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/
hg_5_ has quit [Ping timeout: 245 seconds]
hg_5 has joined #arm-netbook
hg_5_ has joined #arm-netbook
hg_5 has quit [Ping timeout: 255 seconds]
vinifm has joined #arm-netbook
hg_5 has joined #arm-netbook
froese has quit [Quit: Leaving]
hg_5_ has quit [Ping timeout: 276 seconds]
popolon has quit [Quit: Quitte]
hg_5_ has joined #arm-netbook
hg_5 has quit [Ping timeout: 246 seconds]
tinti has quit [Quit: Leaving]
hg_5 has joined #arm-netbook
rz2k has quit []
hg_5__ has joined #arm-netbook
servili007 has joined #arm-netbook
hg_5_ has quit [Ping timeout: 240 seconds]
hg_5 has quit [Ping timeout: 246 seconds]
stefanro has joined #arm-netbook
Jef91|AFK has quit [Quit: Instantbird 1.3 -- http://www.instantbird.com]
stefanro1 has quit [Ping timeout: 255 seconds]
hg_5 has joined #arm-netbook
hg_5_ has joined #arm-netbook
<servili007> Turl: ping
hg_5 has quit [Read error: Connection reset by peer]
hg_5__ has quit [Ping timeout: 276 seconds]
eebrah has joined #arm-netbook
<Turl> servili007: pong
<servili007> Turl: still haven't had much time to play with getting HW accel up and running. Could it be related to using the zatab props?
<Turl> doubt it
<Turl> they're not 'zatab props' in any case
<servili007> I see. I can't even get most other prop sets to give me a display in fact, lol
<servili007> Have you had any recent builds with zatab using the latest setup in the ADT repos?
hg_5_ has quit [Ping timeout: 244 seconds]
<Turl> I've built cm-10.1 but that doesn't have cedar at all
<Turl> (and I haven't pushed all of it yet)
<Turl> servili007: ZaEarl might know
voronaam has quit [Quit: Leaving.]
<servili007> Okay, thanks for the info. Hopefully it won't be too difficult of an issue once I have the time for it
freakazoid0223 has joined #arm-netbook
<servili007> ZaEarl: ping
kaspter has joined #arm-netbook
<ZaEarl> servili007, what are you trying to do?
<servili007> ZaEarl: I made working mk802ii cm10 builds starting from the ADT zatab work, but hw accel seems broken
<ZaEarl> yeah, we stuck with cm9
<servili007> did you run into the same issue on building 10, or just never tried?
<ZaEarl> didn't try very hard, as cm9 runs quite well
eebrah has quit [Read error: Connection reset by peer]
vinifm has quit [Remote host closed the connection]
freakazoid0223 has quit [Ping timeout: 256 seconds]
eebrah has joined #arm-netbook
discopig is now known as sv
Jef91 has joined #arm-netbook
KoH__ has joined #arm-netbook
KoH_ has quit [Ping timeout: 260 seconds]
gimli has joined #arm-netbook
servili007 has quit [Read error: Connection reset by peer]
ManoftheSea has quit [Ping timeout: 255 seconds]
ManoftheSea has joined #arm-netbook
gimli has quit [Ping timeout: 252 seconds]
Quarx has joined #arm-netbook
ZaEarl has quit [Ping timeout: 256 seconds]
Jef91 has quit [Remote host closed the connection]
Jef91 has joined #arm-netbook
Jef91 has quit [Remote host closed the connection]
Jef91 has joined #arm-netbook
rellla has joined #arm-netbook
lawrence has quit [Quit: lawrence]
eebrah has quit [Ping timeout: 252 seconds]
<rm> nice
<rm> what's with the 3.3 GHz though
pcat has joined #arm-netbook
<jammi> maybe 800MHz times four
rellla2 has joined #arm-netbook
rellla has quit [Ping timeout: 276 seconds]
rellla2 is now known as rellla
lawrence has joined #arm-netbook
<orly_owl> 'External 3G'
sv is now known as discopig
arete74 has quit [Quit: leaving]
arete74 has joined #arm-netbook
<gzamboni> does anyone knows the git of this gpio patch: http://www.spinics.net/lists/linux-doc/msg07816.html ?? u-boot ? why does u-boot needs gpio ?
Jef91 has quit [Remote host closed the connection]
<hramrach> mnemoc: yes, that fix to boards.cfg works, thanks
<hramrach> now the u-boot fails to link
<hramrach> I have binutils crash :s
<rm> mnemoc, u here?
<hramrach> lawrence: hello
<hramrach> got any of the case pictures?
<hramrach> or did not get to putting them anywhere?
pcat has quit [Ping timeout: 256 seconds]
<hramrach> ^^ the link error when building u-boot lichewe-dev sun4i_sd
<hramrach> there is linker crash but undefined symbol also
<hramrach> I don't understand how that can happen since init_part is static functtion defined in header included in the file that references it
<slapin> hramrach: lynker crashes because of undefined symbols
<hramrach> possibly CONFIG_PARTITIONS is not defined
<hramrach> but where and how it is defined for the targets that work ..
kaspter has quit [Ping timeout: 240 seconds]
popolon has joined #arm-netbook
pcat has joined #arm-netbook
<oliv3r> mnemoc: I was just talking to tsvetan over at #olimex and he said, that their kernel releases contain binary blobs for certain register access . Ignoring Cedar and Mali, would you know what he specifically is talking about? Pure in the kernel?
