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/
<ssvb> I could not build gnome-shell out of the box, I think it needs some linaro/ubuntu patches to make it GLES friendly
<jelly-home> le sigh, I've managed to stay away from source-based distros since the last millennium
<ssvb> though gnome-shell starts fine from the ubuntu chroot, so in theory it should not be too difficult to get working
<jelly-home> well, replacing the desktop wouldn't have been my first idea on how to use an A10 box, but if it works...
aexl_ has joined #arm-netbook
<ssvb> hramrach: maybe they have a different package for kwin_gles in debian?
<aexl_> lo. does somebody have a statically linked abootimg?
<jelly-home> ssvb: there is none in debian, I checked; Ubuntu however seems to have a http://packages.ubuntu.com/precise-updates/kde-window-manager-gles
<jelly-home> "just" rebuilding kdebase-workspace source for armhf :-)
<jelly-home> so much about binary-based distros
<penguin42> jelly-home: How long does that take on an a10 ?
marcus__ has quit [Remote host closed the connection]
<jelly-home> penguin42: no idea -- personally I wouldn't build kde in 1GB
<penguin42> indeed
<jelly-home> oh, ports.ubuntu.com has it
<jelly-home> well. I _might_ try building it if I had an extra ssd to sacrify for build tree and swap space
<ssvb> jelly-home: well, I just got fed up with this crap and decided to get a clean environment for checking what actually works with GLES and what does not
<ssvb> jelly-home: gentoo is quite convenient, because it is a kinda meta-distribution which can be reconfigured as you wish
<ssvb> jelly-home: I guess there are not many weirdos who would like to completely rip out the full OpenGL from their system :)
aesok has quit [Remote host closed the connection]
<jelly-home> (or ones who don't have it at all like here)
<hramrach> I suspect you can't remove it when you have Qt
<hramrach> some stuff just links with it and the binaries include every feasible feature
anunnaki has quit [Quit: Leaving]
<FergusL> did anybody buy the Wandboard ?
anunnakiii has joined #arm-netbook
<ssvb> hramrach: it's not a problem if I can rebuild the whole system (all the Qt packages, etc.)
anunnakiii is now known as anunnaki
<ssvb> hramrach: and now I also have an x86 system with exactly the same setup, should make it easier porting/debugging/comparing :)
XenGi is now known as XenGi_
bfree has quit [Ping timeout: 252 seconds]
bfree has joined #arm-netbook
bfree_ has joined #arm-netbook
bfree has quit [Ping timeout: 252 seconds]
stefanro has joined #arm-netbook
hp__ has quit [Ping timeout: 264 seconds]
stefanro1 has quit [Ping timeout: 260 seconds]
alcides has joined #arm-netbook
alcides has joined #arm-netbook
alcides has quit [Changing host]
alcides has quit [Quit: /(bb|[^b]{2})/ regular expression junkie + lover of literature...]
penguin42 has quit [Quit: Leaving.]
ln2 has quit [Quit: http://www.kiwiirc.com/ - A hand crafted IRC client]
MMlosh has quit [Quit: Bye...]
focus_it has quit [Remote host closed the connection]
XenGi_ is now known as XenGi
dfletcher is now known as drgreenthumb
bfree__ has joined #arm-netbook
bfree_ has quit [Ping timeout: 252 seconds]
herdingcat has joined #arm-netbook
aexl_ has quit [Quit: Page closed]
<jelly-home> FergusL: this is the first I even heard about it there's one interesting thing going for it -- EDM, 314 pins, existing before EOMA-68 happened
<FergusL> ha jelly-home
<FergusL> is that good ?
<jelly-home> http://www.technexion.com/index.php/products/edm depends on how many vendors end up using it I guess
<FergusL> yes, that's what they say in their wiki
bfree has joined #arm-netbook
bfree__ has quit [Ping timeout: 252 seconds]
XenGi is now known as XenGi_
fragmint has quit [Read error: Connection reset by peer]
fragmint has joined #arm-netbook
aholler has joined #arm-netbook
aholler_ has quit [Ping timeout: 248 seconds]
KoH_ has joined #arm-netbook
KoH__ has quit [Ping timeout: 248 seconds]
eebrah has joined #arm-netbook
bfree has quit [Ping timeout: 252 seconds]
bfree_ has joined #arm-netbook
xymox has quit [Ping timeout: 260 seconds]
xymox has joined #arm-netbook
xymox has quit [Ping timeout: 276 seconds]
xymox has joined #arm-netbook
sky770 has joined #arm-netbook
gimli has joined #arm-netbook
gimli has quit [Read error: Connection reset by peer]
gimli_ has joined #arm-netbook
gimli_ is now known as gimli
anunnaki has quit [Remote host closed the connection]
anunnaki has joined #arm-netbook
<libv> heh, libump build seems quite retarded
<libv> *sigh*
hp__ has joined #arm-netbook
rellla has joined #arm-netbook
<sky770> libv: I thought, it was ARM's PRs
<sky770> :D
hp__ has quit [Ping timeout: 264 seconds]
hp__ has joined #arm-netbook
<libv> tonight, i am going to make the libump repo redudant.
<libv> redundant even
Quarx has joined #arm-netbook
eFfeM has joined #arm-netbook
<sky770> omfg :O ppl at #ubuntu don't even know the diff b/w "ubuntu for phones" & "ubuntu for android"
* sky770 *dammnit* :|
TimeBreaker has joined #arm-netbook
TimeBreaker has quit [Client Quit]
cnxsoft has joined #arm-netbook
anunnaki has quit [Ping timeout: 256 seconds]
eFfeM has quit [Quit: Leaving.]
ganbold__ has joined #arm-netbook
ganbold_ has quit [Ping timeout: 248 seconds]
w00tc0d3 has quit [Remote host closed the connection]
w00tc0d3 has joined #arm-netbook
sky770 has quit []
abesis2 has quit [Ping timeout: 240 seconds]
rz2k has joined #arm-netbook
wingrime has joined #arm-netbook
Quarx has quit []
abesis has joined #arm-netbook
eFfeM has joined #arm-netbook
Guest94707 is now known as fredy
fredy is now known as Guest77256
abesis has quit [Ping timeout: 245 seconds]
abesis has joined #arm-netbook
<mnemoc> libv: you mean a lima version of arm's libump?
cnxsoft has quit [Quit: cnxsoft]
gimli has quit [Ping timeout: 248 seconds]
focus_it has joined #arm-netbook
sky770 has joined #arm-netbook
penguin42 has joined #arm-netbook
* sky770 #ubuntu-phone is coming!! dev touch review this 21st :p :p
gimli has joined #arm-netbook
eFfeM has quit [Ping timeout: 260 seconds]
jukivili has quit [Ping timeout: 245 seconds]
gimli has quit [Ping timeout: 252 seconds]
<specing> sky770: all fine with me if it is not going to be overpriced like neo freerunner
cnxsoft has joined #arm-netbook
eFfeM has joined #arm-netbook
aesok has joined #arm-netbook
OmegaPhil has quit []
OmegaPhil has joined #arm-netbook
OmegaPhil has quit [Changing host]
OmegaPhil has joined #arm-netbook
<sky770> specing: "overpriced" ? lol ..I was talking abt Ubuntu for phones "OS" * :p
<hramrach> they are developing hardware too
<sky770> hramrach: not exactly..but they're roping in OEM's for that thing
<sky770> Huawei being one of them :|
* sky770 suspects chinese low quality built phones ..*nuff*
* sky770 no offence * :D
<sky770> bad thing about the dev. preview this 21st is that there won't be any support apart from nexus/nexus 4 (well..atleast on that day..or even some days after..)
<sky770> devices will eventually get added in the support list depending on indie devs getting "interested" in porting ubuntu to a particualr handset :|
<sky770> well more info to come during/after this MWC :D
<sky770> 25-28th feb :)
<OmegaPhil> 'A signature flourish for Ubuntu and for you.'
