[PATCH] madvise(MADV_REMOVE): remove pages from tmpfs shm backing store

Here is the patch to implement madvise(MADV_REMOVE) - which frees up a
given range of pages & its associated backing store.  Current
implementation supports only shmfs/tmpfs and other filesystems return
-ENOSYS.

"Some app allocates large tmpfs files, then when some task quits and some
client disconnect, some memory can be released.  However the only way to
release tmpfs-swap is to MADV_REMOVE". - Andrea Arcangeli

Databases want to use this feature to drop a section of their bufferpool
(shared memory segments) - without writing back to disk/swap space.

This feature is also useful for supporting hot-plug memory on UML.

Concerns raised by Andrew Morton:

- "We have no plan for holepunching!  If we _do_ have such a plan (or
  might in the future) then what would the API look like?  I think
  sys_holepunch(fd, start, len), so we should start out with that."

- Using madvise is very weird, because people will ask "why do I need to
  mmap my file before I can stick a hole in it?"

- None of the other madvise operations call into the filesystem in this
  manner.  A broad question is: is this capability an MM operation or a
  filesytem operation?  truncate, for example, is a filesystem operation
  which sometimes has MM side-effects.  madvise is an mm operation and with
  this patch, it gains FS side-effects, only they're really, really
  significant ones."

Comments:

- Andrea suggested the fs operation too but then it's more efficient to
  have it as a mm operation with fs side effects, because they don't
  immediatly know fd and physical offset of the range.  It's possible to
  fixup in userland and to use the fs operation but it's more expensive,
  the vmas are already in the kernel and we can use them.

Short term plan &  Future Direction:

- We seem to need this interface only for shmfs/tmpfs files in the short
  term.  We have to add hooks into the filesystem for correctness and
  completeness.  This is what this patch does.

- In the future, plan is to support both fs and mmap apis also.  This
  also involves (other) filesystem specific functions to be implemented.

- Current patch doesn't support VM_NONLINEAR - which can be addressed in
  the future.

Signed-off-by: Badari Pulavarty <pbadari@us.ibm.com>
Cc: Hugh Dickins <hugh@veritas.com>
Cc: Andrea Arcangeli <andrea@suse.de>
Cc: Michael Kerrisk <mtk-manpages@gmx.net>
Cc: Ulrich Drepper <drepper@redhat.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

3 files changed, 83 insertions, 9 deletions

diff --git a/mm/madvise.c b/mm/madvise.c
index 2b7cf0400a21..ae0ae3ea299a 100644
--- a/mm/madvise.c
+++ b/mm/madvise.c
@@ -140,6 +140,36 @@ static long madvise_dontneed(struct vm_area_struct * vma,
         return 0;
 }
 
+/*
+ * Application wants to free up the pages and associated backing store.
+ * This is effectively punching a hole into the middle of a file.
+ *
+ * NOTE: Currently, only shmfs/tmpfs is supported for this operation.
+ * Other filesystems return -ENOSYS.
+ */
+static long madvise_remove(struct vm_area_struct *vma,
+                                unsigned long start, unsigned long end)
+{
+        struct address_space *mapping;
+        loff_t offset, endoff;
+
+        if (vma->vm_flags & (VM_LOCKED|VM_NONLINEAR|VM_HUGETLB))
+                return -EINVAL;
+
+        if (!vma->vm_file || !vma->vm_file->f_mapping
+                || !vma->vm_file->f_mapping->host) {
+                        return -EINVAL;
+        }
+
+        mapping = vma->vm_file->f_mapping;
+
+        offset = (loff_t)(start - vma->vm_start)
+                        + ((loff_t)vma->vm_pgoff << PAGE_SHIFT);
+        endoff = (loff_t)(end - vma->vm_start - 1)
+                        + ((loff_t)vma->vm_pgoff << PAGE_SHIFT);
+        return  vmtruncate_range(mapping->host, offset, endoff);
+}
+
 static long
 madvise_vma(struct vm_area_struct *vma, struct vm_area_struct **prev,
                 unsigned long start, unsigned long end, int behavior)
@@ -152,6 +182,9 @@ madvise_vma(struct vm_area_struct *vma, struct vm_area_struct **prev,
         case MADV_RANDOM:
                 error = madvise_behavior(vma, prev, start, end, behavior);
                 break;
+        case MADV_REMOVE:
+                error = madvise_remove(vma, start, end);
+                break;
 
         case MADV_WILLNEED:
                 error = madvise_willneed(vma, prev, start, end);
@@ -190,6 +223,8 @@ madvise_vma(struct vm_area_struct *vma, struct vm_area_struct **prev,
  *              some pages ahead.
  *  MADV_DONTNEED - the application is finished with the given range,
  *              so the kernel can free resources associated with it.
+ *  MADV_REMOVE - the application wants to free up the given range of
+ *              pages and associated backing store.
  *
  * return values:
  *  zero    - success
diff --git a/mm/memory.c b/mm/memory.c
index d8dde07a3656..e249088908c4 100644
--- a/mm/memory.c
+++ b/mm/memory.c
@@ -1770,9 +1770,32 @@ out_big:
 out_busy:
         return -ETXTBSY;
 }
-
 EXPORT_SYMBOL(vmtruncate);
 
+int vmtruncate_range(struct inode *inode, loff_t offset, loff_t end)
+{
+        struct address_space *mapping = inode->i_mapping;
+
+        /*
+         * If the underlying filesystem is not going to provide
+         * a way to truncate a range of blocks (punch a hole) -
+         * we should return failure right now.
+         */
+        if (!inode->i_op || !inode->i_op->truncate_range)
+                return -ENOSYS;
+
+        down(&inode->i_sem);
+        down_write(&inode->i_alloc_sem);
+        unmap_mapping_range(mapping, offset, (end - offset), 1);
+        truncate_inode_pages_range(mapping, offset, end);
+        inode->i_op->truncate_range(inode, offset, end);
+        up_write(&inode->i_alloc_sem);
+        up(&inode->i_sem);
+
+        return 0;
+}
+EXPORT_SYMBOL(vmtruncate_range);
+
 /* 
  * Primitive swap readahead code. We simply read an aligned block of
  * (1 << page_cluster) entries in the swap area. This method is chosen
diff --git a/mm/shmem.c b/mm/shmem.c
index d9fc277940da..65c148efa2ed 100644
--- a/mm/shmem.c
+++ b/mm/shmem.c
@@ -457,7 +457,7 @@ static void shmem_free_pages(struct list_head *next)
         } while (next);
 }
 
-static void shmem_truncate(struct inode *inode)
+static void shmem_truncate_range(struct inode *inode, loff_t start, loff_t end)
 {
         struct shmem_inode_info *info = SHMEM_I(inode);
         unsigned long idx;
@@ -475,18 +475,27 @@ static void shmem_truncate(struct inode *inode)
         long nr_swaps_freed = 0;
         int offset;
         int freed;
+        int punch_hole = 0;
 
         inode->i_ctime = inode->i_mtime = CURRENT_TIME;
-        idx = (inode->i_size + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT;
+        idx = (start + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT;
         if (idx >= info->next_index)
                 return;
 
         spin_lock(&info->lock);
         info->flags |= SHMEM_TRUNCATE;
-        limit = info->next_index;
-        info->next_index = idx;
+        if (likely(end == (loff_t) -1)) {
+                limit = info->next_index;
+                info->next_index = idx;
+        } else {
+                limit = (end + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT;
+                if (limit > info->next_index)
+                        limit = info->next_index;
+                punch_hole = 1;
+        }
+
         topdir = info->i_indirect;
-        if (topdir && idx <= SHMEM_NR_DIRECT) {
+        if (topdir && idx <= SHMEM_NR_DIRECT && !punch_hole) {
                 info->i_indirect = NULL;
                 nr_pages_to_free++;
                 list_add(&topdir->lru, &pages_to_free);
@@ -573,11 +582,12 @@ static void shmem_truncate(struct inode *inode)
                         set_page_private(subdir, page_private(subdir) - freed);
                         if (offset)
                                 spin_unlock(&info->lock);
-                        BUG_ON(page_private(subdir) > offset);
+                        if (!punch_hole)
+                                BUG_ON(page_private(subdir) > offset);
                 }
                 if (offset)
                         offset = 0;
-                else if (subdir) {
+                else if (subdir && !page_private(subdir)) {
                         dir[diroff] = NULL;
                         nr_pages_to_free++;
                         list_add(&subdir->lru, &pages_to_free);
@@ -594,7 +604,7 @@ done2:
                  * Also, though shmem_getpage checks i_size before adding to
                  * cache, no recheck after: so fix the narrow window there too.
                  */
-                truncate_inode_pages(inode->i_mapping, inode->i_size);
+                truncate_inode_pages_range(inode->i_mapping, start, end);
         }
 
         spin_lock(&info->lock);
@@ -614,6 +624,11 @@ done2:
         }
 }
 
+static void shmem_truncate(struct inode *inode)
+{
+        shmem_truncate_range(inode, inode->i_size, (loff_t)-1);
+}
+
 static int shmem_notify_change(struct dentry *dentry, struct iattr *attr)
 {
         struct inode *inode = dentry->d_inode;
@@ -2083,6 +2098,7 @@ static struct file_operations shmem_file_operations = {
 static struct inode_operations shmem_inode_operations = {
         .truncate       = shmem_truncate,
         .setattr        = shmem_notify_change,
+        .truncate_range = shmem_truncate_range,
 };
 
 static struct inode_operations shmem_dir_inode_operations = {