Skip to contentSkip to conversation

clocksource: sun4i: Save and restore timer registers before and after sleeping

PATCH In review wuyan 2 participants 3 messages Oct 10, 2020 – Dec 12, 2020

Subsystems (2)
lore
Share & tools
lore

wuyan

author
#0 lore raw

PATCH 1/1 clocksource: sun4i: Save and restore timer registers before and after sleeping

Change-Id I7edbc00fd0968d0301757f5a75dbd6f53d6a7cd7
---
 drivers/clocksource/timer-sun4i.c | 45 +++++++++++++++++++++++++++++--
 1 file changed, 43 insertions(+), 2 deletions(-)
drivers/clocksource/timer-sun4i.c +43 -2
index 0ba8155b8287..49fb6b90ec15 100644
@@ -29,6 +29,7 @@
 #define TIMER_IRQ_EN_REG	0x00
 #define TIMER_IRQ_EN(val)		BIT(val)
 #define TIMER_IRQ_ST_REG	0x04
+#define TIMER_IRQ_CLEAR(val)		BIT(val)
 #define TIMER_CTL_REG(val)	(0x10 * val + 0x10)
 #define TIMER_CTL_ENABLE		BIT(0)
 #define TIMER_CTL_RELOAD		BIT(1)
@@ -41,6 +42,19 @@
 
 #define TIMER_SYNC_TICKS	3
 
+/* Registers which needs to be saved and restored before and after sleeping */
+static u32 regs_offset[] = {
+	TIMER_IRQ_EN_REG,
+	TIMER_IRQ_ST_REG,
+	TIMER_CTL_REG(0),
+	TIMER_INTVAL_REG(0),
+	TIMER_CNTVAL_REG(0),
+	TIMER_CTL_REG(1),
+	TIMER_INTVAL_REG(1),
+	TIMER_CNTVAL_REG(1),
+};
+static u32 regs_backup[ARRAY_SIZE(regs_offset)];
+
 /*
  * When we disable a timer, we need to wait at least for 2 cycles of
  * the timer source clock. We will use for that the clocksource timer
@@ -82,10 +96,37 @@ static void sun4i_clkevt_time_start(void __iomem *base, u8 timer,
 	       base + TIMER_CTL_REG(timer));
 }
 
+static inline void save_regs(void __iomem *base)
+{
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(regs_offset); i++)
+		regs_backup[i] = readl(base + regs_offset[i]);
+}
+
+static inline void restore_regs(void __iomem *base)
+{
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(regs_offset); i++)
+		writel(regs_backup[i], base + regs_offset[i]);
+}
+
 static int sun4i_clkevt_shutdown(struct clock_event_device *evt)
 {
 	struct timer_of *to = to_timer_of(evt);
 
+	save_regs(timer_of_base(to));
+	sun4i_clkevt_time_stop(timer_of_base(to), 0);
+
+	return 0;
+}
+
+static int sun4i_tick_resume(struct clock_event_device *evt)
+{
+	struct timer_of *to = to_timer_of(evt);
+
+	restore_regs(timer_of_base(to));
 	sun4i_clkevt_time_stop(timer_of_base(to), 0);
 
 	return 0;
@@ -126,7 +167,7 @@ static int sun4i_clkevt_next_event(unsigned long evt,
 
 static void sun4i_timer_clear_interrupt(void __iomem *base)
 {
-	writel(TIMER_IRQ_EN(0), base + TIMER_IRQ_ST_REG);
+	writel(TIMER_IRQ_CLEAR(0), base + TIMER_IRQ_ST_REG);
 }
 
 static irqreturn_t sun4i_timer_interrupt(int irq, void *dev_id)
@@ -150,7 +191,7 @@ static struct timer_of to = {
 		.set_state_shutdown = sun4i_clkevt_shutdown,
 		.set_state_periodic = sun4i_clkevt_set_periodic,
 		.set_state_oneshot = sun4i_clkevt_set_oneshot,
-		.tick_resume = sun4i_clkevt_shutdown,
+		.tick_resume = sun4i_tick_resume,
 		.set_next_event = sun4i_clkevt_next_event,
 		.cpumask = cpu_possible_mask,
 	},
Signature · 7 lines
2.28.0


_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
#1 lore raw

Hi!

Quoted 1 line · wuyan

A commit log would be welcome here. Also, the last time you contributed you used the name Martin Wu in your Signed-off-by, it would be nice to be consistent there.

Quoted 1 line

Change-Id: I7edbc00fd0968d0301757f5a75dbd6f53d6a7cd7

This should be removed

Quoted 17 lines
---
 drivers/clocksource/timer-sun4i.c | 45 +++++++++++++++++++++++++++++--
 1 file changed, 43 insertions(+), 2 deletions(-)
drivers/clocksource/timer-sun4i.c +1 -0
index 0ba8155b8287..49fb6b90ec15 100644
@@ -29,6 +29,7 @@
 #define TIMER_IRQ_EN_REG	0x00
 #define TIMER_IRQ_EN(val)		BIT(val)
 #define TIMER_IRQ_ST_REG	0x04
+#define TIMER_IRQ_CLEAR(val)		BIT(val)
 #define TIMER_CTL_REG(val)	(0x10 * val + 0x10)
 #define TIMER_CTL_ENABLE		BIT(0)
 #define TIMER_CTL_RELOAD		BIT(1)

drivers/clocksource/timer-sun4i.c line 46

Quoted 5 lines
drivers/clocksource/timer-sun4i.c +2 -0
@@ -41,6 +42,19 @@
 
 #define TIMER_SYNC_TICKS	3
 
+/* Registers which needs to be saved and restored before and after sleeping */
+static u32 regs_offset[] = {

It would be better to have a prefix (like sun4i_timer to be consistent) there so that we know it's less confusing and we know it's not some generic thing.

drivers/clocksource/timer-sun4i.c line 48

Quoted 2 lines
drivers/clocksource/timer-sun4i.c +2 -0
+	TIMER_IRQ_EN_REG,
+	TIMER_IRQ_ST_REG,

Why do you need to save the interrupt status register?

drivers/clocksource/timer-sun4i.c line 56

Quoted 8 lines
drivers/clocksource/timer-sun4i.c +8 -0
+	TIMER_CTL_REG(0),
+	TIMER_INTVAL_REG(0),
+	TIMER_CNTVAL_REG(0),
+	TIMER_CTL_REG(1),
+	TIMER_INTVAL_REG(1),
+	TIMER_CNTVAL_REG(1),
+};
+static u32 regs_backup[ARRAY_SIZE(regs_offset)];

We should store this one in the timer_of struct so that we don't have any issue if there's two timers at some point.

Quoted 13 lines
drivers/clocksource/timer-sun4i.c +6 -0
 /*
  * When we disable a timer, we need to wait at least for 2 cycles of
  * the timer source clock. We will use for that the clocksource timer
@@ -82,10 +96,37 @@ static void sun4i_clkevt_time_start(void __iomem *base, u8 timer,
 	       base + TIMER_CTL_REG(timer));
 }
 
+static inline void save_regs(void __iomem *base)
+{
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(regs_offset); i++)
+		regs_backup[i] = readl(base + regs_offset[i]);

drivers/clocksource/timer-sun4i.c static void sun4i_clkevt_time_start(void __iomem *base, u8 timer, · line 114

Quoted 10 lines
drivers/clocksource/timer-sun4i.c +10 -0
+}
+
+static inline void restore_regs(void __iomem *base)
+{
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(regs_offset); i++)
+		writel(regs_backup[i], base + regs_offset[i]);
+}
+

Same thing here, using the prefix would be nice for those two functions name.

Quoted 19 lines
drivers/clocksource/timer-sun4i.c +11 -0
 static int sun4i_clkevt_shutdown(struct clock_event_device *evt)
 {
 	struct timer_of *to = to_timer_of(evt);
 
+	save_regs(timer_of_base(to));
+	sun4i_clkevt_time_stop(timer_of_base(to), 0);
+
+	return 0;
+}
+
+static int sun4i_tick_resume(struct clock_event_device *evt)
+{
+	struct timer_of *to = to_timer_of(evt);
+
+	restore_regs(timer_of_base(to));
 	sun4i_clkevt_time_stop(timer_of_base(to), 0);
 
 	return 0;

drivers/clocksource/timer-sun4i.c static int sun4i_clkevt_next_event(unsigned long evt, · line 171

Quoted 6 lines
drivers/clocksource/timer-sun4i.c +1 -1
@@ -126,7 +167,7 @@ static int sun4i_clkevt_next_event(unsigned long evt,
 
 static void sun4i_timer_clear_interrupt(void __iomem *base)
 {
-	writel(TIMER_IRQ_EN(0), base + TIMER_IRQ_ST_REG);
+	writel(TIMER_IRQ_CLEAR(0), base + TIMER_IRQ_ST_REG);
 }

This is mostly a cosmetic change right? Either way, it should be in a separate patch.

Thanks!
Maxime

author
#2 lore raw

Hi Maxime,
Sorry for the delay...

Quoted 8 lines · Maxime Ripard

Hi!

Quoted 1 line · wuyan


A commit log would be welcome here. Also, the last time you contributed you used the name Martin Wu in your Signed-off-by, it would be nice to be consistent there.

OK. I'll add the commit log and change my name back to 'Martin Wu' next time. Sorry for the inconvenience.

Quoted 3 lines
Quoted 1 line

Change-Id: I7edbc00fd0968d0301757f5a75dbd6f53d6a7cd7


This should be removed

OK. Thanks for your notice.

Quoted 18 lines
Quoted 18 lines

---
 drivers/clocksource/timer-sun4i.c | 45 +++++++++++++++++++++++++++++--

 1 file changed, 43 insertions(+), 2 deletions(-)
drivers/clocksource/timer-sun4i.c +1 -0
index 0ba8155b8287..49fb6b90ec15 100644
@@ -29,6 +29,7 @@
  #define TIMER_IRQ_EN_REG	0x00
  #define TIMER_IRQ_EN(val)	BIT(val)
  #define TIMER_IRQ_ST_REG	0x04
+#define TIMER_IRQ_CLEAR(val)	BIT(val)
  #define TIMER_CTL_REG(val)	(0x10 * val + 0x10)
  #define TIMER_CTL_ENABLE	BIT(0)
  #define TIMER_CTL_RELOAD	BIT(1)
@@ -41,6 +42,19 @@

Quoted 8 lines

drivers/clocksource/timer-sun4i.c line 46

Quoted 4 lines

 #define TIMER_SYNC_TICKS 3

drivers/clocksource/timer-sun4i.c +2 -0
+/* Registers which needs to be saved and restored before and after sleeping */
+static u32 regs_offset[] = {


It would be better to have a prefix (like sun4i_timer to be consistent) there so that we know it's less confusing and we know it's not some generic thing.

OK. I'll rename 'regs_offset' to 'sun4i_timer_regs_offset'.

Quoted 4 lines

drivers/clocksource/timer-sun4i.c line 48

Quoted 2 lines
drivers/clocksource/timer-sun4i.c +2 -0
+	TIMER_IRQ_EN_REG,
+	TIMER_IRQ_ST_REG,


Why do you need to save the interrupt status register?

The TIMER_IRQ_ST_REG should not be restored. I'll remove this one.

Quoted 11 lines

drivers/clocksource/timer-sun4i.c line 56

Quoted 8 lines
drivers/clocksource/timer-sun4i.c +8 -0
+	TIMER_CTL_REG(0),
+	TIMER_INTVAL_REG(0),
+	TIMER_CNTVAL_REG(0),
+	TIMER_CTL_REG(1),
+	TIMER_INTVAL_REG(1),
+	TIMER_CNTVAL_REG(1),
+};
+static u32 regs_backup[ARRAY_SIZE(regs_offset)];


We should store this one in the timer_of struct so that we don't have any issue if there's two timers at some point.

OK. I'll attach 'regs_backup[]' to '(struct timer_of).private_data'.

Quoted 18 lines
Quoted 18 lines

 /*
  * When we disable a timer, we need to wait at least for 2 cycles of
  * the timer source clock. We will use for that the clocksource timer

drivers/clocksource/timer-sun4i.c +0 -0
@@ -82,10 +96,37 @@ static void sun4i_clkevt_time_start(void __iomem *base, u8 timer,
         base + TIMER_CTL_REG(timer));
  }

drivers/clocksource/timer-sun4i.c +11 -0
+static inline void save_regs(void __iomem *base)
+{
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(regs_offset); i++)
+	regs_backup[i] = readl(base + regs_offset[i]);
+}
+
+static inline void restore_regs(void __iomem *base)
+{
+	int i;
Quoted 8 lines

drivers/clocksource/timer-sun4i.c static void sun4i_clkevt_time_start(void __iomem *base, u8 timer, · line 114

Quoted 5 lines
drivers/clocksource/timer-sun4i.c +5 -0
+
+	for (i = 0; i < ARRAY_SIZE(regs_offset); i++)
+	writel(regs_backup[i], base + regs_offset[i]);
+}
+


Same thing here, using the prefix would be nice for those two functions name.

OK. I'll use 'sun4i_timer_save_regs/sun4i_timer_restore_regs' instead.

Quoted 20 lines
Quoted 20 lines

 static int sun4i_clkevt_shutdown(struct clock_event_device *evt) {
 struct timer_of *to = to_timer_of(evt);

drivers/clocksource/timer-sun4i.c +11 -0
+	save_regs(timer_of_base(to));
+	sun4i_clkevt_time_stop(timer_of_base(to), 0);
+
+	return 0;
+}
+
+static int sun4i_tick_resume(struct clock_event_device *evt)
+{
+	struct timer_of *to = to_timer_of(evt);
+
+	restore_regs(timer_of_base(to));

 sun4i_clkevt_time_stop(timer_of_base(to), 0);

 return 0;

drivers/clocksource/timer-sun4i.c +0 -0
@@ -126,7 +167,7 @@ static int sun4i_clkevt_next_event(unsigned long evt,

Quoted 8 lines
Quoted 5 lines

 static void sun4i_timer_clear_interrupt(void __iomem *base) {

drivers/clocksource/timer-sun4i.c +1 -1
-	writel(TIMER_IRQ_EN(0), base + TIMER_IRQ_ST_REG);
+	writel(TIMER_IRQ_CLEAR(0), base + TIMER_IRQ_ST_REG);

 }


This is mostly a cosmetic change right? Either way, it should be in a separate patch.

Yes. I'll commit it separately.

Quoted 2 lines

Thanks!
Maxime

Thanks for your kindly notice.

Best Regards,
Martin Wu
_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel