Am Fri, 12 Nov 2021 17:37:07 +0100 schrieb Frederic Weisbecker <frederic@xxxxxxxxxx>: > On Fri, Nov 12, 2021 at 04:36:56PM +0100, Sebastian Andrzej Siewior > wrote: > > On 2021-11-04 17:29:08 [+0000], Moessbauer, Felix wrote: > > > Dear subscribers, > > Hi, > > > > I Cced cgroups@vger since thus question fits there better. > > I Cced Frederic in case he has come clues regarding isolcpus and > > cgroups. > > > > > we are currently evaluating how to rework realtime tuning to use > > > cgroup-v2 cpusets instead of the isolcpus kernel parameter. Our > > > use-case are realtime applications with rt and non-rt threads. > > > Hereby, the non-rt thread might create additional non-rt threads: > > > > > > Example (RT CPU=1, 4 CPUs): > > > - Non-RT Thread (A) with default affinity 0xD (1101b) > > > - RT Thread (B) with Affinity 0x2 (0010b, via set_affinity) > > > > > > When using pure isolcpus and cgroup-v1, just setting isolcpus=1 > > > perfectly works: Thread A gets affinity 0xD, Thread B gets 0x2 > > > and additional threads get a default affinity of 0xD. By that, > > > independent of the threads' priorities, we can ensure that > > > nothing is scheduled on our RT cpu (except from kernel threads, > > > etc...). > > > > > > During this journey, we discovered the following: > > > > > > Using cgroup-v2 cpusets and isolcpus together seems to be > > > incompatible: When activating the cpuset controller on a cgroup > > > (for the first time), all default CPU affinities are reset. By > > > that, also the default affinity is set to 0xFFFF..., while with > > > isolcpus we expect it to be (0xFFFF - isolcpus). This breaks the > > > example from above, as now the non-RT thread can also be > > > scheduled on the RT CPU. > > That sounds buggy from the cpuset-v2 side (adding the maintainers in > Cc). We thought the same but ended up saying that it might just be related to the deprecation. But regarding that as a bug could help fix it in stable kernels and distros. Because the patches will need some time and will likely not get backported. Henning > Also please have a look into "[PATCH v8 0/6] cgroup/cpuset: Add new > cpuset partition type & empty effecitve cpus": > > https://lore.kernel.org/lkml/20211018143619.205065-1-longman@xxxxxxxxxx/ > > This stuff adds support for a new "isolated" partition type on > cpuset/cgroup-v2 which should behave just like isolcpus. > > > > > > > When only using cgroup-v2, we can isolate our RT process by > > > placing it in a cgroup with CPUs=0,1 and remove CPU=1 from all > > > other cgroups. However, we do not know of a strategy to set a > > > default affinity: Given the example above, we have no way to > > > ensure that newly created threads are born with an affinity of > > > just 0x2 (without changing the application). > > > > > > Finally, isolcpus itself is deprecated since kernel 5.4. > > > > Where is this the deprecation of isolcpus announced/ written? > > We tried to deprecate it but too many people are still using it. > Better pick an interface that allows you to change the isolated set > at runtime like cpuset.sched_load_balance on cpuset/cgroup-v1 or the > above patchset on v2. > > Thanks.
