Date:   Mon, 28 Mar 2022 11:16:40 -0700
From:   Stanislav Fomichev <sdf@...gle.com>
To:     netdev@...r.kernel.org, bpf@...r.kernel.org
Cc:     ast@...nel.org, daniel@...earbox.net, andrii@...nel.org,
        Stanislav Fomichev <sdf@...gle.com>
Subject: [PATCH bpf-next 3/7] bpf: minimize number of allocated lsm slots per program

Previous patch adds 1:1 mapping between all 211 LSM hooks
and bpf_cgroup program array. Instead of reserving a slot per
possible hook, reserve 10 slots per cgroup for lsm programs.
Those slots are dynamically allocated on demand and reclaimed.
This still adds some bloat to the cgroup and brings us back to
roughly pre-cgroup_bpf_attach_type times.

It should be possible to eventually extend this idea to all hooks if
the memory consumption is unacceptable and shrink overall effective
programs array.

Signed-off-by: Stanislav Fomichev <sdf@...gle.com>
---
 include/linux/bpf-cgroup-defs.h |  4 +-
 include/linux/bpf_lsm.h         |  6 --
 kernel/bpf/bpf_lsm.c            |  9 +--
 kernel/bpf/cgroup.c             | 98 +++++++++++++++++++++++++++++----
 4 files changed, 92 insertions(+), 25 deletions(-)

diff --git a/include/linux/bpf-cgroup-defs.h b/include/linux/bpf-cgroup-defs.h
index 6c661b4df9fa..d42516e86b3a 100644
--- a/include/linux/bpf-cgroup-defs.h
+++ b/include/linux/bpf-cgroup-defs.h
@@ -10,7 +10,9 @@
 
 struct bpf_prog_array;
 
-#define CGROUP_LSM_NUM 211 /* will be addressed in the next patch */
+/* Maximum number of concurrently attachable per-cgroup LSM hooks.
+ */
+#define CGROUP_LSM_NUM 10
 
 enum cgroup_bpf_attach_type {
 	CGROUP_BPF_ATTACH_TYPE_INVALID = -1,
diff --git a/include/linux/bpf_lsm.h b/include/linux/bpf_lsm.h
index 7f0e59f5f9be..613de44aa429 100644
--- a/include/linux/bpf_lsm.h
+++ b/include/linux/bpf_lsm.h
@@ -43,7 +43,6 @@ extern const struct bpf_func_proto bpf_inode_storage_delete_proto;
 void bpf_inode_storage_free(struct inode *inode);
 
 int bpf_lsm_find_cgroup_shim(const struct bpf_prog *prog, bpf_func_t *bpf_func);
-int bpf_lsm_hook_idx(u32 btf_id);
 
 #else /* !CONFIG_BPF_LSM */
 
@@ -74,11 +73,6 @@ static inline int bpf_lsm_find_cgroup_shim(const struct bpf_prog *prog,
 	return -ENOENT;
 }
 
-static inline int bpf_lsm_hook_idx(u32 btf_id)
-{
-	return -EINVAL;
-}
-
 #endif /* CONFIG_BPF_LSM */
 
 #endif /* _LINUX_BPF_LSM_H */
diff --git a/kernel/bpf/bpf_lsm.c b/kernel/bpf/bpf_lsm.c
index 7e4c2a4999de..9cc2f0bf78f1 100644
--- a/kernel/bpf/bpf_lsm.c
+++ b/kernel/bpf/bpf_lsm.c
@@ -57,10 +57,12 @@ static unsigned int __cgroup_bpf_run_lsm_socket(const void *ctx,
 	if (unlikely(!sk))
 		return 0;
 
+	rcu_read_lock(); /* See bpf_lsm_attach_type_get(). */
 	cgrp = sock_cgroup_ptr(&sk->sk_cgrp_data);
 	if (likely(cgrp))
 		ret = BPF_PROG_RUN_ARRAY_CG(cgrp->bpf.effective[prog->aux->cgroup_atype],
 					    ctx, bpf_prog_run, 0);
+	rcu_read_unlock();
 	return ret;
 }
 
@@ -77,7 +79,7 @@ static unsigned int __cgroup_bpf_run_lsm_current(const void *ctx,
 	/*prog = container_of(insn, struct bpf_prog, insnsi);*/
 	prog = (const struct bpf_prog *)((void *)insn - offsetof(struct bpf_prog, insnsi));
 
-	rcu_read_lock();
+	rcu_read_lock(); /* See bpf_lsm_attach_type_get(). */
 	cgrp = task_dfl_cgroup(current);
 	if (likely(cgrp))
 		ret = BPF_PROG_RUN_ARRAY_CG(cgrp->bpf.effective[prog->aux->cgroup_atype],
@@ -122,11 +124,6 @@ int bpf_lsm_find_cgroup_shim(const struct bpf_prog *prog,
 	return 0;
 }
 
-int bpf_lsm_hook_idx(u32 btf_id)
-{
-	return btf_id_set_index(&bpf_lsm_hooks, btf_id);
-}
-
 int bpf_lsm_verify_prog(struct bpf_verifier_log *vlog,
 			const struct bpf_prog *prog)
 {
diff --git a/kernel/bpf/cgroup.c b/kernel/bpf/cgroup.c
index 8fa70bc2aaf7..5ecc680eeff3 100644
--- a/kernel/bpf/cgroup.c
+++ b/kernel/bpf/cgroup.c
@@ -25,6 +25,59 @@
 DEFINE_STATIC_KEY_ARRAY_FALSE(cgroup_bpf_enabled_key, MAX_CGROUP_BPF_ATTACH_TYPE);
 EXPORT_SYMBOL(cgroup_bpf_enabled_key);
 
+/* Readers are protected by rcu+synchronize_rcu.
+ * Writers are protected by cgroup_mutex.
+ */
+refcount_t cgroup_lsm_atype_usecnt[CGROUP_LSM_NUM];
+int cgroup_lsm_atype_btf_id[CGROUP_LSM_NUM];
+
+static enum cgroup_bpf_attach_type bpf_lsm_attach_type_get(u32 attach_btf_id)
+{
+	int i;
+
+	WARN_ON_ONCE(!mutex_is_locked(&cgroup_mutex));
+
+	for (i = 0; i < ARRAY_SIZE(cgroup_lsm_atype_btf_id); i++) {
+		if (cgroup_lsm_atype_btf_id[i] != attach_btf_id)
+			continue;
+
+		refcount_inc(&cgroup_lsm_atype_usecnt[i]);
+		return CGROUP_LSM_START + i;
+	}
+
+	for (i = 0; i < ARRAY_SIZE(cgroup_lsm_atype_usecnt); i++) {
+		if (refcount_read(&cgroup_lsm_atype_usecnt[i]) != 0)
+			continue;
+
+		cgroup_lsm_atype_btf_id[i] = attach_btf_id;
+		refcount_set(&cgroup_lsm_atype_usecnt[i], 1);
+		return CGROUP_LSM_START + i;
+	}
+
+	return -E2BIG;
+}
+
+static void bpf_lsm_attach_type_put(u32 attach_btf_id)
+{
+	int i;
+
+	WARN_ON_ONCE(!mutex_is_locked(&cgroup_mutex));
+
+	for (i = 0; i < ARRAY_SIZE(cgroup_lsm_atype_btf_id); i++) {
+		if (cgroup_lsm_atype_btf_id[i] != attach_btf_id)
+			continue;
+
+		if (refcount_dec_and_test(&cgroup_lsm_atype_usecnt[i])) {
+			/* Wait for any existing users to finish.
+			 */
+			synchronize_rcu();
+		}
+		return;
+	}
+
+	WARN_ON_ONCE(1);
+}
+
 void cgroup_bpf_offline(struct cgroup *cgrp)
 {
 	cgroup_get(cgrp);
@@ -118,6 +171,7 @@ static void bpf_cgroup_lsm_shim_release(struct bpf_prog *prog,
 		return;
 
 	bpf_trampoline_unlink_cgroup_shim(prog);
+	bpf_lsm_attach_type_put(prog->aux->attach_btf_id);
 }
 
 /**
@@ -496,7 +550,9 @@ static int __cgroup_bpf_attach(struct cgroup *cgrp,
 			if (err)
 				return -EINVAL;
 
-			atype = CGROUP_LSM_START + bpf_lsm_hook_idx(p->aux->attach_btf_id);
+			atype = bpf_lsm_attach_type_get(p->aux->attach_btf_id);
+			if (atype < 0)
+				return atype;
 		}
 
 		p->aux->cgroup_atype = atype;
@@ -508,27 +564,37 @@ static int __cgroup_bpf_attach(struct cgroup *cgrp,
 
 	progs = &cgrp->bpf.progs[atype];
 
-	if (!hierarchy_allows_attach(cgrp, atype))
-		return -EPERM;
+	if (!hierarchy_allows_attach(cgrp, atype)) {
+		err = -EPERM;
+		goto cleanup_attach_type;
+	}
 
-	if (!list_empty(progs) && cgrp->bpf.flags[atype] != saved_flags)
+	if (!list_empty(progs) && cgrp->bpf.flags[atype] != saved_flags) {
 		/* Disallow attaching non-overridable on top
 		 * of existing overridable in this cgroup.
 		 * Disallow attaching multi-prog if overridable or none
 		 */
-		return -EPERM;
+		err = -EPERM;
+		goto cleanup_attach_type;
+	}
 
-	if (prog_list_length(progs) >= BPF_CGROUP_MAX_PROGS)
-		return -E2BIG;
+	if (prog_list_length(progs) >= BPF_CGROUP_MAX_PROGS) {
+		err = -E2BIG;
+		goto cleanup_attach_type;
+	}
 
 	pl = find_attach_entry(progs, prog, link, replace_prog,
 			       flags & BPF_F_ALLOW_MULTI);
-	if (IS_ERR(pl))
-		return PTR_ERR(pl);
+	if (IS_ERR(pl)) {
+		err = PTR_ERR(pl);
+		goto cleanup_attach_type;
+	}
 
 	if (bpf_cgroup_storages_alloc(storage, new_storage, type,
-				      prog ? : link->link.prog, cgrp))
-		return -ENOMEM;
+				      prog ? : link->link.prog, cgrp)) {
+		err = -ENOMEM;
+		goto cleanup_attach_type;
+	}
 
 	if (pl) {
 		old_prog = pl->prog;
@@ -536,7 +602,8 @@ static int __cgroup_bpf_attach(struct cgroup *cgrp,
 		pl = kmalloc(sizeof(*pl), GFP_KERNEL);
 		if (!pl) {
 			bpf_cgroup_storages_free(new_storage);
-			return -ENOMEM;
+			err = -ENOMEM;
+			goto cleanup_attach_type;
 		}
 		list_add_tail(&pl->node, progs);
 	}
@@ -581,6 +648,13 @@ static int __cgroup_bpf_attach(struct cgroup *cgrp,
 		list_del(&pl->node);
 		kfree(pl);
 	}
+
+cleanup_attach_type:
+	if (type == BPF_LSM_CGROUP) {
+		struct bpf_prog *p = prog ? : link->link.prog;
+
+		bpf_lsm_attach_type_put(p->aux->attach_btf_id);
+	}
 	return err;
 }
 
-- 
2.35.1.1021.g381101b075-goog

Powered by blists - more mailing lists