7980fb6d3a
and vop_reclaim() methods. They seems to be unused, and the reported situation is normal for the forced unmount. MFC after: 1 week X-MFC-note: keep prtactive symbol in vfs_subr.c
922 lines
25 KiB
C
922 lines
25 KiB
C
/*-
|
|
* Copyright 2000 Hans Reiser
|
|
* See README for licensing and copyright details
|
|
*
|
|
* Ported to FreeBSD by Jean-Sébastien Pédron <dumbbell@FreeBSD.org>
|
|
*
|
|
* $FreeBSD$
|
|
*/
|
|
|
|
#include <gnu/fs/reiserfs/reiserfs_fs.h>
|
|
|
|
static b_strategy_t reiserfs_bufstrategy;
|
|
|
|
/*
|
|
* Buffer operations for ReiserFS vnodes.
|
|
* We punt on VOP_BMAP, so we need to do strategy on the file's vnode
|
|
* rather than the underlying device's.
|
|
*/
|
|
static struct buf_ops reiserfs_vnbufops = {
|
|
.bop_name = "ReiserFS",
|
|
.bop_strategy = reiserfs_bufstrategy,
|
|
};
|
|
|
|
/* Default io size devuned in super.c */
|
|
extern int reiserfs_default_io_size;
|
|
void inode_set_bytes(struct reiserfs_node *ip, off_t bytes);
|
|
|
|
/* Args for the create parameter of reiserfs_get_block */
|
|
#define GET_BLOCK_NO_CREATE 0 /* Don't create new blocks or convert
|
|
tails */
|
|
#define GET_BLOCK_CREATE 1 /* Add anything you need to find block */
|
|
#define GET_BLOCK_NO_HOLE 2 /* Return ENOENT for file holes */
|
|
#define GET_BLOCK_READ_DIRECT 4 /* Read the tail if indirect item not
|
|
found */
|
|
#define GET_BLOCK_NO_ISEM 8 /* i_sem is not held, don't preallocate */
|
|
#define GET_BLOCK_NO_DANGLE 16 /* Don't leave any transactions running */
|
|
|
|
/* -------------------------------------------------------------------
|
|
* vnode operations
|
|
* -------------------------------------------------------------------*/
|
|
|
|
int
|
|
reiserfs_read(struct vop_read_args *ap)
|
|
{
|
|
struct uio *uio;
|
|
struct vnode *vp;
|
|
struct reiserfs_node *ip;
|
|
struct reiserfs_sb_info *sbi;
|
|
|
|
int error;
|
|
long size;
|
|
daddr_t lbn;
|
|
off_t bytesinfile, offset;
|
|
|
|
uio = ap->a_uio;
|
|
vp = ap->a_vp;
|
|
ip = VTOI(vp);
|
|
sbi = ip->i_reiserfs;
|
|
|
|
size = sbi->s_blocksize;
|
|
|
|
for (error = 0; uio->uio_resid > 0;) {
|
|
if ((bytesinfile = ip->i_size - uio->uio_offset) <= 0)
|
|
break;
|
|
|
|
/* Compute the logical block number and its offset */
|
|
lbn = uio->uio_offset / size;
|
|
offset = uio->uio_offset % size;
|
|
reiserfs_log(LOG_DEBUG, "logical block number: %ju\n",
|
|
(intmax_t)lbn);
|
|
reiserfs_log(LOG_DEBUG, "block offset: %ju\n",
|
|
(intmax_t)offset);
|
|
|
|
/* Read file blocks */
|
|
reiserfs_log(LOG_DEBUG, "reiserfs_get_block(%ju)\n",
|
|
(intmax_t)lbn);
|
|
if ((error = reiserfs_get_block(ip, lbn, offset, uio)) != 0) {
|
|
reiserfs_log(LOG_DEBUG,
|
|
"reiserfs_get_block returned the error %d\n",
|
|
error);
|
|
break;
|
|
}
|
|
}
|
|
|
|
return (error);
|
|
}
|
|
|
|
static void
|
|
reiserfs_bufstrategy(struct bufobj *bo, struct buf *bp)
|
|
{
|
|
struct vnode *vp;
|
|
int rc;
|
|
|
|
vp = bo->bo_private;
|
|
KASSERT(bo == &vp->v_bufobj, ("BO/VP mismatch: vp %p bo %p != %p",
|
|
vp, &vp->v_bufobj, bo));
|
|
rc = VOP_STRATEGY(vp, bp);
|
|
KASSERT(rc == 0, ("ReiserFS VOP_STRATEGY failed: bp=%p, "
|
|
"vp=%p, rc=%d", bp, vp, rc));
|
|
}
|
|
|
|
int
|
|
reiserfs_inactive(struct vop_inactive_args *ap)
|
|
{
|
|
int error;
|
|
struct vnode *vp;
|
|
struct thread *td;
|
|
struct reiserfs_node *ip;
|
|
|
|
error = 0;
|
|
vp = ap->a_vp;
|
|
td = ap->a_td;
|
|
ip = VTOI(vp);
|
|
|
|
reiserfs_log(LOG_DEBUG, "deactivating inode used %d times\n",
|
|
vp->v_usecount);
|
|
|
|
#if 0
|
|
/* Ignore inodes related to stale file handles. */
|
|
if (ip->i_mode == 0)
|
|
goto out;
|
|
|
|
out:
|
|
#endif
|
|
|
|
/*
|
|
* If we are done with the inode, reclaim it so that it can be reused
|
|
* immediately.
|
|
*/
|
|
if (ip->i_mode == 0) {
|
|
reiserfs_log(LOG_DEBUG, "recyling\n");
|
|
vrecycle(vp, td);
|
|
}
|
|
|
|
return (error);
|
|
}
|
|
|
|
int
|
|
reiserfs_reclaim(struct vop_reclaim_args *ap)
|
|
{
|
|
struct reiserfs_node *ip;
|
|
struct vnode *vp;
|
|
|
|
vp = ap->a_vp;
|
|
|
|
reiserfs_log(LOG_DEBUG, "reclaiming inode used %d times\n",
|
|
vp->v_usecount);
|
|
ip = VTOI(vp);
|
|
|
|
/* XXX Update this node (write to the disk) */
|
|
|
|
/* Remove the inode from its hash chain. */
|
|
vfs_hash_remove(vp);
|
|
|
|
reiserfs_log(LOG_DEBUG, "free private data\n");
|
|
free(vp->v_data, M_REISERFSNODE);
|
|
vp->v_data = NULL;
|
|
vnode_destroy_vobject(vp);
|
|
|
|
return (0);
|
|
}
|
|
|
|
/* -------------------------------------------------------------------
|
|
* Functions from linux/fs/reiserfs/inode.c
|
|
* -------------------------------------------------------------------*/
|
|
|
|
static void
|
|
_make_cpu_key(struct cpu_key *key, int version,
|
|
uint32_t dirid, uint32_t objectid, off_t offset, int type, int length)
|
|
{
|
|
|
|
key->version = version;
|
|
|
|
key->on_disk_key.k_dir_id = dirid;
|
|
key->on_disk_key.k_objectid = objectid;
|
|
set_cpu_key_k_offset(key, offset);
|
|
set_cpu_key_k_type(key, type);
|
|
key->key_length = length;
|
|
}
|
|
|
|
/*
|
|
* Take base of inode_key (it comes from inode always) (dirid, objectid)
|
|
* and version from an inode, set offset and type of key
|
|
*/
|
|
void
|
|
make_cpu_key(struct cpu_key *key, struct reiserfs_node *ip, off_t offset,
|
|
int type, int length)
|
|
{
|
|
|
|
_make_cpu_key(key, get_inode_item_key_version(ip),
|
|
le32toh(INODE_PKEY(ip)->k_dir_id),
|
|
le32toh(INODE_PKEY(ip)->k_objectid),
|
|
offset, type, length);
|
|
}
|
|
|
|
int
|
|
reiserfs_get_block(struct reiserfs_node *ip, long block, off_t offset,
|
|
struct uio *uio)
|
|
{
|
|
caddr_t blk = NULL, p;
|
|
struct cpu_key key;
|
|
/* unsigned long offset; */
|
|
INITIALIZE_PATH(path);
|
|
struct buf *bp, *blk_bp;
|
|
struct item_head *ih;
|
|
struct reiserfs_sb_info *sbi;
|
|
int blocknr, chars, done = 0, ret = 0, args = 0;
|
|
|
|
sbi = ip->i_reiserfs;
|
|
|
|
/* Prepare the key to look for the 'block'-th block of file */
|
|
reiserfs_log(LOG_DEBUG, "prepare cpu key\n");
|
|
make_cpu_key(&key, ip, (off_t)block * sbi->s_blocksize + 1, TYPE_ANY, 3);
|
|
|
|
/* research: */
|
|
reiserfs_log(LOG_DEBUG, "search for position\n");
|
|
if (search_for_position_by_key(sbi, &key, &path) != POSITION_FOUND) {
|
|
reiserfs_log(LOG_DEBUG, "position not found\n");
|
|
pathrelse(&path);
|
|
#if 0
|
|
if (blk)
|
|
kunmap(bh_result->b_page);
|
|
#endif
|
|
/*
|
|
* We do not return ENOENT if there is a hole but page is
|
|
* uptodate, because it means that there is some MMAPED data
|
|
* associated with it that is yet to be written to disk.
|
|
*/
|
|
if ((args & GET_BLOCK_NO_HOLE)/* &&
|
|
!PageUptodate(bh_result->b_page)*/)
|
|
return (ENOENT);
|
|
return (0);
|
|
}
|
|
reiserfs_log(LOG_DEBUG, "position found\n");
|
|
|
|
bp = get_last_bp(&path);
|
|
ih = get_ih(&path);
|
|
|
|
if (is_indirect_le_ih(ih)) {
|
|
off_t xfersize;
|
|
uint32_t *ind_item = (uint32_t *)B_I_PITEM(bp, ih);
|
|
|
|
reiserfs_log(LOG_DEBUG, "item is INDIRECT\n");
|
|
|
|
blocknr = get_block_num(ind_item, path.pos_in_item);
|
|
reiserfs_log(LOG_DEBUG, "block number: %d "
|
|
"(ind_item=%p, pos_in_item=%u)\n",
|
|
blocknr, ind_item, path.pos_in_item);
|
|
|
|
xfersize = MIN(sbi->s_blocksize - offset,
|
|
ip->i_size - uio->uio_offset);
|
|
xfersize = MIN(xfersize, uio->uio_resid);
|
|
|
|
if (blocknr) {
|
|
ret = bread(sbi->s_devvp,
|
|
blocknr * btodb(sbi->s_blocksize),
|
|
sbi->s_blocksize, NOCRED, &blk_bp);
|
|
reiserfs_log(LOG_DEBUG, "xfersize: %ju\n",
|
|
(intmax_t)xfersize);
|
|
ret = uiomove(blk_bp->b_data + offset, xfersize, uio);
|
|
brelse(blk_bp);
|
|
} else {
|
|
/*
|
|
* We do not return ENOENT if there is a hole but
|
|
* page is uptodate, because it means That there
|
|
* is some MMAPED data associated with it that
|
|
* is yet to be written to disk.
|
|
*/
|
|
if ((args & GET_BLOCK_NO_HOLE)/* &&
|
|
!PageUptodate(bh_result->b_page)*/)
|
|
ret = (ENOENT);
|
|
|
|
/* Skip this hole */
|
|
uio->uio_resid -= xfersize;
|
|
uio->uio_offset += xfersize;
|
|
}
|
|
|
|
pathrelse(&path);
|
|
return (ret);
|
|
}
|
|
|
|
reiserfs_log(LOG_DEBUG, "item should be DIRECT\n");
|
|
|
|
#if 0
|
|
/* Requested data are in direct item(s) */
|
|
if (!(args & GET_BLOCK_READ_DIRECT)) {
|
|
/*
|
|
* We are called by bmap. FIXME: we can not map block of
|
|
* file when it is stored in direct item(s)
|
|
*/
|
|
pathrelse(&path);
|
|
#if 0
|
|
if (blk)
|
|
kunmap(bh_result->b_page);
|
|
#endif
|
|
return (ENOENT);
|
|
}
|
|
#endif
|
|
|
|
#if 0
|
|
/*
|
|
* If we've got a direct item, and the buffer or page was uptodate, we
|
|
* don't want to pull data off disk again. Skip to the end, where we
|
|
* map the buffer and return
|
|
*/
|
|
if (buffer_uptodate(bh_result)) {
|
|
goto finished;
|
|
} else
|
|
/*
|
|
* grab_tail_page can trigger calls to reiserfs_get_block
|
|
* on up to date pages without any buffers. If the page
|
|
* is up to date, we don't want read old data off disk.
|
|
* Set the up to date bit on the buffer instead and jump
|
|
* to the end
|
|
*/
|
|
if (!bh_result->b_page || PageUptodate(bh_result->b_page)) {
|
|
set_buffer_uptodate(bh_result);
|
|
goto finished;
|
|
}
|
|
#endif
|
|
|
|
#if 0
|
|
/* Read file tail into part of page */
|
|
offset = (cpu_key_k_offset(&key) - 1) & (PAGE_CACHE_SIZE - 1);
|
|
fs_gen = get_generation(ip->i_reiserfs);
|
|
copy_item_head(&tmp_ih, ih);
|
|
#endif
|
|
|
|
#if 0
|
|
/*
|
|
* We only want to kmap if we are reading the tail into the page. this
|
|
* is not the common case, so we don't kmap until we are sure we need
|
|
* to. But, this means the item might move if kmap schedules
|
|
*/
|
|
if (!blk) {
|
|
blk = (char *)kmap(bh_result->b_page);
|
|
if (fs_changed (fs_gen, sbi) && item_moved(&tmp_ih, &path))
|
|
goto research;
|
|
}
|
|
blk += offset;
|
|
memset(blk, 0, sbi->s_blocksize);
|
|
#endif
|
|
if (!blk) {
|
|
reiserfs_log(LOG_DEBUG, "allocating buffer\n");
|
|
blk = malloc(ip->i_size, M_REISERFSNODE, M_WAITOK | M_ZERO);
|
|
if (!blk)
|
|
return (ENOMEM);
|
|
}
|
|
/* p += offset; */
|
|
|
|
p = blk;
|
|
do {
|
|
if (!is_direct_le_ih(ih)) {
|
|
reiserfs_log(LOG_ERR, "BUG\n");
|
|
return (ENOENT); /* XXX Wrong error code */
|
|
}
|
|
|
|
/*
|
|
* Make sure we don't read more bytes than actually exist
|
|
* in the file. This can happen in odd cases where i_size
|
|
* isn't correct, and when direct item padding results in
|
|
* a few extra bytes at the end of the direct item
|
|
*/
|
|
if ((le_ih_k_offset(ih) + path.pos_in_item) > ip->i_size)
|
|
break;
|
|
|
|
if ((le_ih_k_offset(ih) - 1 + ih_item_len(ih)) > ip->i_size) {
|
|
chars = ip->i_size - (le_ih_k_offset(ih) - 1) -
|
|
path.pos_in_item;
|
|
done = 1;
|
|
} else {
|
|
chars = ih_item_len(ih) - path.pos_in_item;
|
|
}
|
|
reiserfs_log(LOG_DEBUG, "copying %d bytes\n", chars);
|
|
memcpy(p, B_I_PITEM(bp, ih) + path.pos_in_item, chars);
|
|
if (done) {
|
|
reiserfs_log(LOG_DEBUG, "copy done\n");
|
|
break;
|
|
}
|
|
|
|
p += chars;
|
|
|
|
if (PATH_LAST_POSITION(&path) != (B_NR_ITEMS(bp) - 1))
|
|
/*
|
|
* We done, if read direct item is not the last
|
|
* item of node
|
|
* FIXME: we could try to check right delimiting
|
|
* key to see whether direct item continues in
|
|
* the right neighbor or rely on i_size
|
|
*/
|
|
break;
|
|
|
|
/* Update key to look for the next piece */
|
|
set_cpu_key_k_offset(&key, cpu_key_k_offset(&key) + chars);
|
|
if (search_for_position_by_key(sbi, &key, &path) !=
|
|
POSITION_FOUND)
|
|
/*
|
|
* We read something from tail, even if now we got
|
|
* IO_ERROR
|
|
*/
|
|
break;
|
|
|
|
bp = get_last_bp(&path);
|
|
ih = get_ih(&path);
|
|
} while (1);
|
|
|
|
/* finished: */
|
|
pathrelse(&path);
|
|
/*
|
|
* This buffer has valid data, but isn't valid for io. mapping it to
|
|
* block #0 tells the rest of reiserfs it just has a tail in it
|
|
*/
|
|
ret = uiomove(blk, ip->i_size, uio);
|
|
free(blk, M_REISERFSNODE);
|
|
return (ret);
|
|
}
|
|
|
|
/*
|
|
* Compute real number of used bytes by file
|
|
* Following three functions can go away when we'll have enough space in
|
|
* stat item
|
|
*/
|
|
static int
|
|
real_space_diff(struct reiserfs_node *ip, int sd_size)
|
|
{
|
|
int bytes;
|
|
off_t blocksize = ip->i_reiserfs->s_blocksize;
|
|
|
|
if (S_ISLNK(ip->i_mode) || S_ISDIR(ip->i_mode))
|
|
return (sd_size);
|
|
|
|
/* End of file is also in full block with indirect reference, so round
|
|
* up to the next block.
|
|
*
|
|
* There is just no way to know if the tail is actually packed on the
|
|
* file, so we have to assume it isn't. When we pack the tail, we add
|
|
* 4 bytes to pretend there really is an unformatted node pointer. */
|
|
bytes = ((ip->i_size + (blocksize - 1)) >>
|
|
ip->i_reiserfs->s_blocksize_bits) * UNFM_P_SIZE + sd_size;
|
|
|
|
return (bytes);
|
|
}
|
|
|
|
static inline off_t
|
|
to_real_used_space(struct reiserfs_node *ip, unsigned long blocks, int sd_size)
|
|
{
|
|
|
|
if (S_ISLNK(ip->i_mode) || S_ISDIR(ip->i_mode)) {
|
|
return ip->i_size + (off_t)(real_space_diff(ip, sd_size));
|
|
}
|
|
|
|
return ((off_t)real_space_diff(ip, sd_size)) + (((off_t)blocks) << 9);
|
|
}
|
|
|
|
void
|
|
inode_set_bytes(struct reiserfs_node *ip, off_t bytes)
|
|
{
|
|
|
|
ip->i_blocks = bytes >> 9;
|
|
ip->i_bytes = bytes & 511;
|
|
}
|
|
|
|
/* Called by read_locked_inode */
|
|
static void
|
|
init_inode(struct reiserfs_node *ip, struct path *path)
|
|
{
|
|
struct buf *bp;
|
|
struct item_head *ih;
|
|
uint32_t rdev;
|
|
|
|
bp = PATH_PLAST_BUFFER(path);
|
|
ih = PATH_PITEM_HEAD(path);
|
|
|
|
reiserfs_log(LOG_DEBUG, "copy the key (objectid=%d, dirid=%d)\n",
|
|
ih->ih_key.k_objectid, ih->ih_key.k_dir_id);
|
|
copy_key(INODE_PKEY(ip), &(ih->ih_key));
|
|
/* ip->i_blksize = reiserfs_default_io_size; */
|
|
|
|
reiserfs_log(LOG_DEBUG, "reset some inode structure members\n");
|
|
REISERFS_I(ip)->i_flags = 0;
|
|
#if 0
|
|
REISERFS_I(ip)->i_prealloc_block = 0;
|
|
REISERFS_I(ip)->i_prealloc_count = 0;
|
|
REISERFS_I(ip)->i_trans_id = 0;
|
|
REISERFS_I(ip)->i_jl = NULL;
|
|
REISERFS_I(ip)->i_acl_access = NULL;
|
|
REISERFS_I(ip)->i_acl_default = NULL;
|
|
#endif
|
|
|
|
if (stat_data_v1(ih)) {
|
|
reiserfs_log(LOG_DEBUG, "reiserfs/init_inode: stat data v1\n");
|
|
struct stat_data_v1 *sd;
|
|
unsigned long blocks;
|
|
|
|
sd = (struct stat_data_v1 *)B_I_PITEM(bp, ih);
|
|
|
|
reiserfs_log(LOG_DEBUG,
|
|
"reiserfs/init_inode: filling more members\n");
|
|
set_inode_item_key_version(ip, KEY_FORMAT_3_5);
|
|
set_inode_sd_version(ip, STAT_DATA_V1);
|
|
ip->i_mode = sd_v1_mode(sd);
|
|
ip->i_nlink = sd_v1_nlink(sd);
|
|
ip->i_uid = sd_v1_uid(sd);
|
|
ip->i_gid = sd_v1_gid(sd);
|
|
ip->i_size = sd_v1_size(sd);
|
|
ip->i_atime.tv_sec = sd_v1_atime(sd);
|
|
ip->i_mtime.tv_sec = sd_v1_mtime(sd);
|
|
ip->i_ctime.tv_sec = sd_v1_ctime(sd);
|
|
ip->i_atime.tv_nsec = 0;
|
|
ip->i_ctime.tv_nsec = 0;
|
|
ip->i_mtime.tv_nsec = 0;
|
|
|
|
reiserfs_log(LOG_DEBUG, " mode = %08x\n", ip->i_mode);
|
|
reiserfs_log(LOG_DEBUG, " nlink = %d\n", ip->i_nlink);
|
|
reiserfs_log(LOG_DEBUG, " owner = %d:%d\n", ip->i_uid,
|
|
ip->i_gid);
|
|
reiserfs_log(LOG_DEBUG, " size = %ju\n",
|
|
(intmax_t)ip->i_size);
|
|
reiserfs_log(LOG_DEBUG, " atime = %jd\n",
|
|
(intmax_t)ip->i_atime.tv_sec);
|
|
reiserfs_log(LOG_DEBUG, " mtime = %jd\n",
|
|
(intmax_t)ip->i_mtime.tv_sec);
|
|
reiserfs_log(LOG_DEBUG, " ctime = %jd\n",
|
|
(intmax_t)ip->i_ctime.tv_sec);
|
|
|
|
ip->i_blocks = sd_v1_blocks(sd);
|
|
ip->i_generation = le32toh(INODE_PKEY(ip)->k_dir_id);
|
|
blocks = (ip->i_size + 511) >> 9;
|
|
blocks = _ROUND_UP(blocks, ip->i_reiserfs->s_blocksize >> 9);
|
|
if (ip->i_blocks > blocks) {
|
|
/*
|
|
* There was a bug in <= 3.5.23 when i_blocks could
|
|
* take negative values. Starting from 3.5.17 this
|
|
* value could even be stored in stat data. For such
|
|
* files we set i_blocks based on file size. Just 2
|
|
* notes: this can be wrong for sparce files. On-disk
|
|
* value will be only updated if file's inode will
|
|
* ever change.
|
|
*/
|
|
ip->i_blocks = blocks;
|
|
}
|
|
|
|
rdev = sd_v1_rdev(sd);
|
|
REISERFS_I(ip)->i_first_direct_byte =
|
|
sd_v1_first_direct_byte(sd);
|
|
|
|
/*
|
|
* An early bug in the quota code can give us an odd number
|
|
* for the block count. This is incorrect, fix it here.
|
|
*/
|
|
if (ip->i_blocks & 1) {
|
|
ip->i_blocks++ ;
|
|
}
|
|
inode_set_bytes(ip, to_real_used_space(ip, ip->i_blocks,
|
|
SD_V1_SIZE));
|
|
|
|
/*
|
|
* nopack is initially zero for v1 objects. For v2 objects,
|
|
* nopack is initialised from sd_attrs
|
|
*/
|
|
REISERFS_I(ip)->i_flags &= ~i_nopack_mask;
|
|
reiserfs_log(LOG_DEBUG, "...done\n");
|
|
} else {
|
|
reiserfs_log(LOG_DEBUG, "stat data v2\n");
|
|
/*
|
|
* New stat data found, but object may have old items
|
|
* (directories and symlinks)
|
|
*/
|
|
struct stat_data *sd = (struct stat_data *)B_I_PITEM(bp, ih);
|
|
|
|
reiserfs_log(LOG_DEBUG, "filling more members\n");
|
|
ip->i_mode = sd_v2_mode(sd);
|
|
ip->i_nlink = sd_v2_nlink(sd);
|
|
ip->i_uid = sd_v2_uid(sd);
|
|
ip->i_size = sd_v2_size(sd);
|
|
ip->i_gid = sd_v2_gid(sd);
|
|
ip->i_mtime.tv_sec = sd_v2_mtime(sd);
|
|
ip->i_atime.tv_sec = sd_v2_atime(sd);
|
|
ip->i_ctime.tv_sec = sd_v2_ctime(sd);
|
|
ip->i_ctime.tv_nsec = 0;
|
|
ip->i_mtime.tv_nsec = 0;
|
|
ip->i_atime.tv_nsec = 0;
|
|
|
|
reiserfs_log(LOG_DEBUG, " mode = %08x\n", ip->i_mode);
|
|
reiserfs_log(LOG_DEBUG, " nlink = %d\n", ip->i_nlink);
|
|
reiserfs_log(LOG_DEBUG, " owner = %d:%d\n", ip->i_uid,
|
|
ip->i_gid);
|
|
reiserfs_log(LOG_DEBUG, " size = %ju\n",
|
|
(intmax_t)ip->i_size);
|
|
reiserfs_log(LOG_DEBUG, " atime = %jd\n",
|
|
(intmax_t)ip->i_atime.tv_sec);
|
|
reiserfs_log(LOG_DEBUG, " mtime = %jd\n",
|
|
(intmax_t)ip->i_mtime.tv_sec);
|
|
reiserfs_log(LOG_DEBUG, " ctime = %jd\n",
|
|
(intmax_t)ip->i_ctime.tv_sec);
|
|
|
|
ip->i_blocks = sd_v2_blocks(sd);
|
|
rdev = sd_v2_rdev(sd);
|
|
reiserfs_log(LOG_DEBUG, " blocks = %u\n", ip->i_blocks);
|
|
|
|
if (S_ISCHR(ip->i_mode) || S_ISBLK(ip->i_mode))
|
|
ip->i_generation = le32toh(INODE_PKEY(ip)->k_dir_id);
|
|
else
|
|
ip->i_generation = sd_v2_generation(sd);
|
|
|
|
if (S_ISDIR(ip->i_mode) || S_ISLNK(ip->i_mode))
|
|
set_inode_item_key_version(ip, KEY_FORMAT_3_5);
|
|
else
|
|
set_inode_item_key_version(ip, KEY_FORMAT_3_6);
|
|
|
|
REISERFS_I(ip)->i_first_direct_byte = 0;
|
|
set_inode_sd_version(ip, STAT_DATA_V2);
|
|
inode_set_bytes(ip, to_real_used_space(ip, ip->i_blocks,
|
|
SD_V2_SIZE));
|
|
|
|
/*
|
|
* Read persistent inode attributes from sd and initalise
|
|
* generic inode flags from them
|
|
*/
|
|
REISERFS_I(ip)->i_attrs = sd_v2_attrs(sd);
|
|
sd_attrs_to_i_attrs(sd_v2_attrs(sd), ip);
|
|
reiserfs_log(LOG_DEBUG, "...done\n");
|
|
}
|
|
|
|
pathrelse(path);
|
|
if (S_ISREG(ip->i_mode)) {
|
|
reiserfs_log(LOG_DEBUG, "this inode is a regular file\n");
|
|
//ip->i_op = &reiserfs_file_ip_operations;
|
|
//ip->i_fop = &reiserfs_file_operations;
|
|
//ip->i_mapping->a_ops = &reiserfs_address_space_operations ;
|
|
} else if (S_ISDIR(ip->i_mode)) {
|
|
reiserfs_log(LOG_DEBUG, "this inode is a directory\n");
|
|
//ip->i_op = &reiserfs_dir_ip_operations;
|
|
//ip->i_fop = &reiserfs_dir_operations;
|
|
} else if (S_ISLNK(ip->i_mode)) {
|
|
reiserfs_log(LOG_DEBUG, "this inode is a symlink\n");
|
|
//ip->i_op = &reiserfs_symlink_ip_operations;
|
|
//ip->i_mapping->a_ops = &reiserfs_address_space_operations;
|
|
} else {
|
|
reiserfs_log(LOG_DEBUG, "this inode is something unknown in "
|
|
"this universe\n");
|
|
ip->i_blocks = 0;
|
|
//ip->i_op = &reiserfs_special_ip_operations;
|
|
//init_special_ip(ip, ip->i_mode, new_decode_dev(rdev));
|
|
}
|
|
}
|
|
|
|
/*
|
|
* reiserfs_read_locked_inode is called to read the inode off disk, and
|
|
* it does a make_bad_inode when things go wrong. But, we need to make
|
|
* sure and clear the key in the private portion of the inode, otherwise
|
|
* a corresponding iput might try to delete whatever object the inode
|
|
* last represented.
|
|
*/
|
|
static void
|
|
reiserfs_make_bad_inode(struct reiserfs_node *ip) {
|
|
|
|
memset(INODE_PKEY(ip), 0, KEY_SIZE);
|
|
//make_bad_inode(inode);
|
|
}
|
|
|
|
void
|
|
reiserfs_read_locked_inode(struct reiserfs_node *ip,
|
|
struct reiserfs_iget_args *args)
|
|
{
|
|
INITIALIZE_PATH(path_to_sd);
|
|
struct cpu_key key;
|
|
unsigned long dirino;
|
|
int retval;
|
|
|
|
dirino = args->dirid;
|
|
|
|
/*
|
|
* Set version 1, version 2 could be used too, because stat data
|
|
* key is the same in both versions
|
|
*/
|
|
key.version = KEY_FORMAT_3_5;
|
|
key.on_disk_key.k_dir_id = dirino;
|
|
key.on_disk_key.k_objectid = ip->i_number;
|
|
key.on_disk_key.u.k_offset_v1.k_offset = SD_OFFSET;
|
|
key.on_disk_key.u.k_offset_v1.k_uniqueness = SD_UNIQUENESS;
|
|
|
|
/* Look for the object's stat data */
|
|
retval = search_item(ip->i_reiserfs, &key, &path_to_sd);
|
|
if (retval == IO_ERROR) {
|
|
reiserfs_log(LOG_ERR,
|
|
"I/O failure occured trying to find stat"
|
|
"data %u/%u\n",
|
|
key.on_disk_key.k_dir_id, key.on_disk_key.k_objectid);
|
|
reiserfs_make_bad_inode(ip);
|
|
return;
|
|
}
|
|
if (retval != ITEM_FOUND) {
|
|
/*
|
|
* A stale NFS handle can trigger this without it being
|
|
* an error
|
|
*/
|
|
reiserfs_log(LOG_ERR,
|
|
"item not found (objectid=%u, dirid=%u)\n",
|
|
key.on_disk_key.k_objectid, key.on_disk_key.k_dir_id);
|
|
pathrelse(&path_to_sd);
|
|
reiserfs_make_bad_inode(ip);
|
|
ip->i_nlink = 0;
|
|
return;
|
|
}
|
|
|
|
init_inode(ip, &path_to_sd);
|
|
|
|
/*
|
|
* It is possible that knfsd is trying to access inode of a file
|
|
* that is being removed from the disk by some other thread. As
|
|
* we update sd on unlink all that is required is to check for
|
|
* nlink here. This bug was first found by Sizif when debugging
|
|
* SquidNG/Butterfly, forgotten, and found again after Philippe
|
|
* Gramoulle <philippe.gramoulle@mmania.com> reproduced it.
|
|
*
|
|
* More logical fix would require changes in fs/inode.c:iput() to
|
|
* remove inode from hash-table _after_ fs cleaned disk stuff up and
|
|
* in iget() to return NULL if I_FREEING inode is found in hash-table.
|
|
*/
|
|
/*
|
|
* Currently there is one place where it's ok to meet inode with
|
|
* nlink == 0: processing of open-unlinked and half-truncated files
|
|
* during mount (fs/reiserfs/super.c:finish_unfinished()).
|
|
*/
|
|
if((ip->i_nlink == 0) &&
|
|
!REISERFS_SB(ip->i_reiserfs)->s_is_unlinked_ok ) {
|
|
reiserfs_log(LOG_WARNING, "dead inode read from disk. This is "
|
|
"likely to be race with knfsd. Ignore");
|
|
reiserfs_make_bad_inode(ip);
|
|
}
|
|
|
|
/* Init inode should be relsing */
|
|
reiserfs_check_path(&path_to_sd);
|
|
}
|
|
|
|
int
|
|
reiserfs_iget(
|
|
struct mount *mp, const struct cpu_key *key,
|
|
struct vnode **vpp, struct thread *td)
|
|
{
|
|
int error, flags;
|
|
struct cdev *dev;
|
|
struct vnode *vp;
|
|
struct reiserfs_node *ip;
|
|
struct reiserfs_mount *rmp;
|
|
|
|
struct reiserfs_iget_args args;
|
|
|
|
//restart:
|
|
/* Check if the inode cache contains it */
|
|
// XXX LK_EXCLUSIVE ?
|
|
flags = LK_EXCLUSIVE;
|
|
error = vfs_hash_get(mp, key->on_disk_key.k_objectid, flags,
|
|
td, vpp, NULL, NULL);
|
|
if (error || *vpp != NULL)
|
|
return (error);
|
|
|
|
rmp = VFSTOREISERFS(mp);
|
|
dev = rmp->rm_dev;
|
|
|
|
/*
|
|
* If this malloc() is performed after the getnewvnode() it might
|
|
* block, leaving a vnode with a NULL v_data to be found by
|
|
* reiserfs_sync() if a sync happens to fire right then, which
|
|
* will cause a panic because reiserfs_sync() blindly dereferences
|
|
* vp->v_data (as well it should).
|
|
*/
|
|
reiserfs_log(LOG_DEBUG, "malloc(struct reiserfs_node)\n");
|
|
ip = malloc(sizeof(struct reiserfs_node), M_REISERFSNODE,
|
|
M_WAITOK | M_ZERO);
|
|
|
|
/* Allocate a new vnode/inode. */
|
|
reiserfs_log(LOG_DEBUG, "getnewvnode\n");
|
|
if ((error =
|
|
getnewvnode("reiserfs", mp, &reiserfs_vnodeops, &vp)) != 0) {
|
|
*vpp = NULL;
|
|
free(ip, M_REISERFSNODE);
|
|
reiserfs_log(LOG_DEBUG, "getnewvnode FAILED\n");
|
|
return (error);
|
|
}
|
|
|
|
args.dirid = key->on_disk_key.k_dir_id;
|
|
args.objectid = key->on_disk_key.k_objectid;
|
|
|
|
reiserfs_log(LOG_DEBUG, "filling *ip\n");
|
|
vp->v_data = ip;
|
|
ip->i_vnode = vp;
|
|
ip->i_dev = dev;
|
|
ip->i_number = args.objectid;
|
|
ip->i_ino = args.dirid;
|
|
ip->i_reiserfs = rmp->rm_reiserfs;
|
|
|
|
vp->v_bufobj.bo_ops = &reiserfs_vnbufops;
|
|
vp->v_bufobj.bo_private = vp;
|
|
|
|
/* If this is the root node, set the VV_ROOT flag */
|
|
if (ip->i_number == REISERFS_ROOT_OBJECTID &&
|
|
ip->i_ino == REISERFS_ROOT_PARENT_OBJECTID)
|
|
vp->v_vflag |= VV_ROOT;
|
|
|
|
#if 0
|
|
if (VOP_LOCK(vp, LK_EXCLUSIVE) != 0)
|
|
panic("reiserfs/iget: unexpected lock failure");
|
|
|
|
/*
|
|
* Exclusively lock the vnode before adding to hash. Note, that we
|
|
* must not release nor downgrade the lock (despite flags argument
|
|
* says) till it is fully initialized.
|
|
*/
|
|
lockmgr(vp->v_vnlock, LK_EXCLUSIVE, (struct mtx *)0);
|
|
#endif
|
|
|
|
lockmgr(vp->v_vnlock, LK_EXCLUSIVE, NULL);
|
|
error = insmntque(vp, mp);
|
|
if (error != 0) {
|
|
free(ip, M_REISERFSNODE);
|
|
*vpp = NULL;
|
|
reiserfs_log(LOG_DEBUG, "insmntque FAILED\n");
|
|
return (error);
|
|
}
|
|
error = vfs_hash_insert(vp, key->on_disk_key.k_objectid, flags,
|
|
td, vpp, NULL, NULL);
|
|
if (error || *vpp != NULL)
|
|
return (error);
|
|
|
|
/* Read the inode */
|
|
reiserfs_log(LOG_DEBUG, "call reiserfs_read_locked_inode ("
|
|
"objectid=%d,dirid=%d)\n", args.objectid, args.dirid);
|
|
reiserfs_read_locked_inode(ip, &args);
|
|
|
|
ip->i_devvp = rmp->rm_devvp;
|
|
|
|
switch(vp->v_type = IFTOVT(ip->i_mode)) {
|
|
case VBLK:
|
|
reiserfs_log(LOG_DEBUG, "vnode type VBLK\n");
|
|
vp->v_op = &reiserfs_specops;
|
|
break;
|
|
#if 0
|
|
case VCHR:
|
|
reiserfs_log(LOG_DEBUG, "vnode type VCHR\n");
|
|
vp->v_op = &reiserfs_specops;
|
|
vp = addaliasu(vp, ip->i_rdev);
|
|
ip->i_vnode = vp;
|
|
break;
|
|
case VFIFO:
|
|
reiserfs_log(LOG_DEBUG, "vnode type VFIFO\n");
|
|
vp->v_op = reiserfs_fifoop_p;
|
|
break;
|
|
#endif
|
|
default:
|
|
break;
|
|
}
|
|
|
|
*vpp = vp;
|
|
return (0);
|
|
}
|
|
|
|
void
|
|
sd_attrs_to_i_attrs(uint16_t sd_attrs, struct reiserfs_node *ip)
|
|
{
|
|
|
|
if (reiserfs_attrs(ip->i_reiserfs)) {
|
|
#if 0
|
|
if (sd_attrs & REISERFS_SYNC_FL)
|
|
ip->i_flags |= S_SYNC;
|
|
else
|
|
ip->i_flags &= ~S_SYNC;
|
|
#endif
|
|
if (sd_attrs & REISERFS_IMMUTABLE_FL)
|
|
ip->i_flags |= IMMUTABLE;
|
|
else
|
|
ip->i_flags &= ~IMMUTABLE;
|
|
if (sd_attrs & REISERFS_APPEND_FL)
|
|
ip->i_flags |= APPEND;
|
|
else
|
|
ip->i_flags &= ~APPEND;
|
|
#if 0
|
|
if (sd_attrs & REISERFS_NOATIME_FL)
|
|
ip->i_flags |= S_NOATIME;
|
|
else
|
|
ip->i_flags &= ~S_NOATIME;
|
|
if (sd_attrs & REISERFS_NOTAIL_FL)
|
|
REISERFS_I(ip)->i_flags |= i_nopack_mask;
|
|
else
|
|
REISERFS_I(ip)->i_flags &= ~i_nopack_mask;
|
|
#endif
|
|
}
|
|
}
|
|
|
|
void
|
|
i_attrs_to_sd_attrs(struct reiserfs_node *ip, uint16_t *sd_attrs)
|
|
{
|
|
|
|
if (reiserfs_attrs(ip->i_reiserfs)) {
|
|
#if 0
|
|
if (ip->i_flags & S_SYNC)
|
|
*sd_attrs |= REISERFS_SYNC_FL;
|
|
else
|
|
*sd_attrs &= ~REISERFS_SYNC_FL;
|
|
#endif
|
|
if (ip->i_flags & IMMUTABLE)
|
|
*sd_attrs |= REISERFS_IMMUTABLE_FL;
|
|
else
|
|
*sd_attrs &= ~REISERFS_IMMUTABLE_FL;
|
|
if (ip->i_flags & APPEND)
|
|
*sd_attrs |= REISERFS_APPEND_FL;
|
|
else
|
|
*sd_attrs &= ~REISERFS_APPEND_FL;
|
|
#if 0
|
|
if (ip->i_flags & S_NOATIME)
|
|
*sd_attrs |= REISERFS_NOATIME_FL;
|
|
else
|
|
*sd_attrs &= ~REISERFS_NOATIME_FL;
|
|
if (REISERFS_I(ip)->i_flags & i_nopack_mask)
|
|
*sd_attrs |= REISERFS_NOTAIL_FL;
|
|
else
|
|
*sd_attrs &= ~REISERFS_NOTAIL_FL;
|
|
#endif
|
|
}
|
|
}
|