2018-03-13 03:17:29

by Gustavo A. R. Silva

[permalink] [raw]
Subject: [PATCH v2] netfilter: nf_tables: remove VLA usage

In preparation to enabling -Wvla, remove VLA and replace it
with dynamic memory allocation.

From a security viewpoint, the use of Variable Length Arrays can be
a vector for stack overflow attacks. Also, in general, as the code
evolves it is easy to lose track of how big a VLA can get. Thus, we
can end up having segfaults that are hard to debug.

Also, fixed as part of the directive to remove all VLAs from
the kernel: https://lkml.org/lkml/2018/3/7/621

Signed-off-by: Gustavo A. R. Silva <[email protected]>
---
Changes in v2:
- Use kmalloc_array instead of kcalloc.

net/netfilter/nf_tables_api.c | 23 +++++++++++++++--------
1 file changed, 15 insertions(+), 8 deletions(-)

diff --git a/net/netfilter/nf_tables_api.c b/net/netfilter/nf_tables_api.c
index 3f815b6..5a42e97 100644
--- a/net/netfilter/nf_tables_api.c
+++ b/net/netfilter/nf_tables_api.c
@@ -4357,16 +4357,20 @@ static struct nft_object *nft_obj_init(const struct nft_ctx *ctx,
const struct nft_object_type *type,
const struct nlattr *attr)
{
- struct nlattr *tb[type->maxattr + 1];
+ struct nlattr **tb;
const struct nft_object_ops *ops;
struct nft_object *obj;
- int err;
+ int err = -ENOMEM;
+
+ tb = kmalloc_array(type->maxattr + 1, sizeof(*tb), GFP_KERNEL);
+ if (!tb)
+ goto err1;

if (attr) {
err = nla_parse_nested(tb, type->maxattr, attr, type->policy,
NULL);
if (err < 0)
- goto err1;
+ goto err2;
} else {
memset(tb, 0, sizeof(tb[0]) * (type->maxattr + 1));
}
@@ -4375,7 +4379,7 @@ static struct nft_object *nft_obj_init(const struct nft_ctx *ctx,
ops = type->select_ops(ctx, (const struct nlattr * const *)tb);
if (IS_ERR(ops)) {
err = PTR_ERR(ops);
- goto err1;
+ goto err2;
}
} else {
ops = type->ops;
@@ -4383,18 +4387,21 @@ static struct nft_object *nft_obj_init(const struct nft_ctx *ctx,

err = -ENOMEM;
obj = kzalloc(sizeof(*obj) + ops->size, GFP_KERNEL);
- if (obj == NULL)
- goto err1;
+ if (!obj)
+ goto err2;

err = ops->init(ctx, (const struct nlattr * const *)tb, obj);
if (err < 0)
- goto err2;
+ goto err3;

obj->ops = ops;

+ kfree(tb);
return obj;
-err2:
+err3:
kfree(obj);
+err2:
+ kfree(tb);
err1:
return ERR_PTR(err);
}
--
2.7.4



2018-03-20 12:38:31

by Pablo Neira Ayuso

[permalink] [raw]
Subject: Re: [PATCH v2] netfilter: nf_tables: remove VLA usage

On Mon, Mar 12, 2018 at 10:16:17PM -0500, Gustavo A. R. Silva wrote:
> In preparation to enabling -Wvla, remove VLA and replace it
> with dynamic memory allocation.
>
> From a security viewpoint, the use of Variable Length Arrays can be
> a vector for stack overflow attacks. Also, in general, as the code
> evolves it is easy to lose track of how big a VLA can get. Thus, we
> can end up having segfaults that are hard to debug.
>
> Also, fixed as part of the directive to remove all VLAs from
> the kernel: https://lkml.org/lkml/2018/3/7/621

Applied, thanks.