<speakman>
oliv3r: Usually on MCU's there is zeners which eats anything above it's voltage. That's why they're 5V tolerant. I wonder if there is a simular circuit on A20?
<rellla>
but i'm sure he will get stucked, as there are the most vdpau functions not implemented yet .
<rellla>
jemk ^^ ;)
<juanfont>
... yet :)
<jemk>
rellla: ;)
<rellla>
jemk: i got mplayer work and tried to run VDR via libvdpau. theoretically i should see video - but i first have to solve my channel list issue.
<rellla>
osd could not be displayed because the output_surface or bitmap_surface things aren't implemented.
<juanfont>
i see
<rellla>
is this hard work to do that? i would help but don't know where to go for it...
<jemk>
i'm thinking about how to do all that output stuff, but for full vdpau compatibility we wither have to give up zero copy or figure out a tricky and hard way to
<tgaz>
speakman: about pin mapping. i think that's controlled by the fex script, the gpio_para section.
<rellla>
jemk: getting VDR work on a10/a20 would help to extend the community as it would be the 2nd native arm client besides RPi - for non-interlaced sources atm
<oliv3r>
yeah while I don't like VDR too much (used it for 3 years) it has a rather big following
<oliv3r>
so having support for vdr is a thing
<juanfont>
but, correct me if i'm wrong, this vdpau way has more potential than using cedax, isn't it?
<rellla>
juanfont: vdpau uses cedarx
<juanfont>
yeah, i know.
<rellla>
it's just...
<jemk>
rellla: i have a branch with pretty good working output, but it is terrible slow (not more then 480p) due to many copies
<rellla>
oss
<jemk>
rellla: vdpau uses the cedar hardware, not the cedarx libs
<jemk>
juanfont: ^
<juanfont>
but a working vdpau backend it's more general... vlc, xbmc, mplayer could use it
<oliv3r>
juanfont: vdpau for sunxi is in its extremly early stages and isn't really yet ready for consumers
<oliv3r>
for those who'd like to live on the edge; sure, why not ;)
vicenteH has joined #linux-sunxi
<juanfont>
although it's working nicely... with some samples :)
<juanfont>
so... vdr is broken? (apologies for my ignorance, i've only received a10 and a20 dev boards a few days ago)
<rellla>
jemk: sure. that's what i meant. cedar-engine
<rellla>
oliv3r: don't we all live on the edge here?
<oliv3r>
rellla: absolutly; but sometimes people think it's all done and good :)
<rellla>
because some people don't read the topic: /Allwinner/sunxi development discussion
* rellla
is hungry for jemk's terrible slow branch ;)
<juanfont>
living on the edge is good :)
<oliv3r>
Turl: 3.10 is fixed :)
<oliv3r>
Turl: let me run it for 15 minutes and see if i get the weird stallings; but it boots ocmpletly now
Guest92241 has quit [Ping timeout: 240 seconds]
* mnemoc
wonders what's worse. not been able to connect to the outside world from work. or the android mail client :<
<oliv3r>
k9
<oliv3r>
:)
<oliv3r>
though android mail for 4.2 is nice
<oliv3r>
i like it (though i only briefly used it)
<mnemoc>
does it handle threads and inline replies?
<oliv3r>
one issue i had with it was that it doesn't do imap IDLE
<oliv3r>
it might
<oliv3r>
i know the latest version does support threading etc
<mnemoc>
cool. what about battery penalty?
<n01>
mnemoc: but don't you have internet connection or just irc @ work?
<mnemoc>
n01: irc from the phone
<n01>
web?
<mnemoc>
internet connection is heavily restricted and monitored
<n01>
wow
<mnemoc>
that's actually the business of the company.... so it's expectable
Soru_ has joined #linux-sunxi
<n01>
what do you mean? content filtering and access monitoring?
<mnemoc>
everything is monitored and archived
<speakman>
hm. Is "platform_data" still in use? Or is it/will it be superseeded by device tree?
* n01
thinks mnemoc works for NSA
<n01>
speakman: still in use
<mnemoc>
nah. i believe NSA's cares more able quality
<n01>
ahahaha
<speakman>
n01: ok, thanks. Does sunxi kernel contain board-files too?
<n01>
just one
<speakman>
ok?
<juanfont>
mnemoc, can't you even access to stackoverflow?
<mripard>
n01: no, platform_data are not used anymore
<n01>
for very basic stuff
<mripard>
and we don't have any board file
<oliv3r>
ok going to push 3.10 now
<n01>
uat?
<n01>
let me check
<speakman>
Is the "fex" file superseeding the use of boards?
<n01>
well, mach-sunxi/sunxi.c
<n01>
ok, not really board-specific
<speakman>
can't find such file, n01 :/
<n01>
mripard: which kernel?
<n01>
usp, speakman
massi_ has joined #linux-sunxi
<speakman>
n01: 3.4
<oliv3r>
Turl: pinctrl should be mainline now, diff -u says it's identical.
<n01>
speakman: uh, then I dunno ... usually I work on mainline
<oliv3r>
mnemoc: no IDLE support, so pull only, so no penalty :p
<n01>
mripard: instead of platform_data the data field in of_device_id is used?
<oliv3r>
just make an ssh tunnel through the firewall :D
<oliv3r>
and then run everything via the ssh tunnel :)
<mnemoc>
juanfont: I can. it just gets recorded and processed
<mnemoc>
oliv3r: cool
<mnemoc>
oliv3r: so time to install k9
<oliv3r>
mnemoc: i was talking about the android one
<mnemoc>
oliv3r: sure
<oliv3r>
mnemoc: k9 was forked from android one, does do IMAP-IDLE
<oliv3r>
i use tha tone on my 2.3 phone
<mnemoc>
android -> 3G -> No employer spying
<oliv3r>
also true
<oliv3r>
right, now
<oliv3r>
3.10 needs some merge love from reference mnemoc :)
<mnemoc>
no tag and sunxi-3.0 launching first?
<oliv3r>
i thought you updated reference/3.10
<mnemoc>
not yet
<oliv3r>
that needs to be merged into exp./3.10
<oliv3r>
oh ok
<mnemoc>
reference-3.x is always in sync with sunxi-3-x
<oliv3r>
then i'll await your mergingness
<mnemoc>
so you can git diff/log the range and see sunxi-only stuff
<oliv3r>
for now, i'm gonna look at something else and see how viable it is
<oliv3r>
brb
<mnemoc>
oliv3r: so exp/3.10 becomes sunxi-3.10, 3.10.12-r1 and stage/3.10 ?
<oliv3r>
you think it's un experiminetal allready?/
<oliv3r>
Turl: ^
<mnemoc>
well... it's just like mainline, isn't it?
<oliv3r>
yeah
<oliv3r>
patch set shouldn't be huge
<mnemoc>
so it's a normal dev branch
<oliv3r>
though a review might not hurt
<oliv3r>
i'm not that awesome with git etc ;)
<mnemoc>
just not recommended for normal users yet
<mnemoc>
just for brave ones
<oliv3r>
keep exp. for experimental features? or should that be wip
<mnemoc>
wip/ I think
<mnemoc>
topic branches
<oliv3r>
i agree
<oliv3r>
and someone should review the sunxi_defconfig ;)
<oliv3r>
though i guess our resources are thinnening it seems
<oliv3r>
need some fresh blood
<oliv3r>
mnemoc: btw i replied to phillipe, lets hope he keeps working on his patcha nd is not scared off
<mnemoc>
:)
<mnemoc>
we need to tie arokux1 or he will eat all new commers alive :p
<oliv3r>
*slap*
<mnemoc>
arokux1: btw, I totally agree in the top-posting hate
<oliv3r>
aye, absolutly
<oliv3r>
dhl down for maintanance
<mnemoc>
but slaughtering new contributors might be counter productive
<mnemoc>
oliv3r: all morning. hate them
<oliv3r>
bleh
<oliv3r>
i wanna see the CT status
<mripard>
n01: sunxi.c is not a board file.
<mripard>
n01: and it depends
<mripard>
you might need some values from the DT as well
<mripard>
that you would need to lookup yourself
<speakman>
mripard, n01: in my case I have a device connected to the gpio pins, which is driven by a kernel device driver. Where should I configure on which pins the device is attached?
<n01>
I would suggest DT ... but mripard is the guru here :)
<speakman>
it's still the 3.4 sunxi kernel I'm talking about ;)
<n01>
well, in that case I don't think you have DT
Black_Horseman has quit [Remote host closed the connection]
<speakman>
hm. request_irq(29) returned -16.
<n01>
not good :)
<speakman>
hm, sorry request_irq(gpio_to_irq(29)) returned -16
<n01>
split the two to see which is giving you the error
notmart has joined #linux-sunxi
<speakman>
-16 == -EBUSY
mcbrick has quit [Quit: Leaving]
<speakman>
gpio_to_irq(29) returns -22
<mripard>
which is EINVAL
<mripard>
are you sure you have the interrupt function on that pin ?
<speakman>
mripard: not at all, I just took it for granted there was interrupts on most pins :)
eebrah has quit [Ping timeout: 245 seconds]
<mripard>
speakman: nope
<mripard>
there's only 32 pins that have interrupts on all but A31 SoCs so far
<mripard>
A31 has something like 4 * 32 interrupts, or something like that
<speakman>
ok, "32 external PIO interrupt
<speakman>
sources are supported and interrupt mode can be configured by software."
<speakman>
the docs says
<mripard>
but the idea is the same, not all the pins can be used to trigger interrupts
<speakman>
Ah. Pins with "EINTxx" I guess... hm
<speakman>
...of which 22 of them is sharing pins with LCD. Great...
<oliv3r>
ok before getting side tracked again; time to test the ahci controller!
<oliv3r>
no wait, 3.10 on a20 first :S
<speakman>
PI10 to PI19 has EINTxx. I hope they're available on OLinuXino A20...
* mnemoc
anxiously waiting for tagging sunxi-v3.10.12-r1
<mnemoc>
f* DHL!
<oliv3r>
mnemoc: i think turl atleast has to say something as he was involved too
<oliv3r>
i don't think i'm qualified to pronounce 3.10 taggable
<mnemoc>
Turl!! wake up!
<oliv3r>
:p
<oliv3r>
stupid timezones
<mnemoc>
oliv3r: 3.10 is not supposed to replace 3.4 yet anyway. -r1 just means "equivalent to what you get on mainlined, but LTS"
<oliv3r>
i know
<oliv3r>
but it's not really tested besides bootin git :)
<MadSpark>
hi! any idea why ARM926-EJS is listed as ARMv6 on wiki's main page? it is ARMv5, should i change it?
<oliv3r>
MadSpark: hi; which Allwinner soc does that relate to?
* oliv3r
quietly goes on to read the wiki
<MadSpark>
oliv3r, seems to be all boxchip stuff
<oliv3r>
MadSpark: afaik F20 was ARMv6
<oliv3r>
so quite possible that the title is wrong on which arm core is in that soc :)
<MadSpark>
datasheet says ARM926-EJS that is ARMv5TEJ afaik :)
<mripard>
mnemoc: I'm not buying the "LTS" stuff.
eebrah has joined #linux-sunxi
<oliv3r>
MadSpark: ok in that case, it may very well be armv5; i wonder what sunii was
eebrah is now known as Guest16283
<gzamboni>
fdt was introduced in kernel 3.7 ?
<oliv3r>
gzamboni: for arm, possibly, maybe 3.6, was deff. there in 3.8
<oliv3r>
gzamboni: ppc iirc has used it for ages
<oliv3r>
mripard: what's wrong with LTS? because upstream calls it that? or because android 5.0 is likly to use it
<gzamboni>
thanks oliv3r
<mripard>
oliv3r: who's gonna make the support?
<mripard>
and again, Android doesn't use any kernel version
<mripard>
so it's not a valid argument, I'm sorry
<mripard>
the code people are interested into if they are taking our kernel is the sunxi-specific code. And that part won't get any LTS.
<mnemoc>
mripard: in general or for us?
<mnemoc>
mripard: for me it basically means providing an stable kernel version and someone else backporting fixes
<mnemoc>
not fond to live on the edge
shineworld has joined #linux-sunxi
<oliv3r>
hmm, 3.10 on a20 doesn't work as expected it seems
<oliv3r>
exact same kernel on a10 works just fine
<oliv3r>
might be a dtb issue of course
<oliv3r>
well the only real advantage to me imo, is to be able to forward port the crappy drivers from 3.4 to 3.10
<oliv3r>
i'm even thinking going as far as to add script.bin parser to 3.10 so drivers can work either/or
<Turl>
oliv3r: I cannot say w/o testing :p will give a try later today
<oliv3r>
Turl: everything should be pushed now; remember to set an initramfs though ;)
hramrach has quit [Remote host closed the connection]
<oliv3r>
Turl: but the strangest thing, a10 prints earlyprintk + android log; a20 prints only android log and doesn't give me console
<oliv3r>
so uart is off somehow
<oliv3r>
mripard: AND sunxi-3.10 is the androidized bastard child kernel
hramrach has joined #linux-sunxi
mouchon2 has joined #linux-sunxi
<oliv3r>
Turl: pin mappings appear to be the same :(
<oliv3r>
mripard: what decide what pinmappins are added to the dtsi? for example uart0 only has pina for sun7i, where pina and pinb are defined for sun4i
<oliv3r>
i guess it would have to be tested at the least
shineworld has quit [Quit: Sto andando via]
<mnemoc>
oliv3r: 3.10 is to replace 3.4 as stable, while things get mainline quality
<Turl>
oliv3r: on theory all the muxes would be listed
<Turl>
but there's like 4 per uart
<Turl>
and 90% of them are unused
<Turl>
and that's without mixing and matching pins :)
<Turl>
so we list them when there's a user
<oliv3r>
Turl: so what about the olimexino
<oliv3r>
Turl: i think it exposes all uart pins to some pin on the header
<oliv3r>
so in theory, there's a user right there
<Turl>
oliv3r: one uart is shared with the back SD slot
<Turl>
so that one is not muxed I think
<Turl>
then there's two on board which are
<Turl>
and if someone wants to use one on the pins they can just add it :)
<oliv3r>
anyway, the answer is clear :)
<oliv3r>
i don't hae to go on and spend time adding the default pins :)
<oliv3r>
i'm annoyed that earlyprintk works on a10; and the exact same kernel doesn't printk on a20 (only android printk).
<oliv3r>
and when i say exact same kernel, i mean the binary is the same one, only dtb is different
<mnemoc>
specially in a country which prints it's own money
<oliv3r>
just print more and be done with it
<oliv3r>
whiny pussies
<Turl>
mnemoc: and still manages to have a huge debt, somehow
<mnemoc>
if you have to invest half of your budget in bombing and invading other countries....
<mnemoc>
the oddest part is that the congress is forced by their consititution is accept the f* increase of debt limit
<mnemoc>
but the current sissy playing head of the congress doesn't even want to call the vote
<mnemoc>
because of fear to their extremist wing
<mnemoc>
well... /me bread. bbl
<mnemoc>
+needs
CoSM0 has quit [Read error: Connection reset by peer]
<oliv3r>
bread mmm
<mouchon2>
hi
<mouchon2>
oliver3r: thanks for your review
<oliv3r>
mouchon2: noprob philippe ;)
<oliv3r>
mouchon2: belgian or french?
<mouchon2>
juste some question. If i understand you prefere to have one twiX param (X=1..5)
<mouchon2>
belgian
<mouchon2>
and use 0=unused, -1=default value, Y= freq wanted
<mouchon2>
that's your idea ?
<oliv3r>
yeah your name did sound nederlands :)
<mouchon2>
yes my origine come from Gent
<oliv3r>
im' not sure if i'm qualified enough to say, but yeah, twi0 = freq, twi4 = freq
<oliv3r>
oh, so you speak gewoon nederlands :)
<mouchon2>
ja maar niet wel
\\Mr_C\\ has joined #linux-sunxi
<mouchon2>
ik ben franstalige :-)
<oliv3r>
:p
derethor has quit [Quit: Leaving]
ykchavan has joined #linux-sunxi
<oliv3r>
unset values should default to 'normal' freq i guess
<mouchon2>
yes this is currently the case
<mouchon2>
regarding your loop proposal , i can do this but sunxi_twiX_device should be put in an array then
<mouchon2>
regarding lookup validation values, it make sens to do it, but i am reluctant to do it because some time non standard value can be needed. In my case i need for one device 50khz value that it seem is not standard.
<Tsvetan>
guys, allwinner sent me SDK 1.2 for A13 with Android 4.2.2, the kernel though is still 3.0.8 any interest in this? it's again 3GB and I have no place to host but if you think it's useful let me know
<oliv3r>
mouchon2: yeah the loop may not have been the best idea ;p
<oliv3r>
Tsvetan: if you save it somewhere, that should be good enough. I don't think we have any use for it as the kernel is still the same, and that's the only interesting thing. ANd I doubt there's a newer cedarX
jemk has quit [Ping timeout: 264 seconds]
<oliv3r>
Turl: i'll be damned.
<oliv3r>
Turl: i'm using the ahci branch that ianc got to work, the one he pushed. I get no serial console, after bootconsole disabled, i get nothing (a10). I use the exact same sunxi_defconfig. also my own ahci branch, no console. so 3.10 works perfectly, 3.4 works; anything else fails. I'm flabbergasted :)
<oliv3r>
so now, gonna backport ahci to 3.10 to see if i can get that to spew crap
<Turl>
8250 dw on?
<Turl>
is my clk support for dw uart patch applied?
<oliv3r>
all should be in i suppose
<oliv3r>
and i copied over sunxi_defconfig from 3.10
<oliv3r>
but maybe that's one of the reasons
<oliv3r>
so backporting it is for now and pray
<Turl>
also make sure you pass console=ttyS0,115200
shineworld has joined #linux-sunxi
<Turl>
speed is important or it will die when switching to 9600 default
ykchavan has quit [Read error: Operation timed out]
ZetaNeta has joined #linux-sunxi
<ZetaNeta>
Hi
<ZetaNeta>
i am trying to get the battery away from the screen
<ZetaNeta>
screen is dead
<ZetaNeta>
i want to replace it
<ZetaNeta>
and i need to read the stuff which is hidden behind that battery
<ZetaNeta>
battery is sticked to it with some very good glue XD
<ZetaNeta>
i dont wanna damage the battery... atleast yet
awafaa has quit [Ping timeout: 245 seconds]
<ZetaNeta>
Also. oliv3r, the board got words on it when i finaly opened it
<ZetaNeta>
the one i been talking about
<ZetaNeta>
F727H-MAINBOARD-V2.0.0
<ZetaNeta>
TOPWISE-013
<ZetaNeta>
2012-12-01
<mripard>
mnemoc: again, I'm hearing that "<insert random new thing> while things get mainline quality" for like a year now, without *any* patches. All the drivers that have been merged so far have been rewritten.
hansg has joined #linux-sunxi
hansg has quit [Client Quit]
awafaa has joined #linux-sunxi
<mnemoc>
mripard: it's just an stable stepping stone where crap drivers can mature and be TESTED by the masses before mainlining
<mnemoc>
my cubietruck arrived!
<oliv3r>
that fast?
<oliv3r>
wow
<oliv3r>
well De -> De is probably fast
<oliv3r>
dhl.de still down?
<mnemoc>
yes :-\
jemk has joined #linux-sunxi
<ssvb>
mripard: the hardware is relatively short-lived in modern days, mainline now provides "promise for the future" but is not very suitable for practical use yet
<ssvb>
mripard: the LTS based sunxi kernels allow people to use the hardware right now, and also attract more contributors for the mainline work as a side effect
<mnemoc>
having 3.10 as stable, real life used, it will be trivial to find regretions when improving the code to be mainline ready
<mnemoc>
and it will allow a broader overview of a complete dts/dti
Soru_ has quit [Read error: Connection reset by peer]
naobsd has joined #linux-sunxi
<oliv3r>
also, sunxi-3.10 is androdized
<mnemoc>
more real life users
<mnemoc>
and the oportunity to get more attention from aw
<mnemoc>
so instead of (yet again) hack over their per-soc junk on top of the latest android kernel, they can improve ours...
<ssvb>
mripard: having looked at how omap was doing things (wasting efforts to move between kernels and having lots of regressions in the process), I see that sunxi is doing a much better job
JohnDoe_71Rus has quit [Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org]
<oliv3r>
Turl: while you look; could you also push your (pll6/pll6_sata) clks to 3.10 *wink* :) (no rush)
<mnemoc>
"Hint: Due to improvement works shipment tracking is unfortunately only partly available. We appreciate your understanding."
<mripard>
mnemoc: I'm sorry. but speaking about regressions is funny when no one gives a shit. Have you ever only tested it?
<mnemoc>
^---- DHL BS :p
<mripard>
ssvb: the IPs are mostly the same from one generation to another.
vicenteH has quit [Ping timeout: 240 seconds]
<mripard>
and so far, with 1 year of mainlining, I fail to see how it had attracted developpers
<mripard>
and so far, this 3.10 is only keeping them away.
CoSM0 has joined #linux-sunxi
<oliv3r>
how is it keeping them away?
<mnemoc>
mripard: I'm talking about normal users using the tree and at the same time testing (without knowing) the changes done to use common frameworks, DTS, etc
<mripard>
oliv3r: how much time have you spent on it already ?
<oliv3r>
mripard: fair point, about 6-8 hours i guess
<mnemoc>
mripard: not developers, but just random fellows having a random sunxi-based device
<oliv3r>
mripard: but i went that route because mianline refused to print anyting on the uart :p
<oliv3r>
well after earlyprintk :S
<ssvb>
mripard: just look at the number of people in this channel, without a feature complete and working sunxi kernel, there would be just a few guys at best, and you would be working mostly alone on this hobby project of yours
<oliv3r>
sunxi-3.10 to replace sunxi-3.4 sounds 'reaosnable'
<mripard>
mnemoc: again, for all I know, there is 0 regressions. It's the opposite actually. We have a way better EMAC driver, and a way better i2c driver for example.
<ssvb>
mripard: the current non-mainline sunxi kernel also allows to do the reverse engineering work for the mali and cedar hardware without depending on you to deliver something
<mripard>
ssvb: I'm not questionning the purpose of the 3.4 kernel here.
<mnemoc>
mripard: mainline regressions aren't in general sunxi related, any random patch flying around can affect users
<oliv3r>
i think emac and i2c drivers have been cherry-picked back to 3.10 allready :)
<mnemoc>
mripard: in 3.10 we are only affected by ourselves
<mnemoc>
and we can import crap drivers now, making it usable for normal fellows
<mripard>
mnemoc: oh, please. Any patch added to mainline is way more tested than the patches we could merge. You could say exactly the same about any patch merged in 3.4
<mnemoc>
replacing 3.4
<mnemoc>
mripard: of course not
<mripard>
how so ?
<mnemoc>
I can't tell the same about any patch merged in 3.4, because personally I don't have how/where to test the patches in runtime at all
ganbold_ has joined #linux-sunxi
<mnemoc>
but i don't want to turn this into a religious fight. 3.10 is simply trying to give a *complete* set of drivers so we can encourage people to use it. leaving 3.4 only for unification and comparing with aw trees
<mripard>
we have at least a cubie and a cubie2 doing constant build/boot test, over current mainline, and linux-next
<mripard>
but yeah, sure, regressions.
<mnemoc>
cubieboards
<mnemoc>
1 board
<mnemoc>
one test set
<mripard>
yeah, because your tests are... ?
<mnemoc>
as I told you. I don't have your luxury of having an stable place to live at the moment so I can't test.
<mripard>
anyway... think whatever you want. I'll change my mind when I'll actually see some patches from you guys.
<mnemoc>
I'm saying that the stable branch can be tested by OTHER PEOPLE
<mnemoc>
with other boards
<mnemoc>
and other use cases
<mripard>
mnemoc: I'm not saying about you personnaly, I'm speaking in general here
<mnemoc>
in real life
<mripard>
I don't have the luxury to own a board farm either. I don't host the tests. But they're there anyway.
<mripard>
ok, great. I look forward to see when that 3.10 branch will become usable to every user out there.
<mnemoc>
it's a user testable stepping stone for mainline, not a mainline competitor
<rellla>
The eagle has landed!
<oliv3r>
mripard: i'm trying my bestest :(
leowt has joined #linux-sunxi
<mripard>
mnemoc: it's competing at least for resources.
tinti has quit [Ping timeout: 245 seconds]
<mnemoc>
mripard: the backporting is already done by oliv3r and Turl
<mnemoc>
mripard: so no competing for resources anymore
<mripard>
oliv3r: it's not against you, we all are here because we want to, and you're free to work on whatever you want
<mripard>
mnemoc: which is precisely my point.
<oliv3r>
mripard: i'm working on the 3.10 as I can see it being helpfull for joe-user who just wants a working kernel, but moving away from 3.4-fex-ness; crappy drivers on the new basic framework
<oliv3r>
anyway, atm im' even to stupid to reproduce my ianc's work
<mnemoc>
mripard: my point is that now all the crap drivers can be moved to 3.10 and only DTSized initially. giving a fully functional mainline-compatible base where people can test, while devs can polish them and migrated them to common frameworks
<mripard>
like ssvb pointed out, mainly the three of us contributed to mainline. I'm alone now, and yet, everyone expects me to go even faster
<oliv3r>
mripard: yer not alone!
<mripard>
because you know "mainline support is not sufficient", "there's no IP supported there", "What? No DMA?!"
<oliv3r>
mripard: i just need to get this stuff tested to send to the ML
<mnemoc>
mripard: fewer people works on mainline because it doesn't give you a full system. that's what 3.10 tries to solve.
<mnemoc>
mripard: as soon as 3.10 has all the crap drivers DTSized, mainlining will be faster
<oliv3r>
i'm not even thinking about DTSing the crap drivers
<mnemoc>
because more people will be working in the DTS-based sunxi instead of the fex-based sunxi
<mnemoc>
:)
<oliv3r>
i'm actually concidering adding fex_script_parse to 3.10
ZetaNeta has quit [Ping timeout: 248 seconds]
<oliv3r>
so you can use the 3.4 drivers right now, even when it's not moved to dts yet
<mnemoc>
oliv3r: I have some 3.4 code that scans the script.bin seeing what's enabled and what is not
<mnemoc>
the idea was to spawn platform_devices out of it
tinti has joined #linux-sunxi
<oliv3r>
that's a 2nd step then imo :)
<oliv3r>
anyway, less arguing, more helping oliv3r get things working ;)
<mripard>
mnemoc: again, I'm earing that kind of arguments for too long. I'll change my mind when I'll see some patches coming out of this work.
<mnemoc>
=)
<oliv3r>
well i do see some patches for sunxi landing in the 3.4 from time to time, so people do work on things; not helping mainline obviously
<tgaz>
with "mainline" here, are you (guys) referring to the official linux tree?
<mnemoc>
tgaz: yes
<mnemoc>
torvalds tree
<tgaz>
good. :) then i'm not completely off.
<mnemoc>
just a conflict between those who prefer to iteratively refactor working code and those who prefer to reimplement ;-)
<oliv3r>
actually, you both are right :)
<oliv3r>
it does leech developer time away from mainline effort
<oliv3r>
but i also think it's a good idea generally :)
<oliv3r>
but you know what, if AW would have worked with upstream; we'd all be doing cooler things anyway :)
<mnemoc>
but as a result it should get "everything" mainlined faster than using the same resources to reimplement the rest of the drivers, specially considering the lack of proper documentation
<mnemoc>
that's another of my hopes. if we provide a usable androidized 3.10, always might even decide to join us instead of vomiting their stuff on top of android-3.10
<mripard>
yeah right. because emac was reimplementation, and we could have merged pinctrl or clock drivers just by refactoring it...
<mripard>
anyway, I'll just shut up, if everyone's happy with what we have now.
<oliv3r>
most drivers will probably have to be rewritten, anyway; eitherway you go
<mnemoc>
eventually the result will be completely different than aw's code, yes
<oliv3r>
i don't think you can avoid that; i don't think any of the drivers can easily be refactored
<mnemoc>
but iteratively. so regressions can be noticed faster
<mnemoc>
or reimplement, fine
<mnemoc>
but the rest of the crap drivers still provide a functional system
<ssvb>
oliv3r: I don't quite agree about the importance of AW participation :) look at the pandaboard as an example, it has depended too much on the corporate support from TI, now the corporate support is gone and #pandaboard is mostly dead
<mnemoc>
and so joe can test the reimplemented driver in his random device
<oliv3r>
ssvb: true that; still it makes things frustrating :)
<ssvb>
oliv3r: IMHO too much corporate control over the kernel is driving away independent contributors
<oliv3r>
ssvb: like samsung? :)
<mnemoc>
ssvb: pandaboard was made and subsidiced with TI's marketing budget
<mnemoc>
it is a completely different case
<oliv3r>
anyway, we're all trying to do something good here
<oliv3r>
if only we had more developers :)
<mnemoc>
deprecating 3.0 was good. making 3.10 the stable branch and 3.4 the legacy will be even better
<mnemoc>
and the ultimate goal is still mainline
<tgaz>
i'm in the "give mainline sunxi support ASAP" camp... what's missing? (i know this is a hot topic even without today's ... polite flames)
<oliv3r>
tgaz: what's missing a lot :) and mostly due to lack of dev. time
FDCX has quit [Ping timeout: 240 seconds]
<oliv3r>
whoot! i get console agian
<tgaz>
ah, the Linux mainlining effort page seems to say some stuff is merged.
<oliv3r>
no earlyprintk; but console login
<oliv3r>
*happy dance*
lkcl_ has quit [Ping timeout: 245 seconds]
<oliv3r>
tgaz: there's tons done allready
<tgaz>
sweet
<tgaz>
:)
<oliv3r>
Division by zero in kernel.
<mnemoc>
tgaz: most of the "core" is already mainlined
<oliv3r>
Montjoie: the touchpad controller has a temperature sensor embedded
<oliv3r>
Montjoie: we have some test code, dunno what happened with that
<mnemoc>
.oO
<Montjoie>
I will see that
tinti has quit [Ping timeout: 256 seconds]
fredy has joined #linux-sunxi
<oliv3r>
ohh looks like cubieboard does have a GPIO to turn sata power on/off
<mnemoc>
the phy too iirc
<oliv3r>
i thought only mele had one for the phy
<mnemoc>
maybe
fredy has quit [Ping timeout: 245 seconds]
* mnemoc
wants to sleep
<n01>
mnemoc: no oktoberfest?
slapin has joined #linux-sunxi
<slapin>
hi, all!
tinti has joined #linux-sunxi
<n01>
slapin: 0/
<slapin>
Cheers!
<oliv3r>
slapin: omg!
<mnemoc>
n01: nope.
<mnemoc>
slapin: wb!
<slapin>
hno: btw, got jffs2 working using a10 cubie
<oliv3r>
arokux1: i'm going over the cubieboard specs, and it appears, that atleast for CB1/CB2 VBUS0 is connected to PH3 and VBUS1 is connected to PH6. You have those 2 mixed up. I don't think it matters hugely, since you power on both simultaniously anyway
* slapin
needs some drinking buddies to celebrate gf's first emplyment in her entire life!
<slapin>
where is Marex?
<n01>
in #u-boot :)
<Turl>
I'll have a look later oliv3r
<mnemoc>
slapin: prague iirc
<oliv3r>
Turl: i'm looking too
<oliv3r>
slapin: wb; celeberate hacking on sunxi!
steev has quit [Read error: Operation timed out]
awafaa has quit [Ping timeout: 248 seconds]
<slapin>
oliv3r: of course, but I have another great milestone - the girl made 15 km on public transit without me around, for interview, and then, again, for first work day, that is so great! and I knew all that just a hour ago when she called me, she kept her great success a secret!
jeremb_ has quit [Ping timeout: 248 seconds]
<slapin>
I think celebrating on @lunux-sunxi is safer than on #u-boot
<n01>
gratz dude
ojn has quit [Ping timeout: 264 seconds]
* slapin
just got 6 pints of rum to settle things down
<slapin>
so, cheers!
<oliv3r>
slapin: lol awesome; i'm happy for you
<oliv3r>
i need a milestone like too :)
<oliv3r>
arokux1: nvm the spec has typo's you did it right
\\Mr_C\\ has quit [Ping timeout: 240 seconds]
<oliv3r>
mnemoc: hmm EMAC-PWR-EN suggests cubie needs GPIO for power too
<oliv3r>
PH19 on the cubie it would be
<mnemoc>
unfortunatelly they made some changes the schematics after the version published :<
<mnemoc>
so you can't blindly trust what's published
<oliv3r>
well i do see two SOT23 parts right next to the PHY
<mnemoc>
:)
<oliv3r>
and the schematic does call both transistor and the FET SOT23 parts
<oliv3r>
and i've only found very few sot23 parts
<oliv3r>
2 for the EMAC-PWR-EN
<oliv3r>
and 2 for the SATA
<oliv3r>
and they both are close to those compents
akaizen has quit [Ping timeout: 240 seconds]
<mnemoc>
oliv3r: how do you enable thread view in k9?
<oliv3r>
erm
<oliv3r>
somewhere in the options? i don' tknow
<mnemoc>
but you use it, right?
<mnemoc>
I can't even see all my folders. but if you tell it works. I'll keep searching
<slapin>
I think it won't be bad to full cups and drink to avoid bed prison diseases, and never invent them in our head, so lets be healthy and open-minded! Cheers!
<mnemoc>
cheers slapin!
<Turl>
what are we celebrating? :)
<Turl>
I got a bit late :p
<WarheadsSE>
slapin's gf got a job
<WarheadsSE>
apparently _first_
<Turl>
:)
<slapin>
WarheadsSE: absolutely, thanks!
HeHoPMaJIeH has quit [Remote host closed the connection]
<mnemoc>
but some weeks ago I posted a box for drachensun from my office, and instead of getting it delivered in the us it was delivered to some random person in the building I was living at that time....
FDCX has quit [Remote host closed the connection]
<mnemoc>
and today their tracking service was dead until less than 30m ago
<Turl>
lol
<mnemoc>
but otoh my CT was delivered in less than 24h
<Turl>
yeah the speed is not the issue
<Turl>
it's all the other parts :P
* slapin
remembers russian post and gets another drink
* slapin
means speen AND other parts...
<slapin>
*speed
<Montjoie>
arg, neither send a patch without testing the last minor change...
<mnemoc>
slapin: don't you have better ways to celebrate with your gf than just getting drunk?
