<willmore>
I'm torturing my OpiZ with cpuminer. :) it's doing very well with the stress test. I have a current monitor on it. It idles just north of 0.10A. Occasional peaks to 0.20A. With cpuminer starting--cold system--it drew almost 0.50A.
<willmore>
But, as the SoC got hotter and the clocks slowed things down, current fell with clocks.
<willmore>
At 912MHz, it's getting 1.5KH/s Down from the peak of 1.81KH/s when it was cold.
<willmore>
This is with no heatsinking, nor airflow. Wifi instead of wired.
<willmore>
67-66 degrees.
egbert has quit [Disconnected by services]
egbert has joined #linux-sunxi
<willmore>
A little airflow--no heatsink--drops temps to 61-60 and clocks come up full. That's pretty impressive for a tiny little board.
Da_Coynul has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
cbbc has joined #linux-sunxi
Da_Coynul has joined #linux-sunxi
muvlon has joined #linux-sunxi
cnxsoft has joined #linux-sunxi
lkcl has quit [Ping timeout: 256 seconds]
muvlon_ has joined #linux-sunxi
muvlon has quit [Ping timeout: 268 seconds]
lkcl has joined #linux-sunxi
perr has quit [Ping timeout: 252 seconds]
terra854 has quit [Quit: Connection closed for inactivity]
ninolein_ has quit [Ping timeout: 256 seconds]
ninolein has joined #linux-sunxi
muvlon_ has quit [Quit: Leaving]
<jrg>
I should try it in my opi+2e
<willmore>
jrg, give it a shot.
<willmore>
armbianmonitor -p;minerd --benchmark
<willmore>
KotCzarny, that table of allwinner boards has an error. The eMMC value for the Orange Pi PC 2 says "1GB" but it has no eMMC. Do I fix that on that page or is that generated from some other source?
<willmore>
I think there are other bugs. Some of the SoC lines are redlinks or not even links at all.
Da_Coynul has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
cbbc has quit [Quit: Page closed]
lennyraposo has quit [Ping timeout: 256 seconds]
<dgp>
willmore: you're talking about the schematic?
tom9874 has joined #linux-sunxi
tom9874 has quit [Client Quit]
lennyraposo has joined #linux-sunxi
<wens>
willmore: some of the newer SoCs don't have wiki pages
Amit_T has quit [Read error: Connection reset by peer]
<MoeIcenowy>
jernej: A64 uses TCON1 for HDMI?
<MoeIcenowy>
Wizzup: can you run ls /dev/block on your stock android?
<MoeIcenowy>
jernej: now I met HDMI phy problem :-(
gzamboni has joined #linux-sunxi
[7] has quit [Ping timeout: 258 seconds]
TheSeven has joined #linux-sunxi
leviathanch has joined #linux-sunxi
IgorPec has quit [Ping timeout: 256 seconds]
reinforce has joined #linux-sunxi
scream has joined #linux-sunxi
<KotCzarny>
willmore: it's fully manual work, so go ahead
<miasma>
i already fixed it :)
<miasma>
would it make sense to have a separate page for u-boot/kernel version histories or just cut off really old stuff. e.g. on the u-boot page the version history is already about half of the length of the whole article
<KotCzarny>
any point to run old uboot?
<KotCzarny>
my guess is situation is similar as with old kernels
<miasma>
i think that's another topic. the articles are just getting quite long and the main content is a bit hidden behind the long listing
IgorPec has joined #linux-sunxi
<KotCzarny>
yeah, but for some devices legacy docs are still useful
<miasma>
my suggestion was to move the listing to a uboot/history-subarticle
<miasma>
with few more releases 90% of the article content is just version history :)
<KotCzarny>
maybe just similar table as for the kernel?
<KotCzarny>
with support described per version/soc
<miasma>
possibly. are you kc in the wiki? i saw some unfilled tables there on the user page
<KotCzarny>
hmm?
<miasma>
no it was some other guy
<miasma>
there were lots of 'status' tables on the user page
<miasma>
anyway i'm not sure what the table should contain. there are so many different drivers that need to be covered
<miasma>
and creating/modifying tables for mediawiki is a pita :)
<KotCzarny>
hmm, for example, support for: sdcard/mmc, nand(he he), network(wired, wifi?), usb
<KotCzarny>
maybe hdmi/display too
<KotCzarny>
gpio, spi?
<miasma>
i can collect a list from the items listed so far, but it might not include everything
medvid has quit [Ping timeout: 260 seconds]
scream has quit [Remote host closed the connection]
<KotCzarny>
its a wiki, if anything notices anything missing he/she can add it, but having at least starting point done makes it easier
<miasma>
KotCzarny: it also lists stuff like 'XYZ rewrite'. should there by a note about broken functionality in older versions
<miasma>
let's say the table lists gpio support for platform X since kernel Y. now the next uboot has gpio fixes. there might be multiple fixes in several coming versions :-/
<KotCzarny>
then you can add it in the notes, or just bump the version supported since
<miasma>
ok, let's see. i'll generate an initial version then
sunxi_fan1 has joined #linux-sunxi
<KotCzarny>
also, while you are playing with the wiki, maybe you can add a redirect for uboot->u-boot
<KotCzarny>
or even rename u-boot page into u-boot-legacy and Mainline U-Boot -> u-boot
<KotCzarny>
(but you should consult otherfolks in that regard)
<KotCzarny>
well, u-boot -> u-boot-sunxi
<wens>
you could split it out into a U-boot Mainlining Effort page?
<wens>
which would be a changelog + todo list?
<KotCzarny>
+status table
<Wizzup>
willmore: that is what I was asking as well
<Wizzup>
MoeIcenowy: can do that, later today
valkyr1e has joined #linux-sunxi
<dgp>
Has anyone used JTAG on the H2/H3?
sunxi_fan1 has quit [Ping timeout: 260 seconds]
medvid has joined #linux-sunxi
JohnDoe_71Rus has quit [Read error: Connection reset by peer]
<ssvb>
now I have pretty strong objections and would like to revert your change
<miasma>
ssvb: i'm not sure if it's the optimal way to read it, but my goal was to have all u-boot related bits on the same page
<miasma>
and further configuration for some feature in other pages
<ssvb>
basic U-Boot configuration and compilation is very simple, it makes zero sense to spread it across multiple pages
<miasma>
ssvb: feel free to organize it better. I just found it terribly hard to keep track of all the bits and there already was lots of redundant configs for u-boot through the wiki
<ssvb>
what exactly was hard for you?
<ssvb>
there is one page and it provides step by step instructions
<miasma>
ssvb: ok, in the current form, the mainline article does that
<miasma>
ssvb: but then, you might want to boot from usb, ethernet (nfs, fel, pxe etc.), nand, spi, emmc. it soon becomes quite complex
<KotCzarny>
um
<KotCzarny>
then move tl;dr section to the top
<KotCzarny>
and leave detailed info further in the page
<ssvb>
miasma: what was the purpose of removing the information about the possibility to boot from usb, ethernet (nfs, fel, pxe etc.), nand, spi, emmc?
<ssvb>
miasma: why do the users need to go to another page to find this out?
<miasma>
ssvb: it's not removed. i thought about collecting it to the same page and link to that from articles describing the technologies
<ssvb>
you are forcing the users to do more clicks on the links and have more pages open simultaneously
<miasma>
well.. the original goal was to have them click less
<ssvb>
congratulations, you have achieved exactly the opposite :-)
<miasma>
i could move them back to the uboot article, but it's getting quite long
<miasma>
ok so i'll move them back to the uboot page
<miasma>
ssvb: the confusing part there is, it describes the configuration for both mainline and legacy uboot/kernels, but the page is about mainline u-boot
<ssvb>
it would be great if you actually reported your problem and tried to discuss it rather than making major changes
<ssvb>
the mainline U-Boot can boot legacy 3.4 kernels, what is not clear about this?
<miasma>
ok. so what about the spi/nand boot options, should they be listed on the same article
<miasma>
or on their own pages
<KotCzarny>
imo same page
<miasma>
currently I didn't find any u-boot related configurations for them on the respective pages
<miasma>
ok, so i'll fix that
<miasma>
ssvb: what about the difference in bootm. the legacy article uses the value 0x48000000 and the mainline uboot article 0x42000000. is that hard coded to uboot?
<ssvb>
this value is more or less arbitrary and both are likely fine
<ssvb>
whatever was on the mainline U-Boot page, was providing a known good configuration
<ssvb>
if you want to pick different addresses, be sure to test them first
leviathanch has quit [Remote host closed the connection]
<miasma>
ssvb: ok. i didn't change it. there was this configuration for legacy kernels both on the old and new uboot page. the instructions were basically the same. the address was the only difference
<ssvb>
if you don't understand something, then please ask first and edit later
terra854 has joined #linux-sunxi
<MoeIcenowy>
mripard: if I want to add a simplefb node to H3
<MoeIcenowy>
how should I describe the pipeline?
<MoeIcenowy>
I used "de0-lcd0-hdmi" now
yann has joined #linux-sunxi
Gerwin_J has joined #linux-sunxi
<miasma>
ssvb: I will. My intention was not to delete any information or modify anything especially without testing. I just thought that there might be some duplication of effort when documenting some configuration for both old and new uboot. obviously having a new page just for configuration leads to more clicking. i just feared that the main article is getting too long. it's totally ok for me to leave it that way. I was originally planning to extend and ...
<miasma>
... update the articles on netboot and thought that with all those additions the main uboot page would become quite tiresome to read without linking to other articles
<ssvb>
the old uboot is completely irrelevant
<KotCzarny>
ssvb, is uboot-sunxi relevant or mostly mainlined?
<ssvb>
we already have some warning banners at these pages
Putti has quit [Ping timeout: 258 seconds]
<ssvb>
but if this is not enough, then it might be best to delete them
Worf has joined #linux-sunxi
<ssvb>
KotCzarny: uboot-sunxi is an old unsupported version of u-boot
<ssvb>
KotCzarny: yes, some years ago it was relevant
<KotCzarny>
then it would be best to rename u-boot page to u-boot-legacy and add redirect for u-boot -> u-boot-mainline
<KotCzarny>
also to add redirect uboot->u-boot
<ssvb>
whatever
<KotCzarny>
because right now searching for 'uboot' returns wrong results
<MoeIcenowy>
yes, maybe tag everything -sunxi as legacy is right :-)
florianH has joined #linux-sunxi
<MoeIcenowy>
tkaiser: will you try some experimental H3 HDMI mainline drivers for armbian?
<ssvb>
MoeIcenowy: we can look at it on case by case basis, just the -sunxi suffix in the name does not mean anything
<MoeIcenowy>
yes
<MoeIcenowy>
I mean the old U-Boot 2011.09 and Linux 3.4 guys
<MoeIcenowy>
not simply the suffix :-)
paulk-collins has joined #linux-sunxi
Worf_ has joined #linux-sunxi
Worf has quit [Ping timeout: 250 seconds]
IgorPec has joined #linux-sunxi
<miasma>
KotCzarny: the note says "u-boot-sunxi fork. It is only useful for the devices, which are still not supported by the Mainline U-Boot.". does that still include some devices that are not supported by newer u-boot versions?
<MoeIcenowy>
I think there must be some devices which do not have anyone to port it :-)
<miasma>
ok
mhlavink_afk has joined #linux-sunxi
mhlavink has quit [Ping timeout: 268 seconds]
<ssvb>
miasma: who knows? there may be one or two abandoned devices
<KotCzarny>
isnt it about socs and not devices? is there any device that needs special treatment ?
<ssvb>
KotCzarny: each device needs its own defconfig
<ssvb>
and somebody has to provide it
LargePrime has quit [Ping timeout: 260 seconds]
<KotCzarny>
uhum
Leepty has joined #linux-sunxi
silviop has joined #linux-sunxi
<MoeIcenowy>
some devices may use rare features and needs special treatment.
<wens>
it's actually easy to differentiate, just look at dmesg
ldearquer has joined #linux-sunxi
<wens>
uh, i forgot which of my boards used it
<ldearquer>
I am trying to read uboot.bin from my A33 tablet, but when copying and mounting nanda on my local machine, it comes up with no boot binaries!
<MoeIcenowy>
the script.bin is now unavailable on the nanda.
<MoeIcenowy>
it's now in invisible area of nand.
<ldearquer>
Ok, i thought so
<ldearquer>
same with uboot.bin, isn't it?
<MoeIcenowy>
why do you want uboot.bin?
<MoeIcenowy>
uboot.bin is not accessible inside Linux, no wonder which version.
<ldearquer>
nothing now, just curiosity, only script.bin, because no test point gets me a serial port
<MoeIcenowy>
on A33 boards, the stock uboot uses MicroSD breakout as debugging output
<MoeIcenowy>
you can give up for a dedicated serial :-(
<MoeIcenowy>
for script.bin, dd if=/dev/mem of=/sdcard/script.bin bs=1 count=262144 skip=1124073472
<ldearquer>
thanks a lot for saving me ages of trial and error then! :)
<ldearquer>
Does it need to be enabled on script.bin?
kaspter has joined #linux-sunxi
<MoeIcenowy>
you have no way to change script.bin on A33
<MoeIcenowy>
except have a PheonixSuit image, and modify it
<paulk-collins>
anyone got a known to work tar cmdline to pack a rootfs?
<ldearquer>
I probed the wires into the sd card socket, but saw nothing...
<MoeIcenowy>
try to reboot your device
<ldearquer>
can't i dd from nand, modify and dd again into nand?
<silviop>
By the way , my A33 tablet was fake RAM android and chip ram marking reports 1Gb but it real have only 512Mb !!! So no bug only cinese hardware!!
<ldearquer>
oh, I see, /dev/mem
<KotCzarny>
silviop: lol, so my wild guess was right
<KotCzarny>
you can add that info to wiki now
<ldearquer>
MoeIcenowy: so how do you enable serial?
<silviop>
Yes i daa info to wiky with RED mark!!
<silviop>
Hans de Goedes says that does not exist q8 form factor tablet with 1Gb of ram!!
<MoeIcenowy>
I do have an A33 tablet with 1GB ram
<MoeIcenowy>
but it's not Q8.
<MoeIcenowy>
It's a iPad-like tablet
<silviop>
But i feel you tablet in not 35¬ shipped !!
<MoeIcenowy>
you mean the price in dollar?
<silviop>
Euros
<KotCzarny>
they say you get what you pay for
<MoeIcenowy>
I got it for ￥460 CN, without any shippment (as I bought it from Shenzhen
<miasma>
does q8 mean it's 8" in size?
<jelle>
miasma: I think q8 means the casing
<silviop>
q8 means 7" 19:9
<miasma>
ok
<silviop>
16:9
<silviop>
i use q8 as room interface for domotic
<MoeIcenowy>
q8 is fistly a casing form of A13 tablets
<silviop>
i have five A13 and two A33
bmeneg has quit [Quit: Leaving]
<miasma>
i wouldn't be surprised if a 35eur tablet doesn't come with 1GB RAM
<MoeIcenowy>
but now some Q8 devices slightly broke the casing rules
<silviop>
the problem is that not exist a 40-45¬ 1Gb q8
<MoeIcenowy>
high-end aw vendors design their own models
<MoeIcenowy>
and won't use the public model of Q8
<KotCzarny>
so it was a reference design from aw?
<MoeIcenowy>
e.g. Onda
<MoeIcenowy>
Q8 maybe is not created by aw
<MoeIcenowy>
but at least it's a common design
<silviop>
i think that in fall 2016 1Gb ram and android 5.1 is very low end design
<MoeIcenowy>
to be honest in Shenzhen one can easily get another's design :-)
<miasma>
silviop: sure, but if you also had to pay VAT, the price is indeed very low
<MoeIcenowy>
In the world of A33, 1GB is high-end ;-)
<KotCzarny>
:)
<KotCzarny>
in the world of aw socs
<MoeIcenowy>
not for all aw socs
<silviop>
in '80 handheld electronic calculators are quite expensive but diffused, in '90 they are gifted (during euro transition some government gift euro converter that is electronic calculator), low end tablet way is the same
Da_Coynul has joined #linux-sunxi
popolon has joined #linux-sunxi
bmeneg_ has joined #linux-sunxi
bmeneg_ has quit [Read error: Connection reset by peer]
bmeneg_ has joined #linux-sunxi
bmeneg_ is now known as bmenef
bmenef is now known as bmeneg
bmeneg has quit [Client Quit]
Putti has joined #linux-sunxi
bmeneg has joined #linux-sunxi
<ssvb>
silviop: do you have a picture of these RAM chips?
matthias_bgg has joined #linux-sunxi
ldearquer has quit [Quit: Leaving]
bmeneg has quit [Quit: leaving]
<silviop>
i put tablet in sunxi wiki
<silviop>
et_q8_v2.0
<ssvb>
just deciphering the dram chip markings is not always easy
<silviop>
they are fake too i suppose
<ssvb>
you may have either one x16 chip, or two x8 chips
<silviop>
And if i use an android app like "fill ram memory" it closed by os near 500Mb fill memory
<MoeIcenowy>
silviop: it's really fake ram.
<ldearquer>
Just to confirm, does it mean I won't be able to see uboot/kernel output through serial? uart0 is disabled (and cmdline shows console=ttyS0,115200). No /dev/ttyS0, obviously...
<MoeIcenowy>
ldearquer: yes
<MoeIcenowy>
what do you want to do?
<ldearquer>
...only /dev/ttyS1 exists, which is consitent to the script.bin
<ldearquer>
boot form USB
<ldearquer>
but I need serial
<MoeIcenowy>
I think the ttyS1 is used for bluetooth
<ldearquer>
to see what happens
<MoeIcenowy>
what means "boot from USB" ?
<ldearquer>
FEL
<MoeIcenowy>
mainline guys?
bmeneg has joined #linux-sunxi
<ldearquer>
sorry, don't understand what you mean :(
Ultrasauce has joined #linux-sunxi
<ldearquer>
I'd like to run sunxi-fel uboot, but wanted to make sure first I could read from serial
<MoeIcenowy>
what do you want to boot from USB?
<MoeIcenowy>
mainline U-Boot?
<ldearquer>
ok, mmm, not necessarily, haven't got that far yet. I am following the wiki, and I can enable FEL mode, but before downloading anything, wanted to see if I could read serial
<MoeIcenowy>
do not think about serial on A33
<MoeIcenowy>
please
<MoeIcenowy>
A33 has now U-Boot LCD support
<ldearquer>
But I am seeing I won't be able with the factory sfotware in it
<MoeIcenowy>
(except your board uses MIPI DSI LCD)
Da_Coynul has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
<MoeIcenowy>
sunxi-fel cannot even boot legacy U-Boot/kernel
<ldearquer>
Phew, that's a lot of things...
<MoeIcenowy>
forgot serial first
<ldearquer>
ok
<ldearquer>
so I'll be seeing UBOOT otuput on the LCD?
<MoeIcenowy>
on A33 and Mainline U-Boot, yes
<MoeIcenowy>
Do not try to hack any factory guys
<ldearquer>
ok
<ssvb>
silviop: but I guess Android also needs some amount of RAM for itself, you can probably try to read from /dev/mem and compare data at 0x40000000 and 0x60000000?
<ldearquer>
and what do you mean with legacy uboot/kernel?
<ldearquer>
once I have mainline, what's wrong with enabling serial over sdcard? (sorry for bringin serial back...)
<MoeIcenowy>
you must modify a lot of things
<MoeIcenowy>
and you will lose the most important device -- mmc0
<ldearquer>
But I could get it back just changing a few lines on the device tree file?
<MoeIcenowy>
serial cannot be used along with mmc.
<MoeIcenowy>
on most A33 tablets
<MoeIcenowy>
so if you are not seeing really serious issues on A33, please do not consider to use serial
<ldearquer>
ok thanks a lot for all your help :)
<ldearquer>
may I ask, how do you know that much? Do you work for allwinner?
<MoeIcenowy>
nope, but I have hacked A33 for more than one year ;-)
<jelle>
don't think anyone who works for allwinner is here?
<ldearquer>
:)
<MoeIcenowy>
I think allwinner guys may even have no interest on what we do ;-)
<MoeIcenowy>
they only cares provides new SoC capable of Android, then forgot the SoC and go to another
<jelle>
probably..
<silviop>
ssvb: mem compare is done by mainline uboot that report 512Mb :-( yesterday i loose a day of my life finding a bug in this routine , android can be fooled you ca search in internet about
<ldearquer>
so how do the olinuxino guys create their boards?
<KotCzarny>
ask them?
<ssvb>
silviop: I would still try to do it in Android too, it seems a bit too elaborate to fake it in both the software and the dram chip markings
<MoeIcenowy>
ssvb: simply a busybox free can show the problem :-)
<ssvb>
MoeIcenowy: and what does it report for you?
<ldearquer>
KotCzarny: I meant to ask if, with the resources available at linux-sunxi, could someone create a custom A33 board from scracth?
<silviop>
where free thak info about memory ? is a syscall or /proc /sys ?
<jelle>
silviop: 'free' the linux program
<KotCzarny>
folks here are software guys mostly
<ssvb>
silviop: maybe run 'busybox free' in the adb shell?
<silviop>
i run it in android terminal
<silviop>
why do you tihnk adb shell is different ?
<MoeIcenowy>
and cma can be shared with system memory
<ssvb>
silviop: it was just chopped a bit in the mainline u-boot
<silviop>
i'm not a firmware programming , i can understand how thinghs work but it require a lot of time for me and the manteiner say that i have e fake q8 and he never see a 1Gb q8
Putti has joined #linux-sunxi
<MoeIcenowy>
if busybox free returns a number lower than 512MB in stock firmware
<MoeIcenowy>
you can believe you have really a tablet with 512MB ram
<jelle>
silviop: so what does your vendor u-boot report?
<jelle>
or is there no detected ram size in there?
<MoeIcenowy>
I think there's no
* jelle
has no hw atm
<silviop>
I don't have sd breakout now , but i suppose that original uboot does not output nothing
<jelle>
ohh or solder uart connectors :)
<silviop>
i have sdbreakout at office but now i'm at home
<MoeIcenowy>
uart breakout is more safe
<ssvb>
still the best test is probably to dump data from /dev/mem in android and check if there is anything meaningful there above 512mb
<silviop>
i'm going to modify script_extractor to do this!!
<ssvb>
if there is some unique data at both 0x40000000 and 0x60000000 physical addresses, then you have 1GB RAM
<silviop>
Tomorrow i will update you
<KotCzarny>
is adb blocked on a33 or what?
<KotCzarny>
simple cat /proc/meminfo would be enough usually
<ssvb>
silviop: if you only have 512MB RAM, then it will either wrap around (you will read duplicate data) or it will be empty above 512MB
<silviop>
procmeminfo does not give same result than free
<silviop>
i suppose that free make system call
<silviop>
and not take ram info from /proc
<ssvb>
MoeIcenowy: btw, mali does not need cma
<ssvb>
MoeIcenowy: it has its own mmu
yann has quit [Ping timeout: 260 seconds]
<MoeIcenowy>
ssvb: yes
<silviop>
i strace free in my amd64 linux and it take info from /proc/meminfo
<silviop>
so busybox free is different from standard free
<dgp>
silviop: Android will run statically compiled arm linux binaries just fine if you need to do that
<MoeIcenowy>
I'm going to write an A33 tablet hacking guide on my blog...
<KotCzarny>
make a link in wiki too
<KotCzarny>
or write in both places
<MoeIcenowy>
I'm too lazy to make it suitable for a wiki page...
<miasma>
let's hope they'll update kernels someday so the phones/tablets could just run docker containers :)
<silviop>
i'm trying to crosscompile
<KotCzarny>
MoeIcenowy: then write on your user page?
<MoeIcenowy>
KotCzarny: after me finishing it, maybe I will port it to my user page
<KotCzarny>
:)
<MoeIcenowy>
but firstly I will make it on my blogspot
<KotCzarny>
i guess copy paste will be enough
<silviop>
there is one hardware hack i'm interested , extract USB0 bus from tablet that use SDIO for wifi
<MoeIcenowy>
I'm also interested :-)
<MoeIcenowy>
but I'm not good at hardware hacking...
<MoeIcenowy>
on one of my tablet's board, I saw four points which looks like USB points seen on tablets with USB Wi-Fi
yann-kaelig has joined #linux-sunxi
<willmore>
KotCzarny, NiteHawk, not sure who to tell, but the wiki leaks account email info on the password reset page--it will tell wether an email is used by an account or not. Best practice is to give a generic response regardless of finding a valid user, FWIW.
<silviop>
in A33 i'm working there in a place on pcb for USB Wifi , the difficult is to understand where get D+ D- (GND and 5V are quite simple to find) , configure D+ and D- as gpio and let them "blink" should be possible
<MoeIcenowy>
they are not multiplexed
<KotCzarny>
i guess its not that important, and being too tight lipped might be more user unfriendly
<KotCzarny>
im not admin though
<MoeIcenowy>
but if you got GND and 5V, you can get D+ and D- by the usual USB sequence
<KotCzarny>
imo. it would be nice to fix email filter, to actually match '.' in '*@*', right now with noscript anything near @ is obfuscated
<MoeIcenowy>
the pin close to Vbus (5V) is D-, the pin close to GND is D+
<silviop>
not so simple when yo have a square with 25-64 pad !!!
<silviop>
A33 not use classical rtl with four pad
<silviop>
rtl-pcb
<MoeIcenowy>
do anyone have an A23/33 tablet with USB wifi?
<NiteHawk>
Turl, libv: see willmore's remark above
<willmore>
Okay, now that I read the column header, the PC 2 entry is fine. It says eMMC/SPI FLASH and the PC 2 does have the 16Mb of SPI flash.
<willmore>
Though they advertise it as only 8. Should we reflect the advertised value (since that's what they may change it to in the future) or leave it as the 'what actually ships' value?
<KotCzarny>
willmore: if we ever see 8Mb flash
<KotCzarny>
on opipc2
<willmore>
So, I'll leave it alone.
<KotCzarny>
empirical evidence > specs/docs
* willmore
nods
<willmore>
Leaving it alone. But, on the plus side, I'm logged in and can easily edit other things. ;)
silviop_ has joined #linux-sunxi
silviop has quit [Ping timeout: 260 seconds]
silviop_ is now known as silviop
<tuxillo>
NiteHawk: yeah very true!
<tuxillo>
NiteHawk: crazy busy weekend
Gerwin_J has joined #linux-sunxi
<MoeIcenowy>
silviop: I forgot where can get your board's pic...
<libv>
willmore: if that is an issue, then lots of wikis have that issue...
<libv>
but let me dig a bit
<willmore>
libv, I wouldn't be surprised if lots of wikis have the problem.
<libv>
on the other hand, this is not a massive problem
<willmore>
Nope, that it is not.
<libv>
the email account will get an email anyway, which states, "if this was not your action..."
<willmore>
That's why I marked it "FWIW".
<willmore>
Yep, if the right person enters the right email, they get the email. If they enter the wrong value, they don't get the email. They'll figure it out. An attacker should get nothing.
<libv>
and there might be some people who would like to know which email address it was sent to
<willmore>
Might be.
<libv>
tying a username to an email address usually is not that hard
<willmore>
possibly, but there's no need to make an attackers job easier.
<willmore>
If I run 'poweroff' on my Opi Z, is there a way to restart it short of power cycling it?
<libv>
well, let me go and reset the password on my wikipedia account and see whether it is done there
<KotCzarny>
willmore, buy some relays. ;)
<libv>
and if it is the same there, it will not be fixed at linux-sunxi
cnxsoft has quit [Quit: cnxsoft]
<KotCzarny>
or add some code for uboot/linux to utilize openrisc core to react on buttons
<KotCzarny>
(i was thinking about it, but got sidetracked as usual..)
<libv>
and we've got work to do.
The_Loko has joined #linux-sunxi
<willmore>
thanks for looking into this, libv.
<willmore>
KotCzarny, yeah, I was going to ask apraizel.
lemonzest has quit [Quit: Leaving]
leviathanch has joined #linux-sunxi
yann has quit [Ping timeout: 252 seconds]
<willmore>
Do any of the allwinner chips support Quad-SPI or dual SPI?
<willmore>
Looks like the anwser is no on the H5 at least.
yann has joined #linux-sunxi
Worf has quit [Quit: Konversation terminated!]
tkaiser has joined #linux-sunxi
<tkaiser>
buZz: Also on sale today 'Alpha CHIP 10-pack', that are those with R8M (A13 with DRAM + NAND in a single package showing 300 mW less idle consumption)
<ldearquer>
The A33 tablet I am using doesn't seem to have any LCD chip -it looks like LCD it is wired straight to the cpu. Does it mean it uses MIPI?
<Seppoz>
anyone know a cheap wall mountable table/fram with touchscreen?
<hojnikb>
with customs and vat thats over 50€
<hojnikb>
i got my z3735f 1gb tablet for 40€
<ssvb>
I tried to contact the seller nicely, but got no response
<hojnikb>
thats with ips display, gps and android 5.0
<hojnikb>
and a 3y warranty .)
<silviop>
Is not so simple to detect , android is completely fooled , only syscall to sysinfo give right value but all android tools use /proc/meminfo
leviathanch has joined #linux-sunxi
<silviop>
In italy small packet don't pay customs+vat
<ldearquer>
From http://linux-sunxi.org/LCD, lcd_hv_hspw and lcd_hv_vspw should be used for custom UBoot LCD configuration. My script.fex doesnt' have these, but lcd_hspw and lcd_vspw, are they the same?
reinforce has joined #linux-sunxi
<buZz>
ldearquer: i highly doubt that
<hojnikb>
silviop: isn't everything over 20€ taxed ?
<hojnikb>
at least thats the case in slovenia here
<Seppoz>
silviop: customs will ask you to declare the price. so it depends if they believe the sender and yoz
berkutta has quit [Ping timeout: 248 seconds]
<Seppoz>
silviop: they have a huge databse they can check against
<hojnikb>
i had never have this luck. Everything over 22€ was taxed for me, even if declared price was lower
<buZz>
ldearquer: thats not what that means
<hojnikb>
customs just want a receipt
<hojnikb>
so you can't avoid it unless you fake it
<silviop>
It's italy friend!! Sun,Sea and no customs+vat from china!!
<Seppoz>
yes, but if they dont believe the chinees attached ones they will ask you to send one
<Seppoz>
and if you send them the wrong its fraud
<hojnikb>
exactly
<miasma>
hojnikb: well, people who don't pay the taxes (if they are obliged to) are basically breaking the law
<KotCzarny>
miasma: lets not go into politics
<hojnikb>
agree
<Seppoz>
also they tax is very low
<ldearquer>
buZz: I know nothing of Ruby, but I assumed that picked any of them, with maybe priority to the first
<Seppoz>
couple percent
<buZz>
ldearquer: var.max isnt the same as var
<buZz>
oh you're talking about other variables now
<ldearquer>
buZz: yup, is about the ORed ones
<scelestic>
i dont mind paying tax but an additional 15 euro administration cost is just ridiculous
<buZz>
lcd_hv_hspw instead of lcd_vspw
<ssvb>
buZz: the .max array method returns the largest value in the array
<miasma>
KotCzarny: not sure if abiding by the law or stating the obvious is politics, but ok
<buZz>
ssvb: indeed
<buZz>
but nothing about that line talks about _vspw
<KotCzarny>
miasma: not every law is valid/sincere/honest. and that's politics
<ldearquer>
no, the question raised because the web says
<ssvb>
buZz: there is a similar line for _vspw too
<ldearquer>
Horizontal Sync Length hs lcd_hv_hspw (with a minimum of 1)
<miasma>
KotCzarny: i wasn't arguing about the laws
<buZz>
ssvb: that still does not make me think they are the same number :D
<ldearquer>
but I had no lcd_hv_hspw, only lcd_hspw, but the script shows that they are the same-ish
<buZz>
thats not what you asked
<buZz>
you said > but lcd_hspw and lcd_vspw, are they the
<buZz>
same?
<ssvb>
one is just 'horizontal' and another is 'vertical'
<Seppoz>
notebooks arent even taxed btw
<ssvb>
but essentially they are the same
<ldearquer>
err, yeah, the same than lcd_hv_hspw and lcd_hv_vspw
<ldearquer>
hehe, "respectively"
<ssvb>
KotCzarny: as I mentioned before, I'm a little bit annoyed when these Chinese sellers are lying about the package price even when they don't have to
silviop has quit [Remote host closed the connection]
<ssvb>
KotCzarny: for example, claiming $2 price when the thing actually costs $10
<ErwinH>
They probably aren't lying... :)
<KotCzarny>
maybe they writing what the item cost to THEM ;)
<KotCzarny>
*they're
<jelle>
I ordered a replacement touchscreen which had a different chip :( (so I can't work further on the zeitec touchscreen driver)
<ErwinH>
That's the price they should be charging, or the price without the shipping costs.
<miasma>
i'm a bit annoyed that I need to order the stuff from china. it would make a lot more sense to send a larger shipment overseas so they could package the stuff more efficiently and provide local support
<miasma>
but this way the cost goes up 900%
silviop has joined #linux-sunxi
<KotCzarny>
and that's why we buy aw. the cost.
<miasma>
they actually sell orange pis locally here, but the price is 39,95 eur for opi pc 2
<miasma>
shipping still takes several weeks
<hojnikb>
thats a pretty good margin :)
afaerber has joined #linux-sunxi
<hojnikb>
i wonder whats the BOM on these boards
<ssvb>
when my Cubieboard got stuck at the customs and I needed to "liberate" it, I had to show a printout of the linux-sunxi mailing list post about Cubietech donating free boards to developers :-)
<miasma>
hojnikb: that 39,95 doesn't include the 24 tax
<miasma>
*24%
<hojnikb>
lol
<hojnikb>
so its closer to 50€
<hojnikb>
thats a nice 100% margain
tkaiser has joined #linux-sunxi
<hojnikb>
sell one get one free
<miasma>
hojnikb: still cheaper than rpi3, 79,90 euros (+ 4,10 shipping)
<hojnikb>
masma: wherer do you live ? those are some outrages prices
<dgp>
shipping from china only takes a week so I don't mind it so much :)
<miasma>
hojnikb: both stores have legions of happy customers :)
ErwinH has quit [Ping timeout: 260 seconds]
<KotCzarny>
hojnikb: law of the market, if there are customers, there will be a seller
<hojnikb>
i guess you're right
<ssvb>
I can imagine being really embarrassed if the customs people decide to stop a small shitty $10 package and the price is declared as $2 on the package sticker
<KotCzarny>
many people are afraid to order from anything other than their home country
<hojnikb>
thats pretty sad
<hojnikb>
but true
<KotCzarny>
ssvb: i wouldnt be embarassed if it was unknown to me
<hojnikb>
i really wonder what kind of margins is xunlong working with
foxx_ has quit [Ping timeout: 260 seconds]
<ssvb>
I just don't order anything more expensive than 22 euro from aliexpress for exactly this reason, I don't want to participate in breaking any law :-)
afaerber has quit [Ping timeout: 258 seconds]
<scelestic>
i think of all my aliexpress order as investing in china :P
<KotCzarny>
scelestic: even if you buy electronics from local market, you are investing in china
<KotCzarny>
:P
<scelestic>
KotCzarny: yes but also buying someone a new boat or expensive car ;)
<KotCzarny>
yup
<scelestic>
2 euro's for a 3.5mm plug
<KotCzarny>
time and peace of mind has a value too
<miasma>
ssvb: some stores have an insurance which means that they pay you back if you scan the papers from the customs for them
<miasma>
but it won't necessarily cover the whole tax (e.g. dx.com)
<hramrach>
they are like 200M xz images which you unpack to around 2G raw image, copy to a card, and it resizes itself to match card size on first boot
<hramrach>
grub chainloaded from u-boot, modular distribution kernel with ramdisk - or so it seems
<hramrach>
did not look too closely ;-)
IgorPec has joined #linux-sunxi
<hramrach>
root password is linux. searched for that for a while
<hramrach>
the resize takes a while and only shows output on serial console
Ultrasauce has quit [Ping timeout: 245 seconds]
<ldearquer>
On my A33 tablet, I was trying to make a change on /dev/block/nandd (system) by reading the whole file, mounting it on my host system and writing it back with cat nandd > /dev/block/nandd
<ldearquer>
Permissions are fine, but unlike with nanda and nandc, I am not allowed
<ldearquer>
Could it be disallowed by the kernel at low level?
<ldearquer>
In other words, is it common practice to prevent access to nandd partition?
<NiteHawk>
btw: does any of you know of a CI service that would allow build testing on *BSD?
<tuxillo>
buildbot
<tuxillo>
well at least I did some tests with dragonfly and it worked okay
<NiteHawk>
yes, but that would be self-hosted - right?
<tuxillo>
not sure if there is something on the cloud
<ElBarto>
NiteHawk: yeah it seems that NetBSD needs thoses extra defines
<ElBarto>
don't know any popular CI that do BSD
<ElBarto>
don't know any non-popular too :)
<NiteHawk>
ElBarto: what version did you check that on?
<ElBarto>
12-CURRENT
IgorPec has quit [Ping timeout: 250 seconds]
<ElBarto>
I can test on 11 (stable version if you want)
<ElBarto>
but I know the answer
<ElBarto>
those files haven't been touch in ages :)
<tuxillo>
hehe
<ElBarto>
no problem on 11.0
<ElBarto>
it can't work on 10.3 because our libusb is too old it seems
<tuxillo>
ok, so freebsd is tested
foxx_ has quit [Ping timeout: 260 seconds]
yann-kaelig has quit [Quit: Leaving]
alexxy has quit [Ping timeout: 260 seconds]
alexxy[home] has joined #linux-sunxi
<NiteHawk>
ElBarto: seems strange, sys/endian.h should #define le32toh() etc. iirc, NetBSD needs a special feature marker (#define _NETBSD_SOURCE), maybe that matters here
<ElBarto>
NiteHawk: oh right, I missed the defines in netbsd source, I didn't test on NetBSD, I tought you ask about FreeBSD version
fkluknav has quit [Ping timeout: 260 seconds]
<tuxillo>
I'll be trying in NetBSD in a minute
deskwizard has quit [Ping timeout: 260 seconds]
<NiteHawk>
ElBarto: I don't think anybody is actively using/testing NetBSD sunxi-tools (at least no complaints so far ;)), so my main concern was in fact FreeBSD on your side - as we already "fixed" it once, and I don't want to break it again :D
berkutta has joined #linux-sunxi
<willmore>
ssvb, I was looking at the SPI NOR datasheet and though "hey, can we make this faster?"
Ixnus has joined #linux-sunxi
<ElBarto>
NiteHawk: nobody is really using/testing sunxi-tools on FreeBSD too :)
<tuxillo>
:D
<ElBarto>
NiteHawk: I only needed it for booting on CHIP, never used it before
<buZz>
woo CHIP reference
* buZz
makes hipster sings
dh1tw has joined #linux-sunxi
<buZz>
signs*
<buZz>
:P
<NiteHawk>
kids, steer clear of those drugs :)
Mr__Anderson has joined #linux-sunxi
<buZz>
kids arent made with drugs
<buZz>
although it might help :D
Ixnus has quit [Client Quit]
yann-kaelig has joined #linux-sunxi
Ixnus has joined #linux-sunxi
<willmore>
It help to make them, it helps to keep them....
<MoeIcenowy>
jernej: you can read the code of my fork of your u-boot repo
<tuxillo>
NiteHawk: what is this ? # For functions still missing below, try to substitute 'historic' OpenBSD names
<jernej>
MoeIcenowy: Why do you think so?
<MoeIcenowy>
jernej: A64 DE has LCD output
<jernej>
I think only real difference for that are TCON settings
<MoeIcenowy>
yes...
<MoeIcenowy>
maybe some magic number changed for A64?
<MoeIcenowy>
and for A64 there's a PLL_VIDEO1
<MoeIcenowy>
which is not present on H3
Gerwin_J has quit [Quit: Gerwin_J]
<NiteHawk>
tuxillo: oh, that's supposed to be a comment - i'll fix that. the preprocessor chokes on that, or?
Ixnus has quit [Quit: Page closed]
<tuxillo>
;)
<tuxillo>
yeah
* NiteHawk
searches for the brown paper bag
yann has joined #linux-sunxi
Pepe has joined #linux-sunxi
<MoeIcenowy>
jernej: I think HDMI PHY won't need a working TCON, right?
<jernej>
For initialization and EDID reading, no
Andy-D has joined #linux-sunxi
<MoeIcenowy>
so HDMI PHY init timeout problem is independent to TCON problem, right? :-)
<jernej>
yes
yann has quit [Ping timeout: 265 seconds]
<jernej>
I compared A64 and H3 datasheet and even disassembly of Allwinner binary blob and everything to the PHY init point seems to be equal
nove has joined #linux-sunxi
mossroy has joined #linux-sunxi
<jernej>
ignore for now. during driver writting I found out that for EDID reading this timeout doesn't change anything
<jernej>
you can double check every register setting/address with datasheet (mostly clocks)
<MoeIcenowy>
should I try to make the timeout non-fatal
<MoeIcenowy>
?
<jernej>
it already is
<MoeIcenowy>
oh yes..
IgorPec has joined #linux-sunxi
<tkaiser>
In case anyone is interested, dump of /sys/devices from R40/3.10.65: http://sprunge.us/VcKM
<igraltist>
tkaiser: which board?
<tkaiser>
Banana Pi M2 Ultra, AFAIK the only one so far with R40
<NiteHawk>
tkaiser: any details on SoC ID and SRAM/"swap buffer" layout yet?
<tkaiser>
NiteHawk: Nope, I also do not have physical access to a board
<NiteHawk>
ah, i see. the register dump made me think so
<tkaiser>
NiteHawk: That's just the result of find /sys/devices -type f | while read ; do echo -e "${REPLY}: $(cat "${REPLY}" 2>/dev/null | tr '\n' '|')"; done | curl -F 'sprunge=<-' http://sprunge.us
<MoeIcenowy>
oh R40 have a G2D...
<KotCzarny>
:)
<KotCzarny>
so its good ol' a20 with updated bits
<KotCzarny>
i wonder if they work on replacing cores with a53 ones ;)
<buZz>
a73*
<buZz>
:D
<MoeIcenowy>
but R40 disp is disp2...
<buZz>
lets dream BIGGER
<MoeIcenowy>
buZz: you thought too more
<MoeIcenowy>
allwinner failed on their SoC with a big core -- A80
<buZz>
doesnt ARM licensing include 'plz fix out shitty silicon' ? :D
<buZz>
our*
<MoeIcenowy>
oh R40 continue to have both GMAC and EMAC...
