<focus>
next task is to liven it up with xtal, some capacitors, led and then let it boot a minimal program from sdcard
terryoy_ has joined #linux-sunxi
<focus>
to flash the led
terryoy has quit [Ping timeout: 252 seconds]
<focus>
any thoughts on setting up the compiler environment to do this? I was thinking that if an A10 tablet runs gcc, it should be able to create executable to do this - bearing in mind the executable can't call on anything Linux because there is no ram on this board - it will have to boot from sdcard, and into the ram inside the CPU to do this.
<focus>
boot from sdcard, and load the executable into ram inside cpu
<focus>
any suggestion welcomed, I leave it all documented for anyone else wanting to follow building tablets and its OS from scratch as possible
<RaYmAn>
focus: if you can get it in fel mode, (and have usb?) that's probably easier
<RaYmAn>
because bootrom sets up a stack and stuff automatically
<RaYmAn>
alterntaively, you need arm-none-eabi-, some assembly to setup stack and stuff and call the main function and then it should be fine
<focus>
RaYmAn: how would the compiled executable get from desktop PC to board through the USB. in windoze it looks like livesuit, but i don't do windoze - is there a linux equivalent of livesuit, or can it work from say virtualbox, or is it easier to put it on sdcard and let it boot that up?
terryoy_ has quit [Read error: Connection reset by peer]
<RaYmAn>
something like that
hipboi has joined #linux-sunxi
<focus>
I initially thought fel.c runs on A10 but I think it runs on desktop PC to talk to the A10 through USB after A10 has entered into FEL mode?
<RaYmAn>
yes
<focus>
AHAAAAAAAAAA!!!
<focus>
Got it now!
<RaYmAn>
It's essentially a bootrom low-level usb recovery mode
<hipboi>
yes
<hipboi>
but it seems fel is not working very well
<RaYmAn>
it lets you do very few things - read/write to sram/memory, execute stuff from sram/memory
<hipboi>
hno, haven't seen you for a long time
<hipboi>
now we have u-boot and spl for a20
<RaYmAn>
allwinner released it for a20 officially as well?
<hno>
excellent!
<mnemoc>
hipboi: can you try to integrate it into sunxi-current ?
<hno>
also should integrade the mtd nand support.
<hipboi>
mnemoc: it seems i break a10, i need to check
<mnemoc>
that's the main problem with soc-specific branches. changes tend to break everything else
<hipboi>
well, too many ifdefine...
<focus>
RaYmAn: read/write to sram is all I need to get the miminal exe to flash LED, then I know xtal and CPU is working fine and build from there. Thank you very much for your guidance! :)
<focus>
Hmmm.. actually even less work - if the CPU boots into FEL mode and talks through USB, then its already internally booted! doh! flashing LED becomes extravagant - still having a build environment to make exe and then flash led is still a feat in its own right
BJFreeman has quit [Quit: had a good time]
<hipboi>
hno, i remember long time ago you said the problem of not booting from uSD sometimes
<hipboi>
hno, what's the cause
<mnemoc>
that was fixed by retrying the dram initialization
<mnemoc>
on reboot it tends to work the 2nd or 3rd try
<hipboi>
not that one
<hipboi>
boot from nand instead of uSD
<hipboi>
unplug the power cable and plug in, some time it boots from uSD but sometime from nand
<hipboi>
but reboot can always boot from uSD
<mnemoc>
that would mean the brom is not seeing the card...
<hipboi>
some hardware issue cause that... i guess
<mnemoc>
dust? :)
<hipboi>
:O
<hipboi>
:)
<mnemoc>
iirc the brom uses a more.... primitive method of accessing the uSD than linux
<mnemoc>
my first cubieboard and my olinuxino-a13 weren't booting from uSD because of dust in the slot. solved in both cases by blowing it a bit
<hipboi>
really, i though you were just kidding...
<mnemoc>
nope, I wasn't
<mnemoc>
i really solved the uSD booting problem of those two devices with a breath
vicenteH has joined #linux-sunxi
<hipboi>
LOL, now both cubieboard and cubieboardii works
<mnemoc>
dust?
<hno>
hipboi, there is two separate issues. a) Sometimes booting from NAND on a cold boot where SD boot was expected. This is due to some SD cards taking too long to initialize on power on, making BROM time out and try NAND instead.
<hno>
b) u-boot SPL chokes on some SD cards. Not sure what this is about.
<hno>
a) can not be fixed. b) is fixable once understood what is actually going wrong. Trouble is that all cards I have access to works so not much data to work on.
<hipboi>
hno, ok, most time a happens
<hno>
Holding the CPU reset a little while longer at power on should help I think.
<hno>
I meant "a) can not be fixed in software" above.
<hipboi>
a) understood
tinti has joined #linux-sunxi
Undertasker has joined #linux-sunxi
focus has quit [Remote host closed the connection]
focus has joined #linux-sunxi
<shineworld>
I've often problems with uSD boot. Sometime boot from uSD (90% of cases) but 10% boot from NAND
<shineworld>
I'm in development phase so all my boot should be from uSD till I reach a final product that will be putted on nand
<Undertasker>
maybe timinh problem. Did you try another SD?
<Undertasker>
timing
<ssvb>
mnemoc: btw, I also seem to have a "dust" problem with Mele A2000
<ssvb>
libv: apparently 480MHz memory clock is not so stable on my Mele A2000 after all, this is what was causing failures for "glmark2-es2 + memtester" combo
<shineworld>
I'm using two different brand SD... I've to note who create problem in next boots
<Undertasker>
Also, does your device have a uSD slot, ar are you using an uSD to SD adapter?
<Undertasker>
You must also note, that uSD cards are not meant to be removed very often. Their connector isn't as robust as that of SD/MMC/CF
<Undertasker>
So the connector wears out sooner or later
<ssvb>
libv: now I need to figure out what is the stable maximum, but at least I'm glad it was not something related to removing the mali 64 MiB dedicated memory reservation :)
vgrade has quit [Quit: Lost terminal]
<hipboi>
ssvb, not all the board can run at 480M stable as cubieboard
<hipboi>
i mean other a10 boards
<hipboi>
the same happens on a20
<ssvb>
hipboi: thanks for the info
<ssvb>
what are you using to test stability?
<hipboi>
memory tester
<hipboi>
usually the stock freq is not the max
<hipboi>
usually 10% reserved
<ssvb>
but it also probably depends on the workload, 480MHz memory works fine on Mele A2000 with only CPU load
<ssvb>
loading both CPU and GPU causes some interesting effects
<libv>
ssvb: i also stopped clocking my odroid-x2 at 2GHz
<libv>
ssvb: after building mesa all afternoon yesterday, it ended up becoming flaky
<libv>
temperature in this part of .eu was also hitting 25°C, but with the usb fan attached, that shouldn't have been the core problem
<libv>
ssvb: all hw is crap :)
shineworld has quit [Remote host closed the connection]
vicenteH has quit [Ping timeout: 272 seconds]
krimpsok has joined #linux-sunxi
<krimpsok>
Anyone here using the 3.0 sunxi kernel in combination with android and the cubieboard?
ganbold_ has joined #linux-sunxi
_BJFreeman has joined #linux-sunxi
_BJFreeman is now known as BJFreeman
hipboi has joined #linux-sunxi
hipboi has quit [Quit: Leaving]
<bsdfox_>
hno, I mean using some system calls rather than doing everything through file access
<lunra>
On the Pi, people open /dev/mem and directly write to the registers, you could read the kernel source and do that, unless someone's made an easy library already
jukivili has quit [Quit: "-leaving-"]
jukivili has joined #linux-sunxi
ganbold_ has quit [Ping timeout: 276 seconds]
ganbold_ has joined #linux-sunxi
<krimpsok>
The sunxi kernel should still work with android when I use the sun4i_crane_defconfig right?
krimpsok has quit [Ping timeout: 245 seconds]
rzk has joined #linux-sunxi
tinti_ has quit [Ping timeout: 272 seconds]
rz2k has quit [Ping timeout: 264 seconds]
krimpsok has joined #linux-sunxi
tinti_ has joined #linux-sunxi
<rm>
not a lot of people care about android
<rm>
I'd say anyone buying CB to just use in android, is pretty silly
rellla has joined #linux-sunxi
rellla has quit [Remote host closed the connection]
mdfe has joined #linux-sunxi
paulk-desktop has joined #linux-sunxi
mdfe has quit [Remote host closed the connection]
gzamboni has quit [Ping timeout: 256 seconds]
<techn_>
krimpsok: yes.. It should work
<hno>
bsdfox_, files are used via system calls.
<hno>
lunra, the pio program in sunxi-tools can do that.
<bsdfox_>
hno, ok. I probably shouldn't be concerned with performance issues then?
<hno>
/sys gpio access is slower than direct mapped register i/o, but a hell lot easier and more flexible cross different hardware.
<bsdfox_>
yeah. I don't have a datasheet for this proc (not sunxi) anyway to find out the register addresses
<bsdfox_>
I'll try it out and see how it works
<hno>
plus can be layered to add led functions etc without changing a single line of code.
ganbold_ has quit [Remote host closed the connection]
eebrah has quit [Ping timeout: 248 seconds]
krimpsok has quit [Quit: Page closed]
eebrah has joined #linux-sunxi
ganbold_ has joined #linux-sunxi
vgrade has joined #linux-sunxi
focus_it has joined #linux-sunxi
focus_it has quit [Quit: Leaving]
gzamboni has joined #linux-sunxi
joker_ has joined #linux-sunxi
<joker_>
hy
<joker_>
i've a problem with the marsboard
<joker_>
can someone help me?
<techn_>
"Don't ask to ask. Just ask!" :)
<joker_>
ah ok ... i dont know why but my marsboard dont work ... the power led is up but it seem to be dead
<techn_>
you have serial?
<joker_>
what do you mean?
<techn_>
serial cable.
<techn_>
but if you dont then you could try better power supply
<joker_>
i've a ftdi converter from sparkfun
<techn_>
what boot log says?
<techn_>
My cubie didn't start when it was connected to my monitors usb hub. When connected directly to computer it works well
<joker_>
my problem is that hdmi dont display nothing and usb otg is the same
<techn_>
With serial cable output you can track where it hangs
<joker_>
if i dont put sd card inside it power all of 4 leds near ethernet port but there are the same no output, nothing