Re: [PATCH 1/4] mm: vmscan: Correct check for kswapd sleeping in sleeping_prematurely

[Minchan Kim]

On Fri, Jun 24, 2011 at 11:44 PM, Mel Gorman <mgorman@xxxxxxx> wrote:
> During allocator-intensive workloads, kswapd will be woken frequently
> causing free memory to oscillate between the high and min watermark.
> This is expected behaviour.
>
> A problem occurs if the highest zone is small. Âbalance_pgdat()
> only considers unreclaimable zones when priority is DEF_PRIORITY
> but sleeping_prematurely considers all zones. It's possible for this
> sequence to occur
>
> Â1. kswapd wakes up and enters balance_pgdat()
> Â2. At DEF_PRIORITY, marks highest zone unreclaimable
> Â3. At DEF_PRIORITY-1, ignores highest zone setting end_zone
> Â4. At DEF_PRIORITY-1, calls shrink_slab freeing memory from
> Â Â Â Âhighest zone, clearing all_unreclaimable. Highest zone
> Â Â Â Âis still unbalanced
> Â5. kswapd returns and calls sleeping_prematurely
> Â6. sleeping_prematurely looks at *all* zones, not just the ones
> Â Â being considered by balance_pgdat. The highest small zone
> Â Â has all_unreclaimable cleared but but the zone is not
> Â Â balanced. all_zones_ok is false so kswapd stays awake
>
> This patch corrects the behaviour of sleeping_prematurely to check
> the zones balance_pgdat() checked.
>
> Reported-and-tested-by: PÃdraig Brady <P@xxxxxxxxxxxxxx>
> Signed-off-by: Mel Gorman <mgorman@xxxxxxx>
Reviewed-by: Minchan Kim <minchan.kim@xxxxxxxxx>

-- 
Kind regards,
Minchan Kim
--
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/