[merged mm-stable] sched-numa-implement-access-pid-reset-logic.patch removed from -mm tree

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



The quilt patch titled
     Subject: sched/numa: implement access PID reset logic
has been removed from the -mm tree.  Its filename was
     sched-numa-implement-access-pid-reset-logic.patch

This patch was dropped because it was merged into the mm-stable branch
of git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm

------------------------------------------------------
From: Raghavendra K T <raghavendra.kt@xxxxxxx>
Subject: sched/numa: implement access PID reset logic
Date: Wed, 1 Mar 2023 17:49:02 +0530

This helps to ensure that only recently accessed PIDs scan the VMAs.

Current implementation: (idea supported by PeterZ)

 1. Accessing PID information is maintained in two windows. 
    access_pids[1] being newest.

 2. Reset old access PID info i.e.  access_pid[0] every (4 *
    sysctl_numa_balancing_scan_delay) interval after initial scan delay
    period expires.

The above interval seemed to be experimentally optimum since it avoids
frequent reset of access info as well as helps clearing the old access
info regularly.  The reset logic is implemented in scan path.

Link: https://lkml.kernel.org/r/f7a675f66d1442d048b4216b2baf94515012c405.1677672277.git.raghavendra.kt@xxxxxxx
Signed-off-by: Raghavendra K T <raghavendra.kt@xxxxxxx>
Suggested-by: Mel Gorman <mgorman@xxxxxxxxxxxxxxxxxxx>
Cc: Bharata B Rao <bharata@xxxxxxx>
Cc: David Hildenbrand <david@xxxxxxxxxx>
Cc: Disha Talreja <dishaa.talreja@xxxxxxx>
Cc: Ingo Molnar <mingo@xxxxxxxxxx>
Cc: Mike Rapoport <rppt@xxxxxxxxxx>
Cc: Peter Zijlstra <peterz@xxxxxxxxxxxxx>
Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
---

 include/linux/mm.h       |    4 ++--
 include/linux/mm_types.h |    3 ++-
 kernel/sched/fair.c      |   23 +++++++++++++++++++++--
 3 files changed, 25 insertions(+), 5 deletions(-)

--- a/include/linux/mm.h~sched-numa-implement-access-pid-reset-logic
+++ a/include/linux/mm.h
@@ -1692,8 +1692,8 @@ static inline void vma_set_access_pid_bi
 	unsigned int pid_bit;
 
 	pid_bit = current->pid % BITS_PER_LONG;
-	if (vma->numab_state && !test_bit(pid_bit, &vma->numab_state->access_pids)) {
-		__set_bit(pid_bit, &vma->numab_state->access_pids);
+	if (vma->numab_state && !test_bit(pid_bit, &vma->numab_state->access_pids[1])) {
+		__set_bit(pid_bit, &vma->numab_state->access_pids[1]);
 	}
 }
 #else /* !CONFIG_NUMA_BALANCING */
--- a/include/linux/mm_types.h~sched-numa-implement-access-pid-reset-logic
+++ a/include/linux/mm_types.h
@@ -477,7 +477,8 @@ struct vma_lock {
 
 struct vma_numab_state {
 	unsigned long next_scan;
-	unsigned long access_pids;
+	unsigned long next_pid_reset;
+	unsigned long access_pids[2];
 };
 
 /*
--- a/kernel/sched/fair.c~sched-numa-implement-access-pid-reset-logic
+++ a/kernel/sched/fair.c
@@ -2930,6 +2930,7 @@ static void reset_ptenuma_scan(struct ta
 
 static bool vma_is_accessed(struct vm_area_struct *vma)
 {
+	unsigned long pids;
 	/*
 	 * Allow unconditional access first two times, so that all the (pages)
 	 * of VMAs get prot_none fault introduced irrespective of accesses.
@@ -2939,10 +2940,12 @@ static bool vma_is_accessed(struct vm_ar
 	if (READ_ONCE(current->mm->numa_scan_seq) < 2)
 		return true;
 
-	return test_bit(current->pid % BITS_PER_LONG,
-				&vma->numab_state->access_pids);
+	pids = vma->numab_state->access_pids[0] | vma->numab_state->access_pids[1];
+	return test_bit(current->pid % BITS_PER_LONG, &pids);
 }
 
+#define VMA_PID_RESET_PERIOD (4 * sysctl_numa_balancing_scan_delay)
+
 /*
  * The expensive part of numa migration is done from task_work context.
  * Triggered from task_tick_numa().
@@ -3051,6 +3054,10 @@ static void task_numa_work(struct callba
 
 			vma->numab_state->next_scan = now +
 				msecs_to_jiffies(sysctl_numa_balancing_scan_delay);
+
+			/* Reset happens after 4 times scan delay of scan start */
+			vma->numab_state->next_pid_reset =  vma->numab_state->next_scan +
+				msecs_to_jiffies(VMA_PID_RESET_PERIOD);
 		}
 
 		/*
@@ -3065,6 +3072,18 @@ static void task_numa_work(struct callba
 		if (!vma_is_accessed(vma))
 			continue;
 
+		/*
+		 * RESET access PIDs regularly for old VMAs. Resetting after checking
+		 * vma for recent access to avoid clearing PID info before access..
+		 */
+		if (mm->numa_scan_seq &&
+				time_after(jiffies, vma->numab_state->next_pid_reset)) {
+			vma->numab_state->next_pid_reset = vma->numab_state->next_pid_reset +
+				msecs_to_jiffies(VMA_PID_RESET_PERIOD);
+			vma->numab_state->access_pids[0] = READ_ONCE(vma->numab_state->access_pids[1]);
+			vma->numab_state->access_pids[1] = 0;
+		}
+
 		do {
 			start = max(start, vma->vm_start);
 			end = ALIGN(start + (pages << PAGE_SHIFT), HPAGE_SIZE);
_

Patches currently in -mm which might be from raghavendra.kt@xxxxxxx are





[Index of Archives]     [Kernel Archive]     [IETF Annouce]     [DCCP]     [Netdev]     [Networking]     [Security]     [Bugtraq]     [Yosemite]     [MIPS Linux]     [ARM Linux]     [Linux Security]     [Linux RAID]     [Linux SCSI]

  Powered by Linux