Re: [PATCH bpf-next v3 02/10] rhashtable: Add rhashtable_walk_enter_from()

[Mykyta Yatsenko <mykyta.yatsenko5@gmail.com>]

On 4/28/26 11:35 AM, Herbert Xu wrote:
> On Fri, Apr 24, 2026 at 12:50:44PM -0700, Mykyta Yatsenko wrote:
> .
>> diff --git a/lib/rhashtable.c b/lib/rhashtable.c
>> index 6074ed5f66f3..2a6098edf737 100644
>> --- a/lib/rhashtable.c
>> +++ b/lib/rhashtable.c
>> @@ -692,6 +692,47 @@ void rhashtable_walk_enter(struct rhashtable *ht, struct rhashtable_iter *iter)
>>  }
>>  EXPORT_SYMBOL_GPL(rhashtable_walk_enter);
>>  
>> +/**
>> + * rhashtable_walk_enter_from - Initialise a walk starting at a key's bucket
>> + * @ht:		Table to walk over
>> + * @iter:	Hash table iterator
>> + * @key:	Key whose bucket to start from
>> + * @params:	Hash table parameters
>> + *
>> + * Like rhashtable_walk_enter(), but positions the iterator at the bucket
>> + * containing @key.  If @key is not present in the main table, the iterator is
>> + * positioned at the beginning.
>> + *
>> + * Same constraints as rhashtable_walk_enter() apply.
>> + */
>> +void rhashtable_walk_enter_from(struct rhashtable *ht,
>> +				struct rhashtable_iter *iter,
>> +				const void *key,
>> +				const struct rhashtable_params params)
>> +				__must_hold(RCU)
>> +{
>> +	struct bucket_table *tbl;
>> +	struct rhash_head *he;
>> +
>> +	rhashtable_walk_enter(ht, iter);
>> +
>> +	if (!key)
>> +		return;
>> +
>> +	tbl = iter->walker.tbl;
>> +	if (!tbl)
>> +		return;
>> +
>> +	he = __rhashtable_lookup_one(ht, tbl, key, params,
>> +				     RHT_LOOKUP_NORMAL);
>> +	if (!he)
>> +		return;
>> +
>> +	iter->slot = rht_key_hashfn(ht, tbl, key, params);
>> +	iter->p = he;
>> +}
>> +EXPORT_SYMBOL_GPL(rhashtable_walk_enter_from);
> 
> I still think this is misguided.  I'm not going to stop you from
> doing it if you really wanted to, but this thing will not work well
> if your rhashtable is constantly resizing.
> 
> You also need to consider the possibility that an adversary is
> triggering the constant rehashing.
> 
> Cheers,

Herbert,
We brainstormed with Alexei on how to integrate rhashtable better,
as a BPF map, we have the next proposal:

Introduce rhashtable_next_key() API
```
static inline struct rhash_head *rhashtable_next_key(
	struct rhashtable *ht, const void *prev_key,
	const struct rhashtable_params params);
```

It returns the next element, by finding prev_key in the tbl chain
and returning the next element in whatever tbl the prev one is.
This way we keep rhashtable API tight, less surface, not exposing unnecessary
details, although, this new API will have the same issues with full
iteration support (which we fine to deal with on userspace side).
Appreciate if you share your opinion on this.