[lmyllari]

I uploaded a new build to https://www.dropbox.com/s/1v1wimt221uyw8...r16453.tar

It's built from bf3ba726fbee8ebf9492900612bc3ecc9b95b43e at https://github.com/laurimyllari/OpenELEC...ntel-fixes and includes AE fixes from fritsch and kernel patches to make 5.1 PCM output work, my limited range passthrough and a VC-1 crash fix from fritsch.

[BLKMGK]

I just downloaded one of those images and tried to get it working in virtualbox, but no luck.

---

I am guessing that you got a new kernel with the latest openelec and some devices nodes have changed. I assume you are using the commands in http://forum.xbmc.org/showthread.php?tid...pid1547157

Take a look around in /sys/bus/acpi/devices/ and see where the node lies now (assuming it is there).

Yes, this is a new kernel (3.12.1) and it looks like what has happened was modprobe -r got butchered off of the end of the script file I'm using somewhere along the way. Stupid user error <sigh>

The Intel thread above was started by lmyllari regarding this issue, using the code in his support posting worked but looking closely it differed only slightly than what I'd found here and had used on Frodo. I think the original code will work. Chasing my tail it seems!

Code:

#!/bin/sh

# Work around NUC CIR issue
# see http://forum.xbmc.org/showthread.php?tid=176718&page=17 and
# https://communities.intel.com/thread/46259 for more information

modprobe -r nuvoton-cir
echo "auto" > /sys/bus/acpi/devices/NTN0530\:00/physical_node/resources
modprobe nuvoton-cir

What's interesting is that just now after a reboot this didn't actually work even running it manually and I had to try this from the Intel support post to get it working, running the script above from a shell din't bring the CIR up.

Code:

modprobe -r nuvoton-cir && echo "auto" | tee /sys/bus/acpi/devices/NTN0530\:00/physical_node/resources && modprobe nuvoton-cir

Functionally these appear to be the same thing so this makes no sense. Right now my system is up, remote works, sound works, and it appears to have found my SQL server so my watched marks are synchronized - insert SnoopyDance. I'm going to try rebooting it a few times with the original script and try to be sure of my results.

P.S. I'm on lmyllari's latest build. My thanks to him and others for their work on this!

[Selene]

VC-1 still crash .. though last 1 seconds longer

[BLKMGK]

Alright, it's working consistently through a boot although it's not paying attention to my advanced.xml. It worked once, XBMC's interface disappeared and then came back, all was working. Acted like a crash and restart of the interface. Now after a boot it's not, I'll deal for now as I know this isn't a standard setup many people use.

lmyllari - is there anything specific you would like attempted with this build?

[nickr]

I just downloaded one of those images and tried to get it working in virtualbox, but no luck.

---

I am guessing that you got a new kernel with the latest openelec and some devices nodes have changed. I assume you are using the commands in http://forum.xbmc.org/showthread.php?tid...pid1547157

Take a look around in /sys/bus/acpi/devices/ and see where the node lies now (assuming it is there).

Yes, this is a new kernel (3.12.1) and it looks like what has happened was modprobe -r got butchered off of the end of the script file I'm using somewhere along the way. Stupid user error <sigh>

The Intel thread above was started by lmyllari regarding this issue, using the code in his support posting worked but looking closely it differed only slightly than what I'd found here and had used on Frodo. I think the original code will work. Chasing my tail it seems!

Code:

#!/bin/sh

# Work around NUC CIR issue
# see http://forum.xbmc.org/showthread.php?tid=176718&page=17 and
# https://communities.intel.com/thread/46259 for more information

modprobe -r nuvoton-cir
echo "auto" > /sys/bus/acpi/devices/NTN0530\:00/physical_node/resources
modprobe nuvoton-cir

What's interesting is that just now after a reboot this didn't actually work even running it manually and I had to try this from the Intel support post to get it working, running the script above from a shell din't bring the CIR up.

Code:

modprobe -r nuvoton-cir && echo "auto" | tee /sys/bus/acpi/devices/NTN0530\:00/physical_node/resources && modprobe nuvoton-cir

Functionally these appear to be the same thing so this makes no sense. Right now my system is up, remote works, sound works, and it appears to have found my SQL server so my watched marks are synchronized - insert SnoopyDance. I'm going to try rebooting it a few times with the original script and try to be sure of my results.

P.S. I'm on lmyllari's latest build. My thanks to him and others for their work on this!

Yes and no they are the same thing. The && between the commands ensures that the previous command has completed successfully before the next executes - if any one fails execution stops. But I am not sure why that is making a difference?

[lmyllari]

lmyllari - is there anything specific you would like attempted with this build?

If you see audio issues (besides the HBR streaming hanging audio device), fritsch is probably interested in reports.

I'll prepare another build later with audio debugging so we could maybe start figuring out the HBR audio streaming issue.

---

VC-1 still crash .. though last 1 seconds longer

Can you post logs?

---

What's interesting is that just now after a reboot this didn't actually work even running it manually and I had to try this from the Intel support post to get it working, running the script above from a shell din't bring the CIR up.

Code:

modprobe -r nuvoton-cir && echo "auto" | tee /sys/bus/acpi/devices/NTN0530\:00/physical_node/resources && modprobe nuvoton-cir

Functionally these appear to be the same thing so this makes no sense. Right now my system is up, remote works, sound works, and it appears to have found my SQL server so my watched marks are synchronized - insert SnoopyDance. I'm going to try rebooting it a few times with the original script and try to be sure of my results.

P.S. I'm on lmyllari's latest build. My thanks to him and others for their work on this!

If you post the output of the commands, lsmod and dmesg when it doesn't work, I can try to look into it. You may also want to take a look at the contents of the resources file to see if it was changed successfully.

[BLKMGK]

At this point I believe the remote issues are mostly of my own doing (and stupidity) and appear solved. I do think the Intel rep's attitude about the problem was amusing "it's fine if Windows works" <SMH> If I have remote issues again I'll be sure to grab dumps of logs to assist. I really appreciate your efforts on this for sure!

I'll try to put some audio through it. I did have one boot so far with audio issues but only the one. My biggest remaining issue is that OE seems to ignore my advancedsettings.xml file.I'll do some searching to try and solve this one and attempt to look over logs too. I'm nearly ready to just install this to the SSD and give up the desktop, it really is pretty slick and well done!

[Rezin]

My biggest remaining issue is that OE seems to ignore my advancedsettings.xml file.

What is the owner and group set to for that file? (ls -l advancedsettings.xml)

[micoba]

@Ymillari

First of all: Thank you for your hard work!

I tried your newest build and I don't have sound at all (not even from mp3). Additionaly movies are stuttering (m2ts, own recording, HD; h264 Bluray).
BUT: the picture quality is crystal clear!

Edit: Suddenly it works... I don't understand it...


How do I make a log? Do you have a link to a guide?

[Mangoes]

VC-1 still crash .. though last 1 seconds longer

VC1 doesn't crash. It works fine on its own and with FLAC audio. It's VC1 + DTS-HD that does.

If you use mkv, DTS-HDMA is useless anyway, it only makes your files 1 or 2 Gb bigger (and even more when it is true 16bit filled with "zeros" to make it appear as 24bit, as DTS-HDMA compresses very badly. These same files are 2x smaller with TrueHD, as well as with FLAC when losslessly corrected back to source 16bit with eac3to).

TrueHD is an efficient codec, files are between ~200Mb and 800Mb bigger than FLAC. But DTS-HDMA? What a waste of HD space.

---

I uploaded a new build to https://www.dropbox.com/s/1v1wimt221uyw8...r16453.tar

It's built from bf3ba726fbee8ebf9492900612bc3ecc9b95b43e at https://github.com/laurimyllari/OpenELEC...ntel-fixes and includes AE fixes from fritsch and kernel patches to make 5.1 PCM output work, my limited range passthrough and a VC-1 crash fix from fritsch.

Man, oh man. Do you rock? Yes you do.

[Selene]

That's not the point...it should work, and doesn't. I'm not here to argue codecs :x Also, not only using mkv.

[bry]

about to buy this - any naysayers? http://www.amazon.com/Intel-D54250WYK-Di...D54250WYK1

[Mangoes]

That's not the point...it should work, and doesn't. I'm not here to argue codecs :x Also, not only using mkv.

The point is that VC1 is working (I am watching VC1 movies everyday). Even the sample you sent was working when the DTS-HD was replaced. The problem is with DTS-HD, which already had problems with the new haswell (passthrough due to a buffer problem apparently fixed with a patch, etc...).

Furthermore, several people have already requested a log output from you during the past few days. I'm sure this would help in getting the problem solved?

[BLKMGK]

about to buy this - any naysayers? http://www.amazon.com/Intel-D54250WYK-Di...D54250WYK1

My only reservation would be the price. Vendors are gouging us early adopters. You might want to check and see the stock status of NewEgg, I posted a link a bit ago to their listing of the same hardware for a cheaper price.

[Selene]

Log isn't needed.. fritsch knows what the problem is already. Also, it only happens in VC-1 files it seems not any other ones, because I've got plenty of files that aren't VC-1 (DTS-HD MA) that work just fine. No need to be so hostile.