hno changed the topic of #linux-sunxi to: Allwinner/sunxi development discussion - Don't ask to ask. Just ask! - See | | Logs at
Soru has joined #linux-sunxi
Soru has quit [Ping timeout: 264 seconds]
Soru has joined #linux-sunxi
<vrga> Turl, once more, thank you. the info you provided enabled me to boot the soggy thing into linaro.
<vrga> so its one step towards moving the tab off of android.
tinti_ has quit [Ping timeout: 245 seconds]
alcides has quit [Remote host closed the connection]
egbert has quit [Disconnected by services]
egbert has joined #linux-sunxi
egbert has quit [Disconnected by services]
egbert_ has joined #linux-sunxi
techn_ has joined #linux-sunxi
[hawk] has joined #linux-sunxi
e-ndy_ has joined #linux-sunxi
arete74 has joined #linux-sunxi
Soru_ has joined #linux-sunxi
Soru has quit [*.net *.split]
vincenzo has quit [*.net *.split]
e-ndy has quit [*.net *.split]
techn__ has quit [*.net *.split]
arete74_ has quit [*.net *.split]
bfree has quit [*.net *.split]
ssvb has quit [*.net *.split]
ChanServ has quit [*.net *.split]
vrga has quit [Quit: Leaving]
ssvb has joined #linux-sunxi
MadSpark- has joined #linux-sunxi
^hawk^ has joined #linux-sunxi
zrafa has joined #linux-sunxi
stekern_ has joined #linux-sunxi
heffer_ has joined #linux-sunxi
MadSpark has quit [Ping timeout: 264 seconds]
heffer has quit [Ping timeout: 264 seconds]
[hawk] has quit [Ping timeout: 264 seconds]
stekern has quit [Ping timeout: 264 seconds]
zrafa_ has quit [Ping timeout: 264 seconds]
Turl has quit [Ping timeout: 264 seconds]
rm has quit [Ping timeout: 264 seconds]
rm has joined #linux-sunxi
Turl has joined #linux-sunxi
jlj has quit [Ping timeout: 248 seconds]
jlj has joined #linux-sunxi
ganbold has joined #linux-sunxi
wigyori_ has joined #linux-sunxi
tinti has quit [Quit: Leaving]
hsildaerD has joined #linux-sunxi
ganbold__ has quit [*.net *.split]
TheSeven has quit [*.net *.split]
\\Mr_C\\ has quit [*.net *.split]
andoma has quit [*.net *.split]
wigyori has quit [*.net *.split]
hark_ has quit [*.net *.split]
^vincenzo^ has joined #linux-sunxi
andoma has joined #linux-sunxi
hark_ has joined #linux-sunxi
gzamboni_ has joined #linux-sunxi
cajg_ has joined #linux-sunxi
zumbi_ has joined #linux-sunxi
^hawk^ has quit [*.net *.split]
ssvb has quit [*.net *.split]
specing has quit [*.net *.split]
cajg has quit [*.net *.split]
Dreadlish has quit [*.net *.split]
gzamboni has quit [*.net *.split]
zumbi has quit [*.net *.split]
mnemoc has quit [*.net *.split]
specing has joined #linux-sunxi
ssvb has joined #linux-sunxi
BJfreeman has quit [Quit: had a good time]
ChanServ has joined #linux-sunxi
bfree has joined #linux-sunxi
mnemoc has joined #linux-sunxi
bfree has quit [*.net *.split]
ChanServ has quit [*.net *.split]
\\Mr_C\\ has joined #linux-sunxi
servili007 has quit [Read error: Connection reset by peer]
bfree has joined #linux-sunxi
ChanServ has joined #linux-sunxi
ykchavan has joined #linux-sunxi
jlj has quit [Ping timeout: 245 seconds]
bfree has quit [*.net *.split]
ChanServ has quit [*.net *.split]
\\Mr_C\\ has quit []
jlj has joined #linux-sunxi
hipboi has joined #linux-sunxi
jlj has quit [Ping timeout: 248 seconds]
jlj has joined #linux-sunxi
bfree has joined #linux-sunxi
ChanServ has joined #linux-sunxi
zumbi_ is now known as Guest95491
Turl is now known as Guest90555
rm has quit [Changing host]
rm has joined #linux-sunxi
Guest90555 has joined #linux-sunxi
Guest90555 has quit [Changing host]
Guest90555 is now known as Turl
stekern_ is now known as stekern
hipboi has quit [Ping timeout: 245 seconds]
hipboi has joined #linux-sunxi
Tsvetan2 is now known as Tsvetan
wingrime has joined #linux-sunxi
atiti_ has quit [Ping timeout: 240 seconds]
Soru_ has quit [Read error: Operation timed out]
hipboi has quit [Ping timeout: 268 seconds]
hipboi has joined #linux-sunxi
Soru has joined #linux-sunxi
atiti_ has joined #linux-sunxi
atiti_ is now known as atiti
d9500 has joined #linux-sunxi
TheSeven has joined #linux-sunxi
<oliv3r> morni'
wigyori_ is now known as wigyori
ykchavan has quit [Ping timeout: 264 seconds]
<atiti> morning
dragonn has joined #linux-sunxi
_BJFreeman has joined #linux-sunxi
_BJFreeman is now known as BJfreeman
rellla has joined #linux-sunxi
<oliv3r> Turl: there, sent an e-mail
shineworld has joined #linux-sunxi
<rellla> morning
<rellla> Turl: we need to investigate, how to exclude those f
_BJFreeman has joined #linux-sunxi
<rellla> f*** "credit authorisers"
<rellla> it's quite a mess
BJfreeman is now known as Guest35110
_BJFreeman is now known as BJfreeman
Guest35110 has quit [Ping timeout: 252 seconds]
\\Mr_C\\ has joined #linux-sunxi
<oliv3r> time to replace captcha questions :)
n01 has joined #linux-sunxi
<oliv3r> ok this section on the thumb2 page makes little sense to me
<oliv3r> armv5et is FASTER then armv7-a?
<oliv3r> aren't we ignoring all the armv7 optimizatoins that way? so how can that be faster?
<oliv3r> so armv5te is faster then armv6?
focus has quit [Remote host closed the connection]
focus has joined #linux-sunxi
d9500 has quit [Quit: Leaving]
hsildaerD is now known as Dreadlish
Dreadlish has quit [Changing host]
Dreadlish has joined #linux-sunxi
<wingrime> Tsvetan: ping
\\Mr_C\\ has quit [Quit: sigh]
<Tsvetan> wingrime pong
FunkyPenguin has quit [Remote host closed the connection]
notmart has joined #linux-sunxi
notmart has joined #linux-sunxi
<oliv3r> tabletennis time!
atiti_ has joined #linux-sunxi
<oliv3r> ssvb: just at a curiosity stage right now, what driver should I read to try to understand how normally the 'disp-layer' is done
uwe__ has joined #linux-sunxi
<oliv3r> i'm basically only curious in the final stages of output, e.g. framebuffer -> scale -> output
<wingrime> Tsvetan: i have no any a20 :(
<oliv3r> ssvb: defe i'm guessing in the a1* terms
<oliv3r> wingrime: :(
<Tsvetan> wingrime you can change this easily ;)
<oliv3r> :)
hark has joined #linux-sunxi
<oliv3r> wingrime: how's mpeg reverse engineering going?
<oliv3r> i think after jpeg, mpeg should be easy
<oliv3r> and mpeg shouldn't be too hard with the knowledge you guys have gained so far
<oliv3r> i suppose after mpeg2, h264 will be important (but much harder)
<oliv3r> with jpeg, mpeg2 and h264, we should be able to start working on a decoder library, so that broadcasting decoding should be all possible (dvb-t, dvb-s, dvb-s2)
\\Mr_C\\ has joined #linux-sunxi
FunkyPenguin has joined #linux-sunxi
<wingrime> oliv3r: may bit fields
<oliv3r> from what I read, h254 is like mpeg2 on steroids, isn't it?
atiti has quit [*.net *.split]
hark_ has quit [*.net *.split]
rm has quit [*.net *.split]
uwe_ has quit [*.net *.split]
<wingrime> oliv3r: It diferent engine
<oliv3r> bah :(
<oliv3r> well mpeg2 will give us SD-decoding
<wingrime> oliv3r: blob constain some h264 registeer names
<wingrime> oliv3r: but not all
<oliv3r> right
<oliv3r> i really wish i had the time to integrate the jpeg poc into a something usable, maybe in a few weeks
<oliv3r> i really want to do it, as i'm excited for it :)
<wingrime> oliv3r: mpeg engine do mpeg124 ms-mpeg VP6 xdiv xvid
<oliv3r> okay
<oliv3r> well dvb-s2 braodcast is h264 isn't it
<wingrime> oliv3r: h264 more difficult and complex
<oliv3r> yeah :S
<oliv3r> well vp6 is 'nice to have' as is xvid
<oliv3r> ms-mpeg, mpeg4 isn't that much used I suppose
<oliv3r> still, nice to have too
<wingrime> oliv3r: but mpeg engine now most undestandable
<oliv3r> awesoem
<oliv3r> can't wait till we have some poc for that
<wingrime> oliv3r: soc are not optimized
<oliv3r> remember that bit
<oliv3r> it is belived that the mpeg2 decoder is done with/from that
rm has joined #linux-sunxi
<wingrime> oliv3r: you can see
<wingrime> oliv3r: expectly you also can fix TOC
<wingrime> in Clock page
<atiti_> hey guys
atiti_ is now known as atiti
MadSpark- is now known as MadSpark
FR^2 has joined #linux-sunxi
<wingrime> atti?
<wingrime> atiti?
<atiti> hi :)
<oliv3r> wingrime: oh wow, you did a lot of work on the page, I think you doubled its size since i was working on it :)
<wingrime> oliv3r: VC1 (MS-H264) engine regs still not there\
<atiti> i shall be looking at h264 decoding with cedar today
<atiti> :D
<wingrime> atiti: we have jpeg only without blob
<atiti> ill try it with the blob then
<wingrime> atiti: but this is a much step forward
<atiti> nice :D
<wingrime> it parse jpeg and decode using cedar , than save file into ppm
<atiti> nice work
<atiti> i see its directly talking to /dev/cedar_dev
<wingrime> atiti: re page have jpeg decoding process described
<atiti> cool, right now im at the stage of trying to get cedar + binary blob to work with cvlc
<wingrime> atiti: but problem that cedar_dev using mmio to userspace, but it totaly hack, and can't be on mainline kernel, witch means , we need move most decoder process to kenrel, and mainline it
<wingrime> atiti: total rewrite
<atiti> :/ sounds like a big task
<wingrime> Trul: any news about CM on A13?
BJfreeman has quit [Quit: had a good time]
<oliv3r> ssvb: hm, i guess nvm, defe is only the scaler, tcon and tvenc basically control the output bit
ykchavan has joined #linux-sunxi
<wingrime> ssvb: there is way use vlc and your driver same time
<wingrime> ?
<ssvb> wingrime: it should work
<ssvb> wingrime: except that a13 supports only one scaled layer
<ssvb> wingrime: but a10 should be fine :)
<wingrime> ssvb: so in a13 casee?
<ssvb> wingrime: I need to change the driver to ask for the scaled layer only when it is really needed instead of reserving it all the time
* rellla is looking forward to the recently ordered cubie a20...
<wingrime> ssvb: jemk fixed jpeg PoC , now it display using disp, but I don't know why 'time' app say around 1 sec to run, much slow I think
<ssvb> looks like it only takes 5 seconds to kill a panda with cpuburn - :)
<oliv3r> LOL
<oliv3r> so clocked lsightly to high per default
<oliv3r> ssvb: i'm looking at the tcon registers now, wich are the LCD0 and LCD1 controllers? which are also used for HDMI out etc?
<ssvb> oliv3r: maybe :) there is something fishy about the TI OMAP maximum clock frequency, the nominal is 1.2GHz, but the kernel needs some special power management patches which were missing from the mainline, etc.
gzamboni_ has quit [Read error: Connection reset by peer]
<oliv3r> ssvb: ah, i wouldn't be supprised if it's clocked slightly to high :)
<ssvb> oliv3r: I think the main pandaboard guy (prpplague) hates me, he overreacted the last time when we had a chat :)
<oliv3r> hehehe
<oliv3r> hate is a strong word :)
<ssvb> oliv3r: for hdmi it is better to ask hansg
<ssvb> wingrime: if you are doing uncached reads to grab the decoded jpeg, they are going to be really slow
<oliv3r> i'm just figuring out how the disp engine hangs together
<ssvb> wingrime: have you tried profiling to see where the CPU time is spent?
<oliv3r> i see tcon0, tcon1 and tv0, tv1 registers for output
<oliv3r> so I take it those control how things are output
<oliv3r> Color_Bar_Mode
<oliv3r> 0: The Video Encoder input is coming from the Display Engineer
<oliv3r> 1: The Video Encoder input is coming from an internal standard color
<oliv3r> :D
<oliv3r> interesting
<oliv3r> should help with getting things working
<wingrime> ssvb: not yet, now I want make trace for vp6
<wingrime> ssvb: I just checked mplayer in console and wondered, that mplayer can play mpeg2 with full speed using fbdev
hipboi has quit [Ping timeout: 268 seconds]
Black_Horseman has quit [Ping timeout: 240 seconds]
<wingrime> ssvb: with x11 it was realy slowpocke
<ssvb> wingrime: yes, mplayer has a really crappy x11 vo
<ssvb> I think even Pentium II 300MHz was enough for decoding mpeg2 encoded dvd
<ssvb> so the decoding speed should not be a problem
<wingrime> ssvb: that can describe why any browser so slow
dragonn has quit [Ping timeout: 268 seconds]
dragonn has joined #linux-sunxi
FunkyPenguin has quit [Remote host closed the connection]
FunkyPenguin has joined #linux-sunxi
forcev has joined #linux-sunxi
<wingrime> ssvb: why even with your driver , x11 so slow
forcev has quit [Remote host closed the connection]
<ssvb> wingrime: the driver does not have much to do here, it's basically a XShmPutImage
<ssvb> the rest are redundant buffer copies in mplayer, and also very inefficient scaling (if you enable it)
ykchavan has quit [Quit: Leaving]
<wingrime> ssvb: I just checked mp4 with ac3, it also run without any framedrop
<ssvb> do you have audio enabled?
<wingrime> alsa
<wingrime> yes
<wingrime> ssvb: but framebuffer have lags
<ssvb> ok, just "without any framedrop" may mean "slow motion" sometimes
<ssvb> aha, that's it :)
<wingrime> ssvb: I don't mean speed
<wingrime> ssvb: video are 480p
<ssvb> if audio and video get out of sync, then video is not getting decoded fast enough
<ssvb> or what do you mean by lags?
<wingrime> ssvb: many messages over ssh
<wingrime> [ac3 @ 0x40fc5210]frame sync error
<wingrime> ssvb: 'lag' I mean after mplayer exit, fbconsole still not work
<wingrime> ssvb: mplayer crashes with
<wingrime> *** Error in `mplayer': munmap_chunk(): invalid pointer: 0x43168020 ***
<wingrime> ssvb: framebuffer display last decoded frame
<wingrime> ssvb: if make video out that can do UVY it can reduce overhead (-10% ??) )
Guest95491 is now known as zumbi
atiti_ has joined #linux-sunxi
atiti has quit [Ping timeout: 245 seconds]
cajg_ is now known as cajg
rz2k has joined #linux-sunxi
<atiti_> hm, is there a simple example to use libcedarx?
tinti has joined #linux-sunxi
jlj has quit [Ping timeout: 240 seconds]
atiti_ has quit [*.net *.split]
dragonn has quit [*.net *.split]
bfree has quit [*.net *.split]
ChanServ has quit [*.net *.split]
<oliv3r> atiti there exists examples and we have reference code on github, from aw; but also if you look at vlc that uses libcedarx
atiti_ has joined #linux-sunxi
dragonn has joined #linux-sunxi
jlj has joined #linux-sunxi
<atiti_> does anyone have some documentation of how the cedar workflow works?
<atiti_> i couldnt find too much in the A10/A13 user guides
<atiti_> thx, thats kinda low level :P
<atiti_> im looking at using libcedarx for now
<oliv3r> atiti check your irc history :p
<oliv3r> atiti there exists examples and we have reference code on github, from aw; but also if you look at vlc that uses libcedarx
shineworld has quit [Quit: Leaving]
<atiti_> oliv3r: ye, ive been looking at vlc/xbmc cedar code, but didn't know if there is any other reference code somewhere
<oliv3r> there is some in the libcedar blob github folder
<atiti_> which repo?
<oliv3r> google/wiki are your friends :)
<oliv3r> should have all our stuff
<atiti_> ah nice
<atiti_> thx
bfree has joined #linux-sunxi
egbert_ is now known as egbert
<atiti_> and libcedar is basically a wrapper around libve?
Black_Horseman has joined #linux-sunxi
<oliv3r> i belive so
specing has quit [*.net *.split]
ssvb has quit [*.net *.split]
wigyori has quit [*.net *.split]
arete74 has quit [*.net *.split]
oliv3r has quit [*.net *.split]
drachensun has quit [*.net *.split]
naobsd has joined #linux-sunxi
vrga has joined #linux-sunxi
<vrga> is there any particular multiple that the cpu speed can be for an A13?
<andoma> in showtime (my mediaplayer) i skipped libcedar completely
<andoma> it's just a pile of crap imho
naobsd has quit [Changing host]
naobsd has joined #linux-sunxi
rellla has quit [*.net *.split]
torindel has quit [*.net *.split]
cyp has quit [*.net *.split]
mripard has quit [*.net *.split]
libv has quit [*.net *.split]
techn_ has quit [Read error: No buffer space available]
Black_Horseman has quit [*.net *.split]
dragonn has quit [*.net *.split]
atiti_ has quit [*.net *.split]
tinti has quit [*.net *.split]
FR^2 has quit [*.net *.split]
uwe__ has quit [*.net *.split]
e-ndy_ has quit [*.net *.split]
Tartarus has quit [*.net *.split]
dwilkins has quit [*.net *.split]
tkoskine has quit [*.net *.split]
techn_ has joined #linux-sunxi
rz2k has quit [*.net *.split]
\\Mr_C\\ has quit [*.net *.split]
TheSeven has quit [*.net *.split]
Soru has quit [*.net *.split]
mnemoc has quit [*.net *.split]
vrga has quit [*.net *.split]
focus has quit [*.net *.split]
n01 has quit [*.net *.split]
wingrime has quit [*.net *.split]
cajg has quit [*.net *.split]
Dreadlish has quit [*.net *.split]
jlj has quit [*.net *.split]
hark has quit [*.net *.split]
zumbi has quit [*.net *.split]
andoma has quit [*.net *.split]
^vincenzo^ has quit [*.net *.split]
stekern has quit [*.net *.split]
slapin_nb has quit [*.net *.split]
hramrach__ has quit [*.net *.split]
zrafa has quit [*.net *.split]
Tsvetan has quit [*.net *.split]
Liberator has quit [*.net *.split]
xenoxaos has quit [*.net *.split]
techn_ has quit [*.net *.split]
notmart has quit [*.net *.split]
ganbold has quit [*.net *.split]
jelly-home has quit [*.net *.split]
vicenteH has quit [*.net *.split]
rm has quit [*.net *.split]
Turl has quit [*.net *.split]
mturquette has quit [*.net *.split]
lkcl has quit [*.net *.split]
ojn has quit [*.net *.split]
jukivili has quit [*.net *.split]
MadSpark has quit [*.net *.split]
bfree has quit [*.net *.split]
egbert has quit [*.net *.split]
dl9pf has quit [*.net *.split]
steev has quit [*.net *.split]
derethor has quit [*.net *.split]
CaCtus491 has quit [*.net *.split]
mdp has quit [*.net *.split]
awafaa has quit [*.net *.split]
FunkyPenguin has quit [*.net *.split]
naobsd has quit [*.net *.split]
Soru has joined #linux-sunxi
TheSeven has joined #linux-sunxi
jukivili has joined #linux-sunxi
mnemoc has joined #linux-sunxi
tinti_ has joined #linux-sunxi
jlj_ has joined #linux-sunxi
^vincenzo^ has joined #linux-sunxi
lkcl has joined #linux-sunxi
Guest96720 has joined #linux-sunxi
rm has joined #linux-sunxi
\\Mr_C\\ has joined #linux-sunxi
Turl has joined #linux-sunxi
rz2k has joined #linux-sunxi
MadSpark has joined #linux-sunxi
techn_ has joined #linux-sunxi
Tartarus has joined #linux-sunxi
dwilkins has joined #linux-sunxi
xenoxaos has joined #linux-sunxi
Black_Horseman has joined #linux-sunxi
ojn has joined #linux-sunxi
ChanServ has joined #linux-sunxi
rellla has joined #linux-sunxi
andoma has joined #linux-sunxi
atiti has joined #linux-sunxi
dragonn has joined #linux-sunxi
ssvb has joined #linux-sunxi
uwe__ has joined #linux-sunxi
notmart has joined #linux-sunxi
hark has joined #linux-sunxi
arete74 has joined #linux-sunxi
shineworld has joined #linux-sunxi
wigyori has joined #linux-sunxi
mripard has joined #linux-sunxi
Liberator has joined #linux-sunxi
libv has joined #linux-sunxi
FR^2 has joined #linux-sunxi
drachensun has joined #linux-sunxi
ganbold has joined #linux-sunxi
zumbi has joined #linux-sunxi
torindel has joined #linux-sunxi
zrafa has joined #linux-sunxi
tkoskine has joined #linux-sunxi
specing has joined #linux-sunxi
FunkyPenguin has joined #linux-sunxi
vicenteH has joined #linux-sunxi
Tsvetan has joined #linux-sunxi
stekern has joined #linux-sunxi
mturquette has joined #linux-sunxi
hramrach__ has joined #linux-sunxi
slapin_nb has joined #linux-sunxi
e-ndy_ has joined #linux-sunxi
oliv3r has joined #linux-sunxi
cyp has joined #linux-sunxi
<atiti> ah, so do a cedar backed in for example libva that ffmpeg could use
vrga has joined #linux-sunxi
cajg has joined #linux-sunxi
wingrime has joined #linux-sunxi
n01 has joined #linux-sunxi
Dreadlish has joined #linux-sunxi
focus has joined #linux-sunxi
steev has joined #linux-sunxi
CaCtus491 has joined #linux-sunxi
derethor has joined #linux-sunxi
awafaa has joined #linux-sunxi
mdp has joined #linux-sunxi
tinti_ has quit [Ping timeout: 274 seconds]
\\Mr_C\\ has quit []
tinti_ has joined #linux-sunxi
<rellla> straightest way is imo to make vdpau/vaapi take use of cedar.
bfree has joined #linux-sunxi
dl9pf has joined #linux-sunxi
egbert has joined #linux-sunxi
ZaEarl__ has joined #linux-sunxi
notmart has quit [Quit: notmart terminated!]
ZaEarl__ has quit [Ping timeout: 248 seconds]
shineworld has quit [Quit: Leaving]
ZaEarl__ has joined #linux-sunxi
\\Mr_C\\ has joined #linux-sunxi
leowt has joined #linux-sunxi
Guest96720 is now known as jelly-home
hipboi has joined #linux-sunxi
gzamboni has joined #linux-sunxi
<wingrime> jemk: ping
rellla has quit [Quit: Nettalk6 -]
jlj_ has quit [Ping timeout: 276 seconds]
vrga has quit [Quit: Leaving]
MadSpark- has joined #linux-sunxi
MadSpark has quit [Ping timeout: 276 seconds]
jlj has joined #linux-sunxi
paulk-desktop has joined #linux-sunxi
leowt has quit [Quit: leowt]
ChanServ has quit [shutting down]
ChanServ has joined #linux-sunxi
naobsd has joined #linux-sunxi
<Turl> I enabled akismet on the wiki, let me know if you have trouble editing
<wingrime> User account RellaDoNotBanMe (Talk | contribs) was created ‎
<Turl> wingrime: yes that was me :p
<Turl> used it to repost one of the recent spams and it stopped it
<Turl> so I hope this is where the spammers get blocked and then bored of trying
<wingrime> Turl: how this system work?
<wingrime> I hope see Kalman Filter based
<Turl> wingrime: you know wordpress?
tkoskine has quit [Ping timeout: 256 seconds]
<Turl> wingrime: akismet is wordpress antispam
<wingrime> Turl: I only know than worldpress worst engine ever (and best for hacker)
<Turl> haha
<Turl> but their antispam service is widely used
<Turl> and pretty good
<Turl> basically they run a service
<Turl> where you submit the text and they give you "ham" or "spam" rating
<wingrime> Turl: It look like spamer are mortal not bot
<Turl> wingrime: yes, but akismet checks links for example
<Turl> if anyone spammed same link on another blog or site and it got reported, it is blocked here
<Turl> and spammers don't go and spam a link just on a single site :) so it works good
<wingrime> Turl: good idea
<wingrime> Turl: I think thay make google top using that "creap SEO"
<wingrime> *crap
<Turl> wingrime: not really because wiki has rel="nofollow" on external links so google does not give them pagerank :)
<Turl> but they still spam :P
hipboi has quit [Quit: Leaving]
tkoskine has joined #linux-sunxi
atiti has quit [Read error: Operation timed out]
Superpelican has joined #linux-sunxi
_BJFreeman has joined #linux-sunxi
_BJFreeman is now known as BJfreeman
<mnemoc> short-moin
<Turl> morning mnemoc
<Turl> mnemoc: see above for the record, I installed a new extension on the wiki
<mnemoc> i thought i had restricted the nofollow thing in the config
<mnemoc> but if there is an extension to make that smarter, fine for me
n01 has quit [Remote host closed the connection]
<Turl> mnemoc: yeah but I think these spammers have automat brain
<Turl> they probably never ever checked if it has nofollow or not
<Turl> they just spam spam spam :P
<mnemoc> :)
<Turl> mnemoc: btw, the automatic wiki update script seems to have stalled like some days ago
<Turl> I see git is in there so I wouldn't like to kill it and leave repos in undefined state
<mnemoc> let me look
<mnemoc> f* heat. I had forgotten the feeling of >30C....
