userfaultfd: non-cooperative: userfaultfd_remove revalidate vma in MADV_DONTNEED

userfaultfd_remove() has to be execute before zapping the pagetables or
UFFDIO_COPY could keep filling pages after zap_page_range returned,
which would result in non zero data after a MADV_DONTNEED.

However userfaultfd_remove() may have to release the mmap_sem.  This was
handled correctly in MADV_REMOVE, but MADV_DONTNEED accessed a
potentially stale vma (the very vma passed to zap_page_range(vma, ...)).

The fix consists in revalidating the vma in case userfaultfd_remove()
had to release the mmap_sem.

This also optimizes away an unnecessary down_read/up_read in the
MADV_REMOVE case if UFFD_EVENT_FORK had to be delivered.

It all remains zero runtime cost in case CONFIG_USERFAULTFD=n as
userfaultfd_remove() will be defined as "true" at build time.

Link: http://lkml.kernel.org/r/20170302173738.18994-3-aarcange@redhat.com
Signed-off-by: Andrea Arcangeli <aarcange@redhat.com>
Acked-by: Mike Rapoport <rppt@linux.vnet.ibm.com>
Cc: "Dr. David Alan Gilbert" <dgilbert@redhat.com>
Cc: Mike Kravetz <mike.kravetz@oracle.com>
Cc: Pavel Emelyanov <xemul@parallels.com>
Cc: Hillf Danton <hillf.zj@alibaba-inc.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

diff --git a/fs/userfaultfd.c b/fs/userfaultfd.c
index 2407249998c326e178087f65eb1500954e32abf1..9fd5e51ffb316d64333653ca72e8235d3407d157 100644 (file)
--- a/fs/userfaultfd.c
+++ b/fs/userfaultfd.c
@@ -695,8 +695,7 @@ void mremap_userfaultfd_complete(struct vm_userfaultfd_ctx *vm_ctx,
        userfaultfd_event_wait_completion(ctx, &ewq);
 }
 
-void userfaultfd_remove(struct vm_area_struct *vma,
-                       struct vm_area_struct **prev,
+bool userfaultfd_remove(struct vm_area_struct *vma,
                        unsigned long start, unsigned long end)
 {
        struct mm_struct *mm = vma->vm_mm;
@@ -705,13 +704,11 @@ void userfaultfd_remove(struct vm_area_struct *vma,
 
        ctx = vma->vm_userfaultfd_ctx.ctx;
        if (!ctx || !(ctx->features & UFFD_FEATURE_EVENT_REMOVE))
-               return;
+               return true;
 
        userfaultfd_ctx_get(ctx);
        up_read(&mm->mmap_sem);
 
-       *prev = NULL; /* We wait for ACK w/o the mmap semaphore */
-
        msg_init(&ewq.msg);
 
        ewq.msg.event = UFFD_EVENT_REMOVE;
@@ -720,7 +717,7 @@ void userfaultfd_remove(struct vm_area_struct *vma,
 
        userfaultfd_event_wait_completion(ctx, &ewq);
 
-       down_read(&mm->mmap_sem);
+       return false;
 }
 
 static bool has_unmap_ctx(struct userfaultfd_ctx *ctx, struct list_head *unmaps,

diff --git a/include/linux/userfaultfd_k.h b/include/linux/userfaultfd_k.h
index f2b79bf4c8954cfcf12e4486522e9f9b4beb5bf9..48a3483dccb12360e288ffdd97a9bf8d9d9080a4 100644 (file)
--- a/include/linux/userfaultfd_k.h
+++ b/include/linux/userfaultfd_k.h
@@ -61,8 +61,7 @@ extern void mremap_userfaultfd_complete(struct vm_userfaultfd_ctx *,
                                        unsigned long from, unsigned long to,
                                        unsigned long len);
 
-extern void userfaultfd_remove(struct vm_area_struct *vma,
-                              struct vm_area_struct **prev,
+extern bool userfaultfd_remove(struct vm_area_struct *vma,
                               unsigned long start,
                               unsigned long end);
 
@@ -118,11 +117,11 @@ static inline void mremap_userfaultfd_complete(struct vm_userfaultfd_ctx *ctx,
 {
 }
 
-static inline void userfaultfd_remove(struct vm_area_struct *vma,
-                                     struct vm_area_struct **prev,
+static inline bool userfaultfd_remove(struct vm_area_struct *vma,
                                      unsigned long start,
                                      unsigned long end)
 {
+       return true;
 }
 
 static inline int userfaultfd_unmap_prep(struct vm_area_struct *vma,

diff --git a/mm/madvise.c b/mm/madvise.c
index dc5927c812d3d1f9a209fbdbea3a36a61cbde17d..7a2abf0127aef7a9d4879278293d8cab766133e1 100644 (file)
--- a/mm/madvise.c
+++ b/mm/madvise.c
@@ -513,7 +513,43 @@ static long madvise_dontneed(struct vm_area_struct *vma,
        if (!can_madv_dontneed_vma(vma))
                return -EINVAL;
 
-       userfaultfd_remove(vma, prev, start, end);
+       if (!userfaultfd_remove(vma, start, end)) {
+               *prev = NULL; /* mmap_sem has been dropped, prev is stale */
+
+               down_read(&current->mm->mmap_sem);
+               vma = find_vma(current->mm, start);
+               if (!vma)
+                       return -ENOMEM;
+               if (start < vma->vm_start) {
+                       /*
+                        * This "vma" under revalidation is the one
+                        * with the lowest vma->vm_start where start
+                        * is also < vma->vm_end. If start <
+                        * vma->vm_start it means an hole materialized
+                        * in the user address space within the
+                        * virtual range passed to MADV_DONTNEED.
+                        */
+                       return -ENOMEM;
+               }
+               if (!can_madv_dontneed_vma(vma))
+                       return -EINVAL;
+               if (end > vma->vm_end) {
+                       /*
+                        * Don't fail if end > vma->vm_end. If the old
+                        * vma was splitted while the mmap_sem was
+                        * released the effect of the concurrent
+                        * operation may not cause MADV_DONTNEED to
+                        * have an undefined result. There may be an
+                        * adjacent next vma that we'll walk
+                        * next. userfaultfd_remove() will generate an
+                        * UFFD_EVENT_REMOVE repetition on the
+                        * end-vma->vm_end range, but the manager can
+                        * handle a repetition fine.
+                        */
+                       end = vma->vm_end;
+               }
+               VM_WARN_ON(start >= end);
+       }
        zap_page_range(vma, start, end - start);
        return 0;
 }
@@ -554,8 +590,10 @@ static long madvise_remove(struct vm_area_struct *vma,
         * mmap_sem.
         */
        get_file(f);
-       userfaultfd_remove(vma, prev, start, end);
-       up_read(&current->mm->mmap_sem);
+       if (userfaultfd_remove(vma, start, end)) {
+               /* mmap_sem was not released by userfaultfd_remove() */
+               up_read(&current->mm->mmap_sem);
+       }
        error = vfs_fallocate(f,
                                FALLOC_FL_PUNCH_HOLE | FALLOC_FL_KEEP_SIZE,
                                offset, end - start);