[PATCH 0/1][RFC] prepare_write positive return value V(2)

[Dmitriy Monakhov <dmonakhov@sw.ru>]

Changes from ver(1)
 - __page_symlink(): In order to be on a safe side add explicit zeroing content
                     before fail (just in case).
 -do_lo_send_aops(): If prepare_write can't handle total size of bytes requested
                     from loop_dev it is safer to fail.
 -  pipe_to_file():  Limit the size of the copy to size wich prepare_write ready
                     to handle. instead of failing.
 - ecryptfs: was't changed in this version, because where is no even 
                     AOP_TRUNCATED_PAGE handling code where.

Almost all read/write operation handles data with chunks(segments or pages)
and result has integral behaviour for folowing scenario: 
for_each_chunk() {
     res = op(....);
     if(IS_ERROR(res))
           return progress ? progress : res;
     progress += res;
}
prepare_write may has integral behaviour in case of blksize < pgsize,
for example we successfully allocated/read some blocks, but not all of them,
and than some error happend. Currently we eliminate this progress by doing
vmtrunate() after prepare_has failed.
It is good to have ability to signal about this progress. Interprete positive
prepare_write() ret code as bytes num that fs ready to handle at this moment.
 
BTH: This approach was used in OpenVZ 2.6.9 kernel in order to make FS with 
delayed allocation more correct, and its works well.
I think not everybody will happy about this,  but let's discuss all advantages
and disadvantages of this change.

fixes not dirrectly connected with proposed prepare_write semantic changes:
 __page_symlink : If find_or_create_page has failed on second retry attempt
                  function will exit with wrong error code.
 __generic_cont_expand: Add correct AOP_TRUNCATED_PAGE handling.

Signed-off-by: Dmitriy Monakhov <dmonakhov@openvz.org>
-------------

diff --git a/drivers/block/loop.c b/drivers/block/loop.c
index 6b5b642..dc332dc 100644
--- a/drivers/block/loop.c
+++ b/drivers/block/loop.c
@@ -224,6 +224,7 @@ static int do_lo_send_aops(struct loop_device *lo, struct bio_vec *bvec,
 		sector_t IV;
 		unsigned size;
 		int transfer_result;
+		int partial = 0;
 
 		IV = ((sector_t)index << (PAGE_CACHE_SHIFT - 9))+(offset >> 9);
 		size = PAGE_CACHE_SIZE - offset;
@@ -239,11 +240,20 @@ static int do_lo_send_aops(struct loop_device *lo, struct bio_vec *bvec,
 				page_cache_release(page);
 				continue;
 			}
-			goto unlock;
+			if (ret > 0 && ret < PAGE_CACHE_SIZE) {
+			/* 
+	 	 	 * ->prepare_write() can't handle total size of bytes
+			 * requested. In fact this is not likely to happen, 
+			 * because we request one block only.
+	 	 	 */
+				partial = 1;
+				size = ret;
+			} else
+				goto unlock;
 		}
 		transfer_result = lo_do_transfer(lo, WRITE, page, offset,
 				bvec->bv_page, bv_offs, size, IV);
-		if (unlikely(transfer_result)) {
+		if (unlikely(transfer_result || partial)) {
 			char *kaddr;
 
 			/*
@@ -266,7 +276,7 @@ static int do_lo_send_aops(struct loop_device *lo, struct bio_vec *bvec,
 			}
 			goto unlock;
 		}
-		if (unlikely(transfer_result))
+		if (unlikely(transfer_result || partial))
 			goto unlock;
 		bv_offs += size;
 		len -= size;
diff --git a/fs/buffer.c b/fs/buffer.c
index 1ad674f..ed06560 100644
--- a/fs/buffer.c
+++ b/fs/buffer.c
@@ -2027,13 +2027,20 @@ static int __generic_cont_expand(struct inode *inode, loff_t size,
 	if (size > inode->i_sb->s_maxbytes)
 		goto out;
 
+retry:
 	err = -ENOMEM;
 	page = grab_cache_page(mapping, index);
 	if (!page)
 		goto out;
 	err = mapping->a_ops->prepare_write(NULL, page, offset, offset);
+	if (err == AOP_TRUNCATED_PAGE) {
+		page_cache_release(page);
+		goto retry;
+	}
 	if (err) {
 		/*
+		 * ->prepare_write() called with from==to and must not 
+		 * return positive size of bytes in this case.
 		 * ->prepare_write() may have instantiated a few blocks
 		 * outside i_size.  Trim these off again.
 		 */
@@ -2044,6 +2051,10 @@ static int __generic_cont_expand(struct inode *inode, loff_t size,
 	}
 
 	err = mapping->a_ops->commit_write(NULL, page, offset, offset);
+	if (err == AOP_TRUNCATED_PAGE) {
+		page_cache_release(page);
+		goto retry;
+	}
 
 	unlock_page(page);
 	page_cache_release(page);
diff --git a/fs/namei.c b/fs/namei.c
index e4f108f..6626277 100644
--- a/fs/namei.c
+++ b/fs/namei.c
@@ -2688,20 +2688,36 @@ int __page_symlink(struct inode *inode, const char *symname, int len,
 {
 	struct address_space *mapping = inode->i_mapping;
 	struct page *page;
-	int err = -ENOMEM;
+	int err;
 	char *kaddr;
 
 retry:
+	err = -ENOMEM;	
 	page = find_or_create_page(mapping, 0, gfp_mask);
 	if (!page)
 		goto fail;
 	err = mapping->a_ops->prepare_write(NULL, page, 0, len-1);
-	if (err == AOP_TRUNCATED_PAGE) {
-		page_cache_release(page);
-		goto retry;
-	}
-	if (err)
+	if(unlikely(err)){
+		if (err == AOP_TRUNCATED_PAGE) {
+			page_cache_release(page);
+			goto retry;
+		}
+		if (err > 0 && err < PAGE_CACHE_SIZE) {
+		/* 
+	 	 * ->prepare_write() can't handle total size of bytes requested.
+	 	 * Really this is not likely to happen, because usualy we need 
+		 * one block only. In order to preserve prepare/commit balanced
+		 * invoke commit and fail with ENOSPC.
+	 	 */
+			int ret;
+			kaddr = kmap_atomic(page, KM_USER0);
+			memset(kaddr, 0, err);
+			kunmap_atomic(kaddr, KM_USER0);
+			ret = mapping->a_ops->commit_write(NULL, page, 0, err);
+			err = (ret < 0) ? ret : -ENOSPC;
+		}
 		goto fail_map;
+	}
 	kaddr = kmap_atomic(page, KM_USER0);
 	memcpy(kaddr, symname, len-1);
 	kunmap_atomic(kaddr, KM_USER0);
diff --git a/fs/splice.c b/fs/splice.c
index 2fca6eb..78c3e88 100644
--- a/fs/splice.c
+++ b/fs/splice.c
@@ -649,6 +649,14 @@ find_page:
 	}
 
 	ret = mapping->a_ops->prepare_write(file, page, offset, offset+this_len);
+	if (unlikely(ret > 0 && ret < PAGE_CACHE_SIZE)) {
+	/*
+	 * Limit the size of the copy to size wich prepare_write 
+	 * ready to handle.
+	 */
+		this_len = ret;
+		ret = 0;
+	}
 	if (unlikely(ret)) {
 		loff_t isize = i_size_read(mapping->host);
 
diff --git a/mm/filemap.c b/mm/filemap.c
index 8332c77..8b4789d 100644
--- a/mm/filemap.c
+++ b/mm/filemap.c
@@ -2127,6 +2127,14 @@ generic_file_buffered_write(struct kiocb *iocb, const struct iovec *iov,
 		}
 
 		status = a_ops->prepare_write(file, page, offset, offset+bytes);
+		if (unlikely(status > 0 && status < PAGE_CACHE_SIZE)) {
+		/*
+		 * Limit the size of the copy to size wich prepare_write
+		 * ready to handle.
+		 */
+			bytes = status;
+			status = 0;
+		}
 		if (unlikely(status)) {
 			loff_t isize = i_size_read(inode);