2006-10-04 11:16:22 +02:00
|
|
|
/**
|
|
|
|
* eCryptfs: Linux filesystem encryption layer
|
|
|
|
* Kernel declarations.
|
|
|
|
*
|
|
|
|
* Copyright (C) 1997-2003 Erez Zadok
|
|
|
|
* Copyright (C) 2001-2003 Stony Brook University
|
2008-04-29 09:59:51 +02:00
|
|
|
* Copyright (C) 2004-2008 International Business Machines Corp.
|
2006-10-04 11:16:22 +02:00
|
|
|
* Author(s): Michael A. Halcrow <mahalcro@us.ibm.com>
|
[PATCH] eCryptfs: Public key transport mechanism
This is the transport code for public key functionality in eCryptfs. It
manages encryption/decryption request queues with a transport mechanism.
Currently, netlink is the only implemented transport.
Each inode has a unique File Encryption Key (FEK). Under passphrase, a File
Encryption Key Encryption Key (FEKEK) is generated from a salt/passphrase
combo on mount. This FEKEK encrypts each FEK and writes it into the header of
each file using the packet format specified in RFC 2440. This is all
symmetric key encryption, so it can all be done via the kernel crypto API.
These new patches introduce public key encryption of the FEK. There is no
asymmetric key encryption support in the kernel crypto API, so eCryptfs pushes
the FEK encryption and decryption out to a userspace daemon. After
considering our requirements and determining the complexity of using various
transport mechanisms, we settled on netlink for this communication.
eCryptfs stores authentication tokens into the kernel keyring. These tokens
correlate with individual keys. For passphrase mode of operation, the
authentication token contains the symmetric FEKEK. For public key, the
authentication token contains a PKI type and an opaque data blob managed by
individual PKI modules in userspace.
Each user who opens a file under an eCryptfs partition mounted in public key
mode must be running a daemon. That daemon has the user's credentials and has
access to all of the keys to which the user should have access. The daemon,
when started, initializes the pluggable PKI modules available on the system
and registers itself with the eCryptfs kernel module. Userspace utilities
register public key authentication tokens into the user session keyring.
These authentication tokens correlate key signatures with PKI modules and PKI
blobs. The PKI blobs contain PKI-specific information necessary for the PKI
module to carry out asymmetric key encryption and decryption.
When the eCryptfs module parses the header of an existing file and finds a Tag
1 (Public Key) packet (see RFC 2440), it reads in the public key identifier
(signature). The asymmetrically encrypted FEK is in the Tag 1 packet;
eCryptfs puts together a decrypt request packet containing the signature and
the encrypted FEK, then it passes it to the daemon registered for the
current->euid via a netlink unicast to the PID of the daemon, which was
registered at the time the daemon was started by the user.
The daemon actually just makes calls to libecryptfs, which implements request
packet parsing and manages PKI modules. libecryptfs grabs the public key
authentication token for the given signature from the user session keyring.
This auth tok tells libecryptfs which PKI module should receive the request.
libecryptfs then makes a decrypt() call to the PKI module, and it passes along
the PKI block from the auth tok. The PKI uses the blob to figure out how it
should decrypt the data passed to it; it performs the decryption and passes
the decrypted data back to libecryptfs. libecryptfs then puts together a
reply packet with the decrypted FEK and passes that back to the eCryptfs
module.
The eCryptfs module manages these request callouts to userspace code via
message context structs. The module maintains an array of message context
structs and places the elements of the array on two lists: a free and an
allocated list. When eCryptfs wants to make a request, it moves a msg ctx
from the free list to the allocated list, sets its state to pending, and fires
off the message to the user's registered daemon.
When eCryptfs receives a netlink message (via the callback), it correlates the
msg ctx struct in the alloc list with the data in the message itself. The
msg->index contains the offset of the array of msg ctx structs. It verifies
that the registered daemon PID is the same as the PID of the process that sent
the message. It also validates a sequence number between the received packet
and the msg ctx. Then, it copies the contents of the message (the reply
packet) into the msg ctx struct, sets the state in the msg ctx to done, and
wakes up the process that was sleeping while waiting for the reply.
The sleeping process was whatever was performing the sys_open(). This process
originally called ecryptfs_send_message(); it is now in
ecryptfs_wait_for_response(). When it wakes up and sees that the msg ctx
state was set to done, it returns a pointer to the message contents (the reply
packet) and returns. If all went well, this packet contains the decrypted
FEK, which is then copied into the crypt_stat struct, and life continues as
normal.
The case for creation of a new file is very similar, only instead of a decrypt
request, eCryptfs sends out an encrypt request.
> - We have a great clod of key mangement code in-kernel. Why is that
> not suitable (or growable) for public key management?
eCryptfs uses Howells' keyring to store persistent key data and PKI state
information. It defers public key cryptographic transformations to userspace
code. The userspace data manipulation request really is orthogonal to key
management in and of itself. What eCryptfs basically needs is a secure way to
communicate with a particular daemon for a particular task doing a syscall,
based on the UID. Nothing running under another UID should be able to access
that channel of communication.
> - Is it appropriate that new infrastructure for public key
> management be private to a particular fs?
The messaging.c file contains a lot of code that, perhaps, could be extracted
into a separate kernel service. In essence, this would be a sort of
request/reply mechanism that would involve a userspace daemon. I am not aware
of anything that does quite what eCryptfs does, so I was not aware of any
existing tools to do just what we wanted.
> What happens if one of these daemons exits without sending a quit
> message?
There is a stale uid<->pid association in the hash table for that user. When
the user registers a new daemon, eCryptfs cleans up the old association and
generates a new one. See ecryptfs_process_helo().
> - _why_ does it use netlink?
Netlink provides the transport mechanism that would minimize the complexity of
the implementation, given that we can have multiple daemons (one per user). I
explored the possibility of using relayfs, but that would involve having to
introduce control channels and a protocol for creating and tearing down
channels for the daemons. We do not have to worry about any of that with
netlink.
Signed-off-by: Michael Halcrow <mhalcrow@us.ibm.com>
Cc: David Howells <dhowells@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-02-12 09:53:43 +01:00
|
|
|
* Trevor S. Highland <trevor.highland@gmail.com>
|
|
|
|
* Tyler Hicks <tyhicks@ou.edu>
|
2006-10-04 11:16:22 +02:00
|
|
|
*
|
|
|
|
* This program is free software; you can redistribute it and/or
|
|
|
|
* modify it under the terms of the GNU General Public License as
|
|
|
|
* published by the Free Software Foundation; either version 2 of the
|
|
|
|
* License, or (at your option) any later version.
|
|
|
|
*
|
|
|
|
* This program is distributed in the hope that it will be useful, but
|
|
|
|
* WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
|
|
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
|
|
|
|
* General Public License for more details.
|
|
|
|
*
|
|
|
|
* You should have received a copy of the GNU General Public License
|
|
|
|
* along with this program; if not, write to the Free Software
|
|
|
|
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA
|
|
|
|
* 02111-1307, USA.
|
|
|
|
*/
|
|
|
|
|
|
|
|
#ifndef ECRYPTFS_KERNEL_H
|
|
|
|
#define ECRYPTFS_KERNEL_H
|
|
|
|
|
2016-01-25 03:29:33 +01:00
|
|
|
#include <crypto/skcipher.h>
|
2006-10-04 11:16:22 +02:00
|
|
|
#include <keys/user-type.h>
|
2011-06-27 13:45:45 +02:00
|
|
|
#include <keys/encrypted-type.h>
|
2006-10-04 11:16:22 +02:00
|
|
|
#include <linux/fs.h>
|
2006-12-08 11:36:31 +01:00
|
|
|
#include <linux/fs_stack.h>
|
2006-12-08 11:36:34 +01:00
|
|
|
#include <linux/namei.h>
|
2006-10-04 11:16:22 +02:00
|
|
|
#include <linux/scatterlist.h>
|
2007-02-12 09:53:44 +01:00
|
|
|
#include <linux/hash.h>
|
2008-04-29 09:59:52 +02:00
|
|
|
#include <linux/nsproxy.h>
|
2010-04-22 12:22:04 +02:00
|
|
|
#include <linux/backing-dev.h>
|
2011-06-27 13:45:43 +02:00
|
|
|
#include <linux/ecryptfs.h>
|
2006-10-04 11:16:22 +02:00
|
|
|
|
|
|
|
#define ECRYPTFS_DEFAULT_IV_BYTES 16
|
|
|
|
#define ECRYPTFS_DEFAULT_EXTENT_SIZE 4096
|
|
|
|
#define ECRYPTFS_MINIMUM_HEADER_EXTENT_SIZE 8192
|
[PATCH] eCryptfs: Public key transport mechanism
This is the transport code for public key functionality in eCryptfs. It
manages encryption/decryption request queues with a transport mechanism.
Currently, netlink is the only implemented transport.
Each inode has a unique File Encryption Key (FEK). Under passphrase, a File
Encryption Key Encryption Key (FEKEK) is generated from a salt/passphrase
combo on mount. This FEKEK encrypts each FEK and writes it into the header of
each file using the packet format specified in RFC 2440. This is all
symmetric key encryption, so it can all be done via the kernel crypto API.
These new patches introduce public key encryption of the FEK. There is no
asymmetric key encryption support in the kernel crypto API, so eCryptfs pushes
the FEK encryption and decryption out to a userspace daemon. After
considering our requirements and determining the complexity of using various
transport mechanisms, we settled on netlink for this communication.
eCryptfs stores authentication tokens into the kernel keyring. These tokens
correlate with individual keys. For passphrase mode of operation, the
authentication token contains the symmetric FEKEK. For public key, the
authentication token contains a PKI type and an opaque data blob managed by
individual PKI modules in userspace.
Each user who opens a file under an eCryptfs partition mounted in public key
mode must be running a daemon. That daemon has the user's credentials and has
access to all of the keys to which the user should have access. The daemon,
when started, initializes the pluggable PKI modules available on the system
and registers itself with the eCryptfs kernel module. Userspace utilities
register public key authentication tokens into the user session keyring.
These authentication tokens correlate key signatures with PKI modules and PKI
blobs. The PKI blobs contain PKI-specific information necessary for the PKI
module to carry out asymmetric key encryption and decryption.
When the eCryptfs module parses the header of an existing file and finds a Tag
1 (Public Key) packet (see RFC 2440), it reads in the public key identifier
(signature). The asymmetrically encrypted FEK is in the Tag 1 packet;
eCryptfs puts together a decrypt request packet containing the signature and
the encrypted FEK, then it passes it to the daemon registered for the
current->euid via a netlink unicast to the PID of the daemon, which was
registered at the time the daemon was started by the user.
The daemon actually just makes calls to libecryptfs, which implements request
packet parsing and manages PKI modules. libecryptfs grabs the public key
authentication token for the given signature from the user session keyring.
This auth tok tells libecryptfs which PKI module should receive the request.
libecryptfs then makes a decrypt() call to the PKI module, and it passes along
the PKI block from the auth tok. The PKI uses the blob to figure out how it
should decrypt the data passed to it; it performs the decryption and passes
the decrypted data back to libecryptfs. libecryptfs then puts together a
reply packet with the decrypted FEK and passes that back to the eCryptfs
module.
The eCryptfs module manages these request callouts to userspace code via
message context structs. The module maintains an array of message context
structs and places the elements of the array on two lists: a free and an
allocated list. When eCryptfs wants to make a request, it moves a msg ctx
from the free list to the allocated list, sets its state to pending, and fires
off the message to the user's registered daemon.
When eCryptfs receives a netlink message (via the callback), it correlates the
msg ctx struct in the alloc list with the data in the message itself. The
msg->index contains the offset of the array of msg ctx structs. It verifies
that the registered daemon PID is the same as the PID of the process that sent
the message. It also validates a sequence number between the received packet
and the msg ctx. Then, it copies the contents of the message (the reply
packet) into the msg ctx struct, sets the state in the msg ctx to done, and
wakes up the process that was sleeping while waiting for the reply.
The sleeping process was whatever was performing the sys_open(). This process
originally called ecryptfs_send_message(); it is now in
ecryptfs_wait_for_response(). When it wakes up and sees that the msg ctx
state was set to done, it returns a pointer to the message contents (the reply
packet) and returns. If all went well, this packet contains the decrypted
FEK, which is then copied into the crypt_stat struct, and life continues as
normal.
The case for creation of a new file is very similar, only instead of a decrypt
request, eCryptfs sends out an encrypt request.
> - We have a great clod of key mangement code in-kernel. Why is that
> not suitable (or growable) for public key management?
eCryptfs uses Howells' keyring to store persistent key data and PKI state
information. It defers public key cryptographic transformations to userspace
code. The userspace data manipulation request really is orthogonal to key
management in and of itself. What eCryptfs basically needs is a secure way to
communicate with a particular daemon for a particular task doing a syscall,
based on the UID. Nothing running under another UID should be able to access
that channel of communication.
> - Is it appropriate that new infrastructure for public key
> management be private to a particular fs?
The messaging.c file contains a lot of code that, perhaps, could be extracted
into a separate kernel service. In essence, this would be a sort of
request/reply mechanism that would involve a userspace daemon. I am not aware
of anything that does quite what eCryptfs does, so I was not aware of any
existing tools to do just what we wanted.
> What happens if one of these daemons exits without sending a quit
> message?
There is a stale uid<->pid association in the hash table for that user. When
the user registers a new daemon, eCryptfs cleans up the old association and
generates a new one. See ecryptfs_process_helo().
> - _why_ does it use netlink?
Netlink provides the transport mechanism that would minimize the complexity of
the implementation, given that we can have multiple daemons (one per user). I
explored the possibility of using relayfs, but that would involve having to
introduce control channels and a protocol for creating and tearing down
channels for the daemons. We do not have to worry about any of that with
netlink.
Signed-off-by: Michael Halcrow <mhalcrow@us.ibm.com>
Cc: David Howells <dhowells@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-02-12 09:53:43 +01:00
|
|
|
#define ECRYPTFS_DEFAULT_MSG_CTX_ELEMS 32
|
|
|
|
#define ECRYPTFS_DEFAULT_SEND_TIMEOUT HZ
|
|
|
|
#define ECRYPTFS_MAX_MSG_CTX_TTL (HZ*3)
|
|
|
|
#define ECRYPTFS_DEFAULT_NUM_USERS 4
|
|
|
|
#define ECRYPTFS_MAX_NUM_USERS 32768
|
2007-02-12 09:53:46 +01:00
|
|
|
#define ECRYPTFS_XATTR_NAME "user.ecryptfs"
|
2006-10-04 11:16:22 +02:00
|
|
|
|
|
|
|
void ecryptfs_dump_auth_tok(struct ecryptfs_auth_tok *auth_tok);
|
|
|
|
extern void ecryptfs_to_hex(char *dst, char *src, size_t src_size);
|
|
|
|
extern void ecryptfs_from_hex(char *dst, char *src, int dst_size);
|
|
|
|
|
|
|
|
struct ecryptfs_key_record {
|
|
|
|
unsigned char type;
|
|
|
|
size_t enc_key_size;
|
|
|
|
unsigned char sig[ECRYPTFS_SIG_SIZE];
|
|
|
|
unsigned char enc_key[ECRYPTFS_MAX_ENCRYPTED_KEY_BYTES];
|
|
|
|
};
|
|
|
|
|
|
|
|
struct ecryptfs_auth_tok_list {
|
|
|
|
struct ecryptfs_auth_tok *auth_tok;
|
|
|
|
struct list_head list;
|
|
|
|
};
|
|
|
|
|
|
|
|
struct ecryptfs_crypt_stat;
|
|
|
|
struct ecryptfs_mount_crypt_stat;
|
|
|
|
|
|
|
|
struct ecryptfs_page_crypt_context {
|
|
|
|
struct page *page;
|
|
|
|
#define ECRYPTFS_PREPARE_COMMIT_MODE 0
|
|
|
|
#define ECRYPTFS_WRITEPAGE_MODE 1
|
|
|
|
unsigned int mode;
|
|
|
|
union {
|
|
|
|
struct file *lower_file;
|
|
|
|
struct writeback_control *wbc;
|
|
|
|
} param;
|
|
|
|
};
|
|
|
|
|
2011-06-27 13:45:45 +02:00
|
|
|
#if defined(CONFIG_ENCRYPTED_KEYS) || defined(CONFIG_ENCRYPTED_KEYS_MODULE)
|
|
|
|
static inline struct ecryptfs_auth_tok *
|
|
|
|
ecryptfs_get_encrypted_key_payload_data(struct key *key)
|
|
|
|
{
|
|
|
|
if (key->type == &key_type_encrypted)
|
|
|
|
return (struct ecryptfs_auth_tok *)
|
2015-10-21 15:04:48 +02:00
|
|
|
(&((struct encrypted_key_payload *)key->payload.data[0])->payload_data);
|
2011-06-27 13:45:45 +02:00
|
|
|
else
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline struct key *ecryptfs_get_encrypted_key(char *sig)
|
|
|
|
{
|
|
|
|
return request_key(&key_type_encrypted, sig, NULL);
|
|
|
|
}
|
|
|
|
|
|
|
|
#else
|
|
|
|
static inline struct ecryptfs_auth_tok *
|
|
|
|
ecryptfs_get_encrypted_key_payload_data(struct key *key)
|
|
|
|
{
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline struct key *ecryptfs_get_encrypted_key(char *sig)
|
|
|
|
{
|
|
|
|
return ERR_PTR(-ENOKEY);
|
|
|
|
}
|
|
|
|
|
|
|
|
#endif /* CONFIG_ENCRYPTED_KEYS */
|
|
|
|
|
2006-10-04 11:16:22 +02:00
|
|
|
static inline struct ecryptfs_auth_tok *
|
|
|
|
ecryptfs_get_key_payload_data(struct key *key)
|
|
|
|
{
|
2011-06-27 13:45:45 +02:00
|
|
|
struct ecryptfs_auth_tok *auth_tok;
|
|
|
|
|
|
|
|
auth_tok = ecryptfs_get_encrypted_key_payload_data(key);
|
|
|
|
if (!auth_tok)
|
2015-10-21 15:04:48 +02:00
|
|
|
return (struct ecryptfs_auth_tok *)user_key_payload(key)->data;
|
2011-06-27 13:45:45 +02:00
|
|
|
else
|
|
|
|
return auth_tok;
|
2006-10-04 11:16:22 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
#define ECRYPTFS_MAX_KEYSET_SIZE 1024
|
2015-02-23 12:34:10 +01:00
|
|
|
#define ECRYPTFS_MAX_CIPHER_NAME_SIZE 31
|
2006-10-04 11:16:22 +02:00
|
|
|
#define ECRYPTFS_MAX_NUM_ENC_KEYS 64
|
|
|
|
#define ECRYPTFS_MAX_IV_BYTES 16 /* 128 bits */
|
|
|
|
#define ECRYPTFS_SALT_BYTES 2
|
|
|
|
#define MAGIC_ECRYPTFS_MARKER 0x3c81b7f5
|
|
|
|
#define MAGIC_ECRYPTFS_MARKER_SIZE_BYTES 8 /* 4*2 */
|
2007-10-16 10:28:05 +02:00
|
|
|
#define ECRYPTFS_FILE_SIZE_BYTES (sizeof(u64))
|
2011-05-24 11:56:23 +02:00
|
|
|
#define ECRYPTFS_SIZE_AND_MARKER_BYTES (ECRYPTFS_FILE_SIZE_BYTES \
|
|
|
|
+ MAGIC_ECRYPTFS_MARKER_SIZE_BYTES)
|
2006-10-04 11:16:22 +02:00
|
|
|
#define ECRYPTFS_DEFAULT_CIPHER "aes"
|
|
|
|
#define ECRYPTFS_DEFAULT_KEY_BYTES 16
|
2006-10-31 07:07:17 +01:00
|
|
|
#define ECRYPTFS_DEFAULT_HASH "md5"
|
2009-01-06 23:41:57 +01:00
|
|
|
#define ECRYPTFS_TAG_70_DIGEST ECRYPTFS_DEFAULT_HASH
|
[PATCH] eCryptfs: Public key transport mechanism
This is the transport code for public key functionality in eCryptfs. It
manages encryption/decryption request queues with a transport mechanism.
Currently, netlink is the only implemented transport.
Each inode has a unique File Encryption Key (FEK). Under passphrase, a File
Encryption Key Encryption Key (FEKEK) is generated from a salt/passphrase
combo on mount. This FEKEK encrypts each FEK and writes it into the header of
each file using the packet format specified in RFC 2440. This is all
symmetric key encryption, so it can all be done via the kernel crypto API.
These new patches introduce public key encryption of the FEK. There is no
asymmetric key encryption support in the kernel crypto API, so eCryptfs pushes
the FEK encryption and decryption out to a userspace daemon. After
considering our requirements and determining the complexity of using various
transport mechanisms, we settled on netlink for this communication.
eCryptfs stores authentication tokens into the kernel keyring. These tokens
correlate with individual keys. For passphrase mode of operation, the
authentication token contains the symmetric FEKEK. For public key, the
authentication token contains a PKI type and an opaque data blob managed by
individual PKI modules in userspace.
Each user who opens a file under an eCryptfs partition mounted in public key
mode must be running a daemon. That daemon has the user's credentials and has
access to all of the keys to which the user should have access. The daemon,
when started, initializes the pluggable PKI modules available on the system
and registers itself with the eCryptfs kernel module. Userspace utilities
register public key authentication tokens into the user session keyring.
These authentication tokens correlate key signatures with PKI modules and PKI
blobs. The PKI blobs contain PKI-specific information necessary for the PKI
module to carry out asymmetric key encryption and decryption.
When the eCryptfs module parses the header of an existing file and finds a Tag
1 (Public Key) packet (see RFC 2440), it reads in the public key identifier
(signature). The asymmetrically encrypted FEK is in the Tag 1 packet;
eCryptfs puts together a decrypt request packet containing the signature and
the encrypted FEK, then it passes it to the daemon registered for the
current->euid via a netlink unicast to the PID of the daemon, which was
registered at the time the daemon was started by the user.
The daemon actually just makes calls to libecryptfs, which implements request
packet parsing and manages PKI modules. libecryptfs grabs the public key
authentication token for the given signature from the user session keyring.
This auth tok tells libecryptfs which PKI module should receive the request.
libecryptfs then makes a decrypt() call to the PKI module, and it passes along
the PKI block from the auth tok. The PKI uses the blob to figure out how it
should decrypt the data passed to it; it performs the decryption and passes
the decrypted data back to libecryptfs. libecryptfs then puts together a
reply packet with the decrypted FEK and passes that back to the eCryptfs
module.
The eCryptfs module manages these request callouts to userspace code via
message context structs. The module maintains an array of message context
structs and places the elements of the array on two lists: a free and an
allocated list. When eCryptfs wants to make a request, it moves a msg ctx
from the free list to the allocated list, sets its state to pending, and fires
off the message to the user's registered daemon.
When eCryptfs receives a netlink message (via the callback), it correlates the
msg ctx struct in the alloc list with the data in the message itself. The
msg->index contains the offset of the array of msg ctx structs. It verifies
that the registered daemon PID is the same as the PID of the process that sent
the message. It also validates a sequence number between the received packet
and the msg ctx. Then, it copies the contents of the message (the reply
packet) into the msg ctx struct, sets the state in the msg ctx to done, and
wakes up the process that was sleeping while waiting for the reply.
The sleeping process was whatever was performing the sys_open(). This process
originally called ecryptfs_send_message(); it is now in
ecryptfs_wait_for_response(). When it wakes up and sees that the msg ctx
state was set to done, it returns a pointer to the message contents (the reply
packet) and returns. If all went well, this packet contains the decrypted
FEK, which is then copied into the crypt_stat struct, and life continues as
normal.
The case for creation of a new file is very similar, only instead of a decrypt
request, eCryptfs sends out an encrypt request.
> - We have a great clod of key mangement code in-kernel. Why is that
> not suitable (or growable) for public key management?
eCryptfs uses Howells' keyring to store persistent key data and PKI state
information. It defers public key cryptographic transformations to userspace
code. The userspace data manipulation request really is orthogonal to key
management in and of itself. What eCryptfs basically needs is a secure way to
communicate with a particular daemon for a particular task doing a syscall,
based on the UID. Nothing running under another UID should be able to access
that channel of communication.
> - Is it appropriate that new infrastructure for public key
> management be private to a particular fs?
The messaging.c file contains a lot of code that, perhaps, could be extracted
into a separate kernel service. In essence, this would be a sort of
request/reply mechanism that would involve a userspace daemon. I am not aware
of anything that does quite what eCryptfs does, so I was not aware of any
existing tools to do just what we wanted.
> What happens if one of these daemons exits without sending a quit
> message?
There is a stale uid<->pid association in the hash table for that user. When
the user registers a new daemon, eCryptfs cleans up the old association and
generates a new one. See ecryptfs_process_helo().
> - _why_ does it use netlink?
Netlink provides the transport mechanism that would minimize the complexity of
the implementation, given that we can have multiple daemons (one per user). I
explored the possibility of using relayfs, but that would involve having to
introduce control channels and a protocol for creating and tearing down
channels for the daemons. We do not have to worry about any of that with
netlink.
Signed-off-by: Michael Halcrow <mhalcrow@us.ibm.com>
Cc: David Howells <dhowells@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-02-12 09:53:43 +01:00
|
|
|
#define ECRYPTFS_TAG_1_PACKET_TYPE 0x01
|
2006-10-04 11:16:22 +02:00
|
|
|
#define ECRYPTFS_TAG_3_PACKET_TYPE 0x8C
|
|
|
|
#define ECRYPTFS_TAG_11_PACKET_TYPE 0xED
|
[PATCH] eCryptfs: Public key transport mechanism
This is the transport code for public key functionality in eCryptfs. It
manages encryption/decryption request queues with a transport mechanism.
Currently, netlink is the only implemented transport.
Each inode has a unique File Encryption Key (FEK). Under passphrase, a File
Encryption Key Encryption Key (FEKEK) is generated from a salt/passphrase
combo on mount. This FEKEK encrypts each FEK and writes it into the header of
each file using the packet format specified in RFC 2440. This is all
symmetric key encryption, so it can all be done via the kernel crypto API.
These new patches introduce public key encryption of the FEK. There is no
asymmetric key encryption support in the kernel crypto API, so eCryptfs pushes
the FEK encryption and decryption out to a userspace daemon. After
considering our requirements and determining the complexity of using various
transport mechanisms, we settled on netlink for this communication.
eCryptfs stores authentication tokens into the kernel keyring. These tokens
correlate with individual keys. For passphrase mode of operation, the
authentication token contains the symmetric FEKEK. For public key, the
authentication token contains a PKI type and an opaque data blob managed by
individual PKI modules in userspace.
Each user who opens a file under an eCryptfs partition mounted in public key
mode must be running a daemon. That daemon has the user's credentials and has
access to all of the keys to which the user should have access. The daemon,
when started, initializes the pluggable PKI modules available on the system
and registers itself with the eCryptfs kernel module. Userspace utilities
register public key authentication tokens into the user session keyring.
These authentication tokens correlate key signatures with PKI modules and PKI
blobs. The PKI blobs contain PKI-specific information necessary for the PKI
module to carry out asymmetric key encryption and decryption.
When the eCryptfs module parses the header of an existing file and finds a Tag
1 (Public Key) packet (see RFC 2440), it reads in the public key identifier
(signature). The asymmetrically encrypted FEK is in the Tag 1 packet;
eCryptfs puts together a decrypt request packet containing the signature and
the encrypted FEK, then it passes it to the daemon registered for the
current->euid via a netlink unicast to the PID of the daemon, which was
registered at the time the daemon was started by the user.
The daemon actually just makes calls to libecryptfs, which implements request
packet parsing and manages PKI modules. libecryptfs grabs the public key
authentication token for the given signature from the user session keyring.
This auth tok tells libecryptfs which PKI module should receive the request.
libecryptfs then makes a decrypt() call to the PKI module, and it passes along
the PKI block from the auth tok. The PKI uses the blob to figure out how it
should decrypt the data passed to it; it performs the decryption and passes
the decrypted data back to libecryptfs. libecryptfs then puts together a
reply packet with the decrypted FEK and passes that back to the eCryptfs
module.
The eCryptfs module manages these request callouts to userspace code via
message context structs. The module maintains an array of message context
structs and places the elements of the array on two lists: a free and an
allocated list. When eCryptfs wants to make a request, it moves a msg ctx
from the free list to the allocated list, sets its state to pending, and fires
off the message to the user's registered daemon.
When eCryptfs receives a netlink message (via the callback), it correlates the
msg ctx struct in the alloc list with the data in the message itself. The
msg->index contains the offset of the array of msg ctx structs. It verifies
that the registered daemon PID is the same as the PID of the process that sent
the message. It also validates a sequence number between the received packet
and the msg ctx. Then, it copies the contents of the message (the reply
packet) into the msg ctx struct, sets the state in the msg ctx to done, and
wakes up the process that was sleeping while waiting for the reply.
The sleeping process was whatever was performing the sys_open(). This process
originally called ecryptfs_send_message(); it is now in
ecryptfs_wait_for_response(). When it wakes up and sees that the msg ctx
state was set to done, it returns a pointer to the message contents (the reply
packet) and returns. If all went well, this packet contains the decrypted
FEK, which is then copied into the crypt_stat struct, and life continues as
normal.
The case for creation of a new file is very similar, only instead of a decrypt
request, eCryptfs sends out an encrypt request.
> - We have a great clod of key mangement code in-kernel. Why is that
> not suitable (or growable) for public key management?
eCryptfs uses Howells' keyring to store persistent key data and PKI state
information. It defers public key cryptographic transformations to userspace
code. The userspace data manipulation request really is orthogonal to key
management in and of itself. What eCryptfs basically needs is a secure way to
communicate with a particular daemon for a particular task doing a syscall,
based on the UID. Nothing running under another UID should be able to access
that channel of communication.
> - Is it appropriate that new infrastructure for public key
> management be private to a particular fs?
The messaging.c file contains a lot of code that, perhaps, could be extracted
into a separate kernel service. In essence, this would be a sort of
request/reply mechanism that would involve a userspace daemon. I am not aware
of anything that does quite what eCryptfs does, so I was not aware of any
existing tools to do just what we wanted.
> What happens if one of these daemons exits without sending a quit
> message?
There is a stale uid<->pid association in the hash table for that user. When
the user registers a new daemon, eCryptfs cleans up the old association and
generates a new one. See ecryptfs_process_helo().
> - _why_ does it use netlink?
Netlink provides the transport mechanism that would minimize the complexity of
the implementation, given that we can have multiple daemons (one per user). I
explored the possibility of using relayfs, but that would involve having to
introduce control channels and a protocol for creating and tearing down
channels for the daemons. We do not have to worry about any of that with
netlink.
Signed-off-by: Michael Halcrow <mhalcrow@us.ibm.com>
Cc: David Howells <dhowells@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-02-12 09:53:43 +01:00
|
|
|
#define ECRYPTFS_TAG_64_PACKET_TYPE 0x40
|
|
|
|
#define ECRYPTFS_TAG_65_PACKET_TYPE 0x41
|
|
|
|
#define ECRYPTFS_TAG_66_PACKET_TYPE 0x42
|
|
|
|
#define ECRYPTFS_TAG_67_PACKET_TYPE 0x43
|
2009-01-06 23:41:57 +01:00
|
|
|
#define ECRYPTFS_TAG_70_PACKET_TYPE 0x46 /* FNEK-encrypted filename
|
|
|
|
* as dentry name */
|
|
|
|
#define ECRYPTFS_TAG_71_PACKET_TYPE 0x47 /* FNEK-encrypted filename in
|
|
|
|
* metadata */
|
|
|
|
#define ECRYPTFS_TAG_72_PACKET_TYPE 0x48 /* FEK-encrypted filename as
|
|
|
|
* dentry name */
|
|
|
|
#define ECRYPTFS_TAG_73_PACKET_TYPE 0x49 /* FEK-encrypted filename as
|
|
|
|
* metadata */
|
2012-01-14 16:46:46 +01:00
|
|
|
#define ECRYPTFS_MIN_PKT_LEN_SIZE 1 /* Min size to specify packet length */
|
|
|
|
#define ECRYPTFS_MAX_PKT_LEN_SIZE 2 /* Pass at least this many bytes to
|
|
|
|
* ecryptfs_parse_packet_length() and
|
|
|
|
* ecryptfs_write_packet_length()
|
|
|
|
*/
|
2009-01-06 23:41:57 +01:00
|
|
|
/* Constraint: ECRYPTFS_FILENAME_MIN_RANDOM_PREPEND_BYTES >=
|
|
|
|
* ECRYPTFS_MAX_IV_BYTES */
|
|
|
|
#define ECRYPTFS_FILENAME_MIN_RANDOM_PREPEND_BYTES 16
|
|
|
|
#define ECRYPTFS_NON_NULL 0x42 /* A reasonable substitute for NULL */
|
2006-10-04 11:16:22 +02:00
|
|
|
#define MD5_DIGEST_SIZE 16
|
2009-01-06 23:41:57 +01:00
|
|
|
#define ECRYPTFS_TAG_70_DIGEST_SIZE MD5_DIGEST_SIZE
|
2011-11-05 18:45:08 +01:00
|
|
|
#define ECRYPTFS_TAG_70_MIN_METADATA_SIZE (1 + ECRYPTFS_MIN_PKT_LEN_SIZE \
|
|
|
|
+ ECRYPTFS_SIG_SIZE + 1 + 1)
|
|
|
|
#define ECRYPTFS_TAG_70_MAX_METADATA_SIZE (1 + ECRYPTFS_MAX_PKT_LEN_SIZE \
|
|
|
|
+ ECRYPTFS_SIG_SIZE + 1 + 1)
|
2009-01-06 23:41:57 +01:00
|
|
|
#define ECRYPTFS_FEK_ENCRYPTED_FILENAME_PREFIX "ECRYPTFS_FEK_ENCRYPTED."
|
|
|
|
#define ECRYPTFS_FEK_ENCRYPTED_FILENAME_PREFIX_SIZE 23
|
|
|
|
#define ECRYPTFS_FNEK_ENCRYPTED_FILENAME_PREFIX "ECRYPTFS_FNEK_ENCRYPTED."
|
|
|
|
#define ECRYPTFS_FNEK_ENCRYPTED_FILENAME_PREFIX_SIZE 24
|
|
|
|
#define ECRYPTFS_ENCRYPTED_DENTRY_NAME_LEN (18 + 1 + 4 + 1 + 32)
|
2006-10-04 11:16:22 +02:00
|
|
|
|
2013-02-28 09:39:37 +01:00
|
|
|
#ifdef CONFIG_ECRYPT_FS_MESSAGING
|
|
|
|
# define ECRYPTFS_VERSIONING_MASK_MESSAGING (ECRYPTFS_VERSIONING_DEVMISC \
|
|
|
|
| ECRYPTFS_VERSIONING_PUBKEY)
|
|
|
|
#else
|
|
|
|
# define ECRYPTFS_VERSIONING_MASK_MESSAGING 0
|
|
|
|
#endif
|
|
|
|
|
|
|
|
#define ECRYPTFS_VERSIONING_MASK (ECRYPTFS_VERSIONING_PASSPHRASE \
|
|
|
|
| ECRYPTFS_VERSIONING_PLAINTEXT_PASSTHROUGH \
|
|
|
|
| ECRYPTFS_VERSIONING_XATTR \
|
|
|
|
| ECRYPTFS_VERSIONING_MULTKEY \
|
|
|
|
| ECRYPTFS_VERSIONING_MASK_MESSAGING \
|
|
|
|
| ECRYPTFS_VERSIONING_FILENAME_ENCRYPTION)
|
2007-10-16 10:27:53 +02:00
|
|
|
struct ecryptfs_key_sig {
|
|
|
|
struct list_head crypt_stat_list;
|
2011-03-21 16:00:52 +01:00
|
|
|
char keysig[ECRYPTFS_SIG_SIZE_HEX + 1];
|
2007-10-16 10:27:53 +02:00
|
|
|
};
|
|
|
|
|
2009-01-06 23:41:58 +01:00
|
|
|
struct ecryptfs_filename {
|
|
|
|
struct list_head crypt_stat_list;
|
|
|
|
#define ECRYPTFS_FILENAME_CONTAINS_DECRYPTED 0x00000001
|
|
|
|
u32 flags;
|
|
|
|
u32 seq_no;
|
|
|
|
char *filename;
|
|
|
|
char *encrypted_filename;
|
|
|
|
size_t filename_size;
|
|
|
|
size_t encrypted_filename_size;
|
|
|
|
char fnek_sig[ECRYPTFS_SIG_SIZE_HEX];
|
|
|
|
char dentry_name[ECRYPTFS_ENCRYPTED_DENTRY_NAME_LEN + 1];
|
|
|
|
};
|
|
|
|
|
2006-10-04 11:16:22 +02:00
|
|
|
/**
|
|
|
|
* This is the primary struct associated with each encrypted file.
|
|
|
|
*
|
|
|
|
* TODO: cache align/pack?
|
|
|
|
*/
|
|
|
|
struct ecryptfs_crypt_stat {
|
2009-01-06 23:41:58 +01:00
|
|
|
#define ECRYPTFS_STRUCT_INITIALIZED 0x00000001
|
|
|
|
#define ECRYPTFS_POLICY_APPLIED 0x00000002
|
2011-02-23 07:54:20 +01:00
|
|
|
#define ECRYPTFS_ENCRYPTED 0x00000004
|
|
|
|
#define ECRYPTFS_SECURITY_WARNING 0x00000008
|
|
|
|
#define ECRYPTFS_ENABLE_HMAC 0x00000010
|
|
|
|
#define ECRYPTFS_ENCRYPT_IV_PAGES 0x00000020
|
|
|
|
#define ECRYPTFS_KEY_VALID 0x00000040
|
|
|
|
#define ECRYPTFS_METADATA_IN_XATTR 0x00000080
|
|
|
|
#define ECRYPTFS_VIEW_AS_ENCRYPTED 0x00000100
|
|
|
|
#define ECRYPTFS_KEY_SET 0x00000200
|
|
|
|
#define ECRYPTFS_ENCRYPT_FILENAMES 0x00000400
|
|
|
|
#define ECRYPTFS_ENCFN_USE_MOUNT_FNEK 0x00000800
|
|
|
|
#define ECRYPTFS_ENCFN_USE_FEK 0x00001000
|
|
|
|
#define ECRYPTFS_UNLINK_SIGS 0x00002000
|
2011-03-15 20:54:00 +01:00
|
|
|
#define ECRYPTFS_I_SIZE_INITIALIZED 0x00004000
|
2006-10-04 11:16:22 +02:00
|
|
|
u32 flags;
|
|
|
|
unsigned int file_version;
|
|
|
|
size_t iv_bytes;
|
2010-02-11 12:09:14 +01:00
|
|
|
size_t metadata_size;
|
2006-10-04 11:16:22 +02:00
|
|
|
size_t extent_size; /* Data extent size; default is 4096 */
|
|
|
|
size_t key_size;
|
|
|
|
size_t extent_shift;
|
|
|
|
unsigned int extent_mask;
|
|
|
|
struct ecryptfs_mount_crypt_stat *mount_crypt_stat;
|
2016-01-25 03:29:33 +01:00
|
|
|
struct crypto_skcipher *tfm;
|
|
|
|
struct crypto_shash *hash_tfm; /* Crypto context for generating
|
|
|
|
* the initialization vectors */
|
2015-02-23 12:34:10 +01:00
|
|
|
unsigned char cipher[ECRYPTFS_MAX_CIPHER_NAME_SIZE + 1];
|
2006-10-04 11:16:22 +02:00
|
|
|
unsigned char key[ECRYPTFS_MAX_KEY_BYTES];
|
|
|
|
unsigned char root_iv[ECRYPTFS_MAX_IV_BYTES];
|
2007-10-16 10:27:53 +02:00
|
|
|
struct list_head keysig_list;
|
|
|
|
struct mutex keysig_list_mutex;
|
2006-10-04 11:16:22 +02:00
|
|
|
struct mutex cs_tfm_mutex;
|
2006-10-31 07:07:17 +01:00
|
|
|
struct mutex cs_hash_tfm_mutex;
|
2006-10-04 11:16:22 +02:00
|
|
|
struct mutex cs_mutex;
|
|
|
|
};
|
|
|
|
|
|
|
|
/* inode private data. */
|
|
|
|
struct ecryptfs_inode_info {
|
|
|
|
struct inode vfs_inode;
|
|
|
|
struct inode *wii_inode;
|
eCryptfs: Add reference counting to lower files
For any given lower inode, eCryptfs keeps only one lower file open and
multiplexes all eCryptfs file operations through that lower file. The
lower file was considered "persistent" and stayed open from the first
lookup through the lifetime of the inode.
This patch keeps the notion of a single, per-inode lower file, but adds
reference counting around the lower file so that it is closed when not
currently in use. If the reference count is at 0 when an operation (such
as open, create, etc.) needs to use the lower file, a new lower file is
opened. Since the file is no longer persistent, all references to the
term persistent file are changed to lower file.
Locking is added around the sections of code that opens the lower file
and assign the pointer in the inode info, as well as the code the fputs
the lower file when all eCryptfs users are done with it.
This patch is needed to fix issues, when mounted on top of the NFSv3
client, where the lower file is left silly renamed until the eCryptfs
inode is destroyed.
Signed-off-by: Tyler Hicks <tyhicks@linux.vnet.ibm.com>
2011-04-14 22:35:11 +02:00
|
|
|
struct mutex lower_file_mutex;
|
|
|
|
atomic_t lower_file_count;
|
2007-10-16 10:28:07 +02:00
|
|
|
struct file *lower_file;
|
2006-10-04 11:16:22 +02:00
|
|
|
struct ecryptfs_crypt_stat crypt_stat;
|
|
|
|
};
|
|
|
|
|
|
|
|
/* dentry private data. Each dentry must keep track of a lower
|
|
|
|
* vfsmount too. */
|
|
|
|
struct ecryptfs_dentry_info {
|
2006-12-08 11:36:34 +01:00
|
|
|
struct path lower_path;
|
2013-09-16 02:45:11 +02:00
|
|
|
union {
|
|
|
|
struct ecryptfs_crypt_stat *crypt_stat;
|
|
|
|
struct rcu_head rcu;
|
|
|
|
};
|
2006-10-04 11:16:22 +02:00
|
|
|
};
|
|
|
|
|
2007-10-16 10:28:01 +02:00
|
|
|
/**
|
2007-10-16 10:28:05 +02:00
|
|
|
* ecryptfs_global_auth_tok - A key used to encrypt all new files under the mountpoint
|
|
|
|
* @flags: Status flags
|
|
|
|
* @mount_crypt_stat_list: These auth_toks hang off the mount-wide
|
|
|
|
* cryptographic context. Every time a new
|
|
|
|
* inode comes into existence, eCryptfs copies
|
|
|
|
* the auth_toks on that list to the set of
|
|
|
|
* auth_toks on the inode's crypt_stat
|
|
|
|
* @global_auth_tok_key: The key from the user's keyring for the sig
|
|
|
|
* @global_auth_tok: The key contents
|
|
|
|
* @sig: The key identifier
|
|
|
|
*
|
2007-10-16 10:28:01 +02:00
|
|
|
* ecryptfs_global_auth_tok structs refer to authentication token keys
|
|
|
|
* in the user keyring that apply to newly created files. A list of
|
|
|
|
* these objects hangs off of the mount_crypt_stat struct for any
|
|
|
|
* given eCryptfs mount. This struct maintains a reference to both the
|
|
|
|
* key contents and the key itself so that the key can be put on
|
|
|
|
* unmount.
|
|
|
|
*/
|
2007-10-16 10:27:53 +02:00
|
|
|
struct ecryptfs_global_auth_tok {
|
|
|
|
#define ECRYPTFS_AUTH_TOK_INVALID 0x00000001
|
2009-03-13 21:51:59 +01:00
|
|
|
#define ECRYPTFS_AUTH_TOK_FNEK 0x00000002
|
2007-10-16 10:27:53 +02:00
|
|
|
u32 flags;
|
2007-10-16 10:28:05 +02:00
|
|
|
struct list_head mount_crypt_stat_list;
|
|
|
|
struct key *global_auth_tok_key;
|
|
|
|
unsigned char sig[ECRYPTFS_SIG_SIZE_HEX + 1];
|
2007-10-16 10:27:53 +02:00
|
|
|
};
|
|
|
|
|
2007-10-16 10:28:01 +02:00
|
|
|
/**
|
2007-10-16 10:28:05 +02:00
|
|
|
* ecryptfs_key_tfm - Persistent key tfm
|
|
|
|
* @key_tfm: crypto API handle to the key
|
|
|
|
* @key_size: Key size in bytes
|
|
|
|
* @key_tfm_mutex: Mutex to ensure only one operation in eCryptfs is
|
|
|
|
* using the persistent TFM at any point in time
|
|
|
|
* @key_tfm_list: Handle to hang this off the module-wide TFM list
|
|
|
|
* @cipher_name: String name for the cipher for this TFM
|
|
|
|
*
|
2007-10-16 10:28:01 +02:00
|
|
|
* Typically, eCryptfs will use the same ciphers repeatedly throughout
|
|
|
|
* the course of its operations. In order to avoid unnecessarily
|
|
|
|
* destroying and initializing the same cipher repeatedly, eCryptfs
|
|
|
|
* keeps a list of crypto API contexts around to use when needed.
|
|
|
|
*/
|
2007-10-16 10:27:53 +02:00
|
|
|
struct ecryptfs_key_tfm {
|
2016-01-25 03:29:33 +01:00
|
|
|
struct crypto_skcipher *key_tfm;
|
2007-10-16 10:27:53 +02:00
|
|
|
size_t key_size;
|
|
|
|
struct mutex key_tfm_mutex;
|
2007-10-16 10:28:05 +02:00
|
|
|
struct list_head key_tfm_list;
|
2007-10-16 10:27:53 +02:00
|
|
|
unsigned char cipher_name[ECRYPTFS_MAX_CIPHER_NAME_SIZE + 1];
|
|
|
|
};
|
|
|
|
|
2008-02-06 10:38:37 +01:00
|
|
|
extern struct mutex key_tfm_list_mutex;
|
|
|
|
|
2006-10-04 11:16:22 +02:00
|
|
|
/**
|
|
|
|
* This struct is to enable a mount-wide passphrase/salt combo. This
|
|
|
|
* is more or less a stopgap to provide similar functionality to other
|
|
|
|
* crypto filesystems like EncFS or CFS until full policy support is
|
|
|
|
* implemented in eCryptfs.
|
|
|
|
*/
|
|
|
|
struct ecryptfs_mount_crypt_stat {
|
|
|
|
/* Pointers to memory we do not own, do not free these */
|
|
|
|
#define ECRYPTFS_PLAINTEXT_PASSTHROUGH_ENABLED 0x00000001
|
[PATCH] eCryptfs: xattr flags and mount options
This patch set introduces the ability to store cryptographic metadata into an
lower file extended attribute rather than the lower file header region.
This patch set implements two new mount options:
ecryptfs_xattr_metadata
- When set, newly created files will have their cryptographic
metadata stored in the extended attribute region of the file rather
than the header.
When storing the data in the file header, there is a minimum of 8KB
reserved for the header information for each file, making each file at
least 12KB in size. This can take up a lot of extra disk space if the user
creates a lot of small files. By storing the data in the extended
attribute, each file will only occupy at least of 4KB of space.
As the eCryptfs metadata set becomes larger with new features such as
multi-key associations, most popular filesystems will not be able to store
all of the information in the xattr region in some cases due to space
constraints. However, the majority of users will only ever associate one
key per file, so most users will be okay with storing their data in the
xattr region.
This option should be used with caution. I want to emphasize that the
xattr must be maintained under all circumstances, or the file will be
rendered permanently unrecoverable. The last thing I want is for a user to
forget to set an xattr flag in a backup utility, only to later discover
that their backups are worthless.
ecryptfs_encrypted_view
- When set, this option causes eCryptfs to present applications a
view of encrypted files as if the cryptographic metadata were
stored in the file header, whether the metadata is actually stored
in the header or in the extended attributes.
No matter what eCryptfs winds up doing in the lower filesystem, I want
to preserve a baseline format compatibility for the encrypted files. As of
right now, the metadata may be in the file header or in an xattr. There is
no reason why the metadata could not be put in a separate file in future
versions.
Without the compatibility mode, backup utilities would have to know to
back up the metadata file along with the files. The semantics of eCryptfs
have always been that the lower files are self-contained units of encrypted
data, and the only additional information required to decrypt any given
eCryptfs file is the key. That is what has always been emphasized about
eCryptfs lower files, and that is what users expect. Providing the
encrypted view option will provide a way to userspace applications wherein
they can always get to the same old familiar eCryptfs encrypted files,
regardless of what eCryptfs winds up doing with the metadata behind the
scenes.
This patch:
Add extended attribute support to version bit vector, flags to indicate when
xattr or encrypted view modes are enabled, and support for the new mount
options.
Signed-off-by: Michael Halcrow <mhalcrow@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-02-12 09:53:45 +01:00
|
|
|
#define ECRYPTFS_XATTR_METADATA_ENABLED 0x00000002
|
|
|
|
#define ECRYPTFS_ENCRYPTED_VIEW_ENABLED 0x00000004
|
2007-10-16 10:27:53 +02:00
|
|
|
#define ECRYPTFS_MOUNT_CRYPT_STAT_INITIALIZED 0x00000008
|
2009-01-06 23:41:57 +01:00
|
|
|
#define ECRYPTFS_GLOBAL_ENCRYPT_FILENAMES 0x00000010
|
|
|
|
#define ECRYPTFS_GLOBAL_ENCFN_USE_MOUNT_FNEK 0x00000020
|
|
|
|
#define ECRYPTFS_GLOBAL_ENCFN_USE_FEK 0x00000040
|
2010-10-06 18:31:32 +02:00
|
|
|
#define ECRYPTFS_GLOBAL_MOUNT_AUTH_TOK_ONLY 0x00000080
|
2006-10-04 11:16:22 +02:00
|
|
|
u32 flags;
|
2007-10-16 10:27:53 +02:00
|
|
|
struct list_head global_auth_tok_list;
|
|
|
|
struct mutex global_auth_tok_list_mutex;
|
2006-10-04 11:16:22 +02:00
|
|
|
size_t global_default_cipher_key_size;
|
2009-01-06 23:41:57 +01:00
|
|
|
size_t global_default_fn_cipher_key_bytes;
|
2006-10-04 11:16:22 +02:00
|
|
|
unsigned char global_default_cipher_name[ECRYPTFS_MAX_CIPHER_NAME_SIZE
|
|
|
|
+ 1];
|
2009-01-06 23:41:57 +01:00
|
|
|
unsigned char global_default_fn_cipher_name[
|
|
|
|
ECRYPTFS_MAX_CIPHER_NAME_SIZE + 1];
|
|
|
|
char global_default_fnek_sig[ECRYPTFS_SIG_SIZE_HEX + 1];
|
2006-10-04 11:16:22 +02:00
|
|
|
};
|
|
|
|
|
|
|
|
/* superblock private data. */
|
|
|
|
struct ecryptfs_sb_info {
|
|
|
|
struct super_block *wsi_sb;
|
|
|
|
struct ecryptfs_mount_crypt_stat mount_crypt_stat;
|
2010-04-22 12:22:04 +02:00
|
|
|
struct backing_dev_info bdi;
|
2006-10-04 11:16:22 +02:00
|
|
|
};
|
|
|
|
|
|
|
|
/* file private data. */
|
|
|
|
struct ecryptfs_file_info {
|
|
|
|
struct file *wfi_file;
|
|
|
|
struct ecryptfs_crypt_stat *crypt_stat;
|
|
|
|
};
|
|
|
|
|
|
|
|
/* auth_tok <=> encrypted_session_key mappings */
|
|
|
|
struct ecryptfs_auth_tok_list_item {
|
|
|
|
unsigned char encrypted_session_key[ECRYPTFS_MAX_KEY_BYTES];
|
|
|
|
struct list_head list;
|
|
|
|
struct ecryptfs_auth_tok auth_tok;
|
|
|
|
};
|
|
|
|
|
[PATCH] eCryptfs: Public key transport mechanism
This is the transport code for public key functionality in eCryptfs. It
manages encryption/decryption request queues with a transport mechanism.
Currently, netlink is the only implemented transport.
Each inode has a unique File Encryption Key (FEK). Under passphrase, a File
Encryption Key Encryption Key (FEKEK) is generated from a salt/passphrase
combo on mount. This FEKEK encrypts each FEK and writes it into the header of
each file using the packet format specified in RFC 2440. This is all
symmetric key encryption, so it can all be done via the kernel crypto API.
These new patches introduce public key encryption of the FEK. There is no
asymmetric key encryption support in the kernel crypto API, so eCryptfs pushes
the FEK encryption and decryption out to a userspace daemon. After
considering our requirements and determining the complexity of using various
transport mechanisms, we settled on netlink for this communication.
eCryptfs stores authentication tokens into the kernel keyring. These tokens
correlate with individual keys. For passphrase mode of operation, the
authentication token contains the symmetric FEKEK. For public key, the
authentication token contains a PKI type and an opaque data blob managed by
individual PKI modules in userspace.
Each user who opens a file under an eCryptfs partition mounted in public key
mode must be running a daemon. That daemon has the user's credentials and has
access to all of the keys to which the user should have access. The daemon,
when started, initializes the pluggable PKI modules available on the system
and registers itself with the eCryptfs kernel module. Userspace utilities
register public key authentication tokens into the user session keyring.
These authentication tokens correlate key signatures with PKI modules and PKI
blobs. The PKI blobs contain PKI-specific information necessary for the PKI
module to carry out asymmetric key encryption and decryption.
When the eCryptfs module parses the header of an existing file and finds a Tag
1 (Public Key) packet (see RFC 2440), it reads in the public key identifier
(signature). The asymmetrically encrypted FEK is in the Tag 1 packet;
eCryptfs puts together a decrypt request packet containing the signature and
the encrypted FEK, then it passes it to the daemon registered for the
current->euid via a netlink unicast to the PID of the daemon, which was
registered at the time the daemon was started by the user.
The daemon actually just makes calls to libecryptfs, which implements request
packet parsing and manages PKI modules. libecryptfs grabs the public key
authentication token for the given signature from the user session keyring.
This auth tok tells libecryptfs which PKI module should receive the request.
libecryptfs then makes a decrypt() call to the PKI module, and it passes along
the PKI block from the auth tok. The PKI uses the blob to figure out how it
should decrypt the data passed to it; it performs the decryption and passes
the decrypted data back to libecryptfs. libecryptfs then puts together a
reply packet with the decrypted FEK and passes that back to the eCryptfs
module.
The eCryptfs module manages these request callouts to userspace code via
message context structs. The module maintains an array of message context
structs and places the elements of the array on two lists: a free and an
allocated list. When eCryptfs wants to make a request, it moves a msg ctx
from the free list to the allocated list, sets its state to pending, and fires
off the message to the user's registered daemon.
When eCryptfs receives a netlink message (via the callback), it correlates the
msg ctx struct in the alloc list with the data in the message itself. The
msg->index contains the offset of the array of msg ctx structs. It verifies
that the registered daemon PID is the same as the PID of the process that sent
the message. It also validates a sequence number between the received packet
and the msg ctx. Then, it copies the contents of the message (the reply
packet) into the msg ctx struct, sets the state in the msg ctx to done, and
wakes up the process that was sleeping while waiting for the reply.
The sleeping process was whatever was performing the sys_open(). This process
originally called ecryptfs_send_message(); it is now in
ecryptfs_wait_for_response(). When it wakes up and sees that the msg ctx
state was set to done, it returns a pointer to the message contents (the reply
packet) and returns. If all went well, this packet contains the decrypted
FEK, which is then copied into the crypt_stat struct, and life continues as
normal.
The case for creation of a new file is very similar, only instead of a decrypt
request, eCryptfs sends out an encrypt request.
> - We have a great clod of key mangement code in-kernel. Why is that
> not suitable (or growable) for public key management?
eCryptfs uses Howells' keyring to store persistent key data and PKI state
information. It defers public key cryptographic transformations to userspace
code. The userspace data manipulation request really is orthogonal to key
management in and of itself. What eCryptfs basically needs is a secure way to
communicate with a particular daemon for a particular task doing a syscall,
based on the UID. Nothing running under another UID should be able to access
that channel of communication.
> - Is it appropriate that new infrastructure for public key
> management be private to a particular fs?
The messaging.c file contains a lot of code that, perhaps, could be extracted
into a separate kernel service. In essence, this would be a sort of
request/reply mechanism that would involve a userspace daemon. I am not aware
of anything that does quite what eCryptfs does, so I was not aware of any
existing tools to do just what we wanted.
> What happens if one of these daemons exits without sending a quit
> message?
There is a stale uid<->pid association in the hash table for that user. When
the user registers a new daemon, eCryptfs cleans up the old association and
generates a new one. See ecryptfs_process_helo().
> - _why_ does it use netlink?
Netlink provides the transport mechanism that would minimize the complexity of
the implementation, given that we can have multiple daemons (one per user). I
explored the possibility of using relayfs, but that would involve having to
introduce control channels and a protocol for creating and tearing down
channels for the daemons. We do not have to worry about any of that with
netlink.
Signed-off-by: Michael Halcrow <mhalcrow@us.ibm.com>
Cc: David Howells <dhowells@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-02-12 09:53:43 +01:00
|
|
|
struct ecryptfs_message {
|
2008-04-29 09:59:51 +02:00
|
|
|
/* Can never be greater than ecryptfs_message_buf_len */
|
|
|
|
/* Used to find the parent msg_ctx */
|
|
|
|
/* Inherits from msg_ctx->index */
|
[PATCH] eCryptfs: Public key transport mechanism
This is the transport code for public key functionality in eCryptfs. It
manages encryption/decryption request queues with a transport mechanism.
Currently, netlink is the only implemented transport.
Each inode has a unique File Encryption Key (FEK). Under passphrase, a File
Encryption Key Encryption Key (FEKEK) is generated from a salt/passphrase
combo on mount. This FEKEK encrypts each FEK and writes it into the header of
each file using the packet format specified in RFC 2440. This is all
symmetric key encryption, so it can all be done via the kernel crypto API.
These new patches introduce public key encryption of the FEK. There is no
asymmetric key encryption support in the kernel crypto API, so eCryptfs pushes
the FEK encryption and decryption out to a userspace daemon. After
considering our requirements and determining the complexity of using various
transport mechanisms, we settled on netlink for this communication.
eCryptfs stores authentication tokens into the kernel keyring. These tokens
correlate with individual keys. For passphrase mode of operation, the
authentication token contains the symmetric FEKEK. For public key, the
authentication token contains a PKI type and an opaque data blob managed by
individual PKI modules in userspace.
Each user who opens a file under an eCryptfs partition mounted in public key
mode must be running a daemon. That daemon has the user's credentials and has
access to all of the keys to which the user should have access. The daemon,
when started, initializes the pluggable PKI modules available on the system
and registers itself with the eCryptfs kernel module. Userspace utilities
register public key authentication tokens into the user session keyring.
These authentication tokens correlate key signatures with PKI modules and PKI
blobs. The PKI blobs contain PKI-specific information necessary for the PKI
module to carry out asymmetric key encryption and decryption.
When the eCryptfs module parses the header of an existing file and finds a Tag
1 (Public Key) packet (see RFC 2440), it reads in the public key identifier
(signature). The asymmetrically encrypted FEK is in the Tag 1 packet;
eCryptfs puts together a decrypt request packet containing the signature and
the encrypted FEK, then it passes it to the daemon registered for the
current->euid via a netlink unicast to the PID of the daemon, which was
registered at the time the daemon was started by the user.
The daemon actually just makes calls to libecryptfs, which implements request
packet parsing and manages PKI modules. libecryptfs grabs the public key
authentication token for the given signature from the user session keyring.
This auth tok tells libecryptfs which PKI module should receive the request.
libecryptfs then makes a decrypt() call to the PKI module, and it passes along
the PKI block from the auth tok. The PKI uses the blob to figure out how it
should decrypt the data passed to it; it performs the decryption and passes
the decrypted data back to libecryptfs. libecryptfs then puts together a
reply packet with the decrypted FEK and passes that back to the eCryptfs
module.
The eCryptfs module manages these request callouts to userspace code via
message context structs. The module maintains an array of message context
structs and places the elements of the array on two lists: a free and an
allocated list. When eCryptfs wants to make a request, it moves a msg ctx
from the free list to the allocated list, sets its state to pending, and fires
off the message to the user's registered daemon.
When eCryptfs receives a netlink message (via the callback), it correlates the
msg ctx struct in the alloc list with the data in the message itself. The
msg->index contains the offset of the array of msg ctx structs. It verifies
that the registered daemon PID is the same as the PID of the process that sent
the message. It also validates a sequence number between the received packet
and the msg ctx. Then, it copies the contents of the message (the reply
packet) into the msg ctx struct, sets the state in the msg ctx to done, and
wakes up the process that was sleeping while waiting for the reply.
The sleeping process was whatever was performing the sys_open(). This process
originally called ecryptfs_send_message(); it is now in
ecryptfs_wait_for_response(). When it wakes up and sees that the msg ctx
state was set to done, it returns a pointer to the message contents (the reply
packet) and returns. If all went well, this packet contains the decrypted
FEK, which is then copied into the crypt_stat struct, and life continues as
normal.
The case for creation of a new file is very similar, only instead of a decrypt
request, eCryptfs sends out an encrypt request.
> - We have a great clod of key mangement code in-kernel. Why is that
> not suitable (or growable) for public key management?
eCryptfs uses Howells' keyring to store persistent key data and PKI state
information. It defers public key cryptographic transformations to userspace
code. The userspace data manipulation request really is orthogonal to key
management in and of itself. What eCryptfs basically needs is a secure way to
communicate with a particular daemon for a particular task doing a syscall,
based on the UID. Nothing running under another UID should be able to access
that channel of communication.
> - Is it appropriate that new infrastructure for public key
> management be private to a particular fs?
The messaging.c file contains a lot of code that, perhaps, could be extracted
into a separate kernel service. In essence, this would be a sort of
request/reply mechanism that would involve a userspace daemon. I am not aware
of anything that does quite what eCryptfs does, so I was not aware of any
existing tools to do just what we wanted.
> What happens if one of these daemons exits without sending a quit
> message?
There is a stale uid<->pid association in the hash table for that user. When
the user registers a new daemon, eCryptfs cleans up the old association and
generates a new one. See ecryptfs_process_helo().
> - _why_ does it use netlink?
Netlink provides the transport mechanism that would minimize the complexity of
the implementation, given that we can have multiple daemons (one per user). I
explored the possibility of using relayfs, but that would involve having to
introduce control channels and a protocol for creating and tearing down
channels for the daemons. We do not have to worry about any of that with
netlink.
Signed-off-by: Michael Halcrow <mhalcrow@us.ibm.com>
Cc: David Howells <dhowells@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-02-12 09:53:43 +01:00
|
|
|
u32 index;
|
|
|
|
u32 data_len;
|
|
|
|
u8 data[];
|
|
|
|
};
|
|
|
|
|
|
|
|
struct ecryptfs_msg_ctx {
|
2008-04-29 09:59:51 +02:00
|
|
|
#define ECRYPTFS_MSG_CTX_STATE_FREE 0x01
|
|
|
|
#define ECRYPTFS_MSG_CTX_STATE_PENDING 0x02
|
|
|
|
#define ECRYPTFS_MSG_CTX_STATE_DONE 0x03
|
|
|
|
#define ECRYPTFS_MSG_CTX_STATE_NO_REPLY 0x04
|
|
|
|
u8 state;
|
|
|
|
#define ECRYPTFS_MSG_HELO 100
|
|
|
|
#define ECRYPTFS_MSG_QUIT 101
|
|
|
|
#define ECRYPTFS_MSG_REQUEST 102
|
|
|
|
#define ECRYPTFS_MSG_RESPONSE 103
|
|
|
|
u8 type;
|
|
|
|
u32 index;
|
|
|
|
/* Counter converts to a sequence number. Each message sent
|
|
|
|
* out for which we expect a response has an associated
|
|
|
|
* sequence number. The response must have the same sequence
|
|
|
|
* number as the counter for the msg_stc for the message to be
|
|
|
|
* valid. */
|
|
|
|
u32 counter;
|
|
|
|
size_t msg_size;
|
[PATCH] eCryptfs: Public key transport mechanism
This is the transport code for public key functionality in eCryptfs. It
manages encryption/decryption request queues with a transport mechanism.
Currently, netlink is the only implemented transport.
Each inode has a unique File Encryption Key (FEK). Under passphrase, a File
Encryption Key Encryption Key (FEKEK) is generated from a salt/passphrase
combo on mount. This FEKEK encrypts each FEK and writes it into the header of
each file using the packet format specified in RFC 2440. This is all
symmetric key encryption, so it can all be done via the kernel crypto API.
These new patches introduce public key encryption of the FEK. There is no
asymmetric key encryption support in the kernel crypto API, so eCryptfs pushes
the FEK encryption and decryption out to a userspace daemon. After
considering our requirements and determining the complexity of using various
transport mechanisms, we settled on netlink for this communication.
eCryptfs stores authentication tokens into the kernel keyring. These tokens
correlate with individual keys. For passphrase mode of operation, the
authentication token contains the symmetric FEKEK. For public key, the
authentication token contains a PKI type and an opaque data blob managed by
individual PKI modules in userspace.
Each user who opens a file under an eCryptfs partition mounted in public key
mode must be running a daemon. That daemon has the user's credentials and has
access to all of the keys to which the user should have access. The daemon,
when started, initializes the pluggable PKI modules available on the system
and registers itself with the eCryptfs kernel module. Userspace utilities
register public key authentication tokens into the user session keyring.
These authentication tokens correlate key signatures with PKI modules and PKI
blobs. The PKI blobs contain PKI-specific information necessary for the PKI
module to carry out asymmetric key encryption and decryption.
When the eCryptfs module parses the header of an existing file and finds a Tag
1 (Public Key) packet (see RFC 2440), it reads in the public key identifier
(signature). The asymmetrically encrypted FEK is in the Tag 1 packet;
eCryptfs puts together a decrypt request packet containing the signature and
the encrypted FEK, then it passes it to the daemon registered for the
current->euid via a netlink unicast to the PID of the daemon, which was
registered at the time the daemon was started by the user.
The daemon actually just makes calls to libecryptfs, which implements request
packet parsing and manages PKI modules. libecryptfs grabs the public key
authentication token for the given signature from the user session keyring.
This auth tok tells libecryptfs which PKI module should receive the request.
libecryptfs then makes a decrypt() call to the PKI module, and it passes along
the PKI block from the auth tok. The PKI uses the blob to figure out how it
should decrypt the data passed to it; it performs the decryption and passes
the decrypted data back to libecryptfs. libecryptfs then puts together a
reply packet with the decrypted FEK and passes that back to the eCryptfs
module.
The eCryptfs module manages these request callouts to userspace code via
message context structs. The module maintains an array of message context
structs and places the elements of the array on two lists: a free and an
allocated list. When eCryptfs wants to make a request, it moves a msg ctx
from the free list to the allocated list, sets its state to pending, and fires
off the message to the user's registered daemon.
When eCryptfs receives a netlink message (via the callback), it correlates the
msg ctx struct in the alloc list with the data in the message itself. The
msg->index contains the offset of the array of msg ctx structs. It verifies
that the registered daemon PID is the same as the PID of the process that sent
the message. It also validates a sequence number between the received packet
and the msg ctx. Then, it copies the contents of the message (the reply
packet) into the msg ctx struct, sets the state in the msg ctx to done, and
wakes up the process that was sleeping while waiting for the reply.
The sleeping process was whatever was performing the sys_open(). This process
originally called ecryptfs_send_message(); it is now in
ecryptfs_wait_for_response(). When it wakes up and sees that the msg ctx
state was set to done, it returns a pointer to the message contents (the reply
packet) and returns. If all went well, this packet contains the decrypted
FEK, which is then copied into the crypt_stat struct, and life continues as
normal.
The case for creation of a new file is very similar, only instead of a decrypt
request, eCryptfs sends out an encrypt request.
> - We have a great clod of key mangement code in-kernel. Why is that
> not suitable (or growable) for public key management?
eCryptfs uses Howells' keyring to store persistent key data and PKI state
information. It defers public key cryptographic transformations to userspace
code. The userspace data manipulation request really is orthogonal to key
management in and of itself. What eCryptfs basically needs is a secure way to
communicate with a particular daemon for a particular task doing a syscall,
based on the UID. Nothing running under another UID should be able to access
that channel of communication.
> - Is it appropriate that new infrastructure for public key
> management be private to a particular fs?
The messaging.c file contains a lot of code that, perhaps, could be extracted
into a separate kernel service. In essence, this would be a sort of
request/reply mechanism that would involve a userspace daemon. I am not aware
of anything that does quite what eCryptfs does, so I was not aware of any
existing tools to do just what we wanted.
> What happens if one of these daemons exits without sending a quit
> message?
There is a stale uid<->pid association in the hash table for that user. When
the user registers a new daemon, eCryptfs cleans up the old association and
generates a new one. See ecryptfs_process_helo().
> - _why_ does it use netlink?
Netlink provides the transport mechanism that would minimize the complexity of
the implementation, given that we can have multiple daemons (one per user). I
explored the possibility of using relayfs, but that would involve having to
introduce control channels and a protocol for creating and tearing down
channels for the daemons. We do not have to worry about any of that with
netlink.
Signed-off-by: Michael Halcrow <mhalcrow@us.ibm.com>
Cc: David Howells <dhowells@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-02-12 09:53:43 +01:00
|
|
|
struct ecryptfs_message *msg;
|
|
|
|
struct task_struct *task;
|
|
|
|
struct list_head node;
|
2008-04-29 09:59:51 +02:00
|
|
|
struct list_head daemon_out_list;
|
[PATCH] eCryptfs: Public key transport mechanism
This is the transport code for public key functionality in eCryptfs. It
manages encryption/decryption request queues with a transport mechanism.
Currently, netlink is the only implemented transport.
Each inode has a unique File Encryption Key (FEK). Under passphrase, a File
Encryption Key Encryption Key (FEKEK) is generated from a salt/passphrase
combo on mount. This FEKEK encrypts each FEK and writes it into the header of
each file using the packet format specified in RFC 2440. This is all
symmetric key encryption, so it can all be done via the kernel crypto API.
These new patches introduce public key encryption of the FEK. There is no
asymmetric key encryption support in the kernel crypto API, so eCryptfs pushes
the FEK encryption and decryption out to a userspace daemon. After
considering our requirements and determining the complexity of using various
transport mechanisms, we settled on netlink for this communication.
eCryptfs stores authentication tokens into the kernel keyring. These tokens
correlate with individual keys. For passphrase mode of operation, the
authentication token contains the symmetric FEKEK. For public key, the
authentication token contains a PKI type and an opaque data blob managed by
individual PKI modules in userspace.
Each user who opens a file under an eCryptfs partition mounted in public key
mode must be running a daemon. That daemon has the user's credentials and has
access to all of the keys to which the user should have access. The daemon,
when started, initializes the pluggable PKI modules available on the system
and registers itself with the eCryptfs kernel module. Userspace utilities
register public key authentication tokens into the user session keyring.
These authentication tokens correlate key signatures with PKI modules and PKI
blobs. The PKI blobs contain PKI-specific information necessary for the PKI
module to carry out asymmetric key encryption and decryption.
When the eCryptfs module parses the header of an existing file and finds a Tag
1 (Public Key) packet (see RFC 2440), it reads in the public key identifier
(signature). The asymmetrically encrypted FEK is in the Tag 1 packet;
eCryptfs puts together a decrypt request packet containing the signature and
the encrypted FEK, then it passes it to the daemon registered for the
current->euid via a netlink unicast to the PID of the daemon, which was
registered at the time the daemon was started by the user.
The daemon actually just makes calls to libecryptfs, which implements request
packet parsing and manages PKI modules. libecryptfs grabs the public key
authentication token for the given signature from the user session keyring.
This auth tok tells libecryptfs which PKI module should receive the request.
libecryptfs then makes a decrypt() call to the PKI module, and it passes along
the PKI block from the auth tok. The PKI uses the blob to figure out how it
should decrypt the data passed to it; it performs the decryption and passes
the decrypted data back to libecryptfs. libecryptfs then puts together a
reply packet with the decrypted FEK and passes that back to the eCryptfs
module.
The eCryptfs module manages these request callouts to userspace code via
message context structs. The module maintains an array of message context
structs and places the elements of the array on two lists: a free and an
allocated list. When eCryptfs wants to make a request, it moves a msg ctx
from the free list to the allocated list, sets its state to pending, and fires
off the message to the user's registered daemon.
When eCryptfs receives a netlink message (via the callback), it correlates the
msg ctx struct in the alloc list with the data in the message itself. The
msg->index contains the offset of the array of msg ctx structs. It verifies
that the registered daemon PID is the same as the PID of the process that sent
the message. It also validates a sequence number between the received packet
and the msg ctx. Then, it copies the contents of the message (the reply
packet) into the msg ctx struct, sets the state in the msg ctx to done, and
wakes up the process that was sleeping while waiting for the reply.
The sleeping process was whatever was performing the sys_open(). This process
originally called ecryptfs_send_message(); it is now in
ecryptfs_wait_for_response(). When it wakes up and sees that the msg ctx
state was set to done, it returns a pointer to the message contents (the reply
packet) and returns. If all went well, this packet contains the decrypted
FEK, which is then copied into the crypt_stat struct, and life continues as
normal.
The case for creation of a new file is very similar, only instead of a decrypt
request, eCryptfs sends out an encrypt request.
> - We have a great clod of key mangement code in-kernel. Why is that
> not suitable (or growable) for public key management?
eCryptfs uses Howells' keyring to store persistent key data and PKI state
information. It defers public key cryptographic transformations to userspace
code. The userspace data manipulation request really is orthogonal to key
management in and of itself. What eCryptfs basically needs is a secure way to
communicate with a particular daemon for a particular task doing a syscall,
based on the UID. Nothing running under another UID should be able to access
that channel of communication.
> - Is it appropriate that new infrastructure for public key
> management be private to a particular fs?
The messaging.c file contains a lot of code that, perhaps, could be extracted
into a separate kernel service. In essence, this would be a sort of
request/reply mechanism that would involve a userspace daemon. I am not aware
of anything that does quite what eCryptfs does, so I was not aware of any
existing tools to do just what we wanted.
> What happens if one of these daemons exits without sending a quit
> message?
There is a stale uid<->pid association in the hash table for that user. When
the user registers a new daemon, eCryptfs cleans up the old association and
generates a new one. See ecryptfs_process_helo().
> - _why_ does it use netlink?
Netlink provides the transport mechanism that would minimize the complexity of
the implementation, given that we can have multiple daemons (one per user). I
explored the possibility of using relayfs, but that would involve having to
introduce control channels and a protocol for creating and tearing down
channels for the daemons. We do not have to worry about any of that with
netlink.
