From 7cb9a94c158b956f46cf093ed966d0c1e996dddb Mon Sep 17 00:00:00 2001
From: Thomas Gleixner <tglx@linutronix.de>
Date: Mon, 19 Aug 2019 16:31:45 +0200
Subject: posix-cpu-timers: Fixup stale comment

The comment above cleanup_timers() is outdated. The timers are only removed
from the task/process list heads but not modified in any other way.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://lkml.kernel.org/r/20190819143801.747233612@linutronix.de
---
 kernel/time/posix-cpu-timers.c | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

(limited to 'kernel/time/posix-cpu-timers.c')

diff --git a/kernel/time/posix-cpu-timers.c b/kernel/time/posix-cpu-timers.c
index 0a426f4e3125..742d4a4e6f71 100644
--- a/kernel/time/posix-cpu-timers.c
+++ b/kernel/time/posix-cpu-timers.c
@@ -412,9 +412,10 @@ static void cleanup_timers_list(struct list_head *head)
 }
 
 /*
- * Clean out CPU timers still ticking when a thread exited.  The task
- * pointer is cleared, and the expiry time is replaced with the residual
- * time for later timer_gettime calls to return.
+ * Clean out CPU timers which are still armed when a thread exits. The
+ * timers are only removed from the list. No other updates are done. The
+ * corresponding posix timers are still accessible, but cannot be rearmed.
+ *
  * This must be called with the siglock held.
  */
 static void cleanup_timers(struct list_head *head)
-- 
cgit v1.2.3


From 692117c1f7a6770ed41dd8f277cd9fed1dfb16f1 Mon Sep 17 00:00:00 2001
From: Thomas Gleixner <tglx@linutronix.de>
Date: Mon, 19 Aug 2019 16:31:46 +0200
Subject: posix-cpu-timers: Sanitize bogus WARNONS

Warning when p == NULL and then proceeding and dereferencing p does not
make any sense as the kernel will crash with a NULL pointer dereference
right away.

Bailing out when p == NULL and returning an error code does not cure the
underlying problem which caused p to be NULL. Though it might allow to
do proper debugging.

Same applies to the clock id check in set_process_cpu_timer().

Clean them up and make them return without trying to do further damage.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://lkml.kernel.org/r/20190819143801.846497772@linutronix.de
---
 kernel/time/posix-cpu-timers.c | 20 +++++++++++++-------
 1 file changed, 13 insertions(+), 7 deletions(-)

(limited to 'kernel/time/posix-cpu-timers.c')

diff --git a/kernel/time/posix-cpu-timers.c b/kernel/time/posix-cpu-timers.c
index 742d4a4e6f71..98223d2805c2 100644
--- a/kernel/time/posix-cpu-timers.c
+++ b/kernel/time/posix-cpu-timers.c
@@ -375,7 +375,8 @@ static int posix_cpu_timer_del(struct k_itimer *timer)
 	struct sighand_struct *sighand;
 	struct task_struct *p = timer->it.cpu.task;
 
-	WARN_ON_ONCE(p == NULL);
+	if (WARN_ON_ONCE(!p))
+		return -EINVAL;
 
 	/*
 	 * Protect against sighand release/switch in exit/exec and process/
@@ -581,7 +582,8 @@ static int posix_cpu_timer_set(struct k_itimer *timer, int timer_flags,
 	u64 old_expires, new_expires, old_incr, val;
 	int ret;
 
-	WARN_ON_ONCE(p == NULL);
+	if (WARN_ON_ONCE(!p))
+		return -EINVAL;
 
 	/*
 	 * Use the to_ktime conversion because that clamps the maximum
@@ -716,10 +718,11 @@ static int posix_cpu_timer_set(struct k_itimer *timer, int timer_flags,
 
 static void posix_cpu_timer_get(struct k_itimer *timer, struct itimerspec64 *itp)
 {
-	u64 now;
 	struct task_struct *p = timer->it.cpu.task;
+	u64 now;
 
-	WARN_ON_ONCE(p == NULL);
+	if (WARN_ON_ONCE(!p))
+		return;
 
 	/*
 	 * Easy part: convert the reload time.
@@ -1001,12 +1004,13 @@ static void check_process_timers(struct task_struct *tsk,
  */
 static void posix_cpu_timer_rearm(struct k_itimer *timer)
 {
+	struct task_struct *p = timer->it.cpu.task;
 	struct sighand_struct *sighand;
 	unsigned long flags;
-	struct task_struct *p = timer->it.cpu.task;
 	u64 now;
 
-	WARN_ON_ONCE(p == NULL);
+	if (WARN_ON_ONCE(!p))
+		return;
 
 	/*
 	 * Fetch the current sample and update the timer's expiry time.
@@ -1203,7 +1207,9 @@ void set_process_cpu_timer(struct task_struct *tsk, unsigned int clock_idx,
 	u64 now;
 	int ret;
 
-	WARN_ON_ONCE(clock_idx == CPUCLOCK_SCHED);
+	if (WARN_ON_ONCE(clock_idx >= CPUCLOCK_SCHED))
+		return;
+
 	ret = cpu_timer_sample_group(clock_idx, tsk, &now);
 
 	if (oldval && ret != -EINVAL) {
-- 
cgit v1.2.3


From dce3e8fd039cc1b62760b3ad6822cf04c262cd0e Mon Sep 17 00:00:00 2001
From: Thomas Gleixner <tglx@linutronix.de>
Date: Mon, 19 Aug 2019 16:31:47 +0200
Subject: posix-cpu-timers: Remove tsk argument from run_posix_cpu_timers()

It's always current. Don't give people wrong ideas.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://lkml.kernel.org/r/20190819143801.945469967@linutronix.de
---
 kernel/time/posix-cpu-timers.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

(limited to 'kernel/time/posix-cpu-timers.c')

diff --git a/kernel/time/posix-cpu-timers.c b/kernel/time/posix-cpu-timers.c
index 98223d2805c2..387e0e86e1b8 100644
--- a/kernel/time/posix-cpu-timers.c
+++ b/kernel/time/posix-cpu-timers.c
@@ -1137,11 +1137,12 @@ static inline int fastpath_timer_check(struct task_struct *tsk)
  * already updated our counts.  We need to check if any timers fire now.
  * Interrupts are disabled.
  */
-void run_posix_cpu_timers(struct task_struct *tsk)
+void run_posix_cpu_timers(void)
 {
-	LIST_HEAD(firing);
+	struct task_struct *tsk = current;
 	struct k_itimer *timer, *next;
 	unsigned long flags;
+	LIST_HEAD(firing);
 
 	lockdep_assert_irqs_disabled();
 
-- 
cgit v1.2.3


From 6ae40e3fdcd33a6ff3c490b9302d6a1861093f65 Mon Sep 17 00:00:00 2001
From: Thomas Gleixner <tglx@linutronix.de>
Date: Wed, 21 Aug 2019 21:08:48 +0200
Subject: posix-cpu-timers: Provide task validation functions

The code contains three slightly different copies of validating whether a
given clock resolves to a valid task and whether the current caller has
permissions to access it.

Create central functions. Replace check_clock() as a first step and rename
it to something sensible.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lkml.kernel.org/r/20190821192919.326097175@linutronix.de
---
 kernel/time/posix-cpu-timers.c | 65 ++++++++++++++++++++++++++++--------------
 1 file changed, 44 insertions(+), 21 deletions(-)

(limited to 'kernel/time/posix-cpu-timers.c')

diff --git a/kernel/time/posix-cpu-timers.c b/kernel/time/posix-cpu-timers.c
index 387e0e86e1b8..b06ed8b14861 100644
--- a/kernel/time/posix-cpu-timers.c
+++ b/kernel/time/posix-cpu-timers.c
@@ -35,27 +35,52 @@ void update_rlimit_cpu(struct task_struct *task, unsigned long rlim_new)
 	spin_unlock_irq(&task->sighand->siglock);
 }
 
-static int check_clock(const clockid_t which_clock)
+/*
+ * Functions for validating access to tasks.
+ */
+static struct task_struct *lookup_task(const pid_t pid, bool thread)
 {
-	int error = 0;
 	struct task_struct *p;
-	const pid_t pid = CPUCLOCK_PID(which_clock);
 
-	if (CPUCLOCK_WHICH(which_clock) >= CPUCLOCK_MAX)
-		return -EINVAL;
+	if (!pid)
+		return thread ? current : current->group_leader;
+
+	p = find_task_by_vpid(pid);
+	if (!p || p == current)
+		return p;
+	if (thread)
+		return same_thread_group(p, current) ? p : NULL;
+	if (p == current)
+		return p;
+	return has_group_leader_pid(p) ? p : NULL;
+}
+
+static struct task_struct *__get_task_for_clock(const clockid_t clock,
+						bool getref)
+{
+	const bool thread = !!CPUCLOCK_PERTHREAD(clock);
+	const pid_t pid = CPUCLOCK_PID(clock);
+	struct task_struct *p;
 
-	if (pid == 0)
-		return 0;
+	if (CPUCLOCK_WHICH(clock) >= CPUCLOCK_MAX)
+		return NULL;
 
 	rcu_read_lock();
-	p = find_task_by_vpid(pid);
-	if (!p || !(CPUCLOCK_PERTHREAD(which_clock) ?
-		   same_thread_group(p, current) : has_group_leader_pid(p))) {
-		error = -EINVAL;
-	}
+	p = lookup_task(pid, thread);
+	if (p && getref)
+		get_task_struct(p);
 	rcu_read_unlock();
+	return p;
+}
 
-	return error;
+static inline struct task_struct *get_task_for_clock(const clockid_t clock)
+{
+	return __get_task_for_clock(clock, true);
+}
+
+static inline int validate_clock_permissions(const clockid_t clock)
+{
+	return __get_task_for_clock(clock, false) ? 0 : -EINVAL;
 }
 
 /*
@@ -125,7 +150,8 @@ static inline u64 virt_ticks(struct task_struct *p)
 static int
 posix_cpu_clock_getres(const clockid_t which_clock, struct timespec64 *tp)
 {
-	int error = check_clock(which_clock);
+	int error = validate_clock_permissions(which_clock);
+
 	if (!error) {
 		tp->tv_sec = 0;
 		tp->tv_nsec = ((NSEC_PER_SEC + HZ - 1) / HZ);
@@ -142,20 +168,17 @@ posix_cpu_clock_getres(const clockid_t which_clock, struct timespec64 *tp)
 }
 
 static int
-posix_cpu_clock_set(const clockid_t which_clock, const struct timespec64 *tp)
+posix_cpu_clock_set(const clockid_t clock, const struct timespec64 *tp)
 {
+	int error = validate_clock_permissions(clock);
+
 	/*
 	 * You can never reset a CPU clock, but we check for other errors
 	 * in the call before failing with EPERM.
 	 */
-	int error = check_clock(which_clock);
-	if (error == 0) {
-		error = -EPERM;
-	}
-	return error;
+	return error ? : -EPERM;
 }
 
-
 /*
  * Sample a per-thread clock for the given task.
  */
-- 
cgit v1.2.3


From bfcf3e92c6c07cd1084624bad5622f3dad96328c Mon Sep 17 00:00:00 2001
From: Thomas Gleixner <tglx@linutronix.de>
Date: Wed, 21 Aug 2019 21:08:49 +0200
Subject: posix-cpu-timers: Use common permission check in
 posix_cpu_clock_get()

Replace the next slightly different copy of permission checks. That also
removes the necessarity to check the return value of the sample functions
because the clock id is already validated.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://lkml.kernel.org/r/20190821192919.414813172@linutronix.de
---
 kernel/time/posix-cpu-timers.c | 57 +++++++++++-------------------------------
 1 file changed, 14 insertions(+), 43 deletions(-)

(limited to 'kernel/time/posix-cpu-timers.c')

diff --git a/kernel/time/posix-cpu-timers.c b/kernel/time/posix-cpu-timers.c
index b06ed8b14861..eb11117bf227 100644
--- a/kernel/time/posix-cpu-timers.c
+++ b/kernel/time/posix-cpu-timers.c
@@ -289,53 +289,24 @@ static int cpu_clock_sample_group(const clockid_t which_clock,
 	return 0;
 }
 
-static int posix_cpu_clock_get_task(struct task_struct *tsk,
-				    const clockid_t which_clock,
-				    struct timespec64 *tp)
+static int posix_cpu_clock_get(const clockid_t clock, struct timespec64 *tp)
 {
-	int err = -EINVAL;
-	u64 rtn;
-
-	if (CPUCLOCK_PERTHREAD(which_clock)) {
-		if (same_thread_group(tsk, current))
-			err = cpu_clock_sample(which_clock, tsk, &rtn);
-	} else {
-		if (tsk == current || thread_group_leader(tsk))
-			err = cpu_clock_sample_group(which_clock, tsk, &rtn);
-	}
-
-	if (!err)
-		*tp = ns_to_timespec64(rtn);
+	const clockid_t clkid = CPUCLOCK_WHICH(clock);
+	struct task_struct *tsk;
+	u64 t;
 
-	return err;
-}
-
-
-static int posix_cpu_clock_get(const clockid_t which_clock, struct timespec64 *tp)
-{
-	const pid_t pid = CPUCLOCK_PID(which_clock);
-	int err = -EINVAL;
+	tsk = get_task_for_clock(clock);
+	if (!tsk)
+		return -EINVAL;
 
-	if (pid == 0) {
-		/*
-		 * Special case constant value for our own clocks.
-		 * We don't have to do any lookup to find ourselves.
-		 */
-		err = posix_cpu_clock_get_task(current, which_clock, tp);
-	} else {
-		/*
-		 * Find the given PID, and validate that the caller
-		 * should be able to see it.
-		 */
-		struct task_struct *p;
-		rcu_read_lock();
-		p = find_task_by_vpid(pid);
-		if (p)
-			err = posix_cpu_clock_get_task(p, which_clock, tp);
-		rcu_read_unlock();
-	}
+	if (CPUCLOCK_PERTHREAD(clock))
+		cpu_clock_sample(clkid, tsk, &t);
+	else
+		cpu_clock_sample_group(clkid, tsk, &t);
+	put_task_struct(tsk);
 
-	return err;
+	*tp = ns_to_timespec64(t);
+	return 0;
 }
 
 /*
-- 
cgit v1.2.3


From e5a8b65b4cb2fe024b83bdec0424269949cc0a27 Mon Sep 17 00:00:00 2001
From: Thomas Gleixner <tglx@linutronix.de>
Date: Wed, 21 Aug 2019 21:08:50 +0200
Subject: posix-cpu-timers: Use common permission check in
 posix_cpu_timer_create()

Yet another copy of the same thing gone...

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://lkml.kernel.org/r/20190821192919.505833418@linutronix.de
---
 kernel/time/posix-cpu-timers.c | 35 +++--------------------------------
 1 file changed, 3 insertions(+), 32 deletions(-)

(limited to 'kernel/time/posix-cpu-timers.c')

diff --git a/kernel/time/posix-cpu-timers.c b/kernel/time/posix-cpu-timers.c
index eb11117bf227..4426a0f9c470 100644
--- a/kernel/time/posix-cpu-timers.c
+++ b/kernel/time/posix-cpu-timers.c
@@ -316,44 +316,15 @@ static int posix_cpu_clock_get(const clockid_t clock, struct timespec64 *tp)
  */
 static int posix_cpu_timer_create(struct k_itimer *new_timer)
 {
-	int ret = 0;
-	const pid_t pid = CPUCLOCK_PID(new_timer->it_clock);
-	struct task_struct *p;
+	struct task_struct *p = get_task_for_clock(new_timer->it_clock);
 
-	if (CPUCLOCK_WHICH(new_timer->it_clock) >= CPUCLOCK_MAX)
+	if (!p)
 		return -EINVAL;
 
 	new_timer->kclock = &clock_posix_cpu;
-
 	INIT_LIST_HEAD(&new_timer->it.cpu.entry);
-
-	rcu_read_lock();
-	if (CPUCLOCK_PERTHREAD(new_timer->it_clock)) {
-		if (pid == 0) {
-			p = current;
-		} else {
-			p = find_task_by_vpid(pid);
-			if (p && !same_thread_group(p, current))
-				p = NULL;
-		}
-	} else {
-		if (pid == 0) {
-			p = current->group_leader;
-		} else {
-			p = find_task_by_vpid(pid);
-			if (p && !has_group_leader_pid(p))
-				p = NULL;
-		}
-	}
 	new_timer->it.cpu.task = p;
-	if (p) {
-		get_task_struct(p);
-	} else {
-		ret = -EINVAL;
-	}
-	rcu_read_unlock();
-
-	return ret;
+	return 0;
 }
 
 /*
-- 
cgit v1.2.3


From 19298fbf453c90a6cf72288155f80c6f55e9139d Mon Sep 17 00:00:00 2001
From: Thomas Gleixner <tglx@linutronix.de>
Date: Wed, 21 Aug 2019 21:08:51 +0200
Subject: posix-cpu-timers: Provide quick sample function for itimer

get_itimer() needs a sample of the current thread group cputime. It invokes
thread_group_cputimer() - which is a misnomer. That function also starts
eventually the group cputime accouting which is bogus because the
accounting is already active when a timer is armed.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://lkml.kernel.org/r/20190821192919.599658199@linutronix.de
---
 kernel/time/posix-cpu-timers.c | 21 +++++++++++++++++++++
 1 file changed, 21 insertions(+)

(limited to 'kernel/time/posix-cpu-timers.c')

diff --git a/kernel/time/posix-cpu-timers.c b/kernel/time/posix-cpu-timers.c
index 4426a0f9c470..c22b6b604a95 100644
--- a/kernel/time/posix-cpu-timers.c
+++ b/kernel/time/posix-cpu-timers.c
@@ -232,6 +232,27 @@ static inline void sample_cputime_atomic(struct task_cputime *times,
 	times->sum_exec_runtime = atomic64_read(&atomic_times->sum_exec_runtime);
 }
 
+/**
+ * thread_group_sample_cputime - Sample cputime for a given task
+ * @tsk:	Task for which cputime needs to be started
+ * @iimes:	Storage for time samples
+ *
+ * Called from sys_getitimer() to calculate the expiry time of an active
+ * timer. That means group cputime accounting is already active. Called
+ * with task sighand lock held.
+ *
+ * Updates @times with an uptodate sample of the thread group cputimes.
+ */
+void thread_group_sample_cputime(struct task_struct *tsk,
+				struct task_cputime *times)
+{
+	struct thread_group_cputimer *cputimer = &tsk->signal->cputimer;
+
+	WARN_ON_ONCE(!cputimer->running);
+
+	sample_cputime_atomic(times, &cputimer->cputime_atomic);
+}
+
 void thread_group_cputimer(struct task_struct *tsk, struct task_cputime *times)
 {
 	struct thread_group_cputimer *cputimer = &tsk->signal->cputimer;
-- 
cgit v1.2.3


From a324956fae05d863386c682830e917f6685f1d4f Mon Sep 17 00:00:00 2001
From: Thomas Gleixner <tglx@linutronix.de>
Date: Wed, 21 Aug 2019 21:08:53 +0200
Subject: posix-cpu-timers: Sample directly in timer check

The thread group accounting is active, otherwise the expiry function would
not be running. Sample the thread group time directly.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://lkml.kernel.org/r/20190821192919.780348088@linutronix.de
---
 kernel/time/posix-cpu-timers.c | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

(limited to 'kernel/time/posix-cpu-timers.c')

diff --git a/kernel/time/posix-cpu-timers.c b/kernel/time/posix-cpu-timers.c
index c22b6b604a95..cb736787145b 100644
--- a/kernel/time/posix-cpu-timers.c
+++ b/kernel/time/posix-cpu-timers.c
@@ -914,16 +914,17 @@ static void check_process_timers(struct task_struct *tsk,
 	if (!READ_ONCE(tsk->signal->cputimer.running))
 		return;
 
-        /*
+       /*
 	 * Signify that a thread is checking for process timers.
 	 * Write access to this field is protected by the sighand lock.
 	 */
 	sig->cputimer.checking_timer = true;
 
 	/*
-	 * Collect the current process totals.
+	 * Collect the current process totals. Group accounting is active
+	 * so the sample can be taken directly.
 	 */
-	thread_group_cputimer(tsk, &cputime);
+	sample_cputime_atomic(&cputime, &sig->cputimer.cputime_atomic);
 	utime = cputime.utime;
 	ptime = utime + cputime.stime;
 	sum_sched_runtime = cputime.sum_exec_runtime;
-- 
cgit v1.2.3


From c506bef424ca282f2ad357e86fee940c69018974 Mon Sep 17 00:00:00 2001
From: Thomas Gleixner <tglx@linutronix.de>
Date: Wed, 21 Aug 2019 21:08:54 +0200
Subject: posix-cpu-timers: Rename thread_group_cputimer() and make it static

thread_group_cputimer() is a complete misnomer. The function does two things:

 - For arming process wide timers it makes sure that the atomic time
   storage is up to date. If no cpu timer is armed yet, then the atomic
   time storage is not updated by the scheduler for performance reasons.

   In that case a full summing up of all threads needs to be done and the
   update needs to be enabled.

- Samples the current time into the caller supplied storage.

