[RFC v2 05/10] mm: Add a range lock parameter to userfaultfd_remove()

From: Laurent Dufour
Date: Wed May 24 2017 - 07:22:03 EST


As __mcopy_atomic_hugetlb() called by userfaultfd_remove() may unlock
the mmap_sem, it has to know about the range of lock when dealing with
range lock.

This patch adds a new range_lock pointer parameter to
userfaultfd_remove() and handles it in the callees.

On the otherside, userfaultfd_remove()'s callers are touched to deal
with the range parameter as well.

Signed-off-by: Laurent Dufour <ldufour@xxxxxxxxxxxxxxxxxx>
---
fs/userfaultfd.c | 8 ++++++--
include/linux/userfaultfd_k.h | 28 ++++++++++++++++++++++++----
mm/madvise.c | 42 +++++++++++++++++++++++++++++++++---------
mm/userfaultfd.c | 18 +++++++++++++++---
4 files changed, 78 insertions(+), 18 deletions(-)

diff --git a/fs/userfaultfd.c b/fs/userfaultfd.c
index b3daffc589a2..7d56c21ef65d 100644
--- a/fs/userfaultfd.c
+++ b/fs/userfaultfd.c
@@ -703,8 +703,12 @@ void mremap_userfaultfd_complete(struct vm_userfaultfd_ctx *vm_ctx,
userfaultfd_event_wait_completion(ctx, &ewq);
}

-bool userfaultfd_remove(struct vm_area_struct *vma,
- unsigned long start, unsigned long end)
+bool _userfaultfd_remove(struct vm_area_struct *vma,
+ unsigned long start, unsigned long end
+#ifdef CONFIG_MEM_RANGE_LOCK
+ , struct range_lock *range
+#endif
+ )
{
struct mm_struct *mm = vma->vm_mm;
struct userfaultfd_ctx *ctx;
diff --git a/include/linux/userfaultfd_k.h b/include/linux/userfaultfd_k.h
index 48a3483dccb1..07c3dbddc021 100644
--- a/include/linux/userfaultfd_k.h
+++ b/include/linux/userfaultfd_k.h
@@ -61,9 +61,18 @@ extern void mremap_userfaultfd_complete(struct vm_userfaultfd_ctx *,
unsigned long from, unsigned long to,
unsigned long len);

-extern bool userfaultfd_remove(struct vm_area_struct *vma,
- unsigned long start,
- unsigned long end);
+#ifdef CONFIG_MEM_RANGE_LOCK
+extern bool _userfaultfd_remove(struct vm_area_struct *vma,
+ unsigned long start,
+ unsigned long end,
+ struct range_lock *range);
+#define userfaultfd_remove _userfaultfd_remove
+#else
+extern bool _userfaultfd_remove(struct vm_area_struct *vma,
+ unsigned long start,
+ unsigned long end);
+#define userfaultfd_remove(v, s, e, r) _userfaultfd_remove(v, s, e)
+#endif /* CONFIG_MEM_RANGE_LOCK */

extern int userfaultfd_unmap_prep(struct vm_area_struct *vma,
unsigned long start, unsigned long end,
@@ -117,12 +126,23 @@ static inline void mremap_userfaultfd_complete(struct vm_userfaultfd_ctx *ctx,
{
}

+#ifdef CONFIG_MEM_RANGE_LOCK
static inline bool userfaultfd_remove(struct vm_area_struct *vma,
unsigned long start,
- unsigned long end)
+ unsigned long end,
+ struct range_lock *range)
{
return true;
}
+#else
+static inline bool _userfaultfd_remove(struct vm_area_struct *vma,
+ unsigned long start,
+ unsigned long end)
+{
+ return true;
+}
+#define userfaultfd_remove(v, s, e, r) _userfaultfd_remove(v, s, e)
+#endif

static inline int userfaultfd_unmap_prep(struct vm_area_struct *vma,
unsigned long start, unsigned long end,
diff --git a/mm/madvise.c b/mm/madvise.c
index 25b78ee4fc2c..437f35778f07 100644
--- a/mm/madvise.c
+++ b/mm/madvise.c
@@ -506,13 +506,17 @@ static long madvise_free(struct vm_area_struct *vma,
*/
static long madvise_dontneed(struct vm_area_struct *vma,
struct vm_area_struct **prev,
- unsigned long start, unsigned long end)
+ unsigned long start, unsigned long end
+#ifdef CONFIG_MEM_RANGE_LOCK
+ , struct range_lock *range
+#endif
+ )
{
*prev = vma;
if (!can_madv_dontneed_vma(vma))
return -EINVAL;

- if (!userfaultfd_remove(vma, start, end)) {
+ if (!userfaultfd_remove(vma, start, end, range)) {
*prev = NULL; /* mmap_sem has been dropped, prev is stale */

down_read(&current->mm->mmap_sem);
@@ -558,8 +562,12 @@ static long madvise_dontneed(struct vm_area_struct *vma,
* This is effectively punching a hole into the middle of a file.
*/
static long madvise_remove(struct vm_area_struct *vma,
- struct vm_area_struct **prev,
- unsigned long start, unsigned long end)
+ struct vm_area_struct **prev,
+ unsigned long start, unsigned long end
+#ifdef CONFIG_MEM_RANGE_LOCK
+ , struct range_lock *range
+#endif
+ )
{
loff_t offset;
int error;
@@ -589,7 +597,7 @@ static long madvise_remove(struct vm_area_struct *vma,
* mmap_sem.
*/
get_file(f);
- if (userfaultfd_remove(vma, start, end)) {
+ if (userfaultfd_remove(vma, start, end, NULL)) {
/* mmap_sem was not released by userfaultfd_remove() */
up_read(&current->mm->mmap_sem);
}
@@ -648,17 +656,29 @@ static int madvise_inject_error(int behavior,

static long
madvise_vma(struct vm_area_struct *vma, struct vm_area_struct **prev,
- unsigned long start, unsigned long end, int behavior)
+ unsigned long start, unsigned long end, int behavior
+#ifdef CONFIG_MEM_RANGE_LOCK
+ , struct range_lock *range
+#endif
+ )
{
switch (behavior) {
case MADV_REMOVE:
- return madvise_remove(vma, prev, start, end);
+ return madvise_remove(vma, prev, start, end
+#ifdef CONFIG_MEM_RANGE_LOCK
+ , range
+#endif
+ );
case MADV_WILLNEED:
return madvise_willneed(vma, prev, start, end);
case MADV_FREE:
return madvise_free(vma, prev, start, end);
case MADV_DONTNEED:
- return madvise_dontneed(vma, prev, start, end);
+ return madvise_dontneed(vma, prev, start, end
+#ifdef CONFIG_MEM_RANGE_LOCK
+ , range
+#endif
+ );
default:
return madvise_behavior(vma, prev, start, end, behavior);
}
@@ -826,7 +846,11 @@ SYSCALL_DEFINE3(madvise, unsigned long, start, size_t, len_in, int, behavior)
tmp = end;

/* Here vma->vm_start <= start < tmp <= (end|vma->vm_end). */
- error = madvise_vma(vma, &prev, start, tmp, behavior);
+ error = madvise_vma(vma, &prev, start, tmp, behavior
+#ifdef CONFIG_MEM_RANGE_LOCK
+ , &range
+#endif
+ );
if (error)
goto out;
start = tmp;
diff --git a/mm/userfaultfd.c b/mm/userfaultfd.c
index 8bcb501bce60..ae2babc46fa5 100644
--- a/mm/userfaultfd.c
+++ b/mm/userfaultfd.c
@@ -156,7 +156,11 @@ static __always_inline ssize_t __mcopy_atomic_hugetlb(struct mm_struct *dst_mm,
unsigned long dst_start,
unsigned long src_start,
unsigned long len,
- bool zeropage)
+ bool zeropage
+#ifdef CONFIG_MEM_RANGE_LOCK
+ , struct range_lock *range
+#endif
+ )
{
int vm_alloc_shared = dst_vma->vm_flags & VM_SHARED;
int vm_shared = dst_vma->vm_flags & VM_SHARED;
@@ -368,7 +372,11 @@ extern ssize_t __mcopy_atomic_hugetlb(struct mm_struct *dst_mm,
unsigned long dst_start,
unsigned long src_start,
unsigned long len,
- bool zeropage);
+ bool zeropage
+#ifdef CONFIG_MEM_RANGE_LOCK
+ , struct range_lock *range
+#endif
+ );
#endif /* CONFIG_HUGETLB_PAGE */

static __always_inline ssize_t __mcopy_atomic(struct mm_struct *dst_mm,
@@ -439,7 +447,11 @@ static __always_inline ssize_t __mcopy_atomic(struct mm_struct *dst_mm,
*/
if (is_vm_hugetlb_page(dst_vma))
return __mcopy_atomic_hugetlb(dst_mm, dst_vma, dst_start,
- src_start, len, zeropage);
+ src_start, len, zeropage
+#ifdef CONFIG_MEM_RANGE_LOCK
+ , &range
+#endif
+ );

if (!vma_is_anonymous(dst_vma) && !vma_is_shmem(dst_vma))
goto out_unlock;
--
2.7.4