LSM: Module stacking for AppArmor
Subsystems (31)
A message that is not in the archive (sent off-list or not yet received): [email protected]
Casey Schaufler
maintainerv26 cover LSM: Module stacking for AppArmor
This patchset provides the changes required for the AppArmor security module to stack safely with any other.
v26: Rebase to 5.13-rc1
Include the landlock security module.
Accomodate change from security_task_getsecid() to security_task_getsecid_obj() and security_task_getsecid_subj().
v25: Rebase to 5.12-rc2
Incorporate feedback from v24
- The IMA team suggested improvements to the integrity rule processing.
v24: Rebase to 5.11-rc1
Incorporate feedback from v23
- Address the IMA team's concerns about "label collisions". A label collision occurs when there is ambiguity about which of multiple LSMs is being targeted in the definition of an integrity check rule. A system with Smack and AppArmor would be unable to distinguish which LSM is important to an integrity rule referrencing the label "unconfined" as that label is meaningful to both. Provide a boot option to specify which LSM will be used in IMA rules when multiple LSMs are present. (patch 04) Pull LSM "slot" identification from later audit patches in in support of this (patch 03).
- Pick up a few audit events that need to include supplimental subject context records that had been missed in the previous version.
v23: Rebase to 5.10-rc4
Incorporate feedback from v22
- Change /proc/*/attr/display to /proc/*/attr/interface_lsm to make the purpose clearer. (patch 0012)
- Include ABI documentation. (patch 0012, 0022)
- Introduce LSM documentation updates with the patches where the interfaces are added rather than at the end. (patch 0012, 0022)
Include more maintainers and mail lists in To: and Cc: directives.
v22: Rebase to 5.10-rc1
v21: Rebase to 5.9-rc4
Incorporate feedback from v20
- Further revert UDS SO_PEERSEC to use scaffolding around the interfaces that use lsmblobs and store only a single secid. The possibility of multiple security modules requiring data here is still a future problem.
- Incorporate Richard Guy Briggs' non-syscall auxiliary records patch (patch 0019-0021) in place of my "supplimental" records implementation. [I'm not sure I've given proper attestation. I will correct as appropriate]
v20: Rebase to 5.9-rc1
Change the BPF security module to use the lsmblob data. (patch 0002) Repair length logic in subject label processing (patch 0015) Handle -EINVAL from the empty BPF setprocattr hook (patch 0020) Correct length processing in append_ctx() (patch 0022)
v19: Rebase to 5.8-rc6
Incorporate feedback from v18
- Revert UDS SO_PEERSEC implementation to use lsmblobs directly, rather than allocating as needed. The correct treatment of out-of-memory conditions in the later case is difficult to define. (patch 0005)
- Use a size_t in append_ctx() (patch 0021)
- Fix a memory leak when creating compound contexts. (patch 0021) Fix build error when CONFIG_SECURITY isn't set (patch 0013) Fix build error when CONFIG_SECURITY isn't set (patch 0020) Fix build error when CONFIG_SECURITY isn't set (patch 0021)
v18: Rebase to 5.8-rc3
Incorporate feedback from v17
- Null pointer checking in UDS (patch 0005)
Match changes in IMA code (patch 0012)
Fix the behavior of LSM context supplimental audit
records so that there's always exactly one when it's appropriate for there to be one. This is a substantial change that requires extention of the audit_context beyond syscall events. (patch 0020)
v17: Rebase to 5.7-rc4
v16: Rebase to 5.6
Incorporate feedback from v15 - Thanks Stephen, Mimi and Paul
- Generally improve commit messages WRT scaffolding
- Comment ima_lsm_isset() (patch 0002)
- Some question may remain on IMA warning (patch 0002)
- Mark lsm_slot as __lsm_ro_after_init not __init_data (patch 0002)
- Change name of lsmblob variable in ima_match_rules() (patch 0003)
- Instead of putting a struct lsmblob into the unix_skb_parms structure put a pointer to an allocated instance. There is currently only space for 5 u32's in unix_skb_parms and it is likely to get even tighter. Fortunately, the lifecycle management of the allocated lsmblob is simple. (patch 0005)
- Dropped Acks due to the above change (patch 0005)
- Improved commentary on secmark labeling scaffolding. (patch 0006)
- Reduced secmark related labeling scaffolding. (patch 0006)
- Replace use of the zeroth entry of an lsmblob in scaffolding with a function lsmblob_value() to hopefully make it less obscure. (patch 0006)
- Convert security_secmark_relabel_packet to use lsmblob as this reduces much of the most contentious scaffolding. (patch 0006)
- Dropped Acks due to the above change (patch 0006)
- Added BUILD_BUG_ON() for CIPSO tag 6. (patch 0018)
- Reworked audit subject information. Instead of adding fields in the middle of existing records add a new record to the event. When a separate record is required use subj="?". (patch 0020)
- Dropped Acks due to the above change (patch 0020)
- Reworked audit object information. Instead of adding fields in the middle of existing records add a new record to the event. When a separate record is required use obj="?". (patch 0021)
- Dropped Acks due to the above change (patch 0021)
- Enhanced documentation (patch 0022)
- Removed unnecessary error code check in security_getprocattr() (patch 0021)
v15: Rebase to 5.6-rc1
- Revise IMA data use (patch 0002)
Incorporate feedback from v14
- Fix lockdown module registration naming (patch 0002)
- Revise how /proc/self/attr/context is gathered. (patch 0022)
- Revise access modes on /proc/self/attr/context. (patch 0022)
- Revise documentation on LSM external interfaces. (patch 0022)
v14: Rebase to 5.5-rc5
Incorporate feedback from v13
- Use an array of audit rules (patch 0002)
- Significant change, removed Acks (patch 0002)
- Remove unneeded include (patch 0013)
- Use context.len correctly (patch 0015)
- Reorder code to be more sensible (patch 0016)
- Drop SO_PEERCONTEXT as it's not needed yet (patch 0023)
v13: Rebase to 5.5-rc2
Incorporate feedback from v12
- Print lsmblob size with %z (Patch 0002)
- Convert lockdown LSM initialization. (Patch 0002)
- Restore error check in nft_secmark_compute_secid (Patch 0006)
- Correct blob scaffolding in ima_must_appraise() (Patch 0009)
- Make security_setprocattr() clearer (Patch 0013)
- Use lsm_task_display more widely (Patch 0013)
- Use passed size in lsmcontext_init() (Patch 0014)
- Don't add a smack_release_secctx() hook (Patch 0014)
- Don't print warning in security_release_secctx() (Patch 0014)
- Don't duplicate the label in nfs4_label_init_security() (Patch 0016)
- Remove reviewed-by as code has significant change (Patch 0016)
- Send the entire lsmblob for Tag 6 (Patch 0019)
- Fix description of socket_getpeersec_stream parameters (Patch 0023)
- Retain LSMBLOB_FIRST. What was I thinking? (Patch 0023)
- Add compound context to LSM documentation (Patch 0023)
v12: Rebase to 5.5-rc1
Fixed a couple of incorrect contractions in the text.
v11: Rebase to 5.4-rc6
Incorporate feedback from v10
- Disambiguate reading /proc/.../attr/display by restricting all use of the interface to the current process.
- Fix a merge error in AppArmor's display attribute check
v10: Ask the security modules if the display can be changed.
v9: There is no version 9
v8: Incorporate feedback from v7
- Minor clean-up in display value management
- refactor "compound" context creation to use a common append_ctx() function.
v7: Incorporate feedback from v6
- Make setting the display a privileged operation. The availability of compound contexts reduces the need for setting the display.
v6: Incorporate feedback from v5
- Add subj_<lsm>= and obj_<lsm>= fields to audit records
- Add /proc/.../attr/context to get the full context in lsmname\0value\0... format as suggested by Simon McVittie
- Add SO_PEERCONTEXT for getsockopt() to get the full context in the same format, also suggested by Simon McVittie.
- Add /sys/kernel/security/lsm_display_default to provide the display default value.
v5: Incorporate feedback from v4
- Initialize the lsmcontext in security_secid_to_secctx()
- Clear the lsmcontext in all security_release_secctx() cases
- Don't use the "display" on strictly internal context interfaces.
- The SELinux binder hooks check for cases where the context "display" isn't compatible with SELinux.
v4: Incorporate feedback from v3
- Mark new lsm_<blob>_alloc functions static
- Replace the lsm and slot fields of the security_hook_list with a pointer to a LSM allocated lsm_id structure. The LSM identifies if it needs a slot explicitly. Use the lsm_id rather than make security_add_hooks return the slot value.
- Validate slot values used in security.c
- Reworked the "display" process attribute handling so that it works right and doesn't use goofy list processing.
- fix display value check in dentry_init_security
- Replace audit_log of secids with '?' instead of deleting the audit log
v3: Incorporate feedback from v2
- Make lsmblob parameter and variable names more meaningful, changing "le" and "l" to "blob".
- Improve consistency of constant naming.
- Do more sanity checking during LSM initialization.
- Be a bit clearer about what is temporary scaffolding.
- Rather than clutter security_getpeersec_dgram with otherwise unnecessary checks remove the apparmor
stub, which does nothing useful.
Patch 01 moves management of the sock security blob from the individual modules to the infrastructure.
Patches 02-03 introduce a structure "lsmblob" that will gradually replace the "secid" as a shorthand for security module information. At this point lsmblob contains an array of u32 secids, one "slot" for each of the security modules compiled into the kernel that used secids. A "slot" is allocated when a security module requests one.
Patch 04 provides mechanism for the IMA subsystem to identify explicitly which LSM is subject to IMA policy. This includes a boot option for specifying the default and an additional option in IMA rules "lsm=".
Patches 05-13 change LSM interfaces to use the lsmblob instead of secids. It is important that the lsmblob be a fixed size entity that does not have to be allocated. Several of the places where it is used would have performance and/or locking issues with dynamic allocation.
Patch 14 provides a mechanism for a process to identify which security module's hooks should be used when displaying or converting a security context string. A new interface /proc/self/attr/interface_lsm contains the name of the security module to show. Reading from this file will present the name of the module, while writing to it will set the value. Only names of active security modules are accepted. Internally, the name is translated to the appropriate "slot" number for the module which is then stored in the task security blob. Setting the display requires that all modules using the /proc interfaces allow the transition. The interface LSM of other processess can be neither read nor written. All suggested cases for reading the interface LSM of a different process have race conditions.
Patch 15 Starts the process of changing how a security context is represented. Since it is possible for a security context to have been generated by more than one security module it is now necessary to note which module created a security context so that the correct "release" hook can be called. There are several places where the module that created a security context cannot be inferred.
This is achieved by introducing a "lsmcontext" structure which contains the context string, its length and the "slot" number of the security module that created it. The security_release_secctx() interface is changed, replacing the (string,len) pointer pair with a lsmcontext pointer.
Patches 16-18 convert the security interfaces from (string,len) pointer pairs to a lsmcontext pointer. The slot number identifying the creating module is added by the infrastructure. Where the security context is stored for extended periods the data type is changed.
The Netlabel code is converted to save lsmblob structures instead of secids in Patch 19. This is not strictly necessary as there can only be one security module that uses Netlabel at this point. Using a lsmblob is much cleaner, as the interfaces that use the data have all been converted.
Patch 20 adds checks to the binder hooks which verify that both ends of a transaction use the same interface LSM.
Patches 21-23 add addition audit records for subject and object LSM data when there are multiple security modules with such data. The AUDIT_MAC_TASK_CONTEXTS record is used in conjuction with a "subj=?" field to identify the subject data. The AUDIT_MAC_OBJ_CONTEXTS record is used in conjuction with a "obj=?" field to identify the object data. The AUDIT_MAC_TASK_CONTEXTS record identifies the security module with the data: "subj_selinux=xyz_t subj_apparmor=abc". The AUDIT_MAC_OBJ_CONTEXTS record identifies the security module with the data: "obj_selinux=xyz_t obj_apparmor=abc". While AUDIT_MAC_TASK_CONTEXTS records will always contain an entry for each possible security modules, AUDIT_MAC_OBJ_CONTEXTS records will only contain entries for security modules for which the object in question has data.
An example of the MAC_TASK_CONTEXTS (1420) record is:
type=UNKNOWN[1420]
msg=audit(1600880931.832:113)
subj_apparmor==unconfined
subj_smack=_An example of the MAC_OBJ_CONTEXTS (1421) record is:
type=UNKNOWN[1421]
msg=audit(1601152467.009:1050):
obj_selinux=unconfined_u:object_r:user_home_t:s0Patch 24 adds a new interface for getting the compound security contexts, /proc/self/attr/context. An example of the content of this file is:
selinux\0one_u:one_r:one_t:s0-s0:c0.c1023\0apparmor\0unconfined\0
Finally, with all interference on the AppArmor hooks removed, Patch 25 removes the exclusive bit from AppArmor. An unnecessary stub hook was also removed.
The Ubuntu project is using an earlier version of this patchset in their distribution to enable stacking for containers.
Performance measurements to date have the change within the "noise". The sockperf and dbench results are on the order of 0.2% to 0.8% difference, with better performance being as common as worse. The benchmarks were run with AppArmor and Smack on Ubuntu.
https://github.com/cschaufler/lsm-stacking.git#stack-5.13-rc1-v26
Casey Schaufler (25):
LSM: Infrastructure management of the sock security
LSM: Add the lsmblob data structure.
LSM: provide lsm name and id slot mappings
IMA: avoid label collisions with stacked LSMs
LSM: Use lsmblob in security_audit_rule_match
LSM: Use lsmblob in security_kernel_act_as
LSM: Use lsmblob in security_secctx_to_secid
LSM: Use lsmblob in security_secid_to_secctx
LSM: Use lsmblob in security_ipc_getsecid
LSM: Use lsmblob in security_task_getsecid
LSM: Use lsmblob in security_inode_getsecid
LSM: Use lsmblob in security_cred_getsecid
IMA: Change internal interfaces to use lsmblobs
LSM: Specify which LSM to display
LSM: Ensure the correct LSM context releaser
LSM: Use lsmcontext in security_secid_to_secctx
LSM: Use lsmcontext in security_inode_getsecctx
LSM: security_secid_to_secctx in netlink netfilter
NET: Store LSM netlabel data in a lsmblob
LSM: Verify LSM display sanity in binder
audit: add support for non-syscall auxiliary records
Audit: Add new record for multiple process LSM attributes
Audit: Add a new record for multiple object LSM attributes
LSM: Add /proc attr entry for full LSM context
AppArmor: Remove the exclusive flag
Documentation/ABI/testing/ima_policy | 8 +- Documentation/ABI/testing/procfs-attr-context | 14 + .../ABI/testing/procfs-attr-lsm_display | 22 + Documentation/security/lsm.rst | 28 + drivers/android/binder.c | 26 +- fs/ceph/xattr.c | 6 +- fs/nfs/nfs4proc.c | 8 +- fs/nfsd/nfs4xdr.c | 20 +- fs/proc/base.c | 2 + include/linux/audit.h | 43 +- include/linux/cred.h | 3 +- include/linux/lsm_hooks.h | 36 +- include/linux/security.h | 192 +++++- include/net/netlabel.h | 11 +- include/net/scm.h | 15 +- include/net/xfrm.h | 13 +- include/uapi/linux/audit.h | 2 + kernel/audit.c | 175 ++++-- kernel/audit.h | 11 +- kernel/auditfilter.c | 36 +- kernel/auditsc.c | 193 +++--- kernel/cred.c | 12 +- net/ipv4/cipso_ipv4.c | 26 +- net/ipv4/ip_sockglue.c | 12 +- net/netfilter/nf_conntrack_netlink.c | 24 +- net/netfilter/nf_conntrack_standalone.c | 11 +- net/netfilter/nfnetlink_queue.c | 38 +- net/netfilter/nft_meta.c | 10 +- net/netfilter/xt_SECMARK.c | 7 +- net/netlabel/netlabel_domainhash.c | 4 +- net/netlabel/netlabel_kapi.c | 6 +- net/netlabel/netlabel_unlabeled.c | 106 ++-- net/netlabel/netlabel_unlabeled.h | 2 +- net/netlabel/netlabel_user.c | 23 +- net/netlabel/netlabel_user.h | 2 +- net/xfrm/xfrm_policy.c | 10 +- net/xfrm/xfrm_state.c | 20 +- security/apparmor/include/apparmor.h | 3 +- security/apparmor/include/net.h | 6 +- security/apparmor/include/procattr.h | 2 +- security/apparmor/lsm.c | 105 ++-- security/apparmor/procattr.c | 22 +- security/bpf/hooks.c | 12 +- security/commoncap.c | 7 +- security/integrity/ima/ima.h | 10 +- security/integrity/ima/ima_api.c | 13 +- security/integrity/ima/ima_appraise.c | 9 +- security/integrity/ima/ima_main.c | 60 +- security/integrity/ima/ima_policy.c | 87 ++- security/integrity/integrity_audit.c | 6 +- security/landlock/cred.c | 2 +- security/landlock/fs.c | 2 +- security/landlock/ptrace.c | 2 +- security/landlock/setup.c | 4 + security/landlock/setup.h | 1 + security/loadpin/loadpin.c | 8 +- security/lockdown/lockdown.c | 7 +- security/safesetid/lsm.c | 8 +- security/security.c | 574 ++++++++++++++++-- security/selinux/hooks.c | 99 +-- security/selinux/include/classmap.h | 2 +- security/selinux/include/objsec.h | 5 + security/selinux/include/security.h | 1 + security/selinux/netlabel.c | 25 +- security/selinux/ss/services.c | 4 +- security/smack/smack.h | 6 + security/smack/smack_access.c | 2 +- security/smack/smack_lsm.c | 91 +-- security/smack/smack_netfilter.c | 8 +- security/smack/smackfs.c | 13 +- security/tomoyo/tomoyo.c | 8 +- security/yama/yama_lsm.c | 7 +- 72 files changed, 1758 insertions(+), 640 deletions(-) create mode 100644 Documentation/ABI/testing/procfs-attr-context create mode 100644 Documentation/ABI/testing/procfs-attr-lsm_display
Signature · 1 line
2.29.2
Casey Schaufler
maintainerPATCH v26 1/25 LSM: Infrastructure management of the sock security
Move management of the sock->sk_security blob out of the individual security modules and into the security infrastructure. Instead of allocating the blobs from within the modules the modules tell the infrastructure how much space is required, and the space is allocated there.
--- include/linux/lsm_hooks.h | 1 + security/apparmor/include/net.h | 6 ++- security/apparmor/lsm.c | 38 ++++----------- security/security.c | 36 +++++++++++++- security/selinux/hooks.c | 78 +++++++++++++++---------------- security/selinux/include/objsec.h | 5 ++ security/selinux/netlabel.c | 23 ++++----- security/smack/smack.h | 5 ++ security/smack/smack_lsm.c | 66 ++++++++++++-------------- security/smack/smack_netfilter.c | 8 ++-- 10 files changed, 145 insertions(+), 121 deletions(-)
include/linux/lsm_hooks.h +1 -0
@@ -1588,6 +1588,7 @@ struct lsm_blob_sizes {int lbs_cred;int lbs_file;int lbs_inode; +int lbs_sock;int lbs_superblock;int lbs_ipc;int lbs_msg_msg;
security/apparmor/include/net.h +5 -1
@@ -51,7 +51,11 @@ struct aa_sk_ctx {struct aa_label *peer; }; -#define SK_CTX (X ) ((X)->sk_security) +static inline struct aa_sk_ctx *aa_sock (const struct sock *sk) +{ +return sk->sk_security + apparmor_blob_sizes.lbs_sock; +} +#define SOCK_ctx (X ) SOCK_INODE(X)->i_security#define DEFINE_AUDIT_NET (NAME ,OP ,SK ,F ,T ,P ) \ struct lsm_network_audit NAME ## _net = { .sk = (SK), \
security/apparmor/lsm.c +10 -28
@@ -775,33 +775,15 @@ static int apparmor_task_kill(struct task_struct *target, struct kernel_siginforeturn error; } -/** -* apparmor_sk_alloc_security - allocate and attach the sk_security field -*/ -static int apparmor_sk_alloc_security (struct sock *sk,int family,gfp_t flags) -{ -struct aa_sk_ctx *ctx; - - ctx =kzalloc (sizeof (*ctx), flags); -if (!ctx) -return -ENOMEM ; - -SK_CTX (sk) = ctx; - -return 0 ; -} -/** * apparmor_sk_free_security - free the sk_security field */ static void apparmor_sk_free_security (struct sock *sk) { -struct aa_sk_ctx *ctx =SK_CTX (sk); +struct aa_sk_ctx *ctx =aa_sock (sk); -SK_CTX (sk) =NULL ;aa_put_label (ctx->label);aa_put_label (ctx->peer); -kfree (ctx); }/**
@@ -810,8 +792,8 @@ static void apparmor_sk_free_security(struct sock *sk)static void apparmor_sk_clone_security (const struct sock *sk,struct sock *newsk) { -struct aa_sk_ctx *ctx =SK_CTX (sk); -struct aa_sk_ctx *new =SK_CTX (newsk); +struct aa_sk_ctx *ctx =aa_sock (sk); +struct aa_sk_ctx *new =aa_sock (newsk);if (new->label)aa_put_label (new->label);
@@ -867,7 +849,7 @@ static int apparmor_socket_post_create(struct socket *sock, int family, label =aa_get_current_label ();if (sock->sk) { -struct aa_sk_ctx *ctx =SK_CTX (sock->sk); +struct aa_sk_ctx *ctx =aa_sock (sock->sk);aa_put_label (ctx->label); ctx->label =aa_get_label (label);
@@ -1052,7 +1034,7 @@ static int apparmor_socket_shutdown(struct socket *sock, int how)*/ static int apparmor_socket_sock_rcv_skb (struct sock *sk,struct sk_buff *skb) { -struct aa_sk_ctx *ctx =SK_CTX (sk); +struct aa_sk_ctx *ctx =aa_sock (sk);if (!skb->secmark)return 0 ;
@@ -1065,7 +1047,7 @@ static int apparmor_socket_sock_rcv_skb(struct sock *sk, struct sk_buff *skb)static struct aa_label *sk_peer_label (struct sock *sk) { -struct aa_sk_ctx *ctx =SK_CTX (sk); +struct aa_sk_ctx *ctx =aa_sock (sk);if (ctx->peer)return ctx->peer;
@@ -1149,7 +1131,7 @@ static int apparmor_socket_getpeersec_dgram(struct socket *sock,*/ static void apparmor_sock_graft (struct sock *sk,struct socket *parent) { -struct aa_sk_ctx *ctx =SK_CTX (sk); +struct aa_sk_ctx *ctx =aa_sock (sk);if (!ctx->label) ctx->label =aa_get_current_label ();
@@ -1159,7 +1141,7 @@ static void apparmor_sock_graft(struct sock *sk, struct socket *parent)static int apparmor_inet_conn_request (const struct sock *sk,struct sk_buff *skb,struct request_sock *req) { -struct aa_sk_ctx *ctx =SK_CTX (sk); +struct aa_sk_ctx *ctx =aa_sock (sk);if (!skb->secmark)return 0 ;
@@ -1176,6 +1158,7 @@ struct lsm_blob_sizes apparmor_blob_sizes __lsm_ro_after_init = { .lbs_cred = sizeof(struct aa_task_ctx *), .lbs_file = sizeof(struct aa_file_ctx), .lbs_task = sizeof(struct aa_task_ctx), + .lbs_sock = sizeof(struct aa_sk_ctx), }; static struct security_hook_list apparmor_hooks[] __lsm_ro_after_init = {
@@ -1212,7 +1195,6 @@ static struct security_hook_list apparmor_hooks[] __lsm_ro_after_init = {LSM_HOOK_INIT (getprocattr, apparmor_getprocattr),LSM_HOOK_INIT (setprocattr, apparmor_setprocattr), -LSM_HOOK_INIT (sk_alloc_security, apparmor_sk_alloc_security),LSM_HOOK_INIT (sk_free_security, apparmor_sk_free_security), LSM_HOOK_INIT(sk_clone_security, apparmor_sk_clone_security),
@@ -1764,7 +1746,7 @@ static unsigned int apparmor_ip_postroute(void *priv,if (sk ==NULL )return NF_ACCEPT ; - ctx =SK_CTX (sk); + ctx =aa_sock (sk);if (!apparmor_secmark_check (ctx->label,OP_SENDMSG ,AA_MAY_SEND , skb->secmark, sk))return NF_ACCEPT ;
security/security.c +35 -1
@@ -29,6 +29,7 @@#include <linux/string.h> #include <linux/msg.h> #include <net/flow.h> +#include <net/sock.h> #define MAX_LSM_EVM_XATTR 2
@@ -203,6 +204,7 @@ static void __init lsm_set_blob_sizes(struct lsm_blob_sizes *needed)lsm_set_blob_size (&needed->lbs_inode, &blob_sizes.lbs_inode);lsm_set_blob_size (&needed->lbs_ipc, &blob_sizes.lbs_ipc);lsm_set_blob_size (&needed->lbs_msg_msg, &blob_sizes.lbs_msg_msg); +lsm_set_blob_size (&needed->lbs_sock, &blob_sizes.lbs_sock);lsm_set_blob_size (&needed->lbs_superblock, &blob_sizes.lbs_superblock);lsm_set_blob_size (&needed->lbs_task, &blob_sizes.lbs_task); }
@@ -339,6 +341,7 @@ static void __init ordered_lsm_init(void)init_debug ("inode blob size = %d\n" , blob_sizes.lbs_inode);init_debug ("ipc blob size = %d\n" , blob_sizes.lbs_ipc);init_debug ("msg_msg blob size = %d\n" , blob_sizes.lbs_msg_msg); +init_debug ("sock blob size = %d\n" , blob_sizes.lbs_sock);init_debug ("superblock blob size = %d\n" , blob_sizes.lbs_superblock);init_debug ("task blob size = %d\n" , blob_sizes.lbs_task);
@@ -658,6 +661,28 @@ static int lsm_msg_msg_alloc(struct msg_msg *mp)return 0 ; } +/** +* lsm_sock_alloc - allocate a composite sock blob +* @sock: the sock that needs a blob +* @priority: allocation mode +* +* Allocate the sock blob for all the modules +* +* Returns 0, or -ENOMEM if memory can't be allocated. +*/ +static int lsm_sock_alloc (struct sock *sock,gfp_t priority) +{ +if (blob_sizes.lbs_sock ==0 ) { + sock->sk_security =NULL ; +return 0 ; + } + + sock->sk_security =kzalloc (blob_sizes.lbs_sock, priority); +if (sock->sk_security ==NULL ) +return -ENOMEM ; +return 0 ; +} +/** * lsm_early_task - during initialization allocate a composite task blob * @task: the task that needs a blob
@@ -2258,12 +2283,21 @@ EXPORT_SYMBOL(security_socket_getpeersec_dgram);int security_sk_alloc (struct sock *sk,int family,gfp_t priority) { -return call_int_hook (sk_alloc_security,0 , sk, family, priority); +int rc =lsm_sock_alloc (sk, priority); + +if (unlikely (rc)) +return rc; + rc =call_int_hook (sk_alloc_security,0 , sk, family, priority); +if (unlikely (rc)) +security_sk_free (sk); +return rc; }void security_sk_free (struct sock *sk) {call_void_hook (sk_free_security, sk); +kfree (sk->sk_security); + sk->sk_security =NULL ; }void security_sk_clone (const struct sock *sk,struct sock *newsk)
security/selinux/hooks.c +37 -41
249 lines not on this page · raw message
security/selinux/include/objsec.h +5 -0
@@ -194,4 +194,9 @@ static inline struct superblock_security_struct *selinux_superblock( return superblock->s_security + selinux_blob_sizes.lbs_superblock; } +static inline struct sk_security_struct *selinux_sock (const struct sock *sock) +{ +return sock->sk_security + selinux_blob_sizes.lbs_sock; +} +#endif /* _SELINUX_OBJSEC_H_ */
security/selinux/netlabel.c +12 -11
@@ -17,6 +17,7 @@#include <linux/gfp.h> #include <linux/ip.h> #include <linux/ipv6.h> +#include <linux/lsm_hooks.h> #include <net/sock.h> #include <net/netlabel.h> #include <net/ip.h>
@@ -67,7 +68,7 @@ static int selinux_netlbl_sidlookup_cached(struct sk_buff *skb,static struct netlbl_lsm_secattr *selinux_netlbl_sock_genattr (struct sock *sk) {int rc; -struct sk_security_struct *sksec = sk->sk_security; +struct sk_security_struct *sksec =selinux_sock (sk);struct netlbl_lsm_secattr *secattr;if (sksec->nlbl_secattr !=NULL )
@@ -100,7 +101,7 @@ static struct netlbl_lsm_secattr *selinux_netlbl_sock_getattr( const struct sock *sk, u32 sid) { -struct sk_security_struct *sksec = sk->sk_security; +struct sk_security_struct *sksec =selinux_sock (sk);struct netlbl_lsm_secattr *secattr = sksec->nlbl_secattr;if (secattr ==NULL )
@@ -235,7 +236,7 @@ int selinux_netlbl_skbuff_setsid(struct sk_buff *skb,* being labeled by it's parent socket, if it is just exit */ sk =skb_to_full_sk (skb);if (sk !=NULL ) { -struct sk_security_struct *sksec = sk->sk_security; +struct sk_security_struct *sksec =selinux_sock (sk);if (sksec->nlbl_state !=NLBL_REQSKB )return 0 ;
@@ -273,7 +274,7 @@ int selinux_netlbl_sctp_assoc_request(struct sctp_endpoint *ep, {int rc;struct netlbl_lsm_secattr secattr; -struct sk_security_struct *sksec = ep->base.sk->sk_security; +struct sk_security_struct *sksec =selinux_sock (ep->base.sk);struct sockaddr_in addr4;struct sockaddr_in6 addr6;
@@ -352,7 +353,7 @@ int selinux_netlbl_inet_conn_request(struct request_sock *req, u16 family)*/ void selinux_netlbl_inet_csk_clone (struct sock *sk,u16 family) { -struct sk_security_struct *sksec = sk->sk_security; +struct sk_security_struct *sksec =selinux_sock (sk);if (family ==PF_INET ) sksec->nlbl_state =NLBL_LABELED ;
@@ -370,8 +371,8 @@ void selinux_netlbl_inet_csk_clone(struct sock *sk, u16 family)*/ void selinux_netlbl_sctp_sk_clone (struct sock *sk,struct sock *newsk) { -struct sk_security_struct *sksec = sk->sk_security; -struct sk_security_struct *newsksec = newsk->sk_security; +struct sk_security_struct *sksec =selinux_sock (sk); +struct sk_security_struct *newsksec =selinux_sock (newsk); newsksec->nlbl_state = sksec->nlbl_state; }
@@ -389,7 +390,7 @@ void selinux_netlbl_sctp_sk_clone(struct sock *sk, struct sock *newsk)int selinux_netlbl_socket_post_create (struct sock *sk,u16 family) {int rc; -struct sk_security_struct *sksec = sk->sk_security; +struct sk_security_struct *sksec =selinux_sock (sk);struct netlbl_lsm_secattr *secattr;if (family !=PF_INET && family !=PF_INET6 )
@@ -504,7 +505,7 @@ int selinux_netlbl_socket_setsockopt(struct socket *sock, {int rc =0 ;struct sock *sk = sock->sk; -struct sk_security_struct *sksec = sk->sk_security; +struct sk_security_struct *sksec =selinux_sock (sk);struct netlbl_lsm_secattr secattr; if (selinux_netlbl_option(level, optname) &&
@@ -542,7 +543,7 @@ static int selinux_netlbl_socket_connect_helper(struct sock *sk, struct sockaddr *addr) {int rc; -struct sk_security_struct *sksec = sk->sk_security; +struct sk_security_struct *sksec =selinux_sock (sk);struct netlbl_lsm_secattr *secattr;/* connected sockets are allowed to disconnect when the address family
@@ -581,7 +582,7 @@ static int selinux_netlbl_socket_connect_helper(struct sock *sk,int selinux_netlbl_socket_connect_locked (struct sock *sk,struct sockaddr *addr) { -struct sk_security_struct *sksec = sk->sk_security; +struct sk_security_struct *sksec =selinux_sock (sk);if (sksec->nlbl_state !=NLBL_REQSKB && sksec->nlbl_state !=NLBL_CONNLABELED )
security/smack/smack.h +5 -0
@@ -363,6 +363,11 @@ static inline struct superblock_smack *smack_superblock(return superblock->s_security + smack_blob_sizes.lbs_superblock; } +static inline struct socket_smack *smack_sock (const struct sock *sock) +{ +return sock->sk_security + smack_blob_sizes.lbs_sock; +} +/* * Is the directory transmuting? */
security/smack/smack_lsm.c +31 -35
233 lines not on this page · raw message
security/smack/smack_netfilter.c +4 -4
@@ -28,8 +28,8 @@ static unsigned int smack_ipv6_output(void *priv,struct socket_smack *ssp;struct smack_known *skp; -if (sk && sk->sk_security) { - ssp = sk->sk_security; +if (sk &&smack_sock (sk)) { + ssp =smack_sock (sk); skp = ssp->smk_out; skb->secmark = skp->smk_secid; }
@@ -46,8 +46,8 @@ static unsigned int smack_ipv4_output(void *priv,struct socket_smack *ssp;struct smack_known *skp; -if (sk && sk->sk_security) { - ssp = sk->sk_security; +if (sk &&smack_sock (sk)) { + ssp =smack_sock (sk); skp = ssp->smk_out; skb->secmark = skp->smk_secid; }
Signature · 1 line
2.29.2
Casey Schaufler
maintainerPATCH v26 2/25 LSM: Add the lsmblob data structure.
In reply to Casey SchauflerWhen more than one security module is exporting data to audit and networking sub-systems a single 32 bit integer is no longer sufficient to represent the data. Add a structure to be used instead.
The lsmblob structure is currently an array of u32 "secids". There is an entry for each of the security modules built into the system that would use secids if active. The system assigns the module a "slot" when it registers hooks. If modules are compiled in but not registered there will be unused slots.
A new lsm_id structure, which contains the name of the LSM and its slot number, is created. There is an instance for each LSM, which assigns the name and passes it to the infrastructure to set the slot.
The audit rules data is expanded to use an array of security module data rather than a single instance. Because IMA uses the audit rule functions it is affected as well.
To: Mimi Zohar <[email protected]>
To: Mickaël Salaün <[email protected]>
--- include/linux/audit.h | 4 +- include/linux/lsm_hooks.h | 12 ++++- include/linux/security.h | 67 +++++++++++++++++++++++++-- kernel/auditfilter.c | 24 +++++----- kernel/auditsc.c | 13 +++--- security/apparmor/lsm.c | 7 ++- security/bpf/hooks.c | 12 ++++- security/commoncap.c | 7 ++- security/integrity/ima/ima_policy.c | 40 +++++++++++----- security/landlock/cred.c | 2 +- security/landlock/fs.c | 2 +- security/landlock/ptrace.c | 2 +- security/landlock/setup.c | 4 ++ security/landlock/setup.h | 1 + security/loadpin/loadpin.c | 8 +++- security/lockdown/lockdown.c | 7 ++- security/safesetid/lsm.c | 8 +++- security/security.c | 72 ++++++++++++++++++++++++----- security/selinux/hooks.c | 8 +++- security/smack/smack_lsm.c | 7 ++- security/tomoyo/tomoyo.c | 8 +++- security/yama/yama_lsm.c | 7 ++- 22 files changed, 262 insertions(+), 60 deletions(-)
include/linux/audit.h +3 -1
@@ -11,6 +11,7 @@#include <linux/sched.h> #include <linux/ptrace.h> +#include <linux/security.h> #include <uapi/linux/audit.h> #include <uapi/linux/netfilter/nf_tables.h>
@@ -65,8 +66,9 @@ struct audit_field {kuid_t uid;kgid_t gid;struct { +bool lsm_isset;char *lsm_str; -void *lsm_rule; +void *lsm_rules[LSMBLOB_ENTRIES ]; }; };u32 op;
include/linux/lsm_hooks.h +10 -2
@@ -1570,6 +1570,14 @@ struct security_hook_heads {#undef LSM_HOOK } __randomize_layout; +/* +* Information that identifies a security module. +*/ +struct lsm_id { +const char *lsm;/* Name of the LSM */ +int slot;/* Slot in lsmblob if one is allocated */ +}; +/* * Security module hook list structure. * For use with generic list macros for common operations.
@@ -1578,7 +1586,7 @@ struct security_hook_list {struct hlist_node list;struct hlist_head *head;union security_list_options hook; -char *lsm; +struct lsm_id *lsmid; } __randomize_layout;/*
@@ -1614,7 +1622,7 @@ extern struct security_hook_heads security_hook_heads;extern char *lsm_names;extern void security_add_hooks (struct security_hook_list *hooks,int count, -char *lsm); +struct lsm_id *lsmid);#define LSM_FLAG_LEGACY_MAJOR BIT(0)#define LSM_FLAG_EXCLUSIVE BIT(1)
include/linux/security.h +63 -4
89 lines not on this page · raw message
kernel/auditfilter.c +13 -11
@@ -74,7 +74,7 @@ static void audit_free_lsm_field(struct audit_field *f)case AUDIT_OBJ_LEV_LOW :case AUDIT_OBJ_LEV_HIGH :kfree (f->lsm_str); -security_audit_rule_free (f->lsm_rule); +security_audit_rule_free (f->lsm_rules); } }
@@ -519,9 +519,10 @@ static struct audit_entry *audit_data_to_entry(struct audit_rule_data *data, goto exit_free; } entry->rule.buflen += f_val; + f->lsm_isset = true; f->lsm_str = str; err =security_audit_rule_init (f->type, f->op, str, - (void **)&f->lsm_rule); + f->lsm_rules);/* Keep currently invalid fields around in case they * become valid after a policy reload. */ if (err == -EINVAL ) {
@@ -774,7 +775,7 @@ static int audit_compare_rule(struct audit_krule *a, struct audit_krule *b)return 0 ; } -/* Duplicate LSM field information. The lsm_rule is opaque, so must be +/* Duplicate LSM field information. The lsm_rules is opaque, so must be * re-initialized. */ static inline int audit_dupe_lsm_field (struct audit_field *df,struct audit_field *sf)
@@ -788,9 +789,9 @@ static inline int audit_dupe_lsm_field(struct audit_field *df,return -ENOMEM ; df->lsm_str = lsm_str; -/* our own (refreshed) copy of lsm_rule */ +/* our own (refreshed) copy of lsm_rules */ ret =security_audit_rule_init (df->type, df->op, df->lsm_str, - (void **)&df->lsm_rule); + df->lsm_rules);/* Keep currently invalid fields around in case they * become valid after a policy reload. */ if (ret == -EINVAL ) {
@@ -842,7 +843,7 @@ struct audit_entry *audit_dupe_rule(struct audit_krule *old) new->tree = old->tree;memcpy (new->fields, old->fields,sizeof (struct audit_field ) * fcount); -/* deep copy this information, updating the lsm_rule fields, because +/* deep copy this information, updating the lsm_rules fields, because * the originals will all be freed when the old rule is freed. */ for (i =0 ; i < fcount; i++) {switch (new->fields[i].type) {
@@ -1358,11 +1359,12 @@ int audit_filter(int msgtype, unsigned int listtype)case AUDIT_SUBJ_TYPE :case AUDIT_SUBJ_SEN :case AUDIT_SUBJ_CLR : -if (f->lsm_rule) { +if (f->lsm_isset) {security_task_getsecid_subj (current, &sid); result =security_audit_rule_match (sid, - f->type, f->op, f->lsm_rule); + f->type, f->op, + f->lsm_rules); }break ;case AUDIT_EXE :
@@ -1389,7 +1391,7 @@ int audit_filter(int msgtype, unsigned int listtype)return ret; } -static int update_lsm_rule (struct audit_krule *r) +static int update_lsm_rules (struct audit_krule *r) { struct audit_entry *entry = container_of(r, struct audit_entry, rule);struct audit_entry *nentry;
@@ -1421,7 +1423,7 @@ static int update_lsm_rule(struct audit_krule *r)return err; } -/* This function will re-initialize the lsm_rule field of all applicable rules. +/* This function will re-initialize the lsm_rules field of all applicable rules. * It will traverse the filter lists serarching for rules that contain LSM * specific filter fields. When such a rule is found, it is copied, the * LSM field is re-initialized, and the old rule is replaced with the
@@ -1436,7 +1438,7 @@ int audit_update_lsm_rules(void)for (i =0 ; i <AUDIT_NR_FILTERS ; i++) {list_for_each_entry_safe (r, n, &audit_rules_list[i], list) { -int res =update_lsm_rule (r); +int res =update_lsm_rules (r);if (!err) err = res; }
kernel/auditsc.c +6 -7
@@ -665,14 +665,13 @@ static int audit_filter_rules(struct task_struct *tsk,match for now to avoid losing information that may be wanted. An error message will also be logged upon error */ -if (f->lsm_rule) { +if (f->lsm_isset) {if (need_sid) {security_task_getsecid_subj (tsk, &sid); need_sid =0 ; } result =security_audit_rule_match (sid, f->type, - f->op, - f->lsm_rule); + f->op, f->lsm_rules); }break ;case AUDIT_OBJ_USER :
@@ -682,21 +681,21 @@ static int audit_filter_rules(struct task_struct *tsk,case AUDIT_OBJ_LEV_HIGH :/* The above note for AUDIT_SUBJ_USER...AUDIT_SUBJ_CLR also applies here */ -if (f->lsm_rule) { +if (f->lsm_isset) {/* Find files that match */ if (name) { result =security_audit_rule_match ( name->osid, f->type, f->op, - f->lsm_rule); + f->lsm_rules); }else if (ctx) {list_for_each_entry (n, &ctx->names_list, list) {if (security_audit_rule_match ( n->osid, f->type, f->op, - f->lsm_rule)) { + f->lsm_rules)) { ++result;break ; }
@@ -707,7 +706,7 @@ static int audit_filter_rules(struct task_struct *tsk,break ;if (security_audit_rule_match (ctx->ipc.osid, f->type, f->op, - f->lsm_rule)) + f->lsm_rules)) ++result; }break ;
security/apparmor/lsm.c +6 -1
@@ -1161,6 +1161,11 @@ struct lsm_blob_sizes apparmor_blob_sizes __lsm_ro_after_init = { .lbs_sock = sizeof(struct aa_sk_ctx), }; +static struct lsm_id apparmor_lsmid __lsm_ro_after_init = { + .lsm ="apparmor" , + .slot =LSMBLOB_NEEDED +}; + static struct security_hook_list apparmor_hooks[] __lsm_ro_after_init = {LSM_HOOK_INIT (ptrace_access_check, apparmor_ptrace_access_check),LSM_HOOK_INIT (ptrace_traceme, apparmor_ptrace_traceme),
@@ -1862,7 +1867,7 @@ static int __init apparmor_init(void) goto buffers_out; }security_add_hooks (apparmor_hooks,ARRAY_SIZE (apparmor_hooks), -"apparmor" ); + &apparmor_lsmid);/* Report that AppArmor successfully initialized */ apparmor_initialized =1 ;
security/bpf/hooks.c +11 -1
@@ -15,9 +15,19 @@ static struct security_hook_list bpf_lsm_hooks[] __lsm_ro_after_init = { LSM_HOOK_INIT(task_free, bpf_task_storage_free), }; +/* +* slot has to be LSMBLOB_NEEDED because some of the hooks +* supplied by this module require a slot. +*/ +struct lsm_id bpf_lsmid __lsm_ro_after_init = { + .lsm ="bpf" , + .slot =LSMBLOB_NEEDED +}; +static int __init bpf_lsm_init (void ) { -security_add_hooks (bpf_lsm_hooks,ARRAY_SIZE (bpf_lsm_hooks),"bpf" ); +security_add_hooks (bpf_lsm_hooks,ARRAY_SIZE (bpf_lsm_hooks), + &bpf_lsmid);pr_info ("LSM support for eBPF active\n" );return 0 ; }
security/commoncap.c +6 -1
@@ -1443,6 +1443,11 @@ int cap_mmap_file(struct file *file, unsigned long reqprot,#ifdef CONFIG_SECURITY +static struct lsm_id capability_lsmid __lsm_ro_after_init = { + .lsm ="capability" , + .slot =LSMBLOB_NOT_NEEDED +}; + static struct security_hook_list capability_hooks[] __lsm_ro_after_init = {LSM_HOOK_INIT (capable, cap_capable), LSM_HOOK_INIT(settime, cap_settime),
@@ -1467,7 +1472,7 @@ static struct security_hook_list capability_hooks[] __lsm_ro_after_init = {static int __init capability_init (void ) {security_add_hooks (capability_hooks,ARRAY_SIZE (capability_hooks), -"capability" ); + &capability_lsmid);return 0 ; }
security/integrity/ima/ima_policy.c +29 -11
@@ -80,7 +80,7 @@ struct ima_rule_entry {bool (*fowner_op)(kuid_t ,kuid_t );/* uid_eq(), uid_gt(), uid_lt() */ int pcr;struct { -void *rule;/* LSM file metadata specific */ +void *rules[LSMBLOB_ENTRIES ];/* LSM file metadata specific */ char *args_p;/* audit value */ int type;/* audit type */ } lsm[MAX_LSM_RULES ];
@@ -90,6 +90,22 @@ struct ima_rule_entry {struct ima_template_desc *template; }; +/** +* ima_lsm_isset - Is a rule set for any of the active security modules +* @rules: The set of IMA rules to check +* +* If a rule is set for any LSM return true, otherwise return false. +*/ +static inline bool ima_lsm_isset (void *rules[]) +{ +int i; + +for (i =0 ; i <LSMBLOB_ENTRIES ; i++) +if (rules[i]) +return true; +return false; +} +/* * Without LSM specific knowledge, the default policy can only be * written in terms of .action, .func, .mask, .fsmagic, .uid, and .fowner
@@ -335,9 +351,11 @@ static void ima_free_rule_opt_list(struct ima_rule_opt_list *opt_list)static void ima_lsm_free_rule (struct ima_rule_entry *entry) {int i; +int r;for (i =0 ; i <MAX_LSM_RULES ; i++) { -ima_filter_rule_free (entry->lsm[i].rule); +for (r =0 ; r <LSMBLOB_ENTRIES ; r++) +ima_filter_rule_free (entry->lsm[i].rules[r]);kfree (entry->lsm[i].args_p); } }
@@ -388,8 +406,8 @@ static struct ima_rule_entry *ima_lsm_copy_rule(struct ima_rule_entry *entry)ima_filter_rule_init (nentry->lsm[i].type, Audit_equal, nentry->lsm[i].args_p, - &nentry->lsm[i].rule); -if (!nentry->lsm[i].rule) + &nentry->lsm[i].rules[0 ]); +if (!ima_lsm_isset (nentry->lsm[i].rules))pr_warn ("rule for LSM \'%s\' is undefined\n" , nentry->lsm[i].args_p); }
@@ -578,7 +596,7 @@ static bool ima_match_rules(struct ima_rule_entry *rule,int rc =0 ;u32 osid; -if (!rule->lsm[i].rule) { +if (!ima_lsm_isset (rule->lsm[i].rules)) {if (!rule->lsm[i].args_p)continue ;else
@@ -591,14 +609,14 @@ static bool ima_match_rules(struct ima_rule_entry *rule,security_inode_getsecid (inode, &osid); rc =ima_filter_rule_match (osid, rule->lsm[i].type, Audit_equal, - rule->lsm[i].rule); + rule->lsm[i].rules);break ;case LSM_SUBJ_USER :case LSM_SUBJ_ROLE :case LSM_SUBJ_TYPE : rc =ima_filter_rule_match (secid, rule->lsm[i].type, Audit_equal, - rule->lsm[i].rule); + rule->lsm[i].rules);break ;default :break ;
@@ -994,7 +1012,7 @@ static int ima_lsm_rule_init(struct ima_rule_entry *entry, {int result; -if (entry->lsm[lsm_rule].rule) +if (ima_lsm_isset (entry->lsm[lsm_rule].rules))return -EINVAL ; entry->lsm[lsm_rule].args_p =match_strdup (args);
@@ -1004,8 +1022,8 @@ static int ima_lsm_rule_init(struct ima_rule_entry *entry, entry->lsm[lsm_rule].type = audit_type; result =ima_filter_rule_init (entry->lsm[lsm_rule].type, Audit_equal, entry->lsm[lsm_rule].args_p, - &entry->lsm[lsm_rule].rule); -if (!entry->lsm[lsm_rule].rule) { + &entry->lsm[lsm_rule].rules[0 ]); +if (!ima_lsm_isset (entry->lsm[lsm_rule].rules)) {pr_warn ("rule for LSM \'%s\' is undefined\n" , entry->lsm[lsm_rule].args_p);
@@ -1812,7 +1830,7 @@ int ima_policy_show(struct seq_file *m, void *v) }for (i =0 ; i <MAX_LSM_RULES ; i++) { -if (entry->lsm[i].rule) { +if (ima_lsm_isset (entry->lsm[i].rules)) {switch (i) {case LSM_OBJ_USER : seq_printf(m, pt(Opt_obj_user),
security/landlock/cred.c +1 -1
@@ -42,5 +42,5 @@ static struct security_hook_list landlock_hooks[] __lsm_ro_after_init = {__init void landlock_add_cred_hooks (void ) {security_add_hooks (landlock_hooks,ARRAY_SIZE (landlock_hooks), -LANDLOCK_NAME ); + &landlock_lsmid); }
security/landlock/fs.c +1 -1
@@ -688,5 +688,5 @@ static struct security_hook_list landlock_hooks[] __lsm_ro_after_init = {__init void landlock_add_fs_hooks (void ) {security_add_hooks (landlock_hooks,ARRAY_SIZE (landlock_hooks), -LANDLOCK_NAME ); + &landlock_lsmid); }
security/landlock/ptrace.c +1 -1
@@ -116,5 +116,5 @@ static struct security_hook_list landlock_hooks[] __lsm_ro_after_init = {__init void landlock_add_ptrace_hooks (void ) {security_add_hooks (landlock_hooks,ARRAY_SIZE (landlock_hooks), -LANDLOCK_NAME ); + &landlock_lsmid); }
security/landlock/setup.c +4 -0
@@ -23,6 +23,10 @@ struct lsm_blob_sizes landlock_blob_sizes __lsm_ro_after_init = { .lbs_superblock = sizeof(struct landlock_superblock_security), }; +struct lsm_id landlock_lsmid __lsm_ro_after_init = { + .lsm =LANDLOCK_NAME , +}; +static int __init landlock_init (void ) {landlock_add_cred_hooks ();
security/landlock/setup.h +1 -0
@@ -14,5 +14,6 @@extern bool landlock_initialized;extern struct lsm_blob_sizes landlock_blob_sizes; +extern struct lsm_id landlock_lsmid;#endif /* _SECURITY_LANDLOCK_SETUP_H */
security/loadpin/loadpin.c +7 -1
@@ -192,6 +192,11 @@ static int loadpin_load_data(enum kernel_load_data_id id, bool contents)return loadpin_read_file (NULL , (enum kernel_read_file_id ) id, contents); } +static struct lsm_id loadpin_lsmid __lsm_ro_after_init = { + .lsm ="loadpin" , + .slot =LSMBLOB_NOT_NEEDED +}; + static struct security_hook_list loadpin_hooks[] __lsm_ro_after_init = {LSM_HOOK_INIT (sb_free_security, loadpin_sb_free_security),LSM_HOOK_INIT (kernel_read_file, loadpin_read_file),
@@ -239,7 +244,8 @@ static int __init loadpin_init(void)pr_info ("ready to pin (currently %senforcing)\n" , enforce ?"" :"not " );parse_exclude (); -security_add_hooks (loadpin_hooks,ARRAY_SIZE (loadpin_hooks),"loadpin" ); +security_add_hooks (loadpin_hooks,ARRAY_SIZE (loadpin_hooks), + &loadpin_lsmid);return 0 ; }
security/lockdown/lockdown.c +6 -1
@@ -75,6 +75,11 @@ static struct security_hook_list lockdown_hooks[] __lsm_ro_after_init = { LSM_HOOK_INIT(locked_down, lockdown_is_locked_down), }; +static struct lsm_id lockdown_lsmid __lsm_ro_after_init = { + .lsm ="lockdown" , + .slot =LSMBLOB_NOT_NEEDED +}; +static int __init lockdown_lsm_init (void ) {#if defined(CONFIG_LOCK_DOWN_KERNEL_FORCE_INTEGRITY )
@@ -83,7 +88,7 @@ static int __init lockdown_lsm_init(void)lock_kernel_down ("Kernel configuration" ,LOCKDOWN_CONFIDENTIALITY_MAX );#endif security_add_hooks (lockdown_hooks,ARRAY_SIZE (lockdown_hooks), -"lockdown" ); + &lockdown_lsmid);return 0 ; }
security/safesetid/lsm.c +7 -1
@@ -241,6 +241,11 @@ static int safesetid_task_fix_setgid(struct cred *new,return -EACCES ; } +static struct lsm_id safesetid_lsmid __lsm_ro_after_init = { + .lsm ="safesetid" , + .slot =LSMBLOB_NOT_NEEDED +}; +static struct security_hook_list safesetid_security_hooks[] = {LSM_HOOK_INIT (task_fix_setuid, safesetid_task_fix_setuid),LSM_HOOK_INIT (task_fix_setgid, safesetid_task_fix_setgid),
@@ -250,7 +255,8 @@ static struct security_hook_list safesetid_security_hooks[] = {static int __init safesetid_security_init (void ) {security_add_hooks (safesetid_security_hooks, -ARRAY_SIZE (safesetid_security_hooks),"safesetid" ); +ARRAY_SIZE (safesetid_security_hooks), + &safesetid_lsmid);/* Report that SafeSetID successfully initialized */ safesetid_initialized =1 ;
security/security.c +61 -11
130 lines not on this page · raw message
security/selinux/hooks.c +7 -1
@@ -7111,6 +7111,11 @@ static int selinux_perf_event_write(struct perf_event *event) }#endif +static struct lsm_id selinux_lsmid __lsm_ro_after_init = { + .lsm ="selinux" , + .slot =LSMBLOB_NEEDED +}; +/* * IMPORTANT NOTE: When adding new hooks, please be careful to keep this order: * 1. any hooks that don't belong to (2.) or (3.) below,
@@ -7424,7 +7429,8 @@ static __init int selinux_init(void)hashtab_cache_init (); -security_add_hooks (selinux_hooks,ARRAY_SIZE (selinux_hooks),"selinux" ); +security_add_hooks (selinux_hooks,ARRAY_SIZE (selinux_hooks), + &selinux_lsmid);if (avc_add_callback (selinux_netcache_avc_callback,AVC_CALLBACK_RESET ))panic ("SELinux: Unable to register AVC netcache callback\n" );
security/smack/smack_lsm.c +6 -1
@@ -4694,6 +4694,11 @@ struct lsm_blob_sizes smack_blob_sizes __lsm_ro_after_init = { .lbs_superblock = sizeof(struct superblock_smack), }; +static struct lsm_id smack_lsmid __lsm_ro_after_init = { + .lsm ="smack" , + .slot =LSMBLOB_NEEDED +}; + static struct security_hook_list smack_hooks[] __lsm_ro_after_init = {LSM_HOOK_INIT (ptrace_access_check, smack_ptrace_access_check),LSM_HOOK_INIT (ptrace_traceme, smack_ptrace_traceme),
@@ -4893,7 +4898,7 @@ static __init int smack_init(void)/* * Register with LSM */ -security_add_hooks (smack_hooks,ARRAY_SIZE (smack_hooks),"smack" ); +security_add_hooks (smack_hooks,ARRAY_SIZE (smack_hooks), &smack_lsmid); smack_enabled =1 ;pr_info ("Smack: Initializing.\n" );
security/tomoyo/tomoyo.c +7 -1
@@ -523,6 +523,11 @@ static void tomoyo_task_free(struct task_struct *task) } } +static struct lsm_id tomoyo_lsmid __lsm_ro_after_init = { + .lsm ="tomoyo" , + .slot =LSMBLOB_NOT_NEEDED +}; +/* * tomoyo_security_ops is a "struct security_operations" which is used for * registering TOMOYO.