Rename it to thread_group_start_cputime(), make it static and fixup the
callsite.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://lkml.kernel.org/r/20190821192919.869350319@linutronix.de
---
 kernel/time/posix-cpu-timers.c | 17 +++++++++++++++--
 1 file changed, 15 insertions(+), 2 deletions(-)

(limited to 'kernel/time/posix-cpu-timers.c')

diff --git a/kernel/time/posix-cpu-timers.c b/kernel/time/posix-cpu-timers.c
index cb736787145b..def225ae069a 100644
--- a/kernel/time/posix-cpu-timers.c
+++ b/kernel/time/posix-cpu-timers.c
@@ -253,7 +253,20 @@ void thread_group_sample_cputime(struct task_struct *tsk,
 	sample_cputime_atomic(times, &cputimer->cputime_atomic);
 }
 
-void thread_group_cputimer(struct task_struct *tsk, struct task_cputime *times)
+/**
+ * thread_group_start_cputime - Start cputime and return a sample
+ * @tsk:	Task for which cputime needs to be started
+ * @iimes:	Storage for time samples
+ *
+ * The thread group cputime accouting is avoided when there are no posix
+ * CPU timers armed. Before starting a timer it's required to check whether
+ * the time accounting is active. If not, a full update of the atomic
+ * accounting store needs to be done and the accounting enabled.
+ *
+ * Updates @times with an uptodate sample of the thread group cputimes.
+ */
+static void
+thread_group_start_cputime(struct task_struct *tsk, struct task_cputime *times)
 {
 	struct thread_group_cputimer *cputimer = &tsk->signal->cputimer;
 	struct task_cputime sum;
@@ -536,7 +549,7 @@ static int cpu_timer_sample_group(const clockid_t which_clock,
 {
 	struct task_cputime cputime;
 
-	thread_group_cputimer(p, &cputime);
+	thread_group_start_cputime(p, &cputime);
 	switch (CPUCLOCK_WHICH(which_clock)) {
 	default:
 		return -EINVAL;
-- 
cgit v1.2.3


From 24ab7f5a7b2c917e89fc6a87252f18faff91d6ce Mon Sep 17 00:00:00 2001
From: Thomas Gleixner <tglx@linutronix.de>
Date: Wed, 21 Aug 2019 21:08:55 +0200
Subject: posix-cpu-timers: Consolidate thread group sample code

cpu_clock_sample_group() and cpu_timer_sample_group() are almost the
same. Before the rename one called thread_group_cputimer() and the other
thread_group_cputime(). Really intuitive function names.

Consolidate the functions and also avoid the thread traversal when
the thread group's accounting is already active.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://lkml.kernel.org/r/20190821192919.960966884@linutronix.de
---
 kernel/time/posix-cpu-timers.c | 59 ++++++++++++++----------------------------
 1 file changed, 20 insertions(+), 39 deletions(-)

(limited to 'kernel/time/posix-cpu-timers.c')

diff --git a/kernel/time/posix-cpu-timers.c b/kernel/time/posix-cpu-timers.c
index def225ae069a..a9003b2d614d 100644
--- a/kernel/time/posix-cpu-timers.c
+++ b/kernel/time/posix-cpu-timers.c
@@ -294,29 +294,37 @@ thread_group_start_cputime(struct task_struct *tsk, struct task_cputime *times)
 }
 
 /*
- * Sample a process (thread group) clock for the given group_leader task.
- * Must be called with task sighand lock held for safe while_each_thread()
- * traversal.
+ * Sample a process (thread group) clock for the given task clkid. If the
+ * group's cputime accounting is already enabled, read the atomic
+ * store. Otherwise a full update is required.  Task's sighand lock must be
+ * held to protect the task traversal on a full update.
  */
 static int cpu_clock_sample_group(const clockid_t which_clock,
 				  struct task_struct *p,
-				  u64 *sample)
+				  u64 *sample, bool start)
 {
+	struct thread_group_cputimer *cputimer = &p->signal->cputimer;
 	struct task_cputime cputime;
 
+	if (!READ_ONCE(cputimer->running)) {
+		if (start)
+			thread_group_start_cputime(p, &cputime);
+		else
+			thread_group_cputime(p, &cputime);
+	} else {
+		sample_cputime_atomic(&cputime, &cputimer->cputime_atomic);
+	}
+
 	switch (CPUCLOCK_WHICH(which_clock)) {
 	default:
 		return -EINVAL;
 	case CPUCLOCK_PROF:
-		thread_group_cputime(p, &cputime);
 		*sample = cputime.utime + cputime.stime;
 		break;
 	case CPUCLOCK_VIRT:
-		thread_group_cputime(p, &cputime);
 		*sample = cputime.utime;
 		break;
 	case CPUCLOCK_SCHED:
-		thread_group_cputime(p, &cputime);
 		*sample = cputime.sum_exec_runtime;
 		break;
 	}
@@ -336,7 +344,7 @@ static int posix_cpu_clock_get(const clockid_t clock, struct timespec64 *tp)
 	if (CPUCLOCK_PERTHREAD(clock))
 		cpu_clock_sample(clkid, tsk, &t);
 	else
-		cpu_clock_sample_group(clkid, tsk, &t);
+		cpu_clock_sample_group(clkid, tsk, &t, false);
 	put_task_struct(tsk);
 
 	*tp = ns_to_timespec64(t);
@@ -539,33 +547,6 @@ static void cpu_timer_fire(struct k_itimer *timer)
 	}
 }
 
-/*
- * Sample a process (thread group) timer for the given group_leader task.
- * Must be called with task sighand lock held for safe while_each_thread()
- * traversal.
- */
-static int cpu_timer_sample_group(const clockid_t which_clock,
-				  struct task_struct *p, u64 *sample)
-{
-	struct task_cputime cputime;
-
-	thread_group_start_cputime(p, &cputime);
-	switch (CPUCLOCK_WHICH(which_clock)) {
-	default:
-		return -EINVAL;
-	case CPUCLOCK_PROF:
-		*sample = cputime.utime + cputime.stime;
-		break;
-	case CPUCLOCK_VIRT:
-		*sample = cputime.utime;
-		break;
-	case CPUCLOCK_SCHED:
-		*sample = cputime.sum_exec_runtime;
-		break;
-	}
-	return 0;
-}
-
 /*
  * Guts of sys_timer_settime for CPU timers.
  * This is called with the timer locked and interrupts disabled.
@@ -627,7 +608,7 @@ static int posix_cpu_timer_set(struct k_itimer *timer, int timer_flags,
 	if (CPUCLOCK_PERTHREAD(timer->it_clock)) {
 		cpu_clock_sample(timer->it_clock, p, &val);
 	} else {
-		cpu_timer_sample_group(timer->it_clock, p, &val);
+		cpu_clock_sample_group(timer->it_clock, p, &val, true);
 	}
 
 	if (old) {
@@ -755,7 +736,7 @@ static void posix_cpu_timer_get(struct k_itimer *timer, struct itimerspec64 *itp
 			timer->it.cpu.expires = 0;
 			return;
 		} else {
-			cpu_timer_sample_group(timer->it_clock, p, &now);
+			cpu_clock_sample_group(timer->it_clock, p, &now, false);
 			unlock_task_sighand(p, &flags);
 		}
 	}
@@ -1042,7 +1023,7 @@ static void posix_cpu_timer_rearm(struct k_itimer *timer)
 			/* If the process is dying, no need to rearm */
 			goto unlock;
 		}
-		cpu_timer_sample_group(timer->it_clock, p, &now);
+		cpu_clock_sample_group(timer->it_clock, p, &now, true);
 		bump_cpu_timer(timer, now);
 		/* Leave the sighand locked for the call below.  */
 	}
@@ -1211,7 +1192,7 @@ void set_process_cpu_timer(struct task_struct *tsk, unsigned int clock_idx,
 	if (WARN_ON_ONCE(clock_idx >= CPUCLOCK_SCHED))
 		return;
 
-	ret = cpu_timer_sample_group(clock_idx, tsk, &now);
+	ret = cpu_clock_sample_group(clock_idx, tsk, &now, true);
 
 	if (oldval && ret != -EINVAL) {
 		/*
-- 
cgit v1.2.3


From c7a37c6f4c651a531101c5721814333bae2804ec Mon Sep 17 00:00:00 2001
From: Thomas Gleixner <tglx@linutronix.de>
Date: Wed, 21 Aug 2019 21:08:56 +0200
Subject: posix-cpu-timers: Use clock ID in posix_cpu_timer_set()

Extract the clock ID (PROF/VIRT/SCHED) from the clock selector and use it
as argument to the sample functions. That allows to simplify them once all
callers are fixed.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://lkml.kernel.org/r/20190821192920.050770464@linutronix.de
---
 kernel/time/posix-cpu-timers.c | 11 ++++++-----
 1 file changed, 6 insertions(+), 5 deletions(-)

(limited to 'kernel/time/posix-cpu-timers.c')

diff --git a/kernel/time/posix-cpu-timers.c b/kernel/time/posix-cpu-timers.c
index a9003b2d614d..12561f8ef378 100644
--- a/kernel/time/posix-cpu-timers.c
+++ b/kernel/time/posix-cpu-timers.c
@@ -556,10 +556,11 @@ static void cpu_timer_fire(struct k_itimer *timer)
 static int posix_cpu_timer_set(struct k_itimer *timer, int timer_flags,
 			       struct itimerspec64 *new, struct itimerspec64 *old)
 {
-	unsigned long flags;
-	struct sighand_struct *sighand;
-	struct task_struct *p = timer->it.cpu.task;
+	clockid_t clkid = CPUCLOCK_WHICH(timer->it_clock);
 	u64 old_expires, new_expires, old_incr, val;
+	struct task_struct *p = timer->it.cpu.task;
+	struct sighand_struct *sighand;
+	unsigned long flags;
 	int ret;
 
 	if (WARN_ON_ONCE(!p))
@@ -606,9 +607,9 @@ static int posix_cpu_timer_set(struct k_itimer *timer, int timer_flags,
 	 * check if it's already passed.  In short, we need a sample.
 	 */
 	if (CPUCLOCK_PERTHREAD(timer->it_clock)) {
-		cpu_clock_sample(timer->it_clock, p, &val);
+		cpu_clock_sample(clkid, p, &val);
 	} else {
-		cpu_clock_sample_group(timer->it_clock, p, &val, true);
+		cpu_clock_sample_group(clkid, p, &val, true);
 	}
 
 	if (old) {
-- 
cgit v1.2.3


From 99093c5b81f58815fbfc1fe301c87206c6f5f730 Mon Sep 17 00:00:00 2001
From: Thomas Gleixner <tglx@linutronix.de>
Date: Wed, 21 Aug 2019 21:08:57 +0200
Subject: posix-cpu-timers: Use clock ID in posix_cpu_timer_get()

Extract the clock ID (PROF/VIRT/SCHED) from the clock selector and use it
as argument to the sample functions. That allows to simplify them once all
callers are fixed.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://lkml.kernel.org/r/20190821192920.155487201@linutronix.de
---
 kernel/time/posix-cpu-timers.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

(limited to 'kernel/time/posix-cpu-timers.c')

diff --git a/kernel/time/posix-cpu-timers.c b/kernel/time/posix-cpu-timers.c
index 12561f8ef378..8e4ce8ade88c 100644
--- a/kernel/time/posix-cpu-timers.c
+++ b/kernel/time/posix-cpu-timers.c
@@ -699,6 +699,7 @@ static int posix_cpu_timer_set(struct k_itimer *timer, int timer_flags,
 
 static void posix_cpu_timer_get(struct k_itimer *timer, struct itimerspec64 *itp)
 {
+	clockid_t clkid = CPUCLOCK_WHICH(timer->it_clock);
 	struct task_struct *p = timer->it.cpu.task;
 	u64 now;
 
@@ -717,7 +718,7 @@ static void posix_cpu_timer_get(struct k_itimer *timer, struct itimerspec64 *itp
 	 * Sample the clock to take the difference with the expiry time.
 	 */
 	if (CPUCLOCK_PERTHREAD(timer->it_clock)) {
-		cpu_clock_sample(timer->it_clock, p, &now);
+		cpu_clock_sample(clkid, p, &now);
 	} else {
 		struct sighand_struct *sighand;
 		unsigned long flags;
@@ -737,7 +738,7 @@ static void posix_cpu_timer_get(struct k_itimer *timer, struct itimerspec64 *itp
 			timer->it.cpu.expires = 0;
 			return;
 		} else {
-			cpu_clock_sample_group(timer->it_clock, p, &now, false);
+			cpu_clock_sample_group(clkid, p, &now, false);
 			unlock_task_sighand(p, &flags);
 		}
 	}
-- 
cgit v1.2.3


From da020ce406b2a9b714b82de9123a4c5a4848647b Mon Sep 17 00:00:00 2001
From: Thomas Gleixner <tglx@linutronix.de>
Date: Wed, 21 Aug 2019 21:08:58 +0200
Subject: posix-cpu-timers: Use clock ID in posix_cpu_timer_rearm()

Extract the clock ID (PROF/VIRT/SCHED) from the clock selector and use it
as argument to the sample functions. That allows to simplify them once all
callers are fixed.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://lkml.kernel.org/r/20190821192920.245357769@linutronix.de
---
 kernel/time/posix-cpu-timers.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

(limited to 'kernel/time/posix-cpu-timers.c')

diff --git a/kernel/time/posix-cpu-timers.c b/kernel/time/posix-cpu-timers.c
index 8e4ce8ade88c..aff7e3b64d84 100644
--- a/kernel/time/posix-cpu-timers.c
+++ b/kernel/time/posix-cpu-timers.c
@@ -987,6 +987,7 @@ static void check_process_timers(struct task_struct *tsk,
  */
 static void posix_cpu_timer_rearm(struct k_itimer *timer)
 {
+	clockid_t clkid = CPUCLOCK_WHICH(timer->it_clock);
 	struct task_struct *p = timer->it.cpu.task;
 	struct sighand_struct *sighand;
 	unsigned long flags;
@@ -999,7 +1000,7 @@ static void posix_cpu_timer_rearm(struct k_itimer *timer)
 	 * Fetch the current sample and update the timer's expiry time.
 	 */
 	if (CPUCLOCK_PERTHREAD(timer->it_clock)) {
-		cpu_clock_sample(timer->it_clock, p, &now);
+		cpu_clock_sample(clkid, p, &now);
 		bump_cpu_timer(timer, now);
 		if (unlikely(p->exit_state))
 			return;
@@ -1025,7 +1026,7 @@ static void posix_cpu_timer_rearm(struct k_itimer *timer)
 			/* If the process is dying, no need to rearm */
 			goto unlock;
 		}
-		cpu_clock_sample_group(timer->it_clock, p, &now, true);
+		cpu_clock_sample_group(clkid, p, &now, true);
 		bump_cpu_timer(timer, now);
 		/* Leave the sighand locked for the call below.  */
 	}
-- 
cgit v1.2.3


From 5405d0051f7ca820d1781d87baf4d730ff58f208 Mon Sep 17 00:00:00 2001
From: Thomas Gleixner <tglx@linutronix.de>
Date: Wed, 21 Aug 2019 21:08:59 +0200
Subject: posix-cpu-timers: Remove pointless return value check

set_process_cpu_timer() checks already whether the clock id is valid. No
point in checking the return value of the sample function. That allows to
simplify the sample function later.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://lkml.kernel.org/r/20190821192920.339725769@linutronix.de
---
 kernel/time/posix-cpu-timers.c | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

(limited to 'kernel/time/posix-cpu-timers.c')

diff --git a/kernel/time/posix-cpu-timers.c b/kernel/time/posix-cpu-timers.c
index aff7e3b64d84..5944b7494be7 100644
--- a/kernel/time/posix-cpu-timers.c
+++ b/kernel/time/posix-cpu-timers.c
@@ -1190,14 +1190,13 @@ void set_process_cpu_timer(struct task_struct *tsk, unsigned int clock_idx,
 			   u64 *newval, u64 *oldval)
 {
 	u64 now;
-	int ret;
 
 	if (WARN_ON_ONCE(clock_idx >= CPUCLOCK_SCHED))
 		return;
 
-	ret = cpu_clock_sample_group(clock_idx, tsk, &now, true);
+	cpu_clock_sample_group(clock_idx, tsk, &now, true);
 
-	if (oldval && ret != -EINVAL) {
+	if (oldval) {
 		/*
 		 * We are setting itimer. The *oldval is absolute and we update
 		 * it to be relative, *newval argument is relative and we update
-- 
cgit v1.2.3


From 2092c1d4fed9f279d10600b4c1b5167dd8486a2a Mon Sep 17 00:00:00 2001
From: Thomas Gleixner <tglx@linutronix.de>
Date: Wed, 21 Aug 2019 21:09:00 +0200
Subject: posix-cpu-timers: Simplify sample functions

All callers hand in a valdiated clock id. Remove the return value which was
unchecked in most places anyway.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://lkml.kernel.org/r/20190821192920.430475832@linutronix.de
---
 kernel/time/posix-cpu-timers.c | 28 +++++++++++++---------------
 1 file changed, 13 insertions(+), 15 deletions(-)

(limited to 'kernel/time/posix-cpu-timers.c')

diff --git a/kernel/time/posix-cpu-timers.c b/kernel/time/posix-cpu-timers.c
index 5944b7494be7..cc3d148344d3 100644
--- a/kernel/time/posix-cpu-timers.c
+++ b/kernel/time/posix-cpu-timers.c
@@ -180,14 +180,12 @@ posix_cpu_clock_set(const clockid_t clock, const struct timespec64 *tp)
 }
 
 /*
- * Sample a per-thread clock for the given task.
+ * Sample a per-thread clock for the given task. clkid is validated.
  */
-static int cpu_clock_sample(const clockid_t which_clock,
-			    struct task_struct *p, u64 *sample)
+static void cpu_clock_sample(const clockid_t clkid, struct task_struct *p,
+			     u64 *sample)
 {
-	switch (CPUCLOCK_WHICH(which_clock)) {
-	default:
-		return -EINVAL;
+	switch (clkid) {
 	case CPUCLOCK_PROF:
 		*sample = prof_ticks(p);
 		break;
@@ -197,8 +195,9 @@ static int cpu_clock_sample(const clockid_t which_clock,
 	case CPUCLOCK_SCHED:
 		*sample = task_sched_runtime(p);
 		break;
+	default:
+		WARN_ON_ONCE(1);
 	}
-	return 0;
 }
 
 /*
@@ -297,11 +296,11 @@ thread_group_start_cputime(struct task_struct *tsk, struct task_cputime *times)
  * Sample a process (thread group) clock for the given task clkid. If the
  * group's cputime accounting is already enabled, read the atomic
  * store. Otherwise a full update is required.  Task's sighand lock must be
- * held to protect the task traversal on a full update.
+ * held to protect the task traversal on a full update. clkid is already
+ * validated.
  */
-static int cpu_clock_sample_group(const clockid_t which_clock,
-				  struct task_struct *p,
-				  u64 *sample, bool start)
+static void cpu_clock_sample_group(const clockid_t clkid, struct task_struct *p,
+				   u64 *sample, bool start)
 {
 	struct thread_group_cputimer *cputimer = &p->signal->cputimer;
 	struct task_cputime cputime;
@@ -315,9 +314,7 @@ static int cpu_clock_sample_group(const clockid_t which_clock,
 		sample_cputime_atomic(&cputime, &cputimer->cputime_atomic);
 	}
 
-	switch (CPUCLOCK_WHICH(which_clock)) {
-	default:
-		return -EINVAL;
+	switch (clkid) {
 	case CPUCLOCK_PROF:
 		*sample = cputime.utime + cputime.stime;
 		break;
@@ -327,8 +324,9 @@ static int cpu_clock_sample_group(const clockid_t which_clock,
 	case CPUCLOCK_SCHED:
 		*sample = cputime.sum_exec_runtime;
 		break;
+	default:
+		WARN_ON_ONCE(1);
 	}
-	return 0;
 }
 
 static int posix_cpu_clock_get(const clockid_t clock, struct timespec64 *tp)
-- 
cgit v1.2.3


From 8c2d74f03705c2c993a57673bae8fd535eabede6 Mon Sep 17 00:00:00 2001
From: Thomas Gleixner <tglx@linutronix.de>
Date: Wed, 21 Aug 2019 21:09:01 +0200
Subject: posix-cpu-timers: Get rid of pointer indirection

Now that the sample functions have no return value anymore, the result can
simply be returned instead of using pointer indirection.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://lkml.kernel.org/r/20190821192920.535079278@linutronix.de
---
 kernel/time/posix-cpu-timers.c | 50 +++++++++++++++++++-----------------------
 1 file changed, 22 insertions(+), 28 deletions(-)

(limited to 'kernel/time/posix-cpu-timers.c')

diff --git a/kernel/time/posix-cpu-timers.c b/kernel/time/posix-cpu-timers.c
index cc3d148344d3..a2007ce9322a 100644
--- a/kernel/time/posix-cpu-timers.c
+++ b/kernel/time/posix-cpu-timers.c
@@ -182,22 +182,19 @@ posix_cpu_clock_set(const clockid_t clock, const struct timespec64 *tp)
 /*
  * Sample a per-thread clock for the given task. clkid is validated.
  */
-static void cpu_clock_sample(const clockid_t clkid, struct task_struct *p,
-			     u64 *sample)
+static u64 cpu_clock_sample(const clockid_t clkid, struct task_struct *p)
 {
 	switch (clkid) {
 	case CPUCLOCK_PROF:
-		*sample = prof_ticks(p);
-		break;
+		return prof_ticks(p);
 	case CPUCLOCK_VIRT:
-		*sample = virt_ticks(p);
-		break;
+		return virt_ticks(p);
 	case CPUCLOCK_SCHED:
-		*sample = task_sched_runtime(p);
-		break;
+		return task_sched_runtime(p);
 	default:
 		WARN_ON_ONCE(1);
 	}
+	return 0;
 }
 
 /*
@@ -299,8 +296,8 @@ thread_group_start_cputime(struct task_struct *tsk, struct task_cputime *times)
  * held to protect the task traversal on a full update. clkid is already
  * validated.
  */
-static void cpu_clock_sample_group(const clockid_t clkid, struct task_struct *p,
-				   u64 *sample, bool start)
+static u64 cpu_clock_sample_group(const clockid_t clkid, struct task_struct *p,
+				  bool start)
 {
 	struct thread_group_cputimer *cputimer = &p->signal->cputimer;
 	struct task_cputime cputime;
@@ -316,17 +313,15 @@ static void cpu_clock_sample_group(const clockid_t clkid, struct task_struct *p,
 
 	switch (clkid) {
 	case CPUCLOCK_PROF:
-		*sample = cputime.utime + cputime.stime;
-		break;
+		return cputime.utime + cputime.stime;
 	case CPUCLOCK_VIRT:
-		*sample = cputime.utime;
-		break;
+		return cputime.utime;
 	case CPUCLOCK_SCHED:
-		*sample = cputime.sum_exec_runtime;
-		break;
+		return cputime.sum_exec_runtime;
 	default:
 		WARN_ON_ONCE(1);
 	}
+	return 0;
 }
 
 static int posix_cpu_clock_get(const clockid_t clock, struct timespec64 *tp)
@@ -340,9 +335,9 @@ static int posix_cpu_clock_get(const clockid_t clock, struct timespec64 *tp)
 		return -EINVAL;
 
 	if (CPUCLOCK_PERTHREAD(clock))
-		cpu_clock_sample(clkid, tsk, &t);
+		t = cpu_clock_sample(clkid, tsk);
 	else
-		cpu_clock_sample_group(clkid, tsk, &t, false);
+		t = cpu_clock_sample_group(clkid, tsk, false);
 	put_task_struct(tsk);
 
 	*tp = ns_to_timespec64(t);
@@ -604,11 +599,10 @@ static int posix_cpu_timer_set(struct k_itimer *timer, int timer_flags,
 	 * times (in arm_timer).  With an absolute time, we must
 	 * check if it's already passed.  In short, we need a sample.
 	 */
-	if (CPUCLOCK_PERTHREAD(timer->it_clock)) {
-		cpu_clock_sample(clkid, p, &val);
-	} else {
-		cpu_clock_sample_group(clkid, p, &val, true);
-	}
+	if (CPUCLOCK_PERTHREAD(timer->it_clock))
+		val = cpu_clock_sample(clkid, p);
+	else
+		val = cpu_clock_sample_group(clkid, p, true);
 
 	if (old) {
 		if (old_expires == 0) {
@@ -716,7 +710,7 @@ static void posix_cpu_timer_get(struct k_itimer *timer, struct itimerspec64 *itp
 	 * Sample the clock to take the difference with the expiry time.
 	 */
 	if (CPUCLOCK_PERTHREAD(timer->it_clock)) {
-		cpu_clock_sample(clkid, p, &now);
+		now = cpu_clock_sample(clkid, p);
 	} else {
 		struct sighand_struct *sighand;
 		unsigned long flags;
@@ -736,7 +730,7 @@ static void posix_cpu_timer_get(struct k_itimer *timer, struct itimerspec64 *itp
 			timer->it.cpu.expires = 0;
 			return;
 		} else {
-			cpu_clock_sample_group(clkid, p, &now, false);
+			now = cpu_clock_sample_group(clkid, p, false);
 			unlock_task_sighand(p, &flags);
 		}
 	}
@@ -998,7 +992,7 @@ static void posix_cpu_timer_rearm(struct k_itimer *timer)
 	 * Fetch the current sample and update the timer's expiry time.
 	 */
 	if (CPUCLOCK_PERTHREAD(timer->it_clock)) {
-		cpu_clock_sample(clkid, p, &now);
+		now = cpu_clock_sample(clkid, p);
 		bump_cpu_timer(timer, now);
 		if (unlikely(p->exit_state))
 			return;
@@ -1024,7 +1018,7 @@ static void posix_cpu_timer_rearm(struct k_itimer *timer)
 			/* If the process is dying, no need to rearm */
 			goto unlock;
 		}
-		cpu_clock_sample_group(clkid, p, &now, true);
+		now = cpu_clock_sample_group(clkid, p, true);
 		bump_cpu_timer(timer, now);
 		/* Leave the sighand locked for the call below.  */
 	}
@@ -1192,7 +1186,7 @@ void set_process_cpu_timer(struct task_struct *tsk, unsigned int clock_idx,
 	if (WARN_ON_ONCE(clock_idx >= CPUCLOCK_SCHED))
 		return;
 
-	cpu_clock_sample_group(clock_idx, tsk, &now, true);
+	now = cpu_clock_sample_group(clock_idx, tsk, true);
 
 	if (oldval) {
 		/*
-- 
cgit v1.2.3


From 0476ff2c151ee35bda2f938e8828b6f6113ebf55 Mon Sep 17 00:00:00 2001
From: Thomas Gleixner <tglx@linutronix.de>
Date: Wed, 21 Aug 2019 21:09:02 +0200
Subject: posix-cpu-timers: Sample task times once in expiry check

Sampling the task times twice does not make sense. Do it once.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://lkml.kernel.org/r/20190821192920.639878168@linutronix.de
---
 kernel/time/posix-cpu-timers.c | 10 ++++++----
 1 file changed, 6 insertions(+), 4 deletions(-)

(limited to 'kernel/time/posix-cpu-timers.c')

diff --git a/kernel/time/posix-cpu-timers.c b/kernel/time/posix-cpu-timers.c
index a2007ce9322a..98dab3e4be08 100644
--- a/kernel/time/posix-cpu-timers.c
+++ b/kernel/time/posix-cpu-timers.c
@@ -785,9 +785,9 @@ static inline void check_dl_overrun(struct task_struct *tsk)
 static void check_thread_timers(struct task_struct *tsk,
 				struct list_head *firing)
 {
-	struct list_head *timers = tsk->cpu_timers;
 	struct task_cputime *tsk_expires = &tsk->cputime_expires;
-	u64 expires;
+	struct list_head *timers = tsk->cpu_timers;
+	u64 expires, stime, utime;
 	unsigned long soft;
 
 	if (dl_task(tsk))
@@ -800,10 +800,12 @@ static void check_thread_timers(struct task_struct *tsk,
 	if (task_cputime_zero(&tsk->cputime_expires))
 		return;
 
-	expires = check_timers_list(timers, firing, prof_ticks(tsk));
+	task_cputime(tsk, &utime, &stime);
+
+	expires = check_timers_list(timers, firing, utime + stime);
 	tsk_expires->prof_exp = expires;
 
-	expires = check_timers_list(++timers, firing, virt_ticks(tsk));
+	expires = check_timers_list(++timers, firing, utime);
 	tsk_expires->virt_exp = expires;
 
 	tsk_expires->sched_exp = check_timers_list(++timers, firing,
-- 
cgit v1.2.3


From ab693c5a5e3107f035d5162e6ada9aaf7dd76a1d Mon Sep 17 00:00:00 2001
From: Thomas Gleixner <tglx@linutronix.de>
Date: Wed, 21 Aug 2019 21:09:03 +0200
Subject: posix-cpu-timers: Move prof/virt_ticks into caller

The functions have only one caller left. No point in having them.

Move the almost duplicated code into the caller and simplify it.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://lkml.kernel.org/r/20190821192920.729298382@linutronix.de
---
 kernel/time/posix-cpu-timers.c | 30 +++++++++---------------------
 1 file changed, 9 insertions(+), 21 deletions(-)

(limited to 'kernel/time/posix-cpu-timers.c')

diff --git a/kernel/time/posix-cpu-timers.c b/kernel/time/posix-cpu-timers.c
index 98dab3e4be08..b1c97664a730 100644
--- a/kernel/time/posix-cpu-timers.c
+++ b/kernel/time/posix-cpu-timers.c
@@ -130,23 +130,6 @@ static inline int task_cputime_zero(const struct task_cputime *cputime)
 	return 0;
 }
 
-static inline u64 prof_ticks(struct task_struct *p)
-{
-	u64 utime, stime;
-
-	task_cputime(p, &utime, &stime);
-
-	return utime + stime;
-}
-static inline u64 virt_ticks(struct task_struct *p)
-{
-	u64 utime, stime;
-
-	task_cputime(p, &utime, &stime);
-
-	return utime;
-}
-
 static int
 posix_cpu_clock_getres(const clockid_t which_clock, struct timespec64 *tp)
 {
@@ -184,13 +167,18 @@ posix_cpu_clock_set(const clockid_t clock, const struct timespec64 *tp)
  */
 static u64 cpu_clock_sample(const clockid_t clkid, struct task_struct *p)
 {
+	u64 utime, stime;
+
+	if (clkid == CPUCLOCK_SCHED)
+		return task_sched_runtime(p);
+
+	task_cputime(p, &utime, &stime);
+
 	switch (clkid) {
 	case CPUCLOCK_PROF:
-		return prof_ticks(p);
+		return utime + stime;
 	case CPUCLOCK_VIRT:
-		return virt_ticks(p);
-	case CPUCLOCK_SCHED:
-		return task_sched_runtime(p);
+		return utime;
 	default:
 		WARN_ON_ONCE(1);
 	}
-- 
cgit v1.2.3


From 2b69942f9021bf75bd1b001f53bd2578361fadf3 Mon Sep 17 00:00:00 2001
From: Thomas Gleixner <tglx@linutronix.de>
Date: Wed, 21 Aug 2019 21:09:04 +0200
Subject: posix-cpu-timers: Create a container struct

Per task/process data of posix CPU timers is all over the place which
makes the code hard to follow and requires ifdeffery.

Create a container to hold all this information in one place, so data is
consolidated and the ifdeffery can be confined to the posix timer header
file and removed from places like fork.

As a first step, move the cpu_timers list head array into the new struct
and clean up the initializers and simplify fork. The remaining #ifdef in
fork will be removed later.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://lkml.kernel.org/r/20190821192920.819418976@linutronix.de
---
 kernel/time/posix-cpu-timers.c | 20 ++++++++++----------
 1 file changed, 10 insertions(+), 10 deletions(-)

(limited to 'kernel/time/posix-cpu-timers.c')

diff --git a/kernel/time/posix-cpu-timers.c b/kernel/time/posix-cpu-timers.c
index b1c97664a730..849e2045fb6e 100644
--- a/kernel/time/posix-cpu-timers.c
+++ b/kernel/time/posix-cpu-timers.c
@@ -407,11 +407,11 @@ static void cleanup_timers_list(struct list_head *head)
  *
  * This must be called with the siglock held.
  */
-static void cleanup_timers(struct list_head *head)
+static void cleanup_timers(struct posix_cputimers *pct)
 {
-	cleanup_timers_list(head);
-	cleanup_timers_list(++head);
-	cleanup_timers_list(++head);
+	cleanup_timers_list(&pct->cpu_timers[CPUCLOCK_PROF]);
+	cleanup_timers_list(&pct->cpu_timers[CPUCLOCK_VIRT]);
+	cleanup_timers_list(&pct->cpu_timers[CPUCLOCK_SCHED]);
 }
 
 /*
@@ -421,11 +421,11 @@ static void cleanup_timers(struct list_head *head)
  */
 void posix_cpu_timers_exit(struct task_struct *tsk)
 {
-	cleanup_timers(tsk->cpu_timers);
+	cleanup_timers(&tsk->posix_cputimers);
 }
 void posix_cpu_timers_exit_group(struct task_struct *tsk)
 {
-	cleanup_timers(tsk->signal->cpu_timers);
+	cleanup_timers(&tsk->signal->posix_cputimers);
 }
 
 static inline int expires_gt(u64 expires, u64 new_exp)
@@ -446,10 +446,10 @@ static void arm_timer(struct k_itimer *timer)
 	struct cpu_timer_list *next;
 
 	if (CPUCLOCK_PERTHREAD(timer->it_clock)) {
-		head = p->cpu_timers;
+		head = p->posix_cputimers.cpu_timers;
 		cputime_expires = &p->cputime_expires;
 	} else {
-		head = p->signal->cpu_timers;
+		head = p->signal->posix_cputimers.cpu_timers;
 		cputime_expires = &p->signal->cputime_expires;
 	}
 	head += CPUCLOCK_WHICH(timer->it_clock);
@@ -773,8 +773,8 @@ static inline void check_dl_overrun(struct task_struct *tsk)
 static void check_thread_timers(struct task_struct *tsk,
 				struct list_head *firing)
 {
+	struct list_head *timers = tsk->posix_cputimers.cpu_timers;
 	struct task_cputime *tsk_expires = &tsk->cputime_expires;
-	struct list_head *timers = tsk->cpu_timers;
 	u64 expires, stime, utime;
 	unsigned long soft;
 
@@ -879,9 +879,9 @@ static void check_process_timers(struct task_struct *tsk,
 				 struct list_head *firing)
 {
 	struct signal_struct *const sig = tsk->signal;
+	struct list_head *timers = sig->posix_cputimers.cpu_timers;
 	u64 utime, ptime, virt_expires, prof_expires;
 	u64 sum_sched_runtime, sched_expires;
-	struct list_head *timers = sig->cpu_timers;
 	struct task_cputime cputime;
 	unsigned long soft;
 
-- 
cgit v1.2.3


From 3a245c0f110e2bfcf7f2cd2248a29005c78999e3 Mon Sep 17 00:00:00 2001
From: Thomas Gleixner <tglx@linutronix.de>
Date: Wed, 21 Aug 2019 21:09:06 +0200
Subject: posix-cpu-timers: Move expiry cache into struct posix_cputimers

The expiry cache belongs into the posix_cputimers container where the other
cpu timers information is.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://lkml.kernel.org/r/20190821192921.014444012@linutronix.de
---
 kernel/time/posix-cpu-timers.c | 45 +++++++++++++++++++++++++-----------------
 1 file changed, 27 insertions(+), 18 deletions(-)

(limited to 'kernel/time/posix-cpu-timers.c')

diff --git a/kernel/time/posix-cpu-timers.c b/kernel/time/posix-cpu-timers.c
index 849e2045fb6e..3e29d1692437 100644
--- a/kernel/time/posix-cpu-timers.c
+++ b/kernel/time/posix-cpu-timers.c
@@ -20,11 +20,18 @@
 
 static void posix_cpu_timer_rearm(struct k_itimer *timer);
 
+void posix_cputimers_group_init(struct posix_cputimers *pct, u64 cpu_limit)
+{
+	posix_cputimers_init(pct);
+	if (cpu_limit != RLIM_INFINITY)
+		pct->cputime_expires.prof_exp = cpu_limit * NSEC_PER_SEC;
+}
+
 /*
  * Called after updating RLIMIT_CPU to run cpu timer and update
- * tsk->signal->cputime_expires expiration cache if necessary. Needs
- * siglock protection since other code may update expiration cache as
- * well.
+ * tsk->signal->posix_cputimers.cputime_expires expiration cache if
+ * necessary. Needs siglock protection since other code may update
+ * expiration cache as well.
  */
 void update_rlimit_cpu(struct task_struct *task, unsigned long rlim_new)
 {
@@ -447,10 +454,10 @@ static void arm_timer(struct k_itimer *timer)
 
 	if (CPUCLOCK_PERTHREAD(timer->it_clock)) {
 		head = p->posix_cputimers.cpu_timers;
-		cputime_expires = &p->cputime_expires;
+		cputime_expires = &p->posix_cputimers.cputime_expires;
 	} else {
 		head = p->signal->posix_cputimers.cpu_timers;
-		cputime_expires = &p->signal->cputime_expires;
+		cputime_expires = &p->signal->posix_cputimers.cputime_expires;
 	}
 	head += CPUCLOCK_WHICH(timer->it_clock);
 
@@ -774,7 +781,7 @@ static void check_thread_timers(struct task_struct *tsk,
 				struct list_head *firing)
 {
 	struct list_head *timers = tsk->posix_cputimers.cpu_timers;
-	struct task_cputime *tsk_expires = &tsk->cputime_expires;
+	struct task_cputime *tsk_expires = &tsk->posix_cputimers.cputime_expires;
 	u64 expires, stime, utime;
 	unsigned long soft;
 
@@ -785,7 +792,7 @@ static void check_thread_timers(struct task_struct *tsk,
 	 * If cputime_expires is zero, then there are no active
 	 * per thread CPU timers.
 	 */
-	if (task_cputime_zero(&tsk->cputime_expires))
+	if (task_cputime_zero(tsk_expires))
 		return;
 
 	task_cputime(tsk, &utime, &stime);
@@ -954,10 +961,10 @@ static void check_process_timers(struct task_struct *tsk,
 			prof_expires = x;
 	}
 
-	sig->cputime_expires.prof_exp = prof_expires;
-	sig->cputime_expires.virt_exp = virt_expires;
-	sig->cputime_expires.sched_exp = sched_expires;
-	if (task_cputime_zero(&sig->cputime_expires))
+	sig->posix_cputimers.cputime_expires.prof_exp = prof_expires;
+	sig->posix_cputimers.cputime_expires.virt_exp = virt_expires;
+	sig->posix_cputimers.cputime_expires.sched_exp = sched_expires;
+	if (task_cputime_zero(&sig->posix_cputimers.cputime_expires))
 		stop_process_timers(sig);
 
 	sig->cputimer.checking_timer = false;
@@ -1058,12 +1065,13 @@ static inline int fastpath_timer_check(struct task_struct *tsk)
 {
 	struct signal_struct *sig;
 
-	if (!task_cputime_zero(&tsk->cputime_expires)) {
+	if (!task_cputime_zero(&tsk->posix_cputimers.cputime_expires)) {
 		struct task_cputime task_sample;
 
 		task_cputime(tsk, &task_sample.utime, &task_sample.stime);
 		task_sample.sum_exec_runtime = tsk->se.sum_exec_runtime;
-		if (task_cputime_expired(&task_sample, &tsk->cputime_expires))
+		if (task_cputime_expired(&task_sample,
+					 &tsk->posix_cputimers.cputime_expires))
 			return 1;
 	}
 
@@ -1088,7 +1096,8 @@ static inline int fastpath_timer_check(struct task_struct *tsk)
 
 		sample_cputime_atomic(&group_sample, &sig->cputimer.cputime_atomic);
 
-		if (task_cputime_expired(&group_sample, &sig->cputime_expires))
+		if (task_cputime_expired(&group_sample,
+					 &sig->posix_cputimers.cputime_expires))
 			return 1;
 	}
 
@@ -1204,12 +1213,12 @@ void set_process_cpu_timer(struct task_struct *tsk, unsigned int clock_idx,
 	 */
 	switch (clock_idx) {
 	case CPUCLOCK_PROF:
-		if (expires_gt(tsk->signal->cputime_expires.prof_exp, *newval))
-			tsk->signal->cputime_expires.prof_exp = *newval;
+		if (expires_gt(tsk->signal->posix_cputimers.cputime_expires.prof_exp, *newval))
+			tsk->signal->posix_cputimers.cputime_expires.prof_exp = *newval;
 		break;
 	case CPUCLOCK_VIRT:
-		if (expires_gt(tsk->signal->cputime_expires.virt_exp, *newval))
-			tsk->signal->cputime_expires.virt_exp = *newval;
+		if (expires_gt(tsk->signal->posix_cputimers.cputime_expires.virt_exp, *newval))
+			tsk->signal->posix_cputimers.cputime_expires.virt_exp = *newval;
 		break;
 	}
 
-- 
cgit v1.2.3


From 11b8462f7e1d25f639c88949a2746a9c2667a766 Mon Sep 17 00:00:00 2001
From: Thomas Gleixner <tglx@linutronix.de>
Date: Wed, 21 Aug 2019 21:09:07 +0200
Subject: posix-cpu-timers: Provide array based access to expiry cache

Using struct task_cputime for the expiry cache is a pretty odd choice and
comes with magic defines to rename the fields for usage in the expiry
cache.

struct task_cputime is basically a u64 array with 3 members, but it has
distinct members.

The expiry cache content is different than the content of task_cputime
because

  expiry[PROF]  = task_cputime.stime + task_cputime.utime
  expiry[VIRT]  = task_cputime.utime
  expiry[SCHED] = task_cputime.sum_exec_runtime

So there is no direct mapping between task_cputime and the expiry cache and
the #define based remapping is just a horrible hack.

Having the expiry cache array based allows further simplification of the
expiry code.

To avoid an all in one cleanup which is hard to review add a temporary
anonymous union into struct task_cputime which allows array based access to
it. That requires to reorder the members. Add a build time sanity check to
validate that the members are at the same place.

The union and the build time checks will be removed after conversion.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://lkml.kernel.org/r/20190821192921.105793824@linutronix.de
---
 kernel/time/posix-cpu-timers.c | 12 +++++++++++-
 1 file changed, 11 insertions(+), 1 deletion(-)

(limited to 'kernel/time/posix-cpu-timers.c')

diff --git a/kernel/time/posix-cpu-timers.c b/kernel/time/posix-cpu-timers.c
index 3e29d1692437..a38b6d04e8b5 100644
--- a/kernel/time/posix-cpu-timers.c
+++ b/kernel/time/posix-cpu-timers.c
@@ -18,13 +18,23 @@
 
 #include "posix-timers.h"
 
+static inline void temporary_check(void)
+{
+	BUILD_BUG_ON(offsetof(struct task_cputime, stime) !=
+		     CPUCLOCK_PROF * sizeof(u64));
+	BUILD_BUG_ON(offsetof(struct task_cputime, utime) !=
+		     CPUCLOCK_VIRT * sizeof(u64));
+	BUILD_BUG_ON(offsetof(struct task_cputime, sum_exec_runtime) !=
+		     CPUCLOCK_SCHED * sizeof(u64));
+}
+
 static void posix_cpu_timer_rearm(struct k_itimer *timer);
 
 void posix_cputimers_group_init(struct posix_cputimers *pct, u64 cpu_limit)
 {
 	posix_cputimers_init(pct);
 	if (cpu_limit != RLIM_INFINITY)
-		pct->cputime_expires.prof_exp = cpu_limit * NSEC_PER_SEC;
+		pct->expiries[CPUCLOCK_PROF] = cpu_limit * NSEC_PER_SEC;
 }
 
 /*
-- 
cgit v1.2.3


From 3b495b22d04df3220ccae829bf7c5cadb3059ccf Mon Sep 17 00:00:00 2001
From: Thomas Gleixner <tglx@linutronix.de>
Date: Wed, 21 Aug 2019 21:09:08 +0200
Subject: posix-cpu-timers: Simplify timer queueing

Now that the expiry cache can be accessed as an array, the per clock
checking can be reduced to just comparing the corresponding array elements.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://lkml.kernel.org/r/20190821192921.212129449@linutronix.de
---
 kernel/time/posix-cpu-timers.c | 55 ++++++++++++++++--------------------------
 1 file changed, 21 insertions(+), 34 deletions(-)

(limited to 'kernel/time/posix-cpu-timers.c')

diff --git a/kernel/time/posix-cpu-timers.c b/kernel/time/posix-cpu-timers.c
index a38b6d04e8b5..b13241756fdd 100644
--- a/kernel/time/posix-cpu-timers.c
+++ b/kernel/time/posix-cpu-timers.c
@@ -456,20 +456,20 @@ static inline int expires_gt(u64 expires, u64 new_exp)
  */
 static void arm_timer(struct k_itimer *timer)
 {
+	struct cpu_timer_list *const nt = &timer->it.cpu;
+	int clkidx = CPUCLOCK_WHICH(timer->it_clock);
+	u64 *cpuexp, newexp = timer->it.cpu.expires;
 	struct task_struct *p = timer->it.cpu.task;
 	struct list_head *head, *listpos;
-	struct task_cputime *cputime_expires;
-	struct cpu_timer_list *const nt = &timer->it.cpu;
 	struct cpu_timer_list *next;
 
 	if (CPUCLOCK_PERTHREAD(timer->it_clock)) {
-		head = p->posix_cputimers.cpu_timers;
-		cputime_expires = &p->posix_cputimers.cputime_expires;
+		head = p->posix_cputimers.cpu_timers + clkidx;
+		cpuexp = p->posix_cputimers.expiries + clkidx;
 	} else {
-		head = p->signal->posix_cputimers.cpu_timers;
-		cputime_expires = &p->signal->posix_cputimers.cputime_expires;
+		head = p->signal->posix_cputimers.cpu_timers + clkidx;
+		cpuexp = p->signal->posix_cputimers.expiries + clkidx;
 	}
-	head += CPUCLOCK_WHICH(timer->it_clock);
 
 	listpos = head;
 	list_for_each_entry(next, head, entry) {
@@ -479,35 +479,22 @@ static void arm_timer(struct k_itimer *timer)
 	}
 	list_add(&nt->entry, listpos);
 
-	if (listpos == head) {
-		u64 exp = nt->expires;
+	if (listpos != head)
+		return;
 
-		/*
-		 * We are the new earliest-expiring POSIX 1.b timer, hence
-		 * need to update expiration cache. Take into account that
-		 * for process timers we share expiration cache with itimers
-		 * and RLIMIT_CPU and for thread timers with RLIMIT_RTTIME.
-		 */
+	/*
+	 * We are the new earliest-expiring POSIX 1.b timer, hence
+	 * need to update expiration cache. Take into account that
+	 * for process timers we share expiration cache with itimers
+	 * and RLIMIT_CPU and for thread timers with RLIMIT_RTTIME.
+	 */
+	if (expires_gt(*cpuexp, newexp))
+		*cpuexp = newexp;
 
-		switch (CPUCLOCK_WHICH(timer->it_clock)) {
-		case CPUCLOCK_PROF:
-			if (expires_gt(cputime_expires->prof_exp, exp))
-				cputime_expires->prof_exp = exp;
-			break;
-		case CPUCLOCK_VIRT:
-			if (expires_gt(cputime_expires->virt_exp, exp))
-				cputime_expires->virt_exp = exp;
-			break;
-		case CPUCLOCK_SCHED:
-			if (expires_gt(cputime_expires->sched_exp, exp))
-				cputime_expires->sched_exp = exp;
-			break;
-		}
-		if (CPUCLOCK_PERTHREAD(timer->it_clock))
-			tick_dep_set_task(p, TICK_DEP_BIT_POSIX_TIMER);
-		else
-			tick_dep_set_signal(p->signal, TICK_DEP_BIT_POSIX_TIMER);
-	}
+	if (CPUCLOCK_PERTHREAD(timer->it_clock))
+		tick_dep_set_task(p, TICK_DEP_BIT_POSIX_TIMER);
+	else
+		tick_dep_set_signal(p->signal, TICK_DEP_BIT_POSIX_TIMER);
 }
 
 /*
-- 
cgit v1.2.3


From 1b0dd96d0f07c482bf1d04987cc1b35e376a7518 Mon Sep 17 00:00:00 2001
From: Thomas Gleixner <tglx@linutronix.de>
Date: Wed, 21 Aug 2019 21:09:09 +0200
Subject: posix-cpu-timers: Simplify set_process_cpu_timer()

The expiry cache can now be accessed as an array. Replace the per clock
checks with a simple comparison of the clock indexed array member.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://lkml.kernel.org/r/20190821192921.303316423@linutronix.de
---
 kernel/time/posix-cpu-timers.c | 24 ++++++++----------------
 1 file changed, 8 insertions(+), 16 deletions(-)

(limited to 'kernel/time/posix-cpu-timers.c')

diff --git a/kernel/time/posix-cpu-timers.c b/kernel/time/posix-cpu-timers.c
index b13241756fdd..2c47ce6cab8c 100644
--- a/kernel/time/posix-cpu-timers.c
+++ b/kernel/time/posix-cpu-timers.c
@@ -1174,15 +1174,15 @@ void run_posix_cpu_timers(void)
  * Set one of the process-wide special case CPU timers or RLIMIT_CPU.
  * The tsk->sighand->siglock must be held by the caller.
  */
-void set_process_cpu_timer(struct task_struct *tsk, unsigned int clock_idx,
+void set_process_cpu_timer(struct task_struct *tsk, unsigned int clkid,
 			   u64 *newval, u64 *oldval)
 {
-	u64 now;
+	u64 now, *expiry = tsk->signal->posix_cputimers.expiries + clkid;
 
-	if (WARN_ON_ONCE(clock_idx >= CPUCLOCK_SCHED))
+	if (WARN_ON_ONCE(clkid >= CPUCLOCK_SCHED))
 		return;
 
-	now = cpu_clock_sample_group(clock_idx, tsk, true);
+	now = cpu_clock_sample_group(clkid, tsk, true);
 
 	if (oldval) {
 		/*
@@ -1205,19 +1205,11 @@ void set_process_cpu_timer(struct task_struct *tsk, unsigned int clock_idx,
 	}
 
 	/*
-	 * Update expiration cache if we are the earliest timer, or eventually
-	 * RLIMIT_CPU limit is earlier than prof_exp cpu timer expire.
+	 * Update expiration cache if this is the earliest timer. CPUCLOCK_PROF
+	 * expiry cache is also used by RLIMIT_CPU!.
 	 */
-	switch (clock_idx) {
-	case CPUCLOCK_PROF:
-		if (expires_gt(tsk->signal->posix_cputimers.cputime_expires.prof_exp, *newval))
-			tsk->signal->posix_cputimers.cputime_expires.prof_exp = *newval;
-		break;
-	case CPUCLOCK_VIRT:
-		if (expires_gt(tsk->signal->posix_cputimers.cputime_expires.virt_exp, *newval))
-			tsk->signal->posix_cputimers.cputime_expires.virt_exp = *newval;
-		break;
-	}
+	if (expires_gt(*expiry, *newval))
+		*expiry = *newval;
 
 	tick_dep_set_signal(tsk->signal, TICK_DEP_BIT_POSIX_TIMER);
 }
-- 
cgit v1.2.3


From c02b078e63a6f42029cb655d0aa3c991271637ac Mon Sep 17 00:00:00 2001
From: Thomas Gleixner <tglx@linutronix.de>
Date: Wed, 21 Aug 2019 21:09:10 +0200
Subject: posix-cpu-timers: Switch check_*_timers() to array cache

Use the array based expiry cache in check_thread_timers() and convert the
store in check_process_timers() for consistency.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://lkml.kernel.org/r/20190821192921.408222378@linutronix.de
---
 kernel/time/posix-cpu-timers.c | 26 +++++++++++---------------
 1 file changed, 11 insertions(+), 15 deletions(-)

(limited to 'kernel/time/posix-cpu-timers.c')

diff --git a/kernel/time/posix-cpu-timers.c b/kernel/time/posix-cpu-timers.c
index 2c47ce6cab8c..220e3c7ae849 100644
--- a/kernel/time/posix-cpu-timers.c
+++ b/kernel/time/posix-cpu-timers.c
@@ -778,8 +778,7 @@ static void check_thread_timers(struct task_struct *tsk,
 				struct list_head *firing)
 {
 	struct list_head *timers = tsk->posix_cputimers.cpu_timers;
-	struct task_cputime *tsk_expires = &tsk->posix_cputimers.cputime_expires;
-	u64 expires, stime, utime;
+	u64 stime, utime, *expires = tsk->posix_cputimers.expiries;
 	unsigned long soft;
 
 	if (dl_task(tsk))
@@ -789,19 +788,14 @@ static void check_thread_timers(struct task_struct *tsk,
 	 * If cputime_expires is zero, then there are no active
 	 * per thread CPU timers.
 	 */
-	if (task_cputime_zero(tsk_expires))
+	if (task_cputime_zero(&tsk->posix_cputimers.cputime_expires))
 		return;
 
 	task_cputime(tsk, &utime, &stime);
 
-	expires = check_timers_list(timers, firing, utime + stime);
-	tsk_expires->prof_exp = expires;
-
-	expires = check_timers_list(++timers, firing, utime);
-	tsk_expires->virt_exp = expires;
-
-	tsk_expires->sched_exp = check_timers_list(++timers, firing,
-						   tsk->se.sum_exec_runtime);
+	*expires++ = check_timers_list(timers, firing, utime + stime);
+	*expires++ = check_timers_list(++timers, firing, utime);
+	*expires = check_timers_list(++timers, firing, tsk->se.sum_exec_runtime);
 
 	/*
 	 * Check for the special case thread timers.
@@ -839,7 +833,8 @@ static void check_thread_timers(struct task_struct *tsk,
 			__group_send_sig_info(SIGXCPU, SEND_SIG_PRIV, tsk);
 		}
 	}
-	if (task_cputime_zero(tsk_expires))
+
+	if (task_cputime_zero(&tsk->posix_cputimers.cputime_expires))
 		tick_dep_clear_task(tsk, TICK_DEP_BIT_POSIX_TIMER);
 }
 
@@ -958,9 +953,10 @@ static void check_process_timers(struct task_struct *tsk,
 			prof_expires = x;
 	}
 
-	sig->posix_cputimers.cputime_expires.prof_exp = prof_expires;
-	sig->posix_cputimers.cputime_expires.virt_exp = virt_expires;
-	sig->posix_cputimers.cputime_expires.sched_exp = sched_expires;
+	sig->posix_cputimers.expiries[CPUCLOCK_PROF] = prof_expires;
+	sig->posix_cputimers.expiries[CPUCLOCK_VIRT] = virt_expires;
+	sig->posix_cputimers.expiries[CPUCLOCK_SCHED] = sched_expires;
+
 	if (task_cputime_zero(&sig->posix_cputimers.cputime_expires))
 		stop_process_timers(sig);
 
-- 
cgit v1.2.3


From b0d524f77956eec887b30732af1f5f98cbf62b9f Mon Sep 17 00:00:00 2001
From: Thomas Gleixner <tglx@linutronix.de>
Date: Wed, 21 Aug 2019 21:09:12 +0200
Subject: posix-cpu-timers: Provide array based sample functions

Instead of using task_cputime and doing the addition of utime and stime at
all call sites, it's way simpler to have a sample array which allows
indexed based checks against the expiry cache array.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://lkml.kernel.org/r/20190821192921.590362974@linutronix.de
---
 kernel/time/posix-cpu-timers.c | 26 ++++++++++++++++++++++++++
 1 file changed, 26 insertions(+)

(limited to 'kernel/time/posix-cpu-timers.c')

diff --git a/kernel/time/posix-cpu-timers.c b/kernel/time/posix-cpu-timers.c
index 220e3c7ae849..11c841c6a45d 100644
--- a/kernel/time/posix-cpu-timers.c
+++ b/kernel/time/posix-cpu-timers.c
@@ -202,6 +202,32 @@ static u64 cpu_clock_sample(const clockid_t clkid, struct task_struct *p)
 	return 0;
 }
 
+static inline void store_samples(u64 *samples, u64 stime, u64 utime, u64 rtime)
+{
+	samples[CPUCLOCK_PROF] = stime + utime;
+	samples[CPUCLOCK_VIRT] = utime;
+	samples[CPUCLOCK_SCHED] = rtime;
+}
+
+static void task_sample_cputime(struct task_struct *p, u64 *samples)
+{
+	u64 stime, utime;
+
+	task_cputime(p, &utime, &stime);
+	store_samples(samples, stime, utime, p->se.sum_exec_runtime);
+}
+
+static void proc_sample_cputime_atomic(struct task_cputime_atomic *at,
+				       u64 *samples)
+{
+	u64 stime, utime, rtime;
+
+	utime = atomic64_read(&at->utime);
+	stime = atomic64_read(&at->stime);
+	rtime = atomic64_read(&at->sum_exec_runtime);
+	store_samples(samples, stime, utime, rtime);
+}
+
 /*
  * Set cputime to sum_cputime if sum_cputime > cputime. Use cmpxchg
  * to avoid race conditions with concurrent updates to cputime.
-- 
cgit v1.2.3


From 001f7971433a53bb76443cd8f5827ca27b0bc8ca Mon Sep 17 00:00:00 2001
From: Thomas Gleixner <tglx@linutronix.de>
Date: Wed, 21 Aug 2019 21:09:13 +0200
Subject: posix-cpu-timers: Make expiry checks array based

The expiry cache is an array indexed by clock ids. The new sample functions
allow to retrieve a corresponding array of samples.

Convert the fastpath expiry checks to make use of the new sample functions
and do the comparisons on the sample and the expiry array.

Make the check for the expiry array being zero array based as well.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://lkml.kernel.org/r/20190821192921.695481430@linutronix.de
---
 kernel/time/posix-cpu-timers.c | 85 ++++++++++++++++++------------------------
 1 file changed, 36 insertions(+), 49 deletions(-)

(limited to 'kernel/time/posix-cpu-timers.c')

diff --git a/kernel/time/posix-cpu-timers.c b/kernel/time/posix-cpu-timers.c
index 11c841c6a45d..e159b039e44a 100644
--- a/kernel/time/posix-cpu-timers.c
+++ b/kernel/time/posix-cpu-timers.c
@@ -39,7 +39,7 @@ void posix_cputimers_group_init(struct posix_cputimers *pct, u64 cpu_limit)
 
 /*
  * Called after updating RLIMIT_CPU to run cpu timer and update
- * tsk->signal->posix_cputimers.cputime_expires expiration cache if
+ * tsk->signal->posix_cputimers.expiries expiration cache if
  * necessary. Needs siglock protection since other code may update
  * expiration cache as well.
  */
@@ -132,19 +132,9 @@ static void bump_cpu_timer(struct k_itimer *timer, u64 now)
 	}
 }
 
-/**
- * task_cputime_zero - Check a task_cputime struct for all zero fields.
- *
- * @cputime:	The struct to compare.
- *
- * Checks @cputime to see if all fields are zero.  Returns true if all fields
- * are zero, false if any field is nonzero.
- */
-static inline int task_cputime_zero(const struct task_cputime *cputime)
+static inline bool expiry_cache_is_zero(const u64 *ec)
 {
-	if (!cputime->utime && !cputime->stime && !cputime->sum_exec_runtime)
-		return 1;
-	return 0;
+	return !(ec[CPUCLOCK_PROF] | ec[CPUCLOCK_VIRT] | ec[CPUCLOCK_SCHED]);
 }
 
 static int
@@ -811,10 +801,10 @@ static void check_thread_timers(struct task_struct *tsk,
 		check_dl_overrun(tsk);
 
 	/*
-	 * If cputime_expires is zero, then there are no active
-	 * per thread CPU timers.
+	 * If the expiry cache is zero, then there are no active per thread
+	 * CPU timers.
 	 */
-	if (task_cputime_zero(&tsk->posix_cputimers.cputime_expires))
+	if (expiry_cache_is_zero(tsk->posix_cputimers.expiries))
 		return;
 
 	task_cputime(tsk, &utime, &stime);
@@ -860,7 +850,7 @@ static void check_thread_timers(struct task_struct *tsk,
 		}
 	}
 
-	if (task_cputime_zero(&tsk->posix_cputimers.cputime_expires))
+	if (expiry_cache_is_zero(tsk->posix_cputimers.expiries))
 		tick_dep_clear_task(tsk, TICK_DEP_BIT_POSIX_TIMER);
 }
 
@@ -983,7 +973,7 @@ static void check_process_timers(struct task_struct *tsk,
 	sig->posix_cputimers.expiries[CPUCLOCK_VIRT] = virt_expires;
 	sig->posix_cputimers.expiries[CPUCLOCK_SCHED] = sched_expires;
 
-	if (task_cputime_zero(&sig->posix_cputimers.cputime_expires))
+	if (expiry_cache_is_zero(sig->posix_cputimers.expiries))
 		stop_process_timers(sig);
 
 	sig->cputimer.checking_timer = false;
@@ -1048,26 +1038,23 @@ unlock:
 }
 
 /**
- * task_cputime_expired - Compare two task_cputime entities.
+ * task_cputimers_expired - Compare two task_cputime entities.
  *
- * @sample:	The task_cputime structure to be checked for expiration.
- * @expires:	Expiration times, against which @sample will be checked.
+ * @samples:	Array of current samples for the CPUCLOCK clocks
+ * @expiries:	Array of expiry values for the CPUCLOCK clocks
  *
- * Checks @sample against @expires to see if any field of @sample has expired.
- * Returns true if any field of the former is greater than the corresponding
- * field of the latter if the latter field is set.  Otherwise returns false.
+ * Returns true if any mmember of @samples is greater than the corresponding
+ * member of @expiries if that member is non zero. False otherwise
  */
-static inline int task_cputime_expired(const struct task_cputime *sample,
-					const struct task_cputime *expires)
+static inline bool task_cputimers_expired(const u64 *sample, const u64 *expiries)
 {
-	if (expires->utime && sample->utime >= expires->utime)
-		return 1;
-	if (expires->stime && sample->utime + sample->stime >= expires->stime)
-		return 1;
-	if (expires->sum_exec_runtime != 0 &&
-	    sample->sum_exec_runtime >= expires->sum_exec_runtime)
-		return 1;
-	return 0;
+	int i;
+
+	for (i = 0; i < CPUCLOCK_MAX; i++) {
+		if (expiries[i] && sample[i] >= expiries[i])
+			return true;
+	}
+	return false;
 }
 
 /**
@@ -1080,18 +1067,17 @@ static inline int task_cputime_expired(const struct task_cputime *sample,
  * timers and compare them with the corresponding expiration times.  Return
  * true if a timer has expired, else return false.
  */
-static inline int fastpath_timer_check(struct task_struct *tsk)
+static inline bool fastpath_timer_check(struct task_struct *tsk)
 {
+	u64 *expiries = tsk->posix_cputimers.expiries;
 	struct signal_struct *sig;
 
-	if (!task_cputime_zero(&tsk->posix_cputimers.cputime_expires)) {
-		struct task_cputime task_sample;
+	if (!expiry_cache_is_zero(expiries)) {
+		u64 samples[CPUCLOCK_MAX];
 
-		task_cputime(tsk, &task_sample.utime, &task_sample.stime);
-		task_sample.sum_exec_runtime = tsk->se.sum_exec_runtime;
-		if (task_cputime_expired(&task_sample,
-					 &tsk->posix_cputimers.cputime_expires))
-			return 1;
+		task_sample_cputime(tsk, samples);
+		if (task_cputimers_expired(samples, expiries))
+			return true;
 	}
 
 	sig = tsk->signal;
@@ -1111,19 +1097,20 @@ static inline int fastpath_timer_check(struct task_struct *tsk)
 	 */
 	if (READ_ONCE(sig->cputimer.running) &&
 	    !READ_ONCE(sig->cputimer.checking_timer)) {
-		struct task_cputime group_sample;
+		u64 samples[CPUCLOCK_MAX];
 
-		sample_cputime_atomic(&group_sample, &sig->cputimer.cputime_atomic);
+		proc_sample_cputime_atomic(&sig->cputimer.cputime_atomic,
+					   samples);
 
-		if (task_cputime_expired(&group_sample,
-					 &sig->posix_cputimers.cputime_expires))
-			return 1;
+		if (task_cputimers_expired(samples,
+					   sig->posix_cputimers.expiries))
+			return true;
 	}
 
 	if (dl_task(tsk) && tsk->dl.dl_overrun)
-		return 1;
+		return true;
 
-	return 0;
+	return false;
 }
 
 /*
-- 
cgit v1.2.3


From 46b883995c88520f2c4de6a64cccc04c69d59f83 Mon Sep 17 00:00:00 2001
From: Thomas Gleixner <tglx@linutronix.de>
Date: Wed, 21 Aug 2019 21:09:14 +0200
Subject: posix-cpu-timers: Remove cputime_expires

The last users of the magic struct cputime based expiry cache are
gone. Remove the leftovers.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://lkml.kernel.org/r/20190821192921.790209622@linutronix.de
---
 kernel/time/posix-cpu-timers.c | 10 ----------
 1 file changed, 10 deletions(-)

(limited to 'kernel/time/posix-cpu-timers.c')

diff --git a/kernel/time/posix-cpu-timers.c b/kernel/time/posix-cpu-timers.c
index e159b039e44a..ffd49181e23d 100644
--- a/kernel/time/posix-cpu-timers.c
+++ b/kernel/time/posix-cpu-timers.c
@@ -18,16 +18,6 @@
 
 #include "posix-timers.h"
 
-static inline void temporary_check(void)
-{
-	BUILD_BUG_ON(offsetof(struct task_cputime, stime) !=
-		     CPUCLOCK_PROF * sizeof(u64));
-	BUILD_BUG_ON(offsetof(struct task_cputime, utime) !=
-		     CPUCLOCK_VIRT * sizeof(u64));
-	BUILD_BUG_ON(offsetof(struct task_cputime, sum_exec_runtime) !=
-		     CPUCLOCK_SCHED * sizeof(u64));
-}
-
 static void posix_cpu_timer_rearm(struct k_itimer *timer);
 
 void posix_cputimers_group_init(struct posix_cputimers *pct, u64 cpu_limit)
-- 
cgit v1.2.3


From 87dc64480fb19a6a0fedbdff1e2557be50673287 Mon Sep 17 00:00:00 2001
From: Thomas Gleixner <tglx@linutronix.de>
Date: Mon, 26 Aug 2019 20:22:24 +0200
Subject: posix-cpu-timers: Restructure expiry array

Now that the abused struct task_cputime is gone, it's more natural to
bundle the expiry cache and the list head of each clock into a struct and
have an array of those structs.

Follow the hrtimer naming convention of 'bases' and rename the expiry cache
to 'nextevt' and adapt all usage sites.

Generates also better code .text size shrinks by 80 bytes.

Suggested-by: Ingo Molnar <mingo@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://lkml.kernel.org/r/alpine.DEB.2.21.1908262021140.1939@nanos.tec.linutronix.de
---
 kernel/time/posix-cpu-timers.c | 105 ++++++++++++++++++++++-------------------
 1 file changed, 56 insertions(+), 49 deletions(-)

(limited to 'kernel/time/posix-cpu-timers.c')

diff --git a/kernel/time/posix-cpu-timers.c b/kernel/time/posix-cpu-timers.c
index ffd49181e23d..9ac601abc4c4 100644
--- a/kernel/time/posix-cpu-timers.c
+++ b/kernel/time/posix-cpu-timers.c
@@ -24,13 +24,13 @@ void posix_cputimers_group_init(struct posix_cputimers *pct, u64 cpu_limit)
 {
 	posix_cputimers_init(pct);
 	if (cpu_limit != RLIM_INFINITY)
-		pct->expiries[CPUCLOCK_PROF] = cpu_limit * NSEC_PER_SEC;
+		pct->bases[CPUCLOCK_PROF].nextevt = cpu_limit * NSEC_PER_SEC;
 }
 
 /*
  * Called after updating RLIMIT_CPU to run cpu timer and update
- * tsk->signal->posix_cputimers.expiries expiration cache if
- * necessary. Needs siglock protection since other code may update
+ * tsk->signal->posix_cputimers.bases[clock].nextevt expiration cache if
+ * necessary. Needs siglock protection since other code may update the
  * expiration cache as well.
  */
 void update_rlimit_cpu(struct task_struct *task, unsigned long rlim_new)
@@ -122,9 +122,11 @@ static void bump_cpu_timer(struct k_itimer *timer, u64 now)
 	}
 }
 
-static inline bool expiry_cache_is_zero(const u64 *ec)
+static inline bool expiry_cache_is_zero(const struct posix_cputimers *pct)
 {
-	return !(ec[CPUCLOCK_PROF] | ec[CPUCLOCK_VIRT] | ec[CPUCLOCK_SCHED]);
+	return !(pct->bases[CPUCLOCK_PROF].nextevt |
+		 pct->bases[CPUCLOCK_VIRT].nextevt |
+		 pct->bases[CPUCLOCK_SCHED].nextevt);
 }
 
 static int
@@ -432,9 +434,9 @@ static void cleanup_timers_list(struct list_head *head)
  */
 static void cleanup_timers(struct posix_cputimers *pct)
 {
-	cleanup_timers_list(&pct->cpu_timers[CPUCLOCK_PROF]);
-	cleanup_timers_list(&pct->cpu_timers[CPUCLOCK_VIRT]);
-	cleanup_timers_list(&pct->cpu_timers[CPUCLOCK_SCHED]);
+	cleanup_timers_list(&pct->bases[CPUCLOCK_PROF].cpu_timers);
+	cleanup_timers_list(&pct->bases[CPUCLOCK_VIRT].cpu_timers);
+	cleanup_timers_list(&pct->bases[CPUCLOCK_SCHED].cpu_timers);
 }
 
 /*
@@ -464,21 +466,19 @@ static void arm_timer(struct k_itimer *timer)
 {
 	struct cpu_timer_list *const nt = &timer->it.cpu;
 	int clkidx = CPUCLOCK_WHICH(timer->it_clock);
-	u64 *cpuexp, newexp = timer->it.cpu.expires;
 	struct task_struct *p = timer->it.cpu.task;
+	u64 newexp = timer->it.cpu.expires;
+	struct posix_cputimer_base *base;
 	struct list_head *head, *listpos;
 	struct cpu_timer_list *next;
 
-	if (CPUCLOCK_PERTHREAD(timer->it_clock)) {
-		head = p->posix_cputimers.cpu_timers + clkidx;
-		cpuexp = p->posix_cputimers.expiries + clkidx;
-	} else {
-		head = p->signal->posix_cputimers.cpu_timers + clkidx;
-		cpuexp = p->signal->posix_cputimers.expiries + clkidx;
-	}
+	if (CPUCLOCK_PERTHREAD(timer->it_clock))
+		base = p->posix_cputimers.bases + clkidx;
+	else
+		base = p->signal->posix_cputimers.bases + clkidx;
 
-	listpos = head;
-	list_for_each_entry(next, head, entry) {
+	listpos = head = &base->cpu_timers;
+	list_for_each_entry(next,head, entry) {
 		if (nt->expires < next->expires)
 			break;
 		listpos = &next->entry;
@@ -494,8 +494,8 @@ static void arm_timer(struct k_itimer *timer)
 	 * for process timers we share expiration cache with itimers
 	 * and RLIMIT_CPU and for thread timers with RLIMIT_RTTIME.
 	 */
-	if (expires_gt(*cpuexp, newexp))
-		*cpuexp = newexp;
+	if (expires_gt(base->nextevt, newexp))
+		base->nextevt = newexp;
 
 	if (CPUCLOCK_PERTHREAD(timer->it_clock))
 		tick_dep_set_task(p, TICK_DEP_BIT_POSIX_TIMER);
@@ -783,9 +783,9 @@ static inline void check_dl_overrun(struct task_struct *tsk)
 static void check_thread_timers(struct task_struct *tsk,
 				struct list_head *firing)
 {
-	struct list_head *timers = tsk->posix_cputimers.cpu_timers;
-	u64 stime, utime, *expires = tsk->posix_cputimers.expiries;
+	struct posix_cputimer_base *base = tsk->posix_cputimers.bases;
 	unsigned long soft;
+	u64 stime, utime;
 
 	if (dl_task(tsk))
 		check_dl_overrun(tsk);
@@ -794,14 +794,18 @@ static void check_thread_timers(struct task_struct *tsk,
 	 * If the expiry cache is zero, then there are no active per thread
 	 * CPU timers.
 	 */
-	if (expiry_cache_is_zero(tsk->posix_cputimers.expiries))
+	if (expiry_cache_is_zero(&tsk->posix_cputimers))
 		return;
 
 	task_cputime(tsk, &utime, &stime);
 
-	*expires++ = check_timers_list(timers, firing, utime + stime);
-	*expires++ = check_timers_list(++timers, firing, utime);
-	*expires = check_timers_list(++timers, firing, tsk->se.sum_exec_runtime);
+	base->nextevt = check_timers_list(&base->cpu_timers, firing,
+					  utime + stime);
+	base++;
+	base->nextevt = check_timers_list(&base->cpu_timers, firing, utime);
+	base++;
+	base->nextevt = check_timers_list(&base->cpu_timers, firing,
+					  tsk->se.sum_exec_runtime);
 
 	/*
 	 * Check for the special case thread timers.
@@ -840,7 +844,7 @@ static void check_thread_timers(struct task_struct *tsk,
 		}
 	}
 
-	if (expiry_cache_is_zero(tsk->posix_cputimers.expiries))
+	if (expiry_cache_is_zero(&tsk->posix_cputimers))
 		tick_dep_clear_task(tsk, TICK_DEP_BIT_POSIX_TIMER);
 }
 
@@ -884,7 +888,7 @@ static void check_process_timers(struct task_struct *tsk,
 				 struct list_head *firing)
 {
 	struct signal_struct *const sig = tsk->signal;
-	struct list_head *timers = sig->posix_cputimers.cpu_timers;
+	struct posix_cputimer_base *base = sig->posix_cputimers.bases;
 	u64 utime, ptime, virt_expires, prof_expires;
 	u64 sum_sched_runtime, sched_expires;
 	struct task_cputime cputime;
@@ -912,9 +916,12 @@ static void check_process_timers(struct task_struct *tsk,
 	ptime = utime + cputime.stime;
 	sum_sched_runtime = cputime.sum_exec_runtime;
 
-	prof_expires = check_timers_list(timers, firing, ptime);
-	virt_expires = check_timers_list(++timers, firing, utime);
-	sched_expires = check_timers_list(++timers, firing, sum_sched_runtime);
+	prof_expires = check_timers_list(&base[CPUCLOCK_PROF].cpu_timers,
+					 firing, ptime);
+	virt_expires = check_timers_list(&base[CPUCLOCK_VIRT].cpu_timers,
+					 firing, utime);
+	sched_expires = check_timers_list(&base[CPUCLOCK_SCHED].cpu_timers,
+					  firing, sum_sched_runtime);
 
 	/*
 	 * Check for the special case process timers.
@@ -959,11 +966,11 @@ static void check_process_timers(struct task_struct *tsk,
 			prof_expires = x;
 	}
 
-	sig->posix_cputimers.expiries[CPUCLOCK_PROF] = prof_expires;
-	sig->posix_cputimers.expiries[CPUCLOCK_VIRT] = virt_expires;
-	sig->posix_cputimers.expiries[CPUCLOCK_SCHED] = sched_expires;
+	base[CPUCLOCK_PROF].nextevt = prof_expires;
+	base[CPUCLOCK_VIRT].nextevt = virt_expires;
+	base[CPUCLOCK_SCHED].nextevt = sched_expires;
 
-	if (expiry_cache_is_zero(sig->posix_cputimers.expiries))
+	if (expiry_cache_is_zero(&sig->posix_cputimers))
 		stop_process_timers(sig);
 
 	sig->cputimer.checking_timer = false;
@@ -1028,20 +1035,21 @@ unlock:
 }
 
 /**
- * task_cputimers_expired - Compare two task_cputime entities.
+ * task_cputimers_expired - Check whether posix CPU timers are expired
  *
  * @samples:	Array of current samples for the CPUCLOCK clocks
- * @expiries:	Array of expiry values for the CPUCLOCK clocks
+ * @pct:	Pointer to a posix_cputimers container
  *
- * Returns true if any mmember of @samples is greater than the corresponding
- * member of @expiries if that member is non zero. False otherwise
+ * Returns true if any member of @samples is greater than the corresponding
+ * member of @pct->bases[CLK].nextevt. False otherwise
  */
-static inline bool task_cputimers_expired(const u64 *sample, const u64 *expiries)
+static inline bool
+task_cputimers_expired(const u64 *sample, struct posix_cputimers *pct)
 {
 	int i;
 
 	for (i = 0; i < CPUCLOCK_MAX; i++) {
-		if (expiries[i] && sample[i] >= expiries[i])
+		if (pct->bases[i].nextevt && sample[i] >= pct->bases[i].nextevt)
 			return true;
 	}
 	return false;
@@ -1059,14 +1067,13 @@ static inline bool task_cputimers_expired(const u64 *sample, const u64 *expiries
  */
 static inline bool fastpath_timer_check(struct task_struct *tsk)
 {
-	u64 *expiries = tsk->posix_cputimers.expiries;
 	struct signal_struct *sig;
 
-	if (!expiry_cache_is_zero(expiries)) {
+	if (!expiry_cache_is_zero(&tsk->posix_cputimers)) {
 		u64 samples[CPUCLOCK_MAX];
 
 		task_sample_cputime(tsk, samples);
-		if (task_cputimers_expired(samples, expiries))
+		if (task_cputimers_expired(samples, &tsk->posix_cputimers))
 			return true;
 	}
 
@@ -1092,8 +1099,7 @@ static inline bool fastpath_timer_check(struct task_struct *tsk)
 		proc_sample_cputime_atomic(&sig->cputimer.cputime_atomic,
 					   samples);
 
-		if (task_cputimers_expired(samples,
-					   sig->posix_cputimers.expiries))
+		if (task_cputimers_expired(samples, &sig->posix_cputimers))
 			return true;
 	}
 
@@ -1176,11 +1182,12 @@ void run_posix_cpu_timers(void)
 void set_process_cpu_timer(struct task_struct *tsk, unsigned int clkid,
 			   u64 *newval, u64 *oldval)
 {
-	u64 now, *expiry = tsk->signal->posix_cputimers.expiries + clkid;
+	u64 now, *nextevt;
 
 	if (WARN_ON_ONCE(clkid >= CPUCLOCK_SCHED))
 		return;
 
+	nextevt = &tsk->signal->posix_cputimers.bases[clkid].nextevt;
 	now = cpu_clock_sample_group(clkid, tsk, true);
 
 	if (oldval) {
@@ -1207,8 +1214,8 @@ void set_process_cpu_timer(struct task_struct *tsk, unsigned int clkid,
 	 * Update expiration cache if this is the earliest timer. CPUCLOCK_PROF
 	 * expiry cache is also used by RLIMIT_CPU!.
 	 */
-	if (expires_gt(*expiry, *newval))
-		*expiry = *newval;
+	if (expires_gt(*nextevt, *newval))
+		*nextevt = *newval;
 
 	tick_dep_set_signal(tsk->signal, TICK_DEP_BIT_POSIX_TIMER);
 }
-- 
cgit v1.2.3


From b7be4ef1365dcb56fdffc6689e41058b23f5996d Mon Sep 17 00:00:00 2001
From: Thomas Gleixner <tglx@linutronix.de>
Date: Wed, 21 Aug 2019 21:09:16 +0200
Subject: posix-cpu-timers: Switch thread group sampling to array

That allows more simplifications in various places.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://lkml.kernel.org/r/20190821192921.988426956@linutronix.de
---
 kernel/time/posix-cpu-timers.c | 104 ++++++++++++++++++-----------------------
 1 file changed, 45 insertions(+), 59 deletions(-)

(limited to 'kernel/time/posix-cpu-timers.c')

diff --git a/kernel/time/posix-cpu-timers.c b/kernel/time/posix-cpu-timers.c
index 9ac601abc4c4..e62139a89375 100644
--- a/kernel/time/posix-cpu-timers.c
+++ b/kernel/time/posix-cpu-timers.c
@@ -225,22 +225,14 @@ retry:
 	}
 }
 
-static void update_gt_cputime(struct task_cputime_atomic *cputime_atomic, struct task_cputime *sum)
+static void update_gt_cputime(struct task_cputime_atomic *cputime_atomic,
+			      struct task_cputime *sum)
 {
 	__update_gt_cputime(&cputime_atomic->utime, sum->utime);
 	__update_gt_cputime(&cputime_atomic->stime, sum->stime);
 	__update_gt_cputime(&cputime_atomic->sum_exec_runtime, sum->sum_exec_runtime);
 }
 
-/* Sample task_cputime_atomic values in "atomic_timers", store results in "times". */
-static inline void sample_cputime_atomic(struct task_cputime *times,
-					 struct task_cputime_atomic *atomic_times)
-{
-	times->utime = atomic64_read(&atomic_times->utime);
-	times->stime = atomic64_read(&atomic_times->stime);
-	times->sum_exec_runtime = atomic64_read(&atomic_times->sum_exec_runtime);
-}
-
 /**
  * thread_group_sample_cputime - Sample cputime for a given task
  * @tsk:	Task for which cputime needs to be started
@@ -252,20 +244,19 @@ static inline void sample_cputime_atomic(struct task_cputime *times,
  *
  * Updates @times with an uptodate sample of the thread group cputimes.
  */
-void thread_group_sample_cputime(struct task_struct *tsk,
-				struct task_cputime *times)
+void thread_group_sample_cputime(struct task_struct *tsk, u64 *samples)
 {
 	struct thread_group_cputimer *cputimer = &tsk->signal->cputimer;
 
 	WARN_ON_ONCE(!cputimer->running);
 
-	sample_cputime_atomic(times, &cputimer->cputime_atomic);
+	proc_sample_cputime_atomic(&cputimer->cputime_atomic, samples);
 }
 
 /**
  * thread_group_start_cputime - Start cputime and return a sample
  * @tsk:	Task for which cputime needs to be started
- * @iimes:	Storage for time samples
+ * @samples:	Storage for time samples
  *
  * The thread group cputime accouting is avoided when there are no posix
  * CPU timers armed. Before starting a timer it's required to check whether
@@ -274,14 +265,14 @@ void thread_group_sample_cputime(struct task_struct *tsk,
  *
  * Updates @times with an uptodate sample of the thread group cputimes.
  */
-static void
-thread_group_start_cputime(struct task_struct *tsk, struct task_cputime *times)
+static void thread_group_start_cputime(struct task_struct *tsk, u64 *samples)
 {
 	struct thread_group_cputimer *cputimer = &tsk->signal->cputimer;
-	struct task_cputime sum;
 
 	/* Check if cputimer isn't running. This is accessed without locking. */
 	if (!READ_ONCE(cputimer->running)) {
+		struct task_cputime sum;
+
 		/*
 		 * The POSIX timer interface allows for absolute time expiry
 		 * values through the TIMER_ABSTIME flag, therefore we have
@@ -299,7 +290,15 @@ thread_group_start_cputime(struct task_struct *tsk, struct task_cputime *times)
 		 */
 		WRITE_ONCE(cputimer->running, true);
 	}
-	sample_cputime_atomic(times, &cputimer->cputime_atomic);
+	proc_sample_cputime_atomic(&cputimer->cputime_atomic, samples);
+}
+
+static void __thread_group_cputime(struct task_struct *tsk, u64 *samples)
+{
+	struct task_cputime ct;
+
+	thread_group_cputime(tsk, &ct);
+	store_samples(samples, ct.stime, ct.utime, ct.sum_exec_runtime);
 }
 
 /*
@@ -313,28 +312,18 @@ static u64 cpu_clock_sample_group(const clockid_t clkid, struct task_struct *p,
 				  bool start)
 {
 	struct thread_group_cputimer *cputimer = &p->signal->cputimer;
-	struct task_cputime cputime;
+	u64 samples[CPUCLOCK_MAX];
 
 	if (!READ_ONCE(cputimer->running)) {
 		if (start)
-			thread_group_start_cputime(p, &cputime);
+			thread_group_start_cputime(p, samples);
 		else
-			thread_group_cputime(p, &cputime);
+			__thread_group_cputime(p, samples);
 	} else {
-		sample_cputime_atomic(&cputime, &cputimer->cputime_atomic);
+		proc_sample_cputime_atomic(&cputimer->cputime_atomic, samples);
 	}
 
-	switch (clkid) {
-	case CPUCLOCK_PROF:
-		return cputime.utime + cputime.stime;
-	case CPUCLOCK_VIRT:
-		return cputime.utime;
-	case CPUCLOCK_SCHED:
-		return cputime.sum_exec_runtime;
-	default:
-		WARN_ON_ONCE(1);
-	}
-	return 0;
+	return samples[clkid];
 }
 
 static int posix_cpu_clock_get(const clockid_t clock, struct timespec64 *tp)
@@ -889,9 +878,7 @@ static void check_process_timers(struct task_struct *tsk,
 {
 	struct signal_struct *const sig = tsk->signal;
 	struct posix_cputimer_base *base = sig->posix_cputimers.bases;
-	u64 utime, ptime, virt_expires, prof_expires;
-	u64 sum_sched_runtime, sched_expires;
-	struct task_cputime cputime;
+	u64 virt_exp, prof_exp, sched_exp, samples[CPUCLOCK_MAX];
 	unsigned long soft;
 
 	/*
@@ -911,30 +898,29 @@ static void check_process_timers(struct task_struct *tsk,
 	 * Collect the current process totals. Group accounting is active
 	 * so the sample can be taken directly.
 	 */
-	sample_cputime_atomic(&cputime, &sig->cputimer.cputime_atomic);
-	utime = cputime.utime;
-	ptime = utime + cputime.stime;
-	sum_sched_runtime = cputime.sum_exec_runtime;
-
-	prof_expires = check_timers_list(&base[CPUCLOCK_PROF].cpu_timers,
-					 firing, ptime);
-	virt_expires = check_timers_list(&base[CPUCLOCK_VIRT].cpu_timers,
-					 firing, utime);
-	sched_expires = check_timers_list(&base[CPUCLOCK_SCHED].cpu_timers,
-					  firing, sum_sched_runtime);
+	proc_sample_cputime_atomic(&sig->cputimer.cputime_atomic, samples);
+
+	prof_exp = check_timers_list(&base[CPUCLOCK_PROF].cpu_timers,
+				     firing, samples[CPUCLOCK_PROF]);
+	virt_exp = check_timers_list(&base[CPUCLOCK_VIRT].cpu_timers,
+				     firing, samples[CPUCLOCK_VIRT]);
+	sched_exp = check_timers_list(&base[CPUCLOCK_SCHED].cpu_timers,
+				      firing, samples[CPUCLOCK_SCHED]);
 
 	/*
 	 * Check for the special case process timers.
 	 */
-	check_cpu_itimer(tsk, &sig->it[CPUCLOCK_PROF], &prof_expires, ptime,
-			 SIGPROF);
-	check_cpu_itimer(tsk, &sig->it[CPUCLOCK_VIRT], &virt_expires, utime,
-			 SIGVTALRM);
+	check_cpu_itimer(tsk, &sig->it[CPUCLOCK_PROF], &prof_exp,
+			 samples[CPUCLOCK_PROF], SIGPROF);
+	check_cpu_itimer(tsk, &sig->it[CPUCLOCK_VIRT], &virt_exp,
+			 samples[CPUCLOCK_PROF], SIGVTALRM);
+
 	soft = task_rlimit(tsk, RLIMIT_CPU);
 	if (soft != RLIM_INFINITY) {
-		unsigned long psecs = div_u64(ptime, NSEC_PER_SEC);
+		u64 softns, ptime = samples[CPUCLOCK_PROF];
 		unsigned long hard = task_rlimit_max(tsk, RLIMIT_CPU);
-		u64 x;
+		unsigned long psecs = div_u64(ptime, NSEC_PER_SEC);
+
 		if (psecs >= hard) {
 			/*
 			 * At the hard limit, we just die.
@@ -961,14 +947,14 @@ static void check_process_timers(struct task_struct *tsk,
 				sig->rlim[RLIMIT_CPU].rlim_cur = soft;
 			}
 		}
-		x = soft * NSEC_PER_SEC;
-		if (!prof_expires || x < prof_expires)
-			prof_expires = x;
+		softns = soft * NSEC_PER_SEC;
+		if (!prof_exp || softns < prof_exp)
+			prof_exp = softns;
 	}
 
-	base[CPUCLOCK_PROF].nextevt = prof_expires;
-	base[CPUCLOCK_VIRT].nextevt = virt_expires;
-	base[CPUCLOCK_SCHED].nextevt = sched_expires;
+	base[CPUCLOCK_PROF].nextevt = prof_exp;
+	base[CPUCLOCK_VIRT].nextevt = virt_exp;
+	base[CPUCLOCK_SCHED].nextevt = sched_exp;
 
 	if (expiry_cache_is_zero(&sig->posix_cputimers))
 		stop_process_timers(sig);
-- 
cgit v1.2.3


From fe0517f893d36636de20d0a809fc0c788ca0cade Mon Sep 17 00:00:00 2001
From: Thomas Gleixner <tglx@linutronix.de>
Date: Wed, 21 Aug 2019 21:09:17 +0200
Subject: posix-cpu-timers: Respect INFINITY for hard RTTIME limit

The RTIME limit expiry code does not check the hard RTTIME limit for
INFINITY, i.e. being disabled.  Add it.

While this could be considered an ABI breakage if something would depend on
this behaviour. Though it's highly unlikely to have an effect because
RLIM_INFINITY is at minimum INT_MAX and the RTTIME limit is in seconds, so
the timer would fire after ~68 years.

Adding this obvious correct limit check also allows further consolidation
of that code and is a prerequisite for cleaning up the 0 based checks and
the rlimit setter code.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://lkml.kernel.org/r/20190821192922.078293002@linutronix.de
---
 kernel/time/posix-cpu-timers.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

(limited to 'kernel/time/posix-cpu-timers.c')

diff --git a/kernel/time/posix-cpu-timers.c b/kernel/time/posix-cpu-timers.c
index e62139a89375..a738d7659915 100644
--- a/kernel/time/posix-cpu-timers.c
+++ b/kernel/time/posix-cpu-timers.c
@@ -921,7 +921,7 @@ static void check_process_timers(struct task_struct *tsk,
 		unsigned long hard = task_rlimit_max(tsk, RLIMIT_CPU);
 		unsigned long psecs = div_u64(ptime, NSEC_PER_SEC);
 
-		if (psecs >= hard) {
+		if (hard != RLIM_INFINITY && psecs >= hard) {
 			/*
 			 * At the hard limit, we just die.
 			 * No need to calculate anything else now.
-- 
cgit v1.2.3


From 2bbdbdae05167c688b6d3499a7dab74208b80a22 Mon Sep 17 00:00:00 2001
From: Thomas Gleixner <tglx@linutronix.de>
Date: Wed, 21 Aug 2019 21:09:19 +0200
Subject: posix-cpu-timers: Get rid of zero checks

Deactivation of the expiry cache is done by setting all clock caches to
0. That requires to have a check for zero in all places which update the
expiry cache:

	if (cache == 0 || new < cache)
		cache = new;

Use U64_MAX as the deactivated value, which allows to remove the zero
checks when updating the cache and reduces it to the obvious check:

	if (new < cache)
		cache = new;

This also removes the weird workaround in do_prlimit() which was required
to convert a RLIMIT_CPU value of 0 (immediate expiry) to 1 because handing
in 0 to the posix CPU timer code would have effectively disarmed it.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://lkml.kernel.org/r/20190821192922.275086128@linutronix.de
---
 kernel/time/posix-cpu-timers.c | 38 +++++++++++++++-----------------------
 1 file changed, 15 insertions(+), 23 deletions(-)

(limited to 'kernel/time/posix-cpu-timers.c')

diff --git a/kernel/time/posix-cpu-timers.c b/kernel/time/posix-cpu-timers.c
index a738d7659915..cf85292575c5 100644
--- a/kernel/time/posix-cpu-timers.c
+++ b/kernel/time/posix-cpu-timers.c
@@ -122,11 +122,12 @@ static void bump_cpu_timer(struct k_itimer *timer, u64 now)
 	}
 }
 
-static inline bool expiry_cache_is_zero(const struct posix_cputimers *pct)
+/* Check whether all cache entries contain U64_MAX, i.e. eternal expiry time */
+static inline bool expiry_cache_is_inactive(const struct posix_cputimers *pct)
 {
-	return !(pct->bases[CPUCLOCK_PROF].nextevt |
-		 pct->bases[CPUCLOCK_VIRT].nextevt |
-		 pct->bases[CPUCLOCK_SCHED].nextevt);
+	return !(~pct->bases[CPUCLOCK_PROF].nextevt |
+		 ~pct->bases[CPUCLOCK_VIRT].nextevt |
+		 ~pct->bases[CPUCLOCK_SCHED].nextevt);
 }
 
 static int
@@ -442,11 +443,6 @@ void posix_cpu_timers_exit_group(struct task_struct *tsk)
 	cleanup_timers(&tsk->signal->posix_cputimers);
 }
 
-static inline int expires_gt(u64 expires, u64 new_exp)
-{
-	return expires == 0 || expires > new_exp;
-}
-
 /*
  * Insert the timer on the appropriate list before any timers that
  * expire later.  This must be called with the sighand lock held.
@@ -483,7 +479,7 @@ static void arm_timer(struct k_itimer *timer)
 	 * for process timers we share expiration cache with itimers
 	 * and RLIMIT_CPU and for thread timers with RLIMIT_RTTIME.
 	 */
-	if (expires_gt(base->nextevt, newexp))
+	if (newexp < base->nextevt)
 		base->nextevt = newexp;
 
 	if (CPUCLOCK_PERTHREAD(timer->it_clock))
@@ -753,7 +749,7 @@ check_timers_list(struct list_head *timers,
 		list_move_tail(&t->entry, firing);
 	}
 
-	return 0;
+	return U64_MAX;
 }
 
 static inline void check_dl_overrun(struct task_struct *tsk)
@@ -779,11 +775,7 @@ static void check_thread_timers(struct task_struct *tsk,
 	if (dl_task(tsk))
 		check_dl_overrun(tsk);
 
-	/*
-	 * If the expiry cache is zero, then there are no active per thread
-	 * CPU timers.
-	 */
-	if (expiry_cache_is_zero(&tsk->posix_cputimers))
+	if (expiry_cache_is_inactive(&tsk->posix_cputimers))
 		return;
 
 	task_cputime(tsk, &utime, &stime);
@@ -833,7 +825,7 @@ static void check_thread_timers(struct task_struct *tsk,
 		}
 	}
 
-	if (expiry_cache_is_zero(&tsk->posix_cputimers))
+	if (expiry_cache_is_inactive(&tsk->posix_cputimers))
 		tick_dep_clear_task(tsk, TICK_DEP_BIT_POSIX_TIMER);
 }
 
@@ -864,7 +856,7 @@ static void check_cpu_itimer(struct task_struct *tsk, struct cpu_itimer *it,
 		__group_send_sig_info(signo, SEND_SIG_PRIV, tsk);
 	}
 
-	if (it->expires && (!*expires || it->expires < *expires))
+	if (it->expires && it->expires < *expires)
 		*expires = it->expires;
 }
 
@@ -948,7 +940,7 @@ static void check_process_timers(struct task_struct *tsk,
 			}
 		}
 		softns = soft * NSEC_PER_SEC;
-		if (!prof_exp || softns < prof_exp)
+		if (softns < prof_exp)
 			prof_exp = softns;
 	}
 
@@ -956,7 +948,7 @@ static void check_process_timers(struct task_struct *tsk,
 	base[CPUCLOCK_VIRT].nextevt = virt_exp;
 	base[CPUCLOCK_SCHED].nextevt = sched_exp;
 
-	if (expiry_cache_is_zero(&sig->posix_cputimers))
+	if (expiry_cache_is_inactive(&sig->posix_cputimers))
 		stop_process_timers(sig);
 
 	sig->cputimer.checking_timer = false;
@@ -1035,7 +1027,7 @@ task_cputimers_expired(const u64 *sample, struct posix_cputimers *pct)
 	int i;
 
 	for (i = 0; i < CPUCLOCK_MAX; i++) {
-		if (pct->bases[i].nextevt && sample[i] >= pct->bases[i].nextevt)
+		if (sample[i] >= pct->bases[i].nextevt)
 			return true;
 	}
 	return false;
@@ -1055,7 +1047,7 @@ static inline bool fastpath_timer_check(struct task_struct *tsk)
 {
 	struct signal_struct *sig;
 
-	if (!expiry_cache_is_zero(&tsk->posix_cputimers)) {
+	if (!expiry_cache_is_inactive(&tsk->posix_cputimers)) {
 		u64 samples[CPUCLOCK_MAX];
 
 		task_sample_cputime(tsk, samples);
@@ -1200,7 +1192,7 @@ void set_process_cpu_timer(struct task_struct *tsk, unsigned int clkid,
 	 * Update expiration cache if this is the earliest timer. CPUCLOCK_PROF
 	 * expiry cache is also used by RLIMIT_CPU!.
 	 */
-	if (expires_gt(*nextevt, *newval))
+	if (*newval < *nextevt)
 		*nextevt = *newval;
 
 	tick_dep_set_signal(tsk->signal, TICK_DEP_BIT_POSIX_TIMER);
-- 
cgit v1.2.3


From 1cd07c0b94f2c320270d76edb7dd49bceb09c1df Mon Sep 17 00:00:00 2001
From: Thomas Gleixner <tglx@linutronix.de>
Date: Wed, 21 Aug 2019 21:09:20 +0200
Subject: posix-cpu-timers: Consolidate timer expiry further

With the array based samples and expiry cache, the expiry function can use
a loop to collect timers from the clock specific lists.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://lkml.kernel.org/r/20190821192922.365469982@linutronix.de
---
 kernel/time/posix-cpu-timers.c | 63 ++++++++++++++++++++----------------------
 1 file changed, 30 insertions(+), 33 deletions(-)

(limited to 'kernel/time/posix-cpu-timers.c')

diff --git a/kernel/time/posix-cpu-timers.c b/kernel/time/posix-cpu-timers.c
index cf85292575c5..caafdfdd6f0f 100644
--- a/kernel/time/posix-cpu-timers.c
+++ b/kernel/time/posix-cpu-timers.c
@@ -752,6 +752,18 @@ check_timers_list(struct list_head *timers,
 	return U64_MAX;
 }
 
+static void collect_posix_cputimers(struct posix_cputimers *pct,
+				    u64 *samples, struct list_head *firing)
+{
+	struct posix_cputimer_base *base = pct->bases;
+	int i;
+
+	for (i = 0; i < CPUCLOCK_MAX; i++, base++) {
+		base->nextevt = check_timers_list(&base->cpu_timers, firing,
+						   samples[i]);
+	}
+}
+
 static inline void check_dl_overrun(struct task_struct *tsk)
 {
 	if (tsk->dl.dl_overrun) {
@@ -768,25 +780,18 @@ static inline void check_dl_overrun(struct task_struct *tsk)
 static void check_thread_timers(struct task_struct *tsk,
 				struct list_head *firing)
 {
-	struct posix_cputimer_base *base = tsk->posix_cputimers.bases;
+	struct posix_cputimers *pct = &tsk->posix_cputimers;
+	u64 samples[CPUCLOCK_MAX];
 	unsigned long soft;
-	u64 stime, utime;
 
 	if (dl_task(tsk))
 		check_dl_overrun(tsk);
 
-	if (expiry_cache_is_inactive(&tsk->posix_cputimers))
+	if (expiry_cache_is_inactive(pct))
 		return;
 
-	task_cputime(tsk, &utime, &stime);
-
-	base->nextevt = check_timers_list(&base->cpu_timers, firing,
-					  utime + stime);
-	base++;
-	base->nextevt = check_timers_list(&base->cpu_timers, firing, utime);
-	base++;
-	base->nextevt = check_timers_list(&base->cpu_timers, firing,
-					  tsk->se.sum_exec_runtime);
+	task_sample_cputime(tsk, samples);
+	collect_posix_cputimers(pct, samples, firing);
 
 	/*
 	 * Check for the special case thread timers.
@@ -825,7 +830,7 @@ static void check_thread_timers(struct task_struct *tsk,
 		}
 	}
 
-	if (expiry_cache_is_inactive(&tsk->posix_cputimers))
+	if (expiry_cache_is_inactive(pct))
 		tick_dep_clear_task(tsk, TICK_DEP_BIT_POSIX_TIMER);
 }
 
@@ -869,15 +874,15 @@ static void check_process_timers(struct task_struct *tsk,
 				 struct list_head *firing)
 {
 	struct signal_struct *const sig = tsk->signal;
-	struct posix_cputimer_base *base = sig->posix_cputimers.bases;
-	u64 virt_exp, prof_exp, sched_exp, samples[CPUCLOCK_MAX];
+	struct posix_cputimers *pct = &sig->posix_cputimers;
+	u64 samples[CPUCLOCK_MAX];
 	unsigned long soft;
 
 	/*
 	 * If cputimer is not running, then there are no active
 	 * process wide timers (POSIX 1.b, itimers, RLIMIT_CPU).
 	 */
-	if (!READ_ONCE(tsk->signal->cputimer.running))
+	if (!READ_ONCE(sig->cputimer.running))
 		return;
 
        /*
@@ -891,21 +896,17 @@ static void check_process_timers(struct task_struct *tsk,
 	 * so the sample can be taken directly.
 	 */
 	proc_sample_cputime_atomic(&sig->cputimer.cputime_atomic, samples);
-
-	prof_exp = check_timers_list(&base[CPUCLOCK_PROF].cpu_timers,
-				     firing, samples[CPUCLOCK_PROF]);
-	virt_exp = check_timers_list(&base[CPUCLOCK_VIRT].cpu_timers,
-				     firing, samples[CPUCLOCK_VIRT]);
-	sched_exp = check_timers_list(&base[CPUCLOCK_SCHED].cpu_timers,
-				      firing, samples[CPUCLOCK_SCHED]);
+	collect_posix_cputimers(pct, samples, firing);
 
 	/*
 	 * Check for the special case process timers.
 	 */
-	check_cpu_itimer(tsk, &sig->it[CPUCLOCK_PROF], &prof_exp,
+	check_cpu_itimer(tsk, &sig->it[CPUCLOCK_PROF],
+			 &pct->bases[CPUCLOCK_PROF].nextevt,
 			 samples[CPUCLOCK_PROF], SIGPROF);
-	check_cpu_itimer(tsk, &sig->it[CPUCLOCK_VIRT], &virt_exp,
-			 samples[CPUCLOCK_PROF], SIGVTALRM);
+	check_cpu_itimer(tsk, &sig->it[CPUCLOCK_VIRT],
+			 &pct->bases[CPUCLOCK_VIRT].nextevt,
+			 samples[CPUCLOCK_VIRT], SIGVTALRM);
 
 	soft = task_rlimit(tsk, RLIMIT_CPU);
 	if (soft != RLIM_INFINITY) {
@@ -940,15 +941,11 @@ static void check_process_timers(struct task_struct *tsk,
 			}
 		}
 		softns = soft * NSEC_PER_SEC;
-		if (softns < prof_exp)
-			prof_exp = softns;
+		if (softns < pct->bases[CPUCLOCK_PROF].nextevt)
+			pct->bases[CPUCLOCK_PROF].nextevt = softns;
 	}
 
-	base[CPUCLOCK_PROF].nextevt = prof_exp;
-	base[CPUCLOCK_VIRT].nextevt = virt_exp;
-	base[CPUCLOCK_SCHED].nextevt = sched_exp;
-
-	if (expiry_cache_is_inactive(&sig->posix_cputimers))
+	if (expiry_cache_is_inactive(pct))
 		stop_process_timers(sig);
 
 	sig->cputimer.checking_timer = false;
-- 
cgit v1.2.3


From 8ea1de90a5eccdc18c8f05f8596bae8660a3ff9a Mon Sep 17 00:00:00 2001
From: Thomas Gleixner <tglx@linutronix.de>
Date: Wed, 21 Aug 2019 21:09:21 +0200
Subject: posix-cpu-timers: Get rid of 64bit divisions

Instead of dividing A to match the units of B it's more efficient to
multiply B to match the units of A.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://lkml.kernel.org/r/20190821192922.458286860@linutronix.de
---
 kernel/time/posix-cpu-timers.c | 24 ++++++++++++++----------
 1 file changed, 14 insertions(+), 10 deletions(-)

(limited to 'kernel/time/posix-cpu-timers.c')

diff --git a/kernel/time/posix-cpu-timers.c b/kernel/time/posix-cpu-timers.c
index caafdfdd6f0f..dcdf9c8241b1 100644
--- a/kernel/time/posix-cpu-timers.c
+++ b/kernel/time/posix-cpu-timers.c
@@ -798,10 +798,11 @@ static void check_thread_timers(struct task_struct *tsk,
 	 */
 	soft = task_rlimit(tsk, RLIMIT_RTTIME);
 	if (soft != RLIM_INFINITY) {
+		/* Task RT timeout is accounted in jiffies. RTTIME is usec */
+		unsigned long rtim = tsk->rt.timeout * (USEC_PER_SEC / HZ);
 		unsigned long hard = task_rlimit_max(tsk, RLIMIT_RTTIME);
 
-		if (hard != RLIM_INFINITY &&
-		    tsk->rt.timeout > DIV_ROUND_UP(hard, USEC_PER_SEC/HZ)) {
+		if (hard != RLIM_INFINITY && rtim >= hard) {
 			/*
 			 * At the hard limit, we just die.
 			 * No need to calculate anything else now.
@@ -813,7 +814,7 @@ static void check_thread_timers(struct task_struct *tsk,
 			__group_send_sig_info(SIGKILL, SEND_SIG_PRIV, tsk);
 			return;
 		}
-		if (tsk->rt.timeout > DIV_ROUND_UP(soft, USEC_PER_SEC/HZ)) {
+		if (rtim >= soft) {
 			/*
 			 * At the soft limit, send a SIGXCPU every second.
 			 */
@@ -910,11 +911,13 @@ static void check_process_timers(struct task_struct *tsk,
 
 	soft = task_rlimit(tsk, RLIMIT_CPU);
 	if (soft != RLIM_INFINITY) {
-		u64 softns, ptime = samples[CPUCLOCK_PROF];
+		/* RLIMIT_CPU is in seconds. Samples are nanoseconds */
 		unsigned long hard = task_rlimit_max(tsk, RLIMIT_CPU);
-		unsigned long psecs = div_u64(ptime, NSEC_PER_SEC);
+		u64 ptime = samples[CPUCLOCK_PROF];
+		u64 softns = (u64)soft * NSEC_PER_SEC;
+		u64 hardns = (u64)hard * NSEC_PER_SEC;
 
-		if (hard != RLIM_INFINITY && psecs >= hard) {
+		if (hard != RLIM_INFINITY && ptime >= hardns) {
 			/*
 			 * At the hard limit, we just die.
 			 * No need to calculate anything else now.
@@ -926,7 +929,7 @@ static void check_process_timers(struct task_struct *tsk,
 			__group_send_sig_info(SIGKILL, SEND_SIG_PRIV, tsk);
 			return;
 		}
-		if (psecs >= soft) {
+		if (ptime >= softns) {
 			/*
 			 * At the soft limit, send a SIGXCPU every second.
 			 */
@@ -936,11 +939,12 @@ static void check_process_timers(struct task_struct *tsk,
 			}
 			__group_send_sig_info(SIGXCPU, SEND_SIG_PRIV, tsk);
 			if (soft < hard) {
-				soft++;
-				sig->rlim[RLIMIT_CPU].rlim_cur = soft;
+				sig->rlim[RLIMIT_CPU].rlim_cur = soft + 1;
+				softns += NSEC_PER_SEC;
 			}
 		}
-		softns = soft * NSEC_PER_SEC;
+
+		/* Update the expiry cache */
 		if (softns < pct->bases[CPUCLOCK_PROF].nextevt)
 			pct->bases[CPUCLOCK_PROF].nextevt = softns;
 	}
-- 
cgit v1.2.3


From dd6702241337bcd0bae01d2644b7bae1a496d937 Mon Sep 17 00:00:00 2001
From: Thomas Gleixner <tglx@linutronix.de>
Date: Wed, 21 Aug 2019 21:09:22 +0200
Subject: posix-cpu-timers: Remove pointless comparisons

The soft RLIMIT expiry code checks whether the soft limit is greater than
the hard limit. That's pointless because if the soft RLIMIT is greater than
the hard RLIMIT then that code cannot be reached as the hard RLIMIT check
is before that and already killed the process.

Remove it.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://lkml.kernel.org/r/20190821192922.548747613@linutronix.de
---
 kernel/time/posix-cpu-timers.c | 16 +++++++---------
 1 file changed, 7 insertions(+), 9 deletions(-)

(limited to 'kernel/time/posix-cpu-timers.c')

diff --git a/kernel/time/posix-cpu-timers.c b/kernel/time/posix-cpu-timers.c
index dcdf9c8241b1..115c8dfa4d46 100644
--- a/kernel/time/posix-cpu-timers.c
+++ b/kernel/time/posix-cpu-timers.c
@@ -814,15 +814,14 @@ static void check_thread_timers(struct task_struct *tsk,
 			__group_send_sig_info(SIGKILL, SEND_SIG_PRIV, tsk);
 			return;
 		}
+
 		if (rtim >= soft) {
 			/*
 			 * At the soft limit, send a SIGXCPU every second.
 			 */
-			if (soft < hard) {
-				soft += USEC_PER_SEC;
-				tsk->signal->rlim[RLIMIT_RTTIME].rlim_cur =
-					soft;
-			}
+			soft += USEC_PER_SEC;
+			tsk->signal->rlim[RLIMIT_RTTIME].rlim_cur = soft;
+
 			if (print_fatal_signals) {
 				pr_info("RT Watchdog Timeout (soft): %s[%d]\n",
 					tsk->comm, task_pid_nr(tsk));
@@ -938,10 +937,9 @@ static void check_process_timers(struct task_struct *tsk,
 					tsk->comm, task_pid_nr(tsk));
 			}
 			__group_send_sig_info(SIGXCPU, SEND_SIG_PRIV, tsk);
-			if (soft < hard) {
-				sig->rlim[RLIMIT_CPU].rlim_cur = soft + 1;
-				softns += NSEC_PER_SEC;
-			}
+
+			sig->rlim[RLIMIT_CPU].rlim_cur = soft + 1;
+			softns += NSEC_PER_SEC;
 		}
 
 		/* Update the expiry cache */
-- 
cgit v1.2.3


From 8991afe2640d05a805eba01277856e8549cdc838 Mon Sep 17 00:00:00 2001
From: Thomas Gleixner <tglx@linutronix.de>
Date: Wed, 21 Aug 2019 21:09:23 +0200
Subject: posix-cpu-timers: Deduplicate rlimit handling

Both thread and process expiry functions have the same functionality for
sending signals for soft and hard RLIMITs duplicated in 4 different
ways.

Split it out into a common function and cleanup the callsites.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://lkml.kernel.org/r/20190821192922.653276779@linutronix.de
---
 kernel/time/posix-cpu-timers.c | 67 ++++++++++++++++--------------------------
 1 file changed, 25 insertions(+), 42 deletions(-)

(limited to 'kernel/time/posix-cpu-timers.c')

diff --git a/kernel/time/posix-cpu-timers.c b/kernel/time/posix-cpu-timers.c
index 115c8dfa4d46..ef39a7a4a95c 100644
--- a/kernel/time/posix-cpu-timers.c
+++ b/kernel/time/posix-cpu-timers.c
@@ -772,6 +772,20 @@ static inline void check_dl_overrun(struct task_struct *tsk)
 	}
 }
 
+static bool check_rlimit(u64 time, u64 limit, int signo, bool rt, bool hard)
+{
+	if (time < limit)
+		return false;
+
+	if (print_fatal_signals) {
+		pr_info("%s Watchdog Timeout (%s): %s[%d]\n",
+			rt ? "RT" : "CPU", hard ? "hard" : "soft",
+			current->comm, task_pid_nr(current));
+	}
+	__group_send_sig_info(signo, SEND_SIG_PRIV, current);
+	return true;
+}
+
 /*
  * Check for any per-thread CPU timers that have fired and move them off
  * the tsk->cpu_timers[N] list onto the firing list.  Here we update the
@@ -799,34 +813,18 @@ static void check_thread_timers(struct task_struct *tsk,
 	soft = task_rlimit(tsk, RLIMIT_RTTIME);
 	if (soft != RLIM_INFINITY) {
 		/* Task RT timeout is accounted in jiffies. RTTIME is usec */
-		unsigned long rtim = tsk->rt.timeout * (USEC_PER_SEC / HZ);
+		unsigned long rttime = tsk->rt.timeout * (USEC_PER_SEC / HZ);
 		unsigned long hard = task_rlimit_max(tsk, RLIMIT_RTTIME);
 
-		if (hard != RLIM_INFINITY && rtim >= hard) {
-			/*
-			 * At the hard limit, we just die.
-			 * No need to calculate anything else now.
-			 */
-			if (print_fatal_signals) {
-				pr_info("CPU Watchdog Timeout (hard): %s[%d]\n",
-					tsk->comm, task_pid_nr(tsk));
-			}
-			__group_send_sig_info(SIGKILL, SEND_SIG_PRIV, tsk);
+		/* At the hard limit, send SIGKILL. No further action. */
+		if (hard != RLIM_INFINITY &&
+		    check_rlimit(rttime, hard, SIGKILL, true, true))
 			return;
-		}
 
-		if (rtim >= soft) {
-			/*
-			 * At the soft limit, send a SIGXCPU every second.
-			 */
+		/* At the soft limit, send a SIGXCPU every second */
+		if (check_rlimit(rttime, soft, SIGXCPU, true, false)) {
 			soft += USEC_PER_SEC;
 			tsk->signal->rlim[RLIMIT_RTTIME].rlim_cur = soft;
-
-			if (print_fatal_signals) {
-				pr_info("RT Watchdog Timeout (soft): %s[%d]\n",
-					tsk->comm, task_pid_nr(tsk));
-			}
-			__group_send_sig_info(SIGXCPU, SEND_SIG_PRIV, tsk);
 		}
 	}
 
@@ -916,28 +914,13 @@ static void check_process_timers(struct task_struct *tsk,
 		u64 softns = (u64)soft * NSEC_PER_SEC;
 		u64 hardns = (u64)hard * NSEC_PER_SEC;
 
-		if (hard != RLIM_INFINITY && ptime >= hardns) {
-			/*
-			 * At the hard limit, we just die.
-			 * No need to calculate anything else now.
-			 */
-			if (print_fatal_signals) {
-				pr_info("RT Watchdog Timeout (hard): %s[%d]\n",
-					tsk->comm, task_pid_nr(tsk));
-			}
-			__group_send_sig_info(SIGKILL, SEND_SIG_PRIV, tsk);
+		/* At the hard limit, send SIGKILL. No further action. */
+		if (hard != RLIM_INFINITY &&
+		    check_rlimit(ptime, hardns, SIGKILL, false, true))
 			return;
-		}
-		if (ptime >= softns) {
-			/*
-			 * At the soft limit, send a SIGXCPU every second.
-			 */
-			if (print_fatal_signals) {
-				pr_info("CPU Watchdog Timeout (soft): %s[%d]\n",
-					tsk->comm, task_pid_nr(tsk));
-			}
-			__group_send_sig_info(SIGXCPU, SEND_SIG_PRIV, tsk);
 
+		/* At the soft limit, send a SIGXCPU every second */
+		if (check_rlimit(ptime, softns, SIGXCPU, false, false)) {
 			sig->rlim[RLIMIT_CPU].rlim_cur = soft + 1;
 			softns += NSEC_PER_SEC;
 		}
-- 
cgit v1.2.3


From 244d49e30653658d4e7e9b2b8427777cbbc5affe Mon Sep 17 00:00:00 2001
From: Thomas Gleixner <tglx@linutronix.de>
Date: Wed, 21 Aug 2019 21:09:24 +0200
Subject: posix-cpu-timers: Move state tracking to struct posix_cputimers

Put it where it belongs and clean up the ifdeffery in fork completely.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lkml.kernel.org/r/20190821192922.743229404@linutronix.de
---
 kernel/time/posix-cpu-timers.c | 73 +++++++++++++++++++++++-------------------
 1 file changed, 40 insertions(+), 33 deletions(-)

(limited to 'kernel/time/posix-cpu-timers.c')

diff --git a/kernel/time/posix-cpu-timers.c b/kernel/time/posix-cpu-timers.c
index ef39a7a4a95c..52f4c99c1d60 100644
--- a/kernel/time/posix-cpu-timers.c
+++ b/kernel/time/posix-cpu-timers.c
@@ -23,8 +23,10 @@ static void posix_cpu_timer_rearm(struct k_itimer *timer);
 void posix_cputimers_group_init(struct posix_cputimers *pct, u64 cpu_limit)
 {
 	posix_cputimers_init(pct);
-	if (cpu_limit != RLIM_INFINITY)
+	if (cpu_limit != RLIM_INFINITY) {
 		pct->bases[CPUCLOCK_PROF].nextevt = cpu_limit * NSEC_PER_SEC;
+		pct->timers_active = true;
+	}
 }
 
 /*
@@ -248,8 +250,9 @@ static void update_gt_cputime(struct task_cputime_atomic *cputime_atomic,
 void thread_group_sample_cputime(struct task_struct *tsk, u64 *samples)
 {
 	struct thread_group_cputimer *cputimer = &tsk->signal->cputimer;
+	struct posix_cputimers *pct = &tsk->signal->posix_cputimers;
 
-	WARN_ON_ONCE(!cputimer->running);
+	WARN_ON_ONCE(!pct->timers_active);
 
 	proc_sample_cputime_atomic(&cputimer->cputime_atomic, samples);
 }
@@ -269,9 +272,10 @@ void thread_group_sample_cputime(struct task_struct *tsk, u64 *samples)
 static void thread_group_start_cputime(struct task_struct *tsk, u64 *samples)
 {
 	struct thread_group_cputimer *cputimer = &tsk->signal->cputimer;
+	struct posix_cputimers *pct = &tsk->signal->posix_cputimers;
 
 	/* Check if cputimer isn't running. This is accessed without locking. */
-	if (!READ_ONCE(cputimer->running)) {
+	if (!READ_ONCE(pct->timers_active)) {
 		struct task_cputime sum;
 
 		/*
@@ -283,13 +287,13 @@ static void thread_group_start_cputime(struct task_struct *tsk, u64 *samples)
 		update_gt_cputime(&cputimer->cputime_atomic, &sum);
 
 		/*
-		 * We're setting cputimer->running without a lock. Ensure
-		 * this only gets written to in one operation. We set
-		 * running after update_gt_cputime() as a small optimization,
-		 * but barriers are not required because update_gt_cputime()
+		 * We're setting timers_active without a lock. Ensure this
+		 * only gets written to in one operation. We set it after
+		 * update_gt_cputime() as a small optimization, but
+		 * barriers are not required because update_gt_cputime()
 		 * can handle concurrent updates.
 		 */
-		WRITE_ONCE(cputimer->running, true);
+		WRITE_ONCE(pct->timers_active, true);
 	}
 	proc_sample_cputime_atomic(&cputimer->cputime_atomic, samples);
 }
@@ -313,9 +317,10 @@ static u64 cpu_clock_sample_group(const clockid_t clkid, struct task_struct *p,
 				  bool start)
 {
 	struct thread_group_cputimer *cputimer = &p->signal->cputimer;
+	struct posix_cputimers *pct = &p->signal->posix_cputimers;
 	u64 samples[CPUCLOCK_MAX];
 
-	if (!READ_ONCE(cputimer->running)) {
+	if (!READ_ONCE(pct->timers_active)) {
 		if (start)
 			thread_group_start_cputime(p, samples);
 		else
@@ -834,10 +839,10 @@ static void check_thread_timers(struct task_struct *tsk,
 
 static inline void stop_process_timers(struct signal_struct *sig)
 {
-	struct thread_group_cputimer *cputimer = &sig->cputimer;
+	struct posix_cputimers *pct = &sig->posix_cputimers;
 
-	/* Turn off cputimer->running. This is done without locking. */
-	WRITE_ONCE(cputimer->running, false);
+	/* Turn off the active flag. This is done without locking. */
+	WRITE_ONCE(pct->timers_active, false);
 	tick_dep_clear_signal(sig, TICK_DEP_BIT_POSIX_TIMER);
 }
 
@@ -877,17 +882,17 @@ static void check_process_timers(struct task_struct *tsk,
 	unsigned long soft;
 
 	/*
-	 * If cputimer is not running, then there are no active
-	 * process wide timers (POSIX 1.b, itimers, RLIMIT_CPU).
+	 * If there are no active process wide timers (POSIX 1.b, itimers,
+	 * RLIMIT_CPU) nothing to check.
 	 */
-	if (!READ_ONCE(sig->cputimer.running))
+	if (!READ_ONCE(pct->timers_active))
 		return;
 
        /*
 	 * Signify that a thread is checking for process timers.
 	 * Write access to this field is protected by the sighand lock.
 	 */
-	sig->cputimer.checking_timer = true;
+	pct->timers_active = true;
 
 	/*
 	 * Collect the current process totals. Group accounting is active
@@ -933,7 +938,7 @@ static void check_process_timers(struct task_struct *tsk,
 	if (expiry_cache_is_inactive(pct))
 		stop_process_timers(sig);
 
-	sig->cputimer.checking_timer = false;
+	pct->expiry_active = false;
 }
 
 /*
@@ -1027,39 +1032,41 @@ task_cputimers_expired(const u64 *sample, struct posix_cputimers *pct)
  */
 static inline bool fastpath_timer_check(struct task_struct *tsk)
 {
+	struct posix_cputimers *pct = &tsk->posix_cputimers;
 	struct signal_struct *sig;
 
-	if (!expiry_cache_is_inactive(&tsk->posix_cputimers)) {
+	if (!expiry_cache_is_inactive(pct)) {
 		u64 samples[CPUCLOCK_MAX];
 
 		task_sample_cputime(tsk, samples);
-		if (task_cputimers_expired(samples, &tsk->posix_cputimers))
+		if (task_cputimers_expired(samples, pct))
 			return true;
 	}
 
 	sig = tsk->signal;
+	pct = &sig->posix_cputimers;
 	/*
-	 * Check if thread group timers expired when the cputimer is
-	 * running and no other thread in the group is already checking
-	 * for thread group cputimers. These fields are read without the
-	 * sighand lock. However, this is fine because this is meant to
-	 * be a fastpath heuristic to determine whether we should try to
-	 * acquire the sighand lock to check/handle timers.
+	 * Check if thread group timers expired when timers are active and
+	 * no other thread in the group is already handling expiry for
+	 * thread group cputimers. These fields are read without the
+	 * sighand lock. However, this is fine because this is meant to be
+	 * a fastpath heuristic to determine whether we should try to
+	 * acquire the sighand lock to handle timer expiry.
 	 *
-	 * In the worst case scenario, if 'running' or 'checking_timer' gets
-	 * set but the current thread doesn't see the change yet, we'll wait
-	 * until the next thread in the group gets a scheduler interrupt to
-	 * handle the timer. This isn't an issue in practice because these
-	 * types of delays with signals actually getting sent are expected.
+	 * In the worst case scenario, if concurrently timers_active is set
+	 * or expiry_active is cleared, but the current thread doesn't see
+	 * the change yet, the timer checks are delayed until the next
+	 * thread in the group gets a scheduler interrupt to handle the
+	 * timer. This isn't an issue in practice because these types of
+	 * delays with signals actually getting sent are expected.
 	 */
-	if (READ_ONCE(sig->cputimer.running) &&
-	    !READ_ONCE(sig->cputimer.checking_timer)) {
+	if (READ_ONCE(pct->timers_active) && !READ_ONCE(pct->expiry_active)) {
 		u64 samples[CPUCLOCK_MAX];
 
 		proc_sample_cputime_atomic(&sig->cputimer.cputime_atomic,
 					   samples);
 
-		if (task_cputimers_expired(samples, &sig->posix_cputimers))
+		if (task_cputimers_expired(samples, pct))
 			return true;
 	}
 
-- 
cgit v1.2.3


From 60bda037f1dd8151e0c9ee5b09f0c091a0f643cd Mon Sep 17 00:00:00 2001
From: Thomas Gleixner <tglx@linutronix.de>
Date: Tue, 27 Aug 2019 21:31:02 +0200
Subject: posix-cpu-timers: Utilize timerqueue for storage

Using a linear O(N) search for timer insertion affects execution time and
D-cache footprint badly with a larger number of timers.

Switch the storage to a timerqueue which is already used for hrtimers and
alarmtimers. It does not affect the size of struct k_itimer as it.alarm is
still larger.

The extra list head for the expiry list will go away later once the expiry
is moved into task work context.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://lkml.kernel.org/r/alpine.DEB.2.21.1908272129220.1939@nanos.tec.linutronix.de
---
 kernel/time/posix-cpu-timers.c | 189 +++++++++++++++++++++--------------------
 1 file changed, 96 insertions(+), 93 deletions(-)

(limited to 'kernel/time/posix-cpu-timers.c')

diff --git a/kernel/time/posix-cpu-timers.c b/kernel/time/posix-cpu-timers.c
index 52f4c99c1d60..73c492ce404b 100644
--- a/kernel/time/posix-cpu-timers.c
+++ b/kernel/time/posix-cpu-timers.c
@@ -96,19 +96,19 @@ static inline int validate_clock_permissions(const clockid_t clock)
  * Update expiry time from increment, and increase overrun count,
  * given the current clock sample.
  */
-static void bump_cpu_timer(struct k_itimer *timer, u64 now)
+static u64 bump_cpu_timer(struct k_itimer *timer, u64 now)
 {
+	u64 delta, incr, expires = timer->it.cpu.node.expires;
 	int i;
-	u64 delta, incr;
 
 	if (!timer->it_interval)
-		return;
+		return expires;
 
-	if (now < timer->it.cpu.expires)
-		return;
+	if (now < expires)
+		return expires;
 
 	incr = timer->it_interval;
-	delta = now + incr - timer->it.cpu.expires;
+	delta = now + incr - expires;
 
 	/* Don't use (incr*2 < delta), incr*2 might overflow. */
 	for (i = 0; incr < delta - incr; i++)
@@ -118,10 +118,11 @@ static void bump_cpu_timer(struct k_itimer *timer, u64 now)
 		if (delta < incr)
 			continue;
 
-		timer->it.cpu.expires += incr;
+		timer->it.cpu.node.expires += incr;
 		timer->it_overrun += 1LL << i;
 		delta -= incr;
 	}
+	return timer->it.cpu.node.expires;
 }
 
 /* Check whether all cache entries contain U64_MAX, i.e. eternal expiry time */
@@ -365,7 +366,7 @@ static int posix_cpu_timer_create(struct k_itimer *new_timer)
 		return -EINVAL;
 
 	new_timer->kclock = &clock_posix_cpu;
-	INIT_LIST_HEAD(&new_timer->it.cpu.entry);
+	timerqueue_init(&new_timer->it.cpu.node);
 	new_timer->it.cpu.task = p;
 	return 0;
 }
@@ -378,10 +379,11 @@ static int posix_cpu_timer_create(struct k_itimer *new_timer)
  */
 static int posix_cpu_timer_del(struct k_itimer *timer)
 {
-	int ret = 0;
-	unsigned long flags;
+	struct cpu_timer *ctmr = &timer->it.cpu;
+	struct task_struct *p = ctmr->task;
 	struct sighand_struct *sighand;
-	struct task_struct *p = timer->it.cpu.task;
+	unsigned long flags;
+	int ret = 0;
 
 	if (WARN_ON_ONCE(!p))
 		return -EINVAL;
@@ -393,15 +395,15 @@ static int posix_cpu_timer_del(struct k_itimer *timer)
 	sighand = lock_task_sighand(p, &flags);
 	if (unlikely(sighand == NULL)) {
 		/*
-		 * We raced with the reaping of the task.
-		 * The deletion should have cleared us off the list.
+		 * This raced with the reaping of the task. The exit cleanup
+		 * should have removed this timer from the timer queue.
 		 */
-		WARN_ON_ONCE(!list_empty(&timer->it.cpu.entry));
+		WARN_ON_ONCE(ctmr->head || timerqueue_node_queued(&ctmr->node));
 	} else {
 		if (timer->it.cpu.firing)
 			ret = TIMER_RETRY;
 		else
-			list_del(&timer->it.cpu.entry);
+			cpu_timer_dequeue(ctmr);
 
 		unlock_task_sighand(p, &flags);
 	}
@@ -412,12 +414,16 @@ static int posix_cpu_timer_del(struct k_itimer *timer)
 	return ret;
 }
 
-static void cleanup_timers_list(struct list_head *head)
+static void cleanup_timerqueue(struct timerqueue_head *head)
 {
-	struct cpu_timer_list *timer, *next;
+	struct timerqueue_node *node;
+	struct cpu_timer *ctmr;
 
-	list_for_each_entry_safe(timer, next, head, entry)
-		list_del_init(&timer->entry);
+	while ((node = timerqueue_getnext(head))) {
+		timerqueue_del(head, node);
+		ctmr = container_of(node, struct cpu_timer, node);
+		ctmr->head = NULL;
+	}
 }
 
 /*
@@ -429,9 +435,9 @@ static void cleanup_timers_list(struct list_head *head)
  */
 static void cleanup_timers(struct posix_cputimers *pct)
 {
-	cleanup_timers_list(&pct->bases[CPUCLOCK_PROF].cpu_timers);
-	cleanup_timers_list(&pct->bases[CPUCLOCK_VIRT].cpu_timers);
-	cleanup_timers_list(&pct->bases[CPUCLOCK_SCHED].cpu_timers);
+	cleanup_timerqueue(&pct->bases[CPUCLOCK_PROF].tqhead);
+	cleanup_timerqueue(&pct->bases[CPUCLOCK_VIRT].tqhead);
+	cleanup_timerqueue(&pct->bases[CPUCLOCK_SCHED].tqhead);
 }
 
 /*
@@ -454,28 +460,18 @@ void posix_cpu_timers_exit_group(struct task_struct *tsk)
  */
 static void arm_timer(struct k_itimer *timer)
 {
-	struct cpu_timer_list *const nt = &timer->it.cpu;
 	int clkidx = CPUCLOCK_WHICH(timer->it_clock);
-	struct task_struct *p = timer->it.cpu.task;
-	u64 newexp = timer->it.cpu.expires;
+	struct cpu_timer *ctmr = &timer->it.cpu;
+	u64 newexp = cpu_timer_getexpires(ctmr);
+	struct task_struct *p = ctmr->task;
 	struct posix_cputimer_base *base;
-	struct list_head *head, *listpos;
-	struct cpu_timer_list *next;
 
 	if (CPUCLOCK_PERTHREAD(timer->it_clock))
 		base = p->posix_cputimers.bases + clkidx;
 	else
 		base = p->signal->posix_cputimers.bases + clkidx;
 
-	listpos = head = &base->cpu_timers;
-	list_for_each_entry(next,head, entry) {
-		if (nt->expires < next->expires)
-			break;
-		listpos = &next->entry;
-	}
-	list_add(&nt->entry, listpos);
-
-	if (listpos != head)
+	if (!cpu_timer_enqueue(&base->tqhead, ctmr))
 		return;
 
 	/*
@@ -498,24 +494,26 @@ static void arm_timer(struct k_itimer *timer)
  */
 static void cpu_timer_fire(struct k_itimer *timer)
 {
+	struct cpu_timer *ctmr = &timer->it.cpu;
+
 	if ((timer->it_sigev_notify & ~SIGEV_THREAD_ID) == SIGEV_NONE) {
 		/*
 		 * User don't want any signal.
 		 */
-		timer->it.cpu.expires = 0;
+		cpu_timer_setexpires(ctmr, 0);
 	} else if (unlikely(timer->sigq == NULL)) {
 		/*
 		 * This a special case for clock_nanosleep,
 		 * not a normal timer from sys_timer_create.
 		 */
 		wake_up_process(timer->it_process);
-		timer->it.cpu.expires = 0;
+		cpu_timer_setexpires(ctmr, 0);
 	} else if (!timer->it_interval) {
 		/*
 		 * One-shot timer.  Clear it as soon as it's fired.
 		 */
 		posix_timer_event(timer, 0);
-		timer->it.cpu.expires = 0;
+		cpu_timer_setexpires(ctmr, 0);
 	} else if (posix_timer_event(timer, ++timer->it_requeue_pending)) {
 		/*
 		 * The signal did not get queued because the signal
@@ -539,10 +537,11 @@ static int posix_cpu_timer_set(struct k_itimer *timer, int timer_flags,
 {
 	clockid_t clkid = CPUCLOCK_WHICH(timer->it_clock);
 	u64 old_expires, new_expires, old_incr, val;
-	struct task_struct *p = timer->it.cpu.task;
+	struct cpu_timer *ctmr = &timer->it.cpu;
+	struct task_struct *p = ctmr->task;
 	struct sighand_struct *sighand;
 	unsigned long flags;
-	int ret;
+	int ret = 0;
 
 	if (WARN_ON_ONCE(!p))
 		return -EINVAL;
@@ -562,22 +561,21 @@ static int posix_cpu_timer_set(struct k_itimer *timer, int timer_flags,
 	 * If p has just been reaped, we can no
 	 * longer get any information about it at all.
 	 */
-	if (unlikely(sighand == NULL)) {
+	if (unlikely(sighand == NULL))
 		return -ESRCH;
-	}
 
 	/*
 	 * Disarm any old timer after extracting its expiry time.
 	 */
-
-	ret = 0;
 	old_incr = timer->it_interval;
-	old_expires = timer->it.cpu.expires;
+	old_expires = cpu_timer_getexpires(ctmr);
+
 	if (unlikely(timer->it.cpu.firing)) {
 		timer->it.cpu.firing = -1;
 		ret = TIMER_RETRY;
-	} else
-		list_del_init(&timer->it.cpu.entry);
+	} else {
+		cpu_timer_dequeue(ctmr);
+	}
 
 	/*
 	 * We need to sample the current value to convert the new
@@ -598,18 +596,16 @@ static int posix_cpu_timer_set(struct k_itimer *timer, int timer_flags,
 			old->it_value.tv_nsec = 0;
 		} else {
 			/*
-			 * Update the timer in case it has
-			 * overrun already.  If it has,
-			 * we'll report it as having overrun
-			 * and with the next reloaded timer
-			 * already ticking, though we are
-			 * swallowing that pending
-			 * notification here to install the
-			 * new setting.
+			 * Update the timer in case it has overrun already.
+			 * If it has, we'll report it as having overrun and
+			 * with the next reloaded timer already ticking,
+			 * though we are swallowing that pending
+			 * notification here to install the new setting.
 			 */
-			bump_cpu_timer(timer, val);
-			if (val < timer->it.cpu.expires) {
-				old_expires = timer->it.cpu.expires - val;
+			u64 exp = bump_cpu_timer(timer, val);
+
+			if (val < exp) {
+				old_expires = exp - val;
 				old->it_value = ns_to_timespec64(old_expires);
 			} else {
 				old->it_value.tv_nsec = 1;
@@ -638,7 +634,7 @@ static int posix_cpu_timer_set(struct k_itimer *timer, int timer_flags,
 	 * For a timer with no notification action, we don't actually
 	 * arm the timer (we'll just fake it for timer_gettime).
 	 */
-	timer->it.cpu.expires = new_expires;
+	cpu_timer_setexpires(ctmr, new_expires);
 	if (new_expires != 0 && val < new_expires) {
 		arm_timer(timer);
 	}
@@ -680,8 +676,9 @@ static int posix_cpu_timer_set(struct k_itimer *timer, int timer_flags,
 static void posix_cpu_timer_get(struct k_itimer *timer, struct itimerspec64 *itp)
 {
 	clockid_t clkid = CPUCLOCK_WHICH(timer->it_clock);
-	struct task_struct *p = timer->it.cpu.task;
-	u64 now;
+	struct cpu_timer *ctmr = &timer->it.cpu;
+	u64 now, expires = cpu_timer_getexpires(ctmr);
+	struct task_struct *p = ctmr->task;
 
 	if (WARN_ON_ONCE(!p))
 		return;
@@ -691,7 +688,7 @@ static void posix_cpu_timer_get(struct k_itimer *timer, struct itimerspec64 *itp
 	 */
 	itp->it_interval = ktime_to_timespec64(timer->it_interval);
 
-	if (!timer->it.cpu.expires)
+	if (!expires)
 		return;
 
 	/*
@@ -713,9 +710,9 @@ static void posix_cpu_timer_get(struct k_itimer *timer, struct itimerspec64 *itp
 			/*
 			 * The process has been reaped.
 			 * We can't even collect a sample any more.
-			 * Call the timer disarmed, nothing else to do.
+			 * Disarm the timer, nothing else to do.
 			 */
-			timer->it.cpu.expires = 0;
+			cpu_timer_setexpires(ctmr, 0);
 			return;
 		} else {
 			now = cpu_clock_sample_group(clkid, p, false);
@@ -723,8 +720,8 @@ static void posix_cpu_timer_get(struct k_itimer *timer, struct itimerspec64 *itp
 		}
 	}
 
-	if (now < timer->it.cpu.expires) {
-		itp->it_value = ns_to_timespec64(timer->it.cpu.expires - now);
+	if (now < expires) {
+		itp->it_value = ns_to_timespec64(expires - now);
 	} else {
 		/*
 		 * The timer should have expired already, but the firing
@@ -735,37 +732,41 @@ static void posix_cpu_timer_get(struct k_itimer *timer, struct itimerspec64 *itp
 	}
 }
 
-static unsigned long long
-check_timers_list(struct list_head *timers,
-		  struct list_head *firing,
-		  unsigned long long curr)
-{
-	int maxfire = 20;
-
-	while (!list_empty(timers)) {
-		struct cpu_timer_list *t;
-
-		t = list_first_entry(timers, struct cpu_timer_list, entry);
+#define MAX_COLLECTED	20
 
-		if (!--maxfire || curr < t->expires)
-			return t->expires;
-
-		t->firing = 1;
-		list_move_tail(&t->entry, firing);
+static u64 collect_timerqueue(struct timerqueue_head *head,
+			      struct list_head *firing, u64 now)
+{
+	struct timerqueue_node *next;
+	int i = 0;
+
+	while ((next = timerqueue_getnext(head))) {
+		struct cpu_timer *ctmr;
+		u64 expires;
+
+		ctmr = container_of(next, struct cpu_timer, node);
+		expires = cpu_timer_getexpires(ctmr);
+		/* Limit the number of timers to expire at once */
+		if (++i == MAX_COLLECTED || now < expires)
+			return expires;
+
+		ctmr->firing = 1;
+		cpu_timer_dequeue(ctmr);
+		list_add_tail(&ctmr->elist, firing);
 	}
 
 	return U64_MAX;
 }
 
-static void collect_posix_cputimers(struct posix_cputimers *pct,
-				    u64 *samples, struct list_head *firing)
+static void collect_posix_cputimers(struct posix_cputimers *pct, u64 *samples,
+				    struct list_head *firing)
 {
 	struct posix_cputimer_base *base = pct->bases;
 	int i;
 
 	for (i = 0; i < CPUCLOCK_MAX; i++, base++) {
-		base->nextevt = check_timers_list(&base->cpu_timers, firing,
-						   samples[i]);
+		base->nextevt = collect_timerqueue(&base->tqhead, firing,
+						    samples[i]);
 	}
 }
 
@@ -948,7 +949,8 @@ static void check_process_timers(struct task_struct *tsk,
 static void posix_cpu_timer_rearm(struct k_itimer *timer)
 {
 	clockid_t clkid = CPUCLOCK_WHICH(timer->it_clock);
-	struct task_struct *p = timer->it.cpu.task;
+	struct cpu_timer *ctmr = &timer->it.cpu;
+	struct task_struct *p = ctmr->task;
 	struct sighand_struct *sighand;
 	unsigned long flags;
 	u64 now;
@@ -980,7 +982,7 @@ static void posix_cpu_timer_rearm(struct k_itimer *timer)
 			 * The process has been reaped.
 			 * We can't even collect a sample any more.
 			 */
-			timer->it.cpu.expires = 0;
+			cpu_timer_setexpires(ctmr, 0);
 			return;
 		} else if (unlikely(p->exit_state) && thread_group_empty(p)) {
 			/* If the process is dying, no need to rearm */
@@ -1124,11 +1126,11 @@ void run_posix_cpu_timers(void)
 	 * each timer's lock before clearing its firing flag, so no
 	 * timer call will interfere.
 	 */
-	list_for_each_entry_safe(timer, next, &firing, it.cpu.entry) {
+	list_for_each_entry_safe(timer, next, &firing, it.cpu.elist) {
 		int cpu_firing;
 
 		spin_lock(&timer->it_lock);
-		list_del_init(&timer->it.cpu.entry);
+		list_del_init(&timer->it.cpu.elist);
 		cpu_firing = timer->it.cpu.firing;
 		timer->it.cpu.firing = 0;
 		/*
@@ -1204,6 +1206,7 @@ static int do_cpu_nanosleep(const clockid_t which_clock, int flags,
 	timer.it_overrun = -1;
 	error = posix_cpu_timer_create(&timer);
 	timer.it_process = current;
+
 	if (!error) {
 		static struct itimerspec64 zero_it;
 		struct restart_block *restart;
@@ -1219,7 +1222,7 @@ static int do_cpu_nanosleep(const clockid_t which_clock, int flags,
 		}
 
 		while (!signal_pending(current)) {
-			if (timer.it.cpu.expires == 0) {
+			if (!cpu_timer_getexpires(&timer.it.cpu)) {
 				/*
 				 * Our timer fired and was reset, below
 				 * deletion can not fail.
@@ -1241,7 +1244,7 @@ static int do_cpu_nanosleep(const clockid_t which_clock, int flags,
 		/*
 		 * We were interrupted by a signal.
 		 */
-		expires = timer.it.cpu.expires;
+		expires = cpu_timer_getexpires(&timer.it.cpu);
 		error = posix_cpu_timer_set(&timer, 0, &zero_it, &it);
 		if (!error) {
 			/*
-- 
cgit v1.2.3


From a2ed4fd685cd23e98922f933d5dbccfbe82a4f08 Mon Sep 17 00:00:00 2001
From: Thomas Gleixner <tglx@linutronix.de>
Date: Thu, 29 Aug 2019 12:52:28 +0200
Subject: posix-cpu-timers: Make expiry_active check actually work correctly

The state tracking changes broke the expiry active check by not writing to
it and instead sitting timers_active, which is already set.

That's not a big issue as the actual expiry is protected by sighand lock,
so concurrent handling is not possible. That means that the second task
which invokes that function executes the expiry code for nothing.

Write to the proper flag.

Also add a check whether the flag is set into check_process_timers(). That
check had been missing in the code before the rework already. The check for
another task handling the expiry of process wide timers was only done in
the fastpath check. If the fastpath check returns true because a per task
timer expired, then the checking of process wide timers was done in
parallel which is as explained above just a waste of cycles.

Fixes: 244d49e30653 ("posix-cpu-timers: Move state tracking to struct posix_cputimers")
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Frederic Weisbecker <frederic@kernel.org>
---
 kernel/time/posix-cpu-timers.c | 9 +++++----
 1 file changed, 5 insertions(+), 4 deletions(-)

(limited to 'kernel/time/posix-cpu-timers.c')

diff --git a/kernel/time/posix-cpu-timers.c b/kernel/time/posix-cpu-timers.c
index 73c492ce404b..c3a95b122209 100644
--- a/kernel/time/posix-cpu-timers.c
+++ b/kernel/time/posix-cpu-timers.c
@@ -884,16 +884,17 @@ static void check_process_timers(struct task_struct *tsk,
 
 	/*
 	 * If there are no active process wide timers (POSIX 1.b, itimers,
-	 * RLIMIT_CPU) nothing to check.
+	 * RLIMIT_CPU) nothing to check. Also skip the process wide timer
+	 * processing when there is already another task handling them.
 	 */
-	if (!READ_ONCE(pct->timers_active))
+	if (!READ_ONCE(pct->timers_active) || pct->expiry_active)
 		return;
 
-       /*
+	/*
 	 * Signify that a thread is checking for process timers.
 	 * Write access to this field is protected by the sighand lock.
 	 */
-	pct->timers_active = true;
+	pct->expiry_active = true;
 
 	/*
 	 * Collect the current process totals. Group accounting is active
-- 
cgit v1.2.3


From 77b4b5420422fc037d00b8f3f0e89b2262e4ae29 Mon Sep 17 00:00:00 2001
From: Thomas Gleixner <tglx@linutronix.de>
Date: Thu, 5 Sep 2019 23:15:08 +0200
Subject: posix-cpu-timers: Fix permission check regression

The recent consolidation of the three permission checks introduced a subtle
regression. For timer_create() with a process wide timer it returns the
current task if the lookup through the PID which is encoded into the
clockid results in returning current.

That's broken because it does not validate whether the current task is the
group leader.

That was caused by the two different variants of permission checks:

  - posix_cpu_timer_get() allowed access to the process wide clock when the
    looked up task is current. That's not an issue because the process wide
    clock is in the shared sighand.

  - posix_cpu_timer_create() made sure that the looked up task is the group
    leader.

Restore the previous state.

Note, that these permission checks are more than questionable, but that's
subject to follow up changes.

Fixes: 6ae40e3fdcd3 ("posix-cpu-timers: Provide task validation functions")
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://lkml.kernel.org/r/alpine.DEB.2.21.1909052314110.1902@nanos.tec.linutronix.de
---
 kernel/time/posix-cpu-timers.c | 44 +++++++++++++++++++++++++++++++++---------
 1 file changed, 35 insertions(+), 9 deletions(-)

(limited to 'kernel/time/posix-cpu-timers.c')

diff --git a/kernel/time/posix-cpu-timers.c b/kernel/time/posix-cpu-timers.c
index c3a95b122209..92a431981b1c 100644
--- a/kernel/time/posix-cpu-timers.c
+++ b/kernel/time/posix-cpu-timers.c
@@ -47,25 +47,46 @@ void update_rlimit_cpu(struct task_struct *task, unsigned long rlim_new)
 /*
  * Functions for validating access to tasks.
  */
-static struct task_struct *lookup_task(const pid_t pid, bool thread)
+static struct task_struct *lookup_task(const pid_t pid, bool thread,
+				       bool gettime)
 {
 	struct task_struct *p;
 
+	/*
+	 * If the encoded PID is 0, then the timer is targeted at current
+	 * or the process to which current belongs.
+	 */
 	if (!pid)
 		return thread ? current : current->group_leader;
 
 	p = find_task_by_vpid(pid);
-	if (!p || p == current)
+	if (!p)
 		return p;
+
 	if (thread)
 		return same_thread_group(p, current) ? p : NULL;
-	if (p == current)
-		return p;
+
+	if (gettime) {
+		/*
+		 * For clock_gettime(PROCESS) the task does not need to be
+		 * the actual group leader. tsk->sighand gives
+		 * access to the group's clock.
+		 *
+		 * Timers need the group leader because they take a
+		 * reference on it and store the task pointer until the
+		 * timer is destroyed.
+		 */
+		return (p == current || thread_group_leader(p)) ? p : NULL;
+	}
+
+	/*
+	 * For processes require that p is group leader.
+	 */
 	return has_group_leader_pid(p) ? p : NULL;
 }
 
 static struct task_struct *__get_task_for_clock(const clockid_t clock,
-						bool getref)
+						bool getref, bool gettime)
 {
 	const bool thread = !!CPUCLOCK_PERTHREAD(clock);
 	const pid_t pid = CPUCLOCK_PID(clock);
@@ -75,7 +96,7 @@ static struct task_struct *__get_task_for_clock(const clockid_t clock,
 		return NULL;
 
 	rcu_read_lock();
-	p = lookup_task(pid, thread);
+	p = lookup_task(pid, thread, gettime);
 	if (p && getref)
 		get_task_struct(p);
 	rcu_read_unlock();
@@ -84,12 +105,17 @@ static struct task_struct *__get_task_for_clock(const clockid_t clock,
 
 static inline struct task_struct *get_task_for_clock(const clockid_t clock)
 {
-	return __get_task_for_clock(clock, true);
+	return __get_task_for_clock(clock, true, false);
+}
+
+static inline struct task_struct *get_task_for_clock_get(const clockid_t clock)
+{
+	return __get_task_for_clock(clock, true, true);
 }
 
 static inline int validate_clock_permissions(const clockid_t clock)
 {
-	return __get_task_for_clock(clock, false) ? 0 : -EINVAL;
+	return __get_task_for_clock(clock, false, false) ? 0 : -EINVAL;
 }
 
 /*
@@ -339,7 +365,7 @@ static int posix_cpu_clock_get(const clockid_t clock, struct timespec64 *tp)
 	struct task_struct *tsk;
 	u64 t;
 
-	tsk = get_task_for_clock(clock);
+	tsk = get_task_for_clock_get(clock);
 	if (!tsk)
 		return -EINVAL;
 
-- 
cgit v1.2.3