Received-SPF: pass (google.com: domain of linux-kernel+bounces-116579-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:45e3:2400::1 as permitted sender) client-ip=2604:1380:45e3:2400::1;
From: Adrian Hunter <adrian.hunter@intel.com>
To: Thomas Gleixner <tglx@linutronix.de>
Cc: Michael Ellerman <mpe@ellerman.id.au>,
	Nicholas Piggin <npiggin@gmail.com>,
	Christophe Leroy <christophe.leroy@csgroup.eu>,
	"Aneesh Kumar K.V" <aneesh.kumar@kernel.org>,
	"Naveen N. Rao" <naveen.n.rao@linux.ibm.com>,
	Heiko Carstens <hca@linux.ibm.com>,
	Vasily Gorbik <gor@linux.ibm.com>,
	Alexander Gordeev <agordeev@linux.ibm.com>,
	Christian Borntraeger <borntraeger@linux.ibm.com>,
	Sven Schnelle <svens@linux.ibm.com>,
	Ingo Molnar <mingo@redhat.com>,
	Borislav Petkov <bp@alien8.de>,
	Dave Hansen <dave.hansen@linux.intel.com>,
	x86@kernel.org,
	"H. Peter Anvin" <hpa@zytor.com>,
	Andy Lutomirski <luto@kernel.org>,
	Vincenzo Frascino <vincenzo.frascino@arm.com>,
	John Stultz <jstultz@google.com>,
	Stephen Boyd <sboyd@kernel.org>,
	Peter Zijlstra <peterz@infradead.org>,
	Randy Dunlap <rdunlap@infradead.org>,
	Bjorn Helgaas <bhelgaas@google.com>,
	Arnd Bergmann <arnd@arndb.de>,
	Anna-Maria Behnsen <anna-maria@linutronix.de>,
	linuxppc-dev@lists.ozlabs.org,
	linux-kernel@vger.kernel.org,
	linux-s390@vger.kernel.org
Subject: [PATCH V2 19/19] clocksource: Make watchdog and suspend-timing multiplication overflow safe
Date: Mon, 25 Mar 2024 08:40:23 +0200
Message-Id: <20240325064023.2997-20-adrian.hunter@intel.com>
In-Reply-To: <20240325064023.2997-1-adrian.hunter@intel.com>
References: <20240325064023.2997-1-adrian.hunter@intel.com>
Precedence: bulk
MIME-Version: 1.0
Organization: Intel Finland Oy, Registered Address: PL 281, 00181 Helsinki, Business Identity Code: 0357606 - 4, Domiciled in Helsinki
Content-Transfer-Encoding: 8bit

Kernel timekeeping is designed to keep the change in cycles (since the last
timer interrupt) below max_cycles, which prevents multiplication overflow
when converting cycles to nanoseconds. However, if timer interrupts stop,
the clocksource_cyc2ns() calculation will eventually overflow.

Add protection against that. Simplify by folding together
clocksource_delta() and clocksource_cyc2ns() into cycles_to_nsec_safe().
Check against max_cycles, falling back to a slower higher precision
calculation.

Suggested-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
---
 kernel/time/clocksource.c | 42 +++++++++++++++++++--------------------
 1 file changed, 20 insertions(+), 22 deletions(-)

diff --git a/kernel/time/clocksource.c b/kernel/time/clocksource.c
index e5b260aa0e02..4d50d53ac719 100644
--- a/kernel/time/clocksource.c
+++ b/kernel/time/clocksource.c
@@ -20,6 +20,16 @@
 #include "tick-internal.h"
 #include "timekeeping_internal.h"
 
+static noinline u64 cycles_to_nsec_safe(struct clocksource *cs, u64 start, u64 end)
+{
+	u64 delta = clocksource_delta(end, start, cs->mask);
+
+	if (likely(delta < cs->max_cycles))
+		return clocksource_cyc2ns(delta, cs->mult, cs->shift);
+
+	return mul_u64_u32_shr(delta, cs->mult, cs->shift);
+}
+
 /**
  * clocks_calc_mult_shift - calculate mult/shift factors for scaled math of clocks
  * @mult:	pointer to mult variable
@@ -222,8 +232,8 @@ enum wd_read_status {
 static enum wd_read_status cs_watchdog_read(struct clocksource *cs, u64 *csnow, u64 *wdnow)
 {
 	unsigned int nretries, max_retries;
-	u64 wd_end, wd_end2, wd_delta;
 	int64_t wd_delay, wd_seq_delay;
+	u64 wd_end, wd_end2;
 
 	max_retries = clocksource_get_max_watchdog_retry();
 	for (nretries = 0; nretries <= max_retries; nretries++) {
@@ -234,9 +244,7 @@ static enum wd_read_status cs_watchdog_read(struct clocksource *cs, u64 *csnow,
 		wd_end2 = watchdog->read(watchdog);
 		local_irq_enable();
 
-		wd_delta = clocksource_delta(wd_end, *wdnow, watchdog->mask);
-		wd_delay = clocksource_cyc2ns(wd_delta, watchdog->mult,
-					      watchdog->shift);
+		wd_delay = cycles_to_nsec_safe(watchdog, *wdnow, wd_end);
 		if (wd_delay <= WATCHDOG_MAX_SKEW) {
 			if (nretries > 1 || nretries >= max_retries) {
 				pr_warn("timekeeping watchdog on CPU%d: %s retried %d times before success\n",
@@ -254,8 +262,7 @@ static enum wd_read_status cs_watchdog_read(struct clocksource *cs, u64 *csnow,
 		 * report system busy, reinit the watchdog and skip the current
 		 * watchdog test.
 		 */
-		wd_delta = clocksource_delta(wd_end2, wd_end, watchdog->mask);
-		wd_seq_delay = clocksource_cyc2ns(wd_delta, watchdog->mult, watchdog->shift);
+		wd_seq_delay = cycles_to_nsec_safe(watchdog, wd_end, wd_end2);
 		if (wd_seq_delay > WATCHDOG_MAX_SKEW/2)
 			goto skip_test;
 	}
@@ -366,8 +373,7 @@ void clocksource_verify_percpu(struct clocksource *cs)
 		delta = (csnow_end - csnow_mid) & cs->mask;
 		if (delta < 0)
 			cpumask_set_cpu(cpu, &cpus_ahead);
-		delta = clocksource_delta(csnow_end, csnow_begin, cs->mask);
-		cs_nsec = clocksource_cyc2ns(delta, cs->mult, cs->shift);
+		cs_nsec = cycles_to_nsec_safe(cs, csnow_begin, csnow_end);
 		if (cs_nsec > cs_nsec_max)
 			cs_nsec_max = cs_nsec;
 		if (cs_nsec < cs_nsec_min)
@@ -398,8 +404,8 @@ static inline void clocksource_reset_watchdog(void)
 
 static void clocksource_watchdog(struct timer_list *unused)
 {
-	u64 csnow, wdnow, cslast, wdlast, delta;
 	int64_t wd_nsec, cs_nsec, interval;
+	u64 csnow, wdnow, cslast, wdlast;
 	int next_cpu, reset_pending;
 	struct clocksource *cs;
 	enum wd_read_status read_ret;
@@ -456,12 +462,8 @@ static void clocksource_watchdog(struct timer_list *unused)
 			continue;
 		}
 
-		delta = clocksource_delta(wdnow, cs->wd_last, watchdog->mask);
-		wd_nsec = clocksource_cyc2ns(delta, watchdog->mult,
-					     watchdog->shift);
-
-		delta = clocksource_delta(csnow, cs->cs_last, cs->mask);
-		cs_nsec = clocksource_cyc2ns(delta, cs->mult, cs->shift);
+		wd_nsec = cycles_to_nsec_safe(watchdog, cs->wd_last, wdnow);
+		cs_nsec = cycles_to_nsec_safe(cs, cs->cs_last, csnow);
 		wdlast = cs->wd_last; /* save these in case we print them */
 		cslast = cs->cs_last;
 		cs->cs_last = csnow;
@@ -832,7 +834,7 @@ void clocksource_start_suspend_timing(struct clocksource *cs, u64 start_cycles)
  */
 u64 clocksource_stop_suspend_timing(struct clocksource *cs, u64 cycle_now)
 {
-	u64 now, delta, nsec = 0;
+	u64 now, nsec = 0;
 
 	if (!suspend_clocksource)
 		return 0;
@@ -847,12 +849,8 @@ u64 clocksource_stop_suspend_timing(struct clocksource *cs, u64 cycle_now)
 	else
 		now = suspend_clocksource->read(suspend_clocksource);
 
-	if (now > suspend_start) {
-		delta = clocksource_delta(now, suspend_start,
-					  suspend_clocksource->mask);
-		nsec = mul_u64_u32_shr(delta, suspend_clocksource->mult,
-				       suspend_clocksource->shift);
-	}
+	if (now > suspend_start)
+		nsec = cycles_to_nsec_safe(suspend_clocksource, suspend_start, now);
 
 	/*
 	 * Disable the suspend timer to save power if current clocksource is
-- 
2.34.1