Re: Rawhide nodebug and the 3.12 kernel
Le Mer 6 novembre 2013 21:39, Bruno Wolff III a écrit : On Wed, Nov 06, 2013 at 12:30:37 -0800, Adam Williamson awill...@redhat.com wrote: FWIW the ship has probably sailed now, but I really don't think it'd be much of a problem to have 3.12 in F20 at release time. It's what I've been running on my F20 box here for the last several weeks anyway, and based on my testing it's unlikely to cause us any particular problems. I asked about this last week and the kernel devs didn't feel comfortable about switching after beta or trying to get a freeze exception to get 3.12 into beta. I run rawhide nodebug kernels on three machines and am not seeing any regressions relative to 3.11. I'm seeing system hangs with rawhide kernels now (oops on boot with the latest one). No idea if it's a kernel hang or just dead input, and journalctl's lack of handling of logs that were interupted by a reset is not helping -- Nicolas Mailhot -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Re: Rawhide nodebug and the 3.12 kernel
On Wed, Nov 06, 2013 at 08:00:33 -0600, Justin M. Forbes jmfor...@linuxtx.org wrote: We have a slight issue with the 3.12 kernel timing in that it is too late to push it into Fedora 20, but too far away from the Fedora 20 release to just ignore the 3.13 development cycle until we can push 3.12. As a result, we will be tracking 3.12 and stable updates for it in the rawhide-nodebug repository. This gives us a chance to keep it built and tested on all primary architectures, and make sure we are in good shape to push 3.12 out as an update as soon as possible. Once 3.12 can be pushed to releases, the rawhide-nodebug repository will return to doing non debug builds of rawhide, tracking Linus' tree upstream. I will let everyone know that is happening through the same channels with a couple of days notice. This should work out better than it normally would, since there is an extra week for the 3.13 merge window. -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Re: Rawhide nodebug and the 3.12 kernel
On Wed, 2013-11-06 at 08:00 -0600, Justin M. Forbes wrote: We have a slight issue with the 3.12 kernel timing in that it is too late to push it into Fedora 20, but too far away from the Fedora 20 release to just ignore the 3.13 development cycle until we can push 3.12. As a result, we will be tracking 3.12 and stable updates for it in the rawhide-nodebug repository. This gives us a chance to keep it built and tested on all primary architectures, and make sure we are in good shape to push 3.12 out as an update as soon as possible. Once 3.12 can be pushed to releases, the rawhide-nodebug repository will return to doing non debug builds of rawhide, tracking Linus' tree upstream. I will let everyone know that is happening through the same channels with a couple of days notice. More information on the rawhide-nodebug repository can be found at: http://fedoraproject.org/wiki/RawhideKernelNodebug FWIW the ship has probably sailed now, but I really don't think it'd be much of a problem to have 3.12 in F20 at release time. It's what I've been running on my F20 box here for the last several weeks anyway, and based on my testing it's unlikely to cause us any particular problems. -- Adam Williamson Fedora QA Community Monkey IRC: adamw | Twitter: AdamW_Fedora | XMPP: adamw AT happyassassin . net http://www.happyassassin.net -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Re: Rawhide nodebug and the 3.12 kernel
On Wed, Nov 6, 2013 at 3:30 PM, Adam Williamson awill...@redhat.com wrote: On Wed, 2013-11-06 at 08:00 -0600, Justin M. Forbes wrote: We have a slight issue with the 3.12 kernel timing in that it is too late to push it into Fedora 20, but too far away from the Fedora 20 release to just ignore the 3.13 development cycle until we can push 3.12. As a result, we will be tracking 3.12 and stable updates for it in the rawhide-nodebug repository. This gives us a chance to keep it built and tested on all primary architectures, and make sure we are in good shape to push 3.12 out as an update as soon as possible. Once 3.12 can be pushed to releases, the rawhide-nodebug repository will return to doing non debug builds of rawhide, tracking Linus' tree upstream. I will let everyone know that is happening through the same channels with a couple of days notice. More information on the rawhide-nodebug repository can be found at: http://fedoraproject.org/wiki/RawhideKernelNodebug FWIW the ship has probably sailed now, but I really don't think it'd be much of a problem to have 3.12 in F20 at release time. It's what I've been running on my F20 box here for the last several weeks anyway, and based on my testing it's unlikely to cause us any particular problems. I literally just screamed. The past 3 releases we've been pointedly reminded by QA that when doing fixes they should be scoped to as small as possible. So we've been trying really hard to do that during the Alpha and the Beta freezes. And now you want to ship a Beta (which gets the widest testing feedback of the pre-releases) with 3.11, and shove 3.12 in after that as the F20 release kernel? It could be the massive amount of email and meetings today addling my brain, but can you explain how that makes sense? josh -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Re: Rawhide nodebug and the 3.12 kernel
