linux-stable/security/selinux/avc.c

1225 lines
32 KiB
C
Raw Permalink Normal View History

// SPDX-License-Identifier: GPL-2.0-only
/*
* Implementation of the kernel access vector cache (AVC).
*
* Authors: Stephen Smalley, <sds@tycho.nsa.gov>
* James Morris <jmorris@redhat.com>
*
* Update: KaiGai, Kohei <kaigai@ak.jp.nec.com>
* Replaced the avc_lock spinlock by RCU.
*
* Copyright (C) 2003 Red Hat, Inc., James Morris <jmorris@redhat.com>
*/
#include <linux/types.h>
#include <linux/stddef.h>
#include <linux/kernel.h>
#include <linux/slab.h>
#include <linux/fs.h>
#include <linux/dcache.h>
#include <linux/init.h>
#include <linux/skbuff.h>
#include <linux/percpu.h>
selinux: extended permissions for ioctls Add extended permissions logic to selinux. Extended permissions provides additional permissions in 256 bit increments. Extend the generic ioctl permission check to use the extended permissions for per-command filtering. Source/target/class sets including the ioctl permission may additionally include a set of commands. Example: allowxperm <source> <target>:<class> ioctl unpriv_app_socket_cmds auditallowxperm <source> <target>:<class> ioctl priv_gpu_cmds Where unpriv_app_socket_cmds and priv_gpu_cmds are macros representing commonly granted sets of ioctl commands. When ioctl commands are omitted only the permissions are checked. This feature is intended to provide finer granularity for the ioctl permission that may be too imprecise. For example, the same driver may use ioctls to provide important and benign functionality such as driver version or socket type as well as dangerous capabilities such as debugging features, read/write/execute to physical memory or access to sensitive data. Per-command filtering provides a mechanism to reduce the attack surface of the kernel, and limit applications to the subset of commands required. The format of the policy binary has been modified to include ioctl commands, and the policy version number has been incremented to POLICYDB_VERSION_XPERMS_IOCTL=30 to account for the format change. The extended permissions logic is deliberately generic to allow components to be reused e.g. netlink filters Signed-off-by: Jeff Vander Stoep <jeffv@google.com> Acked-by: Nick Kralevich <nnk@google.com> Signed-off-by: Paul Moore <pmoore@redhat.com>
2015-07-10 21:19:56 +00:00
#include <linux/list.h>
#include <net/sock.h>
#include <linux/un.h>
#include <net/af_unix.h>
#include <linux/ip.h>
#include <linux/audit.h>
#include <linux/ipv6.h>
#include <net/ipv6.h>
#include "avc.h"
#include "avc_ss.h"
selinux: dynamic class/perm discovery Modify SELinux to dynamically discover class and permission values upon policy load, based on the dynamic object class/perm discovery logic from libselinux. A mapping is created between kernel-private class and permission indices used outside the security server and the policy values used within the security server. The mappings are only applied upon kernel-internal computations; similar mappings for the private indices of userspace object managers is handled on a per-object manager basis by the userspace AVC. The interfaces for compute_av and transition_sid are split for kernel vs. userspace; the userspace functions are distinguished by a _user suffix. The kernel-private class indices are no longer tied to the policy values and thus do not need to skip indices for userspace classes; thus the kernel class index values are compressed. The flask.h definitions were regenerated by deleting the userspace classes from refpolicy's definitions and then regenerating the headers. Going forward, we can just maintain the flask.h, av_permissions.h, and classmap.h definitions separately from policy as they are no longer tied to the policy values. The next patch introduces a utility to automate generation of flask.h and av_permissions.h from the classmap.h definitions. The older kernel class and permission string tables are removed and replaced by a single security class mapping table that is walked at policy load to generate the mapping. The old kernel class validation logic is completely replaced by the mapping logic. The handle unknown logic is reworked. reject_unknown=1 is handled when the mappings are computed at policy load time, similar to the old handling by the class validation logic. allow_unknown=1 is handled when computing and mapping decisions - if the permission was not able to be mapped (i.e. undefined, mapped to zero), then it is automatically added to the allowed vector. If the class was not able to be mapped (i.e. undefined, mapped to zero), then all permissions are allowed for it if allow_unknown=1. avc_audit leverages the new security class mapping table to lookup the class and permission names from the kernel-private indices. The mdp program is updated to use the new table when generating the class definitions and allow rules for a minimal boot policy for the kernel. It should be noted that this policy will not include any userspace classes, nor will its policy index values for the kernel classes correspond with the ones in refpolicy (they will instead match the kernel-private indices). Signed-off-by: Stephen Smalley <sds@tycho.nsa.gov> Signed-off-by: James Morris <jmorris@namei.org>
2009-09-30 17:37:50 +00:00
#include "classmap.h"
selinux: add tracepoint on audited events The audit data currently captures which process and which target is responsible for a denial. There is no data on where exactly in the process that call occurred. Debugging can be made easier by being able to reconstruct the unified kernel and userland stack traces [1]. Add a tracepoint on the SELinux denials which can then be used by userland (i.e. perf). Although this patch could manually be added by each OS developer to trouble shoot a denial, adding it to the kernel streamlines the developers workflow. It is possible to use perf for monitoring the event: # perf record -e avc:selinux_audited -g -a ^C # perf report -g [...] 6.40% 6.40% audited=800000 tclass=4 | __libc_start_main | |--4.60%--__GI___ioctl | entry_SYSCALL_64 | do_syscall_64 | __x64_sys_ioctl | ksys_ioctl | binder_ioctl | binder_set_nice | can_nice | capable | security_capable | cred_has_capability.isra.0 | slow_avc_audit | common_lsm_audit | avc_audit_post_callback | avc_audit_post_callback | It is also possible to use the ftrace interface: # echo 1 > /sys/kernel/debug/tracing/events/avc/selinux_audited/enable # cat /sys/kernel/debug/tracing/trace tracer: nop entries-in-buffer/entries-written: 1/1 #P:8 [...] dmesg-3624 [001] 13072.325358: selinux_denied: audited=800000 tclass=4 The tclass value can be mapped to a class by searching security/selinux/flask.h. The audited value is a bit field of the permissions described in security/selinux/av_permissions.h for the corresponding class. [1] https://source.android.com/devices/tech/debug/native_stack_dump Signed-off-by: Thiébaud Weksteen <tweek@google.com> Suggested-by: Joel Fernandes <joelaf@google.com> Reviewed-by: Peter Enderborg <peter.enderborg@sony.com> Acked-by: Stephen Smalley <stephen.smalley.work@gmail.com> Signed-off-by: Paul Moore <paul@paul-moore.com>
2020-08-21 14:08:21 +00:00
#define CREATE_TRACE_POINTS
#include <trace/events/avc.h>
#define AVC_CACHE_SLOTS 512
#define AVC_DEF_CACHE_THRESHOLD 512
#define AVC_CACHE_RECLAIM 16
#ifdef CONFIG_SECURITY_SELINUX_AVC_STATS
#define avc_cache_stats_incr(field) this_cpu_inc(avc_cache_stats.field)
#else
#define avc_cache_stats_incr(field) do {} while (0)
#endif
struct avc_entry {
u32 ssid;
u32 tsid;
u16 tclass;
struct av_decision avd;
selinux: extended permissions for ioctls Add extended permissions logic to selinux. Extended permissions provides additional permissions in 256 bit increments. Extend the generic ioctl permission check to use the extended permissions for per-command filtering. Source/target/class sets including the ioctl permission may additionally include a set of commands. Example: allowxperm <source> <target>:<class> ioctl unpriv_app_socket_cmds auditallowxperm <source> <target>:<class> ioctl priv_gpu_cmds Where unpriv_app_socket_cmds and priv_gpu_cmds are macros representing commonly granted sets of ioctl commands. When ioctl commands are omitted only the permissions are checked. This feature is intended to provide finer granularity for the ioctl permission that may be too imprecise. For example, the same driver may use ioctls to provide important and benign functionality such as driver version or socket type as well as dangerous capabilities such as debugging features, read/write/execute to physical memory or access to sensitive data. Per-command filtering provides a mechanism to reduce the attack surface of the kernel, and limit applications to the subset of commands required. The format of the policy binary has been modified to include ioctl commands, and the policy version number has been incremented to POLICYDB_VERSION_XPERMS_IOCTL=30 to account for the format change. The extended permissions logic is deliberately generic to allow components to be reused e.g. netlink filters Signed-off-by: Jeff Vander Stoep <jeffv@google.com> Acked-by: Nick Kralevich <nnk@google.com> Signed-off-by: Paul Moore <pmoore@redhat.com>
2015-07-10 21:19:56 +00:00
struct avc_xperms_node *xp_node;
};
struct avc_node {
struct avc_entry ae;
struct hlist_node list; /* anchored in avc_cache->slots[i] */
struct rcu_head rhead;
};
selinux: extended permissions for ioctls Add extended permissions logic to selinux. Extended permissions provides additional permissions in 256 bit increments. Extend the generic ioctl permission check to use the extended permissions for per-command filtering. Source/target/class sets including the ioctl permission may additionally include a set of commands. Example: allowxperm <source> <target>:<class> ioctl unpriv_app_socket_cmds auditallowxperm <source> <target>:<class> ioctl priv_gpu_cmds Where unpriv_app_socket_cmds and priv_gpu_cmds are macros representing commonly granted sets of ioctl commands. When ioctl commands are omitted only the permissions are checked. This feature is intended to provide finer granularity for the ioctl permission that may be too imprecise. For example, the same driver may use ioctls to provide important and benign functionality such as driver version or socket type as well as dangerous capabilities such as debugging features, read/write/execute to physical memory or access to sensitive data. Per-command filtering provides a mechanism to reduce the attack surface of the kernel, and limit applications to the subset of commands required. The format of the policy binary has been modified to include ioctl commands, and the policy version number has been incremented to POLICYDB_VERSION_XPERMS_IOCTL=30 to account for the format change. The extended permissions logic is deliberately generic to allow components to be reused e.g. netlink filters Signed-off-by: Jeff Vander Stoep <jeffv@google.com> Acked-by: Nick Kralevich <nnk@google.com> Signed-off-by: Paul Moore <pmoore@redhat.com>
2015-07-10 21:19:56 +00:00
struct avc_xperms_decision_node {
struct extended_perms_decision xpd;
struct list_head xpd_list; /* list of extended_perms_decision */
};
struct avc_xperms_node {
struct extended_perms xp;
struct list_head xpd_head; /* list head of extended_perms_decision */
};
struct avc_cache {
struct hlist_head slots[AVC_CACHE_SLOTS]; /* head for avc_node->list */
spinlock_t slots_lock[AVC_CACHE_SLOTS]; /* lock for writes */
atomic_t lru_hint; /* LRU hint for reclaim scan */
atomic_t active_nodes;
u32 latest_notif; /* latest revocation notification */
};
struct avc_callback_node {
int (*callback) (u32 event);
u32 events;
struct avc_callback_node *next;
};
#ifdef CONFIG_SECURITY_SELINUX_AVC_STATS
DEFINE_PER_CPU(struct avc_cache_stats, avc_cache_stats) = { 0 };
#endif
struct selinux_avc {
unsigned int avc_cache_threshold;
struct avc_cache avc_cache;
};
static struct selinux_avc selinux_avc;
void selinux_avc_init(void)
{
int i;
selinux_avc.avc_cache_threshold = AVC_DEF_CACHE_THRESHOLD;
for (i = 0; i < AVC_CACHE_SLOTS; i++) {
INIT_HLIST_HEAD(&selinux_avc.avc_cache.slots[i]);
spin_lock_init(&selinux_avc.avc_cache.slots_lock[i]);
}
atomic_set(&selinux_avc.avc_cache.active_nodes, 0);
atomic_set(&selinux_avc.avc_cache.lru_hint, 0);
}
unsigned int avc_get_cache_threshold(void)
{
return selinux_avc.avc_cache_threshold;
}
void avc_set_cache_threshold(unsigned int cache_threshold)
{
selinux_avc.avc_cache_threshold = cache_threshold;
}
static struct avc_callback_node *avc_callbacks __ro_after_init;
static struct kmem_cache *avc_node_cachep __ro_after_init;
static struct kmem_cache *avc_xperms_data_cachep __ro_after_init;
static struct kmem_cache *avc_xperms_decision_cachep __ro_after_init;
static struct kmem_cache *avc_xperms_cachep __ro_after_init;
static inline int avc_hash(u32 ssid, u32 tsid, u16 tclass)
{
return (ssid ^ (tsid<<2) ^ (tclass<<4)) & (AVC_CACHE_SLOTS - 1);
}
/**
* avc_init - Initialize the AVC.
*
* Initialize the access vector cache.
*/
void __init avc_init(void)
{
avc_node_cachep = kmem_cache_create("avc_node", sizeof(struct avc_node),
selinux: extended permissions for ioctls Add extended permissions logic to selinux. Extended permissions provides additional permissions in 256 bit increments. Extend the generic ioctl permission check to use the extended permissions for per-command filtering. Source/target/class sets including the ioctl permission may additionally include a set of commands. Example: allowxperm <source> <target>:<class> ioctl unpriv_app_socket_cmds auditallowxperm <source> <target>:<class> ioctl priv_gpu_cmds Where unpriv_app_socket_cmds and priv_gpu_cmds are macros representing commonly granted sets of ioctl commands. When ioctl commands are omitted only the permissions are checked. This feature is intended to provide finer granularity for the ioctl permission that may be too imprecise. For example, the same driver may use ioctls to provide important and benign functionality such as driver version or socket type as well as dangerous capabilities such as debugging features, read/write/execute to physical memory or access to sensitive data. Per-command filtering provides a mechanism to reduce the attack surface of the kernel, and limit applications to the subset of commands required. The format of the policy binary has been modified to include ioctl commands, and the policy version number has been incremented to POLICYDB_VERSION_XPERMS_IOCTL=30 to account for the format change. The extended permissions logic is deliberately generic to allow components to be reused e.g. netlink filters Signed-off-by: Jeff Vander Stoep <jeffv@google.com> Acked-by: Nick Kralevich <nnk@google.com> Signed-off-by: Paul Moore <pmoore@redhat.com>
2015-07-10 21:19:56 +00:00
0, SLAB_PANIC, NULL);
avc_xperms_cachep = kmem_cache_create("avc_xperms_node",
sizeof(struct avc_xperms_node),
0, SLAB_PANIC, NULL);
avc_xperms_decision_cachep = kmem_cache_create(
"avc_xperms_decision_node",
sizeof(struct avc_xperms_decision_node),
0, SLAB_PANIC, NULL);
avc_xperms_data_cachep = kmem_cache_create("avc_xperms_data",
sizeof(struct extended_perms_data),
0, SLAB_PANIC, NULL);
}
int avc_get_hash_stats(char *page)
{
int i, chain_len, max_chain_len, slots_used;
struct avc_node *node;
struct hlist_head *head;
rcu_read_lock();
slots_used = 0;
max_chain_len = 0;
for (i = 0; i < AVC_CACHE_SLOTS; i++) {
head = &selinux_avc.avc_cache.slots[i];
if (!hlist_empty(head)) {
slots_used++;
chain_len = 0;
hlist: drop the node parameter from iterators I'm not sure why, but the hlist for each entry iterators were conceived list_for_each_entry(pos, head, member) The hlist ones were greedy and wanted an extra parameter: hlist_for_each_entry(tpos, pos, head, member) Why did they need an extra pos parameter? I'm not quite sure. Not only they don't really need it, it also prevents the iterator from looking exactly like the list iterator, which is unfortunate. Besides the semantic patch, there was some manual work required: - Fix up the actual hlist iterators in linux/list.h - Fix up the declaration of other iterators based on the hlist ones. - A very small amount of places were using the 'node' parameter, this was modified to use 'obj->member' instead. - Coccinelle didn't handle the hlist_for_each_entry_safe iterator properly, so those had to be fixed up manually. The semantic patch which is mostly the work of Peter Senna Tschudin is here: @@ iterator name hlist_for_each_entry, hlist_for_each_entry_continue, hlist_for_each_entry_from, hlist_for_each_entry_rcu, hlist_for_each_entry_rcu_bh, hlist_for_each_entry_continue_rcu_bh, for_each_busy_worker, ax25_uid_for_each, ax25_for_each, inet_bind_bucket_for_each, sctp_for_each_hentry, sk_for_each, sk_for_each_rcu, sk_for_each_from, sk_for_each_safe, sk_for_each_bound, hlist_for_each_entry_safe, hlist_for_each_entry_continue_rcu, nr_neigh_for_each, nr_neigh_for_each_safe, nr_node_for_each, nr_node_for_each_safe, for_each_gfn_indirect_valid_sp, for_each_gfn_sp, for_each_host; type T; expression a,c,d,e; identifier b; statement S; @@ -T b; <+... when != b ( hlist_for_each_entry(a, - b, c, d) S | hlist_for_each_entry_continue(a, - b, c) S | hlist_for_each_entry_from(a, - b, c) S | hlist_for_each_entry_rcu(a, - b, c, d) S | hlist_for_each_entry_rcu_bh(a, - b, c, d) S | hlist_for_each_entry_continue_rcu_bh(a, - b, c) S | for_each_busy_worker(a, c, - b, d) S | ax25_uid_for_each(a, - b, c) S | ax25_for_each(a, - b, c) S | inet_bind_bucket_for_each(a, - b, c) S | sctp_for_each_hentry(a, - b, c) S | sk_for_each(a, - b, c) S | sk_for_each_rcu(a, - b, c) S | sk_for_each_from -(a, b) +(a) S + sk_for_each_from(a) S | sk_for_each_safe(a, - b, c, d) S | sk_for_each_bound(a, - b, c) S | hlist_for_each_entry_safe(a, - b, c, d, e) S | hlist_for_each_entry_continue_rcu(a, - b, c) S | nr_neigh_for_each(a, - b, c) S | nr_neigh_for_each_safe(a, - b, c, d) S | nr_node_for_each(a, - b, c) S | nr_node_for_each_safe(a, - b, c, d) S | - for_each_gfn_sp(a, c, d, b) S + for_each_gfn_sp(a, c, d) S | - for_each_gfn_indirect_valid_sp(a, c, d, b) S + for_each_gfn_indirect_valid_sp(a, c, d) S | for_each_host(a, - b, c) S | for_each_host_safe(a, - b, c, d) S | for_each_mesh_entry(a, - b, c, d) S ) ...+> [akpm@linux-foundation.org: drop bogus change from net/ipv4/raw.c] [akpm@linux-foundation.org: drop bogus hunk from net/ipv6/raw.c] [akpm@linux-foundation.org: checkpatch fixes] [akpm@linux-foundation.org: fix warnings] [akpm@linux-foudnation.org: redo intrusive kvm changes] Tested-by: Peter Senna Tschudin <peter.senna@gmail.com> Acked-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com> Signed-off-by: Sasha Levin <sasha.levin@oracle.com> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Marcelo Tosatti <mtosatti@redhat.com> Cc: Gleb Natapov <gleb@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-02-28 01:06:00 +00:00
hlist_for_each_entry_rcu(node, head, list)
chain_len++;
if (chain_len > max_chain_len)
max_chain_len = chain_len;
}
}
rcu_read_unlock();
return scnprintf(page, PAGE_SIZE, "entries: %d\nbuckets used: %d/%d\n"
"longest chain: %d\n",
atomic_read(&selinux_avc.avc_cache.active_nodes),
slots_used, AVC_CACHE_SLOTS, max_chain_len);
}
selinux: extended permissions for ioctls Add extended permissions logic to selinux. Extended permissions provides additional permissions in 256 bit increments. Extend the generic ioctl permission check to use the extended permissions for per-command filtering. Source/target/class sets including the ioctl permission may additionally include a set of commands. Example: allowxperm <source> <target>:<class> ioctl unpriv_app_socket_cmds auditallowxperm <source> <target>:<class> ioctl priv_gpu_cmds Where unpriv_app_socket_cmds and priv_gpu_cmds are macros representing commonly granted sets of ioctl commands. When ioctl commands are omitted only the permissions are checked. This feature is intended to provide finer granularity for the ioctl permission that may be too imprecise. For example, the same driver may use ioctls to provide important and benign functionality such as driver version or socket type as well as dangerous capabilities such as debugging features, read/write/execute to physical memory or access to sensitive data. Per-command filtering provides a mechanism to reduce the attack surface of the kernel, and limit applications to the subset of commands required. The format of the policy binary has been modified to include ioctl commands, and the policy version number has been incremented to POLICYDB_VERSION_XPERMS_IOCTL=30 to account for the format change. The extended permissions logic is deliberately generic to allow components to be reused e.g. netlink filters Signed-off-by: Jeff Vander Stoep <jeffv@google.com> Acked-by: Nick Kralevich <nnk@google.com> Signed-off-by: Paul Moore <pmoore@redhat.com>
