Hi All,
A brief update on where things stand in regards to the Fedora kernel.
Greg KH recently announced that the 3.14.y stable kernel will be a longterm kernel release. Given that F19 is evidently very lightly used at this point, we're going to remain with 3.14.y until it goes EOL. At the moment, 3.14.8 is in stable updates with 3.14.11 in updates-testing. 3.14.12 is built in koji. Testing and karma on the updates-testing releases would be appreciated.
F20 is currently at 3.15.4 with 3.15.5 being pushed likely later today. There have been a few regressions with intel video but hopefully those are starting to clear up. F20 will stay on 3.15.y likely until 3.16.1 or 3.16.2 is released.
F21 branched from rawhide last week, but we're keeping the two branches in sync for now. Once 3.16 final is released, we'll evaluate where we are in both the Fedora and upstream kernel release cycles and see if 3.17 is a good candidate for F21. At the moment, we're planning on sticking with 3.16.
Rawhide will continue to churn through upstream releases as usual. It is at 3.16-rc5 today. As side note, I never enabled slub debugging for the 3.16-rcX builds. Nobody noticed. It is possible the people that typically testing Rawhide kernels are always using the NoDebug repo (or booting with slub_debug=-), but I find it interesting that nobody noticed a speedup in the merge window kernels.
Lastly, I recently wrote about the kernel-playground Copr. As of yesterday, it is on 3.16-rc5 and contains the latest kdbus and overlayfs v23. There have been a few questions about an F21 build, and I plan on doing that once F21 is available as a chroot in Copr.
As always, if you have any questions please email the list.
josh
On 15.07.2014 17:41, Josh Boyer wrote:
Hi All,
A brief update on where things stand in regards to the Fedora kernel.
Greg KH recently announced that the 3.14.y stable kernel will be a longterm kernel release. Given that F19 is evidently very lightly used at this point, we're going to remain with 3.14.y until it goes EOL. At the moment, 3.14.8 is in stable updates with 3.14.11 in updates-testing. 3.14.12 is built in koji. Testing and karma on the updates-testing releases would be appreciated.
F20 is currently at 3.15.4 with 3.15.5 being pushed likely later today. There have been a few regressions with intel video but hopefully those are starting to clear up. F20 will stay on 3.15.y likely until 3.16.1 or 3.16.2 is released.
F21 branched from rawhide last week, but we're keeping the two branches in sync for now. Once 3.16 final is released, we'll evaluate where we are in both the Fedora and upstream kernel release cycles and see if 3.17 is a good candidate for F21. At the moment, we're planning on sticking with 3.16.
Rawhide will continue to churn through upstream releases as usual. It is at 3.16-rc5 today. As side note, I never enabled slub debugging for the 3.16-rcX builds. Nobody noticed. It is possible the people that typically testing Rawhide kernels are always using the NoDebug repo (or booting with slub_debug=-), but I find it interesting that nobody noticed a speedup in the merge window kernels.
We noticed, and we know about it! :)
Lastly, I recently wrote about the kernel-playground Copr. As of yesterday, it is on 3.16-rc5 and contains the latest kdbus and overlayfs v23. There have been a few questions about an F21 build, and I plan on doing that once F21 is available as a chroot in Copr.
As always, if you have any questions please email the list.
josh
We appreciate your efforts,
test@lists.fedoraproject.org For testing and quality assurance of Fedora releases
On Tue, Jul 15, 2014 at 11:41:16 -0400, Josh Boyer jwboyer@fedoraproject.org wrote:
Rawhide will continue to churn through upstream releases as usual. It is at 3.16-rc5 today. As side note, I never enabled slub debugging for the 3.16-rcX builds. Nobody noticed. It is possible the people that typically testing Rawhide kernels are always using the NoDebug repo (or booting with slub_debug=-), but I find it interesting that nobody noticed a speedup in the merge window kernels.
I mostly test nodebug kernels, I use slub_debug=- and there were significant regressions on i686 (and there still is one affecting one of my machines). So I wasn't likely to notice changes this time around.
Lastly, I recently wrote about the kernel-playground Copr. As of yesterday, it is on 3.16-rc5 and contains the latest kdbus and overlayfs v23. There have been a few questions about an F21 build, and I plan on doing that once F21 is available as a chroot in Copr.
Can that repo be used for testing fixes that haven't been accepted upstream yet? (I didn't have any particular ones in mind, but it seemed like it might be another possible use for the repo.)
On Tue, Jul 15, 2014 at 1:13 PM, Bruno Wolff III bruno@wolff.to wrote:
On Tue, Jul 15, 2014 at 11:41:16 -0400, Josh Boyer jwboyer@fedoraproject.org wrote:
Lastly, I recently wrote about the kernel-playground Copr. As of yesterday, it is on 3.16-rc5 and contains the latest kdbus and overlayfs v23. There have been a few questions about an F21 build, and I plan on doing that once F21 is available as a chroot in Copr.
Can that repo be used for testing fixes that haven't been accepted upstream yet? (I didn't have any particular ones in mind, but it seemed like it might be another possible use for the repo.)
I guess it depends on the "fix"? Most fixes would be better served with just a normal scratch-build. If it's a new feature that's going to land as a fix to a long-standing issue, then maybe. I can't think of anything that would really fit that description though.
josh
kernel@lists.fedoraproject.org