<tkaiser>
jemk: Idle consumption gets even lower below 408 MHz but the biggest step is between 408 and 456 MHz. And I tested all clockspeeds with tinymemtest and cpuminer. What I did not was testing these lower frequencies with memtester. I only tested DRAM clockspeeds with lima-memtester on NEO above the vendor's 432 MHz (and got deadlocks at 672 MHz after an hour)
Mr__Anderson has quit [Remote host closed the connection]
putti_ has quit [Ping timeout: 244 seconds]
MiLeon has joined #linux-sunxi
reev has joined #linux-sunxi
<reev>
Hi I have banana pi routerboard and i am running centos 7 on it but interface eth0 is showing down after reboot I even ran ip link set dev eth0 up but still it is showing down
<reev>
what could be the problem?
<KotCzarny>
mainline or legacy kernel? also there is a tool to configure onboard switch (which shows as eth0)
NiteHawk has quit [Remote host closed the connection]
imcsk8 has quit [Ping timeout: 260 seconds]
NiteHawk has joined #linux-sunxi
NiteHawk has quit [Changing host]
NiteHawk has joined #linux-sunxi
imcsk8 has joined #linux-sunxi
<reev>
KotCzarny uname -a shows 4.2.3-200.el7.armv7hl
<reev>
what kind of tool ?
<reev>
it doesn't even have lspci
<KotCzarny>
wonder if they backported switch code themselves, as the support for it was only added recently
<reev>
so what should i do now?
<KotCzarny>
also, keep in mind this board is fine for HOME purposes, if you want to use it for real work you're gonna have hard time
<KotCzarny>
are you stuck for centos?
<KotCzarny>
if not, try armbian
iamfrankenstein has quit [Ping timeout: 244 seconds]
<KotCzarny>
you can also try compiling own kernel
<reev>
yes stuck with centos
<reev>
yes i can but i want to try first
<KotCzarny>
tool is named swconfig
<KotCzarny>
its used to configure switch ports
<KotCzarny>
though you might have different problem
<reev>
i tries swconfig is not present too
<reev>
*tried
<KotCzarny>
grab it from somewhere (armbian is fine)
<KotCzarny>
tkaiser, used with battery backup is working nice as a router, though i agree its FUBAR
<tkaiser>
KotCzarny: Unless you add an USB-Ethernet-dongle this can not be considered a 'router' as people want to use it: to separate WAN/LAN. It's *not* possible with this design.
<KotCzarny>
tkaiser, you are going overboard here, for home purposes it does what it says
<KotCzarny>
ie. separation works when configured
<tkaiser>
BS. When you brick the device then your whole network is open since 'separation' does only work when Linux booted and swconfig did it's job. But hey, it's moronic to discuss with people who think a VLAN capable switch would make a router. :(
<tkaiser>
The whole device is crap (so many design flaws) and a router would need an additional Ethernet interface.
<KotCzarny>
that's what i've said
<tkaiser>
KotCzarny: Why do people want to use this device _at home_? Since they believe it could separate WAN from their WLAN. Is this possible with this design? No! Why do say then 'for home use ok'? That's plain BS.
<tkaiser>
s/WLAN/LAN/
<tkaiser>
WLAN is close to unusable with this board
<KotCzarny>
you are saying things from corporate point of view where separation means physical separation
<tkaiser>
Ok, time to stop. It gets moronic again
<KotCzarny>
at home its enough that packets go where they have to
<reev>
tkaiser we running bandwidth management for our network on this bullshit board
<KotCzarny>
tkaiser, i would love if you had the same stance for systemd, which is more threat than this thingie
apritzel has joined #linux-sunxi
IgorPec has joined #linux-sunxi
apritzel has quit [Ping timeout: 244 seconds]
yann has quit [Ping timeout: 276 seconds]
MiLeon has quit [Quit: Leaving]
reinforce has quit [Quit: Leaving.]
yann has joined #linux-sunxi
reinforce has joined #linux-sunxi
ricardocrudo has joined #linux-sunxi
iamfrankenstein has joined #linux-sunxi
jstein_ has joined #linux-sunxi
MiLeon has joined #linux-sunxi
MiLeon has quit [Remote host closed the connection]
jstein_ is now known as jstein
matthias_bgg has joined #linux-sunxi
apritzel has joined #linux-sunxi
ikmaak has joined #linux-sunxi
perr has joined #linux-sunxi
perr has quit [Changing host]
perr has joined #linux-sunxi
iamfrankenstein1 has joined #linux-sunxi
iamfrankenstein has quit [Ping timeout: 276 seconds]
iamfrankenstein1 is now known as iamfrankenstein
popolon has joined #linux-sunxi
apritzel has quit [Ping timeout: 244 seconds]
apritzel has joined #linux-sunxi
<MoeIcenowy>
have anyone tried to use the bluetooth function of RTL8723BS on mainline kernels?
<apritzel>
MoeIcenowy: I was wondering about that as well, access should be pretty easy via UART3, right?
<apritzel>
MoeIcenowy: have you tried talking to it via screen or minicom?
<apritzel>
MoeIcenowy: it should respond to those ancient AT commands
<buZz>
did you guys know that 'pocketchip' uses the mainline kernel
<buZz>
AND wants to get the mali proprietary driver working :P
ikmaak has quit [Ping timeout: 250 seconds]
<wens>
buZz: yes
<wens>
the proprietary driver is a userspace library
<buZz>
was kinda amusing to me :P
<buZz>
wens: isnt it a combination? kernel side interface + userspace lib?
<wens>
buZz: afaik the kernel side is open source
<wens>
see nvidia for an example :p
<buZz>
i havent seen the proprietary mali ever work on a mainline sunxi
perr has quit [Quit: Leaving]
ikmaak has joined #linux-sunxi
matthias_bgg has quit [Quit: Leaving]
matthias_bgg has joined #linux-sunxi
matthias_bgg has quit [Remote host closed the connection]
ikmaak has quit [Ping timeout: 265 seconds]
ikmaak has joined #linux-sunxi
OverCR has joined #linux-sunxi
reinforce has quit [Ping timeout: 250 seconds]
ikmaak has quit [Ping timeout: 276 seconds]
ikmaak has joined #linux-sunxi
reinforce has joined #linux-sunxi
ikmaak has quit [Remote host closed the connection]
iamfrankenstein1 has joined #linux-sunxi
iamfrankenstein has quit [Ping timeout: 276 seconds]
iamfrankenstein1 is now known as iamfrankenstein
ikmaak has joined #linux-sunxi
<ssvb>
buZz: it is not very difficult to get mali working with a *patched* mainline kernel, everything depends on your definition of what is "mainline"
<buZz>
ah k
<buZz>
well patched != mainline, imho
<buZz>
but k, as soon as i get that device i'll give it a go
<buZz>
its kinda gaming oriented hw, should have opengl
<buZz>
-es
ikmaak has quit [Remote host closed the connection]
Mr__Anderson has joined #linux-sunxi
ikmaak has joined #linux-sunxi
fdcx has quit [Remote host closed the connection]
cnxsoft has quit [Read error: Connection reset by peer]
fdcx has joined #linux-sunxi
paulk-collins has joined #linux-sunxi
fdcx has quit [Remote host closed the connection]
fdcx has joined #linux-sunxi
JohnDoe_71Rus has quit [Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org]
dearfibonacci has quit [Ping timeout: 276 seconds]
dearfibonacci has joined #linux-sunxi
iamfrankenstein1 has joined #linux-sunxi
<MoeIcenowy>
buZz: but now the CHIP guys have not get the proper blob
<MoeIcenowy>
maybe we should ask Allwinner for one
iamfrankenstein has quit [Ping timeout: 276 seconds]
iamfrankenstein1 is now known as iamfrankenstein
<buZz>
oh you mean the gpu inside R8 isnt the same as in A13?
<buZz>
i thought it was 100% A13 compatible
<MoeIcenowy>
buZz: it's of course compatible.
<MoeIcenowy>
but we have currently no Utgard r6p0 blobs
reev has quit [Ping timeout: 260 seconds]
<buZz>
is r6p0 the only option? cant i just take an older one?
<MoeIcenowy>
buZz: the EULA is changed in r6p0
<buZz>
i dont see how that matters to the chip
<buZz>
does the gpu need to read the eula?
reev has joined #linux-sunxi
<wens>
buZz: the new EULA is more friendly
<KotCzarny>
no, but for driver to be mainlined/distributed it matters greatly
<wens>
the most important part is you can now redistribute it (i think)
curtis22 has joined #linux-sunxi
<buZz>
alright
<buZz>
i was mostly motivated to 'get it working'
<buZz>
not looking for a lawsuit, redistributory rights, or starting any business :P
<MoeIcenowy>
wens:
<MoeIcenowy>
If I specify aliases serial0 = &r_uart and serial1 = &uart1
<MoeIcenowy>
will r_uart surely become /dev/ttyS0 and uart1 surely become /dev/ttyS1?
<wens>
yes
<MoeIcenowy>
(my board's rtl8723bs bt is at uart1
<MoeIcenowy>
(and the uart1 pinctrl things of A23/33 is in a mess
<MoeIcenowy>
(I doubt that no one have used it
<MoeIcenowy>
I will soon send two patches on it
reinforce has quit [Read error: Connection reset by peer]
mosajjal has joined #linux-sunxi
<rellla>
MoeIcenowy: the available r6p0 utgard userspace driver for mali 450 is not what we want i suppose?
cnxsoft has joined #linux-sunxi
reinforce has joined #linux-sunxi
<MoeIcenowy>
rella: where
dearfibonacci has quit [Ping timeout: 250 seconds]
<MoeIcenowy>
tkaiser: It says "The R40 will be a chip which focuses on Open-Source Community"
<MoeIcenowy>
"All the SDK of R40 will be open-sourced
<FergusL>
tkaiser: yes I've seen you linked to this
<tkaiser>
And then: Xunlong IMO does a good job developing hardware but they obviously fail when it's about everything else. Which does not matter that much to me
<MoeIcenowy>
maybe the R40 will be an A20 with more peripherals and two more cores
<MoeIcenowy>
or maybe an A33 with extra peripherals
mosajjal has joined #linux-sunxi
<jelle>
well it's nice to have a cheap SATA enabled board
<huawei>
MoeIcenowy, but AW says it every times...
<tkaiser>
huawei: But now Tina is there ;) At least Tina is mentioned on R40 product page as well as in Banana Pi gitbook (empty page)
<MoeIcenowy>
huawei: ;-) ;-) ;-) ;-) ;-)
<MoeIcenowy>
<(=╯▽╰=)>
<FergusL>
tkaiser: can you tell from the Air FEX if it will have audiocodec out on the headers? around line 590
<tkaiser>
MoeIcenowy: Based on feature comparison of A33, H3 and R40 I would assume it's a H3 (DDR3L also possible, GMAC) with IP blocks from A20
<tkaiser>
FergusL: Audio should work for sure, they made a kernel patch and fex changes since PA16 that will be used for uart3 (BT part of AP6212 on the Air) conflicts with audio out.
<FergusL>
seems it's commented out, ye
<FergusL>
s
<huawei>
tkaiser, the video say R40 will focus on opensource, all SDK will release fully opensource
<MoeIcenowy>
huawei: they didn't say "fully"
<tkaiser>
FergusL: But take this with a grain of salt. I really don't care about audio and the only time I heard sound coming out of an SBC it scared me almost to death (testing cedrus video decoding without knowing my display has speakers built-in)
<FergusL>
hahaha
<MoeIcenowy>
he said “所有的 SDK 都将开源” not “所有的 SDK 都将完全开源”
<MoeIcenowy>
we can only pray it not to be blobed open-source
<FergusL>
tkaiser: the only thing I do care about is audio on SBC! hehe, not that I expect it to be great though...
<KotCzarny>
FergusL: if you connect i2s audio card you can achieve good audio
<KotCzarny>
but good i2s audio card will cost you much more than sbc ;)
<FergusL>
KotCzarny: that's the end goal and yes, the downside is that it adds cost, specifically that I'd probably make the I2S pcb myself
<KotCzarny>
on the plus side, it will be reusable if you ever require more cpu power and switch sbc
<huawei>
MoeIcenowy, ok. i still wanna a newer kernel :-P
<FergusL>
that's true
<FergusL>
there are some cheap I2S dacs on ebay, I might start with that, but I'd like to have analog in as well
<KotCzarny>
for now my main concern with onboard analog audio is lack of good shielding
<KotCzarny>
usb acrtivity is very audible
<KotCzarny>
*activity
<FergusL>
oh, really? I'll try that on OPi2
<KotCzarny>
its usually noticable on boards which use wifi for connectivity
<FergusL>
now that I have my software working with MIDI and audio, it also uses network
<huawei>
MoeIcenowy, the video also mention that R40 has most powerful IO port in the R series
<huawei>
wishee to have SATA port
imcsk8_ is now known as imcsk8
Nyuutwo has quit [Ping timeout: 240 seconds]
<FergusL>
I'm writing audio software that uses ALSA to play sound, I want to see if I can get really low latency
<FergusL>
<1ms
<huawei>
FergusL, jack is more suitable
<FergusL>
huawei: jack doesn't provide anything more that what you get with standard ALSA, expect for the huge software ecosystem which I'm not planning on using, the audio processing will be my own
<huawei>
FergusL, ok, i have mixed the level of jack and alsa
<tkaiser>
hauwei: SATA is confirmed for R40. But I would like to get T3 instead (like R40 but with AXP221s support)
<KotCzarny>
they should make chip with 8 sata lines and cash on cheapest nas boards ;)
<huawei>
tkaiser, any T3 board released?
<FergusL>
are those Allwinner product codenames? like R8 was for A... A10 ?
<huawei>
FergusL, R8 seems like A...A13?
<tkaiser>
huawei: T3 is for automotive, I guess there will never be a dev board with this now that Allwinner starts with marketing efforts around IoT/Tina/Openness for R40. But maybe R40 can also be combined with AXP221s -- who knows
<tkaiser>
R8 _is_ A13
<GeneralStupid>
audi R8 is a lot faster
<KotCzarny>
but is it opensource?
<GeneralStupid>
i really love all the work you guys do for allwinner stuff -.- but i think i will buy a raspberry pi or something with more support the next time
<KotCzarny>
or stick to h3
<KotCzarny>
which is cheaper and quite well supported
<huawei>
GeneralStupid, but rpi only has USB 2 and 100M ethernet....
<GeneralStupid>
but every usb port is working, not like in orangepi pc
* KotCzarny
looks at his opipc. all ports work
<huawei>
GeneralStupid, you mean usb works in mainline kernel?
<GeneralStupid>
huawei: no 3.4.
<KotCzarny>
GeneralStupid: which ports dont work for you?
<tkaiser>
GeneralStupid: You use wrong settings, it's just that. Are you the one constantly fighting with Android on Oranges?
<GeneralStupid>
no
<GeneralStupid>
i tried that for a very short time and since then i use devuan
<tkaiser>
GeneralStupid: ok, since then it's easy. Wrong script.bin or wrong kernel config. USB works perfectly when you use good OS images or know how to tweak settings ;)
<GeneralStupid>
tkaiser: my kernel incl. modules is from armbian
<KotCzarny>
GeneralStupid: but your kernel is misconfigured
<KotCzarny>
and uboot
<tkaiser>
GeneralStupid: As explained yesterday: Armbian is not a kernel, it's a build system that takes care, that u-boot, kernel and hardware stuff (settings, fex/DT) match
<GeneralStupid>
i use kernel from armbian
<tkaiser>
You use most probably script.bin for OPi 2 or Plus.
<KotCzarny>
then use also armbian's fex
<KotCzarny>
for opipc
<GeneralStupid>
that could be a good idea.
<huawei>
tkaiser, i saw built-in battery support in bananapi's gitbook. maybe it will combine with AXP chip
Mr__Anderson has quit [Remote host closed the connection]
<MoeIcenowy>
huawei: M2 Ultra?
<tkaiser>
huawei: If you look closely in their 'documentation' then this new board has sometimes 2 USB host ports, sometimes 3, DRAM is clocked with 733 MHz (that's copy&paste from BPi M3 which has LPDDR3 RAM) and the new board has just one led (that's copy&paste from BPI M2+). They really don't give a shit about what they provide as documentation or even information. And even if you tell them multiple times that something's wrong
<tkaiser>
they do not correct it. You can't trust in anything they publish since it always starts as copy&paste from somewhere else
<MoeIcenowy>
tkaiser: I even have acquired an AXP813 datasheet, with a title of "AXP288"
<apritzel>
tkaiser: I totally agree on that, you can only believe things you've verified on h/w
<apritzel>
especially when it comes to new SoCs ...
dearfibonacci has quit [Quit: Leaving]
cnxsoft has quit [Quit: cnxsoft]
<tkaiser>
apritzel: Well, based on 'Team BPi' information I thought BPi-M64 would be equipped with LPDDR3 (and expected booting troubles) but had to learn that it's just the same type of DRAM when you created the wiki page after looking at the board in front of you ;)
<tkaiser>
*same as Pine64+
<apritzel>
indeed ;-)
IgorPec has quit [Ping timeout: 276 seconds]
reinforce has quit [Quit: Leaving.]
Mr__Anderson has joined #linux-sunxi
apritzel has quit [Read error: Connection reset by peer]
Leepty has quit [Remote host closed the connection]
<KotCzarny>
more reliable and cheap than finding usb hub with power switching support
yann has quit [Ping timeout: 250 seconds]
<NiteHawk>
topi`: according to the schematics, opi pc can drive the VBUS on the micro usb port (PL2 = USB0-DRVVBUS), but not the standard ones (they seem to have a fixed, "always-on" voltage supply). there's also PL3 = USB1-DRVVBUS, but it's unclear what that's (doesn't seem connected)
<KotCzarny>
topi: you might want relay with bt if you plan on controlling more distant devices, or relays that are networked via power grid, to keep the cables to minimum
<tkaiser>
NiteHawk: Ethernet dongle connected to gets power and switching it to 1 and back to 0 does nothing
<MoeIcenowy>
mripard: do you know how to drive a rtl8723bs's bluetooth part with mainline kernel?
<MoeIcenowy>
(the hadess' out-of-tree driver is built
<topi`>
I need to set up one of my OPi PC's to test out with this usb0-drvvbus signal... can it boot from the OPi Lite image, which I happen to have in one SD card already?
<KotCzarny>
ssvb, are those fexes original allwinner's ones or fixed?
<ssvb>
KotCzarny: it's a mix of both
<ssvb>
normally people push the original fex (which comes from the board manufacturer) and then maybe apply some fixes in the follow up commits
<buZz>
ssvb: TIL :)
<KotCzarny>
im asking, because vendors usually dont have a clue
<KotCzarny>
and while they might be a starting point, often they are suboptimal or plainly broken
<KotCzarny>
so hard to recommend ;)
<ssvb>
well, development boards is a special category of devices, they usually really have bad fex files
<ssvb>
but consumer oriented devices (such as tablets and htpc boxes) are expected to have reasonably usable fex files
<KotCzarny>
[Aug18 00:00] Division by zero in kernel.
<KotCzarny>
wow. nice.
<topi`>
dev boards have always been a niche segment, but the advent of cheap mobile SoCs kind of made them dirt cheap
<topi`>
in the old days, a dev box for a certain PowerPC chip would cost like $2000
<topi`>
because of the small factory runs
<topi`>
I'm surprised ppl aren't very interested in controlling the VBUS of individual usb sockets... that's very easy and no-soldering-required way of controlling different appliances remotely
<KotCzarny>
relays are easier and more flexible
<topi`>
I can figure out 100 different ways to take advantage of that
<KotCzarny>
and capable
<topi`>
to get relays, I need to place an order to dealextreme.com or elsewhere and then figure out how they work
foxx_ has quit [Ping timeout: 240 seconds]
<topi`>
that wouldn't be easy for a person completely unfamiliar with electronics
|Jeroen| has quit [Quit: dada]
<KotCzarny>
you set gpio, bang relay switched
<KotCzarny>
cant get much easier
<topi`>
I agree, it's easy, but it's not very beginner friendly
<KotCzarny>
dont be afraid to learn, there is whole world
<buZz>
22:14:37 < lkcl> buZz: it's a long story. they moved to non-free infrastructure without inviting me to comment on the consequences of doing so.
<buZz>
22:15:03 < lkcl> buZz: the sunxi community formed out of the efforts that i inspired, here on arm-netbook
<buZz>
waw
<topi`>
so he has burned some bridges behind him?
<KotCzarny>
what non-free infra?
<topi`>
does he mean github.com?
<buZz>
22:15:46 < lkcl> buZz: after they set up a git repository on github's non-free proprietary infrastructure (which sells your data to the highest bidder)
<buZz>
22:16:13 < lkcl> buZz: and after they set up the mailing list on google groups (which also sells your data and uses it for advertising purposes)
<buZz>
22:17:00 < lkcl> buZz: i offered to join their community when the source code for both google groups and github are published as software libre projects. preferably under the AGPLv3+
<buZz>
yep
<jelle>
that's constructive :-)
<buZz>
well, now we know ;)
<buZz>
he just went full stallman
<aalm>
lol
<topi`>
well, did he provide any stallman-friendly alternatives to github?
<KotCzarny>
isnt gpl code free, no matter where its hosted?
<topi`>
if github is good enough for Linus, it ought to be good enough for Stallman
IgorPec has quit [Ping timeout: 265 seconds]
<jelle>
huh, nah that's not how stallman works
<topi`>
maybe I should set up a kickstarter for a completely free (stallman-ish) github.com clone :)
<topi`>
jelle: as gitlab is from Atlassian, I think they also have some non-free clauses
<jelle>
bitbucket is from atlassian gitlab is not
MiLeon has quit [Quit: Leaving]
<topi`>
hmm? we have a corporate license from Atlassian and we got gitlab
<topi`>
ok, sorry, Gitlab is from Gitlab Inc
<topi`>
ok so they *also* offer a self-hosted, free app as well as their SaaS service
<jelly>
atlassian's thing used to be called stash
<topi`>
well, we are kind of off-topic. I need to concentrate on my project trying to get a consumer 4-socket 230V thingy I bought from amazon, remotely controlled
<topi`>
it has a USB cable which drives the sockets ON if +5V detected at VBUS, otherwise OFF
<KotCzarny>
make your own cable and drive it from gpio pins directly instead of usb port?
<topi`>
one way to drive VBUS on/off would be to suspend the OPi for a while
<topi`>
do we have any +5V compatible pins?
<topi`>
I thought everything was 3.3V like on RPi
<KotCzarny>
as a bonus you will have your usb ports free
<KotCzarny>
topi, test it, maybe 3.3v will be enough to switch?
<topi`>
oh, there's +5V going to HDMI as well
<topi`>
maybe that could be hacked ;)
<KotCzarny>
o.O
<KotCzarny>
keep current in mind
<topi`>
probably HDMI can be switched off at will?
<KotCzarny>
or you can fry things
<topi`>
it's a major consumer of power in any case
<KotCzarny>
nope, power on hdmi is only for little chips that answer on connect
<topi`>
is it possible to put any of these XUnlong boards to suspend?
<KotCzarny>
and its 50-200mA max
Mr__Anderson has joined #linux-sunxi
Mr__Anderson has quit [Remote host closed the connection]
OneSploit has quit [Ping timeout: 258 seconds]
Mr__Anderson has joined #linux-sunxi
<tkaiser>
topi`: With legacy kernel you can put any Xunlong board to suspend. Won't change anything regarding 5V available on USB
<tkaiser>
topi`: If you want to go this route, choose a Lime2 or buy a simple relay to connect via jumper wires (no soldering and no USB needed)
<topi`>
tkaiser: what about the USB0-DRVVBUS signal in PL2 like NiteHawk suggested? Looking at schematics, and the overcurrent protector chip specs, it sure looks like it's able to switch off VBUS if EN is driven to 0
<tkaiser>
topi`: I was curious, tried it out and... nada, niente, nothing. Doesn't work.
<tkaiser>
USB consumers on every Orange Pi keep powered regardless what you do
<tkaiser>
(possible exception: the first Orange Pi and OPi Mini -- both with A20 and AXP209)
OneSploit has joined #linux-sunxi
OneSploit has quit [Changing host]
OneSploit has joined #linux-sunxi
Guest99570 is now known as fredy
fredy is now known as Guest27100
Guest27100 is now known as fredy
reinforce has quit [Quit: Leaving.]
Mr__Anderson has joined #linux-sunxi
fredy is now known as fredy`2
fredy`2 is now known as fredy
<topi`>
I have some BananaPI's (the original), would that have the AXP209?
<KotCzarny>
yes
<topi`>
let's see the schematics...
<topi`>
hmm, Lemaker or Sinovoip? I'm not entirely sure...
<KotCzarny>
most a20 were bundled with axp209
<tkaiser>
topi`: at least with Banana Pro it doesn't work even if it's described differently in fex file. Not tried with the real Banana Pi
<KotCzarny>
but just read about relays, its really nice, and you can even drop your thingie altogether, and drive 230V directly with gpio
<topi`>
<default> is something that gets inherited from U-Boot?
<tkaiser>
PcDuino 3 Nano uses the same pin mapping and there this is pure BS because the power pins from USB OTG are directly connected with DC-IN. You can't trust the fex files here. Simply try it out ;)
<topi`>
right...
<tkaiser>
As already said: Not tested with original Banana Pi. At least with Banana Pro and PcDuino Nano (and all Orange Pi that are H3 based) you're out of luck.
<topi`>
odd thing, if I look at the orangePI schematics, or the bananaPI schematics, they look mostly the same.
<topi`>
even the layout&orientation of the signals...
<KotCzarny>
the original banana pi was cloned to hell and back
<KotCzarny>
and xunlong dev was designing the board for banana pi
<topi`>
but was the original Banana Pi from Lemaker or Sinovoip?
<topi`>
this history of BPi is such a confusing topic...
OneSploit has quit [Ping timeout: 265 seconds]
<KotCzarny>
citing facebook: it's complicated
<tkaiser>
topi`: Most probably neither nor ;)
<tkaiser>
It started at Foxconn, they let Steven/Xunlong do the ODM work, then SinoVoip did the manufacturing and LeMaker should build a community around.
<topi`>
so, basically Xunlong is where all the skills are?
