linux-stable/fs/fs_context.c

700 lines
17 KiB
C
Raw Normal View History

// SPDX-License-Identifier: GPL-2.0-or-later
/* Provide a way to create a superblock configuration context within the kernel
* that allows a superblock to be set up prior to mounting.
*
* Copyright (C) 2017 Red Hat, Inc. All Rights Reserved.
* Written by David Howells (dhowells@redhat.com)
*/
#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
#include <linux/module.h>
#include <linux/fs_context.h>
vfs: Implement a filesystem superblock creation/configuration context [AV - unfuck kern_mount_data(); we want non-NULL ->mnt_ns on long-living mounts] [AV - reordering fs/namespace.c is badly overdue, but let's keep it separate from that series] [AV - drop simple_pin_fs() change] [AV - clean vfs_kern_mount() failure exits up] Implement a filesystem context concept to be used during superblock creation for mount and superblock reconfiguration for remount. The mounting procedure then becomes: (1) Allocate new fs_context context. (2) Configure the context. (3) Create superblock. (4) Query the superblock. (5) Create a mount for the superblock. (6) Destroy the context. Rather than calling fs_type->mount(), an fs_context struct is created and fs_type->init_fs_context() is called to set it up. Pointers exist for the filesystem and LSM to hang their private data off. A set of operations has to be set by ->init_fs_context() to provide freeing, duplication, option parsing, binary data parsing, validation, mounting and superblock filling. Legacy filesystems are supported by the provision of a set of legacy fs_context operations that build up a list of mount options and then invoke fs_type->mount() from within the fs_context ->get_tree() operation. This allows all filesystems to be accessed using fs_context. It should be noted that, whilst this patch adds a lot of lines of code, there is quite a bit of duplication with existing code that can be eliminated should all filesystems be converted over. Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2018-11-01 23:07:25 +00:00
#include <linux/fs_parser.h>
#include <linux/fs.h>
#include <linux/mount.h>
#include <linux/nsproxy.h>
#include <linux/slab.h>
#include <linux/magic.h>
#include <linux/security.h>
#include <linux/mnt_namespace.h>
#include <linux/pid_namespace.h>
#include <linux/user_namespace.h>
#include <net/net_namespace.h>
#include <asm/sections.h>
#include "mount.h"
#include "internal.h"
vfs: Implement a filesystem superblock creation/configuration context [AV - unfuck kern_mount_data(); we want non-NULL ->mnt_ns on long-living mounts] [AV - reordering fs/namespace.c is badly overdue, but let's keep it separate from that series] [AV - drop simple_pin_fs() change] [AV - clean vfs_kern_mount() failure exits up] Implement a filesystem context concept to be used during superblock creation for mount and superblock reconfiguration for remount. The mounting procedure then becomes: (1) Allocate new fs_context context. (2) Configure the context. (3) Create superblock. (4) Query the superblock. (5) Create a mount for the superblock. (6) Destroy the context. Rather than calling fs_type->mount(), an fs_context struct is created and fs_type->init_fs_context() is called to set it up. Pointers exist for the filesystem and LSM to hang their private data off. A set of operations has to be set by ->init_fs_context() to provide freeing, duplication, option parsing, binary data parsing, validation, mounting and superblock filling. Legacy filesystems are supported by the provision of a set of legacy fs_context operations that build up a list of mount options and then invoke fs_type->mount() from within the fs_context ->get_tree() operation. This allows all filesystems to be accessed using fs_context. It should be noted that, whilst this patch adds a lot of lines of code, there is quite a bit of duplication with existing code that can be eliminated should all filesystems be converted over. Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2018-11-01 23:07:25 +00:00
enum legacy_fs_param {
LEGACY_FS_UNSET_PARAMS,
LEGACY_FS_MONOLITHIC_PARAMS,
LEGACY_FS_INDIVIDUAL_PARAMS,
};
struct legacy_fs_context {
char *legacy_data; /* Data page for legacy filesystems */
size_t data_size;
vfs: Implement a filesystem superblock creation/configuration context [AV - unfuck kern_mount_data(); we want non-NULL ->mnt_ns on long-living mounts] [AV - reordering fs/namespace.c is badly overdue, but let's keep it separate from that series] [AV - drop simple_pin_fs() change] [AV - clean vfs_kern_mount() failure exits up] Implement a filesystem context concept to be used during superblock creation for mount and superblock reconfiguration for remount. The mounting procedure then becomes: (1) Allocate new fs_context context. (2) Configure the context. (3) Create superblock. (4) Query the superblock. (5) Create a mount for the superblock. (6) Destroy the context. Rather than calling fs_type->mount(), an fs_context struct is created and fs_type->init_fs_context() is called to set it up. Pointers exist for the filesystem and LSM to hang their private data off. A set of operations has to be set by ->init_fs_context() to provide freeing, duplication, option parsing, binary data parsing, validation, mounting and superblock filling. Legacy filesystems are supported by the provision of a set of legacy fs_context operations that build up a list of mount options and then invoke fs_type->mount() from within the fs_context ->get_tree() operation. This allows all filesystems to be accessed using fs_context. It should be noted that, whilst this patch adds a lot of lines of code, there is quite a bit of duplication with existing code that can be eliminated should all filesystems be converted over. Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2018-11-01 23:07:25 +00:00
enum legacy_fs_param param_type;
};
static int legacy_init_fs_context(struct fs_context *fc);
