<Turl>
DJJeff: boot an SD image and try fixing it with nand-part and dd then
<DJJeff>
how do I do this?
<DJJeff>
the nand on my tablet is all 0's
<Turl>
DJJeff: given that you don't have a working system it'll be hard to get the original memory config, but if you play it safe, you can probably find a board on uboot which will boot
<Turl>
you have the script.bin on that nanda I suppose
<Turl>
you can then make a sd card image
<DJJeff>
using phoenixcard?
<Turl>
no, using dd, mkfs.vfat, make, the wiki, etc
<DJJeff>
X)X
<DJJeff>
X_X
<Turl>
or something like the fedora image, which should be less painful
<jemk>
rellla: but I can't really believe this, according to ssvbs g2d_bench it does ~80MPix/s blit, that would mean around 25ms per 1080p image
FR^2 has joined #linux-sunxi
leviathanch2 has joined #linux-sunxi
<rellla>
jemk: in old 3.3 sources, timeout was set to 100ms ...
<rellla>
what do you think of that: 23:49 <ssvb> rellla: could it be some sort of a race, caused by several different g2d users trying to do something at once?
<rellla>
if i see it the right way, there is everything handled after each other and waiting fo reach other in libvdpau and softhddevice. except g2d_wait_cmd_finish
<jemk>
rellla: it looks like the g2d driver has locks in the ioctls, so only one can be active at the same time
<jemk>
vdpau isn't thread safe yet (it has to be according to specification), that's why i recommended to build softhddevice without USE_VIDEO_THREAD
<jemk>
but it looks like it isn't relevant for the g2d error
Gerwin_J has quit [Quit: Gerwin_J]
<rellla>
jemk: so may i increase the timeout?
<rellla>
is there any possibility to clear/reset the whole g2d thing without a reboot?
newleaves_ has quit [Read error: Connection reset by peer]
<jemk>
rellla: increase timeout could help, but if we really need more than the 50ms it will be too slow for smooth video
Net147 has joined #linux-sunxi
<jemk>
and to reset it might help to unload the g2d module if you build it as module
<jemk>
but i don't know
<rellla>
jemk: i adressed a 100ms test to someone ;) lets see.
sehraf has joined #linux-sunxi
<rellla>
and how is the g2d timeout chained with smoothness of the video? even 50ms should be too much?
notmart has joined #linux-sunxi
kz1 has joined #linux-sunxi
rz2k has joined #linux-sunxi
t3st3r has quit [Write error: Connection reset by peer]
Black_Horseman has quit [Remote host closed the connection]
<rellla>
jemk: the strange thing is, that g2d totally crashes, so even mplayer has this overlay after the issue appeared within vdr. there is nothing that clears or resets it.
kivutar has joined #linux-sunxi
nabblet has joined #linux-sunxi
<jemk>
rellla: I reproduced the problem with a small test, once g2d had the irq pending timeout it doesn't work anymore. Nothing else is involved, no disp, no threads, no ve, so it's a g2d kernel bug
<rellla>
voila
leviathanch2 has quit [Ping timeout: 264 seconds]
popolon has joined #linux-sunxi
popolon has joined #linux-sunxi
<jemk>
now the hard part, find out how to fix it ;)
<rellla>
so this should go to the mailing list...
<jemk>
but the timeout has to be increased anyway, or we can't use g2d for big images.
<rellla>
yes. but a second time: what has timeout todo with the video's smoothness/ frame rate?
nabblet has quit [Ping timeout: 240 seconds]
<jemk>
g2d draws the overlay, if it needs that long the frame will be delayed till it finishes
<jemk>
and at the moment it also fills the output_surface with transparent black each frame to overwrite possible old osd before drawing the new one
nabblet has joined #linux-sunxi
<rellla>
ah ok. video surface + output surface goes into presentation queue and then are displayed if the frame composed of both is ready. so long timeout -> long frame creation -> low frame rate
<jemk>
exactly, it could be optimized to only draw dirty areas, that speeds up for small osd, but for big overlays like in vdr it won't change too much
<jemk>
normally this function should compose the video into the output surface, using a different layer for video is not the way that is intended by vdpau
<jemk>
there is no such thing as closing osd in vdpau. you draw a video to the output surface and then draw something above it. We don't draw the video to the output_surface but show it on a different layer, so the output surface contains the old osd
<jemk>
you can see what happens by removing the black fill, you will always have the old osd below the new one
<rellla>
but now you call G2D_CMD_FILLRECT AND G2D_CMD_BITBLT each frame, right?
