[PATCH 0/2] bpf: Build with -Wcast-function-type

[PATCH 0/2] bpf: Build with -Wcast-function-type

[Kees Cook]

Hi,

In order to keep ahead of cases in the kernel where Control Flow Integrity
(CFI) may trip over function call casts, enabling -Wcast-function-type
is helpful. To that end, replace BPF_CAST_CALL() as it triggers warnings
with this option and is now one of the last places in the kernel in need
of fixing.

Thanks,

-Kees

Kees Cook (2):
  bpf: Replace "want address" users of BPF_CAST_CALL with BPF_CALL_IMM
  bpf: Replace callers of BPF_CAST_CALL with proper function typedef

 include/linux/bpf.h    |  4 +++-
 include/linux/filter.h |  7 +++----
 kernel/bpf/arraymap.c  |  7 +++----
 kernel/bpf/hashtab.c   | 13 ++++++-------
 kernel/bpf/helpers.c   |  5 ++---
 kernel/bpf/verifier.c  | 26 +++++++++-----------------
 6 files changed, 26 insertions(+), 36 deletions(-)

-- 
2.30.2

[PATCH 1/2] bpf: Replace "want address" users of BPF_CAST_CALL with BPF_CALL_IMM

[Kees Cook]

In order to keep ahead of cases in the kernel where Control Flow
Integrity (CFI) may trip over function call casts, enabling
-Wcast-function-type is helpful. To that end, BPF_CAST_CALL causes
various warnings and is one of the last places in the kernel triggering
this warning.

Most places using BPF_CAST_CALL actually just want a void * to perform
math on. It's not actually performing a call, so just use a different
helper to get the void *, by way of the new BPF_CALL_IMM() helper, which
can clean up a common copy/paste idiom as well.

This change results in no object code difference.

Cc: Alexei Starovoitov <ast@kernel.org>
Cc: Daniel Borkmann <daniel@iogearbox.net>
Cc: Andrii Nakryiko <andrii@kernel.org>
Cc: Martin KaFai Lau <kafai@fb.com>
Cc: Song Liu <songliubraving@fb.com>
Cc: Yonghong Song <yhs@fb.com>
Cc: John Fastabend <john.fastabend@gmail.com>
Cc: KP Singh <kpsingh@kernel.org>
Cc: netdev@vger.kernel.org
Cc: bpf@vger.kernel.org
Cc: Gustavo A. R. Silva <gustavoars@kernel.org>
Link: https://github.com/KSPP/linux/issues/20
Signed-off-by: Kees Cook <keescook@chromium.org>
---
 include/linux/filter.h |  6 +++++-
 kernel/bpf/hashtab.c   |  6 +++---
 kernel/bpf/verifier.c  | 26 +++++++++-----------------
 3 files changed, 17 insertions(+), 21 deletions(-)

diff --git a/include/linux/filter.h b/include/linux/filter.h
index 4a93c12543ee..6c247663d4ce 100644
--- a/include/linux/filter.h
+++ b/include/linux/filter.h
@@ -365,13 +365,17 @@ static inline bool insn_is_zext(const struct bpf_insn *insn)
 #define BPF_CAST_CALL(x)					\
 		((u64 (*)(u64, u64, u64, u64, u64))(x))
 
+/* Convert function address to BPF immediate */
+
+#define BPF_CALL_IMM(x)	((void *)(x) - (void *)__bpf_call_base)
+
 #define BPF_EMIT_CALL(FUNC)					\
 	((struct bpf_insn) {					\
 		.code  = BPF_JMP | BPF_CALL,			\
 		.dst_reg = 0,					\
 		.src_reg = 0,					\
 		.off   = 0,					\
-		.imm   = ((FUNC) - __bpf_call_base) })
+		.imm   = BPF_CALL_IMM(FUNC) })
 
 /* Raw code statement block */
 
diff --git a/kernel/bpf/hashtab.c b/kernel/bpf/hashtab.c
index 32471ba02708..3d8f9d6997d5 100644
--- a/kernel/bpf/hashtab.c
+++ b/kernel/bpf/hashtab.c
@@ -668,7 +668,7 @@ static int htab_map_gen_lookup(struct bpf_map *map, struct bpf_insn *insn_buf)
 
 	BUILD_BUG_ON(!__same_type(&__htab_map_lookup_elem,
 		     (void *(*)(struct bpf_map *map, void *key))NULL));
-	*insn++ = BPF_EMIT_CALL(BPF_CAST_CALL(__htab_map_lookup_elem));
+	*insn++ = BPF_EMIT_CALL(__htab_map_lookup_elem);
 	*insn++ = BPF_JMP_IMM(BPF_JEQ, ret, 0, 1);
 	*insn++ = BPF_ALU64_IMM(BPF_ADD, ret,
 				offsetof(struct htab_elem, key) +
@@ -709,7 +709,7 @@ static int htab_lru_map_gen_lookup(struct bpf_map *map,
 
 	BUILD_BUG_ON(!__same_type(&__htab_map_lookup_elem,
 		     (void *(*)(struct bpf_map *map, void *key))NULL));
-	*insn++ = BPF_EMIT_CALL(BPF_CAST_CALL(__htab_map_lookup_elem));
+	*insn++ = BPF_EMIT_CALL(__htab_map_lookup_elem);
 	*insn++ = BPF_JMP_IMM(BPF_JEQ, ret, 0, 4);
 	*insn++ = BPF_LDX_MEM(BPF_B, ref_reg, ret,
 			      offsetof(struct htab_elem, lru_node) +
@@ -2397,7 +2397,7 @@ static int htab_of_map_gen_lookup(struct bpf_map *map,
 
 	BUILD_BUG_ON(!__same_type(&__htab_map_lookup_elem,
 		     (void *(*)(struct bpf_map *map, void *key))NULL));
-	*insn++ = BPF_EMIT_CALL(BPF_CAST_CALL(__htab_map_lookup_elem));
+	*insn++ = BPF_EMIT_CALL(__htab_map_lookup_elem);
 	*insn++ = BPF_JMP_IMM(BPF_JEQ, ret, 0, 2);
 	*insn++ = BPF_ALU64_IMM(BPF_ADD, ret,
 				offsetof(struct htab_elem, key) +
diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c
index e76b55917905..15068fa7b16d 100644
--- a/kernel/bpf/verifier.c
+++ b/kernel/bpf/verifier.c
@@ -1730,7 +1730,7 @@ static int add_kfunc_call(struct bpf_verifier_env *env, u32 func_id)
 
 	desc = &tab->descs[tab->nr_descs++];
 	desc->func_id = func_id;
-	desc->imm = BPF_CAST_CALL(addr) - __bpf_call_base;
+	desc->imm = BPF_CALL_IMM(addr);
 	err = btf_distill_func_proto(&env->log, btf_vmlinux,
 				     func_proto, func_name,
 				     &desc->func_model);
@@ -12469,8 +12469,7 @@ static int jit_subprogs(struct bpf_verifier_env *env)
 			if (!bpf_pseudo_call(insn))
 				continue;
 			subprog = insn->off;
-			insn->imm = BPF_CAST_CALL(func[subprog]->bpf_func) -
-				    __bpf_call_base;
+			insn->imm = BPF_CALL_IMM(func[subprog]->bpf_func);
 		}
 
 		/* we use the aux data to keep a list of the start addresses
@@ -12950,32 +12949,25 @@ static int do_misc_fixups(struct bpf_verifier_env *env)
 patch_map_ops_generic:
 			switch (insn->imm) {
 			case BPF_FUNC_map_lookup_elem:
-				insn->imm = BPF_CAST_CALL(ops->map_lookup_elem) -
-					    __bpf_call_base;
+				insn->imm = BPF_CALL_IMM(ops->map_lookup_elem);
 				continue;
 			case BPF_FUNC_map_update_elem:
-				insn->imm = BPF_CAST_CALL(ops->map_update_elem) -
-					    __bpf_call_base;
+				insn->imm = BPF_CALL_IMM(ops->map_update_elem);
 				continue;
 			case BPF_FUNC_map_delete_elem:
-				insn->imm = BPF_CAST_CALL(ops->map_delete_elem) -
-					    __bpf_call_base;
+				insn->imm = BPF_CALL_IMM(ops->map_delete_elem);
 				continue;
 			case BPF_FUNC_map_push_elem:
-				insn->imm = BPF_CAST_CALL(ops->map_push_elem) -
-					    __bpf_call_base;
+				insn->imm = BPF_CALL_IMM(ops->map_push_elem);
 				continue;
 			case BPF_FUNC_map_pop_elem:
-				insn->imm = BPF_CAST_CALL(ops->map_pop_elem) -
-					    __bpf_call_base;
+				insn->imm = BPF_CALL_IMM(ops->map_pop_elem);
 				continue;
 			case BPF_FUNC_map_peek_elem:
-				insn->imm = BPF_CAST_CALL(ops->map_peek_elem) -
-					    __bpf_call_base;
+				insn->imm = BPF_CALL_IMM(ops->map_peek_elem);
 				continue;
 			case BPF_FUNC_redirect_map:
-				insn->imm = BPF_CAST_CALL(ops->map_redirect) -
-					    __bpf_call_base;
+				insn->imm = BPF_CALL_IMM(ops->map_redirect);
 				continue;
 			}
 
-- 
2.30.2

[PATCH 2/2] bpf: Replace callers of BPF_CAST_CALL with proper function typedef

[Kees Cook]

In order to keep ahead of cases in the kernel where Control Flow
Integrity (CFI) may trip over function call casts, enabling
-Wcast-function-type is helpful. To that end, BPF_CAST_CALL causes
various warnings and is one of the last places in the kernel
triggering this warning.

For actual function calls, replace BPF_CAST_CALL() with a typedef, which
captures the same details about the given function pointers.

This change results in no object code difference.

Cc: Alexei Starovoitov <ast@kernel.org>
Cc: Daniel Borkmann <daniel@iogearbox.net>
Cc: Andrii Nakryiko <andrii@kernel.org>
Cc: Martin KaFai Lau <kafai@fb.com>
Cc: Song Liu <songliubraving@fb.com>
Cc: Yonghong Song <yhs@fb.com>
Cc: John Fastabend <john.fastabend@gmail.com>
Cc: KP Singh <kpsingh@kernel.org>
Cc: netdev@vger.kernel.org
Cc: bpf@vger.kernel.org
Cc: Gustavo A. R. Silva <gustavoars@kernel.org>
Link: https://github.com/KSPP/linux/issues/20
Signed-off-by: Kees Cook <keescook@chromium.org>
---
 include/linux/bpf.h    | 4 +++-
 include/linux/filter.h | 5 -----
 kernel/bpf/arraymap.c  | 7 +++----
 kernel/bpf/hashtab.c   | 7 +++----
 kernel/bpf/helpers.c   | 5 ++---
 5 files changed, 11 insertions(+), 17 deletions(-)

diff --git a/include/linux/bpf.h b/include/linux/bpf.h
index f4c16f19f83e..ff633f08cb51 100644
--- a/include/linux/bpf.h
+++ b/include/linux/bpf.h
@@ -48,6 +48,7 @@ extern struct idr btf_idr;
 extern spinlock_t btf_idr_lock;
 extern struct kobject *btf_kobj;
 
+typedef u64 (*bpf_callback_t)(u64, u64, u64, u64, u64);
 typedef int (*bpf_iter_init_seq_priv_t)(void *private_data,
 					struct bpf_iter_aux_info *aux);
 typedef void (*bpf_iter_fini_seq_priv_t)(void *private_data);
@@ -142,7 +143,8 @@ struct bpf_map_ops {
 	int (*map_set_for_each_callback_args)(struct bpf_verifier_env *env,
 					      struct bpf_func_state *caller,
 					      struct bpf_func_state *callee);
-	int (*map_for_each_callback)(struct bpf_map *map, void *callback_fn,
+	int (*map_for_each_callback)(struct bpf_map *map,
+				     bpf_callback_t callback_fn,
 				     void *callback_ctx, u64 flags);
 
 	/* BTF name and id of struct allocated by map_alloc */
diff --git a/include/linux/filter.h b/include/linux/filter.h
index 6c247663d4ce..47f80adbe744 100644
--- a/include/linux/filter.h
+++ b/include/linux/filter.h
@@ -360,11 +360,6 @@ static inline bool insn_is_zext(const struct bpf_insn *insn)
 		.off   = 0,					\
 		.imm   = TGT })
 
-/* Function call */
-
-#define BPF_CAST_CALL(x)					\
-		((u64 (*)(u64, u64, u64, u64, u64))(x))
-
 /* Convert function address to BPF immediate */
 
 #define BPF_CALL_IMM(x)	((void *)(x) - (void *)__bpf_call_base)
diff --git a/kernel/bpf/arraymap.c b/kernel/bpf/arraymap.c
index cebd4fb06d19..5e1ccfae916b 100644
--- a/kernel/bpf/arraymap.c
+++ b/kernel/bpf/arraymap.c
@@ -645,7 +645,7 @@ static const struct bpf_iter_seq_info iter_seq_info = {
 	.seq_priv_size		= sizeof(struct bpf_iter_seq_array_map_info),
 };
 
-static int bpf_for_each_array_elem(struct bpf_map *map, void *callback_fn,
+static int bpf_for_each_array_elem(struct bpf_map *map, bpf_callback_t callback_fn,
 				   void *callback_ctx, u64 flags)
 {
 	u32 i, key, num_elems = 0;
@@ -668,9 +668,8 @@ static int bpf_for_each_array_elem(struct bpf_map *map, void *callback_fn,
 			val = array->value + array->elem_size * i;
 		num_elems++;
 		key = i;
-		ret = BPF_CAST_CALL(callback_fn)((u64)(long)map,
-					(u64)(long)&key, (u64)(long)val,
-					(u64)(long)callback_ctx, 0);
+		ret = callback_fn((u64)(long)map, (u64)(long)&key,
+				  (u64)(long)val, (u64)(long)callback_ctx, 0);
 		/* return value: 0 - continue, 1 - stop and return */
 		if (ret)
 			break;
diff --git a/kernel/bpf/hashtab.c b/kernel/bpf/hashtab.c
index 3d8f9d6997d5..d29af9988f37 100644
--- a/kernel/bpf/hashtab.c
+++ b/kernel/bpf/hashtab.c
@@ -2049,7 +2049,7 @@ static const struct bpf_iter_seq_info iter_seq_info = {
 	.seq_priv_size		= sizeof(struct bpf_iter_seq_hash_map_info),
 };
 
-static int bpf_for_each_hash_elem(struct bpf_map *map, void *callback_fn,
+static int bpf_for_each_hash_elem(struct bpf_map *map, bpf_callback_t callback_fn,
 				  void *callback_ctx, u64 flags)
 {
 	struct bpf_htab *htab = container_of(map, struct bpf_htab, map);
@@ -2089,9 +2089,8 @@ static int bpf_for_each_hash_elem(struct bpf_map *map, void *callback_fn,
 				val = elem->key + roundup_key_size;
 			}
 			num_elems++;
-			ret = BPF_CAST_CALL(callback_fn)((u64)(long)map,
-					(u64)(long)key, (u64)(long)val,
-					(u64)(long)callback_ctx, 0);
+			ret = callback_fn((u64)(long)map, (u64)(long)key,
+					  (u64)(long)val, (u64)(long)callback_ctx, 0);
 			/* return value: 0 - continue, 1 - stop and return */
 			if (ret) {
 				rcu_read_unlock();
diff --git a/kernel/bpf/helpers.c b/kernel/bpf/helpers.c
index 9aabf84afd4b..25d7e02ba449 100644
--- a/kernel/bpf/helpers.c
+++ b/kernel/bpf/helpers.c
@@ -1058,7 +1058,7 @@ static enum hrtimer_restart bpf_timer_cb(struct hrtimer *hrtimer)
 	struct bpf_hrtimer *t = container_of(hrtimer, struct bpf_hrtimer, timer);
 	struct bpf_map *map = t->map;
 	void *value = t->value;
-	void *callback_fn;
+	bpf_callback_t callback_fn;
 	void *key;
 	u32 idx;
 
@@ -1083,8 +1083,7 @@ static enum hrtimer_restart bpf_timer_cb(struct hrtimer *hrtimer)
 		key = value - round_up(map->key_size, 8);
 	}
 
-	BPF_CAST_CALL(callback_fn)((u64)(long)map, (u64)(long)key,
-				   (u64)(long)value, 0, 0);
+	callback_fn((u64)(long)map, (u64)(long)key, (u64)(long)value, 0, 0);
 	/* The verifier checked that return value is zero. */
 
 	this_cpu_write(hrtimer_running, NULL);
-- 
2.30.2

Re: [PATCH 0/2] bpf: Build with -Wcast-function-type

[Andrii Nakryiko]

On Mon, Sep 27, 2021 at 11:27 AM Kees Cook <keescook@chromium.org> wrote:
>
> Hi,
>
> In order to keep ahead of cases in the kernel where Control Flow Integrity
> (CFI) may trip over function call casts, enabling -Wcast-function-type
> is helpful. To that end, replace BPF_CAST_CALL() as it triggers warnings
> with this option and is now one of the last places in the kernel in need
> of fixing.
>
> Thanks,
>
> -Kees
>
> Kees Cook (2):
>   bpf: Replace "want address" users of BPF_CAST_CALL with BPF_CALL_IMM
>   bpf: Replace callers of BPF_CAST_CALL with proper function typedef
>

Both patches look good to me. For the series:

Acked-by: Andrii Nakryiko <andrii@kernel.org>

>  include/linux/bpf.h    |  4 +++-
>  include/linux/filter.h |  7 +++----
>  kernel/bpf/arraymap.c  |  7 +++----
>  kernel/bpf/hashtab.c   | 13 ++++++-------
>  kernel/bpf/helpers.c   |  5 ++---
>  kernel/bpf/verifier.c  | 26 +++++++++-----------------
>  6 files changed, 26 insertions(+), 36 deletions(-)
>
> --
> 2.30.2
>

Re: [PATCH 0/2] bpf: Build with -Wcast-function-type

[Alexei Starovoitov]

On Mon, Sep 27, 2021 at 10:24 PM Andrii Nakryiko
<andrii.nakryiko@gmail.com> wrote:
>
> On Mon, Sep 27, 2021 at 11:27 AM Kees Cook <keescook@chromium.org> wrote:
> >
> > Hi,
> >
> > In order to keep ahead of cases in the kernel where Control Flow Integrity
> > (CFI) may trip over function call casts, enabling -Wcast-function-type
> > is helpful. To that end, replace BPF_CAST_CALL() as it triggers warnings
> > with this option and is now one of the last places in the kernel in need
> > of fixing.
> >
> > Thanks,
> >
> > -Kees
> >
> > Kees Cook (2):
> >   bpf: Replace "want address" users of BPF_CAST_CALL with BPF_CALL_IMM
> >   bpf: Replace callers of BPF_CAST_CALL with proper function typedef
> >
>
> Both patches look good to me. For the series:
>
> Acked-by: Andrii Nakryiko <andrii@kernel.org>

It needs a rebase to bpf-next:
In file included from ../lib/test_bpf.c:12:
../lib/test_bpf.c: In function ‘prepare_tail_call_tests’:
../lib/test_bpf.c:12442:27: error: implicit declaration of function
‘BPF_CAST_CALL’; did you mean ‘BPF_EMIT_CALL’?
[-Werror=implicit-function-declaration]
     *insn = BPF_EMIT_CALL(BPF_CAST_CALL(addr));

Please mark the patches as [PATCH bpf-next v2] to help CI pick the right tree.