2015-07-10 21:19:56 +00:00
/*
* using a linked list for extended_perms_decision lookup because the list is
* always small. i.e. less than 5, typically 1
*/
static struct extended_perms_decision *avc_xperms_decision_lookup(u8 driver,
struct avc_xperms_node *xp_node)
{
struct avc_xperms_decision_node *xpd_node;
list_for_each_entry(xpd_node, &xp_node->xpd_head, xpd_list) {
if (xpd_node->xpd.driver == driver)
return &xpd_node->xpd;
}
return NULL;
}
static inline unsigned int
avc_xperms_has_perm(struct extended_perms_decision *xpd,
u8 perm, u8 which)
{
unsigned int rc = 0;
if ((which == XPERMS_ALLOWED) &&
(xpd->used & XPERMS_ALLOWED))
rc = security_xperm_test(xpd->allowed->p, perm);
else if ((which == XPERMS_AUDITALLOW) &&
(xpd->used & XPERMS_AUDITALLOW))
rc = security_xperm_test(xpd->auditallow->p, perm);
else if ((which == XPERMS_DONTAUDIT) &&
(xpd->used & XPERMS_DONTAUDIT))
rc = security_xperm_test(xpd->dontaudit->p, perm);
return rc;
}
static void avc_xperms_allow_perm(struct avc_xperms_node *xp_node,
u8 driver, u8 perm)
{
struct extended_perms_decision *xpd;
security_xperm_set(xp_node->xp.drivers.p, driver);
xpd = avc_xperms_decision_lookup(driver, xp_node);
if (xpd && xpd->allowed)
security_xperm_set(xpd->allowed->p, perm);
}
static void avc_xperms_decision_free(struct avc_xperms_decision_node *xpd_node)
{
struct extended_perms_decision *xpd;
xpd = &xpd_node->xpd;
if (xpd->allowed)
kmem_cache_free(avc_xperms_data_cachep, xpd->allowed);
if (xpd->auditallow)
kmem_cache_free(avc_xperms_data_cachep, xpd->auditallow);
if (xpd->dontaudit)
kmem_cache_free(avc_xperms_data_cachep, xpd->dontaudit);
kmem_cache_free(avc_xperms_decision_cachep, xpd_node);
}
static void avc_xperms_free(struct avc_xperms_node *xp_node)
{
struct avc_xperms_decision_node *xpd_node, *tmp;
if (!xp_node)
return;
list_for_each_entry_safe(xpd_node, tmp, &xp_node->xpd_head, xpd_list) {
list_del(&xpd_node->xpd_list);
avc_xperms_decision_free(xpd_node);
}
kmem_cache_free(avc_xperms_cachep, xp_node);
}
static void avc_copy_xperms_decision(struct extended_perms_decision *dest,
struct extended_perms_decision *src)
{
dest->driver = src->driver;
dest->used = src->used;
if (dest->used & XPERMS_ALLOWED)
memcpy(dest->allowed->p, src->allowed->p,
sizeof(src->allowed->p));
if (dest->used & XPERMS_AUDITALLOW)
memcpy(dest->auditallow->p, src->auditallow->p,
sizeof(src->auditallow->p));
if (dest->used & XPERMS_DONTAUDIT)
memcpy(dest->dontaudit->p, src->dontaudit->p,
sizeof(src->dontaudit->p));
}
/*
* similar to avc_copy_xperms_decision, but only copy decision
* information relevant to this perm
*/
static inline void avc_quick_copy_xperms_decision(u8 perm,
struct extended_perms_decision *dest,
struct extended_perms_decision *src)
{
/*
* compute index of the u32 of the 256 bits (8 u32s) that contain this
* command permission
*/
u8 i = perm >> 5;
dest->used = src->used;
if (dest->used & XPERMS_ALLOWED)
dest->allowed->p[i] = src->allowed->p[i];
if (dest->used & XPERMS_AUDITALLOW)
dest->auditallow->p[i] = src->auditallow->p[i];
if (dest->used & XPERMS_DONTAUDIT)
dest->dontaudit->p[i] = src->dontaudit->p[i];
}
static struct avc_xperms_decision_node
*avc_xperms_decision_alloc(u8 which)
{
struct avc_xperms_decision_node *xpd_node;
struct extended_perms_decision *xpd;
selinux: use __GFP_NOWARN with GFP_NOWAIT in the AVC In the field, we have seen lots of allocation failure from the call path below. 06-03 13:29:12.999 1010315 31557 31557 W Binder : 31542_2: page allocation failure: order:0, mode:0x800(GFP_NOWAIT), nodemask=(null),cpuset=background,mems_allowed=0 ... ... 06-03 13:29:12.999 1010315 31557 31557 W Call trace: 06-03 13:29:12.999 1010315 31557 31557 W : dump_backtrace.cfi_jt+0x0/0x8 06-03 13:29:12.999 1010315 31557 31557 W : dump_stack+0xc8/0x14c 06-03 13:29:12.999 1010315 31557 31557 W : warn_alloc+0x158/0x1c8 06-03 13:29:12.999 1010315 31557 31557 W : __alloc_pages_slowpath+0x9d8/0xb80 06-03 13:29:12.999 1010315 31557 31557 W : __alloc_pages_nodemask+0x1c4/0x430 06-03 13:29:12.999 1010315 31557 31557 W : allocate_slab+0xb4/0x390 06-03 13:29:12.999 1010315 31557 31557 W : ___slab_alloc+0x12c/0x3a4 06-03 13:29:12.999 1010315 31557 31557 W : kmem_cache_alloc+0x358/0x5e4 06-03 13:29:12.999 1010315 31557 31557 W : avc_alloc_node+0x30/0x184 06-03 13:29:12.999 1010315 31557 31557 W : avc_update_node+0x54/0x4f0 06-03 13:29:12.999 1010315 31557 31557 W : avc_has_extended_perms+0x1a4/0x460 06-03 13:29:12.999 1010315 31557 31557 W : selinux_file_ioctl+0x320/0x3d0 06-03 13:29:12.999 1010315 31557 31557 W : __arm64_sys_ioctl+0xec/0x1fc 06-03 13:29:12.999 1010315 31557 31557 W : el0_svc_common+0xc0/0x24c 06-03 13:29:12.999 1010315 31557 31557 W : el0_svc+0x28/0x88 06-03 13:29:12.999 1010315 31557 31557 W : el0_sync_handler+0x8c/0xf0 06-03 13:29:12.999 1010315 31557 31557 W : el0_sync+0x1a4/0x1c0 .. .. 06-03 13:29:12.999 1010315 31557 31557 W SLUB : Unable to allocate memory on node -1, gfp=0x900(GFP_NOWAIT|__GFP_ZERO) 06-03 13:29:12.999 1010315 31557 31557 W cache : avc_node, object size: 72, buffer size: 80, default order: 0, min order: 0 06-03 13:29:12.999 1010315 31557 31557 W node 0 : slabs: 57, objs: 2907, free: 0 06-03 13:29:12.999 1010161 10686 10686 W SLUB : Unable to allocate memory on node -1, gfp=0x900(GFP_NOWAIT|__GFP_ZERO) 06-03 13:29:12.999 1010161 10686 10686 W cache : avc_node, object size: 72, buffer size: 80, default order: 0, min order: 0 06-03 13:29:12.999 1010161 10686 10686 W node 0 : slabs: 57, objs: 2907, free: 0 06-03 13:29:12.999 1010161 10686 10686 W SLUB : Unable to allocate memory on node -1, gfp=0x900(GFP_NOWAIT|__GFP_ZERO) 06-03 13:29:12.999 1010161 10686 10686 W cache : avc_node, object size: 72, buffer size: 80, default order: 0, min order: 0 06-03 13:29:12.999 1010161 10686 10686 W node 0 : slabs: 57, objs: 2907, free: 0 06-03 13:29:12.999 1010161 10686 10686 W SLUB : Unable to allocate memory on node -1, gfp=0x900(GFP_NOWAIT|__GFP_ZERO) 06-03 13:29:12.999 1010161 10686 10686 W cache : avc_node, object size: 72, buffer size: 80, default order: 0, min order: 0 06-03 13:29:12.999 1010161 10686 10686 W node 0 : slabs: 57, objs: 2907, free: 0 06-03 13:29:13.000 1010161 10686 10686 W SLUB : Unable to allocate memory on node -1, gfp=0x900(GFP_NOWAIT|__GFP_ZERO) 06-03 13:29:13.000 1010161 10686 10686 W cache : avc_node, object size: 72, buffer size: 80, default order: 0, min order: 0 06-03 13:29:13.000 1010161 10686 10686 W node 0 : slabs: 57, objs: 2907, free: 0 06-03 13:29:13.000 1010161 10686 10686 W SLUB : Unable to allocate memory on node -1, gfp=0x900(GFP_NOWAIT|__GFP_ZERO) 06-03 13:29:13.000 1010161 10686 10686 W cache : avc_node, object size: 72, buffer size: 80, default order: 0, min order: 0 06-03 13:29:13.000 1010161 10686 10686 W node 0 : slabs: 57, objs: 2907, free: 0 06-03 13:29:13.000 1010161 10686 10686 W SLUB : Unable to allocate memory on node -1, gfp=0x900(GFP_NOWAIT|__GFP_ZERO) 06-03 13:29:13.000 1010161 10686 10686 W cache : avc_node, object size: 72, buffer size: 80, default order: 0, min order: 0 06-03 13:29:13.000 1010161 10686 10686 W node 0 : slabs: 57, objs: 2907, free: 0 06-03 13:29:13.000 10230 30892 30892 W SLUB : Unable to allocate memory on node -1, gfp=0x900(GFP_NOWAIT|__GFP_ZERO) 06-03 13:29:13.000 10230 30892 30892 W cache : avc_node, object size: 72, buffer size: 80, default order: 0, min order: 0 06-03 13:29:13.000 10230 30892 30892 W node 0 : slabs: 57, objs: 2907, free: 0 06-03 13:29:13.000 10230 30892 30892 W SLUB : Unable to allocate memory on node -1, gfp=0x900(GFP_NOWAIT|__GFP_ZERO) 06-03 13:29:13.000 10230 30892 30892 W cache : avc_node, object size: 72, buffer size: 80, default order: 0, min order: 0 Based on [1], selinux is tolerate for failure of memory allocation. Then, use __GFP_NOWARN together. [1] 476accbe2f6e ("selinux: use GFP_NOWAIT in the AVC kmem_caches") Signed-off-by: Minchan Kim <minchan@kernel.org> [PM: subj fix, line wraps, normalized commit refs] Signed-off-by: Paul Moore <paul@paul-moore.com>
2021-06-09 16:37:17 +00:00
xpd_node = kmem_cache_zalloc(avc_xperms_decision_cachep,
GFP_NOWAIT | __GFP_NOWARN);
selinux: extended permissions for ioctls Add extended permissions logic to selinux. Extended permissions provides additional permissions in 256 bit increments. Extend the generic ioctl permission check to use the extended permissions for per-command filtering. Source/target/class sets including the ioctl permission may additionally include a set of commands. Example: allowxperm <source> <target>:<class> ioctl unpriv_app_socket_cmds auditallowxperm <source> <target>:<class> ioctl priv_gpu_cmds Where unpriv_app_socket_cmds and priv_gpu_cmds are macros representing commonly granted sets of ioctl commands. When ioctl commands are omitted only the permissions are checked. This feature is intended to provide finer granularity for the ioctl permission that may be too imprecise. For example, the same driver may use ioctls to provide important and benign functionality such as driver version or socket type as well as dangerous capabilities such as debugging features, read/write/execute to physical memory or access to sensitive data. Per-command filtering provides a mechanism to reduce the attack surface of the kernel, and limit applications to the subset of commands required. The format of the policy binary has been modified to include ioctl commands, and the policy version number has been incremented to POLICYDB_VERSION_XPERMS_IOCTL=30 to account for the format change. The extended permissions logic is deliberately generic to allow components to be reused e.g. netlink filters Signed-off-by: Jeff Vander Stoep <jeffv@google.com> Acked-by: Nick Kralevich <nnk@google.com> Signed-off-by: Paul Moore <pmoore@redhat.com>
2015-07-10 21:19:56 +00:00
if (!xpd_node)
return NULL;
xpd = &xpd_node->xpd;
if (which & XPERMS_ALLOWED) {
xpd->allowed = kmem_cache_zalloc(avc_xperms_data_cachep,
selinux: use __GFP_NOWARN with GFP_NOWAIT in the AVC In the field, we have seen lots of allocation failure from the call path below. 06-03 13:29:12.999 1010315 31557 31557 W Binder : 31542_2: page allocation failure: order:0, mode:0x800(GFP_NOWAIT), nodemask=(null),cpuset=background,mems_allowed=0 ... ... 06-03 13:29:12.999 1010315 31557 31557 W Call trace: 06-03 13:29:12.999 1010315 31557 31557 W : dump_backtrace.cfi_jt+0x0/0x8 06-03 13:29:12.999 1010315 31557 31557 W : dump_stack+0xc8/0x14c 06-03 13:29:12.999 1010315 31557 31557 W : warn_alloc+0x158/0x1c8 06-03 13:29:12.999 1010315 31557 31557 W : __alloc_pages_slowpath+0x9d8/0xb80 06-03 13:29:12.999 1010315 31557 31557 W : __alloc_pages_nodemask+0x1c4/0x430 06-03 13:29:12.999 1010315 31557 31557 W : allocate_slab+0xb4/0x390 06-03 13:29:12.999 1010315 31557 31557 W : ___slab_alloc+0x12c/0x3a4 06-03 13:29:12.999 1010315 31557 31557 W : kmem_cache_alloc+0x358/0x5e4 06-03 13:29:12.999 1010315 31557 31557 W : avc_alloc_node+0x30/0x184 06-03 13:29:12.999 1010315 31557 31557 W : avc_update_node+0x54/0x4f0 06-03 13:29:12.999 1010315 31557 31557 W : avc_has_extended_perms+0x1a4/0x460 06-03 13:29:12.999 1010315 31557 31557 W : selinux_file_ioctl+0x320/0x3d0 06-03 13:29:12.999 1010315 31557 31557 W : __arm64_sys_ioctl+0xec/0x1fc 06-03 13:29:12.999 1010315 31557 31557 W : el0_svc_common+0xc0/0x24c 06-03 13:29:12.999 1010315 31557 31557 W : el0_svc+0x28/0x88 06-03 13:29:12.999 1010315 31557 31557 W : el0_sync_handler+0x8c/0xf0 06-03 13:29:12.999 1010315 31557 31557 W : el0_sync+0x1a4/0x1c0 .. .. 06-03 13:29:12.999 1010315 31557 31557 W SLUB : Unable to allocate memory on node -1, gfp=0x900(GFP_NOWAIT|__GFP_ZERO) 06-03 13:29:12.999 1010315 31557 31557 W cache : avc_node, object size: 72, buffer size: 80, default order: 0, min order: 0 06-03 13:29:12.999 1010315 31557 31557 W node 0 : slabs: 57, objs: 2907, free: 0 06-03 13:29:12.999 1010161 10686 10686 W SLUB : Unable to allocate memory on node -1, gfp=0x900(GFP_NOWAIT|__GFP_ZERO) 06-03 13:29:12.999 1010161 10686 10686 W cache : avc_node, object size: 72, buffer size: 80, default order: 0, min order: 0 06-03 13:29:12.999 1010161 10686 10686 W node 0 : slabs: 57, objs: 2907, free: 0 06-03 13:29:12.999 1010161 10686 10686 W SLUB : Unable to allocate memory on node -1, gfp=0x900(GFP_NOWAIT|__GFP_ZERO) 06-03 13:29:12.999 1010161 10686 10686 W cache : avc_node, object size: 72, buffer size: 80, default order: 0, min order: 0 06-03 13:29:12.999 1010161 10686 10686 W node 0 : slabs: 57, objs: 2907, free: 0 06-03 13:29:12.999 1010161 10686 10686 W SLUB : Unable to allocate memory on node -1, gfp=0x900(GFP_NOWAIT|__GFP_ZERO) 06-03 13:29:12.999 1010161 10686 10686 W cache : avc_node, object size: 72, buffer size: 80, default order: 0, min order: 0 06-03 13:29:12.999 1010161 10686 10686 W node 0 : slabs: 57, objs: 2907, free: 0 06-03 13:29:13.000 1010161 10686 10686 W SLUB : Unable to allocate memory on node -1, gfp=0x900(GFP_NOWAIT|__GFP_ZERO) 06-03 13:29:13.000 1010161 10686 10686 W cache : avc_node, object size: 72, buffer size: 80, default order: 0, min order: 0 06-03 13:29:13.000 1010161 10686 10686 W node 0 : slabs: 57, objs: 2907, free: 0 06-03 13:29:13.000 1010161 10686 10686 W SLUB : Unable to allocate memory on node -1, gfp=0x900(GFP_NOWAIT|__GFP_ZERO) 06-03 13:29:13.000 1010161 10686 10686 W cache : avc_node, object size: 72, buffer size: 80, default order: 0, min order: 0 06-03 13:29:13.000 1010161 10686 10686 W node 0 : slabs: 57, objs: 2907, free: 0 06-03 13:29:13.000 1010161 10686 10686 W SLUB : Unable to allocate memory on node -1, gfp=0x900(GFP_NOWAIT|__GFP_ZERO) 06-03 13:29:13.000 1010161 10686 10686 W cache : avc_node, object size: 72, buffer size: 80, default order: 0, min order: 0 06-03 13:29:13.000 1010161 10686 10686 W node 0 : slabs: 57, objs: 2907, free: 0 06-03 13:29:13.000 10230 30892 30892 W SLUB : Unable to allocate memory on node -1, gfp=0x900(GFP_NOWAIT|__GFP_ZERO) 06-03 13:29:13.000 10230 30892 30892 W cache : avc_node, object size: 72, buffer size: 80, default order: 0, min order: 0 06-03 13:29:13.000 10230 30892 30892 W node 0 : slabs: 57, objs: 2907, free: 0 06-03 13:29:13.000 10230 30892 30892 W SLUB : Unable to allocate memory on node -1, gfp=0x900(GFP_NOWAIT|__GFP_ZERO) 06-03 13:29:13.000 10230 30892 30892 W cache : avc_node, object size: 72, buffer size: 80, default order: 0, min order: 0 Based on [1], selinux is tolerate for failure of memory allocation. Then, use __GFP_NOWARN together. [1] 476accbe2f6e ("selinux: use GFP_NOWAIT in the AVC kmem_caches") Signed-off-by: Minchan Kim <minchan@kernel.org> [PM: subj fix, line wraps, normalized commit refs] Signed-off-by: Paul Moore <paul@paul-moore.com>
2021-06-09 16:37:17 +00:00
GFP_NOWAIT | __GFP_NOWARN);
selinux: extended permissions for ioctls Add extended permissions logic to selinux. Extended permissions provides additional permissions in 256 bit increments. Extend the generic ioctl permission check to use the extended permissions for per-command filtering. Source/target/class sets including the ioctl permission may additionally include a set of commands. Example: allowxperm <source> <target>:<class> ioctl unpriv_app_socket_cmds auditallowxperm <source> <target>:<class> ioctl priv_gpu_cmds Where unpriv_app_socket_cmds and priv_gpu_cmds are macros representing commonly granted sets of ioctl commands. When ioctl commands are omitted only the permissions are checked. This feature is intended to provide finer granularity for the ioctl permission that may be too imprecise. For example, the same driver may use ioctls to provide important and benign functionality such as driver version or socket type as well as dangerous capabilities such as debugging features, read/write/execute to physical memory or access to sensitive data. Per-command filtering provides a mechanism to reduce the attack surface of the kernel, and limit applications to the subset of commands required. The format of the policy binary has been modified to include ioctl commands, and the policy version number has been incremented to POLICYDB_VERSION_XPERMS_IOCTL=30 to account for the format change. The extended permissions logic is deliberately generic to allow components to be reused e.g. netlink filters Signed-off-by: Jeff Vander Stoep <jeffv@google.com> Acked-by: Nick Kralevich <nnk@google.com> Signed-off-by: Paul Moore <pmoore@redhat.com>
2015-07-10 21:19:56 +00:00
if (!xpd->allowed)
goto error;
}
if (which & XPERMS_AUDITALLOW) {
xpd->auditallow = kmem_cache_zalloc(avc_xperms_data_cachep,
selinux: use __GFP_NOWARN with GFP_NOWAIT in the AVC In the field, we have seen lots of allocation failure from the call path below. 06-03 13:29:12.999 1010315 31557 31557 W Binder : 31542_2: page allocation failure: order:0, mode:0x800(GFP_NOWAIT), nodemask=(null),cpuset=background,mems_allowed=0 ... ... 06-03 13:29:12.999 1010315 31557 31557 W Call trace: 06-03 13:29:12.999 1010315 31557 31557 W : dump_backtrace.cfi_jt+0x0/0x8 06-03 13:29:12.999 1010315 31557 31557 W : dump_stack+0xc8/0x14c 06-03 13:29:12.999 1010315 31557 31557 W : warn_alloc+0x158/0x1c8 06-03 13:29:12.999 1010315 31557 31557 W : __alloc_pages_slowpath+0x9d8/0xb80 06-03 13:29:12.999 1010315 31557 31557 W : __alloc_pages_nodemask+0x1c4/0x430 06-03 13:29:12.999 1010315 31557 31557 W : allocate_slab+0xb4/0x390 06-03 13:29:12.999 1010315 31557 31557 W : ___slab_alloc+0x12c/0x3a4 06-03 13:29:12.999 1010315 31557 31557 W : kmem_cache_alloc+0x358/0x5e4 06-03 13:29:12.999 1010315 31557 31557 W : avc_alloc_node+0x30/0x184 06-03 13:29:12.999 1010315 31557 31557 W : avc_update_node+0x54/0x4f0 06-03 13:29:12.999 1010315 31557 31557 W : avc_has_extended_perms+0x1a4/0x460 06-03 13:29:12.999 1010315 31557 31557 W : selinux_file_ioctl+0x320/0x3d0 06-03 13:29:12.999 1010315 31557 31557 W : __arm64_sys_ioctl+0xec/0x1fc 06-03 13:29:12.999 1010315 31557 31557 W : el0_svc_common+0xc0/0x24c 06-03 13:29:12.999 1010315 31557 31557 W : el0_svc+0x28/0x88 06-03 13:29:12.999 1010315 31557 31557 W : el0_sync_handler+0x8c/0xf0 06-03 13:29:12.999 1010315 31557 31557 W : el0_sync+0x1a4/0x1c0 .. .. 06-03 13:29:12.999 1010315 31557 31557 W SLUB : Unable to allocate memory on node -1, gfp=0x900(GFP_NOWAIT|__GFP_ZERO) 06-03 13:29:12.999 1010315 31557 31557 W cache : avc_node, object size: 72, buffer size: 80, default order: 0, min order: 0 06-03 13:29:12.999 1010315 31557 31557 W node 0 : slabs: 57, objs: 2907, free: 0 06-03 13:29:12.999 1010161 10686 10686 W SLUB : Unable to allocate memory on node -1, gfp=0x900(GFP_NOWAIT|__GFP_ZERO) 06-03 13:29:12.999 1010161 10686 10686 W cache : avc_node, object size: 72, buffer size: 80, default order: 0, min order: 0 06-03 13:29:12.999 1010161 10686 10686 W node 0 : slabs: 57, objs: 2907, free: 0 06-03 13:29:12.999 1010161 10686 10686 W SLUB : Unable to allocate memory on node -1, gfp=0x900(GFP_NOWAIT|__GFP_ZERO) 06-03 13:29:12.999 1010161 10686 10686 W cache : avc_node, object size: 72, buffer size: 80, default order: 0, min order: 0 06-03 13:29:12.999 1010161 10686 10686 W node 0 : slabs: 57, objs: 2907, free: 0 06-03 13:29:12.999 1010161 10686 10686 W SLUB : Unable to allocate memory on node -1, gfp=0x900(GFP_NOWAIT|__GFP_ZERO) 06-03 13:29:12.999 1010161 10686 10686 W cache : avc_node, object size: 72, buffer size: 80, default order: 0, min order: 0 06-03 13:29:12.999 1010161 10686 10686 W node 0 : slabs: 57, objs: 2907, free: 0 06-03 13:29:13.000 1010161 10686 10686 W SLUB : Unable to allocate memory on node -1, gfp=0x900(GFP_NOWAIT|__GFP_ZERO) 06-03 13:29:13.000 1010161 10686 10686 W cache : avc_node, object size: 72, buffer size: 80, default order: 0, min order: 0 06-03 13:29:13.000 1010161 10686 10686 W node 0 : slabs: 57, objs: 2907, free: 0 06-03 13:29:13.000 1010161 10686 10686 W SLUB : Unable to allocate memory on node -1, gfp=0x900(GFP_NOWAIT|__GFP_ZERO) 06-03 13:29:13.000 1010161 10686 10686 W cache : avc_node, object size: 72, buffer size: 80, default order: 0, min order: 0 06-03 13:29:13.000 1010161 10686 10686 W node 0 : slabs: 57, objs: 2907, free: 0 06-03 13:29:13.000 1010161 10686 10686 W SLUB : Unable to allocate memory on node -1, gfp=0x900(GFP_NOWAIT|__GFP_ZERO) 06-03 13:29:13.000 1010161 10686 10686 W cache : avc_node, object size: 72, buffer size: 80, default order: 0, min order: 0 06-03 13:29:13.000 1010161 10686 10686 W node 0 : slabs: 57, objs: 2907, free: 0 06-03 13:29:13.000 10230 30892 30892 W SLUB : Unable to allocate memory on node -1, gfp=0x900(GFP_NOWAIT|__GFP_ZERO) 06-03 13:29:13.000 10230 30892 30892 W cache : avc_node, object size: 72, buffer size: 80, default order: 0, min order: 0 06-03 13:29:13.000 10230 30892 30892 W node 0 : slabs: 57, objs: 2907, free: 0 06-03 13:29:13.000 10230 30892 30892 W SLUB : Unable to allocate memory on node -1, gfp=0x900(GFP_NOWAIT|__GFP_ZERO) 06-03 13:29:13.000 10230 30892 30892 W cache : avc_node, object size: 72, buffer size: 80, default order: 0, min order: 0 Based on [1], selinux is tolerate for failure of memory allocation. Then, use __GFP_NOWARN together. [1] 476accbe2f6e ("selinux: use GFP_NOWAIT in the AVC kmem_caches") Signed-off-by: Minchan Kim <minchan@kernel.org> [PM: subj fix, line wraps, normalized commit refs] Signed-off-by: Paul Moore <paul@paul-moore.com>
2021-06-09 16:37:17 +00:00
GFP_NOWAIT | __GFP_NOWARN);
selinux: extended permissions for ioctls Add extended permissions logic to selinux. Extended permissions provides additional permissions in 256 bit increments. Extend the generic ioctl permission check to use the extended permissions for per-command filtering. Source/target/class sets including the ioctl permission may additionally include a set of commands. Example: allowxperm <source> <target>:<class> ioctl unpriv_app_socket_cmds auditallowxperm <source> <target>:<class> ioctl priv_gpu_cmds Where unpriv_app_socket_cmds and priv_gpu_cmds are macros representing commonly granted sets of ioctl commands. When ioctl commands are omitted only the permissions are checked. This feature is intended to provide finer granularity for the ioctl permission that may be too imprecise. For example, the same driver may use ioctls to provide important and benign functionality such as driver version or socket type as well as dangerous capabilities such as debugging features, read/write/execute to physical memory or access to sensitive data. Per-command filtering provides a mechanism to reduce the attack surface of the kernel, and limit applications to the subset of commands required. The format of the policy binary has been modified to include ioctl commands, and the policy version number has been incremented to POLICYDB_VERSION_XPERMS_IOCTL=30 to account for the format change. The extended permissions logic is deliberately generic to allow components to be reused e.g. netlink filters Signed-off-by: Jeff Vander Stoep <jeffv@google.com> Acked-by: Nick Kralevich <nnk@google.com> Signed-off-by: Paul Moore <pmoore@redhat.com>
2015-07-10 21:19:56 +00:00
if (!xpd->auditallow)
goto error;
}
if (which & XPERMS_DONTAUDIT) {
xpd->dontaudit = kmem_cache_zalloc(avc_xperms_data_cachep,
selinux: use __GFP_NOWARN with GFP_NOWAIT in the AVC In the field, we have seen lots of allocation failure from the call path below. 06-03 13:29:12.999 1010315 31557 31557 W Binder : 31542_2: page allocation failure: order:0, mode:0x800(GFP_NOWAIT), nodemask=(null),cpuset=background,mems_allowed=0 ... ... 06-03 13:29:12.999 1010315 31557 31557 W Call trace: 06-03 13:29:12.999 1010315 31557 31557 W : dump_backtrace.cfi_jt+0x0/0x8 06-03 13:29:12.999 1010315 31557 31557 W : dump_stack+0xc8/0x14c 06-03 13:29:12.999 1010315 31557 31557 W : warn_alloc+0x158/0x1c8 06-03 13:29:12.999 1010315 31557 31557 W : __alloc_pages_slowpath+0x9d8/0xb80 06-03 13:29:12.999 1010315 31557 31557 W : __alloc_pages_nodemask+0x1c4/0x430 06-03 13:29:12.999 1010315 31557 31557 W : allocate_slab+0xb4/0x390 06-03 13:29:12.999 1010315 31557 31557 W : ___slab_alloc+0x12c/0x3a4 06-03 13:29:12.999 1010315 31557 31557 W : kmem_cache_alloc+0x358/0x5e4 06-03 13:29:12.999 1010315 31557 31557 W : avc_alloc_node+0x30/0x184 06-03 13:29:12.999 1010315 31557 31557 W : avc_update_node+0x54/0x4f0 06-03 13:29:12.999 1010315 31557 31557 W : avc_has_extended_perms+0x1a4/0x460 06-03 13:29:12.999 1010315 31557 31557 W : selinux_file_ioctl+0x320/0x3d0 06-03 13:29:12.999 1010315 31557 31557 W : __arm64_sys_ioctl+0xec/0x1fc 06-03 13:29:12.999 1010315 31557 31557 W : el0_svc_common+0xc0/0x24c 06-03 13:29:12.999 1010315 31557 31557 W : el0_svc+0x28/0x88 06-03 13:29:12.999 1010315 31557 31557 W : el0_sync_handler+0x8c/0xf0 06-03 13:29:12.999 1010315 31557 31557 W : el0_sync+0x1a4/0x1c0 .. .. 06-03 13:29:12.999 1010315 31557 31557 W SLUB : Unable to allocate memory on node -1, gfp=0x900(GFP_NOWAIT|__GFP_ZERO) 06-03 13:29:12.999 1010315 31557 31557 W cache : avc_node, object size: 72, buffer size: 80, default order: 0, min order: 0 06-03 13:29:12.999 1010315 31557 31557 W node 0 : slabs: 57, objs: 2907, free: 0 06-03 13:29:12.999 1010161 10686 10686 W SLUB : Unable to allocate memory on node -1, gfp=0x900(GFP_NOWAIT|__GFP_ZERO) 06-03 13:29:12.999 1010161 10686 10686 W cache : avc_node, object size: 72, buffer size: 80, default order: 0, min order: 0 06-03 13:29:12.999 1010161 10686 10686 W node 0 : slabs: 57, objs: 2907, free: 0 06-03 13:29:12.999 1010161 10686 10686 W SLUB : Unable to allocate memory on node -1, gfp=0x900(GFP_NOWAIT|__GFP_ZERO) 06-03 13:29:12.999 1010161 10686 10686 W cache : avc_node, object size: 72, buffer size: 80, default order: 0, min order: 0 06-03 13:29:12.999 1010161 10686 10686 W node 0 : slabs: 57, objs: 2907, free: 0 06-03 13:29:12.999 1010161 10686 10686 W SLUB : Unable to allocate memory on node -1, gfp=0x900(GFP_NOWAIT|__GFP_ZERO) 06-03 13:29:12.999 1010161 10686 10686 W cache : avc_node, object size: 72, buffer size: 80, default order: 0, min order: 0 06-03 13:29:12.999 1010161 10686 10686 W node 0 : slabs: 57, objs: 2907, free: 0 06-03 13:29:13.000 1010161 10686 10686 W SLUB : Unable to allocate memory on node -1, gfp=0x900(GFP_NOWAIT|__GFP_ZERO) 06-03 13:29:13.000 1010161 10686 10686 W cache : avc_node, object size: 72, buffer size: 80, default order: 0, min order: 0 06-03 13:29:13.000 1010161 10686 10686 W node 0 : slabs: 57, objs: 2907, free: 0 06-03 13:29:13.000 1010161 10686 10686 W SLUB : Unable to allocate memory on node -1, gfp=0x900(GFP_NOWAIT|__GFP_ZERO) 06-03 13:29:13.000 1010161 10686 10686 W cache : avc_node, object size: 72, buffer size: 80, default order: 0, min order: 0 06-03 13:29:13.000 1010161 10686 10686 W node 0 : slabs: 57, objs: 2907, free: 0 06-03 13:29:13.000 1010161 10686 10686 W SLUB : Unable to allocate memory on node -1, gfp=0x900(GFP_NOWAIT|__GFP_ZERO) 06-03 13:29:13.000 1010161 10686 10686 W cache : avc_node, object size: 72, buffer size: 80, default order: 0, min order: 0 06-03 13:29:13.000 1010161 10686 10686 W node 0 : slabs: 57, objs: 2907, free: 0 06-03 13:29:13.000 10230 30892 30892 W SLUB : Unable to allocate memory on node -1, gfp=0x900(GFP_NOWAIT|__GFP_ZERO) 06-03 13:29:13.000 10230 30892 30892 W cache : avc_node, object size: 72, buffer size: 80, default order: 0, min order: 0 06-03 13:29:13.000 10230 30892 30892 W node 0 : slabs: 57, objs: 2907, free: 0 06-03 13:29:13.000 10230 30892 30892 W SLUB : Unable to allocate memory on node -1, gfp=0x900(GFP_NOWAIT|__GFP_ZERO) 06-03 13:29:13.000 10230 30892 30892 W cache : avc_node, object size: 72, buffer size: 80, default order: 0, min order: 0 Based on [1], selinux is tolerate for failure of memory allocation. Then, use __GFP_NOWARN together. [1] 476accbe2f6e ("selinux: use GFP_NOWAIT in the AVC kmem_caches") Signed-off-by: Minchan Kim <minchan@kernel.org> [PM: subj fix, line wraps, normalized commit refs] Signed-off-by: Paul Moore <paul@paul-moore.com>
2021-06-09 16:37:17 +00:00
GFP_NOWAIT | __GFP_NOWARN);
selinux: extended permissions for ioctls Add extended permissions logic to selinux. Extended permissions provides additional permissions in 256 bit increments. Extend the generic ioctl permission check to use the extended permissions for per-command filtering. Source/target/class sets including the ioctl permission may additionally include a set of commands. Example: allowxperm <source> <target>:<class> ioctl unpriv_app_socket_cmds auditallowxperm <source> <target>:<class> ioctl priv_gpu_cmds Where unpriv_app_socket_cmds and priv_gpu_cmds are macros representing commonly granted sets of ioctl commands. When ioctl commands are omitted only the permissions are checked. This feature is intended to provide finer granularity for the ioctl permission that may be too imprecise. For example, the same driver may use ioctls to provide important and benign functionality such as driver version or socket type as well as dangerous capabilities such as debugging features, read/write/execute to physical memory or access to sensitive data. Per-command filtering provides a mechanism to reduce the attack surface of the kernel, and limit applications to the subset of commands required. The format of the policy binary has been modified to include ioctl commands, and the policy version number has been incremented to POLICYDB_VERSION_XPERMS_IOCTL=30 to account for the format change. The extended permissions logic is deliberately generic to allow components to be reused e.g. netlink filters Signed-off-by: Jeff Vander Stoep <jeffv@google.com> Acked-by: Nick Kralevich <nnk@google.com> Signed-off-by: Paul Moore <pmoore@redhat.com>
2015-07-10 21:19:56 +00:00
if (!xpd->dontaudit)
goto error;
}
return xpd_node;
error:
avc_xperms_decision_free(xpd_node);
return NULL;
}
static int avc_add_xperms_decision(struct avc_node *node,
struct extended_perms_decision *src)
{
struct avc_xperms_decision_node *dest_xpd;
node->ae.xp_node->xp.len++;
dest_xpd = avc_xperms_decision_alloc(src->used);
if (!dest_xpd)
return -ENOMEM;
avc_copy_xperms_decision(&dest_xpd->xpd, src);
list_add(&dest_xpd->xpd_list, &node->ae.xp_node->xpd_head);
return 0;
}
static struct avc_xperms_node *avc_xperms_alloc(void)
{
struct avc_xperms_node *xp_node;
selinux: use __GFP_NOWARN with GFP_NOWAIT in the AVC In the field, we have seen lots of allocation failure from the call path below. 06-03 13:29:12.999 1010315 31557 31557 W Binder : 31542_2: page allocation failure: order:0, mode:0x800(GFP_NOWAIT), nodemask=(null),cpuset=background,mems_allowed=0 ... ... 06-03 13:29:12.999 1010315 31557 31557 W Call trace: 06-03 13:29:12.999 1010315 31557 31557 W : dump_backtrace.cfi_jt+0x0/0x8 06-03 13:29:12.999 1010315 31557 31557 W : dump_stack+0xc8/0x14c 06-03 13:29:12.999 1010315 31557 31557 W : warn_alloc+0x158/0x1c8 06-03 13:29:12.999 1010315 31557 31557 W : __alloc_pages_slowpath+0x9d8/0xb80 06-03 13:29:12.999 1010315 31557 31557 W : __alloc_pages_nodemask+0x1c4/0x430 06-03 13:29:12.999 1010315 31557 31557 W : allocate_slab+0xb4/0x390 06-03 13:29:12.999 1010315 31557 31557 W : ___slab_alloc+0x12c/0x3a4 06-03 13:29:12.999 1010315 31557 31557 W : kmem_cache_alloc+0x358/0x5e4 06-03 13:29:12.999 1010315 31557 31557 W : avc_alloc_node+0x30/0x184 06-03 13:29:12.999 1010315 31557 31557 W : avc_update_node+0x54/0x4f0 06-03 13:29:12.999 1010315 31557 31557 W : avc_has_extended_perms+0x1a4/0x460 06-03 13:29:12.999 1010315 31557 31557 W : selinux_file_ioctl+0x320/0x3d0 06-03 13:29:12.999 1010315 31557 31557 W : __arm64_sys_ioctl+0xec/0x1fc 06-03 13:29:12.999 1010315 31557 31557 W : el0_svc_common+0xc0/0x24c 06-03 13:29:12.999 1010315 31557 31557 W : el0_svc+0x28/0x88 06-03 13:29:12.999 1010315 31557 31557 W : el0_sync_handler+0x8c/0xf0 06-03 13:29:12.999 1010315 31557 31557 W : el0_sync+0x1a4/0x1c0 .. .. 06-03 13:29:12.999 1010315 31557 31557 W SLUB : Unable to allocate memory on node -1, gfp=0x900(GFP_NOWAIT|__GFP_ZERO) 06-03 13:29:12.999 1010315 31557 31557 W cache : avc_node, object size: 72, buffer size: 80, default order: 0, min order: 0 06-03 13:29:12.999 1010315 31557 31557 W node 0 : slabs: 57, objs: 2907, free: 0 06-03 13:29:12.999 1010161 10686 10686 W SLUB : Unable to allocate memory on node -1, gfp=0x900(GFP_NOWAIT|__GFP_ZERO) 06-03 13:29:12.999 1010161 10686 10686 W cache : avc_node, object size: 72, buffer size: 80, default order: 0, min order: 0 06-03 13:29:12.999 1010161 10686 10686 W node 0 : slabs: 57, objs: 2907, free: 0 06-03 13:29:12.999 1010161 10686 10686 W SLUB : Unable to allocate memory on node -1, gfp=0x900(GFP_NOWAIT|__GFP_ZERO) 06-03 13:29:12.999 1010161 10686 10686 W cache : avc_node, object size: 72, buffer size: 80, default order: 0, min order: 0 06-03 13:29:12.999 1010161 10686 10686 W node 0 : slabs: 57, objs: 2907, free: 0 06-03 13:29:12.999 1010161 10686 10686 W SLUB : Unable to allocate memory on node -1, gfp=0x900(GFP_NOWAIT|__GFP_ZERO) 06-03 13:29:12.999 1010161 10686 10686 W cache : avc_node, object size: 72, buffer size: 80, default order: 0, min order: 0 06-03 13:29:12.999 1010161 10686 10686 W node 0 : slabs: 57, objs: 2907, free: 0 06-03 13:29:13.000 1010161 10686 10686 W SLUB : Unable to allocate memory on node -1, gfp=0x900(GFP_NOWAIT|__GFP_ZERO) 06-03 13:29:13.000 1010161 10686 10686 W cache : avc_node, object size: 72, buffer size: 80, default order: 0, min order: 0 06-03 13:29:13.000 1010161 10686 10686 W node 0 : slabs: 57, objs: 2907, free: 0 06-03 13:29:13.000 1010161 10686 10686 W SLUB : Unable to allocate memory on node -1, gfp=0x900(GFP_NOWAIT|__GFP_ZERO) 06-03 13:29:13.000 1010161 10686 10686 W cache : avc_node, object size: 72, buffer size: 80, default order: 0, min order: 0 06-03 13:29:13.000 1010161 10686 10686 W node 0 : slabs: 57, objs: 2907, free: 0 06-03 13:29:13.000 1010161 10686 10686 W SLUB : Unable to allocate memory on node -1, gfp=0x900(GFP_NOWAIT|__GFP_ZERO) 06-03 13:29:13.000 1010161 10686 10686 W cache : avc_node, object size: 72, buffer size: 80, default order: 0, min order: 0 06-03 13:29:13.000 1010161 10686 10686 W node 0 : slabs: 57, objs: 2907, free: 0 06-03 13:29:13.000 10230 30892 30892 W SLUB : Unable to allocate memory on node -1, gfp=0x900(GFP_NOWAIT|__GFP_ZERO) 06-03 13:29:13.000 10230 30892 30892 W cache : avc_node, object size: 72, buffer size: 80, default order: 0, min order: 0 06-03 13:29:13.000 10230 30892 30892 W node 0 : slabs: 57, objs: 2907, free: 0 06-03 13:29:13.000 10230 30892 30892 W SLUB : Unable to allocate memory on node -1, gfp=0x900(GFP_NOWAIT|__GFP_ZERO) 06-03 13:29:13.000 10230 30892 30892 W cache : avc_node, object size: 72, buffer size: 80, default order: 0, min order: 0 Based on [1], selinux is tolerate for failure of memory allocation. Then, use __GFP_NOWARN together. [1] 476accbe2f6e ("selinux: use GFP_NOWAIT in the AVC kmem_caches") Signed-off-by: Minchan Kim <minchan@kernel.org> [PM: subj fix, line wraps, normalized commit refs] Signed-off-by: Paul Moore <paul@paul-moore.com>
2021-06-09 16:37:17 +00:00
xp_node = kmem_cache_zalloc(avc_xperms_cachep, GFP_NOWAIT | __GFP_NOWARN);
selinux: extended permissions for ioctls Add extended permissions logic to selinux. Extended permissions provides additional permissions in 256 bit increments. Extend the generic ioctl permission check to use the extended permissions for per-command filtering. Source/target/class sets including the ioctl permission may additionally include a set of commands. Example: allowxperm <source> <target>:<class> ioctl unpriv_app_socket_cmds auditallowxperm <source> <target>:<class> ioctl priv_gpu_cmds Where unpriv_app_socket_cmds and priv_gpu_cmds are macros representing commonly granted sets of ioctl commands. When ioctl commands are omitted only the permissions are checked. This feature is intended to provide finer granularity for the ioctl permission that may be too imprecise. For example, the same driver may use ioctls to provide important and benign functionality such as driver version or socket type as well as dangerous capabilities such as debugging features, read/write/execute to physical memory or access to sensitive data. Per-command filtering provides a mechanism to reduce the attack surface of the kernel, and limit applications to the subset of commands required. The format of the policy binary has been modified to include ioctl commands, and the policy version number has been incremented to POLICYDB_VERSION_XPERMS_IOCTL=30 to account for the format change. The extended permissions logic is deliberately generic to allow components to be reused e.g. netlink filters Signed-off-by: Jeff Vander Stoep <jeffv@google.com> Acked-by: Nick Kralevich <nnk@google.com> Signed-off-by: Paul Moore <pmoore@redhat.com>
2015-07-10 21:19:56 +00:00
if (!xp_node)
return xp_node;
INIT_LIST_HEAD(&xp_node->xpd_head);
return xp_node;
}
static int avc_xperms_populate(struct avc_node *node,
struct avc_xperms_node *src)
{
struct avc_xperms_node *dest;
struct avc_xperms_decision_node *dest_xpd;
struct avc_xperms_decision_node *src_xpd;
if (src->xp.len == 0)
return 0;
dest = avc_xperms_alloc();
if (!dest)
return -ENOMEM;
memcpy(dest->xp.drivers.p, src->xp.drivers.p, sizeof(dest->xp.drivers.p));
dest->xp.len = src->xp.len;
/* for each source xpd allocate a destination xpd and copy */
list_for_each_entry(src_xpd, &src->xpd_head, xpd_list) {
dest_xpd = avc_xperms_decision_alloc(src_xpd->xpd.used);
if (!dest_xpd)
goto error;
avc_copy_xperms_decision(&dest_xpd->xpd, &src_xpd->xpd);
list_add(&dest_xpd->xpd_list, &dest->xpd_head);
}
node->ae.xp_node = dest;
return 0;
error:
avc_xperms_free(dest);
return -ENOMEM;
}
static inline u32 avc_xperms_audit_required(u32 requested,
struct av_decision *avd,
struct extended_perms_decision *xpd,
u8 perm,
int result,
u32 *deniedp)
{
u32 denied, audited;
denied = requested & ~avd->allowed;
if (unlikely(denied)) {
audited = denied & avd->auditdeny;
if (audited && xpd) {
if (avc_xperms_has_perm(xpd, perm, XPERMS_DONTAUDIT))
audited &= ~requested;
}
} else if (result) {
audited = denied = requested;
} else {
audited = requested & avd->auditallow;
if (audited && xpd) {
if (!avc_xperms_has_perm(xpd, perm, XPERMS_AUDITALLOW))
audited &= ~requested;
}
}
*deniedp = denied;
return audited;
}
static inline int avc_xperms_audit(u32 ssid, u32 tsid, u16 tclass,
u32 requested, struct av_decision *avd,
struct extended_perms_decision *xpd,
u8 perm, int result,
struct common_audit_data *ad)
selinux: extended permissions for ioctls Add extended permissions logic to selinux. Extended permissions provides additional permissions in 256 bit increments. Extend the generic ioctl permission check to use the extended permissions for per-command filtering. Source/target/class sets including the ioctl permission may additionally include a set of commands. Example: allowxperm <source> <target>:<class> ioctl unpriv_app_socket_cmds auditallowxperm <source> <target>:<class> ioctl priv_gpu_cmds Where unpriv_app_socket_cmds and priv_gpu_cmds are macros representing commonly granted sets of ioctl commands. When ioctl commands are omitted only the permissions are checked. This feature is intended to provide finer granularity for the ioctl permission that may be too imprecise. For example, the same driver may use ioctls to provide important and benign functionality such as driver version or socket type as well as dangerous capabilities such as debugging features, read/write/execute to physical memory or access to sensitive data. Per-command filtering provides a mechanism to reduce the attack surface of the kernel, and limit applications to the subset of commands required. The format of the policy binary has been modified to include ioctl commands, and the policy version number has been incremented to POLICYDB_VERSION_XPERMS_IOCTL=30 to account for the format change. The extended permissions logic is deliberately generic to allow components to be reused e.g. netlink filters Signed-off-by: Jeff Vander Stoep <jeffv@google.com> Acked-by: Nick Kralevich <nnk@google.com> Signed-off-by: Paul Moore <pmoore@redhat.com>
2015-07-10 21:19:56 +00:00
{
u32 audited, denied;
audited = avc_xperms_audit_required(
requested, avd, xpd, perm, result, &denied);
if (likely(!audited))
return 0;
return slow_avc_audit(ssid, tsid, tclass, requested,
audited, denied, result, ad);
selinux: extended permissions for ioctls Add extended permissions logic to selinux. Extended permissions provides additional permissions in 256 bit increments. Extend the generic ioctl permission check to use the extended permissions for per-command filtering. Source/target/class sets including the ioctl permission may additionally include a set of commands. Example: allowxperm <source> <target>:<class> ioctl unpriv_app_socket_cmds auditallowxperm <source> <target>:<class> ioctl priv_gpu_cmds Where unpriv_app_socket_cmds and priv_gpu_cmds are macros representing commonly granted sets of ioctl commands. When ioctl commands are omitted only the permissions are checked. This feature is intended to provide finer granularity for the ioctl permission that may be too imprecise. For example, the same driver may use ioctls to provide important and benign functionality such as driver version or socket type as well as dangerous capabilities such as debugging features, read/write/execute to physical memory or access to sensitive data. Per-command filtering provides a mechanism to reduce the attack surface of the kernel, and limit applications to the subset of commands required. The format of the policy binary has been modified to include ioctl commands, and the policy version number has been incremented to POLICYDB_VERSION_XPERMS_IOCTL=30 to account for the format change. The extended permissions logic is deliberately generic to allow components to be reused e.g. netlink filters Signed-off-by: Jeff Vander Stoep <jeffv@google.com> Acked-by: Nick Kralevich <nnk@google.com> Signed-off-by: Paul Moore <pmoore@redhat.com>
2015-07-10 21:19:56 +00:00
}
static void avc_node_free(struct rcu_head *rhead)
{
struct avc_node *node = container_of(rhead, struct avc_node, rhead);
selinux: extended permissions for ioctls Add extended permissions logic to selinux. Extended permissions provides additional permissions in 256 bit increments. Extend the generic ioctl permission check to use the extended permissions for per-command filtering. Source/target/class sets including the ioctl permission may additionally include a set of commands. Example: allowxperm <source> <target>:<class> ioctl unpriv_app_socket_cmds auditallowxperm <source> <target>:<class> ioctl priv_gpu_cmds Where unpriv_app_socket_cmds and priv_gpu_cmds are macros representing commonly granted sets of ioctl commands. When ioctl commands are omitted only the permissions are checked. This feature is intended to provide finer granularity for the ioctl permission that may be too imprecise. For example, the same driver may use ioctls to provide important and benign functionality such as driver version or socket type as well as dangerous capabilities such as debugging features, read/write/execute to physical memory or access to sensitive data. Per-command filtering provides a mechanism to reduce the attack surface of the kernel, and limit applications to the subset of commands required. The format of the policy binary has been modified to include ioctl commands, and the policy version number has been incremented to POLICYDB_VERSION_XPERMS_IOCTL=30 to account for the format change. The extended permissions logic is deliberately generic to allow components to be reused e.g. netlink filters Signed-off-by: Jeff Vander Stoep <jeffv@google.com> Acked-by: Nick Kralevich <nnk@google.com> Signed-off-by: Paul Moore <pmoore@redhat.com>
2015-07-10 21:19:56 +00:00
avc_xperms_free(node->ae.xp_node);
kmem_cache_free(avc_node_cachep, node);
avc_cache_stats_incr(frees);
}
static void avc_node_delete(struct avc_node *node)
{
hlist_del_rcu(&node->list);
call_rcu(&node->rhead, avc_node_free);
atomic_dec(&selinux_avc.avc_cache.active_nodes);
}
static void avc_node_kill(struct avc_node *node)
{
selinux: extended permissions for ioctls Add extended permissions logic to selinux. Extended permissions provides additional permissions in 256 bit increments. Extend the generic ioctl permission check to use the extended permissions for per-command filtering. Source/target/class sets including the ioctl permission may additionally include a set of commands. Example: allowxperm <source> <target>:<class> ioctl unpriv_app_socket_cmds auditallowxperm <source> <target>:<class> ioctl priv_gpu_cmds Where unpriv_app_socket_cmds and priv_gpu_cmds are macros representing commonly granted sets of ioctl commands. When ioctl commands are omitted only the permissions are checked. This feature is intended to provide finer granularity for the ioctl permission that may be too imprecise. For example, the same driver may use ioctls to provide important and benign functionality such as driver version or socket type as well as dangerous capabilities such as debugging features, read/write/execute to physical memory or access to sensitive data. Per-command filtering provides a mechanism to reduce the attack surface of the kernel, and limit applications to the subset of commands required. The format of the policy binary has been modified to include ioctl commands, and the policy version number has been incremented to POLICYDB_VERSION_XPERMS_IOCTL=30 to account for the format change. The extended permissions logic is deliberately generic to allow components to be reused e.g. netlink filters Signed-off-by: Jeff Vander Stoep <jeffv@google.com> Acked-by: Nick Kralevich <nnk@google.com> Signed-off-by: Paul Moore <pmoore@redhat.com>
2015-07-10 21:19:56 +00:00
avc_xperms_free(node->ae.xp_node);
kmem_cache_free(avc_node_cachep, node);
avc_cache_stats_incr(frees);
atomic_dec(&selinux_avc.avc_cache.active_nodes);
}
static void avc_node_replace(struct avc_node *new, struct avc_node *old)
{
hlist_replace_rcu(&old->list, &new->list);
call_rcu(&old->rhead, avc_node_free);
atomic_dec(&selinux_avc.avc_cache.active_nodes);
}
static inline int avc_reclaim_node(void)
{
struct avc_node *node;
int hvalue, try, ecx;
unsigned long flags;
struct hlist_head *head;
spinlock_t *lock;
for (try = 0, ecx = 0; try < AVC_CACHE_SLOTS; try++) {
hvalue = atomic_inc_return(&selinux_avc.avc_cache.lru_hint) &
(AVC_CACHE_SLOTS - 1);
head = &selinux_avc.avc_cache.slots[hvalue];
lock = &selinux_avc.avc_cache.slots_lock[hvalue];
if (!spin_trylock_irqsave(lock, flags))
continue;
rcu_read_lock();
hlist: drop the node parameter from iterators I'm not sure why, but the hlist for each entry iterators were conceived list_for_each_entry(pos, head, member) The hlist ones were greedy and wanted an extra parameter: hlist_for_each_entry(tpos, pos, head, member) Why did they need an extra pos parameter? I'm not quite sure. Not only they don't really need it, it also prevents the iterator from looking exactly like the list iterator, which is unfortunate. Besides the semantic patch, there was some manual work required: - Fix up the actual hlist iterators in linux/list.h - Fix up the declaration of other iterators based on the hlist ones. - A very small amount of places were using the 'node' parameter, this was modified to use 'obj->member' instead. - Coccinelle didn't handle the hlist_for_each_entry_safe iterator properly, so those had to be fixed up manually. The semantic patch which is mostly the work of Peter Senna Tschudin is here: @@ iterator name hlist_for_each_entry, hlist_for_each_entry_continue, hlist_for_each_entry_from, hlist_for_each_entry_rcu, hlist_for_each_entry_rcu_bh, hlist_for_each_entry_continue_rcu_bh, for_each_busy_worker, ax25_uid_for_each, ax25_for_each, inet_bind_bucket_for_each, sctp_for_each_hentry, sk_for_each, sk_for_each_rcu, sk_for_each_from, sk_for_each_safe, sk_for_each_bound, hlist_for_each_entry_safe, hlist_for_each_entry_continue_rcu, nr_neigh_for_each, nr_neigh_for_each_safe, nr_node_for_each, nr_node_for_each_safe, for_each_gfn_indirect_valid_sp, for_each_gfn_sp, for_each_host; type T; expression a,c,d,e; identifier b; statement S; @@ -T b; <+... when != b ( hlist_for_each_entry(a, - b, c, d) S | hlist_for_each_entry_continue(a, - b, c) S | hlist_for_each_entry_from(a, - b, c) S | hlist_for_each_entry_rcu(a, - b, c, d) S | hlist_for_each_entry_rcu_bh(a, - b, c, d) S | hlist_for_each_entry_continue_rcu_bh(a, - b, c) S | for_each_busy_worker(a, c, - b, d) S | ax25_uid_for_each(a, - b, c) S | ax25_for_each(a, - b, c) S | inet_bind_bucket_for_each(a, - b, c) S | sctp_for_each_hentry(a, - b, c) S | sk_for_each(a, - b, c) S | sk_for_each_rcu(a, - b, c) S | sk_for_each_from -(a, b) +(a) S + sk_for_each_from(a) S | sk_for_each_safe(a, - b, c, d) S | sk_for_each_bound(a, - b, c) S | hlist_for_each_entry_safe(a, - b, c, d, e) S | hlist_for_each_entry_continue_rcu(a, - b, c) S | nr_neigh_for_each(a, - b, c) S | nr_neigh_for_each_safe(a, - b, c, d) S | nr_node_for_each(a, - b, c) S | nr_node_for_each_safe(a, - b, c, d) S | - for_each_gfn_sp(a, c, d, b) S + for_each_gfn_sp(a, c, d) S | - for_each_gfn_indirect_valid_sp(a, c, d, b) S + for_each_gfn_indirect_valid_sp(a, c, d) S | for_each_host(a, - b, c) S | for_each_host_safe(a, - b, c, d) S | for_each_mesh_entry(a, - b, c, d) S ) ...+> [akpm@linux-foundation.org: drop bogus change from net/ipv4/raw.c] [akpm@linux-foundation.org: drop bogus hunk from net/ipv6/raw.c] [akpm@linux-foundation.org: checkpatch fixes] [akpm@linux-foundation.org: fix warnings] [akpm@linux-foudnation.org: redo intrusive kvm changes] Tested-by: Peter Senna Tschudin <peter.senna@gmail.com> Acked-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com> Signed-off-by: Sasha Levin <sasha.levin@oracle.com> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Marcelo Tosatti <mtosatti@redhat.com> Cc: Gleb Natapov <gleb@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-02-28 01:06:00 +00:00
hlist_for_each_entry(node, head, list) {
avc_node_delete(node);
avc_cache_stats_incr(reclaims);
ecx++;
if (ecx >= AVC_CACHE_RECLAIM) {
rcu_read_unlock();
spin_unlock_irqrestore(lock, flags);
goto out;
}
}
rcu_read_unlock();
spin_unlock_irqrestore(lock, flags);
}
out:
return ecx;
}
static struct avc_node *avc_alloc_node(void)
{
struct avc_node *node;
selinux: use __GFP_NOWARN with GFP_NOWAIT in the AVC In the field, we have seen lots of allocation failure from the call path below. 06-03 13:29:12.999 1010315 31557 31557 W Binder : 31542_2: page allocation failure: order:0, mode:0x800(GFP_NOWAIT), nodemask=(null),cpuset=background,mems_allowed=0 ... ... 06-03 13:29:12.999 1010315 31557 31557 W Call trace: 06-03 13:29:12.999 1010315 31557 31557 W : dump_backtrace.cfi_jt+0x0/0x8 06-03 13:29:12.999 1010315 31557 31557 W : dump_stack+0xc8/0x14c 06-03 13:29:12.999 1010315 31557 31557 W : warn_alloc+0x158/0x1c8 06-03 13:29:12.999 1010315 31557 31557 W : __alloc_pages_slowpath+0x9d8/0xb80 06-03 13:29:12.999 1010315 31557 31557 W : __alloc_pages_nodemask+0x1c4/0x430 06-03 13:29:12.999 1010315 31557 31557 W : allocate_slab+0xb4/0x390 06-03 13:29:12.999 1010315 31557 31557 W : ___slab_alloc+0x12c/0x3a4 06-03 13:29:12.999 1010315 31557 31557 W : kmem_cache_alloc+0x358/0x5e4 06-03 13:29:12.999 1010315 31557 31557 W : avc_alloc_node+0x30/0x184 06-03 13:29:12.999 1010315 31557 31557 W : avc_update_node+0x54/0x4f0 06-03 13:29:12.999 1010315 31557 31557 W : avc_has_extended_perms+0x1a4/0x460 06-03 13:29:12.999 1010315 31557 31557 W : selinux_file_ioctl+0x320/0x3d0 06-03 13:29:12.999 1010315 31557 31557 W : __arm64_sys_ioctl+0xec/0x1fc 06-03 13:29:12.999 1010315 31557 31557 W : el0_svc_common+0xc0/0x24c 06-03 13:29:12.999 1010315 31557 31557 W : el0_svc+0x28/0x88 06-03 13:29:12.999 1010315 31557 31557 W : el0_sync_handler+0x8c/0xf0 06-03 13:29:12.999 1010315 31557 31557 W : el0_sync+0x1a4/0x1c0 .. .. 06-03 13:29:12.999 1010315 31557 31557 W SLUB : Unable to allocate memory on node -1, gfp=0x900(GFP_NOWAIT|__GFP_ZERO) 06-03 13:29:12.999 1010315 31557 31557 W cache : avc_node, object size: 72, buffer size: 80, default order: 0, min order: 0 06-03 13:29:12.999 1010315 31557 31557 W node 0 : slabs: 57, objs: 2907, free: 0 06-03 13:29:12.999 1010161 10686 10686 W SLUB : Unable to allocate memory on node -1, gfp=0x900(GFP_NOWAIT|__GFP_ZERO) 06-03 13:29:12.999 1010161 10686 10686 W cache : avc_node, object size: 72, buffer size: 80, default order: 0, min order: 0 06-03 13:29:12.999 1010161 10686 10686 W node 0 : slabs: 57, objs: 2907, free: 0 06-03 13:29:12.999 1010161 10686 10686 W SLUB : Unable to allocate memory on node -1, gfp=0x900(GFP_NOWAIT|__GFP_ZERO) 06-03 13:29:12.999 1010161 10686 10686 W cache : avc_node, object size: 72, buffer size: 80, default order: 0, min order: 0 06-03 13:29:12.999 1010161 10686 10686 W node 0 : slabs: 57, objs: 2907, free: 0 06-03 13:29:12.999 1010161 10686 10686 W SLUB : Unable to allocate memory on node -1, gfp=0x900(GFP_NOWAIT|__GFP_ZERO) 06-03 13:29:12.999 1010161 10686 10686 W cache : avc_node, object size: 72, buffer size: 80, default order: 0, min order: 0 06-03 13:29:12.999 1010161 10686 10686 W node 0 : slabs: 57, objs: 2907, free: 0 06-03 13:29:13.000 1010161 10686 10686 W SLUB : Unable to allocate memory on node -1, gfp=0x900(GFP_NOWAIT|__GFP_ZERO) 06-03 13:29:13.000 1010161 10686 10686 W cache : avc_node, object size: 72, buffer size: 80, default order: 0, min order: 0 06-03 13:29:13.000 1010161 10686 10686 W node 0 : slabs: 57, objs: 2907, free: 0 06-03 13:29:13.000 1010161 10686 10686 W SLUB : Unable to allocate memory on node -1, gfp=0x900(GFP_NOWAIT|__GFP_ZERO) 06-03 13:29:13.000 1010161 10686 10686 W cache : avc_node, object size: 72, buffer size: 80, default order: 0, min order: 0 06-03 13:29:13.000 1010161 10686 10686 W node 0 : slabs: 57, objs: 2907, free: 0 06-03 13:29:13.000 1010161 10686 10686 W SLUB : Unable to allocate memory on node -1, gfp=0x900(GFP_NOWAIT|__GFP_ZERO) 06-03 13:29:13.000 1010161 10686 10686 W cache : avc_node, object size: 72, buffer size: 80, default order: 0, min order: 0 06-03 13:29:13.000 1010161 10686 10686 W node 0 : slabs: 57, objs: 2907, free: 0 06-03 13:29:13.000 10230 30892 30892 W SLUB : Unable to allocate memory on node -1, gfp=0x900(GFP_NOWAIT|__GFP_ZERO) 06-03 13:29:13.000 10230 30892 30892 W cache : avc_node, object size: 72, buffer size: 80, default order: 0, min order: 0 06-03 13:29:13.000 10230 30892 30892 W node 0 : slabs: 57, objs: 2907, free: 0 06-03 13:29:13.000 10230 30892 30892 W SLUB : Unable to allocate memory on node -1, gfp=0x900(GFP_NOWAIT|__GFP_ZERO) 06-03 13:29:13.000 10230 30892 30892 W cache : avc_node, object size: 72, buffer size: 80, default order: 0, min order: 0 Based on [1], selinux is tolerate for failure of memory allocation. Then, use __GFP_NOWARN together. [1] 476accbe2f6e ("selinux: use GFP_NOWAIT in the AVC kmem_caches") Signed-off-by: Minchan Kim <minchan@kernel.org> [PM: subj fix, line wraps, normalized commit refs] Signed-off-by: Paul Moore <paul@paul-moore.com>
2021-06-09 16:37:17 +00:00
node = kmem_cache_zalloc(avc_node_cachep, GFP_NOWAIT | __GFP_NOWARN);
if (!node)
goto out;
INIT_HLIST_NODE(&node->list);
avc_cache_stats_incr(allocations);
if (atomic_inc_return(&selinux_avc.avc_cache.active_nodes) >
selinux_avc.avc_cache_threshold)
avc_reclaim_node();
out:
return node;
}
static void avc_node_populate(struct avc_node *node, u32 ssid, u32 tsid, u16 tclass, struct av_decision *avd)
{
node->ae.ssid = ssid;
node->ae.tsid = tsid;
node->ae.tclass = tclass;
memcpy(&node->ae.avd, avd, sizeof(node->ae.avd));
}
static inline struct avc_node *avc_search_node(u32 ssid, u32 tsid, u16 tclass)
{
struct avc_node *node, *ret = NULL;
int hvalue;
struct hlist_head *head;
hvalue = avc_hash(ssid, tsid, tclass);
head = &selinux_avc.avc_cache.slots[hvalue];
hlist: drop the node parameter from iterators I'm not sure why, but the hlist for each entry iterators were conceived list_for_each_entry(pos, head, member) The hlist ones were greedy and wanted an extra parameter: hlist_for_each_entry(tpos, pos, head, member) Why did they need an extra pos parameter? I'm not quite sure. Not only they don't really need it, it also prevents the iterator from looking exactly like the list iterator, which is unfortunate. Besides the semantic patch, there was some manual work required: - Fix up the actual hlist iterators in linux/list.h - Fix up the declaration of other iterators based on the hlist ones. - A very small amount of places were using the 'node' parameter, this was modified to use 'obj->member' instead. - Coccinelle didn't handle the hlist_for_each_entry_safe iterator properly, so those had to be fixed up manually. The semantic patch which is mostly the work of Peter Senna Tschudin is here: @@ iterator name hlist_for_each_entry, hlist_for_each_entry_continue, hlist_for_each_entry_from, hlist_for_each_entry_rcu, hlist_for_each_entry_rcu_bh, hlist_for_each_entry_continue_rcu_bh, for_each_busy_worker, ax25_uid_for_each, ax25_for_each, inet_bind_bucket_for_each, sctp_for_each_hentry, sk_for_each, sk_for_each_rcu, sk_for_each_from, sk_for_each_safe, sk_for_each_bound, hlist_for_each_entry_safe, hlist_for_each_entry_continue_rcu, nr_neigh_for_each, nr_neigh_for_each_safe, nr_node_for_each, nr_node_for_each_safe, for_each_gfn_indirect_valid_sp, for_each_gfn_sp, for_each_host; type T; expression a,c,d,e; identifier b; statement S; @@ -T b; <+... when != b ( hlist_for_each_entry(a, - b, c, d) S | hlist_for_each_entry_continue(a, - b, c) S | hlist_for_each_entry_from(a, - b, c) S | hlist_for_each_entry_rcu(a, - b, c, d) S | hlist_for_each_entry_rcu_bh(a, - b, c, d) S | hlist_for_each_entry_continue_rcu_bh(a, - b, c) S | for_each_busy_worker(a, c, - b, d) S | ax25_uid_for_each(a, - b, c) S | ax25_for_each(a, - b, c) S | inet_bind_bucket_for_each(a, - b, c) S | sctp_for_each_hentry(a, - b, c) S | sk_for_each(a, - b, c) S | sk_for_each_rcu(a, - b, c) S | sk_for_each_from -(a, b) +(a) S + sk_for_each_from(a) S | sk_for_each_safe(a, - b, c, d) S | sk_for_each_bound(a, - b, c) S | hlist_for_each_entry_safe(a, - b, c, d, e) S | hlist_for_each_entry_continue_rcu(a, - b, c) S | nr_neigh_for_each(a, - b, c) S | nr_neigh_for_each_safe(a, - b, c, d) S | nr_node_for_each(a, - b, c) S | nr_node_for_each_safe(a, - b, c, d) S | - for_each_gfn_sp(a, c, d, b) S + for_each_gfn_sp(a, c, d) S | - for_each_gfn_indirect_valid_sp(a, c, d, b) S + for_each_gfn_indirect_valid_sp(a, c, d) S | for_each_host(a, - b, c) S | for_each_host_safe(a, - b, c, d) S | for_each_mesh_entry(a, - b, c, d) S ) ...+> [akpm@linux-foundation.org: drop bogus change from net/ipv4/raw.c] [akpm@linux-foundation.org: drop bogus hunk from net/ipv6/raw.c] [akpm@linux-foundation.org: checkpatch fixes] [akpm@linux-foundation.org: fix warnings] [akpm@linux-foudnation.org: redo intrusive kvm changes] Tested-by: Peter Senna Tschudin <peter.senna@gmail.com> Acked-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com> Signed-off-by: Sasha Levin <sasha.levin@oracle.com> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Marcelo Tosatti <mtosatti@redhat.com> Cc: Gleb Natapov <gleb@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-02-28 01:06:00 +00:00
hlist_for_each_entry_rcu(node, head, list) {
if (ssid == node->ae.ssid &&
tclass == node->ae.tclass &&
tsid == node->ae.tsid) {
ret = node;
break;
}
}
return ret;
}
/**
* avc_lookup - Look up an AVC entry.
* @ssid: source security identifier
* @tsid: target security identifier
* @tclass: target security class
*
* Look up an AVC entry that is valid for the
* (@ssid, @tsid), interpreting the permissions
* based on @tclass. If a valid AVC entry exists,
* then this function returns the avc_node.
* Otherwise, this function returns NULL.
*/
static struct avc_node *avc_lookup(u32 ssid, u32 tsid, u16 tclass)
{
struct avc_node *node;
avc_cache_stats_incr(lookups);
node = avc_search_node(ssid, tsid, tclass);
if (node)
return node;
avc_cache_stats_incr(misses);
return NULL;
}
static int avc_latest_notif_update(int seqno, int is_insert)
{
int ret = 0;
static DEFINE_SPINLOCK(notif_lock);
unsigned long flag;
spin_lock_irqsave(&notif_lock, flag);
if (is_insert) {
if (seqno < selinux_avc.avc_cache.latest_notif) {
pr_warn("SELinux: avc: seqno %d < latest_notif %d\n",
seqno, selinux_avc.avc_cache.latest_notif);
ret = -EAGAIN;
}
} else {
if (seqno > selinux_avc.avc_cache.latest_notif)
selinux_avc.avc_cache.latest_notif = seqno;
}
spin_unlock_irqrestore(&notif_lock, flag);
return ret;
}
/**
* avc_insert - Insert an AVC entry.
* @ssid: source security identifier
* @tsid: target security identifier
* @tclass: target security class
* @avd: resulting av decision
selinux: extended permissions for ioctls Add extended permissions logic to selinux. Extended permissions provides additional permissions in 256 bit increments. Extend the generic ioctl permission check to use the extended permissions for per-command filtering. Source/target/class sets including the ioctl permission may additionally include a set of commands. Example: allowxperm <source> <target>:<class> ioctl unpriv_app_socket_cmds auditallowxperm <source> <target>:<class> ioctl priv_gpu_cmds Where unpriv_app_socket_cmds and priv_gpu_cmds are macros representing commonly granted sets of ioctl commands. When ioctl commands are omitted only the permissions are checked. This feature is intended to provide finer granularity for the ioctl permission that may be too imprecise. For example, the same driver may use ioctls to provide important and benign functionality such as driver version or socket type as well as dangerous capabilities such as debugging features, read/write/execute to physical memory or access to sensitive data. Per-command filtering provides a mechanism to reduce the attack surface of the kernel, and limit applications to the subset of commands required. The format of the policy binary has been modified to include ioctl commands, and the policy version number has been incremented to POLICYDB_VERSION_XPERMS_IOCTL=30 to account for the format change. The extended permissions logic is deliberately generic to allow components to be reused e.g. netlink filters Signed-off-by: Jeff Vander Stoep <jeffv@google.com> Acked-by: Nick Kralevich <nnk@google.com> Signed-off-by: Paul Moore <pmoore@redhat.com>
2015-07-10 21:19:56 +00:00
* @xp_node: resulting extended permissions
*
* Insert an AVC entry for the SID pair
* (@ssid, @tsid) and class @tclass.
* The access vectors and the sequence number are
* normally provided by the security server in
* response to a security_compute_av() call. If the
* sequence number @avd->seqno is not less than the latest
* revocation notification, then the function copies
* the access vectors into a cache entry.
*/
static void avc_insert(u32 ssid, u32 tsid, u16 tclass,
struct av_decision *avd, struct avc_xperms_node *xp_node)
{
struct avc_node *pos, *node = NULL;
int hvalue;
unsigned long flag;
spinlock_t *lock;
struct hlist_head *head;
if (avc_latest_notif_update(avd->seqno, 1))
return;
node = avc_alloc_node();
if (!node)
return;
avc_node_populate(node, ssid, tsid, tclass, avd);
if (avc_xperms_populate(node, xp_node)) {
avc_node_kill(node);
return;
}
hvalue = avc_hash(ssid, tsid, tclass);
head = &selinux_avc.avc_cache.slots[hvalue];
lock = &selinux_avc.avc_cache.slots_lock[hvalue];
spin_lock_irqsave(lock, flag);
hlist_for_each_entry(pos, head, list) {
if (pos->ae.ssid == ssid &&
pos->ae.tsid == tsid &&
pos->ae.tclass == tclass) {
avc_node_replace(node, pos);
goto found;
}
}
hlist_add_head_rcu(&node->list, head);
found:
spin_unlock_irqrestore(lock, flag);
return;
}
/**
* avc_audit_pre_callback - SELinux specific information
* will be called by generic audit code
* @ab: the audit buffer
* @a: audit_data
*/
static void avc_audit_pre_callback(struct audit_buffer *ab, void *a)
{
struct common_audit_data *ad = a;
struct selinux_audit_data *sad = ad->selinux_audit_data;
u32 av = sad->audited;
const char *const *perms;
int i, perm;
audit_log_format(ab, "avc: %s ", sad->denied ? "denied" : "granted");
if (av == 0) {
audit_log_format(ab, " null");
return;
}
perms = secclass_map[sad->tclass-1].perms;
audit_log_format(ab, " {");
i = 0;
perm = 1;
while (i < (sizeof(av) * 8)) {
if ((perm & av) && perms[i]) {
audit_log_format(ab, " %s", perms[i]);
av &= ~perm;
}
i++;
perm <<= 1;
}
if (av)
audit_log_format(ab, " 0x%x", av);
audit_log_format(ab, " } for ");
}
/**
* avc_audit_post_callback - SELinux specific information
* will be called by generic audit code
* @ab: the audit buffer
* @a: audit_data
*/
static void avc_audit_post_callback(struct audit_buffer *ab, void *a)
{
struct common_audit_data *ad = a;
struct selinux_audit_data *sad = ad->selinux_audit_data;
selinux: add basic filtering for audit trace events This patch adds further attributes to the event. These attributes are helpful to understand the context of the message and can be used to filter the events. There are three common items. Source context, target context and tclass. There are also items from the outcome of operation performed. An event is similar to: <...>-1309 [002] .... 6346.691689: selinux_audited: requested=0x4000000 denied=0x4000000 audited=0x4000000 result=-13 scontext=system_u:system_r:cupsd_t:s0-s0:c0.c1023 tcontext=system_u:object_r:bin_t:s0 tclass=file With systems where many denials are occurring, it is useful to apply a filter. The filtering is a set of logic that is inserted with the filter file. Example: echo "tclass==\"file\" " > events/avc/selinux_audited/filter This adds that we only get tclass=file. The trace can also have extra properties. Adding the user stack can be done with echo 1 > options/userstacktrace Now the output will be runcon-1365 [003] .... 6960.955530: selinux_audited: requested=0x4000000 denied=0x4000000 audited=0x4000000 result=-13 scontext=system_u:system_r:cupsd_t:s0-s0:c0.c1023 tcontext=system_u:object_r:bin_t:s0 tclass=file runcon-1365 [003] .... 6960.955560: <user stack trace> => <00007f325b4ce45b> => <00005607093efa57> Signed-off-by: Peter Enderborg <peter.enderborg@sony.com> Reviewed-by: Thiébaud Weksteen <tweek@google.com> Acked-by: Stephen Smalley <stephen.smalley.work@gmail.com> Signed-off-by: Paul Moore <paul@paul-moore.com>
2020-08-21 14:08:22 +00:00
char *scontext = NULL;
char *tcontext = NULL;
const char *tclass = NULL;
u32 scontext_len;
selinux: add basic filtering for audit trace events This patch adds further attributes to the event. These attributes are helpful to understand the context of the message and can be used to filter the events. There are three common items. Source context, target context and tclass. There are also items from the outcome of operation performed. An event is similar to: <...>-1309 [002] .... 6346.691689: selinux_audited: requested=0x4000000 denied=0x4000000 audited=0x4000000 result=-13 scontext=system_u:system_r:cupsd_t:s0-s0:c0.c1023 tcontext=system_u:object_r:bin_t:s0 tclass=file With systems where many denials are occurring, it is useful to apply a filter. The filtering is a set of logic that is inserted with the filter file. Example: echo "tclass==\"file\" " > events/avc/selinux_audited/filter This adds that we only get tclass=file. The trace can also have extra properties. Adding the user stack can be done with echo 1 > options/userstacktrace Now the output will be runcon-1365 [003] .... 6960.955530: selinux_audited: requested=0x4000000 denied=0x4000000 audited=0x4000000 result=-13 scontext=system_u:system_r:cupsd_t:s0-s0:c0.c1023 tcontext=system_u:object_r:bin_t:s0 tclass=file runcon-1365 [003] .... 6960.955560: <user stack trace> => <00007f325b4ce45b> => <00005607093efa57> Signed-off-by: Peter Enderborg <peter.enderborg@sony.com> Reviewed-by: Thiébaud Weksteen <tweek@google.com> Acked-by: Stephen Smalley <stephen.smalley.work@gmail.com> Signed-off-by: Paul Moore <paul@paul-moore.com>
2020-08-21 14:08:22 +00:00
u32 tcontext_len;
int rc;
rc = security_sid_to_context(sad->ssid, &scontext,
&scontext_len);
if (rc)
audit_log_format(ab, " ssid=%d", sad->ssid);
selinux: add basic filtering for audit trace events This patch adds further attributes to the event. These attributes are helpful to understand the context of the message and can be used to filter the events. There are three common items. Source context, target context and tclass. There are also items from the outcome of operation performed. An event is similar to: <...>-1309 [002] .... 6346.691689: selinux_audited: requested=0x4000000 denied=0x4000000 audited=0x4000000 result=-13 scontext=system_u:system_r:cupsd_t:s0-s0:c0.c1023 tcontext=system_u:object_r:bin_t:s0 tclass=file With systems where many denials are occurring, it is useful to apply a filter. The filtering is a set of logic that is inserted with the filter file. Example: echo "tclass==\"file\" " > events/avc/selinux_audited/filter This adds that we only get tclass=file. The trace can also have extra properties. Adding the user stack can be done with echo 1 > options/userstacktrace Now the output will be runcon-1365 [003] .... 6960.955530: selinux_audited: requested=0x4000000 denied=0x4000000 audited=0x4000000 result=-13 scontext=system_u:system_r:cupsd_t:s0-s0:c0.c1023 tcontext=system_u:object_r:bin_t:s0 tclass=file runcon-1365 [003] .... 6960.955560: <user stack trace> => <00007f325b4ce45b> => <00005607093efa57> Signed-off-by: Peter Enderborg <peter.enderborg@sony.com> Reviewed-by: Thiébaud Weksteen <tweek@google.com> Acked-by: Stephen Smalley <stephen.smalley.work@gmail.com> Signed-off-by: Paul Moore <paul@paul-moore.com>
2020-08-21 14:08:22 +00:00
else
audit_log_format(ab, " scontext=%s", scontext);
rc = security_sid_to_context(sad->tsid, &tcontext,
selinux: add basic filtering for audit trace events This patch adds further attributes to the event. These attributes are helpful to understand the context of the message and can be used to filter the events. There are three common items. Source context, target context and tclass. There are also items from the outcome of operation performed. An event is similar to: <...>-1309 [002] .... 6346.691689: selinux_audited: requested=0x4000000 denied=0x4000000 audited=0x4000000 result=-13 scontext=system_u:system_r:cupsd_t:s0-s0:c0.c1023 tcontext=system_u:object_r:bin_t:s0 tclass=file With systems where many denials are occurring, it is useful to apply a filter. The filtering is a set of logic that is inserted with the filter file. Example: echo "tclass==\"file\" " > events/avc/selinux_audited/filter This adds that we only get tclass=file. The trace can also have extra properties. Adding the user stack can be done with echo 1 > options/userstacktrace Now the output will be runcon-1365 [003] .... 6960.955530: selinux_audited: requested=0x4000000 denied=0x4000000 audited=0x4000000 result=-13 scontext=system_u:system_r:cupsd_t:s0-s0:c0.c1023 tcontext=system_u:object_r:bin_t:s0 tclass=file runcon-1365 [003] .... 6960.955560: <user stack trace> => <00007f325b4ce45b> => <00005607093efa57> Signed-off-by: Peter Enderborg <peter.enderborg@sony.com> Reviewed-by: Thiébaud Weksteen <tweek@google.com> Acked-by: Stephen Smalley <stephen.smalley.work@gmail.com> Signed-off-by: Paul Moore <paul@paul-moore.com>
2020-08-21 14:08:22 +00:00
&tcontext_len);
if (rc)
audit_log_format(ab, " tsid=%d", sad->tsid);
selinux: add basic filtering for audit trace events This patch adds further attributes to the event. These attributes are helpful to understand the context of the message and can be used to filter the events. There are three common items. Source context, target context and tclass. There are also items from the outcome of operation performed. An event is similar to: <...>-1309 [002] .... 6346.691689: selinux_audited: requested=0x4000000 denied=0x4000000 audited=0x4000000 result=-13 scontext=system_u:system_r:cupsd_t:s0-s0:c0.c1023 tcontext=system_u:object_r:bin_t:s0 tclass=file With systems where many denials are occurring, it is useful to apply a filter. The filtering is a set of logic that is inserted with the filter file. Example: echo "tclass==\"file\" " > events/avc/selinux_audited/filter This adds that we only get tclass=file. The trace can also have extra properties. Adding the user stack can be done with echo 1 > options/userstacktrace Now the output will be runcon-1365 [003] .... 6960.955530: selinux_audited: requested=0x4000000 denied=0x4000000 audited=0x4000000 result=-13 scontext=system_u:system_r:cupsd_t:s0-s0:c0.c1023 tcontext=system_u:object_r:bin_t:s0 tclass=file runcon-1365 [003] .... 6960.955560: <user stack trace> => <00007f325b4ce45b> => <00005607093efa57> Signed-off-by: Peter Enderborg <peter.enderborg@sony.com> Reviewed-by: Thiébaud Weksteen <tweek@google.com> Acked-by: Stephen Smalley <stephen.smalley.work@gmail.com> Signed-off-by: Paul Moore <paul@paul-moore.com>
2020-08-21 14:08:22 +00:00
else
audit_log_format(ab, " tcontext=%s", tcontext);
selinux: add basic filtering for audit trace events This patch adds further attributes to the event. These attributes are helpful to understand the context of the message and can be used to filter the events. There are three common items. Source context, target context and tclass. There are also items from the outcome of operation performed. An event is similar to: <...>-1309 [002] .... 6346.691689: selinux_audited: requested=0x4000000 denied=0x4000000 audited=0x4000000 result=-13 scontext=system_u:system_r:cupsd_t:s0-s0:c0.c1023 tcontext=system_u:object_r:bin_t:s0 tclass=file With systems where many denials are occurring, it is useful to apply a filter. The filtering is a set of logic that is inserted with the filter file. Example: echo "tclass==\"file\" " > events/avc/selinux_audited/filter This adds that we only get tclass=file. The trace can also have extra properties. Adding the user stack can be done with echo 1 > options/userstacktrace Now the output will be runcon-1365 [003] .... 6960.955530: selinux_audited: requested=0x4000000 denied=0x4000000 audited=0x4000000 result=-13 scontext=system_u:system_r:cupsd_t:s0-s0:c0.c1023 tcontext=system_u:object_r:bin_t:s0 tclass=file runcon-1365 [003] .... 6960.955560: <user stack trace> => <00007f325b4ce45b> => <00005607093efa57> Signed-off-by: Peter Enderborg <peter.enderborg@sony.com> Reviewed-by: Thiébaud Weksteen <tweek@google.com> Acked-by: Stephen Smalley <stephen.smalley.work@gmail.com> Signed-off-by: Paul Moore <paul@paul-moore.com>
2020-08-21 14:08:22 +00:00
tclass = secclass_map[sad->tclass-1].name;
audit_log_format(ab, " tclass=%s", tclass);
if (sad->denied)
audit_log_format(ab, " permissive=%u", sad->result ? 0 : 1);
selinux: log invalid contexts in AVCs In case a file has an invalid context set, in an AVC record generated upon access to such file, the target context is always reported as unlabeled. This patch adds new optional fields to the AVC record (srawcon and trawcon) that report the actual context string if it differs from the one reported in scontext/tcontext. This is useful for diagnosing SELinux denials involving invalid contexts. To trigger an AVC that illustrates this situation: # setenforce 0 # touch /tmp/testfile # setfattr -n security.selinux -v system_u:object_r:banana_t:s0 /tmp/testfile # runcon system_u:system_r:sshd_t:s0 cat /tmp/testfile AVC before: type=AVC msg=audit(1547801083.248:11): avc: denied { open } for pid=1149 comm="cat" path="/tmp/testfile" dev="tmpfs" ino=6608 scontext=system_u:system_r:sshd_t:s0 tcontext=system_u:object_r:unlabeled_t:s15:c0.c1023 tclass=file permissive=1 AVC after: type=AVC msg=audit(1547801083.248:11): avc: denied { open } for pid=1149 comm="cat" path="/tmp/testfile" dev="tmpfs" ino=6608 scontext=system_u:system_r:sshd_t:s0 tcontext=system_u:object_r:unlabeled_t:s15:c0.c1023 tclass=file permissive=1 trawcon=system_u:object_r:banana_t:s0 Note that it is also possible to encounter this situation with the 'scontext' field - e.g. when a new policy is loaded while a process is running, whose context is not valid in the new policy. Link: https://bugzilla.redhat.com/show_bug.cgi?id=1135683 Cc: Daniel Walsh <dwalsh@redhat.com> Signed-off-by: Ondrej Mosnacek <omosnace@redhat.com> Reviewed-by: Stephen Smalley <sds@tycho.nsa.gov> Signed-off-by: Paul Moore <paul@paul-moore.com>
2019-01-25 10:06:51 +00:00
selinux: add basic filtering for audit trace events This patch adds further attributes to the event. These attributes are helpful to understand the context of the message and can be used to filter the events. There are three common items. Source context, target context and tclass. There are also items from the outcome of operation performed. An event is similar to: <...>-1309 [002] .... 6346.691689: selinux_audited: requested=0x4000000 denied=0x4000000 audited=0x4000000 result=-13 scontext=system_u:system_r:cupsd_t:s0-s0:c0.c1023 tcontext=system_u:object_r:bin_t:s0 tclass=file With systems where many denials are occurring, it is useful to apply a filter. The filtering is a set of logic that is inserted with the filter file. Example: echo "tclass==\"file\" " > events/avc/selinux_audited/filter This adds that we only get tclass=file. The trace can also have extra properties. Adding the user stack can be done with echo 1 > options/userstacktrace Now the output will be runcon-1365 [003] .... 6960.955530: selinux_audited: requested=0x4000000 denied=0x4000000 audited=0x4000000 result=-13 scontext=system_u:system_r:cupsd_t:s0-s0:c0.c1023 tcontext=system_u:object_r:bin_t:s0 tclass=file runcon-1365 [003] .... 6960.955560: <user stack trace> => <00007f325b4ce45b> => <00005607093efa57> Signed-off-by: Peter Enderborg <peter.enderborg@sony.com> Reviewed-by: Thiébaud Weksteen <tweek@google.com> Acked-by: Stephen Smalley <stephen.smalley.work@gmail.com> Signed-off-by: Paul Moore <paul@paul-moore.com>
2020-08-21 14:08:22 +00:00
trace_selinux_audited(sad, scontext, tcontext, tclass);
kfree(tcontext);
kfree(scontext);
selinux: log invalid contexts in AVCs In case a file has an invalid context set, in an AVC record generated upon access to such file, the target context is always reported as unlabeled. This patch adds new optional fields to the AVC record (srawcon and trawcon) that report the actual context string if it differs from the one reported in scontext/tcontext. This is useful for diagnosing SELinux denials involving invalid contexts. To trigger an AVC that illustrates this situation: # setenforce 0 # touch /tmp/testfile # setfattr -n security.selinux -v system_u:object_r:banana_t:s0 /tmp/testfile # runcon system_u:system_r:sshd_t:s0 cat /tmp/testfile AVC before: type=AVC msg=audit(1547801083.248:11): avc: denied { open } for pid=1149 comm="cat" path="/tmp/testfile" dev="tmpfs" ino=6608 scontext=system_u:system_r:sshd_t:s0 tcontext=system_u:object_r:unlabeled_t:s15:c0.c1023 tclass=file permissive=1 AVC after: type=AVC msg=audit(1547801083.248:11): avc: denied { open } for pid=1149 comm="cat" path="/tmp/testfile" dev="tmpfs" ino=6608 scontext=system_u:system_r:sshd_t:s0 tcontext=system_u:object_r:unlabeled_t:s15:c0.c1023 tclass=file permissive=1 trawcon=system_u:object_r:banana_t:s0 Note that it is also possible to encounter this situation with the 'scontext' field - e.g. when a new policy is loaded while a process is running, whose context is not valid in the new policy. Link: https://bugzilla.redhat.com/show_bug.cgi?id=1135683 Cc: Daniel Walsh <dwalsh@redhat.com> Signed-off-by: Ondrej Mosnacek <omosnace@redhat.com> Reviewed-by: Stephen Smalley <sds@tycho.nsa.gov> Signed-off-by: Paul Moore <paul@paul-moore.com>
2019-01-25 10:06:51 +00:00
/* in case of invalid context report also the actual context string */
rc = security_sid_to_context_inval(sad->ssid, &scontext,
selinux: log invalid contexts in AVCs In case a file has an invalid context set, in an AVC record generated upon access to such file, the target context is always reported as unlabeled. This patch adds new optional fields to the AVC record (srawcon and trawcon) that report the actual context string if it differs from the one reported in scontext/tcontext. This is useful for diagnosing SELinux denials involving invalid contexts. To trigger an AVC that illustrates this situation: # setenforce 0 # touch /tmp/testfile # setfattr -n security.selinux -v system_u:object_r:banana_t:s0 /tmp/testfile # runcon system_u:system_r:sshd_t:s0 cat /tmp/testfile AVC before: type=AVC msg=audit(1547801083.248:11): avc: denied { open } for pid=1149 comm="cat" path="/tmp/testfile" dev="tmpfs" ino=6608 scontext=system_u:system_r:sshd_t:s0 tcontext=system_u:object_r:unlabeled_t:s15:c0.c1023 tclass=file permissive=1 AVC after: type=AVC msg=audit(1547801083.248:11): avc: denied { open } for pid=1149 comm="cat" path="/tmp/testfile" dev="tmpfs" ino=6608 scontext=system_u:system_r:sshd_t:s0 tcontext=system_u:object_r:unlabeled_t:s15:c0.c1023 tclass=file permissive=1 trawcon=system_u:object_r:banana_t:s0 Note that it is also possible to encounter this situation with the 'scontext' field - e.g. when a new policy is loaded while a process is running, whose context is not valid in the new policy. Link: https://bugzilla.redhat.com/show_bug.cgi?id=1135683 Cc: Daniel Walsh <dwalsh@redhat.com> Signed-off-by: Ondrej Mosnacek <omosnace@redhat.com> Reviewed-by: Stephen Smalley <sds@tycho.nsa.gov> Signed-off-by: Paul Moore <paul@paul-moore.com>
2019-01-25 10:06:51 +00:00
&scontext_len);
if (!rc && scontext) {
if (scontext_len && scontext[scontext_len - 1] == '\0')
scontext_len--;
audit_log_format(ab, " srawcon=");
audit_log_n_untrustedstring(ab, scontext, scontext_len);
selinux: log invalid contexts in AVCs In case a file has an invalid context set, in an AVC record generated upon access to such file, the target context is always reported as unlabeled. This patch adds new optional fields to the AVC record (srawcon and trawcon) that report the actual context string if it differs from the one reported in scontext/tcontext. This is useful for diagnosing SELinux denials involving invalid contexts. To trigger an AVC that illustrates this situation: # setenforce 0 # touch /tmp/testfile # setfattr -n security.selinux -v system_u:object_r:banana_t:s0 /tmp/testfile # runcon system_u:system_r:sshd_t:s0 cat /tmp/testfile AVC before: type=AVC msg=audit(1547801083.248:11): avc: denied { open } for pid=1149 comm="cat" path="/tmp/testfile" dev="tmpfs" ino=6608 scontext=system_u:system_r:sshd_t:s0 tcontext=system_u:object_r:unlabeled_t:s15:c0.c1023 tclass=file permissive=1 AVC after: type=AVC msg=audit(1547801083.248:11): avc: denied { open } for pid=1149 comm="cat" path="/tmp/testfile" dev="tmpfs" ino=6608 scontext=system_u:system_r:sshd_t:s0 tcontext=system_u:object_r:unlabeled_t:s15:c0.c1023 tclass=file permissive=1 trawcon=system_u:object_r:banana_t:s0 Note that it is also possible to encounter this situation with the 'scontext' field - e.g. when a new policy is loaded while a process is running, whose context is not valid in the new policy. Link: https://bugzilla.redhat.com/show_bug.cgi?id=1135683 Cc: Daniel Walsh <dwalsh@redhat.com> Signed-off-by: Ondrej Mosnacek <omosnace@redhat.com> Reviewed-by: Stephen Smalley <sds@tycho.nsa.gov> Signed-off-by: Paul Moore <paul@paul-moore.com>
2019-01-25 10:06:51 +00:00
kfree(scontext);
}
rc = security_sid_to_context_inval(sad->tsid, &scontext,
selinux: log invalid contexts in AVCs In case a file has an invalid context set, in an AVC record generated upon access to such file, the target context is always reported as unlabeled. This patch adds new optional fields to the AVC record (srawcon and trawcon) that report the actual context string if it differs from the one reported in scontext/tcontext. This is useful for diagnosing SELinux denials involving invalid contexts. To trigger an AVC that illustrates this situation: # setenforce 0 # touch /tmp/testfile # setfattr -n security.selinux -v system_u:object_r:banana_t:s0 /tmp/testfile # runcon system_u:system_r:sshd_t:s0 cat /tmp/testfile AVC before: type=AVC msg=audit(1547801083.248:11): avc: denied { open } for pid=1149 comm="cat" path="/tmp/testfile" dev="tmpfs" ino=6608 scontext=system_u:system_r:sshd_t:s0 tcontext=system_u:object_r:unlabeled_t:s15:c0.c1023 tclass=file permissive=1 AVC after: type=AVC msg=audit(1547801083.248:11): avc: denied { open } for pid=1149 comm="cat" path="/tmp/testfile" dev="tmpfs" ino=6608 scontext=system_u:system_r:sshd_t:s0 tcontext=system_u:object_r:unlabeled_t:s15:c0.c1023 tclass=file permissive=1 trawcon=system_u:object_r:banana_t:s0 Note that it is also possible to encounter this situation with the 'scontext' field - e.g. when a new policy is loaded while a process is running, whose context is not valid in the new policy. Link: https://bugzilla.redhat.com/show_bug.cgi?id=1135683 Cc: Daniel Walsh <dwalsh@redhat.com> Signed-off-by: Ondrej Mosnacek <omosnace@redhat.com> Reviewed-by: Stephen Smalley <sds@tycho.nsa.gov> Signed-off-by: Paul Moore <paul@paul-moore.com>
2019-01-25 10:06:51 +00:00
&scontext_len);
if (!rc && scontext) {
if (scontext_len && scontext[scontext_len - 1] == '\0')
scontext_len--;
audit_log_format(ab, " trawcon=");
audit_log_n_untrustedstring(ab, scontext, scontext_len);
selinux: log invalid contexts in AVCs In case a file has an invalid context set, in an AVC record generated upon access to such file, the target context is always reported as unlabeled. This patch adds new optional fields to the AVC record (srawcon and trawcon) that report the actual context string if it differs from the one reported in scontext/tcontext. This is useful for diagnosing SELinux denials involving invalid contexts. To trigger an AVC that illustrates this situation: # setenforce 0 # touch /tmp/testfile # setfattr -n security.selinux -v system_u:object_r:banana_t:s0 /tmp/testfile # runcon system_u:system_r:sshd_t:s0 cat /tmp/testfile AVC before: type=AVC msg=audit(1547801083.248:11): avc: denied { open } for pid=1149 comm="cat" path="/tmp/testfile" dev="tmpfs" ino=6608 scontext=system_u:system_r:sshd_t:s0 tcontext=system_u:object_r:unlabeled_t:s15:c0.c1023 tclass=file permissive=1 AVC after: type=AVC msg=audit(1547801083.248:11): avc: denied { open } for pid=1149 comm="cat" path="/tmp/testfile" dev="tmpfs" ino=6608 scontext=system_u:system_r:sshd_t:s0 tcontext=system_u:object_r:unlabeled_t:s15:c0.c1023 tclass=file permissive=1 trawcon=system_u:object_r:banana_t:s0 Note that it is also possible to encounter this situation with the 'scontext' field - e.g. when a new policy is loaded while a process is running, whose context is not valid in the new policy. Link: https://bugzilla.redhat.com/show_bug.cgi?id=1135683 Cc: Daniel Walsh <dwalsh@redhat.com> Signed-off-by: Ondrej Mosnacek <omosnace@redhat.com> Reviewed-by: Stephen Smalley <sds@tycho.nsa.gov> Signed-off-by: Paul Moore <paul@paul-moore.com>
2019-01-25 10:06:51 +00:00
kfree(scontext);
}
}
/*
* This is the slow part of avc audit with big stack footprint.
* Note that it is non-blocking and can be called from under
* rcu_read_lock().
*/
noinline int slow_avc_audit(u32 ssid, u32 tsid, u16 tclass,
u32 requested, u32 audited, u32 denied, int result,
struct common_audit_data *a)
{
struct common_audit_data stack_data;
struct selinux_audit_data sad;
if (WARN_ON(!tclass || tclass >= ARRAY_SIZE(secclass_map)))
return -EINVAL;
if (!a) {
a = &stack_data;
a->type = LSM_AUDIT_DATA_NONE;
}
sad.tclass = tclass;
sad.requested = requested;
sad.ssid = ssid;
sad.tsid = tsid;
sad.audited = audited;
sad.denied = denied;
sad.result = result;
a->selinux_audit_data = &sad;
common_lsm_audit(a, avc_audit_pre_callback, avc_audit_post_callback);
return 0;
}
/**
* avc_add_callback - Register a callback for security events.
* @callback: callback function
* @events: security events
*
* Register a callback function for events in the set @events.
* Returns %0 on success or -%ENOMEM if insufficient memory
* exists to add the callback.
*/
int __init avc_add_callback(int (*callback)(u32 event), u32 events)
{
struct avc_callback_node *c;
int rc = 0;
c = kmalloc(sizeof(*c), GFP_KERNEL);
if (!c) {
rc = -ENOMEM;
goto out;
}
c->callback = callback;
c->events = events;
c->next = avc_callbacks;
avc_callbacks = c;
out:
return rc;
}
/**
* avc_update_node - Update an AVC entry
* @event : Updating event
* @perms : Permission mask bits
* @driver: xperm driver information
* @xperm: xperm permissions
* @ssid: AVC entry source sid
* @tsid: AVC entry target sid
* @tclass : AVC entry target object class
* @seqno : sequence number when decision was made
selinux: extended permissions for ioctls Add extended permissions logic to selinux. Extended permissions provides additional permissions in 256 bit increments. Extend the generic ioctl permission check to use the extended permissions for per-command filtering. Source/target/class sets including the ioctl permission may additionally include a set of commands. Example: allowxperm <source> <target>:<class> ioctl unpriv_app_socket_cmds auditallowxperm <source> <target>:<class> ioctl priv_gpu_cmds Where unpriv_app_socket_cmds and priv_gpu_cmds are macros representing commonly granted sets of ioctl commands. When ioctl commands are omitted only the permissions are checked. This feature is intended to provide finer granularity for the ioctl permission that may be too imprecise. For example, the same driver may use ioctls to provide important and benign functionality such as driver version or socket type as well as dangerous capabilities such as debugging features, read/write/execute to physical memory or access to sensitive data. Per-command filtering provides a mechanism to reduce the attack surface of the kernel, and limit applications to the subset of commands required. The format of the policy binary has been modified to include ioctl commands, and the policy version number has been incremented to POLICYDB_VERSION_XPERMS_IOCTL=30 to account for the format change. The extended permissions logic is deliberately generic to allow components to be reused e.g. netlink filters Signed-off-by: Jeff Vander Stoep <jeffv@google.com> Acked-by: Nick Kralevich <nnk@google.com> Signed-off-by: Paul Moore <pmoore@redhat.com>
2015-07-10 21:19:56 +00:00
* @xpd: extended_perms_decision to be added to the node
* @flags: the AVC_* flags, e.g. AVC_EXTENDED_PERMS, or 0.
*
* if a valid AVC entry doesn't exist,this function returns -ENOENT.
* if kmalloc() called internal returns NULL, this function returns -ENOMEM.
* otherwise, this function updates the AVC entry. The original AVC-entry object
* will release later by RCU.
*/
static int avc_update_node(u32 event, u32 perms, u8 driver, u8 xperm, u32 ssid,
u32 tsid, u16 tclass, u32 seqno,
struct extended_perms_decision *xpd,
u32 flags)
{
int hvalue, rc = 0;
unsigned long flag;
struct avc_node *pos, *node, *orig = NULL;
struct hlist_head *head;
spinlock_t *lock;
node = avc_alloc_node();
if (!node) {
rc = -ENOMEM;
goto out;
}
/* Lock the target slot */
hvalue = avc_hash(ssid, tsid, tclass);
head = &selinux_avc.avc_cache.slots[hvalue];
lock = &selinux_avc.avc_cache.slots_lock[hvalue];
spin_lock_irqsave(lock, flag);
hlist: drop the node parameter from iterators I'm not sure why, but the hlist for each entry iterators were conceived list_for_each_entry(pos, head, member) The hlist ones were greedy and wanted an extra parameter: hlist_for_each_entry(tpos, pos, head, member) Why did they need an extra pos parameter? I'm not quite sure. Not only they don't really need it, it also prevents the iterator from looking exactly like the list iterator, which is unfortunate. Besides the semantic patch, there was some manual work required: - Fix up the actual hlist iterators in linux/list.h - Fix up the declaration of other iterators based on the hlist ones. - A very small amount of places were using the 'node' parameter, this was modified to use 'obj->member' instead. - Coccinelle didn't handle the hlist_for_each_entry_safe iterator properly, so those had to be fixed up manually. The semantic patch which is mostly the work of Peter Senna Tschudin is here: @@ iterator name hlist_for_each_entry, hlist_for_each_entry_continue, hlist_for_each_entry_from, hlist_for_each_entry_rcu, hlist_for_each_entry_rcu_bh, hlist_for_each_entry_continue_rcu_bh, for_each_busy_worker, ax25_uid_for_each, ax25_for_each, inet_bind_bucket_for_each, sctp_for_each_hentry, sk_for_each, sk_for_each_rcu, sk_for_each_from, sk_for_each_safe, sk_for_each_bound, hlist_for_each_entry_safe, hlist_for_each_entry_continue_rcu, nr_neigh_for_each, nr_neigh_for_each_safe, nr_node_for_each, nr_node_for_each_safe, for_each_gfn_indirect_valid_sp, for_each_gfn_sp, for_each_host; type T; expression a,c,d,e; identifier b; statement S; @@ -T b; <+... when != b ( hlist_for_each_entry(a, - b, c, d) S | hlist_for_each_entry_continue(a, - b, c) S | hlist_for_each_entry_from(a, - b, c) S | hlist_for_each_entry_rcu(a, - b, c, d) S | hlist_for_each_entry_rcu_bh(a, - b, c, d) S | hlist_for_each_entry_continue_rcu_bh(a, - b, c) S | for_each_busy_worker(a, c, - b, d) S | ax25_uid_for_each(a, - b, c) S | ax25_for_each(a, - b, c) S | inet_bind_bucket_for_each(a, - b, c) S | sctp_for_each_hentry(a, - b, c) S | sk_for_each(a, - b, c) S | sk_for_each_rcu(a, - b, c) S | sk_for_each_from -(a, b) +(a) S + sk_for_each_from(a) S | sk_for_each_safe(a, - b, c, d) S | sk_for_each_bound(a, - b, c) S | hlist_for_each_entry_safe(a, - b, c, d, e) S | hlist_for_each_entry_continue_rcu(a, - b, c) S | nr_neigh_for_each(a, - b, c) S | nr_neigh_for_each_safe(a, - b, c, d) S | nr_node_for_each(a, - b, c) S | nr_node_for_each_safe(a, - b, c, d) S | - for_each_gfn_sp(a, c, d, b) S + for_each_gfn_sp(a, c, d) S | - for_each_gfn_indirect_valid_sp(a, c, d, b) S + for_each_gfn_indirect_valid_sp(a, c, d) S | for_each_host(a, - b, c) S | for_each_host_safe(a, - b, c, d) S | for_each_mesh_entry(a, - b, c, d) S ) ...+> [akpm@linux-foundation.org: drop bogus change from net/ipv4/raw.c] [akpm@linux-foundation.org: drop bogus hunk from net/ipv6/raw.c] [akpm@linux-foundation.org: checkpatch fixes] [akpm@linux-foundation.org: fix warnings] [akpm@linux-foudnation.org: redo intrusive kvm changes] Tested-by: Peter Senna Tschudin <peter.senna@gmail.com> Acked-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com> Signed-off-by: Sasha Levin <sasha.levin@oracle.com> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Marcelo Tosatti <mtosatti@redhat.com> Cc: Gleb Natapov <gleb@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-02-28 01:06:00 +00:00
hlist_for_each_entry(pos, head, list) {
if (ssid == pos->ae.ssid &&
tsid == pos->ae.tsid &&
tclass == pos->ae.tclass &&
seqno == pos->ae.avd.seqno){
orig = pos;
break;
}
}
if (!orig) {
rc = -ENOENT;
avc_node_kill(node);
goto out_unlock;
}
/*
* Copy and replace original node.
*/
avc_node_populate(node, ssid, tsid, tclass, &orig->ae.avd);
selinux: extended permissions for ioctls Add extended permissions logic to selinux. Extended permissions provides additional permissions in 256 bit increments. Extend the generic ioctl permission check to use the extended permissions for per-command filtering. Source/target/class sets including the ioctl permission may additionally include a set of commands. Example: allowxperm <source> <target>:<class> ioctl unpriv_app_socket_cmds auditallowxperm <source> <target>:<class> ioctl priv_gpu_cmds Where unpriv_app_socket_cmds and priv_gpu_cmds are macros representing commonly granted sets of ioctl commands. When ioctl commands are omitted only the permissions are checked. This feature is intended to provide finer granularity for the ioctl permission that may be too imprecise. For example, the same driver may use ioctls to provide important and benign functionality such as driver version or socket type as well as dangerous capabilities such as debugging features, read/write/execute to physical memory or access to sensitive data. Per-command filtering provides a mechanism to reduce the attack surface of the kernel, and limit applications to the subset of commands required. The format of the policy binary has been modified to include ioctl commands, and the policy version number has been incremented to POLICYDB_VERSION_XPERMS_IOCTL=30 to account for the format change. The extended permissions logic is deliberately generic to allow components to be reused e.g. netlink filters Signed-off-by: Jeff Vander Stoep <jeffv@google.com> Acked-by: Nick Kralevich <nnk@google.com> Signed-off-by: Paul Moore <pmoore@redhat.com>
2015-07-10 21:19:56 +00:00
if (orig->ae.xp_node) {
rc = avc_xperms_populate(node, orig->ae.xp_node);
if (rc) {
avc_node_kill(node);
selinux: extended permissions for ioctls Add extended permissions logic to selinux. Extended permissions provides additional permissions in 256 bit increments. Extend the generic ioctl permission check to use the extended permissions for per-command filtering. Source/target/class sets including the ioctl permission may additionally include a set of commands. Example: allowxperm <source> <target>:<class> ioctl unpriv_app_socket_cmds auditallowxperm <source> <target>:<class> ioctl priv_gpu_cmds Where unpriv_app_socket_cmds and priv_gpu_cmds are macros representing commonly granted sets of ioctl commands. When ioctl commands are omitted only the permissions are checked. This feature is intended to provide finer granularity for the ioctl permission that may be too imprecise. For example, the same driver may use ioctls to provide important and benign functionality such as driver version or socket type as well as dangerous capabilities such as debugging features, read/write/execute to physical memory or access to sensitive data. Per-command filtering provides a mechanism to reduce the attack surface of the kernel, and limit applications to the subset of commands required. The format of the policy binary has been modified to include ioctl commands, and the policy version number has been incremented to POLICYDB_VERSION_XPERMS_IOCTL=30 to account for the format change. The extended permissions logic is deliberately generic to allow components to be reused e.g. netlink filters Signed-off-by: Jeff Vander Stoep <jeffv@google.com> Acked-by: Nick Kralevich <nnk@google.com> Signed-off-by: Paul Moore <pmoore@redhat.com>
2015-07-10 21:19:56 +00:00
goto out_unlock;
}
}
switch (event) {
case AVC_CALLBACK_GRANT:
node->ae.avd.allowed |= perms;
selinux: extended permissions for ioctls Add extended permissions logic to selinux. Extended permissions provides additional permissions in 256 bit increments. Extend the generic ioctl permission check to use the extended permissions for per-command filtering. Source/target/class sets including the ioctl permission may additionally include a set of commands. Example: allowxperm <source> <target>:<class> ioctl unpriv_app_socket_cmds auditallowxperm <source> <target>:<class> ioctl priv_gpu_cmds Where unpriv_app_socket_cmds and priv_gpu_cmds are macros representing commonly granted sets of ioctl commands. When ioctl commands are omitted only the permissions are checked. This feature is intended to provide finer granularity for the ioctl permission that may be too imprecise. For example, the same driver may use ioctls to provide important and benign functionality such as driver version or socket type as well as dangerous capabilities such as debugging features, read/write/execute to physical memory or access to sensitive data. Per-command filtering provides a mechanism to reduce the attack surface of the kernel, and limit applications to the subset of commands required. The format of the policy binary has been modified to include ioctl commands, and the policy version number has been incremented to POLICYDB_VERSION_XPERMS_IOCTL=30 to account for the format change. The extended permissions logic is deliberately generic to allow components to be reused e.g. netlink filters Signed-off-by: Jeff Vander Stoep <jeffv@google.com> Acked-by: Nick Kralevich <nnk@google.com> Signed-off-by: Paul Moore <pmoore@redhat.com>
2015-07-10 21:19:56 +00:00
if (node->ae.xp_node && (flags & AVC_EXTENDED_PERMS))
avc_xperms_allow_perm(node->ae.xp_node, driver, xperm);
break;
case AVC_CALLBACK_TRY_REVOKE:
case AVC_CALLBACK_REVOKE:
node->ae.avd.allowed &= ~perms;
break;
case AVC_CALLBACK_AUDITALLOW_ENABLE:
node->ae.avd.auditallow |= perms;
break;
case AVC_CALLBACK_AUDITALLOW_DISABLE:
node->ae.avd.auditallow &= ~perms;
break;
case AVC_CALLBACK_AUDITDENY_ENABLE:
node->ae.avd.auditdeny |= perms;
break;
case AVC_CALLBACK_AUDITDENY_DISABLE:
node->ae.avd.auditdeny &= ~perms;
break;
selinux: extended permissions for ioctls Add extended permissions logic to selinux. Extended permissions provides additional permissions in 256 bit increments. Extend the generic ioctl permission check to use the extended permissions for per-command filtering. Source/target/class sets including the ioctl permission may additionally include a set of commands. Example: allowxperm <source> <target>:<class> ioctl unpriv_app_socket_cmds auditallowxperm <source> <target>:<class> ioctl priv_gpu_cmds Where unpriv_app_socket_cmds and priv_gpu_cmds are macros representing commonly granted sets of ioctl commands. When ioctl commands are omitted only the permissions are checked. This feature is intended to provide finer granularity for the ioctl permission that may be too imprecise. For example, the same driver may use ioctls to provide important and benign functionality such as driver version or socket type as well as dangerous capabilities such as debugging features, read/write/execute to physical memory or access to sensitive data. Per-command filtering provides a mechanism to reduce the attack surface of the kernel, and limit applications to the subset of commands required. The format of the policy binary has been modified to include ioctl commands, and the policy version number has been incremented to POLICYDB_VERSION_XPERMS_IOCTL=30 to account for the format change. The extended permissions logic is deliberately generic to allow components to be reused e.g. netlink filters Signed-off-by: Jeff Vander Stoep <jeffv@google.com> Acked-by: Nick Kralevich <nnk@google.com> Signed-off-by: Paul Moore <pmoore@redhat.com>
2015-07-10 21:19:56 +00:00
case AVC_CALLBACK_ADD_XPERMS:
avc_add_xperms_decision(node, xpd);
break;
}
avc_node_replace(node, orig);
out_unlock:
spin_unlock_irqrestore(lock, flag);
out:
return rc;
}
/**
* avc_flush - Flush the cache
*/
static void avc_flush(void)
{
struct hlist_head *head;
struct avc_node *node;
spinlock_t *lock;
unsigned long flag;
int i;
for (i = 0; i < AVC_CACHE_SLOTS; i++) {
head = &selinux_avc.avc_cache.slots[i];
lock = &selinux_avc.avc_cache.slots_lock[i];
spin_lock_irqsave(lock, flag);
/*
* With preemptable RCU, the outer spinlock does not
* prevent RCU grace periods from ending.
*/
rcu_read_lock();
hlist: drop the node parameter from iterators I'm not sure why, but the hlist for each entry iterators were conceived list_for_each_entry(pos, head, member) The hlist ones were greedy and wanted an extra parameter: hlist_for_each_entry(tpos, pos, head, member) Why did they need an extra pos parameter? I'm not quite sure. Not only they don't really need it, it also prevents the iterator from looking exactly like the list iterator, which is unfortunate. Besides the semantic patch, there was some manual work required: - Fix up the actual hlist iterators in linux/list.h - Fix up the declaration of other iterators based on the hlist ones. - A very small amount of places were using the 'node' parameter, this was modified to use 'obj->member' instead. - Coccinelle didn't handle the hlist_for_each_entry_safe iterator properly, so those had to be fixed up manually. The semantic patch which is mostly the work of Peter Senna Tschudin is here: @@ iterator name hlist_for_each_entry, hlist_for_each_entry_continue, hlist_for_each_entry_from, hlist_for_each_entry_rcu, hlist_for_each_entry_rcu_bh, hlist_for_each_entry_continue_rcu_bh, for_each_busy_worker, ax25_uid_for_each, ax25_for_each, inet_bind_bucket_for_each, sctp_for_each_hentry, sk_for_each, sk_for_each_rcu, sk_for_each_from, sk_for_each_safe, sk_for_each_bound, hlist_for_each_entry_safe, hlist_for_each_entry_continue_rcu, nr_neigh_for_each, nr_neigh_for_each_safe, nr_node_for_each, nr_node_for_each_safe, for_each_gfn_indirect_valid_sp, for_each_gfn_sp, for_each_host; type T; expression a,c,d,e; identifier b; statement S; @@ -T b; <+... when != b ( hlist_for_each_entry(a, - b, c, d) S | hlist_for_each_entry_continue(a, - b, c) S | hlist_for_each_entry_from(a, - b, c) S | hlist_for_each_entry_rcu(a, - b, c, d) S | hlist_for_each_entry_rcu_bh(a, - b, c, d) S | hlist_for_each_entry_continue_rcu_bh(a, - b, c) S | for_each_busy_worker(a, c, - b, d) S | ax25_uid_for_each(a, - b, c) S | ax25_for_each(a, - b, c) S | inet_bind_bucket_for_each(a, - b, c) S | sctp_for_each_hentry(a, - b, c) S | sk_for_each(a, - b, c) S | sk_for_each_rcu(a, - b, c) S | sk_for_each_from -(a, b) +(a) S + sk_for_each_from(a) S | sk_for_each_safe(a, - b, c, d) S | sk_for_each_bound(a, - b, c) S | hlist_for_each_entry_safe(a, - b, c, d, e) S | hlist_for_each_entry_continue_rcu(a, - b, c) S | nr_neigh_for_each(a, - b, c) S | nr_neigh_for_each_safe(a, - b, c, d) S | nr_node_for_each(a, - b, c) S | nr_node_for_each_safe(a, - b, c, d) S | - for_each_gfn_sp(a, c, d, b) S + for_each_gfn_sp(a, c, d) S | - for_each_gfn_indirect_valid_sp(a, c, d, b) S + for_each_gfn_indirect_valid_sp(a, c, d) S | for_each_host(a, - b, c) S | for_each_host_safe(a, - b, c, d) S | for_each_mesh_entry(a, - b, c, d) S ) ...+> [akpm@linux-foundation.org: drop bogus change from net/ipv4/raw.c] [akpm@linux-foundation.org: drop bogus hunk from net/ipv6/raw.c] [akpm@linux-foundation.org: checkpatch fixes] [akpm@linux-foundation.org: fix warnings] [akpm@linux-foudnation.org: redo intrusive kvm changes] Tested-by: Peter Senna Tschudin <peter.senna@gmail.com> Acked-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com> Signed-off-by: Sasha Levin <sasha.levin@oracle.com> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Marcelo Tosatti <mtosatti@redhat.com> Cc: Gleb Natapov <gleb@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-02-28 01:06:00 +00:00
hlist_for_each_entry(node, head, list)
avc_node_delete(node);
rcu_read_unlock();
spin_unlock_irqrestore(lock, flag);
}
}
/**
* avc_ss_reset - Flush the cache and revalidate migrated permissions.
* @seqno: policy sequence number
*/
int avc_ss_reset(u32 seqno)
{
struct avc_callback_node *c;
int rc = 0, tmprc;
avc_flush();
for (c = avc_callbacks; c; c = c->next) {
if (c->events & AVC_CALLBACK_RESET) {
tmprc = c->callback(AVC_CALLBACK_RESET);
/* save the first error encountered for the return
value and continue processing the callbacks */
if (!rc)
rc = tmprc;
}
}
avc_latest_notif_update(seqno, 0);
return rc;
}
selinux: uninline unlikely parts of avc_has_perm_noaudit() This is based on earlier patch posted to the list by Linus, his commit description read: "avc_has_perm_noaudit()is one of those hot functions that end up being used by almost all filesystem operations (through "avc_has_perm()") and it's intended to be cheap enough to inline. However, it turns out that the unlikely parts of it (where it doesn't find an existing avc node) need a fair amount of stack space for the automatic replacement node, so if it were to be inlined (at least clang does not) it would just use stack space unnecessarily. So split the unlikely part out of it, and mark that part noinline. That improves the actual likely part." The basic idea behind the patch was reasonable, but there were minor nits (double indenting, etc.) and the RCU read lock unlock/re-lock in avc_compute_av() began to look even more ugly. This patch builds on Linus' first effort by cleaning things up a bit and removing the RCU unlock/lock dance in avc_compute_av(). Removing the RCU lock dance in avc_compute_av() is safe as there are currently two callers of avc_compute_av(): avc_has_perm_noaudit() and avc_has_extended_perms(). The first caller in avc_has_perm_noaudit() does not require a RCU lock as there is no avc_node to protect so the RCU lock can be dropped before calling avc_compute_av(). The second caller, avc_has_extended_perms(), is similar in that there is no avc_node that requires RCU protection, but the code is simplified by holding the RCU look around the avc_compute_av() call, and given that we enter a RCU critical section in security_compute_av() (called from av_compute_av()) the impact will likely be unnoticeable. It is also worth noting that avc_has_extended_perms() is only called from the SELinux ioctl() access control hook at the moment. Acked-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Paul Moore <paul@paul-moore.com>
2023-03-07 19:55:01 +00:00
/**
* avc_compute_av - Add an entry to the AVC based on the security policy
* @ssid: subject
* @tsid: object/target
* @tclass: object class
* @avd: access vector decision
* @xp_node: AVC extended permissions node
*
selinux: uninline unlikely parts of avc_has_perm_noaudit() This is based on earlier patch posted to the list by Linus, his commit description read: "avc_has_perm_noaudit()is one of those hot functions that end up being used by almost all filesystem operations (through "avc_has_perm()") and it's intended to be cheap enough to inline. However, it turns out that the unlikely parts of it (where it doesn't find an existing avc node) need a fair amount of stack space for the automatic replacement node, so if it were to be inlined (at least clang does not) it would just use stack space unnecessarily. So split the unlikely part out of it, and mark that part noinline. That improves the actual likely part." The basic idea behind the patch was reasonable, but there were minor nits (double indenting, etc.) and the RCU read lock unlock/re-lock in avc_compute_av() began to look even more ugly. This patch builds on Linus' first effort by cleaning things up a bit and removing the RCU unlock/lock dance in avc_compute_av(). Removing the RCU lock dance in avc_compute_av() is safe as there are currently two callers of avc_compute_av(): avc_has_perm_noaudit() and avc_has_extended_perms(). The first caller in avc_has_perm_noaudit() does not require a RCU lock as there is no avc_node to protect so the RCU lock can be dropped before calling avc_compute_av(). The second caller, avc_has_extended_perms(), is similar in that there is no avc_node that requires RCU protection, but the code is simplified by holding the RCU look around the avc_compute_av() call, and given that we enter a RCU critical section in security_compute_av() (called from av_compute_av()) the impact will likely be unnoticeable. It is also worth noting that avc_has_extended_perms() is only called from the SELinux ioctl() access control hook at the moment. Acked-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Paul Moore <paul@paul-moore.com>
2023-03-07 19:55:01 +00:00
* Slow-path helper function for avc_has_perm_noaudit, when the avc_node lookup
* fails. Don't inline this, since it's the slow-path and just results in a
* bigger stack frame.
