The quilt patch titled
Subject: mm/khugepaged: bypassing unnecessary scans with MMF_DISABLE_THP check
has been removed from the -mm tree. Its filename was
mm-khugepaged-bypassing-unnecessary-scans-with-mmf_disable_thp-check.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: Lance Yang <ioworker0@xxxxxxxxx>
Subject: mm/khugepaged: bypassing unnecessary scans with MMF_DISABLE_THP check
Date: Mon, 29 Jan 2024 13:45:51 +0800
khugepaged scans the entire address space in the background for each
given mm, looking for opportunities to merge sequences of basic pages
into huge pages. However, when an mm is inserted to the mm_slots list,
and the MMF_DISABLE_THP flag is set later, this scanning process
becomes unnecessary for that mm and can be skipped to avoid redundant
operations, especially in scenarios with a large address space.
On an Intel Core i5 CPU, the time taken by khugepaged to scan the
address space of the process, which has been set with the
MMF_DISABLE_THP flag after being added to the mm_slots list, is as
follows (shorter is better):
VMA Count | Old | New | Change
---------------------------------------
50 | 23us | 9us | -60.9%
100 | 32us | 9us | -71.9%
200 | 44us | 9us | -79.5%
400 | 75us | 9us | -88.0%
800 | 98us | 9us | -90.8%
Once the count of VMAs for the process exceeds page_to_scan, khugepaged
needs to wait for scan_sleep_millisecs ms before scanning the next
process. IMO, unnecessary scans could actually be skipped with a very
inexpensive mm->flags check in this case.
This commit introduces a check before each scanning process to test the
MMF_DISABLE_THP flag for the given mm; if the flag is set, the scanning
process is bypassed, thereby improving the efficiency of khugepaged.
This optimization is not a correctness issue but rather an enhancement
to save expensive checks on each VMA when userspace cannot prctl itself
before spawning into the new process.
On some servers within our company, we deploy a daemon responsible for
monitoring and updating local applications. Some applications prefer
not to use THP, so the daemon calls prctl to disable THP before
fork/exec. Conversely, for other applications, the daemon calls prctl
to enable THP before fork/exec.
Ideally, the daemon should invoke prctl after the fork, but its current
implementation follows the described approach. In the Go standard
library, there is no direct encapsulation of the fork system call;
instead, fork and execve are combined into one through
syscall.ForkExec.
Link: https://lkml.kernel.org/r/20240129054551.57728-1-ioworker0@xxxxxxxxx
Signed-off-by: Lance Yang <ioworker0@xxxxxxxxx>
Acked-by: David Hildenbrand <david@xxxxxxxxxx>
Cc: Michal Hocko <mhocko@xxxxxxxx>
Cc: Minchan Kim <minchan@xxxxxxxxxx>
Cc: Muchun Song <songmuchun@xxxxxxxxxxxxx>
Cc: Peter Xu <peterx@xxxxxxxxxx>
Cc: Zach O'Keefe <zokeefe@xxxxxxxxxx>
Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
---
mm/khugepaged.c | 18 ++++++++++++------
1 file changed, 12 insertions(+), 6 deletions(-)
--- a/mm/khugepaged.c~mm-khugepaged-bypassing-unnecessary-scans-with-mmf_disable_thp-check
+++ a/mm/khugepaged.c
@@ -410,6 +410,12 @@ static inline int hpage_collapse_test_ex
return atomic_read(&mm->mm_users) == 0;
}
+static inline int hpage_collapse_test_exit_or_disable(struct mm_struct *mm)
+{
+ return hpage_collapse_test_exit(mm) ||
+ test_bit(MMF_DISABLE_THP, &mm->flags);
+}
+
void __khugepaged_enter(struct mm_struct *mm)
{
struct khugepaged_mm_slot *mm_slot;
@@ -1422,7 +1428,7 @@ static void collect_mm_slot(struct khuge
lockdep_assert_held(&khugepaged_mm_lock);
- if (hpage_collapse_test_exit(mm)) {
+ if (hpage_collapse_test_exit_or_disable(mm)) {
/* free mm_slot */
hash_del(&slot->hash);
list_del(&slot->mm_node);
@@ -2360,7 +2366,7 @@ static unsigned int khugepaged_scan_mm_s
goto breakouterloop_mmap_lock;
progress++;
- if (unlikely(hpage_collapse_test_exit(mm)))
+ if (unlikely(hpage_collapse_test_exit_or_disable(mm)))
goto breakouterloop;
vma_iter_init(&vmi, mm, khugepaged_scan.address);
@@ -2368,7 +2374,7 @@ static unsigned int khugepaged_scan_mm_s
unsigned long hstart, hend;
cond_resched();
- if (unlikely(hpage_collapse_test_exit(mm))) {
+ if (unlikely(hpage_collapse_test_exit_or_disable(mm))) {
progress++;
break;
}
@@ -2390,7 +2396,7 @@ skip:
bool mmap_locked = true;
cond_resched();
- if (unlikely(hpage_collapse_test_exit(mm)))
+ if (unlikely(hpage_collapse_test_exit_or_disable(mm)))
goto breakouterloop;
VM_BUG_ON(khugepaged_scan.address < hstart ||
@@ -2408,7 +2414,7 @@ skip:
fput(file);
if (*result == SCAN_PTE_MAPPED_HUGEPAGE) {
mmap_read_lock(mm);
- if (hpage_collapse_test_exit(mm))
+ if (hpage_collapse_test_exit_or_disable(mm))
goto breakouterloop;
*result = collapse_pte_mapped_thp(mm,
khugepaged_scan.address, false);
@@ -2450,7 +2456,7 @@ breakouterloop_mmap_lock:
* Release the current mm_slot if this mm is about to die, or
* if we scanned all vmas of this mm.
*/
- if (hpage_collapse_test_exit(mm) || !vma) {
+ if (hpage_collapse_test_exit_or_disable(mm) || !vma) {
/*
* Make sure that if mm_users is reaching zero while
* khugepaged runs here, khugepaged_exit will find
_
Patches currently in -mm which might be from ioworker0@xxxxxxxxx are
