aops.c

ocfs1.4.1 oracle分布式文件系统

			if (ext_flags & OCFS2_EXT_UNWRITTEN)
				*extents_to_split = *extents_to_split + 2;
		} else if (phys) {
			/*
			 * Only increment phys if it doesn't describe
			 * a hole.
			 */
			phys++;
		}

		desc->c_phys = phys;
		if (phys == 0) {
			desc->c_new = 1;
			*clusters_to_alloc = *clusters_to_alloc + 1;
		}
		if (ext_flags & OCFS2_EXT_UNWRITTEN)
			desc->c_unwritten = 1;

		num_clusters--;
	}

	ret = 0;
out:
	return ret;
}

static int ocfs2_write_begin_inline(struct address_space *mapping,
				    struct inode *inode,
				    struct ocfs2_write_ctxt *wc)
{
	int ret;
	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
	struct page *page;
	handle_t *handle;
	struct ocfs2_dinode *di = (struct ocfs2_dinode *)wc->w_di_bh->b_data;

	page = find_or_create_page(mapping, 0, GFP_NOFS);
	if (!page) {
		ret = -ENOMEM;
		mlog_errno(ret);
		goto out;
	}
	/*
	 * If we don't set w_num_pages then this page won't get unlocked
	 * and freed on cleanup of the write context.
	 */
	wc->w_pages[0] = wc->w_target_page = page;
	wc->w_num_pages = 1;

	handle = ocfs2_start_trans(osb, OCFS2_INODE_UPDATE_CREDITS);
	if (IS_ERR(handle)) {
		ret = PTR_ERR(handle);
		mlog_errno(ret);
		goto out;
	}

	ret = ocfs2_journal_access(handle, inode, wc->w_di_bh,
				   OCFS2_JOURNAL_ACCESS_WRITE);
	if (ret) {
		ocfs2_commit_trans(osb, handle);

		mlog_errno(ret);
		goto out;
	}

	if (!(OCFS2_I(inode)->ip_dyn_features & OCFS2_INLINE_DATA_FL))
		ocfs2_set_inode_data_inline(inode, di);

	if (!PageUptodate(page)) {
		ret = ocfs2_read_inline_data(inode, page, wc->w_di_bh);
		if (ret) {
			ocfs2_commit_trans(osb, handle);

			goto out;
		}
	}

	wc->w_handle = handle;
out:
	return ret;
}

int ocfs2_size_fits_inline_data(struct buffer_head *di_bh, u64 new_size)
{
	struct ocfs2_dinode *di = (struct ocfs2_dinode *)di_bh->b_data;

	if (new_size <= le16_to_cpu(di->id2.i_data.id_count))
		return 1;
	return 0;
}

static int ocfs2_try_to_write_inline_data(struct address_space *mapping,
					  struct inode *inode, loff_t pos,
					  unsigned len, struct page *mmap_page,
					  struct ocfs2_write_ctxt *wc)
{
	int ret, written = 0;
	loff_t end = pos + len;
	struct ocfs2_inode_info *oi = OCFS2_I(inode);

	mlog(0, "Inode %llu, write of %u bytes at off %llu. features: 0x%x\n",
	     (unsigned long long)oi->ip_blkno, len, (unsigned long long)pos,
	     oi->ip_dyn_features);

	/*
	 * Handle inodes which already have inline data 1st.
	 */
	if (oi->ip_dyn_features & OCFS2_INLINE_DATA_FL) {
		if (mmap_page == NULL &&
		    ocfs2_size_fits_inline_data(wc->w_di_bh, end))
			goto do_inline_write;

		/*
		 * The write won't fit - we have to give this inode an
		 * inline extent list now.
		 */
		ret = ocfs2_convert_inline_data_to_extents(inode, wc->w_di_bh);
		if (ret)
			mlog_errno(ret);
		goto out;
	}

	/*
	 * Check whether the inode can accept inline data.
	 */
	if (oi->ip_clusters != 0 || i_size_read(inode) != 0)
		return 0;

	/*
	 * Check whether the write can fit.
	 */
	if (mmap_page || end > ocfs2_max_inline_data(inode->i_sb))
		return 0;

do_inline_write:
	ret = ocfs2_write_begin_inline(mapping, inode, wc);
	if (ret) {
		mlog_errno(ret);
		goto out;
	}

	/*
	 * This signals to the caller that the data can be written
	 * inline.
	 */
	written = 1;
out:
	return written ? written : ret;
}

/*
 * This function only does anything for file systems which can't
 * handle sparse files.
 *
 * What we want to do here is fill in any hole between the current end
 * of allocation and the end of our write. That way the rest of the
 * write path can treat it as an non-allocating write, which has no
 * special case code for sparse/nonsparse files.
 */
static int ocfs2_expand_nonsparse_inode(struct inode *inode, loff_t pos,
					unsigned len,
					struct ocfs2_write_ctxt *wc)
{
	int ret;
	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
	loff_t newsize = pos + len;

	if (ocfs2_sparse_alloc(osb))
		return 0;

	if (newsize <= i_size_read(inode))
		return 0;

	ret = ocfs2_extend_no_holes(inode, newsize, newsize - len);
	if (ret)
		mlog_errno(ret);

	return ret;
}

int ocfs2_write_begin_nolock(struct address_space *mapping,
			     loff_t pos, unsigned len, unsigned flags,
			     struct page **pagep, void **fsdata,
			     struct buffer_head *di_bh, struct page *mmap_page)
{
	int ret, credits = OCFS2_INODE_UPDATE_CREDITS;
	unsigned int clusters_to_alloc, extents_to_split;
	struct ocfs2_write_ctxt *wc;
	struct inode *inode = mapping->host;
	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
	struct ocfs2_dinode *di;
	struct ocfs2_alloc_context *data_ac = NULL;
	struct ocfs2_alloc_context *meta_ac = NULL;
	handle_t *handle;

	ret = ocfs2_alloc_write_ctxt(&wc, osb, pos, len, di_bh);
	if (ret) {
		mlog_errno(ret);
		return ret;
	}

	if (ocfs2_supports_inline_data(osb)) {
		ret = ocfs2_try_to_write_inline_data(mapping, inode, pos, len,
						     mmap_page, wc);
		if (ret == 1) {
			ret = 0;
			goto success;
		}
		if (ret < 0) {
			mlog_errno(ret);
			goto out;
		}
	}

	ret = ocfs2_expand_nonsparse_inode(inode, pos, len, wc);
	if (ret) {
		mlog_errno(ret);
		goto out;
	}

	ret = ocfs2_populate_write_desc(inode, wc, &clusters_to_alloc,
					&extents_to_split);
	if (ret) {
		mlog_errno(ret);
		goto out;
	}

	di = (struct ocfs2_dinode *)wc->w_di_bh->b_data;

	/*
	 * We set w_target_from, w_target_to here so that
	 * ocfs2_write_end() knows which range in the target page to
	 * write out. An allocation requires that we write the entire
	 * cluster range.
	 */
	if (clusters_to_alloc || extents_to_split) {
		/*
		 * XXX: We are stretching the limits of
		 * ocfs2_lock_allocators(). It greatly over-estimates
		 * the work to be done.
		 */
		ret = ocfs2_lock_allocators(inode, di, clusters_to_alloc,
					    extents_to_split, &data_ac, &meta_ac);
		if (ret) {
			mlog_errno(ret);
			goto out;
		}

		credits = ocfs2_calc_extend_credits(inode->i_sb, di,
						    clusters_to_alloc);

	}

	ocfs2_set_target_boundaries(osb, wc, pos, len,
				    clusters_to_alloc + extents_to_split);

	handle = ocfs2_start_trans(osb, credits);
	if (IS_ERR(handle)) {
		ret = PTR_ERR(handle);
		mlog_errno(ret);
		goto out;
	}

	wc->w_handle = handle;

	/*
	 * We don't want this to fail in ocfs2_write_end(), so do it
	 * here.
	 */
	ret = ocfs2_journal_access(handle, inode, wc->w_di_bh,
				   OCFS2_JOURNAL_ACCESS_WRITE);
	if (ret) {
		mlog_errno(ret);
		goto out_commit;
	}

	/*
	 * Fill our page array first. That way we've grabbed enough so
	 * that we can zero and flush if we error after adding the
	 * extent.
	 */
	ret = ocfs2_grab_pages_for_write(mapping, wc, wc->w_cpos, pos,
					 clusters_to_alloc + extents_to_split,
					 mmap_page);
	if (ret) {
		mlog_errno(ret);
		goto out_commit;
	}

	ret = ocfs2_write_cluster_by_desc(mapping, data_ac, meta_ac, wc, pos,
					  len);
	if (ret) {
		mlog_errno(ret);
		goto out_commit;
	}

	if (data_ac)
		ocfs2_free_alloc_context(data_ac);
	if (meta_ac)
		ocfs2_free_alloc_context(meta_ac);

success:
	*pagep = wc->w_target_page;
	*fsdata = wc;
	return 0;
out_commit:
	ocfs2_commit_trans(osb, handle);

out:
	ocfs2_free_write_ctxt(wc);

	if (data_ac)
		ocfs2_free_alloc_context(data_ac);
	if (meta_ac)
		ocfs2_free_alloc_context(meta_ac);
	return ret;
}

int ocfs2_write_begin(struct file *file, struct address_space *mapping,
		      loff_t pos, unsigned len, unsigned flags,
		      struct page **pagep, void **fsdata)
{
	int ret;
	struct buffer_head *di_bh = NULL;
	struct inode *inode = mapping->host;

	ret = ocfs2_inode_lock(inode, &di_bh, 1);
	if (ret) {
		mlog_errno(ret);
		return ret;
	}

	/*
	 * Take alloc sem here to prevent concurrent lookups. That way
	 * the mapping, zeroing and tree manipulation within
	 * ocfs2_write() will be safe against ->readpage(). This
	 * should also serve to lock out allocation from a shared
	 * writeable region.
	 */
	down_write(&OCFS2_I(inode)->ip_alloc_sem);

	ret = ocfs2_write_begin_nolock(mapping, pos, len, flags, pagep,
				       fsdata, di_bh, NULL);
	if (ret) {
		mlog_errno(ret);
		goto out_fail;
	}

	brelse(di_bh);

	return 0;

out_fail:
	up_write(&OCFS2_I(inode)->ip_alloc_sem);

	brelse(di_bh);
	ocfs2_inode_unlock(inode, 1);

	return ret;
}

static void ocfs2_write_end_inline(struct inode *inode, loff_t pos,
				   unsigned len, unsigned *copied,
				   struct ocfs2_dinode *di,
				   struct ocfs2_write_ctxt *wc)
{
	void *kaddr;

	if (unlikely(*copied < len)) {
		if (!PageUptodate(wc->w_target_page)) {
			*copied = 0;
			return;
		}
	}

	kaddr = kmap_atomic(wc->w_target_page, KM_USER0);
	memcpy(di->id2.i_data.id_data + pos, kaddr + pos, *copied);
	kunmap_atomic(kaddr, KM_USER0);

	mlog(0, "Data written to inode at offset %llu. "
	     "id_count = %u, copied = %u, i_dyn_features = 0x%x\n",
	     (unsigned long long)pos, *copied,
	     le16_to_cpu(di->id2.i_data.id_count),
	     le16_to_cpu(di->i_dyn_features));
}

int ocfs2_write_end_nolock(struct address_space *mapping,
			   loff_t pos, unsigned len, unsigned copied,
			   struct page *page, void *fsdata)
{
	int i;
	unsigned from, to, start = pos & (PAGE_CACHE_SIZE - 1);
	struct inode *inode = mapping->host;
	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
	struct ocfs2_write_ctxt *wc = fsdata;
	struct ocfs2_dinode *di = (struct ocfs2_dinode *)wc->w_di_bh->b_data;
	handle_t *handle = wc->w_handle;
	struct page *tmppage;

	if (OCFS2_I(inode)->ip_dyn_features & OCFS2_INLINE_DATA_FL) {
		ocfs2_write_end_inline(inode, pos, len, &copied, di, wc);
		goto out_write_size;
	}

	if (unlikely(copied < len)) {
		if (!PageUptodate(wc->w_target_page))
			copied = 0;

		ocfs2_zero_new_buffers(wc->w_target_page, start+copied,
				       start+len);
	}
	flush_dcache_page(wc->w_target_page);

	for(i = 0; i < wc->w_num_pages; i++) {
		tmppage = wc->w_pages[i];

		if (tmppage == wc->w_target_page) {
			from = wc->w_target_from;
			to = wc->w_target_to;

			BUG_ON(from > PAGE_CACHE_SIZE ||
			       to > PAGE_CACHE_SIZE ||
			       to < from);
		} else {
			/*
			 * Pages adjacent to the target (if any) imply
			 * a hole-filling write in which case we want
			 * to flush their entire range.
			 */
			from = 0;
			to = PAGE_CACHE_SIZE;
		}

		if (page_has_buffers(tmppage)) {
			if (ocfs2_should_order_data(inode))
				walk_page_buffers(wc->w_handle,
						  page_buffers(tmppage),
						  from, to, NULL,
						  ocfs2_journal_dirty_data);
			block_commit_write(tmppage, from, to);
		}
	}

out_write_size:
	pos += copied;
	if (pos > inode->i_size) {
		i_size_write(inode, pos);
		mark_inode_dirty(inode);
	}
	inode->i_blocks = ocfs2_inode_sector_count(inode);
	di->i_size = cpu_to_le64((u64)i_size_read(inode));
	inode->i_mtime = inode->i_ctime = CURRENT_TIME;
	di->i_mtime = di->i_ctime = cpu_to_le64(inode->i_mtime.tv_sec);
	di->i_mtime_nsec = di->i_ctime_nsec = cpu_to_le32(inode->i_mtime.tv_nsec);
	ocfs2_journal_dirty(handle, wc->w_di_bh);

	ocfs2_commit_trans(osb, handle);

	ocfs2_run_deallocs(osb, &wc->w_dealloc);

	ocfs2_free_write_ctxt(wc);

	return copied;
}

int ocfs2_write_end(struct file *file, struct address_space *mapping,
		    loff_t pos, unsigned len, unsigned copied,
		    struct page *page, void *fsdata)
{
	int ret;
	struct inode *inode = mapping->host;

	ret = ocfs2_write_end_nolock(mapping, pos, len, copied, page, fsdata);

	up_write(&OCFS2_I(inode)->ip_alloc_sem);
	ocfs2_inode_unlock(inode, 1);

	return ret;
}

#ifdef AOPS_IS_NOT_CONST
struct address_space_operations ocfs2_aops = {
#else
const struct address_space_operations ocfs2_aops = {
#endif
	.readpage	= ocfs2_readpage,
	.readpages	= ocfs2_readpages,
	.writepage	= ocfs2_writepage,
	.bmap		= ocfs2_bmap,
	.sync_page	= block_sync_page,
	.direct_IO	= ocfs2_direct_IO,
	.invalidatepage	= ocfs2_invalidatepage,
	.releasepage	= ocfs2_releasepage,
	.migratepage	= buffer_migrate_page,
};