From: Martin KaFai Lau <martin.lau@linux.dev>
To: Kui-Feng Lee <kuifeng@meta.com>
Cc: bpf@vger.kernel.org, ast@kernel.org, song@kernel.org,
kernel-team@meta.com, andrii@kernel.org, sdf@google.com
Subject: Re: [PATCH bpf-next v3 2/8] bpf: Create links for BPF struct_ops maps.
Date: Mon, 6 Mar 2023 18:11:46 -0800 [thread overview]
Message-ID: <1e7a3b7b-7693-14d9-9eb4-7a516badba95@linux.dev> (raw)
In-Reply-To: <20230303012122.852654-3-kuifeng@meta.com>
On 3/2/23 5:21 PM, Kui-Feng Lee wrote:
> diff --git a/include/linux/bpf.h b/include/linux/bpf.h
> index cb837f42b99d..b845be719422 100644
> --- a/include/linux/bpf.h
> +++ b/include/linux/bpf.h
> @@ -1396,6 +1396,11 @@ struct bpf_link {
> struct work_struct work;
> };
>
> +struct bpf_struct_ops_link {
> + struct bpf_link link;
> + struct bpf_map __rcu *map;
__rcu is only needed after the link_update change in patch 5?
It is fine to keep it in this patch but please leave a comment in the commit
message.
Does 'struct bpf_struct_ops_link' have to be in bpf.h?
> +};
> +
> struct bpf_link_ops {
> void (*release)(struct bpf_link *link);
> void (*dealloc)(struct bpf_link *link);
> @@ -1964,6 +1969,7 @@ int bpf_link_new_fd(struct bpf_link *link);
> struct file *bpf_link_new_file(struct bpf_link *link, int *reserved_fd);
> struct bpf_link *bpf_link_get_from_fd(u32 ufd);
> struct bpf_link *bpf_link_get_curr_or_next(u32 *id);
> +int bpf_struct_ops_link_create(union bpf_attr *attr);
>
> int bpf_obj_pin_user(u32 ufd, const char __user *pathname);
> int bpf_obj_get_user(const char __user *pathname, int flags);
> @@ -2308,6 +2314,11 @@ static inline void bpf_link_put(struct bpf_link *link)
> {
> }
>
> +static inline int bpf_struct_ops_link_create(union bpf_attr *attr)
> +{
> + return -EOPNOTSUPP;
> +}
Is this currently under '#ifdef CONFIG_BPF_SYSCALL' alone?
Not sure if it is correct. Please double check.
ifeq ($(CONFIG_BPF_JIT),y)
obj-$(CONFIG_BPF_SYSCALL) += bpf_struct_ops.o
obj-$(CONFIG_BPF_SYSCALL) += cpumask.o
obj-${CONFIG_BPF_LSM} += bpf_lsm.o
endif
obj-$(CONFIG_BPF_SYSCALL) += syscall.o ...
> +
> static inline int bpf_obj_get_user(const char __user *pathname, int flags)
> {
> return -EOPNOTSUPP;
> diff --git a/include/uapi/linux/bpf.h b/include/uapi/linux/bpf.h
> index 17afd2b35ee5..cd0ff39981e8 100644
> --- a/include/uapi/linux/bpf.h
> +++ b/include/uapi/linux/bpf.h
> @@ -1033,6 +1033,7 @@ enum bpf_attach_type {
> BPF_PERF_EVENT,
> BPF_TRACE_KPROBE_MULTI,
> BPF_LSM_CGROUP,
> + BPF_STRUCT_OPS,
> __MAX_BPF_ATTACH_TYPE
> };
>
> @@ -1266,6 +1267,9 @@ enum {
>
> /* Create a map that is suitable to be an inner map with dynamic max entries */
> BPF_F_INNER_MAP = (1U << 12),
> +
> +/* Create a map that will be registered/unregesitered by the backed bpf_link */
> + BPF_F_LINK = (1U << 13),
> };
>
> /* Flags for BPF_PROG_QUERY. */
> @@ -1507,7 +1511,10 @@ union bpf_attr {
> } task_fd_query;
>
> struct { /* struct used by BPF_LINK_CREATE command */
> - __u32 prog_fd; /* eBPF program to attach */
> + union {
> + __u32 prog_fd; /* eBPF program to attach */
> + __u32 map_fd; /* eBPF struct_ops to attach */
nit. Remove eBPF. "struct_ops to attach"
> + };
> union {
> __u32 target_fd; /* object to attach to */
> __u32 target_ifindex; /* target ifindex */
> @@ -6354,6 +6361,9 @@ struct bpf_link_info {
> struct {
> __u32 ifindex;
> } xdp;
> + struct {
> + __u32 map_id;
> + } struct_ops;
> };
> } __attribute__((aligned(8)));
>
> diff --git a/kernel/bpf/bpf_struct_ops.c b/kernel/bpf/bpf_struct_ops.c
> index bba03b6b010b..9ec675576d97 100644
> --- a/kernel/bpf/bpf_struct_ops.c
> +++ b/kernel/bpf/bpf_struct_ops.c
> @@ -14,6 +14,7 @@
>
> enum bpf_struct_ops_state {
> BPF_STRUCT_OPS_STATE_INIT,
> + BPF_STRUCT_OPS_STATE_READY,
Please add it to the end. Although it is not in uapi, try not to disrupt the
userspace introspection tool if it does not have to.
> BPF_STRUCT_OPS_STATE_INUSE,
> BPF_STRUCT_OPS_STATE_TOBEFREE,
> };
> @@ -494,11 +495,19 @@ static int bpf_struct_ops_map_update_elem(struct bpf_map *map, void *key,
> *(unsigned long *)(udata + moff) = prog->aux->id;
> }
>
> - bpf_map_inc(map);
> -
> set_memory_rox((long)st_map->image, 1);
> + if (st_map->map.map_flags & BPF_F_LINK) {
> + /* Let bpf_link handle registration & unregistration.
> + *
> + * Pair with smp_load_acquire() during lookup_elem().
> + */
> + smp_store_release(&kvalue->state, BPF_STRUCT_OPS_STATE_READY);
> + goto unlock;
> + }
> +
> err = st_ops->reg(kdata);
> if (likely(!err)) {
> + bpf_map_inc(map);
> /* Pair with smp_load_acquire() during lookup_elem().
> * It ensures the above udata updates (e.g. prog->aux->id)
> * can be seen once BPF_STRUCT_OPS_STATE_INUSE is set.
> @@ -514,7 +523,6 @@ static int bpf_struct_ops_map_update_elem(struct bpf_map *map, void *key,
> */
> set_memory_nx((long)st_map->image, 1);
> set_memory_rw((long)st_map->image, 1);
> - bpf_map_put(map);
>
> reset_unlock:
> bpf_struct_ops_map_put_progs(st_map);
> @@ -532,10 +540,15 @@ static int bpf_struct_ops_map_delete_elem(struct bpf_map *map, void *key)
> struct bpf_struct_ops_map *st_map;
>
> st_map = (struct bpf_struct_ops_map *)map;
> + if (st_map->map.map_flags & BPF_F_LINK)
> + return -EOPNOTSUPP;
> +
> prev_state = cmpxchg(&st_map->kvalue.state,
> BPF_STRUCT_OPS_STATE_INUSE,
> BPF_STRUCT_OPS_STATE_TOBEFREE);
> switch (prev_state) {
> + case BPF_STRUCT_OPS_STATE_READY:
> + return -EOPNOTSUPP;
If this case never happens, this case should be removed. The WARN in the default
case at the end is a better handling.
> case BPF_STRUCT_OPS_STATE_INUSE:
> st_map->st_ops->unreg(&st_map->kvalue.data);
> bpf_map_put(map);
> @@ -618,7 +631,7 @@ static void bpf_struct_ops_map_free_rcu(struct bpf_map *map)
> static int bpf_struct_ops_map_alloc_check(union bpf_attr *attr)
> {
> if (attr->key_size != sizeof(unsigned int) || attr->max_entries != 1 ||
> - attr->map_flags || !attr->btf_vmlinux_value_type_id)
> + (attr->map_flags & ~BPF_F_LINK) || !attr->btf_vmlinux_value_type_id)
> return -EINVAL;
> return 0;
> }
> @@ -714,3 +727,100 @@ void bpf_struct_ops_put(const void *kdata)
>
> bpf_map_put(&st_map->map);
> }
> +
> +static void bpf_struct_ops_map_link_dealloc(struct bpf_link *link)
> +{
> + struct bpf_struct_ops_link *st_link;
> + struct bpf_struct_ops_map *st_map;
> +
> + st_link = container_of(link, struct bpf_struct_ops_link, link);
> + if (st_link->map) {
Will map ever be NULL
> + st_map = (struct bpf_struct_ops_map *)st_link->map;
> + st_map->st_ops->unreg(&st_map->kvalue.data);
> + bpf_map_put(st_link->map);
> + }
> + kfree(st_link);
> +}
> +
> +static void bpf_struct_ops_map_link_show_fdinfo(const struct bpf_link *link,
> + struct seq_file *seq)
> +{
> + struct bpf_struct_ops_link *st_link;
> + struct bpf_map *map;
> +
> + st_link = container_of(link, struct bpf_struct_ops_link, link);
> + rcu_read_lock_trace();
Should it be rcu_read_lock()?
> + map = rcu_dereference(st_link->map);
> + if (map)
> + seq_printf(seq, "map_id:\t%d\n", map->id);
> + rcu_read_unlock_trace();
> +}
> +
> +static int bpf_struct_ops_map_link_fill_link_info(const struct bpf_link *link,
> + struct bpf_link_info *info)
> +{
> + struct bpf_struct_ops_link *st_link;
> + struct bpf_map *map;
> +
> + st_link = container_of(link, struct bpf_struct_ops_link, link);
> + rcu_read_lock_trace();
> + map = rcu_dereference(st_link->map);
> + if (map)
> + info->struct_ops.map_id = map->id;
> + rcu_read_unlock_trace();
> + return 0;
> +}
> +
> +static const struct bpf_link_ops bpf_struct_ops_map_lops = {
> + .dealloc = bpf_struct_ops_map_link_dealloc,
> + .show_fdinfo = bpf_struct_ops_map_link_show_fdinfo,
> + .fill_link_info = bpf_struct_ops_map_link_fill_link_info,
> +};
> +
> +int bpf_struct_ops_link_create(union bpf_attr *attr)
> +{
> + struct bpf_struct_ops_link *link = NULL;
> + struct bpf_link_primer link_primer;
> + struct bpf_struct_ops_map *st_map;
> + struct bpf_map *map;
> + int err;
> +
> + map = bpf_map_get(attr->link_create.map_fd);
> + if (!map)
> + return -EINVAL;
> +
> + st_map = (struct bpf_struct_ops_map *)map;
> +
> + if (map->map_type != BPF_MAP_TYPE_STRUCT_OPS || !(map->map_flags & BPF_F_LINK) ||
> + /* Pair with smp_store_release() during map_update */
> + smp_load_acquire(&st_map->kvalue.state) != BPF_STRUCT_OPS_STATE_READY) {
> + err = -EINVAL;
> + goto err_out;
> + }
> +
> + link = kzalloc(sizeof(*link), GFP_USER);
> + if (!link) {
> + err = -ENOMEM;
> + goto err_out;
> + }
> + bpf_link_init(&link->link, BPF_LINK_TYPE_STRUCT_OPS, &bpf_struct_ops_map_lops, NULL);
> + link->map = map;
RCU_INIT_POINTER().
