Forum

Hi, i just upgraded my beta install (apt-get update/upgrade) and CEC stop working on my setup. Last time a used my XBMC was a week ago without such problem. I don't know if it has been already reported.

I use the setting "make xbmc the active source..." and when i force the tv to activate CEC (anynet+ with samsung tv), the tv switch to HDMI3 source but my RPi is connected to an AVR which is connected to the HDMI2 port on TV... There is absolutly nothing connected on the TV HDMI3...

(2nd Oct, 2013 07:39 AM)mk01 Wrote: [ -> ]Dave, so you are telling me that you are missing file which is part of package you just installed ? Ok, I will believe you.

can you post

Code:

dpkg -L ir-keytable

Based on the output of the above I've rerun nano /etc/rc_maps.cfg and I've been able to comment out the line mentioned back on the 17th.. that had been originally quoted as rc_maps.conf!

Wife is now watching something so will check if this has had an effect on rc operation once shes gone to bed..

(2nd Oct, 2013 07:39 AM)mk01 Wrote: [ -> ]

(2nd Oct, 2013 07:31 AM)Dave400 Wrote: [ -> ]within xbmc, top right corner

what is TZ of XBMC ? tzconfig / tzselect will not selects it's TZ. xbian-config in beta2 will change XBMC TZ as well when changing system TZ. But you have to go and set, it is not trying to sync the setting automatically or so. Gotham is finally taking TZ from system settings.

TZ set in xbian-config as europe/london, also manually set the same in xbmc but it's showing 18:03 instead of 23:03..[/b]

@mpnico

I don't know if reported before (so exactly detailed written down) - but I have Samsung TV as well, also AVR in-between and can just confirm it happened already two times to me as well.

Last time maybe 2 months ago. As it happened only twice and during Beta2 preparation there were so much changes at CEC upstream, firmware on RPI, linux kernel, XBMC itself - that I was not even tried to look for a cause.

I don't even remember what was the special step to get it back, but I remember being changing PORT number in SETTING->SYSTEM->INPUT->CEC. Normally most of the time when this happens, removing ~/.xbmc/userdata/peripheral_data/rpi_2708_1001.xml and XBMC reload helps.

(I was currently checking name of the field in INPUT setup and it was PORT number before, with the latest XBMC there is PHYSICAL ADDRESS instead. but this is even better. you get the address when you stop XBMC, in console run
and you will see it on device with osd name "CECTester". )

please report back.
btw: although it happened twice on one RPI with Yamaha AVR and SAMSUNG TV, it never happened on other RPI + YAMAHA + SAMSUNG. the both Yamahas and TVs are not exactly the same models, but YAMAHA almost and have identical FW. SAMSUNG TV are different with year of production - never happened with the older one, the other is almost brand new (early this year).

btw: I'm using "make AS" as well.

---

(2nd Oct, 2013 08:05 AM)Dave400 Wrote: [ -> ]Based on the output of the above I've rerun nano /etc/rc_maps.cfg and I've been able to comment out the line mentioned back on the 17th.. that had been originally quoted as rc_maps.conf!

you see, we all do mistakes. myself included. thanks for pointing this out, will comment this in original post and change - not to confuse others if they will try the same.

and yes, please report.

I'm quite angry for the RCs not being working as it was working with A5. For B1X there were _zero_ changes in the package and before for Beta2 there have been changes two:

1) sysvinit script being copied into upstart template
2) someone from user base was changing /etc/lirc/harware.conf file as fix for something not working. As the commit was accepted I took that.

but problem is that I have no remote to test that (I have CEC on all devices) and for the standard setup (lirc_rpi module load and lirc start) it starts as requested - so lirc creates device, the script recognize lirc_rpi was loaded etc. So even if willing so much, I can't check & fix even the smallest issue on that topic and every question-answer with me is try and see.

So i will buy original swiss chocolate to the one hitting the pot there.

@mk01
You're so quick to bring detailled answer

Here is the result of the command :

I have to use 0.0.0.0 or 2.3.0.0 in the physical adress ?

I will also remove the xml file you provide, but for now, TV is busy, i'll try it later.

Thx.

@mpnico

try as with the delete first. use filling the address only if all other fail.

also have seen on other forums (and other distros) hard reset TV (meaning not only turning off but unplug from wallet for 30s plug back again).

Ok, i'll report back later, thx.

(2nd Oct, 2013 08:14 AM)mk01 Wrote: [ -> ]

(2nd Oct, 2013 08:05 AM)Dave400 Wrote: [ -> ]Based on the output of the above I've rerun nano /etc/rc_maps.cfg and I've been able to comment out the line mentioned back on the 17th.. that had been originally quoted as rc_maps.conf!

you see, we all do mistakes. myself included. thanks for pointing this out, will comment this in original post and change - not to confuse others if they will try the same.

and yes, please report.

I'm quite angry for the RCs not being working as it was working with A5. For B1X there were _zero_ changes in the package and before for Beta2 there have been changes two:

1) sysvinit script being copied into upstart template
2) someone from user base was changing /etc/lirc/harware.conf file as fix for something not working. As the commit was accepted I took that.

but problem is that I have no remote to test that (I have CEC on all devices) and for the standard setup (lirc_rpi module load and lirc start) it starts as requested - so lirc creates device, the script recognize lirc_rpi was loaded etc. So even if willing so much, I can't check & fix even the smallest issue on that topic and every question-answer with me is try and see.

So i will buy original swiss chocolate to the one hitting the pot there.

rc6 line was commented out with a # still just arrow keys and 'enter' button working, nothing else.. uncommented and stopped lirc in case irtables install would have an effect but nada...

time stuff.. any broadcast message (such as after sudo reboot) has correct time, just within xbmc where its wrong even with all timezones set to match..

(2nd Oct, 2013 08:05 AM)Dave400 Wrote: [ -> ]TZ set in xbian-config as europe/london, also manually set the same in xbmc but it's showing 18:03 instead of 23:03..[/b]

I would like to know where the diff of 5h is taken from. This would mean system TZ being set wrong at the moment when when ntpd was setting it. At least it looks to me like this.

the 23:03 was actual time at London I know (it's 2h to myself in winter due to DST - Bratislava), but did you check and confirmed with "date"?. sorry for asking stupid-like questions but you have not confirmed this - but it is important.

btw: you can check GMT in /etc/localtime (cat /etc/localtime - it is binary file, but last characters are TZ in text form), then running "ntpq -p" to see how ntpd is seeing the diff to reference and if you run "ntpdc sysinfo" you will get the reference time in absolute.

@mk01 : removing the xml file did the trick, cec is now working properly, not a beta 2 bug so. Thx again.

(2nd Oct, 2013 08:57 AM)mk01 Wrote: [ -> ]

(2nd Oct, 2013 08:05 AM)Dave400 Wrote: [ -> ]TZ set in xbian-config as europe/london, also manually set the same in xbmc but it's showing 18:03 instead of 23:03..[/b]

I would like to know where the diff of 5h is taken from. This would mean system TZ being set wrong at the moment when when ntpd was setting it. At least it looks to me like this.

the 23:03 was actual time at London I know (it's 2h to myself in winter due to DST - Bratislava), but did you check and confirmed with "date"?. sorry for asking stupid-like questions but you have not confirmed this - but it is important.

in top right of xbmc it's [19:07] and underneath date of Tue, Oct 1, 2013..
(for reference it's now 0:08, Wed, Oct 2, 2013)

(2nd Oct, 2013 08:57 AM)mk01 Wrote: [ -> ]btw: you can check GMT in /etc/localtime (cat /etc/localtime - it is binary file, but last characters are TZ in text form), then running "ntpq -p" to see how ntpd is seeing the diff to reference and if you run "ntpdc sysinfo" you will get the reference time in absolute.

cat /etc/localtime

ntpq -p
ntpdc sysinfo

(2nd Oct, 2013 08:57 AM)Dave400 Wrote: [ -> ]time stuff.. any broadcast message (such as after sudo reboot) has correct time, just within xbmc where its wrong even with all timezones set to match..

ok then Dave I will just ask you to update from apt as today new frodo was compiled and committed. that newest one I checked for correct time, but can't confirm for the one you probably have installed 2.9-10.11 ? I know it was correct for me and without reports to be wrong with 2.9-10.10 - that was XBMC 12.2 dated late August. 2.9-10.10 was Sep-12. the 2.9-10.12 is week newer. if still wrong, lets wait for others to confirm or not and then we look at your system, ok ?

---

(2nd Oct, 2013 09:09 AM)Dave400 Wrote: [ -> ]ntpdc sysinfo

Code:

No address associated with hostname

that all looks good, ntpdc [enter] in it's prompt "sysinfo" [enter]

(2nd Oct, 2013 09:11 AM)mk01 Wrote: [ -> ]ok then Dave I will just ask you to update from apt as today new frodo was compiled and committed. that newest one I checked for correct time, but can't confirm for the one you probably have installed 2.9-10.11 ? I know it was correct for me and without reports to be wrong with 2.9-10.10 - that was XBMC 12.2 dated late August. 2.9-10.10 was Sep-12. the 2.9-10.12 is week newer. if still wrong, lets wait for others to confirm or not and then we look at your system, ok ?

in system info via xmbc it shows at bottom above memory usage: xbmc 12.2 git: 20130829-18397e1 (compiled: Sep 8 2013)

(2nd Oct, 2013 09:11 AM)mk01 Wrote: [ -> ]

(2nd Oct, 2013 09:09 AM)Dave400 Wrote: [ -> ]ntpdc sysinfo

Code:

No address associated with hostname

that all looks good, ntpdc [enter] in it's prompt "sysinfo" [enter]

@Dave

to the RC story. as you are loading mceusb module for lirc (or was using lirc like this before) when lirc is starting it is getting this info from dmesg output (or better told startup script for lirc is doing the search and then it is setting the params for lirc binary).

and the command line which started lircd could be seen when you type "ps ax" as command. standard setup - dummy for me - with lirc_rpi module will produce such line:

can you send yours please?

(2nd Oct, 2013 09:28 AM)mk01 Wrote: [ -> ]@Dave

to the RC story. as you are loading mceusb module for lirc (or was using lirc like this before) when lirc is starting it is getting this info from dmesg output (or better told startup script for lirc is doing the search and then it is setting the params for lirc binary).

and the command line which started lircd could be seen when you type "ps ax" as command. standard setup - dummy for me - with lirc_rpi module will produce such line:

Code:

/usr/local/sbin/lircd --driver=default --device=/dev/lirc0 -u

can you send yours please?

seems to be missing the 'device' string...

full output of ps ax is

(2nd Oct, 2013 09:23 AM)Dave400 Wrote: [ -> ]

Code:

apt-get upgrade
Reading package lists... Done
Building dependency tree
Reading state information... Done
The following packages have been kept back:
  xbian-package-config-shell xbian-package-config-xbmc xbian-package-xbmc
  xbian-update
0 upgraded, 0 newly installed, 0 to remove and 4 not upgraded.

ah, kept back packages. when you do apt-get install xbian-package-config-shell, you get version which fixes the problem not providing kept back packages for further updates.

then you can just update with xbian-config (shell or XBMC) and all will be fine for future, but the new config-shell is essential to this.

newly upgraded systems (from beta1X or A5) will be fine as upgrading via apt-get install xbian-update (as in the instruction in opening post) will install the correct config-shell automatically, but you was upgrading to beta2 before that fix - that's why it stuck. apt-get install xbian-update would help as well, but if you do just config-shell and then using it (to avoid anymore hacking on command line) would confirm the fix. thanks in advance.

probably I should write a communiqué about this because other users with earlier update to B2 are for sure stuck as well.

Matus

---

(2nd Oct, 2013 09:33 AM)Dave400 Wrote: [ -> ]

Code:

753 ?        Ss     0:00 /usr/local/sbin/lircd --driver=default

seems to be missing the 'device' string...

yes, i see that. this could be because the mceusb module is creating different device and this is not fetched correctly from dmesg and probably code inside /etc/lirc/hardware.conf.

I was going through your posts (even in other threads) but could not find actual "dmesg". "lsmod" i found, but please send your dmesg. I will hack this to my system and will debug the process with "simulating" your environment.

then I will recheck the accepted commit on git, maybe wasn't si good after all but this is only assumption. by any chance do you know what was the last time, xbian ver when it was working ?

---

just some general info for all on of you

1) I would like to ask all of you testing/running Beta2. check actual status of "apt-get update; apt-get -s upgrade".

if any packages are being listed as "Kept Back", do

this will install fix and upgradeable packages will never be kept back again. this is quite significant as depending on time when you upgraded to Beta2, the list of kept packages and not being updated as needed can be long and even with some bugs being fixed, posted to repo you won't get them

2) as I was communicating two days ago, today new XBMC FRODO 12.2 was compiled and pushed to repo. So if by accident something stop working don't hesitate and report. I would appreciate also comments whether the few bugs (i remember for instance XBMC crashing on creating bookmark in video) reported are still open or fixed

3) there is also a brand new firmware (dated Oct-1). With firmwares the quality was changing like weather - and even already fixed bugs were appearing again after some time (not very pleasant)
Because of this I have not pushed it into updates, but for those willing to try you can install it with

if there are issues, please report back and reinstall back simply by

thanks for reading this!