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
tomboy64 has quit [Remote host closed the connection]
\\Mr_C\\ has quit [Quit: .]
tomboy64 has joined #linux-sunxi
<arokux2> mickoo: still compiling?
<Turl> arokux2: you can ask benn :)
<arokux2> Turl: ... he has overwritten 2013 code of Broadcom with 2012 one..
<arokux2> what can he tell me you think? :(
<Turl> arokux2: where he got it from :)
<arokux2> Turl: stolen from some android?
<Turl> arokux2: dunno
<Turl> arokux2: I mean you could always just make a repo and host it, but I'd rather be in the clear
<arokux2> Turl: yeah, you are right. but I just see this code created by Benn and understand that this is allwinner code...............
adb has quit [Ping timeout: 252 seconds]
<arokux2> Turl: how can a normal dev overwrite a 2013 year code by 2012 one?
<arokux2> how?!!!!
<Turl> arokux2: cp stable/* tree/*; git commit -a ? :)
<Turl> err, I wrote an extra *
<arokux2> Turl: I bet he never tried if tree/ worked. how do you think...?
<arokux2> Turl: I wonder what should I do.. that code is sprinkled by allwinner
<arokux2> Turl: should we cherry pick that? or should we really go after allwinner changes and put them into bcmdhd that is our kernel?
<Turl> is bcmdhd on mainline?
<Turl> (or another driver for that same hw?)
<arokux2> Turl: no idea, I'm talking about sunxi-3.4 now.
<arokux2> Turl: (plz, ping me)
<wens> mripard: gmac on u-boot is jemk's work
<Turl> arokux2: if there's a driver already we should just use that
<arokux2> wens: I've got wifi working. network manager has spawned his wpa_supplicant and was interfering.
<wens> arokux2: that's why i used server image. less interference from user settings :p
<arokux2> Turl: so you suggest to go over 40k LOC and find AW specific bits needed to get bcmdhd in sunxi-3.4 to work with cubietruck's wlan adapter? :)
<wens> Turl: brcm80211 in mainline
<arokux2> wens: :p
<arokux2> wens: sdio is used by wlan adapter, it is not yet supported in mainline, right?
<arokux2> ( wens: I use Arch Linux - no interference at all)
<wens> arokux2: no it's not. so SD/SDIO is really important
<arokux2> wens: so the question is should we make our sunxi-3.4 working with wlan of ct?
<wens> and i think firmware can only be loaded through SDIO
mickoo has quit [Quit: http://www.kiwiirc.com/ - A hand crafted IRC client]
egbert has quit [Disconnected by services]
egbert has joined #linux-sunxi
<arokux2> interesting situation
<arokux2> Turl: still up?
naobsd has quit [Quit: Page closed]
arokux2 has quit [Remote host closed the connection]
TheSeven has quit [Disconnected by services]
[7] has joined #linux-sunxi
deasy has joined #linux-sunxi
Quarx has joined #linux-sunxi
printallthething has quit [Read error: No route to host]
printallthething has joined #linux-sunxi
Black_Horseman has quit [Ping timeout: 272 seconds]
<kriegerod> arokux: wifi works with stopped network-manager
<kriegerod> tomboy64: didn't before. so what?
<steev> anyone running the latest stage/sunxi-3.4? i'm running into something odd here, where if i try to do... well, i get a kernel panic when i try to do git reset --hard HEAD^
<tomboy64> i just saw in the backlog arokux, wens and Turl discussing the origin of ct's firmware
<kriegerod> steev: i am running that branch, revision of some days ago
<steev> some days ago wouldn't be latest then :)
<steev> or is it...
<steev> kriegerod: do you have the sun7i: pm.....
<steev> 93e38ba7b610e585878aecd0680409319a27c2f9 that commit id
<steev> and is your rootfs on sata, on a cubieboard2
<kriegerod> steev: could you first share your kernel panic log output?
<steev> if i could actually install a system logger, sure
<kriegerod> i have cubietruck on usb hdd.
tzafrir has quit [Ping timeout: 264 seconds]
<steev> but any time i do anything cpu intensive i get a kernel panic
<kriegerod> steev: take a photoshot of monitor :)
<steev> the kernel panic is quite long, but okay
ncrmnt has joined #linux-sunxi
<kriegerod> steev: could it be power supplying issue?
<steev> doubtful, i've been using the system just fine for... quite a while
<kriegerod> or board configuration issue, like frequency of something is set to unstable value
<kriegerod> steev: well, but in your described case you have it unstable
<kriegerod> steev: what is your board power supply?
<steev> kriegerod: yes, since i *rebooted* into the latest committish it's been unstable. i'm using the USB cord connected to a powered usb hub
<steev> ugh, now i have to figure out how to get it off the phone
<steev> maybe that will work
<steev> and that comes from doing "git status" inside my ~/linux-sunxi folder
<steev> s/folder/directory/
ncrmnt has quit [Ping timeout: 264 seconds]
Fusing has joined #linux-sunxi
wingrime has joined #linux-sunxi
paulk-collins has joined #linux-sunxi
rellla3 has joined #linux-sunxi
<libv> has anyone seen those new cheapo netbooks with an A20?
<libv> they are all over ebay.de
<libv> ah, urgh. not allwinner but wm
n01 has joined #linux-sunxi
<libv> nm.
<moofree> i tried to buy this one, but the seller said it wasn't available (and never took down the listing) http://www.aliexpress.com/item/3G-Tablet-PC-7-inch-Pink-Pocket-Tablet-Android-4-1-BOXCHIP-Dual-Core-1-2GHz/1126423782.html
<libv> moofree: hrm, pink :)
<libv> keyboard is seriously subpar and so is the display...
<moofree> true
<moofree> the wondermedia laptops typically max out at 1024x600, i think i've seen such a device with an a10 in it instead of wm
<JohnDoe_71Rus> nice :)
<libv> i wish lkcl would finish the tablet and skip the router, and go straight to a netbook. but i am sure that heard that sort of whingeing before
<libv> although, the router seems like it is a lot less complicated than the tablet
rellla3 has quit [Quit: Nettalk6 - www.ntalk.de]
tzafrir has joined #linux-sunxi
n01_ has joined #linux-sunxi
JohnDoe_71Rus has quit [Quit: KVIrc Insomnia 4.0.2, revision: 402, sources date: 20100627, built on: 2010-08-02 17:44:33 UTC 402 http://www.kvirc.net/]
<wingrime> oliv3r: ping
<lkcl> libv: yup. i reaaally want a damn netbook.
<lkcl> and yes the router's seeeriously less complicated. the main issue there is that it is extremely tight on space and still has to be single-sided
<libv> :)
<lkcl> ... if you don't mind a 10in netbook (or something using the 1024x768 ipad LCD) then we maaay be able to get away with using the flying squirrel PCB.
<lkcl> just.
arokux2 has joined #linux-sunxi
<arokux2> tomboy64: our firmware is not there
naobsd has joined #linux-sunxi
<arokux2> mnemoc: ping
ZetaNeta has joined #linux-sunxi
<steev> you can get a 10" 1280x720 lcd
<libv> steev: i think it is more about the amount of lines and the frequencies possible on the LVDS connection
<steev> you should be able to do 1280x720 on the lvds
<steev> the mx51 can do it
<arokux2> oliv3r: ping
<steev> in fact, my efika has a 1280x720 in it, it's nice
<libv> steev: the SoC might be able to cope, but that doesn't mean that the pcb is able to do so
<steev> don't use a shitty pcb?
<libv> *sigh*
<steev> dude, if you're designing it yourself
<steev> you should be able to overcome it
<libv> read lkcl his last statements again.
<steev> yes, he said MAY, not we're gonna
<steev> it's much cheaper to use someone else's design, but if you want something done right...
<steev> :)
<arokux2> anybody knows in what relation does AP6210 stands to broadcom chips?
<lkcl> steev: the issue is that everyone we ask either wants stupid amounts of money - efika demanded $250,000 - or they simply don't "get it"
<lkcl> it's too simple for them to comprehend.
<lkcl> so... yep, i'm not a PCB design expert but i'm having to do the PCB designs myself. luckily they're quite simple (all high-speed lines are differential pairs)
<steev> lkcl: well it's outside my skillset to design pcbs, but knowing the people who do, there's a reason they fetch the money that they do
notmart has joined #linux-sunxi
notmart has joined #linux-sunxi
notmart has quit [Client Quit]
<steev> lkcl: that isn't to say i agreed with everything that was said or done at genesi, definitely wasn't the case, but i wasn't a spokesperson
<lkcl> steev: i asked them if they'd like to collaborate - *collaborate* on a design - and they asked for $250k. that's not collaboration!
<lkcl> bill's a really nice guy. unfortunately he has an extremely aggressive technical manager.
<steev> eh
<steev> i don't think i'd agree with the latter statement, but i don't like talking too bad about people
tomboy64 has quit [Remote host closed the connection]
<tomee^> [09:44:42] <steev> kriegerod: yes, since i *rebooted* into the latest committish it's been unstable. i'm using the USB cord connected to a powered usb hub
tomboy64 has joined #linux-sunxi
<tomee^> steev: plugged into OTG? don't do that.
<steev> what?
<tomee^> if we're talking about a cubieboard, that is.
<steev> it's a cubieboard2
<tomee^> then don't.
<steev> what otg port are you talking about?
<tomee^> "5VDC input 2A or USB otg input"
<tomee^> the mini-usb port
<steev> uh
<steev> cubieboard2 doesn't have a mini usb port
<tomee^> and that thing facing you
<tomee^> next to the monstrous RJ45, is what?
<steev> that's hdmi
<steev> er, oh, no i suppose it isn't
<steev> that isn't the port i have plugged in
<steev> DC to USB cable x 1
<tomee^> steev: ok, some vibrators also use the same socket form factor, but I assume if it looks like USB, it is USB.
<steev> it never even crossed my mind to attempt to power it via the miniusb
<ijc> that's a micro usb IIRC.and it is the OTG port referred to in various places.
<steev> right - i'm just clarifying, that isn't what i'm using to power it, i'm using the DC to USB cable that comes with the CB2, it's a barrel connector (to the cb2) to USB which is plugged in to a powered usb hub
<tomee^> steev: to the DC jack next to HDMI?
<steev> yes
<steev> which i couldn't find a decent picture of
<tomee^> ok then.
<tomee^> that is fine, but ... isn't.
<tomee^> how much juice can your usb hub deliver?
<tomee^> most hubs I've seen have a 1A power supply and 4 or 5 USB ports.
AreaScout has joined #linux-sunxi
<arokux2> tomee^: you do not need to upload that image anymore.
<tomee^> ijc: to me it is a mini usb type B. same as used in windows mobile/windows ce devices. micro usb is what you see in most androids. see http://upload.wikimedia.org/wikipedia/commons/thumb/c/cb/Types-usb_new.svg/271px-Types-usb_new.svg.png
<tomee^> arokux2: got it working?!
<ijc> tomee^ ack
<arokux2> tomee^: yes. have you used desktop or server image?
<tomee^> arokux2: omg, you're testing my memory... first the server one, for sure.
<tomee^> arokux2: but which kernel I ended up with I don't remember. but you md5-ed it and it was the same that you used, no?
<arokux2> tomee^: kernel is the same
<tomee^> arokux2: so it was userland trouble?
<tomee^> arokux2: network manager kicking in? lack of wpa supplicant? wrong dhcp client configuration hooks? time skew?
<arokux2> tomee^: the problem was network-manager that spawned its own wpa_supplicant which interfered with mine
<tomee^> HAHA
<tomee^> yeah
<tomee^> reminds me of fixing various devices without first checking if they are powered up ;p
rz2k has joined #linux-sunxi
<tomee^> arokux2: anyway, congratulations on the success :) my ct is no longer a unicorn in that area (wifi)
<arokux2> ;)
<tomee^> arokux2: so now we can join our efforts in making the bluetooth and wifi direct work ;)
<tomee^> arokux2: and where do I send this LCD again? so you gain interest in the hardware acceleration part? :)
<arokux2> tomee^: sorry, I cannot promise I have enough time for that :)
<tomee^> neither do I ;-)
eebrah has joined #linux-sunxi
<tomee^> but a common goal is always encouraging ;)
<tomee^> steev: so, have you checked the usb hub part?
<steev> as much as i can - i'm not exactly an EE over here, it powers the BBB just fine, and there isn't anything else plugged in (including the BBB)
<tomee^> steev: google up usb specs. no more than 500mA is allowed or it breaks the standard (and is dangerous to a device)
<steev> tomee^: well i switched to a different powered usb hub - same thing, go into linux-sunxi, run git status, and kernel panic
<tomee^> steev: so your usb hub should not and probably can not deliver more than 500mA. while cb2 power rating is 2A. so my guess would be it needs at least 1A + whatever you plug into its USB ports (2x500mA)
<steev> there is nothing else plugged in to it's usb ports
<steev> i only ever ssh into it though i do have it plugged in to a monitor
<tomee^> steev: decrease scaling_max_freq to like 600-720mhz.
<tomee^> steev: and see if it still panics.
<steev> or, try backing out the change that i don't think is correct in the kernel, and go from there
<tomee^> ... or that ;D
<steev> oh hey, that time just booting it caused the kp :D
<tomee^> which commit do you suspect btw?
<steev> the latest one
<steev> i mentioned the committish of it earlier
tomboy64 has quit [Remote host closed the connection]
<steev> and seriously, it's 2013, who the hell is still building with gcc 4.4, and more importantly, WHY. even android uses 4.6 now
tomboy64 has joined #linux-sunxi
<tomboy64> arokux2: can you give me the device's pci-id?
TomiK has quit [Quit: KVirc a enfin son Script! [AndarisScript 4.3 sur www.andaris.net]]
<arokux2> tomboy64: of what?
<tomboy64> the wlan and the bt device
<steev> hmm, another possibility could be that i built the kernel with 4.8
<arokux2> tomboy64: it is not connected with pci interface
<tomboy64> hm
<tomee^> tomboy64: what do you need it for?
<tomboy64> i'm not awake really
<tomboy64> i was just wondering why the existing bcm drivers didn't support it
<tomboy64> but if it's sdio it makes perfect sense
<tomee^> yeah
<tomee^> it's like talking to your DMA controller over a telephone line ;)
<tomboy64> sound like iscsi
<tomboy64> ^^
maz_ has joined #linux-sunxi
<tomee^> hehe ;)
<tomee^> sas in the first place
<tomee^> how did they manage to pack 68 320MHz wires into than teeny wire...
<wens> tomee^: you mean scsi U-320?
<tomee^> yeah
<tomee^> remember how those LVDS tapes looked like? and what they costed?
JohnDoe_71Rus has joined #linux-sunxi
<wens> i hate those connectors, hard to attach, easy to break a pin
<tomee^> true
<tomee^> and then, somehow, someone managed to wrap that into SAS which is what, 4 pins? 6?
<mnemoc> arokux2: pong
<arokux2> mnemoc: in the leaked source drops the version of bcmdhd is 1.28.23.1, Benn has however 1.28.23.3. I cannot find it on the internet. where did he get it?
<arokux2> mnemoc: in android you can find 1.28.23 and then 1.28.24, nothing in between
<arokux2> wens: can you please explain me what is the ap6210 and what connection does it have to broadcom chips?
<steev> arokux2: it's a mk802 iv device iirc
<steev> the wifi chip in some of them, maybe
<steev> i know it's in mine
<arokux2> iv?
<steev> i suppose it's meant to be written IV
<steev> like 4
<steev> Broadcom AP6210 wireless chip.
<arokux2> steev: ah.. ap6210 is also a ct device
<steev> what is ct?
<arokux2> steev: cubietruck
<steev> ugh
<[7]> does someone know if the livesuit / FEL protocol has been reverse engineered already?
<arokux2> steev: and then this device consists out of two wifi+bt?
<mnemoc> arokux2: no idea. but if you have such drivers, please only send the cover letter to the ML, and I make a pull
<steev> do people NOT do web searches before naming crap they are going to release?
<steev> or is it meant to be confusing on purpose?
<arokux2> mnemoc: the situation is very difficult.....
<arokux2> mnemoc: in our sunxi-3.4 there is a *much* newer version of bcmdhd
<mnemoc> :)
<arokux2> mnemoc: Benn has just overwritten it with the older version...
<mnemoc> doh
<arokux2> mnemoc: we are not going to cherry pick that? :(
<mnemoc> i don't think we should...
<[7]> is it only me, or does that FEL protocol somehow resemble USB mass storage's SCSI transport?
<arokux2> mnemoc: so you think we should go over 30k LOC and search for allwinner changes?
<mnemoc> arokux2: look at the lichee ... -dev branches in my github
<mnemoc> they might have made the changes in different commits
<arokux2> mnemoc: did already. as said this tree contains 1.28.23.1
<arokux2> Benn's code however 1.28.23.3
<arokux2> I suspect Benn has access to some other code?!!
<mnemoc> arokux2: in that world they just need to mail/call the right people to get the .rar of the latest driver...
<mnemoc> ^Med obviusly
<arokux2> mnemoc: do you see CONFIG_ARCH_SUN6I_BCMDHD?
<arokux2> mnemoc: does that indicate the driver was taken from a31 branch?
<mnemoc> probably
Black_Horseman has joined #linux-sunxi
<mnemoc> btw, github supports ?w=1 to ignore whitespace changes in the diff
<mnemoc> but "fixing" a driver by dropping in an older one isn't nice
<arokux2> mnemoc: well.. if you take a look at our a31 code you'll see it is much older than Benn's code
<steev> tomee^: mnemoc: okay, so - reverting 93e38ba7b610e585878aecd0680409319a27c2f9 - does work here. maybe it's because i'm using 4.8.2 for my compiler, i'm not sure, but my test case does not fail here once i've reverted it
<mnemoc> arokux2: our -dev source drop is from feb. 2013
<tomee^> I see.
<tomee^> but 4.8 is bad for the kernel anyway, many people claim so.
<steev> 4.8 pre 4.8.2 yes
<arokux2> mnemoc: which branch is it exactly?
<arokux2> amery/lichee-3.3/sun6i-dev ?
<mnemoc> all the lichee-*/*-dev
<steev> and while some 4.8 may be bad from the kernel, that doesn't mean sticking with a compiler that was released in 2009 :P
<mnemoc> were part of the same leak
<steev> it's sad that open source software can be considered leaked
<steev> makes it sound so ominous
* mnemoc hates when replacing the compiler fixes issues :<
<steev> mnemoc: i didn't replace the compiler!
<mnemoc> steev: I prefer the term "liberated" :)
<steev> i replaced your commit to make gcc 4.4 still work!
<mnemoc> :D
<arokux2> #defineEPI_VERSION_STR"5.90.195.89"
<steev> seriously, come into the modern times bro
<steev> it's nice here
<arokux2> mnemoc: it is still an ancient version on bcmdhd
<arokux2> of*
<mnemoc> feb 2013 is ancient times already
<mnemoc> a20 was not even launched yet
<arokux2> mnemoc: so Benn has access to something new?
<mnemoc> no doubt
<arokux2> mnemoc: ... why is he silent???
<mnemoc> urgent things vs. important things? ... and it seems we scared him
<mnemoc> like with olimex devs...
<arokux2> mnemoc: maybe we should remove there hardware from "featured"?!
<arokux2> mnemoc: scared? with what?
<mnemoc> culture shock
<arokux2> mnemoc: he continues allwinner tradition in his kernel
Taduro has quit [Read error: Connection reset by peer]
<mnemoc> arokux2: cubietech does keep a lot of commits public and helps to get docu and drivers. same as olimex ... even if their hw is not oshw
<mnemoc> they still deserve an special place imo
<arokux2> so either he gives us the code he has based his bcmdhd driver on or I'll send him his ct back and remove it from featured hw
<arokux2> mnemoc: I'm sorry, but when did you last time look at their tree?
<mnemoc> even olimex people, been europeans, neglect to submit their changes. they keep, just like cubietech, their own separated repo and forum
<mnemoc> arokux2: i've never had the time to look into either trees
<arokux2> take a look at its ugliness
<mnemoc> our ML + linux-arm is more than enough for me
<arokux2> mnemoc: olimex ppl never had such kernel as cubietech.. or at least i'm not aware of it.
<arokux2> /AW judge sdio read write timeout, 1s
<arokux2> this is allwinner shit and it is _NOT_ in our lichee
<arokux2> so Benn _has_ new allwinner kernel and just overwrites with it parts of our tree
<arokux2> this is _not_ acceptable in my opinion.
<mnemoc> "ring ring, hi! hi!. hey bro, I have an issue with sdio... do you have a fix for this and that? sure, sending you the patch over email. thanks!"
<arokux2> mnemoc: I doubt it was a patch though.
<arokux2> mnemoc: at least drivers/net/wireless/bcmdhd/*
<wens> arokux2: have no idea. seems like licensing broadcom IP and producing it themselves
<mnemoc> .rar :p
<mnemoc> rm -rf + extract. joy!
<mnemoc> commit, push. "product orient development"
Taduro_ has joined #linux-sunxi
<arokux2> mnemoc: nice. i have nothing against it. but *why* do we have it featured?
<arokux2> what is so cool about it?!
<mnemoc> olimex and cubietech are more open and more helpful than any other
<mnemoc> that doesn't mean that either are perfect. far from it
<mnemoc> but better than the rest
maz_ has quit [Read error: Operation timed out]
<arokux2> mnemoc: as said, when it was the last time you've seen their trees?
<arokux2> mnemoc: olimex does not have their own tree
<mnemoc> they do
<arokux2> mnemoc: show me it please
<steev> the only thing i really like about the cubietruck is that the case will properly hold an ssd/hdd
<[7]> can sunxi uboot access nand these days? (on A20)
<[7]> or do I still have to use lichee?
<libv> arokux2: there's several things that we wouldn't have if it wasn't for the cubie guys, mali binaries are, for me, the best known example
<libv> arokux2: also, i think you can just mail benn and complain to him directly
<arokux2> libv: aren't they passed to you by olimex too?
<arokux2> libv: did that already.
<tomee^> glmark2 Score: 205
<libv> arokux2: olimex is never going to do that with those binaries
<arokux2> libv: why?
<libv> when did you mail benn?
<arokux2> yesterday
<wens> arokux2: it's the weekend :)
<arokux2> yes, time to hack on sunxi :)
<libv> arokux2: of all folks, you should give the cubie guys the benefit of the doubt :)
<arokux2> but how can I cherry pick the code for wifi/bt... maybe you have the idea? do you think we should overwrite the code by the older one?
<arokux2> wens: ^
<mnemoc> arokux2: https://github.com/hehopmajieh?tab=repositories keeps the trees used by olimex
<arokux2> mnemoc: i do not see them abusing our tree.
<libv> fsf thinking strikes again.
<libv> if it were up to the fsf, we would still have pvr as a high priority project, and no single ARM GPU would be actively REed today.
<arokux2> libv: btw, I fully support you and I'm just astonished by the replies on ML
TomiK has joined #linux-sunxi
<arokux2> "we should ask them to comply and *wait*"!
<arokux2> wtf??? maybe they should use the logo once they comply?!
<libv> arokux2: check the latest email in that thread.
<arokux2> libv: i'm sorry, but only people actively participating at sunxi.org can change that wiki. the others can create their own.
<arokux2> libv: " may or may not be an honest mistake" how naive could that be?
<arokux2> libv: most probably iteaduino just decided to put all the logos they could to get all the attention they could achieve
<libv> arokux2: let's give mr wood a chance to fix things as he sees fit
<mnemoc> arokux2: the term abuse is subjective. they do fix stuff and keep it in their tree for us review and fetch... just like cubietech people.
<mnemoc> but both companies help us
<arokux2> mnemoc: "This repository is temporarily unavailable." so I cannot check. but i've seen patches submitted by that guy, and nothing by Benn
<mnemoc> all linux blobs came from cubietech. and a lot of free boards. olimex helps getting docu and boards too. so from my perspective both companies deserve to be in our hall of fame
<arokux2> mnemoc: i'm just angry because as it looks like there is no chance to have support for wifi with our tree.....
<mnemoc> arokux2: mail benn and ask him to submit the needed *fixes*... not bulk import.
<mnemoc> but if our community agrees in downgrading the broadcom driver, so be it
<arokux2> mnemoc: i've mailed him and asked what code base did he use.
<mnemoc> personally I don't like the idea. but we are a community, not a dictatorship
<mnemoc> arokux2: ask him to submit the changes needed to get the CT running on our sunxi-
<mnemoc> 3.4
<arokux2> mnemoc: :)
<arokux2> mnemoc: is that sarcasm?
<arokux2> mnemoc: he'll never do that.
<mnemoc> he needs us
<libv> just as we need him
<mnemoc> AW won't do any newer developmnet for the A20... and I assume that someone selling A20-based devboards expects to be able to provide linux upgrades
<libv> arokux2: you have a cubietruck shaped credit-token on your desk, give benn the benefit of the doubt and assume that it's about lack of time or an error
<mnemoc> it's not as with tablet, sell and forget.
<tomee^> arokux2: what is the policy of linux-sunxi in general? have things working (Ubuntu-like) or use only GPL-compliant stuff (Debian-like, involving running around with pendrives because they won't include a closed-source firmware for a common ethernet card in the installer)
<mnemoc> I do believe Benn *wants* to help us. but he is just too busy, confused, lost, culture shocked, etc...
<mnemoc> so he needs to be guided... softly
<arokux2> tomee^: sunxi.org is about the kernel, not distros in the first place.
<mnemoc> but we do deal with blobs and the RE projects of those blobs too
<tomee^> arokux2: I was asking about your policy, mentioning these 2 distros was only a comparison
<tomee^> mnemoc: exactly.
<arokux2> tomee^: I cannot answer that question accurately, my personal goal - mainline.
<tomee^> mnemoc: that's why I am asking.
<mnemoc> i believe most of the community is on the open source pragmatism than in the fundamentalist religion of openness
<arokux2> mnemoc: and we != lima
<tomee^> arokux2: oh, so that means more like a Debian approach
<libv> tomee^: i am, somewhat, the maintainer of the sunxi-mali repo (although i am a bit lazy)
<tomee^> arokux2: therefore simply transplanting the bcmdhd source from cubietech is out of the question?
<mnemoc> arokux2: but in the case of cedarx, we are their home
<libv> tomee^: i have also packaged some things for ubuntu
<mnemoc> arokux2: lima has other homes too, even if this one is nicer ;-)
<libv> tomee^: but i will not be packaging the mali binaries for instance, people do need to experience some pain when installing binaries
<libv> tomee^: i think it just depends on the different bits and who's handling them
<tomee^> libv: I see. frankly, this is one place when I sort of hit a wall ;)
<arokux2> tomee^: ok, so explaining again. our sunxi-3.4 contains v1.88.45 of bcmdhd, cubie/sunxi-3.4 contains v1.28.23
<libv> tomee^: what do you mean?
<mnemoc> libv: do you think viable to make lima not completely living within mesa to allow android glue?
<tomee^> libv: your git repos and documentation are pretty awesome (although sometimes outdates but that's normal when you develop rapidly)
<arokux2> tomee^: so pulling theirs to ours mean overwriting new code with old one.
<tomee^> libv: but installing the vanilla Mali kernel driver... had no idea how to do that.
<libv> mnemoc: lima will never be a part of mesa
<mnemoc> libv: it would be pretty exciting to be able to have a fully blob-less android 4.4 :)
<mnemoc> gallium3d?
<tomee^> libv: I mean I wanted the r3p2.
<libv> mnemoc: i hate the big monolithic stupidity of mesa
<mnemoc> good
<tomee^> arokux2: yeah... I see.
<libv> mnemoc: mesa can be built for android though
<mnemoc> lima -> mesa -> surfacewhatever ?
<tomee^> arokux2: but in the meantime, wouldn't it be better to have an ugly, but working driver in the tree (or a branch for that matter) ?
<libv> mnemoc: and adding an android build system to lima should be easy as well
<libv> mnemoc: yes
<arokux2> tomee^: this is the question to answer.
<arokux2> tomee^: and I do not know the answer to it.
<tomee^> arokux2: from a purely pragmatic point of view. if there was wifi in the kernel, all those more or (usually) less skilled volunteers would be able to package your kernel, not cubietech's or some wire-wrapped duct-taped god-knows-what
<libv> tomee^: why do you so desperately need r3p2 in your kernel?
<arokux2> tomee^: and from other point of view we should show that *this* kind of shit shouldn't be imposed at us by cubietech or whoever else.
<tomee^> libv: because the 2 pixel processors vs 1 seem to make a tremendous difference in performance in some areas
<mnemoc> tomee^: but downgrading from 1.88.45 to 1.28.23 to "get" a working wifi sounds wrong
<arokux2> tomee^: i'm not going to support their board in *this* way.
<libv> tomee^: what games are you playing that you notice this?
<arokux2> tomee^: i'd better slowly work on mainlining of some drivers for the SoC and let the users decide with money should the Cubietech live or die.
<mnemoc> > Hi, is it possible to rename gmane.comp.hardware.netbook.arm.sunxi
<mnemoc> > into gmane.linux.ports.arm.sunxi ?
<mnemoc> Sorry; group renaming isn't possible.
<arokux2> :)
<mnemoc> .... snif snif snif
<tomee^> mnemoc: you are totally right. it IS wrong. but which is the bigger evil? having a piece of crap in the basket for a while, but the basked be used by 100000 users, or have a shiny basked and not be able to sell it to anyone?
<mnemoc> for me such change cross the line
<mnemoc> fixed have to be fixed by fixing them. not by dropping over something else
<arokux2> tomee^: how long is "for a while" and who is going to fix that?
<tomee^> libv: none. but compare these two: http://heroin.pl/upl/cubietruck/scores/
<tomee^> libv: 2d effects and desktop shading benchmarks are 2 or 3x faster
<tomee^> arokux2: ha, that is a good point ;(
<libv> wow
<libv> they slowed down on the high poly bump...
<tomee^> libv: by 10%, yes.
<AreaScout> i would like to have a limadriver only update, this would be enough for me :)
<tomee^> libv: but I guess what I assembled (r3p2 borrowed from ssvb + the xf86-fbturbo patch), plus the debugging and profiling enabled is far from optimal
<tomee^> arokux2: and about cubietech live or die... is there another board that has all the bells and whistles of a cubietruck in a similar price range?
<wens> arokux2: no idea :(
<arokux2> tomee^: I do not know, but why should I care? (maybe olimex)
<tomee^> I'm confused now ;)
<tomee^> I thought linux-sunxi is more or less about providing support to real-world hardware
<libv> tomee^: but we are humans with limited time and a lot to do
<tomee^> libv: I totally understand that
<tomee^> I wasn't even complaining or even suggesting
<tomee^> I was just asking, trying to understand ;)
<arokux2> tomee^: and we do it because of passion and fun, and it is not fun to overwrite new code by old one.. something like this ppl do only for money
<tomee^> first big surprise for me was the level of support for the board I bought from the vendor = ZERO
<libv> tomee^: also, different people have different priorities
<tomee^> so now I am wondering, maybe there's some secret agenda to linux-sunxi that I am unaware of ;)
<arokux2> tomee^: and paid ppl are sitting in China, Cubietech and they did it for you already
<tomee^> arokux2: their code is useless.
<tomee^> arokux2: as a whole
iamfrankenstein has joined #linux-sunxi
<arokux2> tomee^: my personal agenda is to learn embedded, hardware and kernel development.
<tomee^> ok
<libv> tomee^: 1.5ys ago, you wouldn't even have dreamt of having mali binaries with X11 running
<tomee^> so, the god forsaken wifi in cubietruck... overwriting code would be wrong
<tomee^> how about something like a DKMS drop-in ?
<tomee^> "have cubietruck? ok, so the upstream, proper code won't work because the board is broken. but if you want to, you can download bcmdhd_shit instead"
<arokux2> tomee^: why do you ask us to overwrite? just used their kernel and you are done.
<mnemoc> tomee^: sunxi.org is vendor neutral and ads free. no hidden agendas or business behind. just about improving the free software support of these awesome chips and help people liberating their devices
<arokux2> use*
<arokux2> we are not upstream to cubietechs kernel, I cannot say we are upstream...
<mnemoc> dropping over an old kernel driver in bulk without a clear idea of what is been fixed is not covered by my defition of "improving software"
<tomee^> arokux2: but I don't want their kernel.
<mnemoc> tomee^: then help us fixing our driver to get the wifi+bt of the CT working
<arokux2> tomee^: and what does it have to do with me? :)
<arokux2> or sunxi.org
<arokux2> go complain to them.
<mnemoc> you are very welcomed to find what's wrong with our driver and fix it.
<arokux2> or yes.. go through 30k LOC and help... for free.. doing the work cubietech should have done.
<mnemoc> it's not about cubietech, it's about people using that board
<mnemoc> go throught 30k LOC and help for free... etc etc is basically what we all are doing here regarding Allwinner itself
<tomee^> mnemoc: I might eventually take a look at the wifi issue if noone else does... but I disagree with an attitude of "cubietech should have done that, and since they did not, I am not laying a finger on it, even though I could make it work"
<mnemoc> tomee^: arokux2 (as everyone else here) is free to work on whatever they want. he doesn't want to work on that, it's fine. it's his own free time.
<arokux2> tomee^: there is a long list of work to do, wifi isn't the only issue.
<mnemoc> and I also agree allwinner and cubietech and olimex should be contributing directly
<tomee^> mnemoc: that sounds a lot like the XBMC's team approach. "this patch would allow to use our software in 100000 of devices? so what. it looks ugly and is not based on abstracion layers. go fork yourselves"
<arokux2> tomee^: *this* is real work http://linux-sunxi.org/Linux_mainlining_effort
<mnemoc> tomee^: it's not a patch, it's a downgrade bomb
<arokux2> tomee^: not supporting *some* board.
<mnemoc> tomee^: we do import things. but in this case you want a blind downgrade from 1.88 to 1.23
<tomee^> mnemoc: I don't.
<mnemoc> new drivers we do import as-is. and then try to clean and improve
<mnemoc> but in this particular case we do have a driver
<mnemoc> which needs to be improved to work on that board
<mnemoc> not erased and replaced with another source bomb
<tomee^> mnemoc: I was just wondering, IF the effort to make it RIGHT would be like - 30 man-work-days, and to make it WORK SOMEHOW - 1 man-day...
<tomee^> mnemoc: ok
<tomee^> mnemoc: as long as the changes aren't too big
<mnemoc> to be able to estimate effort, you first need to dive into
<kriegerod> tomee^: some 1-man-day solutions can be still unacceptable for the whole process of development
<libv> tomee^: if i understand you correctly, you are just whining because _you_ cannot get _everything_ to work for _you_ _now_ .
pirea has joined #linux-sunxi
Black_Horseman has quit [Quit: Αποχώρησε]
<mnemoc> first dive into the problem and figure out the real problem.
<tomee^> mnemoc: for example, sometimes you can hack around because, I don't know, say - a class property is public while it should be private.
<mnemoc> first dive into the problem and figure out the real problem :)
<tomee^> mnemoc: then you base your hackery on that (let's say that's what cubietech did)
<tomee^> mnemoc: and in the newer release, the code is cleaned up and there is no way to backport bugs into it ;)
<mnemoc> cubietech probably didn't hack anything. only imported a "known to work" .rar on top of our driver
<mnemoc> that works on their env. but not on ours.
<pirea> arokux / arokux2 why is outdated that package?
