Minki/linux
1
0
Fork 1
mirror of https://github.com/torvalds/linux.git synced 2024-11-27 14:41:39 +00:00
Code Issues Packages Projects Releases Wiki Activity
8050318598
linux/fs/reiserfs/dir.c
Frederic Weisbecker 8ebc423238 reiserfs: kill-the-BKL 
This patch is an attempt to remove the Bkl based locking scheme from
reiserfs and is intended.

It is a bit inspired from an old attempt by Peter Zijlstra:

   http://lkml.indiana.edu/hypermail/linux/kernel/0704.2/2174.html

The bkl is heavily used in this filesystem to prevent from
concurrent write accesses on the filesystem.

Reiserfs makes a deep use of the specific properties of the Bkl:

- It can be acqquired recursively by a same task
- It is released on the schedule() calls and reacquired when schedule() returns

The two properties above are a roadmap for the reiserfs write locking so it's
very hard to simply replace it with a common mutex.

- We need a recursive-able locking unless we want to restructure several blocks
  of the code.
- We need to identify the sites where the bkl was implictly relaxed
  (schedule, wait, sync, etc...) so that we can in turn release and
  reacquire our new lock explicitly.
  Such implicit releases of the lock are often required to let other
  resources producer/consumer do their job or we can suffer unexpected
  starvations or deadlocks.

So the new lock that replaces the bkl here is a per superblock mutex with a
specific property: it can be acquired recursively by a same task, like the
bkl.

For such purpose, we integrate a lock owner and a lock depth field on the
superblock information structure.

The first axis on this patch is to turn reiserfs_write_(un)lock() function
into a wrapper to manage this mutex. Also some explicit calls to
lock_kernel() have been converted to reiserfs_write_lock() helpers.

The second axis is to find the important blocking sites (schedule...(),
wait_on_buffer(), sync_dirty_buffer(), etc...) and then apply an explicit
release of the write lock on these locations before blocking. Then we can
safely wait for those who can give us resources or those who need some.
Typically this is a fight between the current writer, the reiserfs workqueue
(aka the async commiter) and the pdflush threads.

The third axis is a consequence of the second. The write lock is usually
on top of a lock dependency chain which can include the journal lock, the
flush lock or the commit lock. So it's dangerous to release and trying to
reacquire the write lock while we still hold other locks.

This is fine with the bkl:

      T1                       T2

lock_kernel()
    mutex_lock(A)
    unlock_kernel()
    // do something
                            lock_kernel()
                                mutex_lock(A) -> already locked by T1
                                schedule() (and then unlock_kernel())
    lock_kernel()
    mutex_unlock(A)
    ....

This is not fine with a mutex:

      T1                       T2

mutex_lock(write)
    mutex_lock(A)
    mutex_unlock(write)
    // do something
                           mutex_lock(write)
                              mutex_lock(A) -> already locked by T1
                              schedule()

    mutex_lock(write) -> already locked by T2
    deadlock

The solution in this patch is to provide a helper which releases the write
lock and sleep a bit if we can't lock a mutex that depend on it. It's another
simulation of the bkl behaviour.

The last axis is to locate the fs callbacks that are called with the bkl held,
according to Documentation/filesystem/Locking.

Those are:

- reiserfs_remount
- reiserfs_fill_super
- reiserfs_put_super

Reiserfs didn't need to explicitly lock because of the context of these callbacks.
But now we must take care of that with the new locking.

After this patch, reiserfs suffers from a slight performance regression (for now).
On UP, a high volume write with dd reports an average of 27 MB/s instead
of 30 MB/s without the patch applied.

Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
Reviewed-by: Ingo Molnar <mingo@elte.hu>
Cc: Jeff Mahoney <jeffm@suse.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Bron Gondwana <brong@fastmail.fm>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
LKML-Reference: <1239070789-13354-1-git-send-email-fweisbec@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-14 07:17:59 +02:00

313 lines
8.9 KiB
C
Raw Blame History

/*
* Copyright 2000 by Hans Reiser, licensing governed by reiserfs/README
*/
#include <linux/string.h>
#include <linux/errno.h>
#include <linux/fs.h>
#include <linux/reiserfs_fs.h>
#include <linux/stat.h>
#include <linux/buffer_head.h>
#include <asm/uaccess.h>
extern const struct reiserfs_key MIN_KEY;
static int reiserfs_readdir(struct file *, void *, filldir_t);
static int reiserfs_dir_fsync(struct file *filp, struct dentry *dentry,
int datasync);
const struct file_operations reiserfs_dir_operations = {
.read = generic_read_dir,
.readdir = reiserfs_readdir,
.fsync = reiserfs_dir_fsync,
.ioctl = reiserfs_ioctl,
#ifdef CONFIG_COMPAT
.compat_ioctl = reiserfs_compat_ioctl,
#endif
};
static int reiserfs_dir_fsync(struct file *filp, struct dentry *dentry,
int datasync)
{
struct inode *inode = dentry->d_inode;
int err;
reiserfs_write_lock(inode->i_sb);
err = reiserfs_commit_for_inode(inode);
reiserfs_write_unlock(inode->i_sb);
if (err < 0)
return err;
return 0;
}
#define store_ih(where,what) copy_item_head (where, what)
static inline bool is_privroot_deh(struct dentry *dir,
struct reiserfs_de_head *deh)
{
struct dentry *privroot = REISERFS_SB(dir->d_sb)->priv_root;
if (reiserfs_expose_privroot(dir->d_sb))
return 0;
return (dir == dir->d_parent && privroot->d_inode &&
deh->deh_objectid == INODE_PKEY(privroot->d_inode)->k_objectid);
}
int reiserfs_readdir_dentry(struct dentry *dentry, void *dirent,
filldir_t filldir, loff_t *pos)
{
struct inode *inode = dentry->d_inode;
struct cpu_key pos_key; /* key of current position in the directory (key of directory entry) */
INITIALIZE_PATH(path_to_entry);
struct buffer_head *bh;
int item_num, entry_num;
const struct reiserfs_key *rkey;
struct item_head *ih, tmp_ih;
int search_res;
char *local_buf;
loff_t next_pos;
char small_buf[32]; /* avoid kmalloc if we can */
struct reiserfs_dir_entry de;
int ret = 0;
reiserfs_write_lock(inode->i_sb);
reiserfs_check_lock_depth(inode->i_sb, "readdir");
/* form key for search the next directory entry using f_pos field of
file structure */
make_cpu_key(&pos_key, inode, *pos ?: DOT_OFFSET, TYPE_DIRENTRY, 3);
next_pos = cpu_key_k_offset(&pos_key);
path_to_entry.reada = PATH_READA;
while (1) {
research:
/* search the directory item, containing entry with specified key */
search_res =
search_by_entry_key(inode->i_sb, &pos_key, &path_to_entry,
&de);
if (search_res == IO_ERROR) {
// FIXME: we could just skip part of directory which could
// not be read
ret = -EIO;
goto out;
}
entry_num = de.de_entry_num;
bh = de.de_bh;
item_num = de.de_item_num;
ih = de.de_ih;
store_ih(&tmp_ih, ih);
/* we must have found item, that is item of this directory, */
RFALSE(COMP_SHORT_KEYS(&(ih->ih_key), &pos_key),
"vs-9000: found item %h does not match to dir we readdir %K",
ih, &pos_key);
RFALSE(item_num > B_NR_ITEMS(bh) - 1,
"vs-9005 item_num == %d, item amount == %d",
item_num, B_NR_ITEMS(bh));
/* and entry must be not more than number of entries in the item */
RFALSE(I_ENTRY_COUNT(ih) < entry_num,
"vs-9010: entry number is too big %d (%d)",
entry_num, I_ENTRY_COUNT(ih));
if (search_res == POSITION_FOUND
|| entry_num < I_ENTRY_COUNT(ih)) {
/* go through all entries in the directory item beginning from the entry, that has been found */
struct reiserfs_de_head *deh =
B_I_DEH(bh, ih) + entry_num;
for (; entry_num < I_ENTRY_COUNT(ih);
entry_num++, deh++) {
int d_reclen;
char *d_name;
off_t d_off;
ino_t d_ino;
if (!de_visible(deh))
/* it is hidden entry */
continue;
d_reclen = entry_length(bh, ih, entry_num);
d_name = B_I_DEH_ENTRY_FILE_NAME(bh, ih, deh);
if (d_reclen <= 0 ||
d_name + d_reclen > bh->b_data + bh->b_size) {
/* There is corrupted data in entry,
* We'd better stop here */
pathrelse(&path_to_entry);
ret = -EIO;
goto out;
}
if (!d_name[d_reclen - 1])
d_reclen = strlen(d_name);
if (d_reclen >
REISERFS_MAX_NAME(inode->i_sb->
s_blocksize)) {
/* too big to send back to VFS */
continue;
}
/* Ignore the .reiserfs_priv entry */
if (is_privroot_deh(dentry, deh))
continue;
d_off = deh_offset(deh);
*pos = d_off;
d_ino = deh_objectid(deh);
if (d_reclen <= 32) {
local_buf = small_buf;
} else {
local_buf = kmalloc(d_reclen,
GFP_NOFS);
if (!local_buf) {
pathrelse(&path_to_entry);
ret = -ENOMEM;
goto out;
}
if (item_moved(&tmp_ih, &path_to_entry)) {
kfree(local_buf);
goto research;
}
}
// Note, that we copy name to user space via temporary
// buffer (local_buf) because filldir will block if
// user space buffer is swapped out. At that time
// entry can move to somewhere else
memcpy(local_buf, d_name, d_reclen);
/*
* Since filldir might sleep, we can release
* the write lock here for other waiters
*/
reiserfs_write_unlock(inode->i_sb);
if (filldir
(dirent, local_buf, d_reclen, d_off, d_ino,
DT_UNKNOWN) < 0) {
reiserfs_write_lock(inode->i_sb);
if (local_buf != small_buf) {
kfree(local_buf);
}
goto end;
}
reiserfs_write_lock(inode->i_sb);
if (local_buf != small_buf) {
kfree(local_buf);
}
// next entry should be looked for with such offset
next_pos = deh_offset(deh) + 1;
if (item_moved(&tmp_ih, &path_to_entry)) {
goto research;
}
} /* for */
}
if (item_num != B_NR_ITEMS(bh) - 1)
// end of directory has been reached
goto end;
/* item we went through is last item of node. Using right
delimiting key check is it directory end */
rkey = get_rkey(&path_to_entry, inode->i_sb);
if (!comp_le_keys(rkey, &MIN_KEY)) {
/* set pos_key to key, that is the smallest and greater
that key of the last entry in the item */
set_cpu_key_k_offset(&pos_key, next_pos);
continue;
}
if (COMP_SHORT_KEYS(rkey, &pos_key)) {
// end of directory has been reached
goto end;
}
/* directory continues in the right neighboring block */
set_cpu_key_k_offset(&pos_key,
le_key_k_offset(KEY_FORMAT_3_5, rkey));
} /* while */
end:
*pos = next_pos;
pathrelse(&path_to_entry);
reiserfs_check_path(&path_to_entry);
out:
reiserfs_write_unlock(inode->i_sb);
return ret;
}
static int reiserfs_readdir(struct file *file, void *dirent, filldir_t filldir)
{
struct dentry *dentry = file->f_path.dentry;
return reiserfs_readdir_dentry(dentry, dirent, filldir, &file->f_pos);
}
/* compose directory item containing "." and ".." entries (entries are
not aligned to 4 byte boundary) */
/* the last four params are LE */
void make_empty_dir_item_v1(char *body, __le32 dirid, __le32 objid,
__le32 par_dirid, __le32 par_objid)
{
struct reiserfs_de_head *deh;
memset(body, 0, EMPTY_DIR_SIZE_V1);
deh = (struct reiserfs_de_head *)body;
/* direntry header of "." */
put_deh_offset(&(deh[0]), DOT_OFFSET);
/* these two are from make_le_item_head, and are are LE */
deh[0].deh_dir_id = dirid;
deh[0].deh_objectid = objid;
deh[0].deh_state = 0; /* Endian safe if 0 */
put_deh_location(&(deh[0]), EMPTY_DIR_SIZE_V1 - strlen("."));
mark_de_visible(&(deh[0]));
/* direntry header of ".." */
put_deh_offset(&(deh[1]), DOT_DOT_OFFSET);
/* key of ".." for the root directory */
/* these two are from the inode, and are are LE */
deh[1].deh_dir_id = par_dirid;
deh[1].deh_objectid = par_objid;
deh[1].deh_state = 0; /* Endian safe if 0 */
put_deh_location(&(deh[1]), deh_location(&(deh[0])) - strlen(".."));
mark_de_visible(&(deh[1]));
/* copy ".." and "." */
memcpy(body + deh_location(&(deh[0])), ".", 1);
memcpy(body + deh_location(&(deh[1])), "..", 2);
}
/* compose directory item containing "." and ".." entries */
void make_empty_dir_item(char *body, __le32 dirid, __le32 objid,
__le32 par_dirid, __le32 par_objid)
{
struct reiserfs_de_head *deh;
memset(body, 0, EMPTY_DIR_SIZE);
deh = (struct reiserfs_de_head *)body;
/* direntry header of "." */
put_deh_offset(&(deh[0]), DOT_OFFSET);
/* these two are from make_le_item_head, and are are LE */
deh[0].deh_dir_id = dirid;
deh[0].deh_objectid = objid;
deh[0].deh_state = 0; /* Endian safe if 0 */
put_deh_location(&(deh[0]), EMPTY_DIR_SIZE - ROUND_UP(strlen(".")));
mark_de_visible(&(deh[0]));
/* direntry header of ".." */
put_deh_offset(&(deh[1]), DOT_DOT_OFFSET);
/* key of ".." for the root directory */
/* these two are from the inode, and are are LE */
deh[1].deh_dir_id = par_dirid;
deh[1].deh_objectid = par_objid;
deh[1].deh_state = 0; /* Endian safe if 0 */
put_deh_location(&(deh[1]),
deh_location(&(deh[0])) - ROUND_UP(strlen("..")));
mark_de_visible(&(deh[1]));
/* copy ".." and "." */
memcpy(body + deh_location(&(deh[0])), ".", 1);
memcpy(body + deh_location(&(deh[1])), "..", 2);
}
Reference in New Issue View Git Blame Copy Permalink