* [RFC 5/8] NetLabel: SELinux support
From: paul.moore @ 2006-06-22 22:49 UTC (permalink / raw)
To: netdev; +Cc: davem
In-Reply-To: <20060622224910.885573000@flek.zko.hp.com>
[-- Attachment #1: netlabel-selinux-2.6.17 --]
[-- Type: text/plain, Size: 49211 bytes --]
Add NetLabel support to the SELinux LSM and modify the socket_post_create() LSM
hook to return an error code. The most significant part of this patch is the
addition of NetLabel hooks into the following SELinux LSM hooks:
* selinux_file_permission()
* selinux_socket_sendmsg()
* selinux_socket_post_create()
* selinux_socket_post_accept() [NEW]
* selinux_socket_sock_rcv_skb()
* selinux_socket_getpeersec_stream()
* selinux_socket_getpeersec_dgram()
The basic reasoning behind this patch is that outgoing packets are "NetLabel'd"
by labeling their socket and the NetLabel security attributes are checked via
the additional hook in selinux_socket_sock_rcv_skb(). NetLabel itself is only
a labeling mechanism, similar to filesystem extended attributes, it is up to
the SELinux enforcement mechanism to perform the actual access checks.
In addition to the changes outlined above this patch also includes some changes
to the extended bitmap (ebitmap) and multi-level security (mls) code to import
and export SELinux TE/MLS attributes into and out of NetLabel.
---
include/linux/security.h | 25 -
net/socket.c | 13
security/dummy.c | 6
security/selinux/hooks.c | 59 ++
security/selinux/include/av_inherit.h | 1
security/selinux/include/av_perm_to_string.h | 1
security/selinux/include/av_permissions.h | 1
security/selinux/include/flask.h | 1
security/selinux/include/objsec.h | 11
security/selinux/include/security.h | 2
security/selinux/include/selinux_netlabel.h | 94 ++++
security/selinux/nlmsgtab.c | 15
security/selinux/ss/Makefile | 1
security/selinux/ss/ebitmap.c | 155 +++++++
security/selinux/ss/ebitmap.h | 6
security/selinux/ss/mls.c | 160 +++++++
security/selinux/ss/mls.h | 25 +
security/selinux/ss/selinux_netlabel.c | 574 +++++++++++++++++++++++++++
security/selinux/ss/services.c | 12
security/selinux/ss/services.h | 2
20 files changed, 1134 insertions(+), 30 deletions(-)
Index: linux-2.6.17.i686-quilt/include/linux/security.h
===================================================================
--- linux-2.6.17.i686-quilt.orig/include/linux/security.h
+++ linux-2.6.17.i686-quilt/include/linux/security.h
@@ -1267,8 +1267,8 @@ struct security_operations {
int (*unix_may_send) (struct socket * sock, struct socket * other);
int (*socket_create) (int family, int type, int protocol, int kern);
- void (*socket_post_create) (struct socket * sock, int family,
- int type, int protocol, int kern);
+ int (*socket_post_create) (struct socket * sock, int family,
+ int type, int protocol, int kern);
int (*socket_bind) (struct socket * sock,
struct sockaddr * address, int addrlen);
int (*socket_connect) (struct socket * sock,
@@ -2677,13 +2677,13 @@ static inline int security_socket_create
return security_ops->socket_create(family, type, protocol, kern);
}
-static inline void security_socket_post_create(struct socket * sock,
- int family,
- int type,
- int protocol, int kern)
+static inline int security_socket_post_create(struct socket * sock,
+ int family,
+ int type,
+ int protocol, int kern)
{
- security_ops->socket_post_create(sock, family, type,
- protocol, kern);
+ return security_ops->socket_post_create(sock, family, type,
+ protocol, kern);
}
static inline int security_socket_bind(struct socket * sock,
@@ -2809,11 +2809,12 @@ static inline int security_socket_create
return 0;
}
-static inline void security_socket_post_create(struct socket * sock,
- int family,
- int type,
- int protocol, int kern)
+static inline int security_socket_post_create(struct socket * sock,
+ int family,
+ int type,
+ int protocol, int kern)
{
+ return 0;
}
static inline int security_socket_bind(struct socket * sock,
Index: linux-2.6.17.i686-quilt/net/socket.c
===================================================================
--- linux-2.6.17.i686-quilt.orig/net/socket.c
+++ linux-2.6.17.i686-quilt/net/socket.c
@@ -976,11 +976,18 @@ int sock_create_lite(int family, int typ
goto out;
}
- security_socket_post_create(sock, family, type, protocol, 1);
sock->type = type;
+ err = security_socket_post_create(sock, family, type, protocol, 1);
+ if (err)
+ goto out_release;
+
out:
*res = sock;
return err;
+out_release:
+ sock_release(sock);
+ sock = NULL;
+ goto out;
}
/* No kernel lock held - perfect */
@@ -1218,7 +1225,9 @@ static int __sock_create(int family, int
*/
module_put(net_families[family]->owner);
*res = sock;
- security_socket_post_create(sock, family, type, protocol, kern);
+ err = security_socket_post_create(sock, family, type, protocol, kern);
+ if (err)
+ goto out_release;
out:
net_family_read_unlock();
Index: linux-2.6.17.i686-quilt/security/dummy.c
===================================================================
--- linux-2.6.17.i686-quilt.orig/security/dummy.c
+++ linux-2.6.17.i686-quilt/security/dummy.c
@@ -692,10 +692,10 @@ static int dummy_socket_create (int fami
return 0;
}
-static void dummy_socket_post_create (struct socket *sock, int family, int type,
- int protocol, int kern)
+static int dummy_socket_post_create (struct socket *sock, int family, int type,
+ int protocol, int kern)
{
- return;
+ return 0;
}
static int dummy_socket_bind (struct socket *sock, struct sockaddr *address,
Index: linux-2.6.17.i686-quilt/security/selinux/hooks.c
===================================================================
--- linux-2.6.17.i686-quilt.orig/security/selinux/hooks.c
+++ linux-2.6.17.i686-quilt/security/selinux/hooks.c
@@ -12,6 +12,8 @@
* Copyright (C) 2003 Red Hat, Inc., James Morris <jmorris@redhat.com>
* Copyright (C) 2004-2005 Trusted Computer Solutions, Inc.
* <dgoeddel@trustedcs.com>
+ * Copyright (C) 2006 Hewlett-Packard Development Company, L.P.
+ * Paul Moore, <paul.moore@hp.com>
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License version 2,
@@ -74,6 +76,7 @@
#include "objsec.h"
#include "netif.h"
#include "xfrm.h"
+#include "selinux_netlabel.h"
#define XATTR_SELINUX_SUFFIX "selinux"
#define XATTR_NAME_SELINUX XATTR_SECURITY_PREFIX XATTR_SELINUX_SUFFIX
@@ -2293,6 +2296,7 @@ static int selinux_inode_listsecurity(st
static int selinux_file_permission(struct file *file, int mask)
{
+ int rc;
struct inode *inode = file->f_dentry->d_inode;
if (!mask) {
@@ -2304,8 +2308,12 @@ static int selinux_file_permission(struc
if ((file->f_flags & O_APPEND) && (mask & MAY_WRITE))
mask |= MAY_APPEND;
- return file_has_perm(current, file,
- file_mask_to_av(inode->i_mode, mask));
+ rc = file_has_perm(current, file,
+ file_mask_to_av(inode->i_mode, mask));
+ if (rc)
+ return rc;
+
+ return selinux_netlbl_inode_permission(inode, mask);
}
static int selinux_file_alloc_security(struct file *file)
@@ -2922,8 +2930,8 @@ out:
return err;
}
-static void selinux_socket_post_create(struct socket *sock, int family,
- int type, int protocol, int kern)
+static int selinux_socket_post_create(struct socket *sock, int family,
+ int type, int protocol, int kern)
{
struct inode_security_struct *isec;
struct task_security_struct *tsec;
@@ -2935,7 +2943,7 @@ static void selinux_socket_post_create(s
isec->sid = kern ? SECINITSID_KERNEL : tsec->sid;
isec->initialized = 1;
- return;
+ return selinux_netlbl_socket_create(sock, family, isec->sid);
}
/* Range of port numbers used to automatically bind.
@@ -3113,10 +3121,24 @@ static int selinux_socket_accept(struct
return 0;
}
+#ifdef CONFIG_NETLABEL
+static void selinux_socket_post_accept(struct socket *sock,
+ struct socket *newsock)
+{
+ selinux_netlbl_socket_accept(sock, newsock);
+}
+#endif /* CONFIG_NETLABEL */
+
static int selinux_socket_sendmsg(struct socket *sock, struct msghdr *msg,
int size)
{
- return socket_has_perm(current, sock, SOCKET__WRITE);
+ int rc;
+
+ rc = socket_has_perm(current, sock, SOCKET__WRITE);
+ if (rc)
+ return rc;
+
+ return selinux_netlbl_inode_permission(SOCK_INODE(sock), MAY_WRITE);
}
static int selinux_socket_recvmsg(struct socket *sock, struct msghdr *msg,
@@ -3306,10 +3328,15 @@ static int selinux_socket_sock_rcv_skb(s
err = avc_has_perm(sock_sid, port_sid,
sock_class, recv_perm, &ad);
+ if (err)
+ goto out;
}
- if (!err)
- err = selinux_xfrm_sock_rcv_skb(sock_sid, skb);
+ err = selinux_netlbl_sock_rcv_skb(sock_class, sock_sid, skb, &ad);
+ if (err)
+ goto out;
+
+ err = selinux_xfrm_sock_rcv_skb(sock_sid, skb);
out:
return err;
@@ -3333,8 +3360,9 @@ static int selinux_socket_getpeersec_str
peer_sid = ssec->peer_sid;
}
else if (isec->sclass == SECCLASS_TCP_SOCKET) {
- peer_sid = selinux_socket_getpeer_stream(sock->sk);
-
+ err = selinux_netlbl_socket_getpeersec_stream(sock, &peer_sid);
+ if (err || peer_sid == SECSID_NULL)
+ peer_sid = selinux_socket_getpeer_stream(sock->sk);
if (peer_sid == SECSID_NULL) {
err = -ENOPROTOOPT;
goto out;
@@ -3369,8 +3397,12 @@ out:
static int selinux_socket_getpeersec_dgram(struct sk_buff *skb, char **secdata, u32 *seclen)
{
- int err = 0;
- u32 peer_sid = selinux_socket_getpeer_dgram(skb);
+ int err;
+ u32 peer_sid;
+
+ err = selinux_netlbl_socket_getpeersec_dgram(skb, &peer_sid);
+ if (err || peer_sid == SECSID_NULL)
+ peer_sid = selinux_socket_getpeer_dgram(skb);
if (peer_sid == SECSID_NULL)
return -EINVAL;
@@ -4353,6 +4385,9 @@ static struct security_operations selinu
.socket_connect = selinux_socket_connect,
.socket_listen = selinux_socket_listen,
.socket_accept = selinux_socket_accept,
+#ifdef CONFIG_NETLABEL
+ .socket_post_accept = selinux_socket_post_accept,
+#endif
.socket_sendmsg = selinux_socket_sendmsg,
.socket_recvmsg = selinux_socket_recvmsg,
.socket_getsockname = selinux_socket_getsockname,
Index: linux-2.6.17.i686-quilt/security/selinux/include/av_inherit.h
===================================================================
--- linux-2.6.17.i686-quilt.orig/security/selinux/include/av_inherit.h
+++ linux-2.6.17.i686-quilt/security/selinux/include/av_inherit.h
@@ -29,3 +29,4 @@
S_(SECCLASS_NETLINK_IP6FW_SOCKET, socket, 0x00400000UL)
S_(SECCLASS_NETLINK_DNRT_SOCKET, socket, 0x00400000UL)
S_(SECCLASS_NETLINK_KOBJECT_UEVENT_SOCKET, socket, 0x00400000UL)
+ S_(SECCLASS_NETLINK_NETLABEL_SOCKET, socket, 0x00400000UL)
Index: linux-2.6.17.i686-quilt/security/selinux/include/av_perm_to_string.h
===================================================================
--- linux-2.6.17.i686-quilt.orig/security/selinux/include/av_perm_to_string.h
+++ linux-2.6.17.i686-quilt/security/selinux/include/av_perm_to_string.h
@@ -239,3 +239,4 @@
S_(SECCLASS_ASSOCIATION, ASSOCIATION__SENDTO, "sendto")
S_(SECCLASS_ASSOCIATION, ASSOCIATION__RECVFROM, "recvfrom")
S_(SECCLASS_ASSOCIATION, ASSOCIATION__SETCONTEXT, "setcontext")
+ S_(SECCLASS_NETLINK_NETLABEL_SOCKET, NETLINK_NETLABEL_SOCKET__NLMSG_WRITE, "nlmsg_write")
Index: linux-2.6.17.i686-quilt/security/selinux/include/av_permissions.h
===================================================================
--- linux-2.6.17.i686-quilt.orig/security/selinux/include/av_permissions.h
+++ linux-2.6.17.i686-quilt/security/selinux/include/av_permissions.h
@@ -933,3 +933,4 @@
#define NETLINK_KOBJECT_UEVENT_SOCKET__SEND_MSG 0x00100000UL
#define NETLINK_KOBJECT_UEVENT_SOCKET__NAME_BIND 0x00200000UL
+#define NETLINK_NETLABEL_SOCKET__NLMSG_WRITE 0x00000001UL
Index: linux-2.6.17.i686-quilt/security/selinux/include/flask.h
===================================================================
--- linux-2.6.17.i686-quilt.orig/security/selinux/include/flask.h
+++ linux-2.6.17.i686-quilt/security/selinux/include/flask.h
@@ -60,6 +60,7 @@
#define SECCLASS_NSCD 53
#define SECCLASS_ASSOCIATION 54
#define SECCLASS_NETLINK_KOBJECT_UEVENT_SOCKET 55
+#define SECCLASS_NETLINK_NETLABEL_SOCKET 56
/*
* Security identifier indices for initial entities
Index: linux-2.6.17.i686-quilt/security/selinux/include/objsec.h
===================================================================
--- linux-2.6.17.i686-quilt.orig/security/selinux/include/objsec.h
+++ linux-2.6.17.i686-quilt/security/selinux/include/objsec.h
@@ -35,6 +35,14 @@ struct task_security_struct {
u32 ptrace_sid; /* SID of ptrace parent */
};
+struct netlbl_security_struct {
+ u32 netlbl_sid; /* SID used to set the NetLabel */
+ u32 peer_sid; /* SID of the connected peer */
+ u32 req_netlbl:1, /* socket requires a NetLabel label */
+ labeled:1, /* socket is labeled with NetLabel */
+ __unused:30;
+};
+
struct inode_security_struct {
struct inode *inode; /* back pointer to inode object */
struct list_head list; /* list of inode_security_struct */
@@ -44,6 +52,9 @@ struct inode_security_struct {
unsigned char initialized; /* initialization flag */
struct semaphore sem;
unsigned char inherit; /* inherit SID from parent entry */
+#ifdef CONFIG_NETLABEL
+ struct netlbl_security_struct netlbl;
+#endif
};
struct file_security_struct {
Index: linux-2.6.17.i686-quilt/security/selinux/include/security.h
===================================================================
--- linux-2.6.17.i686-quilt.orig/security/selinux/include/security.h
+++ linux-2.6.17.i686-quilt/security/selinux/include/security.h
@@ -8,6 +8,8 @@
#ifndef _SELINUX_SECURITY_H_
#define _SELINUX_SECURITY_H_
+#include <linux/skbuff.h>
+
#include "flask.h"
#define SECSID_NULL 0x00000000 /* unspecified SID */
Index: linux-2.6.17.i686-quilt/security/selinux/include/selinux_netlabel.h
===================================================================
--- /dev/null
+++ linux-2.6.17.i686-quilt/security/selinux/include/selinux_netlabel.h
@@ -0,0 +1,94 @@
+/*
+ * SELinux interface to the NetLabel subsystem
+ *
+ * Author : Paul Moore <paul.moore@hp.com>
+ *
+ */
+
+/*
+ * (c) Copyright Hewlett-Packard Development Company, L.P., 2006
+ *
+ * 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 _SELINUX_NETLABEL_H_
+#define _SELINUX_NETLABEL_H_
+
+#ifdef CONFIG_NETLABEL
+void selinux_netlbl_cache_invalidate(void);
+
+int selinux_netlbl_inode_permission(struct inode *inode, int mask);
+
+int selinux_netlbl_socket_create(struct socket *sock,
+ const int sock_family,
+ u32 sid);
+void selinux_netlbl_socket_accept(struct socket *sock, struct socket *newsock);
+int selinux_netlbl_sock_rcv_skb(const u16 sock_class,
+ const u32 sock_sid,
+ struct sk_buff *skb,
+ struct avc_audit_data *ad);
+int selinux_netlbl_socket_getpeersec_stream(struct socket *sock,
+ u32 *peer_sid);
+int selinux_netlbl_socket_getpeersec_dgram(struct sk_buff *skb, u32 *peer_sid);
+#else
+static inline void selinux_netlbl_cache_invalidate(void)
+{
+ return;
+}
+
+static inline int selinux_netlbl_inode_permission(struct inode *inode,
+ int mask)
+{
+ return 0;
+}
+
+static inline int selinux_netlbl_socket_create(struct socket *sock,
+ const int sock_family,
+ u32 sid)
+{
+ return 0;
+}
+
+static inline void selinux_netlbl_socket_accept(struct socket *sock,
+ struct socket *newsock)
+{
+ return;
+}
+
+static inline int selinux_netlbl_sock_rcv_skb(const u16 sock_class,
+ const u32 sock_sid,
+ struct sk_buff *skb,
+ struct avc_audit_data *ad)
+{
+ return 0;
+}
+
+static inline int selinux_netlbl_socket_getpeersec_stream(struct socket *sock,
+ u32 *peer_sid)
+{
+ *peer_sid = SECSID_NULL;
+ return 0;
+}
+
+static inline int selinux_netlbl_socket_getpeersec_dgram(struct sk_buff *skb,
+ u32 *peer_sid)
+{
+ *peer_sid = SECSID_NULL;
+ return 0;
+}
+#endif /* CONFIG_NETLABEL */
+
+#endif
Index: linux-2.6.17.i686-quilt/security/selinux/nlmsgtab.c
===================================================================
--- linux-2.6.17.i686-quilt.orig/security/selinux/nlmsgtab.c
+++ linux-2.6.17.i686-quilt/security/selinux/nlmsgtab.c
@@ -19,6 +19,7 @@
#include <linux/inet_diag.h>
#include <linux/xfrm.h>
#include <linux/audit.h>
+#include <net/netlabel.h>
#include "flask.h"
#include "av_permissions.h"
@@ -116,6 +117,15 @@ static struct nlmsg_perm nlmsg_audit_per
{ AUDIT_WATCH_LIST, NETLINK_AUDIT_SOCKET__NLMSG_READPRIV },
};
+static struct nlmsg_perm nlmsg_netlabel_perms[] =
+{
+ { NETLBL_NLTYPE_NONE, NETLINK_NETLABEL_SOCKET__NLMSG_WRITE },
+ { NETLBL_NLTYPE_MGMT, NETLINK_NETLABEL_SOCKET__NLMSG_WRITE },
+ { NETLBL_NLTYPE_RIPSO, NETLINK_NETLABEL_SOCKET__NLMSG_WRITE },
+ { NETLBL_NLTYPE_CIPSOV4, NETLINK_NETLABEL_SOCKET__NLMSG_WRITE },
+ { NETLBL_NLTYPE_CIPSOV6, NETLINK_NETLABEL_SOCKET__NLMSG_WRITE },
+ { NETLBL_NLTYPE_UNLABELED, NETLINK_NETLABEL_SOCKET__NLMSG_WRITE },
+};
static int nlmsg_perm(u16 nlmsg_type, u32 *perm, struct nlmsg_perm *tab, size_t tabsize)
{
@@ -169,6 +179,11 @@ int selinux_nlmsg_lookup(u16 sclass, u16
}
break;
+ case SECCLASS_NETLINK_NETLABEL_SOCKET:
+ err = nlmsg_perm(nlmsg_type, perm, nlmsg_netlabel_perms,
+ sizeof(nlmsg_netlabel_perms));
+ break;
+
/* No messaging from userspace, or class unknown/unhandled */
default:
err = -ENOENT;
Index: linux-2.6.17.i686-quilt/security/selinux/ss/Makefile
===================================================================
--- linux-2.6.17.i686-quilt.orig/security/selinux/ss/Makefile
+++ linux-2.6.17.i686-quilt/security/selinux/ss/Makefile
@@ -7,3 +7,4 @@ obj-y := ss.o
ss-y := ebitmap.o hashtab.o symtab.o sidtab.o avtab.o policydb.o services.o conditional.o mls.o
+obj-$(CONFIG_NETLABEL) += selinux_netlabel.o
Index: linux-2.6.17.i686-quilt/security/selinux/ss/ebitmap.c
===================================================================
--- linux-2.6.17.i686-quilt.orig/security/selinux/ss/ebitmap.c
+++ linux-2.6.17.i686-quilt/security/selinux/ss/ebitmap.c
@@ -3,6 +3,14 @@
*
* Author : Stephen Smalley, <sds@epoch.ncsc.mil>
*/
+/*
+ * Updated: Hewlett-Packard <paul.moore@hp.com>
+ *
+ * Added ebitmap_export() and ebitmap_import()
+ *
+ * (c) Copyright Hewlett-Packard Development Company, L.P., 2006
+ */
+
#include <linux/kernel.h>
#include <linux/slab.h>
#include <linux/errno.h>
@@ -59,6 +67,153 @@ int ebitmap_cpy(struct ebitmap *dst, str
return 0;
}
+/**
+ * ebitmap_export - Export an ebitmap to a unsigned char bitmap string
+ * @src: the ebitmap to export
+ * @dst: the resulting bitmap string
+ * @dst_len: length of dst in bytes
+ *
+ * Description:
+ * Allocate a buffer at least src->highbit bits long and export the extensible
+ * bitmap into the buffer. The bitmap string will be in little endian format,
+ * i.e. LSB first. The value returned in dst_len may not the true size of the
+ * buffer as the length of the buffer is rounded up to a multiple of MAPTYPE.
+ * The caller must free the buffer when finished. Returns zero on success,
+ * negative values on failure.
+ *
+ */
+int ebitmap_export(const struct ebitmap *src,
+ unsigned char **dst,
+ u32 *dst_len)
+{
+ u32 bitmap_len;
+ unsigned char *bitmap;
+ struct ebitmap_node *iter_node;
+ MAPTYPE node_val;
+ u32 bitmap_byte;
+ unsigned char bitmask;
+
+ if (src == NULL || dst == NULL || dst_len == NULL)
+ return -EINVAL;
+
+ *dst = NULL;
+ *dst_len = 0;
+
+ bitmap_len = src->highbit / 8;
+ if (src->highbit % 8 > 0)
+ bitmap_len += 1;
+ if (bitmap_len == 0)
+ return -EINVAL;
+
+ bitmap = kzalloc(bitmap_len + sizeof(MAPTYPE) -
+ (bitmap_len % sizeof(MAPTYPE)),
+ GFP_ATOMIC);
+ if (bitmap == NULL)
+ return -ENOMEM;
+
+ /* PM - there _has_ to be a faster way to do this, work on this more */
+ iter_node = src->node;
+ do {
+ bitmap_byte = iter_node->startbit / 8;
+ bitmask = 0x80;
+ node_val = iter_node->map;
+ do {
+ if (bitmask == 0) {
+ bitmap_byte++;
+ bitmask = 0x80;
+ }
+ if (node_val & (MAPTYPE)0x01)
+ bitmap[bitmap_byte] |= bitmask;
+ node_val >>= 1;
+ bitmask >>= 1;
+ } while (node_val > 0);
+ iter_node = iter_node->next;
+ } while (iter_node);
+
+ *dst = bitmap;
+ *dst_len = bitmap_len;
+ return 0;
+}
+
+/**
+ * ebitmap_import - Import an unsigned char bitmap string into an ebitmap
+ * @src: the bitmap string
+ * @src_len: the bitmap length in bytes
+ * @dst: the empty ebitmap
+ *
+ * Description:
+ * This function takes a little endian bitmap string in src and imports it into
+ * the ebitmap pointed to by dst. Returns zero on success, negative values on
+ * failure.
+ *
+ */
+int ebitmap_import(const unsigned char *src,
+ const u32 src_len,
+ struct ebitmap *dst)
+{
+ u32 src_off = 0;
+ struct ebitmap_node *node_new;
+ struct ebitmap_node *node_last = NULL;
+ u32 src_rem = src_len;
+ MAPTYPE tmp_val;
+ u32 iter;
+ u32 iter_bit;
+ unsigned char src_byte;
+
+ if (src == NULL || dst == NULL || src_len == 0)
+ return -EINVAL;
+
+ do {
+ node_new = kzalloc(sizeof(*node_new), GFP_ATOMIC);
+ if (node_new == NULL) {
+ ebitmap_destroy(dst);
+ return -ENOMEM;
+ }
+
+ /* PM - there _has_ to be a faster way to do this,
+ work on this more */
+ if (src_rem >= sizeof(MAPTYPE))
+ iter = sizeof(MAPTYPE);
+ else
+ iter = src_rem;
+ tmp_val = 0;
+ while (iter > 0) {
+ src_byte = src[src_off + --iter];
+ if (src_byte > 0)
+ for (iter_bit = 0; iter_bit < 8; iter_bit++) {
+ tmp_val <<= 1;
+ tmp_val |= src_byte & 0x01;
+ src_byte >>= 1;
+ }
+ else
+ tmp_val <<= 8;
+ }
+ node_new->map = tmp_val;
+ node_new->startbit = src_off * 8;
+
+ if (node_last != NULL)
+ node_last->next = node_new;
+ else
+ dst->node = node_new;
+ node_last = node_new;
+
+ if (src_rem >= sizeof(MAPTYPE)) {
+ src_off += sizeof(MAPTYPE);
+ src_rem -= sizeof(MAPTYPE);
+ } else
+ src_off += src_rem;
+ } while (src_off < src_len);
+
+ tmp_val = node_last->map;
+ dst->highbit = node_last->startbit;
+ while (tmp_val >= 1) {
+ dst->highbit += 1;
+ tmp_val >>= 1;
+ }
+
+ return 0;
+}
+
int ebitmap_contains(struct ebitmap *e1, struct ebitmap *e2)
{
struct ebitmap_node *n1, *n2;
Index: linux-2.6.17.i686-quilt/security/selinux/ss/ebitmap.h
===================================================================
--- linux-2.6.17.i686-quilt.orig/security/selinux/ss/ebitmap.h
+++ linux-2.6.17.i686-quilt/security/selinux/ss/ebitmap.h
@@ -69,6 +69,12 @@ static inline int ebitmap_node_get_bit(s
int ebitmap_cmp(struct ebitmap *e1, struct ebitmap *e2);
int ebitmap_cpy(struct ebitmap *dst, struct ebitmap *src);
+int ebitmap_export(const struct ebitmap *src,
+ unsigned char **dst,
+ u32 *dst_len);
+int ebitmap_import(const unsigned char *src,
+ const u32 src_len,
+ struct ebitmap *dst);
int ebitmap_contains(struct ebitmap *e1, struct ebitmap *e2);
int ebitmap_get_bit(struct ebitmap *e, unsigned long bit);
int ebitmap_set_bit(struct ebitmap *e, unsigned long bit, int value);
Index: linux-2.6.17.i686-quilt/security/selinux/ss/mls.c
===================================================================
--- linux-2.6.17.i686-quilt.orig/security/selinux/ss/mls.c
+++ linux-2.6.17.i686-quilt/security/selinux/ss/mls.c
@@ -10,6 +10,13 @@
*
* Copyright (C) 2004-2006 Trusted Computer Solutions, Inc.
*/
+/*
+ * Updated: Hewlett-Packard <paul.moore@hp.com>
+ *
+ * Added support to import/export the MLS label
+ *
+ * (c) Copyright Hewlett-Packard Development Company, L.P., 2006
+ */
#include <linux/kernel.h>
#include <linux/slab.h>
@@ -585,3 +592,156 @@ int mls_compute_sid(struct context *scon
return -EINVAL;
}
+/**
+ * mls_export_lvl - Export the MLS sensitivity levels
+ * @context: the security context
+ * @lvl_low: the low sensitivity level
+ * @lvl_high: the high sensitivity level
+ *
+ * Description:
+ * Given the security context copy the low MLS sensitivity level into lvl_low
+ * and the high sensitivity level in lvl_high. The MLS levels are only
+ * exported if the pointers are not NULL, if they are NULL then that level is
+ * not exported. Returns zero on success, negative values on failure.
+ *
+ */
+int mls_export_lvl(const struct context *context, u32 *lvl_low, u32 *lvl_high)
+{
+ if (!selinux_mls_enabled)
+ return 0;
+
+ if (lvl_low != NULL)
+ *lvl_low = context->range.level[0].sens - 1;
+ if (lvl_high != NULL)
+ *lvl_high = context->range.level[1].sens - 1;
+
+ return 0;
+}
+
+/**
+ * mls_import_lvl - Import the MLS sensitivity levels
+ * @context: the security context
+ * @lvl_low: the low sensitivity level
+ * @lvl_high: the high sensitivity level
+ *
+ * Description:
+ * Given the security context and the two sensitivty levels, set the MLS levels
+ * in the context according the two given as parameters. Returns zero on
+ * success, negative values on failure.
+ *
+ */
+int mls_import_lvl(struct context *context,
+ const u32 lvl_low,
+ const u32 lvl_high)
+{
+ if (!selinux_mls_enabled)
+ return 0;
+
+ context->range.level[0].sens = lvl_low + 1;
+ context->range.level[1].sens = lvl_high + 1;
+
+ return 0;
+}
+
+/**
+ * mls_export_cat - Export the MLS categories
+ * @context: the security context
+ * @cat_low: the low category
+ * @cat_low_len: length of the cat_low bitmap in bytes
+ * @cat_high: the high category
+ * @cat_high_len: length of the cat_high bitmap in bytes
+ *
+ * Description:
+ * Given the security context export the low MLS category bitmap into cat_low
+ * and the high category bitmap into cat_high. The MLS categories are only
+ * exported if the pointers are not NULL, if they are NULL then that level is
+ * not exported. The caller is responsibile for freeing the memory when
+ * finished. Returns zero on success, negative values on failure.
+ *
+ */
+int mls_export_cat(const struct context *context,
+ unsigned char **cat_low,
+ u32 *cat_low_len,
+ unsigned char **cat_high,
+ u32 *cat_high_len)
+{
+ int ret_val = -EPERM;
+
+ if (!selinux_mls_enabled)
+ return 0;
+
+ if (cat_low != NULL && cat_low_len != NULL) {
+ ret_val = ebitmap_export(&context->range.level[0].cat,
+ cat_low,
+ cat_low_len);
+ if (ret_val != 0)
+ goto export_cat_failure;
+ }
+ if (cat_high != NULL && cat_high_len != NULL) {
+ ret_val = ebitmap_export(&context->range.level[1].cat,
+ cat_high,
+ cat_high_len);
+ if (ret_val != 0)
+ goto export_cat_failure;
+ }
+
+ return 0;
+
+export_cat_failure:
+ if (cat_low != NULL && *cat_low != NULL)
+ kfree(*cat_low);
+ if (cat_high != NULL && *cat_high != NULL)
+ kfree(*cat_high);
+ return ret_val;
+}
+
+/**
+ * mls_import_cat - Import the MLS categories
+ * @context: the security context
+ * @cat_low: the low category
+ * @cat_low_len: length of the cat_low bitmap in bytes
+ * @cat_high: the high category
+ * @cat_high_len: length of the cat_high bitmap in bytes
+ *
+ * Description:
+ * Given the security context and the two category bitmap strings import the
+ * categories into the security context. The MLS categories are only imported
+ * if the pointers are not NULL, if they are NULL they are skipped. Returns
+ * zero on success, negative values on failure.
+ *
+ */
+int mls_import_cat(struct context *context,
+ const unsigned char *cat_low,
+ const u32 cat_low_len,
+ const unsigned char *cat_high,
+ const u32 cat_high_len)
+{
+ int ret_val = -EPERM;
+
+ if (!selinux_mls_enabled)
+ return 0;
+
+ if (cat_low != NULL && cat_low_len > 0) {
+ ret_val = ebitmap_import(cat_low,
+ cat_low_len,
+ &context->range.level[0].cat);
+ if (ret_val != 0)
+ goto import_cat_failure;
+ }
+ if (cat_high != NULL && cat_high_len > 0) {
+ ret_val = ebitmap_import(cat_high,
+ cat_high_len,
+ &context->range.level[1].cat);
+ if (ret_val != 0)
+ goto import_cat_failure;
+ }
+
+ return 0;
+
+import_cat_failure:
+ if (cat_low)
+ ebitmap_destroy(&context->range.level[0].cat);
+ if (cat_high)
+ ebitmap_destroy(&context->range.level[1].cat);
+ return ret_val;
+}
Index: linux-2.6.17.i686-quilt/security/selinux/ss/mls.h
===================================================================
--- linux-2.6.17.i686-quilt.orig/security/selinux/ss/mls.h
+++ linux-2.6.17.i686-quilt/security/selinux/ss/mls.h
@@ -10,6 +10,13 @@
*
* Copyright (C) 2004-2006 Trusted Computer Solutions, Inc.
*/
+/*
+ * Updated: Hewlett-Packard <paul.moore@hp.com>
+ *
+ * Added support to import/export the MLS label
+ *
+ * (c) Copyright Hewlett-Packard Development Company, L.P., 2006
+ */
#ifndef _SS_MLS_H_
#define _SS_MLS_H_
@@ -42,5 +49,23 @@ int mls_compute_sid(struct context *scon
int mls_setup_user_range(struct context *fromcon, struct user_datum *user,
struct context *usercon);
+int mls_export_lvl(const struct context *context,
+ u32 *lvl_low,
+ u32 *lvl_high);
+int mls_import_lvl(struct context *context,
+ const u32 lvl_low,
+ const u32 lvl_high);
+
+int mls_export_cat(const struct context *context,
+ unsigned char **cat_low,
+ u32 *cat_low_len,
+ unsigned char **cat_high,
+ u32 *cat_high_len);
+int mls_import_cat(struct context *context,
+ const unsigned char *cat_low,
+ const u32 cat_low_len,
+ const unsigned char *cat_high,
+ const u32 cat_high_len);
+
#endif /* _SS_MLS_H */
Index: linux-2.6.17.i686-quilt/security/selinux/ss/selinux_netlabel.c
===================================================================
--- /dev/null
+++ linux-2.6.17.i686-quilt/security/selinux/ss/selinux_netlabel.c
@@ -0,0 +1,574 @@
+/*
+ * SELinux interface to the NetLabel subsystem
+ *
+ * Author : Paul Moore <paul.moore@hp.com>
+ *
+ */
+
+/*
+ * (c) Copyright Hewlett-Packard Development Company, L.P., 2006
+ *
+ * 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
+ *
+ */
+
+#include <linux/kernel.h>
+#include <linux/errno.h>
+#include <linux/audit.h>
+#include <net/sock.h>
+#include <net/netlabel.h>
+
+#include "avc.h"
+#include "context.h"
+#include "sidtab.h"
+#include "services.h"
+#include "mls.h"
+#include "objsec.h"
+
+/*
+ * This is the structure we store inside the NetLabel cache block.
+ */
+#define NETLBL_CACHE(x) ((struct netlbl_cache *)(x))
+#define NETLBL_CACHE_T_NONE 0
+#define NETLBL_CACHE_T_SID 1
+#define NETLBL_CACHE_T_MLSLBL 2
+struct netlbl_cache {
+ u32 type;
+ union {
+ u32 sid;
+ struct mls_level mls_label;
+ } data;
+};
+
+/**
+ * selinux_netlbl_cache_free - Free the NetLabel cached data
+ * @data: the data to free
+ *
+ * Description:
+ * This function is intended to be used as the free() callback inside the
+ * netlbl_lsm_cache structure.
+ *
+ */
+static void selinux_netlbl_cache_free(const void *data)
+{
+ struct netlbl_cache *cache = NETLBL_CACHE(data);
+ switch (cache->type) {
+ case NETLBL_CACHE_T_MLSLBL:
+ ebitmap_destroy(&cache->data.mls_label.cat);
+ break;
+ }
+ kfree(data);
+}
+
+/**
+ * selinux_netlbl_cache_add - Add an entry to the NetLabel cache
+ * @skb: the packet
+ * @ctx: the SELinux context
+ *
+ * Description:
+ * Attempt to cache the context in @ctx, which was derived from the packet in
+ * @skb, in the NetLabel subsystem cache.
+ *
+ */
+static void selinux_netlbl_cache_add(struct sk_buff *skb, struct context *ctx)
+{
+ struct netlbl_cache *cache = NULL;
+ struct netlbl_lsm_secattr secattr;
+
+ netlbl_secattr_init(&secattr);
+
+ cache = kzalloc(sizeof(*cache), GFP_ATOMIC);
+ if (cache == NULL)
+ goto netlbl_cache_add_return;
+ secattr.cache.free = selinux_netlbl_cache_free;
+ secattr.cache.data = (void *)cache;
+ secattr.set_cache = 1;
+
+ if (ebitmap_cpy(&cache->data.mls_label.cat,
+ &ctx->range.level[0].cat) != 0)
+ goto netlbl_cache_add_return;
+ cache->data.mls_label.sens = ctx->range.level[0].sens;
+ cache->type = NETLBL_CACHE_T_MLSLBL;
+
+ if (netlbl_cache_add(skb, &secattr) != 0)
+ goto netlbl_cache_add_return;
+ secattr.set_cache = 0;
+
+netlbl_cache_add_return:
+ netlbl_secattr_destroy(&secattr);
+}
+
+/**
+ * selinux_netlbl_cache_invalidate - Invalidate the NetLabel cache
+ *
+ * Description:
+ * Invalidate the NetLabel security attribute mapping cache.
+ *
+ */
+void selinux_netlbl_cache_invalidate(void)
+{
+ netlbl_cache_invalidate();
+}
+
+/**
+ * selinux_netlbl_domain_export - Exports the type of a given context
+ * @context: the security context
+ * @scontext: the resulting type string
+ * @scontext_len: the length of scontext including the NULL byte
+ *
+ * Description:
+ * Allocate a buffer for the type name specified in context and copy the type
+ * name into the buffer. The caller must free the buffer when finished.
+ * Returns zero on success, negative values on failure.
+ *
+ */
+static int selinux_netlbl_domain_export(const struct context *context,
+ char **scontext,
+ u32 *scontext_len)
+{
+ char *str;
+ u32 str_len;
+
+ *scontext = NULL;
+ if (scontext_len != NULL)
+ *scontext_len = 0;
+ str_len = strlen(policydb.p_type_val_to_name[context->type - 1]) + 1;
+ str = kmalloc(str_len, GFP_ATOMIC);
+ if (str == NULL)
+ return -ENOMEM;
+ strcpy(str, policydb.p_type_val_to_name[context->type - 1]);
+
+ *scontext = str;
+ if (scontext_len != NULL)
+ *scontext_len = str_len;
+ return 0;
+}
+
+/**
+ * selinux_netlbl_secattr_to_sid - Convert a NetLabel secattr to a SELinux SID
+ * @skb: the network packet
+ * @secattr: the NetLabel packet security attributes
+ * @base_sid: the SELinux SID to use as a context for MLS only attributes
+ * @sid: the SELinux SID
+ *
+ * Description:
+ * Convert the given NetLabel packet security attributes in @secattr into a
+ * SELinux SID. If the @secattr field does not contain a full SELinux
+ * SID/context then use the context in @base_sid as the foundation. If @skb
+ * is not NULL attempt to cache as much data as possibile. Returns zero on
+ * success, negative values on failure.
+ *
+ */
+static int selinux_netlbl_secattr_to_sid(struct sk_buff *skb,
+ struct netlbl_lsm_secattr *secattr,
+ const u32 base_sid,
+ u32 *sid)
+{
+ int ret_val = -EIDRM;
+ struct context *ctx;
+ struct context ctx_new;
+ struct netlbl_cache *cache;
+ u32 ctx_new_destroy = 0;
+
+ if (secattr->set_cache) {
+ cache = NETLBL_CACHE(secattr->cache.data);
+ switch (cache->type) {
+ case NETLBL_CACHE_T_SID:
+ *sid = cache->data.sid;
+ break;
+ case NETLBL_CACHE_T_MLSLBL:
+ ctx = sidtab_search(&sidtab, base_sid);
+ if (ctx == NULL)
+ goto netlbl_secattr_to_sid_failure;
+ ret_val = context_cpy(&ctx_new, ctx);
+ if (ret_val != 0)
+ goto netlbl_secattr_to_sid_failure;
+ ctx_new_destroy = 1;
+ mls_context_destroy(&ctx_new);
+
+ ctx_new.range.level[0].sens =
+ cache->data.mls_label.sens;
+ ret_val = ebitmap_cpy(&ctx_new.range.level[0].cat,
+ &cache->data.mls_label.cat);
+ if (ret_val != 0)
+ goto netlbl_secattr_to_sid_failure;
+ ctx_new.range.level[1].sens =
+ cache->data.mls_label.sens;
+ ret_val = ebitmap_cpy(&ctx_new.range.level[1].cat,
+ &cache->data.mls_label.cat);
+ if (ret_val != 0)
+ goto netlbl_secattr_to_sid_failure;
+
+ ret_val = sidtab_context_to_sid(&sidtab,
+ &ctx_new,
+ sid);
+ break;
+ default:
+ goto netlbl_secattr_to_sid_failure;
+ }
+ } else if (secattr->set_mls_lvl) {
+ ctx = sidtab_search(&sidtab, base_sid);
+ if (ctx == NULL)
+ goto netlbl_secattr_to_sid_failure;
+ ret_val = context_cpy(&ctx_new, ctx);
+ if (ret_val != 0)
+ goto netlbl_secattr_to_sid_failure;
+ ctx_new_destroy = 1;
+ mls_context_destroy(&ctx_new);
+
+ if (mls_import_lvl(&ctx_new,
+ secattr->mls_lvl,
+ secattr->mls_lvl) != 0)
+ goto netlbl_secattr_to_sid_failure;
+ if (secattr->set_mls_cat) {
+ if (mls_import_cat(&ctx_new,
+ secattr->mls_cat,
+ secattr->mls_cat_len,
+ secattr->mls_cat,
+ secattr->mls_cat_len) != 0)
+ goto netlbl_secattr_to_sid_failure;
+ }
+
+ ret_val = mls_context_isvalid(&policydb, &ctx_new);
+ if (ret_val != 1)
+ goto netlbl_secattr_to_sid_failure;
+
+ if (skb != NULL)
+ selinux_netlbl_cache_add(skb, &ctx_new);
+
+ ret_val = sidtab_context_to_sid(&sidtab, &ctx_new, sid);
+ } else
+ *sid = SECINITSID_UNLABELED;
+
+ ret_val = 0;
+
+netlbl_secattr_to_sid_failure:
+ if (ctx_new_destroy)
+ context_destroy(&ctx_new);
+ return ret_val;
+}
+
+/**
+ * selinux_netlbl_skbuff_getsid - Get the sid of a packet using NetLabel
+ * @skb: the packet
+ * @base_sid: the SELinux SID to use as a context for MLS only attributes
+ * @sid: the SID
+ *
+ * Description:
+ * Call the NetLabel mechanism to get the security attributes of the given
+ * packet and use those attributes to determine the correct context/SID to
+ * assign to the packet. Returns zero on success, negative values on failure.
+ *
+ */
+static int selinux_netlbl_skbuff_getsid(struct sk_buff *skb,
+ const u32 base_sid,
+ u32 *sid)
+{
+ int ret_val;
+ struct netlbl_lsm_secattr secattr;
+
+ netlbl_secattr_init(&secattr);
+ ret_val = netlbl_skbuff_getattr(skb, &secattr);
+ if (ret_val == 0)
+ ret_val = selinux_netlbl_secattr_to_sid(skb,
+ &secattr,
+ base_sid,
+ sid);
+
+ secattr.set_cache = 0;
+ netlbl_secattr_destroy(&secattr);
+
+ return ret_val;
+}
+
+/**
+ * selinux_netlbl_socket_setsid - Label a socket using the NetLabel mechanism
+ * @sock: the socket to label
+ * @sock_family: the socket family
+ * @sid: the SID to use
+ *
+ * Description:
+ * Attempt to label a socket using the NetLabel mechanism using the given
+ * SID. Returns zero values on success, negative values on failure.
+ *
+ */
+static int selinux_netlbl_socket_setsid(struct socket *sock,
+ const int sock_family,
+ u32 sid)
+{
+ int ret_val;
+ struct inode_security_struct *isec = SOCK_INODE(sock)->i_security;
+ struct netlbl_lsm_secattr secattr;
+ struct context *ctx;
+
+ ctx = sidtab_search(&sidtab, sid);
+ if (ctx != NULL) {
+ netlbl_secattr_init(&secattr);
+
+ if (selinux_netlbl_domain_export(ctx,
+ &secattr.domain,
+ NULL) == 0)
+ secattr.set_domain = 1;
+ if (mls_export_lvl(ctx, &secattr.mls_lvl, NULL) == 0)
+ secattr.set_mls_lvl = 1;
+ if (mls_export_cat(ctx,
+ &secattr.mls_cat,
+ &secattr.mls_cat_len,
+ NULL,
+ NULL) == 0)
+ secattr.set_mls_cat = 1;
+
+ ret_val = netlbl_socket_setattr(sock, &secattr);
+ if (ret_val == 0) {
+ isec->netlbl.netlbl_sid = sid;
+ isec->netlbl.labeled = 1;
+ }
+
+ netlbl_secattr_destroy(&secattr);
+ } else
+ ret_val = -ENOENT;
+
+ return ret_val;
+}
+
+/**
+ * selinux_netlbl_socket_peeksid - Get the SID of the first queued packet
+ * @sock: the socket to query
+ * @sid: the packet's SID
+ *
+ * Description:
+ * Examine the first incoming packet in the socket's queue and determine the
+ * packet's SELinux SID. Return zero on success, negative values on failure.
+ *
+ */
+static int selinux_netlbl_socket_peeksid(struct socket *sock, u32 *sid)
+{
+ int ret_val;
+ struct netlbl_lsm_secattr secattr;
+ struct inode_security_struct *isec = SOCK_INODE(sock)->i_security;
+
+ netlbl_secattr_init(&secattr);
+ ret_val = netlbl_socket_peekattr(sock, &secattr);
+ if (ret_val == 0)
+ ret_val = selinux_netlbl_secattr_to_sid(NULL,
+ &secattr,
+ isec->sid,
+ sid);
+
+ secattr.set_cache = 0;
+ netlbl_secattr_destroy(&secattr);
+
+ return ret_val;
+}
+
+/**
+ * selinux_netlbl_socket_create - Label a socket using the NetLabel mechanism
+ * @sock: the socket to label
+ * @sock_family: the socket family
+ * @sid: the SID to use
+ *
+ * Description:
+ * Attempt to label a socket using the NetLabel mechanism using the given
+ * SID. Returns zero values on success, negative values on failure.
+ *
+ */
+int selinux_netlbl_socket_create(struct socket *sock,
+ const int sock_family,
+ u32 sid)
+{
+ struct inode_security_struct *isec;
+
+ if (!ss_initialized || sock_family != PF_INET)
+ return 0;
+
+ isec = SOCK_INODE(sock)->i_security;
+ isec->netlbl.req_netlbl = 1;
+
+ /* PM - should we audit failures, or at the very least do a printk()
+ so users are not left wondering? */
+ return selinux_netlbl_socket_setsid(sock, sock_family, sid);
+}
+
+/**
+ * selinux_netlbl_socket_accept - Handle the labeling of an accept()ed socket
+ * @sock: the original socket
+ * @newsock: the new accept()ed socket
+ *
+ * Description:
+ * Attempt to label a socket using the NetLabel mechanism based on the packets
+ * in the queue and the original socket's SID.
+ *
+ */
+void selinux_netlbl_socket_accept(struct socket *sock, struct socket *newsock)
+{
+ int ret_val;
+ struct inode_security_struct *newisec;
+ struct netlbl_lsm_secattr secattr;
+
+ if (!ss_initialized ||
+ newsock->sk == NULL || newsock->sk->sk_family != PF_INET)
+ return;
+
+ newisec = SOCK_INODE(newsock)->i_security;
+ newisec->netlbl.req_netlbl = 1;
+
+ netlbl_secattr_init(&secattr);
+ ret_val = netlbl_socket_getattr(newsock, &secattr);
+ if (ret_val == 0)
+ ret_val = selinux_netlbl_secattr_to_sid(NULL,
+ &secattr,
+ newisec->sid,
+ &newisec->netlbl.peer_sid);
+ if (ret_val != 0)
+ newisec->netlbl.peer_sid = SECSID_NULL;
+ secattr.set_cache = 0;
+ netlbl_secattr_destroy(&secattr);
+}
+
+/**
+ * selinux_netlbl_file_permission - Very the file/socket is NetLabel labeled
+ * @inode: the file descriptor's inode
+ * @mask: the permission mask
+ *
+ * Description:
+ * Looks at a file's inode and if it is marked as a socket protected by
+ * NetLabel then verify that the socket has been labeled, if not try to label
+ * the socket now with the task's SID. Returns zero on success, negative
+ * values on failure.
+ *
+ */
+int selinux_netlbl_inode_permission(struct inode *inode, int mask)
+{
+ int ret_val = 0;
+ struct inode_security_struct *isec = inode->i_security;
+ struct task_security_struct *tsec;
+ struct netlbl_security_struct *netlbl_sec = &isec->netlbl;
+ struct socket *sock;
+
+ if (netlbl_sec->req_netlbl && (mask & (MAY_WRITE | MAY_APPEND))) {
+ tsec = current->security;
+ if (netlbl_sec->labeled == 0) {
+ sock = SOCKET_I(inode);
+ ret_val = selinux_netlbl_socket_setsid(sock,
+ sock->sk->sk_family,
+ tsec->sid);
+ } else if (netlbl_sec->netlbl_sid != tsec->sid &&
+ netlbl_sec->netlbl_sid != SECINITSID_KERNEL)
+ ret_val = -EACCES;
+ }
+
+ return ret_val;
+}
+
+/**
+ * selinux_netlbl_sock_rcv_skb - Do an inbound access check using NetLabel
+ * @sock_class: the socket class
+ * @sock_sid: the socket's SID
+ * @skb: the incoming packet
+ * @ad: the audit data
+ *
+ * Description:
+ * Fetch the NetLabel security attributes from @skb and perform an access check
+ * against the receiving socket. Returns zero on success, negative values on
+ * error.
+ *
+ */
+int selinux_netlbl_sock_rcv_skb(const u16 sock_class,
+ const u32 sock_sid,
+ struct sk_buff *skb,
+ struct avc_audit_data *ad)
+{
+ int ret_val;
+ u32 netlbl_sid;
+ u32 recv_perm;
+
+ ret_val = selinux_netlbl_skbuff_getsid(skb, sock_sid, &netlbl_sid);
+ if (ret_val)
+ goto netlbl_sock_rcv_skb;
+
+ /* PM - at some point we want to do an unlabeled check here too but
+ we need a way to enable/disable NetLabel checks at runtime first,
+ see RH BZ #195238 for more details */
+ if (netlbl_sid != SECINITSID_UNLABELED) {
+ switch (sock_class) {
+ case SECCLASS_UDP_SOCKET:
+ recv_perm = UDP_SOCKET__RECV_MSG;
+ break;
+ case SECCLASS_TCP_SOCKET:
+ recv_perm = TCP_SOCKET__RECV_MSG;
+ break;
+ default:
+ recv_perm = RAWIP_SOCKET__RECV_MSG;
+ break;
+ }
+
+ ret_val = avc_has_perm(sock_sid,
+ netlbl_sid,
+ sock_class,
+ recv_perm,
+ ad);
+ if (ret_val)
+ netlbl_skbuff_err(skb, ret_val);
+ }
+
+netlbl_sock_rcv_skb:
+ return ret_val;
+}
+
+/**
+ * selinux_netlbl_socket_peersid - Return the peer SID of a connected socket
+ * @sock: the socket
+ * @peer_sid: the peer SID
+ *
+ * Description:
+ * Examine @sock to find the connected peer's SID, if that is not possibile
+ * check the socket's receive queue and take the SID from the incoming
+ * packets. Returns zero on success, negative values on error.
+ *
+ */
+int selinux_netlbl_socket_getpeersec_stream(struct socket *sock, u32 *peer_sid)
+{
+ int ret_val = 0;
+ struct inode_security_struct *isec = SOCK_INODE(sock)->i_security;
+
+ if (isec->netlbl.peer_sid == 0)
+ ret_val = selinux_netlbl_socket_peeksid(sock, peer_sid);
+ else
+ *peer_sid = isec->netlbl.peer_sid;
+
+ return ret_val;
+}
+
+/**
+ * selinux_netlbl_socket_getpeersec_dgram - Return the SID of a NetLabel packet
+ * @skb: the packet
+ * @peer_sid: the packet's SID
+ *
+ * Description:
+ * Examine @skb to find the SID assigned to it by NetLabel. Returns zero on
+ * success, negative values on error.
+ *
+ */
+int selinux_netlbl_socket_getpeersec_dgram(struct sk_buff *skb, u32 *peer_sid)
+{
+ struct inode_security_struct *isec;
+
+ if (skb->sk == NULL || skb->sk->sk_socket == NULL)
+ return -ENOPROTOOPT;
+
+ isec = SOCK_INODE(skb->sk->sk_socket)->i_security;
+
+ return selinux_netlbl_skbuff_getsid(skb, isec->sid, peer_sid);
+}
Index: linux-2.6.17.i686-quilt/security/selinux/ss/services.c
===================================================================
--- linux-2.6.17.i686-quilt.orig/security/selinux/ss/services.c
+++ linux-2.6.17.i686-quilt/security/selinux/ss/services.c
@@ -13,6 +13,11 @@
*
* Added conditional policy language extensions
*
+ * Updated: Hewlett-Packard <paul.moore@hp.com>
+ *
+ * Added support for NetLabel
+ *
+ * Copyright (C) 2006 Hewlett-Packard Development Company, L.P.
* Copyright (C) 2004-2006 Trusted Computer Solutions, Inc.
* Copyright (C) 2003 - 2004 Tresys Technology, LLC
* Copyright (C) 2003 Red Hat, Inc., James Morris <jmorris@redhat.com>
@@ -29,6 +34,7 @@
#include <linux/sched.h>
#include <linux/audit.h>
#include <linux/mutex.h>
+#include <net/sock.h>
#include "flask.h"
#include "avc.h"
@@ -40,6 +46,8 @@
#include "services.h"
#include "conditional.h"
#include "mls.h"
+#include "objsec.h"
+#include "selinux_netlabel.h"
extern void selnl_notify_policyload(u32 seqno);
unsigned int policydb_loaded_version;
@@ -54,7 +62,7 @@ static DEFINE_MUTEX(load_mutex);
#define LOAD_LOCK mutex_lock(&load_mutex)
#define LOAD_UNLOCK mutex_unlock(&load_mutex)
-static struct sidtab sidtab;
+struct sidtab sidtab;
struct policydb policydb;
int ss_initialized = 0;
@@ -1241,6 +1249,7 @@ int security_load_policy(void *data, siz
selinux_complete_init();
avc_ss_reset(seqno);
selnl_notify_policyload(seqno);
+ selinux_netlbl_cache_invalidate();
return 0;
}
@@ -1295,6 +1304,7 @@ int security_load_policy(void *data, siz
avc_ss_reset(seqno);
selnl_notify_policyload(seqno);
+ selinux_netlbl_cache_invalidate();
return 0;
Index: linux-2.6.17.i686-quilt/security/selinux/ss/services.h
===================================================================
--- linux-2.6.17.i686-quilt.orig/security/selinux/ss/services.h
+++ linux-2.6.17.i686-quilt/security/selinux/ss/services.h
@@ -9,7 +9,9 @@
#include "policydb.h"
#include "sidtab.h"
+extern struct sidtab sidtab;
extern struct policydb policydb;
+extern int ss_initialized;
#endif /* _SS_SERVICES_H_ */
--
paul moore
linux security @ hp
^ permalink raw reply
* [RFC 2/8] NetLabel: core network changes
From: paul.moore @ 2006-06-22 22:49 UTC (permalink / raw)
To: netdev; +Cc: davem
In-Reply-To: <20060622224910.885573000@flek.zko.hp.com>
[-- Attachment #1: netlabel-net_core-2.6.17 --]
[-- Type: text/plain, Size: 25376 bytes --]
Changes to the core network stack to support the NetLabel subsystem. This
includes changes to support the NetLabel NETLINK communication mechanism,
changes to the IPv4 option handling to support CIPSO labels, and a new NetLabel
hook in inet_accept() to handle NetLabel attributes across an accept().
---
include/linux/ip.h | 1
include/linux/netlink.h | 1
include/net/cipso_ipv4.h | 251 ++++++++++++++++++++++++
include/net/inet_sock.h | 2
include/net/netlabel.h | 487 +++++++++++++++++++++++++++++++++++++++++++++++
net/ipv4/Makefile | 1
net/ipv4/af_inet.c | 3
net/ipv4/ah4.c | 2
net/ipv4/ip_options.c | 19 +
9 files changed, 765 insertions(+), 2 deletions(-)
Index: linux-2.6.17.i686-quilt/include/linux/ip.h
===================================================================
--- linux-2.6.17.i686-quilt.orig/include/linux/ip.h
+++ linux-2.6.17.i686-quilt/include/linux/ip.h
@@ -57,6 +57,7 @@
#define IPOPT_SEC (2 |IPOPT_CONTROL|IPOPT_COPY)
#define IPOPT_LSRR (3 |IPOPT_CONTROL|IPOPT_COPY)
#define IPOPT_TIMESTAMP (4 |IPOPT_MEASUREMENT)
+#define IPOPT_CIPSO (6 |IPOPT_CONTROL|IPOPT_COPY)
#define IPOPT_RR (7 |IPOPT_CONTROL)
#define IPOPT_SID (8 |IPOPT_CONTROL|IPOPT_COPY)
#define IPOPT_SSRR (9 |IPOPT_CONTROL|IPOPT_COPY)
Index: linux-2.6.17.i686-quilt/include/linux/netlink.h
===================================================================
--- linux-2.6.17.i686-quilt.orig/include/linux/netlink.h
+++ linux-2.6.17.i686-quilt/include/linux/netlink.h
@@ -21,6 +21,7 @@
#define NETLINK_DNRTMSG 14 /* DECnet routing messages */
#define NETLINK_KOBJECT_UEVENT 15 /* Kernel messages to userspace */
#define NETLINK_GENERIC 16
+#define NETLINK_NETLABEL 17 /* Network packet labeling */
#define MAX_LINKS 32
Index: linux-2.6.17.i686-quilt/include/net/cipso_ipv4.h
===================================================================
--- /dev/null
+++ linux-2.6.17.i686-quilt/include/net/cipso_ipv4.h
@@ -0,0 +1,251 @@
+/*
+ * CIPSO - Commercial IP Security Option
+ *
+ * This is an implementation of the CIPSO 2.2 protocol as specified in
+ * draft-ietf-cipso-ipsecurity-01.txt with additional tag types as found in
+ * FIPS-188, copies of both documents can be found in the Documentation
+ * directory. While CIPSO never became a full IETF RFC standard many vendors
+ * have chosen to adopt the protocol and over the years it has become a
+ * de-facto standard for labeled networking.
+ *
+ * Author: Paul Moore <paul.moore@hp.com>
+ *
+ */
+
+/*
+ * (c) Copyright Hewlett-Packard Development Company, L.P., 2006
+ *
+ * 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 _CIPSO_IPV4_H
+#define _CIPSO_IPV4_H
+
+#include <linux/types.h>
+#include <linux/rcupdate.h>
+#include <linux/list.h>
+#include <net/netlabel.h>
+
+/* known doi values */
+#define CIPSO_V4_DOI_UNKNOWN 0x00000000
+
+/* tag types */
+#define CIPSO_V4_TAG_INVALID 0
+#define CIPSO_V4_TAG_RBITMAP 1
+#define CIPSO_V4_TAG_ENUM 2
+#define CIPSO_V4_TAG_RANGE 5
+#define CIPSO_V4_TAG_PBITMAP 6
+#define CIPSO_V4_TAG_FREEFORM 7
+
+/* doi mapping types */
+#define CIPSO_V4_MAP_UNKNOWN 0
+#define CIPSO_V4_MAP_STD 1
+#define CIPSO_V4_MAP_PASS 2
+
+/* limits */
+#define CIPSO_V4_MAX_REM_LVLS 256
+#define CIPSO_V4_INV_LVL 0x80000000
+#define CIPSO_V4_MAX_LOC_LVLS (CIPSO_V4_INV_LVL - 1)
+#define CIPSO_V4_MAX_REM_CATS 65536
+#define CIPSO_V4_INV_CAT 0x80000000
+#define CIPSO_V4_MAX_LOC_CATS (CIPSO_V4_INV_CAT - 1)
+
+/*
+ * CIPSO DOI definitions
+ */
+
+/* DOI definition struct */
+#define CIPSO_V4_TAG_MAXCNT 5
+struct cipso_v4_doi {
+ u32 doi;
+ u32 type;
+ union {
+ struct cipso_v4_std_map_tbl *std;
+ } map;
+ u8 tags[CIPSO_V4_TAG_MAXCNT];
+
+ u32 valid;
+ struct list_head list;
+ struct rcu_head rcu;
+ struct list_head dom_list;
+};
+
+/* Standard CIPSO mapping table */
+/* NOTE: the highest order bit (i.e. 0x80000000) is an 'invalid' flag, if the
+ * bit is set then consider that value as unspecified, meaning the
+ * mapping for that particular level/category is invalid */
+struct cipso_v4_std_map_tbl {
+ struct {
+ u32 *cipso;
+ u32 *local;
+ u32 cipso_size;
+ u32 local_size;
+ } lvl;
+ struct {
+ u32 *cipso;
+ u32 *local;
+ u32 cipso_size;
+ u32 local_size;
+ } cat;
+};
+
+/*
+ * Helper Functions
+ */
+
+#define CIPSO_V4_OPTEXIST(x) (IPCB(x)->opt.cipso != 0)
+#define CIPSO_V4_OPTPTR(x) ((x)->nh.raw + IPCB(x)->opt.cipso)
+
+/*
+ * DOI List Functions
+ */
+
+#ifdef CONFIG_NETLABEL_CIPSOV4
+int cipso_v4_doi_add(struct cipso_v4_doi *doi_def);
+int cipso_v4_doi_remove(const u32 doi,
+ void (*callback) (struct rcu_head * head));
+struct cipso_v4_doi *cipso_v4_doi_getdef(const u32 doi);
+struct sk_buff *cipso_v4_doi_dump(const u32 doi, const u32 headroom);
+int cipso_v4_doi_domhsh_add(struct cipso_v4_doi *doi_def, const char *domain);
+int cipso_v4_doi_domhsh_remove(struct cipso_v4_doi *doi_def,
+ const char *domain);
+#else
+static inline int cipso_v4_doi_add(struct cipso_v4_doi *doi_def)
+{
+ return -ENOSYS;
+}
+
+static inline int cipso_v4_doi_remove(const u32 doi,
+ void (*callback) (struct rcu_head * head))
+{
+ return 0;
+}
+
+static inline struct cipso_v4_doi *cipso_v4_doi_getdef(const u32 doi)
+{
+ return NULL;
+}
+
+static inline struct sk_buff *cipso_v4_doi_dump(const u32 doi,
+ const u32 headroom)
+{
+ return NULL;
+}
+
+static inline int cipso_v4_doi_domhsh_add(struct cipso_v4_doi *doi_def,
+ const char *domain)
+{
+ return -ENOSYS;
+}
+
+static inline int cipso_v4_doi_domhsh_remove(struct cipso_v4_doi *doi_def,
+ const char *domain)
+{
+ return 0;
+}
+#endif /* CONFIG_NETLABEL_CIPSOV4 */
+
+/*
+ * Label Mapping Cache Functions
+ */
+
+#ifdef CONFIG_NETLABEL_CIPSOV4
+void cipso_v4_cache_invalidate(void);
+int cipso_v4_cache_add(const struct sk_buff *skb,
+ const struct netlbl_lsm_secattr *secattr);
+#else
+static inline void cipso_v4_cache_invalidate(void)
+{
+ return;
+}
+
+static inline int cipso_v4_cache_add(const struct sk_buff *skb,
+ const struct netlbl_lsm_secattr *secattr)
+{
+ return 0;
+}
+#endif /* CONFIG_NETLABEL_CIPSOV4 */
+
+/*
+ * Protocol Handling Functions
+ */
+
+#ifdef CONFIG_NETLABEL_CIPSOV4
+int cipso_v4_error(struct sk_buff *skb,
+ const int error,
+ const u32 gateway);
+int cipso_v4_socket_setopt(struct socket *sock,
+ unsigned char *opt,
+ u32 opt_len);
+int cipso_v4_socket_setattr(const struct socket *sock,
+ const struct cipso_v4_doi *doi_def,
+ const struct netlbl_lsm_secattr *secattr);
+int cipso_v4_socket_getopt(const struct socket *sock,
+ unsigned char **opt,
+ u32 *opt_len);
+int cipso_v4_socket_getattr(const struct socket *sock,
+ struct netlbl_lsm_secattr *secattr);
+int cipso_v4_skbuff_getattr(const struct sk_buff *skb,
+ struct netlbl_lsm_secattr *secattr);
+int cipso_v4_validate(unsigned char **option);
+#else
+static inline int cipso_v4_error(struct sk_buff *skb,
+ const int error,
+ const u32 gateway)
+{
+ return 0;
+}
+
+static inline int cipso_v4_socket_setopt(struct socket *sock,
+ unsigned char *opt,
+ u32 opt_len)
+{
+ return -ENOSYS;
+}
+
+static inline int cipso_v4_socket_setattr(const struct socket *sock,
+ const struct cipso_v4_doi *doi_def,
+ const struct netlbl_lsm_secattr *secattr)
+{
+ return -ENOSYS;
+}
+
+static inline int cipso_v4_socket_getopt(const struct socket *sock,
+ unsigned char **opt,
+ u32 *opt_len)
+{
+ return -ENOSYS;
+}
+
+static inline int cipso_v4_socket_getattr(const struct socket *sock,
+ struct netlbl_lsm_secattr *secattr)
+{
+ return -ENOSYS;
+}
+
+static inline int cipso_v4_skbuff_getattr(const struct sk_buff *skb,
+ struct netlbl_lsm_secattr *secattr)
+{
+ return -ENOSYS;
+}
+
+static inline int cipso_v4_validate(unsigned char **option)
+{
+ return -ENOSYS;
+}
+#endif /* CONFIG_NETLABEL_CIPSOV4 */
+
+#endif /* _CIPSO_IPV4_H */
Index: linux-2.6.17.i686-quilt/include/net/inet_sock.h
===================================================================
--- linux-2.6.17.i686-quilt.orig/include/net/inet_sock.h
+++ linux-2.6.17.i686-quilt/include/net/inet_sock.h
@@ -52,7 +52,7 @@ struct ip_options {
ts_needtime:1,
ts_needaddr:1;
unsigned char router_alert;
- unsigned char __pad1;
+ unsigned char cipso;
unsigned char __pad2;
unsigned char __data[0];
};
Index: linux-2.6.17.i686-quilt/include/net/netlabel.h
===================================================================
--- /dev/null
+++ linux-2.6.17.i686-quilt/include/net/netlabel.h
@@ -0,0 +1,487 @@
+/*
+ * NetLabel System
+ *
+ * The NetLabel system manages static and dynamic label mappings for network
+ * protocols such as CIPSO and RIPSO.
+ *
+ * Author: Paul Moore <paul.moore@hp.com>
+ *
+ */
+
+/*
+ * (c) Copyright Hewlett-Packard Development Company, L.P., 2006
+ *
+ * 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 _NETLABEL_H
+#define _NETLABEL_H
+
+#include <linux/types.h>
+#include <linux/skbuff.h>
+#include <net/netlink.h>
+
+/*
+ * NetLabel - A management interface for maintaining network packet label
+ * mapping tables for explicit packet labling protocols.
+ *
+ * Network protocols such as CIPSO and RIPSO require a label translation layer
+ * to convert the label on the packet into something meaningful on the host
+ * machine. In the current Linux implementation these mapping tables live
+ * inside the kernel; NetLabel provides a mechanism for user space applications
+ * to manage these mapping tables.
+ *
+ * NetLabel makes use of NETLINK as a transport layer to send messages between
+ * kernel and user space. The general format of a NetLabel message is shown
+ * below:
+ *
+ * +----------+------------------+--------- --- -- -
+ * | nlmsghdr | subsystem header | payload
+ * +----------+------------------+--------- --- -- -
+ *
+ * The 'nlmsghdr' struct should be dealt with like any other NETLINK 'nlmsghdr'
+ * struct. The subsystem header structure is dependent on the subsystem
+ * specified in the 'nlmsghdr->nlmsg_type' and should be defined below,
+ * supporting functions should be defined in the corresponding
+ * net/netlabel/netlabel_<subsys>.h|c file.
+ *
+ */
+
+/*
+ * NetLabel NETLINK protocol
+ */
+
+#define NETLBL_PROTO_VERSION 1
+
+/* NetLabel NETLINK groups */
+#define NETLBL_NLGRP_NONE 0x00000000
+#define NETLBL_NLGRP_MAX 0x00000000
+
+/* NetLabel NETLINK types */
+#define NETLBL_NLTYPE_NONE 0
+#define NETLBL_NLTYPE_MGMT 1
+#define NETLBL_NLTYPE_RIPSO 2
+#define NETLBL_NLTYPE_CIPSOV4 3
+#define NETLBL_NLTYPE_CIPSOV6 4
+#define NETLBL_NLTYPE_UNLABELED 5
+
+/* NetLabel return codes */
+#define NETLBL_E_OK 0
+
+/*
+ * Helper functions
+ */
+
+/**
+ * netlbl_put_u8 - Write a u8 value into a buffer
+ * @buffer: the buffer
+ * @val: the value
+ *
+ * Description:
+ * Write the value specified in @val into the buffer specified by @buffer.
+ *
+ */
+static inline void netlbl_put_u8(unsigned char *buffer, const u8 val)
+{
+ *(u8 *)buffer = val;
+}
+
+/**
+ * netlbl_put_u32 - Write a u32 value into a buffer
+ * @buffer: the buffer
+ * @val: the value
+ *
+ * Description:
+ * Write the value specified in @val into the buffer specified by @buffer.
+ *
+ */
+static inline void netlbl_put_u32(unsigned char *buffer, const u32 val)
+{
+ *(u32 *)buffer = val;
+}
+
+/**
+ * netlbl_put_str - Write a string into a buffer
+ * @buffer: the buffer
+ * @val: the string
+ *
+ * Description:
+ * Write the string specified in @val into the buffer specified by @buffer.
+ *
+ */
+static inline void netlbl_put_str(unsigned char *buffer, const char *val)
+{
+ strcpy((char *)buffer, val);
+}
+
+/**
+ * netlbl_put_hdr - Write a NETLINK header into a buffer
+ * @buffer: the buffer
+ * @msg_type: the NETLINK message type
+ * @msg_len: the NETLINK message length
+ * @msg_flags: the NETLINK message flags
+ * @msg_pid: the NETLINK message PID
+ * @msg_seq: the NETLINK message sequence number
+ *
+ * Description:
+ * Use the given values to write a NETLINK header into the given buffer.
+ *
+ */
+static inline void netlbl_put_hdr(unsigned char *buffer,
+ const u32 msg_type,
+ const u16 msg_len,
+ const u16 msg_flags,
+ const u32 msg_pid,
+ const u32 msg_seq)
+{
+ struct nlmsghdr *hdr = (struct nlmsghdr *)buffer;
+ hdr->nlmsg_len = msg_len;
+ hdr->nlmsg_type = msg_type;
+ hdr->nlmsg_flags = msg_flags;
+ hdr->nlmsg_seq = msg_seq;
+ hdr->nlmsg_pid = msg_pid;
+}
+
+/**
+ * netlbl_put_u8 - Write a u8 value into a buffer and increment the buffer
+ * @buffer: the buffer
+ * @val: the value
+ *
+ * Description:
+ * Write the value specified in @val into the buffer specified by @buffer
+ * and advance the buffer pointer past the newly written value.
+ *
+ */
+static inline void netlbl_putinc_u8(unsigned char **buffer, const u8 val)
+{
+ netlbl_put_u8(*buffer, val);
+ *buffer += sizeof(u8);
+}
+
+/**
+ * netlbl_put_u32 - Write a u32 value into a buffer and increment the buffer
+ * @buffer: the buffer
+ * @val: the value
+ *
+ * Description:
+ * Write the value specified in @val into the buffer specified by @buffer
+ * and advance the buffer pointer past the newly written value.
+ *
+ */
+static inline void netlbl_putinc_u32(unsigned char **buffer, const u32 val)
+{
+ netlbl_put_u32(*buffer, val);
+ *buffer += sizeof(u32);
+}
+
+/**
+ * netlbl_put_u8 - Write a string into a buffer and increment the buffer
+ * @buffer: the buffer
+ * @val: the value
+ *
+ * Description:
+ * Write the string specified in @val into the buffer specified by @buffer
+ * and advance the buffer pointer past the newly written value.
+ *
+ */
+static inline void netlbl_putinc_str(unsigned char **buffer, const char *val)
+{
+ netlbl_put_str(*buffer, val);
+ *buffer += strlen(val) + 1;
+}
+
+/**
+ * netlbl_put_hdr - Write a NETLINK header into a buffer and increment the ptr
+ * @buffer: the buffer
+ * @msg_type: the NETLINK message type
+ * @msg_len: the NETLINK message length
+ * @msg_flags: the NETLINK message flags
+ * @msg_pid: the NETLINK message PID
+ * @msg_seq: the NETLINK message sequence number
+ *
+ * Description:
+ * Use the given values to write a NETLINK header into the given buffer and
+ * then increment the buffer pointer past the header.
+ *
+ */
+static inline void netlbl_putinc_hdr(unsigned char **buffer,
+ const u32 msg_type,
+ const u16 msg_len,
+ const u16 msg_flags,
+ const u32 msg_pid,
+ const u32 msg_seq)
+{
+ netlbl_put_hdr(*buffer,
+ msg_type,
+ msg_len,
+ msg_flags,
+ msg_pid,
+ msg_seq);
+ *buffer += NLMSG_HDRLEN;
+}
+
+/**
+ * netlbl_get_u8 - Read a u8 value from a buffer
+ * @buffer: the buffer
+ *
+ * Description:
+ * Return a u8 value pointed to by @buffer.
+ *
+ */
+static inline u8 netlbl_get_u8(const unsigned char *buffer)
+{
+ return *(u8 *)buffer;
+}
+
+/**
+ * netlbl_get_u32 - Read a u32 value from a buffer
+ * @buffer: the buffer
+ *
+ * Description:
+ * Return a u8 value pointed to by @buffer.
+ *
+ */
+static inline u32 netlbl_get_u32(const unsigned char *buffer)
+{
+ return *(u32 *)buffer;
+}
+
+/**
+ * netlbl_getinc_u8 - Read a u8 value from a buffer and increment the buffer
+ * @buffer: the buffer
+ *
+ * Description:
+ * Return a u8 value pointed to by @buffer and increment the buffer pointer
+ * past the value.
+ *
+ */
+static inline u8 netlbl_getinc_u8(unsigned char **buffer)
+{
+ u8 val = netlbl_get_u8(*buffer);
+ *buffer += sizeof(u8);
+ return val;
+}
+
+/**
+ * netlbl_getinc_u32 - Read a u32 value from a buffer and increment the buffer
+ * @buffer: the buffer
+ *
+ * Description:
+ * Return a u32 value pointed to by @buffer and increment the buffer pointer
+ * past the value.
+ *
+ */
+static inline u32 netlbl_getinc_u32(unsigned char **buffer)
+{
+ u32 val = netlbl_get_u32(*buffer);
+ *buffer += sizeof(u32);
+ return val;
+}
+
+/*
+ * NetLabel - Kernel API for accessing the network packet label mappings.
+ *
+ * The following functions are provided for use by other kernel modules,
+ * specifically kernel LSM modules, to provide a consistent, transparent API
+ * for dealing with explicit packet labeling protocols such as CIPSO and
+ * RIPSO. The functions defined here are implemented in the
+ * net/netlabel/netlabel_kapi.c file.
+ *
+ */
+
+/* Domain mapping definition struct */
+struct netlbl_dom_map;
+
+/* Domain mapping operations */
+int netlbl_domhsh_remove(const char *domain);
+
+/* LSM security attributes */
+struct netlbl_lsm_cache {
+ void (*free) (const void *data);
+ void *data;
+};
+struct netlbl_lsm_secattr {
+ char *domain;
+
+ u32 mls_lvl;
+ unsigned char *mls_cat;
+ u32 mls_cat_len;
+
+ struct netlbl_lsm_cache cache;
+
+ u32 set_domain:1,
+ set_mls_lvl:1,
+ set_mls_cat:1,
+ set_cache:1,
+ __unused:28;
+};
+
+/*
+ * LSM security attribute operations
+ */
+
+/**
+ * netlbl_secattr_alloc - Allocate and initialize a netlbl_lsm_secattr struct
+ * @flags: the memory allocation flags
+ *
+ * Description:
+ * Allocate and initialize a netlbl_lsm_secattr struct. Returns a valid
+ * pointer on success, or NULL on failure.
+ *
+ */
+static inline struct netlbl_lsm_secattr *netlbl_secattr_alloc(const int flags)
+{
+ return kzalloc(sizeof(struct netlbl_lsm_secattr), flags);
+}
+
+/**
+ * netlbl_secattr_free - Frees a netlbl_lsm_secattr struct
+ * @secattr: the struct to free
+ *
+ * Description:
+ * Frees @secattr including all of the internal buffers.
+ *
+ */
+static inline void netlbl_secattr_free(struct netlbl_lsm_secattr *secattr)
+{
+ BUG_ON(secattr == NULL);
+ if (secattr->set_domain)
+ kfree(secattr->domain);
+ if (secattr->set_mls_cat)
+ kfree(secattr->mls_cat);
+ if (secattr->set_cache && secattr->cache.free)
+ secattr->cache.free(secattr->cache.data);
+ kfree(secattr);
+}
+
+/**
+ * netlbl_secattr_init - Initialize a netlbl_lsm_secattr struct
+ * @secattr: the struct to initialize
+ *
+ * Description:
+ * Initialize an already allocated netlbl_lsm_secattr struct. Returns zero on
+ * success, negative values on error.
+ *
+ */
+static inline int netlbl_secattr_init(struct netlbl_lsm_secattr *secattr)
+{
+ BUG_ON(secattr == NULL);
+ memset(secattr, 0, sizeof(*secattr));
+ return 0;
+}
+
+/**
+ * netlbl_secattr_destroy - Clears a netlbl_lsm_secattr struct
+ * @secattr: the struct to clear
+ *
+ * Description:
+ * Destroys the @secattr struct, including freeing all of the internal buffers.
+ * On return the struct is suitable for reuse.
+ *
+ */
+static inline void netlbl_secattr_destroy(struct netlbl_lsm_secattr *secattr)
+{
+ BUG_ON(secattr == NULL);
+ if (secattr->set_domain)
+ kfree(secattr->domain);
+ if (secattr->set_mls_cat)
+ kfree(secattr->mls_cat);
+ if (secattr->set_cache && secattr->cache.free)
+ secattr->cache.free(secattr->cache.data);
+ memset(secattr, 0, sizeof(*secattr));
+}
+
+/*
+ * LSM protocol operations
+ */
+
+#ifdef CONFIG_NETLABEL
+int netlbl_socket_setattr(const struct socket *sock,
+ const struct netlbl_lsm_secattr *secattr);
+int netlbl_socket_peekattr(const struct socket *sock,
+ struct netlbl_lsm_secattr *secattr);
+int netlbl_socket_getattr(const struct socket *sock,
+ struct netlbl_lsm_secattr *secattr);
+int netlbl_skbuff_getattr(const struct sk_buff *skb,
+ struct netlbl_lsm_secattr *secattr);
+int netlbl_skbuff_err(struct sk_buff *skb, int error);
+#else
+static inline int netlbl_socket_setattr(const struct socket *sock,
+ const struct netlbl_lsm_secattr *secattr)
+{
+ return -ENOSYS;
+}
+
+static inline int netlbl_socket_peekattr(const struct socket *sock,
+ struct netlbl_lsm_secattr *secattr)
+{
+ return -ENOSYS;
+}
+
+static inline int netlbl_socket_getattr(const struct socket *sock,
+ struct netlbl_lsm_secattr *secattr)
+{
+ return -ENOSYS;
+}
+
+static inline int netlbl_skbuff_getattr(const struct sk_buff *skb,
+ struct netlbl_lsm_secattr *secattr)
+{
+ return -ENOSYS;
+}
+
+static inline int netlbl_skbuff_err(struct sk_buff *skb, int error)
+{
+ return 0;
+}
+#endif /* CONFIG_NETLABEL */
+
+/*
+ * LSM label mapping cache operations
+ */
+
+#ifdef CONFIG_NETLABEL
+void netlbl_cache_invalidate(void);
+int netlbl_cache_add(const struct sk_buff *skb,
+ const struct netlbl_lsm_secattr *secattr);
+#else
+static inline void netlbl_cache_invalidate(void)
+{
+ return;
+}
+
+static inline int netlbl_cache_add(const struct sk_buff *skb,
+ const struct netlbl_lsm_secattr *secattr)
+{
+ return 0;
+}
+#endif /* CONFIG_NETLABEL */
+
+/*
+ * Network stack operations
+ */
+
+#ifdef CONFIG_NETLABEL
+void netlbl_socket_inet_accept(struct socket *sock, struct socket *newsock);
+#else
+static inline void netlbl_socket_inet_accept(struct socket *sock,
+ struct socket *newsock)
+{
+ return;
+}
+#endif /* CONFIG_NETLABEL */
+
+#endif /* _NETLABEL_H */
Index: linux-2.6.17.i686-quilt/net/ipv4/Makefile
===================================================================
--- linux-2.6.17.i686-quilt.orig/net/ipv4/Makefile
+++ linux-2.6.17.i686-quilt/net/ipv4/Makefile
@@ -42,6 +42,7 @@ obj-$(CONFIG_TCP_CONG_HYBLA) += tcp_hybl
obj-$(CONFIG_TCP_CONG_HTCP) += tcp_htcp.o
obj-$(CONFIG_TCP_CONG_VEGAS) += tcp_vegas.o
obj-$(CONFIG_TCP_CONG_SCALABLE) += tcp_scalable.o
+obj-$(CONFIG_NETLABEL_CIPSOV4) += cipso_ipv4.o
obj-$(CONFIG_XFRM) += xfrm4_policy.o xfrm4_state.o xfrm4_input.o \
xfrm4_output.o
Index: linux-2.6.17.i686-quilt/net/ipv4/af_inet.c
===================================================================
--- linux-2.6.17.i686-quilt.orig/net/ipv4/af_inet.c
+++ linux-2.6.17.i686-quilt/net/ipv4/af_inet.c
@@ -114,6 +114,7 @@
#ifdef CONFIG_IP_MROUTE
#include <linux/mroute.h>
#endif
+#include <net/netlabel.h>
DEFINE_SNMP_STAT(struct linux_mib, net_statistics) __read_mostly;
@@ -616,6 +617,8 @@ int inet_accept(struct socket *sock, str
sock_graft(sk2, newsock);
+ netlbl_socket_inet_accept(sock, newsock);
+
newsock->state = SS_CONNECTED;
err = 0;
release_sock(sk2);
Index: linux-2.6.17.i686-quilt/net/ipv4/ah4.c
===================================================================
--- linux-2.6.17.i686-quilt.orig/net/ipv4/ah4.c
+++ linux-2.6.17.i686-quilt/net/ipv4/ah4.c
@@ -35,7 +35,7 @@ static int ip_clear_mutable_options(stru
switch (*optptr) {
case IPOPT_SEC:
case 0x85: /* Some "Extended Security" crap. */
- case 0x86: /* Another "Commercial Security" crap. */
+ case IPOPT_CIPSO:
case IPOPT_RA:
case 0x80|21: /* RFC1770 */
break;
Index: linux-2.6.17.i686-quilt/net/ipv4/ip_options.c
===================================================================
--- linux-2.6.17.i686-quilt.orig/net/ipv4/ip_options.c
+++ linux-2.6.17.i686-quilt/net/ipv4/ip_options.c
@@ -24,6 +24,7 @@
#include <net/ip.h>
#include <net/icmp.h>
#include <net/route.h>
+#include <net/cipso_ipv4.h>
/*
* Write options to IP header, record destination address to
@@ -194,6 +195,13 @@ int ip_options_echo(struct ip_options *
dopt->is_strictroute = sopt->is_strictroute;
}
}
+ if (sopt->cipso) {
+ optlen = sptr[sopt->cipso+1];
+ dopt->cipso = dopt->optlen+sizeof(struct iphdr);
+ memcpy(dptr, sptr+sopt->cipso, optlen);
+ dptr += optlen;
+ dopt->optlen += optlen;
+ }
while (dopt->optlen & 3) {
*dptr++ = IPOPT_END;
dopt->optlen++;
@@ -435,6 +443,17 @@ int ip_options_compile(struct ip_options
if (optptr[2] == 0 && optptr[3] == 0)
opt->router_alert = optptr - iph;
break;
+ case IPOPT_CIPSO:
+ if (opt->cipso) {
+ pp_ptr = optptr;
+ goto error;
+ }
+ opt->cipso = optptr - iph;
+ if (cipso_v4_validate(&optptr)) {
+ pp_ptr = optptr;
+ goto error;
+ }
+ break;
case IPOPT_SEC:
case IPOPT_SID:
default:
--
paul moore
linux security @ hp
^ permalink raw reply
* [RFC 4/8] NetLabel: core NetLabel subsystem
From: paul.moore @ 2006-06-22 22:49 UTC (permalink / raw)
To: netdev; +Cc: davem
In-Reply-To: <20060622224910.885573000@flek.zko.hp.com>
[-- Attachment #1: netlabel-core --]
[-- Type: text/plain, Size: 78122 bytes --]
Add a new kernel subsystem, NetLabel, to provide explicit packet labeling
services (CIPSO, RIPSO, etc.) to LSM developers. NetLabel is designed to work
in conjunction with a LSM to intercept and decode security labels on incoming
network packets as well as ensure that outgoing network packets are labeled
according to the security mechanism employed by the LSM. The NetLabel
subsystem is configured through a NETLINK interface described in the header
files included in this patch.
---
net/Makefile | 1
net/netlabel/Kconfig | 47 ++
net/netlabel/Makefile | 16
net/netlabel/netlabel_cipso_v4.h | 209 +++++++++++
net/netlabel/netlabel_domainhash.c | 601 ++++++++++++++++++++++++++++++++
net/netlabel/netlabel_domainhash.h | 64 +++
net/netlabel/netlabel_kapi.c | 373 ++++++++++++++++++++
net/netlabel/netlabel_mgmt.c | 688 +++++++++++++++++++++++++++++++++++++
net/netlabel/netlabel_mgmt.h | 248 +++++++++++++
net/netlabel/netlabel_unlabeled.h | 105 +++++
net/netlabel/netlabel_user.c | 162 ++++++++
net/netlabel/netlabel_user.h | 67 +++
12 files changed, 2581 insertions(+)
Index: linux-2.6.17.i686-quilt/net/Makefile
===================================================================
--- linux-2.6.17.i686-quilt.orig/net/Makefile
+++ linux-2.6.17.i686-quilt/net/Makefile
@@ -47,6 +47,7 @@ obj-$(CONFIG_IP_DCCP) += dccp/
obj-$(CONFIG_IP_SCTP) += sctp/
obj-$(CONFIG_IEEE80211) += ieee80211/
obj-$(CONFIG_TIPC) += tipc/
+obj-$(CONFIG_NETLABEL) += netlabel/
ifeq ($(CONFIG_NET),y)
obj-$(CONFIG_SYSCTL) += sysctl_net.o
Index: linux-2.6.17.i686-quilt/net/netlabel/Kconfig
===================================================================
--- /dev/null
+++ linux-2.6.17.i686-quilt/net/netlabel/Kconfig
@@ -0,0 +1,47 @@
+#
+# NetLabel configuration
+#
+
+config NETLABEL
+ bool "NetLabel subsystem support"
+ depends on NET && SECURITY
+ default n
+ ---help---
+ NetLabel provides support for network packet labeling protocols. For
+ more information see Documentation/netlabel.
+
+ If you are unsure, say N.
+
+config NETLABEL_CIPSOV4
+ bool "CIPSO for IPv4"
+ depends on NETLABEL
+ default n
+ ---help---
+ The Commercial IP Security Option (CIPSO) is commonly used in
+ trusted multi-level security networks. This implementation is based
+ on the IETF draft draft-ietf-cipso-ipsecurity-01.txt and FIPS-188.
+
+ If you are unsure say N.
+
+config NETLABEL_UNLABELED
+ bool "Unlabeled packets"
+ depends on NETLABEL
+ default y
+ ---help---
+ This adds support for sending and receiving unlabeled packets through
+ the NetLabel system.
+
+ If you are unsure say Y.
+
+config NETLABEL_UNLABELED_DEFAULT
+ bool "Enable unlabeled traffic by default"
+ depends on NETLABEL_UNLABELED
+ default y
+ ---help---
+ Enabling this option allows unlabeled (i.e. normal) network traffic
+ on the system without any user configuration at bootup. Disabling
+ this option means the user must configure the NetLabel subsystem
+ before any type of network traffic can be sent or received on the
+ system.
+
+ If you are unsure say Y.
Index: linux-2.6.17.i686-quilt/net/netlabel/Makefile
===================================================================
--- /dev/null
+++ linux-2.6.17.i686-quilt/net/netlabel/Makefile
@@ -0,0 +1,16 @@
+#
+# Makefile for the NetLabel subsystem.
+#
+# Feb 9, 2006, Paul Moore <paul.moore@hp.com>
+#
+
+# base objects
+obj-y := netlabel_user.o netlabel_kapi.o netlabel_domainhash.o
+
+# management objects
+obj-y += netlabel_mgmt.o
+
+# protocol modules
+obj-$(CONFIG_NETLABEL_UNLABELED) += netlabel_unlabeled.o
+obj-$(CONFIG_NETLABEL_CIPSOV4) += netlabel_cipso_v4.o
+
Index: linux-2.6.17.i686-quilt/net/netlabel/netlabel_cipso_v4.h
===================================================================
--- /dev/null
+++ linux-2.6.17.i686-quilt/net/netlabel/netlabel_cipso_v4.h
@@ -0,0 +1,209 @@
+/*
+ * NetLabel CIPSO/IPv4 Support
+ *
+ * This file defines the CIPSO/IPv4 functions for the NetLabel system. The
+ * NetLabel system manages static and dynamic label mappings for network
+ * protocols such as CIPSO and RIPSO.
+ *
+ * Author: Paul Moore <paul.moore@hp.com>
+ *
+ */
+
+/*
+ * (c) Copyright Hewlett-Packard Development Company, L.P., 2006
+ *
+ * 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 _NETLABEL_CIPSO_V4
+#define _NETLABEL_CIPSO_V4
+
+#include <net/netlabel.h>
+
+/*
+ * The following NetLabel payloads are supported by the CIPSO subsystem, all
+ * of which are preceeded by the nlmsghdr struct.
+ *
+ * o ACK:
+ * Sent by the kernel in response to an applications message, applications
+ * should never send this message.
+ *
+ * +----------------------+-----------------------+
+ * | seq number (32 bits) | return code (32 bits) |
+ * +----------------------+-----------------------+
+ *
+ * seq number: the sequence number of the original message, taken from the
+ * nlmsghdr structure
+ * return code: return value, based on errno values
+ *
+ * o ADD:
+ * Sent by an application to add a new DOI mapping table, after completion
+ * of the task the kernel should ACK this message.
+ *
+ * +--------------------+---------------------+
+ * | map type (32 bits) | tag count (32 bits) | ...
+ * +--------------------+---------------------+
+ *
+ * +-----------------+
+ * | tag #X (8 bits) | ... repeated
+ * +-----------------+
+ *
+ * +-------------- ---- --- -- -
+ * | mapping data
+ * +-------------- ---- --- -- -
+ *
+ * map type: the mapping table type (defined in the cipso_ipv4.h header
+ * as CIPSO_V4_MAP_*)
+ * tag count: the number of tags, must be greater than zero
+ * tag: the CIPSO tag for the DOI, tags listed first are given
+ * higher priorirty when sending packets
+ * mapping data: specific to the map type (see below)
+ *
+ * CIPSO_V4_MAP_STD
+ *
+ * +------------------+-----------------------+-----------------------+
+ * | levels (32 bits) | max l level (32 bits) | max r level (32 bits) | ...
+ * +------------------+-----------------------+-----------------------+
+ *
+ * +----------------------+---------------------+---------------------+
+ * | categories (32 bits) | max l cat (32 bits) | max r cat (32 bits) | ...
+ * +----------------------+---------------------+---------------------+
+ *
+ * +--------------------------+--------------------------+
+ * | local level #X (32 bits) | CIPSO level #X (32 bits) | ... repeated
+ * +--------------------------+--------------------------+
+ *
+ * +-----------------------------+-----------------------------+
+ * | local category #X (32 bits) | CIPSO category #X (32 bits) | ... repeated
+ * +-----------------------------+-----------------------------+
+ *
+ * levels: the number of level mappings
+ * max l level: the highest local level
+ * max r level: the highest remote/CIPSO level
+ * categories: the number of category mappings
+ * max l cat: the highest local category
+ * max r cat: the highest remote/CIPSO category
+ * local level: the local part of a level mapping
+ * CIPSO level: the remote/CIPSO part of a level mapping
+ * local category: the local part of a category mapping
+ * CIPSO category: the remote/CIPSO part of a category mapping
+ *
+ * CIPSO_V4_MAP_PASS
+ *
+ * No mapping data is needed for this map type.
+ *
+ * o REMOVE:
+ * Sent by an application to remove a specific DOI mapping table from the
+ * CIPSO V4 system. This message does not contain a payload. The kernel
+ * should ACK this message.
+ *
+ * o LIST:
+ * This message can be sent either from an application or by the kernel in
+ * response to an application generated LIST message. When sent by an
+ * application there is no payload. If the application sets the DOI field
+ * to zero in the CIPSO V4 message header then the kernel should respond
+ * with a list of valid DOIs. If the application sets the DOI field equal to
+ * a non-zero value then the kernel should respond with the matching mapping
+ * table. In the case of an error the kernel should respond with an ACK
+ * message.
+ *
+ * DOI Listing (DOI == 0)
+ *
+ * +---------------------+------------------+-----------------------+
+ * | DOI count (32 bits) | DOI #X (32 bits) | map type #X (32 bits) | ...
+ * +---------------------+------------------+-----------------------+
+ *
+ * DOI count: the number of DOIs
+ * DOI: the DOI value
+ * map type: the DOI mapping table type (defined in the cipso_ipv4.h
+ * header as CIPSO_V4_MAP_*)
+ *
+ * DOI Mapping Table (DOI != 0)
+ *
+ * +--------------------+
+ * | map type (32 bits) | ...
+ * +--------------------+
+ *
+ * map type: the DOI mapping table type (defined in the cipso_ipv4.h
+ * header as CIPSO_V4_MAP_*)
+ *
+ * (map type == CIPSO_V4_MAP_STD)
+ *
+ * +----------------+------------------+----------------------+
+ * | tags (32 bits) | levels (32 bits) | categories (32 bits) | ...
+ * +----------------+------------------+----------------------+
+ *
+ * +-----------------+
+ * | tag #X (8 bits) | ... repeated
+ * +-----------------+
+ *
+ * +--------------------------+--------------------------+
+ * | local level #X (32 bits) | CIPSO level #X (32 bits) | ... repeated
+ * +--------------------------+--------------------------+
+ *
+ * +-----------------------------+-----------------------------+
+ * | local category #X (32 bits) | CIPSO category #X (32 bits) | ... repeated
+ * +-----------------------------+-----------------------------+
+ *
+ * tags: the number of CIPSO tag types
+ * levels: the number of level mappings
+ * categories: the number of category mappings
+ * tag: the tag number, tags listed first are given higher
+ * priority when sending packets
+ * local level: the local part of a level mapping
+ * CIPSO level: the remote/CIPSO part of a level mapping
+ * local category: the local part of a category mapping
+ * CIPSO category: the remote/CIPSO part of a category mapping
+ *
+ * (map type == CIPSO_V4_MAP_PASS)
+ *
+ * +----------------+
+ * | tags (32 bits) | ...
+ * +----------------+
+ *
+ * +-----------------+
+ * | tag #X (8 bits) | ... repeated
+ * +-----------------+
+ *
+ * tags: the number of CIPSO tag types
+ * tag: the tag number, tags listed first are given higher
+ * priority when sending packets
+ *
+ */
+
+/* CIPSO V4 message header */
+struct netlbl_cipsov4_msghdr {
+ enum { NL_CV4_NOOP,
+ NL_CV4_ACK,
+ NL_CV4_ADD,
+ NL_CV4_REMOVE,
+ NL_CV4_LIST
+ } opcode;
+ u32 doi;
+};
+
+/* Process CIPSO V4 NetLabel messages */
+#ifdef CONFIG_NETLABEL_CIPSOV4
+void netlbl_cipsov4_rcv(const struct sk_buff *skb, const unsigned char *msg);
+#else
+static inline void netlbl_cipsov4_rcv(const struct sk_buff *skb,
+ const unsigned char *msg)
+{
+ return;
+}
+#endif
+
+#endif
Index: linux-2.6.17.i686-quilt/net/netlabel/netlabel_domainhash.c
===================================================================
--- /dev/null
+++ linux-2.6.17.i686-quilt/net/netlabel/netlabel_domainhash.c
@@ -0,0 +1,601 @@
+/*
+ * NetLabel Domain Hash Table
+ *
+ * This file manages the domain hash table that NetLabel uses to determine
+ * which network labeling protocol to use for a given domain. The NetLabel
+ * system manages static and dynamic label mappings for network protocols such
+ * as CIPSO and RIPSO.
+ *
+ * Author: Paul Moore <paul.moore@hp.com>
+ *
+ */
+
+/*
+ * (c) Copyright Hewlett-Packard Development Company, L.P., 2006
+ *
+ * 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
+ *
+ */
+
+#include <linux/types.h>
+#include <linux/rcupdate.h>
+#include <linux/list.h>
+#include <linux/skbuff.h>
+#include <linux/spinlock.h>
+#include <linux/string.h>
+#include <net/netlabel.h>
+#include <net/cipso_ipv4.h>
+#include <asm/bug.h>
+
+#include "netlabel_mgmt.h"
+#include "netlabel_domainhash.h"
+
+struct netlbl_domhsh_tbl {
+ struct list_head *tbl;
+ u32 size;
+};
+
+/* Domain hash table */
+/* XXX - updates should be so rare that having one spinlock for the entire
+ hash table should be okay */
+DEFINE_SPINLOCK(netlbl_domhsh_lock);
+static struct netlbl_domhsh_tbl *netlbl_domhsh = NULL;
+
+/* Default domain mapping */
+DEFINE_SPINLOCK(netlbl_domhsh_def_lock);
+static struct netlbl_dom_map *netlbl_domhsh_def = NULL;
+
+/*
+ * Domain Hash Table Helper Functions
+ */
+
+/**
+ * netlbl_domhsh_free_entry - Frees a domain hash table entry
+ * @entry: the entry's RCU field
+ *
+ * Description:
+ * This function is designed to be used as a callback to the call_rcu()
+ * function so that the memory allocated to a hash table entry can be released
+ * safely.
+ *
+ */
+static void netlbl_domhsh_free_entry(struct rcu_head *entry)
+{
+ struct netlbl_dom_map *ptr;
+
+ ptr = container_of(entry, struct netlbl_dom_map, rcu);
+ if (ptr->domain)
+ kfree(ptr->domain);
+ kfree(ptr);
+}
+
+/**
+ * netlbl_domhsh_hash - Hashing function for the domain hash table
+ * @domain: the domain name to hash
+ *
+ * Description:
+ * This is the hashing function for the domain hash table, it returns the
+ * correct bucket number for the domain. The caller is responsibile for
+ * calling the rcu_read_[un]lock() functions.
+ *
+ */
+static u32 netlbl_domhsh_hash(const char *domain)
+{
+ char *p;
+ char *keyp = (char *)domain;
+ u32 size;
+ u32 val = 0;
+
+ /* This is taken (with slight modification) from
+ security/selinux/ss/symtab.c:symhash() */
+
+ size = strlen(keyp);
+ for (p = keyp; (p - keyp) < size; p++)
+ val = (val << 4 | (val >> (8 * sizeof(u32) - 4))) ^ (*p);
+ return val & (rcu_dereference(netlbl_domhsh)->size - 1);
+}
+
+/**
+ * netlbl_domhsh_search - Search for a domain entry
+ * @domain: the domain
+ * @def: return default if no match is found
+ *
+ * Description:
+ * Searches the domain hash table and returns a pointer to the hash table
+ * entry if found, otherwise NULL is returned. If @def is non-zero and a
+ * match is not found in the domain hash table the default mapping is returned
+ * if it exists. The caller is responsibile for the rcu hash table locks
+ * (i.e. the caller much call rcu_read_[un]lock()).
+ *
+ */
+static struct netlbl_dom_map *netlbl_domhsh_search(const char *domain,
+ const u32 def)
+{
+ u32 bkt;
+ struct netlbl_dom_map *iter;
+
+ if (domain != NULL) {
+ bkt = netlbl_domhsh_hash(domain);
+ list_for_each_entry_rcu(iter, &netlbl_domhsh->tbl[bkt], list)
+ if (iter->valid && strcmp(iter->domain, domain) == 0)
+ return iter;
+ }
+
+ if (def != 0) {
+ iter = rcu_dereference(netlbl_domhsh_def);
+ if (iter != NULL && iter->valid)
+ return iter;
+ }
+
+ return NULL;
+}
+
+/*
+ * Domain Hash Table Functions
+ */
+
+/**
+ * netlbl_domhsh_init - Init for the domain hash
+ * @size: the number of bits to use for the hash buckets
+ *
+ * Description:
+ * Initializes the domain hash table, should be called only by
+ * netlbl_user_init() during module load. Returns zero on success, non-zero
+ * values on error.
+ *
+ */
+int netlbl_domhsh_init(const u32 size)
+{
+ u32 iter;
+ struct netlbl_domhsh_tbl *hsh_tbl;
+
+ if (size <= 0)
+ return -EINVAL;
+
+ hsh_tbl = kmalloc(sizeof(*hsh_tbl), GFP_KERNEL);
+ if (hsh_tbl == NULL)
+ return -ENOMEM;
+ hsh_tbl->size = 1 << size;
+ hsh_tbl->tbl = kcalloc(hsh_tbl->size,
+ sizeof(struct list_head),
+ GFP_KERNEL);
+ if (hsh_tbl->tbl == NULL) {
+ kfree(hsh_tbl);
+ return -ENOMEM;
+ }
+ for (iter = 0; iter < hsh_tbl->size; iter++)
+ INIT_LIST_HEAD(&hsh_tbl->tbl[iter]);
+
+ rcu_read_lock();
+ spin_lock(&netlbl_domhsh_lock);
+ rcu_assign_pointer(netlbl_domhsh, hsh_tbl);
+ spin_unlock(&netlbl_domhsh_lock);
+ rcu_read_unlock();
+
+ return 0;
+}
+
+/**
+ * netlbl_domhsh_exit - Cleanup for the domain hash
+ *
+ * Description:
+ * Clears and frees the domain hash table. No real serious effort is made
+ * towards locking but that shouldn't be an issue since we should only be
+ * called when the module is being unloaded. Returns zero on success, non-zero
+ * values on error.
+ *
+ */
+void netlbl_domhsh_exit(void)
+{
+ u32 iter_bkt;
+ struct netlbl_domhsh_tbl *hsh_tbl;
+ struct netlbl_dom_map *iter_list;
+
+ rcu_read_lock();
+ if (rcu_dereference(netlbl_domhsh) == NULL) {
+ rcu_read_unlock();
+ return;
+ }
+ iter_list = rcu_dereference(netlbl_domhsh_def);
+ spin_lock(&netlbl_domhsh_def_lock);
+ rcu_assign_pointer(netlbl_domhsh_def, NULL);
+ spin_unlock(&netlbl_domhsh_def_lock);
+ rcu_read_unlock();
+
+ if (iter_list) {
+ synchronize_rcu();
+ netlbl_domhsh_free_entry(&iter_list->rcu);
+ }
+
+ rcu_read_lock();
+ hsh_tbl = rcu_dereference(netlbl_domhsh);
+ spin_lock(&netlbl_domhsh_lock);
+ rcu_assign_pointer(netlbl_domhsh, NULL);
+ spin_unlock(&netlbl_domhsh_lock);
+ rcu_read_unlock();
+
+ synchronize_rcu();
+
+ /* PM - it probably doesn't matter, but do we need to use the rcu list
+ functions here since we are the only one who has access to the
+ list now? */
+ for (iter_bkt = 0; iter_bkt < hsh_tbl->size; iter_bkt++)
+ list_for_each_entry_rcu(iter_list,
+ &hsh_tbl->tbl[iter_bkt], list) {
+ iter_list->valid = 0;
+ list_del_rcu(&iter_list->list);
+ netlbl_domhsh_free_entry(&iter_list->rcu);
+ }
+ kfree(hsh_tbl->tbl);
+ kfree(hsh_tbl);
+
+ return;
+}
+
+/**
+ * netlbl_domhsh_add - Adds a entry to the domain hash table
+ * @entry: the entry to add
+ *
+ * Description:
+ * Adds a new entry to the domain hash table and handles any updates to the
+ * lower level protocol handler (i.e. CIPSO). Returns zero on success,
+ * negative on failure.
+ *
+ */
+int netlbl_domhsh_add(struct netlbl_dom_map *entry)
+{
+ int ret_val = -EPERM;
+ u32 bkt;
+
+ BUG_ON(entry == NULL);
+
+ switch (entry->type) {
+ case NETLBL_NLTYPE_UNLABELED:
+ ret_val = 0;
+ break;
+ case NETLBL_NLTYPE_CIPSOV4:
+ ret_val = cipso_v4_doi_domhsh_add(entry->type_def.cipsov4,
+ entry->domain);
+ break;
+ default:
+ BUG();
+ }
+ if (ret_val != 0)
+ return ret_val;
+
+ entry->valid = 1;
+ INIT_RCU_HEAD(&entry->rcu);
+
+ rcu_read_lock();
+ if (entry->domain != NULL &&
+ netlbl_domhsh_search(entry->domain, 0) == NULL) {
+ bkt = netlbl_domhsh_hash(entry->domain);
+ spin_lock(&netlbl_domhsh_lock);
+ list_add_tail_rcu(&entry->list, &netlbl_domhsh->tbl[bkt]);
+ spin_unlock(&netlbl_domhsh_lock);
+ } else if (entry->domain == NULL &&
+ rcu_dereference(netlbl_domhsh_def) == NULL) {
+ INIT_LIST_HEAD(&entry->list);
+ spin_lock(&netlbl_domhsh_def_lock);
+ rcu_assign_pointer(netlbl_domhsh_def, entry);
+ spin_unlock(&netlbl_domhsh_def_lock);
+ } else {
+ rcu_read_unlock();
+ switch (entry->type) {
+ case NETLBL_NLTYPE_UNLABELED:
+ break;
+ case NETLBL_NLTYPE_CIPSOV4:
+ if (cipso_v4_doi_domhsh_remove(entry->type_def.cipsov4,
+ entry->domain) != 0)
+ BUG();
+ break;
+ default:
+ BUG();
+ }
+ return -EPERM;
+ }
+ rcu_read_unlock();
+
+ return 0;
+}
+
+/**
+ * netlbl_domhsh_add_default - Adds the default entry to the domain hash table
+ * @entry: the entry to add
+ *
+ * Description:
+ * Adds a new default entry to the domain hash table and handles any updates
+ * to the lower level protocol handler (i.e. CIPSO). Returns zero on success,
+ * negative on failure.
+ *
+ */
+int netlbl_domhsh_add_default(struct netlbl_dom_map *entry)
+{
+ return netlbl_domhsh_add(entry);
+}
+
+/**
+ * netlbl_domhsh_remove - Removes an entry from the domain hash table
+ * @domain: the domain to remove
+ *
+ * Description:
+ * Removes an entry from the domain hash table and handles any updates to the
+ * lower level protocol handler (i.e. CIPSO). Returns zero on success,
+ * negative on failure.
+ *
+ */
+int netlbl_domhsh_remove(const char *domain)
+{
+ int ret_val = -EPERM;
+ struct netlbl_dom_map *entry;
+
+ rcu_read_lock();
+ if (domain != NULL)
+ entry = netlbl_domhsh_search(domain, 0);
+ else
+ entry = netlbl_domhsh_search(domain, 1);
+ if (entry == NULL)
+ goto remove_return;
+ switch (entry->type) {
+ case NETLBL_NLTYPE_UNLABELED:
+ break;
+ case NETLBL_NLTYPE_CIPSOV4:
+ if (cipso_v4_doi_domhsh_remove(entry->type_def.cipsov4,
+ entry->domain) != 0)
+ goto remove_return;
+ break;
+ default:
+ BUG();
+ }
+ if (entry != rcu_dereference(netlbl_domhsh_def)) {
+ spin_lock(&netlbl_domhsh_lock);
+ entry->valid = 0;
+ list_del_rcu(&entry->list);
+ spin_unlock(&netlbl_domhsh_lock);
+ } else {
+ spin_lock(&netlbl_domhsh_def_lock);
+ entry->valid = 0;
+ rcu_assign_pointer(netlbl_domhsh_def, NULL);
+ spin_unlock(&netlbl_domhsh_def_lock);
+ }
+ call_rcu(&entry->rcu, netlbl_domhsh_free_entry);
+
+ ret_val = 0;
+
+remove_return:
+ rcu_read_unlock();
+ return ret_val;
+}
+
+/**
+ * netlbl_domhsh_remove_default - Removes the default entry from the table
+ *
+ * Description:
+ * Removes/resets the default entry for the domain hash table and handles any
+ * updates to the lower level protocol handler (i.e. CIPSO). Returns zero on
+ * success, non-zero on failure.
+ *
+ */
+int netlbl_domhsh_remove_default(void)
+{
+ return netlbl_domhsh_remove(NULL);
+}
+
+/**
+ * netlbl_domhsh_getentry - Get an entry from the domain hash table
+ * @domain: the domain name to search for
+ *
+ * Description:
+ * Look through the domain hash table searching for an entry to match @domain,
+ * return a pointer to a copy of the entry or NULL. The caller is responsibile
+ * for ensuring that rcu_read_[un]lock() is called.
+ *
+ */
+struct netlbl_dom_map *netlbl_domhsh_getentry(const char *domain)
+{
+ return netlbl_domhsh_search(domain, 1);
+}
+
+/**
+ * netlbl_domhsh_dump - Dump the domain hash table into a sk_buff
+ *
+ * Description:
+ * Dump the domain hash table into a buffer suitable for returning to an
+ * application in response to a NetLabel management DOMAIN message. This
+ * function may fail if another process is growing the hash table at the same
+ * time. The returned sk_buff has room at the front of the sk_buff for
+ * a nlmsghdr struct and a netlbl_mgmt_msghdr struct. See netlabel.h for the
+ * DOMAIN message format. Returns a pointer to a sk_buff on success, NULL on
+ * error.
+ *
+ */
+struct sk_buff *netlbl_domhsh_dump(void)
+{
+ struct sk_buff *skb;
+ unsigned char *buf;
+ u32 buf_len;
+ u32 bkt_iter;
+ u32 dom_cnt = 0;
+ struct netlbl_domhsh_tbl *hsh_tbl;
+ struct netlbl_dom_map *list_iter;
+ u32 tmp_len;
+
+ /* XXX - This is kinda ugly as we have to go through the table once to
+ determine how large of a buffer we need, drop the locks, allocate
+ the buffer, grab the locks, and finally fill the buffer. The
+ problem is that there is that open window where the table could
+ grow and we will end up short on space. */
+
+ buf_len = 4;
+ rcu_read_lock();
+ hsh_tbl = rcu_dereference(netlbl_domhsh);
+ for (bkt_iter = 0; bkt_iter < hsh_tbl->size; bkt_iter++)
+ list_for_each_entry_rcu(list_iter,
+ &hsh_tbl->tbl[bkt_iter], list) {
+ buf_len += 9 + strlen(list_iter->domain);
+ switch (list_iter->type) {
+ case NETLBL_NLTYPE_UNLABELED:
+ break;
+ case NETLBL_NLTYPE_CIPSOV4:
+ buf_len += 8;
+ break;
+ default:
+ rcu_read_unlock();
+ BUG();
+ }
+ dom_cnt++;
+ }
+ rcu_read_unlock();
+
+ skb = alloc_skb(NLMSG_SPACE(sizeof(struct netlbl_mgmt_msghdr) +
+ buf_len), GFP_KERNEL);
+ if (skb == NULL)
+ return NULL;
+ skb_reserve(skb, NLMSG_SPACE(sizeof(struct netlbl_mgmt_msghdr)));
+ tmp_len = skb_tailroom(skb);
+ if (tmp_len < buf_len)
+ goto dump_failure;
+ buf = skb_put(skb, buf_len);
+ buf_len -= 4;
+ netlbl_putinc_u32(&buf, dom_cnt);
+
+ rcu_read_lock();
+ hsh_tbl = rcu_dereference(netlbl_domhsh);
+ for (bkt_iter = 0; bkt_iter < hsh_tbl->size; bkt_iter++)
+ list_for_each_entry_rcu(list_iter,
+ &hsh_tbl->tbl[bkt_iter], list) {
+ tmp_len = strlen(list_iter->domain) + 1;
+ if (buf_len < tmp_len + 8) {
+ rcu_read_unlock();
+ goto dump_failure;
+ }
+ buf_len -= tmp_len + 8;
+ netlbl_putinc_u32(&buf, tmp_len);
+ netlbl_putinc_str(&buf, list_iter->domain);
+ netlbl_putinc_u32(&buf, list_iter->type);
+ switch (list_iter->type) {
+ case NETLBL_NLTYPE_UNLABELED:
+ break;
+ case NETLBL_NLTYPE_CIPSOV4:
+ if (buf_len < 8) {
+ rcu_read_unlock();
+ goto dump_failure;
+ }
+ buf_len -= 8;
+ netlbl_putinc_u32(&buf,
+ list_iter->type_def.cipsov4->type);
+ netlbl_putinc_u32(&buf,
+ list_iter->type_def.cipsov4->doi);
+ break;
+ default:
+ rcu_read_unlock();
+ BUG();
+ }
+ }
+ rcu_read_unlock();
+
+ return skb;
+
+dump_failure:
+ kfree_skb(skb);
+ return NULL;
+}
+
+/**
+ * netlbl_domhsh_dump_default - Dump the default domain mapping into a sk_buff
+ *
+ * Description:
+ * Dump the default domain mapping into a buffer suitable for returning to an
+ * application in response to a NetLabel management DEFDOMAIN message. This
+ * function may fail if another process is changing the default domain mapping
+ * at the same time. The returned sk_buff has room at the front of the
+ * skb_buff for a nlmsghdr struct and a netlbl_mgmt_msghdr struct. See
+ * netlabel.h for the DEFDOMAIN message format. Returns a pointer to a
+ * sk_buff on success, NULL on error.
+ *
+ */
+struct sk_buff *netlbl_domhsh_dump_default(void)
+{
+ struct sk_buff *skb;
+ unsigned char *buf;
+ u32 buf_len;
+ u32 tmp_len;
+ struct netlbl_dom_map *entry;
+
+ /* XXX - This is kinda ugly as we have to go look at the default
+ mapping once to determine how large of a buffer we need, drop the
+ locks, allocate the buffer, grab the locks, and finally fill the
+ buffer. The problem is that there is that open window where the
+ default mapping could change on us, if that is the case we fail. */
+
+ buf_len = 4;
+ rcu_read_lock();
+ entry = rcu_dereference(netlbl_domhsh_def);
+ if (entry != NULL)
+ switch (entry->type) {
+ case NETLBL_NLTYPE_UNLABELED:
+ break;
+ case NETLBL_NLTYPE_CIPSOV4:
+ buf_len += 8;
+ break;
+ default:
+ BUG();
+ }
+ rcu_read_unlock();
+
+ skb = alloc_skb(NLMSG_SPACE(sizeof(struct netlbl_mgmt_msghdr) +
+ buf_len), GFP_KERNEL);
+ if (skb == NULL)
+ return NULL;
+ skb_reserve(skb, NLMSG_SPACE(sizeof(struct netlbl_mgmt_msghdr)));
+ tmp_len = skb_tailroom(skb);
+ if (tmp_len < buf_len)
+ goto dump_default_failure;
+ buf = skb_put(skb, buf_len);
+
+ rcu_read_lock();
+ if (entry != rcu_dereference(netlbl_domhsh_def)) {
+ rcu_read_unlock();
+ goto dump_default_failure;
+ }
+ buf_len -= 4;
+ if (entry != NULL) {
+ netlbl_putinc_u32(&buf, entry->type);
+ switch (entry->type) {
+ case NETLBL_NLTYPE_UNLABELED:
+ break;
+ case NETLBL_NLTYPE_CIPSOV4:
+ if (buf_len < 8) {
+ rcu_read_unlock();
+ goto dump_default_failure;
+ }
+ buf_len -= 8;
+ netlbl_putinc_u32(&buf, entry->type_def.cipsov4->type);
+ netlbl_putinc_u32(&buf, entry->type_def.cipsov4->doi);
+ break;
+ default:
+ rcu_read_unlock();
+ BUG();
+ }
+ } else
+ netlbl_putinc_u32(&buf, NETLBL_NLTYPE_NONE);
+ rcu_read_unlock();
+
+ return skb;
+
+dump_default_failure:
+ kfree_skb(skb);
+ return NULL;
+}
Index: linux-2.6.17.i686-quilt/net/netlabel/netlabel_domainhash.h
===================================================================
--- /dev/null
+++ linux-2.6.17.i686-quilt/net/netlabel/netlabel_domainhash.h
@@ -0,0 +1,64 @@
+/*
+ * NetLabel Domain Hash Table
+ *
+ * This file manages the domain hash table that NetLabel uses to determine
+ * which network labeling protocol to use for a given domain. The NetLabel
+ * system manages static and dynamic label mappings for network protocols such
+ * as CIPSO and RIPSO.
+ *
+ * Author: Paul Moore <paul.moore@hp.com>
+ *
+ */
+
+/*
+ * (c) Copyright Hewlett-Packard Development Company, L.P., 2006
+ *
+ * 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 _NETLABEL_DOMAINHASH_H
+#define _NETLABEL_DOMAINHASH_H
+
+/* Domain hash table size */
+/* XXX - currently this number is an uneducated guess */
+#define NETLBL_DOMHSH_BITSIZE 7
+
+/* Domain mapping definition struct */
+struct netlbl_dom_map {
+ char *domain;
+ u32 type;
+ union {
+ struct cipso_v4_doi *cipsov4;
+ } type_def;
+
+ u32 valid;
+ struct list_head list;
+ struct rcu_head rcu;
+};
+
+/* init/exit functions */
+int netlbl_domhsh_init(const u32 size);
+void netlbl_domhsh_exit(void);
+
+/* Manipulate the domain hash table */
+int netlbl_domhsh_add(struct netlbl_dom_map *entry);
+int netlbl_domhsh_add_default(struct netlbl_dom_map *entry);
+int netlbl_domhsh_remove_default(void);
+struct netlbl_dom_map *netlbl_domhsh_getentry(const char *domain);
+struct sk_buff *netlbl_domhsh_dump(void);
+struct sk_buff *netlbl_domhsh_dump_default(void);
+
+#endif
Index: linux-2.6.17.i686-quilt/net/netlabel/netlabel_kapi.c
===================================================================
--- /dev/null
+++ linux-2.6.17.i686-quilt/net/netlabel/netlabel_kapi.c
@@ -0,0 +1,373 @@
+/*
+ * NetLabel Kernel API
+ *
+ * This file defines the kernel API for the NetLabel system. The NetLabel
+ * system manages static and dynamic label mappings for network protocols such
+ * as CIPSO and RIPSO.
+ *
+ * Author: Paul Moore <paul.moore@hp.com>
+ *
+ */
+
+/*
+ * (c) Copyright Hewlett-Packard Development Company, L.P., 2006
+ *
+ * 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
+ *
+ */
+
+#include <linux/init.h>
+#include <linux/types.h>
+#include <net/ip.h>
+#include <net/netlabel.h>
+#include <net/cipso_ipv4.h>
+#include <asm/bug.h>
+
+#include "netlabel_domainhash.h"
+#include "netlabel_user.h"
+#include "netlabel_unlabeled.h"
+
+/* PM - remove a lot of the input verification BUG()s once this gets farther
+ along */
+
+/*
+ * LSM Functions
+ */
+
+/**
+ * netlbl_socket_setattr - Label a socket using the correct protocol
+ * @sock: the socket to label
+ * @secattr: the security attributes
+ *
+ * Description:
+ * Attach the correct label to the given socket using the security attributes
+ * specified in @secattr. Returns zero on success, negative values on failure.
+ *
+ */
+int netlbl_socket_setattr(const struct socket *sock,
+ const struct netlbl_lsm_secattr *secattr)
+{
+ int ret_val = -ENOENT;
+ struct netlbl_dom_map *dom_entry;
+
+ /* PM - split these into individual BUGs so we can pinpoint problems
+ easier */
+ BUG_ON(sock == NULL);
+ BUG_ON(secattr == NULL);
+ BUG_ON(secattr->set_domain == 0);
+
+ rcu_read_lock();
+ dom_entry = netlbl_domhsh_getentry(secattr->domain);
+ if (dom_entry == NULL)
+ goto socket_setattr_return;
+ switch (dom_entry->type) {
+ case NETLBL_NLTYPE_CIPSOV4:
+ ret_val = cipso_v4_socket_setattr(sock,
+ dom_entry->type_def.cipsov4,
+ secattr);
+ break;
+#ifdef CONFIG_NETLABEL_UNLABELED
+ case NETLBL_NLTYPE_UNLABELED:
+ ret_val = 0;
+ break;
+#endif
+ default:
+ ret_val = -ENOENT;
+ }
+
+socket_setattr_return:
+ rcu_read_unlock();
+ return ret_val;
+}
+
+/**
+ * netlbl_socket_peekattr - Get the security attributes on a queued packet
+ * @sock: the socket
+ * @secattr: the security attributes
+ *
+ * Description:
+ * Peek at the incoming packet queue of @sock and return the security
+ * attributes of the first packet in the queue. If there are no packets in
+ * the queue or no packets with security attributes return -ENOMSG. Otherwise
+ * return zero on success, negative values on failure.
+ *
+ */
+int netlbl_socket_peekattr(const struct socket *sock,
+ struct netlbl_lsm_secattr *secattr)
+{
+ int ret_val = -ENOMSG;
+ struct sk_buff_head *sock_queue;
+ struct sk_buff *skb;
+ struct sk_buff *skb_tmp = NULL;
+
+ /* PM - split these into individual BUGs so we can pinpoint problems
+ easier */
+ BUG_ON(sock == NULL);
+ BUG_ON(sock->sk == NULL);
+ BUG_ON(sock->sk->sk_family != PF_INET);
+ BUG_ON(secattr == NULL);
+
+ sock_queue = &sock->sk->sk_receive_queue;
+ /* XXX - We could just take the lock here and do a skb_peek() but we
+ might end up holding the lock for a long time so grab the
+ lock, do a peek, and clone the result */
+ spin_lock(&sock_queue->lock);
+ skb = skb_peek(sock_queue);
+ if (skb != NULL)
+ skb_tmp = skb_clone(skb, GFP_ATOMIC);
+ spin_unlock(&sock_queue->lock);
+ if (skb_tmp != NULL) {
+ ret_val = netlbl_skbuff_getattr(skb_tmp, secattr);
+ kfree_skb(skb_tmp);
+ } else
+ ret_val = -ENOMEM;
+
+ return ret_val;
+}
+
+/**
+ * netlbl_socket_getattr - Determine the security attributes of a socket
+ * @sock: the socket
+ * @secattr: the security attributes
+ *
+ * Description:
+ * Examines the given socket to see any NetLabel style labeling has been
+ * applied to the socket, if so it parses the socket label and returns the
+ * security attributes in @secattr. Returns zero on success, negative values
+ * on failure.
+ *
+ */
+int netlbl_socket_getattr(const struct socket *sock,
+ struct netlbl_lsm_secattr *secattr)
+{
+ int ret_val = -ENOMSG;
+
+ /* PM - split these into individual BUGs so we can pinpoint problems
+ easier */
+ BUG_ON(sock == NULL);
+ BUG_ON(sock->sk == NULL);
+ BUG_ON(sock->sk->sk_family != PF_INET);
+ BUG_ON(secattr == NULL);
+
+ ret_val = cipso_v4_socket_getattr(sock, secattr);
+ if (ret_val == 0)
+ return 0;
+
+ ret_val = netlbl_unlabel_getattr(secattr);
+
+ return ret_val;
+}
+
+/**
+ * netlbl_skbuff_getattr - Determine the security attributes of a packet
+ * @skb: the packet
+ * @secattr: the security attributes
+ *
+ * Description:
+ * Examines the given packet to see if a recognized form of packet labeling
+ * is present, if so it parses the packet label and returns the security
+ * attributes in @secattr. Returns zero on success, negative values on
+ * failure.
+ *
+ */
+int netlbl_skbuff_getattr(const struct sk_buff *skb,
+ struct netlbl_lsm_secattr *secattr)
+{
+ int ret_val = -ENOMSG;
+
+ /* PM - split these into individual BUGs so we can pinpoint problems
+ easier */
+ BUG_ON(skb == NULL);
+ BUG_ON(secattr == NULL);
+
+ ret_val = cipso_v4_skbuff_getattr(skb, secattr);
+ if (ret_val == 0)
+ return 0;
+
+ ret_val = netlbl_unlabel_getattr(secattr);
+
+ return ret_val;
+}
+
+/**
+ * netlbl_skbuff_err - Handle a LSM error on a sk_buff
+ * @skb: the packet
+ * @error: the error code
+ *
+ * Description:
+ * Deal with a LSM problem when handling the packet in @skb, typically this is
+ * a permission denied problem (-EACCES). The correct action is determined
+ * according to the packet's labeling protocol. Returns zero on success,
+ * negative values on failure.
+ *
+ */
+int netlbl_skbuff_err(struct sk_buff *skb, int error)
+{
+ BUG_ON(skb == NULL);
+
+ if (CIPSO_V4_OPTEXIST(skb))
+ return cipso_v4_error(skb, error, 0);
+
+ return -ENOMSG;
+}
+
+/**
+ * netlbl_cache_invalidate - Invalidate all of the NetLabel protocol caches
+ *
+ * Description:
+ * For all of the NetLabel protocols that support some form of label mapping
+ * cache, invalidate the cache. Returns zero on success, negative values on
+ * error.
+ *
+ */
+void netlbl_cache_invalidate(void)
+{
+ cipso_v4_cache_invalidate();
+}
+
+/**
+ * netlbl_cache_add - Add an entry to a NetLabel protocol cache
+ * @skb: the packet
+ * @secattr: the packet's security attributes
+ *
+ * Description:
+ * Add the LSM security attributes for the given packet to the underlying
+ * NetLabel protocol's label mapping cache. Returns zero on success, negative
+ * values on error.
+ *
+ */
+int netlbl_cache_add(const struct sk_buff *skb,
+ const struct netlbl_lsm_secattr *secattr)
+{
+ int ret_val = -ENOMSG;
+
+ /* PM - split these into individual BUGs so we can pinpoint problems
+ easier */
+ BUG_ON(skb == NULL);
+ BUG_ON(secattr == NULL);
+
+ if (secattr->set_cache != 1)
+ return -ENOMSG;
+
+ if (CIPSO_V4_OPTEXIST(skb))
+ ret_val = cipso_v4_cache_add(skb, secattr);
+
+ return ret_val;
+}
+
+/*
+ * Network Stack Functions
+ */
+
+/**
+ * netlbl_socket_inet_accept - Preserve the NetLabel across an accept()
+ * @sock: the old socket
+ * @newsock: the new socket
+ *
+ * Description:
+ * For kernel threads only, clear the NetLabel from @newsock and copy the
+ * NetLabel from @sock. The caller is responsibile for locking and releasing
+ * the sockets.
+ *
+ */
+void netlbl_socket_inet_accept(struct socket *sock, struct socket *newsock)
+{
+ unsigned char opt_null = IPOPT_NOOP, *opt = NULL;
+ u32 opt_len = 0;
+
+ /* We are only interested in preserving the NetLabel on kernel
+ threads across an accept, the LSM should handle user level
+ accept calls. */
+ if (current->mm)
+ return;
+
+ /* PM - we can't return a error here (no way to un-accept a socket)
+ and i'm pretty sure i would be lynched if i put a BUG_ON() here
+ but we should probably do a printk() at the very least ... */
+
+ if (cipso_v4_socket_getopt(sock, &opt, &opt_len) == 0)
+ cipso_v4_socket_setopt(newsock, opt, opt_len);
+ else
+ cipso_v4_socket_setopt(newsock, &opt_null, 1);
+
+ return;
+}
+
+/*
+ * Setup Functions
+ */
+
+/**
+ * netlbl_init - Init for the NetLabel module
+ *
+ * Description:
+ * Perform the initialization of the NetLabel module before first use.
+ *
+ */
+static int __init netlbl_init(void)
+{
+ int ret_val = 0;
+
+ printk(KERN_INFO "NetLabel: Initializing\n");
+ printk(KERN_INFO "NetLabel: domain hash size = %u\n",
+ (1 << NETLBL_DOMHSH_BITSIZE));
+ printk(KERN_INFO "NetLabel: protocols ="
+#ifdef CONFIG_NETLABEL_UNLABELED
+ " UNLABELED"
+#endif
+#ifdef CONFIG_NETLABEL_CIPSOV4
+ " CIPSOv4"
+#endif
+ "\n");
+
+ ret_val = netlbl_domhsh_init(NETLBL_DOMHSH_BITSIZE);
+ if (ret_val != 0)
+ goto init_return;
+
+ ret_val = netlbl_netlink_init();
+ if (ret_val != 0)
+ goto init_return;
+
+#ifdef CONFIG_NETLABEL_UNLABELED_DEFAULT
+ ret_val = netlbl_unlabel_defconf();
+ if (ret_val != 0)
+ goto init_return;
+ printk(KERN_INFO "NetLabel: unlabeled traffic allowed by default\n");
+#endif
+
+init_return:
+ if (ret_val != 0) {
+ netlbl_netlink_exit();
+ netlbl_domhsh_exit();
+ }
+ return ret_val;
+}
+
+/* PM - not sure there is any point to this now, delete this func? */
+/**
+ * netlbl_exit - Cleanup for the NetLabel module
+ *
+ * Description:
+ * Perform any cleanup required before removing the NetLabel module.
+ *
+ */
+static void __exit netlbl_exit(void)
+{
+ netlbl_netlink_exit();
+ netlbl_domhsh_exit();
+}
+
+subsys_initcall(netlbl_init);
+
Index: linux-2.6.17.i686-quilt/net/netlabel/netlabel_mgmt.c
===================================================================
--- /dev/null
+++ linux-2.6.17.i686-quilt/net/netlabel/netlabel_mgmt.c
@@ -0,0 +1,688 @@
+/*
+ * NetLabel Management Support
+ *
+ * This file defines the management functions for the NetLabel system. The
+ * NetLabel system manages static and dynamic label mappings for network
+ * protocols such as CIPSO and RIPSO.
+ *
+ * Author: Paul Moore <paul.moore@hp.com>
+ *
+ */
+
+/*
+ * (c) Copyright Hewlett-Packard Development Company, L.P., 2006
+ *
+ * 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
+ *
+ */
+
+#include <linux/types.h>
+#include <linux/socket.h>
+#include <linux/string.h>
+#include <linux/skbuff.h>
+#include <net/sock.h>
+#include <net/netlink.h>
+#include <net/netlabel.h>
+#include <net/cipso_ipv4.h>
+
+#include "netlabel_domainhash.h"
+#include "netlabel_user.h"
+#include "netlabel_mgmt.h"
+
+/*
+ * Local Prototypes
+ */
+
+static void netlbl_mgmt_send_ack(const struct sk_buff *req_skb,
+ const u32 ret_code);
+
+/*
+ * Helper Functions
+ */
+
+/**
+ * netlbl_mgmt_put_hdr - Write a MGMT NetLabel header into a buffer
+ * @buffer: the buffer
+ * @opcode: the NetLabel Management opcode
+ *
+ * Description:
+ * Use the given values to write a NetLabel Management header into the given
+ * buffer.
+ *
+ */
+static void netlbl_mgmt_put_hdr(unsigned char *buffer, const u32 opcode)
+{
+ ((struct netlbl_mgmt_msghdr *)buffer)->opcode = opcode;
+}
+
+/**
+ * netlbl_mgmt_putinc_hdr - Write a MGMT NetLabel header into a buffer
+ * @buffer: the buffer
+ * @opcode: the NetLabel Management opcode
+ *
+ * Description:
+ * Use the given values to write a NetLabel Management header into the given
+ * buffer and increment the buffer pointer past the header.
+ *
+ */
+static void netlbl_mgmt_putinc_hdr(unsigned char **buffer, const u32 opcode)
+{
+ netlbl_mgmt_put_hdr(*buffer, opcode);
+ *buffer += sizeof(struct netlbl_mgmt_msghdr);
+}
+
+/**
+ * netlbl_mgmt_payload_len - Return the length of the payload
+ * @skb: the NETLINK buffer
+ *
+ * Description:
+ * This function returns the length of the NetLabel management payload.
+ *
+ */
+static u32 netlbl_mgmt_payload_len(const struct sk_buff *skb)
+{
+ const struct nlmsghdr *nl_hdr = (struct nlmsghdr *)skb->data;
+
+ if (nlmsg_len(nl_hdr) <= sizeof(struct netlbl_mgmt_msghdr))
+ return 0;
+ return nlmsg_len(nl_hdr) - sizeof(struct netlbl_mgmt_msghdr);
+}
+
+/**
+ * netlbl_mgmt_payload_data - Returns a pointer to the start of the data
+ * @skb: the NETLINK buffer
+ *
+ * Description:
+ * This function returns a pointer to the start of the NetLabel management
+ * payload.
+ *
+ */
+static unsigned char *netlbl_mgmt_payload_data(const struct sk_buff *skb)
+{
+ return nlmsg_data((struct nlmsghdr *)skb->data) +
+ sizeof(struct netlbl_mgmt_msghdr);
+}
+
+/*
+ * Label Mapping Functions
+ */
+
+/**
+ * netlbl_mgmt_add - Handle an ADD message
+ * @req_skb: the NETLINK buffer
+ * @msg: the NetLabel management message
+ *
+ * Description:
+ * Process a user generated ADD message and add the domains from the message
+ * to the hash table. See netlabel.h for a description of the message format.
+ *
+ */
+static void netlbl_mgmt_add(const struct sk_buff *req_skb,
+ const unsigned char *msg)
+{
+ int ret_val = -EINVAL;
+ unsigned char *msg_ptr = (unsigned char *)msg;
+ u32 msg_len = netlbl_mgmt_payload_len(req_skb);
+ u32 count;
+ struct netlbl_dom_map *entry = NULL;
+ u32 iter;
+ u32 tmp_val;
+ u32 rcu_locked = 0;
+
+ if (msg_len < 4)
+ goto add_failure;
+ count = netlbl_getinc_u32(&msg_ptr);
+ msg_len -= 4;
+
+ for (iter = 0; iter < count && msg_len > 0; iter++, entry = NULL) {
+ if (msg_len < 4)
+ goto add_failure;
+ tmp_val = netlbl_getinc_u32(&msg_ptr);
+ msg_len -= 4;
+ if (tmp_val == 0)
+ goto add_failure;
+ entry = kzalloc(sizeof(*entry), GFP_KERNEL);
+ if (entry == NULL) {
+ ret_val = -ENOMEM;
+ goto add_failure;
+ }
+
+ if (msg_len < tmp_val)
+ goto add_failure;
+ entry->domain = kmalloc(tmp_val, GFP_KERNEL);
+ if (entry->domain == NULL) {
+ ret_val = -ENOMEM;
+ goto add_failure;
+ }
+ strncpy(entry->domain, msg_ptr, tmp_val);
+ entry->domain[tmp_val - 1] = '\0';
+ msg_ptr += tmp_val;
+ msg_len -= tmp_val;
+
+ if (msg_len < 4)
+ goto add_failure;
+ tmp_val = netlbl_getinc_u32(&msg_ptr);
+ msg_len -= 4;
+ entry->type = tmp_val;
+ switch (tmp_val) {
+#ifdef CONFIG_NETLABEL_UNLABELED
+ case NETLBL_NLTYPE_UNLABELED:
+ break;
+#endif
+#ifdef CONFIG_NETLABEL_CIPSOV4
+ case NETLBL_NLTYPE_CIPSOV4:
+ if (msg_len < 4)
+ goto add_failure;
+ tmp_val = netlbl_getinc_u32(&msg_ptr);
+ msg_len -= 4;
+ /* XXX - we should be holding a rcu_read_lock() here
+ while we hold the result but since the entry
+ will always be deleted when the CIPSO DOI
+ is deleted we aren't going to keep the lock */
+ rcu_read_lock();
+ entry->type_def.cipsov4 = cipso_v4_doi_getdef(tmp_val);
+ if (entry->type_def.cipsov4 == NULL) {
+ rcu_read_unlock();
+ goto add_failure;
+ }
+ rcu_locked = 1;
+ break;
+#endif
+ default:
+ goto add_failure;
+ }
+
+ ret_val = netlbl_domhsh_add(entry);
+ if (rcu_locked) {
+ rcu_read_unlock();
+ rcu_locked = 0;
+ }
+ if (ret_val != 0)
+ goto add_failure;
+ }
+
+ netlbl_mgmt_send_ack(req_skb, NETLBL_E_OK);
+ return;
+
+add_failure:
+ if (entry) {
+ if (entry->domain)
+ kfree(entry->domain);
+ kfree(entry);
+ }
+ netlbl_mgmt_send_ack(req_skb, -ret_val);
+}
+
+/**
+ * netlbl_mgmt_remove - Handle a REMOVE message
+ * @req_skb: the NETLINK buffer
+ * @msg: the NetLabel management message
+ *
+ * Description:
+ * Process a user generated REMOVE message and remove the specified domain
+ * mappings.
+ *
+ */
+static void netlbl_mgmt_remove(const struct sk_buff *req_skb,
+ const unsigned char *msg)
+{
+ int ret_val = -EINVAL;
+ unsigned char *msg_ptr = (unsigned char *)msg;
+ u32 msg_len = netlbl_mgmt_payload_len(req_skb);
+ u32 count;
+ u32 iter;
+ u32 tmp_val;
+
+ if (msg_len < 4)
+ goto remove_return;
+ count = netlbl_getinc_u32(&msg_ptr);
+ msg_len -= 4;
+
+ for (iter = 0; iter < count && msg_len > 0; iter++) {
+ if (msg_len < 4)
+ goto remove_return;
+ tmp_val = netlbl_getinc_u32(&msg_ptr);
+ msg_len -= 4;
+ if (tmp_val == 0 ||
+ msg_len < tmp_val || msg_ptr[tmp_val - 1] != '\0')
+ goto remove_return;
+ ret_val = netlbl_domhsh_remove(msg_ptr);
+ msg_ptr += tmp_val;
+ msg_len -= tmp_val;
+ if (ret_val != 0)
+ goto remove_return;
+ }
+
+ ret_val = 0;
+
+remove_return:
+ netlbl_mgmt_send_ack(req_skb, -ret_val);
+}
+
+/**
+ * netlbl_mgmt_list - Handle a LIST message
+ * @req_skb: the NETLINK buffer
+ * @msg: the NetLabel management message
+ *
+ * Description:
+ * Process a user generated LIST message and dumps the domain hash table in a
+ * form suitable for use in a kernel generated LIST message.
+ *
+ */
+static void netlbl_mgmt_list(const struct sk_buff *req_skb,
+ const unsigned char *msg)
+{
+ int ret_val = -ENOMEM;
+ struct sk_buff *skb;
+ unsigned char *buf_ptr;
+
+ skb = netlbl_domhsh_dump();
+ if (skb == NULL)
+ goto list_return;
+ buf_ptr = skb_push(skb,
+ NLMSG_SPACE(sizeof(struct netlbl_mgmt_msghdr)));
+ if (buf_ptr == NULL) {
+ kfree_skb(skb);
+ ret_val = -EAGAIN;
+ goto list_return;
+ }
+ netlbl_putinc_hdr(&buf_ptr,
+ NETLBL_NLTYPE_MGMT,
+ skb->len,
+ 0,
+ NETLINK_CB(req_skb).pid,
+ 0);
+ netlbl_mgmt_putinc_hdr(&buf_ptr, NL_MGMT_LIST);
+
+ ret_val = netlbl_netlink_snd(skb, NETLINK_CB(req_skb).pid);
+
+list_return:
+ if (ret_val != 0)
+ netlbl_mgmt_send_ack(req_skb, -ret_val);
+}
+
+/**
+ * netlbl_mgmt_adddef - Handle an ADDDEF message
+ * @req_skb: the NETLINK buffer
+ * @msg: the NetLabel management message
+ *
+ * Description:
+ * Process a user generated ADDDEF message and respond accordingly.
+ *
+ */
+static void netlbl_mgmt_adddef(const struct sk_buff *req_skb,
+ const unsigned char *msg)
+{
+ int ret_val = -EINVAL;
+ unsigned char *msg_ptr = (unsigned char *)msg;
+ u32 msg_len = netlbl_mgmt_payload_len(req_skb);
+ struct netlbl_dom_map *entry = NULL;
+ u32 tmp_val;
+ u32 rcu_locked = 0;
+
+ if (msg_len < 4)
+ goto adddef_failure;
+ tmp_val = netlbl_getinc_u32(&msg_ptr);
+ msg_len -= 4;
+
+ entry = kzalloc(sizeof(*entry), GFP_KERNEL);
+ if (entry == NULL) {
+ ret_val = -ENOMEM;
+ goto adddef_failure;
+ }
+ entry->type = tmp_val;
+
+ switch (entry->type) {
+#ifdef CONFIG_NETLABEL_UNLABELED
+ case NETLBL_NLTYPE_UNLABELED:
+ break;
+#endif
+#ifdef CONFIG_NETLABEL_CIPSOV4
+ case NETLBL_NLTYPE_CIPSOV4:
+ if (msg_len < 4)
+ goto adddef_failure;
+ tmp_val = netlbl_getinc_u32(&msg_ptr);
+ msg_len -= 4;
+ /* XXX - we should be holding a rcu_read_lock here while we
+ hold the result but since the entry will always be
+ deleted when the CIPSO DOI is deleted we are going
+ to skip the lock */
+ rcu_read_lock();
+ entry->type_def.cipsov4 = cipso_v4_doi_getdef(tmp_val);
+ if (entry->type_def.cipsov4 == NULL) {
+ rcu_read_unlock();
+ goto adddef_failure;
+ }
+ rcu_locked = 1;
+ break;
+#endif
+ default:
+ ret_val = -EINVAL;
+ goto adddef_failure;
+ }
+
+ ret_val = netlbl_domhsh_add_default(entry);
+ if (rcu_locked) {
+ rcu_read_unlock();
+ rcu_locked = 0;
+ }
+ if (ret_val != 0)
+ goto adddef_failure;
+
+ netlbl_mgmt_send_ack(req_skb, NETLBL_E_OK);
+ return;
+
+adddef_failure:
+ if (entry)
+ kfree(entry);
+ netlbl_mgmt_send_ack(req_skb, -ret_val);
+}
+
+/**
+ * netlbl_mgmt_removedef - Handle a REMOVEDEF message
+ * @req_skb: the NETLINK buffer
+ * @msg: the NetLabel management message
+ *
+ * Description:
+ * Process a user generated REMOVEDEF message and remove the default domain
+ * mapping.
+ *
+ */
+static void netlbl_mgmt_removedef(const struct sk_buff *req_skb,
+ const unsigned char *msg)
+{
+ int ret_val;
+
+ ret_val = netlbl_domhsh_remove_default();
+ netlbl_mgmt_send_ack(req_skb, -ret_val);
+}
+
+/**
+ * netlbl_mgmt_listdef - Handle a LISTDEF message
+ * @req_skb: the NETLINK buffer
+ * @msg: the NetLabel management message
+ *
+ * Description:
+ * Process a user generated LISTDEF message and dumps the default domain
+ * mapping in a form suitable for use in a kernel generated LISTDEF message.
+ *
+ */
+static void netlbl_mgmt_listdef(const struct sk_buff *req_skb,
+ const unsigned char *msg)
+{
+ int ret_val = -ENOMEM;
+ struct sk_buff *skb;
+ unsigned char *buf_ptr;
+
+ skb = netlbl_domhsh_dump_default();
+ if (skb == NULL)
+ goto listdef_return;
+ buf_ptr = skb_push(skb,
+ NLMSG_SPACE(sizeof(struct netlbl_mgmt_msghdr)));
+ if (buf_ptr == NULL) {
+ kfree_skb(skb);
+ ret_val = -EAGAIN;
+ goto listdef_return;
+ }
+ netlbl_putinc_hdr(&buf_ptr,
+ NETLBL_NLTYPE_MGMT,
+ skb->len,
+ 0,
+ NETLINK_CB(req_skb).pid,
+ 0);
+ netlbl_mgmt_putinc_hdr(&buf_ptr, NL_MGMT_LISTDEF);
+
+ ret_val = netlbl_netlink_snd(skb, NETLINK_CB(req_skb).pid);
+
+listdef_return:
+ if (ret_val != 0)
+ netlbl_mgmt_send_ack(req_skb, -ret_val);
+}
+
+/**
+ * netlbl_mgmt_modules - Handle a MODULES message
+ * @req_skb: the NETLINK buffer
+ * @msg: the NetLabel management message
+ *
+ * Description:
+ * Process a user generated MODULES message and respond accordingly.
+ *
+ */
+static void netlbl_mgmt_modules(const struct sk_buff *req_skb,
+ const unsigned char *msg)
+{
+ int ret_val = -EINVAL;
+ u32 mod_count = 0;
+ size_t msg_size;
+ size_t data_size;
+ struct sk_buff *skb = NULL;
+ struct nlmsghdr *nl_mgmt_hdr;
+ unsigned char *data;
+
+ if (netlbl_mgmt_payload_len(req_skb) < 4)
+ goto nlmsg_failure;
+ mod_count = netlbl_get_u32(msg);
+ if (mod_count != 0)
+ goto nlmsg_failure;
+
+#ifdef CONFIG_NETLABEL_UNLABELED
+ mod_count += 1;
+#endif
+#ifdef CONFIG_NETLABEL_CIPSOV4
+ mod_count += 1;
+#endif
+
+ data_size = sizeof(struct netlbl_mgmt_msghdr) + 4 + mod_count * 4;
+ msg_size = NLMSG_SPACE(data_size);
+
+ skb = alloc_skb(msg_size, GFP_KERNEL);
+ if (skb == NULL) {
+ ret_val = -ENOMEM;
+ goto nlmsg_failure;
+ }
+
+ nl_mgmt_hdr = NLMSG_PUT(skb,
+ NETLINK_CB(req_skb).pid,
+ 0,
+ NETLBL_NLTYPE_MGMT,
+ data_size);
+ nl_mgmt_hdr->nlmsg_len = msg_size;
+
+ data = NLMSG_DATA(nl_mgmt_hdr);
+ netlbl_mgmt_putinc_hdr(&data, NL_MGMT_MODULES);
+ netlbl_putinc_u32(&data, mod_count);
+#ifdef CONFIG_NETLABEL_UNLABELED
+ netlbl_putinc_u32(&data, NETLBL_NLTYPE_UNLABELED);
+#endif
+#ifdef CONFIG_NETLABEL_CIPSOV4
+ netlbl_putinc_u32(&data, NETLBL_NLTYPE_CIPSOV4);
+#endif
+
+ netlbl_netlink_snd(skb, NETLINK_CB(req_skb).pid);
+ return;
+
+nlmsg_failure:
+ if (skb)
+ kfree_skb(skb);
+ netlbl_mgmt_send_ack(req_skb, -ret_val);
+}
+
+/**
+ * netlbl_mgmt_version - Handle a VERSION message
+ * @req_skb: the NETLINK buffer
+ * @msg: the NetLabel management message
+ *
+ * Description:
+ * Process a user generated VERSION message and respond accordingly.
+ *
+ */
+static void netlbl_mgmt_version(const struct sk_buff *req_skb,
+ const unsigned char *msg)
+{
+ int ret_val = -EINVAL;
+ u32 str_len;
+ size_t msg_size;
+ size_t data_size;
+ struct sk_buff *skb = NULL;
+ struct nlmsghdr *nl_mgmt_hdr;
+ unsigned char *data;
+
+ if (netlbl_mgmt_payload_len(req_skb) < 4)
+ goto nlmsg_failure;
+ str_len = netlbl_get_u32(msg);
+ if (str_len != 0)
+ goto nlmsg_failure;
+
+ data_size = sizeof(struct netlbl_mgmt_msghdr) + 4;
+ msg_size = NLMSG_SPACE(data_size);
+
+ skb = alloc_skb(msg_size, GFP_KERNEL);
+ if (skb == NULL) {
+ ret_val = -ENOMEM;
+ goto nlmsg_failure;
+ }
+
+ nl_mgmt_hdr = NLMSG_PUT(skb,
+ NETLINK_CB(req_skb).pid,
+ 0,
+ NETLBL_NLTYPE_MGMT,
+ data_size);
+ nl_mgmt_hdr->nlmsg_len = msg_size;
+
+ data = NLMSG_DATA(nl_mgmt_hdr);
+ netlbl_mgmt_putinc_hdr(&data, NL_MGMT_VERSION);
+ netlbl_putinc_u32(&data, NETLBL_PROTO_VERSION);
+
+ netlbl_netlink_snd(skb, NETLINK_CB(req_skb).pid);
+ return;
+
+nlmsg_failure:
+ if (skb)
+ kfree_skb(skb);
+ netlbl_mgmt_send_ack(req_skb, -ret_val);
+}
+
+/*
+ * NetLabel Protocol Handlers
+ */
+
+/**
+ * netlbl_mgmt_send_ack - Send an ACK message
+ * @req_skb: the NETLINK buffer
+ * @ret_code: return code to use
+ *
+ * Description:
+ * This function sends an ACK message to the sender of the NETLINK message
+ * specified by @req_skb. Returns negative values on error.
+ *
+ */
+static void netlbl_mgmt_send_ack(const struct sk_buff *req_skb,
+ const u32 ret_code)
+{
+ size_t msg_size;
+ size_t data_size;
+ struct sk_buff *skb;
+ struct nlmsghdr *nl_ack_hdr;
+ unsigned char *data;
+
+ data_size = sizeof(struct netlbl_mgmt_msghdr) + 8;
+ msg_size = NLMSG_SPACE(data_size);
+
+ skb = alloc_skb(msg_size, GFP_KERNEL);
+ if (skb == NULL)
+ return;
+
+ nl_ack_hdr = NLMSG_PUT(skb,
+ NETLINK_CB(req_skb).pid,
+ 0,
+ NETLBL_NLTYPE_MGMT,
+ data_size);
+ nl_ack_hdr->nlmsg_len = msg_size;
+
+ data = NLMSG_DATA(nl_ack_hdr);
+ netlbl_mgmt_putinc_hdr(&data, NL_MGMT_ACK);
+ netlbl_putinc_u32(&data,
+ ((struct nlmsghdr *)req_skb->data)->nlmsg_seq);
+ netlbl_putinc_u32(&data, ret_code);
+
+ netlbl_netlink_snd(skb, NETLINK_CB(req_skb).pid);
+ return;
+
+nlmsg_failure:
+ kfree_skb(skb);
+}
+
+/**
+ * netlbl_mgmt_rcv - Process incoming NetLabel packets
+ * @skb: the NETLINK buffer
+ * @msg: pointer to the start of the NetLabel data
+ *
+ * Description:
+ * This function is reponsibile for reading all of the incoming NetLabel
+ * management traffic and dispatching it to the correct functions.
+ *
+ */
+void netlbl_mgmt_rcv(const struct sk_buff *skb, const unsigned char *msg)
+{
+ int ret_val;
+
+ if (nlmsg_len((struct nlmsghdr *)skb->data) <
+ sizeof(struct netlbl_mgmt_msghdr)) {
+ netlbl_mgmt_send_ack(skb, EINVAL);
+ return;
+ }
+
+ switch (((struct netlbl_mgmt_msghdr *)msg)->opcode) {
+ case NL_MGMT_ADD:
+ ret_val = netlbl_netlink_cap_check(skb, CAP_NET_ADMIN);
+ if (ret_val != 0)
+ netlbl_mgmt_send_ack(skb, ret_val);
+ netlbl_mgmt_add(skb, netlbl_mgmt_payload_data(skb));
+ break;
+ case NL_MGMT_REMOVE:
+ ret_val = netlbl_netlink_cap_check(skb, CAP_NET_ADMIN);
+ if (ret_val != 0)
+ netlbl_mgmt_send_ack(skb, ret_val);
+ netlbl_mgmt_remove(skb, netlbl_mgmt_payload_data(skb));
+ break;
+ case NL_MGMT_LIST:
+ netlbl_mgmt_list(skb, netlbl_mgmt_payload_data(skb));
+ break;
+ case NL_MGMT_ADDDEF:
+ ret_val = netlbl_netlink_cap_check(skb, CAP_NET_ADMIN);
+ if (ret_val != 0)
+ netlbl_mgmt_send_ack(skb, ret_val);
+ netlbl_mgmt_adddef(skb, netlbl_mgmt_payload_data(skb));
+ break;
+ case NL_MGMT_REMOVEDEF:
+ ret_val = netlbl_netlink_cap_check(skb, CAP_NET_ADMIN);
+ if (ret_val != 0)
+ netlbl_mgmt_send_ack(skb, ret_val);
+ netlbl_mgmt_removedef(skb, netlbl_mgmt_payload_data(skb));
+ break;
+ case NL_MGMT_LISTDEF:
+ netlbl_mgmt_listdef(skb, netlbl_mgmt_payload_data(skb));
+ break;
+ case NL_MGMT_MODULES:
+ netlbl_mgmt_modules(skb, netlbl_mgmt_payload_data(skb));
+ break;
+ case NL_MGMT_VERSION:
+ netlbl_mgmt_version(skb, netlbl_mgmt_payload_data(skb));
+ break;
+ default:
+ netlbl_mgmt_send_ack(skb, EINVAL);
+ return;
+ }
+}
Index: linux-2.6.17.i686-quilt/net/netlabel/netlabel_mgmt.h
===================================================================
--- /dev/null
+++ linux-2.6.17.i686-quilt/net/netlabel/netlabel_mgmt.h
@@ -0,0 +1,248 @@
+/*
+ * NetLabel Management Support
+ *
+ * This file defines the management functions for the NetLabel system. The
+ * NetLabel system manages static and dynamic label mappings for network
+ * protocols such as CIPSO and RIPSO.
+ *
+ * Author: Paul Moore <paul.moore@hp.com>
+ *
+ */
+
+/*
+ * (c) Copyright Hewlett-Packard Development Company, L.P., 2006
+ *
+ * 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 _NETLABEL_MGMT_H
+#define _NETLABEL_MGMT_H
+
+#include <net/netlabel.h>
+
+/*
+ * The following NetLabel payloads are supported by the management interface,
+ * all of which are preceeded by the nlmsghdr struct.
+ *
+ * o ACK:
+ * Sent by the kernel in response to an applications message, applications
+ * should never send this message.
+ *
+ * +----------------------+-----------------------+
+ * | seq number (32 bits) | return code (32 bits) |
+ * +----------------------+-----------------------+
+ *
+ * seq number: the sequence number of the original message, taken from the
+ * nlmsghdr structure
+ * return code: return value, based on errno values
+ *
+ * o ADD:
+ * Sent by an application to add a domain mapping to the NetLabel system.
+ * The kernel should respond with an ACK.
+ *
+ * +-------------------+
+ * | domains (32 bits) | ...
+ * +-------------------+
+ *
+ * domains: the number of domains in the message
+ *
+ * +-------------------------+--------------------------+
+ * | domain length (32 bits) | domain string (variable) | ...
+ * +-------------------------+--------------------------+
+ *
+ * +-------------------------+-------------- ---- --- -- -
+ * | protocol type (32 bits) | mapping data ... repeated
+ * +-------------------------+-------------- ---- --- -- -
+ *
+ * domain length: the length of the "domain string" in bytes
+ * domain string: the domain string, NULL terminated
+ * protocol type: the protocol type (defined by NETLBL_NLTYPE_*)
+ * mapping data: specific to the map type (see below)
+ *
+ * NETLBL_NLTYPE_UNLABELED
+ *
+ * No mapping data for this protocol type.
+ *
+ * NETLBL_NLTYPE_CIPSOV4
+ *
+ * +---------------+
+ * | doi (32 bits) |
+ * +---------------+
+ *
+ * doi: the CIPSO DOI value
+ *
+ * o REMOVE:
+ * Sent by an application to remove a domain mapping from the NetLabel
+ * system. The kernel should ACK this message.
+ *
+ * +-------------------+
+ * | domains (32 bits) | ...
+ * +-------------------+
+ *
+ * domains: the number of domains in the message
+ *
+ * +-------------------------+--------------------------+
+ * | domain length (32 bits) | domain string (variable) | ...
+ * +-------------------------+--------------------------+
+ *
+ * domain length: the length of the "domain string" in bytes
+ * domain string: the domain string, NULL terminated
+ *
+ * o LIST:
+ * This message can be sent either from an application or by the kernel in
+ * response to an application generated LIST message. When sent by an
+ * application there is no payload. The kernel should respond to a LIST
+ * message either with a LIST message on success or an ACK message on
+ * failure.
+ *
+ * +-------------------+
+ * | domains (32 bits) | ...
+ * +-------------------+
+ *
+ * domains: the number of domains in the message
+ *
+ * +-------------------------+--------------------------+
+ * | domain length (32 bits) | domain string (variable) | ...
+ * +-------------------------+--------------------------+
+ *
+ * +-------------------------+-------------- ---- --- -- -
+ * | protocol type (32 bits) | mapping data ... repeated
+ * +-------------------------+-------------- ---- --- -- -
+ *
+ * domain length: the length of the "domain string" in bytes
+ * domain string: the domain string, NULL terminated
+ * protocol type: the protocol type (defined by NETLBL_NLTYPE_*)
+ * mapping data: specific to the map type (see below)
+ *
+ * NETLBL_NLTYPE_UNLABELED
+ *
+ * No mapping data for this protocol type.
+ *
+ * NETLBL_NLTYPE_CIPSOV4
+ *
+ * +----------------+---------------+
+ * | type (32 bits) | doi (32 bits) |
+ * +----------------+---------------+
+ *
+ * type: the CIPSO mapping table type (defined in the cipso_ipv4.h header
+ * as CIPSO_V4_MAP_*)
+ * doi: the CIPSO DOI value
+ *
+ * o ADDDEF:
+ * Sent by an application to set the default domain mapping for the NetLabel
+ * system. The kernel should respond with an ACK.
+ *
+ * +-------------------------+-------------- ---- --- -- -
+ * | protocol type (32 bits) | mapping data ... repeated
+ * +-------------------------+-------------- ---- --- -- -
+ *
+ * protocol type: the protocol type (defined by NETLBL_NLTYPE_*)
+ * mapping data: specific to the map type (see below)
+ *
+ * NETLBL_NLTYPE_UNLABELED
+ *
+ * No mapping data for this protocol type.
+ *
+ * NETLBL_NLTYPE_CIPSOV4
+ *
+ * +---------------+
+ * | doi (32 bits) |
+ * +---------------+
+ *
+ * doi: the CIPSO DOI value
+ *
+ * o REMOVEDEF:
+ * Sent by an application to remove the default domain mapping from the
+ * NetLabel system, there is no payload. The kernel should ACK this message.
+ *
+ * o LISTDEF:
+ * This message can be sent either from an application or by the kernel in
+ * response to an application generated LISTDEF message. When sent by an
+ * application there is no payload. The kernel should respond to a
+ * LISTDEF message either with a LISTDEF message on success or an ACK message
+ * on failure.
+ *
+ * +-------------------------+-------------- ---- --- -- -
+ * | protocol type (32 bits) | mapping data ... repeated
+ * +-------------------------+-------------- ---- --- -- -
+ *
+ * protocol type: the protocol type (defined by NETLBL_NLTYPE_*)
+ * mapping data: specific to the map type (see below)
+ *
+ * NETLBL_NLTYPE_UNLABELED
+ *
+ * No mapping data for this protocol type.
+ *
+ * NETLBL_NLTYPE_CIPSOV4
+ *
+ * +----------------+---------------+
+ * | type (32 bits) | doi (32 bits) |
+ * +----------------+---------------+
+ *
+ * type: the CIPSO mapping table type (defined in the cipso_ipv4.h header
+ * as CIPSO_V4_MAP_*)
+ * doi: the CIPSO DOI value
+ *
+ * o MODULES:
+ * Sent by an application to request a list of configured NetLabel modules
+ * in the kernel. This message type is also used by the kernel to respond
+ * to an application's MODULES message.
+ *
+ * +-------------------+
+ * | modules (32 bits) | ...
+ * +-------------------+
+ *
+ * modules: the number of modules in the message, if this is an application
+ * generated message and the value is zero then return a list of
+ * the configured modules
+ *
+ * +------------------+
+ * | module (32 bits) | ... repeated
+ * +------------------+
+ *
+ * module: the module number as defined by NETLBL_NLTYPE_*
+ *
+ * o VERSION:
+ * Sent by an application to request the NetLabel version string. This
+ * message type is also used by the kernel to respond to an VERSION request.
+ *
+ * +-------------------+
+ * | version (32 bits) |
+ * +-------------------+
+ *
+ * version: the protocol version number
+ *
+ */
+
+/* MGMT message header */
+struct netlbl_mgmt_msghdr {
+ enum { NL_MGMT_NOOP,
+ NL_MGMT_ACK,
+ NL_MGMT_ADD,
+ NL_MGMT_REMOVE,
+ NL_MGMT_LIST,
+ NL_MGMT_ADDDEF,
+ NL_MGMT_REMOVEDEF,
+ NL_MGMT_LISTDEF,
+ NL_MGMT_MODULES,
+ NL_MGMT_VERSION
+ } opcode;
+};
+
+/* Process NetLabel management messages */
+void netlbl_mgmt_rcv(const struct sk_buff *skb, const unsigned char *msg);
+
+#endif
Index: linux-2.6.17.i686-quilt/net/netlabel/netlabel_unlabeled.h
===================================================================
--- /dev/null
+++ linux-2.6.17.i686-quilt/net/netlabel/netlabel_unlabeled.h
@@ -0,0 +1,105 @@
+/*
+ * NetLabel Unlabeled Support
+ *
+ * This file defines functions for dealing with unlabeled packets for the
+ * NetLabel system. The NetLabel system manages static and dynamic label
+ * mappings for network protocols such as CIPSO and RIPSO.
+ *
+ * Author: Paul Moore <paul.moore@hp.com>
+ *
+ */
+
+/*
+ * (c) Copyright Hewlett-Packard Development Company, L.P., 2006
+ *
+ * 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 _NETLABEL_UNLABELED_H
+#define _NETLABEL_UNLABELED_H
+
+#include <net/netlabel.h>
+
+/*
+ * The following NetLabel payloads are supported by the Unlabeled subsystem,
+ * all of which are preceeded by the nlmsghdr struct.
+ *
+ * o ACK:
+ * Sent by the kernel in response to an applications message, applications
+ * should never send this message.
+ *
+ * +----------------------+-----------------------+
+ * | seq number (32 bits) | return code (32 bits) |
+ * +----------------------+-----------------------+
+ *
+ * seq number: the sequence number of the original message, taken from the
+ * nlmsghdr structure
+ * return code: return value, based on errno values
+ *
+ * o ACCEPT
+ * This message is sent from an application to specify if the kernel should
+ * allow unlabled packets to pass if they do not match any of the static
+ * mappings defined in the unlabeled module.
+ *
+ * +-----------------+
+ * | allow (32 bits) |
+ * +-----------------+
+ *
+ * allow: if true (1) then allow the packets to pass, if false (0) then
+ * reject the packets
+ *
+ */
+
+/* Unlabeled message header */
+struct netlbl_unlabel_msghdr {
+ enum { NL_UNL_NOOP,
+ NL_UNL_ACK,
+ NL_UNL_ACCEPT
+ } opcode;
+};
+
+/* Process Unlabeled NetLabel messages */
+#ifdef CONFIG_NETLABEL_UNLABELED
+void netlbl_unlabel_rcv(const struct sk_buff *skb, const unsigned char *msg);
+#else
+static inline void netlbl_unlabel_rcv(const struct sk_buff *skb,
+ const unsigned char *msg)
+{
+ return;
+}
+#endif
+
+/* Process Unlabeled incoming network packets */
+#ifdef CONFIG_NETLABEL_UNLABELED
+int netlbl_unlabel_getattr(struct netlbl_lsm_secattr *secattr);
+#else
+static inline int netlbl_unlabel_getattr(struct netlbl_lsm_secattr *secattr)
+{
+ return -ENOSYS;
+}
+#endif
+
+/* Set the default configuration to allow Unlabeled packets */
+#ifdef CONFIG_NETLABEL_UNLABELED
+int netlbl_unlabel_defconf(void);
+#else
+static inline int netlbl_unlabel_defconf(void)
+{
+ return -ENOSYS;
+}
+#endif
+
+#endif
Index: linux-2.6.17.i686-quilt/net/netlabel/netlabel_user.c
===================================================================
--- /dev/null
+++ linux-2.6.17.i686-quilt/net/netlabel/netlabel_user.c
@@ -0,0 +1,162 @@
+/*
+ * NetLabel NETLINK Interface
+ *
+ * This file defines the NETLINK interface for the NetLabel system. The
+ * NetLabel system manages static and dynamic label mappings for network
+ * protocols such as CIPSO and RIPSO.
+ *
+ * Author: Paul Moore <paul.moore@hp.com>
+ *
+ */
+
+/*
+ * (c) Copyright Hewlett-Packard Development Company, L.P., 2006
+ *
+ * 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
+ *
+ */
+
+#include <linux/init.h>
+#include <linux/types.h>
+#include <linux/socket.h>
+#include <net/sock.h>
+#include <net/netlink.h>
+#include <net/netlabel.h>
+#include <asm/bug.h>
+
+#include "netlabel_mgmt.h"
+#include "netlabel_unlabeled.h"
+#include "netlabel_cipso_v4.h"
+#include "netlabel_user.h"
+
+/* NETLINK socket */
+/* PM - do we need a lock (or something) around this? */
+static struct sock *netlbl_nl = NULL;
+
+/*
+ * NetLabel Functions
+ */
+
+/**
+ * netlbl_netlink_rcv - Catch incoming NETLINK packets
+ * @sk: the NETLINK socket
+ * @len: absolutely no clue, socket queue length maybe?
+ *
+ * Description:
+ * Receives the NETLINK packet, inspects the packet to determine the correct
+ * NetLabel subsystem and hands the packet off to the correct handler.
+ *
+ */
+static void netlbl_netlink_rcv(struct sock *sk, const int len)
+{
+ struct sk_buff *skb;
+ struct nlmsghdr *nl_hdr;
+
+ /* XXX - should this use netlink_run_queue() instead? */
+ while ((skb = skb_dequeue(&sk->sk_receive_queue)) != NULL) {
+ nl_hdr = (struct nlmsghdr *)skb->data;
+ switch (nl_hdr->nlmsg_type) {
+ case NETLBL_NLTYPE_MGMT:
+ netlbl_mgmt_rcv(skb, nlmsg_data(nl_hdr));
+ break;
+#ifdef CONFIG_NETLABEL_UNLABELED
+ case NETLBL_NLTYPE_UNLABELED:
+ netlbl_unlabel_rcv(skb, nlmsg_data(nl_hdr));
+ break;
+#endif
+#ifdef CONFIG_NETLABEL_CIPSOV4
+ case NETLBL_NLTYPE_CIPSOV4:
+ netlbl_cipsov4_rcv(skb, nlmsg_data(nl_hdr));
+ break;
+#endif
+ }
+ nlmsg_free(skb);
+ }
+}
+
+/**
+ * netlbl_netlink_init - Initialize the netlink socket
+ *
+ * Description:
+ * Create the netlink socket and do any other setup required. Returns zero on
+ * success and non-zero on failure.
+ *
+ */
+int netlbl_netlink_init(void)
+{
+ /* XXX - it might be a good idea to spawn a thread here at startup to
+ handle the updates to the label mapping databases and have the
+ netlbl_netlink_rcv() function simply poke the thread but i am unsure
+ and this approach is much simpler so it wins for now */
+ netlbl_nl = netlink_kernel_create(NETLINK_NETLABEL,
+ NETLBL_NLGRP_MAX,
+ netlbl_netlink_rcv, THIS_MODULE);
+ if (netlbl_nl == NULL)
+ return -ENOMEM;
+
+ return 0;
+}
+
+/**
+ * netlbl_exit - Release the netlink socket
+ *
+ * Description:
+ * Close the netlink socket and do any other cleanup required.
+ *
+ */
+void netlbl_netlink_exit(void)
+{
+ /* XXX - if we do have a worker thread, see the above comment in
+ netlbl_netlink_init(), we should handle any cleanup here
+ before we go away */
+ if (netlbl_nl != NULL)
+ sock_release(netlbl_nl->sk_socket);
+}
+
+/*
+ * NETLINK I/O Functions
+ */
+
+/**
+ * netlbl_netlink_snd - Send a NetLabel message
+ * @skb: NetLabel message
+ * @pid: destination PID
+ *
+ * Description:
+ * Sends a unicast NetLabel message over the NETLINK socket.
+ *
+ */
+int netlbl_netlink_snd(struct sk_buff *skb, const u32 pid)
+{
+ return nlmsg_unicast(netlbl_nl, skb, pid);
+}
+
+/**
+ * netlbl_netlink_snd - Send a NetLabel message
+ * @skb: NetLabel message
+ * @pid: sending PID
+ * @group: multicast group id
+ *
+ * Description:
+ * Sends a multicast NetLabel message over the NETLINK socket to all members
+ * of @group except @pid.
+ *
+ */
+int netlbl_netlink_snd_multicast(struct sk_buff *skb,
+ const u32 pid,
+ const u32 group)
+{
+ return nlmsg_multicast(netlbl_nl, skb, pid, group);
+}
Index: linux-2.6.17.i686-quilt/net/netlabel/netlabel_user.h
===================================================================
--- /dev/null
+++ linux-2.6.17.i686-quilt/net/netlabel/netlabel_user.h
@@ -0,0 +1,67 @@
+/*
+ * NetLabel NETLINK Interface
+ *
+ * This file defines the NETLINK interface for the NetLabel system. The
+ * NetLabel system manages static and dynamic label mappings for network
+ * protocols such as CIPSO and RIPSO.
+ *
+ * Author: Paul Moore <paul.moore@hp.com>
+ *
+ */
+
+/*
+ * (c) Copyright Hewlett-Packard Development Company, L.P., 2006
+ *
+ * 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 _NETLABEL_USER_H
+#define _NETLABEL_USER_H
+
+#include <linux/skbuff.h>
+#include <linux/capability.h>
+
+/* NetLabel NETLINK helper functions */
+
+/**
+ * netlbl_netlink_cap_check - Check the NETLINK msg capabilities
+ * @skb: the NETLINK buffer
+ * @req_cap: the required capability
+ *
+ * Description:
+ * Check the NETLINK buffer's capabilities against the required capabilities.
+ * Returns zero on success, negative values on failure.
+ *
+ */
+static inline int netlbl_netlink_cap_check(const struct sk_buff *skb,
+ const kernel_cap_t req_cap)
+{
+ if (cap_raised(NETLINK_CB(skb).eff_cap, req_cap))
+ return 0;
+ return -EPERM;
+}
+
+
+/* NetLabel NETLINK I/O functions */
+
+int netlbl_netlink_init(void);
+void netlbl_netlink_exit(void);
+int netlbl_netlink_snd(struct sk_buff *skb, const u32 pid);
+int netlbl_netlink_snd_multicast(struct sk_buff *skb,
+ const u32 pid,
+ const u32 group);
+
+#endif
--
paul moore
linux security @ hp
^ permalink raw reply
* [RFC 1/8] NetLabel: documentation
From: paul.moore @ 2006-06-22 22:49 UTC (permalink / raw)
To: netdev; +Cc: davem
In-Reply-To: <20060622224910.885573000@flek.zko.hp.com>
[-- Attachment #1: netlabel-documentation --]
[-- Type: text/plain, Size: 39964 bytes --]
Documentation for the NetLabel system, this includes a basic overview of how
NetLabel works and how LSM developers can integrate it into their favorite
LSM. Also, due to the difficulty of finding expired IETF drafts, I am
including the IETF CIPSO draft that is the basis of the NetLabel CIPSO
implementation.
---
CREDITS | 7
Documentation/00-INDEX | 2
Documentation/netlabel/00-INDEX | 10
Documentation/netlabel/cipso_ipv4.txt | 48
Documentation/netlabel/draft-ietf-cipso-ipsecurity-01.txt | 791 ++++++++++++++
Documentation/netlabel/introduction.txt | 52
Documentation/netlabel/lsm_interface.txt | 47
7 files changed, 957 insertions(+)
Index: linux-2.6.17.i686-quilt/CREDITS
===================================================================
--- linux-2.6.17.i686-quilt.orig/CREDITS
+++ linux-2.6.17.i686-quilt/CREDITS
@@ -2383,6 +2383,13 @@ N: Thomas Molina
E: tmolina@cablespeed.com
D: bug fixes, documentation, minor hackery
+N: Paul Moore
+E: paul.moore@hp.com
+D: NetLabel author
+S: Hewlett-Packard
+S: 110 Spit Brook Road
+S: Nashua, NH 03062
+
N: James Morris
E: jmorris@namei.org
W: http://namei.org/
Index: linux-2.6.17.i686-quilt/Documentation/00-INDEX
===================================================================
--- linux-2.6.17.i686-quilt.orig/Documentation/00-INDEX
+++ linux-2.6.17.i686-quilt/Documentation/00-INDEX
@@ -184,6 +184,8 @@ mtrr.txt
- how to use PPro Memory Type Range Registers to increase performance.
nbd.txt
- info on a TCP implementation of a network block device.
+netlabel/
+ - directory with information on the NetLabel subsystem.
networking/
- directory with info on various aspects of networking with Linux.
nfsroot.txt
Index: linux-2.6.17.i686-quilt/Documentation/netlabel/00-INDEX
===================================================================
--- /dev/null
+++ linux-2.6.17.i686-quilt/Documentation/netlabel/00-INDEX
@@ -0,0 +1,10 @@
+00-INDEX
+ - this file.
+cipso_ipv4.txt
+ - documentation on the IPv4 CIPSO protocol engine.
+draft-ietf-cipso-ipsecurity-01.txt
+ - IETF draft of the CIPSO protocol, dated 16 July 1992.
+introduction.txt
+ - NetLabel introduction, READ THIS FIRST.
+lsm_interface.txt
+ - documentation on the NetLabel kernel security module API.
Index: linux-2.6.17.i686-quilt/Documentation/netlabel/cipso_ipv4.txt
===================================================================
--- /dev/null
+++ linux-2.6.17.i686-quilt/Documentation/netlabel/cipso_ipv4.txt
@@ -0,0 +1,48 @@
+NetLabel CIPSO/IPv4 Protocol Engine
+==============================================================================
+Paul Moore, paul.moore@hp.com
+
+May 17, 2006
+
+ * Overview
+
+The NetLabel CIPSO/IPv4 protocol engine is based on the IETF Commercial IP
+Security Option (CIPSO) draft from July 16, 1992. A copy of this draft can be
+found in this directory, consult '00-INDEX' for the filename. While the IETF
+draft never made it to an RFC standard it has become a de-facto standard for
+labeled networking and is used in many trusted operating systems.
+
+ * Outbound Packet Processing
+
+The CIPSO/IPv4 protocol engine applies the CIPSO IP option to packets by
+adding the CIPSO label to the socket. This causes all packets leaving the
+system through the socket to have the CIPSO IP option applied. The socket's
+CIPSO label can be changed at any point in time, however, it is recommended
+that it is set upon the socket's creation. The LSM can set the socket's CIPSO
+label by using the NetLabel security module API; if the NetLabel "domain" is
+configured to use CIPSO for packet labeling then a CIPSO IP option will be
+generated and attached to the socket.
+
+ * Inbound Packet Processing
+
+The CIPSO/IPv4 protocol engine validates every CIPSO IP option it finds at the
+IP layer without any special handling required by the LSM. However, in order
+to decode and translate the CIPSO label on the packet the LSM must use the
+NetLabel security module API to extract the security attributes of the packet.
+This is typically done at the socket layer using the 'socket_sock_rcv_skb()'
+LSM hook.
+
+ * Label Translation
+
+The CIPSO/IPv4 protocol engine contains a mechanism to translate CIPSO security
+attributes such as sensitivity level and category to values which are
+appropriate for the host. These mappings are defined as part of a CIPSO
+Domain Of Interpretation (DOI) definition and are configured through the
+NetLabel user space communication layer. Each DOI definition can have a
+different security attribute mapping table.
+
+ * Label Translation Cache
+
+The NetLabel system provides a framework for caching security attribute
+mappings from the network labels to the corresponding LSM identifiers. The
+CIPSO/IPv4 protocol engine supports this caching mechanism.
Index: linux-2.6.17.i686-quilt/Documentation/netlabel/draft-ietf-cipso-ipsecurity-01.txt
===================================================================
--- /dev/null
+++ linux-2.6.17.i686-quilt/Documentation/netlabel/draft-ietf-cipso-ipsecurity-01.txt
@@ -0,0 +1,791 @@
+IETF CIPSO Working Group
+16 July, 1992
+
+
+
+ COMMERCIAL IP SECURITY OPTION (CIPSO 2.2)
+
+
+
+1. Status
+
+This Internet Draft provides the high level specification for a Commercial
+IP Security Option (CIPSO). This draft reflects the version as approved by
+the CIPSO IETF Working Group. Distribution of this memo is unlimited.
+
+This document is an Internet Draft. Internet Drafts are working documents
+of the Internet Engineering Task Force (IETF), its Areas, and its Working
+Groups. Note that other groups may also distribute working documents as
+Internet Drafts.
+
+Internet Drafts are draft documents valid for a maximum of six months.
+Internet Drafts may be updated, replaced, or obsoleted by other documents
+at any time. It is not appropriate to use Internet Drafts as reference
+material or to cite them other than as a "working draft" or "work in
+progress."
+
+Please check the I-D abstract listing contained in each Internet Draft
+directory to learn the current status of this or any other Internet Draft.
+
+
+
+
+2. Background
+
+Currently the Internet Protocol includes two security options. One of
+these options is the DoD Basic Security Option (BSO) (Type 130) which allows
+IP datagrams to be labeled with security classifications. This option
+provides sixteen security classifications and a variable number of handling
+restrictions. To handle additional security information, such as security
+categories or compartments, another security option (Type 133) exists and
+is referred to as the DoD Extended Security Option (ESO). The values for
+the fixed fields within these two options are administered by the Defense
+Information Systems Agency (DISA).
+
+Computer vendors are now building commercial operating systems with
+mandatory access controls and multi-level security. These systems are
+no longer built specifically for a particular group in the defense or
+intelligence communities. They are generally available commercial systems
+for use in a variety of government and civil sector environments.
+
+The small number of ESO format codes can not support all the possible
+applications of a commercial security option. The BSO and ESO were
+designed to only support the United States DoD. CIPSO has been designed
+to support multiple security policies. This Internet Draft provides the
+format and procedures required to support a Mandatory Access Control
+security policy. Support for additional security policies shall be
+defined in future RFCs.
+
+
+
+
+Internet Draft, Expires 15 Jan 93 [PAGE 1]
+
+
+
+CIPSO INTERNET DRAFT 16 July, 1992
+
+
+
+
+3. CIPSO Format
+
+Option type: 134 (Class 0, Number 6, Copy on Fragmentation)
+Option length: Variable
+
+This option permits security related information to be passed between
+systems within a single Domain of Interpretation (DOI). A DOI is a
+collection of systems which agree on the meaning of particular values
+in the security option. An authority that has been assigned a DOI
+identifier will define a mapping between appropriate CIPSO field values
+and their human readable equivalent. This authority will distribute that
+mapping to hosts within the authority's domain. These mappings may be
+sensitive, therefore a DOI authority is not required to make these
+mappings available to anyone other than the systems that are included in
+the DOI.
+
+This option MUST be copied on fragmentation. This option appears at most
+once in a datagram. All multi-octet fields in the option are defined to be
+transmitted in network byte order. The format of this option is as follows:
+
++----------+----------+------//------+-----------//---------+
+| 10000110 | LLLLLLLL | DDDDDDDDDDDD | TTTTTTTTTTTTTTTTTTTT |
++----------+----------+------//------+-----------//---------+
+
+ TYPE=134 OPTION DOMAIN OF TAGS
+ LENGTH INTERPRETATION
+
+
+ Figure 1. CIPSO Format
+
+
+3.1 Type
+
+This field is 1 octet in length. Its value is 134.
+
+
+3.2 Length
+
+This field is 1 octet in length. It is the total length of the option
+including the type and length fields. With the current IP header length
+restriction of 40 octets the value of this field MUST not exceed 40.
+
+
+3.3 Domain of Interpretation Identifier
+
+This field is an unsigned 32 bit integer. The value 0 is reserved and MUST
+not appear as the DOI identifier in any CIPSO option. Implementations
+should assume that the DOI identifier field is not aligned on any particular
+byte boundary.
+
+To conserve space in the protocol, security levels and categories are
+represented by numbers rather than their ASCII equivalent. This requires
+a mapping table within CIPSO hosts to map these numbers to their
+corresponding ASCII representations. Non-related groups of systems may
+
+
+
+Internet Draft, Expires 15 Jan 93 [PAGE 2]
+
+
+
+CIPSO INTERNET DRAFT 16 July, 1992
+
+
+
+have their own unique mappings. For example, one group of systems may
+use the number 5 to represent Unclassified while another group may use the
+number 1 to represent that same security level. The DOI identifier is used
+to identify which mapping was used for the values within the option.
+
+
+3.4 Tag Types
+
+A common format for passing security related information is necessary
+for interoperability. CIPSO uses sets of "tags" to contain the security
+information relevant to the data in the IP packet. Each tag begins with
+a tag type identifier followed by the length of the tag and ends with the
+actual security information to be passed. All multi-octet fields in a tag
+are defined to be transmitted in network byte order. Like the DOI
+identifier field in the CIPSO header, implementations should assume that
+all tags, as well as fields within a tag, are not aligned on any particular
+octet boundary. The tag types defined in this document contain alignment
+bytes to assist alignment of some information, however alignment can not
+be guaranteed if CIPSO is not the first IP option.
+
+CIPSO tag types 0 through 127 are reserved for defining standard tag
+formats. Their definitions will be published in RFCs. Tag types whose
+identifiers are greater than 127 are defined by the DOI authority and may
+only be meaningful in certain Domains of Interpretation. For these tag
+types, implementations will require the DOI identifier as well as the tag
+number to determine the security policy and the format associated with the
+tag. Use of tag types above 127 are restricted to closed networks where
+interoperability with other networks will not be an issue. Implementations
+that support a tag type greater than 127 MUST support at least one DOI that
+requires only tag types 1 to 127.
+
+Tag type 0 is reserved. Tag types 1, 2, and 5 are defined in this
+Internet Draft. Types 3 and 4 are reserved for work in progress.
+The standard format for all current and future CIPSO tags is shown below:
+
++----------+----------+--------//--------+
+| TTTTTTTT | LLLLLLLL | IIIIIIIIIIIIIIII |
++----------+----------+--------//--------+
+ TAG TAG TAG
+ TYPE LENGTH INFORMATION
+
+ Figure 2: Standard Tag Format
+
+In the three tag types described in this document, the length and count
+restrictions are based on the current IP limitation of 40 octets for all
+IP options. If the IP header is later expanded, then the length and count
+restrictions specified in this document may increase to use the full area
+provided for IP options.
+
+
+3.4.1 Tag Type Classes
+
+Tag classes consist of tag types that have common processing requirements
+and support the same security policy. The three tags defined in this
+Internet Draft belong to the Mandatory Access Control (MAC) Sensitivity
+
+
+
+Internet Draft, Expires 15 Jan 93 [PAGE 3]
+
+
+
+CIPSO INTERNET DRAFT 16 July, 1992
+
+
+
+class and support the MAC Sensitivity security policy.
+
+
+3.4.2 Tag Type 1
+
+This is referred to as the "bit-mapped" tag type. Tag type 1 is included
+in the MAC Sensitivity tag type class. The format of this tag type is as
+follows:
+
++----------+----------+----------+----------+--------//---------+
+| 00000001 | LLLLLLLL | 00000000 | LLLLLLLL | CCCCCCCCCCCCCCCCC |
++----------+----------+----------+----------+--------//---------+
+
+ TAG TAG ALIGNMENT SENSITIVITY BIT MAP OF
+ TYPE LENGTH OCTET LEVEL CATEGORIES
+
+ Figure 3. Tag Type 1 Format
+
+
+3.4.2.1 Tag Type
+
+This field is 1 octet in length and has a value of 1.
+
+
+3.4.2.2 Tag Length
+
+This field is 1 octet in length. It is the total length of the tag type
+including the type and length fields. With the current IP header length
+restriction of 40 bytes the value within this field is between 4 and 34.
+
+
+3.4.2.3 Alignment Octet
+
+This field is 1 octet in length and always has the value of 0. Its purpose
+is to align the category bitmap field on an even octet boundary. This will
+speed many implementations including router implementations.
+
+
+3.4.2.4 Sensitivity Level
+
+This field is 1 octet in length. Its value is from 0 to 255. The values
+are ordered with 0 being the minimum value and 255 representing the maximum
+value.
+
+
+3.4.2.5 Bit Map of Categories
+
+The length of this field is variable and ranges from 0 to 30 octets. This
+provides representation of categories 0 to 239. The ordering of the bits
+is left to right or MSB to LSB. For example category 0 is represented by
+the most significant bit of the first byte and category 15 is represented
+by the least significant bit of the second byte. Figure 4 graphically
+shows this ordering. Bit N is binary 1 if category N is part of the label
+for the datagram, and bit N is binary 0 if category N is not part of the
+label. Except for the optimized tag 1 format described in the next section,
+
+
+
+Internet Draft, Expires 15 Jan 93 [PAGE 4]
+
+
+
+CIPSO INTERNET DRAFT 16 July, 1992
+
+
+
+minimal encoding SHOULD be used resulting in no trailing zero octets in the
+category bitmap.
+
+ octet 0 octet 1 octet 2 octet 3 octet 4 octet 5
+ XXXXXXXX XXXXXXXX XXXXXXXX XXXXXXXX XXXXXXXX XXXXXXXX . . .
+bit 01234567 89111111 11112222 22222233 33333333 44444444
+number 012345 67890123 45678901 23456789 01234567
+
+ Figure 4. Ordering of Bits in Tag 1 Bit Map
+
+
+3.4.2.6 Optimized Tag 1 Format
+
+Routers work most efficiently when processing fixed length fields. To
+support these routers there is an optimized form of tag type 1. The format
+does not change. The only change is to the category bitmap which is set to
+a constant length of 10 octets. Trailing octets required to fill out the 10
+octets are zero filled. Ten octets, allowing for 80 categories, was chosen
+because it makes the total length of the CIPSO option 20 octets. If CIPSO
+is the only option then the option will be full word aligned and additional
+filler octets will not be required.
+
+
+3.4.3 Tag Type 2
+
+This is referred to as the "enumerated" tag type. It is used to describe
+large but sparsely populated sets of categories. Tag type 2 is in the MAC
+Sensitivity tag type class. The format of this tag type is as follows:
+
++----------+----------+----------+----------+-------------//-------------+
+| 00000010 | LLLLLLLL | 00000000 | LLLLLLLL | CCCCCCCCCCCCCCCCCCCCCCCCCC |
++----------+----------+----------+----------+-------------//-------------+
+
+ TAG TAG ALIGNMENT SENSITIVITY ENUMERATED
+ TYPE LENGTH OCTET LEVEL CATEGORIES
+
+ Figure 5. Tag Type 2 Format
+
+
+3.4.3.1 Tag Type
+
+This field is one octet in length and has a value of 2.
+
+
+3.4.3.2 Tag Length
+
+This field is 1 octet in length. It is the total length of the tag type
+including the type and length fields. With the current IP header length
+restriction of 40 bytes the value within this field is between 4 and 34.
+
+
+3.4.3.3 Alignment Octet
+
+This field is 1 octet in length and always has the value of 0. Its purpose
+is to align the category field on an even octet boundary. This will
+
+
+
+Internet Draft, Expires 15 Jan 93 [PAGE 5]
+
+
+
+CIPSO INTERNET DRAFT 16 July, 1992
+
+
+
+speed many implementations including router implementations.
+
+
+3.4.3.4 Sensitivity Level
+
+This field is 1 octet in length. Its value is from 0 to 255. The values
+are ordered with 0 being the minimum value and 255 representing the
+maximum value.
+
+
+3.4.3.5 Enumerated Categories
+
+In this tag, categories are represented by their actual value rather than
+by their position within a bit field. The length of each category is 2
+octets. Up to 15 categories may be represented by this tag. Valid values
+for categories are 0 to 65534. Category 65535 is not a valid category
+value. The categories MUST be listed in ascending order within the tag.
+
+
+3.4.4 Tag Type 5
+
+This is referred to as the "range" tag type. It is used to represent
+labels where all categories in a range, or set of ranges, are included
+in the sensitivity label. Tag type 5 is in the MAC Sensitivity tag type
+class. The format of this tag type is as follows:
+
++----------+----------+----------+----------+------------//-------------+
+| 00000101 | LLLLLLLL | 00000000 | LLLLLLLL | Top/Bottom | Top/Bottom |
++----------+----------+----------+----------+------------//-------------+
+
+ TAG TAG ALIGNMENT SENSITIVITY CATEGORY RANGES
+ TYPE LENGTH OCTET LEVEL
+
+ Figure 6. Tag Type 5 Format
+
+
+3.4.4.1 Tag Type
+
+This field is one octet in length and has a value of 5.
+
+
+3.4.4.2 Tag Length
+
+This field is 1 octet in length. It is the total length of the tag type
+including the type and length fields. With the current IP header length
+restriction of 40 bytes the value within this field is between 4 and 34.
+
+
+3.4.4.3 Alignment Octet
+
+This field is 1 octet in length and always has the value of 0. Its purpose
+is to align the category range field on an even octet boundary. This will
+speed many implementations including router implementations.
+
+
+
+
+
+Internet Draft, Expires 15 Jan 93 [PAGE 6]
+
+
+
+CIPSO INTERNET DRAFT 16 July, 1992
+
+
+
+3.4.4.4 Sensitivity Level
+
+This field is 1 octet in length. Its value is from 0 to 255. The values
+are ordered with 0 being the minimum value and 255 representing the maximum
+value.
+
+
+3.4.4.5 Category Ranges
+
+A category range is a 4 octet field comprised of the 2 octet index of the
+highest numbered category followed by the 2 octet index of the lowest
+numbered category. These range endpoints are inclusive within the range of
+categories. All categories within a range are included in the sensitivity
+label. This tag may contain a maximum of 7 category pairs. The bottom
+category endpoint for the last pair in the tag MAY be omitted and SHOULD be
+assumed to be 0. The ranges MUST be non-overlapping and be listed in
+descending order. Valid values for categories are 0 to 65534. Category
+65535 is not a valid category value.
+
+
+3.4.5 Minimum Requirements
+
+A CIPSO implementation MUST be capable of generating at least tag type 1 in
+the non-optimized form. In addition, a CIPSO implementation MUST be able
+to receive any valid tag type 1 even those using the optimized tag type 1
+format.
+
+
+4. Configuration Parameters
+
+The configuration parameters defined below are required for all CIPSO hosts,
+gateways, and routers that support multiple sensitivity labels. A CIPSO
+host is defined to be the origination or destination system for an IP
+datagram. A CIPSO gateway provides IP routing services between two or more
+IP networks and may be required to perform label translations between
+networks. A CIPSO gateway may be an enhanced CIPSO host or it may just
+provide gateway services with no end system CIPSO capabilities. A CIPSO
+router is a dedicated IP router that routes IP datagrams between two or more
+IP networks.
+
+An implementation of CIPSO on a host MUST have the capability to reject a
+datagram for reasons that the information contained can not be adequately
+protected by the receiving host or if acceptance may result in violation of
+the host or network security policy. In addition, a CIPSO gateway or router
+MUST be able to reject datagrams going to networks that can not provide
+adequate protection or may violate the network's security policy. To
+provide this capability the following minimal set of configuration
+parameters are required for CIPSO implementations:
+
+HOST_LABEL_MAX - This parameter contains the maximum sensitivity label that
+a CIPSO host is authorized to handle. All datagrams that have a label
+greater than this maximum MUST be rejected by the CIPSO host. This
+parameter does not apply to CIPSO gateways or routers. This parameter need
+not be defined explicitly as it can be implicitly derived from the
+PORT_LABEL_MAX parameters for the associated interfaces.
+
+
+
+Internet Draft, Expires 15 Jan 93 [PAGE 7]
+
+
+
+CIPSO INTERNET DRAFT 16 July, 1992
+
+
+
+
+HOST_LABEL_MIN - This parameter contains the minimum sensitivity label that
+a CIPSO host is authorized to handle. All datagrams that have a label less
+than this minimum MUST be rejected by the CIPSO host. This parameter does
+not apply to CIPSO gateways or routers. This parameter need not be defined
+explicitly as it can be implicitly derived from the PORT_LABEL_MIN
+parameters for the associated interfaces.
+
+PORT_LABEL_MAX - This parameter contains the maximum sensitivity label for
+all datagrams that may exit a particular network interface port. All
+outgoing datagrams that have a label greater than this maximum MUST be
+rejected by the CIPSO system. The label within this parameter MUST be
+less than or equal to the label within the HOST_LABEL_MAX parameter. This
+parameter does not apply to CIPSO hosts that support only one network port.
+
+PORT_LABEL_MIN - This parameter contains the minimum sensitivity label for
+all datagrams that may exit a particular network interface port. All
+outgoing datagrams that have a label less than this minimum MUST be
+rejected by the CIPSO system. The label within this parameter MUST be
+greater than or equal to the label within the HOST_LABEL_MIN parameter.
+This parameter does not apply to CIPSO hosts that support only one network
+port.
+
+PORT_DOI - This parameter is used to assign a DOI identifier value to a
+particular network interface port. All CIPSO labels within datagrams
+going out this port MUST use the specified DOI identifier. All CIPSO
+hosts and gateways MUST support either this parameter, the NET_DOI
+parameter, or the HOST_DOI parameter.
+
+NET_DOI - This parameter is used to assign a DOI identifier value to a
+particular IP network address. All CIPSO labels within datagrams destined
+for the particular IP network MUST use the specified DOI identifier. All
+CIPSO hosts and gateways MUST support either this parameter, the PORT_DOI
+parameter, or the HOST_DOI parameter.
+
+HOST_DOI - This parameter is used to assign a DOI identifier value to a
+particular IP host address. All CIPSO labels within datagrams destined for
+the particular IP host will use the specified DOI identifier. All CIPSO
+hosts and gateways MUST support either this parameter, the PORT_DOI
+parameter, or the NET_DOI parameter.
+
+This list represents the minimal set of configuration parameters required
+to be compliant. Implementors are encouraged to add to this list to
+provide enhanced functionality and control. For example, many security
+policies may require both incoming and outgoing datagrams be checked against
+the port and host label ranges.
+
+
+4.1 Port Range Parameters
+
+The labels represented by the PORT_LABEL_MAX and PORT_LABEL_MIN parameters
+MAY be in CIPSO or local format. Some CIPSO systems, such as routers, may
+want to have the range parameters expressed in CIPSO format so that incoming
+labels do not have to be converted to a local format before being compared
+against the range. If multiple DOIs are supported by one of these CIPSO
+
+
+
+Internet Draft, Expires 15 Jan 93 [PAGE 8]
+
+
+
+CIPSO INTERNET DRAFT 16 July, 1992
+
+
+
+systems then multiple port range parameters would be needed, one set for
+each DOI supported on a particular port.
+
+The port range will usually represent the total set of labels that may
+exist on the logical network accessed through the corresponding network
+interface. It may, however, represent a subset of these labels that are
+allowed to enter the CIPSO system.
+
+
+4.2 Single Label CIPSO Hosts
+
+CIPSO implementations that support only one label are not required to
+support the parameters described above. These limited implementations are
+only required to support a NET_LABEL parameter. This parameter contains
+the CIPSO label that may be inserted in datagrams that exit the host. In
+addition, the host MUST reject any incoming datagram that has a label which
+is not equivalent to the NET_LABEL parameter.
+
+
+5. Handling Procedures
+
+This section describes the processing requirements for incoming and
+outgoing IP datagrams. Just providing the correct CIPSO label format
+is not enough. Assumptions will be made by one system on how a
+receiving system will handle the CIPSO label. Wrong assumptions may
+lead to non-interoperability or even a security incident. The
+requirements described below represent the minimal set needed for
+interoperability and that provide users some level of confidence.
+Many other requirements could be added to increase user confidence,
+however at the risk of restricting creativity and limiting vendor
+participation.
+
+
+5.1 Input Procedures
+
+All datagrams received through a network port MUST have a security label
+associated with them, either contained in the datagram or assigned to the
+receiving port. Without this label the host, gateway, or router will not
+have the information it needs to make security decisions. This security
+label will be obtained from the CIPSO if the option is present in the
+datagram. See section 4.1.2 for handling procedures for unlabeled
+datagrams. This label will be compared against the PORT (if appropriate)
+and HOST configuration parameters defined in section 3.
+
+If any field within the CIPSO option, such as the DOI identifier, is not
+recognized the IP datagram is discarded and an ICMP "parameter problem"
+(type 12) is generated and returned. The ICMP code field is set to "bad
+parameter" (code 0) and the pointer is set to the start of the CIPSO field
+that is unrecognized.
+
+If the contents of the CIPSO are valid but the security label is
+outside of the configured host or port label range, the datagram is
+discarded and an ICMP "destination unreachable" (type 3) is generated
+and returned. The code field of the ICMP is set to "communication with
+destination network administratively prohibited" (code 9) or to
+
+
+
+Internet Draft, Expires 15 Jan 93 [PAGE 9]
+
+
+
+CIPSO INTERNET DRAFT 16 July, 1992
+
+
+
+"communication with destination host administratively prohibited"
+(code 10). The value of the code field used is dependent upon whether
+the originator of the ICMP message is acting as a CIPSO host or a CIPSO
+gateway. The recipient of the ICMP message MUST be able to handle either
+value. The same procedure is performed if a CIPSO can not be added to an
+IP packet because it is too large to fit in the IP options area.
+
+If the error is triggered by receipt of an ICMP message, the message
+is discarded and no response is permitted (consistent with general ICMP
+processing rules).
+
+
+5.1.1 Unrecognized tag types
+
+The default condition for any CIPSO implementation is that an
+unrecognized tag type MUST be treated as a "parameter problem" and
+handled as described in section 4.1. A CIPSO implementation MAY allow
+the system administrator to identify tag types that may safely be
+ignored. This capability is an allowable enhancement, not a
+requirement.
+
+
+5.1.2 Unlabeled Packets
+
+A network port may be configured to not require a CIPSO label for all
+incoming datagrams. For this configuration a CIPSO label must be
+assigned to that network port and associated with all unlabeled IP
+datagrams. This capability might be used for single level networks or
+networks that have CIPSO and non-CIPSO hosts and the non-CIPSO hosts
+all operate at the same label.
+
+If a CIPSO option is required and none is found, the datagram is
+discarded and an ICMP "parameter problem" (type 12) is generated and
+returned to the originator of the datagram. The code field of the ICMP
+is set to "option missing" (code 1) and the ICMP pointer is set to 134
+(the value of the option type for the missing CIPSO option).
+
+
+5.2 Output Procedures
+
+A CIPSO option MUST appear only once in a datagram. Only one tag type
+from the MAC Sensitivity class MAY be included in a CIPSO option. Given
+the current set of defined tag types, this means that CIPSO labels at
+first will contain only one tag.
+
+All datagrams leaving a CIPSO system MUST meet the following condition:
+
+ PORT_LABEL_MIN <= CIPSO label <= PORT_LABEL_MAX
+
+If this condition is not satisfied the datagram MUST be discarded.
+If the CIPSO system only supports one port, the HOST_LABEL_MIN and the
+HOST_LABEL_MAX parameters MAY be substituted for the PORT parameters in
+the above condition.
+
+The DOI identifier to be used for all outgoing datagrams is configured by
+
+
+
+Internet Draft, Expires 15 Jan 93 [PAGE 10]
+
+
+
+CIPSO INTERNET DRAFT 16 July, 1992
+
+
+
+the administrator. If port level DOI identifier assignment is used, then
+the PORT_DOI configuration parameter MUST contain the DOI identifier to
+use. If network level DOI assignment is used, then the NET_DOI parameter
+MUST contain the DOI identifier to use. And if host level DOI assignment
+is employed, then the HOST_DOI parameter MUST contain the DOI identifier
+to use. A CIPSO implementation need only support one level of DOI
+assignment.
+
+
+5.3 DOI Processing Requirements
+
+A CIPSO implementation MUST support at least one DOI and SHOULD support
+multiple DOIs. System and network administrators are cautioned to
+ensure that at least one DOI is common within an IP network to allow for
+broadcasting of IP datagrams.
+
+CIPSO gateways MUST be capable of translating a CIPSO option from one
+DOI to another when forwarding datagrams between networks. For
+efficiency purposes this capability is only a desired feature for CIPSO
+routers.
+
+
+5.4 Label of ICMP Messages
+
+The CIPSO label to be used on all outgoing ICMP messages MUST be equivalent
+to the label of the datagram that caused the ICMP message. If the ICMP was
+generated due to a problem associated with the original CIPSO label then the
+following responses are allowed:
+
+ a. Use the CIPSO label of the original IP datagram
+ b. Drop the original datagram with no return message generated
+
+In most cases these options will have the same effect. If you can not
+interpret the label or if it is outside the label range of your host or
+interface then an ICMP message with the same label will probably not be
+able to exit the system.
+
+
+6. Assignment of DOI Identifier Numbers =
+
+Requests for assignment of a DOI identifier number should be addressed to
+the Internet Assigned Numbers Authority (IANA).
+
+
+7. Acknowledgements
+
+Much of the material in this RFC is based on (and copied from) work
+done by Gary Winiger of Sun Microsystems and published as Commercial
+IP Security Option at the INTEROP 89, Commercial IPSO Workshop.
+
+
+8. Author's Address
+
+To submit mail for distribution to members of the IETF CIPSO Working
+Group, send mail to: cipso@wdl1.wdl.loral.com.
+
+
+
+Internet Draft, Expires 15 Jan 93 [PAGE 11]
+
+
+
+CIPSO INTERNET DRAFT 16 July, 1992
+
+
+
+
+To be added to or deleted from this distribution, send mail to:
+cipso-request@wdl1.wdl.loral.com.
+
+
+9. References
+
+RFC 1038, "Draft Revised IP Security Option", M. St. Johns, IETF, January
+1988.
+
+RFC 1108, "U.S. Department of Defense Security Options
+for the Internet Protocol", Stephen Kent, IAB, 1 March, 1991.
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+Internet Draft, Expires 15 Jan 93 [PAGE 12]
+
+
+
Index: linux-2.6.17.i686-quilt/Documentation/netlabel/introduction.txt
===================================================================
--- /dev/null
+++ linux-2.6.17.i686-quilt/Documentation/netlabel/introduction.txt
@@ -0,0 +1,52 @@
+NetLabel Introduction
+==============================================================================
+Paul Moore, paul.moore@hp.com
+
+June 16, 2006
+
+ * Overview
+
+NetLabel is a mechanism which can be used by kernel security modules to attach
+security attributes to outgoing network packets generated from user space
+applications and read security attributes from incoming network packets. It
+is composed of four main components, the protocol engines, the communication
+layer, the kernel security module API, and the network stack operations.
+
+ * Protocol Engines
+
+The protocol engines are responsible for both applying and retrieving the
+network packet's security attributes. If any translation between the network
+security attributes and those on the host are required then the protocol
+engine will handle those tasks as well. Other kernel subsystems should
+refrain from calling the protocol engines directly, instead they should use
+the NetLabel kernel security module API described below.
+
+Detailed information about each NetLabel protocol engine can be found in this
+directory, consult '00-INDEX' for filenames.
+
+ * Communication Layer
+
+The communication layer exists to allow NetLabel configuration and monitoring
+from user space. The NetLabel communication layer uses a message based
+protocol built on top of the NETLINK transport mechanism. The exact formatting
+of these NetLabel messages can be found in the the 'net/netlabel/' directory as
+comments in the header files.
+
+ * Security Module API
+
+The purpose of the NetLabel security module API is to provide a protocol
+independent interface to the underlying NetLabel protocol engines. In addition
+to protocol independence, the security module API is designed to be completely
+LSM independent which should allow multiple LSMs to leverage the same code
+base.
+
+Detailed information about the NetLabel security module API can be found in the
+'include/net/netlabel.h' header file as well as the 'lsm_interface.txt' file
+found in this directory.
+
+ * Network Stack Operations
+
+The reason for the network stack operations are to ensure the correct handling
+of the security attributes for kernel generated packets. In general only the
+core network stack should call these functions.
+
Index: linux-2.6.17.i686-quilt/Documentation/netlabel/lsm_interface.txt
===================================================================
--- /dev/null
+++ linux-2.6.17.i686-quilt/Documentation/netlabel/lsm_interface.txt
@@ -0,0 +1,47 @@
+NetLabel Linux Security Module Interface
+==============================================================================
+Paul Moore, paul.moore@hp.com
+
+May 17, 2006
+
+ * Overview
+
+NetLabel is a mechanism which can set and retrieve security attributes from
+network packets. It is intended to be used by LSM developers who want to make
+use of a common code base for several different packet labeling protocols.
+The NetLabel security module API is defined in 'include/net/netlabel.h' but a
+brief overview is given below.
+
+ * NetLabel Security Attributes
+
+Since NetLabel supports multiple different packet labeling protocols and LSMs
+it uses the concept of security attributes to refer to the packet's security
+labels. The NetLabel security attributes are defined by the
+'netlbl_lsm_secattr' structure in the NetLabel header file. Internally the
+NetLabel subsystem converts the security attributes to and from the correct
+low-level packet label depending on the NetLabel build time and run time
+configuration. It is up to the LSM developer to translate the NetLabel
+security attributes into whatever security identifiers are in use for their
+particular LSM.
+
+ * NetLabel LSM Protocol Operations
+
+These are the functions which allow the LSM developer to manipulate the labels
+on outgoing packets as well as read the labels on incoming packets. Functions
+exist to operate both on sockets as well as the sk_buffs directly. These high
+level functions are translated into low level protocol operations based on how
+the administrator has configured the NetLabel subsystem.
+
+ * NetLabel Label Mapping Cache Operations
+
+Depending on the exact configuration, translation between the network packet
+label and the internal LSM security identifier can be time consuming. The
+NetLabel label mapping cache is a caching mechanism which can be used to
+sidestep much of this overhead once a mapping has been established. Once the
+LSM has received a packet, used NetLabel to decode it's security attributes,
+and translated the security attributes into a LSM internal identifier the LSM
+can use the NetLabel caching functions to associate the LSM internal
+identifier with the network packet's label. This means that in the future
+when a incoming packet matches a cached value not only are the internal
+NetLabel translation mechanisms bypassed but the LSM translation mechanisms are
+bypassed as well which should result in a significant reduction in overhead.
--
paul moore
linux security @ hp
^ permalink raw reply
* [RFC 0/8] Minor update to allow a "sane review"
From: paul.moore @ 2006-06-22 22:49 UTC (permalink / raw)
To: netdev; +Cc: davem
This is a minor update to the patch set I set yesterday (June 21st), the
changes include minor Makefile fixes and a patch/file reordering as suggested
by David Miller. In addition, there are some minor changes to fix problems
brought about by different build combinations.
I have not yet made the shift to the generic NETLINK mechanism, but
considering the review problems due to the poor patch/file ordering I thought
it best to re-issue the patch set with that problem fixed so people would be
able to review the patch further.
--
paul moore
linux security @ hp
^ permalink raw reply
* RE: [patch 2.6.17] s2io driver irq fix
From: Ravinandan Arakali @ 2006-06-22 21:50 UTC (permalink / raw)
To: 'Andrew Morton', 'Ananda Raju'
Cc: netdev, linux-kernel, linux-fsdevel, dgc, balbir, viro, neilb,
jblunck, tglx, ananda.raju, leonid.grossman, alicia.pena
In-Reply-To: <20060621211534.b740d0f8.akpm@osdl.org>
Andrew,
My understanding is that MSI-X vectors are not usually shared. We don't want
to spend cycles checking if the interrupt was indeed from our card or
another device on same IRQ.
In fact, current driver shares IRQ for the MSI case which I think is a bug.
That should also be non-shared. Our MSI handler just runs thru' the Tx/Rx
completions and returns IRQ_HANDLED. In case of IRQ sharing, we could be
falsely claiming the interrupt as our own.
Ravi
-----Original Message-----
From: netdev-owner@vger.kernel.org
[mailto:netdev-owner@vger.kernel.org]On Behalf Of Andrew Morton
Sent: Wednesday, June 21, 2006 9:16 PM
To: Ananda Raju
Cc: netdev@vger.kernel.org; linux-kernel@vger.kernel.org;
linux-fsdevel@vger.kernel.org; dgc@sgi.com; balbir@in.ibm.com;
viro@zeniv.linux.org.uk; neilb@suse.de; jblunck@suse.de;
tglx@linutronix.de; ananda.raju@neterion.com;
leonid.grossman@neterion.com; ravinandan.arakali@neterion.com;
alicia.pena@neterion.com
Subject: Re: [patch 2.6.17] s2io driver irq fix
On Wed, 21 Jun 2006 15:50:49 -0400 (EDT)
Ananda Raju <Ananda.Raju@neterion.com> wrote:
> + if (sp->intr_type == MSI_X) {
> + int i;
>
> - free_irq(vector, arg);
> + for (i=1; (sp->s2io_entries[i].in_use == MSIX_FLG); i++) {
> + if (sp->s2io_entries[i].type == MSIX_FIFO_TYPE) {
> + sprintf(sp->desc[i], "%s:MSI-X-%d-TX",
> + dev->name, i);
> + err = request_irq(sp->entries[i].vector,
> + s2io_msix_fifo_handle, 0, sp->desc[i],
> + sp->s2io_entries[i].arg);
Is it usual to prohibit IRQ sharing with msix?
-
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply
* Re: [NET]: Prevent multiple qdisc runs
From: Herbert Xu @ 2006-06-22 22:43 UTC (permalink / raw)
To: jamal; +Cc: netdev, David S. Miller, Robert Olsson
In-Reply-To: <1151004682.5392.97.camel@jzny2>
On Thu, Jun 22, 2006 at 03:31:22PM -0400, jamal wrote:
>
> Your gut feeling is for #1 and my worry is for #2 ;->
> I actually think your change is obviously valuable for scenarios where
> the bus is slower and therefore transmits take longer - my feeling is it
> may not be beneficial for fast buses like PCI-E or high speed PCI/X
> where the possibility of getting access tx collision is lower.
Sure. However, I still don't see the point of transmitting in parallel
even there. The reason is that there is no work being done here by the
CPU between dequeueing the packet and obtaining the TX lock. As such
the cost of doing it in parallel is going to be dominated by the cache
bouncing.
Obviously it is a little different for lockless drivers where we do
dev_queue_xmit_nit (and now GSO) without any locks. However, you don't
want parallelism there because it breaks packet ordering.
> The other reason I mentioned earlier as justification to leave the
> granularity at the level where it was is for good qos clocking. i.e
> to allow incoming packets to be used to clock the tx path - otherwise
> you will be dependent on HZ for your egress rate accuracy. I am not sure
> if this later point made sense - I could elaborate.
I don't understand where HZ comes in. If you find that qdisc_run is
already running, then the packet you've just queued will most likely
be processed by that qdisc_run immediately unless the device is full.
Cheers,
--
Visit Openswan at http://www.openswan.org/
Email: Herbert Xu ~{PmV>HI~} <herbert@gondor.apana.org.au>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt
^ permalink raw reply
* RE: [PATCH 0/2][RFC] Network Event Notifier Mechanism
From: Caitlin Bestler @ 2006-06-22 22:39 UTC (permalink / raw)
To: hadi, Steve Wise; +Cc: netdev, David Miller
netdev-owner@vger.kernel.org wrote:
> On Thu, 2006-22-06 at 15:58 -0500, Steve Wise wrote:
>> On Thu, 2006-06-22 at 16:36 -0400, jamal wrote:
>
>> I created a new notifier block in my patch for these network events.
>> I guess I thought I was using the existing infrastructure to provide
>> this notification service. (I thought my patch was lovely :) But I
>> didn't integrate with netlink for user space notification. Mainly cuz
>> I didn't think these events should be propagated up to users unless
>> there was a need.
>
> I think they will be useful in user space. Typically you only
> propagate them if there's a user space program subscribed to
> listening (there are hooks which will tell you if there's
> anyone listening).
