Skip to contentSkip to conversation

mm-handle-swap-page-faults-if-the-faulting-page-can-be-locked.patch removed from -mm tree

merged mm-stable Andrew Morton 2 participants 3 messages Apr 18, 2023

lore
Share & tools
lore

Andrew Morton

author tree notice Applied
#0 lore raw

The quilt patch titled
Subject: mm: handle swap page faults under vma lock if page is uncontended
has been removed from the -mm tree. Its filename was
mm-handle-swap-page-faults-if-the-faulting-page-can-be-locked.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: Suren Baghdasaryan <[email protected]>
Subject: mm: handle swap page faults under vma lock if page is uncontended
Date: Fri, 14 Apr 2023 11:00:43 -0700

When page fault is handled under VMA lock protection, all swap page faults are retried with mmap_lock because folio_lock_or_retry implementation has to drop and reacquire mmap_lock if folio could not be immediately locked.

Instead of retrying all swapped page faults, retry only when folio locking fails.

Cc Davidlohr Bueso <[email protected]>
Cc Jan Kara <[email protected]>
Cc Johannes Weiner <[email protected]>
Cc Josef Bacik <[email protected]>
Cc Laurent Dufour <[email protected]>
Cc Liam R. Howlett <[email protected]>
Cc Lorenzo Stoakes <[email protected]>
Cc Michal Hocko <[email protected]>
Cc Michel Lespinasse <[email protected]>
Cc Minchan Kim <[email protected]>
Cc Punit Agrawal <[email protected]>
Cc Vlastimil Babka <[email protected]>

---

 mm/filemap.c |    6 ++++++
 mm/memory.c  |    5 -----
 2 files changed, 6 insertions(+), 5 deletions(-)
mm/filemap.c renamed from mm/filemap.c~mm-handle-swap-page-faults-if-the-faulting-page-can-be-locked +6 -0
@@ -1706,6 +1706,8 @@ static int __folio_lock_async(struct fol
  *     mmap_lock has been released (mmap_read_unlock(), unless flags had both
  *     FAULT_FLAG_ALLOW_RETRY and FAULT_FLAG_RETRY_NOWAIT set, in
  *     which case mmap_lock is still held.
+ *     If flags had FAULT_FLAG_VMA_LOCK set, meaning the operation is performed
+ *     with VMA lock only, the VMA lock is still held.
  *
  * If neither ALLOW_RETRY nor KILLABLE are set, will always return true
  * with the folio locked and the mmap_lock unperturbed.
@@ -1713,6 +1715,10 @@ static int __folio_lock_async(struct fol
 bool __folio_lock_or_retry(struct folio *folio, struct mm_struct *mm,
 			 unsigned int flags)
 {
+	/* Can't do this if not holding mmap_lock */
+	if (flags & FAULT_FLAG_VMA_LOCK)
+		return false;
+
 	if (fault_flag_allow_retry_first(flags)) {
 		/*
 		 * CAUTION! In this case, mmap_lock is not released
mm/memory.c renamed from mm/memory.c~mm-handle-swap-page-faults-if-the-faulting-page-can-be-locked +0 -5
@@ -3711,11 +3711,6 @@ vm_fault_t do_swap_page(struct vm_fault
 	if (!pte_unmap_same(vmf))
 		goto out;
 
-	if (vmf->flags & FAULT_FLAG_VMA_LOCK) {
-		ret = VM_FAULT_RETRY;
-		goto out;
-	}
-
 	entry = pte_to_swp_entry(vmf->orig_pte);
 	if (unlikely(non_swap_entry(entry))) {
 		if (is_migration_entry(entry)) {

_

Patches currently in -mm which might be from [email protected] are

#1 lore raw
Quoted 9 lines · Andrew Morton

The quilt patch titled
Subject: mm: handle swap page faults under vma lock if page is uncontended
has been removed from the -mm tree. Its filename was
mm-handle-swap-page-faults-if-the-faulting-page-can-be-locked.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

Andrew, there are cases which are not properly handled in this patch, outlined here: https://lore.kernel.org/all/[email protected]/
Please drop it for now until I post the next version. It's not stable material yet. Thanks!

