This series allows sockops programs to access the network namespace
inode and device via (struct bpf_sock_ops)->netns_ino and ->netns_dev.
This can be useful to apply different policies on different network
namespaces.
In the unlikely case where network namespaces are not compiled in
(CONFIG_NET_NS=n), the verifier will generate code to return netns_dev
as usual and will return 0 for netns_ino.
The generated BPF bytecode for netns_ino is loading the correct
inode number at the time of execution.
However, the generated BPF bytecode for netns_dev is loading an
immediate value determined at BPF-load-time by looking at the
initial network namespace. In practice, this works because all netns
currently use the same virtual device. If this was to change, this
code would need to be updated too.
It also adds sockmap and verifier selftests to cover the new fields.
Partial reads work thanks to commit e2f7fc0ac69 ("bpf: fix undefined
behavior in narrow load handling").
v1 patchset can be found at:
https://lkml.org/lkml/2019/4/12/238
Changes since v1:
- add netns_dev (review from Alexei)
- tools/include/uapi/linux/bpf.h: update with netns_dev
- tools/testing/selftests/bpf/test_sockmap_kern.h: print debugs with
- This is a new selftest (review from Song)
v2 patchest can be found at:
https://lkml.org/lkml/2019/4/18/685
Changes since v2:
- replace __u64 by u64 in kernel code (review from Y Song)
- remove unneeded #else branch: program would be rejected in
is_valid_access (review from Y Song)
- allow partial reads (<u64) (review from Y Song)
- standalone patch for the sync (requested by Y Song)
- update commitmsg to refer to netns_ino
- test partial reads on netns_dev (review from Y Song)
- split in two tests
v3 patchset can be found at:
https://lkml.org/lkml/2019/4/26/740
Changes since v3:
- return netns_dev unconditionally and set netns_ino to 0 if
CONFIG_NET_NS is not enabled (review from Jakub Kicinski)
- use bpf_ctx_record_field_size and bpf_ctx_narrow_access_ok instead of
manually deal with partial reads (review from Y Song)
- update commit message to reflect new code and remove note about
partial reads since it was discussed in the review
- use bpf_ctx_range() and offsetofend()
v4 patchset can be found at:
https://lkml.org/lkml/2019/5/24/714
Changes since v4:
- add netns_dev comment on uapi headers (review from Y Song)
- remove redundant bounds check (review from Y Song)
Alban Crequy (4):
bpf: sock ops: add netns ino and dev in bpf context
bpf: sync bpf.h to tools/ for bpf_sock_ops->netns*
selftests: bpf: read netns_ino from struct bpf_sock_ops
selftests: bpf: verifier: read netns_dev and netns_ino from struct
bpf_sock_ops
include/uapi/linux/bpf.h | 6 ++
net/core/filter.c | 67 +++++++++++++++++++
tools/include/uapi/linux/bpf.h | 6 ++
tools/testing/selftests/bpf/test_sockmap.c | 38 ++++++++++-
.../testing/selftests/bpf/test_sockmap_kern.h | 22 ++++++
.../testing/selftests/bpf/verifier/var_off.c | 53 +++++++++++++++
6 files changed, 189 insertions(+), 3 deletions(-)
--
2.21.0
From: Alban Crequy <[email protected]>
This shows how a sockops program could be restricted to a specific
network namespace. The sockops program looks at the current netns via
(struct bpf_sock_ops)->netns_ino and checks if the value matches the
configuration in the new BPF map "sock_netns".
The test program ./test_sockmap accepts a new parameter "--netns"; the
default value is the current netns found by stat() on /proc/self/ns/net,
so the previous tests still pass:
sudo ./test_sockmap
...
Summary: 412 PASSED 0 FAILED
...
Summary: 824 PASSED 0 FAILED
I run my additional test in the following way:
NETNS=$(readlink /proc/self/ns/net | sed 's/^net:\[\(.*\)\]$/\1/')
CGR=/sys/fs/cgroup/unified/user.slice/user-1000.slice/session-5.scope/
sudo ./test_sockmap --cgroup $CGR --netns $NETNS &
cat /sys/kernel/debug/tracing/trace_pipe
echo foo | nc -l 127.0.0.1 8080 &
echo bar | nc 127.0.0.1 8080
=> the connection goes through the sockmap
When testing with a wrong $NETNS, I get the trace_pipe log:
> not binding connection on netns 4026531992
Signed-off-by: Alban Crequy <[email protected]>
---
Changes since v1:
- tools/include/uapi/linux/bpf.h: update with netns_dev
- tools/testing/selftests/bpf/test_sockmap_kern.h: print debugs with
both netns_dev and netns_ino
Changes since v2:
- update commitmsg to refer to netns_ino
---
tools/testing/selftests/bpf/test_sockmap.c | 38 +++++++++++++++++--
.../testing/selftests/bpf/test_sockmap_kern.h | 22 +++++++++++
2 files changed, 57 insertions(+), 3 deletions(-)
diff --git a/tools/testing/selftests/bpf/test_sockmap.c b/tools/testing/selftests/bpf/test_sockmap.c
index 3845144e2c91..5a1b9c96fca1 100644
--- a/tools/testing/selftests/bpf/test_sockmap.c
+++ b/tools/testing/selftests/bpf/test_sockmap.c
@@ -2,6 +2,7 @@
// Copyright (c) 2017-2018 Covalent IO, Inc. http://covalent.io
#include <stdio.h>
#include <stdlib.h>
+#include <stdint.h>
#include <sys/socket.h>
#include <sys/ioctl.h>
#include <sys/select.h>
@@ -21,6 +22,7 @@
#include <sys/resource.h>
#include <sys/types.h>
#include <sys/sendfile.h>
+#include <sys/stat.h>
#include <linux/netlink.h>
#include <linux/socket.h>
@@ -63,8 +65,8 @@ int s1, s2, c1, c2, p1, p2;
int test_cnt;
int passed;
int failed;
-int map_fd[8];
-struct bpf_map *maps[8];
+int map_fd[9];
+struct bpf_map *maps[9];
int prog_fd[11];
int txmsg_pass;
@@ -84,6 +86,7 @@ int txmsg_ingress;
int txmsg_skb;
int ktls;
int peek_flag;
+uint64_t netns_opt;
static const struct option long_options[] = {
{"help", no_argument, NULL, 'h' },
@@ -111,6 +114,7 @@ static const struct option long_options[] = {
{"txmsg_skb", no_argument, &txmsg_skb, 1 },
{"ktls", no_argument, &ktls, 1 },
{"peek", no_argument, &peek_flag, 1 },
+ {"netns", required_argument, NULL, 'n'},
{0, 0, NULL, 0 }
};
@@ -1585,6 +1589,7 @@ char *map_names[] = {
"sock_bytes",
"sock_redir_flags",
"sock_skb_opts",
+ "sock_netns",
};
int prog_attach_type[] = {
@@ -1619,6 +1624,8 @@ static int populate_progs(char *bpf_file)
struct bpf_object *obj;
int i = 0;
long err;
+ struct stat netns_sb;
+ uint64_t netns_ino;
obj = bpf_object__open(bpf_file);
err = libbpf_get_error(obj);
@@ -1655,6 +1662,28 @@ static int populate_progs(char *bpf_file)
}
}
+ if (netns_opt == 0) {
+ err = stat("/proc/self/ns/net", &netns_sb);
+ if (err) {
+ fprintf(stderr,
+ "ERROR: cannot stat network namespace: %ld (%s)\n",
+ err, strerror(errno));
+ return -1;
+ }
+ netns_ino = netns_sb.st_ino;
+ } else {
+ netns_ino = netns_opt;
+ }
+ i = 1;
+ err = bpf_map_update_elem(map_fd[8], &netns_ino, &i, BPF_ANY);
+ if (err) {
+ fprintf(stderr,
+ "ERROR: bpf_map_update_elem (netns): %ld (%s)\n",
+ err, strerror(errno));
+ return -1;
+ }
+
+
return 0;
}
@@ -1738,7 +1767,7 @@ int main(int argc, char **argv)
if (argc < 2)
return test_suite(-1);
- while ((opt = getopt_long(argc, argv, ":dhvc:r:i:l:t:p:q:",
+ while ((opt = getopt_long(argc, argv, ":dhvc:r:i:l:t:p:q:n:",
long_options, &longindex)) != -1) {
switch (opt) {
case 's':
@@ -1805,6 +1834,9 @@ int main(int argc, char **argv)
return -1;
}
break;
+ case 'n':
+ netns_opt = strtoull(optarg, NULL, 10);
+ break;
case 0:
break;
case 'h':
diff --git a/tools/testing/selftests/bpf/test_sockmap_kern.h b/tools/testing/selftests/bpf/test_sockmap_kern.h
index e7639f66a941..317406dad6cf 100644
--- a/tools/testing/selftests/bpf/test_sockmap_kern.h
+++ b/tools/testing/selftests/bpf/test_sockmap_kern.h
@@ -91,6 +91,13 @@ struct bpf_map_def SEC("maps") sock_skb_opts = {
.max_entries = 1
};
+struct bpf_map_def SEC("maps") sock_netns = {
+ .type = BPF_MAP_TYPE_HASH,
+ .key_size = sizeof(__u64),
+ .value_size = sizeof(int),
+ .max_entries = 16
+};
+
SEC("sk_skb1")
int bpf_prog1(struct __sk_buff *skb)
{
@@ -132,9 +139,24 @@ int bpf_sockmap(struct bpf_sock_ops *skops)
{
__u32 lport, rport;
int op, err = 0, index, key, ret;
+ int i = 0;
+ __u64 netns_dev, netns_ino;
+ int *allowed;
op = (int) skops->op;
+ netns_dev = skops->netns_dev;
+ netns_ino = skops->netns_ino;
+ bpf_printk("bpf_sockmap: netns_dev = %lu netns_ino = %lu\n",
+ netns_dev, netns_ino);
+
+ // Only allow sockmap connection on the configured network namespace
+ allowed = bpf_map_lookup_elem(&sock_netns, &netns_ino);
+ if (allowed == NULL || *allowed == 0) {
+ bpf_printk("not binding connection on netns_ino %lu\n",
+ netns_ino);
+ return 0;
+ }
switch (op) {
case BPF_SOCK_OPS_PASSIVE_ESTABLISHED_CB:
--
2.21.0