On Fri Sep 4, 2026 at 8:03 PM CEST, Andrew Morton wrote: > On Fri, 4 Sep 2026 18:20:19 +0800 "Hui Zhu" <[email protected]> wrote: > >> @@ -159,6 +160,74 @@ __bpf_kfunc void bpf_mem_cgroup_flush_stats(struct >> mem_cgroup *memcg) >> mem_cgroup_flush_stats(memcg); >> } >> >> +/* >> + * Reclaim must not recurse: try_to_free_mem_cgroup_pages() overwrites >> + * current->reclaim_state, so a nested call would corrupt the outer >> + * reclaim state. Reclaim windows are marked with PF_MEMALLOC; >> + * reclaim_state is also checked because it is installed slightly >> + * before PF_MEMALLOC. >> + */ >> +static bool bpf_in_reclaim_context(void) >> +{ >> + return (current->flags & PF_MEMALLOC) || current->reclaim_state; >> +} >> + > > Would life improve if try_to_free_mem_cgroup_pages() didn't do that? > If try_to_free_mem_cgroup_pages() (or some variant of it) were to > permit nesting? > > rs = new_thing(current, &sc.reclaim_state); > ... > set_task_reclaim_state(current, rs); >
My understanding is that given the current restrictions on where this kfunc can be called, this is mostly a defensive and shouldn't be firing (?). Given that, we can probably revisit this when such reentrancy is necessary, even though I doubt it would be the case in the future. It will at most be invoked from a wq callback or a kthread, so should be fine. > ?