> The netdevice events tend to be a lot more usable in a few
> other blocks because they are lower in the hierachy (i.e
> routing depends on ip addresses which depend on netdevices)
> within the kernel unlike in this case where you are the only
> consumer; so it does sound logical to me to do it in user
> space; however, not totally unreasonable to do it in the kernel.
>
These services are relevant to any RDMA connection. The user-space
consumer of RDMA services is no more interested in tracking the
routing of the remote IP address than the consumer of socket
services is.
>>
>>
>> Another issue I see with netlink is that the event notifications
>> aren't reliable. Especially the CONFIG_ARPD stuff because it allocs
>> an sk_buff with ATOMIC. A lost neighbour macaddr change is perhaps
>> fatal for an RDMA connection...
>>
>
> This would happen in the cases where you are short on memory;
> i would suspect you will need to allocate memory in your
> driver as well to update something in the hardware as well -
> so same problem.
> You can however work around issues like these in netlink.
>
A direct notification call to the driver makes the driver responsible
for providing whatever buffering it requires to save the information.
And if there is insufficient memory available at least the driver
is aware of the failure.
Allowing a third component to fail to relay information means that
the driver can no longer be responsible for maintaining its own
consistency with kernel routing, ARP and neighbor tables.
Maintaining that consistency is a matter of correct network
behaviour, not doing status reports. obviously we cannot have
hardware looking at and interpreting these tables directly.
So a *reliable* subscription would seem to be the only option.
If the only subscribers who require reliable notifications are
kernel drivers, does it really mamke sense to make those changes
in code that also supports user space?
>
> I am still unclear:
> You have destination IP address, the dstMAC of the nexthop to
> get the packet to this IP address and i suspect some srcMAC
> address you will use sending out as well as the pathMTU to
> get there correct?
> Because of the IP address it sounds to me like you are
> populating an L3 table How is this info used in hardware? Can
> you explain how an arriving packet would be used by the RDMA
> in conjunction with this info once it is in the hardware?
>
Some packets are associated with established RDMA (or iSCSI)
connections, and are processed on the RDMA (or iSCSI) device.
These packets will also pass through other packets to the
host stack for processing (non-matched Ethernet frames for
IP networks, and IPoIB tunneled frames for IB networks).
The device provides L5 services (RDMA and/or iSCSI) in addition
to L2 services (as an Ethernet device). The rest of the network
rightfully demands that the left hand knows what the right hand
is doing. So information that is provided to a host, ARP/ICMP,
should affect the behaviour of *all* connections from that host.
Do you agree that having the device subsribe to the kernel
maintained tables is a better solution than having it attempt
to guess the correct values in parallel?
^ permalink raw reply
* New Qlogic qla3xxx NIC Driver v2.02.00k34 for upstream inclusion
From: Ron Mercer @ 2006-06-22 22:37 UTC (permalink / raw)
To: jeff, Andrew Morton; +Cc: netdev, linux-driver, Francois Romieu
Jeff/Andrew,
Please find the Qlogic qla3xxx Ethernet driver posted at the URL below.
This is a complementary network driver for our ISP4XXX parts.
There is a concurrent effort underway to get the iSCSI driver (qla4xxx)
integrated upstream as well.
I have added the missing license that Francois pointed out, plus
Andrew's patch and suggestions from yesterday.
This submission is contained in a patch file that does the following:
Adds:
drivers/net/qla3xxx.c
drivers/net/qla3xxx.h
Documentation/networking/LICENSE.qla3xxx
Modifies:
MAINTAINERS
drivers/net/Makefile
drivers/net/Kconfig
Patch file qla3xxxpatch1-v2.02.00-k34.txt is at the following link:
ftp://ftp.qlogic.com/outgoing/linux/network/upstream/2.02.00k34/qla3xxxp
atch1-v2.02.00-k34.txt
Signed-off-by: Ron Mercer <ron.mercer@qlogic.com>
Some notes on the driver/hardware:
- Built and tested using kernel 2.6.17-rc4.
- The chip supports two ethernet and two iSCSI functions.
- The functions ql_sem_lock, ql_sem_spinlock, ql_sem_unlock, and
ql_wait_for_drvr_lock are used to protect resources that are shared
across the network and iSCSI functions. This protection is mostly
during chip initialization and resets, but also include link management.
- The PHY/MII are not exported through ethtool due to the fact that the
iSCSI function will control the common link at least 50% of the time.
Regards,
Ron Mercer
Qlogic Corporation
^ permalink raw reply
* [PATCH] fix net-core kernel-doc
From: Randy.Dunlap @ 2006-06-22 22:14 UTC (permalink / raw)
To: netdev; +Cc: davem
From: Randy Dunlap <rdunlap@xenotime.net>
Warning(/var/linsrc/linux-2617-g4//include/linux/skbuff.h:304): No description found for parameter 'dma_cookie'
Warning(/var/linsrc/linux-2617-g4//include/net/sock.h:1274): No description found for parameter 'copied_early'
Warning(/var/linsrc/linux-2617-g4//net/core/dev.c:3309): No description found for parameter 'chan'
Warning(/var/linsrc/linux-2617-g4//net/core/dev.c:3309): No description found for parameter 'event'
Signed-off-by: Randy Dunlap <rdunlap@xenotime.net>
---
include/linux/skbuff.h | 2 ++
include/net/sock.h | 1 +
net/core/dev.c | 4 ++--
3 files changed, 5 insertions(+), 2 deletions(-)
--- linux-2617-g4.orig/include/linux/skbuff.h
+++ linux-2617-g4/include/linux/skbuff.h
@@ -209,6 +209,8 @@ enum {
* @nf_bridge: Saved data about a bridged frame - see br_netfilter.c
* @tc_index: Traffic control index
* @tc_verd: traffic control verdict
+ * @dma_cookie: a cookie to one of several possible DMA operations
+ * done by skb DMA functions
* @secmark: security marking
*/
--- linux-2617-g4.orig/include/net/sock.h
+++ linux-2617-g4/include/net/sock.h
@@ -1265,6 +1265,7 @@ sock_recv_timestamp(struct msghdr *msg,
* sk_eat_skb - Release a skb if it is no longer needed
* @sk: socket to eat this skb from
* @skb: socket buffer to eat
+ * @copied_early: flag indicating whether DMA operations copied this data early
*
* This routine must be called with interrupts disabled or with the socket
* locked so that the sk_buff queue operation is ok.
--- linux-2617-g4.orig/net/core/dev.c
+++ linux-2617-g4/net/core/dev.c
@@ -3301,8 +3301,8 @@ static void net_dma_rebalance(void)
/**
* netdev_dma_event - event callback for the net_dma_client
* @client: should always be net_dma_client
- * @chan:
- * @event:
+ * @chan: DMA channel for the event
+ * @event: event type
*/
static void netdev_dma_event(struct dma_client *client, struct dma_chan *chan,
enum dma_event event)
---
^ permalink raw reply
* RE: [PATCH 0/2][RFC] Network Event Notifier Mechanism
From: jamal @ 2006-06-22 22:21 UTC (permalink / raw)
To: Caitlin Bestler; +Cc: Steve Wise, David Miller, netdev
In-Reply-To: <54AD0F12E08D1541B826BE97C98F99F15F55E2@NT-SJCA-0751.brcm.ad.broadcom.com>
On Thu, 2006-22-06 at 15:11 -0700, Caitlin Bestler wrote:
> netdev-owner@vger.kernel.org wrote:
>
> These subscriptions are an attempt to cede full control
> of these issues back to one place, the kernel, and to
> guarantee that an offload device can never think that
> the route to to X is Y when the kernel says it is Z.
> Or that it has a different PMTU, etc.
>
Ok, so it is the routing information then that you are syncing,
correct?
> I don't have any strong opinion on the best mechanism
> for implementing these subscriptions, but having correct
> consistent networking behaviour depend on a user-mode
> relay strikes me as odd.
>
And why does it sound odd?
You will need to think about one issue:
Linux caches routing info - it is not just as simple as keeping track of
the FIB and somehow correlating that to the ARP entries.
cheers,
jamal
^ permalink raw reply
* Re: [PATCH 0/2][RFC] Network Event Notifier Mechanism
From: jamal @ 2006-06-22 22:14 UTC (permalink / raw)
To: Steve Wise; +Cc: netdev, David Miller
In-Reply-To: <1151009932.3040.80.camel@stevo-desktop>
On Thu, 2006-22-06 at 15:58 -0500, Steve Wise wrote:
> On Thu, 2006-06-22 at 16:36 -0400, jamal wrote:
> I created a new notifier block in my patch for these network events. I
> guess I thought I was using the existing infrastructure to provide this
> notification service. (I thought my patch was lovely :) But I didn't
> integrate with netlink for user space notification. Mainly cuz I didn't
> think these events should be propagated up to users unless there was a
> need.
I think they will be useful in user space. Typically you only propagate
them if there's a user space program subscribed to listening (there are
hooks which will tell you if there's anyone listening).
The netdevice events tend to be a lot more usable in a few other blocks
because they are lower in the hierachy (i.e routing depends on ip
addresses which depend on netdevices) within the kernel unlike in this
case where you are the only consumer; so it does sound logical to me
to do it in user space; however, not totally unreasonable to do it in
the kernel.
> Just to clarify, you're suggesting I add any needed netlink hooks for
> rt_redirect and the others that don't exist today, and use a NETLINK
> socket in user space to discover these events. Yes?
>
indeed.
> > Your mileage may vary. If you do it in user space you dont have to wait
> > for the next kernel release in case of a bug.
>
> As long as all the events are passed up correctly :-)
>
They have been for years ;->
> > Additionally, it allows
> > for more feature richness that would tend to bloat the kernel/infiniband
> > otherwise.
>
>
> Another issue I see with netlink is that the event notifications aren't
> reliable. Especially the CONFIG_ARPD stuff because it allocs an sk_buff
> with ATOMIC. A lost neighbour macaddr change is perhaps fatal for an
> RDMA connection...
>
This would happen in the cases where you are short on memory; i would
suspect you will need to allocate memory in your driver as well to
update something in the hardware as well - so same problem.
You can however work around issues like these in netlink.
>
> > Out of curiosity - what does RDMA NIC have that would need these events?
> > a route table or L2 table etc? Can you elucidate a little?
> >
>
> Mainly the L2 table, next hop ip addr, and the path mtu. RDMA NICs
> implement the entire RDMA stack in HW. How they deal with L2 and L3
> changes vary to some degree, but what seems to be emerging is that they
> get this information from the native stack because ARP and ICMP, for
> example, are always passed up to the native stack.
>
I am still unclear:
You have destination IP address, the dstMAC of the nexthop to get the
packet to this IP address and i suspect some srcMAC address you will use
sending out as well as the pathMTU to get there correct?
Because of the IP address it sounds to me like you are populating an L3
table
How is this info used in hardware? Can you explain how an arriving
packet would be used by the RDMA in conjunction with this info once it
is in the hardware?
> These devices also act a standard Ethernet NIC btw...
>
Meaning there is no funky hardware processing?
cheers,
jamal
^ permalink raw reply
* RE: [PATCH 0/2][RFC] Network Event Notifier Mechanism
From: Caitlin Bestler @ 2006-06-22 22:11 UTC (permalink / raw)
To: hadi, Steve Wise; +Cc: David Miller, netdev
netdev-owner@vger.kernel.org wrote:
> On Thu, 2006-22-06 at 15:40 -0500, Steve Wise wrote:
>> On Thu, 2006-06-22 at 15:43 -0400, jamal wrote:
>>>
>>> No - what these 2 gents are saying was these events and
>>> infrastructure already exist.
>>
>> Notification of the exact events needed does not exist today.
>>
>
> Ok, so you cant event make use of anything that already exists?
> Or is a subset of what you need already there?
>
>> The key events, again, are:
>>
>> - the neighbour entry mac address has changed.
>>
>>
>> - the next hop ip address (ie the neighbour) for a given dst_entry
>> has changed.
>
>
> I dont see a difference for the above two from an L2 perspective.
> Are you keeping track of IP addresses?
> You didn't answer my question in the previous email as to
> what RDMA needs to keep track of in hardware.
>
The RDMA device is handling L4 or L5 connections that
have L3 Addresses (IP). Subscribing to the information
allows the device to keep its behaviour consistent
with the host stack.
The common alternative before proposing this integration
was to have the RDMA device sniff all incoming packets
and attempt to do parallel procesing on a large set
of lower layer protocols (ICMP, ARP, routing, ...)
Or by simply trusting that the IB network adminstrator
has faithfully replicated all IP-relevent instructions
in two forums (traditional IP nework administration
and IB network administration).
These subscriptions are an attempt to cede full control
of these issues back to one place, the kernel, and to
guarantee that an offload device can never think that
the route to to X is Y when the kernel says it is Z.
Or that it has a different PMTU, etc.
I don't have any strong opinion on the best mechanism
for implementing these subscriptions, but having correct
consistent networking behaviour depend on a user-mode
relay strikes me as odd.
^ permalink raw reply
* Re: New Qlogic qla3xxx NIC Driver v2.02.00k31 for upstream inclusion
From: Jeff Garzik @ 2006-06-22 21:35 UTC (permalink / raw)
To: Andrew Morton; +Cc: Ron Mercer, Linux-Driver, netdev
In-Reply-To: <20060622142217.0abfe030.akpm@osdl.org>
Andrew Morton wrote:
> "Ron Mercer" <ron.mercer@qlogic.com> wrote:
>> Thanks for your feedback. I've fixed everything you mentioned with the
>> exception of the code below. I need to swap every other byte pair of my
>> nvram mac address for big endian systems. Everything else I tried was
>> more messy. Did you have anything specific in mind?
>
> Nope, it was more a question for the other netdev developers. If nobody
> can point at a library function or something then fine, leave it as-is.
It seems sane to me. The kernel doesn't have a standard le48_to_cpu()
and cpu_to_le48()...
Jeff
^ permalink raw reply
* Re: [RFC 2/7] NetLabel: core network changes
From: David Miller @ 2006-06-22 21:32 UTC (permalink / raw)
To: sgrubb
Cc: jmorris, paul.moore, sds, redhat-lspp, linux-security-module,
selinux, netdev
In-Reply-To: <200606221105.00331.sgrubb@redhat.com>
From: Steve Grubb <sgrubb@redhat.com>
Date: Thu, 22 Jun 2006 11:05:00 -0400
> If moving this over to generic is a must, then I think SE Linux will
> have to clip into generic to control its packet flow.
James Morris has already started discussions on this.
It's great that there will be another reason why that work
needs to be done :)
^ permalink raw reply
* Re: [3/5] [NET]: Add software TSOv4
From: David Miller @ 2006-06-22 21:32 UTC (permalink / raw)
To: yoshfuji; +Cc: herbert, netdev
In-Reply-To: <20060623.000403.47759894.yoshfuji@linux-ipv6.org>
From: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
Date: Fri, 23 Jun 2006 00:04:03 +0900 (JST)
> In article <20060622081400.GC22671@gondor.apana.org.au> (at Thu, 22 Jun 2006 18:14:00 +1000), Herbert Xu <herbert@gondor.apana.org.au> says:
>
> > [NET]: Add software TSOv4
> >
> > This patch adds the GSO implementation for IPv4 TCP.
> >
> > Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
>
> I'd appreciate if you code up IPv6 TCP as well. :-)
To my understanding doing IPV6 TCP TSO is a non-trivial task, even in
software.
The header editing is a lot more complicated because things like
routing and other extension headers can sit between IPV6 and TCP
header.
It is probably why IPV6 TSO hardware does not exist yet :)
Do not take this to mean I think it should not be implemented, I think
it should.
^ permalink raw reply
* Re: New Qlogic qla3xxx NIC Driver v2.02.00k31 for upstream inclusion
From: Andrew Morton @ 2006-06-22 21:22 UTC (permalink / raw)
To: Ron Mercer; +Cc: jeff, Linux-Driver, netdev
In-Reply-To: <0BB3E5E7462EEA4295BC02D49691DC07176862@AVEXCH1.qlogic.org>
"Ron Mercer" <ron.mercer@qlogic.com> wrote:
>
> Thanks for your feedback. I've fixed everything you mentioned with the
> exception of the code below. I need to swap every other byte pair of my
> nvram mac address for big endian systems. Everything else I tried was
> more messy. Did you have anything specific in mind?
Nope, it was more a question for the other netdev developers. If nobody
can point at a library function or something then fine, leave it as-is.
^ permalink raw reply
* RE: New Qlogic qla3xxx NIC Driver v2.02.00k31 for upstream inclusion
From: Ron Mercer @ 2006-06-22 21:00 UTC (permalink / raw)
To: Andrew Morton; +Cc: jeff, Linux Driver, netdev
Andrew,
Thanks for your feedback. I've fixed everything you mentioned with the
exception of the code below. I need to swap every other byte pair of my
nvram mac address for big endian systems. Everything else I tried was
more messy. Did you have anything specific in mind?
Ron Mercer
>
> - Is there a better way of doing this?
>
> static void ql_swap_mac_addr(u8 * macAddress) { #ifdef __BIG_ENDIAN
> u8 temp;
> temp = macAddress[0];
> macAddress[0] = macAddress[1];
> macAddress[1] = temp;
> temp = macAddress[2];
> macAddress[2] = macAddress[3];
> macAddress[3] = temp;
> temp = macAddress[4];
> macAddress[4] = macAddress[5];
> macAddress[5] = temp;
> #endif
> }
>
^ permalink raw reply
* Re: [PATCH 0/2][RFC] Network Event Notifier Mechanism
From: Steve Wise @ 2006-06-22 20:58 UTC (permalink / raw)
To: hadi; +Cc: David Miller, netdev
In-Reply-To: <1151008597.5392.118.camel@jzny2>
On Thu, 2006-06-22 at 16:36 -0400, jamal wrote:
> On Thu, 2006-22-06 at 15:18 -0500, Steve Wise wrote:
> > On Thu, 2006-06-22 at 15:43 -0400, jamal wrote:
>
> > > As an example, search for NETDEV_CHANGEADDR,NETDEV_CHANGEMTU etc.
> > > Actually you are probably making this too complicated.
> >
> > NETDEV_CHANGEADDR uses a notifier block, and the network subsystem calls
> > call_netdevice_notifiers() when it sets an addr. And any kernel module
> > can register for these events. That's the model I used to create the
> > netevent_notifier mechanism in the patch I posted.
> >
>
> it also gets emmited as a netlink event.
>
right.
> > I could add the new events to this netdevice notifier, but these aren't
> > really net device events. Their network events.
> >
>
> Different blocks for sure - the point is the infrastructure which
> constitutes using notifiers exists. And it is joined at the hip with
> netlink.
>
I created a new notifier block in my patch for these network events. I
guess I thought I was using the existing infrastructure to provide this
notification service. (I thought my patch was lovely :) But I didn't
integrate with netlink for user space notification. Mainly cuz I didn't
think these events should be propagated up to users unless there was a
need.
> > I can indeed extend the rtnetlink stuff to add the events in question
> > (neighbour mac addr change, route redirect, etc). In fact, there is
> > similar functionality under the CONFIG_ARPD option to support a user
> > space arp daemon. Its not quite the same, and it doesn't cover redirect
> > and routing events, just neighbour events.
> >
>
> CONFIG_ARPD will give you all neighbor events you need.
> => rt_redirect doesnt exist neither do route cache
> creation/updates/deletions. FIB changes exist etc
>
Just to clarify, you're suggesting I add any needed netlink hooks for
rt_redirect and the others that don't exist today, and use a NETLINK
socket in user space to discover these events. Yes?
> > But in the case of the RDMA subsystem, the consumer of these events is
> > in the kernel. Why is it better to propagate events all the way up to
> > user space, then send the event back down into the Infiniband kernel
> > subsystem? That seems very inefficient.
>
> Your mileage may vary. If you do it in user space you dont have to wait
> for the next kernel release in case of a bug.
As long as all the events are passed up correctly :-)
> Additionally, it allows
> for more feature richness that would tend to bloat the kernel/infiniband
> otherwise.
Another issue I see with netlink is that the event notifications aren't
reliable. Especially the CONFIG_ARPD stuff because it allocs an sk_buff
with ATOMIC. A lost neighbour macaddr change is perhaps fatal for an
RDMA connection...
> Out of curiosity - what does RDMA NIC have that would need these events?
> a route table or L2 table etc? Can you elucidate a little?
>
Mainly the L2 table, next hop ip addr, and the path mtu. RDMA NICs
implement the entire RDMA stack in HW. How they deal with L2 and L3
changes vary to some degree, but what seems to be emerging is that they
get this information from the native stack because ARP and ICMP, for
example, are always passed up to the native stack.
These devices also act a standard Ethernet NIC btw...
Steve.
^ permalink raw reply
* Re: [PATCH 0/2][RFC] Network Event Notifier Mechanism
From: jamal @ 2006-06-22 20:56 UTC (permalink / raw)
To: Steve Wise; +Cc: David Miller, netdev
In-Reply-To: <1151008816.3040.60.camel@stevo-desktop>
On Thu, 2006-22-06 at 15:40 -0500, Steve Wise wrote:
> On Thu, 2006-06-22 at 15:43 -0400, jamal wrote:
> >
> > No - what these 2 gents are saying was these events and infrastructure
> > already exist.
>
> Notification of the exact events needed does not exist today.
>
Ok, so you cant event make use of anything that already exists?
Or is a subset of what you need already there?
> The key events, again, are:
>
> - the neighbour entry mac address has changed.
>
>
> - the next hop ip address (ie the neighbour) for a given dst_entry has
> changed.
I dont see a difference for the above two from an L2 perspective.
Are you keeping track of IP addresses?
You didn't answer my question in the previous email as to what RDMA
needs to keep track of in hardware.
>
> - the path mtu for a given dst_entry has changed.
>
Same with this.
cheers,
jamal
^ permalink raw reply
* Re: [RFC] sysfs + configfs on 802.11 wireless drivers
From: Greg KH @ 2006-06-22 20:41 UTC (permalink / raw)
To: Luis R. Rodriguez; +Cc: netdev, John W. Linville
In-Reply-To: <43e72e890606220512h5197d473s5f7da9e734814005@mail.gmail.com>
On Thu, Jun 22, 2006 at 08:12:24AM -0400, Luis R. Rodriguez wrote:
> I know we recently moved WE to netlink but I figured celebrating the
> happy marriage of Mr. sysfs and Mrs. configfs on 2.6.16 by giving them
> offspring would be nice and more appropriate.
What, right after deciding one thing, you want to switch it all over to
something completly different? How about giving netlink a chance and
wait to see how well (or not) that works out, so that you will have more
insight and history into any problems that you had so as to best address
them in the future.
thanks,
greg k-h
^ permalink raw reply
* Re: [PATCH 0/2][RFC] Network Event Notifier Mechanism
From: Steve Wise @ 2006-06-22 20:40 UTC (permalink / raw)
To: hadi; +Cc: netdev, David Miller
In-Reply-To: <1151005381.5392.108.camel@jzny2>
On Thu, 2006-06-22 at 15:43 -0400, jamal wrote:
>
> No - what these 2 gents are saying was these events and infrastructure
> already exist.
Notification of the exact events needed does not exist today.
The key events, again, are:
- the neighbour entry mac address has changed.
- the next hop ip address (ie the neighbour) for a given dst_entry has
changed.
- the path mtu for a given dst_entry has changed.
Steve.
^ permalink raw reply
* Re: [PATCH 0/2][RFC] Network Event Notifier Mechanism
From: jamal @ 2006-06-22 20:36 UTC (permalink / raw)
To: Steve Wise; +Cc: David Miller, netdev
In-Reply-To: <1151007524.3040.48.camel@stevo-desktop>
On Thu, 2006-22-06 at 15:18 -0500, Steve Wise wrote:
> On Thu, 2006-06-22 at 15:43 -0400, jamal wrote:
> > As an example, search for NETDEV_CHANGEADDR,NETDEV_CHANGEMTU etc.
> > Actually you are probably making this too complicated.
>
> NETDEV_CHANGEADDR uses a notifier block, and the network subsystem calls
> call_netdevice_notifiers() when it sets an addr. And any kernel module
> can register for these events. That's the model I used to create the
> netevent_notifier mechanism in the patch I posted.
>
it also gets emmited as a netlink event.
> I could add the new events to this netdevice notifier, but these aren't
> really net device events. Their network events.
>
Different blocks for sure - the point is the infrastructure which
constitutes using notifiers exists. And it is joined at the hip with
netlink.
> I can indeed extend the rtnetlink stuff to add the events in question
> (neighbour mac addr change, route redirect, etc). In fact, there is
> similar functionality under the CONFIG_ARPD option to support a user
> space arp daemon. Its not quite the same, and it doesn't cover redirect
> and routing events, just neighbour events.
>
CONFIG_ARPD will give you all neighbor events you need.
=> rt_redirect doesnt exist neither do route cache
creation/updates/deletions. FIB changes exist etc
> But in the case of the RDMA subsystem, the consumer of these events is
> in the kernel. Why is it better to propagate events all the way up to
> user space, then send the event back down into the Infiniband kernel
> subsystem? That seems very inefficient.
Your mileage may vary. If you do it in user space you dont have to wait
for the next kernel release in case of a bug. Additionally, it allows
for more feature richness that would tend to bloat the kernel/infiniband
otherwise.
Out of curiosity - what does RDMA NIC have that would need these events?
a route table or L2 table etc? Can you elucidate a little?
cheers,
jamal
^ permalink raw reply
* RE: [PATCH 7/32] [TIPC] Multicast link failure now resets all links to "nacking" node.
From: Stephens, Allan @ 2006-06-22 20:33 UTC (permalink / raw)
To: James Morris, Per Liden; +Cc: David Miller, netdev
Nice observation, James. As a relative newcomer to the official Linux
kernel development world, I'm impressed that non-TIPC folks are looking
at TIPC changes closely enough to see things like this!
Per, I'll leave it to you to decide if you want to address James's
concern. But be aware that the link_reset_all() routine is only called
to handle emergency situations when TIPC's multicast link has run into
serious problems and is trying to recover. Most systems will never
follow this path, so the cost of the unnecessary
local_bh_disable()/local_bh_enable() pairing shouldn't have any real
impact on the overall performance of TIPC.
Regards,
Al Stephens
> -----Original Message-----
> From: James Morris [mailto:jmorris@namei.org]
> Sent: Thursday, June 22, 2006 2:51 PM
> To: Per Liden
> Cc: David Miller; netdev@vger.kernel.org; Stephens, Allan
> Subject: Re: [PATCH 7/32] [TIPC] Multicast link failure now
> resets all links to "nacking" node.
>
> On Thu, 22 Jun 2006, Per Liden wrote:
>
> > +static void link_reset_all(unsigned long addr) {
> > + struct node *n_ptr;
> > + char addr_string[16];
> > + u32 i;
> > +
> > + read_lock_bh(&tipc_net_lock);
> > + n_ptr = tipc_node_find((u32)addr);
> > + if (!n_ptr) {
> > + read_unlock_bh(&tipc_net_lock);
> > + return; /* node no longer exists */
> > + }
> > +
> > + tipc_node_lock(n_ptr);
>
> You already have bh's disabled here, and tipc_node_lock()
> also disables them.
>
> Not sure if it's really worth worrying about but if so, you
> could perhaps implement tipc_node_lock_bh() and tipc_node_lock().
>
>
>
> - James
> --
> James Morris
> <jmorris@namei.org>
>
^ permalink raw reply
* [PATCH] WAN: ioremap() failure checks in drivers
From: Krzysztof Halasa @ 2006-06-22 20:29 UTC (permalink / raw)
To: Jeff Garzik; +Cc: netdev
Eric Sesterhenn found that pci200syn initialization lacks return
statement in ioremap() error path (coverity bug id #195). It looks
like more WAN drivers have problems with ioremap().
Signed-off-by: Krzysztof Halasa <khc@pm.waw.pl>
--- a/drivers/net/wan/c101.c
+++ b/drivers/net/wan/c101.c
@@ -326,21 +326,21 @@ static int __init c101_run(unsigned long
if (request_irq(irq, sca_intr, 0, devname, card)) {
printk(KERN_ERR "c101: could not allocate IRQ\n");
c101_destroy_card(card);
- return(-EBUSY);
+ return -EBUSY;
}
card->irq = irq;
if (!request_mem_region(winbase, C101_MAPPED_RAM_SIZE, devname)) {
printk(KERN_ERR "c101: could not request RAM window\n");
c101_destroy_card(card);
- return(-EBUSY);
+ return -EBUSY;
}
card->phy_winbase = winbase;
card->win0base = ioremap(winbase, C101_MAPPED_RAM_SIZE);
if (!card->win0base) {
printk(KERN_ERR "c101: could not map I/O address\n");
c101_destroy_card(card);
- return -EBUSY;
+ return -EFAULT;
}
card->tx_ring_buffers = TX_RING_BUFFERS;
--- a/drivers/net/wan/n2.c
+++ b/drivers/net/wan/n2.c
@@ -387,6 +387,11 @@ static int __init n2_run(unsigned long i
}
card->phy_winbase = winbase;
card->winbase = ioremap(winbase, USE_WINDOWSIZE);
+ if (!card->winbase) {
+ printk(KERN_ERR "n2: ioremap() failed\n");
+ n2_destroy_card(card);
+ return -EFAULT;
+ }
outb(0, io + N2_PCR);
outb(winbase >> 12, io + N2_BAR);
--- a/drivers/net/wan/pci200syn.c
+++ b/drivers/net/wan/pci200syn.c
@@ -358,6 +358,7 @@ #endif
card->rambase == NULL) {
printk(KERN_ERR "pci200syn: ioremap() failed\n");
pci200_pci_remove_one(pdev);
+ return -EFAULT;
}
/* Reset PLX */
--- a/drivers/net/wan/wanxl.c
+++ b/drivers/net/wan/wanxl.c
@@ -634,7 +634,13 @@ #endif
/* set up PLX mapping */
plx_phy = pci_resource_start(pdev, 0);
+
card->plx = ioremap_nocache(plx_phy, 0x70);
+ if (!card->plx) {
+ printk(KERN_ERR "wanxl: ioremap() failed\n");
+ wanxl_pci_remove_one(pdev);
+ return -EFAULT;
+ }
#if RESET_WHILE_LOADING
wanxl_reset(card);
@@ -700,6 +706,12 @@ #endif
}
mem = ioremap_nocache(mem_phy, PDM_OFFSET + sizeof(firmware));
+ if (!mem) {
+ printk(KERN_ERR "wanxl: ioremap() failed\n");
+ wanxl_pci_remove_one(pdev);
+ return -EFAULT;
+ }
+
for (i = 0; i < sizeof(firmware); i += 4)
writel(htonl(*(u32*)(firmware + i)), mem + PDM_OFFSET + i);
^ permalink raw reply
page: next (older) | prev (newer) | latest
- recent:[subjects (threaded)|topics (new)|topics (active)]
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox