hno changed the topic of #linux-sunxi to: /Allwinner/sunxi development discussion - Don't ask to ask. Just ask and wait! - See | | Logs at
\\Mr_C\\ has joined #linux-sunxi
mateobur has joined #linux-sunxi
<mateobur> My cubieborad gives
<mateobur> SDRAM initialization fail
<mateobur> is anything I can do?
<deasy> mateobur, it's with the original uboot from the nand?
<mateobur> yes
<mateobur> it stopped working
<mateobur> second day of use
<woprr> got network back using the olimex A20 micro debian card fex, hid_sunplus working too
<woprr> now fixing (hdmi-)sound
<woprr> got console on hdmi too
<deasy> mateobur, :/ bad bad hope it not happens here
<deasy> mateobur, contact where you have buy it
<mateobur> I tried LiveSuit, but nothing
<woprr> thats the reason I bought without built in nand flash, a sata SSD lasted only a year here
<deasy> hmm woprr you say it for mateobur ?
<woprr> but maybe uboot sdram config error
<woprr> pardon?
<deasy> if he have not touch it it can't be...
* woprr gets back to work
<deasy> woprr, use your warranty for a normal use
<deasy> it's a deffect unit
<deasy> or you write lot of little file
<deasy> dev?
<hno> SDRAM is not NAND. It is the RAM.
<woprr> deasy, I don't know which warranty law applies, cubieboard is chinese, and there are different warranty laws for consumers and practionieers...
<deasy> i talk about your ssd
<deasy> i have buy mine in Germany
<woprr> and those are all evaluation board, so consumer warranty may not apply
<hno> mateobur, try if it works to livesuit without the UART cable attached.
<woprr> deasy, I've got a full refund from OCZ ;-)
<deasy> LOL okay in understand
<deasy> SSD ocz are crap
<deasy> i*
<woprr> ...after "analyzing" their SMART support, etc in their forums ^^
<deasy> some have almost 50% of return
<woprr> well SSD is a chip memory experienced, WD is not...
<woprr> -SSD +OCZ
<woprr> anyway, must fix audio now, bb later or tomorrow
<deasy> crucial intel corsair or some sandisk are okay
<deasy> but stay way of ocz
<deasy> far*
<deasy> octane and petrol ocz have 30-40% return
<deasy> worst in the last 52% oh my !
<deasy> "those are all evaluation board" even the final version?
<deasy> that's sell like normal product
mateobur has quit [Quit: Saliendo]
<woprr> SSD wear off = Floppy Disc :D
<woprr> or nand, SD ...
ganbold has joined #linux-sunxi
egbert has quit [Disconnected by services]
egbert has joined #linux-sunxi
<deasy> woprr, how they can do warranty for life :/
hipboi_ has joined #linux-sunxi
_hipboi_ has quit [Ping timeout: 264 seconds]
BJfreeman has quit [Read error: Connection reset by peer]
deasy has quit [Quit: Quitte]
bamvor has joined #linux-sunxi
bsdfox has quit [Ping timeout: 256 seconds]
\\Mr_C\\ has quit []
xtofury has joined #linux-sunxi
rz2k has quit []
<xtofury> /msg NickServ identify rdjedj
\\Mr_C\\ has joined #linux-sunxi
[7] has quit [Disconnected by services]
TheSeven has joined #linux-sunxi
xtofury has quit []
hipboi_ has quit [Remote host closed the connection]
hipboi has joined #linux-sunxi
xtofury has joined #linux-sunxi
<xtofury> ok I managed to get raw disk access to my sdcard but cannot figure out the section here
<xtofury> I ran this, it worked ok dd if=/dev/sde1 of=/dev/mmcblkX bs=512 count=2047
<xtofury> there's sdb which appears the same as sde1, this is strange and confusing... when i did this sdb said unallocated...sde1 had a partition that occupied nearly the whole card... what gives?
bamvor has quit [Ping timeout: 256 seconds]
bamvor has joined #linux-sunxi
xtofury has quit [Ping timeout: 264 seconds]
Olaff has joined #linux-sunxi
\\Mr_C\\ has quit []
jacq has joined #linux-sunxi
wingrime has joined #linux-sunxi
wingrime has quit [Quit: Lost terminal]
\\Mr_C\\ has joined #linux-sunxi
mcbrick has quit [Quit: Leaving]
\\Mr_C\\ has quit []
jacq has quit [Ping timeout: 240 seconds]
\\Mr_C\\ has joined #linux-sunxi
rellla has joined #linux-sunxi
forcev is now known as Funkypenguin
Funkypenguin has quit [Changing host]
Funkypenguin has joined #linux-sunxi
FR^2 has joined #linux-sunxi
massi has joined #linux-sunxi
panda84kde has joined #linux-sunxi
vicenteH has quit [Ping timeout: 264 seconds]
navlrac has quit [Ping timeout: 245 seconds]
diego_at_work has joined #linux-sunxi
n01 has joined #linux-sunxi
n01 has joined #linux-sunxi
notmart has joined #linux-sunxi
arokux has joined #linux-sunxi
<arokux> hi, any idea who Jussi Kivilinna is?
e-ndy has joined #linux-sunxi
dlan has quit [Ping timeout: 240 seconds]
dlan has joined #linux-sunxi
<mnemoc> arokux: jukivili
<arokux> mnemoc, thx
<arokux> jukivili, ping :)
jacq has joined #linux-sunxi
<Seppoz> hi all
<wigyori> morning
n01 has quit [Read error: Operation timed out]
<arokux> hi Seppoz
<Seppoz> hey :)
<Seppoz> did you manage to get USB working?
<arokux> Seppoz, no)
<Seppoz> hmmm
tzafrir has quit [Ping timeout: 260 seconds]
adrian has joined #linux-sunxi
vicenteH has joined #linux-sunxi
arokux has quit [Ping timeout: 264 seconds]
<jukivili> arokux: hello... (too slow)
shineworld has joined #linux-sunxi
n01 has joined #linux-sunxi
<arokux1> jukivili, still there?
<arokux1> jukivili, I've commented on your patches
<jukivili> hello
<arokux1> jukivili, let's mainline ohci/echi support, I've done a lot for that already, but something is missing and it is not working yet.
<arokux1> jukivili, if you are interested into looking into it, I could bring it into nice form and push to github or so
<jukivili> I checked your patch, it looks ok (except the part that Hans commented). Will you make v2 of it?
<arokux1> jukivili, yes, it will include more cleanups too.
<jukivili> Ok, I'll rebase my patch set on your patch later
<arokux1> jukivili, what about mainlining?
jacq|2 has joined #linux-sunxi
FRQuadrat has joined #linux-sunxi
shineworld has quit [Ping timeout: 240 seconds]
jacq has quit [Ping timeout: 240 seconds]
Turl has quit [Ping timeout: 240 seconds]
FR^2 has quit [Ping timeout: 240 seconds]
Turl has joined #linux-sunxi
<jukivili> does ehci require dma?
fluo75 has quit [Read error: No buffer space available]
fluo75 has joined #linux-sunxi
<arokux1> jukivili, yes, I think, there is dma code
fluo75|2 has joined #linux-sunxi
fluo75|3 has joined #linux-sunxi
fluo75 has quit [Ping timeout: 256 seconds]
<jukivili> well.. I'm working on with musb/otg driver for sunxi now, and have other stuff wip too.. I don't think I have time to start working something new atm
fluo75|2 has quit [Ping timeout: 276 seconds]
zerodamage_ has joined #linux-sunxi
<arokux1> jukivili, on mainlining of musb/otg you mean?
tzafrir has joined #linux-sunxi
<arokux1> jukivili, (please use my name so I get pinged)
<zerodamage_> anyone with 1-hand experience in lvds lcds for cubieboard?
<Seppoz> zerodamage_: are you sure its lvds anyways
<jukivili> arokux1: trying to get musb work with sunxi-3.4 tree
<arokux1> jukivili, I see, I thought it was working already?
<zerodamage_> Seppoz: yes it an chimei innolux N101BGE-L21
<zerodamage_> Seppoz: have used it with an LVDS driver before
zerodamage has quit [Ping timeout: 250 seconds]
<Seppoz> no only ttl
<zerodamage_> ok
<Seppoz> and i havent even played with the ttl yet
<zerodamage_> when i use the default fex, it shows mode 800x480 for fb1
<jukivili> arokux1: I mean the musb driver in mainline that AW copy-pasted and mangled into their own OTG driver
<arokux1> jukivili, I see. what you say is sounds like musb is not yet working in sunxi-3.4. in what version it has worked then?
<jukivili> arokux1: adding sunxi support to mainline musb would allow otg support to be later added to mainline
<arokux1> jukivili, ah, you want to *adjust* the mainline 3.4 musb, so that it supports sunxi hardware too?
<jukivili> arokux1: yes :)
<arokux1> jukivili, ok, I see. why not adding it to mainline 3.11?
<jukivili> arokux1: it's not quite working yet
<wigyori> arokux1: for yesterday's question about the dmesg, i've tried a couple things to no luck, and written it down at . do you have a minute to take a look at it, to see if i'm screwing up something obvious?
<zerodamage_> Seppoz: ok, adding my resolution is what f*** it up
<zerodamage_> Seppoz: ie causes it to reboot forever
<Seppoz> which resolution did you add
<zerodamage_> 1366x768 to script.fex
<Seppoz> oh i see
<Seppoz> maybt they dont set the framebuffer correct
<Seppoz> i meant he kernel
<arokux1> jukivili, ok, so I'll continue to work on usb1 and 2 on my own then. I'll report my work to the list, so that we don't get into conflicts.
<zerodamage_> quite possible i guess
<zerodamage_> maybe i can find an 800x480 lvds screen i my stash, and try that instead
<jukivili> arokux1: ok.. I actually had made those patches months ago, just didn't send them off back then
<arokux1> jukivili, yes, it's a pity, I've seen it on you github today. so you should also post something as soon as it's done :)
<Seppoz> zerodamage_ : what happens if you set your resolution?
<zerodamage_> Seppoz: with fbset?
<Seppoz> no in script.bin
<Seppoz> i doubt you can use fbset with this cpu
<arokux1> jukivili, I'm thinking, taking into account you work on musb, maybe it'd be the best for me to take care of your hci patches. i.e. put them on top of mine and submit everything in one series.
<zerodamage_> it boots ok, with a different resolution under lcd0_para, but it does not change /deb/fb1
<zerodamage_> always reported as 800x480
<Seppoz> hmm
<Seppoz> and you sure your doing the script.bin stuff right?
<zerodamage_> i have fex2bin and back to make sure it is actually done, and it does change behavior if i for example change the disp_mode from hdmi to lcd
<zerodamage_> but the frame buffer does not follow the script.bin, maybe my kernel is too old or has a wrong boot line?
<Seppoz> which one are you using?
<zerodamage_> uname -r 3.0.76
<Seppoz> try modifying the driver in video/sunxi/lcd/lcd_panel_cf.c
<Seppoz> maybt its not using the config
<zerodamage_> then i have to rebuild kernel from scratch?
<Seppoz> ye
<arokux1> wigyori, I do not know much. but why are you using old u-boot with mripards tree? if you want to boot a feature full kernel you'd better get sunxi-3.4? (consider this info could be wrong)
<Seppoz> zerodamage_: there are certain configs in this file, no idea how they are beeing used
<Seppoz> i mean the folder
<wigyori> arokux1: i'm rather into getting mainline up and running, and not necessarily a full-featured kernel
<arokux1> wigyori, I see, then go for new uboot too.
<zerodamage_> Seppoz: yes i have seen them, but did believe it loaded preferentialy from script.bin, maybe it only works for some because the resolution already matches....
<wigyori> arokux1: old u-boot? 2013.07-rc1 counts as old?
<arokux1> wigyori, this is bleeding edge:
<wigyori> or should i go for trunk
<arokux1> wigyori, well, that's upstream uboot
<arokux1> wigyori, the one I've shown to you has more on sunxi
Black_Horseman has quit [Quit: Zwi se logou mou!!!]
<wigyori> ok, i'll give it a shot
<arokux1> wigyori, other than that you'd better find somebody who has the hardware. I own mele a1000 (sun4i, A10).
<wigyori> :)
<wigyori> thanks for your help
tinti has quit [Quit: Leaving]
<arokux1> wigyori, you're welcome!
<Seppoz> zerodamage_: i cant find any trace in the linux 3.4 source that would say so
<Seppoz> all they read is the lcd spi stuff
<Seppoz> but i might just not see it
<arokux1> wigyori, btw, here are the instructions to build uboot:
tinti has joined #linux-sunxi
<zerodamage_> lcd spi?
<zerodamage_> as in EDID?
<Seppoz> no idea
<zerodamage_> can you give a link to the source where you found this?
<Seppoz> its in the 3.4 branch of linux sunxi
<zerodamage_> ok
<Seppoz> infact
<Seppoz> its in the disp_lcd
<Seppoz> so yea it probably uses the config
<zerodamage_> folder?
<Seppoz> video/sunxi/disp
<zerodamage_> thanks, dont know so much of the structure
<Seppoz> ret = script_parser_fetch(primary_key, "lcd_x", &value, 1);
<zerodamage_> yes i see, its very strange that fb1 does not follow then....?
<Seppoz> if you replace all the DE_WRN with prontk you should get alot of debus output
<Seppoz> including all the settings in the script file
<zerodamage_> yes
<zerodamage_> but compiling the kernel is a bit of a task, i have never done this before.... :(
<Seppoz> its quite easy want me to upload the procedure?
<zerodamage_> yes if you could that would be nice
<zerodamage_> is the sunxi-disp not a module anyway?
<Seppoz> based on the defconfig no
<zerodamage_> ok
<arokux1> jukivili, if this is fine for you, you just need to resubmit your patches with my comments taken into account.
<arokux1> jukivili, comments on the mailing list.
<zerodamage_> ty
<zerodamage_> build-essentials and gcc is enough?
<Seppoz> first of yll you want to apt-get install gcc-4.6-arm-linux-gnueabi ncurses-dev uboot-mkimage build-essential git
<zerodamage_> ok
<zerodamage_> answers that :)
<arokux1> Seppoz, there is sunxi-bsp, that does everything for you
<Seppoz> zerodamage_^^^^^^^^^^^^^^^^^^^^^^^^
<zerodamage_> ack
<Seppoz> i prefere building it myself tho
<mnemoc> the bsp doesn't install packages for you
<Seppoz> you can also use that one hf toolchain if you want
<Seppoz> no idea that the name was
<arokux1> Seppoz, ok, just wanted to let you know.
<Seppoz> just change the CROSS_COMPILE to whatever you want to use
<zerodamage_> ack
<arokux1> mnemoc, they belong into rootfs which is a parameter or?
<Seppoz> zerodamage_: the resulting uImage will be in kernel_bin/arch/arm/boot/uimage
<zerodamage_> y thanks
<mnemoc> arokux1: rootfs is about the target, apt-get install is about the host/builder
<mnemoc> the bsp can not assume any particular host/builder OS
<mnemoc> so it can't install packages
<zerodamage_> silly question, but cant i get the DE_WRN out somehow, its it an internal thing?
<arokux1> mnemoc, I mean you'd need to create rootfs first, and then supply it to bsp, or?
<mnemoc> yes
<mnemoc> but you can get rootfs tarballs of different distributions (or make your own) trivially
<zerodamage_> anyway, printk goes to dmesg or where?
<arokux1> Seppoz, make -j16 <--- nice hardware you've got.. 16 cores?!
<Seppoz> 24 actually
<zerodamage_> E P I C
<zerodamage_> well that may not run overhere then on an FX4100
<arokux1> mnemoc, btw, do you know what causes doubling of the kernel bootlog in 3.4?
<zerodamage_> maybe i will try a 3.4 kernel also?
<arokux1> zerodamage_, "The AMD FX-4100 is a quad-core processor is basically a FX-8150 with 4-cores disabled. " :)
<zerodamage_> yup
<mnemoc> arokux1: debug_ll
<zerodamage_> so -j4 ?
<zerodamage_> :P
<arokux1> zerodamage_, -j5, num_cpu + 1
<zerodamage_> kk
* mnemoc does 2*num_core :p
<zerodamage_> hot
<arokux1> zerodamage_, some say num_cpu +2 even. you can test
<Seppoz> its 2 cpus.. 2x 6 cores whith ht
<Seppoz> it only appears as 24
<jelly-home> arokux1: +1 depends on the cpu scheduler and cpu and kernel version... BFS author recommends exactly the number of cores
<jelly-home> (when bfs is in use)
<zerodamage_> if i do -j100 it will get the rest from "the cloud" right?
<arokux1> jelly-home, oh, ok, complicated. hno ---^
<Seppoz> -j is not only the number of cores but the number of processes used
<zerodamage_> surely today this the way it functions
<arokux1> Seppoz, -j is only the number of processes created.
<Seppoz> ye
<arokux1> mnemoc, if DEBUG_LL is on all the messages get duplicated?
<zerodamage_> i will go do manual work outside now, maybe I will have the guts to do the kernel later
eebrah|away is now known as eebrah
<arokux1> zerodamage_, what do you mean?
<arokux1> Seppoz, which motherboard are you using?
