CRED: Use RCU to access another task's creds and to release a task's own creds
Use RCU to access another task's creds and to release a task's own creds.
This means that it will be possible for the credentials of a task to be
replaced without another task (a) requiring a full lock to read them, and (b)
seeing deallocated memory.
Signed-off-by: David Howells <dhowells@redhat.com>
Acked-by: James Morris <jmorris@namei.org>
Acked-by: Serge Hallyn <serue@us.ibm.com>
Signed-off-by: James Morris <jmorris@namei.org>
diff --git a/kernel/auditsc.c b/kernel/auditsc.c
index 2febf51..ae8ef88 100644
--- a/kernel/auditsc.c
+++ b/kernel/auditsc.c
@@ -447,7 +447,7 @@
struct audit_names *name,
enum audit_state *state)
{
- struct cred *cred = tsk->cred;
+ const struct cred *cred = get_task_cred(tsk);
int i, j, need_sid = 1;
u32 sid;
@@ -642,8 +642,10 @@
break;
}
- if (!result)
+ if (!result) {
+ put_cred(cred);
return 0;
+ }
}
if (rule->filterkey && ctx)
ctx->filterkey = kstrdup(rule->filterkey, GFP_ATOMIC);
@@ -651,6 +653,7 @@
case AUDIT_NEVER: *state = AUDIT_DISABLED; break;
case AUDIT_ALWAYS: *state = AUDIT_RECORD_CONTEXT; break;
}
+ put_cred(cred);
return 1;
}
@@ -1229,7 +1232,7 @@
static void audit_log_exit(struct audit_context *context, struct task_struct *tsk)
{
- struct cred *cred = tsk->cred;
+ const struct cred *cred;
int i, call_panic = 0;
struct audit_buffer *ab;
struct audit_aux_data *aux;
@@ -1239,13 +1242,14 @@
context->pid = tsk->pid;
if (!context->ppid)
context->ppid = sys_getppid();
- context->uid = cred->uid;
- context->gid = cred->gid;
- context->euid = cred->euid;
- context->suid = cred->suid;
+ cred = current_cred();
+ context->uid = cred->uid;
+ context->gid = cred->gid;
+ context->euid = cred->euid;
+ context->suid = cred->suid;
context->fsuid = cred->fsuid;
- context->egid = cred->egid;
- context->sgid = cred->sgid;
+ context->egid = cred->egid;
+ context->sgid = cred->sgid;
context->fsgid = cred->fsgid;
context->personality = tsk->personality;
@@ -2088,7 +2092,7 @@
audit_log_format(ab, "login pid=%d uid=%u "
"old auid=%u new auid=%u"
" old ses=%u new ses=%u",
- task->pid, task->cred->uid,
+ task->pid, task_uid(task),
task->loginuid, loginuid,
task->sessionid, sessionid);
audit_log_end(ab);
@@ -2471,7 +2475,7 @@
context->target_pid = t->pid;
context->target_auid = audit_get_loginuid(t);
- context->target_uid = t->cred->uid;
+ context->target_uid = task_uid(t);
context->target_sessionid = audit_get_sessionid(t);
security_task_getsecid(t, &context->target_sid);
memcpy(context->target_comm, t->comm, TASK_COMM_LEN);
@@ -2490,6 +2494,7 @@
struct audit_aux_data_pids *axp;
struct task_struct *tsk = current;
struct audit_context *ctx = tsk->audit_context;
+ uid_t uid = current_uid(), t_uid = task_uid(t);
if (audit_pid && t->tgid == audit_pid) {
if (sig == SIGTERM || sig == SIGHUP || sig == SIGUSR1 || sig == SIGUSR2) {
@@ -2497,7 +2502,7 @@
if (tsk->loginuid != -1)
audit_sig_uid = tsk->loginuid;
else
- audit_sig_uid = tsk->cred->uid;
+ audit_sig_uid = uid;
security_task_getsecid(tsk, &audit_sig_sid);
}
if (!audit_signals || audit_dummy_context())
@@ -2509,7 +2514,7 @@
if (!ctx->target_pid) {
ctx->target_pid = t->tgid;
ctx->target_auid = audit_get_loginuid(t);
- ctx->target_uid = t->cred->uid;
+ ctx->target_uid = t_uid;
ctx->target_sessionid = audit_get_sessionid(t);
security_task_getsecid(t, &ctx->target_sid);
memcpy(ctx->target_comm, t->comm, TASK_COMM_LEN);
@@ -2530,7 +2535,7 @@
axp->target_pid[axp->pid_count] = t->tgid;
axp->target_auid[axp->pid_count] = audit_get_loginuid(t);
- axp->target_uid[axp->pid_count] = t->cred->uid;
+ axp->target_uid[axp->pid_count] = t_uid;
axp->target_sessionid[axp->pid_count] = audit_get_sessionid(t);
security_task_getsecid(t, &axp->target_sid[axp->pid_count]);
memcpy(axp->target_comm[axp->pid_count], t->comm, TASK_COMM_LEN);
diff --git a/kernel/cgroup.c b/kernel/cgroup.c
index e210526..a512a75 100644
--- a/kernel/cgroup.c
+++ b/kernel/cgroup.c
@@ -1279,7 +1279,7 @@
static int attach_task_by_pid(struct cgroup *cgrp, u64 pid)
{
struct task_struct *tsk;
- uid_t euid;
+ const struct cred *cred = current_cred(), *tcred;
int ret;
if (pid) {
@@ -1289,16 +1289,16 @@
rcu_read_unlock();
return -ESRCH;
}
- get_task_struct(tsk);
- rcu_read_unlock();
- euid = current_euid();
- if (euid &&
- euid != tsk->cred->uid &&
- euid != tsk->cred->suid) {
- put_task_struct(tsk);
+ tcred = __task_cred(tsk);
+ if (cred->euid &&
+ cred->euid != tcred->uid &&
+ cred->euid != tcred->suid) {
+ rcu_read_unlock();
return -EACCES;
}
+ get_task_struct(tsk);
+ rcu_read_unlock();
} else {
tsk = current;
get_task_struct(tsk);
diff --git a/kernel/exit.c b/kernel/exit.c
index e0f6e18..bbc2253 100644
--- a/kernel/exit.c
+++ b/kernel/exit.c
@@ -160,7 +160,10 @@
int zap_leader;
repeat:
tracehook_prepare_release_task(p);
- atomic_dec(&p->cred->user->processes);
+ /* don't need to get the RCU readlock here - the process is dead and
+ * can't be modifying its own credentials */
+ atomic_dec(&__task_cred(p)->user->processes);
+
proc_flush_task(p);
write_lock_irq(&tasklist_lock);
tracehook_finish_release_task(p);
@@ -1267,12 +1270,12 @@
unsigned long state;
int retval, status, traced;
pid_t pid = task_pid_vnr(p);
+ uid_t uid = __task_cred(p)->uid;
if (!likely(options & WEXITED))
return 0;
if (unlikely(options & WNOWAIT)) {
- uid_t uid = p->cred->uid;
int exit_code = p->exit_code;
int why, status;
@@ -1393,7 +1396,7 @@
if (!retval && infop)
retval = put_user(pid, &infop->si_pid);
if (!retval && infop)
- retval = put_user(p->cred->uid, &infop->si_uid);
+ retval = put_user(uid, &infop->si_uid);
if (!retval)
retval = pid;
@@ -1458,7 +1461,8 @@
if (!unlikely(options & WNOWAIT))
p->exit_code = 0;
- uid = p->cred->uid;
+ /* don't need the RCU readlock here as we're holding a spinlock */
+ uid = __task_cred(p)->uid;
unlock_sig:
spin_unlock_irq(&p->sighand->siglock);
if (!exit_code)
@@ -1532,10 +1536,10 @@
}
if (!unlikely(options & WNOWAIT))
p->signal->flags &= ~SIGNAL_STOP_CONTINUED;
+ uid = __task_cred(p)->uid;
spin_unlock_irq(&p->sighand->siglock);
pid = task_pid_vnr(p);
- uid = p->cred->uid;
get_task_struct(p);
read_unlock(&tasklist_lock);
diff --git a/kernel/futex.c b/kernel/futex.c
index 28421d8..4fe790e 100644
--- a/kernel/futex.c
+++ b/kernel/futex.c
@@ -439,15 +439,20 @@
static struct task_struct * futex_find_get_task(pid_t pid)
{
struct task_struct *p;
- uid_t euid = current_euid();
+ const struct cred *cred = current_cred(), *pcred;
rcu_read_lock();
p = find_task_by_vpid(pid);
- if (!p || (euid != p->cred->euid &&
- euid != p->cred->uid))
+ if (!p) {
p = ERR_PTR(-ESRCH);
- else
- get_task_struct(p);
+ } else {
+ pcred = __task_cred(p);
+ if (cred->euid != pcred->euid &&
+ cred->euid != pcred->uid)
+ p = ERR_PTR(-ESRCH);
+ else
+ get_task_struct(p);
+ }
rcu_read_unlock();
@@ -1831,7 +1836,7 @@
{
struct robust_list_head __user *head;
unsigned long ret;
- uid_t euid = current_euid();
+ const struct cred *cred = current_cred(), *pcred;
if (!futex_cmpxchg_enabled)
return -ENOSYS;
@@ -1847,8 +1852,9 @@
if (!p)
goto err_unlock;
ret = -EPERM;
- if (euid != p->cred->euid &&
- euid != p->cred->uid &&
+ pcred = __task_cred(p);
+ if (cred->euid != pcred->euid &&
+ cred->euid != pcred->uid &&
!capable(CAP_SYS_PTRACE))
goto err_unlock;
head = p->robust_list;
diff --git a/kernel/futex_compat.c b/kernel/futex_compat.c
index 2c3fd5e..d607a5b 100644
--- a/kernel/futex_compat.c
+++ b/kernel/futex_compat.c
@@ -135,7 +135,7 @@
{
struct compat_robust_list_head __user *head;
unsigned long ret;
- uid_t euid = current_euid();
+ const struct cred *cred = current_cred(), *pcred;
if (!futex_cmpxchg_enabled)
return -ENOSYS;
@@ -151,8 +151,9 @@
if (!p)
goto err_unlock;
ret = -EPERM;
- if (euid != p->cred->euid &&
- euid != p->cred->uid &&
+ pcred = __task_cred(p);
+ if (cred->euid != pcred->euid &&
+ cred->euid != pcred->uid &&
!capable(CAP_SYS_PTRACE))
goto err_unlock;
head = p->compat_robust_list;
diff --git a/kernel/ptrace.c b/kernel/ptrace.c
index 49849d1..b9d5f4e 100644
--- a/kernel/ptrace.c
+++ b/kernel/ptrace.c
@@ -115,7 +115,7 @@
int __ptrace_may_access(struct task_struct *task, unsigned int mode)
{
- struct cred *cred = current->cred, *tcred = task->cred;
+ const struct cred *cred = current_cred(), *tcred;
/* May we inspect the given task?
* This check is used both for attaching with ptrace
@@ -125,19 +125,23 @@
* because setting up the necessary parent/child relationship
* or halting the specified task is impossible.
*/
- uid_t uid = cred->uid;
- gid_t gid = cred->gid;
int dumpable = 0;
/* Don't let security modules deny introspection */
if (task == current)
return 0;
- if ((uid != tcred->euid ||
- uid != tcred->suid ||
- uid != tcred->uid ||
- gid != tcred->egid ||
- gid != tcred->sgid ||
- gid != tcred->gid) && !capable(CAP_SYS_PTRACE))
+ rcu_read_lock();
+ tcred = __task_cred(task);
+ if ((cred->uid != tcred->euid ||
+ cred->uid != tcred->suid ||
+ cred->uid != tcred->uid ||
+ cred->gid != tcred->egid ||
+ cred->gid != tcred->sgid ||
+ cred->gid != tcred->gid) &&
+ !capable(CAP_SYS_PTRACE)) {
+ rcu_read_unlock();
return -EPERM;
+ }
+ rcu_read_unlock();
smp_rmb();
if (task->mm)
dumpable = get_dumpable(task->mm);
diff --git a/kernel/sched.c b/kernel/sched.c
index 733c59e..92992e2 100644
--- a/kernel/sched.c
+++ b/kernel/sched.c
@@ -345,7 +345,9 @@
struct task_group *tg;
#ifdef CONFIG_USER_SCHED
- tg = p->cred->user->tg;
+ rcu_read_lock();
+ tg = __task_cred(p)->user->tg;
+ rcu_read_unlock();
#elif defined(CONFIG_CGROUP_SCHED)
tg = container_of(task_subsys_state(p, cpu_cgroup_subsys_id),
struct task_group, css);
@@ -5121,6 +5123,22 @@
set_load_weight(p);
}
+/*
+ * check the target process has a UID that matches the current process's
+ */
+static bool check_same_owner(struct task_struct *p)
+{
+ const struct cred *cred = current_cred(), *pcred;
+ bool match;
+
+ rcu_read_lock();
+ pcred = __task_cred(p);
+ match = (cred->euid == pcred->euid ||
+ cred->euid == pcred->uid);
+ rcu_read_unlock();
+ return match;
+}
+
static int __sched_setscheduler(struct task_struct *p, int policy,
struct sched_param *param, bool user)
{
@@ -5128,7 +5146,6 @@
unsigned long flags;
const struct sched_class *prev_class = p->sched_class;
struct rq *rq;
- uid_t euid;
/* may grab non-irq protected spin_locks */
BUG_ON(in_interrupt());
@@ -5181,9 +5198,7 @@
return -EPERM;
/* can't change other user's priorities */
- euid = current_euid();
- if (euid != p->cred->euid &&
- euid != p->cred->uid)
+ if (!check_same_owner(p))
return -EPERM;
}
@@ -5394,7 +5409,6 @@
cpumask_t cpus_allowed;
cpumask_t new_mask = *in_mask;
struct task_struct *p;
- uid_t euid;
int retval;
get_online_cpus();
@@ -5415,11 +5429,8 @@
get_task_struct(p);
read_unlock(&tasklist_lock);
- euid = current_euid();
retval = -EPERM;
- if (euid != p->cred->euid &&
- euid != p->cred->uid &&
- !capable(CAP_SYS_NICE))
+ if (!check_same_owner(p) && !capable(CAP_SYS_NICE))
goto out_unlock;
retval = security_task_setscheduler(p, 0, NULL);
diff --git a/kernel/signal.c b/kernel/signal.c
index 80e8a648..8498912 100644
--- a/kernel/signal.c
+++ b/kernel/signal.c
@@ -177,6 +177,11 @@
return sig;
}
+/*
+ * allocate a new signal queue record
+ * - this may be called without locks if and only if t == current, otherwise an
+ * appopriate lock must be held to protect t's user_struct
+ */
static struct sigqueue *__sigqueue_alloc(struct task_struct *t, gfp_t flags,
int override_rlimit)
{
@@ -184,11 +189,12 @@
struct user_struct *user;
/*
- * In order to avoid problems with "switch_user()", we want to make
- * sure that the compiler doesn't re-load "t->user"
+ * We won't get problems with the target's UID changing under us
+ * because changing it requires RCU be used, and if t != current, the
+ * caller must be holding the RCU readlock (by way of a spinlock) and
+ * we use RCU protection here
*/
- user = t->cred->user;
- barrier();
+ user = __task_cred(t)->user;
atomic_inc(&user->sigpending);
if (override_rlimit ||
atomic_read(&user->sigpending) <=
@@ -562,12 +568,13 @@
/*
* Bad permissions for sending the signal
+ * - the caller must hold at least the RCU read lock
*/
static int check_kill_permission(int sig, struct siginfo *info,
struct task_struct *t)
{
+ const struct cred *cred = current_cred(), *tcred;
struct pid *sid;
- uid_t uid, euid;
int error;
if (!valid_signal(sig))
@@ -580,10 +587,11 @@
if (error)
return error;
- uid = current_uid();
- euid = current_euid();
- if ((euid ^ t->cred->suid) && (euid ^ t->cred->uid) &&
- (uid ^ t->cred->suid) && (uid ^ t->cred->uid) &&
+ tcred = __task_cred(t);
+ if ((cred->euid ^ tcred->suid) &&
+ (cred->euid ^ tcred->uid) &&
+ (cred->uid ^ tcred->suid) &&
+ (cred->uid ^ tcred->uid) &&
!capable(CAP_KILL)) {
switch (sig) {
case SIGCONT:
@@ -1011,6 +1019,10 @@
return sighand;
}
+/*
+ * send signal info to all the members of a group
+ * - the caller must hold the RCU read lock at least
+ */
int group_send_sig_info(int sig, struct siginfo *info, struct task_struct *p)
{
unsigned long flags;
@@ -1032,8 +1044,8 @@
/*
* __kill_pgrp_info() sends a signal to a process group: this is what the tty
* control characters do (^C, ^Z etc)
+ * - the caller must hold at least a readlock on tasklist_lock
*/
-
int __kill_pgrp_info(int sig, struct siginfo *info, struct pid *pgrp)
{
struct task_struct *p = NULL;
@@ -1089,6 +1101,7 @@
{
int ret = -EINVAL;
struct task_struct *p;
+ const struct cred *pcred;
if (!valid_signal(sig))
return ret;
@@ -1099,9 +1112,11 @@
ret = -ESRCH;
goto out_unlock;
}
- if ((info == SEND_SIG_NOINFO || (!is_si_special(info) && SI_FROMUSER(info)))
- && (euid != p->cred->suid) && (euid != p->cred->uid)
- && (uid != p->cred->suid) && (uid != p->cred->uid)) {
+ pcred = __task_cred(p);
+ if ((info == SEND_SIG_NOINFO ||
+ (!is_si_special(info) && SI_FROMUSER(info))) &&
+ euid != pcred->suid && euid != pcred->uid &&
+ uid != pcred->suid && uid != pcred->uid) {
ret = -EPERM;
goto out_unlock;
}
@@ -1372,10 +1387,9 @@
*/
rcu_read_lock();
info.si_pid = task_pid_nr_ns(tsk, tsk->parent->nsproxy->pid_ns);
+ info.si_uid = __task_cred(tsk)->uid;
rcu_read_unlock();
- info.si_uid = tsk->cred->uid;
-
thread_group_cputime(tsk, &cputime);
info.si_utime = cputime_to_jiffies(cputime.utime);
info.si_stime = cputime_to_jiffies(cputime.stime);
@@ -1443,10 +1457,9 @@
*/
rcu_read_lock();
info.si_pid = task_pid_nr_ns(tsk, tsk->parent->nsproxy->pid_ns);
+ info.si_uid = __task_cred(tsk)->uid;
rcu_read_unlock();
- info.si_uid = tsk->cred->uid;
-
info.si_utime = cputime_to_clock_t(tsk->utime);
info.si_stime = cputime_to_clock_t(tsk->stime);
@@ -1713,7 +1726,7 @@
info->si_errno = 0;
info->si_code = SI_USER;
info->si_pid = task_pid_vnr(current->parent);
- info->si_uid = current->parent->cred->uid;
+ info->si_uid = task_uid(current->parent);
}
/* If the (new) signal is now blocked, requeue it. */
diff --git a/kernel/sys.c b/kernel/sys.c
index c4d6b59..ccc9eb7 100644
--- a/kernel/sys.c
+++ b/kernel/sys.c
@@ -112,14 +112,17 @@
void (*pm_power_off_prepare)(void);
+/*
+ * set the priority of a task
+ * - the caller must hold the RCU read lock
+ */
static int set_one_prio(struct task_struct *p, int niceval, int error)
{
- uid_t euid = current_euid();
+ const struct cred *cred = current_cred(), *pcred = __task_cred(p);
int no_nice;
- if (p->cred->uid != euid &&
- p->cred->euid != euid &&
- !capable(CAP_SYS_NICE)) {
+ if (pcred->uid != cred->euid &&
+ pcred->euid != cred->euid && !capable(CAP_SYS_NICE)) {
error = -EPERM;
goto out;
}
diff --git a/kernel/tsacct.c b/kernel/tsacct.c
index 6d1ed07..2dc06ab 100644
--- a/kernel/tsacct.c
+++ b/kernel/tsacct.c
@@ -27,6 +27,7 @@
*/
void bacct_add_tsk(struct taskstats *stats, struct task_struct *tsk)
{
+ const struct cred *tcred;
struct timespec uptime, ts;
u64 ac_etime;
@@ -53,10 +54,11 @@
stats->ac_flag |= AXSIG;
stats->ac_nice = task_nice(tsk);
stats->ac_sched = tsk->policy;
- stats->ac_uid = tsk->cred->uid;
- stats->ac_gid = tsk->cred->gid;
stats->ac_pid = tsk->pid;
rcu_read_lock();
+ tcred = __task_cred(tsk);
+ stats->ac_uid = tcred->uid;
+ stats->ac_gid = tcred->gid;
stats->ac_ppid = pid_alive(tsk) ?
rcu_dereference(tsk->real_parent)->tgid : 0;
rcu_read_unlock();
