<DonkeyHotei>
[Mon 2012-07-02 09:06:23 AM PDT] <DonkeyHotei> i will be out of town from july 10 to july 12 and i really hope the delivery is not during that time
<DonkeyHotei>
[Mon 2012-07-02 09:10:08 AM PDT] <Turl> DonkeyHotei: don't worry, it usually takes more like 15-20 days
<DonkeyHotei>
sure enough, delivery was attempted while i was out of town
<hipboi_>
mnemoc: i promise there will be 1GB modekl
<hipboi_>
s/modekl/model/
<ibot>
hipboi_ meant: mnemoc: i promise there will be 1GB model
<specing>
hipboi_ :)
<specing>
hipboi_: Any quad core ARMs being designed?
<Turl>
quad core A15 maybe? :)
<hipboi_>
specing: sorry i can't tell
lkcl has quit [Ping timeout: 252 seconds]
lkcl has joined #arm-netbook
<rm>
received my USB video "card" from China
<rm>
will try it with Mele today
<rm>
should have all fine and dandy, DPMS should work, 1680x1050 over DVI, etc
<lundman>
:)
<rm>
maybe even some sort of DRI/Xv
<Turl>
rm fu, my usb<->parallel cable is still on the mail :(
<rm>
btw, are there any new discoveries regarding defining custom modes on HDMI of A10?
<lundman>
should be easy to add more modes
<lundman>
then write a C program to trigger them
<specing>
hipboi_: :(
Shivan has quit [Ping timeout: 245 seconds]
Shivan has joined #arm-netbook
<rm>
how :S
<rm>
well, I have a program to switch modes
<rm>
(not that it works properly, though)
<rm>
but where to add new ones
hipboi__ has joined #arm-netbook
<hno>
steev, NiH is a natural syndrome of humans. It feels easier to throw away and start over than trying to fix someone elses mess (or your own when memory have faded). We consequently ignore the fact that those last 5%. takes the most effort and starting from scratch only adds to the effort.
hipboi_ has quit [Ping timeout: 248 seconds]
<specing>
heh
<hno>
rm, the hardware is completely flexible. The inflexible part is only the mode switching code based on a fixed set of modes. See TCON1_set_hdmi_mode function (drivers/video/sun4i/disp/de_bsp/de/ebios/de_lcdc.c ) and other uses of those mode constants..
<CIA-122>
rhombus-tech: Peter master * r4aa1ff79e5bf /allwinner_a10/hacking_the_mele_a1000/script_for_installing_debian_on_sdcard.mdwn: Updated download link
<CIA-122>
rhombus-tech: Peter master * r564b525ab8c6 /allwinner_a10/hacking_the_mele_a1000/mele_ubuntu_image.mdwn: Updated download link
<hno>
mnemoc, there was quite much configuration details hiding in boot0/boot1. See bootinfo.c
<hno>
even a complete copy of script.bin if I understand the structs correct.
* hno
did some gdb magics to extract them all in symbolic form.
<Turl>
hno: I was told the recovery keys were hardcoded on bootX, and I saw entries for them on script.bin
<Turl>
might explain why then
eebrah has quit [Ping timeout: 265 seconds]
<CIA-122>
rhombus-tech: Peter master * r73672526949e /allwinner_a10/hacking_the_mele_a1000/android.mdwn: Update download link
<hno>
Turl, no "recovery" keys in the configurable areas.
<CIA-122>
rhombus-tech: Peter master * re618a2596d4f /allwinner_a10/hacking_the_mele_a1000/android.mdwn: Update download link
<Turl>
:)
<lundman>
what are we downloading
<hno>
lundman, who is downloading?
<Shivan>
What whould be requiered to run opengl-es specific apps from framebuffer (not x) on A10?
<hno>
mali driver + libraries.
<Shivan>
i can be done in the currect state? last time i tryied to use the mali driver i was getting "Unable to get secure ID" and it was a 3.0.36 kernel and modules
<mnemoc>
hno: awesome! :D
<mnemoc>
hno: are you fetching the blob using fel?
lkcl has quit [Ping timeout: 255 seconds]
<mnemoc>
<30>udevd[144]: renamed network interface eth0 to eth1
<mnemoc>
<7>wemac wemac.0: enabling eth1
<mnemoc>
wth
<Shivan>
OK im here at this point
<Shivan>
(EE) MALI(0): [maliModifyPixmapHeader:165] UMP failed to retrieve secure id
<ejstacey>
mnemoc: I've seen that (in linux in general) when there are udev rules (which were dynamically created) that had a specific MAC
<ejstacey>
the rule was under /etc/udev
<ejstacey>
or /etc/udev.d, whatever the proper dir is
<mnemoc>
ejstacey: so it's the reaction to a MAC address change?
<Shivan>
maybe is because DISP and HDMI is built-in?
<Shivan>
LCD too
<ejstacey>
mnemoc: yep, something (udev itself?) wrote a config file that said only a specific mac gets eth0
<mnemoc>
ejstacey: hate "smart" tools :<
<ejstacey>
mnemoc: it's an attempt to have proper/static ordering of ethernet interfaces
<ejstacey>
mnemoc: yeah, can bite you sometimes!
<mnemoc>
Shivan: 3.0.36?
<Shivan>
yep
<mnemoc>
zgrep DRM /proc/config.gz
<Shivan>
disp, lcd and hdmi are built in otherwise there is not signal on hdmi
<mnemoc>
Shivan: zgrep DRM /proc/config.gz
<Shivan>
CONFIG_DRM=m CONFIG_DRM_MALI=m
<Shivan>
mmm
<Shivan>
drm=y?
lkcl has joined #arm-netbook
<mnemoc>
try. I've personally never used it
<mnemoc>
also check all the mali modules are loaded
<Shivan>
disp, fsb, lcd and hdmi are not showing up on lsmod, but its because they are built in (i think)
<Shivan>
drm, mali_drm, mali and ump are there
<mnemoc>
just for the sake of testing, can you turn the sun4i-disp bunch to modules?
<Shivan>
im compiling a new one right now
<specing>
mnemoc, ejstacey you can disable that feature
<mnemoc>
specing: first I need to find out why I got a random MAC :)
<mnemoc>
Shivan: if that works I'll have to ask you to test the 3.4 branch too :p but 3.0 first :)
<specing>
You can also write an udev rule to set the mac to deadbeafcafe on bootup
<mnemoc>
I want to change how the mali memory is allocated, but need someone with working X11/mali to test it
<Shivan>
last time i try 3.4 branch on my aurora it did not even load up the kernel (usb keyboard and mouse does turn up once kernel is loaded, that does not happen with 3.4 for me)
<mnemoc>
aurora :(
<mnemoc>
was hoping something with serial console
<Shivan>
:(
RITRedbeard has quit [Read error: Connection reset by peer]
<mnemoc>
I have pending a patch that enables lcd_spi (for aurora), but it also includes some fixes to the disp driver itself.... but it gives Quarx a greenish layer on hdmi
RITRedbeard has joined #arm-netbook
<Shivan>
thats why 3.4 branch does not load on aurora?
<mnemoc>
Shivan: without a serial console hard to say what "does not load" means
<Shivan>
yeah, lets say, kernel does not even power up usb host
<Shivan>
i cant move disp to modules whiout moving lcd and hdmi too if i do that i get not signal on HDMI :(
<mnemoc>
yes, natural. when I asked you for that I was assuming a mele.
<Shivan>
uou known why HDMI does not work on a module on an aurora?
<mnemoc>
no idea
<Shivan>
maybe i should retry using the internal lcd
<mnemoc>
you can also try to boot from nand and use the uSD slot for a serial console
<mnemoc>
and *see* what's going on
hipboi__ has quit [Read error: Connection reset by peer]
<Shivan>
well i give up no idea of what is going on
<mnemoc>
Shivan: just noticed DRM_MALI doesn't build when DRM is built-in
<mnemoc>
meh, no.
<mnemoc>
in 3.4 it needs experimental :<
rellla has quit [Quit: ChatZilla 0.9.88.2 [Firefox 13.0.1/20120614114901]]
Kesskisspass has joined #arm-netbook
Kesskisspass has left #arm-netbook [#arm-netbook]
Kesskisspass has joined #arm-netbook
<Shivan>
OK got it
<Shivan>
the mali driver is loading, and X server starting, but still a black screen on hdmi
<mnemoc>
Shivan: why wasn't it loading?
