On 2019-01-26 01:30:03 [+0000], Hindman, Gavin wrote: > > My understanding is that the dev tree is just that - some work in > progress so could be subject to transient latency and stability > regressions. Certainly, the dev tree generally works, but it's not > guaranteed to remain that way. The stable patch sets just get the > changes considered truly stable. I was waiting with my answer since I didn't know what the missing pieces of information might be. With your text above tl;dr answer is that I currently don't plan to add new rewrites to the dev RT tree. The long version is: v4.19 is currently in the dev-tree. It will be handed over to the stable team once v5.0-RT is released (since it is the v4.19+2 release). This is the basic workflow. Given how long v4.19-RT is out, it should "stable" / it should not have any known bugs / regressions. If so, I am not aware of it. If there is anything, please let me know. If you haven't started using v4.19-RT yet, please do so. The sooner problems are reported, the sooner people are aware of them and may work on them. At this point in time, I don't plan to throw any new patches into the dev tree which would lead to a rewrite / change of RT behaviour. How stable is stable. Around August I backported a handful patches for v4.9-RT because futex was broken in some corner cases. The last v4.19-RT release had a ptrace related fix which is or will be picked up by the RT stable team. The locking latency problem on x86 for v4.1+ was routed via the stable team (not the RT-stable team). So there is that. Hope that answers your question. Sebastian
