fiemap cleanups v4
Subsystems (14)
Hi all,
This series cleans up the fiemap support in ext4 and in general.
Ted or Al, can one of you pick this up? It touches both ext4 and core code, so either tree could work.
Changes since v3:
- dropped the fixes that have been merged int mainline
Changes since v2:
- commit message typo
- doc updates
- use d_inode in cifs
- add a missing return statement in cifs
- remove the filemap_write_and_wait call from ext4_ioctl_get_es_cache
Changes since v1:
- rename fiemap_validate to fiemap_prep
- lift FIEMAP_FLAG_SYNC handling to common code
- add a new linux/fiemap.h header
- remove __generic_block_fiemap
- remove access_ok calls from fiemap and ext4
PATCH 1/9 ext4: split _ext4_fiemap
The fiemap and EXT4_IOC_GET_ES_CACHE cases share almost no code, so split them into entirely separate functions.
--- fs/ext4/extents.c | 72 +++++++++++++++++++++++------------------------ 1 file changed, 35 insertions(+), 37 deletions(-)
fs/ext4/extents.c +35 -37
@@ -4854,11 +4854,9 @@ static int ext4_fiemap_check_ranges(struct inode *inode, u64 start, u64 *len)return 0 ; } -static int _ext4_fiemap (struct inode *inode,struct fiemap_extent_info *fieinfo, -__u64 start,__u64 len,bool from_es_cache) +int ext4_fiemap (struct inode *inode,struct fiemap_extent_info *fieinfo, +u64 start,u64 len) { -ext4_lblk_t start_blk; -u32 ext4_fiemap_flags =FIEMAP_FLAG_SYNC |FIEMAP_FLAG_XATTR ;int error =0 ;if (fieinfo->fi_flags &FIEMAP_FLAG_CACHE ) {
@@ -4868,10 +4866,7 @@ static int _ext4_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo, fieinfo->fi_flags &= ~FIEMAP_FLAG_CACHE ; } -if (from_es_cache) - ext4_fiemap_flags &=FIEMAP_FLAG_XATTR ; - -if (fiemap_check_flags (fieinfo, ext4_fiemap_flags)) +if (fiemap_check_flags (fieinfo,FIEMAP_FLAG_SYNC |FIEMAP_FLAG_XATTR ))return -EBADR ;/*
@@ -4885,40 +4880,20 @@ static int _ext4_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,if (fieinfo->fi_flags &FIEMAP_FLAG_XATTR ) { fieinfo->fi_flags &= ~FIEMAP_FLAG_XATTR ; - error =iomap_fiemap (inode, fieinfo, start, len, - &ext4_iomap_xattr_ops); - }else if (!from_es_cache) { - error =iomap_fiemap (inode, fieinfo, start, len, - &ext4_iomap_report_ops); - }else { -ext4_lblk_t len_blks; -__u64 last_blk; - - start_blk = start >> inode->i_sb->s_blocksize_bits; - last_blk = (start + len -1 ) >> inode->i_sb->s_blocksize_bits; -if (last_blk >=EXT_MAX_BLOCKS ) - last_blk =EXT_MAX_BLOCKS -1 ; - len_blks = ((ext4_lblk_t ) last_blk) - start_blk +1 ; - -/* -* Walk the extent tree gathering extent information -* and pushing extents back to the user. -*/ - error =ext4_fill_es_cache_info (inode, start_blk, len_blks, - fieinfo); +return iomap_fiemap (inode, fieinfo, start, len, + &ext4_iomap_xattr_ops); } -return error; -} -int ext4_fiemap (struct inode *inode,struct fiemap_extent_info *fieinfo, -__u64 start,__u64 len) -{ -return _ext4_fiemap (inode, fieinfo, start, len, false); +return iomap_fiemap (inode, fieinfo, start, len, &ext4_iomap_report_ops); }int ext4_get_es_cache (struct inode *inode,struct fiemap_extent_info *fieinfo,__u64 start,__u64 len) { +ext4_lblk_t start_blk, len_blks; +__u64 last_blk; +int error =0 ; +if (ext4_has_inline_data (inode)) {int has_inline;
@@ -4929,9 +4904,32 @@ int ext4_get_es_cache(struct inode *inode, struct fiemap_extent_info *fieinfo,return 0 ; } -return _ext4_fiemap (inode, fieinfo, start, len, true); -} +if (fieinfo->fi_flags &FIEMAP_FLAG_CACHE ) { + error =ext4_ext_precache (inode); +if (error) +return error; + fieinfo->fi_flags &= ~FIEMAP_FLAG_CACHE ; + } + +if (fiemap_check_flags (fieinfo,FIEMAP_FLAG_SYNC )) +return -EBADR ; + error =ext4_fiemap_check_ranges (inode, start, &len); +if (error) +return error; + + start_blk = start >> inode->i_sb->s_blocksize_bits; + last_blk = (start + len -1 ) >> inode->i_sb->s_blocksize_bits; +if (last_blk >=EXT_MAX_BLOCKS ) + last_blk =EXT_MAX_BLOCKS -1 ; + len_blks = ((ext4_lblk_t ) last_blk) - start_blk +1 ; + +/* +* Walk the extent tree gathering extent information +* and pushing extents back to the user. +*/ +return ext4_fill_es_cache_info (inode, start_blk, len_blks, fieinfo); +}/* * ext4_access_path:
Signature · 1 line
2.26.2
PATCH 2/9 ext4: remove the call to fiemap_check_flags in ext4_fiemap
In reply to Christoph Hellwigiomap_fiemap already calls fiemap_check_flags first thing, so this additional check is redundant.
--- fs/ext4/extents.c | 3 --- 1 file changed, 3 deletions(-)
fs/ext4/extents.c +0 -3
@@ -4866,9 +4866,6 @@ int ext4_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo, fieinfo->fi_flags &= ~FIEMAP_FLAG_CACHE ; } -if (fiemap_check_flags (fieinfo,FIEMAP_FLAG_SYNC |FIEMAP_FLAG_XATTR )) -return -EBADR ; -/* * For bitmap files the maximum size limit could be smaller than * s_maxbytes, so check len here manually instead of just relying on the
Signature · 1 line
2.26.2
PATCH 3/9 fs: mark __generic_block_fiemap static
In reply to Christoph HellwigThere is no caller left outside of ioctl.c.
--- fs/ioctl.c | 4 +--- include/linux/fs.h | 4 ---- 2 files changed, 1 insertion(+), 7 deletions(-)
fs/ioctl.c +1 -3
@@ -307,8 +307,7 @@ static inline loff_t blk_to_logical(struct inode *inode, sector_t blk)* If you use this function directly, you need to do your own locking. Use * generic_block_fiemap if you want the locking done for you. */ - -int __generic_block_fiemap (struct inode *inode, +static int __generic_block_fiemap (struct inode *inode,struct fiemap_extent_info *fieinfo,loff_t start,loff_t len,get_block_t *get_block) {
@@ -453,7 +452,6 @@ int __generic_block_fiemap(struct inode *inode,return ret; } -EXPORT_SYMBOL (__generic_block_fiemap);/** * generic_block_fiemap - FIEMAP for block based inodes
include/linux/fs.h +0 -4
@@ -3299,10 +3299,6 @@ static inline int vfs_fstat(int fd, struct kstat *stat)extern const char *vfs_get_link (struct dentry *,struct delayed_call *);extern int vfs_readlink (struct dentry *,char __user *,int ); -extern int __generic_block_fiemap (struct inode *inode, -struct fiemap_extent_info *fieinfo, -loff_t start,loff_t len, -get_block_t *get_block);extern int generic_block_fiemap (struct inode *inode,struct fiemap_extent_info *fieinfo,u64 start,u64 len,get_block_t *get_block);
Signature · 1 line
2.26.2
PATCH 4/9 fs: move the fiemap definitions out of fs.h
In reply to Christoph HellwigNo need to pull the fiemap definitions into almost every file in the kernel build.
--- fs/bad_inode.c | 1 + fs/btrfs/extent_io.h | 1 + fs/cifs/inode.c | 1 + fs/cifs/smb2ops.c | 1 + fs/ext2/inode.c | 1 + fs/ext4/ext4.h | 1 + fs/f2fs/data.c | 1 + fs/f2fs/inline.c | 1 + fs/gfs2/inode.c | 1 + fs/hpfs/file.c | 1 + fs/ioctl.c | 1 + fs/iomap/fiemap.c | 1 + fs/nilfs2/inode.c | 1 + fs/overlayfs/inode.c | 1 + fs/xfs/xfs_iops.c | 1 + include/linux/fiemap.h | 24 ++++++++++++++++++++++++ include/linux/fs.h | 19 +------------------ include/uapi/linux/fiemap.h | 6 +++--- 18 files changed, 43 insertions(+), 21 deletions(-) create mode 100644 include/linux/fiemap.h
fs/bad_inode.c +1 -0
@@ -15,6 +15,7 @@#include <linux/time.h> #include <linux/namei.h> #include <linux/poll.h> +#include <linux/fiemap.h> static int bad_file_open (struct inode *inode,struct file *filp) {
fs/btrfs/extent_io.h +1 -0
@@ -5,6 +5,7 @@#include <linux/rbtree.h> #include <linux/refcount.h> +#include <linux/fiemap.h> #include "ulist.h" /*
fs/cifs/inode.c +1 -0
@@ -25,6 +25,7 @@#include <linux/freezer.h> #include <linux/sched/signal.h> #include <linux/wait_bit.h> +#include <linux/fiemap.h> #include <asm/div64.h> #include "cifsfs.h"
fs/cifs/smb2ops.c +1 -0
@@ -12,6 +12,7 @@#include <linux/uuid.h> #include <linux/sort.h> #include <crypto/aead.h> +#include <linux/fiemap.h> #include "cifsfs.h" #include "cifsglob.h" #include "smb2pdu.h"
fs/ext2/inode.c +1 -0
@@ -36,6 +36,7 @@#include <linux/iomap.h> #include <linux/namei.h> #include <linux/uio.h> +#include <linux/fiemap.h> #include "ext2.h" #include "acl.h" #include "xattr.h"
fs/ext4/ext4.h +1 -0
@@ -36,6 +36,7 @@#include <crypto/hash.h> #include <linux/falloc.h> #include <linux/percpu-rwsem.h> +#include <linux/fiemap.h> #ifdef __KERNEL__#include <linux/compat.h> #endif
fs/f2fs/data.c +1 -0
@@ -19,6 +19,7 @@#include <linux/uio.h> #include <linux/cleancache.h> #include <linux/sched/signal.h> +#include <linux/fiemap.h> #include "f2fs.h" #include "node.h"
fs/f2fs/inline.c +1 -0
@@ -8,6 +8,7 @@#include <linux/fs.h> #include <linux/f2fs_fs.h> +#include <linux/fiemap.h> #include "f2fs.h" #include "node.h"
fs/gfs2/inode.c +1 -0
@@ -17,6 +17,7 @@#include <linux/crc32.h> #include <linux/iomap.h> #include <linux/security.h> +#include <linux/fiemap.h> #include <linux/uaccess.h> #include "gfs2.h"
fs/hpfs/file.c +1 -0
@@ -9,6 +9,7 @@#include "hpfs_fn.h" #include <linux/mpage.h> +#include <linux/fiemap.h> #define BLOCKS (size) (((size) + 511) >> 9)
fs/ioctl.c +1 -0
@@ -18,6 +18,7 @@#include <linux/buffer_head.h> #include <linux/falloc.h> #include <linux/sched/signal.h> +#include <linux/fiemap.h> #include "internal.h"
fs/iomap/fiemap.c +1 -0
@@ -6,6 +6,7 @@#include <linux/compiler.h> #include <linux/fs.h> #include <linux/iomap.h> +#include <linux/fiemap.h> struct fiemap_ctx {struct fiemap_extent_info *fi;
fs/nilfs2/inode.c +1 -0
@@ -14,6 +14,7 @@#include <linux/pagemap.h> #include <linux/writeback.h> #include <linux/uio.h> +#include <linux/fiemap.h> #include "nilfs.h" #include "btnode.h" #include "segment.h"
fs/overlayfs/inode.c +1 -0
@@ -10,6 +10,7 @@#include <linux/xattr.h> #include <linux/posix_acl.h> #include <linux/ratelimit.h> +#include <linux/fiemap.h> #include "overlayfs.h"
fs/xfs/xfs_iops.c +1 -0
@@ -25,6 +25,7 @@#include <linux/posix_acl.h> #include <linux/security.h> #include <linux/iversion.h> +#include <linux/fiemap.h> /* * Directories have different lock order w.r.t. mmap_sem compared to regular
include/linux/fiemap.h new file +24 -0
@@ -0,0 +1,24 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +#ifndef _LINUX_FIEMAP_H +#define _LINUX_FIEMAP_H 1 + +#include <uapi/linux/fiemap.h> +#include <linux/fs.h> + +struct fiemap_extent_info { +unsigned int fi_flags;/* Flags as passed from user */ +unsigned int fi_extents_mapped;/* Number of mapped extents */ +unsigned int fi_extents_max;/* Size of fiemap_extent array */ +struct fiemap_extent __user *fi_extents_start;/* Start of +fiemap_extent array */ +}; + +int fiemap_fill_next_extent (struct fiemap_extent_info *info,u64 logical, +u64 phys,u64 len,u32 flags); +int fiemap_check_flags (struct fiemap_extent_info *fieinfo,u32 fs_flags); + +int generic_block_fiemap (struct inode *inode, +struct fiemap_extent_info *fieinfo,u64 start,u64 len, +get_block_t *get_block); + +#endif /* _LINUX_FIEMAP_H 1 */
include/linux/fs.h +1 -18
@@ -24,7 +24,6 @@#include <linux/capability.h> #include <linux/semaphore.h> #include <linux/fcntl.h> -#include <linux/fiemap.h> #include <linux/rculist_bl.h> #include <linux/atomic.h> #include <linux/shrinker.h>
@@ -48,6 +47,7 @@ struct backing_dev_info;struct bdi_writeback ;struct bio ;struct export_operations ; +struct fiemap_extent_info ;struct hd_geometry ;struct iovec ;struct kiocb ;
@@ -1745,19 +1745,6 @@ extern long compat_ptr_ioctl(struct file *file, unsigned int cmd,extern void inode_init_owner (struct inode *inode,const struct inode *dir,umode_t mode);extern bool may_open_dev (const struct path *path); -/* -* VFS FS_IOC_FIEMAP helper definitions. -*/ -struct fiemap_extent_info { -unsigned int fi_flags;/* Flags as passed from user */ -unsigned int fi_extents_mapped;/* Number of mapped extents */ -unsigned int fi_extents_max;/* Size of fiemap_extent array */ -struct fiemap_extent __user *fi_extents_start;/* Start of -fiemap_extent array */ -}; -int fiemap_fill_next_extent (struct fiemap_extent_info *info,u64 logical, -u64 phys,u64 len,u32 flags); -int fiemap_check_flags (struct fiemap_extent_info *fieinfo,u32 fs_flags);/* * This is the "filldir" function type, used by readdir() to let
@@ -3299,10 +3286,6 @@ static inline int vfs_fstat(int fd, struct kstat *stat)extern const char *vfs_get_link (struct dentry *,struct delayed_call *);extern int vfs_readlink (struct dentry *,char __user *,int ); -extern int generic_block_fiemap (struct inode *inode, -struct fiemap_extent_info *fieinfo,u64 start, -u64 len,get_block_t *get_block); -extern struct file_system_type *get_filesystem (struct file_system_type *fs);extern void put_filesystem (struct file_system_type *fs);extern struct file_system_type *get_fs_type (const char *name);
include/uapi/linux/fiemap.h +3 -3
@@ -9,8 +9,8 @@* Andreas Dilger <[email protected]> */ -#ifndef _LINUX_FIEMAP_H -#define _LINUX_FIEMAP_H +#ifndef _UAPI_LINUX_FIEMAP_H +#define _UAPI_LINUX_FIEMAP_H#include <linux/types.h>
@@ -67,4 +67,4 @@ struct fiemap {#define FIEMAP_EXTENT_SHARED 0x00002000/* Space shared with other * files. */ -#endif /* _LINUX_FIEMAP_H */ +#endif /* _UAPI_LINUX_FIEMAP_H */
Signature · 1 line
2.26.2
PATCH 5/9 iomap: fix the iomap_fiemap prototype
In reply to Christoph Hellwigiomap_fiemap should take u64 start and len arguments, just like the ->fiemap prototype.
--- fs/iomap/fiemap.c | 2 +- include/linux/iomap.h | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-)
fs/iomap/fiemap.c +1 -1
@@ -66,7 +66,7 @@ iomap_fiemap_actor(struct inode *inode, loff_t pos, loff_t length, void *data, }int iomap_fiemap (struct inode *inode,struct fiemap_extent_info *fi, -loff_t start,loff_t len,const struct iomap_ops *ops) +u64 start,u64 len,const struct iomap_ops *ops) {struct fiemap_ctx ctx;loff_t ret;
include/linux/iomap.h +1 -1
@@ -178,7 +178,7 @@ int iomap_truncate_page(struct inode *inode, loff_t pos, bool *did_zero,vm_fault_t iomap_page_mkwrite (struct vm_fault *vmf,const struct iomap_ops *ops);int iomap_fiemap (struct inode *inode,struct fiemap_extent_info *fieinfo, -loff_t start,loff_t len,const struct iomap_ops *ops); +u64 start,u64 len,const struct iomap_ops *ops);loff_t iomap_seek_hole (struct inode *inode,loff_t offset,const struct iomap_ops *ops); loff_t iomap_seek_data(struct inode *inode, loff_t offset,
Signature · 1 line
2.26.2
PATCH 6/9 fs: move fiemap range validation into the file systems instances
In reply to Christoph HellwigReplace fiemap_check_flags with a fiemap_prep helper that also takes the inode and mapped range, and performs the sanity check and truncation previously done in fiemap_check_range. This way the validation is inside the file system itself and thus properly works for the stacked overlayfs case as well.
--- Documentation/filesystems/fiemap.txt | 12 +++--- fs/btrfs/inode.c | 2 +- fs/cifs/smb2ops.c | 6 ++- fs/ext4/extents.c | 5 ++- fs/f2fs/data.c | 3 +- fs/ioctl.c | 63 +++++++++++----------------- fs/iomap/fiemap.c | 2 +- fs/nilfs2/inode.c | 2 +- fs/ocfs2/extent_map.c | 3 +- include/linux/fiemap.h | 3 +- 10 files changed, 47 insertions(+), 54 deletions(-)
Documentation/filesystems/fiemap.txt +7 -5
@@ -203,16 +203,18 @@ EINTR once fatal signal received. Flag checking should be done at the beginning of the ->fiemap callback via the -fiemap_check_flags() helper: +fiemap_prep() helper: -int fiemap_check_flags(struct fiemap_extent_info *fieinfo, u32 fs_flags); +int fiemap_prep(struct inode *inode, struct fiemap_extent_info *fieinfo, + u64 start, u64 *len, u32 supported_flags); The struct fieinfo should be passed in as received from ioctl_fiemap(). The set of fiemap flags which the fs understands should be passed via fs_flags. If -fiemap_check_flags finds invalid user flags, it will place the bad values in +fiemap_prep finds invalid user flags, it will place the bad values in fieinfo->fi_flags and return -EBADR. If the file system gets -EBADR, from -fiemap_check_flags(), it should immediately exit, returning that error back to -ioctl_fiemap(). +fiemap_prep(), it should immediately exit, returning that error back to +ioctl_fiemap(). Additionally the range is validate against the supported +maximum file size. For each extent in the request range, the file system should call
fs/btrfs/inode.c +1 -1
@@ -8250,7 +8250,7 @@ static int btrfs_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo, {int ret; - ret =fiemap_check_flags (fieinfo,BTRFS_FIEMAP_FLAGS ); + ret =fiemap_prep (inode, fieinfo, start, &len,BTRFS_FIEMAP_FLAGS );if (ret)return ret;
fs/cifs/smb2ops.c +4 -2
@@ -3408,8 +3408,10 @@ static int smb3_fiemap(struct cifs_tcon *tcon,int i, num, rc, flags, last_blob;u64 next; -if (fiemap_check_flags (fei,FIEMAP_FLAG_SYNC )) -return -EBADR ; + rc =fiemap_prep (d_inode (cfile->dentry), fei, start, &len, +FIEMAP_FLAG_SYNC ); +if (rc) +return rc; xid =get_xid (); again:
fs/ext4/extents.c +3 -2
@@ -4908,8 +4908,9 @@ int ext4_get_es_cache(struct inode *inode, struct fiemap_extent_info *fieinfo, fieinfo->fi_flags &= ~FIEMAP_FLAG_CACHE ; } -if (fiemap_check_flags (fieinfo,FIEMAP_FLAG_SYNC )) -return -EBADR ; + error =fiemap_prep (inode, fieinfo, start, &len,FIEMAP_FLAG_SYNC ); +if (error) +return error; error =ext4_fiemap_check_ranges (inode, start, &len);if (error)
fs/f2fs/data.c +2 -1
@@ -1825,7 +1825,8 @@ int f2fs_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,return ret; } - ret =fiemap_check_flags (fieinfo,FIEMAP_FLAG_SYNC |FIEMAP_FLAG_XATTR ); + ret =fiemap_prep (inode, fieinfo, start, &len, +FIEMAP_FLAG_SYNC |FIEMAP_FLAG_XATTR );if (ret)return ret;
fs/ioctl.c +24 -39
110 lines not on this page · raw message
fs/iomap/fiemap.c +1 -1
@@ -75,7 +75,7 @@ int iomap_fiemap(struct inode *inode, struct fiemap_extent_info *fi, ctx.fi = fi; ctx.prev.type =IOMAP_HOLE ; - ret =fiemap_check_flags (fi,FIEMAP_FLAG_SYNC ); + ret =fiemap_prep (inode, fi, start, &len,FIEMAP_FLAG_SYNC );if (ret)return ret;
fs/nilfs2/inode.c +1 -1
@@ -1006,7 +1006,7 @@ int nilfs_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,unsigned int blkbits = inode->i_blkbits;int ret, n; - ret =fiemap_check_flags (fieinfo,FIEMAP_FLAG_SYNC ); + ret =fiemap_prep (inode, fieinfo, start, &len,FIEMAP_FLAG_SYNC );if (ret)return ret;
fs/ocfs2/extent_map.c +2 -1
@@ -746,7 +746,8 @@ int ocfs2_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,struct buffer_head *di_bh =NULL ;struct ocfs2_extent_rec rec; - ret =fiemap_check_flags (fieinfo,OCFS2_FIEMAP_FLAGS ); + ret =fiemap_prep (inode, fieinfo, map_start, &map_len, +OCFS2_FIEMAP_FLAGS );if (ret)return ret;
include/linux/fiemap.h +2 -1
@@ -13,9 +13,10 @@ struct fiemap_extent_info {fiemap_extent array */ }; +int fiemap_prep (struct inode *inode,struct fiemap_extent_info *fieinfo, +u64 start,u64 *len,u32 supported_flags);int fiemap_fill_next_extent (struct fiemap_extent_info *info,u64 logical,u64 phys,u64 len,u32 flags); -int fiemap_check_flags (struct fiemap_extent_info *fieinfo,u32 fs_flags); int generic_block_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo, u64 start, u64 len,
Signature · 1 line
2.26.2
Re: [PATCH 6/9] fs: move fiemap range validation into the file systems instances
Quoted 8 lines · Christoph Hellwig
Replace
fiemap_check_flagswith afiemap_prephelper that also takes the inode and mapped range, and performs the sanity check and truncation previously done infiemap_check_range. This way the validation is inside the file system itself and thus properly works for the stacked overlayfs case as well.Signed-off-by Christoph Hellwig <[email protected]>Reviewed-by Amir Goldstein <[email protected]>
Still looks ok,
--D
Quoted 278 lines
--- Documentation/filesystems/fiemap.txt | 12 +++--- fs/btrfs/inode.c | 2 +- fs/cifs/smb2ops.c | 6 ++- fs/ext4/extents.c | 5 ++- fs/f2fs/data.c | 3 +- fs/ioctl.c | 63 +++++++++++----------------- fs/iomap/fiemap.c | 2 +- fs/nilfs2/inode.c | 2 +- fs/ocfs2/extent_map.c | 3 +- include/linux/fiemap.h | 3 +- 10 files changed, 47 insertions(+), 54 deletions(-)Documentation/filesystems/fiemap.txt +7
-5@@ -203,16 +203,18 @@ EINTR once fatal signal received. Flag checking should be done at the beginning of the ->fiemap callback via the -fiemap_check_flags() helper: +fiemap_prep() helper: -int fiemap_check_flags(struct fiemap_extent_info *fieinfo, u32 fs_flags); +int fiemap_prep(struct inode *inode, struct fiemap_extent_info *fieinfo, + u64 start, u64 *len, u32 supported_flags); The struct fieinfo should be passed in as received from ioctl_fiemap(). The set of fiemap flags which the fs understands should be passed via fs_flags. If -fiemap_check_flags finds invalid user flags, it will place the bad values in +fiemap_prep finds invalid user flags, it will place the bad values in fieinfo->fi_flags and return -EBADR. If the file system gets -EBADR, from -fiemap_check_flags(), it should immediately exit, returning that error back to -ioctl_fiemap(). +fiemap_prep(), it should immediately exit, returning that error back to +ioctl_fiemap(). Additionally the range is validate against the supported +maximum file size. For each extent in the request range, the file system should callfs/btrfs/inode.c +1
-1@@ -8250,7 +8250,7 @@ static int btrfs_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo, {int ret; - ret =fiemap_check_flags (fieinfo,BTRFS_FIEMAP_FLAGS ); + ret =fiemap_prep (inode, fieinfo, start, &len,BTRFS_FIEMAP_FLAGS );if (ret)return ret;fs/cifs/smb2ops.c +4
-2@@ -3408,8 +3408,10 @@ static int smb3_fiemap(struct cifs_tcon *tcon,int i, num, rc, flags, last_blob;u64 next; -if (fiemap_check_flags (fei,FIEMAP_FLAG_SYNC )) -return -EBADR ; + rc =fiemap_prep (d_inode (cfile->dentry), fei, start, &len, +FIEMAP_FLAG_SYNC ); +if (rc) +return rc; xid =get_xid (); again:fs/ext4/extents.c +3
-2@@ -4908,8 +4908,9 @@ int ext4_get_es_cache(struct inode *inode, struct fiemap_extent_info *fieinfo, fieinfo->fi_flags &= ~FIEMAP_FLAG_CACHE ; } -if (fiemap_check_flags (fieinfo,FIEMAP_FLAG_SYNC )) -return -EBADR ; + error =fiemap_prep (inode, fieinfo, start, &len,FIEMAP_FLAG_SYNC ); +if (error) +return error; error =ext4_fiemap_check_ranges (inode, start, &len);if (error)fs/f2fs/data.c +2
-1@@ -1825,7 +1825,8 @@ int f2fs_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,return ret; } - ret =fiemap_check_flags (fieinfo,FIEMAP_FLAG_SYNC |FIEMAP_FLAG_XATTR ); + ret =fiemap_prep (inode, fieinfo, start, &len, +FIEMAP_FLAG_SYNC |FIEMAP_FLAG_XATTR );if (ret)return ret;fs/ioctl.c +24
-39@@ -149,61 +149,50 @@ int fiemap_fill_next_extent(struct fiemap_extent_info *fieinfo, u64 logical,EXPORT_SYMBOL (fiemap_fill_next_extent);/** -* fiemap_check_flags - check validity of requested flags for fiemap +* fiemap_prep - check validity of requested flags for fiemap +* @inode: Inode to operate on * @fieinfo: Fiemap context passed into ->fiemap -* @fs_flags: Set of fiemap flags that the file system understands +* @start: Start of the mapped range +* @len: Length of the mapped range, can be truncated by this function. +* @supported_flags: Set of fiemap flags that the file system understands * -* Called from file system ->fiemap callback. This will compute the -* intersection of valid fiemap flags and those that the fs supports. That -* value is then compared against the user supplied flags. In case of bad user -* flags, the invalid values will be written into the fieinfo structure, and -* -EBADR is returned, which tells ioctl_fiemap() to return those values to -* userspace. For this reason, a return code of -EBADR should be preserved. +* This function must be called from each ->fiemap instance to validate the +* fiemap request against the file system parameters. * -* Returns 0 on success, -EBADR on bad flags. +* Returns 0 on success, or a negative error on failure. */ -int fiemap_check_flags (struct fiemap_extent_info *fieinfo,u32 fs_flags) +int fiemap_prep (struct inode *inode,struct fiemap_extent_info *fieinfo, +u64 start,u64 *len,u32 supported_flags) { +u64 maxbytes = inode->i_sb->s_maxbytes;u32 incompat_flags; - incompat_flags = fieinfo->fi_flags & ~(FIEMAP_FLAGS_COMPAT & fs_flags); -if (incompat_flags) { - fieinfo->fi_flags = incompat_flags; -return -EBADR ; - } -return 0 ; -} -EXPORT_SYMBOL (fiemap_check_flags); - -static int fiemap_check_ranges (struct super_block *sb, -u64 start,u64 len,u64 *new_len) -{ -u64 maxbytes = (u64 ) sb->s_maxbytes; - - *new_len = len; - -if (len ==0 ) +if (*len ==0 )return -EINVAL ; -if (start > maxbytes)return -EFBIG ;/* * Shrink request scope to what the fs can actually handle. */ -if (len > maxbytes || (maxbytes - len) < start) - *new_len = maxbytes - start; +if (*len > maxbytes || (maxbytes - *len) < start) + *len = maxbytes - start; + supported_flags &=FIEMAP_FLAGS_COMPAT ; + incompat_flags = fieinfo->fi_flags & ~supported_flags; +if (incompat_flags) { + fieinfo->fi_flags = incompat_flags; +return -EBADR ; + }return 0 ; } +EXPORT_SYMBOL (fiemap_prep);static int ioctl_fiemap (struct file *filp,struct fiemap __user *ufiemap) {struct fiemap fiemap;struct fiemap_extent_info fieinfo = {0 , };struct inode *inode =file_inode (filp); -struct super_block *sb = inode->i_sb; -u64 len;int error;if (!inode->i_op->fiemap)@@ -215,11 +204,6 @@ static int ioctl_fiemap(struct file *filp, struct fiemap __user *ufiemap)if (fiemap.fm_extent_count >FIEMAP_MAX_EXTENTS )return -EINVAL ; - error =fiemap_check_ranges (sb, fiemap.fm_start, fiemap.fm_length, - &len); -if (error) -return error; - fieinfo.fi_flags = fiemap.fm_flags; fieinfo.fi_extents_max = fiemap.fm_extent_count; fieinfo.fi_extents_start = ufiemap->fm_extents;@@ -232,7 +216,8 @@ static int ioctl_fiemap(struct file *filp, struct fiemap __user *ufiemap)if (fieinfo.fi_flags &FIEMAP_FLAG_SYNC )filemap_write_and_wait (inode->i_mapping); - error = inode->i_op->fiemap (inode, &fieinfo, fiemap.fm_start, len); + error = inode->i_op->fiemap (inode, &fieinfo, fiemap.fm_start, + fiemap.fm_length); fiemap.fm_flags = fieinfo.fi_flags; fiemap.fm_mapped_extents = fieinfo.fi_extents_mapped;if (copy_to_user (ufiemap, &fiemap,sizeof (fiemap)))@@ -320,7 +305,7 @@ static int __generic_block_fiemap(struct inode *inode,bool past_eof = false, whole_file = false;int ret =0 ; - ret =fiemap_check_flags (fieinfo,FIEMAP_FLAG_SYNC ); + ret =fiemap_prep (inode, fieinfo, start, &len,FIEMAP_FLAG_SYNC );if (ret)return ret;fs/iomap/fiemap.c +1
-1@@ -75,7 +75,7 @@ int iomap_fiemap(struct inode *inode, struct fiemap_extent_info *fi, ctx.fi = fi; ctx.prev.type =IOMAP_HOLE ; - ret =fiemap_check_flags (fi,FIEMAP_FLAG_SYNC ); + ret =fiemap_prep (inode, fi, start, &len,FIEMAP_FLAG_SYNC );if (ret)return ret;fs/nilfs2/inode.c +1
-1@@ -1006,7 +1006,7 @@ int nilfs_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,unsigned int blkbits = inode->i_blkbits;int ret, n; - ret =fiemap_check_flags (fieinfo,FIEMAP_FLAG_SYNC ); + ret =fiemap_prep (inode, fieinfo, start, &len,FIEMAP_FLAG_SYNC );if (ret)return ret;fs/ocfs2/extent_map.c +2
-1@@ -746,7 +746,8 @@ int ocfs2_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,struct buffer_head *di_bh =NULL ;struct ocfs2_extent_rec rec; - ret =fiemap_check_flags (fieinfo,OCFS2_FIEMAP_FLAGS ); + ret =fiemap_prep (inode, fieinfo, map_start, &map_len, +OCFS2_FIEMAP_FLAGS );if (ret)return ret;include/linux/fiemap.h +2
-1@@ -13,9 +13,10 @@ struct fiemap_extent_info {fiemap_extent array */ }; +int fiemap_prep (struct inode *inode,struct fiemap_extent_info *fieinfo, +u64 start,u64 *len,u32 supported_flags);int fiemap_fill_next_extent (struct fiemap_extent_info *info,u64 logical,u64 phys,u64 len,u32 flags); -int fiemap_check_flags (struct fiemap_extent_info *fieinfo,u32 fs_flags); int generic_block_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo, u64 start, u64 len,Signature · 1 line
2.26.2
PATCH 7/9 fs: handle FIEMAP_FLAG_SYNC in fiemap_prep
In reply to Christoph HellwigBy moving FIEMAP_FLAG_SYNC handling to fiemap_prep we ensure it is handled once instead of duplicated, but can still be done under fs locks, like xfs/iomap intended with its duplicate handling. Also make sure the error value of filemap_write_and_wait is propagated to user space.
--- fs/btrfs/inode.c | 4 +--- fs/cifs/smb2ops.c | 3 +-- fs/ext4/extents.c | 2 +- fs/ext4/ioctl.c | 3 --- fs/f2fs/data.c | 3 +-- fs/ioctl.c | 10 ++++++---- fs/iomap/fiemap.c | 8 +------- fs/nilfs2/inode.c | 2 +- fs/ocfs2/extent_map.c | 5 +---- fs/overlayfs/inode.c | 4 ---- 10 files changed, 13 insertions(+), 31 deletions(-)
fs/btrfs/inode.c +1 -3
@@ -8243,14 +8243,12 @@ static ssize_t btrfs_direct_IO(struct kiocb *iocb, struct iov_iter *iter)return ret; } -#define BTRFS_FIEMAP_FLAGS (FIEMAP_FLAG_SYNC) -static int btrfs_fiemap (struct inode *inode,struct fiemap_extent_info *fieinfo,__u64 start,__u64 len) {int ret; - ret =fiemap_prep (inode, fieinfo, start, &len,BTRFS_FIEMAP_FLAGS ); + ret =fiemap_prep (inode, fieinfo, start, &len,0 );if (ret)return ret;
fs/cifs/smb2ops.c +1 -2
@@ -3408,8 +3408,7 @@ static int smb3_fiemap(struct cifs_tcon *tcon,int i, num, rc, flags, last_blob;u64 next; - rc =fiemap_prep (d_inode (cfile->dentry), fei, start, &len, -FIEMAP_FLAG_SYNC ); + rc =fiemap_prep (d_inode (cfile->dentry), fei, start, &len,0 );if (rc)return rc;
fs/ext4/extents.c +1 -1
@@ -4908,7 +4908,7 @@ int ext4_get_es_cache(struct inode *inode, struct fiemap_extent_info *fieinfo, fieinfo->fi_flags &= ~FIEMAP_FLAG_CACHE ; } - error =fiemap_prep (inode, fieinfo, start, &len,FIEMAP_FLAG_SYNC ); + error =fiemap_prep (inode, fieinfo, start, &len,0 );if (error)return error;
fs/ext4/ioctl.c +0 -3
@@ -759,9 +759,6 @@ static int ext4_ioctl_get_es_cache(struct file *filp, unsigned long arg) fieinfo.fi_extents_max * sizeof(struct fiemap_extent)))return -EFAULT ; -if (fieinfo.fi_flags &FIEMAP_FLAG_SYNC ) -filemap_write_and_wait (inode->i_mapping); - error =ext4_get_es_cache (inode, &fieinfo, fiemap.fm_start, fiemap.fm_length); fiemap.fm_flags = fieinfo.fi_flags;
fs/f2fs/data.c +1 -2
@@ -1825,8 +1825,7 @@ int f2fs_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,return ret; } - ret =fiemap_prep (inode, fieinfo, start, &len, -FIEMAP_FLAG_SYNC |FIEMAP_FLAG_XATTR ); + ret =fiemap_prep (inode, fieinfo, start, &len,FIEMAP_FLAG_XATTR );if (ret)return ret;
fs/ioctl.c +6 -4
@@ -166,6 +166,7 @@ int fiemap_prep(struct inode *inode, struct fiemap_extent_info *fieinfo, {u64 maxbytes = inode->i_sb->s_maxbytes;u32 incompat_flags; +int ret =0 ;if (*len ==0 )return -EINVAL ;
@@ -178,13 +179,17 @@ int fiemap_prep(struct inode *inode, struct fiemap_extent_info *fieinfo,if (*len > maxbytes || (maxbytes - *len) < start) *len = maxbytes - start; + supported_flags |=FIEMAP_FLAG_SYNC ; supported_flags &=FIEMAP_FLAGS_COMPAT ; incompat_flags = fieinfo->fi_flags & ~supported_flags;if (incompat_flags) { fieinfo->fi_flags = incompat_flags;return -EBADR ; } -return 0 ; + +if (fieinfo->fi_flags &FIEMAP_FLAG_SYNC ) + ret =filemap_write_and_wait (inode->i_mapping); +return ret; }EXPORT_SYMBOL (fiemap_prep);
@@ -213,9 +218,6 @@ static int ioctl_fiemap(struct file *filp, struct fiemap __user *ufiemap) fieinfo.fi_extents_max * sizeof(struct fiemap_extent)))return -EFAULT ; -if (fieinfo.fi_flags &FIEMAP_FLAG_SYNC ) -filemap_write_and_wait (inode->i_mapping); - error = inode->i_op->fiemap (inode, &fieinfo, fiemap.fm_start, fiemap.fm_length); fiemap.fm_flags = fieinfo.fi_flags;
fs/iomap/fiemap.c +1 -7
@@ -75,16 +75,10 @@ int iomap_fiemap(struct inode *inode, struct fiemap_extent_info *fi, ctx.fi = fi; ctx.prev.type =IOMAP_HOLE ; - ret =fiemap_prep (inode, fi, start, &len,FIEMAP_FLAG_SYNC ); + ret =fiemap_prep (inode, fi, start, &len,0 );if (ret)return ret; -if (fi->fi_flags &FIEMAP_FLAG_SYNC ) { - ret =filemap_write_and_wait (inode->i_mapping); -if (ret) -return ret; - } -while (len >0 ) { ret =iomap_apply (inode, start, len,IOMAP_REPORT , ops, &ctx, iomap_fiemap_actor);
fs/nilfs2/inode.c +1 -1
@@ -1006,7 +1006,7 @@ int nilfs_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,unsigned int blkbits = inode->i_blkbits;int ret, n; - ret =fiemap_prep (inode, fieinfo, start, &len,FIEMAP_FLAG_SYNC ); + ret =fiemap_prep (inode, fieinfo, start, &len,0 );if (ret)return ret;
fs/ocfs2/extent_map.c +1 -4
@@ -733,8 +733,6 @@ static int ocfs2_fiemap_inline(struct inode *inode, struct buffer_head *di_bh,return 0 ; } -#define OCFS2_FIEMAP_FLAGS (FIEMAP_FLAG_SYNC) -int ocfs2_fiemap (struct inode *inode,struct fiemap_extent_info *fieinfo,u64 map_start,u64 map_len) {
@@ -746,8 +744,7 @@ int ocfs2_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,struct buffer_head *di_bh =NULL ;struct ocfs2_extent_rec rec; - ret =fiemap_prep (inode, fieinfo, map_start, &map_len, -OCFS2_FIEMAP_FLAGS ); + ret =fiemap_prep (inode, fieinfo, map_start, &map_len,0 );if (ret)return ret;
fs/overlayfs/inode.c +0 -4
@@ -480,10 +480,6 @@ static int ovl_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,return -EOPNOTSUPP ; old_cred =ovl_override_creds (inode->i_sb); - -if (fieinfo->fi_flags &FIEMAP_FLAG_SYNC ) -filemap_write_and_wait (realinode->i_mapping); - err = realinode->i_op->fiemap (realinode, fieinfo, start, len);revert_creds (old_cred);
Signature · 1 line
2.26.2
PATCH 8/9 fs: remove the access_ok() check in ioctl_fiemap
In reply to Christoph Hellwigaccess_ok just checks we are fed a proper user pointer. We also do that in copy_to_user itself, so no need to do this early.
