Jakub Kicinski <jakub.kicinski@xxxxxxxxxxxxx> writes:
>> diff --git a/include/linux/bpf.h b/include/linux/bpf.h
>> index 5b9d22338606..753abfb78c13 100644
>> --- a/include/linux/bpf.h
>> +++ b/include/linux/bpf.h
>> @@ -383,6 +383,7 @@ struct bpf_prog_aux {
>> struct list_head ksym_lnode;
>> const struct bpf_prog_ops *ops;
>> struct bpf_map **used_maps;
>> + struct bpf_array *chain_progs;
>> struct bpf_prog *prog;
>> struct user_struct *user;
>> u64 load_time; /* ns since boottime */
>> @@ -443,6 +444,7 @@ struct bpf_array {
>>
>> #define BPF_COMPLEXITY_LIMIT_INSNS 1000000 /* yes. 1M insns */
>> #define MAX_TAIL_CALL_CNT 32
>> +#define BPF_NUM_CHAIN_SLOTS 8
>
> This could be user arg? Also the behaviour of mapping could be user
> controlled? Perhaps even users could pass the snippet to map the
> return code to the location, one day?
(Forgot to reply to this point).
Yeah, we could make it user-configurable. Or just dynamically increase
the size of the array if we run out. Or do something different with
linked list, as I alluded to in the other reply :)
-Toke
