bpf: Support refcounted local kptrs in existing semantics

A local kptr is considered 'refcounted' when it is of a type that has a
bpf_refcount field. When such a kptr is created, its refcount should be
initialized to 1; when destroyed, the object should be free'd only if a
refcount decr results in 0 refcount.

Existing logic always frees the underlying memory when destroying a
local kptr, and 0-initializes all btf_record fields. This patch adds
checks for "is local kptr refcounted?" and new logic for that case in
the appropriate places.

This patch focuses on changing existing semantics and thus conspicuously
does _not_ provide a way for BPF programs in increment refcount. That
follows later in the series.

__bpf_obj_drop_impl is modified to do the right thing when it sees a
refcounted type. Container types for graph nodes (list, tree, stashed in
map) are migrated to use __bpf_obj_drop_impl as a destructor for their
nodes instead of each having custom destruction code in their _free
paths. Now that "drop" isn't a synonym for "free" when the type is
refcounted it makes sense to centralize this logic.

Signed-off-by: Dave Marchevsky <davemarchevsky@fb.com>
Link: https://lore.kernel.org/r/20230415201811.343116-4-davemarchevsky@fb.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>

include/linux/bpf.h

@@ -370,6 +370,9 @@ static inline void bpf_obj_init(const struct btf_record *rec, void *obj)
 		return;
 	for (i = 0; i < rec->cnt; i++)
 		memset(obj + rec->fields[i].offset, 0, rec->fields[i].size);
+
+	if (rec->refcount_off >= 0)
+		refcount_set((refcount_t *)(obj + rec->refcount_off), 1);
 }
 
 /* 'dst' must be a temporary buffer and should not point to memory that is being

kernel/bpf/helpers.c

@@ -1798,6 +1798,8 @@ bpf_base_func_proto(enum bpf_func_id func_id)
 	}
 }
 
+void __bpf_obj_drop_impl(void *p, const struct btf_record *rec);
+
 void bpf_list_head_free(const struct btf_field *field, void *list_head,
 			struct bpf_spin_lock *spin_lock)
 {
@@ -1828,13 +1830,8 @@ unlock:
 		/* The contained type can also have resources, including a
 		 * bpf_list_head which needs to be freed.
 		 */
-		bpf_obj_free_fields(field->graph_root.value_rec, obj);
-		/* bpf_mem_free requires migrate_disable(), since we can be
-		 * called from map free path as well apart from BPF program (as
-		 * part of map ops doing bpf_obj_free_fields).
-		 */
 		migrate_disable();
-		bpf_mem_free(&bpf_global_ma, obj);
+		__bpf_obj_drop_impl(obj, field->graph_root.value_rec);
 		migrate_enable();
 	}
 }
@@ -1871,10 +1868,9 @@ void bpf_rb_root_free(const struct btf_field *field, void *rb_root,
 		obj = pos;
 		obj -= field->graph_root.node_offset;
 
-		bpf_obj_free_fields(field->graph_root.value_rec, obj);
 
 		migrate_disable();
-		bpf_mem_free(&bpf_global_ma, obj);
+		__bpf_obj_drop_impl(obj, field->graph_root.value_rec);
 		migrate_enable();
 	}
 }
@@ -1897,8 +1893,17 @@ __bpf_kfunc void *bpf_obj_new_impl(u64 local_type_id__k, void *meta__ign)
 	return p;
 }
 
+/* Must be called under migrate_disable(), as required by bpf_mem_free */
 void __bpf_obj_drop_impl(void *p, const struct btf_record *rec)
 {
+	if (rec && rec->refcount_off >= 0 &&
+	    !refcount_dec_and_test((refcount_t *)(p + rec->refcount_off))) {
+		/* Object is refcounted and refcount_dec didn't result in 0
+		 * refcount. Return without freeing the object
+		 */
+		return;
+	}
+
 	if (rec)
 		bpf_obj_free_fields(rec, p);
 	bpf_mem_free(&bpf_global_ma, p);