pmpp has quit [Quit: No Ping reply in 180 seconds.]
pmpp has joined #linux-sunxi
<planix>
Hi I am working on a project that involves a olimex allwinner a20 som board and KSZ8895 5 port ethernet switch. ethernet switch is configured via an eeprom connected to KSZ8895. Switch starts ok. 5th port of the switch is connnected via MII to SOM Board (Same connections are used with Olimex Micro) I have following dts configuration for gmac peripherial.
<planix>
After booting up i can see board starts receiving data (ifconfig shows rx bytes going up) but if I ping a specific internal ip for following stack trace is send to console. Do you guys have any idea.
<planix>
56 data bytes [ 32.084334] ------------[ cut here ]------------ [ 32.088990] WARNING: CPU: 0 PID: 0 at net/sched/sch_generic.c:323 dev_watchdog+0x26c/0x270 [ 32.097266] NETDEV WATCHDOG: eth0 (sun7i-dwmac): transmit queue 0 timed out [ 32.104217] Modules linked in: [ 32.107288] CPU: 0 PID: 0 Comm: swapper/0 Not tainted 4.15.0 #2 [ 32.113197] Hardware name: Allwinner sun7i (A20) Family [ 32.118442] [<c010e6e8>] (un
reinforce has quit [Ping timeout: 240 seconds]
reinforce has joined #linux-sunxi
imcsk8_ has joined #linux-sunxi
imcsk8 has quit [Ping timeout: 240 seconds]
<wens>
embedded ecosystem talk at ELC doesn't mention Mali-4xx
<icenowy[m]>
maybe they're finally abandoned
anarsoul|2 has joined #linux-sunxi
f0xx has quit [Ping timeout: 240 seconds]
planix_ has joined #linux-sunxi
raspo666 has joined #linux-sunxi
hardfalcon has joined #linux-sunxi
planix_ has quit [Ping timeout: 260 seconds]
f0xx has joined #linux-sunxi
vagrantc has quit [Quit: leaving]
hipboi has joined #linux-sunxi
hardfalcon has quit [Quit: Leaving.]
hardfalcon has joined #linux-sunxi
anarsoul|2 has quit [Ping timeout: 265 seconds]
reinforce has quit [Ping timeout: 246 seconds]
reinforce has joined #linux-sunxi
hlauer has joined #linux-sunxi
fkluknav_ has joined #linux-sunxi
IgorPec has quit [Ping timeout: 256 seconds]
planix has quit [Ping timeout: 260 seconds]
Mr__Anderson has quit [Remote host closed the connection]
mkid has joined #linux-sunxi
f0xx has quit [Ping timeout: 268 seconds]
sr-digitronic has joined #linux-sunxi
sr-digitronic has quit [Client Quit]
f0xx has joined #linux-sunxi
<sunxi_fan>
mirko: sorry for the late reply about weston on DRM/KMS and libMali.so r6p2.. yes i have a working setup, feel free to ask for tech details, if you still need help or hints..
<sunxi_fan>
let me tell you the Qt5.9(buildroot)/wayland/weston/libMali.so/DRM/KMS stack is pretty cpu waste..
<sunxi_fan>
IMHO, so i headed over the Qt/EGLFS/libMali.so/DRM/KMS and that's more light .. had both the "hellowindows --multiscreen" and this QML example (https://github.com/alpqr/quickmwtest) fairly working on my multscreen setup (LDC RGB and HDMI on Olimex A20-SOM) but ...
<wens>
icenowy[m]: the lima repos are still seeing work
Wizzup has joined #linux-sunxi
dddddd has joined #linux-sunxi
<pmpp>
sunxi_fan: hi, is QOffscreenSurface available and working ?
andreagus281 has joined #linux-sunxi
BenG83_ has quit [Ping timeout: 264 seconds]
msimpson has joined #linux-sunxi
clemens3 has joined #linux-sunxi
<sunxi_fan>
pmpp: ummhh.. "good question!" :-) i don't know, really.. need to check what you mean. me not really a "Qt guru" but just trying to put up a POC to find the clue; so here the pretty raw and simplest code just showing the issue in my HW setup
<pmpp>
i ask that because mirko's trouble is with some calls to window kind widgets in an app, which is not allowed with eglfs
_whitelogger has quit [K-Lined]
_whitelogger has joined #linux-sunxi
<sunxi_fan>
i see.. indeed my focus is on embedded GUI with two "full screen" displays (maybe owned by a single process, the renderer...) so indeed i can stick with EGLFS constraint..
<sunxi_fan>
for different use cases, of course, some kind of compositor/window manager is due..
<pmpp>
that make two window or 1 ?
<pmpp>
i was wondering(for me) if eglfs + vfb multiscreen is possible
<sunxi_fan>
here indeed we need to "closely" choose the wording!! :-) because AFAICS, the default of the "EGLFS" setting is a "virtual desktop", i.e. "something" that extends smoothly between both screens and where the coordinates are added up..
<sunxi_fan>
for example in my setup LCD=800x480 and HDMI is =1920x108 so the "rectangles" should be:
<pmpp>
xinerama style ?
<libv>
wens: this talk leaves out my work completely
<libv>
as if those free arm gpu projects happened out of thin air
<libv>
i was written out of history on modesetting, freeing ati, and now this
<libv>
quite amazing really.
<libv>
on another note, both the nintendo NES and SNES classic are gpl violating devices
<libv>
they come with a binary libnand for uboot, for starters
raspo666 has quit [Ping timeout: 256 seconds]
<libv>
but if i go out and make noise about it, i will be bashed for doing so, and others will see their business boosted
<sunxi_fan>
...but the simple code snippet, i've put on github, is showing the rectangles both on the second screen (the hdmi one) with an offset, and the LCD stays black.
<Net147>
pmpp: like widget->windowHandle()->setScreen(QGuiApplication::screens()[1]) ?
<pmpp>
dunno i was reading the multiscreen doc and stumble on that
msimpson__ has joined #linux-sunxi
<Net147>
I know it's pretty simple to have different QML on each screen, but I haven't tried widgets
msimpson_ has quit [Ping timeout: 260 seconds]
yann has joined #linux-sunxi
mavkhimenia has joined #linux-sunxi
<oliv3r>
hey, anybody familiar with valgrind crashes on the a20 due to neon instructions? A coleague just reported that they where rebuilding valgrind because the valgrind crashing on the neon stuff. I don't have any more details then that sadly
mavkhimenia has quit [Read error: Connection reset by peer]
mavkhimenia has joined #linux-sunxi
<icenowy[m]>
wens: I mean the Mali-4xx abandoned by some device designers ;-)
<icenowy[m]>
their spec is too low in 2018
<Net147>
oliv3r: is it unhandled instruction? try updating to newer version of valgrind?
<oliv3r>
yeah they are checking it out now, i think they are now just doing a rebuild of the existing debian package to remove neon support alltogether
<libv>
the samsung galaxy s2 was released at MWC in february 2011
<libv>
around the time when i had gotten my hands on a telechips tablet with a mali-200 and i started poking at the android 2.1 on it, where bionic lacked even LD_PRELOAD
<megi>
if anyone's using arch linux arm, don't pacman -Su atm :)
<libv>
7 years.
fkluknav_ has joined #linux-sunxi
<willmore>
megi, oh, now, what happened?
<megi>
results in broken pacman due to missing library
<oliv3r>
mripard: btw congrats on the successfull kickstarter! 31k was a good result I reccon
<megi>
and some unfun time
<willmore>
megi, ouch.
<oliv3r>
libv: shows how old we are getting! :p
<icenowy[m]>
and I think it may be possible for ARM Ltd to discontinue to provide Utgard license if lima's ready
<libv>
icenowy[m]: it's a graphics driver
<libv>
there is no such thing as ready
<icenowy[m]>
P.S. according to Allwinner people, different variants of the mali blob is different soft IP
<libv>
which is why i spent years talking modular drivers
<icenowy[m]>
and need to be purchased independently
<libv>
but that's something the xorg community has never been willing to see (apart from some intel guys who are actively making it easier for people to install the latest and greatest)
<Net147>
oliv3r: no buildroot or OE?
<oliv3r>
not for the standard debian packages; but i just got some more info, appearantly the jessie version of valgrind does not support neon, and their app uses neon; thus they need a rebuilt valgrind WITH neon
<oliv3r>
easy enough
<icenowy[m]>
it seems that Utgard cannot support glamor well...
dddddd has quit [Ping timeout: 256 seconds]
dddddd has joined #linux-sunxi
<Net147>
oliv3r: what is typical rootfs size? I find Debian somewhat large even with minimal installation.
<oliv3r>
for us? extracted we are at 350-ish mb
<oliv3r>
but mind you, our guys insist on python
<oliv3r>
so, debian is perfect :p
<oliv3r>
it's more of a micro-server rather then an emebedded system ):
<oliv3r>
oh andt he UI is py-qt
<oliv3r>
with fonts and resources, we are moving towards 500-ish mb
<Net147>
usually with OE I have image less than 160 MB
<Net147>
that's including fonts and resources
IgorPec has joined #linux-sunxi
jkarlson has joined #linux-sunxi
jkarlson has joined #linux-sunxi
jkarlson has quit [Changing host]
mzki has quit [Ping timeout: 264 seconds]
mzki has joined #linux-sunxi
scream has joined #linux-sunxi
jkarlson is now known as Ke
traw has joined #linux-sunxi
fkluknav_ has quit [Ping timeout: 260 seconds]
mavkhimenia has quit [Ping timeout: 256 seconds]
<oliv3r>
yeah but the debian part is not even our biggest culprit
<oliv3r>
python, python modules
<oliv3r>
numpy
<oliv3r>
all those take up a few hondered mesg
<oliv3r>
then we have the no-tofu font, which by itself is 65mb stripped
<ssvb>
oliv3r: valgrind supports neon, at least it used to be working fine for me the last time I tried it
<Net147>
oliv3r: what is no-tofu is that noto font?
<ssvb>
it would be great if your collegue reports a bug to valgrind developers, unless they already know about it
<oliv3r>
then i don' know what they are doing :p
<oliv3r>
Net147: yes it is :)
juri__ has joined #linux-sunxi
<oliv3r>
we need asian and cyrilic glyphs
juri_ has quit [Ping timeout: 260 seconds]
elros_ has joined #linux-sunxi
BenG83_ has joined #linux-sunxi
<Net147>
oliv3r: one thing I don't like about Python is that a lot of runtime errors would have been caught at compile time if using C++
<oliv3r>
i hate python :p
<oliv3r>
even though i appreciate it and think its quite usefull for a number of things
<oliv3r>
hate being a very strong word obviously
qeed has joined #linux-sunxi
traw has quit [Quit: Leaving]
mavkhimenia has joined #linux-sunxi
kloczek has quit [Remote host closed the connection]
tom_nov has joined #linux-sunxi
elros_ has quit [Remote host closed the connection]
kloczek has joined #linux-sunxi
Nikunj has joined #linux-sunxi
<Nikunj>
Hi
hardfalcon1 has quit [Ping timeout: 240 seconds]
afaerber has quit [Quit: Leaving]
elros_ has joined #linux-sunxi
elros_ has quit [Remote host closed the connection]
fkluknav_ has joined #linux-sunxi
afaerber has joined #linux-sunxi
fkluknav_ has quit [Ping timeout: 240 seconds]
tkaiser has joined #linux-sunxi
hipboi has quit [Quit: Leaving]
raspo666 has joined #linux-sunxi
afaerber has quit [Remote host closed the connection]
Nikunj has quit [Ping timeout: 260 seconds]
juri__ has quit [Read error: Connection reset by peer]
mavkhimenia has quit [Read error: Connection reset by peer]
mavkhimenia has joined #linux-sunxi
return0e has quit [Read error: Connection reset by peer]
return0e has joined #linux-sunxi
juri_ has joined #linux-sunxi
fkluknav_ has joined #linux-sunxi
afaerber has joined #linux-sunxi
nuuuciano__ has joined #linux-sunxi
JohnDoe_71Rus has quit [Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org]
Turl has quit [Ping timeout: 246 seconds]
mavkhime_ has joined #linux-sunxi
Turl has joined #linux-sunxi
mavkhime_ has quit [Read error: Connection reset by peer]
mavkhime_ has joined #linux-sunxi
mavkhimenia has quit [Read error: Connection reset by peer]
leviathan has joined #linux-sunxi
mavkhime_ has quit [Remote host closed the connection]
mavkhimenia has joined #linux-sunxi
kaspter has quit [Read error: Connection reset by peer]
kaspter has joined #linux-sunxi
mavkhime_ has joined #linux-sunxi
mavkhimenia has quit [Ping timeout: 256 seconds]
mavkhime_ has quit [Remote host closed the connection]
mavkhimenia has joined #linux-sunxi
kaspter1 has joined #linux-sunxi
mavkhimenia has quit [Read error: Connection reset by peer]
hanni76 has joined #linux-sunxi
mavkhimenia has joined #linux-sunxi
kaspter has quit [Ping timeout: 268 seconds]
kaspter1 is now known as kaspter
mavkhimenia has quit [Remote host closed the connection]
mavkhimenia has joined #linux-sunxi
cnxsoft has quit [Remote host closed the connection]
mavkhime_ has joined #linux-sunxi
IgorPec has quit [Read error: Connection reset by peer]
mavkhimenia has quit [Ping timeout: 240 seconds]
IgorPec has joined #linux-sunxi
juri_ has quit [Read error: Connection reset by peer]
mavkhimenia has joined #linux-sunxi
mavkhimenia has quit [Remote host closed the connection]
juri_ has joined #linux-sunxi
mavkhimenia has joined #linux-sunxi
mavkhime_ has quit [Ping timeout: 256 seconds]
JohnDoe_71Rus has joined #linux-sunxi
juri_ has quit [Read error: Connection reset by peer]
juri_ has joined #linux-sunxi
mavkhimenia has quit [Remote host closed the connection]
mavkhimenia has joined #linux-sunxi
juri_ has quit [Read error: Connection reset by peer]
juri_ has joined #linux-sunxi
planix has joined #linux-sunxi
mavkhimenia has quit [Remote host closed the connection]
mavkhimenia has joined #linux-sunxi
juri_ has quit [Read error: Connection reset by peer]
elros_ has joined #linux-sunxi
juri_ has joined #linux-sunxi
BenG83_ has quit [Remote host closed the connection]
vinimac has joined #linux-sunxi
matthias_bgg has quit [Ping timeout: 264 seconds]
fkluknav_ has quit [Ping timeout: 256 seconds]
<vinimac>
hi. Does the old sunxi-3.4 have support to allwinner H2?
mavkhime_ has joined #linux-sunxi
<vinimac>
hi. Does the old sunxi-3.4 kernel have support to allwinner H2?
<elros_>
anybody was able to use phoenixcard in startup mode with a33 image? It always fail for me
raspo666 has quit [Ping timeout: 265 seconds]
wasutton3_mobile has quit [Remote host closed the connection]
BenG83_ has quit [Remote host closed the connection]
matthias_bgg has joined #linux-sunxi
<adj_>
vinimac, I don't know, but I would avoid if possible non-mainline kernels for allwinner
BenG83_ has joined #linux-sunxi
f0xx has joined #linux-sunxi
reinforce1 has joined #linux-sunxi
stonegray has joined #linux-sunxi
Putti has joined #linux-sunxi
<vinimac>
adj_, I have to run my system from NAND memory.
reinforce has quit [Quit: Leaving.]
Putti has quit [Remote host closed the connection]
Putti has joined #linux-sunxi
<adj_>
vinimac, I think that the 3.4 kernel is only for suni4, suni5 and suni7, but I'm not sure
planix has joined #linux-sunxi
<KotCzarny>
for sun8i too
<adj_>
ok
<KotCzarny>
and i think for sun9i too
<adj_>
i didn't even spelled sun4i, sun5i and sun7i correctly
<adj_>
:)
<KotCzarny>
but we dont talk about those smelly things here, unless necessary
<adj_>
I would not hold my data on NAND with a very old kernel, without a maintained NAND code
<adj_>
vinimac, can't you use sd/emmc?
hlauer has quit [Ping timeout: 248 seconds]
tkaiser has joined #linux-sunxi
<BenG83_>
^
planix has quit [Ping timeout: 260 seconds]
<lvrp16>
wens: i'm sure you can find someone going to SFO and hitch a ride to monterey together
<lvrp16>
time to get yuq and libv visas and buy plane and admission tickets to ELC
<KotCzarny>
and enjoy the fight? :>
<lvrp16>
people fight over the internet, when you put them in the same room, that rarely happens
<lvrp16>
but fun to see none-the-less ;)
<libv>
you've never seen me with the xorg/fdo powerplayers then ;p
<lvrp16>
maxime: paulk-gagarine>: congrats on the fundraising, awaiting the results :)
<adj_>
lvrp16, can you post a link to a fight between libv and yuq?
<adj_>
I'm truly interested, not trolling
<libv>
adj_: as far as i know, such a thing never happened
<libv>
adj_: i am not sure where KotCzarny got that from
<lvrp16>
didn't hawking just write a paper on the multiverse? definitely happened
<lvrp16>
:P
<lvrp16>
just maybe not in this one
lkcl has quit [Ping timeout: 240 seconds]
BenG83_ has quit [Quit: Leaving]
<adj_>
interuniverse links don't work, HTTP 404
<adj_>
or HTTP 404 NOT FOUND
<lvrp16>
libv: you should do a fundraiser for RE'ed utgard, I can probably help get your first $20K USD
<libv>
lvrp16: afaiu, isn't yuq pretty much well advanced at this point?
<lvrp16>
yeah but a guy using his spare time is not as quick as two people or more doing it
<libv>
but i thought you were paying him?
<lvrp16>
it's still a spare time project
mavkhimenia has joined #linux-sunxi
<libv>
does his employer know about this?
<libv>
because i kind of assumed that he'd no longer work for amd by now
<libv>
which is why i am surprised about your statement about spare time
<lvrp16>
it's more like patreon rather than paying him for work
<libv>
i think mshopf did a tiny bit of consulting on radeonhd stuff, on PLLs, which was what i had tested to death, and he asked me if i was willing to do so, but that was 2009 after the suse layoffs, and i was unable to take in any money besides the miserable dole
<lvrp16>
there's no bound constraints
f0xx has quit [Ping timeout: 240 seconds]
<libv>
he did have to have a chat with his boss about it
<libv>
to make sure that there were no issues there
<wens>
IP / copyright issues come to mind
<libv>
i am less worried about that, at which point is it a spare time project, and at which point does it eat into working time
mavkhime_ has quit [Ping timeout: 264 seconds]
<lvrp16>
REing shaders for 10 y/o hardware hardly qualifies as competing against AMD
<wens>
then it becomes a problem of "is it paid work of the employer" and who the copyright owner is
<libv>
if you're receiving money for it, the lines between spare time and work time become really blurry, and the developer and his manager have to be very clear about things
<libv>
if such a thing is found out about afterwards, it is easy to misconstrue even if time was strictly seperate
<libv>
lvrp16: please warn yuq, maybe he has cleared this and all is well, or maybe he is oblivious and could use the heads up
<lvrp16>
i think he is clear about it from my convos with him, i just don't have the material knowledge to answer your specifics
<libv>
ok
kaspter has quit [Read error: Connection reset by peer]
<lvrp16>
given bootlin's successs, i would expect that an utgard fundraising effort being quite relevant for many people
kaspter has joined #linux-sunxi
<lvrp16>
but it's a community project so fundraising for one group or person seems incorrect way of doing it
<KotCzarny>
assuming people didnt spent their quarterly budget allowances on it already
msimpson__ has quit [Read error: Connection reset by peer]
<lvrp16>
yeah, the big players only contributed half so the other half of the funds came from small donors which is great as it is not a trivial amount
<lvrp16>
i would imagine that upstream utgard would affect 100x more people than upstream allwinner vpu
<KotCzarny>
cctv/video (re)encoder/player is what those boxes are
<KotCzarny>
they quite suck for desktop uses
<wens>
it would be hard to estimate a delivery time if people were only spending their spare time on it though
<lvrp16>
i mean raspberry pi 3 sold 5m units last year
<KotCzarny>
rpi is a scam
<lvrp16>
the market size is hardly trivial anymore
<lvrp16>
i know rpi is bs but the market is not bs
<wens>
as opposed to bootlin as a company devoting resources to a kickstarter project
<wens>
cheap sbcs that don't suck is a market :)
<lvrp16>
yeah that's the painful bit
f0xx has joined #linux-sunxi
<lvrp16>
it's DRAM that is killing cheap sbcs
<lvrp16>
$9 for 1GB of DDR4...pains...
<wens>
has market availability gotten any better?
<KotCzarny>
pity you cant just reuse sodimms from laptops
<KotCzarny>
people have plenty of them in closets
<lvrp16>
sodimms are 8-bit, you can use it with zte's chips
<lvrp16>
but the open source bits are nil
<wens>
nand is probably going to see a small price jump soon
<wens>
samsung had a power outage at one of its fabs last week
<lvrp16>
samsung has moved to ddr4 so ddr3 prices are going through the roof
<lvrp16>
ddr3 1GB went up 30% last month
<wens>
3% of production went down the drain
<wens>
I wonder what market uses more dram, computers (PCs, laptops, servers) or mobile / embedded devices
<lvrp16>
edge computing devices is taking off so they're consuming the most
<lvrp16>
stupid smart speakers
<lvrp16>
useless pieces of garbage in my opinion
<wens>
"the latest trend"
<lvrp16>
also china is blowing up in terms of internal electronics consumption
<KotCzarny>
i still use 15yo thinkpad
Xalius_Ph has joined #linux-sunxi
<lvrp16>
electronics recyclers are also making a killing
<lvrp16>
with ddr and nand prices being so high
<lvrp16>
which is a good thing since people hopefully recycle more
<KotCzarny>
fun part is that thinkpad consumes ~8-15W or power
<KotCzarny>
that's with display
<buZz>
isnt that typical for a decent laptop?
<KotCzarny>
sure, but one would think that over 15 years there would be more innovation
<KotCzarny>
in that area
<lvrp16>
that's idle with power gating
<lvrp16>
full load it's going to be close to 30W
<lvrp16>
bubble busting KotCzarny
<KotCzarny>
lvrp16: depending on the load and setup
<KotCzarny>
x40 was a great piece of engineering
<lvrp16>
for the electronics they had at the time, thinkpads were genuine pieces of art
<lvrp16>
in engineering terms
<lvrp16>
just got a Raspberry Pi 3 Model B+
<tkaiser>
Chrchrchr
<lvrp16>
that things eats power like crazy
<lvrp16>
headless sysbench it consumes 1A.
<icenowy[m]>
thinkpads!
<icenowy[m]>
;-)
<lvrp16>
that's 3x a s905x
<tkaiser>
lvrp16: Idle consumption increase is only related to Gigabit Eternet and Wi-Fi/BT
<lvrp16>
50% more than Raspberry Pi Model B
mavkhimenia has quit [Remote host closed the connection]
<lvrp16>
yeah idle is OK, load is where the new PMIC sucks power like crazy
<megi>
lvrp16: wouldn't soc desolder itself at 5W?
<icenowy[m]>
lvrp16: how is the on-board USB "GbE"?
<icenowy[m]>
and 5G Wi-Fi?
<lvrp16>
they implemented dvfs, drops voltage to 1.2v from 1.35v
<icenowy[m]>
1.2v is also very high
<lvrp16>
320 Mb for receive, 240Mb for transmit, 250Mb for simultaneous
<KotCzarny>
500Mb ?
<KotCzarny>
ahm, nvm
<lvrp16>
no 250Mb combined
<tkaiser>
icenowy[m]: Close to unbelievable but only one LibreELEC dev tested wi-fi so far. No other numbers available
<lvrp16>
transmit causes the thing to slow for some reason
<lvrp16>
contention on USB bus
<tkaiser>
100 Mbits/s with 802.11n at 5GHz
<KotCzarny>
power drops on transmit?
<lvrp16>
they run 1.4GHz at 1.35v
<tkaiser>
lvrp16: You should get 315 Mbits/sec over Gigabit Ethernet
<lvrp16>
tkaiser: yeah for iperf receive only, i can get 320Mb
<lvrp16>
send only 240Mb
<lvrp16>
send and receive is 30/220, i forget split which way
<tkaiser>
But in any real world scenario with disk access in parallel it's 19 MB/s since USB contention issue
<lvrp16>
rpi pulled one straight from orange pi's book and added a IHS
<lvrp16>
orange-pi/banana-pi
<tkaiser>
IHS?
<lvrp16>
integrated heat spreader
<lvrp16>
they no longer have the hotspot issue like before
<lvrp16>
temperature sensor is on GPU, so it never throttles itself and CPU area goes to 100C.
<lvrp16>
for so many unit sales, u would think they'd get their hardware shit together
<lvrp16>
hire a competent hardware engineer
<megi>
ah, not just soc, the board spreads heat better too
JohnDoe8 has joined #linux-sunxi
<tkaiser>
megi: 'learning from cheap competitors like Orange Pi' ;)
<megi>
I think orange pi's are this way too, the connectors can get quite hot
<icenowy[m]>
lvrp16: P.S. how's the temperature of full quad-core @ 1.4GHz?
<megi>
ah
<icenowy[m]>
I wonder whether it will severely overheat
<KotCzarny>
is it real 1.4ghz?
<tkaiser>
Yep
<wens>
tkaiser: 100 Mbits/s? sounds like SDIO limits
<KotCzarny>
ho.
<wens>
oops, past bed time, catch you guys tomorrow
<tkaiser>
wens: I thought the bus they're on is good for 200 Mbits/sec.
<tkaiser>
But single antenna Wi-Fi --> crap
<icenowy[m]>
wens: I also expected SDIO limit ;-)
<wens>
tkaiser: it's 200 MHz at 4 bits SDR or DDR?
<wens>
I don't remember SDIO going that high
<icenowy[m]>
tkaiser: I think 200Mbit/s is a theortical limit
<tkaiser>
I read they exchanged the two SDIO thingies they have with RPi 3 and thought they talked about 200 as maximum.
<icenowy[m]>
and the real speed wlil be far less than it
f0xx has quit [Ping timeout: 264 seconds]
<icenowy[m]>
because of the design of the wlan card
<tkaiser>
But I'm really surprised that nobody so far tested Wi-Fi. Just one external guy in his house with 802.11n only (no 802.11ac numbers anywhere)
JohnDoe_71Rus has quit [Ping timeout: 240 seconds]
<lvrp16>
icenowy: it goes to 80C and throttles voltage and clock down
<buZz>
:) just ordered me a B+ to complete my collection
<buZz>
no sense to it really :P
<tkaiser>
lvrp16: 'echo > /sys/devices/platform/soc/3f980000.usb/buspower' and 'echo $busid > /sys/bus/usb/drivers/smsc95xx/unbind' (the older Pi 3, the new one has another node)
<lvrp16>
turning off the wifi and usb chip?
<tkaiser>
Yes, transforming it into a secure device without connections to the outside!
<tkaiser>
The new Ethernet chip there when used with mainline kernel, active EEE and a switch that is also EEE capable makes already for 400mW less
<tkaiser>
Downgrading to Fast Ethernet also greatly reduces consumption
<tkaiser>
lvrp16: Would be interesting how stable the new Pi is with all the overclocking. Maybe trying out xhpl which led to discovery of crappy under-voltage settings on RPI 3 B two years ago?