* [PATCH bpf-next v5 0/2] bpf: cgroup skb improvements for bpf_prog_test_run
@ 2020-08-02 21:36 Dmitry Yakunin
2020-08-02 21:36 ` [PATCH bpf-next v5 1/2] bpf: setup socket family and addresses in bpf_prog_test_run_skb Dmitry Yakunin
2020-08-02 21:36 ` [PATCH bpf-next v5 2/2] bpf: allow to specify ifindex for skb " Dmitry Yakunin
0 siblings, 2 replies; 5+ messages in thread
From: Dmitry Yakunin @ 2020-08-02 21:36 UTC (permalink / raw)
To: alexei.starovoitov, daniel, netdev, bpf; +Cc: sdf
This patchset contains some improvements for testing cgroup/skb programs
through BPF_PROG_TEST_RUN command.
v2:
- fix build without CONFIG_CGROUP_BPF (kernel test robot <lkp@intel.com>)
v3:
- fix build without CONFIG_IPV6 (kernel test robot <lkp@intel.com>)
v4:
- remove cgroup storage related commits for future rework (Daniel Borkmann)
v5:
- check skb length before access to inet headers (Eric Dumazet)
Dmitry Yakunin (2):
bpf: setup socket family and addresses in bpf_prog_test_run_skb
bpf: allow to specify ifindex for skb in bpf_prog_test_run_skb
net/bpf/test_run.c | 39 ++++++++++++++++++++++--
tools/testing/selftests/bpf/prog_tests/skb_ctx.c | 5 +++
2 files changed, 42 insertions(+), 2 deletions(-)
--
2.7.4
^ permalink raw reply [flat|nested] 5+ messages in thread
* [PATCH bpf-next v5 1/2] bpf: setup socket family and addresses in bpf_prog_test_run_skb
2020-08-02 21:36 [PATCH bpf-next v5 0/2] bpf: cgroup skb improvements for bpf_prog_test_run Dmitry Yakunin
@ 2020-08-02 21:36 ` Dmitry Yakunin
2020-08-02 22:48 ` Alexei Starovoitov
2020-08-02 21:36 ` [PATCH bpf-next v5 2/2] bpf: allow to specify ifindex for skb " Dmitry Yakunin
1 sibling, 1 reply; 5+ messages in thread
From: Dmitry Yakunin @ 2020-08-02 21:36 UTC (permalink / raw)
To: alexei.starovoitov, daniel, netdev, bpf; +Cc: sdf
Now it's impossible to test all branches of cgroup_skb bpf program which
accesses skb->family and skb->{local,remote}_ip{4,6} fields because they
are zeroed during socket allocation. This commit fills socket family and
addresses from related fields in constructed skb.
v2:
- fix build without CONFIG_IPV6 (kernel test robot <lkp@intel.com>)
v3:
- check skb length before access to inet headers (Eric Dumazet)
Signed-off-by: Dmitry Yakunin <zeil@yandex-team.ru>
---
net/bpf/test_run.c | 21 +++++++++++++++++++++
1 file changed, 21 insertions(+)
diff --git a/net/bpf/test_run.c b/net/bpf/test_run.c
index b03c469..8d69295 100644
--- a/net/bpf/test_run.c
+++ b/net/bpf/test_run.c
@@ -449,6 +449,27 @@ int bpf_prog_test_run_skb(struct bpf_prog *prog, const union bpf_attr *kattr,
skb->protocol = eth_type_trans(skb, current->nsproxy->net_ns->loopback_dev);
skb_reset_network_header(skb);
+ switch (skb->protocol) {
+ case htons(ETH_P_IP):
+ sk->sk_family = AF_INET;
+ if (pskb_may_pull(skb, sizeof(struct iphdr))) {
+ sk->sk_rcv_saddr = ip_hdr(skb)->saddr;
+ sk->sk_daddr = ip_hdr(skb)->daddr;
+ }
+ break;
+#if IS_ENABLED(CONFIG_IPV6)
+ case htons(ETH_P_IPV6):
+ sk->sk_family = AF_INET6;
+ if (pskb_may_pull(skb, sizeof(struct ipv6hdr))) {