> +
> + err = bpf_link_prime(&link->link, &link_primer);
> + if (err)
> + goto err_out;
> +
> + err = st_map->st_ops->reg(st_map->kvalue.data);
> + if (err) {
> + bpf_link_cleanup(&link_primer);
> + goto err_out;
> + }
> +
> + return bpf_link_settle(&link_primer);
> +
> +err_out:
> + bpf_map_put(map);
> + kfree(link);
> + return err;
> +}
> +
> diff --git a/kernel/bpf/syscall.c b/kernel/bpf/syscall.c
> index 358a0e40555e..3db4938212d6 100644
> --- a/kernel/bpf/syscall.c
> +++ b/kernel/bpf/syscall.c
> @@ -2743,10 +2743,11 @@ void bpf_link_inc(struct bpf_link *link)
> static void bpf_link_free(struct bpf_link *link)
> {
> bpf_link_free_id(link->id);
> + /* detach BPF program, clean up used resources */
> if (link->prog) {
> - /* detach BPF program, clean up used resources */
This comment move seems unnecessary.
> link->ops->release(link);
> bpf_prog_put(link->prog);
> + /* The struct_ops links clean up map by them-selves. */
This also seems unnecessary to only spell out for struct_ops link. Each specific
link type does its cleanup in ->dealloc.
next prev parent reply other threads:[~2023-03-07 2:12 UTC|newest]
Thread overview: 20+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-03-03 1:21 [PATCH bpf-next v3 0/8] Transit between BPF TCP congestion controls Kui-Feng Lee
2023-03-03 1:21 ` [PATCH bpf-next v3 1/8] bpf: Maintain the refcount of struct_ops maps directly Kui-Feng Lee
2023-03-06 20:10 ` Stanislav Fomichev
2023-03-06 21:45 ` Kui-Feng Lee
2023-03-06 23:16 ` Martin KaFai Lau
2023-03-06 23:54 ` Kui-Feng Lee
2023-03-07 0:36 ` Martin KaFai Lau
2023-03-03 1:21 ` [PATCH bpf-next v3 2/8] bpf: Create links for BPF struct_ops maps Kui-Feng Lee
2023-03-06 20:23 ` Stanislav Fomichev
2023-03-06 22:02 ` Kui-Feng Lee
2023-03-07 2:11 ` Martin KaFai Lau [this message]
2023-03-07 18:04 ` Kui-Feng Lee
2023-03-03 1:21 ` [PATCH bpf-next v3 3/8] net: Update an existing TCP congestion control algorithm Kui-Feng Lee
2023-03-07 2:17 ` Martin KaFai Lau
2023-03-07 19:17 ` Kui-Feng Lee
2023-03-03 1:21 ` [PATCH bpf-next v3 4/8] libbpf: Create a bpf_link in bpf_map__attach_struct_ops() Kui-Feng Lee
2023-03-03 1:21 ` [PATCH bpf-next v3 5/8] bpf: Update the struct_ops of a bpf_link Kui-Feng Lee
2023-03-03 1:21 ` [PATCH bpf-next v3 6/8] libbpf: Update a bpf_link with another struct_ops Kui-Feng Lee
2023-03-03 1:21 ` [PATCH bpf-next v3 7/8] libbpf: Use .struct_ops.link section to indicate a struct_ops with a link Kui-Feng Lee
2023-03-03 1:21 ` [PATCH bpf-next v3 8/8] selftests/bpf: Test switching TCP Congestion Control algorithms Kui-Feng Lee
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1e7a3b7b-7693-14d9-9eb4-7a516badba95@linux.dev \
--to=martin.lau@linux.dev \
--cc=andrii@kernel.org \
--cc=ast@kernel.org \
--cc=bpf@vger.kernel.org \
--cc=kernel-team@meta.com \
--cc=kuifeng@meta.com \
--cc=sdf@google.com \
--cc=song@kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox