static errcode_t lookup_journal_block(ocfs2_filesys *fs,
struct journal_info *ji,
uint64_t blkoff,
uint64_t *blkno,
int check_dup)
{
errcode_t ret;
uint64_t contig;
int was_set;
ret = ocfs2_extent_map_get_blocks(ji->ji_cinode, blkoff, 1, blkno,
&contig, NULL);
if (ret) {
com_err(whoami, ret, "while looking up logical block "
"%"PRIu64" in slot %d's journal", blkoff, ji->ji_slot);
goto out;
}
if (check_dup) {
o2fsck_bitmap_set(ji->ji_used_blocks, *blkno, &was_set);
if (was_set) {
printf("Logical block %"PRIu64" in slot %d's journal "
"maps to block %"PRIu64" which has already "
"been used in another journal.\n", blkoff,
ji->ji_slot, *blkno);
ret = OCFS2_ET_DUPLICATE_BLOCK;
}
}
out:
return ret;
}
static errcode_t check_journals_func(o2fsck_state *ost,
ocfs2_cached_inode *ci,
struct journal_check_context *jc)
{
errcode_t err, ret;
ocfs2_filesys *fs = ost->ost_fs;
uint64_t contig;
uint64_t blkno;
char *buf = NULL;
journal_superblock_t *jsb;
struct journal_check_info *ji = &(jc->jc_info[jc->jc_this_slot]);
ret = ocfs2_malloc_blocks(fs->fs_io, 1, &buf);
if (ret)
goto out;
err = ocfs2_extent_map_get_blocks(ci, 0, 1, &blkno, &contig, NULL);
if (err) {
ji->i_error = err;
goto out;
}
ji->i_clusters = ci->ci_inode->i_clusters;
err = ocfs2_read_journal_superblock(fs, blkno, buf);
if (err) {
ji->i_error = err;
goto out;
}
jsb = (journal_superblock_t *)buf;
ji->i_features.opt_compat = jsb->s_feature_compat;
ji->i_features.opt_ro_compat = jsb->s_feature_ro_compat;
ji->i_features.opt_incompat = jsb->s_feature_incompat;
if (!ji->i_clusters) {
ji->i_error = OCFS2_ET_JOURNAL_TOO_SMALL;
goto out;
}
jc->jc_max_clusters = ocfs2_max(jc->jc_max_clusters,
ci->ci_inode->i_clusters);
jc->jc_max_features.opt_compat |= jsb->s_feature_compat;
jc->jc_max_features.opt_ro_compat |= jsb->s_feature_ro_compat;
jc->jc_max_features.opt_incompat |= jsb->s_feature_incompat;
ji->i_error = 0;
out:
if (buf)
ocfs2_free(&buf);
return ret;
}
int ocfs2_init_slot_info(struct ocfs2_super *osb)
{
int status, i;
u64 blkno;
struct inode *inode = NULL;
struct buffer_head *bh = NULL;
struct ocfs2_slot_info *si;
si = kzalloc(sizeof(struct ocfs2_slot_info), GFP_KERNEL);
if (!si) {
status = -ENOMEM;
mlog_errno(status);
goto bail;
}
spin_lock_init(&si->si_lock);
si->si_num_slots = osb->max_slots;
si->si_size = OCFS2_MAX_SLOTS;
for(i = 0; i < si->si_num_slots; i++)
si->si_global_node_nums[i] = OCFS2_INVALID_SLOT;
inode = ocfs2_get_system_file_inode(osb, SLOT_MAP_SYSTEM_INODE,
OCFS2_INVALID_SLOT);
if (!inode) {
status = -EINVAL;
mlog_errno(status);
goto bail;
}
status = ocfs2_extent_map_get_blocks(inode, 0ULL, 1, &blkno, NULL);
if (status < 0) {
mlog_errno(status);
goto bail;
}
status = ocfs2_read_block(osb, blkno, &bh, 0, inode);
if (status < 0) {
mlog_errno(status);
goto bail;
}
si->si_inode = inode;
si->si_bh = bh;
osb->slot_info = si;
bail:
if (status < 0 && si)
ocfs2_free_slot_info(si);
return status;
}
/*
* JBD Might read a cached version of another nodes journal file. We
* don't want this as this file changes often and we get no
* notification on those changes. The only way to be sure that we've
* got the most up to date version of those blocks then is to force
* read them off disk. Just searching through the buffer cache won't
* work as there may be pages backing this file which are still marked
* up to date. We know things can't change on this file underneath us
* as we have the lock by now :)
*/
static int ocfs2_force_read_journal(struct inode *inode)
{
int status = 0;
int i;
u64 v_blkno, p_blkno, p_blocks, num_blocks;
#define CONCURRENT_JOURNAL_FILL 32ULL
struct buffer_head *bhs[CONCURRENT_JOURNAL_FILL];
mlog_entry_void();
memset(bhs, 0, sizeof(struct buffer_head *) * CONCURRENT_JOURNAL_FILL);
num_blocks = ocfs2_blocks_for_bytes(inode->i_sb, inode->i_size);
v_blkno = 0;
while (v_blkno < num_blocks) {
status = ocfs2_extent_map_get_blocks(inode, v_blkno,
&p_blkno, &p_blocks, NULL);
if (status < 0) {
mlog_errno(status);
goto bail;
}
if (p_blocks > CONCURRENT_JOURNAL_FILL)
p_blocks = CONCURRENT_JOURNAL_FILL;
/* We are reading journal data which should not
* be put in the uptodate cache */
status = ocfs2_read_blocks(OCFS2_SB(inode->i_sb),
p_blkno, p_blocks, bhs, 0,
NULL);
if (status < 0) {
mlog_errno(status);
goto bail;
}
for(i = 0; i < p_blocks; i++) {
brelse(bhs[i]);
bhs[i] = NULL;
}
v_blkno += p_blocks;
}
bail:
for(i = 0; i < CONCURRENT_JOURNAL_FILL; i++)
if (bhs[i])
brelse(bhs[i]);
mlog_exit(status);
return status;
}
static int ocfs2_direct_IO_get_blocks(struct inode *inode, sector_t iblock,
struct buffer_head *bh_result, int create)
{
int ret;
u64 p_blkno, inode_blocks, contig_blocks;
unsigned int ext_flags;
unsigned char blocksize_bits = inode->i_sb->s_blocksize_bits;
unsigned long max_blocks = bh_result->b_size >> inode->i_blkbits;
/* This function won't even be called if the request isn't all
* nicely aligned and of the right size, so there's no need
* for us to check any of that. */
inode_blocks = ocfs2_blocks_for_bytes(inode->i_sb, i_size_read(inode));
/* This figures out the size of the next contiguous block, and
* our logical offset */
ret = ocfs2_extent_map_get_blocks(inode, iblock, &p_blkno,
&contig_blocks, &ext_flags);
if (ret) {
mlog(ML_ERROR, "get_blocks() failed iblock=%llu\n",
(unsigned long long)iblock);
ret = -EIO;
goto bail;
}
/* We should already CoW the refcounted extent in case of create. */
BUG_ON(create && (ext_flags & OCFS2_EXT_REFCOUNTED));
/*
* get_more_blocks() expects us to describe a hole by clearing
* the mapped bit on bh_result().
*
* Consider an unwritten extent as a hole.
*/
if (p_blkno && !(ext_flags & OCFS2_EXT_UNWRITTEN))
map_bh(bh_result, inode->i_sb, p_blkno);
else
clear_buffer_mapped(bh_result);
/* make sure we don't map more than max_blocks blocks here as
that's all the kernel will handle at this point. */
if (max_blocks < contig_blocks)
contig_blocks = max_blocks;
bh_result->b_size = contig_blocks << blocksize_bits;
bail:
return ret;
}
static sector_t ocfs2_bmap(struct address_space *mapping, sector_t block)
{
sector_t status;
u64 p_blkno = 0;
int err = 0;
struct inode *inode = mapping->host;
mlog_entry("(block = %llu)\n", (unsigned long long)block);
/* We don't need to lock journal system files, since they aren't
* accessed concurrently from multiple nodes.
*/
if (!INODE_JOURNAL(inode)) {
err = ocfs2_inode_lock(inode, NULL, 0);
if (err) {
if (err != -ENOENT)
mlog_errno(err);
goto bail;
}
down_read(&OCFS2_I(inode)->ip_alloc_sem);
}
if (!(OCFS2_I(inode)->ip_dyn_features & OCFS2_INLINE_DATA_FL))
err = ocfs2_extent_map_get_blocks(inode, block, &p_blkno, NULL,
NULL);
if (!INODE_JOURNAL(inode)) {
up_read(&OCFS2_I(inode)->ip_alloc_sem);
ocfs2_inode_unlock(inode, 0);
}
if (err) {
mlog(ML_ERROR, "get_blocks() failed, block = %llu\n",
(unsigned long long)block);
mlog_errno(err);
goto bail;
}
bail:
status = err ? 0 : p_blkno;
mlog_exit((int)status);
return status;
}
/*
* TODO: this should probably be merged into ocfs2_get_block
*
* However, you now need to pay attention to the cont_prepare_write()
* stuff in ocfs2_get_block (that is, ocfs2_get_block pretty much
* expects never to extend).
