tinti has quit [Read error: Connection reset by peer]
focus_it has quit [Quit: Leaving]
<j^>
yes
<hno>
slapin_, No idea what you talk about. The clock is configured before AHB gating is enabled, and in lichee tree the file is board.c, not clock.c. Clock settings moved to clock.c in lichee-dev-mmc and inherited from there to sunxi
<hno>
but NAND have never been used in lichee-dev-mmc or sunxi branches.
<rz2k>
j^: libdri2-dev and libdri2-1 both installed miss libdri2.so symlink to libdri2.so.1
<rz2k>
thats linaro bug :p
<hno>
rz2k, that's normally created by ldconfig
<hno>
and should only be needed in -dev package.
<hno>
runtime should use .so.1
<j^>
rz2k: latest linaro created that symlink for me
<rz2k>
dont know why it doesnt work, by the way, mali-400 has its own dri2 stuff like swapbuffers, but name of the functions are different.
<rz2k>
oh, cool
<rz2k>
okay then
<j^>
but can add the comment again and leave it in the packages above if you think its still an issue
popolon has quit [Quit: Quitte]
<rz2k>
lets leave it like that for a moment. also libv did a big patch to mali-libs, so eventually I will rewrite it, or you :p
<j^>
if i can get it to work i might
<hno>
slapin_, gating and nand clock is already enabled when NAND driver is initialized.
<hno>
only that the NAND clock is running a bit slow.
<hno>
slapin_, sr32 macro is being removed. u-boot have other convenience functions for bit field manipulations.
<hno>
slapin_, see [clear][set]bits_le32() family of functions.
<hno>
s/clear/clr/
<ibot>
hno meant: slapin_, see [clr][set]bits_le32() family of functions.
<slapin_>
hno: ok, have you seen NAND_SetClock's wiched attempt to set reserved bits 12 and 13 to 0? which they falsely call DIV_N? is it datasheet bug or code bug?
<hno>
I have not studied that part of the NAND driver.
<hno>
more focused on the controller as such.
<slapin_>
hno: ok, life will show then
<hno>
Probably it does not matter if clock or AHB gate is enabled first, as long as they both are before you even think of touching a NAND controller register.
<hno>
would expect a hard hang if only the AHB gate is enabled.
Brandon15811 has quit [Excess Flood]
Brandon15811 has joined #arm-netbook
<hno>
slapin_, not sure if code or datasheet is wrong on the N bits. But kernel source do the same.
<hno>
But with the default being 0 and everyone setting it to 0 there isn't much that can go wrong.
<hno>
A quick probe.. datasheet is correct code is wrong.
<hno>
sun4i#mw.l 0x1c20080 0x7fffffff
<hno>
sun4i#md.l 0x1c20080 1
<hno>
01c20080: 0303000f ....
ssvb has quit [Quit: Leaving]
<hno>
so now we know Allwinner writes both crappy datasheet and crappy code. Can't trust either.
penguin42 has quit [Quit: Leaving.]
<lundman>
supplies!
<slapin_>
hno: awwwww
<Turl>
hno: at least they're not both wrong simultaneously
<Turl>
:P
stefanro has joined #arm-netbook
stefanro1 has quit [Ping timeout: 252 seconds]
Sv has quit [Read error: Connection reset by peer]
<slapin_>
NAND init sequence works surprizingly well, successfully executed reset command on chip, will read chip ID tomorrow, if all that is not a dream.
<slapin_>
gnight, all! 5:50 here...
Sv has joined #arm-netbook
merbzt has quit [Ping timeout: 268 seconds]
vinifm has quit [Quit: Ex-Chat]
Sv has quit [Quit: Sv]
pwhalen has quit [Ping timeout: 252 seconds]
cnxsoft has joined #arm-netbook
tuliom has quit [Quit: Konversation terminated!]
drachensun has joined #arm-netbook
<Mehhh>
Does anyone know which firmware will allow host / otg so i can short out pin 4 / 5 and have it boot?
Brandon15811 has quit [Ping timeout: 246 seconds]
Brandon15811 has joined #arm-netbook
Brandon15811 has quit [Excess Flood]
Brandon15811 has joined #arm-netbook
Sv has joined #arm-netbook
Sv has quit [Changing host]
Sv has joined #arm-netbook
Sv has quit [Remote host closed the connection]
Sv has joined #arm-netbook
Sv has joined #arm-netbook
Sv has quit [Changing host]
<Turl>
lundman: ping
<lundman>
pong
<Turl>
lundman: is it worth running with compression on on a folder with multimedia?
<lundman>
nope
<Turl>
I'm guessing no
gsilvis has quit [Read error: Connection reset by peer]
gsilvis has joined #arm-netbook
<lundman>
zpool/secure compressratio 1.00x
<Turl>
so I just make another zfs on there and turn off compression right?
<lundman>
it cant do a thing
<lundman>
doesnt hurt either per se
<Turl>
what about dedup lundman?
<lundman>
yeah
<lundman>
i didnt gain anything on dedup for media
<bsdfox>
Turl, you using freebsd or solaris?
<Turl>
bsdfox: linux
<bsdfox>
:(
<lundman>
rpool dedupratio 1.00x -
<lundman>
er wrnog pool
<lundman>
zpool dedupratio 1.00x -
<Turl>
bsdfox: wanted to give Debian GNU/kFreeBSD a try too
<lundman>
zpool/secure encryption aes-256-ccm
<Turl>
you encrypt your multimedia?
<lundman>
yes
<Turl>
why? o.O
<lundman>
having a hell of a time to get ccm-aes to work right on linux though
<Turl>
night
Sv has quit [Remote host closed the connection]
Sv has joined #arm-netbook
Sv has joined #arm-netbook
Sv has quit [Changing host]
Quarx has joined #arm-netbook
rsalveti has quit [Ping timeout: 246 seconds]
rsalveti has joined #arm-netbook
rellla has joined #arm-netbook
ssvb has joined #arm-netbook
gimli has joined #arm-netbook
RITRedbeard_ has joined #arm-netbook
RITRedbeard has quit [Read error: Connection reset by peer]
<libv>
but... not my goal, i am going to clean some things up, push this to github, so you guys can go play further
<sky770>
sorry to interrupt; got 2 questions right here
<sky770>
1st: does recently market hogging dual cores likes amlogic 8726 m/m3 or the freescale quad core ones ahve any significant performance improvement in Linux running on these platforms?
<rz2k>
I have 1080p vga output for testing gles stuff
<sky770>
V/S single cores like AllWinner Tech's A10
<rz2k>
interesting how it will work out
<sky770>
Yo rztk thx for replying; you got 1080p on which SoC?
<sky770>
and OS?
<rz2k>
sky770: that was reply to libv :p
<rz2k>
about quake3 timedemo
<sky770>
:P
<sky770>
but is that on Linux
<rz2k>
yes
<rz2k>
check mali400 page @ wiki
<sky770>
oh wait... :O certainly its on nux..omfg..quake3d?! Waaaaaw :D
<sky770>
on it.
* sky770
*happy face*
<rz2k>
we have it working for some time, I've compiled q3 for nokia couple days ago but I failed
<rz2k>
libv: I've used headers from Snowball mali-400
<rz2k>
package
<rz2k>
something might be broken
<libv>
rz2k: well, if you want framebuffer, this is known working now with the headers i plucked from the web
<libv>
no x11 for me, as that just gets in the way of lima work
<libv>
codewise, this time round, not politics :p
<sky770>
idk, am kinda having *shrugged* feeling towards allwinner's SoC(s)
<sky770>
waiting for other quad cores ones to come aroun like the freescale i.MX6
<sky770>
around*
<libv>
sky770: my guess is that it has the most support from any open source community so far, and that it is most probably the most open source not-unfriendly soc around atm
<libv>
getting your own stuff running on allwinner, from bootloader on, is not too hard
<libv>
and the binary blob situation is better than for instance the pandaboards
<libv>
especially since the 3d accel has things in the pipeline
popolon has joined #arm-netbook
<sky770>
+1 libv
<sky770>
but is there any significant performance difference between the quad core ones coming in V/S the Allwinner A10 SoC in terms of XBMC performance? Well, its obvious that we dont have any practical insights yet..but how do you feel about it ^ ?
<libv>
depends on how well you can hook up xbmc.
Holo_ has quit [Quit: Connection closed for inactivity]
<sky770>
hmm
grimes99 has joined #arm-netbook
<sky770>
Guys, I gotta for now; c ya around
<sky770>
tkc
sky770 has quit []
<mnemoc>
libv: hi, can you push your patchset for mali-libs directly? I can't get the 2/7 applied
<mnemoc>
error: corrupt binary patch at line ....
<libv>
mnemoc: ok
<libv>
mnemoc: yeah, those copied blobs are a pain
<mnemoc>
hopefully we won't need more bin blog patches after that one :)
fluxi has quit [Ping timeout: 276 seconds]
grimes99 has quit []
vinifm has joined #arm-netbook
cnxsoft has quit [Read error: Connection reset by peer]
freakazoid0223 has joined #arm-netbook
<libv>
mnemoc: pushed
<mnemoc>
libv: thanks!
gsilvis has quit [Read error: Connection reset by peer]
gsilvis has joined #arm-netbook
kaspter has joined #arm-netbook
kaspter has quit [Client Quit]
focus_it has joined #arm-netbook
<vgrade>
libv: picked up you new mali-libs. xorg is up with mali drivers \o
<vgrade>
libv: but getting undefined _mali_clz_lut in libEGL.so when running test
<libv>
but the test should only work on the fb libs :)
<vgrade>
libv: ok
<libv>
vgrade: what libs were installed?
<vgrade>
x11
<libv>
what version, what abi?
<vgrade>
armhf
<libv>
standard kernel should have r3p0
<vgrade>
linartolinaro
<vgrade>
r3p0
<libv>
_mali_clz_lut is present twice in libMali.so
<vgrade>
so good
<vgrade>
they named it twice
<libv>
no, inconclusive
<libv>
vgrade: you just ran make to get the test binary, right?
<vgrade>
yes
<vgrade>
just followed the instructions
<libv>
point an editor at /usr/include/EGL/eglplatform.h
<libv>
mali_native_window is really just there for raw fbdev
<libv>
i think one would need to use __unix__
<vgrade>
I'm shareing monitor between irc and mele output , sec while I replumb
<libv>
surely you can read this header file over ssh
<vgrade>
not tried ssh yet
<libv>
install ssh, and relieve yourself of a lot of pain
<libv>
ethernet and wifi on the mele work
<vgrade>
ok header up
<vgrade>
and ssh
<vgrade>
id
<rz2k>
we should probably do our own benchmark page
<rz2k>
like r-pi
Quarx has quit []
<vgrade>
libv: same error with framebuffer libs
freakazoid0223 has quit [Quit: Leaving]
merbzt has joined #arm-netbook
<techn>
vgrade: what you are trying to run?
<vgrade>
test/test in mali-libs
<techn>
are you sure that you have libs installed to correct place.. and not having mesa egl libs?
<vgrade>
techn: the new structure in mali-libs git has a makefile which should install the right libs
<techn>
in other works.. where are your libEGL.so* located?
<techn>
*words
<vgrade>
in /lib
rellla has joined #arm-netbook
<techn>
there could be mesa libs also present.. under /usr/lib, if I remember correctly
<vgrade>
ah right so there are
Sternennebel has quit [Ping timeout: 246 seconds]
<vgrade>
some in /usr/lib as well
Sternennebel has joined #arm-netbook
MMlosh has quit [Quit: Bye...]
<vgrade>
but they look like mali as well, but different version
rellla has quit [Client Quit]
<vgrade>
I wonder if one lot came in with the hwpack
<techn>
hwpack libs were located under /lib
<rz2k>
vgrade: /usr/lib/arm-<tab>/mesa-egl/
<rz2k>
kill those
<techn>
mali-libs seems to install libs under /usr/lib
lerc has joined #arm-netbook
<vgrade>
rz2k: don't have that dir
<vgrade>
rz2k: sorry I do
mnemoc has quit [Excess Flood]
mnemoc has joined #arm-netbook
<vgrade>
ok I have triange
<rz2k>
install mesa-utils-extra and run es2_info
<rz2k>
if you will have it crash - everything is ok
<drachensun>
Yup, we are getting a lot of press right now
<drachensun>
we are over 5k, 10% there
<focus_it>
hmm seems to offer Linux on the NAND in some of the options. Is that easy to do if you had your own SDCard? Could I flash a Cubie board or Mele 2000 from SDCard?
<L84Supper>
drachensun: did you have any problems getting the project up on indiegogo?
<rz2k>
$50k is a bit too much for a ly-f1 (like) resell with linaro/our kernel inside.
<rz2k>
am I right?
<drachensun>
L84Supper: nope, I did have problems trying to get it on kickstarter
<drachensun>
L84Supper: They dismiss the projects with no explanation, indiegogo was easy
ZaEarl_ has joined #arm-netbook
<drachensun>
rz2k: What part is too much? Most of that money is paying for the devices, shipping in, shipping to the contributors and crowd funding fees
<L84Supper>
rz2k: the next couple of days should be a good indication after the story spreads
<drachensun>
rz2k: Oh I gotcha, you think its not gonna happen
<rz2k>
thats the matter of marketing/advertisment
<drachensun>
yeah, thats the challenge right now
<mnemoc>
drachensun: what have you done already on the software side of A10?
<mnemoc>
or you are waiting until you are "hired" to dive in?
<drachensun>
heh, no I've done a lot of the prelim work
<drachensun>
mostly thought its setting a build and provisioning system
<drachensun>
so I can handle flashing all these devices
<drachensun>
and loading onto the internal memory was tricky
<rz2k>
I'm just sort of surprised that someone will buy ly-f1 (unit price http://www.alibaba.com/product-gs/530060157/LY_F1_7_A10Cortex_A8_tablet.html) for $120 considering that you likely will use our repos with known problems (3D acceleration fails, cedar being barely used, countless shit-code by allwinner in kernel) and say that "it works".
<rz2k>
I'd bought your tablet if you will bring G2D x11 support for example.
<drachensun>
yeah, thats my biggest goal, having the repository setup
<drachensun>
so it can just be apt-get upgrade to get the latest fixes as they get working
<mnemoc>
i'm surprised that someone who has shown zero lines of code of anything will get funded to do build the code we do and maintain
<rz2k>
rewriting crap code is your biggest goal, if you want it to go big
<rz2k>
lol
<drachensun>
I've submitted some lines of code
<drachensun>
I got a lot of grief about spacing and comments with my first patch so I thought I would hold on to some of my work until I had time to make it nice and pretty
<drachensun>
and this project has taken lots and lots of hours
<drachensun>
so I haven't gotten to that
<drachensun>
but I will share it
<drachensun>
I will tell you though, more than once I started on something and had someone post a patch fixing it before I got done
<drachensun>
heh I'm gonna have to code faster with all the real pros in this chat room
<rz2k>
spacing and comments will not do the job. support guys who want and can move this project (like mnemoc, hno, libv) and then there is a chances. also allwinner officially dont support linux, by the way, you cant commercially rely on our work.
<rz2k>
s/is a/are/
<ibot>
rz2k meant: spacing and comments will not do the job. support guys who want and can move this project (like mnemoc, hno, libv) and then there are chances. also allwinner officially dont support linux, by the way, you cant commercially rely on our work.
<rz2k>
would be cool to be first guys with Lima running on linux, though.
<mnemoc>
i'm sure we will :)
<rz2k>
we should make a list of bottlenecks in kernel, like NIC being CPU only
<focus_it>
ibot: when spoke with Allwinner they pointed at sunxi for all my Linux needs. indirectly they rely on you!
<rz2k>
lol
<rz2k>
awesome
<Turl>
ha, really? how nice of them
<rz2k>
I'm just worried that people will fund that $50K, buy the devices and then see "barely moving windows in X11" and flood this place with "we've f@cking paid".
<mnemoc>
rz2k: A table of the state of each driver/subsystem/lib and an small TODO/"what's next?" would be nice
<mnemoc>
rz2k: maybe drachensun will indeed surprise us all...
<rz2k>
I hope
<jelly-home>
NIC being CPU only # that's the SoC NIC? What about an USB2 connected one like 9710:7830 MCS7830?
<Turl>
rz2k: at most they'll flood tom I guess, we didn't make the boards
<rz2k>
yeah, SoC NIC called sun4i_wemac
<mnemoc>
focus_it: allwinner point you to our project??? amazing
<rz2k>
nobody tested usb ones
<rz2k>
mnemoc: they know this place since Tom GPL'd kernel, I think
<Turl>
I think rm was using USB NIC rz2k
<rz2k>
jelly-home: if you have one - feel free to run the test
<jelly-home>
I blame the communist party for not having even a MK802 in mygrubby hand yet
<rz2k>
and document on wiki
<mnemoc>
I use a usb/dm9601 for ethernet on the a13-olinuxno.... haven't benchmarked it, but feels decent
<rm>
wa
<rm>
t
<rm>
jelly-home, yes the MCS7830 works exceptionally well
<rm>
I use it on my MK802 VPN/IPv6 router
<rz2k>
rm: could you document it on wiki/USB-Ethernet ?
<jelly-home>
oh, nice
<rz2k>
like with simple test and kernel config needed
<rz2k>
and device name
<rm>
people reported problems with ASIX NICs though
<rm>
hope 3.4 kernel will work better w/those
<mnemoc>
3.4's usb needs love
<rz2k>
we can backport stuff from there if it really needed and dont rely on 3.4 only
<rm>
usb/dm9601 works too, except it's an USB 1.1 device
<focus_it>
AllWinner sales specifically pointed at http://linux-sunxi.org and many other links with sunxi for the relevant information I was seeking
<rm>
that is so nice of them :D
<rz2k>
why they didnt send us their new patches (sun4i_gpio) god damn them?
<rz2k>
if they point people here.
<mnemoc>
very true
<rm>
yep exactly my point
<focus_it>
There is a twist to the AllWinner tale - I bought $1000 PRO-A10 to develop Linux. Originally they said they support Linux. Then after I purchase, the seller of PRO-A10 developer system said they only support Android.
drachenphone has joined #arm-netbook
<rz2k>
lol
<rz2k>
you mean that wits-tech devkit?
<focus_it>
yep
drachenphone has quit [Read error: Connection reset by peer]
drachenphone has joined #arm-netbook
<focus_it>
They insisted they "Only support Linux and not Android"
<focus_it>
The utter fsckwits!
<jelly-home>
and then you returned the product?
<focus_it>
I had to send them a drawing saying Android runs on Dalvik, and Dalvik runs on LINUX!!!!!!!!!!
<rm>
if you buy $1000 dev kits instead of a $70 Mele, I might have a bridge to sell you
<mnemoc>
or a cubieboard :p
<focus_it>
Ignorance on my part
<focus_it>
I wasn't aware of what the PRO-A10 was
<rm>
One Does Not Simply buy a cubieboard
<focus_it>
or was not
<merbzt>
I have the eiffel tower for sale also
<mnemoc>
rm: true
<focus_it>
I feel a clown
<rz2k>
shit happens
<rz2k>
wait
<focus_it>
But can't back down because the company wants to juggernaut forward with Linux plands
<focus_it>
plans
<focus_it>
Already laid off one person for lack of progress
<mnemoc>
focus_it: we got the option of buying 4 pro-a10 kits for $250 each... but hopefully someone found the mele, and then Tom started selling it on aliexpress
<drachenphone>
Did they access to any docs beyond what's currently available on the wiki?
<rz2k>
if you've bought it, you should probably recieve latest A10 SDK?
<drachenphone>
Did they give you access I mean
<mnemoc>
focus_it: and git access to their linux repo
<focus_it>
I believe I have many good things on the CD, but without their cooperation for tech support I'm stuck
<focus_it>
not git access to their Linux repo
<focus_it>
should I ask for it?
<drachenphone>
Definitely
<focus_it>
I ask monday
<drachenphone>
If they sent you here they must want us to have access to all that to help you
<focus_it>
what can i grab / get that might get me to build a Lubuntu on to A10?
<rz2k>
if you will have something interesting, you should probably share it with hno, he found wrong info in datasheets, might be they updated them.
<mnemoc>
focus_it: they used to ask you to sign a GPL-violating NDA of not sharing any of the code in their linux repo
<focus_it>
They not ask me yet
<focus_it>
One first day we get into fight about Linux support
<mnemoc>
focus_it: having a feed that in git form (with history) would help a lot to integrate their imporvements
<mnemoc>
s/a feed//
<ibot>
mnemoc meant: focus_it: having that in git form (with history) would help a lot to integrate their imporvements
<mnemoc>
focus_it: also and having a feed of their improvements/fixes (commit by commit, not bulk)
<mnemoc>
unless they force you to sign a GPL-violating NDA
<focus_it>
so if I get git access I should grab all the stuff in git and grab all the history unless NDA
<focus_it>
if offer NDA, I refuse and argue they must not violate GPL
<focus_it>
since I got off on the wrong side of tech support, I don't think that will make matters any worse
<rz2k>
or you can sign it, say that it actually is not legal in your country to violate GPL and publish the sources, as they dumped you with linux support. :p
drachenphone has quit [Read error: Connection reset by peer]
<specing>
technically their NDA is invalid
drachenphone has joined #arm-netbook
<focus_it>
my company may have a dim view of legal stuff
<mnemoc>
yes, but if you want to make an a10-based product breaking the (invalid) agreement can give you troubles
<mnemoc>
focus_it: ask them for access to their kernel git
<rz2k>
drachenphone: drachensun: what do you think about selling Freescale i.MX6q stick computers/tablets, freescale imx6 is in mainline of the kernel, vivante gpu should have X11 support as they are Linaro members.
<captainigloo>
there is a way to read power on/off button in u-boot on the mele a2000 ?
pawel5870 has joined #arm-netbook
<rz2k>
captainigloo: it was somewhere in /sys, axp drivers export it as gpio or something
<captainigloo>
rz2k: and it's the same for the led ?
<mnemoc>
focus_it: once you get an answer, please let us know
<focus_it>
I will ask about git and kernel - may be they softened their stance after I tell them Android ran on Linux because they ask engineer to give me support - but haven't heard anything afterwards
<drachenphone>
Yeah its available on the i2c bus from the axp209
<rz2k>
captainigloo: dont remember, but it should be exported
<focus_it>
I press buttons for best support and make available as much as I can
<rz2k>
android controls it somehow :p
<captainigloo>
not so simple then :)
<focus_it>
I spend a lot of time with MK802 and make Gambas GUI developer package work on it
<focus_it>
This link is not for public release but it allows Lubuntu to run Gambas
<mnemoc>
focus_it: if you have an mk802, can you share it's script.bin/fex with me? for the bsp
<drachenphone>
Rz2k I think the cost of the tablets made me think I would have trouble selling them but I'll take a look when I am back to my PC.
<focus_it>
I have an off the shelf MK802 - does that have an accessible script.bin/fex
<rz2k>
drachenphone: imx6q is still not released, all imx6q tablets on alibaba/aliexpress use preproduction silicon that is still has higher price than when it will be selling world wide to minor customers (planned in q4 2012, but I dont think they will do it on time).
<mnemoc>
focus_it: sure, just mount /dev/nanda
<mnemoc>
focus_it: /dev/block/nanda in android
<drachenphone>
Rz2k very interesting. All the user space work should translate easily too
<rz2k>
drachenphone: I really would buy imx6q tablet/devboard/stick-pc even at this state just because it is newer and from better manufacturer than allwinner stick that is now "rather old". I think many guys will have same opinion.
<rz2k>
unfortunately, hiapad.com still doesnt manufacture their stick-pc, wandaboard is only preorders and sabrelite is a devkit (and also costs a ton).
<focus_it>
mnemoc could those mount commands be run from Lubuntu?
<rz2k>
might be freescale finnaly started to ship imx6q final silicon
rsalveti has quit [Ping timeout: 246 seconds]
<mnemoc>
charbax promised to sell samples...
<rz2k>
"Dear customer ,if you order more freesacle CPU , the cycle time is 35 days at least."
<mnemoc>
:
<mnemoc>
:<
rsalveti has joined #arm-netbook
<Mehhh>
do the script.bin things have to be run from a microsd or can it apply to nand
<Mehhh>
for mk802's that don't have host mode by default
<mnemoc>
script.bin is a binary data structure, it doesn't run
<Mehhh>
i know this, added, run from, whatever
<focus_it>
mnemoc: I note it - and get the file for you tomorrow after I set it up again
<rz2k>
charbax promised many stuff, would be awesome if he will run a good ARM shop, last time I've checked, he resell mk802II's
<ZaEarl_>
script.bin can be loaded from wherever, nand, sd.
<mnemoc>
focus_it: awesome, thank you :)
<rz2k>
other stuff was "coming soon"
<mnemoc>
Mehhh: u-boot loads it into memory. nand u-boot reads it from nanda, mmc u-boot reads it from the boot partition in the card
<rz2k>
drachenphone: all the userspace work should be already done by linaro (hwpacks), freescale is member of that organization. I might be wrong, though.
<Mehhh>
awesome
<Mehhh>
So i'll have to do it twice
<Mehhh>
granted i assume most linux images have it already
<Mehhh>
the structure is copied over i assume
<rz2k>
drachenphone: also guy named steev from here works at Genesi with imx6q, they previously sold cool imx5 laptop and it was somewhat popular, dont know if they will do the imx6q one.
<mnemoc>
it seems genesi is distracted with other projects :
<mnemoc>
:<
Hexxeh has quit [Quit: Connection closed for inactivity]
<drachensun>
I definitely prefer the idea of working with freescale. I actually know the local rep here so I can probably get a call back about thing that might come up.
<mnemoc>
drachensun: do you plan to build hardware at all or only to resell/rebrand devices?
<drachensun>
mnemoc: I would like to build hardware
<rz2k>
freescale should have reference design and hiapad stuff is pretty simple (just imx6q, two ddr3s, sd card as a nand storage and a usb-wifi from realtek)
<mnemoc>
drachensun: a 11-13" laptop with good screen would be awesome
<drachensun>
design for manufacture was my day job for most of the last 6 years
<drachensun>
but down side of that is I'm well aware of the volumes needed to get the cost anywhere near what people will usually pay
<rz2k>
somehow arm chromebook got his way to people :p
<drachensun>
so I'm really hoping we can get enough interest to justify a manufacturing run of an open hardware design
<mnemoc>
drachensun: make an imx6 eoma68 card, and then share the laptop effort with lkcl
<mnemoc>
rz2k: but the screen is disappointing :
<mnemoc>
:(
<rz2k>
you should also talk with lkcl the Luke, he has a lot of experience in openhw
<drachensun>
Yeah I started talking with him just today actually
<rz2k>
(also I dont believe that EOMA68 without support from big compaines will ever go out, but thats just my two cents)
<focus_it>
drachensun: I'm working on A10 with LCD SO-DIMM 200 pin card running Lubuntu and Gambas. Completely GPL'd. The circuit is going to be done in KiCAD.
<Mehhh>
sorry i found this channel by trying every combo i could think of
<Mehhh>
after banging my head against the wall in electronics and raspberrypi chans
<drachensun>
routing the ram lines is pretty tough. I used to do a lot of routing but my last attempt to route ddr had some trouble, thats how I met this engineer
<rz2k>
yeah we try to maintain the wiki, but it isnt updated to "right now" stuff, but you can get general idea from there.
<drachensun>
focux_it: excellent, I haven't tried out kiCad, I just started playing with it the other day, but I've been pretty busy
<drachensun>
focus_it: Does it have support for creating busses with trace length matching?
<L84Supper>
KiCAD has limited tools for high speed complex routing
<Mehhh>
I wish the MK802 pcb could be made smaller / thinner. Would it still work if i "cut" the PCB where the big USB and power port is?
<rz2k>
you should google for that, but I dont think so, as far as I know only orcad/pads/altium have something like decent lenght matching
<Mehhh>
if i'm only using the side one
<Mehhh>
where the slits are
<rz2k>
its multilayer one
<rz2k>
dont do that
<drachensun>
I almost want to just figure out how to boot leg orcad
<drachensun>
shouldn't say that in public right :-)
<rz2k>
eagle recently got the diffpairs meander
<rz2k>
lol
<focus_it>
drachensun: it has features for creating tracks of specific lengths, but then when you route it, it may not retain that!
<L84Supper>
Allegro is the PCB application by Cadence
<L84Supper>
there used to be Orcad layout for simple stuff
<drachensun>
I thought they were all owned by the same thing now
<drachensun>
I had them all as the same package before
<Mehhh>
ok, so I need the toolchain just to edit or add OTG so i can use a 4pin cable?
<drachensun>
oh how sweet it was...
<L84Supper>
Cadence bough Orcad, but people confuse orcad capture with layout
<focus_it>
dranchensun: personally I believe the lengths issue is oversold, there are so many tricks to get things like this right. I plan on several iterations to get the board right and crash free.
<drachensun>
yeah, it never seems to be right the first time, no matter how many checks
<L84Supper>
it depends on who is routing the PCB, the tools can only do so much
<drachensun>
yup
<L84Supper>
PADS and Altium are the other popular tools for PCB, KiCAD is for simple boards or masochists :)
<focus_it>
L84Supper: KiCad is lightening quick when you know how to work it - all the data is in text files to modify to heart's content
<focus_it>
under linux KiCAD wins hands down if you go against Altium
<rz2k>
who's not going against the Altium after they've moved to China and breaked the whole company to little developer centers around the world?
<focus_it>
For complex designs, like an A10 board, my bet is KiCAD will win - we got both in office.
<focus_it>
I got a CN engineer being trained up on KiCAD because i can't stand Altium
<focus_it>
I hope to have the A10 board designed asap and released for anyone to go make their own boards
<L84Supper>
is Allwinner selling parts in low volume now?
<focus_it>
You can get them from Wits-Tech - about $11 for 20 pieces
<L84Supper>
when did they start selling them?
<rz2k>
yeah, that was unknown info
<Mehhh>
Wait, allwinner sells LV now? Quick, someone kickstart a "design your own board" company
<Maqs>
operating temperature: °C. so F does not work? :-(
<L84Supper>
yeah, only metric heat
<Turl>
Maqs: F has been obsoleted for ages, upgrade your knowledge ;)
<Turl>
:P
<L84Supper>
does the axp209 also synthesize some clocks for A10 or is it only a fancy power switch?
<mnemoc>
the US' hidden love for the empire
<focus_it>
mnemoc: thanks
ZaEarl_ is now known as ZaEarl
<mnemoc>
focus_it: the reference design sends one usb host to the wifi module and the other to a hub
<L84Supper>
oh 45 page axp209 data sheet
<Mehhh>
what is the easiest way to enable OTG so i can solder 4/5 to use a 4pin USB that these chinese vendors sold to me as 5pin
<mnemoc>
Mehhh: only one of the 3 usb controllers is otg capable
<focus_it>
mnemoc: thanks again - I assume the wifi drivers are in sunxi somewhere and it caters for a range of these devices? So I should check for a supported device and then go buy some samples.
pwhalen has joined #arm-netbook
<mnemoc>
focus_it: the rtl8188 is the most common wifi chip.... probably because it's the cheapest :p
<ZaEarl>
and it's the best!
<mnemoc>
:)
<focus_it>
ok work on that chip for wifi
<Turl>
lol @ "it's the best" :P
<focus_it>
about ram: A10 uses 1GB - can it do 2GB?
ibot has joined #arm-netbook
<mnemoc>
focus_it: in theory....
<mnemoc>
focus_it: but allwinner discourages it
<focus_it>
probably because it can then run a proper Linux distro with X!!
<focus_it>
I aim for 2GB
<focus_it>
Any substitute for NAND? What I want is something that functions like SDCard but doesn't have wear issues
<mnemoc>
focus_it: no one has made A10/2GB yet... there might be a reason
<mnemoc>
focus_it: emmc
<L84Supper>
mnemoc: any reason why they discourage 2GB? flaky silicon design?
<mnemoc>
L84Supper: don't know
<focus_it>
is it economical to put emmc - is emmc in volume very expensive?
<focus_it>
only samsung do emmc?
<Mehhh>
mnemoc , on all MK802's? I have the original. Purple top.
<Mehhh>
Someone said i have to enable it in script.bin
<Mehhh>
to use with the mini usb
<mnemoc>
Mehhh: I'm talking about the SoC. I don't know anything about the mk802 in particular
<hno>
mnemoc, sounds like it's time to write Eva a letter again.
* hno
read todays
<hno>
backlog
<L84Supper>
A10 in POP with DDR3 would have been nice
<mnemoc>
hno: yes, if they are using us to justify their gpl-violatings, they should at least give us access to their internal trees
<hno>
or open up data sheet and user guide, and provide some level of technical support on those.
<mnemoc>
absolutely
<specing>
focus_it: debian X11 runs just fine on 256M *trollface*
<hno>
don't really care that much about more code from them. But documentation like the PLL issue with rev 6 would help a lot.
<L84Supper>
register spec
<hno>
s/6/C/
<ibot>
hno meant: don't really care that much about more code from them. But documentation like the PLL issue with rev C would help a lot.
<hno>
the sun5i trees they ship now is a gpl disaster.
<hno>
L84Supper, yes. the "user guide" is register specification. Not complete but coversa lot.
<L84Supper>
hno: I meant it would be nice if they just shared their complete register spec and bug list
<hno>
not sure they have a complete spec..
<mnemoc>
when the user manual was leaked tom said it was the same document he used when working for allwinner
<hno>
It's quite clear Tom had additional source of information on the DRAM controller.
<mnemoc>
maybe chinese "notes" for each controller?
<L84Supper>
heh, having RAM controller specs would be nice
<hno>
L84Supper, yes.
<L84Supper>
memory configs are complicated even with specs
<hno>
mnemoc, there is no comments either, looke like indiegogo is having problems.
merbzt has quit [Ping timeout: 240 seconds]
von_fritz has quit [Quit: vonfritz leaves, don't panic]
gimli has quit [Quit: Verlassend]
merbzt has joined #arm-netbook
revident has joined #arm-netbook
<hno>
L84Supper, I think some of the parameters can only be determined by measuring the board. But default set of timing parameters should be good for boards fulfilling the design requirements.
<hno>
it's not really that many parameters that are board specific.
<lkcl>
L84Supper: the A10's DDR3 it's just that nobody's ever tested it in "double-chained" mode (8 ICs) or put down the ultra-expensive 4Giga-bit ICs.
<lkcl>
so it would need someone with a budget and some balls to actually test it out
<lkcl>
theoretically there's nothing stopping the A10 from using 2Gbytes of RAM....
<hno>
lkcl, is there design recommendations & requirements on dual rank DDR3 from allwinner?
<lkcl>
hno: not that i know of - that's the point :) nobody's ever tried it!
<hno>
I only know of a bit flip in the registers, but have a feeling it should impact some timings as well.
<hno>
plus bus drive level.
avernos has quit [Ping timeout: 246 seconds]
<hno>
lkcl, in the NDA covered material you received early on, was there any specifications on the DRAM controller? It's left out completely in current "User Manual"., and fex guide only refers to "parameters supplied by Allwinner" in everything outside the basic parameters on physical organisation.
Sv has quit [Quit: Sv]
Sv has joined #arm-netbook
<lkcl>
hno: naah. the stuff i got was pretty much exactly what was "leaked" onto the internet at the same time +/- a couple of months
Sv has quit [Client Quit]
<lkcl>
all right - night all. trying to get some earlier nights...
Sv has joined #arm-netbook
Sv has joined #arm-netbook
Sv has quit [Changing host]
Sv has quit [Client Quit]
<libv>
vgrade: framebuffer libs work fine here
<libv>
vgrade: what distribution are you using, where are you running the test from?
Sv has joined #arm-netbook
Sv has quit [Changing host]
Sv has joined #arm-netbook
Omni-Max__ has quit [Ping timeout: 246 seconds]
avernos has joined #arm-netbook
<vgrade>
libv: framebuffer working for me, also x11
<vgrade>
libv: the issue was I followed the new (your) instructions to install the mali libs on a filesystem built using the sunxi hwpack which included the mali libs in /lib
<libv>
ah, not in /usr/lib/
<libv>
with an old copy still sitting around?
<vgrade>
so ended up with two mali lib installs. This should get fixed once the hwpack builders use your new install method
<libv>
so x11 also worked?
<libv>
how?
<vgrade>
when I say worked I was able to boot x with the mali driver
<mnemoc>
i'm blindly copying the .so files into the hwpack
<libv>
ah, so you did not run the test
<libv>
i doubt that it will take much to port it though
<libv>
just get a screen, display and window the X way
<libv>
'just'
pawel5870 has quit [Read error: Operation timed out]
<vgrade>
I ran the test on framebufer drivers no X that worked I saw triangle
<vgrade>
then swapped to X11 drivers
<libv>
ok :)
<vgrade>
booted X ran the test, saw triangle fullscreen
<libv>
?
<libv>
it just overwrote the fb directly?
<vgrade>
yup
<libv>
at the same 480x480 resolution?
<vgrade>
yes
<libv>
in the top left corner, and the rest of the screen was black?
<vgrade>
yes
<libv>
ok
<libv>
X probably overwrote it quickly
<libv>
once you did a few things
revident has quit [Quit: Combustible lemons? Bah, I bring you weaponized asparagus!]
<libv>
well, thanks, that's a nice datapoint
<vgrade>
I'm using the mele hwpack mele-a1000_hwpack_2012.11.10.7z and linaro-precise-alip-20120923-397.tar.gz for the record
cat_x301 has quit [Ping timeout: 256 seconds]
<vgrade>
just booted into Mer as this is where I do most of my work but have to fix the mali xorg driver due to interface changes in 1.13.0