<oliv3r>
hno: i had the issue on lichee-dev-a10, so the same fix needs to go into that tree aswell
<oliv3r>
hno: they should/are seperate patches; i did split them as best as I could; but it's a WIP branch :p
<oliv3r>
i think i've sent the most reasonable work to the ML a while ago, but i could have put it on hold; let me know when your back from vacation and I'll clean up the patchset
<oliv3r>
Turl: mainline suspend? sweet; remember, wingrime did some suspend work on 3.4, so don't forget to pick his brain ;)
<oliv3r>
01:34 < hno> oliv3r, would be good if dramc magics patch could be split to
<oliv3r>
sunxi-current. Almost impossible to review a20 changes now.
<oliv3r>
01:45 -!- utente [~utente@86.33.168.109.dsl.static.ip.kpnqwest.it] has quit
<oliv3r>
[Read error: Operation timed out]
<oliv3r>
Turl: and what are you doing up so late!
<oliv3r>
and i did some copy paste fail there :(
<Turl>
oliv3r: it's called "holidays" :P
<oliv3r>
lol
<oliv3r>
what time is it where you are?
<Turl>
oliv3r: do you have any kind of amperimeter or such to measure power consumption?
<Turl>
oliv3r: almost 4AM
<oliv3r>
ohh, that's not so bad
<oliv3r>
i have a multimeter
<oliv3r>
i'd have to rig something up to measure power draw (which should be not too hard
<oliv3r>
splice some old usb cable
<oliv3r>
insert multimeter
<oliv3r>
done
<Turl>
oliv3r: if you could test a mainline branch and see if power draw goes down when you suspend it'd be great
<Turl>
oliv3r: not now though, the code is pretty green and hacky so far
<oliv3r>
i'm @work
<oliv3r>
so can't do anything like that now anyhow :)
<Turl>
as a sidenote, debugging assembler by poking stuff to the uart via mmio is a great experience :p
<oliv3r>
lol
<Turl>
oliv3r: yeah, think some days/weeks in the future :) prepare it in advance if you wish :p
Soru has quit [Ping timeout: 240 seconds]
<oliv3r>
lol
<oliv3r>
i'll be going on holidays aug 3th for a week
<oliv3r>
and after that i'm a week @home
<oliv3r>
hno: well github is identicla to my local tree; so that'sg ood; but there's 2 seperate patches, 9929 and 9f2ad
<Turl>
with some luck I might get something testable before then
<oliv3r>
one is sun7i specifics, the other is magic values cleanup
<oliv3r>
Turl: that'd be good
<oliv3r>
with a bit of luck, we'll have some SID in the kernel :p
<Turl>
I need to get up to speed on ARM assembler "best practices"
<oliv3r>
your writing suspend assembly?
<Turl>
like, can I (ab)use macros "static inline" style?
<Turl>
oliv3r: yes, the dram code is asm :p
<oliv3r>
i know nothing of suspend, except for what wingrime did/does
<oliv3r>
style wise, hno might know
<Turl>
oliv3r: any idea what does SDR_CCR, DCR, DRR, HPCR and company mean?
<oliv3r>
of course
<oliv3r>
:p
<Turl>
oliv3r: y u no wiki it then :P
<oliv3r>
there's some preliminary info on the rhombus wiki
<oliv3r>
what our current todo is: Test stage/sunxi-3.4
<oliv3r>
fix regressions that can be fixed (without hansg's patchset)
<oliv3r>
once stage/sunxi-3.4 is deemed stable; pull hansg's for-amery into stage/sunxi-3.4
<oliv3r>
though I am not aware of many regressions with the current 3.4
<oliv3r>
ssvb knows of 2 issues; dunno if there's more
<techn_>
oliv3r: android is broken :p
<ssvb>
techn_: is it a regression in stage/sunxi-3.4?
yuuki has joined #linux-sunxi
<yuuki>
a31 rndis function of android_usb gadget driver on 3.3.0 lichee kernel has bug - packets from host to device do not deliver to device. Do someone have patch/fix?
tinti has joined #linux-sunxi
<wingrime>
ssvb: after some hand-made patches 3.4 works normaly on a20, disp,usb,wemac works
n01 has joined #linux-sunxi
<ssvb>
wingrime: good, what about cedarx? ;)
<ssvb>
looks like the SD slot is going to die in my Mele A2000, I'm getting u-boot stuck failing to read mmc increasingly more often :(
<ssvb>
maybe it's a good idea to migrate u-boot to NAND before it is too late
<oliv3r>
yuuki: we know nothing of 3.3; our work is in 3.0 and 3.4, but i don't think a31 code has been touched at all (other then some preliminary mainline work)
<oliv3r>
ssvb: different SD card?
<oliv3r>
ssvb: rz2k was working on the mtd stuff
<ssvb>
oliv3r: yeah, a good idea, maybe it's actually the card dying
<PiyushVerma>
I am having isuse in running es2_info
<PiyushVerma>
x11 sunxi_fb driver workin ok
<PiyushVerma>
but when I run ./es2_info I get Error: eglGetDisplay() failed mali
yuuki has quit [Read error: Connection timed out]
<PiyushVerma>
where If I use egl directly on FB then it works
<PiyushVerma>
ssvb: and last time qt build success but still have runtime issue
hglm has joined #linux-sunxi
<ssvb>
PiyushVerma: check permissions for /dev/mali and /dev/ump
<WarheadsSE>
and maybe membership to video group
<PiyushVerma>
that's ok I am as it's able to run on FB
<hglm>
PiyushVerma: AFAIK the userspace mali libs can onlt
<hglm>
only be intalled for either fb on X, not both at the same time.
<PiyushVerma>
the default test program of mali-libs works
<PiyushVerma>
hglm: thanks to point it out . any idea how to configure for x ?
<hglm>
PiyushVerma: It's in the instructions for the userspace mali libs, it autodetects but you can add a commandline options to make.
<PiyushVerma>
hglm: ok checking option in make file
<hglm>
PiyushVerma: Should be documented in the README.
<PiyushVerma>
done that now it say couldn't open display
<PiyushVerma>
when I install fb lib then sunxi-mali test program work
<PiyushVerma>
but when install for x11 then that not work
<PiyushVerma>
going to read Common pitfalls
<libv>
PiyushVerma: 15:56 < libv> PiyushVerma: how about first using the test program included in sunxi-mali?
<PiyushVerma>
libv: as mention by hglm there there is 2 different binary one for fb and other for x11. when I build for fb then it works when build for x11 then error failed to open x display
<ssvb>
PiyushVerma: how do you run it? maybe you need to "export DISPLAY=:0" if you run it from the ssh shell?
<PiyushVerma>
ssvb: right that was one issue now I run within xwindow and error change to eglInitialise faild
<PiyushVerma>
checking code which section
<PiyushVerma>
eglInitialize call fail
<ssvb>
PiyushVerma: check /var/log/Xorg.0.log and also probably strace the test program
<ssvb>
PiyushVerma: make sure that you have the right permissions/ownership for /dev/mali and /dev/ump
<PiyushVerma>
I am running as root and have enough permission going to debug as u suggested
n01 has quit [Ping timeout: 240 seconds]
<ssvb>
well, there must be something wrong in your system or in what you are doing
<ssvb>
maybe also check that you really have the x11 flavour of libMali.so installed
<PiyushVerma>
there are 100K + lines of entry in log of SUNXIFB(0): FBIOPUTCMAP: Invalid argument
<PiyushVerma>
and with strace it's pollling many time and resource temporarily unavailable
tzafrir has quit [Ping timeout: 240 seconds]
<ssvb>
heh, "FBIOPUTCMAP: Invalid argument" is kinda normal, even though we should really suppress it
<PiyushVerma>
ssvb: how to verify mali lib
<PiyushVerma>
I think lib is change cause bfore fb was working and no fb not working
<ssvb>
strace may show if accessing or opening something may be failing
<PiyushVerma>
ok lib verified
hglm has quit [Quit: leaving]
<PiyushVerma>
md5 sum of installed lib and sunxi-mali/lib/mali/r3p0/armhf/x11 is same
<ssvb>
good, can you also run ldd for the test program?
<ssvb>
just to be sure that the right libraries are resolved as its dependencies
<PiyushVerma>
what could be reson of Resource temporarily unavailable ?
<ssvb>
how the hell are we going to know? you are the one who has the complete logs :)
<PiyushVerma>
wow it's using libEGL and libUMP but no libMali if test with ldd
<hglm>
You probably have to reconfigure/recompile/reinstall sunxifb with the userspace X11 mali libs installed.
<wingrime>
oliv3r: sound works
<oliv3r>
wingrime: yeah hans did say it should
<oliv3r>
good to know :)
<PiyushVerma>
hglm: already did that going to do again
<hglm>
I just installed sunxi-mali on my recent debian install, only problem I encountered was a dependency on XNextEvent in test/test from the sunxi-mali repository. I added -lX11 to fix it.
<hglm>
PiyushVerma: Make sure you rerun autoreconf -vi and configure
<PiyushVerma>
hglm: ok so it's working ur side u able to render ?
<hglm>
Piyush: Yep, test/test works.
<PiyushVerma>
hglm: are u using ssvb repo of linux-sunxi ?
<hglm>
I had to install the seperate DRI2 lib though as described on the wiki.
<hglm>
Piyush: You should use the linux-sunxi repo for Mali, ssvb for sunxifb.
<PiyushVerma>
so 2 thing missed one repo and other DRI2
<PiyushVerma>
as automake was saying libdri missing so I install libdri-dev
<wingrime>
oliv3r: you have pending pull to hasang ?
pitillo has joined #linux-sunxi
<oliv3r>
on the mailing list 1 or 2
<libv>
PiyushVerma: why is DISPLAY not automatically initialized?
<oliv3r>
hopefully we can get that done quickly, so we can work from stage/sunxi-3.4 again
<libv>
PiyushVerma: any terminal program under X sets that automatically.
<wingrime>
oliv3r: I did pull req over github system
<PiyushVerma>
libv: actually early I was using ssh that's why latter I switch to x
eebrah is now known as eebrah|away
<hglm>
Regarding stage/sunxi-3.4 stability, I couldn't get wlan to work on that branch, while it works with hansg's latest for-amery branch.
<libv>
PiyushVerma: you did use the udev rules?
<PiyushVerma>
libv: yes
<libv>
PiyushVerma: and the user is in the right group?
<ssvb>
hglm: not sure if anyone really cares about this stuff, but it would be interesting to bisect to identify when it your wlan broke and when it got fixed again
<libv>
anyway, off for shopping
<hglm>
ssvb: Perhaps, it suspect may have to do with USB not being properly configured when I compiled stage/sunxi-3.4 with the default config.
<PiyushVerma>
ls
<ssvb>
PiyushVerma: can you make sure that you don't have "no 3D acceleration because the driver has been compiled without libUMP" message in your /var/log/Xorg.0.log ?
<PiyushVerma>
ssvb: yes message is there
<PiyushVerma>
no 3D acceleration
paulk-desktop has joined #linux-sunxi
<ssvb>
PiyushVerma: then it is <hglm> You probably have to reconfigure/recompile/reinstall sunxifb with the userspace X11 mali libs installed.
<PiyushVerma>
ssvb: right looks we got the issue going to do that thankyou very much
<PiyushVerma>
ssvb: some more changes needed many distry don't have libdri2-dev
<PiyushVerma>
in my case libdri-dev works without issue
<PiyushVerma>
so can escape libdri2 build issue
<WarheadsSE>
mmm
atiti has quit [Ping timeout: 245 seconds]
n01 has joined #linux-sunxi
<PiyushVerma>
ssvb: one more thing in a qt example top side it's getting line line kind of noice not sure what is that
<PiyushVerma>
but content also rendering properly but just some extra noice
<ssvb>
PiyushVerma: heh, there is some issue related to how the mali dri2 blob requests the front and back buffers
<ssvb>
PiyushVerma: qt5 tries to resize the window just between these requests
<ssvb>
PiyushVerma: this should not be a problem with r3p2
<PiyushVerma>
ssvb: sorry just back
<PiyushVerma>
but I am not resizing window
<ssvb>
PiyushVerma: qt does on your behalf
<PiyushVerma>
ok
<ssvb>
PiyushVerma: are you running the demo in fullscreen mode?
