Re: [RFC PATCH bpf-next v3 02/10] libbpf: Add support for extended bpf syscall

[Leon Hwang <leon.hwang@linux.dev>]

On 7/10/25 07:08, Andrii Nakryiko wrote:
> On Thu, Oct 2, 2025 at 8:49 AM Leon Hwang <leon.hwang@linux.dev> wrote:
>>
>> To support the extended 'bpf()' syscall introduced in the previous commit,
>> introduce the following internal APIs:
>>
>> * 'sys_bpf_ext()'
>> * 'sys_bpf_ext_fd()'
>>   They wrap the raw 'syscall()' interface to support passing extended
>>   attributes.
>> * 'probe_sys_bpf_ext()'
>>   Check whether current kernel supports the extended attributes.
>>
>> Signed-off-by: Leon Hwang <leon.hwang@linux.dev>
>> ---
>>  tools/lib/bpf/bpf.c             | 33 +++++++++++++++++++++++++++++++++
>>  tools/lib/bpf/features.c        |  8 ++++++++
>>  tools/lib/bpf/libbpf_internal.h |  3 +++
>>  3 files changed, 44 insertions(+)
>>
>> diff --git a/tools/lib/bpf/bpf.c b/tools/lib/bpf/bpf.c
>> index 339b197972374..9cd79beb13a2d 100644
>> --- a/tools/lib/bpf/bpf.c
>> +++ b/tools/lib/bpf/bpf.c
>> @@ -69,6 +69,39 @@ static inline __u64 ptr_to_u64(const void *ptr)
>>         return (__u64) (unsigned long) ptr;
>>  }
>>
>> +static inline int sys_bpf_ext(enum bpf_cmd cmd, union bpf_attr *attr,
>> +                             unsigned int size,
>> +                             struct bpf_common_attr *common_attrs,
>> +                             unsigned int size_common)
>> +{
>> +       cmd = common_attrs ? cmd | BPF_COMMON_ATTRS : cmd & ~BPF_COMMON_ATTRS;
>
> nit: put those () two branches of ternary operator, there is no need
> to rely on obscure C operator precedence order here
>

Ack.

>> +       return syscall(__NR_bpf, cmd, attr, size, common_attrs, size_common);
>> +}
>> +
>> +static inline int sys_bpf_ext_fd(enum bpf_cmd cmd, union bpf_attr *attr,
>> +                                unsigned int size,
>> +                                struct bpf_common_attr *common_attrs,
>> +                                unsigned int size_common)
>> +{
>> +       int fd;
>> +
>> +       fd = sys_bpf_ext(cmd, attr, size, common_attrs, size_common);
>> +       return ensure_good_fd(fd);
>> +}
>> +
>> +int probe_sys_bpf_ext(void)
>> +{
>> +       const size_t attr_sz = offsetofend(union bpf_attr, prog_token_fd);
>> +       union bpf_attr attr;
>> +       int fd;
>> +
>> +       memset(&attr, 0, attr_sz);
>> +       fd = syscall(__NR_bpf, BPF_PROG_LOAD | BPF_COMMON_ATTRS, &attr, attr_sz, NULL,
>> +                    sizeof(struct bpf_common_attr));
>> +       fd = errno == EFAULT ? syscall(__NR_memfd_create, "fd", 0) : fd;
>> +       return ensure_good_fd(fd);
>
> why do we need to create FD?..
>
>> +}
>> +
>>  static inline int sys_bpf(enum bpf_cmd cmd, union bpf_attr *attr,
>>                           unsigned int size)
>>  {
>> diff --git a/tools/lib/bpf/features.c b/tools/lib/bpf/features.c
>> index 760657f5224c2..d01df62394f89 100644
>> --- a/tools/lib/bpf/features.c
>> +++ b/tools/lib/bpf/features.c
>> @@ -507,6 +507,11 @@ static int probe_kern_arg_ctx_tag(int token_fd)
>>         return probe_fd(prog_fd);
>>  }
>>
>> +static int probe_kern_extended_syscall(int token_fd)
>> +{
>> +       return probe_fd(probe_sys_bpf_ext());
>> +}
>> +
>
> just do that feature detection right here without creating any new
> FDs... make sys_bpf_ext() exposed just like we do that with sys_bpf()
> and use this here
>

Ack.

Then I'll expose sys_bpf_ext() directly without creating any new FDs.

Thanks,
Leon

[...]