<darkfleet>
look at the log at that time and i have the same error, except it doesnt seem to mention any fix
<slapin>
what is the difference between mk802ii and mk802+?
<slapin>
and do anybody have some recent stock image for mk802ii?
* slapin
found one
svp has joined #arm-netbook
SouL_ has joined #arm-netbook
SouL_ has left #arm-netbook [#arm-netbook]
<specing>
markic: I have tried using pl2303 on A10 and failed
<slapin>
also - does PhoenixCard.exe run under wine?
<specing>
markic: it just didn't register it
<slapin>
specing: probably module is missing
<specing>
slapin: I compiled it into the kernel
<specing>
there was no announcement in dmesg
<specing>
the only thing that worked on that host USB port was USB keyboard and mass-storage, nothing else
<specing>
but this was 5 months ago
<specing>
maybe the kernel mess is sorted out now
<specing>
anyway Im too busy with the stellaris ARM launchpad to care about A10 :)
Skoti_ has joined #arm-netbook
<jinzo>
specing, there's been quite some work done on USBs
<specing>
nice
penguin42 has joined #arm-netbook
luka has joined #arm-netbook
AndChat138129 has joined #arm-netbook
<AndChat138129>
The sun4i kernel on service.i-onik.de is missing a few files ipv4 because of uppercase filenames.
<AndChat138129>
Looks like it came from a windows machine.
eebrah has joined #arm-netbook
<markic>
specing: I failed with any .ko file
<specing>
markic: recent kernel?
<specing>
recent as in todays/yesterdays?
<markic>
specing: 3.0.8
<specing>
Now that really tells me when it was released
<markic>
specing: vermagic is OK, and module symvers match
<specing>
you didn't build this kernel yourself, did you?
<AndChat138129>
I have already compiled the sun4i kernel successsfully and boots under android.
<specing>
AndChat138129: you mean android boots on top of your kernel?
<markic>
specing: yes I built all modules myself
<specing>
markic: I don't care about the modules, Im asking you for the kernel
<specing>
DOOOH
<AndChat138129>
Yes the one from service.i-onik.de
<specing>
AndChat138129: why are you using the unofficial unofficial kernel? Atleast use the official unofficial one
<AndChat138129>
That kernel contains gpio_sw.ko
markic has joined #arm-netbook
<AndChat138129>
and it looks like a later kernel release.
<markic>
specing: my kernel is factory installed
<specing>
markic: that means it is old as the maya prophecy
<markic>
maybe
<AndChat138129>
The kernel source is missing the cedar module.
<specing>
markic: try the git one before complaining about non-working stuff again
<markic>
specing: so, what's your story - have you found out why your modules failed to install?
<specing>
markic: they installed just fine
_whitelogger has joined #arm-netbook
<techn_>
AndChat138129: check our import branches
<techn_>
there is also some newer codes
<markic>
specing: I would like to flash my tablet but many people complain that they can not find img with full hardware support
<markic>
specing: touchscreen is biggest problem
<techn_>
markic: which touchscreen you have?
<markic>
techn_: don't know :(
<techn_>
you got stock rom somethere?
<techn_>
enable usb debug mode and connect to device with adb
<markic>
techn_: no
<markic>
yes, i did that
<techn_>
but since you dont have stock kernel that wont help
<buZz>
maybe find something to flash that _is_ stock
<buZz>
then analyze
<buZz>
AndChat138129: what does the cedar module _do_ actually?
truki has joined #arm-netbook
<truki>
hi all, as a totally noob trying to compile an a13 kernel
<truki>
anyone alive here?
<markic>
what means error message: "HCI socket registration failed" when trying to insmod bluetooth.ko?
<markic>
in dmesg i am getting "HCI device and connection manager initialized" followed with "HCI socket registration failed", and don't know what could be the problem
<buZz>
truki: quite a lot of ppl alive here
<jinzo>
buZz, I would wager a guess that there're currently 107 alive people here.
<buZz>
well
<buZz>
i assume some are bots ;)
<truki>
ok then, ridiculously i could get the lcd working after init starts but external kbd is not recognized, even no numlock led... :/
<jinzo>
and some duplicates, but yes :D
<buZz>
truki: did you compile usb keyboard support? :)
<truki>
but mouse light is blinking...
<truki>
i did a total static build, no modules...
<truki>
lol u mean usb hid?
<buZz>
eh yeah i think
<truki>
its already there
<buZz>
i have no idea what your external kbd is
<buZz>
is it usb? :D
<buZz>
do you have serial console to your device?
<truki>
no serial, jtag, spi
<truki>
it's just a silly a13 tablet
<truki>
btw yeah kbd is a standart usb one
<buZz>
i would be very suprised if that tablet doesnt have uart pins somewhere :)
<buZz>
anyway, hmm, silly, you could run lsusb as init and see the output?
<specing>
yes
<specing>
or throw it into /etc/local.d/
<specing>
(I have done this)
<buZz>
or start ssh and network by default ;)
<truki>
i tried kernel commandlines as /bin/bash but it didnt show up and rebooted
<truki>
hmm
<specing>
buZz: if you have everything set up driver-wise and connection-wise
<buZz>
ofcourse
<specing>
everything I have is wireless which is black magic
<specing>
voodoo
<buZz>
lol
<specing>
and no Rx PIN :(
<penguin42>
wireless is awful - you can't follow your data
<specing>
I'll try to find it with a GPIO banger once I find out how this ARM GPIO stuff works
<specing>
Im currently coding my first LED blinker for the stellaris launchpad
<buZz>
hehe cool
<buZz>
so did i ;)
<buZz>
code a blinker
<slapin>
buZz: you can code far faster blinker on a10, by the way
gimli has joined #arm-netbook
<slapin>
buZz: you can order LED tape of 2000 LEDs and blink them all at onece
<buZz>
hehe
<buZz>
i have a string of WS2801 leds here
<buZz>
fake SPI
<buZz>
supposedly i could do 750fps with them on stellaris
<buZz>
just have them on an arduino now
<buZz>
and already over 50fps
<slapin>
hno: ping
<slapin>
hno: hi, it seems we'll have to resync mtd in u-boot and submit to u-boot ML for merge.
<buZz>
hmm
<buZz>
when is VT capable Cortex becoming available?
<buZz>
wouldnt that be 2012?
<penguin42>
buZz: A15 are supposed to be able to do virtualisation well
<rz2k>
that guy is from redhat, he knows what he does
<rz2k>
:p
<jelly-home>
can't tell if sarcasm
<rz2k>
ssvb: also, I'm not sure, but mali libs *could* be built with dependency on umplock/no umplock
<rz2k>
no idea what will happen if we will enable it
<rz2k>
atleast enabling OS memory backend crashes everything to pieces
<rz2k>
(mali driver can work without UMP with os memory)
hp__ has quit [Ping timeout: 276 seconds]
<ssvb>
rz2k: so there is no real umplock kernel driver available?
Skoti_ has joined #arm-netbook
<rz2k>
in r3p2, but I have no idea if we can get it working with our r3p0 or r3p1 libs
<rz2k>
r3p2 xserver is updated too
<rz2k>
s/xserver/xf86-video-mali/
<ibot>
rz2k meant: r3p2 xf86-video-mali is updated too
<rz2k>
it now has umplocking/caching by default and selection is done by xorg config
<rz2k>
other stuff was updated too
<techn_>
should we start asking new libs from tom? :)
<rz2k>
he's probably busy with cubieboard production
<rz2k>
but I would enjoy seeing them
<rz2k>
especially because they use new drm skeleton
<rz2k>
s/drm/drm-kms/
<ibot>
rz2k meant: especially because they use new drm-kms skeleton
<techn_>
or should we wait until next? :)
<rz2k>
it could happen that there will be no next
<rz2k>
because ARM is moving to T6xx
<techn_>
that's true
<ssvb>
rz2k: ok, I see, just right now xf86-video-mali looks very broken to me, vsync does not seem to work, lots of buffer copies are loading the CPU almost 100% on 3D workloads, 2D performance is horrible
<rz2k>
vsync needs to be wired with /dev/disp
<techn_>
rz2k: ?
<techn_>
it should be in /dev/fb
<rz2k>
fbscalerlayer is in disp?
<techn_>
it's in both
<rz2k>
allwinner at its best
<techn_>
and /dev/disp will eventually be deprecated
<techn_>
and all interfaces moved to sysfs
<rz2k>
i remember that turning scalerlayer causes lots of tearing and no vsync at all
ppord has quit [Read error: Connection reset by peer]
ppord has joined #arm-netbook
<rz2k>
we have a some sort of mess around there
<rz2k>
we can use provided by Allwinner libump OR compile our own one with missing symbols problem fixed
<rz2k>
mali-libs is 1:1 as it came from Allwinner
<ssvb>
yes, libUMP.so from mali-libs has no -ldri2 -ldrm -lXfixes linked and causes a lot of issues
<ssvb>
the GLES2 headers also seem to be incomplete
<libv>
ssvb: i dug them up from the arm sdk anyway
<rz2k>
ssvb: if you have some love to share with mali-libs/xf86-video-mali, we will hugely appreciate it. :p
<ssvb>
libv: I could not compile glmark2 using the headers from mali-libs and just took them from mesa (not sure if that's the right solution, but it seemed to work)
<Dessimat0r>
how do i debug my uhost with allwinner a10 if it doesn't boot after i compiled a new kernel with everything else the same and rebuilt the image and flashed it?
<Dessimat0r>
maybe the uboot config is bad?
<ssvb>
libv: in any case, with the whole mali drivers mess, this GLchar issue is just one more annoyance (sorry if this hurts your purist feelings)
<Dessimat0r>
it was the sunxi 3.4 kernel
<Dessimat0r>
with patch to disable USB gadgets
<libv>
ssvb: i am not a gl standard purist.
<Dessimat0r>
the kernel config was taken from the kernel of an image that works on the device
<libv>
ssvb: i was simply stating that there is a difference between the standards, and that this is the only reason why this went wrong
<Dessimat0r>
i did menuconfig and loaded the existing config
<libv>
ssvb: not because the mali sdk headers are broken, and not because your compiler might be messed up
<techn_>
Dessimat0r: have you enabled CONFIG_SUNXI_IGNORE_ATAG_MEM.. if you are using allwinners u-boot?
<libv>
ssvb: i am actually quite amazed that "i took headers from project XXX, which might be totally different, and it worked" was an option you chose for what is a simple lacking define
<ssvb>
libv: that's why I said "not sure if that's the right solution, but it seemed to work" :) And it was definitely a lot faster to try than looking into the headers, fixing this define and hoping that nothing else breaks
<ssvb>
libv: I'm glad that you found it entertaining
<Dessimat0r>
techn_: i think I am, when i rolled it back into the same image I reused the same boot stuff except for the kernel
<Dessimat0r>
when i had finished compilation of the kernel stuff, it did moan about a missing boot config or something
<Dessimat0r>
but i guess that may use the default
Sv2 has joined #arm-netbook
Sv2 has joined #arm-netbook
Sv2 has quit [Changing host]
<Dessimat0r>
but when I roll it into the same boot just replacing the kernel, is the old boot config used?
<Dessimat0r>
i guess I should try recompiling the kernel with CONFIG_SUNXI_IGNORE_ATAG_MEM
<Dessimat0r>
and flashing it like before
<Dessimat0r>
if it's a kernel config option :)
<techn_>
yes.. It'ss kernel config option
sv has quit [Ping timeout: 276 seconds]
<ssvb>
rz2k: honestly I have no good idea about what can be done with xf86-video-mali, the only really necessary piece of it is the DRI2 part which is enforced for interoperability with the mali binary blobs
<ssvb>
rz2k: but because this driver does a lot of memory copy operations by CPU and has no real vsync, I guess a simple XShmPutImage based implementation for integrating GLES with X11 would work equally well
<ssvb>
rz2k: and be compatible with any DDX
<ssvb>
rz2k: for example, SGX drivers even had libpvrPVR2D_X11WSEGL.so earlier which worked "fine" with xf86-video-fbdev, but then they abandoned it in favour of DRI2, totally killing 2D performance as a side effect :)
<ssvb>
libv: does Mali400 have MMU?
<Dessimat0r>
techn_: ok :D
<libv>
ssvb: depends, but mostly yes
<Dessimat0r>
cheers for that :))
hg_5 has quit [Ping timeout: 264 seconds]
<Dessimat0r>
i will do it right away and tell you what happens with it
<ssvb>
libv: I wonder if mali UMP can be also used for G2D, or is it better not to mix them at all
truki has joined #arm-netbook
<techn_>
ssvb: please dont mix them
<techn_>
or if you mix.. be sure that it works with sun5i
<techn_>
sun5i doesn't have G2D :/
<truki>
hi peeps...any arrowed tree style info for the forks of this a13 hell from leakage till now? got totally confused with all those stuff... lichee/crane/bsp/reference etc
<truki>
btw if there are just sun4i and sun5i platforms why defconfig_a13, etc stuff? redundancy?
<techn_>
truki: derived from allwinner
<truki>
sure but then we have cyanogenmod branch
<ssvb>
techn_: G2D is a separate hardware accelerator, it can be used or disabled without any problems
<truki>
got screwed with the complexity
<ssvb>
libv: xf86-video-mali is allocating pixmaps using UMP, currently this brings nothing but performance problems
<rz2k>
ssvb: by DDX you mean EXA?
<techn_>
it makes some sense.. sun4i - a10. sun5i - a10s, a12, a13.
<techn_>
a10s is a12 could be the same
<techn_>
ssvb: yep
<rz2k>
ssvb: and yes 2d/3d buffering and especially showing big pixmaps/textures makes drivers cry and output 11fps for just constantly showing 1080p image :/
<ssvb>
rz2k: DDX == any X11 driver including xf86-video-fbdev (which is not using EXA or implementing DRI2), but this was mostly a theoretical question because we are stuck with DRI2 and can't interoperate with mali blobs without it
<ssvb>
but EXA can be kicked out if necessary
<rz2k>
oh. so writing EXA<->G2D hooks is pointless? ARM states in readme that we should do the 2D accel. this way, using EXA for 2D and blobs+dri2 for 3D.
<truki>
what is the stablest recommended branch for a13 anyway? otg was not working recently for 3.0 master branch...
<rz2k>
ssvb: could you please try to take r3p2 UMP/UMPLock and build xf86-video-mali with umplock? I would do that myself, but I'm away from my devices :/
<techn_>
truki: 3.0 master
<truki>
eh, otg didnt work in my case... it gives TT error
<techn_>
truki: what you mean with otg? slave or host mode?