add TCB driver for sama5d2
Subsystems (2)
v5 cover add TCB driver for sama5d2
When booting under OP-TEE, the SYSC is secured which means the PIT is also not accessible by non-secure world. The TCB 0 however is always available for non-secure world and thus can be used.
This series add a TCB driver and enables it in sama5d2 configs.
---
Changes in v5:
- Fixed missing include in .dts file for IRQ defines
- Add aic interrupt controller node to avoid DTC warnings
Changes in v4:
- Add missing interrupts property to tcb node
Changes in v3:
- Remove useless defines
- Add printf in case of unsupported timers specified in device-tree
- Add "syscon" compatible to tcb node
Changes in v2:
- Reworked driver to use existing Linux bindings
Clément Léger (4):
timer: atmel_tcb_timer: add atmel_tcb driver
ARM: dts: at91: sama5d2: add AIC node
ARM: dts: at91: sama5d2: add TCB node
configs: sama5d2: enable option CONFIG_ATMEL_TCB_TIMER
MAINTAINERS | 1 + arch/arm/dts/sama5d2.dtsi | 27 ++++ configs/sama5d2_icp_mmc_defconfig | 1 + configs/sama5d2_ptc_ek_mmc_defconfig | 1 + configs/sama5d2_ptc_ek_nandflash_defconfig | 1 + configs/sama5d2_xplained_emmc_defconfig | 1 + configs/sama5d2_xplained_mmc_defconfig | 1 + configs/sama5d2_xplained_qspiflash_defconfig | 1 + configs/sama5d2_xplained_spiflash_defconfig | 1 + drivers/timer/Kconfig | 8 + drivers/timer/Makefile | 1 + drivers/timer/atmel_tcb_timer.c | 155 +++++++++++++++++++ 12 files changed, 199 insertions(+) create mode 100644 drivers/timer/atmel_tcb_timer.c
Signature · 1 line
2.34.1
PATCH v5 1/4 timer: atmel_tcb_timer: add atmel_tcb driver
Add a driver for the timer counter block that can be found on sama5d2. This driver will be used when booting under OP-TEE since the pit timer which is part of the SYSC is secured. Channel 1 & 2 are configured to be chained together which allows to have a 64bits counter.
--- MAINTAINERS | 1 + drivers/timer/Kconfig | 8 ++ drivers/timer/Makefile | 1 + drivers/timer/atmel_tcb_timer.c | 155 ++++++++++++++++++++++++++++++++ 4 files changed, 165 insertions(+) create mode 100644 drivers/timer/atmel_tcb_timer.c
MAINTAINERS +1 -0
@@ -352,6 +352,7 @@ F: arch/arm/mach-at91/ F: board/atmel/ F: drivers/cpu/at91_cpu.c F: drivers/misc/microchip_flexcom.c +F: drivers/timer/atmel_tcb_timer.c F: include/dt-bindings/mfd/atmel-flexcom.h F: drivers/timer/mchp-pit64b-timer.c
drivers/timer/Kconfig +8 -0
@@ -96,6 +96,14 @@ config ATMEL_PIT_TIMER it is designed to offer maximum accuracy and efficient management, even for systems with long response time. +config ATMEL_TCB_TIMER + bool "Atmel timer counter support" + depends on TIMER + depends on ARCH_AT91 + help + Select this to enable the use of the timer counter as a monotonic + counter. + config CADENCE_TTC_TIMER bool "Cadence TTC (Triple Timer Counter)" depends on TIMER
drivers/timer/Makefile +1 -0
@@ -10,6 +10,7 @@ obj-$(CONFIG_ARC_TIMER) += arc_timer.o obj-$(CONFIG_AST_TIMER) += ast_timer.o obj-$(CONFIG_ATCPIT100_TIMER) += atcpit100_timer.o obj-$(CONFIG_ATMEL_PIT_TIMER) += atmel_pit_timer.o +obj-$(CONFIG_ATMEL_TCB_TIMER) += atmel_tcb_timer.o obj-$(CONFIG_CADENCE_TTC_TIMER) += cadence-ttc.o obj-$(CONFIG_DESIGNWARE_APB_TIMER) += dw-apb-timer.o obj-$(CONFIG_MPC83XX_TIMER) += mpc83xx_timer.o
drivers/timer/atmel_tcb_timer.c new file +155 -0
155 lines not on this page · raw message
Signature · 1 line
2.34.1
Re: [PATCH v5 1/4] timer: atmel_tcb_timer: add atmel_tcb driver
Quoted 73 lines · Clément Léger
EXTERNAL EMAIL: Do not click links or open attachments unless you know the content is safe
Add a driver for the timer counter block that can be found on sama5d2. This driver will be used when booting under OP-TEE since the pit timer which is part of the SYSC is secured. Channel 1 & 2 are configured to be chained together which allows to have a 64bits counter.
Signed-off-by Clément Léger <[email protected]>--- MAINTAINERS | 1 + drivers/timer/Kconfig | 8 ++ drivers/timer/Makefile | 1 + drivers/timer/atmel_tcb_timer.c | 155 ++++++++++++++++++++++++++++++++ 4 files changed, 165 insertions(+) create mode 100644 drivers/timer/atmel_tcb_timer.cMAINTAINERS +1
-0@@ -352,6 +352,7 @@ F: arch/arm/mach-at91/ F: board/atmel/ F: drivers/cpu/at91_cpu.c F: drivers/misc/microchip_flexcom.c +F: drivers/timer/atmel_tcb_timer.c F: include/dt-bindings/mfd/atmel-flexcom.h F: drivers/timer/mchp-pit64b-timer.cdrivers/timer/Kconfig +8
-0@@ -96,6 +96,14 @@ config ATMEL_PIT_TIMER it is designed to offer maximum accuracy and efficient management, even for systems with long response time. +config ATMEL_TCB_TIMER + bool "Atmel timer counter support" + depends on TIMER + depends on ARCH_AT91 + help + Select this to enable the use of the timer counter as a monotonic + counter. + config CADENCE_TTC_TIMER bool "Cadence TTC (Triple Timer Counter)" depends on TIMERdrivers/timer/Makefile +1
-0@@ -10,6 +10,7 @@ obj-$(CONFIG_ARC_TIMER) += arc_timer.o obj-$(CONFIG_AST_TIMER) += ast_timer.o obj-$(CONFIG_ATCPIT100_TIMER) += atcpit100_timer.o obj-$(CONFIG_ATMEL_PIT_TIMER) += atmel_pit_timer.o +obj-$(CONFIG_ATMEL_TCB_TIMER) += atmel_tcb_timer.o obj-$(CONFIG_CADENCE_TTC_TIMER) += cadence-ttc.o obj-$(CONFIG_DESIGNWARE_APB_TIMER) += dw-apb-timer.o obj-$(CONFIG_MPC83XX_TIMER) += mpc83xx_timer.odrivers/timer/atmel_tcb_timer.c new file +8
-0@@ -0,0 +1,155 @@ +// SPDX-License-Identifier: GPL-2.0+ +/* +* Copyright (C) 2022 Microchip Corporation +* +* Author: Clément Léger <[email protected]> +*/ + +#include <common.h>
drivers/timer/atmel_tcb_timer.c line 18
Quoted 10 lines
drivers/timer/atmel_tcb_timer.c new file +10
-0+#include <clk.h> +#include <dm.h> +#include <timer.h> +#include <asm/io.h> +#include <linux/bitops.h> + +#define TCB_CHAN (chan) ((chan) * 0x40) + +#define TCB_CCR (chan) (0x0 + TCB_CHAN(chan)) +#define TCB_CCR_SWTRG 0x4
This is not used
drivers/timer/atmel_tcb_timer.c line 19
Quoted 1 line
drivers/timer/atmel_tcb_timer.c new file +1
-0+#define TCB_CCR_CLKEN 0x1
Here
drivers/timer/atmel_tcb_timer.c line 22
Quoted 3 lines
drivers/timer/atmel_tcb_timer.c new file +3
-0+ +#define TCB_CMR (chan) (0x4 + TCB_CHAN(chan)) +#define TCB_CMR_WAVE BIT(15)
here
drivers/timer/atmel_tcb_timer.c line 23
Quoted 1 line
drivers/timer/atmel_tcb_timer.c new file +1
-0+#define TCB_CMR_TIMER_CLOCK0 0
Datasheet names it CLOCK1.
drivers/timer/atmel_tcb_timer.c line 25
Quoted 2 lines
drivers/timer/atmel_tcb_timer.c new file +2
-0+#define TCB_CMR_XC1 6 +#define TCB_CMR_ACPA_SET BIT(16)
here
drivers/timer/atmel_tcb_timer.c line 36
Quoted 11 lines
drivers/timer/atmel_tcb_timer.c new file +11
-0+#define TCB_CMR_ACPC_CLEAR (2 << 18) + +#define TCB_CV (chan) (0x10 + TCB_CHAN(chan)) + +#define TCB_RA (chan) (0x14 + TCB_CHAN(chan)) +#define TCB_RC (chan) (0x1c + TCB_CHAN(chan)) + +#define TCB_IDR (chan) (0x28 + TCB_CHAN(chan)) + +#define TCB_BCR 0xc0 +#define TCB_BCR_SYNC 0x1
and here: there a mixture of using BIT() and using plain bit defines: would be good to stick to one or the other.
Quoted 36 lines
drivers/timer/atmel_tcb_timer.c new file +36
-0+ +#define TCB_BMR 0xc4 +#define TCB_BMR_TC1XC1S_TIOA0 (2 << 2) + +#define TCB_WPMR 0xe4 +#define TCB_WPMR_WAKEY 0x54494d + +struct atmel_tcb_plat { +void __iomem *base; +}; + +static u64 atmel_tcb_get_count (struct udevice *dev) +{ +struct atmel_tcb_plat *plat =dev_get_plat (dev); +u64 cv0 =0 ; +u64 cv1 =0 ; + +do { + cv1 =readl (plat->base +TCB_CV (1 )); + cv0 =readl (plat->base +TCB_CV (0 )); + }while (readl (plat->base +TCB_CV (1 )) != cv1); + + cv0 |= cv1 <<32 ; + +return cv0; +} + +static void atmel_tcb_configure (void __iomem *base) +{ +/* Disable write protection */ +writel (TCB_WPMR_WAKEY , base +TCB_WPMR ); + +/* Disable all irqs for both channel 0 & 1 */ +writel (0xff , base +TCB_IDR (0 )); +writel (0xff , base +TCB_IDR (1 )); +
drivers/timer/atmel_tcb_timer.c line 82
Quoted 10 lines
drivers/timer/atmel_tcb_timer.c new file +10
-0+/* +* In order to avoid wrapping, use a 64 bit counter by chaining +* two channels. +* Channel 0 is configured to generate a clock on TIOA0 which is cleared +* when reaching 0x80000000 and set when reaching 0. +*/ +writel (TCB_CMR_TIMER_CLOCK0 |TCB_CMR_WAVE |TCB_CMR_ACPA_SET + |TCB_CMR_ACPC_CLEAR , base +TCB_CMR (0 )); +writel (0x80000000 , base +TCB_RC (0 )); +writel (0x1 , base +TCB_RA (0 ));
I see this set to zero in Linux. Is it set to 1 to avoid "set when reaching 0" described above?
