mturquette has quit [Read error: Operation timed out]
mturquette has joined #linux-sunxi
Quarx has joined #linux-sunxi
atiti_ has quit [Ping timeout: 248 seconds]
focus has quit [Ping timeout: 248 seconds]
<oliv3r>
gooooooood morning everybody
atiti_ has joined #linux-sunxi
focus has joined #linux-sunxi
_BJFreeman has joined #linux-sunxi
_BJFreeman is now known as BJfreeman
paulk-desktop has joined #linux-sunxi
vicenteH has quit [Ping timeout: 264 seconds]
BJfreeman has quit [Ping timeout: 252 seconds]
_BJFreeman has joined #linux-sunxi
_BJFreeman is now known as BJfreeman
dragonn has joined #linux-sunxi
FR^2 has joined #linux-sunxi
BJfreeman has quit [Ping timeout: 252 seconds]
_BJFreeman has joined #linux-sunxi
_BJFreeman is now known as BJfreeman
vicenteH has joined #linux-sunxi
atiti_ is now known as atiti
_BJFreeman has joined #linux-sunxi
BJfreeman is now known as Guest49515
_BJFreeman is now known as BJfreeman
Guest49515 has quit [Ping timeout: 252 seconds]
shineworld has joined #linux-sunxi
BJfreeman has quit [Quit: had a good time]
<shineworld>
Just a curiosity. Often in Allwinner code products I found the name/term "lichee". What is it ? a nick of developer ? a product name like "vanilla" for android ?
rellla has joined #linux-sunxi
<oliv3r>
shineworld: pretty much
<oliv3r>
lichee is the name for their development tree
<oliv3r>
like vanilla kernel for a kernel.org unpatched default kernel
<oliv3r>
lichee is vanilla + allwinner magic sauce
<shineworld>
thanks for info ... this help me to understand. I'm watching some A20 SDK packages and git trying to figure out hot to move my next steps
<shineworld>
A20 is a new product but I've begin to find more different versions of u-boot, kernels, and so on
<shineworld>
*begun
<shineworld>
My target will be to use A20/Android how I'm already using A10/Android
<shineworld>
*as
<shineworld>
*** bad day ***
paulk-desktop has quit [Quit: Ex-Chat]
wingrime has joined #linux-sunxi
\\Mr_C\\ has quit []
Quarx has quit [Read error: Connection reset by peer]
e-ndy_ is now known as e-ndy
tinti has joined #linux-sunxi
Black_Horseman has quit [Remote host closed the connection]
forcev has quit [Remote host closed the connection]
FunkyPenguin has joined #linux-sunxi
<oliv3r>
mripard_: Turl so ... how do I make platform_drv_getdata/setdata work after using these different attribute groups :S they work fine, but the data i store/retrieve is different (store ok, retrieve NULL). turl you mentioned the 'different object thing' yesterday, but not quite sure how to fix it.
<oliv3r>
Turl: does that mean you are under water?
atiti has joined #linux-sunxi
<hramrach__>
no
<hramrach__>
it means that everything is wet
<hramrach__>
there is a reason most electronics has operating conditions 95% max (non-condensing) or similar
<hramrach__>
to operate at 100% you would need some coating on the pcb at the very least
<Turl>
oliv3r: lol, no
<utente>
what is the best filesystem to use on A10 device? i talk about filesystem to format the flash nand chip for root filesyste,
<hramrach__>
ext4 is the default so most kernels will support that
<Turl>
oliv3r: "The Relative Humidity expresses how much moisture is in the air, as a percentage of the total moisture the air could contain at the current temperature. In other words, if the air has all the water vapor that it can contain at a given temperature, the relative humidity is 100%"
<oliv3r>
Turl: i know it was funny still
<Turl>
:P
<oliv3r>
there, replied to greg!
<hramrach__>
actually it would be probably quite easy to make olinuxino resistant to these conditions if you planned ahead :p
<Turl>
94% now :p
<Turl>
I wonder what's the operating conditions on olimex products wrt humidity
<hramrach__>
sane environment required :p
<Turl>
oliv3r: I don't see it :p
Black_Horseman has joined #linux-sunxi
<oliv3r>
lkml is slow :(
<utente>
hramrach__, i mean, a FS to preserver at maximum the stress caming from to much writing on flash.
<oliv3r>
i think there's like 10 minutes delay
<hramrach__>
utente: it's not easy to do and have a reliable fs at the same time
<utente>
hramrach__, to make hardware resilient and rexixtern to humidity you can use special paint to "tropicalize" the board.
<Turl>
oliv3r: are you subscribed? might be moderated
<oliv3r>
i am not subscribed :)
<oliv3r>
too high volume for me
<utente>
hramrach__, if it was easy, i already had did it. Cos it is not heasy, i came here to ask suggestions .)
<oliv3r>
Turl: i saw the other day, how much callories are in a bag of microwave popcorn; so i best probably stop buying that
jemk has quit [Quit: bye]
<hglm>
utente: You can change your ext4 filesystem from ordered data mode to write-back mode (which is faster) using tune2fs.
