mnemoc changed the topic of #arm-netbook to: EOMA: Embedded Open Modular Architecture - Don't ask to ask. Just ask! - http://elinux.org/Embedded_Open_Modular_Architecture/EOMA-68 - ML arm-netbook@lists.phcomp.co.uk - Logs http://ibot.rikers.org/%23arm-netbook or http://irclog.whitequark.org/arm-netbook/ - http://rhombus-tech.net/
tuliom has quit [Quit: Konversation terminated!]
popolon has quit [Quit: Quitte]
<libv> has arm released the kernel code for the t604 already?
myfluxi has joined #arm-netbook
tinti has joined #arm-netbook
<ZaEarl> libv, shouldn't Google's Chromebook include it?
<libv> which is why i am asking :)
<ZaEarl> they're shipping chromebooks, right? they must have published the gpl sources, right?
<libv> ah, yes, available from their site
<libv> now i see it... listed as "... Android Kernel Device Driver Sources"
<libv> that has the t60x sources listed
<libv> smart.
freakazoid0223 has joined #arm-netbook
<ZaEarl> but where are the chromebook sources?
<libv> rz2k: are there any egl/gles1/gles2 headers that match the mali blobs from arm?
<libv> ah, there are.
<libv> seems like they must be distributed with these blobs
<ZaEarl> Steam beta starting. 2012 is the year of the Linux desktop.
<projectgus> ZaEarl: if there is going to be a year of linux on the desktop, it makes sense it's a beta and less than 2 months long :P
<ZaEarl> oh, so close: http://i.imgur.com/UQaGf.jpg
<lundman> i have steam beta. but it is for osx so probably doesnt carry across :)
Dssence has quit [Ping timeout: 260 seconds]
Dssence has joined #arm-netbook
alcides has quit [Quit: /(bb|[^b]{2})/ regular expression junkie + lover of literature...]
bsdfox has joined #arm-netbook
bsdfox has quit [Changing host]
bsdfox has joined #arm-netbook
bsdfox has quit [Write error: Connection reset by peer]
bsdfox has joined #arm-netbook
bsdfox has quit [Changing host]
bsdfox has joined #arm-netbook
stefanro has joined #arm-netbook
bsdfox has quit [Read error: Connection reset by peer]
stefanro1 has quit [Ping timeout: 244 seconds]
bsdfox has joined #arm-netbook
bsdfox has joined #arm-netbook
bsdfox has quit [Changing host]
bsdfox has quit [Read error: Connection reset by peer]
bsdfox has joined #arm-netbook
bsdfox has quit [Changing host]
bsdfox has joined #arm-netbook
bsdfox has quit [Read error: Connection reset by peer]
bsdfox has joined #arm-netbook
bsdfox has quit [Changing host]
bsdfox has joined #arm-netbook
bsdfox has quit [Read error: Connection reset by peer]
bsdfox has joined #arm-netbook
bsdfox_ has joined #arm-netbook
bsdfox_ has quit [Changing host]
bsdfox_ has joined #arm-netbook
bsdfox has quit [Ping timeout: 252 seconds]
bfree has quit [Ping timeout: 252 seconds]
bsdfox has joined #arm-netbook
bsdfox has quit [Changing host]
bsdfox has joined #arm-netbook
bfree has joined #arm-netbook
bsdfox_ has quit [Ping timeout: 252 seconds]
cheng has joined #arm-netbook
bsdfox_ has joined #arm-netbook
bsdfox_ has quit [Changing host]
bsdfox_ has joined #arm-netbook
bsdfox has quit [Ping timeout: 240 seconds]
tinti has quit [Ping timeout: 246 seconds]
bsdfox_ has quit [Read error: Connection reset by peer]
bsdfox has joined #arm-netbook
bsdfox has quit [Changing host]
bsdfox has joined #arm-netbook
bsdfox_ has joined #arm-netbook
bsdfox_ has quit [Changing host]
bsdfox_ has joined #arm-netbook
bsdfox has quit [Ping timeout: 255 seconds]
mSquare has joined #arm-netbook
bsdfox has joined #arm-netbook
bsdfox has quit [Changing host]
bsdfox has joined #arm-netbook
bsdfox_ has quit [Ping timeout: 246 seconds]
bsdfox has quit [Read error: Connection reset by peer]
bsdfox has joined #arm-netbook
RITRedbeard_ has joined #arm-netbook
benjamin__ has quit [Ping timeout: 268 seconds]
bsdfox has quit [Read error: Connection reset by peer]
bsdfox has joined #arm-netbook
bsdfox has quit [Changing host]
bsdfox has joined #arm-netbook
bsdfox has quit [Write error: Connection reset by peer]
bsdfox has joined #arm-netbook
bsdfox has quit [Changing host]
bsdfox has joined #arm-netbook
slash_random has quit [Ping timeout: 240 seconds]
bsdfox has quit [Read error: Connection reset by peer]
cheng has quit [Quit: Leaving]
bsdfox has joined #arm-netbook
bsdfox has quit [Read error: Connection reset by peer]
bsdfox has joined #arm-netbook
kop has quit [Ping timeout: 260 seconds]
tekzilla has quit [Ping timeout: 255 seconds]
tekzilla has joined #arm-netbook
kop has joined #arm-netbook
wingrime has joined #arm-netbook
freakazoid0223 has quit [Ping timeout: 240 seconds]
freakazoid0223 has joined #arm-netbook
<Turl> mnemoc: I'm really late to the conversation, but poor mips :(
<Turl> mnemoc: I'm learning MIPS asm, way less braindead than x86 :)
Kraln has quit [Quit: No Ping reply in 180 seconds.]
Kraln has joined #arm-netbook
freakazoid0223 has quit [Ping timeout: 240 seconds]
merbanan has quit [Ping timeout: 260 seconds]
freakazoid0223 has joined #arm-netbook
sky770 has joined #arm-netbook
<sky770> Hi all
<sky770> any news on the Cedar X VPU front?
* sky770 wonders if anyone's home..? :|
<lundman> no.
<sky770> BAHAHA..very funny lundman (lmao nice nick :D )
<sky770> anyone having hardware acceleration on ubuntu/lubuntu with full X on ?