<zerodamage_> building kernels has been a bad experience for me, and this lcd buisness is draining my life-force, so i will go work on our boat outside
<zerodamage_> thanks for all the tips and cya later
<Seppoz> arokux1: i cant really remeber, want me to find out
<arokux1> zerodamage_, ;) consider this process as fun
<arokux1> Seppoz, no
<Seppoz> ok
<zerodamage_> sudo halt zerodamage
<wigyori> sudo make sandwich ;p
<Seppoz> sudo go shopping
deasy has joined #linux-sunxi
<n01> zerodamage_: worst password ever :)
<Seppoz> would be fun if wifes name is sudo xD
<Turl> morning
<arokux1> Turl, hi
Soru has quit [Ping timeout: 260 seconds]
<arokux1> Turl, I just want to remember that we wanted to clarify what "reset" means in clock context. mturquette was supposed to provide us a clue
FRQuadrat is now known as FR^2
<Turl> arokux1: yes (was that a question?)
<arokux1> Turl, no, just a reminder and pinging of mturquette :)
Robs58 has quit [Ping timeout: 245 seconds]
Falcon1 has joined #linux-sunxi
Falcon1 has left #linux-sunxi [#linux-sunxi]
_whitelogger has joined #linux-sunxi
ganbold_ has joined #linux-sunxi
heffer has joined #linux-sunxi
heffer has quit [Changing host]
heffer has joined #linux-sunxi
<arokux1> is any of you aware of some hardware upgrade subscription service? i.e. to upgrade hardware regularly for some fee.
<Turl> arokux1: some US carriers are starting to do that
<arokux1> Turl, can you supply a web link, plz?
vinifr has joined #linux-sunxi
fredy has quit [Excess Flood]
<arokux1> Turl, no, I've meant PC upgrades, like new CPU and motherboard
<Turl> ah, no idea then
fredy has joined #linux-sunxi
<Turl> you said hardware on an embedded channel, I imagined phones and the like ;)
<adrian> i need some help
<adrian> i try booting from nfs server
<adrian> ADDRCONF(NETDEV_CHANGE): eth0: link becomes ready
<adrian> [ 6.557232] IP-Config: Guessing netmask
<adrian> [ 6.562207] IP-Config: Gateway not on directly connected network.
<adrian> [ 13.090480] [AXP]init charger->rest_vol = 100
<adrian> [ 101.617679] VFS: Unable to mount root fs via NFS, trying floppy.
<adrian> [ 101.624336] VFS: Cannot open root device "nfs" or unknown-block(2,0)
<adrian> i build the kernel with CONFIG_ROOT_NFS=y and CONFIG_IP_PNP=y
<Turl> adrian: I'd review the network and nfs parameters you're passing to the kernel
<adrian> you refer at the env from u-boot?
<adrian> bootargs=console=ttyS0,115200 root=/dev/nfs nfsroot=${serverip}:/home/nfs/arch ip=${ipaddr}:${netmask}:${gatewayip}:::eth0
<adrian> bootcmd=nfs 0x43000000 ${serverip}:/home/nfs/arch/script.bin; nfs 0x48000000 ${serverip}:/home/nfs/arch/uImage; bootm 0x48000000
n01 has quit [Ping timeout: 245 seconds]
<vinifr> adrian, i guess uboot-sunxi does support nfs boot
<adrian> if i set the rootfs to mmc is working
<vinifr> ...does not support...
<mnemoc> nfs is not bootloader's issue
<adrian> the kernel loads well from nfs
<arokux1> adrian, you supply netmask, but it says it's guessing it...
<adrian> also if i boot same enviroment but setting the rootfs to the mmc , i see no /dev/nfs device
<Turl> google says ip=<client-IP-number>::<gateway-IP-number>:<netmask>:<client-
<Turl> hostname>:eth0:off
<Turl> you seem to have a different order
<adrian> il try it
<arokux1> adrian, tell us if it worked. you could update our wiki then.
<adrian> ok
<Turl> adrian: here is the full list and explanation for each of them
<vinifr> does uboot-sunxi support .dtb load?
<ganbold_> vinifr: it seems so, you can check u-boot readme
<Turl> vinifr: yes
<jukivili> arokux1: sorry, I had to leave for a moment.. I can send new patches with the fixes
<Turl> vinifr: you can load anything really, even mp3 tracks :p
<vinifr> Turl, nice :)
<arokux1> jukivili, ok, do it please. I'll then add your patches to mine and post as one series.
Robs58 has joined #linux-sunxi
<mnemoc> vinifr: the sunxi-current branch does. not sure if that was merged already into the sunxi branch
<vinifr> i will test
<arokux1> vinifr, I do it with fatload and bootm commands all the time
wingrime has joined #linux-sunxi
<vinifr> arokux1, what do you branch you use?
shineworld has joined #linux-sunxi
<adrian> not working, i am recompiling the kernel maybe i did something wrong
adrian has quit [Remote host closed the connection]
adrian has joined #linux-sunxi
shineworld has quit [Remote host closed the connection]
<adrian> same problem, after recompiling the kernel...
<Turl> maybe ssvb can help
<Turl> he was running his systems off nfs
vinifr has quit [Quit: Saindo]
<ssvb> adrian: mine cmdline looks more like "ip=${ipaddr}:${serverip}:${gatewayip}:${netmask}::eth0"
<ssvb> adrian: which seems to be closer to what is described at
<ssvb> adrian: also check that your nfs server is really properly configured, up and running
shineworld has joined #linux-sunxi
<adrian> if i set it like this i get IP-Config: Unable to set interface netmask (-22).
<adrian> bootargs 'console=ttyS0,115200 root=/dev/nfs nfsroot=${serverip}:/home/nfs/arch ip=dhcp'
<ssvb> are these variables properly expanded to ip addresses?
<adrian> this one enable the network
<adrian> [ 11.087115] IP-Config: Got DHCP answer from, my address is
<adrian> [ 11.095738] IP-Config: Complete:
<adrian> [ 11.106193] host=, domain=, nis-domain=(none),
<adrian> [ 11.099018] device=eth0, addr=, mask=, gw=,
<adrian> [ 11.111965] bootserver=, rootserver=, rootpath=
shineworld has quit [Quit: Leaving]
<ssvb> does it now work correctly with "ip=dhcp"?
<adrian> no
<adrian> just hangs, i think is becouse if the dhcp server is not the nfs server...
<adrian> i must set ip manual
n01 has joined #linux-sunxi
<adrian> is there a way to check in the c code of the kernel where is this order of ip setup
<Seppoz> adrian: maybe your udhcp is started as daemon?
<Seppoz> ip address is set by init afaik
<Seppoz> not kernel
<Seppoz> do you need to use uboot to set ip?
<adrian> yes
<Seppoz> can you just use the network scripts?
<adrian> actualy i want to be able to set the mac address from u-boot
<adrian> or from uEnv.txt
<Seppoz> mac address is read from script.bin afaik
<adrian> on this kernel i discovered a typo
<Seppoz> cant you just make a initrd script that reads uEnv?
<Seppoz> and let the distro set it
<Seppoz> as first step of init
<Seppoz> like this
<Seppoz> just use the file instead of eeprom
<adrian> if(SCIRPT_ITEM_VALUE_TYPE_STR != script_get_item("dynameic", "MAC", &emac_mac)){
<adrian> here is the typo in drivers/net/ethernet/sun7i/sun7i_wemac.c
<arokux1> adrian, :) funny. should it be dynamic?
<adrian> yes
<adrian> this read from fex file i think
<arokux1> yes
<arokux1> script_get_item, for script as in script.bin
<adrian> is there a function like this to take from u-boot env?
<arokux1> mnemoc could know
<Seppoz> there is bootot-env tools
<Seppoz> *uboot
<Seppoz> for distro
<Seppoz> you could use those to read it
<Seppoz> and you can allways make your own kernel argument
<Seppoz> and pass it to the kernel from uboot
<Seppoz> and use this instead of the script one
<Seppoz> and the function you showed
<Seppoz> *at
<adrian> i dont know exactly how to make one argument, that would be great btw
<arokux1> there should exist some solution for this already..
<Seppoz> its quite easy:
<Seppoz> just use the function __setup("ethmac=",set_mac); and use static int __init set_mac(char *str)
<Seppoz> here is a dm9000 example
tat has joined #linux-sunxi
<Seppoz> adrian: you might as well just set it in uboot
<Seppoz> adrian: and remove the new setting from the kernel
<Seppoz> maybe it survives reinit from the kernel
<adrian> int that file i have __setup("mac_addr=", set_mac_addr);
<tat> hello i have a the device listed under, mine has an additional cammera but i guess it is the more or les the same, i need the FTDI driver for that device, my kernel version is Linux version 3.0.8+ (root@tom-desktop), unfortunately there is no /proc/config.gz , so i wan't to try to just compile a module with a common config and see if it loads, what kerenl source would be best to use ?
<adrian> mac_addr i used in the uEnv to set mac address
<tat> ah don't forget: it is an android kernel
<Seppoz> adrian: you need to pass mac_addr= as kernela rgument
<adrian> i put it in extraargs
<arokux1> tat, as a wild guess, sunxi-3.0 branch of this:
<arokux1> tat, please report if it worked.
