cnxsoft changed the topic of #imx6-dongle to: i.MX6 HDMI Dongle Software Development. You can also join us at https://groups.google.com/forum/?fromgroups#!forum/imx6-dongle - IRC Logs - http://irclog.whitequark.org/imx6-dongle/
Greg_ has quit [Quit: Greg_]
fossxplorer has quit [Ping timeout: 248 seconds]
hste has quit [Ping timeout: 240 seconds]
rz2k has quit []
cnxsoft has joined #imx6-dongle
Noskcaj has quit [*.net *.split]
wumpus has quit [*.net *.split]
Noskcaj has joined #imx6-dongle
wumpus has joined #imx6-dongle
Noskcaj10 has joined #imx6-dongle
Noskcaj has quit [Ping timeout: 276 seconds]
drachensun has quit [*.net *.split]
drachensun has joined #imx6-dongle
Greg___ has joined #imx6-dongle
ajayr has quit [Remote host closed the connection]
ajayr has joined #imx6-dongle
hste has joined #imx6-dongle
Greg___ has quit [Quit: Greg___]
<projectgus> abrasive: so i just fired up the ubuntu image to try and poke around the mmc io starvation stuff you were talking about
<abrasive> any juice?
<projectgus> I'm sure I could reproduce it on thursday night (I just ran one dd if=/dev/zero of=test.bin and everything got unresponsive)
<projectgus> I can't reproduce it now
<abrasive> heh
<projectgus> even running multiple parallel dds to a file, doing whole-filesystem finds, dd-ing straight from the block device
<projectgus> no matter what I do, both the X session and my ssh session stay pretty solid
<abrasive> changed anything?
<projectgus> nope
<projectgus> not a thing, just shutdown and rebooted
<abrasive> there's a little bit of indeterminism in these things coming from somewhere.
<abrasive> i've had times when you have to give it some serial input during boot or it gets stuck
<projectgus> the kernel?
<abrasive> yeah.
<abrasive> well, maybe early userland actually
<projectgus> weird
<projectgus> ah ok
<projectgus> hmm ok, I kicked off a bonnie++ session (lots of small file writes I think), and things got a little jumpy
<projectgus> what sort of things did you see it choking on before?
<abrasive> small files from memory
<abrasive> it's been a while
<hste> projectgus: could be the cache for the sdcard is filled up
<abrasive> the cache must be <=1gb though.
<projectgus> the vfs cache? yeah I haven't been explicitly dropping it
<projectgus> it's possible that's what happens if you do enough bulk writes, suddenly small reads become very slow because they're queued
<abrasive> i found that unrelated IO (like ssh) became ultra bad
<projectgus> yeah that's what I was trying to reproduce
<abrasive> did you get that when you were having trouble?
<projectgus> and it certainly seemed like i saw it on thursday, switching terminal tabs became quite time-consuming
<projectgus> i didn't have ssh up so I couldn't test anything quite that simple
<projectgus> it's possible switching tabs in gnome terminal actually requires file i/o :)
<projectgus> most of the slowdown I'm seeing now, running bonnie++, I
<projectgus> 'd attribute to slow filesystem reads
<projectgus> just echoes from typing and things like that still seem ok
<projectgus> *shrug*
<projectgus> oh snap, just hit it then
<projectgus> or possibly I've actually locked the whole thing up
<projectgus> got one tab switch in, but my ssh session is hung on 'cat /proc/interrupts'
<projectgus> ah the thermal pad fell off
<projectgus> perhaps this lockup is of the "cpu overheated" variety
<projectgus> but for a minute there is was responding, hjjust very slowly
<abrasive> maybe my issue was overheating too.
<abrasive> my thermal pad was falling off for a while
<projectgus> rebooting, thermal pad now attached with rubber band
<abrasive> heh
<projectgus> unsure if thermal rubber band or not
<projectgus> i have some tiny heat sinks I bought for my 3d printer stepper drivers ($5 for 20), I might test them out on the GK802 first
<projectgus> yeah now I think about it the thermal pad probably fell off on thursday night too
<abrasive> i just yanked the internal SD and put the lid back on.
<abrasive> console cable out the end
<abrasive> boot from USB + external SD
<projectgus> sensible
<projectgus> at least for development purposes
<abrasive> yeah. when uboot's cooked i'll pop it back in
<projectgus> alright, restarted my test rig (one bonnie++ run, plus one dd of the entire sd card's contents)
<projectgus> monitoring temps now, hovering aorund 61-62
<projectgus> system seems reponsive, I/O a little hesistant but not beyond what you'd expect
<projectgus> will let it run for a while, see how warm the rubber band can get before it deforms
<projectgus> anyone have any idea why the load averages are reported wrong btrw?
<projectgus> seem to all be offset by different amounts
<projectgus> still fine
<projectgus> i reckon it's temperature problems
<projectgus> mmc i/o definitely pegs a core without much trouble
<projectgus> uh-oh, 65
<projectgus> weird
<projectgus> got slow
<projectgus> cat /sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_cur_freq
<projectgus> 396000
<projectgus> seems a bit on the slow side
<projectgus> .... and locked up again
<projectgus> some stuff's still working, like it's still responding to pings
<hste> projectgus: in dmesg u should see if the freq goes down
<projectgus> hste: due to thermal throttling you mean?
<hste> yes
<projectgus> my dmesg is being spammed with some noisy debug message
<projectgus> I suspect this might have been the ondemand scaler turning it ruight down
<hste> u could use a kernel with less debug turned on
<projectgus> this looks something someone put in for debugging purposes, but yes
<projectgus> it's just one message, OnAction_p2p, about 500x a minute
<hste> what kernel are you using?
toktok has joined #imx6-dongle
<projectgus> the one from jas-hack's post, the thermal throttling enabled one
MindBeat has quit [Ping timeout: 252 seconds]
<hste> that shouldnt have so much debug on. U could check /proc/config.gz
<hste> And check for debug
<toktok> hi everyone. i just managed to run the ubuntu 12.04 image from jas-hacks. everything works find apart from audio...should have dd'ed the 3.0.35-0269 kernel for 12.04 as well. I would appreciate any suggestions on how to move forward
<projectgus> hste: that's not legitimate debug I don't think, someone's thrown a printk() in when they were testing something
<abrasive> no idea here.
<hste> projectgus: do u have some special devices pluged in?
<abrasive> probably the wifi driver
<abrasive> spammy crap that is
<hste> toktok: what resolution do u have?
<toktok> 720p
<projectgus> hste abrasive: yeah I just found it, it's a stray line in the wifi driver
<hste> toktok: have u done the dbus fix?
<projectgus> vendorstyle debugging
<projectgus> will send a pull request
<toktok> yes the network-manager shows
<projectgus> at some point :)
<toktok> the audio volume i can control
<toktok> and the audio device is shown to be imx-hdmi-soc
<hste> toktok: do u have some noice on the speakers or is it totally quiet?
<abrasive> the wifi driver is awful
<projectgus> unsurprising :)
MindBeat has joined #imx6-dongle
<toktok> it is totally quiet. the tv sound works obviously with other channels
<hste> abrasive: mine is working great after I lost the antenna wire . 3-4MB/s
<abrasive> hilarious.
<hste> and thts B not b
<abrasive> with the case on?
<hste> no the case is off
<hste> but a heatsink mounted on the cpu
<projectgus> hste: that's awesome :)
<projectgus> hste: so you replaced the rubber block?
<hste> just strapped a heatsink onto it
<projectgus> right
<projectgus> have switched scalers from ondemand to performance, running tests again
<projectgus> temps marginally higher, late 60s
<projectgus> first threshold set for the thermal zone is 80 though, so lots of room before it throttles
<projectgus> oddly enough the watch process I have cat-ting the thermal sensor freezes up periodically
<hste> projectgus: There is a patch for more accurate temp but its not on the git yet https://community.freescale.com/docs/DOC-94582
<projectgus> thanks
<projectgus> i'm not too fussed about accuracy assuming everything is relatively OK
<projectgus> i would like whatever causes it to not respond occasionally to not happen, though! :)
<projectgus> well i think this one has run longer than any other
<projectgus> *touch wood*
<hste> you could compile in CONFIG_TASKSTATS=y
<hste> CONFIG_TASK_DELAY_ACCT=y
<hste> CONFIG_TASK_IO_ACCOUNTING=y and try iotop
<hste> CONFIG_TASK_XACCT=y
<projectgus> true, but I'm pretty sure I know which processes are causing I/O
<projectgus> I imagine there's some fancy way I can use those tools to log slow I/O though
<abrasive> probably more effort than it's worth.
<projectgus> still good, I think maybe cpufreq is still a bit unstable
<projectgus> some kind of race condition, maybe with mmc maybe with something else
<projectgus> doesn't really bother me for my use case (yay!)
<hste> abrasive: do u have some fancy dongle-projects going on at moment? :)
<abrasive> have been trying to get yocto going for a while.
<hste> some recieps not working?
<abrasive> haven't figured out what i need to get DRI/DRM up
<projectgus> argh, spoke too soon, everything just locked up then
<projectgus> grumble
<projectgus> next suspect: wifi
<hste> abrasive: [meta-freescale] [meta-fsl-arm PATCH v3 12/13] xf86-dri-vivante: Fix DRI module name ?
<abrasive> where (when) is that from?
<abrasive> i get undefined symbols when trying to load the dri xorg extension.
<hste> its from the mailing list
<abrasive> ours? i should subscribe to that...
fossxplorer has joined #imx6-dongle
fossxplorer has quit [Changing host]
fossxplorer has joined #imx6-dongle
<abrasive> wow, otavio's been busy
<hste> :)
<abrasive> i don't think that bug affects my problem, i think i'm just missing something in my image.
<toktok> what is supposed to be the sound driver in the 12.04 image? I just checked and i am apparently using the 3.0.35-02695-ga09806b-dirty kernel which is supposed to have the sound fixed. but lsmod gives just vivante, drm, 8192cu
<toktok> do i need another driver for sound?
<hste> abrasive: what do u use. danny or master?
<abrasive> danny
<hste> abrasive: try master its more uptodate
<abrasive> okay.
<hste> abrasive: and use the rpm. the others arent tested that well
<abrasive> which rpm?
<hste> if u want it packed into rpm or dep or ipkg
<hste> I think it default is sat up to rpm
<hste> abrasive; you shuold als remove the build but can keep the downloads.
<abrasive> oh right
<abrasive> yes, i use rpm.
<hste> toktok: its compiled into kernel therfore u don't see the module
<toktok> oh ok. thanks.
<toktok> any other ideas to locate the problem?
<hste> abrasive: i tried the danny first also :) Got that advise from otavio
<abrasive> this looks much better
<projectgus> locked again w/ wifi off
<projectgus> going to try one more time without
<projectgus> running a watch process on the temperature
Noskcaj10 has quit [Ping timeout: 256 seconds]
jas-hacks_ has joined #imx6-dongle
toktok has quit [Quit: toktok]
Noskcaj has joined #imx6-dongle
Noskcaj has joined #imx6-dongle
Noskcaj has quit [Changing host]
<hste> jas-hacks_: morning :)
<jas-hacks_> hste: morning
<hste> jas-hacks_: time for fun today? :)
<abrasive> hste: thanks for the tips with yocto
<abrasive> looking forward to next image :)
<hste> jas-hacks_: do u mind if I use the new uboot in scriptinstall or do u want to wait till its ready?
<hste> abrasive: its not me but otavio that should be thanked :) Looks like he is doing a lot of work on it
<jas-hacks_> hste: Only have time in the evening to look at stuff
<abrasive> jas-hacks_: did you have a chance to look at my email?
<jas-hacks_> hste: You can use uboot, needs some clean up/changes based on abrasive's comments
<projectgus> hmm ok, so the lockups appear to have been not due to I/O but because I was checking the temp regularly
<projectgus> must be some kind of issue with reading it from userspace
<projectgus> good news though
<projectgus> haven't seen or reproduced i/o induced slowdowns yet, I think abrasive we both might have seen them due to thermal throttling
<jas-hacks_> abrasive: Haven't had time to refactor the code, will do that tonight
<abrasive> i think you're right, gus
<jas-hacks_> projectgus: is this with ubuntu?
<projectgus> jas-hacks_: yes, on your ubuntu image
<projectgus> I was trying to reproduce I/O induced slowdowns abrasive mentioned (lots of MMC I/O causing other things to become unresponsive)
<projectgus> and started getting lockups
<jas-hacks_> projectgus: In unity?
<projectgus> but I think it was because I was running "watch -n1 cat /sys/class/thermal/thermal_zone0/temp"
<projectgus> jas-hacks_: yes
<projectgus> once I stopped doing the above though I could complete tasks which were locking the whole system midway with that running in a separate terminal
<jas-hacks_> projectgus: There are number of issues :
<jas-hacks_> 1. gpu can look down to 1/64 of its clock speed when thermal temp is too hight
<jas-hacks_> 2. cpu will dramatically lower if freq when temp is too hight
<projectgus> k
<jas-hacks_> 3. There's some 'apt' processing happening in the background that consumes high cpu.
Noskcaj has quit [Ping timeout: 252 seconds]
<projectgus> yeah, I think abrasive and I may have both seen massive slowdowns when the thermal pad had fallen off the CPU, so that explains that
<projectgus> I was just running some tests to confirm, using bonnie & dd to load up the mmc cards
<projectgus> (I want to use the gk802 as a server, runnin RAID1 across the mmc slots, so wanted to be sure it wasn't going to cripple performance)
toktok has joined #imx6-dongle
<projectgus> but as I said, I didn't get any slowdowns
<jas-hacks_> I'm creating a Xubuntu releases that seems to behave better
<projectgus> but there seems to be a race where accessing /sys/class/thermal/thermal_zone0/temp can very occasionally lock the entire system
<projectgus> I'll investigate some more in the next few days
<projectgus> jas-hacks_: thanks very much for the details
<projectgus> off for now
<jas-hacks_> projectgus: I haven't had a lock up occurring, the main problem is the time it takes to lower the temp, it can be minutes
<projectgus> jas-hacks_: have you been 'cat'-ting that sysfs entry very regularly though?
<jas-hacks_> projectgus: no
<projectgus> yeah as I said I think that's what was doing it
<jas-hacks_> projectgus: 1st thing to do is get a proper heatsink
<projectgus> yeah I'll look into that
<projectgus> temps haven't gone above 70 while i've been monitoring so I don't actually think throttling is occuring
<jas-hacks_> projectgus: so you will be running a headless image?
<projectgus> yeah, my plan is to run debian headless
<projectgus> "debian, headless" I should say
<projectgus> although that'd be an awesome if confusing debian release name :P
<jas-hacks_> projectgus: then I expect it to behave well because you can run an armhf build
<projectgus> jas-hacks_: btw do you have any idea why the load averages are reported wrong?
<projectgus> nod
<hste> projectgus: running armhf wheezy works fine on the dongle
<jas-hacks_> projectgus: not sure at the moment
<projectgus> i figured, once you take out video most of the gnarly stuff is gone :)
<projectgus> (also)
<projectgus> OK so that sysfs thing doesn't seem to be enough to hang the system either, by itself
<projectgus> running while [ 1 ]; do cat /sys/class/thermal/thermal_zone0/temp > /dev/null; done & and it seems happy
<projectgus> figures, I imagine it's just reading a register somewhere so why caue issues?
<hste> projectgus: there is also someone made Gentoo on it
<abrasive> which uboot are you running, projectgus?
<projectgus> abrasive: um, at the moment the binary from jas-hacks' pot
<projectgus> post, even
<abrasive> jas-hacks_: does that have the RAM fixes in it?
<projectgus> anyhow I'm done for this evening, will investigate some more tomorrow
<jas-hacks_> abrasive: yes
<abrasive> ok.
<projectgus> thanks jas-hacks_, abrasive, hste
<abrasive> ez.
<toktok> hi again...i am a little bit at a loss trying to solve the sound problem. everything else seems to run fine with 12.04 image. but just cannot get the sound to work. I would appreciate any suggestions for toubleshooting
<abrasive> can you find /sys -name *edid*
<abrasive> and find the edid file there?
