[PATCH v1 2/2] cgroup/bpf: only cgroup v2 can be attached by bpf programs

From: Chen Ridong
Date: Fri Oct 18 2024 - 04:42:04 EST


From: Chen Ridong <chenridong@xxxxxxxxxx>

Only cgroup v2 can be attached by bpf programs, so this patch introduces
that cgroup_bpf_inherit and cgroup_bpf_offline can only be called in
cgroup v2, and this can fix the memleak mentioned by commit 04f8ef5643bc
("cgroup: Fix memory leak caused by missing cgroup_bpf_offline"), which
has been reverted.

Fixes: 2b0d3d3e4fcf ("percpu_ref: reduce memory footprint of percpu_ref in fast path")
Fixes: 4bfc0bb2c60e ("bpf: decouple the lifetime of cgroup_bpf from cgroup itself")
Link: https://lore.kernel.org/cgroups/aka2hk5jsel5zomucpwlxsej6iwnfw4qu5jkrmjhyfhesjlfdw@46zxhg5bdnr7/
Signed-off-by: Chen Ridong <chenridong@xxxxxxxxxx>
---
kernel/cgroup/cgroup.c | 17 +++++++++++------
1 file changed, 11 insertions(+), 6 deletions(-)

diff --git a/kernel/cgroup/cgroup.c b/kernel/cgroup/cgroup.c
index 30444e096027..e275eaf2de7f 100644
--- a/kernel/cgroup/cgroup.c
+++ b/kernel/cgroup/cgroup.c
@@ -2140,8 +2140,10 @@ int cgroup_setup_root(struct cgroup_root *root, u16 ss_mask)
if (ret)
goto exit_stats;

- ret = cgroup_bpf_inherit(root_cgrp);
- WARN_ON_ONCE(ret);
+ if (root == &cgrp_dfl_root) {
+ ret = cgroup_bpf_inherit(root_cgrp);
+ WARN_ON_ONCE(ret);
+ }

trace_cgroup_setup_root(root);

@@ -5708,9 +5710,11 @@ static struct cgroup *cgroup_create(struct cgroup *parent, const char *name,
if (ret)
goto out_kernfs_remove;

- ret = cgroup_bpf_inherit(cgrp);
- if (ret)
- goto out_psi_free;
+ if (cgrp->root == &cgrp_dfl_root) {
+ ret = cgroup_bpf_inherit(cgrp);
+ if (ret)
+ goto out_psi_free;
+ }

/*
* New cgroup inherits effective freeze counter, and
@@ -6024,7 +6028,8 @@ static int cgroup_destroy_locked(struct cgroup *cgrp)

cgroup1_check_for_release(parent);

- cgroup_bpf_offline(cgrp);
+ if (cgrp->root == &cgrp_dfl_root)
+ cgroup_bpf_offline(cgrp);

/* put the base reference */
percpu_ref_kill(&cgrp->self.refcnt);
--
2.34.1