Re: [PATCH] mm: migrate: One less atomic operation

From: Jacobo Giralt
Date: Sat Nov 05 2011 - 10:49:02 EST


I think I used an old e-mail address for Nick (got it from git blame),
I'm changing it now, sorry for the noise!

Regards,
Jacobo Giralt.

2011/11/5 Jacobo Giralt <jacobo.giralt@xxxxxxxxx>:
> From 3754c8617ef4377ce2ca2e3b28bdc28f8de1aa0d Mon Sep 17 00:00:00 2001
> From: Jacobo Giralt <jacobo.giralt@xxxxxxxxx>
> Date: Sat, 5 Nov 2011 13:12:50 +0100
> Subject: [PATCH] mm: migrate: One less atomic operation
>
> migrate_page_move_mapping drops a reference from the
> old page after unfreezing its counter. Both operations
> can be merged into a single atomic operation by
> directly unfreezing to one less reference.
>
> The same applies to migrate_huge_page_move_mapping.
>
> Signed-off-by: Jacobo Giralt <jacobo.giralt@xxxxxxxxx>
> ---
>  mm/migrate.c |   10 ++++------
>  1 files changed, 4 insertions(+), 6 deletions(-)
>
> diff --git a/mm/migrate.c b/mm/migrate.c
> index 33358f8..46d04a0 100644
> --- a/mm/migrate.c
> +++ b/mm/migrate.c
> @@ -269,12 +269,12 @@ static int migrate_page_move_mapping(struct address_space *mapping,
>
>        radix_tree_replace_slot(pslot, newpage);
>
> -       page_unfreeze_refs(page, expected_count);
>        /*
> -        * Drop cache reference from old page.
> +        * Drop cache reference from old page by unfreezing
> +        * to one less reference.
>         * We know this isn't the last reference.
>         */
> -       __put_page(page);
> +       page_unfreeze_refs(page, expected_count - 1);
>
>        /*
>         * If moved to a different zone then also account
> @@ -334,9 +334,7 @@ int migrate_huge_page_move_mapping(struct address_space *mapping,
>
>        radix_tree_replace_slot(pslot, newpage);
>
> -       page_unfreeze_refs(page, expected_count);
> -
> -       __put_page(page);
> +       page_unfreeze_refs(page, expected_count - 1);
>
>        spin_unlock_irq(&mapping->tree_lock);
>        return 0;
> --
> 1.7.5.4
>
>
>
>
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/