aops.c

ocfs1.4.1 oracle分布式文件系统

/* -*- mode: c; c-basic-offset: 8; -*-
 * vim: noexpandtab sw=8 ts=8 sts=0:
 *
 * Copyright (C) 2002, 2004 Oracle.  All rights reserved.
 *
 * This program is free software; you can redistribute it and/or
 * modify it under the terms of the GNU General Public
 * License as published by the Free Software Foundation; either
 * version 2 of the License, or (at your option) any later version.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 * General Public License for more details.
 *
 * You should have received a copy of the GNU General Public
 * License along with this program; if not, write to the
 * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
 * Boston, MA 021110-1307, USA.
 */

#include <linux/fs.h>
#include <linux/slab.h>
#include <linux/highmem.h>
#include <linux/pagemap.h>
#include <asm/byteorder.h>
#include <linux/swap.h>
#include <linux/pipe_fs_i.h>
#include <linux/mpage.h>

#define MLOG_MASK_PREFIX ML_FILE_IO
#include <cluster/masklog.h>

#include "ocfs2.h"

#include "alloc.h"
#include "aops.h"
#include "dlmglue.h"
#include "extent_map.h"
#include "file.h"
#include "inode.h"
#include "journal.h"
#include "suballoc.h"
#include "super.h"
#include "symlink.h"

#include "buffer_head_io.h"

static int ocfs2_symlink_get_block(struct inode *inode, sector_t iblock,
				   struct buffer_head *bh_result, int create)
{
	int err = -EIO;
	int status;
	struct ocfs2_dinode *fe = NULL;
	struct buffer_head *bh = NULL;
	struct buffer_head *buffer_cache_bh = NULL;
	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
	void *kaddr;

	mlog_entry("(0x%p, %llu, 0x%p, %d)\n", inode,
		   (unsigned long long)iblock, bh_result, create);

	BUG_ON(ocfs2_inode_is_fast_symlink(inode));

	if ((iblock << inode->i_sb->s_blocksize_bits) > PATH_MAX + 1) {
		mlog(ML_ERROR, "block offset > PATH_MAX: %llu",
		     (unsigned long long)iblock);
		goto bail;
	}

	status = ocfs2_read_block(OCFS2_SB(inode->i_sb),
				  OCFS2_I(inode)->ip_blkno,
				  &bh, OCFS2_BH_CACHED, inode);
	if (status < 0) {
		mlog_errno(status);
		goto bail;
	}
	fe = (struct ocfs2_dinode *) bh->b_data;

	if (!OCFS2_IS_VALID_DINODE(fe)) {
		mlog(ML_ERROR, "Invalid dinode #%llu: signature = %.*s\n",
		     (unsigned long long)le64_to_cpu(fe->i_blkno), 7,
		     fe->i_signature);
		goto bail;
	}

	if ((u64)iblock >= ocfs2_clusters_to_blocks(inode->i_sb,
						    le32_to_cpu(fe->i_clusters))) {
		mlog(ML_ERROR, "block offset is outside the allocated size: "
		     "%llu\n", (unsigned long long)iblock);
		goto bail;
	}

	/* We don't use the page cache to create symlink data, so if
	 * need be, copy it over from the buffer cache. */
	if (!buffer_uptodate(bh_result) && ocfs2_inode_is_new(inode)) {
		u64 blkno = le64_to_cpu(fe->id2.i_list.l_recs[0].e_blkno) +
			    iblock;
		buffer_cache_bh = sb_getblk(osb->sb, blkno);
		if (!buffer_cache_bh) {
			mlog(ML_ERROR, "couldn't getblock for symlink!\n");
			goto bail;
		}

		/* we haven't locked out transactions, so a commit
		 * could've happened. Since we've got a reference on
		 * the bh, even if it commits while we're doing the
		 * copy, the data is still good. */
		if (buffer_jbd(buffer_cache_bh)
		    && ocfs2_inode_is_new(inode)) {
			kaddr = kmap_atomic(bh_result->b_page, KM_USER0);
			if (!kaddr) {
				mlog(ML_ERROR, "couldn't kmap!\n");
				goto bail;
			}
			memcpy(kaddr + (bh_result->b_size * iblock),
			       buffer_cache_bh->b_data,
			       bh_result->b_size);
			kunmap_atomic(kaddr, KM_USER0);
			set_buffer_uptodate(bh_result);
		}
		brelse(buffer_cache_bh);
	}

	map_bh(bh_result, inode->i_sb,
	       le64_to_cpu(fe->id2.i_list.l_recs[0].e_blkno) + iblock);

	err = 0;

bail:
	if (bh)
		brelse(bh);

	mlog_exit(err);
	return err;
}

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 ocfs2_read_inline_data(struct inode *inode, struct page *page,
			   struct buffer_head *di_bh)
{
	void *kaddr;
	loff_t size;
	struct ocfs2_dinode *di = (struct ocfs2_dinode *)di_bh->b_data;

	if (!(le16_to_cpu(di->i_dyn_features) & OCFS2_INLINE_DATA_FL)) {
		ocfs2_error(inode->i_sb, "Inode %llu lost inline data flag",
			    (unsigned long long)OCFS2_I(inode)->ip_blkno);
		return -EROFS;
	}

	size = i_size_read(inode);

	if (size > PAGE_CACHE_SIZE ||
	    size > ocfs2_max_inline_data(inode->i_sb)) {
		ocfs2_error(inode->i_sb,
			    "Inode %llu has with inline data has bad size: %Lu",
			    (unsigned long long)OCFS2_I(inode)->ip_blkno,
			    (unsigned long long)size);
		return -EROFS;
	}

	kaddr = kmap_atomic(page, KM_USER0);
	if (size)
		memcpy(kaddr, di->id2.i_data.id_data, size);
	/* Clear the remaining part of the page */
	memset(kaddr + size, 0, PAGE_CACHE_SIZE - size);
	flush_dcache_page(page);
	kunmap_atomic(kaddr, KM_USER0);

	SetPageUptodate(page);

	return 0;
}

static int ocfs2_readpage_inline(struct inode *inode, struct page *page)
{
	int ret;
	struct buffer_head *di_bh = NULL;
	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);

	BUG_ON(!PageLocked(page));
	BUG_ON(!(OCFS2_I(inode)->ip_dyn_features & OCFS2_INLINE_DATA_FL));

	ret = ocfs2_read_block(osb, OCFS2_I(inode)->ip_blkno, &di_bh,
			       OCFS2_BH_CACHED, inode);
	if (ret) {
		mlog_errno(ret);
		goto out;
	}

	ret = ocfs2_read_inline_data(inode, page, di_bh);
out:
	unlock_page(page);

	brelse(di_bh);
	return ret;
}

static int ocfs2_readpage(struct file *file, struct page *page)
{
	struct inode *inode = page->mapping->host;
	struct ocfs2_inode_info *oi = OCFS2_I(inode);
	loff_t start = (loff_t)page->index << PAGE_CACHE_SHIFT;
	int ret, unlock = 1;

	mlog_entry("(0x%p, %lu)\n", file, (page ? page->index : 0));

	ret = ocfs2_inode_lock_with_page(inode, NULL, 0, page);
	if (ret != 0) {
		if (ret == AOP_TRUNCATED_PAGE)
			unlock = 0;
		mlog_errno(ret);
		goto out;
	}

	if (down_read_trylock(&oi->ip_alloc_sem) == 0) {
		ret = AOP_TRUNCATED_PAGE;
		goto out_inode_unlock;
	}

	/*
	 * i_size might have just been updated as we grabed the meta lock.  We
	 * might now be discovering a truncate that hit on another node.
	 * block_read_full_page->get_block freaks out if it is asked to read
	 * beyond the end of a file, so we check here.  Callers
	 * (generic_file_read, vm_ops->fault) are clever enough to check i_size
	 * and notice that the page they just read isn't needed.
	 *
	 * XXX sys_readahead() seems to get that wrong?
	 */
	if (start >= i_size_read(inode)) {
		zero_user_page(page, 0, PAGE_SIZE, KM_USER0);
		SetPageUptodate(page);
		ret = 0;
		goto out_alloc;
	}

	if (oi->ip_dyn_features & OCFS2_INLINE_DATA_FL)
		ret = ocfs2_readpage_inline(inode, page);
	else
		ret = block_read_full_page(page, ocfs2_get_block);
	unlock = 0;

out_alloc:
	up_read(&OCFS2_I(inode)->ip_alloc_sem);
out_inode_unlock:
	ocfs2_inode_unlock(inode, 0);
out:
	if (unlock)
		unlock_page(page);
	mlog_exit(ret);
	return ret;
}

