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/
Guest10529 is now known as hp__
<Turl> mnemoc: ping
freakazoid0223 has joined #arm-netbook
stefanro has joined #arm-netbook
stefanro1 has quit [Ping timeout: 255 seconds]
cheng has joined #arm-netbook
popolon has quit [Quit: Quitte]
Jeff91 has joined #arm-netbook
Jeff91 is now known as Jef91
Jef91 has quit [Changing host]
Jef91 has joined #arm-netbook
<ManoftheSea> hmm, okay. Must put aside until I get my USB-TTL-UART
Jef91 is now known as Jef91|AFK
Jef91|AFK has quit [Quit: Leaving]
Jef91 has joined #arm-netbook
freakazoid0223 has quit [Quit: Leaving]
Jef91 has quit [Ping timeout: 244 seconds]
freakazoid0223 has joined #arm-netbook
Quarx has joined #arm-netbook
mtbg has quit [Quit: Konversation terminated!]
mikey_w has quit [Read error: Connection reset by peer]
mikey_w has joined #arm-netbook
Kasbert has quit [Quit: Leaving]
antonl has joined #arm-netbook
rellla has joined #arm-netbook
<Mehhh> Is there a kernel config that already has the x86 stuff removed?
<Mehhh> para ports, fibre channel, scsi, etc
<hno> Mehhh, SCSI is needed if you want to talk to USB sticks.
<Mehhh> Oh sorry, it was iSCSI. Manually unchecked
sspiff has joined #arm-netbook
<rm> iSCSI is not "x86 stuff" it's a network block device protocol
<Mehhh> don't worry rm, i removed fibre channel just to be fair and balanced
<hno> lpt support is needed if you want to connect a USB printer.
<Mehhh> Good to know. Thanks. Have you guys tested with LTO?
<Mehhh> and which version of the malis are the best right now? I tried 0 / 0 and it was Mehhh for regular X usage.
raoulh has joined #arm-netbook
popolon has joined #arm-netbook
<Mehhh> I can't see the folders on GH, did the file / folder situation with BSP get sorted?
<oliv3r> oh wow, just saw hans de goede posted some patches on linux-sunxi, nice to see some 'hardcore' commiters show interest :)
<oliv3r> hno: if you connect a usb printer via libusb, is lpt support still needed?
arete has joined #arm-netbook
ssspiff has joined #arm-netbook
sspiff has quit [Ping timeout: 252 seconds]
<Mehhh> I have commitment issues. By that I mean i can't push it correctly.
<Mehhh> btw, why is the res on the CTP listed as max 800x480 for the mk802 fex?
<Mehhh> Technically it should support up to a 1080p
<Mehhh> Even the stocki Goodix-TS panels can go higher than the stock listed
<libv> ctp?
<slapin> Mehhh: how are you going to add CTP to mk802?
<Mehhh> I'm not, personally. It's possible, but i'm saying that the resolution is off
<slapin> Mehhh: how that is even possible?
<Mehhh> 1024x768 is the norm
<Mehhh> USB otg and hdmi out
<slapin> Mehhh: cool, any links?
<Mehhh> you can get power and data from the same
<Mehhh> no idea, i lapdocked mine
<Mehhh> there are some threads about it on rikomagic and miniand
<libv> ah, capacitive touch panel
<Mehhh> yes
<Mehhh> The confusion was regarding the documentation on their site
<Mehhh> where it says to set to 800x480/600 as an example, but saying that the res = actual res of the screen you're using
arete has quit [Quit: Leaving]
rz2k has joined #arm-netbook
cheng has quit [Quit: Leaving]
<Mehhh> Do you guys want me to make a bash script that places the files that are missing in the BSP folders into there using git pull ?
<slapin> Mehhh: any photos re CTP over usb/hdmi?
<Mehhh> I am not using it on a touch screen
<Mehhh> i'm just pointing out that the values are wrong
<slapin> Mehhh: then why CTP?
<Mehhh> and that the actual base res has a different config area
<Mehhh> because it's listed in there
<Mehhh> i'm not using one, i'm just pointing out the file is wrong
<slapin> Mehhh: so all you talk about is theoretical?
<Mehhh> No, it's a bug
<slapin> Mehhh: my device can do 1366x768 then
<slapin> Mehhh: where is a bug?
<Mehhh> It's artificially capping your res for the TS
<Mehhh> the lcd res is just base, stock, etc. the ctp is a maximum value
<slapin> Mehhh: capacitive touches are actually low-res
<Mehhh> so it doesn't hurt to have the lcd res different and simply jack the max TS to 1080p
<slapin> Mehhh: all hi-res touches are resistive
<Mehhh> slapin, the max isn't 800x480
<Mehhh> most of the tablets that come with that panel are 1024x768
<slapin> Mehhh: your thingers are big enough so you won't notice
<Mehhh> 800x480 should not be the maximum value when the default is far higher
<Mehhh> Also, please see Transformer Prime by Asus
<slapin> Mehhh: screen res is irrelevant too touch res
<Mehhh> regarding your theory
<Mehhh> slapin, please stop trolling.
<slapin> s/too/to/
<ibot> slapin meant: Mehhh: screen res is irrelevant to touch res
<Mehhh> I've already looked it up
* slapin is never troll, well, almost
<Mehhh> See, this is from the a10 tablet listed in the boards
<Mehhh> ctp_screen_max_x = 1024 ctp_screen_max_y = 768
<slapin> Mehhh: any particular link?
<Mehhh> What do you want to see?
<slapin> Mehhh: a link to the appropriate config
<Mehhh> A simple check on any chinese website will show you
<Mehhh> libv, has anyone tried compiling the kernel with LTO flags?
<slapin> Mehhh: these configs are from working devices, why?
<Mehhh> I don't understand your question.
<Mehhh> See
<Mehhh> The max res is just off on certain others so if plugged into a touchscreen it is listed incorrectly
<slapin> Mehhh: so you get config from LCD-less device, and claim, that their touch screen configuration is wrong, am I right?
<Mehhh> Yes
<Mehhh> because it has the capability to support it, and the config is wrong.
<slapin> Mehhh: Ahhh, I've got an idea. I think they got it from default kernel configuration, so didn't touch it.
<Mehhh> Also, the boot clock on the MK802 is 912mhz not 1008mhz
<slapin> Mehhh: that depends on clone
<slapin> Mehhh: I have collected enough variants already...
<Mehhh> So what is yours clocked at?
<Mehhh> The default for major brands seems to be 912
<Mehhh> Rikomagic, MiniAnd, Pengpod
<slapin> Mehhh: mine are nameless, they come in all the same green boxes called android 4.0 mini pc
<Mehhh> I know, I have a few ;)
<Mehhh> Android reports as 1.5ghz but actual clockspeed is 912mhz
<slapin> Mehhh: inside can be amlogic, rockchip, a10, with 512 or 1G memory size, with PMU (axp209) or without, with male or female HDMI connector...
<Mehhh> i'm talking about the MK802, A10
<slapin> Mehhh: what is this 1.5GHz b/s for?
<Mehhh> the regular one
<Mehhh> No idea, fake BS
<Mehhh> They are adding a fake value to the GPU clock I think
<jinzo> all the frequencies combined afaik
<Mehhh> yeh
<slapin> Mehhh: which one 512/1G, with PMU or without is regular?
<jinzo> co CPU + GPU
<Mehhh> I have the 1G
<slapin> jinzo: and how android can report that?
<Mehhh> regular 1gb mk802
<Mehhh> slapin, any benchmarking software
<slapin> mine reports 890MHz for one attached to TV
<Mehhh> Is the govenor on demand?
<jinzo> slapin, they modify the source or something? I wouldn't know.
<Mehhh> Also, because of HDMI the default res i'm setting to 1280x720
<Mehhh> 720p should be the minimum for HDMI
<slapin> Mehhh: I have not touched this one, which one is default there I don't know.
<Mehhh> It's fine i just hope they take the commit. I still need to clone the 512 and make the 1G fex for some guy here i can't remember
<slapin> 720p is not minimum
<Mehhh> but they don't have the right RAM values
<Mehhh> You have an HDMI tv with less than 720p?
<slapin> I have a small TV with 720x460 resolution, HDMI input works well
<Mehhh> Does it downscale?
<slapin> s/720x460/720x480/
<ibot> slapin meant: I have a small TV with 720x480 resolution, HDMI input works well
<slapin> Mehhh: I don't think so, as device is multires and reports the resolution properly
<slapin> my laptop xrandr -q shows from 320x240 to 720x480
<slapin> so I think it doesn't care much for 'p's with output
<slapin> and mk802's output configuration allows selecting such resulution, among with 640x480, which works too
<Mehhh> Yeh, it seems to downscale
<Mehhh> i just tested it
<slapin> Mehhh: what downscale?
<Mehhh> I put 1080p and shoved it in a 1366xwhatever screen
<Mehhh> it did edge-to-edge
<Mehhh> it downscaled
<slapin> Mehhh: have you tried to configure your output resolution in settings?
<Mehhh> it's just shrinking
<Mehhh> most images don't have a settings panel ;)
<slapin> Mehhh: mk802 one does
<Mehhh> i had mine hard set for 1080p
<Mehhh> i mean in linux
<Mehhh> not android
<slapin> Mehhh: as I have never seen such device without these settings
<Mehhh> Linux, not Android
<slapin> Mehhh: I tested with android
NAiL has joined #arm-netbook
<Mehhh> Yes, i was in linux.
<slapin> Mehhh: xrandr?
<Mehhh> Look, I wasn't writing a report on it, I was just testing the ability and norm to downscale
<slapin> anyway, it should work from Linux too
<Mehhh> for non-high res devices
<Mehhh> So it does downscale
<Mehhh> That's all I wanted to know
<slapin> Mehhh: it should work without downscaling, too.
<Mehhh> I'm testing that in android right now
<Mehhh> so far it looks like crap
<slapin> Mehhh: as most dumb panels don't know which resolution they use
<slapin> Mehhh: even if you somehow attach these with HDMI
<Mehhh> ?
<Mehhh> 1080p downscaled
marcan` has joined #arm-netbook
dyoung-away has joined #arm-netbook
Hexxeh has quit [*.net *.split]
dyoung has quit [*.net *.split]
marcan has quit [*.net *.split]
markatto has quit [*.net *.split]
NAiL_ has quit [*.net *.split]
dyoung-away is now known as dyoung
dyoung has quit [Changing host]
dyoung has joined #arm-netbook
<slapin> Mehhh: where have you got this laptop from?
<RaYmAn> slapin: I would imagine it's something like this: http://www.motorola.com/us/consumers/Motorola-Lapdock-100/73945,en_US,pd.html
<Mehhh> In America they are 10$
<Mehhh> So i'm just balling all crazy with random screens
<slapin> Mehhh: when I attach android to my small tv, everything becomes quite big, and resolution shown is 720x480, and when I set it to 1080p, it is just black screen without display.
<Mehhh> Really weird
<Mehhh> 720x480 = weird
<Mehhh> It should do more than 480p tbh
<Mehhh> try @ 720
<Mehhh> 1280x720
<slapin> Mehhh: this is display top resolution
<Mehhh> I know, I'm having you test it
<slapin> Mehhh: 720x480
<slapin> Mehhh: not mk802's
<Mehhh> I know, now try to set it to 720
<slapin> nothing
<Mehhh> k
<slapin> darkness
<slapin> 640x480 and 720x480 both work
<Mehhh> btw, TS works
<slapin> Mehhh: any docs on how you did all this?
<Mehhh> the lapdock?
<RaYmAn> if it's a USB touchscreen it should be relatively obvious to get it working
<Mehhh> Yeh, slapin was just giving me a hard time because he didn't think it was possible ;)
<Mehhh> So i found him proof
<Mehhh> I am forking the mk802 fex
<slapin> I have old ai touchbook, which is even GPLed, but CPU is not starting for some reason, so I'd like to take its screen to use...
markatto has joined #arm-netbook
<RaYmAn> Mehhh: there's a long way from usb touchscreen to directly connected touchscreen :P
<Mehhh> oh rayman, maybe you know this
<Mehhh> Has anyone done LTO flag optimizations to the kernel
<Mehhh> for resting
<Mehhh> teting
<Mehhh> grrr "TESTING"
<slapin> Mehhh: how do you attach lapdock - HDMI?
<Mehhh> Yes
<Mehhh> Micro female hdmi to mini male hdmi adapter
<Mehhh> Then, micro female usb to mini male usb adapter
<slapin> Mehhh: any links to adapters?
<Mehhh> They won't help you
<Mehhh> USA sites
<Mehhh> try ebay or aliexpress.com if you want china / worldwide
<slapin> Mehhh: not found any lapdocks there
<slapin> Id' prefer hdmi->lvds converter actually
<slapin> as I have tons of various screens
<Mehhh> lvds is universal?
<Mehhh> Ugh, I wish Motorola sold these stupid Starconn adapter replacements
<Mehhh> hmmm
<Mehhh> for some reason my git pulls are pulling 3.0.52 not 3.0.8
<Mehhh> for bsp
<slapin> lvds is universal, connectors are not
<Mehhh> It's weird, the lapdock has what looks like an LVDS connector, but soldered to USB and hdmi
niftylettuce_ has quit []
<slapin> Mehhh: as I said, connectors are often non-standard, or even absent.
<libv> i have a ton of epia LVDS encoder modules
<libv> only now are suitable cables being made available over aliexpress
<slapin> libv: any links?
<libv> i would've had to contact an actual cable manufacturer to get them made for me (i think they have some just laying around)
RITRedbeard has quit [Read error: Connection reset by peer]
<libv> or pay iirc 700+ for suitable crimp tools
<libv> now cheap chinese crimp tools are available too
RITRedbeard has joined #arm-netbook
<libv> now that unichrome is a dead end chipset, it doesn't matter anymore
<slapin> libv: why?
<libv> who still buys VIA based hw?
<rm> is there even any VIA-based h/w to buy :)
<libv> of course there is, but there is no point
<libv> atoms are cheaper and way better supported
<libv> more reliable
<libv> better performing
<libv> VIA is as dead as it gets
<libv> and it was always a bit of a playground for a member of the chen family
<slapin> lots of Russian customers buy epia chips packed in PC104 equipment for lots of $$$ because of marketing
<libv> slapin: i was self-employed selling consulting services around my unichrome graphics driver work in 2006 and the first half of 2007
<libv> plenty of contacts, but financially a very very negative endeavour
<libv> it wasn't a useful way of spending anyones time then
<libv> and this was more than a year before the atom was released which killed off VIAs market completely
<slapin> libv: actually quality currently is no concern
<libv> it's always been cheap crap
<libv> and the customers were equally good.
<slapin> libv: as all embedded market here
<slapin> libv: the products found their customers
<slapin> finally
<libv> well, this is how the allwinner stuff should be treated from our point of view as well
<libv> use it as marketing, use it to scratch your own itch
<libv> and if you are lucky, you can make a few bob on the side selling services around it, but only if you are really lucky
<libv> don't sell it full time, or you'll starve
tinti has joined #arm-netbook
<libv> in any case, unichrome was huge in 2006 compared to what it is in 2012/2013
<slapin> they sell a board with a lithium battery, as device complying with -40 +85 C temperature range, without additional requirements, and everything is fine to everyone
<slapin> no support calls ever
<libv> an even though i have a about 35 different unichrome devices, from mini-itx to atx to laptops and netbooks, sitting around in this playroom/office of mine
<slapin> huge government contracts
<libv> i will not waste time developing it anymore
<libv> and the old laptop scrounged panels i have, and the ton of lvds modules, will never get wired up
<libv> i will from time to time, pluck some rare board with interesting display bits off of ebay though, but that's not rational behaviour
* slapin thinks how much will cost for Chinese devs to add -40 +85 C claim to A10 devices, in addition to some cheap GSM and GPS module?
<libv> -40 +85 is a lie
<libv> for VIA, it is a massive lie
<slapin> libv: it is always a lie
<libv> many unichrome devices i have have developed cold joints over the years
<RaYmAn> it probably makes a good deal of assumptions on the enviroment, lol
<slapin> libv: but that's current govt requirement in .ru for equipment
<libv> i should have them reflowed, but again, i cannot be arsed
<slapin> so there is massive lie. if you won't lie, govt agents will buy stuff from your neighbor, who lies better than you :)
<slapin> so at the moment I need some dev who won't be afreid to claim -40 +85C to standard A10 stick design
<libv> you're on your own with that.
<slapin> libv: I claim all the responsibility
<libv> then find out the testing requirements and start testing the cubieboard or something similar
<mnemoc> i think i read that was the range for the SoC itself, not the boards
<slapin> libv: that's too cumbersome for me, I just need customized devices (from 10000 to 100000 pcs.) with that claim (and papers supporting the claim), no testing will occur here (and actuall running environment will be +10 - + 60 C)
<Mehhh> libv, have you ever seen an LVDS look-a-like, possibly LVDS output to HDMI and USB from the same connector?