@@ -575,7 +580,8 @@ static int __init tomoyo_init(void)struct tomoyo_task *s =tomoyo_task (current);/* register ourselves with the security framework */ -security_add_hooks (tomoyo_hooks,ARRAY_SIZE (tomoyo_hooks),"tomoyo" ); +security_add_hooks (tomoyo_hooks,ARRAY_SIZE (tomoyo_hooks), + &tomoyo_lsmid);pr_info ("TOMOYO Linux initialized\n" ); s->domain_info = &tomoyo_kernel_domain;atomic_inc (&tomoyo_kernel_domain.users);
security/yama/yama_lsm.c +6 -1
@@ -421,6 +421,11 @@ static int yama_ptrace_traceme(struct task_struct *parent)return rc; } +static struct lsm_id yama_lsmid __lsm_ro_after_init = { + .lsm ="yama" , + .slot =LSMBLOB_NOT_NEEDED +}; + static struct security_hook_list yama_hooks[] __lsm_ro_after_init = {LSM_HOOK_INIT (ptrace_access_check, yama_ptrace_access_check),LSM_HOOK_INIT (ptrace_traceme, yama_ptrace_traceme),
@@ -477,7 +482,7 @@ static inline void yama_init_sysctl(void) { }static int __init yama_init (void ) {pr_info ("Yama: becoming mindful.\n" ); -security_add_hooks (yama_hooks,ARRAY_SIZE (yama_hooks),"yama" ); +security_add_hooks (yama_hooks,ARRAY_SIZE (yama_hooks), &yama_lsmid);yama_init_sysctl ();return 0 ; }
Signature · 1 line
2.29.2
Mickaël Salaün
maintainerRe: [PATCH v26 02/25] LSM: Add the lsmblob data structure.
I like this design but there is an issue with Landlock though, see below.
Quoted 58 lines · Casey Schaufler
When more than one security module is exporting data to audit and networking sub-systems a single 32 bit integer is no longer sufficient to represent the data. Add a structure to be used instead.
The lsmblob structure is currently an array of u32 "secids". There is an entry for each of the security modules built into the system that would use secids if active. The system assigns the module a "slot" when it registers hooks. If modules are compiled in but not registered there will be unused slots.
A new
lsm_idstructure, which contains the name of the LSM and its slot number, is created. There is an instance for each LSM, which assigns the name and passes it to the infrastructure to set the slot.The audit rules data is expanded to use an array of security module data rather than a single instance. Because IMA uses the audit rule functions it is affected as well.
Acked-by Stephen Smalley <[email protected]>Acked-by Paul Moore <[email protected]>Acked-by John Johansen <[email protected]>Signed-off-by Casey Schaufler <[email protected]>Cc <[email protected]>To: Mimi Zohar <[email protected]>
To: Mickaël Salaün <[email protected]>--- include/linux/audit.h | 4 +- include/linux/lsm_hooks.h | 12 ++++- include/linux/security.h | 67 +++++++++++++++++++++++++-- kernel/auditfilter.c | 24 +++++----- kernel/auditsc.c | 13 +++--- security/apparmor/lsm.c | 7 ++- security/bpf/hooks.c | 12 ++++- security/commoncap.c | 7 ++- security/integrity/ima/ima_policy.c | 40 +++++++++++----- security/landlock/cred.c | 2 +- security/landlock/fs.c | 2 +- security/landlock/ptrace.c | 2 +- security/landlock/setup.c | 4 ++ security/landlock/setup.h | 1 + security/loadpin/loadpin.c | 8 +++- security/lockdown/lockdown.c | 7 ++- security/safesetid/lsm.c | 8 +++- security/security.c | 72 ++++++++++++++++++++++++----- security/selinux/hooks.c | 8 +++- security/smack/smack_lsm.c | 7 ++- security/tomoyo/tomoyo.c | 8 +++- security/yama/yama_lsm.c | 7 ++- 22 files changed, 262 insertions(+), 60 deletions(-)
[...]
security/landlock/setup.c struct lsm_blob_sizes landlock_blob_sizes __lsm_ro_after_init = { · line 27
Quoted 10 lines
security/landlock/setup.c +2
-0@@ -23,6 +23,10 @@ struct lsm_blob_sizes landlock_blob_sizes __lsm_ro_after_init = { .lbs_superblock = sizeof(struct landlock_superblock_security), }; +struct lsm_id landlock_lsmid __lsm_ro_after_init = { + .lsm =LANDLOCK_NAME ,
It is missing: .slot = LSMBLOB_NEEDED,
You can run the Landlock tests please?
make -C tools/testing/selftests TARGETS=landlock gen_tar tar -xf kselftest.tar.gz && ./run_kselftest.sh
Quoted 5 lines
security/landlock/setup.c +2
-0+}; +static int __init landlock_init (void ) {landlock_add_cred_hooks ();
[...]
Quoted 28 lines
security/security.c +7
-1@@ -344,6 +344,7 @@ static void __init ordered_lsm_init(void)init_debug ("sock blob size = %d\n" , blob_sizes.lbs_sock);init_debug ("superblock blob size = %d\n" , blob_sizes.lbs_superblock);init_debug ("task blob size = %d\n" , blob_sizes.lbs_task); +init_debug ("lsmblob size = %zu\n" ,sizeof (struct lsmblob ));/* * Create any kmem_caches needed for blobs @@ -471,21 +472,36 @@ static int lsm_append(const char *new, char **result)return 0 ; } +/* +* Current index to use while initializing the lsmblob secid list. +*/ +static int lsm_slot __lsm_ro_after_init; +/** * security_add_hooks - Add a modules hooks to the hook lists. * @hooks: the hooks to add * @count: the number of hooks to add -* @lsm: the name of the security module +* @lsmid: the identification information for the security module *
security/security.c static int lsm_append(const char *new, char **result) · line 494
Quoted 10 lines
security/security.c +3
-1* Each LSM has to register its hooks with the infrastructure. +* If the LSM is using hooks that export secids allocate a slot +* for it in the lsmblob. */ void __init security_add_hooks (struct security_hook_list *hooks,int count, -char *lsm) +struct lsm_id *lsmid) {int i;
Could you add a WARN_ON(!lsmid->slot || !lsmid->name) here?
Quoted 14 lines
security/security.c +9
-1+if (lsmid->slot ==LSMBLOB_NEEDED ) { +if (lsm_slot >=LSMBLOB_ENTRIES ) +panic ("%s Too many LSMs registered.\n" , __func__); + lsmid->slot = lsm_slot++; +init_debug ("%s assigned lsmblob slot %d\n" , lsmid->lsm, + lsmid->slot); + } +for (i =0 ; i < count; i++) { - hooks[i].lsm = lsm; + hooks[i].lsmid = lsmid;hlist_add_tail_rcu (&hooks[i].list, hooks[i].head); }
Casey Schaufler
maintainerRe: [PATCH v26 02/25] LSM: Add the lsmblob data structure.
Quoted 71 lines · Mickaël Salaün
I like this design but there is an issue with Landlock though, see below.
Quoted 58 lines · Casey Schaufler
When more than one security module is exporting data to audit and networking sub-systems a single 32 bit integer is no longer sufficient to represent the data. Add a structure to be used instead.
The lsmblob structure is currently an array of u32 "secids". There is an entry for each of the security modules built into the system that would use secids if active. The system assigns the module a "slot" when it registers hooks. If modules are compiled in but not registered there will be unused slots.
A new
lsm_idstructure, which contains the name of the LSM and its slot number, is created. There is an instance for each LSM, which assigns the name and passes it to the infrastructure to set the slot.The audit rules data is expanded to use an array of security module data rather than a single instance. Because IMA uses the audit rule functions it is affected as well.
Acked-by Stephen Smalley <[email protected]>Acked-by Paul Moore <[email protected]>Acked-by John Johansen <[email protected]>Signed-off-by Casey Schaufler <[email protected]>Cc <[email protected]>To: Mimi Zohar <[email protected]>
To: Mickaël Salaün <[email protected]>--- include/linux/audit.h | 4 +- include/linux/lsm_hooks.h | 12 ++++- include/linux/security.h | 67 +++++++++++++++++++++++++-- kernel/auditfilter.c | 24 +++++----- kernel/auditsc.c | 13 +++--- security/apparmor/lsm.c | 7 ++- security/bpf/hooks.c | 12 ++++- security/commoncap.c | 7 ++- security/integrity/ima/ima_policy.c | 40 +++++++++++----- security/landlock/cred.c | 2 +- security/landlock/fs.c | 2 +- security/landlock/ptrace.c | 2 +- security/landlock/setup.c | 4 ++ security/landlock/setup.h | 1 + security/loadpin/loadpin.c | 8 +++- security/lockdown/lockdown.c | 7 ++- security/safesetid/lsm.c | 8 +++- security/security.c | 72 ++++++++++++++++++++++++----- security/selinux/hooks.c | 8 +++- security/smack/smack_lsm.c | 7 ++- security/tomoyo/tomoyo.c | 8 +++- security/yama/yama_lsm.c | 7 ++- 22 files changed, 262 insertions(+), 60 deletions(-)[...]
Quoted 8 lines
security/landlock/setup.c +0
-0@@ -23,6 +23,10 @@ struct lsm_blob_sizes landlock_blob_sizes __lsm_ro_after_init = { .lbs_superblock = sizeof(struct landlock_superblock_security), };
Quoted 3 lines
security/landlock/setup.c struct lsm_blob_sizes landlock_blob_sizes __lsm_ro_after_init = { · line 27
Quoted 2 lines
security/landlock/setup.c +2
-0+struct lsm_id landlock_lsmid __lsm_ro_after_init = { + .lsm =LANDLOCK_NAME ,It is missing: .slot =
LSMBLOB_NEEDED,
Sorry for the delay.
Landlock does not provide any of the hooks that use a struct lsmblob. That would be secid_to_secctx, secctx_to_secid, inode_getsecid, cred_getsecid, kernel_act_as task_getsecid_subj task_getsecid_obj and ipc_getsecid. Setting .slot = LSMBLOB_NEEDED indicates that the LSM uses a slot in struct lsmblob. Landlock does not need a slot.
Quoted 4 lines
You can run the Landlock tests please?
make -Ctools/testing/selftestsTARGETS=landlockgen_tartar -xf kselftest.tar.gz && ./run_kselftest.sh
Sure. I'll add them to my routine.
Quoted 40 lines
Quoted 7 lines
security/landlock/setup.c +2
-0+}; +static int __init landlock_init (void ) {landlock_add_cred_hooks ();[...]
Quoted 31 lines
security/security.c +9
-1@@ -344,6 +344,7 @@ static void __init ordered_lsm_init(void)init_debug ("sock blob size = %d\n" , blob_sizes.lbs_sock);init_debug ("superblock blob size = %d\n" , blob_sizes.lbs_superblock);init_debug ("task blob size = %d\n" , blob_sizes.lbs_task); +init_debug ("lsmblob size = %zu\n" ,sizeof (struct lsmblob ));/* * Create any kmem_caches needed for blobs @@ -471,21 +472,36 @@ static int lsm_append(const char *new, char **result)return 0 ; } +/* +* Current index to use while initializing the lsmblob secid list. +*/ +static int lsm_slot __lsm_ro_after_init; +/** * security_add_hooks - Add a modules hooks to the hook lists. * @hooks: the hooks to add * @count: the number of hooks to add -* @lsm: the name of the security module +* @lsmid: the identification information for the security module * * Each LSM has to register its hooks with the infrastructure. +* If the LSM is using hooks that export secids allocate a slot +* for it in the lsmblob.
Quoted 8 lines
security/security.c static int lsm_append(const char *new, char **result) · line 494
Quoted 7 lines
security/security.c +1
-1*/ void __init security_add_hooks (struct security_hook_list *hooks,int count, -char *lsm) +struct lsm_id *lsmid) {int i;Could you add a
WARN_ON(!lsmid->slot || !lsmid->name)here?
Yes. That's reasonable.
Quoted 16 lines
Quoted 16 lines
security/security.c +9
-1+if (lsmid->slot ==LSMBLOB_NEEDED ) { +if (lsm_slot >=LSMBLOB_ENTRIES ) +panic ("%s Too many LSMs registered.\n" , __func__); + lsmid->slot = lsm_slot++; +init_debug ("%s assigned lsmblob slot %d\n" , lsmid->lsm, + lsmid->slot); + } +for (i =0 ; i < count; i++) { - hooks[i].lsm = lsm; + hooks[i].lsmid = lsmid;hlist_add_tail_rcu (&hooks[i].list, hooks[i].head); }
Mickaël Salaün
maintainerRe: [PATCH v26 02/25] LSM: Add the lsmblob data structure.
Quoted 76 lines · Casey Schaufler
Quoted 71 lines · Mickaël Salaün
I like this design but there is an issue with Landlock though, see below.
Quoted 58 lines · Casey Schaufler
When more than one security module is exporting data to audit and networking sub-systems a single 32 bit integer is no longer sufficient to represent the data. Add a structure to be used instead.
The lsmblob structure is currently an array of u32 "secids". There is an entry for each of the security modules built into the system that would use secids if active. The system assigns the module a "slot" when it registers hooks. If modules are compiled in but not registered there will be unused slots.
A new
lsm_idstructure, which contains the name of the LSM and its slot number, is created. There is an instance for each LSM, which assigns the name and passes it to the infrastructure to set the slot.The audit rules data is expanded to use an array of security module data rather than a single instance. Because IMA uses the audit rule functions it is affected as well.
Acked-by Stephen Smalley <[email protected]>Acked-by Paul Moore <[email protected]>Acked-by John Johansen <[email protected]>Signed-off-by Casey Schaufler <[email protected]>Cc <[email protected]>To: Mimi Zohar <[email protected]>
To: Mickaël Salaün <[email protected]>--- include/linux/audit.h | 4 +- include/linux/lsm_hooks.h | 12 ++++- include/linux/security.h | 67 +++++++++++++++++++++++++-- kernel/auditfilter.c | 24 +++++----- kernel/auditsc.c | 13 +++--- security/apparmor/lsm.c | 7 ++- security/bpf/hooks.c | 12 ++++- security/commoncap.c | 7 ++- security/integrity/ima/ima_policy.c | 40 +++++++++++----- security/landlock/cred.c | 2 +- security/landlock/fs.c | 2 +- security/landlock/ptrace.c | 2 +- security/landlock/setup.c | 4 ++ security/landlock/setup.h | 1 + security/loadpin/loadpin.c | 8 +++- security/lockdown/lockdown.c | 7 ++- security/safesetid/lsm.c | 8 +++- security/security.c | 72 ++++++++++++++++++++++++----- security/selinux/hooks.c | 8 +++- security/smack/smack_lsm.c | 7 ++- security/tomoyo/tomoyo.c | 8 +++- security/yama/yama_lsm.c | 7 ++- 22 files changed, 262 insertions(+), 60 deletions(-)[...]
Quoted 8 lines
security/landlock/setup.c +0
-0@@ -23,6 +23,10 @@ struct lsm_blob_sizes landlock_blob_sizes __lsm_ro_after_init = { .lbs_superblock = sizeof(struct landlock_superblock_security), };Quoted 3 lines
security/landlock/setup.c struct lsm_blob_sizes landlock_blob_sizes __lsm_ro_after_init = { · line 27
Quoted 2 lines
security/landlock/setup.c +2
-0+struct lsm_id landlock_lsmid __lsm_ro_after_init = { + .lsm =LANDLOCK_NAME ,It is missing: .slot =
LSMBLOB_NEEDED,
Quoted 7 lines
Sorry for the delay.
Landlock does not provide any of the hooks that use a struct lsmblob. That would be
secid_to_secctx,secctx_to_secid,inode_getsecid,cred_getsecid,kernel_act_astask_getsecid_subjtask_getsecid_objandipc_getsecid. Setting .slot =LSMBLOB_NEEDEDindicates that the LSM uses a slot in struct lsmblob. Landlock does not need a slot.
Indeed, the (generic) "blob" name misled me. Would it make sense to use a name with "secid" in it instead?
Shouldn't the slot field be set to LSMBLOB_NOT_NEEDED (-3) then (instead of the implicit 0)?
Quoted 7 lines
Quoted 5 lines
You can run the Landlock tests please?
make -Ctools/testing/selftestsTARGETS=landlockgen_tartar -xf kselftest.tar.gz && ./run_kselftest.shSure. I'll add them to my routine.
Thanks.
Quoted 50 lines
Quoted 41 lines
Quoted 8 lines
security/landlock/setup.c +2
-0+}; +static int __init landlock_init (void ) {landlock_add_cred_hooks ();[...]
Quoted 31 lines
security/security.c +9
-1@@ -344,6 +344,7 @@ static void __init ordered_lsm_init(void)init_debug ("sock blob size = %d\n" , blob_sizes.lbs_sock);init_debug ("superblock blob size = %d\n" , blob_sizes.lbs_superblock);init_debug ("task blob size = %d\n" , blob_sizes.lbs_task); +init_debug ("lsmblob size = %zu\n" ,sizeof (struct lsmblob ));/* * Create any kmem_caches needed for blobs @@ -471,21 +472,36 @@ static int lsm_append(const char *new, char **result)return 0 ; } +/* +* Current index to use while initializing the lsmblob secid list. +*/ +static int lsm_slot __lsm_ro_after_init; +/** * security_add_hooks - Add a modules hooks to the hook lists. * @hooks: the hooks to add * @count: the number of hooks to add -* @lsm: the name of the security module +* @lsmid: the identification information for the security module * * Each LSM has to register its hooks with the infrastructure. +* If the LSM is using hooks that export secids allocate a slot +* for it in the lsmblob. Quoted 8 lines
security/security.c static int lsm_append(const char *new, char **result) · line 494
Quoted 7 lines
security/security.c +1
-1*/ void __init security_add_hooks (struct security_hook_list *hooks,int count, -char *lsm) +struct lsm_id *lsmid) {int i;Could you add a
WARN_ON(!lsmid->slot || !lsmid->name)here?
Quoted 1 line
Yes. That's reasonable.
I guess my above comment makes sense if lsmid->slot should not be zero but LSMBLOB_NOT_NEEDED instead, otherwise the Landlock lsmid would throw a warning.
Quoted 18 lines
Quoted 17 lines
Quoted 17 lines
security/security.c +9
-1+if (lsmid->slot ==LSMBLOB_NEEDED ) { +if (lsm_slot >=LSMBLOB_ENTRIES ) +panic ("%s Too many LSMs registered.\n" , __func__); + lsmid->slot = lsm_slot++; +init_debug ("%s assigned lsmblob slot %d\n" , lsmid->lsm, + lsmid->slot); + } +for (i =0 ; i < count; i++) { - hooks[i].lsm = lsm; + hooks[i].lsmid = lsmid;hlist_add_tail_rcu (&hooks[i].list, hooks[i].head); }
Casey Schaufler
maintainerPATCH v26 3/25 LSM: provide lsm name and id slot mappings
In reply to Casey SchauflerProvide interfaces to map LSM slot numbers and LSM names. Update the LSM registration code to save this information.
--- include/linux/security.h | 4 ++++ security/security.c | 45 ++++++++++++++++++++++++++++++++++++++++ 2 files changed, 49 insertions(+)
include/linux/security.h +4 -0
@@ -192,6 +192,10 @@ static inline bool lsmblob_equal(struct lsmblob *bloba, struct lsmblob *blobb)return !memcmp (bloba, blobb,sizeof (*bloba)); } +/* Map lsm names to blob slot numbers */ +extern int lsm_name_to_slot (char *name); +extern const char *lsm_slot_to_name (int slot); +/* These functions are in security/commoncap.c */ extern int cap_capable (const struct cred *cred,struct user_namespace *ns,int cap,unsigned int opts);
security/security.c +45 -0
57 lines not on this page · raw message
Signature · 1 line
2.29.2
Re: [PATCH v26 03/25] LSM: provide lsm name and id slot mappings
Quoted 4 lines · Casey Schaufler
Provide interfaces to map LSM slot numbers and LSM names. Update the LSM registration code to save this information.
Signed-off-by Casey Schaufler <[email protected]>
Seems a reasonable set of utilities. :)
Signature · 1 line
Kees Cook
Re: [PATCH v26 03/25] LSM: provide lsm name and id slot mappings
In reply to Casey SchauflerQuoted 9 lines · Casey Schaufler
Provide interfaces to map LSM slot numbers and LSM names. Update the LSM registration code to save this information.
Signed-off-by Casey Schaufler <[email protected]>--- include/linux/security.h | 4 ++++ security/security.c | 45 ++++++++++++++++++++++++++++++++++++++++ 2 files changed, 49 insertions(+)
--
paul moore
www.paul-moore.com
Casey Schaufler
maintainerPATCH v26 4/25 IMA: avoid label collisions with stacked LSMs
In reply to Casey SchauflerIntegrity measurement may filter on security module information and needs to be clear in the case of multiple active security modules which applies. Provide a boot option ima_rules_lsm= to allow the user to specify an active securty module to apply filters to. If not specified, use the first registered module that supports the audit_rule_match() LSM hook. Allow the user to specify in the IMA policy an lsm= option to specify the security module to use for a particular rule.
To: Mimi Zohar <[email protected]>
To: [email protected]
--- Documentation/ABI/testing/ima_policy | 8 ++- security/integrity/ima/ima_policy.c | 77 ++++++++++++++++++++-------- 2 files changed, 62 insertions(+), 23 deletions(-)
Documentation/ABI/testing/ima_policy +7 -1
@@ -25,7 +25,7 @@ Description: base: [[func=] [mask=] [fsmagic=] [fsuuid=] [uid=] [euid=] [fowner=] [fsname=]] lsm: [[subj_user=] [subj_role=] [subj_type=] - [obj_user=] [obj_role=] [obj_type=]] + [obj_user=] [obj_role=] [obj_type=] [lsm=]] option: [[appraise_type=]] [template=] [permit_directio] [appraise_flag=] [keyrings=] base:
@@ -117,6 +117,12 @@ Description: measure subj_user=_ func=FILE_CHECK mask=MAY_READ + It is possible to explicitly specify which security + module a rule applies to using lsm=. If the security + modules specified is not active on the system the rule + will be rejected. If lsm= is not specified the first + security module registered on the system will be assumed. + Example of measure rules using alternate PCRs:: measure func=KEXEC_KERNEL_CHECK pcr=4
security/integrity/ima/ima_policy.c +55 -22
185 lines not on this page · raw message
Signature · 1 line
2.29.2
Re: [PATCH v26 04/25] IMA: avoid label collisions with stacked LSMs
Quoted 69 lines · Casey Schaufler
Integrity measurement may filter on security module information and needs to be clear in the case of multiple active security modules which applies. Provide a boot option ima_rules_lsm= to allow the user to specify an active securty module to apply filters to. If not specified, use the first registered module that supports the
audit_rule_match()LSM hook. Allow the user to specify in the IMA policy an lsm= option to specify the security module to use for a particular rule.Signed-off-by Casey Schaufler <[email protected]>To: Mimi Zohar <[email protected]>
To: [email protected]--- Documentation/ABI/testing/ima_policy | 8 ++- security/integrity/ima/ima_policy.c | 77 ++++++++++++++++++++-------- 2 files changed, 62 insertions(+), 23 deletions(-)Documentation/ABI/testing/ima_policy +7
-1@@ -25,7 +25,7 @@ Description: base: [[func=] [mask=] [fsmagic=] [fsuuid=] [uid=] [euid=] [fowner=] [fsname=]] lsm: [[subj_user=] [subj_role=] [subj_type=] - [obj_user=] [obj_role=] [obj_type=]] + [obj_user=] [obj_role=] [obj_type=] [lsm=]] option: [[appraise_type=]] [template=] [permit_directio] [appraise_flag=] [keyrings=] base:@@ -117,6 +117,12 @@ Description: measure subj_user=_ func=FILE_CHECK mask=MAY_READ + It is possible to explicitly specify which security + module a rule applies to using lsm=. If the security + modules specified is not active on the system the rule + will be rejected. If lsm= is not specified the first + security module registered on the system will be assumed. + Example of measure rules using alternate PCRs:: measure func=KEXEC_KERNEL_CHECK pcr=4security/integrity/ima/ima_policy.c +5
-3@@ -79,8 +79,9 @@ struct ima_rule_entry {bool (*uid_op)(kuid_t ,kuid_t );/* Handlers for operators */ bool (*fowner_op)(kuid_t ,kuid_t );/* uid_eq(), uid_gt(), uid_lt() */ int pcr; +int which_lsm;/* which of the rules to use */ struct { -void *rules[LSMBLOB_ENTRIES ];/* LSM file metadata specific */ +void *rule;/* LSM file metadata specific */ char *args_p;/* audit value */ int type;/* audit type */ } lsm[MAX_LSM_RULES ];@@ -92,17 +93,15 @@ struct ima_rule_entry {/** * ima_lsm_isset - Is a rule set for any of the active security modules -* @rules: The set of IMA rules to check +* @entry: the rule entry to examine +* @lsm_rule: the specific rule type in question * -* If a rule is set for any LSM return true, otherwise return false. +* If a rule is set return true, otherwise return false. */
security/integrity/ima/ima_policy.c struct ima_rule_entry { · line 104
Quoted 10 lines
security/integrity/ima/ima_policy.c +3
-6-static inline bool ima_lsm_isset (void *rules[]) +static inline bool ima_lsm_isset (struct ima_rule_entry *entry,int lsm_rule) { -int i; - -for (i =0 ; i <LSMBLOB_ENTRIES ; i++) -if (rules[i]) -return true; +if (entry->lsm[lsm_rule].rule) +return true;
I'd be happier if this retained the MAX_LSM_RULES bounds-check.
Quoted 172 lines
security/integrity/ima/ima_policy.c +47
-13return false; }@@ -282,6 +281,20 @@ static int __init default_appraise_policy_setup(char *str) }__setup ("ima_appraise_tcb" , default_appraise_policy_setup); +static int ima_rules_lsm__ro_after_init ; + +static int __init ima_rules_lsm_init (char *str) +{ + ima_rules_lsm =lsm_name_to_slot (str); +if (ima_rules_lsm <0 ) { + ima_rules_lsm =0 ; +pr_err ("rule lsm \"%s\" not registered" , str); + } + +return 1 ; +} +__setup ("ima_rules_lsm=" , ima_rules_lsm_init); +static struct ima_rule_opt_list *ima_alloc_rule_opt_list (const substring_t *src) {struct ima_rule_opt_list *opt_list;@@ -351,11 +364,10 @@ static void ima_free_rule_opt_list(struct ima_rule_opt_list *opt_list)static void ima_lsm_free_rule (struct ima_rule_entry *entry) {int i; -int r;for (i =0 ; i <MAX_LSM_RULES ; i++) { -for (r =0 ; r <LSMBLOB_ENTRIES ; r++) -ima_filter_rule_free (entry->lsm[i].rules[r]); +if (entry->lsm[i].rule) +ima_filter_rule_free (entry->lsm[i].rule);kfree (entry->lsm[i].args_p); } }@@ -406,8 +418,8 @@ static struct ima_rule_entry *ima_lsm_copy_rule(struct ima_rule_entry *entry)ima_filter_rule_init (nentry->lsm[i].type, Audit_equal, nentry->lsm[i].args_p, - &nentry->lsm[i].rules[0 ]); -if (!ima_lsm_isset (nentry->lsm[i].rules)) + &nentry->lsm[i].rule); +if (!ima_lsm_isset (nentry, i))pr_warn ("rule for LSM \'%s\' is undefined\n" , nentry->lsm[i].args_p); }@@ -596,7 +608,7 @@ static bool ima_match_rules(struct ima_rule_entry *rule,int rc =0 ;u32 osid; -if (!ima_lsm_isset (rule->lsm[i].rules)) { +if (!ima_lsm_isset (rule, i)) {if (!rule->lsm[i].args_p)continue ;else @@ -609,14 +621,14 @@ static bool ima_match_rules(struct ima_rule_entry *rule,security_inode_getsecid (inode, &osid); rc =ima_filter_rule_match (osid, rule->lsm[i].type, Audit_equal, - rule->lsm[i].rules); + rule->lsm[i].rule);break ;case LSM_SUBJ_USER :case LSM_SUBJ_ROLE :case LSM_SUBJ_TYPE : rc =ima_filter_rule_match (secid, rule->lsm[i].type, Audit_equal, - rule->lsm[i].rules); + rule->lsm[i].rule);break ;default :break ;@@ -966,7 +978,7 @@ enum { Opt_uid_lt, Opt_euid_lt, Opt_fowner_lt, Opt_appraise_type, Opt_appraise_flag, Opt_permit_directio, Opt_pcr, Opt_template, Opt_keyrings, - Opt_label, Opt_err + Opt_lsm, Opt_label, Opt_err }; static const match_table_t policy_tokens = {@@ -1004,6 +1016,7 @@ static const match_table_t policy_tokens = { {Opt_template, "template=%s"}, {Opt_keyrings, "keyrings=%s"}, {Opt_label, "label=%s"}, + {Opt_lsm, "lsm=%s"}, {Opt_err, NULL} };@@ -1012,7 +1025,7 @@ static int ima_lsm_rule_init(struct ima_rule_entry *entry, {int result; -if (ima_lsm_isset (entry->lsm[lsm_rule].rules)) +if (ima_lsm_isset (entry, lsm_rule))return -EINVAL ; entry->lsm[lsm_rule].args_p =match_strdup (args);@@ -1022,8 +1035,8 @@ static int ima_lsm_rule_init(struct ima_rule_entry *entry, entry->lsm[lsm_rule].type = audit_type; result =ima_filter_rule_init (entry->lsm[lsm_rule].type, Audit_equal, entry->lsm[lsm_rule].args_p, - &entry->lsm[lsm_rule].rules[0 ]); -if (!ima_lsm_isset (entry->lsm[lsm_rule].rules)) { + &entry->lsm[lsm_rule].rule); +if (!ima_lsm_isset (entry, lsm_rule)) {pr_warn ("rule for LSM \'%s\' is undefined\n" , entry->lsm[lsm_rule].args_p);@@ -1561,6 +1574,19 @@ static int ima_parse_rule(char *rule, struct ima_rule_entry *entry) &(template_desc->num_fields)); entry->template = template_desc;break ; +case Opt_lsm: + result =lsm_name_to_slot (args[0 ].from); +if (result ==LSMBLOB_INVALID ) { +int i; + +for (i =0 ; i <MAX_LSM_RULES ; i++) + entry->lsm[i].args_p =NULL ; + result = -EINVAL ; +break ; + } + entry->which_lsm = result; + result =0 ; +break ;case Opt_err:ima_log_string (ab,"UNKNOWN" , p); result = -EINVAL ;@@ -1597,6 +1623,7 @@ ssize_t ima_parse_add_rule(char *rule)struct ima_rule_entry *entry;ssize_t result, len;int audit_info =0 ; +int i; p =strsep (&rule,"\n" ); len =strlen (p) +1 ;@@ -1614,6 +1641,9 @@ ssize_t ima_parse_add_rule(char *rule)INIT_LIST_HEAD (&entry->list); +for (i =0 ; i <MAX_LSM_RULES ; i++) + entry->which_lsm = ima_rules_lsm; + result =ima_parse_rule (p, entry);if (result) {ima_free_rule (entry);@@ -1830,7 +1860,7 @@ int ima_policy_show(struct seq_file *m, void *v) }for (i =0 ; i <MAX_LSM_RULES ; i++) { -if (ima_lsm_isset (entry->lsm[i].rules)) { +if (ima_lsm_isset (entry, i)) {switch (i) {case LSM_OBJ_USER : seq_printf(m, pt(Opt_obj_user),@@ -1872,6 +1902,9 @@ int ima_policy_show(struct seq_file *m, void *v)seq_puts (m,"appraise_flag=check_blacklist " );if (entry->flags &IMA_PERMIT_DIRECTIO )seq_puts (m,"permit_directio " ); +if (entry->which_lsm >=0 ) +seq_printf (m,pt (Opt_lsm), +lsm_slot_to_name (entry->which_lsm));rcu_read_unlock ();seq_puts (m,"\n" );return 0 ;Signature · 1 line
2.29.2
Otherwise, this looks correct. With bounds checking:
Signature · 1 line
Kees Cook
Casey Schaufler
maintainerPATCH v26 5/25 LSM: Use lsmblob in security_audit_rule_match
In reply to Casey SchauflerChange the secid parameter of security_audit_rule_match to a lsmblob structure pointer. Pass the entry from the lsmblob structure for the approprite slot to the LSM hook.
Change the users of security_audit_rule_match to use the lsmblob instead of a u32. The scaffolding function lsmblob_init() fills the blob with the value of the old secid, ensuring that it is available to the appropriate module hook. The sources of the secid, security_task_getsecid() and security_inode_getsecid(), will be converted to use the blob structure later in the series. At the point the use of lsmblob_init() is dropped.
To: Mimi Zohar <[email protected]>
--- include/linux/security.h | 7 ++++--- kernel/auditfilter.c | 6 ++++-- kernel/auditsc.c | 16 +++++++++++----- security/integrity/ima/ima.h | 4 ++-- security/integrity/ima/ima_policy.c | 7 +++++-- security/security.c | 10 ++++++++-- 6 files changed, 34 insertions(+), 16 deletions(-)
include/linux/security.h +4 -3
@@ -1944,7 +1944,8 @@ static inline int security_key_getsecurity(struct key *key, char **_buffer)#ifdef CONFIG_SECURITY int security_audit_rule_init (u32 field,u32 op,char *rulestr,void **lsmrule);int security_audit_rule_known (struct audit_krule *krule); -int security_audit_rule_match (u32 secid,u32 field,u32 op,void **lsmrule); +int security_audit_rule_match (struct lsmblob *blob,u32 field,u32 op, +void **lsmrule);void security_audit_rule_free (void **lsmrule);#else
@@ -1960,8 +1961,8 @@ static inline int security_audit_rule_known(struct audit_krule *krule)return 0 ; } -static inline int security_audit_rule_match (u32 secid,u32 field,u32 op, -void **lsmrule) +static inline int security_audit_rule_match (struct lsmblob *blob,u32 field, +u32 op,void **lsmrule) {return 0 ; }
kernel/auditfilter.c +4 -2
@@ -1331,6 +1331,7 @@ int audit_filter(int msgtype, unsigned int listtype)struct audit_field *f = &e->rule.fields[i];pid_t pid;u32 sid; +struct lsmblob blob;switch (f->type) {case AUDIT_PID :
@@ -1362,8 +1363,9 @@ int audit_filter(int msgtype, unsigned int listtype)if (f->lsm_isset) {security_task_getsecid_subj (current, &sid); - result =security_audit_rule_match (sid, - f->type, f->op, +lsmblob_init (&blob, sid); + result =security_audit_rule_match ( + &blob, f->type, f->op, f->lsm_rules); }break ;
kernel/auditsc.c +11 -5
@@ -472,6 +472,7 @@ static int audit_filter_rules(struct task_struct *tsk,const struct cred *cred;int i, need_sid =1 ;u32 sid; +struct lsmblob blob;unsigned int sessionid; cred =rcu_dereference_check (tsk->cred, tsk == current || task_creation);
@@ -670,8 +671,10 @@ static int audit_filter_rules(struct task_struct *tsk,security_task_getsecid_subj (tsk, &sid); need_sid =0 ; } - result =security_audit_rule_match (sid, f->type, - f->op, f->lsm_rules); +lsmblob_init (&blob, sid); + result =security_audit_rule_match (&blob, + f->type, f->op, + f->lsm_rules); }break ;case AUDIT_OBJ_USER :
@@ -684,15 +687,17 @@ static int audit_filter_rules(struct task_struct *tsk,if (f->lsm_isset) {/* Find files that match */ if (name) { +lsmblob_init (&blob, name->osid); result =security_audit_rule_match ( - name->osid, + &blob, f->type, f->op, f->lsm_rules); }else if (ctx) {list_for_each_entry (n, &ctx->names_list, list) { +lsmblob_init (&blob, name->osid);if (security_audit_rule_match ( - n->osid, + &blob, f->type, f->op, f->lsm_rules)) {
@@ -704,7 +709,8 @@ static int audit_filter_rules(struct task_struct *tsk,/* Find ipc objects that match */ if (!ctx || ctx->type !=AUDIT_IPC )break ; -if (security_audit_rule_match (ctx->ipc.osid, +lsmblob_init (&blob, ctx->ipc.osid); +if (security_audit_rule_match (&blob, f->type, f->op, f->lsm_rules)) ++result;
security/integrity/ima/ima.h +2 -2
@@ -433,8 +433,8 @@ static inline void ima_filter_rule_free(void *lsmrule) { } -static inline int ima_filter_rule_match (u32 secid,u32 field,u32 op, -void *lsmrule) +static inline int ima_filter_rule_match (struct lsmblob *blob,u32 field, +u32 op,void *lsmrule) {return -EINVAL ; }
security/integrity/ima/ima_policy.c +5 -2
@@ -607,6 +607,7 @@ static bool ima_match_rules(struct ima_rule_entry *rule,for (i =0 ; i <MAX_LSM_RULES ; i++) {int rc =0 ;u32 osid; +struct lsmblob lsmdata;if (!ima_lsm_isset (rule, i)) {if (!rule->lsm[i].args_p)
@@ -619,14 +620,16 @@ static bool ima_match_rules(struct ima_rule_entry *rule,case LSM_OBJ_ROLE :case LSM_OBJ_TYPE :security_inode_getsecid (inode, &osid); - rc =ima_filter_rule_match (osid, rule->lsm[i].type, +lsmblob_init (&lsmdata, osid); + rc =ima_filter_rule_match (&lsmdata, rule->lsm[i].type, Audit_equal, rule->lsm[i].rule);break ;case LSM_SUBJ_USER :case LSM_SUBJ_ROLE :case LSM_SUBJ_TYPE : - rc =ima_filter_rule_match (secid, rule->lsm[i].type, +lsmblob_init (&lsmdata, secid); + rc =ima_filter_rule_match (&lsmdata, rule->lsm[i].type, Audit_equal, rule->lsm[i].rule);break ;
security/security.c +8 -2
@@ -2669,11 +2669,14 @@ void security_audit_rule_free(void **lsmrule)hlist_for_each_entry (hp, &security_hook_heads.audit_rule_free, list) {if (WARN_ON (hp->lsmid->slot <0 || hp->lsmid->slot >= lsm_slot))continue ; +if (lsmrule[hp->lsmid->slot] ==NULL ) +continue ; hp->hook.audit_rule_free (lsmrule[hp->lsmid->slot]); } } -int security_audit_rule_match (u32 secid,u32 field,u32 op,void **lsmrule) +int security_audit_rule_match (struct lsmblob *blob,u32 field,u32 op, +void **lsmrule) {struct security_hook_list *hp;int rc;
@@ -2681,7 +2684,10 @@ int security_audit_rule_match(u32 secid, u32 field, u32 op, void **lsmrule)hlist_for_each_entry (hp, &security_hook_heads.audit_rule_match, list) {if (WARN_ON (hp->lsmid->slot <0 || hp->lsmid->slot >= lsm_slot))continue ; - rc = hp->hook.audit_rule_match (secid, field, op, +if (lsmrule[hp->lsmid->slot] ==NULL ) +continue ; + rc = hp->hook.audit_rule_match (blob->secid[hp->lsmid->slot], + field, op, &lsmrule[hp->lsmid->slot]);if (rc)return rc;
Signature · 1 line
2.29.2
Casey Schaufler
maintainerPATCH v26 6/25 LSM: Use lsmblob in security_kernel_act_as
In reply to Casey SchauflerChange the security_kernel_act_as interface to use a lsmblob structure in place of the single u32 secid in support of module stacking. Change its only caller, set_security_override, to do the same. Change that one's only caller, set_security_override_from_ctx, to call it with the new parameter type.
The security module hook is unchanged, still taking a secid. The infrastructure passes the correct entry from the lsmblob. lsmblob_init() is used to fill the lsmblob structure, however this will be removed later in the series when security_secctx_to_secid() is updated to provide a lsmblob instead of a secid.
To: David Howells <[email protected]>
--- include/linux/cred.h | 3 ++- include/linux/security.h | 5 +++-- kernel/cred.c | 10 ++++++---- security/security.c | 14 ++++++++++++-- 4 files changed, 23 insertions(+), 9 deletions(-)
include/linux/cred.h +2 -1
@@ -18,6 +18,7 @@struct cred ;struct inode ; +struct lsmblob ;/* * COW Supplementary groups list
@@ -164,7 +165,7 @@ extern const struct cred *override_creds(const struct cred *);extern void revert_creds (const struct cred *);extern struct cred *prepare_kernel_cred (struct task_struct *);extern int change_create_files_as (struct cred *,struct inode *); -extern int set_security_override (struct cred *,u32 ); +extern int set_security_override (struct cred *,struct lsmblob *);extern int set_security_override_from_ctx (struct cred *,const char *);extern int set_create_files_as (struct cred *,struct inode *);extern int cred_fscmp (const struct cred *,const struct cred *);
include/linux/security.h +3 -2
@@ -461,7 +461,7 @@ void security_cred_free(struct cred *cred);int security_prepare_creds (struct cred *new,const struct cred *old,gfp_t gfp);void security_transfer_creds (struct cred *new,const struct cred *old);void security_cred_getsecid (const struct cred *c,u32 *secid); -int security_kernel_act_as (struct cred *new,u32 secid); +int security_kernel_act_as (struct cred *new,struct lsmblob *blob);int security_kernel_create_files_as (struct cred *new,struct inode *inode);int security_kernel_module_request (char *kmod_name);int security_kernel_load_data (enum kernel_load_data_id id,bool contents);
@@ -1103,7 +1103,8 @@ static inline void security_transfer_creds(struct cred *new, { } -static inline int security_kernel_act_as (struct cred *cred,u32 secid) +static inline int security_kernel_act_as (struct cred *cred, +struct lsmblob *blob) {return 0 ; }
kernel/cred.c +6 -4
@@ -733,14 +733,14 @@ EXPORT_SYMBOL(prepare_kernel_cred);/** * set_security_override - Set the security ID in a set of credentials * @new: The credentials to alter -* @secid: The LSM security ID to set +* @blob: The LSM security information to set * * Set the LSM security ID in a set of credentials so that the subjective * security is overridden when an alternative set of credentials is used. */ -int set_security_override (struct cred *new,u32 secid) +int set_security_override (struct cred *new,struct lsmblob *blob) { -return security_kernel_act_as (new, secid); +return security_kernel_act_as (new, blob); }EXPORT_SYMBOL (set_security_override);
@@ -756,6 +756,7 @@ EXPORT_SYMBOL(set_security_override);*/ int set_security_override_from_ctx (struct cred *new,const char *secctx) { +struct lsmblob blob;u32 secid;int ret;
@@ -763,7 +764,8 @@ int set_security_override_from_ctx(struct cred *new, const char *secctx)if (ret <0 )return ret; -return set_security_override (new, secid); +lsmblob_init (&blob, secid); +return set_security_override (new, &blob); }EXPORT_SYMBOL (set_security_override_from_ctx);
security/security.c +12 -2
@@ -1796,9 +1796,19 @@ void security_cred_getsecid(const struct cred *c, u32 *secid) }EXPORT_SYMBOL (security_cred_getsecid); -int security_kernel_act_as (struct cred *new,u32 secid) +int security_kernel_act_as (struct cred *new,struct lsmblob *blob) { -return call_int_hook (kernel_act_as,0 , new, secid); +struct security_hook_list *hp; +int rc; + +hlist_for_each_entry (hp, &security_hook_heads.kernel_act_as, list) { +if (WARN_ON (hp->lsmid->slot <0 || hp->lsmid->slot >= lsm_slot)) +continue ; + rc = hp->hook.kernel_act_as (new, blob->secid[hp->lsmid->slot]); +if (rc !=0 ) +return rc; + } +return 0 ; }int security_kernel_create_files_as (struct cred *new,struct inode *inode)
Signature · 1 line
2.29.2
Casey Schaufler
maintainerPATCH v26 7/25 LSM: Use lsmblob in security_secctx_to_secid
In reply to Casey SchauflerChange the security_secctx_to_secid interface to use a lsmblob structure in place of the single u32 secid in support of module stacking. Change its callers to do the same.
The security module hook is unchanged, still passing back a secid. The infrastructure passes the correct entry from the lsmblob.
To: Pablo Neira Ayuso <[email protected]>
--- include/linux/security.h | 26 ++++++++++++++++++-- kernel/cred.c | 4 +--- net/netfilter/nft_meta.c | 10 ++++---- net/netfilter/xt_SECMARK.c | 7 +++++- net/netlabel/netlabel_unlabeled.c | 23 +++++++++++------- security/security.c | 40 ++++++++++++++++++++++++++----- 6 files changed, 85 insertions(+), 25 deletions(-)
include/linux/security.h +24 -2
@@ -196,6 +196,27 @@ static inline bool lsmblob_equal(struct lsmblob *bloba, struct lsmblob *blobb)extern int lsm_name_to_slot (char *name);extern const char *lsm_slot_to_name (int slot); +/** +* lsmblob_value - find the first non-zero value in an lsmblob structure. +* @blob: Pointer to the data +* +* This needs to be used with extreme caution, as the cases where +* it is appropriate are rare. +* +* Return the first secid value set in the lsmblob. +* There should only be one. +*/ +static inline u32 lsmblob_value (const struct lsmblob *blob) +{ +int i; + +for (i =0 ; i <LSMBLOB_ENTRIES ; i++) +if (blob->secid[i]) +return blob->secid[i]; + +return 0 ; +} +/* These functions are in security/commoncap.c */ extern int cap_capable (const struct cred *cred,struct user_namespace *ns,int cap,unsigned int opts);
@@ -527,7 +548,8 @@ int security_setprocattr(const char *lsm, const char *name, void *value,int security_netlink_send (struct sock *sk,struct sk_buff *skb);int security_ismaclabel (const char *name);int security_secid_to_secctx (u32 secid,char **secdata,u32 *seclen); -int security_secctx_to_secid (const char *secdata,u32 seclen,u32 *secid); +int security_secctx_to_secid (const char *secdata,u32 seclen, +struct lsmblob *blob);void security_release_secctx (char *secdata,u32 seclen);void security_inode_invalidate_secctx (struct inode *inode);int security_inode_notifysecctx (struct inode *inode,void *ctx,u32 ctxlen);
@@ -1382,7 +1404,7 @@ static inline int security_secid_to_secctx(u32 secid, char **secdata, u32 *seclestatic inline int security_secctx_to_secid (const char *secdata,u32 seclen, -u32 *secid) +struct lsmblob *blob) {return -EOPNOTSUPP ; }
kernel/cred.c +1 -3
@@ -757,14 +757,12 @@ EXPORT_SYMBOL(set_security_override);int set_security_override_from_ctx (struct cred *new,const char *secctx) {struct lsmblob blob; -u32 secid;int ret; - ret =security_secctx_to_secid (secctx,strlen (secctx), &secid); + ret =security_secctx_to_secid (secctx,strlen (secctx), &blob);if (ret <0 )return ret; -lsmblob_init (&blob, secid);return set_security_override (new, &blob); }EXPORT_SYMBOL (set_security_override_from_ctx);
net/netfilter/nft_meta.c +5 -5
@@ -809,21 +809,21 @@ static const struct nla_policy nft_secmark_policy[NFTA_SECMARK_MAX + 1] = {static int nft_secmark_compute_secid (struct nft_secmark *priv) { -u32 tmp_secid =0 ; +struct lsmblob blob;int err; - err =security_secctx_to_secid (priv->ctx,strlen (priv->ctx), &tmp_secid); + err =security_secctx_to_secid (priv->ctx,strlen (priv->ctx), &blob);if (err)return err; -if (!tmp_secid) +if (!lsmblob_is_set (&blob))return -ENOENT ; - err =security_secmark_relabel_packet (tmp_secid); + err =security_secmark_relabel_packet (lsmblob_value (&blob));if (err)return err; - priv->secid = tmp_secid; + priv->secid =lsmblob_value (&blob);return 0 ; }
net/netfilter/xt_SECMARK.c +6 -1
@@ -42,13 +42,14 @@ secmark_tg(struct sk_buff *skb, const struct xt_secmark_target_info_v1 *info)static int checkentry_lsm (struct xt_secmark_target_info_v1 *info) { +struct lsmblob blob;int err; info->secctx[SECMARK_SECCTX_MAX -1 ] ='\0' ; info->secid =0 ; err =security_secctx_to_secid (info->secctx,strlen (info->secctx), - &info->secid); + &blob);if (err) {if (err == -EINVAL ) pr_info_ratelimited("invalid security context \'%s\'\n",
@@ -56,6 +57,10 @@ static int checkentry_lsm(struct xt_secmark_target_info_v1 *info)return err; } +/* xt_secmark_target_info can't be changed to use lsmblobs because +* it is exposed as an API. Use lsmblob_value() to get the one +* value that got set by security_secctx_to_secid(). */ + info->secid =lsmblob_value (&blob);if (!info->secid) {pr_info_ratelimited ("unable to map security context \'%s\'\n" , info->secctx);
net/netlabel/netlabel_unlabeled.c +15 -8
@@ -882,7 +882,7 @@ static int netlbl_unlabel_staticadd(struct sk_buff *skb,void *addr;void *mask;u32 addr_len; -u32 secid; +struct lsmblob blob;struct netlbl_audit audit_info;/* Don't allow users to add both IPv4 and IPv6 addresses for a
@@ -906,13 +906,18 @@ static int netlbl_unlabel_staticadd(struct sk_buff *skb, ret_val =security_secctx_to_secid (nla_data (info->attrs[NLBL_UNLABEL_A_SECCTX ]),nla_len (info->attrs[NLBL_UNLABEL_A_SECCTX ]), - &secid); + &blob);if (ret_val !=0 )return ret_val; +/* netlbl_unlhsh_add will be changed to pass a struct lsmblob * +* instead of a u32 later in this patch set. security_secctx_to_secid() +* will only be setting one entry in the lsmblob struct, so it is +* safe to use lsmblob_value() to get that one value. */ +return netlbl_unlhsh_add (&init_net, - dev_name, addr, mask, addr_len, secid, - &audit_info); + dev_name, addr, mask, addr_len, +lsmblob_value (&blob), &audit_info); }/**
@@ -933,7 +938,7 @@ static int netlbl_unlabel_staticadddef(struct sk_buff *skb,void *addr;void *mask;u32 addr_len; -u32 secid; +struct lsmblob blob;struct netlbl_audit audit_info;/* Don't allow users to add both IPv4 and IPv6 addresses for a
@@ -955,13 +960,15 @@ static int netlbl_unlabel_staticadddef(struct sk_buff *skb, ret_val =security_secctx_to_secid (nla_data (info->attrs[NLBL_UNLABEL_A_SECCTX ]),nla_len (info->attrs[NLBL_UNLABEL_A_SECCTX ]), - &secid); + &blob);if (ret_val !=0 )return ret_val; +/* security_secctx_to_secid() will only put one secid into the lsmblob +* so it's safe to use lsmblob_value() to get the secid. */ return netlbl_unlhsh_add (&init_net, -NULL , addr, mask, addr_len, secid, - &audit_info); +NULL , addr, mask, addr_len, +lsmblob_value (&blob), &audit_info); }/**
security/security.c +34 -6
@@ -2191,10 +2191,22 @@ int security_secid_to_secctx(u32 secid, char **secdata, u32 *seclen) }EXPORT_SYMBOL (security_secid_to_secctx); -int security_secctx_to_secid (const char *secdata,u32 seclen,u32 *secid) +int security_secctx_to_secid (const char *secdata,u32 seclen, +struct lsmblob *blob) { - *secid =0 ; -return call_int_hook (secctx_to_secid,0 , secdata, seclen, secid); +struct security_hook_list *hp; +int rc; + +lsmblob_init (blob,0 ); +hlist_for_each_entry (hp, &security_hook_heads.secctx_to_secid, list) { +if (WARN_ON (hp->lsmid->slot <0 || hp->lsmid->slot >= lsm_slot)) +continue ; + rc = hp->hook.secctx_to_secid (secdata, seclen, + &blob->secid[hp->lsmid->slot]); +if (rc !=0 ) +return rc; + } +return 0 ; }EXPORT_SYMBOL (security_secctx_to_secid);
@@ -2345,10 +2357,26 @@ int security_socket_getpeersec_stream(struct socket *sock, char __user *optval, optval, optlen, len); } -int security_socket_getpeersec_dgram (struct socket *sock,struct sk_buff *skb,u32 *secid) +int security_socket_getpeersec_dgram (struct socket *sock,struct sk_buff *skb, +u32 *secid) { -return call_int_hook (socket_getpeersec_dgram, -ENOPROTOOPT , sock, - skb, secid); +struct security_hook_list *hp; +int rc = -ENOPROTOOPT ; + +/* +* Only one security module should provide a real hook for +* this. A stub or bypass like is used in BPF should either +* (somehow) leave rc unaltered or return -ENOPROTOOPT. +*/ +hlist_for_each_entry (hp, &security_hook_heads.socket_getpeersec_dgram, + list) { +if (WARN_ON (hp->lsmid->slot <0 || hp->lsmid->slot >= lsm_slot)) +continue ; + rc = hp->hook.socket_getpeersec_dgram (sock, skb, secid); +if (rc != -ENOPROTOOPT ) +break ; + } +return rc; }EXPORT_SYMBOL (security_socket_getpeersec_dgram);
Signature · 1 line
2.29.2
Re: [PATCH v26 07/25] LSM: Use lsmblob in security_secctx_to_secid
Quoted 8 lines · Casey Schaufler
Change the
security_secctx_to_secidinterface to use a lsmblob structure in place of the single u32 secid in support of module stacking. Change its callers to do the same.The security module hook is unchanged, still passing back a secid. The infrastructure passes the correct entry from the lsmblob.
Signed-off-by Casey Schaufler <[email protected]>
This looks like sane refactoring into the new blob type.
Signature · 1 line
Kees Cook
Re: [PATCH v26 07/25] LSM: Use lsmblob in security_secctx_to_secid
In reply to Casey SchauflerQuoted 12 lines · Casey Schaufler
Change the
security_secctx_to_secidinterface to use a lsmblob structure in place of the single u32 secid in support of module stacking. Change its callers to do the same.The security module hook is unchanged, still passing back a secid. The infrastructure passes the correct entry from the lsmblob.
To: Pablo Neira Ayuso <[email protected]>
Quoted 8 lines
--- include/linux/security.h | 26 ++++++++++++++++++-- kernel/cred.c | 4 +--- net/netfilter/nft_meta.c | 10 ++++---- net/netfilter/xt_SECMARK.c | 7 +++++- net/netlabel/netlabel_unlabeled.c | 23 +++++++++++------- security/security.c | 40 ++++++++++++++++++++++++++----- 6 files changed, 85 insertions(+), 25 deletions(-)
--
paul moore
www.paul-moore.com
Casey Schaufler
maintainerPATCH v26 8/25 LSM: Use lsmblob in security_secid_to_secctx
In reply to Casey SchauflerChange security_secid_to_secctx() to take a lsmblob as input instead of a u32 secid. It will then call the LSM hooks using the lsmblob element allocated for that module. The callers have been updated as well. This allows for the possibility that more than one module may be called upon to translate a secid to a string, as can occur in the audit code.
To: Pablo Neira Ayuso <[email protected]>
To: Paul Moore <[email protected]>
--- drivers/android/binder.c | 12 +++++++++- include/linux/security.h | 5 +++-- include/net/scm.h | 7 +++++- kernel/audit.c | 20 +++++++++++++++-- kernel/auditsc.c | 28 +++++++++++++++++++---- net/ipv4/ip_sockglue.c | 4 +++- net/netfilter/nf_conntrack_netlink.c | 14 ++++++++++-- net/netfilter/nf_conntrack_standalone.c | 4 +++- net/netfilter/nfnetlink_queue.c | 11 +++++++-- net/netlabel/netlabel_unlabeled.c | 30 +++++++++++++++++++++---- net/netlabel/netlabel_user.c | 6 ++--- security/security.c | 11 +++++---- 12 files changed, 123 insertions(+), 29 deletions(-)
drivers/android/binder.c +11 -1
@@ -2711,6 +2711,7 @@ static void binder_transaction(struct binder_proc *proc,if (target_node && target_node->txn_security_ctx) {u32 secid; +struct lsmblob blob;size_t added_size;/*
@@ -2723,7 +2724,16 @@ static void binder_transaction(struct binder_proc *proc,* case well anyway. */ security_task_getsecid_obj (proc->tsk, &secid); - ret =security_secid_to_secctx (secid, &secctx, &secctx_sz); +/* +* Later in this patch set security_task_getsecid() will +* provide a lsmblob instead of a secid. lsmblob_init +* is used to ensure that all the secids in the lsmblob +* get the value returned from security_task_getsecid(), +* which means that the one expected by +* security_secid_to_secctx() will be set. +*/ +lsmblob_init (&blob, secid); + ret =security_secid_to_secctx (&blob, &secctx, &secctx_sz);if (ret) { return_error =BR_FAILED_REPLY ; return_error_param = ret;
include/linux/security.h +3 -2
@@ -547,7 +547,7 @@ int security_setprocattr(const char *lsm, const char *name, void *value, size_t size);int security_netlink_send (struct sock *sk,struct sk_buff *skb);int security_ismaclabel (const char *name); -int security_secid_to_secctx (u32 secid,char **secdata,u32 *seclen); +int security_secid_to_secctx (struct lsmblob *blob,char **secdata,u32 *seclen);int security_secctx_to_secid (const char *secdata,u32 seclen,struct lsmblob *blob);void security_release_secctx (char *secdata,u32 seclen);
@@ -1397,7 +1397,8 @@ static inline int security_ismaclabel(const char *name)return 0 ; } -static inline int security_secid_to_secctx (u32 secid,char **secdata,u32 *seclen) +static inline int security_secid_to_secctx (struct lsmblob *blob, +char **secdata,u32 *seclen) {return -EOPNOTSUPP ; }
include/net/scm.h +6 -1
@@ -92,12 +92,17 @@ static __inline__ int scm_send(struct socket *sock, struct msghdr *msg,#ifdef CONFIG_SECURITY_NETWORK static inline void scm_passec (struct socket *sock,struct msghdr *msg,struct scm_cookie *scm) { +struct lsmblob lb;char *secdata;u32 seclen;int err;if (test_bit (SOCK_PASSSEC , &sock->flags)) { - err =security_secid_to_secctx (scm->secid, &secdata, &seclen); +/* There can only be one security module using the secid, +* and the infrastructure will know which it is. +*/ +lsmblob_init (&lb, scm->secid); + err =security_secid_to_secctx (&lb, &secdata, &seclen);if (!err) {put_cmsg (msg,SOL_SOCKET ,SCM_SECURITY , seclen, secdata);
kernel/audit.c +18 -2
@@ -1442,7 +1442,16 @@ static int audit_receive_msg(struct sk_buff *skb, struct nlmsghdr *nlh)case AUDIT_SIGNAL_INFO : len =0 ;if (audit_sig_sid) { - err =security_secid_to_secctx (audit_sig_sid, &ctx, &len); +struct lsmblob blob; + +/* +* lsmblob_init sets all values in the lsmblob +* to audit_sig_sid. This is temporary until +* audit_sig_sid is converted to a lsmblob, which +* happens later in this patch set. +*/ +lsmblob_init (&blob, audit_sig_sid); + err =security_secid_to_secctx (&blob, &ctx, &len);if (err)return err; }
@@ -2131,12 +2140,19 @@ int audit_log_task_context(struct audit_buffer *ab)unsigned len;int error;u32 sid; +struct lsmblob blob;security_task_getsecid_subj (current, &sid);if (!sid)return 0 ; - error =security_secid_to_secctx (sid, &ctx, &len); +/* +* lsmblob_init sets all values in the lsmblob to sid. +* This is temporary until security_task_getsecid is converted +* to use a lsmblob, which happens later in this patch set. +*/ +lsmblob_init (&blob, sid); + error =security_secid_to_secctx (&blob, &ctx, &len);if (error) {if (error != -EINVAL ) goto error_path;
kernel/auditsc.c +24 -4
@@ -671,6 +671,13 @@ static int audit_filter_rules(struct task_struct *tsk,security_task_getsecid_subj (tsk, &sid); need_sid =0 ; } +/* +* lsmblob_init sets all values in the lsmblob +* to sid. This is temporary until +* security_task_getsecid() is converted to +* provide a lsmblob, which happens later in +* this patch set. +*/ lsmblob_init (&blob, sid); result = security_audit_rule_match(&blob, f->type, f->op,
@@ -687,6 +694,13 @@ static int audit_filter_rules(struct task_struct *tsk,if (f->lsm_isset) {/* Find files that match */ if (name) { +/* +* lsmblob_init sets all values in the +* lsmblob to sid. This is temporary +* until name->osid is converted to a +* lsmblob, which happens later in +* this patch set. +*/ lsmblob_init (&blob, name->osid); result = security_audit_rule_match( &blob,
@@ -993,6 +1007,7 @@ static int audit_log_pid_context(struct audit_context *context, pid_t pid,char *ctx =NULL ;u32 len;int rc =0 ; +struct lsmblob blob; ab =audit_log_start (context,GFP_KERNEL ,AUDIT_OBJ_PID );if (!ab)
@@ -1002,7 +1017,8 @@ static int audit_log_pid_context(struct audit_context *context, pid_t pid,from_kuid (&init_user_ns, auid), from_kuid(&init_user_ns, uid), sessionid);if (sid) { -if (security_secid_to_secctx (sid, &ctx, &len)) { +lsmblob_init (&blob, sid); +if (security_secid_to_secctx (&blob, &ctx, &len)) {audit_log_format (ab," obj=(none)" ); rc =1 ; }else {
@@ -1245,7 +1261,10 @@ static void show_special(struct audit_context *context, int *call_panic)if (osid) {char *ctx =NULL ;u32 len; -if (security_secid_to_secctx (osid, &ctx, &len)) { +struct lsmblob blob; + +lsmblob_init (&blob, osid); +if (security_secid_to_secctx (&blob, &ctx, &len)) {audit_log_format (ab," osid=%u" , osid); *call_panic =1 ; }else {
@@ -1398,9 +1417,10 @@ static void audit_log_name(struct audit_context *context, struct audit_names *n,if (n->osid !=0 ) {char *ctx =NULL ;u32 len; +struct lsmblob blob; -if (security_secid_to_secctx ( - n->osid, &ctx, &len)) { +lsmblob_init (&blob, n->osid); +if (security_secid_to_secctx (&blob, &ctx, &len)) {audit_log_format (ab," osid=%u" , n->osid);if (call_panic) *call_panic =2 ;
net/ipv4/ip_sockglue.c +3 -1
@@ -130,6 +130,7 @@ static void ip_cmsg_recv_checksum(struct msghdr *msg, struct sk_buff *skb,static void ip_cmsg_recv_security (struct msghdr *msg,struct sk_buff *skb) { +struct lsmblob lb;char *secdata;u32 seclen, secid;int err;
@@ -138,7 +139,8 @@ static void ip_cmsg_recv_security(struct msghdr *msg, struct sk_buff *skb)if (err)return ; - err =security_secid_to_secctx (secid, &secdata, &seclen); +lsmblob_init (&lb, secid); + err =security_secid_to_secctx (&lb, &secdata, &seclen);if (err)return ;
net/netfilter/nf_conntrack_netlink.c +12 -2
@@ -338,8 +338,13 @@ static int ctnetlink_dump_secctx(struct sk_buff *skb, const struct nf_conn *ct)struct nlattr *nest_secctx;int len, ret;char *secctx; +struct lsmblob blob; - ret =security_secid_to_secctx (ct->secmark, &secctx, &len); +/* lsmblob_init() puts ct->secmark into all of the secids in blob. +* security_secid_to_secctx() will know which security module +* to use to create the secctx. */ +lsmblob_init (&blob, ct->secmark); + ret =security_secid_to_secctx (&blob, &secctx, &len);if (ret)return 0 ;
@@ -647,8 +652,13 @@ static inline int ctnetlink_secctx_size(const struct nf_conn *ct) {#ifdef CONFIG_NF_CONNTRACK_SECMARK int len, ret; +struct lsmblob blob; - ret =security_secid_to_secctx (ct->secmark,NULL , &len); +/* lsmblob_init() puts ct->secmark into all of the secids in blob. +* security_secid_to_secctx() will know which security module +* to use to create the secctx. */ +lsmblob_init (&blob, ct->secmark); + ret =security_secid_to_secctx (&blob,NULL , &len);if (ret)return 0 ;
net/netfilter/nf_conntrack_standalone.c +3 -1
@@ -175,8 +175,10 @@ static void ct_show_secctx(struct seq_file *s, const struct nf_conn *ct)int ret;u32 len;char *secctx; +struct lsmblob blob; - ret =security_secid_to_secctx (ct->secmark, &secctx, &len); +lsmblob_init (&blob, ct->secmark); + ret =security_secid_to_secctx (&blob, &secctx, &len);if (ret)return ;
net/netfilter/nfnetlink_queue.c +9 -2
@@ -305,13 +305,20 @@ static u32 nfqnl_get_sk_secctx(struct sk_buff *skb, char **secdata) {u32 seclen =0 ;#if IS_ENABLED (CONFIG_NETWORK_SECMARK ) +struct lsmblob blob; +if (!skb || !sk_fullsock (skb->sk))return 0 ;read_lock_bh (&skb->sk->sk_callback_lock); -if (skb->secmark) -security_secid_to_secctx (skb->secmark, secdata, &seclen); +if (skb->secmark) { +/* lsmblob_init() puts ct->secmark into all of the secids in +* blob. security_secid_to_secctx() will know which security +* module to use to create the secctx. */ +lsmblob_init (&blob, skb->secmark); +security_secid_to_secctx (&blob, secdata, &seclen); + }read_unlock_bh (&skb->sk->sk_callback_lock);#endif
net/netlabel/netlabel_unlabeled.c +26 -4
@@ -376,6 +376,7 @@ int netlbl_unlhsh_add(struct net *net,struct audit_buffer *audit_buf =NULL ;char *secctx =NULL ;u32 secctx_len; +struct lsmblob blob;if (addr_len !=sizeof (struct in_addr ) && addr_len !=sizeof (struct in6_addr ))
@@ -438,7 +439,11 @@ int netlbl_unlhsh_add(struct net *net, unlhsh_add_return:rcu_read_unlock ();if (audit_buf !=NULL ) { -if (security_secid_to_secctx (secid, +/* lsmblob_init() puts secid into all of the secids in blob. +* security_secid_to_secctx() will know which security module +* to use to create the secctx. */ +lsmblob_init (&blob, secid); +if (security_secid_to_secctx (&blob, &secctx, &secctx_len) ==0 ) {audit_log_format (audit_buf," sec_obj=%s" , secctx);
@@ -475,6 +480,7 @@ static int netlbl_unlhsh_remove_addr4(struct net *net,struct net_device *dev;char *secctx;u32 secctx_len; +struct lsmblob blob;spin_lock (&netlbl_unlhsh_lock); list_entry = netlbl_af4list_remove(addr->s_addr, mask->s_addr,
@@ -494,8 +500,13 @@ static int netlbl_unlhsh_remove_addr4(struct net *net, addr->s_addr, mask->s_addr);if (dev !=NULL )dev_put (dev); +/* lsmblob_init() puts entry->secid into all of the secids +* in blob. security_secid_to_secctx() will know which +* security module to use to create the secctx. */ +if (entry !=NULL ) +lsmblob_init (&blob, entry->secid);if (entry !=NULL && -security_secid_to_secctx (entry->secid, +security_secid_to_secctx (&blob, &secctx, &secctx_len) ==0 ) {audit_log_format (audit_buf," sec_obj=%s" , secctx);security_release_secctx (secctx, secctx_len);
@@ -537,6 +548,7 @@ static int netlbl_unlhsh_remove_addr6(struct net *net,struct net_device *dev;char *secctx;u32 secctx_len; +struct lsmblob blob;spin_lock (&netlbl_unlhsh_lock); list_entry =netlbl_af6list_remove (addr, mask, &iface->addr6_list);
@@ -555,8 +567,13 @@ static int netlbl_unlhsh_remove_addr6(struct net *net, addr, mask);if (dev !=NULL )dev_put (dev); +/* lsmblob_init() puts entry->secid into all of the secids +* in blob. security_secid_to_secctx() will know which +* security module to use to create the secctx. */ +if (entry !=NULL ) +lsmblob_init (&blob, entry->secid);if (entry !=NULL && -security_secid_to_secctx (entry->secid, +security_secid_to_secctx (&blob, &secctx, &secctx_len) ==0 ) {audit_log_format (audit_buf," sec_obj=%s" , secctx);security_release_secctx (secctx, secctx_len);
@@ -1082,6 +1099,7 @@ static int netlbl_unlabel_staticlist_gen(u32 cmd,u32 secid;char *secctx;u32 secctx_len; +struct lsmblob blob; data = genlmsg_put(cb_arg->skb, NETLINK_CB(cb_arg->nl_cb->skb).portid, cb_arg->seq, &netlbl_unlabel_gnl_family,
@@ -1136,7 +1154,11 @@ static int netlbl_unlabel_staticlist_gen(u32 cmd, secid = addr6->secid; } - ret_val =security_secid_to_secctx (secid, &secctx, &secctx_len); +/* lsmblob_init() secid into all of the secids in blob. +* security_secid_to_secctx() will know which security module +* to use to create the secctx. */ +lsmblob_init (&blob, secid); + ret_val =security_secid_to_secctx (&blob, &secctx, &secctx_len);if (ret_val !=0 ) goto list_cb_failure; ret_val = nla_put(cb_arg->skb,
net/netlabel/netlabel_user.c +3 -3
@@ -86,6 +86,7 @@ struct audit_buffer *netlbl_audit_start_common(int type,struct audit_buffer *audit_buf;char *secctx;u32 secctx_len; +struct lsmblob blob;if (audit_enabled ==AUDIT_OFF )return NULL ;
@@ -98,10 +99,9 @@ struct audit_buffer *netlbl_audit_start_common(int type,from_kuid (&init_user_ns, audit_info->loginuid), audit_info->sessionid); +lsmblob_init (&blob, audit_info->secid);if (audit_info->secid !=0 && -security_secid_to_secctx (audit_info->secid, - &secctx, - &secctx_len) ==0 ) { +security_secid_to_secctx (&blob, &secctx, &secctx_len) ==0 ) {audit_log_format (audit_buf," subj=%s" , secctx);security_release_secctx (secctx, secctx_len); }
security/security.c +5 -6
@@ -2172,17 +2172,16 @@ int security_ismaclabel(const char *name) }EXPORT_SYMBOL (security_ismaclabel); -int security_secid_to_secctx (u32 secid,char **secdata,u32 *seclen) +int security_secid_to_secctx (struct lsmblob *blob,char **secdata,u32 *seclen) {struct security_hook_list *hp;int rc; -/* -* Currently, only one LSM can implement secid_to_secctx (i.e this -* LSM hook is not "stackable"). -*/ hlist_for_each_entry (hp, &security_hook_heads.secid_to_secctx, list) { - rc = hp->hook.secid_to_secctx (secid, secdata, seclen); +if (WARN_ON (hp->lsmid->slot <0 || hp->lsmid->slot >= lsm_slot)) +continue ; + rc = hp->hook.secid_to_secctx (blob->secid[hp->lsmid->slot], + secdata, seclen);if (rc !=LSM_RET_DEFAULT (secid_to_secctx))return rc; }
Signature · 1 line
2.29.2
Re: [PATCH v26 08/25] LSM: Use lsmblob in security_secid_to_secctx
Quoted 9 lines · Casey Schaufler
Change
security_secid_to_secctx()to take a lsmblob as input instead of a u32 secid. It will then call the LSM hooks using the lsmblob element allocated for that module. The callers have been updated as well. This allows for the possibility that more than one module may be called upon to translate a secid to a string, as can occur in the audit code.Signed-off-by Casey Schaufler <[email protected]>
This looks sensible.
Signature · 1 line
Kees Cook
Re: [PATCH v26 08/25] LSM: Use lsmblob in security_secid_to_secctx
In reply to Casey SchauflerQuoted 15 lines · Casey Schaufler
Change
security_secid_to_secctx()to take a lsmblob as input instead of a u32 secid. It will then call the LSM hooks using the lsmblob element allocated for that module. The callers have been updated as well. This allows for the possibility that more than one module may be called upon to translate a secid to a string, as can occur in the audit code.Signed-off-by Casey Schaufler <[email protected]>To: Pablo Neira Ayuso <[email protected]>
To: Paul Moore <[email protected]>
Quoted 14 lines
--- drivers/android/binder.c | 12 +++++++++- include/linux/security.h | 5 +++-- include/net/scm.h | 7 +++++- kernel/audit.c | 20 +++++++++++++++-- kernel/auditsc.c | 28 +++++++++++++++++++---- net/ipv4/ip_sockglue.c | 4 +++- net/netfilter/nf_conntrack_netlink.c | 14 ++++++++++-- net/netfilter/nf_conntrack_standalone.c | 4 +++- net/netfilter/nfnetlink_queue.c | 11 +++++++-- net/netlabel/netlabel_unlabeled.c | 30 +++++++++++++++++++++---- net/netlabel/netlabel_user.c | 6 ++--- security/security.c | 11 +++++---- 12 files changed, 123 insertions(+), 29 deletions(-)
--
paul moore
www.paul-moore.com
Casey Schaufler
maintainerPATCH v26 9/25 LSM: Use lsmblob in security_ipc_getsecid
In reply to Casey SchauflerThere may be more than one LSM that provides IPC data for auditing. Change security_ipc_getsecid() to fill in a lsmblob structure instead of the u32 secid. The audit data structure containing the secid will be updated later, so there is a bit of scaffolding here.
--- include/linux/security.h | 7 ++++--- kernel/auditsc.c | 7 ++++++- security/security.c | 12 +++++++++--- 3 files changed, 19 insertions(+), 7 deletions(-)
include/linux/security.h +4 -3
@@ -518,7 +518,7 @@ int security_task_prctl(int option, unsigned long arg2, unsigned long arg3, unsigned long arg4, unsigned long arg5);void security_task_to_inode (struct task_struct *p,struct inode *inode);int security_ipc_permission (struct kern_ipc_perm *ipcp,short flag); -void security_ipc_getsecid (struct kern_ipc_perm *ipcp,u32 *secid); +void security_ipc_getsecid (struct kern_ipc_perm *ipcp,struct lsmblob *blob);int security_msg_msg_alloc (struct msg_msg *msg);void security_msg_msg_free (struct msg_msg *msg);int security_msg_queue_alloc (struct kern_ipc_perm *msq);
@@ -1275,9 +1275,10 @@ static inline int security_ipc_permission(struct kern_ipc_perm *ipcp,return 0 ; } -static inline void security_ipc_getsecid (struct kern_ipc_perm *ipcp,u32 *secid) +static inline void security_ipc_getsecid (struct kern_ipc_perm *ipcp, +struct lsmblob *blob) { - *secid =0 ; +lsmblob_init (blob,0 ); }static inline int security_msg_msg_alloc (struct msg_msg *msg)
kernel/auditsc.c +6 -1
@@ -2323,11 +2323,16 @@ void __audit_mq_getsetattr(mqd_t mqdes, struct mq_attr *mqstat)void __audit_ipc_obj (struct kern_ipc_perm *ipcp) {struct audit_context *context =audit_context (); +struct lsmblob blob; context->ipc.uid = ipcp->uid; context->ipc.gid = ipcp->gid; context->ipc.mode = ipcp->mode; context->ipc.has_perm =0 ; -security_ipc_getsecid (ipcp, &context->ipc.osid); +security_ipc_getsecid (ipcp, &blob); +/* context->ipc.osid will be changed to a lsmblob later in +* the patch series. This will allow auditing of all the object +* labels associated with the ipc object. */ + context->ipc.osid =lsmblob_value (&blob); context->type =AUDIT_IPC ; }
security/security.c +9 -3
@@ -1992,10 +1992,16 @@ int security_ipc_permission(struct kern_ipc_perm *ipcp, short flag)return call_int_hook (ipc_permission,0 , ipcp, flag); } -void security_ipc_getsecid (struct kern_ipc_perm *ipcp,u32 *secid) +void security_ipc_getsecid (struct kern_ipc_perm *ipcp,struct lsmblob *blob) { - *secid =0 ; -call_void_hook (ipc_getsecid, ipcp, secid); +struct security_hook_list *hp; + +lsmblob_init (blob,0 ); +hlist_for_each_entry (hp, &security_hook_heads.ipc_getsecid, list) { +if (WARN_ON (hp->lsmid->slot <0 || hp->lsmid->slot >= lsm_slot)) +continue ; + hp->hook.ipc_getsecid (ipcp, &blob->secid[hp->lsmid->slot]); + } }int security_msg_msg_alloc (struct msg_msg *msg)
Signature · 1 line
2.29.2
Casey Schaufler
maintainerPATCH v26 10/25 LSM: Use lsmblob in security_task_getsecid
In reply to Casey SchauflerChange the security_task_getsecid_subj() and security_task_getsecid_obj() interfaces to fill in a lsmblob structure instead of a u32 secid in support of LSM stacking. Audit interfaces will need to collect all possible secids for possible reporting.
--- drivers/android/binder.c | 12 +----- include/linux/security.h | 14 ++++--- kernel/audit.c | 16 +++----- kernel/auditfilter.c | 4 +- kernel/auditsc.c | 25 ++++++------ net/netlabel/netlabel_unlabeled.c | 5 ++- net/netlabel/netlabel_user.h | 6 ++- security/integrity/ima/ima_appraise.c | 10 +++-- security/integrity/ima/ima_main.c | 56 +++++++++++++++------------ security/security.c | 25 +++++++++--- 10 files changed, 94 insertions(+), 79 deletions(-)
drivers/android/binder.c +1 -11
@@ -2710,7 +2710,6 @@ static void binder_transaction(struct binder_proc *proc, t->priority =task_nice (current);if (target_node && target_node->txn_security_ctx) { -u32 secid;struct lsmblob blob;size_t added_size;
@@ -2723,16 +2722,7 @@ static void binder_transaction(struct binder_proc *proc,* here; however, it isn't clear that binder would handle that * case well anyway. */ -security_task_getsecid_obj (proc->tsk, &secid); -/* -* Later in this patch set security_task_getsecid() will -* provide a lsmblob instead of a secid. lsmblob_init -* is used to ensure that all the secids in the lsmblob -* get the value returned from security_task_getsecid(), -* which means that the one expected by -* security_secid_to_secctx() will be set. -*/ -lsmblob_init (&blob, secid); +security_task_getsecid_obj (proc->tsk, &blob); ret =security_secid_to_secctx (&blob, &secctx, &secctx_sz);if (ret) { return_error =BR_FAILED_REPLY ;
include/linux/security.h +8 -6
@@ -500,8 +500,8 @@ int security_task_fix_setgid(struct cred *new, const struct cred *old,int security_task_setpgid (struct task_struct *p,pid_t pgid);int security_task_getpgid (struct task_struct *p);int security_task_getsid (struct task_struct *p); -void security_task_getsecid_subj (struct task_struct *p,u32 *secid); -void security_task_getsecid_obj (struct task_struct *p,u32 *secid); +void security_task_getsecid_subj (struct task_struct *p,struct lsmblob *blob); +void security_task_getsecid_obj (struct task_struct *p,struct lsmblob *blob);int security_task_setnice (struct task_struct *p,int nice);int security_task_setioprio (struct task_struct *p,int ioprio);int security_task_getioprio (struct task_struct *p);
@@ -1197,14 +1197,16 @@ static inline int security_task_getsid(struct task_struct *p)return 0 ; } -static inline void security_task_getsecid_subj (struct task_struct *p,u32 *secid) +static inline void security_task_getsecid_subj (struct task_struct *p, +struct lsmblob *blob) { - *secid =0 ; +lsmblob_init (blob,0 ); } -static inline void security_task_getsecid_obj (struct task_struct *p,u32 *secid) +static inline void security_task_getsecid_obj (struct task_struct *p, +struct lsmblob *blob) { - *secid =0 ; +lsmblob_init (blob,0 ); }static inline int security_task_setnice (struct task_struct *p,int nice)
kernel/audit.c +6 -10
@@ -2139,19 +2139,12 @@ int audit_log_task_context(struct audit_buffer *ab)char *ctx =NULL ;unsigned len;int error; -u32 sid;struct lsmblob blob; -security_task_getsecid_subj (current, &sid); -if (!sid) +security_task_getsecid_subj (current, &blob); +if (!lsmblob_is_set (&blob))return 0 ; -/* -* lsmblob_init sets all values in the lsmblob to sid. -* This is temporary until security_task_getsecid is converted -* to use a lsmblob, which happens later in this patch set. -*/ -lsmblob_init (&blob, sid); error =security_secid_to_secctx (&blob, &ctx, &len);if (error) {if (error != -EINVAL )
@@ -2359,6 +2352,7 @@ int audit_set_loginuid(kuid_t loginuid)int audit_signal_info (int sig,struct task_struct *t) {kuid_t uid =current_uid (), auid; +struct lsmblob blob; if (auditd_test_task(t) && (sig == SIGTERM || sig == SIGHUP ||
@@ -2369,7 +2363,9 @@ int audit_signal_info(int sig, struct task_struct *t) audit_sig_uid = auid; else audit_sig_uid = uid; -security_task_getsecid_subj (current, &audit_sig_sid); +security_task_getsecid_subj (current, &blob); +/* scaffolding until audit_sig_sid is converted */ + audit_sig_sid = blob.secid[0 ]; }return audit_signal_info_syscall (t);
kernel/auditfilter.c +1 -3
@@ -1330,7 +1330,6 @@ int audit_filter(int msgtype, unsigned int listtype)for (i =0 ; i < e->rule.field_count; i++) {struct audit_field *f = &e->rule.fields[i];pid_t pid; -u32 sid;struct lsmblob blob;switch (f->type) {
@@ -1362,8 +1361,7 @@ int audit_filter(int msgtype, unsigned int listtype)case AUDIT_SUBJ_CLR :if (f->lsm_isset) {security_task_getsecid_subj (current, - &sid); -lsmblob_init (&blob, sid); + &blob); result =security_audit_rule_match ( &blob, f->type, f->op, f->lsm_rules);
kernel/auditsc.c +12 -13
@@ -471,7 +471,6 @@ static int audit_filter_rules(struct task_struct *tsk, {const struct cred *cred;int i, need_sid =1 ; -u32 sid;struct lsmblob blob;unsigned int sessionid;
@@ -668,17 +667,9 @@ static int audit_filter_rules(struct task_struct *tsk,logged upon error */ if (f->lsm_isset) {if (need_sid) { -security_task_getsecid_subj (tsk, &sid); +security_task_getsecid_subj (tsk, &blob); need_sid =0 ; } -/* -* lsmblob_init sets all values in the lsmblob -* to sid. This is temporary until -* security_task_getsecid() is converted to -* provide a lsmblob, which happens later in -* this patch set. -*/ -lsmblob_init (&blob, sid); result =security_audit_rule_match (&blob, f->type, f->op, f->lsm_rules);
@@ -2422,12 +2413,15 @@ int __audit_sockaddr(int len, void *a)void __audit_ptrace (struct task_struct *t) {struct audit_context *context =audit_context (); +struct lsmblob blob; context->target_pid =task_tgid_nr (t); context->target_auid =audit_get_loginuid (t); context->target_uid =task_uid (t); context->target_sessionid =audit_get_sessionid (t); -security_task_getsecid_obj (t, &context->target_sid); +security_task_getsecid_obj (t, &blob); +/* scaffolding - until target_sid is converted */ + context->target_sid = blob.secid[0 ];memcpy (context->target_comm, t->comm,TASK_COMM_LEN ); }
@@ -2443,6 +2437,7 @@ int audit_signal_info_syscall(struct task_struct *t)struct audit_aux_data_pids *axp;struct audit_context *ctx =audit_context ();kuid_t t_uid =task_uid (t); +struct lsmblob blob;if (!audit_signals ||audit_dummy_context ())return 0 ;
@@ -2454,7 +2449,9 @@ int audit_signal_info_syscall(struct task_struct *t) ctx->target_auid =audit_get_loginuid (t); ctx->target_uid = t_uid; ctx->target_sessionid =audit_get_sessionid (t); -security_task_getsecid_obj (t, &ctx->target_sid); +security_task_getsecid_obj (t, &blob); +/* scaffolding until target_sid is converted */ + ctx->target_sid = blob.secid[0 ];memcpy (ctx->target_comm, t->comm,TASK_COMM_LEN );return 0 ; }
@@ -2475,7 +2472,9 @@ int audit_signal_info_syscall(struct task_struct *t) axp->target_auid[axp->pid_count] =audit_get_loginuid (t); axp->target_uid[axp->pid_count] = t_uid; axp->target_sessionid[axp->pid_count] =audit_get_sessionid (t); -security_task_getsecid_obj (t, &axp->target_sid[axp->pid_count]); +security_task_getsecid_obj (t, &blob); +/* scaffolding until target_sid is converted */ + axp->target_sid[axp->pid_count] = blob.secid[0 ];memcpy (axp->target_comm[axp->pid_count], t->comm,TASK_COMM_LEN ); axp->pid_count++;
net/netlabel/netlabel_unlabeled.c +4 -1
@@ -1564,11 +1564,14 @@ int __init netlbl_unlabel_defconf(void)int ret_val;struct netlbl_dom_map *entry;struct netlbl_audit audit_info; +struct lsmblob blob;/* Only the kernel is allowed to call this function and the only time * it is called is at bootup before the audit subsystem is reporting * messages so don't worry to much about these values. */ -security_task_getsecid_subj (current, &audit_info.secid); +security_task_getsecid_subj (current, &blob); +/* scaffolding until audit_info.secid is converted */ + audit_info.secid = blob.secid[0 ]; audit_info.loginuid =GLOBAL_ROOT_UID ; audit_info.sessionid =0 ;
net/netlabel/netlabel_user.h +5 -1
@@ -34,7 +34,11 @@static inline void netlbl_netlink_auditinfo (struct sk_buff *skb,struct netlbl_audit *audit_info) { -security_task_getsecid_subj (current, &audit_info->secid); +struct lsmblob blob; + +security_task_getsecid_subj (current, &blob); +/* scaffolding until secid is converted */ + audit_info->secid = blob.secid[0 ]; audit_info->loginuid =audit_get_loginuid (current); audit_info->sessionid =audit_get_sessionid (current); }
security/integrity/ima/ima_appraise.c +6 -4
@@ -71,14 +71,16 @@ bool is_ima_appraise_enabled(void)int ima_must_appraise (struct user_namespace *mnt_userns,struct inode *inode,int mask,enum ima_hooks func) { -u32 secid; +struct lsmblob blob;if (!ima_appraise)return 0 ; -security_task_getsecid_subj (current, &secid); -return ima_match_policy (mnt_userns, inode,current_cred (), secid, func, - mask,IMA_APPRAISE |IMA_HASH ,NULL ,NULL ,NULL ); +security_task_getsecid_subj (current, &blob); +/* scaffolding the .secid[0] */ +return ima_match_policy (mnt_userns, inode,current_cred (), + blob.secid[0 ], func, mask, +IMA_APPRAISE |IMA_HASH ,NULL ,NULL ,NULL ); } static int ima_fix_xattr(struct dentry *dentry,
security/integrity/ima/ima_main.c +32 -24
131 lines not on this page · raw message
security/security.c +19 -6
@@ -1902,17 +1902,30 @@ int security_task_getsid(struct task_struct *p)return call_int_hook (task_getsid,0 , p); } -void security_task_getsecid_subj (struct task_struct *p,u32 *secid) +void security_task_getsecid_subj (struct task_struct *p,struct lsmblob *blob) { - *secid =0 ; -call_void_hook (task_getsecid_subj, p, secid); +struct security_hook_list *hp; + +lsmblob_init (blob,0 ); +hlist_for_each_entry (hp, &security_hook_heads.task_getsecid_subj, + list) { +if (WARN_ON (hp->lsmid->slot <0 || hp->lsmid->slot >= lsm_slot)) +continue ; + hp->hook.task_getsecid_subj (p, &blob->secid[hp->lsmid->slot]); + } }EXPORT_SYMBOL (security_task_getsecid_subj); -void security_task_getsecid_obj (struct task_struct *p,u32 *secid) +void security_task_getsecid_obj (struct task_struct *p,struct lsmblob *blob) { - *secid =0 ; -call_void_hook (task_getsecid_obj, p, secid); +struct security_hook_list *hp; + +lsmblob_init (blob,0 ); +hlist_for_each_entry (hp, &security_hook_heads.task_getsecid_obj, list) { +if (WARN_ON (hp->lsmid->slot <0 || hp->lsmid->slot >= lsm_slot)) +continue ; + hp->hook.task_getsecid_obj (p, &blob->secid[hp->lsmid->slot]); + } }EXPORT_SYMBOL (security_task_getsecid_obj);
Signature · 1 line
2.29.2
Casey Schaufler
maintainerPATCH v26 11/25 LSM: Use lsmblob in security_inode_getsecid
In reply to Casey SchauflerChange the security_inode_getsecid() interface to fill in a lsmblob structure instead of a u32 secid. This allows for its callers to gather data from all registered LSMs. Data is provided for IMA and audit.
--- include/linux/security.h | 7 ++++--- kernel/auditsc.c | 6 +++++- security/integrity/ima/ima_policy.c | 4 +--- security/security.c | 11 +++++++++-- 4 files changed, 19 insertions(+), 9 deletions(-)
include/linux/security.h +4 -3
@@ -454,7 +454,7 @@ int security_inode_getsecurity(struct user_namespace *mnt_userns, void **buffer, bool alloc);int security_inode_setsecurity (struct inode *inode,const char *name,const void *value,size_t size,int flags);int security_inode_listsecurity (struct inode *inode,char *buffer,size_t buffer_size); -void security_inode_getsecid (struct inode *inode,u32 *secid); +void security_inode_getsecid (struct inode *inode,struct lsmblob *blob);int security_inode_copy_up (struct dentry *src,struct cred **new);int security_inode_copy_up_xattr (const char *name); int security_kernfs_init_security(struct kernfs_node *kn_dir,
@@ -1005,9 +1005,10 @@ static inline int security_inode_listsecurity(struct inode *inode, char *buffer,return 0 ; } -static inline void security_inode_getsecid (struct inode *inode,u32 *secid) +static inline void security_inode_getsecid (struct inode *inode, +struct lsmblob *blob) { - *secid =0 ; +lsmblob_init (blob,0 ); }static inline int security_inode_copy_up (struct dentry *src,struct cred **new)
kernel/auditsc.c +5 -1
@@ -1962,13 +1962,17 @@ static void audit_copy_inode(struct audit_names *name,const struct dentry *dentry, struct inode *inode, unsigned int flags) { + struct lsmblob blob; + name->ino = inode->i_ino; name->dev = inode->i_sb->s_dev; name->mode = inode->i_mode; name->uid = inode->i_uid; name->gid = inode->i_gid; name->rdev = inode->i_rdev; -security_inode_getsecid (inode, &name->osid); +security_inode_getsecid (inode, &blob); +/* scaffolding until osid is updated */ + name->osid = blob.secid[0 ];if (flags &AUDIT_INODE_NOEVAL ) { name->fcap_ver =-1 ;return ;
security/integrity/ima/ima_policy.c +1 -3
@@ -606,7 +606,6 @@ static bool ima_match_rules(struct ima_rule_entry *rule,return false;for (i =0 ; i <MAX_LSM_RULES ; i++) {int rc =0 ; -u32 osid;struct lsmblob lsmdata;if (!ima_lsm_isset (rule, i)) {
@@ -619,8 +618,7 @@ static bool ima_match_rules(struct ima_rule_entry *rule,case LSM_OBJ_USER :case LSM_OBJ_ROLE :case LSM_OBJ_TYPE : -security_inode_getsecid (inode, &osid); -lsmblob_init (&lsmdata, osid); +security_inode_getsecid (inode, &lsmdata); rc =ima_filter_rule_match (&lsmdata, rule->lsm[i].type, Audit_equal, rule->lsm[i].rule);
security/security.c +9 -2
@@ -1546,9 +1546,16 @@ int security_inode_listsecurity(struct inode *inode, char *buffer, size_t buffer }EXPORT_SYMBOL (security_inode_listsecurity); -void security_inode_getsecid (struct inode *inode,u32 *secid) +void security_inode_getsecid (struct inode *inode,struct lsmblob *blob) { -call_void_hook (inode_getsecid, inode, secid); +struct security_hook_list *hp; + +lsmblob_init (blob,0 ); +hlist_for_each_entry (hp, &security_hook_heads.inode_getsecid, list) { +if (WARN_ON (hp->lsmid->slot <0 || hp->lsmid->slot >= lsm_slot)) +continue ; + hp->hook.inode_getsecid (inode, &blob->secid[hp->lsmid->slot]); + } }int security_inode_copy_up (struct dentry *src,struct cred **new)
Signature · 1 line
2.29.2
Casey Schaufler
maintainerPATCH v26 12/25 LSM: Use lsmblob in security_cred_getsecid
In reply to Casey SchauflerChange the security_cred_getsecid() interface to fill in a lsmblob instead of a u32 secid. The associated data elements in the audit sub-system are changed from a secid to a lsmblob to accommodate multiple possible LSM audit users.
--- include/linux/security.h | 2 +- kernel/audit.c | 25 +++++++---------------- kernel/audit.h | 3 ++- kernel/auditsc.c | 34 ++++++++++++------------------- security/integrity/ima/ima_main.c | 8 ++++---- security/security.c | 12 ++++++++--- 6 files changed, 36 insertions(+), 48 deletions(-)
include/linux/security.h +1 -1
@@ -481,7 +481,7 @@ int security_cred_alloc_blank(struct cred *cred, gfp_t gfp);void security_cred_free (struct cred *cred);int security_prepare_creds (struct cred *new,const struct cred *old,gfp_t gfp);void security_transfer_creds (struct cred *new,const struct cred *old); -void security_cred_getsecid (const struct cred *c,u32 *secid); +void security_cred_getsecid (const struct cred *c,struct lsmblob *blob);int security_kernel_act_as (struct cred *new,struct lsmblob *blob);int security_kernel_create_files_as (struct cred *new,struct inode *inode);int security_kernel_module_request (char *kmod_name);
kernel/audit.c +7 -18
@@ -125,7 +125,7 @@ static u32 audit_backlog_wait_time = AUDIT_BACKLOG_WAIT_TIME;/* The identity of the user shutting down the audit system. */ static kuid_t audit_sig_uid =INVALID_UID ;static pid_t audit_sig_pid =-1 ; -static u32 audit_sig_sid; +struct lsmblob audit_sig_lsm;/* Records can be lost in several ways: 0) [suppressed in audit_alloc]
@@ -1441,29 +1441,21 @@ static int audit_receive_msg(struct sk_buff *skb, struct nlmsghdr *nlh) }case AUDIT_SIGNAL_INFO : len =0 ; -if (audit_sig_sid) { -struct lsmblob blob; - -/* -* lsmblob_init sets all values in the lsmblob -* to audit_sig_sid. This is temporary until -* audit_sig_sid is converted to a lsmblob, which -* happens later in this patch set. -*/ -lsmblob_init (&blob, audit_sig_sid); - err =security_secid_to_secctx (&blob, &ctx, &len); +if (lsmblob_is_set (&audit_sig_lsm)) { + err =security_secid_to_secctx (&audit_sig_lsm, &ctx, + &len);if (err)return err; } sig_data =kmalloc (sizeof (*sig_data) + len,GFP_KERNEL );if (!sig_data) { -if (audit_sig_sid) +if (lsmblob_is_set (&audit_sig_lsm))security_release_secctx (ctx, len);return -ENOMEM ; } sig_data->uid =from_kuid (&init_user_ns, audit_sig_uid); sig_data->pid = audit_sig_pid; -if (audit_sig_sid) { +if (lsmblob_is_set (&audit_sig_lsm)) {memcpy (sig_data->ctx, ctx, len);security_release_secctx (ctx, len); }
@@ -2352,7 +2344,6 @@ int audit_set_loginuid(kuid_t loginuid)int audit_signal_info (int sig,struct task_struct *t) {kuid_t uid =current_uid (), auid; -struct lsmblob blob; if (auditd_test_task(t) && (sig == SIGTERM || sig == SIGHUP ||
@@ -2363,9 +2354,7 @@ int audit_signal_info(int sig, struct task_struct *t) audit_sig_uid = auid; else audit_sig_uid = uid; -security_task_getsecid_subj (current, &blob); -/* scaffolding until audit_sig_sid is converted */ - audit_sig_sid = blob.secid[0 ]; +security_task_getsecid_subj (current, &audit_sig_lsm); }return audit_signal_info_syscall (t);
kernel/audit.h +2 -1
@@ -9,6 +9,7 @@#include <linux/fs.h> #include <linux/audit.h> #include <linux/skbuff.h> +#include <linux/security.h> #include <uapi/linux/mqueue.h> #include <linux/tty.h>
@@ -134,7 +135,7 @@ struct audit_context {kuid_t target_auid;kuid_t target_uid;unsigned int target_sessionid; -u32 target_sid; +struct lsmblob target_lsm;char target_comm[TASK_COMM_LEN ];struct audit_tree_refs *trees, *first_trees;
kernel/auditsc.c +13 -21
@@ -111,7 +111,7 @@ struct audit_aux_data_pids {kuid_t target_auid[AUDIT_AUX_PIDS ];kuid_t target_uid[AUDIT_AUX_PIDS ];unsigned int target_sessionid[AUDIT_AUX_PIDS ]; -u32 target_sid[AUDIT_AUX_PIDS ]; +struct lsmblob target_lsm[AUDIT_AUX_PIDS ];char target_comm[AUDIT_AUX_PIDS ][TASK_COMM_LEN ];int pid_count; };
@@ -991,14 +991,14 @@ static inline void audit_free_context(struct audit_context *context) }static int audit_log_pid_context (struct audit_context *context,pid_t pid, -kuid_t auid,kuid_t uid,unsigned int sessionid, -u32 sid,char *comm) +kuid_t auid,kuid_t uid, +unsigned int sessionid, +struct lsmblob *blob,char *comm) {struct audit_buffer *ab;char *ctx =NULL ;u32 len;int rc =0 ; -struct lsmblob blob; ab =audit_log_start (context,GFP_KERNEL ,AUDIT_OBJ_PID );if (!ab)
@@ -1007,9 +1007,8 @@ static int audit_log_pid_context(struct audit_context *context, pid_t pid,audit_log_format (ab,"opid=%d oauid=%d ouid=%d oses=%d" , pid,from_kuid (&init_user_ns, auid),from_kuid (&init_user_ns, uid), sessionid); -if (sid) { -lsmblob_init (&blob, sid); -if (security_secid_to_secctx (&blob, &ctx, &len)) { +if (lsmblob_is_set (blob)) { +if (security_secid_to_secctx (blob, &ctx, &len)) {audit_log_format (ab," obj=(none)" ); rc =1 ; }else {
@@ -1580,7 +1579,7 @@ static void audit_log_exit(void) axs->target_auid[i], axs->target_uid[i], axs->target_sessionid[i], - axs->target_sid[i], + &axs->target_lsm[i], axs->target_comm[i])) call_panic = 1; }
@@ -1589,7 +1588,7 @@ static void audit_log_exit(void)audit_log_pid_context (context, context->target_pid, context->target_auid, context->target_uid, context->target_sessionid, - context->target_sid, context->target_comm)) + &context->target_lsm, context->target_comm)) call_panic = 1;if (context->pwd.dentry && context->pwd.mnt) {
@@ -1765,7 +1764,7 @@ void __audit_syscall_exit(int success, long return_code) context->aux =NULL ; context->aux_pids =NULL ; context->target_pid =0 ; - context->target_sid =0 ; +lsmblob_init (&context->target_lsm,0 ); context->sockaddr_len =0 ; context->type =0 ; context->fds[0 ] =-1 ;
@@ -2319,6 +2318,7 @@ void __audit_ipc_obj(struct kern_ipc_perm *ipcp) {struct audit_context *context =audit_context ();struct lsmblob blob; + context->ipc.uid = ipcp->uid; context->ipc.gid = ipcp->gid; context->ipc.mode = ipcp->mode;
@@ -2417,15 +2417,12 @@ int __audit_sockaddr(int len, void *a)void __audit_ptrace (struct task_struct *t) {struct audit_context *context =audit_context (); -struct lsmblob blob; context->target_pid =task_tgid_nr (t); context->target_auid =audit_get_loginuid (t); context->target_uid =task_uid (t); context->target_sessionid =audit_get_sessionid (t); -security_task_getsecid_obj (t, &blob); -/* scaffolding - until target_sid is converted */ - context->target_sid = blob.secid[0 ]; +security_task_getsecid_obj (t, &context->target_lsm);memcpy (context->target_comm, t->comm,TASK_COMM_LEN ); }
@@ -2441,7 +2438,6 @@ int audit_signal_info_syscall(struct task_struct *t)struct audit_aux_data_pids *axp;struct audit_context *ctx =audit_context ();kuid_t t_uid =task_uid (t); -struct lsmblob blob;if (!audit_signals ||audit_dummy_context ())return 0 ;
@@ -2453,9 +2449,7 @@ int audit_signal_info_syscall(struct task_struct *t) ctx->target_auid =audit_get_loginuid (t); ctx->target_uid = t_uid; ctx->target_sessionid =audit_get_sessionid (t); -security_task_getsecid_obj (t, &blob); -/* scaffolding until target_sid is converted */ - ctx->target_sid = blob.secid[0 ]; +security_task_getsecid_obj (t, &ctx->target_lsm);memcpy (ctx->target_comm, t->comm,TASK_COMM_LEN );return 0 ; }
@@ -2476,9 +2470,7 @@ int audit_signal_info_syscall(struct task_struct *t) axp->target_auid[axp->pid_count] =audit_get_loginuid (t); axp->target_uid[axp->pid_count] = t_uid; axp->target_sessionid[axp->pid_count] =audit_get_sessionid (t); -security_task_getsecid_obj (t, &blob); -/* scaffolding until target_sid is converted */ - axp->target_sid[axp->pid_count] = blob.secid[0 ]; +security_task_getsecid_obj (t, &axp->target_lsm[axp->pid_count]);memcpy (axp->target_comm[axp->pid_count], t->comm,TASK_COMM_LEN ); axp->pid_count++;
security/integrity/ima/ima_main.c +4 -4
@@ -470,7 +470,6 @@ int ima_file_mprotect(struct vm_area_struct *vma, unsigned long prot)int ima_bprm_check (struct linux_binprm *bprm) {int ret; -u32 secid;struct lsmblob blob;security_task_getsecid_subj (current, &blob);
@@ -480,9 +479,10 @@ int ima_bprm_check(struct linux_binprm *bprm)if (ret)return ret; -security_cred_getsecid (bprm->cred, &secid); -return process_measurement (bprm->file, bprm->cred, secid,NULL ,0 , -MAY_EXEC ,CREDS_CHECK ); +security_cred_getsecid (bprm->cred, &blob); +/* scaffolding until process_measurement changes */ +return process_measurement (bprm->file, bprm->cred, blob.secid[0 ], +NULL ,0 ,MAY_EXEC ,CREDS_CHECK ); }/**
security/security.c +9 -3
@@ -1796,10 +1796,16 @@ void security_transfer_creds(struct cred *new, const struct cred *old)call_void_hook (cred_transfer, new, old); } -void security_cred_getsecid (const struct cred *c,u32 *secid) +void security_cred_getsecid (const struct cred *c,struct lsmblob *blob) { - *secid =0 ; -call_void_hook (cred_getsecid, c, secid); +struct security_hook_list *hp; + +lsmblob_init (blob,0 ); +hlist_for_each_entry (hp, &security_hook_heads.cred_getsecid, list) { +if (WARN_ON (hp->lsmid->slot <0 || hp->lsmid->slot >= lsm_slot)) +continue ; + hp->hook.cred_getsecid (c, &blob->secid[hp->lsmid->slot]); + } }EXPORT_SYMBOL (security_cred_getsecid);
Signature · 1 line
2.29.2
Casey Schaufler
maintainerPATCH v26 13/25 IMA: Change internal interfaces to use lsmblobs
In reply to Casey SchauflerThe IMA interfaces ima_get_action() and ima_match_policy() call LSM functions that use lsmblobs. Change the IMA functions to pass the lsmblob to be compatible with the LSM functions.
To: Mimi Zohar <[email protected]>
--- security/integrity/ima/ima.h | 6 ++--- security/integrity/ima/ima_api.c | 6 ++--- security/integrity/ima/ima_appraise.c | 5 ++-- security/integrity/ima/ima_main.c | 36 +++++++++++---------------- security/integrity/ima/ima_policy.c | 17 ++++++------- 5 files changed, 31 insertions(+), 39 deletions(-)
security/integrity/ima/ima.h +3 -3
@@ -251,7 +251,7 @@ static inline void ima_process_queued_keys(void) {}/* LIM API function definitions */ int ima_get_action (struct user_namespace *mnt_userns,struct inode *inode, -const struct cred *cred,u32 secid,int mask, +const struct cred *cred,struct lsmblob *blob,int mask,enum ima_hooks func,int *pcr,struct ima_template_desc **template_desc,const char *func_data);
@@ -282,8 +282,8 @@ const char *ima_d_path(const struct path *path, char **pathbuf, char *filename);/* IMA policy related functions */ int ima_match_policy (struct user_namespace *mnt_userns,struct inode *inode, -const struct cred *cred,u32 secid,enum ima_hooks func, -int mask,int flags,int *pcr, +const struct cred *cred,struct lsmblob *blob, +enum ima_hooks func,int mask,int flags,int *pcr,struct ima_template_desc **template_desc,const char *func_data);void ima_init_policy (void );
security/integrity/ima/ima_api.c +3 -3
@@ -165,7 +165,7 @@ void ima_add_violation(struct file *file, const unsigned char *filename,* @mnt_userns: user namespace of the mount the inode was found from * @inode: pointer to the inode associated with the object being validated * @cred: pointer to credentials structure to validate -* @secid: secid of the task being validated +* @blob: LSM data of the task being validated * @mask: contains the permission mask (MAY_READ, MAY_WRITE, MAY_EXEC, * MAY_APPEND) * @func: caller identifier
@@ -185,7 +185,7 @@ void ima_add_violation(struct file *file, const unsigned char *filename,* */ int ima_get_action (struct user_namespace *mnt_userns,struct inode *inode, -const struct cred *cred,u32 secid,int mask, +const struct cred *cred,struct lsmblob *blob,int mask,enum ima_hooks func,int *pcr,struct ima_template_desc **template_desc,const char *func_data)
@@ -194,7 +194,7 @@ int ima_get_action(struct user_namespace *mnt_userns, struct inode *inode, flags &= ima_policy_flag; -return ima_match_policy (mnt_userns, inode, cred, secid, func, mask, +return ima_match_policy (mnt_userns, inode, cred, blob, func, mask, flags, pcr, template_desc, func_data); }
security/integrity/ima/ima_appraise.c +2 -3
@@ -77,10 +77,9 @@ int ima_must_appraise(struct user_namespace *mnt_userns, struct inode *inode,return 0 ;security_task_getsecid_subj (current, &blob); -/* scaffolding the .secid[0] */ return ima_match_policy (mnt_userns, inode,current_cred (), - blob.secid[0 ], func, mask, -IMA_APPRAISE |IMA_HASH ,NULL ,NULL ,NULL ); + &blob, func, mask,IMA_APPRAISE |IMA_HASH , +NULL ,NULL ,NULL ); } static int ima_fix_xattr(struct dentry *dentry,
security/integrity/ima/ima_main.c +15 -21
@@ -194,8 +194,8 @@ void ima_file_free(struct file *file) }static int process_measurement (struct file *file,const struct cred *cred, -u32 secid,char *buf,loff_t size,int mask, -enum ima_hooks func) +struct lsmblob *blob,char *buf,loff_t size, +int mask,enum ima_hooks func) {struct inode *inode =file_inode (file);struct integrity_iint_cache *iint =NULL ;
@@ -218,7 +218,7 @@ static int process_measurement(struct file *file, const struct cred *cred,* bitmask based on the appraise/audit/measurement policy. * Included is the appraise submask. */ - action =ima_get_action (file_mnt_user_ns (file), inode, cred, secid, + action =ima_get_action (file_mnt_user_ns (file), inode, cred, blob, mask, func, &pcr, &template_desc,NULL ); violation_check = ((func ==FILE_CHECK || func ==MMAP_CHECK ) && (ima_policy_flag &IMA_MEASURE ));
@@ -392,8 +392,7 @@ int ima_file_mmap(struct file *file, unsigned long prot)if (file && (prot &PROT_EXEC )) {security_task_getsecid_subj (current, &blob); -/* scaffolding - until process_measurement changes */ -return process_measurement (file,current_cred (), blob.secid[0 ], +return process_measurement (file,current_cred (), &blob,NULL ,0 ,MAY_EXEC ,MMAP_CHECK ); }
@@ -434,7 +433,7 @@ int ima_file_mprotect(struct vm_area_struct *vma, unsigned long prot) inode =file_inode (vma->vm_file);/* scaffolding */ action =ima_get_action (file_mnt_user_ns (vma->vm_file), inode, -current_cred (), blob.secid[0 ],MAY_EXEC , +current_cred (), &blob,MAY_EXEC ,MMAP_CHECK , &pcr, &template,0 );/* Is the mmap'ed file in policy? */
@@ -473,16 +472,14 @@ int ima_bprm_check(struct linux_binprm *bprm)struct lsmblob blob;security_task_getsecid_subj (current, &blob); -/* scaffolding until process_measurement changes */ - ret =process_measurement (bprm->file,current_cred (), blob.secid[0 ], -NULL ,0 ,MAY_EXEC ,BPRM_CHECK ); + ret =process_measurement (bprm->file,current_cred (), &blob,NULL ,0 , +MAY_EXEC ,BPRM_CHECK );if (ret)return ret;security_cred_getsecid (bprm->cred, &blob); -/* scaffolding until process_measurement changes */ -return process_measurement (bprm->file, bprm->cred, blob.secid[0 ], -NULL ,0 ,MAY_EXEC ,CREDS_CHECK ); +return process_measurement (bprm->file, bprm->cred, &blob,NULL ,0 , +MAY_EXEC ,CREDS_CHECK ); }/**
@@ -500,8 +497,7 @@ int ima_file_check(struct file *file, int mask)struct lsmblob blob;security_task_getsecid_subj (current, &blob); -/* scaffolding until process_measurement changes */ -return process_measurement (file,current_cred (), blob.secid[0 ],NULL ,0 , +return process_measurement (file,current_cred (), &blob,NULL ,0 , mask & (MAY_READ |MAY_WRITE |MAY_EXEC |MAY_APPEND ),FILE_CHECK ); }
@@ -698,9 +694,8 @@ int ima_read_file(struct file *file, enum kernel_read_file_id read_id,/* Read entire file for all partial reads. */ func = read_idmap[read_id] ?:FILE_CHECK ;security_task_getsecid_subj (current, &blob); -/* scaffolding - until process_measurement changes */ -return process_measurement (file,current_cred (), blob.secid[0 ],NULL , -0 ,MAY_READ , func); +return process_measurement (file,current_cred (), &blob,NULL ,0 , +MAY_READ , func); }const int read_idmap[READING_MAX_ID ] = {
@@ -742,9 +737,8 @@ int ima_post_read_file(struct file *file, void *buf, loff_t size, func = read_idmap[read_id] ?:FILE_CHECK ;security_task_getsecid_subj (current, &blob); -/* scaffolding until process_measurement changes */ -return process_measurement (file,current_cred (), blob.secid[0 ], buf, - size,MAY_READ , func); +return process_measurement (file,current_cred (), &blob, buf, size, +MAY_READ , func); }/**
@@ -889,7 +883,7 @@ void process_buffer_measurement(struct user_namespace *mnt_userns,security_task_getsecid_subj (current, &blob);/* scaffolding */ action =ima_get_action (mnt_userns, inode,current_cred (), - blob.secid[0 ],0 , func, &pcr, &template, + &blob,0 , func, &pcr, &template, func_data);if (!(action &IMA_MEASURE ))return ;
security/integrity/ima/ima_policy.c +8 -9
@@ -546,7 +546,7 @@ static bool ima_match_rule_data(struct ima_rule_entry *rule,* @mnt_userns: user namespace of the mount the inode was found from * @inode: a pointer to an inode * @cred: a pointer to a credentials structure for user validation -* @secid: the secid of the task to be validated +* @blob: the lsm data of the task to be validated * @func: LIM hook identifier * @mask: requested action (MAY_READ | MAY_WRITE | MAY_APPEND | MAY_EXEC) * @func_data: func specific data, may be NULL
@@ -556,8 +556,8 @@ static bool ima_match_rule_data(struct ima_rule_entry *rule,static bool ima_match_rules (struct ima_rule_entry *rule,struct user_namespace *mnt_userns,struct inode *inode,const struct cred *cred, -u32 secid,enum ima_hooks func,int mask, -const char *func_data) +struct lsmblob *blob,enum ima_hooks func, +int mask,const char *func_data) {int i;
@@ -626,8 +626,7 @@ static bool ima_match_rules(struct ima_rule_entry *rule,case LSM_SUBJ_USER :case LSM_SUBJ_ROLE :case LSM_SUBJ_TYPE : -lsmblob_init (&lsmdata, secid); - rc =ima_filter_rule_match (&lsmdata, rule->lsm[i].type, + rc =ima_filter_rule_match (blob, rule->lsm[i].type, Audit_equal, rule->lsm[i].rule);break ;
@@ -671,7 +670,7 @@ static int get_subaction(struct ima_rule_entry *rule, enum ima_hooks func)* @inode: pointer to an inode for which the policy decision is being made * @cred: pointer to a credentials structure for which the policy decision is * being made -* @secid: LSM secid of the task to be validated +* @blob: LSM data of the task to be validated * @func: IMA hook identifier * @mask: requested action (MAY_READ | MAY_WRITE | MAY_APPEND | MAY_EXEC) * @pcr: set the pcr to extend
@@ -686,8 +685,8 @@ static int get_subaction(struct ima_rule_entry *rule, enum ima_hooks func)* than writes so ima_match_policy() is classical RCU candidate. */ int ima_match_policy (struct user_namespace *mnt_userns,struct inode *inode, -const struct cred *cred,u32 secid,enum ima_hooks func, -int mask,int flags,int *pcr, +const struct cred *cred,struct lsmblob *blob, +enum ima_hooks func,int mask,int flags,int *pcr,struct ima_template_desc **template_desc,const char *func_data) {
@@ -703,7 +702,7 @@ int ima_match_policy(struct user_namespace *mnt_userns, struct inode *inode,if (!(entry->action & actmask))continue ; -if (!ima_match_rules (entry, mnt_userns, inode, cred, secid, +if (!ima_match_rules (entry, mnt_userns, inode, cred, blob, func, mask, func_data))continue ;
Signature · 1 line
2.29.2
Casey Schaufler
maintainerPATCH v26 14/25 LSM: Specify which LSM to display
In reply to Casey SchauflerCreate a new entry "interface_lsm" in the procfs attr directory for controlling which LSM security information is displayed for a process. A process can only read or write its own display value.
The name of an active LSM that supplies hooks for human readable data may be written to "interface_lsm" to set the value. The name of the LSM currently in use can be read from "interface_lsm". At this point there can only be one LSM capable of display active. A helper function lsm_task_ilsm() is provided to get the interface lsm slot for a task_struct.
Setting the "interface_lsm" requires that all security modules using setprocattr hooks allow the action. Each security module is responsible for defining its policy.
AppArmor hook provided by John Johansen <[email protected]> SELinux hook provided by Stephen Smalley <[email protected]>
--- .../ABI/testing/procfs-attr-lsm_display | 22 +++ Documentation/security/lsm.rst | 14 ++ fs/proc/base.c | 1 + include/linux/lsm_hooks.h | 17 ++ security/apparmor/include/apparmor.h | 3 +- security/apparmor/lsm.c | 32 ++++ security/security.c | 166 ++++++++++++++++-- security/selinux/hooks.c | 11 ++ security/selinux/include/classmap.h | 2 +- security/smack/smack_lsm.c | 7 + 10 files changed, 256 insertions(+), 19 deletions(-) create mode 100644 Documentation/ABI/testing/procfs-attr-lsm_display
Documentation/ABI/testing/procfs-attr-lsm_display new file +22 -0
@@ -0,0 +1,22 @@ +What: /proc/*/attr/lsm_display +Contact: [email protected], +Description: The name of the Linux security module (LSM) that will + provide information in the /proc/*/attr/current, + /proc/*/attr/prev and /proc/*/attr/exec interfaces. + The details of permissions required to read from + this interface are dependent on the LSMs active on the + system. + A process cannot write to this interface unless it + refers to itself. + The other details of permissions required to write to + this interface are dependent on the LSMs active on the + system. + The format of the data used by this interface is a + text string identifying the name of an LSM. The values + accepted are: + selinux - the SELinux LSM + smack - the Smack LSM + apparmor - The AppArmor LSM + By convention the LSM names are lower case and do not + contain special characters. +Users: LSM user-space
Documentation/security/lsm.rst +14 -0
@@ -129,3 +129,17 @@ to identify it as the first security module to be registered. The capabilities security module does not use the general security blobs, unlike other modules. The reasons are historical and are based on overhead, complexity and performance concerns. + +LSM External Interfaces +======================= + +The LSM infrastructure does not generally provide external interfaces. +The individual security modules provide what external interfaces they +require. + +The file ``/sys/kernel/security/lsm`` provides a comma +separated list of the active security modules. + +The file ``/proc/pid/attr/interface_lsm`` contains the name of the security +module for which the ``/proc/pid/attr/current`` interface will +apply. This interface can be written to.
fs/proc/base.c +1 -0
@@ -2807,6 +2807,7 @@ static const struct pid_entry attr_dir_stuff[] = {ATTR (NULL ,"fscreate" ,0666 ),ATTR (NULL ,"keycreate" ,0666 ),ATTR (NULL ,"sockcreate" ,0666 ), +ATTR (NULL ,"interface_lsm" ,0666 ), #ifdef CONFIG_SECURITY_SMACKDIR ("smack" ,0555 , proc_smack_attr_dir_inode_ops, proc_smack_attr_dir_ops),
include/linux/lsm_hooks.h +17 -0
@@ -1686,4 +1686,21 @@ static inline void security_delete_hooks(struct security_hook_list *hooks,extern int lsm_inode_alloc (struct inode *inode); +/** +* lsm_task_ilsm - the "interface_lsm" for this task +* @task: The task to report on +* +* Returns the task's interface LSM slot. +*/ +static inline int lsm_task_ilsm (struct task_struct *task) +{ +#ifdef CONFIG_SECURITY +int *ilsm = task->security; + +if (ilsm) +return *ilsm; +#endif +return LSMBLOB_INVALID ; +} +#endif /* ! __LINUX_LSM_HOOKS_H */
security/apparmor/include/apparmor.h +2 -1
@@ -28,8 +28,9 @@#define AA_CLASS_SIGNAL 10#define AA_CLASS_NET 14#define AA_CLASS_LABEL 16 +#define AA_CLASS_DISPLAY_LSM 17 -#define AA_CLASS_LAST AA_CLASS_LABEL +#define AA_CLASS_LAST AA_CLASS_DISPLAY_LSM/* Control parameters settable through module/boot flags */ extern enum audit_mode aa_g_audit;
security/apparmor/lsm.c +32 -0
@@ -621,6 +621,25 @@ static int apparmor_getprocattr(struct task_struct *task, char *name,return error; } + +static int profile_interface_lsm (struct aa_profile *profile, +struct common_audit_data *sa) +{ +struct aa_perms perms = { }; +unsigned int state; + + state =PROFILE_MEDIATES (profile,AA_CLASS_DISPLAY_LSM ); +if (state) { +aa_compute_perms (profile->policy.dfa, state, &perms); +aa_apply_modes_to_perms (profile, &perms); +aad (sa)->label = &profile->label; + +return aa_check_perms (profile, &perms,AA_MAY_WRITE , sa,NULL ); + } + +return 0 ; +} +static int apparmor_setprocattr (const char *name,void *value,size_t size) {
@@ -632,6 +651,19 @@ static int apparmor_setprocattr(const char *name, void *value,if (size ==0 )return -EINVAL ; +/* LSM infrastructure does actual setting of interface_lsm if allowed */ +if (!strcmp (name,"interface_lsm" )) { +struct aa_profile *profile; +struct aa_label *label; + +aad (&sa)->info ="set interface lsm" ; + label =begin_current_label_crit_section (); + error =fn_for_each_confined (label, profile, +profile_interface_lsm (profile, &sa)); +end_current_label_crit_section (label); +return error; + } +/* AppArmor requires that the buffer must be null terminated atm */ if (args[size -1 ] !='\0' ) {/* null terminate */
security/security.c +149 -17
251 lines not on this page · raw message
security/selinux/hooks.c +11 -0
@@ -6510,6 +6510,17 @@ static int selinux_setprocattr(const char *name, void *value, size_t size)/* * Basic control over ability to set these attributes at all. */ + +/* +* For setting interface_lsm, we only perform a permission check; +* the actual update to the interface_lsm value is handled by the +* LSM framework. +*/ +if (!strcmp (name,"interface_lsm" )) +return avc_has_perm (&selinux_state, + mysid, mysid,SECCLASS_PROCESS2 , +PROCESS2__SETDISPLAY ,NULL ); +if (!strcmp (name,"exec" )) error = avc_has_perm(&selinux_state, mysid, mysid, SECCLASS_PROCESS,
security/selinux/include/classmap.h +1 -1
@@ -53,7 +53,7 @@ struct security_class_mapping secclass_map[] = {"execmem" ,"execstack" ,"execheap" ,"setkeycreate" , "setsockcreate", "getrlimit", NULL } }, {"process2" , - { "nnp_transition", "nosuid_transition", NULL } }, + { "nnp_transition", "nosuid_transition", "setdisplay", NULL } }, {"system" , { "ipc_info", "syslog_read", "syslog_mod", "syslog_console", "module_request", "module_load", NULL } },
security/smack/smack_lsm.c +7 -0
@@ -3508,6 +3508,13 @@ static int smack_setprocattr(const char *name, void *value, size_t size)struct smack_known_list_elem *sklep;int rc; +/* +* Allow the /proc/.../attr/current and SO_PEERSEC "interface_lsm" +* to be reset at will. +*/ +if (strcmp (name,"interface_lsm" ) ==0 ) +return 0 ; +if (!smack_privileged (CAP_MAC_ADMIN ) &&list_empty (&tsp->smk_relabel))return -EPERM ;
Signature · 1 line
2.29.2
Re: [PATCH v26 14/25] LSM: Specify which LSM to display
Quoted 95 lines · Casey Schaufler
Create a new entry "
interface_lsm" in the procfs attr directory for controlling which LSM security information is displayed for a process. A process can only read or write its own display value.The name of an active LSM that supplies hooks for human readable data may be written to "
interface_lsm" to set the value. The name of the LSM currently in use can be read from "interface_lsm". At this point there can only be one LSM capable of display active. A helper functionlsm_task_ilsm()is provided to get the interface lsm slot for atask_struct.Setting the "
interface_lsm" requires that all security modules using setprocattr hooks allow the action. Each security module is responsible for defining its policy.AppArmor hook provided by John Johansen <[email protected]> SELinux hook provided by Stephen Smalley <[email protected]>
Signed-off-by Casey Schaufler <[email protected]>Cc Kees Cook <[email protected]>Cc Stephen Smalley <[email protected]>Cc Paul Moore <[email protected]>Cc John Johansen <[email protected]>Cc Greg Kroah-Hartman <[email protected]>--- .../ABI/testing/procfs-attr-lsm_display | 22 +++ Documentation/security/lsm.rst | 14 ++ fs/proc/base.c | 1 + include/linux/lsm_hooks.h | 17 ++ security/apparmor/include/apparmor.h | 3 +- security/apparmor/lsm.c | 32 ++++ security/security.c | 166 ++++++++++++++++-- security/selinux/hooks.c | 11 ++ security/selinux/include/classmap.h | 2 +- security/smack/smack_lsm.c | 7 + 10 files changed, 256 insertions(+), 19 deletions(-) create mode 100644 Documentation/ABI/testing/procfs-attr-lsm_displayDocumentation/ABI/testing/procfs-attr-lsm_display new file +22
-0@@ -0,0 +1,22 @@ +What: /proc/*/attr/lsm_display +Contact: [email protected], +Description: The name of the Linux security module (LSM) that will + provide information in the /proc/*/attr/current, + /proc/*/attr/prev and /proc/*/attr/exec interfaces. + The details of permissions required to read from + this interface are dependent on the LSMs active on the + system. + A process cannot write to this interface unless it + refers to itself. + The other details of permissions required to write to + this interface are dependent on the LSMs active on the + system. + The format of the data used by this interface is a + text string identifying the name of an LSM. The values + accepted are: + selinux - the SELinux LSM + smack - the Smack LSM + apparmor - The AppArmor LSM + By convention the LSM names are lower case and do not + contain special characters. +Users: LSM user-spaceDocumentation/security/lsm.rst +14
-0@@ -129,3 +129,17 @@ to identify it as the first security module to be registered. The capabilities security module does not use the general security blobs, unlike other modules. The reasons are historical and are based on overhead, complexity and performance concerns. + +LSM External Interfaces +======================= + +The LSM infrastructure does not generally provide external interfaces. +The individual security modules provide what external interfaces they +require. + +The file ``/sys/kernel/security/lsm`` provides a comma +separated list of the active security modules. + +The file ``/proc/pid/attr/interface_lsm`` contains the name of the security +module for which the ``/proc/pid/attr/current`` interface will +apply. This interface can be written to.fs/proc/base.c +0
-0
fs/proc/base.c static const struct pid_entry attr_dir_stuff[] = { · line 2810
Quoted 4 lines
fs/proc/base.c +1
-0@@ -2807,6 +2807,7 @@ static const struct pid_entry attr_dir_stuff[] = {ATTR (NULL ,"fscreate" ,0666 ),ATTR (NULL ,"keycreate" ,0666 ),ATTR (NULL ,"sockcreate" ,0666 ), + ATTR(NULL, "interface_lsm", 0666),
Anyone can open this file... (continued below)
Quoted 182 lines
fs/proc/base.c +0
-0#ifdef CONFIG_SECURITY_SMACK DIR ("smack" ,0555 , proc_smack_attr_dir_inode_ops, proc_smack_attr_dir_ops),include/linux/lsm_hooks.h +17
-0@@ -1686,4 +1686,21 @@ static inline void security_delete_hooks(struct security_hook_list *hooks,extern int lsm_inode_alloc (struct inode *inode); +/** +* lsm_task_ilsm - the "interface_lsm" for this task +* @task: The task to report on +* +* Returns the task's interface LSM slot. +*/ +static inline int lsm_task_ilsm (struct task_struct *task) +{ +#ifdef CONFIG_SECURITY +int *ilsm = task->security; + +if (ilsm) +return *ilsm; +#endif +return LSMBLOB_INVALID ; +} +#endif /* ! __LINUX_LSM_HOOKS_H */ security/apparmor/include/apparmor.h +2
-1@@ -28,8 +28,9 @@#define AA_CLASS_SIGNAL 10#define AA_CLASS_NET 14#define AA_CLASS_LABEL 16 +#define AA_CLASS_DISPLAY_LSM 17 -#define AA_CLASS_LAST AA_CLASS_LABEL +#define AA_CLASS_LAST AA_CLASS_DISPLAY_LSM/* Control parameters settable through module/boot flags */ extern enum audit_mode aa_g_audit;security/apparmor/lsm.c +32
-0@@ -621,6 +621,25 @@ static int apparmor_getprocattr(struct task_struct *task, char *name,return error; } + +static int profile_interface_lsm (struct aa_profile *profile, +struct common_audit_data *sa) +{ +struct aa_perms perms = { }; +unsigned int state; + + state =PROFILE_MEDIATES (profile,AA_CLASS_DISPLAY_LSM ); +if (state) { +aa_compute_perms (profile->policy.dfa, state, &perms); +aa_apply_modes_to_perms (profile, &perms); +aad (sa)->label = &profile->label; + +return aa_check_perms (profile, &perms,AA_MAY_WRITE , sa,NULL ); + } + +return 0 ; +} +static int apparmor_setprocattr (const char *name,void *value,size_t size) {@@ -632,6 +651,19 @@ static int apparmor_setprocattr(const char *name, void *value,if (size ==0 )return -EINVAL ; +/* LSM infrastructure does actual setting of interface_lsm if allowed */ +if (!strcmp (name,"interface_lsm" )) { +struct aa_profile *profile; +struct aa_label *label; + +aad (&sa)->info ="set interface lsm" ; + label =begin_current_label_crit_section (); + error =fn_for_each_confined (label, profile, +profile_interface_lsm (profile, &sa)); +end_current_label_crit_section (label); +return error; + } +/* AppArmor requires that the buffer must be null terminated atm */ if (args[size -1 ] !='\0' ) {/* null terminate */ security/security.c +43
-4@@ -77,7 +77,16 @@ static struct kmem_cache *lsm_file_cache;static struct kmem_cache *lsm_inode_cache;char *lsm_names; -static struct lsm_blob_sizes blob_sizes __lsm_ro_after_init; + +/* +* The task blob includes the "interface_lsm" slot used for +* chosing which module presents contexts. +* Using a long to avoid potential alignment issues with +* module assigned task blobs. +*/ +static struct lsm_blob_sizes blob_sizes __lsm_ro_after_init = { + .lbs_task =sizeof (long ), +};/* Boot-time LSM user choice */ static __initdata const char *chosen_lsm_order;@@ -669,6 +678,8 @@ int lsm_inode_alloc(struct inode *inode)*/ static int lsm_task_alloc (struct task_struct *task) { +int *ilsm; +if (blob_sizes.lbs_task ==0 ) { task->security =NULL ;return 0 ;@@ -677,6 +688,15 @@ static int lsm_task_alloc(struct task_struct *task) task->security =kzalloc (blob_sizes.lbs_task,GFP_KERNEL );if (task->security ==NULL )return -ENOMEM ; + +/* +* The start of the task blob contains the "interface" LSM slot number. +* Start with it set to the invalid slot number, indicating that the +* default first registered LSM be displayed. +*/ + ilsm = task->security; + *ilsm =LSMBLOB_INVALID ; +return 0 ; }@@ -1732,14 +1752,26 @@ int security_file_open(struct file *file)int security_task_alloc (struct task_struct *task,unsigned long clone_flags) { +int *oilsm = current->security; +int *nilsm;int rc =lsm_task_alloc (task); -if (rc) +if (unlikely (rc))return rc; + rc =call_int_hook (task_alloc,0 , task, clone_flags); -if (unlikely (rc)) +if (unlikely (rc)) {security_task_free (task); -return rc; +return rc; + } + +if (oilsm) { + nilsm = task->security; +if (nilsm) + *nilsm = *oilsm; + } + +return 0 ; }void security_task_free (struct task_struct *task)@@ -2171,23 +2203,110 @@ int security_getprocattr(struct task_struct *p, const char *lsm, char *name, char **value) {struct security_hook_list *hp; +int ilsm =lsm_task_ilsm (current); +int slot =0 ; + +if (!strcmp (name,"interface_lsm" )) { +/* +* lsm_slot will be 0 if there are no displaying modules. +*/
security/security.c int security_getprocattr(struct task_struct *p, const char *lsm, char *name, · line 2222
Quoted 10 lines
security/security.c +10
-0+if (lsm_slot ==0 ) +return -EINVAL ; + +/* +* Only allow getting the current process' interface_lsm. +* There are too few reasons to get another process' +* interface_lsm and too many LSM policy issues. +*/ +if (current != p) +return -EINVAL ;
... but context isn't established by just checking "current", as this file handle may have been given to another process.
I suspect the security_get/setprocattr needs to gain a pointer to "file" so that the f_cred struct can be examined[1] (i.e. compare opener against reader/writer).
[1] https://www.kernel.org/doc/html/latest/security/credentials.html#open-file-credentials
Quoted 51 lines
security/security.c +38
-0+ + ilsm =lsm_task_ilsm (p); +if (ilsm !=LSMBLOB_INVALID ) + slot = ilsm; + *value =kstrdup (lsm_slotlist[slot]->lsm,GFP_KERNEL ); +if (*value) +return strlen (*value); +return -ENOMEM ; + }hlist_for_each_entry (hp, &security_hook_heads.getprocattr, list) {if (lsm !=NULL &&strcmp (lsm, hp->lsmid->lsm))continue ; +if (lsm ==NULL && ilsm !=LSMBLOB_INVALID && + ilsm != hp->lsmid->slot) +continue ;return hp->hook.getprocattr (p, name, value); }return LSM_RET_DEFAULT (getprocattr); } +/** +* security_setprocattr - Set process attributes via /proc +* @lsm: name of module involved, or NULL +* @name: name of the attribute +* @value: value to set the attribute to +* @size: size of the value +* +* Set the process attribute for the specified security module +* to the specified value. Note that this can only be used to set +* the process attributes for the current, or "self" process. +* The /proc code has already done this check. +* +* Returns 0 on success, an appropriate code otherwise. +*/ int security_setprocattr (const char *lsm,const char *name,void *value,size_t size) {struct security_hook_list *hp; +char *termed; +char *copy; +int *ilsm = current->security; +int rc = -EINVAL ; +int slot =0 ; + +if (!strcmp (name,"interface_lsm" )) { +/* +* Change the "interface_lsm" value only if all the security +* modules that support setting a procattr allow it. +* It is assumed that all such security modules will be +* cooperative.
security/security.c int security_getprocattr(struct task_struct *p, const char *lsm, char *name, · line 2283
Quoted 10 lines
security/security.c +10
-0+*/ +if (size ==0 ) +return -EINVAL ; + +hlist_for_each_entry (hp, &security_hook_heads.setprocattr, + list) { + rc = hp->hook.setprocattr (name, value, size); +if (rc <0 ) +return rc; + }
Similarly here -- how do the LSMs reason about who actually _opened_ this file?
-Kees
Quoted 161 lines
security/security.c +48
-13+ + rc = -EINVAL ; + + copy =kmemdup_nul (value, size,GFP_KERNEL ); +if (copy ==NULL ) +return -ENOMEM ; + + termed =strsep (©," \n" ); + +for (slot =0 ; slot < lsm_slot; slot++) +if (!strcmp (termed, lsm_slotlist[slot]->lsm)) { + *ilsm = lsm_slotlist[slot]->slot; + rc = size; +break ; + } + +kfree (termed); +return rc; + }hlist_for_each_entry (hp, &security_hook_heads.setprocattr, list) {if (lsm !=NULL &&strcmp (lsm, hp->lsmid->lsm))continue ; +if (lsm ==NULL && *ilsm !=LSMBLOB_INVALID && + *ilsm != hp->lsmid->slot) +continue ;return hp->hook.setprocattr (name, value, size); }return LSM_RET_DEFAULT (setprocattr);@@ -2207,15 +2326,15 @@ EXPORT_SYMBOL(security_ismaclabel);int security_secid_to_secctx (struct lsmblob *blob,char **secdata,u32 *seclen) {struct security_hook_list *hp; -int rc; +int ilsm =lsm_task_ilsm (current);hlist_for_each_entry (hp, &security_hook_heads.secid_to_secctx, list) {if (WARN_ON (hp->lsmid->slot <0 || hp->lsmid->slot >= lsm_slot))continue ; - rc = hp->hook.secid_to_secctx (blob->secid[hp->lsmid->slot], - secdata, seclen); -if (rc !=LSM_RET_DEFAULT (secid_to_secctx)) -return rc; +if (ilsm ==LSMBLOB_INVALID || ilsm == hp->lsmid->slot) +return hp->hook.secid_to_secctx ( + blob->secid[hp->lsmid->slot], + secdata, seclen); }return LSM_RET_DEFAULT (secid_to_secctx);@@ -2226,16 +2345,15 @@ int security_secctx_to_secid(const char *secdata, u32 seclen, struct lsmblob *blob) {struct security_hook_list *hp; -int rc; +int ilsm =lsm_task_ilsm (current);lsmblob_init (blob,0 );hlist_for_each_entry (hp, &security_hook_heads.secctx_to_secid, list) {if (WARN_ON (hp->lsmid->slot <0 || hp->lsmid->slot >= lsm_slot))continue ; - rc = hp->hook.secctx_to_secid (secdata, seclen, - &blob->secid[hp->lsmid->slot]); -if (rc !=0 ) -return rc; +if (ilsm ==LSMBLOB_INVALID || ilsm == hp->lsmid->slot) +return hp->hook.secctx_to_secid (secdata, seclen, + &blob->secid[hp->lsmid->slot]); }return 0 ; }@@ -2243,7 +2361,14 @@ EXPORT_SYMBOL(security_secctx_to_secid);void security_release_secctx (char *secdata,u32 seclen) { -call_void_hook (release_secctx, secdata, seclen); +struct security_hook_list *hp; +int ilsm =lsm_task_ilsm (current); + +hlist_for_each_entry (hp, &security_hook_heads.release_secctx, list) +if (ilsm ==LSMBLOB_INVALID || ilsm == hp->lsmid->slot) { + hp->hook.release_secctx (secdata, seclen); +return ; + } }EXPORT_SYMBOL (security_release_secctx);@@ -2384,8 +2509,15 @@ EXPORT_SYMBOL(security_sock_rcv_skb);int security_socket_getpeersec_stream (struct socket *sock,char __user *optval,int __user *optlen,unsigned len) { -return call_int_hook (socket_getpeersec_stream, -ENOPROTOOPT , sock, - optval, optlen, len); +int ilsm =lsm_task_ilsm (current); +struct security_hook_list *hp; + +hlist_for_each_entry (hp, &security_hook_heads.socket_getpeersec_stream, + list) +if (ilsm ==LSMBLOB_INVALID || ilsm == hp->lsmid->slot) +return hp->hook.socket_getpeersec_stream (sock, optval, + optlen, len); +return -ENOPROTOOPT ; } int security_socket_getpeersec_dgram(struct socket *sock, struct sk_buff *skb,security/selinux/hooks.c +11
-0@@ -6510,6 +6510,17 @@ static int selinux_setprocattr(const char *name, void *value, size_t size)/* * Basic control over ability to set these attributes at all. */ + +/* +* For setting interface_lsm, we only perform a permission check; +* the actual update to the interface_lsm value is handled by the +* LSM framework. +*/ +if (!strcmp (name,"interface_lsm" )) +return avc_has_perm (&selinux_state, + mysid, mysid,SECCLASS_PROCESS2 , +PROCESS2__SETDISPLAY ,NULL ); +if (!strcmp (name,"exec" )) error = avc_has_perm(&selinux_state, mysid, mysid, SECCLASS_PROCESS,security/selinux/include/classmap.h +1
-1@@ -53,7 +53,7 @@ struct security_class_mapping secclass_map[] = {"execmem" ,"execstack" ,"execheap" ,"setkeycreate" , "setsockcreate", "getrlimit", NULL } }, {"process2" , - { "nnp_transition", "nosuid_transition", NULL } }, + { "nnp_transition", "nosuid_transition", "setdisplay", NULL } }, {"system" , { "ipc_info", "syslog_read", "syslog_mod", "syslog_console", "module_request", "module_load", NULL } },security/smack/smack_lsm.c +7
-0@@ -3508,6 +3508,13 @@ static int smack_setprocattr(const char *name, void *value, size_t size)struct smack_known_list_elem *sklep;int rc; +/* +* Allow the /proc/.../attr/current and SO_PEERSEC "interface_lsm" +* to be reset at will. +*/ +if (strcmp (name,"interface_lsm" ) ==0 ) +return 0 ; +if (!smack_privileged (CAP_MAC_ADMIN ) &&list_empty (&tsp->smk_relabel))return -EPERM ;Signature · 1 line
2.29.2
Signature · 1 line
Kees Cook
Casey Schaufler
maintainerRe: [PATCH v26 14/25] LSM: Specify which LSM to display
Quoted 297 lines · Kees Cook
Quoted 95 lines · Casey Schaufler
Create a new entry "
interface_lsm" in the procfs attr directory for controlling which LSM security information is displayed for a process. A process can only read or write its own display value.The name of an active LSM that supplies hooks for human readable data may be written to "
interface_lsm" to set the value. The name of the LSM currently in use can be read from "interface_lsm". At this point there can only be one LSM capable of display active. A helper functionlsm_task_ilsm()is provided to get the interface lsm slot for atask_struct.Setting the "
interface_lsm" requires that all security modules using setprocattr hooks allow the action. Each security module is responsible for defining its policy.AppArmor hook provided by John Johansen <[email protected]> SELinux hook provided by Stephen Smalley <[email protected]>
Signed-off-by Casey Schaufler <[email protected]>Cc Kees Cook <[email protected]>Cc Stephen Smalley <[email protected]>Cc Paul Moore <[email protected]>Cc John Johansen <[email protected]>Cc Greg Kroah-Hartman <[email protected]>--- .../ABI/testing/procfs-attr-lsm_display | 22 +++ Documentation/security/lsm.rst | 14 ++ fs/proc/base.c | 1 + include/linux/lsm_hooks.h | 17 ++ security/apparmor/include/apparmor.h | 3 +- security/apparmor/lsm.c | 32 ++++ security/security.c | 166 ++++++++++++++++-- security/selinux/hooks.c | 11 ++ security/selinux/include/classmap.h | 2 +- security/smack/smack_lsm.c | 7 + 10 files changed, 256 insertions(+), 19 deletions(-) create mode 100644 Documentation/ABI/testing/procfs-attr-lsm_displayDocumentation/ABI/testing/procfs-attr-lsm_display new file +22
-0@@ -0,0 +1,22 @@ +What: /proc/*/attr/lsm_display +Contact: [email protected], +Description: The name of the Linux security module (LSM) that will + provide information in the /proc/*/attr/current, + /proc/*/attr/prev and /proc/*/attr/exec interfaces. + The details of permissions required to read from + this interface are dependent on the LSMs active on the + system. + A process cannot write to this interface unless it + refers to itself. + The other details of permissions required to write to + this interface are dependent on the LSMs active on the + system. + The format of the data used by this interface is a + text string identifying the name of an LSM. The values + accepted are: + selinux - the SELinux LSM + smack - the Smack LSM + apparmor - The AppArmor LSM + By convention the LSM names are lower case and do not + contain special characters. +Users: LSM user-spaceDocumentation/security/lsm.rst +14
-0@@ -129,3 +129,17 @@ to identify it as the first security module to be registered. The capabilities security module does not use the general security blobs, unlike other modules. The reasons are historical and are based on overhead, complexity and performance concerns. + +LSM External Interfaces +======================= + +The LSM infrastructure does not generally provide external interfaces. +The individual security modules provide what external interfaces they +require. + +The file ``/sys/kernel/security/lsm`` provides a comma +separated list of the active security modules. + +The file ``/proc/pid/attr/interface_lsm`` contains the name of the security +module for which the ``/proc/pid/attr/current`` interface will +apply. This interface can be written to.fs/proc/base.c +0
-0fs/proc/base.c static const struct pid_entry attr_dir_stuff[] = { · line 2810
Quoted 4 lines
fs/proc/base.c +1
-0@@ -2807,6 +2807,7 @@ static const struct pid_entry attr_dir_stuff[] = {ATTR (NULL ,"fscreate" ,0666 ),ATTR (NULL ,"keycreate" ,0666 ),ATTR (NULL ,"sockcreate" ,0666 ), + ATTR(NULL, "interface_lsm", 0666),Anyone can open this file... (continued below)
Quoted 182 lines
fs/proc/base.c +0
-0#ifdef CONFIG_SECURITY_SMACK DIR ("smack" ,0555 , proc_smack_attr_dir_inode_ops, proc_smack_attr_dir_ops),include/linux/lsm_hooks.h +17
-0@@ -1686,4 +1686,21 @@ static inline void security_delete_hooks(struct security_hook_list *hooks,extern int lsm_inode_alloc (struct inode *inode); +/** +* lsm_task_ilsm - the "interface_lsm" for this task +* @task: The task to report on +* +* Returns the task's interface LSM slot. +*/ +static inline int lsm_task_ilsm (struct task_struct *task) +{ +#ifdef CONFIG_SECURITY +int *ilsm = task->security; + +if (ilsm) +return *ilsm; +#endif +return LSMBLOB_INVALID ; +} +#endif /* ! __LINUX_LSM_HOOKS_H */ security/apparmor/include/apparmor.h +2
-1@@ -28,8 +28,9 @@#define AA_CLASS_SIGNAL 10#define AA_CLASS_NET 14#define AA_CLASS_LABEL 16 +#define AA_CLASS_DISPLAY_LSM 17 -#define AA_CLASS_LAST AA_CLASS_LABEL +#define AA_CLASS_LAST AA_CLASS_DISPLAY_LSM/* Control parameters settable through module/boot flags */ extern enum audit_mode aa_g_audit;security/apparmor/lsm.c +32
-0@@ -621,6 +621,25 @@ static int apparmor_getprocattr(struct task_struct *task, char *name,return error; } + +static int profile_interface_lsm (struct aa_profile *profile, +struct common_audit_data *sa) +{ +struct aa_perms perms = { }; +unsigned int state; + + state =PROFILE_MEDIATES (profile,AA_CLASS_DISPLAY_LSM ); +if (state) { +aa_compute_perms (profile->policy.dfa, state, &perms); +aa_apply_modes_to_perms (profile, &perms); +aad (sa)->label = &profile->label; + +return aa_check_perms (profile, &perms,AA_MAY_WRITE , sa,NULL ); + } + +return 0 ; +} +static int apparmor_setprocattr (const char *name,void *value,size_t size) {@@ -632,6 +651,19 @@ static int apparmor_setprocattr(const char *name, void *value,if (size ==0 )return -EINVAL ; +/* LSM infrastructure does actual setting of interface_lsm if allowed */ +if (!strcmp (name,"interface_lsm" )) { +struct aa_profile *profile; +struct aa_label *label; + +aad (&sa)->info ="set interface lsm" ; + label =begin_current_label_crit_section (); + error =fn_for_each_confined (label, profile, +profile_interface_lsm (profile, &sa)); +end_current_label_crit_section (label); +return error; + } +/* AppArmor requires that the buffer must be null terminated atm */ if (args[size -1 ] !='\0' ) {/* null terminate */ security/security.c +43
-4@@ -77,7 +77,16 @@ static struct kmem_cache *lsm_file_cache;static struct kmem_cache *lsm_inode_cache;char *lsm_names; -static struct lsm_blob_sizes blob_sizes __lsm_ro_after_init; + +/* +* The task blob includes the "interface_lsm" slot used for +* chosing which module presents contexts. +* Using a long to avoid potential alignment issues with +* module assigned task blobs. +*/ +static struct lsm_blob_sizes blob_sizes __lsm_ro_after_init = { + .lbs_task =sizeof (long ), +};/* Boot-time LSM user choice */ static __initdata const char *chosen_lsm_order;@@ -669,6 +678,8 @@ int lsm_inode_alloc(struct inode *inode)*/ static int lsm_task_alloc (struct task_struct *task) { +int *ilsm; +if (blob_sizes.lbs_task ==0 ) { task->security =NULL ;return 0 ;@@ -677,6 +688,15 @@ static int lsm_task_alloc(struct task_struct *task) task->security =kzalloc (blob_sizes.lbs_task,GFP_KERNEL );if (task->security ==NULL )return -ENOMEM ; + +/* +* The start of the task blob contains the "interface" LSM slot number. +* Start with it set to the invalid slot number, indicating that the +* default first registered LSM be displayed. +*/ + ilsm = task->security; + *ilsm =LSMBLOB_INVALID ; +return 0 ; }@@ -1732,14 +1752,26 @@ int security_file_open(struct file *file)int security_task_alloc (struct task_struct *task,unsigned long clone_flags) { +int *oilsm = current->security; +int *nilsm;int rc =lsm_task_alloc (task); -if (rc) +if (unlikely (rc))return rc; + rc =call_int_hook (task_alloc,0 , task, clone_flags); -if (unlikely (rc)) +if (unlikely (rc)) {security_task_free (task); -return rc; +return rc; + } + +if (oilsm) { + nilsm = task->security; +if (nilsm) + *nilsm = *oilsm; + } + +return 0 ; }void security_task_free (struct task_struct *task)@@ -2171,23 +2203,110 @@ int security_getprocattr(struct task_struct *p, const char *lsm, char *name, char **value) {struct security_hook_list *hp; +int ilsm =lsm_task_ilsm (current); +int slot =0 ; + +if (!strcmp (name,"interface_lsm" )) { +/* +* lsm_slot will be 0 if there are no displaying modules. +*/ security/security.c int security_getprocattr(struct task_struct *p, const char *lsm, char *name, · line 2222
Quoted 10 lines
security/security.c +10
-0+if (lsm_slot ==0 ) +return -EINVAL ; + +/* +* Only allow getting the current process' interface_lsm. +* There are too few reasons to get another process' +* interface_lsm and too many LSM policy issues. +*/ +if (current != p) +return -EINVAL ;... but context isn't established by just checking "current", as this file handle may have been given to another process.
Quoted 5 lines
I suspect the security_get/setprocattr needs to gain a pointer to "file" so that the
f_credstruct can be examined[1] (i.e. compare opener against reader/writer).[1] https://www.kernel.org/doc/html/latest/security/credentials.html#open-file-credentials
It's not credentials being checked here. The check is whether the task that would be affected is "current". Process A can't open /proc/B/attr/interface_lsm with write access. The only process that can open it for write access is B. If process B opens /proc/B/attr/interface_lsm for write access it could send the file handle to process A, but process A can't write to the file because (current != p) that is, (A != B).
Quoted 227 lines
Quoted 51 lines
security/security.c +38
-0+ + ilsm =lsm_task_ilsm (p); +if (ilsm !=LSMBLOB_INVALID ) + slot = ilsm; + *value =kstrdup (lsm_slotlist[slot]->lsm,GFP_KERNEL ); +if (*value) +return strlen (*value); +return -ENOMEM ; + }hlist_for_each_entry (hp, &security_hook_heads.getprocattr, list) {if (lsm !=NULL &&strcmp (lsm, hp->lsmid->lsm))continue ; +if (lsm ==NULL && ilsm !=LSMBLOB_INVALID && + ilsm != hp->lsmid->slot) +continue ;return hp->hook.getprocattr (p, name, value); }return LSM_RET_DEFAULT (getprocattr); } +/** +* security_setprocattr - Set process attributes via /proc +* @lsm: name of module involved, or NULL +* @name: name of the attribute +* @value: value to set the attribute to +* @size: size of the value +* +* Set the process attribute for the specified security module +* to the specified value. Note that this can only be used to set +* the process attributes for the current, or "self" process. +* The /proc code has already done this check. +* +* Returns 0 on success, an appropriate code otherwise. +*/ int security_setprocattr (const char *lsm,const char *name,void *value,size_t size) {struct security_hook_list *hp; +char *termed; +char *copy; +int *ilsm = current->security; +int rc = -EINVAL ; +int slot =0 ; + +if (!strcmp (name,"interface_lsm" )) { +/* +* Change the "interface_lsm" value only if all the security +* modules that support setting a procattr allow it. +* It is assumed that all such security modules will be +* cooperative. security/security.c int security_getprocattr(struct task_struct *p, const char *lsm, char *name, · line 2283
Quoted 10 lines
security/security.c +10
-0+*/ +if (size ==0 ) +return -EINVAL ; + +hlist_for_each_entry (hp, &security_hook_heads.setprocattr, + list) { + rc = hp->hook.setprocattr (name, value, size); +if (rc <0 ) +return rc; + }Similarly here -- how do the LSMs reason about who actually _opened_ this file?
-Kees
Quoted 161 lines
security/security.c +48
-13+ + rc = -EINVAL ; + + copy =kmemdup_nul (value, size,GFP_KERNEL ); +if (copy ==NULL ) +return -ENOMEM ; + + termed =strsep (©," \n" ); + +for (slot =0 ; slot < lsm_slot; slot++) +if (!strcmp (termed, lsm_slotlist[slot]->lsm)) { + *ilsm = lsm_slotlist[slot]->slot; + rc = size; +break ; + } + +kfree (termed); +return rc; + }hlist_for_each_entry (hp, &security_hook_heads.setprocattr, list) {if (lsm !=NULL &&strcmp (lsm, hp->lsmid->lsm))continue ; +if (lsm ==NULL && *ilsm !=LSMBLOB_INVALID && + *ilsm != hp->lsmid->slot) +continue ;return hp->hook.setprocattr (name, value, size); }return LSM_RET_DEFAULT (setprocattr);@@ -2207,15 +2326,15 @@ EXPORT_SYMBOL(security_ismaclabel);int security_secid_to_secctx (struct lsmblob *blob,char **secdata,u32 *seclen) {struct security_hook_list *hp; -int rc; +int ilsm =lsm_task_ilsm (current);hlist_for_each_entry (hp, &security_hook_heads.secid_to_secctx, list) {if (WARN_ON (hp->lsmid->slot <0 || hp->lsmid->slot >= lsm_slot))continue ; - rc = hp->hook.secid_to_secctx (blob->secid[hp->lsmid->slot], - secdata, seclen); -if (rc !=LSM_RET_DEFAULT (secid_to_secctx)) -return rc; +if (ilsm ==LSMBLOB_INVALID || ilsm == hp->lsmid->slot) +return hp->hook.secid_to_secctx ( + blob->secid[hp->lsmid->slot], + secdata, seclen); }return LSM_RET_DEFAULT (secid_to_secctx);@@ -2226,16 +2345,15 @@ int security_secctx_to_secid(const char *secdata, u32 seclen, struct lsmblob *blob) {struct security_hook_list *hp; -int rc; +int ilsm =lsm_task_ilsm (current);lsmblob_init (blob,0 );hlist_for_each_entry (hp, &security_hook_heads.secctx_to_secid, list) {if (WARN_ON (hp->lsmid->slot <0 || hp->lsmid->slot >= lsm_slot))continue ; - rc = hp->hook.secctx_to_secid (secdata, seclen, - &blob->secid[hp->lsmid->slot]); -if (rc !=0 ) -return rc; +if (ilsm ==LSMBLOB_INVALID || ilsm == hp->lsmid->slot) +return hp->hook.secctx_to_secid (secdata, seclen, + &blob->secid[hp->lsmid->slot]); }return 0 ; }@@ -2243,7 +2361,14 @@ EXPORT_SYMBOL(security_secctx_to_secid);void security_release_secctx (char *secdata,u32 seclen) { -call_void_hook (release_secctx, secdata, seclen); +struct security_hook_list *hp; +int ilsm =lsm_task_ilsm (current); + +hlist_for_each_entry (hp, &security_hook_heads.release_secctx, list) +if (ilsm ==LSMBLOB_INVALID || ilsm == hp->lsmid->slot) { + hp->hook.release_secctx (secdata, seclen); +return ; + } }EXPORT_SYMBOL (security_release_secctx);@@ -2384,8 +2509,15 @@ EXPORT_SYMBOL(security_sock_rcv_skb);int security_socket_getpeersec_stream (struct socket *sock,char __user *optval,int __user *optlen,unsigned len) { -return call_int_hook (socket_getpeersec_stream, -ENOPROTOOPT , sock, - optval, optlen, len); +int ilsm =lsm_task_ilsm (current); +struct security_hook_list *hp; + +hlist_for_each_entry (hp, &security_hook_heads.socket_getpeersec_stream, + list) +if (ilsm ==LSMBLOB_INVALID || ilsm == hp->lsmid->slot) +return hp->hook.socket_getpeersec_stream (sock, optval, + optlen, len); +return -ENOPROTOOPT ; } int security_socket_getpeersec_dgram(struct socket *sock, struct sk_buff *skb,security/selinux/hooks.c +11
-0@@ -6510,6 +6510,17 @@ static int selinux_setprocattr(const char *name, void *value, size_t size)/* * Basic control over ability to set these attributes at all. */ + +/* +* For setting interface_lsm, we only perform a permission check; +* the actual update to the interface_lsm value is handled by the +* LSM framework. +*/ +if (!strcmp (name,"interface_lsm" )) +return avc_has_perm (&selinux_state, + mysid, mysid,SECCLASS_PROCESS2 , +PROCESS2__SETDISPLAY ,NULL ); +if (!strcmp (name,"exec" )) error = avc_has_perm(&selinux_state, mysid, mysid, SECCLASS_PROCESS,security/selinux/include/classmap.h +1
-1@@ -53,7 +53,7 @@ struct security_class_mapping secclass_map[] = {"execmem" ,"execstack" ,"execheap" ,"setkeycreate" , "setsockcreate", "getrlimit", NULL } }, {"process2" , - { "nnp_transition", "nosuid_transition", NULL } }, + { "nnp_transition", "nosuid_transition", "setdisplay", NULL } }, {"system" , { "ipc_info", "syslog_read", "syslog_mod", "syslog_console", "module_request", "module_load", NULL } },security/smack/smack_lsm.c +7
-0@@ -3508,6 +3508,13 @@ static int smack_setprocattr(const char *name, void *value, size_t size)struct smack_known_list_elem *sklep;int rc; +/* +* Allow the /proc/.../attr/current and SO_PEERSEC "interface_lsm" +* to be reset at will. +*/ +if (strcmp (name,"interface_lsm" ) ==0 ) +return 0 ; +if (!smack_privileged (CAP_MAC_ADMIN ) &&list_empty (&tsp->smk_relabel))return -EPERM ;Signature · 1 line
2.29.2
Re: [PATCH v26 14/25] LSM: Specify which LSM to display
Quoted 41 lines · Casey Schaufler
Quoted 40 lines · Kees Cook
Quoted 39 lines · Casey Schaufler
Create a new entry "
interface_lsm" in the procfs attr directory for controlling which LSM security information is displayed for a process. A process can only read or write its own display value.The name of an active LSM that supplies hooks for human readable data may be written to "
interface_lsm" to set the value. The name of the LSM currently in use can be read from "interface_lsm". At this point there can only be one LSM capable of display active. A helper functionlsm_task_ilsm()is provided to get the interface lsm slot for atask_struct.Setting the "
interface_lsm" requires that all security modules using setprocattr hooks allow the action. Each security module is responsible for defining its policy.AppArmor hook provided by John Johansen <[email protected]> SELinux hook provided by Stephen Smalley <[email protected]>
Signed-off-by Casey Schaufler <[email protected]>Cc Kees Cook <[email protected]>Cc Stephen Smalley <[email protected]>Cc Paul Moore <[email protected]>Cc John Johansen <[email protected]>Cc Greg Kroah-Hartman <[email protected]>--- .../ABI/testing/procfs-attr-lsm_display | 22 +++ Documentation/security/lsm.rst | 14 ++ fs/proc/base.c | 1 + include/linux/lsm_hooks.h | 17 ++ security/apparmor/include/apparmor.h | 3 +- security/apparmor/lsm.c | 32 ++++ security/security.c | 166 ++++++++++++++++-- security/selinux/hooks.c | 11 ++ security/selinux/include/classmap.h | 2 +- security/smack/smack_lsm.c | 7 + 10 files changed, 256 insertions(+), 19 deletions(-) create mode 100644 Documentation/ABI/testing/procfs-attr-lsm_display
...
Quoted 28 lines
Quoted 28 lines
Quoted 11 lines
security/security.c +7
-0@@ -2171,23 +2203,110 @@ int security_getprocattr(struct task_struct *p, const char *lsm, char *name, char **value) {struct security_hook_list *hp; +int ilsm =lsm_task_ilsm (current); +int slot =0 ; + +if (!strcmp (name,"interface_lsm" )) { +/* +* lsm_slot will be 0 if there are no displaying modules. +*/ security/security.c int security_getprocattr(struct task_struct *p, const char *lsm, char *name, · line 2222
Quoted 10 lines
security/security.c +10
-0+if (lsm_slot ==0 ) +return -EINVAL ; + +/* +* Only allow getting the current process' interface_lsm. +* There are too few reasons to get another process' +* interface_lsm and too many LSM policy issues. +*/ +if (current != p) +return -EINVAL ;... but context isn't established by just checking "current", as this file handle may have been given to another process.
I suspect the security_get/setprocattr needs to gain a pointer to "file" so that the
f_credstruct can be examined[1] (i.e. compare opener against reader/writer).
Quoted 8 lines
Quoted 1 line
[1] https://www.kernel.org/doc/html/latest/security/credentials.html#open-file-credentials
It's not credentials being checked here. The check is whether the task that would be affected is "current". Process A can't open
/proc/B/attr/interface_lsmwith write access. The only process that can open it for write access is B. If process B opens/proc/B/attr/interface_lsmfor write access it could send the file handle to process A, but process A can't write to the file because (current != p) that is, (A != B).
Agreed.
--
paul moore
www.paul-moore.com
Casey Schaufler
maintainerPATCH v26 15/25 LSM: Ensure the correct LSM context releaser
In reply to Casey SchauflerAdd a new lsmcontext data structure to hold all the information about a "security context", including the string, its size and which LSM allocated the string. The allocation information is necessary because LSMs have different policies regarding the lifecycle of these strings. SELinux allocates and destroys them on each use, whereas Smack provides a pointer to an entry in a list that never goes away.
To: Pablo Neira Ayuso <[email protected]>
Cc: [email protected]
--- drivers/android/binder.c | 10 ++++--- fs/ceph/xattr.c | 6 ++++- fs/nfs/nfs4proc.c | 8 ++++-- fs/nfsd/nfs4xdr.c | 7 +++-- include/linux/security.h | 35 +++++++++++++++++++++++-- include/net/scm.h | 5 +++- kernel/audit.c | 14 +++++++--- kernel/auditsc.c | 12 ++++++--- net/ipv4/ip_sockglue.c | 4 ++- net/netfilter/nf_conntrack_netlink.c | 4 ++- net/netfilter/nf_conntrack_standalone.c | 4 ++- net/netfilter/nfnetlink_queue.c | 13 ++++++--- net/netlabel/netlabel_unlabeled.c | 19 +++++++++++--- net/netlabel/netlabel_user.c | 4 ++- security/security.c | 11 ++++---- 15 files changed, 121 insertions(+), 35 deletions(-)
drivers/android/binder.c +7 -3
@@ -2461,6 +2461,7 @@ static void binder_transaction(struct binder_proc *proc,int t_debug_id =atomic_inc_return (&binder_last_id);char *secctx =NULL ;u32 secctx_sz =0 ; +struct lsmcontext scaff;/* scaffolding */ e =binder_transaction_log_add (&binder_transaction_log); e->debug_id = t_debug_id;
@@ -2772,7 +2773,8 @@ static void binder_transaction(struct binder_proc *proc, t->security_ctx =0 ;WARN_ON (1 ); } -security_release_secctx (secctx, secctx_sz); +lsmcontext_init (&scaff, secctx, secctx_sz,0 ); +security_release_secctx (&scaff); secctx =NULL ; } t->buffer->debug_id = t->debug_id;
@@ -3114,8 +3116,10 @@ static void binder_transaction(struct binder_proc *proc,binder_alloc_free_buf (&target_proc->alloc, t->buffer); err_binder_alloc_buf_failed: err_bad_extra_size: -if (secctx) -security_release_secctx (secctx, secctx_sz); +if (secctx) { +lsmcontext_init (&scaff, secctx, secctx_sz,0 ); +security_release_secctx (&scaff); + } err_get_secctx_failed:kfree (tcomplete);binder_stats_deleted (BINDER_STAT_TRANSACTION_COMPLETE );
fs/ceph/xattr.c +5 -1
@@ -1356,12 +1356,16 @@ int ceph_security_init_secctx(struct dentry *dentry, umode_t mode,void ceph_release_acl_sec_ctx (struct ceph_acl_sec_ctx *as_ctx) { +#ifdef CONFIG_CEPH_FS_SECURITY_LABEL +struct lsmcontext scaff;/* scaffolding */ +#endif #ifdef CONFIG_CEPH_FS_POSIX_ACL posix_acl_release (as_ctx->acl);posix_acl_release (as_ctx->default_acl);#endif #ifdef CONFIG_CEPH_FS_SECURITY_LABEL -security_release_secctx (as_ctx->sec_ctx, as_ctx->sec_ctxlen); +lsmcontext_init (&scaff, as_ctx->sec_ctx, as_ctx->sec_ctxlen,0 ); +security_release_secctx (&scaff);#endif if (as_ctx->pagelist)ceph_pagelist_release (as_ctx->pagelist);
fs/nfs/nfs4proc.c +6 -2
@@ -136,8 +136,12 @@ nfs4_label_init_security(struct inode *dir, struct dentry *dentry,static inline void nfs4_label_release_security (struct nfs4_label *label) { -if (label) -security_release_secctx (label->label, label->len); +struct lsmcontext scaff;/* scaffolding */ + +if (label) { +lsmcontext_init (&scaff, label->label, label->len,0 ); +security_release_secctx (&scaff); + } }static inline u32 *nfs4_bitmask (struct nfs_server *server,struct nfs4_label *label) {
fs/nfsd/nfs4xdr.c +5 -2
@@ -2844,6 +2844,7 @@ nfsd4_encode_fattr(struct xdr_stream *xdr, struct svc_fh *fhp,int err;struct nfs4_acl *acl =NULL ;#ifdef CONFIG_NFSD_V4_SECURITY_LABEL +struct lsmcontext scaff;/* scaffolding */ void *context =NULL ;int contextlen;#endif
@@ -3345,8 +3346,10 @@ nfsd4_encode_fattr(struct xdr_stream *xdr, struct svc_fh *fhp, out:#ifdef CONFIG_NFSD_V4_SECURITY_LABEL -if (context) -security_release_secctx (context, contextlen); +if (context) { +lsmcontext_init (&scaff, context, contextlen,0 );/*scaffolding*/ +security_release_secctx (&scaff); + }#endif /* CONFIG_NFSD_V4_SECURITY_LABEL */ kfree (acl);if (tempfh) {
include/linux/security.h +33 -2
@@ -133,6 +133,37 @@ enum lockdown_reason {extern const char *const lockdown_reasons[LOCKDOWN_CONFIDENTIALITY_MAX +1 ]; +/* +* A "security context" is the text representation of +* the information used by LSMs. +* This structure contains the string, its length, and which LSM +* it is useful for. +*/ +struct lsmcontext { +char *context;/* Provided by the module */ +u32 len; +int slot;/* Identifies the module */ +}; + +/** +* lsmcontext_init - initialize an lsmcontext structure. +* @cp: Pointer to the context to initialize +* @context: Initial context, or NULL +* @size: Size of context, or 0 +* @slot: Which LSM provided the context +* +* Fill in the lsmcontext from the provided information. +* This is a scaffolding function that will be removed when +* lsmcontext integration is complete. +*/ +static inline void lsmcontext_init (struct lsmcontext *cp,char *context, +u32 size,int slot) +{ + cp->slot = slot; + cp->context = context; + cp->len = size; +} +/* * Data exported by the security modules *
@@ -550,7 +581,7 @@ int security_ismaclabel(const char *name);int security_secid_to_secctx (struct lsmblob *blob,char **secdata,u32 *seclen);int security_secctx_to_secid (const char *secdata,u32 seclen,struct lsmblob *blob); -void security_release_secctx (char *secdata,u32 seclen); +void security_release_secctx (struct lsmcontext *cp);void security_inode_invalidate_secctx (struct inode *inode);int security_inode_notifysecctx (struct inode *inode,void *ctx,u32 ctxlen);int security_inode_setsecctx (struct dentry *dentry,void *ctx,u32 ctxlen);
@@ -1414,7 +1445,7 @@ static inline int security_secctx_to_secid(const char *secdata,return -EOPNOTSUPP ; } -static inline void security_release_secctx (char *secdata,u32 seclen) +static inline void security_release_secctx (struct lsmcontext *cp) { }
include/net/scm.h +4 -1
@@ -92,6 +92,7 @@ static __inline__ int scm_send(struct socket *sock, struct msghdr *msg,#ifdef CONFIG_SECURITY_NETWORK static inline void scm_passec (struct socket *sock,struct msghdr *msg,struct scm_cookie *scm) { +struct lsmcontext context;struct lsmblob lb;char *secdata;u32 seclen;
@@ -106,7 +107,9 @@ static inline void scm_passec(struct socket *sock, struct msghdr *msg, struct scif (!err) {put_cmsg (msg,SOL_SOCKET ,SCM_SECURITY , seclen, secdata); -security_release_secctx (secdata, seclen); +/*scaffolding*/ +lsmcontext_init (&context, secdata, seclen,0 ); +security_release_secctx (&context); } } }
kernel/audit.c +10 -4
@@ -1192,6 +1192,7 @@ static int audit_receive_msg(struct sk_buff *skb, struct nlmsghdr *nlh)struct audit_sig_info *sig_data;char *ctx =NULL ;u32 len; +struct lsmcontext scaff;/* scaffolding */ err =audit_netlink_ok (skb, msg_type);if (err)
@@ -1449,15 +1450,18 @@ static int audit_receive_msg(struct sk_buff *skb, struct nlmsghdr *nlh) } sig_data =kmalloc (sizeof (*sig_data) + len,GFP_KERNEL );if (!sig_data) { -if (lsmblob_is_set (&audit_sig_lsm)) -security_release_secctx (ctx, len); +if (lsmblob_is_set (&audit_sig_lsm)) { +lsmcontext_init (&scaff, ctx, len,0 ); +security_release_secctx (&scaff); + }return -ENOMEM ; } sig_data->uid =from_kuid (&init_user_ns, audit_sig_uid); sig_data->pid = audit_sig_pid;if (lsmblob_is_set (&audit_sig_lsm)) {memcpy (sig_data->ctx, ctx, len); -security_release_secctx (ctx, len); +lsmcontext_init (&scaff, ctx, len,0 ); +security_release_secctx (&scaff); }audit_send_reply (skb, seq,AUDIT_SIGNAL_INFO ,0 ,0 , sig_data,sizeof (*sig_data) + len);
@@ -2132,6 +2136,7 @@ int audit_log_task_context(struct audit_buffer *ab)unsigned len;int error;struct lsmblob blob; +struct lsmcontext scaff;/* scaffolding */ security_task_getsecid_subj (current, &blob);if (!lsmblob_is_set (&blob))
@@ -2145,7 +2150,8 @@ int audit_log_task_context(struct audit_buffer *ab) }audit_log_format (ab," subj=%s" , ctx); -security_release_secctx (ctx, len); +lsmcontext_init (&scaff, ctx, len,0 ); +security_release_secctx (&scaff);return 0 ; error_path:
kernel/auditsc.c +9 -3
@@ -996,6 +996,7 @@ static int audit_log_pid_context(struct audit_context *context, pid_t pid, struct lsmblob *blob, char *comm) { struct audit_buffer *ab; +struct lsmcontext lsmcxt;char *ctx =NULL ;u32 len;int rc =0 ;
@@ -1013,7 +1014,8 @@ static int audit_log_pid_context(struct audit_context *context, pid_t pid, rc =1 ; }else {audit_log_format (ab," obj=%s" , ctx); -security_release_secctx (ctx, len); +lsmcontext_init (&lsmcxt, ctx, len,0 );/*scaffolding*/ +security_release_secctx (&lsmcxt); } }audit_log_format (ab," ocomm=" );
@@ -1226,6 +1228,7 @@ static void audit_log_fcaps(struct audit_buffer *ab, struct audit_names *name)static void show_special (struct audit_context *context,int *call_panic) { +struct lsmcontext lsmcxt;struct audit_buffer *ab;int i;
@@ -1259,7 +1262,8 @@ static void show_special(struct audit_context *context, int *call_panic) *call_panic =1 ; }else {audit_log_format (ab," obj=%s" , ctx); -security_release_secctx (ctx, len); +lsmcontext_init (&lsmcxt, ctx, len,0 ); +security_release_secctx (&lsmcxt); } }if (context->ipc.has_perm) {
@@ -1408,6 +1412,7 @@ static void audit_log_name(struct audit_context *context, struct audit_names *n,char *ctx =NULL ;u32 len;struct lsmblob blob; +struct lsmcontext lsmcxt;lsmblob_init (&blob, n->osid);if (security_secid_to_secctx (&blob, &ctx, &len)) {
@@ -1416,7 +1421,8 @@ static void audit_log_name(struct audit_context *context, struct audit_names *n, *call_panic =2 ; }else {audit_log_format (ab," obj=%s" , ctx); -security_release_secctx (ctx, len); +lsmcontext_init (&lsmcxt, ctx, len,0 );/* scaffolding */ +security_release_secctx (&lsmcxt); } }
net/ipv4/ip_sockglue.c +3 -1
@@ -130,6 +130,7 @@ static void ip_cmsg_recv_checksum(struct msghdr *msg, struct sk_buff *skb,static void ip_cmsg_recv_security (struct msghdr *msg,struct sk_buff *skb) { +struct lsmcontext context;struct lsmblob lb;char *secdata;u32 seclen, secid;
@@ -145,7 +146,8 @@ static void ip_cmsg_recv_security(struct msghdr *msg, struct sk_buff *skb)return ;put_cmsg (msg,SOL_IP ,SCM_SECURITY , seclen, secdata); -security_release_secctx (secdata, seclen); +lsmcontext_init (&context, secdata, seclen,0 );/* scaffolding */ +security_release_secctx (&context); }static void ip_cmsg_recv_dstaddr (struct msghdr *msg,struct sk_buff *skb)
net/netfilter/nf_conntrack_netlink.c +3 -1
@@ -339,6 +339,7 @@ static int ctnetlink_dump_secctx(struct sk_buff *skb, const struct nf_conn *ct)int len, ret;char *secctx;struct lsmblob blob; +struct lsmcontext context;/* lsmblob_init() puts ct->secmark into all of the secids in blob. * security_secid_to_secctx() will know which security module
@@ -359,7 +360,8 @@ static int ctnetlink_dump_secctx(struct sk_buff *skb, const struct nf_conn *ct) ret =0 ; nla_put_failure: -security_release_secctx (secctx, len); +lsmcontext_init (&context, secctx, len,0 );/* scaffolding */ +security_release_secctx (&context); return ret; }#else
net/netfilter/nf_conntrack_standalone.c +3 -1
@@ -176,6 +176,7 @@ static void ct_show_secctx(struct seq_file *s, const struct nf_conn *ct)u32 len;char *secctx;struct lsmblob blob; +struct lsmcontext context;lsmblob_init (&blob, ct->secmark); ret =security_secid_to_secctx (&blob, &secctx, &len);
@@ -184,7 +185,8 @@ static void ct_show_secctx(struct seq_file *s, const struct nf_conn *ct)seq_printf (s,"secctx=%s " , secctx); - security_release_secctx(secctx, len); +lsmcontext_init (&context, secctx, len,0 );/* scaffolding */ + security_release_secctx(&context); }#else static inline void ct_show_secctx (struct seq_file *s,const struct nf_conn *ct)
net/netfilter/nfnetlink_queue.c +9 -4
@@ -397,6 +397,7 @@ nfqnl_build_packet_message(struct net *net, struct nfqnl_instance *queue,enum ip_conntrack_info ctinfo;struct nfnl_ct_hook *nfnl_ct;bool csum_verify; +struct lsmcontext scaff;/* scaffolding */ char *secdata =NULL ;u32 seclen =0 ;
@@ -626,8 +627,10 @@ nfqnl_build_packet_message(struct net *net, struct nfqnl_instance *queue, } nlh->nlmsg_len = skb->len; -if (seclen) -security_release_secctx (secdata, seclen); +if (seclen) { +lsmcontext_init (&scaff, secdata, seclen,0 ); +security_release_secctx (&scaff); + }return skb; nla_put_failure:
@@ -635,8 +638,10 @@ nfqnl_build_packet_message(struct net *net, struct nfqnl_instance *queue,kfree_skb (skb);net_err_ratelimited ("nf_queue: error creating packet message\n" ); nlmsg_failure: -if (seclen) -security_release_secctx (secdata, seclen); +if (seclen) { +lsmcontext_init (&scaff, secdata, seclen,0 ); +security_release_secctx (&scaff); + }return NULL ; }
net/netlabel/netlabel_unlabeled.c +15 -4
@@ -374,6 +374,7 @@ int netlbl_unlhsh_add(struct net *net,struct net_device *dev;struct netlbl_unlhsh_iface *iface;struct audit_buffer *audit_buf =NULL ; +struct lsmcontext context;char *secctx =NULL ;u32 secctx_len;struct lsmblob blob;
@@ -447,7 +448,9 @@ int netlbl_unlhsh_add(struct net *net, &secctx, &secctx_len) == 0) {audit_log_format (audit_buf," sec_obj=%s" , secctx); -security_release_secctx (secctx, secctx_len); +/* scaffolding */ +lsmcontext_init (&context, secctx, secctx_len,0 ); +security_release_secctx (&context); }audit_log_format (audit_buf," res=%u" , ret_val ==0 ?1 :0 );audit_log_end (audit_buf);
@@ -478,6 +481,7 @@ static int netlbl_unlhsh_remove_addr4(struct net *net,struct netlbl_unlhsh_addr4 *entry;struct audit_buffer *audit_buf;struct net_device *dev; +struct lsmcontext context;char *secctx;u32 secctx_len;struct lsmblob blob;
@@ -509,7 +513,9 @@ static int netlbl_unlhsh_remove_addr4(struct net *net, security_secid_to_secctx(&blob, &secctx, &secctx_len) == 0) {audit_log_format (audit_buf," sec_obj=%s" , secctx); -security_release_secctx (secctx, secctx_len); +/* scaffolding */ +lsmcontext_init (&context, secctx, secctx_len,0 ); +security_release_secctx (&context); }audit_log_format (audit_buf," res=%u" , entry !=NULL ?1 :0 );audit_log_end (audit_buf);
@@ -546,6 +552,7 @@ static int netlbl_unlhsh_remove_addr6(struct net *net,struct netlbl_unlhsh_addr6 *entry;struct audit_buffer *audit_buf;struct net_device *dev; +struct lsmcontext context;char *secctx;u32 secctx_len;struct lsmblob blob;
@@ -576,7 +583,8 @@ static int netlbl_unlhsh_remove_addr6(struct net *net, security_secid_to_secctx(&blob, &secctx, &secctx_len) == 0) {audit_log_format (audit_buf," sec_obj=%s" , secctx); -security_release_secctx (secctx, secctx_len); +lsmcontext_init (&context, secctx, secctx_len,0 ); +security_release_secctx (&context); }audit_log_format (audit_buf," res=%u" , entry !=NULL ?1 :0 );audit_log_end (audit_buf);
@@ -1095,6 +1103,7 @@ static int netlbl_unlabel_staticlist_gen(u32 cmd,int ret_val = -ENOMEM ;struct netlbl_unlhsh_walk_arg *cb_arg = arg;struct net_device *dev; +struct lsmcontext context;void *data;u32 secid;char *secctx;
@@ -1165,7 +1174,9 @@ static int netlbl_unlabel_staticlist_gen(u32 cmd,NLBL_UNLABEL_A_SECCTX , secctx_len, secctx); -security_release_secctx (secctx, secctx_len); +/* scaffolding */ +lsmcontext_init (&context, secctx, secctx_len,0 ); +security_release_secctx (&context);if (ret_val !=0 ) goto list_cb_failure;
net/netlabel/netlabel_user.c +3 -1
@@ -84,6 +84,7 @@ struct audit_buffer *netlbl_audit_start_common(int type, struct netlbl_audit *audit_info) {struct audit_buffer *audit_buf; +struct lsmcontext context;char *secctx;u32 secctx_len;struct lsmblob blob;
@@ -103,7 +104,8 @@ struct audit_buffer *netlbl_audit_start_common(int type,if (audit_info->secid !=0 &&security_secid_to_secctx (&blob, &secctx, &secctx_len) ==0 ) {audit_log_format (audit_buf," subj=%s" , secctx); -security_release_secctx (secctx, secctx_len); +lsmcontext_init (&context, secctx, secctx_len,0 );/*scaffolding*/ +security_release_secctx (&context); }return audit_buf;
security/security.c +6 -5
@@ -2359,16 +2359,17 @@ int security_secctx_to_secid(const char *secdata, u32 seclen, }EXPORT_SYMBOL (security_secctx_to_secid); -void security_release_secctx (char *secdata,u32 seclen) +void security_release_secctx (struct lsmcontext *cp) {struct security_hook_list *hp; -int ilsm =lsm_task_ilsm (current);hlist_for_each_entry (hp, &security_hook_heads.release_secctx, list) -if (ilsm ==LSMBLOB_INVALID || ilsm == hp->lsmid->slot) { - hp->hook.release_secctx (secdata, seclen); -return ; +if (cp->slot == hp->lsmid->slot) { + hp->hook.release_secctx (cp->context, cp->len); +break ; } + +memset (cp,0 ,sizeof (*cp)); }EXPORT_SYMBOL (security_release_secctx);
Signature · 1 line
2.29.2
Re: [PATCH v26 15/25] LSM: Ensure the correct LSM context releaser
Quoted 29 lines · Casey Schaufler
Add a new lsmcontext data structure to hold all the information about a "security context", including the string, its size and which LSM allocated the string. The allocation information is necessary because LSMs have different policies regarding the lifecycle of these strings. SELinux allocates and destroys them on each use, whereas Smack provides a pointer to an entry in a list that never goes away.
Reviewed-by Kees Cook <[email protected]>Reviewed-by John Johansen <[email protected]>Acked-by Stephen Smalley <[email protected]>Acked-by Chuck Lever <[email protected]>Signed-off-by Casey Schaufler <[email protected]>To: Pablo Neira Ayuso <[email protected]>
Cc: [email protected]--- drivers/android/binder.c | 10 ++++--- fs/ceph/xattr.c | 6 ++++- fs/nfs/nfs4proc.c | 8 ++++-- fs/nfsd/nfs4xdr.c | 7 +++-- include/linux/security.h | 35 +++++++++++++++++++++++-- include/net/scm.h | 5 +++- kernel/audit.c | 14 +++++++--- kernel/auditsc.c | 12 ++++++---
Quoted 8 lines
net/ipv4/ip_sockglue.c | 4 ++- net/netfilter/nf_conntrack_netlink.c | 4 ++- net/netfilter/nf_conntrack_standalone.c | 4 ++- net/netfilter/nfnetlink_queue.c | 13 ++++++--- net/netlabel/netlabel_unlabeled.c | 19 +++++++++++--- net/netlabel/netlabel_user.c | 4 ++- security/security.c | 11 ++++---- 15 files changed, 121 insertions(+), 35 deletions(-)
--
paul moore
www.paul-moore.com
--
Linux-audit mailing list
[email protected]
https://listman.redhat.com/mailman/listinfo/linux-audit
Casey Schaufler
maintainerPATCH v26 16/25 LSM: Use lsmcontext in security_secid_to_secctx
In reply to Casey SchauflerReplace the (secctx,seclen) pointer pair with a single lsmcontext pointer to allow return of the LSM identifier along with the context and context length. This allows security_release_secctx() to know how to release the context. Callers have been modified to use or save the returned data from the new structure.
--- drivers/android/binder.c | 26 +++++++--------- include/linux/security.h | 4 +-- include/net/scm.h | 9 ++---- kernel/audit.c | 39 +++++++++++------------- kernel/auditsc.c | 31 +++++++------------ net/ipv4/ip_sockglue.c | 8 ++--- net/netfilter/nf_conntrack_netlink.c | 18 +++++------ net/netfilter/nf_conntrack_standalone.c | 7 ++--- net/netfilter/nfnetlink_queue.c | 5 +++- net/netlabel/netlabel_unlabeled.c | 40 ++++++++----------------- net/netlabel/netlabel_user.c | 7 ++--- security/security.c | 10 +++++-- 12 files changed, 81 insertions(+), 123 deletions(-)
drivers/android/binder.c +10 -16
@@ -2459,9 +2459,7 @@ static void binder_transaction(struct binder_proc *proc,binder_size_t last_fixup_min_off =0 ;struct binder_context *context = proc->context;int t_debug_id =atomic_inc_return (&binder_last_id); -char *secctx =NULL ; -u32 secctx_sz =0 ; -struct lsmcontext scaff;/* scaffolding */ +struct lsmcontext lsmctx = { }; e =binder_transaction_log_add (&binder_transaction_log); e->debug_id = t_debug_id;
@@ -2724,14 +2722,14 @@ static void binder_transaction(struct binder_proc *proc,* case well anyway. */ security_task_getsecid_obj (proc->tsk, &blob); - ret =security_secid_to_secctx (&blob, &secctx, &secctx_sz); + ret =security_secid_to_secctx (&blob, &lsmctx);if (ret) { return_error =BR_FAILED_REPLY ; return_error_param = ret; return_error_line = __LINE__; goto err_get_secctx_failed; } - added_size =ALIGN (secctx_sz,sizeof (u64)); + added_size =ALIGN (lsmctx.len,sizeof (u64)); extra_buffers_size += added_size;if (extra_buffers_size < added_size) {/* integer overflow of extra_buffers_size */
@@ -2758,24 +2756,22 @@ static void binder_transaction(struct binder_proc *proc, t->buffer =NULL ; goto err_binder_alloc_buf_failed; } -if (secctx) { +if (lsmctx.context) {int err;size_t buf_offset =ALIGN (tr->data_size,sizeof (void *)) +ALIGN (tr->offsets_size,sizeof (void *)) +ALIGN (extra_buffers_size,sizeof (void *)) - -ALIGN (secctx_sz,sizeof (u64)); +ALIGN (lsmctx.len,sizeof (u64)); t->security_ctx = (uintptr_t )t->buffer->user_data + buf_offset; err =binder_alloc_copy_to_buffer (&target_proc->alloc, t->buffer, buf_offset, - secctx, secctx_sz); + lsmctx.context, lsmctx.len);if (err) { t->security_ctx =0 ;WARN_ON (1 ); } -lsmcontext_init (&scaff, secctx, secctx_sz,0 ); -security_release_secctx (&scaff); - secctx =NULL ; +security_release_secctx (&lsmctx); } t->buffer->debug_id = t->debug_id; t->buffer->transaction = t;
@@ -2832,7 +2828,7 @@ static void binder_transaction(struct binder_proc *proc, off_end_offset = off_start_offset + tr->offsets_size; sg_buf_offset =ALIGN (off_end_offset,sizeof (void *)); sg_buf_end_offset = sg_buf_offset + extra_buffers_size - -ALIGN (secctx_sz,sizeof (u64)); +ALIGN (lsmctx.len,sizeof (u64)); off_min =0 ;for (buffer_offset = off_start_offset; buffer_offset < off_end_offset; buffer_offset +=sizeof (binder_size_t)) {
@@ -3116,10 +3112,8 @@ static void binder_transaction(struct binder_proc *proc,binder_alloc_free_buf (&target_proc->alloc, t->buffer); err_binder_alloc_buf_failed: err_bad_extra_size: -if (secctx) { -lsmcontext_init (&scaff, secctx, secctx_sz,0 ); -security_release_secctx (&scaff); - } +if (lsmctx.context) +security_release_secctx (&lsmctx); err_get_secctx_failed:kfree (tcomplete);binder_stats_deleted (BINDER_STAT_TRANSACTION_COMPLETE );
include/linux/security.h +2 -2
@@ -578,7 +578,7 @@ int security_setprocattr(const char *lsm, const char *name, void *value, size_t size);int security_netlink_send (struct sock *sk,struct sk_buff *skb);int security_ismaclabel (const char *name); -int security_secid_to_secctx (struct lsmblob *blob,char **secdata,u32 *seclen); +int security_secid_to_secctx (struct lsmblob *blob,struct lsmcontext *cp);int security_secctx_to_secid (const char *secdata,u32 seclen,struct lsmblob *blob);void security_release_secctx (struct lsmcontext *cp);
@@ -1433,7 +1433,7 @@ static inline int security_ismaclabel(const char *name) }static inline int security_secid_to_secctx (struct lsmblob *blob, -char **secdata,u32 *seclen) +struct lsmcontext *cp) {return -EOPNOTSUPP ; }
include/net/scm.h +3 -6
@@ -94,8 +94,6 @@ static inline void scm_passec(struct socket *sock, struct msghdr *msg, struct sc {struct lsmcontext context;struct lsmblob lb; -char *secdata; -u32 seclen;int err;if (test_bit (SOCK_PASSSEC , &sock->flags)) {
@@ -103,12 +101,11 @@ static inline void scm_passec(struct socket *sock, struct msghdr *msg, struct sc* and the infrastructure will know which it is. */ lsmblob_init (&lb, scm->secid); - err =security_secid_to_secctx (&lb, &secdata, &seclen); + err =security_secid_to_secctx (&lb, &context);if (!err) { -put_cmsg (msg,SOL_SOCKET ,SCM_SECURITY , seclen, secdata); -/*scaffolding*/ -lsmcontext_init (&context, secdata, seclen,0 ); +put_cmsg (msg,SOL_SOCKET ,SCM_SECURITY , context.len, + context.context);security_release_secctx (&context); } }
kernel/audit.c +17 -22
@@ -1190,9 +1190,6 @@ static int audit_receive_msg(struct sk_buff *skb, struct nlmsghdr *nlh)struct audit_buffer *ab;u16 msg_type = nlh->nlmsg_type;struct audit_sig_info *sig_data; -char *ctx =NULL ; -u32 len; -struct lsmcontext scaff;/* scaffolding */ err =audit_netlink_ok (skb, msg_type);if (err)
@@ -1440,33 +1437,34 @@ static int audit_receive_msg(struct sk_buff *skb, struct nlmsghdr *nlh)kfree (new);break ; } -case AUDIT_SIGNAL_INFO : - len =0 ; +case AUDIT_SIGNAL_INFO : { +struct lsmcontext context = { }; +int len =0 ; +if (lsmblob_is_set (&audit_sig_lsm)) { - err =security_secid_to_secctx (&audit_sig_lsm, &ctx, - &len); + err =security_secid_to_secctx (&audit_sig_lsm, + &context);if (err)return err; } - sig_data =kmalloc (sizeof (*sig_data) + len,GFP_KERNEL ); + sig_data =kmalloc (sizeof (*sig_data) + context.len,GFP_KERNEL );if (!sig_data) { -if (lsmblob_is_set (&audit_sig_lsm)) { -lsmcontext_init (&scaff, ctx, len,0 ); -security_release_secctx (&scaff); - } +if (lsmblob_is_set (&audit_sig_lsm)) +security_release_secctx (&context);return -ENOMEM ; } sig_data->uid =from_kuid (&init_user_ns, audit_sig_uid); sig_data->pid = audit_sig_pid;if (lsmblob_is_set (&audit_sig_lsm)) { -memcpy (sig_data->ctx, ctx, len); -lsmcontext_init (&scaff, ctx, len,0 ); -security_release_secctx (&scaff); + len = context.len; +memcpy (sig_data->ctx, context.context, len); +security_release_secctx (&context); }audit_send_reply (skb, seq,AUDIT_SIGNAL_INFO ,0 ,0 , sig_data,sizeof (*sig_data) + len);kfree (sig_data);break ; + }case AUDIT_TTY_GET : {struct audit_tty_status s;unsigned int t;
@@ -2132,26 +2130,23 @@ void audit_log_key(struct audit_buffer *ab, char *key)int audit_log_task_context (struct audit_buffer *ab) { -char *ctx =NULL ; -unsigned len;int error;struct lsmblob blob; -struct lsmcontext scaff;/* scaffolding */ +struct lsmcontext context;security_task_getsecid_subj (current, &blob);if (!lsmblob_is_set (&blob))return 0 ; - error =security_secid_to_secctx (&blob, &ctx, &len); + error =security_secid_to_secctx (&blob, &context);if (error) {if (error != -EINVAL ) goto error_path;return 0 ; } -audit_log_format (ab," subj=%s" , ctx); -lsmcontext_init (&scaff, ctx, len,0 ); -security_release_secctx (&scaff); +audit_log_format (ab," subj=%s" , context.context); +security_release_secctx (&context);return 0 ; error_path:
kernel/auditsc.c +11 -20
@@ -996,9 +996,7 @@ static int audit_log_pid_context(struct audit_context *context, pid_t pid, struct lsmblob *blob, char *comm) { struct audit_buffer *ab; -struct lsmcontext lsmcxt; -char *ctx =NULL ; -u32 len; +struct lsmcontext lsmctx;int rc =0 ; ab =audit_log_start (context,GFP_KERNEL ,AUDIT_OBJ_PID );
@@ -1009,13 +1007,12 @@ static int audit_log_pid_context(struct audit_context *context, pid_t pid,from_kuid (&init_user_ns, auid), from_kuid(&init_user_ns, uid), sessionid);if (lsmblob_is_set (blob)) { -if (security_secid_to_secctx (blob, &ctx, &len)) { +if (security_secid_to_secctx (blob, &lsmctx)) {audit_log_format (ab," obj=(none)" ); rc =1 ; }else { -audit_log_format (ab," obj=%s" , ctx); -lsmcontext_init (&lsmcxt, ctx, len,0 );/*scaffolding*/ -security_release_secctx (&lsmcxt); +audit_log_format (ab," obj=%s" , lsmctx.context); +security_release_secctx (&lsmctx); } }audit_log_format (ab," ocomm=" );
@@ -1228,7 +1225,6 @@ static void audit_log_fcaps(struct audit_buffer *ab, struct audit_names *name)static void show_special (struct audit_context *context,int *call_panic) { -struct lsmcontext lsmcxt;struct audit_buffer *ab;int i;
@@ -1252,17 +1248,15 @@ static void show_special(struct audit_context *context, int *call_panic)from_kgid (&init_user_ns, context->ipc.gid), context->ipc.mode);if (osid) { -char *ctx =NULL ; -u32 len; +struct lsmcontext lsmcxt;struct lsmblob blob;lsmblob_init (&blob, osid); -if (security_secid_to_secctx (&blob, &ctx, &len)) { +if (security_secid_to_secctx (&blob, &lsmcxt)) {audit_log_format (ab," osid=%u" , osid); *call_panic =1 ; }else { -audit_log_format (ab," obj=%s" , ctx); -lsmcontext_init (&lsmcxt, ctx, len,0 ); +audit_log_format (ab," obj=%s" , lsmcxt.context);security_release_secctx (&lsmcxt); } }
@@ -1409,20 +1403,17 @@ static void audit_log_name(struct audit_context *context, struct audit_names *n,MAJOR (n->rdev), MINOR(n->rdev));if (n->osid !=0 ) { -char *ctx =NULL ; -u32 len;struct lsmblob blob; -struct lsmcontext lsmcxt; +struct lsmcontext lsmctx;lsmblob_init (&blob, n->osid); -if (security_secid_to_secctx (&blob, &ctx, &len)) { +if (security_secid_to_secctx (&blob, &lsmctx)) {audit_log_format (ab," osid=%u" , n->osid);if (call_panic) *call_panic =2 ; }else { -audit_log_format (ab," obj=%s" , ctx); -lsmcontext_init (&lsmcxt, ctx, len,0 );/* scaffolding */ -security_release_secctx (&lsmcxt); +audit_log_format (ab," obj=%s" , lsmctx.context); +security_release_secctx (&lsmctx); } }
net/ipv4/ip_sockglue.c +3 -5
@@ -132,8 +132,7 @@ static void ip_cmsg_recv_security(struct msghdr *msg, struct sk_buff *skb) {struct lsmcontext context;struct lsmblob lb; -char *secdata; -u32 seclen, secid; +u32 secid;int err; err =security_socket_getpeersec_dgram (NULL , skb, &secid);
@@ -141,12 +140,11 @@ static void ip_cmsg_recv_security(struct msghdr *msg, struct sk_buff *skb)return ;lsmblob_init (&lb, secid); - err =security_secid_to_secctx (&lb, &secdata, &seclen); + err =security_secid_to_secctx (&lb, &context);if (err)return ; -put_cmsg (msg,SOL_IP ,SCM_SECURITY , seclen, secdata); -lsmcontext_init (&context, secdata, seclen,0 );/* scaffolding */ +put_cmsg (msg,SOL_IP ,SCM_SECURITY , context.len, context.context);security_release_secctx (&context); }
net/netfilter/nf_conntrack_netlink.c +8 -10
@@ -336,8 +336,7 @@ static int ctnetlink_dump_mark(struct sk_buff *skb, const struct nf_conn *ct)static int ctnetlink_dump_secctx (struct sk_buff *skb,const struct nf_conn *ct) {struct nlattr *nest_secctx; -int len, ret; -char *secctx; +int ret;struct lsmblob blob;struct lsmcontext context;
@@ -345,7 +344,7 @@ static int ctnetlink_dump_secctx(struct sk_buff *skb, const struct nf_conn *ct)* security_secid_to_secctx() will know which security module * to use to create the secctx. */ lsmblob_init (&blob, ct->secmark); - ret =security_secid_to_secctx (&blob, &secctx, &len); + ret =security_secid_to_secctx (&blob, &context);if (ret)return 0 ;
@@ -354,13 +353,12 @@ static int ctnetlink_dump_secctx(struct sk_buff *skb, const struct nf_conn *ct)if (!nest_secctx) goto nla_put_failure; -if (nla_put_string (skb,CTA_SECCTX_NAME , secctx)) +if (nla_put_string (skb,CTA_SECCTX_NAME , context.context)) goto nla_put_failure;nla_nest_end (skb, nest_secctx); ret =0 ; nla_put_failure: -lsmcontext_init (&context, secctx, len,0 );/* scaffolding */ security_release_secctx (&context);return ret; }
@@ -655,15 +653,15 @@ static inline int ctnetlink_secctx_size(const struct nf_conn *ct)#ifdef CONFIG_NF_CONNTRACK_SECMARK int len, ret;struct lsmblob blob; +struct lsmcontext context; -/* lsmblob_init() puts ct->secmark into all of the secids in blob. -* security_secid_to_secctx() will know which security module -* to use to create the secctx. */ -lsmblob_init (&blob, ct->secmark); - ret =security_secid_to_secctx (&blob,NULL , &len); + ret =security_secid_to_secctx (&blob, &context);if (ret)return 0 ; + len = context.len; +security_release_secctx (&context); +return nla_total_size (0 )/* CTA_SECCTX */ +nla_total_size (sizeof (char ) * len);/* CTA_SECCTX_NAME */ #else
net/netfilter/nf_conntrack_standalone.c +2 -5
@@ -173,19 +173,16 @@ static void ct_seq_stop(struct seq_file *s, void *v)static void ct_show_secctx (struct seq_file *s,const struct nf_conn *ct) {int ret; -u32 len; -char *secctx;struct lsmblob blob;struct lsmcontext context;lsmblob_init (&blob, ct->secmark); - ret =security_secid_to_secctx (&blob, &secctx, &len); + ret =security_secid_to_secctx (&blob, &context);if (ret)return ; -seq_printf (s,"secctx=%s " , secctx); +seq_printf (s,"secctx=%s " , context.context); -lsmcontext_init (&context, secctx, len,0 );/* scaffolding */ security_release_secctx (&context); }#else
net/netfilter/nfnetlink_queue.c +4 -1
@@ -306,6 +306,7 @@ static u32 nfqnl_get_sk_secctx(struct sk_buff *skb, char **secdata)u32 seclen =0 ;#if IS_ENABLED (CONFIG_NETWORK_SECMARK )struct lsmblob blob; +struct lsmcontext context = { };if (!skb || !sk_fullsock (skb->sk))return 0 ;
@@ -317,10 +318,12 @@ static u32 nfqnl_get_sk_secctx(struct sk_buff *skb, char **secdata)* blob. security_secid_to_secctx() will know which security * module to use to create the secctx. */ lsmblob_init (&blob, skb->secmark); - security_secid_to_secctx(&blob, secdata, &seclen); +security_secid_to_secctx (&blob, &context); + *secdata = context.context; }read_unlock_bh (&skb->sk->sk_callback_lock); + seclen = context.len;#endif return seclen; }
net/netlabel/netlabel_unlabeled.c +12 -28
@@ -375,8 +375,6 @@ int netlbl_unlhsh_add(struct net *net,struct netlbl_unlhsh_iface *iface;struct audit_buffer *audit_buf =NULL ;struct lsmcontext context; -char *secctx =NULL ; -u32 secctx_len;struct lsmblob blob; if (addr_len != sizeof(struct in_addr) &&
@@ -444,12 +442,9 @@ int netlbl_unlhsh_add(struct net *net,* security_secid_to_secctx() will know which security module * to use to create the secctx. */ lsmblob_init (&blob, secid); -if (security_secid_to_secctx (&blob, - &secctx, - &secctx_len) ==0 ) { -audit_log_format (audit_buf," sec_obj=%s" , secctx); -/* scaffolding */ -lsmcontext_init (&context, secctx, secctx_len,0 ); +if (security_secid_to_secctx (&blob, &context) ==0 ) { +audit_log_format (audit_buf," sec_obj=%s" , + context.context);security_release_secctx (&context); }audit_log_format (audit_buf," res=%u" , ret_val ==0 ?1 :0 );
@@ -482,8 +477,6 @@ static int netlbl_unlhsh_remove_addr4(struct net *net,struct audit_buffer *audit_buf;struct net_device *dev;struct lsmcontext context; -char *secctx; -u32 secctx_len;struct lsmblob blob;spin_lock (&netlbl_unlhsh_lock);
@@ -510,11 +503,9 @@ static int netlbl_unlhsh_remove_addr4(struct net *net,if (entry !=NULL )lsmblob_init (&blob, entry->secid);if (entry !=NULL && -security_secid_to_secctx (&blob, - &secctx, &secctx_len) ==0 ) { -audit_log_format (audit_buf," sec_obj=%s" , secctx); -/* scaffolding */ -lsmcontext_init (&context, secctx, secctx_len,0 ); +security_secid_to_secctx (&blob, &context) ==0 ) { +audit_log_format (audit_buf," sec_obj=%s" , + context.context);security_release_secctx (&context); }audit_log_format (audit_buf," res=%u" , entry !=NULL ?1 :0 );
@@ -553,8 +544,6 @@ static int netlbl_unlhsh_remove_addr6(struct net *net,struct audit_buffer *audit_buf;struct net_device *dev;struct lsmcontext context; -char *secctx; -u32 secctx_len;struct lsmblob blob;spin_lock (&netlbl_unlhsh_lock);
@@ -580,10 +569,9 @@ static int netlbl_unlhsh_remove_addr6(struct net *net,if (entry !=NULL )lsmblob_init (&blob, entry->secid);if (entry !=NULL && -security_secid_to_secctx (&blob, - &secctx, &secctx_len) ==0 ) { -audit_log_format (audit_buf," sec_obj=%s" , secctx); -lsmcontext_init (&context, secctx, secctx_len,0 ); +security_secid_to_secctx (&blob, &context) ==0 ) { +audit_log_format (audit_buf," sec_obj=%s" , + context.context);security_release_secctx (&context); }audit_log_format (audit_buf," res=%u" , entry !=NULL ?1 :0 );
@@ -1106,8 +1094,6 @@ static int netlbl_unlabel_staticlist_gen(u32 cmd,struct lsmcontext context;void *data;u32 secid; -char *secctx; -u32 secctx_len;struct lsmblob blob; data = genlmsg_put(cb_arg->skb, NETLINK_CB(cb_arg->nl_cb->skb).portid,
@@ -1167,15 +1153,13 @@ static int netlbl_unlabel_staticlist_gen(u32 cmd,* security_secid_to_secctx() will know which security module * to use to create the secctx. */ lsmblob_init (&blob, secid); - ret_val =security_secid_to_secctx (&blob, &secctx, &secctx_len); + ret_val =security_secid_to_secctx (&blob, &context);if (ret_val !=0 ) goto list_cb_failure; ret_val =nla_put (cb_arg->skb,NLBL_UNLABEL_A_SECCTX , - secctx_len, - secctx); -/* scaffolding */ -lsmcontext_init (&context, secctx, secctx_len,0 ); + context.len, + context.context);security_release_secctx (&context);if (ret_val !=0 ) goto list_cb_failure;
net/netlabel/netlabel_user.c +2 -5
@@ -85,8 +85,6 @@ struct audit_buffer *netlbl_audit_start_common(int type, {struct audit_buffer *audit_buf;struct lsmcontext context; -char *secctx; -u32 secctx_len;struct lsmblob blob;if (audit_enabled ==AUDIT_OFF )
@@ -102,9 +100,8 @@ struct audit_buffer *netlbl_audit_start_common(int type,lsmblob_init (&blob, audit_info->secid);if (audit_info->secid !=0 && -security_secid_to_secctx (&blob, &secctx, &secctx_len) ==0 ) { -audit_log_format (audit_buf," subj=%s" , secctx); -lsmcontext_init (&context, secctx, secctx_len,0 );/*scaffolding*/ +security_secid_to_secctx (&blob, &context) ==0 ) { +audit_log_format (audit_buf," subj=%s" , context.context);security_release_secctx (&context); }
security/security.c +7 -3
@@ -2323,18 +2323,22 @@ int security_ismaclabel(const char *name) }EXPORT_SYMBOL (security_ismaclabel); -int security_secid_to_secctx (struct lsmblob *blob,char **secdata,u32 *seclen) +int security_secid_to_secctx (struct lsmblob *blob,struct lsmcontext *cp) {struct security_hook_list *hp;int ilsm =lsm_task_ilsm (current); +memset (cp,0 ,sizeof (*cp)); +hlist_for_each_entry (hp, &security_hook_heads.secid_to_secctx, list) {if (WARN_ON (hp->lsmid->slot <0 || hp->lsmid->slot >= lsm_slot))continue ; -if (ilsm ==LSMBLOB_INVALID || ilsm == hp->lsmid->slot) +if (ilsm ==LSMBLOB_INVALID || ilsm == hp->lsmid->slot) { + cp->slot = hp->lsmid->slot;return hp->hook.secid_to_secctx ( blob->secid[hp->lsmid->slot], - secdata, seclen); + &cp->context, &cp->len); + } }return LSM_RET_DEFAULT (secid_to_secctx);
Signature · 1 line
2.29.2
Casey Schaufler
maintainerPATCH v26 17/25 LSM: Use lsmcontext in security_inode_getsecctx
In reply to Casey SchauflerChange the security_inode_getsecctx() interface to fill a lsmcontext structure instead of data and length pointers. This provides the information about which LSM created the context so that security_release_secctx() can use the correct hook.
--- fs/nfsd/nfs4xdr.c | 23 +++++++++-------------- include/linux/security.h | 5 +++-- security/security.c | 13 +++++++++++-- 3 files changed, 23 insertions(+), 18 deletions(-)
fs/nfsd/nfs4xdr.c +9 -14
@@ -2727,11 +2727,11 @@ nfsd4_encode_layout_types(struct xdr_stream *xdr, u32 layout_types)#ifdef CONFIG_NFSD_V4_SECURITY_LABEL static inline __be32 nfsd4_encode_security_label (struct xdr_stream *xdr,struct svc_rqst *rqstp, -void *context,int len) +struct lsmcontext *context) {__be32 *p; - p =xdr_reserve_space (xdr, len +4 +4 +4 ); + p =xdr_reserve_space (xdr, context->len +4 +4 +4 );if (!p)return nfserr_resource;
@@ -2741,13 +2741,13 @@ nfsd4_encode_security_label(struct xdr_stream *xdr, struct svc_rqst *rqstp,*/ *p++ =cpu_to_be32 (0 );/* lfs */ *p++ =cpu_to_be32 (0 );/* pi */ - p =xdr_encode_opaque (p, context, len); + p =xdr_encode_opaque (p, context->context, context->len); return 0; }#else static inline __be32 nfsd4_encode_security_label (struct xdr_stream *xdr,struct svc_rqst *rqstp, -void *context,int len) +struct lsmcontext *context) {return 0 ; }#endif
@@ -2844,9 +2844,7 @@ nfsd4_encode_fattr(struct xdr_stream *xdr, struct svc_fh *fhp,int err;struct nfs4_acl *acl =NULL ;#ifdef CONFIG_NFSD_V4_SECURITY_LABEL -struct lsmcontext scaff;/* scaffolding */ -void *context =NULL ; -int contextlen; +struct lsmcontext context = { };#endif bool contextsupport = false;struct nfsd4_compoundres *resp = rqstp->rq_resp;
@@ -2904,7 +2902,7 @@ nfsd4_encode_fattr(struct xdr_stream *xdr, struct svc_fh *fhp, bmval0 & FATTR4_WORD0_SUPPORTED_ATTRS) {if (exp->ex_flags &NFSEXP_SECURITY_LABEL ) err =security_inode_getsecctx (d_inode (dentry), - &context, &contextlen); + &context);else err = -EOPNOTSUPP ; contextsupport = (err ==0 );
@@ -3324,8 +3322,7 @@ nfsd4_encode_fattr(struct xdr_stream *xdr, struct svc_fh *fhp,#ifdef CONFIG_NFSD_V4_SECURITY_LABEL if (bmval2 &FATTR4_WORD2_SECURITY_LABEL ) { - status =nfsd4_encode_security_label (xdr, rqstp, context, - contextlen); + status =nfsd4_encode_security_label (xdr, rqstp, &context);if (status) goto out; }
@@ -3346,10 +3343,8 @@ nfsd4_encode_fattr(struct xdr_stream *xdr, struct svc_fh *fhp, out:#ifdef CONFIG_NFSD_V4_SECURITY_LABEL -if (context) { -lsmcontext_init (&scaff, context, contextlen,0 );/*scaffolding*/ -security_release_secctx (&scaff); - } +if (context.context) +security_release_secctx (&context);#endif /* CONFIG_NFSD_V4_SECURITY_LABEL */ kfree (acl);if (tempfh) {
include/linux/security.h +3 -2
@@ -585,7 +585,7 @@ void security_release_secctx(struct lsmcontext *cp);void security_inode_invalidate_secctx (struct inode *inode);int security_inode_notifysecctx (struct inode *inode,void *ctx,u32 ctxlen);int security_inode_setsecctx (struct dentry *dentry,void *ctx,u32 ctxlen); -int security_inode_getsecctx (struct inode *inode,void **ctx,u32 *ctxlen); +int security_inode_getsecctx (struct inode *inode,struct lsmcontext *cp);int security_locked_down (enum lockdown_reason what);#else /* CONFIG_SECURITY */
@@ -1461,7 +1461,8 @@ static inline int security_inode_setsecctx(struct dentry *dentry, void *ctx, u32 {return -EOPNOTSUPP ; } -static inline int security_inode_getsecctx (struct inode *inode,void **ctx,u32 *ctxlen) +static inline int security_inode_getsecctx (struct inode *inode, +struct lsmcontext *cp) {return -EOPNOTSUPP ; }
security/security.c +11 -2
@@ -2395,9 +2395,18 @@ int security_inode_setsecctx(struct dentry *dentry, void *ctx, u32 ctxlen) }EXPORT_SYMBOL (security_inode_setsecctx); -int security_inode_getsecctx (struct inode *inode,void **ctx,u32 *ctxlen) +int security_inode_getsecctx (struct inode *inode,struct lsmcontext *cp) { -return call_int_hook (inode_getsecctx, -EOPNOTSUPP , inode, ctx, ctxlen); +struct security_hook_list *hp; + +memset (cp,0 ,sizeof (*cp)); + +hlist_for_each_entry (hp, &security_hook_heads.inode_getsecctx, list) { + cp->slot = hp->lsmid->slot; +return hp->hook.inode_getsecctx (inode, (void **)&cp->context, + &cp->len); + } +return -EOPNOTSUPP ; }EXPORT_SYMBOL (security_inode_getsecctx);
Signature · 1 line
2.29.2
Re: [PATCH v26 17/25] LSM: Use lsmcontext in security_inode_getsecctx
Quoted 11 lines · Casey Schaufler
Change the
security_inode_getsecctx()interface to fill a lsmcontext structure instead of data and length pointers. This provides the information about which LSM created the context so thatsecurity_release_secctx()can use the correct hook.Acked-by Stephen Smalley <[email protected]>Acked-by Paul Moore <[email protected]>Acked-by Chuck Lever <[email protected]>Reviewed-by John Johansen <[email protected]>Signed-off-by Casey Schaufler <[email protected]>
Seem good to me.
Signature · 1 line
Kees Cook
Casey Schaufler
maintainerPATCH v26 18/25 LSM: security_secid_to_secctx in netlink netfilter
In reply to Casey SchauflerChange netlink netfilter interfaces to use lsmcontext pointers, and remove scaffolding.
--- net/netfilter/nfnetlink_queue.c | 37 +++++++++++++-------------------- 1 file changed, 14 insertions(+), 23 deletions(-)
net/netfilter/nfnetlink_queue.c +14 -23
@@ -301,15 +301,13 @@ static int nfqnl_put_sk_uidgid(struct sk_buff *skb, struct sock *sk)return -1 ; } -static u32 nfqnl_get_sk_secctx (struct sk_buff *skb,char **secdata) +static void nfqnl_get_sk_secctx (struct sk_buff *skb,struct lsmcontext *context) { -u32 seclen =0 ;#if IS_ENABLED (CONFIG_NETWORK_SECMARK )struct lsmblob blob; -struct lsmcontext context = { };if (!skb || !sk_fullsock (skb->sk)) -return 0 ; +return ;read_lock_bh (&skb->sk->sk_callback_lock);
@@ -318,14 +316,12 @@ static u32 nfqnl_get_sk_secctx(struct sk_buff *skb, char **secdata)* blob. security_secid_to_secctx() will know which security * module to use to create the secctx. */ lsmblob_init (&blob, skb->secmark); -security_secid_to_secctx (&blob, &context); - *secdata = context.context; + security_secid_to_secctx(&blob, context); }read_unlock_bh (&skb->sk->sk_callback_lock); - seclen = context.len;#endif -return seclen; +return ; }static u32 nfqnl_get_bridge_size (struct nf_queue_entry *entry)
@@ -397,12 +393,10 @@ nfqnl_build_packet_message(struct net *net, struct nfqnl_instance *queue,struct net_device *indev;struct net_device *outdev;struct nf_conn *ct =NULL ; +struct lsmcontext context = { };enum ip_conntrack_info ctinfo;struct nfnl_ct_hook *nfnl_ct;bool csum_verify; -struct lsmcontext scaff;/* scaffolding */ -char *secdata =NULL ; -u32 seclen =0 ; size =nlmsg_total_size (sizeof (struct nfgenmsg )) +nla_total_size (sizeof (struct nfqnl_msg_packet_hdr ))
@@ -470,9 +464,9 @@ nfqnl_build_packet_message(struct net *net, struct nfqnl_instance *queue, }if ((queue->flags &NFQA_CFG_F_SECCTX ) && entskb->sk) { - seclen =nfqnl_get_sk_secctx (entskb, &secdata); -if (seclen) - size +=nla_total_size (seclen); +nfqnl_get_sk_secctx (entskb, &context); +if (context.len) + size +=nla_total_size (context.len); } skb =alloc_skb (size,GFP_ATOMIC );
@@ -602,7 +596,8 @@ nfqnl_build_packet_message(struct net *net, struct nfqnl_instance *queue, nfqnl_put_sk_uidgid(skb, entskb->sk) < 0) goto nla_put_failure; -if (seclen &&nla_put (skb,NFQA_SECCTX , seclen, secdata)) +if (context.len && +nla_put (skb,NFQA_SECCTX , context.len, context.context)) goto nla_put_failure;if (ct && nfnl_ct->build (skb, ct, ctinfo,NFQA_CT ,NFQA_CT_INFO ) <0 )
@@ -630,10 +625,8 @@ nfqnl_build_packet_message(struct net *net, struct nfqnl_instance *queue, } nlh->nlmsg_len = skb->len; -if (seclen) { -lsmcontext_init (&scaff, secdata, seclen,0 ); -security_release_secctx (&scaff); - } +if (context.len) +security_release_secctx (&context);return skb; nla_put_failure:
@@ -641,10 +634,8 @@ nfqnl_build_packet_message(struct net *net, struct nfqnl_instance *queue,kfree_skb (skb);net_err_ratelimited ("nf_queue: error creating packet message\n" ); nlmsg_failure: -if (seclen) { -lsmcontext_init (&scaff, secdata, seclen,0 ); -security_release_secctx (&scaff); - } +if (context.len) +security_release_secctx (&context);return NULL ; }
Signature · 1 line
2.29.2
Re: [PATCH v26 18/25] LSM: security_secid_to_secctx in netlink netfilter
Quoted 11 lines · Casey Schaufler
Change netlink netfilter interfaces to use lsmcontext pointers, and remove scaffolding.
Reviewed-by Kees Cook <[email protected]>Reviewed-by John Johansen <[email protected]>Acked-by Stephen Smalley <[email protected]>Signed-off-by Casey Schaufler <[email protected]>
Quoted 3 lines
--- net/netfilter/nfnetlink_queue.c | 37 +++++++++++++-------------------- 1 file changed, 14 insertions(+), 23 deletions(-)
--
paul moore
www.paul-moore.com
--
Linux-audit mailing list
[email protected]
https://listman.redhat.com/mailman/listinfo/linux-audit
Casey Schaufler
maintainerPATCH v26 19/25 NET: Store LSM netlabel data in a lsmblob
In reply to Casey SchauflerNetlabel uses LSM interfaces requiring an lsmblob and the internal storage is used to pass information between these interfaces, so change the internal data from a secid to a lsmblob. Update the netlabel interfaces and their callers to accommodate the change. This requires that the modules using netlabel use the lsm_id.slot to access the correct secid when using netlabel.
--- include/net/netlabel.h | 8 +-- net/ipv4/cipso_ipv4.c | 26 ++++++---- net/netlabel/netlabel_kapi.c | 6 +-- net/netlabel/netlabel_unlabeled.c | 79 +++++++++-------------------- net/netlabel/netlabel_unlabeled.h | 2 +- security/selinux/hooks.c | 2 +- security/selinux/include/security.h | 1 + security/selinux/netlabel.c | 2 +- security/selinux/ss/services.c | 4 +- security/smack/smack.h | 1 + security/smack/smack_access.c | 2 +- security/smack/smack_lsm.c | 11 ++-- security/smack/smackfs.c | 10 ++-- 13 files changed, 68 insertions(+), 86 deletions(-)
include/net/netlabel.h +4 -4
@@ -166,7 +166,7 @@ struct netlbl_lsm_catmap {* @attr.mls: MLS sensitivity label * @attr.mls.cat: MLS category bitmap * @attr.mls.lvl: MLS sensitivity level -* @attr.secid: LSM specific secid token +* @attr.lsmblob: LSM specific data * * Description: * This structure is used to pass security attributes between NetLabel and the
@@ -201,7 +201,7 @@ struct netlbl_lsm_secattr {struct netlbl_lsm_catmap *cat;u32 lvl; } mls; -u32 secid; +struct lsmblob lsmblob; } attr; };
@@ -415,7 +415,7 @@ int netlbl_cfg_unlbl_static_add(struct net *net,const void *addr, const void *mask, u16 family, - u32 secid, + struct lsmblob *lsmblob, struct netlbl_audit *audit_info); int netlbl_cfg_unlbl_static_del(struct net *net, const char *dev_name,
@@ -523,7 +523,7 @@ static inline int netlbl_cfg_unlbl_static_add(struct net *net,const void *addr, const void *mask, u16 family, - u32 secid, + struct lsmblob *lsmblob, struct netlbl_audit *audit_info) { return -ENOSYS;
net/ipv4/cipso_ipv4.c +16 -10
@@ -106,15 +106,17 @@ int cipso_v4_rbm_strictvalid = 1;/* Base length of the local tag (non-standard tag). * Tag definition (may change between kernel versions) * -* 0 8 16 24 32 -* +----------+----------+----------+----------+ -* | 10000000 | 00000110 | 32-bit secid value | -* +----------+----------+----------+----------+ -* | in (host byte order)| -* +----------+----------+ -* +* 0 8 16 16 + sizeof(struct lsmblob) +* +----------+----------+---------------------+ +* | 10000000 | 00000110 | LSM blob data | +* +----------+----------+---------------------+ +* +* All secid and flag fields are in host byte order. +* The lsmblob structure size varies depending on which +* Linux security modules are built in the kernel. +* The data is opaque. */ -#define CIPSO_V4_TAG_LOC_BLEN 6 +#define CIPSO_V4_TAG_LOC_BLEN (2 + sizeof(struct lsmblob))/* * Helper Functions
@@ -1460,7 +1462,11 @@ static int cipso_v4_gentag_loc(const struct cipso_v4_doi *doi_def, buffer[0 ] =CIPSO_V4_TAG_LOCAL ; buffer[1 ] =CIPSO_V4_TAG_LOC_BLEN ; - *(u32 *)&buffer[2 ] = secattr->attr.secid; +/* Ensure that there is sufficient space in the CIPSO header +* for the LSM data. */ +BUILD_BUG_ON (CIPSO_V4_TAG_LOC_BLEN >CIPSO_V4_OPT_LEN_MAX ); +memcpy (&buffer[2 ], &secattr->attr.lsmblob, +sizeof (secattr->attr.lsmblob));return CIPSO_V4_TAG_LOC_BLEN ; }
@@ -1480,7 +1486,7 @@ static int cipso_v4_parsetag_loc(const struct cipso_v4_doi *doi_def,const unsigned char *tag, struct netlbl_lsm_secattr *secattr) { - secattr->attr.secid = *(u32 *)&tag[2]; + memcpy(&secattr->attr.lsmblob, &tag[2], sizeof(secattr->attr.lsmblob)); secattr->flags |=NETLBL_SECATTR_SECID ;return 0 ;
net/netlabel/netlabel_kapi.c +3 -3
@@ -196,7 +196,7 @@ int netlbl_cfg_unlbl_map_add(const char *domain,* @addr: IP address in network byte order (struct in[6]_addr) * @mask: address mask in network byte order (struct in[6]_addr) * @family: address family -* @secid: LSM secid value for the entry +* @lsmblob: LSM data value for the entry * @audit_info: NetLabel audit information * * Description:
@@ -210,7 +210,7 @@ int netlbl_cfg_unlbl_static_add(struct net *net,const void *addr, const void *mask, u16 family, - u32 secid, + struct lsmblob *lsmblob, struct netlbl_audit *audit_info) { u32 addr_len;
@@ -230,7 +230,7 @@ int netlbl_cfg_unlbl_static_add(struct net *net,return netlbl_unlhsh_add (net, dev_name, addr, mask, addr_len, - secid, audit_info); + lsmblob, audit_info); }/**
net/netlabel/netlabel_unlabeled.c +24 -55
227 lines not on this page · raw message
net/netlabel/netlabel_unlabeled.h +1 -1
@@ -211,7 +211,7 @@ int netlbl_unlhsh_add(struct net *net,const void *addr, const void *mask, u32 addr_len, - u32 secid, + struct lsmblob *lsmblob, struct netlbl_audit *audit_info); int netlbl_unlhsh_remove(struct net *net, const char *dev_name,
security/selinux/hooks.c +1 -1
@@ -7122,7 +7122,7 @@ static int selinux_perf_event_write(struct perf_event *event) }#endif -static struct lsm_id selinux_lsmid __lsm_ro_after_init = { +struct lsm_id selinux_lsmid __lsm_ro_after_init = { .lsm ="selinux" , .slot =LSMBLOB_NEEDED };
security/selinux/include/security.h +1 -0
@@ -73,6 +73,7 @@struct netlbl_lsm_secattr ;extern int selinux_enabled_boot; +extern struct lsm_id selinux_lsmid;/* * type_datum properties
security/selinux/netlabel.c +1 -1
@@ -108,7 +108,7 @@ static struct netlbl_lsm_secattr *selinux_netlbl_sock_getattr(return NULL ;if ((secattr->flags &NETLBL_SECATTR_SECID ) && - (secattr->attr.secid == sid)) + (secattr->attr.lsmblob.secid[selinux_lsmid.slot] == sid))return secattr;return NULL ;
security/selinux/ss/services.c +2 -2
@@ -3883,7 +3883,7 @@ int security_netlbl_secattr_to_sid(struct selinux_state *state,if (secattr->flags &NETLBL_SECATTR_CACHE ) *sid = *(u32 *)secattr->cache->data;else if (secattr->flags &NETLBL_SECATTR_SECID ) - *sid = secattr->attr.secid; + *sid = secattr->attr.lsmblob.secid[selinux_lsmid.slot];else if (secattr->flags &NETLBL_SECATTR_MLS_LVL ) { rc = -EIDRM ; ctx =sidtab_search (sidtab,SECINITSID_NETMSG );
@@ -3960,7 +3960,7 @@ int security_netlbl_sid_to_secattr(struct selinux_state *state,if (secattr->domain ==NULL ) goto out; - secattr->attr.secid = sid; + secattr->attr.lsmblob.secid[selinux_lsmid.slot] = sid; secattr->flags |=NETLBL_SECATTR_DOMAIN_CPY |NETLBL_SECATTR_SECID ;mls_export_netlbl_lvl (policydb, ctx, secattr); rc =mls_export_netlbl_cat (policydb, ctx, secattr);
security/smack/smack.h +1 -0
@@ -303,6 +303,7 @@ int smack_populate_secattr(struct smack_known *skp);* Shared data. */ extern int smack_enabled; +extern struct lsm_id smack_lsmid;extern int smack_cipso_direct;extern int smack_cipso_mapped;extern struct smack_known *smack_net_ambient;
security/smack/smack_access.c +1 -1
@@ -522,7 +522,7 @@ int smack_populate_secattr(struct smack_known *skp) {int slen; - skp->smk_netlabel.attr.secid = skp->smk_secid; + skp->smk_netlabel.attr.lsmblob.secid[smack_lsmid.slot] = skp->smk_secid; skp->smk_netlabel.domain = skp->smk_known; skp->smk_netlabel.cache =netlbl_secattr_cache_alloc (GFP_ATOMIC );if (skp->smk_netlabel.cache !=NULL ) {
security/smack/smack_lsm.c +6 -5
@@ -3720,11 +3720,12 @@ static struct smack_known *smack_from_secattr(struct netlbl_lsm_secattr *sap,if ((sap->flags &NETLBL_SECATTR_CACHE ) !=0 )return (struct smack_known *)sap->cache->data; +/* +* Looks like a fallback, which gives us a secid. +*/ if ((sap->flags &NETLBL_SECATTR_SECID ) !=0 ) -/* -* Looks like a fallback, which gives us a secid. -*/ -return smack_from_secid (sap->attr.secid); +return smack_from_secid ( + sap->attr.lsmblob.secid[smack_lsmid.slot]);if ((sap->flags &NETLBL_SECATTR_MLS_LVL ) !=0 ) {/*
@@ -4701,7 +4702,7 @@ struct lsm_blob_sizes smack_blob_sizes __lsm_ro_after_init = { .lbs_superblock = sizeof(struct superblock_smack), }; -static struct lsm_id smack_lsmid __lsm_ro_after_init = { +struct lsm_id smack_lsmid __lsm_ro_after_init = { .lsm ="smack" , .slot =LSMBLOB_NEEDED };
security/smack/smackfs.c +7 -3
@@ -1140,6 +1140,7 @@ static void smk_net4addr_insert(struct smk_net4addr *new)static ssize_t smk_write_net4addr (struct file *file,const char __user *buf,size_t count,loff_t *ppos) { +struct lsmblob lsmblob;struct smk_net4addr *snp;struct sockaddr_in newname;char *smack;
@@ -1271,10 +1272,13 @@ static ssize_t smk_write_net4addr(struct file *file, const char __user *buf,* this host so that incoming packets get labeled. * but only if we didn't get the special CIPSO option */ -if (rc ==0 && skp !=NULL ) +if (rc ==0 && skp !=NULL ) { +lsmblob_init (&lsmblob,0 ); + lsmblob.secid[smack_lsmid.slot] = snp->smk_label->smk_secid; rc =netlbl_cfg_unlbl_static_add (&init_net,NULL , - &snp->smk_host, &snp->smk_mask,PF_INET , - snp->smk_label->smk_secid, &audit_info); + &snp->smk_host, &snp->smk_mask,PF_INET , &lsmblob, + &audit_info); + }if (rc ==0 ) rc = count;
Signature · 1 line
2.29.2
Casey Schaufler
maintainerPATCH v26 20/25 LSM: Verify LSM display sanity in binder
In reply to Casey SchauflerVerify that the tasks on the ends of a binder transaction use the same "interface_lsm" security module. This prevents confusion of security "contexts".
--- security/security.c | 29 +++++++++++++++++++++++++++++ 1 file changed, 29 insertions(+)
security/security.c +29 -0
@@ -857,9 +857,38 @@ int security_binder_set_context_mgr(struct task_struct *mgr)return call_int_hook (binder_set_context_mgr,0 , mgr); } +/** +* security_binder_transaction - Binder driver transaction check +* @from: source of the transaction +* @to: destination of the transaction +* +* Verify that the tasks have the same LSM "display", then +* call the security module hooks. +* +* Returns -EINVAL if the displays don't match, or the +* result of the security module checks. +*/ int security_binder_transaction (struct task_struct *from,struct task_struct *to) { +int from_ilsm =lsm_task_ilsm (from); +int to_ilsm =lsm_task_ilsm (to); + +/* +* If the ilsm is LSMBLOB_INVALID the first module that has +* an entry is used. This will be in the 0 slot. +* +* This is currently only required if the server has requested +* peer contexts, but it would be unwieldly to have too much of +* the binder driver detail here. +*/ +if (from_ilsm ==LSMBLOB_INVALID ) + from_ilsm =0 ; +if (to_ilsm ==LSMBLOB_INVALID ) + to_ilsm =0 ; +if (from_ilsm != to_ilsm) +return -EINVAL ; +return call_int_hook (binder_transaction,0 , from, to); }
Signature · 1 line
2.29.2
Casey Schaufler
maintainerPATCH v26 21/25 audit: add support for non-syscall auxiliary records
In reply to Casey SchauflerStandalone audit records have the timestamp and serial number generated on the fly and as such are unique, making them standalone. This new function audit_alloc_local() generates a local audit context that will be used only for a standalone record and its auxiliary record(s). The context is discarded immediately after the local associated records are produced.
To: Richard Guy Briggs <[email protected]>
--- include/linux/audit.h | 8 ++++++++ kernel/audit.h | 1 + kernel/auditsc.c | 33 ++++++++++++++++++++++++++++----- 3 files changed, 37 insertions(+), 5 deletions(-)
include/linux/audit.h +8 -0
@@ -289,6 +289,8 @@ static inline int audit_signal_info(int sig, struct task_struct *t)/* Public API */ extern int audit_alloc (struct task_struct *task);extern void __audit_free (struct task_struct *task); +extern struct audit_context *audit_alloc_local (gfp_t gfpflags); +extern void audit_free_context (struct audit_context *context);extern void __audit_syscall_entry (int major,unsigned long a0,unsigned long a1,unsigned long a2,unsigned long a3);extern void __audit_syscall_exit (int ret_success,long ret_value);
@@ -552,6 +554,12 @@ static inline void audit_log_nfcfg(const char *name, u8 af,extern int audit_n_rules;extern int audit_signals;#else /* CONFIG_AUDITSYSCALL */ ++static inline struct audit_context *audit_alloc_local(gfp_t gfpflags) +{ +return NULL ; +} +static inline void audit_free_context (struct audit_context *context) +{ }static inline int audit_alloc (struct task_struct *task) {return 0 ;
kernel/audit.h +1 -0
@@ -99,6 +99,7 @@ struct audit_proctitle {struct audit_context {int dummy;/* must be the first element */ int in_syscall;/* 1 if task is in a syscall */ +bool local;/* local context needed */ enum audit_state state, current_state;unsigned int serial;/* serial number for record */ int major;/* syscall number */
kernel/auditsc.c +28 -5
@@ -925,11 +925,13 @@ static inline void audit_free_aux(struct audit_context *context) } } -static inline struct audit_context *audit_alloc_context (enum audit_state state) +static inline struct audit_context *audit_alloc_context (enum audit_state state, +gfp_t gfpflags) {struct audit_context *context; - context =kzalloc (sizeof (*context),GFP_KERNEL ); +/* We can be called in atomic context via audit_tg() */ + context =kzalloc (sizeof (*context), gfpflags);if (!context)return NULL ; context->state = state;
@@ -965,7 +967,8 @@ int audit_alloc(struct task_struct *tsk)return 0 ; } -if (!(context =audit_alloc_context (state))) { + context =audit_alloc_context (state,GFP_KERNEL ); +if (!context) {kfree (key);audit_log_lost ("out of memory in audit_alloc" );return -ENOMEM ;
@@ -977,8 +980,27 @@ int audit_alloc(struct task_struct *tsk)return 0 ; } -static inline void audit_free_context (struct audit_context *context) +struct audit_context *audit_alloc_local (gfp_t gfpflags) { +struct audit_context *context =NULL ; + + context =audit_alloc_context (AUDIT_RECORD_CONTEXT , gfpflags); +if (!context) { +audit_log_lost ("out of memory in audit_alloc_local" ); + goto out; + } + context->serial =audit_serial (); +ktime_get_coarse_real_ts64 (&context->ctime); + context->local = true; +out: +return context; +} +EXPORT_SYMBOL (audit_alloc_local); + +void audit_free_context (struct audit_context *context) +{ +if (!context) +return ;audit_free_module (context);audit_free_names (context);unroll_tree_refs (context,NULL ,0 );
@@ -989,6 +1011,7 @@ static inline void audit_free_context(struct audit_context *context)audit_proctitle_free (context);kfree (context); } +EXPORT_SYMBOL (audit_free_context); static int audit_log_pid_context(struct audit_context *context, pid_t pid, kuid_t auid, kuid_t uid,
@@ -2210,7 +2233,7 @@ EXPORT_SYMBOL_GPL(__audit_inode_child);int auditsc_get_stamp (struct audit_context *ctx,struct timespec64 *t,unsigned int *serial) { -if (!ctx->in_syscall) +if (!ctx->in_syscall && !ctx->local)return 0 ;if (!ctx->serial) ctx->serial =audit_serial ();
Signature · 1 line
2.29.2
Re: [PATCH v26 21/25] audit: add support for non-syscall auxiliary records
Quoted 11 lines · Casey Schaufler
Standalone audit records have the timestamp and serial number generated on the fly and as such are unique, making them standalone. This new function
audit_alloc_local()generates a local audit context that will be used only for a standalone record and its auxiliary record(s). The context is discarded immediately after the local associated records are produced.Signed-off-by Richard Guy Briggs <[email protected]>Signed-off-by Casey Schaufler <[email protected]>
Quoted 6 lines
To: Richard Guy Briggs <[email protected]>
--- include/linux/audit.h | 8 ++++++++ kernel/audit.h | 1 + kernel/auditsc.c | 33 ++++++++++++++++++++++++++++----- 3 files changed, 37 insertions(+), 5 deletions(-)
At some point I suspect we will need to add filtering for these "local" records, but that is a problem for another day.
--
paul moore
www.paul-moore.com
--
Linux-audit mailing list
[email protected]
https://listman.redhat.com/mailman/listinfo/linux-audit
Casey Schaufler
maintainerPATCH v26 22/25 Audit: Add new record for multiple process LSM attributes
In reply to Casey SchauflerCreate a new audit record type to contain the subject information when there are multiple security modules that require such data. This record is linked with the same timestamp and serial number using the audit_alloc_local() mechanism.
The record is produced only in cases where there is more than one security module with a process "context".
In cases where this record is produced the subj= fields of other records in the audit event will be set to "subj=?".
An example of the MAC_TASK_CONTEXTS (1420) record is:
type=UNKNOWN[1420]
msg=audit(1600880931.832:113)
subj_apparmor==unconfined
subj_smack=_There will be a subj_$LSM= entry for each security module LSM that supports the secid_to_secctx and secctx_to_secid hooks. The BPF security module implements secid/secctx translation hooks, so it has to be considered to provide a secctx even though it may not actually do so.
To: [email protected]
To: [email protected]
To: [email protected]
Cc: [email protected]
--- drivers/android/binder.c | 2 +- include/linux/audit.h | 24 ++++++++ include/linux/security.h | 16 ++++- include/net/netlabel.h | 3 +- include/net/scm.h | 2 +- include/net/xfrm.h | 13 +++- include/uapi/linux/audit.h | 1 + kernel/audit.c | 80 ++++++++++++++++++------- kernel/audit.h | 3 + kernel/auditfilter.c | 6 +- kernel/auditsc.c | 75 ++++++++++++++++++++--- net/ipv4/ip_sockglue.c | 2 +- net/netfilter/nf_conntrack_netlink.c | 4 +- net/netfilter/nf_conntrack_standalone.c | 2 +- net/netfilter/nfnetlink_queue.c | 2 +- net/netlabel/netlabel_domainhash.c | 4 +- net/netlabel/netlabel_unlabeled.c | 24 ++++---- net/netlabel/netlabel_user.c | 20 ++++--- net/netlabel/netlabel_user.h | 6 +- net/xfrm/xfrm_policy.c | 10 ++-- net/xfrm/xfrm_state.c | 20 ++++--- security/integrity/ima/ima_api.c | 7 ++- security/integrity/integrity_audit.c | 6 +- security/security.c | 46 +++++++++----- security/smack/smackfs.c | 3 +- 25 files changed, 274 insertions(+), 107 deletions(-)
drivers/android/binder.c +1 -1
@@ -2722,7 +2722,7 @@ static void binder_transaction(struct binder_proc *proc,* case well anyway. */ security_task_getsecid_obj (proc->tsk, &blob); - ret =security_secid_to_secctx (&blob, &lsmctx); + ret =security_secid_to_secctx (&blob, &lsmctx,LSMBLOB_DISPLAY );if (ret) { return_error =BR_FAILED_REPLY ; return_error_param = ret;
include/linux/audit.h +24 -0
@@ -164,6 +164,8 @@ extern struct audit_buffer *audit_log_start(struct audit_context *ctx, gfp_t gfp extern __printf(2, 3) void audit_log_format(struct audit_buffer *ab, const char *fmt, ...); extern void audit_log_end(struct audit_buffer *ab); +extern void audit_log_end_local(struct audit_buffer *ab, + struct audit_context *context); extern bool audit_string_contains_control(const char *string, size_t len); extern void audit_log_n_hex(struct audit_buffer *ab,
@@ -188,6 +190,7 @@ extern void audit_log_lost(const char *message);extern int audit_log_task_context (struct audit_buffer *ab);extern void audit_log_task_info (struct audit_buffer *ab); +extern void audit_log_lsm (struct audit_context *context);extern int audit_update_lsm_rules (void );
@@ -226,6 +229,9 @@ void audit_log_format(struct audit_buffer *ab, const char *fmt, ...) { }static inline void audit_log_end (struct audit_buffer *ab) { } +static inline void audit_log_end_local (struct audit_buffer *ab, +struct audit_context *context) +{ }static inline void audit_log_n_hex (struct audit_buffer *ab,const unsigned char *buf,size_t len) { }
@@ -252,6 +258,8 @@ static inline int audit_log_task_context(struct audit_buffer *ab) }static inline void audit_log_task_info (struct audit_buffer *ab) { } +static void audit_log_lsm (struct audit_context *context) +{ }static inline kuid_t audit_get_loginuid (struct task_struct *tsk) {
@@ -291,6 +299,7 @@ extern int audit_alloc(struct task_struct *task);extern void __audit_free (struct task_struct *task);extern struct audit_context *audit_alloc_local (gfp_t gfpflags);extern void audit_free_context (struct audit_context *context); +extern void audit_free_local (struct audit_context *context);extern void __audit_syscall_entry (int major,unsigned long a0,unsigned long a1,unsigned long a2,unsigned long a3);extern void __audit_syscall_exit (int ret_success,long ret_value);
@@ -386,6 +395,19 @@ static inline void audit_ptrace(struct task_struct *t)__audit_ptrace (t); } +static inline struct audit_context *audit_alloc_for_lsm (gfp_t gfp) +{ +struct audit_context *context =audit_context (); + +if (context) +return context; + +if (lsm_multiple_contexts ()) +return audit_alloc_local (gfp); + +return NULL ; +} +/* Private API (for audit.c only) */ extern void __audit_ipc_obj (struct kern_ipc_perm *ipcp);extern void __audit_ipc_set_perm (unsigned long qbytes,uid_t uid,gid_t gid,umode_t mode);
@@ -560,6 +582,8 @@ extern int audit_signals; }static inline void audit_free_context (struct audit_context *context) { } +static inline void audit_free_local (struct audit_context *context) +{ }static inline int audit_alloc (struct task_struct *task) {return 0 ;
include/linux/security.h +14 -2
@@ -182,6 +182,8 @@ struct lsmblob {#define LSMBLOB_INVALID -1/* Not a valid LSM slot number */ #define LSMBLOB_NEEDED -2/* Slot requested on initialization */ #define LSMBLOB_NOT_NEEDED -3/* Slot not requested */ +#define LSMBLOB_DISPLAY -4/* Use the "display" slot */ +#define LSMBLOB_FIRST -5/* Use the default "display" slot */ /** * lsmblob_init - initialize an lsmblob structure
@@ -248,6 +250,15 @@ static inline u32 lsmblob_value(const struct lsmblob *blob)return 0 ; } +static inline bool lsm_multiple_contexts (void ) +{ +#ifdef CONFIG_SECURITY +return lsm_slot_to_name (1 ) !=NULL ; +#else +return false; +#endif +} +/* These functions are in security/commoncap.c */ extern int cap_capable (const struct cred *cred,struct user_namespace *ns,int cap,unsigned int opts);
@@ -578,7 +589,8 @@ int security_setprocattr(const char *lsm, const char *name, void *value, size_t size);int security_netlink_send (struct sock *sk,struct sk_buff *skb);int security_ismaclabel (const char *name); -int security_secid_to_secctx (struct lsmblob *blob,struct lsmcontext *cp); +int security_secid_to_secctx (struct lsmblob *blob,struct lsmcontext *cp, +int display);int security_secctx_to_secid (const char *secdata,u32 seclen,struct lsmblob *blob);void security_release_secctx (struct lsmcontext *cp);
@@ -1433,7 +1445,7 @@ static inline int security_ismaclabel(const char *name) }static inline int security_secid_to_secctx (struct lsmblob *blob, -struct lsmcontext *cp) +struct lsmcontext *cp,int display) {return -EOPNOTSUPP ; }
include/net/netlabel.h +2 -1
@@ -97,7 +97,8 @@ struct calipso_doi;/* NetLabel audit information */ struct netlbl_audit { -u32 secid; +struct audit_context *localcontext; +struct lsmblob lsmdata;kuid_t loginuid;unsigned int sessionid; };
include/net/scm.h +1 -1
@@ -101,7 +101,7 @@ static inline void scm_passec(struct socket *sock, struct msghdr *msg, struct sc* and the infrastructure will know which it is. */ lsmblob_init (&lb, scm->secid); - err =security_secid_to_secctx (&lb, &context); + err =security_secid_to_secctx (&lb, &context,LSMBLOB_DISPLAY );if (!err) { put_cmsg(msg, SOL_SOCKET, SCM_SECURITY, context.len,
include/net/xfrm.h +11 -2
@@ -669,13 +669,22 @@ struct xfrm_spi_skb_cb {#define XFRM_SPI_SKB_CB (__skb) ((struct xfrm_spi_skb_cb *)&((__skb)->cb[0]))#ifdef CONFIG_AUDITSYSCALL -static inline struct audit_buffer *xfrm_audit_start (const char *op) +static inline struct audit_buffer *xfrm_audit_start (const char *op, +struct audit_context **lac) { +struct audit_context *context;struct audit_buffer *audit_buf =NULL ;if (audit_enabled ==AUDIT_OFF )return NULL ; - audit_buf =audit_log_start (audit_context (),GFP_ATOMIC , + context =audit_context (); +if (lac !=NULL ) { +if (lsm_multiple_contexts () && context ==NULL ) + context =audit_alloc_local (GFP_ATOMIC ); + *lac = context; + } + + audit_buf =audit_log_start (context,GFP_ATOMIC ,AUDIT_MAC_IPSEC_EVENT );if (audit_buf ==NULL )return NULL ;
include/uapi/linux/audit.h +1 -0
@@ -139,6 +139,7 @@#define AUDIT_MAC_UNLBL_STCDEL 1417/* NetLabel: del a static label */ #define AUDIT_MAC_CALIPSO_ADD 1418/* NetLabel: add CALIPSO DOI entry */ #define AUDIT_MAC_CALIPSO_DEL 1419/* NetLabel: del CALIPSO DOI entry */ +#define AUDIT_MAC_TASK_CONTEXTS 1420/* Multiple LSM contexts */ #define AUDIT_FIRST_KERN_ANOM_MSG 1700#define AUDIT_LAST_KERN_ANOM_MSG 1799
kernel/audit.c +57 -23
227 lines not on this page · raw message
kernel/audit.h +3 -0
@@ -100,6 +100,7 @@ struct audit_context {int dummy;/* must be the first element */ int in_syscall;/* 1 if task is in a syscall */ bool local;/* local context needed */ +bool lsmdone;/* multiple security reported */ enum audit_state state, current_state;unsigned int serial;/* serial number for record */ int major;/* syscall number */
@@ -131,6 +132,7 @@ struct audit_context {kgid_t gid, egid, sgid, fsgid;unsigned long personality;int arch; +struct lsmblob lsm;pid_t target_pid;kuid_t target_auid;
@@ -201,6 +203,7 @@ struct audit_context {extern bool audit_ever_enabled;extern void audit_log_session_info (struct audit_buffer *ab); +extern void audit_log_lsm_common (struct audit_context *context);extern int auditd_test_task (struct task_struct *task);
kernel/auditfilter.c +4 -2
@@ -1098,12 +1098,14 @@ static void audit_list_rules(int seq, struct sk_buff_head *q)/* Log rule additions and removals */ static void audit_log_rule_change (char *action,struct audit_krule *rule,int res) { +struct audit_context *context;struct audit_buffer *ab;if (!audit_enabled)return ; - ab =audit_log_start (audit_context (),GFP_KERNEL ,AUDIT_CONFIG_CHANGE ); + context =audit_alloc_for_lsm (GFP_KERNEL ); + ab =audit_log_start (context,GFP_KERNEL ,AUDIT_CONFIG_CHANGE );if (!ab)return ;audit_log_session_info (ab);
@@ -1111,7 +1113,7 @@ static void audit_log_rule_change(char *action, struct audit_krule *rule, int reaudit_log_format (ab," op=%s" , action);audit_log_key (ab, rule->filterkey);audit_log_format (ab," list=%d res=%d" , rule->listnr, res); -audit_log_end (ab); +audit_log_end_local (ab, context); }/**
kernel/auditsc.c +66 -9
164 lines not on this page · raw message
net/ipv4/ip_sockglue.c +1 -1
@@ -140,7 +140,7 @@ static void ip_cmsg_recv_security(struct msghdr *msg, struct sk_buff *skb)return ;lsmblob_init (&lb, secid); - err =security_secid_to_secctx (&lb, &context); + err =security_secid_to_secctx (&lb, &context,LSMBLOB_DISPLAY );if (err)return ;
net/netfilter/nf_conntrack_netlink.c +2 -2
@@ -344,7 +344,7 @@ static int ctnetlink_dump_secctx(struct sk_buff *skb, const struct nf_conn *ct)* security_secid_to_secctx() will know which security module * to use to create the secctx. */ lsmblob_init (&blob, ct->secmark); - ret =security_secid_to_secctx (&blob, &context); + ret =security_secid_to_secctx (&blob, &context,LSMBLOB_DISPLAY );if (ret)return 0 ;
@@ -655,7 +655,7 @@ static inline int ctnetlink_secctx_size(const struct nf_conn *ct)struct lsmblob blob;struct lsmcontext context; - ret =security_secid_to_secctx (&blob, &context); + ret =security_secid_to_secctx (&blob, &context,LSMBLOB_DISPLAY );if (ret)return 0 ;
net/netfilter/nf_conntrack_standalone.c +1 -1
@@ -177,7 +177,7 @@ static void ct_show_secctx(struct seq_file *s, const struct nf_conn *ct)struct lsmcontext context;lsmblob_init (&blob, ct->secmark); - ret =security_secid_to_secctx (&blob, &context); + ret =security_secid_to_secctx (&blob, &context,LSMBLOB_DISPLAY );if (ret)return ;
net/netfilter/nfnetlink_queue.c +1 -1
@@ -316,7 +316,7 @@ static void nfqnl_get_sk_secctx(struct sk_buff *skb, struct lsmcontext *context)* blob. security_secid_to_secctx() will know which security * module to use to create the secctx. */ lsmblob_init (&blob, skb->secmark); -security_secid_to_secctx (&blob, context); +security_secid_to_secctx (&blob, context,LSMBLOB_DISPLAY ); }read_unlock_bh (&skb->sk->sk_callback_lock);
net/netlabel/netlabel_domainhash.c +2 -2
@@ -259,7 +259,7 @@ static void netlbl_domhsh_audit_add(struct netlbl_dom_map *entry,break ; }audit_log_format (audit_buf," res=%u" , result ==0 ?1 :0 ); -audit_log_end (audit_buf); +audit_log_end_local (audit_buf, audit_info->localcontext); } }
@@ -614,7 +614,7 @@ int netlbl_domhsh_remove_entry(struct netlbl_dom_map *entry,audit_log_format (audit_buf," nlbl_domain=%s res=1" , entry->domain ? entry->domain :"(default)" ); -audit_log_end (audit_buf); +audit_log_end_local (audit_buf, audit_info->localcontext); }switch (entry->def.type) {
net/netlabel/netlabel_unlabeled.c +12 -12
@@ -437,13 +437,14 @@ int netlbl_unlhsh_add(struct net *net, unlhsh_add_return:rcu_read_unlock ();if (audit_buf !=NULL ) { -if (security_secid_to_secctx (lsmblob, &context) ==0 ) { +if (security_secid_to_secctx (lsmblob, &context, +LSMBLOB_FIRST ) ==0 ) {audit_log_format (audit_buf," sec_obj=%s" , context.context);security_release_secctx (&context); }audit_log_format (audit_buf," res=%u" , ret_val ==0 ?1 :0 ); -audit_log_end (audit_buf); +audit_log_end_local (audit_buf, audit_info->localcontext); }return ret_val; }
@@ -492,13 +493,14 @@ static int netlbl_unlhsh_remove_addr4(struct net *net,if (dev !=NULL )dev_put (dev);if (entry !=NULL && -security_secid_to_secctx (&entry->lsmblob, &context) ==0 ) { +security_secid_to_secctx (&entry->lsmblob, &context, +LSMBLOB_FIRST ) ==0 ) {audit_log_format (audit_buf," sec_obj=%s" , context.context);security_release_secctx (&context); }audit_log_format (audit_buf," res=%u" , entry !=NULL ?1 :0 ); -audit_log_end (audit_buf); +audit_log_end_local (audit_buf, audit_info->localcontext); }if (entry ==NULL )
@@ -552,13 +554,14 @@ static int netlbl_unlhsh_remove_addr6(struct net *net,if (dev !=NULL )dev_put (dev);if (entry !=NULL && -security_secid_to_secctx (&entry->lsmblob, &context) ==0 ) { +security_secid_to_secctx (&entry->lsmblob, &context, +LSMBLOB_FIRST ) ==0 ) {audit_log_format (audit_buf," sec_obj=%s" , context.context);security_release_secctx (&context); }audit_log_format (audit_buf," res=%u" , entry !=NULL ?1 :0 ); -audit_log_end (audit_buf); +audit_log_end_local (audit_buf, audit_info->localcontext); }if (entry ==NULL )
@@ -741,7 +744,7 @@ static void netlbl_unlabel_acceptflg_set(u8 value,if (audit_buf !=NULL ) {audit_log_format (audit_buf," unlbl_accept=%u old=%u" , value, old_val); -audit_log_end (audit_buf); +audit_log_end_local (audit_buf, audit_info->localcontext); } }
@@ -1122,7 +1125,7 @@ static int netlbl_unlabel_staticlist_gen(u32 cmd, lsmb = (struct lsmblob *)&addr6->lsmblob; } - ret_val =security_secid_to_secctx (lsmb, &context); + ret_val =security_secid_to_secctx (lsmb, &context,LSMBLOB_FIRST );if (ret_val !=0 ) goto list_cb_failure; ret_val = nla_put(cb_arg->skb,
@@ -1528,14 +1531,11 @@ int __init netlbl_unlabel_defconf(void)int ret_val;struct netlbl_dom_map *entry;struct netlbl_audit audit_info; -struct lsmblob blob;/* Only the kernel is allowed to call this function and the only time * it is called is at bootup before the audit subsystem is reporting * messages so don't worry to much about these values. */ -security_task_getsecid_subj (current, &blob); -/* scaffolding until audit_info.secid is converted */ - audit_info.secid = blob.secid[0 ]; +security_task_getsecid_subj (current, &audit_info.lsmdata); audit_info.loginuid =GLOBAL_ROOT_UID ; audit_info.sessionid =0 ;
net/netlabel/netlabel_user.c +13 -7
@@ -83,14 +83,17 @@ int __init netlbl_netlink_init(void)struct audit_buffer *netlbl_audit_start_common (int type,struct netlbl_audit *audit_info) { +struct audit_context *audit_ctx;struct audit_buffer *audit_buf;struct lsmcontext context; -struct lsmblob blob;if (audit_enabled ==AUDIT_OFF )return NULL ; - audit_buf =audit_log_start (audit_context (),GFP_ATOMIC , type); + audit_ctx =audit_alloc_for_lsm (GFP_ATOMIC ); + audit_info->localcontext = audit_ctx; + + audit_buf =audit_log_start (audit_ctx,GFP_ATOMIC , type);if (audit_buf ==NULL )return NULL ;
@@ -98,11 +101,14 @@ struct audit_buffer *netlbl_audit_start_common(int type,from_kuid (&init_user_ns, audit_info->loginuid), audit_info->sessionid); -lsmblob_init (&blob, audit_info->secid); -if (audit_info->secid !=0 && -security_secid_to_secctx (&blob, &context) ==0 ) { -audit_log_format (audit_buf," subj=%s" , context.context); -security_release_secctx (&context); +if (lsmblob_is_set (&audit_info->lsmdata)) { +if (!lsm_multiple_contexts () && +security_secid_to_secctx (&audit_info->lsmdata, &context, +LSMBLOB_FIRST ) ==0 ) { +audit_log_format (audit_buf," subj=%s" , + context.context); +security_release_secctx (&context); + } }return audit_buf;
net/netlabel/netlabel_user.h +1 -5
@@ -34,11 +34,7 @@static inline void netlbl_netlink_auditinfo (struct sk_buff *skb,struct netlbl_audit *audit_info) { -struct lsmblob blob; - -security_task_getsecid_subj (current, &blob); -/* scaffolding until secid is converted */ - audit_info->secid = blob.secid[0 ]; +security_task_getsecid_subj (current, &audit_info->lsmdata); audit_info->loginuid =audit_get_loginuid (current); audit_info->sessionid =audit_get_sessionid (current); }
net/xfrm/xfrm_policy.c +6 -4
@@ -4173,30 +4173,32 @@ static void xfrm_audit_common_policyinfo(struct xfrm_policy *xp,void xfrm_audit_policy_add (struct xfrm_policy *xp,int result,bool task_valid) { +struct audit_context *context;struct audit_buffer *audit_buf; - audit_buf =xfrm_audit_start ("SPD-add" ); + audit_buf =xfrm_audit_start ("SPD-add" , &context);if (audit_buf ==NULL )return ;xfrm_audit_helper_usrinfo (task_valid, audit_buf);audit_log_format (audit_buf," res=%u" , result);xfrm_audit_common_policyinfo (xp, audit_buf); -audit_log_end (audit_buf); +audit_log_end_local (audit_buf, context); }EXPORT_SYMBOL_GPL (xfrm_audit_policy_add);void xfrm_audit_policy_delete (struct xfrm_policy *xp,int result,bool task_valid) { +struct audit_context *context;struct audit_buffer *audit_buf; - audit_buf =xfrm_audit_start ("SPD-delete" ); + audit_buf =xfrm_audit_start ("SPD-delete" , &context);if (audit_buf ==NULL )return ;xfrm_audit_helper_usrinfo (task_valid, audit_buf);audit_log_format (audit_buf," res=%u" , result);xfrm_audit_common_policyinfo (xp, audit_buf); -audit_log_end (audit_buf); +audit_log_end_local (audit_buf, context); }EXPORT_SYMBOL_GPL (xfrm_audit_policy_delete);#endif
net/xfrm/xfrm_state.c +11 -9
@@ -2747,29 +2747,31 @@ static void xfrm_audit_helper_pktinfo(struct sk_buff *skb, u16 family,void xfrm_audit_state_add (struct xfrm_state *x,int result,bool task_valid) { +struct audit_context *context;struct audit_buffer *audit_buf; - audit_buf =xfrm_audit_start ("SAD-add" ); + audit_buf =xfrm_audit_start ("SAD-add" , &context);if (audit_buf ==NULL )return ;xfrm_audit_helper_usrinfo (task_valid, audit_buf);xfrm_audit_helper_sainfo (x, audit_buf);audit_log_format (audit_buf," res=%u" , result); -audit_log_end (audit_buf); +audit_log_end_local (audit_buf, context); }EXPORT_SYMBOL_GPL (xfrm_audit_state_add);void xfrm_audit_state_delete (struct xfrm_state *x,int result,bool task_valid) { +struct audit_context *context;struct audit_buffer *audit_buf; - audit_buf =xfrm_audit_start ("SAD-delete" ); + audit_buf =xfrm_audit_start ("SAD-delete" , &context);if (audit_buf ==NULL )return ;xfrm_audit_helper_usrinfo (task_valid, audit_buf);xfrm_audit_helper_sainfo (x, audit_buf);audit_log_format (audit_buf," res=%u" , result); -audit_log_end (audit_buf); +audit_log_end_local (audit_buf, context); }EXPORT_SYMBOL_GPL (xfrm_audit_state_delete);
@@ -2779,7 +2781,7 @@ void xfrm_audit_state_replay_overflow(struct xfrm_state *x,struct audit_buffer *audit_buf;u32 spi; - audit_buf =xfrm_audit_start ("SA-replay-overflow" ); + audit_buf =xfrm_audit_start ("SA-replay-overflow" ,NULL );if (audit_buf ==NULL )return ;xfrm_audit_helper_pktinfo (skb, x->props.family, audit_buf);
@@ -2797,7 +2799,7 @@ void xfrm_audit_state_replay(struct xfrm_state *x,struct audit_buffer *audit_buf;u32 spi; - audit_buf =xfrm_audit_start ("SA-replayed-pkt" ); + audit_buf =xfrm_audit_start ("SA-replayed-pkt" ,NULL );if (audit_buf ==NULL )return ;xfrm_audit_helper_pktinfo (skb, x->props.family, audit_buf);
@@ -2812,7 +2814,7 @@ void xfrm_audit_state_notfound_simple(struct sk_buff *skb, u16 family) {struct audit_buffer *audit_buf; - audit_buf =xfrm_audit_start ("SA-notfound" ); + audit_buf =xfrm_audit_start ("SA-notfound" ,NULL );if (audit_buf ==NULL )return ;xfrm_audit_helper_pktinfo (skb, family, audit_buf);
@@ -2826,7 +2828,7 @@ void xfrm_audit_state_notfound(struct sk_buff *skb, u16 family,struct audit_buffer *audit_buf;u32 spi; - audit_buf =xfrm_audit_start ("SA-notfound" ); + audit_buf =xfrm_audit_start ("SA-notfound" ,NULL );if (audit_buf ==NULL )return ;xfrm_audit_helper_pktinfo (skb, family, audit_buf);
@@ -2844,7 +2846,7 @@ void xfrm_audit_state_icvfail(struct xfrm_state *x,__be32 net_spi;__be32 net_seq; - audit_buf =xfrm_audit_start ("SA-icv-failure" ); + audit_buf =xfrm_audit_start ("SA-icv-failure" ,NULL );if (audit_buf ==NULL )return ;xfrm_audit_helper_pktinfo (skb, x->props.family, audit_buf);
security/integrity/ima/ima_api.c +4 -3
@@ -342,6 +342,7 @@ void ima_store_measurement(struct integrity_iint_cache *iint,void ima_audit_measurement (struct integrity_iint_cache *iint,const unsigned char *filename) { +struct audit_context *context;struct audit_buffer *ab;char *hash;const char *algo_name = hash_algo_name[iint->ima_hash->algo];
@@ -358,8 +359,8 @@ void ima_audit_measurement(struct integrity_iint_cache *iint,hex_byte_pack (hash + (i *2 ), iint->ima_hash->digest[i]); hash[i *2 ] ='\0' ; - ab =audit_log_start (audit_context (),GFP_KERNEL , -AUDIT_INTEGRITY_RULE ); + context =audit_alloc_for_lsm (GFP_KERNEL ); + ab =audit_log_start (context,GFP_KERNEL ,AUDIT_INTEGRITY_RULE );if (!ab) goto out;
@@ -368,7 +369,7 @@ void ima_audit_measurement(struct integrity_iint_cache *iint,audit_log_format (ab," hash=\"%s:%s\"" , algo_name, hash);audit_log_task_info (ab); -audit_log_end (ab); +audit_log_end_local (ab, context); iint->flags |=IMA_AUDITED ; out:
security/integrity/integrity_audit.c +4 -2
@@ -38,13 +38,15 @@ void integrity_audit_message(int audit_msgno, struct inode *inode, const char *cause, int result, int audit_info, int errno) { + struct audit_context *context;struct audit_buffer *ab;char name[TASK_COMM_LEN ];if (!integrity_audit_info && audit_info ==1 )/* Skip info messages */ return ; - ab =audit_log_start (audit_context (),GFP_KERNEL , audit_msgno); + context =audit_alloc_for_lsm (GFP_KERNEL ); + ab =audit_log_start (context,GFP_KERNEL , audit_msgno); audit_log_format(ab, "pid=%d uid=%u auid=%u ses=%u", task_pid_nr(current), from_kuid(&init_user_ns, current_uid()),
@@ -63,5 +65,5 @@ void integrity_audit_message(int audit_msgno, struct inode *inode,audit_log_format (ab," ino=%lu" , inode->i_ino); }audit_log_format (ab," res=%d errno=%d" , !result, errno); -audit_log_end (ab); +audit_log_end_local (ab, context); }
security/security.c +30 -16
86 lines not on this page · raw message
security/smack/smackfs.c +2 -1
@@ -185,7 +185,8 @@ static void smk_netlabel_audit_set(struct netlbl_audit *nap) nap->loginuid =audit_get_loginuid (current); nap->sessionid =audit_get_sessionid (current); - nap->secid = skp->smk_secid; +lsmblob_init (&nap->lsmdata,0 ); + nap->lsmdata.secid[smack_lsmid.slot] = skp->smk_secid; }/*
Signature · 1 line
2.29.2
Paul Moore
maintainerRe: [PATCH v26 22/25] Audit: Add new record for multiple process LSM attributes
Quoted 5 lines · Casey Schaufler
Create a new audit record type to contain the subject information when there are multiple security modules that require such data. This record is linked with the same timestamp and serial number using the
audit_alloc_local()mechanism.
The record is linked with the other associated records into a single event, it doesn't matter if it gets the timestamp/serial from audit_alloc_local() or an existing audit event, e.g. ongoing syscall.
Quoted 10 lines
The record is produced only in cases where there is more than one security module with a process "context".
In cases where this record is produced the subj= fields of other records in the audit event will be set to "subj=?".An example of the
MAC_TASK_CONTEXTS(1420) record is:type=UNKNOWN[1420] msg=audit(1600880931.832:113) subj_apparmor==unconfined
It should be just a single "=" in the line above.
Quoted 40 lines
subj_smack=_
There will be a subj_$LSM= entry for each security module LSM that supports the
secid_to_secctxandsecctx_to_secidhooks. The BPF security module implements secid/secctx translation hooks, so it has to be considered to provide a secctx even though it may not actually do so.Signed-off-by Casey Schaufler <[email protected]>To: [email protected]
To: [email protected]
To: [email protected]
Cc: [email protected]--- drivers/android/binder.c | 2 +- include/linux/audit.h | 24 ++++++++ include/linux/security.h | 16 ++++- include/net/netlabel.h | 3 +- include/net/scm.h | 2 +- include/net/xfrm.h | 13 +++- include/uapi/linux/audit.h | 1 + kernel/audit.c | 80 ++++++++++++++++++------- kernel/audit.h | 3 + kernel/auditfilter.c | 6 +- kernel/auditsc.c | 75 ++++++++++++++++++++--- net/ipv4/ip_sockglue.c | 2 +- net/netfilter/nf_conntrack_netlink.c | 4 +- net/netfilter/nf_conntrack_standalone.c | 2 +- net/netfilter/nfnetlink_queue.c | 2 +- net/netlabel/netlabel_domainhash.c | 4 +- net/netlabel/netlabel_unlabeled.c | 24 ++++---- net/netlabel/netlabel_user.c | 20 ++++--- net/netlabel/netlabel_user.h | 6 +- net/xfrm/xfrm_policy.c | 10 ++-- net/xfrm/xfrm_state.c | 20 ++++--- security/integrity/ima/ima_api.c | 7 ++- security/integrity/integrity_audit.c | 6 +- security/security.c | 46 +++++++++----- security/smack/smackfs.c | 3 +- 25 files changed, 274 insertions(+), 107 deletions(-)
...
Quoted 10 lines
include/linux/audit.h +2
-0@@ -386,6 +395,19 @@ static inline void audit_ptrace(struct task_struct *t)__audit_ptrace (t); } +static inline struct audit_context *audit_alloc_for_lsm (gfp_t gfp) +{
include/linux/audit.h static inline void audit_ptrace(struct task_struct *t) · line 409
Quoted 10 lines
include/linux/audit.h +10
-0+struct audit_context *context =audit_context (); + +if (context) +return context; + +if (lsm_multiple_contexts ()) +return audit_alloc_local (gfp); + +return NULL ; +}
See my other comments, but this seems wrong at face value. The additional LSM record should happen as part of the existing audit log functions.
Quoted 89 lines
include/linux/security.h +14
-2@@ -182,6 +182,8 @@ struct lsmblob {#define LSMBLOB_INVALID -1/* Not a valid LSM slot number */ #define LSMBLOB_NEEDED -2/* Slot requested on initialization */ #define LSMBLOB_NOT_NEEDED -3/* Slot not requested */ +#define LSMBLOB_DISPLAY -4/* Use the "display" slot */ +#define LSMBLOB_FIRST -5/* Use the default "display" slot */ /** * lsmblob_init - initialize an lsmblob structure @@ -248,6 +250,15 @@ static inline u32 lsmblob_value(const struct lsmblob *blob)return 0 ; } +static inline bool lsm_multiple_contexts (void ) +{ +#ifdef CONFIG_SECURITY +return lsm_slot_to_name (1 ) !=NULL ; +#else +return false; +#endif +} +/* These functions are in security/commoncap.c */ extern int cap_capable (const struct cred *cred,struct user_namespace *ns,int cap,unsigned int opts);@@ -578,7 +589,8 @@ int security_setprocattr(const char *lsm, const char *name, void *value, size_t size);int security_netlink_send (struct sock *sk,struct sk_buff *skb);int security_ismaclabel (const char *name); -int security_secid_to_secctx (struct lsmblob *blob,struct lsmcontext *cp); +int security_secid_to_secctx (struct lsmblob *blob,struct lsmcontext *cp, +int display);int security_secctx_to_secid (const char *secdata,u32 seclen,struct lsmblob *blob);void security_release_secctx (struct lsmcontext *cp);@@ -1433,7 +1445,7 @@ static inline int security_ismaclabel(const char *name) }static inline int security_secid_to_secctx (struct lsmblob *blob, -struct lsmcontext *cp) +struct lsmcontext *cp,int display) {return -EOPNOTSUPP ; }include/net/netlabel.h +2
-1@@ -97,7 +97,8 @@ struct calipso_doi;/* NetLabel audit information */ struct netlbl_audit { -u32 secid; +struct audit_context *localcontext; +struct lsmblob lsmdata;kuid_t loginuid;unsigned int sessionid; };include/net/scm.h +1
-1@@ -101,7 +101,7 @@ static inline void scm_passec(struct socket *sock, struct msghdr *msg, struct sc* and the infrastructure will know which it is. */ lsmblob_init (&lb, scm->secid); - err =security_secid_to_secctx (&lb, &context); + err =security_secid_to_secctx (&lb, &context,LSMBLOB_DISPLAY );if (!err) { put_cmsg(msg, SOL_SOCKET, SCM_SECURITY, context.len,include/net/xfrm.h +3
-1@@ -669,13 +669,22 @@ struct xfrm_spi_skb_cb {#define XFRM_SPI_SKB_CB (__skb) ((struct xfrm_spi_skb_cb *)&((__skb)->cb[0]))#ifdef CONFIG_AUDITSYSCALL -static inline struct audit_buffer *xfrm_audit_start (const char *op) +static inline struct audit_buffer *xfrm_audit_start (const char *op, +struct audit_context **lac) { +struct audit_context *context;struct audit_buffer *audit_buf =NULL ;
include/net/xfrm.h struct xfrm_spi_skb_cb { · line 685
Quoted 10 lines
include/net/xfrm.h +6
-1if (audit_enabled ==AUDIT_OFF )return NULL ; - audit_buf = audit_log_start(audit_context(), GFP_ATOMIC, + context =audit_context (); +if (lac !=NULL ) { +if (lsm_multiple_contexts () && context ==NULL ) + context =audit_alloc_local (GFP_ATOMIC ); + *lac = context; + }
Okay, we've got a disconnect here regarding "audit contexts" and "local contexts", skip down below where I attempt to explain things a little more but basically if there is a place that uses this pattern:
audit_log_start(audit_context(), ...);
... you don't need, or want, a "local context". You might need a local context if you see the following pattern:
audit_log_start(NULL, ...);
The "local context" idea is a hack and should be avoided whenever possible; if you have an existing audit context from a syscall, or something else, you *really* should use it ... or have a *really* good explanation as to why you can not.
Quoted 10 lines
include/net/xfrm.h +1
-0+ audit_buf =audit_log_start (context,GFP_ATOMIC ,AUDIT_MAC_IPSEC_EVENT );if (audit_buf ==NULL )return NULL ;kernel/audit.c +0
-0
kernel/audit.c void audit_log_lost(const char *message) · line 394
Quoted 10 lines
kernel/audit.c +3
-1@@ -386,10 +386,12 @@ void audit_log_lost(const char *message)static int audit_log_config_change (char *function_name,u32 new,u32 old,int allow_changes) { +struct audit_context *context;struct audit_buffer *ab;int rc =0 ; - ab =audit_log_start (audit_context (),GFP_KERNEL ,AUDIT_CONFIG_CHANGE ); + context =audit_alloc_for_lsm (GFP_KERNEL ); + ab =audit_log_start (context,GFP_KERNEL ,AUDIT_CONFIG_CHANGE );
Use the existing context, don't create your own, it breaks the record associations in the audit event stream.
kernel/audit.c static int audit_log_config_change(char *function_name, u32 new, u32 old, · line 403
Quoted 9 lines
kernel/audit.c +1
-1if (unlikely (!ab))return rc;audit_log_format (ab,"op=set %s=%u old=%u " , function_name, new, old);@@ -398,7 +400,7 @@ static int audit_log_config_change(char *function_name, u32 new, u32 old,if (rc) allow_changes =0 ;/* Something weird, deny request */ audit_log_format (ab," res=%d" , allow_changes); -audit_log_end (ab); +audit_log_end_local (ab, context);
More on this below, but we really should just use audit_log_end(), "local contexts" are not special, the are regular audit contexts ... although if they are used properly (limited scope) you do need to free them when you are done.
kernel/audit.c static int audit_receive_msg(struct sk_buff *skb, struct nlmsghdr *nlh) · line 1359
Quoted 11 lines
kernel/audit.c +2
-1return rc; }@@ -1357,7 +1355,8 @@ static int audit_receive_msg(struct sk_buff *skb, struct nlmsghdr *nlh)if (err)break ; } -audit_log_user_recv_msg (&ab, msg_type); + lcontext =audit_alloc_for_lsm (GFP_KERNEL ); +audit_log_common_recv_msg (lcontext, &ab, msg_type);
Same.
kernel/audit.c static int audit_receive_msg(struct sk_buff *skb, struct nlmsghdr *nlh) · line 1372
Quoted 9 lines
kernel/audit.c +1
-1if (msg_type !=AUDIT_USER_TTY ) {/* ensure NULL termination */ str[data_len -1 ] ='\0' ;@@ -1370,7 +1369,7 @@ static int audit_receive_msg(struct sk_buff *skb, struct nlmsghdr *nlh) data_len--;audit_log_n_untrustedstring (ab, str, data_len); } -audit_log_end (ab); +audit_log_end_local (ab, lcontext);
Same.
Quoted 7 lines
kernel/audit.c +0
-0}break ;case AUDIT_ADD_RULE :@@ -1378,13 +1377,14 @@ static int audit_receive_msg(struct sk_buff *skb, struct nlmsghdr *nlh)if (data_len <sizeof (struct audit_rule_data ))return -EINVAL ;if (audit_enabled ==AUDIT_LOCKED ) {
kernel/audit.c static int audit_receive_msg(struct sk_buff *skb, struct nlmsghdr *nlh) · line 1387
Quoted 10 lines
kernel/audit.c +3
-2-audit_log_common_recv_msg (audit_context (), &ab, + lcontext =audit_alloc_for_lsm (GFP_KERNEL ); +audit_log_common_recv_msg (lcontext, &ab,AUDIT_CONFIG_CHANGE );audit_log_format (ab," op=%s audit_enabled=%d res=0" , msg_type ==AUDIT_ADD_RULE ?"add_rule" :"remove_rule" , audit_enabled); -audit_log_end (ab); +audit_log_end_local (ab, lcontext);
Same. I'm going to stop calling these out, I think you get the idea.
Quoted 8 lines
kernel/audit.c +4
-0@@ -2396,6 +2415,21 @@ void audit_log_end(struct audit_buffer *ab)audit_buffer_free (ab); } +/** +* audit_log_end_local - end one audit record with local context +* @ab: the audit_buffer +* @context: the local context
kernel/audit.c void audit_log_end(struct audit_buffer *ab) · line 2431
Quoted 10 lines
kernel/audit.c +10
-0+* +* Emit an LSM context record if appropriate, then end the audit event +* in the usual way. +*/ +void audit_log_end_local (struct audit_buffer *ab,struct audit_context *context) +{ +audit_log_end (ab); +audit_log_lsm_common (context); +audit_free_local (context); +}
Eeesh, no, not this please.
First, some background on audit contexts and the idea of a "local context" as we have been using it in the audit container ID work, which is where this originated. An audit context contains a few things, but likely the most important for this discussion is the audit event timestamp and serial number (I may refer to this combo as just a "timestamp" in the future); this timestamp/serial is shared across all of the audit records that make up this audit event, linking them together. A shared timestamp is what allows you to group an open() SYSCALL record with the PATH record that provides the file's pathname info.
While there are some exceptions in the current code, most audit events occur as a result of a syscall, and their audit context in this case is the syscall's audit context (see the open() example above), but there are some cases being discussed where we have an audit event that does not occur as a result of a syscall but there is a need to group multiple audit records together in a single event. This is where the "local context" comes into play, it allows us to create an audit context outside of a syscall and share that context across multiple audit records, allowing the records to share a timestamp/serial and grouping them together as a single audit event in the audit stream.
While a function like audit_alloc_local() make sense, there really shouldn't be an audit_log_end_local() function, the normal audit_log_end() function should be used.
Does that make sense?
kernel/audit.h struct audit_context { · line 103
Quoted 9 lines
kernel/audit.h +1
-0@@ -100,6 +100,7 @@ struct audit_context {int dummy;/* must be the first element */ int in_syscall;/* 1 if task is in a syscall */ bool local;/* local context needed */ +bool lsmdone;/* multiple security reported */
"lsmdone" doesn't seem consistent with the comment, how about "lsm_multi" or something similar?
Quoted 9 lines
kernel/audit.h +0
-0enum audit_state state, current_state;unsigned int serial;/* serial number for record */ int major;/* syscall number */ kernel/auditsc.c +0
-0
kernel/auditsc.c void audit_free_context(struct audit_context *context) · line 1021
Quoted 9 lines
kernel/auditsc.c +6
-0@@ -1013,6 +1013,13 @@ void audit_free_context(struct audit_context *context) }EXPORT_SYMBOL (audit_free_context); +void audit_free_local (struct audit_context *context) +{ +if (context && context->local) +audit_free_context (context); +} +EXPORT_SYMBOL (audit_free_local);
We don't need this function, just use audit_free_context(). A "local context" is the same as a non-local context; what makes a context "local" is the scope of the audit context (local function scope vs syscall scope) and nothing else.
Quoted 9 lines
kernel/auditsc.c +5
-0@@ -1504,6 +1512,47 @@ static void audit_log_proctitle(void)audit_log_end (ab); } +void audit_log_lsm_common (struct audit_context *context) +{ +struct audit_buffer *ab; +struct lsmcontext lsmdata; +bool sep = false;
kernel/auditsc.c static void audit_log_proctitle(void) · line 1529
Quoted 10 lines
kernel/auditsc.c +10
-0+int error; +int i; + +if (!lsm_multiple_contexts () || context ==NULL || + !lsmblob_is_set (&context->lsm)) +return ; + + ab =audit_log_start (context,GFP_ATOMIC ,AUDIT_MAC_TASK_CONTEXTS ); +if (!ab) +return ;/* audit_panic or being filtered */
We should be consistent with our use of audit_panic() when we bail on error; we use it below, but not here - why?
Quoted 6 lines
kernel/auditsc.c +6
-0+for (i =0 ; i <LSMBLOB_ENTRIES ; i++) { +if (context->lsm.secid[i] ==0 ) +continue ; + error =security_secid_to_secctx (&context->lsm, &lsmdata, i); +if (error && error != -EINVAL ) { +audit_panic ("error in audit_log_lsm" );
kernel/auditsc.c static void audit_log_proctitle(void) · line 1547
Quoted 10 lines
kernel/auditsc.c +10
-0+return ; + } + +audit_log_format (ab,"%ssubj_%s=%s" , sep ?" " :"" , +lsm_slot_to_name (i), lsmdata.context); + sep = true; +security_release_secctx (&lsmdata); + } +audit_log_end (ab); + context->lsmdone = true;
Maybe I missed it, but why do we need this flag?
kernel/auditsc.c static void audit_log_proctitle(void) · line 1554
Quoted 7 lines
kernel/auditsc.c +7
-0+} + +void audit_log_lsm (struct audit_context *context) +{ +if (!context->lsmdone) +audit_log_lsm_common (context); +}
I think I was distracted with the local context issue and I've lost track of the details here, perhaps it's best to fix the local context issue first (that should be a big change to this patch) and then we can take another look.
--
paul moore
www.paul-moore.com
--
Linux-audit mailing list
[email protected]
https://listman.redhat.com/mailman/listinfo/linux-audit
Re: [PATCH v26 22/25] Audit: Add new record for multiple process LSM attributes
Quoted 9 lines · Paul Moore
Quoted 4 lines · Casey Schaufler
Create a new audit record type to contain the subject information when there are multiple security modules that require such data. This record is linked with the same timestamp and serial number using the
audit_alloc_local()mechanism.The record is linked with the other associated records into a single event, it doesn't matter if it gets the timestamp/serial from
audit_alloc_local()or an existing audit event, e.g. ongoing syscall.
I had similar concerns this would be misunderstood...
Quoted 64 lines
Quoted 10 lines
The record is produced only in cases where there is more than one security module with a process "context".
In cases where this record is produced the subj= fields of other records in the audit event will be set to "subj=?".An example of the
MAC_TASK_CONTEXTS(1420) record is:type=UNKNOWN[1420] msg=audit(1600880931.832:113) subj_apparmor==unconfinedIt should be just a single "=" in the line above.
Quoted 40 lines
subj_smack=_
There will be a subj_$LSM= entry for each security module LSM that supports the
secid_to_secctxandsecctx_to_secidhooks. The BPF security module implements secid/secctx translation hooks, so it has to be considered to provide a secctx even though it may not actually do so.Signed-off-by Casey Schaufler <[email protected]>To: [email protected]
To: [email protected]
To: [email protected]
Cc: [email protected]--- drivers/android/binder.c | 2 +- include/linux/audit.h | 24 ++++++++ include/linux/security.h | 16 ++++- include/net/netlabel.h | 3 +- include/net/scm.h | 2 +- include/net/xfrm.h | 13 +++- include/uapi/linux/audit.h | 1 + kernel/audit.c | 80 ++++++++++++++++++------- kernel/audit.h | 3 + kernel/auditfilter.c | 6 +- kernel/auditsc.c | 75 ++++++++++++++++++++--- net/ipv4/ip_sockglue.c | 2 +- net/netfilter/nf_conntrack_netlink.c | 4 +- net/netfilter/nf_conntrack_standalone.c | 2 +- net/netfilter/nfnetlink_queue.c | 2 +- net/netlabel/netlabel_domainhash.c | 4 +- net/netlabel/netlabel_unlabeled.c | 24 ++++---- net/netlabel/netlabel_user.c | 20 ++++--- net/netlabel/netlabel_user.h | 6 +- net/xfrm/xfrm_policy.c | 10 ++-- net/xfrm/xfrm_state.c | 20 ++++--- security/integrity/ima/ima_api.c | 7 ++- security/integrity/integrity_audit.c | 6 +- security/security.c | 46 +++++++++----- security/smack/smackfs.c | 3 +- 25 files changed, 274 insertions(+), 107 deletions(-)...
Quoted 8 lines
include/linux/audit.h +0
-0@@ -386,6 +395,19 @@ static inline void audit_ptrace(struct task_struct *t)__audit_ptrace (t); }
Quoted 16 lines
include/linux/audit.h static inline void audit_ptrace(struct task_struct *t) · line 409
Quoted 12 lines
include/linux/audit.h +12
-0+static inline struct audit_context *audit_alloc_for_lsm (gfp_t gfp) +{ +struct audit_context *context =audit_context (); + +if (context) +return context; + +if (lsm_multiple_contexts ()) +return audit_alloc_local (gfp); + +return NULL ; +}See my other comments, but this seems wrong at face value. The additional LSM record should happen as part of the existing audit log functions.
Right. The only time a local context should be necessary is if there is a group of records that should be divorced from a syscall or that has no syscall (or other similar context).
Quoted 109 lines
Quoted 89 lines
include/linux/security.h +14
-2@@ -182,6 +182,8 @@ struct lsmblob {#define LSMBLOB_INVALID -1/* Not a valid LSM slot number */ #define LSMBLOB_NEEDED -2/* Slot requested on initialization */ #define LSMBLOB_NOT_NEEDED -3/* Slot not requested */ +#define LSMBLOB_DISPLAY -4/* Use the "display" slot */ +#define LSMBLOB_FIRST -5/* Use the default "display" slot */ /** * lsmblob_init - initialize an lsmblob structure @@ -248,6 +250,15 @@ static inline u32 lsmblob_value(const struct lsmblob *blob)return 0 ; } +static inline bool lsm_multiple_contexts (void ) +{ +#ifdef CONFIG_SECURITY +return lsm_slot_to_name (1 ) !=NULL ; +#else +return false; +#endif +} +/* These functions are in security/commoncap.c */ extern int cap_capable (const struct cred *cred,struct user_namespace *ns,int cap,unsigned int opts);@@ -578,7 +589,8 @@ int security_setprocattr(const char *lsm, const char *name, void *value, size_t size);int security_netlink_send (struct sock *sk,struct sk_buff *skb);int security_ismaclabel (const char *name); -int security_secid_to_secctx (struct lsmblob *blob,struct lsmcontext *cp); +int security_secid_to_secctx (struct lsmblob *blob,struct lsmcontext *cp, +int display);int security_secctx_to_secid (const char *secdata,u32 seclen,struct lsmblob *blob);void security_release_secctx (struct lsmcontext *cp);@@ -1433,7 +1445,7 @@ static inline int security_ismaclabel(const char *name) }static inline int security_secid_to_secctx (struct lsmblob *blob, -struct lsmcontext *cp) +struct lsmcontext *cp,int display) {return -EOPNOTSUPP ; }include/net/netlabel.h +2
-1@@ -97,7 +97,8 @@ struct calipso_doi;/* NetLabel audit information */ struct netlbl_audit { -u32 secid; +struct audit_context *localcontext; +struct lsmblob lsmdata;kuid_t loginuid;unsigned int sessionid; };include/net/scm.h +1
-1@@ -101,7 +101,7 @@ static inline void scm_passec(struct socket *sock, struct msghdr *msg, struct sc* and the infrastructure will know which it is. */ lsmblob_init (&lb, scm->secid); - err =security_secid_to_secctx (&lb, &context); + err =security_secid_to_secctx (&lb, &context,LSMBLOB_DISPLAY );if (!err) { put_cmsg(msg, SOL_SOCKET, SCM_SECURITY, context.len,include/net/xfrm.h +3
-1@@ -669,13 +669,22 @@ struct xfrm_spi_skb_cb {#define XFRM_SPI_SKB_CB (__skb) ((struct xfrm_spi_skb_cb *)&((__skb)->cb[0]))#ifdef CONFIG_AUDITSYSCALL -static inline struct audit_buffer *xfrm_audit_start (const char *op) +static inline struct audit_buffer *xfrm_audit_start (const char *op, +struct audit_context **lac) { +struct audit_context *context;struct audit_buffer *audit_buf =NULL ;include/net/xfrm.h struct xfrm_spi_skb_cb { · line 685
Quoted 10 lines
include/net/xfrm.h +6
-1if (audit_enabled ==AUDIT_OFF )return NULL ; - audit_buf = audit_log_start(audit_context(), GFP_ATOMIC, + context =audit_context (); +if (lac !=NULL ) { +if (lsm_multiple_contexts () && context ==NULL ) + context =audit_alloc_local (GFP_ATOMIC ); + *lac = context; + }Okay, we've got a disconnect here regarding "audit contexts" and "local contexts", skip down below where I attempt to explain things a little more but basically if there is a place that uses this pattern:
audit_log_start(audit_context(), ...);... you don't need, or want, a "local context". You might need a local context if you see the following pattern:
Quoted 6 lines
audit_log_start(NULL, ...);The "local context" idea is a hack and should be avoided whenever possible; if you have an existing audit context from a syscall, or something else, you *really* should use it ... or have a *really* good explanation as to why you can not.
Yes, what Paul said. Examples include network events that have no syscall related to them or a user record that isn't really related to the syscall that produced it.
Quoted 17 lines
Quoted 17 lines
include/net/xfrm.h +1
-0+ audit_buf =audit_log_start (context,GFP_ATOMIC ,AUDIT_MAC_IPSEC_EVENT );if (audit_buf ==NULL )return NULL ;kernel/audit.c +1
-0@@ -386,10 +386,12 @@ void audit_log_lost(const char *message)static int audit_log_config_change (char *function_name,u32 new,u32 old,int allow_changes) { +struct audit_context *context;struct audit_buffer *ab;int rc =0 ;
Quoted 6 lines
kernel/audit.c void audit_log_lost(const char *message) · line 394
Quoted 3 lines
kernel/audit.c +2
-1- ab =audit_log_start (audit_context (),GFP_KERNEL ,AUDIT_CONFIG_CHANGE ); + context =audit_alloc_for_lsm (GFP_KERNEL ); + ab =audit_log_start (context,GFP_KERNEL ,AUDIT_CONFIG_CHANGE );Use the existing context, don't create your own, it breaks the record associations in the audit event stream.
Agreed.
Quoted 6 lines
Quoted 6 lines
kernel/audit.c +0
-0if (unlikely (!ab))return rc;audit_log_format (ab,"op=set %s=%u old=%u " , function_name, new, old);@@ -398,7 +400,7 @@ static int audit_log_config_change(char *function_name, u32 new, u32 old,if (rc) allow_changes =0 ;/* Something weird, deny request */
Quoted 8 lines
kernel/audit.c static int audit_log_config_change(char *function_name, u32 new, u32 old, · line 403
Quoted 3 lines
kernel/audit.c +1
-1audit_log_format (ab," res=%d" , allow_changes); -audit_log_end (ab); +audit_log_end_local (ab, context);More on this below, but we really should just use
audit_log_end(), "local contexts" are not special, the are regular audit contexts ... although if they are used properly (limited scope) you do need to free them when you are done.
I was a bit puzzled about this before and had the same concerns.
Quoted 50 lines
kernel/audit.c static int audit_receive_msg(struct sk_buff *skb, struct nlmsghdr *nlh) · line 1359
Quoted 11 lines
kernel/audit.c +2
-1return rc; }@@ -1357,7 +1355,8 @@ static int audit_receive_msg(struct sk_buff *skb, struct nlmsghdr *nlh)if (err)break ; } -audit_log_user_recv_msg (&ab, msg_type); + lcontext =audit_alloc_for_lsm (GFP_KERNEL ); +audit_log_common_recv_msg (lcontext, &ab, msg_type);Same.
kernel/audit.c static int audit_receive_msg(struct sk_buff *skb, struct nlmsghdr *nlh) · line 1372
Quoted 9 lines
kernel/audit.c +1
-1if (msg_type !=AUDIT_USER_TTY ) {/* ensure NULL termination */ str[data_len -1 ] ='\0' ;@@ -1370,7 +1369,7 @@ static int audit_receive_msg(struct sk_buff *skb, struct nlmsghdr *nlh) data_len--;audit_log_n_untrustedstring (ab, str, data_len); } -audit_log_end (ab); +audit_log_end_local (ab, lcontext);Same.
Quoted 7 lines
kernel/audit.c +0
-0}break ;case AUDIT_ADD_RULE :@@ -1378,13 +1377,14 @@ static int audit_receive_msg(struct sk_buff *skb, struct nlmsghdr *nlh)if (data_len <sizeof (struct audit_rule_data ))return -EINVAL ;if (audit_enabled ==AUDIT_LOCKED ) {kernel/audit.c static int audit_receive_msg(struct sk_buff *skb, struct nlmsghdr *nlh) · line 1387
Quoted 10 lines
kernel/audit.c +3
-2-audit_log_common_recv_msg (audit_context (), &ab, + lcontext =audit_alloc_for_lsm (GFP_KERNEL ); +audit_log_common_recv_msg (lcontext, &ab,AUDIT_CONFIG_CHANGE );audit_log_format (ab," op=%s audit_enabled=%d res=0" , msg_type ==AUDIT_ADD_RULE ?"add_rule" :"remove_rule" , audit_enabled); -audit_log_end (ab); +audit_log_end_local (ab, lcontext);Same. I'm going to stop calling these out, I think you get the idea.
Quoted 4 lines
kernel/audit.c +0
-0@@ -2396,6 +2415,21 @@ void audit_log_end(struct audit_buffer *ab)audit_buffer_free (ab); }
Quoted 16 lines
Quoted 4 lines
kernel/audit.c +4
-0+/** +* audit_log_end_local - end one audit record with local context +* @ab: the audit_buffer +* @context: the local context kernel/audit.c void audit_log_end(struct audit_buffer *ab) · line 2431
Quoted 10 lines
kernel/audit.c +10
-0+* +* Emit an LSM context record if appropriate, then end the audit event +* in the usual way. +*/ +void audit_log_end_local (struct audit_buffer *ab,struct audit_context *context) +{ +audit_log_end (ab); +audit_log_lsm_common (context); +audit_free_local (context); +}Eeesh, no, not this please.
It was this that I found hard to follow and justify.
Quoted 10 lines
First, some background on audit contexts and the idea of a "local context" as we have been using it in the audit container ID work, which is where this originated. An audit context contains a few things, but likely the most important for this discussion is the audit event timestamp and serial number (I may refer to this combo as just a "timestamp" in the future); this timestamp/serial is shared across all of the audit records that make up this audit event, linking them together. A shared timestamp is what allows you to group an
open()SYSCALL record with the PATH record that provides the file's pathname info.
How about a "postmark"? :-) (Yes, timestamp works...)
Quoted 11 lines
While there are some exceptions in the current code, most audit events occur as a result of a syscall, and their audit context in this case is the syscall's audit context (see the
open()example above), but there are some cases being discussed where we have an audit event that does not occur as a result of a syscall but there is a need to group multiple audit records together in a single event. This is where the "local context" comes into play, it allows us to create an audit context outside of a syscall and share that context across multiple audit records, allowing the records to share a timestamp/serial and grouping them together as a single audit event in the audit stream.
Quoted 3 lines
While a function like
audit_alloc_local()make sense, there really shouldn't be anaudit_log_end_local()function, the normalaudit_log_end()function should be used.
Exactly.
Quoted 27 lines
Does that make sense?
kernel/audit.h struct audit_context { · line 103
Quoted 9 lines
kernel/audit.h +1
-0@@ -100,6 +100,7 @@ struct audit_context {int dummy;/* must be the first element */ int in_syscall;/* 1 if task is in a syscall */ bool local;/* local context needed */ +bool lsmdone;/* multiple security reported */ "lsmdone" doesn't seem consistent with the comment, how about "
lsm_multi" or something similar?Quoted 12 lines
kernel/audit.h +0
-0enum audit_state state, current_state;unsigned int serial;/* serial number for record */ int major;/* syscall number */ kernel/auditsc.c +0
-0@@ -1013,6 +1013,13 @@ void audit_free_context(struct audit_context *context) }EXPORT_SYMBOL (audit_free_context);
Quoted 11 lines
kernel/auditsc.c void audit_free_context(struct audit_context *context) · line 1021
Quoted 6 lines
kernel/auditsc.c +6
-0+void audit_free_local (struct audit_context *context) +{ +if (context && context->local) +audit_free_context (context); +} +EXPORT_SYMBOL (audit_free_local);We don't need this function, just use
audit_free_context(). A "local context" is the same as a non-local context; what makes a context "local" is the scope of the audit context (local function scope vs syscall scope) and nothing else.
Agreed.
Quoted 40 lines
Quoted 9 lines
kernel/auditsc.c +5
-0@@ -1504,6 +1512,47 @@ static void audit_log_proctitle(void)audit_log_end (ab); } +void audit_log_lsm_common (struct audit_context *context) +{ +struct audit_buffer *ab; +struct lsmcontext lsmdata; +bool sep = false;kernel/auditsc.c static void audit_log_proctitle(void) · line 1529
Quoted 10 lines
kernel/auditsc.c +10
-0+int error; +int i; + +if (!lsm_multiple_contexts () || context ==NULL || + !lsmblob_is_set (&context->lsm)) +return ; + + ab =audit_log_start (context,GFP_ATOMIC ,AUDIT_MAC_TASK_CONTEXTS ); +if (!ab) +return ;/* audit_panic or being filtered */ We should be consistent with our use of
audit_panic()when we bail on error; we use it below, but not here - why?Quoted 6 lines
kernel/auditsc.c +6
-0+for (i =0 ; i <LSMBLOB_ENTRIES ; i++) { +if (context->lsm.secid[i] ==0 ) +continue ; + error =security_secid_to_secctx (&context->lsm, &lsmdata, i); +if (error && error != -EINVAL ) { +audit_panic ("error in audit_log_lsm" );kernel/auditsc.c static void audit_log_proctitle(void) · line 1547
Quoted 10 lines
kernel/auditsc.c +10
-0+return ; + } + +audit_log_format (ab,"%ssubj_%s=%s" , sep ?" " :"" , +lsm_slot_to_name (i), lsmdata.context); + sep = true; +security_release_secctx (&lsmdata); + } +audit_log_end (ab); + context->lsmdone = true;
Quoted 1 line
Maybe I missed it, but why do we need this flag?
This is partly what confused me earlier...
Quoted 4 lines
Quoted 4 lines
kernel/auditsc.c +4
-0+} + +void audit_log_lsm (struct audit_context *context) +{
Quoted 8 lines
kernel/auditsc.c static void audit_log_proctitle(void) · line 1554
Quoted 3 lines
kernel/auditsc.c +3
-0+if (!context->lsmdone) +audit_log_lsm_common (context); +}I think I was distracted with the local context issue and I've lost track of the details here, perhaps it's best to fix the local context issue first (that should be a big change to this patch) and then we can take another look.
I got distracted by the local context issues too earlier... Sorry, this feedback would have been more useful earlier. I never completed that review and got pulled off to other priorities...
Quoted 1 line
paul moore
- RGB
--
Richard Guy Briggs <[email protected]>
Sr. S/W Engineer, Kernel Security, Base Operating Systems Remote, Ottawa, Red Hat Canada
IRC: rgb, SunRaycer
Voice: +1.647.777.2635, Internal: (81) 32635
Casey Schaufler
maintainerRe: [PATCH v26 22/25] Audit: Add new record for multiple process LSM attributes
In reply to Paul MooreQuoted 14 lines · Paul Moore
Quoted 4 lines · Casey Schaufler
Create a new audit record type to contain the subject information when there are multiple security modules that require such data. This record is linked with the same timestamp and serial number using the
audit_alloc_local()mechanism.The record is linked with the other associated records into a single event, it doesn't matter if it gets the timestamp/serial from
audit_alloc_local()or an existing audit event, e.g. ongoing syscall.Quoted 5 lines
The record is produced only in cases where there is more than one security module with a process "context".
In cases where this record is produced the subj= fields of other records in the audit event will be set to "subj=?".
Quoted 6 lines
Quoted 5 lines
An example of the
MAC_TASK_CONTEXTS(1420) record is:type=UNKNOWN[1420] msg=audit(1600880931.832:113) subj_apparmor==unconfinedIt should be just a single "=" in the line above.
AppArmor provides the 2nd "=" as part of the subject context. What's here is correct. I won't argue that it won't case confusion or worse.
Quoted 50 lines
Quoted 40 lines
subj_smack=_
There will be a subj_$LSM= entry for each security module LSM that supports the
secid_to_secctxandsecctx_to_secidhooks. The BPF security module implements secid/secctx translation hooks, so it has to be considered to provide a secctx even though it may not actually do so.Signed-off-by Casey Schaufler <[email protected]>To: [email protected]
To: [email protected]
To: [email protected]
Cc: [email protected]--- drivers/android/binder.c | 2 +- include/linux/audit.h | 24 ++++++++ include/linux/security.h | 16 ++++- include/net/netlabel.h | 3 +- include/net/scm.h | 2 +- include/net/xfrm.h | 13 +++- include/uapi/linux/audit.h | 1 + kernel/audit.c | 80 ++++++++++++++++++------- kernel/audit.h | 3 + kernel/auditfilter.c | 6 +- kernel/auditsc.c | 75 ++++++++++++++++++++--- net/ipv4/ip_sockglue.c | 2 +- net/netfilter/nf_conntrack_netlink.c | 4 +- net/netfilter/nf_conntrack_standalone.c | 2 +- net/netfilter/nfnetlink_queue.c | 2 +- net/netlabel/netlabel_domainhash.c | 4 +- net/netlabel/netlabel_unlabeled.c | 24 ++++---- net/netlabel/netlabel_user.c | 20 ++++--- net/netlabel/netlabel_user.h | 6 +- net/xfrm/xfrm_policy.c | 10 ++-- net/xfrm/xfrm_state.c | 20 ++++--- security/integrity/ima/ima_api.c | 7 ++- security/integrity/integrity_audit.c | 6 +- security/security.c | 46 +++++++++----- security/smack/smackfs.c | 3 +- 25 files changed, 274 insertions(+), 107 deletions(-)...
Quoted 8 lines
include/linux/audit.h +0
-0@@ -386,6 +395,19 @@ static inline void audit_ptrace(struct task_struct *t)__audit_ptrace (t); }
Quoted 15 lines
include/linux/audit.h static inline void audit_ptrace(struct task_struct *t) · line 409
Quoted 12 lines
include/linux/audit.h +12
-0+static inline struct audit_context *audit_alloc_for_lsm (gfp_t gfp) +{ +struct audit_context *context =audit_context (); + +if (context) +return context; + +if (lsm_multiple_contexts ()) +return audit_alloc_local (gfp); + +return NULL ; +}See my other comments, but this seems wrong at face value. The additional LSM record should happen as part of the existing audit log functions.
I'm good with that. But if you defer calling audit_alloc_local() until you know you need it you may be in a place where you can't associate the new context with the event. I think. I will have another go at it.
<snip>
Quoted 4 lines
I think I was distracted with the local context issue and I've lost track of the details here, perhaps it's best to fix the local context issue first (that should be a big change to this patch) and then we can take another look.
I really need to move forward. I'll give allocation of local contexts as necessary in audit_log_task_context() another shot.
Quoted 5 lines
--
paul moore
www.paul-moore.com
Paul Moore
maintainerRe: [PATCH v26 22/25] Audit: Add new record for multiple process LSM attributes
Quoted 17 lines · Casey Schaufler
Quoted 16 lines · Paul Moore
Quoted 4 lines · Casey Schaufler
Create a new audit record type to contain the subject information when there are multiple security modules that require such data. This record is linked with the same timestamp and serial number using the
audit_alloc_local()mechanism.The record is linked with the other associated records into a single event, it doesn't matter if it gets the timestamp/serial from
audit_alloc_local()or an existing audit event, e.g. ongoing syscall.Quoted 7 lines
The record is produced only in cases where there is more than one security module with a process "context".
In cases where this record is produced the subj= fields of other records in the audit event will be set to "subj=?".An example of the
MAC_TASK_CONTEXTS(1420) record is:
Quoted 8 lines
Quoted 4 lines
Quoted 3 lines
type=UNKNOWN[1420] msg=audit(1600880931.832:113) subj_apparmor==unconfinedIt should be just a single "=" in the line above.
AppArmor provides the 2nd "=" as part of the subject context. What's here is correct. I won't argue that it won't case confusion or worse.
Oh, wow, okay. That needs to change at some point but I agree it's out of scope for this patchset. In the meantime I might suggest using something other than AppArmor as an example here.
Quoted 66 lines
Quoted 50 lines
Quoted 40 lines
subj_smack=_
There will be a subj_$LSM= entry for each security module LSM that supports the
secid_to_secctxandsecctx_to_secidhooks. The BPF security module implements secid/secctx translation hooks, so it has to be considered to provide a secctx even though it may not actually do so.Signed-off-by Casey Schaufler <[email protected]>To: [email protected]
To: [email protected]
To: [email protected]
Cc: [email protected]--- drivers/android/binder.c | 2 +- include/linux/audit.h | 24 ++++++++ include/linux/security.h | 16 ++++- include/net/netlabel.h | 3 +- include/net/scm.h | 2 +- include/net/xfrm.h | 13 +++- include/uapi/linux/audit.h | 1 + kernel/audit.c | 80 ++++++++++++++++++------- kernel/audit.h | 3 + kernel/auditfilter.c | 6 +- kernel/auditsc.c | 75 ++++++++++++++++++++--- net/ipv4/ip_sockglue.c | 2 +- net/netfilter/nf_conntrack_netlink.c | 4 +- net/netfilter/nf_conntrack_standalone.c | 2 +- net/netfilter/nfnetlink_queue.c | 2 +- net/netlabel/netlabel_domainhash.c | 4 +- net/netlabel/netlabel_unlabeled.c | 24 ++++---- net/netlabel/netlabel_user.c | 20 ++++--- net/netlabel/netlabel_user.h | 6 +- net/xfrm/xfrm_policy.c | 10 ++-- net/xfrm/xfrm_state.c | 20 ++++--- security/integrity/ima/ima_api.c | 7 ++- security/integrity/integrity_audit.c | 6 +- security/security.c | 46 +++++++++----- security/smack/smackfs.c | 3 +- 25 files changed, 274 insertions(+), 107 deletions(-)...
Quoted 8 lines
include/linux/audit.h +0
-0@@ -386,6 +395,19 @@ static inline void audit_ptrace(struct task_struct *t)__audit_ptrace (t); }Quoted 15 lines
include/linux/audit.h static inline void audit_ptrace(struct task_struct *t) · line 409
Quoted 12 lines
include/linux/audit.h +12
-0+static inline struct audit_context *audit_alloc_for_lsm (gfp_t gfp) +{ +struct audit_context *context =audit_context (); + +if (context) +return context; + +if (lsm_multiple_contexts ()) +return audit_alloc_local (gfp); + +return NULL ; +}See my other comments, but this seems wrong at face value. The additional LSM record should happen as part of the existing audit log functions.
Quoted 4 lines
I'm good with that. But if you defer calling
audit_alloc_local()until you know you need it you may be in a place where you can't associate the new context with the event. I think. I will have another go at it.
I can't think of a case where you would ever not know if you need to allocate a local context at the start. If you are unsure, get in touch and we can work it out.
Quoted 7 lines
Quoted 4 lines
I think I was distracted with the local context issue and I've lost track of the details here, perhaps it's best to fix the local context issue first (that should be a big change to this patch) and then we can take another look.
I really need to move forward. I'll give allocation of local contexts as necessary in
audit_log_task_context()another shot.
I appreciate the desire to move forward, and while I can't speak for everyone, I'll do my best to work with you to find a good solution. If you get stuck or aren't sure you know how to reach me :)
As a start, I might suggest looking at some of the recent audit container ID patchsets from Richard; while they have had some issues, they should serve as a basic example of what we mean when we talk about "local contexts" and how they should be used.
Signature · 2 lines
paul moore www.paul-moore.com
Re: [PATCH v26 22/25] Audit: Add new record for multiple process LSM attributes
Quoted 23 lines · Paul Moore
Quoted 22 lines · Casey Schaufler
Quoted 20 lines · Paul Moore
Quoted 4 lines · Casey Schaufler
Create a new audit record type to contain the subject information when there are multiple security modules that require such data. This record is linked with the same timestamp and serial number using the
audit_alloc_local()mechanism.The record is linked with the other associated records into a single event, it doesn't matter if it gets the timestamp/serial from
audit_alloc_local()or an existing audit event, e.g. ongoing syscall.Quoted 10 lines
The record is produced only in cases where there is more than one security module with a process "context".
In cases where this record is produced the subj= fields of other records in the audit event will be set to "subj=?".An example of the
MAC_TASK_CONTEXTS(1420) record is:type=UNKNOWN[1420] msg=audit(1600880931.832:113) subj_apparmor==unconfinedIt should be just a single "=" in the line above.
Quoted 7 lines
Quoted 3 lines
AppArmor provides the 2nd "=" as part of the subject context. What's here is correct. I won't argue that it won't case confusion or worse.
Oh, wow, okay. That needs to change at some point but I agree it's out of scope for this patchset. In the meantime I might suggest using something other than AppArmor as an example here.
Similar but not identical situation to:
BUG: INTEGRITY_POLICY_RULE violates audit message format · Issue #113 · linux-audit/audit-kernel https://github.com/linux-audit/audit-kernel/issues/113
Quoted 83 lines
Quoted 66 lines
Quoted 50 lines
Quoted 40 lines
subj_smack=_
There will be a subj_$LSM= entry for each security module LSM that supports the
secid_to_secctxandsecctx_to_secidhooks. The BPF security module implements secid/secctx translation hooks, so it has to be considered to provide a secctx even though it may not actually do so.Signed-off-by Casey Schaufler <[email protected]>To: [email protected]
To: [email protected]
To: [email protected]
Cc: [email protected]--- drivers/android/binder.c | 2 +- include/linux/audit.h | 24 ++++++++ include/linux/security.h | 16 ++++- include/net/netlabel.h | 3 +- include/net/scm.h | 2 +- include/net/xfrm.h | 13 +++- include/uapi/linux/audit.h | 1 + kernel/audit.c | 80 ++++++++++++++++++------- kernel/audit.h | 3 + kernel/auditfilter.c | 6 +- kernel/auditsc.c | 75 ++++++++++++++++++++--- net/ipv4/ip_sockglue.c | 2 +- net/netfilter/nf_conntrack_netlink.c | 4 +- net/netfilter/nf_conntrack_standalone.c | 2 +- net/netfilter/nfnetlink_queue.c | 2 +- net/netlabel/netlabel_domainhash.c | 4 +- net/netlabel/netlabel_unlabeled.c | 24 ++++---- net/netlabel/netlabel_user.c | 20 ++++--- net/netlabel/netlabel_user.h | 6 +- net/xfrm/xfrm_policy.c | 10 ++-- net/xfrm/xfrm_state.c | 20 ++++--- security/integrity/ima/ima_api.c | 7 ++- security/integrity/integrity_audit.c | 6 +- security/security.c | 46 +++++++++----- security/smack/smackfs.c | 3 +- 25 files changed, 274 insertions(+), 107 deletions(-)...
Quoted 8 lines
include/linux/audit.h +0
-0@@ -386,6 +395,19 @@ static inline void audit_ptrace(struct task_struct *t)__audit_ptrace (t); }Quoted 15 lines
include/linux/audit.h static inline void audit_ptrace(struct task_struct *t) · line 409
Quoted 12 lines
include/linux/audit.h +12
-0+static inline struct audit_context *audit_alloc_for_lsm (gfp_t gfp) +{ +struct audit_context *context =audit_context (); + +if (context) +return context; + +if (lsm_multiple_contexts ()) +return audit_alloc_local (gfp); + +return NULL ; +}See my other comments, but this seems wrong at face value. The additional LSM record should happen as part of the existing audit log functions.
Quoted 4 lines
I'm good with that. But if you defer calling
audit_alloc_local()until you know you need it you may be in a place where you can't associate the new context with the event. I think. I will have another go at it.I can't think of a case where you would ever not know if you need to allocate a local context at the start. If you are unsure, get in touch and we can work it out.
Quoted 7 lines
Quoted 4 lines
I think I was distracted with the local context issue and I've lost track of the details here, perhaps it's best to fix the local context issue first (that should be a big change to this patch) and then we can take another look.
I really need to move forward. I'll give allocation of local contexts as necessary in
audit_log_task_context()another shot.
Quoted 17 lines
I appreciate the desire to move forward, and while I can't speak for everyone, I'll do my best to work with you to find a good solution. If you get stuck or aren't sure you know how to reach me :)
As a start, I might suggest looking at some of the recent audit container ID patchsets from Richard; while they have had some issues, they should serve as a basic example of what we mean when we talk about "local contexts" and how they should be used.
Signature · 7 lines
paul moore www.paul-moore.com -- Linux-audit mailing list [email protected] https://listman.redhat.com/mailman/listinfo/linux-audit
- RGB
--
Richard Guy Briggs <[email protected]>
Sr. S/W Engineer, Kernel Security, Base Operating Systems Remote, Ottawa, Red Hat Canada
IRC: rgb, SunRaycer
Voice: +1.647.777.2635, Internal: (81) 32635
Re: [PATCH v26 22/25] Audit: Add new record for multiple process LSM attributes
In reply to Casey SchauflerQuoted 11 lines · EDT Casey Schaufler
Quoted 11 lines
Quoted 10 lines
The record is produced only in cases where there is more than one security module with a process "context".
In cases where this record is produced the subj= fields of other records in the audit event will be set to "subj=?".An example of the
MAC_TASK_CONTEXTS(1420) record is:type=UNKNOWN[1420]
msg=audit(1600880931.832:113)
subj_apparmor==unconfined
Quoted 5 lines
Quoted 1 line
It should be just a single "=" in the line above.
AppArmor provides the 2nd "=" as part of the subject context. What's here is correct. I won't argue that it won't case confusion or worse.
We have a specification for a very long time:
https://github.com/linux-audit/audit-documentation/wiki/SPEC-Writing-Good-Events
Everyone has to obey rules or no tools work.
-Steve
--
Linux-audit mailing list
[email protected]
https://listman.redhat.com/mailman/listinfo/linux-audit
Casey Schaufler
maintainerRe: [PATCH v26 22/25] Audit: Add new record for multiple process LSM attributes
Quoted 12 lines · Steve Grubb
Quoted 11 lines · EDT Casey Schaufler
Quoted 11 lines
Quoted 10 lines
The record is produced only in cases where there is more than one security module with a process "context".
In cases where this record is produced the subj= fields of other records in the audit event will be set to "subj=?".An example of the
MAC_TASK_CONTEXTS(1420) record is:type=UNKNOWN[1420]
msg=audit(1600880931.832:113)
subj_apparmor==unconfinedIt should be just a single "=" in the line above.
Quoted 8 lines
Quoted 3 lines
AppArmor provides the 2nd "=" as part of the subject context. What's here is correct. I won't argue that it won't case confusion or worse.
We have a specification for a very long time:
https://github.com/linux-audit/audit-documentation/wiki/SPEC-Writing-Good-Events
Everyone has to obey rules or no tools work.
Right. Would quoting the value subj_apparmor="=unconfined" suffice? I think that's what the rulebook says, but I like to be sure.
--
Linux-audit mailing list
[email protected]
https://listman.redhat.com/mailman/listinfo/linux-audit
Re: [PATCH v26 22/25] Audit: Add new record for multiple process LSM attributes
Hello Casey,
Quoted 19 lines · EDT Casey Schaufler
Quoted 18 lines · Steve Grubb
Quoted 16 lines · EDT Casey Schaufler
Quoted 5 lines
Quoted 5 lines
The record is produced only in cases where there is more than one security module with a process "context".
In cases where this record is produced the subj= fields of other records in the audit event will be set to "subj=?".Quoted 7 lines
Quoted 5 lines
An example of the
MAC_TASK_CONTEXTS(1420) record is:type=UNKNOWN[1420]
msg=audit(1600880931.832:113)
subj_apparmor==unconfinedIt should be just a single "=" in the line above.
AppArmor provides the 2nd "=" as part of the subject context. What's here is correct. I won't argue that it won't case confusion or worse.
Quoted 8 lines
Quoted 6 lines
We have a specification for a very long time:
https://github.com/linux-audit/audit-documentation/wiki/SPEC-Writing-Good
-EventsEveryone has to obey rules or no tools work.
Right. Would quoting the value subj_apparmor="=unconfined" suffice?
Yes, if you really need that '=' as part of the value. I'd try to do away with it entirely if possible. Generally '==' is used to convey "equal to". Where '=' means "is assigned". I'd argue that they are interchangable in meaning and can simply be converted to the log's expected format.
Quoted 1 line
I think that's what the rulebook says, but I like to be sure.
:-)
-Steve
--
Linux-audit mailing list
[email protected]
https://listman.redhat.com/mailman/listinfo/linux-audit
Casey Schaufler
maintainerRe: [PATCH v26 22/25] Audit: Add new record for multiple process LSM attributes
In reply to Paul MooreQuoted 418 lines · Paul Moore
Quoted 4 lines · Casey Schaufler
Create a new audit record type to contain the subject information when there are multiple security modules that require such data. This record is linked with the same timestamp and serial number using the
audit_alloc_local()mechanism.The record is linked with the other associated records into a single event, it doesn't matter if it gets the timestamp/serial from
audit_alloc_local()or an existing audit event, e.g. ongoing syscall.Quoted 10 lines
The record is produced only in cases where there is more than one security module with a process "context".
In cases where this record is produced the subj= fields of other records in the audit event will be set to "subj=?".An example of the
MAC_TASK_CONTEXTS(1420) record is:type=UNKNOWN[1420] msg=audit(1600880931.832:113) subj_apparmor==unconfinedIt should be just a single "=" in the line above.
Quoted 40 lines
subj_smack=_
There will be a subj_$LSM= entry for each security module LSM that supports the
secid_to_secctxandsecctx_to_secidhooks. The BPF security module implements secid/secctx translation hooks, so it has to be considered to provide a secctx even though it may not actually do so.Signed-off-by Casey Schaufler <[email protected]>To: [email protected]
To: [email protected]
To: [email protected]
Cc: [email protected]--- drivers/android/binder.c | 2 +- include/linux/audit.h | 24 ++++++++ include/linux/security.h | 16 ++++- include/net/netlabel.h | 3 +- include/net/scm.h | 2 +- include/net/xfrm.h | 13 +++- include/uapi/linux/audit.h | 1 + kernel/audit.c | 80 ++++++++++++++++++------- kernel/audit.h | 3 + kernel/auditfilter.c | 6 +- kernel/auditsc.c | 75 ++++++++++++++++++++--- net/ipv4/ip_sockglue.c | 2 +- net/netfilter/nf_conntrack_netlink.c | 4 +- net/netfilter/nf_conntrack_standalone.c | 2 +- net/netfilter/nfnetlink_queue.c | 2 +- net/netlabel/netlabel_domainhash.c | 4 +- net/netlabel/netlabel_unlabeled.c | 24 ++++---- net/netlabel/netlabel_user.c | 20 ++++--- net/netlabel/netlabel_user.h | 6 +- net/xfrm/xfrm_policy.c | 10 ++-- net/xfrm/xfrm_state.c | 20 ++++--- security/integrity/ima/ima_api.c | 7 ++- security/integrity/integrity_audit.c | 6 +- security/security.c | 46 +++++++++----- security/smack/smackfs.c | 3 +- 25 files changed, 274 insertions(+), 107 deletions(-)...
Quoted 10 lines
include/linux/audit.h +2
-0@@ -386,6 +395,19 @@ static inline void audit_ptrace(struct task_struct *t)__audit_ptrace (t); } +static inline struct audit_context *audit_alloc_for_lsm (gfp_t gfp) +{include/linux/audit.h static inline void audit_ptrace(struct task_struct *t) · line 409
Quoted 10 lines
include/linux/audit.h +10
-0+struct audit_context *context =audit_context (); + +if (context) +return context; + +if (lsm_multiple_contexts ()) +return audit_alloc_local (gfp); + +return NULL ; +}See my other comments, but this seems wrong at face value. The additional LSM record should happen as part of the existing audit log functions.
Quoted 89 lines
include/linux/security.h +14
-2@@ -182,6 +182,8 @@ struct lsmblob {#define LSMBLOB_INVALID -1/* Not a valid LSM slot number */ #define LSMBLOB_NEEDED -2/* Slot requested on initialization */ #define LSMBLOB_NOT_NEEDED -3/* Slot not requested */ +#define LSMBLOB_DISPLAY -4/* Use the "display" slot */ +#define LSMBLOB_FIRST -5/* Use the default "display" slot */ /** * lsmblob_init - initialize an lsmblob structure @@ -248,6 +250,15 @@ static inline u32 lsmblob_value(const struct lsmblob *blob)return 0 ; } +static inline bool lsm_multiple_contexts (void ) +{ +#ifdef CONFIG_SECURITY +return lsm_slot_to_name (1 ) !=NULL ; +#else +return false; +#endif +} +/* These functions are in security/commoncap.c */ extern int cap_capable (const struct cred *cred,struct user_namespace *ns,int cap,unsigned int opts);@@ -578,7 +589,8 @@ int security_setprocattr(const char *lsm, const char *name, void *value, size_t size);int security_netlink_send (struct sock *sk,struct sk_buff *skb);int security_ismaclabel (const char *name); -int security_secid_to_secctx (struct lsmblob *blob,struct lsmcontext *cp); +int security_secid_to_secctx (struct lsmblob *blob,struct lsmcontext *cp, +int display);int security_secctx_to_secid (const char *secdata,u32 seclen,struct lsmblob *blob);void security_release_secctx (struct lsmcontext *cp);@@ -1433,7 +1445,7 @@ static inline int security_ismaclabel(const char *name) }static inline int security_secid_to_secctx (struct lsmblob *blob, -struct lsmcontext *cp) +struct lsmcontext *cp,int display) {return -EOPNOTSUPP ; }include/net/netlabel.h +2
-1@@ -97,7 +97,8 @@ struct calipso_doi;/* NetLabel audit information */ struct netlbl_audit { -u32 secid; +struct audit_context *localcontext; +struct lsmblob lsmdata;kuid_t loginuid;unsigned int sessionid; };include/net/scm.h +1
-1@@ -101,7 +101,7 @@ static inline void scm_passec(struct socket *sock, struct msghdr *msg, struct sc* and the infrastructure will know which it is. */ lsmblob_init (&lb, scm->secid); - err =security_secid_to_secctx (&lb, &context); + err =security_secid_to_secctx (&lb, &context,LSMBLOB_DISPLAY );if (!err) { put_cmsg(msg, SOL_SOCKET, SCM_SECURITY, context.len,include/net/xfrm.h +3
-1@@ -669,13 +669,22 @@ struct xfrm_spi_skb_cb {#define XFRM_SPI_SKB_CB (__skb) ((struct xfrm_spi_skb_cb *)&((__skb)->cb[0]))#ifdef CONFIG_AUDITSYSCALL -static inline struct audit_buffer *xfrm_audit_start (const char *op) +static inline struct audit_buffer *xfrm_audit_start (const char *op, +struct audit_context **lac) { +struct audit_context *context;struct audit_buffer *audit_buf =NULL ;include/net/xfrm.h struct xfrm_spi_skb_cb { · line 685
Quoted 10 lines
include/net/xfrm.h +6
-1if (audit_enabled ==AUDIT_OFF )return NULL ; - audit_buf = audit_log_start(audit_context(), GFP_ATOMIC, + context =audit_context (); +if (lac !=NULL ) { +if (lsm_multiple_contexts () && context ==NULL ) + context =audit_alloc_local (GFP_ATOMIC ); + *lac = context; + }Okay, we've got a disconnect here regarding "audit contexts" and "local contexts", skip down below where I attempt to explain things a little more but basically if there is a place that uses this pattern:
audit_log_start(audit_context(), ...);... you don't need, or want, a "local context". You might need a local context if you see the following pattern:
audit_log_start(NULL, ...);The "local context" idea is a hack and should be avoided whenever possible; if you have an existing audit context from a syscall, or something else, you *really* should use it ... or have a *really* good explanation as to why you can not.
Quoted 9 lines
include/net/xfrm.h +1
-0+ audit_buf =audit_log_start (context,GFP_ATOMIC ,AUDIT_MAC_IPSEC_EVENT );if (audit_buf ==NULL )return NULL ;kernel/audit.c +0
-0kernel/audit.c void audit_log_lost(const char *message) · line 394
Quoted 10 lines
kernel/audit.c +3
-1@@ -386,10 +386,12 @@ void audit_log_lost(const char *message)static int audit_log_config_change (char *function_name,u32 new,u32 old,int allow_changes) { +struct audit_context *context;struct audit_buffer *ab;int rc =0 ; - ab =audit_log_start (audit_context (),GFP_KERNEL ,AUDIT_CONFIG_CHANGE ); + context =audit_alloc_for_lsm (GFP_KERNEL ); + ab =audit_log_start (context,GFP_KERNEL ,AUDIT_CONFIG_CHANGE );Use the existing context, don't create your own, it breaks the record associations in the audit event stream.
kernel/audit.c static int audit_log_config_change(char *function_name, u32 new, u32 old, · line 403
Quoted 9 lines
kernel/audit.c +1
-1if (unlikely (!ab))return rc;audit_log_format (ab,"op=set %s=%u old=%u " , function_name, new, old);@@ -398,7 +400,7 @@ static int audit_log_config_change(char *function_name, u32 new, u32 old,if (rc) allow_changes =0 ;/* Something weird, deny request */ audit_log_format (ab," res=%d" , allow_changes); -audit_log_end (ab); +audit_log_end_local (ab, context);More on this below, but we really should just use
audit_log_end(), "local contexts" are not special, the are regular audit contexts ... although if they are used properly (limited scope) you do need to free them when you are done.kernel/audit.c static int audit_receive_msg(struct sk_buff *skb, struct nlmsghdr *nlh) · line 1359
Quoted 10 lines
kernel/audit.c +2
-1return rc; }@@ -1357,7 +1355,8 @@ static int audit_receive_msg(struct sk_buff *skb, struct nlmsghdr *nlh)if (err)break ; } -audit_log_user_recv_msg (&ab, msg_type); + lcontext =audit_alloc_for_lsm (GFP_KERNEL ); +audit_log_common_recv_msg (lcontext, &ab, msg_type);Same.
kernel/audit.c static int audit_receive_msg(struct sk_buff *skb, struct nlmsghdr *nlh) · line 1372
Quoted 9 lines
kernel/audit.c +1
-1if (msg_type !=AUDIT_USER_TTY ) {/* ensure NULL termination */ str[data_len -1 ] ='\0' ;@@ -1370,7 +1369,7 @@ static int audit_receive_msg(struct sk_buff *skb, struct nlmsghdr *nlh) data_len--;audit_log_n_untrustedstring (ab, str, data_len); } -audit_log_end (ab); +audit_log_end_local (ab, lcontext);Same.
Quoted 7 lines
kernel/audit.c +0
-0}break ;case AUDIT_ADD_RULE :@@ -1378,13 +1377,14 @@ static int audit_receive_msg(struct sk_buff *skb, struct nlmsghdr *nlh)if (data_len <sizeof (struct audit_rule_data ))return -EINVAL ;if (audit_enabled ==AUDIT_LOCKED ) {kernel/audit.c static int audit_receive_msg(struct sk_buff *skb, struct nlmsghdr *nlh) · line 1387
Quoted 10 lines
kernel/audit.c +3
-2-audit_log_common_recv_msg (audit_context (), &ab, + lcontext =audit_alloc_for_lsm (GFP_KERNEL ); +audit_log_common_recv_msg (lcontext, &ab,AUDIT_CONFIG_CHANGE );audit_log_format (ab," op=%s audit_enabled=%d res=0" , msg_type ==AUDIT_ADD_RULE ?"add_rule" :"remove_rule" , audit_enabled); -audit_log_end (ab); +audit_log_end_local (ab, lcontext);Same. I'm going to stop calling these out, I think you get the idea.
Quoted 8 lines
kernel/audit.c +4
-0@@ -2396,6 +2415,21 @@ void audit_log_end(struct audit_buffer *ab)audit_buffer_free (ab); } +/** +* audit_log_end_local - end one audit record with local context +* @ab: the audit_buffer +* @context: the local context kernel/audit.c void audit_log_end(struct audit_buffer *ab) · line 2431
Quoted 10 lines
kernel/audit.c +10
-0+* +* Emit an LSM context record if appropriate, then end the audit event +* in the usual way. +*/ +void audit_log_end_local (struct audit_buffer *ab,struct audit_context *context) +{ +audit_log_end (ab); +audit_log_lsm_common (context); +audit_free_local (context); +}Eeesh, no, not this please.
First, some background on audit contexts and the idea of a "local context" as we have been using it in the audit container ID work, which is where this originated. An audit context contains a few things, but likely the most important for this discussion is the audit event timestamp and serial number (I may refer to this combo as just a "timestamp" in the future); this timestamp/serial is shared across all of the audit records that make up this audit event, linking them together. A shared timestamp is what allows you to group an
open()SYSCALL record with the PATH record that provides the file's pathname info.While there are some exceptions in the current code, most audit events occur as a result of a syscall, and their audit context in this case is the syscall's audit context (see the
open()example above), but there are some cases being discussed where we have an audit event that does not occur as a result of a syscall but there is a need to group multiple audit records together in a single event. This is where the "local context" comes into play, it allows us to create an audit context outside of a syscall and share that context across multiple audit records, allowing the records to share a timestamp/serial and grouping them together as a single audit event in the audit stream.While a function like
audit_alloc_local()make sense, there really shouldn't be anaudit_log_end_local()function, the normalaudit_log_end()function should be used.Does that make sense?
kernel/audit.h struct audit_context { · line 103
Quoted 9 lines
kernel/audit.h +1
-0@@ -100,6 +100,7 @@ struct audit_context {int dummy;/* must be the first element */ int in_syscall;/* 1 if task is in a syscall */ bool local;/* local context needed */ +bool lsmdone;/* multiple security reported */ "lsmdone" doesn't seem consistent with the comment, how about "
lsm_multi" or something similar?Quoted 8 lines
kernel/audit.h +0
-0enum audit_state state, current_state;unsigned int serial;/* serial number for record */ int major;/* syscall number */ kernel/auditsc.c +0
-0kernel/auditsc.c void audit_free_context(struct audit_context *context) · line 1021
Quoted 9 lines
kernel/auditsc.c +6
-0@@ -1013,6 +1013,13 @@ void audit_free_context(struct audit_context *context) }EXPORT_SYMBOL (audit_free_context); +void audit_free_local (struct audit_context *context) +{ +if (context && context->local) +audit_free_context (context); +} +EXPORT_SYMBOL (audit_free_local);We don't need this function, just use
audit_free_context(). A "local context" is the same as a non-local context; what makes a context "local" is the scope of the audit context (local function scope vs syscall scope) and nothing else.Quoted 9 lines
kernel/auditsc.c +5
-0@@ -1504,6 +1512,47 @@ static void audit_log_proctitle(void)audit_log_end (ab); } +void audit_log_lsm_common (struct audit_context *context) +{ +struct audit_buffer *ab; +struct lsmcontext lsmdata; +bool sep = false;kernel/auditsc.c static void audit_log_proctitle(void) · line 1529
Quoted 10 lines
kernel/auditsc.c +10
-0+int error; +int i; + +if (!lsm_multiple_contexts () || context ==NULL || + !lsmblob_is_set (&context->lsm)) +return ; + + ab =audit_log_start (context,GFP_ATOMIC ,AUDIT_MAC_TASK_CONTEXTS ); +if (!ab) +return ;/* audit_panic or being filtered */ We should be consistent with our use of
audit_panic()when we bail on error; we use it below, but not here - why?Quoted 6 lines
kernel/auditsc.c +6
-0+for (i =0 ; i <LSMBLOB_ENTRIES ; i++) { +if (context->lsm.secid[i] ==0 ) +continue ; + error =security_secid_to_secctx (&context->lsm, &lsmdata, i); +if (error && error != -EINVAL ) { +audit_panic ("error in audit_log_lsm" );kernel/auditsc.c static void audit_log_proctitle(void) · line 1547
Quoted 10 lines
kernel/auditsc.c +10
-0+return ; + } + +audit_log_format (ab,"%ssubj_%s=%s" , sep ?" " :"" , +lsm_slot_to_name (i), lsmdata.context); + sep = true; +security_release_secctx (&lsmdata); + } +audit_log_end (ab); + context->lsmdone = true;Maybe I missed it, but why do we need this flag?
kernel/auditsc.c static void audit_log_proctitle(void) · line 1554
Quoted 7 lines
kernel/auditsc.c +7
-0+} + +void audit_log_lsm (struct audit_context *context) +{ +if (!context->lsmdone) +audit_log_lsm_common (context); +}I think I was distracted with the local context issue and I've lost track of the details here, perhaps it's best to fix the local context issue first (that should be a big change to this patch) and then we can take another look.
--
paul moore
www.paul-moore.com
Casey Schaufler
maintainerRe: [PATCH v26 22/25] Audit: Add new record for multiple process LSM attributes
In reply to Paul MooreOn 5/21/2021 1:19 PM, Paul Moore wrote:
<snip> and trim the CC list.
Quoted 5 lines
Okay, we've got a disconnect here regarding "audit contexts" and "local contexts", skip down below where I attempt to explain things a little more but basically if there is a place that uses this pattern:
audit_log_start(audit_context(), ...);
This pattern isn't helpful. audit_context() returns NULL in about 1 of 4 calls. All of these callers of audit_context() get a NULL result some of the time.
getname_kerneldebugfs_create_dirtracefs_create_filevfs_fchowndo_settimeofday64bprm_execveksys_mmap_pgoffmove_addr_to_kernel__do_pipe_flags__do_sys_capsetsyscall_trace_entercap_bprm_creds_from_fileload_module__x64_sys_fsetxattrbpf_prog_loadaudit_signal_info_syscallsel_write_enforcecommon_lsm_auditaudit_set_loginuid__dev_set_promiscuity
ipcpermsdevpts_pty_new
Quoted 9 lines
... you don't need, or want, a "local context". You might need a local context if you see the following pattern:
audit_log_start(NULL, ...);The "local context" idea is a hack and should be avoided whenever possible; if you have an existing audit context from a syscall, or something else, you *really* should use it ... or have a *really* good explanation as to why you can not.
Almost all audit events want to record the subject context by calling audit_log_task_context(). If multiple contexts need to be recorded there has to be a separate record, which means there has to be an audit context. The only case where an audit context is reliably available is in syscalls. Hence, a "local context" for many of the cases where the first argument to audit_log_start() would otherwise be NULL, either explicitly or because audit_context() returns NULL.
Is there some other way to synchronize the "timestamp" without use of an audit context? My understanding is that this is the primary purpose of the audit context.
Re: [PATCH v26 22/25] Audit: Add new record for multiple process LSM attributes
Quoted 11 lines · Casey Schaufler
On 5/21/2021 1:19 PM, Paul Moore wrote:
<snip> and trim the CC list.
Quoted 5 lines
Okay, we've got a disconnect here regarding "audit contexts" and "local contexts", skip down below where I attempt to explain things a little more but basically if there is a place that uses this pattern:
audit_log_start(audit_context(), ...);This pattern isn't helpful.
audit_context()returns NULL in about 1 of 4 calls.
Ok, this rings a bell... I think we talked about this on an earlier revision...
Quoted 35 lines
All of these callers of
audit_context()get a NULL result some of the time.
getname_kerneldebugfs_create_dirtracefs_create_filevfs_fchowndo_settimeofday64bprm_execveksys_mmap_pgoffmove_addr_to_kernel__do_pipe_flags__do_sys_capsetsyscall_trace_entercap_bprm_creds_from_fileload_module__x64_sys_fsetxattrbpf_prog_loadaudit_signal_info_syscallsel_write_enforcecommon_lsm_auditaudit_set_loginuid__dev_set_promiscuity
ipcpermsdevpts_pty_newQuoted 9 lines
... you don't need, or want, a "local context". You might need a local context if you see the following pattern:
audit_log_start(NULL, ...);The "local context" idea is a hack and should be avoided whenever possible; if you have an existing audit context from a syscall, or something else, you *really* should use it ... or have a *really* good explanation as to why you can not.
Quoted 7 lines
Almost all audit events want to record the subject context by calling
audit_log_task_context(). If multiple contexts need to be recorded there has to be a separate record, which means there has to be an audit context. The only case where an audit context is reliably available is in syscalls. Hence, a "local context" for many of the cases where the first argument toaudit_log_start()would otherwise be NULL, either explicitly or becauseaudit_context()returns NULL.
Ok, so in that case, I think I'd test audit_context() and if it is indeed NULL then create a local context, otherwise use the one that is available. I should really go back and look carefully again at your code to see if it is in fact doing this, but unilaterally creating a local context if a context already exists is going to cause confusion because there will be two events generated for one event.
Is it possible these functions are not actually generating records if the context is NULL?
I'd be digging to find out why the context is NULL in these cases and if any record is even being produced in those cases? Perhaps there is an active filter that indicates that logging is not of interest for that process/task/file/syscall/perm/etc...
Quoted 3 lines
Is there some other way to synchronize the "timestamp" without use of an audit context? My understanding is that this is the primary purpose of the audit context.
What has been done is to call it with a NULL context and it would assign a timestamp and serial number, but those are all single records that don't need synchronization (obviously). This was why I'd come up with this mechanism to solve the need to attach a contid record to a single record associated with a network event (or user record that should not be associated with a syscall). Those were the only two use cases I had up until now.
- RGB
--
Richard Guy Briggs <[email protected]>
Sr. S/W Engineer, Kernel Security, Base Operating Systems Remote, Ottawa, Red Hat Canada
IRC: rgb, SunRaycer
Voice: +1.647.777.2635, Internal: (81) 32635
Casey Schaufler
maintainerRe: [PATCH v26 22/25] Audit: Add new record for multiple process LSM attributes
Quoted 55 lines · Richard Guy Briggs
Quoted 10 lines · Casey Schaufler
On 5/21/2021 1:19 PM, Paul Moore wrote:
<snip> and trim the CC list.
Quoted 5 lines
Okay, we've got a disconnect here regarding "audit contexts" and "local contexts", skip down below where I attempt to explain things a little more but basically if there is a place that uses this pattern:
audit_log_start(audit_context(), ...);This pattern isn't helpful.
audit_context()returns NULL in about 1 of 4 calls.Ok, this rings a bell... I think we talked about this on an earlier revision...
Quoted 34 lines
All of these callers of
audit_context()get a NULL result some of the time.
getname_kerneldebugfs_create_dirtracefs_create_filevfs_fchowndo_settimeofday64bprm_execveksys_mmap_pgoffmove_addr_to_kernel__do_pipe_flags__do_sys_capsetsyscall_trace_entercap_bprm_creds_from_fileload_module__x64_sys_fsetxattrbpf_prog_loadaudit_signal_info_syscallsel_write_enforcecommon_lsm_auditaudit_set_loginuid__dev_set_promiscuity
ipcpermsdevpts_pty_newQuoted 9 lines
... you don't need, or want, a "local context". You might need a local context if you see the following pattern:
audit_log_start(NULL, ...);The "local context" idea is a hack and should be avoided whenever possible; if you have an existing audit context from a syscall, or something else, you *really* should use it ... or have a *really* good explanation as to why you can not.
Quoted 7 lines
Almost all audit events want to record the subject context by calling
audit_log_task_context(). If multiple contexts need to be recorded there has to be a separate record, which means there has to be an audit context. The only case where an audit context is reliably available is in syscalls. Hence, a "local context" for many of the cases where the first argument toaudit_log_start()would otherwise be NULL, either explicitly or becauseaudit_context()returns NULL.
Quoted 3 lines
Ok, so in that case, I think I'd test
audit_context()and if it is indeed NULL then create a local context, otherwise use the one that is available.
audit_alloc_for_lsm() returns the value of audit_context() if it is not NULL. Otherwise it checks to see if a separate record will be required. If it is the value from audit_alloc_local() is returned. Otherwise, it returns NULL.
Quoted 4 lines
I should really go back and look carefully again at your
code to see if it is in fact doing this, but unilaterally creating a local context if a context already exists is going to cause confusion because there will be two events generated for one event.
Indeed. I had discovered that.
Quoted 2 lines
Is it possible these functions are not actually generating records if the context is NULL?
There are definitely cases where records are generated when audit_context() is NULL.
Quoted 2 lines
I'd be digging to find out why the context is NULL in these cases and if any record is even being produced in those cases?
Context is NULL because they're not coming out of syscalls.
Quoted 7 lines
Perhaps there is an
active filter that indicates that logging is not of interest for that process/task/file/syscall/perm/etc...Quoted 3 lines
Is there some other way to synchronize the "timestamp" without use of an audit context? My understanding is that this is the primary purpose of the audit context.
Quoted 7 lines
What has been done is to call it with a NULL context and it would assign a timestamp and serial number, but those are all single records that don't need synchronization (obviously). This was why I'd come up with this mechanism to solve the need to attach a contid record to a single record associated with a network event (or user record that should not be associated with a syscall). Those were the only two use cases I had up until now.
Right. Adding the contid record is a special case. Adding the lsmcontext record is the common case. Thus Paul's lament.
Quoted 11 lines
- RGB
--
Richard Guy Briggs <[email protected]>
Sr. S/W Engineer, Kernel Security, Base Operating Systems Remote, Ottawa, Red Hat Canada
IRC: rgb, SunRaycer
Voice: +1.647.777.2635, Internal: (81) 32635
Re: [PATCH v26 22/25] Audit: Add new record for multiple process LSM attributes
Quoted 75 lines · Casey Schaufler
Quoted 55 lines · Richard Guy Briggs
Quoted 10 lines · Casey Schaufler
On 5/21/2021 1:19 PM, Paul Moore wrote:
<snip> and trim the CC list.
Quoted 5 lines
Okay, we've got a disconnect here regarding "audit contexts" and "local contexts", skip down below where I attempt to explain things a little more but basically if there is a place that uses this pattern:
audit_log_start(audit_context(), ...);This pattern isn't helpful.
audit_context()returns NULL in about 1 of 4 calls.Ok, this rings a bell... I think we talked about this on an earlier revision...
Quoted 34 lines
All of these callers of
audit_context()get a NULL result some of the time.
getname_kerneldebugfs_create_dirtracefs_create_filevfs_fchowndo_settimeofday64bprm_execveksys_mmap_pgoffmove_addr_to_kernel__do_pipe_flags__do_sys_capsetsyscall_trace_entercap_bprm_creds_from_fileload_module__x64_sys_fsetxattrbpf_prog_loadaudit_signal_info_syscallsel_write_enforcecommon_lsm_auditaudit_set_loginuid__dev_set_promiscuity
ipcpermsdevpts_pty_newQuoted 9 lines
... you don't need, or want, a "local context". You might need a local context if you see the following pattern:
audit_log_start(NULL, ...);The "local context" idea is a hack and should be avoided whenever possible; if you have an existing audit context from a syscall, or something else, you *really* should use it ... or have a *really* good explanation as to why you can not.
Quoted 7 lines
Almost all audit events want to record the subject context by calling
audit_log_task_context(). If multiple contexts need to be recorded there has to be a separate record, which means there has to be an audit context. The only case where an audit context is reliably available is in syscalls. Hence, a "local context" for many of the cases where the first argument toaudit_log_start()would otherwise be NULL, either explicitly or becauseaudit_context()returns NULL.Quoted 3 lines
Ok, so in that case, I think I'd test
audit_context()and if it is indeed NULL then create a local context, otherwise use the one that is available.
audit_alloc_for_lsm()returns the value ofaudit_context()if it is not NULL. Otherwise it checks to see if a separate record will be required. If it is the value fromaudit_alloc_local()is returned. Otherwise, it returns NULL.Quoted 4 lines
I should really go back and look carefully again at your
code to see if it is in fact doing this, but unilaterally creating a local context if a context already exists is going to cause confusion because there will be two events generated for one event.Indeed. I had discovered that.
Quoted 2 lines
Is it possible these functions are not actually generating records if the context is NULL?
Quoted 7 lines
There are definitely cases where records are generated when
audit_context()is NULL.Quoted 2 lines
I'd be digging to find out why the context is NULL in these cases and if any record is even being produced in those cases?
Context is NULL because they're not coming out of syscalls.
Where are they coming from then? I'm guessing that they are in fact coming from syscalls, but that it wasn't a syscall rule that triggered the need to record the event.
Quoted 15 lines
Quoted 7 lines
Perhaps there is an
active filter that indicates that logging is not of interest for that process/task/file/syscall/perm/etc...Quoted 3 lines
Is there some other way to synchronize the "timestamp" without use of an audit context? My understanding is that this is the primary purpose of the audit context.
Quoted 7 lines
What has been done is to call it with a NULL context and it would assign a timestamp and serial number, but those are all single records that don't need synchronization (obviously). This was why I'd come up with this mechanism to solve the need to attach a contid record to a single record associated with a network event (or user record that should not be associated with a syscall). Those were the only two use cases I had up until now.
Quoted 2 lines
Right. Adding the contid record is a special case. Adding the lsmcontext record is the common case. Thus Paul's lament.
Yeah, this is a similar sort of accompanying record that needs to be tied into an event which is why I had suggested the similarity behind your local context generation patch and the one in the contid patchset.
Quoted 1 line
Quoted 1 line
- RGB
- RGB
--
Richard Guy Briggs <[email protected]>
Sr. S/W Engineer, Kernel Security, Base Operating Systems Remote, Ottawa, Red Hat Canada
IRC: rgb, SunRaycer
Voice: +1.647.777.2635, Internal: (81) 32635
Casey Schaufler
maintainerRe: [PATCH v26 22/25] Audit: Add new record for multiple process LSM attributes
Quoted 76 lines · Richard Guy Briggs
Quoted 75 lines · Casey Schaufler
Quoted 55 lines · Richard Guy Briggs
Quoted 10 lines · Casey Schaufler
On 5/21/2021 1:19 PM, Paul Moore wrote:
<snip> and trim the CC list.
Quoted 5 lines
Okay, we've got a disconnect here regarding "audit contexts" and "local contexts", skip down below where I attempt to explain things a little more but basically if there is a place that uses this pattern:
audit_log_start(audit_context(), ...);This pattern isn't helpful.
audit_context()returns NULL in about 1 of 4 calls.Ok, this rings a bell... I think we talked about this on an earlier revision...
Quoted 34 lines
All of these callers of
audit_context()get a NULL result some of the time.
getname_kerneldebugfs_create_dirtracefs_create_filevfs_fchowndo_settimeofday64bprm_execveksys_mmap_pgoffmove_addr_to_kernel__do_pipe_flags__do_sys_capsetsyscall_trace_entercap_bprm_creds_from_fileload_module__x64_sys_fsetxattrbpf_prog_loadaudit_signal_info_syscallsel_write_enforcecommon_lsm_auditaudit_set_loginuid__dev_set_promiscuity
ipcpermsdevpts_pty_newQuoted 9 lines
... you don't need, or want, a "local context". You might need a local context if you see the following pattern:
audit_log_start(NULL, ...);The "local context" idea is a hack and should be avoided whenever possible; if you have an existing audit context from a syscall, or something else, you *really* should use it ... or have a *really* good explanation as to why you can not.
Quoted 7 lines
Almost all audit events want to record the subject context by calling
audit_log_task_context(). If multiple contexts need to be recorded there has to be a separate record, which means there has to be an audit context. The only case where an audit context is reliably available is in syscalls. Hence, a "local context" for many of the cases where the first argument toaudit_log_start()would otherwise be NULL, either explicitly or becauseaudit_context()returns NULL.Quoted 3 lines
Ok, so in that case, I think I'd test
audit_context()and if it is indeed NULL then create a local context, otherwise use the one that is available.
audit_alloc_for_lsm()returns the value ofaudit_context()if it is not NULL. Otherwise it checks to see if a separate record will be required. If it is the value fromaudit_alloc_local()is returned. Otherwise, it returns NULL.Quoted 4 lines
I should really go back and look carefully again at your
code to see if it is in fact doing this, but unilaterally creating a local context if a context already exists is going to cause confusion because there will be two events generated for one event.Indeed. I had discovered that.
Quoted 2 lines
Is it possible these functions are not actually generating records if the context is NULL?
There are definitely cases where records are generated when
audit_context()is NULL.
Quoted 6 lines
Quoted 3 lines
Quoted 2 lines
I'd be digging to find out why the context is NULL in these cases and if any record is even being produced in those cases?
Context is NULL because they're not coming out of syscalls.
Where are they coming from then? I'm guessing that they are in fact coming from syscalls, but that it wasn't a syscall rule that triggered the need to record the event.
audit_receive_msg() is one place. If you look at my patch you'll see others, I only put audit_alloc_for_lsm() calls in where they were needed. There are plenty of places.
Quoted 15 lines
Quoted 15 lines
Quoted 15 lines
Perhaps there is an
active filter that indicates that logging is not of interest for that process/task/file/syscall/perm/etc...Quoted 3 lines
Is there some other way to synchronize the "timestamp" without use of an audit context? My understanding is that this is the primary purpose of the audit context.
What has been done is to call it with a NULL context and it would assign a timestamp and serial number, but those are all single records that don't need synchronization (obviously). This was why I'd come up with this mechanism to solve the need to attach a contid record to a single record associated with a network event (or user record that should not be associated with a syscall). Those were the only two use cases I had up until now.
Quoted 5 lines
Quoted 2 lines
Right. Adding the contid record is a special case. Adding the lsmcontext record is the common case. Thus Paul's lament.
Yeah, this is a similar sort of accompanying record that needs to be tied into an event which is why I had suggested the similarity behind your local context generation patch and the one in the contid patchset.
Just so. I think the $2 point is that the audit system seems oriented to have multiple records per event be unusual. I need to make it normal.
Quoted 11 lines
Quoted 2 lines
Quoted 2 lines
- RGB
- RGB
--
Richard Guy Briggs <[email protected]>
Sr. S/W Engineer, Kernel Security, Base Operating Systems Remote, Ottawa, Red Hat Canada
IRC: rgb, SunRaycer
Voice: +1.647.777.2635, Internal: (81) 32635
Casey Schaufler
maintainerPATCH v26 23/25 Audit: Add a new record for multiple object LSM attributes
In reply to Casey SchauflerCreate a new audit record type to contain the object information when there are multiple security modules that may require such data. This record is linked with the same timestamp and serial number.
An example of the MAC_OBJ_CONTEXTS (1421) record is:
type=UNKNOWN[1421]
msg=audit(1601152467.009:1050):
obj_selinux=unconfined_u:object_r:user_home_t:s0Not all security modules that can provide object information do so in all cases. It is possible that a security module won't apply an object attribute in all cases.
To: Paul Moore <[email protected]>
--- include/linux/audit.h | 7 ++++ include/uapi/linux/audit.h | 1 + kernel/audit.c | 53 +++++++++++++++++++++++++++ kernel/audit.h | 4 +-- kernel/auditsc.c | 73 +++++++------------------------------- 5 files changed, 75 insertions(+), 63 deletions(-)
include/linux/audit.h +7 -0
@@ -189,6 +189,8 @@ extern void audit_log_path_denied(int type,extern void audit_log_lost (const char *message);extern int audit_log_task_context (struct audit_buffer *ab); +extern int audit_log_object_context (struct audit_buffer *ab, +struct lsmblob *blob);extern void audit_log_task_info (struct audit_buffer *ab);extern void audit_log_lsm (struct audit_context *context);
@@ -256,6 +258,11 @@ static inline int audit_log_task_context(struct audit_buffer *ab) {return 0 ; } +static inline int audit_log_object_context (struct audit_buffer *ab, +struct lsmblob *blob) +{ +return 0 ; +}static inline void audit_log_task_info (struct audit_buffer *ab) { }static void audit_log_lsm (struct audit_context *context)
include/uapi/linux/audit.h +1 -0
@@ -140,6 +140,7 @@#define AUDIT_MAC_CALIPSO_ADD 1418/* NetLabel: add CALIPSO DOI entry */ #define AUDIT_MAC_CALIPSO_DEL 1419/* NetLabel: del CALIPSO DOI entry */ #define AUDIT_MAC_TASK_CONTEXTS 1420/* Multiple LSM contexts */ +#define AUDIT_MAC_OBJ_CONTEXTS 1421/* Multiple LSM object contexts */ #define AUDIT_FIRST_KERN_ANOM_MSG 1700#define AUDIT_LAST_KERN_ANOM_MSG 1799
kernel/audit.c +53 -0
59 lines not on this page · raw message
kernel/audit.h +2 -2
@@ -78,7 +78,7 @@ struct audit_names {kuid_t uid;kgid_t gid;dev_t rdev; -u32 osid; +struct lsmblob oblob;struct audit_cap_data fcap;unsigned int fcap_ver;unsigned char type;/* record type */
@@ -155,7 +155,7 @@ struct audit_context {kuid_t uid;kgid_t gid;umode_t mode; -u32 osid; +struct lsmblob oblob;int has_perm;uid_t perm_uid;gid_t perm_gid;
kernel/auditsc.c +12 -61
143 lines not on this page · raw message
Signature · 1 line
2.29.2
Casey Schaufler
maintainerPATCH v26 24/25 LSM: Add /proc attr entry for full LSM context
In reply to Casey SchauflerAdd an entry /proc/.../attr/context which displays the full process security "context" in compound format:
lsm1\0value\0lsm2\0value\0...
This entry is not writable.
A security module may decide that its policy does not allow this information to be displayed. In this case none of the information will be displayed.
--- Documentation/ABI/testing/procfs-attr-context | 14 ++++ Documentation/security/lsm.rst | 14 ++++ fs/proc/base.c | 1 + include/linux/lsm_hooks.h | 6 ++ security/apparmor/include/procattr.h | 2 +- security/apparmor/lsm.c | 8 +- security/apparmor/procattr.c | 22 +++--- security/security.c | 79 +++++++++++++++++++ security/selinux/hooks.c | 2 +- security/smack/smack_lsm.c | 2 +- 10 files changed, 135 insertions(+), 15 deletions(-) create mode 100644 Documentation/ABI/testing/procfs-attr-context
Documentation/ABI/testing/procfs-attr-context new file +14 -0
@@ -0,0 +1,14 @@ +What: /proc/*/attr/context +Contact: [email protected], +Description: The current security information used by all Linux + security module (LSMs) that are active on the system. + The details of permissions required to read from + this interface and hence obtain the security state + of the task identified is dependent on the LSMs that + are active on the system. + A process cannot write to this interface. + The data provided by this interface will have the form: + lsm_name\0lsm_data\0[lsm_name\0lsm_data\0]... + where lsm_name is the name of the LSM and the following + lsm_data is the process data for that LSM. +Users: LSM user-space
Documentation/security/lsm.rst +14 -0
@@ -143,3 +143,17 @@ separated list of the active security modules. The file ``/proc/pid/attr/interface_lsm`` contains the name of the security module for which the ``/proc/pid/attr/current`` interface will apply. This interface can be written to. + +The infrastructure does provide an interface for the special +case where multiple security modules provide a process context. +This is provided in compound context format. + +- `lsm\0value\0lsm\0value\0` + +The `lsm` and `value` fields are NUL-terminated bytestrings. +Each field may contain whitespace or non-printable characters. +The NUL bytes are included in the size of a compound context. +The context ``Bell\0Secret\0Biba\0Loose\0`` has a size of 23. + +The file ``/proc/pid/attr/context`` provides the security +context of the identified process.
fs/proc/base.c +1 -0
@@ -2808,6 +2808,7 @@ static const struct pid_entry attr_dir_stuff[] = {ATTR (NULL ,"keycreate" ,0666 ),ATTR (NULL ,"sockcreate" ,0666 ),ATTR (NULL ,"interface_lsm" ,0666 ), +ATTR (NULL ,"context" ,0444 ), #ifdef CONFIG_SECURITY_SMACKDIR ("smack" ,0555 , proc_smack_attr_dir_inode_ops, proc_smack_attr_dir_ops),
include/linux/lsm_hooks.h +6 -0
@@ -1397,6 +1397,12 @@* @pages contains the number of pages. * Return 0 if permission is granted. * +* @getprocattr: +* Provide the named process attribute for display in special files in +* the /proc/.../attr directory. Attribute naming and the data displayed +* is at the discretion of the security modules. The exception is the +* "context" attribute, which will contain the security context of the +* task as a nul terminated text string without trailing whitespace. * @ismaclabel: * Check if the extended attribute specified by @name * represents a MAC label. Returns 1 if name is a MAC
security/apparmor/include/procattr.h +1 -1
@@ -11,7 +11,7 @@#ifndef __AA_PROCATTR_H#define __AA_PROCATTR_H -int aa_getprocattr (struct aa_label *label,char **string); +int aa_getprocattr (struct aa_label *label,char **string,bool newline);int aa_setprocattr_changehat (char *args,size_t size,int flags);#endif /* __AA_PROCATTR_H */
security/apparmor/lsm.c +6 -2
@@ -602,6 +602,7 @@ static int apparmor_getprocattr(struct task_struct *task, char *name,const struct cred *cred =get_task_cred (task);struct aa_task_ctx *ctx =task_ctx (current);struct aa_label *label =NULL ; +bool newline = true;if (strcmp (name,"current" ) ==0 ) label =aa_get_newest_label (cred_label (cred));
@@ -609,11 +610,14 @@ static int apparmor_getprocattr(struct task_struct *task, char *name, label =aa_get_newest_label (ctx->previous); else if (strcmp(name, "exec") == 0 && ctx->onexec) label =aa_get_newest_label (ctx->onexec); - else + else if (strcmp(name, "context") == 0) { + label = aa_get_newest_label(cred_label(cred)); + newline = false; + } else error = -EINVAL; if (label) - error =aa_getprocattr (label, value); + error = aa_getprocattr(label, value, newline); aa_put_label(label); put_cred(cred);
security/apparmor/procattr.c +12 -10
@@ -20,6 +20,7 @@* aa_getprocattr - Return the profile information for @profile * @profile: the profile to print profile info about (NOT NULL) * @string: Returns - string containing the profile info (NOT NULL) +* @newline: Should a newline be added to @string. * * Returns: length of @string on success else error on failure *
@@ -30,20 +31,21 @@* * Returns: size of string placed in @string else error code on failure */ -int aa_getprocattr (struct aa_label *label,char **string) +int aa_getprocattr (struct aa_label *label,char **string,bool newline) {struct aa_ns *ns =labels_ns (label);struct aa_ns *current_ns =aa_get_current_ns (); +int flags =FLAG_VIEW_SUBNS |FLAG_HIDDEN_UNCONFINED ;int len;if (!aa_ns_visible (current_ns, ns, true)) {aa_put_ns (current_ns);return -EACCES ; } +if (newline) + flags |=FLAG_SHOW_MODE ; - len =aa_label_snxprint (NULL ,0 , current_ns, label, -FLAG_SHOW_MODE |FLAG_VIEW_SUBNS | -FLAG_HIDDEN_UNCONFINED ); + len =aa_label_snxprint (NULL ,0 , current_ns, label, flags);AA_BUG (len <0 ); *string =kmalloc (len +2 ,GFP_KERNEL );
@@ -52,19 +54,19 @@ int aa_getprocattr(struct aa_label *label, char **string)return -ENOMEM ; } - len =aa_label_snxprint (*string, len +2 , current_ns, label, -FLAG_SHOW_MODE |FLAG_VIEW_SUBNS | -FLAG_HIDDEN_UNCONFINED ); + len =aa_label_snxprint (*string, len +2 , current_ns, label, flags);if (len <0 ) {aa_put_ns (current_ns);return len; } - (*string)[len] ='\n' ; - (*string)[len +1 ] =0 ; +if (newline) { + (*string)[len] ='\n' ; + (*string)[++len] =0 ; + }aa_put_ns (current_ns); -return len +1 ; +return len; }/**
security/security.c +79 -0
97 lines not on this page · raw message
security/selinux/hooks.c +1 -1
@@ -6468,7 +6468,7 @@ static int selinux_getprocattr(struct task_struct *p, goto bad; } -if (!strcmp (name,"current" )) +if (!strcmp (name,"current" ) || !strcmp (name,"context" )) sid = __tsec->sid;else if (!strcmp (name,"prev" )) sid = __tsec->osid;
security/smack/smack_lsm.c +1 -1
@@ -3477,7 +3477,7 @@ static int smack_getprocattr(struct task_struct *p, char *name, char **value)char *cp;int slen; -if (strcmp (name,"current" ) !=0 ) +if (strcmp (name,"current" ) !=0 &&strcmp (name,"context" ) !=0 )return -EINVAL ; cp =kstrdup (skp->smk_known,GFP_KERNEL );
Signature · 1 line
2.29.2
Casey Schaufler
maintainerPATCH v26 25/25 AppArmor: Remove the exclusive flag
In reply to Casey SchauflerWith the inclusion of the interface LSM process attribute mechanism AppArmor no longer needs to be treated as an "exclusive" security module. Remove the flag that indicates it is exclusive. Remove the stub getpeersec_dgram AppArmor hook as it has no effect in the single LSM case and interferes in the multiple LSM case.
--- security/apparmor/lsm.c | 20 +------------------- 1 file changed, 1 insertion(+), 19 deletions(-)
security/apparmor/lsm.c +1 -19
@@ -1138,22 +1138,6 @@ static int apparmor_socket_getpeersec_stream(struct socket *sock,return error; } -/** -* apparmor_socket_getpeersec_dgram - get security label of packet -* @sock: the peer socket -* @skb: packet data -* @secid: pointer to where to put the secid of the packet -* -* Sets the netlabel socket state on sk from parent -*/ -static int apparmor_socket_getpeersec_dgram (struct socket *sock, -struct sk_buff *skb,u32 *secid) - -{ -/* TODO: requires secid support */ -return -ENOPROTOOPT ; -} -/** * apparmor_sock_graft - Initialize newly created socket * @sk: child sock
@@ -1257,8 +1241,6 @@ static struct security_hook_list apparmor_hooks[] __lsm_ro_after_init = {#endif LSM_HOOK_INIT (socket_getpeersec_stream, apparmor_socket_getpeersec_stream), -LSM_HOOK_INIT (socket_getpeersec_dgram, - apparmor_socket_getpeersec_dgram),LSM_HOOK_INIT (sock_graft, apparmor_sock_graft), #ifdef CONFIG_NETWORK_SECMARK LSM_HOOK_INIT(inet_conn_request, apparmor_inet_conn_request),
@@ -1928,7 +1910,7 @@ static int __init apparmor_init(void)DEFINE_LSM (apparmor) = { .name ="apparmor" , - .flags =LSM_FLAG_LEGACY_MAJOR |LSM_FLAG_EXCLUSIVE , + .flags =LSM_FLAG_LEGACY_MAJOR , .enabled = &apparmor_enabled, .blobs = &apparmor_blob_sizes, .init = apparmor_init,
Signature · 1 line
2.29.2