2006-01-16 09:50:04 -07:00
|
|
|
/*
|
|
|
|
* Copyright (C) Sistina Software, Inc. 1997-2003 All rights reserved.
|
2008-01-28 10:24:35 -07:00
|
|
|
* Copyright (C) 2004-2008 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/pagemap.h>
|
2006-05-05 14:59:11 -06:00
|
|
|
#include <linux/pagevec.h>
|
2006-01-30 04:55:32 -07:00
|
|
|
#include <linux/mpage.h>
|
2006-02-14 04:54:42 -07:00
|
|
|
#include <linux/fs.h>
|
2007-01-15 06:52:17 -07:00
|
|
|
#include <linux/writeback.h>
|
2007-10-16 02:25:07 -06:00
|
|
|
#include <linux/swap.h>
|
2006-02-27 15:23:27 -07:00
|
|
|
#include <linux/gfs2_ondisk.h>
|
2007-10-18 04:15:50 -06:00
|
|
|
#include <linux/backing-dev.h>
|
2013-05-07 17:19:08 -06:00
|
|
|
#include <linux/aio.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 "bmap.h"
|
|
|
|
#include "glock.h"
|
|
|
|
#include "inode.h"
|
|
|
|
#include "log.h"
|
|
|
|
#include "meta_io.h"
|
|
|
|
#include "quota.h"
|
|
|
|
#include "trans.h"
|
2006-02-08 04:50:51 -07:00
|
|
|
#include "rgrp.h"
|
2007-05-14 11:42:18 -06:00
|
|
|
#include "super.h"
|
2006-02-27 15:23:27 -07:00
|
|
|
#include "util.h"
|
2006-07-11 07:46:33 -06:00
|
|
|
#include "glops.h"
|
2006-01-16 09:50:04 -07:00
|
|
|
|
2006-07-26 09:27:10 -06:00
|
|
|
|
2012-04-11 10:58:07 -06:00
|
|
|
static void gfs2_page_add_databufs(struct gfs2_inode *ip, struct page *page,
|
|
|
|
unsigned int from, unsigned int to)
|
2006-07-26 09:27:10 -06:00
|
|
|
{
|
|
|
|
struct buffer_head *head = page_buffers(page);
|
|
|
|
unsigned int bsize = head->b_size;
|
|
|
|
struct buffer_head *bh;
|
|
|
|
unsigned int start, end;
|
|
|
|
|
|
|
|
for (bh = head, start = 0; bh != head || !start;
|
|
|
|
bh = bh->b_this_page, start = end) {
|
|
|
|
end = start + bsize;
|
|
|
|
if (end <= from || start >= to)
|
|
|
|
continue;
|
[GFS2] fix jdata issues
This is a patch for the first three issues of RHBZ #238162
The first issue is that when you allocate a new page for a file, it will not
start off uptodate. This makes sense, since you haven't written anything to that
part of the file yet. Unfortunately, gfs2_pin() checks to make sure that the
buffers are uptodate. The solution to this is to mark the buffers uptodate in
gfs2_commit_write(), after they have been zeroed out and have the data written
into them. I'm pretty confident with this fix, although it's not completely
obvious that there is no problem with marking the buffers uptodate here.
The second issue is simply that you can try to pin a data buffer that is already
on the incore log, and thus, already pinned. This patch checks to see if this
buffer is already on the log, and exits databuf_lo_add() if it is, just like
buf_lo_add() does.
The third issue is that gfs2_log_flush() doesn't do it's block accounting
correctly. Both metadata and journaled data are logged, but gfs2_log_flush()
only compares the number of metadata blocks with the number of blocks to commit
to the ondisk journal. This patch also counts the journaled data blocks.
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
2007-06-01 13:21:38 -06:00
|
|
|
if (gfs2_is_jdata(ip))
|
|
|
|
set_buffer_uptodate(bh);
|
2012-12-14 05:36:02 -07:00
|
|
|
gfs2_trans_add_data(ip->i_gl, bh);
|
2006-07-26 09:27:10 -06:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2006-01-16 09:50:04 -07:00
|
|
|
/**
|
2006-09-18 15:18:23 -06:00
|
|
|
* gfs2_get_block_noalloc - Fills in a buffer head with details about a block
|
2006-01-16 09:50:04 -07:00
|
|
|
* @inode: The inode
|
|
|
|
* @lblock: The block number to look up
|
|
|
|
* @bh_result: The buffer head to return the result in
|
|
|
|
* @create: Non-zero if we may add block to the file
|
|
|
|
*
|
|
|
|
* Returns: errno
|
|
|
|
*/
|
|
|
|
|
2006-09-18 15:18:23 -06:00
|
|
|
static int gfs2_get_block_noalloc(struct inode *inode, sector_t lblock,
|
|
|
|
struct buffer_head *bh_result, int create)
|
2006-01-16 09:50:04 -07:00
|
|
|
{
|
|
|
|
int error;
|
|
|
|
|
2007-12-10 13:13:27 -07:00
|
|
|
error = gfs2_block_map(inode, lblock, bh_result, 0);
|
2006-01-16 09:50:04 -07:00
|
|
|
if (error)
|
|
|
|
return error;
|
2007-09-18 07:19:13 -06:00
|
|
|
if (!buffer_mapped(bh_result))
|
2006-09-18 15:18:23 -06:00
|
|
|
return -EIO;
|
|
|
|
return 0;
|
2006-01-16 09:50:04 -07:00
|
|
|
}
|
|
|
|
|
2006-09-18 15:18:23 -06:00
|
|
|
static int gfs2_get_block_direct(struct inode *inode, sector_t lblock,
|
|
|
|
struct buffer_head *bh_result, int create)
|
2006-08-31 10:14:44 -06:00
|
|
|
{
|
2007-12-10 13:13:27 -07:00
|
|
|
return gfs2_block_map(inode, lblock, bh_result, 0);
|
2006-08-31 10:14:44 -06:00
|
|
|
}
|
2006-09-18 15:18:23 -06:00
|
|
|
|
2006-01-16 09:50:04 -07:00
|
|
|
/**
|
2007-09-28 06:49:05 -06:00
|
|
|
* gfs2_writepage_common - Common bits of writepage
|
|
|
|
* @page: The page to be written
|
|
|
|
* @wbc: The writeback control
|
2006-01-16 09:50:04 -07:00
|
|
|
*
|
2007-09-28 06:49:05 -06:00
|
|
|
* Returns: 1 if writepage is ok, otherwise an error code or zero if no error.
|
2006-01-16 09:50:04 -07:00
|
|
|
*/
|
|
|
|
|
2007-09-28 06:49:05 -06:00
|
|
|
static int gfs2_writepage_common(struct page *page,
|
|
|
|
struct writeback_control *wbc)
|
2006-01-16 09:50:04 -07:00
|
|
|
{
|
2006-02-08 04:50:51 -07:00
|
|
|
struct inode *inode = page->mapping->host;
|
2006-08-08 11:23:19 -06:00
|
|
|
struct gfs2_inode *ip = GFS2_I(inode);
|
|
|
|
struct gfs2_sbd *sdp = GFS2_SB(inode);
|
2006-02-08 04:50:51 -07:00
|
|
|
loff_t i_size = i_size_read(inode);
|
|
|
|
pgoff_t end_index = i_size >> PAGE_CACHE_SHIFT;
|
|
|
|
unsigned offset;
|
2006-01-16 09:50:04 -07:00
|
|
|
|
2007-09-28 06:49:05 -06:00
|
|
|
if (gfs2_assert_withdraw(sdp, gfs2_glock_is_held_excl(ip->i_gl)))
|
|
|
|
goto out;
|
2006-02-27 15:23:27 -07:00
|
|
|
if (current->journal_info)
|
2007-09-28 06:49:05 -06:00
|
|
|
goto redirty;
|
2006-02-08 04:50:51 -07:00
|
|
|
/* Is the page fully outside i_size? (truncate in progress) */
|
2007-09-28 06:49:05 -06:00
|
|
|
offset = i_size & (PAGE_CACHE_SIZE-1);
|
2006-05-02 10:09:42 -06:00
|
|
|
if (page->index > end_index || (page->index == end_index && !offset)) {
|
2013-05-21 21:17:23 -06:00
|
|
|
page->mapping->a_ops->invalidatepage(page, 0, PAGE_CACHE_SIZE);
|
2007-09-28 06:49:05 -06:00
|
|
|
goto out;
|
2006-01-16 09:50:04 -07:00
|
|
|
}
|
2007-09-28 06:49:05 -06:00
|
|
|
return 1;
|
|
|
|
redirty:
|
|
|
|
redirty_page_for_writepage(wbc, page);
|
|
|
|
out:
|
|
|
|
unlock_page(page);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
2013-08-27 14:22:07 -06:00
|
|
|
* gfs2_writepage - Write page for writeback mappings
|
2007-09-28 06:49:05 -06:00
|
|
|
* @page: The page
|
|
|
|
* @wbc: The writeback control
|
|
|
|
*
|
|
|
|
*/
|
|
|
|
|
2013-08-27 14:22:07 -06:00
|
|
|
static int gfs2_writepage(struct page *page, struct writeback_control *wbc)
|
2007-09-28 06:49:05 -06:00
|
|
|
{
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
ret = gfs2_writepage_common(page, wbc);
|
|
|
|
if (ret <= 0)
|
|
|
|
return ret;
|
|
|
|
|
2010-06-14 02:58:41 -06:00
|
|
|
return nobh_writepage(page, gfs2_get_block_noalloc, wbc);
|
2007-09-28 06:49:05 -06:00
|
|
|
}
|
|
|
|
|
2007-10-17 02:04:24 -06:00
|
|
|
/**
|
|
|
|
* __gfs2_jdata_writepage - The core of jdata writepage
|
|
|
|
* @page: The page to write
|
|
|
|
* @wbc: The writeback control
|
|
|
|
*
|
|
|
|
* This is shared between writepage and writepages and implements the
|
|
|
|
* core of the writepage operation. If a transaction is required then
|
|
|
|
* PageChecked will have been set and the transaction will have
|
|
|
|
* already been started before this is called.
|
|
|
|
*/
|
|
|
|
|
|
|
|
static int __gfs2_jdata_writepage(struct page *page, struct writeback_control *wbc)
|
|
|
|
{
|
|
|
|
struct inode *inode = page->mapping->host;
|
|
|
|
struct gfs2_inode *ip = GFS2_I(inode);
|
|
|
|
struct gfs2_sbd *sdp = GFS2_SB(inode);
|
|
|
|
|
|
|
|
if (PageChecked(page)) {
|
|
|
|
ClearPageChecked(page);
|
|
|
|
if (!page_has_buffers(page)) {
|
|
|
|
create_empty_buffers(page, inode->i_sb->s_blocksize,
|
|
|
|
(1 << BH_Dirty)|(1 << BH_Uptodate));
|
|
|
|
}
|
|
|
|
gfs2_page_add_databufs(ip, page, 0, sdp->sd_vfs->s_blocksize-1);
|
|
|
|
}
|
|
|
|
return block_write_full_page(page, gfs2_get_block_noalloc, wbc);
|
|
|
|
}
|
|
|
|
|
2007-09-28 06:49:05 -06:00
|
|
|
/**
|
|
|
|
* gfs2_jdata_writepage - Write complete page
|
|
|
|
* @page: Page to write
|
|
|
|
*
|
|
|
|
* Returns: errno
|
|
|
|
*
|
|
|
|
*/
|
|
|
|
|
|
|
|
static int gfs2_jdata_writepage(struct page *page, struct writeback_control *wbc)
|
|
|
|
{
|
|
|
|
struct inode *inode = page->mapping->host;
|
|
|
|
struct gfs2_sbd *sdp = GFS2_SB(inode);
|
2008-10-15 02:46:39 -06:00
|
|
|
int ret;
|
2007-09-28 06:49:05 -06:00
|
|
|
int done_trans = 0;
|
|
|
|
|
2007-10-17 01:35:19 -06:00
|
|
|
if (PageChecked(page)) {
|
2007-10-17 02:04:24 -06:00
|
|
|
if (wbc->sync_mode != WB_SYNC_ALL)
|
|
|
|
goto out_ignore;
|
2008-10-15 02:46:39 -06:00
|
|
|
ret = gfs2_trans_begin(sdp, RES_DINODE + 1, 0);
|
|
|
|
if (ret)
|
2006-02-08 04:50:51 -07:00
|
|
|
goto out_ignore;
|
|
|
|
done_trans = 1;
|
|
|
|
}
|
2008-10-15 02:46:39 -06:00
|
|
|
ret = gfs2_writepage_common(page, wbc);
|
|
|
|
if (ret > 0)
|
|
|
|
ret = __gfs2_jdata_writepage(page, wbc);
|
2006-02-08 04:50:51 -07:00
|
|
|
if (done_trans)
|
|
|
|
gfs2_trans_end(sdp);
|
2008-10-15 02:46:39 -06:00
|
|
|
return ret;
|
2006-02-08 04:50:51 -07:00
|
|
|
|
|
|
|
out_ignore:
|
|
|
|
redirty_page_for_writepage(wbc, page);
|
|
|
|
unlock_page(page);
|
|
|
|
return 0;
|
2006-01-16 09:50:04 -07:00
|
|
|
}
|
|
|
|
|
2007-01-15 06:52:17 -07:00
|
|
|
/**
|
2013-01-28 02:30:07 -07:00
|
|
|
* gfs2_writepages - Write a bunch of dirty pages back to disk
|
2007-01-15 06:52:17 -07:00
|
|
|
* @mapping: The mapping to write
|
|
|
|
* @wbc: Write-back control
|
|
|
|
*
|
2013-01-28 02:30:07 -07:00
|
|
|
* Used for both ordered and writeback modes.
|
2007-01-15 06:52:17 -07:00
|
|
|
*/
|
2013-01-28 02:30:07 -07:00
|
|
|
static int gfs2_writepages(struct address_space *mapping,
|
|
|
|
struct writeback_control *wbc)
|
2007-01-15 06:52:17 -07:00
|
|
|
{
|
2007-10-17 01:47:38 -06:00
|
|
|
return mpage_writepages(mapping, wbc, gfs2_get_block_noalloc);
|
2007-01-15 06:52:17 -07:00
|
|
|
}
|
|
|
|
|
2007-10-17 02:04:24 -06:00
|
|
|
/**
|
|
|
|
* gfs2_write_jdata_pagevec - Write back a pagevec's worth of pages
|
|
|
|
* @mapping: The mapping
|
|
|
|
* @wbc: The writeback control
|
|
|
|
* @writepage: The writepage function to call for each page
|
|
|
|
* @pvec: The vector of pages
|
|
|
|
* @nr_pages: The number of pages to write
|
|
|
|
*
|
|
|
|
* Returns: non-zero if loop should terminate, zero otherwise
|
|
|
|
*/
|
|
|
|
|
|
|
|
static int gfs2_write_jdata_pagevec(struct address_space *mapping,
|
|
|
|
struct writeback_control *wbc,
|
|
|
|
struct pagevec *pvec,
|
|
|
|
int nr_pages, pgoff_t end)
|
|
|
|
{
|
|
|
|
struct inode *inode = mapping->host;
|
|
|
|
struct gfs2_sbd *sdp = GFS2_SB(inode);
|
|
|
|
loff_t i_size = i_size_read(inode);
|
|
|
|
pgoff_t end_index = i_size >> PAGE_CACHE_SHIFT;
|
|
|
|
unsigned offset = i_size & (PAGE_CACHE_SIZE-1);
|
|
|
|
unsigned nrblocks = nr_pages * (PAGE_CACHE_SIZE/inode->i_sb->s_blocksize);
|
|
|
|
int i;
|
|
|
|
int ret;
|
|
|
|
|
2008-03-06 16:43:52 -07:00
|
|
|
ret = gfs2_trans_begin(sdp, nrblocks, nrblocks);
|
2007-10-17 02:04:24 -06:00
|
|
|
if (ret < 0)
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
for(i = 0; i < nr_pages; i++) {
|
|
|
|
struct page *page = pvec->pages[i];
|
|
|
|
|
|
|
|
lock_page(page);
|
|
|
|
|
|
|
|
if (unlikely(page->mapping != mapping)) {
|
|
|
|
unlock_page(page);
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (!wbc->range_cyclic && page->index > end) {
|
|
|
|
ret = 1;
|
|
|
|
unlock_page(page);
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (wbc->sync_mode != WB_SYNC_NONE)
|
|
|
|
wait_on_page_writeback(page);
|
|
|
|
|
|
|
|
if (PageWriteback(page) ||
|
|
|
|
!clear_page_dirty_for_io(page)) {
|
|
|
|
unlock_page(page);
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Is the page fully outside i_size? (truncate in progress) */
|
|
|
|
if (page->index > end_index || (page->index == end_index && !offset)) {
|
2013-05-21 21:17:23 -06:00
|
|
|
page->mapping->a_ops->invalidatepage(page, 0,
|
|
|
|
PAGE_CACHE_SIZE);
|
2007-10-17 02:04:24 -06:00
|
|
|
unlock_page(page);
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
|
|
|
ret = __gfs2_jdata_writepage(page, wbc);
|
|
|
|
|
|
|
|
if (ret || (--(wbc->nr_to_write) <= 0))
|
|
|
|
ret = 1;
|
|
|
|
}
|
|
|
|
gfs2_trans_end(sdp);
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* gfs2_write_cache_jdata - Like write_cache_pages but different
|
|
|
|
* @mapping: The mapping to write
|
|
|
|
* @wbc: The writeback control
|
|
|
|
* @writepage: The writepage function to call
|
|
|
|
* @data: The data to pass to writepage
|
|
|
|
*
|
|
|
|
* The reason that we use our own function here is that we need to
|
|
|
|
* start transactions before we grab page locks. This allows us
|
|
|
|
* to get the ordering right.
|
|
|
|
*/
|
|
|
|
|
|
|
|
static int gfs2_write_cache_jdata(struct address_space *mapping,
|
|
|
|
struct writeback_control *wbc)
|
|
|
|
{
|
|
|
|
int ret = 0;
|
|
|
|
int done = 0;
|
|
|
|
struct pagevec pvec;
|
|
|
|
int nr_pages;
|
|
|
|
pgoff_t index;
|
|
|
|
pgoff_t end;
|
|
|
|
int scanned = 0;
|
|
|
|
int range_whole = 0;
|
|
|
|
|
|
|
|
pagevec_init(&pvec, 0);
|
|
|
|
if (wbc->range_cyclic) {
|
|
|
|
index = mapping->writeback_index; /* Start from prev offset */
|
|
|
|
end = -1;
|
|
|
|
} else {
|
|
|
|
index = wbc->range_start >> PAGE_CACHE_SHIFT;
|
|
|
|
end = wbc->range_end >> PAGE_CACHE_SHIFT;
|
|
|
|
if (wbc->range_start == 0 && wbc->range_end == LLONG_MAX)
|
|
|
|
range_whole = 1;
|
|
|
|
scanned = 1;
|
|
|
|
}
|
|
|
|
|
|
|
|
retry:
|
|
|
|
while (!done && (index <= end) &&
|
|
|
|
(nr_pages = pagevec_lookup_tag(&pvec, mapping, &index,
|
|
|
|
PAGECACHE_TAG_DIRTY,
|
|
|
|
min(end - index, (pgoff_t)PAGEVEC_SIZE-1) + 1))) {
|
|
|
|
scanned = 1;
|
|
|
|
ret = gfs2_write_jdata_pagevec(mapping, wbc, &pvec, nr_pages, end);
|
|
|
|
if (ret)
|
|
|
|
done = 1;
|
|
|
|
if (ret > 0)
|
|
|
|
ret = 0;
|
|
|
|
|
|
|
|
pagevec_release(&pvec);
|
|
|
|
cond_resched();
|
|
|
|
}
|
|
|
|
|
|
|
|
if (!scanned && !done) {
|
|
|
|
/*
|
|
|
|
* We hit the last page and there is more work to be done: wrap
|
|
|
|
* back to the start of the file
|
|
|
|
*/
|
|
|
|
scanned = 1;
|
|
|
|
index = 0;
|
|
|
|
goto retry;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (wbc->range_cyclic || (range_whole && wbc->nr_to_write > 0))
|
|
|
|
mapping->writeback_index = index;
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
* gfs2_jdata_writepages - Write a bunch of dirty pages back to disk
|
|
|
|
* @mapping: The mapping to write
|
|
|
|
* @wbc: The writeback control
|
|
|
|
*
|
|
|
|
*/
|
|
|
|
|
|
|
|
static int gfs2_jdata_writepages(struct address_space *mapping,
|
|
|
|
struct writeback_control *wbc)
|
|
|
|
{
|
|
|
|
struct gfs2_inode *ip = GFS2_I(mapping->host);
|
|
|
|
struct gfs2_sbd *sdp = GFS2_SB(mapping->host);
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
ret = gfs2_write_cache_jdata(mapping, wbc);
|
|
|
|
if (ret == 0 && wbc->sync_mode == WB_SYNC_ALL) {
|
|
|
|
gfs2_log_flush(sdp, ip->i_gl);
|
|
|
|
ret = gfs2_write_cache_jdata(mapping, wbc);
|
|
|
|
}
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
2006-01-16 09:50:04 -07:00
|
|
|
/**
|
|
|
|
* stuffed_readpage - Fill in a Linux page with stuffed file data
|
|
|
|
* @ip: the inode
|
|
|
|
* @page: the page
|
|
|
|
*
|
|
|
|
* Returns: errno
|
|
|
|
*/
|
|
|
|
|
|
|
|
static int stuffed_readpage(struct gfs2_inode *ip, struct page *page)
|
|
|
|
{
|
|
|
|
struct buffer_head *dibh;
|
2010-03-25 08:32:43 -06:00
|
|
|
u64 dsize = i_size_read(&ip->i_inode);
|
2006-01-16 09:50:04 -07:00
|
|
|
void *kaddr;
|
|
|
|
int error;
|
|
|
|
|
2007-04-20 02:18:30 -06:00
|
|
|
/*
|
2008-04-28 03:12:10 -06:00
|
|
|
* Due to the order of unstuffing files and ->fault(), we can be
|
2007-04-20 02:18:30 -06:00
|
|
|
* asked for a zero page in the case of a stuffed file being extended,
|
|
|
|
* so we need to supply one here. It doesn't happen often.
|
|
|
|
*/
|
|
|
|
if (unlikely(page->index)) {
|
2008-02-04 23:28:29 -07:00
|
|
|
zero_user(page, 0, PAGE_CACHE_SIZE);
|
2009-01-07 15:03:37 -07:00
|
|
|
SetPageUptodate(page);
|
2007-04-20 02:18:30 -06:00
|
|
|
return 0;
|
|
|
|
}
|
2006-05-05 14:59:11 -06:00
|
|
|
|
2006-01-16 09:50:04 -07:00
|
|
|
error = gfs2_meta_inode_buffer(ip, &dibh);
|
|
|
|
if (error)
|
|
|
|
return error;
|
|
|
|
|
2011-11-25 08:14:30 -07:00
|
|
|
kaddr = kmap_atomic(page);
|
2010-03-25 08:32:43 -06:00
|
|
|
if (dsize > (dibh->b_size - sizeof(struct gfs2_dinode)))
|
|
|
|
dsize = (dibh->b_size - sizeof(struct gfs2_dinode));
|
|
|
|
memcpy(kaddr, dibh->b_data + sizeof(struct gfs2_dinode), dsize);
|
|
|
|
memset(kaddr + dsize, 0, PAGE_CACHE_SIZE - dsize);
|
2011-11-25 08:14:30 -07:00
|
|
|
kunmap_atomic(kaddr);
|
2007-04-20 02:18:30 -06:00
|
|
|
flush_dcache_page(page);
|
2006-01-16 09:50:04 -07:00
|
|
|
brelse(dibh);
|
|
|
|
SetPageUptodate(page);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/**
|
2007-10-15 07:42:35 -06:00
|
|
|
* __gfs2_readpage - readpage
|
|
|
|
* @file: The file to read a page for
|
2006-01-16 09:50:04 -07:00
|
|
|
* @page: The page to read
|
|
|
|
*
|
2007-10-15 07:42:35 -06:00
|
|
|
* This is the core of gfs2's readpage. Its used by the internal file
|
|
|
|
* reading code as in that case we already hold the glock. Also its
|
|
|
|
* called by gfs2_readpage() once the required lock has been granted.
|
|
|
|
*
|
2006-01-16 09:50:04 -07:00
|
|
|
*/
|
|
|
|
|
2007-10-15 07:42:35 -06:00
|
|
|
static int __gfs2_readpage(void *file, struct page *page)
|
2006-01-16 09:50:04 -07:00
|
|
|
{
|
2006-06-14 13:32:57 -06:00
|
|
|
struct gfs2_inode *ip = GFS2_I(page->mapping->host);
|
|
|
|
struct gfs2_sbd *sdp = GFS2_SB(page->mapping->host);
|
2006-01-16 09:50:04 -07:00
|
|
|
int error;
|
|
|
|
|
2006-02-08 04:50:51 -07:00
|
|
|
if (gfs2_is_stuffed(ip)) {
|
2006-05-05 14:59:11 -06:00
|
|
|
error = stuffed_readpage(ip, page);
|
|
|
|
unlock_page(page);
|
2007-10-15 07:42:35 -06:00
|
|
|
} else {
|
2007-12-10 13:13:27 -07:00
|
|
|
error = mpage_readpage(page, gfs2_block_map);
|
2007-10-15 07:42:35 -06:00
|
|
|
}
|
2006-01-16 09:50:04 -07:00
|
|
|
|
|
|
|
if (unlikely(test_bit(SDF_SHUTDOWN, &sdp->sd_flags)))
|
2007-10-15 07:42:35 -06:00
|
|
|
return -EIO;
|
2006-01-16 09:50:04 -07:00
|
|
|
|
2007-10-15 07:42:35 -06:00
|
|
|
return error;
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* gfs2_readpage - read a page of a file
|
|
|
|
* @file: The file to read
|
|
|
|
* @page: The page of the file
|
|
|
|
*
|
2008-06-02 02:14:54 -06:00
|
|
|
* This deals with the locking required. We have to unlock and
|
|
|
|
* relock the page in order to get the locking in the right
|
|
|
|
* order.
|
2007-10-15 07:42:35 -06:00
|
|
|
*/
|
|
|
|
|
|
|
|
static int gfs2_readpage(struct file *file, struct page *page)
|
|
|
|
{
|
2008-06-02 02:14:54 -06:00
|
|
|
struct address_space *mapping = page->mapping;
|
|
|
|
struct gfs2_inode *ip = GFS2_I(mapping->host);
|
2008-05-21 10:03:22 -06:00
|
|
|
struct gfs2_holder gh;
|
2007-10-15 07:42:35 -06:00
|
|
|
int error;
|
|
|
|
|
2008-06-02 02:14:54 -06:00
|
|
|
unlock_page(page);
|
2008-09-18 06:53:59 -06:00
|
|
|
gfs2_holder_init(ip->i_gl, LM_ST_SHARED, 0, &gh);
|
|
|
|
error = gfs2_glock_nq(&gh);
|
2008-06-02 02:14:54 -06:00
|
|
|
if (unlikely(error))
|
2008-05-21 10:03:22 -06:00
|
|
|
goto out;
|
2008-06-02 02:14:54 -06:00
|
|
|
error = AOP_TRUNCATED_PAGE;
|
|
|
|
lock_page(page);
|
|
|
|
if (page->mapping == mapping && !PageUptodate(page))
|
|
|
|
error = __gfs2_readpage(file, page);
|
|
|
|
else
|
|
|
|
unlock_page(page);
|
2008-05-21 10:03:22 -06:00
|
|
|
gfs2_glock_dq(&gh);
|
2006-02-08 04:50:51 -07:00
|
|
|
out:
|
2008-05-21 10:03:22 -06:00
|
|
|
gfs2_holder_uninit(&gh);
|
2008-06-02 02:14:54 -06:00
|
|
|
if (error && error != AOP_TRUNCATED_PAGE)
|
|
|
|
lock_page(page);
|
2007-10-15 07:42:35 -06:00
|
|
|
return error;
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* gfs2_internal_read - read an internal file
|
|
|
|
* @ip: The gfs2 inode
|
|
|
|
* @buf: The buffer to fill
|
|
|
|
* @pos: The file position
|
|
|
|
* @size: The amount to read
|
|
|
|
*
|
|
|
|
*/
|
|
|
|
|
2012-04-16 09:40:55 -06:00
|
|
|
int gfs2_internal_read(struct gfs2_inode *ip, char *buf, loff_t *pos,
|
|
|
|
unsigned size)
|
2007-10-15 07:42:35 -06:00
|
|
|
{
|
|
|
|
struct address_space *mapping = ip->i_inode.i_mapping;
|
|
|
|
unsigned long index = *pos / PAGE_CACHE_SIZE;
|
|
|
|
unsigned offset = *pos & (PAGE_CACHE_SIZE - 1);
|
|
|
|
unsigned copied = 0;
|
|
|
|
unsigned amt;
|
|
|
|
struct page *page;
|
|
|
|
void *p;
|
|
|
|
|
|
|
|
do {
|
|
|
|
amt = size - copied;
|
|
|
|
if (offset + size > PAGE_CACHE_SIZE)
|
|
|
|
amt = PAGE_CACHE_SIZE - offset;
|
|
|
|
page = read_cache_page(mapping, index, __gfs2_readpage, NULL);
|
|
|
|
if (IS_ERR(page))
|
|
|
|
return PTR_ERR(page);
|
2011-11-25 08:14:30 -07:00
|
|
|
p = kmap_atomic(page);
|
2007-10-15 07:42:35 -06:00
|
|
|
memcpy(buf + copied, p + offset, amt);
|
2011-11-25 08:14:30 -07:00
|
|
|
kunmap_atomic(p);
|
2007-10-15 07:42:35 -06:00
|
|
|
mark_page_accessed(page);
|
|
|
|
page_cache_release(page);
|
|
|
|
copied += amt;
|
|
|
|
index++;
|
|
|
|
offset = 0;
|
|
|
|
} while(copied < size);
|
|
|
|
(*pos) += size;
|
|
|
|
return size;
|
2006-05-05 14:59:11 -06:00
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* gfs2_readpages - Read a bunch of pages at once
|
|
|
|
*
|
|
|
|
* Some notes:
|
|
|
|
* 1. This is only for readahead, so we can simply ignore any things
|
|
|
|
* which are slightly inconvenient (such as locking conflicts between
|
|
|
|
* the page lock and the glock) and return having done no I/O. Its
|
|
|
|
* obviously not something we'd want to do on too regular a basis.
|
|
|
|
* Any I/O we ignore at this time will be done via readpage later.
|
2006-12-15 14:49:51 -07:00
|
|
|
* 2. We don't handle stuffed files here we let readpage do the honours.
|
2006-05-05 14:59:11 -06:00
|
|
|
* 3. mpage_readpages() does most of the heavy lifting in the common case.
|
2007-12-10 13:13:27 -07:00
|
|
|
* 4. gfs2_block_map() is relied upon to set BH_Boundary in the right places.
|
2006-05-05 14:59:11 -06:00
|
|
|
*/
|
2007-10-15 08:40:33 -06:00
|
|
|
|
2006-05-05 14:59:11 -06:00
|
|
|
static int gfs2_readpages(struct file *file, struct address_space *mapping,
|
|
|
|
struct list_head *pages, unsigned nr_pages)
|
|
|
|
{
|
|
|
|
struct inode *inode = mapping->host;
|
2006-06-14 13:32:57 -06:00
|
|
|
struct gfs2_inode *ip = GFS2_I(inode);
|
|
|
|
struct gfs2_sbd *sdp = GFS2_SB(inode);
|
2006-05-05 14:59:11 -06:00
|
|
|
struct gfs2_holder gh;
|
2007-10-15 08:40:33 -06:00
|
|
|
int ret;
|
2006-05-05 14:59:11 -06:00
|
|
|
|
2008-09-18 06:53:59 -06:00
|
|
|
gfs2_holder_init(ip->i_gl, LM_ST_SHARED, 0, &gh);
|
|
|
|
ret = gfs2_glock_nq(&gh);
|
2007-10-15 07:42:35 -06:00
|
|
|
if (unlikely(ret))
|
2007-10-15 08:40:33 -06:00
|
|
|
goto out_uninit;
|
2006-12-15 14:49:51 -07:00
|
|
|
if (!gfs2_is_stuffed(ip))
|
2007-12-10 13:13:27 -07:00
|
|
|
ret = mpage_readpages(mapping, pages, nr_pages, gfs2_block_map);
|
2007-10-15 08:40:33 -06:00
|
|
|
gfs2_glock_dq(&gh);
|
|
|
|
out_uninit:
|
|
|
|
gfs2_holder_uninit(&gh);
|
2006-05-05 14:59:11 -06:00
|
|
|
if (unlikely(test_bit(SDF_SHUTDOWN, &sdp->sd_flags)))
|
|
|
|
ret = -EIO;
|
|
|
|
return ret;
|
2006-01-16 09:50:04 -07:00
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
2007-10-16 02:25:07 -06:00
|
|
|
* gfs2_write_begin - Begin to write to a file
|
2006-01-16 09:50:04 -07:00
|
|
|
* @file: The file to write to
|
2007-10-16 02:25:07 -06:00
|
|
|
* @mapping: The mapping in which to write
|
|
|
|
* @pos: The file offset at which to start writing
|
|
|
|
* @len: Length of the write
|
|
|
|
* @flags: Various flags
|
|
|
|
* @pagep: Pointer to return the page
|
|
|
|
* @fsdata: Pointer to return fs data (unused by GFS2)
|
2006-01-16 09:50:04 -07:00
|
|
|
*
|
|
|
|
* Returns: errno
|
|
|
|
*/
|
|
|
|
|
2007-10-16 02:25:07 -06:00
|
|
|
static int gfs2_write_begin(struct file *file, struct address_space *mapping,
|
|
|
|
loff_t pos, unsigned len, unsigned flags,
|
|
|
|
struct page **pagep, void **fsdata)
|
2006-01-16 09:50:04 -07:00
|
|
|
{
|
2007-10-16 02:25:07 -06:00
|
|
|
struct gfs2_inode *ip = GFS2_I(mapping->host);
|
|
|
|
struct gfs2_sbd *sdp = GFS2_SB(mapping->host);
|
2009-06-25 14:09:51 -06:00
|
|
|
struct gfs2_inode *m_ip = GFS2_I(sdp->sd_statfs_inode);
|
2008-12-10 03:28:10 -07:00
|
|
|
unsigned int data_blocks = 0, ind_blocks = 0, rblocks;
|
2012-07-30 07:53:19 -06:00
|
|
|
unsigned requested = 0;
|
2006-02-08 04:50:51 -07:00
|
|
|
int alloc_required;
|
2006-01-16 09:50:04 -07:00
|
|
|
int error = 0;
|
2007-10-16 02:25:07 -06:00
|
|
|
pgoff_t index = pos >> PAGE_CACHE_SHIFT;
|
|
|
|
unsigned from = pos & (PAGE_CACHE_SIZE - 1);
|
|
|
|
struct page *page;
|
2006-10-09 11:11:54 -06:00
|
|
|
|
2008-09-18 06:53:59 -06:00
|
|
|
gfs2_holder_init(ip->i_gl, LM_ST_EXCLUSIVE, 0, &ip->i_gh);
|
|
|
|
error = gfs2_glock_nq(&ip->i_gh);
|
2007-10-16 02:25:07 -06:00
|
|
|
if (unlikely(error))
|
2006-02-08 04:50:51 -07:00
|
|
|
goto out_uninit;
|
2009-06-25 14:09:51 -06:00
|
|
|
if (&ip->i_inode == sdp->sd_rindex) {
|
|
|
|
error = gfs2_glock_nq_init(m_ip->i_gl, LM_ST_EXCLUSIVE,
|
|
|
|
GL_NOCACHE, &m_ip->i_gh);
|
|
|
|
if (unlikely(error)) {
|
|
|
|
gfs2_glock_dq(&ip->i_gh);
|
|
|
|
goto out_uninit;
|
|
|
|
}
|
|
|
|
}
|
2006-01-16 09:50:04 -07:00
|
|
|
|
2010-06-24 17:21:20 -06:00
|
|
|
alloc_required = gfs2_write_alloc_required(ip, pos, len);
|
2006-02-08 04:50:51 -07:00
|
|
|
|
2008-12-10 03:28:10 -07:00
|
|
|
if (alloc_required || gfs2_is_jdata(ip))
|
|
|
|
gfs2_write_calc_reserv(ip, len, &data_blocks, &ind_blocks);
|
|
|
|
|
2006-02-08 04:50:51 -07:00
|
|
|
if (alloc_required) {
|
2013-10-02 04:13:25 -06:00
|
|
|
struct gfs2_alloc_parms ap = { .aflags = 0, };
|
2008-03-10 09:34:50 -06:00
|
|
|
error = gfs2_quota_lock_check(ip);
|
2006-02-08 04:50:51 -07:00
|
|
|
if (error)
|
2012-05-18 07:28:23 -06:00
|
|
|
goto out_unlock;
|
2006-02-08 04:50:51 -07:00
|
|
|
|
2012-07-30 07:53:19 -06:00
|
|
|
requested = data_blocks + ind_blocks;
|
2013-10-02 04:13:25 -06:00
|
|
|
ap.target = requested;
|
|
|
|
error = gfs2_inplace_reserve(ip, &ap);
|
2006-02-08 04:50:51 -07:00
|
|
|
if (error)
|
|
|
|
goto out_qunlock;
|
|
|
|
}
|
|
|
|
|
|
|
|
rblocks = RES_DINODE + ind_blocks;
|
|
|
|
if (gfs2_is_jdata(ip))
|
|
|
|
rblocks += data_blocks ? data_blocks : 1;
|
|
|
|
if (ind_blocks || data_blocks)
|
|
|
|
rblocks += RES_STATFS + RES_QUOTA;
|
2009-06-25 14:09:51 -06:00
|
|
|
if (&ip->i_inode == sdp->sd_rindex)
|
|
|
|
rblocks += 2 * RES_STATFS;
|
2010-09-27 15:00:04 -06:00
|
|
|
if (alloc_required)
|
2012-07-30 07:53:19 -06:00
|
|
|
rblocks += gfs2_rg_blocks(ip, requested);
|
2006-02-08 04:50:51 -07:00
|
|
|
|
2007-09-17 03:59:52 -06:00
|
|
|
error = gfs2_trans_begin(sdp, rblocks,
|
|
|
|
PAGE_CACHE_SIZE/sdp->sd_sb.sb_bsize);
|
2006-02-08 04:50:51 -07:00
|
|
|
if (error)
|
2007-07-17 03:29:02 -06:00
|
|
|
goto out_trans_fail;
|
2006-02-08 04:50:51 -07:00
|
|
|
|
2007-10-17 07:05:41 -06:00
|
|
|
error = -ENOMEM;
|
2009-01-06 03:08:33 -07:00
|
|
|
flags |= AOP_FLAG_NOFS;
|
fs: symlink write_begin allocation context fix
With the write_begin/write_end aops, page_symlink was broken because it
could no longer pass a GFP_NOFS type mask into the point where the
allocations happened. They are done in write_begin, which would always
assume that the filesystem can be entered from reclaim. This bug could
cause filesystem deadlocks.
The funny thing with having a gfp_t mask there is that it doesn't really
allow the caller to arbitrarily tinker with the context in which it can be
called. It couldn't ever be GFP_ATOMIC, for example, because it needs to
take the page lock. The only thing any callers care about is __GFP_FS
anyway, so turn that into a single flag.
Add a new flag for write_begin, AOP_FLAG_NOFS. Filesystems can now act on
this flag in their write_begin function. Change __grab_cache_page to
accept a nofs argument as well, to honour that flag (while we're there,
change the name to grab_cache_page_write_begin which is more instructive
and does away with random leading underscores).
This is really a more flexible way to go in the end anyway -- if a
filesystem happens to want any extra allocations aside from the pagecache
ones in ints write_begin function, it may now use GFP_KERNEL (rather than
GFP_NOFS) for common case allocations (eg. ocfs2_alloc_write_ctxt, for a
random example).
[kosaki.motohiro@jp.fujitsu.com: fix ubifs]
[kosaki.motohiro@jp.fujitsu.com: fix fuse]
Signed-off-by: Nick Piggin <npiggin@suse.de>
Reviewed-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: <stable@kernel.org> [2.6.28.x]
Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
[ Cleaned up the calling convention: just pass in the AOP flags
untouched to the grab_cache_page_write_begin() function. That
just simplifies everybody, and may even allow future expansion of the
logic. - Linus ]
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-04 13:00:53 -07:00
|
|
|
page = grab_cache_page_write_begin(mapping, index, flags);
|
2007-10-17 07:05:41 -06:00
|
|
|
*pagep = page;
|
|
|
|
if (unlikely(!page))
|
|
|
|
goto out_endtrans;
|
|
|
|
|
2006-02-08 04:50:51 -07:00
|
|
|
if (gfs2_is_stuffed(ip)) {
|
2007-10-17 07:05:41 -06:00
|
|
|
error = 0;
|
2007-10-16 02:25:07 -06:00
|
|
|
if (pos + len > sdp->sd_sb.sb_bsize - sizeof(struct gfs2_dinode)) {
|
2006-07-26 08:51:20 -06:00
|
|
|
error = gfs2_unstuff_dinode(ip, page);
|
2006-02-15 05:26:19 -07:00
|
|
|
if (error == 0)
|
|
|
|
goto prepare_write;
|
2007-10-17 07:05:41 -06:00
|
|
|
} else if (!PageUptodate(page)) {
|
2006-01-16 09:50:04 -07:00
|
|
|
error = stuffed_readpage(ip, page);
|
2007-10-17 07:05:41 -06:00
|
|
|
}
|
2006-02-15 05:26:19 -07:00
|
|
|
goto out;
|
2006-02-08 04:50:51 -07:00
|
|
|
}
|
|
|
|
|
2006-02-15 05:26:19 -07:00
|
|
|
prepare_write:
|
2010-10-06 02:47:23 -06:00
|
|
|
error = __block_write_begin(page, from, len, gfs2_block_map);
|
2006-02-08 04:50:51 -07:00
|
|
|
out:
|
2007-10-17 07:05:41 -06:00
|
|
|
if (error == 0)
|
|
|
|
return 0;
|
|
|
|
|
2011-03-14 07:19:21 -06:00
|
|
|
unlock_page(page);
|
2007-10-17 07:05:41 -06:00
|
|
|
page_cache_release(page);
|
2010-05-26 09:05:34 -06:00
|
|
|
|
2010-08-11 02:37:53 -06:00
|
|
|
gfs2_trans_end(sdp);
|
2007-10-17 07:05:41 -06:00
|
|
|
if (pos + len > ip->i_inode.i_size)
|
2010-08-11 02:37:53 -06:00
|
|
|
gfs2_trim_blocks(&ip->i_inode);
|
|
|
|
goto out_trans_fail;
|
|
|
|
|
2007-10-17 07:05:41 -06:00
|
|
|
out_endtrans:
|
|
|
|
gfs2_trans_end(sdp);
|
2007-07-17 03:29:02 -06:00
|
|
|
out_trans_fail:
|
2007-10-17 07:05:41 -06:00
|
|
|
if (alloc_required) {
|
|
|
|
gfs2_inplace_release(ip);
|
2006-02-08 04:50:51 -07:00
|
|
|
out_qunlock:
|
2007-10-17 07:05:41 -06:00
|
|
|
gfs2_quota_unlock(ip);
|
|
|
|
}
|
2006-02-08 04:50:51 -07:00
|
|
|
out_unlock:
|
2009-06-25 14:09:51 -06:00
|
|
|
if (&ip->i_inode == sdp->sd_rindex) {
|
|
|
|
gfs2_glock_dq(&m_ip->i_gh);
|
|
|
|
gfs2_holder_uninit(&m_ip->i_gh);
|
|
|
|
}
|
2007-10-17 07:05:41 -06:00
|
|
|
gfs2_glock_dq(&ip->i_gh);
|
2006-02-08 04:50:51 -07:00
|
|
|
out_uninit:
|
2007-10-17 07:05:41 -06:00
|
|
|
gfs2_holder_uninit(&ip->i_gh);
|
2006-01-16 09:50:04 -07:00
|
|
|
return error;
|
|
|
|
}
|
|
|
|
|
[GFS2] kernel changes to support new gfs2_grow command
This is another revision of my gfs2 kernel patch that allows
gfs2_grow to function properly.
Steve Whitehouse expressed some concerns about the previous
patch and I restructured it based on his comments.
The previous patch was doing the statfs_change at file close time,
under its own transaction. The current patch does the statfs_change
inside the gfs2_commit_write function, which keeps it under the
umbrella of the inode transaction.
I can't call ri_update to re-read the rindex file during the
transaction because the transaction may have outstanding unwritten
buffers attached to the rgrps that would be otherwise blown away.
So instead, I created a new function, gfs2_ri_total, that will
re-read the rindex file just to total the file system space
for the sake of the statfs_change. The ri_update will happen
later, when gfs2 realizes the version number has changed, as it
happened before my patch.
Since the statfs_change is happening at write_commit time and there
may be multiple writes to the rindex file for one grow operation.
So one consequence of this restructuring is that instead of getting
one kernel message to indicate the change, you may see several.
For example, before when you did a gfs2_grow, you'd get a single
message like:
GFS2: File system extended by 247876 blocks (968MB)
Now you get something like:
GFS2: File system extended by 207896 blocks (812MB)
GFS2: File system extended by 39980 blocks (156MB)
This version has also been successfully run against the hours-long
"gfs2_fsck_hellfire" test that does several gfs2_grow and gfs2_fsck
while interjecting file system damage. It does this repeatedly
under a variety Resource Group conditions.
Signed-off-By: Bob Peterson <rpeterso@redhat.com>
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
2007-05-09 08:37:57 -06:00
|
|
|
/**
|
|
|
|
* adjust_fs_space - Adjusts the free space available due to gfs2_grow
|
|
|
|
* @inode: the rindex inode
|
|
|
|
*/
|
|
|
|
static void adjust_fs_space(struct inode *inode)
|
|
|
|
{
|
|
|
|
struct gfs2_sbd *sdp = inode->i_sb->s_fs_info;
|
2009-06-25 14:09:51 -06:00
|
|
|
struct gfs2_inode *m_ip = GFS2_I(sdp->sd_statfs_inode);
|
|
|
|
struct gfs2_inode *l_ip = GFS2_I(sdp->sd_sc_inode);
|
[GFS2] kernel changes to support new gfs2_grow command
This is another revision of my gfs2 kernel patch that allows
gfs2_grow to function properly.
Steve Whitehouse expressed some concerns about the previous
patch and I restructured it based on his comments.
The previous patch was doing the statfs_change at file close time,
under its own transaction. The current patch does the statfs_change
inside the gfs2_commit_write function, which keeps it under the
umbrella of the inode transaction.
I can't call ri_update to re-read the rindex file during the
transaction because the transaction may have outstanding unwritten
buffers attached to the rgrps that would be otherwise blown away.
So instead, I created a new function, gfs2_ri_total, that will
re-read the rindex file just to total the file system space
for the sake of the statfs_change. The ri_update will happen
later, when gfs2 realizes the version number has changed, as it
happened before my patch.
Since the statfs_change is happening at write_commit time and there
may be multiple writes to the rindex file for one grow operation.
So one consequence of this restructuring is that instead of getting
one kernel message to indicate the change, you may see several.
For example, before when you did a gfs2_grow, you'd get a single
message like:
GFS2: File system extended by 247876 blocks (968MB)
Now you get something like:
GFS2: File system extended by 207896 blocks (812MB)
GFS2: File system extended by 39980 blocks (156MB)
This version has also been successfully run against the hours-long
"gfs2_fsck_hellfire" test that does several gfs2_grow and gfs2_fsck
while interjecting file system damage. It does this repeatedly
under a variety Resource Group conditions.
Signed-off-By: Bob Peterson <rpeterso@redhat.com>
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
2007-05-09 08:37:57 -06:00
|
|
|
struct gfs2_statfs_change_host *m_sc = &sdp->sd_statfs_master;
|
|
|
|
struct gfs2_statfs_change_host *l_sc = &sdp->sd_statfs_local;
|
2009-06-25 14:09:51 -06:00
|
|
|
struct buffer_head *m_bh, *l_bh;
|
[GFS2] kernel changes to support new gfs2_grow command
This is another revision of my gfs2 kernel patch that allows
gfs2_grow to function properly.
Steve Whitehouse expressed some concerns about the previous
patch and I restructured it based on his comments.
The previous patch was doing the statfs_change at file close time,
under its own transaction. The current patch does the statfs_change
inside the gfs2_commit_write function, which keeps it under the
umbrella of the inode transaction.
I can't call ri_update to re-read the rindex file during the
transaction because the transaction may have outstanding unwritten
buffers attached to the rgrps that would be otherwise blown away.
So instead, I created a new function, gfs2_ri_total, that will
re-read the rindex file just to total the file system space
for the sake of the statfs_change. The ri_update will happen
later, when gfs2 realizes the version number has changed, as it
happened before my patch.
Since the statfs_change is happening at write_commit time and there
may be multiple writes to the rindex file for one grow operation.
So one consequence of this restructuring is that instead of getting
one kernel message to indicate the change, you may see several.
For example, before when you did a gfs2_grow, you'd get a single
message like:
GFS2: File system extended by 247876 blocks (968MB)
Now you get something like:
GFS2: File system extended by 207896 blocks (812MB)
GFS2: File system extended by 39980 blocks (156MB)
This version has also been successfully run against the hours-long
"gfs2_fsck_hellfire" test that does several gfs2_grow and gfs2_fsck
while interjecting file system damage. It does this repeatedly
under a variety Resource Group conditions.
Signed-off-By: Bob Peterson <rpeterso@redhat.com>
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
2007-05-09 08:37:57 -06:00
|
|
|
u64 fs_total, new_free;
|
|
|
|
|
|
|
|
/* Total up the file system space, according to the latest rindex. */
|
|
|
|
fs_total = gfs2_ri_total(sdp);
|
2009-06-25 14:09:51 -06:00
|
|
|
if (gfs2_meta_inode_buffer(m_ip, &m_bh) != 0)
|
|
|
|
return;
|
[GFS2] kernel changes to support new gfs2_grow command
This is another revision of my gfs2 kernel patch that allows
gfs2_grow to function properly.
Steve Whitehouse expressed some concerns about the previous
patch and I restructured it based on his comments.
The previous patch was doing the statfs_change at file close time,
under its own transaction. The current patch does the statfs_change
inside the gfs2_commit_write function, which keeps it under the
umbrella of the inode transaction.
I can't call ri_update to re-read the rindex file during the
transaction because the transaction may have outstanding unwritten
buffers attached to the rgrps that would be otherwise blown away.
So instead, I created a new function, gfs2_ri_total, that will
re-read the rindex file just to total the file system space
for the sake of the statfs_change. The ri_update will happen
later, when gfs2 realizes the version number has changed, as it
happened before my patch.
Since the statfs_change is happening at write_commit time and there
may be multiple writes to the rindex file for one grow operation.
So one consequence of this restructuring is that instead of getting
one kernel message to indicate the change, you may see several.
For example, before when you did a gfs2_grow, you'd get a single
message like:
GFS2: File system extended by 247876 blocks (968MB)
Now you get something like:
GFS2: File system extended by 207896 blocks (812MB)
GFS2: File system extended by 39980 blocks (156MB)
This version has also been successfully run against the hours-long
"gfs2_fsck_hellfire" test that does several gfs2_grow and gfs2_fsck
while interjecting file system damage. It does this repeatedly
under a variety Resource Group conditions.
Signed-off-By: Bob Peterson <rpeterso@redhat.com>
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
2007-05-09 08:37:57 -06:00
|
|
|
|
|
|
|
spin_lock(&sdp->sd_statfs_spin);
|
2009-06-25 14:09:51 -06:00
|
|
|
gfs2_statfs_change_in(m_sc, m_bh->b_data +
|
|
|
|
sizeof(struct gfs2_dinode));
|
[GFS2] kernel changes to support new gfs2_grow command
This is another revision of my gfs2 kernel patch that allows
gfs2_grow to function properly.
Steve Whitehouse expressed some concerns about the previous
patch and I restructured it based on his comments.
The previous patch was doing the statfs_change at file close time,
under its own transaction. The current patch does the statfs_change
inside the gfs2_commit_write function, which keeps it under the
umbrella of the inode transaction.
I can't call ri_update to re-read the rindex file during the
transaction because the transaction may have outstanding unwritten
buffers attached to the rgrps that would be otherwise blown away.
So instead, I created a new function, gfs2_ri_total, that will
re-read the rindex file just to total the file system space
for the sake of the statfs_change. The ri_update will happen
later, when gfs2 realizes the version number has changed, as it
happened before my patch.
Since the statfs_change is happening at write_commit time and there
may be multiple writes to the rindex file for one grow operation.
So one consequence of this restructuring is that instead of getting
one kernel message to indicate the change, you may see several.
For example, before when you did a gfs2_grow, you'd get a single
message like:
GFS2: File system extended by 247876 blocks (968MB)
Now you get something like:
GFS2: File system extended by 207896 blocks (812MB)
GFS2: File system extended by 39980 blocks (156MB)
This version has also been successfully run against the hours-long
"gfs2_fsck_hellfire" test that does several gfs2_grow and gfs2_fsck
while interjecting file system damage. It does this repeatedly
under a variety Resource Group conditions.
Signed-off-By: Bob Peterson <rpeterso@redhat.com>
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
2007-05-09 08:37:57 -06:00
|
|
|
if (fs_total > (m_sc->sc_total + l_sc->sc_total))
|
|
|
|
new_free = fs_total - (m_sc->sc_total + l_sc->sc_total);
|
|
|
|
else
|
|
|
|
new_free = 0;
|
|
|
|
spin_unlock(&sdp->sd_statfs_spin);
|
2007-05-10 15:54:38 -06:00
|
|
|
fs_warn(sdp, "File system extended by %llu blocks.\n",
|
|
|
|
(unsigned long long)new_free);
|
[GFS2] kernel changes to support new gfs2_grow command
This is another revision of my gfs2 kernel patch that allows
gfs2_grow to function properly.
Steve Whitehouse expressed some concerns about the previous
patch and I restructured it based on his comments.
The previous patch was doing the statfs_change at file close time,
under its own transaction. The current patch does the statfs_change
inside the gfs2_commit_write function, which keeps it under the
umbrella of the inode transaction.
I can't call ri_update to re-read the rindex file during the
transaction because the transaction may have outstanding unwritten
buffers attached to the rgrps that would be otherwise blown away.
So instead, I created a new function, gfs2_ri_total, that will
re-read the rindex file just to total the file system space
for the sake of the statfs_change. The ri_update will happen
later, when gfs2 realizes the version number has changed, as it
happened before my patch.
Since the statfs_change is happening at write_commit time and there
may be multiple writes to the rindex file for one grow operation.
So one consequence of this restructuring is that instead of getting
one kernel message to indicate the change, you may see several.
For example, before when you did a gfs2_grow, you'd get a single
message like:
GFS2: File system extended by 247876 blocks (968MB)
Now you get something like:
GFS2: File system extended by 207896 blocks (812MB)
GFS2: File system extended by 39980 blocks (156MB)
This version has also been successfully run against the hours-long
"gfs2_fsck_hellfire" test that does several gfs2_grow and gfs2_fsck
while interjecting file system damage. It does this repeatedly
under a variety Resource Group conditions.
Signed-off-By: Bob Peterson <rpeterso@redhat.com>
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
2007-05-09 08:37:57 -06:00
|
|
|
gfs2_statfs_change(sdp, new_free, new_free, 0);
|
2009-06-25 14:09:51 -06:00
|
|
|
|
|
|
|
if (gfs2_meta_inode_buffer(l_ip, &l_bh) != 0)
|
|
|
|
goto out;
|
|
|
|
update_statfs(sdp, m_bh, l_bh);
|
|
|
|
brelse(l_bh);
|
|
|
|
out:
|
|
|
|
brelse(m_bh);
|
[GFS2] kernel changes to support new gfs2_grow command
This is another revision of my gfs2 kernel patch that allows
gfs2_grow to function properly.
Steve Whitehouse expressed some concerns about the previous
patch and I restructured it based on his comments.
The previous patch was doing the statfs_change at file close time,
under its own transaction. The current patch does the statfs_change
inside the gfs2_commit_write function, which keeps it under the
umbrella of the inode transaction.
I can't call ri_update to re-read the rindex file during the
transaction because the transaction may have outstanding unwritten
buffers attached to the rgrps that would be otherwise blown away.
So instead, I created a new function, gfs2_ri_total, that will
re-read the rindex file just to total the file system space
for the sake of the statfs_change. The ri_update will happen
later, when gfs2 realizes the version number has changed, as it
happened before my patch.
Since the statfs_change is happening at write_commit time and there
may be multiple writes to the rindex file for one grow operation.
So one consequence of this restructuring is that instead of getting
one kernel message to indicate the change, you may see several.
For example, before when you did a gfs2_grow, you'd get a single
message like:
GFS2: File system extended by 247876 blocks (968MB)
Now you get something like:
GFS2: File system extended by 207896 blocks (812MB)
GFS2: File system extended by 39980 blocks (156MB)
This version has also been successfully run against the hours-long
"gfs2_fsck_hellfire" test that does several gfs2_grow and gfs2_fsck
while interjecting file system damage. It does this repeatedly
under a variety Resource Group conditions.
Signed-off-By: Bob Peterson <rpeterso@redhat.com>
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
2007-05-09 08:37:57 -06:00
|
|
|
}
|
|
|
|
|
2006-01-16 09:50:04 -07:00
|
|
|
/**
|
2007-10-16 02:25:07 -06:00
|
|
|
* gfs2_stuffed_write_end - Write end for stuffed files
|
|
|
|
* @inode: The inode
|
|
|
|
* @dibh: The buffer_head containing the on-disk inode
|
|
|
|
* @pos: The file position
|
|
|
|
* @len: The length of the write
|
|
|
|
* @copied: How much was actually copied by the VFS
|
|
|
|
* @page: The page
|
|
|
|
*
|
|
|
|
* This copies the data from the page into the inode block after
|
|
|
|
* the inode data structure itself.
|
|
|
|
*
|
|
|
|
* Returns: errno
|
|
|
|
*/
|
|
|
|
static int gfs2_stuffed_write_end(struct inode *inode, struct buffer_head *dibh,
|
|
|
|
loff_t pos, unsigned len, unsigned copied,
|
|
|
|
struct page *page)
|
|
|
|
{
|
|
|
|
struct gfs2_inode *ip = GFS2_I(inode);
|
|
|
|
struct gfs2_sbd *sdp = GFS2_SB(inode);
|
2009-06-25 14:09:51 -06:00
|
|
|
struct gfs2_inode *m_ip = GFS2_I(sdp->sd_statfs_inode);
|
2007-10-16 02:25:07 -06:00
|
|
|
u64 to = pos + copied;
|
|
|
|
void *kaddr;
|
|
|
|
unsigned char *buf = dibh->b_data + sizeof(struct gfs2_dinode);
|
|
|
|
|
|
|
|
BUG_ON((pos + len) > (dibh->b_size - sizeof(struct gfs2_dinode)));
|
2011-11-25 08:14:30 -07:00
|
|
|
kaddr = kmap_atomic(page);
|
2007-10-16 02:25:07 -06:00
|
|
|
memcpy(buf + pos, kaddr + pos, copied);
|
|
|
|
memset(kaddr + pos + copied, 0, len - copied);
|
|
|
|
flush_dcache_page(page);
|
2011-11-25 08:14:30 -07:00
|
|
|
kunmap_atomic(kaddr);
|
2007-10-16 02:25:07 -06:00
|
|
|
|
|
|
|
if (!PageUptodate(page))
|
|
|
|
SetPageUptodate(page);
|
|
|
|
unlock_page(page);
|
|
|
|
page_cache_release(page);
|
|
|
|
|
2009-05-12 10:16:20 -06:00
|
|
|
if (copied) {
|
2010-08-11 02:53:11 -06:00
|
|
|
if (inode->i_size < to)
|
2009-05-12 10:16:20 -06:00
|
|
|
i_size_write(inode, to);
|
2007-10-16 02:25:07 -06:00
|
|
|
mark_inode_dirty(inode);
|
|
|
|
}
|
|
|
|
|
2009-11-10 11:54:56 -07:00
|
|
|
if (inode == sdp->sd_rindex) {
|
2007-10-16 02:25:07 -06:00
|
|
|
adjust_fs_space(inode);
|
2012-04-10 06:56:04 -06:00
|
|
|
sdp->sd_rindex_uptodate = 0;
|
2009-11-10 11:54:56 -07:00
|
|
|
}
|
2007-10-16 02:25:07 -06:00
|
|
|
|
|
|
|
brelse(dibh);
|
|
|
|
gfs2_trans_end(sdp);
|
2009-06-25 14:09:51 -06:00
|
|
|
if (inode == sdp->sd_rindex) {
|
|
|
|
gfs2_glock_dq(&m_ip->i_gh);
|
|
|
|
gfs2_holder_uninit(&m_ip->i_gh);
|
|
|
|
}
|
2007-10-16 02:25:07 -06:00
|
|
|
gfs2_glock_dq(&ip->i_gh);
|
|
|
|
gfs2_holder_uninit(&ip->i_gh);
|
|
|
|
return copied;
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* gfs2_write_end
|
2006-01-16 09:50:04 -07:00
|
|
|
* @file: The file to write to
|
2007-10-16 02:25:07 -06:00
|
|
|
* @mapping: The address space to write to
|
|
|
|
* @pos: The file position
|
|
|
|
* @len: The length of the data
|
|
|
|
* @copied:
|
|
|
|
* @page: The page that has been written
|
|
|
|
* @fsdata: The fsdata (unused in GFS2)
|
|
|
|
*
|
|
|
|
* The main write_end function for GFS2. We have a separate one for
|
|
|
|
* stuffed files as they are slightly different, otherwise we just
|
|
|
|
* put our locking around the VFS provided functions.
|
2006-01-16 09:50:04 -07:00
|
|
|
*
|
|
|
|
* Returns: errno
|
|
|
|
*/
|
|
|
|
|
2007-10-16 02:25:07 -06:00
|
|
|
static int gfs2_write_end(struct file *file, struct address_space *mapping,
|
|
|
|
loff_t pos, unsigned len, unsigned copied,
|
|
|
|
struct page *page, void *fsdata)
|
2006-01-16 09:50:04 -07:00
|
|
|
{
|
|
|
|
struct inode *inode = page->mapping->host;
|
2006-06-14 13:32:57 -06:00
|
|
|
struct gfs2_inode *ip = GFS2_I(inode);
|
|
|
|
struct gfs2_sbd *sdp = GFS2_SB(inode);
|
2009-06-25 14:09:51 -06:00
|
|
|
struct gfs2_inode *m_ip = GFS2_I(sdp->sd_statfs_inode);
|
2006-02-08 04:50:51 -07:00
|
|
|
struct buffer_head *dibh;
|
2007-10-16 02:25:07 -06:00
|
|
|
unsigned int from = pos & (PAGE_CACHE_SIZE - 1);
|
|
|
|
unsigned int to = from + len;
|
|
|
|
int ret;
|
2013-09-03 15:59:42 -06:00
|
|
|
struct gfs2_trans *tr = current->journal_info;
|
|
|
|
BUG_ON(!tr);
|
2006-01-16 09:50:04 -07:00
|
|
|
|
2008-02-22 09:07:18 -07:00
|
|
|
BUG_ON(gfs2_glock_is_locked_by_me(ip->i_gl) == NULL);
|
2006-02-08 04:50:51 -07:00
|
|
|
|
2007-10-16 02:25:07 -06:00
|
|
|
ret = gfs2_meta_inode_buffer(ip, &dibh);
|
|
|
|
if (unlikely(ret)) {
|
|
|
|
unlock_page(page);
|
|
|
|
page_cache_release(page);
|
|
|
|
goto failed;
|
|
|
|
}
|
2006-02-08 04:50:51 -07:00
|
|
|
|
2007-10-16 02:25:07 -06:00
|
|
|
if (gfs2_is_stuffed(ip))
|
|
|
|
return gfs2_stuffed_write_end(inode, dibh, pos, len, copied, page);
|
2006-01-16 09:50:04 -07:00
|
|
|
|
2007-10-17 01:35:19 -06:00
|
|
|
if (!gfs2_is_writeback(ip))
|
2007-10-16 02:25:07 -06:00
|
|
|
gfs2_page_add_databufs(ip, page, from, to);
|
2006-01-16 09:50:04 -07:00
|
|
|
|
2007-10-16 02:25:07 -06:00
|
|
|
ret = generic_write_end(file, mapping, pos, len, copied, page, fsdata);
|
2013-09-03 15:59:42 -06:00
|
|
|
if (tr->tr_num_buf_new)
|
|
|
|
__mark_inode_dirty(inode, I_DIRTY_DATASYNC);
|
|
|
|
else
|
|
|
|
gfs2_trans_add_meta(ip->i_gl, dibh);
|
|
|
|
|
2006-10-02 10:39:19 -06:00
|
|
|
|
2009-11-10 11:54:56 -07:00
|
|
|
if (inode == sdp->sd_rindex) {
|
[GFS2] kernel changes to support new gfs2_grow command
This is another revision of my gfs2 kernel patch that allows
gfs2_grow to function properly.
Steve Whitehouse expressed some concerns about the previous
patch and I restructured it based on his comments.
The previous patch was doing the statfs_change at file close time,
under its own transaction. The current patch does the statfs_change
inside the gfs2_commit_write function, which keeps it under the
umbrella of the inode transaction.
I can't call ri_update to re-read the rindex file during the
transaction because the transaction may have outstanding unwritten
buffers attached to the rgrps that would be otherwise blown away.
So instead, I created a new function, gfs2_ri_total, that will
re-read the rindex file just to total the file system space
for the sake of the statfs_change. The ri_update will happen
later, when gfs2 realizes the version number has changed, as it
happened before my patch.
Since the statfs_change is happening at write_commit time and there
may be multiple writes to the rindex file for one grow operation.
So one consequence of this restructuring is that instead of getting
one kernel message to indicate the change, you may see several.
For example, before when you did a gfs2_grow, you'd get a single
message like:
GFS2: File system extended by 247876 blocks (968MB)
Now you get something like:
GFS2: File system extended by 207896 blocks (812MB)
GFS2: File system extended by 39980 blocks (156MB)
This version has also been successfully run against the hours-long
"gfs2_fsck_hellfire" test that does several gfs2_grow and gfs2_fsck
while interjecting file system damage. It does this repeatedly
under a variety Resource Group conditions.
Signed-off-By: Bob Peterson <rpeterso@redhat.com>
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
2007-05-09 08:37:57 -06:00
|
|
|
adjust_fs_space(inode);
|
2012-04-10 06:56:04 -06:00
|
|
|
sdp->sd_rindex_uptodate = 0;
|
2009-11-10 11:54:56 -07:00
|
|
|
}
|
[GFS2] kernel changes to support new gfs2_grow command
This is another revision of my gfs2 kernel patch that allows
gfs2_grow to function properly.
Steve Whitehouse expressed some concerns about the previous
patch and I restructured it based on his comments.
The previous patch was doing the statfs_change at file close time,
under its own transaction. The current patch does the statfs_change
inside the gfs2_commit_write function, which keeps it under the
umbrella of the inode transaction.
I can't call ri_update to re-read the rindex file during the
transaction because the transaction may have outstanding unwritten
buffers attached to the rgrps that would be otherwise blown away.
So instead, I created a new function, gfs2_ri_total, that will
re-read the rindex file just to total the file system space
for the sake of the statfs_change. The ri_update will happen
later, when gfs2 realizes the version number has changed, as it
happened before my patch.
Since the statfs_change is happening at write_commit time and there
may be multiple writes to the rindex file for one grow operation.
So one consequence of this restructuring is that instead of getting
one kernel message to indicate the change, you may see several.
For example, before when you did a gfs2_grow, you'd get a single
message like:
GFS2: File system extended by 247876 blocks (968MB)
Now you get something like:
GFS2: File system extended by 207896 blocks (812MB)
GFS2: File system extended by 39980 blocks (156MB)
This version has also been successfully run against the hours-long
"gfs2_fsck_hellfire" test that does several gfs2_grow and gfs2_fsck
while interjecting file system damage. It does this repeatedly
under a variety Resource Group conditions.