Quoted 12 lines
drivers/timer/atmel_tcb_timer.c new file +12
-0+writel (TCB_CCR_CLKEN , base +TCB_CCR (0 )); + +/* Channel 1 is configured to use TIOA0 as input */ +writel (TCB_CMR_XC1 |TCB_CMR_WAVE , base +TCB_CMR (1 )); +writel (TCB_CCR_CLKEN , base +TCB_CCR (1 )); + +/* Set XC1 input to be TIOA0 (ie output of Channel 0) */ +writel (TCB_BMR_TC1XC1S_TIOA0 , base +TCB_BMR ); + +/* Sync & start all timers */ +writel (TCB_BCR_SYNC , base +TCB_BCR ); +}
drivers/timer/atmel_tcb_timer.c line 104
Quoted 10 lines
drivers/timer/atmel_tcb_timer.c new file +10
-0+ +static int atmel_tcb_probe (struct udevice *dev) +{ +struct atmel_tcb_plat *plat =dev_get_plat (dev); +struct timer_dev_priv *uc_priv =dev_get_uclass_priv (dev); +struct clk clk; +ulong clk_rate; +int ret; + +if (!device_is_compatible (dev->parent,"atmel,sama5d2-tcb" ))
Is this necessary?
Quoted 54 lines
drivers/timer/atmel_tcb_timer.c new file +51
-1+return -EINVAL ; + +/* Currently, we only support channel 0 and 1 to be chained */ +if (dev_read_addr_index (dev,0 ) !=0 && +dev_read_addr_index (dev,1 ) !=1 ) { +printf ("Error: only chained timers 0 and 1 are supported\n" ); +return -EINVAL ; + } + + ret =clk_get_by_name (dev->parent,"gclk" , &clk); +if (ret) +return -EINVAL ; + + clk_rate =clk_get_rate (&clk); +if (!clk_rate) +return -EINVAL ; + + uc_priv->clock_rate = clk_rate; + +atmel_tcb_configure (plat->base); + +return 0 ; +} + +static int atmel_tcb_of_to_plat (struct udevice *dev) +{ +struct atmel_tcb_plat *plat =dev_get_plat (dev); + + plat->base =dev_read_addr_ptr (dev->parent); + +return 0 ; +} + +static const struct timer_ops atmel_tcb_ops = { + .get_count = atmel_tcb_get_count, +}; + +static const struct udevice_id atmel_tcb_ids[] = { + { .compatible ="atmel,tcb-timer" }, + { } +}; + +U_BOOT_DRIVER (atmel_tcb) = { + .name ="atmel_tcb" , + .id =UCLASS_TIMER , + .of_match = atmel_tcb_ids, + .of_to_plat = atmel_tcb_of_to_plat, + .plat_auto =sizeof (struct atmel_tcb_plat ), + .probe = atmel_tcb_probe, + .ops = &atmel_tcb_ops, +}; --2.34.1
Re: [PATCH v5 1/4] timer: atmel_tcb_timer: add atmel_tcb driver
Quoted 97 lines · Le Mon, 7 Mar 2022 10:57:02 +0000, <[email protected]> a écrit :
Quoted 73 lines · Clément Léger
EXTERNAL EMAIL: Do not click links or open attachments unless you know the content is safe
Add a driver for the timer counter block that can be found on sama5d2. This driver will be used when booting under OP-TEE since the pit timer which is part of the SYSC is secured. Channel 1 & 2 are configured to be chained together which allows to have a 64bits counter.
Signed-off-by Clément Léger <[email protected]>--- MAINTAINERS | 1 + drivers/timer/Kconfig | 8 ++ drivers/timer/Makefile | 1 + drivers/timer/atmel_tcb_timer.c | 155 ++++++++++++++++++++++++++++++++ 4 files changed, 165 insertions(+) create mode 100644 drivers/timer/atmel_tcb_timer.cMAINTAINERS +1
-0@@ -352,6 +352,7 @@ F: arch/arm/mach-at91/ F: board/atmel/ F: drivers/cpu/at91_cpu.c F: drivers/misc/microchip_flexcom.c +F: drivers/timer/atmel_tcb_timer.c F: include/dt-bindings/mfd/atmel-flexcom.h F: drivers/timer/mchp-pit64b-timer.cdrivers/timer/Kconfig +8
-0@@ -96,6 +96,14 @@ config ATMEL_PIT_TIMER it is designed to offer maximum accuracy and efficient management, even for systems with long response time. +config ATMEL_TCB_TIMER + bool "Atmel timer counter support" + depends on TIMER + depends on ARCH_AT91 + help + Select this to enable the use of the timer counter as a monotonic + counter. + config CADENCE_TTC_TIMER bool "Cadence TTC (Triple Timer Counter)" depends on TIMERdrivers/timer/Makefile +1
-0@@ -10,6 +10,7 @@ obj-$(CONFIG_ARC_TIMER) += arc_timer.o obj-$(CONFIG_AST_TIMER) += ast_timer.o obj-$(CONFIG_ATCPIT100_TIMER) += atcpit100_timer.o obj-$(CONFIG_ATMEL_PIT_TIMER) += atmel_pit_timer.o +obj-$(CONFIG_ATMEL_TCB_TIMER) += atmel_tcb_timer.o obj-$(CONFIG_CADENCE_TTC_TIMER) += cadence-ttc.o obj-$(CONFIG_DESIGNWARE_APB_TIMER) += dw-apb-timer.o obj-$(CONFIG_MPC83XX_TIMER) += mpc83xx_timer.odrivers/timer/atmel_tcb_timer.c new file +8
-0@@ -0,0 +1,155 @@ +// SPDX-License-Identifier: GPL-2.0+ +/* +* Copyright (C) 2022 Microchip Corporation +* +* Author: Clément Léger <[email protected]> +*/ + +#include <common.h> drivers/timer/atmel_tcb_timer.c line 18
Quoted 10 lines
drivers/timer/atmel_tcb_timer.c new file +10
-0+#include <clk.h> +#include <dm.h> +#include <timer.h> +#include <asm/io.h> +#include <linux/bitops.h> + +#define TCB_CHAN (chan) ((chan) * 0x40) + +#define TCB_CCR (chan) (0x0 + TCB_CHAN(chan)) +#define TCB_CCR_SWTRG 0x4This is not used
drivers/timer/atmel_tcb_timer.c line 19
Quoted 1 line
drivers/timer/atmel_tcb_timer.c new file +1
-0+#define TCB_CCR_CLKEN 0x1Here
drivers/timer/atmel_tcb_timer.c line 22
Quoted 3 lines
drivers/timer/atmel_tcb_timer.c new file +3
-0+ +#define TCB_CMR (chan) (0x4 + TCB_CHAN(chan)) +#define TCB_CMR_WAVE BIT(15)here
Quoted 3 lines
drivers/timer/atmel_tcb_timer.c line 23
Quoted 1 line
drivers/timer/atmel_tcb_timer.c new file +1
-0+#define TCB_CMR_TIMER_CLOCK0 0Datasheet names it CLOCK1.
Indeed, I'll rename that.
Quoted 6 lines
drivers/timer/atmel_tcb_timer.c line 25
Quoted 3 lines
drivers/timer/atmel_tcb_timer.c new file +2
-0+#define TCB_CMR_XC1 6 +#define TCB_CMR_ACPA_SET BIT(16)here
Quoted 14 lines
drivers/timer/atmel_tcb_timer.c line 36
Quoted 11 lines
drivers/timer/atmel_tcb_timer.c new file +11
-0+#define TCB_CMR_ACPC_CLEAR (2 << 18) + +#define TCB_CV (chan) (0x10 + TCB_CHAN(chan)) + +#define TCB_RA (chan) (0x14 + TCB_CHAN(chan)) +#define TCB_RC (chan) (0x1c + TCB_CHAN(chan)) + +#define TCB_IDR (chan) (0x28 + TCB_CHAN(chan)) + +#define TCB_BCR 0xc0 +#define TCB_BCR_SYNC 0x1and here: there a mixture of using
BIT()and using plain bit defines: would be good to stick to one or the other.
Acked, I guess I'll go with plain bit defines.
Quoted 42 lines
Quoted 42 lines
drivers/timer/atmel_tcb_timer.c new file +41
-0+ +#define TCB_BMR 0xc4 +#define TCB_BMR_TC1XC1S_TIOA0 (2 << 2) + +#define TCB_WPMR 0xe4 +#define TCB_WPMR_WAKEY 0x54494d + +struct atmel_tcb_plat { +void __iomem *base; +}; + +static u64 atmel_tcb_get_count (struct udevice *dev) +{ +struct atmel_tcb_plat *plat =dev_get_plat (dev); +u64 cv0 =0 ; +u64 cv1 =0 ; + +do { + cv1 =readl (plat->base +TCB_CV (1 )); + cv0 =readl (plat->base +TCB_CV (0 )); + }while (readl (plat->base +TCB_CV (1 )) != cv1); + + cv0 |= cv1 <<32 ; + +return cv0; +} + +static void atmel_tcb_configure (void __iomem *base) +{ +/* Disable write protection */ +writel (TCB_WPMR_WAKEY , base +TCB_WPMR ); + +/* Disable all irqs for both channel 0 & 1 */ +writel (0xff , base +TCB_IDR (0 )); +writel (0xff , base +TCB_IDR (1 )); + +/* +* In order to avoid wrapping, use a 64 bit counter by chaining +* two channels. +* Channel 0 is configured to generate a clock on TIOA0 which is cleared +* when reaching 0x80000000 and set when reaching 0.
Quoted 8 lines
drivers/timer/atmel_tcb_timer.c line 82
Quoted 5 lines
drivers/timer/atmel_tcb_timer.c new file +5
-0+*/ +writel (TCB_CMR_TIMER_CLOCK0 |TCB_CMR_WAVE |TCB_CMR_ACPA_SET + |TCB_CMR_ACPC_CLEAR , base +TCB_CMR (0 )); +writel (0x80000000 , base +TCB_RC (0 )); +writel (0x1 , base +TCB_RA (0 ));I see this set to zero in Linux. Is it set to 1 to avoid "set when reaching 0" described above?
Yes, do you want me to add a comment ?
Quoted 17 lines
Quoted 17 lines
drivers/timer/atmel_tcb_timer.c new file +16
-0+writel (TCB_CCR_CLKEN , base +TCB_CCR (0 )); + +/* Channel 1 is configured to use TIOA0 as input */ +writel (TCB_CMR_XC1 |TCB_CMR_WAVE , base +TCB_CMR (1 )); +writel (TCB_CCR_CLKEN , base +TCB_CCR (1 )); + +/* Set XC1 input to be TIOA0 (ie output of Channel 0) */ +writel (TCB_BMR_TC1XC1S_TIOA0 , base +TCB_BMR ); + +/* Sync & start all timers */ +writel (TCB_BCR_SYNC , base +TCB_BCR ); +} + +static int atmel_tcb_probe (struct udevice *dev) +{ +struct atmel_tcb_plat *plat =dev_get_plat (dev);
Quoted 8 lines
drivers/timer/atmel_tcb_timer.c line 104
Quoted 6 lines
drivers/timer/atmel_tcb_timer.c new file +6
-0+struct timer_dev_priv *uc_priv =dev_get_uclass_priv (dev); +struct clk clk; +ulong clk_rate; +int ret; + +if (!device_is_compatible (dev->parent,"atmel,sama5d2-tcb" ))Is this necessary?
If we want to be pedantic about the bindings, yes... THis ensure that we respect the node hierarchy that have been described in the Linux bindings.
Thanks for the review,
Signature · 3 lines
Clément Léger, Embedded Linux and Kernel engineer at Bootlin https://bootlin.com
PATCH v5 2/4 ARM: dts: at91: sama5d2: add AIC node
In reply to Clément LégerWhen using interrupts property, a global interrupt controller needs to be added to avoid warnings when compiling device-tree:
arch/arm/dts/at91-sama5d2_xplained.dtb: Warning (interrupts_property):
/ahb/apb/timer@f800c000: Missing interrupt-parent
Add AIC node as the sama5d2 global interrupt controller.
--- arch/arm/dts/sama5d2.dtsi | 10 ++++++++++ 1 file changed, 10 insertions(+)
arch/arm/dts/sama5d2.dtsi +10 -0
@@ -3,6 +3,7 @@ / { model = "Atmel SAMA5D2 family SoC"; compatible = "atmel,sama5d2"; + interrupt-parent = <&aic>; aliases { spi0 = &spi0;
@@ -700,6 +701,15 @@ clocks = <&h32ck>; }; + aic: interrupt-controller@fc020000 { + #interrupt-cells = <3>; + compatible = "atmel,sama5d2-aic"; + interrupt-controller; + reg = <0xfc020000 0x200>; + atmel,external-irqs = <49>; + status = "disabled"; + }; + watchdog@f8048040 { compatible = "atmel,sama5d4-wdt"; reg = <0xf8048040 0x10>;
Signature · 1 line
2.34.1
Re: [PATCH v5 2/4] ARM: dts: at91: sama5d2: add AIC node
Quoted 27 lines · Clément Léger
When using interrupts property, a global interrupt controller needs to be added to avoid warnings when compiling device-tree:
arch/arm/dts/at91-sama5d2_xplained.dtb: Warning (interrupts_property):
/ahb/apb/timer@f800c000: Missing interrupt-parentAdd AIC node as the sama5d2 global interrupt controller.
Signed-off-by Clément Léger <[email protected]>--- arch/arm/dts/sama5d2.dtsi | 10 ++++++++++ 1 file changed, 10 insertions(+)arch/arm/dts/sama5d2.dtsi +1
-0@@ -3,6 +3,7 @@ / { model = "Atmel SAMA5D2 family SoC"; compatible = "atmel,sama5d2"; + interrupt-parent = <&aic>; aliases { spi0 = &spi0;@@ -700,6 +701,15 @@ clocks = <&h32ck>;
arch/arm/dts/sama5d2.dtsi line 711
Quoted 10 lines
arch/arm/dts/sama5d2.dtsi +8
-0}; + aic: interrupt-controller@fc020000 { + #interrupt-cells = <3>; + compatible = "atmel,sama5d2-aic"; + interrupt-controller; + reg = <0xfc020000 0x200>; + atmel,external-irqs = <49>; + status = "disabled"; + };
DT nodes should be sorted by base address.
Check patch 3/4 as well.
Quoted 7 lines
arch/arm/dts/sama5d2.dtsi +1
-1+ watchdog@f8048040 { compatible = "atmel,sama5d4-wdt"; reg = <0xf8048040 0x10>; --2.34.1
Re: [PATCH v5 2/4] ARM: dts: at91: sama5d2: add AIC node
In reply to Clément LégerQuoted 28 lines · Clément Léger
EXTERNAL EMAIL: Do not click links or open attachments unless you know the content is safe
When using interrupts property, a global interrupt controller needs to be added to avoid warnings when compiling device-tree:
arch/arm/dts/at91-sama5d2_xplained.dtb: Warning (interrupts_property):
/ahb/apb/timer@f800c000: Missing interrupt-parentAdd AIC node as the sama5d2 global interrupt controller.
Signed-off-by Clément Léger <[email protected]>--- arch/arm/dts/sama5d2.dtsi | 10 ++++++++++ 1 file changed, 10 insertions(+)arch/arm/dts/sama5d2.dtsi +1
-0@@ -3,6 +3,7 @@ / { model = "Atmel SAMA5D2 family SoC"; compatible = "atmel,sama5d2"; + interrupt-parent = <&aic>; aliases { spi0 = &spi0;
arch/arm/dts/sama5d2.dtsi line 708
Quoted 8 lines
arch/arm/dts/sama5d2.dtsi +5
-0@@ -700,6 +701,15 @@ clocks = <&h32ck>; }; + aic: interrupt-controller@fc020000 { + #interrupt-cells = <3>; + compatible = "atmel,sama5d2-aic"; + interrupt-controller; + reg = <0xfc020000 0x200>;
Most of the nodes in this DT has compatible and reg at the beginning. For clarity would be good to keep the rule for this one as well.
Quoted 10 lines
arch/arm/dts/sama5d2.dtsi +4
-1+ atmel,external-irqs = <49>; + status = "disabled"; + }; + watchdog@f8048040 { compatible = "atmel,sama5d4-wdt"; reg = <0xf8048040 0x10>; --2.34.1
PATCH v5 3/4 ARM: dts: at91: sama5d2: add TCB node
In reply to Clément LégerAdd the device-tree node to describe the TCB timer.
--- arch/arm/dts/sama5d2.dtsi | 17 +++++++++++++++++ 1 file changed, 17 insertions(+)
arch/arm/dts/sama5d2.dtsi +17 -0
@@ -1,4 +1,5 @@ #include "skeleton.dtsi" +#include <dt-bindings/interrupt-controller/irq.h> / { model = "Atmel SAMA5D2 family SoC";
@@ -710,6 +711,22 @@ status = "disabled"; }; + tcb0: timer@f800c000 { + compatible = "atmel,sama5d2-tcb", "simple-mfd", + "syscon"; + reg = <0xf800c000 0x100>; + interrupts = <35 IRQ_TYPE_LEVEL_HIGH 0>; + clocks = <&tcb0_clk>, <&tcb0_gclk>, <&clk32k>; + clock-names = "t0_clk", "gclk", "slow_clk"; + #address-cells = <1>; + #size-cells = <0>; + + timer0: timer@0 { + compatible = "atmel,tcb-timer"; + reg = <0>, <1>; + }; + }; + watchdog@f8048040 { compatible = "atmel,sama5d4-wdt"; reg = <0xf8048040 0x10>;
Signature · 1 line
2.34.1
Re: [PATCH v5 3/4] ARM: dts: at91: sama5d2: add TCB node
Quoted 18 lines · Clément Léger
Add the device-tree node to describe the TCB timer.
Signed-off-by Clément Léger <[email protected]>--- arch/arm/dts/sama5d2.dtsi | 17 +++++++++++++++++ 1 file changed, 17 insertions(+)arch/arm/dts/sama5d2.dtsi +1
-0@@ -1,4 +1,5 @@ #include "skeleton.dtsi" +#include <dt-bindings/interrupt-controller/irq.h> / { model = "Atmel SAMA5D2 family SoC";
arch/arm/dts/sama5d2.dtsi line 716
Quoted 6 lines
arch/arm/dts/sama5d2.dtsi +3
-0@@ -710,6 +711,22 @@ status = "disabled"; }; + tcb0: timer@f800c000 { + compatible = "atmel,sama5d2-tcb", "simple-mfd", + "syscon";
No need to line break the previous line. DT files can have more than 80 chars per line, it's fine.
If there are no other comments I will fix this when applying, no need to send a v6 for now.
Quoted 19 lines
arch/arm/dts/sama5d2.dtsi +13
-1+ reg = <0xf800c000 0x100>; + interrupts = <35 IRQ_TYPE_LEVEL_HIGH 0>; + clocks = <&tcb0_clk>, <&tcb0_gclk>, <&clk32k>; + clock-names = "t0_clk", "gclk", "slow_clk"; + #address-cells = <1>; + #size-cells = <0>; + + timer0: timer@0 { + compatible = "atmel,tcb-timer"; + reg = <0>, <1>; + }; + }; + watchdog@f8048040 { compatible = "atmel,sama5d4-wdt"; reg = <0xf8048040 0x10>; --2.34.1
Re: [PATCH v5 3/4] ARM: dts: at91: sama5d2: add TCB node
Quoted 22 lines · Le Fri, 4 Mar 2022 14:31:08 +0000, <[email protected]> a écrit :
Quoted 21 lines · Clément Léger
Add the device-tree node to describe the TCB timer.
Signed-off-by Clément Léger <[email protected]>--- arch/arm/dts/sama5d2.dtsi | 17 +++++++++++++++++ 1 file changed, 17 insertions(+)arch/arm/dts/sama5d2.dtsi +1
-0@@ -1,4 +1,5 @@ #include "skeleton.dtsi" +#include <dt-bindings/interrupt-controller/irq.h> / { model = "Atmel SAMA5D2 family SoC";@@ -710,6 +711,22 @@ status = "disabled"; };
Quoted 6 lines
arch/arm/dts/sama5d2.dtsi line 716
Quoted 3 lines
arch/arm/dts/sama5d2.dtsi +3
-0+ tcb0: timer@f800c000 { + compatible = "atmel,sama5d2-tcb", "simple-mfd", + "syscon";No need to line break the previous line. DT files can have more than 80 chars per line, it's fine.
Ok, wasn't sure of that, found some of them that had these line breaks. Good to know for next patches.
Quoted 3 lines
If there are no other comments I will fix this when applying, no need to send a v6 for now.
Ok, thanks Eugen.
Quoted 21 lines
Quoted 20 lines
arch/arm/dts/sama5d2.dtsi +13
-1+ reg = <0xf800c000 0x100>; + interrupts = <35 IRQ_TYPE_LEVEL_HIGH 0>; + clocks = <&tcb0_clk>, <&tcb0_gclk>, <&clk32k>; + clock-names = "t0_clk", "gclk", "slow_clk"; + #address-cells = <1>; + #size-cells = <0>; + + timer0: timer@0 { + compatible = "atmel,tcb-timer"; + reg = <0>, <1>; + }; + }; + watchdog@f8048040 { compatible = "atmel,sama5d4-wdt"; reg = <0xf8048040 0x10>; --2.34.1
Signature · 3 lines
Clément Léger, Embedded Linux and Kernel engineer at Bootlin https://bootlin.com
PATCH v5 4/4 configs: sama5d2: enable option CONFIG_ATMEL_TCB_TIMER
In reply to Clément LégerThis will allow using the TCB timer instead of the PIT one when running under OP-TEE.
--- configs/sama5d2_icp_mmc_defconfig | 1 + configs/sama5d2_ptc_ek_mmc_defconfig | 1 + configs/sama5d2_ptc_ek_nandflash_defconfig | 1 + configs/sama5d2_xplained_emmc_defconfig | 1 + configs/sama5d2_xplained_mmc_defconfig | 1 + configs/sama5d2_xplained_qspiflash_defconfig | 1 + configs/sama5d2_xplained_spiflash_defconfig | 1 + 7 files changed, 7 insertions(+)
configs/sama5d2_icp_mmc_defconfig +1 -0
@@ -83,5 +83,6 @@ CONFIG_ATMEL_USART=y CONFIG_TIMER=y CONFIG_SPL_TIMER=y CONFIG_ATMEL_PIT_TIMER=y +CONFIG_ATMEL_TCB_TIMER=y CONFIG_OF_LIBFDT_OVERLAY=y # CONFIG_EFI_LOADER_HII is not set
configs/sama5d2_ptc_ek_mmc_defconfig +1 -0
@@ -72,6 +72,7 @@ CONFIG_DEBUG_UART_ANNOUNCE=y CONFIG_ATMEL_USART=y CONFIG_TIMER=y CONFIG_ATMEL_PIT_TIMER=y +CONFIG_ATMEL_TCB_TIMER=y CONFIG_USB=y CONFIG_USB_EHCI_HCD=y CONFIG_USB_STORAGE=y
configs/sama5d2_ptc_ek_nandflash_defconfig +1 -0
@@ -72,6 +72,7 @@ CONFIG_DEBUG_UART_ANNOUNCE=y CONFIG_ATMEL_USART=y CONFIG_TIMER=y CONFIG_ATMEL_PIT_TIMER=y +CONFIG_ATMEL_TCB_TIMER=y CONFIG_USB=y CONFIG_USB_EHCI_HCD=y CONFIG_USB_STORAGE=y
configs/sama5d2_xplained_emmc_defconfig +1 -0
@@ -91,6 +91,7 @@ CONFIG_ATMEL_QSPI=y CONFIG_TIMER=y CONFIG_SPL_TIMER=y CONFIG_ATMEL_PIT_TIMER=y +CONFIG_ATMEL_TCB_TIMER=y CONFIG_USB=y CONFIG_USB_EHCI_HCD=y CONFIG_USB_STORAGE=y
configs/sama5d2_xplained_mmc_defconfig +1 -0
@@ -93,6 +93,7 @@ CONFIG_ATMEL_QSPI=y CONFIG_TIMER=y CONFIG_SPL_TIMER=y CONFIG_ATMEL_PIT_TIMER=y +CONFIG_ATMEL_TCB_TIMER=y CONFIG_USB=y CONFIG_USB_EHCI_HCD=y CONFIG_USB_STORAGE=y
configs/sama5d2_xplained_qspiflash_defconfig +1 -0
@@ -92,6 +92,7 @@ CONFIG_ATMEL_QSPI=y CONFIG_TIMER=y CONFIG_SPL_TIMER=y CONFIG_ATMEL_PIT_TIMER=y +CONFIG_ATMEL_TCB_TIMER=y CONFIG_USB=y CONFIG_USB_EHCI_HCD=y CONFIG_USB_STORAGE=y
configs/sama5d2_xplained_spiflash_defconfig +1 -0
@@ -96,6 +96,7 @@ CONFIG_ATMEL_QSPI=y CONFIG_TIMER=y CONFIG_SPL_TIMER=y CONFIG_ATMEL_PIT_TIMER=y +CONFIG_ATMEL_TCB_TIMER=y CONFIG_USB=y CONFIG_USB_EHCI_HCD=y CONFIG_USB_STORAGE=y
Signature · 1 line
2.34.1
Re: [PATCH v5 4/4] configs: sama5d2: enable option CONFIG_ATMEL_TCB_TIMER
Hi Clément,
Quoted 21 lines · Clément Léger
EXTERNAL EMAIL: Do not click links or open attachments unless you know the content is safe
This will allow using the TCB timer instead of the PIT one when running under OP-TEE.
Signed-off-by Clément Léger <[email protected]>--- configs/sama5d2_icp_mmc_defconfig | 1 + configs/sama5d2_ptc_ek_mmc_defconfig | 1 + configs/sama5d2_ptc_ek_nandflash_defconfig | 1 + configs/sama5d2_xplained_emmc_defconfig | 1 + configs/sama5d2_xplained_mmc_defconfig | 1 + configs/sama5d2_xplained_qspiflash_defconfig | 1 + configs/sama5d2_xplained_spiflash_defconfig | 1 + 7 files changed, 7 insertions(+)configs/sama5d2_icp_mmc_defconfig +0
-0
configs/sama5d2_icp_mmc_defconfig CONFIG_ATMEL_USART=y · line 86
Quoted 4 lines
configs/sama5d2_icp_mmc_defconfig +1
-0@@ -83,5 +83,6 @@ CONFIG_ATMEL_USART=y CONFIG_TIMER=y CONFIG_SPL_TIMER=y CONFIG_ATMEL_PIT_TIMER=y +CONFIG_ATMEL_TCB_TIMER=y
As far as I can tell, dm_timer_init() will probe the first available timer if there is no CONFIG_OF_REAL enabled + tick-timer DT entry. With this, having 2 timers enabled in config might not lead to using TCB all the time as it might be expected.
Thank you,
Claudiu Beznea
Quoted 77 lines
configs/sama5d2_icp_mmc_defconfig +0
-0CONFIG_OF_LIBFDT_OVERLAY=y # CONFIG_EFI_LOADER_HII is not setconfigs/sama5d2_ptc_ek_mmc_defconfig +1
-0@@ -72,6 +72,7 @@ CONFIG_DEBUG_UART_ANNOUNCE=y CONFIG_ATMEL_USART=y CONFIG_TIMER=y CONFIG_ATMEL_PIT_TIMER=y +CONFIG_ATMEL_TCB_TIMER=y CONFIG_USB=y CONFIG_USB_EHCI_HCD=y CONFIG_USB_STORAGE=yconfigs/sama5d2_ptc_ek_nandflash_defconfig +1
-0@@ -72,6 +72,7 @@ CONFIG_DEBUG_UART_ANNOUNCE=y CONFIG_ATMEL_USART=y CONFIG_TIMER=y CONFIG_ATMEL_PIT_TIMER=y +CONFIG_ATMEL_TCB_TIMER=y CONFIG_USB=y CONFIG_USB_EHCI_HCD=y CONFIG_USB_STORAGE=yconfigs/sama5d2_xplained_emmc_defconfig +1
-0@@ -91,6 +91,7 @@ CONFIG_ATMEL_QSPI=y CONFIG_TIMER=y CONFIG_SPL_TIMER=y CONFIG_ATMEL_PIT_TIMER=y +CONFIG_ATMEL_TCB_TIMER=y CONFIG_USB=y CONFIG_USB_EHCI_HCD=y CONFIG_USB_STORAGE=yconfigs/sama5d2_xplained_mmc_defconfig +1
-0@@ -93,6 +93,7 @@ CONFIG_ATMEL_QSPI=y CONFIG_TIMER=y CONFIG_SPL_TIMER=y CONFIG_ATMEL_PIT_TIMER=y +CONFIG_ATMEL_TCB_TIMER=y CONFIG_USB=y CONFIG_USB_EHCI_HCD=y CONFIG_USB_STORAGE=yconfigs/sama5d2_xplained_qspiflash_defconfig +1
-0@@ -92,6 +92,7 @@ CONFIG_ATMEL_QSPI=y CONFIG_TIMER=y CONFIG_SPL_TIMER=y CONFIG_ATMEL_PIT_TIMER=y +CONFIG_ATMEL_TCB_TIMER=y CONFIG_USB=y CONFIG_USB_EHCI_HCD=y CONFIG_USB_STORAGE=yconfigs/sama5d2_xplained_spiflash_defconfig +1
-1@@ -96,6 +96,7 @@ CONFIG_ATMEL_QSPI=y CONFIG_TIMER=y CONFIG_SPL_TIMER=y CONFIG_ATMEL_PIT_TIMER=y +CONFIG_ATMEL_TCB_TIMER=y CONFIG_USB=y CONFIG_USB_EHCI_HCD=y CONFIG_USB_STORAGE=y --2.34.1
Re: [PATCH v5 4/4] configs: sama5d2: enable option CONFIG_ATMEL_TCB_TIMER
Quoted 6 lines · Le Mon, 7 Mar 2022 11:06:17 +0000, <[email protected]> a écrit :
Quoted 6 lines
configs/sama5d2_icp_mmc_defconfig +0
-0@@ -83,5 +83,6 @@ CONFIG_ATMEL_USART=y CONFIG_TIMER=y
Quoted 8 lines
configs/sama5d2_icp_mmc_defconfig CONFIG_ATMEL_USART=y · line 86
Quoted 3 lines
configs/sama5d2_icp_mmc_defconfig +1
-0CONFIG_SPL_TIMER=y CONFIG_ATMEL_PIT_TIMER=y +CONFIG_ATMEL_TCB_TIMER=yAs far as I can tell,
dm_timer_init()will probe the first available timer if there is noCONFIG_OF_REALenabled + tick-timer DT entry. With this, having 2 timers enabled in config might not lead to using TCB all the time as it might be expected.
Hi Claudiu,
Hum indeed. Should I remove this commit and let the user select the TCB when needed ?
Thanks,
Quoted 3 lines
Thank you,
Claudiu Beznea
Signature · 3 lines
Clément Léger, Embedded Linux and Kernel engineer at Bootlin https://bootlin.com
Re: [PATCH v5 4/4] configs: sama5d2: enable option CONFIG_ATMEL_TCB_TIMER
Quoted 20 lines · Clément Léger
EXTERNAL EMAIL: Do not click links or open attachments unless you know the content is safe
Quoted 6 lines · Le Mon, 7 Mar 2022 11:06:17 +0000, <[email protected]> a écrit :
Quoted 6 lines
configs/sama5d2_icp_mmc_defconfig +0
-0@@ -83,5 +83,6 @@ CONFIG_ATMEL_USART=y CONFIG_TIMER=yQuoted 8 lines
configs/sama5d2_icp_mmc_defconfig CONFIG_ATMEL_USART=y · line 86
Quoted 3 lines
configs/sama5d2_icp_mmc_defconfig +1
-0CONFIG_SPL_TIMER=y CONFIG_ATMEL_PIT_TIMER=y +CONFIG_ATMEL_TCB_TIMER=yAs far as I can tell,
dm_timer_init()will probe the first available timer if there is noCONFIG_OF_REALenabled + tick-timer DT entry. With this, having 2 timers enabled in config might not lead to using TCB all the time as it might be expected.
Quoted 4 lines
Hi Claudiu,
Hum indeed. Should I remove this commit and let the user select the TCB when needed ?
I would select only one timer in config to be able to work also with Linux DT (if any) as in Linux SAMA5D2 works with TCB as the active/prefered clocksource/clockevent so it is already available in DT. Eugen, is this OK?
Thank you,
Claudiu Beznea
Quoted 12 lines
Thanks,
Quoted 3 lines
Thank you,
Claudiu Beznea--
Clément Léger,
Embedded Linux and Kernel engineer at Bootlin
https://bootlin.com
Re: [PATCH v5 4/4] configs: sama5d2: enable option CONFIG_ATMEL_TCB_TIMER
Quoted 21 lines · On 3/7/22 3:06 PM, Claudiu Beznea - M18063 wrote:
Quoted 20 lines · Clément Léger
EXTERNAL EMAIL: Do not click links or open attachments unless you know the content is safe
Quoted 14 lines · Le Mon, 7 Mar 2022 11:06:17 +0000, <[email protected]> a écrit :
configs/sama5d2_icp_mmc_defconfig CONFIG_ATMEL_USART=y · line 86
Quoted 9 lines
configs/sama5d2_icp_mmc_defconfig +1
-0@@ -83,5 +83,6 @@ CONFIG_ATMEL_USART=y CONFIG_TIMER=y CONFIG_SPL_TIMER=y CONFIG_ATMEL_PIT_TIMER=y +CONFIG_ATMEL_TCB_TIMER=yAs far as I can tell,
dm_timer_init()will probe the first available timer if there is noCONFIG_OF_REALenabled + tick-timer DT entry. With this, having 2 timers enabled in config might not lead to using TCB all the time as it might be expected.
Quoted 8 lines
Quoted 4 lines
Hi Claudiu,
Hum indeed. Should I remove this commit and let the user select the TCB when needed ?
I would select only one timer in config to be able to work also with Linux DT (if any) as in Linux SAMA5D2 works with TCB as the active/prefered clocksource/clockevent so it is already available in DT. Eugen, is this OK?
Hi,
In defconfig you are free to select which driver you want to build. (both, only one, or none).
In DT, you have to describe the board and SoC as in Linux.
Quoted 17 lines
Thank you,
Claudiu BezneaQuoted 12 lines
Thanks,
Quoted 3 lines
Thank you,
Claudiu Beznea--
Clément Léger,
Embedded Linux and Kernel engineer at Bootlin
https://bootlin.com