From: Namjae Jeon Subject: [PATCH 1/3] fs: Add FALLOC_FL_INSERT_RANGE flags for fallocate Date: Mon, 31 Mar 2014 23:54:13 +0900 Message-ID: <1396277653-10827-1-git-send-email-linkinjeon@gmail.com> Cc: linux-fsdevel@vger.kernel.org, xfs@oss.sgi.com, linux-ext4@vger.kernel.org, linux-kernel@vger.kernel.org, Namjae Jeon , Namjae Jeon , Ashish Sangwan To: viro@zeniv.linux.org.uk, david@fromorbit.com, hch@infradead.org, tytso@mit.edu, adilger.kernel@dilger.ca, jack@suse.cz, lczerner@redhat.com Return-path: Received: from mail-pa0-f50.google.com ([209.85.220.50]:57675 "EHLO mail-pa0-f50.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752804AbaCaOyX (ORCPT ); Mon, 31 Mar 2014 10:54:23 -0400 Sender: linux-ext4-owner@vger.kernel.org List-ID: From: Namjae Jeon FALLOC_FL_INSERT_RANGE command is the opposite command of FALLOC_FL_COLLAPSE_RANGE that is needed for advertisers or someone who want to add some data in the middle of file. FALLOC_FL_INSERT_RANGE will insert a range of a file after shifting extents to right as given length. and this command also has same limitation as FALLOC_FL_COLLAPSE_RANGE, block boundary and use ftruncate(2) for crosses EOF. Signed-off-by: Namjae Jeon Signed-off-by: Ashish Sangwan --- fs/open.c | 21 ++++++++++++++++++--- include/uapi/linux/falloc.h | 17 +++++++++++++++++ 2 files changed, 35 insertions(+), 3 deletions(-) diff --git a/fs/open.c b/fs/open.c index 631aea8..34727ba 100644 --- a/fs/open.c +++ b/fs/open.c @@ -232,7 +232,8 @@ int do_fallocate(struct file *file, int mode, loff_t offset, loff_t len) /* Return error if mode is not supported */ if (mode & ~(FALLOC_FL_KEEP_SIZE | FALLOC_FL_PUNCH_HOLE | - FALLOC_FL_COLLAPSE_RANGE | FALLOC_FL_ZERO_RANGE)) + FALLOC_FL_COLLAPSE_RANGE | FALLOC_FL_ZERO_RANGE | + FALLOC_FL_INSERT_RANGE)) return -EOPNOTSUPP; /* Punch hole and zero range are mutually exclusive */ @@ -250,6 +251,11 @@ int do_fallocate(struct file *file, int mode, loff_t offset, loff_t len) (mode & ~FALLOC_FL_COLLAPSE_RANGE)) return -EINVAL; + /* Insert range should only be used exclusively. */ + if ((mode & FALLOC_FL_INSERT_RANGE) && + (mode & ~FALLOC_FL_INSERT_RANGE)) + return -EINVAL; + if (!(file->f_mode & FMODE_WRITE)) return -EBADF; @@ -257,8 +263,8 @@ int do_fallocate(struct file *file, int mode, loff_t offset, loff_t len) * It's not possible to punch hole or perform collapse range * on append only file */ - if (mode & (FALLOC_FL_PUNCH_HOLE | FALLOC_FL_COLLAPSE_RANGE) - && IS_APPEND(inode)) + if (mode & (FALLOC_FL_PUNCH_HOLE | FALLOC_FL_COLLAPSE_RANGE | + FALLOC_FL_INSERT_RANGE) && IS_APPEND(inode)) return -EPERM; if (IS_IMMUTABLE(inode)) @@ -286,6 +292,11 @@ int do_fallocate(struct file *file, int mode, loff_t offset, loff_t len) if (((offset + len) > inode->i_sb->s_maxbytes) || ((offset + len) < 0)) return -EFBIG; + /* Check for wrap through zero in case of insert range. */ + if ((mode & FALLOC_FL_INSERT_RANGE) && (inode->i_size + len) > + inode->i_sb->s_maxbytes) + return -EFBIG; + /* * There is no need to overlap collapse range with EOF, in which case * it is effectively a truncate operation @@ -294,6 +305,10 @@ int do_fallocate(struct file *file, int mode, loff_t offset, loff_t len) (offset + len >= i_size_read(inode))) return -EINVAL; + /* Offset should be less than i_size in case of insert range. */ + if ((mode & FALLOC_FL_INSERT_RANGE) && (offset >= i_size_read(inode))) + return -EINVAL; + if (!file->f_op->fallocate) return -EOPNOTSUPP; diff --git a/include/uapi/linux/falloc.h b/include/uapi/linux/falloc.h index d1197ae..f4d12eb 100644 --- a/include/uapi/linux/falloc.h +++ b/include/uapi/linux/falloc.h @@ -41,4 +41,21 @@ */ #define FALLOC_FL_ZERO_RANGE 0x10 +/* + * FALLOC_FL_INSERT_RANGE is use to allocate disk space within the file size + * without overwriting any existing data. The contents of the file beyond offset + * are shifted towards right by len bytes to create space for inserting + * new disk blocks and in this space new zeroed out disk blocks are inserted. + * As such, this operation will increase the size of the file by len bytes. + * Different filesystems may implement different limitations on the granularity + * of the operation. Most will limit operations to filesystem block size + * boundaries, but this boundary may be larger or smaller depending on + * the filesystem and/or the configuration of the filesystem or file. + * + * Attempting to allocate new blocks using this flag at OR beyond the end of + * the file is considered an illegal operation - just use ftruncate(2) or + * fallocate(2) with mode 0 for such type of operations. + */ +#define FALLOC_FL_INSERT_RANGE 0x20 + #endif /* _UAPI_FALLOC_H_ */ -- 1.7.11-rc0