<Entropy512>
I only experience the issue 1-2 times a day
<Entropy512>
so no clue if I'm fixed or not
<xplodwild_work>
Entropy512: will test on SGS2 and Note once I'm home
<Entropy512>
but the "fix" involves pulling in a driver known to be at the root of the infamous "superbrick" fiasco
<Entropy512>
I THINK I made the change needed to render it safe
<Entropy512>
but after weeks of Superbrick I'm paranoid
<xZain>
you brick your Phone
<Entropy512>
xplod did, I've avoided it
<Entropy512>
but many other Note owners
<xZain>
o.o
<Entropy512>
and anyone who has used MMC drivers other than those from I9100 Update4 has had bad shit happen
<Entropy512>
people have been bricking right and left since N7000 ICS went official
<Entropy512>
I9100 has the defect but no kernels that trigger it (other than Siyah 3.1rc6 - because it pulled in an MMC driver from a dangerous source)
<Entropy512>
however we think we know which change renders kernels dangerous and I reverted that one out
<Entropy512>
:)
<xZain>
lol
<xZain>
nice
<Entropy512>
xplodwild_work: might want to consider a test build, I won't be able to push one out before I need to go to work though
<stickyboy>
Ok. Flashed with Odin in Windows VM.
<xplodwild_work>
Entropy512: I'll do one this afternoon if I've got time
<Entropy512>
k
<Entropy512>
whee lockscreen calendar
<Entropy512>
I wonder why Infuse users don't seem to have any issue with the new driver
datagutt has joined #teamhacksung
datagutt has quit [Changing host]
datagutt has joined #teamhacksung
sbrissen has joined #teamhacksung
sbrissen_away has quit [Ping timeout: 265 seconds]
<xZain>
xplodwild_work you know about the bug upon receiving notification the music volume goes down
<Entropy512>
dude we know and it probably won't get fixed, libaudio is sucking
<Entropy512>
who removed that from known issues...
<xplodwild_work>
Thanks Yamaha.
supercurio has quit [Quit: et Hop !]
Rebell_ is now known as Rebellos
<Entropy512>
xplodwild_work: yamaha and Samsung in general
<Entropy512>
but yeah that MC1N2 driver is scary
<xplodwild_work>
Entropy512: well Samsung itself isn't that bad at opensourcing
cdesai has joined #teamhacksung
<xplodwild_work>
You find almost everything on Linaro
<xplodwild_work>
(mfc, basic camera hal, etc)
<xplodwild_work>
but the audio hal is yamaha's proprietary
lolwat has quit [Read error: Connection reset by peer]
<Espenfjo>
Sensors are done in two parts. libakm.so which is the main interface against the kernel. This one reads raw values from the kernel, does gain adjustments, does magic, etc. etc, and writes the values to an uinput/uevent handler.
<Espenfjo>
Then there is libsensros
<Espenfjo>
Then there is libsensors (AkmSensor.cpp) that reads the values from the uinput/uevent handler and does unit conversion
<Espenfjo>
Like mPendingEvents[Orientation].orientation.azimuth = value * CONVERT_O_A;
<xZain>
WTF is ??/100 Charging ? ??
<xZain>
and my phone seems super HOT
<Entropy512>
no clue - battery hardware failure?
<xZain>
it is from codeworkx kernel and nightly
<Entropy512>
Espenfjo: ah, so the "magic" is in libakm?
<Espenfjo>
Most of it, yes
<Espenfjo>
And libakm is very closed source
<Espenfjo>
So in theory it should be possible to fix this, since libakm is the same all the way
<xZain>
:\ can someone mirror the latest Nightly ?
<Entropy512>
hmm, so where does the compass tilt compensation occur?
<Entropy512>
it's not anywhere in libsensors
<Espenfjo>
AkmSensor.cpp
<Espenfjo>
yaw/pitch/roll
<Entropy512>
Espenfjo: compass tiltcomp isn't in there, it's just getting raw x/y/z mag readings
<Entropy512>
the conversion into a heading isn't there
<Entropy512>
crud, a bit late, afk shower
<Espenfjo>
There is two ways of doing orientation. 1. applying math to accelerometer + magnetometer. 2. Using values from orientation sensor (yaw,pitch,roll).
<Entropy512>
hmm so the akm is doing internal y/p/r fusion?
<Entropy512>
annoying
<xZain>
I need the latest Nightly
<xZain>
can someone ?
<xplodwild_work>
xZain: get.cm ?
sbrissen has quit [Quit: Konversation terminated!]
<xZain>
nvm I uTorrent it :
<xZain>
damn the kernel with old wifi driver + nightly = some funny shit
<xZain>
going to update to latest nightly
<xZain>
phone is uber HOT !
sbrissen has joined #teamhacksung
<Entropy512>
Espenfjo: is there a chance that Samsung's libsensors ignores y/p/r from libakm and does the y/p/r math from accel/mag themselves?
<Espenfjo>
It might
<Espenfjo>
It is the preferred way to do it in ics
<Espenfjo>
But then again, I get the same, broken result doing it that way as well
<Entropy512>
oh so even your raw accel/mag readings are busted?
<Espenfjo>
They have to be, altough it is worth double checking that one
<Espenfjo>
Doing a compass app using both methods
<Entropy512>
Where did we get our libakm header from?
<Entropy512>
Is there a possibility our libakm uses a slightly different interface?
<Entropy512>
(see the whole "extra field in libgps sv stats" mess...)
<Entropy512>
ok I have to go now
rodries has joined #teamhacksung
Rebellos is now known as Rebellos|afk
rtfpessoa has quit [Ping timeout: 245 seconds]
rtfpessoa has joined #teamhacksung
sbrissen_ has joined #teamhacksung
sbrissen has quit [Ping timeout: 246 seconds]
sbrissen_ is now known as sbrissen
Rebellos has joined #teamhacksung
Rebellos|afk has quit [Ping timeout: 246 seconds]
Rebell_ has joined #teamhacksung
Rebellos has quit [Ping timeout: 256 seconds]
modeme has joined #teamhacksung
xZain has quit [Ping timeout: 245 seconds]
peterperfect has joined #teamhacksung
supercurio has joined #teamhacksung
supercurio has joined #teamhacksung
maltloaf has quit [Read error: Connection reset by peer]
toastcfh_ has joined #teamhacksung
Rebell_ has quit [Ping timeout: 240 seconds]
koud_ has quit [Read error: Connection reset by peer]
<Espenfjo>
Entropy512: It(libsensors loading libakm) comes from the Crespo
koud has joined #teamhacksung
maltloaf has joined #teamhacksung
FRLinux has quit [Ping timeout: 248 seconds]
peterperfect has quit [Ping timeout: 245 seconds]
xplodwild_work has quit [Quit: Page closed]
FRLinux has joined #teamhacksung
devz3ro has quit [Read error: Connection reset by peer]
devz3ro has joined #teamhacksung
nebkat has quit [Remote host closed the connection]
davidism has joined #teamhacksung
nebkat has joined #teamhacksung
stickyboy is now known as sticky|away
autoprime has quit [Read error: Connection reset by peer]
autoprime has joined #teamhacksung
cdesai has quit [Remote host closed the connection]
koud has quit [Read error: Connection reset by peer]
koud has joined #teamhacksung
<dhiru1602>
EXPORT_SYMBOL requires changes to Makefiles?
autoprime has quit [Read error: Connection reset by peer]
autoprim_ has joined #teamhacksung
T6720K34 has joined #teamhacksung
a90342 has quit [Remote host closed the connection]
techmik has quit [Read error: Connection reset by peer]
techmik has joined #teamhacksung
techmik has quit [Read error: Connection reset by peer]
techmik has joined #teamhacksung
squadzone has joined #teamhacksung
luminoso has joined #teamhacksung
peterperfect has joined #teamhacksung
koud has quit [Read error: Connection reset by peer]
<davidism>
Anyone know how to force USB Host Mode for SGS2?
<Espenfjo>
Still no, i would guess :)
<Espenfjo>
Entropy512: One thing though. The amount of data read with the opensource sensors is about ten times the amount per second.. Alltough just applying a sleep, or even a low-pass filter do not help.
<Espenfjo>
Entropy512: Sleeps will just get the device to crash
<steinex>
using a rom based on CM9 that is using CM kernel by default, wonder if it may be enough to extract kernel from latest nightly and be done with it
<steinex>
Espenfjo: ok, thanks
M_N is now known as M_N-Away
<Espenfjo>
Not sure what is implementing it, but it is somewhere in the framework
<Espenfjo>
It can be set at build time in the device overlay
<Kaik541>
Espenfjo: so if I'm reading this correctly, CM's official standpoint is "fuck the deaf"
<Kaik541>
:P
<Espenfjo>
Kaik541: Definately :P
peterperfect has quit [Ping timeout: 245 seconds]
<codeworkx>
shit fucking device
<codeworkx>
looks like it only accepts every seconds vibe pattern
<Kaik541>
codeworkx: well there's your problem, why are you trying to make it fuck shit
<Espenfjo>
So ill submit a patch to hardcode it to /usr/bin/stat :P
sbrissen is now known as sbrissen_gone
Turtuga_ has quit [Quit: Leaving]
Turtuga has joined #teamhacksung
dhiru1602 has quit [Ping timeout: 248 seconds]
dhiru1602 has joined #teamhacksung
autoprim_ has quit [Read error: Connection reset by peer]
autoprime has joined #teamhacksung
koud has quit [Remote host closed the connection]
koud has joined #teamhacksung
M_N has left #teamhacksung [#teamhacksung]
davidism has quit [Quit: Leaving]
<Entropy512>
<codeworkx> Entropy512: try pershots wifi driver. (galaxy-2636 repo) - getting that to run on Exynos/hbird devices is easier said than done - tried on Infuse and it didn't play well
<Entropy512>
give the new MMC code a try - I only -1ed it out of paranoia - want others to test it more
<Entropy512>
will try pershoot's this weekend but I suspect it'll need some work to not break on our devices
h8rift has joined #teamhacksung
<Entropy512>
<codeworkx> looks like it only accepts every seconds vibe pattern - just give up on getting vibedroid to work
<Parth>
yeah I've never gotten the vibe pattern for the MMS app to work right
<Parth>
like, sometimes it works, but most of the time it like spazzes and doesn't do what I tell it to do
Rebellos has joined #teamhacksung
<codeworkx>
my vibrator is very very weak
<codeworkx>
and i dunno why
<codeworkx>
switched to tuna driver
<codeworkx>
still weak -.-
<Parth>
my vibrator is noisy
<Parth>
it's strong, but it doesn't follow any consistent pattern set in the MMS app settings (for Vibe Pattern)
<Parth>
oh, i9300 sources have a newer bcm4330 wifi driver
<Parth>
want to give that a spin? :P
<Entropy512>
hmm, maybe, wonder if it's similar to N7K one
<Entropy512>
you gotta leech it
<Entropy512>
and then you diff it
<Entropy512>
and then you go insane
<Parth>
no it's newer than N7k's driver
<Parth>
well, similar yeah
<Parth>
these driver updates are making me excited. i hope they fixed compatibility with the wifi at school
<Entropy512>
goddamnit, another spam email from Mitt "my dog is on the roof" Romney
<Parth>
never mind that gokhan already fixed the n7k's driver to get rid of the lockscreen issue :P
<Parth>
lol
<Entropy512>
Parth: he didn't actually fix it, he pulled in the N7K MMC drivers
<Entropy512>
which people are claiming fixes it
<Parth>
oh idk his post on xda was that he changed some line somewhere
<Parth>
dunno if that's what he did
<Entropy512>
he has no changes other than region code stuff in the wifi drivers
<Entropy512>
but he pulled in the N7000 drivers
<Parth>
FUCKKKKK
<Entropy512>
(minus MMC_CAP_ERASE of course)
<Parth>
5Ghz wifi is broken again
<Entropy512>
N7000 MMC drivers that is
<Entropy512>
working fine for me here
<Parth>
no i mean on Siyah
<Parth>
(this is just me talking)
<Parth>
and NVM, it took longer than usual to bring up the 5Ghz networks.
<Parth>
Entropy512: you guys are reverting the wifi driver in CM kernel? :O
<Entropy512>
no
<Entropy512>
pulling in the N7000 MMC driver to match it
<Entropy512>
(same thing GM did)
<Entropy512>
I'm not sure if it'll fully fix it, but we won't know until more people beat on it
<Parth>
so many confusing posts on xda
<Entropy512>
I've seen reports of N7000 touchwizz kernels having the same issue
<Parth>
does that wifi battery draining bug still exist? i know my phone literally dies in 6 hours with wifi on even if it's sleeping most of the time
devatwork has joined #teamhacksung
<Entropy512>
cannot reproduce, mine is fine with wifi on all the time
<Entropy512>
I think the I9100 update4 drivers had shitty packet filters that would cause problems on dirty networks
<Entropy512>
def. no in the N7000 driver
<Entropy512>
can arpdrain test all day with no wakeups
<Parth>
=\
rtfpessoa has quit [Read error: Connection reset by peer]