<tomee^> libv: whining? no, I am showing great respect for the whole linux-sunxi community. I am really in awe.
<wens> hack-ish stuff won't get mainlined though
<tomee^> libv: you do stuff that the vendors should have done
<libv> tomee^: then why keep on hammering on that one point?
<arokux2> pirea: what package?
<mnemoc> this wifi driver won't be mainlined at all
<tomee^> libv: I am sorry and I apologize if it sounded like I am whining and complaining on you
<mnemoc> the linux-sunxi thing is between two goals. mainlining allwinner IP support, and giving people a kernel they can use on their devices today
<arokux2> pirea: why do you think I thing it is outdated?
<tomee^> mnemoc: I see.
<pirea> somebody has marked my package as outdated :|
<pirea> fuck :|
<pirea> and i don't know why
<pirea> because have no bug...
<mnemoc> this wifi thing is not for us to mainline. but "it would be nice" if we fix that not-mainlinable driver to improve the experience of normal people using the full featured kernel
<pirea> i use that pkgbuild for my kernel and everything is fine
<mnemoc> but a bulk and blind downgrade doesn't qualify as "fix"
<libv> tomee^: each individual contributor here is pulling into the directions that interests this contributor
<libv> tomee^: if you see room for improvement, then it is up to you to improve it
<wens> mnemoc: reminds me of my current problem... GbE works on cubietech kernel, not gmac on mainline, nor stmmac
<wens> will try 3.4 kernel tomorrow
<tomee^> err?
<tomee^> gmac works in mainline
<mnemoc> wens: gmac is on stage/sunxi-3.4 until someone adds a decent runtime hook to decide which mii to use
<arokux2> wens: GbE as opposed to 100MbE?
<mnemoc> mainline might also have a mii issue
<wens> tomee^: torbenh3 has a working tree.
<wens> mnemoc: working on that :)
<mnemoc> depending on the board
<mnemoc> wens: :)
<wens> arokux2: 100 MbE works, GbE doesn't. nothing gets transmitted :(
<wens> either the transmit clock, or DMA setting is wrong
<arokux2> I see...
<wens> mnemoc: torbenh3 cleaned up gmac code enough to use DT for clocks and phy mode i think, though only ct support
<wens> mnemoc: i have stmmac working *wink*, but need to clean up the code to be merged
<wens> i recall seeing some option in the driver that applies to GbE only, so that i will try
<tomee^> hmm
<tomee^> so the stage/sunxi-3.4 gmac driver work but only in 100mbit mode?
<tomee^> and breaks when trying to establish a 1000 link?
<tomee^> or did I understand wrong?
<arokux2> the mainline one, wens haven't tried stage/sunxi-3.4 but it should work i think.
rzk has joined #linux-sunxi
rz2k has quit [Ping timeout: 240 seconds]
jinzo has joined #linux-sunxi
bennhuang has joined #linux-sunxi
<bennhuang> itead even use cubieboard's firmware http://itead.cc/down/html/?22.html :(
<bennhuang> ping arokux
<arokux2> bennhuang: hey
<arokux2> bennhuang: i've just read your e-mail
<Turl> bennhuang: :| they didn't even bother to make own firmware?
<arokux2> libv: ^
ZetaNeta has quit [Quit: Leaving]
<libv> i'm sorry. patrick wood is dealing with itead from now on.
<libv> ;)
<bennhuang> arokux2: yes. You can find out on the link
<arokux2> bennhuang: I'm downloading the sdk right now.
<libv> i believe that they are allowed to copy that image, right?
<libv> that's just plain laziness, nothing we can really fault in that
<bennhuang> libv: yes. I just want to say. in china, they will copy every thing
<libv> bennhuang: stop being such a prejudist against the chinese :p
<bennhuang> libv: even let their user to get support from cubietech.
wingrime has quit [Ping timeout: 246 seconds]
<Turl> bennhuang: how seriously does your legal system take mark infringement (fake CE, FCC logo for example)?
<arokux2> bennhuang: is China your main market?
<bennhuang> arokux2: no. we even not willing to sale the board in china :D
<arokux2> bennhuang: so all the others have been warned: http://linux-sunxi.org/Iteaduino
<bennhuang> arokux2: :P. too much such competitors, and I don't even want to say more about that.
<arokux2> I can only say that we are aware of this and inform people that come along.
<arokux2> bennhuang: but I also would like you to stop developing the fork of sunxi-3.4 if you want sunxi.org to support your kenel/board.
<arokux2> bennhuang: at least tell us what you are fixing and what the problems are.
<bennhuang> arokux2: cubietech will keep dev board available as much long as we can. but obviously, cubietech want to survive, we need to do some other work :D
<arokux2> bennhuang: and so you do not even have time to communicate what your fixes are for?
<ganbold_> bennhuang: does Cubietech have plan to develop a board with 2 gigabit ethernet port in near future?
<tomee^> one question if I may ask. why is the r3p2-01rel1 X11 libMali.so blob only 1/2 the size of the others?
<tomee^> and lacks e.g. the symbol _mali_clz_lut ?
<arokux2> bennhuang: do you think I have time to go to your tree and spend days on it just to understand that you have used some code which is not even freely available?
<libv> tomee^: is this a unified blob or not?
<bennhuang> arokux2: I know we need to follow the community kernel
<kriegerod> tomee^: did you check libEGL and libGLES* for this symbol?
* [7] decides to go ahead and port allwinner NAND support to latest sunxi uboot
<tomee^> libv: the x11 one
<bennhuang> arokux2: Make the wifi modules work is easier than make it get merged IMO
<arokux2> bennhuang: I do not expect you to do everything.. we are there to help. but currently there is no communication at all. this should improve.
<bennhuang> arokux2: first work, then get merged is good way
<arokux2> bennhuang: I agree with you. but talk to us too.
<kriegerod> +1 to arokux
<tomee^> libv: root@cubietruck:/sata/build/2/sunxi-mali# du -h ./lib/mali/r3p0/armhf/x11/libMali.so; strings ./lib/mali/r3p0/armhf/x11/libMali.so | grep mali_clz
<tomee^> 920K ./lib/mali/r3p0/armhf/x11/libMali.so
<tomee^> _mali_clz_lut
<tomee^> root@cubietruck:/sata/build/2/sunxi-mali# du -h ./lib/mali/r3p2-01rel1/armhf/x11/libMali.so; strings ./lib/mali/r3p2-01rel1/armhf/x11/libMali.so | grep mali_clz || echo "not there"
<tomee^> 560K ./lib/mali/r3p2-01rel1/armhf/x11/libMali.so
<tomee^> not there
<kriegerod> tomee^: grep also libEGL and libGLES*
<libv> tomee^: the other libMali.so is an all encompassing blob
<arokux2> bennhuang: I understand you are fighting to survive. but if you want sunxi.org or me personally to hep with maintaining of sunxi-3.4 you should help *somehow*.
<libv> tomee^: the r3p2-01rel1 one is one with all the different bits split out
<arokux2> bennhuang: by *somehow* I do not mean submitting perfect code.
<libv> tomee^: hence "unified blob"
<Turl> arokux2: bennhuang a wiki page should suffice, don't you think?
<arokux2> Turl: a wiki page about what?
<Turl> arokux2: bennhuang make a wiki page and when bennhuang adds a patch to the tree, he should write a small comment about why it is there, or what plans there are for it
<Turl> or something like that
<tomee^> kriegerod: it is in libEGL.so
<tomee^> libv: hm, ok, I see
<arokux2> bennhuang: yes, thi sis the minimal help I expect from you ----^
<arokux2> good idea Turl
<bennhuang> arokux2: sound good
<bennhuang> arokux2: or even mail me with TODO list :D
<arokux2> bennhuang: mails are not good - wiki is better where everyone sees it and can help
<kriegerod> Turl: i think meaningful commit message is easier and better
<tomee^> oh damn it
<tomee^> now it works
eebrah has quit [Ping timeout: 240 seconds]
<kriegerod> arokux2: what about public mailing list dedicated to these issues?
<tomee^> must have been the ccache or ldconfig messing up
<tomee^> kriegerod: thank you.
<tomee^> libv: thank you.
<Turl> kriegerod: that could work too :)
<arokux2> info tend to get lost at ML, a wiki is better imho, but it doen't matter..
eebrah has joined #linux-sunxi
<kriegerod> i just think the collaboration will more naturely fit in conversation manner rather than book
<tomee^> ... or, not...
<kriegerod> or article
<arokux2> bennhuang: so go ahead: http://linux-sunxi.org/CubietechKernel :)
<bennhuang> my biggest advantage is to community with allwinner, and let them to help us :D
<bennhuang> we are very near to allwinner office
<arokux2> bennhuang: is allwinner helping you?
<bennhuang> arokux2: no
<arokux2> I doubt it, since you buy 20-30k chips, whereas they are selling millions
<tomee^> libv: does this mean the ABI is different between those two? like in need-to-recompile-everything different?
