2006-01-16 09:50:04 -07:00
|
|
|
/*
|
|
|
|
* Copyright (C) Sistina Software, Inc. 1997-2003 All rights reserved.
|
2006-05-18 13:09:15 -06:00
|
|
|
* Copyright (C) 2004-2006 Red Hat, Inc. All rights reserved.
|
2006-01-16 09:50:04 -07:00
|
|
|
*
|
|
|
|
* This copyrighted material is made available to anyone wishing to use,
|
|
|
|
* modify, copy, or redistribute it subject to the terms and conditions
|
2006-09-01 09:05:15 -06:00
|
|
|
* of the GNU General Public License version 2.
|
2006-01-16 09:50:04 -07:00
|
|
|
*/
|
|
|
|
|
|
|
|
#include <linux/sched.h>
|
|
|
|
#include <linux/slab.h>
|
|
|
|
#include <linux/spinlock.h>
|
|
|
|
#include <linux/completion.h>
|
|
|
|
#include <linux/buffer_head.h>
|
|
|
|
#include <linux/mm.h>
|
|
|
|
#include <linux/pagemap.h>
|
|
|
|
#include <linux/writeback.h>
|
|
|
|
#include <linux/swap.h>
|
|
|
|
#include <linux/delay.h>
|
2006-10-02 09:38:25 -06:00
|
|
|
#include <linux/bio.h>
|
2006-02-27 15:23:27 -07:00
|
|
|
#include <linux/gfs2_ondisk.h>
|
2006-09-18 23:56:29 -06:00
|
|
|
#include <linux/lm_interface.h>
|
2006-01-16 09:50:04 -07:00
|
|
|
|
|
|
|
#include "gfs2.h"
|
2006-02-27 15:23:27 -07:00
|
|
|
#include "incore.h"
|
2006-01-16 09:50:04 -07:00
|
|
|
#include "glock.h"
|
|
|
|
#include "glops.h"
|
|
|
|
#include "inode.h"
|
|
|
|
#include "log.h"
|
|
|
|
#include "lops.h"
|
|
|
|
#include "meta_io.h"
|
|
|
|
#include "rgrp.h"
|
|
|
|
#include "trans.h"
|
2006-02-27 15:23:27 -07:00
|
|
|
#include "util.h"
|
2006-07-11 07:46:33 -06:00
|
|
|
#include "ops_address.h"
|
2006-01-16 09:50:04 -07:00
|
|
|
|
|
|
|
static int aspace_get_block(struct inode *inode, sector_t lblock,
|
|
|
|
struct buffer_head *bh_result, int create)
|
|
|
|
{
|
2006-02-27 15:23:27 -07:00
|
|
|
gfs2_assert_warn(inode->i_sb->s_fs_info, 0);
|
2006-01-16 09:50:04 -07:00
|
|
|
return -EOPNOTSUPP;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int gfs2_aspace_writepage(struct page *page,
|
|
|
|
struct writeback_control *wbc)
|
|
|
|
{
|
|
|
|
return block_write_full_page(page, aspace_get_block, wbc);
|
|
|
|
}
|
|
|
|
|
2006-07-03 11:37:30 -06:00
|
|
|
static const struct address_space_operations aspace_aops = {
|
2006-01-16 09:50:04 -07:00
|
|
|
.writepage = gfs2_aspace_writepage,
|
2006-07-11 07:46:33 -06:00
|
|
|
.releasepage = gfs2_releasepage,
|
2006-01-16 09:50:04 -07:00
|
|
|
};
|
|
|
|
|
|
|
|
/**
|
|
|
|
* gfs2_aspace_get - Create and initialize a struct inode structure
|
|
|
|
* @sdp: the filesystem the aspace is in
|
|
|
|
*
|
|
|
|
* Right now a struct inode is just a struct inode. Maybe Linux
|
|
|
|
* will supply a more lightweight address space construct (that works)
|
|
|
|
* in the future.
|
|
|
|
*
|
|
|
|
* Make sure pages/buffers in this aspace aren't in high memory.
|
|
|
|
*
|
|
|
|
* Returns: the aspace
|
|
|
|
*/
|
|
|
|
|
|
|
|
struct inode *gfs2_aspace_get(struct gfs2_sbd *sdp)
|
|
|
|
{
|
|
|
|
struct inode *aspace;
|
|
|
|
|
|
|
|
aspace = new_inode(sdp->sd_vfs);
|
|
|
|
if (aspace) {
|
2006-07-11 07:50:54 -06:00
|
|
|
mapping_set_gfp_mask(aspace->i_mapping, GFP_NOFS);
|
2006-01-16 09:50:04 -07:00
|
|
|
aspace->i_mapping->a_ops = &aspace_aops;
|
|
|
|
aspace->i_size = ~0ULL;
|
2006-09-27 02:50:31 -06:00
|
|
|
aspace->i_private = NULL;
|
2006-01-16 09:50:04 -07:00
|
|
|
insert_inode_hash(aspace);
|
|
|
|
}
|
|
|
|
return aspace;
|
|
|
|
}
|
|
|
|
|
|
|
|
void gfs2_aspace_put(struct inode *aspace)
|
|
|
|
{
|
|
|
|
remove_inode_hash(aspace);
|
|
|
|
iput(aspace);
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* gfs2_meta_inval - Invalidate all buffers associated with a glock
|
|
|
|
* @gl: the glock
|
|
|
|
*
|
|
|
|
*/
|
|
|
|
|
|
|
|
void gfs2_meta_inval(struct gfs2_glock *gl)
|
|
|
|
{
|
|
|
|
struct gfs2_sbd *sdp = gl->gl_sbd;
|
|
|
|
struct inode *aspace = gl->gl_aspace;
|
|
|
|
struct address_space *mapping = gl->gl_aspace->i_mapping;
|
|
|
|
|
|
|
|
gfs2_assert_withdraw(sdp, !atomic_read(&gl->gl_ail_count));
|
|
|
|
|
|
|
|
atomic_inc(&aspace->i_writecount);
|
|
|
|
truncate_inode_pages(mapping, 0);
|
|
|
|
atomic_dec(&aspace->i_writecount);
|
|
|
|
|
|
|
|
gfs2_assert_withdraw(sdp, !mapping->nrpages);
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* gfs2_meta_sync - Sync all buffers associated with a glock
|
|
|
|
* @gl: The glock
|
|
|
|
*
|
|
|
|
*/
|
|
|
|
|
2006-09-21 15:05:23 -06:00
|
|
|
void gfs2_meta_sync(struct gfs2_glock *gl)
|
2006-01-16 09:50:04 -07:00
|
|
|
{
|
|
|
|
struct address_space *mapping = gl->gl_aspace->i_mapping;
|
2006-09-21 15:05:23 -06:00
|
|
|
int error;
|
2006-01-16 09:50:04 -07:00
|
|
|
|
2006-09-21 15:05:23 -06:00
|
|
|
filemap_fdatawrite(mapping);
|
|
|
|
error = filemap_fdatawait(mapping);
|
2006-01-16 09:50:04 -07:00
|
|
|
|
|
|
|
if (error)
|
|
|
|
gfs2_io_error(gl->gl_sbd);
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* getbuf - Get a buffer with a given address space
|
2006-11-23 09:16:32 -07:00
|
|
|
* @gl: the glock
|
2006-01-16 09:50:04 -07:00
|
|
|
* @blkno: the block number (filesystem scope)
|
|
|
|
* @create: 1 if the buffer should be created
|
|
|
|
*
|
|
|
|
* Returns: the buffer
|
|
|
|
*/
|
|
|
|
|
2006-11-23 09:16:32 -07:00
|
|
|
static struct buffer_head *getbuf(struct gfs2_glock *gl, u64 blkno, int create)
|
2006-01-16 09:50:04 -07:00
|
|
|
{
|
2006-11-23 09:16:32 -07:00
|
|
|
struct address_space *mapping = gl->gl_aspace->i_mapping;
|
|
|
|
struct gfs2_sbd *sdp = gl->gl_sbd;
|
2006-01-16 09:50:04 -07:00
|
|
|
struct page *page;
|
|
|
|
struct buffer_head *bh;
|
|
|
|
unsigned int shift;
|
|
|
|
unsigned long index;
|
|
|
|
unsigned int bufnum;
|
|
|
|
|
|
|
|
shift = PAGE_CACHE_SHIFT - sdp->sd_sb.sb_bsize_shift;
|
|
|
|
index = blkno >> shift; /* convert block to page */
|
|
|
|
bufnum = blkno - (index << shift); /* block buf index within page */
|
|
|
|
|
|
|
|
if (create) {
|
|
|
|
for (;;) {
|
2006-11-23 09:16:32 -07:00
|
|
|
page = grab_cache_page(mapping, index);
|
2006-01-16 09:50:04 -07:00
|
|
|
if (page)
|
|
|
|
break;
|
|
|
|
yield();
|
|
|
|
}
|
|
|
|
} else {
|
2006-11-23 09:16:32 -07:00
|
|
|
page = find_lock_page(mapping, index);
|
2006-01-16 09:50:04 -07:00
|
|
|
if (!page)
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (!page_has_buffers(page))
|
|
|
|
create_empty_buffers(page, sdp->sd_sb.sb_bsize, 0);
|
|
|
|
|
|
|
|
/* Locate header for our buffer within our page */
|
|
|
|
