notmart has quit [Remote host closed the connection]
notmart has joined #linux-sunxi
notmart has joined #linux-sunxi
bsdfox_ has joined #linux-sunxi
bsdfox_ has quit [Changing host]
bsdfox_ has joined #linux-sunxi
bsdfox has quit [Ping timeout: 256 seconds]
<oliv3r>
hno: spam
hno has quit [Excess Flood]
hno has joined #linux-sunxi
<oliv3r>
excess flood heh :)
rz2k has joined #linux-sunxi
rz2k has quit []
<soul>
Good morning
<oliv3r>
lo
<soul>
Reading this http://linux-sunxi.org/Hyundai_A7HD (I have the A7 without the HD) I see that he did soldering. Without that it's impossible to do it? Is necessary for booting a new system?
<oliv3r>
depends :)
<oliv3r>
i think the soldering is for the UART connection
<oliv3r>
if you have and make it work, you can use a SD -> Uart connection
<soul>
I don't know what UART is but If I don't have a boot menu (or I don't know how to show it)...
<oliv3r>
uart = rs232, e.g. 'serial port'
<oliv3r>
/dev/ttyS0 (com1)
<oliv3r>
you connect your device via serial connection to your PC
<oliv3r>
(or via a rs232 -> usb /dev/ttyUSB0) connector
<oliv3r>
once you have a uart connection, you can see the console, and thus interact with u-boot
hramrach_ has quit [Ping timeout: 240 seconds]
<soul>
Oh, so it's impossible to do it without that
<soul>
Hmm... I don't know more people with my tablet, this is going to be looooooong
<soul>
xD
hramrach_ has joined #linux-sunxi
<rm>
you could have just taken some image, afaik there are some for Pengpod or what was its name
<rm>
and tried maybe it just works on your tablet too
<lkcl>
derethor: no i don't work for allwinner. baseband is *very* hard to do right: high-frequency analog on the same die as digital is very very tricky.
<derethor>
i am totally unable to boot my a13 with vga support... may I ask you guys for help?
<derethor>
I dont have an uart console either
<derethor>
the only response I have is the green led of the a13 olinuxino micro
<derethor>
i tried almost everything... I think the kernel, 3.0, it has the framebuffer enabled
<derethor>
the script.bin, i guess it is correct
<derethor>
and i wrote all uboot parameters I found
<oliv3r>
if you use common sense (fex files can be very wrong by default)
<oliv3r>
you should be able to find one that is extremly similar :)
<oliv3r>
secondly, take your a10-meminfo data and fill in the dram_para section in your script.fex
<soul>
You go very fast! haha Let me try to understand it all
<derethor>
fbmem=32M@0x5a000000 is okey for the fb?
<oliv3r>
derethor: no clue; ssvb ^
<oliv3r>
hno: ping
<ssvb>
derethor: hansg ^
<ssvb>
derethor: more like sunxi_fb_mem_reserve=32 with the new kernels
<derethor>
i will try the 3.4 branch then
<hansg>
derethor, what ssvb said
<ssvb>
derethor: about the failure to initialize framebuffer, unless you have CONFIG_FB_SUNXI_RESERVED_MEM=y in your kernel config, you do not want to allocate large framebuffers
<oliv3r>
hno: a few things; in dramc it was added that we reset the GPS for sun7i? that really can't be right, can it? also why would SUN4I do it?
<ssvb>
derethor: kmalloc can easily fail for large buffers and bring down the system
hansg has quit [Quit: Leaving]
<soul>
Sorry for what I'm going to say and for what I said, but I'm not into this so that's why I don't know a lot of things.
<soul>
[13:24:02] <oliv3r> so now you have something to mail to the ML :p ← What I have to do exactly?
<oliv3r>
soul well for starters, you can mail the .fex (with dram-para filled in from meminfo) and your meminfo dump. If only to have it archived publically :)
<soul>
oliv3r: So I don't have to do anything in that case? I mean, It's done.
<oliv3r>
it's done :)
<soul>
Great, now I'm trying to subscribe
<oliv3r>
calhemp_: well mele a100, mk802s should be well supported :)
<soul>
me*
<hark>
hi, is it possible to use the trustzone with cryptodev on the A20?
<calhemp_>
Oliv3r, ohh thanks, i read months ago that putting the fex files from others a10 devices, can help, i'm not linux expert but if this can help, i'll put my two cents
<calhemp_>
i like to see every week, effort of this channel to make a10 work with linux more friendly
<oliv3r>
calhemp_: fex file is just a configuration file. e.g. enable/disable a driver, configure a driver. it doesn't configure your chip in any way
<oliv3r>
hno: i'm trying to work away more magic values (in dram.c) but now i'm having a little blackout :)
<oliv3r>
hno: I have #define CPU_CFG_CHIP_VER (0x3 << 6)
<oliv3r>
#define CPU_CFG_CHIP_VER (0x3 << 6)
<oliv3r>
crap stupid mouse :S
<oliv3r>
anyway, that define should be fine for reading that specifc register
<oliv3r>
(with a reg_val & CHIP_VER)
<oliv3r>
writing to single bit registers is easy too, (0x1 << bitnr); negating it with &= ~() of course.
<oliv3r>
but is there a 'pretty' way to write a register this way? reg = CHIP_VER | value; won't work obviously; so add new defines for each value of CHIP_VER?
<soul>
You have joined the group linux-sunxi@googlegroups.com.
<soul>
Now I have to send an email attaching the two files to linux-sunxi@googlegroups.com?
<oliv3r>
soul: yep
calhemp_ has quit [Quit: calhemp_]
<soul>
oliv3r: ok! done! :D
<soul>
I'm going to eat, I'll come back for kill your patience!
<oliv3r>
hno: ohh of course, #define WDT_MODE_TIMEOUT(n) ((n) << 3) i remember seeind in the wdt
<oliv3r>
stupid blackout! stupid alergies and immense heat :S
<oliv3r>
i really should get eye-drops for this :S
eebrah is now known as eebrah|away
stekern has quit [Ping timeout: 276 seconds]
stekern has joined #linux-sunxi
<oliv3r>
hno: how many bus widths do you think there are in the dram controller? the DCR config reguster says bus-width is 3 bits wide, which allows for a lot of densities. I assume that the number of DLLCR[x] register is related? x=3 for 8 bit with, x=5 for 16 bit width?
vicenteH has quit [Read error: Operation timed out]
<hno>
The numer of DLLCR is related. width/8+1.
<hno>
supported widths are 8,16,32 from what I remember.
<hno>
DLLCR0 is special.
<hno>
oliv3r, there is u-boot macros for dealing with clearing/setting registerbits.
piyushverma has quit [Quit: Konversation terminated!]
<soul>
I can install the distro with a 1GB microSD?
<soul>
Can I*
<soul>
Maybe is too little
<oliv3r>
hno: yes, but those get passed magic values now :)
<oliv3r>
hno: i was messing about with macro's for those magic values
piyushverma has joined #linux-sunxi
rz2k has joined #linux-sunxi
rz2k has quit []
<hark>
hi, does anybody know about if it is possible to use the crypto acceleration of A10 / A20 / A31 on linux, or if somebody is working on it?
\\Mr_C\\ has joined #linux-sunxi
piyushverma has quit [Read error: Connection reset by peer]
<Turl>
hark: I have some PoC code that I wrote to learn how it works, I hope to write a proper driver when I get a chance http://goo.gl/DVl4R
<rm>
...is there anything that the Allwinner chips *don't* have? :)
<n01>
coffee machine \0/
<rm>
looks like with a proper driver they could probably make sandwiches and do laundry
<Turl>
rm: heh, yeah, they have all the things :)
<Turl>
rm: they even have a tv analog decoder lol
_BJFreeman has joined #linux-sunxi
_BJFreeman is now known as BJfreeman
<hark>
Turl: nice
<vincenzo>
Hello. I'm still looking at the goodix (BADix IMHO) 801 ts driver issue...
<vincenzo>
I have a couple of ideas that I would to share before I start... any advice is welcome.
<vincenzo>
Idea 1) use objdump to disasseble the probe function (for the working driver) and try to see if it does the same as the source driver
<vincenzo>
Idea 2) modify the i2c-core functions in order to printk all the i2c traffic to&from the binary/working driver
<vincenzo>
I admit I have little experience in reverse engineering and i2c comunication... so I dare to ask for yor advice before wasting time in unproductive ways
<rm>
I guess the reason why that phrase is there, is so that the listing appears in search results for "Raspberry Pi"
Soru has quit [Ping timeout: 264 seconds]
<oliv3r>
well its much better then the pi
<hno>
Cubieboard A20 arrived today.
<techn_>
mine is stuck in customs.. again :(
soul has joined #linux-sunxi
<wingrime>
hno: cool
<wingrime>
hno: tom not send me any emails about shipment-trace
<wingrime>
hno: so only wait...
<hno>
And arrived in a box for a change. The A10 board was shipped before there was boxes. And also included serial cable, sata cable and uSD breakout.
<hno>
wingrime, I did not get a tracking number either.
<wingrime>
huo: promicing
ganbold has quit [*.net *.split]
Dreadlish has quit [*.net *.split]
n01|away has quit [*.net *.split]
<wingrime>
huo: russian mail some times very slow
n01|away has joined #linux-sunxi
<oliv3r>
nothing yet
ganbold has joined #linux-sunxi
<hno>
wingrime, are you using a mirror?
<oliv3r>
was hoping it would arrive today; maybe tomorrow
<wingrime>
huo: ??
<hno>
why the u?
<oliv3r>
hNo, not hUo
<wingrime>
hno: sorry
<ssvb>
hno: was it labeled as a gift? any indication of a price?
<oliv3r>
the PCB is the 09 09 version?
<hno>
gift. no customs fee.
Dreadlish has joined #linux-sunxi
<hno>
yes, 2012 09 09 version.
<wingrime>
hno: NXP send to my mail some adv with samples
<wingrime>
hno: but thay not send to russia now
<techn_>
hno: did it have any price on package?
<wingrime>
hno: but it first time thay have some intersing stuff
<wingrime>
hno: High speed switches to 10G
<oliv3r>
wingrime: where in russia do you live?
<techn_>
hno: and how about.. who was sender? company or private person?
<wingrime>
hno: hno: it can be used as quadrature demodulator to SDR
<wingrime>
oliver: yekaterinburg
<Turl>
hno: my A10 one was in a box
<oliv3r>
my a10 was in a foam box
<wingrime>
it realy crap, with this ICs realy simple make Software radio using sound card
<oliv3r>
wingrime: oh wow, that's very close to shenzen!
<wingrime>
oliv3r: omg
<oliv3r>
?
<hno>
My cubieboard A10 was in an envelope with some pieces of foam taped around it.
<wingrime>
oliv3r: russian mail have strange logistics
<oliv3r>
hno: same
<oliv3r>
'foam box'
<wingrime>
oliv3r: bypass moscow
<oliv3r>
yekaterinburg is far far east though isn't it
<hno>
the Cubie A20 package was quite anonymous. Some private person as sender. Can't find the wrapping now so not sure what price was stated.
<wingrime>
oliv3r: asia border here
<techn_>
hno: great.. since finnish customs needs send to be private and value must be under 45e
<techn_>
sender
notmart has quit [Quit: notmart terminated!]
<oliv3r>
wingrime: yeah sounds cool :D
<oliv3r>
45 is lucky, officially it's 25E here
<techn_>
oliv3r: for gifts?
<Turl>
25US$ here :P
<Turl>
and they don't buy the gift marking ;)
<techn_>
here it's 45e for gifts and non gifts 22e :(
<lkcl>
oliv3r: allo :)
<lkcl>
and what the heck is an ar100?
wingrime has quit [Read error: Operation timed out]
<hno>
lkcl, a small OpenRISC core embedded in the A31.
<hno>
lkcl, it's used in deep powersave mode where the ARM cores are completely shut off.
<lkcl>
ack.
<hno>
techn_, found the wrapping envelope. It's sent from what looks like a company PO Box, Gift box ticked, ad content description "PCB", 0.2KG, USD30.
<hno>
oliv3r, no. The A10/A20 have 32-bit memory data bus. The I/O width is the smallest part that can be individually addressed. I..e if you have a quad chip design then the data bus is divided in 4 * 8bit. A dual chip configuration then 2 * 16 bit.
n01 has joined #linux-sunxi
\\Mr_C\\ has joined #linux-sunxi
<oliv3r>
ah ok
<oliv3r>
i'll have to let it settle in; i'm really dislectic with numbers
eebrah is now known as eebrah|away
eebrah has joined #linux-sunxi
eebrah is now known as Guest13732
Guest13732 has quit [Client Quit]
<Turl>
oliv3r: I see you got some more comments from andy :P
<n01>
uhm any idea on how to test rtc alarm?
<techn_>
n01: shutdown device and wait until alarm?
<n01>
uhm I have no backup battery
<oliv3r>
Turl: andy, yeah, it made me go 'seriously dude'
<oliv3r>
Turl: while I know those bracers are extra, I think it helps on readability without hurting by bracers overload
<Turl>
oliv3r: it's arguable :p the dev_ comment is a good catch though
<Turl>
n01: suspend device and wait? :P
<Turl>
I think you cannot suspend on mainline yet though
<Turl>
n01: I suppose it triggers an interrupt on the alarm right? you could chech the count on /proc/interrupts before and after
<n01>
good idea
<oliv3r>
Turl: yes, DRV_NAME was good; allready removed it :) but the sysfs thing gregh mentioned, no clue what to do with that, but mripard been afk all day!
<oliv3r>
or do a little print when the interrupt triggers :)
<Turl>
oliv3r: yeah, I've seen talk about that race a couple of times, let me see if I can find any example on the mail
<derethor>
i am from spain, we are living the same crash here
stekern has quit [Ping timeout: 264 seconds]
stekern has joined #linux-sunxi
<oliv3r>
Turl: will read it later/tomorrow
<oliv3r>
preppin' a patch for hno :)
paulk-desktop has quit [Quit: Ex-Chat]
<oliv3r>
hno: now my branch is broken :S
<oliv3r>
i push to github; but nothing happens and it says 'all up to day'
<oliv3r>
all up to date*
\\Mr_C\\ has quit [Ping timeout: 276 seconds]
nove has quit [Quit: nove]
<oliv3r>
hno: if you sometime have 20 minutes or so, i'd like to talk to you about dram.c cleanup (not entirly related to the patch I just sent you)
<oliv3r>
hno: also, sorry for making it seem there's 6 patches in that mail; haven't figured out yet how to single select a commit for e-mail yet, it does everything FROM there, so i simply didn't send the rest :)
<hno>
oliv3r, regarding git.. is it a new branch? Then you need to name it when pushing.
<oliv3r>
i tried to --force push it into wip/a20 (on mine)
<oliv3r>
it just said 'success' but didn't do anything
<oliv3r>
i did try to push (current) wip/cleanup -> wip/a20 however
<hno>
how?
<oliv3r>
probably wrong
<hno>
the syntax for pushing into another branch is quite explicit.
<oliv3r>
git push --set-upstream oliver wip/a20
<oliv3r>
--force
<hno>
that's a force push of wip/a20.
<oliv3r>
i tried to cherry pick my commit to my a20 branch, but that failed due to missing stuff (i've resend that)
<hno>
doesn't matter which branch you have checked out currently.
<oliv3r>
ahh
<oliv3r>
ok
<oliv3r>
i'll look it up and try again
<oliv3r>
if you could take a glance at the v2 patch i sent last, i think that's missing in your repo
<hno>
ok.
<oliv3r>
(i could be wrong)
<Turl>
to push to a diff name you can push remote mylocalbranch:myremotebranch
<oliv3r>
Turl: testing
<hno>
To push into another branch you do local_branchname:remote_branchname
<hno>
or use suitable patterns for the same in .git/config
<oliv3r>
where do i tell the remote repo, push remote olivergithub mylocal:myremote?
<oliv3r>
kk
<hno>
yes
<hno>
or one word less... git push remote_name local_branch:remote_branch
<oliv3r>
git push remote oliver wip/cleanup:wip/a20 --force
<oliv3r>
fatal: Could not read from remote repository.
<oliv3r>
fatal: 'remote' does not appear to be a git repository
<hno>
but you really should fix it locally and then push.
<oliv3r>
ah without 'remote'
<oliv3r>
duh
<Turl>
oliv3r: remote is your remote name
<Turl>
:P
<oliv3r>
yay
<oliv3r>
my heroes
<mripard>
oliv3r: I don't know what greg meant, and I won't have time to look it up this week presumably, sorry
<hno>
oliv3r, the DLLCR bits is not right. You got it the oppisite way. bus_width field value 3 is 32 bits. If it's not 3 then the code assumes 16 bit bus width. And the dllcr number is the number of DLLs to have active.
<mripard>
oliv3r: it's quite outdated though
<mripard>
I think it doesn't even mention the device tree
<mripard>
let alone pinctrl, CCF, etc.
<Turl>
mripard: I just saw a dt mention
<Turl>
slide 431
<Turl>
430 rather
<mripard>
yeah, my point
<mripard>
*a* dt mention :)
<oliv3r>
430 slides, ouch
<oliv3r>
mripard: tell your collegue to update it
<oliv3r>
:D
<mripard>
it's currently being done :):
<mripard>
we will also switch to a new board, new labs, and so on.
<mripard>
so it takes quite some time
<Turl>
sunxi boards? :P
<mripard>
beaglebone black
<mripard>
I tried to talk them into using an olinuxino or something, but apparently "the mainline support is not there yet"
<mripard>
I don't really know what they mean ;)
<Turl>
BBB's mainline support isn't there either :)
<Turl>
actually I applied for GSoC on mainlining BBB support but they didn't take me :p
<mripard>
yeah, but at least it has a working DT kernel
<mripard>
oh ?
<mripard>
too bad.
<Turl>
yeah, but not so bad either
<mripard>
that means you will have more time to work on sunxi \o/
<soul>
oliv3r: No, after sending the files to the mailing list, I did nothing (I have to know what I have to do now)
<hno>
plus what I said about DLLCR above.
<mripard>
oliv3r: ssvb was on fire in the comments :)
<oliv3r>
hno: oh yeah it needs some overlooking; but generally speaking should be 'ok'
<oliv3r>
mripard: lol aye
<oliv3r>
hno: the DLLCR thing i did hours before :p
<mripard>
Turl: could you send your Acked-by if you find some time on the A10s-olinuxino?
<Turl>
mripard: I thought a tested-by sufficed? :P but sure
wingrime has quit [Ping timeout: 264 seconds]
<mripard>
Turl: well, you reviewed it and had some comments :)
<mripard>
Tested means it works, not that it's clean :)
<oliv3r>
hno @l383; oh shit good spot; yes you are correct, i'll double check all those statements, I forgot to go back to that one specifically, as I later started to define the values to and 'fixed' most of those possible errors
<Turl>
mripard: ok, I'll send it in a minute then
<Turl>
mripard: I don't usually test stuff if it looks uber ugly :P