<tomee^> libv: I mean everything that links against EGL
<bennhuang> arokux2: they think is good to have a company like cubietech
<libv> tomee^: no, gles/egl should be stable
<tomee^> libv: /usr/local/bin/mplayer: symbol lookup error: /usr/lib/libEGL.so: undefined symbol: _mali_clz_lut
<libv> bennhuang: so you guys can provide images for all those other allwinner customers
<arokux2> :)
<tomee^> libv: root@cubietruck:~# strings /usr/lib/libEGL.so|grep mali_clz -> _mali_clz_lut
<bennhuang> libv: yes. we can get sdk, and get some support from allwinner. they also let us to give suggestion to their next chip
<libv> tomee^: strings is not what you should be using
<libv> bennhuang: i was just joking about the itead situation :)
<libv> tomee^: nm is what you need
<bennhuang> arokux2: dev board can bring more potential customers
<bennhuang> arokux2: for allwinner
<arokux2> bennhuang: I see
<tomee^> libv: root@cubietruck:~# nm /usr/lib/libEGL.so|grep mali_clz
<tomee^> 0000defc t _mali_clz
<tomee^> U _mali_clz_lut
<tomee^> libv: ouch...
<arokux2> bennhuang: so do you agree to communicate what is going on at your tree here? http://linux-sunxi.org/CubietechKernel
<libv> tomee^: there does seem to be some further dependency brokenness there. yes
<bennhuang> arokux2: yes
<arokux2> bennhuang: good!
<rzk> tomee^: ld_preload=libglesv2.so
<rzk> or libmali.so, mplayer just doesnt preload everything needed.
<tomee^> rzk: yeah!
<tomee^> ;]
<bennhuang> arokux2: actually, I will follow the sunxi commit stytle for the code I am familliar with :)
<libv> it's just further dependency brokenness, like with the X11 stuff in r3p0
<tomee^> rzk: this seems to do the trick. but why did the dependencies break, hmm.
<rzk> this problem comes from r2p4 ages and still going
<arokux2> bennhuang: the more you do the better, but I understand you do not have time to do thing properly. so at minimum please comment on the commits at that wiki page.
<bennhuang> arokux2: get it
<libv> rzk: quite amazing how arm has managed to endless keep on breaking those things
eebrah has quit [Read error: Connection reset by peer]
<arokux2> bennhuang: you have also used some very fresh code in bcmdhd, where is it from? I mean this: https://github.com/cubieboard/linux-sunxi/tree/cubie/sunxi-3.4/drivers/net/wireless/bcmdhd/include/proto
<tomee^> libv: ... one of the reasons you're sticking to r3p0 ?
<libv> tomee^: bennhuang got me a cubietruck, but i haven't gotten round to it yet. the devices i have running today are either A10 or exynos based
eebrah has joined #linux-sunxi
<techn__> tomee^: we need android libs for r3p2
<arokux2> bennhuang: oh, I see it is also from sdk.
<arokux2> techn__: what is the state of hackberry usb power problem. is it but in hw or otg driver?
<techn__> has someone asked from hackberry or measured how detect pin is wired?
<arokux2> techn__: yes, there were some users of hackberry here.
<wens> "bcmdhd is what broadcom open-sourced for android. Separately, we started working on brcmfmac for the mainline linux kernel. Both drivers come from the same proprietary driver. The goal is to have brcmfmac work on Android as well and we have most features on board for that. "
<wens> arokux2: found this on the linux-wireless ML
<arokux2> techn__: can you maybe write down what should be measured?
<arokux2> techn__: take a look at the last bullet here: http://linux-sunxi.org/USB#Known_issues
<arokux2> wens: thanks
<wens> basically means you need to dig through all the android repositories :(
<arokux2> wens: my humble goal is to make sunxi-3.4 usable with ct in a reasonable way
<arokux2> wens: then i go back to my almost ready patches for usb host
<arokux2> bennhuang: where is the android sdk coming from? it contains some code which is not available from google
<wens> arokux2: about usb, i know otg is mentor graphics IP, what about the hosts?
<arokux2> google's android only has 1.28.23 and then 1.23.24 versions of bcmdhd. the sdk contains 1.28.23.3
<arokux2> wens: nothing special. some clocks should be turned on and it works with the general code available in linux
eebrah has quit [Read error: No route to host]
<wens> arokux2: ah i see
<wens> arokux2: basic glue code, then
<arokux2> wens: yes. but it touches lots of things......
<arokux2> wens: clk, gpio, resets
<wens> arokux2: ouch
<arokux2> wens: all of that is very easy but it take time to learn how it works and how it should be used.
<wens> arokux2: tell me about it. i'm trying to figure out the proper way to register platform driver
<arokux2> wens: the proper was is dt
bennhuang has quit [Ping timeout: 245 seconds]
<wens> arokux2: i meant to make the driver/extension known to the kernel, so it will associate dt with driver
<arokux2> wens: there is compatible field in dt and of_table in the driver
<wens> arokux2: that part i get
<arokux2> wens: so what is left?
<wens> arokux2: registering driver/module with the kernel. i've seen 2 or 3 ways already
<wens> arokux2: thanks
<wens> arokux2: cubietech
<arokux2> cubietech?
<wens> arokux2: cubietech's bcmdhd has wierd version macros. increment is different from the string
tomboy64 has quit [Remote host closed the connection]
<wens> maybe it was modifed by some third party
tomboy64 has joined #linux-sunxi
wingrime has joined #linux-sunxi
<arokux2> wens: and this third party is allwinner :)
<arokux2> mnemoc: ping
mickoo has joined #linux-sunxi
<wingrime> wens: are you currenty AW stuff?
\\Mr_C\\ has joined #linux-sunxi
<mickoo> Hi :), i have some troubles with u-boot-sunxi. I created boot.scr but i think that's it's ignored (i moved uImage and script.bin inside a /boot/ directory inside the first partition and changed boot.scr to load uImage and script.bin from there but it boots only if i place them in the root dir)
<mickoo> i can't access to serial port, my board is an hackberry
<mickoo> is there a way to write uboot logs on the SD?
tomboy64 has quit [Quit: Uhh ... gotta go.]
<mnemoc> arokux2:
<arokux2> mnemoc: can you please import lichee-3.2.tar.gz ?
<arokux2> mickoo: have you solved your USB problem?
Gerwin_J has joined #linux-sunxi
<mickoo> arokux2:hi :)
ZetaNeta has joined #linux-sunxi
<mickoo> arokux2: i got some problems to make hackberry to load new uImage
<mickoo> :(
<arokux2> mickoo: what about USB powering problem you had yesterday?
<mnemoc> arokux2: url?
<arokux2> mnemoc: also.. do you know how the stuff is versioned?
<arokux2> android-3.2.tar.gz vs anroid-4.2? :(
<mickoo> arokux: I have not tryied it yet, i'm trying to understand why my uImage isn't loaded correctly
<arokux2> ok
<mnemoc> arokux2: no idea what you are talking about. there is nothing like android 3.2. android 4.2 existed.
<arokux2> mnemoc: how can I identify lichee kernel?
<arokux2> the 3.2 i'm talking about was given by drachensun once
<mnemoc> look into the Makefile
<arokux2> mnemoc: the kernel version is 3.3.0
<mnemoc> then it's old
<arokux2> mnemoc: and it is the same cubietech is having in theirs A20-android-4.2.tar.xz
<arokux2> mnemoc: but we do not have it in imports
<mnemoc> the last SDKs tsvetan shared are based upon linux 3.4
<mnemoc> arokux2: imports is not merely about dumping in. it's also identifying the right base
<mnemoc> we do have lichee-3.3 stuff
<mnemoc> specially the -dev branches
<arokux2> mnemoc: this is older: import/lichee-3.3/a31-dev
<arokux2> mnemoc: although also linux-3.3.0
<mnemoc> lichee-3.3/* are pristine imports of .git from allwinner
<mnemoc> import/lichee-3.3/* are sanitized and rebased upon a refenrece branch
<mnemoc> but those source dumps are usually destructive, so importing upon the head of import/lichee-3.3/a31-dev requires reviewing. to be sure to not delete stuff shouldn't be deleted
<arokux2> mnemoc: I see, but all ours 3.3 doesn't have a newer bcmdhd 1.28.23.3
<arokux2> they yave 1.28.23.1
<arokux2> have*
<arokux2> and tar xf A20-android-4.2.tar.xz ct-droid1/lichee/linux-3.3/
<mnemoc> what's the point? I don't get it
<mnemoc> why do we want yet another 3.3 dump?
<arokux2> mnemoc: not sure :) it has the code that is not in the available dumps
<mnemoc> weirdly that 1.2 is newer than the other 2.0
<arokux2> mnemoc: yeah, the last one doesn't have bcmdhd 1.28.23.3
<mnemoc> and?
<mnemoc> if you want to import a driver from a dump, just import THAT driver
<arokux2> mnemoc: ok, I was unsure....
eebrah has joined #linux-sunxi
<wingrime> lkcl: how much layers used in tablet PCB
<mickoo> arokux: thanks, enabling sunxi-musb usb device are powered
rzk has quit []
<mickoo> arokux: now do you know if there is a way to log uboot on the sd card?
