aboutsummaryrefslogtreecommitdiff
path: root/fs/logfs
diff options
context:
space:
mode:
Diffstat (limited to 'fs/logfs')
-rw-r--r--fs/logfs/Kconfig4
-rw-r--r--fs/logfs/compr.c2
-rw-r--r--fs/logfs/dev_bdev.c102
-rw-r--r--fs/logfs/dev_mtd.c114
-rw-r--r--fs/logfs/dir.c120
-rw-r--r--fs/logfs/file.c68
-rw-r--r--fs/logfs/gc.c60
-rw-r--r--fs/logfs/inode.c113
-rw-r--r--fs/logfs/journal.c47
-rw-r--r--fs/logfs/logfs.h70
-rw-r--r--fs/logfs/logfs_abi.h10
-rw-r--r--fs/logfs/readwrite.c275
-rw-r--r--fs/logfs/segment.c134
-rw-r--r--fs/logfs/super.c179
14 files changed, 717 insertions, 581 deletions
diff --git a/fs/logfs/Kconfig b/fs/logfs/Kconfig
index daf9a9b32dd..09ed066c022 100644
--- a/fs/logfs/Kconfig
+++ b/fs/logfs/Kconfig
@@ -1,6 +1,6 @@
config LOGFS
- tristate "LogFS file system (EXPERIMENTAL)"
- depends on (MTD || BLOCK) && EXPERIMENTAL
+ tristate "LogFS file system"
+ depends on (MTD || BLOCK)
select ZLIB_INFLATE
select ZLIB_DEFLATE
select CRC32
diff --git a/fs/logfs/compr.c b/fs/logfs/compr.c
index 44bbfd249ab..961f02b86d9 100644
--- a/fs/logfs/compr.c
+++ b/fs/logfs/compr.c
@@ -81,7 +81,7 @@ error:
int __init logfs_compr_init(void)
{
- size_t size = max(zlib_deflate_workspacesize(),
+ size_t size = max(zlib_deflate_workspacesize(MAX_WBITS, MAX_MEM_LEVEL),
zlib_inflate_workspacesize());
stream.workspace = vmalloc(size);
if (!stream.workspace)
diff --git a/fs/logfs/dev_bdev.c b/fs/logfs/dev_bdev.c
index 9718c22f186..76279e11982 100644
--- a/fs/logfs/dev_bdev.c
+++ b/fs/logfs/dev_bdev.c
@@ -9,38 +9,28 @@
#include <linux/bio.h>
#include <linux/blkdev.h>
#include <linux/buffer_head.h>
+#include <linux/gfp.h>
+#include <linux/prefetch.h>
#define PAGE_OFS(ofs) ((ofs) & (PAGE_SIZE-1))
-static void request_complete(struct bio *bio, int err)
-{
- complete((struct completion *)bio->bi_private);
-}
-
static int sync_request(struct page *page, struct block_device *bdev, int rw)
{
struct bio bio;
struct bio_vec bio_vec;
- struct completion complete;
bio_init(&bio);
+ bio.bi_max_vecs = 1;
bio.bi_io_vec = &bio_vec;
bio_vec.bv_page = page;
bio_vec.bv_len = PAGE_SIZE;
bio_vec.bv_offset = 0;
bio.bi_vcnt = 1;
- bio.bi_idx = 0;
- bio.bi_size = PAGE_SIZE;
bio.bi_bdev = bdev;
- bio.bi_sector = page->index * (PAGE_SIZE >> 9);
- init_completion(&complete);
- bio.bi_private = &complete;
- bio.bi_end_io = request_complete;
-
- submit_bio(rw, &bio);
- generic_unplug_device(bdev_get_queue(bdev));
- wait_for_completion(&complete);
- return test_bit(BIO_UPTODATE, &bio.bi_flags) ? 0 : -EIO;
+ bio.bi_iter.bi_sector = page->index * (PAGE_SIZE >> 9);
+ bio.bi_iter.bi_size = PAGE_SIZE;
+
+ return submit_bio_wait(rw, &bio);
}
static int bdev_readpage(void *_sb, struct page *page)
@@ -66,21 +56,18 @@ static DECLARE_WAIT_QUEUE_HEAD(wq);
static void writeseg_end_io(struct bio *bio, int err)
{
const int uptodate = test_bit(BIO_UPTODATE, &bio->bi_flags);
- struct bio_vec *bvec = bio->bi_io_vec + bio->bi_vcnt - 1;
+ struct bio_vec *bvec;
+ int i;
struct super_block *sb = bio->bi_private;
struct logfs_super *super = logfs_super(sb);
- struct page *page;
BUG_ON(!uptodate); /* FIXME: Retry io or write elsewhere */
BUG_ON(err);
- BUG_ON(bio->bi_vcnt == 0);
- do {
- page = bvec->bv_page;
- if (--bvec >= bio->bi_io_vec)
- prefetchw(&bvec->bv_page->flags);
-
- end_page_writeback(page);
- } while (bvec >= bio->bi_io_vec);
+
+ bio_for_each_segment_all(bvec, bio, i) {
+ end_page_writeback(bvec->bv_page);
+ page_cache_release(bvec->bv_page);
+ }
bio_put(bio);
if (atomic_dec_and_test(&super->s_pending_writes))
wake_up(&wq);
@@ -93,21 +80,21 @@ static int __bdev_writeseg(struct super_block *sb, u64 ofs, pgoff_t index,
struct address_space *mapping = super->s_mapping_inode->i_mapping;
struct bio *bio;
struct page *page;
- struct request_queue *q = bdev_get_queue(sb->s_bdev);
- unsigned int max_pages = queue_max_hw_sectors(q) >> (PAGE_SHIFT - 9);
+ unsigned int max_pages;
int i;
+ max_pages = min(nr_pages, (size_t) bio_get_nr_vecs(super->s_bdev));
+
bio = bio_alloc(GFP_NOFS, max_pages);
- BUG_ON(!bio); /* FIXME: handle this */
+ BUG_ON(!bio);
for (i = 0; i < nr_pages; i++) {
if (i >= max_pages) {
/* Block layer cannot split bios :( */
bio->bi_vcnt = i;
- bio->bi_idx = 0;
- bio->bi_size = i * PAGE_SIZE;
+ bio->bi_iter.bi_size = i * PAGE_SIZE;
bio->bi_bdev = super->s_bdev;
- bio->bi_sector = ofs >> 9;
+ bio->bi_iter.bi_sector = ofs >> 9;
bio->bi_private = sb;
bio->bi_end_io = writeseg_end_io;
atomic_inc(&super->s_pending_writes);
@@ -132,10 +119,9 @@ static int __bdev_writeseg(struct super_block *sb, u64 ofs, pgoff_t index,
unlock_page(page);
}
bio->bi_vcnt = nr_pages;
- bio->bi_idx = 0;
- bio->bi_size = nr_pages * PAGE_SIZE;
+ bio->bi_iter.bi_size = nr_pages * PAGE_SIZE;
bio->bi_bdev = super->s_bdev;
- bio->bi_sector = ofs >> 9;
+ bio->bi_iter.bi_sector = ofs >> 9;
bio->bi_private = sb;
bio->bi_end_io = writeseg_end_io;
atomic_inc(&super->s_pending_writes);
@@ -164,7 +150,6 @@ static void bdev_writeseg(struct super_block *sb, u64 ofs, size_t len)
}
len = PAGE_ALIGN(len);
__bdev_writeseg(sb, ofs, ofs >> PAGE_SHIFT, len >> PAGE_SHIFT);
- generic_unplug_device(bdev_get_queue(logfs_super(sb)->s_bdev));
}
@@ -187,21 +172,21 @@ static int do_erase(struct super_block *sb, u64 ofs, pgoff_t index,
{
struct logfs_super *super = logfs_super(sb);
struct bio *bio;
- struct request_queue *q = bdev_get_queue(sb->s_bdev);
- unsigned int max_pages = queue_max_hw_sectors(q) >> (PAGE_SHIFT - 9);
+ unsigned int max_pages;
int i;
+ max_pages = min(nr_pages, (size_t) bio_get_nr_vecs(super->s_bdev));
+
bio = bio_alloc(GFP_NOFS, max_pages);
- BUG_ON(!bio); /* FIXME: handle this */
+ BUG_ON(!bio);
for (i = 0; i < nr_pages; i++) {
if (i >= max_pages) {
/* Block layer cannot split bios :( */
bio->bi_vcnt = i;
- bio->bi_idx = 0;
- bio->bi_size = i * PAGE_SIZE;
+ bio->bi_iter.bi_size = i * PAGE_SIZE;
bio->bi_bdev = super->s_bdev;
- bio->bi_sector = ofs >> 9;
+ bio->bi_iter.bi_sector = ofs >> 9;
bio->bi_private = sb;
bio->bi_end_io = erase_end_io;
atomic_inc(&super->s_pending_writes);
@@ -220,10 +205,9 @@ static int do_erase(struct super_block *sb, u64 ofs, pgoff_t index,
bio->bi_io_vec[i].bv_offset = 0;
}
bio->bi_vcnt = nr_pages;
- bio->bi_idx = 0;
- bio->bi_size = nr_pages * PAGE_SIZE;
+ bio->bi_iter.bi_size = nr_pages * PAGE_SIZE;
bio->bi_bdev = super->s_bdev;
- bio->bi_sector = ofs >> 9;
+ bio->bi_iter.bi_sector = ofs >> 9;
bio->bi_private = sb;
bio->bi_end_io = erase_end_io;
atomic_inc(&super->s_pending_writes);
@@ -292,9 +276,14 @@ static int bdev_write_sb(struct super_block *sb, struct page *page)
return sync_request(page, bdev, WRITE);
}
-static void bdev_put_device(struct super_block *sb)
+static void bdev_put_device(struct logfs_super *s)
{
- close_bdev_exclusive(logfs_super(sb)->s_bdev, FMODE_READ|FMODE_WRITE);
+ blkdev_put(s->s_bdev, FMODE_READ|FMODE_WRITE|FMODE_EXCL);
+}
+
+static int bdev_can_write_buf(struct super_block *sb, u64 ofs)
+{
+ return 0;
}
static const struct logfs_device_ops bd_devops = {
@@ -304,24 +293,29 @@ static const struct logfs_device_ops bd_devops = {
.readpage = bdev_readpage,
.writeseg = bdev_writeseg,
.erase = bdev_erase,
+ .can_write_buf = bdev_can_write_buf,
.sync = bdev_sync,
.put_device = bdev_put_device,
};
-int logfs_get_sb_bdev(struct file_system_type *type, int flags,
- const char *devname, struct vfsmount *mnt)
+int logfs_get_sb_bdev(struct logfs_super *p, struct file_system_type *type,
+ const char *devname)
{
struct block_device *bdev;
- bdev = open_bdev_exclusive(devname, FMODE_READ|FMODE_WRITE, type);
+ bdev = blkdev_get_by_path(devname, FMODE_READ|FMODE_WRITE|FMODE_EXCL,
+ type);
if (IS_ERR(bdev))
return PTR_ERR(bdev);
if (MAJOR(bdev->bd_dev) == MTD_BLOCK_MAJOR) {
int mtdnr = MINOR(bdev->bd_dev);
- close_bdev_exclusive(bdev, FMODE_READ|FMODE_WRITE);
- return logfs_get_sb_mtd(type, flags, mtdnr, mnt);
+ blkdev_put(bdev, FMODE_READ|FMODE_WRITE|FMODE_EXCL);
+ return logfs_get_sb_mtd(p, mtdnr);
}
- return logfs_get_sb_device(type, flags, NULL, bdev, &bd_devops, mnt);
+ p->s_bdev = bdev;
+ p->s_mtd = NULL;
+ p->s_devops = &bd_devops;
+ return 0;
}
diff --git a/fs/logfs/dev_mtd.c b/fs/logfs/dev_mtd.c
index cafb6ef2e05..9c501449450 100644
--- a/fs/logfs/dev_mtd.c
+++ b/fs/logfs/dev_mtd.c
@@ -9,16 +9,18 @@
#include <linux/completion.h>
#include <linux/mount.h>
#include <linux/sched.h>
+#include <linux/slab.h>
#define PAGE_OFS(ofs) ((ofs) & (PAGE_SIZE-1))
-static int mtd_read(struct super_block *sb, loff_t ofs, size_t len, void *buf)
+static int logfs_mtd_read(struct super_block *sb, loff_t ofs, size_t len,
+ void *buf)
{
struct mtd_info *mtd = logfs_super(sb)->s_mtd;
size_t retlen;
int ret;
- ret = mtd->read(mtd, ofs, len, &retlen, buf);
+ ret = mtd_read(mtd, ofs, len, &retlen, buf);
BUG_ON(ret == -EINVAL);
if (ret)
return ret;
@@ -30,7 +32,8 @@ static int mtd_read(struct super_block *sb, loff_t ofs, size_t len, void *buf)
return 0;
}
-static int mtd_write(struct super_block *sb, loff_t ofs, size_t len, void *buf)
+static int loffs_mtd_write(struct super_block *sb, loff_t ofs, size_t len,
+ void *buf)
{
struct logfs_super *super = logfs_super(sb);
struct mtd_info *mtd = super->s_mtd;
@@ -46,7 +49,7 @@ static int mtd_write(struct super_block *sb, loff_t ofs, size_t len, void *buf)
BUG_ON(len > PAGE_CACHE_SIZE);
page_start = ofs & PAGE_CACHE_MASK;
page_end = PAGE_CACHE_ALIGN(ofs + len) - 1;
- ret = mtd->write(mtd, ofs, len, &retlen, buf);
+ ret = mtd_write(mtd, ofs, len, &retlen, buf);
if (ret || (retlen != len))
return -EIO;
@@ -59,14 +62,15 @@ static int mtd_write(struct super_block *sb, loff_t ofs, size_t len, void *buf)
* asynchronous properties. So just to prevent the first implementor of such
* a thing from breaking logfs in 2350, we do the usual pointless dance to
* declare a completion variable and wait for completion before returning
- * from mtd_erase(). What an excercise in futility!
+ * from logfs_mtd_erase(). What an exercise in futility!