shineworld has quit [Quit: Leaving]
<PiyushVerma>
so changing mali drive in kernle is just replacing a folder ? should I try that
_BJFreeman has joined #linux-sunxi
<PiyushVerma>
ssvb: not in full screen
_BJFreeman is now known as BJfreeman
<ssvb>
PiyushVerma: still the window manager may reposition/resize right after start
<ssvb>
PiyushVerma: upgrading mali is not so easy, you also need libMali.so blob with a matching version
<PiyushVerma>
ssvb: every mili second ?
<ssvb>
PiyushVerma: if you have mismatched back and front dri2 buffers, the picture changes every second frame
<ssvb>
PiyushVerma: for example normal pictures on even frames and garbage on odd frames
<PiyushVerma>
ok
<ssvb>
PiyushVerma: do you have any contacts with Allwinner or whoever might potentially provide a new version of libMali.so blob?
<PiyushVerma>
not yet but will check with indirect source
<ssvb>
if the framebuffer mali drivers work good for your applications, you should probably go for it
<ssvb>
XBMC is using the framebuffer mali drivers
<PiyushVerma>
actually qt5 default code is for x11 so to use fremebuffer with qt5 need a bit patch for qt5
n01 has quit [Ping timeout: 240 seconds]
<Offshore>
greetings, guys
<Offshore>
where to post the feature request for "halt" command in u-boot? :)
n01 has joined #linux-sunxi
wingrime has quit [Ping timeout: 260 seconds]
yuuki has joined #linux-sunxi
<libv>
WarheadsSE: the wiki page you mean, or potential installation issues?
<libv>
or packaging?
* libv
still doesn't get why noone has shown up to package sunxi-mali
<WarheadsSE>
libv: getting back to cleanly packaging it
<WarheadsSE>
no one has tried aside from me, and I have really had my hands full, plus my office is packed at the moment thanks to a tree falling into it
<libv>
cool
<libv>
:(
<WarheadsSE>
yeah, fun
<libv>
trees are terrible. they always are in the way on beautiful forest paths. The block the view from nice rocks on top of hills in the middle of the forest. _And_ they fall on things
<libv>
the eu should cut them all down!
<WarheadsSE>
I rent, and landlord is 1) dragging his feet 2) possibly being negligent 3) even more slightly intentionally witholding info from the insurance agency
<WarheadsSE>
anything inside is covered cost x2 by my own insurance.
<libv>
but that doesn't fix the building...
<libv>
and endless paperwork...
<WarheadsSE>
right
<WarheadsSE>
I am out my office space, work bench, massive compile box, etc etc
<WarheadsSE>
and this isn't the only thing wrong recently with the property
<WarheadsSE>
not touching that.
<WarheadsSE>
needless to say, I have refused to pay rent until he clearly communicates with me.
<WarheadsSE>
(and I am within my rights to do so without his ability to evict)
<libv>
i had good hopes for mdfe and his obs/mer packaging work
<WarheadsSE>
Well, I can make it work in Arch without too much of an issue
<WarheadsSE>
I just haven't had the time to come back to it.
<libv>
but mdfe is too knackered from working 2 tasks at open exchange, and hasn't been able to find time for his tablet since march
<WarheadsSE>
but now I have a bunch of the AW Axx series that have the mali400
alcides has quit [Remote host closed the connection]
soul has joined #linux-sunxi
rings has joined #linux-sunxi
wingrime has joined #linux-sunxi
rings has left #linux-sunxi [#linux-sunxi]
_BJFreeman has joined #linux-sunxi
BJfreeman is now known as Guest8492
_BJFreeman is now known as BJfreeman
Guest8492 has quit [Read error: Connection reset by peer]
arete74 has quit [Ping timeout: 246 seconds]
arete74 has joined #linux-sunxi
BJfreeman has quit [Read error: Connection reset by peer]
ibrah has joined #linux-sunxi
_BJFreeman has joined #linux-sunxi
_BJFreeman is now known as BJfreeman
rings_IIV has joined #linux-sunxi
_BJFreeman has joined #linux-sunxi
BJfreeman has quit [Read error: Connection reset by peer]
_BJFreeman is now known as BJfreeman
<Turl>
oliv3r> but don't tell him i forgot ;) <- I can read :p
rz2k has quit []
tzafrir has joined #linux-sunxi
BJfreeman has quit [Read error: Connection reset by peer]
_BJFreeman has joined #linux-sunxi
_BJFreeman is now known as BJfreeman
ibrah has quit [Ping timeout: 240 seconds]
<oliv3r>
Turl: i know d'oh i even made sure your name was highlighted a few lines above :)
<oliv3r>
you where supposed to say 'oh but i never asked that'
<oliv3r>
which means i dind't forget anything!
<oliv3r>
wingrime: please send small patches over the mailinglist (additional if anything) github is nice to do a pull; but reviewing or being aware that there is a patch that needs reviewing is easier over the ML
<oliv3r>
but priority task 1 still stands; we need to see if stage 3.4 is good to go; so we can merge hansg's patches in
<oliv3r>
maybe even move current stage to non-stage
<Turl>
oliv3r: :P
ibrah has joined #linux-sunxi
<wingrime>
oliv3r: I know, but untill mnemoc not returned yet
<oliv3r>
yeah :(
<oliv3r>
hopefully monday he said
<wingrime>
oliv3r: i can't send it becose it on hsang top
<oliv3r>
well we can still review it
rings_IIV has quit []
<Turl>
sleep $((15*60)) && ping mripard
<oliv3r>
so mripard should respond to you in 15 h ours?
<ssvb>
oliv3r: "VDPAU is a slice-level API. Put another way, VDPAU implementations accept "slice" data from the bitstream, and perform all required processing of those slices (e.g VLD decoding, IDCT, motion compensation, in-loop deblocking, etc.)."
<oliv3r>
hmm, that's not what the wiki says
<oliv3r>
well it does say the slice bit
<oliv3r>
a slice contains macro blocks
<oliv3r>
and i think, from how i understand it, you feed a slice to the engine (e.g. a group of macro blocks) and then offload 1 operation
<oliv3r>
get it back, do the next one
<oliv3r>
but that's from the wiki
<oliv3r>
vaapi does the same incidentally
<ssvb>
I prefer reading the sources :) and vdpau actually provides a nice tracer of api calls, which greatly helps in understanding how it works
Soru has joined #linux-sunxi
<ssvb>
I played with nvidia vdpau traces and dumped parts of the bitstream when debugging the cedarx h.264 artifacts problem
<oliv3r>
ah; very cool
Offshore has quit [Ping timeout: 240 seconds]
<oliv3r>
and vaapi doesn't hae all that at all?
Offshore has joined #linux-sunxi
<ssvb>
vdpau also provides simple bitmap surfaces and 2d functionality (for subtitles and osd), where g2d fits quite nicely
ibrah has quit [Ping timeout: 276 seconds]
<oliv3r>
ah right, and vaapi needs this all composited via the 3d engine?
<ssvb>
that's what I can see in the current implementations of the video players
<oliv3r>
all our chips have g2d? even a13?
<ssvb>
maybe not a13
paulk-desktop has quit [Quit: Ex-Chat]
<ssvb>
I don't have any a13 hardware to verify for sure, but all the information indicates that a13 does not have g2d
<oliv3r>
so how would a13 hardware do that?
<oliv3r>
osd and subs
<ssvb>
I guess using the cpu, or maybe mali
<ssvb>
anyway, I think it's better to just implement it and then see how it works in practice instead of speculating :)
Offshore has quit [Ping timeout: 268 seconds]
<oliv3r>
well in that regard, libva sounds slightly more interesting