JohnDoe_71Rus has quit [Quit: KVIrc 4.0.4 Insomnia http://www.kvirc.net/]
<arokux2> mickoo: log u-boot?
ZetaNeta has quit [Ping timeout: 272 seconds]
<mickoo> arokux2: yes.. i don't have access to hackberry serial port so i was asking if it's possible somehow to write uboot output to a file on the SD
<arokux2> mickoo: I see. sorry, I do not know if this is possible.
<mnemoc> arokux2: importing 3rd drivers is much easier because we don't need to find the changes, they are standalone things
<mnemoc> and in this case not even sunxi related
<arokux2> mnemoc: ok
eebrah has quit [Read error: Connection reset by peer]
<mnemoc> arokux2: I am updating that tar.xz into dl.linux-sunxi.org/SDK ... but unsure about it's value
<mnemoc> s/updating/downloading/
<mnemoc> eta 5h
<arokux2> mnemoc: ok, for the one it contains the bcmdhd of 1.28.23.3 not found anywhere else
<mnemoc> Content-Length: 3393723104
eebrah has joined #linux-sunxi
wingrime has quit [Ping timeout: 245 seconds]
hglm has joined #linux-sunxi
<hglm> arokux: I noticed you changed my USB power problem reference in linux-sunxi.org/USB to refer to Hackberry only, but it happens on my A20 tablet as well and I think it happens on most tablets even with A10...
tomboy64 has joined #linux-sunxi
<mickoo> sorry guys, i installed mali binary drivers but when i start xorg i get this
<mickoo> (EE) AIGLX: reverting to software rendering
<mickoo> (EE) AIGLX error: dlopen of /usr/lib/arm-linux-gnueabihf/dri/lima_dri.so failed (/usr/lib/arm-linux-gnueabihf/dri/lima_dri.so: cannot open shared object file: No such file or directory)
<Wizzup> I see mali and lima, you're doing something odd
<mickoo> i think so :)
<mickoo> Wizzup:let me check if i have something compiled against lima
<mickoo> if i paste xorg.log could you give me some hints? :)
<Wizzup> mickoo: I couldn't, I am still waiting on my cubieboard.
<Wizzup> But someone else might.
<hglm> mickoo: Did you compile sunxi-mali after first installing the seperate libdri2?
\\Mr_C\\ has quit [Quit: .]
<tomboy64> yay
eebrah has quit [Ping timeout: 248 seconds]
<tomboy64> first attempt to flash - and i bricked it
<tomboy64> <- hero of the day
eebrah has joined #linux-sunxi
<hglm> mickoo: I also get the AIGLX error in my X server logs, but Mali GLES2 works. AIGLX refers to regular OpenGL drivers which are not supported with hardware acceleration.
eebrah has quit [Read error: Connection reset by peer]
\\Mr_C\\ has joined #linux-sunxi
ZetaNeta has joined #linux-sunxi
<buZz> yeah
<buZz> opengl is -not- supported
eebrah has joined #linux-sunxi
popolon has joined #linux-sunxi
<mickoo> hglm: Yes, sunxi mali compiled after libdri2
<hglm> bUZz: Even though OpenGL ES2.0 is pretty close API-wise to modern use somewhere between OpenGL 2.0 and 3.0 -- so a GLES2 back-end is often possible given some effort.
<lkcl> wingrime: only 4.
<buZz> hglm: sure, its just that nearly nobody is putting in that effort
<mickoo> hglm: is that wrong?
<buZz> enlightenment has good opengles support though
<lkcl> 1.2mm thickness. could be less.
<hglm> mickoo: the message doesn't mean anything, Mali GLES2 should work...try test/test.
<lkcl> wingrime: 6 wasn't necessary. 2 would be virtually impossible given the single-sided design and the very small area available for components
<mickoo> hglm: okk thanks.. so if sunxi-mali "test" triangle is shown then it should be allright..
<hglm> mickoo: Yes, there are other GLES2 apps but not many. glmark2-es2 is nice but can be hard to compile.
woprr has quit [Remote host closed the connection]
<mickoo> hglm: ok.. i'll try to compile it.. where can i find source?
<hglm> buZZ: I have written a 3D engine (developed on PC OpenGL) that works pretty well on GLES2 with the same code base and shaders (and some #ifdefs). And Mali is reasonably fast as long as you don't do per pixel lighting for the whole scene.
<buZz> hglm: cool, now make it as popular as unity!
<buZz> :D
<buZz> my last 3d engine was all software, 80x50 textmode output, and written in qbasic ;)
<hglm> buZz: It isn't public atm, but I though about it. Unity seems to OK judging from some Android games using it with good results.
Fusing2 has joined #linux-sunxi
Fusing has quit [Ping timeout: 248 seconds]
<hglm> buZz: I also started with a 100% software renderer -- then rewrote for OpenGL 3.0.
tzafrir has quit [Ping timeout: 272 seconds]
eebrah has quit [Read error: Connection reset by peer]
Fusing2 has quit [Ping timeout: 240 seconds]
zeRez has joined #linux-sunxi
<buZz> well this was .. 18 years ago?
eebrah has joined #linux-sunxi
<buZz> i'm getting old :P
tomboy64 has quit [Remote host closed the connection]
Gerwin_J has quit [Quit: Gerwin_J]
tomboy64 has joined #linux-sunxi
<techn__> uups.. typo in commit text
<arokux2> techn__: you want me to review?
<techn__> arokux2: that needs testing
<techn__> from hackberry owners
ZetaNeta has quit [Ping timeout: 248 seconds]
tomboy64 has quit [Quit: Uhh ... gotta go.]
<arokux2> techn__: hglm has hackberry.
Gerwin_J has joined #linux-sunxi
<arokux2> hglm: can you please test?
<hglm> arokux2: I don't have a Hackberry...I only have an A20 tablet (and a A10)...the USB power problem seems to occur with many tablets as well.
<techn__> commit affects on these devices auxtek-t003.fex, mini-x_a10s.fex, megafeis_a08.fex, mk802_a10s.fex
<techn__> has anyone got one of these devices+
<techn__> ?
<Turl> techn__: wrap the commit message :p
eebrah_ has joined #linux-sunxi
<Turl> techn__: and (personal opinion here) use some other paste site, ubuntu one doesn't let you download plaintext easily
<hglm> techn__: Is the patch for suni-3.4 or stage/sunxi-3.4 branch?
<techn__> Turl: pastebinit is easy to use
<techn__> hglm: should apply on both?
<techn__> hglm: which device you got?
<hglm> techn: OK, I have a A20 tablet.
<techn__> hglm: fex file somewhere?
<techn__> original
eebrah has quit [Ping timeout: 245 seconds]
<hglm> techn: Yes, I edited already (fixed WiFi and memory clock), what should I look for?
eebrah_ is now known as eebrah
<techn__> hglm: usb_port_type on usb0
<hglm> usb_port_type = 2
<techn__> hglm: then patch should have on effect
<techn__> 0:device only;1:host only;2:OTG
<hglm> OK. usbc1 is type 1 though (usbc2 is WiFi on my device).
Quarx has quit [Read error: Connection reset by peer]
<techn__> hglm: usb works with stick rom?
<hglm> Hmm, haven't tried USB stick yet...only mouse/keyboard with mini-hub (non-powered). Will try now.
<techn__> stock rom I ment :)
eebrah has quit [Read error: Connection reset by peer]
<techn__> hglm: I tried A20 OTG port.. looks like it doesnt work
<techn__> hglm: oh.. It's disabled in default a20 defconfig
eebrah has joined #linux-sunxi
hglm has quit [Read error: Connection reset by peer]
hglm has joined #linux-sunxi
<hglm> I have A20 kernel with Inventra sunxi-musb (OTG) enabled (sunxi-3.4 branch). If driver is not enabled, USB has no power.
<hglm> However when I just inserted a USB stick the kernel/USB driver crashed...USB didn't work anymore.
<mnemoc> libv: do you still consider the Iteaduino page to be protected?
