#include <linux/splice.h>
#include <linux/mount.h>
#include <linux/writeback.h>
+#include <linux/falloc.h>
#define MLOG_MASK_PREFIX ML_INODE
#include <cluster/masklog.h>
int ret;
struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
handle_t *handle;
+ struct ocfs2_dinode *di = (struct ocfs2_dinode *) bh->b_data;
mlog_entry_void();
goto out;
}
+ ret = ocfs2_journal_access(handle, inode, bh,
+ OCFS2_JOURNAL_ACCESS_WRITE);
+ if (ret) {
+ mlog_errno(ret);
+ goto out_commit;
+ }
+
+ /*
+ * Don't use ocfs2_mark_inode_dirty() here as we don't always
+ * have i_mutex to guard against concurrent changes to other
+ * inode fields.
+ */
inode->i_atime = CURRENT_TIME;
- ret = ocfs2_mark_inode_dirty(handle, inode, bh);
+ di->i_atime = cpu_to_le64(inode->i_atime.tv_sec);
+ di->i_atime_nsec = cpu_to_le32(inode->i_atime.tv_nsec);
+
+ ret = ocfs2_journal_dirty(handle, bh);
if (ret < 0)
mlog_errno(ret);
+out_commit:
ocfs2_commit_trans(OCFS2_SB(inode->i_sb), handle);
out:
mlog_exit(ret);
struct ocfs2_alloc_context **data_ac,
struct ocfs2_alloc_context **meta_ac)
{
- int ret, num_free_extents;
+ int ret = 0, num_free_extents;
unsigned int max_recs_needed = clusters_to_add + 2 * extents_to_split;
struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
*meta_ac = NULL;
- *data_ac = NULL;
+ if (data_ac)
+ *data_ac = NULL;
+
+ BUG_ON(clusters_to_add != 0 && data_ac == NULL);
mlog(0, "extend inode %llu, i_size = %lld, di->i_clusters = %u, "
"clusters_to_add = %u, extents_to_split = %u\n",
}
}
+ if (clusters_to_add == 0)
+ goto out;
+
ret = ocfs2_reserve_clusters(osb, clusters_to_add, data_ac);
if (ret < 0) {
if (ret != -ENOSPC)
}
if (size_change && attr->ia_size != i_size_read(inode)) {
+ if (attr->ia_size > sb->s_maxbytes) {
+ status = -EFBIG;
+ goto bail_unlock;
+ }
+
if (i_size_read(inode) > attr->ia_size)
status = ocfs2_truncate_file(inode, bh, attr->ia_size);
else
return ret;
}
-static int ocfs2_write_remove_suid(struct inode *inode)
+static int __ocfs2_write_remove_suid(struct inode *inode,
+ struct buffer_head *bh)
{
int ret;
- struct buffer_head *bh = NULL;
- struct ocfs2_inode_info *oi = OCFS2_I(inode);
handle_t *handle;
struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
struct ocfs2_dinode *di;
mlog_entry("(Inode %llu, mode 0%o)\n",
- (unsigned long long)oi->ip_blkno, inode->i_mode);
+ (unsigned long long)OCFS2_I(inode)->ip_blkno, inode->i_mode);
handle = ocfs2_start_trans(osb, OCFS2_INODE_UPDATE_CREDITS);
if (handle == NULL) {
goto out;
}
- ret = ocfs2_read_block(osb, oi->ip_blkno, &bh, OCFS2_BH_CACHED, inode);
- if (ret < 0) {
- mlog_errno(ret);
- goto out_trans;
- }
-
ret = ocfs2_journal_access(handle, inode, bh,
OCFS2_JOURNAL_ACCESS_WRITE);
if (ret < 0) {
mlog_errno(ret);
- goto out_bh;
+ goto out_trans;
}
inode->i_mode &= ~S_ISUID;
ret = ocfs2_journal_dirty(handle, bh);
if (ret < 0)
mlog_errno(ret);
-out_bh:
- brelse(bh);
+
out_trans:
ocfs2_commit_trans(osb, handle);
out:
return ret;
}
+static int ocfs2_write_remove_suid(struct inode *inode)
+{
+ int ret;
+ struct buffer_head *bh = NULL;
+ struct ocfs2_inode_info *oi = OCFS2_I(inode);
+
+ ret = ocfs2_read_block(OCFS2_SB(inode->i_sb),
+ oi->ip_blkno, &bh, OCFS2_BH_CACHED, inode);
+ if (ret < 0) {
+ mlog_errno(ret);
+ goto out;
+ }
+
+ ret = __ocfs2_write_remove_suid(inode, bh);
+out:
+ brelse(bh);
+ return ret;
+}
+
/*
* Allocate enough extents to cover the region starting at byte offset
* start for len bytes. Existing extents are skipped, any extents
return ret;
}
+static int __ocfs2_remove_inode_range(struct inode *inode,
+ struct buffer_head *di_bh,
+ u32 cpos, u32 phys_cpos, u32 len,
+ struct ocfs2_cached_dealloc_ctxt *dealloc)
+{
+ int ret;
+ u64 phys_blkno = ocfs2_clusters_to_blocks(inode->i_sb, phys_cpos);
+ struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
+ struct inode *tl_inode = osb->osb_tl_inode;
+ handle_t *handle;
+ struct ocfs2_alloc_context *meta_ac = NULL;
+ struct ocfs2_dinode *di = (struct ocfs2_dinode *)di_bh->b_data;
+
+ ret = ocfs2_lock_allocators(inode, di, 0, 1, NULL, &meta_ac);
+ if (ret) {
+ mlog_errno(ret);
+ return ret;
+ }
+
+ mutex_lock(&tl_inode->i_mutex);
+
+ if (ocfs2_truncate_log_needs_flush(osb)) {
+ ret = __ocfs2_flush_truncate_log(osb);
+ if (ret < 0) {
+ mlog_errno(ret);
+ goto out;
+ }
+ }
+
+ handle = ocfs2_start_trans(osb, OCFS2_REMOVE_EXTENT_CREDITS);
+ if (handle == NULL) {
+ ret = -ENOMEM;
+ mlog_errno(ret);
+ goto out;
+ }
+
+ ret = ocfs2_journal_access(handle, inode, di_bh,
+ OCFS2_JOURNAL_ACCESS_WRITE);
+ if (ret) {
+ mlog_errno(ret);
+ goto out;
+ }
+
+ ret = ocfs2_remove_extent(inode, di_bh, cpos, len, handle, meta_ac,
+ dealloc);
+ if (ret) {
+ mlog_errno(ret);
+ goto out_commit;
+ }
+
+ OCFS2_I(inode)->ip_clusters -= len;
+ di->i_clusters = cpu_to_le32(OCFS2_I(inode)->ip_clusters);
+
+ ret = ocfs2_journal_dirty(handle, di_bh);
+ if (ret) {
+ mlog_errno(ret);
+ goto out_commit;
+ }
+
+ ret = ocfs2_truncate_log_append(osb, handle, phys_blkno, len);
+ if (ret)
+ mlog_errno(ret);
+
+out_commit:
+ ocfs2_commit_trans(osb, handle);
+out:
+ mutex_unlock(&tl_inode->i_mutex);
+
+ if (meta_ac)
+ ocfs2_free_alloc_context(meta_ac);
+
+ return ret;
+}
+
+/*
+ * Truncate a byte range, avoiding pages within partial clusters. This
+ * preserves those pages for the zeroing code to write to.
+ */
+static void ocfs2_truncate_cluster_pages(struct inode *inode, u64 byte_start,
+ u64 byte_len)
+{
+ struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
+ loff_t start, end;
+ struct address_space *mapping = inode->i_mapping;
+
+ start = (loff_t)ocfs2_align_bytes_to_clusters(inode->i_sb, byte_start);
+ end = byte_start + byte_len;
+ end = end & ~(osb->s_clustersize - 1);
+
+ if (start < end) {
+ unmap_mapping_range(mapping, start, end - start, 0);
+ truncate_inode_pages_range(mapping, start, end - 1);
+ }
+}
+
+static int ocfs2_zero_partial_clusters(struct inode *inode,
+ u64 start, u64 len)
+{
+ int ret = 0;
+ u64 tmpend, end = start + len;
+ struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
+ unsigned int csize = osb->s_clustersize;
+ handle_t *handle;
+
+ /*
+ * The "start" and "end" values are NOT necessarily part of
+ * the range whose allocation is being deleted. Rather, this
+ * is what the user passed in with the request. We must zero
+ * partial clusters here. There's no need to worry about
+ * physical allocation - the zeroing code knows to skip holes.
+ */
+ mlog(0, "byte start: %llu, end: %llu\n",
+ (unsigned long long)start, (unsigned long long)end);
+
+ /*
+ * If both edges are on a cluster boundary then there's no
+ * zeroing required as the region is part of the allocation to
+ * be truncated.
+ */
+ if ((start & (csize - 1)) == 0 && (end & (csize - 1)) == 0)
+ goto out;
+
+ handle = ocfs2_start_trans(osb, OCFS2_INODE_UPDATE_CREDITS);
+ if (handle == NULL) {
+ ret = -ENOMEM;
+ mlog_errno(ret);
+ goto out;
+ }
+
+ /*
+ * We want to get the byte offset of the end of the 1st cluster.
+ */
+ tmpend = (u64)osb->s_clustersize + (start & ~(osb->s_clustersize - 1));
+ if (tmpend > end)
+ tmpend = end;
+
+ mlog(0, "1st range: start: %llu, tmpend: %llu\n",
+ (unsigned long long)start, (unsigned long long)tmpend);
+
+ ret = ocfs2_zero_range_for_truncate(inode, handle, start, tmpend);
+ if (ret)
+ mlog_errno(ret);
+
+ if (tmpend < end) {
+ /*
+ * This may make start and end equal, but the zeroing
+ * code will skip any work in that case so there's no
+ * need to catch it up here.
+ */
+ start = end & ~(osb->s_clustersize - 1);
+
+ mlog(0, "2nd range: start: %llu, end: %llu\n",
+ (unsigned long long)start, (unsigned long long)end);
+
+ ret = ocfs2_zero_range_for_truncate(inode, handle, start, end);
+ if (ret)
+ mlog_errno(ret);
+ }
+
+ ocfs2_commit_trans(osb, handle);
+out:
+ return ret;
+}
+
+static int ocfs2_remove_inode_range(struct inode *inode,
+ struct buffer_head *di_bh, u64 byte_start,
+ u64 byte_len)
+{
+ int ret = 0;
+ u32 trunc_start, trunc_len, cpos, phys_cpos, alloc_size;
+ struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
+ struct ocfs2_cached_dealloc_ctxt dealloc;
+
+ ocfs2_init_dealloc_ctxt(&dealloc);
+
+ if (byte_len == 0)
+ return 0;
+
+ trunc_start = ocfs2_clusters_for_bytes(osb->sb, byte_start);
+ trunc_len = (byte_start + byte_len) >> osb->s_clustersize_bits;
+ if (trunc_len >= trunc_start)
+ trunc_len -= trunc_start;
+ else
+ trunc_len = 0;
+
+ mlog(0, "Inode: %llu, start: %llu, len: %llu, cstart: %u, clen: %u\n",
+ (unsigned long long)OCFS2_I(inode)->ip_blkno,
+ (unsigned long long)byte_start,
+ (unsigned long long)byte_len, trunc_start, trunc_len);
+
+ ret = ocfs2_zero_partial_clusters(inode, byte_start, byte_len);
+ if (ret) {
+ mlog_errno(ret);
+ goto out;
+ }
+
+ cpos = trunc_start;
+ while (trunc_len) {
+ ret = ocfs2_get_clusters(inode, cpos, &phys_cpos,
+ &alloc_size, NULL);
+ if (ret) {
+ mlog_errno(ret);
+ goto out;
+ }
+
+ if (alloc_size > trunc_len)
+ alloc_size = trunc_len;
+
+ /* Only do work for non-holes */
+ if (phys_cpos != 0) {
+ ret = __ocfs2_remove_inode_range(inode, di_bh, cpos,
+ phys_cpos, alloc_size,
+ &dealloc);
+ if (ret) {
+ mlog_errno(ret);
+ goto out;
+ }
+ }
+
+ cpos += alloc_size;
+ trunc_len -= alloc_size;
+ }
+
+ ocfs2_truncate_cluster_pages(inode, byte_start, byte_len);
+
+out:
+ ocfs2_schedule_truncate_log_flush(osb, 1);
+ ocfs2_run_deallocs(osb, &dealloc);
+
+ return ret;
+}
+
+/*
+ * Parts of this function taken from xfs_change_file_space()
+ */
+static int __ocfs2_change_file_space(struct file *file, struct inode *inode,
+ loff_t f_pos, unsigned int cmd,
+ struct ocfs2_space_resv *sr,
+ int change_size)
+{
+ int ret;
+ s64 llen;
+ loff_t size;
+ struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
+ struct buffer_head *di_bh = NULL;
+ handle_t *handle;
+ unsigned long long max_off = inode->i_sb->s_maxbytes;
+
+ if (ocfs2_is_hard_readonly(osb) || ocfs2_is_soft_readonly(osb))
+ return -EROFS;
+
+ mutex_lock(&inode->i_mutex);
+
+ /*
+ * This prevents concurrent writes on other nodes
+ */
+ ret = ocfs2_rw_lock(inode, 1);
+ if (ret) {
+ mlog_errno(ret);
+ goto out;
+ }
+
+ ret = ocfs2_meta_lock(inode, &di_bh, 1);
+ if (ret) {
+ mlog_errno(ret);
+ goto out_rw_unlock;
+ }
+
+ if (inode->i_flags & (S_IMMUTABLE|S_APPEND)) {
+ ret = -EPERM;
+ goto out_meta_unlock;
+ }
+
+ switch (sr->l_whence) {
+ case 0: /*SEEK_SET*/
+ break;
+ case 1: /*SEEK_CUR*/
+ sr->l_start += f_pos;
+ break;
+ case 2: /*SEEK_END*/
+ sr->l_start += i_size_read(inode);
+ break;
+ default:
+ ret = -EINVAL;
+ goto out_meta_unlock;
+ }
+ sr->l_whence = 0;
+
+ llen = sr->l_len > 0 ? sr->l_len - 1 : sr->l_len;
+
+ if (sr->l_start < 0
+ || sr->l_start > max_off
+ || (sr->l_start + llen) < 0
+ || (sr->l_start + llen) > max_off) {
+ ret = -EINVAL;
+ goto out_meta_unlock;
+ }
+ size = sr->l_start + sr->l_len;
+
+ if (cmd == OCFS2_IOC_RESVSP || cmd == OCFS2_IOC_RESVSP64) {
+ if (sr->l_len <= 0) {
+ ret = -EINVAL;
+ goto out_meta_unlock;
+ }
+ }
+
+ if (file && should_remove_suid(file->f_path.dentry)) {
+ ret = __ocfs2_write_remove_suid(inode, di_bh);
+ if (ret) {
+ mlog_errno(ret);
+ goto out_meta_unlock;
+ }
+ }
+
+ down_write(&OCFS2_I(inode)->ip_alloc_sem);
+ switch (cmd) {
+ case OCFS2_IOC_RESVSP:
+ case OCFS2_IOC_RESVSP64:
+ /*
+ * This takes unsigned offsets, but the signed ones we
+ * pass have been checked against overflow above.
+ */
+ ret = ocfs2_allocate_unwritten_extents(inode, sr->l_start,
+ sr->l_len);
+ break;
+ case OCFS2_IOC_UNRESVSP:
+ case OCFS2_IOC_UNRESVSP64:
+ ret = ocfs2_remove_inode_range(inode, di_bh, sr->l_start,
+ sr->l_len);
+ break;
+ default:
+ ret = -EINVAL;
+ }
+ up_write(&OCFS2_I(inode)->ip_alloc_sem);
+ if (ret) {
+ mlog_errno(ret);
+ goto out_meta_unlock;
+ }
+
+ /*
+ * We update c/mtime for these changes
+ */
+ handle = ocfs2_start_trans(osb, OCFS2_INODE_UPDATE_CREDITS);
+ if (IS_ERR(handle)) {
+ ret = PTR_ERR(handle);
+ mlog_errno(ret);
+ goto out_meta_unlock;
+ }
+
+ if (change_size && i_size_read(inode) < size)
+ i_size_write(inode, size);
+
+ inode->i_ctime = inode->i_mtime = CURRENT_TIME;
+ ret = ocfs2_mark_inode_dirty(handle, inode, di_bh);
+ if (ret < 0)
+ mlog_errno(ret);
+
+ ocfs2_commit_trans(osb, handle);
+
+out_meta_unlock:
+ brelse(di_bh);
+ ocfs2_meta_unlock(inode, 1);
+out_rw_unlock:
+ ocfs2_rw_unlock(inode, 1);
+
+ mutex_unlock(&inode->i_mutex);
+out:
+ return ret;
+}
+
+int ocfs2_change_file_space(struct file *file, unsigned int cmd,
+ struct ocfs2_space_resv *sr)
+{
+ struct inode *inode = file->f_path.dentry->d_inode;
+ struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);;
+
+ if ((cmd == OCFS2_IOC_RESVSP || cmd == OCFS2_IOC_RESVSP64) &&
+ !ocfs2_writes_unwritten_extents(osb))
+ return -ENOTTY;
+ else if ((cmd == OCFS2_IOC_UNRESVSP || cmd == OCFS2_IOC_UNRESVSP64) &&
+ !ocfs2_sparse_alloc(osb))
+ return -ENOTTY;
+
+ if (!S_ISREG(inode->i_mode))
+ return -EINVAL;
+
+ if (!(file->f_mode & FMODE_WRITE))
+ return -EBADF;
+
+ return __ocfs2_change_file_space(file, inode, file->f_pos, cmd, sr, 0);
+}
+
+static long ocfs2_fallocate(struct inode *inode, int mode, loff_t offset,
+ loff_t len)
+{
+ struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
+ struct ocfs2_space_resv sr;
+ int change_size = 1;
+
+ if (!ocfs2_writes_unwritten_extents(osb))
+ return -EOPNOTSUPP;
+
+ if (S_ISDIR(inode->i_mode))
+ return -ENODEV;
+
+ if (mode & FALLOC_FL_KEEP_SIZE)
+ change_size = 0;
+
+ sr.l_whence = 0;
+ sr.l_start = (s64)offset;
+ sr.l_len = (s64)len;
+
+ return __ocfs2_change_file_space(NULL, inode, offset,
+ OCFS2_IOC_RESVSP64, &sr, change_size);
+}
+
static int ocfs2_prepare_inode_for_write(struct dentry *dentry,
loff_t *ppos,
size_t count,
loff_t pos;
const struct iovec *cur_iov = iov;
struct page *user_page, *page;
- char *buf, *dst;
+ char * uninitialized_var(buf);
+ char *dst;
void *fsdata;
/*
}
dst = kmap_atomic(page, KM_USER0);
- memcpy(dst + (pos & (PAGE_CACHE_SIZE - 1)), buf, bytes);
+ memcpy(dst + (pos & (loff_t)(PAGE_CACHE_SIZE - 1)), buf, bytes);
kunmap_atomic(dst, KM_USER0);
flush_dcache_page(page);
ocfs2_put_write_source(user_page);
src = buf->ops->map(pipe, buf, 1);
dst = kmap_atomic(page, KM_USER1);
memcpy(dst + offset, src + buf->offset, count);
- kunmap_atomic(page, KM_USER1);
+ kunmap_atomic(dst, KM_USER1);
buf->ops->unmap(pipe, buf, src);
copied = ocfs2_write_end(file, file->f_mapping, sd->pos, count, count,
.setattr = ocfs2_setattr,
.getattr = ocfs2_getattr,
.permission = ocfs2_permission,
+ .fallocate = ocfs2_fallocate,
};
const struct inode_operations ocfs2_special_file_iops = {