<popolon>
libv, what do you think about pc-on-a-stick category
<popolon>
some was on HTPC category
<libv>
some?
<libv>
_all_
<popolon>
the name is not really clear to
<popolon>
some of _existing_
<libv>
where does pc on a stick start, and where does it end?
<lukas2511>
"PC-on-a-stick is a term describing a PC in a USB stick format, generally using ARM system on a chip."
<lukas2511>
so basically all systems in "usb-stick" format
<libv>
the original mk802 was clearly that
<popolon>
that's a stick shaped thing that is attached to HDMI port
<libv>
but recent ones are a lot fatter
<libv>
and have female hdmi connectors
<lukas2511>
than it's not a stick anymore
<popolon>
like a PC in a HDMI dongle/stick
<libv>
lukas2511: compare mk802 and mk802-iii
<libv>
which one is what?
<libv>
and where do you draw the line?
<libv>
in the end they are all meant to be connected to a TV
<lukas2511>
mh...
<libv>
hence a unified htpc category
<popolon>
generally that's one hard piece that is attached to HDMI port
<libv>
perhaps this category should be further clarified
<libv>
anyway, enough of this for now.
<popolon>
libv, not I didn't removed from HTPC, it don't make sense
<popolon>
this is a notion commonly used in press about arm devices
<lukas2511>
yea well, i think as long as you basically stick it directly into the hdmi port of a tv it is a "stick", even if it's quite big. female hdmi ports would require a cable, and therefore wouldn't be a "stick" anymore
<Turl>
uh, wall of text, let me read
<libv>
popolon: but then you have devices in 2 different classification categories
<libv>
anyway, not now
<popolon>
is this a problem ?
<popolon>
there are devices in A31 devices and tablets at the same time
<libv>
Turl: i hope popolon keeps his word, and doesn't muddle things up further until i have time to sort things out, so i will not have to suspend his access
<libv>
devices is the super category
<popolon>
generally, in wiki, category is not for sticking pure definition with pure ideas
<popolon>
but instead to help to find an easier way things
<popolon>
libv, if you blocked my acount I can't "muddle" anymore anyway
<libv>
popolon: what about listing, collecting, classifying do you not understand?
<popolon>
that's the goal yes
<popolon>
collecting indormation, classifing and listing them
<libv>
most of the work of classification is in finding structure in the data.
<popolon>
sure
<libv>
and why the fuck am i having to repeat myself again while i should be doing other things?
<lukas2511>
i think instead of a lot of (small?) categories and (near-)empty pages something like http://wiki.openwrt.org/toh/start would be a better thing
<popolon>
there was several categories by OS on A20-Cubieboard/Nand Images
<popolon>
I tried to unify names and typo
<popolon>
all changes with big black N are new pages / categories
<popolon>
easy to catch for re-reading changes
<maksimlin>
hi, has anyone built rhombus's u-bbot for sunxi6 before? it built fine for me uboot.bin but not the SPL - I suspect because SPL_CONFIG was not enabled but I have no idea about u-boots build setup so I could be barkign up the wornong tree?
<maksimlin>
CONFIG_SPL I mean - looking at the README it looks like it needs to be set to build the spl loader ?
<Turl>
maksimlin: allwinner doesn't use uboot SPL
<Turl>
they have their own bootloaders to initialize the hardware, boot0/boot1
<maksimlin>
Turl: oh! I think I've gotten mixed up then - I was trying to make a bootable sdcard ...
<Turl>
maksimlin: u-boot-sunxi doesn't support sun6i yet
<maksimlin>
Turl: also I should add its eventually for a A31 tab but for now Im happy to start with an A10
<Turl>
A10 is ok, that's sun4i
<maksimlin>
Turl: oh so the u-boot from http://git.hands.com/u-boot.git is too diff to use per the instruction on that wiki page?
<maksimlin>
Turl: would it help if I explain more what Im trying to achieve?
<maksimlin>
well I'll do it anyways in case someone here spots I'm doing something wrong or plain impossible ... :-)
<maksimlin>
so I currently have a A10 tab and also will soon get a A31 tab, the A31 is the one Moz is using for their tablet contributor program for FxOS
<popolon>
good night
popolon has quit [Quit: Quitte]
<Turl>
maksimlin: basically sun6i (A31) support on our software is scarce/nonexistent
Deskwizard has quit [Read error: Connection reset by peer]
<maksimlin>
Turl: also the reason I was trying to use uboot was because allwinners fork has fastboot mode in it, so that would give an easy way for people to easily flash new system/boot imgs into nand of a bricked device
Deskwizard has joined #linux-sunxi
<maksimlin>
because these tabs do not seem to have any hardware key combo wired up for the uboot thats already in the nand to go into fastboot mode!
<maksimlin>
Turl: sorry a couple more questions if you have time... 1. should I be able to build lunix-sunxi u-boot for sdcard boot for a random A10 tab or would there be board specific config I would need to find for my tab?
<maksimlin>
and 2. I bought a couple of Toms uSDcard uart breakouts http://linux-sunxi.org/MicroSD_Breakout hoping to get at uboot on my A10 tab but would you/anyone know what serial config u-boot might use? 115200N8 ?
akaizen has joined #linux-sunxi
Andy-D has joined #linux-sunxi
mmarker has joined #linux-sunxi
Black_Horseman has quit [Ping timeout: 245 seconds]
protoCall7 has joined #linux-sunxi
egbert has quit [Disconnected by services]
egbert has joined #linux-sunxi
<Turl>
maksimlin: you can ask your question on the list, no need to ask tom specially
akaizen has quit [Remote host closed the connection]
akaizen has joined #linux-sunxi
akaizen has quit [Ping timeout: 240 seconds]
Andy-D has quit [Ping timeout: 244 seconds]
apo_ has joined #linux-sunxi
mmarker has quit [Remote host closed the connection]
akaizen has joined #linux-sunxi
TheSeven has quit [Ping timeout: 272 seconds]
TheSeven has joined #linux-sunxi
payne has joined #linux-sunxi
cajg has quit [Ping timeout: 260 seconds]
cajg has joined #linux-sunxi
maksimlin has joined #linux-sunxi
xavia has quit [Quit: Leaving.]
payne has quit [Remote host closed the connection]
bbrezillon has quit [Ping timeout: 252 seconds]
ChALkeR has joined #linux-sunxi
<ChALkeR>
Hi all.
Quarx has joined #linux-sunxi
<ChALkeR>
Have anyone tried to run Qt QML apps on top of framebuffer with sunxi-mali userspace drivers?
<ChALkeR>
Somewhy, minimalegl is a lot slower than eglfs for me.
<ChALkeR>
An eglfs has terrible tearing.
<ChALkeR>
I'm using debian jessie packages for Qt.
<ChALkeR>
minimalegl gives something about 1fps.
<ChALkeR>
eglfs gives better results, but still not 60fps and it has tearing.
akaizen has quit [Remote host closed the connection]
akaizen has joined #linux-sunxi
akaizen has quit [Ping timeout: 240 seconds]
<maksimlin>
mripard: ping! a question about a31 and u-boot spl if you have a min?
<ChALkeR>
I used an LD_PRELOAD hack to override eglCreateWindowSurface, because Qt tries to create it with window=0, which is incorrect for mali binary drivers.
<mripard>
you'll have to use the boot0/boot1 bootloaders from Allwinner
<maksimlin>
mripard: thanks for confirming that! ok I'm fine to do that but would you know where to get boot0/boot1 for a31 in wiki I only found link to A20 ones ?
shineworld has joined #linux-sunxi
shineworld has quit [Changing host]
shineworld has joined #linux-sunxi
<mripard>
I don't know if its' publicly available
<maksimlin>
mripard: oh :-(
<maksimlin>
mripard: sorry to bug you, but would you have any ideas on how I could boot a custom uboot build on a a31 tab where I can only get it into FEL ?
<maksimlin>
ie. its 'bricked' by flashing bad boot or system.img into its nand
<maksimlin>
mripard: for instance I read about the usb booting via FEL in the wiki so I thought I could use that... ?
diego_r has joined #linux-sunxi
<maksimlin>
but I'm not sure what it means to hae a "FEL enabled" u-boot as I'm building the uboot from above repo which I'm pretty sure is the one that is working on the device and importantly for us has fastboot impl in it
<mripard>
yes, but you need to have a first stage bootloader, and u-boot isn't that.
<mripard>
dd it on your sd card, and try to boot it
<mripard>
it might work, it might not.
<maksimlin>
mripard: ok thanks! what is in that img?
<mripard>
a bootable image, with boot0, boot1 and u-boot
<maksimlin>
fantastic thanks!
<mripard>
but probably not for your board
<maksimlin>
I dont have a tablet myself yet but I can get others to try it
<maksimlin>
mripard: probably dumb question - but how does the prop. boot1 knwo where to find u-boot ?
<maksimlin>
mripard: I ask because what comes to mind is trying to swap out your u-boot with my custom one, which would go straight into fastboot mode
HeHoPMaJIeH has joined #linux-sunxi
HeHoPMaJIeH has joined #linux-sunxi
<Montjoie>
Could someone point me where is hans u-boot repo for having smp on mainline ? I cannot find it on wiki
<maksimlin>
mripard: I have to go now, but THANKS for your help, really appreciate it! I'll check in again later on in the channel logs to see if its possible to swap out the uboot
maksimlin has quit [Quit: ChatZilla 0.9.90.1 [Firefox 29.0/20140428193838]]
<alexvf>
rellla: unfortunately, without any success
<alexvf>
rellla: i don't really know where to start
<rellla>
alexvf: i haven't noticed such a bug, but i did not try as much videos ;) canyou post some info about the video, e.g. with mediainfo to get some infos about exact codec etc...
<alexvf>
rellla: the video stream is reported by ffmpeg as Stream #0.0[0x100]: Video: h264 (High), yuv420p, 1920x1080, 24 fps, 2k tbr, 90k tbn, 48 tbc
<alexvf>
rellla: maybe i can upload the video somewhere
<rellla>
unfortunately i can't do tests by myself atm.
<rellla>
do that, so others can reproduce this.
<alexvf>
rellla: the only special thing i see is that it has sei messages but that should not affect decoding, just optimize decoding
<rellla>
and remember, the whole libvdpau is poc and not ready for use ;)
<alexvf>
rellla: do you know when can i upload it? is there a server for that purpose or just use some external web service?
<rellla>
"sei" messages?
<rellla>
easily use some external space
<alexvf>
rellla: sei messages is an h264 concept
avsm has joined #linux-sunxi
<alexvf>
rellla: i know libvdpau is poc but i want to improve it, or improve whatever it will turn into when it left the "poc" status
<rellla>
ah. ok. you see i'm not a h264 guru. and as you said, jemk would be the one to ask for this mainly - if he is interested in that anymore.
<rellla>
improving is very good, but it will probably a road with dead end ...
<alexvf>
rellla: why would it be a road with dead end?
<rellla>
iirc jemk was working on some kind of cedrus driver or v4l device from scratch.
<alexvf>
rellla: yes, i hear something but i don't know if he even started
<rellla>
i don't now, too. have to go now. bbl
<alexvf>
rellla: my hope is that improvements in libvdpau will translate to that video driver also
ganbold has quit [Remote host closed the connection]
ganbold has joined #linux-sunxi
mmarker has joined #linux-sunxi
mmarker has quit [Changing host]
mmarker has joined #linux-sunxi
r000n has joined #linux-sunxi
enrico_ has joined #linux-sunxi
<popolon>
if someone know if the open-source NAND driver check for badblocks on nand (like livesuit does), and is able to avoid them ?
<popolon>
and, will a dd on block device file avoid them in this case ?
rz2k has quit [Read error: Connection reset by peer]
smotocel69 has joined #linux-sunxi
<smotocel69>
some news about libvdpau-sunxi?
<smotocel69>
jemk?
<smotocel69>
jemk ?
<smotocel69>
rellla?
64MAAYQ1J has quit [Ping timeout: 264 seconds]
Andy-D has quit [Ping timeout: 244 seconds]
<alexvf>
smotocel69: no news as far as i know
<alexvf>
smotocel69: what are you looking for exactly?