<utente>
hglm, thanks.
<hglm>
See my blog at hglm.github.io for some more performance hints.
<utente>
hglm, what is the difference between the 2 modes
<utente>
ok oi go to read
<hglm>
utente: Write back doesn't preserve ordering and delays and combines some writes I guess. It's less safe in case of a hard crash but not really bad.
<utente>
mmm ok
<utente>
nice blog page. i already know many of those tricks, but not all. thansk.
<Turl>
oliv3r: heh
<hglm>
utente: Thanks, performance can be improved a lot on small flash-based devices because most distros still enable all logging etc which is only needed for PC servers and kills performance on a small flash-based device.
<oliv3r>
Turl: well atleast greg took my 1 line 'fixerup' patch for the binary sysfs thing
<oliv3r>
so when that gets pushed, I'm officially a kernel license holder
<oliv3r>
or whatever its called
<oliv3r>
there should be something else in there from me in the media bit, but haven't found my email in the commit log
<oliv3r>
d67ceb3398d7ae7fda78ec753639296f442ecf66
<oliv3r>
ohh that was my first commit!
<hglm>
I am also thinking about making some contributions to the mainline kernel -- but they affect core functionality across multiple platforms so I have to be careful.
<oliv3r>
good, i'm not the only one that feels stupid then
<Turl>
oliv3r: must be because I only wrote a clk driver so far and that doesn't need all that driver stuff, it's just called form the core :P
<oliv3r>
well from what I understand, greg said you don't want to create any sysfs entries in your probe function due to races with userspace
<oliv3r>
you should populate the .group entries in the proper struct
<oliv3r>
then the whole argument came that platform drivers where poop and then they werne't so it was all good
<oliv3r>
but i think you still aren't allowed to put stuff in probe
<Turl>
yeah, from what I gathered, probe == bad
<Turl>
I still don't fully understand why though
<Turl>
isn't probe supposed to decide if the device is there or not?
<oliv3r>
but then maxime says, he would put it in probe, att the attributes to the bin_attrs field in device_driver, and call device_register, but youa re allready being probed so its to late?
<Turl>
mripard_: ^^^
<oliv3r>
i think, the kernel first creates sysfs entries (first the default)
<oliv3r>
don't call hiM! he'll get mad at me again for being stupid :(
paulk-desktop has quit [Quit: Ex-Chat]
<oliv3r>
then it creates whatever you have in your .group struct entry
<oliv3r>
and then calls probe
<oliv3r>
and if i'm not mistaken, but here i could be wrong, all this only happens if dt says this module belongs to your platform
<Turl>
yeah, if a compatible matching your driver's was found right?
<oliv3r>
hmm
<oliv3r>
yeah absolutly
<oliv3r>
as you define your compatlibes in the _register() function via the struct you pass
<oliv3r>
so the kernel knows whether to call probe or not
<oliv3r>
(and thus knows wether to create sysfs entries or not)
<oliv3r>
Turl, mnemoc: btw, i closed some issues on github today for linux-sunxi; most where months old and just silly
<oliv3r>
some == most
<Turl>
oliv3r: :)
<oliv3r>
so if you object on any, reopen them :)
<Turl>
oliv3r: I'll push an updated sunxi-devel later today
wingrime_ has quit [Ping timeout: 256 seconds]
<oliv3r>
i think my mainline tree is pretty foobared anyway
<oliv3r>
i added greg's driver-core-linus as a remote and rebased based on that
<oliv3r>
4000 patches later; having had to skip a few in usb taht failed to apply, i have something ... recent-ish
<Turl>
ouch
<Turl>
I always rebase against torvalds/master
<oliv3r>
i was at 3.10 something from one of maxime's emac trees
<oliv3r>
yeah i will from now on :)
<Turl>
and cherrypick stuff from other branches
<oliv3r>
once this fucking sid driver gets out
<Turl>
n01's watchdog patch was pretty much ready too
<oliv3r>
only a few fixes needed
\\Mr_C\\ has quit []
<Turl>
the maintainer just commented on a variable that needed renaming iirc
<oliv3r>
well once i know how to get platform_read_data to work properly
<oliv3r>
or register my attributes properly, it should be 'done deal'
<Turl>
:)
<oliv3r>
on the upside
<oliv3r>
we made greg write a blog post AND revise sysfs :)
<Turl>
:)
<oliv3r>
which is pretty amaxing for a 4 register read only drivr
<Turl>
oliv3r: you can add that on your CV :P
<oliv3r>
lol
<oliv3r>
i'll have a patch for it too!
<oliv3r>
well i hope it does make me more hireable :)
<oliv3r>
'linux kernel patches'
<oliv3r>
how bad ass is that
<Turl>
you need to write at least two for that :P
<Turl>
ah nvm, dt one is split :)
<jelly-home>
s/patches/driver development/ <- use more long trendy words