<sky770> am talking about 1080p GPU HW acc. :D
<lundman> what the
<lundman> its my name! :)
* sky770 facepalm :D
<lundman> I think everyone is voting, or something
* sky770 feels zzzZZZ..over here
eFfeM has joined #arm-netbook
<sky770> hey effem
RITRedbeard_ has quit [Ping timeout: 246 seconds]
<sky770> yo lundman
<sky770> where you from?
<sv> sky770, last i remember reading, armhf libs were released for cedarx
<sv> not sure if work was put into it yet though
<lundman> like, originally? long time ago? sweden
<sky770> Yeah Sv, even I have had a look at the github pages
<sv> i,ve tried to find out more
<sv> but the infos are scarce
<sky770> yep :|
<sky770> I thought someone around here might have a clue/fix on that..so.. :D
<sky770> Oh sweden! lundman that's a pretty country you're from :)
<lundman> not after I left it!
<lundman> hah
<sky770> Sv, I suppose guys at Pivos DS have figured it out atleast for their XBMC coz they are using the A10 boxpack :)
<sky770> lundamn meeh :D
<sv> yeah i saw that
<sky770> lundman meeh :D
eFfeM has quit [Quit: Leaving.]
<sky770> Its all about $$ nowadays.. damn :(
<lundman> A10 boxpack?
<sky770> the Allwinner tech's A10 SoC :) a cortex A8 based SoC
<sky770> bahaha :D its the chinese SoC crap that got us here :D
<sky770> oh btw' amlogic's 8726 came up pretty good but now the chinese firms are aware of the open source drivers/src that amlogic gave up and thats a threat to the ultra cheap allwinner's A10 SoC
sky770 has quit [Ping timeout: 252 seconds]
rellla has joined #arm-netbook
<hno> XMBC is operational. So is MALI but not entirely stable. Nothing hidden or secret. http://linux-sunxi.org/XBMC http://linux-sunxi.org/CedarX http://linux-sunxi.org/Mali400
Entropy512 has quit [Quit: No Ping reply in 180 seconds.]
Entropy512 has joined #arm-netbook
rellla has quit [Ping timeout: 256 seconds]
rellla has joined #arm-netbook
Quarx has joined #arm-netbook
Quarx has quit [Read error: Connection reset by peer]
Quarx has joined #arm-netbook
rellla has quit [Quit: rellla]
rellla has joined #arm-netbook
cheng has joined #arm-netbook
pawel5870 has joined #arm-netbook
andoma__ is now known as andoma
cat_x301 has joined #arm-netbook
wingrime has quit [Ping timeout: 255 seconds]
<hno> Yay, even VLC is operational now with CedarX. https://groups.google.com/d/msg/linux-sunxi/iS2RFgLI3JY/OkGfC-o3sHsJ
<rm> I didn't pay much attention, is that under Xorg?
<orly_owl> progress! \o/
<andoma> great
<hno> rm, directfb
<cheng> how about mplayer, can we use that with CedarX?
<hno> cheng, not done yet, but should be easy based on either xmbc or vlc integration.
<cheng> that's good, hope to look into that in near future
<mnemoc> libva please
<mnemoc> then every player will be able to use it
popolon has joined #arm-netbook
<sv> hi
<sv> yess
pawel5870 has quit [Ping timeout: 252 seconds]
pawel5870 has joined #arm-netbook
tinti has joined #arm-netbook
bsdfox has quit [Ping timeout: 260 seconds]
graffiti has joined #arm-netbook
<graffiti> rellla: are you able to complile xbmc with the latest git?
<rellla> Graffiti: my last compile succeeded. iirc, i used latest git. do you have any issues?
<mnemoc> rellla: can you take a look at https://github.com/willswang/libcedarx ? maybe we can male xbmc and vlc and others share a common base
<mnemoc> make*
<rellla> mnemoc: saw it already. this could be interesting for empat0 - he's xbmca10-man. hopefully he noticed this repo. next i try to compile willswangs wrapper.
<mnemoc> it would be also awesome if that libcedarx could be refactored to not include the closed libs
<mnemoc> but the author ignred my request :|
<mnemoc> a git submodule maybe
<mnemoc> but it's kind of risky to include files with unclear license in open source trees
<rellla> this libcedarx is only a wrapper around the closed binarys, right? so nothing new on open driver front? so whats the alternative to closed libs at this time?
<mnemoc> we have no alternative to closed libs
<mnemoc> I only mean to decouple the free code from the closed libs, repo-wise
<mnemoc> even if we inevitably need to link against the closed libs
<andoma> cedarx is just a memory allocator
<andoma> strictly it's not even needed
<mnemoc> the last libs use a different allocator lib
<andoma> libve (which is what's closed) only needs a few function pointer for it to help allocate memory and communicate with hardware
<slapin_> mnemoc: how it is even possible to get non-android libs for video? is there some connections with AllWinner?
tinti has quit [Ping timeout: 245 seconds]
<mnemoc> slapin_: yes, hipboi
<slapin_> that's really awesome
<mnemoc> it's not easy, but we have debian/ubuntu armel and armhf cedarx libs thanks to him
<rellla> ok. so new libcedarx should better refer to /use closed libs in linux-sunxi/cedar-libs tree and binarys can be sorted out of that repo. and whats the main goal of this wrapper? creating a wrapping-library that all apps (vlc, xbmc, mplayer ...) can base on?
<mnemoc> rellla: having a good-api wrapping library around the closed libs is a good thing... but ideally it should be made in the form for a libva driver
<mnemoc> and taking the closed libraries from an external/submodule
<rellla> because most apps can handle libva already?
<mnemoc> yes
<mnemoc> it's also called "VA API"
<slapin_> what is libva?
<slapin_> aaaaa
<slapin_> ah
<slapin_> thanks
<slapin_> I remember...
<mnemoc> omx made right :p
<rellla> anybody here, knowing or using http://www.tvdr.de? <-can stream to xbmc
<rellla> it would be awesome, to get vdr running on a10. but there's no output device for a10 atm.
cheng has quit [Quit: Leaving]
rz2k has quit [Read error: Connection reset by peer]
<Maqs> rellla: vdr + xbmc (with pvr addons) should work
rz2k has joined #arm-netbook
<rellla> Maqs: yes, i streamed already ;-) but for me, xbmc is too oversized. i don't need music,video,scraping etc.
<Maqs> tvheadend and showtime, once andoma releases it to the public ;-)
<Maqs> that is.. the a10-version of showtime..
<rellla> Maqs: there even are some successful attemps to get vdr running on rpi - without xbmc.
<andoma> i think tvheadend should run fine on rpi
<Maqs> i had tvheadend + xbmc on the pi
<andoma> don't see why it would not work on a10 out of the box
<Maqs> andoma: i mean showtime ;)
<andoma> Yeah that still WIP :)
<Maqs> i don't want xbmc, as it is just to big, slow etc.
<Maqs> btw, is there a trick to get showtime working with tvheadend on another host?
<andoma> if you have avahi installed they will find each other
<andoma> otherwise you can just type htsp://ip.address.of.tvheadend in showtime's search box
<andoma> and once opened you can bookmark it
<andoma> lunch .. bbl
<rellla> didn't use tvheadend nor showtime. tvheadend<-backend, showtime<-frontend like xbmc, but mor lightweight?
<rellla> right?
tzafrir_laptop has quit [Ping timeout: 245 seconds]
<Maqs> thanks
<Maqs> rellla: right
Almamuetya10 has joined #arm-netbook
<Maqs> and it supports anything else that's important for me.. and, surprise, superb spotify integration
tinti has joined #arm-netbook
graffiti has quit [Quit: Page closed]
sv is now known as sv
slash_random has joined #arm-netbook
tekzilla has left #arm-netbook [#arm-netbook]
merbanan has joined #arm-netbook
mSquare has left #arm-netbook [#arm-netbook]
focus_well has joined #arm-netbook
ssvb has quit [Ping timeout: 255 seconds]
ssvb has joined #arm-netbook
<hno> stefanro, ok, PHY power mystery solved. It's always powered on Cubieboard so no need for powering it on via GPIO there. But needed on other boards.
<lundman> ah interesting
<mnemoc> hno: got corrected schematics? or only that particular answer?
<mnemoc> ah, ok. saw the reply
<stefanro> hno: thx for the notice
tzafrir_laptop has joined #arm-netbook
<hno> Hackberry need PH19 set. Mele another pin. Now just need to figure out best place for that info.
<mnemoc> boards.cfg ?
<hno> Yes that's a primary candidate.
<lundman> isnt it in script.bin?
<mnemoc> and when defined empty meaning, has ethernet but no phy power
<lundman> guess you dont want uboot to read it
<mnemoc> lundman: script.bin is not read by uboot
<mnemoc> and we need to fix the kernel driver to not ignore the phy power gpio
<mnemoc> if the field is set, it has to be used
<hno> The kernel is ignoring it?
<Turl> PHY_POWER is defined to 0 iirc
<mnemoc> exactly
<mnemoc> blame lundman :p
<hno> So just set that to 1. iirc the code should handle blank value correct.
<mnemoc> that code needs to be refactored to use the existance of of emac_power as reference
<hno> From what I can tell it should just work. But might want to silent an KERN_ERR printk when not set or blank.
<mnemoc> maybe freeing that memory of mos_pin_handler wasn't defined
<mnemoc> s/of/if/
<ibot> mnemoc meant: maybe freeing that memory if mos_pin_handler wasn't defined
<mnemoc> will try to find time to do that change and test it today. but if someone else can, much better.
<hno> It's never freed at all. And you meant L1796 I think.
<mnemoc> true, chrome is showing me the page f*ed up
<mnemoc> I have line one at #include <linux/io.h>
eFfeM has joined #arm-netbook
mikey_w has quit [Ping timeout: 244 seconds]
<hno> Gah.. why is script_parser_fetch using an (int *) for the result?
<mnemoc> it's an extremely awful API. I made an alternative in 3.4 but still unhappy
eFfeM has quit [Client Quit]
<Turl> script.bin is all f'ed up :<
<Turl> kernel module parameters could be used to replace them
tzafrir_laptop has quit [Ping timeout: 240 seconds]
<libv> are there any debian packaging wizards around?
<libv> because it might make sense to have debian/ trees for some of our bits
<mnemoc> http://packages.linux-sunxi.org/ still looks for maintainer
<libv> (it's been 5ys since i last dealt with it myself)
<Turl> libv: I know a thing or two, haven't touched it for like one or two years though
<Turl> what bits would you want to package?
<mnemoc> x11 stuff is specially important
<libv> currently looking at the mali binaries
<libv> as they are missing headers
<libv> which i am about to add
<mnemoc> libv: we have "read to use" git repos for those
<mnemoc> with headers, patches, and stuff
<libv> mnemoc: gl headers?
<mnemoc> made by rz2k
<mnemoc> i believe so, but haven't looked myself
<libv> not in mali-libs
<mnemoc> there is a libump repo, xf86-video-mali and another mali one I don't remember
<mnemoc> rz2k knows
<libv> libump does not contain, for instance, egl.h
ZaEarl has quit [Ping timeout: 256 seconds]
<libv> anyway, i am experimenting around a bit as to how i can get some of the lima tests building and running on mali libs + fb
<libv> i will manually install these headers (which seem pretty universal - as ARM does not provide them with their own GL extensions)
<libv> this to verify whether there is a viable path that way
<libv> if so, then others definitely will want to have these headers installed as well
<mnemoc> shouldn't we add these headers to the mali-libs repo?
<libv> yes, they should be added there
<mnemoc> you can push access ;-)
<libv> together with a Makefile to install things properly into the os
<mnemoc> but try to keep the structure, so we don't kill the integration the xbmc people is using
<libv> the best solution also adds debian packaging on top
<mnemoc> and .rpm probably :p
<mnemoc> but that is another story
<mnemoc> there are some fedora/opensuse fellows around
<Turl> rpm still exists? :P
<Turl> j/k, not wanting to start a flamewar :)
<libv> with all .debs created per architecture, with each providing a general target which the -devel package (egl/gl headers) depending on the general target
<libv> i do not know anything about rpm, even though i worked at suse for a bit :)
<libv> and even though i use suse as my main desktop
<mnemoc> 1) Makefile and .h in mali-libs, 2) .deb, 3) ...
<libv> but yeah, if people are willing to do the work, of course
<libv> yup
<libv> but let me first get to the stage where some lima egl tests are running
tzafrir_laptop has joined #arm-netbook
kaspter has joined #arm-netbook
<Turl> this AN7G3 GPL release has more goodix touch drivers
<Turl> and the ctp_platform_ops.h is different
QingPei has joined #arm-netbook
pawel5870 has quit [Ping timeout: 252 seconds]
arnd_ has joined #arm-netbook
gsilvis has quit [Read error: Connection reset by peer]
gsilvis has joined #arm-netbook
<mnemoc> easy way it to reset to that commit, branch, rm -rf *, and then copy over the new files
<Turl> mnemoc: want me to reset, commit all the new files and push to github so we can all compare?
<mnemoc> almost
<mnemoc> do it per-driver and sanitized if possible
<Turl> sanitized?
kaspter has quit [Read error: Connection reset by peer]
<mnemoc> you will see sun5i stuff removed for example
<mnemoc> we don't want that part in the commit
<mnemoc> only relevant changes
<Turl> ah, ok
<mnemoc> also, if you can do re-encoding, you might prefer 7ea850073ec6c1f91047676723d21677bce82eba as base
<Turl> hm, sprunge broke :< didn't return any url
<mnemoc> too large
QingPei has quit [Ping timeout: 256 seconds]
<mnemoc> Turl: http://sprunge.us/dNGh?sh is my sanitizer script
<mnemoc> fixes permissions, encoding and trailing whitespace
<mnemoc> modules/gps/ !
<mnemoc> we didn't have a gps .ko for 3.0
<mnemoc> btw, you might need to clear .gitignore
<Turl> "gps.ko_"
<mnemoc> yes, gpl violating as usual
<Turl> $ strings modules/gps/gps.ko_ |grep magic
<Turl> vrrmagic=3.0.8+ preempt mod_unload modversions ARMv7
<Turl> vrrmagic?
QingPei has joined #arm-netbook
<mnemoc> we then need to do the same for the A10s SDK :<
<mnemoc> use 7ea850073ec6c1f91047676723d21677bce82eba as base, then apply http://sprunge.us/dNGh?sh to all files, and then per-driver commits
<pucko> wat? there is a sdk for a10/mele?
<mnemoc> A10s != A10
<mnemoc> if you notice renames (like the for the gpio driver) they go in the same commit
<mnemoc> but be sure to empty .gitignore
<mnemoc> they include files normally ignored in the SDKs
<mnemoc> so we would miss them
<mnemoc> like .bin or .fw
<mnemoc> or .ko in the case of the gps or cedarx
<mnemoc> or .o in mach-sun5i :'(
<mnemoc> Turl: can you paste arch/arm/mach-sun4i/include/mach/gps.h ?
<Turl> mnemoc: sure, sec
<Turl> I got reply from gandi btw, paste in /query
<mnemoc> uh
<mnemoc> this people is weird
<Turl> tom cubie?
<Turl> lol
<Turl> eh
<Turl> the header is the gps module? wtf
<mnemoc> nah, only the gpio parts
<Turl> half of it rather
<Turl> yeah
<mnemoc> if the .ko has symbols someone can try to RE it
<mnemoc> that's why it's important to not get it lost
<Turl> mnemoc: how do you run your sanitizer thingy, find kernel/ -exec ~/sanitize {} \;
<Turl> ?
<mnemoc> or -print0 | xargs -r0 sanitize
<mnemoc> find -type f
<drachensun> turl: I keep getting corrupted objects when I call repo sync, is that normal?
<Turl> drachensun: no
<drachensun> heh
<drachensun> figured, any idea what would it could be?
* Turl never uses xargs :/
<Turl> drachensun: corrupted objects? :)
<mnemoc> Turl: don't know what it can do to non-files
<Turl> mnemoc: good idea
<drachensun> turl: error: inflate: data stream error (incorrect data check)%
<mnemoc> i'm not a fan of -exec :p
<Turl> drachensun: maybe your intertubes is corrupting/has corrupted data
<drachensun> turl: so it seems but its happening over and over
<drachensun> turl: ugh
<drachensun> turl: I'm not sure how to clear it either, once its downloaded a bad one it just insist on trying to unpack it again and again
<Turl> drachensun: remove it
<Turl> mnemoc: sanitizing all the things!
<Turl> mnemoc: sanitizr: 94: enca: not found
<Turl> missing deps? :P
<drachensun> turl: Ok,its complaining about the repo version to, do I need the latest version? I was having trouble finding it
<Turl> repo version?
<mnemoc> Turl: sudo apt-get install enca :)
<Turl> repo is self updating
<Turl> mnemoc: done, resanitizing all the things :)
<drachensun> turl: thats what I thought, but it keeps saying "... A new repo command ( 1.18) is available.... You should upgrade soon:"
QingPei has quit [Quit: Leaving.]
<RaYmAn> it tells you how to upgrade
<RaYmAn> just do the command it says :P
<Turl> drachensun: there's a cp command just under the message
<Turl> the bsp runs it after a successful sync too so you shouldn't worry there
<drachensun> so there is
<drachensun> I'm obviously doing too many things in parallel here ;-)
<mnemoc> Turl: did you like my toy? :)
<Turl> mnemoc: `sanitize' ? :P it's still working
<mnemoc> :)
<Turl> mnemoc: reimplement it on assembler so it's faster
<Turl> :P
<mnemoc> :)
jelly has quit [Quit: Client exiting]
<Turl> mnemoc: there's sun5i cedar code on there
<mnemoc> the fun part comes after that
<mnemoc> let see how it compares to the code we had
<mnemoc> that base commit doesn't include anything from the A13 SDK
<mnemoc> so the leave the sun5i part until the end
<Turl> mnemoc: how does your script behave wrt upstream code?
<Turl> does it sanitize too?
<mnemoc> no
<mnemoc> it only removes trailing whitespace from files with weird encoding
<mnemoc> upstream files should be unaffected by the enc, perm and crlf filters
<Turl> uhm mnemoc
<Turl> it's sanitizing .git :<
<Turl> I hope it doesn't corrupt it
<mnemoc> it might...
<mnemoc> you was *your* find call, not my poor little sanitizer
<Turl> but your sanitizer didn't have a sanity check :<
<Turl> git status works, good
<mnemoc> finished?
<Turl> let me check out all the deleted sun5i stuff so I can start diffing
<mnemoc> git diff Makefile ?
* hno is a fedora fellow
<Turl> mnemoc: vendor hacked arch and crosscompiler for their OS drop :< http://paste.ubuntu.com/1340037/
<mnemoc> doh
<mnemoc> Turl: those should be reverted too
<mnemoc> weird they deleted all .gitignore files
<mnemoc> or that was you? :)
<Turl> I rm'd all the tree and copied theirs on top
<mnemoc> but theirs didn't come with .gitignores
<mnemoc> so they deleted them
<mnemoc> keep them deleted
<mnemoc> so you can see all the files
<mnemoc> just don't commit
<mnemoc> many changes can be headers-only
<Turl> new file -> drivers/misc/sunxi-reg.c
<Turl> :P
<mnemoc> we have that one from A13 iirc
<Turl> sun5i audio has changes
<mnemoc> leave sun5i to the end
<mnemoc> we need to cross diff with A13 and A10s SDKs :<
<Turl> just sanitized whitespace mnemoc :<
<Turl> git diff -w sound/soc/sun5i/ is empty
<mnemoc> revert all the noise
<mnemoc> and add all files
<Turl> yeah I'm doing that
<mnemoc> :)
<mnemoc> hno: that's why I mentioned it ;-)
<Turl> okay, no sun5i changes
<Turl> at least not on files named *sun5i*
<mnemoc> good
<Turl> docs is noise
<mnemoc> as expected
<Turl> mnemoc: hack level: allwinner http://paste.ubuntu.com/1340069/
tinti has quit [Ping timeout: 252 seconds]
<mnemoc> god
tinti has joined #arm-netbook
<mnemoc> Turl: evil hacks to core mainline code in a separated commit please
<Turl> any hint on where to start mnemoc? http://paste.ubuntu.com/1340087/
<mnemoc> with the drives with new files maybe
<mnemoc> gps, sunxi-reg, sunxi-ps2, gpio
<mnemoc> to reduce the noise in `git status`
<mnemoc> but be sure all .gitignore files are empty
<mnemoc> there are tons of changes within scripts/ which are probably just noise
<mnemoc> usi-bcm4329 is mostly blob-only
<mnemoc> same for nano
<mnemoc> those are easy commits
pawel5870 has joined #arm-netbook
arnd_ has quit [Ping timeout: 260 seconds]
cat_x3011 has joined #arm-netbook
<libv> urgh, crap, nfsv3 and nfsv4 are not compiled in
<mnemoc> defconfig patches welcomed ;-)
cat_x301 has quit [Ping timeout: 256 seconds]
<libv> mnemoc: yeahyeah :p
<Turl> mnemoc: +obj-$(CONFIG_TOUCHSCREEN_FT5X06_TS_2PLUS1) += ft5x06_ts_2plus1.o
<Turl> the file doesn't exist however
<Turl> :P
<mnemoc> typical :<
<Turl> mnemoc: should I commit it?
<mnemoc> it's only that line? then I would say no
<Turl> mnemoc: + kconfig
<mnemoc> -1
<mnemoc> useless
<mnemoc> and the Kconfig part doesn't tell anything we can't add ourselves if we eventually get the driver
<Turl> + Say Y here if you have Softwinner University Program PS/2 ports.
<Turl> wtf is a softwinner university program PS/2 o.O
<mnemoc> i've seen that before
<mnemoc> maybe in the 2.6.36 tree
<mnemoc> but the driver gives us registers, and that is a good thing to saveguard
<Turl> MODULE_AUTHOR("Aaron.maoye<leafy.myeh@allwinnertech.com");
<Turl> MODULE_DESCRIPTION("SW F3X PS2 controller driver");
<Turl> guy derped his author tag >.<
<techn> :)
<Turl> I'll resume comitting after lunch :)
<mnemoc> should have been wip :p
<mnemoc> but forgiven, for now
slapin has joined #arm-netbook
<Turl> mnemoc: I copied name to match other lichee branch :<
<mnemoc> this one isn't a normal lichee branch
<mnemoc> it has sun5i
<mnemoc> it's more of a unified lichee3-sunxi
<mnemoc> lichee-3.0.8-sun4i had history
projectgus has quit [Ping timeout: 248 seconds]
slapin has quit [Ping timeout: 276 seconds]
bsdfox has joined #arm-netbook
cat_x3011 has quit [Read error: Connection reset by peer]
slapin has joined #arm-netbook
<mnemoc> Turl: :D
projectgus has joined #arm-netbook
<mnemoc> Turl: btw, try git diff --check before commiting :)
<mnemoc> sanitize -f will remove trailing white space for you
<mnemoc> but don't run it massively
<Turl> mnemoc: I'm a fan of sed 's/\s*$//g'
<mnemoc> good enough :) but the /g is kind of meaningless on expressions with $ ;-)
<mnemoc> you can't have more than one match per line
vinifm has joined #arm-netbook
slash_random has quit [Ping timeout: 240 seconds]
rellla has quit [Remote host closed the connection]
<Turl> mnemoc: can't hurt right? :)
<Turl> mnemoc: I hate when I forget it on other expressions and only replaces the first instance :<
<Turl> mnemoc: sunxi_gpio driver :o
<Turl> mnemoc: looks like an actual proper gpio driver by the name
<mnemoc> Turl: how does it compare with our current crapver?
<mnemoc> ./drivers/misc/sun4i-gpio.c
<mnemoc> it really can't be worse
<rz2k> it atleast are more than one file
<rz2k> :3
<mnemoc> nice :)
<mnemoc> Turl: can you commit it within misc/ ?