<smotocel69>
vc1 decodign
<smotocel69>
decoding
<alexvf>
smotocel69: i think jemk has left libvdpau development
<alexvf>
smotocel69: he is aimed towards an v4l device driver
<smotocel69>
it's sad
<alexvf>
smotocel69: which i don't know if is wip or not
<alexvf>
smotocel69: you can try vc1 decoding by yourself, but as i am told when i ask about it, libvdpau-sunxi is poc code
<smotocel69>
alexvf i know :)
<smotocel69>
libvdpau sunxi must be reimplemented
<smotocel69>
as openmax
<alexvf>
smotocel69: omx woud be great too
<alexvf>
smotocel69: in fact, any of them, but i think there is not much people with the knowledge to do it
<alexvf>
smotocel69: i find myself trying to fix playback of some h264 files with libvdpau
<smotocel69>
alexvf right
<alexvf>
smotocel69: just to realize that setting random bits does not work :D
<smotocel69>
:))
<smotocel69>
i am waiting cubieboard 8 :)
<buZz>
so am i :)
<smotocel69>
will be fun
<buZz>
my fun will require sata though :P
<smotocel69>
:)
rz2k has joined #linux-sunxi
smotocel69 has quit [Quit: Leaving]
smotocel69 has joined #linux-sunxi
<rico1>
popolon: nand badblocks are not handled at the nand controller level, but by UBI or JFFS2 or yaffs...
ChALkeR has quit [Ping timeout: 265 seconds]
r000n has quit [Quit: Page closed]
Black_Horseman has quit [Ping timeout: 252 seconds]
shineworld has quit [Quit: Leaving]
[xeros] has joined #linux-sunxi
Andy-D has joined #linux-sunxi
Renard has joined #linux-sunxi
<Renard>
Would a Mele M6 be of use to any dev in here?
vindolin has joined #linux-sunxi
ganbold_ has joined #linux-sunxi
<mmarker>
Hmm. Anyone playing with 3.4.90 on a A20? I get a nice kernel panic when doing a lot of write I/O on a USB hard drive. Trying to get a backtrace, but want to see if I'm the only one alone on this.
<mmarker>
(Doesn't get logged in syslog - so I need to C&P in minicom when I get local to the board)
<popolon>
rico1, so a dd should not manage bad blocks and is not a good method to flash the nand ?
<popolon>
livesuit detect the bad blocks and seems to avoid them during flashing process
<rico1>
nope, to flash the nand, you should use flash_erase / nandwrite
<popolon>
sadly, there is nothing about that on linux-sunxi wiki
<Montjoie>
Does some people are trying the eudyptula challenge here ?
fredy has quit [*.net *.split]
<Renard>
MeleM5: "For building u-boot, use the "Mele_M5" target." → seems this target doesn't exist in the uboot git
<popolon>
mmmh mtd is the driver, not a tool
<Renard>
(only A1000, A1000G and A3700)
fredy has joined #linux-sunxi
<Renard>
also: "The .fex file can be found in sunxi-boards as mele_m5.fex" → 404
shineworld has joined #linux-sunxi
<popolon>
I perhaps start to understand
<popolon>
Are the mtd-utils compatible with linux-sunxi drivers ?
<rico1>
mtd-utils are for mtd devices (/dev/mtdx), linux-sunxi driver are far from being mtd device, (they simulate a block device AFAIK), so it's completely different.
<popolon>
ok, thanks
<rico1>
main visible difference : you can use ext4 (for example) on linux-sunxi driver, whereas with mtd, you can use UBI/UBIFS or jffs2, yaffs
<rico1>
(and you can't use ext4 with mtd)
issueat has joined #linux-sunxi
<popolon>
as far I understand you first need to mad a UBI/JFFS/YAFFS partition at low layer and you can add ext4 (for example) at higher level
<popolon>
(in case of the use of mtd)
<rico1>
yes, you could have ext4 on top of UBI, but only with the recent ubiblock patch (and read-only)
<rico1>
but I would not recommend that
<rico1>
AFAIK, the best solution on mtd is to use UBI + UBIFS.
<popolon>
but anyway that's not for sunxi devices today.
<popolon>
perhaps this incompatibility should be added somewhere in wiki, so user don't destroy or lost time with mtd utils on sunxi ?
<rz2k>
using ext above ubi is useless
<rz2k>
ubi needs a mtd aware fs above it
ChALkeR has joined #linux-sunxi
<rz2k>
popolon: bbrezillon developed MTD drivers and tested them on cubietruck board
avsm has quit [Quit: Leaving.]
<rz2k>
mainline ofcourse
<rz2k>
you can find discussion around mailing list
<arete74>
for boot on cubie1 i use same kernel but different dtb
<mripard>
you didn't set fdt_high
<arete74>
mripard: yes the uEnv.txt is same, modify only dtb
<wens>
i'm guessing we'll have -rc1 this weekend?
<mripard>
arete74: no, you didn't, otherwise, the DTB wouldn't be relocated
<mripard>
and it is relocated.
<lukas2511>
"fdt_high=ffffffff;" is that a valid command?
Andy-D has quit [Ping timeout: 240 seconds]
<lukas2511>
arete74: i think as command it should be something like "env set fdt_high ffffffff"
<mmarker>
mripard: I had the same issue with my 3.15 build last night, but I thought I did set fdt_high. Basically if I see that "relocation" line as in that error, I've screwed up somewhere?
<mripard>
I don't know if it's what's causing your issue, but it surely doesn't help
<mripard>
the reason we have to set this, is that if you look at the last line of the u-boot output, it relocates the DTB to 40ff8000
<mripard>
which is quite close to the kernel itself
<mripard>
and if the kernel is big, whenever it decompresses itself, it will overwrite the device tree.
<mmarker>
And things get clobbered. Been there, done that before.
ChALkeR has quit [Ping timeout: 245 seconds]
<arete74>
mripard: how you see what dtb is relocated?
<mmarker>
Arete74: last line from uboot in your error.
<lukas2511>
arete74: " Loading Device Tree to 40ff8000, end 40fff78d ... OK"
<arete74>
ok thanks
leviathanch2 has joined #linux-sunxi
patapovich has joined #linux-sunxi
<arete74>
is possible what uboot version is old?
HeHoPMaJIeH has quit [Quit: Leaving]
<mripard>
newer u-boot version don't need the fdt_high thing
<mripard>
but you can set it on all the versions we had
ChALkeR has joined #linux-sunxi
<patapovich>
Renard: i removed that picture from mele m6 wiki on purpose
<arete74>
mripard: ok re-check the uEnv.txt
<Renard>
patapovich, oh - what did I do wrong?
<Renard>
"Picture is not released under the Creative Commons Attribution"
<patapovich>
Renard: all material in wiki has to be your Intellectual property
<Renard>
YEs, I repalced it with one of my (crappy) shots
<patapovich>
ok
ChALkeR has quit [Ping timeout: 264 seconds]
patapovich has quit [Ping timeout: 246 seconds]
mmarker has quit [Remote host closed the connection]
<jemk>
alexvf: your artifacts are because of wrong reference picture lists, I simpliefied that part when writing libvdpau bacause it was good enough for my videos
<jemk>
alexvf: I started fixing this for interlaced h264, because there it leads to real problems, but never had time to continue
<alexvf>
jemk: ok, is there any field in the h264 headers to let me know what files are problematic?
<alexvf>
jemk: i can continue the work you started
<jemk>
alexvf: no, but you shouldn't need it. I'm currently trying to seperate that part from the interleaced experiments and hopefully push a working version
<alexvf>
jemk: that would be great
<alexvf>
jemk: but please, let me know if you need/want help
<alexvf>
jemk: i'm no expert in sunxi internals, but have read enough h264 pdf's to implement some of the algorithms and i'm willing to help
avsm1 has joined #linux-sunxi
avsm has quit [Ping timeout: 245 seconds]
<jemk>
alexvf: help is always welcome, but my h264 code is really hacky, it started small and got extended a lot without cleaning up
<jemk>
alexvf: i hoped to clean this while adding interlaced decoding, but the time..
<alexvf>
jemk: i could try to fix the reference pic list in the current version
<alexvf>
jemk: without you trying to separate that part from the other experiments
ChALkeR has joined #linux-sunxi
<alexvf>
jemk: just if you can give some indication of which is the point that should be fixed ...
<alexvf>
jemk: is the RefPicList0 and RefPicList1 generation from input reference_frames?
<jemk>
alexvf: yes, that part is responsible for your problem i guess, but it should be fixed now, see git
<alexvf>
jemk: one part of me wants to understand the changes but the other one can't wait to test it with my problematic file! :D
<alexvf>
jemk: it works well!
shineworld has quit [Quit: Leaving]
ChALkeR has quit [Ping timeout: 240 seconds]
xavia has joined #linux-sunxi
<jemk>
alexvf: whew, at least a small step
<alexvf>
jemk: thank you very much
<alexvf>
jemk: i had implemented a couple of reorder functions in my player
<alexvf>
jemk: but never integrated them because i though libvdpau took care of it
<jemk>
alexvf: if you want to fight with h264 algorithms, you could try to add the missing parts of ref_pic_list_modification(), i guess that will be the next thing leading to similar problems as soon as someone finds a video that makes use of it
libcg has quit [Ping timeout: 252 seconds]
<jemk>
yes, vdpau should take care of this, but I sometimes silently ignored this
<alexvf>
jemk: do you mean the mmco part?
<alexvf>
i never found a file with these
<jemk>
alexvf: that part is done by vlc/mplayer/whatever
<alexvf>
jemk: mmm, then i should do it too
<jemk>
alexvf: or to be correct, its done by ffmpeg/libav
<alexvf>
jemk: yes, i don't use ffmpeg/libav to decode so i should be doing it, which i don't
<jemk>
alexvf: oh, ok, have fun :D
protoCall7 has joined #linux-sunxi
<alexvf>
jemk: then, the missing parts you are talking about are long_term_refs?
<alexvf>
jemk: it continues to need the extra check on the surface being not null
<Turl>
Renard: judging by the picture, it has tSD inside
<Renard>
Turl: you're welcome? what's tSD?
<Renard>
s/?/!
<Turl>
Renard: it's like a nand chip, but not really, it has its own wear levelling and it's exposed as if it were mmc to the outside world
<jemk>
alexvf: ok, i'll try to check that again, but i think i never managed to reproduce that
<alexvf>
jemk: i only have to play a file with the slave interface and call switch ratio command
<Renard>
also, I complained to the vendor than the thing doesn't run XBMC as expected (didn't know about vdpau support on Allwinner beforehand); he's offering me to get another M6 for €15 shipped, which I could probably donate (right now, I'm awaiting an answer to see if he'd ship me a board not based on an Allwinner CPU first)
<alexvf>
jemk: i don't know if it depends on the file but i think it does not
<Renard>
Turl: at least there's an (internal) /sdcard mounted in Android by default, haven't explored much though
<alexvf>
jemk: but for sure, the file i uploaded also exhibits that behaviour
<Renard>
the actual external sd card mounts on /mnt/something
<alexvf>
jemk: the problem is that when changing aspect ratio reference frames are released, but the decoding code does not take that into account
protoCall7 has quit [Quit: protoCall7]
<alexvf>
jemk: iirc
<jemk>
alexvf: why the hell thry change anything at the decoder, aspect ratio is only relevant for output...
vindolin has quit [Ping timeout: 252 seconds]
<alexvf>
jemk: well i didn't go that far, it is just a thought
Andy-D has joined #linux-sunxi
<alexvf>
jemk: but what is true is that reference frames get released (or corrupted)
<alexvf>
jemk: i checked with the debugger
<alexvf>
jemk: but i didn't check why that happen
leviathanch2 has quit [Ping timeout: 240 seconds]
leviathanch2 has joined #linux-sunxi
<jemk>
alexvf: we must have different mplayers, mine doesn't crash with exact same commandline and file...
Skaag has quit [Ping timeout: 245 seconds]
Andy-D has quit [Ping timeout: 252 seconds]
<alexvf>
jemk: ii mplayer2 2.0-554-gf63dbad-1+b1 armhf next generation movie player for Unix-like systems
avsm1 has quit [Read error: Connection reset by peer]
<jemk>
alexvf: mplayer2... isn't it dead? it had lots of bugs last time i tried, in the meantime it got even removed from the repository
protoCall7 has joined #linux-sunxi
<alexvf>
which one do you use?
<alexvf>
mplayer doesn't work with vdpau afaik
<jemk>
mpv for normal usage, original mplayer to test osd
akaizen has joined #linux-sunxi
Andy-D has joined #linux-sunxi
bbrezillon has quit [Ping timeout: 264 seconds]
bonbons has joined #linux-sunxi
enrico_ has quit [Quit: Bye]
<alexvf>
jemk: i cannot easily install mpv in the board i using
avsm has joined #linux-sunxi
akaizen has quit [Ping timeout: 240 seconds]
Andy-D has quit [Ping timeout: 240 seconds]
<alexvf>
jemk: we can assume then that it is a bug in mplayer2
<alexvf>
jemk: anyway i guess the check is harmless, the decision on whether to pull the change is yours, feel free to close the pull request if you don't want to apply the workaround
<jemk>
alexvf: i guess so, the only path i can find (but i could miss something) that would introduce a invalid surface to a non-NULL RefPicList entry is getting an invalid handle from the player, but i thougt it would crash earlier then
avsm has quit [Client Quit]
akaizen has joined #linux-sunxi
<alexvf>
jemk: i guess the handle is ok, but not the pointer to the surface
<alexvf>
jemk: for whatever reason mplayer2 has
<alexvf>
jemk: btw when are mplayer guys stopping nonsense branching? :)
<jemk>
alexvf: the reason i didn't pull it is that we should never get a non-NULL RefPicList entry without valid surface, if we get this there must be a bug somewhere else in libvdpau
vindolin has joined #linux-sunxi
<jemk>
alexvf: and i didn't want to work around my own bugs that way
<alexvf>
jemk: i get your point
<alexvf>
alexvf: i will keep the workaround in my copy as i need it while i cannot test mpv
<alexvf>
lol i'm speaking to myself
vindolin has quit [Remote host closed the connection]
<alexvf>
jemk: if i find a legitimate situation where an invalid surface may get to libvdpau i will let you know but let's assume there isn't
rico1 has quit [Quit: Leaving.]
popolon_ has joined #linux-sunxi
popolon_ has quit [Read error: Connection reset by peer]
<jemk>
alexvf: well, a buggy player is a legitimate situation, we should handle this, but at the correct place. but libvdpau-sunxi is full of places where a buggy/different player could crash it all
<jemk>
alexvf: fixing all these bugs would need a lot of work, remember libvdpau was only a test framework for me, i never paid much attention on this sort of bugs, my player didn't trigger them
<alexvf>
jemk: i understand focusing on making work that will be useful for a final implementation
<alexvf>
jemk: this kind of minor bugs doesn't provide nothing meaningful
<alexvf>
jemk: and will have to be handled in another way in v4l device driver or whatever it gets done
Quarx has joined #linux-sunxi
<alexvf>
jemk: implementing h264 functionality does provide something meaningful and reusable (or at least the knowledge needed to implement them)
avsm has joined #linux-sunxi
avsm has quit [Client Quit]
avsm has joined #linux-sunxi
avsm has quit [Client Quit]
avsm has joined #linux-sunxi
diego_r has quit [Ping timeout: 264 seconds]
ChALkeR has joined #linux-sunxi
bbrezillon has joined #linux-sunxi
FR^2 has quit [Quit: Connection reset by peer]
marcin_ has quit [Remote host closed the connection]
avsm has quit [Quit: Leaving.]
marcin_ has joined #linux-sunxi
CaptHindsight has quit [Quit: gone]
CaptHindsight has joined #linux-sunxi
issueat has joined #linux-sunxi
bbrezillon has quit [Ping timeout: 240 seconds]
bbrezillon has joined #linux-sunxi
Netlynx has joined #linux-sunxi
Netlynx has joined #linux-sunxi
Netlynx has quit [Changing host]
bertrik has joined #linux-sunxi
libcg has joined #linux-sunxi
shineworld has joined #linux-sunxi
ninolein_ has quit [Ping timeout: 240 seconds]
ninolein has joined #linux-sunxi
Andy-D has joined #linux-sunxi
ChALkeR has quit [Ping timeout: 240 seconds]
shineworld has quit [Quit: Leaving]
shineworld has joined #linux-sunxi
libcg has quit [Ping timeout: 240 seconds]
Turl has quit [*.net *.split]
Turl has joined #linux-sunxi
FDCX has quit [*.net *.split]
heffer has quit [*.net *.split]
bluse has quit [*.net *.split]
MackBoy has quit [*.net *.split]
heffer has joined #linux-sunxi
bluse has joined #linux-sunxi
FDCX has joined #linux-sunxi
MackBoy has joined #linux-sunxi
formruga has joined #linux-sunxi
bbrezillon has quit [Ping timeout: 264 seconds]
shineworld has quit [Quit: Leaving]
notmart has quit [Quit: notmart terminated!]
tomboy64 has quit [Remote host closed the connection]