<techn__>
Turl or anyone: have you tried to compile every ts screen in?
<Turl>
techn__: I tried to bundle them some time ago
<Turl>
but mine is not supported so idk if it works or not :)
<techn__>
ok.. I have device with gt811 and another with ft5x_ts .. they seem to work together when modprobed simultanious
<bsdfox>
anyone know how I can change the bus speed of /dev/i2c-1? I get write errors early on if debug isn't enabled (debugging appears to slow it down a lot and writes fine reliably)
<bsdfox>
it looks like the chip says i2c-1 and i2c-2 support 200kHz but mine is running at 400kHz
<Turl>
let me check the clock docs
<Turl>
bsdfox: does i2c have some other name? I don't see it on a quick look
<Turl>
bsdfox: twi, can it be?
<Turl>
twi is gated from apb1 which is 24Mhz
<bsdfox>
Turl, yeah twi
<bsdfox>
that'd be wayyy too high
<Turl>
yeah it's divided :)
<Turl>
bsdfox: it all traces back to the speeds defined on arch/arm/mach-sun4i/include/mach/i2c.h
<Turl>
that's used to configure the divisors to make it match those freqs
<Turl>
1 and 2 should be running at 200 and 0 at 400
<bsdfox>
yeah that's what it looks like to me. I'll have to check it at my office next week. I wasn't getting a good trigger at home
<bsdfox>
there are lots of i2c devices that can't handle 200kHz though.. I wonder how we limit the speed for these
<Turl>
they controller supports 100khz too
rz2k has joined #linux-sunxi
<Turl>
so you just need to lower the clock :)
<bsdfox>
I'm thinking there is a way to do it in userland
<Turl>
no idea
<Turl>
bsdfox: mripard was working on a new i2c driver, he may know
<bsdfox>
ok I iwll ping him
<bsdfox>
thanks
<bsdfox>
hmm maybe it's not a speed issue. hard coding it to 100kHz did slow it down but writes still fail
<bsdfox>
yeah strange.. it's reproducable that writes work reliably with debug enabled
<Turl>
damn Heisenbugs :P
<libv>
bsdfox: the it is timing
<libv>
then even
<bsdfox>
I'm looping reads right now and watch scl/sda on a scope and the signals look good
<bsdfox>
so I don't think it's hardware related
egbert has quit [Disconnected by services]
egbert has joined #linux-sunxi
techn_ has joined #linux-sunxi
techn__ has quit [Ping timeout: 240 seconds]
torqu3e has quit [Quit: torqu3e]
ZaEarl has quit [Ping timeout: 264 seconds]
wingrime has joined #linux-sunxi
torqu3e has joined #linux-sunxi
<wingrime>
anyone tested pm patch ?
<wingrime>
mnemoc: are you here
hipboi has joined #linux-sunxi
rellla has joined #linux-sunxi
<wingrime>
trul: are you here ?
<rz2k>
wingrime: dude, its sunday noon, everybody drinking beer or preparing for it. also one mention of your patches is enough :p who wants to test will test when they will have time.
<rz2k>
не надойдай тут мне!
<wingrime>
rz2k: I have find bug in disp
<rz2k>
great, report it to ML
<wingrime>
it simply ask direcly ssvb or trul or techn
paulk-desktop has joined #linux-sunxi
shineworld has joined #linux-sunxi
<paulk-desktop>
hi
<paulk-desktop>
Amongst all the A1x boards (Cubieboard, hackberry, gooseberry, A13-OLinuXino), do some have a fully free software boot process
<paulk-desktop>
ie: is uboot not chainloaded by a non-free 1st stage bootloader on some boards?
<wingrime>
mmc boot
<paulk-desktop>
what about it?
<wingrime>
any board can boot using mmc card
<paulk-desktop>
is it possible to use uboot as first bootloader with mmc boot?
<paulk-desktop>
so then I guess it would be possible to flash u-boto to NAND and use it as first bootloader as well
<paulk-desktop>
u-boot*
<wingrime>
I may be have outdated information, but nand have cosed source part
<wingrime>
for boot
<wingrime>
only mmc have full opensource boot
n01 has joined #linux-sunxi
<paulk-desktop>
ah ok
<paulk-desktop>
that's why everyone is using mmc
<paulk-desktop>
but does nand work from kernel?
<wingrime>
yep
<paulk-desktop>
ok, then I could build a free kernel and userspace, flash it to nand and use the non-released uboot to boot it
<wingrime>
paulk: you not understand me fully , uboot can boot from nand , but it need some "first stage loader" that not opensource
eebrah has joined #linux-sunxi
<paulk-desktop>
wingrime, like to do some early initialization that uboot can't do?
<wingrime>
yes, you have init DRAM at least
<paulk-desktop>
ok well, I'll see
<paulk-desktop>
thanks
<rz2k>
<paulk-desktop> but does nand work from kernel? - from kernel yes, but our u-boot cant boot from it
<paulk-desktop>
okay
<rz2k>
you still need uboot on sdcard
<paulk-desktop>
that's a shame
<rz2k>
slapin & hno was working on it
<paulk-desktop>
but to get things clear: does that mean that the free uboot cannot be booted from nand or that it cannot boot a kernel stored on nand?
<paulk-desktop>
ie: does it start when flashed on nand
<rz2k>
uboot cant be stored in nand and booted from it
<paulk-desktop>
ok, but it can boot a kernel stored in NAND if it's booted from sdcard?
<rz2k>
and obviously it cant read nand, so you need to place the kernel somewhere else
<paulk-desktop>
ah ok
<rz2k>
when kernel is loaded you can access nand
<rz2k>
it has driver for it
<rz2k>
we have lichee-dev branch of u-boot that can be loaded from nand, this is bare u-boot from sourcedrop, it has ton of bugs and eventually all nand stuff was removed for some reason
<rz2k>
but this makes the boot process not a gpl one
<rz2k>
because A10 loads the boot0/1 from nand first and then loads u-boot
<paulk-desktop>
ok I see
hipboi has quit [Remote host closed the connection]
<ol1ver>
Turl: what ts dop you have thats unsupported
<ol1ver>
paulk-desktop: in the future, we do hope to make u-boot support nand boot
<ol1ver>
paulk-desktop: just not there yet
<paulk-desktop>
nice
<ol1ver>
first, we'd need to replace boot0 with u-boot-SPL; right now, u-boot-spl (as mentioned above) can only boot mmc (it only has mmc driver)
<ol1ver>
once that has Nand support; we could build u-boot-SPL with nand support and once that is possible, we're golden ;)
<ol1ver>
boot1 can be replaced by u-boot (without-spl) and that could then boot anything
<rz2k>
lets boot qnx and fly to the moon.
n01 has quit [Ping timeout: 245 seconds]
<wingrime>
Trul: are you here ?
<mnemoc>
techn_: pushed a fix to soc-detect. sure you didn't see any "Allwinner A13 revision X" or similar before BROM Ver. or immediatelly after?
<wingrime>
mnemoc: What about "my pm" patch?
<wingrime>
mnemoc: I tested if myself
<wingrime>
*it
<wingrime>
no problems for me (except known bugs) , as code 100% same it must work
<wingrime>
mnemoc: can I merge "devices.c" to plat ?
n01 has joined #linux-sunxi
<mnemoc>
wingrime: haven't had time to look at the ML. but unless something is a bugfix I give it a couple of days for people to comment before merging to the stage branches.
<wingrime>
mnemoc: ok, but I still have old patches in ML not merged
<mnemoc>
wingrime: remind me of those monday evening please
<wingrime>
mnemoc: ok
<wingrime>
mnemoc: I can merge "devices.c" to single one
<wingrime>
mnemoc: what do you think about it
<mnemoc>
go ahead :)
<wingrime>
mnemoc: last demand dma,clock , includes
<wingrime>
mnemoc: If I finish this you can rename plat-sunxi to march-sunxi and remove sun4i and march-sun5i what do you think about it
<mnemoc>
slow it down :p
<mnemoc>
we still need to merge mach-sun6i and mach-sun7i ;-)
<mnemoc>
and we still need a CONFIG_ to tell what sunNi is
<wingrime>
we need hardware first
<mnemoc>
let's keep it with plat-sunxi for a while ;-)
<mnemoc>
but mach-sunNi minimal
<wingrime>
mnemoc: it not difficut to move dma and devices. clock need still investegate
<wingrime>
mnemoc: I not prefer CONFIG_ i prefer dynamic detection for single kernel for a13 an a10
<wingrime>
using machine_is_sun4i()
<mnemoc>
i didn't say I prefer CONFIG_, it's just that they make unification simpler
mdfe has joined #linux-sunxi
<mnemoc>
once drivers are unified we can switch to machine_is_sunNi() and the new (in my soc-detect branch, still under testing) sunxi_is_{a10,a13,a12,a10s}()
<mnemoc>
but first we need to unify the different mach-sunNi and sunNi-foo drivers
<mnemoc>
(including sun[67]i)
rellla has quit [Remote host closed the connection]
<wingrime>
mnemoc: are you still here ?
<wingrime>
I have question about mali
<wingrime>
[ 19.400000] Mali: mali clock set completed, clock is 180000000 Mhz
<wingrime>
^ ^
<wingrime>
witch means 180 EHz - gamma rays ?
<techn_>
mnemoc: If that is new print.. no I didn't see it
<techn_>
mnemoc: how hard it would be to add ATAG support to lichee-dev or should it be there?
<techn_>
but missing from boota command?
<wingrime>
techn_: I have strange bug when x11 in 16 bit mode after resume it not returns to 16 bit and I see two same small view on screen
<wingrime>
techn_: go to console than back fixes it
<techn_>
wingrime: sounds like it goes to scaler mode after resume
<techn_>
or maybe goes to 32bit mode on normal layer
<wingrime>
techn_: try repeat bug
<wingrime>
techn_: config xorg.conf to 16 bit color
<wingrime>
techn_: than echo "mem" > /sys/power/state
<wingrime>
techn__: will you join g2d to one for sun4i and sun5i ?
<techn_>
wingrime: there is no g2d in sun5i ?:/
<wingrime>
sorry I mean cedarx
<techn_>
not in plans
hramrach has quit [Remote host closed the connection]
<ol1ver>
Turl: are u-boot enviroment variables in any way usefull to know why bootimg booting fails? for example boot_normal=nand read 4000780 boot; boota 40007800
<mnemoc>
techn_: lichee-dev *does* support ATAG. stock doesn't. that's why as people to replace stock with lichee-dev
<techn_>
mnemoc: I tried to remove that line.. it worked without ATAG hack
<techn_>
but
<mnemoc>
A13 doesn't have g2d, but all the other socs do
<mnemoc>
including A10S (also sun5i)
<techn_>
mnemoc: "[ 0.000000] Ignoring unrecognised tag 0x00000000" with aboot
<techn_>
but it works though
<mnemoc>
techn_: I don't understand why you don't see the "sunxi: Allwinner A13 revision B" line
<mnemoc>
techn_: that one is normal
<mnemoc>
just wrong end-of-atag sentinel
<techn_>
ok.. great.. I'll send patch for atag support for boota
shineworld has joined #linux-sunxi
<Turl>
ol1ver: GT801NI
<Turl>
ol1ver: I don't understand your other question
<mnemoc>
techn_: can you try again removing the (void) in core.c's sw_core_map_io() ?
<techn_>
mnemoc: ok
<ssvb>
wingrime: the screen split into two greenish halves indeed looks like the framebuffer working in 32bit mode, but X server treating it as 16bit
<mnemoc>
techn_: and updated branch obviusly :)
<techn_>
mnemoc: yes
<ssvb>
wingrime: I have seen it once on a10, but then it disappeared after kernel upgrade, maybe I need to try running some more tests
<techn_>
mnemoc: I'm trying to make mk_livesuit.sh work with linux rootfs.. I have problem that it wont launch init from correct root :/
<mnemoc>
techn_: and nothing of the sort "sunxi: Allwinner A13 ..." ?
<techn_>
mnemoc: no Allwinner print
<mnemoc>
wtf
<mnemoc>
techn_: might it be the const attribute? deef9d8f2cd98c456ba03aeea1201ba61bef485f
<techn_>
cant understand how const functions work with c :p
<mnemoc>
if you get the Allwinner line it will mean I don't understand it either
<Turl>
what are you discussing guys? :)
<mnemoc>
Turl: wip/sunxi-3.0/soc-detect in my github
<mnemoc>
sunxi_is_a10(), _a13(), _a10s(), ....
<mnemoc>
for soc dependent code
<mnemoc>
but somehow it's not printing when it should....
<wingrime>
ssvb: try suspend using console
<wingrime>
ssvb: I explaned it in backlog
rz2k has quit []
<Turl>
mnemoc: where is this stuff called?
ganbold___ has joined #linux-sunxi
ganbold__ has quit [Remote host closed the connection]
<Turl>
nvm, found it on core.c
<mnemoc>
Turl: sw_get_ic_ver() itself is called in all places where the soc revision matters
<mnemoc>
I'm extending it to support different sunNi in the same code
<ol1ver>
mnemoc: i've added the stock nanda for my binary to my local clone of sunxi-bin-archive; is it ok if i push those directly? No point in having those go over the mL
<mnemoc>
and the first time it's called it should print a message telling what SoC/rev was found
<mnemoc>
ol1ver: sure, go ahead
<ol1ver>
Turl: I'm still trying to figure out why my u-boot won't boot my kernel. So I went over the u-boot enviroment partition and saw (what i posted up). Could it be that those address are somehow specific to my tablet and that's whats causing the kernel not to load? (i know i know, uart)
<mnemoc>
Turl: i need it for the gpio_eint code, but sunxi_is_foo() should be useful in other parts as well (clocks?)
