Re: [PATCH V2] Add the pagefault count into memcg stats

[Minchan Kim <minchan.kim@gmail.com>]

On Mon, Mar 28, 2011 at 12:26 PM, Ying Han <yinghan@google.com> wrote:
> Two new stats in per-memcg memory.stat which tracks the number of
> page faults and number of major page faults.
>
> "pgfault"
> "pgmajfault"
>
> They are different from "pgpgin"/"pgpgout" stat which count number of
> pages charged/discharged to the cgroup and have no meaning of reading/
> writing page to disk.
>
> It is valuable to track the two stats for both measuring application's
> performance as well as the efficiency of the kernel page reclaim path.
> Counting pagefaults per process is useful, but we also need the aggregated
> value since processes are monitored and controlled in cgroup basis in memcg.
>
> Functional test: check the total number of pgfault/pgmajfault of all
> memcgs and compare with global vmstat value:
>
> $ cat /proc/vmstat | grep fault
> pgfault 1070751
> pgmajfault 553
>
> $ cat /dev/cgroup/memory.stat | grep fault
> pgfault 1071138
> pgmajfault 553
> total_pgfault 1071142
> total_pgmajfault 553
>
> $ cat /dev/cgroup/A/memory.stat | grep fault
> pgfault 199
> pgmajfault 0
> total_pgfault 199
> total_pgmajfault 0
>
> Performance test: run page fault test(pft) wit 16 thread on faulting in 15G
> anon pages in 16G container. There is no regression noticed on the "flt/cpu/s"
>
> Sample output from pft:
> TAG pft:anon-sys-default:
>  Gb  Thr CLine   User     System     Wall    flt/cpu/s fault/wsec
>  15   16   1     0.69s   230.99s    14.62s   16972.539 268876.196
>
> +-------------------------------------------------------------------------+
>    N           Min           Max        Median           Avg        Stddev
> x  10     16682.962     17344.027     16913.524     16928.812      166.5362
> +  10      16718.92     17023.453     16907.164     16902.399     88.468851
> No difference proven at 95.0% confidence
>
> Signed-off-by: Ying Han <yinghan@google.com>
> ---
>  Documentation/cgroups/memory.txt |    4 +++
>  fs/ncpfs/mmap.c                  |    2 +
>  include/linux/memcontrol.h       |   18 ++++++++++++++
>  mm/filemap.c                     |    1 +
>  mm/memcontrol.c                  |   47 ++++++++++++++++++++++++++++++++++++++
>  mm/memory.c                      |    2 +
>  mm/shmem.c                       |    2 +
>  7 files changed, 76 insertions(+), 0 deletions(-)
>
> diff --git a/Documentation/cgroups/memory.txt b/Documentation/cgroups/memory.txt
> index b6ed61c..2db6103 100644
> --- a/Documentation/cgroups/memory.txt
> +++ b/Documentation/cgroups/memory.txt
> @@ -385,6 +385,8 @@ mapped_file - # of bytes of mapped file (includes tmpfs/shmem)
>  pgpgin         - # of pages paged in (equivalent to # of charging events).
>  pgpgout                - # of pages paged out (equivalent to # of uncharging events).
>  swap           - # of bytes of swap usage
> +pgfault                - # of page faults.
> +pgmajfault     - # of major page faults.
>  inactive_anon  - # of bytes of anonymous memory and swap cache memory on
>                LRU list.
>  active_anon    - # of bytes of anonymous and swap cache memory on active
> @@ -406,6 +408,8 @@ total_mapped_file   - sum of all children's "cache"
>  total_pgpgin           - sum of all children's "pgpgin"
>  total_pgpgout          - sum of all children's "pgpgout"
>  total_swap             - sum of all children's "swap"
> +total_pgfault          - sum of all children's "pgfault"
> +total_pgmajfault       - sum of all children's "pgmajfault"
>  total_inactive_anon    - sum of all children's "inactive_anon"
>  total_active_anon      - sum of all children's "active_anon"
>  total_inactive_file    - sum of all children's "inactive_file"
> diff --git a/fs/ncpfs/mmap.c b/fs/ncpfs/mmap.c
> index a7c07b4..e5d71b2 100644
> --- a/fs/ncpfs/mmap.c
> +++ b/fs/ncpfs/mmap.c
> @@ -16,6 +16,7 @@
>  #include <linux/mman.h>
>  #include <linux/string.h>
>  #include <linux/fcntl.h>
> +#include <linux/memcontrol.h>
>
>  #include <asm/uaccess.h>
>  #include <asm/system.h>
> @@ -92,6 +93,7 @@ static int ncp_file_mmap_fault(struct vm_area_struct *area,
>         * -- wli
>         */
>        count_vm_event(PGMAJFAULT);
> +       mem_cgroup_count_vm_event(area->vm_mm, PGMAJFAULT);
>        return VM_FAULT_MAJOR;
>  }
>
> diff --git a/include/linux/memcontrol.h b/include/linux/memcontrol.h
> index 5a5ce70..45e5268 100644
> --- a/include/linux/memcontrol.h
> +++ b/include/linux/memcontrol.h
> @@ -24,6 +24,7 @@ struct mem_cgroup;
>  struct page_cgroup;
>  struct page;
>  struct mm_struct;
> +enum vm_event_item;
>
>  /* Stats that can be updated by kernel. */
>  enum mem_cgroup_page_stat_item {
> @@ -147,6 +148,9 @@ unsigned long mem_cgroup_soft_limit_reclaim(struct zone *zone, int order,
>                                                gfp_t gfp_mask);
>  u64 mem_cgroup_get_limit(struct mem_cgroup *mem);
>
> +void mem_cgroup_pgfault(struct mem_cgroup *memcg, int val);
> +void mem_cgroup_pgmajfault(struct mem_cgroup *memcg, int val);

Do we have to expose above two functions?
Isn't it enough to only mem_cgroup_count_vm_event?

> +void mem_cgroup_count_vm_event(struct mm_struct *mm, enum vm_event_item idx);


-- 
Kind regards,
Minchan Kim

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Fight unfair telecom internet charges in Canada: sign http://stopthemeter.ca/
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>