<libv> Mehhh: no, and if so, there would be dedicated electronics for it
<slapin> Mehhh: can you take picture of it somewhere?
tinti has quit [Ping timeout: 250 seconds]
tinti_ has joined #arm-netbook
<Mehhh> Need to update the MK802 fex to this: https://github.com/Miniand/mk802-u-boot-config/blob/mele/script.fex
stefanro has quit [Quit: Leaving.]
<slapin> could anybody sell me the lapdock?
<Mehhh> Good luck, people are going nuts tracking them down to sell on ebay
<Mehhh> prob your best bet
stefanro has joined #arm-netbook
<rz2k> hardkernel sent first shipment of odroid-x2's.
tinti_ has quit [Quit: Leaving]
<Mehhh> slapin
<Mehhh> are those figures correct for your mk802.c file
<slapin> Mehhh: ?
<slapin> Mehhh: what do you mean?
<Mehhh> That mk802.c file you made
<Mehhh> on github
<slapin> you mean memory settings?
<slapin> that one was for original 512MB version
<slapin> these are quite pessimistic values, so these do work for all variants I have
<Mehhh> Awesome.
<Mehhh> I'm using some of them to clean up this mess
<Mehhh> then i'll overwrite my own 1GB to get a solid 1GB file up
<slapin> Mehhh: better add new machine name for different settings
RITRedbeard has quit [Read error: Connection reset by peer]
RITRedbeard has joined #arm-netbook
<Mehhh> nononono, i only took the ram settings
<Mehhh> ;)
<slapin> Mehhh: don't do dangerous things with them, please
<Mehhh> I'm not
<Mehhh> They were using really out-dated fex files and had tons of missing values for the MK
<slapin> Mehhh: that's cool, but it'd be better if you helped with new things, like new spl-based u-boot, DT, machine support, and not playing with random things, buy who am I to tell people what to do
* slapin is back to $$$ work
Hexxeh has joined #arm-netbook
<Mehhh> :P
<Mehhh> Holy crap... Hexxeh?
<Mehhh> Someone peel me away before i scare him off
<orly_owl> nah
<Mehhh> slapin, i'm just here for the MK802, so i'm trying to automate the creation. You'll be happy to know i'm making a bash to populate the bsp folders
<Mehhh> This wack-a-mole file slapping has to be more organized
<slapin> Mehhh: what bash?
<rz2k> someone should check new patches @ ml and approve them to main branch or to staging/
<Mehhh> Bash script
* rz2k too noobish in git for that
<Mehhh> The BSP should just be pulling from the other repos entirely
<slapin> Mehhh: what automation? You make script to build device image or what? can you help somewhat building OE image for MK802's? It requires some compiling, but builds from source, not tossing binary files...
<Mehhh> I'm talking about populating the folders
<Mehhh> which aren't
<Mehhh> notice how the board folder on the BSP repo has older files
<Mehhh> if you simply run a bash script that populates the folders with the latest, it's never a problem
<mnemoc> rz2k: I can do the merging, but after people acks them
* slapin does have that list subscription problems
<mnemoc> rz2k: the ts patches for example have been 2 weeks waiting for someone to test on android
<slapin> mnemoc: can you subscribe me to the list? google group opens ok over web, but not subscription
<mnemoc> slapin: try sending an empty mail to linux-sunxi+subscribe@googlegroups.com
<mnemoc> rz2k: the stage branch probed itself useless. zero feedback and several problems reported after merging it to the main branches
<libv> too few people spread too thinly :(
<slapin> mnemoc: mail sent, but no reply at all, will wait a few hours, but it seems it won't work this way too.
<Mehhh> Are you guys against having a bash script to populate each?
<Mehhh> for the BSP
<Mehhh> for example, one in the main folder, or perhaps each subdir
<mnemoc> slapin: give me your name/address on /q to send you an invitation
<Mehhh> for cloning the latest from the official section
<mnemoc> Mehhh: the bsp clones the subprojects itself, when needed
<Mehhh> How do I make it need it?
<mnemoc> `make foo`
<mnemoc> if foo is not cloned, it will do it
<Mehhh> If you do a file compare, it's just not the case
<Mehhh> ahhh kk
<mnemoc> considering how fat some repos are it makes no sense to force people to clone them when they aren't going to be used
<Mehhh> So you're saying I can prune the kernel ? ;)
<slapin> can we somehow git gc repo on github?
<Mehhh> 500MB :<
<mnemoc> slapin: github seems to keep a single repo for all forks and only distinguish heads
<slapin> some original Allwinner repos are extremely large
<mnemoc> you can tell git to only clone certain heads
<slapin> mnemoc: I think it just uses --reference
<slapin> mnemoc: see pm
<mnemoc> slapin: pick a hash of any random fork or subfork of linus' tree
<mnemoc> slapin: and then access it from any other random fork or subfork of it
<slapin> mnemoc: --reference
<mnemoc> :)
<slapin> mnemoc: that doesn't removes the need to git gc repo from crap
<slapin> mnemoc: because AW repo does have too much crap it downloads for too long time, unlike Linus's or mine
<Mehhh> yes
<Mehhh> they don't even exclude any folders
<Mehhh> rofl
<slapin> mnemoc: if you push+ a lot, or rebase, your repo contains unlimited amounts of crap
<mnemoc> i suppose the ideal is to clone only the `sunxi-3.0` head
<mnemoc> slapin: /q
<slapin> mnemoc: replied
<Mehhh> ok so
<Mehhh> does anyone mind if i place the git info in the subdirs
<Mehhh> of the bsp
<Mehhh> so you can at least pop a term up and pull
<Mehhh> lots of things aren't
<mnemoc> Mehhh: it feels like you are trying to solve the wrong problem
<Mehhh> just trying to make it automagic
<mnemoc> it is automagic
<Mehhh> Automated
<Mehhh> I have one target device so I need to make the steps as fast as possible for crapping out images on the fly
<mnemoc> `./configure mydevice; make`
<Mehhh> I still have to figure out an easier way to add commonly requested features...
<Mehhh> prob python + GTK
<Mehhh> mnemoc, that doesn't do all the Mali crap
<mnemoc> `make` alone will only do `make hwpack`
<Mehhh> yeh
<mnemoc> and that includes mali libs
<Mehhh> that imports the rp0/0 ones or whatever though
<Mehhh> which suck w/X
<mnemoc> as long as the kernel has p0, the bsp will have p0
<mnemoc> so again, you are "fixing" the wrong problem
<mnemoc> bbl
cheng has joined #arm-netbook
<Mehhh> now i'm getting errors
<Mehhh> Odd, after config'd: make[1]: *** No rule to make target `mk802'. Stop.
<mnemoc> that means u-boot doesn't know about that board
<slapin> well, people have you ever read u-boot README. which recomments doing make board_config then make all?
<slapin> I don't see any changes preventing you from doing this
* slapin does a lot of typoes today, but speacs a lot due to caffeine
<slapin> s/speacs/speaks/
<slapin> samn
<slapin> *damn
<mnemoc> mk802_config will fail in the same way as there is no mk802 board
<mnemoc> but yes, I agree that needs to be change in the bsp's Makefile
<mnemoc> ${board}_config instead of ${board}
<oliv3r> slapin: have you tried to unsubscribe and resubscrube to the googlegroups list? I've had the same issue (different group) and that seems to work
tinti has joined #arm-netbook
<Mehhh> $(Q)sunxi-tools/fex2bin sunxi-boards/sys_config/$(SOC)/$(BOARD).fex > $(BUILD_PATH)/$(BOARD).bin
<Mehhh> that line, right?
<slapin> oliv3r: mail interface ignores my another email address
<slapin> oliv3r: at all, while it is already subscribed to another google group, which works and mail arrives.
<slapin> oliv3r: this is wicked bug of google groups
orly_owl has quit [Ping timeout: 255 seconds]
DEAT_ has quit [Ping timeout: 240 seconds]
DEAT_ has joined #arm-netbook
<Mehhh> slapin do you know which line he was talking about? I'm like sleepy zombie status atm
* rz2k done reading g2d stuff
<rz2k> should be easy to wire it with EXA, only 4 ioctls + bit masks.
<rz2k> anyone wants to help&
<rz2k> ?
<slapin> Mehhh: which line of what? I'm a bit out of context here
* slapin got a monday problem
<Mehhh> [08:36] <mnemoc> mk802_config will fail in the same way as there is no mk802 board [08:37] <mnemoc> but yes, I agree that needs to be change in the bsp's Makefile [08:37] <mnemoc> ${board}_config instead of ${board}
<Mehhh> i have a sleep problem :P
<Turl> slapin: tried unsubscribing and then subscribing again?
<slapin> Mehhh: the line, which builds u-boot
<slapin> Mehhh: dunno which one
<slapin> but that line looks weird
hg_5 has joined #arm-netbook
<slapin> Turl: I can't do it from web interface and from mail interface, it totally ignores all attempts.
<rz2k> thats an interesting statement from ARM: http://forums.arm.com/index.php?/topic/16334-opencl-for-arndale-board-and-nexus-10/page__p__40152#entry40152
<oliv3r> linux-sunxi+unsubscribe!
<oliv3r> @googlegroups ..
<Turl> slapin: email to linux-sunxi+unsubscribe@googlegroups.com doesn't work?
<slapin> oliv3r: totally ignored
<slapin> oliv3r: as subscription email
<Turl> slapin: maybe your mail server is spammy? :)
<Turl> try using another smtp if you can
<slapin> Turl: mail is reported delivered
<Turl> slapin: I think mnemoc can invite people from the admin too I think
<Turl> s/I think$//
<slapin> Turl: and I receive another google group eithout problems, and mail otherwise works including google's
<oliv3r> i have had the same happen to me on a different google groups list. I could sent it anything I wanted, but it would not post anything to the list. My mail server reported all was fine. I received other peoples mails just fine. linux-sunxi worked just fine in the meantime
<oliv3r> so i re-subscribed and now it works again :D
<slapin> oliv3r: I will wait for a few days. I'm paying google's customer and will try to find where is some kind of support, if that will not fix it then
<ssvb> rz2k: let's see if you can make a g2d driver faster than me :)
<rz2k> ssvb: I'm copypasting https://www.codeaurora.org/gitweb/quic/xwin/?p=xf86-video-msm.git;a=blob;f=src/msm-exa-c2d.c;hb=refs/heads/chromium by the way :p
<focus_well> Hi, releasing first SoM board - 2 x 100 pin ARMs on SO-DIMM - its just a stepping stone to A10 SoM board later http://www.gplsquared.com/SoM1/SoM1.html
<focus_well> Its in KiCAD - play with it to heart's content - feedback appreciated
<ssvb> rz2k: I'm trying a totally different approach, but don't want to announce anything before I'm satisfied with how it works
<mnemoc> slapin: want me to remove your @gmail from the group?
<focus_well> The idea is KiCAD is a game changer and we can all learn KiCAD and communicate PCBs through exchanging GPL'd design files
<slapin> mnemoc: don't
<rz2k> ssvb: just looking at their c2d it seems like they did it somewhat right, so.. any ETA/status atleast?
<mnemoc> slapin: ok
<rz2k> ssvb: I'm trying to do it in xf86-video-mali, by the way.
<rz2k> so we have 3d and 2d at same time.
popolon has quit [Ping timeout: 240 seconds]
popolon has joined #arm-netbook
orly_owl has joined #arm-netbook
<ssvb> rz2k: hopefully before Christmas
<ssvb> g2d kernel driver is quite buggy and poorly designed
orly_owl1 has joined #arm-netbook
orly_owl1 has quit [Client Quit]
simosx has joined #arm-netbook
simosx has quit [Changing host]
simosx has joined #arm-netbook
<Mehhh> mnemoc, which line from the makefile? This right? $(Q)sunxi-tools/fex2bin sunxi-boards/sys_config/$(SOC)/$(BOARD).fex > $(BUILD_PATH)/$(BOARD).bin
<slapin> Mehhh: it should be something with make in it.
<slapin> Mehhh: like make board -> make board_config\n\tmake all or something
<rz2k> ssvb: cool, I will wait, then. bugs in allwinner code is standard thing.
<mnemoc> Mehhh: without context or any hint of what you consider invalid there. that like is ok
<mnemoc> s/like/line/
<ibot> mnemoc meant: Mehhh: without context or any hint of what you consider invalid there. that line is ok
<Mehhh> oh sorry i'm exhausted
<Mehhh> i need to output the correct config
<Mehhh> for the mk802 which apparently is missing or the args are contextually inaccurate
<mnemoc> Mehhh: rest a bit ;-) and then tell what is the *real* problem you are trying to solve
<slapin> Mehhh: go to sleep then
<Mehhh> The real problem is i need to get this done :)
<mnemoc> Mehhh: and?
<slapin> Mehhh: it seems you nees to start from beginning
<mnemoc> Mehhh: blaiming random lines in a Makefile you don't understand won't take you anywhere near solving anything
<Mehhh> I already did this fine manually
<Mehhh> if it was up to me i'd redo this all in python, but i have days not months
<Mehhh> to use
<Mehhh> I'm not here for a history lesson i'm here to get this ironed out so that anyone can get their mk802 done
<slapin> Mehhh: what is what you want to get, what expected output and what goes wrong, what git repo are you playing with and what file to look at? I've got about 10 mins to look
<Mehhh> I need the uboot to build from BSP
<Mehhh> but it's passing the wrong args
<slapin> and?
<slapin> Mehhh: what to clone and which file to look at?
<mnemoc> Mehhh: I've told you 3 times already that you need to add mk802 board support on u-boot
<Mehhh> and i think i had to add to boards
<Mehhh> earlier
<slapin> Mehhh: are we speak about leechie stuff?
<slapin> why don't use hno/uboot-allwinner.git?
<Mehhh> ...
<Mehhh> Here's exactly what I did to waste more time
<mnemoc> slapin: that's the development tree
<Mehhh> I first git the entire project
<Mehhh> did it manually
<Mehhh> fine
<slapin> Mehhh: it is always important to show all the context.
<Mehhh> Then, knowing it was all fine in my own hands I then went to the BSP
<Mehhh> and jesus christ overload holes, which i am more than willing to debug and fix
<mnemoc> but not willing to listen
<Mehhh> I did what you said
<Mehhh> exactly what you said
<mnemoc> and what's the problem?
<Mehhh> I already told you
<mnemoc> no, you pasted the line that generates script.bin
<Mehhh> i already added to boards
<mnemoc> that doesn't count as telling what's failing on you
<Mehhh> and for some reason now it's trying to arch armv5
<Mehhh> rofl
<slapin> Mehhh: could you get full build output of what is wrong somewhere on paste?
<slapin> Mehhh: learn to make proper reports full of context
<Mehhh> It's an entire page
<Mehhh> hold
<mnemoc> .oO(what's a page in this context?)o
<Mehhh> of errors
<mnemoc> what's the lenght of what you call a page?
<slapin> Mehhh: it is very hard to help without even knowing what you're working on
<Mehhh> 50 terminal lines ~
<slapin> Mehhh: paste it on fpaste.org
<Mehhh> slapin, all you need to know is i am automating BSP because it is fail
<Mehhh> I've already fixed a few things already
<Mehhh> I am here for one device
<mnemoc> Mehhh: you still haven't said what is failing. just whining
<Mehhh> the MK802 / MK802 1GB
<rm> <mnemoc> what's the lenght of what you call a page? <- a page is typically 4K on ARM
<rm> :)
<mnemoc> rm: :)
<Mehhh> I don't know why you're more concerned with fighting
<Mehhh> instead of fixing
<Mehhh> i'm in git atm
<Mehhh> so unless you have a direct line please
<Mehhh> leave me be
<mnemoc> Mehhh: we want to help you. but you don't give proper information about the problem
<Mehhh> slow and steady is better than nothing and not ready
<slapin> Mehhh: I don't use BSP at all, but I could help if you show what a problem is a proper way - terminal outputs, files, and repository links.
<Mehhh> By that point i'd have it done myself
<Mehhh> if you can't see the obviously broken parts i just don't even.
<Mehhh> It's even set to the wrong arch in some places
<mnemoc> the BSP assumes there is a $board.fex in sunxi-boards for a valid board $board in u-boot-sunxi
<Mehhh> I did that
<Mehhh> did it did it did it
<rm> I wonder wtf are you even doing :) the MK802 is a solved problem, it just works with ready-made stuff, bootloaders and etc; it's not like you are on the cutting edge of anything and need to patch stuff manually and then rebuild
<mnemoc> show your diff, and the complete output
<orly_owl> which is the best mk802/latest revision? with power protection? mk802 II?
<Mehhh> rm this isn't for me
<rm> wtf is even a BSP
<Mehhh> or the public
<mnemoc> rm: a bunch of scripts to help lazy people. sunxi-bsp repo
<Mehhh> lazy? i already made my own image manually
<oliv3r> BSP = Board Support Package :p
<Mehhh> and it is crude to do that each time if you have to pump them out
<rm> and it presumably works
<slapin> BSP is marketspeak gibberish usually in Linux/android context = crappy boot loader/nadly ported kernel/crappy rootfs full of obsolete stuff
<slapin> s/andly/badly/
<Mehhh> i'm doing none of that, heh
<slapin> s/nadly/badly/
<ibot> slapin meant: BSP is marketspeak gibberish usually in Linux/android context = crappy boot loader/badly ported kernel/crappy rootfs full of obsolete stuff
<Mehhh> Ok, why can't i edit bsp files on git
<Mehhh> ah ah
<Mehhh> amazing you hotlinked
<mnemoc> what?
<mnemoc> can you please make a paste of the error you get and the changes you have done?
<mnemoc> rm: the bsp products "hardware packages" for each board, just like the nightlys
<mnemoc> s/products/produces/
<ibot> mnemoc meant: rm: the bsp produces "hardware packages" for each board, just like the nightlys
<slapin> orly_owl: dunno about these, I've got devices labelled as mk802ii which were made with newer board (PMU) and old one (NO PMU), so I wonder
<orly_owl> oh ok
<orly_owl> i guess buying from tom is safest
<Mehhh> refresh
<slapin> orly_owl: I buy cheap devices for $38 and hope for best. lucky at 50% cases.
<Mehhh> added MK802 and MK802GB (GB for the 1GB version)
<Mehhh> change it if u want but mk8021GB looks ugly
<mnemoc> how and where
<orly_owl> Mehhh: refresh?
<orly_owl> slapin: i hope for more then 50%
<Mehhh> The next one i buy will be the A20 stick if it comes out
<slapin> Mehhh: this name is very ugly, and there is different 1Gb versions with different hw settings
<orly_owl> i dont care how it looks
<mnemoc> Mehhh: so this is what you did? https://github.com/linux-sunxi/u-boot-sunxi/pull/21.diff ?
<mnemoc> Mehhh: nothing else?
<Mehhh> no, i have to edit the other file
<Mehhh> wow
<Mehhh> you know what
<slapin> Mehhh: use 802ii for device with PMU
<Mehhh> this is where i just sit back
<Mehhh> because you're more interested in feeling your ego up and down
<mnemoc> Mehhh: boards.cfg is just an index. you need proper board support in code
cheng has quit [Ping timeout: 255 seconds]
<Mehhh> guess what is being done
<Mehhh> right now
<popolon> Mehhh, do you know the max memory supported by A20/A31 ?
<Mehhh> no idea, I saw one device with 2GB but who knows
* slapin never seen anything with A20/A31
<mnemoc> Mehhh: I was trying to help you. but well... not my problem. back to work
<Mehhh> also slapin, if they want to add say the MK802+ which i believe you are speaking of, they can
<Mehhh> A comparo wiki entry is prob needed at some point
<slapin> Mehhh is not very friendly it seems
<Mehhh> honestly, i may just leach a bit from the slickdeals mk roundup
<Mehhh> it's fairly well written
tinti has quit [Ping timeout: 264 seconds]
<mnemoc> slapin: sunxi-bsp changed to use make ${board}_config; make all instead of make ${board}
<slapin> mnemoc: thanks a lot!
<slapin> mnemoc: can you provide full URL for the repo? (I can't access my browser at the moment)
<mnemoc> git clone git@github.com:linux-sunxi/sunxi-bsp.git
<Mehhh> Has anyone used the linaro tc and LTO flag on this kernel?
<rm> this one looks okay
cheng has joined #arm-netbook
<rm> ubercheap + good feedback + explicitly specifies "II" and the differences from the old one
<orly_owl> II is the latest? or MK802+?
<orly_owl> which is best
<orly_owl> hmm II seems best
<mnemoc> orly_owl: + has pmu. the ii also has more ports
<orly_owl> aha ok
<rm> crazy
Lantizia has joined #arm-netbook
<orly_owl> cheep
<Lantizia> Hey I think I have a netbook based on something called s3c2450
<Mehhh> Doesn't the 512 not have two holes?
tinti has joined #arm-netbook
<Lantizia> according to the chat log for this channel in 2010 - there was no support (says bjdooks)... http://ibot.rikers.org/%23arm-netbook/20100419.html.gz
Maqs has quit [Ping timeout: 240 seconds]
<Lantizia> any progress been made since then?
aholler_ has joined #arm-netbook
<mnemoc> Lantizia: samsung arm9. a classic from ancient times
<mnemoc> not sure if anyone cares about s3c24xx anymore
Maqs has joined #arm-netbook
<Lantizia> mnemoc, well I can't seem to find any instructions to load anything else on it
<Lantizia> it's a Datawind Ubisurfer, it already has Linux on it (came with it) but a crippled version of it
<Lantizia> the later versions of this model came with WinCE instead
<mnemoc> kernel and u-boot support is all mainlined afaik
<rm> afaik Rikomagic is "the official" vendor
<rm> rest are clones/OEMs
<orly_owl> ok
<rm> (which mostly do work perfectly, though)
<hp__> rm indeed a very good price
<Mehhh> that 28$ one also has a 15$ coupon
<orly_owl> heh
<hp__> rm the MK III has Rockchip RK3066 not to bad
<hp__> $57
<rm> the coupon is $15 off $300
<Mehhh> that's what you think
<rm> hp__, not too bad? that's horrible
<orly_owl> MK III isnt a10 soc then
<Mehhh> check out the glitched shipping too
<Mehhh> roflll
<hp__> rm its a dual core arm cpu..
<hp__> for $57
<rm> does it run Debian?
<rm> from an SD card?
<Mehhh> hp, slickdeals has a deal on the mk808 for like 34.99~
<hp__> indeed not bad
<rm> I see it for around $50 free shipping on aliexpress
<rm> but like I said RK3066 is uninteresting
popolon has quit [Ping timeout: 240 seconds]
popolon has joined #arm-netbook
<rm> and that's in pre-alpha right now
<rm> even if it works, GOOD LUCK getting any sort of 2D or 3D acceleration there
<rm> although no
<rm> this one does seem to use Mali 400MP
<Lantizia> mnemoc, does SMDK2450 ring any bells? apparently mine has that in it
<Lantizia> apparently I somehow need to work out if it is wm8505 or vt8500 ?!
<Mehhh> "Allwinner’s A20 is a dual core chip with 512MB of L2 cache"
<mnemoc> Lantizia: I honestly don't remember anything about it
<rm> Samsung S3C2450 Mobile Processor
<rm> this does not seem to have anything to do with wm8505 or vt8500
<rm> which are VIA/WonderMedia chips
<Lantizia> hmm well I seem to have what this guy has at the top of this page... http://ubuntuforums.org/printthread.php?t=1349626&pp=75&page=14
<Lantizia> i'm just looking to put something different on it - and instructions how
Manofthe1ea has joined #arm-netbook
Manofthe1ea has left #arm-netbook [#arm-netbook]
<Mehhh> Bro, you didn't allow the original mk802 title in. Jesus, I thought politics with stupid stuff was bad in the USA.
antonl has quit [Ping timeout: 240 seconds]
<Lantizia> rm, i'd be very grateful for any help you or anyone else might be able to give to someone new to doing this kind of thing
<Mehhh> I just have the need to get things done.
<rm> Lantizia, my advice would be to throw it away and buy that $100 netbook based on the A10
<Lantizia> I already was throwing it away lol - this is just an excuse to play, I don't really need a netbook
<rm> it's likely to be an order of magnitude faster, will have a better screen (both size and resolution)
<rm> Lantizia, what's the OS currently on it?
<rm> WinCE or Android?
<Lantizia> what gets me is it *already runs linux* and since it was shipped
<Lantizia> it's got some kind of distro on here, but although it has iceweasel it's definately not debian
<rm> WELL
<rm> if it already runs GNU/Linux
<rm> this kind of makes things an order of magnitude easier
<Lantizia> yeah but an extremely retarded copy of it!
<rm> don't throw it away then :D
<Lantizia> ok so what are my options?
<rm> try replacing it with debian, fail, brick it, and THEN throw away
<Lantizia> right *how* would I go about putting proper debian on it?
<rm> tbh I don't know how to begin explaining what you'd do
<rm> if you have zero experience with this
hg_5 has quit [Ping timeout: 264 seconds]
<Lantizia> zero experience with putting another OS on a Datawind Ubisurfer yeah... but not zero experience with linux, open source and generally doing weird shit with firmware
<rm> first of all, find out what's the storage in there
<Lantizia> been doing that stuff for years
<mnemoc> Mehhh: I just pushed support for two boards into u-boot-linux. look at the commits
<rm> does it just run a normal distro from its internal flash in an FS like Ext3, or are there some weird tricks with MTD / cramfs / unionfs involved
<rm> how much free space does that flash has
<Lantizia> lemme check...
<rm> then debootstrap a minimal "armel" architecture debian system and try replacing the one you already have, with ti
<rm> it*
<rm> EXCEPT for the kernel
<rm> most likely you will absolutely need to keep the original kernel
<Lantizia> ok its slowly booted up and i'm on the terminal,,, /dev/root seems to be 1gb
<Lantizia> nothing unusual mounted - no sign of cramfs or uniionfs
<rm> uname -a
<rm> ?
<Lantizia> oh but /dev/root is fs type yaffs2
<Lantizia> Linux ubisurfer2 2.6.21.5-cfs-v19 #423 Tue Nov 17 15:28:50 EST 2009 armv5tejl unknown
<Lantizia> mount basically says.... rootfs on / type rootfs (rw), /dev/root on / type yaffs2 (rw), proc on /proc type proc (rw), sysfs on /sys type sysfs (rw), devpts on /dev/pts type devpts (rw), tmpfs on /tmp type tmpfs (rw)
cheng has quit [Quit: Leaving]
tinti_ has joined #arm-netbook
<Lantizia> /etc/issue claims it's something called "PocketSurfer Linux"
tinti_ has quit [Max SendQ exceeded]
<Lantizia> rm: it's also got xfce's thunar on it, unfortunately no openssh-server :(
<rm> thing is, I don't know how would you recover if after some experiments it no longer boots
<Lantizia> thing is, I won't give a *uck :P
<rm> but that would be disappointing
<rm> since from what you describe it's basically 95% "there"
<rm> ("there" being running a modern lightweight Debian system)
tinti_ has joined #arm-netbook
tinti_ has quit [Remote host closed the connection]
<Lantizia> nah it's not debian
<Lantizia> all i said is it has iceweasel
<rm> so what
tinti_ has joined #arm-netbook
<rm> did you check if you have "dpkg" and "apt-get"?
tinti has quit [Remote host closed the connection]
<Lantizia> yup it has neither
tinti_ has quit [Remote host closed the connection]
tinti has joined #arm-netbook
<Lantizia> rm, what would your suggestion have been... simply replce all the files on the yaffs2 /dev/root filesystem with other files - somehow?
vinifm has joined #arm-netbook
tinti has quit [Max SendQ exceeded]
<Lantizia> i've got it on the network, but without an ssh server or ftp server or anything on it - i don't see how i'm gonna get such files over
<rm> not somehow but with cp and mv
<rm> :)
<rm> but wait
<rm> maybe a better approach would be to slowly convert it into debian
<rm> you may end up with a mess, but since it has iceweasel perhaps it doesn't deviate from debian much
<Lantizia> to be fair i just did a bit of digging in /var/lib and it looks like aptitude 'used' to be installed
<Lantizia> but isn't no longer
<rm> (i.e. based on it)
<rm> so as the first step, get some old armel package with "dpkg"
<Lantizia> i can ship files over to it using my sd card :)
<rm> unpack it and place all files where they have to be manually
<rm> to */bin/, */lib/ etc
<rm> then check if it runs
<rm> then try installing some simple package with it
<rm> etc
<Lantizia> it'd first help if I can determine approx which debian release it is based on
<Mehhh> Wheezy babyyyy
<rm> well does it have bash?
<Lantizia> maybe - just had to reboot it as it spazzed when i took out the sd while mounted :S
<Lantizia> the terminal drops you in to busybox though
<rm> bash --version; and http://packages.debian.org/search?keywords=bash squeeze had bash 4.1
<rm> I have a feeling it might be older than squeeze
<Lantizia> bash is 3.2.0
Turl has quit [Excess Flood]
<Lantizia> GNU bash, version 3.2.0(1)-release (arm-unknown-linux-gnu)
<Lantizia> Copyright (C) 2005 Free Software Foundation, Inc.
Turl has joined #arm-netbook
<rm> Debian Lenny -> GNU bash, version 3.2.39(1)-release (x86_64-pc-linux-gnu)
gimli has joined #arm-netbook
<rm> armel
<rm> or... was there no armel back then
<rm> yep there was
<rm> but which version is that
<rm> from which debian release
<Lantizia> not sure if it's the right version, i've no way of checking
<Lantizia> i'm just looking in the pool dir of the archives repo
<rm> because if it's too new, it just won't run due to incomaptible libs, first and foremost the glibc
<Lantizia> the earlier versioon is 1.13.26 and has no armel version
Guest43338 has joined #arm-netbook
<Lantizia> i'm just gonna extract the .deb onto the filesystenm and see what happens :)
<rm> like I said before, try getting yourself a debootstrap'ed modern Debian armel system
<rm> then place it onto your SD card
<rm> and you should be able to chroot to it
Guest43338 has quit [Max SendQ exceeded]
<rm> also can mount --bind some stuff from the original system
<Lantizia> don't want to rely on the sd really
Guest43338 has joined #arm-netbook
Guest43338 has quit [Max SendQ exceeded]
Guest43338 has joined #arm-netbook
Guest43338 has quit [Max SendQ exceeded]
<xxiao> google play store is dead?
<xxiao> nexus7 can't access it now
Guest43338 has joined #arm-netbook
Guest43338 has quit [Max SendQ exceeded]
rellla has quit [Ping timeout: 248 seconds]
Guest43338 has joined #arm-netbook
Guest43338 has quit [Max SendQ exceeded]
<vinifm> in i2c driver client, how do I tell the bus that will be used?
Quarx has quit []
DEAT_ has quit [Read error: Connection reset by peer]
DEAT_ has joined #arm-netbook
L84Supper has quit [Ping timeout: 240 seconds]
L84Supper has joined #arm-netbook
voronaam has joined #arm-netbook
<Turl> rz2k: ping
freakazoid0223 has quit [Quit: Leaving]
<rz2k> Turl: yeah?
Manofthe1ea has joined #arm-netbook
<Turl> rz2k: is the framebuffer vsync ioctl broken?
<techn_> Turl: you have to have scaler enabled
<techn_> .. in .fex
<Turl> it 'works' but not correctly (it's not vsyncing but painting a frame a second or such)
<Turl> techn_: so if I set it to 1 it will vsync correctly?
<techn_> Turl: I think that yes.. vsync is tied to scalers interrupt
<rz2k> yeah
<rz2k> fb ioctl does nothing, you need to set scaler layer to 1 to get vsync always on
<Turl> meanwhile /me hacked vsync support using usleep
<techn_> hmm.. that ioctl should return error if scaler is not enabled :/
<Turl> and it should return intantly
<Turl> not block execution for a crapton of time :)
<techn_> or even better.. it should use some other method
<rz2k> if you have a solution, that would be awesome
<techn_> Turl: how come you don't have scaler enabled.. you don't have hdmi output on that tablet?
<Turl> yeah it has
<Turl> never used it though, I don't have the cable
<Turl> print an informative warning too if you'd like
<Turl> techn_: enabling scaler didn't work from what I see
<techn_> :(
<Turl> do I need to enable it for all outputs?
<Turl> I only did on fb0
<techn_> you can try.. I havent studied that so far
<Turl> nope no luck
<techn_> Turl: and no code modifications?
<techn_> hmm.. could it be that it works only with hdmi :/
<Turl> I'll keep on doing fake vsync on an userspace thread then
gimli has quit [Quit: Verlassend]
gimli has joined #arm-netbook
<Turl> techn_: lol, with scaler on my recovery system shows 4 screens with all fail colors :P
<techn_> strange :)
<Lantizia> rm, still about? :D you'll never guess what I've done :P
<techn_> Turl: Oh.. I know.. it must be becouse recovery system has some other colorspace
<techn_> scaler layer supports only 32bpp and some yuv formats :/
<Lantizia> rm, I googled for any old debian armel rootfs and got this... http://tinyurl.com/cr6axkz - I've then literally untarred the lot straight on to / :D took a bit of effort but now seems stable again and i've got dpkg/apt-get too (it's installing openssh-server right now :P)
<Lantizia> trouble is I'm now a tad worried what'll happen if I reboot :D
mentis has joined #arm-netbook
<mentis> hi
<mentis> can somebody please help me with this http://pastebin.com/Hi916uAP
<mentis> im trying to boot my a710 from sdcard...
<mentis> with no luck
simosx has quit [Quit: Αποχώρησε]
<Manofthe1ea> mentis, what is /system?
<Manofthe1ea> I'm guessing when you installed, you mounted your sdcard on /system, and then installed the rootfs?
<mentis> not exactly
<Manofthe1ea> oh, this is android. Sorry, I don't know android.
<mentis> i've cloned android-jb from sunxi git
<Manofthe1ea> It found init, but all the paths are pointed at /system/bin/xyz
<Manofthe1ea> I would expect just /bin/xyz
<mentis> dont know how exactly to deploy it on sdcard
<Manofthe1ea> [ 18.250000] init: cannot find '/system/bin/sh', disabling 'console'
<mentis> thing is that this file exists
Mehhh has quit [Ping timeout: 245 seconds]
<Manofthe1ea> So, the problem is your rootfs, not your card.
<Manofthe1ea> where does that file exist?
<mentis> the same others
<Manofthe1ea> does it exist at /bin/sh on the memory card?
<mentis> no
<Manofthe1ea> also... you have a lot of partitions. boot, root, and 4 logical?
<mentis> exactly
<Manofthe1ea> not that that matters.
<mentis> i have uImage, evb.bin in root
<mentis> in root
<Manofthe1ea> So, when you mount the mmcblk0p2, you have a /system/bin/sh?
<mentis> i have everything that was in android-jb/out/root folder meaning all init.rc init.sunxi.rc stuff
<mentis> my uboot loads mmcblk0p2 as /
<mentis> i modified the init.rc to mount mmcblk0p5 as /system
<Manofthe1ea> hmm... your log doesn't show that happening before init tries to access stuff.
<mentis> true
<mentis> i donkt know how to make it more verbouse
<mentis> verbose*
<Manofthe1ea> more verbose? I think that indicates that it is NOT mounting p5
<mentis> this is also my guess
<Manofthe1ea> what's on your root partition, if not the base files to bring up the system?
<Manofthe1ea> have you tried copying /system back to p2?
<mentis> root = mmcblk0p2
<mentis> create default.prop init init.goldfish.rc init.rc init.sun4i.usb.rc init.usb.rc sbin system ueventd.rc
<mentis> data dev init.cm.rc initlogo.rle init.sun4i.rc init.trace.rc proc sys ueventd.goldfish.rc ueventd.sun4i.rc
<mentis> no I didnt
<Manofthe1ea> again, I don't know Android, but this sounds like the same issue as a linux system would have with a seperate /usr partition.
<Manofthe1ea> aka, it fails.
<Manofthe1ea> where is the mount command? /system?
<mentis> in init.sun4i.rc
<Manofthe1ea> no, I mean, the actual command, not just the call.
<mentis> mount ext4 /dev/mmcblk0p5 /system wait ro
<Turl> /dev/block/*
<mentis> was mount ext4 /dev/block/nandd /system wait ro, i changed it to boot from sd
<mentis> Turl:?
<mentis> sure?
<mentis> earlier today i managed to run lubuntu on this tablet
<Manofthe1ea> I mean, /system/bin/mount?
<Manofthe1ea> is on p5, which isn't mounted, and thus can't be mounted.
<mentis> i have mmc* in /dev
<mentis> but i'll check it
<mentis> oh!
<mentis> oh!!!!!!!!!!!
<mentis> true
<mentis> omg !
<mentis> LOL
<mentis> that would explain it
<mentis> there is no mount command avaliable!
<mentis> ok i will revormat the sd to 2 partition layout and give it a try ...
<Manofthe1ea> Needs more Tiger Blood.
traeak has joined #arm-netbook
Manofthe1ea has quit [Quit: ask ManoftheSea]
<traeak> hmm i bail for a copuleo f weeks and then there's new allwinner soc's
<mnemoc> i'm hoping for an A40
<mnemoc> same as the A20 but quad A
<mnemoc> A7
<traeak> wonder how a7 compares with a15 core
<traeak> samsung chromebook looks appealing
<traeak> and the move to powervr is...interesting (a41)
<mnemoc> an a7 should be like the half of a a15
<jelly-home> whaat
<mnemoc> http://en.wikipedia.org/wiki/List_of_ARM_microprocessor_cores says the A7 is 1.9 DMIPS/MHz and A15 between 3.5 and 4
<mnemoc> i don't find the A31 interesting at all currently
<mnemoc> unless tom manages to release some core code
<mnemoc> otoh the A20 feels better
<traeak> sry a31 is the quad powervr
<techn_> hmm.. what is the purpose of that interrupt controlelr patch :/
<traeak> of course with no acceleration for powervr or mali i guess it doesn't matter either way
<mripard> mnemoc: hi
<mnemoc> mripard: hi
<mripard> I have a pinctrl driver working finally
<mripard> a few cleanups and I'll post it for reviews
<mnemoc> awesome
<mentis> no luck :(
<techn_> also.. anyone know why sunxi uart codes are trying to use U7_16550A driver.. shouldn't it use 16550A
<mripard> one question though, do you know what the value set in the 2 bits for the dlevel and the pullup registers corresponds to ?
<mentis> maybe uboot is wrong? 0_o or something in evb.bin, but i copied it from my tablet ...
<mripard> (and what value are possible?)
<mnemoc> mripard: levels are 0,1,2,3. pull is 0 for disable, 1 for up, and 2 for down
<mripard> ok
<mripard> and the levels correspond to what exactly? voltage? current?
<mnemoc> not sure
<mripard> ok
rz2k has quit []
<mripard> mnemoc: and do you want to be in cc for these patches ?
<mnemoc> sure :) thanks!
<mripard> what is your mail address ? :)
<mnemoc> amery geeks cl
<mripard> k
vinifm has quit [Quit: Saindo]
<mentis> how is the amount of ram defined in allwinner a10?
<mnemoc> what do you mean?
<Turl> mentis: uboot provides the information
<mentis> Turl, also for sth like AW16XX I2C adapter base_Addr = 0xe88b8000 ?
<mnemoc> no. that's fixed
<mentis> mnemoc, no
<Turl> mentis: re. earlier, block devices are on /dev/block/mmcblk*, not /dev/mmcblk*
<mnemoc> mentis: the register addresses are fixed by soc. fortunatelly they repeat in f20, a10 and a13. hopefully also in the other allwinner socs
<mentis> Turl, doesnt matter now i reformated my sdcard to 2 partitions only, i have 1st vfast and 2nd ext4, im passing root=mmcblk0p2 and it boots
<mentis> mnemoc, i have to go through u-boot conf then :/
<mentis> tank you :)
<mnemoc> we have a list in the wiki
<mnemoc> might save you some time
<mnemoc> http://linux-sunxi.org/I2C has some extra info regarding the I2C controller in particular
<mentis> mnemoc, is there a way to see how its set up in working - default kernel ?
<mentis> than you
<mentis> thank you*
<mnemoc> my pleasure
<ManoftheSea> Well, I found the config for uboot in hno's tree has changes for the mk802ii, so I'm going to try with that one.
<techn_> Turl: /bin/bash: prebuilts/gcc/linux-x86/arm/arm-linux-androideabi-4.6/bin/arm-linux-androideabi-gcc: cannot execute binary file
<Turl> bad libc? just a long shot guess though
<mentis> what says 'file; ?
<Turl> where are you running it?
<techn_> sunxi-bsp and make android
<techn_> ubuntu something..
<mentis> techn_: try what is the output of $file prebuilts/gcc/linux-x86/arm/arm-linux-androideabi-4.6/bin/arm-linux-androideabi-gcc ?
<techn_> Turl: mentis: yeah.. it wont work, even if called directly from bash
<mentis> techn_: what is the output of command 'file' ?
<mentis> type 'file' in console followed by path to that arm-linux-androideabi-gcc
<techn_> prebuilts/gcc/linux-x86/arm/arm-linux-androideabi-4.6/bin/arm-linux-androideabi-gcc-4.6.x-google: ELF 64-bit LSB executable, x86-64, version 1 (SYSV), dynamically linked (uses shared libs), for GNU/Linux 2.6.15, BuildID[sha1]=0xf8e11f40cf6b10fa26e6fea774d9d94058c53595, stripped
<mentis> ty
<mnemoc> techn_: maybe you have a 32b host?
<techn_> I have 32bit host
<mnemoc> *g*
<mentis> so u go the anserw
<mnemoc> techn_: you need 64b to compile android
<techn_> uaah
<mnemoc> techn_: i can give you access to one if you like
<techn_> mnemoc: no need.. I'll download it
<mnemoc> ok
<ManoftheSea> gg mnemoc
<mnemoc> ManoftheSea: what "gg" stands for?
<ManoftheSea> here, good guess. Sometimes also a generic term of congratulations.
<mnemoc> ManoftheSea: :)
<techn_> that was easy
<Turl> mnemoc: I was talking with a friend yesterday, he works on embedded systems, and the GSoC topic came up
<Turl> maybe we should organize some projects for GSoC to get more stuff done
gimli has quit [Ping timeout: 240 seconds]
<mnemoc> don't think any of the "pending" stuff we have is worth of a GSoC project
<mnemoc> maybe REing cedarx
<Turl> ok. 4.2 running on A10 :)
<ManoftheSea> Jelly Belly? Yay!
<mnemoc> AOSP or CM?
<ManoftheSea> will it run on my mk802ii? Am I the only person with that that's interested in it?
<Turl> ManoftheSea: jelly bean 4.2 yeah
* jelly-home hates Google and their highlighting shitty codenames
<ManoftheSea> when I say 4.2 Jelly bean, it makes jelly jelly(ous)
<Turl> mnemoc: CM, AOSP is unpolished by definition
<Turl> ManoftheSea: if you make a device tree for it, sure
<jelly-home> ManoftheSea: just annoyed... if you said JB that would be much appreciated
<ManoftheSea> Ok, jelly.
<mnemoc> Turl: I can't digest CM at all
<Turl> mnemoc: why?
<Turl> mnemoc: CM is "what AOSP should have been"
<mnemoc> + a disgusting default look and feel
<mnemoc> and an even
<ManoftheSea> unpolished, but it's buttered.
<mnemoc> more disgusting splash logo
<ManoftheSea> but it's blue, right?
<Turl> mnemoc: it looks and feels the same as AOSP
<Turl> mnemoc: and you can replace the splash logo if you don't like it (it's just a ring with a lead moving on it and the text 'cyanogenmod'
<mnemoc> i won't troll about. it's just a personal opinion. I wouldn't use android at all if it looks like CM by default
<Turl> then you wouldn't use AOSP
<Turl> CM9+ looks the same as AOSP, it doesn't have any themes or anything by default
<mnemoc> haven't tried recent CMs. last was on my nexus one
<mnemoc> first and last actually
<Turl> it has changed a lot since the CM7 days
<Turl> I'd suggest you give it a second chance, you'll like it
<mnemoc> do you have a mele image? ;-)
<Turl> no, but feel free to make a device tree :)
<mnemoc> is your common 4.2-ready?
<mnemoc> one device tree repo for each sunxi is absurd by any mean
<mnemoc> s/sunxi/sunxi device/
<ibot> mnemoc meant: one device tree repo for each sunxi device is absurd by any mean
<ManoftheSea> I thought the tree was dependent on the board.
<mnemoc> and the diff between sunxi boards is marginal
<Turl> mnemoc: but simple stuff like keyboard mappings changes
<mnemoc> by far not worthy of a repo
<Turl> you cannot make one size fits all
<ManoftheSea> oh, sorry, repo? No.
<Turl> we discussed this already several times
<ManoftheSea> I thought you meant a different tree. Like... src/device-tree/device_name.bin
<ManoftheSea> but you'd only need one repo to hold all the trees, right/
<mnemoc> ManoftheSea: Turl insists in one .git to clone for each device
<mnemoc> instead of a single device/allwinner/ with a dir per device
<Turl> we discussed the problems with overlapping repos too
* ManoftheSea believes turl has reasons, even if he doesn't understand the issue yet.
<mnemoc> Turl: so it has to be not-overlapping ;-) device/allwinner/
<mnemoc> Turl: without sub-repos
<Turl> yeah and device/allwinner/common overlays it :P
<Turl> unless you want to put it all on a huge bucket
<mnemoc> Turl: i do
<ManoftheSea> why not?
<mnemoc> Turl: common/ has almost everything. mkdir common; git mv * common/; git commit
<mnemoc> and rename to android_device_allwinner.git
<mnemoc> the uniqueness of device/allwinner/$device/ are minimal
<mnemoc> zero value on having one repo for each
<mnemoc> but we already discussed that several times and we won't agree
<ManoftheSea> is there a goal of mainlining those? Wouldn't they need to be in one repo to end up in THE linux repo?
<mnemoc> ManoftheSea: this is android device tree. not linux kernel related
<ManoftheSea> Also, the device-tree shouldn't change often... right? The board+revision is released
<mnemoc> ManoftheSea: to build android's userspace
<ManoftheSea> it's not the same device tree?
<mnemoc> no
<Turl> it's a set of makefiels and files
<Turl> s/fiels/files/
<ibot> Turl meant: it's a set of makefiles and files
<ManoftheSea> I thought the device tree was OS-agnostic, and described the stuff on the board
<mnemoc> ManoftheSea: same name, different things
<ManoftheSea> Who decided to overload that operator?
<ManoftheSea> Google? Jerks.
<mnemoc> ManoftheSea: it's a file system tree of devices. therefor device tree
<mnemoc> device/$vendor/$device/Makefile
<ManoftheSea> ok. Well, I don't do much with A-droid, so I retract my opinions.
mentis has quit [Quit: Konversation terminated!]
<mnemoc> Turl: having a 4.2-ready device/allwinner/common and a sample device/allwinner/foo using it I can fork and do the heretic things :)
<Turl> https://github.com/allwinner-dev-team/android_device_allwinner_common/commit/cfdba3b882ef7c6482105d11c6dd8cfc1362dbf7 <- what I used to do vsync but it doesn't work :| I hacked it up with usleep
<mnemoc> Turl: what branch is that on? jellybean seems 1M old...
<Turl> mr1-staging
* Turl needs to push other repos with mr1-staging code
<mnemoc> what mr1 stands for?
<RaYmAn> mr1 is a name google uses
<RaYmAn> mr1-staging is CM's integration branch for mr1
<mnemoc> i see
* RaYmAn shrugs at what it stands for - it's just what google use ;)
<mnemoc> :)
<mnemoc> smells to "master 1"
<Turl> main release maybe
<mnemoc> +1
<Turl> 4.1 was mr0 I think
<RaYmAn> they also had a HC mr1 (hc 3.1)
<RaYmAn> though, no source for that one
<mnemoc> their internal mess to avoid showing too much for their secret next release is probably similar to allwinner's
<mnemoc> open development is much simpler...
<mnemoc> they should learn a bit from the chromeos people
<RaYmAn> oh yes
<RaYmAn> and start not locking down their frigging devices
* RaYmAn curses at n7 being semi-locked
<RaYmAn> chromebook is fully open :/
<ManoftheSea> what about the nex4?
<RaYmAn> not sure about that one
<ManoftheSea> Can't get it. It's so out of stock
* mnemoc fears all that glass all over the phone
<ManoftheSea> but it's GORILLA GLASS 2 (tm)
<mnemoc> I drop my galaxy nexus at least 3 times per day
* ManoftheSea boggles
<mnemoc> the headphones cable likes to get traped on my chair
<ManoftheSea> When is someone going to create the skull-mount EOMA i/o-board?
<mnemoc> and then it jumps down the desk
* mnemoc gave up on eoma
<Turl> ManoftheSea: the back isn't as far as I heard
<ManoftheSea> the back?
<Turl> the N4 glass back
<Turl> ManoftheSea: you can skull-mount a mele
<mnemoc> ManoftheSea: it literally has glass all around
<ManoftheSea> but... gorilla!
<mnemoc> :)
<ManoftheSea> ook ook?
<Turl> the hw quality is rather bad from what I see from people talking about it
<ManoftheSea> "we are out of inventory"
<ManoftheSea> the 16 GB is 5-6 weeks.
<Turl> (internals-wise)
<mnemoc> ManoftheSea: for that price/specs of course it's out of inventory :)
<ManoftheSea> well, yeah.
<ManoftheSea> Who'd a thunk it?
<ManoftheSea> I want a J Butterfly for ~ $300. Also, a flying unicorn.
<jelly-home> with a rainbow mane?
<jelly-home> wait, no
<jelly-home> ManoftheSea: unicorn + wings = alicorn
<jelly-home> (in certain fandoms)
<ManoftheSea> yes, jelly, but I wasn't gonna identify myself like that again.
<ManoftheSea> Also, Rarity is a flying unicorn, not an alicorn.