/*
 * This is used only for read-ahead. Failures or difficult to handle
 * situations are safe to ignore.
 *
 * Right now, we don't bother with BH_Boundary - in-inode extent lists
 * are quite large (243 extents on 4k blocks), so most inodes don't
 * grow out to a tree. If need be, detecting boundary extents could
 * trivially be added in a future version of ocfs2_get_block().
 */
static int ocfs2_readpages(struct file *filp, struct address_space *mapping,
			   struct list_head *pages, unsigned nr_pages)
{
	int ret, err = -EIO;
	struct inode *inode = mapping->host;
	struct ocfs2_inode_info *oi = OCFS2_I(inode);
	loff_t start;
	struct page *last;

	/*
	 * Use the nonblocking flag for the dlm code to avoid page
	 * lock inversion, but don't bother with retrying.
	 */
	ret = ocfs2_inode_lock_full(inode, NULL, 0, OCFS2_LOCK_NONBLOCK);
	if (ret)
		return err;

	if (down_read_trylock(&oi->ip_alloc_sem) == 0) {
		ocfs2_inode_unlock(inode, 0);
		return err;
	}

	/*
	 * Don't bother with inline-data. There isn't anything
	 * to read-ahead in that case anyway...
	 */
	if (oi->ip_dyn_features & OCFS2_INLINE_DATA_FL)
		goto out_unlock;

	/*
	 * Check whether a remote node truncated this file - we just
	 * drop out in that case as it's not worth handling here.
	 */
	last = list_entry(pages->prev, struct page, lru);
	start = (loff_t)last->index << PAGE_CACHE_SHIFT;
	if (start >= i_size_read(inode))
		goto out_unlock;

	err = mpage_readpages(mapping, pages, nr_pages, ocfs2_get_block);

out_unlock:
	up_read(&oi->ip_alloc_sem);
	ocfs2_inode_unlock(inode, 0);

	return err;
}

/* Note: Because we don't support holes, our allocation has
 * already happened (allocation writes zeros to the file data)
 * so we don't have to worry about ordered writes in
 * ocfs2_writepage.
 *
 * ->writepage is called during the process of invalidating the page cache
 * during blocked lock processing.  It can't block on any cluster locks
 * to during block mapping.  It's relying on the fact that the block
 * mapping can't have disappeared under the dirty pages that it is
 * being asked to write back.
 */
static int ocfs2_writepage(struct page *page, struct writeback_control *wbc)
{
	int ret;

	mlog_entry("(0x%p)\n", page);

	ret = block_write_full_page(page, ocfs2_get_block, wbc);

	mlog_exit(ret);

	return ret;
}

/*
 * This is called from ocfs2_write_zero_page() which has handled it's
 * own cluster locking and has ensured allocation exists for those
 * blocks to be written.
 */
int ocfs2_prepare_write_nolock(struct inode *inode, struct page *page,
			       unsigned from, unsigned to)
{
	int ret;

	ret = block_prepare_write(page, from, to, ocfs2_get_block);

	return ret;
}

/* Taken from ext3. We don't necessarily need the full blown
 * functionality yet, but IMHO it's better to cut and paste the whole
 * thing so we can avoid introducing our own bugs (and easily pick up
 * their fixes when they happen) --Mark */
int walk_page_buffers(	handle_t *handle,
			struct buffer_head *head,
			unsigned from,
			unsigned to,
			int *partial,
			int (*fn)(	handle_t *handle,
					struct buffer_head *bh))
{
	struct buffer_head *bh;
	unsigned block_start, block_end;
	unsigned blocksize = head->b_size;
	int err, ret = 0;
	struct buffer_head *next;

	for (	bh = head, block_start = 0;
		ret == 0 && (bh != head || !block_start);
	    	block_start = block_end, bh = next)
	{
		next = bh->b_this_page;
		block_end = block_start + blocksize;
		if (block_end <= from || block_start >= to) {
			if (partial && !buffer_uptodate(bh))
				*partial = 1;
			continue;
		}
		err = (*fn)(handle, bh);
		if (!ret)
			ret = err;
	}
	return ret;
}

handle_t *ocfs2_start_walk_page_trans(struct inode *inode,
							 struct page *page,
							 unsigned from,
							 unsigned to)
{
	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
	handle_t *handle;
	int ret = 0;

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

	if (ocfs2_should_order_data(inode)) {
		ret = walk_page_buffers(handle,
					page_buffers(page),
					from, to, NULL,
					ocfs2_journal_dirty_data);
		if (ret < 0) 
			mlog_errno(ret);
	}
out:
	if (ret) {