mirror of
https://github.com/torvalds/linux.git
synced 2024-11-15 08:31:55 +00:00
a11f7e63c5
Fix a corruption that can happen when we have (two or more) outstanding
aio's to an overlapping unaligned region. Ext4
(e9e3bcecf4) and xfs recently had to fix
similar issues.
In our case what happens is that we can have an outstanding aio on a region
and if a write comes in with some bytes overlapping the original aio we may
decide to read that region into a page before continuing (typically because
of buffered-io fallback). Since we have no ordering guarantees with the
aio, we can read stale or bad data into the page and then write it back out.
If the i/o is page and block aligned, then we avoid this issue as there
won't be any need to read data from disk.
I took the same approach as Eric in the ext4 patch and introduced some
serialization of unaligned async direct i/o. I don't expect this to have an
effect on the most common cases of AIO. Unaligned aio will be slower
though, but that's far more acceptable than data corruption.
Signed-off-by: Mark Fasheh <mfasheh@suse.com>
Signed-off-by: Joel Becker <jlbec@evilplan.org>
184 lines
6.1 KiB
C
184 lines
6.1 KiB
C
/* -*- mode: c; c-basic-offset: 8; -*-
|
|
* vim: noexpandtab sw=8 ts=8 sts=0:
|
|
*
|
|
* inode.h
|
|
*
|
|
* Function prototypes
|
|
*
|
|
* Copyright (C) 2002, 2004 Oracle. All rights reserved.
|
|
*
|
|
* This program is free software; you can redistribute it and/or
|
|
* modify it under the terms of the GNU General Public
|
|
* License as published by the Free Software Foundation; either
|
|
* version 2 of the License, or (at your option) any later version.
|
|
*
|
|
* This program is distributed in the hope that it will be useful,
|
|
* but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
|
|
* General Public License for more details.
|
|
*
|
|
* You should have received a copy of the GNU General Public
|
|
* License along with this program; if not, write to the
|
|
* Free Software Foundation, Inc., 59 Temple Place - Suite 330,
|
|
* Boston, MA 021110-1307, USA.
|
|
*/
|
|
|
|
#ifndef OCFS2_INODE_H
|
|
#define OCFS2_INODE_H
|
|
|
|
#include "extent_map.h"
|
|
|
|
/* OCFS2 Inode Private Data */
|
|
struct ocfs2_inode_info
|
|
{
|
|
u64 ip_blkno;
|
|
|
|
struct ocfs2_lock_res ip_rw_lockres;
|
|
struct ocfs2_lock_res ip_inode_lockres;
|
|
struct ocfs2_lock_res ip_open_lockres;
|
|
|
|
/* protects allocation changes on this inode. */
|
|
struct rw_semaphore ip_alloc_sem;
|
|
|
|
/* protects extended attribute changes on this inode */
|
|
struct rw_semaphore ip_xattr_sem;
|
|
|
|
/* Number of outstanding AIO's which are not page aligned */
|
|
atomic_t ip_unaligned_aio;
|
|
|
|
/* These fields are protected by ip_lock */
|
|
spinlock_t ip_lock;
|
|
u32 ip_open_count;
|
|
struct list_head ip_io_markers;
|
|
u32 ip_clusters;
|
|
|
|
u16 ip_dyn_features;
|
|
struct mutex ip_io_mutex;
|
|
u32 ip_flags; /* see below */
|
|
u32 ip_attr; /* inode attributes */
|
|
|
|
/* protected by recovery_lock. */
|
|
struct inode *ip_next_orphan;
|
|
|
|
struct ocfs2_caching_info ip_metadata_cache;
|
|
struct ocfs2_extent_map ip_extent_map;
|
|
struct inode vfs_inode;
|
|
struct jbd2_inode ip_jinode;
|
|
|
|
u32 ip_dir_start_lookup;
|
|
|
|
/* Only valid if the inode is the dir. */
|
|
u32 ip_last_used_slot;
|
|
u64 ip_last_used_group;
|
|
u32 ip_dir_lock_gen;
|
|
|
|
struct ocfs2_alloc_reservation ip_la_data_resv;
|
|
};
|
|
|
|
/*
|
|
* Flags for the ip_flags field
|
|
*/
|
|
/* System file inodes */
|
|
#define OCFS2_INODE_SYSTEM_FILE 0x00000001
|
|
#define OCFS2_INODE_JOURNAL 0x00000002
|
|
#define OCFS2_INODE_BITMAP 0x00000004
|
|
/* This inode has been wiped from disk */
|
|
#define OCFS2_INODE_DELETED 0x00000008
|
|
/* Another node is deleting, so our delete is a nop */
|
|
#define OCFS2_INODE_SKIP_DELETE 0x00000010
|
|
/* Has the inode been orphaned on another node?
|
|
*
|
|
* This hints to ocfs2_drop_inode that it should clear i_nlink before
|
|
* continuing.
|
|
*
|
|
* We *only* set this on unlink vote from another node. If the inode
|
|
* was locally orphaned, then we're sure of the state and don't need
|
|
* to twiddle i_nlink later - it's either zero or not depending on
|
|
* whether our unlink succeeded. Otherwise we got this from a node
|
|
* whose intention was to orphan the inode, however he may have
|
|
* crashed, failed etc, so we let ocfs2_drop_inode zero the value and
|
|
* rely on ocfs2_delete_inode to sort things out under the proper
|
|
* cluster locks.
|
|
*/
|
|
#define OCFS2_INODE_MAYBE_ORPHANED 0x00000020
|
|
/* Does someone have the file open O_DIRECT */
|
|
#define OCFS2_INODE_OPEN_DIRECT 0x00000040
|
|
/* Tell the inode wipe code it's not in orphan dir */
|
|
#define OCFS2_INODE_SKIP_ORPHAN_DIR 0x00000080
|
|
|
|
static inline struct ocfs2_inode_info *OCFS2_I(struct inode *inode)
|
|
{
|
|
return container_of(inode, struct ocfs2_inode_info, vfs_inode);
|
|
}
|
|
|
|
#define INODE_JOURNAL(i) (OCFS2_I(i)->ip_flags & OCFS2_INODE_JOURNAL)
|
|
#define SET_INODE_JOURNAL(i) (OCFS2_I(i)->ip_flags |= OCFS2_INODE_JOURNAL)
|
|
|
|
extern struct kmem_cache *ocfs2_inode_cache;
|
|
|
|
extern const struct address_space_operations ocfs2_aops;
|
|
extern const struct ocfs2_caching_operations ocfs2_inode_caching_ops;
|
|
|
|
static inline struct ocfs2_caching_info *INODE_CACHE(struct inode *inode)
|
|
{
|
|
return &OCFS2_I(inode)->ip_metadata_cache;
|
|
}
|
|
|
|
void ocfs2_evict_inode(struct inode *inode);
|
|
int ocfs2_drop_inode(struct inode *inode);
|
|
|
|
/* Flags for ocfs2_iget() */
|
|
#define OCFS2_FI_FLAG_SYSFILE 0x1
|
|
#define OCFS2_FI_FLAG_ORPHAN_RECOVERY 0x2
|
|
struct inode *ocfs2_ilookup(struct super_block *sb, u64 feoff);
|
|
struct inode *ocfs2_iget(struct ocfs2_super *osb, u64 feoff, unsigned flags,
|
|
int sysfile_type);
|
|
int ocfs2_inode_init_private(struct inode *inode);
|
|
int ocfs2_inode_revalidate(struct dentry *dentry);
|
|
void ocfs2_populate_inode(struct inode *inode, struct ocfs2_dinode *fe,
|
|
int create_ino);
|
|
void ocfs2_read_inode(struct inode *inode);
|
|
void ocfs2_read_inode2(struct inode *inode, void *opaque);
|
|
ssize_t ocfs2_rw_direct(int rw, struct file *filp, char *buf,
|
|
size_t size, loff_t *offp);
|
|
void ocfs2_sync_blockdev(struct super_block *sb);
|
|
void ocfs2_refresh_inode(struct inode *inode,
|
|
struct ocfs2_dinode *fe);
|
|
int ocfs2_mark_inode_dirty(handle_t *handle,
|
|
struct inode *inode,
|
|
struct buffer_head *bh);
|
|
int ocfs2_aio_read(struct file *file, struct kiocb *req, struct iocb *iocb);
|
|
int ocfs2_aio_write(struct file *file, struct kiocb *req, struct iocb *iocb);
|
|
struct buffer_head *ocfs2_bread(struct inode *inode,
|
|
int block, int *err, int reada);
|
|
|
|
void ocfs2_set_inode_flags(struct inode *inode);
|
|
void ocfs2_get_inode_flags(struct ocfs2_inode_info *oi);
|
|
|
|
static inline blkcnt_t ocfs2_inode_sector_count(struct inode *inode)
|
|
{
|
|
int c_to_s_bits = OCFS2_SB(inode->i_sb)->s_clustersize_bits - 9;
|
|
|
|
return (blkcnt_t)(OCFS2_I(inode)->ip_clusters << c_to_s_bits);
|
|
}
|
|
|
|
/* Validate that a bh contains a valid inode */
|
|
int ocfs2_validate_inode_block(struct super_block *sb,
|
|
struct buffer_head *bh);
|
|
/*
|
|
* Read an inode block into *bh. If *bh is NULL, a bh will be allocated.
|
|
* This is a cached read. The inode will be validated with
|
|
* ocfs2_validate_inode_block().
|
|
*/
|
|
int ocfs2_read_inode_block(struct inode *inode, struct buffer_head **bh);
|
|
/* The same, but can be passed OCFS2_BH_* flags */
|
|
int ocfs2_read_inode_block_full(struct inode *inode, struct buffer_head **bh,
|
|
int flags);
|
|
|
|
static inline struct ocfs2_inode_info *cache_info_to_inode(struct ocfs2_caching_info *ci)
|
|
{
|
|
return container_of(ci, struct ocfs2_inode_info, ip_metadata_cache);
|
|
}
|
|
|
|
#endif /* OCFS2_INODE_H */
|