<OmegaPhil> jesus christ
<sky770> ? :(
<sky770> amen.. ? :D
<OmegaPhil> I'll get him on the phone and have a talk about what I read
cnxsoft has quit [Ping timeout: 256 seconds]
hansg has joined #arm-netbook
<ssvb> libv: what's the motivation for making libump redundant?
FergusL has left #arm-netbook ["Quit"]
gimli has joined #arm-netbook
sky770 has quit []
rellla2 has joined #arm-netbook
rellla1 has joined #arm-netbook
<libv> ssvb: i want to fold it in with mali-libs, in such a way that things get built and installed in one go
<libv> mnemoc: nope
<libv> the current setup is a mess, and leads to issues all around
<libv> libump will be built from source, and will only get the X11 dependencies when make x11 is run
<libv> for framebuffer, it will be built without those dependencies
<libv> all the existing libump.sos will be removed from mali-libs
<libv> suddenly, this thing will make sense and will be useful.
<libv> oh, libump will be built from source, and the depending on the kernel version, the right version of it will be built too
<libv> no more of this multiple branches nonsense
<libv> there really is no reason to carry both the binaries and the source around
<libv> we need to build from source anyway to work around linking issues, so the binaries go
<ssvb> libv: ok, I just thought that we already build it from sources
<libv> ssvb: no, we first install the binary
<libv> and then the howto tells you to build from source
<libv> which then gets you the X11 libs linked in
<ssvb> but you fixed it already?
<libv> only for r3p0/hf/x11
<ssvb> ok
<libv> so the current setup is a complete mess.
<techn__> true
<ssvb> yes, no disagreement here :)
<libv> what we will get is a one-shot install.
<ssvb> what's going to happen to https://github.com/linux-sunxi/libump ?
<libv> master should become an empty commit
<libv> it is not as if it is tracking ARM code releases anyway
<techn__> apt installable package would be nice too.. which will tell packaging management not to install mesa stuff
<libv> techn__: as the main author of the lima driver, i find it amazing that i am having to spend so much time making sense in all of this myself already
ganbold_ has joined #arm-netbook
<libv> surely others see the problems with this too
<libv> then why do i end up being the one fixing such brokenness here?
<libv> so no, i am not going to provide a debian subdirectory
<ssvb> sorry, you lost me here
<libv> i already am pretty unhappy about the fact that noone else could've been arsed to fix up the test for x11
<libv> apart from a binary hack to libMali.so, making that test also run on x11 was really trivial
<ssvb> yes, but apparently nobody was interested in this test
ganbold__ has quit [Ping timeout: 252 seconds]
<ssvb> you are interested, so you fixed it, problem solved
<ssvb> thanks, that's surely appreciated
<libv> ssvb: yet techn__ merged in a single commit to help explain why it was not working on X11
<libv> and every week or so, someone would come round and explain
<libv> s/explain/complain/
<ibot> libv meant: and every week or so, someone would come round and complain
<libv> but no, i am not wasting my time on debian packaging this thing as well
<ssvb> neither do I :)
<techn__> .. was it sky770 who is taking initiative as package maintainer :/
<hramrach> packaging this is pretty much impossible with two different sets of libraries
<ssvb> packaging is problematic because of the missing license notice for the mali binaries, I would guess a lot of people are avoiding to touch this stuff simply for this reason
<libv> nope, people are just lazy
<ssvb> that's also true, but surely not the only reason
<libv> in this case, it is
<ssvb> let's ask hansg about what he thinks about packaging mali drivers for fedora ;)
<libv> ssvb: hansg has better things to do
<ssvb> we all have better things to do
<libv> heh
<libv> ssvb: how many people have come round and said "why is there no packages for this"
<hramrach> never heard anyone saying it
<libv> ssvb: how many of those were even considering doin the busywork that is creating packages?
