NetLabel fixes and reworked Netlink interface
cover NetLabel fixes and reworked Netlink interface
This patchset, made against net-2.6.19, contains a collection of fixes for bugs found over the past few weeks as well as reworked Netlink interface which should bring NetLabel more inline with common Netlink methods. My apologies for not getting this out sooner, but the Netlink interface rewrite was a bit more involved than I originally thought. For those wishing to test/use these patches you will need a new copy of netlabel_tools; you can get a working version from the SVN respository at http://netlabel.sf.net.
I have tested these patches and verified that they work and solve all known issues. Please review and consider them for 2.6.19.
Thanks.
--
paul moore
linux security @ hp
PATCH 1/6 NetLabel: correct improper handling of non-NetLabel peer contexts
Fix a problem where NetLabel would always set the value of sk_security_struct->peer_sid in selinux_netlbl_sock_graft() to the context of the socket, causing problems when users would query the context of the connection. This patch fixes this so that the value in sk_security_struct->peer_sid is only set when the connection is NetLabel based, otherwise the value is untouched.
--- include/net/cipso_ipv4.h | 7 +++++ include/net/netlabel.h | 8 ++++++ net/ipv4/cipso_ipv4.c | 48 ++++++++++++++++++++++++++++------------- net/netlabel/netlabel_kapi.c | 23 +++++++++++++++++++ security/selinux/ss/services.c | 12 +++++++++- 5 files changed, 82 insertions(+), 16 deletions(-)
include/net/cipso_ipv4.h +7 -0
@@ -205,6 +205,7 @@ void cipso_v4_error(struct sk_buff *skb,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_sock_getattr (struct sock *sk,struct netlbl_lsm_secattr *secattr);int cipso_v4_socket_getattr (const struct socket *sock,struct netlbl_lsm_secattr *secattr); int cipso_v4_skbuff_getattr(const struct sk_buff *skb,
@@ -225,6 +226,12 @@ static inline int cipso_v4_socket_setattreturn -ENOSYS ; } +static inline int cipso_v4_sock_getattr (struct sock *sk, +struct netlbl_lsm_secattr *secattr) +{ +return -ENOSYS ; +} +static inline int cipso_v4_socket_getattr (const struct socket *sock,struct netlbl_lsm_secattr *secattr) {
include/net/netlabel.h +8 -0
@@ -238,6 +238,8 @@ static inline void netlbl_secattr_free(s#ifdef CONFIG_NETLABEL int netlbl_socket_setattr (const struct socket *sock,const struct netlbl_lsm_secattr *secattr); +int netlbl_sock_getattr (struct sock *sk, +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,
@@ -250,6 +252,12 @@ static inline int netlbl_socket_setattr(return -ENOSYS ; } +static inline int netlbl_sock_getattr (struct sock *sk, +struct netlbl_lsm_secattr *secattr) +{ +return -ENOSYS ; +} +static inline int netlbl_socket_getattr (const struct socket *sock,struct netlbl_lsm_secattr *secattr) {
net/ipv4/cipso_ipv4.c +33 -15
84 lines not on this page · raw message
net/netlabel/netlabel_kapi.c +23 -0
@@ -85,6 +85,29 @@ socket_setattr_return: }/** +* netlbl_sock_getattr - Determine the security attributes of a sock +* @sk: the sock +* @secattr: the security attributes +* +* Description: +* Examines the given sock to see any NetLabel style labeling has been +* applied to the sock, if so it parses the socket label and returns the +* security attributes in @secattr. Returns zero on success, negative values +* on failure. +* +*/ +int netlbl_sock_getattr (struct sock *sk,struct netlbl_lsm_secattr *secattr) +{ +int ret_val; + + ret_val =cipso_v4_sock_getattr (sk, secattr); +if (ret_val ==0 ) +return 0 ; + +return netlbl_unlabel_getattr (secattr); +} + +/** * netlbl_socket_getattr - Determine the security attributes of a socket * @sock: the socket * @secattr: the security attributes
security/selinux/ss/services.c +11 -1
@@ -2502,14 +2502,24 @@ void selinux_netlbl_sock_graft(struct so {struct inode_security_struct *isec =SOCK_INODE (sock)->i_security;struct sk_security_struct *sksec = sk->sk_security; +struct netlbl_lsm_secattr secattr; +u32 nlbl_peer_sid; sksec->sclass = isec->sclass;if (sk->sk_family !=PF_INET )return ; +netlbl_secattr_init (&secattr); +if (netlbl_sock_getattr (sk, &secattr) ==0 && +selinux_netlbl_secattr_to_sid (NULL , + &secattr, + sksec->sid, + &nlbl_peer_sid) ==0 ) + sksec->peer_sid = nlbl_peer_sid; +netlbl_secattr_destroy (&secattr,0 ); + sksec->nlbl_state =NLBL_REQUIRE ; - sksec->peer_sid = sksec->sid;/* Try to set the NetLabel on the socket to save time later, if we fail * here we will pick up the pieces in later calls to
--
paul moore
linux security @ hp
Re: [PATCH 1/6] NetLabel: correct improper handling of non-NetLabel peer contexts
Quoted 6 lines · On Thu, 21 Sep 2006, [email protected] wrote:
Fix a problem where NetLabel would always set the value of
sk_security_struct->peer_sidinselinux_netlbl_sock_graft()to the context of the socket, causing problems when users would query the context of the connection. This patch fixes this so that the value insk_security_struct->peer_sidis only set when the connection is NetLabel based, otherwise the value is untouched.
I'll let Thomas comment on the Netlink changes, as he's been working with you on them.
These changes otherwise seem ok. How much testing has this had with and without Netlabel enabled?
- James
Signature · 2 lines
James Morris <[email protected]>
Re: [PATCH 1/6] NetLabel: correct improper handling of non-NetLabel peer contexts
Quoted 10 lines · James Morris
Quoted 6 lines · On Thu, 21 Sep 2006, [email protected] wrote:
Fix a problem where NetLabel would always set the value of
sk_security_struct->peer_sidinselinux_netlbl_sock_graft()to the context of the socket, causing problems when users would query the context of the connection. This patch fixes this so that the value insk_security_struct->peer_sidis only set when the connection is NetLabel based, otherwise the value is untouched.
Quoted 2 lines
I'll let Thomas comment on the Netlink changes, as he's been working with you on them.
Hey, third times the charm right?
;)
Quoted 2 lines
These changes otherwise seem ok. How much testing has this had with and without Netlabel enabled?
Joshua Brindle started a thread on the SELinux list which uncovered at least two problems that I am aware of, one with NetLabel overriding the peer's context and one with IPsec labeling returning the context from the wrong SA (I may have the details of this wrong, Venkat posted about this in the same thread). This patch correct the NetLabel problem and was tested using the simple reproducer provided by Joshua; I tested both on a NetLabel'd and non-NetLabel'd connection and the results were correct. I have also tested this patch with the getpeercon() enabled version of xinetd in FC/Rawhide; when configured to use labeled connections and used with NetLabel the daemon is spawned with the correct MLS label, when configured to use labeled connections and NetLabel is not used xinetd reports an error because getpeercon() returns an error (desired behavior as there is no network label present for the connection).
If there is some other test you would like to see run let me know and I'll give it a shot and report the results.
Signature · 2 lines
paul moore linux security @ hp
PATCH 2/6 NetLabel: make the CIPSOv4 cache spinlocks bottom half safe
In reply to [email protected]The CIPSOv4 cache traversal routines are triggered both the userspace events (cache invalidation due to DOI removal or updated SELinux policy) and network packet processing events. As a result there is a problem with the existing CIPSOv4 cache spinlocks as they are not bottom-half/softirq safe. This patch converts the CIPSOv4 cache spin_[un]lock() calls into spin_[un]lock_bh() calls to address this problem.
--- net/ipv4/cipso_ipv4.c | 16 ++++++++-------- 1 files changed, 8 insertions(+), 8 deletions(-)
net/ipv4/cipso_ipv4.c +8 -8
@@ -259,7 +259,7 @@ void cipso_v4_cache_invalidate(void)u32 iter;for (iter =0 ; iter <CIPSO_V4_CACHE_BUCKETS ; iter++) { -spin_lock (&cipso_v4_cache[iter].lock); +spin_lock_bh (&cipso_v4_cache[iter].lock);list_for_each_entry_safe (entry, tmp_entry, &cipso_v4_cache[iter].list, list) {
@@ -267,7 +267,7 @@ void cipso_v4_cache_invalidate(void)cipso_v4_cache_entry_free (entry); } cipso_v4_cache[iter].size =0 ; -spin_unlock (&cipso_v4_cache[iter].lock); +spin_unlock_bh (&cipso_v4_cache[iter].lock); }return ;
@@ -309,7 +309,7 @@ static int cipso_v4_cache_check(const un hash =cipso_v4_map_cache_hash (key, key_len); bkt = hash & (CIPSO_V4_CACHE_BUCKETBITS -1 ); -spin_lock (&cipso_v4_cache[bkt].lock); +spin_lock_bh (&cipso_v4_cache[bkt].lock);list_for_each_entry (entry, &cipso_v4_cache[bkt].list, list) { if (entry->hash == hash && entry->key_len == key_len &&
@@ -318,7 +318,7 @@ static int cipso_v4_cache_check(const un secattr->cache.free = entry->lsm_data.free; secattr->cache.data = entry->lsm_data.data;if (prev_entry ==NULL ) { -spin_unlock (&cipso_v4_cache[bkt].lock); +spin_unlock_bh (&cipso_v4_cache[bkt].lock);return 0 ; }
@@ -333,12 +333,12 @@ static int cipso_v4_cache_check(const un &prev_entry->list); } -spin_unlock (&cipso_v4_cache[bkt].lock); +spin_unlock_bh (&cipso_v4_cache[bkt].lock);return 0 ; } prev_entry = entry; } -spin_unlock (&cipso_v4_cache[bkt].lock); +spin_unlock_bh (&cipso_v4_cache[bkt].lock);return -ENOENT ; }
@@ -387,7 +387,7 @@ int cipso_v4_cache_add(const struct sk_b entry->lsm_data.data = secattr->cache.data; bkt = entry->hash & (CIPSO_V4_CACHE_BUCKETBITS -1 ); -spin_lock (&cipso_v4_cache[bkt].lock); +spin_lock_bh (&cipso_v4_cache[bkt].lock);if (cipso_v4_cache[bkt].size < cipso_v4_cache_bucketsize) {list_add (&entry->list, &cipso_v4_cache[bkt].list); cipso_v4_cache[bkt].size +=1 ;
@@ -398,7 +398,7 @@ int cipso_v4_cache_add(const struct sk_blist_add (&entry->list, &cipso_v4_cache[bkt].list);cipso_v4_cache_entry_free (old_entry); } -spin_unlock (&cipso_v4_cache[bkt].lock); +spin_unlock_bh (&cipso_v4_cache[bkt].lock);return 0 ;
--
paul moore
linux security @ hp
PATCH 3/6 NetLabel: change the SELinux permissions
In reply to [email protected]Change NetLabel to use the 'recvfrom' socket permission and the SECINITSID_NETMSG SELinux SID as the NetLabel base SID for incoming packets. This patch effectively makes the old, and currently unused, SELinux NETMSG permissions NetLabel permissions.
--- security/selinux/ss/services.c | 8 ++++---- 1 files changed, 4 insertions(+), 4 deletions(-)
security/selinux/ss/services.c +4 -4
@@ -2611,7 +2611,7 @@ int selinux_netlbl_sock_rcv_skb(struct su32 netlbl_sid;u32 recv_perm; - rc =selinux_netlbl_skbuff_getsid (skb, sksec->sid, &netlbl_sid); + rc =selinux_netlbl_skbuff_getsid (skb,SECINITSID_NETMSG , &netlbl_sid);if (rc !=0 )return rc;
@@ -2620,13 +2620,13 @@ int selinux_netlbl_sock_rcv_skb(struct sswitch (sksec->sclass) {case SECCLASS_UDP_SOCKET : - recv_perm =UDP_SOCKET__RECV_MSG ; + recv_perm =UDP_SOCKET__RECVFROM ;break ;case SECCLASS_TCP_SOCKET : - recv_perm =TCP_SOCKET__RECV_MSG ; + recv_perm =TCP_SOCKET__RECVFROM ;break ;default : - recv_perm =RAWIP_SOCKET__RECV_MSG ; + recv_perm =RAWIP_SOCKET__RECVFROM ; } rc = avc_has_perm(sksec->sid,
--
paul moore
linux security @ hp
PATCH 4/6 NetLabel: rework the Netlink attribute handling (part 1)
In reply to [email protected]At the suggestion of Thomas Graf, rewrite NetLabel's use of Netlink attributes to better follow the common Netlink attribute usage.
--- include/net/cipso_ipv4.h | 16 +- include/net/netlabel.h | 49 -------- net/ipv4/cipso_ipv4.c | 203 ++++--------------------------------- net/netlabel/netlabel_domainhash.c | 183 ++++++--------------------------- net/netlabel/netlabel_domainhash.h | 6 - net/netlabel/netlabel_user.c | 65 ++--------- net/netlabel/netlabel_user.h | 146 +------------------------- 7 files changed, 93 insertions(+), 575 deletions(-)
include/net/cipso_ipv4.h +7 -9
@@ -130,8 +130,9 @@ extern int cipso_v4_rbm_strictvalid;int cipso_v4_doi_add (struct cipso_v4_doi *doi_def);int cipso_v4_doi_remove (u32 doi,void (*callback) (struct rcu_head * head));struct cipso_v4_doi *cipso_v4_doi_getdef (u32 doi); -struct sk_buff *cipso_v4_doi_dump_all (size_t headroom); -struct sk_buff *cipso_v4_doi_dump (u32 doi,size_t headroom); +int cipso_v4_doi_walk (u32 *skip_cnt, +int (*callback) (struct cipso_v4_doi *doi_def,void *arg), +void *cb_arg);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);
@@ -152,14 +153,11 @@ static inline struct cipso_v4_doi *cipsoreturn NULL ; } -static inline struct sk_buff *cipso_v4_doi_dump_all (size_t headroom) +static inline int cipso_v4_doi_walk (u32 *skip_cnt, +int (*callback) (struct cipso_v4_doi *doi_def,void *arg), +void *cb_arg) { -return NULL ; -} - -static inline struct sk_buff *cipso_v4_doi_dump (u32 doi,size_t headroom) -{ -return NULL ; +return 0 ; } static inline int cipso_v4_doi_domhsh_add(struct cipso_v4_doi *doi_def,
include/net/netlabel.h +2 -47
61 lines not on this page · raw message
net/ipv4/cipso_ipv4.c +24 -179
221 lines not on this page · raw message
net/netlabel/netlabel_domainhash.c +37 -146
197 lines not on this page · raw message
net/netlabel/netlabel_domainhash.h +4 -2
@@ -61,7 +61,9 @@ int netlbl_domhsh_add(struct netlbl_dom_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 (size_t headroom); -struct sk_buff *netlbl_domhsh_dump_default (size_t headroom); +int netlbl_domhsh_walk (u32 *skip_bkt, +u32 *skip_chain, +int (*callback) (struct netlbl_dom_map *entry,void *arg), +void *cb_arg);#endif
net/netlabel/netlabel_user.c +12 -53
95 lines not on this page · raw message
net/netlabel/netlabel_user.h +7 -139
170 lines not on this page · raw message
--
paul moore
linux security @ hp
Re: [PATCH 4/6] NetLabel: rework the Netlink attribute handling (part 1)
* [email protected] <[email protected]> 2006-09-21 12:57
net/netlabel/netlabel_user.c int netlbl_netlink_init(void) · line 117
Quoted 9 lines
net/netlabel/netlabel_user.c +5
-1-int netlbl_netlink_snd_multicast (struct sk_buff *skb,u32 pid,u32 group) +int netlbl_netlink_snd_multicast (struct sk_buff *skb, +void *buf, +u32 pid, +u32 group) { +genlmsg_end (skb, buf);return genlmsg_multicast (skb, pid, group,GFP_KERNEL ); }
This is a bit weird at first, one would expect that the message is terminated when it is build but I guess this is ok for now.
Rest seems fine.
PATCH 5/6 NetLabel: rework the Netlink attribute handling (part 2)
In reply to [email protected]At the suggestion of Thomas Graf, rewrite NetLabel's use of Netlink attributes to better follow the common Netlink attribute usage.
--- net/netlabel/netlabel_cipso_v4.c | 593 +++++++++++++++++++++++++------------- net/netlabel/netlabel_cipso_v4.h | 235 +++++---------- net/netlabel/netlabel_mgmt.c | 531 +++++++++++++++++----------------- net/netlabel/netlabel_mgmt.h | 213 ++++--------- net/netlabel/netlabel_unlabeled.c | 77 ++-- net/netlabel/netlabel_unlabeled.h | 41 +- 6 files changed, 875 insertions(+), 815 deletions(-)
net/netlabel/netlabel_cipso_v4.c +396 -197
@@ -41,15 +41,37 @@#include "netlabel_user.h" #include "netlabel_cipso_v4.h" +/* Argument struct for cipso_v4_doi_walk() */ +struct netlbl_cipsov4_doiwalk_arg { +struct netlink_callback *nl_cb; +struct sk_buff *skb; +u32 seq; +}; +/* NetLabel Generic NETLINK CIPSOv4 family */ static struct genl_family netlbl_cipsov4_gnl_family = { .id =GENL_ID_GENERATE , .hdrsize =0 , .name =NETLBL_NLTYPE_CIPSOV4_NAME , .version =NETLBL_PROTO_VERSION , - .maxattr =0 , + .maxattr =NLBL_CIPSOV4_A_MAX , }; +/* NetLabel Netlink attribute policy */ +static struct nla_policy netlbl_cipsov4_genl_policy[NLBL_CIPSOV4_A_MAX +1 ] = { + [NLBL_CIPSOV4_A_DOI ] = { .type =NLA_U32 }, + [NLBL_CIPSOV4_A_MTYPE ] = { .type =NLA_U32 }, + [NLBL_CIPSOV4_A_TAG ] = { .type =NLA_U8 }, + [NLBL_CIPSOV4_A_TAGLST ] = { .type =NLA_NESTED }, + [NLBL_CIPSOV4_A_MLSLVLLOC ] = { .type =NLA_U32 }, + [NLBL_CIPSOV4_A_MLSLVLREM ] = { .type =NLA_U32 }, + [NLBL_CIPSOV4_A_MLSLVL ] = { .type =NLA_NESTED }, + [NLBL_CIPSOV4_A_MLSLVLLST ] = { .type =NLA_NESTED }, + [NLBL_CIPSOV4_A_MLSCATLOC ] = { .type =NLA_U32 }, + [NLBL_CIPSOV4_A_MLSCATREM ] = { .type =NLA_U32 }, + [NLBL_CIPSOV4_A_MLSCAT ] = { .type =NLA_NESTED }, + [NLBL_CIPSOV4_A_MLSCATLST ] = { .type =NLA_NESTED }, +};/* * Helper Functions
@@ -81,16 +103,13 @@ static void netlbl_cipsov4_doi_free(strukfree (ptr); } -/* * NetLabel Command Handlers */ /** * netlbl_cipsov4_add_std - Adds a CIPSO V4 DOI definition -* @doi: the DOI value -* @msg: the ADD message data -* @msg_size: the size of the ADD message buffer +* @info: the Generic NETLINK info block * * Description: * Create a new CIPSO_V4_MAP_STD DOI definition based on the given ADD message
@@ -98,22 +117,31 @@ static void netlbl_cipsov4_doi_free(stru* error. * */ -static int netlbl_cipsov4_add_std (u32 doi,struct nlattr *msg,size_t msg_size) +static int netlbl_cipsov4_add_std (struct genl_info *info) {int ret_val = -EINVAL ; -int msg_len = msg_size; -u32 num_tags; -u32 num_lvls; -u32 num_cats;struct cipso_v4_doi *doi_def =NULL ; -u32 iter; -u32 tmp_val_a; -u32 tmp_val_b; - -if (msg_len <NETLBL_LEN_U32 ) - goto add_std_failure; - num_tags =netlbl_getinc_u32 (&msg, &msg_len); -if (num_tags ==0 || num_tags >CIPSO_V4_TAG_MAXCNT ) +struct nlattr *nla_a; +struct nlattr *nla_b; +struct nlattr *nla_c; +int nla_a_rem; +int nla_b_rem; +u32 iter =0 ; + +if (!info->attrs[NLBL_CIPSOV4_A_DOI ] || + !info->attrs[NLBL_CIPSOV4_A_TAGLST ] || + !info->attrs[NLBL_CIPSOV4_A_MLSLVLLST ]) + goto add_std_failure; + +if (nla_validate (nla_data (info->attrs[NLBL_CIPSOV4_A_TAGLST ]), +nla_len (info->attrs[NLBL_CIPSOV4_A_TAGLST ]), +NLBL_CIPSOV4_A_MAX , + netlbl_cipsov4_genl_policy) !=0 ) + goto add_std_failure; +if (nla_validate (nla_data (info->attrs[NLBL_CIPSOV4_A_MLSLVLLST ]), +nla_len (info->attrs[NLBL_CIPSOV4_A_MLSLVLLST ]), +NLBL_CIPSOV4_A_MAX , + netlbl_cipsov4_genl_policy) !=0 ) goto add_std_failure; doi_def =kmalloc (sizeof (*doi_def),GFP_KERNEL );
@@ -128,28 +156,44 @@ static int netlbl_cipsov4_add_std(u32 do } doi_def->type =CIPSO_V4_MAP_STD ; -for (iter =0 ; iter < num_tags; iter++) { -if (msg_len <NETLBL_LEN_U8 ) - goto add_std_failure; - doi_def->tags[iter] =netlbl_getinc_u8 (&msg, &msg_len); -switch (doi_def->tags[iter]) { -case CIPSO_V4_TAG_RBITMAP : -break ; -default : - goto add_std_failure; +nla_for_each_attr (nla_a, +nla_data (info->attrs[NLBL_CIPSOV4_A_TAGLST ]), +nla_len (info->attrs[NLBL_CIPSOV4_A_TAGLST ]), + nla_a_rem) +if (nla_a->nla_type ==NLBL_CIPSOV4_A_TAG ) { +if (iter >CIPSO_V4_TAG_MAXCNT ) + goto add_std_failure; + doi_def->tags[iter++] =nla_get_u8 (nla_a); } - }if (iter <CIPSO_V4_TAG_MAXCNT ) doi_def->tags[iter] =CIPSO_V4_TAG_INVALID ; -if (msg_len <6 *NETLBL_LEN_U32 ) - goto add_std_failure; - - num_lvls =netlbl_getinc_u32 (&msg, &msg_len); -if (num_lvls ==0 ) - goto add_std_failure; - doi_def->map.std->lvl.local_size =netlbl_getinc_u32 (&msg, &msg_len); -if (doi_def->map.std->lvl.local_size >CIPSO_V4_MAX_LOC_LVLS ) +nla_for_each_attr (nla_a, +nla_data (info->attrs[NLBL_CIPSOV4_A_MLSLVLLST ]), +nla_len (info->attrs[NLBL_CIPSOV4_A_MLSLVLLST ]), + nla_a_rem) +if (nla_a->nla_type ==NLBL_CIPSOV4_A_MLSLVL ) { +nla_for_each_attr (nla_b, +nla_data (nla_a), +nla_len (nla_a), + nla_b_rem) +switch (nla_b->nla_type) { +case NLBL_CIPSOV4_A_MLSLVLLOC : +if (nla_get_u32 (nla_b) >= + doi_def->map.std->lvl.local_size) + doi_def->map.std->lvl.local_size = +nla_get_u32 (nla_b) +1 ; +break ; +case NLBL_CIPSOV4_A_MLSLVLREM : +if (nla_get_u32 (nla_b) >= + doi_def->map.std->lvl.cipso_size) + doi_def->map.std->lvl.cipso_size = +nla_get_u32 (nla_b) +1 ; +break ; + } + } +if (doi_def->map.std->lvl.local_size >CIPSO_V4_MAX_LOC_LVLS || + doi_def->map.std->lvl.cipso_size >CIPSO_V4_MAX_REM_LVLS ) goto add_std_failure; doi_def->map.std->lvl.local = kcalloc(doi_def->map.std->lvl.local_size, sizeof(u32),
@@ -158,9 +202,6 @@ static int netlbl_cipsov4_add_std(u32 do ret_val = -ENOMEM ; goto add_std_failure; } - doi_def->map.std->lvl.cipso_size =netlbl_getinc_u8 (&msg, &msg_len); -if (doi_def->map.std->lvl.cipso_size >CIPSO_V4_MAX_REM_LVLS ) - goto add_std_failure; doi_def->map.std->lvl.cipso =kcalloc (doi_def->map.std->lvl.cipso_size,sizeof (u32),GFP_KERNEL );
@@ -168,68 +209,96 @@ static int netlbl_cipsov4_add_std(u32 do ret_val = -ENOMEM ; goto add_std_failure; } +nla_for_each_attr (nla_a, +nla_data (info->attrs[NLBL_CIPSOV4_A_MLSLVLLST ]), +nla_len (info->attrs[NLBL_CIPSOV4_A_MLSLVLLST ]), + nla_a_rem) +if (nla_a->nla_type ==NLBL_CIPSOV4_A_MLSLVL ) { + nla_b =nla_find_nested (nla_a, +NLBL_CIPSOV4_A_MLSLVLLOC ); + nla_c =nla_find_nested (nla_a, +NLBL_CIPSOV4_A_MLSLVLREM ); +if (nla_b ==NULL || nla_c ==NULL ) + goto add_std_failure; + doi_def->map.std->lvl.local[nla_get_u32 (nla_b)] = +nla_get_u32 (nla_c); + doi_def->map.std->lvl.cipso[nla_get_u32 (nla_c)] = +nla_get_u32 (nla_b); + } - num_cats =netlbl_getinc_u32 (&msg, &msg_len); - doi_def->map.std->cat.local_size =netlbl_getinc_u32 (&msg, &msg_len); -if (doi_def->map.std->cat.local_size >CIPSO_V4_MAX_LOC_CATS ) - goto add_std_failure; - doi_def->map.std->cat.local =kcalloc (doi_def->map.std->cat.local_size, +if (info->attrs[NLBL_CIPSOV4_A_MLSCATLST ]) { +if (nla_validate ( +nla_data (info->attrs[NLBL_CIPSOV4_A_MLSCATLST ]), +nla_len (info->attrs[NLBL_CIPSOV4_A_MLSCATLST ]), +NLBL_CIPSOV4_A_MAX , + netlbl_cipsov4_genl_policy) !=0 ) + goto add_std_failure; + +nla_for_each_attr (nla_a, +nla_data (info->attrs[NLBL_CIPSOV4_A_MLSCATLST ]), +nla_len (info->attrs[NLBL_CIPSOV4_A_MLSCATLST ]), + nla_a_rem) +if (nla_a->nla_type ==NLBL_CIPSOV4_A_MLSCAT ) { +nla_for_each_attr (nla_b, +nla_data (nla_a), +nla_len (nla_a), + nla_b_rem) +switch (nla_b->nla_type) { +case NLBL_CIPSOV4_A_MLSCATLOC : +if (nla_get_u32 (nla_b) >= + doi_def->map.std->cat.local_size) + doi_def->map.std->cat.local_size = +nla_get_u32 (nla_b) +1 ; +break ; +case NLBL_CIPSOV4_A_MLSCATREM : +if (nla_get_u32 (nla_b) >= + doi_def->map.std->cat.cipso_size) + doi_def->map.std->cat.cipso_size = +nla_get_u32 (nla_b) +1 ; +break ; + } + } +if (doi_def->map.std->cat.local_size >CIPSO_V4_MAX_LOC_CATS || + doi_def->map.std->cat.cipso_size >CIPSO_V4_MAX_REM_CATS ) + goto add_std_failure; + doi_def->map.std->cat.local =kcalloc ( + doi_def->map.std->cat.local_size,sizeof (u32),GFP_KERNEL ); -if (doi_def->map.std->cat.local ==NULL ) { - ret_val = -ENOMEM ; - goto add_std_failure; - } - doi_def->map.std->cat.cipso_size =netlbl_getinc_u16 (&msg, &msg_len); -if (doi_def->map.std->cat.cipso_size >CIPSO_V4_MAX_REM_CATS ) - goto add_std_failure; - doi_def->map.std->cat.cipso =kcalloc (doi_def->map.std->cat.cipso_size, +if (doi_def->map.std->cat.local ==NULL ) { + ret_val = -ENOMEM ; + goto add_std_failure; + } + doi_def->map.std->cat.cipso =kcalloc ( + doi_def->map.std->cat.cipso_size,sizeof (u32),GFP_KERNEL ); -if (doi_def->map.std->cat.cipso ==NULL ) { - ret_val = -ENOMEM ; - goto add_std_failure; - } - -if (msg_len < - num_lvls * (NETLBL_LEN_U32 +NETLBL_LEN_U8 ) + - num_cats * (NETLBL_LEN_U32 +NETLBL_LEN_U16 )) - goto add_std_failure; - -for (iter =0 ; iter < doi_def->map.std->lvl.cipso_size; iter++) - doi_def->map.std->lvl.cipso[iter] =CIPSO_V4_INV_LVL ; -for (iter =0 ; iter < doi_def->map.std->lvl.local_size; iter++) - doi_def->map.std->lvl.local[iter] =CIPSO_V4_INV_LVL ; -for (iter =0 ; iter < doi_def->map.std->cat.cipso_size; iter++) - doi_def->map.std->cat.cipso[iter] =CIPSO_V4_INV_CAT ; -for (iter =0 ; iter < doi_def->map.std->cat.local_size; iter++) - doi_def->map.std->cat.local[iter] =CIPSO_V4_INV_CAT ; - -for (iter =0 ; iter < num_lvls; iter++) { - tmp_val_a =netlbl_getinc_u32 (&msg, &msg_len); - tmp_val_b =netlbl_getinc_u8 (&msg, &msg_len); - -if (tmp_val_a >= doi_def->map.std->lvl.local_size || - tmp_val_b >= doi_def->map.std->lvl.cipso_size) +if (doi_def->map.std->cat.cipso ==NULL ) { + ret_val = -ENOMEM ; goto add_std_failure; - - doi_def->map.std->lvl.cipso[tmp_val_b] = tmp_val_a; - doi_def->map.std->lvl.local[tmp_val_a] = tmp_val_b; + } +nla_for_each_attr (nla_a, +nla_data (info->attrs[NLBL_CIPSOV4_A_MLSCATLST ]), +nla_len (info->attrs[NLBL_CIPSOV4_A_MLSCATLST ]), + nla_a_rem) +if (nla_a->nla_type ==NLBL_CIPSOV4_A_MLSCAT ) { + nla_b =nla_find_nested (nla_a, +NLBL_CIPSOV4_A_MLSCATLOC ); + nla_c =nla_find_nested (nla_a, +NLBL_CIPSOV4_A_MLSCATREM ); +if (nla_b ==NULL || nla_c ==NULL ) + goto add_std_failure; + doi_def->map.std->cat.local[ +nla_get_u32 (nla_b)] = +nla_get_u32 (nla_c); + doi_def->map.std->cat.cipso[ +nla_get_u32 (nla_c)] = +nla_get_u32 (nla_b); + } } -for (iter =0 ; iter < num_cats; iter++) { - tmp_val_a =netlbl_getinc_u32 (&msg, &msg_len); - tmp_val_b =netlbl_getinc_u16 (&msg, &msg_len); - -if (tmp_val_a >= doi_def->map.std->cat.local_size || - tmp_val_b >= doi_def->map.std->cat.cipso_size) - goto add_std_failure; + doi_def->doi =nla_get_u32 (info->attrs[NLBL_CIPSOV4_A_DOI ]); - doi_def->map.std->cat.cipso[tmp_val_b] = tmp_val_a; - doi_def->map.std->cat.local[tmp_val_a] = tmp_val_b; - } - - doi_def->doi = doi; ret_val =cipso_v4_doi_add (doi_def);if (ret_val !=0 ) goto add_std_failure;
@@ -243,9 +312,7 @@ add_std_failure:/** * netlbl_cipsov4_add_pass - Adds a CIPSO V4 DOI definition -* @doi: the DOI value -* @msg: the ADD message data -* @msg_size: the size of the ADD message buffer +* @info: the Generic NETLINK info block * * Description: * Create a new CIPSO_V4_MAP_PASS DOI definition based on the given ADD message
@@ -253,20 +320,22 @@ add_std_failure:* error. * */ -static int netlbl_cipsov4_add_pass (u32 doi, -struct nlattr *msg, -size_t msg_size) +static int netlbl_cipsov4_add_pass (struct genl_info *info) {int ret_val = -EINVAL ; -int msg_len = msg_size; -u32 num_tags;struct cipso_v4_doi *doi_def =NULL ; -u32 iter; +struct nlattr *nla; +int nla_rem; +u32 iter =0 ; -if (msg_len <NETLBL_LEN_U32 ) +if (!info->attrs[NLBL_CIPSOV4_A_DOI ] || + !info->attrs[NLBL_CIPSOV4_A_TAGLST ]) goto add_pass_failure; - num_tags =netlbl_getinc_u32 (&msg, &msg_len); -if (num_tags ==0 || num_tags >CIPSO_V4_TAG_MAXCNT ) + +if (nla_validate (nla_data (info->attrs[NLBL_CIPSOV4_A_TAGLST ]), +nla_len (info->attrs[NLBL_CIPSOV4_A_TAGLST ]), +NLBL_CIPSOV4_A_MAX , + netlbl_cipsov4_genl_policy) !=0 ) goto add_pass_failure; doi_def =kmalloc (sizeof (*doi_def),GFP_KERNEL );
@@ -276,21 +345,20 @@ static int netlbl_cipsov4_add_pass(u32 d } doi_def->type =CIPSO_V4_MAP_PASS ; -for (iter =0 ; iter < num_tags; iter++) { -if (msg_len <NETLBL_LEN_U8 ) - goto add_pass_failure; - doi_def->tags[iter] =netlbl_getinc_u8 (&msg, &msg_len); -switch (doi_def->tags[iter]) { -case CIPSO_V4_TAG_RBITMAP : -break ; -default : - goto add_pass_failure; +nla_for_each_attr (nla, +nla_data (info->attrs[NLBL_CIPSOV4_A_TAGLST ]), +nla_len (info->attrs[NLBL_CIPSOV4_A_TAGLST ]), + nla_rem) +if (nla->nla_type ==NLBL_CIPSOV4_A_TAG ) { +if (iter >CIPSO_V4_TAG_MAXCNT ) + goto add_pass_failure; + doi_def->tags[iter++] =nla_get_u8 (nla); } - }if (iter <CIPSO_V4_TAG_MAXCNT ) doi_def->tags[iter] =CIPSO_V4_TAG_INVALID ; - doi_def->doi = doi; + doi_def->doi =nla_get_u32 (info->attrs[NLBL_CIPSOV4_A_DOI ]); + ret_val =cipso_v4_doi_add (doi_def);if (ret_val !=0 ) goto add_pass_failure;
@@ -316,34 +384,21 @@ static int netlbl_cipsov4_add(struct sk_ {int ret_val = -EINVAL ; -u32 doi;u32 map_type; -int msg_len =netlbl_netlink_payload_len (skb); -struct nlattr *msg =netlbl_netlink_payload_data (skb); - ret_val =netlbl_netlink_cap_check (skb,CAP_NET_ADMIN ); -if (ret_val !=0 ) - goto add_return; - -if (msg_len <2 *NETLBL_LEN_U32 ) - goto add_return; +if (!info->attrs[NLBL_CIPSOV4_A_MTYPE ]) +return -EINVAL ; - doi =netlbl_getinc_u32 (&msg, &msg_len); - map_type =netlbl_getinc_u32 (&msg, &msg_len); + map_type =nla_get_u32 (info->attrs[NLBL_CIPSOV4_A_MTYPE ]);switch (map_type) {case CIPSO_V4_MAP_STD : - ret_val =netlbl_cipsov4_add_std (doi, msg, msg_len); + ret_val =netlbl_cipsov4_add_std (info);break ;case CIPSO_V4_MAP_PASS : - ret_val =netlbl_cipsov4_add_pass (doi, msg, msg_len); + ret_val =netlbl_cipsov4_add_pass (info);break ; } -add_return: -netlbl_netlink_send_ack (info, - netlbl_cipsov4_gnl_family.id, -NLBL_CIPSOV4_C_ACK , - -ret_val);return ret_val; }
@@ -353,84 +408,237 @@ add_return:* @info: the Generic NETLINK info block * * Description: -* Process a user generated LIST message and respond accordingly. Returns -* zero on success and negative values on error. +* Process a user generated LIST message and respond accordingly. While the +* response message generated by the kernel is straightforward, determining +* before hand the size of the buffer to allocate is not (we have to generate +* the message to know the size). In order to keep this function sane what we +* do is allocate a buffer of NLMSG_GOODSIZE and try to fit the response in +* that size, if we fail then we restart with a larger buffer and try again. +* We continue in this manner until we hit a limit of failed attempts then we +* give up and just send an error message. Returns zero on success and +* negative values on error. * */ static int netlbl_cipsov4_list (struct sk_buff *skb,struct genl_info *info) { -int ret_val = -EINVAL ; +int ret_val; +struct sk_buff *ans_skb =NULL ; +u32 nlsze_mult =1 ; +void *data;u32 doi; -struct nlattr *msg =netlbl_netlink_payload_data (skb); -struct sk_buff *ans_skb; +struct nlattr *nla_a; +struct nlattr *nla_b; +struct cipso_v4_doi *doi_def; +u32 iter; -if (netlbl_netlink_payload_len (skb) !=NETLBL_LEN_U32 ) +if (!info->attrs[NLBL_CIPSOV4_A_DOI ]) { + ret_val = -EINVAL ; goto list_failure; + } - doi =nla_get_u32 (msg); - ans_skb =cipso_v4_doi_dump (doi,NLMSG_SPACE (GENL_HDRLEN )); +list_start: + ans_skb =nlmsg_new (NLMSG_GOODSIZE * nlsze_mult,GFP_KERNEL );if (ans_skb ==NULL ) { ret_val = -ENOMEM ; goto list_failure; } -netlbl_netlink_hdr_push (ans_skb, - info->snd_pid, -0 , - netlbl_cipsov4_gnl_family.id, -NLBL_CIPSOV4_C_LIST ); + data =netlbl_netlink_hdr_put (ans_skb, + info->snd_pid, + info->snd_seq, + netlbl_cipsov4_gnl_family.id, +0 , +NLBL_CIPSOV4_C_LIST ); +if (data ==NULL ) { + ret_val = -ENOMEM ; + goto list_failure; + } + + doi =nla_get_u32 (info->attrs[NLBL_CIPSOV4_A_DOI ]); + +rcu_read_lock (); + doi_def =cipso_v4_doi_getdef (doi); +if (doi_def ==NULL ) { + ret_val = -EINVAL ; + goto list_failure; + } + + ret_val =nla_put_u32 (ans_skb,NLBL_CIPSOV4_A_MTYPE , doi_def->type); +if (ret_val !=0 ) + goto list_failure_lock; + + nla_a =nla_nest_start (ans_skb,NLBL_CIPSOV4_A_TAGLST ); +if (nla_a ==NULL ) { + ret_val = -ENOMEM ; + goto list_failure_lock; + } +for (iter =0 ; + iter <CIPSO_V4_TAG_MAXCNT && + doi_def->tags[iter] !=CIPSO_V4_TAG_INVALID ; + iter++) { + ret_val =nla_put_u8 (ans_skb, +NLBL_CIPSOV4_A_TAG , + doi_def->tags[iter]); +if (ret_val !=0 ) + goto list_failure_lock; + } +nla_nest_end (ans_skb, nla_a); - ret_val =netlbl_netlink_snd (ans_skb, info->snd_pid); +switch (doi_def->type) { +case CIPSO_V4_MAP_STD : + nla_a =nla_nest_start (ans_skb,NLBL_CIPSOV4_A_MLSLVLLST ); +if (nla_a ==NULL ) { + ret_val = -ENOMEM ; + goto list_failure_lock; + } +for (iter =0 ; + iter < doi_def->map.std->lvl.local_size; + iter++) { +if (doi_def->map.std->lvl.local[iter] == +CIPSO_V4_INV_LVL ) +continue ; + + nla_b =nla_nest_start (ans_skb,NLBL_CIPSOV4_A_MLSLVL ); +if (nla_b ==NULL ) { + ret_val = -ENOMEM ; + goto list_retry; + } + ret_val =nla_put_u32 (ans_skb, +NLBL_CIPSOV4_A_MLSLVLLOC , + iter); +if (ret_val !=0 ) + goto list_retry; + ret_val =nla_put_u32 (ans_skb, +NLBL_CIPSOV4_A_MLSLVLREM , + doi_def->map.std->lvl.local[iter]); +if (ret_val !=0 ) + goto list_retry; +nla_nest_end (ans_skb, nla_b); + } +nla_nest_end (ans_skb, nla_a); + + nla_a =nla_nest_start (ans_skb,NLBL_CIPSOV4_A_MLSCATLST ); +if (nla_a ==NULL ) { + ret_val = -ENOMEM ; + goto list_retry; + } +for (iter =0 ; + iter < doi_def->map.std->cat.local_size; + iter++) { +if (doi_def->map.std->cat.local[iter] == +CIPSO_V4_INV_CAT ) +continue ; + + nla_b =nla_nest_start (ans_skb,NLBL_CIPSOV4_A_MLSCAT ); +if (nla_b ==NULL ) { + ret_val = -ENOMEM ; + goto list_retry; + } + ret_val =nla_put_u32 (ans_skb, +NLBL_CIPSOV4_A_MLSCATLOC , + iter); +if (ret_val !=0 ) + goto list_retry; + ret_val =nla_put_u32 (ans_skb, +NLBL_CIPSOV4_A_MLSCATREM , + doi_def->map.std->cat.local[iter]); +if (ret_val !=0 ) + goto list_retry; +nla_nest_end (ans_skb, nla_b); + } +nla_nest_end (ans_skb, nla_a); + +break ; + } +rcu_read_unlock (); + + ret_val =netlbl_netlink_snd (ans_skb, data, info->snd_pid);if (ret_val !=0 ) goto list_failure;return 0 ; +list_retry: +/* XXX - this limit is a guesstimate */ +if (nlsze_mult <4 ) { +rcu_read_unlock (); +kfree_skb (ans_skb); + nlsze_mult++; + goto list_start; + } +list_failure_lock: +rcu_read_unlock (); list_failure: -netlbl_netlink_send_ack (info, - netlbl_cipsov4_gnl_family.id, -NLBL_CIPSOV4_C_ACK , - -ret_val); +kfree_skb (ans_skb); +return ret_val; +} + +/** +* netlbl_cipsov4_listall_cb - cipso_v4_doi_walk() callback for LISTALL +* @doi_def: the CIPSOv4 DOI definition +* @arg: the netlbl_cipsov4_doiwalk_arg structure +* +* Description: +* This function is designed to be used as a callback to the +* cipso_v4_doi_walk() function for use in generating a response for a LISTALL +* message. Returns the size of the message on success, negative values on +* failure. +* +*/ +static int netlbl_cipsov4_listall_cb (struct cipso_v4_doi *doi_def,void *arg) +{ +int ret_val = -ENOMEM ; +struct netlbl_cipsov4_doiwalk_arg *cb_arg = arg; +void *data; + + data =netlbl_netlink_hdr_put (cb_arg->skb, +NETLINK_CB (cb_arg->nl_cb->skb).pid, + cb_arg->seq, + netlbl_cipsov4_gnl_family.id, +NLM_F_MULTI , +NLBL_CIPSOV4_C_LISTALL ); +if (data ==NULL ) + goto listall_cb_failure; + + ret_val =nla_put_u32 (cb_arg->skb,NLBL_CIPSOV4_A_DOI , doi_def->doi); +if (ret_val !=0 ) + goto listall_cb_failure; + ret_val =nla_put_u32 (cb_arg->skb, +NLBL_CIPSOV4_A_MTYPE , + doi_def->type); +if (ret_val !=0 ) + goto listall_cb_failure; + +return genlmsg_end (cb_arg->skb, data); + +listall_cb_failure: +genlmsg_cancel (cb_arg->skb, data);return ret_val; }/** * netlbl_cipsov4_listall - Handle a LISTALL message * @skb: the NETLINK buffer -* @info: the Generic NETLINK info block +* @cb: the NETLINK callback * * Description: * Process a user generated LISTALL message and respond accordingly. Returns * zero on success and negative values on error. * */ -static int netlbl_cipsov4_listall (struct sk_buff *skb,struct genl_info *info) +static int netlbl_cipsov4_listall (struct sk_buff *skb, +struct netlink_callback *cb) { -int ret_val = -EINVAL ; -struct sk_buff *ans_skb; +struct netlbl_cipsov4_doiwalk_arg cb_arg; +int doi_skip = cb->args[0 ]; - ans_skb =cipso_v4_doi_dump_all (NLMSG_SPACE (GENL_HDRLEN )); -if (ans_skb ==NULL ) { - ret_val = -ENOMEM ; - goto listall_failure; - } -netlbl_netlink_hdr_push (ans_skb, - info->snd_pid, -0 , - netlbl_cipsov4_gnl_family.id, -NLBL_CIPSOV4_C_LISTALL ); + cb_arg.nl_cb = cb; + cb_arg.skb = skb; + cb_arg.seq = cb->nlh->nlmsg_seq; - ret_val =netlbl_netlink_snd (ans_skb, info->snd_pid); -if (ret_val !=0 ) - goto listall_failure; - -return 0 ; +cipso_v4_doi_walk (&doi_skip, netlbl_cipsov4_listall_cb, &cb_arg); -listall_failure: -netlbl_netlink_send_ack (info, - netlbl_cipsov4_gnl_family.id, -NLBL_CIPSOV4_C_ACK , - -ret_val); -return ret_val; + cb->args[0 ] = doi_skip; +return skb->len; }/**
@@ -445,27 +653,14 @@ listall_failure:*/ static int netlbl_cipsov4_remove (struct sk_buff *skb,struct genl_info *info) { -int ret_val; +int ret_val = -EINVAL ;u32 doi; -struct nlattr *msg =netlbl_netlink_payload_data (skb); - ret_val =netlbl_netlink_cap_check (skb,CAP_NET_ADMIN ); -if (ret_val !=0 ) - goto remove_return; - -if (netlbl_netlink_payload_len (skb) !=NETLBL_LEN_U32 ) { - ret_val = -EINVAL ; - goto remove_return; +if (info->attrs[NLBL_CIPSOV4_A_DOI ]) { + doi =nla_get_u32 (info->attrs[NLBL_CIPSOV4_A_DOI ]); + ret_val =cipso_v4_doi_remove (doi, netlbl_cipsov4_doi_free); } - doi =nla_get_u32 (msg); - ret_val =cipso_v4_doi_remove (doi, netlbl_cipsov4_doi_free); - -remove_return: -netlbl_netlink_send_ack (info, - netlbl_cipsov4_gnl_family.id, -NLBL_CIPSOV4_C_ACK , - -ret_val);return ret_val; }
@@ -475,14 +670,16 @@ remove_return:static struct genl_ops netlbl_cipsov4_genl_c_add = { .cmd =NLBL_CIPSOV4_C_ADD , - .flags =0 , + .flags =GENL_ADMIN_PERM , + .policy = netlbl_cipsov4_genl_policy, .doit = netlbl_cipsov4_add, .dumpit =NULL , };static struct genl_ops netlbl_cipsov4_genl_c_remove = { .cmd =NLBL_CIPSOV4_C_REMOVE , - .flags =0 , + .flags =GENL_ADMIN_PERM , + .policy = netlbl_cipsov4_genl_policy, .doit = netlbl_cipsov4_remove, .dumpit =NULL , };
@@ -490,6 +687,7 @@ static struct genl_ops netlbl_cipsov4_gestatic struct genl_ops netlbl_cipsov4_genl_c_list = { .cmd =NLBL_CIPSOV4_C_LIST , .flags =0 , + .policy = netlbl_cipsov4_genl_policy, .doit = netlbl_cipsov4_list, .dumpit =NULL , };
@@ -497,8 +695,9 @@ static struct genl_ops netlbl_cipsov4_gestatic struct genl_ops netlbl_cipsov4_genl_c_listall = { .cmd =NLBL_CIPSOV4_C_LISTALL , .flags =0 , - .doit = netlbl_cipsov4_listall, - .dumpit =NULL , + .policy = netlbl_cipsov4_genl_policy, + .doit =NULL , + .dumpit = netlbl_cipsov4_listall, };/*
net/netlabel/netlabel_cipso_v4.h +92 -143
273 lines not on this page · raw message
net/netlabel/netlabel_mgmt.c +269 -262
748 lines not on this page · raw message
net/netlabel/netlabel_mgmt.h +69 -144
281 lines not on this page · raw message
net/netlabel/netlabel_unlabeled.c +33 -44
135 lines not on this page · raw message
net/netlabel/netlabel_unlabeled.h +16 -25
72 lines not on this page · raw message
--
paul moore
linux security @ hp
Re: [PATCH 5/6] NetLabel: rework the Netlink attribute handling (part 2)
* [email protected] <[email protected]> 2006-09-21 12:57
Quoted 126 lines
At the suggestion of Thomas Graf, rewrite NetLabel's use of Netlink attributes to better follow the common Netlink attribute usage.
Signed-off-by Paul Moore <[email protected]>--- net/netlabel/netlabel_cipso_v4.c | 593 +++++++++++++++++++++++++------------- net/netlabel/netlabel_cipso_v4.h | 235 +++++---------- net/netlabel/netlabel_mgmt.c | 531 +++++++++++++++++----------------- net/netlabel/netlabel_mgmt.h | 213 ++++--------- net/netlabel/netlabel_unlabeled.c | 77 ++-- net/netlabel/netlabel_unlabeled.h | 41 +- 6 files changed, 875 insertions(+), 815 deletions(-)Index:
net-2.6.19/net/netlabel/netlabel_cipso_v4.c
===================================================================net/netlabel/netlabel_cipso_v4.c +46
-21@@ -41,15 +41,37 @@#include "netlabel_user.h" #include "netlabel_cipso_v4.h" +/* Argument struct for cipso_v4_doi_walk() */ +struct netlbl_cipsov4_doiwalk_arg { +struct netlink_callback *nl_cb; +struct sk_buff *skb; +u32 seq; +}; +/* NetLabel Generic NETLINK CIPSOv4 family */ static struct genl_family netlbl_cipsov4_gnl_family = { .id =GENL_ID_GENERATE , .hdrsize =0 , .name =NETLBL_NLTYPE_CIPSOV4_NAME , .version =NETLBL_PROTO_VERSION , - .maxattr =0 , + .maxattr =NLBL_CIPSOV4_A_MAX , }; +/* NetLabel Netlink attribute policy */ +static struct nla_policy netlbl_cipsov4_genl_policy[NLBL_CIPSOV4_A_MAX +1 ] = { + [NLBL_CIPSOV4_A_DOI ] = { .type =NLA_U32 }, + [NLBL_CIPSOV4_A_MTYPE ] = { .type =NLA_U32 }, + [NLBL_CIPSOV4_A_TAG ] = { .type =NLA_U8 }, + [NLBL_CIPSOV4_A_TAGLST ] = { .type =NLA_NESTED }, + [NLBL_CIPSOV4_A_MLSLVLLOC ] = { .type =NLA_U32 }, + [NLBL_CIPSOV4_A_MLSLVLREM ] = { .type =NLA_U32 }, + [NLBL_CIPSOV4_A_MLSLVL ] = { .type =NLA_NESTED }, + [NLBL_CIPSOV4_A_MLSLVLLST ] = { .type =NLA_NESTED }, + [NLBL_CIPSOV4_A_MLSCATLOC ] = { .type =NLA_U32 }, + [NLBL_CIPSOV4_A_MLSCATREM ] = { .type =NLA_U32 }, + [NLBL_CIPSOV4_A_MLSCAT ] = { .type =NLA_NESTED }, + [NLBL_CIPSOV4_A_MLSCATLST ] = { .type =NLA_NESTED }, +};/* * Helper Functions @@ -81,16 +103,13 @@ static void netlbl_cipsov4_doi_free(strukfree (ptr); } -/* * NetLabel Command Handlers */ /** * netlbl_cipsov4_add_std - Adds a CIPSO V4 DOI definition -* @doi: the DOI value -* @msg: the ADD message data -* @msg_size: the size of the ADD message buffer +* @info: the Generic NETLINK info block * * Description: * Create a new CIPSO_V4_MAP_STD DOI definition based on the given ADD message @@ -98,22 +117,31 @@ static void netlbl_cipsov4_doi_free(stru* error. * */ -static int netlbl_cipsov4_add_std (u32 doi,struct nlattr *msg,size_t msg_size) +static int netlbl_cipsov4_add_std (struct genl_info *info) {int ret_val = -EINVAL ; -int msg_len = msg_size; -u32 num_tags; -u32 num_lvls; -u32 num_cats;struct cipso_v4_doi *doi_def =NULL ; -u32 iter; -u32 tmp_val_a; -u32 tmp_val_b; - -if (msg_len <NETLBL_LEN_U32 ) - goto add_std_failure; - num_tags =netlbl_getinc_u32 (&msg, &msg_len); -if (num_tags ==0 || num_tags >CIPSO_V4_TAG_MAXCNT ) +struct nlattr *nla_a; +struct nlattr *nla_b; +struct nlattr *nla_c; +int nla_a_rem; +int nla_b_rem; +u32 iter =0 ; + +if (!info->attrs[NLBL_CIPSOV4_A_DOI ] || + !info->attrs[NLBL_CIPSOV4_A_TAGLST ] || + !info->attrs[NLBL_CIPSOV4_A_MLSLVLLST ]) + goto add_std_failure; + +if (nla_validate (nla_data (info->attrs[NLBL_CIPSOV4_A_TAGLST ]), +nla_len (info->attrs[NLBL_CIPSOV4_A_TAGLST ]), +NLBL_CIPSOV4_A_MAX , + netlbl_cipsov4_genl_policy) !=0 ) + goto add_std_failure; +if (nla_validate (nla_data (info->attrs[NLBL_CIPSOV4_A_MLSLVLLST ]), +nla_len (info->attrs[NLBL_CIPSOV4_A_MLSLVLLST ]), +NLBL_CIPSOV4_A_MAX , + netlbl_cipsov4_genl_policy) !=0 ) goto add_std_failure; doi_def =kmalloc (sizeof (*doi_def),GFP_KERNEL );@@ -128,28 +156,44 @@ static int netlbl_cipsov4_add_std(u32 do } doi_def->type =CIPSO_V4_MAP_STD ; -for (iter =0 ; iter < num_tags; iter++) { -if (msg_len <NETLBL_LEN_U8 ) - goto add_std_failure;
net/netlabel/netlabel_cipso_v4.c static int netlbl_cipsov4_add_std(u32 do · line 162
Quoted 10 lines
net/netlabel/netlabel_cipso_v4.c +4
-6- doi_def->tags[iter] =netlbl_getinc_u8 (&msg, &msg_len); -switch (doi_def->tags[iter]) { -case CIPSO_V4_TAG_RBITMAP : -break ; -default : - goto add_std_failure; +nla_for_each_attr (nla_a, +nla_data (info->attrs[NLBL_CIPSOV4_A_TAGLST ]), +nla_len (info->attrs[NLBL_CIPSOV4_A_TAGLST ]), + nla_a_rem)
You can use nla_for_each_nested() here.