*/
struct buffer_head *ocfs2_bread(struct inode *inode,
int block, int *err, int reada)
{
struct buffer_head *bh = NULL;
int tmperr;
u64 p_blkno;
int readflags = OCFS2_BH_CACHED;
if (reada)
readflags |= OCFS2_BH_READAHEAD;
if (((u64)block << inode->i_sb->s_blocksize_bits) >=
i_size_read(inode)) {
BUG_ON(!reada);
return NULL;
}
down_read(&OCFS2_I(inode)->ip_alloc_sem);
tmperr = ocfs2_extent_map_get_blocks(inode, block, &p_blkno, NULL,
NULL);
up_read(&OCFS2_I(inode)->ip_alloc_sem);
if (tmperr < 0) {
mlog_errno(tmperr);
goto fail;
}
tmperr = ocfs2_read_block(OCFS2_SB(inode->i_sb), p_blkno, &bh,
readflags, inode);
if (tmperr < 0)
goto fail;
tmperr = 0;
*err = 0;
return bh;
fail:
if (bh) {
brelse(bh);
bh = NULL;
}
*err = -EIO;
return NULL;
}
static sector_t ocfs2_bmap(struct address_space *mapping, sector_t block)
{
sector_t status;
u64 p_blkno = 0;
int err = 0;
struct inode *inode = mapping->host;
trace_ocfs2_bmap((unsigned long long)OCFS2_I(inode)->ip_blkno,
(unsigned long long)block);
if (!INODE_JOURNAL(inode)) {
err = ocfs2_inode_lock(inode, NULL, 0);
if (err) {
if (err != -ENOENT)
mlog_errno(err);
goto bail;
}
down_read(&OCFS2_I(inode)->ip_alloc_sem);
}
if (!(OCFS2_I(inode)->ip_dyn_features & OCFS2_INLINE_DATA_FL))
err = ocfs2_extent_map_get_blocks(inode, block, &p_blkno, NULL,
NULL);
if (!INODE_JOURNAL(inode)) {
up_read(&OCFS2_I(inode)->ip_alloc_sem);
ocfs2_inode_unlock(inode, 0);
}
if (err) {
mlog(ML_ERROR, "get_blocks() failed, block = %llu\n",
(unsigned long long)block);
mlog_errno(err);
goto bail;
}
bail:
status = err ? 0 : p_blkno;
return status;
}
static int ocfs2_direct_IO_get_blocks(struct inode *inode, sector_t iblock,
struct buffer_head *bh_result, int create)
{
int ret;
u64 p_blkno, inode_blocks, contig_blocks;
unsigned int ext_flags;
unsigned char blocksize_bits = inode->i_sb->s_blocksize_bits;
unsigned long max_blocks = bh_result->b_size >> inode->i_blkbits;
inode_blocks = ocfs2_blocks_for_bytes(inode->i_sb, i_size_read(inode));
ret = ocfs2_extent_map_get_blocks(inode, iblock, &p_blkno,
&contig_blocks, &ext_flags);
if (ret) {
mlog(ML_ERROR, "get_blocks() failed iblock=%llu\n",
(unsigned long long)iblock);
ret = -EIO;
goto bail;
}
BUG_ON(create && (ext_flags & OCFS2_EXT_REFCOUNTED));
/*
* get_more_blocks() expects us to describe a hole by clearing
* the mapped bit on bh_result().
*
* Consider an unwritten extent as a hole.
*/
if (p_blkno && !(ext_flags & OCFS2_EXT_UNWRITTEN))
map_bh(bh_result, inode->i_sb, p_blkno);
else
clear_buffer_mapped(bh_result);
if (max_blocks < contig_blocks)
contig_blocks = max_blocks;
bh_result->b_size = contig_blocks << blocksize_bits;
bail:
return ret;
}
/*
* TODO: Make this into a generic get_blocks function.
*
* From do_direct_io in direct-io.c:
* "So what we do is to permit the ->get_blocks function to populate
* bh.b_size with the size of IO which is permitted at this offset and
* this i_blkbits."
*
* This function is called directly from get_more_blocks in direct-io.c.
*
* called like this: dio->get_blocks(dio->inode, fs_startblk,
* fs_count, map_bh, dio->rw == WRITE);
*/
static int ocfs2_direct_IO_get_blocks(struct inode *inode, sector_t iblock,
struct buffer_head *bh_result, int create)
{
int ret;
u64 p_blkno, inode_blocks, contig_blocks;
unsigned int ext_flags;
unsigned char blocksize_bits = inode->i_sb->s_blocksize_bits;
unsigned long max_blocks = bh_result->b_size >> inode->i_blkbits;
/* This function won't even be called if the request isn't all
* nicely aligned and of the right size, so there's no need
* for us to check any of that. */
inode_blocks = ocfs2_blocks_for_bytes(inode->i_sb, i_size_read(inode));
/*
* Any write past EOF is not allowed because we'd be extending.
*/
if (create && (iblock + max_blocks) > inode_blocks) {
ret = -EIO;
goto bail;
}
/* This figures out the size of the next contiguous block, and
* our logical offset */
ret = ocfs2_extent_map_get_blocks(inode, iblock, &p_blkno,
&contig_blocks, &ext_flags);
if (ret) {
mlog(ML_ERROR, "get_blocks() failed iblock=%llu\n",
(unsigned long long)iblock);
ret = -EIO;
goto bail;
}
if (!ocfs2_sparse_alloc(OCFS2_SB(inode->i_sb)) && !p_blkno && create) {
ocfs2_error(inode->i_sb,
"Inode %llu has a hole at block %llu\n",
(unsigned long long)OCFS2_I(inode)->ip_blkno,
(unsigned long long)iblock);
ret = -EROFS;
goto bail;
}
/*
* get_more_blocks() expects us to describe a hole by clearing
* the mapped bit on bh_result().
*
* Consider an unwritten extent as a hole.
*/
if (p_blkno && !(ext_flags & OCFS2_EXT_UNWRITTEN))
map_bh(bh_result, inode->i_sb, p_blkno);
else {
/*
* ocfs2_prepare_inode_for_write() should have caught
* the case where we'd be filling a hole and triggered
* a buffered write instead.
