On 12/11/18 11:46 AM, Jens Axboe wrote:
> On 12/11/18 11:45 AM, Matthew Wilcox wrote:
>> On Tue, Dec 11, 2018 at 11:41:55AM -0700, Jens Axboe wrote:
>>> On Wed, Nov 28, 2018 at 11:35 AM Matthew Wilcox <willy@xxxxxxxxxxxxx> wrote:
>>>>
>>>> rcu_read_lock();
>>>> - table = rcu_dereference(mm->ioctx_table);
>>>> -
>>>> - if (!table || id >= table->nr)
>>>> - goto out;
>>>> -
>>>> - ctx = rcu_dereference(table->table[id]);
>>>> + ctx = xa_load(&mm->ioctx, id);
>>>> if (ctx && ctx->user_id == ctx_id) {
>>>> if (percpu_ref_tryget_live(&ctx->users))
>>>> ret = ctx;
>>>> }
>>>
>>> Question on this part - do we need that RCU read lock around this now? I
>>> don't think we do.
>>
>> I think we need the rcu read lock here to prevent ctx from being freed
>> under us by free_ioctx().
>
> Then that begs the question, how about __xa_load() that is already called
> under RCU read lock?
Something like this, mem remap has an existing user that can use this
too already.
--
Jens Axboe
