anunnaki has quit [Remote host closed the connection]
<mnemoc>
I'm moving dl.linux-sunxi.org to another server, please don't upload anything via sftp/scp for now
<hno>
hozer, the 1 during boot for USB NAND access is part of the Allwinner bootloader, not u-boot. be warned however that II have found that this USB drive emulation is not entirely stable.
<hno>
often not really writing the changes to NAND.
<hno>
hi mnemoc
<mnemoc>
hno: hi, when you have a minute can you test if you can ssh to linux-sunxi.org and sudo ?
<hno>
ssh works, but don't remember what my sudo password is.
<hno>
mnemoc^
<mnemoc>
hno: /q
<hno>
thanks. verified.
<mnemoc>
great
<hno>
wow, mtd drivers have emerged for both u-boot and linux-3.4. Nice.
* hno
have been dormantly lurkering for too long.
paulk-desktop has joined #linux-sunxi
jelly has quit [Read error: Connection reset by peer]
tzafrir has joined #linux-sunxi
tinti has joined #linux-sunxi
tinti has quit [Read error: Connection reset by peer]
vicenteH has joined #linux-sunxi
tinti has joined #linux-sunxi
* rellla
is looking forward - willswang forked libhybris ;)
_BJFreeman has joined #linux-sunxi
BJFreeman has quit [Ping timeout: 256 seconds]
_BJFreeman is now known as BJFreeman
_BJFreeman has joined #linux-sunxi
BJFreeman has quit [Ping timeout: 252 seconds]
_BJFreeman is now known as BJFreeman
xlab has joined #linux-sunxi
hansg is now known as hansg_lunch
<hno>
mnemoc, what is the status of kernel developments past 3.4?
xlab has quit [Ping timeout: 272 seconds]
_BJFreeman has joined #linux-sunxi
BJFreeman has quit [Ping timeout: 256 seconds]
_BJFreeman is now known as BJFreeman
_BJFreeman has joined #linux-sunxi
tinti has quit [Ping timeout: 252 seconds]
BJFreeman has quit [Ping timeout: 252 seconds]
_BJFreeman is now known as BJFreeman
shineworld has quit [Remote host closed the connection]
hansg_lunch is now known as hansg
<mnemoc>
hno: mripard is leading the mainlining process.
<mnemoc>
hno: personally I've had very little time. but my idea is to obsolete 3.0 as soon as android people confirms they can use 3.4
<mnemoc>
hno: and then integrate sun[67]i (3.3 based) there
<mnemoc>
i don't really see much point in making a legacy >3.4 branch
<mnemoc>
legacy as of script.bin-based
<mnemoc>
it would be more of a distraction. 3.4 is a good middle ground for unifying/refactoring the whole sunXi world, and then providing decent drivers that can be DTSized and mainlined
<mnemoc>
i'll try to merge the stage branches today and make another stable-version jump before adding the rest of the pending commits
<mnemoc>
hipboi promised a20-based prototype cubieboards (ii) for early may. don't know about a31 devices with uart yet
<mnemoc>
but he is planning to launch a "cubiebox" (mele-ish but with all pins available) with a31 soon
<mnemoc>
hopefully there will be a a20-based equivalent of that too
tinti has joined #linux-sunxi
shineworld has joined #linux-sunxi
<shineworld>
Have someone tried/used android open accessory with cubieboard or any other Allwinner powered device ?
<shineworld>
I've tried with cubieboard and openbox but doesen't work.
<shineworld>
So for test I've purchased a tablet based on Allwinner A13 and running 4.0.4 (5.0.8+) but don't work yet...
<shineworld>
AOA = android open accessory is fully supported starting from honeycomb
<shineworld>
In my Nexus 7 works perfectly without problems
<shineworld>
I need to use it with cubieboard and a tablet A1x powered
shineworld has quit [Quit: Leaving]
shineworld has joined #linux-sunxi
<shineworld>
that's strange because in repo there is AOA ... but don't work
<shineworld>
and I've bought a tablet just to try if there are bugs on github.com/cubieboard sources but same result
shineworld has quit [Remote host closed the connection]
shineworld has joined #linux-sunxi
<ssvb>
mnemoc: who are these android people? maybe Turl?
<ssvb>
Turl: the only theoretical problem that I can see is that the UI code in android might be written in such a way that it needs a certain amount of GPU memory being always available and never failing allocations
<ssvb>
android does not use swap, right?
<Turl>
no swap
<Turl>
no mem? let's hunt some processes!
