<mnemoc>
the PCB board shows the 4-pin uart header in the right-bottom
<FergusL>
and who is tom ? :) tom == cnxsoft ?
<mnemoc>
tom = hipboi = cubieboard
<FergusL>
oh ! sure ! okay
<mnemoc>
the cubiebox will expose all spare pins, the mele doesn't
<FergusL>
the white header ? I noted it also have a lot of unpopulated pads all over the board !
deasy has joined #linux-sunxi
<mnemoc>
cnxsoft = jean-luc aufranc
<mnemoc>
FergusL: order one and document it in the wiki ;-)
<mnemoc>
we all want to know what's on those pads ;-)
<mnemoc>
as soon as devs start to get a20 and a31 boards with uart (olimex is finally sending a10s olinuxinos to devs) we can start integrating/cleaning/improving the sun[67]i code we got
<mnemoc>
due to mali and sata a20 is obviusly more interesting, but a31 is nice too
<FergusL>
(is cnxsoft french or french-canadian by chance ? if you happen to know it, Paris here)
<FergusL>
I see, thanks for the explanation
<mnemoc>
G+ says he studied in bristol and lives in thailand :p
<FergusL>
I've just read the G+ too hehe :p
<mnemoc>
I would bet for cnxsoft been french-french, while charbax is french-swiss iirc
<deasy>
salut les français
<FergusL>
salut !
* deasy
pointe grevaillot ha merde il semble belge
tinti has joined #linux-sunxi
<mripard_>
deasy: pire, picard.
<FergusL>
were frequency infos about A20 released ?
<FergusL>
mais c'est quoi tous ces français ! hah
<mnemoc>
FergusL: see the import/lichee-3.3/a20-dev branch in amery's github
<mnemoc>
FergusL: you'll find the cpufreq tables in mach-sun6i
<mnemoc>
note that code is not maintained, it's only a reference
<deasy>
FergusL: non non belgique ici
<FergusL>
pardon, je devrais dire francophone ouais
<FergusL>
thanks mnemoc
<mripard_>
deasy: FergusL: do you happen to know who wrote the article on linuxfr this morning about linux-sunxi?
<mnemoc>
another french!
<FergusL>
no, the user name doesn't ring a bell
<deasy>
cubie soon mine...MINE and i dominate the world with my blog on it...muahahahaha
<FergusL>
I need to test puredata on a cubie
_BJFreeman has joined #linux-sunxi
_BJFreeman is now known as BJfreeman
mturquette has joined #linux-sunxi
mturquette has quit [Changing host]
mturquette has joined #linux-sunxi
sanka has quit [Quit: Leaving]
ganbold__ has quit [Remote host closed the connection]
<drachensun>
So I've read more about the EGON bootloader and the header and all
<drachensun>
and since the A31 BROM is blowing right by my SPL without even the courtesy of crashing, I'm guessing thats what I need to fix for the sun6i
<drachensun>
anyone know how I can extract the header from boot0 or boot1 on the fast to see if maybe the magic number has changed or anything?
<drachensun>
on the flash I mean
<mnemoc>
hno extracted boot0/boot1 from the livesuit images
<mnemoc>
as they live in the first *raw* block of nand, they aren't reachable using /dev/nand
<drachensun>
gotcha
<drachensun>
Mele hasn't released the firmware image yet, but I think there are some for the ainol A31 tablets
<mnemoc>
a u-boot or fel app knowing to talk sunxi dma/nand would be able to extract them
<mnemoc>
but /dev/mem is easier.... don't know if allwinner's 4.2 comes with it enabled
<drachensun>
its in there
<drachensun>
I was looking at that dram settings extractor
<mnemoc>
then adjust a10meminfo based on mach-sun6i data and read the pre-initialized dram registry from android
<mnemoc>
iirc the base address is different than on sun[3457]i
<drachensun>
ok, so is the whole ram accessible in /dev/mem or is it just findable because during boot it was loaded in?
<drachensun>
I meant to say
<drachensun>
the whole nand accessible
<mnemoc>
you read the dram registry within the soc, not the boot0 header
<drachensun>
duh, that makes sense
<mnemoc>
to read the raw nand you have to use dma and use some randomization stuff, not directly from a memory address
<mnemoc>
so reading the dramc is much easier
<mnemoc>
don't know if the mtd driver allows us raw access
wingrime has quit [Ping timeout: 260 seconds]
shineworld has quit [Quit: Leaving]
vinifm has joined #linux-sunxi
vinifm has quit [Ping timeout: 245 seconds]
ssvb has joined #linux-sunxi
vinifm has joined #linux-sunxi
<Turl>
mripard_: ping
n01 is now known as n01|afk
<mripard_>
Turl: pong
<Turl>
mripard_: hi
<Turl>
mripard_: I was going to ask you something but I forgot already :/
eebrah|away is now known as eebrah
<mripard_>
Turl: why are you all doing this to me :)
mdfe has joined #linux-sunxi
<mripard_>
oliv3r: did it yesterday already :)
<Turl>
mripard_: haha
<Turl>
mripard_: ah, now I recall :)
<Turl>
mripard_: did you get any info regarding M1?
<mripard_>
M1?
<Turl>
mripard_: the shady undocumented divisor I found :)
<mripard_>
aaaah
<mripard_>
no, I totally forgot to ask
<mripard_>
sorry :(
<Turl>
np
deasy has quit [Remote host closed the connection]
<oliv3r>
lol
<oliv3r>
i still forgot
<oliv3r>
but! i should have a patch for the BSP ready
<oliv3r>
gotta test it a bit more
<oliv3r>
Turl: where do you keep your initramfs? :)
<Turl>
oliv3r: on my spinning rust
<oliv3r>
ssvb: high level wrapper? i am not qualified to say. vpdau is becomine the 'community standard' but that's only because nvidia pushed it at the start, and it worked. I think vaapi was more general and for 'all', if that makes sense
<oliv3r>
personally, i'd hope that openmax becomes more wide spread; though it does look like an over engineerd and rather complex approach
<Turl>
vdpau is nvidia only though, isn't it?
<Turl>
there's vaapi wrappers to use vdpau and whatever ati engineered
<ssvb>
"These video streams may be combined (composited) with bitmap content, to implement OSDs and other application user interfaces." - that's what can be used for subtitles
<ssvb>
vdpau is surely supported by more video players than the direct use cedarx libve ;)
<rellla>
ssvb: have you taken a look into empat0's xbmca10 yet?
<ssvb>
rellla: as a user, I tried XBMC but did not like it in general (mplayer FTW :) )
<rellla>
i mean, if you tried to understand how he implemented cedarx, not xbmc in general :)
<ssvb>
not yet, but cedarx libve seems to have a relatively simple api
<rellla>
as i'm a noob i try to understand it. i read the libve-user-guide, and if i understand correctly, then you can use the libve_*** methods, and that should be all...
<rellla>
vinifm: call me a "try to understand code be reading and logical thinking" ;)
<rellla>
no developer. no C. most of what is written in this channel i don't understand at all :p
<rellla>
and willswang is doing it the second way. can anyone point out me the main difference?
simosx has joined #linux-sunxi
simosx has quit [Changing host]
simosx has joined #linux-sunxi
<Turl>
mturquette: hi, are you around?
deasy has joined #linux-sunxi
<ssvb>
rellla: I'm also a hardware accelerated video decoding noob :)
<rellla>
i don't want to discuss, who is there greater noob ;)
<ssvb>
rellla: still what I mean is that libve ducumentation is the primary source of information, willswang's cedarx vlc and empat0's xbmc code are the secondary source
<vinifm>
Turl, gpio-sunxi.c replace sun4i-gpio.c?
<rellla>
so i think, willswang or huceke, too, did it more "correct" and are using the libve as described in the user guide.
<mnemoc>
vinifm: yes
<mnemoc>
sun4i-gpio.c should have been removed
<vinifm>
how i use it in userspace?
<rellla>
so there could be a issue of the maybe faulty reset which results in the problems, hramrach_ mentioned.
<mnemoc>
vinifm: standard /sys/class/gpio interface
<ssvb>
rellla: yes, I still need to reproduce and try to debug this xmbc/libhybris issue (I was trying to compile xbmc for gentoo, but not very successfully so far)
<rellla>
let's see if there's something happening the next days. i think, empat0's xbmca10 implementation should be rewritten to use the libve like described
<vinifm>
so can i use: open("/sys/class/gpio",FLAG);
<rellla>
i wasn't able to try gimli's xbmc yet. i'm not a friend of this whole buildroot-thing yet.
<mnemoc>
vinifm: basically you'll need to open each /sys/class/gpio/gpioN/value
<mnemoc>
don't know why there is not a c-friendly interface for gpiolib
<vinifm>
open each one :0
<mnemoc>
yes, annoying as hell
<mnemoc>
but I believe epoll should do it's magic after that
<mnemoc>
libev or similar would help
n01 has joined #linux-sunxi
<n01>
oliv3r: can u give me an update about DT support in the next 3.10? It's not clear to me if I have to use sunxi-...or sun4i-...
<n01>
I'm preparing the final patch
<mripard_>
have you tested it on 3.10?
<n01>
compatible = "allwinner,sun4i-wdt"; got it
<n01>
BTW, at the end I'm not supporting reboot code in wdt driver and IRQ managing ... it came out a very slim driver actually
<n01>
at the moment I prefer this way
<n01>
eventually I'll change it in the future (when I have a clearer idea on how to do it)
n01 has quit [Quit: leaving]
<mnemoc>
anyone from the us here? does t-mobile block or punishes voip, ssh, imap, .... or any protocol in their "unlimited" plans?
<mnemoc>
i've seen they block desktop-ish user-agents, but that's a browser-only thing
<drachensun>
I've got tmobile, I haven't seen them block ssh or imap, though I rarely even exceed my '4G' soft cap
<drachensun>
I haven't tried voip
<mnemoc>
any tethering issues?
<mnemoc>
drachensun: or not doing that?
<drachensun>
I've only used that a few times
<drachensun>
but it worked fine
<mnemoc>
cool. thanks :)
jabjoe1 has joined #linux-sunxi
<drachensun>
no problem, fair warning, their customer service is totally incompetent, took 2-3 days to actually get my number transferred and once a month I get text saying 'Your bill was auto paid" followed by several "your service has been cut off for non payment" texts but thats really more annoying than anything
<jabjoe1>
libv: You there? Not really getting any where with booting a kernel I've built on my A13B tablet. I've got uboot I've built booting, but still not the kernel. The 3.4 image I can boot has a /proc/config.gz, but using that config to build the 3.4 sunxi branch still doesn't build me a kernel I can boot. :-(
<mnemoc>
drachensun: ouch
<jabjoe1>
libv: I've unclipped it the case, and taken some pictures. I can't see where to get serial onto. BenD is back soon, so I might ask him when he is. BenB keeps talking about using the screen i2c.
<libv>
jabjoe1: where are the pictures?
<drachensun>
one other thing, at&t is the only one whose network will support the higher speed data if you have one of the those MTK chinese type phones, Tmobile will work but you will be stuck at 2G speed
<drachensun>
but At&t is way more expensive of course
<libv>
jabjoe1: screen i2c???
* mnemoc
wonders if the bands supported by galaxy nexus sold in europe is different than those sold in the us
<libv>
ah, screen as in display
<jabjoe1>
libv: yer
<libv>
jabjoe1: uart usually is available directly
eebrah is now known as eebrah|away
rellla has quit [Remote host closed the connection]
<libv>
jabjoe1: how do you know that your u-boot is booting?
<jabjoe1>
libv: It's the one I've put onto the mico sdcard.
<jabjoe1>
On to the magic region of nothingness.
<libv>
again, how do you know that _your_ uboot is booting?
<libv>
just because it stops doesn'T mean that it is booting
<jabjoe1>
libv: Well if I put nothing into that region, it doesn't boot. If I put the wrong uboot into that region, it doesn't boot.
<drachensun>
mnemoc: I think you are fine, I just checked wikipedia and they say some software was the only difference. My galaxy nexus runs great on T Mobile
<drachensun>
so I have confirmed the sun6i boot0 header is different
<libv>
jabjoe1: worse case scenario is that ben solders directly to the legs of the chip
<Turl>
mnemoc: they're the 'uncarrier' now, I'd be surprised if they did
<libv>
jabjoe1: did you try looking at the other side of the board?
<jabjoe1>
libv: Not yet, was hoping to have to take it apart as little as possible.
<libv>
jabjoe1: reserve a few h with bdooks and that datasheet, and a 3.3V uart, and you should have yourself serial
<libv>
jabjoe1: about booting, again, i am not clear how you are telling that your uboot is booting
<jabjoe1>
libv: cheers, I'll ask Ben when he is back.
<libv>
what does "booting" mean for you?
<jabjoe1>
libv: I'm putting the uboot I built on the mico sdcard with the prebuilt kernel that boots.
<jabjoe1>
libv: If I try the kernel I build, nothing happens.
<libv>
prebuilt being the android one, or from a sunxi bsp?
<jabjoe1>
the sunxi bsp one
<libv>
ooh, success then, no need for a uart :)
<libv>
are you sure that you are using sun5i_defconfig?
<libv>
sun4i_defconfig is for a10
<jabjoe1>
Tried that. The config from the 3.4 image that boots (/proc/config.gz) and a13_nuclear
<jabjoe1>
a13_mid too
<jabjoe1>
If I leave it when "nothing has happened" it does seam to flatten the battery. So I think it might be there in some way, just not touching the screen.
<jabjoe1>
flatten the battery is left that is.
<libv>
where is the kernel that is known good?
* libv
has never used the bsp
<jabjoe1>
libv: "a13_mid_hwpack_3.0" and "a13_mid_hwpack_3.4"
<drachensun>
mnemoc: when you got that A31 release, was there anything else other than the u-boot and kernel? U-boot doesn't have an updated mksunxiboot attached and I am wondering if maybe it is in something else that was released
<mnemoc>
drachensun: buildroot and allwinner tools
<mnemoc>
I handed over the u-boot -dev stuff to u-boot people
<drachensun>
sunxi-bsp is that the buildroot you mean?
<drachensun>
and so I am seeing similarities and where they match up to the new A31 dram header
<drachensun>
I was just hoping to avoid all that crawling :) oh well
<drachensun>
mnemoc if you are getting a US cell plan are you going to be in the US? I've got 1 extra A20 tablet with a uart on it
<mnemoc>
drachensun: yes, I might move to that side of the big pond soon. need to reset my life
<jabjoe1>
cool, build it all with gcc 4.7 and all works. :-) Now I just have to do some config tweaking and I'm going to be pretty happy. Then to try Bodhi Linux armhf tablet UI...... :-)
<drachensun>
I can ship it to you elsewhere as well but in Europe shipping plus brokerage plus vat will probably cost double what the tablet does
<mnemoc>
drachensun: If I hop there I'll remind you of your offer ;-)
<mnemoc>
was hoping on a20 and a31 cubiebox prototypes from hipboi|cubie soon
<libv>
jabjoe1: really?
<libv>
techn_: nice one
<libv>
seems like we need to add to the wiki
<jabjoe1>
libv: Can't hurt to have a look.
simosx has quit [Quit: Αποχώρησε]
jabjoe1 has left #linux-sunxi [#linux-sunxi]
vinifm has quit [Quit: Saindo]
deasy has joined #linux-sunxi
Kadosch has quit [Read error: Connection reset by peer]
<oliv3r>
drachensun: boot0 and boot1 for both sun6i, sun6i-a31s aswell as sun7i all still read: #define BOOT0_MAGIC "eGON.BT0" #define BOOT1_MAGIC "eGON.BT1"
<oliv3r>
drachensun: so they shouldn't have changed those
<oliv3r>
(yes i'm really slowly backreading tons and responding before reading all :p)
<oliv3r>
mnemoc: drachensun: also, the entire memory code has changed for sun6i; sun7i is the same iirc, so it's not just changing base address, its the entire struct thats different
<oliv3r>
Turl: can you load your initramfs somewhere so I can steel it :p I don't want to make one myself; im' very lazy and want something known to work :p
<Turl>
oliv3r: sure
<Turl>
I just built a buildroot tho :)
tinti has quit [Read error: Connection reset by peer]
<BJfreeman>
just did a pull from the respository, the makefile is referencing files and directories not in the pull
<BJfreeman>
3/linux-sunxi/scripts/Kbuild.include: No such file or directory