<draconian>
guys can i use olinuxino sd car images to boot a generic A13 tablet ?
Netlynx has joined #linux-sunxi
Netlynx has joined #linux-sunxi
<draconian>
any clue ?
<draconian>
tried rebuilding the kernel and didnot help
Andy-D has joined #linux-sunxi
<libv>
draconian: manual build howto
<draconian>
i did check that n it didnt boot at all
<draconian>
it says nothing about generic A13 Tablets only refers to specific development boards of the same SoC
<draconian>
the only boot i got was from mele debian with bash command not found for all commands !
<adj_>
libv, do you have a script to recommend reading "new device howto" and "manual build howto"?
<adj_>
could save you a lot of key strokes
<adj_>
;)
<libv>
adj_: each time should come with a situation specific lart that smarts just enough that people actually listen and then read up carefully
<libv>
because otherwise people just ignore such things
bonbons has joined #linux-sunxi
<libv>
oh, but good point
<libv>
draconian: there is no thing called generic
<mrnuke>
hi libv
<libv>
draconian: you need to work through the new device howto
<libv>
draconian: and the word is "work" not "check"
<draconian>
re-checking it now
<libv>
...
<linkmauve1>
libv, since you already did quite a lot of Mali RE, do you know why there seem to be an implicit black glClear() at every frame on my Mali-400 with the blob?
<linkmauve1>
And also, why rendering to a texture through a framebuffer then displaying it to the screen, then redrawing to it without clearing the previous frame, halves or worse the framerate I got by not using a framebuffer at all?
<linkmauve1>
My guess would be I’m breaking asyncness, but I have no way to check that in the driver, right?
<libv>
why do you need this?
<linkmauve1>
I’m just running my game engine on an A10 Lime.
<linkmauve1>
Aside from a small mistake, it was entirely GLESv2 compatible. :)
<linkmauve1>
The longer goal is to provide a free GLES engine to try free drivers on.
<libv>
well, first of all, you incur twice the work
<libv>
one to render out the framebuffer, one to render out that framebuffer to the screen
<linkmauve1>
Yeah.
<libv>
even with the simple geometry, the fragment shaders will have to re-render every pixel, and read in the texture you put in the framebuffer
<linkmauve1>
But the first one is at a fixed 640×480 resolution, rendering many elements, while the second is at the size of the window.
<linkmauve1>
On desktop I use ARB_framebuffer_blit to lower the load, but it doesn’t seem available in libMali.
<libv>
linkmauve1: why not use display layers?
<linkmauve1>
First time I hear that name. :)
<linkmauve1>
Are they like DRM planes?
<libv>
linkmauve1: on the other hand, not much is wasted in mali when doing a trivial full frame render
<libv>
linkmauve1: drm planes are the shortsighted intellisation by someone who hadn't done display drivers before
<libv>
like most things drm
<libv>
but yes
<linkmauve1>
Ah btw, when I’m doing only the first rendering pass directly to the screen, the game is taking about 70~80% of the CPU, and when I add the second one it goes as low as 30%.
<linkmauve1>
So display layers are a proprietary (?) API to display a scaled texture?
<libv>
display layers, or overlays, show different areas of the framebuffer in different places, and often do scaling and/or colour conversion
<libv>
you could get away with 16bit colour for instance
<libv>
or you could combine different parts of the userinterface more nicely
<linkmauve1>
Get a request, open the layer, SET_SRC_WINDOW and SET_SCN_WINDOW, and then render every frame normally?
<linkmauve1>
Does it integrate nicely with X11, like with window size change or such?
<libv>
you will have to take care of that yourself
<linkmauve1>
Ok.
domidumont has joined #linux-sunxi
<linkmauve1>
That means it’d work fine even on non-X11. \o/
nicksydney has quit [Remote host closed the connection]
formruga has joined #linux-sunxi
nicksydney has joined #linux-sunxi
VargaD has quit [Ping timeout: 250 seconds]
VargaD has joined #linux-sunxi
<ssvb>
linkmauve1: assuming that you don't mind adding a special code path, which works only on Allwinner hardware and only with the sunxi-3.4 kernels
<ssvb>
linkmauve1: but some developers are doing the same with the raspberrypi dispmanx :)
<linkmauve1>
Well, ideally that’d be handled by the underlying toolkit, SDL2 in my case. :p
quitte has joined #linux-sunxi
<linkmauve1>
And they seem quite open to platform specifics paths.
<ssvb>
jemk: tried to experiment to find a replacement for the use of Mali hardware in the lima-memtester program
<ssvb>
jemk: appears that the 90 degrees rotated blit in g2d is also efficient for exposing memory reliability problems
<ssvb>
jemk: now running some statistical tests, to compare the efficiency of "lima spinning cube" vs. "g2d 90 degrees rotated blit" for this task :)
<ssvb>
jemk: anyway, it looks like we should be able to test memory reliability on A31 too
konradoo77 has quit [Ping timeout: 245 seconds]
paulk-collins has joined #linux-sunxi
<ssvb>
jemk: so even though Allwinner A13 does not have G2D and Allwinner A31 does not have Mali, all devices can be still covered
<ssvb>
ijc: a good catch with that u-boot psci disabling option
whitesn has quit [Quit: WeeChat 1.0.1]
whitesn has joined #linux-sunxi
whitesn has quit [Changing host]
whitesn has joined #linux-sunxi
domidumont has quit [Ping timeout: 246 seconds]
viccuad has joined #linux-sunxi
konradoo77 has joined #linux-sunxi
barbarin has joined #linux-sunxi
barbarin has quit [Client Quit]
barbarin has joined #linux-sunxi
domidumont has joined #linux-sunxi
<barbarin>
Hi, I follow the New device howto and create a wiki page for my Allwinner device.
<barbarin>
But some guy name Rellla deleted it.
<barbarin>
My device is a netbook, and I tried H6 the only netbook boot image allready on sunxi. The H6 won't work as this netbook looks different from mind
ricardocrudo has joined #linux-sunxi
ricardocrudo has quit [Remote host closed the connection]
ricardocrudo has joined #linux-sunxi
draconian has quit [Remote host closed the connection]
<barbarin>
As any image work and some one delete the wiki page I took time to create, Can some one tell me, what assistance the sunxi community could provide ? Tanks
Renard has joined #linux-sunxi
Black_Horseman has quit [Ping timeout: 264 seconds]
konradoo77 has quit [Ping timeout: 258 seconds]
<ssvb>
barbarin: there is a standard template for the device information pages, just take it into use and populate with information
<ssvb>
barbarin: this makes it
<ssvb>
1) easier for you, because you don't need to invent your own way of representing the information
<ssvb>
2) provides uniform look across all pages and makes it easier for everyone to compare features of different devices
<rellla>
This was the reason why this guy rellla deleted it :p
<barbarin>
Ok I found the "Device Page Example", It is more clear now. I will do it the right way, Thank you guys
konradoo77 has joined #linux-sunxi
<ssvb>
barbarin: thanks!
<ssvb>
btw, where did you find that non-working image?
<ssvb>
it does not provide a complete linux system yet, but eventually we may have something that is easy to install for people like you :)
<barbarin>
Ok, I try these links
popolon has joined #linux-sunxi
konradoo77 has quit [Ping timeout: 258 seconds]
konradoo77 has joined #linux-sunxi
deasy has joined #linux-sunxi
deasy has quit [Read error: No route to host]
deasy has joined #linux-sunxi
<barbarin>
I'm lost with the numerous files needed to boot a device. I try uboot tools with sun4i but it didn't create spl/sunxi-spl.bin. Is it essential ?
konradoo77 has quit [Ping timeout: 272 seconds]
afaerber_ has quit [Ping timeout: 240 seconds]
Andy-D has quit [Ping timeout: 256 seconds]
afaerber_ has joined #linux-sunxi
konradoo77 has joined #linux-sunxi
cnxsoft has quit [Quit: cnxsoft]
<linkmauve1>
Urgh, “chmod 777 /dev/disp /dev/cedar_dev /dev/ump /dev/mali” in the /etc/rc.local of that minimal Wheezy image…
<linkmauve1>
And the author left his public ssh key in root’s authorized_keys.
Akagi201 has joined #linux-sunxi
Andy-D has joined #linux-sunxi
<ssvb>
linkmauve1: the joy of relying on random binary images from the Internet
<linkmauve1>
Indeed. :D
<ssvb>
does mali at least work out of the box?
<linkmauve1>
It does.
<linkmauve1>
Well, except for some reason he commented mali and mali_drm in /etc/modules
<linkmauve1>
It works after uncommenting them.
<ssvb>
:)
<ssvb>
barbarin: it's hard to say without knowing which commands you run and what output you get
<Wizzup>
So the USB OTG in A10 and A20 is one that already has a driver in mainline, is what I read from the mainline page
<Wizzup>
But afaik it doesn't work yet on mainline, right?
konradoo77 has quit [Ping timeout: 240 seconds]
leviathanch2 has joined #linux-sunxi
konradoo77 has joined #linux-sunxi
leviathanch2 has quit [Ping timeout: 258 seconds]
<ssvb>
Wizzup: it is listed in the "Work In Progress" section
<Wizzup>
yes, it is :)
<Wizzup>
There's a patch, but the text below it on the wiki is confusing me
<ssvb>
Wizzup: the hardware itself has a driver, but still needs some extra glue code to make use of it
<Wizzup>
alright - I just wanted to use it as extra 'normal' usb port
<ssvb>
Wizzup: ping arokux, maybe he can share some of his work-in-progress code for testing
<adj_>
does anyone know what boxchip means? I only found on linux-sunxi web that the F-series are named Boxchip xxx
<adj_>
some tables/netbooks says it has a Boxchip A10 Cortex-a8 multicore (or Boxchip/Allwinner A10 Cortex-a8 multicore)
<adj_>
is it allwinner A10? whay they say cortex-a8 multicore?