Signed-off-by: Michael Halcrow <mhalcrow@us.ibm.com>
Cc: David Howells <dhowells@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-02-12 09:53:43 +01:00
|
|
|
struct mutex mux;
|
|
|
|
};
|
|
|
|
|
2008-04-29 09:59:51 +02:00
|
|
|
struct ecryptfs_daemon {
|
|
|
|
#define ECRYPTFS_DAEMON_IN_READ 0x00000001
|
|
|
|
#define ECRYPTFS_DAEMON_IN_POLL 0x00000002
|
|
|
|
#define ECRYPTFS_DAEMON_ZOMBIE 0x00000004
|
|
|
|
#define ECRYPTFS_DAEMON_MISCDEV_OPEN 0x00000008
|
|
|
|
u32 flags;
|
|
|
|
u32 num_queued_msg_ctx;
|
2012-06-11 18:47:47 +02:00
|
|
|
struct file *file;
|
2008-04-29 09:59:51 +02:00
|
|
|
struct mutex mux;
|
|
|
|
struct list_head msg_ctx_out_queue;
|
|
|
|
wait_queue_head_t wait;
|
|
|
|
struct hlist_node euid_chain;
|
[PATCH] eCryptfs: Public key transport mechanism
This is the transport code for public key functionality in eCryptfs. It
manages encryption/decryption request queues with a transport mechanism.
Currently, netlink is the only implemented transport.
Each inode has a unique File Encryption Key (FEK). Under passphrase, a File
Encryption Key Encryption Key (FEKEK) is generated from a salt/passphrase
combo on mount. This FEKEK encrypts each FEK and writes it into the header of
each file using the packet format specified in RFC 2440. This is all
symmetric key encryption, so it can all be done via the kernel crypto API.
These new patches introduce public key encryption of the FEK. There is no
asymmetric key encryption support in the kernel crypto API, so eCryptfs pushes
the FEK encryption and decryption out to a userspace daemon. After
considering our requirements and determining the complexity of using various
transport mechanisms, we settled on netlink for this communication.
eCryptfs stores authentication tokens into the kernel keyring. These tokens
correlate with individual keys. For passphrase mode of operation, the
authentication token contains the symmetric FEKEK. For public key, the
authentication token contains a PKI type and an opaque data blob managed by
individual PKI modules in userspace.
Each user who opens a file under an eCryptfs partition mounted in public key
mode must be running a daemon. That daemon has the user's credentials and has
access to all of the keys to which the user should have access. The daemon,
when started, initializes the pluggable PKI modules available on the system
and registers itself with the eCryptfs kernel module. Userspace utilities
register public key authentication tokens into the user session keyring.
These authentication tokens correlate key signatures with PKI modules and PKI
blobs. The PKI blobs contain PKI-specific information necessary for the PKI
module to carry out asymmetric key encryption and decryption.
When the eCryptfs module parses the header of an existing file and finds a Tag
1 (Public Key) packet (see RFC 2440), it reads in the public key identifier
(signature). The asymmetrically encrypted FEK is in the Tag 1 packet;
eCryptfs puts together a decrypt request packet containing the signature and
the encrypted FEK, then it passes it to the daemon registered for the
current->euid via a netlink unicast to the PID of the daemon, which was
registered at the time the daemon was started by the user.
The daemon actually just makes calls to libecryptfs, which implements request
packet parsing and manages PKI modules. libecryptfs grabs the public key
authentication token for the given signature from the user session keyring.
This auth tok tells libecryptfs which PKI module should receive the request.
libecryptfs then makes a decrypt() call to the PKI module, and it passes along
the PKI block from the auth tok. The PKI uses the blob to figure out how it
should decrypt the data passed to it; it performs the decryption and passes
the decrypted data back to libecryptfs. libecryptfs then puts together a
reply packet with the decrypted FEK and passes that back to the eCryptfs
module.
The eCryptfs module manages these request callouts to userspace code via
message context structs. The module maintains an array of message context
structs and places the elements of the array on two lists: a free and an
allocated list. When eCryptfs wants to make a request, it moves a msg ctx
from the free list to the allocated list, sets its state to pending, and fires
off the message to the user's registered daemon.
When eCryptfs receives a netlink message (via the callback), it correlates the
msg ctx struct in the alloc list with the data in the message itself. The
msg->index contains the offset of the array of msg ctx structs. It verifies
that the registered daemon PID is the same as the PID of the process that sent
the message. It also validates a sequence number between the received packet
and the msg ctx. Then, it copies the contents of the message (the reply
packet) into the msg ctx struct, sets the state in the msg ctx to done, and
wakes up the process that was sleeping while waiting for the reply.
The sleeping process was whatever was performing the sys_open(). This process
originally called ecryptfs_send_message(); it is now in
ecryptfs_wait_for_response(). When it wakes up and sees that the msg ctx
state was set to done, it returns a pointer to the message contents (the reply
packet) and returns. If all went well, this packet contains the decrypted
FEK, which is then copied into the crypt_stat struct, and life continues as
normal.
The case for creation of a new file is very similar, only instead of a decrypt
request, eCryptfs sends out an encrypt request.
> - We have a great clod of key mangement code in-kernel. Why is that
> not suitable (or growable) for public key management?
eCryptfs uses Howells' keyring to store persistent key data and PKI state
information. It defers public key cryptographic transformations to userspace
code. The userspace data manipulation request really is orthogonal to key
management in and of itself. What eCryptfs basically needs is a secure way to
communicate with a particular daemon for a particular task doing a syscall,
based on the UID. Nothing running under another UID should be able to access
that channel of communication.
> - Is it appropriate that new infrastructure for public key
> management be private to a particular fs?
The messaging.c file contains a lot of code that, perhaps, could be extracted
into a separate kernel service. In essence, this would be a sort of
request/reply mechanism that would involve a userspace daemon. I am not aware
of anything that does quite what eCryptfs does, so I was not aware of any
existing tools to do just what we wanted.
> What happens if one of these daemons exits without sending a quit
> message?
There is a stale uid<->pid association in the hash table for that user. When
the user registers a new daemon, eCryptfs cleans up the old association and
generates a new one. See ecryptfs_process_helo().
> - _why_ does it use netlink?
Netlink provides the transport mechanism that would minimize the complexity of
the implementation, given that we can have multiple daemons (one per user). I
explored the possibility of using relayfs, but that would involve having to
introduce control channels and a protocol for creating and tearing down
channels for the daemons. We do not have to worry about any of that with
netlink.
Signed-off-by: Michael Halcrow <mhalcrow@us.ibm.com>
Cc: David Howells <dhowells@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-02-12 09:53:43 +01:00
|
|
|
};
|
|
|
|
|
2013-02-28 09:39:37 +01:00
|
|
|
#ifdef CONFIG_ECRYPT_FS_MESSAGING
|
2008-04-29 09:59:51 +02:00
|
|
|
extern struct mutex ecryptfs_daemon_hash_mux;
|
2013-02-28 09:39:37 +01:00
|
|
|
#endif
|
2008-04-29 09:59:51 +02:00
|
|
|
|
2010-02-11 14:10:38 +01:00
|
|
|
static inline size_t
|
|
|
|
ecryptfs_lower_header_size(struct ecryptfs_crypt_stat *crypt_stat)
|
|
|
|
{
|
|
|
|
if (crypt_stat->flags & ECRYPTFS_METADATA_IN_XATTR)
|
|
|
|
return 0;
|
2010-02-11 12:09:14 +01:00
|
|
|
return crypt_stat->metadata_size;
|
2010-02-11 14:10:38 +01:00
|
|
|
}
|
|
|
|
|
2006-10-04 11:16:22 +02:00
|
|
|
static inline struct ecryptfs_file_info *
|
|
|
|
ecryptfs_file_to_private(struct file *file)
|
|
|
|
{
|
2010-09-05 03:52:48 +02:00
|
|
|
return file->private_data;
|
2006-10-04 11:16:22 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
static inline void
|
|
|
|
ecryptfs_set_file_private(struct file *file,
|
|
|
|
struct ecryptfs_file_info *file_info)
|
|
|
|
{
|
|
|
|
file->private_data = file_info;
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline struct file *ecryptfs_file_to_lower(struct file *file)
|
|
|
|
{
|
|
|
|
return ((struct ecryptfs_file_info *)file->private_data)->wfi_file;
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline void
|
|
|
|
ecryptfs_set_file_lower(struct file *file, struct file *lower_file)
|
|
|
|
{
|
|
|
|
((struct ecryptfs_file_info *)file->private_data)->wfi_file =
|
|
|
|
lower_file;
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline struct ecryptfs_inode_info *
|
|
|
|
ecryptfs_inode_to_private(struct inode *inode)
|
|
|
|
{
|
|
|
|
return container_of(inode, struct ecryptfs_inode_info, vfs_inode);
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline struct inode *ecryptfs_inode_to_lower(struct inode *inode)
|
|
|
|
{
|
|
|
|
return ecryptfs_inode_to_private(inode)->wii_inode;
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline void
|
|
|
|
ecryptfs_set_inode_lower(struct inode *inode, struct inode *lower_inode)
|
|
|
|
{
|
|
|
|
ecryptfs_inode_to_private(inode)->wii_inode = lower_inode;
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline struct ecryptfs_sb_info *
|
|
|
|
ecryptfs_superblock_to_private(struct super_block *sb)
|
|
|
|
{
|
|
|
|
return (struct ecryptfs_sb_info *)sb->s_fs_info;
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline void
|
|
|
|
ecryptfs_set_superblock_private(struct super_block *sb,
|
|
|
|
struct ecryptfs_sb_info *sb_info)
|
|
|
|
{
|
|
|
|
sb->s_fs_info = sb_info;
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline struct super_block *
|
|
|
|
ecryptfs_superblock_to_lower(struct super_block *sb)
|
|
|
|
{
|
|
|
|
return ((struct ecryptfs_sb_info *)sb->s_fs_info)->wsi_sb;
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline void
|
|
|
|
ecryptfs_set_superblock_lower(struct super_block *sb,
|
|
|
|
struct super_block *lower_sb)
|
|
|
|
{
|
|
|
|
((struct ecryptfs_sb_info *)sb->s_fs_info)->wsi_sb = lower_sb;
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline struct ecryptfs_dentry_info *
|
|
|
|
ecryptfs_dentry_to_private(struct dentry *dentry)
|
|
|
|
{
|
|
|
|
return (struct ecryptfs_dentry_info *)dentry->d_fsdata;
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline void
|
|
|
|
ecryptfs_set_dentry_private(struct dentry *dentry,
|
|
|
|
struct ecryptfs_dentry_info *dentry_info)
|
|
|
|
{
|
|
|
|
dentry->d_fsdata = dentry_info;
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline struct dentry *
|
|
|
|
ecryptfs_dentry_to_lower(struct dentry *dentry)
|
|
|
|
{
|
2006-12-08 11:36:34 +01:00
|
|
|
return ((struct ecryptfs_dentry_info *)dentry->d_fsdata)->lower_path.dentry;
|
2006-10-04 11:16:22 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
static inline struct vfsmount *
|
|
|
|
ecryptfs_dentry_to_lower_mnt(struct dentry *dentry)
|
|
|
|
{
|
2006-12-08 11:36:34 +01:00
|
|
|
return ((struct ecryptfs_dentry_info *)dentry->d_fsdata)->lower_path.mnt;
|
2006-10-04 11:16:22 +02:00
|
|
|
}
|
|
|
|
|
2013-01-24 08:18:08 +01:00
|
|
|
static inline struct path *
|
|
|
|
ecryptfs_dentry_to_lower_path(struct dentry *dentry)
|
|
|
|
{
|
|
|
|
return &((struct ecryptfs_dentry_info *)dentry->d_fsdata)->lower_path;
|
|
|
|
}
|
|
|
|
|
2006-10-04 11:16:22 +02:00
|
|
|
#define ecryptfs_printk(type, fmt, arg...) \
|
2008-04-29 09:59:48 +02:00
|
|
|
__ecryptfs_printk(type "%s: " fmt, __func__, ## arg);
|
2011-11-01 01:11:33 +01:00
|
|
|
__printf(1, 2)
|
2006-10-04 11:16:22 +02:00
|
|
|
void __ecryptfs_printk(const char *fmt, ...);
|
|
|
|
|
|
|
|
extern const struct file_operations ecryptfs_main_fops;
|
|
|
|
extern const struct file_operations ecryptfs_dir_fops;
|
2007-02-12 09:55:38 +01:00
|
|
|
extern const struct inode_operations ecryptfs_main_iops;
|
|
|
|
extern const struct inode_operations ecryptfs_dir_iops;
|
|
|
|
extern const struct inode_operations ecryptfs_symlink_iops;
|
2007-02-12 09:55:41 +01:00
|
|
|
extern const struct super_operations ecryptfs_sops;
|
2009-02-20 06:57:52 +01:00
|
|
|
extern const struct dentry_operations ecryptfs_dops;
|
2009-09-22 02:01:10 +02:00
|
|
|
extern const struct address_space_operations ecryptfs_aops;
|
2006-10-04 11:16:22 +02:00
|
|
|
extern int ecryptfs_verbosity;
|
[PATCH] eCryptfs: Public key transport mechanism
This is the transport code for public key functionality in eCryptfs. It
manages encryption/decryption request queues with a transport mechanism.
Currently, netlink is the only implemented transport.
Each inode has a unique File Encryption Key (FEK). Under passphrase, a File
Encryption Key Encryption Key (FEKEK) is generated from a salt/passphrase
combo on mount. This FEKEK encrypts each FEK and writes it into the header of
each file using the packet format specified in RFC 2440. This is all
symmetric key encryption, so it can all be done via the kernel crypto API.
These new patches introduce public key encryption of the FEK. There is no
asymmetric key encryption support in the kernel crypto API, so eCryptfs pushes
the FEK encryption and decryption out to a userspace daemon. After
considering our requirements and determining the complexity of using various
transport mechanisms, we settled on netlink for this communication.
eCryptfs stores authentication tokens into the kernel keyring. These tokens
correlate with individual keys. For passphrase mode of operation, the
authentication token contains the symmetric FEKEK. For public key, the
authentication token contains a PKI type and an opaque data blob managed by
individual PKI modules in userspace.
Each user who opens a file under an eCryptfs partition mounted in public key
mode must be running a daemon. That daemon has the user's credentials and has
access to all of the keys to which the user should have access. The daemon,
when started, initializes the pluggable PKI modules available on the system
and registers itself with the eCryptfs kernel module. Userspace utilities
register public key authentication tokens into the user session keyring.
These authentication tokens correlate key signatures with PKI modules and PKI
blobs. The PKI blobs contain PKI-specific information necessary for the PKI
module to carry out asymmetric key encryption and decryption.
When the eCryptfs module parses the header of an existing file and finds a Tag
1 (Public Key) packet (see RFC 2440), it reads in the public key identifier
(signature). The asymmetrically encrypted FEK is in the Tag 1 packet;
eCryptfs puts together a decrypt request packet containing the signature and
the encrypted FEK, then it passes it to the daemon registered for the
current->euid via a netlink unicast to the PID of the daemon, which was
registered at the time the daemon was started by the user.
The daemon actually just makes calls to libecryptfs, which implements request
packet parsing and manages PKI modules. libecryptfs grabs the public key
authentication token for the given signature from the user session keyring.
This auth tok tells libecryptfs which PKI module should receive the request.
libecryptfs then makes a decrypt() call to the PKI module, and it passes along
the PKI block from the auth tok. The PKI uses the blob to figure out how it
should decrypt the data passed to it; it performs the decryption and passes
the decrypted data back to libecryptfs. libecryptfs then puts together a
reply packet with the decrypted FEK and passes that back to the eCryptfs
module.
The eCryptfs module manages these request callouts to userspace code via
message context structs. The module maintains an array of message context
structs and places the elements of the array on two lists: a free and an
allocated list. When eCryptfs wants to make a request, it moves a msg ctx
from the free list to the allocated list, sets its state to pending, and fires
off the message to the user's registered daemon.
When eCryptfs receives a netlink message (via the callback), it correlates the
msg ctx struct in the alloc list with the data in the message itself. The
msg->index contains the offset of the array of msg ctx structs. It verifies
that the registered daemon PID is the same as the PID of the process that sent
the message. It also validates a sequence number between the received packet
and the msg ctx. Then, it copies the contents of the message (the reply
packet) into the msg ctx struct, sets the state in the msg ctx to done, and
wakes up the process that was sleeping while waiting for the reply.
The sleeping process was whatever was performing the sys_open(). This process
originally called ecryptfs_send_message(); it is now in
ecryptfs_wait_for_response(). When it wakes up and sees that the msg ctx
state was set to done, it returns a pointer to the message contents (the reply
packet) and returns. If all went well, this packet contains the decrypted
FEK, which is then copied into the crypt_stat struct, and life continues as
normal.
The case for creation of a new file is very similar, only instead of a decrypt
request, eCryptfs sends out an encrypt request.
> - We have a great clod of key mangement code in-kernel. Why is that
> not suitable (or growable) for public key management?
eCryptfs uses Howells' keyring to store persistent key data and PKI state
information. It defers public key cryptographic transformations to userspace
code. The userspace data manipulation request really is orthogonal to key
management in and of itself. What eCryptfs basically needs is a secure way to
communicate with a particular daemon for a particular task doing a syscall,
based on the UID. Nothing running under another UID should be able to access
that channel of communication.
> - Is it appropriate that new infrastructure for public key
> management be private to a particular fs?
The messaging.c file contains a lot of code that, perhaps, could be extracted
into a separate kernel service. In essence, this would be a sort of
request/reply mechanism that would involve a userspace daemon. I am not aware
of anything that does quite what eCryptfs does, so I was not aware of any
existing tools to do just what we wanted.
> What happens if one of these daemons exits without sending a quit
> message?
There is a stale uid<->pid association in the hash table for that user. When
the user registers a new daemon, eCryptfs cleans up the old association and
generates a new one. See ecryptfs_process_helo().
> - _why_ does it use netlink?
Netlink provides the transport mechanism that would minimize the complexity of
the implementation, given that we can have multiple daemons (one per user). I
explored the possibility of using relayfs, but that would involve having to
introduce control channels and a protocol for creating and tearing down
channels for the daemons. We do not have to worry about any of that with
netlink.
Signed-off-by: Michael Halcrow <mhalcrow@us.ibm.com>
Cc: David Howells <dhowells@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-02-12 09:53:43 +01:00
|
|
|
extern unsigned int ecryptfs_message_buf_len;
|
|
|
|
extern signed long ecryptfs_message_wait_timeout;
|
|
|
|
extern unsigned int ecryptfs_number_of_users;
|
2006-10-04 11:16:22 +02:00
|
|
|
|
|
|
|
extern struct kmem_cache *ecryptfs_auth_tok_list_item_cache;
|
|
|
|
extern struct kmem_cache *ecryptfs_file_info_cache;
|
|
|
|
extern struct kmem_cache *ecryptfs_dentry_info_cache;
|
|
|
|
extern struct kmem_cache *ecryptfs_inode_info_cache;
|
|
|
|
extern struct kmem_cache *ecryptfs_sb_info_cache;
|
2011-05-24 12:11:12 +02:00
|
|
|
extern struct kmem_cache *ecryptfs_header_cache;
|
2007-02-12 09:53:46 +01:00
|
|
|
extern struct kmem_cache *ecryptfs_xattr_cache;
|
2007-02-16 10:28:40 +01:00
|
|
|
extern struct kmem_cache *ecryptfs_key_record_cache;
|
2007-10-16 10:27:53 +02:00
|
|
|
extern struct kmem_cache *ecryptfs_key_sig_cache;
|
|
|
|
extern struct kmem_cache *ecryptfs_global_auth_tok_cache;
|
|
|
|
extern struct kmem_cache *ecryptfs_key_tfm_cache;
|
2006-10-04 11:16:22 +02:00
|
|
|
|
2011-05-24 04:18:20 +02:00
|
|
|
struct inode *ecryptfs_get_inode(struct inode *lower_inode,
|
|
|
|
struct super_block *sb);
|
2011-03-15 20:54:00 +01:00
|
|
|
void ecryptfs_i_size_init(const char *page_virt, struct inode *inode);
|
2012-06-21 08:50:59 +02:00
|
|
|
int ecryptfs_initialize_file(struct dentry *ecryptfs_dentry,
|
|
|
|
struct inode *ecryptfs_inode);
|
2009-01-06 23:41:58 +01:00
|
|
|
int ecryptfs_decode_and_decrypt_filename(char **decrypted_name,
|
|
|
|
size_t *decrypted_name_size,
|
2013-06-16 18:05:38 +02:00
|
|
|
struct super_block *sb,
|
2009-01-06 23:41:58 +01:00
|
|
|
const char *name, size_t name_size);
|
2006-10-04 11:16:22 +02:00
|
|
|
int ecryptfs_fill_zeros(struct file *file, loff_t new_length);
|
2009-01-06 23:41:58 +01:00
|
|
|
int ecryptfs_encrypt_and_encode_filename(
|
|
|
|
char **encoded_name,
|
|
|
|
size_t *encoded_name_size,
|
|
|
|
struct ecryptfs_mount_crypt_stat *mount_crypt_stat,
|
|
|
|
const char *name, size_t name_size);
|
2006-10-04 11:16:22 +02:00
|
|
|
struct dentry *ecryptfs_lower_dentry(struct dentry *this_dentry);
|
|
|
|
void ecryptfs_dump_hex(char *data, int bytes);
|
|
|
|
int virt_to_scatterlist(const void *addr, int size, struct scatterlist *sg,
|
|
|
|
int sg_size);
|
|
|
|
int ecryptfs_compute_root_iv(struct ecryptfs_crypt_stat *crypt_stat);
|
|
|
|
void ecryptfs_rotate_iv(unsigned char *iv);
|
|
|
|
void ecryptfs_init_crypt_stat(struct ecryptfs_crypt_stat *crypt_stat);
|
2007-10-16 10:28:01 +02:00
|
|
|
void ecryptfs_destroy_crypt_stat(struct ecryptfs_crypt_stat *crypt_stat);
|
|
|
|
void ecryptfs_destroy_mount_crypt_stat(
|
2006-10-04 11:16:22 +02:00
|
|
|
struct ecryptfs_mount_crypt_stat *mount_crypt_stat);
|
|
|
|
int ecryptfs_init_crypt_ctx(struct ecryptfs_crypt_stat *crypt_stat);
|
2007-10-16 10:28:13 +02:00
|
|
|
int ecryptfs_write_inode_size_to_metadata(struct inode *ecryptfs_inode);
|
2007-10-16 10:28:08 +02:00
|
|
|
int ecryptfs_encrypt_page(struct page *page);
|
|
|
|
int ecryptfs_decrypt_page(struct page *page);
|
2011-11-22 00:31:02 +01:00
|
|
|
int ecryptfs_write_metadata(struct dentry *ecryptfs_dentry,
|
|
|
|
struct inode *ecryptfs_inode);
|
2007-10-16 10:28:10 +02:00
|
|
|
int ecryptfs_read_metadata(struct dentry *ecryptfs_dentry);
|
2011-11-22 00:31:02 +01:00
|
|
|
int ecryptfs_new_file_context(struct inode *ecryptfs_inode);
|
2010-02-11 07:02:32 +01:00
|
|
|
void ecryptfs_write_crypt_stat_flags(char *page_virt,
|
|
|
|
struct ecryptfs_crypt_stat *crypt_stat,
|
|
|
|
size_t *written);
|
2011-05-24 11:56:23 +02:00
|
|
|
int ecryptfs_read_and_validate_header_region(struct inode *inode);
|
|
|
|
int ecryptfs_read_and_validate_xattr_region(struct dentry *dentry,
|
2011-05-24 10:49:02 +02:00
|
|
|
struct inode *inode);
|
2009-01-06 23:41:57 +01:00
|
|
|
u8 ecryptfs_code_for_cipher_string(char *cipher_name, size_t key_bytes);
|
2008-02-06 10:38:36 +01:00
|
|
|
int ecryptfs_cipher_code_to_string(char *str, u8 cipher_code);
|
2006-10-04 11:16:22 +02:00
|
|
|
void ecryptfs_set_default_sizes(struct ecryptfs_crypt_stat *crypt_stat);
|
|
|
|
int ecryptfs_generate_key_packet_set(char *dest_base,
|
|
|
|
struct ecryptfs_crypt_stat *crypt_stat,
|
|
|
|
struct dentry *ecryptfs_dentry,
|
|
|
|
size_t *len, size_t max);
|
|
|
|
int
|
|
|
|
ecryptfs_parse_packet_set(struct ecryptfs_crypt_stat *crypt_stat,
|
|
|
|
unsigned char *src, struct dentry *ecryptfs_dentry);
|
|
|
|
int ecryptfs_truncate(struct dentry *dentry, loff_t new_length);
|
2007-10-16 10:28:10 +02:00
|
|
|
ssize_t
|
|
|
|
ecryptfs_getxattr_lower(struct dentry *lower_dentry, const char *name,
|
|
|
|
void *value, size_t size);
|
2007-02-12 09:53:46 +01:00
|
|
|
int
|
|
|
|
ecryptfs_setxattr(struct dentry *dentry, const char *name, const void *value,
|
|
|
|
size_t size, int flags);
|
2007-10-16 10:28:10 +02:00
|
|
|
int ecryptfs_read_xattr_region(char *page_virt, struct inode *ecryptfs_inode);
|
2013-02-28 09:39:37 +01:00
|
|
|
#ifdef CONFIG_ECRYPT_FS_MESSAGING
|
2012-06-11 18:47:47 +02:00
|
|
|
int ecryptfs_process_response(struct ecryptfs_daemon *daemon,
|
|
|
|
struct ecryptfs_message *msg, u32 seq);
|
2008-10-16 07:02:51 +02:00
|
|
|
int ecryptfs_send_message(char *data, int data_len,
|
[PATCH] eCryptfs: Public key transport mechanism
This is the transport code for public key functionality in eCryptfs. It
manages encryption/decryption request queues with a transport mechanism.
Currently, netlink is the only implemented transport.
Each inode has a unique File Encryption Key (FEK). Under passphrase, a File
Encryption Key Encryption Key (FEKEK) is generated from a salt/passphrase
combo on mount. This FEKEK encrypts each FEK and writes it into the header of
each file using the packet format specified in RFC 2440. This is all
symmetric key encryption, so it can all be done via the kernel crypto API.
These new patches introduce public key encryption of the FEK. There is no
asymmetric key encryption support in the kernel crypto API, so eCryptfs pushes
the FEK encryption and decryption out to a userspace daemon. After
considering our requirements and determining the complexity of using various
transport mechanisms, we settled on netlink for this communication.
eCryptfs stores authentication tokens into the kernel keyring. These tokens
correlate with individual keys. For passphrase mode of operation, the
authentication token contains the symmetric FEKEK. For public key, the
authentication token contains a PKI type and an opaque data blob managed by
individual PKI modules in userspace.
Each user who opens a file under an eCryptfs partition mounted in public key
mode must be running a daemon. That daemon has the user's credentials and has
access to all of the keys to which the user should have access. The daemon,
when started, initializes the pluggable PKI modules available on the system
and registers itself with the eCryptfs kernel module. Userspace utilities
register public key authentication tokens into the user session keyring.
These authentication tokens correlate key signatures with PKI modules and PKI
blobs. The PKI blobs contain PKI-specific information necessary for the PKI
module to carry out asymmetric key encryption and decryption.
When the eCryptfs module parses the header of an existing file and finds a Tag
1 (Public Key) packet (see RFC 2440), it reads in the public key identifier
(signature). The asymmetrically encrypted FEK is in the Tag 1 packet;
eCryptfs puts together a decrypt request packet containing the signature and
the encrypted FEK, then it passes it to the daemon registered for the
current->euid via a netlink unicast to the PID of the daemon, which was
registered at the time the daemon was started by the user.
The daemon actually just makes calls to libecryptfs, which implements request
packet parsing and manages PKI modules. libecryptfs grabs the public key
authentication token for the given signature from the user session keyring.
This auth tok tells libecryptfs which PKI module should receive the request.
libecryptfs then makes a decrypt() call to the PKI module, and it passes along
the PKI block from the auth tok. The PKI uses the blob to figure out how it
should decrypt the data passed to it; it performs the decryption and passes
the decrypted data back to libecryptfs. libecryptfs then puts together a
reply packet with the decrypted FEK and passes that back to the eCryptfs
module.
The eCryptfs module manages these request callouts to userspace code via
message context structs. The module maintains an array of message context
structs and places the elements of the array on two lists: a free and an
allocated list. When eCryptfs wants to make a request, it moves a msg ctx
from the free list to the allocated list, sets its state to pending, and fires
off the message to the user's registered daemon.
When eCryptfs receives a netlink message (via the callback), it correlates the
msg ctx struct in the alloc list with the data in the message itself. The
msg->index contains the offset of the array of msg ctx structs. It verifies
that the registered daemon PID is the same as the PID of the process that sent
the message. It also validates a sequence number between the received packet
and the msg ctx. Then, it copies the contents of the message (the reply
packet) into the msg ctx struct, sets the state in the msg ctx to done, and
wakes up the process that was sleeping while waiting for the reply.
The sleeping process was whatever was performing the sys_open(). This process
originally called ecryptfs_send_message(); it is now in
ecryptfs_wait_for_response(). When it wakes up and sees that the msg ctx
state was set to done, it returns a pointer to the message contents (the reply
packet) and returns. If all went well, this packet contains the decrypted
FEK, which is then copied into the crypt_stat struct, and life continues as
normal.
The case for creation of a new file is very similar, only instead of a decrypt
request, eCryptfs sends out an encrypt request.
> - We have a great clod of key mangement code in-kernel. Why is that
> not suitable (or growable) for public key management?
eCryptfs uses Howells' keyring to store persistent key data and PKI state
information. It defers public key cryptographic transformations to userspace
code. The userspace data manipulation request really is orthogonal to key
management in and of itself. What eCryptfs basically needs is a secure way to
communicate with a particular daemon for a particular task doing a syscall,
based on the UID. Nothing running under another UID should be able to access
that channel of communication.
> - Is it appropriate that new infrastructure for public key
> management be private to a particular fs?
The messaging.c file contains a lot of code that, perhaps, could be extracted
into a separate kernel service. In essence, this would be a sort of
request/reply mechanism that would involve a userspace daemon. I am not aware
of anything that does quite what eCryptfs does, so I was not aware of any
existing tools to do just what we wanted.
> What happens if one of these daemons exits without sending a quit
> message?
There is a stale uid<->pid association in the hash table for that user. When
the user registers a new daemon, eCryptfs cleans up the old association and
generates a new one. See ecryptfs_process_helo().
> - _why_ does it use netlink?
Netlink provides the transport mechanism that would minimize the complexity of
the implementation, given that we can have multiple daemons (one per user). I
explored the possibility of using relayfs, but that would involve having to
introduce control channels and a protocol for creating and tearing down
channels for the daemons. We do not have to worry about any of that with
netlink.
Signed-off-by: Michael Halcrow <mhalcrow@us.ibm.com>
Cc: David Howells <dhowells@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-02-12 09:53:43 +01:00
|
|
|
struct ecryptfs_msg_ctx **msg_ctx);
|
|
|
|
int ecryptfs_wait_for_response(struct ecryptfs_msg_ctx *msg_ctx,
|
|
|
|
struct ecryptfs_message **emsg);
|
2008-10-16 07:02:51 +02:00
|
|
|
int ecryptfs_init_messaging(void);
|
|
|
|
void ecryptfs_release_messaging(void);
|
2013-02-28 09:39:37 +01:00
|
|
|
#else
|
|
|
|
static inline int ecryptfs_init_messaging(void)
|
|
|
|
{
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
static inline void ecryptfs_release_messaging(void)
|
|
|
|
{ }
|
|
|
|
static inline int ecryptfs_send_message(char *data, int data_len,
|
|
|
|
struct ecryptfs_msg_ctx **msg_ctx)
|
|
|
|
{
|
|
|
|
return -ENOTCONN;
|
|
|
|
}
|
|
|
|
static inline int ecryptfs_wait_for_response(struct ecryptfs_msg_ctx *msg_ctx,
|
|
|
|
struct ecryptfs_message **emsg)
|
|
|
|
{
|
|
|
|
return -ENOMSG;
|
|
|
|
}
|
|
|
|
#endif
|
2008-10-16 07:02:51 +02:00
|
|
|
|
2007-02-12 09:53:47 +01:00
|
|
|
void
|
|
|
|
ecryptfs_write_header_metadata(char *virt,
|
|
|
|
struct ecryptfs_crypt_stat *crypt_stat,
|
|
|
|
size_t *written);
|
2007-10-16 10:27:53 +02:00
|
|
|
int ecryptfs_add_keysig(struct ecryptfs_crypt_stat *crypt_stat, char *sig);
|
|
|
|
int
|
|
|
|
ecryptfs_add_global_auth_tok(struct ecryptfs_mount_crypt_stat *mount_crypt_stat,
|
2009-03-13 21:51:59 +01:00
|
|
|
char *sig, u32 global_auth_tok_flags);
|
2007-10-16 10:27:53 +02:00
|
|
|
int ecryptfs_get_global_auth_tok_for_sig(
|
|
|
|
struct ecryptfs_global_auth_tok **global_auth_tok,
|
|
|
|
struct ecryptfs_mount_crypt_stat *mount_crypt_stat, char *sig);
|
|
|
|
int
|
|
|
|
ecryptfs_add_new_key_tfm(struct ecryptfs_key_tfm **key_tfm, char *cipher_name,
|
|
|
|
size_t key_size);
|
|
|
|
int ecryptfs_init_crypto(void);
|
2007-10-16 10:28:01 +02:00
|
|
|
int ecryptfs_destroy_crypto(void);
|
2008-02-06 10:38:37 +01:00
|
|
|
int ecryptfs_tfm_exists(char *cipher_name, struct ecryptfs_key_tfm **key_tfm);
|
2016-01-25 03:29:33 +01:00
|
|
|
int ecryptfs_get_tfm_and_mutex_for_cipher_name(struct crypto_skcipher **tfm,
|
2007-10-16 10:27:53 +02:00
|
|
|
struct mutex **tfm_mutex,
|
|
|
|
char *cipher_name);
|
|
|
|
int ecryptfs_keyring_auth_tok_for_sig(struct key **auth_tok_key,
|
|
|
|
struct ecryptfs_auth_tok **auth_tok,
|
|
|
|
char *sig);
|
2007-10-16 10:28:07 +02:00
|
|
|
int ecryptfs_write_lower(struct inode *ecryptfs_inode, char *data,
|
|
|
|
loff_t offset, size_t size);
|
|
|
|
int ecryptfs_write_lower_page_segment(struct inode *ecryptfs_inode,
|
|
|
|
struct page *page_for_lower,
|
|
|
|
size_t offset_in_page, size_t size);
|
2010-05-21 17:09:58 +02:00
|
|
|
int ecryptfs_write(struct inode *inode, char *data, loff_t offset, size_t size);
|
2007-10-16 10:28:07 +02:00
|
|
|
int ecryptfs_read_lower(char *data, loff_t offset, size_t size,
|
|
|
|
struct inode *ecryptfs_inode);
|
|
|
|
int ecryptfs_read_lower_page_segment(struct page *page_for_ecryptfs,
|
|
|
|
pgoff_t page_index,
|
|
|
|
size_t offset_in_page, size_t size,
|
|
|
|
struct inode *ecryptfs_inode);
|
2010-05-21 17:02:14 +02:00
|
|
|
struct page *ecryptfs_get_locked_page(struct inode *inode, loff_t index);
|
2008-04-29 09:59:51 +02:00
|
|
|
int ecryptfs_parse_packet_length(unsigned char *data, size_t *size,
|
|
|
|
size_t *length_size);
|
|
|
|
int ecryptfs_write_packet_length(char *dest, size_t size,
|
|
|
|
size_t *packet_size_length);
|
2013-02-28 09:39:37 +01:00
|
|
|
#ifdef CONFIG_ECRYPT_FS_MESSAGING
|
2008-04-29 09:59:51 +02:00
|
|
|
int ecryptfs_init_ecryptfs_miscdev(void);
|
|
|
|
void ecryptfs_destroy_ecryptfs_miscdev(void);
|
|
|
|
int ecryptfs_send_miscdev(char *data, size_t data_size,
|
|
|
|
struct ecryptfs_msg_ctx *msg_ctx, u8 msg_type,
|
|
|
|
u16 msg_flags, struct ecryptfs_daemon *daemon);
|
|
|
|
void ecryptfs_msg_ctx_alloc_to_free(struct ecryptfs_msg_ctx *msg_ctx);
|
|
|
|
int
|
2012-06-11 18:47:47 +02:00
|
|
|
ecryptfs_spawn_daemon(struct ecryptfs_daemon **daemon, struct file *file);
|
2013-02-28 09:39:37 +01:00
|
|
|
int ecryptfs_exorcise_daemon(struct ecryptfs_daemon *daemon);
|
|
|
|
int ecryptfs_find_daemon_by_euid(struct ecryptfs_daemon **daemon);
|
|
|
|
#endif
|
2008-07-24 06:30:02 +02:00
|
|
|
int ecryptfs_init_kthread(void);
|
|
|
|
void ecryptfs_destroy_kthread(void);
|
|
|
|
int ecryptfs_privileged_open(struct file **lower_file,
|
|
|
|
struct dentry *lower_dentry,
|
2008-11-14 00:39:22 +01:00
|
|
|
struct vfsmount *lower_mnt,
|
|
|
|
const struct cred *cred);
|
2011-05-24 10:49:02 +02:00
|
|
|
int ecryptfs_get_lower_file(struct dentry *dentry, struct inode *inode);
|
eCryptfs: Add reference counting to lower files
For any given lower inode, eCryptfs keeps only one lower file open and
multiplexes all eCryptfs file operations through that lower file. The
lower file was considered "persistent" and stayed open from the first
lookup through the lifetime of the inode.
This patch keeps the notion of a single, per-inode lower file, but adds
reference counting around the lower file so that it is closed when not
currently in use. If the reference count is at 0 when an operation (such
as open, create, etc.) needs to use the lower file, a new lower file is
opened. Since the file is no longer persistent, all references to the
term persistent file are changed to lower file.
Locking is added around the sections of code that opens the lower file
and assign the pointer in the inode info, as well as the code the fputs
the lower file when all eCryptfs users are done with it.
This patch is needed to fix issues, when mounted on top of the NFSv3
client, where the lower file is left silly renamed until the eCryptfs
inode is destroyed.
Signed-off-by: Tyler Hicks <tyhicks@linux.vnet.ibm.com>
2011-04-14 22:35:11 +02:00
|
|
|
void ecryptfs_put_lower_file(struct inode *inode);
|
2009-01-06 23:41:57 +01:00
|
|
|
int
|
|
|
|
ecryptfs_write_tag_70_packet(char *dest, size_t *remaining_bytes,
|
|
|
|
size_t *packet_size,
|
|
|
|
struct ecryptfs_mount_crypt_stat *mount_crypt_stat,
|
|
|
|
char *filename, size_t filename_size);
|
|
|
|
int
|
|
|
|
ecryptfs_parse_tag_70_packet(char **filename, size_t *filename_size,
|
|
|
|
size_t *packet_size,
|
|
|
|
struct ecryptfs_mount_crypt_stat *mount_crypt_stat,
|
|
|
|
char *data, size_t max_packet_size);
|
2011-11-05 18:45:08 +01:00
|
|
|
int ecryptfs_set_f_namelen(long *namelen, long lower_namelen,
|
|
|
|
struct ecryptfs_mount_crypt_stat *mount_crypt_stat);
|
2009-01-06 23:41:58 +01:00
|
|
|
int ecryptfs_derive_iv(char *iv, struct ecryptfs_crypt_stat *crypt_stat,
|
|
|
|
loff_t offset);
|
[PATCH] eCryptfs: Public key transport mechanism
This is the transport code for public key functionality in eCryptfs. It
manages encryption/decryption request queues with a transport mechanism.
Currently, netlink is the only implemented transport.
Each inode has a unique File Encryption Key (FEK). Under passphrase, a File
Encryption Key Encryption Key (FEKEK) is generated from a salt/passphrase
combo on mount. This FEKEK encrypts each FEK and writes it into the header of
each file using the packet format specified in RFC 2440. This is all
symmetric key encryption, so it can all be done via the kernel crypto API.
These new patches introduce public key encryption of the FEK. There is no
asymmetric key encryption support in the kernel crypto API, so eCryptfs pushes
the FEK encryption and decryption out to a userspace daemon. After
considering our requirements and determining the complexity of using various
transport mechanisms, we settled on netlink for this communication.
eCryptfs stores authentication tokens into the kernel keyring. These tokens
correlate with individual keys. For passphrase mode of operation, the
authentication token contains the symmetric FEKEK. For public key, the
authentication token contains a PKI type and an opaque data blob managed by
individual PKI modules in userspace.
Each user who opens a file under an eCryptfs partition mounted in public key
mode must be running a daemon. That daemon has the user's credentials and has
access to all of the keys to which the user should have access. The daemon,
when started, initializes the pluggable PKI modules available on the system
and registers itself with the eCryptfs kernel module. Userspace utilities
register public key authentication tokens into the user session keyring.
These authentication tokens correlate key signatures with PKI modules and PKI
blobs. The PKI blobs contain PKI-specific information necessary for the PKI
module to carry out asymmetric key encryption and decryption.
When the eCryptfs module parses the header of an existing file and finds a Tag
1 (Public Key) packet (see RFC 2440), it reads in the public key identifier
(signature). The asymmetrically encrypted FEK is in the Tag 1 packet;
eCryptfs puts together a decrypt request packet containing the signature and
the encrypted FEK, then it passes it to the daemon registered for the
current->euid via a netlink unicast to the PID of the daemon, which was
registered at the time the daemon was started by the user.
The daemon actually just makes calls to libecryptfs, which implements request
packet parsing and manages PKI modules. libecryptfs grabs the public key
authentication token for the given signature from the user session keyring.
This auth tok tells libecryptfs which PKI module should receive the request.
libecryptfs then makes a decrypt() call to the PKI module, and it passes along
the PKI block from the auth tok. The PKI uses the blob to figure out how it
should decrypt the data passed to it; it performs the decryption and passes
the decrypted data back to libecryptfs. libecryptfs then puts together a
reply packet with the decrypted FEK and passes that back to the eCryptfs
module.
The eCryptfs module manages these request callouts to userspace code via
message context structs. The module maintains an array of message context
structs and places the elements of the array on two lists: a free and an
allocated list. When eCryptfs wants to make a request, it moves a msg ctx
from the free list to the allocated list, sets its state to pending, and fires
off the message to the user's registered daemon.
When eCryptfs receives a netlink message (via the callback), it correlates the
msg ctx struct in the alloc list with the data in the message itself. The
msg->index contains the offset of the array of msg ctx structs. It verifies
that the registered daemon PID is the same as the PID of the process that sent
the message. It also validates a sequence number between the received packet
and the msg ctx. Then, it copies the contents of the message (the reply
packet) into the msg ctx struct, sets the state in the msg ctx to done, and
wakes up the process that was sleeping while waiting for the reply.
The sleeping process was whatever was performing the sys_open(). This process
originally called ecryptfs_send_message(); it is now in
ecryptfs_wait_for_response(). When it wakes up and sees that the msg ctx
state was set to done, it returns a pointer to the message contents (the reply
packet) and returns. If all went well, this packet contains the decrypted
FEK, which is then copied into the crypt_stat struct, and life continues as
normal.
The case for creation of a new file is very similar, only instead of a decrypt
request, eCryptfs sends out an encrypt request.
> - We have a great clod of key mangement code in-kernel. Why is that
> not suitable (or growable) for public key management?
eCryptfs uses Howells' keyring to store persistent key data and PKI state
information. It defers public key cryptographic transformations to userspace
code. The userspace data manipulation request really is orthogonal to key
management in and of itself. What eCryptfs basically needs is a secure way to
communicate with a particular daemon for a particular task doing a syscall,
based on the UID. Nothing running under another UID should be able to access
that channel of communication.
> - Is it appropriate that new infrastructure for public key
> management be private to a particular fs?
The messaging.c file contains a lot of code that, perhaps, could be extracted
into a separate kernel service. In essence, this would be a sort of
request/reply mechanism that would involve a userspace daemon. I am not aware
of anything that does quite what eCryptfs does, so I was not aware of any
existing tools to do just what we wanted.
> What happens if one of these daemons exits without sending a quit
> message?
There is a stale uid<->pid association in the hash table for that user. When
the user registers a new daemon, eCryptfs cleans up the old association and
generates a new one. See ecryptfs_process_helo().
> - _why_ does it use netlink?
Netlink provides the transport mechanism that would minimize the complexity of
the implementation, given that we can have multiple daemons (one per user). I
explored the possibility of using relayfs, but that would involve having to
introduce control channels and a protocol for creating and tearing down
channels for the daemons. We do not have to worry about any of that with
netlink.
Signed-off-by: Michael Halcrow <mhalcrow@us.ibm.com>
Cc: David Howells <dhowells@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-02-12 09:53:43 +01:00
|
|
|
|
2006-10-04 11:16:22 +02:00
|
|
|
#endif /* #ifndef ECRYPTFS_KERNEL_H */
|