<mnemoc> we need nice banner templates for these things
Taduro_ has quit [Read error: Connection reset by peer]
eebrah has quit [Read error: Connection reset by peer]
tomboy64 has joined #linux-sunxi
Taduro has joined #linux-sunxi
eebrah has joined #linux-sunxi
ZetaNeta has joined #linux-sunxi
eebrah has quit [Ping timeout: 272 seconds]
eebrah has joined #linux-sunxi
eebrah has quit [Ping timeout: 252 seconds]
ZetaNeta has quit [Ping timeout: 272 seconds]
eebrah has joined #linux-sunxi
fredy has quit [Excess Flood]
fredy has joined #linux-sunxi
eebrah has quit [Read error: Connection reset by peer]
hglm has quit [Quit: leaving]
eebrah has joined #linux-sunxi
eebrah has quit [Read error: Connection reset by peer]
eebrah has joined #linux-sunxi
zeRez has quit [Quit: zeRez]
deasy has quit [Quit: Nom d'un quark, c'est Edmonton !]
eebrah has quit [Ping timeout: 240 seconds]
tzafrir has joined #linux-sunxi
eebrah has joined #linux-sunxi
eebrah_ has joined #linux-sunxi
adb has joined #linux-sunxi
eebrah has quit [Read error: Connection reset by peer]
jinzo has quit [Read error: Connection reset by peer]
woprr has joined #linux-sunxi
jinzo has joined #linux-sunxi
cajg has quit [Ping timeout: 272 seconds]
tomboy64 has quit [Quit: Uhh ... gotta go.]
eebrah_ has quit [Read error: Connection reset by peer]
n01_ has quit [Read error: Connection reset by peer]
cajg has joined #linux-sunxi
eebrah has joined #linux-sunxi
n01_ has joined #linux-sunxi
<tomee^> food for thought: I've got a 2A PSU (in between "total crap" and "reliable") and a 7200rpm SATA drive. with r3p2 mali, I got random system shutdowns as soon as graphics was in use.
<tomee^> now I shortened the cord to ~30cm and... no more crashes so far
<tomee^> and given the quality of wires I probably lose 20% on every 50cm of wire
<Turl> tomee^: there was a similar issue with the cables that shipped with the first cubieboards
<tomee^> take a look at those chargers and their cables, and compare that to what you see in your car fueling e.g. the tail lights
<tomee^> Turl: really? ha, didn't know
<tomee^> Turl: anyway this somewhat backs up my theory that you need JUICE to fuel such a board+SATA drive+TOSlink+whatever
<tomee^> Turl: and that some of people's problems might be PSU-related
eebrah has quit [Ping timeout: 245 seconds]
<techn__> tomee^: I had problems with usb until I changed my power cord
<Turl> tomee^: just the disk itself often needs quite a bit of power
<techn__> with cubietruck.. cubie2 worked ok with previous setup
<Turl> and if the PSU's voltage dips too low when the disk starts spinning, you'll get trouble
eebrah has joined #linux-sunxi
<tomee^> Turl: this is exactly how it looked like
<tomee^> Turl: I imagine that a 7200 laptop drive can draw 3 to 5-8W in peak
<tomee^> Turl: which would mean 0,6 to 1A
ganbold has quit [Ping timeout: 245 seconds]
<tomee^> Turl: and I look at the board again... WHERE IS THE CAPACITOR FOR VOLTAGE STABILIZATION? :)
<Turl> tomee^: it's possible yep, certainly it's >0.5A in some cases
<tomee^> techn__: what kind of PSU do you use now?
<Turl> tomee^: on your PSU? :P
eebrah has quit [Read error: Connection reset by peer]
<tomee^> Turl: on my psu my ass.
<techn__> tomee^: some old Nokia charger.. It says 2A
<Turl> tomee^: one capacitor less saves quite some money on the long run :p
<tomee^> Turl: I got one transformer-based stabilized voltage-regulated PSU at home. 5V 1000mA. weighs about half a kilogram.
<Turl> :p
<tomee^> Turl: ... and yields extra money on repairs. yeah ;)
<tomee^> now, those 750-1000mA cell phone chargers...
<Turl> tomee^: I use one of those to power my CB2
<tomee^> how come they are the size of a matchbox (and weigh as much as 20 matches)...? :-)
<Turl> but I run full headless
ganbold has joined #linux-sunxi
<tomee^> Turl: headless = no Mali I guess?
ZetaNeta has joined #linux-sunxi
montjoie[home] has quit [Quit: Quitte]
<tomee^> "in response to various trouble reports, cubietech officially recommends a Fluke PM28xx power suppply for the cubietruck, starting at $900 USD. get a 3% discount with a CUBIETRUCK coupon!"
<Turl> tomee^: no mali, hdmi or sata
<Turl> just the board with serial and ethernet
<tomee^> oh, so you could probably power it through electromagnetic interference ;)
AreaScout has quit []
<Turl> tomee^: :p
<tomee^> (as a matter of fact, I have seen for myself, that with a 100m long 2x2 UTP and 2 DSLs on it, you can sync on the other DSL if the first one is offline)
<tomee^> 100% galvanic insulation. just Tesla's laws in play ;-)
<tomee^> (meaning: you can sync on DSL1 with the modem plugged to DSL2)
tomboy64 has joined #linux-sunxi
<Turl> haha
n01 has quit [Ping timeout: 248 seconds]
n01_ has quit [Ping timeout: 272 seconds]
naobsd has quit [Ping timeout: 250 seconds]
naobsd_ has joined #linux-sunxi
Deskwizard has joined #linux-sunxi
<tomee^> [d]
hglm has joined #linux-sunxi
Black_Horseman has joined #linux-sunxi
<hglm> tomee: Did you encounter any "wavy screen" issues when running Mali in X? Looks like a DRAM contention issue with the disp layer interface. It only shows when the HDMI resolution is set to 1920x1080.
tomboy64 has quit [Quit: Uhh ... gotta go.]
<Turl> hglm: screen tearing?
<hglm> Well, not tearing within the Mali window, but just waviness/oscillation over the whole screen because the chip can't read memory fast enough for video output (HDMI).
Deskwizard has quit [Read error: Connection reset by peer]
techn_ has joined #linux-sunxi
techn__ has quit [Ping timeout: 272 seconds]
buZz has quit [Read error: Connection reset by peer]
buZz has joined #linux-sunxi
paulk-collins has quit [Remote host closed the connection]
TehCaptain has quit [Read error: Connection reset by peer]
hglm has quit [Quit: leaving]
techn_ has quit [Ping timeout: 264 seconds]
techn_ has joined #linux-sunxi
<TomiK> hello :)
<TomiK> for those that it could interest
tomboy64 has joined #linux-sunxi
<TomiK> after 3 days of work
<TomiK> I boot my sun3i tablet on a dummy bootloader which send some bytes on the serial port at 577Mhz
<TomiK> \o/
naobsd_ has quit [Quit: Page closed]
FDCX has joined #linux-sunxi
<tomee^> hglm: My TV has a native res of 1280x720 so I can't say about 1080p
<[7]> someone around who knows the A20 NAND controller well?
<[7]> I've merged the lichee A20 nand driver and android/fastboot support into the sunxi branch
<[7]> just like I did for the A10 about a year ago
<[7]> I've managed to make it compile, but it doesn't work yet
<[7]> I'm getting a "nand _wait_cmd_finish time out, status:0xf11"
hglm has joined #linux-sunxi
<[7]> ...whatever that means
tomboy64 has quit [Remote host closed the connection]
<hglm> tomee: Mali works fine me at 1280x720 (also faster even with same window size when compared to driving the monitor at 1080p). If you want to see more effect of the two pixel processors with the latest mali driver it should show up with a higher window resoluton -- for example running glmark2-es2 with the --size argument (default is 800x600).
tomboy64 has joined #linux-sunxi
<Turl> TomiK: \o/ great!
<TomiK> Turl> :)
<TomiK> now the biggest job I think is adapt u-boot to it
<TomiK> sorry for my bad english, I'm french
<TomiK> Turl> thx
<TomiK> :)
Black_Horseman has quit [Quit: Αποχώρησε]
<Turl> TomiK: have you seen this? http://www.sysfwlab.com/?cat=8
Black_Horseman has joined #linux-sunxi
<TomiK> Turl> yes, I have the same tablet and this blog is my main documentation source
<TomiK> :)
<TomiK> the dummy bootloader I use comes from the great job of this guy :)
<Turl> :)
jinzo has quit [Quit: Leaving]
geecko has joined #linux-sunxi
ganbold_ has quit [Remote host closed the connection]
pfdm has joined #linux-sunxi
* [7] hates undocumented hardware
* TomiK too :)
<[7]> that just makes diagnosing and fixing problems extremely time consuming
<TomiK> amen
<TomiK> yes unfortunatly
tomboy64 has quit [Remote host closed the connection]
<[7]> it's even worse if you're messing with code that you haven't written yourself
tomboy64 has joined #linux-sunxi
<TomiK> yep
<[7]> so why does that stupid nand controller never signal completion?
<[7]> I'm not even sure what it's supposed to do
<[7]> the code is probably attempting to read some page, but I don't really know
<TomiK> [7]> I can't help you, I don't know enough to help you, but could you try to write a simple code to test and understand by yourself what it's supposed to do ?
<TomiK> it's only a suggestion, it's what I did on my sun3i
arokux2 has quit [Remote host closed the connection]
<[7]> well, I'm messing with uboot here
<[7]> kinda hard to strip that down to a minimal test case if you don't have a clue how that hardware even works
<[7]> the problems is that the code is the best documentation that's available... which is kinda nasty if the code has bugs
<popolon> Does someone have information about the rumour that speak of a Mali-T6xx or Mali-T7xx AllWinner A80 ?
<Turl> [7]: is the controller correctly clocked? did any other previous operation succeed?
adb has quit [Ping timeout: 252 seconds]
hglm has quit [Quit: leaving]
popolon has quit [Quit: Quitte]