for (bh = page_buffers(page); bufnum--; bh = bh->b_this_page)
|
|
|
|
/* Do nothing */;
|
|
|
|
get_bh(bh);
|
|
|
|
|
|
|
|
if (!buffer_mapped(bh))
|
|
|
|
map_bh(bh, sdp->sd_vfs, blkno);
|
|
|
|
|
|
|
|
unlock_page(page);
|
|
|
|
mark_page_accessed(page);
|
|
|
|
page_cache_release(page);
|
|
|
|
|
|
|
|
return bh;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void meta_prep_new(struct buffer_head *bh)
|
|
|
|
{
|
|
|
|
struct gfs2_meta_header *mh = (struct gfs2_meta_header *)bh->b_data;
|
|
|
|
|
|
|
|
lock_buffer(bh);
|
|
|
|
clear_buffer_dirty(bh);
|
|
|
|
set_buffer_uptodate(bh);
|
|
|
|
unlock_buffer(bh);
|
|
|
|
|
|
|
|
mh->mh_magic = cpu_to_be32(GFS2_MAGIC);
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* gfs2_meta_new - Get a block
|
|
|
|
* @gl: The glock associated with this block
|
|
|
|
* @blkno: The block number
|
|
|
|
*
|
|
|
|
* Returns: The buffer
|
|
|
|
*/
|
|
|
|
|
2006-09-04 10:49:07 -06:00
|
|
|
struct buffer_head *gfs2_meta_new(struct gfs2_glock *gl, u64 blkno)
|
2006-01-16 09:50:04 -07:00
|
|
|
{
|
|
|
|
struct buffer_head *bh;
|
2006-11-23 09:16:32 -07:00
|
|
|
bh = getbuf(gl, blkno, CREATE);
|
2006-01-16 09:50:04 -07:00
|
|
|
meta_prep_new(bh);
|
|
|
|
return bh;
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* gfs2_meta_read - Read a block from disk
|
|
|
|
* @gl: The glock covering the block
|
|
|
|
* @blkno: The block number
|
2006-09-21 15:05:23 -06:00
|
|
|
* @flags: flags
|
2006-01-16 09:50:04 -07:00
|
|
|
* @bhp: the place where the buffer is returned (NULL on failure)
|
|
|
|
*
|
|
|
|
* Returns: errno
|
|
|
|
*/
|
|
|
|
|
2006-09-04 10:49:07 -06:00
|
|
|
int gfs2_meta_read(struct gfs2_glock *gl, u64 blkno, int flags,
|
2006-01-16 09:50:04 -07:00
|
|
|
struct buffer_head **bhp)
|
|
|
|
{
|
2006-11-23 09:16:32 -07:00
|
|
|
*bhp = getbuf(gl, blkno, CREATE);
|
2006-09-21 15:05:23 -06:00
|
|
|
if (!buffer_uptodate(*bhp))
|
2006-10-02 09:38:25 -06:00
|
|
|
ll_rw_block(READ_META, 1, bhp);
|
2006-09-21 15:05:23 -06:00
|
|
|
if (flags & DIO_WAIT) {
|
|
|
|
int error = gfs2_meta_wait(gl->gl_sbd, *bhp);
|
|
|
|
if (error) {
|
|
|
|
brelse(*bhp);
|
|
|
|
return error;
|
|
|
|
}
|
|
|
|
}
|
2006-01-16 09:50:04 -07:00
|
|
|
|
2006-09-21 15:05:23 -06:00
|
|
|
return 0;
|
2006-01-16 09:50:04 -07:00
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
2006-09-21 15:05:23 -06:00
|
|
|
* gfs2_meta_wait - Reread a block from disk
|
2006-01-16 09:50:04 -07:00
|
|
|
* @sdp: the filesystem
|
2006-09-21 15:05:23 -06:00
|
|
|
* @bh: The block to wait for
|
2006-01-16 09:50:04 -07:00
|
|
|
*
|
|
|
|
* Returns: errno
|
|
|
|
*/
|
|
|
|
|
2006-09-21 15:05:23 -06:00
|
|
|
int gfs2_meta_wait(struct gfs2_sbd *sdp, struct buffer_head *bh)
|
2006-01-16 09:50:04 -07:00
|
|
|
{
|
|
|
|
if (unlikely(test_bit(SDF_SHUTDOWN, &sdp->sd_flags)))
|
|
|
|
return -EIO;
|
|
|
|
|
2006-09-21 15:05:23 -06:00
|
|
|
wait_on_buffer(bh);
|
2006-01-16 09:50:04 -07:00
|
|
|
|
2006-09-21 15:05:23 -06:00
|
|
|
if (!buffer_uptodate(bh)) {
|
|
|
|
struct gfs2_trans *tr = current->journal_info;
|
|
|
|
if (tr && tr->tr_touched)
|
|
|
|
gfs2_io_error_bh(sdp, bh);
|
|
|
|
return -EIO;
|
2006-01-16 09:50:04 -07:00
|
|
|
}
|
2006-09-21 15:05:23 -06:00
|
|
|
if (unlikely(test_bit(SDF_SHUTDOWN, &sdp->sd_flags)))
|
|
|
|
return -EIO;
|
2006-01-16 09:50:04 -07:00
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
2006-01-18 04:32:00 -07:00
|
|
|
* gfs2_attach_bufdata - attach a struct gfs2_bufdata structure to a buffer
|
2006-01-16 09:50:04 -07:00
|
|
|
* @gl: the glock the buffer belongs to
|
|
|
|
* @bh: The buffer to be attached to
