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

[Matthew Wilcox]

On Wed, Jan 12, 2022 at 01:51:55PM +0530, 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.

> > Since every value entry found in the xarray requires me to do the
> > xas_pause(), I do agree that xa_for_each() is the appropriate call here.
> > Will switch to this in the next spin. Waiting for further review
> > comments on this patch.
> 
> I also found the below documentation:
> xa_for_each() will spin if it hits a retry entry; if you intend to see
> retry entries, you should use the xas_for_each() iterator instead.
> 
> Since retry entries are expected, I should be using the xas_for_each()
> with the corrections you had pointed out. Isn't it?

No.  You aren't handling retry entries at all; you clearly don't
expect to see them.  Just let the XArray code handle them itself.