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

[Ying Han <yinghan@google.com>]

On Sun, Mar 27, 2011 at 5:07 PM, KAMEZAWA Hiroyuki
<kamezawa.hiroyu@jp.fujitsu.com> wrote:
> On Sat, 26 Mar 2011 17:14:44 -0700
> 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"
>>
>> 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.
>>
>> 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 1069962
>> pgmajfault 553
>> total_pgfault 1069966
>> 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     0.67s   232.11s    14.68s   16892.130 267796.518
>>
>> $ ./ministat mmotm.txt mmotm_fault.txt
>> x mmotm.txt (w/o patch)
>> + mmotm_fault.txt (w/ patch)
>> +-------------------------------------------------------------------------+
>>     N           Min           Max        Median           Avg        Stddev
>> x  10     16682.962     17344.027     16913.524     16928.812      166.5362
>> +  10      16696.49      17480.09     16949.143     16951.448     223.56288
>> No difference proven at 95.0% confidence
>>
>> Signed-off-by: Ying Han <yinghan@google.com>
>
> Hmm, maybe useful ? (It's good to describe what is difference with PGPGIN)
> Especially, you should show why this is useful than per process pgfault count.
> What I thought of this, I thought that I need per-process information, finally...
> and didn't add this.
>
> Anyway, I have a request for the style of the function. (see below)

Thanks for your comment, and I will post V2 shortly.

>
>
>> ---
>>  Documentation/cgroups/memory.txt |    4 +++
>>  fs/ncpfs/mmap.c                  |    2 +
>>  include/linux/memcontrol.h       |   22 +++++++++++++++
>>  mm/filemap.c                     |    1 +
>>  mm/memcontrol.c                  |   54 ++++++++++++++++++++++++++++++++++++++
>>  mm/memory.c                      |    2 +
>>  mm/shmem.c                       |    1 +
>>  7 files changed, 86 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..adb3f45 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_pgmajfault_from_mm(area->vm_mm);
>
> Could you do this as  mem_cgroup_count_vm_event(area->vm_mm, PGMAJFAULT) ?

will be included in V2.


>
> <snip>
>
>> +void mem_cgroup_pgfault_from_mm(struct mm_struct *mm)
>> +{
>> +     struct mem_cgroup *mem;
>> +
>> +     if (!mm)
>> +             return;
>> +
>> +     rcu_read_lock();
>> +     mem = mem_cgroup_from_task(rcu_dereference(mm->owner));
>> +     if (unlikely(!mem))
>> +             goto out;
>> +     mem_cgroup_pgfault(mem, 1);
>> +
>> +out:
>> +     rcu_read_unlock();
>> +}
>> +
>> +void mem_cgroup_pgmajfault_from_mm(struct mm_struct *mm)
>> +{
>> +     struct mem_cgroup *mem;
>> +
>> +     if (!mm)
>> +             return;
>> +
>> +     rcu_read_lock();
>> +     mem = mem_cgroup_from_task(rcu_dereference(mm->owner));
>> +     if (unlikely(!mem))
>> +             goto out;
>> +     mem_cgroup_pgmajfault(mem, 1);
>> +out:
>> +     rcu_read_unlock();
>> +}
>> +EXPORT_SYMBOL(mem_cgroup_pgmajfault_from_mm);
>> +
>
> Then, you can do above 2 in a function.

--Ying
>
>
> Thanks,
> -Kame
>
>

--
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>