netdev.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH bpf-next] libbpf: fix error when st-prefix_ops and ops from differ btf
@ 2025-09-26  7:17 D. Wythe
  2025-09-26 16:53 ` Martin KaFai Lau
  0 siblings, 1 reply; 3+ messages in thread
From: D. Wythe @ 2025-09-26  7:17 UTC (permalink / raw)
  To: ast, daniel, andrii, martin.lau, pabeni, song, sdf, haoluo, yhs,
	edumazet, john.fastabend, kpsingh, jolsa, mjambigi, wenjia,
	wintera, dust.li, tonylu, guwen
  Cc: bpf, davem, kuba, netdev, sidraya, jaka

When a module registers a struct_ops, the struct_ops type and its
corresponding map_value type ("bpf_struct_ops_") may reside in different
btf objects, here are four possible case:

+--------+---------------+-------------+---------------------------------+
|        |bpf_struct_ops_| xxx_ops     |                                 |
+--------+---------------+-------------+---------------------------------+
| case 0 | btf_vmlinux   | bft_vmlinux | be used and reg only in vmlinux |
+--------+---------------+-------------+---------------------------------+
| case 1 | btf_vmlinux   | mod_btf     | INVALID                         |
+--------+---------------+-------------+---------------------------------+
| case 2 | mod_btf       | btf_vmlinux | reg in mod but be used both in  |
|        |               |             | vmlinux and mod.                |
+--------+---------------+-------------+---------------------------------+
| case 3 | mod_btf       | mod_btf     | be used and reg only in mod     |
+--------+---------------+-------------+---------------------------------+

Currently we figure out the mod_btf by searching with the struct_ops type,
which makes it impossible to figure out the mod_btf when the struct_ops
type is in btf_vmlinux while it's corresponding map_value type is in
mod_btf (case 2).

The fix is to use the corresponding map_value type ("bpf_struct_ops_")
as the lookup anchor instead of the struct_ops type to figure out the
`btf` and `mod_btf` via find_ksym_btf_id(), and then we can locate
the kern_type_id via btf__find_by_name_kind() with the `btf` we just
obtained from find_ksym_btf_id().

With this change the lookup obtains the correct btf and mod_btf for case 2,
preserves correct behavior for other valid cases, and still fails as
expected for the invalid scenario (case 1).

Fixes: 590a00888250 ("bpf: libbpf: Add STRUCT_OPS support")
Signed-off-by: D. Wythe <alibuda@linux.alibaba.com>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
---
 tools/lib/bpf/libbpf.c | 37 ++++++++++++++++++-------------------
 1 file changed, 18 insertions(+), 19 deletions(-)

diff --git a/tools/lib/bpf/libbpf.c b/tools/lib/bpf/libbpf.c
index 5161c2b39875..a93eed660404 100644
--- a/tools/lib/bpf/libbpf.c
+++ b/tools/lib/bpf/libbpf.c
@@ -1018,35 +1018,34 @@ find_struct_ops_kern_types(struct bpf_object *obj, const char *tname_raw,
 	const struct btf_member *kern_data_member;
 	struct btf *btf = NULL;
 	__s32 kern_vtype_id, kern_type_id;
-	char tname[256];
+	char tname[256], stname[256];
 	__u32 i;
 
 	snprintf(tname, sizeof(tname), "%.*s",
 		 (int)bpf_core_essential_name_len(tname_raw), tname_raw);
 
-	kern_type_id = find_ksym_btf_id(obj, tname, BTF_KIND_STRUCT,
-					&btf, mod_btf);
-	if (kern_type_id < 0) {
-		pr_warn("struct_ops init_kern: struct %s is not found in kernel BTF\n",
-			tname);
-		return kern_type_id;
-	}
-	kern_type = btf__type_by_id(btf, kern_type_id);
+	snprintf(stname, sizeof(stname), "%s%.*s", STRUCT_OPS_VALUE_PREFIX,
+		 (int)strlen(tname), tname);
 
-	/* Find the corresponding "map_value" type that will be used
-	 * in map_update(BPF_MAP_TYPE_STRUCT_OPS).  For example,
-	 * find "struct bpf_struct_ops_tcp_congestion_ops" from the
-	 * btf_vmlinux.
+	/* Look for the corresponding "map_value" type that will be used
+	 * in map_update(BPF_MAP_TYPE_STRUCT_OPS) first, figure out the btf
+	 * and the mod_btf.
+	 * For example, find "struct bpf_struct_ops_tcp_congestion_ops".
 	 */
-	kern_vtype_id = find_btf_by_prefix_kind(btf, STRUCT_OPS_VALUE_PREFIX,
-						tname, BTF_KIND_STRUCT);
+	kern_vtype_id = find_ksym_btf_id(obj, stname, BTF_KIND_STRUCT, &btf, mod_btf);
 	if (kern_vtype_id < 0) {
-		pr_warn("struct_ops init_kern: struct %s%s is not found in kernel BTF\n",
-			STRUCT_OPS_VALUE_PREFIX, tname);
+		pr_warn("struct_ops init_kern: struct %s is not found in kernel BTF\n", stname);
 		return kern_vtype_id;
 	}
 	kern_vtype = btf__type_by_id(btf, kern_vtype_id);
 
+	kern_type_id = btf__find_by_name_kind(btf, tname, BTF_KIND_STRUCT);
+	if (kern_type_id < 0) {
+		pr_warn("struct_ops init_kern: struct %s is not found in kernel BTF\n", tname);
+		return kern_type_id;
+	}
+	kern_type = btf__type_by_id(btf, kern_type_id);
+
 	/* Find "struct tcp_congestion_ops" from
 	 * struct bpf_struct_ops_tcp_congestion_ops {
 	 *	[ ... ]
@@ -1059,8 +1058,8 @@ find_struct_ops_kern_types(struct bpf_object *obj, const char *tname_raw,
 			break;
 	}
 	if (i == btf_vlen(kern_vtype)) {
-		pr_warn("struct_ops init_kern: struct %s data is not found in struct %s%s\n",
-			tname, STRUCT_OPS_VALUE_PREFIX, tname);
+		pr_warn("struct_ops init_kern: struct %s data is not found in struct %s\n",
+			tname, stname);
 		return -EINVAL;
 	}
 
-- 
2.45.0


^ permalink raw reply related	[flat|nested] 3+ messages in thread

* Re: [PATCH bpf-next] libbpf: fix error when st-prefix_ops and ops from differ btf
  2025-09-26  7:17 [PATCH bpf-next] libbpf: fix error when st-prefix_ops and ops from differ btf D. Wythe
@ 2025-09-26 16:53 ` Martin KaFai Lau
  2025-09-26 20:12   ` Andrii Nakryiko
  0 siblings, 1 reply; 3+ messages in thread
From: Martin KaFai Lau @ 2025-09-26 16:53 UTC (permalink / raw)
  To: D. Wythe
  Cc: ast, daniel, andrii, pabeni, song, sdf, haoluo, yhs, edumazet,
	john.fastabend, kpsingh, jolsa, mjambigi, wenjia, wintera,
	dust.li, tonylu, guwen, bpf, davem, kuba, netdev, sidraya, jaka

On 9/26/25 12:17 AM, D. Wythe wrote:
> When a module registers a struct_ops, the struct_ops type and its
> corresponding map_value type ("bpf_struct_ops_") may reside in different
> btf objects, here are four possible case:
> 
> +--------+---------------+-------------+---------------------------------+
> |        |bpf_struct_ops_| xxx_ops     |                                 |
> +--------+---------------+-------------+---------------------------------+
> | case 0 | btf_vmlinux   | bft_vmlinux | be used and reg only in vmlinux |

s/bft/btf/

> +--------+---------------+-------------+---------------------------------+
> | case 1 | btf_vmlinux   | mod_btf     | INVALID                         |
> +--------+---------------+-------------+---------------------------------+
> | case 2 | mod_btf       | btf_vmlinux | reg in mod but be used both in  |
> |        |               |             | vmlinux and mod.                |
> +--------+---------------+-------------+---------------------------------+
> | case 3 | mod_btf       | mod_btf     | be used and reg only in mod     |
> +--------+---------------+-------------+---------------------------------+
> 
> Currently we figure out the mod_btf by searching with the struct_ops type,
> which makes it impossible to figure out the mod_btf when the struct_ops
> type is in btf_vmlinux while it's corresponding map_value type is in
> mod_btf (case 2).
> 
> The fix is to use the corresponding map_value type ("bpf_struct_ops_")
> as the lookup anchor instead of the struct_ops type to figure out the
> `btf` and `mod_btf` via find_ksym_btf_id(), and then we can locate
> the kern_type_id via btf__find_by_name_kind() with the `btf` we just
> obtained from find_ksym_btf_id().
> 
> With this change the lookup obtains the correct btf and mod_btf for case 2,
> preserves correct behavior for other valid cases, and still fails as
> expected for the invalid scenario (case 1).
> 
> Fixes: 590a00888250 ("bpf: libbpf: Add STRUCT_OPS support")
> Signed-off-by: D. Wythe <alibuda@linux.alibaba.com>
> Acked-by: Andrii Nakryiko <andrii@kernel.org>
> ---
>   tools/lib/bpf/libbpf.c | 37 ++++++++++++++++++-------------------
>   1 file changed, 18 insertions(+), 19 deletions(-)
> 
> diff --git a/tools/lib/bpf/libbpf.c b/tools/lib/bpf/libbpf.c
> index 5161c2b39875..a93eed660404 100644
> --- a/tools/lib/bpf/libbpf.c
> +++ b/tools/lib/bpf/libbpf.c
> @@ -1018,35 +1018,34 @@ find_struct_ops_kern_types(struct bpf_object *obj, const char *tname_raw,
>   	const struct btf_member *kern_data_member;
>   	struct btf *btf = NULL;
>   	__s32 kern_vtype_id, kern_type_id;
> -	char tname[256];
> +	char tname[256], stname[256];
>   	__u32 i;
>   
>   	snprintf(tname, sizeof(tname), "%.*s",
>   		 (int)bpf_core_essential_name_len(tname_raw), tname_raw);
>   
> -	kern_type_id = find_ksym_btf_id(obj, tname, BTF_KIND_STRUCT,
> -					&btf, mod_btf);
> -	if (kern_type_id < 0) {
> -		pr_warn("struct_ops init_kern: struct %s is not found in kernel BTF\n",
> -			tname);
> -		return kern_type_id;
> -	}
> -	kern_type = btf__type_by_id(btf, kern_type_id);
> +	snprintf(stname, sizeof(stname), "%s%.*s", STRUCT_OPS_VALUE_PREFIX,
> +		 (int)strlen(tname), tname);

nit. strlen(tname) should not be needed. Others lgtm.

Acked-by: Martin KaFai Lau <martin.lau@kernel.org>



^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [PATCH bpf-next] libbpf: fix error when st-prefix_ops and ops from differ btf
  2025-09-26 16:53 ` Martin KaFai Lau
@ 2025-09-26 20:12   ` Andrii Nakryiko
  0 siblings, 0 replies; 3+ messages in thread
From: Andrii Nakryiko @ 2025-09-26 20:12 UTC (permalink / raw)
  To: Martin KaFai Lau
  Cc: D. Wythe, ast, daniel, andrii, pabeni, song, sdf, haoluo, yhs,
	edumazet, john.fastabend, kpsingh, jolsa, mjambigi, wenjia,
	wintera, dust.li, tonylu, guwen, bpf, davem, kuba, netdev,
	sidraya, jaka

On Fri, Sep 26, 2025 at 9:53 AM Martin KaFai Lau <martin.lau@linux.dev> wrote:
>
> On 9/26/25 12:17 AM, D. Wythe wrote:
> > When a module registers a struct_ops, the struct_ops type and its
> > corresponding map_value type ("bpf_struct_ops_") may reside in different
> > btf objects, here are four possible case:
> >
> > +--------+---------------+-------------+---------------------------------+
> > |        |bpf_struct_ops_| xxx_ops     |                                 |
> > +--------+---------------+-------------+---------------------------------+
> > | case 0 | btf_vmlinux   | bft_vmlinux | be used and reg only in vmlinux |
>
> s/bft/btf/
>
> > +--------+---------------+-------------+---------------------------------+
> > | case 1 | btf_vmlinux   | mod_btf     | INVALID                         |
> > +--------+---------------+-------------+---------------------------------+
> > | case 2 | mod_btf       | btf_vmlinux | reg in mod but be used both in  |
> > |        |               |             | vmlinux and mod.                |
> > +--------+---------------+-------------+---------------------------------+
> > | case 3 | mod_btf       | mod_btf     | be used and reg only in mod     |
> > +--------+---------------+-------------+---------------------------------+
> >
> > Currently we figure out the mod_btf by searching with the struct_ops type,
> > which makes it impossible to figure out the mod_btf when the struct_ops
> > type is in btf_vmlinux while it's corresponding map_value type is in
> > mod_btf (case 2).
> >
> > The fix is to use the corresponding map_value type ("bpf_struct_ops_")
> > as the lookup anchor instead of the struct_ops type to figure out the
> > `btf` and `mod_btf` via find_ksym_btf_id(), and then we can locate
> > the kern_type_id via btf__find_by_name_kind() with the `btf` we just
> > obtained from find_ksym_btf_id().
> >
> > With this change the lookup obtains the correct btf and mod_btf for case 2,
> > preserves correct behavior for other valid cases, and still fails as
> > expected for the invalid scenario (case 1).
> >
> > Fixes: 590a00888250 ("bpf: libbpf: Add STRUCT_OPS support")
> > Signed-off-by: D. Wythe <alibuda@linux.alibaba.com>
> > Acked-by: Andrii Nakryiko <andrii@kernel.org>
> > ---
> >   tools/lib/bpf/libbpf.c | 37 ++++++++++++++++++-------------------
> >   1 file changed, 18 insertions(+), 19 deletions(-)
> >
> > diff --git a/tools/lib/bpf/libbpf.c b/tools/lib/bpf/libbpf.c
> > index 5161c2b39875..a93eed660404 100644
> > --- a/tools/lib/bpf/libbpf.c
> > +++ b/tools/lib/bpf/libbpf.c
> > @@ -1018,35 +1018,34 @@ find_struct_ops_kern_types(struct bpf_object *obj, const char *tname_raw,
> >       const struct btf_member *kern_data_member;
> >       struct btf *btf = NULL;
> >       __s32 kern_vtype_id, kern_type_id;
> > -     char tname[256];
> > +     char tname[256], stname[256];

I adjusted tname size down to 192 to avoid compiler complains that
snprintf below can truncate output (useless warning, but oh well)

> >       __u32 i;
> >
> >       snprintf(tname, sizeof(tname), "%.*s",
> >                (int)bpf_core_essential_name_len(tname_raw), tname_raw);
> >
> > -     kern_type_id = find_ksym_btf_id(obj, tname, BTF_KIND_STRUCT,
> > -                                     &btf, mod_btf);
> > -     if (kern_type_id < 0) {
> > -             pr_warn("struct_ops init_kern: struct %s is not found in kernel BTF\n",
> > -                     tname);
> > -             return kern_type_id;
> > -     }
> > -     kern_type = btf__type_by_id(btf, kern_type_id);
> > +     snprintf(stname, sizeof(stname), "%s%.*s", STRUCT_OPS_VALUE_PREFIX,
> > +              (int)strlen(tname), tname);
>
> nit. strlen(tname) should not be needed. Others lgtm.

yeah, dropped that while applying, thanks. this was a copy/paste from
tname's snprint, but there we do actually need a prefix of a type, not
here, so strlen() makes no sense


applied to bpf-next

>
> Acked-by: Martin KaFai Lau <martin.lau@kernel.org>
>
>
>

^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2025-09-26 20:13 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2025-09-26  7:17 [PATCH bpf-next] libbpf: fix error when st-prefix_ops and ops from differ btf D. Wythe
2025-09-26 16:53 ` Martin KaFai Lau
2025-09-26 20:12   ` Andrii Nakryiko

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).