net/netlabel/netlabel_cipso_v4.c static int netlbl_cipsov4_add_std(u32 do · line 167
Quoted 5 lines
net/netlabel/netlabel_cipso_v4.c +4
-0+if (nla_a->nla_type ==NLBL_CIPSOV4_A_TAG ) { +if (iter >CIPSO_V4_TAG_MAXCNT ) + goto add_std_failure; + doi_def->tags[iter++] =nla_get_u8 (nla_a); }
Perfectly validated against generic policy yet flexible, I like this a lot.
net/netlabel/netlabel_cipso_v4.c static int netlbl_cipsov4_add_std(u32 do · line 174
Quoted 4 lines
net/netlabel/netlabel_cipso_v4.c +4
-0+nla_for_each_attr (nla_a, +nla_data (info->attrs[NLBL_CIPSOV4_A_MLSLVLLST ]), +nla_len (info->attrs[NLBL_CIPSOV4_A_MLSLVLLST ]), + nla_a_rem)
Could use nla_for_each_nested() here as well.
net/netlabel/netlabel_cipso_v4.c static int netlbl_cipsov4_add_std(u32 do · line 179
Quoted 5 lines
net/netlabel/netlabel_cipso_v4.c +5
-0+if (nla_a->nla_type ==NLBL_CIPSOV4_A_MLSLVL ) { +nla_for_each_attr (nla_b, +nla_data (nla_a), +nla_len (nla_a), + nla_b_rem)
... and again :-)
net/netlabel/netlabel_cipso_v4.c static int netlbl_cipsov4_add_std(u32 do · line 182
Quoted 3 lines
net/netlabel/netlabel_cipso_v4.c +3
-0+switch (nla_b->nla_type) { +case NLBL_CIPSOV4_A_MLSLVLLOC : + if (nla_get_u32(nla_b) >=
Attributes on this level are not yet validated.
Quoted 16 lines
net/netlabel/netlabel_cipso_v4.c +12
-0+ doi_def->map.std->lvl.local_size) + doi_def->map.std->lvl.local_size = +nla_get_u32 (nla_b) +1 ; +break ; +case NLBL_CIPSOV4_A_MLSLVLREM : +if (nla_get_u32 (nla_b) >= + doi_def->map.std->lvl.cipso_size) + doi_def->map.std->lvl.cipso_size = +nla_get_u32 (nla_b) +1 ; +break ; + } + }@@ -168,68 +209,96 @@ static int netlbl_cipsov4_add_std(u32 do ret_val = -ENOMEM ; goto add_std_failure; }
net/netlabel/netlabel_cipso_v4.c static int netlbl_cipsov4_add_std(u32 do · line 221
Quoted 10 lines
net/netlabel/netlabel_cipso_v4.c +9
-0+nla_for_each_attr (nla_a, +nla_data (info->attrs[NLBL_CIPSOV4_A_MLSLVLLST ]), +nla_len (info->attrs[NLBL_CIPSOV4_A_MLSLVLLST ]), + nla_a_rem) +if (nla_a->nla_type ==NLBL_CIPSOV4_A_MLSLVL ) { + nla_b =nla_find_nested (nla_a, +NLBL_CIPSOV4_A_MLSLVLLOC ); + nla_c =nla_find_nested (nla_a, +NLBL_CIPSOV4_A_MLSLVLREM );
Maybe find better names for nla_a and nla_b, would enhance readability
Quoted 9 lines
net/netlabel/netlabel_cipso_v4.c +7
-1+if (nla_b ==NULL || nla_c ==NULL ) + goto add_std_failure; + doi_def->map.std->lvl.local[nla_get_u32 (nla_b)] = +nla_get_u32 (nla_c); + doi_def->map.std->lvl.cipso[nla_get_u32 (nla_c)] = +nla_get_u32 (nla_b); + } - num_cats =netlbl_getinc_u32 (&msg, &msg_len);
net/netlabel/netlabel_cipso_v4.c static int netlbl_cipsov4_add_std(u32 do · line 234
Quoted 10 lines
net/netlabel/netlabel_cipso_v4.c +6
-4- doi_def->map.std->cat.local_size =netlbl_getinc_u32 (&msg, &msg_len); -if (doi_def->map.std->cat.local_size >CIPSO_V4_MAX_LOC_CATS ) - goto add_std_failure; - doi_def->map.std->cat.local = kcalloc(doi_def->map.std->cat.local_size, +if (info->attrs[NLBL_CIPSOV4_A_MLSCATLST ]) { +if (nla_validate ( +nla_data (info->attrs[NLBL_CIPSOV4_A_MLSCATLST ]), +nla_len (info->attrs[NLBL_CIPSOV4_A_MLSCATLST ]), +NLBL_CIPSOV4_A_MAX , + netlbl_cipsov4_genl_policy) !=0 )
We might want to add nla_validate_nested() to the interface as this seems to become a common usage.
Quoted 12 lines
net/netlabel/netlabel_cipso_v4.c +0
-8@@ -276,21 +345,20 @@ static int netlbl_cipsov4_add_pass(u32 d } doi_def->type =CIPSO_V4_MAP_PASS ; -for (iter =0 ; iter < num_tags; iter++) { -if (msg_len <NETLBL_LEN_U8 ) - goto add_pass_failure; - doi_def->tags[iter] =netlbl_getinc_u8 (&msg, &msg_len); -switch (doi_def->tags[iter]) { -case CIPSO_V4_TAG_RBITMAP : -break ; -default :
net/netlabel/netlabel_cipso_v4.c static int netlbl_cipsov4_add_pass(u32 d · line 356
Quoted 10 lines
net/netlabel/netlabel_cipso_v4.c +8
-1- goto add_pass_failure; +nla_for_each_attr (nla, +nla_data (info->attrs[NLBL_CIPSOV4_A_TAGLST ]), +nla_len (info->attrs[NLBL_CIPSOV4_A_TAGLST ]), + nla_rem) +if (nla->nla_type ==NLBL_CIPSOV4_A_TAG ) { +if (iter >CIPSO_V4_TAG_MAXCNT ) + goto add_pass_failure; + doi_def->tags[iter++] =nla_get_u8 (nla); }
This is duplicated code, put this into netlbl_cipsov4_get_tags()?
Quoted 96 lines
net/netlabel/netlabel_cipso_v4.c +65
-14@@ -353,84 +408,237 @@ add_return:* @info: the Generic NETLINK info block * * Description: -* Process a user generated LIST message and respond accordingly. Returns -* zero on success and negative values on error. +* Process a user generated LIST message and respond accordingly. While the +* response message generated by the kernel is straightforward, determining +* before hand the size of the buffer to allocate is not (we have to generate +* the message to know the size). In order to keep this function sane what we +* do is allocate a buffer of NLMSG_GOODSIZE and try to fit the response in +* that size, if we fail then we restart with a larger buffer and try again. +* We continue in this manner until we hit a limit of failed attempts then we +* give up and just send an error message. Returns zero on success and +* negative values on error. * */ static int netlbl_cipsov4_list (struct sk_buff *skb,struct genl_info *info) { -int ret_val = -EINVAL ; +int ret_val; +struct sk_buff *ans_skb =NULL ; +u32 nlsze_mult =1 ; +void *data;u32 doi; -struct nlattr *msg =netlbl_netlink_payload_data (skb); -struct sk_buff *ans_skb; +struct nlattr *nla_a; +struct nlattr *nla_b; +struct cipso_v4_doi *doi_def; +u32 iter; -if (netlbl_netlink_payload_len (skb) !=NETLBL_LEN_U32 ) +if (!info->attrs[NLBL_CIPSOV4_A_DOI ]) { + ret_val = -EINVAL ; goto list_failure; + } - doi =nla_get_u32 (msg); - ans_skb =cipso_v4_doi_dump (doi,NLMSG_SPACE (GENL_HDRLEN )); +list_start: + ans_skb =nlmsg_new (NLMSG_GOODSIZE * nlsze_mult,GFP_KERNEL );if (ans_skb ==NULL ) { ret_val = -ENOMEM ; goto list_failure; } -netlbl_netlink_hdr_push (ans_skb, - info->snd_pid, -0 , - netlbl_cipsov4_gnl_family.id, -NLBL_CIPSOV4_C_LIST ); + data =netlbl_netlink_hdr_put (ans_skb, + info->snd_pid, + info->snd_seq, + netlbl_cipsov4_gnl_family.id, +0 , +NLBL_CIPSOV4_C_LIST ); +if (data ==NULL ) { + ret_val = -ENOMEM ; + goto list_failure; + } + + doi =nla_get_u32 (info->attrs[NLBL_CIPSOV4_A_DOI ]); + +rcu_read_lock (); + doi_def =cipso_v4_doi_getdef (doi); +if (doi_def ==NULL ) { + ret_val = -EINVAL ; + goto list_failure; + } + + ret_val =nla_put_u32 (ans_skb,NLBL_CIPSOV4_A_MTYPE , doi_def->type); +if (ret_val !=0 ) + goto list_failure_lock; + + nla_a =nla_nest_start (ans_skb,NLBL_CIPSOV4_A_TAGLST ); +if (nla_a ==NULL ) { + ret_val = -ENOMEM ; + goto list_failure_lock; + } +for (iter =0 ; + iter <CIPSO_V4_TAG_MAXCNT && + doi_def->tags[iter] !=CIPSO_V4_TAG_INVALID ; + iter++) { + ret_val =nla_put_u8 (ans_skb, +NLBL_CIPSOV4_A_TAG , + doi_def->tags[iter]); +if (ret_val !=0 ) + goto list_failure_lock; + } +nla_nest_end (ans_skb, nla_a); - ret_val =netlbl_netlink_snd (ans_skb, info->snd_pid); +switch (doi_def->type) { +case CIPSO_V4_MAP_STD : + nla_a =nla_nest_start (ans_skb,NLBL_CIPSOV4_A_MLSLVLLST );
net/netlabel/netlabel_cipso_v4.c add_return: · line 498
Quoted 10 lines
net/netlabel/netlabel_cipso_v4.c +10
-0+if (nla_a ==NULL ) { + ret_val = -ENOMEM ; + goto list_failure_lock; + } +for (iter =0 ; + iter < doi_def->map.std->lvl.local_size; + iter++) { +if (doi_def->map.std->lvl.local[iter] == +CIPSO_V4_INV_LVL ) +continue ;
Can you estimate the number of entries being dumped here and in the cat list below?
Quoted 44 lines
net/netlabel/netlabel_cipso_v4.c +44
-0+ + nla_b =nla_nest_start (ans_skb,NLBL_CIPSOV4_A_MLSLVL ); +if (nla_b ==NULL ) { + ret_val = -ENOMEM ; + goto list_retry; + } + ret_val =nla_put_u32 (ans_skb, +NLBL_CIPSOV4_A_MLSLVLLOC , + iter); +if (ret_val !=0 ) + goto list_retry; + ret_val =nla_put_u32 (ans_skb, +NLBL_CIPSOV4_A_MLSLVLREM , + doi_def->map.std->lvl.local[iter]); +if (ret_val !=0 ) + goto list_retry; +nla_nest_end (ans_skb, nla_b); + } +nla_nest_end (ans_skb, nla_a); + + nla_a =nla_nest_start (ans_skb,NLBL_CIPSOV4_A_MLSCATLST ); +if (nla_a ==NULL ) { + ret_val = -ENOMEM ; + goto list_retry; + } +for (iter =0 ; + iter < doi_def->map.std->cat.local_size; + iter++) { +if (doi_def->map.std->cat.local[iter] == +CIPSO_V4_INV_CAT ) +continue ; + + nla_b =nla_nest_start (ans_skb,NLBL_CIPSOV4_A_MLSCAT ); +if (nla_b ==NULL ) { + ret_val = -ENOMEM ; + goto list_retry; + } + ret_val =nla_put_u32 (ans_skb, +NLBL_CIPSOV4_A_MLSCATLOC , + iter); +if (ret_val !=0 ) + goto list_retry; + ret_val = nla_put_u32(ans_skb, + NLBL_CIPSOV4_A_MLSCATREM,
net/netlabel/netlabel_cipso_v4.c add_return: · line 552
Quoted 10 lines
net/netlabel/netlabel_cipso_v4.c +10
-0+ doi_def->map.std->cat.local[iter]); +if (ret_val !=0 ) + goto list_retry; +nla_nest_end (ans_skb, nla_b); + } +nla_nest_end (ans_skb, nla_a); + +break ; + } +rcu_read_unlock ();
What I meant in the previous posting is to put genlmsg_end() in here like the rest of the code to be clear and symmetric.
The rest of the code has simliar issues, nevertheless this looks really great. Everything is extendable and relatively easy to maintain. This should definitely go into 2.6.19. Good work.
Re: [PATCH 5/6] NetLabel: rework the Netlink attribute handling (part 2)
Quoted 21 lines · Thomas Graf
* [email protected] <[email protected]> 2006-09-21 12:57
Quoted 19 lines
At the suggestion of Thomas Graf, rewrite NetLabel's use of Netlink attributes to better follow the common Netlink attribute usage.
Signed-off-by Paul Moore <[email protected]>---
net/netlabel/netlabel_cipso_v4.c | 593 +++++++++++++++++++++++++------------- net/netlabel/netlabel_cipso_v4.h | 235 +++++---------- net/netlabel/netlabel_mgmt.c | 531 +++++++++++++++++----------------- net/netlabel/netlabel_mgmt.h | 213 ++++--------- net/netlabel/netlabel_unlabeled.c | 77 ++-- net/netlabel/netlabel_unlabeled.h | 41 +- 6 files changed, 875 insertions(+), 815 deletions(-){snip}
net/netlabel/netlabel_cipso_v4.c +0
-0@@ -128,28 +156,44 @@ static int netlbl_cipsov4_add_std(u32 do } doi_def->type =CIPSO_V4_MAP_STD ;
Quoted 16 lines
net/netlabel/netlabel_cipso_v4.c static int netlbl_cipsov4_add_std(u32 do · line 162
Quoted 13 lines
net/netlabel/netlabel_cipso_v4.c +4
-9-for (iter =0 ; iter < num_tags; iter++) { -if (msg_len <NETLBL_LEN_U8 ) - goto add_std_failure; - doi_def->tags[iter] =netlbl_getinc_u8 (&msg, &msg_len); -switch (doi_def->tags[iter]) { -case CIPSO_V4_TAG_RBITMAP : -break ; -default : - goto add_std_failure; +nla_for_each_attr (nla_a, +nla_data (info->attrs[NLBL_CIPSOV4_A_TAGLST ]), +nla_len (info->attrs[NLBL_CIPSOV4_A_TAGLST ]), + nla_a_rem)You can use
nla_for_each_nested()here.
Huh, look at that, guess I didn't scroll down far enough in the header file :) Thanks.
Quoted 31 lines
net/netlabel/netlabel_cipso_v4.c static int netlbl_cipsov4_add_std(u32 do · line 167
Quoted 5 lines
net/netlabel/netlabel_cipso_v4.c +4
-0+if (nla_a->nla_type ==NLBL_CIPSOV4_A_TAG ) { +if (iter >CIPSO_V4_TAG_MAXCNT ) + goto add_std_failure; + doi_def->tags[iter++] =nla_get_u8 (nla_a); }Perfectly validated against generic policy yet flexible, I like this a lot.
net/netlabel/netlabel_cipso_v4.c static int netlbl_cipsov4_add_std(u32 do · line 174
Quoted 4 lines
net/netlabel/netlabel_cipso_v4.c +4
-0+nla_for_each_attr (nla_a, +nla_data (info->attrs[NLBL_CIPSOV4_A_MLSLVLLST ]), +nla_len (info->attrs[NLBL_CIPSOV4_A_MLSLVLLST ]), + nla_a_rem)Could use
nla_for_each_nested()here as well.net/netlabel/netlabel_cipso_v4.c static int netlbl_cipsov4_add_std(u32 do · line 179
Quoted 5 lines
net/netlabel/netlabel_cipso_v4.c +5
-0+if (nla_a->nla_type ==NLBL_CIPSOV4_A_MLSLVL ) { +nla_for_each_attr (nla_b, +nla_data (nla_a), +nla_len (nla_a), + nla_b_rem)... and again :-)
Quoted 7 lines
net/netlabel/netlabel_cipso_v4.c static int netlbl_cipsov4_add_std(u32 do · line 182
Quoted 3 lines
net/netlabel/netlabel_cipso_v4.c +3
-0+switch (nla_b->nla_type) { +case NLBL_CIPSOV4_A_MLSLVLLOC : + if (nla_get_u32(nla_b) >=Attributes on this level are not yet validated.
Yes, you're right, same problem with the _MLSCAT attribute. Thanks.
Quoted 22 lines
Quoted 20 lines
net/netlabel/netlabel_cipso_v4.c +16
-0+ doi_def->map.std->lvl.local_size) + doi_def->map.std->lvl.local_size = +nla_get_u32 (nla_b) +1 ; +break ; +case NLBL_CIPSOV4_A_MLSLVLREM : +if (nla_get_u32 (nla_b) >= + doi_def->map.std->lvl.cipso_size) + doi_def->map.std->lvl.cipso_size = +nla_get_u32 (nla_b) +1 ; +break ; + } + }@@ -168,68 +209,96 @@ static int netlbl_cipsov4_add_std(u32 do ret_val = -ENOMEM ; goto add_std_failure; } +nla_for_each_attr (nla_a, +nla_data (info->attrs[NLBL_CIPSOV4_A_MLSLVLLST ]), +nla_len (info->attrs[NLBL_CIPSOV4_A_MLSLVLLST ]), + nla_a_rem)
Quoted 9 lines
net/netlabel/netlabel_cipso_v4.c static int netlbl_cipsov4_add_std(u32 do · line 220
Quoted 5 lines
net/netlabel/netlabel_cipso_v4.c +5
-0+if (nla_a->nla_type ==NLBL_CIPSOV4_A_MLSLVL ) { + nla_b =nla_find_nested (nla_a, +NLBL_CIPSOV4_A_MLSLVLLOC ); + nla_c =nla_find_nested (nla_a, +NLBL_CIPSOV4_A_MLSLVLREM );Maybe find better names for
nla_aandnla_b, would enhance readability
Fair enough.
Quoted 8 lines
Quoted 8 lines
net/netlabel/netlabel_cipso_v4.c +7
-0+if (nla_b ==NULL || nla_c ==NULL ) + goto add_std_failure; + doi_def->map.std->lvl.local[nla_get_u32 (nla_b)] = +nla_get_u32 (nla_c); + doi_def->map.std->lvl.cipso[nla_get_u32 (nla_c)] = +nla_get_u32 (nla_b); + }
Quoted 14 lines
net/netlabel/netlabel_cipso_v4.c static int netlbl_cipsov4_add_std(u32 do · line 234
Quoted 11 lines
net/netlabel/netlabel_cipso_v4.c +6
-5- num_cats =netlbl_getinc_u32 (&msg, &msg_len); - doi_def->map.std->cat.local_size =netlbl_getinc_u32 (&msg, &msg_len); -if (doi_def->map.std->cat.local_size >CIPSO_V4_MAX_LOC_CATS ) - goto add_std_failure; - doi_def->map.std->cat.local = kcalloc(doi_def->map.std->cat.local_size, +if (info->attrs[NLBL_CIPSOV4_A_MLSCATLST ]) { +if (nla_validate ( +nla_data (info->attrs[NLBL_CIPSOV4_A_MLSCATLST ]), +nla_len (info->attrs[NLBL_CIPSOV4_A_MLSCATLST ]), +NLBL_CIPSOV4_A_MAX , + netlbl_cipsov4_genl_policy) !=0 )We might want to add
nla_validate_nested()to the interface as this seems to become a common usage.
Easy enough. I'll make that a separate patch in the next patchset.
Quoted 16 lines
Quoted 16 lines
net/netlabel/netlabel_cipso_v4.c +3
-9@@ -276,21 +345,20 @@ static int netlbl_cipsov4_add_pass(u32 d } doi_def->type =CIPSO_V4_MAP_PASS ; -for (iter =0 ; iter < num_tags; iter++) { -if (msg_len <NETLBL_LEN_U8 ) - goto add_pass_failure; - doi_def->tags[iter] =netlbl_getinc_u8 (&msg, &msg_len); -switch (doi_def->tags[iter]) { -case CIPSO_V4_TAG_RBITMAP : -break ; -default : - goto add_pass_failure; +nla_for_each_attr (nla, +nla_data (info->attrs[NLBL_CIPSOV4_A_TAGLST ]), + nla_len(info->attrs[NLBL_CIPSOV4_A_TAGLST]),
Quoted 9 lines
net/netlabel/netlabel_cipso_v4.c static int netlbl_cipsov4_add_pass(u32 d · line 356
Quoted 6 lines
net/netlabel/netlabel_cipso_v4.c +5
-0+ nla_rem) +if (nla->nla_type ==NLBL_CIPSOV4_A_TAG ) { +if (iter >CIPSO_V4_TAG_MAXCNT ) + goto add_pass_failure; + doi_def->tags[iter++] =nla_get_u8 (nla); }This is duplicated code, put this into
netlbl_cipsov4_get_tags()?
There is actually a little bit more than that too, I see your point. I'll make a netlbl_cipsov4_add_common().
Quoted 102 lines
Quoted 102 lines
net/netlabel/netlabel_cipso_v4.c +71
-14@@ -353,84 +408,237 @@ add_return:* @info: the Generic NETLINK info block * * Description: -* Process a user generated LIST message and respond accordingly. Returns -* zero on success and negative values on error. +* Process a user generated LIST message and respond accordingly. While the +* response message generated by the kernel is straightforward, determining +* before hand the size of the buffer to allocate is not (we have to generate +* the message to know the size). In order to keep this function sane what we +* do is allocate a buffer of NLMSG_GOODSIZE and try to fit the response in +* that size, if we fail then we restart with a larger buffer and try again. +* We continue in this manner until we hit a limit of failed attempts then we +* give up and just send an error message. Returns zero on success and +* negative values on error. * */ static int netlbl_cipsov4_list (struct sk_buff *skb,struct genl_info *info) { -int ret_val = -EINVAL ; +int ret_val; +struct sk_buff *ans_skb =NULL ; +u32 nlsze_mult =1 ; +void *data;u32 doi; -struct nlattr *msg =netlbl_netlink_payload_data (skb); -struct sk_buff *ans_skb; +struct nlattr *nla_a; +struct nlattr *nla_b; +struct cipso_v4_doi *doi_def; +u32 iter; -if (netlbl_netlink_payload_len (skb) !=NETLBL_LEN_U32 ) +if (!info->attrs[NLBL_CIPSOV4_A_DOI ]) { + ret_val = -EINVAL ; goto list_failure; + } - doi =nla_get_u32 (msg); - ans_skb =cipso_v4_doi_dump (doi,NLMSG_SPACE (GENL_HDRLEN )); +list_start: + ans_skb =nlmsg_new (NLMSG_GOODSIZE * nlsze_mult,GFP_KERNEL );if (ans_skb ==NULL ) { ret_val = -ENOMEM ; goto list_failure; } -netlbl_netlink_hdr_push (ans_skb, - info->snd_pid, -0 , - netlbl_cipsov4_gnl_family.id, -NLBL_CIPSOV4_C_LIST ); + data =netlbl_netlink_hdr_put (ans_skb, + info->snd_pid, + info->snd_seq, + netlbl_cipsov4_gnl_family.id, +0 , +NLBL_CIPSOV4_C_LIST ); +if (data ==NULL ) { + ret_val = -ENOMEM ; + goto list_failure; + } + + doi =nla_get_u32 (info->attrs[NLBL_CIPSOV4_A_DOI ]); + +rcu_read_lock (); + doi_def =cipso_v4_doi_getdef (doi); +if (doi_def ==NULL ) { + ret_val = -EINVAL ; + goto list_failure; + } + + ret_val =nla_put_u32 (ans_skb,NLBL_CIPSOV4_A_MTYPE , doi_def->type); +if (ret_val !=0 ) + goto list_failure_lock; + + nla_a =nla_nest_start (ans_skb,NLBL_CIPSOV4_A_TAGLST ); +if (nla_a ==NULL ) { + ret_val = -ENOMEM ; + goto list_failure_lock; + } +for (iter =0 ; + iter <CIPSO_V4_TAG_MAXCNT && + doi_def->tags[iter] !=CIPSO_V4_TAG_INVALID ; + iter++) { + ret_val =nla_put_u8 (ans_skb, +NLBL_CIPSOV4_A_TAG , + doi_def->tags[iter]); +if (ret_val !=0 ) + goto list_failure_lock; + } +nla_nest_end (ans_skb, nla_a); - ret_val =netlbl_netlink_snd (ans_skb, info->snd_pid); +switch (doi_def->type) { +case CIPSO_V4_MAP_STD : + nla_a =nla_nest_start (ans_skb,NLBL_CIPSOV4_A_MLSLVLLST ); +if (nla_a ==NULL ) { + ret_val = -ENOMEM ; + goto list_failure_lock; + } + for (iter = 0; + iter < doi_def->map.std->lvl.local_size;
Quoted 9 lines
net/netlabel/netlabel_cipso_v4.c add_return: · line 498
Quoted 4 lines
net/netlabel/netlabel_cipso_v4.c +4
-0+ iter++) { +if (doi_def->map.std->lvl.local[iter] == +CIPSO_V4_INV_LVL ) +continue ;Can you estimate the number of entries being dumped here and in the cat list below?
It's too hard to come up with a reasonable estimate without going through the entire list before hand, which in previous messages (might of been off-list) you pointed out as a bad thing. If you would prefer I can go back to doing it that way?
Quoted 49 lines
Quoted 49 lines
net/netlabel/netlabel_cipso_v4.c +49
-0+ + nla_b =nla_nest_start (ans_skb,NLBL_CIPSOV4_A_MLSLVL ); +if (nla_b ==NULL ) { + ret_val = -ENOMEM ; + goto list_retry; + } + ret_val =nla_put_u32 (ans_skb, +NLBL_CIPSOV4_A_MLSLVLLOC , + iter); +if (ret_val !=0 ) + goto list_retry; + ret_val =nla_put_u32 (ans_skb, +NLBL_CIPSOV4_A_MLSLVLREM , + doi_def->map.std->lvl.local[iter]); +if (ret_val !=0 ) + goto list_retry; +nla_nest_end (ans_skb, nla_b); + } +nla_nest_end (ans_skb, nla_a); + + nla_a =nla_nest_start (ans_skb,NLBL_CIPSOV4_A_MLSCATLST ); +if (nla_a ==NULL ) { + ret_val = -ENOMEM ; + goto list_retry; + } +for (iter =0 ; + iter < doi_def->map.std->cat.local_size; + iter++) { +if (doi_def->map.std->cat.local[iter] == +CIPSO_V4_INV_CAT ) +continue ; + + nla_b =nla_nest_start (ans_skb,NLBL_CIPSOV4_A_MLSCAT ); +if (nla_b ==NULL ) { + ret_val = -ENOMEM ; + goto list_retry; + } + ret_val =nla_put_u32 (ans_skb, +NLBL_CIPSOV4_A_MLSCATLOC , + iter); +if (ret_val !=0 ) + goto list_retry; + ret_val =nla_put_u32 (ans_skb, +NLBL_CIPSOV4_A_MLSCATREM , + doi_def->map.std->cat.local[iter]); +if (ret_val !=0 ) + goto list_retry; +nla_nest_end (ans_skb, nla_b); + }
Quoted 8 lines
net/netlabel/netlabel_cipso_v4.c add_return: · line 552
Quoted 5 lines
net/netlabel/netlabel_cipso_v4.c +5
-0+nla_nest_end (ans_skb, nla_a); + +break ; + } +rcu_read_unlock ();What I meant in the previous posting is to put
genlmsg_end()in here like the rest of the code to be clear and symmetric.
Okay, I put it there because it was a common trend in the code, i.e. end the message then do a write, I thought I would combine them. I'll split it back out.
Quoted 3 lines
The rest of the code has simliar issues, nevertheless this looks really great. Everything is extendable and relatively easy to maintain. This should definitely go into 2.6.19. Good work.
Thanks for your patience as I got the hang of the genetlink stuff. I'll make the changes you suggested and get a new version out soon.
Signature · 2 lines
paul moore linux security @ hp
Re: [PATCH 5/6] NetLabel: rework the Netlink attribute handling (part 2)
* Paul Moore <[email protected]> 2006-09-25 10:13
Quoted 20 lines
Quoted 10 lines
Quoted 10 lines
net/netlabel/netlabel_cipso_v4.c +9
-1- ret_val =netlbl_netlink_snd (ans_skb, info->snd_pid); +switch (doi_def->type) { +case CIPSO_V4_MAP_STD : + nla_a =nla_nest_start (ans_skb,NLBL_CIPSOV4_A_MLSLVLLST ); +if (nla_a ==NULL ) { + ret_val = -ENOMEM ; + goto list_failure_lock; + } + for (iter = 0; + iter < doi_def->map.std->lvl.local_size;Quoted 9 lines
net/netlabel/netlabel_cipso_v4.c add_return: · line 498
Quoted 4 lines
net/netlabel/netlabel_cipso_v4.c +4
-0+ iter++) { +if (doi_def->map.std->lvl.local[iter] == +CIPSO_V4_INV_LVL ) +continue ;Can you estimate the number of entries being dumped here and in the cat list below?
Quoted 4 lines
It's too hard to come up with a reasonable estimate without going through the entire list before hand, which in previous messages (might of been off-list) you pointed out as a bad thing. If you would prefer I can go back to doing it that way?
Sorry, I didn't make myself clear. I didn't mean to estimate the size of the message. The way you're doing it right now is perfectly fine. I'm interested in a estimate on how many items are being dumped in practice. Less than 1K or a couple of K?
High order allocations are likely to fail if under pressure while using dumpit() allows to reuse memory ressources.
Re: [PATCH 5/6] NetLabel: rework the Netlink attribute handling (part 2)
Quoted 27 lines · Thomas Graf
* Paul Moore <[email protected]> 2006-09-25 10:13
Quoted 20 lines
Quoted 10 lines
Quoted 10 lines
net/netlabel/netlabel_cipso_v4.c +9
-1- ret_val =netlbl_netlink_snd (ans_skb, info->snd_pid); +switch (doi_def->type) { +case CIPSO_V4_MAP_STD : + nla_a =nla_nest_start (ans_skb,NLBL_CIPSOV4_A_MLSLVLLST ); +if (nla_a ==NULL ) { + ret_val = -ENOMEM ; + goto list_failure_lock; + } + for (iter = 0; + iter < doi_def->map.std->lvl.local_size;Quoted 9 lines
net/netlabel/netlabel_cipso_v4.c add_return: · line 498
Quoted 4 lines
net/netlabel/netlabel_cipso_v4.c +4
-0+ iter++) { +if (doi_def->map.std->lvl.local[iter] == +CIPSO_V4_INV_LVL ) +continue ;Can you estimate the number of entries being dumped here and in the cat list below?
Quoted 4 lines
It's too hard to come up with a reasonable estimate without going through the entire list before hand, which in previous messages (might of been off-list) you pointed out as a bad thing. If you would prefer I can go back to doing it that way?
Quoted 7 lines
Sorry, I didn't make myself clear. I didn't mean to estimate the size of the message. The way you're doing it right now is perfectly fine. I'm interested in a estimate on how many items are being dumped in practice. Less than 1K or a couple of K?
High order allocations are likely to fail if under pressure while using
dumpit()allows to reuse memory ressources.
It's hard to get a good idea of the likely usage scenario as I'm not sure how users will want to configure NetLabel. However, there is a possibility of this growing larger than NLMSG_GOODSIZE, which makes things a bit more difficult. From what I can see dumpit() requires the message be within NLMSG_GOODSIZE while a doit() response can be of arbitrary length; this is why I chose a doit() response. Even if the system is under memory pressure I think a failure here is okay as this is not what I would consider a critical message.
Signature · 2 lines
paul moore linux security @ hp
PATCH 6/6 NetLabel: update docs with website information
In reply to [email protected]Now that all of the supporting pieces of NetLabel have a home at SourceForge update the Kconfig help text and add an entry to the MAINTAINERS file.
--- MAINTAINERS | 7 +++++++ net/netlabel/Kconfig | 5 ++++- 2 files changed, 11 insertions(+), 1 deletion(-)
MAINTAINERS +7 -0
@@ -2008,6 +2008,13 @@ L: [email protected] L: [email protected] S: Supported +NETLABEL +P: Paul Moore +M: [email protected] +W: http://netlabel.sf.net +L: [email protected] +S: Supported + NETROM NETWORK LAYER P: Ralf Baechle M: [email protected]
net/netlabel/Kconfig +4 -1
@@ -9,6 +9,9 @@ config NETLABEL ---help--- NetLabel provides support for explicit network packet labeling protocols such as CIPSO and RIPSO. For more information see - Documentation/netlabel. + Documentation/netlabel as well as the NetLabel SourceForge project + for configuration tools and additional documentation. + + * http://netlabel.sf.net If you are unsure, say N.
--
paul moore
linux security @ hp