Skip to contentSkip to conversation

fix pagewalk minor issues

PATCH New KOSAKI Motohiro 1 participant 5 messages May 25, 2011

Subsystems (2)
Memory management MM core
lore
Share & tools
lore

KOSAKI Motohiro

author
#0 lore raw

cover fix pagewalk minor issues

While reviewing Stephen's numa_maps improvement, I've found recent pagewalk changes made some minor issues.

This series fix them. I think only [1/4] need to backport to -stable.

KOSAKI Motohiro (4):
pagewalk: Fix walk_page_range() don't check find_vma() result
properly
pagewalk: don't look up vma if walk->hugetlb_entry is unused
pagewalk: add locking-rule commnets
pagewalk: fix code comment for THP

 include/linux/mm.h |    1 +
 mm/pagewalk.c      |   49 ++++++++++++++++++++++++++++++++++++++++++-------
 2 files changed, 43 insertions(+), 7 deletions(-)
Signature · 1 line
1.7.3.1
author
#1 lore raw

PATCH 1/4 pagewalk: Fix walk_page_range() don't check find_vma() result properly

The doc of find_vma() says,

    /* Look up the first VMA which satisfies  addr < vm_end,  NULL if none. */
    struct vm_area_struct *find_vma(struct mm_struct *mm, unsigned long addr)
    {
     (snip)

Thus, caller should confirm whether the returned vma matches a desired one.

Signed-off-by KOSAKI Motohiro <[email protected]>
---
 mm/pagewalk.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)
mm/pagewalk.c +1 -1
index c3450d5..606bbb4 100644
@@ -176,7 +176,7 @@ int walk_page_range(unsigned long addr, unsigned long end,
 		 * we can't handled it in the same manner as non-huge pages.
 		 */
 		vma = find_vma(walk->mm, addr);
-		if (vma && is_vm_hugetlb_page(vma)) {
+		if (vma && vma->vm_start <= addr && is_vm_hugetlb_page(vma)) {
 			if (vma->vm_end < next)
 				next = vma->vm_end;
 			/*
Signature · 1 line
1.7.3.1
author
#2 lore raw

PATCH 2/4 pagewalk: don't look up vma if walk->hugetlb_entry is unused

In reply to KOSAKI Motohiro

Currently, walk_page_range() call find_vma() every page table walk iteration. but it's completely unnecessary if walk->hugetlb_entry is unused. And we don't have to assume find_vma() is lightweight operation. Thus this patch check walk->hugetlb_entry and avoid find_vma() call if possible.

This patch also makes some cleanups. 1) remove ugly uninitialized_var() and 2) #ifdef in function body.

Signed-off-by KOSAKI Motohiro <[email protected]>
---
 mm/pagewalk.c |   43 +++++++++++++++++++++++++++++++++++++------
 1 files changed, 37 insertions(+), 6 deletions(-)
mm/pagewalk.c +37 -6
index 606bbb4..ee4ff87 100644
@@ -126,7 +126,39 @@ static int walk_hugetlb_range(struct vm_area_struct *vma,
 
 	return 0;
 }
-#endif
+
+static struct vm_area_struct* hugetlb_vma(unsigned long addr, struct mm_walk *walk)
+{
+	struct vm_area_struct *vma;
+
+	/* We don't need vma lookup at all. */
+	if (!walk->hugetlb_entry)
+		return NULL;
+
+	VM_BUG_ON(!rwsem_is_locked(&walk->mm->mmap_sem));
+	vma = find_vma(walk->mm, addr);
+	if (vma && vma->vm_start <= addr && is_vm_hugetlb_page(vma))
+		return vma;
+
+	return NULL;
+}
+
+#else /* CONFIG_HUGETLB_PAGE */
+static struct vm_area_struct* hugetlb_vma(unsigned long addr, struct mm_walk *walk)
+{
+	return NULL;
+}
+
+static int walk_hugetlb_range(struct vm_area_struct *vma,
+			      unsigned long addr, unsigned long end,
+			      struct mm_walk *walk)
+{
+	return 0;
+}
+
+#endif /* CONFIG_HUGETLB_PAGE */
+
+
 