|
2006-01-18 04:32:00 -07:00
|
|
|
* @meta: Flag to indicate whether its metadata or not
|
2006-01-16 09:50:04 -07:00
|
|
|
*/
|
|
|
|
|
2006-02-27 10:00:42 -07:00
|
|
|
void gfs2_attach_bufdata(struct gfs2_glock *gl, struct buffer_head *bh,
|
|
|
|
int meta)
|
2006-01-16 09:50:04 -07:00
|
|
|
{
|
|
|
|
struct gfs2_bufdata *bd;
|
|
|
|
|
2006-02-08 04:50:51 -07:00
|
|
|
if (meta)
|
|
|
|
lock_page(bh->b_page);
|
2006-01-16 09:50:04 -07:00
|
|
|
|
2006-02-27 15:23:27 -07:00
|
|
|
if (bh->b_private) {
|
2006-02-08 04:50:51 -07:00
|
|
|
if (meta)
|
|
|
|
unlock_page(bh->b_page);
|
2006-01-16 09:50:04 -07:00
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
2007-02-10 02:45:03 -07:00
|
|
|
bd = kmem_cache_zalloc(gfs2_bufdata_cachep, GFP_NOFS | __GFP_NOFAIL),
|
2006-01-16 09:50:04 -07:00
|
|
|
bd->bd_bh = bh;
|
|
|
|
bd->bd_gl = gl;
|
|
|
|
|
|
|
|
INIT_LIST_HEAD(&bd->bd_list_tr);
|
2006-09-04 12:47:06 -06:00
|
|
|
if (meta)
|
2006-01-18 04:32:00 -07:00
|
|
|
lops_init_le(&bd->bd_le, &gfs2_buf_lops);
|
2006-09-04 12:47:06 -06:00
|
|
|
else
|
2006-01-18 04:32:00 -07:00
|
|
|
lops_init_le(&bd->bd_le, &gfs2_databuf_lops);
|
2006-02-27 15:23:27 -07:00
|
|
|
bh->b_private = bd;
|
2006-01-16 09:50:04 -07:00
|
|
|
|
2006-02-08 04:50:51 -07:00
|
|
|
if (meta)
|
|
|
|
unlock_page(bh->b_page);
|
2006-01-16 09:50:04 -07:00
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
2006-01-18 06:38:44 -07:00
|
|
|
* gfs2_pin - Pin a buffer in memory
|
2006-01-16 09:50:04 -07:00
|
|
|
* @sdp: the filesystem the buffer belongs to
|
|
|
|
* @bh: The buffer to be pinned
|
|
|
|
*
|
|
|
|
*/
|
|
|
|
|
2006-01-18 06:38:44 -07:00
|
|
|
void gfs2_pin(struct gfs2_sbd *sdp, struct buffer_head *bh)
|
2006-01-16 09:50:04 -07:00
|
|
|
{
|
2006-02-27 15:23:27 -07:00
|
|
|
struct gfs2_bufdata *bd = bh->b_private;
|
2006-01-16 09:50:04 -07:00
|
|
|
|
|
|
|
gfs2_assert_withdraw(sdp, test_bit(SDF_JOURNAL_LIVE, &sdp->sd_flags));
|
|
|
|
|
|
|
|
if (test_set_buffer_pinned(bh))
|
|
|
|
gfs2_assert_withdraw(sdp, 0);
|
|
|
|
|
|
|
|
wait_on_buffer(bh);
|
|
|
|
|
|
|
|
/* If this buffer is in the AIL and it has already been written
|
|
|
|
to in-place disk block, remove it from the AIL. */
|
|
|
|
|
|
|
|
gfs2_log_lock(sdp);
|
|
|
|
if (bd->bd_ail && !buffer_in_io(bh))
|
|
|
|
list_move(&bd->bd_ail_st_list, &bd->bd_ail->ai_ail2_list);
|
|
|
|
gfs2_log_unlock(sdp);
|
|
|
|
|
|
|
|
clear_buffer_dirty(bh);
|
|
|
|
wait_on_buffer(bh);
|
|
|
|
|
|
|
|
if (!buffer_uptodate(bh))
|
|
|
|
gfs2_io_error_bh(sdp, bh);
|
|
|
|
|
|
|
|
get_bh(bh);
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
2006-01-18 06:38:44 -07:00
|
|
|
* gfs2_unpin - Unpin a buffer
|
2006-01-16 09:50:04 -07:00
|
|
|
* @sdp: the filesystem the buffer belongs to
|
|
|
|
* @bh: The buffer to unpin
|
|
|
|
* @ai:
|
|
|
|
*
|
|
|
|
*/
|
|
|
|
|
2006-01-18 06:38:44 -07:00
|
|
|
void gfs2_unpin(struct gfs2_sbd *sdp, struct buffer_head *bh,
|
|
|
|
struct gfs2_ail *ai)
|
2006-01-16 09:50:04 -07:00
|
|
|
{
|
2006-02-27 15:23:27 -07:00
|
|
|
struct gfs2_bufdata *bd = bh->b_private;
|
2006-01-16 09:50:04 -07:00
|
|
|
|
|
|
|
gfs2_assert_withdraw(sdp, buffer_uptodate(bh));
|
|
|
|
|
|
|
|
if (!buffer_pinned(bh))
|
|
|
|
gfs2_assert_withdraw(sdp, 0);
|
|
|
|
|
|
|
|
mark_buffer_dirty(bh);
|
|
|
|
clear_buffer_pinned(bh);
|
|
|
|
|
|
|
|
gfs2_log_lock(sdp);
|
|
|
|
if (bd->bd_ail) {
|
|
|
|
list_del(&bd->bd_ail_st_list);
|
|
|
|
brelse(bh);
|
|
|
|
} else {
|
|
|
|
struct gfs2_glock *gl = bd->bd_gl;
|
|
|
|
list_add(&bd->bd_ail_gl_list, &gl->gl_ail_list);
|
|
|
|
atomic_inc(&gl->gl_ail_count);
|
|
|
|
}
|
|
|
|
