[PATCHv17,bpf-next,2/6] bpf: add a new bpf argument type ARG_CONST_MAP_PTR_OR_NULL

Commit Message

Hangbin Liu Jan. 25, 2021, 12:45 p.m. UTC

Add a new bpf argument type ARG_CONST_MAP_PTR_OR_NULL which could be
used when we want to allow NULL pointer for map parameter. The bpf helper
need to take care and check if the map is NULL when use this type.

Acked-by: Toke Høiland-Jørgensen <[e-mail]>
Signed-off-by: Hangbin Liu <[e-mail]>

---
v13-v17: no update
v11-v12: rebase the patch to latest bpf-next
v10: remove useless CONST_PTR_TO_MAP_OR_NULL and Copy-paste comment.
v9: merge the patch from [1] in to this series.
v1-v8: no this patch

[1] https://lore.kernel.org/bpf/[e-mail]/
---
 include/linux/bpf.h   |  1 +
 kernel/bpf/verifier.c | 10 ++++++----
 2 files changed, 7 insertions(+), 4 deletions(-)

Comments

John Fastabend Jan. 27, 2021, 9:50 p.m. UTC | #1

Hangbin Liu wrote:
> Add a new bpf argument type ARG_CONST_MAP_PTR_OR_NULL which could be
> used when we want to allow NULL pointer for map parameter. The bpf helper
> need to take care and check if the map is NULL when use this type.
> 
> Acked-by: Toke Høiland-Jørgensen <[e-mail]>
> Signed-off-by: Hangbin Liu <[e-mail]>
> 
> ---
> v13-v17: no update
> v11-v12: rebase the patch to latest bpf-next
> v10: remove useless CONST_PTR_TO_MAP_OR_NULL and Copy-paste comment.
> v9: merge the patch from [1] in to this series.
> v1-v8: no this patch
> 
> [1] https://lore.kernel.org/bpf/[e-mail]/
> ---

LGTM

Acked-by: John Fastabend <[e-mail]>

Patch

diff --git a/include/linux/bpf.h b/include/linux/bpf.h
index 1aac2af12fed..b75207a2484c 100644
--- a/include/linux/bpf.h
+++ b/include/linux/bpf.h
@@ -296,6 +296,7 @@  enum bpf_arg_type {
 	ARG_CONST_ALLOC_SIZE_OR_ZERO,	/* number of allocated bytes requested */
 	ARG_PTR_TO_BTF_ID_SOCK_COMMON,	/* pointer to in-kernel sock_common or bpf-mirrored bpf_sock */
 	ARG_PTR_TO_PERCPU_BTF_ID,	/* pointer to in-kernel percpu type */
+	ARG_CONST_MAP_PTR_OR_NULL,	/* const argument used as pointer to bpf_map or NULL */
 	__BPF_ARG_TYPE_MAX,
 };
 
diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c
index d0eae51b31e4..4d147f39bfe7 100644
--- a/kernel/bpf/verifier.c
+++ b/kernel/bpf/verifier.c
@@ -445,7 +445,8 @@  static bool arg_type_may_be_null(enum bpf_arg_type type)
 	       type == ARG_PTR_TO_MEM_OR_NULL ||
 	       type == ARG_PTR_TO_CTX_OR_NULL ||
 	       type == ARG_PTR_TO_SOCKET_OR_NULL ||
-	       type == ARG_PTR_TO_ALLOC_MEM_OR_NULL;
+	       type == ARG_PTR_TO_ALLOC_MEM_OR_NULL ||
+	       type == ARG_CONST_MAP_PTR_OR_NULL;
 }
 
 /* Determine whether the function releases some resources allocated by another
@@ -4108,6 +4109,7 @@  static const struct bpf_reg_types *compatible_reg_types[__BPF_ARG_TYPE_MAX] = {
 	[ARG_CONST_SIZE_OR_ZERO]	= &scalar_types,
 	[ARG_CONST_ALLOC_SIZE_OR_ZERO]	= &scalar_types,
 	[ARG_CONST_MAP_PTR]		= &const_map_ptr_types,
+	[ARG_CONST_MAP_PTR_OR_NULL]	= &const_map_ptr_types,
 	[ARG_PTR_TO_CTX]		= &context_types,
 	[ARG_PTR_TO_CTX_OR_NULL]	= &context_types,
 	[ARG_PTR_TO_SOCK_COMMON]	= &sock_types,
@@ -4253,9 +4255,9 @@  static int check_func_arg(struct bpf_verifier_env *env, u32 arg,
 		meta->ref_obj_id = reg->ref_obj_id;
 	}
 
-	if (arg_type == ARG_CONST_MAP_PTR) {
-		/* bpf_map_xxx(map_ptr) call: remember that map_ptr */
-		meta->map_ptr = reg->map_ptr;
+	if (arg_type == ARG_CONST_MAP_PTR ||
+	    arg_type == ARG_CONST_MAP_PTR_OR_NULL) {
+		meta->map_ptr = register_is_null(reg) ? NULL : reg->map_ptr;
 	} else if (arg_type == ARG_PTR_TO_MAP_KEY) {
 		/* bpf_map_xxx(..., map_ptr, ..., key) call:
 		 * check that [key, key + map->key_size) are within