<MoeIcenowy>
jernej: how much progress have you made on mainlining your H3 U-Boot HDMI?
<jernej>
MoeIcenowy: I think I removed most of unnecessary code. Now I'm renaming functions and making defines
<jernej>
MoeIcenowy: What's the rush?
<MoeIcenowy>
just be curious
<MoeIcenowy>
as I want to get my U-Boot-Linux combined build updated after 4.9 is released
<MoeIcenowy>
and I want a patch clean enough at that time :-)
<jernej>
anyway, first round will be RFC anyway, because I'm not sure how to properly integrate changes in clocks and if my division between new and old DE is right
<jernej>
have to go
<jernej>
bye
<MoeIcenowy>
bye
<MoeIcenowy>
if you need, I can test on my legacy devices
jernej has quit [Ping timeout: 250 seconds]
kaspter has joined #linux-sunxi
lkcl has quit [Read error: Connection reset by peer]
lkcl has joined #linux-sunxi
IgorPec has joined #linux-sunxi
<tuxillo>
moin
terra854 has quit [Quit: Connection closed for inactivity]
terra854 has joined #linux-sunxi
tkaiser has joined #linux-sunxi
<tkaiser>
H5 datasheet (705 pages) sent by Xunlong. Exceeds 16 MB upload limit (25 MB) :(
<ErwinH>
Had my Orange Pi One running for 35 minutes, idling at 40 degrees celcius. The surface temperature of the SoC is now 48 degrees celcius according to my IR thermometer.
Ultrasauce has quit [Ping timeout: 256 seconds]
<KotCzarny>
you might see what mainline a20 thermo driver does, maybe formula is fitting too
Ultrasauce has joined #linux-sunxi
tkaiser has joined #linux-sunxi
<tkaiser>
KotCzarny: C'mon, please stop this.
<tkaiser>
KotCzarny: thermal sensor in A10/A20 sits there by accident, this is different with H3 and all more recent Allwinner SoCs. We also know that we get thermal readouts that are somewhat off (too low) with mainline u-boot compared to BSP u-boot
<KotCzarny>
tkaiser, so what do you propose? finding formula by 'calibrating' ?
<tkaiser>
KotCzarny: A10, A13 and A20 needed a formula. With H3 it's easy: It's 10¡C to 15¡C more than reported by sensor
<KotCzarny>
so yeah, 'calibrating'
<tkaiser>
At least that was the difference when I tried this out in early 2016. Same kernel/image only one time used Allwinner's 2011.09 u-boot, the 2nd time mainline u-boot
<tkaiser>
KotCzarny: If you have too much time, then send patches. I prefer to don't give a sh*t and adjust throttling values accordingly
Putti has quit [Ping timeout: 248 seconds]
fkluknav has joined #linux-sunxi
florianH has joined #linux-sunxi
<tkaiser>
KotCzarny: I don't know whether any person using u-boot 2011.09 did thermal measurements on the surface so all we currently know is that numbers differ based on u-boot version for reasons unknown (clock stuff?)
Christos_ has joined #linux-sunxi
<KotCzarny>
heck, thermal readout depending on clocks? thats crazy talk
<tkaiser>
KotCzarny: Feel free to find the real reason. That was a quick though by ssvb. I don't have any clue here at all. I just wasted some of my time to do the comparison.
LargePrime has quit [Ping timeout: 246 seconds]
Christos_ has quit [Client Quit]
<tkaiser>
And then decided to use 'max temp' for throttling somewhat lower than normal
<KotCzarny>
does throttle happen in software or there is throttling settable in soc?
<tkaiser>
To have some safety headroom regarding H3's 125¡C max even if we know that readouts are too low.
<tkaiser>
KotCzarny: Read the famous user manual ;) With BSP kernel maybe ARISC is involved, with mainline it's kernel code (as megi implemented it)
afaerber has joined #linux-sunxi
juri_ has quit [Ping timeout: 250 seconds]
gumblex has quit [Quit: Bye.]
<tkaiser>
IMO the only (real9 purpose of those thermal sensors in more recent SoCs is to prevent damage/overheating so lets simply use them for that and choose values that ensure not exceeding some tresholds. No matter how the actual numbers the driver produces look like.
juri_ has joined #linux-sunxi
<tkaiser>
ErwinH: Did you also run some lightweight stuff (sysbench) or something heavy (cpuburn-a7 for example)?
<ErwinH>
No, because it's running without heatsink and I don't trust my image at the moment. Somehow megi's branch doesn't give me a good working image.
Putti has joined #linux-sunxi
Putti has quit [Remote host closed the connection]
Putti has joined #linux-sunxi
gumblex has joined #linux-sunxi
<tkaiser>
ErwinH: Hehe, just let 'sudo armbianmonitor -m' run in another shell and stop if you feel temperatures increase too much. And do the 'thumb check' first.
<tkaiser>
If you press your thumb on H3 and reported temperature gets below 37¡C you know something's wrong (either with you or the numbers a driver is spitting out ;) )
LargePrime has joined #linux-sunxi
Worf has joined #linux-sunxi
<ErwinH>
I'll try that in a bit.
juri_ has quit [Ping timeout: 246 seconds]
juri_ has joined #linux-sunxi
<tkaiser>
ErwinH: megi's throttling tresholds are pretty low, there's not that much to fear
<dgp>
The temp reading from the THS driver for mainline seem to be totally out of whack on the h2.. so I start throttling at 40c I think
<dgp>
I can compile a kernel on all 4 cores without it melting down like that :)
<ErwinH>
Ok, but I can't get megi's branch to create a usable image :(
<tkaiser>
dgp: I would assume H2+ being identical with H3 here. But my tests date a bit back. When testing temperatures did not exceed 70¡C and cpufreq jumped between 816 and 1008 MHz (without using the possible 912 MHz in between)
<tkaiser>
dgp: Yes, orangepi-4.9 branch
<KotCzarny>
i would assume h2+ to be cooler than h3 because of few components removed/disabled
<KotCzarny>
but it might be hotter because of small board
<dgp>
tkaiser: I felt it getting really hot but still reading 45 - 50 which didn't seem right
juri_ has joined #linux-sunxi
<tkaiser>
ErwinH: Do the thumb test. You body is a huge heatsink and heat dissipation is more effective.
<KotCzarny>
hehe, thumb test with soc running cpuburn might hurt
<tkaiser>
dgp: What's really hot? It hurts when you touch the SoC?
<dgp>
ErwinH: I have that repo + rc6 merged
<ErwinH>
That was a thumb test, or index finger actually :)
<dgp>
tkaiser: yeah, like almost burning skin hot
<tkaiser>
KotCzarny: Idle temperature
<KotCzarny>
dgp, grab ir thermometer, they usually have 'object temperature' setting which is not that bad, they are cheap too
<jelle>
oh nice u-boot has 2 new maintainers for sunxi
<KotCzarny>
and there is a high chance you have one already
<tkaiser>
dgp: Then there's something really wrong.
<dgp>
KotCzarny: I don't have one at the moment but I will get one from banggood if I remember
<dgp>
tkaiser: yeah that's what I thought. I found the temp reading for the highest temp I thought would be acceptable and adjusted the trip points for that
<tkaiser>
dgp: If it's too hot then this won't work. But normally the thumb test is pretty simply, start temperature monitoring and press your thumb on the SoC and watch temp behaviour
<tkaiser>
dgp: Ah, by fiddling around with trip points wrongly (BSP kernel) I managed a few times to do the wrong thing. Always feeding 1.3V even at lowest clockspeeds.
juri_ has quit [Ping timeout: 248 seconds]
<tkaiser>
dgp: But I doubt that can be possible with mainline
<tkaiser>
But with BSP kernel with wrong settings ARISC got crazy in the background
<dgp>
It's possible, but the pi zero only has 1.1v or 1.3v. My regulator setup seems to work as I'm reading 1.1v most of the time
<tkaiser>
dgp: Multimeter?
<dgp>
yeah, there is a test point for the cpu voltage regulator output
<dgp>
I added regulators in my dt for the dram etc too so something else can't grab the gpios.. but for some reason they get turned off at boot so everything stops ;)
<MoeIcenowy>
dgp: did you specify regulator-boot-on?
<MoeIcenowy>
or regulator-always-on (I think this is proper
juri_ has joined #linux-sunxi
<dgp>
MoeIcenowy: yes, but the gpio seems to go into the disabled state when the pinmux is configured
<dgp>
so it gets set to output, goes low, DRAM turns off
silviop has joined #linux-sunxi
<MoeIcenowy>
my RMB ￥19.9 UASP-capable USB-to-SATA bridge arrived
<MoeIcenowy>
there's all the docs from aw, but, most is in Chinese
<MoeIcenowy>
KotCzarny: I will reboot to Windows to test it
<MoeIcenowy>
My friend has made use of UASP of this bridge under Windows
<silviop>
Ok perfect for me , i'm investigating about USB0 that is not used for wifi in A33 tablet
scream has joined #linux-sunxi
<silviop>
In this reference USB0 is used for wifi , i'm looking reference with wifi on SDIO
f0xx has quit [Quit: terminated!]
f0xx has joined #linux-sunxi
<silviop>
ok this is my tablet :https://github.com/pine64-android/device-softwinner/blob/master/astar-common/Documents/A33/Hardware/%E6%A0%87%E6%A1%88%E5%8E%9F%E7%90%86%E5%9B%BE/a33(a23)_tablet_std_v1_0_162.pdf
juri_ has quit [Ping timeout: 256 seconds]
gaby has joined #linux-sunxi
juri_ has joined #linux-sunxi
lkcl has quit [Ping timeout: 256 seconds]
lkcl has joined #linux-sunxi
<MoeIcenowy>
silviop: the reference provided both way
juri_ has quit [Ping timeout: 252 seconds]
<MoeIcenowy>
yes it's a ns1066 chip
juri_ has joined #linux-sunxi
<tkaiser>
MoeIcenowy: Most probably no UAS support in kernel. But it also depends on your USB host controller in your PC too. My first test would be therefore mainline kernel with A20 or H3 instead ;)
<KotCzarny>
MoeIcenowy: hack the usb id and try again
bugzc has quit [Ping timeout: 260 seconds]
juri_ has quit [Ping timeout: 246 seconds]
juri_ has joined #linux-sunxi
<MoeIcenowy>
I do not want to test UAS now...
<MoeIcenowy>
what I need now is only a cheap bridge :-)
<MoeIcenowy>
as I got a old SATA 2.5" harddisk :-)
<KotCzarny>
unless its really old or broken, im sure it could profit from additional 5MB/s and lower cpu usage :)
<KotCzarny>
and we can profit from confirming ns1066 is worth buying too
<MoeIcenowy>
but many documents say that ns1066 is not capable of uasp...
<KotCzarny>
good to know then
<KotCzarny>
now someone should buy and test ns1068
<MoeIcenowy>
I asked the supporter for info about the UASP support
<KotCzarny>
as this one is advertised to be uasp capable
gumblex has quit [Remote host closed the connection]
gumblex has joined #linux-sunxi
mdsrv has quit [Ping timeout: 250 seconds]
juri_ has quit [Ping timeout: 250 seconds]
juri_ has joined #linux-sunxi
<MoeIcenowy>
oh my bridge is also ORICO
kaspter has quit [Ping timeout: 250 seconds]
<scelestic>
meh, i should have hooked up my esp8266 with esp-link, i rebooted my orange pi and i didnt come back up
juri_ has quit [Ping timeout: 250 seconds]
<scelestic>
i/it*
juri_ has joined #linux-sunxi
tkaiser has quit [Ping timeout: 250 seconds]
f0xx has quit [Quit: terminated!]
f0xx has joined #linux-sunxi
tuxillo has joined #linux-sunxi
<tuxillo>
hi
juri_ has quit [Ping timeout: 240 seconds]
Turl has quit [Remote host closed the connection]
juri_ has joined #linux-sunxi
Turl has joined #linux-sunxi
yann-kaelig has joined #linux-sunxi
Putti has quit [Ping timeout: 248 seconds]
afaerber has quit [Quit: Ex-Chat]
Putti has joined #linux-sunxi
afaerber has joined #linux-sunxi
afaerber has quit [Ping timeout: 240 seconds]
mzki has joined #linux-sunxi
afaerber has joined #linux-sunxi
silviop has quit [Ping timeout: 260 seconds]
chomwitt has joined #linux-sunxi
markon has joined #linux-sunxi
maa has joined #linux-sunxi
markon has quit [Client Quit]
JohnDoe_71Rus has quit [Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org]
tkaiser has joined #linux-sunxi
paulk-minnie has joined #linux-sunxi
willmore has quit [Ping timeout: 248 seconds]
willmore has joined #linux-sunxi
cnxsoft has quit [Quit: cnxsoft]
Putti has quit [Ping timeout: 250 seconds]
av_ has joined #linux-sunxi
gumblex has quit [Ping timeout: 245 seconds]
gumblex has joined #linux-sunxi
Seppoz has joined #linux-sunxi
<Seppoz>
anyone knows if a20 has a built in tempertaure sensor?
libv_ has joined #linux-sunxi
<tkaiser>
Seppoz: there is one inside the touch panel controller.
<av_>
hi everyone, I've got an A10 working fine with u-boot and kernel 4.8, but out of three peripherals I need, only GPIO works: audio and SPI don't.
<Seppoz>
any way to read it?
<Seppoz>
any documentation?
<tkaiser>
Seppoz: mainline kernel?
<Seppoz>
no 3.4
<tkaiser>
Seppoz: Armbian?
<Seppoz>
no custom
<tkaiser>
Seppoz: Ok, there's a method from Tom Cubie which is not that great since it spams to syslog and a binary available.
<Seppoz>
any link?
<Seppoz>
is it open source?
<av_>
how can I get those to work -- do I need a new DTB file? my current dtb appears to include definitions for SPI, but /sys/class/spidev and spi_master are empty
libv has quit [Ping timeout: 260 seconds]
<tkaiser>
Seppoz: Yes, started in LeMaker forum 100 years ago. I would search for sunxi-temp/sunxi_tp_temp
<dan0_0>
Yea, at these smaller sizes, people will pay for more ram
<KotCzarny>
Seppoz: neat
<tkaiser>
KotCzarny: product development, if you look at Pine64 they made their money with clueless people thinking they would *need* as much DRAM as possible and accessories. The Wi-Fi/BT module is the same as on the CHIP but the whole CHIP costs $1 less ;)
f0xx has quit [Ping timeout: 258 seconds]
<tkaiser>
dan0_0: do you know whether it would require 6Gb/12Gb modules for 3GB DRAM? Or can 2 x 4Gb combines with 2 x 8Gb?
<tkaiser>
dan0_0: Me too, but maybe that's the reason. No 6Gb/12Gb DDR3/DDR3L available. But I've to admit that I'm absolutely clueless when it comes to hardware :)
<tkaiser>
Seppoz: Use mainline kernel, thermal readouts there are fine (there were only problems with 4.0 or 4.1 when readouts were below ambient temp or even 0¡C)
<Seppoz>
ok
<tkaiser>
Seppoz: But don
<Seppoz>
its not that easy to switch to mainline kernel tho
<Seppoz>
with all the ios i use
<tkaiser>
't trust too much in it, the thermal sensor in A20 is part of another IP block and not designed to protect from overheating. It's there and it can be used and some linux-sunxi folks did calibration some years ago
<Seppoz>
yea i figured from what i just read the last hafe hour
<tuxillo>
NiteHawk: but just by looking at the patch it looks like a nicer approach imho
<NiteHawk>
tuxillo: yes, it should be more foolproof, especially with regards to older BSD versions that still might require those #defines from the original gist
<NiteHawk>
tuxillo: you may just leave some feedback in the PR discussion on GitHub. if you approve the new patch, i could pull it into sunxi-tools and then close your request
<tuxillo>
yeah
<tuxillo>
will try it soon enough
leviathanch has joined #linux-sunxi
|Jeroen| has joined #linux-sunxi
dh1tw has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
<NiteHawk>
tuxillo: thanks! :)
iamfrankenstein1 has quit [Ping timeout: 256 seconds]
xes_ has joined #linux-sunxi
cptG has joined #linux-sunxi
xes has quit [Ping timeout: 244 seconds]
cptG_ has quit [Ping timeout: 240 seconds]
Putti has quit [Ping timeout: 240 seconds]
xenoxaos has quit [Ping timeout: 256 seconds]
massi has quit [Remote host closed the connection]
xenoxaos has joined #linux-sunxi
|Jeroen| has quit [Quit: dada]
xes_ is now known as xes
reinforce has quit [Quit: Leaving.]
terra854 has quit [Quit: Connection closed for inactivity]
Putti has joined #linux-sunxi
y0g1 has joined #linux-sunxi
hojnikb has joined #linux-sunxi
<hojnikb>
can anyone share debian or armbian desktop image for pc2 ?
<hojnikb>
xunlong image on gdrive is wrong
<tkaiser>
hojnikb: Full GPU and video acceleration needed?
<hojnikb>
i just need desktop
y0g1 has quit [Client Quit]
<hojnikb>
i can use android for video stuff
<hojnikb>
so mali and cedarx is not needed for now.
<hojnikb>
i just want to see if it boots so i can play around with it
<rellla>
cedarx is not vailable anyway
<rellla>
*available
<hojnikb>
well in h3 it seems to be working under linux or atleast something comaptabile to it.
<tkaiser>
hojnikb: 'Needed for now'... I merged the 'Debian Server' stuff from Xunlong with an Armbian desktop image. Worked but already deleted since why?
<hojnikb>
otherwise 4k playback wouldn't be possible under openelec on a measly a7 cores :)
paulk-collins has quit [Remote host closed the connection]
<hojnikb>
why not
<hojnikb>
this is literally retarded
<hojnikb>
there is not a signle image on xunlong that can be downloaded
<rellla>
i only want to state once more, that cedarx is aw crappy code and cedarx != cedrus
<KotCzarny>
at first look there is only cedarx visible
<KotCzarny>
anyone up to add a notice or shall i cook up something?
<KotCzarny>
done.
<KotCzarny>
also, in regard of recent sunxi-tools rename, there is still ' sunxi-tools: Tools to help hacking sunxi devices' on https://linux-sunxi.org/Main_Page
<KotCzarny>
ssvb: ^
mihlit has joined #linux-sunxi
vagrantc has joined #linux-sunxi
iamfrankenstein1 has joined #linux-sunxi
iamfrankenstein has quit [Ping timeout: 252 seconds]
iamfrankenstein1 is now known as iamfrankenstein
<freemangordon>
Wizzup: seems I bought the right one by chance :) - Q8H_HD , 1024x600, quad core A33
<freemangordon>
jelle: does ^^^ model seems about right to you?
<Wizzup>
freemangordon: I think that may be similar to what I got today (from plesio)
<Wizzup>
I will unpack it tomorrow
<freemangordon>
waht is weird is there is no manufacturer name written on the device, or in the user manual :)
<freemangordon>
ok, lets find the quick howto-boot-linux :)
<KotCzarny>
fmg: 1/ uart pads
<KotCzarny>
(just in case).
<KotCzarny>
2/ sdcard with proper uboot, kernel and the rest
<hojnikb>
hmmm, debian image seems pretty snappy
<MoeIcenowy>
KotCzarny: I think UART pads on A33 devices are usually useless
<MoeIcenowy>
as all my seen A33 devices are designed to use PF2/PF4 as UART pads
<MoeIcenowy>
which can be easily retrieved by uSD Breakout
<freemangordon>
so, is there a "test" image I can do a quick test with?
<hojnikb>
a armbian image would be great
<hojnikb>
an*
<freemangordon>
KotCzarny: I guess I have to disassemble to access uart pads?
<hojnikb>
maybe tkaiser will find his image somewhere :)
<MoeIcenowy>
verified on: (A33) Aoson M751s (no UART pads, but official firmware's boot0/u-boot debugging info is outputed via PF*), iNet D978 Rev02 (UART pads are PF2/PF4); (R16) Parrot (UART socket is PF2/PF4), Nintendo NES Mini (UART test pads are PF2/PF4)
<MoeIcenowy>
freemangordon: maybe it will be better to purchase a uSD Breakout, as what I said
<freemangordon>
MoeIcenowy: sorry, I am new to all this stuff, could you elaborate
<freemangordon>
all this == allwinner
<MoeIcenowy>
you can build a A33 image manually
<MoeIcenowy>
with mainline
<MoeIcenowy>
are you familiar with U-Boot and Mainline ARM Linux?
<freemangordon>
MoeIcenowy: yes, but this is step 2 or 3, I first want to know if I can boot something at all
<freemangordon>
MoeIcenowy: yes
<MoeIcenowy>
ok you can boot just a simple U-Boot.
<freemangordon>
mainline?
<MoeIcenowy>
yes
<freemangordon>
good
<MoeIcenowy>
you can try q8_a33_tablet_1024x600_defconfig
<KotCzarny>
fmg: as MoeIcenowy said, those tablets usually mux the uart. hum.
<MoeIcenowy>
but before going to mainline guys, you should check your touchscreen
<KotCzarny>
MoeIcenowy: is there other way to see serial messages?
<freemangordon>
yeah, I was about to ask for dtb :)
<MoeIcenowy>
freemangordon: for dtb? sun8i-a33-q8-tablet.dtb
<freemangordon>
ok
<freemangordon>
thanks
<MoeIcenowy>
BUT BEFORE YOU GOING TO MAINLINE GUYS
<MoeIcenowy>
CHECK YOUR TOUCHSCREEN
<MoeIcenowy>
IF IT'S A SILEAD GSLX68X
<MoeIcenowy>
PLEASE BACKUP THE ANDROID VENDOR DRIVER
<MoeIcenowy>
A FIRMWARE LIES IN IT
<KotCzarny>
MoeIcenowy: can it boot from sdcard without touching nand?
<MoeIcenowy>
of course yes
<freemangordon>
MoeIcenowy: how to check TS from non-rooted android?
<MoeIcenowy>
try to lsmod?
apritzel has joined #linux-sunxi
<MoeIcenowy>
I think many AW tablets have adb root access
<MoeIcenowy>
especially brandless ones
<freemangordon>
keep in mind I try to not touch android no matter what, so I have no android SDK around :)
<MoeIcenowy>
to be honest it's not so easy to touch android guys from mainline :-)
<MoeIcenowy>
although I did wipe my NAND, for testing mainline A33 NAND support
bugzc has joined #linux-sunxi
<MoeIcenowy>
apritzel: morning
<apritzel>
MoeIcenowy: hi
<MoeIcenowy>
is A64 eMMC support stable now?
<freemangordon>
MoeIcenowy: see, what I am trying to do is not to replace android on nand, at least not initially
<apritzel>
not really
hojnikb has quit [Quit: Page closed]
<apritzel>
MoeIcenowy: it was a hack only anyway
<MoeIcenowy>
freemangordon: to be honest, currently no one can replace it on nand with mainline guys :-)
<freemangordon>
:)
<MoeIcenowy>
I did the experiment, and failed :-(
<apritzel>
MoeIcenowy: also I still see it hang (for up to half a minute, sometimes), but at least it always continues eventually
<MoeIcenowy>
I cannot even make a SPL which can read the main U-Boot binary :-(
<MoeIcenowy>
apritzel: mysterious...
<apritzel>
MoeIcenowy: I have no proof, but I guess this is eMMC driver related
<MoeIcenowy>
is H3 eMMC boards working properly?
<MoeIcenowy>
is A83T eMMC boards working properly?
<apritzel>
MoeIcenowy: I don't have any, so cannot say
<MoeIcenowy>
I think these two revisions do have all small advances
<apritzel>
MoeIcenowy: but I don't remember any complaints
<MoeIcenowy>
montjoie seems to have both
<MoeIcenowy>
so do I
<freemangordon>
well, when it comes to it, I will try, I have some experience with onenand on omap3, however now my aim is userspace so SD card image should suffice
<apritzel>
MoeIcenowy: A64/H5 eMMC is quite a difference, though
<MoeIcenowy>
freemangordon: I think you should surely retrieve enough info from Android
<MoeIcenowy>
at least, script.bin
<MoeIcenowy>
which can be only read by reading /dev/mem now
<MoeIcenowy>
which needs root
<apritzel>
MoeIcenowy: do you want to send out any U-Boot bits for the OPi Zero?
<MoeIcenowy>
but you can try "rootmydevice", the famous aw backdoor
<apritzel>
MoeIcenowy: if not, I could send out what I use here ...
<MoeIcenowy>
apritzel: after the Linux dt is mainlined
<freemangordon>
if I only know how . Seems I will have to RTFM :)
<apritzel>
MoeIcenowy: yeah, OK
<MoeIcenowy>
as Orange Pi One ones works properly here :-)
<apritzel>
MoeIcenowy: yeah, but I have a SPI flash soldered, so need some changes anyway
<MoeIcenowy>
for rootmydevice backdoor, you can google it :-)
<MoeIcenowy>
apritzel: not everyone have it soldered...
<MoeIcenowy>
I still don't dare to solder it by myself
<apritzel>
MoeIcenowy: but the CONFIG_ options will not hurt
<MoeIcenowy>
yes
<MoeIcenowy>
freemangordon: for any problems related to A33 tablet hacking, feel free to ask me ;-)
Andy-D has quit [Ping timeout: 268 seconds]
<tkaiser>
MoeIcenowy: Do you use a special kernel config with your latest A64 USB branch?
<nove>
rellla: one thing, the changes at the Category:Video_Engine should provable be at http://linux-sunxi.org/Video_Engine itself at the top, and not in the middle of the how to name things
<nove>
rellla: is because of caching, the cache for the main page has to be invalidated or just by a edit
<mihlit>
Guys, did I waste money? I bought friendlyarm's NanoPi M2 which uses samsung's S5P4418. I've read that this soc is obsolete, it is not supported in mainline kernel and never will be. Is it true?
<KotCzarny>
mihlit: who knows, this is allwinner socs related channel
<tkaiser>
mihlit: This is a Samsung/Nexell SoC
<KotCzarny>
you can try returning it
<rellla>
nove: ok about the caching. but i don't get the other thing... can you fix it by yourself probably?
<mihlit>
Features looked good, only later I found out, it's not H3 like NanoPi Neo.
<mihlit>
I don't return things I buy if it is me to blame. I did the mistake, so I will have to live with it :)
<mihlit>
on allwinner topic, I made another mistake two years ago, when I bought Phoenix A20. That's another candidate for trash bin, right?
<tkaiser>
mihlit: In case you bought also a FriendlyARM LCD that might fun since this really works flawlessly. And while the kernel is rubbish the OS images aren't.
<mihlit>
tkaiser: yes, their os image works well, but I don't have use for hw with no future (mainline support)
<KotCzarny>
mihlit: dont be confused, a20 based devices arent obsolete
<tkaiser>
mihlit: Regarding this Nexell stuff your only hope is Samsung since they use some of the SoCs in their ARTIK IoT line. Maybe you've luck there. But not that much hope
<KotCzarny>
and keep in mind sata on a20 is 5x faster than on any other allwinner device
<mihlit>
KotCzarny: well, I have other devices with a20, but for Phoenix, there is no device tree available, afaik
<KotCzarny>
try default a20 one?
<nove>
rellla: another thing
<tkaiser>
mihlit: If you have a fex/script.bin and some knowledge then it's not that much of a problem
<KotCzarny>
or just make one using script.fex as a source for pins
<KotCzarny>
and assuming it has vga port, maybe its olimex clone? ;)
<mihlit>
KotCzarny: I've tried image for cubieboard2 which has same nic chip and other similarities, but it boots with usb not working, sata error messages filling console,... not good
<KotCzarny>
then just compare pin setups
<tkaiser>
mihlit: CB2 images are bad anyway since DRAM is clocked way too high always.
<mihlit>
tkaiser: I don't have fex file, they have some images for this board available, but they are for nand and mounting that image does not work
<nove>
rellla: in the description of cedrus, what was reverse engineered was the hardware
leviathanch has quit [Remote host closed the connection]
<tkaiser>
mihlit: There are instructions somewhere to dump this sysconfig.fex stuff from memory, then you can convert it to a normal fex file and have all the pin mappings
<mihlit>
did anyone try to create devicetree from fex file? is it difficult? If it should take me weeks, it won't be worth it. I have just one board...
<tkaiser>
mihlit: Only with 10-15 H3 devices, but there's is super simple because they're all the same more or less. But it shouldn't be that hard if you choose a board that's close to yours.
<nove>
rellla: is not that, what shoudl be moved was only the changes that you made, the left to explain the naming should stay
Mr__Anderson has joined #linux-sunxi
<rellla>
nove: i'm ok with it, by dividing software naming and hardware naming. but feel free to change it if you're not ok with it
afaerber has quit [Quit: Ex-Chat]
deskwizard has quit [Ping timeout: 240 seconds]
<freemangordon>
ok, got u-boot up and running, now need to figure out how to make it run the kernel on uSD card
<tkaiser>
LOL, the next H3 board around the corner
netlynx has quit [Quit: Ex-Chat]
<apritzel>
tkaiser: btw: thanks for digging out the H5 user manual, even without a CONFIDENTIAL watermark!
<jelle>
apritzel: the patched you posted on the u-boot list, do they require the same build process?
<apritzel>
jelle: the split 32/64 build?
iamfrankenstein has quit [Quit: iamfrankenstein]
<apritzel>
jelle: depends ;-)
<apritzel>
actually there is a 64-bit SPL (that's the reason for the new post)
iamfrankenstein has joined #linux-sunxi
<jelle>
ok
<apritzel>
so you just do make pine64_plus_defconfig && make with an AArch64 cross-compiler and get a sunxi-spl.bin plus a u-boot.img
<jelle>
apritzel: oh but I have an orange pi pc 2
<apritzel>
jelle: those patches are just for A64
<apritzel>
I will post the H5 patches later on top of that
<jelle>
yup so I assumed they apply for the orange pi pc 2
<jelle>
oh ahh I see
<apritzel>
jemk has made some progress on the H5 DRAM stuff, I just need to put it on top of this branch (which isn't trivial)
<jelle>
nice
<apritzel>
let me just implement SW4 as a FEL button in the existing H5 SPL ;-)
<apritzel>
(because if you have something in the NOR flash you lose the easy FEL access)
megi has joined #linux-sunxi
<apritzel>
tkaiser: what was that button on the Opi PC 2 used for again on that AW firmware?
<MoeIcenowy>
freemangordon: yes mainline touchscreen driver needs it
<tkaiser>
apritzel: Huh? Which button?
<tkaiser>
apritzel: SW4?
<apritzel>
tkaiser: yes
<apritzel>
didn't you mention something the other day?
<apritzel>
that they have some kind of power off functionality or the like?
<apritzel>
(or what that another board?)
<tkaiser>
apritzel: 'power button', when used with BSP u-boot/kernel then it gets defined as 'sunxi key' and can then wake up from suspend to ram.
<nove>
rellla: again i am nitpicking about wording, at "which aims at becoming a fully 100%", it does not _aims_ or will _become_ it is already is 100%
<tkaiser>
Since 'poweroff' reboots the board and then u-boot sits there waiting for the press of a button.
<apritzel>
tkaiser: so it's not "poweroff", but "pm-suspend" then?
<tkaiser>
apritzel: But as far as I understand this could be used for anything.
<apritzel>
tkaiser: yeah, it's just connected to PL3
<apritzel>
tkaiser: I have it hacked to be a FEL button
<rellla>
nove: i think the "aims" is referred to "will become" , but we can change that to "is"
<tkaiser>
apritzel: Nope, it's really combined functionality. With a specific kernel settings 'poweroff' will set a flag before and issue a reboot. But u-boot will then wait for event sources (button or IR or maybe later also WiFi/BT)
<tkaiser>
apritzel: And on H3 boards with BSP stuff this works pretty well but currently with the BSP stuff for OPi PC 2 consumption is way to high in this mode.
<tkaiser>
apritzel: And what kind of hack is this to get FEL functionality?
<apritzel>
tkaiser: so it will keep the DRAM around and save all the state as "pm-suspend" is supposed to?
<tkaiser>
apritzel: In theory, yes.
<apritzel>
tkaiser: so it's broken, because poweroff means poweroff and pm-suspend means pm-suspend
<apritzel>
tkaiser: which is UNIX: you do as I told you
<apritzel>
tkaiser: if you have something bootable in NP}
<apritzel>
NOR flash, you loose easy FEL access
<apritzel>
because you can't remove the SD card and boot in FEL
<apritzel>
because NOR flash is between SD and FEL in boot order
<apritzel>
you can put in a magic SD card which drops you to FEL
<apritzel>
but this code in SPL will check the state of the button and branch to 0x20 if it is pressed
<apritzel>
so if you have this specially prepared SPL in the NOR flash, it acts like a FEL button
<tkaiser>
apritzel: Now i understand, it's a software and not a hardware hack. I thought you're running around with solder iron in your hand ;)
<apritzel>
tkaiser: I had my share of soldering for the month already with putting that SPI flash on the OPi Zero ;-)
<tkaiser>
apritzel: But that's smart :)
<rellla>
nove: i'm done for today. feel free to change like you want :)
<apritzel>
EVERYONE: rellla releases the Wiki!
<apritzel>
;-)
<rellla>
haha apritzel... who is ever reading the changelog? :p
<apritzel>
me, constantly
<tkaiser>
rellla: Bookmarked, only linux-sunxi bookmark ;)
<rellla>
me, too :)
<apritzel>
I subscribed the RSS feed, actually
<rellla>
apritzel: there is a "minor changes" filter afaik :p
<rellla>
so then there are only 7 edits left
<apritzel>
rellla: well, the "minor" flags seems to be treated differently than in Wikipedia
<apritzel>
AFAIK in WP it's for formatting changes or typos or the like
<apritzel>
here people seem to use it for: "not so important"
<rellla>
-> me, yes
<tkaiser>
me2
<rellla>
to don't nag on apritzel's rss feed ...
<apritzel>
especially tkaiser is the master of understatement here ;-)
<apritzel>
a new Pine64 board? minor stuff ... ;-)
<tkaiser>
apritzel: The only interesting thing with that is LPDDR3 IMO
<tkaiser>
As soon as anyone announces a base board for 4 of them with a switch IC with the same count of RGMII interfaces it might be different ;)
Ntemis has quit [Remote host closed the connection]
<apritzel>
tkaiser, well, and the SPI flash and eMMC ...
deskwizard has joined #linux-sunxi
<apritzel>
tkaiser: wasn't it 8? :-D
<tkaiser>
apritzel: Combining 4 will already be challenging due to thermal problems.
mzki has quit [Ping timeout: 245 seconds]
<tkaiser>
At least they can not be powered from a single USB port ;) http://clusterhat.com
<apritzel>
tkaiser: no way!1!1!! this is a low power ARM chip which displays videos @ 5 W and 35C !1!1
<tkaiser>
apritzel: OMG, you remembered *both* numbers!!1!11!
<nove>
rellla: anyway, i still think the Video Engine page should be similar as before any changes, itis the page for hardware, the software should be explained in their respective pages
<apritzel>
tkaiser: seriously? they power four of those via USB?
mzki has joined #linux-sunxi
<tkaiser>
apritzel: The Pi Zero consumes almost nothing even under 'load'. I measures 800 mW with moronic sysbench test.
<rellla>
nove: but it had cedarx already in the naming scheme list before?! I just added cedrus and put bothe to the top.
<apritzel>
tkaiser: but that's only true for the BCM2835, right?
<tkaiser>
apritzel: Sure. And what helps here is that the Zero saves the USB hub + Fast Ethernet thingie which consumes 360 mW alone while doing nothing
dh1tw has joined #linux-sunxi
<nove>
rellla: cedarx was there because of the usual confusion of how to name the hardware
<apritzel>
tkaiser: neat little idea to use USB OTG as the interface. Slow, but clever
<tkaiser>
apritzel: But way slower than Allwinner Android implementation! ;)
<rellla>
nove: so lets just link cedarx and cedrus on top of the page with a small expanation and put the extended version to the specific sites?