On Wed, Dec 09, 2015 at 10:43:23AM +0800, Herbert Xu wrote:
>
> OK after thinking about it a little bit more I think your approach
> is safe and we won't need to play any more games to get this to
> work. So I'm fine with this patch.
One suggestion though: your inline function doesn't actually
in
On Tue, Dec 08, 2015 at 09:14:24AM -0800, Tom Herbert wrote:
>
> That creates one more level of indirection. I don't see how add an
> atomic replace operation adds any complexity to the rhashtable, none
> of the semantics for rhashtable need to be changed.
OK after thinking about it a little bit
On Tue, Dec 8, 2015 at 1:39 AM, Herbert Xu wrote:
> David Miller wrote:
>> From: Tom Herbert
>> Date: Tue, 1 Dec 2015 15:11:09 -0800
>>
>>> + lock = rht_bucket_lock(tbl, hash);
>>> +
>>> + spin_lock_bh(lock);
>>> +
>>> + pprev = &tbl->buckets[hash];
>>> + rht_for_each(he, tbl, ha
David Miller wrote:
> From: Tom Herbert
> Date: Tue, 1 Dec 2015 15:11:09 -0800
>
>> + lock = rht_bucket_lock(tbl, hash);
>> +
>> + spin_lock_bh(lock);
>> +
>> + pprev = &tbl->buckets[hash];
>> + rht_for_each(he, tbl, hash) {
>> + if (he != obj_old) {
>> +
From: Tom Herbert
Date: Tue, 1 Dec 2015 15:11:09 -0800
> + lock = rht_bucket_lock(tbl, hash);
> +
> + spin_lock_bh(lock);
> +
> + pprev = &tbl->buckets[hash];
> + rht_for_each(he, tbl, hash) {
> + if (he != obj_old) {
> + pprev = &he->next;
> +