<Turl> mnemoc: :)
<Turl> mnemoc: high humidity?
<mnemoc> too
<mnemoc> my nomrla summers are 20C with high muditify. but this feels like twice that
<Turl> you need a bigger notebook :)
<Turl> 20C is perfect temperature :P
<mnemoc> i need a large ventilator :p
<Turl> for 30C, yeah
<Turl> mnemoc: /q
notmart has joined #linux-sunxi
notmart has joined #linux-sunxi
<oliv3r> i got a large ventilator :)
<oliv3r> 16" i guess
Superpelican has quit [Ping timeout: 276 seconds]
<oliv3r> it's not in use atm
<oliv3r> Mauro Carvalho Chehab moved from Red Hat to Samsung!
<Turl> oliv3r: that's the media maintainer on linus rants :)
<oliv3r> lol
<oliv3r> yeah i know mauro
<oliv3r> i'm subscribed to linux-media actually
<oliv3r> i've sent a few patches in the past
<Turl> oliv3r: moved to korea?
<Turl> he was in brazil I think
<oliv3r> R&D Director
<oliv3r> Samsung Electronics
atiti has joined #linux-sunxi
Black_Horseman has quit [Quit: Zwi se logou mou!!!]
<oliv3r> paulk-desktop: ping
<paulk-desktop> hi
<paulk-desktop> oliv3r, pong
<oliv3r> oh that was fast :)
<oliv3r> are you the paulk that's listed on the replicant samsung SG2 page?
<paulk-desktop> yeah
<oliv3r> ohh lol
<paulk-desktop> I'm the main Replicant developer
<oliv3r> hi :)
<oliv3r> oh really
<oliv3r> I did not know :)
<oliv3r> but good to know :)
<paulk-desktop> basically I started porting Replicant to an allwinner tab but I stopped because I have to focus on other devices
<oliv3r> how many people are working on the S2?
<paulk-desktop> there is me and the team at #teamhacksung
<paulk-desktop> however they are dropping support for it
<oliv3r> yeah i've been there :)
<oliv3r> they are?
<oliv3r> i'm about to buy one
<paulk-desktop> you won't get new CyanogenMod updates
<oliv3r> because you said to get the S2 over the S3 :)
<paulk-desktop> S3 will be supported by Replicant soon
<paulk-desktop> it should be just as good
<oliv3r> so get an S3 instead?
<paulk-desktop> well I still don't know if the camera will work with free software
<paulk-desktop> and anyway they will drop support for it too
BJfreeman has quit [Ping timeout: 252 seconds]
<paulk-desktop> well, your call :)
<oliv3r> then the 2 it is :p
<oliv3r> so what IS the best replicant device to get then?
<paulk-desktop> oh you want to run replicant ?
<oliv3r> as much OSS as possible
<paulk-desktop> okay
<paulk-desktop> galaxy s2 is best supported
<paulk-desktop> and there is a chance we get its GPS supported too
<paulk-desktop> while it's a dead-end for galaxy s3
<paulk-desktop> so galaxy s2 is probably best
<wingrime> paulk: CM droped less 512 MB phones
<wingrime> paulk: when CM9 Comes out
<paulk-desktop> wingrime, yeah I know
<paulk-desktop> but that's legitimate
<oliv3r> well s2 has 1G
<oliv3r> so if it has gps support
<paulk-desktop> the thing is that the developers complain about Exynos devices
<oliv3r> why drop it?
<oliv3r> i can understand that
<Turl> oliv3r: insignal crap code :p
<paulk-desktop> they claim they don't have enough documentation
<oliv3r> Turl: insignal?
<paulk-desktop> anyway I don't share their views
<oliv3r> so CM drops it, but not replicant? :)
<paulk-desktop> oliv3r, right
<oliv3r> i think they are just shouting mostly
<Turl> oliv3r: insignal does the software for exynos
<paulk-desktop> but it'll make it hard for us to keep up to new versions
<paulk-desktop> Turl, mostly, they write it themselves
<paulk-desktop> (samsung)
<Turl> paulk-desktop: that's the problems
<Turl> the problem*
<paulk-desktop> insignal does the boards
<oliv3r> i guess they hope that samsung releases more code/support by shouting of dropping
<wingrime> paulk-desktop: what problem GT540 have unoffical support even for jb
<paulk-desktop> Turl, are you aware of the exynos issues the hacksung developers complain about ?
<paulk-desktop> Because I don't agree with their conclusions at all
<oliv3r> i'm not, share :)
<Turl> paulk-desktop: the code you get is insignal's and the blobs you get are samsung's, so getting them to play together is a mess
<wingrime> paulk-desktop: so you can use CM as base
<Turl> paulk-desktop: I am, /whois me :)
<paulk-desktop> wingrime, We do use CM as a base
<paulk-desktop> makes things a lot easier for us
<paulk-desktop> we replace low-level stuff
<paulk-desktop> Turl, right
<paulk-desktop> so you're a teamhacksung dev?
<wingrime> paulk-desktop: make AW opensoure much simpler
<oliv3r> lol yeah
<paulk-desktop> sure
<paulk-desktop> but it's not the same
<paulk-desktop> allwinner only has "easy" components
<Turl> paulk-desktop: no, not teamhacksung
<Turl> but I know the guys and have seen their complaints
<paulk-desktop> ie: no modem with unknown protocol, no GPS, etc
<wingrime> paulk-desktop: are you doing something with qcom &
<oliv3r> modem = samsungril
<Turl> paulk-desktop: there's code for all the android bits except mali and cedarx (hw video dec/enc)
<paulk-desktop> oliv3r, guess who wrote that?
<Turl> (on sunxi that is)
<paulk-desktop> Turl, yeah
<paulk-desktop> Anyway Samsung releases kernel sources, that's as much as we should except from them
<wingrime> paulk-desktop: do you have any deal with qcom ril?
<paulk-desktop> if their blobs use a fucked-up API, well, that's legit
<Turl> paulk-desktop: yeah, which works with their blobs and break with insignal's often :(
<paulk-desktop> they are blobs, the problem at the root is software being non-free
<paulk-desktop> wingrime, no, I gave up on qcom phones
<paulk-desktop> but I know they use a binary protocol that wasn't reversed
<oliv3r> paulk-desktop: you :p
<paulk-desktop> oliv3r, right
<wingrime> paulk-desktop: thats starange, i think thay use sun remote procedure call
<Turl> paulk-desktop: how do you handle GL on newer android versions on replicant?
<paulk-desktop> Turl, we have tweaks to make it work without hw EGL
<paulk-desktop> and it's pretty good
<oliv3r> so with gps being 'almost', and samsungril being worked on continuesly, what is there that's 'badly' supported right now?
<oliv3r> (well mali, duh')
<wingrime> paulk-desktop: witch RTOS use samsung modem core ?
<paulk-desktop> oliv3r, GPS is far from almost, it's closer to early stages, but everything will depend on how much doc we can get
<paulk-desktop> wingrime, I didn't check that, we don't intend to replace the modem's firmware
<paulk-desktop> but they surely modified it from intel to implement their samsung ipc protocol
<oliv3r> 'how much doc we can get' so is there anything being worked out in that regard?
<paulk-desktop> not currently, but I started weeks ago
<oliv3r> ok, good good
<oliv3r> what phone do you own/use?
<paulk-desktop> and I wrote to CSR (who bought SiRF) about it
<paulk-desktop> to ask for doc and they simply refused
<wingrime> paulk-desktop: I can't consider phone free when that phone modem core can do what he want ,
<oliv3r> oh :(
<paulk-desktop> the galaxy s2 GPS chips is a CSR GSD4t
FR^2 has quit [Quit: Connection reset by peer]
<paulk-desktop> same as galaxy nexus btw
<wingrime> paulk-desktop: with full access to system memory
<paulk-desktop> wingrime, right, that's why we try to find out whether it can access the phone's memory, gps, audio, etc
<paulk-desktop> and no, on galaxy s2, it doesn't access system memory
<paulk-desktop> it's connected to the AP over HSIC, which is high-speed USB
<paulk-desktop> on the galaxy s, it was over shared memory
<wingrime> paulk-desktop: QCOM, use shared memory , witch means modem have full access
<paulk-desktop> wingrime, that's for sure
<paulk-desktop> that's why we dropped QCOM phones
<paulk-desktop> Samsung phones are much better on this regard
<oliv3r> paulk-desktop: so what phone do you own/use?
<paulk-desktop> oliv3r, sorry, skipped the question
<paulk-desktop> I have that:
<paulk-desktop> * nexus s, galaxy s, galaxy s2, galaxy tab 2 10.1 3g, htc dream, nexus one, geeksphone one, the allwinner tab, and galaxy s3
<paulk-desktop> and gta02 and gta04
<paulk-desktop> I was using galaxy s2 daily for long
<oliv3r> but switched to s3?
<paulk-desktop> just today I switched to galaxys3 with replicant
<oliv3r> ah :)
<oliv3r> go use the s2
<paulk-desktop> to test whether telephony is stable
<oliv3r> it's better ;)
<paulk-desktop> in what regard?
<paulk-desktop> for daily use, I don't care much
<wingrime> paulk-desktop: you should think about AW, we now have workable jpeg decoder poc
<oliv3r> that you'll be using the same one as I will :)
<paulk-desktop> wingrime, don't worry, we thing about it
<paulk-desktop> think*
<oliv3r> i wouldn't worry too much :)
<oliv3r> don't spread your resources :)
<oliv3r> i mean, we'd welcome a good developer here of coruse
<paulk-desktop> thanks
<oliv3r> but with more and more being supported via us here (etc)
<oliv3r> once a fully free implementation is done
_BJFreeman has joined #linux-sunxi
<oliv3r> having support in replicant sounds like a piece of cake
<paulk-desktop> I will add replicant support for my a13 tablet at some point anyway
<paulk-desktop> and others are interested in allwinner
<oliv3r> it should be really easy imo
<oliv3r> with the sunxi-bsp
<oliv3r> you allready get a nice hwpack
<paulk-desktop> most of the work is to integrate android
<oliv3r> that has all the files that are allwinner 'unieque'
<paulk-desktop> which I've already pretty much done
_BJFreeman is now known as BJfreeman
<oliv3r> and i think turl did the CM - allwinner work
<oliv3r> so combining the sunxi-bsp, tur's allwinner repo and replicant shouldn't take much
<paulk-desktop> we'll have matter to discuss when I get back to working on allwinner
<wingrime> paulk-desktop: CM on a10 fully workable
<paulk-desktop> sure
<paulk-desktop> ah also, quick technical question
<paulk-desktop> is there an ADC to detect jack plug?
<paulk-desktop> I had the impression that was done by the hardware (sound routing)
<wingrime> paulk-desktop: I have impression that plug have harware switch
<paulk-desktop> right
<paulk-desktop> that's my impression too
<paulk-desktop> but it's a shame
<paulk-desktop> because we can't regulate the volume properly
<wingrime> paulk-desktop: china have intension make all stuff cheap, but we have GPIO
<paulk-desktop> usually, we lower the volume when going into headphones
<paulk-desktop> here it can't be done since there is no detection mechanism
<oliv3r> current sunxi-alsa driver works, but needs a rewrite
<paulk-desktop> so it fucks up your ears
<oliv3r> hehe
<oliv3r> i'll check the sound driver and datasheet if there's such a detection circuit
<oliv3r> i can't remember
<paulk-desktop> thanks
<oliv3r> and if you design something, you could add a gpio for that
<wingrime> paulk-desktop: It depend tablet vendor, becose use GPIO are realy easy for detection
<oliv3r> but right now, nada
<paulk-desktop> ok
<oliv3r> i'll check paul
<paulk-desktop> also the Replicant co-developer is really interested in reverse engineering cedarx
<oliv3r> you go use S2 as main development target :p
<oliv3r> cause i know i'm getting an old used phone, but i'm getting it so it's properly supported with oss stuff
<oliv3r> i don't care that it's a little slower
<paulk-desktop> oliv3r, best I can do is to add GPS support
<oliv3r> it looks like a fine phone (i got a white one for my gf allready) :)
<paulk-desktop> the rest is there
<oliv3r> what else do we need at that point ;)
<oliv3r> libril will slowly improve i'm sure
<paulk-desktop> and for the RIL, it's not device-specific
<paulk-desktop> right
<paulk-desktop> but if it improves for s3 it'll improve for s2 as well
<oliv3r> and firmware blobs ... who knows maybe we get lucky
<wingrime> paulk-desktop: we have strange GPS hardware support but now one saw it
<paulk-desktop> I doubt so
<oliv3r> and mali, we have libv in the far future
<wingrime> * no noe
<paulk-desktop> right
<wingrime> * no one
<oliv3r> ok, research mode :)
<paulk-desktop> for firmwares, I use them when I need them
<paulk-desktop> I know they are nonfree but lots of other firmwares run in chips
<paulk-desktop> we just don't distribute them
<wingrime> paulk-desktop: we have usualy usb connected wifi sometime with firmware
<paulk-desktop> wingrime, yeah I know
<wingrime> paulk-desktop: also TS can go with firmware , but it not nessesory
<paulk-desktop> but wifi's not the only thing
<paulk-desktop> wingrime, that means it's not loaded by the CPU but it's there, in the chip
<paulk-desktop> and since replicant is about free software, we don't distribute the firmwares
<wingrime> paulk-desktop: yes, ts ic - sime mcu with firmware
<wingrime> *simple
<paulk-desktop> for instance my ektf2k has a firmware pre-installed
<paulk-desktop> (that's the touchscreen controller)
<wingrime> I have "zet" touchsceen , and vendor supricely give us gpl source by request
<paulk-desktop> driver source I guess
<paulk-desktop> not firmware source
<wingrime> yes
<wingrime> but after I make replacemet
<paulk-desktop> turns out I found the missing bits of the driver I needed only right after reverse engineering it…
<wingrime> paulk-desktop: TS firmware are very related glass configuration
<wingrime> paulk-desktop: so it uniqe for tablet
<paulk-desktop> yep
<wingrime> paulk-desktop: "zet" vendor according their user manual have some windows app for low-level firmware config
<paulk-desktop> huh ok
<wingrime> paulk-desktop: so use universal thing not work
<paulk-desktop> that doesn't mean free firmwares aren't possible
<wingrime> paulk-desktop: thay use some MCU-core so it can be disassembled and rewrited
<wingrime> paulk-desktop: thay can use very simple MCU like intel 51
<paulk-desktop> ok
<paulk-desktop> I know it's possible, it's just a bit hard
<paulk-desktop> I'm not that good at reading ASM already
<paulk-desktop> I hope I'll have classes about it when I get in university
<wingrime> paulk-desktop: But witch means , we have only 3 - problem 1) wifi (sometimes) 2) cedarx 3) mali
<paulk-desktop> right
<paulk-desktop> about cedarx, the replicant co-developer is interested in reverse engineering it
<wingrime> paulk-desktop: he can join channel anytime
<paulk-desktop> I think some work has been done already
<paulk-desktop> do you think it's doable?
<wingrime> paulk-desktop: we have tracer
<paulk-desktop> generally speaking (by looking at samsung devices) it's not a big deal
<wingrime> paulk-desktop: take a look to my results
<paulk-desktop> ok
<paulk-desktop> how is the kernel driver also?
<wingrime> paulk-desktop:
<paulk-desktop> oh yeah, nice
<wingrime> paulk-desktop: I get some short register names
<wingrime> paulk-desktop: from blob
<paulk-desktop> nice
<wingrime> paulk-desktop: but 100% only for mpeg engine and VC1 and some h264 engine
<paulk-desktop> what's wrong with the rest?
<wingrime> paulk-desktop: thay use structures for that regsiters
<wingrime> paulk-desktop: thay put into structure all registers in some engines
<wingrime> paulk-desktop: but new code use macroses
Superpelican has joined #linux-sunxi
<paulk-desktop> so it's harder to figure out?
<paulk-desktop> anyway it seems that you have seriously started the work
<wingrime> paulk-desktop: by trace and thay use copy plaste for " engines" I mean engines have simular regs placment sometimes
<wingrime> paulk-desktop: we have tracer that can trace all access to registers
<paulk-desktop> ok
<paulk-desktop> that's nice
<paulk-desktop> it works?
<wingrime> paulk-desktop: yes
<paulk-desktop> nice
<paulk-desktop> that's a good step
<wingrime> paulk-desktop: but it still have no real usage
<paulk-desktop> ok
<paulk-desktop> that can be used for camera implementation
<paulk-desktop> on replicant for instance
<wingrime> paulk-desktop: we must make some code for Stagefright to make it fly, but It not clear will it work faster that software decoding , becose engine used generaly for mjpeg
<wingrime> mjpeg= many jpegs
<oliv3r> doesn't stagefright use libjpeg-(turbo) as backend?
<wingrime> oliv3r: it have many backends for software decoding
<wingrime> oliv3r: but it clear , that some devices on android use jpeg hw accel
<paulk-desktop> wingrime, not on camera, it doesn't use stagefright
<Turl> spammers keep signing up but at least now they don't spam :D
<paulk-desktop> basically most of the time the images out of v4l2 come as ycbcr
<paulk-desktop> and you need to quickly convert to jpeg
<Turl> paulk-desktop: most use libjpeg-turbo for that
<paulk-desktop> ok
<Turl> (jpeg encoding that is, not sure if it handles ycbcr or not)
<wingrime> paulk-desktop: in qcom case it done using hw + modem core
<paulk-desktop> yes I remember
<paulk-desktop> it was already like that on htc dream
<paulk-desktop> exynos has an embedded jpeg converter
<paulk-desktop> so we use that on samsung android phones
<oliv3r> wingrime: well as I said earlier, i really wanna port jpeg poc ->
<oliv3r> if we change to do hardware decoding, all libjpeg 'users' are accelerated
<wingrime> oliv3r: good intesion
<oliv3r> paulk-desktop: i saw s2 uses the 'mfc' firmware for media accelerator
<oliv3r> is that exynos 'cedar' or is that the arm one
<paulk-desktop> mfc is what cedar is to allwinner
<paulk-desktop> but the jpeg thing is something else that is in the SOC
<paulk-desktop> oliv3r, I don't think it is Cedar
<Turl> paulk-desktop: and what's fimc?
<paulk-desktop> Turl, that one I tried to figure out for long
<wingrime> paulk-desktop: we need many people here for cedar , becose we need rewite driver for mainline, all decoding stuff must go to kernel and vaapi
<paulk-desktop> not any precise hardware component it seems
<paulk-desktop> wingrime, I understand
<Turl> paulk-desktop: can it be the equivalent to G2D?
<paulk-desktop> what's G2D?
<Turl> paulk-desktop: 2D accelerator
<Turl> blitting and that kind of stuff
<wingrime> paulk-desktop: cedar have 6 engines so we need support them
<paulk-desktop> ah ok. No I don't think so
<paulk-desktop> wingrime, ok
<paulk-desktop> I'll tell that to GNUtoo who's the one interested in doing it
Tartarus has quit [Quit: ZNC -]
<oliv3r> better focus development then dupilicate spread it?
Tartarus has joined #linux-sunxi
<paulk-desktop> yeah
<wingrime> oliv3r: thats why I focus on wiki with registers more than coding
<wingrime> oliv3r: oliv3r: you also can help, if you write manual for new tracer
<oliv3r> wingrime: yeah, wiki (documentatino) is very important now
<oliv3r> i spy every now and then :)
zrafa has left #linux-sunxi [#linux-sunxi]
<wingrime> oliv3r: also do you like new TOC style (wiki replacment for registers)
<wingrime> ?
Tartarus has quit [Ping timeout: 248 seconds]
<oliv3r> yeah it's good :)
<Turl> great, i2c landed on mainline
<oliv3r> Turl: sweet; did you read my message to greg?
<Turl> I think the only missing thing now for us is the emac patches
<Turl> oliv3r: yes, saw it some minutes ago
Tartarus has joined #linux-sunxi
<wingrime> Turl: whats going on with mmc and dma drivers
<wingrime> ?
<wingrime> in mainline
<Turl> wingrime: I haven't ever heard back from the guy doing the DMA driver, mripard might know more :)
<Turl> wingrime: mripard is working on SDIO driver now
<wingrime> but mmc most important
<Turl> wingrime: isn't SDIO MMC?
<wingrime> Turl: I always thinking SDIO - wifi over mmc connector
<Turl> maybe, I'm not sure
<Turl> let's have mripard clarify :P
<Turl> wingrime: do you want to write a useful driver?
<Turl> nobody is doing the AXP drivers
<Turl> hans submitted a pretty clean driver to linux-sunxi for one of the new axp you could use as a base
Superpelican has quit [Quit: Konversation terminated!]
<atiti> i wanna write a useful driver
<wingrime> Turl: good idea in case that I not see working batt indication on tablet (kpower work strange)
Tartarus has quit [Read error: Connection reset by peer]
<oliv3r> Turl: hushy wingrime is busy RE-ing :)0
<wingrime> oliv3r: cedar very longterm project
<wingrime> also how we will call new driver for cedar?
<mripard> Turl: wingrime: SDIO is storage + extensions to support devices
<mripard> so it covers both what you were talking about actually :)
<mripard> and yep, no news from matt.
<mripard> I guess everyone wanting to start working either on DMA or SPI can start
<wingrime> mripard: dma are realy important
<mripard> we've waited too long for this already.
<mripard> wingrime: I know.
<mripard> but there's too much important things to do
<mripard> Turl: wanting to do DMA (or SDIO, I didn't really started working on it)
<mripard> +?
<oliv3r> paulk-desktop: it looks like there is 'something' in hardware, from the a13 usermanual:
<oliv3r> HPCOM output protection enable
<Turl> mripard: not for now, go ahead :)
<oliv3r> paulk-desktop: with a debounce delay, so I think it does the volume limiting in hardware
<oliv3r> wingrime: cederus :)
<Turl> mripard: I still have the clocks patches sitting there
<mripard> but yeah, I guess the next items to tackle are A20/A31/DMA/MMC
<oliv3r> wingrime: cedar is a tree, cederus is the latin name, mnemoc came up with it :)
<Turl> mripard: I'll work on A20 when I get my cubie :)
<oliv3r> mripard: a31 u-boot is a LOT of work
<mripard> but I don't have enough brain/hands to work on all of this.
<oliv3r> but p2mi driver is done and pmu pretty much
<mripard> oliv3r: I was talking mostly about the kernel, but it sure is :)
<oliv3r> and i've asked greg how to do sid, cause i don't know
<wingrime> mripard: I understand how dma work, but I have not understand dma-engine api
<mripard> wingrime: I have the same problem currently :)
<Turl> oliv3r: I'd go with "not possible", at least with that way :)
<oliv3r> wingrime: AXP is really important too and quite independant
<Turl> oliv3r: it really expects a show and write op pointer under the attr
<oliv3r> Turl: show and store
<oliv3r> but i want read and writ e;)
<oliv3r> i'll rewrite it to the mtd layer otherwise
<Turl> wingrime: nothing that can't be fixed with a bit of reading :)
<Turl> oliv3r: I was thinking maybe a soc bus driver would be a perfect fit for sid
<mripard> Turl: actually, there's not much to read
<oliv3r> dma is reasonably well documented
<wingrime> mripard: I can wite 'custom' api for dma but I don't think it ever acepted
<mripard> wingrime: yep, it has to be with dmaengine
<oliv3r> Turl: soc bus driver?!
<wingrime> mripard: current aw dirver use callbacks for end of work (In IRQ context!!! crap) and emac use it for next transmit((
<wingrime> it totaly crap
<wingrime> mripard: how much glue code we need for make usb work?
<oliv3r> wingrime: you mean usb-otg?
<wingrime> mainline usb
<Turl> oliv3r: see Documentation/ABI/testing/sysfs-devices-soc
Tsvetan is now known as TsvetanAway
<Turl> soc_id could be SID
<oliv3r> Turl: ah, hmm, maybe your right :)
<oliv3r> so does socbus bring lsusb?
_BJFreeman has joined #linux-sunxi
<Turl> lsusb? what?
BJfreeman is now known as Guest35572
_BJFreeman is now known as BJfreeman
<paulk-desktop> oliv3r, thanks
<Turl> oliv3r: you can see implementations in arch/arm/mach-omap2/id.c
<oliv3r> lssoc*
<oliv3r> Turl: rgr
<oliv3r> Turl: soc ID is sometimes empty
<Turl> arch/arm/mach-ux500/cpu.c
Guest35572 has quit [Ping timeout: 252 seconds]
<Turl> oliv3r: or you could keep sid driver as is and get a free internal api user with this :)
<wingrime> and where mnemoc?
<Turl> wingrime: he's charging his phone :)
<wingrime> Turl: you can begin merge some patches from mail list
<oliv3r> Turl: i think the eeprom driver allready lives under the soc tree though?
<Turl> oliv3r: yes, I think so
pirea has joined #linux-sunxi
<Turl> oliv3r: do you want to write the soc bus stuff? I can do it otherwise
<pirea> hy all
<pirea> who want to help me to make a open cedarx and open libve?
<wingrime> pirea: are you from replicant?
<Turl> pirea: ^ this guy :)
<pirea> i am not from replicant
<wingrime> pirea: base info here
<wingrime> pirea: take a look
<oliv3r> Turl: it's probably a good idea to port SID to SocID
<oliv3r> Turl: i'll drop sid and change it to soc bus stuff
<oliv3r> Turl: can 2 drivers access 1 register?
<wingrime> Turl: can you begin aprove some patches from maillist while mnemoc busy ?
<Turl> wingrime: he should be back later today
<wingrime> Turl: I very hope
<Turl> oliv3r: if you use of_iomap, yes, but if you're writing you'll need a lock
<Turl> if you use the more elaborate stuff I don't think you can map the same thing twice, I'll let mripard correct me on this
<oliv3r> Turl: need to write yeah; ok good
<oliv3r> Turl: if socID is blank (or wrong) i'll use sramc id method
<pirea> wingrime i saw that information
<oliv3r> pirea: step 1) documenting, tracing, replying traces and Prove of concept code
<mripard> wingrime: I have no idea. But you can look at other layers to get an idea
<pirea> oliv3r okey :)
<mripard> oliv3r: Turl: no, two drivers should not access the same registers (for obvious reasons)
<oliv3r> mripard: but what if you absolutly have to.
<oliv3r> (not sure if the kernel needs to write to sramc)
<mripard> like, when exactly?
<pirea> a10 can decode 1080p with neon?
<pirea> oliv3r?
<oliv3r> pirea: noway
<pirea> why?
<wingrime> slow memory
<Turl> pirea: can nexus S decode 1080p woth neon?
<Turl> with*
<pirea> idk
<wingrime> pirea: what extractly you want/can do?
jlj has quit [Ping timeout: 248 seconds]
<pirea> i want do make an open cedarx lib :)
<wingrime> pirea: are you have any deal with ARM ASM?
<pirea> wingrime i know kow to read arm asm
<pirea> but
<pirea> i don't understand disassembled code
<wingrime> pirea: now we have figure bit fields for mpeg regs
<wingrime> pirea: it mean find relations with video types / frame types
<pirea> wingrime i know how to disassmebled all libve object files
<wingrime> pirea: I can help with it
<wingrime> * You
jlj has joined #linux-sunxi
<pirea> wingrime is a little problem
<pirea> i dont have net at home
<pirea> my only one device with net is a blackberry 8520
<wingrime> pirea: you can do it with pice of paper
<pirea> :)
<pirea> i cant connect with that device on irc
<pirea> just mail
<atiti> tether your blackberry to your pc
<pirea> atiti doesn't work
<pirea> my internet option is BIS
<pirea> and doesn't work with pc
<pirea> only with phone
<pirea> but i can recive mails
<wingrime> pirea: you can write results here
<wingrime> pirea: it realy not a problem
<pirea> wingrime if you can help me you can give me your mail?
<wingrime> pirea: I have better solution
<pirea> what?
<wingrime> pirea: linux-sunxi mail list
<pirea> where is that mail list?
<pirea> :)
<wingrime> oliv3r: ^
<pirea> the solution is not ok
<wingrime> pirea: mail list realy good for offline disscusion over email
<pirea> i will try
<oliv3r> pirea: has a page about it :)
<pirea> i found it
<wingrime> pirea: from what country are you?
<pirea> romania
<pirea> wingrime what toolchain are you use?
<oliv3r> doesn't romania have best internet of europe :)
<pirea> oliv3r yes :D
<oliv3r> so where is your home interne t;)
<pirea> i complicated at me
<pirea> :D
<oliv3r> how are you on the internet now? :)
<pirea> i am in a bar
<pirea> pub
<pirea> something like this
<pirea> :P
<pirea> but is away from me
<pirea> i can't come here dayli
<wingrime> pirea: have you any a10/a13 hw?
<pirea> yes
<pirea> a cubieboard
<pirea> wingrime you know how to give g2d more memory?
<oliv3r> pirea: boot flag
<pirea> from script.bin doesn't work
<oliv3r> kernel reservation or kernel flags it he name of the page
Black_Horseman has joined #linux-sunxi
Black_Horseman has joined #linux-sunxi
<wingrime> pirea: current task is anyway run many files on cedarx and make traces / figure how regs bits realated with file
<pirea> let away cedar
<pirea> :)
<pirea> openVG libs for mali400 exists?
tinti_ has quit [Ping timeout: 264 seconds]
jlj has quit [Remote host closed the connection]
tinti_ has joined #linux-sunxi
jlj has joined #linux-sunxi
<pirea> wingrime how to compile code to use thumb2 + arm instructions?
<pirea> with -mtumb-interwork ?
<pirea> and -marm -mthumb?
<wingrime> yes
hglm has joined #linux-sunxi
<wingrime> we have page for Tumb2 on wiki
<pirea> i have readed that
<hglm> pirea: thumb2 is default when you compile C code on Debian/Ubuntu
<pirea> hglm i use archlinux arm crosscompiler
<hglm> ok, -mthumb should work. But it may depend on another things in archlinux userland.
<oliv3r> -march=arm -mtune=thumb2
<pirea> i can't understand that now :)
<hglm> oliv3r: I don't think that really makes sense, thumb2 is a different instruction set.
<oliv3r> pirea: arch should/might have crosscompilertoolchain, i know gentoo does :p
<pirea> oliv3r i will use gentoo soon
<pirea> on my cubieboard
<oliv3r> hglm: your right:
<oliv3r> lists it all :)
<oliv3r> export CFLAGS=”-mthumb -march=armv7-a”
<hglm> But you have to be careful that your userland environment knows how to deal with Thumb2 code.
<oliv3r> optimizing to thumb2 is something you would wanna do if 'normal' mode owrks stabaly
<ssvb> pirea: reserving memory for g2d does not have any use at the moment (in linux)
<hglm> ssvb: I thought so, so you disable it and still enable G2D in sunxifb I guess. Must be a relic from Allwinner Android.
<ssvb> hglm: g2d has some sort of a weird memory allocation ioctl api, which is simply too bad to be used
<pirea> ssvb i use g2d accel
<ssvb> pirea: with what software?
<pirea> ssvb why cpu is at 100% when i use gpu?
<pirea> ssvb sunxifb
<ssvb> because we are only using g2d to accelerate the operations inside of the framebuffer, it does not need to touch/reserve g2d memory
<ssvb> pirea: ^
<ssvb> pirea: it's a bit complicated, not everything can be done with the gpu
<pirea> ssvb i not about g2d
<pirea> is about when xbmc run on framebuffer with opengles libs
<pirea> :|
<pirea> cpu is used too much
<pirea> on raspberry pi this is not happening
<hglm> pirea: I don't think xbmc can use opengles for video acceleration, you are probably seeing xbmc using 100% CPU video decode.
<ssvb> pirea: I guess it just uses xbmc in a different way
<pirea> hglm this is happening when i am not playing videos
<ssvb> pirea: are you sure you are using this XBMC: ?
<pirea> yes
<hglm> Maybe the Mali driver does busy-waiting when executing SwapBuffers waiting for vsync.
<ssvb> pirea: in any case, seems like something is really broken in your setup
<pirea> ssvb idk
<ssvb> pirea: is it at least fast?
<ssvb> I mean, do you only worry about the cpu usage?
<pirea> version r3p2 where is?
<pirea> ssvb yes
<pirea> everytime when i use gpu cpu is used
<ssvb> pirea: there are no binaries for r3p2 from allwinner
<pirea> and all task managers are not showing corect cpu usage on processes
<oliv3r> wasn't there a bug a while ago with high CPU usage
<oliv3r> should be on mailing list
<ssvb> hglm: the framebuffer mali drivers have zero cpu usage when waiting for vsync and running simple gles demos
<hglm> pirea: Maybe you are seeing the effect of CPU frequency scaling. If little CPU is used, CPu runs at 60 MHz or 400 Mhz or whatever your cpufreq configuration is, giving erreonous CPU load stats.
<pirea> hglm cpu is set on performance
<ssvb> how do you know that 'all task managers are not showing corect cpu usage'?
<ssvb> maybe you are imagining your problem? ;)
<pirea> ssvb idk how to explaing that
<pirea> :)
<ssvb> you can use 'htop' tool to get the basic cpu and memory usage numbers
<pirea> if i am using htop or top or gnome-system-monitor or xfce4-task-manager or everything
<pirea> cpu usage for processes is 0
<ssvb> sounds good
<pirea> is not ok
<pirea> in that bar
<pirea> where is showed all cpu usage is not 0
<pirea> is more than 0
<ssvb> how much more than 0?
<pirea> 60-70
<ssvb> can you run 'perf top'?
<pirea> all processes is on 0
<pirea> i will make some screenshots if i can
<pirea> :D
<ssvb> can you run 'perf top'?
<pirea> now no
<ssvb> I guess this sucks
<pirea> ssvb give me your mail
<pirea> one of that processes have cpu usage -1%
<oliv3r> pirea: you can always discuss onthe mailinglist :)
<pirea> olive3r i will try
<pirea> oky
<ssvb> using the right tools helps
<pirea> oky
<pirea> ssvb is nowere openVG libs?
<pirea> on isn't
<ssvb> pirea: here is the list of the mali blobs, which are available for sunxi devices -
<ssvb> openvg does not get much love
tinti_ has quit [Ping timeout: 276 seconds]
<ssvb> pirea: what kind of software are you going to run with OpenVG?
<hglm> ssvb: Did you know that ARM7MP (i.e. Cortex A7 used on A20) has a "Preload to write" instruction? Maybe that's what is required to get optimal memory throughput on A20.
<pirea> ssvb a little jpeg decoder
<ssvb> hglm: yes, such instruction exists
<ssvb> hglm: I have seen some people who managed to use it to regress the performance significantly on Cortex-A9 :)
<ssvb> hglm: that's what happens when one is coding from the book and doesn't verify the results
<hglm> ssvb: I guess so, preload stuff always needs to be benchmarked carefully.
<ssvb> hglm: they added this instruction to ISA (probably as reserved for future use), but none of the real CPU cores manages to benefit from it
<ssvb> pirea: how does OpenVG help for jpeg decoding?
<hglm> ssvb: So it does not help when implemented in (say) memset? Maybe it depends on the particular details of the L2 cache architecture.
<pirea> ssvb idk how to explain
<pirea> my english is limited
<pirea> :_
<ssvb> hglm: PLDW does regress memset performance significantly
<hglm> ssvb: Good to know that.
<ssvb> hglm: MIPS has PrepareForStore prefetch, but it is a bit dangerous because it discards the data from the cache line
<ssvb> hglm: so one needs to know the exact cache line size in order not to corrupt data :)
tinti_ has joined #linux-sunxi
<pirea> ssvb userland libs can be modified for mali400?
<ssvb> pirea: sorry, it looks like I'm having problems trying to understand what you want to fix or improve
<pirea> ssvb know that :)
<pirea> now just to make open libve/cedarx libs
<pirea> :)
<oliv3r> many today that want to work on that today :)
<oliv3r> pirea: check out todays irc log :) (se title)
<ssvb> pirea: linux-sunxi wiki has most of the information you need to get started
hramrach__ has quit [Remote host closed the connection]
hramrach__ has joined #linux-sunxi
<pirea> cya
pirea has quit [Quit: Leaving]
tinti_ has quit [Quit: Leaving]
<oliv3r> mripard: oh nice e-mail :D but what about turls idea, socID for sid
<mripard> that would just be a customer of it
<oliv3r> ohh
<mripard> you'd have a SID driver in drivers/MTD
<oliv3r> mtd driver with soc ID 'using' it
<oliv3r> but that means it can't be a module, can it?
<mripard> and your socID driver saying "hhhm, I want that range of the SID"
<mripard> and that's it
<mripard> (and you don't have to worry about the dual-mapping thing you were talking about :))
<mripard> oliv3r: ah. I haven't really looked at the exact SoC bus details
rellla has joined #linux-sunxi
<mripard> but yeah, possibly.
<oliv3r> mripard: also true
<oliv3r> actually, dual mapping might notbe avoided though :)
<mripard> why?
<mripard> sram?
<oliv3r> yeah
<oliv3r> soc ID wants socID (other then sid)
<oliv3r> and unless we have some sram controller
<oliv3r> it might want it too
<oliv3r> i don't think anybody uses the sramc though
<oliv3r> well not register 24 anyhow
<mripard> yeah, and we can just use of_iomap
<mripard> without requesting the region
<mripard> just like we do for the watchdog
<oliv3r> so it might not be a huge issue
<oliv3r> but depending on greg's answers to both our mails
<oliv3r> i'll end up rewriting to mtd
<oliv3r> shouldn't be too hard
<oliv3r> i'll start maybe this weekend, mostlikly monday
_BJFreeman has joined #linux-sunxi
BJfreeman is now known as Guest46467
_BJFreeman is now known as BJfreeman
Guest46467 has quit [Ping timeout: 252 seconds]
Yaku-noob has joined #linux-sunxi
<rellla> Turl hates me ;(
<rellla> this username made me nervous until i backread logs.
<Yaku-noob> hey there, does someone know if there is a forum with more information to the xbmc/lamp build of martin wild, i was looking if there is cvbs support in that xbmc image and how well it might work
<Yaku-noob> <- mele a2000 user
<rellla> haha, well done, Turl. But not very effective as i can see :(
<Turl> rellla: :P
<Turl> spam is back? D:
wingrime has quit [Ping timeout: 264 seconds]
<Turl> rellla: I don't see any spam
<rellla> at least a few registrations
<Turl> rellla: yeah, akismet won't help wrt registrations
<Turl> rellla: ban at will :D
<Turl> rellla: it's a nice activity to relieve stress :P
<rellla> Turl: :)
<Turl> rellla: at least they don't write spam on pages so far :) let's hope it remains effective
<rellla> nevertheless i think, accounts should be banned after a week with zero edits
<Turl> +1
<Turl> I linked to a script to do that the other day
<Turl> I lost the link then :P
<Turl> we would have to back up database before running though, those things tend to be fragile
<rellla> backreading is also good for stress
<rellla> :)
<Turl> :)
<Turl> I'll be back later
<rellla> oh yeah backup. i didn't have some of a joomla website. some indonesian hacked it ;p cu
<rellla> Yaku-noob: try it here or on the mailing list
<Yaku-noob> well that´s why i just asked directly, for android i have 1-2 threads where i follow how bad the xbmc ipa is working, but now i would really like to give the image a try, but i don´t have a flat tv i need cvbs so i don´t know if i would still have a picture, but a server i guess after i install the image
<Yaku-noob> xbmc and the lamp sounded tempting
BJfreeman has quit [Ping timeout: 252 seconds]
Yaku-noob has quit [Remote host closed the connection]
vrga has joined #linux-sunxi
e-ndy_ has quit [Read error: Operation timed out]
rellla has quit [Quit: Nettalk6 -]
e-ndy has joined #linux-sunxi
rzk has joined #linux-sunxi
rz2k has quit [Ping timeout: 264 seconds]
<vrga> has anyone seen somebody run a gentoo rootfs for these things?
rzk has quit [Ping timeout: 264 seconds]
notmart has quit [Quit: notmart terminated!]
vrga has quit [Quit: Leaving]
hglm has quit [Quit: leaving]
dragonn has quit [Read error: Operation timed out]
dragonn has joined #linux-sunxi
dragonn has quit [Client Quit]
dragonn has joined #linux-sunxi
awafaa has quit [Remote host closed the connection]
ojn has quit [Remote host closed the connection]
steev has quit [Remote host closed the connection]
* Turl spots a bug on mripard's changelog
<Turl> oliv3r: btw, when/if you send a new version of the sid driver, add a node to the a10s dtsi :)
<Turl> TsvetanAway: I suppose you're not here right? :)
paulk-desktop has quit [Quit: Ex-Chat]
jlj has quit [Ping timeout: 264 seconds]
jlj has joined #linux-sunxi
naobsd has quit [Quit: Page closed]
jlj has quit [Ping timeout: 276 seconds]
jlj has joined #linux-sunxi