[arch-dev-public] Kernel 2.6.29 series moved to testing
Hi Changes: http://kernelnewbies.org/LinuxChanges added this request: http://bugs.archlinux.org/task/13486 broken binary modules: - aufs - lirc Please maintainers of those packages fix those. It seems the 2.6.29 kernel has a network issue, which will be fixed in .29.1 http://article.gmane.org/gmane.linux.kernel/811417 At the moment keep an downgrade kernel in your cache if you will be affected by this bug. greetings tpowa -- Tobias Powalowski Archlinux Developer & Package Maintainer (tpowa) http://www.archlinux.org tpowa@archlinux.org
Tobias Powalowski schrieb:
added this request: http://bugs.archlinux.org/task/13486
I thought we said that we only add smack if we can compile it =m (which we couldn't).
broken binary modules: - aufs - lirc Please maintainers of those packages fix those.
Are we still adding these weird aufs patches? IMO, the aufs/unionfs situation should be solved more cleanly. They both require us to patch the kernel and both are not too clean. There is an ongoing effort to push aufs2 to mainline, so maybe it is better to use aufs2 instead of aufs?
On Wed, 25 Mar 2009 11:47:08 +0100 Tobias Powalowski <t.powa@gmx.de> wrote:
Hi Changes: http://kernelnewbies.org/LinuxChanges
added this request: http://bugs.archlinux.org/task/13486
broken binary modules: - aufs - lirc Please maintainers of those packages fix those.
It seems the 2.6.29 kernel has a network issue, which will be fixed in .29.1 http://article.gmane.org/gmane.linux.kernel/811417 At the moment keep an downgrade kernel in your cache if you will be affected by this bug.
The new kernel runs perfectly on my system. I'm not affected by the network issue mentioned above. It looks like that NICs with a forcedeth chipset are affected. I have a Broadcom Corporation NetLink BCM5789 Gigabit Ethernet card and it works like before. Cheers, Daniel
Daniel Isenmann wrote:
On Wed, 25 Mar 2009 11:47:08 +0100 Tobias Powalowski <t.powa@gmx.de> wrote:
Hi Changes: http://kernelnewbies.org/LinuxChanges
added this request: http://bugs.archlinux.org/task/13486
broken binary modules: - aufs - lirc Please maintainers of those packages fix those.
It seems the 2.6.29 kernel has a network issue, which will be fixed in .29.1 http://article.gmane.org/gmane.linux.kernel/811417 At the moment keep an downgrade kernel in your cache if you will be affected by this bug.
The new kernel runs perfectly on my system. I'm not affected by the network issue mentioned above. It looks like that NICs with a forcedeth chipset are affected. I have a Broadcom Corporation NetLink BCM5789 Gigabit Ethernet card and it works like before.
Cheers, Daniel
Apart from the FPS in games going from ~50 to ~10, I have no issues (with an Intel 945GM). Going to try KMS but I haven't had much success with that during the RC stage... Allan
Am Mittwoch 25 März 2009 schrieb Allan McRae:
Daniel Isenmann wrote:
On Wed, 25 Mar 2009 11:47:08 +0100
Tobias Powalowski <t.powa@gmx.de> wrote:
Hi Changes: http://kernelnewbies.org/LinuxChanges
added this request: http://bugs.archlinux.org/task/13486
broken binary modules: - aufs - lirc Please maintainers of those packages fix those.
It seems the 2.6.29 kernel has a network issue, which will be fixed in .29.1 http://article.gmane.org/gmane.linux.kernel/811417 At the moment keep an downgrade kernel in your cache if you will be affected by this bug.
The new kernel runs perfectly on my system. I'm not affected by the network issue mentioned above. It looks like that NICs with a forcedeth chipset are affected. I have a Broadcom Corporation NetLink BCM5789 Gigabit Ethernet card and it works like before.
Cheers, Daniel
Apart from the FPS in games going from ~50 to ~10, I have no issues (with an Intel 945GM). Going to try KMS but I haven't had much success with that during the RC stage...
Allan By the way do we still need aufs?
-- Tobias Powalowski Archlinux Developer & Package Maintainer (tpowa) http://www.archlinux.org tpowa@archlinux.org
On Wed, Mar 25, 2009 at 6:47 AM, Tobias Powalowski <t.powa@gmx.de> wrote:
Am Mittwoch 25 März 2009 schrieb Allan McRae:
Daniel Isenmann wrote:
On Wed, 25 Mar 2009 11:47:08 +0100
Tobias Powalowski <t.powa@gmx.de> wrote:
Hi Changes: http://kernelnewbies.org/LinuxChanges
added this request: http://bugs.archlinux.org/task/13486
broken binary modules: - aufs - lirc Please maintainers of those packages fix those.
It seems the 2.6.29 kernel has a network issue, which will be fixed in .29.1 http://article.gmane.org/gmane.linux.kernel/811417 At the moment keep an downgrade kernel in your cache if you will be affected by this bug.
The new kernel runs perfectly on my system. I'm not affected by the network issue mentioned above. It looks like that NICs with a forcedeth chipset are affected. I have a Broadcom Corporation NetLink BCM5789 Gigabit Ethernet card and it works like before.
Cheers, Daniel
Apart from the FPS in games going from ~50 to ~10, I have no issues (with an Intel 945GM). Going to try KMS but I haven't had much success with that during the RC stage...
Allan By the way do we still need aufs?
Yeah, we do - lots of external projects use it (larch, Chakra, etc), and I do want to move the ISOs from unionfs to aufs eventually
Allan McRae schrieb:
Apart from the FPS in games going from ~50 to ~10, I have no issues (with an Intel 945GM). Going to try KMS but I haven't had much success with that during the RC stage...
I have the same chipset, and with 2.6.29-rc8, compositing in KDE was unusable (with 2.6.28, intel 2.6.3 and UXA enabled it was fine). About the colors: I found a message on a mailing list about it and a patch that was supposed to fix it, but it didn't help at all. Also, KMS would freeze the display (or the whole kernel, who knows) when exiting X sometimes and it would cause kernel oopses. KMS is only cool on the surface (the native resolution framebuffer, non-flickering X start and instant console switch are nice), but the whole implementation is immature.
Tobias Powalowski schrieb:
broken binary modules: - aufs - lirc Please maintainers of those packages fix those.
http://aufs.sourceforge.net/README.aufs2 Again about aufs: I have no idea if this is supposed to be built in-kernel, but the patches to the VFS seem to be minimal and I guess we can build it as a module. If we are going to use aufs, we should use aufs2. It is even supposed to be better than unionfs!
Am Mittwoch, 25. März 2009 11:47:08 schrieb Tobias Powalowski:
It seems the 2.6.29 kernel has a network issue, which will be fixed in .29.1 http://article.gmane.org/gmane.linux.kernel/811417 At the moment keep an downgrade kernel in your cache if you will be affected by this bug.
I want to underline this. I have serious problems on two different computers. * the intel xorg driver just produces garbage (tested UXA and EXA; vesa works fine) * forcedeth stops working after a few minutes * ipw2200 does not work (it does not find any networks) * on every boot it complains about broken superblock features and wants to run a fs check (ext4; different hds; totally random) -> make backups ;-) * s626 has to be blacklisted; otherwise udev will take a long time and a modporbe process keeps hanging and cannot be killed * there might be a problem with the nvidia graphics driver So you better shouldn't use it unless you don't need network access or you have backups of all your data. :-) -- Pierre Schmitz Clemens-August-Straße 76 53115 Bonn Telefon 0228 9716608 Mobil 0160 95269831 Jabber pierre@jabber.archlinux.de WWW http://www.archlinux.de
On Wed, 25 Mar 2009 16:27:45 +0100 Pierre Schmitz <pierre@archlinux.de> wrote:
Am Mittwoch, 25. März 2009 11:47:08 schrieb Tobias Powalowski:
It seems the 2.6.29 kernel has a network issue, which will be fixed in .29.1 http://article.gmane.org/gmane.linux.kernel/811417 At the moment keep an downgrade kernel in your cache if you will be affected by this bug.
I want to underline this. I have serious problems on two different computers.
* forcedeth stops working after a few minutes
See thread in the kernel ML.
* on every boot it complains about broken superblock features and wants to run a fs check (ext4; different hds; totally random) -> make backups ;-)
No problems here on ext4. No complains about broken superblocks, etc. Made several reboots to test.
* s626 has to be blacklisted; otherwise udev will take a long time and a modporbe process keeps hanging and cannot be killed
Can't be reproduced here, booting is normal as always.
* there might be a problem with the nvidia graphics driver
Works here with this controller: 02:00.0 VGA compatible controller [0300]: nVidia Corporation G70 [GeForce 7600 GS] [10de:0392] (rev a1)
So you better shouldn't use it unless you don't need network access or you have backups of all your data. :-)
Good advice as always ;)
IMHO we should wait for a .1 release (with the network fix) before move this kernel to [core]. -- Hugo
-------- Original-Nachricht --------
Datum: Thu, 26 Mar 2009 11:02:42 -0300 Von: Hugo Doria <hugodoria@gmail.com> An: Public mailing list for Arch Linux development <arch-dev-public@archlinux.org> Betreff: Re: [arch-dev-public] Kernel 2.6.29 series moved to testing
IMHO we should wait for a .1 release (with the network fix) before move this kernel to [core].
-- Hugo Network is fixed in -2 kernel, i try now if smack causes here freenx segfaults. building a -3 kernel withot smack no.
before we start signoff we should get aufs working too. greetings tpowa -- Neu: GMX FreeDSL Komplettanschluss mit DSL 6.000 Flatrate + Telefonanschluss für nur 17,95 Euro/mtl.!* http://dsl.gmx.de/?ac=OM.AD.PD003K11308T4569a
On Thu, Mar 26, 2009 at 9:53 AM, Tobias Powalowski <t.powa@gmx.de> wrote:
-------- Original-Nachricht --------
Datum: Thu, 26 Mar 2009 11:02:42 -0300 Von: Hugo Doria <hugodoria@gmail.com> An: Public mailing list for Arch Linux development <arch-dev-public@archlinux.org> Betreff: Re: [arch-dev-public] Kernel 2.6.29 series moved to testing
IMHO we should wait for a .1 release (with the network fix) before move this kernel to [core].
-- Hugo Network is fixed in -2 kernel, i try now if smack causes here freenx segfaults. building a -3 kernel withot smack no.
before we start signoff we should get aufs working too.
Running 2.6.29-4 on both of my machines and it seems to be stable, so a psuedo-signoff for i686 and x86_64. Is 2.6.29.1 supposed to be in the repos? -Dan
Am Montag 06 April 2009 schrieb Dan McGee:
On Thu, Mar 26, 2009 at 9:53 AM, Tobias Powalowski <t.powa@gmx.de> wrote:
-------- Original-Nachricht --------
Datum: Thu, 26 Mar 2009 11:02:42 -0300 Von: Hugo Doria <hugodoria@gmail.com> An: Public mailing list for Arch Linux development <arch-dev-public@archlinux.org> Betreff: Re: [arch-dev-public] Kernel 2.6.29 series moved to testing
IMHO we should wait for a .1 release (with the network fix) before move this kernel to [core].
-- Hugo
Network is fixed in -2 kernel, i try now if smack causes here freenx segfaults. building a -3 kernel withot smack no.
before we start signoff we should get aufs working too.
Running 2.6.29-4 on both of my machines and it seems to be stable, so a psuedo-signoff for i686 and x86_64.
Is 2.6.29.1 supposed to be in the repos?
-Dan It is now in the repos, haven't run db yesterday. As long as aufs/unionsfs issue is not solved it cannot move in [core], i don't use any of those so some help here would be appreciated.
greetings tpowa -- Tobias Powalowski Archlinux Developer & Package Maintainer (tpowa) http://www.archlinux.org tpowa@archlinux.org
Tobias Powalowski schrieb:
It is now in the repos, haven't run db yesterday. As long as aufs/unionsfs issue is not solved it cannot move in [core], i don't use any of those so some help here would be appreciated.
This is what I think we should do: - Drop unionfs - Apply aufs-standalone.patch from http://git.c3sl.ufpr.br/gitweb?p=aufs/aufs2-standalone.git;a=tree;h=88c15502... (this is already the 2.6.29 version, and the patch does not do anything bad, it only exports some functions which weren't exported before) - Now, it should be possible to compile aufs2 as an external module by checking out the above repo and running "make". - Build aufs2 utils which are also available on the same server. I think Aaron agreed with me on that in Jabber. Opinions? Problems: devtools and ISOs need to be moved from unionfs to aufs2.
Thomas Bächler wrote:
Tobias Powalowski schrieb:
It is now in the repos, haven't run db yesterday. As long as aufs/unionsfs issue is not solved it cannot move in [core], i don't use any of those so some help here would be appreciated.
This is what I think we should do:
- Drop unionfs - Apply aufs-standalone.patch from http://git.c3sl.ufpr.br/gitweb?p=aufs/aufs2-standalone.git;a=tree;h=88c15502...
(this is already the 2.6.29 version, and the patch does not do anything bad, it only exports some functions which weren't exported before) - Now, it should be possible to compile aufs2 as an external module by checking out the above repo and running "make". - Build aufs2 utils which are also available on the same server.
I think Aaron agreed with me on that in Jabber. Opinions?
Problems: devtools and ISOs need to be moved from unionfs to aufs2.
How hard would it be to convert from using unionfs to aufs2 in devtools? Allan
On Mon, Apr 6, 2009 at 8:23 AM, Allan McRae <allan@archlinux.org> wrote:
Thomas Bächler wrote:
Tobias Powalowski schrieb:
It is now in the repos, haven't run db yesterday. As long as aufs/unionsfs issue is not solved it cannot move in [core], i don't use any of those so some help here would be appreciated.
This is what I think we should do:
- Drop unionfs - Apply aufs-standalone.patch from http://git.c3sl.ufpr.br/gitweb?p=aufs/aufs2-standalone.git;a=tree;h=88c15502... (this is already the 2.6.29 version, and the patch does not do anything bad, it only exports some functions which weren't exported before) - Now, it should be possible to compile aufs2 as an external module by checking out the above repo and running "make". - Build aufs2 utils which are also available on the same server.
I think Aaron agreed with me on that in Jabber. Opinions?
Problems: devtools and ISOs need to be moved from unionfs to aufs2.
How hard would it be to convert from using unionfs to aufs2 in devtools?
Very easy. Minor changes to the mount syntax
Aaron Griffin wrote:
On Mon, Apr 6, 2009 at 8:23 AM, Allan McRae <allan@archlinux.org> wrote:
Thomas Bächler wrote:
Tobias Powalowski schrieb:
It is now in the repos, haven't run db yesterday. As long as aufs/unionsfs issue is not solved it cannot move in [core], i don't use any of those so some help here would be appreciated.
This is what I think we should do:
- Drop unionfs - Apply aufs-standalone.patch from http://git.c3sl.ufpr.br/gitweb?p=aufs/aufs2-standalone.git;a=tree;h=88c15502... (this is already the 2.6.29 version, and the patch does not do anything bad, it only exports some functions which weren't exported before) - Now, it should be possible to compile aufs2 as an external module by checking out the above repo and running "make". - Build aufs2 utils which are also available on the same server.
I think Aaron agreed with me on that in Jabber. Opinions?
Problems: devtools and ISOs need to be moved from unionfs to aufs2.
How hard would it be to convert from using unionfs to aufs2 in devtools?
Very easy. Minor changes to the mount syntax.
Great. Any chance of a patch now the kernel without unionfs is in [core]? I don't have time to look into this at the moment but would like to continue using the build tools... Allan
On Wed, Mar 25, 2009 at 11:27 AM, Pierre Schmitz <pierre@archlinux.de> wrote:
Am Mittwoch, 25. März 2009 11:47:08 schrieb Tobias Powalowski:
It seems the 2.6.29 kernel has a network issue, which will be fixed in .29.1 http://article.gmane.org/gmane.linux.kernel/811417 At the moment keep an downgrade kernel in your cache if you will be affected by this bug.
I want to underline this. I have serious problems on two different computers.
* the intel xorg driver just produces garbage (tested UXA and EXA; vesa works fine)
I am also experiencing this problem (on i686). I'm using XAA. I haven't tried UXA and EXA. I'll just downgrade to the kernel26 in core for now.
* forcedeth stops working after a few minutes * ipw2200 does not work (it does not find any networks) * on every boot it complains about broken superblock features and wants to run a fs check (ext4; different hds; totally random) -> make backups ;-) * s626 has to be blacklisted; otherwise udev will take a long time and a modporbe process keeps hanging and cannot be killed * there might be a problem with the nvidia graphics driver
So you better shouldn't use it unless you don't need network access or you have backups of all your data. :-)
--
Pierre Schmitz
Clemens-August-Straße 76 53115 Bonn
Telefon 0228 9716608 Mobil 0160 95269831 Jabber pierre@jabber.archlinux.de WWW http://www.archlinux.de
On Thu, Mar 26, 2009 at 6:08 PM, Eric Bélanger <snowmaniscool@gmail.com> wrote:
On Wed, Mar 25, 2009 at 11:27 AM, Pierre Schmitz <pierre@archlinux.de> wrote:
Am Mittwoch, 25. März 2009 11:47:08 schrieb Tobias Powalowski:
It seems the 2.6.29 kernel has a network issue, which will be fixed in .29.1 http://article.gmane.org/gmane.linux.kernel/811417 At the moment keep an downgrade kernel in your cache if you will be affected by this bug.
I want to underline this. I have serious problems on two different computers.
* the intel xorg driver just produces garbage (tested UXA and EXA; vesa works fine)
I am also experiencing this problem (on i686). I'm using XAA. I haven't tried UXA and EXA. I'll just downgrade to the kernel26 in core for now.
I forgot: I'm getting this problem for both kernel26 2.6.29-2 and -3.
* forcedeth stops working after a few minutes * ipw2200 does not work (it does not find any networks) * on every boot it complains about broken superblock features and wants to run a fs check (ext4; different hds; totally random) -> make backups ;-) * s626 has to be blacklisted; otherwise udev will take a long time and a modporbe process keeps hanging and cannot be killed * there might be a problem with the nvidia graphics driver
So you better shouldn't use it unless you don't need network access or you have backups of all your data. :-)
--
Pierre Schmitz
Clemens-August-Straße 76 53115 Bonn
Telefon 0228 9716608 Mobil 0160 95269831 Jabber pierre@jabber.archlinux.de WWW http://www.archlinux.de
Eric Bélanger schrieb:
I want to underline this. I have serious problems on two different computers.
* the intel xorg driver just produces garbage (tested UXA and EXA; vesa works fine)
I am also experiencing this problem (on i686). I'm using XAA. I haven't tried UXA and EXA. I'll just downgrade to the kernel26 in core for now.
XAA didn't work for me in months. I always had to use EXA or UXA.
Thomas Bächler wrote:
Eric Bélanger schrieb:
I want to underline this. I have serious problems on two different computers.
* the intel xorg driver just produces garbage (tested UXA and EXA; vesa works fine)
I am also experiencing this problem (on i686). I'm using XAA. I haven't tried UXA and EXA. I'll just downgrade to the kernel26 in core for now.
XAA didn't work for me in months. I always had to use EXA or UXA.
Also, try getting rid of your xorg.conf. Most of my troubles centered around stuff in there that needed removed... Allan
- just fixing the broken nouveau-drm driver (kernel modules should be built against the new kernel...) - I get a weird segfault in my x86_64 chroot here when trying to "mv" files. It leaves a .part file in the dest folder. Haven't seen this before. "cp" works well. [andyrtr@workstation64 trunk]$ mv xf86-video-nouveau-20090325.tar.bz2 ~/sources/ Speicherzugriffsfehler [andyrtr@workstation64 trunk]$ [andyrtr@workstation64 trunk]$ ls -lh ~/sources/xf86-video-nouveau-20090325.tar.bz2* -rw------- 1 andyrtr users 168K 25. Mär 17:15 /home/andyrtr/sources/xf86-video-nouveau-20090325.tar.bz2 -rw-r--r-- 1 andyrtr users 0 25. Mär 17:14 /home/andyrtr/sources/xf86-video-nouveau-20090325.tar.bz2.part dmesg: mv[5254]: segfault at 7ffec9988558 ip 00007f96c13643d6 sp 00007ffec9988560 error 6 in libattr.so.1.1.0[7f96c1362000+4000] mv[5267]: segfault at 7ffeeda9a8f8 ip 00007fb3e54753d6 sp 00007ffeeda9a900 error 6 in libattr.so.1.1.0[7fb3e5473000+4000] mv[5285]: segfault at 7ffe379cc828 ip 00007f962f3a73d6 sp 00007ffe379cc830 error 6 in libattr.so.1.1.0[7f962f3a5000+4000] mv[5332]: segfault at 7ffe014dc338 ip 00007f67f8eb73d6 sp 00007ffe014dc340 error 6 in libattr.so.1.1.0[7f67f8eb5000+4000] mv[5603]: segfault at 7ffe3235e198 ip 00007fe829d3b3d6 sp 00007ffe3235e1a0 error 6 in libattr.so.1.1.0[7fe829d39000+4000] mv[5930]: segfault at 7ffe2a0b6ee8 ip 00007fb721a943d6 sp 00007ffe2a0b6ef0 error 6 in libattr.so.1.1.0[7fb721a92000+4000] no idea if this is coreutils or kernel related. can somebody confirm this? -Andy
more issues here with file system operations also my 32bit chroot. [andyrtr@workstation64 tmp]$ touch foo [andyrtr@workstation64 tmp]$ mv foo foo1 [andyrtr@workstation64 tmp]$ mv foo1 ~/sources/ mv: listing attributes of „foo1“: Das Argument ist ungültig [andyrtr@workstation64 tmp]$ mv foo1 ~/sources/foo2 mv: Aufruf von stat für „foo1“ nicht möglich: Datei oder Verzeichnis nicht gefunden ... ==> Erstelle Paket ... -> Erstelle .PKGINFO-Datei... -> Füge Installations-Skript hinzu... -> Komprimiere Paket ... bsdtar: Couldn't list extended attributes: Das Argument ist ungültig bsdtar: Couldn't list extended attributes: Das Argument ist ungültig bsdtar: Couldn't list extended attributes: Das Argument ist ungültig bsdtar: Couldn't list extended attributes: Das Argument ist ungültig bsdtar: Couldn't list extended attributes: Das Argument ist ungültig bsdtar: Couldn't list extended attributes: Das Argument ist ungültig bsdtar: Couldn't list extended attributes: Das Argument ist ungültig bsdtar: Couldn't list extended attributes: Das Argument ist ungültig bsdtar: Couldn't list extended attributes: Das Argument ist ungültig bsdtar: Couldn't list extended attributes: Das Argument ist ungültig bsdtar: Couldn't list extended attributes: Das Argument ist ungültig bsdtar: Couldn't list extended attributes: Das Argument ist ungültig bsdtar: Couldn't list extended attributes: Das Argument ist ungültig bsdtar: Couldn't list extended attributes: Das Argument ist ungültig ==> Verlasse fakeroot-Umgebung. my /tmp is mounted as tmpfs and /home is reiserfs(3.6) -Andy
Status update: aufs is a real bitch and is not working now, Thomas will look at it on weekend. Buidling now a fixed kernel with working network subsystem, tomorrow morning network should work again as it should. greetings tpowa -- Tobias Powalowski Archlinux Developer & Package Maintainer (tpowa) http://www.archlinux.org tpowa@archlinux.org
participants (10)
-
Aaron Griffin
-
Allan McRae
-
Andreas Radke
-
Dan McGee
-
Daniel Isenmann
-
Eric Bélanger
-
Hugo Doria
-
Pierre Schmitz
-
Thomas Bächler
-
Tobias Powalowski