Files
linux-stable-mirror/fs/afs/dir_silly.c
T
NeilBrownandAl Viro 0e0e490f5d VFS: use wait_var_event for waiting in d_alloc_parallel()
Parallel lookup starts with a call of d_alloc_parallel().  That primitive
either returns a matching hashed dentry or allocates a new one in the
in-lookup state and returns it to the caller.  Once the caller is done
with lookup, it indicates so either by call of d_{splice_alias,add}()
or by call of d_done_lookup(); at that point dentry leaves the in-lookup
state.

If d_alloc_parallel() finds a matching in-lookup dentry, it must wait for
that dentry to leave the in-lookup state, one way or another.  Currently
by supplying wait_queue_head to d_alloc_parallel().  If d_alloc_parallel()
creates a new in-lookup dentry, the address of that wait_queue_head is stored
in ->d_wait of new dentry and stays there while it's in the in-lookup;
subsequent d_alloc_parallel() will wait on the queue found in the matching
in-lookup dentry.  Transition out of in-lookup state wakes waiters on that
queue (if any).

That works, but the calling conventions are inconvenient - the caller must
supply wait_queue_head and make sure that it survives at least until the new
in-lookup dentry leaves the in-lookup state.  That amounts to boilerplate
in the d_alloc_parallel() callers that are followed by a call of d_lookup_done()
in the same function; in cases like nfs asynchronous unlink it gets worse than
that.

This patch changes d_alloc_parallel() to use wake_up_var_locked() to
wake up waiters, and wait_var_event_spinlock() to wait.  dentry->d_lock
is used for synchronisation as it is already held and the relevant
times.

That eliminates the need of caller-supplied wait_queue_head, simplifying
the calling conventions.  Better yet, we only need one bit of information
stored in dentry itself: whether there are any waiters to be woken up,
and that can be easily stored in ->d_flags; ->d_wait goes away.

The reason we need that bit (DCACHE_LOOKUP_WAITERS) is that with wait_var
machinery the queues are shared with all kinds of stuff and there's
no way tell if any of the waiters have anything to do with our dentry;
most of the time none of them will be relevant, so we need to avoid the
pointless wakeups.

Another benefit of the new scheme comes from the fact that wakeups
have to be done outside of write-side critical areas of ->i_dir_seq;
with the old scheme we need to carry the value picked from ->d_wait from
__d_lookup_unhash() to the place where we actually wake the waiters up.
Now we can just leave DCACHE_LOOKUP_WAITERS in ->d_flags until we get
to doing wakeups - that's done within the same ->d_lock scope, so we
are fine; new bit is accessed only under ->d_lock and it's seen only
on dentries with DCACHE_PAR_LOOKUP in ->d_flags.

__d_lookup_unhash() no longer needs to re-init ->d_lru.  That was
previously shared (in a union) with ->d_wait but ->d_wait is now gone
so it no longer corrupts ->d_lru.

Co-developed-by: Al Viro <viro@zeniv.linux.org.uk> # saner handling of flags
Signed-off-by: NeilBrown <neil@brown.name>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2026-06-05 00:34:54 -04:00

290 lines
7.7 KiB
C

// SPDX-License-Identifier: GPL-2.0-or-later
/* AFS silly rename handling
*
* Copyright (C) 2019 Red Hat, Inc. All Rights Reserved.
* Written by David Howells (dhowells@redhat.com)
* - Derived from NFS's sillyrename.
*/
#include <linux/kernel.h>
#include <linux/fs.h>
#include <linux/namei.h>
#include <linux/fsnotify.h>
#include "internal.h"
static void afs_silly_rename_success(struct afs_operation *op)
{
_enter("op=%08x", op->debug_id);
afs_check_dir_conflict(op, &op->file[0]);
afs_vnode_commit_status(op, &op->file[0]);
}
static void afs_silly_rename_edit_dir(struct afs_operation *op)
{
struct afs_vnode_param *dvp = &op->file[0];
struct afs_vnode *dvnode = dvp->vnode;
struct afs_vnode *vnode = AFS_FS_I(d_inode(op->dentry));
struct dentry *old = op->dentry;
struct dentry *new = op->dentry_2;
spin_lock(&old->d_lock);
old->d_flags |= DCACHE_NFSFS_RENAMED;
spin_unlock(&old->d_lock);
if (dvnode->silly_key != op->key) {
key_put(dvnode->silly_key);
dvnode->silly_key = key_get(op->key);
}
down_write(&dvnode->validate_lock);
if (test_bit(AFS_VNODE_DIR_VALID, &dvnode->flags) &&
dvnode->status.data_version == dvp->dv_before + dvp->dv_delta) {
afs_edit_dir_remove(dvnode, &old->d_name,
afs_edit_dir_for_silly_0);
afs_edit_dir_add(dvnode, &new->d_name,
&vnode->fid, afs_edit_dir_for_silly_1);
}
up_write(&dvnode->validate_lock);
}
static const struct afs_operation_ops afs_silly_rename_operation = {
.issue_afs_rpc = afs_fs_rename,
.issue_yfs_rpc = yfs_fs_rename,
.success = afs_silly_rename_success,
.edit_dir = afs_silly_rename_edit_dir,
};
/*
* Actually perform the silly rename step.
*/
static int afs_do_silly_rename(struct afs_vnode *dvnode, struct afs_vnode *vnode,
struct dentry *old, struct dentry *new,
struct key *key)
{
struct afs_operation *op;
_enter("%pd,%pd", old, new);
op = afs_alloc_operation(key, dvnode->volume);
if (IS_ERR(op))
return PTR_ERR(op);
op->more_files = kvzalloc_objs(struct afs_vnode_param, 2);
if (!op->more_files) {
afs_put_operation(op);
return -ENOMEM;
}
afs_op_set_vnode(op, 0, dvnode);
afs_op_set_vnode(op, 1, dvnode);
op->file[0].dv_delta = 1;
op->file[1].dv_delta = 1;
op->file[0].modification = true;
op->file[1].modification = true;
op->file[0].update_ctime = true;
op->file[1].update_ctime = true;
op->more_files[0].vnode = AFS_FS_I(d_inode(old));
op->more_files[0].speculative = true;
op->more_files[1].vnode = AFS_FS_I(d_inode(new));
op->more_files[1].speculative = true;
op->nr_files = 4;
op->dentry = old;
op->dentry_2 = new;
op->ops = &afs_silly_rename_operation;
trace_afs_silly_rename(vnode, false);
return afs_do_sync_operation(op);
}
/*
* Perform silly-rename of a dentry.
*
* AFS is stateless and the server doesn't know when the client is holding a
* file open. To prevent application problems when a file is unlinked while
* it's still open, the client performs a "silly-rename". That is, it renames
* the file to a hidden file in the same directory, and only performs the
* unlink once the last reference to it is put.
*
* The final cleanup is done during dentry_iput.
*/
int afs_sillyrename(struct afs_vnode *dvnode, struct afs_vnode *vnode,
struct dentry *dentry, struct key *key)
{
static unsigned int sillycounter;
struct dentry *sdentry = NULL;
unsigned char silly[16];
int ret = -EBUSY;
_enter("");
/* We don't allow a dentry to be silly-renamed twice. */
if (dentry->d_flags & DCACHE_NFSFS_RENAMED)
return -EBUSY;
sdentry = NULL;
do {
dput(sdentry);
sillycounter++;
/* Create a silly name. Note that the ".__afs" prefix is
* understood by the salvager and must not be changed.
*/
scnprintf(silly, sizeof(silly), ".__afs%04X", sillycounter);
sdentry = lookup_noperm(&QSTR(silly), dentry->d_parent);
/* N.B. Better to return EBUSY here ... it could be dangerous
* to delete the file while it's in use.
*/
if (IS_ERR(sdentry))
goto out;
} while (!d_is_negative(sdentry));
ihold(&vnode->netfs.inode);
ret = afs_do_silly_rename(dvnode, vnode, dentry, sdentry, key);
switch (ret) {
case 0:
/* The rename succeeded. */
set_bit(AFS_VNODE_SILLY_DELETED, &vnode->flags);
d_move(dentry, sdentry);
break;
case -ERESTARTSYS:
/* The result of the rename is unknown. Play it safe by forcing
* a new lookup.
*/
d_drop(dentry);
d_drop(sdentry);
}
iput(&vnode->netfs.inode);
dput(sdentry);
out:
_leave(" = %d", ret);
return ret;
}
static void afs_silly_unlink_success(struct afs_operation *op)
{
_enter("op=%08x", op->debug_id);
afs_check_dir_conflict(op, &op->file[0]);
afs_vnode_commit_status(op, &op->file[0]);
afs_vnode_commit_status(op, &op->file[1]);
afs_update_dentry_version(op, &op->file[0], op->dentry);
}
static void afs_silly_unlink_edit_dir(struct afs_operation *op)
{
struct afs_vnode_param *dvp = &op->file[0];
struct afs_vnode *dvnode = dvp->vnode;
_enter("op=%08x", op->debug_id);
down_write(&dvnode->validate_lock);
if (test_bit(AFS_VNODE_DIR_VALID, &dvnode->flags) &&
dvnode->status.data_version == dvp->dv_before + dvp->dv_delta)
afs_edit_dir_remove(dvnode, &op->dentry->d_name,
afs_edit_dir_for_unlink);
up_write(&dvnode->validate_lock);
}
static const struct afs_operation_ops afs_silly_unlink_operation = {
.issue_afs_rpc = afs_fs_remove_file,
.issue_yfs_rpc = yfs_fs_remove_file,
.success = afs_silly_unlink_success,
.aborted = afs_check_for_remote_deletion,
.edit_dir = afs_silly_unlink_edit_dir,
};
/*
* Tell the server to remove a sillyrename file.
*/
static int afs_do_silly_unlink(struct afs_vnode *dvnode, struct afs_vnode *vnode,
struct dentry *dentry, struct key *key)
{
struct afs_operation *op;
_enter("");
op = afs_alloc_operation(NULL, dvnode->volume);
if (IS_ERR(op))
return PTR_ERR(op);
afs_op_set_vnode(op, 0, dvnode);
afs_op_set_vnode(op, 1, vnode);
op->file[0].dv_delta = 1;
op->file[0].modification = true;
op->file[0].update_ctime = true;
op->file[1].op_unlinked = true;
op->file[1].update_ctime = true;
op->dentry = dentry;
op->ops = &afs_silly_unlink_operation;
trace_afs_silly_rename(vnode, true);
afs_begin_vnode_operation(op);
afs_wait_for_operation(op);
/* If there was a conflict with a third party, check the status of the
* unlinked vnode.
*/
if (op->cumul_error.error == 0 && (op->flags & AFS_OPERATION_DIR_CONFLICT)) {
op->file[1].update_ctime = false;
op->fetch_status.which = 1;
op->ops = &afs_fetch_status_operation;
afs_begin_vnode_operation(op);
afs_wait_for_operation(op);
}
return afs_put_operation(op);
}
/*
* Remove sillyrename file on iput.
*/
int afs_silly_iput(struct dentry *dentry, struct inode *inode)
{
struct afs_vnode *dvnode = AFS_FS_I(d_inode(dentry->d_parent));
struct afs_vnode *vnode = AFS_FS_I(inode);
struct dentry *alias;
int ret;
_enter("%p{%pd},%llx", dentry, dentry, vnode->fid.vnode);
down_read(&dvnode->rmdir_lock);
alias = d_alloc_parallel(dentry->d_parent, &dentry->d_name);
if (IS_ERR(alias)) {
up_read(&dvnode->rmdir_lock);
return 0;
}
if (!d_in_lookup(alias)) {
/* We raced with lookup... See if we need to transfer the
* sillyrename information to the aliased dentry.
*/
ret = 0;
spin_lock(&alias->d_lock);
if (d_really_is_positive(alias) &&
!(alias->d_flags & DCACHE_NFSFS_RENAMED)) {
alias->d_flags |= DCACHE_NFSFS_RENAMED;
ret = 1;
}
spin_unlock(&alias->d_lock);
up_read(&dvnode->rmdir_lock);
dput(alias);
return ret;
}
/* Stop lock-release from complaining. */
spin_lock(&vnode->lock);
vnode->lock_state = AFS_VNODE_LOCK_DELETED;
trace_afs_flock_ev(vnode, NULL, afs_flock_silly_delete, 0);
spin_unlock(&vnode->lock);
afs_do_silly_unlink(dvnode, vnode, dentry, dvnode->silly_key);
up_read(&dvnode->rmdir_lock);
d_lookup_done(alias);
dput(alias);
return 1;
}