linux-stable/fs/afs/dynroot.c
David Howells e49c7b2f6d afs: Build an abstraction around an "operation" concept
Turn the afs_operation struct into the main way that most fileserver
operations are managed.  Various things are added to the struct, including
the following:

 (1) All the parameters and results of the relevant operations are moved
     into it, removing corresponding fields from the afs_call struct.
     afs_call gets a pointer to the op.

 (2) The target volume is made the main focus of the operation, rather than
     the target vnode(s), and a bunch of op->vnode->volume are made
     op->volume instead.

 (3) Two vnode records are defined (op->file[]) for the vnode(s) involved
     in most operations.  The vnode record (struct afs_vnode_param)
     contains:

	- The vnode pointer.

	- The fid of the vnode to be included in the parameters or that was
          returned in the reply (eg. FS.MakeDir).

	- The status and callback information that may be returned in the
     	  reply about the vnode.

	- Callback break and data version tracking for detecting
          simultaneous third-parth changes.

 (4) Pointers to dentries to be updated with new inodes.

 (5) An operations table pointer.  The table includes pointers to functions
     for issuing AFS and YFS-variant RPCs, handling the success and abort
     of an operation and handling post-I/O-lock local editing of a
     directory.

To make this work, the following function restructuring is made:

 (A) The rotation loop that issues calls to fileservers that can be found
     in each function that wants to issue an RPC (such as afs_mkdir()) is
     extracted out into common code, in a new file called fs_operation.c.

 (B) The rotation loops, such as the one in afs_mkdir(), are replaced with
     a much smaller piece of code that allocates an operation, sets the
     parameters and then calls out to the common code to do the actual
     work.

 (C) The code for handling the success and failure of an operation are
     moved into operation functions (as (5) above) and these are called
     from the core code at appropriate times.

 (D) The pseudo inode getting stuff used by the dynamic root code is moved
     over into dynroot.c.

 (E) struct afs_iget_data is absorbed into the operation struct and
     afs_iget() expects to be given an op pointer and a vnode record.

 (F) Point (E) doesn't work for the root dir of a volume, but we know the
     FID in advance (it's always vnode 1, unique 1), so a separate inode
     getter, afs_root_iget(), is provided to special-case that.

 (G) The inode status init/update functions now also take an op and a vnode
     record.

 (H) The RPC marshalling functions now, for the most part, just take an
     afs_operation struct as their only argument.  All the data they need
     is held there.  The result delivery functions write their answers
     there as well.

 (I) The call is attached to the operation and then the operation core does
     the waiting.

And then the new operation code is, for the moment, made to just initialise
the operation, get the appropriate vnode I/O locks and do the same rotation
loop as before.

This lays the foundation for the following changes in the future:

 (*) Overhauling the rotation (again).

 (*) Support for asynchronous I/O, where the fileserver rotation must be
     done asynchronously also.

Signed-off-by: David Howells <dhowells@redhat.com>
2020-06-04 15:37:17 +01:00

396 lines
8.8 KiB
C

// SPDX-License-Identifier: GPL-2.0-or-later
/* AFS dynamic root handling
*
* Copyright (C) 2018 Red Hat, Inc. All Rights Reserved.
* Written by David Howells (dhowells@redhat.com)
*/
#include <linux/fs.h>
#include <linux/namei.h>
#include <linux/dns_resolver.h>
#include "internal.h"
static atomic_t afs_autocell_ino;
/*
* iget5() comparator for inode created by autocell operations
*
* These pseudo inodes don't match anything.
*/
static int afs_iget5_pseudo_test(struct inode *inode, void *opaque)
{
return 0;
}
/*
* iget5() inode initialiser
*/
static int afs_iget5_pseudo_set(struct inode *inode, void *opaque)
{
struct afs_super_info *as = AFS_FS_S(inode->i_sb);
struct afs_vnode *vnode = AFS_FS_I(inode);
struct afs_fid *fid = opaque;
vnode->volume = as->volume;
vnode->fid = *fid;
inode->i_ino = fid->vnode;
inode->i_generation = fid->unique;
return 0;
}
/*
* Create an inode for a dynamic root directory or an autocell dynamic
* automount dir.
*/
struct inode *afs_iget_pseudo_dir(struct super_block *sb, bool root)
{
struct afs_super_info *as = AFS_FS_S(sb);
struct afs_vnode *vnode;
struct inode *inode;
struct afs_fid fid = {};
_enter("");
if (as->volume)
fid.vid = as->volume->vid;
if (root) {
fid.vnode = 1;
fid.unique = 1;
} else {
fid.vnode = atomic_inc_return(&afs_autocell_ino);
fid.unique = 0;
}
inode = iget5_locked(sb, fid.vnode,
afs_iget5_pseudo_test, afs_iget5_pseudo_set, &fid);
if (!inode) {
_leave(" = -ENOMEM");
return ERR_PTR(-ENOMEM);
}
_debug("GOT INODE %p { ino=%lu, vl=%llx, vn=%llx, u=%x }",
inode, inode->i_ino, fid.vid, fid.vnode, fid.unique);
vnode = AFS_FS_I(inode);
/* there shouldn't be an existing inode */
BUG_ON(!(inode->i_state & I_NEW));
inode->i_size = 0;
inode->i_mode = S_IFDIR | S_IRUGO | S_IXUGO;
if (root) {
inode->i_op = &afs_dynroot_inode_operations;
inode->i_fop = &simple_dir_operations;
} else {
inode->i_op = &afs_autocell_inode_operations;
}
set_nlink(inode, 2);
inode->i_uid = GLOBAL_ROOT_UID;
inode->i_gid = GLOBAL_ROOT_GID;
inode->i_ctime = inode->i_atime = inode->i_mtime = current_time(inode);
inode->i_blocks = 0;
inode->i_generation = 0;
set_bit(AFS_VNODE_PSEUDODIR, &vnode->flags);
if (!root) {
set_bit(AFS_VNODE_MOUNTPOINT, &vnode->flags);
inode->i_flags |= S_AUTOMOUNT;
}
inode->i_flags |= S_NOATIME;
unlock_new_inode(inode);
_leave(" = %p", inode);
return inode;
}
/*
* Probe to see if a cell may exist. This prevents positive dentries from
* being created unnecessarily.
*/
static int afs_probe_cell_name(struct dentry *dentry)
{
struct afs_cell *cell;
struct afs_net *net = afs_d2net(dentry);
const char *name = dentry->d_name.name;
size_t len = dentry->d_name.len;
int ret;
/* Names prefixed with a dot are R/W mounts. */
if (name[0] == '.') {
if (len == 1)
return -EINVAL;
name++;
len--;
}
cell = afs_lookup_cell_rcu(net, name, len);
if (!IS_ERR(cell)) {
afs_put_cell(net, cell);
return 0;
}
ret = dns_query(net->net, "afsdb", name, len, "srv=1",
NULL, NULL, false);
if (ret == -ENODATA)
ret = -EDESTADDRREQ;
return ret;
}
/*
* Try to auto mount the mountpoint with pseudo directory, if the autocell
* operation is setted.
*/
struct inode *afs_try_auto_mntpt(struct dentry *dentry, struct inode *dir)
{
struct afs_vnode *vnode = AFS_FS_I(dir);
struct inode *inode;
int ret = -ENOENT;
_enter("%p{%pd}, {%llx:%llu}",
dentry, dentry, vnode->fid.vid, vnode->fid.vnode);
if (!test_bit(AFS_VNODE_AUTOCELL, &vnode->flags))
goto out;
ret = afs_probe_cell_name(dentry);
if (ret < 0)
goto out;
inode = afs_iget_pseudo_dir(dir->i_sb, false);
if (IS_ERR(inode)) {
ret = PTR_ERR(inode);
goto out;
}
_leave("= %p", inode);
return inode;
out:
_leave("= %d", ret);
return ret == -ENOENT ? NULL : ERR_PTR(ret);
}
/*
* Look up @cell in a dynroot directory. This is a substitution for the
* local cell name for the net namespace.
*/
static struct dentry *afs_lookup_atcell(struct dentry *dentry)
{
struct afs_cell *cell;
struct afs_net *net = afs_d2net(dentry);
struct dentry *ret;
unsigned int seq = 0;
char *name;
int len;
if (!net->ws_cell)
return ERR_PTR(-ENOENT);
ret = ERR_PTR(-ENOMEM);
name = kmalloc(AFS_MAXCELLNAME + 1, GFP_KERNEL);
if (!name)
goto out_p;
rcu_read_lock();
do {
read_seqbegin_or_lock(&net->cells_lock, &seq);
cell = rcu_dereference_raw(net->ws_cell);
if (cell) {
len = cell->name_len;
memcpy(name, cell->name, len + 1);
}
} while (need_seqretry(&net->cells_lock, seq));
done_seqretry(&net->cells_lock, seq);
rcu_read_unlock();
ret = ERR_PTR(-ENOENT);
if (!cell)
goto out_n;
ret = lookup_one_len(name, dentry->d_parent, len);
/* We don't want to d_add() the @cell dentry here as we don't want to
* the cached dentry to hide changes to the local cell name.
*/
out_n:
kfree(name);
out_p:
return ret;
}
/*
* Look up an entry in a dynroot directory.
*/
static struct dentry *afs_dynroot_lookup(struct inode *dir, struct dentry *dentry,
unsigned int flags)
{
_enter("%pd", dentry);
ASSERTCMP(d_inode(dentry), ==, NULL);
if (flags & LOOKUP_CREATE)
return ERR_PTR(-EOPNOTSUPP);
if (dentry->d_name.len >= AFSNAMEMAX) {
_leave(" = -ENAMETOOLONG");
return ERR_PTR(-ENAMETOOLONG);
}
if (dentry->d_name.len == 5 &&
memcmp(dentry->d_name.name, "@cell", 5) == 0)
return afs_lookup_atcell(dentry);
return d_splice_alias(afs_try_auto_mntpt(dentry, dir), dentry);
}
const struct inode_operations afs_dynroot_inode_operations = {
.lookup = afs_dynroot_lookup,
};
/*
* Dirs in the dynamic root don't need revalidation.
*/
static int afs_dynroot_d_revalidate(struct dentry *dentry, unsigned int flags)
{
return 1;
}
/*
* Allow the VFS to enquire as to whether a dentry should be unhashed (mustn't
* sleep)
* - called from dput() when d_count is going to 0.
* - return 1 to request dentry be unhashed, 0 otherwise
*/
static int afs_dynroot_d_delete(const struct dentry *dentry)
{
return d_really_is_positive(dentry);
}
const struct dentry_operations afs_dynroot_dentry_operations = {
.d_revalidate = afs_dynroot_d_revalidate,
.d_delete = afs_dynroot_d_delete,
.d_release = afs_d_release,
.d_automount = afs_d_automount,
};
/*
* Create a manually added cell mount directory.
* - The caller must hold net->proc_cells_lock
*/
int afs_dynroot_mkdir(struct afs_net *net, struct afs_cell *cell)
{
struct super_block *sb = net->dynroot_sb;
struct dentry *root, *subdir;
int ret;
if (!sb || atomic_read(&sb->s_active) == 0)
return 0;
/* Let the ->lookup op do the creation */
root = sb->s_root;
inode_lock(root->d_inode);
subdir = lookup_one_len(cell->name, root, cell->name_len);
if (IS_ERR(subdir)) {
ret = PTR_ERR(subdir);
goto unlock;
}
/* Note that we're retaining an extra ref on the dentry */
subdir->d_fsdata = (void *)1UL;
ret = 0;
unlock:
inode_unlock(root->d_inode);
return ret;
}
/*
* Remove a manually added cell mount directory.
* - The caller must hold net->proc_cells_lock
*/
void afs_dynroot_rmdir(struct afs_net *net, struct afs_cell *cell)
{
struct super_block *sb = net->dynroot_sb;
struct dentry *root, *subdir;
if (!sb || atomic_read(&sb->s_active) == 0)
return;
root = sb->s_root;
inode_lock(root->d_inode);
/* Don't want to trigger a lookup call, which will re-add the cell */
subdir = try_lookup_one_len(cell->name, root, cell->name_len);
if (IS_ERR_OR_NULL(subdir)) {
_debug("lookup %ld", PTR_ERR(subdir));
goto no_dentry;
}
_debug("rmdir %pd %u", subdir, d_count(subdir));
if (subdir->d_fsdata) {
_debug("unpin %u", d_count(subdir));
subdir->d_fsdata = NULL;
dput(subdir);
}
dput(subdir);
no_dentry:
inode_unlock(root->d_inode);
_leave("");
}
/*
* Populate a newly created dynamic root with cell names.
*/
int afs_dynroot_populate(struct super_block *sb)
{
struct afs_cell *cell;
struct afs_net *net = afs_sb2net(sb);
int ret;
mutex_lock(&net->proc_cells_lock);
net->dynroot_sb = sb;
hlist_for_each_entry(cell, &net->proc_cells, proc_link) {
ret = afs_dynroot_mkdir(net, cell);
if (ret < 0)
goto error;
}
ret = 0;
out:
mutex_unlock(&net->proc_cells_lock);
return ret;
error:
net->dynroot_sb = NULL;
goto out;
}
/*
* When a dynamic root that's in the process of being destroyed, depopulate it
* of pinned directories.
*/
void afs_dynroot_depopulate(struct super_block *sb)
{
struct afs_net *net = afs_sb2net(sb);
struct dentry *root = sb->s_root, *subdir, *tmp;
/* Prevent more subdirs from being created */
mutex_lock(&net->proc_cells_lock);
if (net->dynroot_sb == sb)
net->dynroot_sb = NULL;
mutex_unlock(&net->proc_cells_lock);
inode_lock(root->d_inode);
/* Remove all the pins for dirs created for manually added cells */
list_for_each_entry_safe(subdir, tmp, &root->d_subdirs, d_child) {
if (subdir->d_fsdata) {
subdir->d_fsdata = NULL;
dput(subdir);
}
}
inode_unlock(root->d_inode);
}