<MoeIcenowy>
and continue to have PS/2 ... (and two PS/2 !)
<KotCzarny>
ps/2 ?
<MoeIcenowy>
PS/2 mouse/keyboard interface ;-)
<KotCzarny>
it still lives?
<jelle>
good thing I still have a PS/2 keyboard :D
<Pepe>
me too and typing from it!
<MoeIcenowy>
I've heard some game players who continue to use PS/2 keyboards because of USB HID protocol cannot support >5 keys pressed at a time
<MoeIcenowy>
but I don't think Banana Pi M2 Ultra have the PS/2 interface wired
<KotCzarny>
rrrright, as if anyone going to press 5 buttons at once
<MoeIcenowy>
and will anyone make a board which uses the two MACs at the same time...
<ssvb>
but you have 10 fingers, no?
<KotCzarny>
20
<ssvb>
right :-)
<KotCzarny>
21 if i count few other things
<MoeIcenowy>
you may need two keyboards to use all the 20 fingers :-)
deskwizard has joined #linux-sunxi
<nove>
also there is the nose
<KotCzarny>
good player cant take eyes away from the screen
BenG83 has joined #linux-sunxi
<igraltist>
i ordered a orangepi-pc will see how this work. searching for boards for a small cluster
<tkaiser>
igraltist: Raspberry Pi Zero instead? Slow as hell so you can use more cluster nodes! ;)
* dgp
isn't sure he get's the point to chaining together multiple machines with relatively low CPU performance and crap IO
<igraltist>
i will see
<igraltist>
as compilestation this would work good i think
<dgp>
You could add all the nodes you like, my old 4790k with 32GB of ram will beat it
<igraltist>
this is not my goal to compare with others
<dgp>
I just don't get it really... If I was going to tape together a bunch of slow CPUs it would be something like m68k or z80 where you could actually write everything running on it
foxx_ has joined #linux-sunxi
<igraltist>
its not your playground for me its fine
<buZz>
:)
<KotCzarny>
dgp: low power compilation cluster?
<KotCzarny>
still needs one good central machine though
<dgp>
KotCzarny: I doubt that it'll save all that much power when you have all the glue to make it work
<KotCzarny>
dgp, i made me decacore one ;) bpi-r1 (that works overshifts as a server/router too) + opipc + opi+2e
<dgp>
I have some Armada370 based thing for a router. One of the only ARM devices I've seen that actually has decent network throughput
<KotCzarny>
for home use banana enough
<KotCzarny>
it really only has to fit usecase
<dgp>
Can it handle 200 or 300mbit through software NAT?
<KotCzarny>
i only have 30/30mbit net pipe
<igraltist>
i have only 5/0.4 Mbit/s
<dgp>
I have gigabit but max in either way is usually only 300mbit
<igraltist>
therefore netspeed is always slow on my side
netlynx has quit [Quit: Ex-Chat]
terra854 has quit [Quit: Connection closed for inactivity]
silviop has quit [Ping timeout: 260 seconds]
<KotCzarny>
dgp, and you should ask tkaiser about software nat performance, as he is the network guy, dont mention bpi-r1 though. ;)
<tkaiser>
KotCzarny: Who mentioned Crapboard R1!? And this is stuff for an 'engine' since those ARM cores are too slow anyway. Look at Marvell.
<dgp>
Armada is the only way to go if you want ARM and decent network performance IMHO
<KotCzarny>
tkaiser: did you test any other board in gmac + usb-network though?
<dgp>
The Armada370 I have has been running for ~200 days without any heat sink :)
<dgp>
and it's in a case the size of two or three CD cases with no air flow
<tkaiser>
For anyone playing around with R40, BPi M2 Ultra and legacy kernel: RPi-Monitor template to monitor device: http://pastebin.com/rSPvsV5p (only value missing/wrong is VDD_CPUX/vcorevoltage -- that's the sysfs node from AXP803/Pine64)
<igraltist>
does the latest u-boot boot the cubietruck from sata
<tkaiser>
Ah, I should add DRAM clockspeed since frequency is adjusted depending on 'budget cooling'. DRAM clocks currently with just 576 MHz and will throttle down to 288 or even 192 MHz (but by default 576 and not the 733 MHz clueless SinoVoip folks mention on their 'product page')
<jernej>
jelle: weird, your board doesn't read EDID correctly. That didn't happen for a long time to me.
mossroy has quit [Ping timeout: 250 seconds]
<jernej>
jelle: I can't find any description of your monitor on the net
<jelle>
jernej: well I have a HDMI => DVI conversion thing which might be borkey
<jelle>
it's passive
<jernej>
jelle: you can try also to revert latest commit, if it makes any difference
<jelle>
I'll try my tv :)
sunxi_fan1 has quit [Ping timeout: 258 seconds]
yann has joined #linux-sunxi
<jernej>
jelle: I tried same monitor with passive HDMI -> DVI cable and it works
<jernej>
seems that it depends on the monitor controller and/or cable
<jelle>
jernej: I'm expecting text in u-boot bootup right?
<jernej>
yes
<jelle>
jernej: hmm blue screen on my tv too
<jernej>
did you compile my branch as-is?
<jelle>
jernej: yup
mossroy has joined #linux-sunxi
<jernej>
can you revert latest commit and try again?
<jelle>
jernej: tv seems to read and detect the correct resolution
<jelle>
jernej: 2d913901506050e84b3dee51569d819b02dc35e7 this one
<jernej>
jelle: yes
<jelle>
ok
vagrantc has quit [Quit: leaving]
<beeble>
if the picture is pink and youbare using a dvi adapter it could be wrong edid parsing. dvi compatible output should always be possible with rgb444 iirc. unfortunatly i cant browse code at the moment
<jelle>
jernej: hmm I'll have to go, I'll check it tommorow
<jernej>
jelle: no problem
<beeble>
thats at least a problem with sunxi bsp on most platforms i have seen
<beeble>
the older the screen the more often i have seen that issue
<jernej>
that would be HDMI/DVI issue. But at least that seems to work
sunxi_fan1 has joined #linux-sunxi
silviop_ has joined #linux-sunxi
silviop has quit [Ping timeout: 260 seconds]
silviop_ is now known as silviop
leviathanch has quit [Remote host closed the connection]