<Turl>
you just put the pin high and the corresponding relay is activated
<mrnuke>
that's worthless. All brawns and no brains
<Turl>
mrnuke: you just hook it to an arduino and bam :) it's smart
<mrnuke>
ooh. I don't do arduino
<Turl>
sunxi board then
<mrnuke>
I'm more of a baremetal type of sucker
<mrnuke>
Nice one! :)
<Turl>
arduino is convenient :p the clike code and one click deploy on the ide is nice
<mrnuke>
I'm thinking of building an expansion board for cubie so that it can fit with the plastic case. You have to take it out of the case to put it on existing expansion boards
<Turl>
mrnuke: what cb and what case? what expansion board?
<mrnuke>
holy fuck radeon performance is skyrocketing
<orly_owl>
yup :)
<orly_owl>
and nouveau has been decent for a while
<mrnuke>
at least AMD is contributing to FLOSS. Can't say that of the nvidia (high-end GPU) side
<orly_owl>
radeon cards still need a nonfree binary though
<orly_owl>
amd contributes to coreboot a bit too
<orly_owl>
so does intel, mostly via google's laptops
<mrnuke>
intel doesn't contribute. It's google that puts in purchasing power to get documentation, and money to pay developers
<mrnuke>
intel just gets a free ride, the motherfuckers
hipboi has joined #linux-sunxi
<orly_owl>
fair point
<orly_owl>
intel's coreboot code relies on nonfree microcode also
<mrnuke>
you know what intel's answer was to coreboot's growing market share?
<orly_owl>
what?
<mrnuke>
Firmware Support Package (FSP)
<mrnuke>
A (maybe) redistributable package of blobs, that only let you configure PCI devices (coreboot's no 1 strength)
<mrnuke>
mind you, it's a no redistribution click-through license
<mrnuke>
so they can't use our resource allocator, due to it being GPL'd, but they figured those suckers (us) will just take their shitty binaries, figure out the licensing issues, and give them free coreboot
<Turl>
mrnuke: sunxi-devel has tons on stuff not on 3.13
<wens>
davidm just took the stmmac core changes into net-next
<Turl>
wens: great
<Turl>
now it's just the clocks missing right?
<wens>
Turl: yep, clocks and DT
<Turl>
it should all land nicely on 3.15
<wens>
Turl: I'm not sure whether he took it in for 3.14 or 3.15 though, since he replied after 3.13 was released
<wens>
and net-next is generally closed during the merge window
<Turl>
we'll see soon when the pull req goes out
JohnDoe_71Rus has joined #linux-sunxi
[7] has quit [Disconnected by services]
TheSeven has joined #linux-sunxi
<mrnuke>
LMFAO!!!! "git show 2b844ba"
<wens>
WTF? weren't the BayTrail IDs added in 3.13?
<mrnuke>
apparently they were removed just before the release
<wens>
I see
<wens>
kernelnewbies lists this in 3.13 changes
<mrnuke>
BTW how does sunxi-devel work? Does it get rebased on 3.13, or does work just continue and it gets merged or cherry-picked as deemed appropriate ?
<wens>
AFAIK, work gets cherry-picked
<wens>
hansg will probably rebase when rc1 comes out
<wens>
he updates his own sunxi-devel branch often, probably forgot to push it to linux-sunxi
<mrnuke>
I'm gonna try to cherry-pick mripard_'s SPI work. That will be interesting to see working
<libv>
oliv3r: try stage/sunxi-3.4
newleaves has joined #linux-sunxi
orly_owl has quit [Ping timeout: 272 seconds]
<wens>
urgh, musb uses the same "register a child device" approach as ahci-imx
iamfrankenstein has joined #linux-sunxi
iamfrankenstein has quit [Read error: Connection reset by peer]
mouchon2 has quit [Ping timeout: 252 seconds]
prasannapete has joined #linux-sunxi
FreezingCold has quit [Ping timeout: 272 seconds]
FreezingCold has joined #linux-sunxi
iamfrankenstein has joined #linux-sunxi
atiti has quit [Ping timeout: 265 seconds]
iamfrankenstein has quit [Ping timeout: 252 seconds]
<gzamboni>
Turl nice relay card, i will buy one :)
shineworld has joined #linux-sunxi
ssvb has quit [Quit: Leaving]
atiti has joined #linux-sunxi
HeHoPMaJIeH has joined #linux-sunxi
Quarx has joined #linux-sunxi
orly_owl has joined #linux-sunxi
<oliv3r>
libv: i'll pull the latest; but i am allready on staging
<mnemoc>
oliv3r: where do you suggest to put the legacy-support code in 3.10 if not in misc/ withou breaking multiplatform?
mouchon2 has joined #linux-sunxi
<oliv3r>
nothing new
<oliv3r>
mnemoc: i dunno :p mach or plat?
<mnemoc>
uhm... will the needed -I be added in multi?
<oliv3r>
technically its our 3.10 so we can hack at will imo
<mnemoc>
ok
<oliv3r>
might aswell keep it in misc?
<oliv3r>
if we hack it anyway
<oliv3r>
i just fiugred having it in mach/plat, we keep it out of the 'main' stuff
<mnemoc>
i'll go for mach-sunxi/include/mach then
<mnemoc>
btw,the reason to enable soc-detect that early is for having the info available by the time clocks get setup in not-well-known socs
<oliv3r>
yeah
nedko has quit [Quit: kernel panic]
nedko has joined #linux-sunxi
nedko has quit [Changing host]
nedko has joined #linux-sunxi
nedko has quit [Client Quit]
nedko has joined #linux-sunxi
nedko has quit [Changing host]
nedko has joined #linux-sunxi
_massi_ has joined #linux-sunxi
<mnemoc>
helps hooking hacks and exceptions to easily import code from dumps and RE
<mnemoc>
my ssd will be shipped (from uk) the 24 :/
<mnemoc>
but 65€ cheaper than from .de
<oliv3r>
wow that's a big diff
<mnemoc>
141€ vs. 208€ for the same intel 530 240GB
<oliv3r>
strange
<oliv3r>
150 E in NL
bbrezillon has joined #linux-sunxi
<mnemoc>
i thought I was buying it from de... i-store-online.de ... but the notification tells it will be shipped from .uk
<mnemoc>
the 24 :|
<oliv3r>
ah ok; so no extranous shipping rates
<oliv3r>
good
<oliv3r>
3 days of air travel; maybe get it by the 27th
<mnemoc>
hope so
<mnemoc>
even if I found an old 40GB spinning drive and temporarily installed that, it's frustratingly slow :( ... not at slow and usb-killing as the stick I used before, but you can still feel the latency even while typing on a terminal
<nedko>
typing in terminal involves disk io? o.0
<mnemoc>
no, but X does...
<mnemoc>
and love to freeze itself
<mnemoc>
loves*
<mnemoc>
i could indeed switch to good old fbcon
<nedko>
a terminal emulator for x11?
<oliv3r>
mnemoc: you just got spoiled :p
<nedko>
i'm pretty sure my x server and x terminal don't do disk io when i type in terminal
<oliv3r>
mnemoc: its sata, it can't be THAT slow
<oliv3r>
i have IDE in my laptop :)
<oliv3r>
though that's an SLC sdd :p
sspiff has joined #linux-sunxi
sspiff has quit [Changing host]
sspiff has joined #linux-sunxi
paulk-collins has joined #linux-sunxi
<mnemoc>
oliv3r: yes, I agree. I was spoiled by too many years of ssd
<oliv3r>
let me know if you did not receive your ssd in time; i have a few 'fast' 100g0 - 500 gb sata drives :p
<mnemoc>
nedko: when thunderbird starts syncing even my mouse pointer halts... so something weird is going on which didn't happen with the ssd
<mnemoc>
oliv3r: I ordered a lexer p10 usb stick too... just for the case
<oliv3r>
if you won't get your stuff in time, i'll bring one with me :)
<mnemoc>
thanks
<oliv3r>
7200RPM if i'm not mistaken
<oliv3r>
60gb
<gzamboni>
i think i need an SDD as my thunderbird is horribly slow even if i do cleanup and tweak not to be
<sspiff>
hi, what's the current status of A20 & A31 support?
<mnemoc>
sspiff: A20 is pretty good. A31... we don't even know how to boot it :(
foubarre has joined #linux-sunxi
<oliv3r>
specing: a31 has no support
<oliv3r>
specing: maxime does some mainline patches though
<foubarre>
wens: hi. Feeling fine. Sorry i missed your message last time.
<nedko>
mnemoc: either you have no enough ram or your system is misconfigured...
<sspiff>
mnemoc: I saw that Mozilla got Firefox OS to run on an A31 device. I'm sure they'd have kernel sources, maybe they'd be willing to share?
<oliv3r>
sspiff: we have those too
FR^2 has joined #linux-sunxi
<libv>
sspiff: nobody just spent time on it
<libv>
oliv3r: your crashing problems are really strange.
<sspiff>
oliv3r, libv: oh ok, never mind then :)
<oliv3r>
libv: whenever i move the mouse, BAM
<libv>
oliv3r: i have no such issues here
<libv>
oliv3r: disable AIGLX ?
<oliv3r>
libv: but let me hook everything up first :) just got settled at work
<oliv3r>
libv: the compositing stuff isn't used by lightdm yet is it?
<libv>
although, it should not be used at all
<gzamboni>
i'm trying to get information from dts and i'm quite confusing, in my dts i have the res base address + lenght and my driver get them properly but it doenst get the irq i did specify in the dts, im using irq_of_parse_and_map, in my dts i have irq 26 and it gets always irq22
<libv>
but why does moving the binaries help?
<sspiff>
I've got a few A10 devices (MK802 1GB, MiniX and Ainal Novo7 Aurora), and I'd like to be able to build Linux images for them, can someone point me to a suitable wiki page?
<libv>
sspiff: new_device_howto
<libv>
sspiff: website is called linux-sunxi.org
<oliv3r>
any any point moving only 1 of the two binaries back? limare.so or lima_dri.so?
<libv>
sspiff: filled with all sorts of good stuff.
<sspiff>
libv: thank you!
<oliv3r>
gzamboni: hex vs dec?
<libv>
oliv3r: i really don't think it should have any effect
<oliv3r>
libv: ok i modified my memory reservations to what ssvb suggested, 0 for g2d, 80 for ve; no mention of mali; 16mb for fb
<oliv3r>
i'll make a xorg.conf now that disables AIGLX
<oliv3r>
ah fbturbo did install an xorg.conf
JohnDoe_71Rus has quit [Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org]
<oliv3r>
swapbufferswait = true
<libv>
oliv3r: you are on stage/3.4 right?
<libv>
oliv3r: don't meddle with xorg.conf for now
<libv>
oliv3r: one step at a time.
<oliv3r>
libv: yep
<oliv3r>
i thoguht you said disable aiglx
<oliv3r>
ok i'll move the bins back; to test the new memory reservations
<libv>
disabling aiglx might be something worth trying
<libv>
although it really doesn't seem to fit
<oliv3r>
right, restarting X
<gzamboni>
oliv3r, afaik for the irq it gets an int value from dts(b)
notmart has joined #linux-sunxi
<oliv3r>
gzamboni: was just guessing to why you saw a difference :)
JohnDoe_71Rus has joined #linux-sunxi
<gzamboni>
i will dig it :)
<oliv3r>
libv: ok rebooted and no problem yet. I did disable composting in xfce4
vpelletier has joined #linux-sunxi
eyal has quit [Ping timeout: 245 seconds]
Black_Horseman has quit [Remote host closed the connection]
<rm>
composting!
eyal has joined #linux-sunxi
<mnemoc>
sspiff: mozilla will just use the proprietary bootchain from allwinner's sdk
rz2k has joined #linux-sunxi
<mnemoc>
sspiff: we can boot A31 after their boot0/boot1 too
<oliv3r>
kriegerod: sspiff: libv only thing i can imagine that might cause this other then a driver bug; is not enough power; but i can't imagine lightdm/xfce4 crashing but nothing else (as the rest runs fine)
<oliv3r>
mozilla probably will use 3.3 kernel; allwinner default sdk etc
<oliv3r>
libv: nvm, managed to get it to crash again
<oliv3r>
with libs in place
enrico_ has joined #linux-sunxi
<oliv3r>
libv: but now it doesn't crash when i move the mouse, so that's good :p
<libv>
oliv3r: i assume that you tried running es2gears and used cedrus?
<oliv3r>
libv: entered password; started to load, moved mouse whiel it was loading, mouse moved choppyly, and back to lightdm
<libv>
?
<oliv3r>
libv: cedrus works fine; well before limare changes; i'll do cedrus now, if i can get in :p
<oliv3r>
libv: cedrus running right now
<oliv3r>
i'll try running es2gears now in parallel
<oliv3r>
both running
<oliv3r>
90 fps
<libv>
oliv3r: you get crashes while running a sata disk powered off the board, right?
<oliv3r>
only when limare is in place
<oliv3r>
but it's running now fine
<mnemoc>
where does the imx community live on irc?
<oliv3r>
i'll wiggle the mouse for a bit now
arete74 has quit [Ping timeout: 245 seconds]
<oliv3r>
0.75 amps btw
<oliv3r>
0.67 - 0.75 amps
<oliv3r>
but only 90-100fps
<oliv3r>
crash
<libv>
oliv3r: what crashes?
<libv>
oliv3r: X ?
<libv>
oliv3r: look at the Xorg logs
<oliv3r>
i'm tailing the file now; let me wiggle a bit
<oliv3r>
right, this might be harder then i thought :p
<oliv3r>
X server crashes, gets restarted and overwrites Xorg.0.log :p
<libv>
and allwinner has firmly chosen to stick with imagination
<libv>
oliv3r: that's not from a crash, is it?
arete74 has joined #linux-sunxi
<Tsvetan>
I really do not understand what so much secret they will revail if open the registers and settings
<oliv3r>
libv: well X crashed and restarted
<oliv3r>
libv: lightdm restarted
<Tsvetan>
it's after all their hardware implementation which matters
<oliv3r>
Tsvetan: nothing really; it's oldskool fear
<libv>
Tsvetan: noone has so far, so noone will.
<libv>
Tsvetan: these are all embedded companies, secret is their middle name
<libv>
much much worse than the PC world
<libv>
and you know how long it took for the PC world to open up
<libv>
(again, as in the early days it was open)
<libv>
and i daresay that i played some role in that. proper display stuff for intel was trailblazed on via, and without me, ati would've kept completely shut, and intel wouldn't be producing docs today
<oliv3r>
__libc_do_syscall () at ../ports/sysdeps/unix/sysv/linux/arm/libc-do-syscall.S:43
<oliv3r>
Program received signal SIGTERM, Terminated.
<oliv3r>
43 in ../ports/sysdeps/unix/sysv/linux/arm/libc-do-syscall.S
<oliv3r>
so i take it lightdm just gets killed because X dies under it?
rellla has joined #linux-sunxi
<libv>
oliv3r: yes, debugging lightdm is not helping anything
<oliv3r>
i ran it in parallel
<libv>
heh, i just had X die on me as well
<oliv3r>
i let lightdm sit their idling
<oliv3r>
for 5 mins? moved mouse; < 10 seconds, bam
<oliv3r>
i'll try a palin X only server
<libv>
and it just terminated "nicely"
<libv>
how great.
<oliv3r>
how nice?
<oliv3r>
plain X shows a black screen now, not the 'grey' one with the X mouse?
popolon has joined #linux-sunxi
popolon has joined #linux-sunxi
<libv>
oliv3r: thank a radeonhd repo hacking redhat idiot for that
<libv>
oliv3r: you never know whether your driver failed or whether X actually started
<oliv3r>
it looks 'prettier'
<oliv3r>
but so useless
<libv>
it's because they couldn't be bothered to fix up the enabling of the display output properly
<libv>
2007/2008. keithp travels the world telling everyone KMS is the future. "We will all get flicker free booting!!! hurray!!!"
<oliv3r>
i still don't have that :p
<oliv3r>
so
<oliv3r>
X isn't crashing now
<libv>
next opensuse release, some manager comes running in "Help! Help! i think something is ruining my LCD!"
<libv>
and what it was was the leftover crap in the FB showing
newleaves has quit [Quit: Leaving]
<libv>
so now ajax has decide to fix that, with memset.
<oliv3r>
heh
<libv>
and for good measure also disabled the stipple, so you wouldn't see the changeover
<libv>
and now we cannot debug the X server anymore.
<libv>
the story of the self-declared awesome Xorg community
<mnemoc>
what about freedesktop? :p
foubarre has quit [Ping timeout: 252 seconds]
<libv>
mnemoc: you mean that not-for-profit organization that has no chosen board of any kind and which is purely used by keithp as his own private piggy bank?
<libv>
mnemoc: or do you mean the server in some cupboard in portland state university?
<libv>
where admins regularly maim repositories of people they don't like
jinzo has joined #linux-sunxi
jinzo has joined #linux-sunxi
<mnemoc>
*g*
<oliv3r>
so there's no way to debug the X server?
<libv>
oliv3r: there should be, install debugging symbols
<oliv3r>
heh i tried using -cursor (isntead of -nocursor)
<libv>
oliv3r: actually... i think this smells of fbturbo running out of memory or something
foubarre has joined #linux-sunxi
<oliv3r>
libv: i can try giving the fb double the mem?
<oliv3r>
see if it takes longer to crash?
<libv>
it does take long to crash with me
<libv>
i have given it a huge fb
<oliv3r>
i have 16m
<libv>
but now it went
<oliv3r>
for 800x600; what's the minimum?
<libv>
now starting with the binary driver.
<libv>
so i can rule out lima
<libv>
and there it went.
<libv>
resizing windows.
<libv>
1 es2gears, 1 lxterminal
orly_owl has quit [Ping timeout: 272 seconds]
* libv
is going to try again, but with no 3d stuff
paulk-collins has quit [Quit: Ex-Chat]
<libv>
in any case, the blame seems to point at fbturbo
orly_owl has joined #linux-sunxi
<oliv3r>
800x600 requires 2mb fb right?
<libv>
and with the 120fps display sync, the missing buffer invalidation and the flicker at lower fps, i think it's not too far fetched to blame fbturbo
<oliv3r>
ssvb!
<libv>
swrast crashes the server too
<oliv3r>
32mb fb causes equal crashes
<libv>
i had noticed it crashing with screensaver kicking in...
<libv>
before
<libv>
anyway, so much for me trying to fix textured based fbos
diego_r has joined #linux-sunxi
<oliv3r>
but if fbturbo is to blame; why don't i notice it with just cedrus and just limare out of the way
<libv>
oliv3r: did you try resizing windows while using cedrus?
<oliv3r>
i will now
<oliv3r>
should i move liblimare out ofthe equation?
<libv>
oliv3r: also, it might be because memory is allocated from a different pool in fbturbo
<libv>
oliv3r: definitely
<oliv3r>
ok liblimare.so moved; rebootedto make sure
<libv>
move /usr/lib/arm-linux-gnueabihf/mesa-egl out of the way as well
<oliv3r>
'full screen' counts as resize right
<libv>
yes
<oliv3r>
i recall that working; but will test
<libv>
it could even be the fact that with the resize, the cursor changes
<oliv3r>
libGLESv2 and friend
<libv>
right
<libv>
put a dot in front of it
<oliv3r>
what baout lima_dri.so/
<oliv3r>
gone! all gone!
<libv>
it won't be triggered
<oliv3r>
so i should have plain fbturbo only; without any form of egl
<oliv3r>
right, playing sintel; resizing from really small; to really large; full screen
<oliv3r>
keeps working fine
<oliv3r>
minimize, maximize, small window, maximize, full screen
<oliv3r>
i'll let it loop for a few minutes/hours and check back later
<libv>
sspiff: lima is not the bit that is crashing.
<libv>
it happens with swrast.
<libv>
and the mali binaries.
shineworld has quit [Quit: Leaving]
<sspiff>
and it works with lima, or lima is not in a state to test this stuff yet?
<libv>
sspiff: stop whining.
<libv>
sspiff: use the binary
<sspiff>
libv: I'm just asking, sorry!
<libv>
sspiff: there is this website that explains how to do that...
<oliv3r>
libv: mplayer still looping
<libv>
i forgot where it was though...
<libv>
could it be?
<libv>
ah! linux-sunxi.org
<libv>
oliv3r: so it must be something with the dri2 code
<sspiff>
libv: that doesn't really say much about the current state of affairs of lima though, only that it's WIP and not ready yet.
<libv>
sspiff: yes, it is very much wip
<libv>
sspiff: live with it
<sspiff>
libv: also, really nice attitude. I realize you must get a lot of this and are making a lot of useful contributions, and I'm grateful for that, but you really should be a little less aggressive.
<libv>
sspiff: i am just sick and tired of whiny people.
<libv>
sspiff: now head to linux-sunxi.org and read up
prasannapete has quit [Quit: prasannapete]
<mnemoc>
does the fork of the fork of mplayer work with cedrus?
<oliv3r>
manKind: so you found something at cubieboard.org; and come to #linux-sunxi.org for help and not #cubieboard? :p
<oliv3r>
manKind: you did compile your boot.cmd to a boot.scr file right?
<oliv3r>
manKind: or copy pasted the content sans ${extra} and see if that does work?
mankind_ has joined #linux-sunxi
<mankind_>
oliv3r: apologiz but if u can help, u'll be welcome
manKind has quit [Ping timeout: 272 seconds]
<oliv3r>
mnemoc: mpv and mplayer2 work; mplayer1 for ubuntu-arm/debian-arm doesn't have vdpau support
<oliv3r>
does firefox OS run with the android stuff?
<oliv3r>
e.g. android drivers etc?
foubarre has quit [Ping timeout: 272 seconds]
<libv>
i don't know, but i guess they use libhybris like the rest of them
<oliv3r>
porting guide looks a lot like the CM stuff
<oliv3r>
AndroidBoard.mk; extract-files.sh
<libv>
hah
<libv>
stop your distro wars everyone.
rz2k has quit [Ping timeout: 252 seconds]
<libv>
it's all android now.
<diego_r>
Firefox OS is 95% Android
<libv>
thank you jolla, firefox-os, and canonical
<libv>
and who still cares about webos or tizen
<mnemoc>
aka "don't disturb the vendor, lower your head and accept the blobs"
<mnemoc>
oddly most "automative" talks in fosdem are about tizen
<libv>
diego_r: sure
<libv>
"open source libraries"
<libv>
that's a laugh.
<wens>
RILD (listed under open source libraries): It's a proprietary piece of code that's implemented by the hardware vendor to talk to their modem hardware.
<diego_r>
At least Jolla only has that wrapper. FirefoxOS is a real Android under the hood.
<libv>
"oem drivers"
<libv>
"oem libs"
<libv>
the oem drivers sit next to the kernel
<libv>
which then states: "we are perfectly happy if you do an nvidia here"
<libv>
and then.... iggy kicks in, telling me what i should be doing instead.
<diego_r>
mnemoc: Once Meego and now Tizen are ahead of other solutions for IVI (automotive)
<oliv3r>
time to read e-mail; then finish my chapter on my book
<oliv3r>
has to be done by tomorrow!
* wens
cheers
<oliv3r>
yay!
<oliv3r>
why are we happy?
<mnemoc>
oliv3r: Creative Commons? :p
<wens>
oliv3r: because you're finishing the a chapter?
mankind_ has quit [Ping timeout: 272 seconds]
<wens>
seems like lkml.org is down
<Tsvetan>
oliv3r I asked Eva about boot0 for A31 and she reply that she will check why it's not opened
<oliv3r>
oh we're happy about me working :p
<oliv3r>
Tsvetan: awesome; we do have it, just not officially sanctioned
<Tsvetan>
:-)
<oliv3r>
Tsvetan: so we have it illegally :p
<oliv3r>
usermanual same;
<oliv3r>
for a23 we ahve absolutly nothing
<oliv3r>
where a23 is really interesting chip (for tablets)
<oliv3r>
it performs pretty well
<Tsvetan>
a23 is not interesting at all
<libv>
Tsvetan: why not?
<Tsvetan>
no sata no ethernet
<libv>
ouch :(
<Tsvetan>
a13 at least is TQFP and with good power dissipation, A23 is BGA
<Tsvetan>
basically A23 is like A13 but in BGA package
<libv>
seems like i need to move away from working on allwinner stuff then
<libv>
the future is all powervr apparently
<oliv3r>
Tsvetan: a23 is more power efficient (they claim) and is probably 50% faster!
<Tsvetan>
libv rockchip seems to keep Mali
<oliv3r>
but the reason it's really interesting; it uses the a31 memory controller
<Tsvetan>
oliv3r are you sure? A31 memory controller with 64 bit
<oliv3r>
a80 will be powervr, but a60 a70 will be mali
<mnemoc>
rumors tell there is a quad variant of the a23 coming
<mnemoc>
and having mali it sounds like a nice alternative to rk3188
<libv>
oliv3r: i lost that hope
<libv>
oliv3r: if they had not gone with the rogue, i could've been hopeful, but they went with the rogue and are now even deeper in bed with imgtec
<oliv3r>
libv: i think it was confirmed
<oliv3r>
i belive the reason they went with powervr rogue, was as they invested some money in powervr; so they have to recoup that first
<libv>
a former nokia colleague sent me some imgtec swag a month ago, to give me something to have fun with, err, make fun of, for the holidays :p
<oliv3r>
:p
cedric has joined #linux-sunxi
<lkcl>
remember i spoke to allwinner's CTO about 6-8 months ago. they were already well into the A80 so it was too late by then.
<oliv3r>
ah that too
<oliv3r>
and hi lkcl
<oliv3r>
did you get my e-mail?
<lkcl>
but i made it absolutely clear that they would be in serious difficulties to continue to use it
<lkcl>
oliv3r: i did
<lkcl>
the concern that he had was performance. MALI puts a *huge* load on the main ARM CPU.
<lkcl>
i explained that that was due to the software
<lkcl>
so, libv: he _really_ wants to see the results of the work that you're doing.
<libv>
lkcl: thank him for his help.
<libv>
oh wait.
<mnemoc>
lkcl: what about asking allwinner's CTO to sponsor libv?
<lkcl>
i also wanted to put you in touch with him but my associates over-rode me - they want to broker a contract where QiMod Ltd would gain financially from the arrangement.
<libv>
i'll tell my gf that there's a ton of folk who will happily use my work in future, and that she should feel all warm and fuzzy inside.
<lkcl>
mnemoc: yes, exactly.
<oliv3r>
libv is cheap anyway :p
<lkcl>
libv: i know how this works :)
<oliv3r>
*ducks8
<libv>
lkcl: yeah, this is not aimed at you personally
<lkcl>
yehhh
<lkcl>
libv: no i get it. been there. done $80,000 worth of reverse-engineering over 5 months and been offered $6k for it.
<libv>
it seems that there's a load of people making a load of money off of stuff others do for free, but they couldn't be arsed to contribute back, at best you get people whining "is it ready yet?"
<libv>
and that's the ones we should be grateful for
<hno>
libv, that's mostly because it's missing a visibe pricetag.
<libv>
hno: explain
<mnemoc>
lkcl: libv's presentations might get a nice "thanks to AW" logo if you hurry ;-)
<lkcl>
libv: you're mentioning this to someone who's done at least five or six *really* significant strategic contributions to software libre over the past 18 years.
<hno>
libv, it's an off pysological effect. Easier to buy than give.
<libv>
mnemoc: quite a few people in here should get that actually, and we all know that that is not going to happen
<hno>
even if you get more by giving.
<libv>
then there is the fact that they are happily in bed with img
<mnemoc>
tell that to samsung and android :p
<libv>
and that the last thing any of these tiny companies (and that includes mozilla) wants to do is piss off arm and hurt their potential future marketspace
<lkcl>
what hno means is: normally there would be an opportunity for people to appreciate the value of the goods at the time they want them, by virtue of there being an exchange.
<lkcl>
in the case of software libre, that timely exchange is *NOT* present.
rz2k has quit [Ping timeout: 272 seconds]
<oliv3r>
libv: cedrus still running smoothly
<libv>
oliv3r: yeah, it's something to do with the dri2 code
<lkcl>
so the only way - really - to get that back is to sell HARDWARE.... with software-libre pre-installed
<lkcl>
and to work _with_ people who fully understand and appreciate the true value of what you're doing.
<oliv3r>
lkcl: do you have any marketing pics of eoma to put on some slides?
<libv>
lkcl: my issue with that is that improv is about 2ys too late, i know you stuck a lot of your heart and soul into it, but there are too many development boards out there already for the same hw
<lkcl>
who *WILL* honour the implicit contract of paying for the value of what you're doing
<libv>
and most of the people just want to put android on it anyway
foubarre has quit [Remote host closed the connection]
<oliv3r>
you can still have a fully libre system, it being android
<oliv3r>
there's replicant after all
<lkcl>
libv: that statement *immediately* tells me that you've missed the point of EOMA68. and yes, i know it's late.
<oliv3r>
with allwinner chips, should be pretty open
<lkcl>
for example: i'm currently evaluating intel's 37xx series SoCs to put into an EOMA68 CPU Card.
<libv>
lkcl: i get the point of it, and i really hope that you get your 2.5k orders
<oliv3r>
well improv is too late; but eoma a60 or whatever is allwinners next chip, or whatever chip; can still catch up
<sspiff>
oliv3r: most people/companies don't see added value in a system being fully libre though. They don't see how that would improve their earnings
foubarre has joined #linux-sunxi
<libv>
lkcl: but the fact that this order wasn't placed yet tells me that demand is not as great as was hoped
<oliv3r>
sspiff: hopefully the world will see a little change in the next 2-5 years
<lkcl>
oliv3r: _if_ they do one that provides mali and can have its arm twisted into providing SATA and Ethernet.
<libv>
and i do not see aaron making a lot of noise either
<oliv3r>
sspiff: with snowden, NSA etc etc; people slowly start to care about their privacy etc
<lkcl>
libv (and others) then help get the word out! :)
<sspiff>
oliv3r: I fear that won't push the masses in the direction we want them to go - software will always be a black box to them. And I know very few people (outside of the software industry) that associate their vulnerability with software and cloud products and seek to remedy this.
<hno>
oliv3r, I have a feeling most just surrender not even wanting to think of the long term implications.
<oliv3r>
lkcl: a little boring :p but ok i'll put that ona slide :)
<oliv3r>
sspiff: give it time
<oliv3r>
i personally think, that 10, 20 years from now, libre software will be predominant
<oliv3r>
it's the only 'secure' way really
<hno>
always falling back on the idiotic "why should I care? I have nothing to hide from them"
<mnemoc>
lkcl: why intel and not an amd apu?
<lkcl>
oliv3r: yeah i know. use joe's pictures as well.
<sspiff>
oliv3r: it's more secure, but how many free software packages are really audited? And by who?
<oliv3r>
hno: that is true, but recently, i notice a very small shift
<oliv3r>
lkcl: do you have a link? I don't know joe's site :p
<lkcl>
mnemoc: because a) AMD's SoCs run too hot b) AMD still haven't responded to enquiries!
<lkcl>
mnemoc: he's also done a VGA converter board which is awesome
* mnemoc
forgot his nick...
<oliv3r>
is there a dev board for eoma yet?
<mnemoc>
the improv
<oliv3r>
that's it then right?
deasy has joined #linux-sunxi
<lkcl>
there's two. there's the original MEB (which has some faults, but basically can be cut up and made to work)
Coweater has joined #linux-sunxi
<nedko>
lkcl: i really like eoma68 format, and i'l like to buy some cards that are at least a20. but so far it looks to me that getting olimex a20-som is more realistic
rz2k has joined #linux-sunxi
<lkcl>
then there's Improv which is an independent project
<mnemoc>
pretty overpriced for a not-oshw independent project :(
<nedko>
all i want from the board pins is power, i2c and ethernet
<lkcl>
nedko: well, that's down to you. if you want to help make eoma68 a reality, then give our client your order.
* mnemoc
still needs a vesa mountable thinclient :(
<nedko>
lkcl: your client?
<lkcl>
nedko: we don't sell direct unless you want to place an order for 10k units.
<mnemoc>
nedko: the people behind the improv board
<lkcl>
nedko: if you go straight to http://www.vaultechnology.com/store/buyimprov you'll end up at the back of the queue... *after* everyone who has been on the rhombus-tech.net preorders list has been served.
AreaScout has joined #linux-sunxi
<oliv3r>
Includes a dual-core ARM EOMA68 processor,
<oliv3r>
*shiver*
<oliv3r>
improv purchasing information @ vaultttech
<jinzo>
:
<jinzo>
:)*
<nedko>
lkcl: can i preorder only cpu board(s)?
nieuwbie has joined #linux-sunxi
<nieuwbie>
hey. Can someone explain me what's the difference between sun4i and sun4i_sdcon target in u-boot-sunxi?
<oliv3r>
nieuwbie: but that is all from the lichee-u-boot branch; which you only should be using if you need nand support
<nieuwbie>
I bit confused here cause first of all I found dram file which maches my polaroid configuration but during the compilation I didn't specify which dram to use
<libv>
polaroid?
<nieuwbie>
second of all I compiled it without problems but I can't find spl binary
<libv>
because there isn't an spl binary with the lichee-dev branch
<nieuwbie>
libv: Yeah. Its a tablet you could buy in kijkshop (Holland)
<nieuwbie>
Polaroid MIDC
<libv>
nieuwbie: nonsense.
<nieuwbie>
libv: but in tutorial you have to copy spl to sdcard
<nieuwbie>
libv: what exactly?
<libv>
nieuwbie: the ly-f1 page wasn't touched since that New_Device_Howto page came into existence
<oliv3r>
libv: polaroid makes allwinner tablets
<libv>
nieuwbie: or did you decide to be purely selfish
<mnemoc>
wow... ly-f1... the very first a10-based tablet
<nieuwbie>
libv: Excuse me?
<mnemoc>
(available to the west)
<oliv3r>
nieuwbie: you are using the wrong u-boot branch
<libv>
nieuwbie: the new_device_howto states to create a wiki page, do you really need it to tell you to fix up the wiki page while at it?
<nieuwbie>
oliv3r: ok. so which should I use then? I though it's only one.
<oliv3r>
nieuwbie: http://linux-sunxi.org/New_Device_howto follow that guide, and note that your device should be noted as 'simiar to ly-f1' so probably don't have to create an entirely new page :)
<oliv3r>
nieuwbie: there's about 20 u-boot branches, the default being sunxi, you switched to a20-lichee-dev or something
<libv>
there is no other mention of polaroid on our wiki
<libv>
than this page
<libv>
so nieuwbie, you never worked through that page at all
JohnDoe_71Rus has quit [Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org]
<libv>
and you never read up on how to install u-boot
<libv>
nieuwbie: why should we bother to try to help you, if you are not telling the truth?
<nieuwbie>
oliv3r: but how do you know that my board is similiar to ly-f1?
<nieuwbie>
libv: I sense negative energy from you. What happened?
<nieuwbie>
:)
<libv>
nieuwbie: you plainly lied
<nieuwbie>
libv: about?
<jinzo>
following the wiki?
<nieuwbie>
I didn't.
<libv>
nieuwbie: and everyone complains and whines, but no-one bothers to do even the basics described in the new_device_howto
<nieuwbie>
I worked it out.
<libv>
nieuwbie: good, then you can work it out for yourself, somewhere else
<oliv3r>
nieuwbie: did you create a new page for your device
<nieuwbie>
libv: I did the basics. I determined that my dram is already in repository.
<libv>
oliv3r: he never did read anything on the wiki
<libv>
otherwise he wouldn't be having any of these issues.
<nieuwbie>
oliv3r: No I didn't cause I thought it's not necessery cause my config is already there.
<libv>
nieuwbie: good for you. goodbye.
<oliv3r>
nieuwbie: and which boards are associated with those dram settings? probably the ly-f1 :p so you should have at least edited the ly-f1 page and added 'similar to polaroid' :p
<nieuwbie>
oliv3r: let me check it.
<nieuwbie>
libv: lets not get emotional here. Please.
<libv>
nieuwbie: ...
<libv>
nieuwbie: come back when you've done what's described on the wiki, and don't expect any of us to waste time on your questions, questions that would've been answered if you had followed the wiki.
<oliv3r>
nieuwbie: check the makefile which boards use that one
<oliv3r>
nieuwbie: but libv is right, all info should be on the wiki :p
<libv>
oliv3r: don't encourage him.
<sspiff>
libv: you really should try a different rethoric
<lkcl>
nedko: ask our client that question. i can introduce you. remember, they're buying the bulk order from us: we're just a distributor.
<libv>
sspiff: hrm, really? why?
<sspiff>
libv: because it makes you look like an arrogant ass, perhaps? I'm not saying the message is wrong or should change, mind you, just the way you bring it across. Your hostility is not encouraging anyone to follow your advice, I think.
<libv>
sspiff: so you can keep on not reading the fucking wiki and don't feel bad about it?
<libv>
sspiff: ok... so you come in here, and ask a RTFW question...
<libv>
a wiki which i spend quite a bit of my time on
<libv>
and then you go off and join #lima and ask "is it ready yet?"
rings_IIV has joined #linux-sunxi
<sspiff>
libv: You see, most of us read the wiki. Sometimes we don't find shit. Sometimes we look in the wrong place. Being shot down for every question you ask isn't encouraging anyone to make life better for you.
<sspiff>
libv: I went into #lima and asked "can the current lima driver run opengl es 2 software?", all you had to do was say "no".
<sspiff>
And mind you, this information isn't clearly available on limadriver.org or the linux-sunxi.org wiki
<libv>
sspiff: it indeed isn't available.
Black_Horseman has joined #linux-sunxi
Black_Horseman has joined #linux-sunxi
<libv>
sspiff: as i am too bloody busy answering RTFW questions or filling out the wiki
<sspiff>
libv: ignoring our questions is a better policy than what you're doing now, and would cost you less time.
<libv>
and then nieuwbie lies about following the new device howto, and i should not take offence?
<libv>
sspiff: yes, indeed, and the same goes for lima
<libv>
i shouldn't bother.
* libv
sticks iggy on again
<hno>
lkcl, that was a quite annoying order page (VAULTech).
<nieuwbie>
libv: I'm sorry. I have to go to the restroom. Be right back.
<hno>
almost impossible to complete checkout.
<sspiff>
libv: never mind, I see we're not getting anywhere
<sspiff>
keep on keeping on, best of luck with that
<oliv3r>
stop the bickering, you are too distracting for me to write my book!
<libv>
hno: not that i intended to spend money i don't have, but i already frowned when makeplaylive referred to vaultech, i think that that would scare away some of the actual buyers as well
<libv>
hno: but i do not think lkcl is in control of any of that
<hno>
libv, I know he is not in control of that.
<AreaScout>
oohhoo specialist's like sspiff and libv struggle together instead they should work together
<AreaScout>
shouldn't struggle ^
<libv>
...
<oliv3r>
what's sspiff's background?
<sspiff>
I'm no expert on allwinner socs and mali graphics like libv, I have no problem admitting that.
<oliv3r>
sspiff: i'm curious
<oliv3r>
sspiff: no judgement
<sspiff>
oliv3r: my job is an embedded linux developer, though embedded should be taken with a grain of salt most of the time lately.
<nieuwbie>
^^
<sspiff>
I've been using openembedded etc since the zaurus days, and I'd like to get my allwinner boards/tablet running to work on the ARM backend of a specific compiler.
<sspiff>
so I'm mostly a consumer, not a producer here.
<nieuwbie>
oliv3r: My board doesn't look like that of ly-f1. As far as I know.
<oliv3r>
nieuwbie: then it warrants a new page :)
<sspiff>
oliv3r: it's for something other than C/C++ and it's seeing a lot of problems on ARM
<sspiff>
mainly because we don't test much on ARM, yet
<oliv3r>
sspiff: ah; *shiver*
<hno>
sspiff, which language / compiler=
<nieuwbie>
oliv3r: Would you like to see the board yourself?
<oliv3r>
nieuwbie: if you have a pic; great
<sspiff>
hno: opendylan. It's not very well know or anything.
<nieuwbie>
oliv3r: I can make it. Just to make libv happy. ^^
<sspiff>
It's a staticly typed lisp dialect
kivutar has joined #linux-sunxi
<oliv3r>
nieuwbie: great
Black_Horseman has quit [Ping timeout: 264 seconds]
iamfrankenstein has joined #linux-sunxi
iamfrankenstein has quit [Read error: Connection reset by peer]
<hno>
sspiff, odd langauge. And why yet another backend? Isn't one of the gcc or llvm backends suffucuent? Maintaining compiler backends across a number of architectures is a great deal of pain.
foubarre has quit [Remote host closed the connection]
<sspiff>
hno: yeah, I know. We're working on using LLVM as a backend as well, but we'll need hardware to test for those as well, and while that one is unfinished we might want to look at fixing outstanding issues.
<sspiff>
hno: yeah, I know. We're working on using LLVM as a backend as well, but we'll need hardware to test for those as well, and while that one is unfinished we might want to look at fixing outstanding issues.
<hno>
sspiff, do Open Dylan need OpenGL ES?
<sspiff>
hno: no, that inquiry was unrelated.
<sspiff>
although some people are doing OpenGL stuff with opendylan, I'm not.
<libv>
nieuwbie: are you that self-centered and/or egoistic that you do realize that by working through that howto and creating a matching wiki page you are mostly helping yourself, and giving yourself something solid to fall back to in future? Or is the fear of accidentally helping others in the process keeping you from doing that?
ganbold_ has joined #linux-sunxi
<nieuwbie>
libv: No. But I do realize that you have no idea what I'm trying to do right now.
<oliv3r>
nieuwbie: your tablet is the 'inet-3f revision 06 from 2012
<nieuwbie>
libv: Yes I've checked that one.
<libv>
seems like a new page is in order
<libv>
nieuwbie: don't bother with the rev bit though
<oliv3r>
nieuwbie: could you go through the new device howto and create a new page for the inet-3f;
<nieuwbie>
oliv3r: Sure.
foubarre has joined #linux-sunxi
<oliv3r>
btw, that serial port is solderd reasonably nice actually
<oliv3r>
concidering the size of the pads
<oliv3r>
and it has anice big battery
<nieuwbie>
oliv3r: I would do that long time ago but I wasn't sure if it's really needed cause I have tendency to not see thing which are in front of me
<oliv3r>
how's the display? TN? or IPS
<oliv3r>
and what is your final goal?
<nieuwbie>
oliv3r: I have no idea.
<oliv3r>
heh, you can see the missing chips for bluetooth support
foubarre has quit [Read error: Connection reset by peer]
<oliv3r>
a USB hub chip, 1 bt chip and possibly 3rd usb chip
<nieuwbie>
oliv3r: well. its broken. Nande is dying of so I heard so i want to be able to boot it from sdram
foubarre has joined #linux-sunxi
<oliv3r>
boot from sdram
<nieuwbie>
oliv3r: Where I can check what kind of display do I have?
<oliv3r>
well TN displays are really horrible to look at
<oliv3r>
if you tilt it up/down even slightly, the colors are way off
<oliv3r>
to bright/dark
<nieuwbie>
I guess its TN
<oliv3r>
bah :)
<oliv3r>
same as mine then
<oliv3r>
you can even add a 2nd speaker for stereo; the pads are even hooked up :)
<oliv3r>
not sure what '3D is for
<nieuwbie>
Yes. It's quite nicely labeled.
<libv>
inet boards are nice
<libv>
nieuwbie: the two inet_* pages were created in the last month
mouchon2 has quit [Ping timeout: 264 seconds]
<libv>
they are reasonably close to what we currently think is a good device page
<libv>
but start with the example page as described in the howto
rz2k has joined #linux-sunxi
<nieuwbie>
libv: sure thing
atiti has joined #linux-sunxi
<nieuwbie>
There is only one small problem. It crashed I can't boot android anymore. Unfotunately those informations I can't provide anymore.
<libv>
nieuwbie: did you wipe your nand?
<libv>
nieuwbie: or are you talking about identification strings?
<nieuwbie>
libv: No. But system partition is corrupted and I can't repair it.
<libv>
let me fire up my inet k70...
<libv>
ouch
<libv>
i think there are some text files on perhaps system which have those strings... so yeah :(
<nieuwbie>
yeah. my ultimate goal is to replace the nand with hot air soldering.
<libv>
so your nand is broken?
pwhalen has joined #linux-sunxi
<libv>
not just the content, but the actual flash chip
<nieuwbie>
only /system partition
<nieuwbie>
I believe so. But once again I'm an amateur so I can't be wrong.
<nieuwbie>
*can
<oliv3r>
nieuwbie: boot from SD card, and mount nanda; atleast you can obtain your script.bin from there; you allready know your memory timings
<oliv3r>
and memory timings aren't crucial to get perfect anyway
<oliv3r>
you can always tune over/underclock it :)
<nieuwbie>
oliv3r: I already did it with full success.
<nieuwbie>
even backed up for future use.
<oliv3r>
ah so the most important part you have, script.bin
<nieuwbie>
yes.
<nieuwbie>
I made bootinfo and script.fex following the tutorial.
foubarre has quit [Remote host closed the connection]
<nieuwbie>
in FEL mode I mean.
<libv>
well, those identification strings are just there to help people find their device easier and move away from android with more confidence, without them having to open their device
<libv>
so it's not the end of the world if they aren't provided
<libv>
nieuwbie: no, there are plenty of similar devices which just have more or less nand
<libv>
those do not require a separate page :)
<nieuwbie>
oh so simply 8 will do?
<libv>
yup
<libv>
i hope that it is clearer now
<nieuwbie>
well the thing is that I'm doing the wiki page the first time in my life and I want to be most accurate as possible.
<libv>
nieuwbie: accuracy is good
<plaes>
yay.. olimex-a20 ordered..
jinzo has quit [Read error: Operation timed out]
deasy has quit [Quit: Nom d'un quark, c'est Edmonton !]
foubarre has joined #linux-sunxi
MadSpark has quit [Quit: ...]
MadSpark has joined #linux-sunxi
<nieuwbie>
libv: I created the page http://linux-sunxi.org/INet-3f but I'm not happy about the content yet which is mostly empty.
<nieuwbie>
and need to make better pictures. unsolder jump wires and clean it with wik then make better photos of the pads.
<nieuwbie>
but returning to my main problem
<nieuwbie>
I was able to boot linux from sd. the thing is that lcd is not working at all. I checked script.bin and replaced it with that located on nanda but with no luck.
<nieuwbie>
everything except lcd and touch screen works fine.
mouchon2 has quit [Ping timeout: 264 seconds]
<libv>
nieuwbie: work through the manual build howto. you will be guided through the standard uboot build
<libv>
lichee-dev is not going to help you much
<libv>
nieuwbie: what kernel are you using?
<libv>
nieuwbie: also, add support to the individual sunxi repositories, and send in the patches
<libv>
again, work through the manual build howto
<libv>
nieuwbie: the photo of the board is fine
<nieuwbie>
give me a second
<libv>
no need to remove the wires
<nieuwbie>
well probably not but it will look nicer for other people to see how those pads look like
<libv>
they can figure it out
<libv>
it's more important to get other things looking better
<nieuwbie>
Kernel 3.0.52
<libv>
nieuwbie: go through the full manual build howto, please
<nieuwbie>
ok that was my intention
<libv>
nieuwbie: this will hopefully fix all the issues, but it at least gives you a good base to work off from
<nieuwbie>
here's the thing I builded u-boot using sun4i target
<nieuwbie>
and the dram configuration is already there
<nieuwbie>
Does it mean that I have to edit boards.cfg and add new device to it?
<libv>
nieuwbie: yup, as described
<libv>
nieuwbie: on the proper branch of u-boot
<nieuwbie>
but howto doesn'tmetion any branch at all.
<nieuwbie>
well when think about it it metion only main branch
<libv>
nieuwbie: but how did you end up with the lichee-dev branch then?
<oliv3r>
libv: i just read that a23 supports 1/2/4 lane 'mipi'
<oliv3r>
controller is maxed at 1280x800 though (for rgb/lcd or mipi)
<vector80>
oliv3r: How they can do HW acceleration under android ?
<libv>
oliv3r: yeah, i saw mipi back in december
<plaes>
vector80: binary blob
<oliv3r>
vector80: libcedarX.a
<vector80>
plaes: How about their original linux BSP ? That one is also binary blob ?
<oliv3r>
libv: i think that's a new addition to their disp engine isn't it
<oliv3r>
1280x800 isn't a horrible res for tablet/phone either
<oliv3r>
i wonder where this limitation is from though; since their hdmi should still do 1080p
<oliv3r>
vector80: AW basp uses cedarX blob
<plaes>
vector80: yes
<vector80>
So, as I understand, nobody is using linux on AW SoC for either decoding or encoding ?
<oliv3r>
vector80: check linux-sunxi.org/Cedrus
<oliv3r>
but. HIGHLY EXPERIMENTAL!
<oliv3r>
you will burn and die if you ask for support questions :)
<oliv3r>
patches wlecome!
<vector80>
oliv3r: I will try my best as a customer to be able to get some info regarding this
<vector80>
As I understand, above library is just for decoding, not for encoding, right ?
<mnemoc>
libv: lichee-dev exists to provide a fixed u-boot for nand
<mnemoc>
libv: so people running from nand and use a non-allwinner kernel
<mnemoc>
s/and/can/
<oliv3r>
vector80: we have encoding PoC; but not merged yet
<libv>
mnemoc: i know, i spent many days getting that working
<oliv3r>
vector80: you probably get the same 'shit' we allready have
<vector80>
oliv3r: BTW, I have been able to compile latest kernel, and run via NFS on my board...