bd->bd_ail = ai;
|
|
|
|
list_add(&bd->bd_ail_st_list, &ai->ai_ail1_list);
|
|
|
|
gfs2_log_unlock(sdp);
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* gfs2_meta_wipe - make inode's buffers so they aren't dirty/pinned anymore
|
|
|
|
* @ip: the inode who owns the buffers
|
|
|
|
* @bstart: the first buffer in the run
|
|
|
|
* @blen: the number of buffers in the run
|
|
|
|
*
|
|
|
|
*/
|
|
|
|
|
2006-09-04 10:49:07 -06:00
|
|
|
void gfs2_meta_wipe(struct gfs2_inode *ip, u64 bstart, u32 blen)
|
2006-01-16 09:50:04 -07:00
|
|
|
{
|
2006-06-14 13:32:57 -06:00
|
|
|
struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
|
2006-01-16 09:50:04 -07:00
|
|
|
struct buffer_head *bh;
|
|
|
|
|
|
|
|
while (blen) {
|
2006-11-23 09:16:32 -07:00
|
|
|
bh = getbuf(ip->i_gl, bstart, NO_CREATE);
|
2006-01-16 09:50:04 -07:00
|
|
|
if (bh) {
|
2006-02-27 15:23:27 -07:00
|
|
|
struct gfs2_bufdata *bd = bh->b_private;
|
2006-01-16 09:50:04 -07:00
|
|
|
|
|
|
|
if (test_clear_buffer_pinned(bh)) {
|
2006-02-27 15:23:27 -07:00
|
|
|
struct gfs2_trans *tr = current->journal_info;
|
[GFS2] assertion failure after writing to journaled file, umount
This patch passes all my nasty tests that were causing the code to
fail under one circumstance or another. Here is a complete summary
of all changes from today's git tree, in order of appearance:
1. There are now separate variables for metadata buffer accounting.
2. Variable sd_log_num_hdrs is no longer needed, since the header
accounting is taken care of by the reserve/refund sequence.
3. Fixed a tiny grammatical problem in a comment.
4. Added a new function "calc_reserved" to calculate the reserved
log space. This isn't entirely necessary, but it has two benefits:
First, it simplifies the gfs2_log_refund function greatly.
Second, it allows for easier debugging because I could sprinkle the
code with calls to this function to make sure the accounting is
proper (by adding asserts and printks) at strategic point of the code.
5. In log_pull_tail there apparently was a kludge to fix up the
accounting based on a "pull" parameter. The buffer accounting is
now done properly, so the kludge was removed.
6. File sync operations were making a call to gfs2_log_flush that
writes another journal header. Since that header was unplanned
for (reserved) by the reserve/refund sequence, the free space had
to be decremented so that when log_pull_tail gets called, the free
space is be adjusted properly. (Did I hear you call that a kludge?
well, maybe, but a lot more justifiable than the one I removed).
7. In the gfs2_log_shutdown code, it optionally syncs the log by
specifying the PULL parameter to log_write_header. I'm not sure
this is necessary anymore. It just seems to me there could be
cases where shutdown is called while there are outstanding log
buffers.
8. In the (data)buf_lo_before_commit functions, I changed some offset
values from being calculated on the fly to being constants. That
simplified some code and we might as well let the compiler do the
calculation once rather than redoing those cycles at run time.
9. This version has my rewritten databuf_lo_add function.
This version is much more like its predecessor, buf_lo_add, which
makes it easier to understand. Again, this might not be necessary,
but it seems as if this one works as well as the previous one,
maybe even better, so I decided to leave it in.
10. In databuf_lo_before_commit, a previous data corruption problem
was caused by going off the end of the buffer. The proper solution
is to have the proper limit in place, rather than stopping earlier.
(Thus my previous attempt to fix it is wrong).
If you don't wrap the buffer, you're stopping too early and that
causes more log buffer accounting problems.
11. In lops.h there are two new (previously mentioned) constants for
figuring out the data offset for the journal buffers.
12. There are also two new functions, buf_limit and databuf_limit to
calculate how many entries will fit in the buffer.
13. In function gfs2_meta_wipe, it needs to distinguish between pinned
metadata buffers and journaled data buffers for proper journal buffer
accounting. It can't use the JDATA gfs2_inode flag because it's
sometimes passed the "real" inode and sometimes the "metadata
inode" and the inode flags will be random bits in a metadata
gfs2_inode. It needs to base its decision on which was passed in.
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
2007-06-18 13:50:20 -06:00
|
|
|
struct gfs2_inode *bh_ip =
|
|
|
|
GFS2_I(bh->b_page->mapping->host);
|
|
|
|
|
2006-01-16 09:50:04 -07:00
|
|
|
gfs2_log_lock(sdp);
|
|
|
|
list_del_init(&bd->bd_le.le_list);
|
|
|
|
gfs2_assert_warn(sdp, sdp->sd_log_num_buf);
|
|
|
|
sdp->sd_log_num_buf--;
|
|
|
|
gfs2_log_unlock(sdp);
|
[GFS2] assertion failure after writing to journaled file, umount
This patch passes all my nasty tests that were causing the code to
fail under one circumstance or another. Here is a complete summary
of all changes from today's git tree, in order of appearance:
1. There are now separate variables for metadata buffer accounting.
2. Variable sd_log_num_hdrs is no longer needed, since the header
accounting is taken care of by the reserve/refund sequence.
3. Fixed a tiny grammatical problem in a comment.
4. Added a new function "calc_reserved" to calculate the reserved
log space. This isn't entirely necessary, but it has two benefits:
First, it simplifies the gfs2_log_refund function greatly.
Second, it allows for easier debugging because I could sprinkle the
code with calls to this function to make sure the accounting is
proper (by adding asserts and printks) at strategic point of the code.
5. In log_pull_tail there apparently was a kludge to fix up the
accounting based on a "pull" parameter. The buffer accounting is
now done properly, so the kludge was removed.
6. File sync operations were making a call to gfs2_log_flush that
writes another journal header. Since that header was unplanned
for (reserved) by the reserve/refund sequence, the free space had
to be decremented so that when log_pull_tail gets called, the free
space is be adjusted properly. (Did I hear you call that a kludge?
well, maybe, but a lot more justifiable than the one I removed).
7. In the gfs2_log_shutdown code, it optionally syncs the log by
specifying the PULL parameter to log_write_header. I'm not sure
this is necessary anymore. It just seems to me there could be
cases where shutdown is called while there are outstanding log
buffers.
8. In the (data)buf_lo_before_commit functions, I changed some offset
values from being calculated on the fly to being constants. That
simplified some code and we might as well let the compiler do the
calculation once rather than redoing those cycles at run time.
9. This version has my rewritten databuf_lo_add function.
This version is much more like its predecessor, buf_lo_add, which
makes it easier to understand. Again, this might not be necessary,
but it seems as if this one works as well as the previous one,
maybe even better, so I decided to leave it in.
10. In databuf_lo_before_commit, a previous data corruption problem
was caused by going off the end of the buffer. The proper solution
is to have the proper limit in place, rather than stopping earlier.
(Thus my previous attempt to fix it is wrong).
If you don't wrap the buffer, you're stopping too early and that
causes more log buffer accounting problems.
11. In lops.h there are two new (previously mentioned) constants for
figuring out the data offset for the journal buffers.
12. There are also two new functions, buf_limit and databuf_limit to
calculate how many entries will fit in the buffer.
13. In function gfs2_meta_wipe, it needs to distinguish between pinned
metadata buffers and journaled data buffers for proper journal buffer
accounting. It can't use the JDATA gfs2_inode flag because it's
sometimes passed the "real" inode and sometimes the "metadata
inode" and the inode flags will be random bits in a metadata
gfs2_inode. It needs to base its decision on which was passed in.
