pwhalen_ has quit [Read error: Connection reset by peer]
deasy has joined #linux-sunxi
<Froolap>
Using `yum update kernel` is not a good idea - it will brick your device! WHAT???? please explain....
egbert has quit [Disconnected by services]
egbert has joined #linux-sunxi
wenbin has quit [Quit: Leaving]
deasy has quit [Quit: Nom d'un quark, c'est Edmonton !]
wenbin has joined #linux-sunxi
<Seppoz>
samfashi: check the AXP code
<Seppoz>
also what kind of driver and why uboot
<samfashi>
but axp use rsb not i2c. I want to use TWI0.
<kurain>
I am working on porting ap6234 wifi&bt module to sun6i, now wifi module is ok, but bluetooth still hangs. log as below http://paste.ubuntu.com/9137881/
<kurain>
AFAIK, sun6i officially supports ap6210, and I port ap6234 according to ap6210 with small hacks
<kurain>
according to the log, it seems that after we have fw pachfile uploaded by ttyS2 but no respond from AP6234 module
physis has joined #linux-sunxi
physis has quit [Ping timeout: 265 seconds]
Andy-D has joined #linux-sunxi
gianMOD has joined #linux-sunxi
akaizen has joined #linux-sunxi
akaizen has quit [Remote host closed the connection]
akaizen has joined #linux-sunxi
n0wl__ has joined #linux-sunxi
kurain has quit [Quit: kurain]
n0wl_ has quit [Ping timeout: 255 seconds]
TheSeven has quit [Ping timeout: 272 seconds]
TheSeven has joined #linux-sunxi
naobsd has joined #linux-sunxi
JohnDoe_71Rus has joined #linux-sunxi
Andy-D has quit [Ping timeout: 264 seconds]
Andy-D has joined #linux-sunxi
wenbin_ has joined #linux-sunxi
wenbin has quit [Ping timeout: 250 seconds]
gianMOD has quit [Remote host closed the connection]
wenbin_ has quit [Quit: Leaving]
Andy-D has quit [Quit: Alive/Dead]
kurain has joined #linux-sunxi
melonipoika has joined #linux-sunxi
domidumont has joined #linux-sunxi
gianMOD has joined #linux-sunxi
diego_r has joined #linux-sunxi
gianMOD has quit [Ping timeout: 244 seconds]
arete74 has quit [Ping timeout: 256 seconds]
lerc has quit [Ping timeout: 244 seconds]
lerc has joined #linux-sunxi
arete74 has joined #linux-sunxi
sehraf has joined #linux-sunxi
sehraf has quit [Client Quit]
sehraf has joined #linux-sunxi
nemunaire has joined #linux-sunxi
afaerber has joined #linux-sunxi
<rellla>
deinterlacer question: what is the "tile flag address" and the "tile flag line stride" which is needed when using motion adaptive filtering?
FR^2 has joined #linux-sunxi
gianMOD has joined #linux-sunxi
gianMOD has quit [Ping timeout: 258 seconds]
samfashi1 has joined #linux-sunxi
Uninstall_ is now known as Uninstall
Uninstall has quit [Changing host]
Uninstall has joined #linux-sunxi
kurain1 has joined #linux-sunxi
samfashi has quit [Ping timeout: 264 seconds]
kurain has quit [Ping timeout: 255 seconds]
arete74 has quit [Ping timeout: 250 seconds]
arete74 has joined #linux-sunxi
shineworld has joined #linux-sunxi
shineworld has left #linux-sunxi [#linux-sunxi]
gianMOD has joined #linux-sunxi
akaizen has quit []
<Froolap>
is the x in sunxi supposed to be a variable for a number or is it a literal?
gianMOD has quit [Ping timeout: 240 seconds]
<rellla>
Froolap: number
cubear has joined #linux-sunxi
libcg has joined #linux-sunxi
<Froolap>
wanted to make certain. I'm looking at a Makefile that has sun6i and sunxi and wasn't sure if I was supposed to change the x to 7 and the =n to =y.
<Froolap>
Thanks
jemk has joined #linux-sunxi
<jemk>
rellla: as far as i understand this tile flags are produced by the ve, register 0x30-0x4c, but i never tried to use it.
kurain1 has quit [Remote host closed the connection]
kurain has joined #linux-sunxi
<rellla>
jemk: thanks. according to our wiki these seem to be the maf registers. is there anything more known about this registers than the offsets? if we have a possibility to get the buffer address and the line stride out of ve, we can send them to defe.
<rellla>
btw i had success with bob deinterlacer and mpv :p
<rellla>
have you noticed my last code pushes about fixing the deinterlacer and the queue handling?
<jemk>
rellla: no more info yet, we would first need to do some traces with the blob using it
<jemk>
rellla: and i don't know what it does, if thats of much use
<jemk>
yes, but why glib... ;)
<rellla>
jemk: because it's easier for a noob to use an existing library :p
<rellla>
do you think it's basically the right way to handle the workflow?
<jemk>
I started a better output for vdpau with presentation thread a year ago but put it on hold when libv anounced kms...
<libv>
jemk: :(
<libv>
jemk: i am a bit busy on a customer project atm, but at least i am coding again, and wondering on whether i will try to have a talk at fosdem again
<jemk>
I found too many missing parts in disp and thought it would be better to help extending kms once it is there
<libv>
i did figure out the pwm a few weeks ago, so lcd support was advancing
<rellla>
libv: how was your night? nightmares?
<libv>
rellla: ?
<rellla>
i followed #lima
<libv>
heh
<libv>
no, i slept well
<rellla>
forget it :p
<jemk>
rellla: I imagined a different design, but i don't know if it would work and how to explain it
<jemk>
rellla: your way is good enough to get it running i think, but i wanted to have an x event loop in the presentation thread to get move events and so on for non-fullscreen players
<jemk>
rellla: ultimate goal would be to get rid of the direct disp access and use some dri or presentation api or how it is called
<rellla>
jemk: yes it would. but thats too complicated to do for me atm. i'm happy with understanding disp more and more
<jemk>
rellla: and the two queues seem a bit overkill to me, why one isn't enough?
<jemk>
aaaand, vdpau-sunxi is not thread safe in any way. you could get funny errors in the presentation thread
<rellla>
jemk: i would have to add some mutexes!?
libcg has quit [Quit: libcg]
deasy has joined #linux-sunxi
deasy_ has joined #linux-sunxi
<jemk>
rellla: well, it should be made thread safe anyway. the specification requires this.
deasy_ has quit [Read error: Connection reset by peer]
<rellla>
jemk: vdp_presentation_queue_display puts the frames in the outer queue, which is locked while access. presentation_thread gets them an fills his internal queue.
<jemk>
yes, but why? why not get the next frame from the first queue and simply display it?
<rellla>
iirc the two queues ensure, that the displaying order is correct.
<rellla>
but don't see a problem when everything is handled fifo
<rellla>
the first input from presentation_queue_display should be displayed first.
<jemk>
i don't think it is valid to call vdp_presentation_queue_display in a different order than the display order
<rellla>
maybe if presentation_thread is too slow, frames could be skipped. but that would not happen, if the queue is threadsafe.
<rellla>
perhaps it's more safe to let vdpau not get locked if displaying is too slow?
<jemk>
vdpau specification is very inaccurate in many points, but catching every posible misuse is overkill in my opinion
<rellla>
i just tried to copy the vaapi-gl vdpau backend :p
<jemk>
they try to catch every posible missuse... :p
<jemk>
rellla: i would try to keep it simple, if there happen to be players that don't work it still can be extended. but i don't think anyone will try to display frames in different order
<jemk>
rellla: put the frame to the queue in vdp_presentation_queue_display, and in the presentation thread while(running){get one from the queue;display it; wait for vsync;} or something like that
<jemk>
rellla: and add a sleep till earliest_presentation_time befor display
<rellla>
jemk: yes. that was, what i first thought of. much simplier and easier to implement vsync.
gianMOD has joined #linux-sunxi
<jemk>
rellla: it only lacks a way to handle x events and it wont get you correct vsync timestamps to put in first_presentation_time
<jemk>
but to get timestamps disp would need to be extended
<rellla>
can't we easily add a FBIO_WAITFORVSYNC before displaying?
gianMOD has quit [Ping timeout: 240 seconds]
<rellla>
additionally we have to handle frames, that are already too late to display, in case displaying is too slow
<jemk>
rellla: what qualifies a frame as to late?
