[PATCH v4 0/4] clocksource: Add MCT support for ARTPEC-8

[PATCH v4 0/4] clocksource: Add MCT support for ARTPEC-8

[Vincent Whitchurch]

This series add supports for the timer block on ARTPEC-8.  The block itself is
fully compatible with the existing exynos4210-mct driver.  The ARTPEC-8 SoC
uses this block from two separate processors running Linux (AMP) so it needs
some extra code to allow this sharing.

v4:
- Rebase on v5.19-rc1 where all pre-requisites are merged
- Minor rework of patch 4 as requested, see patch for details

v3:
- Split and rename devicetree properties
- Add vendor prefix to devicetree properties
- Change descriptions of properties to hopefully describe hardware
- Remove addition of more global variables to the driver

v2:
- The series is now rebased on top of Krzysztof's patch "dt-bindings: timer:
  exynos4210-mct: describe known hardware and its interrupts".
- Combine the Kconfig change and the local timer change into one series
- Use devicetree property rather than module parameter for the local timer handling
- Add specific compatible with the correct number of interrupts.

Cc: linux-kernel@vger.kernel.org
Cc: linux-arm-kernel@lists.infradead.org

Cc: linux-samsung-soc@vger.kernel.org
Cc: alim.akhtar@samsung.com

Cc: devicetree@vger.kernel.org
Cc: robh+dt@kernel.org

Vincent Whitchurch (4):
  dt-bindings: timer: exynos4210-mct: Add ARTPEC-8 MCT support
  clocksource/drivers/exynos_mct: Support frc-shared property
  clocksource/drivers/exynos_mct: Support local-timers property
  clocksource/drivers/exynos_mct: Enable building on ARTPEC

 .../timer/samsung,exynos4210-mct.yaml         | 26 ++++++
 drivers/clocksource/Kconfig                   |  2 +-
 drivers/clocksource/exynos_mct.c              | 83 +++++++++++++++++--
 3 files changed, 101 insertions(+), 10 deletions(-)


base-commit: f2906aa863381afb0015a9eb7fefad885d4e5a56
-- 
2.34.1


_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

[PATCH v4 1/4] dt-bindings: timer: exynos4210-mct: Add ARTPEC-8 MCT support

[Vincent Whitchurch]

The ARTPEC-8 has an MCT with 4 global and 8 local timer interrupts.

The SoC has a quad-core Cortex-A53 and a single-core Cortex-A5 which
share one MCT with one global and eight local timers.  The Cortex-A53
and Cortex-A5 do not have cache-coherency between them, and therefore
run two separate kernels.

The Cortex-A53 boots first and starts the global free-running counter
and also registers a clock events device using the global timer.  (This
global timer clock events is usually replaced by arch timer clock events
for each of the cores.)

When the A5 boots (via the A53), it should not use the global timer
interrupts or write to the global timer registers.  This is because even
if there are four global comparators, the control bits for all four are
in the same registers, and we would need to synchronize between the
cpus.  Instead, the global timer FRC (already started by the A53) should
be used as the clock source, and one of the local timers which are not
used by the A53 can be used for clock events on the A5.

To support this hardware, add a compatible for the MCT as well as two
new properties to describe the hardware-mandated sharing of the FRC and
dedicating local timers to specific processors.

Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Signed-off-by: Vincent Whitchurch <vincent.whitchurch@axis.com>
---

Notes:
    v4:
    - Add Krzysztof's Reviewed-by.
    
    v3:
    - Add all required bindings for ARTPEC-8 in one patch
    - Rename and split local-timer-only to samsung,local-timers and
      samsung,frc-shared
    - Restrict above properties to the ARTPEC-8 compatible.
    - Rewrite descriptions of properties to hopefully describe hardware.
    
    v2:
    - Use devicetree property instead of module parameter.

 .../timer/samsung,exynos4210-mct.yaml         | 26 +++++++++++++++++++
 1 file changed, 26 insertions(+)

diff --git a/Documentation/devicetree/bindings/timer/samsung,exynos4210-mct.yaml b/Documentation/devicetree/bindings/timer/samsung,exynos4210-mct.yaml
index 9c81d00b12e0..829bd2227f7c 100644
--- a/Documentation/devicetree/bindings/timer/samsung,exynos4210-mct.yaml
+++ b/Documentation/devicetree/bindings/timer/samsung,exynos4210-mct.yaml
@@ -25,6 +25,7 @@ properties:
           - samsung,exynos4412-mct
       - items:
           - enum:
+              - axis,artpec8-mct
               - samsung,exynos3250-mct
               - samsung,exynos5250-mct
               - samsung,exynos5260-mct
@@ -45,6 +46,19 @@ properties:
   reg:
     maxItems: 1
 
+  samsung,frc-shared:
+    type: boolean
+    description: |
+      Indicates that the hardware requires that this processor share the
+      free-running counter with a different (main) processor.
+
+  samsung,local-timers:
+    $ref: /schemas/types.yaml#/definitions/uint32-array
+    minItems: 1
+    maxItems: 16
+    description: |
+      List of indices of local timers usable from this processor.
+
   interrupts:
     description: |
       Interrupts should be put in specific order. This is, the local timer
@@ -74,6 +88,17 @@ required:
   - reg
 
 allOf:
+  - if:
+      not:
+        properties:
+          compatible:
+            contains:
+              enum:
+                - axis,artpec8-mct
+    then:
+      properties:
+        samsung,local-timers: false
+        samsung,frc-shared: false
   - if:
       properties:
         compatible:
@@ -101,6 +126,7 @@ allOf:
         compatible:
           contains:
             enum:
+              - axis,artpec8-mct
               - samsung,exynos5260-mct
               - samsung,exynos5420-mct
               - samsung,exynos5433-mct
-- 
2.34.1


_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

[PATCH v4 2/4] clocksource/drivers/exynos_mct: Support frc-shared property

[Vincent Whitchurch]

When the FRC is shared with another main processor, the other processor
is assumed to have started it and this processor should not write to the
global registers.

Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Signed-off-by: Vincent Whitchurch <vincent.whitchurch@axis.com>
---

Notes:
    v4:
    - Add Krzysztof's Reviewed-by.
    
    v3:
    - Split FRC sharing handling from local timer indices handling
    - Remove addition of global variable.

 drivers/clocksource/exynos_mct.c | 21 ++++++++++++++++++---
 1 file changed, 18 insertions(+), 3 deletions(-)

diff --git a/drivers/clocksource/exynos_mct.c b/drivers/clocksource/exynos_mct.c
index f29c812b70c9..12023831dedf 100644
--- a/drivers/clocksource/exynos_mct.c
+++ b/drivers/clocksource/exynos_mct.c
@@ -233,9 +233,16 @@ static cycles_t exynos4_read_current_timer(void)
 }
 #endif
 
-static int __init exynos4_clocksource_init(void)
+static int __init exynos4_clocksource_init(bool frc_shared)
 {
-	exynos4_mct_frc_start();
+	/*
+	 * When the frc is shared, the main processer should have already
+	 * turned it on and we shouldn't be writing to TCON.
+	 */
+	if (frc_shared)
+		mct_frc.resume = NULL;
+	else
+		exynos4_mct_frc_start();
 
 #if defined(CONFIG_ARM)
 	exynos4_delay_timer.read_current_timer = &exynos4_read_current_timer;
@@ -605,6 +612,7 @@ static int __init exynos4_timer_interrupts(struct device_node *np,
 
 static int __init mct_init_dt(struct device_node *np, unsigned int int_type)
 {
+	bool frc_shared = of_property_read_bool(np, "samsung,frc-shared");
 	int ret;
 
 	ret = exynos4_timer_resources(np);
@@ -615,10 +623,17 @@ static int __init mct_init_dt(struct device_node *np, unsigned int int_type)
 	if (ret)
 		return ret;
 
-	ret = exynos4_clocksource_init();
+	ret = exynos4_clocksource_init(frc_shared);
 	if (ret)
 		return ret;
 
+	/*
+	 * When the FRC is shared with a main processor, this secondary
+	 * processor cannot use the global comparator.
+	 */
+	if (frc_shared)
+		return ret;
+
 	return exynos4_clockevent_init();
 }
 
-- 
2.34.1


_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

[PATCH v4 3/4] clocksource/drivers/exynos_mct: Support local-timers property

[Vincent Whitchurch]

If the device tree indicates that the hardware requires that the
processor only use certain local timers, respect that.

Signed-off-by: Vincent Whitchurch <vincent.whitchurch@axis.com>
---

Notes:
    v4:
    - Add function documentation
    - Add const for local_idx
    - Rename irqidx -> irq_idx
    - Error out on invalid number of CPUs
    
    v3:
    - Use array in devicetree
    - Remove addition of global variable
    - Split out FRC sharing changes

 drivers/clocksource/exynos_mct.c | 62 ++++++++++++++++++++++++++++----
 1 file changed, 56 insertions(+), 6 deletions(-)

diff --git a/drivers/clocksource/exynos_mct.c b/drivers/clocksource/exynos_mct.c
index 12023831dedf..bfd60093ee1c 100644
--- a/drivers/clocksource/exynos_mct.c
+++ b/drivers/clocksource/exynos_mct.c
@@ -33,7 +33,7 @@
 #define EXYNOS4_MCT_G_INT_ENB		EXYNOS4_MCTREG(0x248)
 #define EXYNOS4_MCT_G_WSTAT		EXYNOS4_MCTREG(0x24C)
 #define _EXYNOS4_MCT_L_BASE		EXYNOS4_MCTREG(0x300)
-#define EXYNOS4_MCT_L_BASE(x)		(_EXYNOS4_MCT_L_BASE + (0x100 * x))
+#define EXYNOS4_MCT_L_BASE(x)		(_EXYNOS4_MCT_L_BASE + (0x100 * (x)))
 #define EXYNOS4_MCT_L_MASK		(0xffffff00)
 
 #define MCT_L_TCNTB_OFFSET		(0x00)
@@ -66,6 +66,8 @@
 #define MCT_L0_IRQ	4
 /* Max number of IRQ as per DT binding document */
 #define MCT_NR_IRQS	20
+/* Max number of local timers */
+#define MCT_NR_LOCAL	(MCT_NR_IRQS - MCT_L0_IRQ)
 
 enum {
 	MCT_INT_SPI,
@@ -456,7 +458,6 @@ static int exynos4_mct_starting_cpu(unsigned int cpu)
 		per_cpu_ptr(&percpu_mct_tick, cpu);
 	struct clock_event_device *evt = &mevt->evt;
 
-	mevt->base = EXYNOS4_MCT_L_BASE(cpu);
 	snprintf(mevt->name, sizeof(mevt->name), "mct_tick%d", cpu);
 
 	evt->name = mevt->name;
@@ -527,8 +528,17 @@ static int __init exynos4_timer_resources(struct device_node *np)
 	return 0;
 }
 
+/**
+ * exynos4_timer_interrupts - initialize MCT interrupts
+ * @np: device node for MCT
+ * @int_type: interrupt type, MCT_INT_PPI or MCT_INT_SPI
+ * @local_idx: array mapping CPU numbers to local timer indices
+ * @nr_local: size of @local_idx array
+ */
 static int __init exynos4_timer_interrupts(struct device_node *np,
-					   unsigned int int_type)
+					   unsigned int int_type,
+					   const u32 *local_idx,
+					   size_t nr_local)
 {
 	int nr_irqs, i, err, cpu;
 
@@ -561,13 +571,21 @@ static int __init exynos4_timer_interrupts(struct device_node *np,
 	} else {
 		for_each_possible_cpu(cpu) {
 			int mct_irq;
+			unsigned int irq_idx;
 			struct mct_clock_event_device *pcpu_mevt =
 				per_cpu_ptr(&percpu_mct_tick, cpu);
 
+			if (cpu >= nr_local) {
+				err = -EINVAL;
+				goto out_irq;
+			}
+
+			irq_idx = MCT_L0_IRQ + local_idx[cpu];
+
 			pcpu_mevt->evt.irq = -1;
-			if (MCT_L0_IRQ + cpu >= ARRAY_SIZE(mct_irqs))
+			if (irq_idx >= ARRAY_SIZE(mct_irqs))
 				break;
-			mct_irq = mct_irqs[MCT_L0_IRQ + cpu];
+			mct_irq = mct_irqs[irq_idx];
 
 			irq_set_status_flags(mct_irq, IRQ_NOAUTOEN);
 			if (request_irq(mct_irq,
@@ -583,6 +601,17 @@ static int __init exynos4_timer_interrupts(struct device_node *np,
 		}
 	}
 
+	for_each_possible_cpu(cpu) {
+		struct mct_clock_event_device *mevt = per_cpu_ptr(&percpu_mct_tick, cpu);
+
+		if (cpu >= nr_local) {
+			err = -EINVAL;
+			goto out_irq;
+		}
+
+		mevt->base = EXYNOS4_MCT_L_BASE(local_idx[cpu]);
+	}
+
 	/* Install hotplug callbacks which configure the timer on this CPU */
 	err = cpuhp_setup_state(CPUHP_AP_EXYNOS4_MCT_TIMER_STARTING,
 				"clockevents/exynos4/mct_timer:starting",
@@ -613,13 +642,34 @@ static int __init exynos4_timer_interrupts(struct device_node *np,
 static int __init mct_init_dt(struct device_node *np, unsigned int int_type)
 {
 	bool frc_shared = of_property_read_bool(np, "samsung,frc-shared");
+	u32 local_idx[MCT_NR_LOCAL] = {0};
+	int nr_local;
 	int ret;
 
+	nr_local = of_property_count_u32_elems(np, "samsung,local-timers");
+	if (nr_local == 0)
+		return -EINVAL;
+	if (nr_local > 0) {
+		if (nr_local > ARRAY_SIZE(local_idx))
+			return -EINVAL;
+
+		ret = of_property_read_u32_array(np, "samsung,local-timers",
+						 local_idx, nr_local);
+		if (ret)
+			return ret;
+	} else {
+		int i;
+
+		nr_local = ARRAY_SIZE(local_idx);
+		for (i = 0; i < nr_local; i++)
+			local_idx[i] = i;
+	}
+
 	ret = exynos4_timer_resources(np);
 	if (ret)
 		return ret;
 
-	ret = exynos4_timer_interrupts(np, int_type);
+	ret = exynos4_timer_interrupts(np, int_type, local_idx, nr_local);
 	if (ret)
 		return ret;
 
-- 
2.34.1


_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

[PATCH v4 4/4] clocksource/drivers/exynos_mct: Enable building on ARTPEC

[Vincent Whitchurch]

This timer block is used on ARTPEC-8.

Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Signed-off-by: Vincent Whitchurch <vincent.whitchurch@axis.com>
---

Notes:
    v3:
    - Add Krzysztof's Reviewed-by.

 drivers/clocksource/Kconfig | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/clocksource/Kconfig b/drivers/clocksource/Kconfig
index 3c0ee102fe73..b275583f946c 100644
--- a/drivers/clocksource/Kconfig
+++ b/drivers/clocksource/Kconfig
@@ -419,7 +419,7 @@ config ATMEL_TCB_CLKSRC
 config CLKSRC_EXYNOS_MCT
 	bool "Exynos multi core timer driver" if COMPILE_TEST
 	depends on ARM || ARM64
-	depends on ARCH_EXYNOS || COMPILE_TEST
+	depends on ARCH_ARTPEC || ARCH_EXYNOS || COMPILE_TEST
 	help
 	  Support for Multi Core Timer controller on Exynos SoCs.
 
-- 
2.34.1


_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Re: [PATCH v4 3/4] clocksource/drivers/exynos_mct: Support local-timers property

[Krzysztof Kozlowski]

On 09/06/2022 13:27, Vincent Whitchurch wrote:
> If the device tree indicates that the hardware requires that the
> processor only use certain local timers, respect that.
> 
> Signed-off-by: Vincent Whitchurch <vincent.whitchurch@axis.com>

Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>

Daniel,
All patches got my review. Do you need here anything more? It seems it's
only me who reviews such code, so I don't expect more Rb-tags. :)

Best regards,
Krzysztof

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Re: [PATCH v4 3/4] clocksource/drivers/exynos_mct: Support local-timers property

[Vincent Whitchurch]

On Tue, Jun 21, 2022 at 03:11:15PM +0200, Krzysztof Kozlowski wrote:
> On 09/06/2022 13:27, Vincent Whitchurch wrote:
> > If the device tree indicates that the hardware requires that the
> > processor only use certain local timers, respect that.
> > 
> > Signed-off-by: Vincent Whitchurch <vincent.whitchurch@axis.com>
> 
> Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
> 
> Daniel,
> All patches got my review. Do you need here anything more? It seems it's
> only me who reviews such code, so I don't expect more Rb-tags. :)

It's been a couple of months, and unless I'm missing something, it looks
like this series did not get merged, and nor did it receive any further
comments.  Should I resend it with Krzysztof's Reviewed-by tags?

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Re: [PATCH v4 3/4] clocksource/drivers/exynos_mct: Support local-timers property

[Daniel Lezcano]

On 07/09/2022 10:59, Vincent Whitchurch wrote:
> On Tue, Jun 21, 2022 at 03:11:15PM +0200, Krzysztof Kozlowski wrote:
>> On 09/06/2022 13:27, Vincent Whitchurch wrote:
>>> If the device tree indicates that the hardware requires that the
>>> processor only use certain local timers, respect that.
>>>
>>> Signed-off-by: Vincent Whitchurch <vincent.whitchurch@axis.com>
>>
>> Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
>>
>> Daniel,
>> All patches got my review. Do you need here anything more? It seems it's
>> only me who reviews such code, so I don't expect more Rb-tags. :)
> 
> It's been a couple of months, and unless I'm missing something, it looks
> like this series did not get merged, and nor did it receive any further
> comments.  Should I resend it with Krzysztof's Reviewed-by tags?

No, need to resend. I'll pick the series


-- 
<http://www.linaro.org/> Linaro.org │ Open source software for ARM SoCs

Follow Linaro:  <http://www.facebook.com/pages/Linaro> Facebook |
<http://twitter.com/#!/linaroorg> Twitter |
<http://www.linaro.org/linaro-blog/> Blog

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Re: [PATCH v4 0/4] clocksource: Add MCT support for ARTPEC-8

[Daniel Lezcano]

On 09/06/2022 13:27, Vincent Whitchurch wrote:
> This series add supports for the timer block on ARTPEC-8.  The block itself is
> fully compatible with the existing exynos4210-mct driver.  The ARTPEC-8 SoC
> uses this block from two separate processors running Linux (AMP) so it needs
> some extra code to allow this sharing.

Applied, thanks


-- 
<http://www.linaro.org/> Linaro.org │ Open source software for ARM SoCs

Follow Linaro:  <http://www.facebook.com/pages/Linaro> Facebook |
<http://twitter.com/#!/linaroorg> Twitter |
<http://www.linaro.org/linaro-blog/> Blog

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel