[bpf,v3,2/3] bpf: Fix hashtab overflow check on 32-bit arches

Commit Message

Toke Høiland-Jørgensen March 7, 2024, 12:03 p.m. UTC

The hashtab code relies on roundup_pow_of_two() to compute the number of
hash buckets, and contains an overflow check by checking if the
resulting value is 0. However, on 32-bit arches, the roundup code itself
can overflow by doing a 32-bit left-shift of an unsigned long value,
which is undefined behaviour, so it is not guaranteed to truncate
neatly. This was triggered by syzbot on the DEVMAP_HASH type, which
contains the same check, copied from the hashtab code. So apply the same
fix to hashtab, by moving the overflow check to before the roundup.

Fixes: daaf427c6ab3 ("bpf: fix arraymap NULL deref and missing overflow and zero size checks")
Signed-off-by: Toke Høiland-Jørgensen <[e-mail]>
---
 kernel/bpf/hashtab.c | 14 +++++++++-----
 1 file changed, 9 insertions(+), 5 deletions(-)

Comments

Alexei Starovoitov March 8, 2024, 4:09 a.m. UTC | #1

On Thu, Mar 7, 2024 at 4:03 AM Toke Høiland-Jørgensen <[e-mail]> wrote:
>
> -       err = -E2BIG;
> -       /* prevent zero size kmalloc and check for u32 overflow */
> -       if (htab->n_buckets == 0 ||
> -           htab->n_buckets > U32_MAX / sizeof(struct bucket))
> +       /* prevent zero size kmalloc */
> +       if (htab->n_buckets > U32_MAX / sizeof(struct bucket))

The comment carried over the wrong half of the comment.
I fixed it up while applying... to bpf-next.
There probably will be no bpf pr,
since the merge window is about to start.

Thanks!

Toke Høiland-Jørgensen March 11, 2024, 10:58 a.m. UTC | #2

Alexei Starovoitov <[e-mail]> writes:

> On Thu, Mar 7, 2024 at 4:03 AM Toke Høiland-Jørgensen <[e-mail]> wrote:
>>
>> -       err = -E2BIG;
>> -       /* prevent zero size kmalloc and check for u32 overflow */
>> -       if (htab->n_buckets == 0 ||
>> -           htab->n_buckets > U32_MAX / sizeof(struct bucket))
>> +       /* prevent zero size kmalloc */
>> +       if (htab->n_buckets > U32_MAX / sizeof(struct bucket))
>
> The comment carried over the wrong half of the comment.
> I fixed it up while applying... to bpf-next.
> There probably will be no bpf pr,
> since the merge window is about to start.

Oops, yeah, completely misread that comment; thanks for fixing! :)

-Toke

Patch

diff --git a/kernel/bpf/hashtab.c b/kernel/bpf/hashtab.c
index 03a6a2500b6a..0cac6a65235c 100644
--- a/kernel/bpf/hashtab.c
+++ b/kernel/bpf/hashtab.c
@@ -499,7 +499,13 @@  static struct bpf_map *htab_map_alloc(union bpf_attr *attr)
 							  num_possible_cpus());
 	}
 
-	/* hash table size must be power of 2 */
+	/* hash table size must be power of 2; roundup_pow_of_two() can overflow
+	 * into UB on 32-bit arches, so check that first
+	 */
+	err = -E2BIG;
+	if (htab->map.max_entries > 1UL << 31)
+		goto free_htab;
+
 	htab->n_buckets = roundup_pow_of_two(htab->map.max_entries);
 
 	htab->elem_size = sizeof(struct htab_elem) +
@@ -509,10 +515,8 @@  static struct bpf_map *htab_map_alloc(union bpf_attr *attr)
 	else
 		htab->elem_size += round_up(htab->map.value_size, 8);
 
-	err = -E2BIG;
-	/* prevent zero size kmalloc and check for u32 overflow */
-	if (htab->n_buckets == 0 ||
-	    htab->n_buckets > U32_MAX / sizeof(struct bucket))
+	/* prevent zero size kmalloc */
+	if (htab->n_buckets > U32_MAX / sizeof(struct bucket))
 		goto free_htab;
 
 	err = bpf_map_init_elem_count(&htab->map);