mnemoc changed the topic of #arm-netbook to: EOMA: Embedded Open Modular Architecture - Don't ask to ask. Just ask! - http://elinux.org/Embedded_Open_Modular_Architecture/EOMA-68 - ML arm-netbook@lists.phcomp.co.uk - Logs http://ibot.rikers.org/%23arm-netbook or http://irclog.whitequark.org/arm-netbook/ - http://rhombus-tech.net/
<Turl> mnemoc: ever seen this? [ 0.000000] Memory: 760MB = 760MB total
<Turl> I disabled mali, cedar and all the things
<mnemoc> show the rest
<mnemoc> mali is clearly disabled, there is no gap for it there
<mnemoc> but still too much stuff reserved
<Turl> [ 0.000000] Truncating RAM at 40000000-7fffffff to -6f7fffff (vmalloc region overlap).
<Turl> :<
<mnemoc> eh?
<mnemoc> what bootloader?
focus_it has quit [Quit: Leaving]
<Turl> hno's
<mnemoc> mmc or nand?
<mnemoc> what board?
<mnemoc> also, which hno branch
<Turl> mnemoc: mmc, cubie, whatever you built for the hwpacks
<Turl> mnemoc: apparently it's the 3G/1G split so the max mem I can have is 1G-vmalloc
Sternennebel has quit [Quit: Leaving.]
<mnemoc> uhm
benjamin__ has quit [Ping timeout: 268 seconds]
<slapin_> mnemoc: I don't have resources to boot android on this board now
<mnemoc> slapin_: no problem. but you can mount nanda from linux too
<slapin_> mnemoc: busily hacking u-boot at the moment
<mnemoc> no rush
<slapin_> mnemoc: mtest proves very good for memory settings
<mnemoc> it's only to keep consistency between the boards supported by u-boot-sunxi and sunxi-boards/sunxi-bsp
<slapin_> what is sunxi-boards/sunxi-bsp?
tuliom has quit [Quit: Konversation terminated!]
<mnemoc> the first a repository of .fex files per board, the second a tool to generate hardware packs per board.
<mnemoc> there are quite a few repos at https://github.com/linux-sunxi
<Turl> mnemoc: VMSPLIT_2G and a recompile after I can use all 1GB of it as lowmem :)
<mnemoc> and those .fex will become .dts some day in future
<mnemoc> Turl: i don't understand why you had that problem
<Turl> mnemoc: I have HIGHMEM disabled
<mnemoc> ok
popolon has quit [Quit: Quitte]
vinifm has joined #arm-netbook
<slapin_> what is this stuff: http://pastebin.ubuntu.com/1346624/ ? it doesn't match settings I see in nand_sunxi dir...
gsilvis has quit [Read error: Connection reset by peer]
gsilvis has joined #arm-netbook
tzafrir_laptop has quit [Ping timeout: 246 seconds]
<Turl> mnemoc: do cubies have a preset mac addr?
<libv> hrm, got the base ioquake3 commit and am now reducing the changes in the rpi code
<libv> in a few days we hopefully have a gles1 capable port of ioquake3 running on our allwinner with mali libs :)
<Turl> can rpi run quake? :o
<orly_owl> Is anyone using an A10 tablet or just hacking/developing on it?
<libv> its videocore mess is emulating gles1/gles2
<Turl> libv: have you seen http://git.linaro.org/gitweb?p=people/fboudra/ioquake3-gles.git;a=summary ?
<Turl> orly_owl: o/
<slapin_> libv: /me votes for q2 as well...
<libv> Turl: oh no, another idiot who pulled the same shit :(
<libv> Turl: but thanks
<libv> maybe this is a quicker route indeed
<orly_owl> Turl: gday mate
<orly_owl> A7HD vs Novo7 Aurora?
<Turl> other than the idiocy that he commited temp *~ files >.<
<Turl> looks saner
<orly_owl> A10-MID is 7 or 10"?
<orly_owl> http://linux-sunxi.org/A10-MID has no info
<orly_owl> ah 7"
<libv> Turl: i have the rpi code at 5.5kloc+ 123loc-
DEAT has quit [Read error: Connection reset by peer]
<libv> once some further broadcom idiocy has been removed, the changes look kind of good
<libv> oh, and i need to remove etc support :)
<libv> that just gets in the way of lima work :)
vinifm has left #arm-netbook ["Ex-Chat"]
<libv> ah, more like 2800loc+, as one is a copy of a makefile
tinti has quit [Quit: Leaving]
<libv> so i got the diff down to something workable, and the gles port seems more trustworthy than the code i am seeing in the thatsnotright (which is what this guys nick translates to) repository
DEAT has joined #arm-netbook
<Turl> libv: https://github.com/turl/ioquake3/commit/5ff94ad5b1497cd601a9c3d9135c3c362121a65f is +3.3K-32 after nuking all the *~ crap
avernos has joined #arm-netbook
avernos has quit [Changing host]
avernos has joined #arm-netbook
stefanro1 has joined #arm-netbook
stefanro has quit [Ping timeout: 246 seconds]
<jlj> why does ioq3 need changes to run? cause the a10 doesn't have full opengl support?
<Turl> jlj: mali supports GLES, not GL
<jlj> yeah, kinda figured. thanks
bsdfox has quit [Ping timeout: 255 seconds]
<slapin_> ahahaha, first part of NAND init sequence seems to work. >9000 to go.
<slapin_> gnight, all!
jquip has joined #arm-netbook
longsitaizi has joined #arm-netbook
jquip has quit [Quit: jquip]
bsdfox has joined #arm-netbook
bsdfox has quit [Changing host]
bsdfox has joined #arm-netbook
itamarjp has joined #arm-netbook
ibot_ has joined #arm-netbook
ibot has quit [Read error: Connection reset by peer]
ibot_ is now known as ibot
avernos has quit [Quit: Leaving]
ibot has quit [Remote host closed the connection]
ibot has joined #arm-netbook
bsdfox has quit [Ping timeout: 245 seconds]
Quarx has joined #arm-netbook
itamarjp has quit [Read error: Connection reset by peer]
longsitaizi has quit [Ping timeout: 245 seconds]
slash_random has quit [Ping timeout: 240 seconds]
freakazoid0223 has quit [Quit: Leaving]
jquip has joined #arm-netbook
RITRedbeard_ has joined #arm-netbook
RITRedbeard has quit [Ping timeout: 246 seconds]
avernos has joined #arm-netbook
avernos has quit [Changing host]
avernos has joined #arm-netbook
bsdfox has joined #arm-netbook
bsdfox has quit [Changing host]
bsdfox has joined #arm-netbook
jquip has quit [Ping timeout: 240 seconds]
bsdfox has quit [Ping timeout: 260 seconds]
RITRedbeard_ has quit [Read error: Connection reset by peer]
bsdfox has joined #arm-netbook
RITRedbeard_ has joined #arm-netbook
bsdfox has quit [Changing host]
bsdfox has joined #arm-netbook
gimli has joined #arm-netbook
arnd_ has joined #arm-netbook
bsdfox has quit [Ping timeout: 255 seconds]
RITRedbeard has joined #arm-netbook
RITRedbeard_ has quit [Ping timeout: 246 seconds]
rellla has joined #arm-netbook
ibot has quit [*.net *.split]
jeremb has quit [*.net *.split]
ibot has joined #arm-netbook
jeremb has joined #arm-netbook
rellla has quit [Quit: rellla]
Sternennebel has joined #arm-netbook
rz2k has joined #arm-netbook
Sternennebel has quit [Quit: Leaving.]
cat_x301 has joined #arm-netbook
kop has quit [Read error: Connection reset by peer]
popolon has joined #arm-netbook
tzafrir_laptop has joined #arm-netbook
kop has joined #arm-netbook
tuliom has joined #arm-netbook
<mnemoc> http://dpaste.com/828619/plain/ <--- what is the purpose of this asm chunk? I don't now asm but it looks like if (*(NULL) != 0) return *(NULL); which is kind of wtfing from a C perspective
<mnemoc> but I've seen it very often
wim has quit [Ping timeout: 272 seconds]
<Triffid_Hunter> mnemoc: the end result of a chain of macros?
<mnemoc> .oO
<mnemoc> but if the var behind r3 is hardcoded to 0, a C compiler wouldn't accept to dereference it
<mnemoc> and that is supposed to come from C code
* mnemoc scratches his arm-analphabet head
<mnemoc> asm*
<Triffid_Hunter> mnemoc: looks like a lookup for an unresolved symbol.. is that in an object or a final linked elf? (or something else?)
<mnemoc> a gpl-violating built-in.o
<mnemoc> mach-sun5i/built-in.o
<mnemoc> from the function to distinguish between A13, A12 and A10s
<mnemoc> (note A10s != A10)
<mnemoc> if it wasn't hardcoded to 0 I would tell it's a cache
<mnemoc> memoizing
<mnemoc> but it wouldn't be unresolved
<jinzo> maybe it relies on some quirk in the hardware for detection?
rellla has joined #arm-netbook
<Turl> mnemoc: that code seems to r3=0; r0=*r3; if(r0<op>0) jump somewhere
<RaYmAn> when is that code running? before or after mmu setup?
<mnemoc> after
<mnemoc> it's used by /proc/cpuinfo and some drivers
<RaYmAn> weird
Mazon has quit [Read error: Connection reset by peer]
pawel5870 has joined #arm-netbook
<RaYmAn> you almost won't if it relies on specific behavior from npr..
Mazon has joined #arm-netbook
akakcolin_ has joined #arm-netbook
pawel5870 has quit [Ping timeout: 256 seconds]
Sternennebel has joined #arm-netbook
<rz2k> hm
<rz2k> lkcl: you didnt know that kernel\drivers discussion for linux-sunxi moved to our own ML because arm-netbook became A10 place only (which it must not be)?
<mnemoc> i guess he missed a sort of "official" announcement
<mnemoc> that list has never been "announced" anywhere....
<rz2k> it was announced for people who watch linux-sunxi at github in one of the issues opened. :3
<mnemoc> :)
<mnemoc> yes, it's more of a "those who care know it"
<mnemoc> but there are different levels of "caring"
<mnemoc> i doubt luke is following the issue trackers
<mnemoc> he is busy on a completely different layer
pawel5870 has joined #arm-netbook
rz2k has quit [Quit: ZNC - http://znc.in]
akakcolin_ has left #arm-netbook [#arm-netbook]
<rellla> to help improving playing videos on a10, i decided to create kind of a matrix, listing container|codec|profile|bitrate....|audiocodec....|demo-file|working yes/no|issues. test could be done with xbmc or vlc (if working already). would linux-sunxi-wiki be the right place to start, or would that be too xbmc-specific?
<mnemoc> I would say the sunxi wiki
<mnemoc> it's far more relevant to people reading the sunxi wiki than to people reading xbmc specific places
<rellla> main goal would be to collect infos. atm there is no concept in testing. xbmc-forum-users only say works, works not, i have artefacts. this infos could grow more relevant if they are listed together with demo files and codec infos.
<mnemoc> +1
<mnemoc> we can host the demo files in dl.linux-sunxi.org too
<rellla> so it's easier for developers to improve, i think. issue page on github is to confusingly imho.
<rellla> there are a few on samplemedia.linaro.org already. but not too much different codecs.
<mnemoc> want a samplemedia.linux-sunxi.org ? give me your ssh public key
<rellla> let me first do the wiki entry and check if it's needed. thanks anyway.
<mnemoc> ok
<rellla> very strange, these chinese :-(
<rellla> i checked out http://mediainfo.sourceforge.net for video analysis. any suggestions?
pawel5870 has quit [Ping timeout: 252 seconds]
<Gumboot> 10 01:08 <Turl> can rpi run quake? :o
<Gumboot> I think that was one of the original test cases in getting the driver up.
gimli has quit [Quit: Verlassend]
<rellla> mnemoc: samplemedia.linux-sunxi.org would be cool. mediainfo can produce .html analysis overviews. those could be linked to the matrix/samplefiles. i'll call you later...
popolon has quit [Quit: Quitte]
rellla has quit [Quit: rellla]
penguin42 has joined #arm-netbook
<rm> <Gumboot> 10 01:08 <Turl> can rpi run quake? :o
<rm> RPi may run Quake, but A10 will (hopefully) run OpenArena! :)
<orly_owl> rm: link to your site/blog?
<orly_owl> i want the mk802 page
<orly_owl> ty
<rm> OpenArena is a fully free (GPLed) derivative/clone of Quake 3
<hno> mnemoc, Can you make a hires picture of the cubieboard backside please? Mine is a bit hidden now.
<mnemoc> hno: light is pretty bad at the moment but I'll try
vinifm has joined #arm-netbook
ZaEarl_ has quit [Quit: Ex-Chat]
ZaEarl_ has joined #arm-netbook
<libv> hrm, ioquake3.arm has linked ok now....
ZaEarl_ has quit [Ping timeout: 260 seconds]
kop has quit [Ping timeout: 256 seconds]
<libv> still needs to be hacked for mali_native_window, but otherwise things seem to be ok
<libv> broadcom specific code seems to be quite limited
<mnemoc> libv: will merge your patchset for mali-libs now unless you want me to wait for something
<libv> nope, nothing there yet, Turl can still add his own additions when he gets round to testing it
<libv> i was hoping for some more feedback though, as this really makes things easier for a lot of people
<libv> and it definitely will not do the right thing for quite a few people
<mnemoc> what about using two .s for testing if it's armhf or arm at all
<mnemoc> without fancy flags
<libv> .s means gcc
<mnemoc> ok
<libv> gcc -dumpmachine is pretty reliable when gcc is there
<libv> Turl his route sounds ok too, but will also need lots of system dependant tuning
<libv> a combination of the two will be used in the end
<libv> first gcc -dumpmachine (with even more triplets added)
<libv> when that fails, looking at ldconfig naming
<libv> gcc, when there, is very reliable, so we patch up the other case only when gcc is not there
<libv> the version code is also likely to fail for various reasons
<libv> the mali module might not be loaded, /dev/mali might not be accessible
<libv> a c compiler might not be installed, libc-dev might not be installed, etc
<libv> but it still is a very sane route to take, save DKMS mali modules
tinti has joined #arm-netbook
<specing> RoHS compliant...
<specing> Oh my.
<mnemoc> libv: eventually we will have packages and the problem will be solved
<mnemoc> specing: you expected it wasn't?
<libv> mnemoc: true
<libv> hrm, i will need to find a way to first make ioquake3 happy, and then to reduce the dataset, the zipped data of the original q3a (which i own since half a year or so, for lima) is currently like 500MB
<mnemoc> uh
<libv> the timedemo requires original q3a though
<libv> as the map is not available in the q3a demo data
Almamuetya11 has joined #arm-netbook
RITRedbeard has quit [Read error: Connection reset by peer]
merbzt has joined #arm-netbook
<rm> why not OpenArena, which I mentioned just earlier...
<libv> rm: because most of the work has already been done.
<libv> rm: and because i can clearly spot the code where ioquake3 was altered
<libv> rm: you might be missing the point here
<libv> rm: i am not doing this to play quake
<libv> rm: i couldn't care less about gaming
<libv> rm: i am, however, intending to do lima driver development
<rm> I am just sick of how Raspberry Pi gives so much spotlight to a proprietary game
<libv> if _you_ prefer openarena, then by all means, _you_ can port openarena
<libv> proprietary?
<rm> when there's a perfect Free Software version of it (better in many ways)
<libv> it's GPLed.
<rm> I am afraid that the situation repeats with A10
<libv> wtf are you rambling on about?
<rm> "Cubieboard now runs Quake3, yaaaay!"
<rm> libv, data files are proprietary
<rm> engine might be GPLed
<libv> rm: and?
<rm> but you can't play the game without having proprietary data files
<libv> rm: you do know that i have done quite a lot of insane open source work right?
<rm> so the whole product is in effect proprietary, because without the data files you have no game
<rm> .....so?
<libv> rm: from REing bioses to REing whole graphics driver stacks
<rm> you need to stop pushing yourself at every opportunity
<libv> rm: and i am perfectly happy with some of q3a data being proprietary
<rm> I don't care who you are and the point we are discussing has nothing to do with you, your personality or your past work
<libv> again, _you_ are missing the point
<libv> i am doing this to do lima work
<libv> not to run q3a on this hw
<rm> I already had that "you know who I am right" from you before
<rm> and back then it was also not relevant to the discussion
<rm> that's what I mean, nothing more
<libv> running q3a in a hackable tree is an intermediate goal for me
<libv> and i do not care whether that fulfills your goal of free data
<rm> so you are just as misguided as the Raspberry Pi people, this is sad
<rm> carry on
<libv> music is copyrighted, as it should be.
<rm> afk
MMlosh has quit [Quit: Bye...]
<libv> hah.
<libv> rm: go crusade as much as you want, but do not stop people from doing real work.
<libv> cool, without mali_native_window code, at least hello triangle finally vanished :)
Sternennebel has quit [Quit: Leaving.]
vinifm has quit [Quit: Ex-Chat]
tinti has quit [Ping timeout: 246 seconds]
Almamuetya11 has quit [Quit: Nettalk6 - www.ntalk.de]
Sternennebel has joined #arm-netbook
arete has joined #arm-netbook
<libv> woohoo.
<libv> :( no input
Sternennebel has quit [Quit: Leaving.]
gsilvis has quit [Read error: Connection reset by peer]
gsilvis has joined #arm-netbook
cheng has joined #arm-netbook
<libv> yay, now q3a is set up enough to run, but there are... issues :)
<libv> ah, textures are broken, we'll soon fix that :)
tzafrir has joined #arm-netbook
bsdfox has joined #arm-netbook
bsdfox has quit [Changing host]
bsdfox has joined #arm-netbook
tzafrir_laptop has quit [Ping timeout: 246 seconds]
<arete> hi, ubuntu 12.04 is good for build android?
<hno> mnemoc, thanks, perfect.
wim has joined #arm-netbook
focus_it has joined #arm-netbook
rz2k has joined #arm-netbook
arete has quit [Quit: Leaving]
<penguin42> arete74: Yep
freakazoid0223 has joined #arm-netbook
<cheng> last time i saw ubuntu12.04 is armel, but nowadays seems like change to armhf
cheng has quit [Quit: Leaving]
<sv> yesyes
<jelly-home> don't they have both?
<sv> yes*
cromo has joined #arm-netbook
<cromo> Hi, two problems I have here with Mele A1000. I wanted to try the new xbmc for armhf, so I did compile the newest one along the sunxi-3.0 kernel. The work, except for the xbmc rendering slowly using software ("A10: using software rendering" in xbmc logs) and my script.bin not setting the 1080 resolution anymore. Any ideas? I could see the software rendering was already discussed here with empat0, however no solution was mentioned
rz2k has quit []
tuliom_ has joined #arm-netbook
phh has quit [Ping timeout: 248 seconds]
tuliom has quit [Ping timeout: 246 seconds]
phh has joined #arm-netbook
<cromo> OK, I got the first issue solved (rendering) - did not have the A10HWR environment variable set properly
popolon has joined #arm-netbook
Quarx has quit []
phh_ has joined #arm-netbook
phh has quit [Ping timeout: 246 seconds]
phh_ has quit [Quit: No Ping reply in 180 seconds.]
phh has joined #arm-netbook
cromo has quit [Ping timeout: 245 seconds]
rellla has joined #arm-netbook
xxiao has joined #arm-netbook
<vgrade> mele a1000 hangs at DRAM: with mele-a1000_hwpack_2012.11.10.7z any ideas guys?
<vgrade> my a1000 is quite an old one does that metter?
<vgrade> matter?
<hno> vgrade, first version of mele should work fine.
<hno> what does it say on the console?
<vgrade> U-Boot SPL 2012.10.g832a8e5-dirty (Nov 10 2012 - 04:00:35)
<vgrade> DRAM:
bsdfox has quit [Ping timeout: 252 seconds]
<hno> odd. that should work.'
<vgrade> will try an earlier hw pack
<vgrade> ok different result
<vgrade> U-Boot SPL 2011.09-00000-g8d13bc8-dirty (Jul 20 2012 - 14:00:25)
<vgrade> MMC: SUNXI SD/MMC: 0
<vgrade> Loading U-Boot... OK!
<vgrade> Jumping to U-Boot...
<vgrade> hang
alcides has quit [Read error: Connection reset by peer]
alcides has joined #arm-netbook
alcides has joined #arm-netbook
alcides has quit [Changing host]
focus_it has quit [Remote host closed the connection]
focus_it has joined #arm-netbook
phh has quit [Ping timeout: 248 seconds]
phh has joined #arm-netbook
freakazoid0223 has quit [Quit: Leaving]
<vgrade> ls
<libv> vgrade: i had massive issues with my mele, but that was due to too weak a power supply
<libv> vgrade: building uboot worked well for me
<vgrade> libv: I did some work with it when Tom posted the initial ubuntu image and there was always something I could not put my finger on with it.
ZaEarl_ has joined #arm-netbook
<vgrade> I'll have a look at power but I have a 5A voltage selectable unit
<libv> vgrade: it needs to deliver 2A
<libv> err, right :)
<libv> i of course read that as 5V :)
<libv> vgrade: clearing the first sectors is crucial, and dding in the right bits at the right place
empat0 has joined #arm-netbook
<libv> vgrade: you know mo (maurice), right? from basyskom?
<empat0> help! got a new a10 android box (stv-301), u-boot+sunxi-3.0 from today, box boots, i can ssh into it, but no hdmi signal! any ideas?
<vgrade> libv: I worked with him on Plasma yes, I met you at fosdem last year also :)
<libv> empat0: modprobe disp; modprobe lcd; modprobe hdmi
<vgrade> I was with mdfe
<libv> vgrade: i remember meeting you, but i am not sure who introduced us
<libv> ah, he did indeed
<libv> last fosdem was a bit of a blur :)
<empat0> libv: all loaded - no signal#
<vgrade> yea it was maurice, I hear he's moved on from basyskom. Do you still see him in the pub ?
<libv> empat0: can't help you any further then, i think i am the closest to disp stuff today, and i still have barely a clue of what is going on :)
<libv> vgrade: yeah, well, it's been a few weeks (i had flu, and now he is off to cuba)
<libv> but yeah, he went to openexchange
<vgrade> good guy
<libv> openexchange was bumped from suse as it was in competition with novell products
<libv> and now it is doing quite well
<libv> and vacuming up quite a lot of good people here in the nue area :)
<rellla> empat0: no/wrong output settings in script.bin?
<libv> apparently has a good atmosphere, and i am sure that mo will be happy there
<empat0> rella: i used the one from android nanda. looks reasonable.
<vgrade> libv: good to hear
<vgrade> hope to see him at a conf soon
<libv> sadly he is lost to linux on embedded devices now, but he will be happy
<libv> i'll whine at him when he gets back from cuba :)
<libv> i am sure that he can be badgered into coming to fosdem again :)
<vgrade> We had a good night out there last year :)
rellla1 has joined #arm-netbook
rellla has quit [Ping timeout: 265 seconds]
<libv> mo moved to nue about two years ago, and somehow he fit right in with the usual (ex)suse drinking suspects :)
<libv> so high probability of 3 fun nights in brussels in february :)
<hno> is mnemoc around?
<vgrade> lets try doing the sdcard creation wihout the script
rellla1 is now known as rellla
phh has quit [Read error: Connection reset by peer]
phh has joined #arm-netbook
<empat0> hno: i am playing with a new a10 device (stv-301). i got u-boot+kernel running, but i do not have any uart access. can i somehow provide you the exact data for u-boot? is dram.c the only board specific file or are there more?
arnd_ has quit [Remote host closed the connection]
Mehhh has joined #arm-netbook
<Mehhh> Holy god yes
<Mehhh> Is anyone here doing MK802 / lapdock
padpad has joined #arm-netbook
padpad has quit [Quit: Bye]
<libv> empat0: have you opened it up already?
<libv> empat0: hno was able to provide the missing information for a7hd, once the make & model and count of dram chips was known
<libv> empat0: since this is a stb, chances are that you can solder serial to it
<empat0> libv: yes, i opened it. nothing that looks like uart.
<Mehhh> Hey guys
<Mehhh> doing internal
phh has quit [Ping timeout: 252 seconds]
<empat0> libv: i used the dram.c from mini-x, it seems to run fine - besides hdmi
rellla has quit [Quit: rellla]
sv is now known as Genki
<hno> empat0, boards.cfg and dram_boardname.c is the only board specific details currently See sunxi-current branch in my u-boot repository.
<hno> Layout is a bit different in the "stable" sunxi branch.
Genki is now known as Sv
<hno> empat0, provide your dram configuration, what chips you have on the board, and your script.bin. And preferably also boot1 header as instructed in sunxi u-boot wiki.
<empat0> hno: boot1 is a little difficult - i do not have uart, sd-breakout or a livesuit image. is there any chance to get it from nand within android?
<empat0> hno: or linux-sunxi? i guessed the mini-x configuration and it seems to work ...
rz2k has joined #arm-netbook
<hno> empat0, have no other way at the moment. But hope to get ready to extract it from NAND shortly, in which case there will be a FEL tool to extract the NAND boot block.
<hno> empat0, please check what dram_para you have in script.bin. It's not 100% reliable, but often good enough.
<Mehhh> When you short pin4/5, why doesn't the screen come on?
fluxi has joined #arm-netbook
myfluxi has quit [Ping timeout: 260 seconds]
<empat0> hno: thats what i used. problem is it boots, i can ssh into it, but i do not get hdmi output - i used the script.bin from android.
<hno> There is nothing in u-boot related to HDMI.
<jlj> so I tried ./configure gooseberry_A721 with sunxi-bsp and it said No rule to make target `gooseberry_A721'
<jlj> am I doing something wrong? seems to work for the other boards
<empat0> hno: so, from u-boot point of view, i'm ready?
<hno> jlj, Is that listed as a supported board? there is no gooseberry_A721 in sunxi u-boot.
<hno> empat0, yes.
<hno> only not verified.
<empat0> hno: shall i mail script.bin and dram.c?
<hno> empat0, yes, and a note of what DRAM chips you have on the board.
<j^> i'm able to boot my mk802ii with the sunxi-3.4 kernel but get only 448MB of my 1G ram, with the same u-boot setup i can boot an older kernel with 1G, any poitners what needs to be changed?
<hno> j^, you need a newer u-boot.
<jlj> hno: when I do a /sunxi-bsp$ ./configure it is listen in the supported boards
<jlj> Usage: ./configure <board> supported boards:
<jlj> * gooseberry_A721 gooseberry_A721-android
<hno> j^, but so far no one have provided required MK802ii data for u-boot.
<j^> hno: so what would be needed?
<hno> j^, do you have UART?
<j^> no
<hno> j^, are you booting from NAND or SD?
<j^> SD
<hno> Which u-boot version are you using?
<empat0> hno: will try to find a magnifying glass for the rams. thanks and good night.
<hno> empat0, camera at slightly tilted angle usually works best.
empat0 has quit [Quit: irc2go]
<j^> trying to create a mk802ii config based on script.bin from http://blog.dryft.net/2012/08/v4-linaro-armhf-image-for-mk802.html
<j^> but that did not boot
<hno> j^, better use your own script.bin.
<j^> hno: from nand? will have a look
<slapin_> hno: ping
<hno> slapin_, half pong
<hno> j^, yes from nanda.
<focus_it> can cubieboard boot from its SATA?
<focus_it> Also can Mele A2000 boot from sata?
<j^> hno: from script.bin on nanda i get http://paste.ubuntu.com/1349019/ but does not boot with it
<hno> focus_it, kind of. You need to boot Linux kernel from SD/NAND. You can then kexec another Linux krenel from SATA if you prefer, or only have root on sata.
<focus_it> hno: sounds reasonable - if booting from SD, will that Linux need to write to SD before kexec? i.e. end up wearing the SD card?
Mehhh has quit [Ping timeout: 245 seconds]
<hno> focus_it, not at all.
<focus_it> hno: that is a good answer - thanks!
MMlosh has joined #arm-netbook
<hno> j^, you need to fill in the NULL parameters. Which you can do if you know what RAM chips you have.
<hno> j^, how many RAM chips is there? 2 or 4?
cat_x301 has quit [Quit: Leaving.]
vinifm has joined #arm-netbook
<Gumboot> rm: I thought the quake business model was a free engine and commercial maps but with freedom for others to create their own maps if they wanted.
<Gumboot> Has nobody created free maps, or is that not the model for q3a?
<rm> that's what people did, created a complete dataset (not just maps are the issue! also textures, models, sounds), paired it with the freed engine and the result is OpenArena
<hno> j^, if 2 DRAM chips then try copying those NULL parameters from cubieboard.
<hno> j^, 4 chips then.
Mehhh has joined #arm-netbook
<Mehhh> Hello
<Mehhh> Any lapdock people here, or mk802 owners
cat_x301 has joined #arm-netbook
<slapin_> hno: please apply my pull request re mk802
<hno> j^, density = 2048, io_width=8, bus_width=32, size=1024, rank = 1.
<hno> slapin_, I have a pull request?
<slapin_> hno: github pull request
<hno> I see it now. It won't work for the MK802ii discussed tonight.
<slapin_> it is not mk802ii
<slapin_> it is mk802 with 512MB RAM
<slapin_> I don't have any clue about mk802ii
<slapin_> it is separate hardware
<hno> slapin_, merged.
<slapin_> some of mk802s are based on rockchip, not A10.
<slapin_> hno: thanks
<hno> This one is A10.
<hno> slapin_: See <j^> hno: hno http://r-w-x.org/mk802ii.jpg
<Mehhh> Why isn't the A10 mk802 going into OTG when i short pin4/5?
<hno> Mehhh, what do your script.bin say?
<hno> it is possible the port is configured as device only.
<Mehhh> I';m running stock
<Mehhh> Huh?
* hno have no idea what stock means.
<Mehhh> stick firmware
<Mehhh> stock*
<Mehhh> from the actory
<Mehhh> factory*
<hno> I got that, but have no idea what the factory put in your firmware.
<j^> hno: ok http://paste.ubuntu.com/1349089/ boots it up with 448+512=960, now need some way to test if it actually works
<Mehhh> ok, so do you have a firmware you know that works with OTG for keyboard AND power?
<Mehhh> I'm trying to connect to lapdock off one cable
<Mehhh> errr
<Mehhh> one USB
<hno> j^, when tested please submit is as a patch or github pull request, based on sunxi-current branch.
<slapin_> hno: I also would like to discuss the clock part on http://pastebin.ubuntu.com/1346624 it is quite different from setup on http://ossfans.org/nand_registers.ps
<hno> Mehhh, with the right script.bin you can set the port as a HOST port if you like.
<Mehhh> Is there a list of them somewhere or do I have to manually edit?
<hno> OTG is also supported, but there is some reports on OTG mode not being entirely stable when unplugging devices.
<Mehhh> Also, will the script.bin work for android on nand AND linux on microsd?
<slapin_> ah mk802ii is so cool, it have axp209 part. how hot it is?
<slapin_> any links where to buy it?
<hno> slapin_, seen it on aliexpress many times.
<slapin_> how hard it is to solder battery pins to axp209
<hno> I do not dare to solder on axp209. Far too small pins for me.
<slapin_> hno: I do too, but there are too many variants, most without axp209, and many without A10
<slapin_> hno: if they had pads a bit longer...
<slapin_> ah
<slapin_> this one seems solderable
<slapin_> I dunno why they don;t make 4 1mm pads for battery attachment, I'd be very thankful..., even if price increased for $10...
<j^> hno: sunxi-current is in your uboot-allwinner? why not https://github.com/linux-sunxi/u-boot-sunxi
<slapin_> j^: do you still have link where you got mk802ii with nice and shiny axp209?
<slapin_> j^: I mean where you bought it
<slapin_> my board have CPU NAND and RAM chips on the same side, among with wifi module
<Mehhh> does anyone have a script.bin that supports host
<Mehhh> or know where to get one
oliv3r has quit [Ping timeout: 240 seconds]
<hno> j^, sunxi-current is only in my repository, and is where current work is being done. The other is the "stable" branch. They differ a bit in the DRAM part.
<hno> j^, you are welcome to submit to both. Saves me a bit of work.
<hno> the same goes for slapin_
<j^> hno: how do i build in your branch? make 'boardtype' CROSS_COMPILE=... does not work
<hno> j^, tjat e2comp link says it's Rikomagic now.
<hno> j^, that works. Have you added the board to boards.cfg and the dram file in board/sunxi/
<hno> plus matching board/sunxi/Makefile change.
<slapin_> hno: about clocks...
<slapin_> 16:58 < slapin_> hno: I also would like to discuss the clock part on http://pastebin.ubuntu.com/1346624 it is quite different from setup on http://ossfans.org/nand_registers.ps
<hno> The part left in my u-boot tree is not verified.
<j^> hno: mk802 is from Rikomagic no?
<hno> Right.
<hno> a bit tired, both Rockchip and Rikomagic starts with R..
<hno> slapin_, but I don't see any clock setup in your .ps. Only the gating.
rz2k has quit [Read error: Connection reset by peer]
tuliom_ has quit [Quit: Konversation terminated!]
<hno> slapin_?
tuliom has joined #arm-netbook
<j^> hno: pull for your branch https://github.com/hno/uboot-allwinner/pull/2
<slapin_> hno: see init sequence
<slapin_> hno: I try to understand why so different
bsdfox has joined #arm-netbook
bsdfox has quit [Changing host]
bsdfox has joined #arm-netbook
<slapin_> in init sequence NAND clock seems to be sourced from 24 * PLL5
<slapin_> hno:
<slapin_> hno: see 1.2
<slapin_> AHB_GATING_REG0 |= (1<<13) <--- this enables clock gating
<hno> Of the AHB clock yes. Also needs the NAND clock.
<slapin_> but next clock value is calculated from PLL5 values and stored to NAND_SCLK_CFG_REG
<slapin_> which is NAND clock
<hno> In the PS you only have XX = PLL5_CFG_REG, NAND_SCLK_CFG_REG = YY
<slapin_> hno: refresh it please
<hno> I have only seen NAD_SCLK_CFG_REG set to use OSC24M as reference.
<slapin_> hno: you might be right, but I need full understanding
<slapin_> hno: could you please see updated .ps
<hno> it can use OSC24/PLL6/PLL5 as upstream clock source.
<slapin_> in original u-boot and your nand branch I see what I have in .ps
<slapin_> so I want to understand what happens there.
<hno> http://fpaste.org/ifkY/ is what I have in lichee-dev.
<hno> slapin_, which of my nand branches?
<hno> Same in nandtrace branch.
<hno> Where did you find anything else?
<slapin_> hno: I'll check
<slapin_> nand_sunxi/nand_drv/nand_osal_uboot.c, function NAND_GetCmuClk()
<slapin_> hno:
<slapin_> hno: branch - explorenand
<slapin_> hno: it seems nand_sunxi dir assumes clocking from pll5
<slapin_> or there is some global synchronisation problem
<slapin_> even with this settings the controller allows to reset itself and do various setings
<slapin_> hno: as I see, your code provides gating, so I don't need to do my own, isn't it?
<hno> slapin_, it is possible nand_sunxi also sets up the clock. Have not looked in detail there.
<hno> It does not mattre very much. The clock just needs to be <200MHz or so.
<hno> AHB and NAND clocks are separate. Both needs to be enabled.
<hno> The AHB clock is system wide, the NAND clock only used by the NAND controller internally.
<slapin_> hno: it doesn't set it up, it assumes pll5 is configured.
<hno> where?
<slapin_> hno: do I need to do AHB_GATING_REG0 |= (1<<13)? or it is done already?
<slapin_> hno: look at init sequence
<slapin_> hno: it just configures GPIO, enables clock gating and configures NAND clock.
<slapin_> it soesn't configure PLL5 and AHB clock.
<slapin_> if NAND clock is enabled, would reset sequence work for the controller?
<hno> Are we talking about NAND_SetClock call in the driver?
<slapin_> hno: yes
<slapin_> hno: http://paste.ubuntu.com/1349289/ this seems to work, so I just want to be sure is clocking allright...
<vgrade> hno: looks like mu serial converter is the issue, not getting any serial now but latest hwpack boot to hdmi console
<Turl> vgrade: is it connected correctly?
<hno> As I said the clocking is not that critical. Too low and performance suffer. Too high and things fail.
<vgrade> tried all the usual swap 2/3, with without gnd
<vgrade> 115200 baud right
<Turl> yes
<Turl> make sure they're connected correctly on the board
<Turl> and don't plug vcc
<slapin_> hno: so in my case it seems everything is done and all necessary clock work is made, so I just need to proceed with NAND initialization, right?
<vgrade> yea looked at the triangle as pin 1 vcc, pin 2 / 3 tx/rx, pin4 gnd
<hno> slapin_, but yes seem the driver uses PLL5. No idea how that can work unless something enabled PLL5 first.
<Turl> vgrade: try unplugging board and uart from your computer
<Turl> turn board on
<slapin_> hno: could you please investigate into pll5 issue, since you know the code better than me?
<Turl> then plug and launch screen /dev/ttyUSB0 115200
<Turl> or whatever you use
<hno> slapin_, I have never tried that code when booted from SPL. It's quite likely left configured by the Allwinner bootloader when chainloading u-boot.
<vgrade> Turl: ok
<hno> slapin_, md.l 0x1c20020 1
<slapin_> hno: I tried, it works and displays proper NAND configuration.
<hno> Odd.. I get b1059491, not the expected default 0x11049280
<slapin_> sun4i#md.l 0x1c20020 1
<slapin_> 01c20020: b1058f91 ....
<hno> how did you boot?
<slapin_> probably bootrom configures it?
<slapin_> hno: from SD card
<slapin_> hno: via SPL
<hno> Odd. Why do ours differ.
<slapin_> my board's uptime is 2 days already
<hno> Ah, wait. Ofcourse it's configured. PLL5 is the DRAM clock.
<slapin_> hno: then it is very good source to clock NAND from
<slapin_> hno: I just have trouble understanding why bits are different in nand_sunxi and your code to configure NAND clock, regardless of clock source
<hno> Well, the bits in board.c is not used. overwritten by NAND_SetClock
<hno> It's there in all versions. Only have not deleted it in sunxi branch.
<hno> just #if 0 commented it out as the NAND driver is not there.
bfree has quit [Ping timeout: 252 seconds]
rz2k has joined #arm-netbook
<hno> slapin_, a guress is that it's good if the nand clock is enabled when the ahb gate is opened.
<slapin_> hno: did you mean clock.c?
<slapin_> hno: I quite don't understand why gate is enabled before clock is enabled
<slapin_> hno: I might try a bit different approach
<slapin_> hno: how does sr32 macro work? I loke it more than shifting values directly...
<slapin_> s/loke/like/
<ibot> slapin_ meant: hno: how does sr32 macro work? I like it more than shifting values directly...
<libv> vgrade: is this a run of the rs232 to usb?
<libv> vgrade: or is this a specific ttl to usb adapter?
<libv> run of the mill
<libv> even
<vgrade> libv: its the ttl to usb adaptor which tom sent with the early mele
<libv> :(
<libv> hrm, weren't you seeing something on it previously?
<vgrade> I can progress without serial for now but its awlays better to have
<vgrade> yes but very intermittent, a few chars then looked like it had hung.
<vgrade> One time I went for a drink and I had console on hdmi
<vgrade> so thing was booting but something wrong with ttl to usb. probably blown it up swapping tx/rx
RITRedbeard has joined #arm-netbook