<Shivan>
i had the modprobes on rc.local and where not working
<Shivan>
and if i build in disp, lcd and hdmi,the driver cant get secure ID, thats has to be some bug on the kernel code
<mnemoc>
you probably need to use an ioctl() against /dev/disp to actually enable the hdmi
<mnemoc>
don't forget they made that code thinking only in android support, and implemented only the minimal code to get android working
<mnemoc>
the disp driver can't be uglier
<mnemoc>
and so, until it's cleaned, there is no way of fixing anything
<Shivan>
haha i did the DISPLAY=:0 xcalc and xcalc shows up on hdmi screen
<Shivan>
its x server who is not working then...
<mnemoc>
DISPLAY=:0 xcalc shouldn't work without an x server
<furan>
iit's probably your window server. does the xcalc have border/titlebar?
<Shivan>
nop, just a white thing with black letter and lines
<traeak>
rm: you'll have to checkout hte proper branch. that gets you on 3.0.36
<mnemoc>
rm: switch to the wip/linux-3.4-sunxi/forward branch
<traeak>
looks like rtl8192cu is still busted in the 3.4branch
<mnemoc>
traeak: fix it ;-)
<Shivan>
mmm im getting a EGL error init display:272 when running es2_info, im guessing that means it works, at least no error 12291 this time...
<traeak>
hehe, i've got to get this lidar stuff working. i had to cook up my own image processing algorithms, i hate when i have to do that (instead of steal other people's good ideas)
<traeak>
mnemoc: i know i should take a shot at it
<traeak>
mnemoc: does 3.4 at least boot without?
<mnemoc>
sure
<mnemoc>
Welcome to Linaro 12.06 (GNU/Linux 3.4.4+ armv7l)
<hno>
mnemoc, which blob?
<mnemoc>
hno: boot0/boot1
<hno>
No, that's from firmware images, SD images in my case.
<mnemoc>
meh. nothing is perfect :<
<traeak>
hmm...i hsould seriously considering putting the whole damn userspace on the ssd...
<mnemoc>
hno: it would have been nice to be able to script the whole upgrade :)
<mnemoc>
hno: but it's great news anyway
<traeak>
mnemoc: didn't even start booting, time to clean up everything and try again
<mnemoc>
$ uname -a
<mnemoc>
Linux mele-sd 3.4.4+ #30 PREEMPT Fri Jul 13 16:08:02 CEST 2012 armv7l armv7l armv7l GNU/Linux
<traeak>
removal of some includes in drm.h and turning off the unnecessary sata driver config settings
<mnemoc>
filetea.me is sort of peer-to-peer over https
<mnemoc>
just drag a file, and immediately get a url to share
rellla has quit [Ping timeout: 255 seconds]
<mnemoc>
and nothing is uploaded until someone starts downloading
rellla_ is now known as rellla
<rm>
sounds nice
drgreenthumb has quit [Read error: Connection reset by peer]
drgreenthumb has joined #arm-netbook
<libv>
hrm, ch342 serial device... same stuff.
rellla has quit [Quit: ChatZilla 0.9.88.2 [Firefox 13.0.1/20120614114901]]
rz2k has joined #arm-netbook
<libv>
actual pc serial port, still garbage...
<libv>
i guess i will have to wait for the ttl uart -> usb converter to arrive
<hno>
mnemoc, bootinfo is nothing about installation, only firmware investigation.
<hno>
verifying which parameters a firmware is using.
<mnemoc>
hno: yes, but I was thinking about the possiblity to generate a .fex chunk and a header for uboot out of it
drgreenthumb has quit [Changing host]
drgreenthumb has joined #arm-netbook
<mnemoc>
and ultimately replacing boot0/boot1 with our uboot
<hno>
Yes, it means we can reuse the same header for u-boot if we want, to reuse the tools. Or we make our own header for our own tools.
mikie_w has joined #arm-netbook
gimli has joined #arm-netbook
<hno>
replacing boot1 is easy. The hard part is boot0 where we need to figure out how simple raw NAND access works to fit in the available SRAM space.
<hno>
but next goal is GPIO testing via FEL.
<mnemoc>
:)
<hno>
Which is interesting. Enabling testing most pins, only requiring USB & power working.
<hno>
well, DRAM pins don't support GPIO so those can't be tested that way.
<hno>
but all I/O pins including NAND pins.
<hno>
Also need to figure out how simple raw NAND access works to be able to flash our boot0 replacement and u-boot.
<hno>
The raw nand access code in the u-boot nand driver intended for u-boot to use for this purpose do not seem to work proprer, or I am not understanding how to use it.
kaspter has quit [Ping timeout: 246 seconds]
<traeak>
mnemoc: thx that did the trick
<traeak>
heh
<hackandfab>
I've asked gimli how did he managed to use cedar on the mele2000 and he has told me to look at amery git but he thinks newer code is floating around. Anyone knows what code is that?
<mnemoc>
the mess tarball from the SDK
<mnemoc>
Iain Bullard (who seems to be working in REing it) told me he will sort out a good set of files to push to the repo during this weekend
<mnemoc>
because the old code there seems to even cause the kernel to panic :|
<hackandfab>
thx mnemoc, I'll try again next week then
<mnemoc>
if you plan to add cedarx bindings to libva I can point you to the stuff that's floating around... so you can start this weekend ;-)
<gimli>
mnemoc: is Iain also on irc ?
<gimli>
mnemoc: or do you have a mail adress from him, wann contact him
<hackandfab>
thx mnemoc, I'm beginning with arm and the mele (I've done embedded things with coldfire V2 and uClinux) but I wouldn't probably know how to implement the bindings yet.
eFfeM has joined #arm-netbook
<mnemoc>
gimli: 1m, I'll give you his address on /q
<mnemoc>
hackandfab: somehow I thought you were into making a product
<hackandfab>
I'm a student doing my final project on a videoconferencing app and I want to use the mele2000 with a H264 webcam
<hackandfab>
the fab of the nic is because that's what I like to do (I did CNC stuff with a mill, a minilathe and a prusa mendel)
<mnemoc>
:)
<mnemoc>
hackandfab: check your mail
<mnemoc>
sent you a link
eFfeM has quit [Quit: Leaving.]
<hackandfab>
:)
<hackandfab>
:) :) :)
<mnemoc>
but keep waiting the repo, it will probably get shaped up soon
<mnemoc>
s/waiting/watching/
<ibot>
mnemoc meant: but keep watching the repo, it will probably get shaped up soon
<Shivan>
3.4.4 does not boot on aurora, for wharever reason, but i have no way to know why either
Turl has quit [Excess Flood]
Turl has joined #arm-netbook
<steev>
Shivan: only way to find out is to get serial output
<Shivan>
yeah i know, but for some reason the aurora is tricky to get intro a successfull boot, some times it requires to power it on and off several times until the kernel load up successfully.
<Shivan>
and by successfully i mean, it sends power to the usb host
<mnemoc>
get serial output ;-)
<Shivan>
nah ill probably get a mele
<mnemoc>
:)
<Shivan>
there no much point in running linux in a tablet anyway
<steev>
there is for some specific purposes
<Shivan>
a very touch friendly UI is needed... or you mean using it with a keyboard+mouse?
<hno>
Shivan, I can think of many Linux uses of a tablet, both with and without extra input devices.
mpthompson has quit [Ping timeout: 240 seconds]
mikie_w has quit [Quit: Leaving]
mikie_w has joined #arm-netbook
<mnemoc>
it would be nice to see tizen or mer on A10
mikie_w has quit [Remote host closed the connection]
<mnemoc>
but we are trapped by the binary libraries
<mnemoc>
maybe mer's UI can be installed upon ubuntu
<steev>
oh patches, how i love thee
orly_owl has joined #arm-netbook
n6pfk has joined #arm-netbook
mikey_w has quit [Disconnected by services]
mikey_w has joined #arm-netbook
mikey_w has quit [Disconnected by services]
n6pfk has quit [Client Quit]
mikey_w has joined #arm-netbook
n6pfk has joined #arm-netbook
nibb_ has joined #arm-netbook
nibb has quit [Read error: Connection reset by peer]
n6pfk has quit [Read error: Connection reset by peer]
<specing>
the electronics in the modern drives costs about 25$ and the motor/stepper another 25$ probably
<specing>
so the platters are like 50$?
<specing>
so the worst case scenario would be 25$ electro + 40$ motors + 500$ for 10T platters
<specing>
versus 10*100$ for 10T on normal disks
<specing>
datacenters would be all over it in no time :)
<mnemoc>
adding plates also increases the failure rate of the drive
<hno>
not sure it does compared to stacking a number of drives to reach same capacity.
<specing>
even if they failed 125% more often than 3" ones they would still pay off
<mnemoc>
if one of the stack fails, you replace that one. if one plate of the mega-disk fails, you have to replace the complete beast
<hno>
but the higher the spindle the harder to manage vibrations.
<mnemoc>
doesn't sound like a good deal
<hno>
building good disk arrays is complex, and is a lot about selecting the right drives for the job. Most desktop drives do not like that environment.
<traeak>
could be the size makes for harder manufacture as well
<traeak>
shrinking stuff means manufacturing equipment and drive arms have less distnace to move
<traeak>
anything angular is punished by arm distance
<specing>
yeah but track length is 2*pi*r
<specing>
It would be worth it, trust me
* hno
is a bit sad.. my N900 had a violent death the other day.
<specing>
Did it implode? explode?
<hno>
took a bath.
<hno>
water and powered on electronics do not play well together.
<hno>
so I urgently need to fins a replacement
<mnemoc>
a lumia would look... similar :p
<specing>
lol
<libv>
mnemoc: n900 is the brick, with a hw keyboard
<specing>
hno: Why did you bath your N900?
<traeak>
hno: you are fortunate. my n900 had a slow death from crappy microusb
<libv>
where the keyboard switch would wear out quickly, and the micro-usb would break off
<mnemoc>
ah, yes... confused with the n9
<hno>
specing, you don't want to know the details. It's a messy story.
* libv
still needs to disable serial and other things on his n9 so that it stops blinking as suspiciously
<hno>
libv, mine surived quite well until the bath. But yes the usb had started to malfunction a bit. Have not had keyboard switch issues on either of our N900.
* traeak
needs to probably pick a new phone
<libv>
hno: i dropped mine when clumsily flipping it open
<libv>
i still miss it, it seemed much more practical in many ways compared to the n9
<libv>
if only the n950 had actually been put into the shops
<hno>
mine had hit hard floor and streets multiple times without complaining, only a bit of scratches in the paint on the metal around the screen.
<specing>
heh, hammers are the only things nokias play well with ;P
<libv>
with me it just fell flat and nice, didn't seem that catastrophic, but it never really fully booted again
* mnemoc
loves his gnex
<traeak>
i'm sure a 9mm or better would do a good job on a nokia as well
<traeak>
gnex?
<mnemoc>
galaxy nexus
<mnemoc>
and iirc mer supports
<hno>
been looking a bit at that one. but no keyboard and it's android.. but at least the android part is fixable.
<libv>
about mer...
<traeak>
don't care much about hard keyboards
<libv>
it's become rather silent around the vivaldi
t0dbld|work has joined #arm-netbook
lkcl has quit [Ping timeout: 248 seconds]
<traeak>
kde tablet?
<libv>
yeah
ka6sox-away is now known as ka6sox
lkcl has joined #arm-netbook
arete has quit [Quit: Leaving]
nibb_ has quit [Read error: Connection reset by peer]
tom_say has quit [Remote host closed the connection]
Shivan has quit [Ping timeout: 245 seconds]
nibb_ has joined #arm-netbook
mpthompson has joined #arm-netbook
gimli has quit [Quit: Verlassend]
newbie has joined #arm-netbook
newbie is now known as Guest65047
hp_ has quit [Ping timeout: 276 seconds]
t0dbld|work has quit [Ping timeout: 255 seconds]
<rz2k>
anyone tried usb gadgets on a10? g_file_storage seems to work, but g_serial does nothing and g_ether doesnt even compile.
<specing>
yeah
<specing>
it didnt work
<specing>
g_ether compiles though
<specing>
I was hoping to get a serial port through... no workie
RITRedbeard has quit [Read error: Connection reset by peer]
Shivan has joined #arm-netbook
<rz2k>
g_ether source looks like some didnt push all changes.
<Shivan>
i recompiled xserver to disable glx and aiglx and now im missing kbd and mouse modules haha -.-
<Shivan>
i think i miss something i never compiled xserver before
<specing>
rz2k: It would be cool to have it since I haven't found the rx line yet
<rz2k>
s/some/someone/
<ibot>
rz2k meant: g_ether source looks like someone didnt push all changes.
<rz2k>
yeah, I've used g_ether when I was playing with ingenic boards