vfs: Implement a filesystem superblock creation/configuration context [AV - unfuck kern_mount_data(); we want non-NULL ->mnt_ns on long-living mounts] [AV - reordering fs/namespace.c is badly overdue, but let's keep it separate from that series] [AV - drop simple_pin_fs() change] [AV - clean vfs_kern_mount() failure exits up] Implement a filesystem context concept to be used during superblock creation for mount and superblock reconfiguration for remount. The mounting procedure then becomes: (1) Allocate new fs_context context. (2) Configure the context. (3) Create superblock. (4) Query the superblock. (5) Create a mount for the superblock. (6) Destroy the context. Rather than calling fs_type->mount(), an fs_context struct is created and fs_type->init_fs_context() is called to set it up. Pointers exist for the filesystem and LSM to hang their private data off. A set of operations has to be set by ->init_fs_context() to provide freeing, duplication, option parsing, binary data parsing, validation, mounting and superblock filling. Legacy filesystems are supported by the provision of a set of legacy fs_context operations that build up a list of mount options and then invoke fs_type->mount() from within the fs_context ->get_tree() operation. This allows all filesystems to be accessed using fs_context. It should be noted that, whilst this patch adds a lot of lines of code, there is quite a bit of duplication with existing code that can be eliminated should all filesystems be converted over. Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2018-11-01 23:07:25 +00:00
static const struct constant_table common_set_sb_flag[] = {
{ "dirsync", SB_DIRSYNC },
{ "lazytime", SB_LAZYTIME },
{ "mand", SB_MANDLOCK },
{ "ro", SB_RDONLY },
{ "sync", SB_SYNCHRONOUS },
{ },
vfs: Implement a filesystem superblock creation/configuration context [AV - unfuck kern_mount_data(); we want non-NULL ->mnt_ns on long-living mounts] [AV - reordering fs/namespace.c is badly overdue, but let's keep it separate from that series] [AV - drop simple_pin_fs() change] [AV - clean vfs_kern_mount() failure exits up] Implement a filesystem context concept to be used during superblock creation for mount and superblock reconfiguration for remount. The mounting procedure then becomes: (1) Allocate new fs_context context. (2) Configure the context. (3) Create superblock. (4) Query the superblock. (5) Create a mount for the superblock. (6) Destroy the context. Rather than calling fs_type->mount(), an fs_context struct is created and fs_type->init_fs_context() is called to set it up. Pointers exist for the filesystem and LSM to hang their private data off. A set of operations has to be set by ->init_fs_context() to provide freeing, duplication, option parsing, binary data parsing, validation, mounting and superblock filling. Legacy filesystems are supported by the provision of a set of legacy fs_context operations that build up a list of mount options and then invoke fs_type->mount() from within the fs_context ->get_tree() operation. This allows all filesystems to be accessed using fs_context. It should be noted that, whilst this patch adds a lot of lines of code, there is quite a bit of duplication with existing code that can be eliminated should all filesystems be converted over. Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2018-11-01 23:07:25 +00:00
};
static const struct constant_table common_clear_sb_flag[] = {
{ "async", SB_SYNCHRONOUS },
{ "nolazytime", SB_LAZYTIME },
{ "nomand", SB_MANDLOCK },
{ "rw", SB_RDONLY },
{ },
vfs: Implement a filesystem superblock creation/configuration context [AV - unfuck kern_mount_data(); we want non-NULL ->mnt_ns on long-living mounts] [AV - reordering fs/namespace.c is badly overdue, but let's keep it separate from that series] [AV - drop simple_pin_fs() change] [AV - clean vfs_kern_mount() failure exits up] Implement a filesystem context concept to be used during superblock creation for mount and superblock reconfiguration for remount. The mounting procedure then becomes: (1) Allocate new fs_context context. (2) Configure the context. (3) Create superblock. (4) Query the superblock. (5) Create a mount for the superblock. (6) Destroy the context. Rather than calling fs_type->mount(), an fs_context struct is created and fs_type->init_fs_context() is called to set it up. Pointers exist for the filesystem and LSM to hang their private data off. A set of operations has to be set by ->init_fs_context() to provide freeing, duplication, option parsing, binary data parsing, validation, mounting and superblock filling. Legacy filesystems are supported by the provision of a set of legacy fs_context operations that build up a list of mount options and then invoke fs_type->mount() from within the fs_context ->get_tree() operation. This allows all filesystems to be accessed using fs_context. It should be noted that, whilst this patch adds a lot of lines of code, there is quite a bit of duplication with existing code that can be eliminated should all filesystems be converted over. Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2018-11-01 23:07:25 +00:00
};
/*
* Check for a common mount option that manipulates s_flags.
*/
static int vfs_parse_sb_flag(struct fs_context *fc, const char *key)
{
unsigned int token;
token = lookup_constant(common_set_sb_flag, key, 0);
if (token) {
fc->sb_flags |= token;
fc->sb_flags_mask |= token;
return 0;
}
token = lookup_constant(common_clear_sb_flag, key, 0);
if (token) {
fc->sb_flags &= ~token;
fc->sb_flags_mask |= token;
return 0;
}
return -ENOPARAM;
}
/**
* vfs_parse_fs_param - Add a single parameter to a superblock config
* @fc: The filesystem context to modify
* @param: The parameter
*
* A single mount option in string form is applied to the filesystem context
* being set up. Certain standard options (for example "ro") are translated
* into flag bits without going to the filesystem. The active security module
* is allowed to observe and poach options. Any other options are passed over
* to the filesystem to parse.
*
* This may be called multiple times for a context.
*
* Returns 0 on success and a negative error code on failure. In the event of
* failure, supplementary error information may have been set.
*/
int vfs_parse_fs_param(struct fs_context *fc, struct fs_parameter *param)
{
int ret;
if (!param->key)
return invalf(fc, "Unnamed parameter\n");
ret = vfs_parse_sb_flag(fc, param->key);
if (ret != -ENOPARAM)
return ret;
ret = security_fs_context_parse_param(fc, param);
if (ret != -ENOPARAM)
/* Param belongs to the LSM or is disallowed by the LSM; so
* don't pass to the FS.
*/
return ret;
if (fc->ops->parse_param) {
ret = fc->ops->parse_param(fc, param);
if (ret != -ENOPARAM)
return ret;
}
/* If the filesystem doesn't take any arguments, give it the
* default handling of source.
*/
if (strcmp(param->key, "source") == 0) {
if (param->type != fs_value_is_string)
return invalf(fc, "VFS: Non-string source");
if (fc->source)
return invalf(fc, "VFS: Multiple sources");
fc->source = param->string;
param->string = NULL;
return 0;
}
return invalf(fc, "%s: Unknown parameter '%s'",
fc->fs_type->name, param->key);
}
EXPORT_SYMBOL(vfs_parse_fs_param);
/**
* vfs_parse_fs_string - Convenience function to just parse a string.
