diff --git a/arch/arm/include/asm/arch_timer.h b/arch/arm/include/asm/arch_timer.h index 62e75475e57..d40229d9a1c 100644 --- a/arch/arm/include/asm/arch_timer.h +++ b/arch/arm/include/asm/arch_timer.h @@ -2,11 +2,12 @@ #define __ASMARM_ARCH_TIMER_H #include +#include #ifdef CONFIG_ARM_ARCH_TIMER -#define ARCH_HAS_READ_CURRENT_TIMER int arch_timer_of_register(void); int arch_timer_sched_clock_init(void); +struct timecounter *arch_timer_get_timecounter(void); #else static inline int arch_timer_of_register(void) { @@ -17,6 +18,11 @@ static inline int arch_timer_sched_clock_init(void) { return -ENXIO; } + +static inline struct timecounter *arch_timer_get_timecounter(void) +{ + return NULL; +} #endif #endif diff --git a/arch/arm/include/asm/delay.h b/arch/arm/include/asm/delay.h index dc6145120de..ab98fdd083b 100644 --- a/arch/arm/include/asm/delay.h +++ b/arch/arm/include/asm/delay.h @@ -15,6 +15,11 @@ #ifndef __ASSEMBLY__ +struct delay_timer { + unsigned long (*read_current_timer)(void); + unsigned long freq; +}; + extern struct arm_delay_ops { void (*delay)(unsigned long); void (*const_udelay)(unsigned long); @@ -56,6 +61,10 @@ extern void __loop_delay(unsigned long loops); extern void __loop_udelay(unsigned long usecs); extern void __loop_const_udelay(unsigned long); +/* Delay-loop timer registration. */ +#define ARCH_HAS_READ_CURRENT_TIMER +extern void register_current_timer_delay(const struct delay_timer *timer); + #endif /* __ASSEMBLY__ */ #endif /* defined(_ARM_DELAY_H) */ diff --git a/arch/arm/include/asm/timex.h b/arch/arm/include/asm/timex.h index 5e711722ebf..9acc135dad9 100644 --- a/arch/arm/include/asm/timex.h +++ b/arch/arm/include/asm/timex.h @@ -12,13 +12,9 @@ #ifndef _ASMARM_TIMEX_H #define _ASMARM_TIMEX_H -#include #include -#ifdef ARCH_HAS_READ_CURRENT_TIMER +typedef unsigned long cycles_t; #define get_cycles() ({ cycles_t c; read_current_timer(&c) ? 0 : c; }) -#endif - -#include #endif diff --git a/arch/arm/kernel/arch_timer.c b/arch/arm/kernel/arch_timer.c index cf258807160..c8ef20747ee 100644 --- a/arch/arm/kernel/arch_timer.c +++ b/arch/arm/kernel/arch_timer.c @@ -21,18 +21,28 @@ #include #include +#include #include #include #include #include static unsigned long arch_timer_rate; -static int arch_timer_ppi; -static int arch_timer_ppi2; + +enum ppi_nr { + PHYS_SECURE_PPI, + PHYS_NONSECURE_PPI, + VIRT_PPI, + HYP_PPI, + MAX_TIMER_PPI +}; + +static int arch_timer_ppi[MAX_TIMER_PPI]; static struct clock_event_device __percpu **arch_timer_evt; +static struct delay_timer arch_delay_timer; -extern void init_current_timer_delay(unsigned long freq); +static bool arch_timer_use_virtual = true; /* * Architected system timer support. @@ -46,50 +56,104 @@ extern void init_current_timer_delay(unsigned long freq); #define ARCH_TIMER_REG_FREQ 1 #define ARCH_TIMER_REG_TVAL 2 -static void arch_timer_reg_write(int reg, u32 val) +#define ARCH_TIMER_PHYS_ACCESS 0 +#define ARCH_TIMER_VIRT_ACCESS 1 + +/* + * These register accessors are marked inline so the compiler can + * nicely work out which register we want, and chuck away the rest of + * the code. At least it does so with a recent GCC (4.6.3). + */ +static inline void arch_timer_reg_write(const