Skip to contentSkip to conversation

libbpf: split BTF presence checks into libbpf- and kernel-specific parts

PATCH Applied Andrii Nakryiko 3 participants 3 messages Mar 12, 2020

Subsystems (4)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 libbpf: split BTF presence checks into libbpf- and kernel-specific parts

Needs for application BTF being present differs between user-space libbpf needs and kernel needs. Currently, BTF is mandatory only in kernel only when BPF application is using STRUCT_OPS. While libbpf itself relies more heavily on presense of BTF:
- for BTF-defined maps;
- for Kconfig externs;
- for STRUCT_OPS as well.

Thus, checks for presence and validness of bpf_object's BPF needs to be performed separately, which is patch does.

Fixes 5327644614a1 ("libbpf: Relax check whether BTF is mandatory")
Cc Quentin Monnet <[email protected]>
---
 tools/lib/bpf/libbpf.c | 17 ++++++++++++-----
 1 file changed, 12 insertions(+), 5 deletions(-)
tools/lib/bpf/libbpf.c +12 -5
index 223be01dc466..1a787a2faf58 100644
@@ -2284,9 +2284,16 @@ static void bpf_object__sanitize_btf_ext(struct bpf_object *obj)
 	}
 }
 
-static bool bpf_object__is_btf_mandatory(const struct bpf_object *obj)
+static bool libbpf_needs_btf(const struct bpf_object *obj)
 {
-	return obj->efile.st_ops_shndx >= 0 || obj->nr_extern > 0;
+	return obj->efile.btf_maps_shndx >= 0 ||
+	       obj->efile.st_ops_shndx >= 0 ||
+	       obj->nr_extern > 0;
+}
+
+static bool kernel_needs_btf(const struct bpf_object *obj)
+{
+	return obj->efile.st_ops_shndx >= 0;
 }
 
 static int bpf_object__init_btf(struct bpf_object *obj,
@@ -2322,7 +2329,7 @@ static int bpf_object__init_btf(struct bpf_object *obj,
 		}
 	}
 out:
-	if (err && bpf_object__is_btf_mandatory(obj)) {
+	if (err && libbpf_needs_btf(obj)) {
 		pr_warn("BTF is required, but is missing or corrupted.\n");
 		return err;
 	}
@@ -2346,7 +2353,7 @@ static int bpf_object__finalize_btf(struct bpf_object *obj)
 	btf_ext__free(obj->btf_ext);
 	obj->btf_ext = NULL;
 
-	if (bpf_object__is_btf_mandatory(obj)) {
+	if (libbpf_needs_btf(obj)) {
 		pr_warn("BTF is required, but is missing or corrupted.\n");
 		return -ENOENT;
 	}
@@ -2410,7 +2417,7 @@ static int bpf_object__sanitize_and_load_btf(struct bpf_object *obj)
 			obj->btf_ext = NULL;
 		}
 
-		if (bpf_object__is_btf_mandatory(obj))
+		if (kernel_needs_btf(obj))
 			return err;
 	}
 	return 0;
Signature · 1 line
2.17.1
Acked-by on v1 1/1
#1 lore raw
Quoted 1 line · Andrii Nakryiko

Needs for application BTF being present differs between user-space libbpf needs and kernel

Nit. This line looks too long for commit message.

Quoted 8 lines

needs. Currently, BTF is mandatory only in kernel only when BPF application is using STRUCT_OPS. While libbpf itself relies more heavily on presense of BTF:
- for BTF-defined maps;
- for Kconfig externs;
- for STRUCT_OPS as well.

Thus, checks for presence and validness of bpf_object's BPF needs to be performed separately, which is patch does.

maintainer Applied
#2 lore raw
In reply to Andrii Nakryiko
Quoted 7 lines · Andrii Nakryiko

Needs for application BTF being present differs between user-space libbpf needs and kernel needs. Currently, BTF is mandatory only in kernel only when BPF application is using STRUCT_OPS. While libbpf itself relies more heavily on presense of BTF:
- for BTF-defined maps;
- for Kconfig externs;
- for STRUCT_OPS as well.

Quoted 7 lines

Thus, checks for presence and validness of bpf_object's BPF needs to be performed separately, which is patch does.

Fixes 5327644614a1 ("libbpf: Relax check whether BTF is mandatory")
Cc Quentin Monnet <[email protected]>

Applied, thanks!