On Mon, Oct 9, 2023 at 7:32 PM Michal Koutný <mkoutny@xxxxxxxx> wrote: > > Hello. > > On Sat, Oct 07, 2023 at 02:02:58PM +0000, Yafang Shao <laoar.shao@xxxxxxxxx> wrote: > > Two new helpers are added for cgroup1 hierarchy: > > > > - task_cgroup1_id_within_hierarchy > > Retrieves the associated cgroup ID of a task within a specific cgroup1 > > hierarchy. The cgroup1 hierarchy is identified by its hierarchy ID. > > - task_ancestor_cgroup1_id_within_hierarchy > > Retrieves the associated ancestor cgroup ID of a task whithin a > > specific cgroup1 hierarchy. The specific ancestor level is determined by > > its ancestor level. > > > > These helper functions have been added to facilitate the tracing of tasks > > within a particular container or cgroup in BPF programs. It's important to > > note that these helpers are designed specifically for cgroup1. > > Are this helpers need for any 3rd party task? Yes, for example, we can check if the *next* task in sched_switch belongs to a specific cgroup. Hao also pointed out the use case of a 3rd partry task[1]. [1]. https://lore.kernel.org/bpf/CA+khW7hah317_beZ7QDA1R=sWi5q7TanRC+efMhivPKtWzbA4w@xxxxxxxxxxxxxx/ > > I *think* operating on `current` would be simpler wrt assumptions needed > for object presense. > > > > +u64 task_cgroup1_id_within_hierarchy(struct task_struct *tsk, int hierarchy_id) > > +{ > > + struct cgroup_root *root; > > + struct cgroup *cgrp; > > + u64 cgid = 0; > > + > > + spin_lock_irq(&css_set_lock); > > + list_for_each_entry(root, &cgroup_roots, root_list) { > > This should be for_each_root() macro for better uniform style. will use it in the next version. -- Regards Yafang