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
2007-06-18 13:50:20 -06:00
|
|
|
if (bh_ip->i_inode.i_private != NULL)
|
|
|
|
tr->tr_num_databuf_rm++;
|
|
|
|
else
|
|
|
|
tr->tr_num_buf_rm++;
|
2006-01-16 09:50:04 -07:00
|
|
|
brelse(bh);
|
|
|
|
}
|
|
|
|
if (bd) {
|
|
|
|
gfs2_log_lock(sdp);
|
|
|
|
if (bd->bd_ail) {
|
2006-09-04 10:49:07 -06:00
|
|
|
u64 blkno = bh->b_blocknr;
|
2006-01-16 09:50:04 -07:00
|
|
|
bd->bd_ail = NULL;
|
|
|
|
list_del(&bd->bd_ail_st_list);
|
|
|
|
list_del(&bd->bd_ail_gl_list);
|
|
|
|
atomic_dec(&bd->bd_gl->gl_ail_count);
|
|
|
|
brelse(bh);
|
|
|
|
gfs2_log_unlock(sdp);
|
|
|
|
gfs2_trans_add_revoke(sdp, blkno);
|
|
|
|
} else
|
|
|
|
gfs2_log_unlock(sdp);
|
|
|
|
}
|
|
|
|
|
|
|
|
lock_buffer(bh);
|
|
|
|
clear_buffer_dirty(bh);
|
|
|
|
clear_buffer_uptodate(bh);
|
|
|
|
unlock_buffer(bh);
|
|
|
|
|
|
|
|
brelse(bh);
|
|
|
|
}
|
|
|
|
|
|
|
|
bstart++;
|
|
|
|
blen--;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* gfs2_meta_cache_flush - get rid of any references on buffers for this inode
|
|
|
|
* @ip: The GFS2 inode
|
|
|
|
*
|
|
|
|
* This releases buffers that are in the most-recently-used array of
|
|
|
|
* blocks used for indirect block addressing for this inode.
|
|
|
|
*/
|
|
|
|
|
|
|
|
void gfs2_meta_cache_flush(struct gfs2_inode *ip)
|
|
|
|
{
|
|
|
|
struct buffer_head **bh_slot;
|
|
|
|
unsigned int x;
|
|
|
|
|
|
|
|
spin_lock(&ip->i_spin);
|
|
|
|
|
|
|
|
for (x = 0; x < GFS2_MAX_META_HEIGHT; x++) {
|
|
|
|
bh_slot = &ip->i_cache[x];
|
|
|
|
if (!*bh_slot)
|
|
|
|
break;
|
|
|
|
brelse(*bh_slot);
|
|
|
|
*bh_slot = NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
spin_unlock(&ip->i_spin);
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* gfs2_meta_indirect_buffer - Get a metadata buffer
|
|
|
|
* @ip: The GFS2 inode
|
|
|
|
* @height: The level of this buf in the metadata (indir addr) tree (if any)
|
|
|
|
* @num: The block number (device relative) of the buffer
|
|
|
|
* @new: Non-zero if we may create a new buffer
|
|
|
|
* @bhp: the buffer is returned here
|
|
|
|
*
|
|
|
|
* Try to use the gfs2_inode's MRU metadata tree cache.
|
|
|
|
*
|
|
|
|
* Returns: errno
|
|
|
|
*/
|
|
|
|
|
2006-09-04 10:49:07 -06:00
|
|
|
int gfs2_meta_indirect_buffer(struct gfs2_inode *ip, int height, u64 num,
|
2006-01-16 09:50:04 -07:00
|
|
|
int new, struct buffer_head **bhp)
|
|
|
|
{
|
2006-09-21 15:05:23 -06:00
|
|
|
struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
|
|
|
|
struct gfs2_glock *gl = ip->i_gl;
|
|
|
|
struct buffer_head *bh = NULL, **bh_slot = ip->i_cache + height;
|
|
|
|
int in_cache = 0;
|
2006-01-16 09:50:04 -07:00
|
|
|
|
2006-11-23 08:51:34 -07:00
|
|
|
BUG_ON(!gl);
|
|
|
|
BUG_ON(!sdp);
|
|
|
|
|
2006-01-16 09:50:04 -07:00
|
|
|
spin_lock(&ip->i_spin);
|
2006-09-21 15:05:23 -06:00
|
|
|
if (*bh_slot && (*bh_slot)->b_blocknr == num) {
|
|
|
|
bh = *bh_slot;
|
|
|
|
get_bh(bh);
|
|
|
|
in_cache = 1;
|
2006-01-16 09:50:04 -07:00
|
|
|
}
|
|
|
|
spin_unlock(&ip->i_spin);
|
|
|
|
|
2006-09-21 15:05:23 -06:00
|
|
|
if (!bh)
|
2006-11-23 09:16:32 -07:00
|
|
|
bh = getbuf(gl, num, CREATE);
|
2006-01-16 09:50:04 -07:00
|
|
|
|
2006-09-21 15:05:23 -06:00
|
|
|
if (!bh)
|
|
|
|
return -ENOBUFS;
|
2006-01-16 09:50:04 -07:00
|
|
|
|
|
|
|
if (new) {
|
2006-09-21 15:05:23 -06:00
|
|
|
if (gfs2_assert_warn(sdp, height))
|
|
|
|
goto err;
|
|
|
|
meta_prep_new(bh);
|
2006-01-18 04:19:28 -07:00
|
|
|
gfs2_trans_add_bh(ip->i_gl, bh, 1);
|
2006-01-16 09:50:04 -07:00
|
|
|
gfs2_metatype_set(bh, GFS2_METATYPE_IN, GFS2_FORMAT_IN);
|
|
|
|
gfs2_buffer_clear_tail(bh, sizeof(struct gfs2_meta_header));
|
2006-09-21 15:05:23 -06:00
|
|
|
} else {
|
|
|
|
u32 mtype = height ? GFS2_METATYPE_IN : GFS2_METATYPE_DI;
|
|
|
|
if (!buffer_uptodate(bh)) {
|
2006-10-02 09:38:25 -06:00
|
|
|
ll_rw_block(READ_META, 1, &bh);
|
2006-09-21 15:05:23 -06:00
|
|
|
if (gfs2_meta_wait(sdp, bh))
|
|
|
|
goto err;
|
|
|
|
}
|
|
|
|
if (gfs2_metatype_check(sdp, bh, mtype))
|
|
|
|
goto err;
|
|
|
|
}
|
2006-01-16 09:50:04 -07:00
|
|
|
|
2006-09-21 15:05:23 -06:00
|
|
|
if (!in_cache) {
|
|
|
|
spin_lock(&ip->i_spin);
|
|
|
|
if (*bh_slot)
|
|
|
|
brelse(*bh_slot);
|
|
|
|
*bh_slot = bh;
|
|
|
|
get_bh(bh);
|
|
|
|
spin_unlock(&ip->i_spin);
|
2006-01-16 09:50:04 -07:00
|
|
|
}
|
|
|
|
|
|
|
|
*bhp = bh;
|
|
|
|
return 0;
|
2006-09-21 15:05:23 -06:00
|
|
|
err:
|
|
|
|
brelse(bh);
|
|
|
|
return -EIO;
|
2006-01-16 09:50:04 -07:00
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* gfs2_meta_ra - start readahead on an extent of a file
|
|
|
|
* @gl: the glock the blocks belong to
|
|
|
|
* @dblock: the starting disk block
|
|
|
|
* @extlen: the number of blocks in the extent
|
|
|
|
*
|
2006-09-21 15:05:23 -06:00
|
|
|
* returns: the first buffer in the extent
|
2006-01-16 09:50:04 -07:00
|
|
|
*/
|
|
|
|
|
2006-09-21 15:05:23 -06:00
|
|
|
struct buffer_head *gfs2_meta_ra(struct gfs2_glock *gl, u64 dblock, u32 extlen)
|
2006-01-16 09:50:04 -07:00
|
|
|
{
|
|
|
|
struct gfs2_sbd *sdp = gl->gl_sbd;
|
|
|
|
struct buffer_head *first_bh, *bh;
|
2006-09-04 10:49:07 -06:00
|
|
|
u32 max_ra = gfs2_tune_get(sdp, gt_max_readahead) >>
|
2006-02-27 10:00:42 -07:00
|
|
|
sdp->sd_sb.sb_bsize_shift;
|
2006-01-16 09:50:04 -07:00
|
|
|
|
2006-09-21 15:05:23 -06:00
|
|
|
BUG_ON(!extlen);
|
|
|
|
|
|
|
|
if (max_ra < 1)
|
|
|
|
max_ra = 1;
|
2006-01-16 09:50:04 -07:00
|
|
|
if (extlen > max_ra)
|
|
|
|
extlen = max_ra;
|
|
|
|
|
2006-11-23 09:16:32 -07:00
|
|
|
first_bh = getbuf(gl, dblock, CREATE);
|
2006-01-16 09:50:04 -07:00
|
|
|
|
|
|
|
if (buffer_uptodate(first_bh))
|
|
|
|
goto out;
|
2006-09-21 15:05:23 -06:00
|
|
|
if (!buffer_locked(first_bh))
|
2006-10-02 09:38:25 -06:00
|
|
|
ll_rw_block(READ_META, 1, &first_bh);
|
2006-01-16 09:50:04 -07:00
|
|
|
|
|
|
|
dblock++;
|
|
|
|
extlen--;
|
|
|
|
|
|
|
|
while (extlen) {
|
2006-11-23 09:16:32 -07:00
|
|
|
bh = getbuf(gl, dblock, CREATE);
|
2006-01-16 09:50:04 -07:00
|
|
|
|
2006-09-21 15:05:23 -06:00
|
|
|
if (!buffer_uptodate(bh) && !buffer_locked(bh))
|
|
|
|
ll_rw_block(READA, 1, &bh);
|
|
|
|
brelse(bh);
|
2006-01-16 09:50:04 -07:00
|
|
|
dblock++;
|
|
|
|
extlen--;
|
2006-09-21 15:05:23 -06:00
|
|
|
if (!buffer_locked(first_bh) && buffer_uptodate(first_bh))
|
|
|
|
goto out;
|
2006-01-16 09:50:04 -07:00
|
|
|
}
|
|
|
|
|
2006-09-21 15:05:23 -06:00
|
|
|
wait_on_buffer(first_bh);
|
2006-09-04 10:04:26 -06:00
|
|
|
out:
|
2006-09-21 15:05:23 -06:00
|
|
|
return first_bh;
|
2006-01-16 09:50:04 -07:00
|
|
|
}
|
|
|
|
|