<KotCzarny>
you can say that looking at later bananas ;)
<topi`>
:)
<topi`>
Xunlong is definitely leading in bang-for-the-buck department
<tkaiser>
The LeMaker guys got greedy and applied for trademarks and are now out of business (talking themselve into Linaro last year and claiming they would be the community people which _was_ their only asset)
<topi`>
well, greed is one of the 7 sins :)
<tkaiser>
According to wens Steven still does the ODM work and the famous 'Team BPi' is only responsible for worst software and support ever. And for documentation/information you can never trust.
<topi`>
I was somewhat tempted by Lemaker
<topi`>
Lemaker's A1100 board, but I decided not to
OneSploit has joined #linux-sunxi
OneSploit has joined #linux-sunxi
OneSploit has quit [Changing host]
<topi`>
is there anything 64-bit in Steven's pipeline?
<KotCzarny>
opi3
<KotCzarny>
;)
<topi`>
probably based on one of the 64bit wares from Allwinner
<tkaiser>
topi`: LeMaker's A1100 board is most probably made by circuitco (Huskyboard with few modifications)
<topi`>
H64 got cancelled, I think, so it's either A64 or the newly released H5
<tkaiser>
H5, already confirmed
<aalm>
any allwinner SoC w/more than single CAN?
<KotCzarny>
but you should stick to supported socs, and buy 64bit board in a year or two
<topi`>
tkaiser: I decided to preorder the Marvell 8040 board from Solidrun
<topi`>
decent price, probably very well performing board
<topi`>
KotCzarny: H5 probably borrows quite a lot from H3 and friends...
<topi`>
only the cpu cores will be decidedly different
<KotCzarny>
remember h3 is not supported fully in mainline
<topi`>
I'm OK with missing Mali and HDMI and stuff
<topi`>
as long as it can drive the etherne t:)
<tkaiser>
aalm: Is there any Allwinner SoC other than A20 that supports CAN at all?
<topi`>
there was an A64 based board coming from Olinuxino guys, what's up with that one?
<topi`>
yeah probably the A20 is the last SoC that supports Everything :)
<aalm>
tkaiser, idk.
<topi`>
AW noticed that, hey, all those bits are not needed in tablets
<tkaiser>
aalm: Famous 'Team BPi' wrote Banana Pi M2+ would support CAN in the beginning. But that was just 'copy& paste gone wrong' from Banana Pi M1
<topi`>
I was genuinely surprised to find that the R40 intends to bring SATA back
<KotCzarny>
lets wait and see. i hope its not usb-sata bridge IN soc ;)
<topi`>
if I look at the R40 specs, SATA is there, but no CAN
<topi`>
KotCzarny: it would be easiest for them to just reuse the SATA from A20
<topi`>
so you get the same 40MB/s cap on writes ;)
<topi`>
it probably won't be very detrimental, since Linux buffers writes extensively to RAM in any case
Mr__Anderson has quit [Remote host closed the connection]
<topi`>
I wonder how much the H3 SoC costs if the OPi One costs $11 in single pieces?
<topi`>
probably the H5 won't break the bank either
pg12_ has quit [Ping timeout: 264 seconds]
<tkaiser>
topi`: NanoPi NEO with 256MiB is sold for $8 (with some shipping costs added of course)
pg12 has joined #linux-sunxi
<topi`>
at those prices, individual components must surely be <$1
<KotCzarny>
dont forget about shipping charges ;)
<topi`>
some bigger SoCs, like i.MX6 dual, cost like $13 apiece, so a big difference in price
<topi`>
the SATA and Gbit ether are not perfect in i.MX6, either
<topi`>
I'm really looking forward what the Marvell 8040 can do...
OneSploit has quit [Quit: Leaving]
<topi`>
at least they know their shit when it comes to I/O...
<KotCzarny>
yeah. but price also shows their r&d costs
<topi`>
the chinese have some super skill on getting by with very small r&d costs :)
<tkaiser>
topi`: Marvell Armada 38x is already awesome. Only dual-core A9 but able to saturate a few GbE connections and SATA 3.0 lines in parallel. It's about SoC design... and prices ;)