ChanServ changed the topic of #linux-sunxi to: Allwinner/sunxi development discussion - Don't ask to ask. Just ask! - See http://linux-sunxi.org | https://github.com/linux-sunxi/ | Logs at http://irclog.whitequark.org/linux-sunxi
ganbold has joined #linux-sunxi
<calris> Can anyone help me out with a monitor resolution / colour depth problem on sunxi Fedora 18?
<calris> I'm only getting the native LCD screen resolution which is way too high
<calris> monitor-get-edid does not give any output, so I'm a bit stuck on finding available mode lines
<hramrach> native is the only thing you get by default
<calris> I know - but I want to manually add a lower resolution
<hramrach> there is a way to override by kernel argument
<hramrach> but why lower res?
<calris> Its too slow
<calris> And I'm not in a position to recompile u-boot with faster memory timings
<calris> And what about dropping the colour depth to 16bpp? I can't figure out how to do that even
<hramrach> in xorg.conf
<hramrach> or setting -bpp argument when starting X server
<hramrach> you can see this for kernel options http://linux-sunxi.org/KernelArguments
<hramrach> but I have no idea what modes are supported and how to specify them
<calris> I was hoping EDID would tell me
<hramrach> you get the edid block hex dump in kernel log
<hramrach> you can try to pass that to decode-edid
<hramrach> hmm, not dumped anymore
<hramrach> well, the edid thing is kind of work in progress.
<calris> I have EDID bank 0 and EDID bank 1 outputs in dmesg
<calris> Are they any help
<hramrach> that's what you can use with decode-edid
<hramrach> jus pip it through something that does hex to binary data conversion
<hramrach> cat /sys/class/graphics/fb0/modes
<calris> :-)
<calris> Nice
<calris> Any idea what to first letter means?
<hramrach> no
<calris> Ah, looks like KMS needs to be supported first
<hramrach> well, you can set one of those modes with kernel arguments
<calris> Yes, I am just trying that - but I need to figure out if the first letter is important...
<hramrach> it's not used
<hramrach> you set something like 1080p60 or 640x480p50
<ssvb> hramrach: seems like fedora has scaler mode enabled by default, so 16bpp does not work
<hramrach> sucks
<hramrach> somebody who knows what it is used for and what it breaks should write something about the scaler
<ssvb> calris: what happens if you remove "EDID:" part from uEnv.txt?
<calris> ssvb: I removed it at set it to 1280x720 - works :-)
<ssvb> hramrach: IIRC scaler mode is needed for interlaced video modes with some TVs
<ssvb> calris: it should be somewhat faster, but don't expect miracles yet :)
<calris> ssvb: it is tolerable now
<ssvb> calris: it can be much better :)
<calris> ssvb: I hope to do a bit of U-Boot hacking soon
<calris> My plan is to add a filsystem to the SPL so the DRAM settings can be read from FTD
calris has quit [Read error: Connection reset by peer]
techn__ has quit [Read error: Connection reset by peer]
<Wetomelo> Hi, i download a supposed Pineriver MiniX image, but kernel log shows strange stuff
<Wetomelo> sometimes doesn't boot completly :wemac_init_wemac
<Wetomelo> does wemac the wired interface? but MiniX doesn's have one!
egbert has quit [Disconnected by services]
egbert has joined #linux-sunxi
<Wetomelo> I have a little lost. Where i should put sunxi-spl? which address?
<Wetomelo> I downloaded u-boot.bin and sunxi-spl.bin but don't know the offset in the sd card
<Wetomelo> I'm using a MiniX
<bsdfox> are these instructions valid for gpio-sunxi? http://linux-sunxi.org/GPIO
<bsdfox> I don't seem to have the same directory structure
egbert has quit [Disconnected by services]
egbert has joined #linux-sunxi
<Wetomelo> oh come on i'm doing something wrong. I'm trying to save bootloader from an sdcard
<Wetomelo> dd if=spl/sunxi-spl.bin of=/dev/sdc bs=1024 seek=8
<Wetomelo> dd if=u-boot.bin of=/dev/sdc bs=1024 seek=32
<Wetomelo> nope
<Wetomelo> this one
<Wetomelo> if=/dev/sdb of=/sunxi-spl-old.bin
<Wetomelo> =1024 seek=8
<Wetomelo> but stays forever, and the file grows and grws never stop!!
<bsdfox> Wetomelo, are you sure that's what you want?
<bsdfox> also is / on the sd card?
<Wetomelo> buuh doing mistakes sorry
<Wetomelo> i found the problem
Wetomelo has quit [Quit: Ex-Chat]
hramrach has quit [Ping timeout: 276 seconds]
hramrach has joined #linux-sunxi
ibrah has joined #linux-sunxi
ZaEarl has quit [Ping timeout: 264 seconds]
ibrah has quit [Read error: Connection reset by peer]
ibrah has joined #linux-sunxi
ibrah has quit [Ping timeout: 256 seconds]
ibrah has joined #linux-sunxi
ibrah has quit [Ping timeout: 252 seconds]
theOzzieRat has quit [Read error: Connection reset by peer]
theOzzieRat has joined #linux-sunxi
rellla has joined #linux-sunxi
n01 has joined #linux-sunxi
<Epsylon3> dd if=/boot.img of=/dev/block/nandc bs=4096
<Epsylon3> but nothing on serial :
<Epsylon3> NAND read: device 0 offset 0x3000000, size 0x2000000
<Epsylon3> 33554432 bytes read: OK
<Epsylon3> Starting kernel ...
<Epsylon3> for the 3.4
hipboi has joined #linux-sunxi
bsdfox has quit [Ping timeout: 252 seconds]
Epsylon3 is now known as E3-work
<mripard> Turl: \o/
<mripard> and he seems happy with the code
<E3-work> kernel/allwinner/common/drivers/misc/sun4i-gpio.c:249: warning: ignoring return value of 'sysfs_create_group', declared with attribute warn_unused_result
<E3-work> else awusb seems ok
shineworld has joined #linux-sunxi
shineworld has quit [Read error: Connection reset by peer]
christopher has quit [Remote host closed the connection]
vicenteH has joined #linux-sunxi
rz2k has joined #linux-sunxi
<mnemoc> E3-work: was that commit accepted upstream?
rellla2 has joined #linux-sunxi
rellla has quit [Ping timeout: 260 seconds]
rellla2 is now known as rellla
shineworld has joined #linux-sunxi
<hipboi> is there a vlc package?
<Dreadlish> for what.
<hipboi> sunxi vlc deb
<hipboi> i found sunxi vlc is usable
<hipboi> in some way
<Dreadlish> looks like yes.
<hipboi> :O
<hipboi> i mean the sunxi version
<Dreadlish> packages.debian.org will tell you everything you want ;)
<Dreadlish> yep, you can also build it yourself
<Dreadlish> and it can also do everything faster ;)
<hipboi> i built it
<Dreadlish> and?
<hipboi> and found it's usable, in some way
<hipboi> so i am wondering if there is pre-built one
<hipboi> if not, i will pack one
<Dreadlish> but it don't have to work with cedarx
<Dreadlish> (packaged one)
<hipboi> hmm
<rellla> hipboi: willswang made the version with cedarx-support: https://github.com/willswang - based on the latest binarys
<rellla> further improvement will only have success, if community gets newer/ fixed binarys or even some sources. same for xbmc.
<hipboi> rella: i know, willswang is a nice guy, Chinese
<rellla> if so, you should invite him to come here ;)
rz2k has quit []
merbanan has quit [Remote host closed the connection]
eebrah has joined #linux-sunxi
<hramrach> well, you could also hack the wrapper to set up the output to diffeerent sizes
<hramrach> so that you can do windowed vlc
<hramrach> and perhaps figure out if the output can be resized dynamically
<hramrach> so when user resizes window video follows
<hramrach> I heard some rumours about CedarX OpenMAX
<hramrach> but not where it is available and what it actually supports :s
<rellla> hramrach: around a year ago i heard some rumours, that aw thinks about supporting the community way better ... ;)
eebrah has quit [Read error: Connection reset by peer]
<Turl> mripard: yup :)
<Turl> mripard: do you have a branch I can rebase to with the latest uart stuff?
<oliv3r> mripard: what is latest kernel status? can I boot it on my tablet? console only is fine; i have usb keyboard.... a wait, no usb support, so i'd need to ssh in, which requires wifi, which is usb :S crap nvm
<Dreadlish> oliv3r: root on sd /etc/rc.local
<Dreadlish> oliv3r: etc.
<oliv3r> Dreadlish: what do you mean?
<Dreadlish> oliv3r: eh, if you have root on sd, you can add something to /etc/rc.local and it will run on boot ;)
gzamboni has quit [Ping timeout: 245 seconds]
<hipboi> glmark2 score 144
<Turl> oliv3r: you can use uart
<mripard> oliv3r: for now, it has UART, timers, IRQ, muxing and GPIOs working
<Turl> (as i/o, so no usb keyboard needed)
<mripard> clock should be here pretty soon thanks to Turl
<mripard> and hopefully emac as well, if I find some time to work on this
<Turl> mripard: see my question above
brian_ has joined #linux-sunxi
<mripard> Turl: ah yes, sorry
<mripard> hmmm
<mripard> 2s
ganbold_ has joined #linux-sunxi
gzamboni has joined #linux-sunxi
<mripard> Turl: just pushed it on github
<Turl> mripard: thanks
<Turl> mripard: I'll send the gates series in a minute
<mripard> ok :)
<Turl> I wanted to rebase because of uart<->serial
<Turl> mnemoc: bitbucket feels a tiny bit speedier
<Turl> mripard: are you planning to merge all these to your tree and then send a pull req to arm-soc?
Dave77 has joined #linux-sunxi
<mripard> Turl: I'll merge your clock stuff through my tree yes
<mripard> if Mike's ok with this
<mripard> and the dt part of the uart serie as well
<mripard> I'll probably send a first pull request today
<oliv3r> tablet doesn't have uart :S and i haven't received my uart breakout board
<mripard> oh, you're sure it doesn't have a UART?
<mripard> even some unsoldered pins?
<oliv3r> there are some
<oliv3r> but haven't been able to probeit
<oliv3r> i mean i have scoped it, but they all just give 250 mV - 300 mV
<oliv3r> gonna open up the back (put foto's on wiki)
<mripard> ok
<oliv3r> if i only had the breakout board :p
<mripard> what SoC is in it?
<oliv3r> A10
<oliv3r> linux-sunxi.org/BORA
<mripard> ok, to bad :)
<oliv3r> (it's not a bora but that's the easiest link)
<oliv3r> what's wrong with the A10?
<mripard> nothing
<mripard> I'm just searching for someone that could test my A31 patches
<mripard> :)
<oliv3r> i wanted to start work on my touchscreen and the G sensor; i figured best start would be the 3.8 tree
<Turl> mripard: ZaEarl maybe
<oliv3r> i want a an a20 tablet
<Turl> but he probably has tablets, not dev boards
<mnemoc> hopefully hipboi will make some prototypes of his a31 device soon
<oliv3r> forget a31!! :p
<Turl> quad core sells 2x :p
<oliv3r> ok, i don't get this, i installed a crosscompiler on my desktop now, trying to build (over nfs) the same tree I always build. did a make clean of acourse, but now get this: http://pastebin.com/KcyGS64w
<oliv3r> Turl: i'm praying for A40, quad core A10
<mripard> oliv3r: you forgot CROSS_COMPILE maybe ?
<oliv3r> exact same gcc version
<oliv3r> nope, i learned that mistake :)
<Turl> it's a header file on your PC that's broken
<Turl> according to the error
<oliv3r> yeah, maybe file conflict?
<oliv3r> i normally build on my server
<oliv3r> so no gui stuff
<mnemoc> oliv3r: we can't forget A31 even if it has a crap gpu
<mripard> the point is: who cares about GPU ? :)
<oliv3r> mnemoc: of course not, it even so still uses a very similar soc, so support is almost free
<oliv3r> mripard: xbmc/tablet :(
<oliv3r> mripard: libv :p
<mnemoc> oliv3r: the ar100 core will give us a big headache
<oliv3r> mnemoc: true
<oliv3r> Turl: ok my server doesn't have image.h; on my desktop, it comes from 'lensfun'
<oliv3r> so u-boot probably checks for its existance
<mnemoc> MAGIC_VER_A10SA = 0xA10A, <--- f* allwinner's per-soc tree
<hipboi> i have an A31 tablet at hand
<mnemoc> obviusly the sun5i maintainer chosen to ignore that A10 exists
<oliv3r> *facepalm*
<hipboi> mripard: what patch
<oliv3r> a208dc4d09a79a
<oliv3r> know bug :(
<hipboi> mripard: ok, will test tomorrow
<mripard> hipboi: I can send you a kernel image to boot, if you tell me to what UART you have access to
<hipboi> i uart0
<mripard> ok
<hipboi> need to go now
<mripard> it's not in a hurry, I'll ping once I have something that might have a chance to boot :)
<hipboi> ok
<mripard> thanks :)
lorenx has joined #linux-sunxi
<hipboi> back later
hipboi has left #linux-sunxi ["Leaving"]
gzamboni has quit [Ping timeout: 264 seconds]
<oliv3r> mnemoc: it turns out the openwrt guys fell into the same trap, i don't think it's abug with either; but it could possibly be a u-boot bug or sunxi-bsp bug
<oliv3r> Turl: if i want to upgrade my kernel on my tablet, which runs 3.0.26+ right now
<oliv3r> i'll probably have to extract boot.img, safe the initramfs, replace the kernel to begin with
<oliv3r> if the initramfs also holds modules, replace those aswell obviously (never seen the android initramfs)
<oliv3r> any userspace files that need to be replaced?
<Turl> oliv3r: get boot.img, unpack, repack with new kernel, copy to device
<Turl> no, the modules are not on initramfs
<oliv3r> Turl: and how much can be used from sunxi-bsp; since it's broken atm?
<Turl> they're on /system/lib/modules usually
<oliv3r> yeah saw those; i guess our 3.0.62+ kernel would be compatible load older 3.0.26+ modules?
<oliv3r> wanna replace modules 1 at a time (not all are liberated yet)
<Turl> you'll also need to upgrade libMali, libUMP and the rest of its friends on /system/lib/egl to match the new mali kernel module
<Turl> oliv3r: no, as a rule of thumb, you can't load old modules
gzamboni has joined #linux-sunxi
<oliv3r> Turl: well i'll have to :p no touchscreen otherwise! :p
<Turl> there's some really ugly hacks you can do to do it, but preferably just skip those who are not liberated for a first run
<oliv3r> Turl: do i have to configure the bsp with the _android profile?
<oliv3r> well yeah
<Turl> the _android profile is to build android
<Turl> but I suppose your tablet is not supported
<oliv3r> ah, kernel is the same?
<oliv3r> Turl: i run into the boot.scr error
<Turl> personally I'd just build the kernel manually :)
<oliv3r> Turl: my tablet is a copy of, so I can use a supported profile
<Turl> you don't need uboot or any other thing built
<oliv3r> Turl: well i think it does build the kernel just fine, u-boot too; but fails at some later stage
<Turl> oliv3r: unless it's a copy of zatab it's not supported :p
<oliv3r> lol
vicenteH has quit [Ping timeout: 260 seconds]
<oliv3r> what is the boot.scr error anyway?
<Turl> idk, what boot.scr?
<Turl> you don't need to build uboot anyway
<oliv3r> Turl: well if i want to run android from SD-card i do :p but i have u-boot and u-boot SPL building fine
<oliv3r> i'll paste it once my kernel is done compiling :)
<mripard> Turl: if you have some time, cooking a patch to address mike's comment might be good too
<mripard> no hurry, it will be a followup patch
<mripard> but still
<Turl> mripard: not today, but be sure I will
<mripard> I told you, you have some time :)
<Turl> mripard: It'll be something resembling the composite clock but for sunxi, to keep code duplication to a minimum
<mripard> even if it's in a month, it's fine for me
<oliv3r> Turl: cp: cannot stat 'build/boot.scr': No such file or directory; just after Android hwpack
<mripard> or we can wait for the clk-composite to be merged
<Turl> mripard: yeah but it doesn't support fixed rate :) just mux, gate and div
<mripard> ah, I see
<mripard> well, do it your way then :)
<mripard> but I guess, for everything but the oscillators, we can use it, right?
<Turl> mripard: not really either hm
<Turl> mripard: the other duplicated clocks if you wish are gates and factor clocks
<mripard> ah, so it's div, mux and gate, not div, mux or gate ?
<Turl> there are one or two that are gate+div, we could use it for those I suppose
<Turl> mripard: it's div and/or mux and/or gate iirc
<Turl> ie, any 1,2 or 3 out of the 3
<mripard> well, it's fine then, all the clocks we have are div+gate, right?
<mripard> (even though gate only have one child clock)
<mripard> +some
<Turl> I need to look again into it, but I'm pretty sure some were factor clocks and not divs
* Turl is writing cover letter :)
<mripard> aaaah, yes, I forgot factors, you're right
<oliv3r> i documented the A10 CCM on the registers page
rellla has quit [Quit: Nettalk6 - www.ntalk.de]
<oliv3r> i addeda ll sorts of fancy formulas and tried to be as verbose as possible about the clocks in the A10
<Turl> oliv3r: yeah I used the page
<oliv3r> of course i forgot the details :p but it's an easy place to refresh memory
<mripard> Turl: or you can send a patch to extend clk-composite to have fixed-rate and factor clocks :)
<Turl> mripard: yeah, it could be for fixed rate
<Turl> factor clocks are sunxi specific though
<Turl> we need it merged first anyway :)
<Turl> mripard: I also need to find out if there's anything that feeds directly from axi/ahb/apb and not from a gate
<mripard> ah, yes
<Turl> mripard: ok, gates sent
<mripard> yep, saw that :)
<Turl> hm, I forgot to cc the aw people
<Turl> I knew I was missing something :)
rellla has joined #linux-sunxi
gzamboni_ has joined #linux-sunxi
ZaEarl has joined #linux-sunxi
<oliv3r> Turl: why would however made this image, store all modules in /system/lib/modules; and not in their own subdirs? any advantage?
<Turl> oliv3r: ease of use? you can just insmod /system/lib/modules/$name.ko and be sure that's it :)
hipboi has joined #linux-sunxi
<oliv3r> i see the sunxi-bsp hwpack does it too
<oliv3r> well with lib/modules/kernel.version/ i could have two sets of modules :P
<Turl> wb hipboi
gzamboni has quit [Ping timeout: 264 seconds]
bsdfox has joined #linux-sunxi
<hipboi> Turl: back at home
shineworld has quit [Quit: Leaving]
<Turl> :)
<Turl> it's my turn to leave now :p
<Turl> see you all later
lorenx has quit [Quit: irc2go]
techn_ has joined #linux-sunxi
<oliv3r> but no turl! i have one more question!
<oliv3r> if it's ok to dd the created file to nandd
<oliv3r> oh wow, i finally 'fixed' my broken boot image. on nanda/linux/linux.bmp there was a boot image, but it was 'broken' I always thought that's the way it should be, so now i just replaced the image, and tada, it works
<oliv3r> there's more images, like you could multiboot based on an image you can click
eebrah has joined #linux-sunxi
rellla has quit [Remote host closed the connection]
mdfe has joined #linux-sunxi
eebrah has quit [Ping timeout: 252 seconds]
<bsdfox> anyone using gpio with a10?
n01 has quit [Quit: leaving]
<oliv3r> think so
ganbold__ has joined #linux-sunxi
ganbold_ has quit [Remote host closed the connection]
<libv> mdfe: ah, right, i also cannot access this document directly
<libv> let me get an indirect route
<libv> then "internal photos"
<libv> So direct dl is not possible, i first thought this was because of Turl his location that he was not allowed to access it
<libv> mdfe: did the device arrive yet?
<mdfe> yep
<libv> nice :)
<mdfe> but just the device
<libv> no time tonight to gut it though :)
<libv> mdfe: with the one usb->uart adapter we should be able to locate rx/tx already :)
<libv> and solder some cabling in place
<mdfe> usb->uart stuck somewhere in china
<libv> mdfe: what are your first impressions on the mid7042 btw?
<mdfe> hm
<libv> mdfe: really? ouch.
<mdfe> damn german seller
<libv> i bought mine in june or so, so i do not remember what happened there
<mdfe> however, I will try to deploy it usb otg will work out of box
<mdfe> with some luck
Dave77 has quit []
ibrah has joined #linux-sunxi
vicenteH has joined #linux-sunxi
hipboi has quit [Quit: Leaving]
<E3-work> mnemoc: accepted ive no clue, but that fix problems here too
<E3-work> (two different ones)
E3-work is now known as Epsylon3
eebrah has joined #linux-sunxi
<Epsylon3> trying CONFIG_DEBUG_INFO_REDUCED on 3.0.62
<Epsylon3> look like this problem has a large range
<Epsylon3> yep, its the same
traeak has quit [Ping timeout: 264 seconds]
ibrah has quit [Ping timeout: 245 seconds]
traeak has joined #linux-sunxi
<mnemoc> Epsylon3: as it's not mainlined, can you please submit it to our ML? can't pick it directly in that case
<Epsylon3> add .patch to the url
<Epsylon3> but hmm
<Epsylon3> only the gadget makefile is required
<Epsylon3> because that doesnt fix 3.0.62 with CONFIG_DEBUG_INFO_REDUCED :p
<mnemoc> Epsylon3: I know how to get a patch from github. The thing is that it's not "my repo", so commits need to go to the mailinglist so other devs have the chance to comment
<mnemoc> if it's a mainlined "critical bugfix", I can skip that
<mnemoc> but it's not the case
<Epsylon3> yep nevermind
<Epsylon3> will remove the reduced one
<Epsylon3> only keep gadget
<Epsylon3> so, no i dont send it on ML
<mnemoc> but don't get me wrong, I do want the fix
<mnemoc> ok
<Epsylon3> reduce one is another bug, related but not the dwarf one
<Epsylon3> and this flag is not our default
<Epsylon3> and this patch is only required on the kernel with f_fs, not 3.0
<Epsylon3> hmm
<Epsylon3> the file is here, but seems not used
<Epsylon3> g_ffs.c:66:#include "f_fs.c"
<Epsylon3> :p
<Epsylon3> weird that only happens on 3.4 branch
<Epsylon3> no, its because :
<Epsylon3> # CONFIG_USB_FUNCTIONFS is not set
<Epsylon3> :p
ibrah has joined #linux-sunxi
eebrah has quit [Read error: No route to host]
techn__ has joined #linux-sunxi
<Epsylon3> hmm is there a way to put the device in FEL mode from uboot console ?
techn_ has quit [Ping timeout: 246 seconds]
<Epsylon3> nvm, got it, look like we need to press during boot, and not hold it
von_fritz has joined #linux-sunxi
<mnemoc> Epsylon3: afaik u-boot is already too late for jumping to FEL
ibrah has quit [Ping timeout: 264 seconds]
<mnemoc> Epsylon3: but there is a mini bootloader for mmc that jumps directly
<Epsylon3> tx, i understand better now
<mnemoc> Epsylon3: see fel-boot in sunxi-tools
<Epsylon3> awusb
n01 has joined #linux-sunxi
<mnemoc> once in fel mode you can use either livesuit with the retard kernel driver for fel, or the libusb based tools
<mnemoc> fel-boot is only work around to trigger FEL in devices which it isn't trivial
paulk-desktop has joined #linux-sunxi
<Epsylon3> and awusb ?
<Epsylon3> kernel module
eebrah has joined #linux-sunxi
ganbold___ has joined #linux-sunxi
ganbold__ has quit [Ping timeout: 245 seconds]
<mnemoc> Epsylon3: yes, awusb is the "why? oh, why?" kernel module to enable livesuit for ubuntu64
<Epsylon3> that switch to fel mode too ?
<Epsylon3> from os ?
<mnemoc> no no
<mnemoc> you always need to boot into fel yourself
<Epsylon3> ohhh
<mnemoc> *once* in fel mode, you can pick how to deal with fel
<Epsylon3> its for host ? lol
<Epsylon3> hmm
<Epsylon3> will read the code
<mnemoc> "fel" is a usb slave thing, that lets you read/write/call mem from your computer using usb
<Epsylon3> yep i see, motorola use that for upgrades
<Epsylon3> some kind of serial mode
<Epsylon3> at bootloader stage
<mnemoc> awusb is a kernel driver that detects the id of an allwinner device in fel mode nad "enables" livesuit
<Epsylon3> ok, for host so
<mnemoc> yes
<mnemoc> in sunxi-tools we have a command line `./fel` that uses libusb to let you script your way into the device
traeak has quit [Ping timeout: 248 seconds]
<mnemoc> the idea was to split a lib out of it and then use that to implement a (free) flashing tool
<Epsylon3> i need to doublecheck, but i suspect some adb problems with the sunxi:udc: fix gadget init after loading as module
<Epsylon3> but i maybe have also some usb communications problems
<Epsylon3> fastboot locks sometimes
<Epsylon3> adb push also
<Epsylon3> so im on the rel2 system.img with my modules only
<Epsylon3> (and kernel)
<Epsylon3> [130481.551121] usb 3-2: not running at top speed; connect to a high speed hub
<Epsylon3> [130481.551124] usb 3-2: no configurations
<Epsylon3> [130481.557203] usb 3-2: can't read configurations, error -22
traeak has joined #linux-sunxi
christopher has joined #linux-sunxi
<Epsylon3> on device : [ 174.200000] WRN:L1367(/repos/cubieboard-tv-sdk/kernel/allwinner/common/drivers/usb/sun4i_usb/udc/sw_udc.c):handle_ep0: ep0 setup end
ibrah has joined #linux-sunxi
rz2k has joined #linux-sunxi
<mnemoc> (assuming it's genuine)
<Epsylon3> same price for the samsung 16gb uhs 1 on amazon
<Epsylon3> 48MB/s
<Epsylon3> :p
<Epsylon3> and got free shipping
<mnemoc> real?
<mnemoc> uhm... nice prices has amazon on sdhc cards these days
<Epsylon3> 12+3 to 16
<Epsylon3> so yes, its in the range
<Epsylon3> some transcend are cheaper
<Epsylon3> 11 + free shipping
<mnemoc> no free shipping in .es
<Epsylon3> oh right, its wrote free shipping above 15€
<Epsylon3> its why i got the 16€ one :p
<Epsylon3> but i need to test it in the cubie for real perf checks
<Epsylon3> lets do it
<Epsylon3> arf, cant do it
<Epsylon3> media_rw
<Epsylon3> sigh
<Epsylon3> benchmark tools only handle sdcard_rw
<Epsylon3> well later... after the system.img upload
<Epsylon3> need to download it :p
<Epsylon3> via ethernet
<Epsylon3> # busybox wget http://192.168.10.120/system.img
<Epsylon3> Connecting to 192.168.10.120 (192.168.10.120:80)
<Epsylon3> system.img 27% |******** | 49637k 00:00:13 ETA
<Epsylon3> lol
<Epsylon3> during android system run :
<Epsylon3> # dd if=system.img of=/dev/block/nandd bs=4096
<Epsylon3> lets see the result :p
<Epsylon3> 184543480 bytes transferred in 23.283 secs (7926104 bytes/sec)
<Epsylon3> not good
<Epsylon3> [ 6.660000] init: cannot find '/system/bin/sh', disabling 'console'
<Dreadlish> Epsylon3: run recovery and push busybox ;d
<Epsylon3> yea ;)
<Epsylon3> tx for the tip
<gzamboni_> is github.com/allwinner the official's allwinner git ?
<gzamboni_> i cant understand its dates. for the CSI driver it shows v1.64 but all the changelog doesnt match with other gits changelog.
<mnemoc> tom created /allwinner
<mnemoc> in the stone age when the android 2.3 sdk was released
<mnemoc> https://github.com/allwinnerwk otoh is used by allwinner to share code with their customers
<gzamboni_> so the csi version 1.11 is the continued v1.64 from allwinnners
<n01> hno, slapin_nb any news for the MTD driver?
<mnemoc> n01: https://github.com/yuq/sunxi-nfc-mtd is the closest thing to an mtd driver we have
<n01> mnemoc: yes I know that module, but without specification for NFC it is really hard :)
<mnemoc> hno and slapin invested months in trying to read the nand without dma
<gzamboni_> yes, i think i understand now, the v 1 from your git is the CSI:Initial version for linux 3.0.8 1) Ported from linux2.3.36
<mnemoc> and then it seems they ran out of steam
<n01> I can imagine
<mnemoc> gzamboni_: we had a 2.6.36-based repo, and then a forward port for that to 3.0
<mnemoc> gzamboni_: but when qware released their sdk we trashed the past and started over
<gzamboni_> ok
<mnemoc> stuff still in my github
<gzamboni_> mnemoc: i see
<gzamboni_> mnemoc: so you use your personnal git for testing before merging it to the testing branch of the sunxi ?
shineworld has joined #linux-sunxi
<mnemoc> linux-sunxi is less than a year old. everything used to be in my github
<mnemoc> and I keep there the stuff that is for reference only, and my personal mess ups
<gzamboni_> i remember, just want to understand why the csi in your git is on v. 1.5 and in the sunxi is on v. 1.1
<mnemoc> what branch has 1.5?
<mnemoc> gzamboni_: that's a source drop
<gzamboni_> import/lichee-a10-3.0
<mnemoc> for reference to be reviewed and relevant stuff brought into sunxi's kernet
<mnemoc> from i-oink's source drop
<mnemoc> but now with the -dev branches (which have history) that's mostly historic junk
<mnemoc> lichee*/ and import/ branches are there for people (like you) to review allwinner changes and submit patches for the important bits to the sunxi ml
<techn__> anyone seen <3>>>> no handle, treat it handle over, iir 07
<mnemoc> techn__: that reminds me an old uart bug
<mnemoc> gzamboni_: explore the history of your favourite subsystem/driver across all the lichee*/*-dev branches, and use that to improve our tree
<gzamboni_> mnemoc: ok, will do so
<mnemoc> gzamboni_: keep in mind that the first goal is to unify all sunNi-foo drivers into sunxi-foo equivalents
<mnemoc> if you fix something in one sunNi-foo it's critical you do the same in the other
<mnemoc> unifying first makes that simpler
<gzamboni_> you're right
<mnemoc> sun[67]i changes are most likely also relevant to sun[345]i
<mnemoc> but maintained. they already abandoned the 3.0 tree
<techn__> mnemoc: found reason.. allwinner-dev-teams kernel is old and has UART defined in devices.c
<techn__> which is wrong type
<techn__> .. have to test it still
<mnemoc> can you change the manifest to use ours?
<techn__> Turl has added some pathes top of ours
<mnemoc> any you need?
<mnemoc> also, as we "want" to deprecate 3.0, it would be important to test/fix sunxi-3.4 to work on android 4.0
<techn__> there is some lowmemorykiller patch
<techn__> but dunno where it effects
<paulk-desktop> hi
<paulk-desktop> I see that A13 tabs have an android recovery partition -- how can I boot to it?
<techn__> paulk-desktop: I think it should work same as with a10
<Epsylon3> uboot :
<Epsylon3> nand read 40007800 recovery;boota 40007800
<techn__> but I dont know how it works with a10 :p
<Epsylon3> the command is in printenv
<Epsylon3> oh a13
<paulk-desktop> I don't have uboot prompt
<Epsylon3> adb reboot recovery
<paulk-desktop> (I only have USB here)
<mnemoc> iirc you can also write recovery to one of the dummy raw partition
<paulk-desktop> as a string?
<Epsylon3> yes
<Epsylon3> echo > /dev/...
<Epsylon3> the misc part
<Epsylon3> dangerous, if not the right one ;p
<paulk-desktop> nandf here
<paulk-desktop> it's full of 0xff
<mnemoc> in the worst case you can always reflash it ;-)
<Turl> techn__: the memory patch can be dropped
<mnemoc> Turl: so submit the stuff
<mnemoc> Turl: but I suppose your binary driver will go to hell if we deprecate 3.0....
shineworld has quit [Quit: Leaving]
<paulk-desktop> also, is the free uboot capable of booting from NAND?
<paulk-desktop> and can it boot a kernel on the sdcard?
<Turl> mnemoc: yeah :/ need to get source and/or RE it
<techn__> paulk-desktop: lichee-dev with couple modifications
<paulk-desktop> mhh ok
<techn__> paulk-desktop: if you talking about a13 and android
<paulk-desktop> yes
<mnemoc> techn__: will you submit those changes for lichee-dev?
<techn__> I'll try to send patches soon as I get things running
<techn__> and cleaned.. :p
<techn__> I'll add working livesuit support to BSP at same time
<mnemoc> :)
ibrah has quit [Ping timeout: 248 seconds]
von_fritz has quit [Quit: vonfritz leaves, don't panic]
Liberato1 is now known as Liberator
<oliv3r> i've build 3.0-android kernel, but it won't boot ( can't see why of course)
<oliv3r> mkbootimg seems to work, using split-bootimg.pl splits it and mkbootimg to merge it works too, so that step of the process is fine
<oliv3r> it just hangs at boot1 I guess (i see the uboot 'image')
<oliv3r> but it won't continue with a blank screen (booting the kernel)
<oliv3r> if I can get that to work, i can load a 3.4 kernel to see what that does
<Turl> oliv3r: does your screen clear or does it have the bootN ugly bitmap still?
<oliv3r> bootN ugly bitmap
<oliv3r> the little blue android thing
<oliv3r> well green and blue
<Turl> maybe you need to specify the base address to mkbootimg
<oliv3r> i suspect that the screen clear is the kernel booting, and the cm animation is actually android booting
<oliv3r> well i made a new image
<oliv3r> from the split image
<oliv3r> and that works without that
<Turl> if you have serial try enabling earlyprintk then
<oliv3r> tablet :(
<oliv3r> and dont't
<oliv3r> i do have a 'working' boot.img
<oliv3r> can't we extract it from there (if needed?)
<techn__> oliv3r: your u-boot propably doesnt support boota
<oliv3r> bootwhat?
<techn__> 'boota' method to boot android boot.img
<oliv3r> and we do it differently than this 3rd party kernel? :)
<techn__> bootimg is ICS method :/
<oliv3r> i have jb
<oliv3r> it's a Christian Troy's CM10 build
<oliv3r> Linux version 3.0.36+ (alan@Audrey) (gcc version 4.5.1 (Sourcery G++ Lite 2010.09-50) ) #46 PREEMPT Wed Jun 27 07:26:31 CEST 2012
<oliv3r> that's my kernel
<oliv3r> so it's 3rd party compiled
<oliv3r> so I should be able to compile it!
<oliv3r> ah, i do have a /proc/config.gz; would that help?
<oliv3r> hsould be able to roll my own kernel based on that config at the least
<oliv3r> can i see the offset on a running kernel?
gzamboni_ has quit [Ping timeout: 264 seconds]
gzamboni has joined #linux-sunxi
n01 has quit [Ping timeout: 256 seconds]
<Epsylon3> kinnect sdk ^^
<Epsylon3> with a 3D scanner
<Epsylon3> i think i will buy one
<oliv3r> ok using the same .config; i got a kernel half the size. hexdumping the first few lines, gives me a nice linux header with version info, the one I extracted has no such thing. it is a kernel, but it's 8mb and the kernel header doesn't show up till much much later in the file :S
<xenoxaos> debugging built in?
<oliv3r> must be
<oliv3r> he forked mnemoc's tree so that's no different
<oliv3r> but would that change the header?
<mnemoc> problem with forks is that rarely get sync'ed...
<oliv3r> well he could have used different build options; fine
<oliv3r> i used his .config
<oliv3r> so now the kernel _should_ boot
<traeak> Epsylon3: not kinect..there's that other new 3d thing that sits in front of your computer monitor pointing up
<traeak> that is precise
<Epsylon3> a new kind of kinect ?
<traeak> i can't even remeber the name of it...wai tone
<traeak> there we go
<Epsylon3> oh :)
<Epsylon3> hmmm
<Epsylon3> preorder since 1 year ?
<Epsylon3> could be a fake
paulk-desktop has quit [Quit: Ex-Chat]
<Epsylon3> Hype or Hope? We’ll Soon Find Out – Leap Motion “Minority Report” Controller Ships In May
<traeak> i suspect coding the firmware in this device or the driver is the headache
<Epsylon3> glasses ?
<Epsylon3> ive seen google glasses could be at 1500$ too
<Epsylon3> Those who are chosen through the promotion will need to purchase the Glass Explorer Edition for $1,500 and travel to one of three events to pick it up, which will take place in New York, San Francisco, and Los Angeles.
<Epsylon3> hmm ok
<Epsylon3> applications are closed
<Epsylon3> like the war thing new mmorpg
<Epsylon3> dont remember the name
<Epsylon3> open beta was the 9 march
<Epsylon3> no, was a charge test
<Epsylon3> oups, wrong chan ;p