Quoted 76 lines

------------------------------------------------------
From: Suren Baghdasaryan <[email protected]>
Subject: mm: handle swap page faults under vma lock if page is uncontended
Date: Fri, 14 Apr 2023 11:00:43 -0700

When page fault is handled under VMA lock protection, all swap page faults are retried with mmap_lock because folio_lock_or_retry implementation has to drop and reacquire mmap_lock if folio could not be immediately locked.

Instead of retrying all swapped page faults, retry only when folio locking fails.

Cc Davidlohr Bueso <[email protected]>
Cc Jan Kara <[email protected]>
Cc Johannes Weiner <[email protected]>
Cc Josef Bacik <[email protected]>
Cc Laurent Dufour <[email protected]>
Cc Liam R. Howlett <[email protected]>
Cc Lorenzo Stoakes <[email protected]>
Cc Michal Hocko <[email protected]>
Cc Michel Lespinasse <[email protected]>
Cc Minchan Kim <[email protected]>
Cc Punit Agrawal <[email protected]>
Cc Vlastimil Babka <[email protected]>

---

 mm/filemap.c |    6 ++++++
 mm/memory.c  |    5 -----
 2 files changed, 6 insertions(+), 5 deletions(-)
mm/filemap.c renamed from mm/filemap.c~mm-handle-swap-page-faults-if-the-faulting-page-can-be-locked +6 -0
@@ -1706,6 +1706,8 @@ static int __folio_lock_async(struct fol
  *     mmap_lock has been released (mmap_read_unlock(), unless flags had both
  *     FAULT_FLAG_ALLOW_RETRY and FAULT_FLAG_RETRY_NOWAIT set, in
  *     which case mmap_lock is still held.
+ *     If flags had FAULT_FLAG_VMA_LOCK set, meaning the operation is performed
+ *     with VMA lock only, the VMA lock is still held.
  *
  * If neither ALLOW_RETRY nor KILLABLE are set, will always return true
  * with the folio locked and the mmap_lock unperturbed.
@@ -1713,6 +1715,10 @@ static int __folio_lock_async(struct fol
 bool __folio_lock_or_retry(struct folio *folio, struct mm_struct *mm,
                         unsigned int flags)
 {
+       /* Can't do this if not holding mmap_lock */
+       if (flags & FAULT_FLAG_VMA_LOCK)
+               return false;
+
        if (fault_flag_allow_retry_first(flags)) {
                /*
                 * CAUTION! In this case, mmap_lock is not released
mm/memory.c renamed from mm/memory.c~mm-handle-swap-page-faults-if-the-faulting-page-can-be-locked +0 -5
@@ -3711,11 +3711,6 @@ vm_fault_t do_swap_page(struct vm_fault
        if (!pte_unmap_same(vmf))
                goto out;
 
-       if (vmf->flags & FAULT_FLAG_VMA_LOCK) {
-               ret = VM_FAULT_RETRY;
-               goto out;
-       }
-
        entry = pte_to_swp_entry(vmf->orig_pte);
        if (unlikely(non_swap_entry(entry))) {
                if (is_migration_entry(entry)) {

_

Patches currently in -mm which might be from [email protected] are

#2 lore raw
Quoted 8 lines · On Tue, 18 Apr 2023 16:39:40 -0700 Suren Baghdasaryan <[email protected]> wrote:
Quoted 7 lines · Andrew Morton

The quilt patch titled
Subject: mm: handle swap page faults under vma lock if page is uncontended
has been removed from the -mm tree. Its filename was
mm-handle-swap-page-faults-if-the-faulting-page-can-be-locked.patch

Quoted 7 lines
Quoted 2 lines

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

Andrew, there are cases which are not properly handled in this patch, outlined here: https://lore.kernel.org/all/[email protected]/
Please drop it for now until I post the next version. It's not stable material yet. Thanks!

OK, this patch has been dropped.