<mnemoc> :)
<mnemoc> from MODULE_LICENSE("GPL"); we can hopefully assume they "forgot" the GPL header :)
<Turl> tom can probably find one on his computer
<Turl> :P
gimli has joined #arm-netbook
Quarx has quit []
<mnemoc> Turl: interestingly it doesn't use script.bin info like the old one
<Turl> the less script.bin the better :)
<Turl> mnemoc: tom moved sunxi-dbgreg -> sunxi-reg and changed some code
<Turl> should I do it all in one shot, or commit modifications then change makefiles?
<mnemoc> one shot
<mnemoc> been an import we can't expect fine grained history
<mnemoc> it was this and became that
<mnemoc> why? who knows :|
<mnemoc> when merging into 3.0 it will be different
<mnemoc> but lichee3-sunxi is expected to be messy
<mnemoc> Turl: do you have anything on wemac?
pawel5870 has quit [Ping timeout: 252 seconds]
<Turl> mnemoc: let me check
<Turl> mnemoc: http://paste.ubuntu.com/1340435/ thoughts on axp?
<Turl> mnemoc: net/ is unchanged
jas-hacks has joined #arm-netbook
vinifm has quit [Ping timeout: 246 seconds]
vinifm has joined #arm-netbook
RITRedbeard_ has joined #arm-netbook
rellla has joined #arm-netbook
gzamboni has quit [Read error: Connection reset by peer]
traeak has left #arm-netbook [#arm-netbook]
jas-hacks has left #arm-netbook [#arm-netbook]
rsalveti has quit [Ping timeout: 245 seconds]
<vinifm> which parts of the kernel are being rewritten?
<Turl> vinifm: if you see something sunxi related, it's probably going to be rewritten :P
rsalveti has joined #arm-netbook
eFfeM has joined #arm-netbook
vinifm has quit [Read error: Connection timed out]
vinifm has joined #arm-netbook
<hno> Turl, can you diff the pmu change with a little more context lines please
arnd_ has joined #arm-netbook
focus_it has joined #arm-netbook
<mnemoc> :D
<mnemoc> very allwinner
<mnemoc> something doesn't work, drop an msleep()
arnd_ has quit [Read error: Operation timed out]
<hno> but there?
<mnemoc> you saw the original "sata fix".... he sent down the clock down to like 100MHz for all cases
<mnemoc> thread_cpu_clock_get looks like a neat place to slow everything down :<
vinifm has quit [Read error: Connection timed out]
vinifm has joined #arm-netbook
<mnemoc> vinifm: there is a new gpio driver from allwinner if you want to experiment
<vinifm> what brach?
<mnemoc> it's from a recent source drop
<mnemoc> still untested
bsdfox has quit [Read error: Connection reset by peer]
bsdfox has joined #arm-netbook
<RaYmAn> You kind of have to wonder why they choose to rewrite it, yet not make it standardized *again*
graffiti has joined #arm-netbook
<mnemoc> i have to admit the standard way is annoying as hell
<mnemoc> they want you to name each pin, tell the functions of each, and define all valid combinations
<mnemoc> they don't even do that on the documentation
bsdfox has quit [Read error: Connection reset by peer]
bsdfox has joined #arm-netbook
bsdfox has quit [Changing host]
bsdfox has joined #arm-netbook
<mnemoc> allwinner drivers just accept blindly whatever is thrown at them
axion has joined #arm-netbook
<axion> hi :)
<axion> i'm looking for information on the rockchip kernel
bsdfox has quit [Read error: Connection reset by peer]
bsdfox has joined #arm-netbook
bsdfox has quit [Changing host]
<mnemoc> axion: look at cnxsoft's blog
bsdfox has joined #arm-netbook
benjamin__ has joined #arm-netbook
markatto has joined #arm-netbook
<axion> thanks found it
<rellla> hi kernel guys, how can i check, whether a mali.ko is r3p0 without having he sources?
bsdfox has quit [Ping timeout: 252 seconds]
arnd_ has joined #arm-netbook
graffiti has quit [Ping timeout: 245 seconds]
<rz2k> rellla: I believe that is impossible
<rz2k> we can fix mali to print "mali driver version %s loaded" instead of "mali driver loaded" if that is really needed.
<mnemoc> +1
<mnemoc> I added something similar about the SoC
<mnemoc> [ 0.000000] chip-id: A10 (AW1623 revision C)
<rellla> rz2k: i thought so :-(.... anyone knowing then which mali drivers the linaro images are using?
<rellla> rz2k: maybe worth it
<rellla> s/maybe/may be/
<ibot> rellla meant: rz2k: may be worth it
<rz2k> linaro images?
<mnemoc> rellla: if built from our tree, check the date in /proc/version
<rz2k> we have hwpacks built, but linaro is standard.
<rellla> there's a guy in xbmc forum and has problem starting xbmc. i think kernel-version doesn't fit to binary-version...
<rellla> and he uses empat0's xbmca10 upon a linaro-somthing-image
<rz2k> I never did get xbmc working on linaro with xf86-video-mali, he might be confused using it with xbmc that uses framebuffer libs'
focus_it has quit [Quit: Leaving]
<drachensun> I didn't think you needed the mali userspace libs for xbmc
RITRedbeard has quit [Quit: Leaving]
<drachensun> unless its auto installed by the dependencies and I missed it
RITRedbeard has joined #arm-netbook
<drachensun> well yeah, they are dumped in the that libs directory
<rz2k> you dont need them, but if you have incompatible type in /lib, you will end with glitches
<mnemoc> from the hwpack? they are x11 libs
<techn> mnemoc: You removed framebuffer libs there :(
<techn> I added them there since xbmc guys are using framebuffer ones :p
<mnemoc> I asked you.... never got an answer, assumed it was a bug
<mnemoc> specially because you had a broken symlink in the middle :p
<mnemoc> 1m
tinti has quit [Quit: Leaving]
gzamboni has joined #arm-netbook
<rellla> mnemoc,techn: removed from hwpacks, but still staying in repo. i don't use hwpacks. i more like it to do kernel etc. from scratch ;-)
<techn> rellla: yep.. I'm using hwpack as dev tool
<vinifm> mnemoc: patch -pN, N=?
<mnemoc> 1 normally
<rellla> rz2k: so if he uses x11-libs, xbmc won't start, because it waits for the fb-libs.
<Turl> hno: sure, if you tell me how to specify # of context on git diff
<techn> I mean sunxi-bsp is my dev tool.. easy.. modify code and then make hwpack-install ..
<rellla> techn: maybe i should take a look at it, if it makes life easier
<mnemoc> techn: http://sprunge.us/YgFD ?
<mnemoc> techn: is that the intended result?
<techn> mnemoc: yes.. that I was thinking
<techn> you got then bin backup and framebuffer/x11 support at same time
<mnemoc> techn: that makes it ---^
<techn> yep
<mnemoc> should I commit?
<techn> mnemoc: sure :)
<techn> btw.. Is here anyone who wants to test cubieboard linux livesuit image?
<techn> mnemoc: or did that require different u-boot?
<techn> from lichee branch?
<mnemoc> lichee-dev
<mnemoc> I have a prebuilt in the testing-3.4 dir
<mnemoc> it doesn't support O= so it's not bsp friendly yet
<techn> that works with 3.0 kernel?
<mnemoc> yes
<rm> http://www.phoronix.com/scan.php?page=news_item&px=MTIyMzQ Linux Kernel Ported To The Synopsys ARC700
<rm> interesting
<rm> a whole new ISA and CPUs
<rz2k> rellla: it will fill the screen black and segfault / halt hardly
cat_x301 has joined #arm-netbook
Almamuetya10 has quit [Ping timeout: 246 seconds]
freakazoid0223 has quit [Ping timeout: 246 seconds]
sspiff_ has quit [Ping timeout: 245 seconds]
<rz2k> if somebody wants to poke mali on gentoo/archlinux arm or whatever with X.org 1.13 here is what needs to be fixed to compile mali https://github.com/rzk/xf86-video-mali/commit/93ff10156b9941bd7464e8db85e2d379649b74ff https://github.com/rzk/xf86-video-mali/commit/8a8a9b14e50d2c13653cbbec29bb8170d5d19b75
<rz2k> (this is most likely are not compatible with Linaro's versions, so it will stay away from linux-sunxi branches)
<mnemoc> not pushed to linux-sunxi orga?
<mnemoc> ok
<mnemoc> append .patch and link them from the wiki ;-)
<rellla> what is the main goal of the linaro-images in general? i don't understand the advantages?! i mostly do everyhing on debian...
<mnemoc> we do hwpacks, not images
<mnemoc> upon the hwpack you can add any rootfs
gx260 has joined #arm-netbook
<gx260> Going to compile the kernel for my A13 tablet. What config would be the best to use?
<mnemoc> if android a13_nuclear_defconfig, if linux a13_
<mnemoc> defconfig
tuliom has joined #arm-netbook
<mnemoc> rz2k: can't the code be made support both versions of xorg?
<rz2k> no
<rz2k> they've changed the api
<rellla> mnemoc: possibly i meant linaro-rootfs.
<rellla> many people ever talk about "i use linaro to....", "i have a linaro system", tom compiled the mali libs on "linaro", too. did they only mean the rootfs?
<mnemoc> yes
<rz2k> Tom compiled mali libs on ubuntu-arm
<gx260> Thanks! I tried this one but would not compile. sun4i_defconfig
<mnemoc> sun4i = A10
<mnemoc> rellla: linaro does the arm port of ubuntu
<mnemoc> so ubuntu-arm and linaro/ubuntu are the same
<mnemoc> except the name of the default user
<mnemoc> rz2k: #ifdef ?
<mnemoc> #ifdef SOMETHING_THAT_ONLY_EXISTS_SINCE_1_13
<mnemoc> #else
<mnemoc> #endif
<gx260> What is the option -j4 mean?
<mnemoc> run 4 jobs in parallel
<mnemoc> don't do more than the double of the cores you have
<rz2k> I know, need to find xorg devel defines and if they did change when that thing happen
<rellla> mnemoc: ok understand. but better said "a" arm port of ubuntu?
<gx260> Cool thanks!
<mnemoc> rellla: no. canonical pays linaro to do their ubuntu arm port
<mnemoc> it's not a branch
<vinifm> nowadays a driver for linux runs in android and vice versa?
<mnemoc> no
<rz2k> libhybris tried to solve that, dont know if succeded
<rz2k> not interesting detail: if you place fbdev string in xorg.conf for mali longer than /dev/fb0, it will crash
<Turl> lol
<rz2k> there is a part of awesome code there
<rz2k> like char *dev= /dev/fb0, then dev= string_from_config
<Turl> strcpy(...)
<rz2k> not the
<rz2k> :3
<gx260> I'm getting this error: gcc: /drivers/gpu/mali/mali/arch/config.h: No such file or directory
<mnemoc> Turl: I think I saw EARLYSUSPEND_DELAY on the 2.6.36 tree
<gx260> The file is there?
<mnemoc> gx260: did you download a .zip file?
<rellla> mnemoc: thanks. another day, that lets me be more clever at the end of it ;-)
<rellla> gn
<gx260> No used git
<mnemoc> weird
<gx260> arch in the path is a link.
<mnemoc> using make -C ?
<mnemoc> http://dl.linux-sunxi.org/amery/sunxi-3.0/20121107-2117/ <--- this hwpacks have mali fb and x11 libs
<mnemoc> these*
<gx260> No here is my command line: make ARCH=arm CROSS_COMPILE=arm-linux-gnueabi- -j4 uImage modules
<rz2k> interesting
<mnemoc> looks fine....
<rz2k> what distro do you use?
<gx260> #!
<mnemoc> i'm intriged by the / before drivers
<mnemoc> would have expected the kernel dir there
<gx260> this is the actual first error: /usr/src/linux-sunxi/drivers/gpu/mali/mali/Kbuild:48: "USING_PROFILING not supported, disabling."
<mnemoc> that should be a warning, not an error
<mnemoc> run it without -j4 and paste the real thing :)
<gx260> ok, hang on
<gx260> should I put this on pastbin?
<techn> huaah.. what is the easyest way to make filesystem file from directory?
<mnemoc> there is no easy way
<mnemoc> dd + mkfs + loop mount
vinifm has quit [Quit: Saindo]
rellla has quit [Quit: rellla]
<techn> mnemoc: so that's the way I used already
slash_random has joined #arm-netbook
<techn> I checked from build root and they are using genext2fs
<techn> *buildroot
<gx260> Here is the full error: http://pastebin.com/0H1p62EB
<mnemoc> techn: why do you need a filesystem image?
<techn> mnemoc: livesuit image eats them
<mnemoc> :<
<mnemoc> the only thing different than genext2fs thing can do is to make the initial file sparse instead of solid
<mnemoc> ah, doesn't need sudo... nice
pawel5870 has joined #arm-netbook
<mnemoc> gx260: can you add more lines on top?
<gx260> sure, hang on.
Dssence has quit [Ping timeout: 268 seconds]
<hno> Turl, -U8 for 8 lines
cat_x301 has quit [Ping timeout: 240 seconds]
axion has left #arm-netbook ["WeeChat 0.4.0-dev"]
<gx260> I did a make clean and, it is compiling now. be just a min.
<gx260> Ok, here it is: http://pastebin.com/57Bk89HC
eFfeM has quit [Quit: Leaving.]
<mnemoc> gx260: git log HEAD^.. ?
<hno> A / too many making it fail when not using O= ?
<mnemoc> but he isn't using -C ...
* mnemoc puzzled
<hno> He is builting in-tree from what I understand.
<mnemoc> yes
<mnemoc> i don't get why mali's Kbuild gets confused
cat_x301 has joined #arm-netbook
<hno> If there is a / too many then in-tree builds gets confused byt out-of-tree builds works.
tinti has joined #arm-netbook
<mnemoc> gx260: try: mkdir build; make mrproper; make ARCH=arm CROSS_COMPILE=arm-linux-gnueabi- O=build sun4i_defconfig; make ARCH=arm CROSS_COMPILE=arm-linux-gnueabi- O=build -j4 uImage modules ?
tinti has quit [Client Quit]
<gx260> git log HEAD^: http://pastebin.com/v5uHassL
<libv> rz2k: those differences are X version dependant
<libv> rz2k: i think it is dave's massive api breakage
<libv> rz2k: there must be ways to work around this at build time
<mnemoc> libv: any suggested #ifdef ?
<libv> i would have to spend some time finding that myself
<libv> check the trees of "upstream" drivers
<rz2k> thats why I gave up for now, will dig there at weekend
<libv> or check the api/abi version, or xorg version
<libv> there must be several routes in this case
<mnemoc> #if XORG_VERSION > 112 :p
<libv> something like that, yes
<libv> but there must be something more succinct than that
tinti has joined #arm-netbook
arnd_ has quit [Ping timeout: 256 seconds]
cat_x301 has quit [Ping timeout: 240 seconds]
rz2k has quit [Ping timeout: 245 seconds]
traeak has joined #arm-netbook
tom__11 has joined #arm-netbook
tom__11 is now known as sequence
rz2k has joined #arm-netbook
<gx260> ok, seems to be ok now. Thanks!
<mnemoc> hno: anything we can do to solve that "confusion" without restructing the mali driver?
<mnemoc> gx260: it has a good side effect, your tree is always clean ;-)
<mnemoc> gx260: keep using O= for now
pawel5870 has quit [Ping timeout: 252 seconds]
gimli has quit [Quit: Verlassend]
<sequence> Does anyone have a recommendation for a cheap ARM-based device that has ethernet, SATA and USB support? I just bought a Cubieboard, is there something cheaper out there?
<sequence> It was like $50 USD
<hno> mnemoc, find where that / is. $(...)/drivers should be $(...)drivers somewhere.
<mnemoc> hno: ah, cool. I thought it was about levels, not a double /
<hno> a stray /, not double.
pawel5870 has joined #arm-netbook
rz2k has quit [Ping timeout: 245 seconds]
traeak has quit [Ping timeout: 246 seconds]
<hno> mnemoc, mali Kbuild files looks overly complicated and odd..
<mnemoc> mali/Kbuild:submodule_enabled = $(shell gcc $(DEFINES) -E $(PWD)/$(DRIVER_DIR)/arch/config.h | grep type | grep -c $(2))
<mnemoc> is the only candidate
<mnemoc> yes, very
<hno> Looks kind of "Ok, we have this old legacy Makefile full of crap. Lets sed translate it to a Kbuild file.
traeak has joined #arm-netbook
pawel5870 has quit [Quit: Leaving.]
Entropy512 has quit [Remote host closed the connection]
Entropy512 has joined #arm-netbook
mikey_w has joined #arm-netbook
<mnemoc> tried an O=-less build here and worked fine :|
<mnemoc> -C without O= fails misserably.... in mali obviusly
pawel5870 has joined #arm-netbook
sequence has quit [Remote host closed the connection]
<mnemoc> -C issue fixed
<mnemoc> gx260: can you try without O= now?
<mnemoc> update to the head of sunxi-3.0 first obviusly
alcides has joined #arm-netbook
alcides has quit [Changing host]
alcides has joined #arm-netbook
pawel5870 has quit [Ping timeout: 252 seconds]
<libv> mnemoc: why is EGL, GLESv1 and GLESv2 lumped together in libMali.so?
<libv> is this the way linaro has been doing it?
<libv> because on both telechips and amlogic android tablets, i had separated libs
<mnemoc> probably because it's the only way tom found to build them
<mnemoc> in the Mali400 page there is a table with all the failed tries
<libv> ah, so hipboi is responsible for these binaries
<mnemoc> he gets them, yes
<mnemoc> but without allwinner's support
<mnemoc> or arm's support
<libv> right
<libv> i am thinking about a big rename
fredy has quit [Ping timeout: 260 seconds]
<libv> libMali_fb.so.3.0
<libv> and then having symlinks do the rest
<mnemoc> the the .so will still keep the old name inside...
<libv> hrm, true
<mnemoc> ldconfig will destroy everything
<libv> and the space for those strings is badly packed
<libv> s/badly //
<ibot> libv meant: and the space for those strings is packed
* libv spent a few weeks playing with elf headers in Q2 2011 before his mali tablet gained android-2.3 and therefor LD_PRELOAD
<libv> ever so long ago though :)
<mnemoc> :)