<jemk>
yes, it could be optimized with some intelligent logic, but i wasn't motivated enough to figure that out
<mnemoc>
so it's a "normal linux" built using AOSP build system
<nove>
the b2g source is not depend of android, and works in a gnu/linux as is shown by the b2g simulator that isn't any more than b2g compiled to use X11
<nove>
so that gsoc project is possible and not really hard, and if complete
<Gerwin_J>
nove: First the said to me March, but chinese partners get a80 stuff first
kivutar has quit [Quit: Ex-Chat]
<nove>
Gerwin_J: as expected, but why the optimus board, cubietech could making one even better
<mnemoc>
Benn said that the final version will be better (hopefully including more pins and sata) than the one shown at CES
* mnemoc
wonders if allwinnertech 2.0 will send us free boards
<mnemoc>
(and docu :p)
tgaz has joined #linux-sunxi
<Gerwin_J>
Allwinner want to see first Tablets with a80...
<Gerwin_J>
But i want to use this soc for a complete other product
<Gerwin_J>
I have also talk with WITS to make a prototype board with a80
<Gerwin_J>
The get more information this month
t3st3r has joined #linux-sunxi
leviathanch2 has quit [Ping timeout: 264 seconds]
pacopad_ has joined #linux-sunxi
pacopad has quit [Read error: Operation timed out]
pacopad_ is now known as pacopad
sehraf has quit [Ping timeout: 245 seconds]
HeHoPMaJIeH has quit [Remote host closed the connection]
leviathanch2 has joined #linux-sunxi
sehraf has joined #linux-sunxi
pacopad has quit [Ping timeout: 240 seconds]
xeros has quit [Remote host closed the connection]
pacopad has joined #linux-sunxi
rz2k has joined #linux-sunxi
xeros has joined #linux-sunxi
toplessninja has joined #linux-sunxi
<nove>
poor little binary, now has to eat fakes register values
pacopad has quit [Ping timeout: 265 seconds]
pacopad has joined #linux-sunxi
popolon has quit [Quit: Quitte]
lauri has left #linux-sunxi [#linux-sunxi]
<nove>
jemk, this confirms at getbits VE_H264_STATUS is bit 8 that makes the glob to go infinite loop
<nove>
jemk, but maybe instead checking for it, would be better to check for an unexpected value and print a error message if the case
<nove>
jemk, like if(VE_H264_STATUS & 0x0ffeffff) error_happened
bertrik has joined #linux-sunxi
pacopad_ has joined #linux-sunxi
FR^2 has quit [Quit: Connection reset by peer]
pacopad has quit [Ping timeout: 244 seconds]
diego_r has quit [Ping timeout: 240 seconds]
pacopad_ has quit [Ping timeout: 240 seconds]
leviathanch2 has quit [Ping timeout: 240 seconds]
FreezingCold has joined #linux-sunxi
deasy has quit [Remote host closed the connection]
heffer_ has joined #linux-sunxi
pacopad has joined #linux-sunxi
heffer has quit [*.net *.split]
montjoie[home] has quit [*.net *.split]
montjoie[home] has joined #linux-sunxi
pacopad_ has joined #linux-sunxi
_massi_ has quit [Remote host closed the connection]
toplessninja has quit [Quit: Page closed]
pacopad has quit [Ping timeout: 264 seconds]
pacopad_ has quit [Ping timeout: 265 seconds]
pacopad has joined #linux-sunxi
orly_owl has quit [Read error: Connection reset by peer]
orly_owl_ has joined #linux-sunxi
pacopad has quit [Ping timeout: 264 seconds]
pacopad has joined #linux-sunxi
leviathanch2 has joined #linux-sunxi
notmart has quit [Quit: notmart terminated!]
leviathanch2 has quit [Ping timeout: 240 seconds]
pacopad_ has joined #linux-sunxi
pacopad has quit [Ping timeout: 244 seconds]
pacopad has joined #linux-sunxi
pacopad_ has quit [Ping timeout: 240 seconds]
andhe_ has quit [Ping timeout: 264 seconds]
netlynx has joined #linux-sunxi
netlynx has joined #linux-sunxi
netlynx has quit [Changing host]
jemk has joined #linux-sunxi
t3st3r has quit [Remote host closed the connection]