/*
* get next entry in a directory.
*/
struct direct *
dreaddir(struct dirstuff *dirp)
{
struct direct *dp;
daddr_t lbn, d;
for (;;) {
if (dirp->loc >= (int)dirp->ip->di_size)
return (NULL);
if (blkoff(&sblock, dirp->loc) == 0) {
lbn = lblkno(&sblock, dirp->loc);
d = bmap(lbn);
if (d == 0)
return (NULL);
bread(fsbtodb(&sblock, d), dirp->dbuf,
(int)dblksize(&sblock, dirp->ip, (int)lbn));
}
dp = (struct direct *)
(dirp->dbuf + blkoff(&sblock, dirp->loc));
dirp->loc += dp->d_reclen;
if (dp->d_ino == 0) {
continue;
}
return (dp);
}
}
int
ext2fs_blkatoff(struct vnode *vp, off_t offset, char **res,
#undef struct
struct buf **bpp)
{
struct inode *ip;
struct m_ext2fs *fs;
struct buf *bp;
daddr_t lbn;
int error;
ip = VTOI(vp);
fs = ip->i_e2fs;
lbn = lblkno(fs, offset);
*bpp = NULL;
if ((error = bread(vp, lbn, fs->e2fs_bsize, NOCRED, 0, &bp)) != 0) {
brelse(bp, 0);
return (error);
}
if (res)
*res = (char *)bp->b_data + blkoff(fs, offset);
*bpp = bp;
return (0);
}
int
readit(char *buffer, int count)
{
int logno, off, size;
int cnt2, bnum2;
struct fs *fs_copy;
int n = 0;
if (poff + count > inode.i_size)
count = inode.i_size - poff;
while (count > 0 && poff < inode.i_size) {
fs_copy = fs;
off = blkoff(fs_copy, poff);
logno = lblkno(fs_copy, poff);
cnt2 = size = blksize(fs_copy, &inode, logno);
bnum2 = fsbtodb(fs_copy, block_map(logno)) + boff;
if ( (!off) && (size <= count)) {
devread(buffer, bnum2, cnt2);
} else {
size -= off;
if (size > count)
size = count;
devread(iobuf, bnum2, cnt2);
bcopy(iobuf+off, buffer, size);
}
buffer += size;
count -= size;
poff += size;
n += size;
}
return n;
}
/*
* Return buffer with the contents of block "offset" from the beginning of
* vnode "vp". If "res" is non-zero, fill it in with a pointer to the
* remaining space in the vnode.
*/
int
ffs_blkatoff(struct vnode *vp, off_t uoffset, char **res, struct buf **bpp)
{
struct inode *ip;
struct fs *fs;
struct buf *bp;
ufs_daddr_t lbn;
int bsize, error;
ip = VTOI(vp);
fs = ip->i_fs;
lbn = lblkno(fs, uoffset);
bsize = blksize(fs, ip, lbn);
*bpp = NULL;
error = bread(vp, lblktodoff(fs, lbn), bsize, &bp);
if (error) {
brelse(bp);
return (error);
}
if (res)
*res = (char *)bp->b_data + blkoff(fs, uoffset);
*bpp = bp;
return (0);
}
static int
ioread(void *addr, int count, int phys)
{
int logno, off, size;
while (count) {
off = blkoff(fs, poff);
logno = lblkno(fs, poff);
cnt = size = blksize(fs, &inode, logno);
bnum = block_map(logno);
if (bnum == -1)
return(1);
bnum = fsbtodb(fs, bnum) + boff;
size -= off;
if (size > count)
size = count;
if (disk_read(bnum, cnt, (vm_offset_t)iobuf))
return(1);
if (phys)
pcpy(iobuf+off,addr,size);
else
bcopy(iobuf+off,addr,size);
addr = (char *)addr + size;
count -= size;
poff += size;
}
return(0);
}
/*
* Return buffer with the contents of block "offset" from the beginning of
* directory "ip". If "res" is non-zero, fill it in with a pointer to the
* remaining space in the directory.
*/
int
cd9660_blkatoff(vnode_t vp, off_t offset, char **res, buf_t *bpp)
{
struct iso_node *ip;
register struct iso_mnt *imp;
buf_t bp;
daddr_t lbn;
int bsize, error;
ip = VTOI(vp);
imp = ip->i_mnt;
lbn = lblkno(imp, offset);
bsize = blksize(imp, ip, lbn);
if ((bsize != imp->im_sector_size) &&
(offset & (imp->im_sector_size - 1)) == 0) {
bsize = imp->im_sector_size;
}
if ( (error = (int)buf_bread(vp, (daddr64_t)((unsigned)lbn), bsize, NOCRED, &bp)) ) {
buf_brelse(bp);
*bpp = NULL;
return (error);
}
if (res)
*res = (char *)0 + buf_dataptr(bp) + blkoff(imp, offset);
*bpp = bp;
return (0);
}
/* ARGSUSED */
static int
ext2_read(struct vop_read_args *ap)
{
struct vnode *vp;
struct inode *ip;
struct uio *uio;
FS *fs;
struct buf *bp;
daddr_t lbn, nextlbn;
off_t nextloffset;
off_t bytesinfile;
long size, xfersize, blkoffset;
int error, orig_resid;
int seqcount = ap->a_ioflag >> 16;
vp = ap->a_vp;
ip = VTOI(vp);
uio = ap->a_uio;
#ifdef DIAGNOSTIC
if (uio->uio_rw != UIO_READ)
panic("ext2_read: mode");
if (vp->v_type == VLNK) {
if ((int)ip->i_size < vp->v_mount->mnt_maxsymlinklen)
panic("ext2_read: short symlink");
} else if (vp->v_type != VREG && vp->v_type != VDIR)
panic("ext2_read: type %d", vp->v_type);
#endif
fs = ip->I_FS;
#if 0
if ((u_quad_t)uio->uio_offset > fs->fs_maxfilesize)
return (EFBIG);
#endif
orig_resid = uio->uio_resid;
for (error = 0, bp = NULL; uio->uio_resid > 0; bp = NULL) {
if ((bytesinfile = ip->i_size - uio->uio_offset) <= 0)
break;
lbn = lblkno(fs, uio->uio_offset);
nextlbn = lbn + 1;
nextloffset = lblktodoff(fs, nextlbn);
size = BLKSIZE(fs, ip, lbn);
blkoffset = blkoff(fs, uio->uio_offset);
xfersize = fs->s_frag_size - blkoffset;
if (uio->uio_resid < xfersize)
xfersize = uio->uio_resid;
if (bytesinfile < xfersize)
xfersize = bytesinfile;
if (nextloffset >= ip->i_size) {
error = bread(vp, lblktodoff(fs, lbn), size, &bp);
} else if ((vp->v_mount->mnt_flag & MNT_NOCLUSTERR) == 0) {
error = cluster_read(vp, (off_t)ip->i_size,
lblktodoff(fs, lbn), size,
uio->uio_resid,
(ap->a_ioflag >> 16) * BKVASIZE,
&bp);
} else if (seqcount > 1) {
static int
udf_read(struct vop_read_args *ap)
{
struct vnode *vp = ap->a_vp;
struct uio *uio = ap->a_uio;
struct udf_node *node = VTON(vp);
struct udf_mnt *udfmp;
struct file_entry *fentry;
struct buf *bp;
uint8_t *data;
daddr_t lbn, rablock;
off_t diff, fsize;
ssize_t n;
int error = 0;
long size, on;
if (uio->uio_resid == 0)
return (0);
if (uio->uio_offset < 0)
return (EINVAL);
if (is_data_in_fentry(node)) {
fentry = node->fentry;
data = &fentry->data[le32toh(fentry->l_ea)];
fsize = le32toh(fentry->l_ad);
n = uio->uio_resid;
diff = fsize - uio->uio_offset;
if (diff <= 0)
return (0);
if (diff < n)
n = diff;
error = uiomove(data + uio->uio_offset, (int)n, uio);
return (error);
}
fsize = le64toh(node->fentry->inf_len);
udfmp = node->udfmp;
do {
lbn = lblkno(udfmp, uio->uio_offset);
on = blkoff(udfmp, uio->uio_offset);
n = min((u_int)(udfmp->bsize - on),
uio->uio_resid);
diff = fsize - uio->uio_offset;
if (diff <= 0)
return (0);
if (diff < n)
n = diff;
size = udfmp->bsize;
rablock = lbn + 1;
if ((vp->v_mount->mnt_flag & MNT_NOCLUSTERR) == 0) {
if (lblktosize(udfmp, rablock) < fsize) {
error = cluster_read(vp, fsize, lbn, size,
NOCRED, uio->uio_resid,
(ap->a_ioflag >> 16), 0, &bp);
} else {
error = bread(vp, lbn, size, NOCRED, &bp);
}
} else {
/*
* Read a portion of a file into an internal buffer.
* Return the location in the buffer and the amount in the buffer.
*/
static int
buf_read_file(struct open_file *f, char **buf_p, size_t *size_p)
{
struct file *fp = (struct file *)f->f_fsdata;
struct fs *fs = fp->f_fs;
long off;
indp_t file_block;
indp_t disk_block;
size_t block_size;
int rc;
off = blkoff(fs, fp->f_seekp);
file_block = lblkno(fs, fp->f_seekp);
#ifdef LIBSA_LFS
block_size = dblksize(fs, &fp->f_di, file_block);
#else
block_size = sblksize(fs, (int64_t)fp->f_di.di_size, file_block);
#endif
if (file_block != fp->f_buf_blkno) {
rc = block_map(f, file_block, &disk_block);
if (rc)
return rc;
if (disk_block == 0) {
memset(fp->f_buf, 0, block_size);
fp->f_buf_size = block_size;
} else {
twiddle();
rc = DEV_STRATEGY(f->f_dev)(f->f_devdata, F_READ,
FSBTODB(fs, disk_block),
block_size, fp->f_buf, &fp->f_buf_size);
if (rc)
return rc;
}
fp->f_buf_blkno = file_block;
}
/*
* Return address of byte in buffer corresponding to
* offset, and size of remainder of buffer after that
* byte.
*/
*buf_p = fp->f_buf + off;
*size_p = block_size - off;
/*
* But truncate buffer at end of file.
*/
if (*size_p > fp->f_di.di_size - fp->f_seekp)
*size_p = fp->f_di.di_size - fp->f_seekp;
return 0;
}
/*
* Read a portion of a file into an internal buffer. Return
* the location in the buffer and the amount in the buffer.
*
* Parameters:
* buf_p: out
* size_p: out
*/
static int
buf_read_file(struct open_file *f, char **buf_p, size_t *size_p)
{
struct file *fp = (struct file *)f->f_fsdata;
struct fs *fs = fp->f_fs;
long off;
daddr_t file_block;
daddr_t disk_block;
size_t block_size;
int rc;
off = blkoff(fs, fp->f_seekp);
file_block = lblkno(fs, fp->f_seekp);
block_size = dblksize(fs, &fp->f_di, file_block);
if (file_block != fp->f_buf_blkno) {
rc = block_map(f, file_block, &disk_block);
if (rc)
return (rc);
if (fp->f_buf == NULL)
fp->f_buf = malloc(fs->fs_bsize);
if (disk_block == 0) {
bzero(fp->f_buf, block_size);
fp->f_buf_size = block_size;
} else {
twiddle();
rc = (f->f_dev->dv_strategy)(f->f_devdata, F_READ,
fsbtodb(fs, disk_block),
block_size, fp->f_buf, &fp->f_buf_size);
if (rc)
return (rc);
}
fp->f_buf_blkno = file_block;
}
/*
* Return address of byte in buffer corresponding to
* offset, and size of remainder of buffer after that
* byte.
*/
*buf_p = fp->f_buf + off;
*size_p = block_size - off;
/*
* But truncate buffer at end of file.
*/
if (*size_p > fp->f_di.di_size - fp->f_seekp)
*size_p = fp->f_di.di_size - fp->f_seekp;
return (0);
}
/*
* Read a portion of a file into an internal buffer.
* Return the location in the buffer and the amount in the buffer.
*/
static int
buf_read_file(struct open_file *f, char **buf_p, size_t *size_p)
{
struct file *fp = (struct file *)f->f_fsdata;
struct m_ext2fs *fs = fp->f_fs;
long off;
indp_t file_block;
indp_t disk_block;
size_t block_size;
int rc;
off = blkoff(fs, fp->f_seekp);
file_block = lblkno(fs, fp->f_seekp);
block_size = fs->e2fs_bsize; /* no fragment */
if (file_block != fp->f_buf_blkno) {
rc = block_map(f, file_block, &disk_block);
if (rc)
return rc;
if (disk_block == 0) {
memset(fp->f_buf, 0, block_size);
fp->f_buf_size = block_size;
} else {
twiddle();
rc = DEV_STRATEGY(f->f_dev)(f->f_devdata, F_READ,
FSBTODB(fs, disk_block),
block_size, fp->f_buf, &fp->f_buf_size);
if (rc)
return rc;
}
fp->f_buf_blkno = file_block;
}
/*
* Return address of byte in buffer corresponding to
* offset, and size of remainder of buffer after that
* byte.
*/
*buf_p = fp->f_buf + off;
*size_p = block_size - off;
/*
* But truncate buffer at end of file.
*/
/* XXX should handle LARGEFILE */
if (*size_p > fp->f_di.e2di_size - fp->f_seekp)
*size_p = fp->f_di.e2di_size - fp->f_seekp;
return 0;
}
/*
* do_write()
* Local routine to loop over a buffer and write it to a file
*
* Returns 0 on success, 1 on error.
*/
static int
do_write(struct openfile *o, int pos, char *buf, int cnt)
{
uint x, step, blk, boff;
char *blkp;
/*
* Loop across each block, putting our data into place
*/
for (x = 0; x < cnt; x += step) {
/*
* Calculate how much to take out of current block
*/
blkoff(pos, &boff, &step);
if (step > (cnt - x)) {
step = (cnt - x);
}
/*
* Get next block. If not present yet, allocate
* and add to the file.
*/
blk = blknum(pos);
blkp = hash_lookup(o->o_blocks, blk);
if (blkp == 0) {
uint sz;
sz = (blk ? BLOCKSIZE : BLOCKMIN);
blkp = malloc(sz);
if (blkp == 0) {
return(1);
}
bzero(blkp, sz);
if (hash_insert(o->o_blocks, blk, blkp)) {
free(blkp);
return(1);
}
}
/*
* Put contents into block
*/
memcpy(blkp+boff, buf+x, step);
/*
* Advance to next chunk
*/
pos += step;
}
return(0);
}
/*
* Allocate a block in the filesystem.
*
* A preference may be optionally specified. If a preference is given
* the following hierarchy is used to allocate a block:
* 1) allocate the requested block.
* 2) allocate a rotationally optimal block in the same cylinder.
* 3) allocate a block in the same cylinder group.
* 4) quadradically rehash into other cylinder groups, until an
* available block is located.
* If no block preference is given the following hierarchy is used
* to allocate a block:
* 1) allocate a block in the cylinder group that contains the
* inode for the file.
* 2) quadradically rehash into other cylinder groups, until an
* available block is located.
*/
int
ext2_alloc(struct inode *ip, daddr_t lbn, e4fs_daddr_t bpref, int size,
struct ucred *cred, e4fs_daddr_t *bnp)
{
struct m_ext2fs *fs;
struct ext2mount *ump;
int32_t bno;
int cg;
*bnp = 0;
fs = ip->i_e2fs;
ump = ip->i_ump;
mtx_assert(EXT2_MTX(ump), MA_OWNED);
#ifdef INVARIANTS
if ((u_int)size > fs->e2fs_bsize || blkoff(fs, size) != 0) {
vn_printf(ip->i_devvp, "bsize = %lu, size = %d, fs = %s\n",
(long unsigned int)fs->e2fs_bsize, size, fs->e2fs_fsmnt);
panic("ext2_alloc: bad size");
}
if (cred == NOCRED)
panic("ext2_alloc: missing credential");
#endif /* INVARIANTS */
if (size == fs->e2fs_bsize && fs->e2fs->e2fs_fbcount == 0)
goto nospace;
if (cred->cr_uid != 0 &&
fs->e2fs->e2fs_fbcount < fs->e2fs->e2fs_rbcount)
goto nospace;
if (bpref >= fs->e2fs->e2fs_bcount)
bpref = 0;
if (bpref == 0)
cg = ino_to_cg(fs, ip->i_number);
else
cg = dtog(fs, bpref);
bno = (daddr_t)ext2_hashalloc(ip, cg, bpref, fs->e2fs_bsize,
ext2_alloccg);
if (bno > 0) {
/* set next_alloc fields as done in block_getblk */
ip->i_next_alloc_block = lbn;
ip->i_next_alloc_goal = bno;
ip->i_blocks += btodb(fs->e2fs_bsize);
ip->i_flag |= IN_CHANGE | IN_UPDATE;
*bnp = bno;
return (0);
}
nospace:
EXT2_UNLOCK(ump);
ext2_fserr(fs, cred->cr_uid, "filesystem full");
uprintf("\n%s: write failed, filesystem is full\n", fs->e2fs_fsmnt);
return (ENOSPC);
}
static int
find(char *path)
{
char *rest, ch;
int block, off, loc, ino = ROOTINO;
struct dirent *dp;
char list_only;
list_only = (path[0] == '?' && path[1] == '\0');
loop:
devread(iobuf, fsbtodb(fs, ino_to_fsba(fs, ino)) + boff, fs->fs_bsize);
bcopy((void *)&((struct dinode *)iobuf)[ino % fs->fs_inopb],
(void *)&inode.i_din,
sizeof (struct dinode));
if (!*path)
return 1;
while (*path == '/')
path++;
if (!inode.i_size || ((inode.i_mode&IFMT) != IFDIR))
return 0;
for (rest = path; (ch = *rest) && ch != '/'; rest++) ;
*rest = 0;
loc = 0;
do {
if (loc >= inode.i_size) {
if (list_only) {
putchar('\n');
return -1;
} else {
return 0;
}
}
if (!(off = blkoff(fs, loc))) {
block = lblkno(fs, loc);
devread(iobuf, fsbtodb(fs, block_map(block)) + boff,
blksize(fs, &inode, block));
}
dp = (struct dirent *)(iobuf + off);
loc += dp->d_reclen;
if (dp->d_fileno && list_only) {
puts(dp->d_name);
putchar(' ');
}
} while (!dp->d_fileno || strcmp(path, dp->d_name));
ino = dp->d_fileno;
*(path = rest) = ch;
goto loop;
}
static int getch(int fdesc)
{
register struct iob *io;
struct fs *fs;
char *p;
int c, lbn, off, size, diff;
if ((io = iob_from_fdesc(fdesc)) == 0) {
return (-1);
}
p = io->i_ma;
if (io->i_cc <= 0) {
if ((io->i_flgs & F_FILE) != 0) {
diff = io->i_ino.i_size - io->i_offset;
if (diff <= 0)
return (-1);
fs = io->i_fs;
lbn = lblkno(fs, io->i_offset);
io->i_bn = fsbtodb(fs, sbmap(io, lbn)) + io->i_boff;
off = blkoff(fs, io->i_offset);
size = blksize(fs, &io->i_ino, lbn);
} else {
diff = 0;
#ifndef SMALL
io->i_bn = io->i_offset / DEV_BSIZE;
off = 0;
size = DEV_BSIZE;
#endif SMALL
}
io->i_ma = io->i_buf;
io->i_cc = size;
if (devread(io) < 0) {
return (-1);
}
if ((io->i_flgs & F_FILE) != 0) {
if (io->i_offset - off + size >= io->i_ino.i_size)
io->i_cc = diff + off;
io->i_cc -= off;
}
p = &io->i_buf[off];
}
io->i_cc--;
io->i_offset++;
c = (unsigned)*p++;
io->i_ma = p;
return (c);
}
int
find(const char *path)
{
const char *rest;
int ch;
int block, off, loc, ino = ROOTINO;
struct dirent *dp;
for (;;) {
cnt = fs->fs_bsize;
bnum = fsbtodb(fs,itod(fs,ino)) + boff;
if (disk_read(bnum, cnt, (vm_offset_t)iobuf))
return 0;
bcopy((char *)&((struct dinode *)iobuf)[ino % fs->fs_inopb],
(char *)&inode.i_di,
sizeof (struct dinode));
if (!*path)
return 1;
while (*path == '/')
path++;
if (!inode.i_size || ((inode.i_mode&IFMT) != IFDIR))
return 0;
for (rest = path; (ch = *rest) && ch != '/'; rest++) ;
loc = 0;
for (;;) {
if (loc >= inode.i_size)
return 0;
if (!(off = blkoff(fs, loc))) {
block = lblkno(fs, loc);
cnt = blksize(fs, &inode, block);
bnum = fsbtodb(fs, block_map(block)) + boff;
if (disk_read(bnum, cnt, (vm_offset_t)iobuf))
return 0;
}
dp = (struct dirent *)(iobuf + off);
loc += dp->d_reclen;
if (dp->d_ino == 0)
continue;
if (strncmp(path, dp->d_name, rest - path) == 0 &&
dp->d_name[rest - path] == 0)
break;
}
ino = dp->d_ino;
path = rest;
}
}
unsigned long
ufs2_read (char *buf, unsigned long len, unsigned long write)
{
unsigned long logno, off, size, ret = 0;
grub_int64_t map;
while (len && !errnum)
{
off = blkoff (SUPERBLOCK, filepos);
logno = lblkno (SUPERBLOCK, filepos);
size = blksize (SUPERBLOCK, INODE_UFS2, logno);
if ((map = block_map (logno)) < 0)
break;
size -= off;
if (size > len)
size = len;
disk_read_func = disk_read_hook;
devread (fsbtodb (SUPERBLOCK, map), off, size, buf, write);
disk_read_func = NULL;
if (buf)
buf += size;
len -= size; /* len always >= 0 */
filepos += size;
ret += size;
}
if (errnum)
ret = 0;
return ret;
}
int
ufs2_read (char *buf, int len)
{
int logno, off, size, ret = 0;
grub_int64_t map;
while (len && !errnum)
{
off = blkoff (SUPERBLOCK, filepos);
logno = lblkno (SUPERBLOCK, filepos);
size = blksize (SUPERBLOCK, INODE_UFS2, logno);
if ((map = block_map (logno)) < 0)
break;
size -= off;
if (size > len)
size = len;
disk_read_func = disk_read_hook;
devread (fsbtodb (SUPERBLOCK, map), off, size, buf);
disk_read_func = NULL;
buf += size;
len -= size;
filepos += size;
ret += size;
}
if (errnum)
ret = 0;
return ret;
}
/*
* Return buffer with the contents of block "offset" from the beginning of
* directory "ip". If "res" is non-zero, fill it in with a pointer to the
* remaining space in the directory.
*/
int
cd9660_blkatoff(struct vnode *vp, off_t offset, char **res, struct buf **bpp)
{
struct iso_node *ip;
struct iso_mnt *imp;
struct buf *bp;
daddr_t lbn;
int bsize, error;
ip = VTOI(vp);
imp = ip->i_mnt;
lbn = lblkno(imp, offset);
bsize = blksize(imp, ip, lbn);
if ((error = bread(vp, lbn, bsize, NOCRED, 0, &bp)) != 0) {
brelse(bp, 0);
*bpp = NULL;
return (error);
}
if (res)
*res = (char *)bp->b_data + blkoff(imp, offset);
*bpp = bp;
return (0);
}
/*
* Print the block pointers for one inode.
*/
static void
printblocks(ino_t inum, struct ufs1_dinode *dp)
{
char *bufp;
int i, nfrags;
long ndb, offset;
printf("Blocks for inode %d:\n", inum);
printf("Direct blocks:\n");
ndb = howmany(dp->di_size, sblock.fs_bsize);
for (i = 0; i < NDADDR; i++) {
if (dp->di_db[i] == 0) {
putchar('\n');
return;
}
if (i > 0)
printf(", ");
printf("%d", dp->di_db[i]);
if (--ndb == 0 && (offset = blkoff(&sblock, dp->di_size)) != 0) {
nfrags = numfrags(&sblock, fragroundup(&sblock, offset));
printf(" (%d frag%s)", nfrags, nfrags > 1? "s": "");
}
}
putchar('\n');
if (dp->di_ib[0] == 0)
return;
bufp = malloc((unsigned int)sblock.fs_bsize);
if (bufp == NULL)
errx(EEXIT, "cannot allocate indirect block buffer");
printf("Indirect blocks:\n");
for (i = 0; i < NIADDR; i++)
if (printindir(dp->di_ib[i], i, bufp) == 0)
break;
free(bufp);
}
/*
* Print the block pointers for one inode.
*/
static void
printblocks(ino_t inum, union dinode *dp)
{
char *bufp;
int i, nfrags;
long ndb, offset;
ufs2_daddr_t blkno;
printf("Blocks for inode %d:\n", inum);
printf("Direct blocks:\n");
ndb = howmany(DIP(dp, di_size), sblock.fs_bsize);
for (i = 0; i < NDADDR && i < ndb; i++) {
if (i > 0)
printf(", ");
blkno = DIP(dp, di_db[i]);
printf("%jd", (intmax_t)blkno);
}
if (ndb <= NDADDR) {
offset = blkoff(&sblock, DIP(dp, di_size));
if (offset != 0) {
nfrags = numfrags(&sblock, fragroundup(&sblock, offset));
printf(" (%d frag%s)", nfrags, nfrags > 1? "s": "");
}
}
putchar('\n');
if (ndb <= NDADDR)
return;
bufp = malloc((unsigned int)sblock.fs_bsize);
if (bufp == 0)
errx(EEXIT, "cannot allocate indirect block buffer");
printf("Indirect blocks:\n");
for (i = 0; i < NIADDR; i++)
printindir(DIP(dp, di_ib[i]), i, bufp);
free(bufp);
}
/*
* Return buffer with the contents of block "offset" from the beginning of
* directory "ip". If "res" is non-zero, fill it in with a pointer to the
* remaining space in the directory.
*/
int
ext2_blkatoff(struct vnode *vp, off_t offset, char **res, struct buf **bpp)
{
struct inode *ip;
struct ext2_sb_info *fs;
struct buf *bp;
daddr_t lbn;
int bsize, error;
ip = VTOI(vp);
fs = ip->i_e2fs;
lbn = lblkno(fs, offset);
bsize = blksize(fs, ip, lbn);
*bpp = NULL;
if ((error = bread(vp, lblktodoff(fs, lbn), bsize, &bp)) != 0) {
brelse(bp);
return (error);
}
if (res)
*res = (char *)bp->b_data + blkoff(fs, offset);
*bpp = bp;
return (0);
}
int
read(int fdesc, char *buf, int count)
{
int i, size;
register struct iob *file;
struct fs *fs;
int lbn, off;
if ((file = iob_from_fdesc(fdesc)) == 0) {
return (-1);
}
if ((file->i_flgs&F_READ) == 0) {
return (-1);
}
#ifndef SMALL
if ((file->i_flgs & F_FILE) == 0) {
file->i_cc = count;
file->i_ma = buf;
file->i_bn = file->i_boff + (file->i_offset / DEV_BSIZE);
i = devread(file);
file->i_offset += count;
return (i);
}
#endif SMALL
if (file->i_offset+count > file->i_ino.i_size)
count = file->i_ino.i_size - file->i_offset;
if ((i = count) <= 0)
return (0);
/*
* While reading full blocks, do I/O into user buffer.
* Anything else uses getc().
*/
fs = file->i_fs;
while (i) {
off = blkoff(fs, file->i_offset);
lbn = lblkno(fs, file->i_offset);
size = blksize(fs, &file->i_ino, lbn);
if (off == 0 && size <= i) {
file->i_bn = fsbtodb(fs, sbmap(file, lbn)) +
file->i_boff;
file->i_cc = size;
file->i_ma = buf;
if (devread(file) < 0) {
return (-1);
}
file->i_offset += size;
file->i_cc = 0;
buf += size;
i -= size;
} else {
size -= off;
if (size > i)
size = i;
i -= size;
do {
*buf++ = getch(fdesc);
} while (--size);
}
}
return (count);
}
/*
* get next entry in a directory.
*/
struct direct *
readdir(struct dirstuff *dirp)
{
struct direct *dp;
register struct iob *io;
daddr_t lbn, d;
int off;
#if DCACHE
char *bp;
int dirblkno;
if (dcache == 0)
dcache = cacheInit(DCACHE_SIZE, DIRBLKSIZ);
#endif DCACHE
io = dirp->io;
for(;;)
{
if (dirp->loc >= io->i_ino.i_size)
return (NULL);
off = blkoff(io->i_fs, dirp->loc);
lbn = lblkno(io->i_fs, dirp->loc);
#if DCACHE
dirblkno = dirp->loc / DIRBLKSIZ;
if (cacheFind(dcache, io->i_ino.i_number, dirblkno, &bp)) {
dp = (struct direct *)(bp + (dirp->loc % DIRBLKSIZ));
} else
#else DCACHE
if (io->dirbuf_blkno != lbn)
#endif DCACHE
{
if((d = sbmap(io, lbn)) == 0)
return NULL;
io->i_bn = fsbtodb(io->i_fs, d) + io->i_boff;
io->i_ma = io->i_buf;
io->i_cc = blksize(io->i_fs, &io->i_ino, lbn);
if (devread(io) < 0)
{
#if SYS_MESSAGES
error("bn %d: directory read error\n",
io->i_bn);
#endif
return (NULL);
}
#if BIG_ENDIAN_FS
byte_swap_dir_block_in(io->i_buf, io->i_cc);
#endif BIG_ENDIAN_FS
#if DCACHE
bcopy(io->i_buf + dirblkno * DIRBLKSIZ, bp, DIRBLKSIZ);
dp = (struct direct *)(io->i_buf + off);
#endif
}
#if !DCACHE
dp = (struct direct *)(io->i_buf + off);
#endif
dirp->loc += dp->d_reclen;
if (dp->d_ino != 0) return (dp);
}
}
/*
* Vnode op for writing.
*/
static int
ext2_write(struct vop_write_args *ap)
{
struct vnode *vp;
struct uio *uio;
struct inode *ip;
struct m_ext2fs *fs;
struct buf *bp;
daddr_t lbn;
off_t osize;
int blkoffset, error, flags, ioflag, resid, size, seqcount, xfersize;
ioflag = ap->a_ioflag;
uio = ap->a_uio;
vp = ap->a_vp;
seqcount = ioflag >> IO_SEQSHIFT;
ip = VTOI(vp);
#ifdef INVARIANTS
if (uio->uio_rw != UIO_WRITE)
panic("%s: mode", "ext2_write");
#endif
switch (vp->v_type) {
case VREG:
if (ioflag & IO_APPEND)
uio->uio_offset = ip->i_size;
if ((ip->i_flags & APPEND) && uio->uio_offset != ip->i_size)
return (EPERM);
/* FALLTHROUGH */
case VLNK:
break;
case VDIR:
/* XXX differs from ffs -- this is called from ext2_mkdir(). */
if ((ioflag & IO_SYNC) == 0)
panic("ext2_write: nonsync dir write");
break;
default:
panic("ext2_write: type %p %d (%jd,%jd)", (void *)vp,
vp->v_type, (intmax_t)uio->uio_offset,
(intmax_t)uio->uio_resid);
}
KASSERT(uio->uio_resid >= 0, ("ext2_write: uio->uio_resid < 0"));
KASSERT(uio->uio_offset >= 0, ("ext2_write: uio->uio_offset < 0"));
fs = ip->i_e2fs;
if ((uoff_t)uio->uio_offset + uio->uio_resid > fs->e2fs_maxfilesize)
return (EFBIG);
/*
* Maybe this should be above the vnode op call, but so long as
* file servers have no limits, I don't think it matters.
*/
if (vn_rlimit_fsize(vp, uio, uio->uio_td))
return (EFBIG);
resid = uio->uio_resid;
osize = ip->i_size;
if (seqcount > BA_SEQMAX)
flags = BA_SEQMAX << BA_SEQSHIFT;
else
flags = seqcount << BA_SEQSHIFT;
if ((ioflag & IO_SYNC) && !DOINGASYNC(vp))
flags |= IO_SYNC;
for (error = 0; uio->uio_resid > 0;) {
lbn = lblkno(fs, uio->uio_offset);
blkoffset = blkoff(fs, uio->uio_offset);
xfersize = fs->e2fs_fsize - blkoffset;
if (uio->uio_resid < xfersize)
xfersize = uio->uio_resid;
if (uio->uio_offset + xfersize > ip->i_size)
vnode_pager_setsize(vp, uio->uio_offset + xfersize);
/*
* We must perform a read-before-write if the transfer size
* does not cover the entire buffer.
*/
if (fs->e2fs_bsize > xfersize)
flags |= BA_CLRBUF;
else
flags &= ~BA_CLRBUF;
error = ext2_balloc(ip, lbn, blkoffset + xfersize,
ap->a_cred, &bp, flags);
if (error != 0)
break;
/*
* If the buffer is not valid and we did not clear garbage
* out above, we have to do so here even though the write
* covers the entire buffer in order to avoid a mmap()/write
* race where another process may see the garbage prior to
* the uiomove() for a write replacing it.
*/
if ((bp->b_flags & B_CACHE) == 0 && fs->e2fs_bsize <= xfersize)
vfs_bio_clrbuf(bp);
if ((ioflag & (IO_SYNC|IO_INVAL)) == (IO_SYNC|IO_INVAL))
bp->b_flags |= B_NOCACHE;
if (uio->uio_offset + xfersize > ip->i_size)
ip->i_size = uio->uio_offset + xfersize;
size = blksize(fs, ip, lbn) - bp->b_resid;
if (size < xfersize)
xfersize = size;
error =
uiomove((char *)bp->b_data + blkoffset, (int)xfersize, uio);
if (ioflag & (IO_VMIO|IO_DIRECT)) {
bp->b_flags |= B_RELBUF;
}
/*
* If IO_SYNC each buffer is written synchronously. Otherwise
* if we have a severe page deficiency write the buffer
* asynchronously. Otherwise try to cluster, and if that
* doesn't do it then either do an async write (if O_DIRECT),
* or a delayed write (if not).
*/
if (ioflag & IO_SYNC) {
(void)bwrite(bp);
} else if (vm_page_count_severe() ||
buf_dirty_count_severe() ||
(ioflag & IO_ASYNC)) {
bp->b_flags |= B_CLUSTEROK;
bawrite(bp);
} else if (xfersize + blkoffset == fs->e2fs_fsize) {
if ((vp->v_mount->mnt_flag & MNT_NOCLUSTERW) == 0) {
bp->b_flags |= B_CLUSTEROK;
cluster_write(vp, bp, ip->i_size, seqcount, 0);
} else {
bawrite(bp);
}
} else if (ioflag & IO_DIRECT) {
bp->b_flags |= B_CLUSTEROK;
bawrite(bp);
} else {
bp->b_flags |= B_CLUSTEROK;
bdwrite(bp);
}
if (error || xfersize == 0)
break;
}
/*
* If we successfully wrote any data, and we are not the superuser
* we clear the setuid and setgid bits as a precaution against
* tampering.
*/
if ((ip->i_mode & (ISUID | ISGID)) && resid > uio->uio_resid &&
ap->a_cred) {
if (priv_check_cred(ap->a_cred, PRIV_VFS_RETAINSUGID, 0))
ip->i_mode &= ~(ISUID | ISGID);
}
if (error) {
if (ioflag & IO_UNIT) {
(void)ext2_truncate(vp, osize,
ioflag & IO_SYNC, ap->a_cred, uio->uio_td);
uio->uio_offset -= resid - uio->uio_resid;
uio->uio_resid = resid;
}
}
if (uio->uio_resid != resid) {
ip->i_flag |= IN_CHANGE | IN_UPDATE;
if (ioflag & IO_SYNC)
error = ext2_update(vp, 1);
}
return (error);
}
/*
* Vnode op for reading.
*/
static int
ext2_read(struct vop_read_args *ap)
{
struct vnode *vp;
struct inode *ip;
struct uio *uio;
struct m_ext2fs *fs;
struct buf *bp;
daddr_t lbn, nextlbn;
off_t bytesinfile;
long size, xfersize, blkoffset;
int error, orig_resid, seqcount;
int ioflag;
vp = ap->a_vp;
uio = ap->a_uio;
ioflag = ap->a_ioflag;
seqcount = ap->a_ioflag >> IO_SEQSHIFT;
ip = VTOI(vp);
#ifdef INVARIANTS
if (uio->uio_rw != UIO_READ)
panic("%s: mode", "ext2_read");
if (vp->v_type == VLNK) {
if ((int)ip->i_size < vp->v_mount->mnt_maxsymlinklen)
panic("%s: short symlink", "ext2_read");
} else if (vp->v_type != VREG && vp->v_type != VDIR)
panic("%s: type %d", "ext2_read", vp->v_type);
#endif
orig_resid = uio->uio_resid;
KASSERT(orig_resid >= 0, ("ext2_read: uio->uio_resid < 0"));
if (orig_resid == 0)
return (0);
KASSERT(uio->uio_offset >= 0, ("ext2_read: uio->uio_offset < 0"));
fs = ip->i_e2fs;
if (uio->uio_offset < ip->i_size &&
uio->uio_offset >= fs->e2fs_maxfilesize)
return (EOVERFLOW);
for (error = 0, bp = NULL; uio->uio_resid > 0; bp = NULL) {
if ((bytesinfile = ip->i_size - uio->uio_offset) <= 0)
break;
lbn = lblkno(fs, uio->uio_offset);
nextlbn = lbn + 1;
size = blksize(fs, ip, lbn);
blkoffset = blkoff(fs, uio->uio_offset);
xfersize = fs->e2fs_fsize - blkoffset;
if (uio->uio_resid < xfersize)
xfersize = uio->uio_resid;
if (bytesinfile < xfersize)
xfersize = bytesinfile;
if (lblktosize(fs, nextlbn) >= ip->i_size)
error = bread(vp, lbn, size, NOCRED, &bp);
else if ((vp->v_mount->mnt_flag & MNT_NOCLUSTERR) == 0) {
error = cluster_read(vp, ip->i_size, lbn, size,
NOCRED, blkoffset + uio->uio_resid, seqcount,
0, &bp);
} else if (seqcount > 1) {
int nextsize = blksize(fs, ip, nextlbn);
error = breadn(vp, lbn,
size, &nextlbn, &nextsize, 1, NOCRED, &bp);
} else
error = bread(vp, lbn, size, NOCRED, &bp);
if (error) {
brelse(bp);
bp = NULL;
break;
}
/*
* If IO_DIRECT then set B_DIRECT for the buffer. This
* will cause us to attempt to release the buffer later on
* and will cause the buffer cache to attempt to free the
* underlying pages.
*/
if (ioflag & IO_DIRECT)
bp->b_flags |= B_DIRECT;
/*
* We should only get non-zero b_resid when an I/O error
* has occurred, which should cause us to break above.
* However, if the short read did not cause an error,
* then we want to ensure that we do not uiomove bad
* or uninitialized data.
*/
size -= bp->b_resid;
if (size < xfersize) {
if (size == 0)
break;
xfersize = size;
}
error = uiomove((char *)bp->b_data + blkoffset,
(int)xfersize, uio);
if (error)
break;
if (ioflag & (IO_VMIO|IO_DIRECT)) {
/*
* If it's VMIO or direct I/O, then we don't
* need the buf, mark it available for
* freeing. If it's non-direct VMIO, the VM has
* the data.
*/
bp->b_flags |= B_RELBUF;
brelse(bp);
} else {
/*
* Otherwise let whoever
* made the request take care of
* freeing it. We just queue
* it onto another list.
*/
bqrelse(bp);
}
}
/*
* This can only happen in the case of an error
* because the loop above resets bp to NULL on each iteration
* and on normal completion has not set a new value into it.
* so it must have come from a 'break' statement
*/
if (bp != NULL) {
if (ioflag & (IO_VMIO|IO_DIRECT)) {
bp->b_flags |= B_RELBUF;
brelse(bp);
} else {
bqrelse(bp);
}
}
if ((error == 0 || uio->uio_resid != orig_resid) &&
(vp->v_mount->mnt_flag & MNT_NOATIME) == 0)
ip->i_flag |= IN_ACCESS;
return (error);
}
// ffs文件系统的写入操作
int
ffs_write(void *v)
{
struct vop_write_args *ap = v;
struct vnode *vp;
struct uio *uio;
struct inode *ip;
struct fs *fs;
struct buf *bp;
daddr_t lbn;
off_t osize;
int blkoffset, error, extended, flags, ioflag, size, xfersize;
ssize_t resid, overrun;
extended = 0;
ioflag = ap->a_ioflag;
uio = ap->a_uio;
vp = ap->a_vp;
ip = VTOI(vp);
#ifdef DIAGNOSTIC
if (uio->uio_rw != UIO_WRITE)
panic("ffs_write: mode");
#endif
/*
* If writing 0 bytes, succeed and do not change
* update time or file offset (standards compliance)
*/
if (uio->uio_resid == 0)
return (0);
switch (vp->v_type) {
case VREG:
if (ioflag & IO_APPEND)
uio->uio_offset = DIP(ip, size);
if ((DIP(ip, flags) & APPEND) && uio->uio_offset != DIP(ip, size))
return (EPERM);
/* FALLTHROUGH */
case VLNK:
break;
case VDIR:
if ((ioflag & IO_SYNC) == 0)
panic("ffs_write: nonsync dir write");
break;
default:
panic("ffs_write: type");
}
fs = ip->i_fs;
if (uio->uio_offset < 0 ||
(u_int64_t)uio->uio_offset + uio->uio_resid > fs->fs_maxfilesize)
return (EFBIG);
/* do the filesize rlimit check */
if ((error = vn_fsizechk(vp, uio, ioflag, &overrun)))
return (error);
resid = uio->uio_resid;
osize = DIP(ip, size);
flags = ioflag & IO_SYNC ? B_SYNC : 0;
for (error = 0; uio->uio_resid > 0;) {
lbn = lblkno(fs, uio->uio_offset);
blkoffset = blkoff(fs, uio->uio_offset);
xfersize = fs->fs_bsize - blkoffset;
if (uio->uio_resid < xfersize)
xfersize = uio->uio_resid;
if (fs->fs_bsize > xfersize)
flags |= B_CLRBUF;
else
flags &= ~B_CLRBUF;
if ((error = UFS_BUF_ALLOC(ip, uio->uio_offset, xfersize,
ap->a_cred, flags, &bp)) != 0)
break;
if (uio->uio_offset + xfersize > DIP(ip, size)) {
DIP_ASSIGN(ip, size, uio->uio_offset + xfersize);
uvm_vnp_setsize(vp, DIP(ip, size));
extended = 1;
}
(void)uvm_vnp_uncache(vp);
size = blksize(fs, ip, lbn) - bp->b_resid;
if (size < xfersize)
xfersize = size;
error =
uiomovei(bp->b_data + blkoffset, xfersize, uio);
if (error != 0)
memset(bp->b_data + blkoffset, 0, xfersize);
#if 0
if (ioflag & IO_NOCACHE)
bp->b_flags |= B_NOCACHE;
#endif
if (ioflag & IO_SYNC)
(void)bwrite(bp);
else if (xfersize + blkoffset == fs->fs_bsize) {
if (doclusterwrite)
cluster_write(bp, &ip->i_ci, DIP(ip, size));
else
bawrite(bp);
} else
bdwrite(bp);
if (error || xfersize == 0)
break;
ip->i_flag |= IN_CHANGE | IN_UPDATE;
}
/*
* If we successfully wrote any data, and we are not the superuser
* we clear the setuid and setgid bits as a precaution against
* tampering.
*/
if (resid > uio->uio_resid && ap->a_cred && ap->a_cred->cr_uid != 0)
DIP_ASSIGN(ip, mode, DIP(ip, mode) & ~(ISUID | ISGID));
if (resid > uio->uio_resid)
VN_KNOTE(vp, NOTE_WRITE | (extended ? NOTE_EXTEND : 0));
if (error) {
if (ioflag & IO_UNIT) {
(void)UFS_TRUNCATE(ip, osize,
ioflag & IO_SYNC, ap->a_cred);
uio->uio_offset -= resid - uio->uio_resid;
uio->uio_resid = resid;
}
} else if (resid > uio->uio_resid && (ioflag & IO_SYNC)) {
error = UFS_UPDATE(ip, 1);
}
/* correct the result for writes clamped by vn_fsizechk() */
uio->uio_resid += overrun;
return (error);
}
/*
* Vnode op for reading.
*/
int
ffs_read(void *v)
{
struct vop_read_args *ap = v;
struct vnode *vp;
struct inode *ip;
struct uio *uio;
struct fs *fs;
struct buf *bp;
daddr_t lbn, nextlbn;
off_t bytesinfile;
long size, xfersize, blkoffset;
mode_t mode;
int error;
vp = ap->a_vp;
ip = VTOI(vp);
mode = DIP(ip, mode);
uio = ap->a_uio;
#ifdef DIAGNOSTIC
if (uio->uio_rw != UIO_READ)
panic("ffs_read: mode");
if (vp->v_type == VLNK) {
if ((int)DIP(ip, size) < vp->v_mount->mnt_maxsymlinklen ||
(vp->v_mount->mnt_maxsymlinklen == 0 &&
DIP(ip, blocks) == 0))
panic("ffs_read: short symlink");
} else if (vp->v_type != VREG && vp->v_type != VDIR)
panic("ffs_read: type %d", vp->v_type);
#endif
fs = ip->i_fs;
if ((u_int64_t)uio->uio_offset > fs->fs_maxfilesize)
return (EFBIG);
if (uio->uio_resid == 0)
return (0);
for (error = 0, bp = NULL; uio->uio_resid > 0; bp = NULL) {
if ((bytesinfile = DIP(ip, size) - uio->uio_offset) <= 0)
break;
lbn = lblkno(fs, uio->uio_offset);
nextlbn = lbn + 1;
size = fs->fs_bsize; /* WAS blksize(fs, ip, lbn); */
blkoffset = blkoff(fs, uio->uio_offset);
xfersize = fs->fs_bsize - blkoffset;
if (uio->uio_resid < xfersize)
xfersize = uio->uio_resid;
if (bytesinfile < xfersize)
xfersize = bytesinfile;
if (lblktosize(fs, nextlbn) >= DIP(ip, size))
error = bread(vp, lbn, size, &bp);
else if (lbn - 1 == ip->i_ci.ci_lastr) {
error = bread_cluster(vp, lbn, size, &bp);
} else
error = bread(vp, lbn, size, &bp);
if (error)
break;
ip->i_ci.ci_lastr = lbn;
/*
* We should only get non-zero b_resid when an I/O error
* has occurred, which should cause us to break above.
* However, if the short read did not cause an error,
* then we want to ensure that we do not uiomove bad
* or uninitialized data.
*/
size -= bp->b_resid;
if (size < xfersize) {
if (size == 0)
break;
xfersize = size;
}
error = uiomovei(bp->b_data + blkoffset, (int)xfersize, uio);
if (error)
break;
brelse(bp);
}
if (bp != NULL)
brelse(bp);
if (!(vp->v_mount->mnt_flag & MNT_NOATIME) ||
(ip->i_flag & (IN_CHANGE | IN_UPDATE))) {
ip->i_flag |= IN_ACCESS;
}
return (error);
}
/*
* Truncate the inode oip to at most length size, freeing the
* disk blocks.
*/
int
ffs_truncate(struct vnode *vp, off_t length, int flags, struct ucred *cred)
{
struct vnode *ovp = vp;
ufs_daddr_t lastblock;
struct inode *oip;
ufs_daddr_t bn, lbn, lastiblock[NIADDR], indir_lbn[NIADDR];
ufs_daddr_t oldblks[NDADDR + NIADDR], newblks[NDADDR + NIADDR];
struct fs *fs;
struct buf *bp;
int offset, size, level;
long count, nblocks, blocksreleased = 0;
int i;
int aflags, error, allerror;
off_t osize;
oip = VTOI(ovp);
fs = oip->i_fs;
if (length < 0)
return (EINVAL);
if (length > fs->fs_maxfilesize)
return (EFBIG);
if (ovp->v_type == VLNK &&
(oip->i_size < ovp->v_mount->mnt_maxsymlinklen || oip->i_din.di_blocks == 0)) {
#ifdef DIAGNOSTIC
if (length != 0)
panic("ffs_truncate: partial truncate of symlink");
#endif /* DIAGNOSTIC */
bzero((char *)&oip->i_shortlink, (uint)oip->i_size);
oip->i_size = 0;
oip->i_flag |= IN_CHANGE | IN_UPDATE;
return (ffs_update(ovp, 1));
}
if (oip->i_size == length) {
oip->i_flag |= IN_CHANGE | IN_UPDATE;
return (ffs_update(ovp, 0));
}
if (fs->fs_ronly)
panic("ffs_truncate: read-only filesystem");
#ifdef QUOTA
error = ufs_getinoquota(oip);
if (error)
return (error);
#endif
ovp->v_lasta = ovp->v_clen = ovp->v_cstart = ovp->v_lastw = 0;
if (DOINGSOFTDEP(ovp)) {
if (length > 0 || softdep_slowdown(ovp)) {
/*
* If a file is only partially truncated, then
* we have to clean up the data structures
* describing the allocation past the truncation
* point. Finding and deallocating those structures
* is a lot of work. Since partial truncation occurs
* rarely, we solve the problem by syncing the file
* so that it will have no data structures left.