<arokux1> mnemoc, what is the reason to keep 3.0.x kernel?
<mnemoc> 3.4 is preferred these days
<mnemoc> but afaik no one has cared to test it with android yet
<mnemoc> from the current stage/3.4 nothing has been backported to 3.0..... implicitly deprecating it
<arokux1> mnemoc, so 3.4 could power android and X-based systems?
<mnemoc> in theory
<arokux1> mnemoc, can mainline power android (assuming all the soc drivers in place)?
<mnemoc> our local Android expert (aka Turl) doesn't seem to care about it anymore....
<arokux1> mnemoc, he's busy with mainlining
<mnemoc> arokux1: mainline doesn't have any storage driver
<mnemoc> eventually, sure
<mnemoc> today. no
<Turl> mnemoc: :P I can still advise though, and with a bit of time I might retake it when cb2 arrives
<mnemoc> mainline doesn't have a disp driver for sunxi either.... so pretty useless
<arokux1> mnemoc, yes, but assuming there were storage driver. can android run on top of mainline kernel?
<mnemoc> arokux1: no display either
<Turl> arokux1: you'll need to patch it a bit to get wakelocks, binder, paranoid networking et al but I don't see why not
<arokux1> mnemoc, i thought android needs some special features, added by google.
<mnemoc> once all drivers gets cleaned, unifed and mainlined, android will too :p
<arokux1> ok, so some patching is still required. there was an effort to mainline android kernel. so it means it was not 100% finished...
<tat> arokux1: so how would i go on without a config.gz, is there a common config for the A13 ?
<arokux1> tat, why do you need a full config, if you are going to build one module only?
<mnemoc> arokux1: by the time sunxi is fully mainlined, it's most likely android is (mostly) mainlined too
<tat> arokux1: so i just go into the subdirectory and say "make MODULE", is that all ?
<arokux1> mnemoc, ah, so android is mainlined as we speak..
<mnemoc> arokux1: android is not publicly developed
<arokux1> tat, yes, something like this, you'd need to google for a more exact command. it probably needs CROSS_COMP... too etc.
<mnemoc> arokux1: but they added branches closer to mainline
<mnemoc> arokux1: which makes one hope they are indeed prepering their stuff for submitting it
<tat> arokux1: ok i was sure about the arm-cross but well i thought i need a working config for the A13 first
<mnemoc> but unless you are G's CEO, probably no one knows
<arokux1> mnemoc, I have friends who work in google, but not on android kernel stuff..
<arokux1> mnemoc, another question. there is dma driver missing. a lot of stuff depends on it. is there somebody working on dma?
<jukivili> arokux1: do I send patches to mailing-list or to you directly?
<Turl> arokux1: wingrime
<arokux1> jukivili, to mailing list.
<arokux1> Turl, hm, i thought he'll do SATA
<Turl> arokux1: and IR ;)
<arokux1> Turl, that one is finished now (except of some clock issue)
<Turl> arokux1: I know
bsdfox has joined #linux-sunxi
<arokux1> mnemoc, I've got hit by a spam filter :(
<mnemoc> only jukivili's posts there
<Turl> arokux1: try again now
<mnemoc> arokux1: ah, wiki
<arokux1> Turl, works.
<mnemoc> arokux1: I thought you were talking abot ML
n01 has quit [Ping timeout: 264 seconds]
<Turl> mnemoc: wiki spam has slowed down, maybe it's time to disable it
n01 has joined #linux-sunxi
diego_at_work has quit [Quit: Page closed]
<wingrime> Turl: have we any docs about SATA
<wingrime> Turl: Without it I forced do unknown initialization sequence
<wingrime> Turl: register name in driver
<wingrime> *names
<wingrime> Turl: but bis not known
vicenteH has joined #linux-sunxi
\\Mr_C\\ has quit []
jacq|2 has quit [Ping timeout: 276 seconds]
<Turl> wingrime: p822 on A20 user manual, but no info, just simple description
<wingrime> Turl: I have init and release configuration (sunxi on)
<wingrime> Turl: some phy code
<wingrime> Turl: realy simple
rellla has quit [Quit: Nettalk6 -]
\\Mr_C\\ has joined #linux-sunxi
<adrian> i think there is a problem with the argument passing to the kernel
<adrian> if i use like this bootargs 'console=ttyS0,115200 root=/dev/nfs nfsroot=${serverip}:/home/nfs/arch ip=${ipaddr}:${serverip}:${gatewayip}:${netmask}::eth0'
<adrian> is not working
<adrian> but if i use like this bootargs 'console=ttyS0,115200 root=/dev/nfs nfsroot=${serverip}:/home/nfs/arch mac_addr=${ethaddr} ip='
<adrian> the network is configured
<arokux1> adrian, so the variable are not expanded correctly?
<adrian> also if i set the mac_addr=${ethaddr} like this it gives me not random but different mac address
<adrian> if i put it manualy mac_addr=xx:xx:xx:xx:xx is correct
<adrian> i thik yes is not read correctly
<arokux1> adrian, "not random but different" how do you know that "different" wasn't chosen randomly? :)
<arokux1> adrian, (for records, what kernel and uboot have you used?)
<adrian> i pinged it and than arp see same ip every time
<adrian> but not the one i set
<adrian> is not random
<adrian> U-Boot 2013.04-06942-gc057dc2-dirty (Jun 06 2013 - 16:51:55) Allwinner Technology
<adrian> arm-none-linux-gnueabi-gcc (Sourcery G++ Lite 2010.09-50) 4.5.1
<adrian> GNU ld (Sourcery G++ Lite 2010.09-50)
n01 has quit [Ping timeout: 260 seconds]
<arokux1> adrian, Hm.. arp see same IP or same MAC?
<arokux1> adrian, (kernel version?)
<adrian> same mac
<adrian> wait i dont remember
<arokux1> adrian, ok, the mac is set once and than does not change. but it was probably chosen randomly....
<adrian> now is booting :D
n01 has joined #linux-sunxi
<adrian> it booted but, i must disable dhcp
<adrian> 00:00:0e:00:dd:00
<adrian> this mac every time
<arokux1> adrian, every time after you reboot, you mean?
<adrian> yes
<adrian> everytime board is booting
<arokux1> adrian, (have you checked you kernel version? "uname -a")
<adrian> now
<adrian> it is 3.3
<adrian> but i dont remember which branch i used
<arokux1> adrian, 3.3 is the branch, you've used linux-sunxi, or?
<adrian> yes
<arokux1> adrian, hm.. if you used your command with ${serverip}, did you actually set this variable somewhere?
<adrian> yes
<arokux1> adrian, like setenv serverip
<arokux1> ?
<adrian> i set all of them
<adrian> now i wirte mac_addr manualy and is the one i selected
<arokux1> adrian, like "setenv serverip" ?
<adrian> Linux A20 3.3.0+ #1 SMP PREEMPT Mon Aug 5 16:33:42 EEST 2013 armv7l GNU/Linux
<adrian> uname -a
<arokux1> adrian, you can check at the end of this page for: Booting Arch Linux from an NFS share
<adrian> serverip=
<arokux1> adrian, you do not need "=" :)
<adrian> yes from there i started
<adrian> this is from printenv serverip
<arokux1> adrian, ah, ok
<adrian> sun7i#printenv serverip
<adrian> serverip=
<arokux1> adrian, the link I've shown to you uses newer bootloader, yours is quite old.
<adrian> i compiled new bootloader but.. there was a problem with spl
<arokux1> adrian, what was wrong?
<adrian> freezes after detecting mmc
fredy has quit [Excess Flood]
<adrian> i used the olimex provided spl, ( i copied from the sd image) and just writed the u-boot, but there was a problem booting the kernel
<adrian> i will try again
<arokux1> adrian, what board are u sing?
<adrian> a20 olinuxino
<adrian> micro
<adrian> i used A20-OLinuXino_MICRO from boards.cfg
fredy has joined #linux-sunxi
<adrian> now is booting just fine, but with all the ip manualy assinged
<arokux1> adrian, wait, but have u set ipaddr, gatewayip etc. previously?
<adrian> yes
<adrian> with dhcp actualy
<adrian> than save env
<adrian> saveenv
<arokux1> if set with dhcp, and printenv, do you see these variables assigned?
<arokux1> adrian, ^
<adrian> sun7i#printenv gatewayip ipaddr netmask
<adrian> gatewayip=
<adrian> netmask=
<adrian> ipaddr=
<arokux1> adrian, gatewayip is wrong, or?
<arokux1> adrian, you have used in you bootargs.
<adrian> gateway is assigned by dhcp server
<adrian> which is one wireless router
<arokux1> adrian, I see, but you set it to and not 50, manually, why?
<adrian> i set the serverip to
<adrian> aka nfs server
<adrian> sun7i#printenv gatewayip ipaddr netmask serverip
<adrian> gatewayip=
<adrian> ipaddr=
<adrian> serverip=
<adrian> netmask=
<arokux1> adrian> but if i use like this bootargs 'console=ttyS0,115200 root=/dev/nfs nfsroot=${serverip}:/home/nfs/arch mac_addr=${ethaddr} ip='
<adrian> the gateway i did not evan set in my bootargs
<adrian> is blank
<adrian> the problem is that is not reading correctly &{env} variable
<adrian> sorry ${env}
<adrian> i think i made a mistake....
<adrian> i used ' instead of "
<arokux1> adrian, I've noticed your command, where you use variables has not enough ":" at the end.
vinifr has joined #linux-sunxi
<arokux1> adrian, but it should't be a problem, I think
<adrian> but i set the mac addr like that and is working
<adrian> i will set all like before and test
<adrian> ip=<client-ip>:<server-ip>:<gw-ip>:<netmask>:<hostname>:<device>:<autoconf>: <dns0-ip>:<dns1-ip>
eebrah is now known as eebrah|away
<adrian> there are 8 ":" here mine also have 8 ":"
<arokux1> adrian, just try ip=dhcp
<vinifr> hi, ​​which values I could use for 'bootm 0x<kernel-address> 0x<initramfs-address> 0x<dtb-address>'
<arokux1> vinifr, have you done fatload already?
<vinifr> arokux1, yes, fatload mmc 0 0x41000000 sun5i-a13-olinuxino.dtb
<adrian> bootargs "console=ttyS0,115200 root=/dev/nfs nfsroot=${serverip}:/home/nfs/arch mac_addr=${ethaddr} ip=${ipaddr}:${serverip}:${gatewayip}:${netmask}::eth0:::"
<arokux1> vinifr, so you use this for <dtb-address>
<adrian> this one is working
<adrian> it was the -> ' <-
<adrian> must use double quote
<arokux1> adrian, have you tried ip=dhcp?
<adrian> yes
<arokux1> adrian, now it could work too, will save you some typing work
<adrian> if i dont provide same mac address is not working, but i think now it must work,
<adrian> actually is thre a way to start dhcp first time on u-boot
<arokux1> adrian, first time?
<adrian> yes before loading kernel
<adrian> on bootcm
<adrian> on bootcmd
<vinifr> i do 'ARCH=arm CROSS_COMPILE=<toolchain-prefix> LOADADDR=0x40008000 make uImage dtbs', so kernel-address=0x40008000?
<arokux1> adrian, well, you just do "dhcp"
<adrian> i will try now
<arokux1> vinifr, no, it is the address you'll use in fatload .... uImage
<arokux1> vinifr, please report your success/failure.
<vinifr> arokux1, hum, is it needed load .dtb and uImage?
<arokux1> vinifr, yes. dtb is supplied as kernel parameter.
<arokux1> vinifr, read documentaion on bootm, "help bootm"
<arokux1> vinifr, it's quite slim, some things are ommited there. I was going to update it.
<vinifr> arokux1, thanks :)
<arokux1> vinifr, you need to fatload all three things and then bootm ....
<vinifr> then it is mandatory to use initramfs too? is it not optional?
<arokux1> vinifr, which kernel do u use?
<adrian> back to original problem
<arokux1> vinifr, so mainline?
<adrian> it is not a solution...
<arokux1> adrian, what's wrong now?
<vinifr> arokux1, yes
<adrian> if i set it with double quote is writing the values of the variables
<arokux1> vinifr, ok, mainline does not have support for any storage. so your initramfs should be built into the kernel
<adrian> not the variables
<arokux1> vinifr, have you done it?
<adrian> bootargs=console=ttyS0,115200 root=/dev/nfs nfsroot=${serverip}:/home/nfs/arch mac_addr=${ethaddr} ip=${ipaddr}:${serverip}:${gatewayip}:${netmask}::eth0:::
<adrian> this one
<arokux1> adrian, i thought it was working with ----> "
<vinifr> arokux1, no, i am going do it
<arokux1> vinifr, so do it first.
<adrian> if i write it with double quote it become bootargs "console=ttyS0,115200 root=/dev/nfs nfsroot=${serverip}:/home/nfs/arch mac_addr=${ethaddr} ip=${ipaddr}:${serverip}:${gatewayip}:${netmask}::eth0:::"
<arokux1> vinifr, i'll update this wikipage tonight.
<adrian> did nto check it
<adrian> sun7i#printenv bootargs
<adrian> bootargs=console=ttyS0,115200 root=/dev/nfs nfsroot= mac_addr=00:4e:56:a1:45:01 ip=:
<vinifr> ok
<adrian> this is how it looks if i put with double quote
<adrian> i think i have to upgrade the u-boot...
mcbrick has joined #linux-sunxi
<arokux1> adrian, I see, have you tried with ip=dhcp, it will save us from trouble maybe
<adrian> it works only if i set the mac_addr manualy
<adrian> the thing is that i want to set only ethaddr in the u-boot
<adrian> not the entire bootargs
<arokux1> adrian, but how it will know you're going to boot from NFS?
<arokux1> bootargs "console=ttyS0,115200 root=/dev/nfs nfsroot=${serverip}:/home/nfs/arch mac_addr=${ethaddr} ip=dhcp"
<arokux1> will this work?
<adrian> with bootcmd
<adrian> bootargs=console=ttyS0,115200 root=/dev/nfs nfsroot= mac_addr=00:4e:56:a1:45:01 ip=
<adrian> bootcmd=dhcp nfs 0x43000000; nfs 0x48000000; bootm 0x48000000
<adrian> this is the only thing i modify to boot from nfs
<adrian> it is actualy from that web page you gave earlyer
<arokux1> adrian, shouldn't you have ";" after dhcp?
<adrian> all i want is to make the sd card last longer, the nfs is actualy mounted on ram file system
<adrian> yes
<adrian> i forgot it
<adrian> i want to make one some more client networked boards like this, booting same os
<arokux1> adrian, i see.
<adrian> but i dont want the sd card to get damaged
<arokux1> adrian, does it work with ";" after dhcp?
<adrian> it works the dhcp for the u-boot
<adrian> but
<adrian> Using wemac device
<adrian> File transfer via NFS from server; our IP address is
<adrian> Filename '/home/nfs/arch/uImage'.
<adrian> i get this if i set the boot cmd to
<adrian> sun7i#printenv bootcmd
<adrian> bootcmd=dhcp; nfs 0x43000000 ${serverip}:/home/nfs/arch/script.bin; nfs 0x48000000$ {serverip}:/home/nfs/arch/uImage; bootm 0x48000000
<arokux1> adrian, have you set serverip?
<adrian> yes
<arokux1> adrian, try setting it directly. if it works, then you messed up with variables again
<adrian> yes i forgot $ from {serverip}
<adrian> for this one is ok... i wander why is not for passing args
<adrian> maybe i need to run them separatly
<arokux1> adrian, you just need to set bootargs.
<adrian> like bootargs set myargs
<arokux1> adrian, hm.. I think you need a script, because variable will only exist after dhcp command was run
<arokux1> adrian, so you need to set serverip, run dhcp and then set two variables bootcmd and bootargs, then run "boot"
<arokux1> mnemoc, how to add contents to a wiki page?
<vinifr> Create a account and just edit
<adrian> finaly
<adrian> bootargs=
<adrian> bootcmd=run nfsargs; dhcp; nfs 0x43000000 ${serverip}:/home/nfs/arch/script.bin; nfs 0x48000000$ ${serverip}:/home/nfs/arch/uImage; bootm 0x48000000
<adrian> nfsargs=setenv bootargs console=ttyS0,115200 root=/dev/nfs nfsroot=${serverip}:/home/nfs/arch mac_addr=${ethaddr} ip=${ipaddr}:${serverip}:${gatewayip}:${netmask}::eth0:::
<adrian> this is working
<adrian> i put empty bootargs
<adrian> than run nfsargs
<adrian> and is working
<adrian> so bootargs must not contain variables
<adrian> good to know
<arokux1> adrian, it could, but only after dhcp was done, or?
<adrian> yes after dhcp
<arokux1> adrian, can you please add your commands to this page:
<adrian> please if is possible, can i give you the printenv output and you can write it?
<arokux1> adrian, I can do this for you, but you should give me list of commands you have used, not printenv, also use some pastebin website.
<arokux1> hno, ping
FR^2 has quit [Quit: Connection reset by peer]
blunden_ has joined #linux-sunxi
blunden has quit [Ping timeout: 245 seconds]
<arokux1> adrian, so? :)
<adrian> now i write it
<adrian> here it is
<arokux1> adrian, you have used ' instead of " :)
<adrian> yes i used single quote
<adrian> and is ok like this
<arokux1> adrian, ah, ok.
<adrian> if i use double quote is putting the values not the variables
<arokux1> adrian, aha.. this is good to know.
<adrian> by the way i wander is there a problem if multiple pc share same nfs server?
<adrian> like file write concurrency
massi has quit [Quit: Sto andando via]
<adrian> maybe is better to make one path variable also
<adrian> so if anyone want to change the path from the env setup, just change one place
<adrian> in one place ^
<arokux1> adrian, well, some parts of the rootfs are readonly, like /usr/bin, some are rw
<arokux1> adrian, yes, good idea. just change the wiki page:
<adrian> by the way i had no ideea about nfs servrs untill 2 days ago, so i dont know if i do it correctly
<vinifr> which the contents of initramfs to linux-sunxi?
<arokux1> vinifr, what?
<vinifr> what do i put into initramfs?
<vinifr> initramfs=rootfs?
<arokux1> vinifr, your rootfs, it should be small one
<arokux1> vinifr, why do you want to boot mainline, btw?
<vinifr> i wanna test maxine's i2c driver and my lradc driver
<vinifr> lradc need iio framework
<arokux1> vinifr, cool
<Turl> vinifr: I can give you a copy of mine, should have i2c userspace tools
<Turl> you can build your own with buildroot
<vinifr> My rootfs is too large, i am using ubuntu-rootfs
<arokux1> Turl, can we create some tutorial how to do in on wiki? with prebuilt images for different scenarios?
<vinifr> Turl, I would be grateful
<Turl> arokux1: buildroot already has good docs, no need to duplicate imo
<arokux1> Turl, ah, ok then.
<vinifr> Do you generated it using arm-linux-gnueabihf- ?
<Turl> vinifr: buildroot downloads toolchain for itself
<hno> arokux1, yes?
<arokux1> hno, i'm thinking could we move your u-boot wiki to the central wiki?
<vinifr> Turl, thanks
<arokux1> hno, right now i'm cleaning and generalizing our central wiki. there are a lot of people asking same questions every time, so if you write them down in wiki we could answer them quickly.
<hno> I don't like wiki servers without a good manageable backend... but feel free to copy stuff if you like, but please leave a reference.
<arokux1> hno, so you want your wiki to stay too?
<arokux1> hno, I didn't get it.
<hno> arokux1, until linux-sunxi wiki is using a backend that I can "git clone" or equivalent yes.
tat has quit [Remote host closed the connection]
<arokux1> hno, why do you want git clone for a wiki? :)
<hno> because I don't trust mediawiki, or shared hosting. Only shared(distributed) storage of knowledge.
<arokux1> hno, oh... ok
<hno> and also do not see my browser as a usable editor
<arokux1> hno, there is new editor from wikimedia, there you can edit contents directly
<hno> the github wiki is just a git repository. clone/push/merge etc as any other source.
<hno> same for is rhombustech wiki.
<arokux1> hno, I do not care much about backend, just about contents and absence of duplication.
<hno> I care about knowing that any data I produce will be safe and available.
<hno> part of that is encouraging duplication in a controlled manner.
<arokux1> hno, hopefully mnemoc creates backups of our wiki
<hno> wii information really is no different than kernel source. Only access method differ.
<hno> yes I think mnemoc creates backsups of the wiki. But that is only a small part of the picture.
<arokux1> vinifr, this might be useful for u
<arokux1> hno, is this a correct command? make cubieboard_config
<arokux1> hno, on your wiki it is: make 'boardtype' CROSS_COMPILE=arm-linux-gnueabihf-
<hno> arokux1, make cubieboard CROSS_COMPILE=...
<wingrime> arokux1: any success today
<adrian> i found one issue with having multiple pc on one nfs server... on debian, they want to edit /etc/udev/rules.d/70-persistent-net.rules
<hno> arokux1, where CROSS_COMPILE=... is sett to your toolchain, just as when you compile kernel or most other projects.
<adrian> how can i disable this so it will never try to write this file, can i delete it?
<arokux1> hno, yeah, just wanted to know if _config can be there.
<arokux1> adrian, disable what/
<arokux1> ?
<hno> adrian, you should in general have a separate /etc per server. But you can disable most things such as that udev rule thing.
<adrian> i dont know udev is writing to this file /etc/udev/rules.d/70-persistent-net.rules
<hno> it's not udev directly no.
<adrian> so how i disable it?
<adrian> # Unknown net device (/devices/platform/wemac.0/net/eth0) (wemac)
<adrian> SUBSYSTEM=="net", ACTION=="add", DRIVERS=="?*", ATTR{address}=="00:4e:56:a1:45:01", ATTR{dev_id}=="0x0", ATTR{type}=="1", KERNEL=="eth*", NAME="eth0"
<adrian> this is what is writing in it, so having multiple mac address for each board, maybe it will be a problem
<hno> That's an udev event. Many things reacts on those.
<arokux1> wingrime, haven't tried anything yet.
FR^2 has joined #linux-sunxi
<hno> adrian, there should be a udev rule that performs the file update.
<hno> Fedora does not have that file. lets NetworkManager worry about interfaces.
<arokux1> hno, you run fedora?
_BJFreeman has joined #linux-sunxi
_BJFreeman is now known as BJfreeman
rz2k has joined #linux-sunxi
<vinifr> after booting can I use regular ubuntu-rootfs? 'root=/dev/mmcblk0p2'
<arokux1> vinifr, no, as there is no driver for storage
<Turl> vinifr: on mainline? no, no mmc driver
<vinifr> huum, it is need mmc driver first
<arokux1> Turl, any idea how i can move Contents to the top?
<vinifr> thanks to all
<Turl> arokux1: by putting a proper title to the first section
vinifr has quit [Quit: Saindo]
<arokux1> Turl, thanks
<arokux1> Turl, could it be nice idea to add a mainline branch to linux-sunxi?
heffer has joined #linux-sunxi
heffer has quit [Read error: Connection reset by peer]
heffer has joined #linux-sunxi
<arokux1> arokux1, so that we have everything in one place?
_whitelogger_ has joined #linux-sunxi
_whitelogger_ has joined #linux-sunxi
<mnemoc> think so
<Turl> arokux1: I just pushed a fresh sunxi-devel, so it's not that out of date
<arokux1> mnemoc, ok, I think we need a mirror of torvalds tree too, for completeness
<Turl> I still need to go through all the stuff mripard managed to send though
<Turl> arokux1: mirror/master is there
<mnemoc> *g*
<arokux1> Turl, how can I update it?
<Seppoz> good evening guys :)
<deasy> mouarf, wordpress trouble solved for directory uploads : selinux
<Turl> arokux1: update what, your copy of sunxi-devel?
<arokux1> Turl, no, mirror/master on github
<mnemoc> arokux1: it's updated by a bot
<mnemoc> together with the other mirrror/android-*
<Turl> mnemoc: it's like 10d out of date though
<arokux1> mnemoc, when does the bot do it?
<arokux1> Turl, yep :)
<Turl> I just pushed torvalds' head for you
<mnemoc> my N7 decided to download the upgrade to 4.3.... eating all my 3G bw
<mnemoc> i'll check why it's frozen after that
<Turl> mnemoc: haha
<hno> mnemoc, where is the bot?
<mnemoc> hno: cron task in my build server
<mnemoc> running manually now
<mnemoc> hope to move that to the new btrfs-less server today or tomorrow
<hno> ok.
<mnemoc> io in thius machine sucks badly
<Turl> mnemoc: betterfs? :P
<mnemoc> sync'ed
<mnemoc> Turl: ext4 over raid10. tired of fanciness
<Turl> agreed :)
<hno> less thinking needed in such setup.
<wigyori> hno: do you happen to have an a13 board?
<hno> btrfs is nice when used right, but it's perhaps not the best choice for a full blocklevel container storage.
<jelly-home> also less worrying about untested new code
<hno> wigyori, I have 3 I think.
<wigyori> hno: that's cool - could i ask you to test out the uboot and kernel builds i have? still no luck with them apart from the starting kernel message, and i start to suspect that it's due to the board being an old rev. b olimex board
zoobab_ has quit [Quit: Lost terminal]
tzafrir has quit [Read error: Operation timed out]
<adrian> i managed to make nfs read only
<adrian> actualy that was easy but i made it start my application
<adrian> which is based on x11 lib
<hno> wigyori, I can try, but have not experienced any such oddnes with either rev-a or rev-b.
<hno> hmm.. correction, it's rev-a and rev-c I have.
<wigyori> hno: that's fine - i'd just like to see if it boots on another a13 and if it's an issue with my board, or if my kernel/uboot is screwed up completely
<wingrime> hno: it possible use ext2 in boot part or even single partition image ?
<deasy> yup question for hackers, nand driver into linux is always bad or it's better? they have say on #cubieboard than it's bad
<deasy> (i don't know the software it is maybe it's not "nand driver")
arokux has joined #linux-sunxi
<arokux> adrian, still there?
<deasy> ha arokux was not here... hmm
<arokux> can one have some bots on github to pull mirrors, for example
xtofury has joined #linux-sunxi
<xtofury> thank you hno those instructs are much more clear.
<hno> wingrime, yes both works.
<hno> deasy, fuzzy subject.
<wingrime> hno: about, fixed offsets
<wingrime> hno: in dd
<wingrime> hno: thay will not interfere with filesystem?
<deasy> hno, they say trouble with distro and use of the nand
<wingrime> hno: also ext4 should works fine?
fredy has quit [Excess Flood]
fredy has joined #linux-sunxi
xtofury has left #linux-sunxi [#linux-sunxi]
<arokux> anybody here with gmail and downloading (selected) patches automatically and applying them?
<hno> wingrime, just make sure partition starts at 1MB (block 2048)
<hno> or later.
tzafrir has joined #linux-sunxi
<hno> fat, ext2,ext3,ext4 is supported. And esy to add xfs, btrfs or any other fs supported by u-boot.
tat has joined #linux-sunxi
<tat> i have compiled custom kernel mdules ftdi driver for a13, i don't have the original kernel source, my modules don't load, i think it is the srcversion: in the module elf header, is there a good way to forge that string during compile ?
<hno> deasy, are you referring to my and Patrick Wood discussion about the pros and cons of nand vs sd?
<Dreadlish> its plain string
<Dreadlish> you're able to edit it after compilation
<hno> or just "modprobe -f"..
<tat> Dreadlish: what tool to use best
<tat> hno: i only have insmod on my android
<Dreadlish> tat: do you have busybox here?
<tat> Dreadlish: i want forst make a backup before i install busybox, i have a busybox but it misses a lot of tools,
<hno> tat, insmod -f then.
<hno> but maybe android insod do not have -f..
<tat> hno: insmod: can't open '-f'
<Dreadlish> there is no insmod which have -f parameter
<arokux> ok, another question then. what approach (client etc.) do you use to apply a patch from an email?
<tat> Dreadlish: how would i change the srcversion string after compile ? hexedit
<Dreadlish> yup
<arokux> modprobe -f
<Dreadlish> just plain hexeditor
<Dreadlish> arokux: not everytime ;)
<Dreadlish> modprobe -f doesn't skip inkernel checks
<arokux> Dreadlish, what will it skip then?
Soul has joined #linux-sunxi
<arokux> are you guys downloading patches by hand..? o_O
<Dreadlish> arokux: you have to disable them in kernel config
<arokux> Dreadlish, what is this option good for, then?
<Dreadlish> arokux: have you heard about versioning systems?
<Dreadlish> -f just forces to load module second time
<Dreadlish> for example
<Dreadlish> when it crashes or sth like that
<arokux> Dreadlish, i'm asking how patches are applied from e-mail.. without much of fiddling with browser of e-mail client
<Dreadlish> arokux: fetchmail, rest of magic can be done locally
<Dreadlish> if you can get gmail working with fetchmail you can also google for rest
<Turl> arokux: on gmail?
<arokux> Turl, yes
<Turl> go to the menu inside the mail and choose 'view original' or so
<Turl> then save that text to a file
<Turl> then git am file
bsdfox has quit [Ping timeout: 276 seconds]
<arokux> Turl, hm.. any tool for that? that will give me a list of e-mails with patches, so that I just choose from it?
<arokux> Turl, preferably under terminal :)
Black_Horseman has joined #linux-sunxi
<wingrime> include/config/auto.conf:129:warning: symbol value 'm' invalid for USB_SW_SUNXI_UDC0
<wingrime> mnemoc: !!
<Turl> arokux: I dunno, personally I hate console email clients
notmart has quit [Quit: notmart terminated!]
<Turl> I just use thunderbird, choose as many patches as I want, save them as files and then use git am
<wingrime> Turl: nice , linux-sunxi still charset at build
<arokux> Turl, well, me too. I'm not asking for the whole client..
<Turl> wingrime: charset?
<wingrime> /media/archive1/sunxi/linux-sunxi/arch/arm/mach-sun7i/pm/standby/mem_printk.c:168: undefined reference to `memcpy'
<arokux> Turl, found something:
<wingrime> Turl: no (crash)
<wingrime> Turl: auto complite reflex ))
<Turl> wingrime: pm standby image always have derp buildsystem from aw
<Turl> we fixed it for sun4/5i I think
<Turl> maybe not sun7i
<Turl> mnemoc: ^
<arokux> :(
<arokux> I cannot get the workflow right.
<wingrime> Turl: I make a patch for hasng branch
<wingrime> Turl: but Have no idea will it apply
<arokux> I have my tree and committed something. the other guy also committed something on his tree. the third one applied the patches of the second guy and asked me to rebase mine.
<tat> i get for my module "disagrees about version of symbol module_layout" so is there a way to get arround that
<Turl> tat: yes, build kernel and module together/with same config
<Turl> (and from same code)
<wingrime> drivers/ata/sw_ahci_platform.c:42:29: fatal error: plat/sys_config.h: Нет такого файла или каталога
<wingrime> Turl: nice
<Seppoz> nice error!
<tat> Turl: i don't have the original source, but the bigger problem i don't have the original config
<wingrime> mnemoc:!!
<Turl> wingrime: can't read russian
<Turl> tat: then grab the source, roll your own config and use matching kernel and modules
tat has quit [Remote host closed the connection]
bsdfox has joined #linux-sunxi
<Turl> mripard_: :)
<Turl> mripard_: I believe MS stands for Memory Stick
<deasy> hno, sorry i was eating, nop except if it was 2-3days ago on cubieboard i don't though it
<hno> deasy, discussion has been ongoing for several days under subject A little hardware hack to add SDC2 ontop of NAND.
<arokux> mripard_, is working hard
<hno> Turl, MS where? In Allwinner docs yes.
<hno> and code.
<Turl> hno: yes, docs
<Turl> hno: is there code for it? I never saw it
eebrah is now known as eebrah|away
<hno> Turl, no code seen outside maybe a clock gate.
<hno> but it's confirmed it is the memory stick controller.
<Seppoz> does 3.4 have EInt support?
<Seppoz> like GPIO_BUTTONS
fredy has quit [Excess Flood]
<hno> Seppoz, not sure.
<Seppoz> is there a funciton to aquie interrupts form GPIO?
<Seppoz> and is it registered to the kernel
fredy has joined #linux-sunxi
<hno> Seppoz, GPIO driver talks alot about EINT so I suppose it handles it.
<Seppoz> ok :)
<zerodamage_> Seppoz: can i use the 3.4 branch even if the current kernel on SD is 3.0? or will that mess the modules or something?
<zerodamage_> im having an issue where an wxwidgets application seems to crash the display.... if i start it the hdmi disappears, but ssh still works
<arokux> oops, I've spammed into our mailing list..
rellla has quit [Remote host closed the connection]
<arokux> oh no, I didn't
<zerodamage_> ohh noo you did'ant
\\Mr_C\\ has quit []
<libv> arokux: is there a reason why you do not use your real name?
<arokux> libv, yes, if I use it it will remain forever...
<libv> ?
<libv> arokux: so you do not want your name linked to some software work you did?
<arokux> libv, in all the possible lists etc. I do not like people can easily find out what I was doing and when simply by using google.
<libv> it will show that you've been spending your spare time doing something good
<arokux> libv, it's enough my nickname is link. if needed i'll disclose my name then
<libv> i am not sure whether there is an actual need
<libv> but i just find it strange
zerodamage_ has quit [Quit: Page closed]
<arokux> libv, I just do not want my name is everywhere on the internet, that's all.
uwe_ has quit [Ping timeout: 245 seconds]
<arokux> hm, how frequently will gmane synchronize?
uwe_ has joined #linux-sunxi
<arokux> libv, how is work on lima going?
<libv> arokux: too slow, i am not as code-motivated as i should be
<libv> now reworking the limare linking code so the rudimentary mesa driver can use it too
<rz2k> libv: arokux: i bet that you cant push to mainline without your realname.
<libv> rz2k: at least it is "just not done"
<arokux> rz2k, everybody uses there names can't believe that. also: this was patch to the 3.4 :p
<Turl> rz2k: I believe so
<Turl> rz2k: iirc it's for a legal reason (S-o-b tags have some legal meaning)
<rz2k> yep
<arokux> S-o-b?
<rz2k> signed off by
<rz2k> and then your realname
<rz2k> there's also repoted by, tested by and other markings
<rz2k> if needed
<libv> arokux: maybe it is time to either tie arokux to your realname for eternity (the nsa knows who you are anyway :ppp), or to come up with a new nickname, and rejoin freenode with that
<rz2k> nsa knows everything
<arokux> rz2k, yes, I've heard they have some legal meaning, but consider S-o-b: John Doe... will that help? I think no.
<rz2k> lol
<arokux> libv, new nickname? I've actually spent a lot of time thinking this one out. google didn't know it before.
<libv> arokux: i personally use goatsaresexy for my less morally acceptable online activities (not betraying which though :p)
<Turl> unless there's some weird country where free software developers are prosecuted, I dunno why you wouldn't like to tie your name to it
<Turl> in fact it's going to get you better job posts and stuff on the SW industry
<rz2k> if you are doing job on something that might have legal impact (ip cores, patented algos, whetever), linux foundation will want to contact you as you are now part of group of people who "authored" linux.
\\Mr_C\\ has joined #linux-sunxi
<hno> Turl, you would not beleive the odd requests I get for removing old posts from mailinglists...
<rz2k> reverse something broadcom and then try not to go to jail :p
<arokux> hno, :)
<Turl> hno: do you comply though? :)
<arokux> question, do you use gmail? if so, are the mails grouped by thread from our ML?
<rz2k> we have ML at google groups
<Turl> arokux: I do (via thunderbird), yes
<arokux> they say it should, but it doesn't work, although i have those in-reply-to and references.
<arokux> Turl, what about web interface i.e. native one.
<Turl> let me check
<hno> Turl, for most part I can't, most of the archives is not under our control.
<arokux> hno, you see.. then it's there forever.
<Turl> arokux: stuff like [0/1] and [1/1] you just sent aren't, but stuff with the same title is
<libv> arokux: why is it wrong to have your name and email address known forever?
<hno> yes, so do what I practioned for years, write so much so anyone who goes looking drowns.
<rz2k> maybe arokux1 plans to blow the world up soon :p
<libv> there is only limited data that can be mined from open source activities
<Turl> hno: yeah, trying to kill email in public lists is like trying to extinguish a house on fire with a glass of water
<arokux> libv, just because it will be there with no possibility for you to stop it later.
<Turl> hno: haha good one :)
<libv> arokux: choose another nick, another email address and another provider
<libv> and then the world will not be able to tie you to your name as quickly
<arokux> libv, why? i have one already :)
<Turl> arokux: xnkol9, it worked for allwinner :)
<rz2k> lol
<rz2k> reumillatech
<arokux> Turl, xnkol9? whos that? :)
<Turl> "arokux" upside down and reversed
<hno> Was quite funny, one day there was an article on who in sweden was most famous in google search, and they have tried a number of known and less known swedish persons. Tried googling on me and it gave several orders of magniture larger hit..
<hno> but googla have improved their search result processing considerably since then so results makes much more sense today.
<Turl> :)
<arokux> Turl, gmail will sort threaded messages together only if subjects are the same, too bad. cannot be used for patch series, each patch has a different subject.
<Turl> arokux: btw, was looking @ wiki, are you working on usb otg?
<Turl> arokux: if not, you wrote your name under the wrong section
<arokux> Turl, no, jukivili is. I'm working on usb host.
<Turl> (well, nickname rather :P)
<arokux> Turl, :p thanks, will correct it.
<arokux> Turl, I've did some clean ups on the wiki today, not sure you'll like it
<arokux> created a page toolchain and put all info about toolchains on it.
<arokux> also moved stuff to bootable sd card and u-boot
<arokux> removing duplication
<Turl> arokux: wikis have a revert button for a reason
mcbrick has quit [Remote host closed the connection]
<arokux> Turl, so who was xnkol9? some allwinner spy in our irc?
<Turl> codesourcery page could be merged into toolchain too
<arokux> Turl, yep
<Turl> arokux: it's just my attempt at writing "arokux" rotated 180º
<Turl> just like allwinner did on copyright notices once :)
mcbrick has joined #linux-sunxi
<hno> Turl, that was Tom...
<hno> and i respect him for it. stupid managers deserve to be mocked.
<arokux> how would I delete a page on wiki?
<hno> arokux, normally you redirect to where the content is now
<Turl> put a redirect in place
<Turl> then use "what links here" and fix the other pages linking to it
<arokux> Turl, and then, can it be deleted now?
<Turl> no, you need the redirect for people coming from other sites/google
<arokux> Turl, you see, once it's there it is there forever :(
<Turl> :p
<hramrach> ssvb: any point using sunxifb on exynos?
<hramrach> I am running fbdev and it works fine
<hno> arokux, that should be a :)
<hramrach> or better than the armsoc driver, reportedly. And good enough to get graphics
<arokux> hno, a?
<hno> your last statement.
<ssvb> hramrach: exynos5?
<hramrach> yes, exynos5
<arokux> hno, sorry, didn't get it.
<hno> You typed :( where you meant :)
<arokux> ah, well, it depends
<hramrach> there is some stupidity with dual output handling. fbdev clones which is somewhat usable and armsoc crashes
<hramrach> and armsoc has dubious performance advantages. Sometimes it works sometimes not
<arokux> question: when will the patches that Hans catches on your ML lend either into stage/sunxi-3.4 or sunxi-3.4 itself?
<arokux> on our* ML
<hno> hramrach, how would wunxifb work on exynos5?
<hramrach> like every other hardware without g2d. use noen or w/e
<hramrach> neon
<ssvb> hramrach: armsoc is only useful for 3d
FR^2 has quit [Quit: und weg...]
<hramrach> how useful? there is only gles and that is X independent, right?
<hno> how is sunxifb different from generic fbdev then?
<ssvb> hramrach: armsoc is using x11 and dri2 for gles
<ssvb> hno: there are some extra optimizations
<ssvb> hno: I really need to write some blog post about this, just to clarify what exactly is better than fbdev
<hramrach> is there some dual output handling?
<hno> should any of those optizations go into fbdev so other Xorg drivers can use them?
<ssvb> hramrach: for exynos5 the perceived performance difference should not be that significant, but for raspberry pi it is
<hramrach> yes, it seems fast enough with fbdev
<hramrach> good cores there I guess
<ssvb> hramrach: armsoc supports xrandr, but fbdev (and sunxifb) does not
<ssvb> hno: I really need to write some blog post about this :)
<hramrach> and armsoc reportedly crashes when connecting hdmi
<hno> ssvb, ok, i'll wait for your blog post :)
<hramrach> plus I don't have fb1 so god knows if that's going to work
<hramrach> an exynos4 guide says there should be fb0+fb1
<libv> exynos4 hdmi works fine with sunxifb
<libv> needs to be pointed at fb6 through a xorg.conf option, but that's all
<libv> works better with the early lima mesa driver than for instance arm's own "driver"
<ssvb> as long as there are /dev/fbX and /dev/fbY device files for different monitors, fbdev driver supports multi monitor configuration at least in zaphod mode
jelly-home has quit [Ping timeout: 245 seconds]
<ssvb> but afaik not hotplug
hurtigbuffer has joined #linux-sunxi
<libv> we have a todo list from here to the moon anyway, so it is good enough for now
<arokux> hno or Turl , can you address my question please?
<libv> arokux: i think the answer is, when mnemoc gets around to it
<hramrach> arokux: extra testers improve confidence in the patches ;-)
<arokux> hramrach, yep, but it's difficult to know on what to base new patches
<hramrach> see the github repo for whole branches
<arokux> hramrach, won't help, as there are forks all over the place
<hramrach> git://
<arokux> hramrach, well yes, but maybe there is another guy X, that has even more patches?
<hramrach> you were asking about hansg patches and what to base them on
<arokux> hramrach, ok, my question is more general however. and it is: assume I want to do something. what tree should I take as base?
<hramrach> either you need to base it on some patchset so use a repo that has it or use the stage branch as base
<arokux> hramrach, well, you'd think this is the way to go. however later some person X shows up and tells you "hey! I had some patches already and I cannot apply yours, so rebase them please." and you start to rebase, which could be lots of work
<hramrach> or you base your patch on tons of other patches and some person X comes and says 'hey, these patches are nonsense, drop them'
<hramrach> and there you go rebasing either way
<Turl> arokux: hasn't mnemoc merged hansg branch already?
<arokux> Turl, no
<arokux> hramrach, exactly. you want to say this is normal?
<Turl> arokux: I can see all the sun7i patches hansg collected on stage/sunxi-3.4
<Turl> what am I missing?
<arokux> Turl, some usb stuff he has merged today :)
<Turl> ah, so he started a new branch. I see
<arokux> Turl, also some other patches, see:
<Turl> well, unless you need something from there, working on stage/sunxi-3.4 is a safe bet
Robs58 has joined #linux-sunxi
<arokux> Turl, how stuff lends on stage/sunxi-3.4?
<Turl> arokux: mnemoc lands it there
ganbold_ has quit [Remote host closed the connection]
tinti_ has joined #linux-sunxi
_BJFreeman has joined #linux-sunxi
_BJFreeman is now known as BJfreeman
naobsd has quit [Quit: Page closed]
fredy has quit [Excess Flood]
fredy has joined #linux-sunxi
wingrime has quit [Ping timeout: 245 seconds]