Signed-off-By: Bob Peterson <rpeterso@redhat.com>
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
2007-05-09 08:37:57 -06:00
|
|
|
|
2006-02-08 04:50:51 -07:00
|
|
|
brelse(dibh);
|
2007-10-16 02:25:07 -06:00
|
|
|
failed:
|
2011-03-16 14:32:39 -06:00
|
|
|
gfs2_trans_end(sdp);
|
2012-07-30 07:53:19 -06:00
|
|
|
gfs2_inplace_release(ip);
|
2012-05-18 07:28:23 -06:00
|
|
|
if (ip->i_res->rs_qa_qd_num)
|
2006-02-08 04:50:51 -07:00
|
|
|
gfs2_quota_unlock(ip);
|
2009-06-25 14:09:51 -06:00
|
|
|
if (inode == sdp->sd_rindex) {
|
|
|
|
gfs2_glock_dq(&m_ip->i_gh);
|
|
|
|
gfs2_holder_uninit(&m_ip->i_gh);
|
|
|
|
}
|
2007-10-16 02:25:07 -06:00
|
|
|
gfs2_glock_dq(&ip->i_gh);
|
2006-02-08 04:50:51 -07:00
|
|
|
gfs2_holder_uninit(&ip->i_gh);
|
2007-10-16 02:25:07 -06:00
|
|
|
return ret;
|
2006-01-16 09:50:04 -07:00
|
|
|
}
|
|
|
|
|
2007-06-12 10:24:36 -06:00
|
|
|
/**
|
|
|
|
* gfs2_set_page_dirty - Page dirtying function
|
|
|
|
* @page: The page to dirty
|
|
|
|
*
|
|
|
|
* Returns: 1 if it dirtyed the page, or 0 otherwise
|
|
|
|
*/
|
|
|
|
|
|
|
|
static int gfs2_set_page_dirty(struct page *page)
|
|
|
|
{
|
2007-10-17 01:47:38 -06:00
|
|
|
SetPageChecked(page);
|
2007-06-12 10:24:36 -06:00
|
|
|
return __set_page_dirty_buffers(page);
|
|
|
|
}
|
|
|
|
|
2006-01-16 09:50:04 -07:00
|
|
|
/**
|
|
|
|
* gfs2_bmap - Block map function
|
|
|
|
* @mapping: Address space info
|
|
|
|
* @lblock: The block to map
|
|
|
|
*
|
|
|
|
* Returns: The disk address for the block or 0 on hole or error
|
|
|
|
*/
|
|
|
|
|
|
|
|
static sector_t gfs2_bmap(struct address_space *mapping, sector_t lblock)
|
|
|
|
{
|
2006-06-14 13:32:57 -06:00
|
|
|
struct gfs2_inode *ip = GFS2_I(mapping->host);
|
2006-01-16 09:50:04 -07:00
|
|
|
struct gfs2_holder i_gh;
|
|
|
|
sector_t dblock = 0;
|
|
|
|
int error;
|
|
|
|
|
|
|
|
error = gfs2_glock_nq_init(ip->i_gl, LM_ST_SHARED, LM_FLAG_ANY, &i_gh);
|
|
|
|
if (error)
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
if (!gfs2_is_stuffed(ip))
|
2007-12-10 13:13:27 -07:00
|
|
|
dblock = generic_block_bmap(mapping, lblock, gfs2_block_map);
|
2006-01-16 09:50:04 -07:00
|
|
|
|
|
|
|
gfs2_glock_dq_uninit(&i_gh);
|
|
|
|
|
|
|
|
return dblock;
|
|
|
|
}
|
|
|
|
|
2007-09-02 03:48:13 -06:00
|
|
|
static void gfs2_discard(struct gfs2_sbd *sdp, struct buffer_head *bh)
|
|
|
|
{
|
|
|
|
struct gfs2_bufdata *bd;
|
|
|
|
|
|
|
|
lock_buffer(bh);
|
|
|
|
gfs2_log_lock(sdp);
|
|
|
|
clear_buffer_dirty(bh);
|
|
|
|
bd = bh->b_private;
|
|
|
|
if (bd) {
|
2012-05-01 10:00:34 -06:00
|
|
|
if (!list_empty(&bd->bd_list) && !buffer_pinned(bh))
|
|
|
|
list_del_init(&bd->bd_list);
|
2007-09-17 03:59:52 -06:00
|
|
|
else
|
|
|
|
gfs2_remove_from_journal(bh, current->journal_info, 0);
|
2007-09-02 03:48:13 -06:00
|
|
|
}
|
|
|
|
bh->b_bdev = NULL;
|
|
|
|
clear_buffer_mapped(bh);
|
|
|
|
clear_buffer_req(bh);
|
|
|
|
clear_buffer_new(bh);
|
|
|
|
gfs2_log_unlock(sdp);
|
|
|
|
unlock_buffer(bh);
|
|
|
|
}
|
|
|
|
|
2013-05-21 21:17:23 -06:00
|
|
|
static void gfs2_invalidatepage(struct page *page, unsigned int offset,
|
|
|
|
unsigned int length)
|
2006-01-16 09:50:04 -07:00
|
|
|
{
|
2007-09-02 03:48:13 -06:00
|
|
|
struct gfs2_sbd *sdp = GFS2_SB(page->mapping->host);
|
2013-05-21 21:58:49 -06:00
|
|
|
unsigned int stop = offset + length;
|
|
|
|
int partial_page = (offset || length < PAGE_CACHE_SIZE);
|
2007-09-02 03:48:13 -06:00
|
|
|
struct buffer_head *bh, *head;
|
|
|
|
unsigned long pos = 0;
|
|
|
|
|
2006-01-16 09:50:04 -07:00
|
|
|
BUG_ON(!PageLocked(page));
|
2013-05-21 21:58:49 -06:00
|
|
|
if (!partial_page)
|
2007-06-12 10:24:36 -06:00
|
|
|
ClearPageChecked(page);
|
2007-09-02 03:48:13 -06:00
|
|
|
if (!page_has_buffers(page))
|
|
|
|
goto out;
|
2006-01-16 09:50:04 -07:00
|
|
|
|
2007-09-02 03:48:13 -06:00
|
|
|
bh = head = page_buffers(page);
|
|
|
|
do {
|
2013-05-21 21:58:49 -06:00
|
|
|
if (pos + bh->b_size > stop)
|
|
|
|
return;
|
|
|
|
|
2007-09-02 03:48:13 -06:00
|
|
|
if (offset <= pos)
|
|
|
|
gfs2_discard(sdp, bh);
|
|
|
|
pos += bh->b_size;
|
|
|
|
bh = bh->b_this_page;
|
|
|
|
} while (bh != head);
|
|
|
|
out:
|
2013-05-21 21:58:49 -06:00
|
|
|
if (!partial_page)
|
2007-09-02 03:48:13 -06:00
|
|
|
try_to_release_page(page, 0);
|
2006-01-16 09:50:04 -07:00
|
|
|
}
|
|
|
|
|
2006-12-14 11:24:26 -07:00
|
|
|
/**
|
|
|
|
* gfs2_ok_for_dio - check that dio is valid on this file
|
|
|
|
* @ip: The inode
|
|
|
|
* @rw: READ or WRITE
|
|
|
|
* @offset: The offset at which we are reading or writing
|
|
|
|
*
|
|
|
|
* Returns: 0 (to ignore the i/o request and thus fall back to buffered i/o)
|
|
|
|
* 1 (to accept the i/o request)
|
|
|
|
*/
|
|
|
|
static int gfs2_ok_for_dio(struct gfs2_inode *ip, int rw, loff_t offset)
|
|
|
|
{
|
|
|
|
/*
|
|
|
|
* Should we return an error here? I can't see that O_DIRECT for
|
2007-10-17 01:47:38 -06:00
|
|
|
* a stuffed file makes any sense. For now we'll silently fall
|
|
|
|
* back to buffered I/O
|
2006-12-14 11:24:26 -07:00
|
|
|
*/
|
|
|
|
if (gfs2_is_stuffed(ip))
|
|
|
|
return 0;
|
|
|
|
|
2008-09-11 13:35:37 -06:00
|
|
|
if (offset >= i_size_read(&ip->i_inode))
|
2006-12-14 11:24:26 -07:00
|
|
|
return 0;
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
|
2006-07-25 15:24:12 -06:00
|
|
|
static ssize_t gfs2_direct_IO(int rw, struct kiocb *iocb,
|
|
|
|
const struct iovec *iov, loff_t offset,
|
|
|
|
unsigned long nr_segs)
|
2006-02-14 04:54:42 -07:00
|
|
|
{
|
|
|
|
struct file *file = iocb->ki_filp;
|
|
|
|
struct inode *inode = file->f_mapping->host;
|
2013-12-18 07:14:52 -07:00
|
|
|
struct address_space *mapping = inode->i_mapping;
|
2006-06-14 13:32:57 -06:00
|
|
|
struct gfs2_inode *ip = GFS2_I(inode);
|
2006-02-14 04:54:42 -07:00
|
|
|
struct gfs2_holder gh;
|
|
|
|
int rv;
|
|
|
|
|
|
|
|
/*
|
2006-12-14 11:24:26 -07:00
|
|
|
* Deferred lock, even if its a write, since we do no allocation
|
|
|
|
* on this path. All we need change is atime, and this lock mode
|
|
|
|
* ensures that other nodes have flushed their buffered read caches
|
|
|
|
* (i.e. their page cache entries for this inode). We do not,
|
|
|
|
* unfortunately have the option of only flushing a range like
|
|
|
|
* the VFS does.
|
2006-02-14 04:54:42 -07:00
|
|
|
*/
|
2008-09-18 06:53:59 -06:00
|
|
|
gfs2_holder_init(ip->i_gl, LM_ST_DEFERRED, 0, &gh);
|
|
|
|
rv = gfs2_glock_nq(&gh);
|
2006-02-14 04:54:42 -07:00
|
|
|
if (rv)
|
2006-12-14 11:24:26 -07:00
|
|
|
return rv;
|
|
|
|
rv = gfs2_ok_for_dio(ip, rw, offset);
|
|
|
|
if (rv != 1)
|
|
|
|
goto out; /* dio not valid, fall back to buffered i/o */
|
|
|
|
|
2013-12-18 07:14:52 -07:00
|
|
|
/*
|
|
|
|
* Now since we are holding a deferred (CW) lock at this point, you
|
|
|
|
* might be wondering why this is ever needed. There is a case however
|
|
|
|
* where we've granted a deferred local lock against a cached exclusive
|
|
|
|
* glock. That is ok provided all granted local locks are deferred, but
|
|
|
|
* it also means that it is possible to encounter pages which are
|
|
|
|
* cached and possibly also mapped. So here we check for that and sort
|
|
|
|
* them out ahead of the dio. The glock state machine will take care of
|
|
|
|
* everything else.
|
|
|
|
*
|
|
|
|
* If in fact the cached glock state (gl->gl_state) is deferred (CW) in
|
|
|
|
* the first place, mapping->nr_pages will always be zero.
|
|
|
|
*/
|
|
|
|
if (mapping->nrpages) {
|
|
|
|
loff_t lstart = offset & (PAGE_CACHE_SIZE - 1);
|
|
|
|
loff_t len = iov_length(iov, nr_segs);
|
|
|
|
loff_t end = PAGE_ALIGN(offset + len) - 1;
|
|
|
|
|
|
|
|
rv = 0;
|
|
|
|
if (len == 0)
|
|
|
|
goto out;
|
|
|
|
if (test_and_clear_bit(GIF_SW_PAGED, &ip->i_flags))
|
|
|
|
unmap_shared_mapping_range(ip->i_inode.i_mapping, offset, len);
|
|
|
|
rv = filemap_write_and_wait_range(mapping, lstart, end);
|
|
|
|
if (rv)
|
2014-01-14 06:46:51 -07:00
|
|
|
goto out;
|
|
|
|
if (rw == WRITE)
|
|
|
|
truncate_inode_pages_range(mapping, lstart, end);
|
2013-12-18 07:14:52 -07:00
|
|
|
}
|
|
|
|
|
2010-06-04 03:29:53 -06:00
|
|
|
rv = __blockdev_direct_IO(rw, iocb, inode, inode->i_sb->s_bdev, iov,
|
|
|
|
offset, nr_segs, gfs2_get_block_direct,
|
|
|
|
NULL, NULL, 0);
|
2006-02-14 04:54:42 -07:00
|
|
|
out:
|
2012-08-09 11:48:42 -06:00
|
|
|
gfs2_glock_dq(&gh);
|
2006-02-14 04:54:42 -07:00
|
|
|
gfs2_holder_uninit(&gh);
|
|
|
|
return rv;
|
|
|
|
}
|
|
|
|
|
2006-07-11 07:46:33 -06:00
|
|
|
/**
|
2006-08-31 10:14:44 -06:00
|
|
|
* gfs2_releasepage - free the metadata associated with a page
|
2006-07-11 07:46:33 -06:00
|
|
|
* @page: the page that's being released
|
|
|
|
* @gfp_mask: passed from Linux VFS, ignored by us
|
|
|
|
*
|
|
|
|
* Call try_to_free_buffers() if the buffers in this page can be
|
|
|
|
* released.
|
|
|
|
*
|
|
|
|
* Returns: 0
|
|
|
|
*/
|
|
|
|
|
|
|
|
int gfs2_releasepage(struct page *page, gfp_t gfp_mask)
|
|
|
|
{
|
2009-12-08 05:12:13 -07:00
|
|
|
struct address_space *mapping = page->mapping;
|
|
|
|
struct gfs2_sbd *sdp = gfs2_mapping2sbd(mapping);
|
2006-07-11 07:46:33 -06:00
|
|
|
struct buffer_head *bh, *head;
|
|
|
|
struct gfs2_bufdata *bd;
|
|
|
|
|
|
|
|
if (!page_has_buffers(page))
|
2007-09-20 08:26:33 -06:00
|
|
|
return 0;
|
2006-07-11 07:46:33 -06:00
|
|
|
|
2007-08-16 09:03:57 -06:00
|
|
|
gfs2_log_lock(sdp);
|
2011-07-14 01:59:44 -06:00
|
|
|
spin_lock(&sdp->sd_ail_lock);
|
2006-07-11 07:46:33 -06:00
|
|
|
head = bh = page_buffers(page);
|
|
|
|
do {
|
2007-08-16 09:03:57 -06:00
|
|
|
if (atomic_read(&bh->b_count))
|
|
|
|
goto cannot_release;
|
|
|
|
bd = bh->b_private;
|
GFS2: replace gfs2_ail structure with gfs2_trans
In order to allow transactions and log flushes to happen at the same
time, gfs2 needs to move the transaction accounting and active items
list code into the gfs2_trans structure. As a first step toward this,
this patch removes the gfs2_ail structure, and handles the active items
list in the gfs_trans structure. This keeps gfs2 from allocating an ail
structure on log flushes, and gives us a struture that can later be used
to store the transaction accounting outside of the gfs2 superblock
structure.
With this patch, at the end of a transaction, gfs2 will add the
gfs2_trans structure to the superblock if there is not one already.
This structure now has the active items fields that were previously in
gfs2_ail. This is not necessary in the case where the transaction was
simply used to add revokes, since these are never written outside of the
journal, and thus, don't need an active items list.
Also, in order to make sure that the transaction structure is not
removed while it's still in use by gfs2_trans_end, unlocking the
sd_log_flush_lock has to happen slightly later in ending the
transaction.
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
2013-04-05 19:31:46 -06:00
|
|
|
if (bd && bd->bd_tr)
|
2007-08-16 09:03:57 -06:00
|
|
|
goto cannot_release;
|
2011-05-03 04:49:19 -06:00
|
|
|
if (buffer_pinned(bh) || buffer_dirty(bh))
|
|
|
|
goto not_possible;
|
2007-08-16 09:03:57 -06:00
|
|
|
bh = bh->b_this_page;
|
|
|
|
} while(bh != head);
|
2011-07-14 01:59:44 -06:00
|
|
|
spin_unlock(&sdp->sd_ail_lock);
|
2006-07-11 07:46:33 -06:00
|
|
|
|
2007-08-16 09:03:57 -06:00
|
|
|
head = bh = page_buffers(page);
|
|
|
|
do {
|
2006-07-11 07:46:33 -06:00
|
|
|
bd = bh->b_private;
|
|
|
|
if (bd) {
|
|
|
|
gfs2_assert_warn(sdp, bd->bd_bh == bh);
|
2013-11-26 06:21:08 -07:00
|
|
|
if (!list_empty(&bd->bd_list))
|
|
|
|
list_del_init(&bd->bd_list);
|
|
|
|
bd->bd_bh = NULL;
|
2006-07-11 07:46:33 -06:00
|
|
|
bh->b_private = NULL;
|
2006-08-31 10:14:44 -06:00
|
|
|
kmem_cache_free(gfs2_bufdata_cachep, bd);
|
2013-11-26 06:21:08 -07:00
|
|
|
}
|
2006-07-11 07:46:33 -06:00
|
|
|
|
|
|
|
bh = bh->b_this_page;
|
2006-08-24 13:59:40 -06:00
|
|
|
} while (bh != head);
|
2013-11-26 06:21:08 -07:00
|
|
|
gfs2_log_unlock(sdp);
|
2006-07-11 07:46:33 -06:00
|
|
|
|
|
|
|
return try_to_free_buffers(page);
|
2011-05-03 04:49:19 -06:00
|
|
|
|
|
|
|
not_possible: /* Should never happen */
|
|
|
|
WARN_ON(buffer_dirty(bh));
|
|
|
|
WARN_ON(buffer_pinned(bh));
|
2007-08-16 09:03:57 -06:00
|
|
|
cannot_release:
|
2011-07-14 01:59:44 -06:00
|
|
|
spin_unlock(&sdp->sd_ail_lock);
|
2007-08-16 09:03:57 -06:00
|
|
|
gfs2_log_unlock(sdp);
|
|
|
|
return 0;
|
2006-07-11 07:46:33 -06:00
|
|
|
}
|
|
|
|
|
2007-10-17 01:47:38 -06:00
|
|
|
static const struct address_space_operations gfs2_writeback_aops = {
|
2013-08-27 14:22:07 -06:00
|
|
|
.writepage = gfs2_writepage,
|
2013-01-28 02:30:07 -07:00
|
|
|
.writepages = gfs2_writepages,
|
2007-10-17 01:47:38 -06:00
|
|
|
.readpage = gfs2_readpage,
|
|
|
|
.readpages = gfs2_readpages,
|
|
|
|
.write_begin = gfs2_write_begin,
|
|
|
|
.write_end = gfs2_write_end,
|
|
|
|
.bmap = gfs2_bmap,
|
|
|
|
.invalidatepage = gfs2_invalidatepage,
|
|
|
|
.releasepage = gfs2_releasepage,
|
|
|
|
.direct_IO = gfs2_direct_IO,
|
2008-01-03 04:31:38 -07:00
|
|
|
.migratepage = buffer_migrate_page,
|
2009-03-02 19:45:20 -07:00
|
|
|
.is_partially_uptodate = block_is_partially_uptodate,
|
2009-09-16 03:50:16 -06:00
|
|
|
.error_remove_page = generic_error_remove_page,
|
2007-10-17 01:47:38 -06:00
|
|
|
};
|
|
|
|
|
|
|
|
static const struct address_space_operations gfs2_ordered_aops = {
|
2013-08-27 14:22:07 -06:00
|
|
|
.writepage = gfs2_writepage,
|
2013-01-28 02:30:07 -07:00
|
|
|
.writepages = gfs2_writepages,
|
2006-01-16 09:50:04 -07:00
|
|
|
.readpage = gfs2_readpage,
|
2006-05-05 14:59:11 -06:00
|
|
|
.readpages = gfs2_readpages,
|
2007-10-16 02:25:07 -06:00
|
|
|
.write_begin = gfs2_write_begin,
|
|
|
|
.write_end = gfs2_write_end,
|
2007-06-12 10:24:36 -06:00
|
|
|
.set_page_dirty = gfs2_set_page_dirty,
|
2006-01-16 09:50:04 -07:00
|
|
|
.bmap = gfs2_bmap,
|
|
|
|
.invalidatepage = gfs2_invalidatepage,
|
2006-07-11 07:46:33 -06:00
|
|
|
.releasepage = gfs2_releasepage,
|
2006-01-16 09:50:04 -07:00
|
|
|
.direct_IO = gfs2_direct_IO,
|
2008-01-03 04:31:38 -07:00
|
|
|
.migratepage = buffer_migrate_page,
|
2009-03-02 19:45:20 -07:00
|
|
|
.is_partially_uptodate = block_is_partially_uptodate,
|
2009-09-16 03:50:16 -06:00
|
|
|
.error_remove_page = generic_error_remove_page,
|
2006-01-16 09:50:04 -07:00
|
|
|
};
|
|
|
|
|
2007-10-17 01:47:38 -06:00
|
|
|
static const struct address_space_operations gfs2_jdata_aops = {
|
2007-09-28 06:49:05 -06:00
|
|
|
.writepage = gfs2_jdata_writepage,
|
2007-10-17 02:04:24 -06:00
|
|
|
.writepages = gfs2_jdata_writepages,
|
2007-10-17 01:47:38 -06:00
|
|
|
.readpage = gfs2_readpage,
|
|
|
|
.readpages = gfs2_readpages,
|
|
|
|
.write_begin = gfs2_write_begin,
|
|
|
|
.write_end = gfs2_write_end,
|
|
|
|
.set_page_dirty = gfs2_set_page_dirty,
|
|
|
|
.bmap = gfs2_bmap,
|
|
|
|
.invalidatepage = gfs2_invalidatepage,
|
|
|
|
.releasepage = gfs2_releasepage,
|
2009-03-02 19:45:20 -07:00
|
|
|
.is_partially_uptodate = block_is_partially_uptodate,
|
2009-09-16 03:50:16 -06:00
|
|
|
.error_remove_page = generic_error_remove_page,
|
2007-10-17 01:47:38 -06:00
|
|
|
};
|
|
|
|
|
|
|
|
void gfs2_set_aops(struct inode *inode)
|
|
|
|
{
|
|
|
|
struct gfs2_inode *ip = GFS2_I(inode);
|
|
|
|
|
|
|
|
if (gfs2_is_writeback(ip))
|
|
|
|
inode->i_mapping->a_ops = &gfs2_writeback_aops;
|
|
|
|
else if (gfs2_is_ordered(ip))
|
|
|
|
inode->i_mapping->a_ops = &gfs2_ordered_aops;
|
|
|
|
else if (gfs2_is_jdata(ip))
|
|
|
|
inode->i_mapping->a_ops = &gfs2_jdata_aops;
|
|
|
|
else
|
|
|
|
BUG();
|
|
|
|
}
|
|
|
|
|