fs: Protect write paths by sb_start_write - sb_end_write

There are several entry points which dirty pages in a filesystem.  mmap
(handled by block_page_mkwrite()), buffered write (handled by
__generic_file_aio_write()), splice write (generic_file_splice_write),
truncate, and fallocate (these can dirty last partial page - handled inside
each filesystem separately). Protect these places with sb_start_write() and
sb_end_write().

->page_mkwrite() calls are particularly complex since they are called with
mmap_sem held and thus we cannot use standard sb_start_write() due to lock
ordering constraints. We solve the problem by using a special freeze protection
sb_start_pagefault() which ranks below mmap_sem.

BugLink: https://bugs.launchpad.net/bugs/897421
Tested-by: Kamal Mostafa <kamal@canonical.com>
Tested-by: Peter M. Petrakis <peter.petrakis@canonical.com>
Tested-by: Dann Frazier <dann.frazier@canonical.com>
Tested-by: Massimo Morana <massimo.morana@canonical.com>
Signed-off-by: Jan Kara <jack@suse.cz>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>

2 files changed, 13 insertions, 4 deletions

diff --git a/mm/filemap.c b/mm/filemap.c
index 51efee6..fa5ca30 100644
--- a/mm/filemap.c
+++ b/mm/filemap.c
@@ -1718,6 +1718,7 @@ int filemap_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf)
 	struct inode *inode = vma->vm_file->f_path.dentry->d_inode;
 	int ret = VM_FAULT_LOCKED;
 
+	sb_start_pagefault(inode->i_sb);
 	file_update_time(vma->vm_file);
 	lock_page(page);
 	if (page->mapping != inode->i_mapping) {
@@ -1725,7 +1726,14 @@ int filemap_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf)
 		ret = VM_FAULT_NOPAGE;
 		goto out;
 	}
+	/*
+	 * We mark the page dirty already here so that when freeze is in
+	 * progress, we are guaranteed that writeback during freezing will
+	 * see the dirty page and writeprotect it again.
+	 */
+	set_page_dirty(page);
 out:
+	sb_end_pagefault(inode->i_sb);
 	return ret;
 }
 EXPORT_SYMBOL(filemap_page_mkwrite);
@@ -2426,8 +2434,6 @@ ssize_t __generic_file_aio_write(struct kiocb *iocb, const struct iovec *iov,
 	count = ocount;
 	pos = *ppos;
 
-	vfs_check_frozen(inode->i_sb, SB_FREEZE_WRITE);
-
 	/* We can write back this queue in page reclaim */
 	current->backing_dev_info = mapping->backing_dev_info;
 	written = 0;
@@ -2526,6 +2532,7 @@ ssize_t generic_file_aio_write(struct kiocb *iocb, const struct iovec *iov,
 
 	BUG_ON(iocb->ki_pos != pos);
 
+	sb_start_write(inode->i_sb);
 	mutex_lock(&inode->i_mutex);
 	blk_start_plug(&plug);
 	ret = __generic_file_aio_write(iocb, iov, nr_segs, &iocb->ki_pos);
@@ -2539,6 +2546,7 @@ ssize_t generic_file_aio_write(struct kiocb *iocb, const struct iovec *iov,
 			ret = err;
 	}
 	blk_finish_plug(&plug);
+	sb_end_write(inode->i_sb);
 	return ret;
 }
 EXPORT_SYMBOL(generic_file_aio_write);
diff --git a/mm/filemap_xip.c b/mm/filemap_xip.c
index 80b34ef..13e013b 100644
--- a/mm/filemap_xip.c
+++ b/mm/filemap_xip.c
@@ -402,6 +402,8 @@ xip_file_write(struct file *filp, const char __user *buf, size_t len,
 	loff_t pos;
 	ssize_t ret;
 
+	sb_start_write(inode->i_sb);
+
 	mutex_lock(&inode->i_mutex);
 
 	if (!access_ok(VERIFY_READ, buf, len)) {
@@ -412,8 +414,6 @@ xip_file_write(struct file *filp, const char __user *buf, size_t len,
 	pos = *ppos;
 	count = len;
 
-	vfs_check_frozen(inode->i_sb, SB_FREEZE_WRITE);
-
 	/* We can write back this queue in page reclaim */
 	current->backing_dev_info = mapping->backing_dev_info;
 
@@ -437,6 +437,7 @@ xip_file_write(struct file *filp, const char __user *buf, size_t len,
 	current->backing_dev_info = NULL;
  out_up:
 	mutex_unlock(&inode->i_mutex);
+	sb_end_write(inode->i_sb);
 	return ret;
 }
 EXPORT_SYMBOL_GPL(xip_file_write);