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
<Seppoz> did you try to add /tmp to fstab?
<oliv3r> I know!
<Seppoz> i dont think ext4 is a problem
<oliv3r> i don't think so either; it mounts it (in ro) mode fine
<oliv3r> it shouldn't be a missing fstab either
<oliv3r> and i really doubt debootstrap didn't create /tmp
<oliv3r> oh i'll try removing ro from my uenv
<oliv3r> make it RW
<Seppoz> it sounds like somehtng went strange in your 2nd stage
<oliv3r> same shit
<oliv3r> stops after doing mmc detection
<oliv3r> but <6>VFS: Mounted root (ext4 filesystem) on device 8:3.
techn_ has quit [Read error: Connection reset by peer]
<Seppoz> did you pass the rigth startup script?
<Seppoz> i think in ubuntu its sbin init
<Seppoz> not /init
<oliv3r> i passed no init; so i'll adapt that
<oliv3r> (i passed the right init tons of times)
<Seppoz> use /sbin/init
<Seppoz> ok
<oliv3r> nogo
<Seppoz> first thing on firts boot you just use /bin/bash than do the 2nd stage debootstrap
<Seppoz> than change to init
<Seppoz> did you do that?
<oliv3r> i can do a --versbose; but it just runs tons of thigns; then still just halts
<oliv3r> i tried; i can run bash
<oliv3r> but after that, didn't do anything
<oliv3r> if you can walk me through that
<Seppoz> try
<Seppoz> init=/bin/sh
<Seppoz> and when you in sh type
<oliv3r> ok got root shell :)
<oliv3r> without mounted /dev/sda3
<Seppoz> try the 2nd stage debootstrapp now
<Seppoz> like /debootstrap/debootstrap --second-stage , not sure baout exact location
<oliv3r> no such file or dir
<oliv3r> i'll admit, i DID do a minbase install
<oliv3r> maybe I should have done a different variant?
<oliv3r> (i did a find, not ran the command)
<Seppoz> id use the debootstrap way
<oliv3r> yeah i did deboostrap -variant minbase
<Seppoz> ah ok
<oliv3r> i followed that mostly
<oliv3r> without schroot
<Seppoz> then you shoudl ahev atleast debootstrapp command in rfs
<Seppoz> find / | grep debootstrap
<oliv3r> i'll redoo the deboostrap with a 'bigger' variant
<oliv3r> find / -name "*deboot*"
<oliv3r> i did that; which dind't find it
<oliv3r> same
<oliv3r> i'll give that tomorrow a go
<oliv3r> witha bigger variant
<oliv3r> --variant=minbase|buildd|fakechroot|scratchbox
<Seppoz> --foreign DISTRO should do
<WarheadsSE> Turl: okay, what shall everyone blame you for? besides everything?
<oliv3r> guess i'll go witha buildd
<oliv3r> i figured, get a minbase going first; then go from there
<oliv3r> as i could apt-get anything i'd want
<oliv3r> but since i can't even log in/boot the damn thing :)
<oliv3r> --second-stage-target=DIR
<oliv3r> i'll try that from within fedora
<oliv3r> sleep now
<Seppoz> are you using chroot method? qith qemu?
<Seppoz> or just normal debootstrapping procedure
\\Mr_C\\ has quit [Quit: .]
<Seppoz> sudo debootstrap --variant=buildd --arch armel hardy /DIR http://archive.ubuntu.com/ubuntu/
<Seppoz> should do
<Seppoz> sec i forgot --foreighn
<oliv3r> i literally did that but with armjf, saucy and minbase
<oliv3r> i did not ise fpreigm
<Seppoz> let me grab my marsboard
<Turl> WarheadsSE: :p I was just kidding
<oliv3r> im in bed now:p
<WarheadsSE> Turl: :P
<oliv3r> WarheadsSE: doesnt explain why yhe same goes wrong with the alip omage though ...
<WarheadsSE> oliv3r: I didn't read much surrounding it.
<WarheadsSE> what now
<oliv3r> i ment Seppoz ;p
<Seppoz> 1 sec :)
<Seppoz> you cant keep me awake and go to bed lol
<oliv3r> lol
<oliv3r> if i understand the mampage right, minbase is reasonable for setting up a desktop
<oliv3r> hmm foreign might be not needed but ill do it anyway
<Seppoz> did you use --foreign option?
<Turl> you don't need foreign if you're debootstrapping the same arch
<Seppoz> but he isnt
<Turl> foreign is for eg when you want to debootstrap armhf on a x86 box
<Seppoz> thats what hes doing
<Turl> he is I think - he said he was doing it on fedora on the device?
<Seppoz> woot
<Turl> or did I misread? :p
<Turl> oliv3r: ^
<Turl> bbiab
<Seppoz> oliv3r: deboostrapping stage 2 now on device
<Seppoz> in like 5 mins we should know...
markvdb has quit [Quit: Page closed]
enrico_ has quit [Quit: Bye]
<Seppoz> oliv3r: here are my instructions for uboot: http://pastebin.com/yaEN7m7h - i just tested it and it works fine, let me know if you have any issues
<Seppoz> infact i missed to delete one line, here is the most recent:
<Seppoz> you might also want to use an armhf veryion istead of armel
bfree__ has joined #linux-sunxi
bfree_ has quit [Ping timeout: 260 seconds]
egbert has quit [Disconnected by services]
egbert has joined #linux-sunxi
naobsd has quit [Quit: Page closed]
deasy has quit [Quit: Nom d'un quark, c'est Edmonton !]
RomAbptTHN has joined #linux-sunxi
RomAbptTHN has left #linux-sunxi [#linux-sunxi]
TheSeven has quit [Disconnected by services]
[7] has joined #linux-sunxi
Gerwin_J has quit [Quit: Gerwin_J]
<libv> oliv3r: why are you passing init= ?
zIWYIiWPIS has joined #linux-sunxi
zIWYIiWPIS has left #linux-sunxi [#linux-sunxi]
pacopad_ has joined #linux-sunxi
pacopad has quit [Ping timeout: 272 seconds]
pacopad_ is now known as pacopad
Quarx has joined #linux-sunxi
nedko has joined #linux-sunxi
Black_Horseman has quit [Ping timeout: 272 seconds]
jelly-home has quit [Ping timeout: 264 seconds]
mnemoc has quit [Ping timeout: 264 seconds]
FreezingCold has quit [Ping timeout: 264 seconds]
MadSpark has quit [Ping timeout: 264 seconds]
MadSpark has joined #linux-sunxi
mnemoc has joined #linux-sunxi
FreezingCold has joined #linux-sunxi
hurtigbuffer has joined #linux-sunxi
pacopad has quit [Ping timeout: 272 seconds]
pacopad has joined #linux-sunxi
pacopad_ has joined #linux-sunxi
pacopad has quit [Ping timeout: 252 seconds]
pacopad has joined #linux-sunxi
pacopad_ has quit [Ping timeout: 272 seconds]
pacopad_ has joined #linux-sunxi
pacopad has quit [Ping timeout: 272 seconds]
pacopad_ is now known as pacopad
pacopad_ has joined #linux-sunxi
pacopad has quit [Ping timeout: 272 seconds]
pacopad_ is now known as pacopad
FreezingCold has quit [Ping timeout: 260 seconds]
rz2k has joined #linux-sunxi
AreaScout has joined #linux-sunxi
gzamboni has quit [Read error: Connection reset by peer]
gzamboni has joined #linux-sunxi
pacopad_ has joined #linux-sunxi
pacopad has quit [Ping timeout: 272 seconds]
pacopad_ is now known as pacopad
paulk-collins has joined #linux-sunxi
Relkin has quit [Quit: ""]
Relkin has joined #linux-sunxi
<libv> ssvb: have you had any time with dri2 yet?
<mnemoc> jikes, 5124 unread mails in my linux-sunxi folder :\
<mnemoc> libv: hi, is your kms driver somehow usable yet?
<libv> mnemoc: very very limited
<libv> mnemoc: and ugly atm
<libv> mnemoc: on hold until fosdem
<libv> mnemoc: it needs: i2c on both vga and hdmi (hdmi work started), lcd support (some poking was done, but nothing more), then LVDS, palette and interrupts need to be implemented
<libv> then a round if clean-up and then it can be included in sunxi-3.4
<libv> so not much work really
<libv> 1-2 weeks
<libv> then porting to mainline and devicetreeing...
<mnemoc> i see
<libv> vga and hdmi are working fine (both still lack ddc though), hw cursor and planes are working as well, iirc about 6kloc atm
<libv> ah, right, needs testing on a20 and needs fixing on A13
<libv> so 2/3rds of the way there i'd say, at least for sunxi-3.4
<libv> ump integration might be nice to have
<mnemoc> nice
<plaes> oh sweet5
<oliv3r> Seppoz: turl; yeah i was debootstrapping on the same arch
<oliv3r> libv: so i can add --verbose to see hwere it hangs
<libv> oliv3r: well, there is something you are doing different than me or Turl
<oliv3r> cp -ar binary/* /mnt/; copy script.bin, kernel and boot.* uEnv.txt
<oliv3r> that's pretty much it
<oliv3r> but appearantly there's voodoo magic involved
<libv> oliv3r: which boot.* and which uEnv,txt
<libv> oliv3r: pastebin them
<libv> oliv3r: change "root=/dev/sda1 rw rootfstype=ext4" to "root=/dev/sda1 rootwait"
<libv> oliv3r: does this even work without ' around it?
<oliv3r> sure does!
<oliv3r> so remove rw, add rootwait
<oliv3r> but if it has trouble mounting the root, without rootwait, it should barf heavily
<oliv3r> but i'm trying anything now
HdkR has joined #linux-sunxi
<oliv3r> O.
<oliv3r> bobs' your uncle
<oliv3r> removed rootwait; I cannot belive that's the culprit
<oliv3r> it's not
<oliv3r> rootwait doesn't matter
<oliv3r> rw matters
<oliv3r> ro/rw are offlimits
<oliv3r> don't even try to explain why
<oliv3r> or it's the quotes
<oliv3r> well atleast the ubuntu rootfs still crashes; but got alip now so that's good
<libv> crashes?
jinzo has joined #linux-sunxi
rellla has joined #linux-sunxi
rellla has quit [Ping timeout: 259 seconds]
rellla2 has joined #linux-sunxi
Gerwin_J has joined #linux-sunxi
Gerwin_J has quit [Client Quit]
<atiti> are you guys around fosdem_
<atiti> ?
hani has joined #linux-sunxi
<oliv3r> libv: well probably same as before; it doesn't resond to keyboard/serial but prints when i eject sd card
<libv> oliv3r: so this alip image does not work?
<libv> or are you talking about another rootfs?
<ssvb> libv: I got a bit sidetracked implementing dlopen based wrappers for mali blobs
<libv> ssvb: why would you need that?
<oliv3r> libv: yeah alip does; debootstrap ubuntu won't
<oliv3r> but i'm happy
<libv> oliv3r: ok
<libv> oliv3r: are you getting anywhere with lima or with cedrus?
<hani> is there a quick and easy way to export a series of patches in a newsgroup (linux-sunxi for example) and apply it to a local repo
lioka has quit [Changing host]
lioka has joined #linux-sunxi
<libv> hani: s/newsgroup/mailing list/
<libv> hani: you can save the mails directly, and then git am them
<libv> hani: or you can set up some hook to your mail client, google will tell you more
<ssvb> libv: the original mali blobs just cause too many problems if packaged as-is
<libv> ssvb: i would even be very happy with a special branch of fbturbo just for lima
<libv> ssvb: if not, i will have to demo without overlays, which will make it very slow
<hani> oh I use to forward each post into my email and then export yo mbox format and then use git am
<ssvb> libv: yes, I have not given up on that yet
<hani> libv: thanx
Faisal_ has joined #linux-sunxi
<oliv3r> libv: well now that i have a booting system; it's time to install allt he requirements and re-setup cedrus; but under fedora it worked; so a small step to get that ported over to alip
Faisal has quit [Ping timeout: 252 seconds]
Faisal_ has quit [Remote host closed the connection]
hurtigbuffer is now known as jelly-home
ZetaNeta has joined #linux-sunxi
hani has quit [Quit: Page closed]
<ZetaNeta> hi
<oliv3r> hey
<ZetaNeta> Yesterday i bought a mk802.....
<ZetaNeta> Are all mk802 based on allwinner?
<ZetaNeta> Because mine came with a AMLogic AML8726-m....
<ZetaNeta> And i dont know what to do with it now
<ZetaNeta> I can return it (maybe), but the 2 bad points will be: 1) I am again without a MK802 2) The shop is very far from here (Yes, i dont buy through inet)
<ZetaNeta> I know this is allwinner related channel... but, is it possible to "do something" with what i bought?
<oliv3r> ZetaNeta: most
<oliv3r> ZetaNeta: but some are a mistery
<oliv3r> ZetaNeta: amlogic is very linux unfriendly
HdkR has quit [Ping timeout: 272 seconds]
<ZetaNeta> Well, mine got a "burn through plastic" amlogic
<oliv3r> btp?
<ZetaNeta> Well, on its "default" speed, its still more hot than a overclocked a10
<oliv3r> libv: yay my saucy ubuntu image boots :D
<oliv3r> libv: the ubuntu issue was very likly a --second-stage; thanks to seppos
<oliv3r> maybe also due to ro/rw
<oliv3r> but alip + ubuntu now owrk :)
<ZetaNeta> owrk... i need to remember that word
<oliv3r> ZetaNeta: it's like work; but owrk
<ZetaNeta> it's like owned; but pwned
<ZetaNeta> :D
<ZetaNeta> Who want to laugh hard? http://www.amlogic.com/product03.htm
<ZetaNeta> HDML5... marketing!
<libv> ZetaNeta: it's a bit too over the top for me :p
<ZetaNeta> ?
<libv> ZetaNeta: nah, the best one i still the rk3288
<libv> with it's cortex A17 and mali t674
<libv> its even
<libv> should be cortex a12 and mali t624 :)
<ZetaNeta> Who wants the stick i got? I can exchange.
<libv> ZetaNeta: nobody, it's amlogic
<ZetaNeta> )
<ZetaNeta> s/)/(
<oliv3r> ZetaNeta: lkcl investigated amlogic for its use with his eoma stuff; but they where GPL unfriendly etc so that chip got 'nonoed'
<oliv3r> not sure if there's an active community around it
<ZetaNeta> Eh... and i am left with that chip
<oliv3r> or any community at all
<oliv3r> i DO know, that amlogic is used in the pivos box
<oliv3r> and various other xbmc based bnoxes
<oliv3r> pivos does (financially) support xbmc dev on these chips
<ZetaNeta> Its worser than A10 in all places, its more expencive than a a10 mk802, and i am not sure if i can return it
<oliv3r> you should be able to
<oliv3r> it's a fake chinese mk802
<oliv3r> mk802 IV is a RK based chip
<ZetaNeta> Can i use something from here to make gentoo run there? http://openlinux.amlogic.com/
<lkcl> ZetaNeta: whatever you do do *not* get involved with amlogic products.
<ZetaNeta> I know this... but i dont think i have a choice
<ZetaNeta> Maximum i can do is to return it. But i dont remember that one came with a warranty thingy
<lkcl> ZetaNeta: apologies but consider it to be a lesson learned. throw it away, sell it on ebay and then come back here and ask *before* spending money.
<oliv3r> lkcl: in his defence, mk802's usually are a10 based
<oliv3r> lkcl: luck of buying black boxes
<ZetaNeta> lkcl, I came in a shop "real far away", and thought quite a bit before buying it, but there were no wifi around :3
<lkcl> ZetaNeta: return it, sell it, whatever: anything else you have several months worth of reverse-engineering ahead of you if you attempt to stick with it.
<oliv3r> lkcl: don't you think amlogic changed their stance?
<lkcl> oliv3r: they don't have to. like rockchip they have big customers stuck with them, as long as they can keep on supplying them with GPL-violating binary-only but *functional* OSes, what do they care
<oliv3r> lkcl: i did read some time ago that most sources where released though; execpt for vpu
<ZetaNeta> FSF is a powerful foundation....
<lkcl> it's the usual story. ZetaNeta you are extremely unusual in that you want to do something different
<oliv3r> lkcl: ZetaNeta IS unusual :)
<lkcl> ZetaNeta: but they need to actually have a copyright case
<lkcl> :)
<oliv3r> lkcl: will you be attending fosdem2014?
<ZetaNeta> nop
<ZetaNeta> lkcl, "Its all about linux"
<lkcl> oliv3r: i'd love to - but i no longer have a permanent job so i now have to think carefully about what to do.
<mnemoc> :/
<oliv3r> ouch
<oliv3r> well there's live-streams
<oliv3r> :p
<oliv3r> and video's :p
<mnemoc> really sorry to hear about your job :(
<oliv3r> lkcl: quit to focus on rhombus?
netlynx has joined #linux-sunxi
netlynx has joined #linux-sunxi
<lkcl> oliv3r: to do that i need money, don't i. there's not been *any* funding - at all - since the project started.
<lkcl> oliv3r: if i had funding the project would have been a lot further along, wouldn't it?
<mnemoc> lkcl: indiegogo campaign?
* ZetaNeta wants a80. /me wants a80
<mnemoc> ZetaNeta: PVR6 and sata support is uncertain
<oliv3r> lkcl: i thought you had a customer/client that wanted to build his stuff around eoma?
<lkcl> mnemoc: indiegogo campaigns require a working prototype. a working prototype at this kind of level means you need funding, doesn't it?
<oliv3r> ZetaNeta: a80 is powervr
<oliv3r> ZetaNeta: and it likley doesn't have sata
<lkcl> oliv3r: yes... and they chose not to follow our advice so it's delaying the project.
<mnemoc> lkcl: you do have the PCB designs for the tablet, extend that to a good laptop, show renderings, sell
<oliv3r> lkcl: you have the eoma68 cards don't you? and there's the inventus thing or whatever it was called?
<mnemoc> improv
<lkcl> mnemoc: that requires funding to make the PCB so that it can be proved to be correct.
<ZetaNeta> oliv3r, Is powervr bad?
<lkcl> oliv3r: yes, and yes.
<lkcl> ZetaNeta: ask anyone who has a laptop with an Intel Atom GMA chipset.
<libv> lkcl: how is improv selling so far?
<lkcl> libv: i don't know because it's our client's project. we're waiting for them to place the order for the 2,500 units.
<mnemoc> lkcl: sure you need to prototype... but that is after having the indiegogo money
<libv> lkcl: the client being makeplaylive.com ?
<mnemoc> vaultlalala
<libv> mnemoc: ?
<lkcl> libv: yes.
<ZetaNeta> and... eh.... why did they dislike sata?
<mnemoc> the company that sells the improv is called vault something
<oliv3r> lkcl: i thought imrpov was a hobby thing by aron?
<libv> ah :)
<mnemoc> ZetaNeta: they aim at tablets and "phablets".... no use for sata there
<ZetaNeta> mnemoc, I am still waiting for a allwinner phone
<mnemoc> ZetaNeta: but Benn said that the soc supports SATA even if they didn't add it to the optimus board
<Seppoz> oliv3r: seeny my howto?
<mnemoc> ZetaNeta: unfortunatelly everything still in rumour stage
<oliv3r> Seppoz: aye, even used parts of it
<ZetaNeta> I am waiting for a80, because its powerful enough for what i need it
<lkcl> ZetaNeta: the cost of the chipsets to do the R.F. baseband etc, which in almost all cases means having an ARM SoC embedded in them, mean that it's cost-ineffective.
<Seppoz> is it working now?
<oliv3r> Seppoz: i fixed it all; i used alip to do a fresh bootstrap; which left me with a deboostrap dir; which i could use to do a 2nd stage from; i needed to do the 2nd stage; it works now
<lkcl> ZetaNeta: allwinner would need to license the baseband DSPs and put them *into* allwinner SoCs in order to be cost-effective.
<Seppoz> great
* mnemoc hopes this optimus board will help/encourage people to RE the pvr
<oliv3r> nah
<libv> lkcl: so basically makeplaylive hasn't gotten 2.5k pre-orders yet...
<mnemoc> lkcl: the A31 has a fifth core, openrisc. called ar100
<lkcl> mnemoc: there's a project set up for that. _some_ work's been started. enough to know that it's a f*****g complex task.
<ZetaNeta> The SATA support, or atleast USB 3.0 is really important in my case.
<lkcl> libv: 2.5k _orders_. they're committed.
<oliv3r> lkcl: a23 has a diff one; arisc :p
<mnemoc> lkcl: this ar100 might be intended for baseband
<ZetaNeta> Also powerful encoding...
<lkcl> mnemoc: i heard about that. awesome idea.
<mnemoc> lkcl: sure it's a complex task, but this is an amazing board WITH jtag, uart, etc.... not available before
rellla2 has quit [Read error: Connection timed out]
<mnemoc> lkcl: it's easier to do that on a devboard than on a CE device
<libv> lkcl: in that case, why hasn't the order been placed yet?
<mnemoc> or... less amazingly hard :\
<libv> lkcl: are they waiting until all orders are in?
<libv> lkcl: which also means, waiting until the people who expected their board to be delivered in january start demanding their money back?
<lkcl> libv: because they're collecting people's money. they're basically doing a "kickstarter"-like campaign
rellla has joined #linux-sunxi
<oliv3r> eitherway, I do hope the eoma68 standard becomes a success. i still think the idea behind it is not bad at all
<libv> makeplaylive.com forums seem pretty empty
<lkcl> i gotta drop off some deliveries before lunch.
<lkcl> apologies gotta go folks. back in a couple of hours.
<oliv3r> and that at 13:37
<mnemoc> cu lkcl
<libv> the fact that people get the runaround between makeplaylive.com and vaultechnologies.com does not bring confidence...
<oliv3r> added the ubuntu repo's from packages.linux-sunxi and it pulled in mesa packages etc
<oliv3r> getting exited now :)
HdkR has joined #linux-sunxi
<ZetaNeta> hi HdkR, welcome to the club of anonymous allwinner users
<ZetaNeta> Lets all say hi to HdkR
<HdkR> What
<oliv3r> ZetaNeta: i'm not anonymous
<HdkR> I'm not anonymous or a AllWinner user :p
<ZetaNeta> HdkR, "Why then came this channel you to?"
<HdkR> Couple months ago libv posted a discussion that sounded like it was a fun time, so I joined in
<HdkR> I may get a AllWinner A80 just to be ashamed of IMGTec's drivers though
<mnemoc> isn't the A80 announced for Q2 2014?
* HdkR shrugs
<HdkR> Whenever it comes out
<HdkR> I've got my K1 dev board to play with until that time rolls around
<pacopad> @ssvb : Hi , i got a little question for you , i'm looking for a software that transmit image form /dev/disp remotly like a vncserver
<pacopad> @ssvb does it exist ?
<ssvb> pacopad: leaving alone /dev/disp for now, what exactly are you trying to achieve?
<pacopad> i'm playing with vdpau and xbmc, i'dont have always access to my TV because of my kids
<pacopad> and i'd like to see if an image is dsplayed
<pacopad> but with vncserver i got only black screen even if an image is displayed
<ssvb> pacopad: is the performance important for you?
rZr has quit [Read error: Connection reset by peer]
<pacopad> not at all
<pacopad> i just need to some images and the xbmc debug info
<oliv3r> so basically, you want 'occasional' screenshots (0.5 fps) to see whats on tv
<oliv3r> like a nanny-cam I suppose
<oliv3r> i was thinking about this as a feature actually too
<oliv3r> but for the android remote control
rZr has joined #linux-sunxi
<oliv3r> it would be nice to be able to control xbmc from your phone and actually see whats going on on screen
<oliv3r> the remote now is ... meh
<pacopad> @oliv3r : it's exactly what i want , 1 fps would be perfect
ZetaNeta has quit [Ping timeout: 248 seconds]
<ssvb> oliv3r: this means that we need to query the status of all the disp layers, grab the needed data, composite layers together and make a file with a nice screenshot
<oliv3r> ssvb: do all vnc solutions work this way?
<oliv3r> ssvb: i know under 'that other os' programs like dameware install something called a 'mirror driver'
<oliv3r> since they can't hack that into their drivers
<libv> ssvb: nope.
<ccaione> habemus regulators
<ssvb> libv: there is also another way
<ssvb> libv: disp writeback pipeline
<libv> ssvb: indeed
<libv> it might be interesting for miracast/widi in future
<libv> if it works
<oliv3r> that'll be a significant performance hog though, no?
<oliv3r> with the crammed memory bus as it is
<rm> install an $5 webcam
<rm> point it to the TV
<ssvb> rm: :)
<rm> low tech solution!
<oliv3r> bah no f2fs in 3.4
<libv> probably, but the hw seems to have this provision
<oliv3r> oh really? then that's very cool
<oliv3r> miracast is something exciting surely
<pacopad> @rm the TV is actually displaying cartoons for my kid :)
<oliv3r> or so you hope :p
<oliv3r> in a udev rule, how do i figure out what the KERNEL name is supposed to be?
<oliv3r> i tried cedar and cedar_dev, for the sunxi_cedar_dev module
<libv> did lima work out?
<oliv3r> almost there
<oliv3r> i'm installing build-esentials + co now
<oliv3r> but since i was copying the mali/ump udev rules, as i probably need them; i figured it was a good time to add disp + cedar to that
<libv> oliv3r: i have no idea about what that string needs to be though
<libv> module name, dev name... dunno
<oliv3r> yeah, i'll figure that out tomorrow
<oliv3r> first compiling some stuffs :)
<ssvb> oliv3r: for libvdpau this is kind of the first step to make it more user friendly - https://github.com/ssvb/xf86-video-fbturbo/commit/ef851c6207f484196dedf90a92523549c40a4444
<oliv3r> ssvb: ok i have to play catchup here first :)
<ssvb> oliv3r: but more stuff is needed for flawless libvdpau integration
<oliv3r> ssvb: i had vdpau running with .. i don't even know which driver on hansg's fedora image
<oliv3r> i just compiled libvdpau, installed it and used it with mpv
<oliv3r> with this fresh ubuntu image, i installed fbturbo
<oliv3r> so do I need anythign specific for fbturbo?
<oliv3r> [ 12.026] (II) FBTURBO(0): enabled G2D acceleration
<oliv3r> sweet
<oliv3r> i think the overlay branch needs g2d so that's great
<ssvb> oliv3r: the key part for libvdpau integration is 'flawless' :) in the sense that the CPU overhead and memory bandwidth use still can be reduced
<oliv3r> so if i have fbturbo loaded; i can just continue doing what i did before
<oliv3r> export VDPAU_DRIVER=sunxi
<oliv3r> mpv file -optionstoenablevdpau
<Seppoz> is there any sunxi optimized toolchain to start from?
<Seppoz> including cpu opti flags
<oliv3r> Seppoz: most use the linaro stuff
rz2k has quit []
<oliv3r> Seppoz: i use the gentoo version ;)
<Seppoz> but thats just generic arm isnt it?
<oliv3r> armhf i think
<Seppoz> do they even use the hardware floating layer?
<ssvb> oliv3r: yes, fbturbo is not any different from fbdev for libvdpau-sunxi right now, so all the current instructions work
<Seppoz> im searching for an optimized distro/toolchain combo, otherwise i gottam amke one which sucks
<oliv3r> bah no mpv in ubuntu :(
<Seppoz> cant you just apt-get it?
<oliv3r> nopers
<oliv3r> it is in trusty
<oliv3r> but not yet in saucy or saucy-backports
<Seppoz> mplayrt is very old
<Seppoz> should be in others too
<Seppoz> sec
<oliv3r> mplayer2 is in saucy, mpv isn't
<Seppoz> ok its not original mplayer
<Seppoz> well just apt-get build essential and compile it?
<Seppoz> should be quite easy
<oliv3r> effort effort
<Seppoz> mpv in trusty also isnt for arm i think
<oliv3r> i'm using mplayer2 for now
<Seppoz> lazy boy
<Seppoz> :)
<ccaione> oliv3r: do we have the voltage in output from axp? I want to see if what I get from /sys/ is what really is in output
<ccaione> (and here I don't have a multimeter)
<oliv3r> ccaione: you want me to measure the axp for you? is that your question?
<ccaione> :D just 3 pins if you have time
<oliv3r> what board
<ccaione> A20
<Seppoz> its basically set by cpu so you should either have it exported in sysfs or should be quite easy to do
<oliv3r> cb2 or cb3
<ccaione> cb2
<oliv3r> which pins/points
<oliv3r> tkae a pic and point out what you want measured
<ccaione> argh, ok
<ccaione> oliv3r: otherwise I can compare with linux-sunxi kernel. it is easier this way
<oliv3r> whatever is easier for you; i don't mind hooking up my scope to a a few pins if you tell me which pins :)
jemk has joined #linux-sunxi
<ccaione> oliv3r: thank you... let me check with sunxi-kernel before
<pacopad> so no way to transmit /dev/disp output ?
techn_ has joined #linux-sunxi
<oliv3r> not yet :p[
<pacopad> ok thank you
<oliv3r> jemk: mplayer1 doesn't work with vpdau does it? only mplayer2?
<jemk> oliv3r: it should work, but debian mplayer1 had no vdpau support compiled
<oliv3r> jemk: with the latest osd branch, with 4k patches, i can't open sintel-4k
<oliv3r> jemk: yeah i think that's why it fails with invalid vo
<oliv3r> mplayaer2 works fine, so i'll use that
<libv> aha, that french guy who hacked up some a13 tablet responded now
<libv> let's wait and see whether he will really work the wiki
<oliv3r> jemk: ffodivxvdpau is working properly too? or is that a lie in -vc help
<jemk> oliv3r: no mp4/divx yet, so it won't work. mplayer1+2 don't check if vdpau features are supported, they simply assume they are
<jemk> and the 4k sintel, what doesn't work? strange picture or nothing at all. but 4k isn't very useful anyways, too slow
<oliv3r> 4k just stops 'unable to decode intime'
<oliv3r> ok then i'll remove ffodivxvdpau
<oliv3r> do you know if mplayer2 checks /etc/mplayer/mplayer.conf?
<oliv3r> so i can set the vc by default
<oliv3r> seems to ignore it atm
<oliv3r> ok it does print it 'forced vc: ffh264vdpau' and VO [vdpau] but still 100% cpu usage
<oliv3r> blasted typo's! vdapu is something else :)
<jemk> oliv3r: can you remove the XClearWindow in presentation_queue.c:228 and see if it still has 100% cpu?
<oliv3r> it was a typo; i used vdapu instead of vdpau
<jemk> ah, ok
<oliv3r> any clue what parameters i should use to get a cedar rule going?
<oliv3r> i tried KERNEL== now trying DEVNAME=="cedar_dev"
<oliv3r> all won't fly
<jemk> udevadm info /dev/cedar_dev
<oliv3r> E: DEVNAME=/dev/cedar_dev
<oliv3r> N: cedar_dev
<oliv3r> E: SUBSYSTEM=cedar_dev
<jemk> works here: KERNEL=="cedar_dev",MODE="0666"
<oliv3r> that's exactly what i have
<oliv3r> works for disp
<oliv3r> not for cedar_dev
<oliv3r> lemme try 1 more thing
<oliv3r> reboot!
<jemk> and my udevadm output is exactly like yours...
<oliv3r> ok udev restart isn't enough
<oliv3r> reboot is required
<oliv3r> adding to wiki
<jemk> udevadm control --reload-rules && udevadm trigger
<jemk> but don't ask me if this is the recommended way
<woprr> lo
<woprr> u guys are just testing the VE ?
<libv> oliv3r: udevadm trigger ?
<libv> whut? then something else is probably
<libv> off
iamfrankenstein has joined #linux-sunxi
<libv> i had some issues when splitting up the udev from the howto, into ump and mali
<libv> it took ages to find out that it hadn't used the proper suffix for the file
<oliv3r> i took your syntax for disp and cedar
<oliv3r> maybe combine it all into 50-sunxi-video.rules
<oliv3r> right, time to compile liam
<libv> nah, in the end each of these bits comes with its own userspace library
<libv> and this library should also install the necessary udev rules
<oliv3r> too true
<woprr> the aw briefs state h.264 MVC decoding support, here're some testing streams but the fraunhofer hhi JMVC 8.5 extractor/decoder fails on them (requires some SEI and NAL extensions not in those samples or streams from BD), could You try some of the streams and report to list? A friend wants to know for his decision if he should try another effort to implement MVC in ffmpeg, but he doesn't want to do those huge work
<woprr> if we've got hardware decoder support for MVC in AW or Rockchip SoCs already...
iamfrankenstein has quit [Ping timeout: 252 seconds]
iamfrankenstein has joined #linux-sunxi
iamfrankenstein has quit [Read error: Connection reset by peer]
<woprr> heres the sample code to understand MVC (much easier than reading the papers ;) ) cvs –d :pserver:jvtuser@garcon.ient.rwth-aachen.de:/cvs/jvt checkout jmvc
kivutar has joined #linux-sunxi
<oliv3r> which file should I dl and test?
