On Thu,  8 Nov 2012 15:25:35 -0800
Sonny Rao <sonny...@chromium.org> wrote:

> The system uses global_dirtyable_memory() to calculate
> number of dirtyable pages/pages that can be allocated
> to the page cache.  A bug causes an underflow thus making
> the page count look like a big unsigned number.  This in turn
> confuses the dirty writeback throttling to aggressively write
> back pages as they become dirty (usually 1 page at a time).
> 
> Fix is to ensure there is no underflow while doing the math.
> 
> Signed-off-by: Sonny Rao <sonny...@chromium.org>
> Signed-off-by: Puneet Kumar <puneets...@chromium.org>
> ---
>  mm/page-writeback.c |   17 +++++++++++++----
>  1 files changed, 13 insertions(+), 4 deletions(-)
> 
> diff --git a/mm/page-writeback.c b/mm/page-writeback.c
> index 830893b..2a6356c 100644
> --- a/mm/page-writeback.c
> +++ b/mm/page-writeback.c
> @@ -194,11 +194,19 @@ static unsigned long highmem_dirtyable_memory(unsigned 
> long total)
>       unsigned long x = 0;
>  
>       for_each_node_state(node, N_HIGH_MEMORY) {
> +             unsigned long nr_pages;
>               struct zone *z =
>                       &NODE_DATA(node)->node_zones[ZONE_HIGHMEM];
>  
> -             x += zone_page_state(z, NR_FREE_PAGES) +
> -                  zone_reclaimable_pages(z) - z->dirty_balance_reserve;
> +             nr_pages = zone_page_state(z, NR_FREE_PAGES) +
> +                     zone_reclaimable_pages(z);
> +             /*
> +              * Unreclaimable memory (kernel memory or anonymous memory
> +              * without swap) can bring down the dirtyable pages below
> +              * the zone's dirty balance reserve.
> +              */
> +             if (nr_pages >= z->dirty_balance_reserve)
> +                     x += nr_pages - z->dirty_balance_reserve;

If the system has two nodes and one is below its dirty_balance_reserve,
we could end up with something like:

        x = 0;
        ...
        x += 1000;
        ...
        x += -100;

In this case, your fix would cause highmem_dirtyable_memory() to return
1000.  Would it be better to instead return 900?

IOW, we instead add logic along the lines of

        if ((long)x < 0)
                x = 0;
        return x;

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

Reply via email to