Re: [PATCH v3 RESEND] mm: shmem: implement POSIX_FADV_[WILL|DONT]NEED for shmem

[Charan Teja Kalla <quic_charante@quicinc.com>]

Hello Matthew,

On 1/12/2022 7:05 PM, Charan Teja Kalla wrote:
>>>>> +       rcu_read_lock();
>>>>> +       xas_for_each(&xas, page, end) {
>>>>> +               if (!xa_is_value(page))
>>>>> +                       continue;
>>>>> +               xas_pause(&xas);
>>>>> +               rcu_read_unlock();
>>>>> +
>>>>> +               page = shmem_read_mapping_page(mapping, xas.xa_index);
>>>>> +               if (!IS_ERR(page))
>>>>> +                       put_page(page);
>>>>> +
>>>>> +               rcu_read_lock();
>>>>> +               if (need_resched()) {
>>>>> +                       xas_pause(&xas);
>>>>> +                       cond_resched_rcu();
>>>>> +               }
>>>>> +       }
>>>>> +       rcu_read_unlock();
>>> Even the xarray documentation says that: If most entries found during a
>>> walk require you to call xas_pause(), the xa_for_each() iterator may be
>>> more appropriate.
> Yes.  This should obviously be an xa_for_each() loop.

In one of your patch[1], where we used xarray iterator, though most of
the entries found requires to call xas_pause() but still endup in using
xas_for_each() rather than xa_for_each(). Then, Should this code be
changed to use xa_for_each()? The documentation also says that "The
xas_for_each() iterator will expand into more inline code than
xa_for_each()."

[1]https://patchwork.kernel.org/project/linux-mm/patch/20200819184850.24779-4-willy@infradead.org/