<Turl>
that's basically how android works
<ssvb>
hmm, interesting
<Turl>
look at lowmemorykiller on staging/android
<ssvb>
well, I wonder what would happen if the memory is full and the critical UI process tries to allocate some GPU memory in android
<Turl>
ssvb: GL OOMs aren't usually well received, but they don't occur with the current memory allocation
<ssvb>
will this lowmemorykiller be invoked in this case or the GPU allocation will simply fail?
<Turl>
if you have a patch I can apply and test I can give it a try tonight
<Turl>
one with instrumentation to print allocation sizes would be even better :)
Quarx has joined #linux-sunxi
<Turl>
ssvb: LMK has kind of 'buckets', if free ram goes below X, it kills from the first bucket, if ram goes below Y, from this other bucket, etc
<ssvb>
Turl: the patches for tweaking mali GPU memory allocation are in the linux-sunxi ML, posted several days ago
<Turl>
ssvb: this 3 patch series? [PATCH 1/3] gpu: mali: Use real framebuffer location instead of magic constants
<rellla>
ssvb: you mean, it would be worth to think about, if init_data is passed correctly
ZaEarl has joined #linux-sunxi
<ssvb>
Turl: yes
<ssvb>
rellla: I'm going to check how this weighted_pred_flag flag is handled in nvidia vdpau implementation
<ssvb>
rellla: and compare with cedarx vlc
<Turl>
ssvb: ok, I'll give it a try tonight
<ssvb>
Turl: thanks
BJFreeman has quit [Ping timeout: 252 seconds]
yuq825 has joined #linux-sunxi
<yuq825>
hello?
yuq825 has quit [Client Quit]
<Turl>
people these days, so impatient
<lunra>
People hear 'instant messaging' and assume it means 'instant response'
yuq825_ has joined #linux-sunxi
<yuq825_>
Sergey?
<Turl>
yuq825_: -> slapin_n1
<Turl>
hno: ^
<Turl>
got to go, see you guys later
<hno>
hi yuq825_
<yuq825_>
hi
<hno>
welcome here
<yuq825_>
thank you
<hno>
haven't seen slapin_n1 today. But guess he is not too far away, responding to email less than an hour ago.
<yuq825_>
I haven't use this before
<hno>
here you find everyone that matters :)
<hno>
I haven't dissected your u-boot MTD changes yet.
<hno>
gah.. staging/sunxi-3.4 panics during kernel startup for me.
<mnemoc>
uh
<yuq825_>
Is this because of my MTD driver?
<mnemoc>
hno: and not-stage?
<mnemoc>
hno: if booting from nand you need to repace nanda's linux/u-boot.bin with one that properly passes ATAG stuff (like your lichee-dev branch ;-)
<paulk-desktop>
techn_, is it related to the wifi driver too?
<techn_>
paulk-desktop: No.. its related to hno's gcc problem
<paulk-desktop>
ok :)
<paulk-desktop>
thanks!
<hno>
techn_, that's an optimization.
* slapin_n1
looks for self-motivation guides, do anybody have some?
<hno>
there is many, but can't stand any of them so don't have one.
rellla has joined #linux-sunxi
tzafrir has quit [Read error: Operation timed out]
eebrah has joined #linux-sunxi
tinti has quit [Ping timeout: 268 seconds]
tinti has joined #linux-sunxi
<paulk-desktop>
libv, hi
<paulk-desktop>
congrats for running quake3 with compiled shaders!
<ssvb>
rellla: setting weighted_pred_flag to 0 here introduces blocky artifacts for nvidia vdpau - http://git.videolan.org/?p=ffmpeg.git;a=blob;f=libavcodec/vdpau.c;h=6df7f4af9594a32eb87ffc0d8329fea7d6bba931;hb=refs/heads/release/1.2#l208 :)
<ssvb>
rellla: so this indeed means that this flag most likely needs to be passed to the hardware decoder from the client side
<ssvb>
rellla: now it's time to look how this is handled with cedarx
<rellla>
ssvb: so we need to figure, how we can do that.
<rellla>
dito :)
<rellla>
maybe the links i posted would be helpful. there is some conversion done.
<Turl>
idk what you guys are doing, but maybe the android glue code holds the clue?
<rellla>
perhaps we have the possibility to get a libve user guide with the last page filled out. not only /*TODO*/
<ssvb>
:)
<rellla>
Turl: idk, too :p
<rellla>
i wish we could get empatzero and willswang on board again...
shineworld has joined #linux-sunxi
<Turl>
mripard_: :)
<mripard_>
Turl: ?
<Turl>
mripard_: I just got 7 emails from you
<mripard_>
ah
<mripard_>
sorry for the mail bombing :)
<Turl>
no problem :) keep me in cc, I like to see the patches
<mripard_>
that's what I thought :)
<Turl>
mripard_: the clock bit on the cover letter doesn't hold true anymore btw :)
<mripard_>
oh, I forgot to remove it ?
<mripard_>
damn
<mripard_>
And I added the clock support to it as of tonight...
_BJFreeman has joined #linux-sunxi
BJFreeman has quit [Ping timeout: 246 seconds]
<ssvb>
rellla: we could possibly try to sniff this missing init_data information for h.264/avc1 from the android player
tzafrir has joined #linux-sunxi
<ssvb>
rellla: but I don't have any cables for the Internal USB device port for my Mele A2000 device, which are probably needed for adb and android development :(
<ssvb>
Turl: is android fully open source for sunxi (including the hardware accelerated video player)?
<Turl>
ssvb: you can adb over wifi/ethernet
<Turl>
ssvb: cedar uses a couple of blobs, so no
<Turl>
and there's mali
<Turl>
the rest is open
<Turl>
mnemoc: ping
Turl has quit [Ping timeout: 258 seconds]
tinti has quit [Quit: Leaving]
<shineworld>
I'm not able to use Android Open Accessory with cubieboard openbox. It lunch intent, recognize manufacturer, id, version but then don't work.
<shineworld>
in dmesg I've got that: <4>[ 1860.960000] WRN:L1367(/home/shine/android/build/openbox/kernel/allwinner/common/drivers/usb/sun4i_usb/udc/sw_udc.c):handle_ep0: ep0 setup end
<shineworld>
<
<shineworld>
do you know what's mean and happen ?
<techn_>
I'll push my current usb dev stuff
<techn_>
you could try with it
<shineworld>
a new kernel ? what version ?
<techn_>
It's top of 3.0
<shineworld>
good to know...
<shineworld>
I've waited a good reason to step from 3.0.52+ to linux-sunxi 3.0.6x
<shineworld>
saturday I've purchased also a low-cost tablet A13 powered to check if Android Open Accessory works there (in cubie after some days of tries nothing works)
<shineworld>
but also in tablet don't work (I guess same sources and kernel 3.0.8+)
<shineworld>
I've connected my Nexus 7, Galaxy S Advance (JB powered) and SGS S2 and all works fine with accessory device
<shineworld>
so could be a problems with sun4i / sun5i device drivers (I guess)
<techn_>
that usb brach has sun4i and sun5i unified
<techn_>
(that branch contains some other shit ) ;)
<shineworld>
could be I copy/paste only usb section ?
<shineworld>
or just meld mine with that to find differences ?
<techn_>
Or cherry-pick what you need
<techn_>
6 commits
<shineworld>
yeah... thanks very much for sharing
<shineworld>
I will try now !
<shineworld>
what seems very incredible is that none ALLWinner product user (all use same ICS image) catch that problem with AOA ... Or I'm only that use a FT311D chip in the world
<shineworld>
there are at least hundred tablet models which use same chip/android distro
<shineworld>
I've planed to purchase another tablet of different brand but always allwinner a10/13 powered to enlarge the test
<shineworld>
hey, a lot of changes
<techn_>
shineworld: yes :)
<shineworld>
I guess main problem for me is:
<shineworld>
/* Enable udc */
<shineworld>
//sw_udc_enable(udc); //remove by 2011.11.25
<shineworld>
vs your
<shineworld>
/* Enable udc */
<shineworld>
sw_udc_set_pullup(udc, 1);
<techn_>
shineworld: It wont work yet :(
rellla has quit [Remote host closed the connection]
<techn_>
That is the main reason why I havent submitted patch set
<shineworld>
so I can't have openaccessory working with AllWinnder ? (just to ask so I move to other brands like rockchip, ...)
<techn_>
but I think It's much better there.. but not good enough
<techn_>
as said.. I hope I have time to fix it
<shineworld>
like I dumb, I've already ordered 100 tablets to supply a customer which would use an android app to manage GUI interface of a machinery...
<shineworld>
I've thinked that all ICS tablet can support AOA
<techn_>
shineworld: did you try to comment that row out.. it atleast should make android adb working
<shineworld>
ok, tomorrow I will try to suspend the order till I will be able to understand the question
<shineworld>
thanks for support
<techn_>
hmm.. you have same usb problem with 0.52 kernel too?
<techn_>
3.0.52?
<shineworld>
Tablet run 3.0.8+ and AOA don't works. Cubieboard run github.com/cubieboard (3.0.52+) and same problems
<shineworld>
I haven't tried yet github.com/linux-sunxi 3.0.62