+ sk->sk_v6_rcv_saddr = ipv6_hdr(skb)->saddr;
+ sk->sk_v6_daddr = ipv6_hdr(skb)->daddr;
+ }
+ break;
+#endif
+ default:
+ break;
+ }
+
if (is_l2)
__skb_push(skb, hh_len);
if (is_direct_pkt_access)
--
2.7.4
^ permalink raw reply related [flat|nested] 5+ messages in thread
* [PATCH bpf-next v5 2/2] bpf: allow to specify ifindex for skb in bpf_prog_test_run_skb
2020-08-02 21:36 [PATCH bpf-next v5 0/2] bpf: cgroup skb improvements for bpf_prog_test_run Dmitry Yakunin
2020-08-02 21:36 ` [PATCH bpf-next v5 1/2] bpf: setup socket family and addresses in bpf_prog_test_run_skb Dmitry Yakunin
@ 2020-08-02 21:36 ` Dmitry Yakunin
1 sibling, 0 replies; 5+ messages in thread
From: Dmitry Yakunin @ 2020-08-02 21:36 UTC (permalink / raw)
To: alexei.starovoitov, daniel, netdev, bpf; +Cc: sdf
Now skb->dev is unconditionally set to the loopback device in current net
namespace. But if we want to test bpf program which contains code branch
based on ifindex condition (eg filters out localhost packets) it is useful
to allow specifying of ifindex from userspace. This patch adds such option
through ctx_in (__sk_buff) parameter.
Signed-off-by: Dmitry Yakunin <zeil@yandex-team.ru>
---
net/bpf/test_run.c | 22 ++++++++++++++++++++--
tools/testing/selftests/bpf/prog_tests/skb_ctx.c | 5 +++++
2 files changed, 25 insertions(+), 2 deletions(-)
diff --git a/net/bpf/test_run.c b/net/bpf/test_run.c
index 8d69295..369ce90 100644
--- a/net/bpf/test_run.c
+++ b/net/bpf/test_run.c
@@ -327,6 +327,12 @@ static int convert___skb_to_skb(struct sk_buff *skb, struct __sk_buff *__skb)
/* priority is allowed */
if (!range_is_zero(__skb, offsetofend(struct __sk_buff, priority),
+ offsetof(struct __sk_buff, ifindex)))
+ return -EINVAL;
+
+ /* ifindex is allowed */
+
+ if (!range_is_zero(__skb, offsetofend(struct __sk_buff, ifindex),
offsetof(struct __sk_buff, cb)))
return -EINVAL;
@@ -381,6 +387,7 @@ static void convert_skb_to___skb(struct sk_buff *skb, struct __sk_buff *__skb)
__skb->mark = skb->mark;
__skb->priority = skb->priority;
+ __skb->ifindex = skb->dev->ifindex;
__skb->tstamp = skb->tstamp;
memcpy(__skb->cb, &cb->data, QDISC_CB_PRIV_LEN);
__skb->wire_len = cb->pkt_len;
@@ -391,6 +398,8 @@ int bpf_prog_test_run_skb(struct bpf_prog *prog, const union bpf_attr *kattr,
union bpf_attr __user *uattr)
{
bool is_l2 = false, is_direct_pkt_access = false;
+ struct net *net = current->nsproxy->net_ns;
+ struct net_device *dev = net->loopback_dev;
u32 size = kattr->test.data_size_in;
u32 repeat = kattr->test.repeat;
struct __sk_buff *ctx = NULL;
@@ -432,7 +441,7 @@ int bpf_prog_test_run_skb(struct bpf_prog *prog, const union bpf_attr *kattr,
kfree(ctx);
return -ENOMEM;
}
- sock_net_set(sk, current->nsproxy->net_ns);
+ sock_net_set(sk, net);
sock_init_data(NULL, sk);
skb = build_skb(data, 0);
@@ -446,7 +455,14 @@ int bpf_prog_test_run_skb(struct bpf_prog *prog, const union bpf_attr *kattr,
skb_reserve(skb, NET_SKB_PAD + NET_IP_ALIGN);
__skb_put(skb, size);
- skb->protocol = eth_type_trans(skb, current->nsproxy->net_ns->loopback_dev);
+ if (ctx && ctx->ifindex > 1) {
+ dev = dev_get_by_index(net, ctx->ifindex);
+ if (!dev) {
+ ret = -ENODEV;
+ goto out;
+ }
+ }
+ skb->protocol = eth_type_trans(skb, dev);
skb_reset_network_header(skb);
switch (skb->protocol) {
@@ -502,6 +518,8 @@ int bpf_prog_test_run_skb(struct bpf_prog *prog, const union bpf_attr *kattr,
ret = bpf_ctx_finish(kattr, uattr, ctx,
sizeof(struct __sk_buff));
out:
+ if (dev && dev != net->loopback_dev)
+ dev_put(dev);
kfree_skb(skb);
bpf_sk_storage_free(sk);
kfree(sk);
diff --git a/tools/testing/selftests/bpf/prog_tests/skb_ctx.c b/tools/testing/selftests/bpf/prog_tests/skb_ctx.c
index 7021b92..25de86a 100644
--- a/tools/testing/selftests/bpf/prog_tests/skb_ctx.c
+++ b/tools/testing/selftests/bpf/prog_tests/skb_ctx.c
@@ -11,6 +11,7 @@ void test_skb_ctx(void)
.cb[3] = 4,
.cb[4] = 5,
.priority = 6,
+ .ifindex = 1,
.tstamp = 7,
.wire_len = 100,
.gso_segs = 8,
@@ -92,6 +93,10 @@ void test_skb_ctx(void)
"ctx_out_priority",
"skb->priority == %d, expected %d\n",
skb.priority, 7);
+ CHECK_ATTR(skb.ifindex != 1,
+ "ctx_out_ifindex",
+ "skb->ifindex == %d, expected %d\n",
+ skb.ifindex, 1);
CHECK_ATTR(skb.tstamp != 8,
"ctx_out_tstamp",
"skb->tstamp == %lld, expected %d\n",
--
2.7.4
^ permalink raw reply related [flat|nested] 5+ messages in thread
* Re: [PATCH bpf-next v5 1/2] bpf: setup socket family and addresses in bpf_prog_test_run_skb
2020-08-02 21:36 ` [PATCH bpf-next v5 1/2] bpf: setup socket family and addresses in bpf_prog_test_run_skb Dmitry Yakunin
@ 2020-08-02 22:48 ` Alexei Starovoitov
0 siblings, 0 replies; 5+ messages in thread
From: Alexei Starovoitov @ 2020-08-02 22:48 UTC (permalink / raw)
To: Dmitry Yakunin; +Cc: daniel, netdev, bpf, sdf
On Mon, Aug 03, 2020 at 12:36:30AM +0300, Dmitry Yakunin wrote:
> Now it's impossible to test all branches of cgroup_skb bpf program which
> accesses skb->family and skb->{local,remote}_ip{4,6} fields because they
> are zeroed during socket allocation. This commit fills socket family and
> addresses from related fields in constructed skb.
>
> v2:
> - fix build without CONFIG_IPV6 (kernel test robot <lkp@intel.com>)
>
> v3:
> - check skb length before access to inet headers (Eric Dumazet)
>
> Signed-off-by: Dmitry Yakunin <zeil@yandex-team.ru>
> ---
> net/bpf/test_run.c | 21 +++++++++++++++++++++
> 1 file changed, 21 insertions(+)
>
> diff --git a/net/bpf/test_run.c b/net/bpf/test_run.c
> index b03c469..8d69295 100644
> --- a/net/bpf/test_run.c
> +++ b/net/bpf/test_run.c
> @@ -449,6 +449,27 @@ int bpf_prog_test_run_skb(struct bpf_prog *prog, const union bpf_attr *kattr,
> skb->protocol = eth_type_trans(skb, current->nsproxy->net_ns->loopback_dev);
> skb_reset_network_header(skb);
>
> + switch (skb->protocol) {
> + case htons(ETH_P_IP):
> + sk->sk_family = AF_INET;
> + if (pskb_may_pull(skb, sizeof(struct iphdr))) {
skb was just inited with __skb_put(skb, size);
Looking at pskb_may_pull() messes with my brain too much,
since it should never go into __pskb_pull_tail path.
Can you open code the skb->len check instead?
if (sizeof(struct iphdr) <= skb_headlen(skb)) {
> + sk->sk_rcv_saddr = ip_hdr(skb)->saddr;
> + sk->sk_daddr = ip_hdr(skb)->daddr;
> + }
> + break;
> +#if IS_ENABLED(CONFIG_IPV6)
> + case htons(ETH_P_IPV6):
> + sk->sk_family = AF_INET6;
> + if (pskb_may_pull(skb, sizeof(struct ipv6hdr))) {
> + sk->sk_v6_rcv_saddr = ipv6_hdr(skb)->saddr;
> + sk->sk_v6_daddr = ipv6_hdr(skb)->daddr;
> + }
> + break;
> +#endif
> + default:
> + break;
> + }
> +
> if (is_l2)
> __skb_push(skb, hh_len);
> if (is_direct_pkt_access)
> --
> 2.7.4
>
^ permalink raw reply [flat|nested] 5+ messages in thread
* [PATCH bpf-next v5 1/2] bpf: setup socket family and addresses in bpf_prog_test_run_skb
2020-08-02 21:30 [PATCH bpf-next v4 0/2] bpf: cgroup skb improvements for bpf_prog_test_run Dmitry Yakunin
@ 2020-08-02 21:30 ` Dmitry Yakunin
0 siblings, 0 replies; 5+ messages in thread
From: Dmitry Yakunin @ 2020-08-02 21:30 UTC (permalink / raw)
To: alexei.starovoitov, daniel, netdev, bpf; +Cc: sdf
Now it's impossible to test all branches of cgroup_skb bpf program which
accesses skb->family and skb->{local,remote}_ip{4,6} fields because they
are zeroed during socket allocation. This commit fills socket family and
addresses from related fields in constructed skb.
v2:
- fix build without CONFIG_IPV6 (kernel test robot <lkp@intel.com>)
v3:
- check skb length before access to inet headers (Eric Dumazet)
Signed-off-by: Dmitry Yakunin <zeil@yandex-team.ru>
---
net/bpf/test_run.c | 21 +++++++++++++++++++++
1 file changed, 21 insertions(+)
diff --git a/net/bpf/test_run.c b/net/bpf/test_run.c
index b03c469..8d69295 100644
--- a/net/bpf/test_run.c
+++ b/net/bpf/test_run.c
@@ -449,6 +449,27 @@ int bpf_prog_test_run_skb(struct bpf_prog *prog, const union bpf_attr *kattr,
skb->protocol = eth_type_trans(skb, current->nsproxy->net_ns->loopback_dev);
skb_reset_network_header(skb);
+ switch (skb->protocol) {
+ case htons(ETH_P_IP):
+ sk->sk_family = AF_INET;
+ if (pskb_may_pull(skb, sizeof(struct iphdr))) {
+ sk->sk_rcv_saddr = ip_hdr(skb)->saddr;
+ sk->sk_daddr = ip_hdr(skb)->daddr;
+ }
+ break;
+#if IS_ENABLED(CONFIG_IPV6)
+ case htons(ETH_P_IPV6):
+ sk->sk_family = AF_INET6;
+ if (pskb_may_pull(skb, sizeof(struct ipv6hdr))) {
+ sk->sk_v6_rcv_saddr = ipv6_hdr(skb)->saddr;
+ sk->sk_v6_daddr = ipv6_hdr(skb)->daddr;
+ }
+ break;
+#endif
+ default:
+ break;
+ }
+
if (is_l2)
__skb_push(skb, hh_len);
if (is_direct_pkt_access)
--
2.7.4
^ permalink raw reply related [flat|nested] 5+ messages in thread
end of thread, other threads:[~2020-08-02 22:48 UTC | newest]
Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-08-02 21:36 [PATCH bpf-next v5 0/2] bpf: cgroup skb improvements for bpf_prog_test_run Dmitry Yakunin
2020-08-02 21:36 ` [PATCH bpf-next v5 1/2] bpf: setup socket family and addresses in bpf_prog_test_run_skb Dmitry Yakunin
2020-08-02 22:48 ` Alexei Starovoitov
2020-08-02 21:36 ` [PATCH bpf-next v5 2/2] bpf: allow to specify ifindex for skb " Dmitry Yakunin
-- strict thread matches above, loose matches on Subject: below --
2020-08-02 21:30 [PATCH bpf-next v4 0/2] bpf: cgroup skb improvements for bpf_prog_test_run Dmitry Yakunin
2020-08-02 21:30 ` [PATCH bpf-next v5 1/2] bpf: setup socket family and addresses in bpf_prog_test_run_skb Dmitry Yakunin
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.