linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH v2] libbpf: fix BPF_MAP_TYPE_PERF_EVENT_ARRAY auto-pinning
@ 2022-02-25 15:23 Stijn Tintel
  2022-02-26  6:11 ` Song Liu
  2022-02-28 16:40 ` patchwork-bot+netdevbpf
  0 siblings, 2 replies; 3+ messages in thread
From: Stijn Tintel @ 2022-02-25 15:23 UTC (permalink / raw)
  To: bpf, songliubraving, andrii, toke
  Cc: netdev, linux-kernel, kpsingh, yhs, kafai, daniel, ast

When a BPF map of type BPF_MAP_TYPE_PERF_EVENT_ARRAY doesn't have the
max_entries parameter set, the map will be created with max_entries set
to the number of available CPUs. When we try to reuse such a pinned map,
map_is_reuse_compat will return false, as max_entries in the map
definition differs from max_entries of the existing map, causing the
following error:

libbpf: couldn't reuse pinned map at '/sys/fs/bpf/m_logging': parameter mismatch

Fix this by overwriting max_entries in the map definition. For this to
work, we need to do this in bpf_object__create_maps, before calling
bpf_object__reuse_map.

Fixes: 57a00f41644f ("libbpf: Add auto-pinning of maps when loading BPF objects")
Signed-off-by: Stijn Tintel <stijn@linux-ipv6.be>
---
v2: overwrite max_entries in the map definition instead of adding an
    extra check in map_is_reuse_compat, and introduce a helper function
    for this as suggested by Song.
---
 tools/lib/bpf/libbpf.c | 44 ++++++++++++++++++++++++------------------
 1 file changed, 25 insertions(+), 19 deletions(-)

diff --git a/tools/lib/bpf/libbpf.c b/tools/lib/bpf/libbpf.c
index 7f10dd501a52..133462637b09 100644
--- a/tools/lib/bpf/libbpf.c
+++ b/tools/lib/bpf/libbpf.c
@@ -4854,7 +4854,6 @@ static int bpf_object__create_map(struct bpf_object *obj, struct bpf_map *map, b
 	LIBBPF_OPTS(bpf_map_create_opts, create_attr);
 	struct bpf_map_def *def = &map->def;
 	const char *map_name = NULL;
-	__u32 max_entries;
 	int err = 0;
 
 	if (kernel_supports(obj, FEAT_PROG_NAME))
@@ -4864,21 +4863,6 @@ static int bpf_object__create_map(struct bpf_object *obj, struct bpf_map *map, b
 	create_attr.numa_node = map->numa_node;
 	create_attr.map_extra = map->map_extra;
 
-	if (def->type == BPF_MAP_TYPE_PERF_EVENT_ARRAY && !def->max_entries) {
-		int nr_cpus;
-
-		nr_cpus = libbpf_num_possible_cpus();
-		if (nr_cpus < 0) {
-			pr_warn("map '%s': failed to determine number of system CPUs: %d\n",
-				map->name, nr_cpus);
-			return nr_cpus;
-		}
-		pr_debug("map '%s': setting size to %d\n", map->name, nr_cpus);
-		max_entries = nr_cpus;
-	} else {
-		max_entries = def->max_entries;
-	}
-
 	if (bpf_map__is_struct_ops(map))
 		create_attr.btf_vmlinux_value_type_id = map->btf_vmlinux_value_type_id;
 
@@ -4928,7 +4912,7 @@ static int bpf_object__create_map(struct bpf_object *obj, struct bpf_map *map, b
 
 	if (obj->gen_loader) {
 		bpf_gen__map_create(obj->gen_loader, def->type, map_name,
-				    def->key_size, def->value_size, max_entries,
+				    def->key_size, def->value_size, def->max_entries,
 				    &create_attr, is_inner ? -1 : map - obj->maps);
 		/* Pretend to have valid FD to pass various fd >= 0 checks.
 		 * This fd == 0 will not be used with any syscall and will be reset to -1 eventually.
@@ -4937,7 +4921,7 @@ static int bpf_object__create_map(struct bpf_object *obj, struct bpf_map *map, b
 	} else {
 		map->fd = bpf_map_create(def->type, map_name,
 					 def->key_size, def->value_size,
-					 max_entries, &create_attr);
+					 def->max_entries, &create_attr);
 	}
 	if (map->fd < 0 && (create_attr.btf_key_type_id ||
 			    create_attr.btf_value_type_id)) {
@@ -4954,7 +4938,7 @@ static int bpf_object__create_map(struct bpf_object *obj, struct bpf_map *map, b
 		map->btf_value_type_id = 0;
 		map->fd = bpf_map_create(def->type, map_name,
 					 def->key_size, def->value_size,
-					 max_entries, &create_attr);
+					 def->max_entries, &create_attr);
 	}
 
 	err = map->fd < 0 ? -errno : 0;
@@ -5058,6 +5042,24 @@ static int bpf_object_init_prog_arrays(struct bpf_object *obj)
 	return 0;
 }
 
+static int map_set_def_max_entries(struct bpf_map *map)
+{
+	if (map->def.type == BPF_MAP_TYPE_PERF_EVENT_ARRAY && !map->def.max_entries) {
+		int nr_cpus;
+
+		nr_cpus = libbpf_num_possible_cpus();
+		if (nr_cpus < 0) {
+			pr_warn("map '%s': failed to determine number of system CPUs: %d\n",
+				map->name, nr_cpus);
+			return nr_cpus;
+		}
+		pr_debug("map '%s': setting size to %d\n", map->name, nr_cpus);
+		map->def.max_entries = nr_cpus;
+	}
+
+	return 0;
+}
+
 static int
 bpf_object__create_maps(struct bpf_object *obj)
 {
@@ -5090,6 +5092,10 @@ bpf_object__create_maps(struct bpf_object *obj)
 			continue;
 		}
 
+		err = map_set_def_max_entries(map);
+		if (err)
+			goto err_out;
+
 		retried = false;
 retry:
 		if (map->pin_path) {
-- 
2.34.1


^ permalink raw reply related	[flat|nested] 3+ messages in thread
* Re: [PATCH v2] libbpf: fix BPF_MAP_TYPE_PERF_EVENT_ARRAY auto-pinning
  2022-02-25 15:23 [PATCH v2] libbpf: fix BPF_MAP_TYPE_PERF_EVENT_ARRAY auto-pinning Stijn Tintel
@ 2022-02-26  6:11 ` Song Liu
  2022-02-28 16:40 ` patchwork-bot+netdevbpf
  1 sibling, 0 replies; 3+ messages in thread
From: Song Liu @ 2022-02-26  6:11 UTC (permalink / raw)
  To: Stijn Tintel
  Cc: bpf, Song Liu, Andrii Nakryiko, Toke Høiland-Jørgensen,
	Networking, open list, KP Singh, Yonghong Song, Martin KaFai Lau,
	Daniel Borkmann, Alexei Starovoitov

On Fri, Feb 25, 2022 at 7:24 AM Stijn Tintel <stijn@linux-ipv6.be> wrote:
>
> When a BPF map of type BPF_MAP_TYPE_PERF_EVENT_ARRAY doesn't have the
> max_entries parameter set, the map will be created with max_entries set
> to the number of available CPUs. When we try to reuse such a pinned map,
> map_is_reuse_compat will return false, as max_entries in the map
> definition differs from max_entries of the existing map, causing the
> following error:
>
> libbpf: couldn't reuse pinned map at '/sys/fs/bpf/m_logging': parameter mismatch
>
> Fix this by overwriting max_entries in the map definition. For this to
> work, we need to do this in bpf_object__create_maps, before calling
> bpf_object__reuse_map.
>
> Fixes: 57a00f41644f ("libbpf: Add auto-pinning of maps when loading BPF objects")
> Signed-off-by: Stijn Tintel <stijn@linux-ipv6.be>

Acked-by: Song Liu <songliubraving@fb.com>

> ---
> v2: overwrite max_entries in the map definition instead of adding an
>     extra check in map_is_reuse_compat, and introduce a helper function
>     for this as suggested by Song.
> ---
>  tools/lib/bpf/libbpf.c | 44 ++++++++++++++++++++++++------------------
>  1 file changed, 25 insertions(+), 19 deletions(-)
>
> diff --git a/tools/lib/bpf/libbpf.c b/tools/lib/bpf/libbpf.c
> index 7f10dd501a52..133462637b09 100644
> --- a/tools/lib/bpf/libbpf.c
> +++ b/tools/lib/bpf/libbpf.c
> @@ -4854,7 +4854,6 @@ static int bpf_object__create_map(struct bpf_object *obj, struct bpf_map *map, b
>         LIBBPF_OPTS(bpf_map_create_opts, create_attr);
>         struct bpf_map_def *def = &map->def;
>         const char *map_name = NULL;
> -       __u32 max_entries;
>         int err = 0;
>
>         if (kernel_supports(obj, FEAT_PROG_NAME))
> @@ -4864,21 +4863,6 @@ static int bpf_object__create_map(struct bpf_object *obj, struct bpf_map *map, b
>         create_attr.numa_node = map->numa_node;
>         create_attr.map_extra = map->map_extra;
>
> -       if (def->type == BPF_MAP_TYPE_PERF_EVENT_ARRAY && !def->max_entries) {
> -               int nr_cpus;
> -
> -               nr_cpus = libbpf_num_possible_cpus();
> -               if (nr_cpus < 0) {
> -                       pr_warn("map '%s': failed to determine number of system CPUs: %d\n",
> -                               map->name, nr_cpus);
> -                       return nr_cpus;
> -               }
> -               pr_debug("map '%s': setting size to %d\n", map->name, nr_cpus);
> -               max_entries = nr_cpus;
> -       } else {
> -               max_entries = def->max_entries;
> -       }
> -
>         if (bpf_map__is_struct_ops(map))
>                 create_attr.btf_vmlinux_value_type_id = map->btf_vmlinux_value_type_id;
>
> @@ -4928,7 +4912,7 @@ static int bpf_object__create_map(struct bpf_object *obj, struct bpf_map *map, b
>
>         if (obj->gen_loader) {
>                 bpf_gen__map_create(obj->gen_loader, def->type, map_name,
> -                                   def->key_size, def->value_size, max_entries,
> +                                   def->key_size, def->value_size, def->max_entries,
>                                     &create_attr, is_inner ? -1 : map - obj->maps);
>                 /* Pretend to have valid FD to pass various fd >= 0 checks.
>                  * This fd == 0 will not be used with any syscall and will be reset to -1 eventually.
> @@ -4937,7 +4921,7 @@ static int bpf_object__create_map(struct bpf_object *obj, struct bpf_map *map, b
>         } else {
>                 map->fd = bpf_map_create(def->type, map_name,
>                                          def->key_size, def->value_size,
> -                                        max_entries, &create_attr);
> +                                        def->max_entries, &create_attr);
>         }
>         if (map->fd < 0 && (create_attr.btf_key_type_id ||
>                             create_attr.btf_value_type_id)) {
> @@ -4954,7 +4938,7 @@ static int bpf_object__create_map(struct bpf_object *obj, struct bpf_map *map, b
>                 map->btf_value_type_id = 0;
>                 map->fd = bpf_map_create(def->type, map_name,
>                                          def->key_size, def->value_size,
> -                                        max_entries, &create_attr);
> +                                        def->max_entries, &create_attr);
>         }
>
>         err = map->fd < 0 ? -errno : 0;
> @@ -5058,6 +5042,24 @@ static int bpf_object_init_prog_arrays(struct bpf_object *obj)
>         return 0;
>  }
>
> +static int map_set_def_max_entries(struct bpf_map *map)
> +{
> +       if (map->def.type == BPF_MAP_TYPE_PERF_EVENT_ARRAY && !map->def.max_entries) {
> +               int nr_cpus;
> +
> +               nr_cpus = libbpf_num_possible_cpus();
> +               if (nr_cpus < 0) {
> +                       pr_warn("map '%s': failed to determine number of system CPUs: %d\n",
> +                               map->name, nr_cpus);
> +                       return nr_cpus;
> +               }
> +               pr_debug("map '%s': setting size to %d\n", map->name, nr_cpus);
> +               map->def.max_entries = nr_cpus;
> +       }
> +
> +       return 0;
> +}
> +
>  static int
>  bpf_object__create_maps(struct bpf_object *obj)
>  {
> @@ -5090,6 +5092,10 @@ bpf_object__create_maps(struct bpf_object *obj)
>                         continue;
>                 }
>
> +               err = map_set_def_max_entries(map);
> +               if (err)
> +                       goto err_out;
> +
>                 retried = false;
>  retry:
>                 if (map->pin_path) {
> --
> 2.34.1
>

^ permalink raw reply	[flat|nested] 3+ messages in thread
* Re: [PATCH v2] libbpf: fix BPF_MAP_TYPE_PERF_EVENT_ARRAY auto-pinning
  2022-02-25 15:23 [PATCH v2] libbpf: fix BPF_MAP_TYPE_PERF_EVENT_ARRAY auto-pinning Stijn Tintel
  2022-02-26  6:11 ` Song Liu
@ 2022-02-28 16:40 ` patchwork-bot+netdevbpf
  1 sibling, 0 replies; 3+ messages in thread
From: patchwork-bot+netdevbpf @ 2022-02-28 16:40 UTC (permalink / raw)
  To: Stijn Tintel
  Cc: bpf, songliubraving, andrii, toke, netdev, linux-kernel, kpsingh,
	yhs, kafai, daniel, ast

Hello:

This patch was applied to bpf/bpf-next.git (master)
by Daniel Borkmann <daniel@iogearbox.net>:

On Fri, 25 Feb 2022 17:23:55 +0200 you wrote:
> When a BPF map of type BPF_MAP_TYPE_PERF_EVENT_ARRAY doesn't have the
> max_entries parameter set, the map will be created with max_entries set
> to the number of available CPUs. When we try to reuse such a pinned map,
> map_is_reuse_compat will return false, as max_entries in the map
> definition differs from max_entries of the existing map, causing the
> following error:
> 
> [...]

Here is the summary with links:
  - [v2] libbpf: fix BPF_MAP_TYPE_PERF_EVENT_ARRAY auto-pinning
    https://git.kernel.org/bpf/bpf-next/c/a4fbfdd7a160

You are awesome, thank you!
-- 
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html



^ permalink raw reply	[flat|nested] 3+ messages in thread
end of thread, other threads:[~2022-02-28 16:40 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2022-02-25 15:23 [PATCH v2] libbpf: fix BPF_MAP_TYPE_PERF_EVENT_ARRAY auto-pinning Stijn Tintel
2022-02-26  6:11 ` Song Liu
2022-02-28 16:40 ` patchwork-bot+netdevbpf

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).