Re: [PATCH bpf-next v13 7/8] libbpf: Add syscall common attributes support for map_create

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

On 13/5/26 07:56, sashiko-bot@kernel.org wrote:
[...]
>>  
>> -	fd = sys_bpf_fd(BPF_MAP_CREATE, &attr, attr_sz);
>> +	log_opts = OPTS_GET(opts, log_opts, NULL);
>> +	if (log_opts && feat_supported(NULL, FEAT_BPF_SYSCALL_COMMON_ATTRS)) {
> 
> Should there be an OPTS_VALID() check for the nested log_opts structure
> before we start extracting its fields?
> 
> While the parent opts structure is correctly validated with OPTS_VALID()
> at the start of bpf_map_create(), log_opts is a newly introduced option
> structure with its own sz field for forward and backward compatibility.
> 
> If an application compiled against a future libbpf header passes a
> bpf_log_opts struct with new, non-zero fields, wouldn't this silently
> ignore those fields instead of returning -EINVAL?
> 

Good catch.

OPTS_VALID() is missed for log_opts here.

Will post a follow-up patch after the sashiko reviews for v14.

Thanks,
Leon

>> +		memset(&attr_common, 0, attr_common_sz);
>> +		attr_common.log_buf = ptr_to_u64(OPTS_GET(log_opts, buf, NULL));
>> +		attr_common.log_size = OPTS_GET(log_opts, size, 0);
>> +		attr_common.log_level = OPTS_GET(log_opts, level, 0);
>