int access, const int reg, u32 val) { - switch (reg) { - case ARCH_TIMER_REG_CTRL: - asm volatile("mcr p15, 0, %0, c14, c2, 1" : : "r" (val)); - break; - case ARCH_TIMER_REG_TVAL: - asm volatile("mcr p15, 0, %0, c14, c2, 0" : : "r" (val)); - break; + if (access == ARCH_TIMER_PHYS_ACCESS) { + switch (reg) { + case ARCH_TIMER_REG_CTRL: + asm volatile("mcr p15, 0, %0, c14, c2, 1" : : "r" (val)); + break; + case ARCH_TIMER_REG_TVAL: + asm volatile("mcr p15, 0, %0, c14, c2, 0" : : "r" (val)); + break; + } + } + + if (access == ARCH_TIMER_VIRT_ACCESS) { + switch (reg) { + case ARCH_TIMER_REG_CTRL: + asm volatile("mcr p15, 0, %0, c14, c3, 1" : : "r" (val)); + break; + case ARCH_TIMER_REG_TVAL: + asm volatile("mcr p15, 0, %0, c14, c3, 0" : : "r" (val)); + break; + } } isb(); } -static u32 arch_timer_reg_read(int reg) +static inline u32 arch_timer_reg_read(const int access, const int reg) { - u32 val; + u32 val = 0; - switch (reg) { - case ARCH_TIMER_REG_CTRL: - asm volatile("mrc p15, 0, %0, c14, c2, 1" : "=r" (val)); - break; - case ARCH_TIMER_REG_FREQ: - asm volatile("mrc p15, 0, %0, c14, c0, 0" : "=r" (val)); - break; - case ARCH_TIMER_REG_TVAL: - asm volatile("mrc p15, 0, %0, c14, c2, 0" : "=r" (val)); - break; - default: - BUG(); + if (access == ARCH_TIMER_PHYS_ACCESS) { + switch (reg) { + case ARCH_TIMER_REG_CTRL: + asm volatile("mrc p15, 0, %0, c14, c2, 1" : "=r" (val)); + break; + case ARCH_TIMER_REG_TVAL: + asm volatile("mrc p15, 0, %0, c14, c2, 0" : "=r" (val)); + break; + case ARCH_TIMER_REG_FREQ: + asm volatile("mrc p15, 0, %0, c14, c0, 0" : "=r" (val)); + break; + } + } + + if (access == ARCH_TIMER_VIRT_ACCESS) { + switch (reg) { + case ARCH_TIMER_REG_CTRL: + asm volatile("mrc p15, 0, %0, c14, c3, 1" : "=r" (val)); + break; + case ARCH_TIMER_REG_TVAL: + asm volatile("mrc p15, 0, %0, c14, c3, 0" : "=r" (val)); + break; + } } return val; } -static irqreturn_t arch_timer_handler(int irq, void *dev_id) +static inline cycle_t arch_timer_counter_read(const int access) { - struct clock_event_device *evt = *(struct clock_event_device **)dev_id; - unsigned long ctrl; + cycle_t cval = 0; - ctrl = arch_timer_reg_read(ARCH_TIMER_REG_CTRL); + if (access == ARCH_TIMER_PHYS_ACCESS) + asm volatile("mrrc p15, 0, %Q0, %R0, c14" : "=r" (cval)); + + if (access == ARCH_TIMER_VIRT_ACCESS) + asm volatile("mrrc p15, 1, %Q0, %R0, c14" : "=r" (cval)); + + return cval; +} + +static inline cycle_t arch_counter_get_cntpct(void) +{ + return arch_timer_counter_read(ARCH_TIMER_PHYS_ACCESS); +} + +static inline cycle_t arch_counter_get_cntvct(void) +{ + return arch_timer_counter_read(ARCH_TIMER_VIRT_ACCESS); +} + +static irqreturn_t inline timer_handler(const int access, + struct clock_event_device *evt) +{ + unsigned long ctrl; + ctrl = arch_timer_reg_read(access, ARCH_TIMER_REG_CTRL); if (ctrl & ARCH_TIMER_CTRL_IT_STAT) { ctrl |= ARCH_TIMER_CTRL_IT_MASK; - arch_timer_reg_write(ARCH_TIMER_REG_CTRL, ctrl); + arch_timer_reg_write(access, ARCH_TIMER_REG_CTRL, ctrl); evt->event_handler(evt); return IRQ_HANDLED; } @@ -97,63 +161,100 @@ static irqreturn_t arch_timer_handler(int irq, void *dev_id) return IRQ_NONE; } -static void arch_timer_disable(void) +static irqreturn_t arch_timer_handler_virt(int irq, void *dev_id) { - unsigned long ctrl; + struct clock_event_device *evt = *(struct clock_event_device **)dev_id; - ctrl = arch_timer_reg_read(ARCH_TIMER_REG_CTRL); - ctrl &= ~ARCH_TIMER_CTRL_ENABLE; - arch_timer_reg_write(ARCH_TIMER_REG_CTRL, ctrl); + return timer_handler(ARCH_TIMER_VIRT_ACCESS, evt); } -static void arch_timer_set_mode(enum clock_event_mode mode, - struct clock_event_device *clk) +static irqreturn_t arch_timer_handler_phys(int irq, void *dev_id) { + struct clock_event_device *evt = *(struct clock_event_device **)dev_id; + + return timer_handler(ARCH_TIMER_PHYS_ACCESS, evt); +} + +static inline void timer_set_mode(const int access, int mode) +{ + unsigned long ctrl; switch (mode) { case CLOCK_EVT_MODE_UNUSED: case CLOCK_EVT_MODE_SHUTDOWN: - arch_timer_disable(); + ctrl = arch_timer_reg_read(access, ARCH_TIMER_REG_CTRL); + ctrl &= ~ARCH_TIMER_CTRL_ENABLE; + arch_timer_reg_write(access, ARCH_TIMER_REG_CTRL, ctrl); break; default: break; } } -static int arch_timer_set_next_event(unsigned long evt, - struct clock_event_device *unused) +static void arch_timer_set_mode_virt(enum clock_event_mode mode, + struct clock_event_device *clk) +{ + timer_set_mode(ARCH_TIMER_VIRT_ACCESS, mode); +} + +static void arch_timer_set_mode_phys(enum clock_event_mode mode, + struct clock_event_device *clk) +{ + timer_set_mode(ARCH_TIMER_PHYS_ACCESS, mode); +} + +static inline void set_next_event(const int access, unsigned long evt) { unsigned long ctrl; - - ctrl = arch_timer_reg_read(ARCH_TIMER_REG_CTRL); + ctrl = arch_timer_reg_read(access, ARCH_TIMER_REG_CTRL); ctrl |= ARCH_TIMER_CTRL_ENABLE; ctrl &= ~ARCH_TIMER_CTRL_IT_MASK; + arch_timer_reg_write(access, ARCH_TIMER_REG_TVAL, evt); + arch_timer_reg_write(access, ARCH_TIMER_REG_CTRL, ctrl); +} - arch_timer_reg_write(ARCH_TIMER_REG_TVAL, evt); - arch_timer_reg_write(ARCH_TIMER_REG_CTRL, ctrl); +static int arch_timer_set_next_event_virt(unsigned long evt, + struct clock_event_device *unused) +{ + set_next_event(ARCH_TIMER_VIRT_ACCESS, evt); + return 0; +} +static int arch_timer_set_next_event_phys(unsigned long evt, + struct clock_event_device *unused) +{ + set_next_event(ARCH_TIMER_PHYS_ACCESS, evt); return 0; } static int __cpuinit arch_timer_setup(struct clock_event_device *clk) { - /* Be safe... */ - arch_timer_disable(); - clk->features = CLOCK_EVT_FEAT_ONESHOT | CLOCK_EVT_FEAT_C3STOP; clk->name = "arch_sys_timer"; clk->rating = 450; - clk->set_mode = arch_timer_set_mode; - clk->set_next_event = arch_timer_set_next_event; - clk->irq = arch_timer_ppi; + if (arch_timer_use_virtual) { + clk->irq = arch_timer_ppi[VIRT_PPI]; + clk->set_mode = arch_timer_set_mode_virt; + clk->set_next_event = arch_timer_set_next_event_virt; + } else { + clk->irq = arch_timer_ppi[PHYS_SECURE_PPI]; + clk->set_mode = arch_timer_set_mode_phys; + clk->set_next_event = arch_timer_set_next_event_phys; + } + + clk->set_mode(CLOCK_EVT_MODE_SHUTDOWN, NULL); clockevents_config_and_register(clk, arch_timer_rate, 0xf, 0x7fffffff); *__this_cpu_ptr(arch_timer_evt) = clk; - enable_percpu_irq(clk->irq, 0); - if (arch_timer_ppi2) - enable_percpu_irq(arch_timer_ppi2, 0); + if (arch_timer_use_virtual) + enable_percpu_irq(arch_timer_ppi[VIRT_PPI], 0); + else { + enable_percpu_irq(arch_timer_ppi[PHYS_SECURE_PPI], 0); + if (arch_timer_ppi[PHYS_NONSECURE_PPI]) + enable_percpu_irq(arch_timer_ppi[PHYS_NONSECURE_PPI], 0); + } return 0; } @@ -173,8 +274,8 @@ static int arch_timer_available(void) return -ENXIO; if (arch_timer_rate == 0) { - arch_timer_reg_write(ARCH_TIMER_REG_CTRL, 0); - freq = arch_timer_reg_read(ARCH_TIMER_REG_FREQ); + freq = arch_timer_reg_read(ARCH_TIMER_PHYS_ACCESS, + ARCH_TIMER_REG_FREQ); /* Check the timer frequency. */ if (freq == 0) { @@ -185,52 +286,57 @@ static int arch_timer_available(void) arch_timer_rate = freq; } - pr_info_once("Architected local timer running at %lu.%02luMHz.\n", - arch_timer_rate / 1000000, (arch_timer_rate / 10000) % 100); + pr_info_once("Architected local timer running at %lu.%02luMHz (%s).\n", + arch_timer_rate / 1000000, (arch_timer_rate / 10000) % 100, + arch_timer_use_virtual ? "virt" : "phys"); return 0; } -static inline cycle_t arch_counter_get_cntpct(void) +static u32 notrace arch_counter_get_cntpct32(void) { - u32 cvall, cvalh; - - asm volatile("mrrc p15, 0, %0, %1, c14" : "=r" (cvall), "=r" (cvalh)); - - return ((cycle_t) cvalh << 32) | cvall; -} - -static inline cycle_t arch_counter_get_cntvct(void) -{ - u32 cvall, cvalh; - - asm volatile("mrrc p15, 1, %0, %1, c14" : "=r" (cvall), "=r" (cvalh)); - - return ((cycle_t) cvalh << 32) | cvall; -} - -static u32 notrace arch_counter_get_cntvct32(void) -{ - cycle_t cntvct = arch_counter_get_cntvct(); + cycle_t cnt = arch_counter_get_cntpct(); /* * The sched_clock infrastructure only knows about counters * with at most 32bits. Forget about the upper 24 bits for the * time being... */ - return (u32)(cntvct & (u32)~0); + return (u32)cnt; +} + +static u32 notrace arch_counter_get_cntvct32(void) +{ + cycle_t cnt = arch_counter_get_cntvct(); + + /* + * The sched_clock infrastructure only knows about counters + * with at most 32bits. Forget about the upper 24 bits for the + * time being... + */ + return (u32)cnt; } static cycle_t arch_counter_read(struct clocksource *cs) { + /* + * Always use the physical counter for the clocksource. + * CNTHCTL.PL1PCTEN must be set to 1. + */ return arch_counter_get_cntpct(); } -int read_current_timer(unsigned long *timer_val) +static unsigned long arch_timer_read_current_timer(void) { - if (!arch_timer_rate) - return -ENXIO; - *timer_val = arch_counter_get_cntpct(); - return 0; + return arch_counter_get_cntpct(); +} + +static cycle_t arch_counter_read_cc(const struct cyclecounter *cc) +{ + /* + * Always use the physical counter for the clocksource. + * CNTHCTL.PL1PCTEN must be set to 1. + */ + return arch_counter_get_cntpct(); } static struct clocksource clocksource_counter = { @@ -241,14 +347,32 @@ static struct clocksource clocksource_counter = { .flags = CLOCK_SOURCE_IS_CONTINUOUS, }; +static struct cyclecounter cyclecounter = { + .read = arch_counter_read_cc, + .mask = CLOCKSOURCE_MASK(56), +}; + +static struct timecounter timecounter; + +struct timecounter *arch_timer_get_timecounter(void) +{ + return &timecounter; +} + static void __cpuinit arch_timer_stop(struct clock_event_device *clk) { pr_debug("arch_timer_teardown disable IRQ%d cpu #%d\n", clk->irq, smp_processor_id()); - disable_percpu_irq(clk->irq); - if (arch_timer_ppi2) - disable_percpu_irq(arch_timer_ppi2); - arch_timer_set_mode(CLOCK_EVT_MODE_UNUSED, clk); + + if (arch_timer_use_virtual) + disable_percpu_irq(arch_timer_ppi[VIRT_PPI]); + else { + disable_percpu_irq(arch_timer_ppi[PHYS_SECURE_PPI]); + if (arch_timer_ppi[PHYS_NONSECURE_PPI]) + disable_percpu_irq(arch_timer_ppi[PHYS_NONSECURE_PPI]); + } + + clk->set_mode(CLOCK_EVT_MODE_UNUSED, clk); } static struct local_timer_ops arch_timer_ops __cpuinitdata = { @@ -261,36 +385,48 @@ static struct clock_event_device arch_timer_global_evt; static int __init arch_timer_register(void) { int err; + int ppi; err = arch_timer_available(); if (err) - return err; + goto out; arch_timer_evt = alloc_percpu(struct clock_event_device *); - if (!arch_timer_evt) - return -ENOMEM; - - clocksource_register_hz(&clocksource_counter, arch_timer_rate); - - err = request_percpu_irq(arch_timer_ppi, arch_timer_handler, - "arch_timer", arch_timer_evt); - if (err) { - pr_err("arch_timer: can't register interrupt %d (%d)\n", - arch_timer_ppi, err); - goto out_free; + if (!arch_timer_evt) { + err = -ENOMEM; + goto out; } - if (arch_timer_ppi2) { - err = request_percpu_irq(arch_timer_ppi2, arch_timer_handler, + clocksource_register_hz(&clocksource_counter, arch_timer_rate); + cyclecounter.mult = clocksource_counter.mult; + cyclecounter.shift = clocksource_counter.shift; + timecounter_init(&timecounter, &cyclecounter, + arch_counter_get_cntpct()); + + if (arch_timer_use_virtual) { + ppi = arch_timer_ppi[VIRT_PPI]; + err = request_percpu_irq(ppi, arch_timer_handler_virt, "arch_timer", arch_timer_evt); - if (err) { - pr_err("arch_timer: can't register interrupt %d (%d)\n", - arch_timer_ppi2, err); - arch_timer_ppi2 = 0; - goto out_free_irq; + } else { + ppi = arch_timer_ppi[PHYS_SECURE_PPI]; + err = request_percpu_irq(ppi, arch_timer_handler_phys, + "arch_timer", arch_timer_evt); + if (!err && arch_timer_ppi[PHYS_NONSECURE_PPI]) { + ppi = arch_timer_ppi[PHYS_NONSECURE_PPI]; + err = request_percpu_irq(ppi, arch_timer_handler_phys, + "arch_timer", arch_timer_evt); + if (err) + free_percpu_irq(arch_timer_ppi[PHYS_SECURE_PPI], + arch_timer_evt); } } + if (err) { + pr_err("arch_timer: can't register interrupt %d (%d)\n", + ppi, err); + goto out_free; + } + err = local_timer_register(&arch_timer_ops); if (err) { /* @@ -302,21 +438,29 @@ static int __init arch_timer_register(void) arch_timer_global_evt.cpumask = cpumask_of(0); err = arch_timer_setup(&arch_timer_global_evt); } - if (err) goto out_free_irq; - init_current_timer_delay(arch_timer_rate); + /* Use the architected timer for the delay loop. */ + arch_delay_timer.read_current_timer = &arch_timer_read_current_timer; + arch_delay_timer.freq = arch_timer_rate; + register_current_timer_delay(&arch_delay_timer); return 0; out_free_irq: - free_percpu_irq(arch_timer_ppi, arch_timer_evt); - if (arch_timer_ppi2) - free_percpu_irq(arch_timer_ppi2, arch_timer_evt); + if (arch_timer_use_virtual) + free_percpu_irq(arch_timer_ppi[VIRT_PPI], arch_timer_evt); + else { + free_percpu_irq(arch_timer_ppi[PHYS_SECURE_PPI], + arch_timer_evt); + if (arch_timer_ppi[PHYS_NONSECURE_PPI]) + free_percpu_irq(arch_timer_ppi[PHYS_NONSECURE_PPI], + arch_timer_evt); + } out_free: free_percpu(arch_timer_evt); - +out: return err; } @@ -329,6 +473,7 @@ int __init arch_timer_of_register(void) { struct device_node *np; u32 freq; + int i; np = of_find_matching_node(NULL, arch_timer_of_match); if (!np) { @@ -340,22 +485,40 @@ int __init arch_timer_of_register(void) if (!of_property_read_u32(np, "clock-frequency", &freq)) arch_timer_rate = freq; - arch_timer_ppi = irq_of_parse_and_map(np, 0); - arch_timer_ppi2 = irq_of_parse_and_map(np, 1); - pr_info("arch_timer: found %s irqs %d %d\n", - np->name, arch_timer_ppi, arch_timer_ppi2); + for (i = PHYS_SECURE_PPI; i < MAX_TIMER_PPI; i++) + arch_timer_ppi[i] = irq_of_parse_and_map(np, i); + + /* + * If no interrupt provided for virtual timer, we'll have to + * stick to the physical timer. It'd better be accessible... + */ + if (!arch_timer_ppi[VIRT_PPI]) { + arch_timer_use_virtual = false; + + if (!arch_timer_ppi[PHYS_SECURE_PPI] || + !arch_timer_ppi[PHYS_NONSECURE_PPI]) { + pr_warn("arch_timer: No interrupt available, giving up\n"); + return -EINVAL; + } + } return arch_timer_register(); } int __init arch_timer_sched_clock_init(void) { + u32 (*cnt32)(void); int err; err = arch_timer_available(); if (err) return err; - setup_sched_clock(arch_counter_get_cntvct32, 32, arch_timer_rate); + if (arch_timer_use_virtual) + cnt32 = arch_counter_get_cntvct32; + else + cnt32 = arch_counter_get_cntpct32; + + setup_sched_clock(cnt32, 32, arch_timer_rate); return 0; } diff --git a/arch/arm/lib/delay.c b/arch/arm/lib/delay.c index 395d5fbb8fa..9d0a30032d7 100644 --- a/arch/arm/lib/delay.c +++ b/arch/arm/lib/delay.c @@ -34,7 +34,18 @@ struct arm_delay_ops arm_delay_ops = { .udelay = __loop_udelay, }; -#ifdef ARCH_HAS_READ_CURRENT_TIMER +static const struct delay_timer *delay_timer; +static bool delay_calibrated; + +int read_current_timer(unsigned long *timer_val) +{ + if (!delay_timer) + return -ENXIO; + + *timer_val = delay_timer->read_current_timer(); + return 0; +} + static void __timer_delay(unsigned long cycles) { cycles_t start = get_cycles(); @@ -55,18 +66,24 @@ static void __timer_udelay(unsigned long usecs) __timer_const_udelay(usecs * UDELAY_MULT); } -void __init init_current_timer_delay(unsigned long freq) +void __init register_current_timer_delay(const struct delay_timer *timer) { - pr_info("Switching to timer-based delay loop\n"); - lpj_fine = freq / HZ; - loops_per_jiffy = lpj_fine; - arm_delay_ops.delay = __timer_delay; - arm_delay_ops.const_udelay = __timer_const_udelay; - arm_delay_ops.udelay = __timer_udelay; + if (!delay_calibrated) { + pr_info("Switching to timer-based delay loop\n"); + delay_timer = timer; + lpj_fine = timer->freq / HZ; + loops_per_jiffy = lpj_fine; + arm_delay_ops.delay = __timer_delay; + arm_delay_ops.const_udelay = __timer_const_udelay; + arm_delay_ops.udelay = __timer_udelay; + delay_calibrated = true; + } else { + pr_info("Ignoring duplicate/late registration of read_current_timer delay\n"); + } } unsigned long __cpuinit calibrate_delay_is_known(void) { + delay_calibrated = true; return lpj_fine; } -#endif