*/
int vfs_parse_fs_string(struct fs_context *fc, const char *key,
const char *value, size_t v_size)
{
int ret;
struct fs_parameter param = {
.key = key,
.type = fs_value_is_flag,
vfs: Implement a filesystem superblock creation/configuration context [AV - unfuck kern_mount_data(); we want non-NULL ->mnt_ns on long-living mounts] [AV - reordering fs/namespace.c is badly overdue, but let's keep it separate from that series] [AV - drop simple_pin_fs() change] [AV - clean vfs_kern_mount() failure exits up] Implement a filesystem context concept to be used during superblock creation for mount and superblock reconfiguration for remount. The mounting procedure then becomes: (1) Allocate new fs_context context. (2) Configure the context. (3) Create superblock. (4) Query the superblock. (5) Create a mount for the superblock. (6) Destroy the context. Rather than calling fs_type->mount(), an fs_context struct is created and fs_type->init_fs_context() is called to set it up. Pointers exist for the filesystem and LSM to hang their private data off. A set of operations has to be set by ->init_fs_context() to provide freeing, duplication, option parsing, binary data parsing, validation, mounting and superblock filling. Legacy filesystems are supported by the provision of a set of legacy fs_context operations that build up a list of mount options and then invoke fs_type->mount() from within the fs_context ->get_tree() operation. This allows all filesystems to be accessed using fs_context. It should be noted that, whilst this patch adds a lot of lines of code, there is quite a bit of duplication with existing code that can be eliminated should all filesystems be converted over. Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2018-11-01 23:07:25 +00:00
.size = v_size,
};
if (value) {
vfs: Implement a filesystem superblock creation/configuration context [AV - unfuck kern_mount_data(); we want non-NULL ->mnt_ns on long-living mounts] [AV - reordering fs/namespace.c is badly overdue, but let's keep it separate from that series] [AV - drop simple_pin_fs() change] [AV - clean vfs_kern_mount() failure exits up] Implement a filesystem context concept to be used during superblock creation for mount and superblock reconfiguration for remount. The mounting procedure then becomes: (1) Allocate new fs_context context. (2) Configure the context. (3) Create superblock. (4) Query the superblock. (5) Create a mount for the superblock. (6) Destroy the context. Rather than calling fs_type->mount(), an fs_context struct is created and fs_type->init_fs_context() is called to set it up. Pointers exist for the filesystem and LSM to hang their private data off. A set of operations has to be set by ->init_fs_context() to provide freeing, duplication, option parsing, binary data parsing, validation, mounting and superblock filling. Legacy filesystems are supported by the provision of a set of legacy fs_context operations that build up a list of mount options and then invoke fs_type->mount() from within the fs_context ->get_tree() operation. This allows all filesystems to be accessed using fs_context. It should be noted that, whilst this patch adds a lot of lines of code, there is quite a bit of duplication with existing code that can be eliminated should all filesystems be converted over. Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2018-11-01 23:07:25 +00:00
param.string = kmemdup_nul(value, v_size, GFP_KERNEL);
if (!param.string)
return -ENOMEM;
param.type = fs_value_is_string;
vfs: Implement a filesystem superblock creation/configuration context [AV - unfuck kern_mount_data(); we want non-NULL ->mnt_ns on long-living mounts] [AV - reordering fs/namespace.c is badly overdue, but let's keep it separate from that series] [AV - drop simple_pin_fs() change] [AV - clean vfs_kern_mount() failure exits up] Implement a filesystem context concept to be used during superblock creation for mount and superblock reconfiguration for remount. The mounting procedure then becomes: (1) Allocate new fs_context context. (2) Configure the context. (3) Create superblock. (4) Query the superblock. (5) Create a mount for the superblock. (6) Destroy the context. Rather than calling fs_type->mount(), an fs_context struct is created and fs_type->init_fs_context() is called to set it up. Pointers exist for the filesystem and LSM to hang their private data off. A set of operations has to be set by ->init_fs_context() to provide freeing, duplication, option parsing, binary data parsing, validation, mounting and superblock filling. Legacy filesystems are supported by the provision of a set of legacy fs_context operations that build up a list of mount options and then invoke fs_type->mount() from within the fs_context ->get_tree() operation. This allows all filesystems to be accessed using fs_context. It should be noted that, whilst this patch adds a lot of lines of code, there is quite a bit of duplication with existing code that can be eliminated should all filesystems be converted over. Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2018-11-01 23:07:25 +00:00
}
ret = vfs_parse_fs_param(fc, &param);
kfree(param.string);
return ret;
}
EXPORT_SYMBOL(vfs_parse_fs_string);
/**
* generic_parse_monolithic - Parse key[=val][,key[=val]]* mount data
* @ctx: The superblock configuration to fill in.
* @data: The data to parse
*
* Parse a blob of data that's in key[=val][,key[=val]]* form. This can be
* called from the ->monolithic_mount_data() fs_context operation.
*
* Returns 0 on success or the error returned by the ->parse_option() fs_context
* operation on failure.
*/
int generic_parse_monolithic(struct fs_context *fc, void *data)
{
char *options = data, *key;
int ret = 0;
if (!options)
return 0;
ret = security_sb_eat_lsm_opts(options, &fc->security);
if (ret)
return ret;
while ((key = strsep(&options, ",")) != NULL) {
if (*key) {
size_t v_len = 0;
char *value = strchr(key, '=');
if (value) {
if (value == key)
continue;
*value++ = 0;
v_len = strlen(value);
}
ret = vfs_parse_fs_string(fc, key, value, v_len);
if (ret < 0)
break;
}
}
return ret;
}
EXPORT_SYMBOL(generic_parse_monolithic);
/**
* alloc_fs_context - Create a filesystem context.
* @fs_type: The filesystem type.
* @reference: The dentry from which this one derives (or NULL)
* @sb_flags: Filesystem/superblock flags (SB_*)
* @sb_flags_mask: Applicable members of @sb_flags
* @purpose: The purpose that this configuration shall be used for.
*
* Open a filesystem and create a mount context. The mount context is
* initialised with the supplied flags and, if a submount/automount from
* another superblock (referred to by @reference) is supplied, may have
* parameters such as namespaces copied across from that superblock.
*/
static struct fs_context *alloc_fs_context(struct file_system_type *fs_type,
struct dentry *reference,
unsigned int sb_flags,
unsigned int sb_flags_mask,
enum fs_context_purpose purpose)
{
int (*init_fs_context)(struct fs_context *);
struct fs_context *fc;
int ret = -ENOMEM;
fc = kzalloc(sizeof(struct fs_context), GFP_KERNEL);
if (!fc)
return ERR_PTR(-ENOMEM);
fc->purpose = purpose;
fc->sb_flags = sb_flags;
fc->sb_flags_mask = sb_flags_mask;
fc->fs_type = get_filesystem(fs_type);
fc->cred = get_current_cred();
fc->net_ns = get_net(current->nsproxy->net_ns);
fc->log.prefix = fs_type->name;
vfs: syscall: Add fsopen() to prepare for superblock creation Provide an fsopen() system call that starts the process of preparing to create a superblock that will then be mountable, using an fd as a context handle. fsopen() is given the name of the filesystem that will be used: int mfd = fsopen(const char *fsname, unsigned int flags); where flags can be 0 or FSOPEN_CLOEXEC. For example: sfd = fsopen("ext4", FSOPEN_CLOEXEC); fsconfig(sfd, FSCONFIG_SET_PATH, "source", "/dev/sda1", AT_FDCWD); fsconfig(sfd, FSCONFIG_SET_FLAG, "noatime", NULL, 0); fsconfig(sfd, FSCONFIG_SET_FLAG, "acl", NULL, 0); fsconfig(sfd, FSCONFIG_SET_FLAG, "user_xattr", NULL, 0); fsconfig(sfd, FSCONFIG_SET_STRING, "sb", "1", 0); fsconfig(sfd, FSCONFIG_CMD_CREATE, NULL, NULL, 0); fsinfo(sfd, NULL, ...); // query new superblock attributes mfd = fsmount(sfd, FSMOUNT_CLOEXEC, MS_RELATIME); move_mount(mfd, "", sfd, AT_FDCWD, "/mnt", MOVE_MOUNT_F_EMPTY_PATH); sfd = fsopen("afs", -1); fsconfig(fd, FSCONFIG_SET_STRING, "source", "#grand.central.org:root.cell", 0); fsconfig(fd, FSCONFIG_CMD_CREATE, NULL, NULL, 0); mfd = fsmount(sfd, 0, MS_NODEV); move_mount(mfd, "", sfd, AT_FDCWD, "/mnt", MOVE_MOUNT_F_EMPTY_PATH); If an error is reported at any step, an error message may be available to be read() back (ENODATA will be reported if there isn't an error available) in the form: "e <subsys>:<problem>" "e SELinux:Mount on mountpoint not permitted" Once fsmount() has been called, further fsconfig() calls will incur EBUSY, even if the fsmount() fails. read() is still possible to retrieve error information. The fsopen() syscall creates a mount context and hangs it of the fd that it returns. Netlink is not used because it is optional and would make the core VFS dependent on the networking layer and also potentially add network namespace issues. Note that, for the moment, the caller must have SYS_CAP_ADMIN to use fsopen(). Signed-off-by: David Howells <dhowells@redhat.com> cc: linux-api@vger.kernel.org Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2018-11-01 23:33:31 +00:00
mutex_init(&fc->uapi_mutex);
switch (purpose) {
case FS_CONTEXT_FOR_MOUNT:
fc->user_ns = get_user_ns(fc->cred->user_ns);
break;
case FS_CONTEXT_FOR_SUBMOUNT:
fc->user_ns = get_user_ns(reference->d_sb->s_user_ns);
break;
case FS_CONTEXT_FOR_RECONFIGURE:
atomic_inc(&reference->d_sb->s_active);
vfs: set fs_context::user_ns for reconfigure fs_context::user_ns is used by fuse_parse_param(), even during remount, so it needs to be set to the existing value for reconfigure. Reproducer: #include <fcntl.h> #include <sys/mount.h> int main() { char opts[128]; int fd = open("/dev/fuse", O_RDWR); sprintf(opts, "fd=%d,rootmode=040000,user_id=0,group_id=0", fd); mkdir("mnt", 0777); mount("foo", "mnt", "fuse.foo", 0, opts); mount("foo", "mnt", "fuse.foo", MS_REMOUNT, opts); } Crash: BUG: kernel NULL pointer dereference, address: 0000000000000000 #PF: supervisor read access in kernel mode #PF: error_code(0x0000) - not-present page PGD 0 P4D 0 Oops: 0000 [#1] SMP CPU: 0 PID: 129 Comm: syz_make_kuid Not tainted 5.3.0-rc5-next-20190821 #3 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.12.0-20181126_142135-anatol 04/01/2014 RIP: 0010:map_id_range_down+0xb/0xc0 kernel/user_namespace.c:291 [...] Call Trace: map_id_down kernel/user_namespace.c:312 [inline] make_kuid+0xe/0x10 kernel/user_namespace.c:389 fuse_parse_param+0x116/0x210 fs/fuse/inode.c:523 vfs_parse_fs_param+0xdb/0x1b0 fs/fs_context.c:145 vfs_parse_fs_string+0x6a/0xa0 fs/fs_context.c:188 generic_parse_monolithic+0x85/0xc0 fs/fs_context.c:228 parse_monolithic_mount_data+0x1b/0x20 fs/fs_context.c:708 do_remount fs/namespace.c:2525 [inline] do_mount+0x39a/0xa60 fs/namespace.c:3107 ksys_mount+0x7d/0xd0 fs/namespace.c:3325 __do_sys_mount fs/namespace.c:3339 [inline] __se_sys_mount fs/namespace.c:3336 [inline] __x64_sys_mount+0x20/0x30 fs/namespace.c:3336 do_syscall_64+0x4a/0x1a0 arch/x86/entry/common.c:290 entry_SYSCALL_64_after_hwframe+0x49/0xbe Reported-by: syzbot+7d6a57304857423318a5@syzkaller.appspotmail.com Fixes: 408cbe695350 ("vfs: Convert fuse to use the new mount API") Cc: David Howells <dhowells@redhat.com> Cc: Miklos Szeredi <miklos@szeredi.hu> Signed-off-by: Eric Biggers <ebiggers@google.com> Reviewed-by: David Howells <dhowells@redhat.com> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2019-08-22 05:16:33 +00:00
fc->user_ns = get_user_ns(reference->d_sb->s_user_ns);
fc->root = dget(reference);
break;
}
/* TODO: Make all filesystems support this unconditionally */
init_fs_context = fc->fs_type->init_fs_context;
if (!init_fs_context)
init_fs_context = legacy_init_fs_context;
ret = init_fs_context(fc);
if (ret < 0)
goto err_fc;
fc->need_free = true;
return fc;
err_fc:
put_fs_context(fc);
return ERR_PTR(ret);
}
struct fs_context *fs_context_for_mount(struct file_system_type *fs_type,
unsigned int sb_flags)
{
return alloc_fs_context(fs_type, NULL, sb_flags, 0,
FS_CONTEXT_FOR_MOUNT);
}
EXPORT_SYMBOL(fs_context_for_mount);
struct fs_context *fs_context_for_reconfigure(struct dentry *dentry,
unsigned int sb_flags,
unsigned int sb_flags_mask)
{
return alloc_fs_context(dentry->d_sb->s_type, dentry, sb_flags,
sb_flags_mask, FS_CONTEXT_FOR_RECONFIGURE);
}
EXPORT_SYMBOL(fs_context_for_reconfigure);
struct fs_context *fs_context_for_submount(struct file_system_type *type,
struct dentry *reference)
{
return alloc_fs_context(type, reference, 0, 0, FS_CONTEXT_FOR_SUBMOUNT);
}
EXPORT_SYMBOL(fs_context_for_submount);
vfs_get_tree(): evict the call of security_sb_kern_mount() Right now vfs_get_tree() calls security_sb_kern_mount() (i.e. mount MAC) unless it gets MS_KERNMOUNT or MS_SUBMOUNT in flags. Doing it that way is both clumsy and imprecise. Consider the callers' tree of vfs_get_tree(): vfs_get_tree() <- do_new_mount() <- vfs_kern_mount() <- simple_pin_fs() <- vfs_submount() <- kern_mount_data() <- init_mount_tree() <- btrfs_mount() <- vfs_get_tree() <- nfs_do_root_mount() <- nfs4_try_mount() <- nfs_fs_mount() <- vfs_get_tree() <- nfs4_referral_mount() do_new_mount() always does need MAC (we are guaranteed that neither MS_KERNMOUNT nor MS_SUBMOUNT will be passed there). simple_pin_fs(), vfs_submount() and kern_mount_data() pass explicit flags inhibiting that check. So does nfs4_referral_mount() (the flags there are ulimately coming from vfs_submount()). init_mount_tree() is called too early for anything LSM-related; it doesn't matter whether we attempt those checks, they'll do nothing. Finally, in case of btrfs_mount() and nfs_fs_mount(), doing MAC is pointless - either the caller will do it, or the flags are such that we wouldn't have done it either. In other words, the one and only case when we want that check done is when we are called from do_new_mount(), and there we want it unconditionally. So let's simply move it there. The superblock is still locked, so nobody is going to get access to it (via ustat(2), etc.) until we get a chance to apply the checks - we are free to move them to any point up to where we drop ->s_umount (in do_new_mount_fc()). Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2018-12-20 20:04:50 +00:00
void fc_drop_locked(struct fs_context *fc)
{
struct super_block *sb = fc->root->d_sb;
dput(fc->root);
fc->root = NULL;
deactivate_locked_super(sb);
}
static void legacy_fs_context_free(struct fs_context *fc);
/**
* vfs_dup_fc_config: Duplicate a filesystem context.
* @src_fc: The context to copy.
*/
struct fs_context *vfs_dup_fs_context(struct fs_context *src_fc)
{
struct fs_context *fc;
int ret;
if (!src_fc->ops->dup)
return ERR_PTR(-EOPNOTSUPP);
fc = kmemdup(src_fc, sizeof(struct fs_context), GFP_KERNEL);
if (!fc)
return ERR_PTR(-ENOMEM);
vfs: syscall: Add fsopen() to prepare for superblock creation Provide an fsopen() system call that starts the process of preparing to create a superblock that will then be mountable, using an fd as a context handle. fsopen() is given the name of the filesystem that will be used: int mfd = fsopen(const char *fsname, unsigned int flags); where flags can be 0 or FSOPEN_CLOEXEC. For example: sfd = fsopen("ext4", FSOPEN_CLOEXEC); fsconfig(sfd, FSCONFIG_SET_PATH, "source", "/dev/sda1", AT_FDCWD); fsconfig(sfd, FSCONFIG_SET_FLAG, "noatime", NULL, 0); fsconfig(sfd, FSCONFIG_SET_FLAG, "acl", NULL, 0); fsconfig(sfd, FSCONFIG_SET_FLAG, "user_xattr", NULL, 0); fsconfig(sfd, FSCONFIG_SET_STRING, "sb", "1", 0); fsconfig(sfd, FSCONFIG_CMD_CREATE, NULL, NULL, 0); fsinfo(sfd, NULL, ...); // query new superblock attributes mfd = fsmount(sfd, FSMOUNT_CLOEXEC, MS_RELATIME); move_mount(mfd, "", sfd, AT_FDCWD, "/mnt", MOVE_MOUNT_F_EMPTY_PATH); sfd = fsopen("afs", -1); fsconfig(fd, FSCONFIG_SET_STRING, "source", "#grand.central.org:root.cell", 0); fsconfig(fd, FSCONFIG_CMD_CREATE, NULL, NULL, 0); mfd = fsmount(sfd, 0, MS_NODEV); move_mount(mfd, "", sfd, AT_FDCWD, "/mnt", MOVE_MOUNT_F_EMPTY_PATH); If an error is reported at any step, an error message may be available to be read() back (ENODATA will be reported if there isn't an error available) in the form: "e <subsys>:<problem>" "e SELinux:Mount on mountpoint not permitted" Once fsmount() has been called, further fsconfig() calls will incur EBUSY, even if the fsmount() fails. read() is still possible to retrieve error information. The fsopen() syscall creates a mount context and hangs it of the fd that it returns. Netlink is not used because it is optional and would make the core VFS dependent on the networking layer and also potentially add network namespace issues. Note that, for the moment, the caller must have SYS_CAP_ADMIN to use fsopen(). Signed-off-by: David Howells <dhowells@redhat.com> cc: linux-api@vger.kernel.org Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2018-11-01 23:33:31 +00:00
mutex_init(&fc->uapi_mutex);
fc->fs_private = NULL;
fc->s_fs_info = NULL;
fc->source = NULL;
fc->security = NULL;
get_filesystem(fc->fs_type);
get_net(fc->net_ns);
get_user_ns(fc->user_ns);
get_cred(fc->cred);
if (fc->log.log)
refcount_inc(&fc->log.log->usage);
/* Can't call put until we've called ->dup */
ret = fc->ops->dup(fc, src_fc);
if (ret < 0)
goto err_fc;
ret = security_fs_context_dup(fc, src_fc);
if (ret < 0)
goto err_fc;
return fc;
err_fc:
put_fs_context(fc);
return ERR_PTR(ret);
}
EXPORT_SYMBOL(vfs_dup_fs_context);
/**
* logfc - Log a message to a filesystem context
* @fc: The filesystem context to log to.
* @fmt: The format of the buffer.
*/
void logfc(struct fc_log *log, const char *prefix, char level, const char *fmt, ...)
{
va_list va;
struct va_format vaf = {.fmt = fmt, .va = &va};
va_start(va, fmt);
if (!log) {
switch (level) {
case 'w':
printk(KERN_WARNING "%s%s%pV\n", prefix ? prefix : "",
prefix ? ": " : "", &vaf);
break;
case 'e':
printk(KERN_ERR "%s%s%pV\n", prefix ? prefix : "",
prefix ? ": " : "", &vaf);
break;
default:
printk(KERN_NOTICE "%s%s%pV\n", prefix ? prefix : "",
prefix ? ": " : "", &vaf);
break;
}
} else {
unsigned int logsize = ARRAY_SIZE(log->buffer);
u8 index;
char *q = kasprintf(GFP_KERNEL, "%c %s%s%pV\n", level,
prefix ? prefix : "",
prefix ? ": " : "", &vaf);
index = log->head & (logsize - 1);
BUILD_BUG_ON(sizeof(log->head) != sizeof(u8) ||
sizeof(log->tail) != sizeof(u8));
if ((u8)(log->head - log->tail) == logsize) {
/* The buffer is full, discard the oldest message */
if (log->need_free & (1 << index))
kfree(log->buffer[index]);
log->tail++;
}
log->buffer[index] = q ? q : "OOM: Can't store error string";
if (q)
log->need_free |= 1 << index;
else
log->need_free &= ~(1 << index);
log->head++;
}
va_end(va);
}
EXPORT_SYMBOL(logfc);
/*
* Free a logging structure.
*/
static void put_fc_log(struct fs_context *fc)
{
struct fc_log *log = fc->log.log;
int i;
if (log) {
if (refcount_dec_and_test(&log->usage)) {
fc->log.log = NULL;
for (i = 0; i <= 7; i++)
if (log->need_free & (1 << i))
kfree(log->buffer[i]);
kfree(log);
}
}
}
/**
* put_fs_context - Dispose of a superblock configuration context.
* @fc: The context to dispose of.
*/
void put_fs_context(struct fs_context *fc)
{
struct super_block *sb;
if (fc->root) {
sb = fc->root->d_sb;
dput(fc->root);
fc->root = NULL;
deactivate_super(sb);
}
if (fc->need_free && fc->ops && fc->ops->free)
fc->ops->free(fc);
security_free_mnt_opts(&fc->security);
put_net(fc->net_ns);
put_user_ns(fc->user_ns);
put_cred(fc->cred);
put_fc_log(fc);
put_filesystem(fc->fs_type);
kfree(fc->source);
kfree(fc);
}
EXPORT_SYMBOL(put_fs_context);
/*
* Free the config for a filesystem that doesn't support fs_context.
*/
static void legacy_fs_context_free(struct fs_context *fc)
{
vfs: Implement a filesystem superblock creation/configuration context [AV - unfuck kern_mount_data(); we want non-NULL ->mnt_ns on long-living mounts] [AV - reordering fs/namespace.c is badly overdue, but let's keep it separate from that series] [AV - drop simple_pin_fs() change] [AV - clean vfs_kern_mount() failure exits up] Implement a filesystem context concept to be used during superblock creation for mount and superblock reconfiguration for remount. The mounting procedure then becomes: (1) Allocate new fs_context context. (2) Configure the context. (3) Create superblock. (4) Query the superblock. (5) Create a mount for the superblock. (6) Destroy the context. Rather than calling fs_type->mount(), an fs_context struct is created and fs_type->init_fs_context() is called to set it up. Pointers exist for the filesystem and LSM to hang their private data off. A set of operations has to be set by ->init_fs_context() to provide freeing, duplication, option parsing, binary data parsing, validation, mounting and superblock filling. Legacy filesystems are supported by the provision of a set of legacy fs_context operations that build up a list of mount options and then invoke fs_type->mount() from within the fs_context ->get_tree() operation. This allows all filesystems to be accessed using fs_context. It should be noted that, whilst this patch adds a lot of lines of code, there is quite a bit of duplication with existing code that can be eliminated should all filesystems be converted over. Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2018-11-01 23:07:25 +00:00
struct legacy_fs_context *ctx = fc->fs_private;
if (ctx) {
if (ctx->param_type == LEGACY_FS_INDIVIDUAL_PARAMS)
kfree(ctx->legacy_data);
kfree(ctx);
}
}
/*
* Duplicate a legacy config.
*/
static int legacy_fs_context_dup(struct fs_context *fc, struct fs_context *src_fc)
{
struct legacy_fs_context *ctx;
struct legacy_fs_context *src_ctx = src_fc->fs_private;
ctx = kmemdup(src_ctx, sizeof(*src_ctx), GFP_KERNEL);
if (!ctx)
return -ENOMEM;
if (ctx->param_type == LEGACY_FS_INDIVIDUAL_PARAMS) {
ctx->legacy_data = kmemdup(src_ctx->legacy_data,
src_ctx->data_size, GFP_KERNEL);
if (!ctx->legacy_data) {
kfree(ctx);
return -ENOMEM;
}
}
fc->fs_private = ctx;
return 0;
}
vfs: Implement a filesystem superblock creation/configuration context [AV - unfuck kern_mount_data(); we want non-NULL ->mnt_ns on long-living mounts] [AV - reordering fs/namespace.c is badly overdue, but let's keep it separate from that series] [AV - drop simple_pin_fs() change] [AV - clean vfs_kern_mount() failure exits up] Implement a filesystem context concept to be used during superblock creation for mount and superblock reconfiguration for remount. The mounting procedure then becomes: (1) Allocate new fs_context context. (2) Configure the context. (3) Create superblock. (4) Query the superblock. (5) Create a mount for the superblock. (6) Destroy the context. Rather than calling fs_type->mount(), an fs_context struct is created and fs_type->init_fs_context() is called to set it up. Pointers exist for the filesystem and LSM to hang their private data off. A set of operations has to be set by ->init_fs_context() to provide freeing, duplication, option parsing, binary data parsing, validation, mounting and superblock filling. Legacy filesystems are supported by the provision of a set of legacy fs_context operations that build up a list of mount options and then invoke fs_type->mount() from within the fs_context ->get_tree() operation. This allows all filesystems to be accessed using fs_context. It should be noted that, whilst this patch adds a lot of lines of code, there is quite a bit of duplication with existing code that can be eliminated should all filesystems be converted over. Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2018-11-01 23:07:25 +00:00
/*
* Add a parameter to a legacy config. We build up a comma-separated list of
* options.
*/
static int legacy_parse_param(struct fs_context *fc, struct fs_parameter *param)
{
struct legacy_fs_context *ctx = fc->fs_private;
unsigned int size = ctx->data_size;
size_t len = 0;
if (strcmp(param->key, "source") == 0) {
if (param->type != fs_value_is_string)
return invalf(fc, "VFS: Legacy: Non-string source");
if (fc->source)
return invalf(fc, "VFS: Legacy: Multiple sources");
fc->source = param->string;
param->string = NULL;
return 0;
}
if (ctx->param_type == LEGACY_FS_MONOLITHIC_PARAMS)
return invalf(fc, "VFS: Legacy: Can't mix monolithic and individual options");
switch (param->type) {
case fs_value_is_string:
len = 1 + param->size;
/* Fall through */
case fs_value_is_flag:
len += strlen(param->key);
break;
default:
return invalf(fc, "VFS: Legacy: Parameter type for '%s' not supported",
param->key);
}
if (len > PAGE_SIZE - 2 - size)
return invalf(fc, "VFS: Legacy: Cumulative options too large");
if (strchr(param->key, ',') ||
(param->type == fs_value_is_string &&
memchr(param->string, ',', param->size)))
return invalf(fc, "VFS: Legacy: Option '%s' contained comma",
param->key);
if (!ctx->legacy_data) {
ctx->legacy_data = kmalloc(PAGE_SIZE, GFP_KERNEL);
if (!ctx->legacy_data)
return -ENOMEM;
}
ctx->legacy_data[size++] = ',';
len = strlen(param->key);
memcpy(ctx->legacy_data + size, param->key, len);
size += len;
if (param->type == fs_value_is_string) {
ctx->legacy_data[size++] = '=';
memcpy(ctx->legacy_data + size, param->string, param->size);
size += param->size;
}
ctx->legacy_data[size] = '\0';
ctx->data_size = size;
ctx->param_type = LEGACY_FS_INDIVIDUAL_PARAMS;
return 0;
}
/*
* Add monolithic mount data.
*/
static int legacy_parse_monolithic(struct fs_context *fc, void *data)
{
struct legacy_fs_context *ctx = fc->fs_private;
vfs: Implement a filesystem superblock creation/configuration context [AV - unfuck kern_mount_data(); we want non-NULL ->mnt_ns on long-living mounts] [AV - reordering fs/namespace.c is badly overdue, but let's keep it separate from that series] [AV - drop simple_pin_fs() change] [AV - clean vfs_kern_mount() failure exits up] Implement a filesystem context concept to be used during superblock creation for mount and superblock reconfiguration for remount. The mounting procedure then becomes: (1) Allocate new fs_context context. (2) Configure the context. (3) Create superblock. (4) Query the superblock. (5) Create a mount for the superblock. (6) Destroy the context. Rather than calling fs_type->mount(), an fs_context struct is created and fs_type->init_fs_context() is called to set it up. Pointers exist for the filesystem and LSM to hang their private data off. A set of operations has to be set by ->init_fs_context() to provide freeing, duplication, option parsing, binary data parsing, validation, mounting and superblock filling. Legacy filesystems are supported by the provision of a set of legacy fs_context operations that build up a list of mount options and then invoke fs_type->mount() from within the fs_context ->get_tree() operation. This allows all filesystems to be accessed using fs_context. It should be noted that, whilst this patch adds a lot of lines of code, there is quite a bit of duplication with existing code that can be eliminated should all filesystems be converted over. Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2018-11-01 23:07:25 +00:00
if (ctx->param_type != LEGACY_FS_UNSET_PARAMS) {
pr_warn("VFS: Can't mix monolithic and individual options\n");
return -EINVAL;
}
ctx->legacy_data = data;
vfs: Implement a filesystem superblock creation/configuration context [AV - unfuck kern_mount_data(); we want non-NULL ->mnt_ns on long-living mounts] [AV - reordering fs/namespace.c is badly overdue, but let's keep it separate from that series] [AV - drop simple_pin_fs() change] [AV - clean vfs_kern_mount() failure exits up] Implement a filesystem context concept to be used during superblock creation for mount and superblock reconfiguration for remount. The mounting procedure then becomes: (1) Allocate new fs_context context. (2) Configure the context. (3) Create superblock. (4) Query the superblock. (5) Create a mount for the superblock. (6) Destroy the context. Rather than calling fs_type->mount(), an fs_context struct is created and fs_type->init_fs_context() is called to set it up. Pointers exist for the filesystem and LSM to hang their private data off. A set of operations has to be set by ->init_fs_context() to provide freeing, duplication, option parsing, binary data parsing, validation, mounting and superblock filling. Legacy filesystems are supported by the provision of a set of legacy fs_context operations that build up a list of mount options and then invoke fs_type->mount() from within the fs_context ->get_tree() operation. This allows all filesystems to be accessed using fs_context. It should be noted that, whilst this patch adds a lot of lines of code, there is quite a bit of duplication with existing code that can be eliminated should all filesystems be converted over. Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2018-11-01 23:07:25 +00:00
ctx->param_type = LEGACY_FS_MONOLITHIC_PARAMS;
if (!ctx->legacy_data)
return 0;
vfs: Implement a filesystem superblock creation/configuration context [AV - unfuck kern_mount_data(); we want non-NULL ->mnt_ns on long-living mounts] [AV - reordering fs/namespace.c is badly overdue, but let's keep it separate from that series] [AV - drop simple_pin_fs() change] [AV - clean vfs_kern_mount() failure exits up] Implement a filesystem context concept to be used during superblock creation for mount and superblock reconfiguration for remount. The mounting procedure then becomes: (1) Allocate new fs_context context. (2) Configure the context. (3) Create superblock. (4) Query the superblock. (5) Create a mount for the superblock. (6) Destroy the context. Rather than calling fs_type->mount(), an fs_context struct is created and fs_type->init_fs_context() is called to set it up. Pointers exist for the filesystem and LSM to hang their private data off. A set of operations has to be set by ->init_fs_context() to provide freeing, duplication, option parsing, binary data parsing, validation, mounting and superblock filling. Legacy filesystems are supported by the provision of a set of legacy fs_context operations that build up a list of mount options and then invoke fs_type->mount() from within the fs_context ->get_tree() operation. This allows all filesystems to be accessed using fs_context. It should be noted that, whilst this patch adds a lot of lines of code, there is quite a bit of duplication with existing code that can be eliminated should all filesystems be converted over. Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2018-11-01 23:07:25 +00:00
if (fc->fs_type->fs_flags & FS_BINARY_MOUNTDATA)
return 0;
return security_sb_eat_lsm_opts(ctx->legacy_data, &fc->security);
}
/*
* Get a mountable root with the legacy mount command.
*/
static int legacy_get_tree(struct fs_context *fc)
{
struct legacy_fs_context *ctx = fc->fs_private;
struct super_block *sb;
struct dentry *root;
root = fc->fs_type->mount(fc->fs_type, fc->sb_flags,
fc->source, ctx->legacy_data);
if (IS_ERR(root))
return PTR_ERR(root);
sb = root->d_sb;
BUG_ON(!sb);
fc->root = root;
return 0;
}
/*
* Handle remount.
*/
static int legacy_reconfigure(struct fs_context *fc)
{
struct legacy_fs_context *ctx = fc->fs_private;
struct super_block *sb = fc->root->d_sb;
if (!sb->s_op->remount_fs)
return 0;
return sb->s_op->remount_fs(sb, &fc->sb_flags,
ctx ? ctx->legacy_data : NULL);
}
const struct fs_context_operations legacy_fs_context_ops = {
.free = legacy_fs_context_free,
.dup = legacy_fs_context_dup,
vfs: Implement a filesystem superblock creation/configuration context [AV - unfuck kern_mount_data(); we want non-NULL ->mnt_ns on long-living mounts] [AV - reordering fs/namespace.c is badly overdue, but let's keep it separate from that series] [AV - drop simple_pin_fs() change] [AV - clean vfs_kern_mount() failure exits up] Implement a filesystem context concept to be used during superblock creation for mount and superblock reconfiguration for remount. The mounting procedure then becomes: (1) Allocate new fs_context context. (2) Configure the context. (3) Create superblock. (4) Query the superblock. (5) Create a mount for the superblock. (6) Destroy the context. Rather than calling fs_type->mount(), an fs_context struct is created and fs_type->init_fs_context() is called to set it up. Pointers exist for the filesystem and LSM to hang their private data off. A set of operations has to be set by ->init_fs_context() to provide freeing, duplication, option parsing, binary data parsing, validation, mounting and superblock filling. Legacy filesystems are supported by the provision of a set of legacy fs_context operations that build up a list of mount options and then invoke fs_type->mount() from within the fs_context ->get_tree() operation. This allows all filesystems to be accessed using fs_context. It should be noted that, whilst this patch adds a lot of lines of code, there is quite a bit of duplication with existing code that can be eliminated should all filesystems be converted over. Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2018-11-01 23:07:25 +00:00
.parse_param = legacy_parse_param,
.parse_monolithic = legacy_parse_monolithic,
.get_tree = legacy_get_tree,
.reconfigure = legacy_reconfigure,
};
/*
* Initialise a legacy context for a filesystem that doesn't support
* fs_context.
*/
static int legacy_init_fs_context(struct fs_context *fc)
{
fc->fs_private = kzalloc(sizeof(struct legacy_fs_context), GFP_KERNEL);
if (!fc->fs_private)
return -ENOMEM;
fc->ops = &legacy_fs_context_ops;
return 0;
}
int parse_monolithic_mount_data(struct fs_context *fc, void *data)
{
int (*monolithic_mount_data)(struct fs_context *, void *);
vfs: Implement a filesystem superblock creation/configuration context [AV - unfuck kern_mount_data(); we want non-NULL ->mnt_ns on long-living mounts] [AV - reordering fs/namespace.c is badly overdue, but let's keep it separate from that series] [AV - drop simple_pin_fs() change] [AV - clean vfs_kern_mount() failure exits up] Implement a filesystem context concept to be used during superblock creation for mount and superblock reconfiguration for remount. The mounting procedure then becomes: (1) Allocate new fs_context context. (2) Configure the context. (3) Create superblock. (4) Query the superblock. (5) Create a mount for the superblock. (6) Destroy the context. Rather than calling fs_type->mount(), an fs_context struct is created and fs_type->init_fs_context() is called to set it up. Pointers exist for the filesystem and LSM to hang their private data off. A set of operations has to be set by ->init_fs_context() to provide freeing, duplication, option parsing, binary data parsing, validation, mounting and superblock filling. Legacy filesystems are supported by the provision of a set of legacy fs_context operations that build up a list of mount options and then invoke fs_type->mount() from within the fs_context ->get_tree() operation. This allows all filesystems to be accessed using fs_context. It should be noted that, whilst this patch adds a lot of lines of code, there is quite a bit of duplication with existing code that can be eliminated should all filesystems be converted over. Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2018-11-01 23:07:25 +00:00
monolithic_mount_data = fc->ops->parse_monolithic;
vfs: Implement a filesystem superblock creation/configuration context [AV - unfuck kern_mount_data(); we want non-NULL ->mnt_ns on long-living mounts] [AV - reordering fs/namespace.c is badly overdue, but let's keep it separate from that series] [AV - drop simple_pin_fs() change] [AV - clean vfs_kern_mount() failure exits up] Implement a filesystem context concept to be used during superblock creation for mount and superblock reconfiguration for remount. The mounting procedure then becomes: (1) Allocate new fs_context context. (2) Configure the context. (3) Create superblock. (4) Query the superblock. (5) Create a mount for the superblock. (6) Destroy the context. Rather than calling fs_type->mount(), an fs_context struct is created and fs_type->init_fs_context() is called to set it up. Pointers exist for the filesystem and LSM to hang their private data off. A set of operations has to be set by ->init_fs_context() to provide freeing, duplication, option parsing, binary data parsing, validation, mounting and superblock filling. Legacy filesystems are supported by the provision of a set of legacy fs_context operations that build up a list of mount options and then invoke fs_type->mount() from within the fs_context ->get_tree() operation. This allows all filesystems to be accessed using fs_context. It should be noted that, whilst this patch adds a lot of lines of code, there is quite a bit of duplication with existing code that can be eliminated should all filesystems be converted over. Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2018-11-01 23:07:25 +00:00
if (!monolithic_mount_data)
monolithic_mount_data = generic_parse_monolithic;
return monolithic_mount_data(fc, data);
}
vfs: syscall: Add fsconfig() for configuring and managing a context Add a syscall for configuring a filesystem creation context and triggering actions upon it, to be used in conjunction with fsopen, fspick and fsmount. long fsconfig(int fs_fd, unsigned int cmd, const char *key, const void *value, int aux); Where fs_fd indicates the context, cmd indicates the action to take, key indicates the parameter name for parameter-setting actions and, if needed, value points to a buffer containing the value and aux can give more information for the value. The following command IDs are proposed: (*) FSCONFIG_SET_FLAG: No value is specified. The parameter must be boolean in nature. The key may be prefixed with "no" to invert the setting. value must be NULL and aux must be 0. (*) FSCONFIG_SET_STRING: A string value is specified. The parameter can be expecting boolean, integer, string or take a path. A conversion to an appropriate type will be attempted (which may include looking up as a path). value points to a NUL-terminated string and aux must be 0. (*) FSCONFIG_SET_BINARY: A binary blob is specified. value points to the blob and aux indicates its size. The parameter must be expecting a blob. (*) FSCONFIG_SET_PATH: A non-empty path is specified. The parameter must be expecting a path object. value points to a NUL-terminated string that is the path and aux is a file descriptor at which to start a relative lookup or AT_FDCWD. (*) FSCONFIG_SET_PATH_EMPTY: As fsconfig_set_path, but with AT_EMPTY_PATH implied. (*) FSCONFIG_SET_FD: An open file descriptor is specified. value must be NULL and aux indicates the file descriptor. (*) FSCONFIG_CMD_CREATE: Trigger superblock creation. (*) FSCONFIG_CMD_RECONFIGURE: Trigger superblock reconfiguration. For the "set" command IDs, the idea is that the file_system_type will point to a list of parameters and the types of value that those parameters expect to take. The core code can then do the parse and argument conversion and then give the LSM and FS a cooked option or array of options to use. Source specification is also done the same way same way, using special keys "source", "source1", "source2", etc.. [!] Note that, for the moment, the key and value are just glued back together and handed to the filesystem. Every filesystem that uses options uses match_token() and co. to do this, and this will need to be changed - but not all at once. Example usage: fd = fsopen("ext4", FSOPEN_CLOEXEC); fsconfig(fd, fsconfig_set_path, "source", "/dev/sda1", AT_FDCWD); fsconfig(fd, fsconfig_set_path_empty, "journal_path", "", journal_fd); fsconfig(fd, fsconfig_set_fd, "journal_fd", "", journal_fd); fsconfig(fd, fsconfig_set_flag, "user_xattr", NULL, 0); fsconfig(fd, fsconfig_set_flag, "noacl", NULL, 0); fsconfig(fd, fsconfig_set_string, "sb", "1", 0); fsconfig(fd, fsconfig_set_string, "errors", "continue", 0); fsconfig(fd, fsconfig_set_string, "data", "journal", 0); fsconfig(fd, fsconfig_set_string, "context", "unconfined_u:...", 0); fsconfig(fd, fsconfig_cmd_create, NULL, NULL, 0); mfd = fsmount(fd, FSMOUNT_CLOEXEC, MS_NOEXEC); or: fd = fsopen("ext4", FSOPEN_CLOEXEC); fsconfig(fd, fsconfig_set_string, "source", "/dev/sda1", 0); fsconfig(fd, fsconfig_cmd_create, NULL, NULL, 0); mfd = fsmount(fd, FSMOUNT_CLOEXEC, MS_NOEXEC); or: fd = fsopen("afs", FSOPEN_CLOEXEC); fsconfig(fd, fsconfig_set_string, "source", "#grand.central.org:root.cell", 0); fsconfig(fd, fsconfig_cmd_create, NULL, NULL, 0); mfd = fsmount(fd, FSMOUNT_CLOEXEC, MS_NOEXEC); or: fd = fsopen("jffs2", FSOPEN_CLOEXEC); fsconfig(fd, fsconfig_set_string, "source", "mtd0", 0); fsconfig(fd, fsconfig_cmd_create, NULL, NULL, 0); mfd = fsmount(fd, FSMOUNT_CLOEXEC, MS_NOEXEC); Signed-off-by: David Howells <dhowells@redhat.com> cc: linux-api@vger.kernel.org Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2018-11-01 23:36:09 +00:00
/*
* Clean up a context after performing an action on it and put it into a state
* from where it can be used to reconfigure a superblock.
*
* Note that here we do only the parts that can't fail; the rest is in
* finish_clean_context() below and in between those fs_context is marked
* FS_CONTEXT_AWAITING_RECONF. The reason for splitup is that after
* successful mount or remount we need to report success to userland.
* Trying to do full reinit (for the sake of possible subsequent remount)
* and failing to allocate memory would've put us into a nasty situation.
* So here we only discard the old state and reinitialization is left
* until we actually try to reconfigure.
*/
void vfs_clean_context(struct fs_context *fc)
{
if (fc->need_free && fc->ops && fc->ops->free)
fc->ops->free(fc);
fc->need_free = false;
fc->fs_private = NULL;
fc->s_fs_info = NULL;
fc->sb_flags = 0;
security_free_mnt_opts(&fc->security);
kfree(fc->source);
fc->source = NULL;
fc->purpose = FS_CONTEXT_FOR_RECONFIGURE;
fc->phase = FS_CONTEXT_AWAITING_RECONF;
}
int finish_clean_context(struct fs_context *fc)
{
int error;
if (fc->phase != FS_CONTEXT_AWAITING_RECONF)
return 0;
if (fc->fs_type->init_fs_context)
error = fc->fs_type->init_fs_context(fc);
else
error = legacy_init_fs_context(fc);
if (unlikely(error)) {
fc->phase = FS_CONTEXT_FAILED;
return error;
}
fc->need_free = true;
fc->phase = FS_CONTEXT_RECONF_PARAMS;
return 0;
}