Turl changed the topic of #linux-sunxi to: Allwinner/sunxi development discussion - Don't ask to ask. Just ask and wait! - See http://linux-sunxi.org | https://github.com/linux-sunxi/ | Logs at http://irclog.whitequark.org/linux-sunxi
kivutar has joined #linux-sunxi
Night-Shade has joined #linux-sunxi
<chil> libv compiling... but its taking forever. I'll go get some sleep and continue tomorrow! thx for help :)
techn__ has quit [Ping timeout: 252 seconds]
chil has quit [Remote host closed the connection]
rose28357 has quit [Remote host closed the connection]
FDCX_ has joined #linux-sunxi
newleaves has joined #linux-sunxi
FDCX_ has quit [Remote host closed the connection]
Seppoz has joined #linux-sunxi
<Seppoz> when i use usb 0 as otg host only i have to set drive manually for the usb to work, is that normakl? we do not have a detect circiiut its basically connected like all the other 2 usb
<Seppoz> also what is the actual purpose of the drive and hos is it used?
fredy has quit [Excess Flood]
<Seppoz> is the drive defined by usb standard?
<Seppoz> or is this some additional feature that i can use if i want to
<Seppoz> like resetting devices by taking away the power
fredy has joined #linux-sunxi
<oliv3r> wtf a nother sunxk talk
<oliv3r> tizen on allwinner devkces
<oliv3r> why havent i heard of this beford?
deasy has joined #linux-sunxi
rz2k has joined #linux-sunxi
parabyte has joined #linux-sunxi
_whitelogger has joined #linux-sunxi
FDCX has quit [Ping timeout: 260 seconds]
egbert has quit [Disconnected by services]
egbert has joined #linux-sunxi
Night-Shade has quit [Ping timeout: 272 seconds]
naobsd has joined #linux-sunxi
FDCX has joined #linux-sunxi
<mrnuke> mripard: hmm. booting your mainline kernel seems to either hang or have no console
parabyte has quit [Quit: Leaving]
earny has quit [Remote host closed the connection]
earny has joined #linux-sunxi
deasy has quit [Quit: Nom d'un quark, c'est Edmonton !]
samul has joined #linux-sunxi
<samul> A20?
bsdfox has quit [Ping timeout: 272 seconds]
bsdfox has joined #linux-sunxi
bsdfox has quit [Ping timeout: 276 seconds]
pwhalen has quit [Ping timeout: 260 seconds]
bsdfox has joined #linux-sunxi
pwhalen has joined #linux-sunxi
pwhalen has quit [Changing host]
pwhalen has joined #linux-sunxi
[7] has quit [Disconnected by services]
TheSeven has joined #linux-sunxi
bsdfox has quit [Ping timeout: 248 seconds]
bsdfox has joined #linux-sunxi
<libv> oliv3r: same thing for the coreboot port
<libv> the tizen-sunxi community?
Gerwin_J has quit [Quit: Gerwin_J]
bsdfox has quit [Ping timeout: 246 seconds]
bsdfox has joined #linux-sunxi
hero100 has joined #linux-sunxi
mrnuke has quit [Ping timeout: 248 seconds]
mrnuke has joined #linux-sunxi
bsdfox has quit [Ping timeout: 248 seconds]
bsdfox has joined #linux-sunxi
Quarx has joined #linux-sunxi
pacopad has left #linux-sunxi [#linux-sunxi]
sgo11 has joined #linux-sunxi
<sgo11> hi, with cubietruck, the cubieboard lubuntu release has all the modules bulit-in, but they are not loaded at boot. I added them to /etc/modules and reboot, here is the current lsmod output: http://paste.ubuntu.com/6731217/ There are still many problems. first, the 2d/3d acceleration is not working. glxinfo shows "OpenGL renderer string: Software Rasterizer". how to fix?
pwhalen has quit [Ping timeout: 276 seconds]
deasy has joined #linux-sunxi
bsdfox has quit [Ping timeout: 272 seconds]
bsdfox has joined #linux-sunxi
bsdfox has quit [Ping timeout: 252 seconds]
<sgo11> The sound does not work properly either. "espeak hello" works. but vlc gave me errors "cannot open ALSA device "default": No such file or directory" "annot open audio device (/dev/dsp)" "no suitable audio output module".
<sgo11> alsamixer gave me "cannot open mixer: No such file or directory". I am in the audio group.
bsdfox has joined #linux-sunxi
bsdfox has quit [Ping timeout: 272 seconds]
bsdfox has joined #linux-sunxi
<sgo11> ok, for the sound issue, after changing asound.conf from card 1 to card 0. alsamixer can be run. vlc gave me a different error "segmentation fault". try to figure out why.
techn_ has joined #linux-sunxi
notmart has joined #linux-sunxi
<sgo11> sound issue is fixed. 2d/3d acceleration problem still exists
netlynx has joined #linux-sunxi
fredy has quit [Excess Flood]
cubear has joined #linux-sunxi
newleaves has quit [Quit: Leaving]
fredy has joined #linux-sunxi
<sgo11> can anyone help me on 2d/3d hw acceleration please?
<rz2k> glxinfo shows "OpenGL renderer string: Software Rasterizer". how to fix?
<rz2k> why do you think that glxinfo showing that is wrong?
<rz2k> your sunxi device has no OpenGL acceleration
<rz2k> it has OpenGLES acceleration
<rz2k> also to run cubietruck with Mali-400 you need latest stage kernel, cubietruck had couple interesting issues with memory that were fixed couple weeks ago, you will likely have no 3d in the images from cubietech.
<sgo11> rz2k, thanks a lot for your reply. I am newbie on this. so how to make sure my hw acceleration works? how to check if i have the latest stage kernel or not? my current uname output "Linux cubietruck 3.4.61+"
<sgo11> rz2k, I am just about to chmod 777 to /dev/ump and /dev/mali. should I do that?
hani has joined #linux-sunxi
<rz2k> for OpenGL you should leave the software rasterizer, if you want OpenGLES - you need latest kernel https://github.com/linux-sunxi/linux-sunxi/tree/stage/sunxi-3.4
<rz2k> are you sure you need GLES?
<sgo11> rz2k, I just want to play 1080p videos. currently, I even can't play 720p video. I know nothing about hardware. I just want to play videos in both mplayer and chromium and faster lxde etc.. so do I need gles? thanks a lot.
<rz2k> hmmm
<sgo11> currently, my free memory is 573 out of 1825. I don't understand this either. my lxde is too slow. I don't run anything. the memory used up very quickly.
<hani> rz2k : i've some Qs about the yuq mtd driver since u had some experience with it in that past
<rz2k> first, to play 1080p you need CedarX acceleration, second, you can install one of three options: VLC media player with CedarX support, XBMC with CedarX support (needs OpenGLES), libvdpau-sunxi library and any vdpau player (mplayer)
<rz2k> hani: ask freely
<rz2k> sgo11: or if you need only the videoplayer, you can just run the Android
<sgo11> rz2k, my vlc currently has segmentation fault error when play mp3 music. I have no idea why. mplayer can play mp3 fine. PS: my goal of cubietruck is to replace my laptop. so android is not sufficient for this purpose. btw, how to install/enable CedarX acceleration? thanks a lot.
<hani> I'm trying to use the u-boot in the yuq repo I'm facing an issue with the implementation of the PACK image wondering if u know anything about it
<sgo11> rz2k, I just cvh
<rz2k> sgo11: you need to compile your own VLC there are page at wiki, for cedarx acceleartion you only need the programs who can use cedarx.
<rz2k> s/there are page/there are page about it/
<rz2k> hani: honestly no idea, didnt try the images with mtd
<rz2k> sgo11: try libvdpau-sunxi first probably, it can only run h264 and mpeg12, but it will work in any player (most common is mplayer).
<hani> r2zk: I've manged to build u-boot and spl and flash it into the nand but was not able to load the kernel
<rz2k> hani: sorry, I only debugged mtd and faced some issues there for NANDs that are use on Olinuxino boards
<rz2k> it is too early to do SPL support when we have MTD failing on most popular board.
<sgo11> rz2k, thanks. how to install libvdpau-sunxi ? and besides the 1080p play, how to make lxde run faster? do I need the latest kernel or GLES? thanks.
<sgo11> I also want to play videos in chromium. youtube etc... will that lib help? thanks.
<rz2k> no
<rz2k> I have no idea what chromium uses to play videos
<rz2k> for lxde you probably have a cpufreq issue, read http://linux-sunxi.org/Cpufreq
<sgo11> rz2k, sorry about too many questions. I am new to arm and hw. my current free mem is 83M. Total is 1968M. no idea. Does cpufre issue cause this? thanks.
<rz2k> install htop and check who eats the RAM
<sgo11> I only run xchat and chromium. ok. let me see.
<hani> rz2k : Can u please elaborate on those failures?
<rz2k> hani: 1) it is not clear how to properly use the OOB area and how to represent it to MTD layer when running hardware ECC. 2) badblock detection does not always detect bad blocks correctly
<rz2k> first issue causes UBIFS and other filesystem using OOB to fail after 10-50 reboots
<rz2k> filesystems*
<rz2k> some guys from Italy mailed me with exactly the same issue with Yuq driver, but timeframe for them was 100 reboots
<sgo11> rz2k, I tried the second script from Cpufreq wiki page. lxde does not seem faster.
<hani> r2zk : So it is not ready for production yet, how about the current NFC is there any limitation?
<hani> rz2k I ment lichee implemenation
<rz2k> sgo11: unfortunately, I cant be responsible for the cubietech images because they did not consult with us when they made them. I dont know what version of Linaro/Ubuntu they use, what modifications they made and what eats your RAM there.
fredy has quit [Excess Flood]
<rz2k> sgo11: I did not encounter any hardcore slowness (response >2sec) with just standrad linaro-alip and selfcompiled kernel. also, do you run your system from SD card or NAND?
<rz2k> hani: no, current NFC works ok, but it is sort of slow and ugly coded.
<rz2k> hani: example - you cant have partition 4GB+ because it uses something like FAT16 as partition table
fredy has joined #linux-sunxi
<sgo11> rz2k, I am very sorry. the memory usage might be fine. I read the wrong number. I used free -m, I think I should read -/+ buffers/cache line instead of the first line. it's 785M out of 1968M. I am running from sandisk class10 SD. When I say it's slow, it's slower than my 4 years old laptop. When I open chromium tabs, input texts in chromium, when I resize windows, they are slow.
<hani> rz2k : h've had a chance to test Boris BREZILLON patches
<rz2k> sgo11: try to complete the installation of Xorg drivers wiki page http://linux-sunxi.org/Binary_drivers, they help with the resizing issues and they also provide support for Xvideo
<rz2k> 785M out of 2G is likely your chromium
<rz2k> sgo11: also read a paragraph under "Setting up X", it will answer your 2d/3d questions
<sgo11> rz2k, thank you very much for the help. I still have some silly questions if you can answer. (1) if I followed the wiki and install sunxifb driver, do I need the latest stage kernel? you said I need the latest kernel to enable GLES. I don't really know what it is. (2) does 2d/3d hw acceleration make lxde and UI apps faster? (3) I modified the default /etc/modules, http://paste.ubuntu.com/6731778/ can you please have look, is this OK? thanks.
<sgo11> PS: (4) Do I need to sudo chmod 777 /dev/ump and /dev/mali ? I read some posts about this. their current permission are "crw-rw----"
sgo11 is now known as sgo11_away
FunkyPenguin has quit [Ping timeout: 260 seconds]
<rz2k> etc/modules cant hurt much
bsdfox has quit [Ping timeout: 246 seconds]
<rz2k> lxde does not depend on 3d support, you need Xorg driver from that page which has 2d acceleration support
<rz2k> you should have latest kernel, with your current you will likely have 2d acceleration, but 3d acceleration will not work
FunkyPenguin has joined #linux-sunxi
<rz2k> since you only need the first one, you can leave it
<sgo11_away> rz2k, ok. got it. for the latest kernel, Do I need to compile it from source? are there any binary version that I can use directly? thanks.
sgo11_away is now known as sgo11
<rz2k> mnemoc:
<rz2k> mnemoc: where was the automated builder again? :p
<rz2k> sgo11: we had an automated builder somewhere, but I forgot where.
<sgo11> rz2k, ok. thank you very much for your help. ^_^
<rz2k> you are welcome, usually there are more guys around here, but today its seems like its only me.
<rz2k> you can also ask at our mailing list
<rz2k> this is the latest kernel for sun7i (A20)
<sgo11> rz2k, thanks a lot. Do I just "cp -a" to replace the old ones?
<rz2k> back up the old one
bsdfox has joined #linux-sunxi
<rz2k> if it wont work for you for some reason
<rz2k> you will be able to revert
<sgo11> rz2k, yeah, will do that. thanks.
<rz2k> hani: sorry, lost your message, and how was the performance from Boris patches?
<sgo11> mnemoc, thanks a lot too.
<rz2k> hani: I read that they dont use DMA and HW ECC
<mnemoc> yw
<rz2k> sgo11: for the chmod 777 /dev/mali, you need to setup the udev rules, read here http://linux-sunxi.org/Binary_drivers#Mali_kernel_driver
<rz2k> it will do the needed chmod/chgrp for you at module load
<sgo11> rz2k, oh, I have one last question. once I have the latest kernel, do I need to do anything to enable GLES?
<rz2k> go exactly as binary drivers page says
<sgo11> rz2k, ok. got it. I will follow all the steps in that Binary_drivers wiki page. thanks.
sgo11 is now known as sgo11_away
FreezingCold has quit [Ping timeout: 272 seconds]
rellla has joined #linux-sunxi
<hramrach> hello
<hramrach> if we had a mainline nand driver
<hramrach> what would be a resonable way to represent the bootloader area?
<hramrach> like use a commandline partition table? or make it part of the sunxi driver ?
<hramrach> would it be reasonable to provide an option to stora data on whole nand without the boot area - like when you have another flash you boot from or w/e
<hramrach> you could detect the bootloader signature in the driver and reserve the bootloader are a when present but how do you write the boot area on a blank nand then?
enrico__ has joined #linux-sunxi
bsdfox has quit [Ping timeout: 246 seconds]
_massi_ has joined #linux-sunxi
Night-Shade has joined #linux-sunxi
bsdfox has joined #linux-sunxi
<hani> rz2k : I'm doing some testing now will give some feed back soon
bsdfox has quit [Ping timeout: 252 seconds]
_massi_ has quit [Quit: Leaving]
<hani> hramrach : I suggest you read about the boot process of the Allwinner chips
<hani> hramrach :BROM BOOT0 BOOT1 UBOOT
bsdfox has joined #linux-sunxi
<mnemoc> goal is to replace boot0/boot1 with u-boot-spl after mtd-sunxi matures enough
wingrime has joined #linux-sunxi
<rz2k> we atleast now have something working horribly slow but working
<rz2k> thanks to Boris
<mnemoc> and it is technically possible to make a sunxi-nand block device and ans sunxi partition table to keep compatibility with allwinner's proprietary system, but the benefit of that compared to the required effort is very questionable
<hani> rz2k : He seems to be in the right path
bsdfox has quit [Ping timeout: 272 seconds]
<hani> r2zk :he made a good progress but it is too early to compare against yug's
<rz2k> we need dmaengine
<rz2k> and all the channels from the softwinner dma
Black_Horseman has joined #linux-sunxi
Black_Horseman has quit [Changing host]
Black_Horseman has joined #linux-sunxi
<mnemoc> and then the wearing protection layer
rose28357 has joined #linux-sunxi
<wingrime> rz2k: have you saw new AW dma engine driver?
bsdfox has joined #linux-sunxi
<rz2k> nope
<rz2k> some guy said that he will write it, but then he disappeared
FunkyPenguin has quit [Ping timeout: 252 seconds]
_newleaves_ has joined #linux-sunxi
FunkyPenguin has joined #linux-sunxi
FunkyPenguin has quit [Excess Flood]
FunkyPenguin has joined #linux-sunxi
<_newleaves_> test
ddc has joined #linux-sunxi
ddc has quit [Ping timeout: 252 seconds]
<montjoie[home]> still no comment on my driver:'(
<wingrime> rz2k:good news for you, in last a23 source code drop, aw implemented nice dmaengine driver, and all devices already integrated with it
<wingrime> rz2k: If you want you can only add dts , handyeye and mainline it
<wens> mripard said he would do it next?
Black_Horseman has quit [Ping timeout: 276 seconds]
jemk has joined #linux-sunxi
maz_ has joined #linux-sunxi
nove has joined #linux-sunxi
kivutar has quit [Ping timeout: 245 seconds]
Guest16470 has joined #linux-sunxi
Guest16470 has quit [Changing host]
Guest16470 is now known as netchip
chil has joined #linux-sunxi
<chil> libv I posted the 2 dmesg logs with debug enabled for the EDID clk issue: http://pastebin.com/gj8HU3Yk (1080p), http://pastebin.com/Dwbwu0KP (1200p)
<hramrach> what I want right now is stable version of the drivers we have - sata, stmmac, mmc, smp/cpufreq, usb, all the clocks required for that
<hramrach> so we have a workable baseline for the mainline drivers
kivutar has joined #linux-sunxi
sgo11_away is now known as sgo11
sgo11 has quit [Quit: Leaving]
<wingrime> hramrach: most important have working clock tree for that
<wingrime> hramrach: also, my IR driver are whait so
<libv> chil: i think the best option is to wait a bit, there is a much much easier driver to debug in the pipeline
<libv> chil: i sadly do not have access to a 1900x1200 display, but it might actually be that my kms driver "Just works"
<libv> as i use the video0 pll for the dotclock, per default
<libv> chil: the noisy log didn't really reveal anything new :(
<chil> :(
<chil> Well, at least is good to know that there are some updates comming
<chil> how can I follow the kms driver devel status?
hani has quit [Quit: Page closed]
<mripard> wingrime: what's missing?
<mripard> hramrach: it's a chicken-egg problem. We won't work on clocks that are needed, and they won't be needed unless we have a driver that requires it
<wingrime> mripard: IR clocks , I currenly no idea whats going on with mainline clocks, last time I used direct code to play with regs
<chil> libv thx! I can test it with my a20 board + display when the kms driver is available
sgo11 has joined #linux-sunxi
<libv> chil: i have a cubietruck and a 7" tablet with a20
<libv> chil: i just do not have a display capable of showing 1900x1200 :)
<rose28357> Chil,libv: looking far ahead we will have to deal with 4K HDMI @30 hz. The A20 is supposed to provide 4K !
<rose28357> ... And i want a UHDTV ;-)
<libv> rose28357: buy me a 4k monitor.
<libv> chil: ah, looking at the registers... you are still trying to run 1920x1200
parabyte has joined #linux-sunxi
<libv> chil: it's 1900x1200, not 1920x1200
<parabyte> u
<rose28357> It should be 1920x1200. Look here http://en.m.wikipedia.org/wiki/16:10
<libv> ah, you are right
<libv> chil: there's nothing obvious in the registers
<libv> at least not the ones dumped in these logs
_newleaves_ has quit [Quit: 离开]
<wingrime> libv: are you use regs directly ?
<wingrime> libv: without any aw code interaction?
<libv> wingrime: i am using the clock code
sgo11 has quit [Quit: Leaving]
<wingrime> libv: as I remeber disp have some clocks abstract layer..
<HdkR> libv: Apparently Seiki is selling a 39" 4K television that works well as a monitor for $500 :P
hero100 has quit [Remote host closed the connection]
<chil> libv thats why I said I can do testing if you need :)
<buZz> HdkR: seiko?
<buZz> weird
<libv> wingrime: i am using the standard aw clock infrastructure atm
<buZz> eh
<chil> libv also, if you need more debug/logs let me know
<libv> HdkR: does 4k at 30Hz apparently
<buZz> no hdmi input?
<HdkR> libv: Yea, bit disappointing but you take what you can get right? :P
<libv> HdkR: prices are dropping fast now
<libv> HdkR: proper 4k is available for under 1k
<libv> or will be soon
<HdkR> Apparently displayport can't support 60hz 4K until 1.3
<libv> heh, displayport
<libv> dell claimed displayport support for a 30"er back in 2007
<wingrime> libv: modern monitors looks ugly, in comparea with phablet DPI, 4k will be cool
<HdkR> HDMI 4K at 60hz is only available in 2.0
<libv> we ordered it at suse
<HdkR> Right
<libv> 1.6k...
<libv> on hold at all the time
<HdkR> jeez
<libv> 4 months later, in the weekly phone call with ati...
<libv> "oh guys... you know that monitor you ordered from dell..."
<libv> "dell apparently contacted us and asked whether they could use our displayport verification infrastructure"
<libv> oops :)
<libv> 3 months later, we got the monitor, and what a buggy PoS it was indeed :)
<parabyte> okay most random question of the year
<parabyte> Any of you guys seen a Bluetooth or usb coffee machine?
<HdkR> haha
<parabyte> if not i am sooooooo attaching a A13 or A10 to a coffee machine :O
<HdkR> I ordered a 21" monitor from dell with DP about 2 years ago, works fine now :P
<jelly-home> because everything's better with bluetooth?
<parabyte> bluetooth coffee machine ;)
<libv> HdkR: yes, they probably have set up their own testing equipment in the meantime :)
<parabyte> I got a viewsonic ips display less then a month old and its PSU is dying
<parabyte> luckily its external
<parabyte> !
<libv> parabyte: surely you can get that rma
<libv> -ed
<parabyte> i just shoved another psu on it
<parabyte> just surprised it failed
<libv> :)
<parabyte> the caps was hissing
<parabyte> was freaking my cat out!
<parabyte> like my A13 tablet its battery popped!
<parabyte> around the seams and let out some lovely battery juice
<buZz> parabyte: are you building one? (BT coffee machine)
<parabyte> buZz, looks that way
<parabyte> ;)
<buZz> parabyte: i always wanted to implement RFC2324
<buZz> is that your goal?
<libv> parabyte: if you've ever worked in any american style led tech company, and you know what stupidity goes on at management level, and what cheap idiocy has to be done at the bottom level to make up for the wastage... then you wonder why there isn't a fukushima every year and why we humans are alive and from time to time have something working
<parabyte> Actually i was reading about RFC2324 last week and yes i guess it is
<buZz> awesome :)
<buZz> the world needs RFC2324 implementations
<parabyte> lol
<parabyte> someone needs to do it
<parabyte> im thinking applet in my xfce panel that tells me when my coffee is done, and i press the applet to make a brew :)
<HdkR> haha
<HdkR> I don't know if that is smart or the epitamy of laziness
<HdkR> epitome even
<wingrime> jemk: ping
<wingrime> jemk: I just thinked, that between TRIG commands present in each sub-engine must be #reset or clean SRAM routine
ddc has joined #linux-sunxi
FunkyPenguin has quit [Ping timeout: 272 seconds]
FunkyPenguin has joined #linux-sunxi
ddc has quit [Ping timeout: 252 seconds]
enrico__ has quit [Read error: Operation timed out]
<libv> wtf...
enrico__ has joined #linux-sunxi
<libv> nand-part says 32768 blocks
<libv> vfat says 128MB
<libv> there is an 8x discrepancy there
<hramrach> mripard: I don;t think there are clocks missing. It's just there is not a tree with stable versions of the drivers. Especially the clock and smp patches floating around and randomly getting included and excluded in various test trees make merging the drivers into one tree troublesome
<libv> and if i will try to write to the nanda, i might actually kill the nandb fs
<hramrach> libv: did you re-read partition/reboot?
<libv> hramrach: many times.
<hramrach> :s
<libv> hramrach: i have been running off nandb for a few days already
shineworld has joined #linux-sunxi
<libv> it does explain why installing u-boot.bin writing sometimes f-ed up really badly though
enrico___ has joined #linux-sunxi
<hramrach> what does sfdisk say?
<libv> hramrach: on nand?
<hramrach> iirc it does have an option to just list sizes without interpreting the mbr
<libv> hramrach: you've never used nand, have you :)
<libv> ok :)
<hramrach> libv: I wrote the install tutorial
<libv> :)
<libv> ah, 16mb, as expected
<libv> but why does the fs think it is 128mb
<hramrach> so the contained filesystem is just wrong size
<libv> i never touched this partition or the fs on it
enrico__ has quit [Ping timeout: 252 seconds]
<hramrach> maybe it was 128MB at the start but nand_part helpfully resized it to the defeualt 32mb
<hramrach> do you have backup of the original nand content?
<libv> oh, yeah, i have all the data i need, even have a full dump of /dev/nand
<libv> pretty awful though, to discover this.
<libv> how many other devices are broken like this?
<hramrach> you should be able to mount it as loop and have the kernel scan the partition table with new enough mount(8)
<hramrach> you need util-linux 2.21 or 2.22 for that
<libv> ah, true
<hramrach> the 2.20 in debian don't suffice
<libv> :)
<jemk> wingrime: what do you mean?
<libv> hramrach: seems i will need to use offsets then
<hramrach> libv: better get the new mount
<hramrach> so you can see how the kernel interprets the stuff
<hramrach> ok, running nand_part on the dump as readonly lopp should suffice
<hramrach> libv: you are on the a23 tablet?
<libv> the a13, where i had no other option but to use nand
<hramrach> hmm, so far all devices had 16mb boot so nand_part enforces that
<hramrach> but I guess it is not a good idea
<libv> it's a quite severely bad idea
<hramrach> libv: what is the command that gives you the vfat filesystem size?
<libv> df
<hramrach> ok
<hramrach> libv: there is a mode in nand_part already that allows you to specify the boot size according to the nand installation guide
<hramrach> does it work for you?
<libv> hramrach: ...
<libv> hramrach: i am not killing that installation anymore at this point, i have wasted an enormous amount of time on figuring this shit out to begin with
<hramrach> somebody updated with new stuff since I wrote it ...
<wingrime> jemk: I mean, must be some trigger for erase current state of decoder sub engine
<libv> hramrach: when typing it up properly, i discovered the discrepancy
<jemk> wingrime: why? i don't think this exists.
<hramrach> libv: it's just matter of tarring up the / and then untarring it again after repartitioning
<jemk> wingrime: all state is in registers, you simply can write what you want there
<libv> hramrach: yeah, that's my plan, but first i want to find out what really went wrong
<libv> heh... my dump of /dev/nand is not finding vfat at logical offsets
<hramrach> try the new mount :P
<libv> i believe i have mounted from offsets before
<libv> and i am not updating my main system and open up a whole new can of worms in the process.
<hramrach> it would only work on the sunxi kernel anyway
<libv> ?
<libv> surely filesystems can be mounted from any point in an image, without needing to have to be able to read the partition table
<wingrime> jemk: regs, sure, but internal sram should be cleanable with one command
<hramrach> libv: the sunxi partition scheme is not supported by your main system
<jemk> wingrime: why would you want to clean sram?
<libv> hah
<libv> it's at 16MB
<libv> thank you nand-part.
<hramrach> so they put a 128MB filesystem into a 16MB partition in the factory?
<libv> no
rz2k has quit []
<libv> the partition of the original nand content start at 16MB
<libv> ah, no, it still starts there. good
<hramrach> yes. it's the default
<libv> mixed it up with the mele a1000 in the nand example
<libv> that offset worked now
<libv> 128MB
<hramrach> and the size reported by nand_part?
<oliv3r> libv: what's the tizen-sunxi community! never heard o it :(
<wingrime> jemk: for example I want switch video on the fly
<libv> hramrach: it could be that i myself badly wrote the new partition table
<hramrach> nand_part defaults to making the boot partition 16mb regardless of original size I think
<hramrach> but was it like that on the original nand or was it different?
<ssvb> oliv3r: I guess this 'community' is just a single guy who made tizen run on a sunxi device, here is his video - https://www.youtube.com/watch?v=vMQmf4hype4
<jemk> wingrime: why do you want to clean sram then? sram contains only data you have to write and temporary things
<ssvb> oliv3r: It seems to run rather sluggish, looks like there is a lot of work for 'them' before tizen is really usable
enrico___ has quit [Ping timeout: 248 seconds]
<buZz> if tizen will ever get usable at all :D
<buZz> besides the porting effort
enrico__ has joined #linux-sunxi
<libv> nand-part of course refuses to read a file
<libv> but from what i can tell, it says 32M offset and 32M size in the table
kivutar has quit [Quit: Ex-Chat]
<libv> (from looking at nand-part code and a hexdump
<libv> pretty worryingly so
Gerwin_J has joined #linux-sunxi
<nove> wingrime: how certain you are that MACC_AVC_CTRL irq enable is bit 7, in the test that i did it is bit 0 that make VE_WAIT timeout
<oliv3r> jemk: is your vdpau easily cross compileable? or does it rely on to many things?
<jemk> nove: I also use bit 0, more precise bit 0-2, so mask 0x7, maybe thats where the 7 comes from
<oliv3r> i should just get a precompiled lib :
<jemk> oliv3r: i don't know :D always compiled native
sgo11 has joined #linux-sunxi
<libv> ah, crap, i had forgotten about replacing script.bin
<libv> still boots though
<libv> hramrach: thanks for helping out, i will document that one should mount, backup, and check fs size before rewriting the nand
<hramrach> libv: you can set up loop device with the dump and run nand_part on that
<hramrach> libv: the guide is quite explicit about backing up the nand and checking what you have before and after partitioning
<hramrach> libv: df and sfdisk shows kilobytesn and nand_part sectors so nand_part should show twice what df and sfdisk shows
<wingrime> nove: you can check a23 kernel for that, IRQ for AVC form it
<hramrach> so if it starts with 32 it's probably 16M
<wingrime> nove: and don't forget , current drivers looks like not support irq from AVC or ISP,
<wingrime> nove: I not realy sure, but looks like they added it only in a23 code
<sgo11> mnemoc, I am very sorry to interrupt you. I just used linux-sunxi-stage-3.4-sun7i-latest.tar.xz. I simply "cp -a" to my existing system. "uname -a" shows the latest kernel. but I lost my wlan0 wireless interface. I do keep "/lib/firmware/ap6210/" which the latest kernel does not have. do you have any ideas why I lost my wlan0? lsmod shows bcmdhd is loaded. thanks.
<wingrime> nove: or they not used AVC irq
<wingrime> nove: before a23 new blob
<sgo11> since rz2k is not here, I don't know who I can ask.
<nove> wingrime: a23 cedar hardware is also updated, so that is one difference
<oliv3r> ssvb: yeah quite choppy; but it shouldn't really be a huge deal to 'port', I mean, tizen runs on linux, we have a well running kernel; so any distro ontop of that should work reasonably well
bfree__ has joined #linux-sunxi
qlex has joined #linux-sunxi
qlex has left #linux-sunxi [#linux-sunxi]
<sgo11> oliv3r, ha, you are here. can you please help me on that? thanks.
bfree_ has quit [Ping timeout: 260 seconds]
<oliv3r> jemk: can you host a precompiled vdpau bin somewhere I can download?
<oliv3r> sgo11: i have no clue; the ap6210 wifi is highly experimental
<sgo11> oliv3r, actually, I don't know what ap6210 is. rz2k told me my kernel didn't support 3d acceleration. there was a problem. I have to use the latest kernel. then I download the latest nightly build. I cp -a uImage, firmware, modules to my existing system. I saw ap6210 is not in the nightly build kernel, thus, I keep it in firmware. that's all I did. then wlan0 is gone.
<oliv3r> /lib/firmware is only the firmware it's not the actualy driver
paulk-collins has joined #linux-sunxi
<wingrime> nove: I think IRQ should be in AVC already as logical part of cedar design style
<sgo11> oliv3r, so where is the actual driver? maybe I should diff the two modules to see what is lost.
<oliv3r> sgo11: i'm not sure if it has been fully ported yet
<oliv3r> wens: do we have that driver yet?
<wens> oliv3r: BCM43662? was submitted by Arend, should be in wireless-next for 3.14
<nove> oliv3r: binaries are evil, why can't you compile?
<wens> compile what? the firmware?
<nove> vdpau-sunxi
<oliv3r> wens: yeah but we don't have that in our 3.4 do we
<jemk> but nove is right ;)
<oliv3r> jemk: thank you! make install will put that in /usr/lib right
<jemk> yep
<jemk> no
<jemk> /usr/lib/vdpau
<oliv3r> ok
<wens> oliv3r: I thought bcmdhd in 3.4 works?
<oliv3r> dunno; hence my question :)
<wens> I only used cubietech's tree :(
<sgo11> bcmdhd in 3.4 works. it works in my old kernel. today, when I try to use the nightly build kernel. I just lost my wlan0.
<wens> oliv3r: i remember something about cubietech replacing bcmdhd with an older version?
bbrezillon has joined #linux-sunxi
Gerwin_J has quit [Quit: Gerwin_J]
shineworld has quit [Ping timeout: 260 seconds]
<hramrach> they took an old version of the driver and patched it to work with the hardware
<hramrach> so you get either old buggy driver incompatible with current kernel that can use the hardware or completely different driver in sunxi which does not support the hardware and the diff between the two is huge and has lots of architectural changes
<hramrach> maybe somebody patched it in, can't keep up with all the stuff
<sgo11> I tried simply cp the old bcmdhd.ko. it does not work. maybe I have to stick with the old kernel. wifi is important to me. thank all for the help. gotta have some sleep first. thanks.
sgo11 is now known as sgo11_sleep
<hramrach> yes, it's important. But he drivers are really crappy so looks like nobody managed to merge it so far :/
<hramrach> also it seems bcm supports this on mainline
<hramrach> but not ancient kernels ..
maz_ has quit [Ping timeout: 246 seconds]
<libv> hramrach: what do you think of the u-boot patches i provided on the ml?
<libv> hramrach: since you wrote your howto for sun4i, how hard will it be for you to copy and test patch 4 on sun4i?
