Siduction Forum

Siduction Forum => Scripting & Kernelhacking => Topic started by: towo on 2012/05/05, 18:02:32

Title: linux-3.4 rc5 for testing in our experimental repo
Post by: towo on 2012/05/05, 18:02:32: I have uploaded linux 3.4-rc5 for testing in our experimental repo.

If this repo is active
Code: [Select]
apt-get install linux-image-34-siduction-$arch linux-headers-34-siduction-$arch
will install that Kernel.
Title: RE: linux-3.4 rc5 for testing in our experimental repo
Post by: dibl on 2012/05/05, 18:46:31: @towo, please where is the location of the siduction experimental repo, bitte und vielen dank?
Title: linux-3.4 rc5 for testing in our experimental repo
Post by: vindeliker on 2012/05/05, 18:55:42: @towo, what's wrong here?
Code: [Select]
Vorkonfiguration der Pakete ... Vormals nicht ausgewähltes Paket linux-image-34-3.4-rc5-siduction-amd64 wird gewählt. (Lese Datenbank ... 294568 Dateien und Verzeichnisse sind derzeit installiert.) Entpacken von linux-image-34-3.4-rc5-siduction-amd64 (aus .../linux-image-34-3.4-rc5-siduction-amd64_3.4~rc5-1_amd64.deb) ... Vormals nicht ausgewähltes Paket linux-headers-34-3.4-rc5-siduction-amd64 wird gewählt. Entpacken von linux-headers-34-3.4-rc5-siduction-amd64 (aus .../linux-headers-34-3.4-rc5-siduction-amd64_3.4~rc5-1_amd64.deb) ... Vormals nicht ausgewähltes Paket linux-headers-34-siduction-amd64 wird gewählt. Entpacken von linux-headers-34-siduction-amd64 (aus .../linux-headers-34-siduction-amd64_3.4~rc5-1_amd64.deb) ... Vormals nicht ausgewähltes Paket linux-image-34-siduction-amd64 wird gewählt. Entpacken von linux-image-34-siduction-amd64 (aus .../linux-image-34-siduction-amd64_3.4~rc5-1_amd64.deb) ... linux-image-34-3.4-rc5-siduction-amd64 (3.4~rc5-1) wird eingerichtet ... Running depmod. Error setting debconf substitutions in linux-image-3.4-rc5-siduction-amd64/postinst/missing-firmware-3.4-rc5-siduction-amd64: linux-image-3.4-rc5-siduction-amd64/postinst/missing-firmware-3.4-rc5-siduction-amd64 doesn't exist at /var/lib/dpkg/info/linux-image-34-3.4-rc5-siduction-amd64.postinst line 602, <STDIN> line 3. dpkg: Fehler beim Bearbeiten von linux-image-34-3.4-rc5-siduction-amd64 (--configure): Unterprozess installiertes post-installation-Skript gab den Fehlerwert 255 zurück linux-headers-34-3.4-rc5-siduction-amd64 (3.4~rc5-1) wird eingerichtet ... Examining /etc/kernel/header_postinst.d. run-parts: executing /etc/kernel/header_postinst.d/dkms 3.4-rc5-siduction-amd64 linux-headers-34-siduction-amd64 (3.4~rc5-1) wird eingerichtet ... dpkg: Abhängigkeitsprobleme verhindern Konfiguration von linux-image-34-siduction-amd64: linux-image-34-siduction-amd64 hängt ab von linux-image-34-3.4-rc5-siduction-amd64 (= 3.4~rc5-1); aber: Paket linux-image-34-3.4-rc5-siduction-amd64 ist noch nicht konfiguriert. dpkg: Fehler beim Bearbeiten von linux-image-34-siduction-amd64 (--configure): Abhängigkeitsprobleme - verbleibt unkonfiguriert Fehler traten auf beim Bearbeiten von: linux-image-34-3.4-rc5-siduction-amd64 linux-image-34-siduction-amd64 E: Sub-process /usr/bin/dpkg returned an error code (1)

@dibl, have a look at http://packages.siduction.org
Title: linux-3.4 rc5 for testing in our experimental repo
Post by: dibl on 2012/05/05, 19:17:40: Quote from: "vindeliker"

@dibl, have a look at http://packages.siduction.org

Got it -- thank you.
Title: linux-3.4 rc5 for testing in our experimental repo
Post by: towo on 2012/05/05, 19:40:13: There was a typo in some templates, i will opload a fixed version soon.
Title: linux-3.4 rc5 for testing in our experimental repo
Post by: dibl on 2012/05/05, 19:49:15: Meanwhile I have installed it and booted it on a Dell Dimension 4700. I looked at dmesg and all looks good. This CPU will not support virtualization, so no VM is possible. I anticipate there will be the usual trouble with VMware on a 3.4 kernel.