 /**
  * walk_page_range - walk a memory map's page tables with a callback

@@ -165,18 +197,17 @@ int walk_page_range(unsigned long addr, unsigned long end,
 
 	pgd = pgd_offset(walk->mm, addr);
 	do {
-		struct vm_area_struct *uninitialized_var(vma);
+		struct vm_area_struct *vma;
 
 		next = pgd_addr_end(addr, end);
 
-#ifdef CONFIG_HUGETLB_PAGE
 		/*
 		 * handle hugetlb vma individually because pagetable walk for
 		 * the hugetlb page is dependent on the architecture and
 		 * we can't handled it in the same manner as non-huge pages.
 		 */
-		vma = find_vma(walk->mm, addr);
-		if (vma && vma->vm_start <= addr && is_vm_hugetlb_page(vma)) {
+		vma = hugetlb_vma(addr, walk);
+		if (vma) {
 			if (vma->vm_end < next)
 				next = vma->vm_end;
 			/*

@@ -189,7 +220,7 @@ int walk_page_range(unsigned long addr, unsigned long end,
 			pgd = pgd_offset(walk->mm, next);
 			continue;
 		}
-#endif
+
 		if (pgd_none_or_clear_bad(pgd)) {
 			if (walk->pte_hole)
 				err = walk->pte_hole(addr, next, walk);
Signature · 10 lines
1.7.3.1



--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to [email protected].  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Fight unfair telecom internet charges in Canada: sign http://stopthemeter.ca/
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>
author
#3 lore raw

PATCH 3/4 pagewalk: add locking-rule commnets

In reply to KOSAKI Motohiro

Originally, walk_hugetlb_range() didn't require a caller take any lock. But commit d33b9f45bd (mm: hugetlb: fix hugepage memory leak in walk_page_range) changed its rule. Because it added find_vma() call in walk_hugetlb_range().

Any locking-rule change commit should write a doc too.

Signed-off-by KOSAKI Motohiro <[email protected]>
---
 include/linux/mm.h |    1 +
 mm/pagewalk.c      |    3 +++
 2 files changed, 4 insertions(+), 0 deletions(-)
include/linux/mm.h +1 -0
index dd87a78..7337b66 100644
@@ -921,6 +921,7 @@ unsigned long unmap_vmas(struct mmu_gather **tlb,
  * @pte_entry: if set, called for each non-empty PTE (4th-level) entry
  * @pte_hole: if set, called for each hole at all levels
  * @hugetlb_entry: if set, called for each hugetlb entry
+ *                 *Caution*: The caller must hold mmap_sem() if it's used.
  *
  * (see walk_page_range for more details)
  */
mm/pagewalk.c +3 -0
index ee4ff87..f792940 100644
@@ -181,6 +181,9 @@ static int walk_hugetlb_range(struct vm_area_struct *vma,
  *
  * If any callback returns a non-zero value, the walk is aborted and
  * the return value is propagated back to the caller. Otherwise 0 is returned.
+ *
+ * walk->mm->mmap_sem must be held for at least read if walk->hugetlb_entry
+ * is !NULL.
  */
 int walk_page_range(unsigned long addr, unsigned long end,
 		    struct mm_walk *walk)
Signature · 1 line
1.7.3.1
author
#4 lore raw

PATCH 4/4 pagewalk: fix code comment for THP

In reply to KOSAKI Motohiro

commit bae9c19bf1 (thp: split_huge_page_mm/vma) changed locking behavior of walk_page_range(). Thus this patch changes comment too.

Signed-off-by KOSAKI Motohiro <[email protected]>
---
 mm/pagewalk.c |    3 ++-
 1 files changed, 2 insertions(+), 1 deletions(-)
mm/pagewalk.c +2 -1
index f792940..2f5cf10 100644
@@ -176,7 +176,8 @@ static int walk_hugetlb_range(struct vm_area_struct *vma,
  * associated range, and a copy of the original mm_walk for access to
  * the ->private or ->mm fields.
  *
- * No locks are taken, but the bottom level iterator will map PTE
+ * Usually no locks are taken, but splitting transparent huge page may
+ * take page table lock. And the bottom level iterator will map PTE
  * directories from highmem if necessary.
  *
  * If any callback returns a non-zero value, the walk is aborted and
Signature · 10 lines
1.7.3.1



--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to [email protected].  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Fight unfair telecom internet charges in Canada: sign http://stopthemeter.ca/
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>