3 files changed, 43 insertions, 15 deletions

diff --git a/drivers/clocksource/hyperv_timer.c b/drivers/clocksource/hyperv_timer.c
index 2edc13ca184e..10356d4ec55c 100644
--- a/drivers/clocksource/hyperv_timer.c
+++ b/drivers/clocksource/hyperv_timer.c
@@ -549,14 +549,22 @@ static void __init hv_init_tsc_clocksource(void)
 	union hv_reference_tsc_msr tsc_msr;
 
 	/*
+	 * When running as a guest partition:
+	 *
 	 * If Hyper-V offers TSC_INVARIANT, then the virtualized TSC correctly
 	 * handles frequency and offset changes due to live migration,
 	 * pause/resume, and other VM management operations.  So lower the
 	 * Hyper-V Reference TSC rating, causing the generic TSC to be used.
 	 * TSC_INVARIANT is not offered on ARM64, so the Hyper-V Reference
 	 * TSC will be preferred over the virtualized ARM64 arch counter.
+	 *
+	 * When running as the root partition:
+	 *
+	 * There is no HV_ACCESS_TSC_INVARIANT feature. Always lower the rating
+	 * of the Hyper-V Reference TSC.
 	 */
-	if (ms_hyperv.features & HV_ACCESS_TSC_INVARIANT) {
+	if ((ms_hyperv.features & HV_ACCESS_TSC_INVARIANT) ||
+	    hv_root_partition()) {
 		hyperv_cs_tsc.rating = 250;
 		hyperv_cs_msr.rating = 245;
 	}
diff --git a/drivers/clocksource/timer-rtl-otto.c b/drivers/clocksource/timer-rtl-otto.c
index 8a3068b36e75..24c4aa6a3013 100644
--- a/drivers/clocksource/timer-rtl-otto.c
+++ b/drivers/clocksource/timer-rtl-otto.c
@@ -38,6 +38,7 @@
 #define RTTM_BIT_COUNT		28
 #define RTTM_MIN_DELTA		8
 #define RTTM_MAX_DELTA		CLOCKSOURCE_MASK(28)
+#define RTTM_MAX_DIVISOR	GENMASK(15, 0)
 
 /*
  * Timers are derived from the LXB clock frequency. Usually this is a fixed
@@ -112,6 +113,22 @@ static irqreturn_t rttm_timer_interrupt(int irq, void *dev_id)
 	return IRQ_HANDLED;
 }
 
+static void rttm_bounce_timer(void __iomem *base, u32 mode)
+{
+	/*
+	 * When a running timer has less than ~5us left, a stop/start sequence
+	 * might fail. While the details are unknown the most evident effect is
+	 * that the subsequent interrupt will not be fired.
+	 *
+	 * As a workaround issue an intermediate restart with a very slow
+	 * frequency of ~3kHz keeping the target counter (>=8). So the follow
+	 * up restart will always be issued outside the critical window.
+	 */
+
+	rttm_disable_timer(base);
+	rttm_enable_timer(base, mode, RTTM_MAX_DIVISOR);
+}
+
 static void rttm_stop_timer(void __iomem *base)
 {
 	rttm_disable_timer(base);
@@ -129,7 +146,8 @@ static int rttm_next_event(unsigned long delta, struct clock_event_device *clkev
 	struct timer_of *to = to_timer_of(clkevt);
 
 	RTTM_DEBUG(to->of_base.base);
-	rttm_stop_timer(to->of_base.base);
+	rttm_bounce_timer(to->of_base.base, RTTM_CTRL_COUNTER);
+	rttm_disable_timer(to->of_base.base);
 	rttm_set_period(to->of_base.base, delta);
 	rttm_start_timer(to, RTTM_CTRL_COUNTER);
 
@@ -141,7 +159,8 @@ static int rttm_state_oneshot(struct clock_event_device *clkevt)
 	struct timer_of *to = to_timer_of(clkevt);
 
 	RTTM_DEBUG(to->of_base.base);
-	rttm_stop_timer(to->of_base.base);
+	rttm_bounce_timer(to->of_base.base, RTTM_CTRL_COUNTER);
+	rttm_disable_timer(to->of_base.base);
 	rttm_set_period(to->of_base.base, RTTM_TICKS_PER_SEC / HZ);
 	rttm_start_timer(to, RTTM_CTRL_COUNTER);
 
@@ -153,7 +172,8 @@ static int rttm_state_periodic(struct clock_event_device *clkevt)
 	struct timer_of *to = to_timer_of(clkevt);
 
 	RTTM_DEBUG(to->of_base.base);
-	rttm_stop_timer(to->of_base.base);
+	rttm_bounce_timer(to->of_base.base, RTTM_CTRL_TIMER);
+	rttm_disable_timer(to->of_base.base);
 	rttm_set_period(to->of_base.base, RTTM_TICKS_PER_SEC / HZ);
 	rttm_start_timer(to, RTTM_CTRL_TIMER);
 
diff --git a/drivers/clocksource/timer-vf-pit.c b/drivers/clocksource/timer-vf-pit.c
index 911c92146eca..8041a8f62d1f 100644
--- a/drivers/clocksource/timer-vf-pit.c
+++ b/drivers/clocksource/timer-vf-pit.c
@@ -35,30 +35,30 @@ static unsigned long cycle_per_jiffy;
 
 static inline void pit_timer_enable(void)
 {
-	__raw_writel(PITTCTRL_TEN | PITTCTRL_TIE, clkevt_base + PITTCTRL);
+	writel(PITTCTRL_TEN | PITTCTRL_TIE, clkevt_base + PITTCTRL);
 }
 
 static inline void pit_timer_disable(void)
 {
-	__raw_writel(0, clkevt_base + PITTCTRL);
+	writel(0, clkevt_base + PITTCTRL);
 }
 
 static inline void pit_irq_acknowledge(void)
 {
-	__raw_writel(PITTFLG_TIF, clkevt_base + PITTFLG);
+	writel(PITTFLG_TIF, clkevt_base + PITTFLG);
 }
 
 static u64 notrace pit_read_sched_clock(void)
 {
-	return ~__raw_readl(clksrc_base + PITCVAL);
+	return ~readl(clksrc_base + PITCVAL);
 }
 
 static int __init pit_clocksource_init(unsigned long rate)
 {
 	/* set the max load value and start the clock source counter */
-	__raw_writel(0, clksrc_base + PITTCTRL);
-	__raw_writel(~0UL, clksrc_base + PITLDVAL);
-	__raw_writel(PITTCTRL_TEN, clksrc_base + PITTCTRL);
+	writel(0, clksrc_base + PITTCTRL);
+	writel(~0UL, clksrc_base + PITLDVAL);
+	writel(PITTCTRL_TEN, clksrc_base + PITTCTRL);
 
 	sched_clock_register(pit_read_sched_clock, 32, rate);
 	return clocksource_mmio_init(clksrc_base + PITCVAL, "vf-pit", rate,
@@ -76,7 +76,7 @@ static int pit_set_next_event(unsigned long delta,
 	 * hardware requirement.
 	 */
 	pit_timer_disable();
-	__raw_writel(delta - 1, clkevt_base + PITLDVAL);
+	writel(delta - 1, clkevt_base + PITLDVAL);
 	pit_timer_enable();
 
 	return 0;
@@ -125,8 +125,8 @@ static struct clock_event_device clockevent_pit = {
 
 static int __init pit_clockevent_init(unsigned long rate, int irq)
 {
-	__raw_writel(0, clkevt_base + PITTCTRL);
-	__raw_writel(PITTFLG_TIF, clkevt_base + PITTFLG);
+	writel(0, clkevt_base + PITTCTRL);
+	writel(PITTFLG_TIF, clkevt_base + PITTFLG);
 
 	BUG_ON(request_irq(irq, pit_timer_interrupt, IRQF_TIMER | IRQF_IRQPOLL,
 			   "VF pit timer", &clockevent_pit));
@@ -183,7 +183,7 @@ static int __init pit_timer_init(struct device_node *np)
 	cycle_per_jiffy = clk_rate / (HZ);
 
 	/* enable the pit module */
-	__raw_writel(~PITMCR_MDIS, timer_base + PITMCR);
+	writel(~PITMCR_MDIS, timer_base + PITMCR);
 
 	ret = pit_clocksource_init(clk_rate);
 	if (ret)