<Turl>
slapin: wingrime got his CB2 before I did :D
akaizen has quit [Read error: Connection reset by peer]
<slapin>
mnemoc: she's not here, we'll meet on weekend
<slapin>
Turl: I've got one, too, but it seems to be slightly different than A10 one regarding NAND and CedarX so I am still trying to make something interesting out of it.
atiti has quit [Ping timeout: 248 seconds]
lkcl_ has quit [Ping timeout: 240 seconds]
<mnemoc>
slapin: ok. then go ahead and get drunk!
<slapin>
mnemoc: don't tell me...
shineworld has quit [Remote host closed the connection]
* slapin
remembers no obscenity told by him in drunk state, and finds no reason to spoil the party
mouchon2 has quit [Ping timeout: 245 seconds]
<slapin>
mnemoc: bth, do you know how should I deliver package from Gemarny so that it will not be delivered by russian DHL?
<slapin>
mnemoc: as I need to explain this to friend by phone, as he knows about this subject the same as me
<mnemoc>
slapin: don't know. here the national post office forces you to use DHL.... (they are the same)
<mnemoc>
maybe using UPS instead :p
Gerwin_J has quit [Quit: Gerwin_J]
<slapin>
mnemoc: both are wrong as I'm going to deliver devboard and will be subject to commercial customs regulations in case of delivery by DHL or UPS
<drachensun>
I've never gotten DHL to deliver in to Russia
<drachensun>
slapin: are you going to Russia from Germany or to Germany from Russia?
<slapin>
drachensun: not me, my friend is in Munich now
<drachensun>
he is sending?
<slapin>
drachensun: yes
<slapin>
drachensun: russian dhl and ups work only with companies, not with regular citizens, and they do various cheats to deliver packages (asking for nearby company office), and the only category they allow to deliver is so called "Document"
<slapin>
drachensun: which is some abstraction
<slapin>
drachensun: in russia they make business from customs service and don't care much for actual delivery
<drachensun>
yeah, that is my experience to
<drachensun>
they wont deliver straight to a person
fredy has joined #linux-sunxi
<slapin>
drachensun: and by law, the packages delivered to companies are subject to commercial customs service, so no-no in my case
<slapin>
drachensun: packages delivered by post avoid this and have less trouble (except for post errors and post's unpredictable slugishness)
<drachensun>
slapin: so the post office would work, eventually
<drachensun>
I've never tried this
FDCX has joined #linux-sunxi
<drachensun>
we just took Russia off of our deliverable country list after the first problems
<slapin>
drachensun: so life hack is needed - post from Germany not via DHL
<drachensun>
I see
<slapin>
drachensun: I know lots of people do this, well, even Siemens do this occasionally, so there is some unknown secret
<drachensun>
well they have mail forwarders
<drachensun>
before mnemoc and I went through that debacle with DHL I used one, mailbox.de
<drachensun>
they are just slowwwwwwwwwwww
<drachensun>
and expensive
<drachensun>
but they will forward the package on and they had a few different options for delivery to the USA, not all DHL
<slapin>
drachensun: so I think the only way to reveal all secrets is for somebody native speaking German to call DHL office and ask on how to deliver to Russia in a way so that package aboard will be delivered by russian post, not by russian DHL?
<slapin>
I will compensate the amount spent via paypal
<drachensun>
eh, I would send with someone else
vicenteH has joined #linux-sunxi
<drachensun>
DHL will tell you whatever, take that package, forget what you said and happily screw up the delivery
wingrime has quit [Ping timeout: 256 seconds]
<drachensun>
not that another carrier is better in that regard, they are all a bunch of bureaucrats and when it doesn't go perfect they have a mental breakdown
<slapin>
and the level2 life hack is to know how to make DHL-delivered package to be picked-up by russian post EMS service, which is slightly better than regular post as it delivers to the door
<drachensun>
but I think you have to find one that just plans to go the way you need from the start
<slapin>
drachensun: I know people do this via DHL, but people don't want to tell how :(
<drachensun>
ah well if you figure it out, tell me too
<drachensun>
my rep took me out for drinks with his boss, the regional sales director once
<drachensun>
this subject came up
<drachensun>
and neither of them knew of a way around it
<drachensun>
they told me there was once a year when DHL took Russia off the list altogether as a deliverable location
<slapin>
drachensun: DHL Päckchen <- have anybidy tested this?
<drachensun>
because customs was refusing everything from all their customers even proper commercial ones, and they didn't know why
<drachensun>
so they seemed to have some experience with it
<drachensun>
slapin: I haven't, I can't use their 'mail' type services from the US, only express
<slapin>
drachensun: well, I'll ask him to send via DHL's DHL Weltpaket and we'll see, I think.
<slapin>
drachensun: as he's already in post office
<slapin>
drachensun: ah, I think from US USPS service is the best, as all services work, and delivery is 2-3 times faster than from Europe and I never had single failure
<drachensun>
yeah, if we ever need to try again I think we will go that way
Guest16283 has quit [Ping timeout: 240 seconds]
<slapin>
drachensun: and they have services which are picked by EMS locally, which makes them even faster(?)/easier
<wingrime>
jemk: ping
eebrah has joined #linux-sunxi
<slapin>
I rarely get anything from US via USPS longer than 3 weeks, while I get 6-12 weeks for UK and 12-24 weeks from China
eebrah is now known as Guest31284
n01 has quit [Ping timeout: 240 seconds]
<buZz>
slapin: 12 weeks from china? wtf
<buZz>
slapin: which country do you live in?
<slapin>
buZz: Russia
<buZz>
my god
<drachensun>
seriously, 12-24 weeks its nuts, its 4 at the worst here
<buZz>
i am in netherlands and china only takes 3 weeks here
<drachensun>
and we dont share a border with them :-)
<slapin>
buZz: either they deliver packeges by trained cockraches, or russian customs office plus russian post do their job well
<slapin>
buZz: why this doesn't affect US delivery is beyound my understanding
<buZz>
slapin: maybe it would be easier to find a russian close to chinese border, and a chinese close to russian border
* slapin
lost the feeling of forms, please ignore typoes
<buZz>
slapin: and let them walk ~4km and exchange packages
<buZz>
i assume inside-russia packages dont take 10 weeks?
<slapin>
buZz: and deliver via paper plains?
<buZz>
no, just ship it with normal post
<drachensun>
some of my customers in St. Petersburg would just have it sent to Finland
akaizen_ has quit [Ping timeout: 260 seconds]
<drachensun>
but one got really pissed about having to pay their duties and blamed me
<buZz>
my friend lives in belgium and gladly drives to netherlands to post his packages, to avoid any customs delay
<drachensun>
as if I told him to send it to a foreign country for pick up
<slapin>
buZz: if you beleive tracking, the package leaves China after 1-2 weeks, then leaves our reality for several months, not displaying in tracking services, then it magically appears and teleports through several cities at magis speeds (staying no more than week in each) then appear at your post office for you to pick up
<buZz>
slapin: yeah customs dont want you to know they have your packages ;)
awafaa has joined #linux-sunxi
jeremb_ has joined #linux-sunxi
Soru has quit [Read error: Connection reset by peer]
<aexl>
arokux1: i've got mail from Gennie Peng <sales10@mele.cn> confirming that the Mele M5 has a UART interface. she also sent a circuit diagram.
<aexl>
... or rather a cut-out of the uart ...
<jemk>
wingrime: pong
<slapin>
buZz: sent on the middle of July
<buZz>
still not loaded
<buZz>
i am giving up
<buZz>
i'll just believe you, ok?
leowt has quit [Quit: leowt]
<slapin>
buZz: a problem is that it left customs on 2th of august (as I called them)
<wingrime>
jemk: strange cedar x sun2i header have 8 for mpeg selector (others known) but not one chip from aw family not claimed mpeg encoding support
rz2k has joined #linux-sunxi
<wingrime>
jemk: can you try selectors and fingure with register base for 0x8 , and others unknown codes?
montjoie[home] has joined #linux-sunxi
<slapin>
buZz: since 30th of september it is Sorting somewhere near Moscow..., and shows imported on 18th of september
<slapin>
buZz: where it have been for 2 months?
<buZz>
slapin: i bet 'some agency
<buZz>
looking at all YOUR packages
<slapin>
buZz: it was out of customs on 2th of August
Soru has quit [Ping timeout: 240 seconds]
<wingrime>
jemk: I mean we can try selectors and find witch reg base for it
<slapin>
buZz: it was 'imported' on 18th of september
<jemk>
wingrime: i see, i will try if i can find out
<slapin>
and the same damn thing with all packages from China, and it is for everybody here
<buZz>
slapin: yeah you just said this
<wingrime>
jemk: at least must something be for 4,5,6
<jemk>
wingrime: but could be good if you verify it on your a13 to see if it's same
<slapin>
so there is some alternate reality, where there are stores of packages which left customs but were not imported....
<wingrime>
jemk: yes, but I think it same
<wingrime>
jemk: also, do you have any idea how we can use counters?
<slapin>
buZz: Europe too, but China is more mainstream, Europe is for 3D printer kind ans China is for toys
<slapin>
buZz: US os for HP oscilloscopes and digital calibers
<slapin>
s/os/is
<buZz>
i like analog calipers
<slapin>
buZz: they don't have enough precision
Soru_ has joined #linux-sunxi
<buZz>
they have two digits behind the . , enough for me
<slapin>
buZz: not enough for me. I didn't say they are not nice though
<buZz>
i dont need to know the difference between 2.01mm and 2.015mm
<slapin>
buZz: I do
<buZz>
all measurements i do are just for reproducing objects in 3D models , to print later
<slapin>
buZz: this is ok, for this task digital caliber with 0.03 precision is even cheaper than proper analog one
<slapin>
buZz: analog calibers are ok, thay just don't suit my needs
<buZz>
0.03 ?
<buZz>
i am telling you, my analog ones have 0.01 precision
<slapin>
buZz: cheap ones, yes.
<slapin>
buZz: I'm very glad
<buZz>
0.03 is actually -less- precise
<slapin>
buZz: I perfectly understand what I told you and what you told me.
<buZz>
anyway, measurement stuff like this, you should buy Mitutoyo ;)
<slapin>
buZz: $10 caliber is with 0.03 precision
<buZz>
Mitutoyo analog caliper can be 100 usd easily
<slapin>
buZz: $100 is fine by me
<slapin>
buZz: mine had the price of $230 when I bought it
<aexl>
ot much?
<slapin>
I know it might be overpriced, but I needed the tool
<slapin>
everybody told me that for fallic measurement I could buy for $3 normal caliber, but they don't understand, they just don't understand that if damn filter film will have defects I will have to redo a month of work and there is no other person around who'd I beleive he will check it reliably
<oliv3r>
Turl: i haven't looked exactly yet, but should be related to sata/pll6 i pasted the log earlier
Soru_ has quit [Ping timeout: 248 seconds]
Gerwin_J has joined #linux-sunxi
<jemk>
wingrime: looks interesting, note that 0xa00 is used often: http://sprunge.us/VLXa
<Turl>
oliv3r: try rebasing onto the latest sunxi-clk
<montjoie[home]>
I see patch for A20 SMP support, does the current sunxi 3.4 does not handle it ?
<wingrime>
jemk: are you checked others bits?
<wingrime>
jemk: 0x500
<wingrime>
-0x900
<jemk>
wingrime: ranges not listed are never writeable
akaizen_ has joined #linux-sunxi
<wingrime>
jemk: thats understandable
<wingrime>
jemk: but why so bit gap
<wingrime>
*big
Soru_ has joined #linux-sunxi
aexl has quit [Quit: Leaving.]
<jemk>
wingrime: dont know, but verified with reset, so it's pretty sure they are unused. when bits in 0x004 are set i cant read any of the used areas without crash, but the unused work and return zero
akaizen has quit [Ping timeout: 248 seconds]
<wingrime>
jemk: see, 0xa00 - ISP, can encode jpeg and more, but mpeg and jpeg are simular
<oliv3r>
Tue, 01.10.2013 01:38 h
<wingrime>
jemk: so they joined
<oliv3r>
K?ln, Germany !! so close
<wingrime>
jemk: and isp are used in blob with h264 encoder....
panda84kde has quit [Ping timeout: 248 seconds]
<wingrime>
jemk: isp can resize with non-integer ratio
<wingrime>
jemk: thats for direct recoding from camera
popolon has quit [Quit: Quitte]
<wingrime>
jemk: can you add findings to wiki....
geecko_ has joined #linux-sunxi
Gerwin_J has quit [Quit: Gerwin_J]
Gerwin_J has joined #linux-sunxi
FR^2 has quit [Quit: Connection reset by peer]
<wingrime>
jemk: do you played with CACHEREG_WNUM?
n01 has joined #linux-sunxi
<wingrime>
thats can be readed as cache register write num ? may be this is sram writes using magic regs?
<slapin>
drachensun: thanks for link, the package was sent, thanks! now will wait for delivery
<jemk>
wingrime: not yet, i'd like to do more reverse engineering, but i have to add more output functions to vdpau or it won't be of much use.
<wingrime>
jemk: than mpeg4
<wingrime>
jemk: last of interested , not difficult, like h264
<jemk>
wingrime: i don't need more work, but less ;)
<jemk>
or anyone here willing to take some work and help me with output?
<wingrime>
jemk: h264 encoder and ISP, I played with it , I stll not managed to get encoder trace
<wingrime>
jemk: so I only plays with dissassembler,
<wingrime>
jemk: maybe Paulo can help
<wingrime>
jemk: we can currently add only VC1 and MPEG4 to libvdpau
<wingrime>
jemk: other features can't done such way
<wingrime>
jemk: Isp can be interested, as it looks can alpha-mix
<wingrime>
jemk: so, we can use OSD, and subtitles with it
<jemk>
wingrime: that also can be done with g2d (at least on non a13), but i'm totaly stuck on how to do proper output without many memcpy
<wingrime>
jemk: you can help users with speed, adding hw down-scale
<wingrime>
jemk: thats make less pressure to DRAM
<wingrime>
jemk: I not know, but cedar can have own cache
<wingrime>
jemk: some exportabe env variable for hw downscale
<ssvb>
Montjoie: what is needed to get your axp20 thermal sensor patch to work on cubieboard2?
<jemk>
wingrime: that's all no problem, but take an 1080p video shown on an 1080p screen and the useless copies and your dram bandwidth is gone
<jemk>
s/useless/needed but bad/
<wingrime>
jemk: some tables have no 1080p
<wingrime>
ssvb: there termal sensor in a10/a20 internaly
<jemk>
wingrime: yes, then it works there, but vdpau has to work on desktop too
<ssvb>
wingrime: is ax20 an external chip unrelated to a20 soc?
<ssvb>
*axp20
<wingrime>
ssvb: nearly present
<wingrime>
ssvb: axp20x usualy for a10,a20
<ssvb>
wingrime: I mean, is its temperature sensor providing relevant data?
<montjoie[home]>
How do you have the temperature of thermoncouple , does it is the gsensor ?
<ssvb>
in what way are the changes important? the sensor initially showed some increase, but then seemed to settle around 47C
<ssvb>
it's a thermocouple from my multimeter
<ssvb>
and I believe that the core of a20 must be a bit hotter than that
<montjoie[home]>
My AXP209 is at 30°C
<ssvb>
can you try cpuburn?
<montjoie[home]>
later, I try to find some doc about the touchPad that spoked wingrime
<montjoie[home]>
oh I see that BMA250 gsensor also have a thermal sensor:)
mcbrick has joined #linux-sunxi
<wingrime>
montjoie[home]: in a20/a10 manual
<wingrime>
ssvb: useless measue axp209, its far from cpu
<ssvb>
wingrime: yeah, I guess so, we need to support the thermal sensor inside the a20 chip
<montjoie[home]>
It have one ?
Seppoz has quit [Remote host closed the connection]
<ssvb>
wingrime said it has
<jemk>
it has, i tested with uboot
ykchavan has quit [Read error: Connection reset by peer]
ykchavan has joined #linux-sunxi
_whitelogger____ has joined #linux-sunxi
<n01>
anybody booted the 3.12.0-rc2 lately?
hno has quit [*.net *.split]
\\Mr_C\\ has quit [Write error: Connection reset by peer]
_whitelogger___ has quit [*.net *.split]
speakman has quit [*.net *.split]
eagles0513875 has quit [*.net *.split]
<oliv3r>
n01: why?
<n01>
this motherfucker is not booting
<n01>
still dunno why
<oliv3r>
story of my life
<n01>
sunxi-life. It doesn't boot. The end.
hno has joined #linux-sunxi
geecko_ has quit [Read error: Connection reset by peer]
akaizen has quit [Ping timeout: 264 seconds]
eebrah has quit [Read error: Connection reset by peer]
Seppoz has joined #linux-sunxi
akaizen has joined #linux-sunxi
tinti has quit [Ping timeout: 248 seconds]
speakman has joined #linux-sunxi
jemk has quit [Ping timeout: 245 seconds]
pacopad has joined #linux-sunxi
tinti has joined #linux-sunxi
tinti has quit [Max SendQ exceeded]
tinti has joined #linux-sunxi
tinti has quit [Max SendQ exceeded]
tinti has joined #linux-sunxi
akaizen_ has joined #linux-sunxi
rz2k has quit []
tinti has quit [Max SendQ exceeded]
akaizen has quit [Ping timeout: 245 seconds]
tinti has joined #linux-sunxi
iamfrankenstein has joined #linux-sunxi
cajg has quit [Ping timeout: 248 seconds]
tinti has quit [Max SendQ exceeded]
n01 has quit [Read error: Connection reset by peer]
tinti has joined #linux-sunxi
Black_Horseman has joined #linux-sunxi
Black_Horseman has joined #linux-sunxi
n01 has joined #linux-sunxi
<n01>
oliv3r: the tty for mainline is ttyS0 right?
pacopad has quit [Quit: pacopad]
<tkoskine>
ttyS0 console worked for me when I just rebuild sunxi-3.4 kernel (commit 9ee9fc5f0988df5677f0f142b5b88a8988d283d7) and Debian sid image for pcDuino (A10).
<oliv3r>
n01: aye
popolon has joined #linux-sunxi
<oliv3r>
mripard: what does this mean: reg-fixed-voltage ahci-5v.3: could not find pctldev for node /soc@01c00000/pinctrl@01c20800/ahci_pwr_pin@0, deferring probe
<oliv3r>
seems to work however
ykchavan has quit [Read error: No route to host]
<n01>
mripard: is there any problem with uart on 3.12.0-rc2?
<oliv3r>
n01: 3.12.0-rc1 still works
<oliv3r>
but i have to now force console output, otherwise i only get console terminal
<oliv3r>
env set bootargs=console=ttyS0,115200
<oliv3r>
make sure to enable the designware quircks
<oliv3r>
and double check debuging options :)
<n01>
oliv3r: designware quircks
<n01>
?
<oliv3r>
our uart is based on designware IP
<oliv3r>
there's an option in the kernel, designware 8250 quircks
<n01>
3.11.0-rc7 boots fine
<oliv3r>
CONFIG_SERIAL_8250_DW
<n01>
oliv3r: isn't it enabled in multi_v7_defconfig?
<oliv3r>
possibly
<oliv3r>
check the sunxi_defconfig on the ML
<n01>
Symbol: SERIAL_8250_DW [=y]
<n01>
jesus, how to waste the evening ...
<oliv3r>
i hear ya!
<oliv3r>
i've done the same thing the last few days
<oliv3r>
for example, exact identical u-boot booting 3.10 results in a nice and verbose kernel
<oliv3r>
when booting 3.12-rc1 i suddenly need bootargs=console
<oliv3r>
or i've had quite some kernels output nothing at all (maybe they crash?)
<n01>
I also fought with gf "you are always in front of pc ... bla ... stuff ..." gh
cajg has joined #linux-sunxi
tzafrir_laptop has quit [Ping timeout: 240 seconds]
Sonicadvance1 has quit [Ping timeout: 260 seconds]
Sonicadvance1 has joined #linux-sunxi
n01 has quit [Read error: Operation timed out]
akaizen has joined #linux-sunxi
Soru_ has quit [Ping timeout: 256 seconds]
akaizen_ has quit [Ping timeout: 240 seconds]
cajg has quit [Ping timeout: 252 seconds]
ganbold_ has quit [Remote host closed the connection]
Gerwin_J has joined #linux-sunxi
hno has quit [Ping timeout: 260 seconds]
focus_it has quit [Quit: Leaving]
hno has joined #linux-sunxi
tinti has quit [Remote host closed the connection]
<Turl>
oliv3r: are you testing on a CB1 or a CB2?
<Turl>
that spewing on the mail is from CPUfreq, which is only 'supported' on sun4i with that code I ported from 3.4