<libv> ssvb: how many of the remaining, which is none, were then stopped by the fact that there were binaries in there which do not come with a license?
<hramrach> presumably the stuff is pacakged for gentoo
<ssvb> hramrach: not officially :) but I'm ready to provide help to anyone trying to use this stuff
<hramrach> which is pretty much the only distro which can bear the current quality of the code
<bfree_> I've done some work on deb packaging, but so far I've only worked on the kernel and uboot (ok I started to look at sunxi-tools the other day). I've not even touched any video stuff myself yet (apart from just what is in Debian to confirm X will come up), let alone considered attempting to package it
bfree_ is now known as bfree
vinifm has joined #arm-netbook
<hramrach> the driver released by hardkernel is even better
<hramrach> it has copyright notice
<hramrach> it says you can run it on snowball - an Ericsson evb
<hramrach> and only that board
<hramrach> not, say, an Odroid or mk802
<ssvb> lol, have you reported it to hardkernel people?
herdingcat has quit [Remote host closed the connection]
<ssvb> looks like somebody copy-pasted too much
<ssvb> but at least they are trying to do it right
rellla has quit [Remote host closed the connection]
<ssvb> looks like so far only raspberry pi has done the right thing with the blob gpu drivers for arm hardware (if we ignore their shameless marketing)
<ssvb> for raspberry pi we have all the firmware releases backed up in a freely accessible git repository, accompanied with a license which explicitly allows redistribution. IMHO that's about the best possible way to handle it, assuming that the firmware sources have to be kept closed.
<hramrach> no, I did not report it
<libv> ssvb: does your xf86 driver use libump headers?
<ssvb> libv: yes
<libv> ok
eebrah has quit [Ping timeout: 276 seconds]
blindcoder has joined #arm-netbook
<blindcoder> hello
<blindcoder> am I right here for the pengpod?
<blindcoder> I want to compile my own kernel for it (need the /dev/snd/seq device) and wonder where to get the kernel sources
<ssvb> hramrach: but assuming that they announce it as "You can enjoy the Mali400 OpenGL-ES 3D software develoment", there is little hope for the casual end users who actually want to use this shit, and not "develop" for it :-/
<RaYmAn> blindcoder: Surely "True Linux Tablets and MiniPCs" provide kernel source along with the device.
<blindcoder> Turl: thanks!
<blindcoder> Turl: I want to connect it to my midi keyboard using pianobooster, and that needs the /dev/snd/seq device.
<RaYmAn> There's no really any pengpod specific development. They just use the general A10 development happening.
rellla1 has quit [Ping timeout: 240 seconds]
rellla2 has quit [Ping timeout: 248 seconds]
ganbold_ has quit [Remote host closed the connection]
<drachensun> blindcoder: we have a #pengpod channel too
hansg has quit [Remote host closed the connection]
<blindcoder> drachensun: ah, cool
roric_ has joined #arm-netbook
eFfeM has quit [Quit: Leaving.]
panurge has joined #arm-netbook
anunnaki has joined #arm-netbook
abesis has quit [Ping timeout: 252 seconds]
abesis2 has joined #arm-netbook
hramrach has quit [Remote host closed the connection]
ZaEarl has joined #arm-netbook
panurgebr has joined #arm-netbook
panurge has quit [Ping timeout: 256 seconds]
wingrime has quit [Ping timeout: 255 seconds]
<slapin> hi, all?
<anunnaki> im testing crosscompiling a sample file from a guide im following. and it says to emerge qemu. should i do 'USE="arm -x86_64" emerge qemu' or will simply emerge qemu suffice?
