The patch titled
Subject: khugepaged: simplify khugepaged vs. __mmput
has been removed from the -mm tree. Its filename was
khugepaged-simplify-khugepaged-vs-__mmput.patch
This patch was dropped because it was withdrawn
------------------------------------------------------
From: Michal Hocko <mhocko@xxxxxxxx>
Subject: khugepaged: simplify khugepaged vs. __mmput
__khugepaged_exit is called during the final __mmput and it employs a
complex synchronization dances to make sure it doesn't race with the
khugepaged which might be scanning this mm at the same time. This is all
caused by the fact that khugepaged doesn't pin mm_users. Things would
simplify considerably if we simply check the mm at khugepaged_scan_mm_slot
and if mm_users was already 0 then we know it is dead and we can unhash
the mm_slot and move on to another one. This will also guarantee that
__khugepaged_exit cannot race with khugepaged and so we can free up the
slot if it is still hashed.
Link: http://lkml.kernel.org/r/20160603134934.GJ20676@xxxxxxxxxxxxxx
Signed-off-by: Michal Hocko <mhocko@xxxxxxxx>
Tested-by: Sergey Senozhatsky <sergey.senozhatsky.work@xxxxxxxxx>
Cc: Andrea Arcangeli <aarcange@xxxxxxxxxx>
Cc: Sergey Senozhatsky <sergey.senozhatsky.work@xxxxxxxxx>
Cc: Vlastimil Babka <vbabka@xxxxxxx>
Cc: "Kirill A. Shutemov" <kirill.shutemov@xxxxxxxxxxxxxxx>
Cc: Stephen Rothwell <sfr@xxxxxxxxxxxxxxxx>
Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
---
mm/huge_memory.c | 90 +++++++++++++++++++++------------------------
1 file changed, 42 insertions(+), 48 deletions(-)
diff -puN mm/huge_memory.c~khugepaged-simplify-khugepaged-vs-__mmput mm/huge_memory.c
--- a/mm/huge_memory.c~khugepaged-simplify-khugepaged-vs-__mmput
+++ a/mm/huge_memory.c
@@ -1937,7 +1937,8 @@ static void insert_to_mm_slots_hash(stru
static inline int khugepaged_test_exit(struct mm_struct *mm)
{
- return atomic_read(&mm->mm_users) == 0;
+ /* the only pin is from khugepaged_scan_mm_slot */
+ return atomic_read(&mm->mm_users) <= 1;
}
int __khugepaged_enter(struct mm_struct *mm)
@@ -1949,8 +1950,6 @@ int __khugepaged_enter(struct mm_struct
if (!mm_slot)
return -ENOMEM;
- /* __khugepaged_exit() must not run from under us */
- VM_BUG_ON_MM(khugepaged_test_exit(mm), mm);
if (unlikely(test_and_set_bit(MMF_VM_HUGEPAGE, &mm->flags))) {
free_mm_slot(mm_slot);
return 0;
@@ -1993,36 +1992,43 @@ int khugepaged_enter_vma_merge(struct vm
return 0;
}
-void __khugepaged_exit(struct mm_struct *mm)
+static void collect_mm_slot(struct mm_slot *mm_slot)
{
- struct mm_slot *mm_slot;
- int free = 0;
+ struct mm_struct *mm = mm_slot->mm;
- spin_lock(&khugepaged_mm_lock);
- mm_slot = get_mm_slot(mm);
- if (mm_slot && khugepaged_scan.mm_slot != mm_slot) {
+ VM_BUG_ON(NR_CPUS != 1 && !spin_is_locked(&khugepaged_mm_lock));
+
+ if (khugepaged_test_exit(mm)) {
+ /* free mm_slot */
hash_del(&mm_slot->hash);
list_del(&mm_slot->mm_node);
- free = 1;
- }
- spin_unlock(&khugepaged_mm_lock);
- if (free) {
- clear_bit(MMF_VM_HUGEPAGE, &mm->flags);
- free_mm_slot(mm_slot);
- mmdrop(mm);
- } else if (mm_slot) {
/*
- * This is required to serialize against
- * khugepaged_test_exit() (which is guaranteed to run
- * under mmap sem read mode). Stop here (after we
- * return all pagetables will be destroyed) until
- * khugepaged has finished working on the pagetables
- * under the mmap_sem.
+ * Not strictly needed because the mm exited already.
+ *
+ * clear_bit(MMF_VM_HUGEPAGE, &mm->flags);
*/
- down_write(&mm->mmap_sem);
- up_write(&mm->mmap_sem);
+
+ /* khugepaged_mm_lock actually not necessary for the below */
+ free_mm_slot(mm_slot);
+ mmdrop(mm);
+
+ if (khugepaged_scan.mm_slot == mm_slot)
+ khugepaged_scan.mm_slot = NULL;
+ }
+}
+
+void __khugepaged_exit(struct mm_struct *mm)
+{
+ struct mm_slot *mm_slot;
+
+ spin_lock(&khugepaged_mm_lock);
+ mm_slot = get_mm_slot(mm);
+ if (mm_slot) {
+ collect_mm_slot(mm_slot);
+ clear_bit(MMF_VM_HUGEPAGE, &mm->flags);
}
+ spin_unlock(&khugepaged_mm_lock);
}
static void release_pte_page(struct page *page)
@@ -2742,29 +2748,6 @@ out:
return ret;
}
-static void collect_mm_slot(struct mm_slot *mm_slot)
-{
- struct mm_struct *mm = mm_slot->mm;
-
- VM_BUG_ON(NR_CPUS != 1 && !spin_is_locked(&khugepaged_mm_lock));
-
- if (khugepaged_test_exit(mm)) {
- /* free mm_slot */
- hash_del(&mm_slot->hash);
- list_del(&mm_slot->mm_node);
-
- /*
- * Not strictly needed because the mm exited already.
- *
- * clear_bit(MMF_VM_HUGEPAGE, &mm->flags);
- */
-
- /* khugepaged_mm_lock actually not necessary for the below */
- free_mm_slot(mm_slot);
- mmdrop(mm);
- }
-}
-
static unsigned int khugepaged_scan_mm_slot(unsigned int pages,
struct page **hpage)
__releases(&khugepaged_mm_lock)
@@ -2786,6 +2769,16 @@ static unsigned int khugepaged_scan_mm_s
khugepaged_scan.address = 0;
khugepaged_scan.mm_slot = mm_slot;
}
+
+ /*
+ * Do not even try to do anything if the current mm is already
+ * dead. khugepaged_mm_lock will make sure only this or
+ * __khugepaged_exit does the unhasing.
+ */
+ if (!atomic_inc_not_zero(&mm_slot->mm->mm_users)) {
+ collect_mm_slot(mm_slot);
+ return progress;
+ }
spin_unlock(&khugepaged_mm_lock);
mm = mm_slot->mm;
@@ -2869,6 +2862,7 @@ breakouterloop_mmap_sem:
collect_mm_slot(mm_slot);
}
+ mmput_async(mm);
return progress;
}
_
Patches currently in -mm which might be from mhocko@xxxxxxxx are
tree-wide-get-rid-of-__gfp_repeat-for-order-0-allocations-part-i.patch
x86-get-rid-of-superfluous-__gfp_repeat.patch
x86-efi-get-rid-of-superfluous-__gfp_repeat.patch
arm-get-rid-of-superfluous-__gfp_repeat.patch
arm64-get-rid-of-superfluous-__gfp_repeat.patch
arc-get-rid-of-superfluous-__gfp_repeat.patch
mips-get-rid-of-superfluous-__gfp_repeat.patch
nios2-get-rid-of-superfluous-__gfp_repeat.patch
parisc-get-rid-of-superfluous-__gfp_repeat.patch
score-get-rid-of-superfluous-__gfp_repeat.patch
powerpc-get-rid-of-superfluous-__gfp_repeat.patch
sparc-get-rid-of-superfluous-__gfp_repeat.patch
s390-get-rid-of-superfluous-__gfp_repeat.patch
sh-get-rid-of-superfluous-__gfp_repeat.patch
tile-get-rid-of-superfluous-__gfp_repeat.patch
unicore32-get-rid-of-superfluous-__gfp_repeat.patch
jbd2-get-rid-of-superfluous-__gfp_repeat.patch
--
To unsubscribe from this list: send the line "unsubscribe mm-commits" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
