kaspter has quit [Read error: Connection reset by peer]
kaspter has joined #linux-sunxi
cnxsoft has joined #linux-sunxi
kaspter has quit [Remote host closed the connection]
kaspter has joined #linux-sunxi
ganbold has joined #linux-sunxi
cnxsoft has quit [Ping timeout: 240 seconds]
cnxsoft has joined #linux-sunxi
p1u3sch1 has quit [Ping timeout: 244 seconds]
p1u3sch1 has joined #linux-sunxi
vagrantc has quit [Quit: leaving]
kaspter has quit [Ping timeout: 264 seconds]
reinforce has quit [Quit: Leaving.]
IgorPec has joined #linux-sunxi
JohnDoe_71Rus has joined #linux-sunxi
IgorPec has quit [Ping timeout: 240 seconds]
|Jeroen| has joined #linux-sunxi
mossroy has joined #linux-sunxi
IgorPec has joined #linux-sunxi
kaspter has joined #linux-sunxi
mosterta has joined #linux-sunxi
mosterta has quit [Ping timeout: 240 seconds]
Netlynx has joined #linux-sunxi
Netlynx has quit [Changing host]
Netlynx has joined #linux-sunxi
Amit_t_ has joined #linux-sunxi
staplr has quit [Remote host closed the connection]
staplr has joined #linux-sunxi
apritzel has joined #linux-sunxi
|Jeroen| has quit [Quit: dada]
apritzel has quit [Ping timeout: 244 seconds]
zuikis has joined #linux-sunxi
Amit_t_ has quit [Ping timeout: 250 seconds]
jstein has joined #linux-sunxi
TheSeven has quit [Ping timeout: 258 seconds]
[7] has joined #linux-sunxi
<GeneralStupid>
i cannot start X i get DRM: mali_lastclose, is there a way to fix this?
tkaiser has joined #linux-sunxi
Netlynx has quit [Quit: Leaving]
apritzel has joined #linux-sunxi
staplr has quit [Remote host closed the connection]
kaspter has quit [Ping timeout: 240 seconds]
apritzel has quit [Ping timeout: 244 seconds]
maz_ has joined #linux-sunxi
jernej has joined #linux-sunxi
Amit_t_ has joined #linux-sunxi
<tkaiser>
GeneralStupid: Is this on AMD or Intel? Win95 or OS/2?
<GeneralStupid>
tkaiser: allwinner h3, devuan :)
apritzel has joined #linux-sunxi
<GeneralStupid>
this is after i tried armbian
maz_ has quit [Ping timeout: 264 seconds]
IgorPec has quit [Ping timeout: 246 seconds]
mosterta has joined #linux-sunxi
<tkaiser>
GeneralStupid: Well maybe choosing another/better kernel would be an idea?
<GeneralStupid>
-.- i used that image before
jernej has quit [Ping timeout: 258 seconds]
<tkaiser>
GeneralStupid: No idea what that means. At least with latest Armbian desktop images starting HW accelerated X is not an issue at all.
<GeneralStupid>
tkaiser: i already had a kernel with hardware acclerated x and stuff. i just switched back from armbian to my debian
apritzel has quit [Ping timeout: 244 seconds]
Seppoz has joined #linux-sunxi
Amit_t_ has quit [Ping timeout: 250 seconds]
<GeneralStupid>
tkaiser: you want to told me that i could not use my old modules because i used newer ones?
<tkaiser>
GeneralStupid: Nope, since I've no idea about all this GUI stuff but used Armbian desktop builds the last few days testing DRAM on BPi M2+, NanoPi M1 and Beelink X2, at least I know that it works with Armbian. So problem solved ;)
<tkaiser>
GeneralStupid: But using modules for 3.4.39 with kernel 3.4.112 and vice versa should not work :)
<GeneralStupid>
iam using two completely built kernels incl. modules. no mixups there
<GeneralStupid>
But i will try to use the armbian kernel on my debian. Should work, i guess :D
<tkaiser>
GeneralStupid: And to be honest: I've not the slightest idea whether Mali BLOBs also have to be considered (never dealt with this desktop stuff, not even in Armbian itself)
<GeneralStupid>
tkaiser: if u knew that before i had payed more for an raspberry pi ...
<tkaiser>
GeneralStupid: Don't understand a single word. Who had payed for what why?
<GeneralStupid>
i bought an orangepi because of the cheap price compared to a raspberry pi
<tkaiser>
I bought several OPi because of the superiour performance compared to any RPi (more available network/IO bandwidth)
<GeneralStupid>
tkaiser: thats absolutely correct. But if you want to use video it is not the best decission.
<GeneralStupid>
tkaiser: i bought two. One without GUI. that one works fine :D
jernej has joined #linux-sunxi
JohnDoe5 has joined #linux-sunxi
<tkaiser>
GeneralStupid: Well my OPi PC plays video HW accelerated without any problems that not even an RPi 3 can play without overclocking. So obviously it depends on software/settings. But as already said: I've no idea, just made some tests and looked at stuff like http://linux-sunxi.org/User:Rellla/Armbian
<tkaiser>
GeneralStupid: Apart from that people seem to be happy with jernej's unofficial OpenELEC fork (which is sort of a problem since CedarX is still used there)
<GeneralStupid>
tkaiser: i bought it before armbian had support. so i built everything by myself -.-
JohnDoe_71Rus has quit [Ping timeout: 252 seconds]
<GeneralStupid>
tkaiser: maybe it is a creepy setup because of that :D
<tkaiser>
GeneralStupid: IIRC libvdpau with Cedrus support was available in Feb. So before most probably there were many problems with this and that. Also the folks in Orange Pi forums tend to reinvent the wheel for no reasons.
apritzel has joined #linux-sunxi
zuikis has left #linux-sunxi [#linux-sunxi]
<GeneralStupid>
tkaiser: ok you wont believe... but its the truth
zuikis has joined #linux-sunxi
<GeneralStupid>
tkaiser: i just pushed my old image to the sd card again and it works...
<GeneralStupid>
didnt change anything
lemonzest has joined #linux-sunxi
<tkaiser>
GeneralStupid: Since I've to deal with crappy SD cards too often (bug reports in Armbian forum) I wouldn't writing an OS image to SD card 'didnt change anything' ;)
<tkaiser>
*call
jernej has quit [Ping timeout: 260 seconds]
<GeneralStupid>
i actually have an samsung EVO -.-
jernej has joined #linux-sunxi
cnxsoft has quit [Read error: Connection reset by peer]
cnxsoft1 has joined #linux-sunxi
vagrantc has joined #linux-sunxi
cnxsoft1 is now known as cnxsoft
Igor2 has joined #linux-sunxi
* Igor2
is back with the USB OTG gadget/device problem on H3
reinforce has joined #linux-sunxi
<Igor2>
tkaiser, I went on to try armbian kernel - it didn't boot; I decided to use the armbian kernel source and patches, basically the source tree the build script left there, with my customized loboris .config
<Igor2>
so now it boots all fine, 3.4.112, basically the same version the usboot thing worked with
<Igor2>
otg/usb-gadget fails the same way as it did with the oldish loboris kernel
<Igor2>
I think this narrowes it down: it's not a kernel version orpatch problem, it's some configuration issue
<Igor2>
changed usb_port_type to 2, but no luck :/
staplr has joined #linux-sunxi
<GeneralStupid>
hmm, is there a way to debug the boot process?
<Igor2>
(is there a way to debug usb gadget mode?)
apritzel has joined #linux-sunxi
<GeneralStupid>
Igor2: i have usb issues, too.
<GeneralStupid>
Igor2: only 2 usb ports are working usb2 mode one just works in usb 1 mode.
<GeneralStupid>
the 3.4.112 fixed this for me
<Igor2>
i'm tryiong 3.4.112 too, but it doesn't fix my otg problem
<tkaiser>
Igor2: Yeah, you need to change definition of the port in script.bin. And I remembered another detail. Give me a moment...
<GeneralStupid>
Igor2: did you already search the linux-sunxi wiki?
<Igor2>
meanwhile I did the script.bin change, copying the settings from usboot's
<tkaiser>
GeneralStupid: You're using most probably wrong script.bin (for 2/Plus with PC/One/Lite or vice versa)
<GeneralStupid>
maybe this is worth an entry
<Igor2>
yup, searched all wikis and web and everything
<Igor2>
hmm
cnxsoft has quit [Quit: cnxsoft]
<Igor2>
ok, let me check
<GeneralStupid>
tkaiser: i mounted the image which works fine and copied the kernel, the bin and the modules
<GeneralStupid>
(i changed the bin from mmblk0p1 to mmblk0p2, because i needed to)
<GeneralStupid>
but did not boot
<tkaiser>
Igor2: Do a google search for: g_ether site:armbian.com
<GeneralStupid>
iam trying to build an uImage right now
<Igor2>
what I have is "orange pi PC v1.2" (on the PCB)
<tkaiser>
Igor2: In the OPi Lite thread someone reported that he's using an Orange Pi One soldered to a router to offload OpenVPN encryption through g_ether stuff
<Igor2>
when I have problems like this with other hardware, I usually turn on some debugging/verbosity and look at the process to see how far it gets
<Igor2>
with this OTG problem I couldn't get anything like that, instead I am just trying random config settings and kernels without any chance to know if I got closer or further (assuming there are more than one things that can break)
<ssvb>
hmm, is opi+2e board unavailable on aliexpress now?
<GeneralStupid>
tkaiser: the bin file is uboot, right?
<Igor2>
hehe, this russian openvpn hack guy you directed me to, did a code modification in the kernel
<Igor2>
(I found something similar by someone else, a few hours ago)
<Igor2>
is_udc_enable
<Igor2>
[sic]
<Igor2>
well, yolo, I can try to patch the code to permanently enable udc
<tkaiser>
ssvb: According to the numbers shown in the past the 1st batch was probably only 1000 pieces.
<Igor2>
(although I think I'd need to call that pullup() function somehow - after all the USB device needs to pull up the data line to signal presence and speed to the host)
Amit_t_ has joined #linux-sunxi
<Igor2>
hmm, before I mess with the kernel code, what's the standard, user-way of getting this sunxi_udc pullup called?
<Igor2>
is it a module parameter?
<ssvb>
tkaiser: the statistics shows only 110 orders there, also "no longer available" seems to be a more discouraging message than "temporarily out of stock"
<tkaiser>
Igor2: To increase verbosity use loglevel=7 in boot.cmd (has to be converterd to boot.scr afterwards)
<Igor2>
how do I compile cmd to scr?
<tkaiser>
ssvb: I know but IIRC when thirty-something orders were shown then it was ~650 items available
* Igor2
is totally new to this arm world
<tkaiser>
ssvb: So maybe the orders only show up when the orderers confirmed shipping?
<Igor2>
(the patch didn't help - I guess I really need proper pullup)
<Igor2>
hmm, or am I using the wrong uImage...
iaglium_ has quit [Quit: Bed Time]
<tkaiser>
ssvb: Regarding DRAM reliability testing: In the meantime we tested also with NanoPi M1 and Beelink X2 and came to the conclusion to go with 624 MHz on all H3 devices.
<tkaiser>
NanoPi M1 and X2 are also horribly overheating (with lima-memtester reducing CPU clockspeed down to 240 MHz and even killing CPU cores)
<ssvb>
ouch, this seems to be pretty abnormal
iaglium has joined #linux-sunxi
<ssvb>
I mean the overheating
<tkaiser>
ssvb: Yes but it's confirmed from a couple of users/devices. I thought this would be due to DDR3 vs. DDR3L just to learn that OPi One/Lite also use DDR3 (and not DDR3L as I've written wrongly in the wiki)
<ssvb>
was mali clocked at 252mhz or 600mhz in this test?
<tkaiser>
600 MHz
<Igor2>
ahh, I don't even have boot.scr
<tkaiser>
And I doubt the budget cooling stuff in the new BSP kernel we're using with Armbian downclocks GPU cores (at least dmesg isn't outputting anything regarding GPU which differs from the old BSP kernel)
<Igor2>
(the system I run is dietpi, I guess it has different conventions)
<ssvb>
tkaiser: but doesn't OPi One/Lite also overheat somewhat more than OPi PC?
<tkaiser>
Igor2: Yes, this is based on loboris and uses u-boot 2011.09
<Igor2>
(the is_udc_enable = 1 didn't help)
<tkaiser>
ssvb: Only slightly, just a little bit.
<tkaiser>
ssvb: One/Lite with cpuburn-a7 without heatsink run with ~850 MHz, Plus 2E is at 1008 MHz, PC Plus somewhere in between and these other devices throttle down to 480 MHz or even lower
<tkaiser>
So maybe size/material (copper inside?) play some role
<Igor2>
large internal ground/vcc planes help a lot distributing the heat
<tkaiser>
Igor2: That would match my experiences touching components like Ethernet Jack. Pretty warm on the new Oranges, cold on BPi M2+
<Igor2>
btw, top/bottom planes are even better
<Igor2>
some chip datasheets, especially for dc/dc controllers with embedded switch place requirements on how big plane you need to have under the component and how many vias you need to place under the chip (in a matrix) so you can transfer the heat to the plane
<Igor2>
(meanwhile: placed a couple of printk()'s in the code)
<Igor2>
it does skips the code that'd set the puppul
<Igor2>
s/puppul/pullup/
<Igor2>
ohh, fsck....
<Igor2>
tkaiser, figured it
<Igor2>
I need the patch _and_ a magic order of commands
<Igor2>
modprobe g_ether first, and then enable sunxi_udc's device mode
<Igor2>
I thought it wouldn't matter
<Igor2>
but it seem the sunxi_udc driver checks if there's already a g_ user and enables the device accordingly...
<Igor2>
g_ether seems to work now
jernej has quit [Ping timeout: 276 seconds]
<Igor2>
tkaiser, thank you for your persistent support!
ganbold has quit [Quit: This computer has gone to sleep]
ganbold has joined #linux-sunxi
<Igor2>
I think I found like 2 or 3 documents about this sunxi OTG - thye probably all had the order right, but as far as I remember none of them mentioned the order mattered
<Igor2>
just tested again - it does not work without the one-line patch
<tkaiser>
Igor2: Care to write a small tutorial to be included into documentation ($somewhere)?
<Igor2>
sure
<Igor2>
is plain text uploaded to an url on my server ok?
<Igor2>
(just reverting all the hacks i tried layer by layer, to see which ones actually matter and which ones don't)
p1u3sch1 has quit [Ping timeout: 272 seconds]
<Igor2>
not sure if anyone is interested, but let me share why I am fighting with this OTG thing
<Igor2>
the main reason is that I am replacing an old linux home server
<Igor2>
it's a very old PC with 3 NICs and raid; it does all the routing/NAT for my home lan, toward 2 independent uplink ISPs (redundancy ftw!) and is sort of a NAS in the same time
<Igor2>
the replacement is a "cluster" of a raspberry PI and an orange PI
<Igor2>
uplink (cable modem) is connected to their primary ethernet port and an USB->eth dongle is used for the LAN
p1u3sch1 has joined #linux-sunxi
<Igor2>
I want to have an emergency backup link between them, which could be just a serial line (even with SLIP), but as the orange PI has this OTG port I thought it'd be funnier to have a faster link
<tkaiser>
Igor2: Plain text somewhere would be totally fine (and outlining the use case there too). I will push it from there into web.archive.org and link to it
<Igor2>
the other use will be in my hosted server, which is x86_64; there will be an auxiliary server, an orange PI installed in the same box, and I need to connect them - and there's no serial line on the PC MOBO :(
<Igor2>
you can also copy it to a wiki of a related project
<Igor2>
(rebooted without the printk's testing if it still works)
<Igor2>
(reverting script.bin)
apritzel has quit [Ping timeout: 244 seconds]
<Igor2>
cool, script.bin didn't matter
<Igor2>
time to write the file
<ssvb>
tkaiser: cpuburn-a7 does not touch the DRAM at all, it runs entirely inside of the L1 cache
<ssvb>
tkaiser: so it probably can't show the DDR3 vs. DDR3L power consumption difference at all, while lima-memtester definitely stresses the DRAM a lot
camh has quit [Ping timeout: 276 seconds]
<GeneralStupid>
how can i boot zImage with uboot from a FAT partition?
camh has joined #linux-sunxi
<ssvb>
tkaiser: if you want to maximize the power consumption, it's probably best to run cpuburn-a7 together with lima-textured-cube (not lima-memtester)
<ssvb>
tkaiser: and we may try to load the Cedar VPU with some work too :-)
<tkaiser>
ssvb: Good to know. But currently I'm pretty happy to have some results regarding DRAM reliability so that we can simply rely on the 624 MHz we use for all other H3 boards right now.
<MoeIcenowy>
and in the aw sdk driver the code to deal with such chips are blobs
<MoeIcenowy>
(gsl1680 is such an unstable chip that lots of things should be dealt with the driver
<MoeIcenowy>
such as edge fix
<ssvb>
tkaiser: still what about the DRAM reliability after FEL boot? has anyone managed to figure out something?
<MoeIcenowy>
and dithering
<ssvb>
tkaiser: I'm really worried about this, because it is an unresolved mystery which may mean that we have some latent reliability problems
staplr has quit [Remote host closed the connection]
<ssvb>
tkaiser: was it limited to only one poor banana board?
<MoeIcenowy>
KotCzarny: My router is going to suffer high CPU load, for it's a low-end atheros router, only capable of running basic functionality of OpenWRT
<KotCzarny>
um, can i ask you how many unused allwinner boards you own?
<MoeIcenowy>
KotCzarny: no any boards cap of ethernet yet
<KotCzarny>
experiment with dual usb-eth ?
<bbrezillon>
MoeIcenowy: I can't find the NAND datasheet :-/
<bbrezillon>
MoeIcenowy: hm, it seems to be the same generation of NAND we have on the CHIP, so u-boot and linux version should work for you