<discopig>
is ethernet support broken for hackberry a10 board on 3.4.61? when I compare dmesg from 3.4.43 to 3.4.61 ethernet initializes properly on 3.4.43 and connects but on 3.4.61 there is absolutely no activity nor any message related to it and it seems to just not react
tinti has joined #linux-sunxi
TheSeven has quit [Disconnected by services]
[7] has joined #linux-sunxi
JohnDoe_71Rus has joined #linux-sunxi
Black_Horseman has joined #linux-sunxi
Soru has quit [Read error: Connection reset by peer]
rz2k has joined #linux-sunxi
akaizen has joined #linux-sunxi
zetaneta has joined #linux-sunxi
zetaneta has quit [Remote host closed the connection]
Black_Horseman has quit [Quit: Zwi se logou mou!!!]
HeHoPMaJIeH has joined #linux-sunxi
juanfont has quit [Ping timeout: 240 seconds]
juanfont has joined #linux-sunxi
\\Mr_C\\ has joined #linux-sunxi
deasy has joined #linux-sunxi
<steev>
discopig: modprobe sunxi_emac ?
<discopig>
let me try
FR^2 has joined #linux-sunxi
panda84kde has joined #linux-sunxi
<discopig>
nope
<discopig>
it just doesn't react at all
<discopig>
there is an eth0 interface but it does not do anything
<discopig>
on 3.4.43 it all works perfectly
wingrime has joined #linux-sunxi
massi has joined #linux-sunxi
wingrime has quit [Ping timeout: 240 seconds]
rellla has joined #linux-sunxi
<steev>
discopig: pastebin your dmesg output.... somehow?
<discopig>
ok
<discopig>
i'll paste the one from 3.4.43 and the one from 3.4.61
<JohnDoe_71Rus>
I thought phoneix image for installation in nand from sd card
Tsvetan2 has quit [Ping timeout: 252 seconds]
Tsvetan3 has quit [Quit: Leaving]
atiti has joined #linux-sunxi
heffer has quit [Remote host closed the connection]
notmart has joined #linux-sunxi
notmart has quit [Changing host]
notmart has joined #linux-sunxi
heffer has joined #linux-sunxi
heffer has quit [Changing host]
heffer has joined #linux-sunxi
<shineworld>
I'm comparing their img with mine
vicenteH has joined #linux-sunxi
arokux2 has joined #linux-sunxi
<arokux2>
discopig: there was a change in config name. WEMAC -> EMAC
<discopig>
yeah
<discopig>
but i don't undestand why eth0 does absolutely nothing on recent kernel
<discopig>
the interface exists but it doesnt do anything at all, both with dhcp or static ip
<discopig>
and doesn't seem to work at all
<discopig>
but on any older kernel it connects within seconds with dhcp
<oliv3r>
i submitted that rename patch like 3 -6 months ago :)
<oliv3r>
discopig: remember, that the emac alone isn't an entire network stack, you also need the PHY driver for your specific PHY. Mostlikly the RTL PHY
<discopig>
yeah
<oliv3r>
and the MII driver to talk mii between the two
<discopig>
i'm using an old .config from back when i compiled 3.4.43
<discopig>
do you think that could be the issue
<discopig>
i didnt modify it since
tinti has quit [Quit: Leaving]
<oliv3r>
very likly
<oliv3r>
sun4i_wemac could have transistioned to sunxi_emac as noted above
<oliv3r>
phy could be missing
<oliv3r>
double check allt hat
<oliv3r>
in kernel dhcp could be gone
<discopig>
yeah i think that may be it, thanks
<shineworld>
JohnDoe_71Rus, I don't understand... when extracted and checked the olimax A20 SD images contains yet the references to nandx how the image was made for nand, but in the initial sentence "Runnning Android from SD card is about 2-3 times slower than from NAND but for these who choose A20-OLinuXino without NAND this is the only way they can run Android on this board." suppose that this is a image to create a bootable SD...
<shineworld>
I've surely misunderstand something about their post
arokux2 has quit [Remote host closed the connection]
<JohnDoe_71Rus>
shineworld: phoneix card for burn NAND image from SD. It can rewrite data drom nand. "PhoenixSuit will detect the board and would ask for the method of writing the image. Choose method of writing the image and confirm your wish to write the image"
<Tsvetan>
oliv3r if you prove me you can ship from where you are to Bulgaria at lower prices using same courier/service I will send your order free of charge
<Tsvetan>
UPS shipping is next day delivery almost everywhere in EU and cost EUR 13-14 for 0.5KG parcel
<oliv3r>
Tsvetan: i can't :p
<arokux1>
Tsvetan: you've sent me the board with ordinary post, right?
<Tsvetan>
aorkux1 yes, this is DHL Global Mail
<Tsvetan>
about 1-2 weeks to any EU destination
<oliv3r>
Tsvetan: UPS -> Uniterruptable Power Supply
<oliv3r>
Tsvetan: i'll buy some @ fosdem from you ;)
<arokux1>
AXP152 wasn't removed from kernel afaik
<Tsvetan>
arokux1 try to compile a10s board then :)
<oliv3r>
I wonder how long the A20 can run from 130 mAh lipo
<oliv3r>
2Euro's with shipping from ebay
<Tsvetan>
oliv3r these are probably chinese milliampers
<Tsvetan>
I bought from local shop battery marked as 4000mAh
<Tsvetan>
we measured it 500mAh
<Tsvetan>
the size was same as our 1400mAh so I wondered how they made 4000mAh in same size
<arokux1>
:)
<arokux1>
Tsvetan: and your batteries are fine?
<Tsvetan>
arokux1 they are with capacity which is marked ;)
<arokux1>
I wonder if there is a battery imitator device, which can be used to verify that charger works correctly
<arokux1>
Tsvetan: (I'm compiling sun5i now.)
<arokux1>
Tsvetan: sun5i_defconfig builds just fine.
<Tsvetan>
arokux1 try to build it for a10s olinuxino
<oliv3r>
Tsvetan: i wonder what 130 chinese mAh are; 50? :p
<oliv3r>
but if A20 can run 5 - 10 minutes from it; that's actually plenty
<Tsvetan>
test points
<arokux1>
in the schematic you have your components spitted apart from each other. so I wonder how to know how they are connected between each other. for example, if you take a look at battery, the outputs of this module are P10 and BAT
<arokux1>
AXP has pins 38, 39 which go to BAT, so this is the connection then?
<oliv3r>
GND and BAT is my guess ;)
<arokux1>
oliv3r: what GND and BAT?
<Tsvetan>
arokux1 in all CADs you can place wire or to label wires, the wires with same label connects together
<arokux1>
I see, thanks Tsvetan
<Tsvetan>
so all wires in the schematic with label BAT connects together
<arokux1>
Tsvetan: small remark, it would be awesome if PDF were searchable
<Tsvetan>
this is convenient as if you had to place wire drawings all around to all signals it will be messy and not readable
wingrime has joined #linux-sunxi
<Tsvetan>
they are
<Tsvetan>
oops
<Tsvetan>
:)
<Tsvetan>
its very easy to make them searchable
<Tsvetan>
one small checkbox in Eagle says use TEXT when print
<oliv3r>
ctrl-f in firefox seems to work if i remember correctly
<Tsvetan>
oliv3r on the A13 I opened its not searchable
<arokux1>
oliv3r: we are talking about PDFs here.
<oliv3r>
yes
<oliv3r>
what else would i reffer too?
<arokux1>
oliv3r: how firefox and pdf are connected? or you open it embedded in ff?
<oliv3r>
firefox has embedded PDF reader
<arokux1>
oliv3r: ok, now it is clear.
<oliv3r>
pdf.js i think it was called?
<arokux1>
oliv3r: I switch it off, to slow.
<oliv3r>
its ok most of the time
<Tsvetan>
arokux1 try again ;)
<arokux1>
Tsvetan: perfect. it looks better even.
<arokux1>
Tsvetan: any special reason you've decided to have an on-board regulator and be 6-16V tolerant?
<Tsvetan>
I do wonder why Eagle default is to print as graphics instead text in PDF
<Tsvetan>
arokux1 to may use any adapter
<arokux1>
Tsvetan: not 5V one, though :(
<arokux1>
Tsvetan: and any won't fit the power jack.
<oliv3r>
faster rendering probably
<oliv3r>
i think my wrt plug fits too
<mouchon2>
hi
<oliv3r>
hey philiippe
<mouchon2>
i am still working on the i2c and i have a question
<oliv3r>
shoot
<mouchon2>
it seem that in the i2c_adap_sunxi_init there is a check to see if we are in sun7i
<mouchon2>
but this check is not done in exit
<mouchon2>
so some unregistred device will try to be unregistred
<mouchon2>
at this time seem to be a bug for me
<oliv3r>
very quite possible
<mouchon2>
oliver i splited the array as you suggested
granite_crusher has joined #linux-sunxi
<granite_crusher>
arokux1: 6-12V is very useful to use it in car as example
<arokux1>
thanks granite_crusher I'm only slowly realizing what a zoo a DC power supplies and connectors is.
<mouchon2>
hum this is not a bug, this is the twi_used_mask that do the check finally seem that my assertion was wrong
<granite_crusher>
Tsvetan: I have question about A13-olinuxino, is it intentional that 5V is always on?
<Tsvetan>
what do you mean always on?
<Tsvetan>
from 5V are generated all other supplies
<granite_crusher>
Tsvetan: even if board is shutdown (3.3 rail drops down to 1V, ... but 5 stays 5)
<Tsvetan>
sure
<Tsvetan>
it should be there in case you push power button
<granite_crusher>
So what is the purpoise of jumper "5V_E" between P9 and P6?
<Tsvetan>
to make the board as much universal as possible
<Tsvetan>
if for instance we have application where Lipo is not needed, we will not populate some components and short this jumper
<granite_crusher>
Tsvetan: so 5V on GPIO1 GPIO2 always on is intentional, and for usage I should use additional FET enabled by gpio?
<Tsvetan>
where you see these GPIO1/2
<Tsvetan>
axp?
<granite_crusher>
I mean 5V pins on GPIO-1 and GPIO-2 on A13 board, they are on, always
<Tsvetan>
ah GPIO connectors, yes
<wingrime>
granite_crusher: you want FETs on it?
<granite_crusher>
wingrime: I was thinking to power GPS on that GPIO connector 5V; but if it is on always it makes no sense to attach anything to that pin, untill it would be able to cut power to device (as GPS in example)
<granite_crusher>
unless*
<granite_crusher>
wingrime: with help of control gpio and FET
wingrime has quit [Ping timeout: 240 seconds]
<oliv3r>
granite_crusher: you are saying that the A13 always does power IN on GPIO-1? so when you read gpio1 you always get high? that sounds like a mistake; oth if gpio1 is always outputting a voltage, well that's a software thing and can be changed/toggled
<oliv3r>
granite_crusher: u-boot console allows you to change gpio's via the gpio set/clear commands
<hno>
hi
<granite_crusher>
oliv3r: not data GPIO but 5V in gpio connector
ganbold_ has joined #linux-sunxi
<oliv3r>
hno: hi!
<granite_crusher>
oliv3r: it is inpossible to put it of in software (unless somebody would make so strung stuff it would melt solder :P )
<oliv3r>
granite_crusher: then i'm not sure i quite understand what you mean
<oliv3r>
granite_crusher: you mean 5V powersupply
<granite_crusher>
*impossible *strong
<arokux1>
hno: you survived? :)
<oliv3r>
granite_crusher: well i'm not supprised that's always on
<granite_crusher>
oliv3r: yes
<hno>
I think so.
<oliv3r>
granite_crusher: for SATA for example (on A10) there's a gpio for 5V power out
<hno>
but not sure the brain actually thinks..
<oliv3r>
hno: uh oh :p
<granite_crusher>
oliv3r: so if you power off board, your SATA just juice out your bat?
<oliv3r>
hno: did you receive your cubieboard3 allready?
<arokux1>
hno: will you have more time for sunxi now?
<oliv3r>
granite_crusher: how do you power off? i don't think a13 has a power off state, nor does the AXP :)
<oliv3r>
granite_crusher: iirc it reboots into a 'battery charge' mode
<oliv3r>
granite_crusher: actually, i think you CAN turn off power of the AXP. if you push the power button for 10 seconds I think you actually put it in super standby mode
<oliv3r>
granite_crusher: but some part of the AXP is still active, as it powers on when you push the button ;)
<oliv3r>
granite_crusher: otherwise, it's quite possile that the 5V rail is simply connected to the LM7805 before the AXP even gets involved
<granite_crusher>
oliv3r: it have kind of at least it puts off 3.3V by cuting off EXTEN(AXP) which goes to DC-DC converters
<granite_crusher>
pin 20 on AXP
<oliv3r>
granite_crusher: 5V connected to the DC-DC of the AXP, or is it connected to VIN
Soru_ has quit [Read error: Connection reset by peer]
<granite_crusher>
oliv3r: which of 5V, there are few DC DC converters, one converts AC to 5V and that goes to ACIN on AXP... IPSOUT is 5V on AXP? this is connected nearlly to half of pins of AXP
<granite_crusher>
*"AC" because it is already converted by adapter to DC
Black_Horseman has joined #linux-sunxi
<granite_crusher>
other DC-DC converters gets IPSOUT and converts it either to 5V, to 3.3V,1.5V
<oliv3r>
granite_crusher: yeah; could be that it does go through the DC-DC, but that should be controllerable if you can control the AXP
<oliv3r>
but that all I odn't know
docice has joined #linux-sunxi
<granite_crusher>
I cant understand what for is SY7208(SOT23-6), if it comes 5
<granite_crusher>
5V and goes out 5V
<granite_crusher>
it is always enabled
<granite_crusher>
because part of IPSOUT goes around that SY72...
<rellla>
oliv3r: do you have a kernel running already on cb3?
<oliv3r>
granite_crusher: stabilizer
<oliv3r>
rellla: yeap
<oliv3r>
rellla: mainline works beuatifully :)
<arokux1>
rellla: have you got your cb3?
<oliv3r>
once we figure out the name i'll push the dts
<rellla>
arokux1: yeap
<oliv3r>
rellla: but I did tell benn; only send it to rellla if he promisses to release a xbmc image soon
<oliv3r>
rellla: no pressure
<oliv3r>
:p
<granite_crusher>
oliv3r: so if board is off (EXTEN = 0; ) SY7208 do not stabilize, andwe get 5V unstable?
<oliv3r>
granite_crusher: i have absolutly no clue :p
<oliv3r>
granite_crusher: it's a boost regulator, it's generally used to step up 3.3 -> 5V
<rellla>
oliv3r: haha. pressure. what is pressure? not for spare-time-things ;) btw. VDR shows up some "picture" with vdpau. but only some funky screensaver with columns and rows :p
<rellla>
oliv3r: did you add been's patches for gmac etc.?
<oliv3r>
rellla: :p well vdpau is 'nice' but for later ;)
<oliv3r>
rellla: no, not yet
<oliv3r>
rellla: much to do sitll; adding 2GB support to u-boot atm
<rellla>
i'm building cubieboard2/linux-sunxi sunxi-3.4-ct-v101 atm
<rellla>
and then it's time to prepare a sd card for cb3...
<rellla>
arokux1: puh. don't remember. maybe i'm reading too much forums and chat-logs... maybe from cubieforum?
<rellla>
oliv3r: and for the xbmc-thing: i think it's better to invest the time in the open source way via libvdpau-sunxi and mainline xbmc... unfortunately i have no clue about vdpau for now :(
<arokux1>
they must have had a doc to add gmac support?
<arokux1>
the driver is developed by some Shuge
<arokux1>
shuge <shugeLinux@gmail.com>
<arokux1>
any idea who is this guy?
libv has quit [Ping timeout: 260 seconds]
<oliv3r>
arokux1: not supprising, CT is in fact, a20 device, so it's pretty much identical
shineworld has quit [Quit: Leaving]
<oliv3r>
rellla: but vpdau needs a lot of work still; having a working A20 xbmc image brings in potential developers
<oliv3r>
arokux1: that's AW employee; but we have gmac driver allready in a31-dev tree
<oliv3r>
a diff between those 2 is very interesting however
<arokux1>
oliv3r: interesting why he uses: <shugeLinux@gmail.com>
eebrah|noPC has joined #linux-sunxi
<oliv3r>
arokux1: yeah very strange, but you can grep for shuge in the aw tree's
<oliv3r>
arokux1: maybe he did this outside of aw
<oliv3r>
arokux1: OR its one of the guys who left AW to work for cubietech
Black_Horseman has quit [Remote host closed the connection]
<arokux1>
oliv3r: ah.. in a31-dev tree the author is also: shuge <shugeLinux@gmail.com>
<oliv3r>
then he still works for AW
<oliv3r>
i thin
libv has joined #linux-sunxi
<arokux1>
a31-dev should be present in our central tree, how do you think?
<oliv3r>
we only have sanitized properly gpl licensed code
<oliv3r>
amery's repo is pure for reference, hence the AW drops
<arokux1>
oliv3r: ah, so that is the feature?
<oliv3r>
yeah; the dumps end there ;)
<oliv3r>
mnemoc: said he would add the 3.4 bare import for now, but he's been awol for 2 days
vinifr has joined #linux-sunxi
granite_crusher has joined #linux-sunxi
<granite_crusher>
oliv3r: thanks, at last I got peace in my minds about circuit on that board :)
jemk has joined #linux-sunxi
<granite_crusher>
but I still do not undertand what for this 5V is needed while board is off if 5V do not get feedbacked to AXP?
JohnDoe_71Rus has quit [Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org]
<oliv3r>
granite_crusher: no clue :)
vinifr has quit [Ping timeout: 245 seconds]
<arokux1>
here is a diff between cb3 gmac and a31-dev gmac, lots of noise there, though: http://sprunge.us/QEEL?diff
<arokux1>
oliv3r, Tsvetan ^
vicenteH has quit [Ping timeout: 252 seconds]
<granite_crusher>
I got enlighted :) 5V is consumed by USB host aUSB otg, LCD... these have SY6280 chips which ether have EN pin atached to 3.3V or some signal going from GPIO to enble 5V....
eebrah|noPC has quit [Read error: Connection reset by peer]
<granite_crusher>
if this GPIO connector 5V rail vould have the same thing...
<oliv3r>
arokux1: oh he renamed it to sunxio; nice
<oliv3r>
ah some whitespace things
<oliv3r>
i think there's a diff option
<arokux1>
oliv3r: see the last link I sent
<oliv3r>
oh i used the first one
<mouchon2>
i finished v3 of my i2c patch before sending it i would like to be sure to do it correctly. So i make " git format-patch; scripts/checkpatch.pl 0001-V3-Add-i2c-sunxi-bus-frequency-as-param-to-i2c-sunxi.patch; " the next step should be " git send-email 0001-V3-Add-i2c-sunxi-bus-frequency-as-param-to-i2c-sunxi.patch" to dev@linux-sunxi.org that is the correct workflow ?
<oliv3r>
mouchon2: or just use your mail client
<oliv3r>
and attach the patch file ;)
<oliv3r>
but yeah that sounds about right
<arokux1>
oliv3r: ?????
<arokux1>
oliv3r: patches must be inlined, also sign-off tag is needed
<oliv3r>
i have checkpatch as a pre-commit hookl so when i commit something bad, git complains; i then just use git send-email -1 -s --annotate --coverletter
<oliv3r>
arokux1: sign off tag shoudl have been done on the commit ;)
<Tsvetan>
arokux1 the problem with A20 gmac is that pinmux is undocumented
<oliv3r>
arokux1: and a format-patch generated patch opens in clients normally
<oliv3r>
Tsvetan: not anymore
<oliv3r>
Tsvetan: the fex file has it :)
<Tsvetan>
oliv3r?
<Tsvetan>
no documentation
granite_crusher has quit [Quit: Page closed]
<arokux1>
Tsvetan: there are some other changes as you see in diff.
<DarkGamer0690>
can you direct me to an irc channel that handles rooting allwinner a1x android devices running 4.2.2
<oliv3r>
DarkGamer0690: doubt that exists; there's always #android for general android stuff, as that's what your asking, but i pretty much gave you everything you need
<oliv3r>
this is completly unrelated to allwinner's SoC btw, it's a pure android question :)
<DarkGamer0690>
it is saying that su needs to be updated
<oliv3r>
you though be able to almost copy paste that ;)
<oliv3r>
or atleast give you enough content to search with :)
<DarkGamer0690>
i try to install it and it says there was an error installing superuser.
<oliv3r>
DarkGamer0690: is your tablet in recovery mode?
<oliv3r>
e.g. CWM?
<DarkGamer0690>
no
<oliv3r>
i told you this 3 times allready, press the shut down button, say 'reboot -> recovery'
<DarkGamer0690>
it only has Android Recovery <3e>
<oliv3r>
or
<oliv3r>
do adb reboot recovery
<DarkGamer0690>
cwn isn't installed
<DarkGamer0690>
cwm*
<DarkGamer0690>
i'll be back soon
<oliv3r>
adb reboot recovery doesn't do anything?
rellla has joined #linux-sunxi
<oliv3r>
it should put the phone into recovery mode; doesn't matter what recovery mode it is, as long as you can do adb shell and get a root shell your good
vicenteH has joined #linux-sunxi
Soru has joined #linux-sunxi
docice has quit [Quit: docice]
vinifr has quit [Remote host closed the connection]
HeHoPMaJIeH has quit [Remote host closed the connection]
<arokux1>
ssvb: from the description it can run on battery. do you know which one?
zetaneta has quit [Read error: Connection reset by peer]
zetaneta has joined #linux-sunxi
massi has quit [Quit: Sto andando via]
rz2k has joined #linux-sunxi
zetaneta has quit [Quit: Leaving]
deasy has quit [Quit: Nom d'un quark, c'est Edmonton !]
geecko_ has joined #linux-sunxi
drachensun has quit [Ping timeout: 248 seconds]
deasy has joined #linux-sunxi
geecko_ has quit [Quit: Quitte]
arokux2 has joined #linux-sunxi
drachensun has joined #linux-sunxi
geecko_ has joined #linux-sunxi
panda84kde has quit [Quit: Konversation terminated!]
<oliv3r>
quiet day today
pingo has joined #linux-sunxi
geecko_ has quit [Read error: Connection reset by peer]
jemk has quit [Ping timeout: 264 seconds]
focus_it has joined #linux-sunxi
deasy has quit [Quit: Nom d'un quark, c'est Edmonton !]
<ssvb>
oliv3r, mnemoc: did you get your CT up and running?
<oliv3r>
running fine, having issues?
<oliv3r>
i posted the fex file and initial u-boot support
<oliv3r>
2GB support i'll push to a personal wip branch later
boycottg00gle has joined #linux-sunxi
eebrah|noPC has joined #linux-sunxi
<ssvb>
oliv3r: I just tried to boot the SD card taken from cubieboard2, and it booted, but the ethernet does not work right
<ssvb>
oliv3r: thanks, will try to check the updated u-boot and fex
zetaneta has joined #linux-sunxi
zetaneta has joined #linux-sunxi
zetaneta has quit [Changing host]
<popolon>
hi
<popolon>
do you now where is the G2D option in menuconfig ?
<popolon>
does virtio drivers and devfreq will works on an a7 ?
<popolon>
cotex-a7
<ssvb>
popolon: you need stage/sunxi-3.4 branch to enable g2d in sun7i hardware
<ssvb>
popolon: it is not in sunxi-3.4 yet
<popolon>
oh, ok, thanks
mcbrick has joined #linux-sunxi
iamfrankenstein has quit [Ping timeout: 240 seconds]
iamfrankenstein has joined #linux-sunxi
jemk has joined #linux-sunxi
boycottg00gle has quit [Remote host closed the connection]
<ssvb>
oliv3r: ok, updated u-boot and fex
<ssvb>
oliv3r: does emac/gmac work for you on CT?
<oliv3r>
i've only tried mainline emac driver, and that didn't work
<oliv3r>
it loaded fine, mii did detect changes to the cable, but i got timeouts
<ssvb>
ok, will try to use usb ethernet dongle for now :)
<oliv3r>
i read the datasheet for the rtl PHY, and i think it's only GMII and RGMII (e.g. both Gigabit mii) which isn't compatible with the MII interface of the emac
<oliv3r>
the gmac IS capable of servicing (r)mii and (r)gmii
<oliv3r>
ssvb: have you tried 3.4?
<oliv3r>
and why update the fex? i only added memory timings
<arokux2>
jukivili: I think those usb errors can be harmless, but not sure. investigating....
<oliv3r>
the fex file seems wrong
<oliv3r>
jemk ohh wait
<jukivili>
arokux2: I think I have seen similar errors with the integraded rtl8192cu with mk802, with v3.4 kernel
<jukivili>
arokux2: after unloading and reloading rtl8192cu module
<arokux2>
jukivili: yeah, such error appear in 3.4 too, but seldom.
<arokux2>
jukivili: my question to you will be why usb is so much faster in mainline
<oliv3r>
jemk: GNULL/ERXERR For example
<oliv3r>
PA9 is unconnected for gbit PHY
<oliv3r>
well i haven't wrapped my head around it all the way
<Turl>
oliv3r: maybe the phy supports both GMII and MII and those are the matching muxings for AW's favorite phy so one soldered chip works with both? :P
<oliv3r>
Turl: GMII is supposed to be backwards compatible according to jemk
<oliv3r>
i thought so too btw
<jukivili>
arokux2: small changes over long time.. what I find odd is that external device works without problems.. maybe integraded devices are wired somehow differently
<arokux2>
jukivili: maybe I'm configuring it incorrectly. did you use nl80211?
<Turl>
jukivili: are you connecting it directly to the device or is a hub in the middle?
<Turl>
(internal or external)
<arokux2>
jukivili: it works now!
<arokux2>
it fucking works!
<oliv3r>
what I think, for example that RXERR and TXERR where completly dropped from the GMII spec; hence, GNULL for gmac
<jukivili>
arokux2: yes.. however wext / nl80211 should not matter. The wext compatibility layer within kernel uses cfg80211/nl80211.
<jukivili>
Turl: directly
<jukivili>
arokux2: ok ok :)
<arokux2>
jukivili: ppl from #linux-wireless recommend to use nl80211.
<arokux2>
jukivili: i'll try both
<arokux2>
jukivili: so why do you think usb transfers got faster?
<oliv3r>
jemk: i don't quite understand the swap of pa14 and pa15
<oliv3r>
pa14 is GNULL/ETXCK; PA15 is GTXCK/ECRS
<arokux2>
jukivili: I know exactly what part causes problems
<oliv3r>
3.3 uses the gmac driver and it appears to work normally
techn__ has joined #linux-sunxi
<popolon>
ssvb, oh, ok, didn't use the good command to obtain the stage branche then
techn_ has quit [Ping timeout: 246 seconds]
jinzo has quit [Quit: Leaving]
deasy has quit [Quit: Nom d'un quark, c'est Edmonton !]
iamfrankenstein has quit [Ping timeout: 252 seconds]
<oliv3r>
Turl: around still?
<Turl>
oliv3r: sup
wingrime has quit [Ping timeout: 245 seconds]
<arokux2>
Turl: wlan adapter works ;)
<arokux2>
Turl: worked all the time actually...
<Turl>
arokux2: how come?
<oliv3r>
Turl: ok, for dts and u-boot naming, cubietruck or cubieboard3
<arokux2>
oliv3r: CT is supposed to run on battery?
<oliv3r>
Turl: the box says 'cubieboard3' and below cubietruck
<oliv3r>
arokux2: no, but it potentially can
<oliv3r>
arokux2: but so can olimexino a20
<oliv3r>
cb1 can't, they didn't add the pins, but in theory it can
discopig has quit [Read error: Connection reset by peer]
<arokux2>
oliv3r: yeah, my question if the same type of the battery will do.
<Turl>
oliv3r: ask benn what name does he prefer? :)
<oliv3r>
with a lipo batt, even a small one, you have a UPS :)
<oliv3r>
LiPo yep
<oliv3r>
Turl: he's gone till the 8th or 9th
<arokux2>
oliv3r: 3.7V? same connector?
