alcides has quit [Quit: > SELECT * FROM users WHERE clue > 0]
<ZaEarl>
WARNING IRQ 23 crazy!
<ZaEarl>
lol
revident has quit [Ping timeout: 244 seconds]
revident has joined #arm-netbook
<lundman>
it amuses me alcides leaves as his select statement is to return only users with clue, so I guess he has none
ccssnet has quit [Read error: Connection reset by peer]
tekzilla has quit [Ping timeout: 260 seconds]
cheng has joined #arm-netbook
tekzilla has joined #arm-netbook
ccssnet has joined #arm-netbook
jinzo has quit [*.net *.split]
DEAT_ has quit [*.net *.split]
blahee has quit [*.net *.split]
zub has quit [*.net *.split]
Triffid_Hunter has quit [*.net *.split]
zub has joined #arm-netbook
DEAT_ has joined #arm-netbook
blahee has joined #arm-netbook
Triffid_Hunter has joined #arm-netbook
t0dbld has quit [Quit: Leaving]
mSquare has joined #arm-netbook
revident has quit [Quit: Sleep!]
avernos has joined #arm-netbook
avernos has quit [Changing host]
avernos has joined #arm-netbook
jquip has joined #arm-netbook
ccssnet has quit [Read error: Connection reset by peer]
drachensun has quit [Quit: Leaving]
<Turl>
kexec to 3.4 works too, awesome
jquip has quit [Ping timeout: 240 seconds]
cheng_ has joined #arm-netbook
cheng has quit [Read error: Connection reset by peer]
jquip has joined #arm-netbook
rellla has joined #arm-netbook
mysteryname has joined #arm-netbook
jquip has quit [Ping timeout: 240 seconds]
pawel5870 has joined #arm-netbook
ZaEarl has quit [Ping timeout: 240 seconds]
pawel5870 has quit [Ping timeout: 260 seconds]
Quarx has joined #arm-netbook
pawel5870 has joined #arm-netbook
CaCtus491 has quit [Remote host closed the connection]
CaCtus491 has joined #arm-netbook
oliv3r has quit [Remote host closed the connection]
oliv3r has joined #arm-netbook
jquip has joined #arm-netbook
<zub>
hno: aaah, I recall seeing something about that; got more info? also, from a practical standpoint, how can I access these... I guess I'd need some "connector-style" uSD
<zub>
anyway, I got my mk802-turned-headless-wifi-router running :)
<lundman>
and fixed nic driver for more smooth throughput
<rm>
it was extremely hot in Android while driving a HDMI display
<zub>
I tried to kill some config options in kernel, but ran into missing dependencies... fixed these, then had a kernel that freezes; but yup, I'll work on slimming it down to remove unneeded garbage
<rm>
lundman, that's mk802, no wemac here :p
<lundman>
huh
<lundman>
i assumed it has some ether
<lundman>
er same
<rm>
it has if you add an USB dongle
<zub>
what's wemac? I remember seeing some such config option in kernel, but it was 3am, so I don't remember much :)
<rm>
it's the A10's built in NIC
<lundman>
ah
<lundman>
yeah the usb ethers ran pretty well
<zub>
rm: so that's the eth0... something partially present in CPU but not connected on the mk802?
<rm>
yeah
<rm>
it's fully present in the CPU, but not connected to anything out of it
<zub>
ok, I'll remove it from config to get rid of the ghost eth0 :)
<andoma>
hi, i boot my allwinner-v3.0-android-v2 kernel with console=ttyS0,115200 to gain console access for login. "problem" is that the kernel outputs all messages twice. seems once from some low level console output and once from the ttyS0 console driver
<andoma>
any clues how to fix this?
<andoma>
also it panics unless i disable some PM code in drivers/tty/serial/8250_sunxi.c
<andoma>
those things might be related. i'll continue to dig but i just wanted to check if anyone knows anything about it
mysteryname has quit [Ping timeout: 248 seconds]
<hno>
andoma, the double output is a known problem. iirc it's due to how allwinner added their low level debug output support in combination with other allwinner hacks needing low level debug output to always be compiled i.
<hno>
not familiar with PM related panics there.
<rz2k>
<rm> it's fully present in the CPU - still needs realtek rtl8201cp :3
<rz2k>
check cubieboard\mele pcb's.
<zub>
what is the origin of git://github.com/amery/linux-allwinner.git? and is there any support/participation from Allwinner? Or is it just based on tarballs they have released over the time?
mysteryname has joined #arm-netbook
<hno>
zub, there is currently no active participation by Allwinner. But it's a long story.
<rm>
lists a number of irrelevant firmware features like "3G support" and "Skype"
<lundman>
i ran skype on mele, whats the big deal
<mnemoc>
i only looked at the pictures and specs table :< ... should have lmgtfy instead :<
<lundman>
got camera working too
<mnemoc>
above 1.2 the A10 is unusable. chine marketing adds up the Hz of the arm core and the gpu, that's how the come up with the 1.5
<mnemoc>
chinese*
<mnemoc>
thing becomes funny with multicore chips
<jelly>
oh, just like my vSphere cluster runs on 160GHz
<mnemoc>
:)
<oliv3r>
Time to work on some more doc wiki-ing :)
<oliv3r>
How long will the A10 be usefull at all? (Maybe assuming the A15 will be quite the same) how much of this doc work will be in feign?
<mnemoc>
oliv3r: don't forget the A10/PIO tables :)
<mnemoc>
oliv3r: there are many uses for inexpensive cortex-a8. ... not in fancy tablets or laptops, but in industrial applications
<mnemoc>
the next soc from allwinner will most likely keep everything from the a10 except the gpu and arm parts
<mnemoc>
if you look at the datasheet of it's predecessor, the f20, they are very similar
<mnemoc>
same comparing sun3i and sun4i drivers
<oliv3r>
i figure, but i don't wanna spend all this time wikifying it, just to have nobody actually look at it :
<mnemoc>
oliv3r: we need to rewrite almost every single piece of code. believe me, we will look at it
<oliv3r>
good good
<oliv3r>
is there a 'todo/done' page (thought of) allready?
cheng_ has left #arm-netbook ["Leaving"]
<mnemoc>
there is so much to do that is hard to make a list :|
<oliv3r>
well, there's two seperate things right? if not 3, u-boot, nand-boot and linux-kernel :)
<mnemoc>
the first two are the same thing
<mnemoc>
nand comes with a proprietary bootloader (boo0/boot1) which eventually we will replace
<mysteryname>
I can host a redmine for the A10 to create a todo list
<mnemoc>
thanks, but both have issue trackers on github already
<mysteryname>
Ah true
<rm>
mnemoc, what about merging next_mali
<rm>
will that be done and when?
<mnemoc>
i can't before there are working libs for linux and android...
<mnemoc>
assuming, obviusly, it works at least as "good" as the current
avernos has quit [Quit: Leaving]
<mnemoc>
rm: it's mostly a decision that techn_ and rz2k will do, not me
<mnemoc>
s/do/take/
<ibot>
mnemoc meant: rm: it's mostly a decision that techn_ and rz2k will take, not me
<oliv3r>
do we know of any hardware using the A10 PWM?
<mnemoc>
afaik the axp209 (included in all devices) was specially designed for working with A1X, so I would say... all of them?
<mnemoc>
unless i'm misunderstanding the PWM term
<RaYmAn>
you're thinking of PMU :P
<RaYmAn>
PWM is Pulse-Width modulation usually..
<mnemoc>
ah, ok
<mnemoc>
neat feature
<mnemoc>
the cubieboard has a pin called pwm-bl on the lcd segment, and the a13 olinuxino has one called pwm0 also in the lcd connector
<RaYmAn>
it's usually used for controlling backlight, yes
<RaYmAn>
or servos
mysteryname has quit [Ping timeout: 245 seconds]
<oliv3r>
they probably control brightness of the backlight using the PWM then
<oliv3r>
what is a good english technical term for 'does nothing, no effect' etc?
<mnemoc>
NOP?
<mnemoc>
"No Operation"
<oliv3r>
yes! that is good
<oliv3r>
I prefer noop and the full string :) but yes
<mnemoc>
NOP is assembly :p
<lundman>
best opcode ever!
plan_b has joined #arm-netbook
mSquare has quit [Ping timeout: 240 seconds]
Almamuetya has joined #arm-netbook
avernos has joined #arm-netbook
avernos has quit [Changing host]
avernos has joined #arm-netbook
<oliv3r>
this one time, at computercamp, I designed an entire Exaflop supercomputer chip. it was super small and very power efficient. only catch was, you had to do everything with one op-code. noop.
avernos has quit [Read error: No route to host]
tuliom has joined #arm-netbook
<oliv3r>
Help! :(
<oliv3r>
The datasheet mentions that each PWM is a dedicated internal 16bit up counter. The value to which it needs to count is stored in the PWM_CH0_PERIOD register. The PWM Channel0 period register is actually split up into 4 sections however 8bits for the active cycles (no idea what is ment there) which is odd, as it should be 16 bits? Then 8 unused bits. Followed by 8bits for something called 'entire cycles'. also again followed by 8 unused bits. Which is
<oliv3r>
also it is mentioned that the pulse-width depends on the period stored in the register at [0:15], so either there's a bug in the datasheet (likly) or I don't understand how a 16bit counter works (even more likly)
avernos has joined #arm-netbook
avernos has quit [Changing host]
avernos has joined #arm-netbook
avernos has quit [Remote host closed the connection]
mysteryname has joined #arm-netbook
luka has quit [Quit: Changing server]
luka has joined #arm-netbook
luka has quit [Changing host]
luka has joined #arm-netbook
avernos has joined #arm-netbook
avernos has quit [Changing host]
avernos has joined #arm-netbook
<oliv3r>
don't go all silent on me :(
avernos has quit [Read error: Connection reset by peer]
<oliv3r>
mnemoc: I may need a page rename again :)
<oliv3r>
I have created the A10/TIMER section. (Proofreading required! :D) but i'm a little puzzled still. The Timer register is actually split up into two parts, the first 512Bytes is for the Timers etc, while the second half is for the pwm
<oliv3r>
i've split them up as such in the register document, as it makes sense. what is the best way programatically, I don't know. E.g. reading/writing 4k pages, aligning at 1k for better reads, or using it as it is, at the 512Bytes marker.
mysteryname has quit [Ping timeout: 245 seconds]
drachensun has joined #arm-netbook
plan_b has quit [Quit: plan_b]
rvalles has quit [Read error: Connection reset by peer]
rvalles has joined #arm-netbook
<mnemoc>
oliv3r: what do I rename?
<oliv3r>
if you agree, A10/TIMER -> A10/PWM
<mnemoc>
done
<mnemoc>
but there are links from the register page
<oliv3r>
yeah, I allready edited that page to point to them :)
<oliv3r>
but split timer/pwm or not?
<oliv3r>
do I*
<oliv3r>
the rename depends on that :)
<mnemoc>
sure
<oliv3r>
so anybody got an idea on the PWM thing i mentioned above? :)
<mnemoc>
not me
<RaYmAn>
I'd imagine it should be testable?
<oliv3r>
I don't have any A10 hardware :)
<RaYmAn>
ah
<oliv3r>
and there's only reference to the PWM in sunxi-4i.c i think in drivers/rtc on an old branch
sspiff has quit [Remote host closed the connection]
Mazon has quit [Read error: Connection reset by peer]
Mazon has joined #arm-netbook
gimli has joined #arm-netbook
pawel5870 has quit [Remote host closed the connection]
jquip has joined #arm-netbook
gimli has quit [Remote host closed the connection]
gimli has joined #arm-netbook
<jquip>
Hallo there gentlemen :) ... A Question: I received an h6 netbook, which had a chainloaded uboot( linux.ini pointed to uboot, not kernel Image) And yet, it did not boot up when I put the sdcard, it did not boot to android either... That means the boot process has been rerouted right? I dont have a serial logger... I wonder how to debug this...
revident has joined #arm-netbook
drachensun has joined #arm-netbook
RITRedbeard_ has joined #arm-netbook
RITRedbeard has quit [Read error: Connection reset by peer]
plan_b has quit [Quit: Verlassend]
rvalles has quit [Quit: brb]
jquip has quit [Ping timeout: 256 seconds]
ZaEarl has joined #arm-netbook
<RITRedbeard_>
Question: Does MK802/A10 downclock itself it isn't receiving sufficent power from USB or is this image of Puppy Linux just really buggy?
RITRedbeard_ is now known as RITRedbeard
<mnemoc>
by default `ondemand` keeps A10 devices at 60MHz when idling
<drachensun>
hno: you had a hex dump from the AXP the other day, how did you get that? I'm trying i2cdump but I keep getting device busy.
<drachensun>
I can get a nice set of data from the gsensor though
ssvb has quit [Ping timeout: 248 seconds]
lerc has joined #arm-netbook
lerc has quit [Read error: Connection reset by peer]
lerc_ has quit [Ping timeout: 255 seconds]
<jelly>
well, ondemand picks the lowest available freq; it doesn't mean that's the optimal for any specific hardware setup
<mnemoc>
traeak: not really.... pretty busy with $work$ at the moment :<
<mnemoc>
but I have xubuntu running there for the sake of it... and recompiled the kernel and tools a couple of times
<mnemoc>
nothing exciting
<traeak>
sure, well it's probably simpler than the mele i would gather...no wireless
<traeak>
anyawys
<traeak>
work is a bitch (and so are kids!)
<mnemoc>
:)
<mnemoc>
I use my laptop to share "internet" to an ethernet hub on my desk
<mnemoc>
so i prefer wired in general.... wifi setup kinds of sucks
<traeak>
at work here everything is wired
<traeak>
well laptop isn't it, i carry it around and plug stuff into it (usb)
<traeak>
at home everythign is wireless
* mnemoc
have both worlds merged in the same room
<mnemoc>
half office, half bedroom
<mnemoc>
sucks too :-\
<traeak>
ahh
<traeak>
one house we looked at directly adjoining the bedroom was an elevated room almost hte same size
<traeak>
would have been awesome for a family without little kids (older or teenagers)
<traeak>
i prefer an extra open area where the bedrooms are...place to do stuff in the lateevening/early morning
* mnemoc
too
<Turl>
mnemoc: patch sent, got git send-email set up now :)
<mnemoc>
wiki wiki :)
<Turl>
it's too person-dependant a config :P you need an smtp server
<mnemoc>
:(
<Turl>
mnemoc: if you give me an URL I can write a couple of lines on how to go about getting it running anyways
<mnemoc>
/Sending_Patches ?
<Turl>
ok :)
<mnemoc>
OT: what does people use (on linux) to make models for custom cases?
<cat1>
techn_: looking into include/linux/drv_display_sun4i.h at __disp_init_mode_t: what is SCREEN0/1 there? physical head? area in fb? I am trying to make some sense from comments, but fail, please help :)
captainigloo has quit [Ping timeout: 255 seconds]
<techn_>
cat1: SCREEN0/1 loops are support for two heads and one framebuffer
<cat1>
techn_: so it is simply area in video memory?
<techn_>
SCREEN0/1 and "sel" defines different heads
<cat1>
techn_: still do not get it. this is how i see it:
<cat1>
0 -- single head, screen0
<cat1>
1 -- single head, screen1
<cat1>
2 -- dual head, different screens (fb0, fb1)
<cat1>
3 -- dual head, cloned screen (fb0 up and fb1 down)
<cat1>
techn_: i mean that i can understand this that in case of dual head it would imply two screens seen on sepearate head, but in case you have this sort of mapping: fb0 --> screen0 or fb0 --> screen1 -- fail to imagine any practical result of it :)
<cat1>
techn_: but it might be so that I somewhat overcomplicate my vision of all this..
<drachensun>
just a heads up, I tried video mode 3 and 4 and there is some bug in the FB array initialization
<drachensun>
it initialzes buffer 0 then tries to initialize buffer -123012312312
<techn_>
mnemoc: has propably knowledge how framebuffer memory is mapped to display.. but it seems really complicated.. part of display chain is display(hdmi/lcd/tv/vga), mali, g2d(on a10), FE(scaler), BE(colorspace converter), layers and framebuffer
<drachensun>
so its corrupting the pointer somewhere
<hno>
drachensun, I used a logic probe on the i2c bus.
<drachensun>
ahhh
<drachensun>
I figured out the kernel power driver is talking to the AXP
<drachensun>
and bumping me out of the way, and I cant boot without it
<techn_>
cat1: + I'm propably even missing some stuff on that chain :p
<Turl>
mnemoc: maybe it's stuck on moderation?
<drachensun>
so I figured it out it does have a GPIO interface like I really wanted in the first place but I can't seem to get good values out of it
<mnemoc>
Turl: nope
<techn_>
cat1: I'll check if I can find how fb memory is mapped.. but first idea is that memory is mapped to layers which are runned through BE and FE, which propably are some hardware accelerators for mali 2d
<Turl>
mnemoc: want me to resend without the alias? :|
<mnemoc>
try to send it to your own alias
<mnemoc>
to see if it works
<Turl>
mnemoc: the sending did take place because I cc'd another email of mine on tuxfamily and it arrived just fine
<mnemoc>
ok, let's wait then
<cat1>
techn_: ok, but what actually confuses me is that they opeare with SCREEN definition in terms of Xorg. but who promised that it is going to be easy, heh..
<hno>
drachensun, kernel should boot without AXP provided it's configured by the bootloader. u-boot don't yet but the allwinner bootloader do.
<techn_>
cat1: so you are testing Xorgs dual head?
<techn_>
cat1: Does it operate/need two framebuffers or just one
<hno>
it's still a mystery how u-boot SPL can work on mele v1.3 without kernel crash at boot.
<drachensun>
hno: it boots it just gets an endless string of kernel null point errors afterward, I think something ELSE in the kernel is hardcoded for it to be there even if removed in config. I'm not sure though, it was so many dumps I couldn't save it and see the original errors.
<cat1>
techn_: nope, i am trying to start writing disp_core module that would provide infra for others, no real testing.
<drachensun>
techn_: Im interested in having that, but like I said my first attempts ran into an init error, do you want the log from it?
<drachensun>
techn_: I'm willing to help test, I've got a tablet with hdmi out and LCD and I would like to have both at the same time
<cat1>
drachensun: curious what tablet you have. i probably need to order something similar as i am too interested in the same stuff :)
<techn_>
drachensun: I'm currently implementing EDID support.. while waiting mali libs to get gles2 for X.. but dual head stuff is on the list ;)
<drachensun>
cat1: Its a generic one, I import this stuff directly, I'm willing to sell it at cost if you are in the US since its for development.
<cat1>
drachensun: nope, i am on the other end of the world, and just a bit up north :)
<hno>
drachensun, did not see any such hardcoding when doing my minimal linux-3.3 branch long ago.
<hno>
how are you booting?
<drachensun>
SD card
<hno>
using which u-boot version?
<drachensun>
with my older version of you sunxi branch
<drachensun>
I'm basically just before the dram changes now
<hno>
then all bets are off with no pmu support in kernel. should not work.
<drachensun>
gotcha
<hno>
as i said it's still a mysery why thar spl versin at all works on mele c1.3 hardware. should crash, and do crash on many other boards.
<hno>
s/c1/v1/
<ibot>
hno meant: as i said it's still a mysery why thar spl versin at all works on mele v1.3 hardware. should crash, and do crash on many other boards.
<gzamboni>
hno meant: mystery
<hno>
a method that should work is to boot u-boot from nand, insert the sdcard and ask u-boot to reboot.
<hno>
yes
<techn_>
drachensun: "p, I tried video mode 3 and 4 and there is some bug in the FB array initialization" whats that? :p
<drachensun>
its more a curiosity now, I figured out my tablet buttons feed through the AXP chip so first I wanted direct i2c access but the driver makes them available
<drachensun>
so now I just have to figure out if I am using it wrong or it needs a patch
<Turl>
hm
<Turl>
where's wemac on 3.4?
<Turl>
nvm found it
<drachensun>
cat1: gotcha, its probably cheaper shipping to buy direct on aliexpress or something but I don't know who sells them in small volumes, I couldn't find them that way last time I looked
<drachensun>
techn_: I tried the dual monitor thing before, mode 3 and 4 and it was crashing. Its specific to those modes because its starts running a loop to initial all the buffers and screws up the pointer after the first one. I didn't dig any farther than that
<techn_>
drachensun: when? :)
captainigloo has joined #arm-netbook
<drachensun>
techn_: looking through my log captures now....
<techn_>
drachensun: I'm asking becouse today came patch which could have some effect to it
TomNL has quit [Ping timeout: 246 seconds]
<drachensun>
techn_: Ok, it was last week sometime, I'll try that since I can't seem to find the log anyway
<drachensun>
techn_: Are they in allwinner-v3.0-android-v2?
<techn_>
drachensun: yes
<drachensun>
hno: So you were saying you weren't sure why it could boot without PMU settings. So does that mean you added PMU configuration along with the DRAM changes?
<techn_>
cat1: yes.. framebuffer memory pointer it goes to colorspace converter first.. dunno how it goes from there..
jquip has joined #arm-netbook
<hno>
drachensun, the current SPL keeps the CPU running at lower speed matching the default voltage levels for both CPU and DRAM.
<hno>
when PMU support have been added which is not far away now then clocks will be trimmed up again to full speed.
<drachensun>
ok, I was hoping I had another lead for why mine isn't booting. I'll probably dig into that next week.
<hno>
mostly focusing on A13 until I get PMU in shape, as that's the board I have a good debugging setup running for.
<hno>
so it's quite possible there is issues for A10 at the moment.
<drachensun>
techn_: That one crashes too at BSP_disp_get_screen_width
<mnemoc>
"greeting, this is me, and this is my homepage *puff* spam link"
<hno>
"Got nothing to say about me really.<br>Hurrey Im here and a member of this community.<br>I really hope I am useful in some way here.<br><br>My web-site: [****** Aktueller Strompreis]"
<hno>
Next time they pick something sensible as link description as well.
<mnemoc>
i only reacted after seeing the "billiger" word in the url
Quarx has quit []
<hno>
Confused. That is a link to a spam in a blog. Ah, the link makes the page public. It's otherwise not reachable (held for moderation).
<mnemoc>
rz2k: thank you for the replies :)
<rz2k>
I've wrote him to post in english, he seems to have problems with writing on it, so he saw a russian name and wrote on russian.
<hno>
so pure SEO tacticts to make search engines find the page.
* hno
prefer russian over silence.
<mnemoc>
rz2k: but if everyone starts writting in their own language communication would be far worse than if bad english is used
<mnemoc>
hno: true
<rz2k>
yeah
<rz2k>
good night.
<mnemoc>
good night rz2k
Almamuetya has quit [Ping timeout: 245 seconds]
<cat1>
maybe i have asked already, but do we have fix for this: