[RFC,1/3] fs: Add folio_may_straddle_isize helper

Commit Message

Andreas Gruenbacher Dec. 1, 2022, 4:06 p.m. UTC

Add a folio_may_straddle_isize() helper as a replacement for
pagecache_isize_extended() when we have a locked folio.

Use the new helper in generic_write_end(), iomap_write_end(),
ext4_write_end(), and ext4_journalled_write_end().

Signed-off-by: Andreas Gruenbacher <[e-mail]>
---
 fs/buffer.c            |  5 ++---
 fs/ext4/inode.c        | 13 ++++++-------
 fs/iomap/buffered-io.c | 14 +++++++-------
 include/linux/mm.h     |  2 ++
 mm/truncate.c          | 34 ++++++++++++++++++++++++++++++++++
 5 files changed, 51 insertions(+), 17 deletions(-)

Patch

diff --git a/fs/buffer.c b/fs/buffer.c
index d9c6d1fbb6dd..bbae1437994b 100644
--- a/fs/buffer.c
+++ b/fs/buffer.c
@@ -2164,16 +2164,15 @@  int generic_write_end(struct file *file, struct address_space *mapping,
 	 * But it's important to update i_size while still holding page lock:
 	 * page writeout could otherwise come in and zero beyond i_size.
 	 */
-	if (pos + copied > inode->i_size) {
+	if (pos + copied > old_size) {
 		i_size_write(inode, pos + copied);
 		i_size_changed = true;
+		folio_may_straddle_isize(inode, page_folio(page), old_size, pos);
 	}
 
 	unlock_page(page);
 	put_page(page);
 
-	if (old_size < pos)
-		pagecache_isize_extended(inode, old_size, pos);
 	/*
 	 * Don't mark the inode dirty under page lock. First, it unnecessarily
 	 * makes the holding time of page lock longer. Second, it forces lock
diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c
index 2b5ef1b64249..48e6b4716415 100644
--- a/fs/ext4/inode.c
+++ b/fs/ext4/inode.c
@@ -1320,13 +1320,13 @@  static int ext4_write_end(struct file *file,
 	 * If FS_IOC_ENABLE_VERITY is running on this inode, then Merkle tree
 	 * blocks are being written past EOF, so skip the i_size update.
 	 */
-	if (!verity)
+	if (!verity) {
 		i_size_changed = ext4_update_inode_size(inode, pos + copied);
+		folio_may_straddle_isize(inode, page_folio(page), old_size, pos);
+	}
 	unlock_page(page);
 	put_page(page);
 
-	if (old_size < pos && !verity)
-		pagecache_isize_extended(inode, old_size, pos);
 	/*
 	 * Don't mark the inode dirty under page lock. First, it unnecessarily
 	 * makes the holding time of page lock longer. Second, it forces lock
@@ -1432,16 +1432,15 @@  static int ext4_journalled_write_end(struct file *file,
 		if (!partial)
 			SetPageUptodate(page);
 	}
-	if (!verity)
+	if (!verity) {
 		size_changed = ext4_update_inode_size(inode, pos + copied);
+		folio_may_straddle_isize(inode, page_folio(page), old_size, pos);
+	}
 	ext4_set_inode_state(inode, EXT4_STATE_JDATA);
 	EXT4_I(inode)->i_datasync_tid = handle->h_transaction->t_tid;
 	unlock_page(page);
 	put_page(page);
 
-	if (old_size < pos && !verity)
-		pagecache_isize_extended(inode, old_size, pos);
-
 	if (size_changed) {
 		ret2 = ext4_mark_inode_dirty(handle, inode);
 		if (!ret)
diff --git a/fs/iomap/buffered-io.c b/fs/iomap/buffered-io.c
index 91ee0b308e13..5e751125e35e 100644
--- a/fs/iomap/buffered-io.c
+++ b/fs/iomap/buffered-io.c
@@ -713,14 +713,14 @@  static size_t iomap_write_end(struct iomap_iter *iter, loff_t pos, size_t len,
 	if (pos + ret > old_size) {
 		i_size_write(iter->inode, pos + ret);
 		iter->iomap.flags |= IOMAP_F_SIZE_CHANGED;
+		folio_may_straddle_isize(iter->inode, folio, old_size, pos);
+	}
+	if (folio_ops && folio_ops->folio_done) {
+		folio_ops->folio_done(iter->inode, folio, pos, ret);
+	} else {
+		folio_unlock(folio);
+		folio_put(folio);
 	}
-	folio_unlock(folio);
-
-	if (old_size < pos)
-		pagecache_isize_extended(iter->inode, old_size, pos);
-	if (page_ops && page_ops->page_done)
-		page_ops->page_done(iter->inode, pos, ret, &folio->page);
-	folio_put(folio);
 
 	if (ret < len)
 		iomap_write_failed(iter->inode, pos + ret, len - ret);
diff --git a/include/linux/mm.h b/include/linux/mm.h
index 8bbcccbc5565..e1f03c9ed8df 100644
--- a/include/linux/mm.h
+++ b/include/linux/mm.h
@@ -1890,6 +1890,8 @@  int generic_access_phys(struct vm_area_struct *vma, unsigned long addr,
 
 extern void truncate_pagecache(struct inode *inode, loff_t new);
 extern void truncate_setsize(struct inode *inode, loff_t newsize);
+void folio_may_straddle_isize(struct inode *inode, struct folio *folio,
+			      loff_t old_size, loff_t start);
 void pagecache_isize_extended(struct inode *inode, loff_t from, loff_t to);
 void truncate_pagecache_range(struct inode *inode, loff_t offset, loff_t end);
 int generic_error_remove_page(struct address_space *mapping, struct page *page);
diff --git a/mm/truncate.c b/mm/truncate.c
index c0be77e5c008..9a5d3c3c12d0 100644
--- a/mm/truncate.c
+++ b/mm/truncate.c
@@ -779,6 +779,40 @@  void truncate_setsize(struct inode *inode, loff_t newsize)
 }
 EXPORT_SYMBOL(truncate_setsize);
 
+/**
+ * folio_may_straddle_isize - update pagecache after extending i_size
+ * @inode:	inode for which i_size was extended
+ * @folio:	folio to maybe mark read-only
+ * @old_size:	original inode size
+ * @start:	start of the write
+ *
+ * Handle extending an inode by a write that starts behind the old inode size.
+ * If a block-aligned hole exists between the old inode size and the start of
+ * the write, we mark the folio read-only so that page_mkwrite() is called on
+ * the nearest write access to the page.  That way, the filesystem can be sure
+ * that page_mkwrite() is called on the page before a user writes to the page
+ * via mmap.
+ *
+ * This function must be called while we still hold i_rwsem - this not only
+ * makes sure i_size is stable but also that userspace cannot observe the new
+ * i_size value before we are prepared to handle mmap writes there.
+ */
+void folio_may_straddle_isize(struct inode *inode, struct folio *folio,
+			      loff_t old_size, loff_t start)
+{
+	unsigned int blocksize = i_blocksize(inode);
+
+	if (round_up(old_size, blocksize) >= round_down(start, blocksize))
+		return;
+
+	/*
+	 * See clear_page_dirty_for_io() for details why folio_set_dirty()
+	 * is needed.
+	 */
+	if (folio_mkclean(folio))
+		folio_set_dirty(folio);
+}
+
 /**
  * pagecache_isize_extended - update pagecache after extension of i_size
  * @inode:	inode for which i_size was extended