diff mbox series

[bpf,v2] bpf, lpm: fix lookup bug in map_delete_elem

Message ID 20190222131908.6207-1-alban@kinvolk.io
State Accepted
Delegated to: BPF Maintainers
Headers show
Series [bpf,v2] bpf, lpm: fix lookup bug in map_delete_elem | expand

Commit Message

Alban Crequy Feb. 22, 2019, 1:19 p.m. UTC
From: Alban Crequy <alban@kinvolk.io>

trie_delete_elem() was deleting an entry even though it was not matching
if the prefixlen was correct. This patch adds a check on matchlen.

Reproducer:

$ sudo bpftool map create /sys/fs/bpf/mylpm type lpm_trie key 8 value 1 entries 128 name mylpm flags 1
$ sudo bpftool map update pinned /sys/fs/bpf/mylpm key hex 10 00 00 00 aa bb cc dd value hex 01
$ sudo bpftool map dump pinned /sys/fs/bpf/mylpm
key: 10 00 00 00 aa bb cc dd  value: 01
Found 1 element
$ sudo bpftool map delete pinned /sys/fs/bpf/mylpm key hex 10 00 00 00 ff ff ff ff
$ echo $?
0
$ sudo bpftool map dump pinned /sys/fs/bpf/mylpm
Found 0 elements

A similar reproducer is added in the selftests.

Without the patch:

$ sudo ./tools/testing/selftests/bpf/test_lpm_map
test_lpm_map: test_lpm_map.c:485: test_lpm_delete: Assertion `bpf_map_delete_elem(map_fd, key) == -1 && errno == ENOENT' failed.
Aborted

With the patch: test_lpm_map runs without errors.

Fixes: e454cf595853 ("bpf: Implement map_delete_elem for BPF_MAP_TYPE_LPM_TRIE")
Cc: Craig Gallek <kraig@google.com>
Signed-off-by: Alban Crequy <alban@kinvolk.io>

---

Changes v1 to v2:
- add selftest (review from Martin)
- update commitmsg tags (review from Martin)
- rebase on "bpf" tree and test again (review from Martin)
---
 kernel/bpf/lpm_trie.c                      |  1 +
 tools/testing/selftests/bpf/test_lpm_map.c | 10 ++++++++++
 2 files changed, 11 insertions(+)

Comments

Alban Crequy Feb. 22, 2019, 1:38 p.m. UTC | #1
adding cc: Craig Gallek <kraig@google.com>

I got the "git send-email" command wrong, and the cc to Craig was
removed, despite being listed in the commitmsg. Sorry for my mistake.
There is a copy of the patch at this URL, if needed:
https://github.com/kinvolk/linux/commit/5c522b02ee447f2eb060f840ba709af6b425f932

On Fri, Feb 22, 2019 at 2:19 PM Alban Crequy <alban.crequy@gmail.com> wrote:
>
> From: Alban Crequy <alban@kinvolk.io>
>
> trie_delete_elem() was deleting an entry even though it was not matching
> if the prefixlen was correct. This patch adds a check on matchlen.
>
> Reproducer:
>
> $ sudo bpftool map create /sys/fs/bpf/mylpm type lpm_trie key 8 value 1 entries 128 name mylpm flags 1
> $ sudo bpftool map update pinned /sys/fs/bpf/mylpm key hex 10 00 00 00 aa bb cc dd value hex 01
> $ sudo bpftool map dump pinned /sys/fs/bpf/mylpm
> key: 10 00 00 00 aa bb cc dd  value: 01
> Found 1 element
> $ sudo bpftool map delete pinned /sys/fs/bpf/mylpm key hex 10 00 00 00 ff ff ff ff
> $ echo $?
> 0
> $ sudo bpftool map dump pinned /sys/fs/bpf/mylpm
> Found 0 elements
>
> A similar reproducer is added in the selftests.
>
> Without the patch:
>
> $ sudo ./tools/testing/selftests/bpf/test_lpm_map
> test_lpm_map: test_lpm_map.c:485: test_lpm_delete: Assertion `bpf_map_delete_elem(map_fd, key) == -1 && errno == ENOENT' failed.
> Aborted
>
> With the patch: test_lpm_map runs without errors.
>
> Fixes: e454cf595853 ("bpf: Implement map_delete_elem for BPF_MAP_TYPE_LPM_TRIE")
> Cc: Craig Gallek <kraig@google.com>
> Signed-off-by: Alban Crequy <alban@kinvolk.io>
>
> ---
>
> Changes v1 to v2:
> - add selftest (review from Martin)
> - update commitmsg tags (review from Martin)
> - rebase on "bpf" tree and test again (review from Martin)
> ---
>  kernel/bpf/lpm_trie.c                      |  1 +
>  tools/testing/selftests/bpf/test_lpm_map.c | 10 ++++++++++
>  2 files changed, 11 insertions(+)
>
> diff --git a/kernel/bpf/lpm_trie.c b/kernel/bpf/lpm_trie.c
> index abf1002080df..93a5cbbde421 100644
> --- a/kernel/bpf/lpm_trie.c
> +++ b/kernel/bpf/lpm_trie.c
> @@ -471,6 +471,7 @@ static int trie_delete_elem(struct bpf_map *map, void *_key)
>         }
>
>         if (!node || node->prefixlen != key->prefixlen ||
> +           node->prefixlen != matchlen ||
>             (node->flags & LPM_TREE_NODE_FLAG_IM)) {
>                 ret = -ENOENT;
>                 goto out;
> diff --git a/tools/testing/selftests/bpf/test_lpm_map.c b/tools/testing/selftests/bpf/test_lpm_map.c
> index 147e34cfceb7..02d7c871862a 100644
> --- a/tools/testing/selftests/bpf/test_lpm_map.c
> +++ b/tools/testing/selftests/bpf/test_lpm_map.c
> @@ -474,6 +474,16 @@ static void test_lpm_delete(void)
>         assert(bpf_map_lookup_elem(map_fd, key, &value) == -1 &&
>                 errno == ENOENT);
>
> +       key->prefixlen = 30; // unused prefix so far
> +       inet_pton(AF_INET, "192.255.0.0", key->data);
> +       assert(bpf_map_delete_elem(map_fd, key) == -1 &&
> +               errno == ENOENT);
> +
> +       key->prefixlen = 16; // same prefix as the root node
> +       inet_pton(AF_INET, "192.255.0.0", key->data);
> +       assert(bpf_map_delete_elem(map_fd, key) == -1 &&
> +               errno == ENOENT);
> +
>         /* assert initial lookup */
>         key->prefixlen = 32;
>         inet_pton(AF_INET, "192.168.0.1", key->data);
> --
> 2.20.1
>
Craig Gallek Feb. 22, 2019, 3:03 p.m. UTC | #2
On Fri, Feb 22, 2019 at 8:19 AM Alban Crequy <alban.crequy@gmail.com> wrote:
>
> From: Alban Crequy <alban@kinvolk.io>
>
> trie_delete_elem() was deleting an entry even though it was not matching
> if the prefixlen was correct. This patch adds a check on matchlen.
>
> Reproducer:
>
> $ sudo bpftool map create /sys/fs/bpf/mylpm type lpm_trie key 8 value 1 entries 128 name mylpm flags 1
> $ sudo bpftool map update pinned /sys/fs/bpf/mylpm key hex 10 00 00 00 aa bb cc dd value hex 01
> $ sudo bpftool map dump pinned /sys/fs/bpf/mylpm
> key: 10 00 00 00 aa bb cc dd  value: 01
> Found 1 element
> $ sudo bpftool map delete pinned /sys/fs/bpf/mylpm key hex 10 00 00 00 ff ff ff ff
> $ echo $?
> 0
> $ sudo bpftool map dump pinned /sys/fs/bpf/mylpm
> Found 0 elements
>
> A similar reproducer is added in the selftests.
>
> Without the patch:
>
> $ sudo ./tools/testing/selftests/bpf/test_lpm_map
> test_lpm_map: test_lpm_map.c:485: test_lpm_delete: Assertion `bpf_map_delete_elem(map_fd, key) == -1 && errno == ENOENT' failed.
> Aborted
>
> With the patch: test_lpm_map runs without errors.
>
> Fixes: e454cf595853 ("bpf: Implement map_delete_elem for BPF_MAP_TYPE_LPM_TRIE")
> Cc: Craig Gallek <kraig@google.com>
> Signed-off-by: Alban Crequy <alban@kinvolk.io>
>
> ---
Acked-by: Craig Gallek <kraig@google.com>

Good catch, thanks for the fix!
Daniel Borkmann Feb. 22, 2019, 3:22 p.m. UTC | #3
On 02/22/2019 02:19 PM, Alban Crequy wrote:
> From: Alban Crequy <alban@kinvolk.io>
> 
> trie_delete_elem() was deleting an entry even though it was not matching
> if the prefixlen was correct. This patch adds a check on matchlen.
> 
> Reproducer:
> 
> $ sudo bpftool map create /sys/fs/bpf/mylpm type lpm_trie key 8 value 1 entries 128 name mylpm flags 1
> $ sudo bpftool map update pinned /sys/fs/bpf/mylpm key hex 10 00 00 00 aa bb cc dd value hex 01
> $ sudo bpftool map dump pinned /sys/fs/bpf/mylpm
> key: 10 00 00 00 aa bb cc dd  value: 01
> Found 1 element
> $ sudo bpftool map delete pinned /sys/fs/bpf/mylpm key hex 10 00 00 00 ff ff ff ff
> $ echo $?
> 0
> $ sudo bpftool map dump pinned /sys/fs/bpf/mylpm
> Found 0 elements
> 
> A similar reproducer is added in the selftests.
> 
> Without the patch:
> 
> $ sudo ./tools/testing/selftests/bpf/test_lpm_map
> test_lpm_map: test_lpm_map.c:485: test_lpm_delete: Assertion `bpf_map_delete_elem(map_fd, key) == -1 && errno == ENOENT' failed.
> Aborted
> 
> With the patch: test_lpm_map runs without errors.
> 
> Fixes: e454cf595853 ("bpf: Implement map_delete_elem for BPF_MAP_TYPE_LPM_TRIE")
> Cc: Craig Gallek <kraig@google.com>
> Signed-off-by: Alban Crequy <alban@kinvolk.io>

Applied, thanks!
diff mbox series

Patch

diff --git a/kernel/bpf/lpm_trie.c b/kernel/bpf/lpm_trie.c
index abf1002080df..93a5cbbde421 100644
--- a/kernel/bpf/lpm_trie.c
+++ b/kernel/bpf/lpm_trie.c
@@ -471,6 +471,7 @@  static int trie_delete_elem(struct bpf_map *map, void *_key)
 	}
 
 	if (!node || node->prefixlen != key->prefixlen ||
+	    node->prefixlen != matchlen ||
 	    (node->flags & LPM_TREE_NODE_FLAG_IM)) {
 		ret = -ENOENT;
 		goto out;
diff --git a/tools/testing/selftests/bpf/test_lpm_map.c b/tools/testing/selftests/bpf/test_lpm_map.c
index 147e34cfceb7..02d7c871862a 100644
--- a/tools/testing/selftests/bpf/test_lpm_map.c
+++ b/tools/testing/selftests/bpf/test_lpm_map.c
@@ -474,6 +474,16 @@  static void test_lpm_delete(void)
 	assert(bpf_map_lookup_elem(map_fd, key, &value) == -1 &&
 		errno == ENOENT);
 
+	key->prefixlen = 30; // unused prefix so far
+	inet_pton(AF_INET, "192.255.0.0", key->data);
+	assert(bpf_map_delete_elem(map_fd, key) == -1 &&
+		errno == ENOENT);
+
+	key->prefixlen = 16; // same prefix as the root node
+	inet_pton(AF_INET, "192.255.0.0", key->data);
+	assert(bpf_map_delete_elem(map_fd, key) == -1 &&
+		errno == ENOENT);
+
 	/* assert initial lookup */
 	key->prefixlen = 32;
 	inet_pton(AF_INET, "192.168.0.1", key->data);