<woprr> wam...
<oliv3r> downloading mvcds1.264 right now
<woprr> for f in h264-MVC/*; do mediainfo $f 2>&1 ; done ...
<oliv3r> works fine; a little fast
<woprr> yes ok
<oliv3r> dancing chinese
<woprr> really? cool... so the base 2D stream works at least, fine thhx
<oliv3r> mvcsd123 all work
<woprr> well then I try to get the friend to join sunxi devs :-)
<woprr> thanks
<oliv3r> any other file i should test?
<woprr> no thanks :)
<woprr> I need to upgrade and rebuild my sunxi for the Olimex A20 , out of sync for months, the devel repository and branch stated on list by PL is correct?
<oliv3r> woprr: i just installed a saucy system using debootrap; add sunxi kernel; go
fredy has quit [Excess Flood]
<oliv3r> jemk: have you tried playing back 720p25_mobcal_ter.mkv?
<woprr> i've got a debian / already, thx, but kernel, uboot + fex, etc needs upgrading, many changes
fredy has joined #linux-sunxi
<jemk> oliv3r: not yet, i will
<oliv3r> it works
<oliv3r> but is very choppy
<oliv3r> and i get an artifact
<oliv3r> could of course be the encoding
<oliv3r> the ducks take off; into tree, park joy all work fine
<jemk> looks like it's the video, i can't see any difference to software decoding on pc
<oliv3r> ok
<oliv3r> see the tiny artifact too?
<oliv3r> it's a small green line
<oliv3r> flashing
iamfrankenstein has joined #linux-sunxi
<oliv3r> it's short, about 3 cm?
<oliv3r> 2 mm high
<jemk> yes, i see it on pc too
<oliv3r> ok
<oliv3r> sothe video is bad
<oliv3r> i'll delete it :p
<jemk> never trust sample media :p
<oliv3r> :p
<oliv3r> the rest works really well
<oliv3r> not sure what the diff is between my sintel-hd.avi and sintel.2010.1080p.mkv though; besides 700mb diff
<oliv3r> btw, what's needed to get osd working? i did install the osd branch
<oliv3r> presentationtime not supported
<jemk> presentation time is different problem, my part of osd only works with mplayer1, the bitmap part i don't know
<oliv3r> ah ok; then i won't worry about it
<jemk> I didn't put much time in the osd part, as it destroys other things with its current way of doing the second layer
<oliv3r> no prob
<oliv3r> so i should install master branch to get the most advanced stuff?
<jemk> to do it right we need some driver integration
<jemk> yes, master
<jemk> some fixes are even missing in osd
<oliv3r> rgr
iamfrankenstein has quit [Ping timeout: 272 seconds]
<oliv3r> libv: where does libUMP.so come from?
<oliv3r> ah i probably don't need it as i only need limare/lib
<oliv3r> ok installed that; now i'm stuck :p; i think you say to install dri/? but i guess / needs to be made first
<oliv3r> and there i need libUMP.so for; which i'm not sure yet where to get from? I did add/install the packages from packages.linux-sunxi.org
<oliv3r> building libUMP as per binary drivers guide
iamfrankenstein has joined #linux-sunxi
iamfrankenstein has quit [Ping timeout: 252 seconds]
<libv> oliv3r: it's in the email
<libv> and no, you do not need libUMP
<oliv3r> yeah i got libmali installed for the shader;
<oliv3r> but then i'm a little confused
<libv> build lima/limare/lib
<libv> and install it
<oliv3r> did that
<oliv3r> that's all good
<oliv3r> i have packages.linux-sunxi.org added as repo; so i'm guessing all mesa stuff should have come from there anyway
<libv> ok, then you go off and ./configure, make, make install in dri/
<libv> yes, that one dev package is only available in that repo
<oliv3r> yeah but i can't
<oliv3r> since there'sMakefile.am configure.ac src test
<oliv3r> they need to be autotooled
<oliv3r> which I guess make int he toplevel does?
<libv> no
<libv> autoreconf does it i think
<oliv3r> hmm, that's not installed with build-essentials :(
<oliv3r> autoconf is a build-essential imo
<oliv3r> throwing some random combination of automake; aclocal still causes autoreconf to barf errors about undefined LIBTOOL's but i got a configure now
<oliv3r> dricommon not found; that should be in the packages dir i guess
<libv> oliv3r: that's the dev package mentioned in the email.
<oliv3r> aye
<oliv3r> gotta cook dinner; but i did make install in limare/lib
<oliv3r> yet, configure: error: Limare library is missing.
<oliv3r> and it's happily in /usr/lib/liblimare.so
<oliv3r> after dinner, i will sort this
Nikolas has quit [Ping timeout: 272 seconds]
iamfrankenstein has joined #linux-sunxi
iamfrankenstein has quit [Read error: Connection reset by peer]
kivutar has quit [Ping timeout: 248 seconds]
kivutar has joined #linux-sunxi
rz2k has joined #linux-sunxi
tgaz has quit [Ping timeout: 272 seconds]
tgaz has joined #linux-sunxi
Quarx has quit []
HdkR_ has joined #linux-sunxi
HdkR has quit [Remote host closed the connection]
HdkR_ is now known as HdkR
<Seppoz> oliv3r: you did you autoconf right to get the ./configure script? and in the configure you sent the cross compile options right?
deasy has joined #linux-sunxi
<Seppoz> wow they are making alot of tables with the A31 now
<Seppoz> how is the mainlining project going? havent eard anything of it in ages?
<Seppoz> did you device to put it all in the 3-4 or is it still on?
deasy has quit [Quit: Nom d'un quark, c'est Edmonton !]
<ccaione> Seppoz: still going, pretty well if you ask me
<Seppoz> but its all with dtd right?
<ccaione> yes
<Seppoz> i hate dtd, cant use it
<Seppoz> did linux generally switch to dtd?
<ccaione> well, your problem :)
<ccaione> for mainline you must use dt
<Seppoz> ok
<oliv3r> for arm, yes, for x86, you don't have to yet
<mnemoc> found an ooooooold 40GB/2.5" drive from the dell laptop I had back in 2004 :)
uwe_ has quit [Ping timeout: 252 seconds]
<Seppoz> cool
<mnemoc> DT has haters even among some key kernel devs... but there is no better solution, so DT is here to stay :\
<Seppoz> well i find DT very strange for handling drivers
<Seppoz> ever tried to setup something like ads1015 with dt? have fun
<mnemoc> today I was reading some ranting and how the DT/led implementation made impossible to use pwm controlled leds and reversed values
<Seppoz> what i hate most ist dtaht dt has th be though the whole system, including drivers
<Seppoz> i was absolutely fine with the mach files
<Seppoz> what was wrong about them?
<mnemoc> multiplatform
<Seppoz> well you just had a mach file per board
<Seppoz> when speaking about embedde devices you alway shave fiderent systems anways
<mnemoc> everything needs to be a driver to be able to decide things on runtime
<Seppoz> i disagree as i dont see an advantage
<Seppoz> either somehting is physically there ot nor
<Seppoz> *or not
<Seppoz> ever seen a plug and play psu for example?
<Seppoz> or a plug and play ADC via i2c?
<Seppoz> or a plug and play eeprom?
<Seppoz> either its on the board or it isnt
<mnemoc> they want to be able to use the same bin in any armv7 board
<Seppoz> yes well happy setting up lcd via dtd
<Seppoz> or configuring lcd drivers over dt
<Seppoz> its gonna get a real nightmare
<Seppoz> dont you think?
<Seppoz> also what about old drivers?
<Seppoz> will take twice as long to port now
<mnemoc> I also hate that you are not allowed to detect things programatically either. you MUST trust the dtb even for things you can find out yourself... and more reliably
Fusing has joined #linux-sunxi
uwe_ has joined #linux-sunxi
<Seppoz> i really dont know what was wrong about configuring things in the kernel and so on
<Seppoz> who need generic kernels on embedded systems?
<mnemoc> I told you. they want single-bin
<Seppoz> you cant start them anyways because of the bootloader situation
<Seppoz> so you have to dig in anways
<Seppoz> it will never be like adding a cd and start
<mnemoc> bootloader is expected to pass the right dtb to the generic uImage
<Seppoz> unless they make a generic uboot lol
<mnemoc> it makes for generic distributions...
<Seppoz> yes no i get that
netlynx has quit [Remote host closed the connection]
<Seppoz> i mean the concept is ok when its completed, but it really changes alot
<Seppoz> also it requires porting of ALL drivers if i understoot that right
<mnemoc> the facepalm part comes when you realize the dtb parser is too complex for been used in u-boot-spl... so you still need to hardcode the dram initialization and core pinmuxing in u-boot
<Turl> mnemoc: you can do programatic detection on babelfish if you so desire :)
<Seppoz> well you will never be able to have a customized bootloader so i think think it matters much
<mnemoc> Turl: did you see the post about A20 rev B not booting with A20 rev A code?
Fusing2 has joined #linux-sunxi
Fusing has quit [Ping timeout: 246 seconds]
<Seppoz> by the way, how do i detect the ram speed when not getting information from producer?
<Seppoz> or basically ram settings
<Seppoz> speaking about the witstech core boards
<Turl> mnemoc: it was a matter of axp derp I thought? just needed a power cycle or sth
HdkR_ has joined #linux-sunxi
<mnemoc> Turl: and so a different .dtb per board AND soc revision... if you are lucky enough to know which soc rev you got
HdkR has quit [Remote host closed the connection]
<mnemoc> Seppoz: it is believed that livesuit does some trial and error
HdkR_ is now known as HdkR
<Turl> no need for soc revision thing
<mnemoc> good
<mnemoc> Turl: I would have loved DT if it would have been possible to read the board details in u-boot-spl
<mnemoc> everything feels half cooked
<Turl> mnemoc: blame aw for putting too little sram on there :p
<ccaione> at work I'm working with a SoC with 20MB sram on-die
<mnemoc> blame the world for not been compatible with my perfect design!
<Seppoz> cant you just make a parser that extracts the dram settings by scanning trought the dtd
<Seppoz> like reading segments of the dtd
<mnemoc> Seppoz: also the pinmux
<mnemoc> and understanding the references
<Seppoz> you should be abe to set up une by one, yet i bet it would be a pain
<Turl> ccaione: thar's a lot heh :)
<Turl> that's*
<ccaione> Turl: yeah :)
Black_Horseman has joined #linux-sunxi
Black_Horseman has quit [Changing host]
Black_Horseman has joined #linux-sunxi
<Turl> ccaione: it must be a pretty expensive soc
<ccaione> dunno the price, it's not on the market yet
<Seppoz> yes what is it? some fpga?
<ccaione> renesas SoC
<Seppoz> ah ok
<Seppoz> looks interessting
<ccaione> cooking time, bbl
<Seppoz> what ya cookin?
<Seppoz> ;P
<ccaione> meat :D
<Seppoz> xD
Faisal has joined #linux-sunxi
<Faisal> anyone knows why bcmdhd creates not only wlan0, but also p2p0 and bond0 ?
<Seppoz> i think that depends on the setup of it
<mnemoc> tsvetan's talk conflicts with one about kernel debugging :\
<Turl> Faisal: p2p0 is the wifi direct thing iirc
<Faisal> Turl: it was working before with just wlan0 dont know what happened now i got those 3 devices and i cant scan wireless networks for some reason
<Seppoz> what happens if you scan?
<Seppoz> you have wireless tools on the device?
<Seppoz> did you try to scna manually?
<Seppoz> iwlist wlan0 scan
<Faisal> cubie@Cubian:~$ iwlist wlan0 scan
<Faisal> wlan0 Interface doesn't support scanning.
<Faisal> cubie@Cubian:~$ iwlist p2p0 scan
<Faisal> p2p0 No scan results
<Seppoz> cat /etc/wpa_supplicant.conf
<Faisal> sorry
<Faisal> didnt do sudo
<Faisal> working actually
<Faisal> lol
<Seppoz> xD
<Faisal> well i can see the networks at least
<Faisal> any ideas on how i can get that configured once and for all
<Faisal> i need to it to access my AP
<Seppoz> using wpa supplicant
kivutar has quit [Read error: Connection reset by peer]
kivutar has joined #linux-sunxi
<Faisal> isnt there a way to use /etc/network/interfaces ?
<Seppoz> thats basically what it does only via wpa supplicant
paulk-collins has quit [Remote host closed the connection]
<Faisal> Seppoz: Awesome !
<Faisal> you're the man
<Seppoz> yw
Fusing2 has quit [Ping timeout: 272 seconds]
tomboy64 has quit [Ping timeout: 264 seconds]
iamfrankenstein has joined #linux-sunxi
lioka has quit [Quit: leaving]
Fusing has joined #linux-sunxi
lioka has joined #linux-sunxi
kivutar has quit [Ping timeout: 252 seconds]
VargaD_ has joined #linux-sunxi
VargaD has quit [Ping timeout: 248 seconds]
VargaD_ is now known as VargaD
kivutar has joined #linux-sunxi
Vince has joined #linux-sunxi
Faisal_ has joined #linux-sunxi
Faisal has quit [Ping timeout: 260 seconds]
deasy has joined #linux-sunxi
uwe_ has quit [Ping timeout: 272 seconds]
iamfrankenstein has quit [Ping timeout: 260 seconds]
jemk has quit [Ping timeout: 252 seconds]
rellla has quit [Remote host closed the connection]
uwe_ has joined #linux-sunxi
<oliv3r> mnemoc: wasn that old news? that mozilla was using the a31, propriatery or not
<oliv3r> i onder if they show up to fosem again; they did a few years
<mnemoc> oliv3r: i had no idea
pacopad has quit [Quit: pacopad]
<oliv3r> ah, lucky you!
<oliv3r> kidna sucks big time :(
<oliv3r> if I copy a shared so, to /usr/lib/; do i need to do anything for the linker/configure to make it find it?
<libv> oliv3r: are you sure that your issue is not with the libtool warnings you got?
<oliv3r> this is the AC_CHECK_LIBS(); thing
<oliv3r> but i'll fix those too
<oliv3r> src/Makefile.am:4: The usual way to define 'LIBTOOL' is to add 'LT_INIT'
<oliv3r> src/Makefile.am:4: to 'configure.ac' and run 'aclocal' and 'autoconf' again.
<oliv3r> that's what i'm missing
<oliv3r> heh, but that's even there
<oliv3r> i'll reinstall build-essential; though i'm 110% sure i installed them
<oliv3r> how else did i get gcc
<libv> oliv3r: build-essential probably does not include libtool
<oliv3r> yeah it doesn't seem to
<oliv3r> strange
<Faisal_> ok guys, i now have wifi and bluetooth working on the cubietruck, next step is to get 2d acceleration, im using a cubian image r4 from the cubieboard.org website, is there a way to check if maybe this is already enabled, (i doubt it)
<oliv3r> ok running aclocal and autoconf fixes it
<oliv3r> well creates configure without errors
rz2k has quit []
<oliv3r> now lets see if autoreconf runs without errors
<oliv3r> it does not
<oliv3r> are you sure i need to run autorecon? aclocal + autoconf seem to do the trick
<oliv3r> aclocal; autoconf; libtoolize; automake --add-missing; autoreconf
<oliv3r> that's all
<oliv3r> configure: error: Limare library is missing.
<oliv3r> FU!
<libv> oliv3r: ldconfig -v | grep limare
<oliv3r> /sbin/ldconfig.real: Path `/usr/lib/arm-linux-gnueabihf' given more than once liblimare.so -> liblimare.so
<oliv3r> seems to be there
<oliv3r> which makes sense, as make install ran fine
<oliv3r> and the file seems to be happily in /usr/lib/liblimare.so
<oliv3r> only think i can imagine is that limare_init isn't in that .so
<libv> AC_CHECK_LIB([limare],[lima_init],[]
<oliv3r> grep limare_init /usr/lib/liblimare.so
<oliv3r> Binary file /usr/lib//liblimare.so matches
<oliv3r> oh lima_init
<oliv3r> also matches; both exist
<libv> oliv3r: find /usr/ -iname "liblimare.*"
<libv> although, this should be a clean install
paulk-collins has joined #linux-sunxi
<libv> oliv3r: check the config.log
<libv> is there a liblimare dependency that's missing?
<libv> perhaps /usr/lib/libMali.so ?
<mrnuke> I don't know what to do about this: http://fpaste.org/69849/64028139/
<mrnuke> (sunxi-devel branch)
<oliv3r> libUMP probably then
<oliv3r> good point config.log
<libv> oliv3r: no, there is no dependency on libUMO
<libv> oliv3r: the limited ump code i need i actually have that built in
<oliv3r> /usr/bin/ld: warning: libUMP.so, needed by /usr/lib/libMali.so, not found (try using -rpath or -rpath-link)
<oliv3r> /usr/lib/libMali.so: undefined reference to `ump_close'
<oliv3r> I downloaded libMali from github as per mail
<libv> ah
<libv> right, libMali.so needs it
<oliv3r> /usr/lib/libMali.so: ELF 32-bit LSB shared object, ARM, version 1 (SYSV), dynamically linked, BuildID[sha1]=0x08de77ba08bcb840ef6f8f68770b7e430a8b5c3e, not stripped
<libv> oliv3r: did you install fbturbo from a package?
<libv> oliv3r: anyway, apt-get install libump
tomboy64 has joined #linux-sunxi
<oliv3r> libump is already the newest version.
<libv> ah
<libv> hah
<libv> oliv3r: apt-get install libump-dev
<libv> oliv3r: the normal package contains libUMP.so.3.0.0, and nothing more
<libv> libMali.so depends on libUMP.so however
<libv> ah, it also installs libUMP.so.3
<oliv3r> ok that's pulling something in
<libv> but only -dev installs libUMP.so
<oliv3r> /usr/include/mesa/main/glheader.h:53:19: fatal error: GL/gl.h: No such file or irectory
<oliv3r> that should be easy to fix :p
<libv> mesa-common-dev
<oliv3r> probably should just instlal all packages from packages.linux-sunxi.org
<libv> oliv3r: no
<libv> oliv3r: there is only a limited set of these that you will end up needing
<libv> oliv3r: there is logic to this
<libv> and if everything were packaged up correctly, it all would be automated
<libv> anyway, the GL/gl.h thing is a shortcoming of my packaging
<HdkR> gl.h woo
<oliv3r> i think it worked
Fusing has quit [Ping timeout: 252 seconds]
<libv> find out by running es2gears
Fusing has joined #linux-sunxi
<libv> from mesa-utils-extra
<oliv3r> configure worked :p compiling now
<oliv3r> lima_dri.c:141:3: warning: implicit declaration of function 'lima_frame_new' [-Wimplicit-function-declaration]
<oliv3r> usually that's not good :p
<libv> oliv3r: it exists, just not exported in this code
<oliv3r> ok
<libv> oliv3r: well, not in any header
<oliv3r> compil worked error free
<libv> oliv3r: it's completely written out now anyway
<libv> now that i have some fbos working
<Seppoz> is the ram settings in script.bin actually used or is it only set by uboot
<oliv3r> libv: just reporting as i see things; so that atleast someone heard it :p
<oliv3r> Seppoz: only for shits n giggles
<Seppoz> what does that mean?
<oliv3r> livesuit uses that to change boot1
<oliv3r> boot0
<Seppoz> you talking about ram parameters?
<Seppoz> by the way is it possible to disable axp in the kernel yet?
<Seppoz> last time i tried there was some problems with gpio pins of it beeing used somewhere else
<oliv3r> Seppoz: livesuit uses script.bin to 'calculate' the ram parameters for boot0
<oliv3r> or so we expect
<oliv3r> how do you want to disable axp
<oliv3r> or what do you expect to gain
<Seppoz> im not talking about uboot i mean the kernel
<oliv3r> afaik axp 'just works'
<Seppoz> im wondering if the kernel setups the ram again or if only uboot does
<Seppoz> i dont have axpon my hardware
<Seppoz> and whenever i boot i get a kernel failture on axp
<oliv3r> Seppoz: standby mode changes the ram
<Seppoz> kernel still keeps going but looks crappy
<oliv3r> on a10-a20
<oliv3r> a31 and a23 have a seperated risc core; that takes care of ram
<Seppoz> i mean the settings ins cript.bin
<oliv3r> big issue here is, it's all 'hidden' so we don't know how to setup ram
<oliv3r> and thus, no u-boot
<Seppoz> does the kenrel setup the ram again
<Turl> Seppoz: did you disable it on script.bin?
<oliv3r> Seppoz: partially, it gets it from auto refresh, back to 'normal' mode
<Seppoz> im not sure let me check
<Seppoz> yes
<Seppoz> pmu_used = 0
<oliv3r> libv: i don't think it's working yet, oh permission denied on mali
<oliv3r> i can fix that
<Faisal_> hey guys, when i run sunxi_disp_g2dbench, it tells me i have no g2d accelration, how can I get that please ?
rompopov has joined #linux-sunxi
rompopov has quit [Remote host closed the connection]
<oliv3r> libv: lots of debug talk; but SUCCESS! 125 FPS
<Seppoz> please check the sunxi wiki: http://linux-sunxi.org/Binary_drivers
<Seppoz> oliv3r: cool oliver
lioka has quit [Quit: leaving]
<libv> oliv3r: echo "performance" > /sys/devices/system/cpu/cpu0/cpufreq/scaling_governor
<oliv3r> libv: compiled by default :)
<libv> and of course there is lots of debug talk
<oliv3r> we did some power analsys
<libv> oliv3r: what's the cpu usage like in top?
tatinik has joined #linux-sunxi
<libv> i have double that performance on my cubietruck though
<oliv3r> 2977 root 20 0 46832 31m 8256 S 38.9 1.6 3:42.28 Xorg
<oliv3r> 27990 fosdem20 20 0 57860 7668 5420 R 37.6 0.4 0:02.85 es2gears
<libv> check your Xorg.0.log
<libv> it probably is not using overlays
<libv> should be like 50% es2gears, 8% Xorg
<libv> as long as you're not using glmark2-es2, it will look ok
<libv> well, at least until ssvb fixes the buffering :p
tatinik has quit [Ping timeout: 272 seconds]
<oliv3r> q[ 1691.949] (II) FBTURBO(0): enabled display controller hardware overlays for DRI2
<libv> oliv3r: also, resizing is broken
<libv> oliv3r: again, linked to the buffering
<oliv3r> i'm not touching resize though
<oliv3r> i leave it at its default 300x300
<oliv3r> [ 1692.000] (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)
<Faisal_> Seppoz: I did what was in http://linux-sunxi.org/Binary_drivers successfully but when i run for example sunxi_disp_vsyncdemo it tell me explicitely g2d_accel=no
<Faisal_> [ 44.089] (II) FBTURBO(0): enabled display controller hardware overlays for DRI2
<Faisal_> i dont know what is wrong with it, but I really can tell it is not accelerated actually
Fusing has quit [Quit: Nettalk6 - www.ntalk.de]
<ssvb> Faisal_: check if you have ondemand cpufreq governor in use
<ssvb> Faisal_: as for g2d, the permissions for /dev/g2d might be not set properly
<Turl> ssvb: do you still need to specify the accel method as g2d?
<Faisal_> ssvb: ok, quick question how can i check that ?
<Turl> I recall having to set something up on xorg.conf some time ago
<ssvb> Turl: the use of g2d should be enabled by default now
<Faisal_> ssvb: yes /dev/g2d is set at the moment for root only RW, is it ok to change it to 777 ?
<ssvb> Faisal_: you can check for G2D messages in /var/log/Xorg.0.log (X server runs as root and is not affected by restrictive permissions)
<Faisal_> ssvb: cubie@Cubian:/dev$ sudo less /var/log/Xorg.0.log | grep G2D
<Faisal_> [ 43.805] (II) FBTURBO(0): enabled G2D acceleration
<ssvb> Faisal_: looks good
<Faisal_> so what could be wrong i havent check that governor thing
<ssvb> Faisal_: try "echo performance > /sys/devices/system/cpu/cpu0/cpufreq/scaling_governor"
<ssvb> Faisal_: or check what are the current settings by reading it
<oliv3r> g2d permission
<oliv3r> !!
<ssvb> oliv3r: for libvdpau-sunxi?
<Faisal_> ssvb: i dont have the scaling_governor file, i have a stats folder after cpu_freq
<Faisal_> oops my bad
<Faisal_> i have it
<Faisal_> it says fantasy, wtf is that
<Faisal_> some sort of policy for the frequency
<Faisal_> i guess
<ssvb> it's the cpufreq governor from allwinner
<ssvb> but anyway, just change it to "performance" and the system is going to be a lot more responsive
<Faisal_> oliv3r: what you mean by "g2d permission !!" ssvb said it was fine
<Faisal_> ssvb: sounds good do i need to refresh or restart something once i change that ?
<oliv3r> mine isn't
<ssvb> Faisal_: the changes should take effect immediately
<ssvb> other than that, what is the resolution of your monitor? and what kind of hardware are you using?
<oliv3r> ssvb: 160 fps; 40:40 cpu load
<oliv3r> 38:38
FreezingCold has joined #linux-sunxi
<Faisal_> ssvb: alright done, res is 1980x1080 thats a tv im using i only got that monitor
<Faisal_> and cubietruck
jinzo has quit [Quit: Leaving]
lioka has joined #linux-sunxi
<Faisal_> just for testing i chmoded 777 dev/g2d and still the same
<Faisal_> g2daccel=no
<Faisal_> oh i restarted x
<Faisal_> and it worked
<Faisal_> g2d_accel=yes
<Faisal_> let me see if the changes are persistant after a reboot
<ssvb> the chmod changes to dev/g2d are not persistent
<Faisal_> nice catch oliv3r
<ssvb> you need to use udev rules for that
<Faisal_> ssvb: oh snap :(
<Faisal_> udevwhat
<Faisal_> (googling)
<ssvb> Faisal_: the stuff which is done to for ump and mali at http://linux-sunxi.org/Binary_drivers
<oliv3r> libv: can't seem to improve it; 40:40
<Faisal_> ssvb: great
<ssvb> Faisal_: the same can be done for g2d, but in any case it has no real effect for the X server
<oliv3r> ssvb: the 2gb bug will bite me
<ssvb> oliv3r: what kernel are you using?
<Faisal_> ssvb: i will do it, do you know what file i should edit ? 50-mali.rules or 50-sunxi-video-rules
<ssvb> Faisal_: any name is fine for *.rules file
<Faisal_> ssvb: ok, i have added a line with KERNEL=="g2d", MODE="0660", GROUP="video"
<Faisal_> i hope that fine
<Faisal_> im part of the video group already
loreta has joined #linux-sunxi
<Faisal_> but i am not sur if the string "g2d" is what i should have used for sure
<Faisal_> ok so now even after i reboot i get the g2d accel
<Faisal_> doesnt feel like it but thats a start
<ssvb> Faisal_: the use of G2D is limited to just acceleration of moving and scrolling windows
<Faisal_> ssvb: oh ok, let me try if i find scrolling better
<oliv3r> ssvb: 3.4.75
loreta has quit [Remote host closed the connection]
<ssvb> Faisal_: and don't forget about the cpufreq governor, it is likely reset back to 'fantasy' after reboot
<oliv3r> i was running es2gears; went fine for 3 or 4 mintes, then i got auto-logged out (crash i guess)
<oliv3r> and the login screen kept crashing every 5 or so seconds
<oliv3r> i saw in the console some ump messages
<Faisal_> ssvb: yes, scrolling is much smoother now thats right
<Faisal_> ssvb: ok let me check its value now
<Faisal_> ssvb: yes, snap. it went back to fantasy, how can i make that a fixed thing, you got an idea ?
<oliv3r> hmm just ump; new session opened
<oliv3r> libv: what could cause a simple terminal resize to cause a crash?
<oliv3r> composition from the WM?
<Seppoz> is btrfs stable yet? or still experimental?
<oliv3r> experimental
<Seppoz> :(
<Seppoz> its been years...
<Turl> Faisal_: you can add the line to /etc/rc.local or whatever your distro uses
<oliv3r> well a FS is something you rely your life on
<Turl> Seppoz: you can use zfs :)
<Seppoz> never seen it :D
<oliv3r> libv: whenever i move my mouse things crash!
<Seppoz> does uboot support it? lol
<ssvb> oliv3r: is it a 2gb cubietruck without the 2gb support fix for ump?
<Faisal_> Turl: do you know what i could enter there because i have no clue
<Turl> Seppoz: lundman had added some zfs support to uboot I think
<Turl> Faisal_: it's a shell script, so whatever you typed on the terminal to change it the other time
<Faisal_> Turl: sounds good
<oliv3r> ssvb: i'm guessing; i'm running stage/sunxi-3.4
<oliv3r> but i don't even have the mouse plugged in and it crashes
<ssvb> oliv3r: this does not look very good
<ssvb> oliv3r: do you have the same issues with the proprietary mali driver?
<oliv3r> never had that installed :p
<oliv3r> i DO have libmali.so for the shader ocmpilaer
<oliv3r> actually nvm i'm talking poop
<oliv3r> this is the plain xfce4 desktop that's crashing
<oliv3r> randomly
<oliv3r> witout running any 3d stuff
<oliv3r> i do see ump session opened messages
<oliv3r> i did chmod 0 /dev/ump; but still crashing
<oliv3r> i did do a no_mali_reserve
<ssvb> oliv3r: hmm, "no_mali_reserve" is a generally bad option which can cause crashes, but it should be already ignored in stage/sunxi-3.4
<oliv3r> i can remove it; just in case
<oliv3r> but mali isn't even used with a plain composited desktop, right?
<oliv3r> and it worked fine for hours
<oliv3r> as soon as i touch my mouse; bam, even lightdm crashes
<ssvb> composited desktop?
<ssvb> oliv3r: if you have some sort of a smartass compositing window manager, then it might try to use gles acceleration for compositing windows
<oliv3r> lightdm +_ xfce4
<oliv3r> but lightdm shouldn't have any of that
<oliv3r> and lightdm crashes when i move the mouse
<oliv3r> i can keyboard login
<oliv3r> but it crashes after a while
<oliv3r> mouse movement guarantees
<oliv3r> a crash
<ssvb> oliv3r: can you remove (or rename) lima/mali libraries and try again?
<oliv3r> lima or ump
<oliv3r> how is lima used by lightdm?
<ssvb> we don't know, but if something is trying to use EGL/GLES under the hood, then it could possibly explain problems
<oliv3r> accelarated mouse?
<oliv3r> lsof | grep lima!
ganbold_ has quit [Remote host closed the connection]
<oliv3r> sunxi_g2d_mem_reserve=64 sunxi_ve_mem_reserve=128 sunxi_fb_mem_reserve=16
<oliv3r> that enough?
<oliv3r> btw, the kernel doesn't crash; lightdm restarts; but thats' it
<ssvb> "sunxi_g2d_mem_reserve=0 sunxi_ve_mem_reserve=80" would be better
<oliv3r> 0 mem for g2d?
<oliv3r> and what's wrong with 128 for ve :p i got tons of ram!
<oliv3r> limare lima_dri .so moved to tmp fixes all
<oliv3r> sleep time now
<ssvb> yes, 0 is better for g2d, because afaik nobody using the custom g2d memory allocator
<oliv3r> ah ok
<oliv3r> i'll test tomorrow iwth liblimare back
<ssvb> and for VE, we only have 256MB for it in the lower addresses of physical RAM
<ssvb> and we really don't want to cross this 256MB boundary for the VE memory reservation :)
<ssvb> but stage/sunxi-3.4 already should use CMA, and the sunxi_ve_mem_reserve option is supposed to be ignored there
<Faisal_> Turl: perfect i've added : cat /sys/devices/system/cpu/cpu0/cpufreq/scaling_governor | sed 's/fantasy/performance/' > /sys/devices/system/cpu/cpu0/cpufreq/scaling_governor
<Faisal_> i actually had to figure that one out
<ssvb> Faisal_: simple echo did not work?
<Faisal_> ssvb: i didnt know i could do it with echo, im just starting this whole thing lol
<Faisal_> but i tought echo wouldnt replace
deasy has quit [Quit: Nom d'un quark, c'est Edmonton !]
tomboy64 has quit [Ping timeout: 252 seconds]
Faisal_ has quit [Quit: quit]
Faisal has joined #linux-sunxi
tomboy64 has joined #linux-sunxi
uwe_ has quit [Ping timeout: 252 seconds]
kivutar has quit [Ping timeout: 272 seconds]
paulk-collins has quit [Quit: Ex-Chat]
uwe_ has joined #linux-sunxi
AreaScout has quit []
iamfrankenstein has joined #linux-sunxi