<slapin> is there some tool to route ddr3 running under Linux?
jukivili has joined #arm-netbook
<slapin> as I find it too boring to calculate impedance by hand
<anunnaki> my target arch is arm, this host pc is x86_64
blindcoder has left #arm-netbook [#arm-netbook]
<anunnaki> please use my name if you answer my question for ill be jumping channels. thanks guys
w00tc0d3 has quit [Remote host closed the connection]
panurgebr has quit [Quit: Saindo]
rellla has joined #arm-netbook
eebrah has joined #arm-netbook
FergusL has joined #arm-netbook
eebrah has quit [Ping timeout: 255 seconds]
pwhalen has quit [Quit: Leaving]
<Turl> pogoplugs for 19$ each http://www.adorama.com/COCPOGOP21.html
<jinzo> yeah, a bit expensive shipping if you're not in US
<jinzo> but still, a nice deal.
<Turl> apparently that specific model is not mainline supported though
muts has quit [Remote host closed the connection]
muts has joined #arm-netbook
anunnaki has quit [Ping timeout: 248 seconds]
<jinzo> I stumbled on the new pogoplugs, the one with USB 3.0 and SATA looks quite interesting at a nice price point
anunnaki has joined #arm-netbook
anunnaki has quit [Read error: Connection reset by peer]
rellla has quit [Quit: rellla]
w00tc0d3 has joined #arm-netbook
penguin42 has quit [Quit: Leaving.]
anunnaki has joined #arm-netbook
vinifm has quit [Quit: Ex-Chat]
ZaEarl has quit [Ping timeout: 246 seconds]
gimli has quit [Remote host closed the connection]
hramrach has joined #arm-netbook
<buZz> woa,, A31 being released?
<jinzo> buZz, shipping for quite some time now. ~mid jan.
<jinzo> or somewhere there.
<buZz> nice
<buZz> is A20 then also released?
<jinzo> haven't seen anything shipping with it yet.
<jinzo> didn't even notice any preorders floating around (but I'm not following this stuff that much nowadays)
<techn__> got simplest drm driver skeleton working.. with no hw calls.. but should help things to get started :/
<techn__> .. but it's still not ready to spread around.. but if there is volunteers to participate developing.. I could try to push it during next week
eFfeM has joined #arm-netbook
<libv> techn__: drm driver to do what with?
<techn__> libv: Currently it creates dummy bindings for /dev/fb and sysfs
ZaEarl has joined #arm-netbook
<libv> techn__: and the goal is?
<techn__> short term.. to get simple framebuffer which uses normal layer (no scaling) and hdmi
<techn__> long term.. whole disp to mainline
<techn__> but.. if someone wants to implement TCON0 and LCD, RGB or tvout for short term, its welcomed too :p
<libv> it's the wrong approach
<libv> one that leads to dumbing down of our userbase and the loss of register level information
<libv> techn__: the correct way to approach this is to clean up and restructure disp and then add drm kms support on top of the then clean and structured code
<hramrach> how is register level information lost?
<techn__> If we want refactor current driver we need to do it twice.. with huge regression risk.. first to get it cleaned top of old arch.. then move it to drm/kms..
<libv> nope
<hramrach> no move if you just add kms glue on top
<libv> you are, at best, creating a situation where you have two different drivers, both broken in different ways, and no possibility to fix either
<techn__> and this current driver is suffering issues which comes free for drm/kms
<libv> and there is little regression risk when cleaning/restructuring properly
<libv> but that does require patience and clue
<techn__> like improved page flipping and architecture for dual display
<libv> for free?
<libv> really?
<hramrach> the architecture for dual display does not come out of nowhere
<libv> you'll have to explain the mechanics of that "for free" to me then
<hramrach> you still ahve to figure out how to configure it with the actual hardware
<libv> techn__: in any case, if you want to do something useful, and want to play with kms
<libv> try bolting kms on top of the current disp and try to sensibly restructure disp underneath it
<libv> the approach you are taking now is just stupid, and will only lead to further stagnation
FergusL has left #arm-netbook ["Quit"]
<libv> it is however something one thinks off on a sunday afternoon
eFfeM has quit [Quit: Leaving.]
<libv> but believe me, there are much better things to do, which actually will help us along long and short term
<libv> techn__: for an example of the sort of kamikaze stupidity you are running into: the openchrome driver
<techn__> I just hate to see that if I do something for current module I have to do it again in future
<techn__> libv: I understand that
<libv> they forked away from me because i did not want to accept VBE based modesetting
<libv> today, they have 3 different modesetting paths and the few remaining openchrome users try each of the three until one of them sort of works
<libv> nothing ever gets fixed.
<libv> techn__: use your spare time/energy elsewhere then
sv has joined #arm-netbook
discopig has quit [Ping timeout: 276 seconds]
drgreenthumb has quit [Read error: Connection reset by peer]
sv is now known as discopig
rz2k has quit []
<ssvb> techn__: I agree with libv, there way too many examples which show that the drm/kms, wayland, ... <insert some other trendy new thing here> .... is not the silver bullet
<ssvb> techn__: there is a good explanation in "Fire And Motion" by Joel Spolsky - http://www.joelonsoftware.com/articles/fog0000000339.html
<ssvb> techn__: the part about "Watch out when your competition fires at you. Do they just want to force you to keep busy reacting to their volleys, so you can't move forward?" :)
<ssvb> techn__: and just substitute "ODBC, RDO, DAO, ADO, OLEDB, now ADO.NET - All New!" with the things relevant to us - "DRM/KMS, V4L2, Wayland, ..."
<libv> disp is a dead-end, it should die
<libv> but it should die slowly, through dissection
<libv> there is loads of very relevant information in there
<libv> which we simply cannot afford to lose
<ssvb> yes, exactly
<libv> the way to do it is to restructure it, and then bolt something better on top
<libv> not by starting afresh
<libv> you get absolutely nothing for free when starting afresh
<libv> and you throw out the baby with the bathwater while doing so
<libv> and only then does the openchrome syndrome kick in
<libv> which then kills _everything_
<anunnaki> hi need help with qemu. im following this guide http://psas.pdx.edu/GentooCrossComipilerHowTo/
<anunnaki> and when i get to Run the test program step. i typed qemu-arm ./hello i get "qemu-arm command not found"
<anunnaki> i also tried qemu-armv4 ./hello as well.. same result
<anunnaki> i also emerged qemu with the arm use flag.. no change in result
<libv> hrm...
<ssvb> libv: many of the remaining bugs in disp are still fixable, and it is possible to get a somewhat sane graphics stack running on top of it, this requires less efforts than full rewrite
<mnemoc> +1
<libv> ssvb: to what extent does xf86-video-mali depend on ump, does it depend on any of the newer symbols that r3p0 (and up, as that's all the same) provides?
<ssvb> libv: the real problem is that libMali.so depends on ump
<libv> ssvb: i am struggling with introducing versioning on the headers
<libv> i haven't figured out what the best route is
<libv> oh, that one is easy
<libv> my current code just installs the r3 headers
<libv> as there are no abi breakages in there
<libv> there are only additions to the api
<libv> and things will break trying to link to an r2 ump while assuming that you are building against the full api exposed in these headers
<libv> i am currently thinking about echoing a version into ump_version.h
<libv> but this then requires proper dependency tracking when building libUMP.so
<ssvb> r3 version of libump introduces buffers cacheability control, but I think this should be already covered - https://github.com/ssvb/xf86-video-sunxifb/blob/master/configure.ac#L89
<libv> aha, check_lib
<libv> nice
<ssvb> well, maybe not totally nice and reliable, but we are yet to see a bugreport about it failing
<libv> still, it would be better if some #define introduced from ump.h would help delineate this
<libv> oh, it is reliable, but it is not the cleanest solution
<libv> it could make life real easy for me though
<ssvb> there always could be a mismatch between the headers and the libraries or something like this
aholler has quit [Quit: 3.7.9]
aholler has joined #arm-netbook
FergusL has joined #arm-netbook
<libv> heh, exploding complexity...
<libv> so i guess i will keep version out of the headers for now
aesok has quit [Remote host closed the connection]