<toktok> yes
Noskcaj has joined #imx6-dongle
Noskcaj has quit [Changing host]
Noskcaj has joined #imx6-dongle
<abrasive> can you hexdump -C | head it? is it all zeroes/
<toktok> should i type "hexdump -C | head" did i unserstand correctly
<toktok> if i do "cat /sys/devices/platform/mxc_hdmi/edid"
<abrasive> hexdump -C /sys/path/to/file | head
<toktok> it is all zeros
<abrasive> oh.
<abrasive> okay
<abrasive> that's why your sound isn't working.
<abrasive> sorry, i thought it was binary.
<toktok> no problem..thanks for the help by the way
<abrasive> some GK802s (or some TVs?) don't seem to talk properly over DDC
<toktok> is there way to fix it
<abrasive> which is necessary both for video mode setting and for audio
<abrasive> we don't know yet...
<abrasive> haven't been able to reproduce the problem
<abrasive> i will be testing my tv out shortly though so hopefully it fails there ;)
<toktok> :-) strange wish to have....:-) but i like it
<abrasive> i can't help it, i'm a sucker for debugging.
<toktok> world needs ppl like u...thanks...so i should check back in once in a while.....to see your device failed...:-)
<abrasive> yeah :)
<toktok> just an idea...i have another dongle....uhost u1a
<toktok> and with that i get sound
<toktok> i presume there will be an edid file there too
<toktok> is it at all an option to copy that file here?
<toktok> or is this file automatically generated?
<abrasive> automatic, i'm afraid
<toktok> ok thanks...
<hste> toktok: you could try this files:
<hste> dd if=u-boot_multi_boot.imx of=/dev/sdb bs=1K seek=1
<toktok> cd I run dd from within? or do i need to take the sd card out and back in
<hste> if your card is /dev/sdb
<hste> you could try
<hste> if your card is /dev/mmcblk0
<toktok> also with jas' image just dd'ed the uboot and kernel
<toktok> what is the location for the boot.scr
<hste> and before trying boot copy uImage_cpufreq till /boot/uImage
<toktok> ok...
<hste> and cp boot.scr-internalsdcard /boot.scr
<toktok> oh ok...so i do not need to dd it....
<hste> and run a sync
<hste> u have to dd the uboot file
<hste> dd if=u-boot_multi_boot.imx of=/dev/mmcblk0 bs=1K seek=1
<toktok> yes....should dd the kernel or just copy it to /boot/uImage
<hste> but this uboot reeds /boot.scr and kernel from /boot/uImage
<hste> just copy the kernel not dd
<toktok> ok...so actually if i just copy it, it should work then.. thanks let me give it a shot...thanks...
<hste> do a sync before reboot
<toktok> ok
Noskcaj has quit [Remote host closed the connection]
<toktok> i rebooted after doing the copying...but the edid file is still all zeros
<toktok> the first time i tried to reboot it did not run ...the tv said unsupported video format....but i unplugged it and plugged it in again and this time the desktop came back...
<toktok> i presume the kernel names are the same....uname -r still gives 3.0.35-02695-ga09806b-dirty
<toktok> so i don't know if it is running the kernel which is in the first part of the sd card, or the one which i copied to /boot
<abrasive> if uboot updated it will use the one from /boot
<abrasive> unfortunately it's harder to check if uboot updated...
<hste> toktok: what resolution does your monitor handle?
<toktok> it is a tv can handle 1080p
<toktok> in fact with the android on the sdcard i was able to do 1080p and 720p both
<abrasive> android doesn't read the EDID as far as i know
<abrasive> or ignores it
<abrasive> well, my yocto build is nearly cooked, i can test this soon :)
<toktok> fingers crossed
<abrasive> what brand is your tv?
<toktok> phillips
<abrasive> ok.
<abrasive> i think the last one we got was a panasonic
<hste> toktok: Try this http://stende.no-ip.info/files/boot.scr and place it on / it Has hardcode 720P
<abrasive> you can't override modes if edid is broken, can you?
<hste> I'm not sure :)
<toktok> i will replace and report back
<hste> but I got sound when setting the values. But I got edid
<hste> I get a diffrent fbmodes file when setting it
jas-hacks_ has left #imx6-dongle [#imx6-dongle]
<hste> abrasive: what are u going to use the yocto build for?
<abrasive> xbmc
<abrasive> so i'll need a gstreamer patch i think.
<hste> nice
<hste> hack in gstreamer?
<abrasive> there is a gstreamer fork afaik
<abrasive> not sure how stable
<abrasive> but i'm not in a hurry, i have a miniITX xbmc box under my tv atm ;)
<abrasive> too bad i can only buid yocto at work
<hste> faster rig at work?
<abrasive> i don't have a desktop up at home atm actually.
<abrasive> i stole the PSU from mine to run my ZFS RAID...
<abrasive> and yes, a 16-core box with stacks of memory and fast disk.
<abrasive> -j16 is nice.
<toktok> ok it rebooted fine.....but still no audio....and the edid is again all zeros
<hste> I just got a new portable workstation from work. Only 8 cores i7, but 16GB and fast ssd
<abrasive> nice.
<abrasive> "only" 8 cores :p
<hste> :)
<abrasive> you make my desktop feel old.
<abrasive> (4 cores)
<hste> Actually I run the system in a vmware workstation giving it 10GB and 6 cores
<hste> have to have some windows stuff for work on it....
<hste> your dongles has 4 cores :)
<abrasive> yes well. my atom has 2 cores/4 threads. but it's really shit as a compiel box
<abrasive> also, getting yocto to build on arch is really hard
<abrasive> there are a bunch of bugs that show up with newer texinfo and gcc
<abrasive> and then gcc segfaults on some of the libgcc stuff *shudder*
<hste> yes I guess arch its not otavios build plattform :)
<hste> but building at work is not a big problem if the lines are fast for copy the result back
<toktok> hste: with a proper boot.scr would it be possible to boot from the external sd too? this would make testing extremely easy
<hste> toktok: yes
<toktok> great...so i put this at the root if the internal card and it sends it to the external card?
<hste> toktok: but it reads the internal first if you have a boot.scr there
<toktok> so i should put the boot.scr that boots the external on the internal card
<hste> toktok: or remove it from internal
<hste> toktok: you can switch putting the boot.scr on the external
<hste> toktok: its still a preview of the uboot jas-hack_ work on
<toktok> ok...so i remove the internal card. and the put the boot.scr you just send to the root of the external, and it should put...this would be great....since i fear that i am going to break something everytime i open it
<hste> toktok: u need to have the uboot there
<toktok> so actually i can put the sd card in the internal card to the
<toktok> external slot
<toktok> or should the uboot remain on an sd card in the internal slot
<hste> toktok: u need the uboot on the internal sdcard. Then u can put everything elsre on external or even usb
<toktok> ok...will try this....thanks for the help...then i should also learn how to customize boot.scr where can i read up on that
<hste> toktok: U have to compile the mkimage tool to make a boot.scr from a text file
<toktok> i have mkimage in my ubuntu system already
<toktok> the format/syntax is the same as the stuff for allwinner devices? i did it there a couple of times
<hste> toktok: mkimage -A arm -O linux -T script -n "boot" -d boot.txt boot.scr
<toktok> thanks again...i could i presume decode one of the boot.scr files you sent to see how to redirect where to look for uImage and the root system
<hste> toktok: yes
<toktok> thanks again...
<abrasive> okay! one open source fix submitted
<abrasive> it only took me all day...
<hp__> anyone have the cifs module available maybe ?
<hp__> or i need to build it myself hehe
<hp__> not yet made the toolchain
<hste> hp__: what kernel do u use?
fossxplorer has quit [Ping timeout: 264 seconds]
<hste> hp__: I think I compiled in cifsmodule in this one: http://stende.no-ip.info/files/kernel-latest.tgz
<hste> abrasive: what did u fix?
<hp__> hste i use the 1204 one
hste has quit [Remote host closed the connection]
hste has joined #imx6-dongle
<abrasive> hste: just an android kernel for my phone :)
<hste> :)
<abrasive> now it won't boot :(
<abrasive> cos i installed the wrong rom. genius.
<hste> what phone?
<abrasive> droid 4 (xt894)
hste has quit [Remote host closed the connection]
hste has joined #imx6-dongle
<abrasive> hste: do you have any experience getting vivante dri up?
<abrasive> i'm getting AIGLX messages saying the screen isn't dri or dri2 capable :(
<hste> abrasive: that only for aiglx part so its not important
<abrasive> i definitely don't have acceleration... the qt demos run horribly
<toktok> hste: just to report....booting from external sd worked. thanks a lot for pointing out the way...no sound still.... but i feel better now with playing around. hopefully you guys can solve the edid issue....
<hste> abrasive: how does xorg log look?
<abrasive> nothing obvious.
<abrasive> trying again with clean fs
<abrasive> wait, it was clean.
<hste> abrasive: here is a working xorg http://pastebin.com/PtvRTewU
<abrasive> thanks
<abrasive> so dri/dri2 *should* be there
<hste> yes but not in aiglx
<hste> /dev/dri/card0 is there?
<abrasive> oh, they're supposed to be integrated, according to the recipe.
<abrasive> i go see. it might just be my kernel source huh.
<hste> do you have it compiled the vivante stuff as modules or into the kernel?
<abrasive> i think it was compiled in. i go see.
<abrasive> CONFIG_DRM_VIVANTE=y
<hste> u got the vivante/dri.so files in the libs?
<abrasive> they got baked in.
<abrasive> there are no lib errors
<abrasive> the couldn't open dri message probably indicates it's not getting baked in like it's supposed to
<hste> I think there is one version with dri and one without
<abrasive> i don't think sO?
<abrasive> i ended up with /usr/lib/xorg/modules/extensions/libvivante_dri.so in my fs
<abrasive> so if that's not dri enabled i don't know what is :)
<hste> yes that looks like you got it :)
<abrasive> it just doesn't work :)
<abrasive> oh well, i'll sort it out some other time
<abrasive> i think i've hacked enough today
<hste> but it was some problems with naming of the file The Xorg Vivante driver has a hardcoded module name as 'vivante_dri'
<hste> so we need to rename it for full Xorg acceleration to work.
<hste> but it looks like u got the right file
<hste> but the one I got is /usr/lib/xorg/modules/drivers/vivante_drv.so
<abrasive> yes, i ahve that in my recipe
<abrasive> i'll strace it later
<hste> ok good night
<hste> :)
<abrasive> goodnight!
<hste> perhaps otavio is there when u start tomorrow :) I guess he could give u some good answers
<abrasive> yes, he'll probably have some clues
toktok has quit [Quit: toktok]
<abrasive> thanks for the hints anyway, hste :)
<hste> no probs :) u got firmware btw?
<abrasive> er... it comes with the bins, right?
<hste> yes it should: /lib/firmware
<hste> but I think the gpu stuff doesn't need firmware
rz2k has joined #imx6-dongle
hste has quit [Remote host closed the connection]
hste has joined #imx6-dongle
cnxsoft has quit [Remote host closed the connection]
toktok has joined #imx6-dongle
toktok has quit [Client Quit]
hp__ has quit [Ping timeout: 264 seconds]
hp__ has joined #imx6-dongle
toktok has joined #imx6-dongle
ajayr has quit [Ping timeout: 256 seconds]
ajayr has joined #imx6-dongle
toktok has quit [Quit: toktok]
Noskcaj has joined #imx6-dongle
Noskcaj has quit [Changing host]
Noskcaj has joined #imx6-dongle
Noskcaj has quit [Ping timeout: 256 seconds]
Noskcaj has joined #imx6-dongle
rz2k has quit []
iamfrankenstein has joined #imx6-dongle
fossxplorer has joined #imx6-dongle
fossxplorer has quit [Changing host]
fossxplorer has joined #imx6-dongle
hste has quit [Ping timeout: 240 seconds]
Noskcaj has quit [Ping timeout: 252 seconds]
fossxplorer has quit [Quit: Leaving]
rz2k has joined #imx6-dongle
<rz2k> otavio: fsl-image-gui-sdk doesnt have dri/dri2 included in xorg modules build for some reason and all 3D fails because of it
<rz2k> since -sdk inherits the fsl-image-gui, probably that xf86driproto dependency missing post at ML was right.