Skip to contentSkip to conversation

mm: softdirty: addresses before VMAs in PTE holes aren't softdirty

PATCH Superseded Peter Feiner 1 participant 2 messages Sep 10, 2014 – Sep 15, 2014

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 mm: softdirty: addresses before VMAs in PTE holes aren't softdirty

In PTE holes that contain VM_SOFTDIRTY VMAs, unmapped addresses before VM_SOFTDIRTY VMAs are reported as softdirty by /proc/pid/pagemap. This bug was introduced in 68b5a652485682f67eacdee3deae640fb7845b63. The aforementioned patch made /proc/pid/pagemap look at VM_SOFTDIRTY in PTE holes but neglected to observe the start of VMAs returned by find_vma.

Tested:
Wrote a selftest that creates a PMD-sized VMA then unmaps the first page and asserts that the page is not softdirty. I'm going to send the pagemap selftest in a later commit.

---
 fs/proc/task_mmu.c | 22 ++++++++++++++++------
 1 file changed, 16 insertions(+), 6 deletions(-)
fs/proc/task_mmu.c +16 -6
index dfc791c..256dbe9 100644
@@ -931,16 +931,26 @@ static int pagemap_pte_hole(unsigned long start, unsigned long end,
 	while (addr < end) {
 		struct vm_area_struct *vma = find_vma(walk->mm, addr);
 		pagemap_entry_t pme = make_pme(PM_NOT_PRESENT(pm->v2));
-		unsigned long vm_end;
+		unsigned long vm_start = end;
+		unsigned long vm_end = end;
+		unsigned long vm_flags = 0;
 
-		if (!vma) {
-			vm_end = end;
-		} else {
+		if (vma) {
+			vm_start = min(end, vma->vm_start);
 			vm_end = min(end, vma->vm_end);
-			if (vma->vm_flags & VM_SOFTDIRTY)
-				pme.pme |= PM_STATUS2(pm->v2, __PM_SOFT_DIRTY);
+			vm_flags = vma->vm_flags;
+		}
+
+		/* Addresses before the VMA. */
+		for (; addr < vm_start; addr += PAGE_SIZE) {
+			err = add_to_pagemap(addr, &pme, pm);
+			if (err)
+				goto out;
 		}
 
+		/* Addresses in the VMA. */
+		if (vm_flags & VM_SOFTDIRTY)
+			pme.pme |= PM_STATUS2(pm->v2, __PM_SOFT_DIRTY);
 		for (; addr < vm_end; addr += PAGE_SIZE) {
 			err = add_to_pagemap(addr, &pme, pm);
 			if (err)
Signature · 7 lines
2.1.0.rc2.206.gedb03e5

--
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/ .
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>
#1 lore raw

PATCH v2 1/1 mm: softdirty: addresses before VMAs in PTE holes aren't softdirty

In PTE holes that contain VM_SOFTDIRTY VMAs, unmapped addresses before VM_SOFTDIRTY VMAs are reported as softdirty by /proc/pid/pagemap. This bug was introduced in 68b5a652485682f67eacdee3deae640fb7845b63. The aforementioned patch made /proc/pid/pagemap look at VM_SOFTDIRTY in PTE holes but neglected to observe the start of VMAs returned by find_vma.

Tested:
Wrote a selftest that creates a PMD-sized VMA then unmaps the first page and asserts that the page is not softdirty. I'm going to send the pagemap selftest in a later commit.

---

v1 -> v2:
Rewrote to make logic easier to follow.

---
 fs/proc/task_mmu.c | 27 ++++++++++++++++++---------
 1 file changed, 18 insertions(+), 9 deletions(-)
fs/proc/task_mmu.c +18 -9
index dfc791c..c341568 100644
@@ -931,23 +931,32 @@ static int pagemap_pte_hole(unsigned long start, unsigned long end,
 	while (addr < end) {
 		struct vm_area_struct *vma = find_vma(walk->mm, addr);
 		pagemap_entry_t pme = make_pme(PM_NOT_PRESENT(pm->v2));
-		unsigned long vm_end;
+		/* End of address space hole, which we mark as non-present. */
+		unsigned long hole_end;
 
-		if (!vma) {
-			vm_end = end;
-		} else {
-			vm_end = min(end, vma->vm_end);
-			if (vma->vm_flags & VM_SOFTDIRTY)
-				pme.pme |= PM_STATUS2(pm->v2, __PM_SOFT_DIRTY);
+		if (vma)
+			hole_end = min(end, vma->vm_start);
+		else
+			hole_end = end;
+
+		for (; addr < hole_end; addr += PAGE_SIZE) {
+			err = add_to_pagemap(addr, &pme, pm);
+			if (err)
+				goto out;
 		}
 
-		for (; addr < vm_end; addr += PAGE_SIZE) {
+		if (!vma)
+			break;
+
+		/* Addresses in the VMA. */
+		if (vma->vm_flags & VM_SOFTDIRTY)
+			pme.pme |= PM_STATUS2(pm->v2, __PM_SOFT_DIRTY);
+		for (; addr < min(end, vma->vm_end); addr += PAGE_SIZE) {
 			err = add_to_pagemap(addr, &pme, pm);
 			if (err)
 				goto out;
 		}
 	}
-
 out:
 	return err;
 }
Signature · 7 lines
2.1.0.rc2.206.gedb03e5

--
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/ .
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>