Code: [Select]
System: Host: siductionbox Kernel: 3.4-rc5-siduction-686 i686 (32 bit, gcc: 4.7.0) Desktop: KDE 4.7.4 (Qt 4.8.1) Distro: siduction 11.1-rc One Step Beyond - kde-lite - (201112212155) Machine: System: Dell product: Dimension 4700 Mobo: Dell model: 0DH682 Bios: Dell version: A10 date: 01/04/2006 CPU: Single core Intel Pentium 4 CPU (-UP-) cache: 1024 KB flags: (nx sse sse2 sse3) bmips: 5984.99 clocked at 2992.497 MHz Graphics: Card: Intel 82915G/GV/910GL Integrated Graphics Controller bus-ID: 00:02.0 X.Org: 1.12.1 drivers: intel (unloaded: fbdev,vesa) Resolution: 1280x1024@60.0hz GLX Renderer: Mesa DRI Intel 915G x86/MMX/SSE2 GLX Version: 1.4 Mesa 8.0.2 Direct Rendering: Yes Network: Card: Intel 82562ET/EZ/GT/GZ - PRO/100 VE (LOM) Ethernet Controller driver: e100 ver: 3.5.24-k2-NAPI port: dcc0 bus-ID: 03:08.0 IF: eth0 state: up speed: 100 Mbps duplex: full mac: 00:11:11:c1:68:93 Drives: HDD Total Size: 74.4GB (-) 1: WDC_WD740GD Info: Processes: 199 Uptime: 8 min Memory: 661.6/2016.4MB Runlevel: 5 Gcc sys: 4.6.3 Client: Shell inxi: 1.7.29
Title: linux-3.4 rc5 for testing in our experimental repo
Post by: towo on 2012/05/05, 20:42:29: I have uploaded (hopefully) fixed packages, please remove the older packages before installing those new version.
Title: linux-3.4 rc5 for testing in our experimental repo
Post by: ralul on 2012/05/05, 20:59:44: @towo, please hold linux-3.3 kernel for a while, perhaps as a meta package:
linux-image-old

The new Kernel linux-3.4 has:

- abi x32,
which nobody having enough RAM needs
- btrfs better performance
which nobody should use (because of heavy changes daily in the code)
- xen and other virtualization featueres
which the simple minded of the desktop users don't need
- dm features,
same as above for the simple desktop user

But actual coming linux-3.3.5
is most promising performant and stable (*) ...

(*) tested without any bfs patches, because of some bad time accounting of bfs-for-linux-3.3 observed - I dont trust that patch anymore.
Title: linux-3.4 rc5 for testing in our experimental repo
Post by: dibl on 2012/05/05, 21:13:53: Quote from: "ralul"

- btrfs better performance
which nobody should use (because of heavy changes daily in the code)

@ralul -- question:

Is this a new enhancement to existing btrfs support, or is this the actual btrfs support in the kernel?
Title: linux-3.4 rc5 for testing in our experimental repo
Post by: devil on 2012/05/05, 21:19:18: Quote

Is this a new enhancement to existing btrfs support, or is this the actual btrfs support in the kernel?

it's an enhancement

greetz
devil
Title: linux-3.4 rc5 for testing in our experimental repo
Post by: ralul on 2012/05/05, 21:30:04: In linux-3.4 they actualy do better btrfs
- meta-data performance
a) when to commit optimized
b) larger chunks of bytes - less b-trees
- SuSE Jeff Mahony work for error prevention, in the sence of just stopping any more writes at the latest atomic transactions commit, when errors begin ...
Title: linux-3.4 rc5 for testing in our experimental repo
Post by: dibl on 2012/05/05, 21:30:59: OK -- thanks devil. I know there is/was a problem with VM performance on btrfs.
Title: linux-3.4 rc5 for testing in our experimental repo
Post by: ralul on 2012/05/05, 21:37:08: I dont think this VM performance issue is tangled by this work yet. This issue is a much more fundamental:

What to do with a cow system on really large files when changes happen on that files ...

I cannot imaging there is a real solution of this question. Seems to me, btrfs will be the better filesystem but not any more for all purpose (as ext4).
Title: linux-3.4 rc5 for testing in our experimental repo
Post by: dibl on 2012/05/05, 22:18:16: Quote from: "ralul"

What to do with a cow system on really large files when changes happen on that files ...

Yep -- well, my Win 7 VM is a really large file -- about 30GB. So, I think you are right -- maybe btrfs will never be suitable for such applications.