[v1] ksm: add ksm involvement information for each process

[PATCH linux-next] ksm: add ksm involvement information for each process

Posted by xu.xin16@zte.com.cn 1 year, 9 months ago

From: xu xin <xu.xin16@zte.com.cn>

In /proc/<pid>/ksm_stat, Add two extra ksm involvement items including
MMF_VM_MERGEABLE and MMF_VM_MERGE_ANY. It helps administrators to
better know the system's KSM behavior at process level.

MMF_VM_MERGEABLE: yes/no
	whether a process'mm is added by madvise() into the candidate list
	of KSM or not.
MMF_VM_MERGE_ANY: yes/no
	whether a process'mm is added by prctl at process level into the
candidate list of KSM or not.

Signed-off-by: xu xin <xu.xin16@zte.com.cn>
---
 fs/proc/base.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/fs/proc/base.c b/fs/proc/base.c
index 18550c071d71..421594b8510c 100644
--- a/fs/proc/base.c
+++ b/fs/proc/base.c
@@ -3217,6 +3217,10 @@ static int proc_pid_ksm_stat(struct seq_file *m, struct pid_namespace *ns,
 		seq_printf(m, "ksm_zero_pages %lu\n", mm->ksm_zero_pages);
 		seq_printf(m, "ksm_merging_pages %lu\n", mm->ksm_merging_pages);
 		seq_printf(m, "ksm_process_profit %ld\n", ksm_process_profit(mm));
+		seq_printf(m, "MMF_VM_MERGEABLE: %s\n",
+				test_bit(MMF_VM_MERGEABLE, &mm->flags) ? "yes" : "no");
+		seq_printf(m, "MMF_VM_MERGE_ANY: %s\n",
+				test_bit(MMF_VM_MERGE_ANY, &mm->flags) ? "yes" : "no");
 		mmput(mm);
 	}

-- 
2.15.2

Re: [PATCH linux-next] ksm: add ksm involvement information for each process

Posted by David Hildenbrand 1 year, 9 months ago

On 25.04.24 14:49, xu.xin16@zte.com.cn wrote:
> From: xu xin <xu.xin16@zte.com.cn>
> 
> In /proc/<pid>/ksm_stat, Add two extra ksm involvement items including
> MMF_VM_MERGEABLE and MMF_VM_MERGE_ANY. It helps administrators to
> better know the system's KSM behavior at process level.
> 
> MMF_VM_MERGEABLE: yes/no
> 	whether a process'mm is added by madvise() into the candidate list
> 	of KSM or not.
> MMF_VM_MERGE_ANY: yes/no
> 	whether a process'mm is added by prctl at process level into the
> candidate list of KSM or not.
> 
> Signed-off-by: xu xin <xu.xin16@zte.com.cn>
> ---
>   fs/proc/base.c | 4 ++++
>   1 file changed, 4 insertions(+)
> 
> diff --git a/fs/proc/base.c b/fs/proc/base.c
> index 18550c071d71..421594b8510c 100644
> --- a/fs/proc/base.c
> +++ b/fs/proc/base.c
> @@ -3217,6 +3217,10 @@ static int proc_pid_ksm_stat(struct seq_file *m, struct pid_namespace *ns,
>   		seq_printf(m, "ksm_zero_pages %lu\n", mm->ksm_zero_pages);
>   		seq_printf(m, "ksm_merging_pages %lu\n", mm->ksm_merging_pages);
>   		seq_printf(m, "ksm_process_profit %ld\n", ksm_process_profit(mm));
> +		seq_printf(m, "MMF_VM_MERGEABLE: %s\n",
> +				test_bit(MMF_VM_MERGEABLE, &mm->flags) ? "yes" : "no");
> +		seq_printf(m, "MMF_VM_MERGE_ANY: %s\n",
> +				test_bit(MMF_VM_MERGE_ANY, &mm->flags) ? "yes" : "no");

Not sure if exposing these internal flag names is appropriate. Better 
describe what they do.

-- 
Cheers,

David / dhildenb