Linaro-acpi May 2013

linaro-acpi@lists.linaro.org

4 participants
10 discussions

[PATCH 0/5] Prototype i2c bus probe from ACPI

by Graeme Gregory

This patch series implements probing the s3c24xx i2c bus device from ACPI tables. I have included an update to the DSDT for the correct information for I2C0 which is the port connected to the PMIC so is easy to test with i2c tools. On boot I can successfully communicate with the slave at 0x66 on this bus. root@arndale:~# i2cdump -f -y 0 0x66 No size specified (using byte-data access) 0 1 2 3 4 5 6 7 8 9 a b c d e f 0123456789abcdef 00: 03 00 00 00 c0 78 00 01 17 07 01 0e e5 85 4f de ?...?x.???????O? 10: 03 03 03 23 23 23 23 23 23 23 23 23 23 23 23 23 ???############# 20: 23 23 23 23 23 23 23 23 23 23 23 23 23 23 23 00 ###############. 30: 00 45 d9 38 fa 34 00 50 50 50 50 50 50 db 40 00 .E?8?4.PPPPPP?@. 40: 40 40 40 40 40 40 3a 40 00 40 40 40 40 40 40 d9 @@@@@@:@.@@@@@@? 50: 58 70 88 b8 18 58 d8 68 d8 30 98 a8 cc d0 16 1c Xp???X?h?0?????? 60: 28 d4 d4 d4 cc cc c8 2c d4 d4 ec d4 d4 c8 d4 e8 (??????,???????? 70: 28 2c 2c 2c 32 28 2c 08 14 14 14 00 00 00 00 00 (,,,2(,????..... 80: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................ 90: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................ a0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................ b0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................ c0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................ d0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................ e0: 01 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ?............... f0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................ can see the correct IRQ allocated root@arndale:~# cat /proc/interrupts CPU0 CPU1 29: 0 0 GIC arch_timer 30: 24177 6574 GIC arch_timer 66: 0 0 GIC dma-pl330.0 67: 0 0 GIC dma-pl330.1 85: 538 0 GIC exynos4210-uart.2 88: 1024 0 GIC LINA0001:00 97: 0 0 GIC exynos-tmu 103: 11813 0 GIC ehci_hcd:usb1, ohci_hcd:usb2 104: 0 0 GIC xhci-hcd:usb3 107: 432 0 GIC dw-mci 109: 67422 0 GIC dw-mci 126: 0 0 GIC drm_mixer 152: 0 0 GIC mct_tick0_irq 153: 0 0 GIC mct_tick1_irq 156: 0 0 GIC dma-pl330.2 347: 0 0 COMBINER mct_comp_irq IPI0: 0 1 CPU wakeup interrupts IPI1: 0 0 Timer broadcast interrupts IPI2: 5233 6766 Rescheduling interrupts IPI3: 0 0 Function call interrupts IPI4: 98 44 Single function call interrupts IPI5: 0 0 CPU stop interrupts IPI6: 0 0 CPU backtrace Err: 0 and can see the correct GPIO allocated. GPIOs 42-45, GPB3: gpio-42 (i2c-bus ) in hi gpio-43 (i2c-bus ) in hi The GPIO allocation is not how it would be in a final implementation as similar to FDT the GPIO providers (like GIC) can be hierachical so you would normally use the device name + offset to locate the correct GPIO. This is very similar to how it works in FDT as well. Thanks Graeme

10 years, 10 months

[PATCH 0/4] ARM:ACPI: CPU topology support for ACPI processor driver

by Hanjun Guo

This patch set is about to implement the CPU topology support for ACPI processor driver, it is compiled ok, but not tested on the hardware; This part of the patch set is the os lever code for ACPI processor driver, and ASL code will send out in another patch set. (not finished yet) For the ACPI processor driver, when cpu is hot added, arch_fix_phys_package_id(), acpi_map_lsapic() and arch_register_cpu() are arch specific; When cpu is hot removed, arch_unregister_cpu()/acpi_unmap_lsapic() are arch specific. arch_register_cpu() will register the logic cpu on sysfs, create topology information interface, and create online/offline interface for cpu hotplug. unregister is doing the reverse. Hanjun Guo (4): ARM:ACPI: introduce arch_fix_phys_package_id() for armv7 ARM:ACPI: move topology_init() to arch/arm/kernel/topology.c ARM:ACPI: introduce arch_register_cpu() and arch_unregister_cpu() ARM:ACPI: remove CONFIG_X86 in processor_driver.c arch/arm/include/asm/cpu.h | 6 ++++ arch/arm/include/asm/topology.h | 1 + arch/arm/kernel/setup.c | 21 --------------- arch/arm/kernel/topology.c | 54 +++++++++++++++++++++++++++++++++++++++ drivers/acpi/processor_driver.c | 9 ------ 5 files changed, 61 insertions(+), 30 deletions(-)

10 years, 10 months

[PATCH 2/4] ARM:ACPI: move topology_init() to arch/arm/kernel/topology.c

by Hanjun Guo

move topology_init() to where it belongs to. Signed-off-by: Hanjun Guo <hanjun.guo(a)linaro.org> --- arch/arm/kernel/setup.c | 21 --------------------- arch/arm/kernel/topology.c | 21 +++++++++++++++++++++ 2 files changed, 21 insertions(+), 21 deletions(-) diff --git a/arch/arm/kernel/setup.c b/arch/arm/kernel/setup.c index 7b3d3d3..b00bb7a 100644 --- a/arch/arm/kernel/setup.c +++ b/arch/arm/kernel/setup.c @@ -859,27 +859,6 @@ void __init setup_arch(char **cmdline_p) #endif } - -static int __init topology_init(void) -{ - int cpu; - -#ifdef CONFIG_NUMA - int node; - for_each_online_node(node) - register_one_node(node); -#endif - - for_each_possible_cpu(cpu) { - struct cpuinfo_arm *cpuinfo = &per_cpu(cpu_data, cpu); - cpuinfo->cpu.hotpluggable = 1; - register_cpu(&cpuinfo->cpu, cpu); - } - - return 0; -} -subsys_initcall(topology_init); - #ifdef CONFIG_HAVE_PROC_CPU static int __init proc_cpu_init(void) { diff --git a/arch/arm/kernel/topology.c b/arch/arm/kernel/topology.c index d951ae2..dcc551f 100644 --- a/arch/arm/kernel/topology.c +++ b/arch/arm/kernel/topology.c @@ -25,6 +25,7 @@ #include <asm/cputype.h> #include <asm/smp_plat.h> #include <asm/topology.h> +#include <asm/cpu.h> /* * cpu power scale management @@ -455,3 +456,23 @@ void __init init_cpu_topology(void) parse_dt_topology(); } + +static int __init topology_init(void) +{ + int cpu; + +#ifdef CONFIG_NUMA + int node; + for_each_online_node(node) + register_one_node(node); +#endif + + for_each_possible_cpu(cpu) { + struct cpuinfo_arm *cpuinfo = &per_cpu(cpu_data, cpu); + cpuinfo->cpu.hotpluggable = 1; + register_cpu(&cpuinfo->cpu, cpu); + } + + return 0; +} +subsys_initcall(topology_init); -- 1.7.1

10 years, 10 months

[PATCH 4/4] ARM:ACPI: remove CONFIG_X86 in processor_driver.c

by Hanjun Guo

Since the fuctions needed by the ACPI processor driver are implemented, just remove CONFIG_X86. Signed-off-by: Hanjun Guo <hanjun.guo(a)linaro.org> --- drivers/acpi/processor_driver.c | 9 --------- 1 files changed, 0 insertions(+), 9 deletions(-) diff --git a/drivers/acpi/processor_driver.c b/drivers/acpi/processor_driver.c index 0fe57d3..bec717f 100644 --- a/drivers/acpi/processor_driver.c +++ b/drivers/acpi/processor_driver.c @@ -374,12 +374,9 @@ static int acpi_processor_get_info(struct acpi_device *device) * ensure we get the right value in the "physical id" field * of /proc/cpuinfo */ -#ifdef CONFIG_X86 - /* BOZO: abstract out? */ status = acpi_evaluate_object(pr->handle, "_SUN", NULL, &buffer); if (ACPI_SUCCESS(status)) arch_fix_phys_package_id(pr->id, object.integer.value); -#endif return 0; } @@ -841,13 +838,10 @@ static acpi_status acpi_processor_hotadd_init(struct acpi_processor *pr) if (acpi_map_lsapic(handle, &pr->id)) return AE_ERROR; -#ifdef CONFIG_X86 - /* BOZO: abstract out? */ if (arch_register_cpu(pr->id)) { acpi_unmap_lsapic(pr->id); return AE_ERROR; } -#endif /* CPU got hot-plugged, but cpu_data is not initialized yet * Set flag to delay cpu_idle/throttling initialization @@ -881,10 +875,7 @@ static int acpi_processor_handle_eject(struct acpi_processor *pr) "brought the CPU back online\n", pr->id); return -EAGAIN; } -#ifdef CONFIG_X86 - /* BOZO: abstract out? */ arch_unregister_cpu(pr->id); -#endif acpi_unmap_lsapic(pr->id); put_online_cpus(); return (0); -- 1.7.1

10 years, 10 months

[PATCH 3/4] ARM:ACPI: introduce arch_register_cpu() and arch_unregister_cpu()

by Hanjun Guo

Introduce arch_register_cpu() and arch_unregister_cpu() for ACPI processor driver. Signed-off-by: Hanjun Guo <hanjun.guo(a)linaro.org> --- arch/arm/include/asm/cpu.h | 6 ++++++ arch/arm/kernel/topology.c | 33 ++++++++++++++++++++++++++++----- 2 files changed, 34 insertions(+), 5 deletions(-) diff --git a/arch/arm/include/asm/cpu.h b/arch/arm/include/asm/cpu.h index 2744f06..a72f974 100644 --- a/arch/arm/include/asm/cpu.h +++ b/arch/arm/include/asm/cpu.h @@ -12,6 +12,7 @@ #include <linux/percpu.h> #include <linux/cpu.h> +#include <linux/topology.h> struct cpuinfo_arm { struct cpu cpu; @@ -21,6 +22,11 @@ struct cpuinfo_arm { #endif }; +#ifdef CONFIG_HOTPLUG_CPU +extern int arch_register_cpu(int cpu); +extern void arch_unregister_cpu(int cpu); +#endif + DECLARE_PER_CPU(struct cpuinfo_arm, cpu_data); #endif diff --git a/arch/arm/kernel/topology.c b/arch/arm/kernel/topology.c index dcc551f..4e03ee4 100644 --- a/arch/arm/kernel/topology.c +++ b/arch/arm/kernel/topology.c @@ -457,6 +457,32 @@ void __init init_cpu_topology(void) parse_dt_topology(); } +#ifdef CONFIG_HOTPLUG_CPU +int __ref arch_register_cpu(int cpu) +{ + struct cpuinfo_arm *cpuinfo = &per_cpu(cpu_data, cpu); + + /* BSP cann't be taken down on arm */ + if (cpu) + cpuinfo->cpu.hotpluggable = 1; + + return register_cpu(&cpuinfo->cpu, cpu); +} +EXPORT_SYMBOL(arch_register_cpu); + +void arch_unregister_cpu(int cpu) +{ + unregister_cpu(&per_cpu(cpu_data, cpu).cpu); +} +EXPORT_SYMBOL(arch_unregister_cpu); +#else /* CONFIG_HOTPLUG_CPU */ + +static int __init arch_register_cpu(int cpu) +{ + return register_cpu(&per_cpu(cpu_data, cpu).cpu, cpu); +} +#endif /* CONFIG_HOTPLUG_CPU */ + static int __init topology_init(void) { int cpu; @@ -467,11 +493,8 @@ static int __init topology_init(void) register_one_node(node); #endif - for_each_possible_cpu(cpu) { - struct cpuinfo_arm *cpuinfo = &per_cpu(cpu_data, cpu); - cpuinfo->cpu.hotpluggable = 1; - register_cpu(&cpuinfo->cpu, cpu); - } + for_each_possible_cpu(cpu) + arch_register_cpu(cpu); return 0; } -- 1.7.1

10 years, 10 months

[PATCH 1/4] ARM:ACPI: introduce arch_fix_phys_package_id() for armv7

by Hanjun Guo

arch_fix_phys_package_id() will use the slot number provided by ACPI, then we can get the right value in the "physical id" field of /proc/cpuinfo. Signed-off-by: Hanjun Guo <hanjun.guo(a)linaro.org> --- arch/arm/include/asm/topology.h | 1 + arch/arm/kernel/topology.c | 10 ++++++++++ 2 files changed, 11 insertions(+), 0 deletions(-) diff --git a/arch/arm/include/asm/topology.h b/arch/arm/include/asm/topology.h index 5357eb1..932b75b 100644 --- a/arch/arm/include/asm/topology.h +++ b/arch/arm/include/asm/topology.h @@ -27,6 +27,7 @@ void init_cpu_topology(void); void store_cpu_topology(unsigned int cpuid); const struct cpumask *cpu_coregroup_mask(int cpu); int cluster_to_logical_mask(unsigned int socket_id, cpumask_t *cluster_mask); +extern void arch_fix_phys_package_id(int num, u32 slot); #ifdef CONFIG_DISABLE_CPU_SCHED_DOMAIN_BALANCE /* Common values for CPUs */ diff --git a/arch/arm/kernel/topology.c b/arch/arm/kernel/topology.c index 8df74d6..d951ae2 100644 --- a/arch/arm/kernel/topology.c +++ b/arch/arm/kernel/topology.c @@ -290,6 +290,16 @@ void store_cpu_topology(unsigned int cpuid) cpu_topology[cpuid].socket_id, mpidr); } +void arch_fix_phys_package_id(int num, u32 slot) +{ +#ifdef CONFIG_SMP + struct cputopo_arm *cpuid_topo = &cpu_topology[num]; + + if (cpuid_topo->socket_id == -1) + cpuid_topo->socket_id = slot; +#endif +} +EXPORT_SYMBOL_GPL(arch_fix_phys_package_id); #ifdef CONFIG_SCHED_HMP -- 1.7.1

10 years, 10 months

[PATCH 0/2] Add handling of more fields in s3c24xx-i2c

by Graeme Gregory

There were three fields used in the DT version of probing this driver which were not yet in ACPI so handle those fields as well. samsung,i2c-sda-delay = <100>; samsung,i2c-max-bus-freq = <20000>; samsung,i2c-slave-addr = <0x66>; becomes Method (DLAY, 0x0, NotSerialized) { Return (100) } Method (SADD, 0x0, NotSerialized) { Return (0x66) } Method (FREQ, 0x0, NotSerialized) { Return (20000) } Thanks Graeme

10 years, 11 months

[PATCH] ACPICA: Remove redundant macro defines for ACPI_DEBUG_PRINT()

by Hanjun Guo

In commit 0377b5acba2 (ACPICA: Merge all debug output macros into a single file, acoutput.), the patch introdices redundant macro defines for ACPI_DEBUG_PRINT() and ACPI_DEBUG_PRINT_RAW(), just remove it. I will send this to upstream. Signed-off-by: Hanjun Guo <hanjun.guo(a)linaro.org> --- include/acpi/acoutput.h | 2 -- 1 file changed, 2 deletions(-) diff --git a/include/acpi/acoutput.h b/include/acpi/acoutput.h index 4f52ea7..e1c7474 100644 --- a/include/acpi/acoutput.h +++ b/include/acpi/acoutput.h @@ -447,8 +447,6 @@ #define ACPI_DUMP_TABLES(a, b) #define ACPI_DUMP_PATHNAME(a, b, c, d) #define ACPI_DUMP_BUFFER(a, b) -#define ACPI_DEBUG_PRINT(pl) -#define ACPI_DEBUG_PRINT_RAW(pl) #define ACPI_IS_DEBUG_ENABLED(level, component) 0 /* Return macros must have a return statement at the minimum */ -- 1.8.2.1

10 years, 11 months

[PATCH 0/1] start to make use of irq domains

by Graeme Gregory

This series is just a start of attaching the ACPI code to the IRQ domain so we map IRQs in a method compatible with FDT booting. Firstly get the GIC to register itself as the default domain. Then change the ARM ACPI code to use default domain to map IRQs. Graeme

10 years, 11 months

Patches for Al to help debug memory alloc issue!

by Graeme Gregory

Hi Al, These two patches will give some extra ACPI debug output in u-boot so you can see the start/end of the ACPI address allocation. You can also see the size and physical address of each ram bank. Hopefully this helps you track down the issue. My DSDT changes are not big enough to cross the 0x50000000 boundary yet. Thanks Graeme

10 years, 11 months

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

Linaro-acpi May 2013