Le 21/08/2024 à 13:23, Soma Nakata a écrit :
`tname` is returned by `btf__name_by_offset()` as well as `var_name`,
and these addresses point to strings in the btf. Since their locations
may change while loading the bpf program, using `strdup()` ensures
`tname` is safely stored.
Signed-off-by: Soma Nakata <soma.nakata01@xxxxxxxxx>
---
tools/lib/bpf/libbpf.c | 7 +++++--
1 file changed, 5 insertions(+), 2 deletions(-)
diff --git a/tools/lib/bpf/libbpf.c b/tools/lib/bpf/libbpf.c
index a3be6f8fac09..f4ad1b993ec5 100644
--- a/tools/lib/bpf/libbpf.c
+++ b/tools/lib/bpf/libbpf.c
@@ -496,7 +496,7 @@ struct bpf_program {
};
struct bpf_struct_ops {
- const char *tname;
+ char *tname;
const struct btf_type *type;
struct bpf_program **progs;
__u32 *kern_func_off;
@@ -1423,7 +1423,9 @@ static int init_struct_ops_maps(struct bpf_object *obj, const char *sec_name,
memcpy(st_ops->data,
data->d_buf + vsi->offset,
type->size);
- st_ops->tname = tname;
+ st_ops->tname = strdup(tname);
+ if (!st_ops->tname)
+ return -ENOMEM;
Certainly a matter of taste, but I would personally move it just after
"st_ops->kern_func_off = malloc()" and add the NULL check with the
existing ones.
BTW, there are some memory leaks if 1 or more allocations fail in this
function.
Not sure if it is an issue or not, and what should be done in this case.
CJ
st_ops->type = type;
st_ops->type_id = type_id;
@@ -8984,6 +8986,7 @@ static void bpf_map__destroy(struct bpf_map *map)
map->mmaped = NULL;
if (map->st_ops) {
+ zfree(&map->st_ops->tname);
zfree(&map->st_ops->data);
zfree(&map->st_ops->progs);
zfree(&map->st_ops->kern_func_off);