<ol1ver>
mnemoc: i think i may not have access to that sunxi-bin-archive though
<ol1ver>
i can push it to my own cgit and you can pull it from there if you want
<Turl>
ol1ver: it looks pretty generic, compare with what Epsylon3 uses
<Turl>
mnemoc: clocks are per sunNi so no need for macros
<mnemoc>
Turl: sure A10S and A13 clocks are identical?
<mnemoc>
I've not seen anything merged from the "elite" (A10S) repo...
<mnemoc>
ol1ver: try now
<mnemoc>
Turl: anyhow, if you can try it on real hw it would be awesome
<ol1ver>
works
<Turl>
mnemoc: I didn't see those
<Turl>
mnemoc: but sun5i's are mostly a subset of sun4i's
<ol1ver>
Turl: then i simply need uart; and i'm gonna order one when cubieboard a20 comes out :p
<ol1ver>
mnemoc: done pushing
<mnemoc>
ol1ver: :)
ganbold___ has quit [Remote host closed the connection]
ganbold_ has joined #linux-sunxi
n01 has quit [Ping timeout: 260 seconds]
ganbold_ has quit [Remote host closed the connection]
<ol1ver>
mnemoc: ok; i've re-thought the whole generic fex file thing, and got a solution :p I'll post it to the mailinglist in a moment; just one thing
<ol1ver>
mnemoc: do we want patches on a per 'clone' basis (which will contain camera, touchscreen, battery configuration)
<ol1ver>
mnemoc: or have patches for a touchscreen, camera and battery seperated
<ol1ver>
mnemoc: to which the 'clone' will be a patch set, standard board + camera module.patch +touschscreen.patch +battery.patch
<ol1ver>
the latter makes more sense to me; since many clones can just have various combinations; the only thing tricky of course, is the battery
<ol1ver>
mine for example is 'unbranded' (unless the stickers are at the bottom)
<mnemoc>
the .fex should belong to a real device, and one .patch per variant
<mnemoc>
clones qualify as variants
<ol1ver>
mnemoc: well in the case of my tablet, the PCB, 'THE' tablet is standard amongst a few clones
<ol1ver>
the touch-sensor, camera module and battery are all three external
<ol1ver>
clone == variant :p
<mnemoc>
:)
<mnemoc>
need to go
<ol1ver>
mnemoc: ok
<mnemoc>
back in 2~h
<mnemoc>
~2h
<mnemoc>
i use the term variant because a cubieboard 512 is not a clone of the 1GB model ;-)
<mnemoc>
bbl
<ol1ver>
true, and then the issue becomes interesting; the cubieboard.fex would be for the original 512; the 1GB would be a rev2 .patch I suppose
<ol1ver>
but what if 2 variants would emerge from 2 vendors, one with a cam1 module, the other with a cam2 module; have 2 patches, for each variant 1, fine. but then a 3rd vendor comes in, and also has cam1; is that a new vendor3.patch? it won't ever be pretty, rest assured :p
n01 has joined #linux-sunxi
rellla has joined #linux-sunxi
<ol1ver>
i've noticed that if I put usb_used = 0 under usbc1; usbc2 won't come up on hansg's fedora 18 image. Did the usb controller code change to stop checking if it finds the first disabled port?
<wingrime>
It realy strange ..... that manualy debootstraped debian armhf works worser than linaro image, xorg behaves strange with touchscreen, it fells toouch but only at coursor and not moving it, hung on resume
<wingrime>
what patches have linaro?
<drachensun>
I think I have seen the hung touch screen on Linaro resume
<drachensun>
there will be an error on the screen saver "mouse pointer already captured" or something like that and it wont work afterward until X is restarted
<drachensun>
or a reboot of course
<drachensun>
has anyone else had that happen?
<wingrime>
for me linaro works good , but clear armhf .....
rellla has quit [Remote host closed the connection]
paulk_ has joined #linux-sunxi
paulk-desktop has quit [Ping timeout: 264 seconds]
<Turl>
ol1ver: why .patch?
<ol1ver>
Turl: i think mnemoc wants to work it into sunxi-bsp; that the patch + original fex will create a 'variant'.fex or something
<ol1ver>
wingrime: including GUI?
<wingrime>
ollver?
<ol1ver>
wingrime: i've tried the linaro alip images, but couldn't get a gui running, Xorg failed to start :S
<wingrime>
ollver: I simply try make own-build env
<Turl>
ol1ver: why not #include + dev specific stuff? :)
<wingrime>
ollver: For me on a13 linaro works good after my patches
<wingrime>
ollver: Same kernel same modules behaves strange on debian unstable
<wingrime>
ollver: 1) hung on resume
<wingrime>
ollver: 2) ts not move point only press
<wingrime>
ollver: linaro added some patches .....
<wingrime>
ssvb: Look like resume on crash when I use x11 framebuffer driver on debian unstable
<wingrime>
ssvb: I stoped xorg than try to suspend and it works , when xorg started it firstly start than reboot (watchdog?)
<wingrime>
ssvb: it likly bug somewhere in disp
<wingrime>
*looks like
<ssvb>
wingrime: was it exactly the same kernel for linaro and debian?
<wingrime>
yes
<wingrime>
but on linaro sunxi-fb
<wingrime>
on debian simple framebuffer driver (xorg-video-...)
<ssvb>
wingrime: it would be interesting to check just basic Xorg without any desktop environment, for example start the X server alone by running "Xorg" in one ssh shell, and then something simple like xterm in another shell "DISPLAY=:0 xterm"
<ssvb>
wingrime: and then try to suspend/resume
<wingrime>
ssvb: I tryed simular way
<wingrime>
ssvb: bug presend even you on Terminal (Alt+Clt+1)
<wingrime>
ssvb: only stoping Xorg saved me
<wingrime>
form bug
<wingrime>
ssvb: I suspended using echo "mem" > /sys/power/state
<wingrime>
ssvb: on debian I used only IceWM
<wingrime>
ssvb: can you reproduce such bugs I have ?
<ol1ver>
wingrime: ah, i tried to use the linaro alip but xorg wouldn't start due to some mali probs, so gave up and tried hansg's image instead, which so far works :)
<ol1ver>
Turl: you can't include in fex files, can you :)
<wingrime>
olver: linaro alip not worked for me with pre-builded kernel
<wingrime>
olver: soon I replaced kernel+moduler all begin to work
<ol1ver>
wingrime: i'll try the alip rootfs again sometime
<ol1ver>
the alip was my first attempt at sunxi-bsp; so didn't know a lot, i should be able to get it to work this time around
<Turl>
ol1ver: if you run them through cpp it should be possible
<ol1ver>
Turl: dunno what is more ugly :p
<ol1ver>
bah; i hope this won't byte me later: fixdep: error opening depfile: drivers/gpu/mali/ump/../mali/linux/.mali_osk_math.o.d: No such file or directory
<wingrime>
Turl: crypto device are possible for us?
<ssvb>
wingrime: yes, I also have the wrong video mode after suspend (it happened before, then somehow disappeared, and now happens again)
<ssvb>
does linaro ship some sort of kernel for allwinner hardware?
<Turl>
nice, it's way more complete than the A10 one on SS :)
<wingrime>
Turl: you can try write support
<wingrime>
mnemoc: nand flash controler on page 152
<wingrime>
mnemoc: thay put full decsription to other file
<wingrime>
Turl: It enough it write crypto
<wingrime>
*for
<Turl>
wingrime: it might be, but I'm not too familiar with crypto
<Turl>
I'll have a go at writing a driver though
<wingrime>
Turl: you have full register description, it realy enough to wire at least PIO mode driver
<wingrime>
Turl: I more interested where nand docs...
<wingrime>
Turl: It can be some nand for OpenCores
torindel has quit [Ping timeout: 245 seconds]
<Turl>
yeah I don't plan on touching dma any time soon :)
<wingrime>
dma It a great thing
<wingrime>
we must learn profiling tools
<wingrime>
and try get some speed from it
<wingrime>
Turl: dma simpler than Crypto core
<wingrime>
Turl: it simply copy mem
eebrah has quit [Ping timeout: 256 seconds]
wingrime has quit [Ping timeout: 240 seconds]
paulk_ has quit [Quit: Ex-Chat]
paulk-desktop has joined #linux-sunxi
mdfe_ has joined #linux-sunxi
gzamboni has quit [Quit: Saindo]
mdfe has quit [Ping timeout: 245 seconds]
<ol1ver>
so much to do still :(
<ol1ver>
i've put the nand images on the wiki a while ago; but still no registers described in those docs
<ol1ver>
i did try to find out as much as I could about the registers though and put htose up on the wiki; the guy who wrote a mpd driver for his boss expanded it
Dave77 has joined #linux-sunxi
<n01>
ol1ver: where on the wiki?
<ol1ver>
n01: linux-sunxi.org/NFC is my initial guess :)
<ol1ver>
linux-sunxi.org/A10/NFC is my second :)
<ol1ver>
and i just noticed wingrime was long gone
<n01>
ol1ver: tnx :)
<Turl>
yeah, I got doubts now on the crypto stuff :P
<Turl>
the fifo is 4 bytes, but what if you want to hash a non multiple of 4?
<ol1ver>
Turl: you think it's not doable?
<ol1ver>
Turl: pad the rest :)
<Turl>
I thought of padding with 0s
<Turl>
but idk if it'd produce the same md5/sha1
n01 has quit [Read error: No route to host]
n01 has joined #linux-sunxi
<ol1ver>
Turl: hmm, good question
<ol1ver>
well first a driver, then see how to deal with that, but i don't think it'll be impossible
<Turl>
I'm writing a POC from userspace nwo
<Turl>
now*
<ol1ver>
have you ever used 'make hwpack-install' on sunxi-bsp
<ol1ver>
POC?
<Turl>
proof of concept
<Turl>
nope
<ol1ver>
oh duh
<ol1ver>
the bsp uses sfdisk to re-read the partition table; but i don't think it works on all systems :)
<ol1ver>
partprobe does though
<ol1ver>
bah, linaro alip stll doesn't work; lets hope it's int he logs
<techn_>
could someone pastebin boot log from latest kernel?
<techn_>
I have problem that SUNXI_NAND_COMPAT_DEV wont work :(
<ol1ver>
3.0 or 3.4
<ol1ver>
stage or non-stage
<techn_>
stage
<ol1ver>
if you give me 20 minutes or so :)
<techn_>
specially this row: nand: nand1 nand2 nand3 nand4 nand5
<techn_>
it should be nand: nanda nandb..
<techn_>
it that compat works
<ol1ver>
ok so stage/sunxi-3.4
<ol1ver>
pull isn't work :S gimme a sec
<techn_>
Let see if this works if (!strcmp(dname, "nand")) ==> if (strcmp(dname, "nand"))
<techn_>
propably not
<ol1ver>
right, should be at newest; 6fe7041a76f650eed50c68fcdba37cc4da911159
<ol1ver>
building
<ol1ver>
ok i don't get git though; i never done anything with stage, let alone stage/sunxi-3.4; so i check it out and pull; and it spews tons of conflicts
<ol1ver>
so i figure, fine, i'll reset to somewhere in my log and pull; same deal
<ol1ver>
so i reset --hard HEAD
<ol1ver>
but it own't go to th elatest commit? i just pulled the id from github and reset to that, and no problem
<techn_>
oliv3r: I usually use 'git fetch --all' then 'git rebase' to get my own commits top of that branch
<ol1ver>
git pull :p but i should do that instead yeah
<ol1ver>
(i'm a git newb, so that explains a lot :p)
<techn_>
pull usually messes history :p
<techn_>
also rebase can do that.. and even worse :)
shineworld has quit [Ping timeout: 260 seconds]
<ol1ver>
:p
<ol1ver>
ok writing sd
<ol1ver>
ok since i can't boot into X (alip thing) lets see if it stored the bootlog
<ol1ver>
techn_: you want the nand drive bit right?
<ol1ver>
removing mali and mali_drm from /etc/modules brings up FB linaro :)
<ssvb>
in any case, linaro just has xf86-video-modesetting driver preinstalled
<ol1ver>
slightly dissapointed though
<ssvb>
it is not useful for sunxi hardware, but gets automatically loaded by mistake and blows up the x server because it does not like the mali kernel modules
<ol1ver>
hansg's fedora image brings up a nice installer on first boot; much slicker :p
<ol1ver>
i thought the mali driver sorta llready worked
<ol1ver>
but, it's star t:)
<ol1ver>
sleep time now :(
<ol1ver>
thanks :D
<ssvb>
mali just needs a different xorg driver
<ssvb>
mali + xf86-video-modesetting = non-working xorg