* out in so that future reads from that region will get
* zero's.
*/
- struct page *w_pages[OCFS2_MAX_CTXT_PAGES];
unsigned int w_num_pages;
+ struct page *w_pages[OCFS2_MAX_CTXT_PAGES];
struct page *w_target_page;
/*
return ret;
}
- int ocfs2_write_begin_nolock(struct address_space *mapping,
+ int ocfs2_write_begin_nolock(struct file *filp,
+ struct address_space *mapping,
loff_t pos, unsigned len, unsigned flags,
struct page **pagep, void **fsdata,
struct buffer_head *di_bh, struct page *mmap_page)
mlog_errno(ret);
goto out;
} else if (ret == 1) {
- ret = ocfs2_refcount_cow(inode, di_bh,
+ ret = ocfs2_refcount_cow(inode, filp, di_bh,
wc->w_cpos, wc->w_clen, UINT_MAX);
if (ret) {
mlog_errno(ret);
*/
down_write(&OCFS2_I(inode)->ip_alloc_sem);
- ret = ocfs2_write_begin_nolock(mapping, pos, len, flags, pagep,
+ ret = ocfs2_write_begin_nolock(file, mapping, pos, len, flags, pagep,
fsdata, di_bh, NULL);
if (ret) {
mlog_errno(ret);
#include <linux/writeback.h>
#include <linux/falloc.h>
#include <linux/quotaops.h>
+#include <linux/blkdev.h>
#define MLOG_MASK_PREFIX ML_INODE
#include <cluster/masklog.h>
#include "buffer_head_io.h"
-static int ocfs2_sync_inode(struct inode *inode)
-{
- filemap_fdatawrite(inode->i_mapping);
- return sync_mapping_buffers(inode->i_mapping);
-}
-
static int ocfs2_init_file_private(struct inode *inode, struct file *file)
{
struct ocfs2_file_private *fp;
mlog_entry("(0x%p, 0x%p, %d, '%.*s')\n", file, dentry, datasync,
dentry->d_name.len, dentry->d_name.name);
- err = ocfs2_sync_inode(dentry->d_inode);
- if (err)
- goto bail;
-
- if (datasync && !(inode->i_state & I_DIRTY_DATASYNC))
+ if (datasync && !(inode->i_state & I_DIRTY_DATASYNC)) {
+ /*
+ * We still have to flush drive's caches to get data to the
+ * platter
+ */
+ if (osb->s_mount_opt & OCFS2_MOUNT_BARRIER)
+ blkdev_issue_flush(inode->i_sb->s_bdev, GFP_KERNEL,
+ NULL, BLKDEV_IFL_WAIT);
goto bail;
+ }
journal = osb->journal->j_journal;
err = jbd2_journal_force_commit(journal);
if (!(ext_flags & OCFS2_EXT_REFCOUNTED))
goto out;
- return ocfs2_refcount_cow(inode, fe_bh, cpos, 1, cpos+1);
+ return ocfs2_refcount_cow(inode, NULL, fe_bh, cpos, 1, cpos+1);
out:
return status;
BUG_ON(abs_to > (((u64)index + 1) << PAGE_CACHE_SHIFT));
BUG_ON(abs_from & (inode->i_blkbits - 1));
- page = grab_cache_page(mapping, index);
+ page = find_or_create_page(mapping, index, GFP_NOFS);
if (!page) {
ret = -ENOMEM;
mlog_errno(ret);
zero_clusters = last_cpos - zero_cpos;
if (needs_cow) {
- rc = ocfs2_refcount_cow(inode, di_bh, zero_cpos, zero_clusters,
- UINT_MAX);
+ rc = ocfs2_refcount_cow(inode, NULL, di_bh, zero_cpos,
+ zero_clusters, UINT_MAX);
if (rc) {
mlog_errno(rc);
goto out;
}
static int ocfs2_prepare_inode_for_refcount(struct inode *inode,
+ struct file *file,
loff_t pos, size_t count,
int *meta_level)
{
*meta_level = 1;
- ret = ocfs2_refcount_cow(inode, di_bh, cpos, clusters, UINT_MAX);
+ ret = ocfs2_refcount_cow(inode, file, di_bh, cpos, clusters, UINT_MAX);
if (ret)
mlog_errno(ret);
out:
return ret;
}
- static int ocfs2_prepare_inode_for_write(struct dentry *dentry,
+ static int ocfs2_prepare_inode_for_write(struct file *file,
loff_t *ppos,
size_t count,
int appending,
int *has_refcount)
{
int ret = 0, meta_level = 0;
+ struct dentry *dentry = file->f_path.dentry;
struct inode *inode = dentry->d_inode;
loff_t saved_pos, end;
meta_level = -1;
ret = ocfs2_prepare_inode_for_refcount(inode,
+ file,
saved_pos,
count,
&meta_level);
}
can_do_direct = direct_io;
- ret = ocfs2_prepare_inode_for_write(file->f_path.dentry, ppos,
+ ret = ocfs2_prepare_inode_for_write(file, ppos,
iocb->ki_left, appending,
&can_do_direct, &has_refcount);
if (ret < 0) {
written = generic_file_direct_write(iocb, iov, &nr_segs, *ppos,
ppos, count, ocount);
if (written < 0) {
- /*
- * direct write may have instantiated a few
- * blocks outside i_size. Trim these off again.
- * Don't need i_size_read because we hold i_mutex.
- *
- * XXX(truncate): this looks buggy because ocfs2 did not
- * actually implement ->truncate. Take a look at
- * the new truncate sequence and update this accordingly
- */
- if (*ppos + count > inode->i_size)
- truncate_setsize(inode, inode->i_size);
ret = written;
goto out_dio;
}
BUG_ON(ret == -EIOCBQUEUED && !(file->f_flags & O_DIRECT));
if (((file->f_flags & O_DSYNC) && !direct_io) || IS_SYNC(inode) ||
- ((file->f_flags & O_DIRECT) && has_refcount)) {
+ ((file->f_flags & O_DIRECT) && !direct_io)) {
ret = filemap_fdatawrite_range(file->f_mapping, pos,
pos + count - 1);
if (ret < 0)
{
int ret;
- ret = ocfs2_prepare_inode_for_write(out->f_path.dentry, &sd->pos,
+ ret = ocfs2_prepare_inode_for_write(out, &sd->pos,
sd->total_len, 0, NULL, NULL);
if (ret < 0) {
mlog_errno(ret);
return ret;
}
- static int __ocfs2_page_mkwrite(struct inode *inode, struct buffer_head *di_bh,
+ static int __ocfs2_page_mkwrite(struct file *file, struct buffer_head *di_bh,
struct page *page)
{
int ret;
+ struct inode *inode = file->f_path.dentry->d_inode;
struct address_space *mapping = inode->i_mapping;
loff_t pos = page_offset(page);
unsigned int len = PAGE_CACHE_SIZE;
/*
* Another node might have truncated while we were waiting on
* cluster locks.
+ * We don't check size == 0 before the shift. This is borrowed
+ * from do_generic_file_read.
*/
- last_index = size >> PAGE_CACHE_SHIFT;
- if (page->index > last_index) {
+ last_index = (size - 1) >> PAGE_CACHE_SHIFT;
+ if (unlikely(!size || page->index > last_index)) {
ret = -EINVAL;
goto out;
}
* because the "write" would invalidate their data.
*/
if (page->index == last_index)
- len = size & ~PAGE_CACHE_MASK;
+ len = ((size - 1) & ~PAGE_CACHE_MASK) + 1;
- ret = ocfs2_write_begin_nolock(mapping, pos, len, 0, &locked_page,
+ ret = ocfs2_write_begin_nolock(file, mapping, pos, len, 0, &locked_page,
&fsdata, di_bh, page);
if (ret) {
if (ret != -ENOSPC)
*/
down_write(&OCFS2_I(inode)->ip_alloc_sem);
- ret = __ocfs2_page_mkwrite(inode, di_bh, page);
+ ret = __ocfs2_page_mkwrite(vma->vm_file, di_bh, page);
up_write(&OCFS2_I(inode)->ip_alloc_sem);
struct ocfs2_cow_context {
struct inode *inode;
+ struct file *file;
u32 cow_start;
u32 cow_len;
struct ocfs2_extent_tree data_et;
len = min((u64)cpos + clusters, le64_to_cpu(rec.r_cpos) +
le32_to_cpu(rec.r_clusters)) - cpos;
/*
- * If the refcount rec already exist, cool. We just need
- * to check whether there is a split. Otherwise we just need
- * to increase the refcount.
- * If we will insert one, increases recs_add.
- *
* We record all the records which will be inserted to the
* same refcount block, so that we can tell exactly whether
* we need a new refcount block or not.
+ *
+ * If we will insert a new one, this is easy and only happens
+ * during adding refcounted flag to the extent, so we don't
+ * have a chance of spliting. We just need one record.
+ *
+ * If the refcount rec already exists, that would be a little
+ * complicated. we may have to:
+ * 1) split at the beginning if the start pos isn't aligned.
+ * we need 1 more record in this case.
+ * 2) split int the end if the end pos isn't aligned.
+ * we need 1 more record in this case.
+ * 3) split in the middle because of file system fragmentation.
+ * we need 2 more records in this case(we can't detect this
+ * beforehand, so always think of the worst case).
*/
if (rec.r_refcount) {
+ recs_add += 2;
/* Check whether we need a split at the beginning. */
if (cpos == start_cpos &&
cpos != le64_to_cpu(rec.r_cpos))
u64 new_block = ocfs2_clusters_to_blocks(sb, new_cluster);
struct page *page;
pgoff_t page_index;
- unsigned int from, to;
+ unsigned int from, to, readahead_pages;
loff_t offset, end, map_end;
struct address_space *mapping = context->inode->i_mapping;
mlog(0, "old_cluster %u, new %u, len %u at offset %u\n", old_cluster,
new_cluster, new_len, cpos);
+ readahead_pages =
+ (ocfs2_cow_contig_clusters(sb) <<
+ OCFS2_SB(sb)->s_clustersize_bits) >> PAGE_CACHE_SHIFT;
offset = ((loff_t)cpos) << OCFS2_SB(sb)->s_clustersize_bits;
end = offset + (new_len << OCFS2_SB(sb)->s_clustersize_bits);
/*
if (map_end & (PAGE_CACHE_SIZE - 1))
to = map_end & (PAGE_CACHE_SIZE - 1);
- page = grab_cache_page(mapping, page_index);
+ page = find_or_create_page(mapping, page_index, GFP_NOFS);
/*
* In case PAGE_CACHE_SIZE <= CLUSTER_SIZE, This page
if (PAGE_CACHE_SIZE <= OCFS2_SB(sb)->s_clustersize)
BUG_ON(PageDirty(page));
+ if (PageReadahead(page) && context->file) {
+ page_cache_async_readahead(mapping,
+ &context->file->f_ra,
+ context->file,
+ page, page_index,
+ readahead_pages);
+ }
+
if (!PageUptodate(page)) {
ret = block_read_full_page(page, ocfs2_get_block);
if (ret) {
if (map_end > end)
map_end = end;
- page = grab_cache_page(context->inode->i_mapping, page_index);
+ page = find_or_create_page(context->inode->i_mapping,
+ page_index, GFP_NOFS);
BUG_ON(!page);
wait_on_page_writeback(page);
return ret;
}
+ static void ocfs2_readahead_for_cow(struct inode *inode,
+ struct file *file,
+ u32 start, u32 len)
+ {
+ struct address_space *mapping;
+ pgoff_t index;
+ unsigned long num_pages;
+ int cs_bits = OCFS2_SB(inode->i_sb)->s_clustersize_bits;
+
+ if (!file)
+ return;
+
+ mapping = file->f_mapping;
+ num_pages = (len << cs_bits) >> PAGE_CACHE_SHIFT;
+ if (!num_pages)
+ num_pages = 1;
+
+ index = ((loff_t)start << cs_bits) >> PAGE_CACHE_SHIFT;
+ page_cache_sync_readahead(mapping, &file->f_ra, file,
+ index, num_pages);
+ }
+
/*
* Starting at cpos, try to CoW write_len clusters. Don't CoW
* past max_cpos. This will stop when it runs into a hole or an
* unrefcounted extent.
*/
static int ocfs2_refcount_cow_hunk(struct inode *inode,
+ struct file *file,
struct buffer_head *di_bh,
u32 cpos, u32 write_len, u32 max_cpos)
{
BUG_ON(cow_len == 0);
+ ocfs2_readahead_for_cow(inode, file, cow_start, cow_len);
+
context = kzalloc(sizeof(struct ocfs2_cow_context), GFP_NOFS);
if (!context) {
ret = -ENOMEM;
context->ref_root_bh = ref_root_bh;
context->cow_duplicate_clusters = ocfs2_duplicate_clusters_by_page;
context->get_clusters = ocfs2_di_get_clusters;
+ context->file = file;
ocfs2_init_dinode_extent_tree(&context->data_et,
INODE_CACHE(inode), di_bh);
* clusters between cpos and cpos+write_len are safe to modify.
*/
int ocfs2_refcount_cow(struct inode *inode,
+ struct file *file,
struct buffer_head *di_bh,
u32 cpos, u32 write_len, u32 max_cpos)
{
num_clusters = write_len;
if (ext_flags & OCFS2_EXT_REFCOUNTED) {
- ret = ocfs2_refcount_cow_hunk(inode, di_bh, cpos,
+ ret = ocfs2_refcount_cow_hunk(inode, file, di_bh, cpos,
num_clusters, max_cpos);
if (ret) {
mlog_errno(ret);
struct rb_node rf_node;
u64 rf_blkno;
u32 rf_generation;
+ struct kref rf_getcnt;
struct rw_semaphore rf_sem;
struct ocfs2_lock_res rf_lockres;
- struct kref rf_getcnt;
int rf_removed;
/* the following 4 fields are used by caching_info. */
- struct ocfs2_caching_info rf_ci;
spinlock_t rf_lock;
+ struct ocfs2_caching_info rf_ci;
struct mutex rf_io_mutex;
struct super_block *rf_sb;
};
u32 clusters,
int *credits,
int *ref_blocks);
- int ocfs2_refcount_cow(struct inode *inode, struct buffer_head *di_bh,
+ int ocfs2_refcount_cow(struct inode *inode,
+ struct file *filep, struct buffer_head *di_bh,
u32 cpos, u32 write_len, u32 max_cpos);
typedef int (ocfs2_post_refcount_func)(struct inode *inode,