*/
if (create) {
ret = -EIO;
mlog_errno(ret);
goto bail;
}
clear_buffer_mapped(bh_result);
}
/* make sure we don't map more than max_blocks blocks here as
that's all the kernel will handle at this point. */
if (max_blocks < contig_blocks)
contig_blocks = max_blocks;
bh_result->b_size = contig_blocks << blocksize_bits;
bail:
return ret;
}
static int ocfs2_get_block(struct inode *inode, sector_t iblock,
struct buffer_head *bh_result, int create)
{
int err = 0;
unsigned int ext_flags;
u64 max_blocks = bh_result->b_size >> inode->i_blkbits;
u64 p_blkno, count, past_eof;
struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
mlog_entry("(0x%p, %llu, 0x%p, %d)\n", inode,
(unsigned long long)iblock, bh_result, create);
if (OCFS2_I(inode)->ip_flags & OCFS2_INODE_SYSTEM_FILE)
mlog(ML_NOTICE, "get_block on system inode 0x%p (%lu)\n",
inode, inode->i_ino);
if (S_ISLNK(inode->i_mode)) {
/* this always does I/O for some reason. */
err = ocfs2_symlink_get_block(inode, iblock, bh_result, create);
goto bail;
}
err = ocfs2_extent_map_get_blocks(inode, iblock, &p_blkno, &count,
&ext_flags);
if (err) {
mlog(ML_ERROR, "Error %d from get_blocks(0x%p, %llu, 1, "
"%llu, NULL)\n", err, inode, (unsigned long long)iblock,
(unsigned long long)p_blkno);
goto bail;
}
if (max_blocks < count)
count = max_blocks;
/*
* ocfs2 never allocates in this function - the only time we
* need to use BH_New is when we're extending i_size on a file
* system which doesn't support holes, in which case BH_New
* allows block_prepare_write() to zero.
*
* If we see this on a sparse file system, then a truncate has
* raced us and removed the cluster. In this case, we clear
* the buffers dirty and uptodate bits and let the buffer code
* ignore it as a hole.
*/
if (create && p_blkno == 0 && ocfs2_sparse_alloc(osb)) {
clear_buffer_dirty(bh_result);
clear_buffer_uptodate(bh_result);
goto bail;
}
/* Treat the unwritten extent as a hole for zeroing purposes. */
if (p_blkno && !(ext_flags & OCFS2_EXT_UNWRITTEN))
map_bh(bh_result, inode->i_sb, p_blkno);
bh_result->b_size = count << inode->i_blkbits;
if (!ocfs2_sparse_alloc(osb)) {
if (p_blkno == 0) {
err = -EIO;
mlog(ML_ERROR,
"iblock = %llu p_blkno = %llu blkno=(%llu)\n",
(unsigned long long)iblock,
(unsigned long long)p_blkno,
(unsigned long long)OCFS2_I(inode)->ip_blkno);
mlog(ML_ERROR, "Size %llu, clusters %u\n", (unsigned long long)i_size_read(inode), OCFS2_I(inode)->ip_clusters);
dump_stack();
}
past_eof = ocfs2_blocks_for_bytes(inode->i_sb, i_size_read(inode));
mlog(0, "Inode %lu, past_eof = %llu\n", inode->i_ino,
(unsigned long long)past_eof);
if (create && (iblock >= past_eof))
set_buffer_new(bh_result);
}
bail:
if (err < 0)
err = -EIO;
mlog_exit(err);
return err;
}
int main(int argc, char *argv[])
{
errcode_t ret;
uint64_t blkno, contig, blkoff = 0;
uint16_t ext_flags;
int count = 0;
int c, op = 0;
char *filename;
ocfs2_filesys *fs;
ocfs2_cached_inode *cinode;
blkno = OCFS2_SUPER_BLOCK_BLKNO;
initialize_ocfs_error_table();
while ((c = getopt(argc, argv, "i:b:")) != EOF) {
switch (c) {
case 'i':
blkno = read_number(optarg);
if (blkno <= OCFS2_SUPER_BLOCK_BLKNO) {
fprintf(stderr,
"Invalid inode block: %s\n",
optarg);
print_usage();
return 1;
}
break;
case 'b':
if (op) {
fprintf(stderr, "Cannot specify more than one operation\n");
print_usage();
return 1;
}
if (read_b_numbers(optarg,
&blkoff, &count)) {
fprintf(stderr, "Invalid block range: %s\n", optarg);
print_usage();
return 1;
}
op = OP_LOOKUP_BLOCK;
break;
default:
print_usage();
return 1;
break;
}
}
if (!op) {
fprintf(stderr, "Missing operation\n");
print_usage();
return 1;
}
if (optind >= argc) {
fprintf(stderr, "Missing filename\n");
print_usage();
return 1;
}
filename = argv[optind];
ret = ocfs2_open(filename, OCFS2_FLAG_RO, 0, 0, &fs);
if (ret) {
com_err(argv[0], ret,
"while opening file \"%s\"", filename);
goto out;
}
ret = ocfs2_read_cached_inode(fs, blkno, &cinode);
if (ret) {
com_err(argv[0], ret, "while reading inode %"PRIu64, blkno);
goto out_close;
}
fprintf(stdout, "OCFS2 inode %"PRIu64" on \"%s\" has depth %"PRId16"\n",
blkno, filename,
cinode->ci_inode->id2.i_list.l_tree_depth);
