Re: [PATCH v7 04/11] readahead: allocate folios with mapping_min_order in readahead

[Hannes Reinecke <hare@suse.de>]

On 6/21/24 14:19, Pankaj Raghav (Samsung) wrote:
> On Tue, Jun 18, 2024 at 08:56:53AM +0200, Hannes Reinecke wrote:
>> On 6/17/24 18:39, Pankaj Raghav (Samsung) wrote:
>>> On Mon, Jun 17, 2024 at 05:10:15PM +0100, Matthew Wilcox wrote:
>>>> On Mon, Jun 17, 2024 at 04:04:20PM +0000, Pankaj Raghav (Samsung) wrote:
>>>>> On Mon, Jun 17, 2024 at 01:32:42PM +0100, Matthew Wilcox wrote:
>>>>> So the following can still be there from Hannes patch as we have a
>>>>> stable reference:
>>>>>
>>>>>    		ractl->_workingset |= folio_test_workingset(folio);
>>>>> -		ractl->_nr_pages++;
>>>>> +		ractl->_nr_pages += folio_nr_pages(folio);
>>>>> +		i += folio_nr_pages(folio);
>>>>>    	}
>>>>
>>>> We _can_, but we just allocated it, so we know what size it is already.
>>> Yes.
>>>
>>>> I'm starting to feel that Hannes' patch should be combined with this
>>>> one.
>>>
>>> Fine by me. @Hannes, is that ok with you?
>>
>> Sure. I was about to re-send my patchset anyway, so feel free to wrap it in.
> Is it ok if I add your Co-developed and Signed-off tag?
> This is what I have combining your patch with mine and making willy's
> changes:
> 
> diff --git a/mm/readahead.c b/mm/readahead.c
> index 389cd802da63..f56da953c130 100644
> --- a/mm/readahead.c
> +++ b/mm/readahead.c
> @@ -247,9 +247,7 @@ void page_cache_ra_unbounded(struct readahead_control *ractl,
>                  struct folio *folio = xa_load(&mapping->i_pages, index + i);
>                  int ret;
>   
> -
>                  if (folio && !xa_is_value(folio)) {
> -                       long nr_pages = folio_nr_pages(folio);
>                          /*
>                           * Page already present?  Kick off the current batch
>                           * of contiguous pages before continuing with the
> @@ -259,18 +257,7 @@ void page_cache_ra_unbounded(struct readahead_control *ractl,
>                           * not worth getting one just for that.
>                           */
>                          read_pages(ractl);
> -
> -                       /*
> -                        * Move the ractl->_index by at least min_pages
> -                        * if the folio got truncated to respect the
> -                        * alignment constraint in the page cache.
> -                        *
> -                        */
> -                       if (mapping != folio->mapping)
> -                               nr_pages = min_nrpages;
> -
> -                       VM_BUG_ON_FOLIO(nr_pages < min_nrpages, folio);
> -                       ractl->_index += nr_pages;
> +                       ractl->_index += min_nrpages;
>                          i = ractl->_index + ractl->_nr_pages - index;
>                          continue;
>                  }
> @@ -293,8 +280,8 @@ void page_cache_ra_unbounded(struct readahead_control *ractl,
>                  if (i == mark)
>                          folio_set_readahead(folio);
>                  ractl->_workingset |= folio_test_workingset(folio);
> -               ractl->_nr_pages += folio_nr_pages(folio);
> -               i += folio_nr_pages(folio);
> +               ractl->_nr_pages += min_nrpages;
> +               i += min_nrpages;
>          }
>   
>          /*
> 
Yes, that looks fine.
Go.

Cheers,

Hannes
-- 
Dr. Hannes Reinecke                  Kernel Storage Architect
hare@suse.de                                +49 911 74053 688
SUSE Software Solutions GmbH, Frankenstr. 146, 90461 Nürnberg
HRB 36809 (AG Nürnberg), GF: I. Totev, A. McDonald, W. Knoblich