On Wed, Nov 06, 2013 at 12:30:37 -0800, Adam Williamson awill...@redhat.com wrote: FWIW the ship has probably sailed now, but I really don't think it'd be much of a problem to have 3.12 in F20 at release time. It's what I've been running on my F20 box here for the last several weeks anyway, and based on my testing it's unlikely to cause us any particular problems. I asked about this last week and the kernel devs didn't feel comfortable about switching after beta or trying to get a freeze exception to get 3.12 into beta. I run rawhide nodebug kernels on three machines and am not seeing any regressions relative to 3.11. -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Re: Rawhide nodebug and the 3.12 kernel
Am 06.11.2013 21:37, schrieb Josh Boyer: On Wed, Nov 6, 2013 at 3:30 PM, Adam Williamson awill...@redhat.com wrote: FWIW the ship has probably sailed now, but I really don't think it'd be much of a problem to have 3.12 in F20 at release time. It's what I've been running on my F20 box here for the last several weeks anyway, and based on my testing it's unlikely to cause us any particular problems. I literally just screamed. The past 3 releases we've been pointedly reminded by QA that when doing fixes they should be scoped to as small as possible. So we've been trying really hard to do that during the Alpha and the Beta freezes. And now you want to ship a Beta (which gets the widest testing feedback of the pre-releases) with 3.11, and shove 3.12 in after that as the F20 release kernel? It could be the massive amount of email and meetings today addling my brain, but can you explain how that makes sense? may i also bring in this discussion that the images are not rebuilt in the worst case thi smeans live-CD/DVD images are unusable on recent hardware a real-world case for me was F15 - unusable on SandyBridge machines by horrible frrezing desktop all day long, these day snot that critical because we got new kernels after release, but they does not change live-media after it signature.asc Description: OpenPGP digital signature -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Re: Rawhide nodebug and the 3.12 kernel
On Wed, Nov 6, 2013 at 3:39 PM, Bruno Wolff III br...@wolff.to wrote: On Wed, Nov 06, 2013 at 12:30:37 -0800, Adam Williamson awill...@redhat.com wrote: FWIW the ship has probably sailed now, but I really don't think it'd be much of a problem to have 3.12 in F20 at release time. It's what I've been running on my F20 box here for the last several weeks anyway, and based on my testing it's unlikely to cause us any particular problems. I asked about this last week and the kernel devs didn't feel comfortable about switching after beta or trying to get a freeze exception to get 3.12 into beta. Right. Also, at this point I'm not sure what 3.12 actually buys us over a known 3.11. I run rawhide nodebug kernels on three machines and am not seeing any regressions relative to 3.11. Are you running any ARM machines? My understanding is that our F20 kernel has patches that enable important ARM stuff that isn't in rawhide (3.12) because it was conflicting with the churn. So that would need to be added and tested, given ARM is primary on F20. josh -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Re: Rawhide nodebug and the 3.12 kernel
On Wed, Nov 06, 2013 at 15:42:56 -0500, Josh Boyer jwbo...@fedoraproject.org wrote: Are you running any ARM machines? My understanding is that our F20 kernel has patches that enable important ARM stuff that isn't in rawhide (3.12) because it was conflicting with the churn. So that would need to be added and tested, given ARM is primary on F20. Not with 3.12. The only ARM machine I have is an XO and my understanding is that Fedora kernels don't work on that yet. I most have older stuff. I have one that is i686 Xeon, one that is i686 Athlon MP and one that is x86_64 on a night quite as Xeon machine. -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Re: Rawhide nodebug and the 3.12 kernel
On Wed, Nov 6, 2013 at 8:42 PM, Josh Boyer jwbo...@fedoraproject.org wrote: On Wed, Nov 6, 2013 at 3:39 PM, Bruno Wolff III br...@wolff.to wrote: On Wed, Nov 06, 2013 at 12:30:37 -0800, Adam Williamson awill...@redhat.com wrote: FWIW the ship has probably sailed now, but I really don't think it'd be much of a problem to have 3.12 in F20 at release time. It's what I've been running on my F20 box here for the last several weeks anyway, and based on my testing it's unlikely to cause us any particular problems. I asked about this last week and the kernel devs didn't feel comfortable about switching after beta or trying to get a freeze exception to get 3.12 into beta. Right. Also, at this point I'm not sure what 3.12 actually buys us over a known 3.11. I run rawhide nodebug kernels on three machines and am not seeing any regressions relative to 3.11. Are you running any ARM machines? My understanding is that our F20 kernel has patches that enable important ARM stuff that isn't in rawhide (3.12) because it was conflicting with the churn. So that would need to be added and tested, given ARM is primary on F20. The main issue here is BBBlack rebasing which I've started to do (not sure if kyle has looked at this at all) but as it stands the BBB on 3.11 has issues with USB/Display anyway so from this PoV it shouldn't be hard to get us to as good as or better experience with 3.12 on the BBBlack. The advantage of moving to 3.12 from the ARM PoV would be a much better and expanded experience from the i.MX PoV which covers Utilite and Wandboard which are some of our best working devices at the moment and it would improve that experience greatly. From my PoV I'm certainly not opposed to moving to 3.12 from the ARM side of things. Peter -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Re: Rawhide nodebug and the 3.12 kernel
Are you running any ARM machines? My understanding is that our F20 kernel has patches that enable important ARM stuff that isn't in rawhide (3.12) because it was conflicting with the churn. So that would need to be added and tested, given ARM is primary on F20. Not with 3.12. The only ARM machine I have is an XO and my understanding is that Fedora kernels don't work on that yet. Nope, there's a lot of Marvell stuff for their SoC not upstream yet. Peter -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Re: Rawhide nodebug and the 3.12 kernel
On Wed, Nov 6, 2013 at 3:55 PM, Peter Robinson pbrobin...@gmail.com wrote: On Wed, Nov 6, 2013 at 8:42 PM, Josh Boyer jwbo...@fedoraproject.org wrote: On Wed, Nov 6, 2013 at 3:39 PM, Bruno Wolff III br...@wolff.to wrote: On Wed, Nov 06, 2013 at 12:30:37 -0800, Adam Williamson awill...@redhat.com wrote: FWIW the ship has probably sailed now, but I really don't think it'd be much of a problem to have 3.12 in F20 at release time. It's what I've been running on my F20 box here for the last several weeks anyway, and based on my testing it's unlikely to cause us any particular problems. I asked about this last week and the kernel devs didn't feel comfortable about switching after beta or trying to get a freeze exception to get 3.12 into beta. Right. Also, at this point I'm not sure what 3.12 actually buys us over a known 3.11. I run rawhide nodebug kernels on three machines and am not seeing any regressions relative to 3.11. Are you running any ARM machines? My understanding is that our F20 kernel has patches that enable important ARM stuff that isn't in rawhide (3.12) because it was conflicting with the churn. So that would need to be added and tested, given ARM is primary on F20. The main issue here is BBBlack rebasing which I've started to do (not sure if kyle has looked at this at all) but as it stands the BBB on 3.11 has issues with USB/Display anyway so from this PoV it shouldn't be hard to get us to as good as or better experience with 3.12 on the BBBlack. The advantage of moving to 3.12 from the ARM PoV would be a much better and expanded experience from the i.MX PoV which covers Utilite and Wandboard which are some of our best working devices at the moment and it would improve that experience greatly. Has there been a set of devices decided as blocking for F20? Are any of those you mentioned in that set? I don't want to go add more devices people have to test at this point. From my PoV I'm certainly not opposed to moving to 3.12 from the ARM side of things. That's good to know, but I'm still uncomfortable doing it post-Beta. I don't see how that would be reasonable, but I'm willing to be enlightened. josh -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Re: Rawhide nodebug and the 3.12 kernel
Are you running any ARM machines? My understanding is that our F20 kernel has patches that enable important ARM stuff that isn't in rawhide (3.12) because it was conflicting with the churn. So that would need to be added and tested, given ARM is primary on F20. The main issue here is BBBlack rebasing which I've started to do (not sure if kyle has looked at this at all) but as it stands the BBB on 3.11 has issues with USB/Display anyway so from this PoV it shouldn't be hard to get us to as good as or better experience with 3.12 on the BBBlack. The advantage of moving to 3.12 from the ARM PoV would be a much better and expanded experience from the i.MX PoV which covers Utilite and Wandboard which are some of our best working devices at the moment and it would improve that experience greatly. Has there been a set of devices decided as blocking for F20? Are any of those you mentioned in that set? I don't want to go add more devices people have to test at this point. I'm not sure what, if any, specific are classed as blocking from the ARM PoV. From the devices we're actively supporting the Wandboard will have a better experience and the BBB I believe will as well but I'll know that in the next couple of days. From my PoV I'm certainly not opposed to moving to 3.12 from the ARM side of things. That's good to know, but I'm still uncomfortable doing it post-Beta. I don't see how that would be reasonable, but I'm willing to be enlightened. I don't see it as unreasonable, I'm just updating the current status on ARM for 3.12 so you and others don't see us as blocking. Peter -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Re: Rawhide nodebug and the 3.12 kernel
On Wed, 2013-11-06 at 15:37 -0500, Josh Boyer wrote: On Wed, Nov 6, 2013 at 3:30 PM, Adam Williamson awill...@redhat.com wrote: On Wed, 2013-11-06 at 08:00 -0600, Justin M. Forbes wrote: We have a slight issue with the 3.12 kernel timing in that it is too late to push it into Fedora 20, but too far away from the Fedora 20 release to just ignore the 3.13 development cycle until we can push 3.12. As a result, we will be tracking 3.12 and stable updates for it in the rawhide-nodebug repository. This gives us a chance to keep it built and tested on all primary architectures, and make sure we are in good shape to push 3.12 out as an update as soon as possible. Once 3.12 can be pushed to releases, the rawhide-nodebug repository will return to doing non debug builds of rawhide, tracking Linus' tree upstream. I will let everyone know that is happening through the same channels with a couple of days notice. More information on the rawhide-nodebug repository can be found at: http://fedoraproject.org/wiki/RawhideKernelNodebug FWIW the ship has probably sailed now, but I really don't think it'd be much of a problem to have 3.12 in F20 at release time. It's what I've been running on my F20 box here for the last several weeks anyway, and based on my testing it's unlikely to cause us any particular problems. I literally just screamed. The past 3 releases we've been pointedly reminded by QA that when doing fixes they should be scoped to as small as possible. So we've been trying really hard to do that during the Alpha and the Beta freezes. And now you want to ship a Beta (which gets the widest testing feedback of the pre-releases) with 3.11, and shove 3.12 in after that as the F20 release kernel? It could be the massive amount of email and meetings today addling my brain, but can you explain how that makes sense? Sorry, on re-reading that it could've been clearer =) By 'the ship has probably sailed now' I meant 'beta's almost done'. I was trying to say we could probably have safely got it in before Beta (though, admittedly, we didn't know we were going to slip two weeks...) I wasn't really suggesting any change, just noting that 3.12's actually fine for F20 atm if anyone feels like using it. The other funny thing is that the kernel is actually a relatively reliable component, even though it's so vital, because a) it's usually pretty damn obvious if anything terrible is wrong and b) it's extensively and competently tested upstream. I'd actually be _less_ concerned about changing the kernel post-beta than changing, say, I dunno, NetworkManager (not to pick on NM, just an example, many things are in the same boat). Still, I'm entirely fine with sticking with 3.11 and happy the kernel team is considering stability/quality in making that choice. -- Adam Williamson Fedora QA Community Monkey IRC: adamw | Twitter: AdamW_Fedora | XMPP: adamw AT happyassassin . net http://www.happyassassin.net -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Re: Rawhide nodebug and the 3.12 kernel
On Wed, 2013-11-06 at 15:42 -0500, Josh Boyer wrote: On Wed, Nov 6, 2013 at 3:39 PM, Bruno Wolff III br...@wolff.to wrote: On Wed, Nov 06, 2013 at 12:30:37 -0800, Adam Williamson awill...@redhat.com wrote: FWIW the ship has probably sailed now, but I really don't think it'd be much of a problem to have 3.12 in F20 at release time. It's what I've been running on my F20 box here for the last several weeks anyway, and based on my testing it's unlikely to cause us any particular problems. I asked about this last week and the kernel devs didn't feel comfortable about switching after beta or trying to get a freeze exception to get 3.12 into beta. Right. Also, at this point I'm not sure what 3.12 actually buys us over a known 3.11. It makes the sound on my laptop work. THAT SHOULD BE REASON ENOUGH! ;) -- Adam Williamson Fedora QA Community Monkey IRC: adamw | Twitter: AdamW_Fedora | XMPP: adamw AT happyassassin . net http://www.happyassassin.net -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Re: Rawhide nodebug and the 3.12 kernel
On Wed, Nov 6, 2013 at 4:16 PM, Adam Williamson awill...@redhat.com wrote: On Wed, 2013-11-06 at 15:37 -0500, Josh Boyer wrote: On Wed, Nov 6, 2013 at 3:30 PM, Adam Williamson awill...@redhat.com wrote: On Wed, 2013-11-06 at 08:00 -0600, Justin M. Forbes wrote: We have a slight issue with the 3.12 kernel timing in that it is too late to push it into Fedora 20, but too far away from the Fedora 20 release to just ignore the 3.13 development cycle until we can push 3.12. As a result, we will be tracking 3.12 and stable updates for it in the rawhide-nodebug repository. This gives us a chance to keep it built and tested on all primary architectures, and make sure we are in good shape to push 3.12 out as an update as soon as possible. Once 3.12 can be pushed to releases, the rawhide-nodebug repository will return to doing non debug builds of rawhide, tracking Linus' tree upstream. I will let everyone know that is happening through the same channels with a couple of days notice. More information on the rawhide-nodebug repository can be found at: http://fedoraproject.org/wiki/RawhideKernelNodebug FWIW the ship has probably sailed now, but I really don't think it'd be much of a problem to have 3.12 in F20 at release time. It's what I've been running on my F20 box here for the last several weeks anyway, and based on my testing it's unlikely to cause us any particular problems. I literally just screamed. The past 3 releases we've been pointedly reminded by QA that when doing fixes they should be scoped to as small as possible. So we've been trying really hard to do that during the Alpha and the Beta freezes. And now you want to ship a Beta (which gets the widest testing feedback of the pre-releases) with 3.11, and shove 3.12 in after that as the F20 release kernel? It could be the massive amount of email and meetings today addling my brain, but can you explain how that makes sense? Sorry, on re-reading that it could've been clearer =) By 'the ship has probably sailed now' I meant 'beta's almost done'. I was trying to say we could probably have safely got it in before Beta (though, admittedly, we didn't know we were going to slip two weeks...) I wasn't really suggesting any change, just noting that 3.12's actually fine for F20 atm if anyone feels like using it. I wasn't released upstream until this past Sunday. We could have gone with a late RC, but didn't think it was prudent. The other funny thing is that the kernel is actually a relatively reliable component, even though it's so vital, because a) it's usually pretty damn obvious if anything terrible is wrong and b) it's Well, depending on the machine. That's the coverage part I'm worried about. Stuff like backlights, etc. extensively and competently tested upstream. I'd actually be _less_ concerned about changing the kernel post-beta than changing, say, I dunno, NetworkManager (not to pick on NM, just an example, many things are in the same boat). Still, I'm entirely fine with sticking with 3.11 and happy the kernel team is considering stability/quality in making that choice. OK, great. Much less confused, thanks. josh -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Re: Rawhide nodebug and the 3.12 kernel
On Nov 6, 2013, at 1:37 PM, Josh Boyer jwbo...@fedoraproject.org wrote: On Wed, Nov 6, 2013 at 3:30 PM, Adam Williamson awill...@redhat.com wrote: On Wed, 2013-11-06 at 08:00 -0600, Justin M. Forbes wrote: We have a slight issue with the 3.12 kernel timing in that it is too late to push it into Fedora 20, but too far away from the Fedora 20 release to just ignore the 3.13 development cycle until we can push 3.12. As a result, we will be tracking 3.12 and stable updates for it in the rawhide-nodebug repository. This gives us a chance to keep it built and tested on all primary architectures, and make sure we are in good shape to push 3.12 out as an update as soon as possible. Once 3.12 can be pushed to releases, the rawhide-nodebug repository will return to doing non debug builds of rawhide, tracking Linus' tree upstream. I will let everyone know that is happening through the same channels with a couple of days notice. More information on the rawhide-nodebug repository can be found at: http://fedoraproject.org/wiki/RawhideKernelNodebug FWIW the ship has probably sailed now, but I really don't think it'd be much of a problem to have 3.12 in F20 at release time. It's what I've been running on my F20 box here for the last several weeks anyway, and based on my testing it's unlikely to cause us any particular problems. I literally just screamed. The past 3 releases we've been pointedly reminded by QA that when doing fixes they should be scoped to as small as possible. So we've been trying really hard to do that during the Alpha and the Beta freezes. And now you want to ship a Beta (which gets the widest testing feedback of the pre-releases) with 3.11, and shove 3.12 in after that as the F20 release kernel? Whatever major version ships with beta should ship with final. There are some significant changes in 3.12. I have no problem with final being 3.11.7 or .8 or whatever. Otherwise WTF is the point of all the selective backport efforts, and avoidance of major 3.12 changes possibly causing regressions? Chris Murphy -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
