Commit e6be8cd5 authored by Kui-Feng Lee's avatar Kui-Feng Lee Committed by Martin KaFai Lau

bpf: Fix error checks against bpf_get_btf_vmlinux().

In bpf_struct_ops_map_alloc, it needs to check for NULL in the returned
pointer of bpf_get_btf_vmlinux() when CONFIG_DEBUG_INFO_BTF is not set.
ENOTSUPP is used to preserve the same behavior before the
struct_ops kmod support.

In the function check_struct_ops_btf_id(), instead of redoing the
bpf_get_btf_vmlinux() that has already been done in syscall.c, the fix
here is to check for prog->aux->attach_btf_id.
BPF_PROG_TYPE_STRUCT_OPS must require attach_btf_id and syscall.c
guarantees a valid attach_btf as long as attach_btf_id is set.
When attach_btf_id is not set, this patch returns -ENOTSUPP
because it is what the selftest in test_libbpf_probe_prog_types()
and libbpf_probes.c are expecting for feature probing purpose.

Changes from v1:

 - Remove an unnecessary NULL check in check_struct_ops_btf_id()

Reported-by: syzbot+88f0aafe5f950d7489d7@syzkaller.appspotmail.com
Closes: https://lore.kernel.org/bpf/00000000000040d68a060fc8db8c@google.com/
Reported-by: syzbot+1336f3d4b10bcda75b89@syzkaller.appspotmail.com
Closes: https://lore.kernel.org/bpf/00000000000026353b060fc21c07@google.com/
Fixes: fcc2c1fb ("bpf: pass attached BTF to the bpf_struct_ops subsystem")
Signed-off-by: default avatarKui-Feng Lee <thinker.li@gmail.com>
Link: https://lore.kernel.org/r/20240126023113.1379504-1-thinker.li@gmail.comSigned-off-by: default avatarMartin KaFai Lau <martin.lau@kernel.org>
parent be4840b3
...@@ -669,6 +669,8 @@ static struct bpf_map *bpf_struct_ops_map_alloc(union bpf_attr *attr) ...@@ -669,6 +669,8 @@ static struct bpf_map *bpf_struct_ops_map_alloc(union bpf_attr *attr)
btf = bpf_get_btf_vmlinux(); btf = bpf_get_btf_vmlinux();
if (IS_ERR(btf)) if (IS_ERR(btf))
return ERR_CAST(btf); return ERR_CAST(btf);
if (!btf)
return ERR_PTR(-ENOTSUPP);
} }
st_ops_desc = bpf_struct_ops_find_value(btf, attr->btf_vmlinux_value_type_id); st_ops_desc = bpf_struct_ops_find_value(btf, attr->btf_vmlinux_value_type_id);
......
...@@ -20298,7 +20298,10 @@ static int check_struct_ops_btf_id(struct bpf_verifier_env *env) ...@@ -20298,7 +20298,10 @@ static int check_struct_ops_btf_id(struct bpf_verifier_env *env)
return -EINVAL; return -EINVAL;
} }
btf = prog->aux->attach_btf ?: bpf_get_btf_vmlinux(); if (!prog->aux->attach_btf_id)
return -ENOTSUPP;
btf = prog->aux->attach_btf;
if (btf_is_module(btf)) { if (btf_is_module(btf)) {
/* Make sure st_ops is valid through the lifetime of env */ /* Make sure st_ops is valid through the lifetime of env */
env->attach_btf_mod = btf_try_get_module(btf); env->attach_btf_mod = btf_try_get_module(btf);
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment