too dramatic...does the CPU voltage scaling in the kernel affect dcdc3?
<Turl> hglm: no, dvfs is done on dcdc2 iirc
<Turl> (that's the cpu voltage)
<hglm> Ah OK
<hglm> So dcdc3 is basically the DRAM voltage?
<Turl> I suppose so
<Turl> maybe the dram controller's too
notmart has quit [Quit: notmart terminated!]
<hglm> I was also able to reduce the CAS timing for the DRAM from 9 to 6 cycles (u-boot DRAM configuration), at 432 MHz -- gives another small boost in performance.
<Turl> hglm: I'll send a series of patches for uboot in a bit
<Turl> (mbus)
<hglm> Sound good.
<hglm> It would be fun to try DRAM at 480 MHz (which is very good for performance) while keeping MBUS lower at about 400 MHz -- if that's stable that would be nice.
eebrah has quit [Ping timeout: 265 seconds]
eebrah has joined #linux-sunxi
<hglm> On the subject of performance, I am not quite sure what the point is of Allwinner's fantasy (fantasys on Android) CPU speed governor...all it does is pretend the CPU keeps running at 912 MHz, but actually slows down every 10 seconds or so (which is not reported in the /sys fs). I get the impression it was an ugly hack to combat ill-perceived hardware problems, but in my experience is completely unnecessary...for example a tweaked ondemand
eebrah has quit [Read error: Connection reset by peer]
printallthething has joined #linux-sunxi
eebrah has joined #linux-sunxi
eebrah has quit [Read error: Connection reset by peer]
Black_Horseman has quit [Quit: Αποχώρησε]
<tomee^> hglm: maybe a stupid hack to circumvent all those "cpu tweakers" for android? no idea, didn't look into the code.
<tomee^> hglm: btw I took the liberty of editing some of the wiki pages you edited recently
Black_Horseman has joined #linux-sunxi
Black_Horseman has quit [Changing host]
Black_Horseman has joined #linux-sunxi
<hglm> tomee: In Anroid I think the fantasys governor's main task is to maximize the CPU speed when a benchmark (such as Antutu) is detected...my tablet had a completely broken Android chip configuration, runs at 720 MHz but when Antutu runs (and only then) it goes to 1008 MHz...
<hglm> My tablet was rooted, and I can easily the change the governor, but if it is not "fantasys" the kernel keeps spewing "Governor is not fantasys" warnings...
<tomee^> so it sounds like a marketing trick one way or the other, oh well... ehh
<tomee^> I, on the other hand, managed to configure spawning of vdpau-enabled mplayer from xbmc (not so easy as one would have thought). So now I have a set top box capable of playing almost everything I want ;)
<tomee^> If someone doesn't come up with a cleaner solution, I think I might post my build configs and scripts within a couple days
eebrah has joined #linux-sunxi
<hglm> tomee: I haven't tried video decoding on Allwinner yet, only disp layer color conversion backends, but it does sound interesting...does VDPAU work OK?
<tomee^> well, the credit goes to the author, I just compiled and messed around with it
<tomee^> but looks like ~60fps for 720p 1500kbps h264/mkv
eebrah has quit [Read error: Connection reset by peer]
<tomee^> I even accidentaly had a SD video run in the background and mplayer+VDPAU on top of it at the same time ;)
<tomee^> there are issues with sound, though, my apli won't pick up on ac3 passthrough when seeking in the video file (randomly)
<hglm> I have played mainly with gstreamer with xv backend (which is accelerate YUV conversion with fbturbo driver) and it's usable on an A20, because chip has reasonable horsepower to decode video up to a reasonable sizde.
<tomee^> that would be all for now
<tomee^> o/
tomee^ has quit [Quit: I Quit.]
naobsd has joined #linux-sunxi
eebrah has joined #linux-sunxi
<[7]> what the heck is "upfly"?
<[7]> some kind of LUA plugin for livesuit?
<[7]> are there different versions of livesuit that are only compatible with different sets of image files?
* [7] slaps olimex hard
<[7]> why does the alleged source code for olinuxino a20 android 4.2.2 contain a livesuite packer that generates a fundamentally different image format than the binary image that they released!?
adb_ has quit [Ping timeout: 252 seconds]
naobsd has quit [Ping timeout: 250 seconds]
eebrah has quit [Read error: Connection reset by peer]
naobsd has joined #linux-sunxi
enrico_ has quit [Quit: Bye]
* [7] is about to explode
eebrah has joined #linux-sunxi
eebrah has quit [Read error: Connection reset by peer]
eebrah has joined #linux-sunxi
ganbold_ has quit [Remote host closed the connection]
eebrah has quit [Ping timeout: 252 seconds]
TomiK has quit [Quit: KVirc a enfin son Script! [AndarisScript 4.3 sur www.andaris.net]]
hramrach has quit [Remote host closed the connection]
* [7] is getting a little bit closer, and will push a whole bunch of fixes to awimage later...
froy has quit [Quit: kablam!]
techn__ has joined #linux-sunxi
techn_ has quit [Ping timeout: 252 seconds]