--- fs/ioctl.c | 6 +----- 1 file changed, 1 insertion(+), 5 deletions(-)
fs/ioctl.c +1 -5
@@ -213,13 +213,9 @@ static int ioctl_fiemap(struct file *filp, struct fiemap __user *ufiemap) fieinfo.fi_extents_max = fiemap.fm_extent_count; fieinfo.fi_extents_start = ufiemap->fm_extents; -if (fiemap.fm_extent_count !=0 && - !access_ok (fieinfo.fi_extents_start, - fieinfo.fi_extents_max *sizeof (struct fiemap_extent ))) -return -EFAULT ; - error = inode->i_op->fiemap (inode, &fieinfo, fiemap.fm_start, fiemap.fm_length); + fiemap.fm_flags = fieinfo.fi_flags; fiemap.fm_mapped_extents = fieinfo.fi_extents_mapped;if (copy_to_user (ufiemap, &fiemap,sizeof (fiemap)))
Signature · 1 line
2.26.2
PATCH 9/9 ext4: remove the access_ok() check in ext4_ioctl_get_es_cache
In reply to Christoph Hellwigaccess_ok just checks we are fed a proper user pointer. We also do that in copy_to_user itself, so no need to do this early.
--- fs/ext4/ioctl.c | 5 ----- 1 file changed, 5 deletions(-)
fs/ext4/ioctl.c +0 -5
@@ -754,11 +754,6 @@ static int ext4_ioctl_get_es_cache(struct file *filp, unsigned long arg) fieinfo.fi_extents_max = fiemap.fm_extent_count; fieinfo.fi_extents_start = ufiemap->fm_extents; -if (fiemap.fm_extent_count !=0 && - !access_ok (fieinfo.fi_extents_start, - fieinfo.fi_extents_max *sizeof (struct fiemap_extent ))) -return -EFAULT ; - error =ext4_get_es_cache (inode, &fieinfo, fiemap.fm_start, fiemap.fm_length); fiemap.fm_flags = fieinfo.fi_flags;
Signature · 1 line
2.26.2
Al Viro
maintainerQuoted 18 lines · Christoph Hellwig
Hi all,
This series cleans up the fiemap support in ext4 and in general.
Ted or Al, can one of you pick this up? It touches both ext4 and core code, so either tree could work.
Changes since v3:
- dropped the fixes that have been merged int mainlineChanges since v2:
- commit message typo
- doc updates
- used_inodein cifs
- add a missing return statement in cifs
- remove thefilemap_write_and_waitcall fromext4_ioctl_get_es_cache
Quoted 6 lines
Changes since v1:
- renamefiemap_validatetofiemap_prep
- liftFIEMAP_FLAG_SYNChandling to common code
- add a newlinux/fiemap.hheader
- remove__generic_block_fiemap
- removeaccess_okcalls from fiemap and ext4
Hmmm... I can do an immutable shared branch, no problem. What would you prefer for a branchpoint for that one?
Quoted 2 lines · Al Viro
Hmmm... I can do an immutable shared branch, no problem. What would you prefer for a branchpoint for that one?
This needs pretty much latest Linus' tree as it sits on top of the bug fixes merged this week. But then again I'm not sure we even need a shared tree, as there shouldn't be any conflicts. But looking at linux-next it seems Ted has actually pulled in the whole series, so I think we actually are done.
But while I've got you attention: Can you take a look at my "clean up kernel_{read,write} & friends v2" series?
Theodore Y. Ts'o
maintainerQuoted 26 lines · Al Viro
Quoted 18 lines · Christoph Hellwig
Hi all,
This series cleans up the fiemap support in ext4 and in general.
Ted or Al, can one of you pick this up? It touches both ext4 and core code, so either tree could work.
Changes since v3:
- dropped the fixes that have been merged int mainlineChanges since v2:
- commit message typo
- doc updates
- used_inodein cifs
- add a missing return statement in cifs
- remove thefilemap_write_and_waitcall fromext4_ioctl_get_es_cacheQuoted 6 lines
Changes since v1:
- renamefiemap_validatetofiemap_prep
- liftFIEMAP_FLAG_SYNChandling to common code
- add a newlinux/fiemap.hheader
- remove__generic_block_fiemap
- removeaccess_okcalls from fiemap and ext4
Quoted 2 lines
Hmmm... I can do an immutable shared branch, no problem. What would you prefer for a branchpoint for that one?
I thought we had already agreed to run these patches through the ext4 git tree, since most of the changes affect the ext4 tree (and there aren't any other iomap fiemap changes pending as far as I know).
The v3 versions of these patches have been part of the ext4 dev tree since May 19th. Since the ext4 dev tree is rewinding, I can easily update it fiemap-fixes patch to be on top of the first two patches which Linus has already accepted, and then merge it into the ext4 dev branch.
- Ted
Quoted 7 lines · Theodore Y. Ts'o
Quoted 2 lines
Hmmm... I can do an immutable shared branch, no problem. What would you prefer for a branchpoint for that one?
I thought we had already agreed to run these patches through the ext4 git tree, since most of the changes affect the ext4 tree (and there aren't any other iomap fiemap changes pending as far as I know).
Quoted 5 lines
The v3 versions of these patches have been part of the ext4 dev tree since May 19th. Since the ext4 dev tree is rewinding, I can easily update it fiemap-fixes patch to be on top of the first two patches which Linus has already accepted, and then merge it into the ext4 dev branch.
I think we're all fine and set, unless Al really needs a shared branch. Based on what is in linux-next and out on the lists I don't see any real need, though.