<vector80>
oliv3r: Where did you get this "shit" ?
<vector80>
From whom?
<oliv3r>
vector80: i think the latest blob-dump was from tsvetsan
<vector80>
Ok, did you see my new linuxBSP in my folder?
<vector80>
Is that still "shit" ?
<oliv3r>
haha, that's the good shit :p
<oliv3r>
u-boot + kernel was very good
<vector80>
Because, as I remember, there was an encoding example even...
<oliv3r>
the binary blob should be somewhere in that pile
<oliv3r>
yeah i think we have that one allready; it's not extremly helpfull; but best as jemk/nove/wingrime how usefull it is
<vector80>
I see
<vector80>
I will try more on this
<vector80>
As I see, only missing thing is that encoding / decoding things..
<vector80>
Seems all other is fine ?
<oliv3r>
technically we have jpeg encoder; and h264 decoder
<oliv3r>
so we have everything :p
<oliv3r>
but more will follow in the next year :)
<vector80>
yes but those decoders are in ALPHA stage, so you we may say.... they are not mature ?
<vector80>
I mean, for example, we can't even encode a simple camera, right ?
<nieuwbie>
seriosly I have no idea how to configure the boards.cfg
<nieuwbie>
I have like arch cpu soc vendor Im getting confused when comes to board name but I guess it would be INet_3f no idea about target and options
<oliv3r>
vector80: not yet; you need allwinner code for that
kondratiev has joined #linux-sunxi
<vector80>
oliv3r: clear, so you mean, I can't use that encoding example in their BSP..
<oliv3r>
the example should work
<vector80>
Alright. I will give a try on that.
<oliv3r>
good luck with that :)
<oliv3r>
see ML for more info
<vector80>
Sorry, what is ML ?
<plaes>
mailing list
<vector80>
ok
<vector80>
thank you, I am really bad on such short names, sorry again.
<vector80>
oh, I forget to ask something...
<vector80>
How about the audio?
<vector80>
Encoding / decoding ... You are doing all this via SW enc/dec ?
Marin-Cro has joined #linux-sunxi
Marin-Cro has quit [Read error: Connection reset by peer]
Marin-Cro has joined #linux-sunxi
predlog has joined #linux-sunxi
predlog has quit [Ping timeout: 252 seconds]
kondratiev has quit [Ping timeout: 272 seconds]
Marin-Cro has quit [Remote host closed the connection]
<nove>
vector80: is simples, with binaries blobs you are dependent of what Allwinner gives, as those binaries are not integrated with openmax/ffmpeg/gstreamer, you will have to do this work
<nove>
vector80: was already was found, the hardware is simples and there is a active Reversing engineering effort going that is gives results
<HdkR>
Some places claim that you can decode MP3 files full speed with NEON at 10mhz or so :P
<buZz>
man, i remember my old 486, had to drop down to DOS to decode MP3s without lowering bitrate :D
<nove>
vector80: so you have to make a choose, spend time making the binaries works without knowing what allwinner will release next
<buZz>
or stick to vdpau and at least get something portable in the end ;)
<nove>
vector80: or to take the more future proof way, and use the time instead by working in making open source drivers possible
pwhalen has quit [Ping timeout: 264 seconds]
<nove>
vector80: so with binaries or with the open source, it ends to be almost the same amount of work
<nove>
vector80: by one you will be free to do what you want, with the other you have to ask allwinner
<vector80>
nove: , buZz, thank you very much both
<vector80>
I think this is the responsibility of AW, they have to do this, or at least the must give us some clue on how to do this..
<vector80>
Anyway, thank you very much for being very clear
<vector80>
Good evening, bye
vector80 has quit [Remote host closed the connection]
<libv>
oliv3r: did you have libMali or swrast running before, on fedora?
<libv>
oliv3r: or were you just using xf86-video-fbdev ?
slonesja has joined #linux-sunxi
slonesja has quit [Remote host closed the connection]
fredy has quit [Excess Flood]
mouchon2 has joined #linux-sunxi
fredy has joined #linux-sunxi
edikkk has joined #linux-sunxi
h13rrr has joined #linux-sunxi
massi_ has joined #linux-sunxi
kso has joined #linux-sunxi
h13rrr has quit [Ping timeout: 252 seconds]
h13rrr has joined #linux-sunxi
edikkk has quit [Ping timeout: 252 seconds]
kso_ has joined #linux-sunxi
burmat has joined #linux-sunxi
burmat has left #linux-sunxi [#linux-sunxi]
kso has quit [Ping timeout: 265 seconds]
vpelletier has quit [Ping timeout: 264 seconds]
Quarx has quit []
vpelletier has joined #linux-sunxi
<nove>
a crazy idea, lets we sunxi officially send a email to mozilla, asking if they want firefox os run in those generic tablets, we would not do the work from one side, but together, and mozilla has to also push allwinner to contribute
<nove>
as alllwinner does the software for those tablets, they could also start to offer firefox os
lioka has quit [Quit: leaving]
<Turl>
(backlog mode: on)
<Turl>
gzamboni: how did you specify the irq?
<Turl>
oliv3r: lol composting
enrico_ has quit [Quit: Bye]
bbrezillon has quit [Quit: Ex-Chat]
<Turl>
wens: rild is open. It just happens to be only useful to dlopen a blob and run code from it :p
<gzamboni>
Turl, yes like this: interrupts = <26>; in the dtsi file, i added a subnode for the dma, i'm playing around with the dma code from the sdk's linux 3.4 kernel of a23
<gzamboni>
the reg = <0x01c02000 0x1000>; and the clocks = <&ahb_gates 17>; seems to work fine
<gzamboni>
Turl, it seems its getting the next no used irq
notmart has quit [Quit: notmart terminated!]
lioka has joined #linux-sunxi
fredy has quit [Excess Flood]
fredy has joined #linux-sunxi
kso_ has quit [Ping timeout: 252 seconds]
kso has joined #linux-sunxi
netlynx has joined #linux-sunxi
netlynx has joined #linux-sunxi
popolon has quit [Quit: Quitte]
<Turl>
gzamboni: is this on sun4/5i or sun7i?
techn_ has quit [Remote host closed the connection]
<Turl>
gzamboni: sun7i uses the GIC so interrupts look like triplets
Fusing has joined #linux-sunxi
<hramrach>
Turl: compositing works, mostly
* hramrach
was running gnome3 on an a10
<Turl>
hramrach: go read again
<gzamboni>
sun4i
<gzamboni>
Turl but i havent digged it out yet, i will continue tomorow
<RaYmAn>
hramrach: was reading some logs - did you ever get anywhere with sun3i? I got myself an inkcase(.com) which uses the F200 chip as well. Unfortunately, there's no way to put software on it to dump it (other than FEL mode which works)
<Tsvetan>
Turl we have 4 SOMs which one you asking for?
<Turl>
Tsvetan: A20 specially :)
<Turl>
Tsvetan: someone was asking here yesterday
<Turl>
Tsvetan: or maybe in #olimex, don't remember
<Turl>
it is not listed on your site
<Turl>
RaYmAn: no sd slot?
<Tsvetan>
sure we use them now only in custom projects
<RaYmAn>
Turl: indeed
<Turl>
RaYmAn: :(
<Tsvetan>
but will list them soon on the web too
<RaYmAn>
Turl: it has a bluetooth api and a usb otg port - and that's about it
<Turl>
RaYmAn: you'll have to use fel then
<RaYmAn>
Turl: though, I took it partially apart and there does appear to be an UART port
<Turl>
RaYmAn: what's inside, nand?
<Turl>
Tsvetan: :)
<RaYmAn>
Turl: hm, I didn't look that carefully, but I guess - I'm a bit afraid to take it out of the case because of the eink screen - don't want to ruin it :P
<Turl>
Tsvetan: I found one distributor(?) selling them when looking in google
<Turl>
Tsvetan: so I was surprised it wasn't on the site :)
<Turl>
RaYmAn: :)
<Tsvetan>
Turl none exept us have them
<Tsvetan>
this slovakian webshop have none in stock
<gzamboni>
if i comment the interrupts = <27>; line in dtsi i get an error when loading up the driver
<ccaione>
irq_of_parse_and_map expects an interrupt = <> in dts
<ccaione>
what's the problem?
<gzamboni>
it returns always 22, i supose it should return the irq number that is specified in my dts ( 27 )
kivutar has quit [Ping timeout: 246 seconds]
<ccaione>
gzamboni: the int 22 you get is the virtual irq that could be mapped the any hard irq
<gzamboni>
humm, thats what i imagined, i was just looking into the irq.c that it returns: return irq_create_of_mapping(&oirq)
<gzamboni>
so its good, i can continue :)
<ccaione>
it usually depends on .xlate hook of the irq_domain_ops struct
<ccaione>
(and .map)
FR^2 has quit [Remote host closed the connection]
kso has quit [Ping timeout: 246 seconds]
<gzamboni>
gotta study more all that
rellla has joined #linux-sunxi
<oliv3r>
libv: xf86-video-fbdev i guess? maybe swrast but i never tested any 3D; i didn't install fbturbo nor had that
<oliv3r>
libv but ubuntu works 'fine' without 3d
<oliv3r>
Turl: rild has an opensource variant in replicant :)
<Turl>
oliv3r: rild is opensource on aosp too
<Turl>
it's just the lib it loads that isn't
<oliv3r>
Turl: samsung's libril has been RE'ed then
<oliv3r>
paulk-collins probably know best how to explain :)
<Turl>
yeah that's likely :)
<oliv3r>
nove: my decoding isn't working flawlessly anymore
<oliv3r>
nove: after 3 or 4 loops (i don't know when exactly, first i think it was after screen blanking) decoding turns slow; like 1fps and the next itteration of -loop 0 stays that way, until i restart the video
<oliv3r>
nove: it MIGHT be an mplayer2 bug; as it worked with mpv; but i don't know
<oliv3r>
Turl: i belive it was the layer between hardware, and rild
<paulk-collins>
let's roll back
<oliv3r>
paulk-collins: the man that knows the magic
vpelletier has quit [Quit: leaving]
<mnemoc>
oliv3r: it seems all the blame of my slugish system running spinning storage goes to thunderbird
<Turl>
(that's a like 1cm x tablet length display. not really much you can use it for)
<hramrach>
the missing statusbar you never want to take up space on the main display
<hramrach>
still 1280x800 ..
<hramrach>
same as the I ?
<hramrach>
they could up that a bit. OR sell for really cheap but I suspect they do not
ssvb has joined #linux-sunxi
tomboy64 has joined #linux-sunxi
<ccaione>
Turl: for babelfish isn't it better to put them in dts? (no answer on #armlinux yet)
<Turl>
hramrach: it's like 260 usd (in indian rupees)
<Turl>
(wifi only)
<oliv3r>
ssvb: the crashes seem to stem from dri2/fbturbo and some memory issue; libv can describe it better
<Turl>
ccaione: I think the real question here is how many people actually care about such option
<Turl>
ccaione: imo it isn't a hardware characteristic so it shouldn't be on dt
<ccaione>
Turl: honestly I care. It's annoying to wait 8s for shutting it down
<oliv3r>
it's a configurable isn't it
<Turl>
ok then :)
<oliv3r>
is it stored inside the axp's flash? e.g. will it survive power cycle
<oliv3r>
4+ seconds :)
<oliv3r>
though you should never have to do it
<ccaione>
oliv3r: let me check if it survives (I have no battery)
<oliv3r>
it should; otherwise it won't help greatly
<ccaione>
yep, it is persistent
hunterhu has joined #linux-sunxi
orly_owl has quit [Read error: Connection reset by peer]
orly_owl has joined #linux-sunxi
<ccaione>
Turl: 2 vs 1, ok for sysfs :)
<hunterhu>
hello, I am trying to make a not so popular solomom tech ssd2533 touchscreen working on a iview435tpc tablet; the linux kernel driver seems working fine and reporting touchscreen x,y,z data, but I still cannot use it after log into the ubuntu desktop. Does anyone have some idea what needs to be done after kernel driver working in order to have a fully functioning touchscreen?
<Turl>
ccaione: worst case you'll have to rewrite that bit
<Turl>
:p
<Turl>
hunterhu: maybe some x config missing
<hunterhu>
any idea how to do that?
<ccaione>
Turl: BTW, do you have comments on my irqchip driver? I'm waiting for mripard_ be back from holydays :)
<hunterhu>
ubuntu seems have a easy calibration tool, but seems not quite working on the linaro-quantal-rootfs image, does anyone have similar experience before?
<RaYmAn>
doneill: the sinking ship? ;) Nah, still waiting to get my adam2 any day now ;)
<Turl>
ccaione: let me have a look :)
<RaYmAn>
Turl: it's not an eink spine display. It's uhm, STN or something - same tech as those good old HD44780-based character lcd displays
<doneill>
i dropped my adam, lost interest after that.
<RaYmAn>
heh
<Turl>
hunterhu: dunno
<RaYmAn>
did they ever publish what SoC they use?
<doneill>
in the 2? i haven't heard.
<Turl>
ccaione: v3 right?
<RaYmAn>
(after they scrapped using OMAP.)
<ccaione>
Turl: yes
<RaYmAn>
Turl: it's such a pity. NI has some great ideas, both hardware and software, but execution never seems to quite work out :P
<RaYmAn>
doneill: they promised me an Adam2 back in the day, so it's only cause it's free i'm getting it ;)
<Turl>
ccaione: other than what maxime pointed out, it looks ok to me
<Turl>
ccaione: but I'm just talking 'code' wise, I cannot comment on actual irq usage and the like, I'm not familiar with that at all
<ccaione>
Turl: thank you :)
<Turl>
ccaione: also, it's not a big deal, but I think it'd be better if documentation came before the actual usage
<ccaione>
Turl: you mean as patches order?
<Turl>
yeah
<Turl>
it makes more sense to me to first document it and then use it
<ccaione>
got it. fix in v4
<hramrach>
hmm, the test leads for SMD parts are really expensive, at least the ones I found
<hramrach>
so much for working around unknown/broken OS in the flash
<doneill>
rayman, that's a pretty sweet deal.
netlynx has quit [Remote host closed the connection]
<hramrach>
RaYmAn: I was thinking of making use of the Adam I
<hramrach>
but it's really heavy
<mnemoc>
the weight it's the worst problem
<hramrach>
and there is no current OS for it. not even android. and the camera does not work with anything but the ancient original ROM and accel does not work with anything but the beta rom. And without neon the accel is needed even for drawing the UI on the larger than 4" screen :/
doneill has quit [Quit: Page closed]
<hramrach>
well, it has HDMI and big battery so you might use it as semi-mobile since it's too heavy for mobile
<hramrach>
but still no OS
djmaestro has quit [Ping timeout: 272 seconds]
kso has joined #linux-sunxi
jinzo has joined #linux-sunxi
jinzo has quit [Changing host]
jinzo has joined #linux-sunxi
Black_Horseman has joined #linux-sunxi
Black_Horseman has quit [Changing host]
Black_Horseman has joined #linux-sunxi
nieuwbie has quit [Remote host closed the connection]
paulk-collins has quit [Quit: Ex-Chat]
FunkyPenguin has quit [Quit: No Ping reply in 180 seconds.]
FunkyPenguin has joined #linux-sunxi
rellla has quit [Remote host closed the connection]
<WarheadsSE>
Q: cedarx RE effort.. there a clean package for that yet?
<WarheadsSE>
And still no sane answer I have recv'd regarding if everyting on the A20 is supported in sunxi-3.4
<WarheadsSE>
I'd like to get off that crappy 3.3 POS.
<hramrach>
WarheadsSE: no CT WiFi afaik
<Turl>
WarheadsSE: most stuff people care about is supported
<hramrach>
but a20 itself is supported
<WarheadsSE>
Yeah, i know the A20 itself is
<Turl>
main issues are with wifi and maybe some of the less used stuff
<WarheadsSE>
the entire sun{4,5,7} pile is due to bump
<WarheadsSE>
And I never recieved a CT, so
<Turl>
WarheadsSE: there's libvdpau-sunxi for the cedar stuff btw
* WarheadsSE
shrug
<Turl>
WarheadsSE: but it's experimental
<WarheadsSE>
Turl: K. So? :P
<WarheadsSE>
so was the sunxi kernel for a long while
<hramrach>
well, Cedar support in anything is experimental ..
<Turl>
WarheadsSE: maybe oliv3r can tell you how usable it is
<Turl>
he's been testing it today
<hramrach>
vdpau should work with mplayer2 or the next mplayer fork .. mvp?
popolon has joined #linux-sunxi
popolon has joined #linux-sunxi
<WarheadsSE>
So, sane to pull a build @ sunxi-3.4 HEAD for sun{4,5,7}i ?