*/
static void logfs_erase_callback(struct erase_info *ei)
{
complete((struct completion *)ei->priv);
}
-static int mtd_erase_mapping(struct super_block *sb, loff_t ofs, size_t len)
+static int logfs_mtd_erase_mapping(struct super_block *sb, loff_t ofs,
+ size_t len)
{
struct logfs_super *super = logfs_super(sb);
struct address_space *mapping = super->s_mapping_inode->i_mapping;
@@ -83,7 +87,7 @@ static int mtd_erase_mapping(struct super_block *sb, loff_t ofs, size_t len)
return 0;
}
-static int mtd_erase(struct super_block *sb, loff_t ofs, size_t len,
+static int logfs_mtd_erase(struct super_block *sb, loff_t ofs, size_t len,
int ensure_write)
{
struct mtd_info *mtd = logfs_super(sb)->s_mtd;
@@ -101,32 +105,32 @@ static int mtd_erase(struct super_block *sb, loff_t ofs, size_t len,
ei.len = len;
ei.callback = logfs_erase_callback;
ei.priv = (long)&complete;
- ret = mtd->erase(mtd, &ei);
+ ret = mtd_erase(mtd, &ei);
if (ret)
return -EIO;
wait_for_completion(&complete);
if (ei.state != MTD_ERASE_DONE)
return -EIO;
- return mtd_erase_mapping(sb, ofs, len);
+ return logfs_mtd_erase_mapping(sb, ofs, len);
}
-static void mtd_sync(struct super_block *sb)
+static void logfs_mtd_sync(struct super_block *sb)
{
struct mtd_info *mtd = logfs_super(sb)->s_mtd;
- if (mtd->sync)
- mtd->sync(mtd);
+ mtd_sync(mtd);
}
-static int mtd_readpage(void *_sb, struct page *page)
+static int logfs_mtd_readpage(void *_sb, struct page *page)
{
struct super_block *sb = _sb;
int err;
- err = mtd_read(sb, page->index << PAGE_SHIFT, PAGE_SIZE,
+ err = logfs_mtd_read(sb, page->index << PAGE_SHIFT, PAGE_SIZE,
page_address(page));
- if (err == -EUCLEAN) {
+ if (err == -EUCLEAN || err == -EBADMSG) {
+ /* -EBADMSG happens regularly on power failures */
err = 0;
/* FIXME: force GC this segment */
}
@@ -141,18 +145,15 @@ static int mtd_readpage(void *_sb, struct page *page)
return err;
}
-static struct page *mtd_find_first_sb(struct super_block *sb, u64 *ofs)
+static struct page *logfs_mtd_find_first_sb(struct super_block *sb, u64 *ofs)
{
struct logfs_super *super = logfs_super(sb);
struct address_space *mapping = super->s_mapping_inode->i_mapping;
- filler_t *filler = mtd_readpage;
+ filler_t *filler = logfs_mtd_readpage;
struct mtd_info *mtd = super->s_mtd;
- if (!mtd->block_isbad)
- return NULL;
-
*ofs = 0;
- while (mtd->block_isbad(mtd, *ofs)) {
+ while (mtd_block_isbad(mtd, *ofs)) {
*ofs += mtd->erasesize;
if (*ofs >= mtd->size)
return NULL;
@@ -161,18 +162,15 @@ static struct page *mtd_find_first_sb(struct super_block *sb, u64 *ofs)
return read_cache_page(mapping, *ofs >> PAGE_SHIFT, filler, sb);
}
-static struct page *mtd_find_last_sb(struct super_block *sb, u64 *ofs)
+static struct page *logfs_mtd_find_last_sb(struct super_block *sb, u64 *ofs)
{
struct logfs_super *super = logfs_super(sb);
struct address_space *mapping = super->s_mapping_inode->i_mapping;
- filler_t *filler = mtd_readpage;
+ filler_t *filler = logfs_mtd_readpage;
struct mtd_info *mtd = super->s_mtd;
- if (!mtd->block_isbad)
- return NULL;
-
*ofs = mtd->size - mtd->erasesize;
- while (mtd->block_isbad(mtd, *ofs)) {
+ while (mtd_block_isbad(mtd, *ofs)) {
*ofs -= mtd->erasesize;
if (*ofs <= 0)
return NULL;
@@ -182,7 +180,7 @@ static struct page *mtd_find_last_sb(struct super_block *sb, u64 *ofs)
return read_cache_page(mapping, *ofs >> PAGE_SHIFT, filler, sb);
}
-static int __mtd_writeseg(struct super_block *sb, u64 ofs, pgoff_t index,
+static int __logfs_mtd_writeseg(struct super_block *sb, u64 ofs, pgoff_t index,
size_t nr_pages)
{
struct logfs_super *super = logfs_super(sb);
@@ -194,8 +192,8 @@ static int __mtd_writeseg(struct super_block *sb, u64 ofs, pgoff_t index,
page = find_lock_page(mapping, index + i);
BUG_ON(!page);
- err = mtd_write(sb, page->index << PAGE_SHIFT, PAGE_SIZE,
- page_address(page));
+ err = loffs_mtd_write(sb, page->index << PAGE_SHIFT, PAGE_SIZE,
+ page_address(page));
unlock_page(page);
page_cache_release(page);
if (err)
@@ -204,7 +202,7 @@ static int __mtd_writeseg(struct super_block *sb, u64 ofs, pgoff_t index,
return 0;
}
-static void mtd_writeseg(struct super_block *sb, u64 ofs, size_t len)
+static void logfs_mtd_writeseg(struct super_block *sb, u64 ofs, size_t len)
{
struct logfs_super *super = logfs_super(sb);
int head;
@@ -225,30 +223,52 @@ static void mtd_writeseg(struct super_block *sb, u64 ofs, size_t len)
len += head;
}
len = PAGE_ALIGN(len);
- __mtd_writeseg(sb, ofs, ofs >> PAGE_SHIFT, len >> PAGE_SHIFT);
+ __logfs_mtd_writeseg(sb, ofs, ofs >> PAGE_SHIFT, len >> PAGE_SHIFT);
+}
+
+static void logfs_mtd_put_device(struct logfs_super *s)
+{
+ put_mtd_device(s->s_mtd);
}
-static void mtd_put_device(struct super_block *sb)
+static int logfs_mtd_can_write_buf(struct super_block *sb, u64 ofs)
{
- put_mtd_device(logfs_super(sb)->s_mtd);
+ struct logfs_super *super = logfs_super(sb);
+ void *buf;
+ int err;
+
+ buf = kmalloc(super->s_writesize, GFP_KERNEL);
+ if (!buf)
+ return -ENOMEM;
+ err = logfs_mtd_read(sb, ofs, super->s_writesize, buf);
+ if (err)
+ goto out;
+ if (memchr_inv(buf, 0xff, super->s_writesize))
+ err = -EIO;
+ kfree(buf);
+out:
+ return err;
}
static const struct logfs_device_ops mtd_devops = {
- .find_first_sb = mtd_find_first_sb,
- .find_last_sb = mtd_find_last_sb,
- .readpage = mtd_readpage,
- .writeseg = mtd_writeseg,
- .erase = mtd_erase,
- .sync = mtd_sync,
- .put_device = mtd_put_device,
+ .find_first_sb = logfs_mtd_find_first_sb,
+ .find_last_sb = logfs_mtd_find_last_sb,
+ .readpage = logfs_mtd_readpage,
+ .writeseg = logfs_mtd_writeseg,
+ .erase = logfs_mtd_erase,
+ .can_write_buf = logfs_mtd_can_write_buf,
+ .sync = logfs_mtd_sync,
+ .put_device = logfs_mtd_put_device,
};
-int logfs_get_sb_mtd(struct file_system_type *type, int flags,
- int mtdnr, struct vfsmount *mnt)
+int logfs_get_sb_mtd(struct logfs_super *s, int mtdnr)
{
- struct mtd_info *mtd;
- const struct logfs_device_ops *devops = &mtd_devops;
+ struct mtd_info *mtd = get_mtd_device(NULL, mtdnr);
+ if (IS_ERR(mtd))
+ return PTR_ERR(mtd);
- mtd = get_mtd_device(NULL, mtdnr);
- return logfs_get_sb_device(type, flags, mtd, NULL, devops, mnt);
+ s->s_bdev = NULL;
+ s->s_mtd = mtd;
+ s->s_devops = &mtd_devops;
+ return 0;
}
diff --git a/fs/logfs/dir.c b/fs/logfs/dir.c
index 56a8bfbb012..6bdc347008f 100644
--- a/fs/logfs/dir.c
+++ b/fs/logfs/dir.c
@@ -6,13 +6,13 @@
* Copyright (c) 2005-2008 Joern Engel <joern@logfs.org>
*/
#include "logfs.h"
-
+#include <linux/slab.h>
/*
* Atomic dir operations
*
* Directory operations are by default not atomic. Dentries and Inodes are
- * created/removed/altered in seperate operations. Therefore we need to do
+ * created/removed/altered in separate operations. Therefore we need to do
* a small amount of journaling.
*
* Create, link, mkdir, mknod and symlink all share the same function to do
@@ -71,7 +71,7 @@ static int write_dir(struct inode *dir, struct logfs_disk_dentry *dd,
static int write_inode(struct inode *inode)
{
- return __logfs_write_inode(inode, WF_LOCK);
+ return __logfs_write_inode(inode, NULL, WF_LOCK);
}
static s64 dir_seek_data(struct inode *inode, s64 pos)
@@ -92,7 +92,7 @@ static int beyond_eof(struct inode *inode, loff_t bix)
* so short names (len <= 9) don't even occupy the complete 32bit name
* space. A prime >256 ensures short names quickly spread the 32bit
* name space. Add about 26 for the estimated amount of information
- * of each character and pick a prime nearby, preferrably a bit-sparse
+ * of each character and pick a prime nearby, preferably a bit-sparse
* one.
*/
static u32 hash_32(const char *s, int len, u32 seed)
@@ -177,17 +177,17 @@ static struct page *logfs_get_dd_page(struct inode *dir, struct dentry *dentry)
(filler_t *)logfs_readpage, NULL);
if (IS_ERR(page))
return page;
- dd = kmap_atomic(page, KM_USER0);
+ dd = kmap_atomic(page);
BUG_ON(dd->namelen == 0);
if (name->len != be16_to_cpu(dd->namelen) ||
memcmp(name->name, dd->name, name->len)) {
- kunmap_atomic(dd, KM_USER0);
+ kunmap_atomic(dd);
page_cache_release(page);
continue;
}
- kunmap_atomic(dd, KM_USER0);
+ kunmap_atomic(dd);
return page;
}
return NULL;
@@ -197,7 +197,7 @@ static int logfs_remove_inode(struct inode *inode)
{
int ret;
- inode->i_nlink--;
+ drop_nlink(inode);
ret = write_inode(inode);
LOGFS_BUG_ON(ret, inode->i_sb);
return ret;
@@ -281,17 +281,23 @@ static int logfs_rmdir(struct inode *dir, struct dentry *dentry)
/* FIXME: readdir currently has it's own dir_walk code. I don't see a good
* way to combine the two copies */
-#define IMPLICIT_NODES 2
-static int __logfs_readdir(struct file *file, void *buf, filldir_t filldir)
+static int logfs_readdir(struct file *file, struct dir_context *ctx)
{
- struct inode *dir = file->f_dentry->d_inode;
- loff_t pos = file->f_pos - IMPLICIT_NODES;
+ struct inode *dir = file_inode(file);
+ loff_t pos;
struct page *page;
struct logfs_disk_dentry *dd;
- int full;
+ if (ctx->pos < 0)
+ return -EINVAL;
+
+ if (!dir_emit_dots(file, ctx))
+ return 0;
+
+ pos = ctx->pos - 2;
BUG_ON(pos < 0);
- for (;; pos++) {
+ for (;; pos++, ctx->pos++) {
+ bool full;
if (beyond_eof(dir, pos))
break;
if (!logfs_exist_block(dir, pos)) {
@@ -303,45 +309,20 @@ static int __logfs_readdir(struct file *file, void *buf, filldir_t filldir)
(filler_t *)logfs_readpage, NULL);
if (IS_ERR(page))
return PTR_ERR(page);
- dd = kmap_atomic(page, KM_USER0);
+ dd = kmap(page);
BUG_ON(dd->namelen == 0);
- full = filldir(buf, (char *)dd->name, be16_to_cpu(dd->namelen),
- pos, be64_to_cpu(dd->ino), dd->type);
- kunmap_atomic(dd, KM_USER0);
+ full = !dir_emit(ctx, (char *)dd->name,
+ be16_to_cpu(dd->namelen),
+ be64_to_cpu(dd->ino), dd->type);
+ kunmap(page);
page_cache_release(page);
if (full)
break;
}
-
- file->f_pos = pos + IMPLICIT_NODES;
return 0;
}
-static int logfs_readdir(struct file *file, void *buf, filldir_t filldir)
-{
- struct inode *inode = file->f_dentry->d_inode;
- ino_t pino = parent_ino(file->f_dentry);
- int err;
-
- if (file->f_pos < 0)
- return -EINVAL;
-
- if (file->f_pos == 0) {
- if (filldir(buf, ".", 1, 1, inode->i_ino, DT_DIR) < 0)
- return 0;
- file->f_pos++;
- }
- if (file->f_pos == 1) {
- if (filldir(buf, "..", 2, 2, pino, DT_DIR) < 0)
- return 0;
- file->f_pos++;
- }
-
- err = __logfs_readdir(file, buf, filldir);
- return err;
-}
-
static void logfs_set_name(struct logfs_disk_dentry *dd, struct qstr *name)
{
dd->namelen = cpu_to_be16(name->len);
@@ -349,7 +330,7 @@ static void logfs_set_name(struct logfs_disk_dentry *dd, struct qstr *name)
}
static struct dentry *logfs_lookup(struct inode *dir, struct dentry *dentry,
- struct nameidata *nd)
+ unsigned int flags)
{
struct page *page;
struct logfs_disk_dentry *dd;
@@ -365,17 +346,15 @@ static struct dentry *logfs_lookup(struct inode *dir, struct dentry *dentry,
return NULL;
}
index = page->index;
- dd = kmap_atomic(page, KM_USER0);
+ dd = kmap_atomic(page);
ino = be64_to_cpu(dd->ino);
- kunmap_atomic(dd, KM_USER0);
+ kunmap_atomic(dd);
page_cache_release(page);
inode = logfs_iget(dir->i_sb, ino);
- if (IS_ERR(inode)) {
+ if (IS_ERR(inode))
printk(KERN_ERR"LogFS: Cannot read inode #%llx for dentry (%lx, %lx)n",
ino, dir->i_ino, index);
- return ERR_CAST(inode);
- }
return d_splice_alias(inode, dentry);
}
@@ -404,12 +383,12 @@ static int logfs_write_dir(struct inode *dir, struct dentry *dentry,
if (!page)
return -ENOMEM;
- dd = kmap_atomic(page, KM_USER0);
+ dd = kmap_atomic(page);
memset(dd, 0, sizeof(*dd));
dd->ino = cpu_to_be64(inode->i_ino);
dd->type = logfs_type(inode);
logfs_set_name(dd, &dentry->d_name);
- kunmap_atomic(dd, KM_USER0);
+ kunmap_atomic(dd);
err = logfs_write_buf(dir, page, WF_LOCK);
unlock_page(page);
@@ -434,8 +413,11 @@ static int __logfs_create(struct inode *dir, struct dentry *dentry,
int ret;
ta = kzalloc(sizeof(*ta), GFP_KERNEL);
- if (!ta)
+ if (!ta) {
+ drop_nlink(inode);
+ iput(inode);
return -ENOMEM;
+ }
ta->state = CREATE_1;
ta->ino = inode->i_ino;
@@ -455,7 +437,7 @@ static int __logfs_create(struct inode *dir, struct dentry *dentry,
abort_transaction(inode, ta);
li->li_flags |= LOGFS_IF_STILLBORN;
/* FIXME: truncate symlink */
- inode->i_nlink--;
+ drop_nlink(inode);
iput(inode);
goto out;
}
@@ -481,7 +463,7 @@ out:
return ret;
}
-static int logfs_mkdir(struct inode *dir, struct dentry *dentry, int mode)
+static int logfs_mkdir(struct inode *dir, struct dentry *dentry, umode_t mode)
{
struct inode *inode;
@@ -500,8 +482,8 @@ static int logfs_mkdir(struct inode *dir, struct dentry *dentry, int mode)
return __logfs_create(dir, dentry, inode, NULL, 0);
}
-static int logfs_create(struct inode *dir, struct dentry *dentry, int mode,
- struct nameidata *nd)
+static int logfs_create(struct inode *dir, struct dentry *dentry, umode_t mode,
+ bool excl)
{
struct inode *inode;
@@ -516,7 +498,7 @@ static int logfs_create(struct inode *dir, struct dentry *dentry, int mode,
return __logfs_create(dir, dentry, inode, NULL, 0);
}
-static int logfs_mknod(struct inode *dir, struct dentry *dentry, int mode,
+static int logfs_mknod(struct inode *dir, struct dentry *dentry, umode_t mode,
dev_t rdev)
{
struct inode *inode;
@@ -552,22 +534,14 @@ static int logfs_symlink(struct inode *dir, struct dentry *dentry,
return __logfs_create(dir, dentry, inode, target, destlen);
}
-static int logfs_permission(struct inode *inode, int mask)
-{
- return generic_permission(inode, mask, NULL);
-}
-
static int logfs_link(struct dentry *old_dentry, struct inode *dir,
struct dentry *dentry)
{
struct inode *inode = old_dentry->d_inode;
- if (inode->i_nlink >= LOGFS_LINK_MAX)
- return -EMLINK;
-
inode->i_ctime = dir->i_ctime = dir->i_mtime = CURRENT_TIME;
- atomic_inc(&inode->i_count);
- inode->i_nlink++;
+ ihold(inode);
+ inc_nlink(inode);
mark_inode_dirty_sync(inode);
return __logfs_create(dir, dentry, inode, NULL, 0);
@@ -583,9 +557,9 @@ static int logfs_get_dd(struct inode *dir, struct dentry *dentry,
if (IS_ERR(page))
return PTR_ERR(page);
*pos = page->index;
- map = kmap_atomic(page, KM_USER0);
+ map = kmap_atomic(page);
memcpy(dd, map, sizeof(*dd));
- kunmap_atomic(map, KM_USER0);
+ kunmap_atomic(map);
page_cache_release(page);
return 0;
}
@@ -815,13 +789,13 @@ const struct inode_operations logfs_dir_iops = {
.mknod = logfs_mknod,
.rename = logfs_rename,
.rmdir = logfs_rmdir,
- .permission = logfs_permission,
.symlink = logfs_symlink,
.unlink = logfs_unlink,
};
const struct file_operations logfs_dir_fops = {
.fsync = logfs_fsync,
- .ioctl = logfs_ioctl,
- .readdir = logfs_readdir,
+ .unlocked_ioctl = logfs_ioctl,
+ .iterate = logfs_readdir,
.read = generic_read_dir,
+ .llseek = default_llseek,
};
diff --git a/fs/logfs/file.c b/fs/logfs/file.c
index 370f367a933..8538752df2f 100644
--- a/fs/logfs/file.c
+++ b/fs/logfs/file.c
@@ -159,9 +159,20 @@ static int logfs_writepage(struct page *page, struct writeback_control *wbc)
return __logfs_writepage(page);
}
-static void logfs_invalidatepage(struct page *page, unsigned long offset)
+static void logfs_invalidatepage(struct page *page, unsigned int offset,
+ unsigned int length)
{
- move_page_to_btree(page);
+ struct logfs_block *block = logfs_block(page);
+
+ if (block->reserved_bytes) {
+ struct super_block *sb = page->mapping->host->i_sb;
+ struct logfs_super *super = logfs_super(sb);
+
+ super->s_dirty_pages -= block->reserved_bytes;
+ block->ops->free_block(sb, block);
+ BUG_ON(bitmap_weight(block->alias_map, LOGFS_BLOCK_FACTOR));
+ } else
+ move_page_to_btree(page);
BUG_ON(PagePrivate(page) || page->private);
}
@@ -171,9 +182,9 @@ static int logfs_releasepage(struct page *page, gfp_t only_xfs_uses_this)
}
-int logfs_ioctl(struct inode *inode, struct file *file, unsigned int cmd,
- unsigned long arg)
+long logfs_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
{
+ struct inode *inode = file_inode(file);
struct logfs_inode *li = logfs_inode(inode);
unsigned int oldflags, flags;
int err;
@@ -186,7 +197,7 @@ int logfs_ioctl(struct inode *inode, struct file *file, unsigned int cmd,
if (IS_RDONLY(inode))
return -EROFS;
- if (!is_owner_or_cap(inode))
+ if (!inode_owner_or_capable(inode))
return -EACCES;
err = get_user(flags, (int __user *)arg);
@@ -209,13 +220,22 @@ int logfs_ioctl(struct inode *inode, struct file *file, unsigned int cmd,
}
}
-int logfs_fsync(struct file *file, struct dentry *dentry, int datasync)
+int logfs_fsync(struct file *file, loff_t start, loff_t end, int datasync)
{
- struct super_block *sb = dentry->d_inode->i_sb;
- struct logfs_super *super = logfs_super(sb);
+ struct super_block *sb = file->f_mapping->host->i_sb;
+ struct inode *inode = file->f_mapping->host;
+ int ret;
+
+ ret = filemap_write_and_wait_range(inode->i_mapping, start, end);
+ if (ret)
+ return ret;
+
+ mutex_lock(&inode->i_mutex);
+ logfs_get_wblocks(sb, NULL, WF_LOCK);
+ logfs_write_anchor(sb);
+ logfs_put_wblocks(sb, NULL, WF_LOCK);
+ mutex_unlock(&inode->i_mutex);
- /* FIXME: write anchor */
- super->s_devops->sync(sb);
return 0;
}
@@ -224,15 +244,19 @@ static int logfs_setattr(struct dentry *dentry, struct iattr *attr)
struct inode *inode = dentry->d_inode;
int err = 0;
- if (attr->ia_valid & ATTR_SIZE)
+ err = inode_change_ok(inode, attr);
+ if (err)
+ return err;
+
+ if (attr->ia_valid & ATTR_SIZE) {
err = logfs_truncate(inode, attr->ia_size);
- attr->ia_valid &= ~ATTR_SIZE;
+ if (err)
+ return err;
+ }
- if (!err)
- err = inode_change_ok(inode, attr);
- if (!err)
- err = inode_setattr(inode, attr);
- return err;
+ setattr_copy(inode, attr);
+ mark_inode_dirty(inode);
+ return 0;
}
const struct inode_operations logfs_reg_iops = {
@@ -240,15 +264,15 @@ const struct inode_operations logfs_reg_iops = {
};
const struct file_operations logfs_reg_fops = {
- .aio_read = generic_file_aio_read,
- .aio_write = generic_file_aio_write,
+ .read_iter = generic_file_read_iter,
+ .write_iter = generic_file_write_iter,
.fsync = logfs_fsync,
- .ioctl = logfs_ioctl,
+ .unlocked_ioctl = logfs_ioctl,
.llseek = generic_file_llseek,
.mmap = generic_file_readonly_mmap,
.open = generic_file_open,
- .read = do_sync_read,
- .write = do_sync_write,
+ .read = new_sync_read,
+ .write = new_sync_write,
};
const struct address_space_operations logfs_reg_aops = {
diff --git a/fs/logfs/gc.c b/fs/logfs/gc.c
index 92949f95a90..d4efb061bdc 100644
--- a/fs/logfs/gc.c
+++ b/fs/logfs/gc.c
@@ -7,6 +7,7 @@
*/
#include "logfs.h"
#include <linux/sched.h>
+#include <linux/slab.h>
/*
* Wear leveling needs to kick in when the difference between low erase
@@ -121,7 +122,7 @@ static void logfs_cleanse_block(struct super_block *sb, u64 ofs, u64 ino,
logfs_safe_iput(inode, cookie);
}
-static u32 logfs_gc_segment(struct super_block *sb, u32 segno, u8 dist)
+static u32 logfs_gc_segment(struct super_block *sb, u32 segno)
{
struct logfs_super *super = logfs_super(sb);
struct logfs_segment_header sh;
@@ -366,7 +367,7 @@ static struct gc_candidate *get_candidate(struct super_block *sb)
int i, max_dist;
struct gc_candidate *cand = NULL, *this;
- max_dist = min(no_free_segments(sb), LOGFS_NO_AREAS);
+ max_dist = min(no_free_segments(sb), LOGFS_NO_AREAS - 1);
for (i = max_dist; i >= 0; i--) {
this = first_in_list(&super->s_low_list[i]);
@@ -400,7 +401,7 @@ static int __logfs_gc_once(struct super_block *sb, struct gc_candidate *cand)
segno, (u64)segno << super->s_segshift,
dist, no_free_segments(sb), valid,
super->s_free_bytes);
- cleaned = logfs_gc_segment(sb, segno, dist);
+ cleaned = logfs_gc_segment(sb, segno);
log_gc("GC segment #%02x complete - now %x valid\n", segno,
valid - cleaned);
BUG_ON(cleaned != valid);
@@ -458,6 +459,14 @@ static void __logfs_gc_pass(struct super_block *sb, int target)
struct logfs_block *block;
int round, progress, last_progress = 0;
+ /*
+ * Doing too many changes to the segfile at once would result
+ * in a large number of aliases. Write the journal before
+ * things get out of hand.
+ */
+ if (super->s_shadow_tree.no_shadowed_segments >= MAX_OBJ_ALIASES)
+ logfs_write_anchor(sb);
+
if (no_free_segments(sb) >= target &&
super->s_no_object_aliases < MAX_OBJ_ALIASES)
return;
@@ -623,38 +632,31 @@ static int check_area(struct super_block *sb, int i)
{
struct logfs_super *super = logfs_super(sb);
struct logfs_area *area = super->s_area[i];
- struct logfs_object_header oh;
+ gc_level_t gc_level;
+ u32 cleaned, valid, ec;
u32 segno = area->a_segno;
- u32 ofs = area->a_used_bytes;
- __be32 crc;
- int err;
+ u64 ofs = dev_ofs(sb, area->a_segno, area->a_written_bytes);
if (!area->a_is_open)
return 0;
- for (ofs = area->a_used_bytes;
- ofs <= super->s_segsize - sizeof(oh);
- ofs += (u32)be16_to_cpu(oh.len) + sizeof(oh)) {
- err = wbuf_read(sb, dev_ofs(sb, segno, ofs), sizeof(oh), &oh);
- if (err)
- return err;
-
- if (!memchr_inv(&oh, 0xff, sizeof(oh)))
- break;
+ if (super->s_devops->can_write_buf(sb, ofs) == 0)
+ return 0;
- crc = logfs_crc32(&oh, sizeof(oh) - 4, 4);
- if (crc != oh.crc) {
- printk(KERN_INFO "interrupted header at %llx\n",
- dev_ofs(sb, segno, ofs));
- return 0;
- }
- }
- if (ofs != area->a_used_bytes) {
- printk(KERN_INFO "%x bytes unaccounted data found at %llx\n",
- ofs - area->a_used_bytes,
- dev_ofs(sb, segno, area->a_used_bytes));
- area->a_used_bytes = ofs;
- }
+ printk(KERN_INFO"LogFS: Possibly incomplete write at %llx\n", ofs);
+ /*
+ * The device cannot write back the write buffer. Most likely the
+ * wbuf was already written out and the system crashed at some point
+ * before the journal commit happened. In that case we wouldn't have
+ * to do anything. But if the crash happened before the wbuf was
+ * written out correctly, we must GC this segment. So assume the
+ * worst and always do the GC run.
+ */
+ area->a_is_open = 0;
+ valid = logfs_valid_bytes(sb, segno, &ec, &gc_level);
+ cleaned = logfs_gc_segment(sb, segno);
+ if (cleaned != valid)
+ return -EIO;
return 0;
}
diff --git a/fs/logfs/inode.c b/fs/logfs/inode.c
index 33ec1aeaeec..af49e2d6941 100644
--- a/fs/logfs/inode.c
+++ b/fs/logfs/inode.c
@@ -6,6 +6,7 @@
* Copyright (c) 2005-2008 Joern Engel <joern@logfs.org>
*/
#include "logfs.h"
+#include <linux/slab.h>
#include <linux/writeback.h>
#include <linux/backing-dev.h>
@@ -32,7 +33,7 @@
* are being written out - and waiting for GC to make progress, naturally.
*
* So we cannot just call iget() or some variant of it, but first have to check
- * wether the inode in question might be in I_FREEING state. Therefore we
+ * whether the inode in question might be in I_FREEING state. Therefore we
* maintain our own per-sb list of "almost deleted" inodes and check against
* that list first. Normally this should be at most 1-2 entries long.
*
@@ -92,7 +93,7 @@ static struct inode *__logfs_iget(struct super_block *sb, ino_t ino)
/* inode->i_nlink == 0 can be true when called from
* block validator */
/* set i_nlink to 0 to prevent caching */
- inode->i_nlink = 0;
+ clear_nlink(inode);
logfs_inode(inode)->li_flags |= LOGFS_IF_ZOMBIE;
iget_failed(inode);
if (!err)
@@ -140,19 +141,41 @@ struct inode *logfs_safe_iget(struct super_block *sb, ino_t ino, int *is_cached)
return __logfs_iget(sb, ino);
}
+static void logfs_i_callback(struct rcu_head *head)
+{
+ struct inode *inode = container_of(head, struct inode, i_rcu);
+ kmem_cache_free(logfs_inode_cache, logfs_inode(inode));
+}
+
static void __logfs_destroy_inode(struct inode *inode)
{
struct logfs_inode *li = logfs_inode(inode);
BUG_ON(li->li_block);
list_del(&li->li_freeing_list);
- kmem_cache_free(logfs_inode_cache, li);
+ call_rcu(&inode->i_rcu, logfs_i_callback);
+}
+
+static void __logfs_destroy_meta_inode(struct inode *inode)
+{
+ struct logfs_inode *li = logfs_inode(inode);
+ BUG_ON(li->li_block);
+ call_rcu(&inode->i_rcu, logfs_i_callback);
}
static void logfs_destroy_inode(struct inode *inode)
{
struct logfs_inode *li = logfs_inode(inode);
+ if (inode->i_ino < LOGFS_RESERVED_INOS) {
+ /*
+ * The reserved inodes are never destroyed unless we are in
+ * unmont path.
+ */
+ __logfs_destroy_meta_inode(inode);
+ return;
+ }
+
BUG_ON(list_empty(&li->li_freeing_list));
spin_lock(&logfs_inode_lock);
li->li_refcount--;
@@ -185,13 +208,13 @@ static void logfs_init_inode(struct super_block *sb, struct inode *inode)
li->li_height = 0;
li->li_used_bytes = 0;
li->li_block = NULL;
- inode->i_uid = 0;
- inode->i_gid = 0;
+ i_uid_write(inode, 0);
+ i_gid_write(inode, 0);
inode->i_size = 0;
inode->i_blocks = 0;
inode->i_ctime = CURRENT_TIME;
inode->i_mtime = CURRENT_TIME;
- inode->i_nlink = 1;
+ li->li_refcount = 1;
INIT_LIST_HEAD(&li->li_freeing_list);
for (i = 0; i < LOGFS_EMBEDDED_FIELDS; i++)
@@ -233,33 +256,21 @@ static struct inode *logfs_alloc_inode(struct super_block *sb)
* purpose is to create a new inode that will not trigger the warning if such
* an inode is still in use. An ugly hack, no doubt. Suggections for
* improvement are welcome.
+ *
+ * AV: that's what ->put_super() is for...
*/
struct inode *logfs_new_meta_inode(struct super_block *sb, u64 ino)
{
struct inode *inode;
- inode = logfs_alloc_inode(sb);
+ inode = new_inode(sb);
if (!inode)
return ERR_PTR(-ENOMEM);
inode->i_mode = S_IFREG;
inode->i_ino = ino;
- inode->i_sb = sb;
-
- /* This is a blatant copy of alloc_inode code. We'd need alloc_inode
- * to be nonstatic, alas. */
- {
- struct address_space * const mapping = &inode->i_data;
-
- mapping->a_ops = &logfs_reg_aops;
- mapping->host = inode;
- mapping->flags = 0;
- mapping_set_gfp_mask(mapping, GFP_NOFS);
- mapping->assoc_mapping = NULL;
- mapping->backing_dev_info = &default_backing_dev_info;
- inode->i_mapping = mapping;
- inode->i_nlink = 1;
- }
+ inode->i_data.a_ops = &logfs_reg_aops;
+ mapping_set_gfp_mask(&inode->i_data, GFP_NOFS);
return inode;
}
@@ -275,7 +286,7 @@ struct inode *logfs_read_meta_inode(struct super_block *sb, u64 ino)
err = logfs_read_inode(inode);
if (err) {
- destroy_meta_inode(inode);
+ iput(inode);
return ERR_PTR(err);
}
logfs_inode_setops(inode);
@@ -291,23 +302,13 @@ static int logfs_write_inode(struct inode *inode, struct writeback_control *wbc)
if (logfs_inode(inode)->li_flags & LOGFS_IF_STILLBORN)
return 0;
- ret = __logfs_write_inode(inode, flags);
+ ret = __logfs_write_inode(inode, NULL, flags);
LOGFS_BUG_ON(ret, inode->i_sb);
return ret;
}
-void destroy_meta_inode(struct inode *inode)
-{
- if (inode) {
- if (inode->i_data.nrpages)
- truncate_inode_pages(&inode->i_data, 0);
- logfs_clear_inode(inode);
- kmem_cache_free(logfs_inode_cache, logfs_inode(inode));
- }
-}
-
-/* called with inode_lock held */
-static void logfs_drop_inode(struct inode *inode)
+/* called with inode->i_lock held */
+static int logfs_drop_inode(struct inode *inode)
{
struct logfs_super *super = logfs_super(inode->i_sb);
struct logfs_inode *li = logfs_inode(inode);
@@ -315,7 +316,7 @@ static void logfs_drop_inode(struct inode *inode)
spin_lock(&logfs_inode_lock);
list_move(&li->li_freeing_list, &super->s_freeing_list);
spin_unlock(&logfs_inode_lock);
- generic_drop_inode(inode);
+ return generic_drop_inode(inode);
}
static void logfs_set_ino_generation(struct super_block *sb,
@@ -325,7 +326,7 @@ static void logfs_set_ino_generation(struct super_block *sb,
u64 ino;
mutex_lock(&super->s_journal_mutex);
- ino = logfs_seek_hole(super->s_master_inode, super->s_last_ino);
+ ino = logfs_seek_hole(super->s_master_inode, super->s_last_ino + 1);
super->s_last_ino = ino;
super->s_inos_till_wrap--;
if (super->s_inos_till_wrap < 0) {
@@ -338,7 +339,7 @@ static void logfs_set_ino_generation(struct super_block *sb,
mutex_unlock(&super->s_journal_mutex);
}
-struct inode *logfs_new_inode(struct inode *dir, int mode)
+struct inode *logfs_new_inode(struct inode *dir, umode_t mode)
{
struct super_block *sb = dir->i_sb;
struct inode *inode;
@@ -356,14 +357,7 @@ struct inode *logfs_new_inode(struct inode *dir, int mode)
inode->i_mode = mode;
logfs_set_ino_generation(sb, inode);
- inode->i_uid = current_fsuid();
- inode->i_gid = current_fsgid();
- if (dir->i_mode & S_ISGID) {
- inode->i_gid = dir->i_gid;
- if (S_ISDIR(mode))
- inode->i_mode |= S_ISGID;
- }
-
+ inode_init_owner(inode, dir, mode);
logfs_inode_setops(inode);
insert_inode_hash(inode);
@@ -385,17 +379,27 @@ static void logfs_init_once(void *_li)
static int logfs_sync_fs(struct super_block *sb, int wait)
{
- /* FIXME: write anchor */
- logfs_super(sb)->s_devops->sync(sb);
+ logfs_get_wblocks(sb, NULL, WF_LOCK);
+ logfs_write_anchor(sb);
+ logfs_put_wblocks(sb, NULL, WF_LOCK);
return 0;
}
+static void logfs_put_super(struct super_block *sb)
+{
+ struct logfs_super *super = logfs_super(sb);
+ /* kill the meta-inodes */
+ iput(super->s_segfile_inode);
+ iput(super->s_master_inode);
+ iput(super->s_mapping_inode);
+}
+
const struct super_operations logfs_super_operations = {
.alloc_inode = logfs_alloc_inode,
- .clear_inode = logfs_clear_inode,
- .delete_inode = logfs_delete_inode,
.destroy_inode = logfs_destroy_inode,
+ .evict_inode = logfs_evict_inode,
.drop_inode = logfs_drop_inode,
+ .put_super = logfs_put_super,
.write_inode = logfs_write_inode,
.statfs = logfs_statfs,
.sync_fs = logfs_sync_fs,
@@ -413,5 +417,10 @@ int logfs_init_inode_cache(void)
void logfs_destroy_inode_cache(void)
{
+ /*
+ * Make sure all delayed rcu free inodes are flushed before we
+ * destroy cache.
+ */
+ rcu_barrier();
kmem_cache_destroy(logfs_inode_cache);
}
diff --git a/fs/logfs/journal.c b/fs/logfs/journal.c
index 6ad30a4c905..2a09b8d7398 100644
--- a/fs/logfs/journal.c
+++ b/fs/logfs/journal.c
@@ -6,6 +6,7 @@
* Copyright (c) 2005-2008 Joern Engel <joern@logfs.org>
*/
#include "logfs.h"
+#include <linux/slab.h>
static void logfs_calc_free(struct super_block *sb)
{
@@ -131,10 +132,9 @@ static int read_area(struct super_block *sb, struct logfs_je_area *a)
ofs = dev_ofs(sb, area->a_segno, area->a_written_bytes);
if (super->s_writesize > 1)
- logfs_buf_recover(area, ofs, a + 1, super->s_writesize);
+ return logfs_buf_recover(area, ofs, a + 1, super->s_writesize);
else
- logfs_buf_recover(area, ofs, NULL, 0);
- return 0;
+ return logfs_buf_recover(area, ofs, NULL, 0);
}
static void *unpack(void *from, void *to)
@@ -244,7 +244,7 @@ static int read_je(struct super_block *sb, u64 ofs)
read_erasecount(sb, unpack(jh, scratch));
break;
case JE_AREA:
- read_area(sb, unpack(jh, scratch));
+ err = read_area(sb, unpack(jh, scratch));
break;
case JE_OBJ_ALIAS:
err = logfs_load_object_aliases(sb, unpack(jh, scratch),
@@ -388,7 +388,10 @@ static void journal_get_erase_count(struct logfs_area *area)
static int journal_erase_segment(struct logfs_area *area)
{
struct super_block *sb = area->a_sb;
- struct logfs_segment_header sh;
+ union {
+ struct logfs_segment_header sh;
+ unsigned char c[ALIGN(sizeof(struct logfs_segment_header), 16)];
+ } u;
u64 ofs;
int err;
@@ -396,20 +399,21 @@ static int journal_erase_segment(struct logfs_area *area)
if (err)
return err;
- sh.pad = 0;
- sh.type = SEG_JOURNAL;
- sh.level = 0;
- sh.segno = cpu_to_be32(area->a_segno);
- sh.ec = cpu_to_be32(area->a_erase_count);
- sh.gec = cpu_to_be64(logfs_super(sb)->s_gec);
- sh.crc = logfs_crc32(&sh, sizeof(sh), 4);
+ memset(&u, 0, sizeof(u));
+ u.sh.pad = 0;
+ u.sh.type = SEG_JOURNAL;
+ u.sh.level = 0;
+ u.sh.segno = cpu_to_be32(area->a_segno);
+ u.sh.ec = cpu_to_be32(area->a_erase_count);
+ u.sh.gec = cpu_to_be64(logfs_super(sb)->s_gec);
+ u.sh.crc = logfs_crc32(&u.sh, sizeof(u.sh), 4);
/* This causes a bug in segment.c. Not yet. */
//logfs_set_segment_erased(sb, area->a_segno, area->a_erase_count, 0);
ofs = dev_ofs(sb, area->a_segno, 0);
- area->a_used_bytes = ALIGN(sizeof(sh), 16);
- logfs_buf_write(area, ofs, &sh, sizeof(sh));
+ area->a_used_bytes = sizeof(u);
+ logfs_buf_write(area, ofs, &u, sizeof(u));
return 0;
}
@@ -493,6 +497,8 @@ static void account_shadows(struct super_block *sb)
btree_grim_visitor64(&tree->new, (unsigned long)sb, account_shadow);
btree_grim_visitor64(&tree->old, (unsigned long)sb, account_shadow);
+ btree_grim_visitor32(&tree->segment_map, 0, NULL);
+ tree->no_shadowed_segments = 0;
if (li->li_block) {
/*
@@ -559,7 +565,7 @@ static void write_wbuf(struct super_block *sb, struct logfs_area *area,
index = ofs >> PAGE_SHIFT;
page_ofs = ofs & (PAGE_SIZE - 1);
- page = find_lock_page(mapping, index);
+ page = find_or_create_page(mapping, index, GFP_NOFS);
BUG_ON(!page);
memcpy(wbuf, page_address(page) + page_ofs, super->s_writesize);
unlock_page(page);
@@ -608,7 +614,6 @@ static size_t __logfs_write_je(struct super_block *sb, void *buf, u16 type,
compr_len = logfs_compress(buf, data, len, sb->s_blocksize);
if (compr_len < 0 || type == JE_ANCHOR) {
- BUG_ON(len > sb->s_blocksize);
memcpy(data, buf, len);
compr_len = len;
compr = COMPR_NONE;
@@ -660,6 +665,7 @@ static int logfs_write_je_buf(struct super_block *sb, void *buf, u16 type,
if (ofs < 0)
return ofs;
logfs_buf_write(area, ofs, super->s_compressed_je, len);
+ BUG_ON(super->s_no_je >= MAX_JOURNAL_ENTRIES);
super->s_je_array[super->s_no_je++] = cpu_to_be64(ofs);
return 0;
}
@@ -800,6 +806,7 @@ void do_logfs_journal_wl_pass(struct super_block *sb)
{
struct logfs_super *super = logfs_super(sb);
struct logfs_area *area = super->s_journal_area;
+ struct btree_head32 *head = &super->s_reserved_segments;
u32 segno, ec;
int i, err;
@@ -807,6 +814,7 @@ void do_logfs_journal_wl_pass(struct super_block *sb)
/* Drop old segments */
journal_for_each(i)
if (super->s_journal_seg[i]) {
+ btree_remove32(head, super->s_journal_seg[i]);
logfs_set_segment_unreserved(sb,
super->s_journal_seg[i],
super->s_journal_ec[i]);
@@ -819,8 +827,13 @@ void do_logfs_journal_wl_pass(struct super_block *sb)
super->s_journal_seg[i] = segno;
super->s_journal_ec[i] = ec;
logfs_set_segment_reserved(sb, segno);
+ err = btree_insert32(head, segno, (void *)1, GFP_NOFS);
+ BUG_ON(err); /* mempool should prevent this */
+ err = logfs_erase_segment(sb, segno, 1);
+ BUG_ON(err); /* FIXME: remount-ro would be nicer */
}
/* Manually move journal_area */
+ freeseg(sb, area->a_segno);
area->a_segno = super->s_journal_seg[0];
area->a_is_open = 0;
area->a_used_bytes = 0;
@@ -875,8 +888,6 @@ void logfs_cleanup_journal(struct super_block *sb)
struct logfs_super *super = logfs_super(sb);
btree_grim_visitor32(&super->s_reserved_segments, 0, NULL);
- destroy_meta_inode(super->s_master_inode);
- super->s_master_inode = NULL;
kfree(super->s_compressed_je);
kfree(super->s_je);
diff --git a/fs/logfs/logfs.h b/fs/logfs/logfs.h
index 12977943137..5f093760946 100644
--- a/fs/logfs/logfs.h
+++ b/fs/logfs/logfs.h
@@ -136,6 +136,7 @@ struct logfs_area_ops {
int (*erase_segment)(struct logfs_area *area);
};
+struct logfs_super; /* forward */
/**
* struct logfs_device_ops - device access operations
*
@@ -144,6 +145,7 @@ struct logfs_area_ops {
* @erase: erase one segment
* @read: read from the device
* @erase: erase part of the device
+ * @can_write_buf: decide whether wbuf can be written to ofs
*/
struct logfs_device_ops {
struct page *(*find_first_sb)(struct super_block *sb, u64 *ofs);
@@ -153,8 +155,9 @@ struct logfs_device_ops {
void (*writeseg)(struct super_block *sb, u64 ofs, size_t len);
int (*erase)(struct super_block *sb, loff_t ofs, size_t len,
int ensure_write);
+ int (*can_write_buf)(struct super_block *sb, u64 ofs);
void (*sync)(struct super_block *sb);
- void (*put_device)(struct super_block *sb);
+ void (*put_device)(struct logfs_super *s);
};
/**
@@ -257,10 +260,14 @@ struct logfs_shadow {
* struct shadow_tree
* @new: shadows where old_ofs==0, indexed by new_ofs
* @old: shadows where old_ofs!=0, indexed by old_ofs
+ * @segment_map: bitfield of segments containing shadows
+ * @no_shadowed_segment: number of segments containing shadows
*/
struct shadow_tree {
struct btree_head64 new;
struct btree_head64 old;
+ struct btree_head32 segment_map;
+ int no_shadowed_segments;
};
struct object_alias_item {
@@ -305,13 +312,14 @@ typedef int write_alias_t(struct super_block *sb, u64 ino, u64 bix,
level_t level, int child_no, __be64 val);
struct logfs_block_ops {
void (*write_block)(struct logfs_block *block);
- gc_level_t (*block_level)(struct logfs_block *block);
void (*free_block)(struct super_block *sb, struct logfs_block*block);
int (*write_alias)(struct super_block *sb,
struct logfs_block *block,
write_alias_t *write_one_alias);
};
+#define MAX_JOURNAL_ENTRIES 256
+
struct logfs_super {
struct mtd_info *s_mtd; /* underlying device */
struct block_device *s_bdev; /* underlying device */
@@ -378,7 +386,7 @@ struct logfs_super {
u32 s_journal_ec[LOGFS_JOURNAL_SEGS]; /* journal erasecounts */
u64 s_last_version;
struct logfs_area *s_journal_area; /* open journal segment */
- __be64 s_je_array[64];
+ __be64 s_je_array[MAX_JOURNAL_ENTRIES];
int s_no_je;
int s_sum_index; /* for the 12 summaries */
@@ -389,6 +397,7 @@ struct logfs_super {
int s_lock_count;
mempool_t *s_block_pool; /* struct logfs_block pool */
mempool_t *s_shadow_pool; /* struct logfs_shadow pool */
+ struct list_head s_writeback_list; /* writeback pages */
/*
* Space accounting:
* - s_used_bytes specifies space used to store valid data objects.
@@ -463,11 +472,13 @@ void logfs_compr_exit(void);
/* dev_bdev.c */
#ifdef CONFIG_BLOCK
-int logfs_get_sb_bdev(struct file_system_type *type, int flags,
- const char *devname, struct vfsmount *mnt);
+int logfs_get_sb_bdev(struct logfs_super *s,
+ struct file_system_type *type,
+ const char *devname);
#else
-static inline int logfs_get_sb_bdev(struct file_system_type *type, int flags,
- const char *devname, struct vfsmount *mnt)
+static inline int logfs_get_sb_bdev(struct logfs_super *s,
+ struct file_system_type *type,
+ const char *devname)
{
return -ENODEV;
}
@@ -475,11 +486,9 @@ static inline int logfs_get_sb_bdev(struct file_system_type *type, int flags,
/* dev_mtd.c */
#ifdef CONFIG_MTD
-int logfs_get_sb_mtd(struct file_system_type *type, int flags,
- int mtdnr, struct vfsmount *mnt);
+int logfs_get_sb_mtd(struct logfs_super *s, int mtdnr);
#else
-static inline int logfs_get_sb_mtd(struct file_system_type *type, int flags,
- int mtdnr, struct vfsmount *mnt)
+static inline int logfs_get_sb_mtd(struct logfs_super *s, int mtdnr)
{
return -ENODEV;
}
@@ -496,9 +505,8 @@ extern const struct inode_operations logfs_reg_iops;
extern const struct file_operations logfs_reg_fops;
extern const struct address_space_operations logfs_reg_aops;
int logfs_readpage(struct file *file, struct page *page);
-int logfs_ioctl(struct inode *inode, struct file *file, unsigned int cmd,
- unsigned long arg);
-int logfs_fsync(struct file *file, struct dentry *dentry, int datasync);
+long logfs_ioctl(struct file *file, unsigned int cmd, unsigned long arg);
+int logfs_fsync(struct file *file, loff_t start, loff_t end, int datasync);
/* gc.c */
u32 get_best_cand(struct super_block *sb, struct candidate_list *list, u32 *ec);
@@ -512,18 +520,16 @@ extern const struct super_operations logfs_super_operations;
struct inode *logfs_iget(struct super_block *sb, ino_t ino);
struct inode *logfs_safe_iget(struct super_block *sb, ino_t ino, int *cookie);
void logfs_safe_iput(struct inode *inode, int cookie);
-struct inode *logfs_new_inode(struct inode *dir, int mode);
+struct inode *logfs_new_inode(struct inode *dir, umode_t mode);
struct inode *logfs_new_meta_inode(struct super_block *sb, u64 ino);
struct inode *logfs_read_meta_inode(struct super_block *sb, u64 ino);
int logfs_init_inode_cache(void);
void logfs_destroy_inode_cache(void);
-void destroy_meta_inode(struct inode *inode);
void logfs_set_blocks(struct inode *inode, u64 no);
/* these logically belong into inode.c but actually reside in readwrite.c */
int logfs_read_inode(struct inode *inode);
-int __logfs_write_inode(struct inode *inode, long flags);
-void logfs_delete_inode(struct inode *inode);
-void logfs_clear_inode(struct inode *inode);
+int __logfs_write_inode(struct inode *inode, struct page *, long flags);
+void logfs_evict_inode(struct inode *inode);
/* journal.c */
void logfs_write_anchor(struct super_block *sb);
@@ -571,6 +577,8 @@ void initialize_block_counters(struct page *page, struct logfs_block *block,
__be64 *array, int page_is_empty);
int logfs_exist_block(struct inode *inode, u64 bix);
int get_page_reserve(struct inode *inode, struct page *page);
+void logfs_get_wblocks(struct super_block *sb, struct page *page, int lock);
+void logfs_put_wblocks(struct super_block *sb, struct page *page, int lock);
extern struct logfs_block_ops indirect_block_ops;
/* segment.c */
@@ -587,35 +595,33 @@ void move_page_to_btree(struct page *page);
int logfs_init_mapping(struct super_block *sb);
void logfs_sync_area(struct logfs_area *area);
void logfs_sync_segments(struct super_block *sb);
+void freeseg(struct super_block *sb, u32 segno);
+void free_areas(struct super_block *sb);
/* area handling */
int logfs_init_areas(struct super_block *sb);
void logfs_cleanup_areas(struct super_block *sb);
int logfs_open_area(struct logfs_area *area, size_t bytes);
-void __logfs_buf_write(struct logfs_area *area, u64 ofs, void *buf, size_t len,
+int __logfs_buf_write(struct logfs_area *area, u64 ofs, void *buf, size_t len,
int use_filler);
-static inline void logfs_buf_write(struct logfs_area *area, u64 ofs,
+static inline int logfs_buf_write(struct logfs_area *area, u64 ofs,
void *buf, size_t len)
{
- __logfs_buf_write(area, ofs, buf, len, 0);
+ return __logfs_buf_write(area, ofs, buf, len, 0);
}
-static inline void logfs_buf_recover(struct logfs_area *area, u64 ofs,
+static inline int logfs_buf_recover(struct logfs_area *area, u64 ofs,
void *buf, size_t len)
{
- __logfs_buf_write(area, ofs, buf, len, 1);
+ return __logfs_buf_write(area, ofs, buf, len, 1);
}
/* super.c */
struct page *emergency_read_begin(struct address_space *mapping, pgoff_t index);
void emergency_read_end(struct page *page);
void logfs_crash_dump(struct super_block *sb);
-void *memchr_inv(const void *s, int c, size_t n);
int logfs_statfs(struct dentry *dentry, struct kstatfs *stats);
-int logfs_get_sb_device(struct file_system_type *type, int flags,
- struct mtd_info *mtd, struct block_device *bdev,
- const struct logfs_device_ops *devops, struct vfsmount *mnt);
int logfs_check_ds(struct logfs_disk_super *ds);
int logfs_write_sb(struct super_block *sb);
@@ -698,7 +704,7 @@ static inline gc_level_t expand_level(u64 ino, level_t __level)
u8 level = (__force u8)__level;
if (ino == LOGFS_INO_MASTER) {
- /* ifile has seperate areas */
+ /* ifile has separate areas */
level += LOGFS_MAX_LEVELS;
}
return (__force gc_level_t)level;
@@ -721,4 +727,10 @@ static inline struct logfs_area *get_area(struct super_block *sb,
return logfs_super(sb)->s_area[(__force u8)gc_level];
}
+static inline void logfs_mempool_destroy(mempool_t *pool)
+{
+ if (pool)
+ mempool_destroy(pool);
+}
+
#endif
diff --git a/fs/logfs/logfs_abi.h b/fs/logfs/logfs_abi.h
index f674725663f..ae960519c54 100644
--- a/fs/logfs/logfs_abi.h
+++ b/fs/logfs/logfs_abi.h
@@ -50,9 +50,9 @@ static inline void check_##type(void) \
* 12 - gc recycled blocks, long-lived data
* 13 - replacement blocks, short-lived data
*
- * Levels 1-11 are necessary for robust gc operations and help seperate
+ * Levels 1-11 are necessary for robust gc operations and help separate
* short-lived metadata from longer-lived file data. In the future,
- * file data should get seperated into several segments based on simple
+ * file data should get separated into several segments based on simple
* heuristics. Old data recycled during gc operation is expected to be
* long-lived. New data is of uncertain life expectancy. New data
* used to replace older blocks in existing files is expected to be
@@ -117,7 +117,7 @@ static inline void check_##type(void) \
#define pure_ofs(ofs) (ofs & ~LOGFS_FULLY_POPULATED)
/*
- * LogFS needs to seperate data into levels. Each level is defined as the
+ * LogFS needs to separate data into levels. Each level is defined as the
* maximal possible distance from the master inode (inode of the inode file).
* Data blocks reside on level 0, 1x indirect block on level 1, etc.
* Inodes reside on level 6, indirect blocks for the inode file on levels 7-11.
@@ -204,7 +204,7 @@ SIZE_CHECK(logfs_segment_header, LOGFS_SEGMENT_HEADERSIZE);
* @ds_crc: crc32 of structure starting with the next field
* @ds_ifile_levels: maximum number of levels for ifile
* @ds_iblock_levels: maximum number of levels for regular files
- * @ds_data_levels: number of seperate levels for data
+ * @ds_data_levels: number of separate levels for data
* @pad0: reserved, must be 0
* @ds_feature_incompat: incompatible filesystem features
* @ds_feature_ro_compat: read-only compatible filesystem features
@@ -456,7 +456,7 @@ enum logfs_vim {
* @vim: life expectancy of data
*
* "Areas" are segments currently being used for writing. There is at least
- * one area per GC level. Several may be used to seperate long-living from
+ * one area per GC level. Several may be used to separate long-living from
* short-living data. If an area with unknown vim is encountered, it can
* simply be closed.
* The write buffer immediately follow this header.
diff --git a/fs/logfs/readwrite.c b/fs/logfs/readwrite.c
index 7a23b3e7c0a..48140315f62 100644
--- a/fs/logfs/readwrite.c
+++ b/fs/logfs/readwrite.c
@@ -18,6 +18,7 @@
*/
#include "logfs.h"
#include <linux/sched.h>
+#include <linux/slab.h>
static u64 adjust_bix(u64 bix, level_t level)
{
@@ -118,14 +119,14 @@ static void logfs_disk_to_inode(struct logfs_disk_inode *di, struct inode*inode)
inode->i_mode = be16_to_cpu(di->di_mode);
li->li_height = di->di_height;
li->li_flags = be32_to_cpu(di->di_flags);
- inode->i_uid = be32_to_cpu(di->di_uid);
- inode->i_gid = be32_to_cpu(di->di_gid);
+ i_uid_write(inode, be32_to_cpu(di->di_uid));
+ i_gid_write(inode, be32_to_cpu(di->di_gid));
inode->i_size = be64_to_cpu(di->di_size);
logfs_set_blocks(inode, be64_to_cpu(di->di_used_bytes));
inode->i_atime = be64_to_timespec(di->di_atime);
inode->i_ctime = be64_to_timespec(di->di_ctime);
inode->i_mtime = be64_to_timespec(di->di_mtime);
- inode->i_nlink = be32_to_cpu(di->di_refcount);
+ set_nlink(inode, be32_to_cpu(di->di_refcount));
inode->i_generation = be32_to_cpu(di->di_generation);
switch (inode->i_mode & S_IFMT) {
@@ -155,8 +156,8 @@ static void logfs_inode_to_disk(struct inode *inode, struct logfs_disk_inode*di)
di->di_height = li->li_height;
di->di_pad = 0;
di->di_flags = cpu_to_be32(li->li_flags);
- di->di_uid = cpu_to_be32(inode->i_uid);
- di->di_gid = cpu_to_be32(inode->i_gid);
+ di->di_uid = cpu_to_be32(i_uid_read(inode));
+ di->di_gid = cpu_to_be32(i_gid_read(inode));
di->di_size = cpu_to_be64(i_size_read(inode));
di->di_used_bytes = cpu_to_be64(li->li_used_bytes);
di->di_atime = timespec_to_be64(inode->i_atime);
@@ -243,8 +244,7 @@ static void preunlock_page(struct super_block *sb, struct page *page, int lock)
* is waiting for s_write_mutex. We annotate this fact by setting PG_pre_locked
* in addition to PG_locked.
*/
-static void logfs_get_wblocks(struct super_block *sb, struct page *page,
- int lock)
+void logfs_get_wblocks(struct super_block *sb, struct page *page, int lock)
{
struct logfs_super *super = logfs_super(sb);
@@ -259,8 +259,7 @@ static void logfs_get_wblocks(struct super_block *sb, struct page *page,
}
}
-static void logfs_put_wblocks(struct super_block *sb, struct page *page,
- int lock)
+void logfs_put_wblocks(struct super_block *sb, struct page *page, int lock)
{
struct logfs_super *super = logfs_super(sb);
@@ -423,31 +422,12 @@ static void inode_write_block(struct logfs_block *block)
if (inode->i_ino == LOGFS_INO_MASTER)
logfs_write_anchor(inode->i_sb);
else {
- ret = __logfs_write_inode(inode, 0);
+ ret = __logfs_write_inode(inode, NULL, 0);
/* see indirect_write_block comment */
BUG_ON(ret);
}
}
-static gc_level_t inode_block_level(struct logfs_block *block)
-{
- BUG_ON(block->inode->i_ino == LOGFS_INO_MASTER);
- return GC_LEVEL(LOGFS_MAX_LEVELS);
-}
-
-static gc_level_t indirect_block_level(struct logfs_block *block)
-{
- struct page *page;
- struct inode *inode;
- u64 bix;
- level_t level;
-
- page = block->page;
- inode = page->mapping->host;
- logfs_unpack_index(page->index, &bix, &level);
- return expand_level(inode->i_ino, level);
-}
-
/*
* This silences a false, yet annoying gcc warning. I hate it when my editor
* jumps into bitops.h each time I recompile this file.
@@ -499,7 +479,7 @@ static int inode_write_alias(struct super_block *sb,
val = inode_val0(inode);
break;
case INODE_USED_OFS:
- val = cpu_to_be64(li->li_used_bytes);;
+ val = cpu_to_be64(li->li_used_bytes);
break;
case INODE_SIZE_OFS:
val = cpu_to_be64(i_size_read(inode));
@@ -537,9 +517,9 @@ static int indirect_write_alias(struct super_block *sb,
ino = page->mapping->host->i_ino;
logfs_unpack_index(page->index, &bix, &level);
- child = kmap_atomic(page, KM_USER0);
+ child = kmap_atomic(page);
val = child[pos];
- kunmap_atomic(child, KM_USER0);
+ kunmap_atomic(child);
err = write_one_alias(sb, ino, bix, level, pos, val);
if (err)
return err;
@@ -578,22 +558,25 @@ static void inode_free_block(struct super_block *sb, struct logfs_block *block)
static void indirect_free_block(struct super_block *sb,
struct logfs_block *block)
{
- ClearPagePrivate(block->page);
- block->page->private = 0;
+ struct page *page = block->page;
+
+ if (PagePrivate(page)) {
+ ClearPagePrivate(page);
+ page_cache_release(page);
+ set_page_private(page, 0);
+ }
__free_block(sb, block);
}
static struct logfs_block_ops inode_block_ops = {
.write_block = inode_write_block,
- .block_level = inode_block_level,
.free_block = inode_free_block,
.write_alias = inode_write_alias,
};
struct logfs_block_ops indirect_block_ops = {
.write_block = indirect_write_block,
- .block_level = indirect_block_level,
.free_block = indirect_free_block,
.write_alias = indirect_write_alias,
};
@@ -670,8 +653,11 @@ static void alloc_data_block(struct inode *inode, struct page *page)
logfs_unpack_index(page->index, &bix, &level);
block = __alloc_block(inode->i_sb, inode->i_ino, bix, level);
block->page = page;
+
SetPagePrivate(page);
- page->private = (unsigned long)block;
+ page_cache_get(page);
+ set_page_private(page, (unsigned long) block);
+
block->ops = &indirect_block_ops;
}
@@ -687,9 +673,9 @@ static void alloc_indirect_block(struct inode *inode, struct page *page,
alloc_data_block(inode, page);
block = logfs_block(page);
- array = kmap_atomic(page, KM_USER0);
+ array = kmap_atomic(page);
initialize_block_counters(page, block, array, page_is_empty);
- kunmap_atomic(array, KM_USER0);
+ kunmap_atomic(array);
}
static void block_set_pointer(struct page *page, int index, u64 ptr)
@@ -699,10 +685,10 @@ static void block_set_pointer(struct page *page, int index, u64 ptr)
u64 oldptr;
BUG_ON(!block);
- array = kmap_atomic(page, KM_USER0);
+ array = kmap_atomic(page);
oldptr = be64_to_cpu(array[index]);
array[index] = cpu_to_be64(ptr);
- kunmap_atomic(array, KM_USER0);
+ kunmap_atomic(array);
SetPageUptodate(page);
block->full += !!(ptr & LOGFS_FULLY_POPULATED)
@@ -715,9 +701,9 @@ static u64 block_get_pointer(struct page *page, int index)
__be64 *block;
u64 ptr;
- block = kmap_atomic(page, KM_USER0);
+ block = kmap_atomic(page);
ptr = be64_to_cpu(block[index]);
- kunmap_atomic(block, KM_USER0);
+ kunmap_atomic(block);
return ptr;
}
@@ -864,7 +850,7 @@ static u64 seek_holedata_loop(struct inode *inode, u64 bix, int data)
}
slot = get_bits(bix, SUBLEVEL(level));
- rblock = kmap_atomic(page, KM_USER0);
+ rblock = kmap_atomic(page);
while (slot < LOGFS_BLOCK_FACTOR) {
if (data && (rblock[slot] != 0))
break;
@@ -875,12 +861,12 @@ static u64 seek_holedata_loop(struct inode *inode, u64 bix, int data)
bix &= ~(increment - 1);
}
if (slot >= LOGFS_BLOCK_FACTOR) {
- kunmap_atomic(rblock, KM_USER0);
+ kunmap_atomic(rblock);
logfs_put_read_page(page);
return bix;
}
bofs = be64_to_cpu(rblock[slot]);
- kunmap_atomic(rblock, KM_USER0);
+ kunmap_atomic(rblock);
logfs_put_read_page(page);
if (!bofs) {
BUG_ON(data);
@@ -912,6 +898,8 @@ u64 logfs_seek_hole(struct inode *inode, u64 bix)
return bix;
else if (li->li_data[INDIRECT_INDEX] & LOGFS_FULLY_POPULATED)
bix = maxbix(li->li_height);
+ else if (bix >= maxbix(li->li_height))
+ return bix;
else {
bix = seek_holedata_loop(inode, bix, 0);
if (bix < maxbix(li->li_height))
@@ -1113,17 +1101,25 @@ static int logfs_reserve_bytes(struct inode *inode, int bytes)
int get_page_reserve(struct inode *inode, struct page *page)
{
struct logfs_super *super = logfs_super(inode->i_sb);
+ struct logfs_block *block = logfs_block(page);
int ret;
- if (logfs_block(page) && logfs_block(page)->reserved_bytes)
+ if (block && block->reserved_bytes)
return 0;
logfs_get_wblocks(inode->i_sb, page, WF_LOCK);
- ret = logfs_reserve_bytes(inode, 6 * LOGFS_MAX_OBJECTSIZE);
+ while ((ret = logfs_reserve_bytes(inode, 6 * LOGFS_MAX_OBJECTSIZE)) &&
+ !list_empty(&super->s_writeback_list)) {
+ block = list_entry(super->s_writeback_list.next,
+ struct logfs_block, alias_list);
+ block->ops->write_block(block);
+ }
if (!ret) {
alloc_data_block(inode, page);
- logfs_block(page)->reserved_bytes += 6 * LOGFS_MAX_OBJECTSIZE;
+ block = logfs_block(page);
+ block->reserved_bytes += 6 * LOGFS_MAX_OBJECTSIZE;
super->s_dirty_pages += 6 * LOGFS_MAX_OBJECTSIZE;
+ list_move_tail(&block->alias_list, &super->s_writeback_list);
}
logfs_put_wblocks(inode->i_sb, page, WF_LOCK);
return ret;
@@ -1240,6 +1236,18 @@ static void free_shadow(struct inode *inode, struct logfs_shadow *shadow)
mempool_free(shadow, super->s_shadow_pool);
}
+static void mark_segment(struct shadow_tree *tree, u32 segno)
+{
+ int err;
+
+ if (!btree_lookup32(&tree->segment_map, segno)) {
+ err = btree_insert32(&tree->segment_map, segno, (void *)1,
+ GFP_NOFS);
+ BUG_ON(err);
+ tree->no_shadowed_segments++;
+ }
+}
+
/**
* fill_shadow_tree - Propagate shadow tree changes due to a write
* @inode: Inode owning the page
@@ -1287,6 +1295,8 @@ static void fill_shadow_tree(struct inode *inode, struct page *page,
super->s_dirty_used_bytes += shadow->new_len;
super->s_dirty_free_bytes += shadow->old_len;
+ mark_segment(tree, shadow->old_ofs >> super->s_segshift);
+ mark_segment(tree, shadow->new_ofs >> super->s_segshift);
}
}
@@ -1566,11 +1576,15 @@ int logfs_write_buf(struct inode *inode, struct page *page, long flags)
static int __logfs_delete(struct inode *inode, struct page *page)
{
long flags = WF_DELETE;
+ int err;
inode->i_ctime = inode->i_mtime = CURRENT_TIME;
if (page->index < I0_BLOCKS)
return logfs_write_direct(inode, page, flags);
+ err = grow_inode(inode, page->index, 0);
+ if (err)
+ return err;
return logfs_write_rec(inode, page, page->index, 0, flags);
}
@@ -1594,7 +1608,6 @@ int logfs_delete(struct inode *inode, pgoff_t index,
return ret;
}
-/* Rewrite cannot mark the inode dirty but has to write it immediatly. */
int logfs_rewrite_block(struct inode *inode, u64 bix, u64 ofs,
gc_level_t gc_level, long flags)
{
@@ -1611,6 +1624,18 @@ int logfs_rewrite_block(struct inode *inode, u64 bix, u64 ofs,
if (level != 0)
alloc_indirect_block(inode, page, 0);
err = logfs_write_buf(inode, page, flags);
+ if (!err && shrink_level(gc_level) == 0) {
+ /* Rewrite cannot mark the inode dirty but has to
+ * write it immediately.
+ * Q: Can't we just create an alias for the inode
+ * instead? And if not, why not?
+ */
+ if (inode->i_ino == LOGFS_INO_MASTER)
+ logfs_write_anchor(inode->i_sb);
+ else {
+ err = __logfs_write_inode(inode, page, flags);
+ }
+ }
}
logfs_put_write_page(page);
return err;
@@ -1833,20 +1858,44 @@ static int __logfs_truncate(struct inode *inode, u64 size)
return logfs_truncate_direct(inode, size);
}
-int logfs_truncate(struct inode *inode, u64 size)
+/*
+ * Truncate, by changing the segment file, can consume a fair amount
+ * of resources. So back off from time to time and do some GC.
+ * 8 or 2048 blocks should be well within safety limits even if
+ * every single block resided in a different segment.
+ */
+#define TRUNCATE_STEP (8 * 1024 * 1024)
+int logfs_truncate(struct inode *inode, u64 target)
{
struct super_block *sb = inode->i_sb;
- int err;
+ u64 size = i_size_read(inode);
+ int err = 0;
- logfs_get_wblocks(sb, NULL, 1);
- err = __logfs_truncate(inode, size);
- if (!err)
- err = __logfs_write_inode(inode, 0);
- logfs_put_wblocks(sb, NULL, 1);
+ size = ALIGN(size, TRUNCATE_STEP);
+ while (size > target) {
+ if (size > TRUNCATE_STEP)
+ size -= TRUNCATE_STEP;
+ else
+ size = 0;
+ if (size < target)
+ size = target;
+
+ logfs_get_wblocks(sb, NULL, 1);
+ err = __logfs_truncate(inode, size);
+ if (!err)
+ err = __logfs_write_inode(inode, NULL, 0);
+ logfs_put_wblocks(sb, NULL, 1);
+ }
- if (!err)
- err = vmtruncate(inode, size);
+ if (!err) {
+ err = inode_newsize_ok(inode, target);
+ if (err)
+ goto out;
+
+ truncate_setsize(inode, target);
+ }
+ out:
/* I don't trust error recovery yet. */
WARN_ON(err);
return err;
@@ -1868,8 +1917,11 @@ static void move_page_to_inode(struct inode *inode, struct page *page)
li->li_block = block;
block->page = NULL;
- page->private = 0;
- ClearPagePrivate(page);
+ if (PagePrivate(page)) {
+ ClearPagePrivate(page);
+ page_cache_release(page);
+ set_page_private(page, 0);
+ }
}
static void move_inode_to_page(struct page *page, struct inode *inode)
@@ -1885,8 +1937,12 @@ static void move_inode_to_page(struct page *page, struct inode *inode)
BUG_ON(PagePrivate(page));
block->ops = &indirect_block_ops;
block->page = page;
- page->private = (unsigned long)block;
- SetPagePrivate(page);
+
+ if (!PagePrivate(page)) {
+ SetPagePrivate(page);
+ page_cache_get(page);
+ set_page_private(page, (unsigned long) block);
+ }
block->inode = NULL;
li->li_block = NULL;
@@ -1911,9 +1967,9 @@ int logfs_read_inode(struct inode *inode)
if (IS_ERR(page))
return PTR_ERR(page);
- di = kmap_atomic(page, KM_USER0);
+ di = kmap_atomic(page);
logfs_disk_to_inode(di, inode);
- kunmap_atomic(di, KM_USER0);
+ kunmap_atomic(di);
move_page_to_inode(inode, page);
page_cache_release(page);
return 0;
@@ -1932,38 +1988,13 @@ static struct page *inode_to_page(struct inode *inode)
if (!page)
return NULL;
- di = kmap_atomic(page, KM_USER0);
+ di = kmap_atomic(page);
logfs_inode_to_disk(inode, di);
- kunmap_atomic(di, KM_USER0);
+ kunmap_atomic(di);
move_inode_to_page(page, inode);
return page;
}
-/* Cheaper version of write_inode. All changes are concealed in
- * aliases, which are moved back. No write to the medium happens.
- */
-void logfs_clear_inode(struct inode *inode)
-{
- struct super_block *sb = inode->i_sb;
- struct logfs_inode *li = logfs_inode(inode);
- struct logfs_block *block = li->li_block;
- struct page *page;
-
- /* Only deleted files may be dirty at this point */
- BUG_ON(inode->i_state & I_DIRTY && inode->i_nlink);
- if (!block)
- return;
- if ((logfs_super(sb)->s_flags & LOGFS_SB_FLAG_SHUTDOWN)) {
- block->ops->free_block(inode->i_sb, block);
- return;
- }
-
- BUG_ON(inode->i_ino < LOGFS_RESERVED_INOS);
- page = inode_to_page(inode);
- BUG_ON(!page); /* FIXME: Use emergency page */
- logfs_put_write_page(page);
-}
-
static int do_write_inode(struct inode *inode)
{
struct super_block *sb = inode->i_sb;
@@ -1986,6 +2017,9 @@ static int do_write_inode(struct inode *inode)
/* FIXME: transaction is part of logfs_block now. Is that enough? */
err = logfs_write_buf(master_inode, page, 0);
+ if (err)
+ move_page_to_inode(inode, page);
+
logfs_put_write_page(page);
return err;
}
@@ -2013,13 +2047,13 @@ static void logfs_mod_segment_entry(struct super_block *sb, u32 segno,
if (write)
alloc_indirect_block(inode, page, 0);
- se = kmap_atomic(page, KM_USER0);
+ se = kmap_atomic(page);
change_se(se + child_no, arg);
if (write) {
logfs_set_alias(sb, logfs_block(page), child_no);
BUG_ON((int)be32_to_cpu(se[child_no].valid) > super->s_segsize);
}
- kunmap_atomic(se, KM_USER0);
+ kunmap_atomic(se);
logfs_put_write_page(page);
}
@@ -2095,14 +2129,14 @@ void logfs_set_segment_unreserved(struct super_block *sb, u32 segno, u32 ec)
ec_level);
}
-int __logfs_write_inode(struct inode *inode, long flags)
+int __logfs_write_inode(struct inode *inode, struct page *page, long flags)
{
struct super_block *sb = inode->i_sb;
int ret;
- logfs_get_wblocks(sb, NULL, flags & WF_LOCK);
+ logfs_get_wblocks(sb, page, flags & WF_LOCK);
ret = do_write_inode(inode);
- logfs_put_wblocks(sb, NULL, flags & WF_LOCK);
+ logfs_put_wblocks(sb, page, flags & WF_LOCK);
return ret;
}
@@ -2131,18 +2165,39 @@ static int do_delete_inode(struct inode *inode)
* ZOMBIE inodes have already been deleted before and should remain dead,
* if it weren't for valid checking. No need to kill them again here.
*/
-void logfs_delete_inode(struct inode *inode)
+void logfs_evict_inode(struct inode *inode)
{
+ struct super_block *sb = inode->i_sb;
struct logfs_inode *li = logfs_inode(inode);
+ struct logfs_block *block = li->li_block;
+ struct page *page;
- if (!(li->li_flags & LOGFS_IF_ZOMBIE)) {
- li->li_flags |= LOGFS_IF_ZOMBIE;
- if (i_size_read(inode) > 0)
- logfs_truncate(inode, 0);
- do_delete_inode(inode);
+ if (!inode->i_nlink) {
+ if (!(li->li_flags & LOGFS_IF_ZOMBIE)) {
+ li->li_flags |= LOGFS_IF_ZOMBIE;
+ if (i_size_read(inode) > 0)
+ logfs_truncate(inode, 0);
+ do_delete_inode(inode);
+ }
}
- truncate_inode_pages(&inode->i_data, 0);
+ truncate_inode_pages_final(&inode->i_data);
clear_inode(inode);
+
+ /* Cheaper version of write_inode. All changes are concealed in
+ * aliases, which are moved back. No write to the medium happens.
+ */
+ /* Only deleted files may be dirty at this point */
+ BUG_ON(inode->i_state & I_DIRTY && inode->i_nlink);
+ if (!block)
+ return;
+ if ((logfs_super(sb)->s_flags & LOGFS_SB_FLAG_SHUTDOWN)) {
+ block->ops->free_block(inode->i_sb, block);
+ return;
+ }
+
+ page = inode_to_page(inode);
+ BUG_ON(!page); /* FIXME: Use emergency page */
+ logfs_put_write_page(page);
}
void btree_write_block(struct logfs_block *block)
@@ -2195,10 +2250,10 @@ int logfs_inode_write(struct inode *inode, const void *buf, size_t count,
if (!page)
return -ENOMEM;
- pagebuf = kmap_atomic(page, KM_USER0);
+ pagebuf = kmap_atomic(page);
memcpy(pagebuf, buf, count);
flush_dcache_page(page);
- kunmap_atomic(pagebuf, KM_USER0);
+ kunmap_atomic(pagebuf);
if (i_size_read(inode) < pos + LOGFS_BLOCKSIZE)
i_size_write(inode, pos + LOGFS_BLOCKSIZE);
@@ -2226,6 +2281,7 @@ int logfs_init_rw(struct super_block *sb)
int min_fill = 3 * super->s_no_blocks;
INIT_LIST_HEAD(&super->s_object_alias);
+ INIT_LIST_HEAD(&super->s_writeback_list);
mutex_init(&super->s_write_mutex);
super->s_block_pool = mempool_create_kmalloc_pool(min_fill,
sizeof(struct logfs_block));
@@ -2238,9 +2294,6 @@ void logfs_cleanup_rw(struct super_block *sb)
{
struct logfs_super *super = logfs_super(sb);
- destroy_meta_inode(super->s_segfile_inode);
- if (super->s_block_pool)
- mempool_destroy(super->s_block_pool);
- if (super->s_shadow_pool)
- mempool_destroy(super->s_shadow_pool);
+ logfs_mempool_destroy(super->s_block_pool);
+ logfs_mempool_destroy(super->s_shadow_pool);
}
diff --git a/fs/logfs/segment.c b/fs/logfs/segment.c
index 1a14f9910d5..7f9b096d8d5 100644
--- a/fs/logfs/segment.c
+++ b/fs/logfs/segment.c
@@ -10,6 +10,7 @@
* three kinds of objects: inodes, dentries and blocks, both data and indirect.
*/
#include "logfs.h"
+#include <linux/slab.h>
static int logfs_mark_segment_bad(struct super_block *sb, u32 segno)
{
@@ -61,12 +62,13 @@ static struct page *get_mapping_page(struct super_block *sb, pgoff_t index,
page = read_cache_page(mapping, index, filler, sb);
else {
page = find_or_create_page(mapping, index, GFP_NOFS);
- unlock_page(page);
+ if (page)
+ unlock_page(page);
}
return page;
}
-void __logfs_buf_write(struct logfs_area *area, u64 ofs, void *buf, size_t len,
+int __logfs_buf_write(struct logfs_area *area, u64 ofs, void *buf, size_t len,
int use_filler)
{
pgoff_t index = ofs >> PAGE_SHIFT;
@@ -80,10 +82,16 @@ void __logfs_buf_write(struct logfs_area *area, u64 ofs, void *buf, size_t len,
copylen = min((ulong)len, PAGE_SIZE - offset);
page = get_mapping_page(area->a_sb, index, use_filler);
- SetPageUptodate(page);
+ if (IS_ERR(page))
+ return PTR_ERR(page);
BUG_ON(!page); /* FIXME: reserve a pool */
+ SetPageUptodate(page);
memcpy(page_address(page) + offset, buf, copylen);
- SetPagePrivate(page);
+
+ if (!PagePrivate(page)) {
+ SetPagePrivate(page);
+ page_cache_get(page);
+ }
page_cache_release(page);
buf += copylen;
@@ -91,52 +99,67 @@ void __logfs_buf_write(struct logfs_area *area, u64 ofs, void *buf, size_t len,
offset = 0;
index++;
} while (len);
+ return 0;
}
-/*
- * bdev_writeseg will write full pages. Memset the tail to prevent data leaks.
- */
-static void pad_wbuf(struct logfs_area *area, int final)
+static void pad_partial_page(struct logfs_area *area)
{
struct super_block *sb = area->a_sb;
- struct logfs_super *super = logfs_super(sb);
struct page *page;
u64 ofs = dev_ofs(sb, area->a_segno, area->a_used_bytes);
pgoff_t index = ofs >> PAGE_SHIFT;
long offset = ofs & (PAGE_SIZE-1);
u32 len = PAGE_SIZE - offset;
- if (len == PAGE_SIZE) {
- /* The math in this function can surely use some love */
- len = 0;
- }
- if (len) {
- BUG_ON(area->a_used_bytes >= super->s_segsize);
-
- page = get_mapping_page(area->a_sb, index, 0);
+ if (len % PAGE_SIZE) {
+ page = get_mapping_page(sb, index, 0);
BUG_ON(!page); /* FIXME: reserve a pool */
memset(page_address(page) + offset, 0xff, len);
- SetPagePrivate(page);
+ if (!PagePrivate(page)) {
+ SetPagePrivate(page);
+ page_cache_get(page);
+ }
page_cache_release(page);
}
+}
- if (!final)
- return;
+static void pad_full_pages(struct logfs_area *area)
+{
+ struct super_block *sb = area->a_sb;
+ struct logfs_super *super = logfs_super(sb);
+ u64 ofs = dev_ofs(sb, area->a_segno, area->a_used_bytes);
+ u32 len = super->s_segsize - area->a_used_bytes;
+ pgoff_t index = PAGE_CACHE_ALIGN(ofs) >> PAGE_CACHE_SHIFT;
+ pgoff_t no_indizes = len >> PAGE_CACHE_SHIFT;
+ struct page *page;
- area->a_used_bytes += len;
- for ( ; area->a_used_bytes < super->s_segsize;
- area->a_used_bytes += PAGE_SIZE) {
- /* Memset another page */
- index++;
- page = get_mapping_page(area->a_sb, index, 0);
+ while (no_indizes) {
+ page = get_mapping_page(sb, index, 0);
BUG_ON(!page); /* FIXME: reserve a pool */
- memset(page_address(page), 0xff, PAGE_SIZE);
- SetPagePrivate(page);
+ SetPageUptodate(page);
+ memset(page_address(page), 0xff, PAGE_CACHE_SIZE);
+ if (!PagePrivate(page)) {
+ SetPagePrivate(page);
+ page_cache_get(page);
+ }
page_cache_release(page);
+ index++;
+ no_indizes--;
}
}
/*
+ * bdev_writeseg will write full pages. Memset the tail to prevent data leaks.
+ * Also make sure we allocate (and memset) all pages for final writeout.
+ */
+static void pad_wbuf(struct logfs_area *area, int final)
+{
+ pad_partial_page(area);
+ if (final)
+ pad_full_pages(area);
+}
+
+/*
* We have to be careful with the alias tree. Since lookup is done by bix,
* it needs to be normalized, so 14, 15, 16, etc. all match when dealing with
* indirect blocks. So always use it through accessor functions.
@@ -174,14 +197,8 @@ static int btree_write_alias(struct super_block *sb, struct logfs_block *block,
return 0;
}
-static gc_level_t btree_block_level(struct logfs_block *block)
-{
- return expand_level(block->ino, block->level);
-}
-
static struct logfs_block_ops btree_block_ops = {
.write_block = btree_write_block,
- .block_level = btree_block_level,
.free_block = __free_block,
.write_alias = btree_write_alias,
};
@@ -479,8 +496,12 @@ static void move_btree_to_page(struct inode *inode, struct page *page,
mempool_free(item, super->s_alias_pool);
}
block->page = page;
- SetPagePrivate(page);
- page->private = (unsigned long)block;
+
+ if (!PagePrivate(page)) {
+ SetPagePrivate(page);
+ page_cache_get(page);
+ set_page_private(page, (unsigned long) block);
+ }
block->ops = &indirect_block_ops;
initialize_block_counters(page, block, data, 0);
}
@@ -523,15 +544,19 @@ void move_page_to_btree(struct page *page)
BUG_ON(!item); /* mempool empty */
memset(item, 0, sizeof(*item));
- child = kmap_atomic(page, KM_USER0);
+ child = kmap_atomic(page);
item->val = child[pos];
- kunmap_atomic(child, KM_USER0);
+ kunmap_atomic(child);
item->child_no = pos;
list_add(&item->list, &block->item_list);
}
block->page = NULL;
- ClearPagePrivate(page);
- page->private = 0;
+
+ if (PagePrivate(page)) {
+ ClearPagePrivate(page);
+ page_cache_release(page);
+ set_page_private(page, 0);
+ }
block->ops = &btree_block_ops;
err = alias_tree_insert(block->sb, block->ino, block->bix, block->level,
block);
@@ -683,7 +708,7 @@ int logfs_segment_delete(struct inode *inode, struct logfs_shadow *shadow)
return 0;
}
-static void freeseg(struct super_block *sb, u32 segno)
+void freeseg(struct super_block *sb, u32 segno)
{
struct logfs_super *super = logfs_super(sb);
struct address_space *mapping = super->s_mapping_inode->i_mapping;
@@ -696,7 +721,10 @@ static void freeseg(struct super_block *sb, u32 segno)
page = find_get_page(mapping, ofs >> PAGE_SHIFT);
if (!page)
continue;
- ClearPagePrivate(page);
+ if (PagePrivate(page)) {
+ ClearPagePrivate(page);
+ page_cache_release(page);
+ }
page_cache_release(page);
}
}
@@ -835,6 +863,16 @@ static void free_area(struct logfs_area *area)
kfree(area);
}
+void free_areas(struct super_block *sb)
+{
+ struct logfs_super *super = logfs_super(sb);
+ int i;
+
+ for_each_area(i)
+ free_area(super->s_area[i]);
+ free_area(super->s_journal_area);
+}
+
static struct logfs_area *alloc_area(struct super_block *sb)
{
struct logfs_area *area;
@@ -847,9 +885,10 @@ static struct logfs_area *alloc_area(struct super_block *sb)
return area;
}
-static void map_invalidatepage(struct page *page, unsigned long l)
+static void map_invalidatepage(struct page *page, unsigned int o,
+ unsigned int l)
{
- BUG();
+ return;
}
static int map_releasepage(struct page *page, gfp_t g)
@@ -910,18 +949,13 @@ err:
for (i--; i >= 0; i--)
free_area(super->s_area[i]);
free_area(super->s_journal_area);
- mempool_destroy(super->s_alias_pool);
+ logfs_mempool_destroy(super->s_alias_pool);
return -ENOMEM;
}
void logfs_cleanup_areas(struct super_block *sb)
{
struct logfs_super *super = logfs_super(sb);
- int i;
btree_grim_visitor128(&super->s_object_alias_tree, 0, kill_alias);
- for_each_area(i)
- free_area(super->s_area[i]);
- free_area(super->s_journal_area);
- destroy_meta_inode(super->s_mapping_inode);
}
diff --git a/fs/logfs/super.c b/fs/logfs/super.c
index c66beab78de..54360293bcb 100644
--- a/fs/logfs/super.c
+++ b/fs/logfs/super.c
@@ -11,6 +11,9 @@
*/
#include "logfs.h"
#include <linux/bio.h>
+#include <linux/slab.h>
+#include <linux/blkdev.h>
+#include <linux/module.h>
#include <linux/mtd/mtd.h>
#include <linux/statfs.h>
#include <linux/buffer_head.h>
@@ -89,28 +92,6 @@ void logfs_crash_dump(struct super_block *sb)
}
/*
- * TODO: move to lib/string.c
- */
-/**
- * memchr_inv - Find a character in an area of memory.
- * @s: The memory area
- * @c: The byte to search for
- * @n: The size of the area.
- *
- * returns the address of the first character other than @c, or %NULL
- * if the whole buffer contains just @c.
- */
-void *memchr_inv(const void *s, int c, size_t n)
-{
- const unsigned char *p = s;
- while (n-- != 0)
- if ((unsigned char)c != *p++)
- return (void *)(p - 1);
-
- return NULL;
-}
-
-/*
* FIXME: There should be a reserve for root, similar to ext2.
*/
int logfs_statfs(struct dentry *dentry, struct kstatfs *stats)
@@ -136,6 +117,14 @@ static int logfs_sb_set(struct super_block *sb, void *_super)
sb->s_fs_info = super;
sb->s_mtd = super->s_mtd;
sb->s_bdev = super->s_bdev;
+#ifdef CONFIG_BLOCK
+ if (sb->s_bdev)
+ sb->s_bdi = &bdev_get_queue(sb->s_bdev)->backing_dev_info;
+#endif
+#ifdef CONFIG_MTD
+ if (sb->s_mtd)
+ sb->s_bdi = sb->s_mtd->backing_dev_info;
+#endif
return 0;
}
@@ -277,7 +266,7 @@ static int logfs_recover_sb(struct super_block *sb)
}
if (valid0 && valid1 && ds_cmp(ds0, ds1)) {
printk(KERN_INFO"Superblocks don't match - fixing.\n");
- return write_one_sb(sb, super->s_devops->find_last_sb);
+ return logfs_write_sb(sb);
}
/* If neither is valid now, something's wrong. Didn't we properly
* check them before?!? */
@@ -289,6 +278,10 @@ static int logfs_make_writeable(struct super_block *sb)
{
int err;
+ err = logfs_open_segfile(sb);
+ if (err)
+ return err;
+
/* Repair any broken superblock copies */
err = logfs_recover_sb(sb);
if (err)
@@ -299,10 +292,6 @@ static int logfs_make_writeable(struct super_block *sb)
if (err)
return err;
- err = logfs_open_segfile(sb);
- if (err)
- return err;
-
/* Do one GC pass before any data gets dirtied */
logfs_gc_pass(sb);
@@ -315,7 +304,7 @@ static int logfs_make_writeable(struct super_block *sb)
return 0;
}
-static int logfs_get_sb_final(struct super_block *sb, struct vfsmount *mnt)
+static int logfs_get_sb_final(struct super_block *sb)
{
struct logfs_super *super = logfs_super(sb);
struct inode *rootdir;
@@ -326,30 +315,30 @@ static int logfs_get_sb_final(struct super_block *sb, struct vfsmount *mnt)
if (IS_ERR(rootdir))
goto fail;
- sb->s_root = d_alloc_root(rootdir);
+ sb->s_root = d_make_root(rootdir);
if (!sb->s_root)
goto fail;
+ /* at that point we know that ->put_super() will be called */
super->s_erase_page = alloc_pages(GFP_KERNEL, 0);
if (!super->s_erase_page)
- goto fail2;
+ return -ENOMEM;
memset(page_address(super->s_erase_page), 0xFF, PAGE_SIZE);
/* FIXME: check for read-only mounts */
err = logfs_make_writeable(sb);
- if (err)
- goto fail3;
+ if (err) {
+ __free_page(super->s_erase_page);
+ return err;
+ }
log_super("LogFS: Finished mounting\n");
- simple_set_mnt(mnt, sb);
return 0;
-fail3:
- __free_page(super->s_erase_page);
-fail2:
- iput(rootdir);
fail:
- iput(logfs_super(sb)->s_master_inode);
+ iput(super->s_master_inode);
+ iput(super->s_segfile_inode);
+ iput(super->s_mapping_inode);
return -EIO;
}
@@ -376,7 +365,7 @@ static struct page *find_super_block(struct super_block *sb)
if (!first || IS_ERR(first))
return NULL;
last = super->s_devops->find_last_sb(sb, &super->s_sb_ofs[1]);
- if (!last || IS_ERR(first)) {
+ if (!last || IS_ERR(last)) {
page_cache_release(first);
return NULL;
}
@@ -407,7 +396,7 @@ static int __logfs_read_sb(struct super_block *sb)
page = find_super_block(sb);
if (!page)
- return -EIO;
+ return -EINVAL;
ds = page_address(page);
super->s_size = be64_to_cpu(ds->ds_filesystem_size);
@@ -451,6 +440,8 @@ static int logfs_read_sb(struct super_block *sb, int read_only)
btree_init_mempool64(&super->s_shadow_tree.new, super->s_btree_pool);
btree_init_mempool64(&super->s_shadow_tree.old, super->s_btree_pool);
+ btree_init_mempool32(&super->s_shadow_tree.segment_map,
+ super->s_btree_pool);
ret = logfs_init_mapping(sb);
if (ret)
@@ -466,10 +457,6 @@ static int logfs_read_sb(struct super_block *sb, int read_only)
!read_only)
return -EIO;
- mutex_init(&super->s_dirop_mutex);
- mutex_init(&super->s_object_alias_mutex);
- INIT_LIST_HEAD(&super->s_freeing_list);
-
ret = logfs_init_rw(sb);
if (ret)
return ret;
@@ -497,14 +484,15 @@ static void logfs_kill_sb(struct super_block *sb)
/* Alias entries slow down mount, so evict as many as possible */
sync_filesystem(sb);
logfs_write_anchor(sb);
+ free_areas(sb);
/*
* From this point on alias entries are simply dropped - and any
* writes to the object store are considered bugs.
*/
- super->s_flags |= LOGFS_SB_FLAG_SHUTDOWN;
log_super("LogFS: Now in shutdown\n");
generic_shutdown_super(sb);
+ super->s_flags |= LOGFS_SB_FLAG_SHUTDOWN;
BUG_ON(super->s_dirty_used_bytes || super->s_dirty_free_bytes);
@@ -514,43 +502,37 @@ static void logfs_kill_sb(struct super_block *sb)
logfs_cleanup_rw(sb);
if (super->s_erase_page)
__free_page(super->s_erase_page);
- super->s_devops->put_device(sb);
- mempool_destroy(super->s_btree_pool);
- mempool_destroy(super->s_alias_pool);
+ super->s_devops->put_device(super);
+ logfs_mempool_destroy(super->s_btree_pool);
+ logfs_mempool_destroy(super->s_alias_pool);
kfree(super);
log_super("LogFS: Finished unmounting\n");
}
-int logfs_get_sb_device(struct file_system_type *type, int flags,
- struct mtd_info *mtd, struct block_device *bdev,
- const struct logfs_device_ops *devops, struct vfsmount *mnt)
+static struct dentry *logfs_get_sb_device(struct logfs_super *super,
+ struct file_system_type *type, int flags)
{
- struct logfs_super *super;
struct super_block *sb;
int err = -ENOMEM;
static int mount_count;
log_super("LogFS: Start mount %x\n", mount_count++);
- super = kzalloc(sizeof(*super), GFP_KERNEL);
- if (!super)
- goto err0;
- super->s_mtd = mtd;
- super->s_bdev = bdev;
err = -EINVAL;
- sb = sget(type, logfs_sb_test, logfs_sb_set, super);
- if (IS_ERR(sb))
- goto err0;
+ sb = sget(type, logfs_sb_test, logfs_sb_set, flags | MS_NOATIME, super);
+ if (IS_ERR(sb)) {
+ super->s_devops->put_device(super);
+ kfree(super);
+ return ERR_CAST(sb);
+ }
if (sb->s_root) {
/* Device is already in use */
- err = 0;
- simple_set_mnt(mnt, sb);
- goto err0;
+ super->s_devops->put_device(super);
+ kfree(super);
+ return dget(sb->s_root);
}
- super->s_devops = devops;
-
/*
* sb->s_maxbytes is limited to 8TB. On 32bit systems, the page cache
* only covers 16TB and the upper 8TB are used for indirect blocks.
@@ -558,57 +540,75 @@ int logfs_get_sb_device(struct file_system_type *type, int flags,
* the filesystem incompatible with 32bit systems.
*/
sb->s_maxbytes = (1ull << 43) - 1;
+ sb->s_max_links = LOGFS_LINK_MAX;
sb->s_op = &logfs_super_operations;
- sb->s_flags = flags | MS_NOATIME;
err = logfs_read_sb(sb, sb->s_flags & MS_RDONLY);
if (err)
goto err1;
sb->s_flags |= MS_ACTIVE;
- err = logfs_get_sb_final(sb, mnt);
- if (err)
- goto err1;
- return 0;
+ err = logfs_get_sb_final(sb);
+ if (err) {
+ deactivate_locked_super(sb);
+ return ERR_PTR(err);
+ }
+ return dget(sb->s_root);
err1:
- up_write(&sb->s_umount);
- deactivate_super(sb);
- return err;
-err0:
- kfree(super);
- //devops->put_device(sb);
- return err;
+ /* no ->s_root, no ->put_super() */
+ iput(super->s_master_inode);
+ iput(super->s_segfile_inode);
+ iput(super->s_mapping_inode);
+ deactivate_locked_super(sb);
+ return ERR_PTR(err);
}
-static int logfs_get_sb(struct file_system_type *type, int flags,
- const char *devname, void *data, struct vfsmount *mnt)
+static struct dentry *logfs_mount(struct file_system_type *type, int flags,
+ const char *devname, void *data)
{
ulong mtdnr;
+ struct logfs_super *super;
+ int err;
- if (!devname)
- return logfs_get_sb_bdev(type, flags, devname, mnt);
- if (strncmp(devname, "mtd", 3))
- return logfs_get_sb_bdev(type, flags, devname, mnt);
+ super = kzalloc(sizeof(*super), GFP_KERNEL);
+ if (!super)
+ return ERR_PTR(-ENOMEM);
- {
+ mutex_init(&super->s_dirop_mutex);
+ mutex_init(&super->s_object_alias_mutex);
+ INIT_LIST_HEAD(&super->s_freeing_list);
+
+ if (!devname)
+ err = logfs_get_sb_bdev(super, type, devname);
+ else if (strncmp(devname, "mtd", 3))
+ err = logfs_get_sb_bdev(super, type, devname);
+ else {
char *garbage;
mtdnr = simple_strtoul(devname+3, &garbage, 0);
if (*garbage)
- return -EINVAL;
+ err = -EINVAL;
+ else
+ err = logfs_get_sb_mtd(super, mtdnr);
}
- return logfs_get_sb_mtd(type, flags, mtdnr, mnt);
+ if (err) {
+ kfree(super);
+ return ERR_PTR(err);
+ }
+
+ return logfs_get_sb_device(super, type, flags);
}
static struct file_system_type logfs_fs_type = {
.owner = THIS_MODULE,
.name = "logfs",
- .get_sb = logfs_get_sb,
+ .mount = logfs_mount,
.kill_sb = logfs_kill_sb,
.fs_flags = FS_REQUIRES_DEV,
};
+MODULE_ALIAS_FS("logfs");
static int __init logfs_init(void)
{
@@ -626,7 +626,10 @@ static int __init logfs_init(void)
if (ret)
goto out2;
- return register_filesystem(&logfs_fs_type);
+ ret = register_filesystem(&logfs_fs_type);
+ if (!ret)
+ return 0;
+ logfs_destroy_inode_cache();
out2:
logfs_compr_exit();
out1:
generated by cgit v1.2.3-70-g09d2 (git 2.46.0) at 2025-12-06 19:34:04 +0000