00:55
gavlee has quit [Read error: Connection reset by peer]
00:57
gavlee has joined #linux-amlogic
01:10
zkrx has quit [Ping timeout: 260 seconds]
01:16
zkrx has joined #linux-amlogic
01:46
<
gavlee >
big bunch of errors running that ^^
01:48
<
gavlee >
this is from the stock firmware on x96 air, had to install termux and dtc knew there was a way :D
02:47
camus1 has joined #linux-amlogic
02:48
kaspter has quit [Ping timeout: 258 seconds]
02:48
camus1 is now known as kaspter
02:50
gavlee has quit [Excess Flood]
02:50
gavlee has joined #linux-amlogic
04:58
Barada has joined #linux-amlogic
05:33
<
chewitt >
oh wow .. that's like 45+ seconds of boot on Android
05:34
<
chewitt >
mainline is normally done in under 10 seconds :)
07:03
random_yanek has quit [Ping timeout: 246 seconds]
07:07
warpme_ has joined #linux-amlogic
07:10
random_yanek has joined #linux-amlogic
08:00
<
narmstrong >
yeah android does lots of weird and complex stuff like recompiling apps before loading them
08:51
Elpaulo has quit [Read error: Connection reset by peer]
08:51
Elpaulo has joined #linux-amlogic
09:35
<
lvrp16 >
narmstrong: sometimes I think it's way too over-engineered. build takes an hour on 16 threads and 32GB of DRAM.
09:56
mastertheknife has quit [Ping timeout: 265 seconds]
09:58
mastertheknife has joined #linux-amlogic
10:09
<
mace >
two n2's arrived yesterday, shiny
10:09
<
mace >
thought I'd have a go at building for that before i get on with the sztomato boards
10:10
<
mace >
(which, it turns out, also don't have mipi dsi exposed :( )
11:26
sputnik_ has quit [Ping timeout: 256 seconds]
11:33
ldevulder__ has joined #linux-amlogic
11:35
ldevulder_ has quit [Ping timeout: 265 seconds]
11:38
ldevulder has joined #linux-amlogic
11:40
ldevulder_ has joined #linux-amlogic
11:41
ldevulder__ has quit [Ping timeout: 264 seconds]
11:44
ldevulder has quit [Ping timeout: 264 seconds]
11:55
ldevulder_ is now known as ldevulder
12:12
<
narmstrong >
only the vim3 & vim3l has the mipi dsi exposed
12:21
<
mace >
narmstrong: nod - annoyed with sztomato for selling me these boards as having it
12:22
yann|work is now known as yann
12:28
<
mace >
how critical is it to have a FIP from the vendor/amlogic to get a booting board?
12:28
<
narmstrong >
mace: what do you mean by critical ? is it an issue ?
12:29
<
narmstrong >
not really, they are perfectly functional so far
12:29
<
mace >
sztomato want a few k usd and an nda to give me the sdk for these boards
12:29
<
mace >
i'm not wildly convinced i want to spend that
12:30
<
narmstrong >
it depends how much time they spent on the sdk, it theyr only clones the amlogic code and changed the DDR freq, it's not worth it
12:30
<
mace >
having booted them up, they look virtually stock
12:30
<
mace >
both have amlogic splash screens at boot
12:31
<
narmstrong >
so you can take the khadas or hardkernel releases, strip out their changes and you have the vanilla amlogic sdk
12:31
<
mace >
right, cool, thanks
12:31
<
narmstrong >
take the amlogic model from the u-boot prompt and rebuild it in amlogic sdk
12:32
<
mace >
have khadas/hardkernel been pushing their changes to the kernel back upstream?
12:34
<
narmstrong >
virtually no, we have feedback from khadas and they give samples, but recently both started providing upstream based images
12:36
<
mace >
these boards are a311d and s905d3 based
12:36
<
mace >
would be nice to get them usable
12:36
jelly has quit [Ping timeout: 246 seconds]
12:36
cyrozap has quit [Ping timeout: 264 seconds]
12:36
The_CooIest has quit [Ping timeout: 264 seconds]
12:36
<
mace >
try and salvage the money/time they've cost me already
12:39
jelly-home has joined #linux-amlogic
12:39
<
chewitt >
do the boards boot from emmc or SD card, i.e. where is the boot firmware ?
12:40
<
mace >
on board emmc
12:40
<
mace >
i've managed to get a uart console but not been able to break into uboot yet
12:40
<
chewitt >
put that on an SD card and see whether any of the G12B (A311D) or SM1 (S905X3) device-trees work
12:41
<
chewitt >
edit the dtb name in uEnv.ini
12:42
<
mace >
just dd that (after gunzip) straight onto a card aye?
12:42
<
chewitt >
if you have uart working just hold down space or enter during boot
12:42
<
chewitt >
correct, dd
12:43
<
mace >
have tried holding down space etc, but timeout is set to 0. i need to try a wired keyboard, in case that makes a difference
13:01
gaspode has quit [Quit: Woof bloody woof.]
13:02
gaspode has joined #linux-amlogic
13:06
<
gavlee >
I have trouble compiling android on this puny box that's just :o
13:07
<
gavlee >
but yeah ,when android takes that and I can do some much more with a LE image in 100MB :D
13:13
<
mace >
chewitt: ok dug the board out - doesn't appear to even attempt to boot from sd. plugged in wired keyboard - uboot blasts straight past the hold down space prompt
13:13
<
chewitt >
try enter, shift, etc.
13:15
<
mace >
i'm an idiot :D
13:15
<
mace >
helps if you press space on the uart o.O
13:16
<
mace >
does it change the baud rate once you've broken in? getting gibberish after the 'space' prompt
13:17
<
chewitt >
not that i've experienced .. should be 115200,8,n,1
13:17
<
mace >
i'll try it again then
13:17
<
mace >
might be a bad connection
13:18
Barada has quit [Quit: Barada]
13:22
gaspode has quit [Quit: Woof bloody woof.]
13:24
gaspode has joined #linux-amlogic
13:26
<
mace >
chewitt: bad connection. can probe the mmc from uboot anyhow
13:26
<
chewitt >
printenv .. and c/p the output to pastebin
13:27
<
chewitt >
this will show the bootscript, so we can see what autoscript files it looks foor
13:28
<
mace >
righto, one moment
13:42
<
mace >
looking for aml_autoscript, maybe?
13:47
<
mace >
yup, got it booting off mmc, although it seems to be stuck in a panic bootloop now. will need to see what it's upset about and tweak
14:40
<
chewitt >
g12a_u202_v1 <= worth remembering
14:41
<
chewitt >
aml_dt=sm1_ac202_4g
14:41
random_yanek has quit [Quit: random_yanek]
14:42
<
chewitt >
it supports the IR remote (which you prob. don't have) being used to trigger 'update' mode
14:42
<
mace >
nope don't have - i do have a generic one for something else though
14:43
random_yanek has joined #linux-amlogic
14:44
<
chewitt >
upgrade_key=if gpio input GPIOAO_6; then echo detect upgrade key; run update;fi;
14:44
<
chewitt >
so there's a reset button somewhere
14:44
<
mace >
yup right next to the 12v input
14:45
<
chewitt >
and it should look for aml_autoscript
14:45
<
chewitt >
so the LE boot scripts on the SD card should work, if you trigger update mode
14:45
<
chewitt >
then it's all down to "guess the device tree"
14:45
<
mace >
there's another header labelled key with 8 pins, key/gnd/ledg/ledr/key/gnd/ir/5v
14:46
<
mace >
i manually loaded aml_autoscript and booted it by hand, seemed to work well enough
14:46
<
mace >
ignoring kernel panic
14:46
<
chewitt >
which dtb was that?
14:47
<
chewitt >
this is the S905X3 device I guess
14:47
<
mace >
it was handy
14:47
<
mace >
dtb/meson-g12b-gtking.dtb
14:47
<
mace >
i literally grabbed one at random
14:48
<
mace >
u200 might be worth a try next
14:48
<
chewitt >
S905X3 = SM1
14:49
<
chewitt >
pick VIM3L or C4, then X96-air/A95XF3-air
14:49
<
chewitt >
still SM1
14:49
<
chewitt >
VIM3L is the closes match by SoC, but SBCs are more likely to have different plumbing
14:51
<
chewitt >
the reset button is something you press, then power on, and when you release the button the signal is set to enter update
14:52
<
chewitt >
aka "toothpick" update, as on Android boxes the button is normally hidden and you need a tootpick or unfolded paper-clip to press the button
14:52
<
mace >
i didn't imagine it would be that easy :D
14:53
<
chewitt >
it's also fiddly as arse to get the timing right when the bootdelay is set to zero
14:53
<
chewitt >
but eventually you get it right
14:53
<
chewitt >
the SEI610 is also an SM1 device
14:53
<
chewitt >
(remembering)
14:54
<
chewitt >
there are known FIP sources for SEI610 and VIM3L, and C4, but Hardkernel normally modifies theres so they won't run elsewhere
14:55
<
chewitt >
but experimenting with those (and mainline u-boot) requires you to backup and then erase emmc
14:56
<
mace >
is there an easyish way to do that?
14:56
<
chewitt >
yes, once you get a booting (or partially booting) dtb
14:56
<
mace >
i saw mention somewhere of some amlogic tool for flashing
14:57
<
mace >
ah just dump from linux?
14:57
<
chewitt >
Amlogic Burning Tool
14:57
<
chewitt >
but that's for writing not capturing the image
14:57
<
chewitt >
dd works fine
14:57
<
chewitt >
slow.. but then everything will be slow
15:49
The_Coolest has joined #linux-amlogic
16:05
<
mace >
chewitt: silly question, but won't there be a dtb on the emmc already?
16:27
armoon has joined #linux-amlogic
16:55
<
chewitt >
yes, but it looks for aml_autoscript first, so we can override that with our own boot scripts
16:58
<
mace >
i was more wondering if the dtb can be used to boot libreelec
17:06
sputnik_ has joined #linux-amlogic
17:12
sputnik_ has quit [Ping timeout: 265 seconds]
17:46
<
chewitt >
nope, device-trees are kernel version dependent
17:54
armoon has quit [Ping timeout: 245 seconds]
19:00
ldevulder has quit [Quit: Leaving]
21:55
sputnik_ has joined #linux-amlogic
22:00
warpme_ has quit [Quit: Connection closed for inactivity]
22:17
<
mace >
chewitt: got a few minutes to try all the configs you suggested, none boot
22:20
<
mace >
chewitt: all of them seem to crash in video_keeper_init
22:20
<
mace >
null pointer dereference
22:35
narmstrong has quit [Ping timeout: 272 seconds]
22:36
steev has quit [Ping timeout: 260 seconds]
22:36
ldts has quit [Ping timeout: 260 seconds]
22:37
khilman has quit [Ping timeout: 272 seconds]
22:37
saintdev has quit [Ping timeout: 260 seconds]
22:37
lvrp16 has quit [Ping timeout: 260 seconds]
22:37
jbrunet has quit [Ping timeout: 272 seconds]
22:38
jbrunet has joined #linux-amlogic
22:39
khilman has joined #linux-amlogic
22:41
ldts has joined #linux-amlogic
22:41
lvrp16 has joined #linux-amlogic
22:52
yann has quit [Ping timeout: 258 seconds]
22:54
narmstrong has joined #linux-amlogic
22:54
steev has joined #linux-amlogic