<lawrence> @hramrach - haven't had time yet
<lawrence> i'll take a quick snap of some of the mags together
<lawrence> but i need to find an issue which focuses on laptop casings
<lawrence> as they generally focus on 1 specific item each issue - eg mp4 players etc
<hramrach> yeah, laptop and tablet
<hramrach> if you go with the thinkpad case then no need for pictures. they are all around the web
<lawrence> i'll take a quick snap of the type of magazine i get so you have an idea
<hramrach> but maybe some of the oem cases are really good in some other ways
<hramrach> yes, would be interesting to have some idea what you can get
<lawrence> these aren't what you are looking for
<lawrence> but examples of sort of stuff i can find
<lawrence> i need to get the issues that have laptop casings etc in and then they'll have ads from appropriate factories
<hramrach> bike case ;-)
<hramrach> is that for flashlight?
<hramrach> I guess for anything you can stuff inside
<lawrence> i'm just photographing semi random stuff so you see
<lawrence> well, i was till china just said screw you lawrence, no flickr for you
hansg has joined #arm-netbook
Ershov has joined #arm-netbook
<hramrach> I don't really understand what the small bits are for
<hramrach> it's all chinese
<libv> odroid-x2 might get mali binary drivers for a gnu/linux (ubuntu) soon
<libv> finally.
<hramrach> cool
<mnemoc> rm: pong
<rm> mnemoc, a question not actual anymore :)
<mnemoc> rm: :)
Ershov has left #arm-netbook [#arm-netbook]
<mnemoc> oliv3r: the SDKs olimex received included mach-sun5i, /dev/cedar and some other parts as .ko/.o only
<hramrach> is it possible to set uart options someplace other than board.cfg?
<hramrach> the cubieboard config on sunxi branch seems to use the defeault there
<hramrach> and so does sun4i on lychee-dev
<hramrach> but no output
<mnemoc> hramrach: on boards.cfg you set defines, then the defiens are used to pick pinmuxes
<mnemoc> the cubieboard has dedicated uart0 pins so the default fits fine, same for the mele
<hramrach> the only estra define is SPL
<mnemoc> lichee-dev doesn't use the spl part, it relies in boot0/boot1
SouL_ has joined #arm-netbook
SouL_ has left #arm-netbook [#arm-netbook]
<hramrach> yes, that's why it is not defined
<hramrach> I guess the sun4i_sd errorneously defines MMC which includes libmmc and fails to link with it then
<hramrach> in lichee-dev
<mnemoc> one is to boot from nand, with the default uart pins and the SD slot as SD slot
<mnemoc> and the other to boot from nand, with the SD slot as uart/jtag breakout
<hramrach> that seems to be sun4i_sd
<hramrach> but does not link
<mnemoc> odd
<lawrence> hramrach - added some descriptions
<oliv3r> mnemoc: I just wrote a post to gpl-violations to see what they think we can do to get AW to comply ... better. We as users usually can't get much done, but olimex as a 'user' of AW chips has more legal chances
<hramrach> not that I need the sd uart
<mnemoc> oliv3r: allwinner doesn't give the sdk directly, it's via wits-tech
<buZz> oliv3r: is olimex called olimex because you are named oliv3r ?
<mnemoc> oliv3r: they have such a tangle of sister companies that there is no way on earth you can force them to comply
<oliv3r> buZz: lol i have no affilition with olimex :p
<buZz> hehe ok
<mnemoc> oliv3r: an even if they were a single company, they are in china and their sole market is domestic
<hramrach> lawrence: nice, thanks
<mnemoc> oliv3r: they really don't care about "the rest of the world"
<oliv3r> mnemoc: i didn't say it would be easy or even possble :p
<oliv3r> atleast i'm getting a little worried about sunxi's future is all, and maybe there are better SoC's to go after
<lawrence> i wouldn't worry
<lawrence> they're selling more than enough of these at the price point
<hramrach> I would worry about their deals with MS if anything
<buZz> allwinner is doing deals with microsoft?
<lawrence> i just found some gorgeous propaganda
<hramrach> WinRT support
<buZz> lol winRT is a joke
<buZz> just the latest winCE fart
<buZz> from a dying corpse
<lawrence> + all that secure boot crap for it
<lawrence> no-one wants that in the industry
<buZz> i dno
<hramrach> but you need to have MS support contract to sell devices with WInRT and AW is/was working on some
<buZz> EFI is better than legacy BIOS
<lawrence> no.
<buZz> no? :)
<lawrence> no, no no no no.
<buZz> well, on x86 its an improvement
<hramrach> EFI is just broken differently
<oliv3r> so what about freescale i.MX
<buZz> computers boot faster
<hramrach> more features, more bugs
<hramrach> still poor design
<buZz> still better than what was there before
<lawrence> BIOS worked fine.
<hramrach> there was na YT video by a guy who wrote the Linux EFI boot support
<oliv3r> that one complying better? :)
<oliv3r> lawrence: coreboot works better :p
<lawrence> EFI is a complete fark up as far as i'm concerned
<hramrach> buZz: no. broken before, broken after. not better
<buZz> change ;)
<lawrence> i've had machines that i seriously considered returning as faulty due to EFI issues
<hramrach> that happens with BIOS too
<hramrach> but that's implementation issue
<lawrence> i haven't had to spend > 2 days trying to get win64 working on HP's prior to EFI bios
<lawrence> then HP getting involved, and going, er yeah, its a bug, we'll ship a fixed bios to you.
<hramrach> hehe
<lawrence> i don't enjoy wasting days of my life on stupidity like that
<lawrence> plus a lot of computers you can't even turn off efi
<lawrence> i'm venting here
<hramrach> it's not like BIOS implementations are bug-free
<lawrence> i've had MORE THAN MY FAIR SHARE OF PAIN
<lawrence> deep breath in, out.
<lawrence> ok.
<hramrach> and both BIOS and EFI is flawed by design
<hramrach> no sane interface for the OS .
SouL_ has joined #arm-netbook
hg_5 has joined #arm-netbook
<hramrach> invalid script :s
<hramrach> and for linux the script is valid and the kernel crashes
<hramrach> need a valid script for u-boot I guess
<hansg> mnemoc, can you please merge my patch titled; "[PATCH 3.4 4/8] Added support for Olinuxino A13_Micro" into stage/sunxi-3.4 I understand we don't want per board defconfig-s but (for now) this one really is necessary because the micro only has 256M of ram, so if you do a default defconfig you almost have no ram left
<hansg> And people are asking for instructions how to build a Olinuxino A13_Micro kernel ...
SouL_ has left #arm-netbook [#arm-netbook]
<mnemoc> hansg: what about a sun5i_headless_defconfig ? (same for sun4i)
* RaYmAn just got his gk802 and ponders what to do with it
<mnemoc> or "server" instead of headless
<mnemoc> RaYmAn: poke lawrence until we get the open lapdock campaign started :p
<RaYmAn> heh
<lawrence> i tell ya, nothing till after cny
<lawrence> then stuff can proceed
<lawrence> so research research research
<mnemoc> :)
<hansg> mnemoc, I understand the desire for something more generic, my medium term plan is to make the reservation code check the kernel cmdline, allowing to dynamically set the reservation size (down to 0 bytes), and then check the size in the drivers using said reservation, and fail module_init in those drivers when the reservation is 0.
<hansg> mnemoc, so I don't want to spend too much time on short plan solutions. I just checked and the def_config is a single file, so I'll just mail it to the list as attachment, and tell people asking about the micro to copy it into place.
<mnemoc> considering mali support needs all those fat reserves a headless/server defconfig isn't really a short term solution soon death solution...
<mnemoc> s/solution//
<hramrach> lawrence: any idea if there is LVDS -> DP adaptor?
<hansg> mnemoc, answered the user by attaching the micro config to the mail, so fixed :)
<mnemoc> :)
SouL_ has joined #arm-netbook
SouL_ has left #arm-netbook [#arm-netbook]
fragmint has quit [Read error: Connection reset by peer]
fragmint has joined #arm-netbook
SouL_ has joined #arm-netbook
SouL_ has left #arm-netbook [#arm-netbook]
ganbold_ has joined #arm-netbook
vinifm has joined #arm-netbook
tinti has joined #arm-netbook
SouL_ has joined #arm-netbook
SouL_ has left #arm-netbook [#arm-netbook]
tinti has quit [Ping timeout: 276 seconds]
hg_5 has quit [Read error: Connection reset by peer]
<libv> hrm, my mele does not seem too stable at 1200MHz
<Gumboot> You mean the GPU at 1200MHz?
<libv> cpu
<Gumboot> Not trying hard enough!
<libv> hitting assertions all over the place.
matt__ has joined #arm-netbook
fragmint has quit [Read error: Connection reset by peer]
<Gumboot> I need to talk about cheese for a minute.
<Gumboot> Are there any expat EU or UK people in China who could give me some cheese-acquisition advice?
<ganbold_> cheese from china? interesting
<Gumboot> Hell no!
<Gumboot> Imported.
fragmint has joined #arm-netbook
<Gumboot> I know someone over there right now, and he doesn't have any cheese.
<Gumboot> The logistics of shipping cheese to him just aren't working out.
<Gumboot> I found some jarlsberg with a use-by date I could work with, but it's sliced. I don't think he'll thank me for that.
hg_5 has joined #arm-netbook
<Gumboot> I ask in here because there seem to be a few people speaking English and talking like they're in China.
<ganbold_> like lawrence seems in China
<Gumboot> lawrence: Spell colour.
<Gumboot> (Not taking cheese advice from an American!)
* mnemoc bets for .za or .nz
<Gumboot> Sorry, is that racist?
rsalveti has quit [Ping timeout: 256 seconds]
<Gumboot> .nz also has a non-ideal cheese reputation.
<Gumboot> (it's where I'm from, and I had to be reeducated when I moved to the UK)
<mnemoc> do south africans eat cheese?
<Gumboot> If the accent is anything to go by, then yes.
<mnemoc> i mean.... varieties
<mnemoc> hehe
<Gumboot> I'm relying on the Dutch influence to qualify them as cheese experts.
rsalveti has joined #arm-netbook
<libv> 1152MHz seems stable
<mnemoc> libv: open case?
<libv> closed
<libv> this thing needs to be transportable for fosdem
<libv> it's still a bit disappointment, not being able to reach 1.2GHz
<libv> +of a
<Gumboot> ssh-ing into my cubie starts with a long pause. Is that executing a function which could be NEON-optimised? If so, is that in openssl?
<libv> Gumboot: cat /proc/cpuinfo
<libv> or cat /sys/devices/system/cpu/cpu0/cpufreq/scaling_cur_freq
<mnemoc> Gumboot: 60MHz issue, or reverse dns
<hansg> Gumboot, that is likely something hitting the sdcard
<Gumboot> Hm. I thought I patched that.
<hansg> hansg, I've "optimized" a raspi ssh login time by mounting / ro and putting a tmpfs on /var/log, now it flies
<hansg> people always complain about how slow various arm devices are, but they forget that the sucky storage is often a big problem
<Gumboot> Why so much storage activity?
<hansg> libv, do you have NO_HZ set in your kernel .config ? I've been debugging a hang on A10s for 3-4 days in a row now, and it seems the periodic mode of the main timer is broken
<hansg> Reading the code, it really seems broken too, so things match up
<Gumboot> reverse dns, actually.
* Gumboot added an /etc/hosts entry.
<libv> i am running defconfig with fb compiled in atm
<hansg> I've it stable now by removing CLOCK_EVT_FEAT_ONESHOT from the timer struct in arch/arm/mach-sun5i/core.c, you may want to do the same for the sun4i, as that seems affected too
_jm has joined #arm-netbook
<mnemoc> hansg: rm also can't use NO_HZ on A10
<libv> ah, yes, NO_HZ
<libv> seems to be defconfig
<mnemoc> works fine for mostly everyone
<libv> and this is what might randomly kill a benchmark?
<mnemoc> test :)
<hansg> yep, as said better edit arch/arm/mach-sun4i/core.c and remove the CLOCK_EVT_FEAT_ONESHOT, the code for that is pretty fscked up (writes to the counter register, then tells the timer to do a reload from the reload-register, so the passed in value gets ignored). And it does not stop the timer before doing this as it should ...
<libv> at 1.2GHz, whereas the thing is stable at 1152MHz?
<_jm> Gumboot: for cheese in Beijing, you can try http://www.lefromagerdepekin.com/cheese_en.html . locally made, not imported.
<libv> unstable as in: hitting random assertions and segfaults
<hansg> libv, no, just trying to safe you from a possible cause of hangs. I've seen mysterious A10 (not s) hangs which could be explained by this too. Either way, the code is broken, and NO_HZ triggers its usage
<hansg> libv, iow this is unrelated from your overclock problem, but still something worth working around before your fosdem demo :)
<jinzo> Regarding various ARM SoC's and performance, a year and a half ago I worked for a company and we used the IGEP platform (the DM38xx SoC from TI) for some "server" work. We benchmarked and tweaked the ass out of that board
<jinzo> and yes, the relatively low amount of RAM kills you once the device can't write fast enough to the SDCard
<jinzo> even with top-of-the-cream Class 10 SDCards, the performance is sucky and rarely reliable (at least when you push them)
<Gumboot> _jm: This guy's in someplace near Hong Kong which isn't Shenzhen; but I know someone going from Beijing to Shenzhen soonish, so maybe they'll do a mercy run.
<jinzo> I really have to get something with eMMC to see how it pans out
<hansg> Just import some old amsterdam :)
<hansg> jinzo, isn't eMMC just an sdcard without the casing ? IOW won't it suck just as hard ?
<Gumboot> guangzhou, I think.
<jinzo> nope, it has advertised "up to 2x the performance of Class 10 SD Cards"
<jinzo> I don't expect miracles, but I would like to see how much claims like that actually hold water.
<_jm> Gumboot: DHL it :)
<Gumboot> That'll baffle them on two levels!
<hansg> The problem is cheese needs to be transported cooled, and also importing foods likely needs to meet some extra rules, so it likely won't get past customs
<_jm> hansg: LeFromagerdepekin is in Beijing, so at least the customs won't be a problem.
<hansg> true
<Gumboot> I think regulation (or at least obedience) is pretty lax in China anyway.
pcat has quit [Ping timeout: 256 seconds]
valhalla has quit [Read error: No route to host]
diego71 has quit [Read error: Connection reset by peer]
<oliv3r> gumboot, you are near or from holland? :)
<Gumboot> I'm in the UK. Originally from NZ.
<oliv3r> <- from holland, cheese capital of the world
<Gumboot> But I do like cheese.
<buZz> NEEEEEEEDERLAND
pcat has joined #arm-netbook
<hramrach> so long as the cheese does not explode ..
<hramrach> jinzo: at least on A10 the speed of SD is limited to just below 10MB/s it seems
<hramrach> by the mmc controller
<oliv3r> jinzo: Class 10 doesn't mean fast random access, class 10 only 'guarantees' fast performance in certain situations (writing sequential video usually)
<hramrach> so no matter hopw fast the card you do not get any more out of it
<hramrach> class6 and U1 perform the same
<jinzo> I wouldn't say they perform the same, we saw major performance differences in between the classes and manufacturers etc...
<hramrach> well, the U1 is supposedly faster for writing and random i/o
<hramrach> but not by much in practice with such mmc controller
<hramrach> I did not do any real thoutgh benchmarks
<jinzo> hramrach, I am more aiming at Exynos based products that frequenr eMMC (didn't see that much eMMC on the Allwinners SoC)
<hramrach> yes, eMMC or faster controller coud be better
<hramrach> also cards evolve
<hramrach> the U1 spec cards also state iops rating
<hramrach> something you would not see on a SD card a year ago
<hramrach> and was abysmal on most
<hramrach> but with A10 the single core severely limits what you can do to the card
<jinzo> yes I know that the resoults/experiance is old, but just wanted to share a bit. Hopefully stuff got a lot better, because in that time it was really limited use (and not even that chip)
<hramrach> with A10 the limit is the SoC
<hramrach> the cards can do better than the chip
valhalla has joined #arm-netbook
<jammi> some multi-core arm's with enough ram (basically 64bit required) might make even pretty nice servers
<hramrach> it depends on the server purpose
<jinzo> that's good to hear, but I would bet it's still quite hard to source a quality SD Card consistently.
<hramrach> an A10 would make nice server for some uses
<jammi> well, an A10 might be a replacement for a 15 years old server
<jinzo> with all the GPU/VPU overhead, not imo.
<buZz> or a 1 week old raspi
<buZz> :P
<jammi> or something requiring what an average server provided back then
<jinzo> jammi, debatable.
<hramrach> jinzo: if they have IOPS rating they at least try to answer the question how the card will perform
<jinzo> but that's far from what you can trust in a production run :D
<jinzo> if I would trust everything that the vendors say, just to sell their products... :D
<jammi> jinzo: you'll need several servers anyway if you want some actual redundancy
<hramrach> you can benchamrk them on something that actually can reach those speeds on mmc controller
<hramrach> neither an A10 nor an USB reader comes anywhere close so I can't say anthing about validity of their claims
<oliv3r> shipping will be more expensive then the cheese in any case
<hramrach> likely
<oliv3r> Get some cheep (aldi) vaccum seeled cheese and ship it 'express'
<oliv3r> or ask someone from a proper cheese country to do it for you. expect to pay around 50E for 2kg of cheese I guess
<hramrach> that's why finding local reseller is way better
<oliv3r> Keeping it cool, not so sure, specially this time of the year. It's not cooled where it's sold anyway. And 'best before'-date, non-issue, cheese sits on shelves aging for months. Really good cheese is usually 2+ years old :)
<oliv3r> yeah but our 'cheap crappy' cheese may still be miles better then expensive local cheese :)
<hramrach> I don't want to know the shipping costs for a loaf of cheese
rz2k has joined #arm-netbook
<oliv3r> shipping usually goes by the kg :p
<hramrach> yeah, they tend to be large
<oliv3r> cheese? it's quite heavy per volume
<mnemoc> a full piece of cheese is quite heave
<mnemoc> heavy
<oliv3r> expect a small box for 2kg :)
<mnemoc> so import probably happens in containers by ship
<jammi> just a little more dense than water, on average
<RaYmAn> oliv3r: aging cheese is probably in a rather temperature & moisture controlled environment though =P
<jammi> isn't that what the wax is for
<oliv3r> yeab but it usually is encased in wax and vacuum seeled, its deffinatly worth a try :p
<hramrach> that helps with moisture
<oliv3r> mnemoc: 'per a-mail'
<oliv3r> :p
<mnemoc> .oO(/me tries to guess what the 'a' stands for)o
<oliv3r> air*
<jammi> air?
<mnemoc> :)
<oliv3r> air-mail!
<oliv3r> mail per plane :p
<hramrach> solar powered airplane :p
<hramrach> maybe would make it this year ;-)
<oliv3r> ... cheese powered airplane!
* mnemoc has a pretty busy freight harbor right out of his window
<hramrach> unlike cheese powered plane solar powered planes are real
<hramrach> the speed is not stunning, though
<mnemoc> :)
<Gumboot> What is the most efficient way to convert cheese into electrical energy?
<Gumboot> And will there be enough power to run a cubieboard?
<oliv3r> burn it
<hramrach> to reduce power consumption the travel speed is reduced and they cannot keep in air indefinitely so need to land and recharge
<oliv3r> bigger solar panels
<hramrach> heavy
<hramrach> they are basically full panel plated
<hramrach> in good weather and with good batteries you might make it and fly overnight too but you need to plan for halts as well
<mnemoc> and where are the solar powered space elevators?!
<hramrach> where are any space elevaters?
vinifm has quit [Quit: Ex-Chat]
<mnemoc> they should have been around decades ago :<
<hramrach> not enough demand
<hramrach> too mauch material involved in construction
<hramrach> probably material reseearch required to figure out how to make something that high that does not fall apart
<hramrach> a solar plane is something you can build in your garage with some effort
<Gumboot> It'd really be better to build a space elevator outdoors.
<mnemoc> they are supposed to "hang" from a geostationary platform
<Gumboot> If you built it in your garage, it could only hang from the ceiling.
<mnemoc> :)
<libv> hrm, getting the mele to boot with vga should be as simple as altering the .fex, right?
<oliv3r> Gumboot: anyway, if you want a cheese sending service, I don't mind shipping some in your name :p (no guarantee's and warranties of course)
<mnemoc> libv: yes, just change the default screen
hg_5 has quit [Read error: Connection reset by peer]
<libv> rebuild the script.bin, stick it into the partition with boot.scr and uImage, and done...
<libv> yet it still seems to want to use the previous resolution
<libv> should i clear some of the space in the 1st mb for these changes to take effect?
<mnemoc> nah
<mnemoc> the u-boot env doesn't affect this
<libv> ah, i simply reboot, i do not stop the device
<mnemoc> :)
<mnemoc> you can also configure disp via /proc/cmdline these days
<libv> heh, you really do need a cold start
<mnemoc> libv: you can ioctrl() too
<libv> i know, but i need this thing to boot straight to vga&1024x768, so that it is projector friendly :)
<mnemoc> :)
_jm has quit [Quit: Loqui powered]
<slapin> libv: 3D projector?
* slapin waits for one
<ssvb> libv: I think I had to manually load "lcd" module (instead of compiling it into kernel) to get VGA output properly working some time ago, this stuff is really fragile
ZaEarl has joined #arm-netbook
<libv> ssvb: i think hansg must've fixed that one :)
<libv> ssvb: as it works fine now that i did a cold restart :)
<ssvb> I hope so :)
<ssvb> it's good that it works now
Gumboot has quit [Read error: Connection reset by peer]
iriki has joined #arm-netbook
Gumboot has joined #arm-netbook
L84Supper2 has joined #arm-netbook
L84Supper2 has quit [Changing host]
L84Supper2 has joined #arm-netbook
ganbold_ has quit [Remote host closed the connection]
<hansg> Hmm, HMM, I've found why A10 sun4i users have not been hitting the NO_HZ issue, the sun4i core code declares timer0 (the code for which is seriously bugged in periodic mode) as a timer with a priority of 100, where as the sun5i decleares it with a prio of 300
<hansg> So sun4i is using timer1 in NO_HZ mode and sun5i timer0
<hansg> And they have different code-paths (even though they are identical hardware wise)
<hansg> !@$%@&&@#$!!! this cost me bloody <beep> days to debug
<buZz> awesome that you found it!
Marex has quit [Ping timeout: 264 seconds]
Marex has joined #arm-netbook
<hansg> buZz, thanks, I wish I had a better feeling about it though. And now to clean up that bloody mess, well at least that will be more productive then banging my head against the same wall for days
<iriki> hey im losing _weeks_ and gaining white hairs in hacking a kernel in order to enable charging over otg, so you're not alone there
<buZz> :) just keep in mind, we are all making this here software better for everybody to use
<hansg> I know, it comes with the territory, still a guy has to vent his frustration every know and then :)
<buZz> yes, vent all you want ;)
<hansg> The timer0 oneshot mode is so broken it is unbelievable, esp. since there is a perfectly fine copy of the same code in aw_clocksource.c for timer1
tinti has joined #arm-netbook
<drachensun> iriki: I wasn't aware it wouldn't take a charge over the OTG. I'm pretty sure the Android devices will charge from OTG
<drachensun> iriki: Have you looked in the latest sources for fixes? I'm finding a lot of good stuff there
<drachensun> well
<drachensun> its ugly as sin but it generally fixes when I can see Android is working and the community kernel is not
matt__ has quit [Quit: Leaving]
hansg has quit [Quit: Leaving]
SouL_ has joined #arm-netbook
<SPG> Dimitris master 16b9227 rhombus allwinner_a10/orders/jimmis98.mdwn * http://git.hands.com/?p=rhombus.git;a=commitdiff;h=16b9227
gzamboni has quit [Ping timeout: 246 seconds]
gzamboni has joined #arm-netbook
<mnemoc> meh, hansg left :<
<TestModule> :\
<mnemoc> techn_: if you have a minute can you take a look at https://github.com/linux-sunxi/linux-sunxi/issues/113 ?
SouL_ has left #arm-netbook [#arm-netbook]
grog has joined #arm-netbook
pcat has quit [Read error: Connection reset by peer]
rellla has quit [Remote host closed the connection]
Jef91 has joined #arm-netbook
rz2k has quit [Read error: Connection reset by peer]
Quarx has quit []
<techn_> mnemoc: could that be becouse clocks
<techn_> no.. I have same div
<techn_> or maybe it is
<mnemoc> it's weird^3
<slapin> is there any original mk802 or mk802ii with working USB gadget?
pcat has joined #arm-netbook
rz2k has joined #arm-netbook
<slapin> mnemoc:
<mnemoc> slapin:
<hramrach> people:
<xenoxaos> ponies:
voronaam has joined #arm-netbook
<iriki> drachensun:
<iriki> im sorry I now only noticed what u wrote
<mnemoc> A10 != A10S
<Jef91> Alrighty, I wasn't sure that is why I asked :)
<Jef91> Are they totally different chips then? Different GPU and stuff?
<mnemoc> A10S is sun5i. like an A13 with ethernet and hdmi
<mnemoc> same mali400
<hramrach> any real difference in the arch?
<mnemoc> the core is different
<hramrach> they use different power controller but that's external part, presumably any would work
<mnemoc> i mean, very very low initialization and clocks
<mnemoc> sun4i != sun5i
<hramrach> so something that will make you compile a different kernel or u-boot but not seen i nthe resutling system
<mnemoc> axp15 is on the stage branches
<hramrach> yes, seen patches for that
iriki has quit [Quit: cya tomorrow]
<hramrach> why did they not use it with A10?
<mnemoc> A10S is cheaper and smaller
<mnemoc> AXP15 is also cheaper and smaller
<hramrach> what is 10s missing?
<mnemoc> over 100 pins :p
<hramrach> you can't use that many on most boards so no loss
<mnemoc> sata among them
<hramrach> oh, sata
<mnemoc> A10S is hdmi dongle oriented
<hramrach> there goes fast disk
<mnemoc> A13 and A31 are tablet oriented
<hramrach> most users of A10 never connected it so makes sense
<hramrach> even in tablet
<hramrach> I don't use it either
<mnemoc> the A10 is almighty
<hramrach> as far as AW goes
<mnemoc> :)
<hramrach> compare to exynos ;-)
<mnemoc> but even the cubieboard left half of the awesomeness out of the headers :<
<mnemoc> exynos has sata?
<mnemoc> iirc exynos4 doesn't even have native usb
<hramrach> I heard it has sata
<hramrach> but don't think there is borad that uses it
<hramrach> only andriod devkit so useless anyway :<
<TestModule> Were the repos updated for that timing bug earlier?
<hramrach> hmm, maybe only exynos5 has
<slapin> mnemoc: is there any original mk802 or mk802ii with working USB gadget?
<hramrach> exynos 4210 is supposed to have 2 sata channels
<mnemoc> slapin: i assume usb gadget works on 3.0/android
<mnemoc> slapin: i haven't tested it. and don't have any hdmi dongle :|
<slapin> mnemoc: I mean stock firmware, I can attach keyboard over converter, but no USB device, and no android debugging
<mnemoc> no clue, sorry
<TestModule> slapin, the default 1gb mk802 can host devices over the lapdock over the otg
* slapin tries to make ad-hoc children-entertainment device to be able to do some hacking without interrupt storm
<TestModule> but only one at a time so you have to remove the kb/trackpad
<slapin> TestModule: host works, I try to make USB device (gadget) mode work
<TestModule> Ah
<TestModule> Isn't that in the kernel configs?
<slapin> TestModule: host port is ok, OTG port works only in host mode for some reason
<TestModule> Ahhhh
<TestModule> I had to do a solder mod originally
<TestModule> shorting 4->5
<slapin> USB pins?
<TestModule> yeh
<slapin> weird...
<TestModule> flipped the back portion up with a knife, then stuck a super small bit of tin-foil between, then solder
<slapin> ID to GND? strange
<TestModule> oh wait, possibly flip that
<TestModule> the ones on the far left
<slapin> The type of the plug inserted was detected by the state of the pin ID (the mini-A plug has the ID pin grounded while the ID in the mini-B plug was floating).
<slapin> hmmmm
<slapin> ID to gnd = host, ID = floating = device, so ID to GND means it goes to host mode (and it already works like that).
<TestModule> No idea why, it didn't make sense to me, but it was posted as a fix.
<TestModule> It works
<TestModule> I know :)
<TestModule> very nice diagrams though, jesus
revident has joined #arm-netbook
<Jef91> Anyone know if it is possible to get GPU accelerated video playback on the A10 devices under linux yet?
<mnemoc> sure
<techn_> Jef91: vpu accelerated
<mnemoc> :)
<Jef91> Does that work with players like vlc/mplayer? What do I need to do to go about getting that working?
<jinzo> Jef91, there were some efforts with VLC, search the mailing list
<jinzo> also, you're looking for cedarx
<jinzo> oh techn_ link ^
hark has quit [Ping timeout: 276 seconds]
Xlab has joined #arm-netbook
SouL_ has joined #arm-netbook
SouL_ has left #arm-netbook [#arm-netbook]
<TestModule> Jef, they get it under XBMC didn't they?
<Jef91> techn_: which players have the best support for this decoder?
<techn_> Jef91: currently only ones are vlc and xbmc
gimli has joined #arm-netbook
<techn_> no one has implemented support below generic api
<Jef91> kk techn_ I'll take a stab at compiling VLC with support for it then
* mnemoc still hopes someone will start writting the libva plugin
<Xlab> Hi all. Could anybody explain, why I getting this in my custom android fw for A10:
<Xlab> E/gralloc (   80): UMP open failed
<Xlab> E/FramebufferNativeWindow(   80): couldn't open gralloc HAL (Operation not permitted)
<Xlab> E/FramebufferNativeWindow(   80): couldn't open framebuffer HAL (Operation not permitted)
<Xlab> E/gralloc (   80): UMP open failed
<Xlab> Kernel modules for mali are r2p4 and all the libraries libMali, libUMP, libEGL, libGL* have been taken from the original rom (they're r2p4 too).
<Xlab> I give up :(
<Jef91> Has anyone played around with formatting the nand in the MK802 and installing Linux to it isntead of Android?
<GeorgeIoak> anyone else working on getting touchscreen working? i've got tslib loaded and ts_calibrate works and then i installed https://launchpad.net/ubuntu/precise/armhf/xserver-xorg-input-tslib/0.0.6-7build2 but am a bit stuck at this point
<techn_> GeorgeIoak: depends on your ts driver
<Xlab> I've got goodix_touch working, one finger - leftclick, two fingers - right click
<techn_> which one you are using.. I have tested gt811 and ft5x drivers
<techn_> and they work
<techn_> but not like Xlab's :p
<GeorgeIoak> i'm just using a resistive screen with the sun4i_ts module
hark has joined #arm-netbook
<GeorgeIoak> without a conf file in xorg.confd everything works except the touchscreen, if i make a 20-ts.conf file then i loose keyboard, mouse, and touch still doesn't work
Xlab has quit [Ping timeout: 255 seconds]
Xlab has joined #arm-netbook
<Xlab> My goodix_touch sends MT events, so I used https://gitorious.org/gabrbedd/xorg-x11-drv-mtev
<GeorgeIoak> hmmm, i don't think 4-wire touch can send multi-touch
<GeorgeIoak> do you have any hints how i can debug the touch in xorg? since ts_calibrate works it seems like i am close?
<Xlab> nope :(
<specing> 4 wire as in ... SPI?
<xenoxaos> 4 wire as in resistive?
TestModule has quit [Ping timeout: 252 seconds]
focus_it has joined #arm-netbook
TestModule has joined #arm-netbook
<GeorgeIoak> 4 wire as in resistive so it directly connects to the A10
ZaEarl has quit [Ping timeout: 256 seconds]
Xlab has quit [Quit: Leaving.]
traeak has quit [Remote host closed the connection]
traeak has joined #arm-netbook
datagutt has quit [Quit: kthxbai]
Jef91 is now known as Jef91|AFK
freakazoid0223 has joined #arm-netbook
dyoung-away is now known as dyoung
eebrah has joined #arm-netbook
focus_it has quit [Remote host closed the connection]
bfree_ has joined #arm-netbook
bfree has quit [Ping timeout: 252 seconds]
eebrah has quit [Read error: Connection reset by peer]
Jef91|AFK is now known as Jef91
eFfeM has joined #arm-netbook
hg_5 has joined #arm-netbook
eebrah has joined #arm-netbook
eebrah has quit [Ping timeout: 264 seconds]
eebrah has joined #arm-netbook
Jef91 is now known as Jef91|AFK
tinti has quit [Read error: Connection reset by peer]
eebrah has quit [Ping timeout: 252 seconds]
gimli has quit [Ping timeout: 272 seconds]
revident has quit [Quit: Laywoo-D3, look it up, it's awesome.]
<TestModule> Does the BSP disable / remove android-only features from the kernel configs when a non-android kernel is being compiled from the sun**_defconfig ?
<mnemoc> different defconfig is used for android and not-android
<TestModule> awesome, thanks
eFfeM has quit [Quit: Leaving.]
dyoung is now known as dyoung-away
ludde is now known as brakew
freakazoid0223 has quit [Quit: Leaving]
<hramrach> so succesfully restored original linux.ini, replaced u-boot with lychee-dev sun4i
<hramrach> now it complains there is no boot partition
<hramrach> sp I guess the default command is different with this one?
<hramrach> is boot.scr supported?
marcan has quit [Quit: Coyote finally caught me]
eebrah has joined #arm-netbook
DEAT has quit [Read error: Connection reset by peer]
DEAT has joined #arm-netbook
cheng has joined #arm-netbook
<mnemoc> hramrach: no uEnv.txt or boot.scr supported by lichee-dev
<mnemoc> hramrach: you need to tweak the env yourself
<hramrach> and where is the env?
<hramrach> some random part of flash before first partition?
<mnemoc> "*** Warning - bad CRC, using default environment"
<mnemoc> Hit any key to stop autoboot: will get you to u-boot's prompt
<mnemoc> there you can play
marcan has joined #arm-netbook
<hramrach> it gives propmt anyway since it cannot execute the default command
<hramrach> will try to build with something more useful
<hramrach> thanks
<mnemoc> play on the prompt first
<mnemoc> once you are happy, you can make your own "default env"
<mnemoc> you basically need: fatload nand 0 0x48000000 uImage
<mnemoc> bootm 0x48000000
<mnemoc> instead of reading the raw partition
<WarheadsSE> root
<WarheadsSE> (wrong window derp)
<mnemoc> at least it wasn't the next line
<WarheadsSE> :P
<WarheadsSE> it would have been root too
<WarheadsSE> fresh rootfs :P
<mnemoc> =)
grevaillot has quit [Ping timeout: 248 seconds]
eebrah has quit [Quit: ChatZilla 0.9.89 [Firefox 18.0/20130109143300]]
<hramrach> btw the defaut bootargs are bogus
<hramrach> missing space after root
<WarheadsSE> nice
<hramrach> now they are not passed
<hramrach> need to build cammndline into kernel again
<mnemoc> hramrach: the linux-sunxi ml is eager to get your patches :)
<hramrach> mnemoc: is that expected that when you bootm from prompt the bootargs from default env are passed
<hramrach> but when you bootm from the default command they are not
<mnemoc> something in your script needs to setenv bootargs ....
<hramrach> it's the deafult env
<hramrach> I did not setenv on the prompt either
cheng has quit [Quit: Leaving]
<hramrach> hmm, it is somewhat more complex
<hramrach> the default env sets setargs which contains setenv bootargs
<hramrach> anyway it's compiled in and recompiling linux is fater than recompiling u-boot
<hramrach> faster
<WarheadsSE> mnemoc: inbound pull request for the a13-olinuxino-micro for fex :)
<mnemoc> \o/