lkml.org 
[lkml]   [2004]   [Sep]   [5]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
    Patch in this message
    /
    Date
    From
    Subject[RFC][PATCH 3/3] teach kswapd about watermarks
    3/3


    Teach kswapd to free memory on behalf of higher order allocators. This could
    be important for higher order atomic allocations because they otherwise have
    no means to free the memory themselves.

    Signed-off-by: Nick Piggin <nickpiggin@yahoo.com.au>


    ---

    linux-2.6-npiggin/include/linux/mmzone.h | 5 ++--
    linux-2.6-npiggin/mm/page_alloc.c | 3 +-
    linux-2.6-npiggin/mm/vmscan.c | 33 ++++++++++++++++++++++---------
    3 files changed, 29 insertions(+), 12 deletions(-)

    diff -puN mm/vmscan.c~vm-kswapd-heed-order-watermarks mm/vmscan.c
    --- linux-2.6/mm/vmscan.c~vm-kswapd-heed-order-watermarks 2004-09-05 15:12:04.000000000 +1000
    +++ linux-2.6-npiggin/mm/vmscan.c 2004-09-05 15:12:04.000000000 +1000
    @@ -985,7 +985,7 @@ out:
    * the page allocator fallback scheme to ensure that aging of pages is balanced
    * across the zones.
    */
    -static int balance_pgdat(pg_data_t *pgdat, int nr_pages)
    +static int balance_pgdat(pg_data_t *pgdat, int nr_pages, int order)
    {
    int to_free = nr_pages;
    int priority;
    @@ -1023,7 +1023,8 @@ static int balance_pgdat(pg_data_t *pgda
    priority != DEF_PRIORITY)
    continue;

    - if (zone->free_pages <= zone->pages_high) {
    + if (!zone_watermark_ok(zone, order,
    + zone->pages_high, 0, 0, 0)) {
    end_zone = i;
    goto scan;
    }
    @@ -1055,7 +1056,8 @@ scan:
    continue;

    if (nr_pages == 0) { /* Not software suspend */
    - if (zone->free_pages <= zone->pages_high)
    + if (!zone_watermark_ok(zone, order,
    + zone->pages_high, end_zone, 0, 0))
    all_zones_ok = 0;
    }
    zone->temp_priority = priority;
    @@ -1146,13 +1148,22 @@ int kswapd(void *p)
    tsk->flags |= PF_MEMALLOC|PF_KSWAPD;

    for ( ; ; ) {
    + unsigned long order = 0, new_order;
    if (current->flags & PF_FREEZE)
    refrigerator(PF_FREEZE);
    +
    prepare_to_wait(&pgdat->kswapd_wait, &wait, TASK_INTERRUPTIBLE);
    - schedule();
    + new_order = atomic_read(&pgdat->kswapd_max_order);
    + atomic_set(&pgdat->kswapd_max_order, 0);
    + if (order < new_order) {
    + order = new_order;
    + } else {
    + schedule();
    + order = atomic_read(&pgdat->kswapd_max_order);
    + }
    finish_wait(&pgdat->kswapd_wait, &wait);

    - balance_pgdat(pgdat, 0);
    + balance_pgdat(pgdat, 0, order);
    }
    return 0;
    }
    @@ -1160,12 +1171,16 @@ int kswapd(void *p)
    /*
    * A zone is low on free memory, so wake its kswapd task to service it.
    */
    -void wakeup_kswapd(struct zone *zone)
    +void wakeup_kswapd(struct zone *zone, int order)
    {
    - if (zone->free_pages > zone->pages_low)
    - return;
    + pg_data_t *pgdat = zone->zone_pgdat;
    +
    if (!cpuset_zone_allowed(zone))
    return;
    +
    + if (atomic_read(&pgdat->kswapd_max_order) < order)
    + return;
    + atomic_set(&pgdat->kswapd_max_order, order);
    if (!waitqueue_active(&zone->zone_pgdat->kswapd_wait))
    return;
    wake_up_interruptible(&zone->zone_pgdat->kswapd_wait);
    @@ -1188,7 +1203,7 @@ int shrink_all_memory(int nr_pages)
    current->reclaim_state = &reclaim_state;
    for_each_pgdat(pgdat) {
    int freed;
    - freed = balance_pgdat(pgdat, nr_to_free);
    + freed = balance_pgdat(pgdat, nr_to_free, 0);
    ret += freed;
    nr_to_free -= freed;
    if (nr_to_free <= 0)
    diff -puN mm/page_alloc.c~vm-kswapd-heed-order-watermarks mm/page_alloc.c
    --- linux-2.6/mm/page_alloc.c~vm-kswapd-heed-order-watermarks 2004-09-05 15:12:04.000000000 +1000
    +++ linux-2.6-npiggin/mm/page_alloc.c 2004-09-05 15:12:04.000000000 +1000
    @@ -775,7 +775,7 @@ __alloc_pages(unsigned int gfp_mask, uns
    }

    for (i = 0; (z = zones[i]) != NULL; i++)
    - wakeup_kswapd(z);
    + wakeup_kswapd(z, order);

    /*
    * Go through the zonelist again. Let __GFP_HIGH and allocations
    @@ -1613,6 +1613,7 @@ static void __init free_area_init_core(s

    pgdat->nr_zones = 0;
    init_waitqueue_head(&pgdat->kswapd_wait);
    + atomic_set(&pgdat->kswapd_max_order, 0);

    for (j = 0; j < MAX_NR_ZONES; j++) {
    struct zone *zone = pgdat->node_zones + j;
    diff -puN include/linux/mmzone.h~vm-kswapd-heed-order-watermarks include/linux/mmzone.h
    --- linux-2.6/include/linux/mmzone.h~vm-kswapd-heed-order-watermarks 2004-09-05 15:12:04.000000000 +1000
    +++ linux-2.6-npiggin/include/linux/mmzone.h 2004-09-05 15:12:04.000000000 +1000
    @@ -263,8 +263,9 @@ typedef struct pglist_data {
    range, including holes */
    int node_id;
    struct pglist_data *pgdat_next;
    - wait_queue_head_t kswapd_wait;
    + wait_queue_head_t kswapd_wait;
    struct task_struct *kswapd;
    + atomic_t kswapd_max_order;
    } pg_data_t;

    #define node_present_pages(nid) (NODE_DATA(nid)->node_present_pages)
    @@ -278,7 +279,7 @@ void __get_zone_counts(unsigned long *ac
    void get_zone_counts(unsigned long *active, unsigned long *inactive,
    unsigned long *free);
    void build_all_zonelists(void);
    -void wakeup_kswapd(struct zone *zone);
    +void wakeup_kswapd(struct zone *zone, int order);
    int zone_watermark_ok(struct zone *z, int order, unsigned long mark,
    int alloc_type, int can_try_harder, int gfp_high);

    _
    \
     
     \ /
      Last update: 2005-03-22 14:05    [W:3.365 / U:0.104 seconds]
    ©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site