<jemk>
the only thing vdpau does is display a frame as soon as possible after earliest_presentation_time, or if this is 0, immediately.
<rellla>
what happens to the queue, if it is filled more quickly than presentation_queue takes fram out and displays it based on the vsync?
<rellla>
is this a thinkable scenario anyway?
cnxsoft has quit [Ping timeout: 255 seconds]
<jemk>
then the player is buggy
<jemk>
or i misunderstand vdpau
<rellla>
what if display engine is too slow? maybe i have some error in thinking ...
<jemk>
the display engine shows one frame per frame ;)
<jemk>
not more and not less
vitorcarvalhoml has joined #linux-sunxi
<rellla>
ok. so that should not get a problem. i'll try that with some easier queue.
naobsd has quit [Quit: Page closed]
<jemk>
rellla: i'm not sure, but i would not expect problems that are unlikely and make my work much harder.
<jemk>
rellla: if they still turn up we can still find a solution
furmuwon has joined #linux-sunxi
<furmuwon>
hi members
<jemk>
rellla: and don't do handle_get in the thread, it's not thread safe. I should really fix this... But i don't like the easy way of adding mutexes everywhere. It might not be noticeable performance wise, but i don't like it
<furmuwon>
Is the CedarX(Video Engine) supported to encode multiple instance?
<furmuwon>
CSI0 -> VE -> BITSTREAM at same time CSI1 -> VE -> BITSTREAM.
<furmuwon>
is it possible?
samfashi1 has quit [Quit: samfashi1]
<libv>
MY123: why would you want to adapt non-free software?
<libv>
MY123: do you think that that magically makes it free software?
<MY123>
libv, currently writing docs
<MY123>
t
<MY123>
libv, I know that it would be still be propretary
<libv>
MY123: ...
<libv>
MY123: how do you a see any pvr open source project work from this point on?
<MY123>
libv: Writing docs for VC4 and PowerVR at the *same* time ...
<libv>
will the stigma from the leaked source ever not be present?
<libv>
docs on what?
<MY123>
libv: PowerVR ISA and uKernel communication
<libv>
really, do you want to go write docs on horrible code on a horrible engine, on something that is horribly complex?
<libv>
ukernel communication changes slightly for every vendor.
<libv>
which is why i kept on telling people to not bother with pvr and instead spend their time more usefully on other GPUS to let img clean up their own mess
<mnemoc>
horribleness might overflow...
<MY123>
libv: At least, I got a sample working on a BBB
<libv>
i do not have to worry about dealing with a doomed project
<libv>
which is what an open source driver for pvr now has become.
<libv>
MY123: i am quite aware of that.
<libv>
pvr was an insane project before, with much better targets available
<libv>
now it became a doomed project
<MY123>
libv, why doomed?
<libv>
MY123: how many years do you want to spend writing documentation?
<libv>
who is going to write the actual driver?
<MY123>
libv: olerem will finish writing docs
<libv>
and how is that person going to make sure that the proper flow of information was used for any code going into the actual driver
leviathanch has joined #linux-sunxi
<MY123>
and afaerber will help about writing code
<libv>
and that no actual pvr code flows back?
<libv>
yeah sure
<Wizzup>
I'm surprised that the code is still online, by now, surely, imgtec must know about it
<libv>
everyone just pitches in a few hours of their time
<MY123>
libv, simple, Don't touch the code yourself but get another one to write docs
<libv>
graphics drivers are trivial and small.
<MY123>
libv, not for PowerVR
<libv>
and there are millions of people willing and capable of doing the work.
<libv>
this is why projects like lima, freedreno, etnaviv and grate have 1000s of developers
<libv>
think about it for a second.
<libv>
if you think that the REing is the hard bit, and that that now magically got solved, then you are wrong.
<libv>
you more than doubled the workload here
<libv>
and even then you will spend your days defending yourself against even just accusations of having stolen code.
<libv>
this is not a 1000line media decoder or so
<libv>
how many years do you personally want to dedicated to this?
<libv>
and who else is willing to do that as well?
<libv>
and then know that you will have to be _extremely_ careful about any code from outside this process flowing in.
<afaerber>
MY123, note that I never said I would write such a driver alone, I lack the time to do so even if I had the documentation. I just said that I don't want to taint myself by reading leaked code
<libv>
heh, and so it start
<libv>
starts
<libv>
MY123: think before you start.
<libv>
of all the great things my maths teacher taught me, then it was that.
<afaerber>
libv, did you know whether that uKernel could be replaced with something custom?
<afaerber>
*do
<libv>
it's almost always custom.
<libv>
custom to the revision of the sgx, and custom to the customer project.
<libv>
which is why it is such an insane moving target
<afaerber>
I mean, can we - based on RE'd opcodes - write a trivial uKernel to start with and have that loaded as firmware? or is that "in the silicon"?
<furmuwon>
jemk : Could you tell me about my question?
ricardocrudo has joined #linux-sunxi
<jemk>
furmuwon: If your streams are small enough it is possible for sure, but i don't know how to do it with the cedarx blobs
<jemk>
furmuwon: maybe you can simply run everything twice
<furmuwon>
I will try CSI0(30fps 1920x1080) CSI1(30fps, 1280x720)
Black_Horseman has joined #linux-sunxi
Black_Horseman has joined #linux-sunxi
<libv>
afaerber: it's microcode, so per definition it is not in silicon
<jemk>
furmuwon: that is to much, encoder only does 1080p@30fps for h264
<afaerber>
okay, got it
<furmuwon>
jemk : thank.. I am developer for CDR
<jemk>
furmuwon: both 1280x720 could be possible, but risky
<furmuwon>
jemk : we will make the CDR using the A20. front channel(CSI0) FULLHD@30fps, back channel(CSI1) HD@30fps
bengal has quit [Quit: Leaving]
<furmuwon>
jemk : I want to hear your opinion.
<furmuwon>
jemk : CSI0 is encoded by 1920x1080@30fps using CedarX, CS1 is encoded by 1280x720@30fps using ffmpeg.
<furmuwon>
jemk : is it possible?
<jemk>
furmuwon: if ffmpeg is fast enough to encode 720p in software it could be possible, but i doubt its fast enough.
<furmuwon>
jemk : Thank you. I will try that way.
<jemk>
furmuwon: A31 and a23 can do 1920x1080@60fps, that would be enough
<jemk>
furmuwon: at least the manual claims it is that fast, I guess nobody tried it yet
<furmuwon>
jemk : I will see but A20 price is cheap... maybe.. :)
leviathanch has quit [Ping timeout: 256 seconds]
gianMOD has joined #linux-sunxi
Quarx has joined #linux-sunxi
gianMOD has quit [Ping timeout: 272 seconds]
Black_Horseman has quit [Ping timeout: 255 seconds]
buZz has quit [Ping timeout: 272 seconds]
gianMOD has joined #linux-sunxi
libcg has joined #linux-sunxi
cnxsoft has joined #linux-sunxi
JohnDoe_71Rus has quit [Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org]
buZz has joined #linux-sunxi
buZz is now known as Guest81559
Guest81559 is now known as buZz
gianMOD has quit [Remote host closed the connection]
naobsd has joined #linux-sunxi
popolon has joined #linux-sunxi
leviathanch has joined #linux-sunxi
TheQuestionmark has quit [Ping timeout: 272 seconds]
lioka has quit [Ping timeout: 272 seconds]
souther has quit [Ping timeout: 272 seconds]
TheSeven has quit [Read error: Connection reset by peer]
doppo has quit [Ping timeout: 272 seconds]
doppo has joined #linux-sunxi
souther has joined #linux-sunxi
lioka has joined #linux-sunxi
TheSeven has joined #linux-sunxi
gianMOD has joined #linux-sunxi
TheQuestionmark has joined #linux-sunxi
viccuad has joined #linux-sunxi
lioka has quit [Ping timeout: 258 seconds]
lioka has joined #linux-sunxi
<Seppoz>
so merrrii does NOT sell a10 boards anymore?
<Seppoz>
because allwinner dont sell them A10 cpu no more?
<Seppoz>
WTF
leviathanch has quit [Ping timeout: 264 seconds]
<Wizzup>
Does allwinner still produce A10 cpus?
<Seppoz>
i wrote them and email and they said its still availiable, whatever that means
<Wizzup>
That may mean in stock, but not actively producing?
leviathanch has joined #linux-sunxi
Gerwin_J has joined #linux-sunxi
lioka has quit [Changing host]
lioka has joined #linux-sunxi
sleeply4cat has joined #linux-sunxi
<sleeply4cat>
Hello again.
<sleeply4cat>
Now I'm trying to enable Wi-Fi on my tablet. I have an 8188eu module.
<libv>
sleeply4cat: wiki.
<sleeply4cat>
I have compiled a kernel with built-in driver.
leviathanch has quit [Ping timeout: 264 seconds]
<Seppoz>
hey sleepy <3
<Seppoz>
sleeply4cat: wifi is tricky
<sleeply4cat>
...should I use exactly the same method which is written in the wiki?
<Seppoz>
first of all you need uboot with firmware support (which is not enabled by default)
<sleeply4cat>
whae.
<Seppoz>
and thenn you also need the firmware
<Seppoz>
what distro are you using?
<sleeply4cat>
Debian wheezy
<Seppoz>
then uboot has firmware support
<Seppoz>
make sure to install the debian package linux-firmware
<sleeply4cat>
eww..
<Seppoz>
and all the 3rd party stuff
<Seppoz>
google linux firmware debian package
<Seppoz>
and then you also need kernel driver
<Seppoz>
ofcourse
<sleeply4cat>
wait pls >_<
<sleeply4cat>
Is u-boot depends on choiced distro? 0_o
<Seppoz>
no
<sleeply4cat>
[17:38] <Seppoz> what distro are you using? [17:39] <sleeply4cat> Debian wheezy [17:39] <Seppoz> then uboot has firmware support
<Seppoz>
only effect uboot can have is the boot parameter on what to boot the distro from
<Seppoz>
understand plz
<Seppoz>
ooooo
<Seppoz>
sorry
<Seppoz>
i meant UDEV
<Seppoz>
soooooory
<sleeply4cat>
ok; now I need to modify a .fex file?
<sleeply4cat>
Going to trace connections on PCB =\
deasy has quit [Quit: Nom d'un quark, c'est Edmonton !]
<Seppoz>
you sure your using the right fex?
<sleeply4cat>
yup.
<sleeply4cat>
I have script0.bin
<sleeply4cat>
and script.bin
<sleeply4cat>
I used script.bin to get a fex
<Seppoz>
but you are booting from sd right?
<sleeply4cat>
+
<Seppoz>
why you load script.bin from annd then
<Seppoz>
start uboot
<Seppoz>
and get me printenv content
<Seppoz>
i mena buring boot
<sleeply4cat>
Sorry, can't translate correctly.
npcomp_ is now known as npcomp
<Seppoz>
get into the uboot console and type printenv
<sleeply4cat>
wait plz. Hw problems =\
<sleeply4cat>
Anyway, I think, I can unsolder a WIFI module and trace connections..
<Seppoz>
your not using the right fex file man
<Seppoz>
[ 1.102354] [sw-ehci1]: Set USB Power ON
<Seppoz>
[ 1.285110] [sw-ehci2]: Set USB Power ON
<Seppoz>
is how its suppose to look
<Seppoz>
you dont have ehci2
<sleeply4cat>
0_o
<Seppoz>
show me the uboot env ill check where its loaded
<Seppoz>
but if you use the uboot from git it should be loaded from sdcard
<sleeply4cat>
wait plz
<sleeply4cat>
checking voltage on modules..
<sleeply4cat>
3.3v after a few seconds.
<sleeply4cat>
it's ok.
<Seppoz>
yea but the usb port its attached to is not enabled
<sleeply4cat>
maybe. Checking..
<Seppoz>
def :P
<sleeply4cat>
Should I take some photos before assembling a tablet back?
<Seppoz>
yes
<sleeply4cat>
Are there any components under motherboard?
<sleeply4cat>
hmm, seems glued.
<sleeply4cat>
no way.
FR^2 has quit [Quit: Leaving]
<hno>
sleeply4cat, the usb-OTG port and usb-HOST ports are very different. Different hardware and different drivers, and somewhat different content in the Allwinner fex file.