* [Patch net-next v2] sock_map: dump socket map id via diag
@ 2023-03-13 4:16 Cong Wang
2023-03-14 17:58 ` Stanislav Fomichev
2023-03-15 6:14 ` John Fastabend
0 siblings, 2 replies; 3+ messages in thread
From: Cong Wang @ 2023-03-13 4:16 UTC (permalink / raw)
To: netdev; +Cc: bpf, Cong Wang, John Fastabend, Jakub Sitnicki
From: Cong Wang <cong.wang@bytedance.com>
Currently there is no way to know which sockmap a socket has been added
to from outside, especially for that a socket can be added to multiple
sockmap's. We could dump this via socket diag, as shown below.
Sample output:
# ./iproute2/misc/ss -tnaie --sockmap
ESTAB 0 344329 127.0.0.1:1234 127.0.0.1:40912 ino:21098 sk:5 cgroup:/user.slice/user-0.slice/session-c1.scope <-> sockmap: 1
# bpftool map
1: sockmap flags 0x0
key 4B value 4B max_entries 2 memlock 4096B
pids echo-sockmap(549)
4: array name pid_iter.rodata flags 0x480
key 4B value 4B max_entries 1 memlock 4096B
btf_id 10 frozen
pids bpftool(624)
In the future, we could dump other sockmap related stats too, hence I
make it a nested attribute.
Cc: John Fastabend <john.fastabend@gmail.com>
Cc: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Cong Wang <cong.wang@bytedance.com>
---
v2: rename enum's with more generic names
sock_map_idiag_dump -> sock_map_diag_dump()
make sock_map_diag_dump() return number of maps
include/linux/bpf.h | 1 +
include/uapi/linux/inet_diag.h | 1 +
include/uapi/linux/sock_diag.h | 8 ++++++
include/uapi/linux/unix_diag.h | 1 +
net/core/sock_map.c | 51 ++++++++++++++++++++++++++++++++++
net/ipv4/inet_diag.c | 5 ++++
net/unix/diag.c | 6 ++++
7 files changed, 73 insertions(+)
diff --git a/include/linux/bpf.h b/include/linux/bpf.h
index 6792a7940e1e..4cc315ce26a9 100644
--- a/include/linux/bpf.h
+++ b/include/linux/bpf.h
@@ -2638,6 +2638,7 @@ int sock_map_bpf_prog_query(const union bpf_attr *attr,
void sock_map_unhash(struct sock *sk);
void sock_map_destroy(struct sock *sk);
void sock_map_close(struct sock *sk, long timeout);
+int sock_map_diag_dump(struct sock *sk, struct sk_buff *skb, int attr);
#else
static inline int bpf_dev_bound_kfunc_check(struct bpf_verifier_log *log,
struct bpf_prog_aux *prog_aux)
diff --git a/include/uapi/linux/inet_diag.h b/include/uapi/linux/inet_diag.h
index 50655de04c9b..d1f1e4522633 100644
--- a/include/uapi/linux/inet_diag.h
+++ b/include/uapi/linux/inet_diag.h
@@ -161,6 +161,7 @@ enum {
INET_DIAG_SK_BPF_STORAGES,
INET_DIAG_CGROUP_ID,
INET_DIAG_SOCKOPT,
+ INET_DIAG_BPF_MAP,
__INET_DIAG_MAX,
};
diff --git a/include/uapi/linux/sock_diag.h b/include/uapi/linux/sock_diag.h
index 5f74a5f6091d..7c961940b408 100644
--- a/include/uapi/linux/sock_diag.h
+++ b/include/uapi/linux/sock_diag.h
@@ -62,4 +62,12 @@ enum {
#define SK_DIAG_BPF_STORAGE_MAX (__SK_DIAG_BPF_STORAGE_MAX - 1)
+enum {
+ SK_DIAG_BPF_MAP_NONE,
+ SK_DIAG_BPF_MAP_IDS,
+ __SK_DIAG_BPF_MAP_MAX,
+};
+
+#define SK_DIAG_BPF_MAP_MAX (__SK_DIAG_BPF_MAP_MAX - 1)
+
#endif /* _UAPI__SOCK_DIAG_H__ */
diff --git a/include/uapi/linux/unix_diag.h b/include/uapi/linux/unix_diag.h
index a1988576fa8a..b95a2b33521d 100644
--- a/include/uapi/linux/unix_diag.h
+++ b/include/uapi/linux/unix_diag.h
@@ -42,6 +42,7 @@ enum {
UNIX_DIAG_MEMINFO,
UNIX_DIAG_SHUTDOWN,
UNIX_DIAG_UID,
+ UNIX_DIAG_BPF_MAP,
__UNIX_DIAG_MAX,
};
diff --git a/net/core/sock_map.c b/net/core/sock_map.c
index 9b854e236d23..8c4b3044e7a9 100644
--- a/net/core/sock_map.c
+++ b/net/core/sock_map.c
@@ -1656,6 +1656,57 @@ void sock_map_close(struct sock *sk, long timeout)
}
EXPORT_SYMBOL_GPL(sock_map_close);
+int sock_map_diag_dump(struct sock *sk, struct sk_buff *skb, int attrtype)
+{
+ struct sk_psock_link *link;
+ struct nlattr *nla, *attr;
+ int nr_links = 0, ret = 0;
+ struct sk_psock *psock;
+ u32 *ids;
+
+ rcu_read_lock();
+ psock = sk_psock_get(sk);
+ if (unlikely(!psock)) {
+ rcu_read_unlock();
+ return 0;
+ }
+
+ nla = nla_nest_start_noflag(skb, attrtype);
+ if (!nla) {
+ sk_psock_put(sk, psock);
+ rcu_read_unlock();
+ return -EMSGSIZE;
+ }
+ spin_lock_bh(&psock->link_lock);
+ list_for_each_entry(link, &psock->link, list)
+ nr_links++;
+
+ attr = nla_reserve(skb, SK_DIAG_BPF_MAP_IDS,
+ sizeof(link->map->id) * nr_links);
+ if (!attr) {
+ ret = -EMSGSIZE;
+ goto unlock;
+ }
+
+ ids = nla_data(attr);
+ list_for_each_entry(link, &psock->link, list) {
+ *ids = link->map->id;
+ ids++;
+ }
+unlock:
+ spin_unlock_bh(&psock->link_lock);
+ sk_psock_put(sk, psock);
+ rcu_read_unlock();
+ if (ret) {
+ nla_nest_cancel(skb, nla);
+ } else {
+ ret = nr_links;
+ nla_nest_end(skb, nla);
+ }
+ return ret;
+}
+EXPORT_SYMBOL_GPL(sock_map_diag_dump);
+
static int sock_map_iter_attach_target(struct bpf_prog *prog,
union bpf_iter_link_info *linfo,
struct bpf_iter_aux_info *aux)
diff --git a/net/ipv4/inet_diag.c b/net/ipv4/inet_diag.c
index b812eb36f0e3..0949909d5b46 100644
--- a/net/ipv4/inet_diag.c
+++ b/net/ipv4/inet_diag.c
@@ -197,6 +197,11 @@ int inet_diag_msg_attrs_fill(struct sock *sk, struct sk_buff *skb,
&inet_sockopt))
goto errout;
+#ifdef CONFIG_BPF_SYSCALL
+ if (sock_map_diag_dump(sk, skb, INET_DIAG_BPF_MAP) < 0)
+ goto errout;
+#endif
+
return 0;
errout:
return 1;
diff --git a/net/unix/diag.c b/net/unix/diag.c
index 616b55c5b890..54aa8da2831e 100644
--- a/net/unix/diag.c
+++ b/net/unix/diag.c
@@ -6,6 +6,7 @@
#include <linux/skbuff.h>
#include <linux/module.h>
#include <linux/uidgid.h>
+#include <linux/bpf.h>
#include <net/netlink.h>
#include <net/af_unix.h>
#include <net/tcp_states.h>
@@ -172,6 +173,11 @@ static int sk_diag_fill(struct sock *sk, struct sk_buff *skb, struct unix_diag_r
sk_diag_dump_uid(sk, skb, user_ns))
goto out_nlmsg_trim;
+#ifdef CONFIG_BPF_SYSCALL
+ if (sock_map_diag_dump(sk, skb, UNIX_DIAG_BPF_MAP) < 0)
+ goto out_nlmsg_trim;
+#endif
+
nlmsg_end(skb, nlh);
return 0;
--
2.34.1
^ permalink raw reply related [flat|nested] 3+ messages in thread
* Re: [Patch net-next v2] sock_map: dump socket map id via diag
2023-03-13 4:16 [Patch net-next v2] sock_map: dump socket map id via diag Cong Wang
@ 2023-03-14 17:58 ` Stanislav Fomichev
2023-03-15 6:14 ` John Fastabend
1 sibling, 0 replies; 3+ messages in thread
From: Stanislav Fomichev @ 2023-03-14 17:58 UTC (permalink / raw)
To: Cong Wang; +Cc: netdev, bpf, Cong Wang, John Fastabend, Jakub Sitnicki
On 03/12, Cong Wang wrote:
> From: Cong Wang <cong.wang@bytedance.com>
> Currently there is no way to know which sockmap a socket has been added
> to from outside, especially for that a socket can be added to multiple
> sockmap's. We could dump this via socket diag, as shown below.
> Sample output:
> # ./iproute2/misc/ss -tnaie --sockmap
> ESTAB 0 344329 127.0.0.1:1234 127.0.0.1:40912 ino:21098
> sk:5 cgroup:/user.slice/user-0.slice/session-c1.scope <-> sockmap: 1
> # bpftool map
> 1: sockmap flags 0x0
> key 4B value 4B max_entries 2 memlock 4096B
> pids echo-sockmap(549)
> 4: array name pid_iter.rodata flags 0x480
> key 4B value 4B max_entries 1 memlock 4096B
> btf_id 10 frozen
> pids bpftool(624)
> In the future, we could dump other sockmap related stats too, hence I
> make it a nested attribute.
> Cc: John Fastabend <john.fastabend@gmail.com>
> Cc: Jakub Sitnicki <jakub@cloudflare.com>
> Signed-off-by: Cong Wang <cong.wang@bytedance.com>
> ---
> v2: rename enum's with more generic names
> sock_map_idiag_dump -> sock_map_diag_dump()
> make sock_map_diag_dump() return number of maps
> include/linux/bpf.h | 1 +
> include/uapi/linux/inet_diag.h | 1 +
> include/uapi/linux/sock_diag.h | 8 ++++++
> include/uapi/linux/unix_diag.h | 1 +
> net/core/sock_map.c | 51 ++++++++++++++++++++++++++++++++++
> net/ipv4/inet_diag.c | 5 ++++
> net/unix/diag.c | 6 ++++
> 7 files changed, 73 insertions(+)
> diff --git a/include/linux/bpf.h b/include/linux/bpf.h
> index 6792a7940e1e..4cc315ce26a9 100644
> --- a/include/linux/bpf.h
> +++ b/include/linux/bpf.h
> @@ -2638,6 +2638,7 @@ int sock_map_bpf_prog_query(const union bpf_attr
> *attr,
> void sock_map_unhash(struct sock *sk);
> void sock_map_destroy(struct sock *sk);
> void sock_map_close(struct sock *sk, long timeout);
> +int sock_map_diag_dump(struct sock *sk, struct sk_buff *skb, int attr);
> #else
> static inline int bpf_dev_bound_kfunc_check(struct bpf_verifier_log *log,
> struct bpf_prog_aux *prog_aux)
> diff --git a/include/uapi/linux/inet_diag.h
> b/include/uapi/linux/inet_diag.h
> index 50655de04c9b..d1f1e4522633 100644
> --- a/include/uapi/linux/inet_diag.h
> +++ b/include/uapi/linux/inet_diag.h
> @@ -161,6 +161,7 @@ enum {
> INET_DIAG_SK_BPF_STORAGES,
> INET_DIAG_CGROUP_ID,
> INET_DIAG_SOCKOPT,
> + INET_DIAG_BPF_MAP,
> __INET_DIAG_MAX,
> };
> diff --git a/include/uapi/linux/sock_diag.h
> b/include/uapi/linux/sock_diag.h
> index 5f74a5f6091d..7c961940b408 100644
> --- a/include/uapi/linux/sock_diag.h
> +++ b/include/uapi/linux/sock_diag.h
> @@ -62,4 +62,12 @@ enum {
> #define SK_DIAG_BPF_STORAGE_MAX (__SK_DIAG_BPF_STORAGE_MAX - 1)
> +enum {
> + SK_DIAG_BPF_MAP_NONE,
> + SK_DIAG_BPF_MAP_IDS,
> + __SK_DIAG_BPF_MAP_MAX,
> +};
> +
> +#define SK_DIAG_BPF_MAP_MAX (__SK_DIAG_BPF_MAP_MAX - 1)
> +
> #endif /* _UAPI__SOCK_DIAG_H__ */
> diff --git a/include/uapi/linux/unix_diag.h
> b/include/uapi/linux/unix_diag.h
> index a1988576fa8a..b95a2b33521d 100644
> --- a/include/uapi/linux/unix_diag.h
> +++ b/include/uapi/linux/unix_diag.h
> @@ -42,6 +42,7 @@ enum {
> UNIX_DIAG_MEMINFO,
> UNIX_DIAG_SHUTDOWN,
> UNIX_DIAG_UID,
> + UNIX_DIAG_BPF_MAP,
> __UNIX_DIAG_MAX,
> };
> diff --git a/net/core/sock_map.c b/net/core/sock_map.c
> index 9b854e236d23..8c4b3044e7a9 100644
> --- a/net/core/sock_map.c
> +++ b/net/core/sock_map.c
> @@ -1656,6 +1656,57 @@ void sock_map_close(struct sock *sk, long timeout)
> }
> EXPORT_SYMBOL_GPL(sock_map_close);
> +int sock_map_diag_dump(struct sock *sk, struct sk_buff *skb, int
> attrtype)
> +{
> + struct sk_psock_link *link;
> + struct nlattr *nla, *attr;
> + int nr_links = 0, ret = 0;
> + struct sk_psock *psock;
> + u32 *ids;
> +
[..]
> + rcu_read_lock();
Looks like sk_psock_get already starts RCU read section and returns
a refcnt'ed psock. So maybe drop rcu_read_lock/rcu_read_unlock?
> + psock = sk_psock_get(sk);
> + if (unlikely(!psock)) {
> + rcu_read_unlock();
> + return 0;
> + }
> +
> + nla = nla_nest_start_noflag(skb, attrtype);
> + if (!nla) {
> + sk_psock_put(sk, psock);
> + rcu_read_unlock();
> + return -EMSGSIZE;
> + }
> + spin_lock_bh(&psock->link_lock);
> + list_for_each_entry(link, &psock->link, list)
> + nr_links++;
> +
> + attr = nla_reserve(skb, SK_DIAG_BPF_MAP_IDS,
> + sizeof(link->map->id) * nr_links);
> + if (!attr) {
> + ret = -EMSGSIZE;
> + goto unlock;
> + }
> +
> + ids = nla_data(attr);
> + list_for_each_entry(link, &psock->link, list) {
> + *ids = link->map->id;
> + ids++;
> + }
> +unlock:
> + spin_unlock_bh(&psock->link_lock);
> + sk_psock_put(sk, psock);
> + rcu_read_unlock();
> + if (ret) {
> + nla_nest_cancel(skb, nla);
> + } else {
> + ret = nr_links;
> + nla_nest_end(skb, nla);
> + }
> + return ret;
> +}
> +EXPORT_SYMBOL_GPL(sock_map_diag_dump);
> +
> static int sock_map_iter_attach_target(struct bpf_prog *prog,
> union bpf_iter_link_info *linfo,
> struct bpf_iter_aux_info *aux)
> diff --git a/net/ipv4/inet_diag.c b/net/ipv4/inet_diag.c
> index b812eb36f0e3..0949909d5b46 100644
> --- a/net/ipv4/inet_diag.c
> +++ b/net/ipv4/inet_diag.c
> @@ -197,6 +197,11 @@ int inet_diag_msg_attrs_fill(struct sock *sk, struct
> sk_buff *skb,
> &inet_sockopt))
> goto errout;
> +#ifdef CONFIG_BPF_SYSCALL
> + if (sock_map_diag_dump(sk, skb, INET_DIAG_BPF_MAP) < 0)
> + goto errout;
> +#endif
> +
> return 0;
> errout:
> return 1;
> diff --git a/net/unix/diag.c b/net/unix/diag.c
> index 616b55c5b890..54aa8da2831e 100644
> --- a/net/unix/diag.c
> +++ b/net/unix/diag.c
> @@ -6,6 +6,7 @@
> #include <linux/skbuff.h>
> #include <linux/module.h>
> #include <linux/uidgid.h>
> +#include <linux/bpf.h>
> #include <net/netlink.h>
> #include <net/af_unix.h>
> #include <net/tcp_states.h>
> @@ -172,6 +173,11 @@ static int sk_diag_fill(struct sock *sk, struct
> sk_buff *skb, struct unix_diag_r
> sk_diag_dump_uid(sk, skb, user_ns))
> goto out_nlmsg_trim;
> +#ifdef CONFIG_BPF_SYSCALL
> + if (sock_map_diag_dump(sk, skb, UNIX_DIAG_BPF_MAP) < 0)
> + goto out_nlmsg_trim;
> +#endif
> +
> nlmsg_end(skb, nlh);
> return 0;
> --
> 2.34.1
^ permalink raw reply [flat|nested] 3+ messages in thread
* RE: [Patch net-next v2] sock_map: dump socket map id via diag
2023-03-13 4:16 [Patch net-next v2] sock_map: dump socket map id via diag Cong Wang
2023-03-14 17:58 ` Stanislav Fomichev
@ 2023-03-15 6:14 ` John Fastabend
1 sibling, 0 replies; 3+ messages in thread
From: John Fastabend @ 2023-03-15 6:14 UTC (permalink / raw)
To: Cong Wang, netdev; +Cc: bpf, Cong Wang, John Fastabend, Jakub Sitnicki
Cong Wang wrote:
> From: Cong Wang <cong.wang@bytedance.com>
>
> Currently there is no way to know which sockmap a socket has been added
> to from outside, especially for that a socket can be added to multiple
> sockmap's. We could dump this via socket diag, as shown below.
>
> Sample output:
>
> # ./iproute2/misc/ss -tnaie --sockmap
> ESTAB 0 344329 127.0.0.1:1234 127.0.0.1:40912 ino:21098 sk:5 cgroup:/user.slice/user-0.slice/session-c1.scope <-> sockmap: 1
>
> # bpftool map
> 1: sockmap flags 0x0
> key 4B value 4B max_entries 2 memlock 4096B
> pids echo-sockmap(549)
> 4: array name pid_iter.rodata flags 0x480
> key 4B value 4B max_entries 1 memlock 4096B
> btf_id 10 frozen
> pids bpftool(624)
>
> In the future, we could dump other sockmap related stats too, hence I
> make it a nested attribute.
>
> Cc: John Fastabend <john.fastabend@gmail.com>
> Cc: Jakub Sitnicki <jakub@cloudflare.com>
> Signed-off-by: Cong Wang <cong.wang@bytedance.com>
> ---
> v2: rename enum's with more generic names
> sock_map_idiag_dump -> sock_map_diag_dump()
> make sock_map_diag_dump() return number of maps
>
> include/linux/bpf.h | 1 +
> include/uapi/linux/inet_diag.h | 1 +
> include/uapi/linux/sock_diag.h | 8 ++++++
> include/uapi/linux/unix_diag.h | 1 +
> net/core/sock_map.c | 51 ++++++++++++++++++++++++++++++++++
> net/ipv4/inet_diag.c | 5 ++++
> net/unix/diag.c | 6 ++++
> 7 files changed, 73 insertions(+)
>
[...]
> +int sock_map_diag_dump(struct sock *sk, struct sk_buff *skb, int attrtype)
> +{
> + struct sk_psock_link *link;
> + struct nlattr *nla, *attr;
> + int nr_links = 0, ret = 0;
> + struct sk_psock *psock;
> + u32 *ids;
> +
> + rcu_read_lock();
> + psock = sk_psock_get(sk);
> + if (unlikely(!psock)) {
wont this be the common case because we call this for any sk from
inet_diag_msg_attrs_fill(sk, ...)? Probably drop the unlikely?
> + rcu_read_unlock();
> + return 0;
> + }
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2023-03-15 6:14 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2023-03-13 4:16 [Patch net-next v2] sock_map: dump socket map id via diag Cong Wang
2023-03-14 17:58 ` Stanislav Fomichev
2023-03-15 6:14 ` John Fastabend
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).