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/
dyoung has quit [Ping timeout: 256 seconds]
dyoung has joined #arm-netbook
lerc has quit [Ping timeout: 255 seconds]
lerc_ has joined #arm-netbook
lerc_ has quit [Read error: Connection reset by peer]
Jef91 has quit [Quit: Instantbird 1.3 -- http://www.instantbird.com]
ssspiff has quit [Ping timeout: 272 seconds]
ssspiff has joined #arm-netbook
ssspiff has quit [Changing host]
ssspiff has joined #arm-netbook
<TestModule> How can i just clone the stage branch of linux-sunxi?
Alex1269 has joined #arm-netbook
stefanro1 has joined #arm-netbook
stefanro has quit [Ping timeout: 252 seconds]
soooga has joined #arm-netbook
Alex1269 has quit [Ping timeout: 245 seconds]
<libv> ssvb: any further hints besides boosting memory clock by 33%?
<libv> ssvb: i am really getting killed by lack of memory bandwidth here
soooga has quit [Quit: Leaving]
freakazoid0223 has quit [Quit: Leaving]
Jef91 has joined #arm-netbook
ganbold has quit [Remote host closed the connection]
ganbold has joined #arm-netbook
aholler_ has joined #arm-netbook
aholler has quit [Ping timeout: 248 seconds]
ZaEarl has quit [Quit: Ex-Chat]
Quarx has joined #arm-netbook
toxicpsion has quit [Ping timeout: 272 seconds]
aholler_ is now known as aholler
toxicpsion has joined #arm-netbook
dyoung is now known as dyoung-away
<ssvb> libv: using 16bpp color depth may be a reasonable trade-off, or switching to a faster hardware altogether (something like ODROID-U2 looks really nice)
<libv> ssvb: nope, it has to be this hw, and it has to be 32bpp :)
<libv> ssvb: how did you achieve the higher memory clock, the .fex?
<libv> how did overclocking the mali go?
<ssvb> libv: memory clock can be changed in u-boot, fex does not seem to have any effect
<ssvb> I have not tried overclock mali for real, but just tried changing mali_clkdiv in .fex between 3 and 4
<ssvb> techn_ mentioned earlier that mali was already unstable at 400MHz when somebody got this clock frequency by mistake
ganbold has quit [Remote host closed the connection]
gimli has joined #arm-netbook
ganbold has joined #arm-netbook
servili007 has quit [Read error: Connection reset by peer]
<libv> ok, will see what i can find over the next few weeks
Alex1269 has joined #arm-netbook
<TestModule> Has anyone tried 1.2 @ 1.4v?
<TestModule> CPU
Alex1269 has quit [Quit: Page closed]
hipboi has joined #arm-netbook
ssspiff has quit [Ping timeout: 272 seconds]
rz2k has joined #arm-netbook
rellla has joined #arm-netbook
eebrah has joined #arm-netbook
Jef91 is now known as Jef91|AFK
ZaEarl_ has quit [Ping timeout: 248 seconds]
pcat has joined #arm-netbook
ssspiff has joined #arm-netbook
lerc has joined #arm-netbook
hansg has joined #arm-netbook
wingrime has joined #arm-netbook
<libv> anyone read this one?
<oliv3r> sounds interesting
<oliv3r> I still applaud the effort of the R-Pi. I just cry at the design choices concerning the CPU
<libv> well, mostly the fact that they like to claim that this is an open platform while it clearly isn't
<specing> it is open to them, lol
* mnemoc cries at via choices with their APC boards
<libv> their whole claim of "our GLES drivers are now fully and completely open, we are the first industry and like to challenge the others to do the same"
<libv> was quite an afront
<libv> but more on that at fosdem :)
<mnemoc> :)
<libv> cubieboard is mentioned in that article btw
<libv> heh, i get quoted in there too.
<libv> no reference though
<libv> "I'm tempted to do a Kickstarter and say 'I'm going to produce an open-source GPU'. I want $2m from all the people who've criticised me,"
<libv> he's so getting my middle finger at FOSDEM
<libv> for all out lying, and then later on crying when he gets corrected.
<Maqs> a10 multimedia performance is far worse? O.o
Ershov has joined #arm-netbook
Ershov has left #arm-netbook [#arm-netbook]
KoH has joined #arm-netbook
rellla has quit [Remote host closed the connection]
tinti_ has quit [Quit: Leaving]
pcat has quit [Ping timeout: 246 seconds]
rellla has joined #arm-netbook
<TestModule> libv, couldn't you just become an ARM partner and get the source and release it?
<libv> TestModule: what world do you live in?
<mnemoc> *g*
<TestModule> 2m wouldn't cover any licensing issues?
<libv> 2m is the wage of about 50 people for one year, i think arm needs a bit more than that (in pure monetary terms) before its lawyers allow the freeing of this code
<libv> TestModule: once again, what world do you live in, and have you thought your statements through?
<mnemoc> TestModule: assuming you are large enough to be accepted as partner, you have to sign an nda to get sources, which 1) won't allow you to release anything except binaries, 2) will make it impositble for you to work on anything related ever after
<TestModule> Yes, and as an investor of ARM Holdings, I know their profit vs earnings is horrid.
<libv> and the relationship of this sentence to what i or mnemoc just said is?
<TestModule> 2m flat out may not seem so small if requested correctly.
<TestModule> Even if they only did it as a PR stunt.
<mnemoc> TestModule: they only care about $$
<mnemoc> that's their one an only interest. money.
<mnemoc> if they would have moral or any interest in improving their public image they would talk their partners into gpl compliance
<mnemoc> but they don't
<TestModule> GPL =/= ARM vs Intel hobbyist market
mac- has joined #arm-netbook
<mac-> hey
<mnemoc> TestModule: they don't need to open anything to kill intel
<mnemoc> TestModule: the % of buyers who care is irrelevant
<mac-> do you use any variant of MK802 device based on AllWinner A10 with Linux ?
<TestModule> I know this, but that doesn't mean the board does as well, or trusts.
<TestModule> I'd personally phrase it as a destination that must be met, eventually.
<TestModule> With Steam for instance going to Ubuntu, and Blizzard heading that direction.
<TestModule> You can't tell me ARM isn't going to support its own GPUs on Linux other than a binary blob, forever.
<mnemoc> it's very possible ARM doesn't own full rights over the mali code
<mnemoc> libv: do they?
<TestModule> So then, we find out who does.
<oliv3r> mnemoc: via uses the WM because it's their own soc? but yes, it's cryable :)
<libv> mnemoc: i would think they do
<mnemoc> oliv3r: not only that. they chose a crappy soc on both models, and an absurd amount of ram
<oliv3r> libv: yeah i saw your quote and cubieboard. He's trying to sweet talk it :) oh well was to be expected. I also find it interesting that he says the Pi performs just as good as the A10, when it clearly doesn't. It kinda sucks in performance.
<libv> mnemoc: mali was done by demosceners, who, i expect, did not take much from others
soooga has joined #arm-netbook
<libv> mnemoc: those guys developed the m400 class hardware in several iterations from the ground up
<mnemoc> libv: nice
<libv> i really doubt there is ip in there from others, but maybe they did violate some patents unknowingly
<mnemoc> f* patents
<mnemoc> :(
<libv> oh well, we are changing things, for real
<libv> arm will sit up and listen after fosdem :)
<mnemoc> the same fear of been sued is probably blocking many other ip owners from opening up
<andoma> \o/
<mnemoc> libv: :)
<TestModule> Technically, if a partner was about to go out of business, you could buy the licenses (by buying the company) and peek at the source, no? Just as others code around the GPL (by remaking GPLed libs/etc), you too could code around their own simply by changing it.
<libv> TestModule: yes, peek at the source and know that you will never do free software on this hw again.
<libv> eart to TestModule.
<libv> earth even
<TestModule> I'm spit-balling here.
<TestModule> What specifically do you need that isn't already provided from them as open-source?
<mnemoc> TestModule: the user space library to use the hardware
eebrah has quit [Ping timeout: 255 seconds]
<TestModule> Heh, seems my partner bribing idea may not be so far off: http://forums.arm.com/index.php?/topic/16133-openvg-lib-for-mali400/
merbzt has joined #arm-netbook
pcat has joined #arm-netbook
popolon has joined #arm-netbook
<oliv3r> libv: you really want me to go to fosdem even more then I allready do!
<oliv3r> i'll just have to wait for the video
<oliv3r> make sure the video turns out good :)
<libv> i believe that the fosdem organizers are taking care of that
<oliv3r> they tend to yeah, but sometimes they turn out to be crappy :(
<libv> it won't be michael larabel stuck in front of the devroom with his portable camera this time
merbzt has quit [Read error: Operation timed out]
tinti has joined #arm-netbook
<hansg> mac-, yes a lot of people use the original mk802 or the mk802ii with Linux, works great! (as long as you don't want 3d / video decoding)
<TestModule> anyone on the 3.4 stage?
<hansg> Better double check the mk802 you get has an A10 in there though, some models also come with other SOC-s ie the mk802iii has an rk3066 for which Linux support is not so good (yet)
<hansg> TestModule, yes I run 3.4 staging branch all the time, why ?
<TestModule> did you use the linaro toolchain?
<hansg> TestModule, no I'm using Fedora-18 both as development platform, as well as userspace on the a1x device itself. Fedora-18 comes with a packaged cross-compiler called gcc-arm-linux-gnu which I use:
<hansg> [hans@shalem virt-viewer]$ rpm -q gcc-arm-linux-gnu
<hansg> gcc-arm-linux-gnu-4.7.2-2.aa.20121114svn.fc18.x86_64
<hansg> Note this is all gnu-eabi hfp (of course)
<TestModule> hansg , by any chance did you -lfto ?
<TestModule> errr <->
<TestModule> -flto
<hansg> TestModule, I'm not familiar with that option, I'm only cross building the kernel and u-boot using the default CFLAGS from their buildsystems. The userspace gets build natively on the Fedora buildsys
<TestModule> cool cool
* hansg just steals the official panda Fedora images and then drops in the A10 kernel
<TestModule> So you used 3.4 for FS compatibility? BTW clever.
<hansg> I used 3.4, because Fedora userspace tends to depend on newer kernel ABI quite heavily, so I just choose the latest :)
<hansg> Which seems to work well, I had to do a small fixup wrt selinux, but now I've full F-18 including selinux in enforcing mode running nicely on the A10
<TestModule> are you using the original or the ii?
<hansg> TestModule, both, as well as a mele_a1000g, a mini-x-1gb, and a gooseberry-a721
<hansg> Since I'm making images intended for consumption by a wide audience I like to have a diverse set of hw to test with :)
<TestModule> Oh wow. Does your mk802 only use 512 as well since no 1gb fex?
<hansg> Yes my original has only 512MB, I've a 1GB version on its way (through the slow but cheap chinese post), once I've that I'll do a u-boot + fex for the original with 1GB.
<hansg> TestModule, if you've a 1GB version, and you can run a10meminfo on there, then we (me or mnemoc for example) can make a test fex and u-boot for you ...
<hansg> TestModule, which android does your mk802 come with ?
<TestModule> Nice. I think I saw a guy trying to do a bunch of commits for the generic fex, but he has like four posted and it's not organized.
<TestModule> ICS
<TestModule> i have a linux image already loaded if need be on a microsd
<hansg> Normally we also want the script.bin from the nanda partition, but with ICS images that seems to be a copy from another image, and, well, wrong.
<TestModule> Hrmmm
<hansg> So I plan to take the script.bin from the 512M version and modify that. Note my 512M has a clearly wrong script.bin too, but the sunxi-boards git repo has a .fex from an original mk802 with android 2.x, and that is correct. So my plan is to just drop your meminfo in there and use that
<hansg> So to get the meminfo, do the following:
<hansg> 1) Download a10-meminfo-static from: https://github.com/maxnet/a10-meminfo
<hansg> 2) Put it on a fat formatted usb-stick or micro-sd
<hansg> 3) Start the mk802 without an sdcard, so that you boot the original android
<mnemoc> we need a submiting fex files page on the wiki :p
<hansg> 4) install a terminal from the app store
<hansg> 5) start terminal
<hansg> 6) su -
<hansg> 7) insert sd-card
<TestModule> busybox needed?
<hansg> 8) do "mount" to see where the sd-card (or usb-stick) got mounted
<hansg> 9) do "cat <path-to>/a10-meminfo-static > /a10-meminfo-static"
<mnemoc> /a10-meminfo-static ?
<hansg> 10) do "chmod 755 /a10-meminfo-static"
<hansg> mnemoc, yes, sdcard is mounted noexec
<mnemoc> ok
Ershov has joined #arm-netbook
<TestModule> cool, doing now
<hansg> Could be some other path to but / works for me sofar
<hansg> 11) /a10-meminfo-static
<hansg> 12) Write down output
<TestModule> i'm using the stock FW, hope that's ok... 4.0.1 i think?
<oliv3r> i used /data/cache
<hansg> 13) Send email to list with output
<oliv3r> writeable by the user (/ may not) and is exec mounted
<hansg> oliv3r, notice the "su -" at step 6, and we need root-rights for a10-meminfo itself anyways
<hansg> And for the script.bin:
<hansg> 14) Do "mount"
<hansg> 15) See where nanda is mounted
<hansg> 16) if not mounted do: "mkdir /mnt/nanda" and "mount /dev/nanda /mnt/nanda"
<hansg> 17) do "cat <path-to-nanda>/script.bin > <path-to-sd-or-usb>/script.bin"
<hansg> 18) do "umount <path-to-sd-or-usb>"
<hansg> 19) put script.bin somewhere where you can hyperlink it and send a link to the list
<hansg> TestModule, as said I expect your script.bin too be foobar, but still would be good to collect it. Also no busybox needed in my experience
<hansg> mnemoc, I agree we need a wiki, how about copy pasting my above step-by-step as a start ?
<TestModule> awesome, getting all that together now
<mnemoc> hansg: yes, it's a good start
<hansg> TestModule, also the stick firmware is exactly what we want! With non stock the original memory settings may have been replaced ...
<hansg> s/stick/stock/
<ibot> hansg meant: TestModule, also the stock firmware is exactly what we want! With non stock the original memory settings may have been replaced ...
<hansg> back to work for me now ...
<oliv3r> hansg: what if / is ro mounted :p su - wont' help :)
SouL_ has joined #arm-netbook
<hansg> TestModule, btw, does your original have an A10 or an A10s in there? Recently original mk802-s (from the outside atleast) have popped up with an A10s soc in there... I've one of those ont the way too, I expect them to be using an axp20x pmu unlike the A10 ones which have omitted that...
<hansg> And it needs a different kernel build ...
<hansg> oliv3r, true (although I've not encountered that yet), as said I think someone should copy my step by step from above to the wiki. Once it is there, well it is a wiki. So if you think there is a better place to copy it ...
<hansg> Sorry, but -ENOTIME to do the wiki thing, really back to work now.
<TestModule> A10
<hansg> good
<TestModule> ok so should i use the /mnt/ target or the /dev/ target
<hansg> The /mnt
<TestModule> heh, can't chmod it "cd /mnt/extsd/","chmod 755 a10-meminfo-static","Operation not permitted"
<mnemoc> chmod it after copying out of the card
<hansg> TestModule, right that is why you first need to copy it, but there usually is no cp, hence the use of cat.
<hansg> IOW you're forgetting step 9
<TestModule> awesome, just su past everything
<TestModule> ok someone shoot me an email or something the dram output just scrolled
<TestModule> errrr, nm i can SS it
<mnemoc> mripard: fyi someone just reports on github that his A31 .fex uses port banks up to the M
gimli has quit [Ping timeout: 272 seconds]
<TestModule> hansg ibot http://pastebin.com/ZhLqpz9D
<mnemoc> TestModule: ibot is a bot, not a person
<TestModule> ah
bsdfox\ has quit [Ping timeout: 260 seconds]
<mnemoc> TestModule: use those to edit the [dram_para] section of the decomplied script.bin
<TestModule> Ok so, i see nandd, nande,nandh,nandi, no nanda
<mnemoc> uhm
<mnemoc> the first vfat
<mnemoc> whatever letter it has
<TestModule> kk
<hansg> TestModule, it is likely not mounted then, try doing step 16
<hansg> and then "ls /mnt/nanda"
* hansg think it is almost never mounted under android, maybe the wiki version of the instructions should omit step 14+15, and make 16 unconditional
<mnemoc> ack
<hansg> Also on the wiki probably split in separate instructions for meminfo + script.bin, copy-pasting the first few steps
<mnemoc> hansg: btw, how do you feel with the dynamic disp stuff in stage yet?
<mnemoc> hansg: the good thing with wikis is that they can be freely improved over time :)
<hansg> mnemoc, with the "fix pixelclock harder" patch it seems ok
<TestModule> -r -w ?
<TestModule> prob read
<hansg> TestModule, ???
<TestModule> not seeing nanda
<mnemoc> ls /dev/block/nand* ?
<TestModule> d / e / h / i
<mnemoc> allwinner is becoming very good at annoying :<
<TestModule> Yes :<. It's making me pass flags or incorrect usage for mount
<specing> maybe we should just periodically send someone to china to break into their HQ and take the sources... easier
<mnemoc> specing: they probably changed the nand driver to hide certain partition types
<hansg> TestModule, ah yes you need to specify -t vfat to mount, it does not auto-detect fs for some reason
<mnemoc> if it doesn't auto-detect, it's probably not vfat
<TestModule> says device or resource busy for the first one
<TestModule> you know, i can hop on ubuntu and just rip the files
<mnemoc> can you?
<hansg> mnemoc, no, it is the android re-written mount utility (so as to be bsd) it needs a -t in my experience, otherwise it just gives its --help output without even trying
<TestModule> yeh
<mnemoc> hansg: :<
<mnemoc> TestModule: then do so from there
<mnemoc> TestModule: we only need the a10meminfo from android
<TestModule> kk
<hansg> TestModule, yes getting the script.bin from ubuntu is fine, as mnemoc said we only need the a10meminfo from the *stock* android.
<TestModule> lol, i promise you this is stock
<TestModule> it was pre-rooted so, no need
<hansg> We want the script.bin from *stock* too, but getting that while running something else is fine, it is a file, versus probing runtime settings
<hansg> About your a10-meminfo output: 1) Thanks 2) looks very non exciting, exactly what I expected -> good :)
soooga has quit [Quit: Leaving]
<TestModule> Hopefully the bins are more exciting
* hansg expects the bin to be full of nonsense, like having an lcd in there, and a tv-out, and sata, and wired ethernet, etc ...
<hansg> lunch time for me
hansg is now known as hansg_lunch
Guest50854 is now known as fredy
Ershov has quit [Ping timeout: 246 seconds]
SouL_ has left #arm-netbook [#arm-netbook]
<TestModule> heh, nanda exists
<TestModule> says unallocated
<specing> hansg_lunch: put '*' into /etc/filesystems for mount to try all the filesystems
<specing> # This file defines the filesystems search order used by a
<specing> # 'mount -t auto' command.
<specing> # Keep the last '*' intact as it directs mount to use the
<specing> # filesystems list available at /proc/filesystems also.
<specing> # Don't remove it unless you REALLY know what you are doing!
<specing> *
<TestModule> gotcha gotcha
Ershov has joined #arm-netbook
SouL_ has joined #arm-netbook
SouL_ has left #arm-netbook [#arm-netbook]
<oliv3r> mnemoc: hansg_lunch: http://linux-sunxi.org/A10_Meminfo
* mnemoc hates the abuse of the "a10" token
xman has joined #arm-netbook
<oliv3r> mnemoc: rename and move
<oliv3r> :p
<oliv3r> i did the hard work!
<oliv3r> mnemoc: and for now, the meminfo tool is only for the A1x
<TestModule> ok so the folder should have
<oliv3r> oh! new token to replace a10 token :p
<TestModule> boot.axf drv_de.drv
<TestModule> etc
<oliv3r> s/folder/directory/g
<oliv3r> TestModule: that'st he one, there should be a file called script.bin there
<TestModule> got it
<TestModule> mmcblk0p1
TheLarch has joined #arm-netbook
<TestModule> wait, but the mmc, i wonder if it's just seeing the linux one
<mnemoc> oliv3r: same dram registers in sun3i, and probably also in sun6i+
<TestModule> says nanda is not partitioned, is 17MB
<oliv3r> mnemoc: you can rename it! I just quickly came up with _some_ think
<oliv3r> TestModule: I know that allwinnertech changed how their nanda driver works in later android releases
<oliv3r> TestModule: I started a partition comparission once on the wiki, but never got around to finish it
<TestModule> but the correct one only has three files right
<oliv3r> the correct one has script.bin
<oliv3r> the rest is irrelevant
<TestModule> ok well i have two that have that
<TestModule> ok think i found it
<TestModule> i have script.bin and script0.bin just btw
<specing> somehow I question their choice for ext4 as a NAND flash filesystem...
zumbi has joined #arm-netbook
<TestModule> how so?
<TestModule> You can use ext4 w/o journaling
<specing> I doubt they turned it off...
<TestModule> This, this is true
<specing> I run plain ext2 on my flash stuff though
<specing> sd cards, CF
<TestModule> I read through a bunch of stuff, basically the benchmark guys were saying that more recent optimizations have been done to ext4 vs ext2, so it holds an advantage even when both are "equal".
TheLarch has quit [Quit: Leaving]
<oliv3r> the problem with ext4's journaling is, that the journal trashes the nand
<TestModule> yeh
<specing> oliv3r: Sherlock
<oliv3r> if you have a driver that does flash-load balancing, that works fine
<oliv3r> specing: was @ testmodule :)
<slapin> oliv3r: you will do lots of writes in this case anyway
<TestModule> noatime in /etc/fstab
<slapin> if you want journaling you have to use e.g. ubifs or jffs2
<slapin> as this is designed specially for flash-like storage
<specing> last time I used jffs2 it nuked the boot partition on my phone
<specing> fun stuff.
<specing> tunefs -O ^has_journal is no rocket science
<TestModule> I linked that regarding the claimed memory bandwidth improvement
<TestModule> which libv was complaining about earlier (strapped for membandwidth)
<specing> the sad thing is that the thread goes for miles and people clamped benchmarks into it,... and other useless crap
<oliv3r> well you increase ext4 performance if you disable the journal :)
<specing> Sherlock.
