Linux Archive

Linux Archive (
-   64 Studio User (
-   -   Fw: where is: linux-headers-2.6.29-1-common-multimedia (

Mathias Krause 01-31-2010 02:17 PM

Fw: where is: linux-headers-2.6.29-1-common-multimedia
Michael Jarosch wrote:
> Am Sonntag, den 31.01.2010, 13:02 +0100 schrieb Mathias Krause:
>> Hi list,
>> in the meantime, i did some more extensive searches through the web.
>> That leads me to that thread:
>> So, i installed
>> linux-image- and linux-source-
>> I booted into that kernel, but still NVIDIA installer says:
>> ERROR: Unable to find the kernel source tree for the currently running
>> kernel
> Unfortunately, I haven't got a nvidia card, but I'm trying to remember
> my experiences some time ago:
> I _guess_, you can tell the nvidia-installer to take a certain path as
> the source tree. You can find out, how it's done, by using the --help
> option.
Hi Mitsch,

i tried that out, too. But for some reasons that didn't help a lot. (See
other mail with nvidia installer log file)
> Otherwise: Why not just using the "nv" or "nouveau" driver by editing
> xorg.conf? Generally, the opensource-drivers are causing less problems -
> at least the "nv" one. Unfortunately, "nv" doesn't support 3D-hardware
> acceleration.
Oh ... at the moment i am online, with the nv driver. But i cannot get
the second monitor to run. I used 64studio 2.0 before, with a compiled
nvidia driver and that worked very very perfect, including 3D acceleration.

64studio-users mailing list

Ralf Mardorf 02-06-2010 12:22 PM

Fw: where is: linux-headers-2.6.29-1-common-multimedia
Mathias Krause wrote:
>> And because of the sources, I run Synaptic and saw that there aren't the
>> common headers in the amd64 repository. In any case I guess it is good
>> to compile the latest kernel and to add latest rtirq, when there's the
>> need to build the graphics module ... not generally, but at the moment
>> there seems to be a big step for the real-time kernel (and rtirq).
> That's an important thing i did not yet understand. What is the difference
> between the kernel-multimedia-amd64 and rt19-20100103? When looking at the
> naming of that kernels, the last one seems that rt19 has nothing todo with
> multimedia ... and the rt kernel seems to be to be more a snapshot, then a
> tested version.... (and remember, i got that rt19 kernel from a privat
> repository, not from the official one...)

There is the basic kernel, called vanilla kernel, it has got a major
number, e.g. 2.6.31 and a minor number e.g. .6.

You can get the source for 2.6.31 and a patch .6 separated or the
complete source

For real-time you need an additional patch, e.g. -rt19. The complete
name than will be

Older patches are available here:

I build a kernel for my 64 Studio. The patch needs to be
especially for the kernel version. The actual one is
patch-, this means patch -rt20 is the micro number and the
patch is only for kernel, available at

Today it's common only to use this rt patches and I'm nearly sure the 64
Studio guys are using it too. In the past there were 2 different patches

Anyway, even similar kernel versions, using the same patch could have
different configurations.

spinymouse-sudo@64studio:~/Desktop$ ls /boot/config*
/boot/config- /boot/config-

I could build a new kernel, but with another configuration
and call it e.g.

For example, there is a snd-hrtimer module, that did exist for 2.6.29
kernels and it exists also for 2.6.31 kernels. The multimedia kernel
isn't configured with it.

spinymouse-sudo@64studio:~/Desktop$ cat
/boot/config-2.6.29-1-multimedia-amd64 | grep CONFIG_SND_HRTIMER
spinymouse-sudo@64studio:~/Desktop$ cat /boot/config- |

You also could build a kernel 2.6.29-1-multimedia-amd64 with, resp. a
kernel without this option. Even if SND_HRTIMER is more
experimental than the real-time patch already is, you don't need to load
and use the module, so it doesn't have impact to the stability of the

But for example the number of CPUs could have impact.

spinymouse-sudo@64studio:~/Desktop$ cat
/boot/config-2.6.29-1-multimedia-amd64 | grep CONFIG_NR_CPUS
spinymouse-sudo@64studio:~/Desktop$ cat /boot/config- |

While CONFIG_NR_CPUS is set up to 512, there shouldn't be trouble, when
using a quad-core CPU, but when the value is 2, I guess there could be

A new kernel and a new rt patch are less proved than older once, but the
new once are the latest releases and they aren't just snapshots for
testing only.

Differences will be because of different kernel versions and because of
the configuration, but for same kernel versions differences aren't
because of snapshot developer directory versions, but only because of
configurations, or because there are additional patches, e.g. because of
an additional patch for a special kind of suspend function.

For the distro there could be dependencies too, e.g. if you'll use a
kernel instead of a kernel 2.6.29-1-multimedia =
2.6.29.x-rt, you also need latest rtirq script, because of changes for
setting up preemtion/real-time values.

You need to ask Quentin about possible "special configurations" or
additional patches, but I'm sure his kernel is a preemtion/real-time
kernel and I'm nearly sure that the multimedia kernel don't comes with
additional patches, but the regular rt patch.
64studio-users mailing list

Ralf Mardorf 02-06-2010 12:40 PM

Fw: where is: linux-headers-2.6.29-1-common-multimedia
> Anyway, even _similar_ kernel versions, using the same patch could
> have different configurations.

Oops, it should be:

Anyway, even *same* kernel versions, using the same patch could
have different configurations.
64studio-users mailing list

All times are GMT. The time now is 10:39 PM.

VBulletin, Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.
Content Relevant URLs by vBSEO ©2007, Crawlability, Inc.