LMDE BREAKAGES - tracking TESTING

[anigramp]

I'm just pointing out that a new user to LMDE might get a bit 'worried' or get the wrong impression if they see a 48 page thread titled LMDE BREAKAGES.

I see that as a good thing. ONE thread on ONE site covers most issues. Just how many threads are there about "Windows Breakages"?

On-topic: I had a massive download last week (including the dnet nonsense); I took this to be GTK3 and possibly GNOME 3. Haven't rebooted yet Could be about to try Debian's KDE

[Chris M]

I found the problem of youtube over here, its not a flash problem.
Its a problem whit youtube it self, whit vimeo.com the vids are working fine also whit other sites.

Until recently, I had no problem. And like you, it does not happen for all video playback. The vimeo site works fine. CNBC videos work fine. Even Frontline videos work fine.

Jordan at This Week in Linux moved his videos off youtube and onto his site. Same problem there as with youtube.

To be specific, video stops dead, but audio keeps moving along.

If you move the web page in your browser, the video goes again, but then stops as the sound rolls on.

I'm using a 3+ year old Dell Vostro 1500 laptop with integrated Intel graphics, Intel everything else - chipset, and 4 gigs of good ram.

Like I said, this is a new situation. I followed Clem's advice, and still no go. And the linked advice about sound wasn't applicable.

Is anyone else having this issue?

Thanks

Looking back over some of the pages on this thread (starting around page 33) where folks were talking about compiz problems related to Intel graphics, Xorg, and I think the 2.6.39-2-686-pae kernal, I stopped looking at a flash issue. I've been a bit frustrated over the last couple of days - considered ways to downgrade the 686 kernel.

This morning, the 3.0.0-1-686-pae kernel hit the testing repo!! Oh yeah!!

Not that the new kernel is better, but I think its installation may have re-compiled my Intel graphics driver. The updates took care of this issue. There were a few, but I'm pretty sure the new kernel took care of it.

/var/log/apt/history.log

Code: Select all

Start-Date: 2011-08-09  02:37:38
Commandline: apt-get upgrade
Upgrade: xsltproc:i386 (1.1.26-7+b1, 1.1.26-8), base-files:i386 (6.4, 6.5), libdbus-1-3:i386 (1.4.12-5, 1.4.14-1), linux-image-2.6-686:i386 (2.6.39+35.1, 3.0.0+39), python-libxml2:i386 (2.7.8.dfsg-3, 2.7.8.dfsg-4), libxslt1.1:i386 (1.1.26-7+b1, 1.1.26-8), linux-libc-dev:i386 (2.6.39-3, 3.0.0-1), virtuoso-minimal:i386 (6.1.2+dfsg1-1, 6.1.3+dfsg1-1), libpcsclite1:i386 (1.7.2-2, 1.7.4-1), openssh-client:i386 (5.5p1-6, 5.8p1-7), dbus:i386 (1.4.12-5, 1.4.14-1), libvirtodbc0:i386 (6.1.2+dfsg1-1+b2, 6.1.3+dfsg1-1), libxml2:i386 (2.7.8.dfsg-3, 2.7.8.dfsg-4), lxmenu-data:i386 (0.1.1-1, 0.1.2-1), mint-artwork-xfce:i386 (2011.07.30.1, 2011.08.08), libencode-locale-perl:i386 (1.02-1, 1.02-2), libxml2-utils:i386 (2.7.8.dfsg-3, 2.7.8.dfsg-4), virtuoso-opensource-6.1-common:i386 (6.1.2+dfsg1-1+b2, 6.1.3+dfsg1-1), dbus-x11:i386 (1.4.12-5, 1.4.14-1), pcscd:i386 (1.7.2-2, 1.7.4-1), virtuoso-opensource-6.1-bin:i386 (6.1.2+dfsg1-1+b2, 6.1.3+dfsg1-1), libltdl7:i386 (2.4-2, 2.4-3)
End-Date: 2011-08-09  02:38:33

Start-Date: 2011-08-09  02:40:44
Commandline: apt-get dist-upgrade
Install: linux-image-3.0.0-1-686-pae:i386 (3.0.0-1, automatic)
Upgrade: linux-image-686-pae:i386 (2.6.39+35.1, 3.0.0+39)
End-Date: 2011-08-09  02:41:14

kruijf, upgrade to the new kernel.