<oliv3r>
arokux2: yep
discopig has joined #linux-sunxi
<Turl>
oliv3r: ah right, there's some kind of chinese holidays these days
<oliv3r>
Turl: i thought it was a new 'branch' but it actually is, cubieboard3
<oliv3r>
so title:subtitle kind of thing
<Turl>
well, we can always rename it I suppose :)
<oliv3r>
Turl: i think of renaming it to cubieboard3
<Turl>
cb3 should be ok
<arokux2>
Turl: "link is not ready" was just some IPv6 stuff, those usb errors although present but seem to be harmless... I'm using wifi module with myrootfs.. so it works.
<arokux2>
oliv3r: think about users
<arokux2>
oliv3r: they buy a cubietruck...
<arokux2>
do you want to explain to each of them that they really need a cb3 dts?
<oliv3r>
they buy a product, that works; they don't hack with stuff ;)
rz2k has quit []
<oliv3r>
with that reasoning, how will they even find the git tree, the gittree is called cubieboard2
<arokux2>
Turl: I've spotted another issue though. 8192cu (realtek's driver) is 2x faster then in kernel one: tested with sunxi-3.4
<oliv3r>
also, we can always add a cubietruck dts that only includes cubieboard3, like an alias :)
<arokux2>
Turl: in mainline rtl8192cu is as fast as in sunxi-3.4, that means still 2x slower
FDCX has joined #linux-sunxi
<Turl>
oliv3r: you got mainline working already? :)
<Turl>
with all leds? :D
<oliv3r>
Turl: yep
<Turl>
and 2G ram? :P
<oliv3r>
yep
<oliv3r>
i pushed the u-boot patches allready
<Turl>
emac?
* Turl
expects no
<oliv3r>
well the 1gb version; 2gb version is pushed to my wip branch
<oliv3r>
emac 'works' but i get no ip
<oliv3r>
and tx timeouts
<oliv3r>
so either emac driver doesn't work with this PHY
<oliv3r>
or needs some sprinkles
<Turl>
on linux or uboot?
<oliv3r>
but other then that, it does 'work'
<oliv3r>
oh u-boot i haven't tested
<arokux2>
Turl: ok, maybe the difference is 1.5x
<Turl>
oliv3r: what is cubietech shipping for ethernet?
<oliv3r>
rtl8211e
<Turl>
oliv3r: I mean driver :P
<oliv3r>
3.3 gmac
<Turl>
is the driver too different from emac?
<oliv3r>
we only have 3.4 gmac
<oliv3r>
or 3.3 sun6i
<oliv3r>
gotta look still; but shouldn't be hugely diff
<oliv3r>
i'll go read the rtl manaul more to see if i spot anything interesting with regards to regular MII
<oliv3r>
mripard: what do you prefer for the name, cubieboard3 or cubietruck. The title is cubieboard3 with a subtitle of cubietruck
<Turl>
oliv3r: don't you need the gateway too?
<hno>
oliv3r, I suspect that my cubietruck is waiting at the post office. haven't been able to collect the package yet.
<hno>
arokux1, next week is still crazy here.
<arokux2>
hno: I've got everything working with usb
<hno>
arokux2, excellent!
<arokux2>
hno: the in-kernel driver performs badly as compared to realtek's though :(
<arokux2>
1.5x difference.
<hno>
realtek? wireless?
<arokux2>
hno: yeah.
<arokux2>
hno: this comparison is in 3.4. but nothing changes in mainline. rtl8192cu shows same speed - around 600-800KB/s
<arokux2>
hno: 8192cu - 1.2MB/s and a bit more sometimes.
<arokux2>
it is really strange, because the code of 8192cu is open for ages.
<hno>
any difference in firmware used perhaps?
<arokux2>
hno: no, it is the same.
<hno>
ok. odd then I think.
<arokux2>
hno: I'm going to report it to linux-wireless. this is just a byproduct of my usb hacking. however I'd like wlan adapter to perform fast, since I'm going to use mele as "production" machine
<hno>
goodnight. Time to get some hours of sleep.
<arokux2>
goodnight hno
<arokux2>
oliv3r: you there?
deasy has joined #linux-sunxi
arokux2 has quit [Remote host closed the connection]