*/
if ((error = VOP_FSYNC(ovp, MNT_WAIT, 0)) != 0)
return (error);
} else {
#ifdef QUOTA
(void) ufs_chkdq(oip, -oip->i_blocks, NOCRED, 0);
#endif
softdep_setup_freeblocks(oip, length);
vinvalbuf(ovp, 0, 0, 0);
nvnode_pager_setsize(ovp, 0, fs->fs_bsize, 0);
oip->i_flag |= IN_CHANGE | IN_UPDATE;
return (ffs_update(ovp, 0));
}
}
osize = oip->i_size;
/*
* Lengthen the size of the file. We must ensure that the
* last byte of the file is allocated. Since the smallest
* value of osize is 0, length will be at least 1.
*
* nvextendbuf() only breads the old buffer. The blocksize
* of the new buffer must be specified so it knows how large
* to make the VM object.
*/
if (osize < length) {
nvextendbuf(vp, osize, length,
blkoffsize(fs, oip, osize), /* oblksize */
blkoffresize(fs, length), /* nblksize */
blkoff(fs, osize),
blkoff(fs, length),
0);
aflags = B_CLRBUF;
if (flags & IO_SYNC)
aflags |= B_SYNC;
/* BALLOC will reallocate the fragment at the old EOF */
error = VOP_BALLOC(ovp, length - 1, 1, cred, aflags, &bp);
if (error)
return (error);
oip->i_size = length;
if (bp->b_bufsize == fs->fs_bsize)
bp->b_flags |= B_CLUSTEROK;
if (aflags & B_SYNC)
bwrite(bp);
else
bawrite(bp);
oip->i_flag |= IN_CHANGE | IN_UPDATE;
return (ffs_update(ovp, 1));
}
/*
* Shorten the size of the file.
*
* NOTE: The block size specified in nvtruncbuf() is the blocksize
* of the buffer containing length prior to any reallocation
* of the block.
*/
allerror = nvtruncbuf(ovp, length, blkoffsize(fs, oip, length),
blkoff(fs, length), 0);
offset = blkoff(fs, length);
if (offset == 0) {
oip->i_size = length;
} else {
lbn = lblkno(fs, length);
aflags = B_CLRBUF;
if (flags & IO_SYNC)
aflags |= B_SYNC;
error = VOP_BALLOC(ovp, length - 1, 1, cred, aflags, &bp);
if (error)
return (error);
/*
* When we are doing soft updates and the UFS_BALLOC
* above fills in a direct block hole with a full sized
* block that will be truncated down to a fragment below,
* we must flush out the block dependency with an FSYNC
* so that we do not get a soft updates inconsistency
* when we create the fragment below.
*
* nvtruncbuf() may have re-dirtied the underlying block
* as part of its truncation zeroing code. To avoid a
* 'locking against myself' panic in the second fsync we
* can simply undirty the bp since the redirtying was
* related to areas of the buffer that we are going to
* throw away anyway, and we will b*write() the remainder
* anyway down below.
*/
if (DOINGSOFTDEP(ovp) && lbn < NDADDR &&
fragroundup(fs, blkoff(fs, length)) < fs->fs_bsize) {
bundirty(bp);
error = VOP_FSYNC(ovp, MNT_WAIT, 0);
if (error) {
bdwrite(bp);
return (error);
}
}
oip->i_size = length;
size = blksize(fs, oip, lbn);
#if 0
/* remove - nvtruncbuf deals with this */
if (ovp->v_type != VDIR)
bzero((char *)bp->b_data + offset,
(uint)(size - offset));
#endif
/* Kirk's code has reallocbuf(bp, size, 1) here */
allocbuf(bp, size);
if (bp->b_bufsize == fs->fs_bsize)
bp->b_flags |= B_CLUSTEROK;
if (aflags & B_SYNC)
bwrite(bp);
else
bawrite(bp);
}
/*
* Calculate index into inode's block list of
* last direct and indirect blocks (if any)
* which we want to keep. Lastblock is -1 when
* the file is truncated to 0.
*/
lastblock = lblkno(fs, length + fs->fs_bsize - 1) - 1;
lastiblock[SINGLE] = lastblock - NDADDR;
lastiblock[DOUBLE] = lastiblock[SINGLE] - NINDIR(fs);
lastiblock[TRIPLE] = lastiblock[DOUBLE] - NINDIR(fs) * NINDIR(fs);
nblocks = btodb(fs->fs_bsize);
/*
* Update file and block pointers on disk before we start freeing
* blocks. If we crash before free'ing blocks below, the blocks
* will be returned to the free list. lastiblock values are also
* normalized to -1 for calls to ffs_indirtrunc below.
*/
bcopy((caddr_t)&oip->i_db[0], (caddr_t)oldblks, sizeof oldblks);
for (level = TRIPLE; level >= SINGLE; level--)
if (lastiblock[level] < 0) {
oip->i_ib[level] = 0;
lastiblock[level] = -1;
}
for (i = NDADDR - 1; i > lastblock; i--)
oip->i_db[i] = 0;
oip->i_flag |= IN_CHANGE | IN_UPDATE;
error = ffs_update(ovp, 1);
if (error && allerror == 0)
allerror = error;
/*
* Having written the new inode to disk, save its new configuration
* and put back the old block pointers long enough to process them.
* Note that we save the new block configuration so we can check it
* when we are done.
*/
bcopy((caddr_t)&oip->i_db[0], (caddr_t)newblks, sizeof newblks);
bcopy((caddr_t)oldblks, (caddr_t)&oip->i_db[0], sizeof oldblks);
oip->i_size = osize;
if (error && allerror == 0)
allerror = error;
/*
* Indirect blocks first.
*/
indir_lbn[SINGLE] = -NDADDR;
indir_lbn[DOUBLE] = indir_lbn[SINGLE] - NINDIR(fs) - 1;
indir_lbn[TRIPLE] = indir_lbn[DOUBLE] - NINDIR(fs) * NINDIR(fs) - 1;
for (level = TRIPLE; level >= SINGLE; level--) {
bn = oip->i_ib[level];
if (bn != 0) {
error = ffs_indirtrunc(oip, indir_lbn[level],
fsbtodb(fs, bn), lastiblock[level], level, &count);
if (error)
allerror = error;
blocksreleased += count;
if (lastiblock[level] < 0) {
oip->i_ib[level] = 0;
ffs_blkfree(oip, bn, fs->fs_bsize);
blocksreleased += nblocks;
}
}
if (lastiblock[level] >= 0)
goto done;
}
/*
* All whole direct blocks or frags.
*/
for (i = NDADDR - 1; i > lastblock; i--) {
long bsize;
bn = oip->i_db[i];
if (bn == 0)
continue;
oip->i_db[i] = 0;
bsize = blksize(fs, oip, i);
ffs_blkfree(oip, bn, bsize);
blocksreleased += btodb(bsize);
}
if (lastblock < 0)
goto done;
/*
* Finally, look for a change in size of the
* last direct block; release any frags.
*/
bn = oip->i_db[lastblock];
if (bn != 0) {
long oldspace, newspace;
/*
* Calculate amount of space we're giving
* back as old block size minus new block size.
*/
oldspace = blksize(fs, oip, lastblock);
oip->i_size = length;
newspace = blksize(fs, oip, lastblock);
if (newspace == 0)
panic("ffs_truncate: newspace");
if (oldspace - newspace > 0) {
/*
* Block number of space to be free'd is
* the old block # plus the number of frags
* required for the storage we're keeping.
*/
bn += numfrags(fs, newspace);
ffs_blkfree(oip, bn, oldspace - newspace);
blocksreleased += btodb(oldspace - newspace);
}
}
done:
#ifdef DIAGNOSTIC
for (level = SINGLE; level <= TRIPLE; level++)
if (newblks[NDADDR + level] != oip->i_ib[level])
panic("ffs_truncate1");
for (i = 0; i < NDADDR; i++)
if (newblks[i] != oip->i_db[i])
panic("ffs_truncate2");
if (length == 0 && !RB_EMPTY(&ovp->v_rbdirty_tree))
panic("ffs_truncate3");
#endif /* DIAGNOSTIC */
/*
* Put back the real size.
*/
oip->i_size = length;
oip->i_blocks -= blocksreleased;
if (oip->i_blocks < 0) /* sanity */
oip->i_blocks = 0;
oip->i_flag |= IN_CHANGE;
#ifdef QUOTA
(void) ufs_chkdq(oip, -blocksreleased, NOCRED, 0);
#endif
return (allerror);
}
/*
* Balloc defines the structure of filesystem storage
* by allocating the physical blocks on a device given
* the inode and the logical block number in a file.
*/
int
ext2_balloc(struct inode *ip, e2fs_lbn_t lbn, int size, struct ucred *cred,
struct buf **bpp, int flags)
{
struct m_ext2fs *fs;
struct ext2mount *ump;
struct buf *bp, *nbp;
struct vnode *vp = ITOV(ip);
struct indir indirs[NIADDR + 2];
e4fs_daddr_t nb, newb;
e2fs_daddr_t *bap, pref;
int osize, nsize, num, i, error;
*bpp = NULL;
if (lbn < 0)
return (EFBIG);
fs = ip->i_e2fs;
ump = ip->i_ump;
/*
* check if this is a sequential block allocation.
* If so, increment next_alloc fields to allow ext2_blkpref
* to make a good guess
*/
if (lbn == ip->i_next_alloc_block + 1) {
ip->i_next_alloc_block++;
ip->i_next_alloc_goal++;
}
/*
* The first NDADDR blocks are direct blocks
*/
if (lbn < NDADDR) {
nb = ip->i_db[lbn];
/* no new block is to be allocated, and no need to expand
the file */
if (nb != 0 && ip->i_size >= (lbn + 1) * fs->e2fs_bsize) {
error = bread(vp, lbn, fs->e2fs_bsize, NOCRED, &bp);
if (error) {
brelse(bp);
return (error);
}
bp->b_blkno = fsbtodb(fs, nb);
*bpp = bp;
return (0);
}
if (nb != 0) {
/*
* Consider need to reallocate a fragment.
*/
osize = fragroundup(fs, blkoff(fs, ip->i_size));
nsize = fragroundup(fs, size);
if (nsize <= osize) {
error = bread(vp, lbn, osize, NOCRED, &bp);
if (error) {
brelse(bp);
return (error);
}
bp->b_blkno = fsbtodb(fs, nb);
} else {
/* Godmar thinks: this shouldn't happen w/o fragments */
printf("nsize %d(%d) > osize %d(%d) nb %d\n",
(int)nsize, (int)size, (int)osize,
(int)ip->i_size, (int)nb);
panic(
"ext2_balloc: Something is terribly wrong");
/*
* please note there haven't been any changes from here on -
* FFS seems to work.
*/
}
} else {
if (ip->i_size < (lbn + 1) * fs->e2fs_bsize)
nsize = fragroundup(fs, size);
else
nsize = fs->e2fs_bsize;
EXT2_LOCK(ump);
error = ext2_alloc(ip, lbn,
ext2_blkpref(ip, lbn, (int)lbn, &ip->i_db[0], 0),
nsize, cred, &newb);
if (error)
return (error);
bp = getblk(vp, lbn, nsize, 0, 0, 0);
bp->b_blkno = fsbtodb(fs, newb);
if (flags & BA_CLRBUF)
vfs_bio_clrbuf(bp);
}
ip->i_db[lbn] = dbtofsb(fs, bp->b_blkno);
ip->i_flag |= IN_CHANGE | IN_UPDATE;
*bpp = bp;
return (0);
}
/*
* Determine the number of levels of indirection.
*/
pref = 0;
if ((error = ext2_getlbns(vp, lbn, indirs, &num)) != 0)
return (error);
#ifdef INVARIANTS
if (num < 1)
panic ("ext2_balloc: ext2_getlbns returned indirect block");
#endif
/*
* Fetch the first indirect block allocating if necessary.
*/
--num;
nb = ip->i_ib[indirs[0].in_off];
if (nb == 0) {
EXT2_LOCK(ump);
pref = ext2_blkpref(ip, lbn, indirs[0].in_off +
EXT2_NDIR_BLOCKS, &ip->i_db[0], 0);
if ((error = ext2_alloc(ip, lbn, pref, fs->e2fs_bsize, cred,
&newb)))
return (error);
nb = newb;
bp = getblk(vp, indirs[1].in_lbn, fs->e2fs_bsize, 0, 0, 0);
bp->b_blkno = fsbtodb(fs, newb);
vfs_bio_clrbuf(bp);
/*
* Write synchronously so that indirect blocks
* never point at garbage.
*/
if ((error = bwrite(bp)) != 0) {
ext2_blkfree(ip, nb, fs->e2fs_bsize);
return (error);
}
ip->i_ib[indirs[0].in_off] = newb;
ip->i_flag |= IN_CHANGE | IN_UPDATE;
}
/*
* Fetch through the indirect blocks, allocating as necessary.
*/
for (i = 1;;) {
error = bread(vp,
indirs[i].in_lbn, (int)fs->e2fs_bsize, NOCRED, &bp);
if (error) {
brelse(bp);
return (error);
}
bap = (e2fs_daddr_t *)bp->b_data;
nb = bap[indirs[i].in_off];
if (i == num)
break;
i += 1;
if (nb != 0) {
bqrelse(bp);
continue;
}
EXT2_LOCK(ump);
if (pref == 0)
pref = ext2_blkpref(ip, lbn, indirs[i].in_off, bap,
bp->b_lblkno);
error = ext2_alloc(ip, lbn, pref, (int)fs->e2fs_bsize, cred, &newb);
if (error) {
brelse(bp);
return (error);
}
nb = newb;
nbp = getblk(vp, indirs[i].in_lbn, fs->e2fs_bsize, 0, 0, 0);
nbp->b_blkno = fsbtodb(fs, nb);
vfs_bio_clrbuf(nbp);
/*
* Write synchronously so that indirect blocks
* never point at garbage.
*/
if ((error = bwrite(nbp)) != 0) {
ext2_blkfree(ip, nb, fs->e2fs_bsize);
EXT2_UNLOCK(ump);
brelse(bp);
return (error);
}
bap[indirs[i - 1].in_off] = nb;
/*
* If required, write synchronously, otherwise use
* delayed write.
*/
if (flags & IO_SYNC) {
bwrite(bp);
} else {
if (bp->b_bufsize == fs->e2fs_bsize)
bp->b_flags |= B_CLUSTEROK;
bdwrite(bp);
}
}
/*
* Get the data block, allocating if necessary.
*/
if (nb == 0) {
EXT2_LOCK(ump);
pref = ext2_blkpref(ip, lbn, indirs[i].in_off, &bap[0],
bp->b_lblkno);
if ((error = ext2_alloc(ip,
lbn, pref, (int)fs->e2fs_bsize, cred, &newb)) != 0) {
brelse(bp);
return (error);
}
nb = newb;
nbp = getblk(vp, lbn, fs->e2fs_bsize, 0, 0, 0);
nbp->b_blkno = fsbtodb(fs, nb);
if (flags & BA_CLRBUF)
vfs_bio_clrbuf(nbp);
bap[indirs[i].in_off] = nb;
/*
* If required, write synchronously, otherwise use
* delayed write.
*/
if (flags & IO_SYNC) {
bwrite(bp);
} else {
if (bp->b_bufsize == fs->e2fs_bsize)
bp->b_flags |= B_CLUSTEROK;
bdwrite(bp);
}
*bpp = nbp;
return (0);
}
brelse(bp);
if (flags & BA_CLRBUF) {
int seqcount = (flags & BA_SEQMASK) >> BA_SEQSHIFT;
if (seqcount && (vp->v_mount->mnt_flag & MNT_NOCLUSTERR) == 0) {
error = cluster_read(vp, ip->i_size, lbn,
(int)fs->e2fs_bsize, NOCRED,
MAXBSIZE, seqcount, 0, &nbp);
} else {
error = bread(vp, lbn, (int)fs->e2fs_bsize, NOCRED, &nbp);
}
if (error) {
brelse(nbp);
return (error);
}
} else {