Edit: I just noticed that it may not have been the new kernel 3 that fixed my Intel graphics issue (apt-get dist-upgrade). It looks like the preceeding apt-get upgrade upgraded the old 2.6.39 kernal first. That may have done it.

[Gerd50]

With the new kernel 3.0.0-1-686-pae my system won't boot. There are several Fatal Errors in the boot text, driver modules
not loading.

[viking777]

With the new kernel 3.0.0-1-686-pae my system won't boot. There are several Fatal Errors in the boot text, driver modules
not loading.

Yep I got that too. It boots into a terminal login, but no graphics and a whole load of kernel modules that could not be inserted. These are the ones that I managed to record from the last page of the login screen (the only ones I could see)

cpufreq_nforce2
longrun
speedstep_ich
p4_clockmod
powewnow_k7
speedstep_centrino

There were at least a couple more pages of them that were out of view (and one of those must have been nvidia I guess). I don't have boot logging enabled so I can't give you the whole list.

Comprehensively broken would be my first impression.

EDIT. I noticed that the upgrade as offered by synaptic did not include the headers file so I downloaded that separately (it is in synaptic) but it still won't boot.

[viking777]

Could this be the problem?

Code: Select all

Setting up linux-image-3.0.0-1-686-pae (3.0.0-1) ...
Running depmod.
Examining /etc/kernel/postinst.d.
run-parts: executing /etc/kernel/postinst.d/dkms 3.0.0-1-686-pae /boot/vmlinuz-3.0.0-1-686-pae
dkms.conf: Error! No 'BUILT_MODULE_NAME' directive specified for record #0.
Error! Bad conf file.
File: 
does not represent a valid dkms.conf file.

Noticed on a reinstallation, don't know what to do about it though. (except wait for somebody to rewrite dkms.conf).

[GeneC]

Could this be the problem?

Code: Select all

Setting up linux-image-3.0.0-1-686-pae (3.0.0-1) ...
Running depmod.
Examining /etc/kernel/postinst.d.
run-parts: executing /etc/kernel/postinst.d/dkms 3.0.0-1-686-pae /boot/vmlinuz-3.0.0-1-686-pae
dkms.conf: Error! No 'BUILT_MODULE_NAME' directive specified for record #0.
Error! Bad conf file.
File: 
does not represent a valid dkms.conf file.

Noticed on a reinstallation, don't know what to do about it though.

Viking

That is the same problem I had yesterday, viewed here=>
http://forums.linuxmint.com/viewtopic.php?f=141&t=78844

For the first time since LMDE install last Dec. Nvidia did not rebuild on kernel upgrade (the debian dkms method -- borked).

The solution I found was to use SMXI to install the new kernel, and then reinstall Nvidia drivers.

http://smxi.org/

A good read of the "docs" is important

http://smxi.org/docs/

[viking777]

Thanks for that Gene, but dkms will install the Nvidia driver eventually, I am patient enough to wait. However I have found a temporary way around it as you can see here:

Code: Select all

uname -a
Linux acer 3.0.0-1-686-pae #1 SMP Sun Jul 24 14:27:32 UTC 2011 i686 GNU/Linux

I am not sure if it is worth it but if you want to do it this is how.

Boot the new kernel to the terminal login prompt, login then su to root. If you have an xorg.conf file on your system that uses nvidia or ati or something else then do this:

Code: Select all

mv /etc/X11/xorg.conf /etc/X11/xorg.conf.old

then

Code: Select all

modprobe nouveau

NB This command is only temporary, it won't survive a reboot.

You can then issue

Code: Select all

startx

But you will be running as root so better off first doing:

Code: Select all

su yourusername

Then

Code: Select all

startx

But really the whole thing is not worth it - especially with the nouveau driver. You could follow Gene's suggestion and install the nvidia driver with smxi, or you could do it the old way - manually, but personally I think I will wait for it to fix itself, although if it takes too long I might change my mind about that. From what I have read the 3.0 kernel is not a major upgrade despite what the numbers might suggest.

Before you boot back into the old kernel don't forget to

Code: Select all

sudo mv /etc/X11/xorg.conf/old /etc/X11/xorg.conf

[secipolla]

This thread should be removed from the 'sticky' threads and maybe even have its name changed. LMDE now uses the 'incoming/latest' repos and these 'breakages' don't reflect that.

[terdon]

Hi all, I have been trying to install LMDE on my desktop from a multisystem USB drive. Although the live session loads with no problem, both keyboard and mouse are unresponsive. I have posted the details here

I thought I'd repost here since this seems to be a 'LMDE breakage'. Sorry for the double post but this is really annoying. Any ideas?

[viking777]

LMDE now uses the 'incoming/latest' repos and these 'breakages' don't reflect that.

That is completely wrong.

If you read Clem's blog about LMDE you will see that it uses incoming, latest or testing. Nothing goes into incoming and latest without having gone through testing first. People who track testing get to test for breakages well before those on 'incoming' do. If you feel there aren't enough threads on 'incoming/latest' then start your own - that is what zerozero did with this thread.

This thread should be removed from the 'sticky' threads and maybe even have its name changed.

I personally have no problem with changing its name if it confuses people and I don't suppose zerozero would either.

[zerozero]

just a few brief notes:

1- LMDE uses as-of now as default the testing debian repos as default, this will change with the respin, but is didn't yet.

2- if we don't have a thread tracking the testing breakages, how do we help improving the incoming/latest repos?

3- i would like someone to explain to me (in simple words if possible) what's the issue with this thread after all? LMDE follows testing, testing sometimes breaks, we have a place to report the problems; it seems to me like a good plan, or not?

4- other distros do the same, in different ways, but they also breakages records, so...

[GeneC]

I think most would agree that this thread is of great value to all LMDE users.
Certainly those who have changed repos to the new Mint "incoming" or "latest" will want to check here first to see what the problems are, and those tracking "Debian testing" (or SID, or Experimental) will catch problems first.

Incoming and latest do have their own thread on their own forum, as is.
I guess it will just take a bit for folks to find their preferred solution.

[viking777]

3- i would like someone to explain to me (in simple words if possible) what's the issue with this thread after all? LMDE follows testing, testing sometimes breaks, we have a place to report the problems; it seems to me like a good plan, or not?

No issue with this thread whatsoever zerozero it has been, and still is, invaluable. But I think what secipolla means it that it should have a title like:

LMDE BREAKAGES (tracking Debian Testing) - read here first

to distinguish it from breakages in the incoming and latest branches. I don't think there is a problem with that suggestion if it makes things clearer for people. But it is still your thread, so call it what you like, whatever it is called, for me it is about the most important thread on the forum, the only one I am subscribed to and the first one I read every day.

[Chris M]

Could this be the problem?

Code: Select all

Setting up linux-image-3.0.0-1-686-pae (3.0.0-1) ...
Running depmod.
Examining /etc/kernel/postinst.d.
run-parts: executing /etc/kernel/postinst.d/dkms 3.0.0-1-686-pae /boot/vmlinuz-3.0.0-1-686-pae
dkms.conf: Error! No 'BUILT_MODULE_NAME' directive specified for record #0.
Error! Bad conf file.
File: 
does not represent a valid dkms.conf file.

Noticed on a reinstallation, don't know what to do about it though.

Viking

That is the same problem I had yesterday, viewed here=>
http://forums.linuxmint.com/viewtopic.php?f=141&t=78844

For the first time since LMDE install last Dec. Nvidia did not rebuild on kernel upgrade (the debian dkms method -- borked).

The solution I found was to use SMXI to install the new kernel, and then reinstall Nvidia drivers.

http://smxi.org/

A good read of the "docs" is important

http://smxi.org/docs/

Well that sucks. My Intel issue goes away, and a nouveau one takes its place.

I got the same thing on my desktop that uses the nouveau. During the installation, I also got this popup warning:

│ Linux 3.0.0-1-686-pae. In the new version some of the
│ drivers used on this system may require additional firmware files:
│
│ nouveau: nouveau/fuc41ad, nouveau/fuc41ac, nouveau/fuc409d,
│ nouveau/fuc409c, nouveau/nvc4_fuc41ad, nouveau/nvc4_fuc41ac,
│ nouveau/nvc4_fuc409d, nouveau/nvc4_fuc409c, nouveau/nvc3_fuc41ad,
│ nouveau/nvc3_fuc41ac, nouveau/nvc3_fuc409d, nouveau/nvc3_fuc409c,
│ nouveau/nvc0_fuc41ad, nouveau/nvc0_fuc41ac, nouveau/nvc0_fuc409d,
│ nouveau/nvc0_fuc409c

Most firmware files are not included in the system because they do not
│ conform to the Debian Free Software Guidelines. You may need to
│ reconfigure the package manager to include the contrib and non-free
│ sections of the package archive before you can install these firmware
│ files.


Full kernel 3 update result:

Code: Select all

Unpacking replacement linux-image-686-pae ...
Setting up linux-image-3.0.0-1-686-pae (3.0.0-1) ...
Running depmod.
Examining /etc/kernel/postinst.d.
run-parts: executing /etc/kernel/postinst.d/dkms 3.0.0-1-686-pae /boot/vmlinuz-3.0.0-1-686-pae
dkms.conf: Error! No 'BUILT_MODULE_NAME' directive specified for record #0.
Error! Bad conf file.

Unpacking replacement linux-image-686-pae ...
Setting up linux-image-3.0.0-1-686-pae (3.0.0-1) ...
Running depmod.
Examining /etc/kernel/postinst.d.
run-parts: executing /etc/kernel/postinst.d/dkms 3.0.0-1-686-pae /boot/vmlinuz-3.0.0-1-686-pae
dkms.conf: Error! No 'BUILT_MODULE_NAME' directive specified for record #0.
Error! Bad conf file.
File: 
does not represent a valid dkms.conf file.
run-parts: executing /etc/kernel/postinst.d/initramfs-tools 3.0.0-1-686-pae /boot/vmlinuz-3.0.0-1-686-pae
update-initramfs: Generating /boot/initrd.img-3.0.0-1-686-pae
run-parts: executing /etc/kernel/postinst.d/pm-utils 3.0.0-1-686-pae /boot/vmlinuz-3.0.0-1-686-pae
run-parts: executing /etc/kernel/postinst.d/zz-update-grub 3.0.0-1-686-pae /boot/vmlinuz-3.0.0-1-686-pae
Generating grub.cfg ...
Found background image: linuxmint.png
Found Debian background: linuxmint.png
Found linux image: /boot/vmlinuz-3.0.0-1-686-pae
Found initrd image: /boot/initrd.img-3.0.0-1-686-pae
Found linux image: /boot/vmlinuz-2.6.39-2-686-pae
Found initrd image: /boot/initrd.img-2.6.39-2-686-pae
Found Microsoft Windows XP Professional on /dev/sda1
done
Setting up gcc-4.5-base (4.5.3-3) ...
Setting up libelfg0 (0.8.13-3) ...
Setting up cpp-4.5 (4.5.3-3) ...
Setting up gcc-4.5 (4.5.3-3) ...
Setting up linux-headers-3.0.0-1-common (3.0.0-1) ...
Setting up linux-kbuild-3.0.0 (3.0.0-2) ...
Setting up linux-headers-3.0.0-1-686-pae (3.0.0-1) ...
Examining /etc/kernel/header_postinst.d.
run-parts: executing /etc/kernel/header_postinst.d/dkms 3.0.0-1-686-pae
dkms.conf: Error! No 'BUILT_MODULE_NAME' directive specified for record #0.
Error! Bad conf file.
File: 
does not represent a valid dkms.conf file.
Setting up linux-headers-686-pae (3.0.0+39) ...
Setting up linux-headers-2.6-686-pae (3.0.0+39) ...
Setting up linux-image-686-pae (3.0.0+39) ...

Viking, when you say that this will fix itself, that dkms will install the Nvidia driver eventually, what do you mean?

[viking777]

Viking, when you say that this will fix itself, that dkms will install the Nvidia driver eventually, what do you mean?

I can't tell you exactly how this will happen I am not that knowledgeable on packaging matters. If you want me to guess then I suppose that someone will rewrite the dkms.conf file that seems to me to be causing the problem, repackage it and release it in a working condition. But don't ask me how long it will take, it could be anything between hours and weeks.

I didn't quite understand from your post what this means:

My Intel issue goes away, and a nouveau one takes its place.

Are you saying that you have the new kernel on an intel machine that works but on a nouveau one it doesn't?

[Chris M]

Viking, when you say that this will fix itself, that dkms will install the Nvidia driver eventually, what do you mean?

I can't tell you exactly how this will happen I am not that knowledgeable on packaging matters. If you want me to guess then I suppose that someone will rewrite the dkms.conf file that seems to me to be causing the problem, repackage it and release it in a working condition. But don't ask me how long it will take, it could be anything between hours and weeks.

I didn't quite understand from your post what this means:

My Intel issue goes away, and a nouveau one takes its place.

Are you saying that you have the new kernel on an intel machine that works but on a nouveau one it doesn't?

Issue 1: Thanks for that. That's what I considered - that a fix would come down the pipe from Debian, and that would do the fixing. I guess I could try the SMXI work-around, or continue booting the older image. I'm beginning to think that the SMXI script is the only way to go. There are far too many graphics driver issues to deal with in testing. But, I have to believe that a Debian fix for the new kernel 3 issue will come through kind of quickly. Especially since nouveau is involved. Maybe not.

Issue 2: I edited my original post. I just noticed that it may not have been the new kernel 3 that fixed my Intel graphics issue (apt-get dist-upgrade). It looks like the preceding apt-get upgrade upgraded the old 2.6.39 kernel first. That may have done it. I don't know.

But yes, I'm having no issue with the kernel 3 and Intel now. My Intel graphics laptop boots up without issue, and my youtube problem is done.

I've got to "go to work". But if there's some log you want me to post, please ask and I do it later today. Thanks

[kmb42vt]

just a few brief notes:

1- LMDE uses as-of now as default the testing debian repos as default, this will change with the respin, but is didn't yet.

2- if we don't have a thread tracking the testing breakages, how do we help improving the incoming/latest repos?

3- i would like someone to explain to me (in simple words if possible) what's the issue with this thread after all? LMDE follows testing, testing sometimes breaks, we have a place to report the problems; it seems to me like a good plan, or not?

4- other distros do the same, in different ways, but they also breakages records, so...

This thread should remain as it is for more than obvious reasons. There has to be an ongoing thread for those who are and will be in the future continuing to test, report problems and troubleshoot updates coming straight from Debian testing. All the info garnished from this thread can be funneled to Clem and Co. to be used in putting together the next Incoming update package. You know this as well as I do and now hopefully others do now as well.

That being said, I agree with viking777 in that the title for the thread should be altered to...

LMDE BREAKAGES (tracking Debian Testing) - read here first

...in order to avoid any confusion. It also should remain as a "sticky" so testers don't have to hunt for it. You know how non-sticky threads end up getting buried rather quickly these days and this thread is vital for LMDE in my opinion.

[viking777]

But yes, I'm having no issue with the kernel 3 and Intel now

You see, now that is interesting. If an Intel driver can be inserted by the present dkms.conf file then that says to me that there isn't that much wrong with it and maybe it is the nvidia driver that is at fault here not the dkms conf file (the lesson here is treat all computer error messages with a pinch of salt). But that is only a guess, I really don't know for sure. If it is a problem with the Nvidia driver it may take longer to get fixed, maybe I should do a manual install of the latest Nvidia driver onto this kernel and see what happens.

I'll think about it.

[gavinhc]

Everyone who's posted in this thread about issues with the new kernel recompiling their graphics drivers appears to be using the i686 (32 bit) kernel. Has anyone using the amd64 kernel upgraded and experienced the same problem?

[viking777]

Everyone who's posted in this thread about issues with the new kernel recompiling their graphics drivers appears to be using the i686 (32 bit) kernel. Has anyone using the amd64 kernel upgraded and experienced the same problem?

You ask a very good question there gavin. I would like to know the answer to that too.

For years I have avoided 64bit because it has always been less stable than 32bit and never discernibly faster. Maybe the tipping point has arrived?