<oliv3r> xda has good info on certain area's but also has a lot of tools
<oliv3r> tools as in noobs
<libv> forums
<libv> attracts the worst kind
<libv> and those who might normally crystalize out tend to not be able to escape
<TestModule> Would this be extreme for our case usages? http://oldpapyrus.wordpress.com/2012/06/02/tune-ext4-for-the-best-performance/
<specing> "For example, do not use data=writeback for root partitions with ext4. My system didn't boot on Ubuntu 12.04"
<specing> I'd use writeback just to prove those 1337 ubuntu noobs wrong
Ershov has quit [Ping timeout: 240 seconds]
<specing> Oh, and Im going to migrate stuff from ext4 to squashfs (+ AUFS) images on my x86 computer
<specing> Now that should be blazingly fast
<hansg_lunch> oliv3r, thanks for doing the wiki
hansg_lunch is now known as hansg
<TestModule> hansg , i put both links back-to-back above
<specing> Have to figure out a way to compile coreutils into a single binary (like busybox)
<br-> heh, good luck with that
Ershov has joined #arm-netbook
<hansg> TestModule, you got the script.bin from the disk which has both script.bin and script0.bin, right ?
<TestModule> yup
Ershov has left #arm-netbook [#arm-netbook]
<hansg> Yep looks like it is made of fail. I suggest to just use the mk802.fex file from sunxi-boards
<hansg> (make of fail, ie: claims to have the axp20x pmu chip, while that is no where to be found on the board)
<TestModule> :(
<TestModule> So just insert my DRAM settings after the fact?
<TestModule> to the mk802 fex
<hansg> Yes you could do that, and set size to 1024. Not really necessary though, AFAIK the fex dram settings are completely ignored the dram controller is programmed by u-boot and then left alone AFAIK
<slapin> wow, as I see, all mk802s have different DRAM settings?
<hansg> So we just put them there for completenes
<hansg> slapin, different models can have different settings yes
<slapin> hansg: and within same model?
* slapin wonders about memory calibration mechanism
<hansg> TestModule, for u-boot use mini-x-1gb as model for now. That has the exact same dram settings and the dram settings is the only thing that differentiates different *A10* u-boot builds
<hansg> slapin, within the same model (and same memory amount) they usually stay the same. The could tweak them per series when they use different dram chips in a different production run, but they don't seem to do that
<hansg> They're going for a conservative 360 Mhz bus speed with a cas of 6, which should work for any dd3 chips they throw on there.
<hansg> There is this magic bit in the emr1 register which they sometimes flip, usually when using 4 dram chips instead of 2. But not always ...
<hansg> Anyways thanks for the meminfo and the script.bin. I've really got to go and do some payed-for-work now.
<TestModule> kk, thanks man
<TestModule> Does anyone here know if the WiFi on these is considered USB ?
<hansg> TestModule, one last note, please let us know if it works, then we can add an official mk802-1gb to the sunxi-u-boot repo and a fex file to sunxi-boards
<TestModule> kk :)
<TestModule> The sun4i defconfig works OOTB ?
<TestModule> that is generated, etc
<hansg> TestModule, yes it is an usb wifi card. If it is not showing up try "modprobe 8192cu"
<hansg> TestModule, yes the defconfig should work ootb
<TestModule> kk, I see it now in lsusb. Nice.
<hansg> TestModule, the modprobe being necessary is a fex file issue, fixed recently, see: https://github.com/linux-sunxi/sunxi-boards/commit/9be2858e2f3a06522b90691d4b96a2c5299edff9
<TestModule> I just re-downloaded everything from scratch about an hour ago w/ stage 3.4 & bsp
<TestModule> Ok cool, so usbc2 = wifi, 1 = MiniUSB 0 = Full ?
<hansg> Nope, usbc0 = mini/micro and is otg, which is generally extremely broken, ok for plugging in a usb-stick, but not much else, better use a hub in the full usb port if you need more ports. usbc1 is the full port usbc2 is internal usb used by wifi
ganbold_ has joined #arm-netbook
<TestModule> Clone of 'https://github.com/linux-sunxi/linux-sunxi.git' into submodule path 'linux-sunxi' failed make: *** [linux-sunxi/.git] Error 1 (I unpacked the stage 3.4 in that dir)
gimli has joined #arm-netbook
rellla2 has joined #arm-netbook
rellla has quit [Ping timeout: 244 seconds]
<TestModule> nm fixed, tricked it
xman has quit [Quit: Ex-Chat]
datagutt has joined #arm-netbook
hipboi has quit [Quit: Leaving]
<jlj> hansg: do you think there is any hope of getting the otg port to work properly in the future? or is the unreliability due to hardware?
<TestModule> my otg works fine? granted, i shoved tin foil to short it in the back
<jlj> mine mostly works but not for anything image related such as webcams or capture devices
<jlj> did the tinfoil trick make it more reliable?
iriki has joined #arm-netbook
<TestModule> Yes
<TestModule> eventually i soldered over it though
<TestModule> how can i target a different branch with the bsp?
<mripard> mnemoc: ok
<iriki> Hi all
iriki has quit [Ping timeout: 244 seconds]
adminuser has joined #arm-netbook
rellla2 is now known as rellly
rellly is now known as rellla
<TestModule> figured it out if anyone cares: git clone git://github.com/linux-sunxi/linux-sunxi.git -b stage-sunxi-3.4 ./linux-sunxi/
<mnemoc> the branch is called stage/sunxi-3.4, not stage-
<TestModule> how are you pulling it without the others
sspiff has quit [Remote host closed the connection]
rellla has quit [Quit: Nettalk6 - www.ntalk.de]
ssspiff has quit [Remote host closed the connection]
bsdfox has joined #arm-netbook
bsdfox has quit [Changing host]
bsdfox has joined #arm-netbook
iriki has joined #arm-netbook
sspiff has joined #arm-netbook
sspiff has quit [Changing host]
sspiff has joined #arm-netbook
iriki has quit [Client Quit]
iriki has joined #arm-netbook
pcat has quit [Read error: Connection reset by peer]
<WarheadsSE> mnemoc: read that email. I've been tweaking with the doozan reconfig myself
<mnemoc> WarheadsSE: iirc that tool sends ioctrl()s to switch from the blessed list of resolutions only
<WarheadsSE> yeah.
<WarheadsSE> works :P
<mnemoc> :)
<iriki> Hello y`all
<buZz> woa
<buZz> thats the rk3306 device, right?
<buZz> that price is insane O_o
<rm> scam
ZaEarl_ has joined #arm-netbook
<buZz> not so sure
<iriki> mby it is
<rm> 276 sold 144 watchers <- idiots
<buZz> rm <- negative
sspiff has quit [Ping timeout: 240 seconds]
<mnemoc> on aliexpress i would try, on ebay. no way
<buZz> :D
<buZz> man, why so harsh :)
<rm> you are being "not so sure" that this is a scam, and calling me "negative"?
<rm> maybe you are one of those 276 sold 144 watchers then
<buZz> and then you are jumping to sexual advances
<buZz> but we only just met!
<rm> get a couple of tablets from the same seller yeah. http://www.ebay.com/sch/i.html?item=281052567836&_ssn=top.seller888&rt=nc&LH_BIN=1
<rm> not expensive at all!
<rm> ...
<iriki> how r u so sure? Doesnt paypal return the money?
<rm> perhaps they hope not everyone will file a dispute
<rm> or hope to b/s some into waiting 40+ days (dispute deadline)
<rm> or some will not know how/where to file
<rm> etc
<iriki> And will they send u stone
<rm> or that
<iriki> *a stone or a brick
piezo has quit [Ping timeout: 264 seconds]
<iriki> Anyway its too good to be true,thats for sure
SouL_ has joined #arm-netbook
SouL_ has left #arm-netbook [#arm-netbook]
rellla has joined #arm-netbook
<mnemoc> the time limit to open paypal dispute is a pain
<TestModule> they won't do the stone ebcause of shipping
<TestModule> to ship a stone to usa or eu would be insane
<TestModule> you're talking, at least 5lbs
<TestModule> TBH, they could be liquidating for the quad RK? ;)
<iriki> nah dont think so :-)
aexl_ has joined #arm-netbook
TestModule has quit [Quit: Page closed]
<aexl_> lo. i get a "Error: password reset denied" on rhombus when i click the link in the requested mail.
<aexl_> i guess it's due to my nich äxl
piezo has joined #arm-netbook
<hramrach> the store has some feedback but way less than the claimed products sold
<wingrime> who here works on xbmc ?
<rm> I believe it's some trick method of pumping fake feedback
<hramrach> it's of cource possible
<mnemoc> wingrime: rellla
<hramrach> you make fake accounts, buy from yourself, post feedback
<wingrime> mnemoc: thanks
<rm> in fact on most feedback scores the product is not shown
<hramrach> Since the e-bay policies about feedbacke requirement became insane this has become the standard practice for ebayers it seems
<rm> (i.e. removed); and on the rest, it's some unrelated stuff like T-Shirts and Jewelry
<hramrach> interesting
<hramrach> the t-shirts are also more expensive than the tablets
cargo122 has joined #arm-netbook
cargo122 has quit [Client Quit]
<hramrach> as a person with fresh e-bay account you have besically no chance of buying or selling anything
<hramrach> all the guides tell you to not trade with such person, and e-bay even has tools to exclude them
<hramrach> how is such system supposed to work is beyond comprehenition
<mnemoc> i've been succesful buying from zero-feedback sellers. but "reasonable" price and after some messages asking questions
<mnemoc> but it's clearly hard to get started
<aexl_> is there problem with the rhombus site password reset or is it me (äxl)?
<jinzo> hramrach, you selling anything related to this chan?
<mnemoc> aexl_: it's likely it doesn't like not-ascii names
<aexl_> okay. thanks!
<specing> hramrach: they buy from themselves and give excellent feedback?
<hramrach> specing: thre is nothing relly stopping you from doing that
<hramrach> jinzo: no, if I was selling something I probably would not choose ebay
<hramrach> but you can sometimes find stuff on ebay that is no longer available new
tinti has quit [Remote host closed the connection]
tinti has joined #arm-netbook
pcat has joined #arm-netbook
hansg has quit [Remote host closed the connection]
iriki has quit [Quit: AndroIRC - Android IRC Client ( http://www.androirc.com )]
<drachensun> hramrach: You would be paying a lot of ebays fees to pump feedback with a product like this, they charge 9% these days
<drachensun> I think the way to get cheap feedback is the 1 cent coupons
<drachensun> I tried to sell them for a previous business to market it but the feedback hounds kept buying them making it just a waste of time
<drachensun> they would seriously buy 6 of them them send you emails about non delivery and demand their money back
<drachensun> 6 cents
Mist__ has joined #arm-netbook
<jlj> feedback hounds?
<drachensun> people trying to get their feedback score up
<drachensun> without performing the actual commerce expected
<drachensun> one guy had 10K positive feedback so he could hide his ripoffs
<jlj> so they buy lots of really cheap stuff to get feedback
<drachensun> the negative comments were damning to "sent me a box of different colored broken plastic pieces, said to file a claim with the shipper, what I ordered was supposed to be 'new in box'"
<drachensun> well his case he was buying cheap stuff
<drachensun> like 1 cent coupons
<drachensun> all his auctions were for high priced, unique collectables and the feedback implied he didn't really ever send them
<jlj> guess they should separate buyer and seller feedbacks
<drachensun> this was a few years ago I did all that, they might have by now
<jlj> I've mostly seen sellers sell stuff "without profit" to get their review scores up
<jlj> the world of ebay feedback is a little weird
<techn_> whee.. got tracking id for cubie :)
Jef91|AFK is now known as Jef91
tinti has quit [Quit: Leaving]
tinti has joined #arm-netbook
aexl_ has quit [Ping timeout: 245 seconds]
tester2 has joined #arm-netbook
tester2 has quit [Quit: Page closed]
pwhalen has joined #arm-netbook
iriki has joined #arm-netbook
rellla has quit [Read error: Connection reset by peer]
sspiff has joined #arm-netbook
popolon has quit [Quit: Quitte]
rellla has joined #arm-netbook
<iriki> Getting problems with %rand A10 slate.... Arghhhh
deathz0r- has quit [Read error: Connection reset by peer]
deathz0r- has joined #arm-netbook
RITRedbeard_ has quit [Read error: Connection reset by peer]
RITRedbeard_ has joined #arm-netbook
iriki_x has joined #arm-netbook
iriki has quit [Quit: AndroIRC - Android IRC Client ( http://www.androirc.com )]
iriki_x is now known as iriki
leming has joined #arm-netbook
iriki has quit [Client Quit]
Mist__ has quit [Ping timeout: 245 seconds]
vinifm has joined #arm-netbook
Quarx has quit []
voronaam has joined #arm-netbook
rsalveti has quit [Ping timeout: 248 seconds]
jukivili has joined #arm-netbook
rsalveti has joined #arm-netbook
<vinifm> hi, there are a AD driver sunxi?
rz2k has quit [Read error: Connection reset by peer]
<WarheadsSE> <.< "AD" driver
wingrime has quit [Ping timeout: 252 seconds]
<vinifm> hi, are there a internal AD converter driver sunxi?
pwhalen has quit [Quit: Leaving]
<WarheadsSE> please, what is "AD"
<vinifm> sorry, LRADC
<vinifm> ADC = analogic/digital conveter
<WarheadsSE> Low Resolution A/D Converter OKay
<techn_> is that gpio implementation ?
<WarheadsSE> Does the A10 chip have that in the block?
<WarheadsSE> I see the imx mxs-mach has a specific implementation of it
<vinifm> techn: I think not, because it is internal
<techn_> I havent seen any references of ADC stuff
<vinifm> it would be a platform driver, specific for sunxi?
<techn_> rechecked datasheet
<techn_> from uboot sources, #define SUNXI_LRADC_BASE0X01C22800
<mnemoc> the A10 is so awesomely complete
<mnemoc> and wasted
<techn_> yep :D
<jinzo> mnemoc, maybe A20 will be very simmilar :P
<mnemoc> i have big hopes in A20's A10 software compatiblity
<mnemoc> with 2x arm and mali units as only diff. but it's just a day dream
<vinifm> LRADC needs driver?
<techn_> yes. but you need to implement it :)
<techn_> or someone
<vinifm> hm, lradc_sunxi.c :)
<mnemoc> first would be to get the register details from allwinner
<mnemoc> we have it?
<jukivili> stage/sunxi-3.4 no longer boot up :| .. was fine few days ago
<techn_> ooh.. a10s based dongle
<mnemoc> techn_: new mk802+ are A10S based too. pictures on the wiki
<techn_> does those boot with our codes?
<mnemoc> tsvetan mentioned he booted his a10s-olinuxino prototype with our code...
<mnemoc> "dimitar" is supposed to submit stuff... eventually
<vinifm> sorry, internal registers for all ARM is not standard? like uart
<mnemoc> vinifm: each IP is different. soc makers buy some here and some there. others are "own"
<techn_> and if IP is same, base address could vary?
<mnemoc> right
<vinifm> IP?
<mnemoc> itelectual poverty
<vinifm> ok, thanks
<mnemoc> vinifm: each unit/controller
<mnemoc> socs put together a ton of separated units in the same chip
iriki has joined #arm-netbook
<vinifm> hm, ADC, for example
<mnemoc> we know there is an IP implementing LRADC and that it lives at 0X01C22800
<techn_> btw. how compatible apple chips are nowadays with ARM ? since I read somewhere that they bought some bigger licencing from ARM.. and modified even ARM architecture
<WarheadsSE> ARM is jsut the IP really
<WarheadsSE> its the everta stuffs & the layouts that make them really different
<WarheadsSE> wow.. s/exerta/extra/
<traeak> latrest crap i read, whether right or not is that apple's biggest thing is increasing the memory throughput for gpu access
<techn_> yes.. but ARM cpu is standard still, but apple jumped of from that? :/
<WarheadsSE> mah, probably just using a large bus to the GPU/memory
<traeak> i guess samsung may have similar tech
<mnemoc> afaik marvell's and apply bought the license to implement something compatible with, not the actual unit design
<WarheadsSE> some of these ARM chips use a 16-bit to DDR2
<mnemoc> s/apply/apple/
<ibot> mnemoc meant: afaik marvell's and apple bought the license to implement something compatible with, not the actual unit design
<mnemoc> same with that company showing of ARMv8 prototypes
<WarheadsSE> aka marvell's PJ4B
iriki has quit [Ping timeout: 248 seconds]
adminuser has left #arm-netbook ["Ex-Chat"]
pwhalen has joined #arm-netbook
<mnemoc> they showed a real server "this" year too
lerc has quit [Ping timeout: 256 seconds]
lerc has joined #arm-netbook
pwhalen has quit [Remote host closed the connection]
DumDum_ has joined #arm-netbook
<DumDum_> Hello everyone, or anyone. Is this the right place to get info on upgrading me Mele A2000? I would like to put the latest android on it.
<specing> DumDum_: android is hated here
<DumDum_> lol, what do i need then to make my mele a2000 work great?
<specing> well
<DumDum_> even the mMele people do not have any suport files listed
<specing> you need mnemoc's kernel
<specing> a distro of your choice
<specing> cedarx drivers (nonexistant)?
<DumDum_> I had high hopes of installing Ubuntu way back when I bought this, but it seems like the info is all over the place and way above my head.
<specing> make that "a decent distro of your choice"
<specing> mali drivers (nonexistant?)
<DumDum_> I have ben looking for an easy instal for Ubuntu or any linux for a long time, months, now I will settle for android jellybean.
<DumDum_> there is nothing on Mele site
<specing> the chinese will not make it easy
<specing> they can only make it cheap
<jelly-home> and fast ;-)
<specing> actually they could have made it easy, but they don't care
<DumDum_> seems like Mele does not have the big following I was counting on. I was wanting a raspberry pi at the time, settled for this
<specing> PI is a different extreme.. its easy but it sucks
<DumDum_> lol
<DumDum_> i need cheap, easy and non-sucking
<specing> that combination does not exist
<DumDum_> you would think that updating android would be easy, but not on the mele
<specing> either it is cheap and easy (RPI) or cheap and non-sucking(A10) or easy and non-sucking (imx6?)
<specing> pick your choice
<specing> DumDum_: if you want it easy then write to ARM holding to finaly opensource the fucking graphics drivers
pwhalen has joined #arm-netbook
<DumDum_> it does not seem ready for the average joe. I bet 90% of us need a file we can just double click and , voila, sd card ready, boot with it.
<specing> well it took me 4 days to find the tx serial pin with an oscilloscope so I could see it boot, so no
<specing> DumDum_: there is a lot of work to be done, so you can start doing it...
<DumDum_> Did you take note of my Nick?
<specing> the sooner you finish, the sooner the average joe will be able to use it
<DumDum_> I can bring you coffees
<DumDum_> I am going to try this http://lundman.net/wiki/index.php/MeLe_A2000
<DumDum_> Android 4 IceCreamSandwich
<DumDum_> scroll down
<techn_> DumDum_: if you still want linux: http://linux-sunxi.org/FirstSteps
<DumDum_> Thanks techn_, but I am not capable of that.
<specing> He has out-of-order execution, thus he cannot follow a step-by-step guide :D
<DumDum_> specing "We of course do not build a whole distribution, we only build u-boot, the kernel and a handful of tools, and then use an existing rootfs to get a useful system." Sounds like it is only half way there. Not exactly a step by step guide. But at any rate what ever you call it, yes I would have trouble with it.
<specing> DumDum_: yup
<specing> But atleast you have a serial port!
<specing> I don't even have that on my tablet
<specing> (only the tx pin)
freakazoid0223 has joined #arm-netbook
<DumDum_> I just got a Samsung Note @, I am blown away by it.
<specing> point being?
<DumDum_> I am easily blown ?
<buZz> thats what she said
<rm> http://linux-sunxi.org/FirstSteps <- LOOOOOOOOOOOOL
<rm> who seriously calls this first steps
<rm> first steps in what?...
<rm> that's like giving people the LFS handbook instead of an end-user distro like Ubuntu
<buZz> :)
<rm> and calling that "here, your first steps"
<buZz> i have to say, i learned a lot while failing my first LFS :P
<specing> Who is collecting the DRAM output?
<techn_> rm: your guide works only for one specific device :p
<mnemoc> specing: we keep a sunxi-boards repo with verified .fex files
<rm> techn_, it works on a number of devices
<rm> assuming that whatever uboot/script.bin I have on there is able to boot on your device without locking up :p
<rm> and it *is* able for a majority of people, it seems
<DumDum_> Thanks RM
<rm> it will be ssh-in only with this image
<rm> if you look for GUI usage, try https://www.miniand.com/forums/forums/2/topics/1
<mnemoc> the next step is to get the dram info using a10meminfo and use that to verify/complete the [dram_para] section
<mnemoc> and then disable the bogus sections
<mnemoc> specing: we have something called "a10_mid" already, can you compare yours with https://github.com/linux-sunxi/sunxi-boards/blob/master/sys_config/a10/a10_mid_1gb.fex ?
<mnemoc> it's probably the same
<mnemoc> but maybe not
<mnemoc> specing: if it is, it would be awesome if you could review the pull requests I got to clean it :|
<mnemoc> the guy has sent 3 pull requests already.... but as it's a very bogus board name and i can't test it I have prefered to ignore
<specing> mnemoc: different
<specing> < dram_chip_density = 4096
<specing> ---
<specing> > dram_chip_density = 2048
popolon has joined #arm-netbook
<specing> has he published any pictures of the board?
ZaEarl_ has quit [Ping timeout: 244 seconds]
tinti has quit [Quit: Leaving]
TestModule has joined #arm-netbook
<traeak> ugh, anyone know where the current mele a1000 official firmware is ?
<DumDum_> traeak, frustrating isn't it.
<traeak> yeah, i suspect i can use the a200 firmware
<DumDum_> I guess?? lol,
ZaEarl_ has joined #arm-netbook
Jef91 has quit [Remote host closed the connection]
Jef91 has joined #arm-netbook
<mnemoc> specing: not that I know of
vinifm has quit [Quit: Saindo]
rellla has quit [Quit: rellla]
rellla has joined #arm-netbook
<specing> mnemoc: bug him to do so... then we shall see
L84Supper has quit [Ping timeout: 256 seconds]
arokus_h has joined #arm-netbook
pwhalen has quit [Remote host closed the connection]
simosx has joined #arm-netbook
<mnemoc> specing: done
<techn_> drachensun: I think I wont do that larger restructure.. atleast for 3.0
<techn_> on 3.7 kernel there is quite a lot moved to general linux codes. mt-input api works there
lkcl_ has quit [Ping timeout: 260 seconds]
freakazoid0223 has quit [Quit: Leaving]
SouL_ has joined #arm-netbook
SouL_ has left #arm-netbook [#arm-netbook]
DumDum_ has quit [Quit: Page closed]
arokus_h has quit [Ping timeout: 244 seconds]
Jef91 has quit [Quit: Instantbird 1.3 -- http://www.instantbird.com]
popolon has quit [Quit: Quitte]
servili007 has joined #arm-netbook