*/
static noinline void avc_compute_av(u32 ssid, u32 tsid, u16 tclass,
struct av_decision *avd,
struct avc_xperms_node *xp_node)
{
selinux: extended permissions for ioctls Add extended permissions logic to selinux. Extended permissions provides additional permissions in 256 bit increments. Extend the generic ioctl permission check to use the extended permissions for per-command filtering. Source/target/class sets including the ioctl permission may additionally include a set of commands. Example: allowxperm <source> <target>:<class> ioctl unpriv_app_socket_cmds auditallowxperm <source> <target>:<class> ioctl priv_gpu_cmds Where unpriv_app_socket_cmds and priv_gpu_cmds are macros representing commonly granted sets of ioctl commands. When ioctl commands are omitted only the permissions are checked. This feature is intended to provide finer granularity for the ioctl permission that may be too imprecise. For example, the same driver may use ioctls to provide important and benign functionality such as driver version or socket type as well as dangerous capabilities such as debugging features, read/write/execute to physical memory or access to sensitive data. Per-command filtering provides a mechanism to reduce the attack surface of the kernel, and limit applications to the subset of commands required. The format of the policy binary has been modified to include ioctl commands, and the policy version number has been incremented to POLICYDB_VERSION_XPERMS_IOCTL=30 to account for the format change. The extended permissions logic is deliberately generic to allow components to be reused e.g. netlink filters Signed-off-by: Jeff Vander Stoep <jeffv@google.com> Acked-by: Nick Kralevich <nnk@google.com> Signed-off-by: Paul Moore <pmoore@redhat.com>
2015-07-10 21:19:56 +00:00
INIT_LIST_HEAD(&xp_node->xpd_head);
security_compute_av(ssid, tsid, tclass, avd, &xp_node->xp);
avc_insert(ssid, tsid, tclass, avd, xp_node);
}
static noinline int avc_denied(u32 ssid, u32 tsid,
u16 tclass, u32 requested,
u8 driver, u8 xperm, unsigned int flags,
struct av_decision *avd)
{
if (flags & AVC_STRICT)
return -EACCES;
if (enforcing_enabled() &&
!(avd->flags & AVD_FLAGS_PERMISSIVE))
return -EACCES;
avc_update_node(AVC_CALLBACK_GRANT, requested, driver,
xperm, ssid, tsid, tclass, avd->seqno, NULL, flags);
return 0;
}
selinux: extended permissions for ioctls Add extended permissions logic to selinux. Extended permissions provides additional permissions in 256 bit increments. Extend the generic ioctl permission check to use the extended permissions for per-command filtering. Source/target/class sets including the ioctl permission may additionally include a set of commands. Example: allowxperm <source> <target>:<class> ioctl unpriv_app_socket_cmds auditallowxperm <source> <target>:<class> ioctl priv_gpu_cmds Where unpriv_app_socket_cmds and priv_gpu_cmds are macros representing commonly granted sets of ioctl commands. When ioctl commands are omitted only the permissions are checked. This feature is intended to provide finer granularity for the ioctl permission that may be too imprecise. For example, the same driver may use ioctls to provide important and benign functionality such as driver version or socket type as well as dangerous capabilities such as debugging features, read/write/execute to physical memory or access to sensitive data. Per-command filtering provides a mechanism to reduce the attack surface of the kernel, and limit applications to the subset of commands required. The format of the policy binary has been modified to include ioctl commands, and the policy version number has been incremented to POLICYDB_VERSION_XPERMS_IOCTL=30 to account for the format change. The extended permissions logic is deliberately generic to allow components to be reused e.g. netlink filters Signed-off-by: Jeff Vander Stoep <jeffv@google.com> Acked-by: Nick Kralevich <nnk@google.com> Signed-off-by: Paul Moore <pmoore@redhat.com>
2015-07-10 21:19:56 +00:00
/*
* The avc extended permissions logic adds an additional 256 bits of
* permissions to an avc node when extended permissions for that node are
* specified in the avtab. If the additional 256 permissions is not adequate,
* as-is the case with ioctls, then multiple may be chained together and the
* driver field is used to specify which set contains the permission.
*/
int avc_has_extended_perms(u32 ssid, u32 tsid, u16 tclass, u32 requested,
u8 driver, u8 xperm, struct common_audit_data *ad)
selinux: extended permissions for ioctls Add extended permissions logic to selinux. Extended permissions provides additional permissions in 256 bit increments. Extend the generic ioctl permission check to use the extended permissions for per-command filtering. Source/target/class sets including the ioctl permission may additionally include a set of commands. Example: allowxperm <source> <target>:<class> ioctl unpriv_app_socket_cmds auditallowxperm <source> <target>:<class> ioctl priv_gpu_cmds Where unpriv_app_socket_cmds and priv_gpu_cmds are macros representing commonly granted sets of ioctl commands. When ioctl commands are omitted only the permissions are checked. This feature is intended to provide finer granularity for the ioctl permission that may be too imprecise. For example, the same driver may use ioctls to provide important and benign functionality such as driver version or socket type as well as dangerous capabilities such as debugging features, read/write/execute to physical memory or access to sensitive data. Per-command filtering provides a mechanism to reduce the attack surface of the kernel, and limit applications to the subset of commands required. The format of the policy binary has been modified to include ioctl commands, and the policy version number has been incremented to POLICYDB_VERSION_XPERMS_IOCTL=30 to account for the format change. The extended permissions logic is deliberately generic to allow components to be reused e.g. netlink filters Signed-off-by: Jeff Vander Stoep <jeffv@google.com> Acked-by: Nick Kralevich <nnk@google.com> Signed-off-by: Paul Moore <pmoore@redhat.com>
2015-07-10 21:19:56 +00:00
{
struct avc_node *node;
struct av_decision avd;
u32 denied;
struct extended_perms_decision local_xpd;
struct extended_perms_decision *xpd = NULL;
struct extended_perms_data allowed;
struct extended_perms_data auditallow;
struct extended_perms_data dontaudit;
struct avc_xperms_node local_xp_node;
struct avc_xperms_node *xp_node;
int rc = 0, rc2;
xp_node = &local_xp_node;
if (WARN_ON(!requested))
return -EACCES;
selinux: extended permissions for ioctls Add extended permissions logic to selinux. Extended permissions provides additional permissions in 256 bit increments. Extend the generic ioctl permission check to use the extended permissions for per-command filtering. Source/target/class sets including the ioctl permission may additionally include a set of commands. Example: allowxperm <source> <target>:<class> ioctl unpriv_app_socket_cmds auditallowxperm <source> <target>:<class> ioctl priv_gpu_cmds Where unpriv_app_socket_cmds and priv_gpu_cmds are macros representing commonly granted sets of ioctl commands. When ioctl commands are omitted only the permissions are checked. This feature is intended to provide finer granularity for the ioctl permission that may be too imprecise. For example, the same driver may use ioctls to provide important and benign functionality such as driver version or socket type as well as dangerous capabilities such as debugging features, read/write/execute to physical memory or access to sensitive data. Per-command filtering provides a mechanism to reduce the attack surface of the kernel, and limit applications to the subset of commands required. The format of the policy binary has been modified to include ioctl commands, and the policy version number has been incremented to POLICYDB_VERSION_XPERMS_IOCTL=30 to account for the format change. The extended permissions logic is deliberately generic to allow components to be reused e.g. netlink filters Signed-off-by: Jeff Vander Stoep <jeffv@google.com> Acked-by: Nick Kralevich <nnk@google.com> Signed-off-by: Paul Moore <pmoore@redhat.com>
2015-07-10 21:19:56 +00:00
rcu_read_lock();
node = avc_lookup(ssid, tsid, tclass);
selinux: extended permissions for ioctls Add extended permissions logic to selinux. Extended permissions provides additional permissions in 256 bit increments. Extend the generic ioctl permission check to use the extended permissions for per-command filtering. Source/target/class sets including the ioctl permission may additionally include a set of commands. Example: allowxperm <source> <target>:<class> ioctl unpriv_app_socket_cmds auditallowxperm <source> <target>:<class> ioctl priv_gpu_cmds Where unpriv_app_socket_cmds and priv_gpu_cmds are macros representing commonly granted sets of ioctl commands. When ioctl commands are omitted only the permissions are checked. This feature is intended to provide finer granularity for the ioctl permission that may be too imprecise. For example, the same driver may use ioctls to provide important and benign functionality such as driver version or socket type as well as dangerous capabilities such as debugging features, read/write/execute to physical memory or access to sensitive data. Per-command filtering provides a mechanism to reduce the attack surface of the kernel, and limit applications to the subset of commands required. The format of the policy binary has been modified to include ioctl commands, and the policy version number has been incremented to POLICYDB_VERSION_XPERMS_IOCTL=30 to account for the format change. The extended permissions logic is deliberately generic to allow components to be reused e.g. netlink filters Signed-off-by: Jeff Vander Stoep <jeffv@google.com> Acked-by: Nick Kralevich <nnk@google.com> Signed-off-by: Paul Moore <pmoore@redhat.com>
2015-07-10 21:19:56 +00:00
if (unlikely(!node)) {
avc_compute_av(ssid, tsid, tclass, &avd, xp_node);
selinux: extended permissions for ioctls Add extended permissions logic to selinux. Extended permissions provides additional permissions in 256 bit increments. Extend the generic ioctl permission check to use the extended permissions for per-command filtering. Source/target/class sets including the ioctl permission may additionally include a set of commands. Example: allowxperm <source> <target>:<class> ioctl unpriv_app_socket_cmds auditallowxperm <source> <target>:<class> ioctl priv_gpu_cmds Where unpriv_app_socket_cmds and priv_gpu_cmds are macros representing commonly granted sets of ioctl commands. When ioctl commands are omitted only the permissions are checked. This feature is intended to provide finer granularity for the ioctl permission that may be too imprecise. For example, the same driver may use ioctls to provide important and benign functionality such as driver version or socket type as well as dangerous capabilities such as debugging features, read/write/execute to physical memory or access to sensitive data. Per-command filtering provides a mechanism to reduce the attack surface of the kernel, and limit applications to the subset of commands required. The format of the policy binary has been modified to include ioctl commands, and the policy version number has been incremented to POLICYDB_VERSION_XPERMS_IOCTL=30 to account for the format change. The extended permissions logic is deliberately generic to allow components to be reused e.g. netlink filters Signed-off-by: Jeff Vander Stoep <jeffv@google.com> Acked-by: Nick Kralevich <nnk@google.com> Signed-off-by: Paul Moore <pmoore@redhat.com>
2015-07-10 21:19:56 +00:00
} else {
memcpy(&avd, &node->ae.avd, sizeof(avd));
xp_node = node->ae.xp_node;
}
/* if extended permissions are not defined, only consider av_decision */
if (!xp_node || !xp_node->xp.len)
goto decision;
local_xpd.allowed = &allowed;
local_xpd.auditallow = &auditallow;
local_xpd.dontaudit = &dontaudit;
xpd = avc_xperms_decision_lookup(driver, xp_node);
if (unlikely(!xpd)) {
/*
* Compute the extended_perms_decision only if the driver
* is flagged
*/
if (!security_xperm_test(xp_node->xp.drivers.p, driver)) {
avd.allowed &= ~requested;
goto decision;
}
rcu_read_unlock();
security_compute_xperms_decision(ssid, tsid, tclass,
driver, &local_xpd);
selinux: extended permissions for ioctls Add extended permissions logic to selinux. Extended permissions provides additional permissions in 256 bit increments. Extend the generic ioctl permission check to use the extended permissions for per-command filtering. Source/target/class sets including the ioctl permission may additionally include a set of commands. Example: allowxperm <source> <target>:<class> ioctl unpriv_app_socket_cmds auditallowxperm <source> <target>:<class> ioctl priv_gpu_cmds Where unpriv_app_socket_cmds and priv_gpu_cmds are macros representing commonly granted sets of ioctl commands. When ioctl commands are omitted only the permissions are checked. This feature is intended to provide finer granularity for the ioctl permission that may be too imprecise. For example, the same driver may use ioctls to provide important and benign functionality such as driver version or socket type as well as dangerous capabilities such as debugging features, read/write/execute to physical memory or access to sensitive data. Per-command filtering provides a mechanism to reduce the attack surface of the kernel, and limit applications to the subset of commands required. The format of the policy binary has been modified to include ioctl commands, and the policy version number has been incremented to POLICYDB_VERSION_XPERMS_IOCTL=30 to account for the format change. The extended permissions logic is deliberately generic to allow components to be reused e.g. netlink filters Signed-off-by: Jeff Vander Stoep <jeffv@google.com> Acked-by: Nick Kralevich <nnk@google.com> Signed-off-by: Paul Moore <pmoore@redhat.com>
2015-07-10 21:19:56 +00:00
rcu_read_lock();
avc_update_node(AVC_CALLBACK_ADD_XPERMS, requested,
driver, xperm, ssid, tsid, tclass, avd.seqno,
&local_xpd, 0);
selinux: extended permissions for ioctls Add extended permissions logic to selinux. Extended permissions provides additional permissions in 256 bit increments. Extend the generic ioctl permission check to use the extended permissions for per-command filtering. Source/target/class sets including the ioctl permission may additionally include a set of commands. Example: allowxperm <source> <target>:<class> ioctl unpriv_app_socket_cmds auditallowxperm <source> <target>:<class> ioctl priv_gpu_cmds Where unpriv_app_socket_cmds and priv_gpu_cmds are macros representing commonly granted sets of ioctl commands. When ioctl commands are omitted only the permissions are checked. This feature is intended to provide finer granularity for the ioctl permission that may be too imprecise. For example, the same driver may use ioctls to provide important and benign functionality such as driver version or socket type as well as dangerous capabilities such as debugging features, read/write/execute to physical memory or access to sensitive data. Per-command filtering provides a mechanism to reduce the attack surface of the kernel, and limit applications to the subset of commands required. The format of the policy binary has been modified to include ioctl commands, and the policy version number has been incremented to POLICYDB_VERSION_XPERMS_IOCTL=30 to account for the format change. The extended permissions logic is deliberately generic to allow components to be reused e.g. netlink filters Signed-off-by: Jeff Vander Stoep <jeffv@google.com> Acked-by: Nick Kralevich <nnk@google.com> Signed-off-by: Paul Moore <pmoore@redhat.com>
2015-07-10 21:19:56 +00:00
} else {
avc_quick_copy_xperms_decision(xperm, &local_xpd, xpd);
}
xpd = &local_xpd;
if (!avc_xperms_has_perm(xpd, xperm, XPERMS_ALLOWED))
avd.allowed &= ~requested;
decision:
denied = requested & ~(avd.allowed);
if (unlikely(denied))
rc = avc_denied(ssid, tsid, tclass, requested,
driver, xperm, AVC_EXTENDED_PERMS, &avd);
selinux: extended permissions for ioctls Add extended permissions logic to selinux. Extended permissions provides additional permissions in 256 bit increments. Extend the generic ioctl permission check to use the extended permissions for per-command filtering. Source/target/class sets including the ioctl permission may additionally include a set of commands. Example: allowxperm <source> <target>:<class> ioctl unpriv_app_socket_cmds auditallowxperm <source> <target>:<class> ioctl priv_gpu_cmds Where unpriv_app_socket_cmds and priv_gpu_cmds are macros representing commonly granted sets of ioctl commands. When ioctl commands are omitted only the permissions are checked. This feature is intended to provide finer granularity for the ioctl permission that may be too imprecise. For example, the same driver may use ioctls to provide important and benign functionality such as driver version or socket type as well as dangerous capabilities such as debugging features, read/write/execute to physical memory or access to sensitive data. Per-command filtering provides a mechanism to reduce the attack surface of the kernel, and limit applications to the subset of commands required. The format of the policy binary has been modified to include ioctl commands, and the policy version number has been incremented to POLICYDB_VERSION_XPERMS_IOCTL=30 to account for the format change. The extended permissions logic is deliberately generic to allow components to be reused e.g. netlink filters Signed-off-by: Jeff Vander Stoep <jeffv@google.com> Acked-by: Nick Kralevich <nnk@google.com> Signed-off-by: Paul Moore <pmoore@redhat.com>
2015-07-10 21:19:56 +00:00
rcu_read_unlock();
rc2 = avc_xperms_audit(ssid, tsid, tclass, requested,
selinux: extended permissions for ioctls Add extended permissions logic to selinux. Extended permissions provides additional permissions in 256 bit increments. Extend the generic ioctl permission check to use the extended permissions for per-command filtering. Source/target/class sets including the ioctl permission may additionally include a set of commands. Example: allowxperm <source> <target>:<class> ioctl unpriv_app_socket_cmds auditallowxperm <source> <target>:<class> ioctl priv_gpu_cmds Where unpriv_app_socket_cmds and priv_gpu_cmds are macros representing commonly granted sets of ioctl commands. When ioctl commands are omitted only the permissions are checked. This feature is intended to provide finer granularity for the ioctl permission that may be too imprecise. For example, the same driver may use ioctls to provide important and benign functionality such as driver version or socket type as well as dangerous capabilities such as debugging features, read/write/execute to physical memory or access to sensitive data. Per-command filtering provides a mechanism to reduce the attack surface of the kernel, and limit applications to the subset of commands required. The format of the policy binary has been modified to include ioctl commands, and the policy version number has been incremented to POLICYDB_VERSION_XPERMS_IOCTL=30 to account for the format change. The extended permissions logic is deliberately generic to allow components to be reused e.g. netlink filters Signed-off-by: Jeff Vander Stoep <jeffv@google.com> Acked-by: Nick Kralevich <nnk@google.com> Signed-off-by: Paul Moore <pmoore@redhat.com>
2015-07-10 21:19:56 +00:00
&avd, xpd, xperm, rc, ad);
if (rc2)
return rc2;
return rc;
}
selinux: uninline unlikely parts of avc_has_perm_noaudit() This is based on earlier patch posted to the list by Linus, his commit description read: "avc_has_perm_noaudit()is one of those hot functions that end up being used by almost all filesystem operations (through "avc_has_perm()") and it's intended to be cheap enough to inline. However, it turns out that the unlikely parts of it (where it doesn't find an existing avc node) need a fair amount of stack space for the automatic replacement node, so if it were to be inlined (at least clang does not) it would just use stack space unnecessarily. So split the unlikely part out of it, and mark that part noinline. That improves the actual likely part." The basic idea behind the patch was reasonable, but there were minor nits (double indenting, etc.) and the RCU read lock unlock/re-lock in avc_compute_av() began to look even more ugly. This patch builds on Linus' first effort by cleaning things up a bit and removing the RCU unlock/lock dance in avc_compute_av(). Removing the RCU lock dance in avc_compute_av() is safe as there are currently two callers of avc_compute_av(): avc_has_perm_noaudit() and avc_has_extended_perms(). The first caller in avc_has_perm_noaudit() does not require a RCU lock as there is no avc_node to protect so the RCU lock can be dropped before calling avc_compute_av(). The second caller, avc_has_extended_perms(), is similar in that there is no avc_node that requires RCU protection, but the code is simplified by holding the RCU look around the avc_compute_av() call, and given that we enter a RCU critical section in security_compute_av() (called from av_compute_av()) the impact will likely be unnoticeable. It is also worth noting that avc_has_extended_perms() is only called from the SELinux ioctl() access control hook at the moment. Acked-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Paul Moore <paul@paul-moore.com>
2023-03-07 19:55:01 +00:00
/**
* avc_perm_nonode - Add an entry to the AVC
* @ssid: subject
* @tsid: object/target
* @tclass: object class
* @requested: requested permissions
* @flags: AVC flags
* @avd: access vector decision
*
* This is the "we have no node" part of avc_has_perm_noaudit(), which is
* unlikely and needs extra stack space for the new node that we generate, so
* don't inline it.
*/
static noinline int avc_perm_nonode(u32 ssid, u32 tsid, u16 tclass,
selinux: uninline unlikely parts of avc_has_perm_noaudit() This is based on earlier patch posted to the list by Linus, his commit description read: "avc_has_perm_noaudit()is one of those hot functions that end up being used by almost all filesystem operations (through "avc_has_perm()") and it's intended to be cheap enough to inline. However, it turns out that the unlikely parts of it (where it doesn't find an existing avc node) need a fair amount of stack space for the automatic replacement node, so if it were to be inlined (at least clang does not) it would just use stack space unnecessarily. So split the unlikely part out of it, and mark that part noinline. That improves the actual likely part." The basic idea behind the patch was reasonable, but there were minor nits (double indenting, etc.) and the RCU read lock unlock/re-lock in avc_compute_av() began to look even more ugly. This patch builds on Linus' first effort by cleaning things up a bit and removing the RCU unlock/lock dance in avc_compute_av(). Removing the RCU lock dance in avc_compute_av() is safe as there are currently two callers of avc_compute_av(): avc_has_perm_noaudit() and avc_has_extended_perms(). The first caller in avc_has_perm_noaudit() does not require a RCU lock as there is no avc_node to protect so the RCU lock can be dropped before calling avc_compute_av(). The second caller, avc_has_extended_perms(), is similar in that there is no avc_node that requires RCU protection, but the code is simplified by holding the RCU look around the avc_compute_av() call, and given that we enter a RCU critical section in security_compute_av() (called from av_compute_av()) the impact will likely be unnoticeable. It is also worth noting that avc_has_extended_perms() is only called from the SELinux ioctl() access control hook at the moment. Acked-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Paul Moore <paul@paul-moore.com>
2023-03-07 19:55:01 +00:00
u32 requested, unsigned int flags,
struct av_decision *avd)
{
u32 denied;
struct avc_xperms_node xp_node;
avc_compute_av(ssid, tsid, tclass, avd, &xp_node);
selinux: uninline unlikely parts of avc_has_perm_noaudit() This is based on earlier patch posted to the list by Linus, his commit description read: "avc_has_perm_noaudit()is one of those hot functions that end up being used by almost all filesystem operations (through "avc_has_perm()") and it's intended to be cheap enough to inline. However, it turns out that the unlikely parts of it (where it doesn't find an existing avc node) need a fair amount of stack space for the automatic replacement node, so if it were to be inlined (at least clang does not) it would just use stack space unnecessarily. So split the unlikely part out of it, and mark that part noinline. That improves the actual likely part." The basic idea behind the patch was reasonable, but there were minor nits (double indenting, etc.) and the RCU read lock unlock/re-lock in avc_compute_av() began to look even more ugly. This patch builds on Linus' first effort by cleaning things up a bit and removing the RCU unlock/lock dance in avc_compute_av(). Removing the RCU lock dance in avc_compute_av() is safe as there are currently two callers of avc_compute_av(): avc_has_perm_noaudit() and avc_has_extended_perms(). The first caller in avc_has_perm_noaudit() does not require a RCU lock as there is no avc_node to protect so the RCU lock can be dropped before calling avc_compute_av(). The second caller, avc_has_extended_perms(), is similar in that there is no avc_node that requires RCU protection, but the code is simplified by holding the RCU look around the avc_compute_av() call, and given that we enter a RCU critical section in security_compute_av() (called from av_compute_av()) the impact will likely be unnoticeable. It is also worth noting that avc_has_extended_perms() is only called from the SELinux ioctl() access control hook at the moment. Acked-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Paul Moore <paul@paul-moore.com>
2023-03-07 19:55:01 +00:00
denied = requested & ~(avd->allowed);
if (unlikely(denied))
return avc_denied(ssid, tsid, tclass, requested, 0, 0,
selinux: uninline unlikely parts of avc_has_perm_noaudit() This is based on earlier patch posted to the list by Linus, his commit description read: "avc_has_perm_noaudit()is one of those hot functions that end up being used by almost all filesystem operations (through "avc_has_perm()") and it's intended to be cheap enough to inline. However, it turns out that the unlikely parts of it (where it doesn't find an existing avc node) need a fair amount of stack space for the automatic replacement node, so if it were to be inlined (at least clang does not) it would just use stack space unnecessarily. So split the unlikely part out of it, and mark that part noinline. That improves the actual likely part." The basic idea behind the patch was reasonable, but there were minor nits (double indenting, etc.) and the RCU read lock unlock/re-lock in avc_compute_av() began to look even more ugly. This patch builds on Linus' first effort by cleaning things up a bit and removing the RCU unlock/lock dance in avc_compute_av(). Removing the RCU lock dance in avc_compute_av() is safe as there are currently two callers of avc_compute_av(): avc_has_perm_noaudit() and avc_has_extended_perms(). The first caller in avc_has_perm_noaudit() does not require a RCU lock as there is no avc_node to protect so the RCU lock can be dropped before calling avc_compute_av(). The second caller, avc_has_extended_perms(), is similar in that there is no avc_node that requires RCU protection, but the code is simplified by holding the RCU look around the avc_compute_av() call, and given that we enter a RCU critical section in security_compute_av() (called from av_compute_av()) the impact will likely be unnoticeable. It is also worth noting that avc_has_extended_perms() is only called from the SELinux ioctl() access control hook at the moment. Acked-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Paul Moore <paul@paul-moore.com>
2023-03-07 19:55:01 +00:00
flags, avd);
return 0;
}
/**
* avc_has_perm_noaudit - Check permissions but perform no auditing.
* @ssid: source security identifier
* @tsid: target security identifier
* @tclass: target security class
* @requested: requested permissions, interpreted based on @tclass
* @flags: AVC_STRICT or 0
* @avd: access vector decisions
*
* Check the AVC to determine whether the @requested permissions are granted
* for the SID pair (@ssid, @tsid), interpreting the permissions
* based on @tclass, and call the security server on a cache miss to obtain
* a new decision and add it to the cache. Return a copy of the decisions
* in @avd. Return %0 if all @requested permissions are granted,
* -%EACCES if any permissions are denied, or another -errno upon
* other errors. This function is typically called by avc_has_perm(),
* but may also be called directly to separate permission checking from
* auditing, e.g. in cases where a lock must be held for the check but
* should be released for the auditing.
*/
inline int avc_has_perm_noaudit(u32 ssid, u32 tsid,
u16 tclass, u32 requested,
unsigned int flags,
struct av_decision *avd)
{
u32 denied;
selinux: uninline unlikely parts of avc_has_perm_noaudit() This is based on earlier patch posted to the list by Linus, his commit description read: "avc_has_perm_noaudit()is one of those hot functions that end up being used by almost all filesystem operations (through "avc_has_perm()") and it's intended to be cheap enough to inline. However, it turns out that the unlikely parts of it (where it doesn't find an existing avc node) need a fair amount of stack space for the automatic replacement node, so if it were to be inlined (at least clang does not) it would just use stack space unnecessarily. So split the unlikely part out of it, and mark that part noinline. That improves the actual likely part." The basic idea behind the patch was reasonable, but there were minor nits (double indenting, etc.) and the RCU read lock unlock/re-lock in avc_compute_av() began to look even more ugly. This patch builds on Linus' first effort by cleaning things up a bit and removing the RCU unlock/lock dance in avc_compute_av(). Removing the RCU lock dance in avc_compute_av() is safe as there are currently two callers of avc_compute_av(): avc_has_perm_noaudit() and avc_has_extended_perms(). The first caller in avc_has_perm_noaudit() does not require a RCU lock as there is no avc_node to protect so the RCU lock can be dropped before calling avc_compute_av(). The second caller, avc_has_extended_perms(), is similar in that there is no avc_node that requires RCU protection, but the code is simplified by holding the RCU look around the avc_compute_av() call, and given that we enter a RCU critical section in security_compute_av() (called from av_compute_av()) the impact will likely be unnoticeable. It is also worth noting that avc_has_extended_perms() is only called from the SELinux ioctl() access control hook at the moment. Acked-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Paul Moore <paul@paul-moore.com>
2023-03-07 19:55:01 +00:00
struct avc_node *node;
if (WARN_ON(!requested))
return -EACCES;
rcu_read_lock();
node = avc_lookup(ssid, tsid, tclass);
selinux: uninline unlikely parts of avc_has_perm_noaudit() This is based on earlier patch posted to the list by Linus, his commit description read: "avc_has_perm_noaudit()is one of those hot functions that end up being used by almost all filesystem operations (through "avc_has_perm()") and it's intended to be cheap enough to inline. However, it turns out that the unlikely parts of it (where it doesn't find an existing avc node) need a fair amount of stack space for the automatic replacement node, so if it were to be inlined (at least clang does not) it would just use stack space unnecessarily. So split the unlikely part out of it, and mark that part noinline. That improves the actual likely part." The basic idea behind the patch was reasonable, but there were minor nits (double indenting, etc.) and the RCU read lock unlock/re-lock in avc_compute_av() began to look even more ugly. This patch builds on Linus' first effort by cleaning things up a bit and removing the RCU unlock/lock dance in avc_compute_av(). Removing the RCU lock dance in avc_compute_av() is safe as there are currently two callers of avc_compute_av(): avc_has_perm_noaudit() and avc_has_extended_perms(). The first caller in avc_has_perm_noaudit() does not require a RCU lock as there is no avc_node to protect so the RCU lock can be dropped before calling avc_compute_av(). The second caller, avc_has_extended_perms(), is similar in that there is no avc_node that requires RCU protection, but the code is simplified by holding the RCU look around the avc_compute_av() call, and given that we enter a RCU critical section in security_compute_av() (called from av_compute_av()) the impact will likely be unnoticeable. It is also worth noting that avc_has_extended_perms() is only called from the SELinux ioctl() access control hook at the moment. Acked-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Paul Moore <paul@paul-moore.com>
2023-03-07 19:55:01 +00:00
if (unlikely(!node)) {
rcu_read_unlock();
return avc_perm_nonode(ssid, tsid, tclass, requested,
selinux: uninline unlikely parts of avc_has_perm_noaudit() This is based on earlier patch posted to the list by Linus, his commit description read: "avc_has_perm_noaudit()is one of those hot functions that end up being used by almost all filesystem operations (through "avc_has_perm()") and it's intended to be cheap enough to inline. However, it turns out that the unlikely parts of it (where it doesn't find an existing avc node) need a fair amount of stack space for the automatic replacement node, so if it were to be inlined (at least clang does not) it would just use stack space unnecessarily. So split the unlikely part out of it, and mark that part noinline. That improves the actual likely part." The basic idea behind the patch was reasonable, but there were minor nits (double indenting, etc.) and the RCU read lock unlock/re-lock in avc_compute_av() began to look even more ugly. This patch builds on Linus' first effort by cleaning things up a bit and removing the RCU unlock/lock dance in avc_compute_av(). Removing the RCU lock dance in avc_compute_av() is safe as there are currently two callers of avc_compute_av(): avc_has_perm_noaudit() and avc_has_extended_perms(). The first caller in avc_has_perm_noaudit() does not require a RCU lock as there is no avc_node to protect so the RCU lock can be dropped before calling avc_compute_av(). The second caller, avc_has_extended_perms(), is similar in that there is no avc_node that requires RCU protection, but the code is simplified by holding the RCU look around the avc_compute_av() call, and given that we enter a RCU critical section in security_compute_av() (called from av_compute_av()) the impact will likely be unnoticeable. It is also worth noting that avc_has_extended_perms() is only called from the SELinux ioctl() access control hook at the moment. Acked-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Paul Moore <paul@paul-moore.com>
2023-03-07 19:55:01 +00:00
flags, avd);
}
denied = requested & ~node->ae.avd.allowed;
memcpy(avd, &node->ae.avd, sizeof(*avd));
rcu_read_unlock();
if (unlikely(denied))
return avc_denied(ssid, tsid, tclass, requested, 0, 0,
selinux: uninline unlikely parts of avc_has_perm_noaudit() This is based on earlier patch posted to the list by Linus, his commit description read: "avc_has_perm_noaudit()is one of those hot functions that end up being used by almost all filesystem operations (through "avc_has_perm()") and it's intended to be cheap enough to inline. However, it turns out that the unlikely parts of it (where it doesn't find an existing avc node) need a fair amount of stack space for the automatic replacement node, so if it were to be inlined (at least clang does not) it would just use stack space unnecessarily. So split the unlikely part out of it, and mark that part noinline. That improves the actual likely part." The basic idea behind the patch was reasonable, but there were minor nits (double indenting, etc.) and the RCU read lock unlock/re-lock in avc_compute_av() began to look even more ugly. This patch builds on Linus' first effort by cleaning things up a bit and removing the RCU unlock/lock dance in avc_compute_av(). Removing the RCU lock dance in avc_compute_av() is safe as there are currently two callers of avc_compute_av(): avc_has_perm_noaudit() and avc_has_extended_perms(). The first caller in avc_has_perm_noaudit() does not require a RCU lock as there is no avc_node to protect so the RCU lock can be dropped before calling avc_compute_av(). The second caller, avc_has_extended_perms(), is similar in that there is no avc_node that requires RCU protection, but the code is simplified by holding the RCU look around the avc_compute_av() call, and given that we enter a RCU critical section in security_compute_av() (called from av_compute_av()) the impact will likely be unnoticeable. It is also worth noting that avc_has_extended_perms() is only called from the SELinux ioctl() access control hook at the moment. Acked-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Paul Moore <paul@paul-moore.com>
2023-03-07 19:55:01 +00:00
flags, avd);
return 0;
}
/**
* avc_has_perm - Check permissions and perform any appropriate auditing.
* @ssid: source security identifier
* @tsid: target security identifier
* @tclass: target security class
* @requested: requested permissions, interpreted based on @tclass
* @auditdata: auxiliary audit data
*
* Check the AVC to determine whether the @requested permissions are granted
* for the SID pair (@ssid, @tsid), interpreting the permissions
* based on @tclass, and call the security server on a cache miss to obtain
* a new decision and add it to the cache. Audit the granting or denial of
* permissions in accordance with the policy. Return %0 if all @requested
* permissions are granted, -%EACCES if any permissions are denied, or
* another -errno upon other errors.
*/
int avc_has_perm(u32 ssid, u32 tsid, u16 tclass,
u32 requested, struct common_audit_data *auditdata)
{
struct av_decision avd;
int rc, rc2;
rc = avc_has_perm_noaudit(ssid, tsid, tclass, requested, 0,
&avd);
rc2 = avc_audit(ssid, tsid, tclass, requested, &avd, rc,
auditdata);
if (rc2)
return rc2;
return rc;
}
u32 avc_policy_seqno(void)
{
return selinux_avc.avc_cache.latest_notif;
}
void avc_disable(void)
{
SELinux: do not destroy the avc_cache_nodep The security_ops reset done when SELinux is disabled at run time is done after the avc cache is freed and after the kmem_cache for the avc is also freed. This means that between the time the selinux disable code destroys the avc_node_cachep another process could make a security request and could try to allocate from the cache. We are just going to leave the cachep around, like we always have. SELinux: Disabled at runtime. BUG: unable to handle kernel NULL pointer dereference at (null) IP: [<ffffffff81122537>] kmem_cache_alloc+0x9a/0x185 PGD 0 Oops: 0000 [#1] PREEMPT SMP DEBUG_PAGEALLOC last sysfs file: CPU 1 Modules linked in: Pid: 12, comm: khelper Not tainted 2.6.31-tip-05525-g0eeacc6-dirty #14819 System Product Name RIP: 0010:[<ffffffff81122537>] [<ffffffff81122537>] kmem_cache_alloc+0x9a/0x185 RSP: 0018:ffff88003f9258b0 EFLAGS: 00010086 RAX: 0000000000000001 RBX: 0000000000000000 RCX: 0000000078c0129e RDX: 0000000000000000 RSI: ffffffff8130b626 RDI: ffffffff81122528 RBP: ffff88003f925900 R08: 0000000078c0129e R09: 0000000000000001 R10: 0000000000000000 R11: 0000000078c0129e R12: 0000000000000246 R13: 0000000000008020 R14: ffff88003f8586d8 R15: 0000000000000001 FS: 0000000000000000(0000) GS:ffff880002b00000(0000) knlGS:0000000000000000 CS: 0010 DS: 0018 ES: 0018 CR0: 000000008005003b CR2: 0000000000000000 CR3: 0000000001001000 CR4: 00000000000006e0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: ffffffff827bd420 DR6: 00000000ffff0ff0 DR7: 0000000000000400 Process khelper (pid: 12, threadinfo ffff88003f924000, task ffff88003f928000) Stack: 0000000000000246 0000802000000246 ffffffff8130b626 0000000000000001 <0> 0000000078c0129e 0000000000000000 ffff88003f925a70 0000000000000002 <0> 0000000000000001 0000000000000001 ffff88003f925960 ffffffff8130b626 Call Trace: [<ffffffff8130b626>] ? avc_alloc_node+0x36/0x273 [<ffffffff8130b626>] avc_alloc_node+0x36/0x273 [<ffffffff8130b545>] ? avc_latest_notif_update+0x7d/0x9e [<ffffffff8130b8b4>] avc_insert+0x51/0x18d [<ffffffff8130bcce>] avc_has_perm_noaudit+0x9d/0x128 [<ffffffff8130bf20>] avc_has_perm+0x45/0x88 [<ffffffff8130f99d>] current_has_perm+0x52/0x6d [<ffffffff8130fbb2>] selinux_task_create+0x2f/0x45 [<ffffffff81303bf7>] security_task_create+0x29/0x3f [<ffffffff8105c6ba>] copy_process+0x82/0xdf0 [<ffffffff81091578>] ? register_lock_class+0x2f/0x36c [<ffffffff81091a13>] ? mark_lock+0x2e/0x1e1 [<ffffffff8105d596>] do_fork+0x16e/0x382 [<ffffffff81091578>] ? register_lock_class+0x2f/0x36c [<ffffffff810d9166>] ? probe_workqueue_execution+0x57/0xf9 [<ffffffff81091a13>] ? mark_lock+0x2e/0x1e1 [<ffffffff810d9166>] ? probe_workqueue_execution+0x57/0xf9 [<ffffffff8100cdb2>] kernel_thread+0x82/0xe0 [<ffffffff81078b1f>] ? ____call_usermodehelper+0x0/0x139 [<ffffffff8100ce10>] ? child_rip+0x0/0x20 [<ffffffff81078aea>] ? __call_usermodehelper+0x65/0x9a [<ffffffff8107a5c7>] run_workqueue+0x171/0x27e [<ffffffff8107a573>] ? run_workqueue+0x11d/0x27e [<ffffffff81078a85>] ? __call_usermodehelper+0x0/0x9a [<ffffffff8107a7bc>] worker_thread+0xe8/0x10f [<ffffffff810808e2>] ? autoremove_wake_function+0x0/0x63 [<ffffffff8107a6d4>] ? worker_thread+0x0/0x10f [<ffffffff8108042e>] kthread+0x91/0x99 [<ffffffff8100ce1a>] child_rip+0xa/0x20 [<ffffffff8100c754>] ? restore_args+0x0/0x30 [<ffffffff8108039d>] ? kthread+0x0/0x99 [<ffffffff8100ce10>] ? child_rip+0x0/0x20 Code: 0f 85 99 00 00 00 9c 58 66 66 90 66 90 49 89 c4 fa 66 66 90 66 66 90 e8 83 34 fb ff e8 d7 e9 26 00 48 98 49 8b 94 c6 10 01 00 00 <48> 8b 1a 44 8b 7a 18 48 85 db 74 0f 8b 42 14 48 8b 04 c3 ff 42 RIP [<ffffffff81122537>] kmem_cache_alloc+0x9a/0x185 RSP <ffff88003f9258b0> CR2: 0000000000000000 ---[ end trace 42f41a982344e606 ]--- Reported-by: Ingo Molnar <mingo@elte.hu> Signed-off-by: Eric Paris <eparis@redhat.com> Signed-off-by: James Morris <jmorris@namei.org>
2009-09-21 01:21:10 +00:00
/*
* If you are looking at this because you have realized that we are
* not destroying the avc_node_cachep it might be easy to fix, but
* I don't know the memory barrier semantics well enough to know. It's
* possible that some other task dereferenced security_ops when
* it still pointed to selinux operations. If that is the case it's
* possible that it is about to use the avc and is about to need the
* avc_node_cachep. I know I could wrap the security.c security_ops call
* in an rcu_lock, but seriously, it's not worth it. Instead I just flush
* the cache and get that memory back.
*/
if (avc_node_cachep) {
avc_flush();
SELinux: do not destroy the avc_cache_nodep The security_ops reset done when SELinux is disabled at run time is done after the avc cache is freed and after the kmem_cache for the avc is also freed. This means that between the time the selinux disable code destroys the avc_node_cachep another process could make a security request and could try to allocate from the cache. We are just going to leave the cachep around, like we always have. SELinux: Disabled at runtime. BUG: unable to handle kernel NULL pointer dereference at (null) IP: [<ffffffff81122537>] kmem_cache_alloc+0x9a/0x185 PGD 0 Oops: 0000 [#1] PREEMPT SMP DEBUG_PAGEALLOC last sysfs file: CPU 1 Modules linked in: Pid: 12, comm: khelper Not tainted 2.6.31-tip-05525-g0eeacc6-dirty #14819 System Product Name RIP: 0010:[<ffffffff81122537>] [<ffffffff81122537>] kmem_cache_alloc+0x9a/0x185 RSP: 0018:ffff88003f9258b0 EFLAGS: 00010086 RAX: 0000000000000001 RBX: 0000000000000000 RCX: 0000000078c0129e RDX: 0000000000000000 RSI: ffffffff8130b626 RDI: ffffffff81122528 RBP: ffff88003f925900 R08: 0000000078c0129e R09: 0000000000000001 R10: 0000000000000000 R11: 0000000078c0129e R12: 0000000000000246 R13: 0000000000008020 R14: ffff88003f8586d8 R15: 0000000000000001 FS: 0000000000000000(0000) GS:ffff880002b00000(0000) knlGS:0000000000000000 CS: 0010 DS: 0018 ES: 0018 CR0: 000000008005003b CR2: 0000000000000000 CR3: 0000000001001000 CR4: 00000000000006e0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: ffffffff827bd420 DR6: 00000000ffff0ff0 DR7: 0000000000000400 Process khelper (pid: 12, threadinfo ffff88003f924000, task ffff88003f928000) Stack: 0000000000000246 0000802000000246 ffffffff8130b626 0000000000000001 <0> 0000000078c0129e 0000000000000000 ffff88003f925a70 0000000000000002 <0> 0000000000000001 0000000000000001 ffff88003f925960 ffffffff8130b626 Call Trace: [<ffffffff8130b626>] ? avc_alloc_node+0x36/0x273 [<ffffffff8130b626>] avc_alloc_node+0x36/0x273 [<ffffffff8130b545>] ? avc_latest_notif_update+0x7d/0x9e [<ffffffff8130b8b4>] avc_insert+0x51/0x18d [<ffffffff8130bcce>] avc_has_perm_noaudit+0x9d/0x128 [<ffffffff8130bf20>] avc_has_perm+0x45/0x88 [<ffffffff8130f99d>] current_has_perm+0x52/0x6d [<ffffffff8130fbb2>] selinux_task_create+0x2f/0x45 [<ffffffff81303bf7>] security_task_create+0x29/0x3f [<ffffffff8105c6ba>] copy_process+0x82/0xdf0 [<ffffffff81091578>] ? register_lock_class+0x2f/0x36c [<ffffffff81091a13>] ? mark_lock+0x2e/0x1e1 [<ffffffff8105d596>] do_fork+0x16e/0x382 [<ffffffff81091578>] ? register_lock_class+0x2f/0x36c [<ffffffff810d9166>] ? probe_workqueue_execution+0x57/0xf9 [<ffffffff81091a13>] ? mark_lock+0x2e/0x1e1 [<ffffffff810d9166>] ? probe_workqueue_execution+0x57/0xf9 [<ffffffff8100cdb2>] kernel_thread+0x82/0xe0 [<ffffffff81078b1f>] ? ____call_usermodehelper+0x0/0x139 [<ffffffff8100ce10>] ? child_rip+0x0/0x20 [<ffffffff81078aea>] ? __call_usermodehelper+0x65/0x9a [<ffffffff8107a5c7>] run_workqueue+0x171/0x27e [<ffffffff8107a573>] ? run_workqueue+0x11d/0x27e [<ffffffff81078a85>] ? __call_usermodehelper+0x0/0x9a [<ffffffff8107a7bc>] worker_thread+0xe8/0x10f [<ffffffff810808e2>] ? autoremove_wake_function+0x0/0x63 [<ffffffff8107a6d4>] ? worker_thread+0x0/0x10f [<ffffffff8108042e>] kthread+0x91/0x99 [<ffffffff8100ce1a>] child_rip+0xa/0x20 [<ffffffff8100c754>] ? restore_args+0x0/0x30 [<ffffffff8108039d>] ? kthread+0x0/0x99 [<ffffffff8100ce10>] ? child_rip+0x0/0x20 Code: 0f 85 99 00 00 00 9c 58 66 66 90 66 90 49 89 c4 fa 66 66 90 66 66 90 e8 83 34 fb ff e8 d7 e9 26 00 48 98 49 8b 94 c6 10 01 00 00 <48> 8b 1a 44 8b 7a 18 48 85 db 74 0f 8b 42 14 48 8b 04 c3 ff 42 RIP [<ffffffff81122537>] kmem_cache_alloc+0x9a/0x185 RSP <ffff88003f9258b0> CR2: 0000000000000000 ---[ end trace 42f41a982344e606 ]--- Reported-by: Ingo Molnar <mingo@elte.hu> Signed-off-by: Eric Paris <eparis@redhat.com> Signed-off-by: James Morris <jmorris@namei.org>
2009-09-21 01:21:10 +00:00
/* kmem_cache_destroy(avc_node_cachep); */
}
}