On Thu, Sep 4, 2025 at 7:32 AM Roman Gushchin <roman.gushchin@xxxxxxxxx> wrote: > > Tejun Heo <tj@xxxxxxxxxx> writes: > > > Hello, > > > > On Wed, Sep 03, 2025 at 04:30:16PM -0700, Roman Gushchin wrote: > > ... > >> > - I'm passing in cgroup_id as an optional field in struct_ops and then in > >> > enable path, look up the matching cgroup, verify it can attach there and > >> > insert and update data structures accordingly: > >> > > >> > https://git.kernel.org/pub/scm/linux/kernel/git/tj/sched_ext.git/tree/kernel/sched/ext.c?h=scx-hier-prototype#n5280 > >> > >> Yeah, we discussed this option with Martin up in this thread. It doesn't > >> look as the best possible solution, but maybe the best we have at the moment. > >> > >> Ideally, I want something like this: > >> > >> void test_oom(void) > >> { > >> struct test_oom *skel; > >> int err, cgroup_fd; > >> > >> cgroup_fd = open(...); > >> if (cgroup_fd < 0) > >> goto cleanup; > >> > >> skel = test_oom__open_and_load(); > >> if (!skel) > >> goto cleanup; > >> > >> err = test_oom__attach_cgroup(skel, cgroup_fd); > >> if (CHECK_FAIL(err)) > >> goto cleanup; > > > > Yeah, that'd look better but are there practical differences? The only one I > > can think of is fs based permission check but that can be done separately > > too. > > The practical difference is that a single struct ops can be attached > to multiple cgroups. +1 Attaching the same scheduler to multiple cgroups also sounds useful. I feel sched-ext should use cgroup_fd too and do scx_sub_enable() at attach time instead of load time. Then scx_sub_disable() can happen at link detach. Looks more flexible from user pov.
