Re: [PATCH bpf-next v5 1/2] bpf: support BPF_PROG_QUERY for progs attached to sockmap

From: zhudi (E)
Date: Thu Nov 04 2021 - 02:36:07 EST




> -----邮件原件-----
> 发件人: Yonghong Song [mailto:yhs@xxxxxx]
> 发送时间: 2021年11月4日 14:31
> 收件人: zhudi (E) <zhudi2@xxxxxxxxxx>; davem@xxxxxxxxxxxxx;
> ast@xxxxxxxxxx; daniel@xxxxxxxxxxxxx; andrii@xxxxxxxxxx; kafai@xxxxxx;
> songliubraving@xxxxxx; john.fastabend@xxxxxxxxx; kpsingh@xxxxxxxxxx;
> jakub@xxxxxxxxxxxxxx
> 抄送: bpf@xxxxxxxxxxxxxxx; linux-kernel@xxxxxxxxxxxxxxx
> 主题: Re: [PATCH bpf-next v5 1/2] bpf: support BPF_PROG_QUERY for progs
> attached to sockmap
>
>
>
> On 11/3/21 11:07 PM, zhudi (E) wrote:
> > > On 11/3/21 6:07 PM, Di Zhu wrote:
> >>> Right now there is no way to query whether BPF programs are
> >>> attached to a sockmap or not.
> >>>
> >>> we can use the standard interface in libbpf to query, such as:
> >>> bpf_prog_query(mapFd, BPF_SK_SKB_STREAM_PARSER, 0, NULL, ...);
> >>> the mapFd is the fd of sockmap.
> >>>
> >>> Signed-off-by: Di Zhu <zhudi2@xxxxxxxxxx>
> >>> ---
> >>> /* v2 */
> >>> - John Fastabend <john.fastabend@xxxxxxxxx>
> >>> - add selftest code
> >>>
> >>> /* v3 */
> >>> - avoid sleeping caused by copy_to_user() in rcu critical zone
> >>>
> >>> /* v4 */
> >>> - Alexei Starovoitov <alexei.starovoitov@xxxxxxxxx>
> >>> -split into two patches, one for core code and one for selftest.
> >>>
> >>> /* v5 */
> >>> - Yonghong Song <yhs@xxxxxx>
> >>> -Some naming and formatting changes
> >>> ---
> >>> include/linux/bpf.h | 9 ++++++
> >>> kernel/bpf/syscall.c | 5 +++
> >>> net/core/sock_map.c | 74
> >> +++++++++++++++++++++++++++++++++++++++-----
> >>> 3 files changed, 81 insertions(+), 7 deletions(-)
> >>>
> >>> diff --git a/include/linux/bpf.h b/include/linux/bpf.h
> >>> index d604c8251d88..235ea7fc5fd8 100644
> >>> --- a/include/linux/bpf.h
> >>> +++ b/include/linux/bpf.h
> >>> @@ -1961,6 +1961,9 @@ int bpf_prog_test_run_syscall(struct bpf_prog
> >> *prog,
> >>> int sock_map_get_from_fd(const union bpf_attr *attr, struct bpf_prog
> >> *prog);
> >>> int sock_map_prog_detach(const union bpf_attr *attr, enum
> bpf_prog_type
> >> ptype);
> >>> int sock_map_update_elem_sys(struct bpf_map *map, void *key, void
> >> *value, u64 flags);
> >>> +int sock_map_bpf_prog_query(const union bpf_attr *attr,
> >>> + union bpf_attr __user *uattr);
> >>> +
> >>> void sock_map_unhash(struct sock *sk);
> >>> void sock_map_close(struct sock *sk, long timeout);
> >>> #else
> >>> @@ -2014,6 +2017,12 @@ static inline int
> sock_map_update_elem_sys(struct
> >> bpf_map *map, void *key, void
> >>> {
> >>> return -EOPNOTSUPP;
> >>> }
> >>> +
> >>> +static inline int sock_map_bpf_prog_query(const union bpf_attr *attr,
> >>> + union bpf_attr __user *uattr)
> >>> +{
> >>> + return -EINVAL;
> >>> +}
> >>> #endif /* CONFIG_BPF_SYSCALL */
> >>> #endif /* CONFIG_NET && CONFIG_BPF_SYSCALL */
> >>>
> >>> diff --git a/kernel/bpf/syscall.c b/kernel/bpf/syscall.c
> >>> index 4e50c0bfdb7d..748102c3e0c9 100644
> >>> --- a/kernel/bpf/syscall.c
> >>> +++ b/kernel/bpf/syscall.c
> >>> @@ -3275,6 +3275,11 @@ static int bpf_prog_query(const union bpf_attr
> >> *attr,
> >>> case BPF_FLOW_DISSECTOR:
> >>> case BPF_SK_LOOKUP:
> >>> return netns_bpf_prog_query(attr, uattr);
> >>> + case BPF_SK_SKB_STREAM_PARSER:
> >>> + case BPF_SK_SKB_STREAM_VERDICT:
> >>> + case BPF_SK_MSG_VERDICT:
> >>> + case BPF_SK_SKB_VERDICT:
> >>> + return sock_map_bpf_prog_query(attr, uattr);
> >>> default:
> >>> return -EINVAL;
> >>> }
> >>> diff --git a/net/core/sock_map.c b/net/core/sock_map.c
> >>> index e252b8ec2b85..0320d27550fe 100644
> >>> --- a/net/core/sock_map.c
> >>> +++ b/net/core/sock_map.c
> >>> @@ -1412,38 +1412,50 @@ static struct sk_psock_progs
> >> *sock_map_progs(struct bpf_map *map)
> >>> return NULL;
> >>> }
> >>>
> >>> -static int sock_map_prog_update(struct bpf_map *map, struct bpf_prog
> >> *prog,
> >>> - struct bpf_prog *old, u32 which)
> >>> +static int sock_map_prog_lookup(struct bpf_map *map, struct bpf_prog
> >> ***pprog,
> >>> + u32 which)
> >>> {
> >>> struct sk_psock_progs *progs = sock_map_progs(map);
> >>> - struct bpf_prog **pprog;
> >>>
> >>> if (!progs)
> >>> return -EOPNOTSUPP;
> >>>
> >>> switch (which) {
> >>> case BPF_SK_MSG_VERDICT:
> >>> - pprog = &progs->msg_parser;
> >>> + *pprog = &progs->msg_parser;
> >>> break;
> >>> #if IS_ENABLED(CONFIG_BPF_STREAM_PARSER)
> >>> case BPF_SK_SKB_STREAM_PARSER:
> >>> - pprog = &progs->stream_parser;
> >>> + *pprog = &progs->stream_parser;
> >>> break;
> >>> #endif
> >>> case BPF_SK_SKB_STREAM_VERDICT:
> >>> if (progs->skb_verdict)
> >>> return -EBUSY;
> >>> - pprog = &progs->stream_verdict;
> >>> + *pprog = &progs->stream_verdict;
> >>> break;
> >>> case BPF_SK_SKB_VERDICT:
> >>> if (progs->stream_verdict)
> >>> return -EBUSY;
> >>> - pprog = &progs->skb_verdict;
> >>> + *pprog = &progs->skb_verdict;
> >>> break;
> >>> default:
> >>> return -EOPNOTSUPP;
> >>> }
> >>>
> >>> + return 0;
> >>> +}
> >>> +
> >>> +static int sock_map_prog_update(struct bpf_map *map, struct bpf_prog
> >> *prog,
> >>> + struct bpf_prog *old, u32 which)
> >>> +{
> >>> + struct bpf_prog **pprog;
> >>> + int ret;
> >>> +
> >>> + ret = sock_map_prog_lookup(map, &pprog, which);
> >>> + if (ret)
> >>> + return ret;
> >>> +
> >>> if (old)
> >>> return psock_replace_prog(pprog, prog, old);
> >>>
> >>> @@ -1451,6 +1463,54 @@ static int sock_map_prog_update(struct
> bpf_map
> >> *map, struct bpf_prog *prog,
> >>> return 0;
> >>> }
> >>>
> >>> +int sock_map_bpf_prog_query(const union bpf_attr *attr,
> >>> + union bpf_attr __user *uattr)
> >>> +{
> >>> + __u32 __user *prog_ids = u64_to_user_ptr(attr->query.prog_ids);
> >>> + u32 prog_cnt = 0, flags = 0, ufd = attr->target_fd;
> >>> + struct bpf_prog **pprog;
> >>> + struct bpf_prog *prog;
> >>> + struct bpf_map *map;
> >>> + struct fd f;
> >>> + u32 id = 0;
> >>
> >> There is no need to initialize 'id = 0'. id will be assigned later.
> >
> >
> > if (!attr->query.prog_cnt || !prog_ids || !prog_cnt) is met, the id will not be
> assigned later.
> > At the end of the function, we judge whether to copy the program ID by the
> value of the id.
>
> Ya, that is true. id = 0 is indeed needed.
>
> >
> >
> >>
> >>> + int ret;
> >>> +
> >>> + if (attr->query.query_flags)
> >>> + return -EINVAL;
> >>> +
> >>> + f = fdget(ufd);
> >>> + map = __bpf_map_get(f);
> >>> + if (IS_ERR(map))
> >>> + return PTR_ERR(map);
> >>> +
> >>> + rcu_read_lock();
> >>> +
> >>> + ret = sock_map_prog_lookup(map, &pprog, attr->query.attach_type);
> >>> + if (ret)
> >>> + goto end;
> >>> +
> >>> + prog = *pprog;
> >>> + prog_cnt = (!prog) ? 0 : 1;
> >>
> >> (!prog) => !prog ?
> >
> > Yes, It's just my habit
> >
> >>
> >>> +
> >>> + if (!attr->query.prog_cnt || !prog_ids || !prog_cnt)
> >>> + goto end;
> >>> +
> >>> + id = prog->aux->id;
> >>> + if (id == 0)
> >>> + prog_cnt = 0;
> >>
> >>
> >> id will never be 0, see function bpf_prog_alloc_id() is syscall.c.
> >> So 'if (id == 0)' check is not needed.
> >
> >
> > Because we do not hold the reference count, the program may be released
> synchronously
>
> synchronously => asynchronously
>
> > and its ID will be set to 0 in bpf_prog_free_id().
> >
> > Is that right?
>
> Just checked the code again. You are right. Maybe add a comment here to
> make it clear why we check id == 0 here?


Yes, I'll add a comment


> >
> >
> >>
> >>> +
> >>> +end:
> >>> + rcu_read_unlock();
> >>> +
> >>> + if (copy_to_user(&uattr->query.attach_flags, &flags, sizeof(flags)) ||
> >>> + (id != 0 && copy_to_user(prog_ids, &id, sizeof(u32))) ||
> >>> + copy_to_user(&uattr->query.prog_cnt, &prog_cnt,
> sizeof(prog_cnt)))
> >>> + ret = -EFAULT;
> >>> +
> >>> + fdput(f);
> >>> + return ret;
> >>> +}
> >>> +
> >>> static void sock_map_unlink(struct sock *sk, struct sk_psock_link *link)
> >>> {
> >>> switch (link->map->map_type) {
> >>>