<libv> if i get your ack, then when that code makes it to the lichee-dev branch it will severely simplify this howto
<buZz> wow, that Rex board looks cool
<buZz> DM3730
Gerwin_J has joined #linux-sunxi
Seppoz_work has joined #linux-sunxi
<Seppoz_work> hi, im trying to set UART 6 and 7 on PI pins, can somebody check if the fex settings are correct as we cant receive any data: http://pastebin.com/2SXsxhAX
bbrezillon has quit [Quit: Ex-Chat]
bfree__ has quit [Quit: Konversation terminated!]
bfree__ has joined #linux-sunxi
<libv> hramrach: on my hyundai a7hd, the boot partition starts at 1MB on nand
<libv> amazingly
Seppoz_work has quit [Read error: No route to host]
Seppoz_work has joined #linux-sunxi
fredy has quit [Excess Flood]
<HdkR> yes
fredy has joined #linux-sunxi
<wingrime> Turl: thats very interesting to get it in hands
<wingrime> Turl: but power vr - again
<Turl> yeah
<Turl> but usb 3 and dual band wifi
<Turl> interesting for router appliances maybe
<Turl> AC1000 is interesting too
<Turl> AC100*
<wingrime> jemk: better update wiki
<wingrime> jemk: code it self - refernce , but docs are better
<jemk> wingrime: I know, but without code to try now knowledge about registers
mart has joined #linux-sunxi
notmart has quit [Ping timeout: 245 seconds]
<wingrime> jemk: any way, I see many things can be varired..
<hramrach> libv there are patches on the ML?
<wingrime> jemk: are you saw a23 trace?
<wingrime> jemk: for jpeg encoder
* nove still didn't started the poc jpeg encoder
<wingrime> jemk: also, there thing, you should do - make normal, shared header whith registers and bit sets
<hramrach> anyway, it is not clear how much is hardcoded in the bootloader and how much is read from the nand so the code from ancient times tried to make the boot partition consistent with existing device
<wingrime> jemk: curently I see , regs defenition are not in single place, thats problematic
<hramrach> but since devices with different layout are now known it's probably better to update it to keep the partition as is instead
Quarx has quit [Read error: Connection reset by peer]
<wingrime> jemk: >>It is *imortant* that the --- fix typo
<wingrime> jemk: in readme
<hramrach> last thing i see is something about pcduino
<hramrach> on ml
<nove> wingrime: is okay, pocs are only for testing, and is good to have something simples to be a base to find more about the hardware
<jemk> wingrime: this is only a first demo, so stop complaining or make a better one
Seppoz_work has quit [Ping timeout: 245 seconds]
<nove> jemk: good work, but i can only test it monday
<wingrime> jemk: nice work
eFfeM has joined #linux-sunxi
FDCX has quit [Remote host closed the connection]
<wingrime> nove: your replay tool is workable?
<wingrime> currently
<nove> wingrime, jemk, maybe we should have a wiki page with a who is doing what, and what is to do
<wingrime> nove: realy there such place
FDCX has joined #linux-sunxi
<nove> wingrime: i think not, that replay tool was only to check that tracer was working
<nove> wingrime: yes but that page needs to be filled
<eFfeM> Hi I just dusted off my mele a2000, and I was wondering what the best fw is to put on it (it still has the original sw afaik)
<wingrime> nove: we have much HW capabalies that's we currently not know
<nove> wingrime, jemk, also would be good if we make announces in the maillist, i think there are some people that don't know that is being REed
jemk has quit [Ping timeout: 245 seconds]
<hramrach> libv: ok, did some tests with nand-part and it leaves partition 1 as is unless you use -f
<hramrach> so nand-part is ok. should not do anything bogus unless you go out of your way to force it
<wingrime> nove: ... cool anonunces better make with luke and tsvetan
<hramrach> btw it works fine on plain file, just complains it cannot re-read partition into kernel
<hramrach> nove: what status is h264 dec?
<wingrime> hramrach: works, times ago
<hramrach> is there a media player that can use it?
<wingrime> hramrach: but 10-bit video wonder ever be support
<hramrach> with the stuff the AMD guy did it should but not sure these SoCs have enough CPU power frob the stream on the fly
<nove> wingrime: only a email to our maillist, saying things like a poc h264 encoder was started.
<wingrime> hramrach: mplayer
<wingrime> nove: thats too early
<nove> wingrime: see hramrach doesn't know
<wingrime> nove: only I - frames
<hramrach> but seriously, I did not see that driver working at all. It uses OpenMAX interface and I don't have anything linked with OpenMAX
<wingrime> nove: hramrach tested it as I remeber, but failed by some reason
<hramrach> oh, yeah. I had it compiled and it would segfault on start
<wingrime> nove: ^
<ssvb> hramrach: have you reported the details of this crash somewhere?
<hramrach> I reported it here
<hramrach> wingrime told me it WokrsForHim(tm)
jemk has joined #linux-sunxi
<ssvb> hramrach: by details I mean the information about the distro, versions of the relevant software, kernel version, and some sort of a backtrace
<wingrime> hramrach: there problems with debian's arm mplayer vdpau
<hramrach> what kind of problems?
<wingrime> hramrach: repo's version vdpau didn't ever try to run
<wingrime> hramrach: more preciesly, mplayer builded without vdpau but, still claims , that he have
<wingrime> hramrach: I rebuilded it by hand and it workedforme(tm) and many peoples here(tm)
<hramrach> it WorksForMe on radeon
<hramrach> that is it uses vdpau output and when I add a bogus vdpau driver it complains decode failed
<ssvb> hramrach: are you using the debian arm mplayer package on radeon?
<nove> wingrime: mplayer2 from wheezy is working in my case
<hramrach> ssvb: presumably the package should be built the same on each platform
<hramrach> but yes, there may be exceptions
<hramrach> ok, so mplayer2 and not mplayer
<ssvb> hramrach: if mplayer2 works, but mplayer does not, then it might perhaps make sense to notify the debian arm maintainers
<hramrach> hmm, I should figure out how to boot the old kernel again
<ssvb> old kernel?
<nove> hramrach: yes, mplayer wasn't compiled with vdpau support, you should try again
<hramrach> 3.4
<wingrime> ssvb: vdpau on arm are not so common I think
<hramrach> I swapped the cards a lot when trying out different u-boot netbooting hacks so don't even know what is where anymore :s
<wingrime> nove: I think is too early for h264 anoncement, but it will be I think luke poke phoronix guys and there will be nice anonce
<ssvb> wingrime: h264 encoding support announcement?
<wingrime> ssvb: PoC
<wingrime> ssvb: Jemk just pushed PoC with it, but encoder use only I-frames
<nove> wingrime: no, not that, only some email to *our* list, announcing the current status
<ssvb> also it's better to be careful with luke and phoronix, or the announcement may be full of bullshit
<wingrime> ssvb: indeed, we should explain carefully about whats are happend
<jemk> mplayer is bad for vdpau i noticed in the meantime, i should fix my readmes. better use mpv, it works much nicer
<nove> wingrime: like when we have something new, write a email
<wingrime> jemk: thats repo's mplayer bug, we must poke debian arm maintrainer
bsdfox has quit [Ping timeout: 260 seconds]
<jemk> wingrime: not only, mplayer also has a very bad vdpau implementation
<wingrime> jemk: I did't like movements with project forks, thats looks strange, mplayer->mplayer2->mpv
<wingrime> jemk: NIH syndrome
<jemk> but mpv is really nice, it even plays the stream from my encoder, mplayer only complains
<jemk> ok, my stream is broken somewhere, but anyway
<wingrime> jemk: thats becose mpv are just ffmpeg's gui
<wingrime> jemk: try add -vvvv to mplayer
<wingrime> jemk: ffmpeg are too should have error tolerance options, so, make it scream
jemk has quit [Read error: Operation timed out]
wolfy has joined #linux-sunxi
HdkR has quit [Ping timeout: 272 seconds]
wolfy has left #linux-sunxi [#linux-sunxi]
jemk has joined #linux-sunxi
bsdfox has joined #linux-sunxi
Gerwin_J has quit [Quit: Gerwin_J]
HdkR has joined #linux-sunxi
nove has quit [Quit: nove]
netlynx has quit [Quit: Leaving]
parabyte has quit [Quit: Leaving]
Gerwin_J has joined #linux-sunxi
Gerwin_J has quit [Ping timeout: 252 seconds]
popolon has joined #linux-sunxi
popolon has joined #linux-sunxi
mart has quit [Ping timeout: 272 seconds]
FreezingCold has joined #linux-sunxi
pwhalen has joined #linux-sunxi
eFfeM has quit [Read error: Connection reset by peer]
rellla has quit [Ping timeout: 260 seconds]
wingrime has quit [Ping timeout: 272 seconds]
Black_Horseman has joined #linux-sunxi
Black_Horseman has joined #linux-sunxi
chil has quit [Remote host closed the connection]
yon_ has joined #linux-sunxi
yon_ has quit [Client Quit]
jemk has quit [Remote host closed the connection]
physis has joined #linux-sunxi
Gerwin_J has joined #linux-sunxi
pwhalen has quit [Ping timeout: 252 seconds]
Gerwin_J has quit [Ping timeout: 252 seconds]
<mrnuke> oliv3r: I have a big-endian safe version of mksunxiboot. Interested ?
iamfrankenstein has joined #linux-sunxi
<mrnuke> It's also using only standard C lib calls, so it should (?) be portable
pwhalen has joined #linux-sunxi
iamfrankenstein has quit [Quit: Nettalk6 - www.ntalk.de]
<hramrach> mrnuke: what's the advantage over the stuff in u-boot-sunxi?
<hramrach> I build u-boot regularly on x86 and arm and it boots built on either. Are they same endian?
rose28357 has quit [Quit: AndroIRC - Android IRC Client ( http://www.androirc.com )]
Fresh has joined #linux-sunxi
<mrnuke> hramrach: it will compile on non-linux systems, and will produce correct results on BE machines
<mrnuke> arm (by default) and x86 are LE
naobsd has quit [Ping timeout: 272 seconds]
<mrnuke> I did some work last sumer where incorrect byte order, and ubpacked structs bit me really hard
<mrnuke> *unpacked
physis has quit []
<mrnuke> I picked picked up the habit of being careful :p
<Nyuutwo> mrnuke: unpacked structs can be algined, it hit me in some code ported to arm
<mrnuke> Nyuutwo: I feel your pain, brother
<Nyuutwo> now I just use uint8_t table and write byte by byte
<mrnuke> I use the idea of stream and local storage. stream is void*, local is stdint types, and use helpers to insert to/extract from stream
t3st3r has quit [Remote host closed the connection]
pirea has joined #linux-sunxi
<pirea> hy
paulk-collins has quit [Quit: Ex-Chat]
<pirea> arokux you can help me with something?
Fresh has quit [Quit: Leaving]
<pirea> WarheadsSE why you don't update linux kernel for sun*i
<pirea> sunxi
<pirea> ?
<pirea> where i can fiind nightly builds of kernel?
<mrnuke> nightly builds? seriously?
<Nyuutwo> pirea: try downloading linux-sunxi and build it
<Nyuutwo> crosscompiling kernel isn't so hard
<Nyuutwo> and also we have sunxi-bsp which makes a lot easier to deploy new u-boot,fex,kernel to sd card
<mrnuke> Nyuutwo: cross-compiling is easy. Getting it to boot... not so much
bfree_ has joined #linux-sunxi
bfree__ has quit [Ping timeout: 276 seconds]
<Nyuutwo> mrnuke: tried use sunxi-bsp?
<mrnuke> nope. What's that?
<mrnuke> hold on
<mrnuke> don't google that for me
<mrnuke> wait, does sunxi-bsp build mripard's mainline or linux-sunxi ?
<Nyuutwo> linux-sunxi
<mrnuke> aaah. I'm more interested in mainline
<mrnuke> I heard there's some form of initial SPI support there
enrico__ has quit [Quit: Bye]