- openSUSE Kernel

[opensuse-kernel] Need help with a kernel .spec file - its for building a patched kernel
by doiggl＠velocitynet.com.au 29 Feb '08

29 Feb '08

Hello, Can you provide me with a .spec file I can use in the Opensuse Build Service to do the following things: - To build a customised, patched kernel in a installable rpm [i586 & x86_64] formats. I ask because I want to build the package properly. And I do not have advanced experience to build the .spec file myself. Could you provide me with a .spec file to handle the following items: Sources used: ------------- linux-2.6.24.3.tar.bz2 [file copied into home/userid/source/linux-2.6.24.3.tar.bz2] reiser4-for-2.6.24.patch [file copied into home/userid/source/reiser4-for-2.6.24.patch] .config [file copied into home/userid/source/.config] [This has the selected customised kernel configuration options on the following] ["append to kernel release" e.g -default-reiser4 and [File systems / Reiser4 (EXPERIMENTAL)]] Things to do: ------------- - Unpack the kernel source [home/userid/source/linux-2.6.24.3.tar.bz2] - Copy in a .config into the unpacked kernel either a i586 or x86_64 .config file from [home/userid/source/.config] so it will compile. - Apply patch to the unpacked kernel source [patch -p1 < home/userid/source/reiser4-for-2.6.24.patch] - Make installable rpm [i586 & x86_64] formats. - cleanup temporary files I know its a big learning experience for me, and I would like to contribute in some way. Thankyou for your help and time. Glenn -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe(a)opensuse.org For additional commands, e-mail: opensuse-kernel+help(a)opensuse.org

2 2

[opensuse-kernel] Xen kernel bug: new GENERIC_TIME based code breaks live migration
by Michael Marineau 28 Feb '08

28 Feb '08

Greetings, So I've been trying to get together a new xen kernel for Gentoo based on the Suse 2.6.22 xen patches in 2.6.22.17-0.1. So far I've hit two bugs. The easier of the two is below, I'll post the second bug in a second email since they are not related. Note that this is on i386, I haven't even tried x86_64 yet. Currently domU live migration between two hosts is broken. The guest's system clock (implemented in arch/i386/kernel/time-xen.c) is pulled directly from xen so time will typically change significantly when moving between two xen instances. In older versions of the xen patches this works fine but these 2.6.22 patches have replaced the old do_gettimeofday implementation with the GENERIC_TIME implementation which gets the time from the new xen_clocksource_read function. The problem is that xen_clocksource_read makes sure that time never goes backwards, either returning the new time if it went forwards or the previous read if time appears to have gone backwards. So when the system time jumps backwards during a move to a new physical machine suddenly time stops. I've worked around the issue for now by replacing time-xen.c with an older version from redhat's 2.6.21 xen kernel and disabling GENERIC_TIME. The attached patch does this for i386 but it leaves x86_64 broken since I haven't gotten to that yet. For kicks I tried letting xen_clocksource_read return a time in the past but that caused the kernel to get lost in an endless loop somewhere. Perhaps the system time should be kept locally within the kernel instead of pulling directly from xen and updated during the timer tick's wall clock update. Then xen_clocksource_read would use that time plus the change in TSC since the last tick instead of xen's time. Cheers, -- Michael Marineau Oregon State University mike(a)marineau.org -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe(a)opensuse.org For additional commands, e-mail: opensuse-kernel+help(a)opensuse.org

2 2

[opensuse-kernel] Xen kernel bug: oops when systems starts swapping
by Michael Marineau 28 Feb '08

28 Feb '08

Greetings once again, My second and far more nasty bug is that the kernel will oops when things start to swap in/out. I have had trouble finding a cause but here are some easy steps to force the machine into swap to reproduce it. Note that this only happens when running a normal build of xen, when xen is built with it's debugging option (suse's xen-dbg.gz or xen-pae-dbg.gz) I cannot trigger the oops. # replace 1700m with most any value larger than the ammount of ram available mount -t tmpfs -o size=1700m none /mnt/tmp1/ cat /dev/zero >/mnt/tmp1/dump cat /mnt/tmp1/dump >/dev/null The kernel will oops for me during one of the two cat commands depending on the system. On one system running a basic install of Suse it will oops almost instantly during the first cat, on another system running my own patched up kernel it will oops during the second cat. I tried running my custom kernel on the Suse system which results in a hardlock during the first cat. I don't know why the behavior is slightly different between the systems and kernels but the behavior is consistent when switching between pae and no pae and the amount of ram the system booted with. Switching to a debug build of xen always fixes the problem. I don't have any good clues as to why the xen build changes things but I assume it is a race condition and the debug build alters the timing slightly. A couple oops messages are attached, oops-custom is running my custom kernel and during the second cat. The two oops-suse messages are from the other system running suse and happen during the first cat command. Thanks for any ideas, -- Michael Marineau Oregon State University mike(a)marineau.org

1 0

[opensuse-kernel] HEAD update to 2.6.25-rc2-git6
by Jeff Mahoney 26 Feb '08

26 Feb '08

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi all - Just a quick FYI. As some of you may have already noticed, the HEAD kernel is now updated to 2.6.25-rc2-git6. I'll be bumping that to 2.6.25-rc3 later today. Xen and RT have been disabled due to severe conflicts that I'd prefer to leave those teams to solve. Testing is appreciated, but I'd prefer not to check this into Factory until Xen and RT are resynched. Caveats: 2.6.25-rc removed ->read_inode, which squashfs used. I did a quick merge, but haven't tested it yet. New "tainted fields were added, so anything that translates the bitfield will misreport things. - -Jeff - -- Jeff Mahoney SUSE Labs -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.4-svn0 (GNU/Linux) Comment: Using GnuPG with SUSE - http://enigmail.mozdev.org iD8DBQFHwyc8LPWxlyuTD7IRAsZCAJ46GcvaBz1lyIPm5RzASji8Z4AyzACeOOIo ZLrigRbIyi0fYJWrJmcyVVY= =mgnw -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe(a)opensuse.org For additional commands, e-mail: opensuse-kernel+help(a)opensuse.org

2 1

[opensuse-kernel] Re: HEAD update to 2.6.25-rc2-git6
by Jeff Mahoney 25 Feb '08

25 Feb '08

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Sven-Thorsten Dietrich wrote: > On Mon, 2008-02-25 at 15:38 -0500, Jeff Mahoney wrote: > > Hi all - > > Just a quick FYI. > > As some of you may have already noticed, the HEAD kernel is now updated > to 2.6.25-rc2-git6. I'll be bumping that to 2.6.25-rc3 later today. Xen > and RT have been disabled due to severe conflicts that I'd prefer to > leave those teams to solve. > > Testing is appreciated, but I'd prefer not to check this into Factory > until Xen and RT are resynched. > >> I have started the forward-port of RT, I'm probably 25% done with >> porting to -rc1 !! with 6 hours invested - so its going slow. > >> A lot of the tedium is from the ongoing merge of *_[32|64].[ch] files >> into common files, other issues are from recent HRT changes, etc. > >> I hope to be done end of this week, given some other red things that >> need to be done. Yeah, I hear you. I had hoped the regular merge would be a pretty quick thing, and it ended up taking about a day and a half, including all the build failure shakeout messages following the initial commit. Good times. :) - -Jeff - -- Jeff Mahoney SUSE Labs -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.4-svn0 (GNU/Linux) Comment: Using GnuPG with SUSE - http://enigmail.mozdev.org iD8DBQFHwzPELPWxlyuTD7IRAgcyAKCG96Q18+adlU2xtco00GfaMnJhdgCeNb5c iZHnUH+ogJh5ejvCgCal9Sc= =2wFV -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe(a)opensuse.org For additional commands, e-mail: opensuse-kernel+help(a)opensuse.org

1 0

[opensuse-kernel] exec-shield
by richard -rw- weinberger 22 Feb '08

22 Feb '08

Hi, Are there any plans to include exec-shield into the OpenSuSE kernel? I would like to have this feature. :-) Thanks, //richard -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe(a)opensuse.org For additional commands, e-mail: opensuse-kernel+help(a)opensuse.org

4 4

[opensuse-kernel] Re: [opensuse-factory] Version of kernel in 11.0
by Andreas Jaeger 15 Feb '08

15 Feb '08

Hi Kernel hacker, Rafał asks the following on the factory list and asked me to forward this to you. What are your plans moving forward? "Rafał Miłecki" <zajec5polish(a)gmail.com> writes: > Hi, > > I wonder if there is enought time for switching to 2.6.25 kernel in > factory? Some stats: > 9 weeks 1 day between 2.6.21-rc1 and 2.6.21 (stable) > 8 weeks between 2.6.22-rc1 and 2.6.22 > 11 weeks 2 days between 2.6.23-rc1 and 2.6.23 > 13 weeks 1 day between 2.6.24-rc1 and 2.6.24 > This makes me quite sure we can expect 2.6.25 last release candidate > (or even stable) exactly for Beta 1. I think it's perfect time as a > lot of ppl start to test openSUSE development versions in this stage. > We could get feedback about any problems just in time before bigger > changes (one update, or none). > > It would be nice to provide 2.6.25 because of interesting changes: > 1) http://lkml.org/lkml/2008/2/10/327 > 2) improvements in wireless b43 driver (Broadcom 4311 rev 02 support) > 3) ath5k driver (fully open and free replacement for madwifi) > which make me want to see this kernel in 11.0. I help ppl on polish > board and belive me, using madwifi is quite common problem. > > So is there any decision about kernel for 11.0? May we expect .25? Andreas -- Andreas Jaeger, Director Platform/openSUSE, aj(a)suse.de SUSE LINUX Products GmbH, GF: Markus Rex, HRB 16746 (AG Nürnberg) Maxfeldstr. 5, 90409 Nürnberg, Germany GPG fingerprint = 93A3 365E CE47 B889 DF7F FED1 389A 563C C272 A126

2 1

[opensuse-kernel] conflict between kernel-flavor and kernel-syms
by Hans-Peter Jansen 08 Feb '08

08 Feb '08

Hi, I typically manage my setups with yum, where I'm still on openSUSE 10.2, but one thing is really disturbing: Transaction Check Error: file /lib/modules/2.6.18.8-0.8-bigsmp/modules.alias conflicts between attempted installs of kernel-bigsmp-2.6.18.8-0.8 and kernel-syms-2.6.18.8-0.8 which isn't a big surprise, since: > rpm -qvlp i586/kernel-bigsmp-2.6.18.8-0.8.i586.rpm | grep modules.alias -rw-r--r-- 1 root root 307579 Jan 21 17:36 /lib/modules/2.6.18.8-0.8-bigsmp/modules.alias > rpm -qvlp i586/kernel-syms-2.6.18.8-0.8.i586.rpm | grep modules.alias -rw-r--r-- 1 root root 307579 Jan 21 17:52 /lib/modules/2.6.18.8-0.8-bigsmp/modules.alias -rw-r--r-- 1 root root 308322 Jan 21 17:52 /lib/modules/2.6.18.8-0.8-debug/modules.alias -rw-r--r-- 1 root root 306176 Jan 21 17:52 /lib/modules/2.6.18.8-0.8-default/modules.alias -rw-r--r-- 1 root root 306441 Jan 21 17:53 /lib/modules/2.6.18.8-0.8-kdump/modules.alias -rw-r--r-- 1 root root 4086 Jan 21 17:53 /lib/modules/2.6.18.8-0.8-um/modules.alias -rw-r--r-- 1 root root 296385 Jan 21 17:53 /lib/modules/2.6.18.8-0.8-vanilla/modules.alias -rw-r--r-- 1 root root 290340 Jan 21 17:53 /lib/modules/2.6.18.8-0.8-xen/modules.alias -rw-r--r-- 1 root root 290259 Jan 21 17:53 /lib/modules/2.6.18.8-0.8-xenpae/modules.alias Aren't these packages supposed to be installed together? Pete -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe(a)opensuse.org For additional commands, e-mail: opensuse-kernel+help(a)opensuse.org

1 0

[opensuse-kernel] Re: opensuse 11.0 kernel patch justification needed from you
by Greg KH 08 Feb '08

08 Feb '08

On Wed, Feb 06, 2008 at 01:55:00PM +0100, Kurt Garloff wrote: > Greg, > > On Tue, Feb 05, 2008 at 02:10:07PM -0800, Greg Kroah-Hartman wrote: > > Kurt Garloff: > > Why isn't this one upstream? It's been in our tree for ages, which > > tends to make me think it's no longer needed for us? > > patches.suse/scsi-error-test-unit-ready-timeout > > I sent it once (a looooong time ago), probably should send it again. > It did fix a real problem. That would be great. Or do you want me to push it with the others in this series I'm going to try to get there? > > We should be able to drop this now: > > patches.suse/setuid-dumpable-wrongdir > > Upstream got it wrong and set a wrong de-facto standard :-( > > As we had it right in SLES9, this patch keeps us compatible > with SLES9. We're also compatible with mainline: The file > exists twice :-( That's a PITA as well, but the smallest of > all evils I believe. Ok, we're stuck with this for forever then? Ugh, no fun... > > No more maintainer, or no valid "Signed-off-by:" from a Novell developer: > > I'm going to drop these, unless someone really feels that they should > > be in our kernel tree. If so, please let me know by the end of the > > week. Maybe I should just throw them at lkml, as we are using them > > for some reason. If they don't stick, they should probably be > > gone... > > patches.fixes/bridge-module-get-put.patch > > Jan Beulich owns it > > > patches.fixes/do_anonymous_page-race > > > > patches.fixes/grab-swap-token-oops > > Hmmm, if this is real, then it should go upstream/ > > > patches.fixes/ieee1394-sbp2_long_sysfs_ieee1394_id.patch > > patches.fixes/ipv6-no-autoconf > > patches.fixes/oom-warning > > This saved real support load. > > > patches.fixes/parport-mutex > > patches.fixes/remount-no-shrink-dcache > > Should go upstream IMHO > > > patches.fixes/seccomp-disable-tsc-option > > patches.fixes/tiocgdev > > Our boot logging daemon (blogd) needs it and this SUSE patch exists > since the last century or so (!). Werner submitted it numerous times > but some people were arguing that you could achieve similar things > without the patch. I think Werner has not treid for a few years now ... > Please talk to Werner Fink. Will do. > > patches.fixes/tulip-quad-NIC-ifdown > > patches.suse/connector-read-mostly > > patches.suse/crasher-26.diff > > patches.suse/filp-slab-rcu > > patches.suse/lockd-max-hosts-dynamic (and other lockd stuff, > > should all be dropped I > > feel, as we don't have > > anyone who cares about > > it anymore here...) > > I would suggest that Neil Brown reviews them and decides on dropping vs > submission. > > > patches.suse/netfilter-ipt_LOG-mac > > patches.suse/sysctl-add-affinity_load_balancing > > Could be a real significant performance tweak. > I suggest Andi looks at it. > > > patches.suse/twofish-2.6 (We gotta be able to drop this one > > now, right? If not, why isn't it > > upstream?) > > patches.suse/unmap_vmas-lat > > As we have a real RT kernel, we can drop this one. I'll do that. thanks, greg k-h -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe(a)opensuse.org For additional commands, e-mail: opensuse-kernel+help(a)opensuse.org

1 0

[opensuse-kernel] Re: opensuse 11.0 kernel patch justification needed from you
by Greg KH 08 Feb '08

08 Feb '08

On Wed, Feb 06, 2008 at 08:34:52AM +0900, Tejun Heo wrote: > Greg KH wrote: > > Are these upstream for 2.6.25?: > > patches.drivers/libata-add-waits-for-govault > > This one we need to keep for now. > > > patches.drivers/libata-force-cable-type > > This should be replaced by libata-implement-force-parameter which is > acked. Dunno whether it will make 25 or 26. Will update HEAD. (it's > still HEAD right?) Yes, it's still HEAD :) > > patches.drivers/libata-sata_nv-disable-ADMA > > This one we need to keep for now. ADMA still has quite a few issues and > I don't see them being resolved soon. Upstream might disable ADMA by > default for 2.6.25 if solutions don't come up soon. > > > patches.drivers/libata-unlock-hpa-by-default > > Upstream will never change this behavior. We'll just have to keep this > around. > > > patches.drivers/scsi-throttle-SG_DXFER_TO_FROM_DEV-warning-better > > Please keep this for now. I'll forward it upstream. Thanks for the descriptions, greg k-h -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe(a)opensuse.org For additional commands, e-mail: opensuse-kernel+help(a)opensuse.org

1 0