7832738ae9
Add a k_timeout_t type, and use it everywhere that kernel API functions were accepting a millisecond timeout argument. Instead of forcing milliseconds everywhere (which are often not integrally representable as system ticks), do the conversion to ticks at the point where the timeout is created. This avoids an extra unit conversion in some application code, and allows us to express the timeout in units other than milliseconds to achieve greater precision. The existing K_MSEC() et. al. macros now return initializers for a k_timeout_t. The K_NO_WAIT and K_FOREVER constants have now become k_timeout_t values, which means they cannot be operated on as integers. Applications which have their own APIs that need to inspect these vs. user-provided timeouts can now use a K_TIMEOUT_EQ() predicate to test for equality. Timer drivers, which receive an integer tick count in ther z_clock_set_timeout() functions, now use the integer-valued K_TICKS_FOREVER constant instead of K_FOREVER. For the initial release, to preserve source compatibility, a CONFIG_LEGACY_TIMEOUT_API kconfig is provided. When true, the k_timeout_t will remain a compatible 32 bit value that will work with any legacy Zephyr application. Some subsystems present timeout (or timeout-like) values to their own users as APIs that would re-use the kernel's own constants and conventions. These will require some minor design work to adapt to the new scheme (in most cases just using k_timeout_t directly in their own API), and they have not been changed in this patch, instead selecting CONFIG_LEGACY_TIMEOUT_API via kconfig. These subsystems include: CAN Bus, the Microbit display driver, I2S, LoRa modem drivers, the UART Async API, Video hardware drivers, the console subsystem, and the network buffer abstraction. k_sleep() now takes a k_timeout_t argument, with a k_msleep() variant provided that works identically to the original API. Most of the changes here are just type/configuration management and documentation, but there are logic changes in mempool, where a loop that used a timeout numerically has been reworked using a new z_timeout_end_calc() predicate. Also in queue.c, a (when POLL was enabled) a similar loop was needlessly used to try to retry the k_poll() call after a spurious failure. But k_poll() does not fail spuriously, so the loop was removed. Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
241 lines
6.5 KiB
C
241 lines
6.5 KiB
C
/*
|
|
* Copyright (c) 2019 Intel Corporation
|
|
* SPDX-License-Identifier: Apache-2.0
|
|
*/
|
|
|
|
#include <drivers/timer/system_timer.h>
|
|
#include <sys_clock.h>
|
|
#include <spinlock.h>
|
|
#include <drivers/interrupt_controller/loapic.h>
|
|
|
|
BUILD_ASSERT(!IS_ENABLED(CONFIG_SMP), "APIC timer doesn't support SMP");
|
|
|
|
/*
|
|
* Overview:
|
|
*
|
|
* This driver enables the local APIC as the Zephyr system timer. It supports
|
|
* both legacy ("tickful") mode as well as TICKLESS_KERNEL. The driver will
|
|
* work with any APIC that has the ARAT "always running APIC timer" feature
|
|
* (CPUID 0x06, EAX bit 2); for the more accurate z_timer_cycle_get_32(),
|
|
* the invariant TSC feature (CPUID 0x80000007: EDX bit 8) is also required.
|
|
* (Ultimately systems with invariant TSCs should use a TSC-based driver,
|
|
* and the TSC-related parts should be stripped from this implementation.)
|
|
*
|
|
* Configuration:
|
|
*
|
|
* CONFIG_APIC_TIMER=y enables this timer driver.
|
|
* CONFIG_APIC_TIMER_IRQ=<irq> which IRQ to configure for the timer.
|
|
* CONFIG_APIC_TIMER_IRQ_PRIORITY=<p> priority for IRQ_CONNECT()
|
|
*
|
|
* CONFIG_SYS_CLOCK_HW_CYCLES_PER_SEC=<hz> must contain the frequency seen
|
|
* by the local APIC timer block (before it gets to the timer divider).
|
|
*
|
|
* CONFIG_APIC_TIMER_TSC=y enables the more accurate TSC-based cycle counter
|
|
* for z_timer_cycle_get_32(). This also requires the next options be set.
|
|
*
|
|
* CONFIG_APIC_TIMER_TSC_N=<n>
|
|
* CONFIG_APIC_TIMER_TSC_M=<m>
|
|
* When CONFIG_APIC_TIMER_TSC=y, these are set to indicate the ratio of
|
|
* the TSC frequency to CONFIG_SYS_CLOCK_HW_CYCLES_PER_SEC. This can be
|
|
* found via CPUID 0x15 (n = EBX, m = EAX) on most CPUs.
|
|
*/
|
|
|
|
/* These should be merged into include/drivers/interrupt_controller/loapic.h. */
|
|
|
|
#define DCR_DIVIDER_MASK 0x0000000F /* divider bits */
|
|
#define DCR_DIVIDER 0x0000000B /* divide by 1 */
|
|
#define LVT_MODE_MASK 0x00060000 /* timer mode bits */
|
|
#define LVT_MODE 0x00000000 /* one-shot */
|
|
|
|
/*
|
|
* CYCLES_PER_TICK must always be at least '2', otherwise MAX_TICKS
|
|
* will overflow s32_t, which is how 'ticks' are currently represented.
|
|
*/
|
|
|
|
#define CYCLES_PER_TICK \
|
|
(CONFIG_SYS_CLOCK_HW_CYCLES_PER_SEC / CONFIG_SYS_CLOCK_TICKS_PER_SEC)
|
|
|
|
BUILD_ASSERT(CYCLES_PER_TICK >= 2, "APIC timer: bad CYCLES_PER_TICK");
|
|
|
|
/* max number of ticks we can load into the timer in one shot */
|
|
|
|
#define MAX_TICKS (0xFFFFFFFFU / CYCLES_PER_TICK)
|
|
|
|
/*
|
|
* The spinlock protects all access to the local APIC timer registers,
|
|
* as well as 'total_cycles', 'last_announcement', and 'cached_icr'.
|
|
*
|
|
* One important invariant that must be observed: `total_cycles` + `cached_icr`
|
|
* is always an integral multiple of CYCLE_PER_TICK; this is, timer interrupts
|
|
* are only ever scheduled to occur at tick boundaries.
|
|
*/
|
|
|
|
static struct k_spinlock lock;
|
|
static u64_t total_cycles;
|
|
static u32_t cached_icr = CYCLES_PER_TICK;
|
|
|
|
#ifdef CONFIG_TICKLESS_KERNEL
|
|
|
|
static u64_t last_announcement; /* last time we called z_clock_announce() */
|
|
|
|
void z_clock_set_timeout(s32_t n, bool idle)
|
|
{
|
|
ARG_UNUSED(idle);
|
|
|
|
u32_t ccr;
|
|
int full_ticks; /* number of complete ticks we'll wait */
|
|
u32_t full_cycles; /* full_ticks represented as cycles */
|
|
u32_t partial_cycles; /* number of cycles to first tick boundary */
|
|
|
|
if (n < 1) {
|
|
full_ticks = 0;
|
|
} else if ((n == K_TICKS_FOREVER) || (n > MAX_TICKS)) {
|
|
full_ticks = MAX_TICKS - 1;
|
|
} else {
|
|
full_ticks = n - 1;
|
|
}
|
|
|
|
full_cycles = full_ticks * CYCLES_PER_TICK;
|
|
|
|
/*
|
|
* There's a wee race condition here. The timer may expire while
|
|
* we're busy reprogramming it; an interrupt will be queued at the
|
|
* local APIC and the ISR will be called too early, roughly right
|
|
* after we unlock, and not because the count we just programmed has
|
|
* counted down. Luckily this situation is easy to detect, which is
|
|
* why the ISR actually checks to be sure the CCR is 0 before acting.
|
|
*/
|
|
|
|
k_spinlock_key_t key = k_spin_lock(&lock);
|
|
|
|
ccr = x86_read_loapic(LOAPIC_TIMER_CCR);
|
|
total_cycles += (cached_icr - ccr);
|
|
partial_cycles = CYCLES_PER_TICK - (total_cycles % CYCLES_PER_TICK);
|
|
cached_icr = full_cycles + partial_cycles;
|
|
x86_write_loapic(LOAPIC_TIMER_ICR, cached_icr);
|
|
|
|
k_spin_unlock(&lock, key);
|
|
}
|
|
|
|
u32_t z_clock_elapsed(void)
|
|
{
|
|
u32_t ccr;
|
|
u32_t ticks;
|
|
|
|
k_spinlock_key_t key = k_spin_lock(&lock);
|
|
ccr = x86_read_loapic(LOAPIC_TIMER_CCR);
|
|
ticks = total_cycles - last_announcement;
|
|
ticks += cached_icr - ccr;
|
|
k_spin_unlock(&lock, key);
|
|
ticks /= CYCLES_PER_TICK;
|
|
|
|
return ticks;
|
|
}
|
|
|
|
static void isr(void *arg)
|
|
{
|
|
ARG_UNUSED(arg);
|
|
|
|
u32_t cycles;
|
|
s32_t ticks;
|
|
|
|
k_spinlock_key_t key = k_spin_lock(&lock);
|
|
|
|
/*
|
|
* If we get here and the CCR isn't zero, then this interrupt is
|
|
* stale: it was queued while z_clock_set_timeout() was setting
|
|
* a new counter. Just ignore it. See above for more info.
|
|
*/
|
|
|
|
if (x86_read_loapic(LOAPIC_TIMER_CCR) != 0) {
|
|
k_spin_unlock(&lock, key);
|
|
return;
|
|
}
|
|
|
|
/* Restart the timer as early as possible to minimize drift... */
|
|
x86_write_loapic(LOAPIC_TIMER_ICR, MAX_TICKS * CYCLES_PER_TICK);
|
|
|
|
cycles = cached_icr;
|
|
cached_icr = MAX_TICKS * CYCLES_PER_TICK;
|
|
total_cycles += cycles;
|
|
ticks = (total_cycles - last_announcement) / CYCLES_PER_TICK;
|
|
last_announcement = total_cycles;
|
|
k_spin_unlock(&lock, key);
|
|
z_clock_announce(ticks);
|
|
}
|
|
|
|
#else
|
|
|
|
static void isr(void *arg)
|
|
{
|
|
ARG_UNUSED(arg);
|
|
|
|
k_spinlock_key_t key = k_spin_lock(&lock);
|
|
total_cycles += CYCLES_PER_TICK;
|
|
x86_write_loapic(LOAPIC_TIMER_ICR, cached_icr);
|
|
k_spin_unlock(&lock, key);
|
|
|
|
z_clock_announce(1);
|
|
}
|
|
|
|
u32_t z_clock_elapsed(void)
|
|
{
|
|
return 0U;
|
|
}
|
|
|
|
#endif /* CONFIG_TICKLESS_KERNEL */
|
|
|
|
#ifdef CONFIG_APIC_TIMER_TSC
|
|
|
|
u32_t z_timer_cycle_get_32(void)
|
|
{
|
|
u64_t tsc = z_tsc_read();
|
|
u32_t cycles;
|
|
|
|
cycles = (tsc * CONFIG_APIC_TIMER_TSC_M) / CONFIG_APIC_TIMER_TSC_N;
|
|
return cycles;
|
|
}
|
|
|
|
#else
|
|
|
|
u32_t z_timer_cycle_get_32(void)
|
|
{
|
|
u32_t ret;
|
|
u32_t ccr;
|
|
|
|
k_spinlock_key_t key = k_spin_lock(&lock);
|
|
ccr = x86_read_loapic(LOAPIC_TIMER_CCR);
|
|
ret = total_cycles + (cached_icr - ccr);
|
|
k_spin_unlock(&lock, key);
|
|
|
|
return ret;
|
|
}
|
|
|
|
#endif
|
|
|
|
int z_clock_driver_init(struct device *device)
|
|
{
|
|
u32_t val;
|
|
|
|
val = x86_read_loapic(LOAPIC_TIMER_CONFIG); /* set divider */
|
|
val &= ~DCR_DIVIDER_MASK;
|
|
val |= DCR_DIVIDER;
|
|
x86_write_loapic(LOAPIC_TIMER_CONFIG, val);
|
|
|
|
val = x86_read_loapic(LOAPIC_TIMER); /* set timer mode */
|
|
val &= ~LVT_MODE_MASK;
|
|
val |= LVT_MODE;
|
|
x86_write_loapic(LOAPIC_TIMER, val);
|
|
|
|
/* remember, wiring up the interrupt will mess with the LVT, too */
|
|
|
|
IRQ_CONNECT(CONFIG_APIC_TIMER_IRQ,
|
|
CONFIG_APIC_TIMER_IRQ_PRIORITY,
|
|
isr, 0, 0);
|
|
|
|
x86_write_loapic(LOAPIC_TIMER_ICR, cached_icr);
|
|
irq_enable(CONFIG_APIC_TIMER_IRQ);
|
|
|
|
return 0;
|
|
}
|