ret = ocfs2_extent_map_get_blocks(cinode,
blkoff,
count,
&blkno,
&contig,
&ext_flags);
if (ret) {
com_err(argv[0], ret,
"looking up block range %"PRIu64":%d", blkoff, count);
goto out_free;
}
fprintf(stdout, "Lookup of block range %"PRIu64":%d returned %"PRIu64":%"PRIu64"\n",
blkoff, count, blkno, contig);
out_free:
ocfs2_free_cached_inode(fs, cinode);
out_close:
ret = ocfs2_close(fs);
if (ret) {
com_err(argv[0], ret,
"while closing file \"%s\"", filename);
}
out:
return 0;
}
errcode_t ocfs2_allocate_unwritten_extents(ocfs2_filesys *fs, uint64_t ino,
uint64_t offset, uint64_t len)
{
errcode_t ret = 0;
uint32_t n_clusters = 0, cpos;
uint64_t p_blkno = 0, v_blkno, v_end, contig_blocks, wanted_blocks;
ocfs2_cached_inode *ci = NULL;
if (!(fs->fs_flags & OCFS2_FLAG_RW))
return OCFS2_ET_RO_FILESYS;
if (!ocfs2_writes_unwritten_extents(OCFS2_RAW_SB(fs->fs_super)))
return OCFS2_ET_RO_UNSUPP_FEATURE;
ret = ocfs2_read_cached_inode(fs, ino, &ci);
if (ret)
goto out;
if (!(ci->ci_inode->i_flags & OCFS2_VALID_FL))
return OCFS2_ET_INODE_NOT_VALID;
if (ci->ci_inode->i_flags & OCFS2_SYSTEM_FL)
return OCFS2_ET_INVALID_ARGUMENT;
if (!S_ISREG(ci->ci_inode->i_mode))
return OCFS2_ET_INVALID_ARGUMENT;
v_blkno = offset / fs->fs_blocksize;
v_end = (offset + len - 1) / fs->fs_blocksize;
while (v_blkno <= v_end) {
ret = ocfs2_extent_map_get_blocks(ci, v_blkno, 1,
&p_blkno, &contig_blocks,
NULL);
if (ret)
continue;
if (p_blkno) {
v_blkno += contig_blocks;
continue;
}
/*
* There is a hole, so we have to allocate the space and
* insert the unwritten extents.
*/
wanted_blocks = ocfs2_min(contig_blocks, v_end - v_blkno + 1);
n_clusters = ocfs2_clusters_in_blocks(fs, wanted_blocks);
ret = ocfs2_new_clusters(fs, 1, n_clusters, &p_blkno,
&n_clusters);
if (ret || n_clusters == 0)
break;
cpos = ocfs2_blocks_to_clusters(fs, v_blkno);
ret = ocfs2_cached_inode_insert_extent(ci, cpos,
p_blkno, n_clusters,
OCFS2_EXT_UNWRITTEN);
if (ret) {
/*
* XXX: We don't wan't to overwrite the error
* from insert_extent(). But we probably need
* to BE LOUDLY UPSET.
*/
ocfs2_free_clusters(fs, n_clusters, p_blkno);
goto out;
}
/* save up what we have done. */
ret = ocfs2_write_cached_inode(fs, ci);
if (ret)
goto out;
v_blkno = ocfs2_clusters_to_blocks(fs, cpos + n_clusters);
}
if (ci->ci_inode->i_size <= offset + len) {
ci->ci_inode->i_size = offset + len;
ret = ocfs2_write_cached_inode(fs, ci);
}
out:
if (ci)
ocfs2_free_cached_inode(fs, ci);
return ret;
}
/*
* Zero the area past i_size but still within an allocated
* cluster. This avoids exposing nonzero data on subsequent file
* extends.
*/
static errcode_t ocfs2_zero_tail_for_truncate(ocfs2_cached_inode *ci,
uint64_t new_size)
{
errcode_t ret;
char *buf = NULL;
ocfs2_filesys *fs = ci->ci_fs;
uint64_t start_blk, p_blkno, contig_blocks, start_off;
int count, byte_counts, bpc = fs->fs_clustersize /fs->fs_blocksize;
uint16_t ext_flags;
if (new_size == 0)
return 0;
start_blk = new_size / fs->fs_blocksize;
ret = ocfs2_extent_map_get_blocks(ci, start_blk, 1,
&p_blkno, &contig_blocks, &ext_flags);
if (ret)
goto out;
/* Tail is a hole. */
if (!p_blkno)
goto out;
if (ext_flags & OCFS2_EXT_REFCOUNTED) {
uint32_t cpos = ocfs2_blocks_to_clusters(fs, start_blk);
ret = ocfs2_refcount_cow(ci, cpos, 1, cpos + 1);
if (ret)
goto out;
ret = ocfs2_extent_map_get_blocks(ci, start_blk, 1,
&p_blkno, &contig_blocks,
&ext_flags);
if (ret)
goto out;
assert(!(ext_flags & OCFS2_EXT_REFCOUNTED) && p_blkno);
}
/* calculate the total blocks we need to empty. */
count = bpc - (p_blkno & (bpc - 1));
ret = ocfs2_malloc_blocks(fs->fs_io, count, &buf);
if (ret)
goto out;
ret = ocfs2_read_blocks(fs, p_blkno, count, buf);
if (ret)
goto out;
/* empty the content after the new_size and within the same cluster. */
start_off = new_size % fs->fs_blocksize;
byte_counts = count * fs->fs_blocksize - start_off;
memset(buf + start_off, 0, byte_counts);
ret = io_write_block(fs->fs_io, p_blkno, count, buf);
out:
if (buf)
ocfs2_free(&buf);
return ret;
}
/*
* We only need to replay the journals if the inode's flag is set and s_start
* indicates that there is actually pending data in the journals.
*
* In the simple case of an unclean shutdown we don't want to have to build up
* enough state to be able to truncate the inodes waiting in the orphan dir.
* ocfs2 in the kernel only fixes up the orphan dirs if the journal dirty flag
* is set. So after replaying the journals we clear s_startin the journals to
* stop a second journal replay but leave the dirty bit set so that the kernel
* will truncate the orphaned inodes.
*/
errcode_t o2fsck_should_replay_journals(ocfs2_filesys *fs, int *should,
int *has_dirty)
{
uint16_t i, max_slots;
char *buf = NULL;
uint64_t blkno;
errcode_t ret;
ocfs2_cached_inode *cinode = NULL;
int is_dirty;
uint64_t contig;
journal_superblock_t *jsb;
*should = 0;
max_slots = OCFS2_RAW_SB(fs->fs_super)->s_max_slots;
ret = ocfs2_malloc_block(fs->fs_io, &buf);
if (ret) {
com_err(whoami, ret, "while allocating room to read journal "
"blocks");
goto out;
}
jsb = (journal_superblock_t *)buf;
for (i = 0; i < max_slots; i++) {
ret = ocfs2_lookup_system_inode(fs, JOURNAL_SYSTEM_INODE, i,
&blkno);
if (ret) {
com_err(whoami, ret, "while looking up the journal "
"inode for slot %d", i);
goto out;
}
if (cinode) {
ocfs2_free_cached_inode(fs, cinode);
cinode = NULL;
}
ret = ocfs2_read_cached_inode(fs, blkno, &cinode);
if (ret) {
com_err(whoami, ret, "while reading cached inode "
"%"PRIu64" for slot %d's journal", blkno, i);
goto out;
}
is_dirty = cinode->ci_inode->id1.journal1.ij_flags &
OCFS2_JOURNAL_DIRTY_FL;
verbosef("slot %d JOURNAL_DIRTY_FL: %d\n", i, is_dirty);
if (!is_dirty)
continue;
else
*has_dirty = 1;
ret = ocfs2_extent_map_get_blocks(cinode, 0, 1, &blkno,
&contig, NULL);
if (ret) {
com_err(whoami, ret, "while looking up the journal "
"super block in slot %d's journal", i);
goto out;
}
/* XXX be smarter about reading in the whole super block if it
* spans multiple blocks */
ret = ocfs2_read_journal_superblock(fs, blkno, buf);
if (ret) {
com_err(whoami, ret, "while reading the journal "
"super block in slot %d's journal", i);
goto out;
}
if (jsb->s_start)
*should = 1;
}
out:
if (buf)
ocfs2_free(&buf);
if (cinode)
ocfs2_free_cached_inode(fs, cinode);
return ret;
}
errcode_t ocfs2_file_read(ocfs2_cached_inode *ci, void *buf, uint32_t count,
uint64_t offset, uint32_t *got)
{
ocfs2_filesys *fs = ci->ci_fs;
errcode_t ret = 0;
char *ptr = (char *) buf;
uint32_t wanted_blocks;
uint64_t contig_blocks;
uint64_t v_blkno;
uint64_t p_blkno;
uint32_t tmp;
uint64_t num_blocks;
uint16_t extent_flags;
if (ci->ci_inode->i_dyn_features & OCFS2_INLINE_DATA_FL)
return ocfs2_inline_data_read(ci->ci_inode, buf, count,
offset, got);
/* o_direct requires aligned io */
tmp = fs->fs_blocksize - 1;
if ((count & tmp) || (offset & (uint64_t)tmp) ||
((unsigned long)ptr & tmp))
return OCFS2_ET_INVALID_ARGUMENT;
wanted_blocks = count >> OCFS2_RAW_SB(fs->fs_super)->s_blocksize_bits;
v_blkno = offset >> OCFS2_RAW_SB(fs->fs_super)->s_blocksize_bits;
*got = 0;
num_blocks = (ci->ci_inode->i_size + fs->fs_blocksize - 1) >>
OCFS2_RAW_SB(fs->fs_super)->s_blocksize_bits;
if (v_blkno >= num_blocks)
return 0;
if (v_blkno + wanted_blocks > num_blocks)
wanted_blocks = (uint32_t) (num_blocks - v_blkno);
while(wanted_blocks) {
ret = ocfs2_extent_map_get_blocks(ci, v_blkno, 1,
&p_blkno, &contig_blocks,
&extent_flags);
if (ret)
return ret;
if (contig_blocks > wanted_blocks)
contig_blocks = wanted_blocks;
if (!p_blkno || extent_flags & OCFS2_EXT_UNWRITTEN) {
/*
* we meet with a hole or an unwritten extent,
* so just empty the content.
*/
memset(ptr, 0, contig_blocks * fs->fs_blocksize);
} else {
ret = ocfs2_read_blocks(fs, p_blkno, contig_blocks,
ptr);
if (ret)
return ret;
}
*got += (contig_blocks <<
OCFS2_RAW_SB(fs->fs_super)->s_blocksize_bits);
wanted_blocks -= contig_blocks;
if (wanted_blocks) {
ptr += (contig_blocks <<
OCFS2_RAW_SB(fs->fs_super)->s_blocksize_bits);
v_blkno += (uint64_t)contig_blocks;
} else {
if (*got + offset > ci->ci_inode->i_size)
*got = (uint32_t) (ci->ci_inode->i_size - offset);
/* break */
}
}
return ret;
}
static errcode_t ocfs2_file_block_write(ocfs2_cached_inode *ci,
void *buf, uint32_t count,
uint64_t offset, uint32_t *wrote)
{
ocfs2_filesys *fs = ci->ci_fs;
errcode_t ret = 0;
char *ptr = (char *) buf;
uint32_t wanted_blocks;
uint64_t contig_blocks;
uint64_t v_blkno;
uint64_t p_blkno, p_start, p_end;
uint64_t begin_blocks = 0, end_blocks = 0;
uint32_t tmp;
uint64_t num_blocks;
int bs_bits = OCFS2_RAW_SB(fs->fs_super)->s_blocksize_bits;
uint64_t ino = ci->ci_blkno;
uint32_t n_clusters, cluster_begin, cluster_end;
uint64_t bpc = fs->fs_clustersize/fs->fs_blocksize;
int insert = 0;
uint16_t extent_flags = 0;
/* o_direct requires aligned io */
tmp = fs->fs_blocksize - 1;
if ((count & tmp) || (offset & (uint64_t)tmp) ||
((unsigned long)ptr & tmp))
return OCFS2_ET_INVALID_ARGUMENT;
wanted_blocks = count >> bs_bits;
v_blkno = offset >> bs_bits;
*wrote = 0;
num_blocks = (ci->ci_inode->i_size + fs->fs_blocksize - 1) >> bs_bits;
if (v_blkno >= num_blocks)
return 0;
if (v_blkno + wanted_blocks > num_blocks)
wanted_blocks = (uint32_t) (num_blocks - v_blkno);
while(wanted_blocks) {
ret = ocfs2_extent_map_get_blocks(ci, v_blkno, 1,
&p_blkno, &contig_blocks,
&extent_flags);
if (ret)
return ret;
if (contig_blocks > wanted_blocks)
contig_blocks = wanted_blocks;
begin_blocks = 0;
end_blocks = 0;
p_end = 0;
if (!p_blkno) {
/*
* We meet with a hole here, so we allocate clusters
* and empty the both ends in case.
*
* We will postpone the extent insertion after we
* successfully write the extent block, so that and
* problems happens in block writing would not affect
* the file.
*/
cluster_begin = ocfs2_blocks_to_clusters(fs, v_blkno);
cluster_end = ocfs2_blocks_to_clusters(fs,
v_blkno + contig_blocks -1);
n_clusters = cluster_end - cluster_begin + 1;
ret = ocfs2_new_clusters(fs, 1, n_clusters, &p_start,
&n_clusters);
if (ret || n_clusters == 0)
return ret;
begin_blocks = v_blkno & (bpc - 1);
p_blkno = p_start + begin_blocks;
contig_blocks = n_clusters * bpc - begin_blocks;
if (contig_blocks > wanted_blocks) {
end_blocks = contig_blocks - wanted_blocks;
contig_blocks = wanted_blocks;
p_end = p_blkno + wanted_blocks;
}
insert = 1;
} else if (extent_flags & OCFS2_EXT_UNWRITTEN) {
begin_blocks = v_blkno & (bpc - 1);
p_start = p_blkno - begin_blocks;
p_end = p_blkno + wanted_blocks;
end_blocks = (p_end & (bpc - 1)) ?
bpc - (p_end & (bpc - 1 )) : 0;
}
if (begin_blocks) {
/*
* The user don't write the first blocks,
* so we have to empty them.
*/
ret = empty_blocks(fs, p_start, begin_blocks);
if (ret)
return ret;
}
if (end_blocks) {
/*
* we don't need to write that many blocks,
* so empty the blocks at the bottom.
*/
ret = empty_blocks(fs, p_end, end_blocks);
if (ret)
return ret;
}
ret = io_write_block(fs->fs_io, p_blkno, contig_blocks, ptr);
if (ret)
return ret;
if (insert) {
ret = ocfs2_cached_inode_insert_extent(ci,
ocfs2_blocks_to_clusters(fs,v_blkno),
p_start, n_clusters, 0);
if (ret) {
/*
* XXX: We don't wan't to overwrite the error
* from insert_extent(). But we probably need
* to BE LOUDLY UPSET.
*/
ocfs2_free_clusters(fs, n_clusters, p_start);
return ret;
}
/* save up what we have done. */
ret = ocfs2_write_cached_inode(fs, ci);
if (ret)
return ret;
ret = ocfs2_extent_map_get_blocks(ci, v_blkno, 1,
&p_blkno, NULL, NULL);
/* now we shouldn't find a hole. */
if (!p_blkno || p_blkno != p_start + begin_blocks)
ret = OCFS2_ET_INTERNAL_FAILURE;
if (ret)
return ret;
insert = 0;
} else if (extent_flags & OCFS2_EXT_UNWRITTEN) {
cluster_begin = ocfs2_blocks_to_clusters(fs, v_blkno);
cluster_end = ocfs2_blocks_to_clusters(fs,
v_blkno + contig_blocks -1);
n_clusters = cluster_end - cluster_begin + 1;
ret = ocfs2_mark_extent_written(fs, ci->ci_inode,
cluster_begin, n_clusters,
p_blkno & ~(bpc - 1));
if (ret)
return ret;
ocfs2_free_cached_inode(fs, ci);
ocfs2_read_cached_inode(fs,ino, &ci);
}
*wrote += (contig_blocks << bs_bits);
wanted_blocks -= contig_blocks;
if (wanted_blocks) {
ptr += (contig_blocks << bs_bits);
v_blkno += (uint64_t)contig_blocks;
} else {
if (*wrote + offset > ci->ci_inode->i_size)
*wrote = (uint32_t) (ci->ci_inode->i_size - offset);
/* break */
}
}
return ret;
}
static int ocfs2_map_slot_buffers(struct ocfs2_super *osb,
struct ocfs2_slot_info *si)
{
int status = 0;
u64 blkno;
unsigned long long blocks, bytes = 0;
unsigned int i;
struct buffer_head *bh;
status = ocfs2_slot_map_physical_size(osb, si->si_inode, &bytes);
if (status)
goto bail;
blocks = ocfs2_blocks_for_bytes(si->si_inode->i_sb, bytes);
BUG_ON(blocks > UINT_MAX);
si->si_blocks = blocks;
if (!si->si_blocks)
goto bail;
if (si->si_extended)
si->si_slots_per_block =
(osb->sb->s_blocksize /
sizeof(struct ocfs2_extended_slot));
else
si->si_slots_per_block = osb->sb->s_blocksize / sizeof(__le16);
/* The size checks above should ensure this */
BUG_ON((osb->max_slots / si->si_slots_per_block) > blocks);
trace_ocfs2_map_slot_buffers(bytes, si->si_blocks);
si->si_bh = kcalloc(si->si_blocks, sizeof(struct buffer_head *),
GFP_KERNEL);
if (!si->si_bh) {
status = -ENOMEM;
mlog_errno(status);
goto bail;
}
for (i = 0; i < si->si_blocks; i++) {
status = ocfs2_extent_map_get_blocks(si->si_inode, i,
&blkno, NULL, NULL);
if (status < 0) {
mlog_errno(status);
goto bail;
}
trace_ocfs2_map_slot_buffers_block((unsigned long long)blkno, i);
bh = NULL; /* Acquire a fresh bh */
status = ocfs2_read_blocks(INODE_CACHE(si->si_inode), blkno,
1, &bh, OCFS2_BH_IGNORE_CACHE, NULL);
if (status < 0) {
mlog_errno(status);
goto bail;
}
si->si_bh[i] = bh;
}
bail:
return status;
}
int ocfs2_get_block(struct inode *inode, sector_t iblock,
struct buffer_head *bh_result, int create)
{
int err = 0;
unsigned int ext_flags;
u64 max_blocks = bh_result->b_size >> inode->i_blkbits;
u64 p_blkno, count, past_eof;
struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
trace_ocfs2_get_block((unsigned long long)OCFS2_I(inode)->ip_blkno,
(unsigned long long)iblock, bh_result, create);
if (OCFS2_I(inode)->ip_flags & OCFS2_INODE_SYSTEM_FILE)
mlog(ML_NOTICE, "get_block on system inode 0x%p (%lu)\n",
inode, inode->i_ino);
if (S_ISLNK(inode->i_mode)) {
err = ocfs2_symlink_get_block(inode, iblock, bh_result, create);
goto bail;
}
err = ocfs2_extent_map_get_blocks(inode, iblock, &p_blkno, &count,
&ext_flags);
if (err) {
mlog(ML_ERROR, "Error %d from get_blocks(0x%p, %llu, 1, "
"%llu, NULL)\n", err, inode, (unsigned long long)iblock,
(unsigned long long)p_blkno);
goto bail;
}
if (max_blocks < count)
count = max_blocks;
if (create && p_blkno == 0 && ocfs2_sparse_alloc(osb)) {
clear_buffer_dirty(bh_result);
clear_buffer_uptodate(bh_result);
goto bail;
}
/* Treat the unwritten extent as a hole for zeroing purposes. */
if (p_blkno && !(ext_flags & OCFS2_EXT_UNWRITTEN))
map_bh(bh_result, inode->i_sb, p_blkno);
bh_result->b_size = count << inode->i_blkbits;
if (!ocfs2_sparse_alloc(osb)) {
if (p_blkno == 0) {
err = -EIO;
mlog(ML_ERROR,
"iblock = %llu p_blkno = %llu blkno=(%llu)\n",
(unsigned long long)iblock,
(unsigned long long)p_blkno,
(unsigned long long)OCFS2_I(inode)->ip_blkno);
mlog(ML_ERROR, "Size %llu, clusters %u\n", (unsigned long long)i_size_read(inode), OCFS2_I(inode)->ip_clusters);
dump_stack();
goto bail;
}
}
past_eof = ocfs2_blocks_for_bytes(inode->i_sb, i_size_read(inode));
trace_ocfs2_get_block_end((unsigned long long)OCFS2_I(inode)->ip_blkno,
(unsigned long long)past_eof);
if (create && (iblock >= past_eof))
set_buffer_new(bh_result);
bail:
if (err < 0)
err = -EIO;
return err;
}