f5899b5d4f
The udelay() inline for ia64 uses the ITC. If CONFIG_PREEMPT is enabled and the platform has unsynchronized ITCs and the calling task migrates to another CPU while doing the udelay loop, then the effective delay may be too short or very, very long. This patch disables preemption around 100 usec chunks of the overall desired udelay time. This minimizes preemption-holdoffs. udelay() is now too big to be inline, move it out of line and export it. Signed-off-by: John Hawkes <hawkes@sgi.com> Signed-off-by: Tony Luck <tony.luck@intel.com>
280 lines
7.8 KiB
C
280 lines
7.8 KiB
C
/*
|
|
* linux/arch/ia64/kernel/time.c
|
|
*
|
|
* Copyright (C) 1998-2003 Hewlett-Packard Co
|
|
* Stephane Eranian <eranian@hpl.hp.com>
|
|
* David Mosberger <davidm@hpl.hp.com>
|
|
* Copyright (C) 1999 Don Dugger <don.dugger@intel.com>
|
|
* Copyright (C) 1999-2000 VA Linux Systems
|
|
* Copyright (C) 1999-2000 Walt Drummond <drummond@valinux.com>
|
|
*/
|
|
#include <linux/config.h>
|
|
|
|
#include <linux/cpu.h>
|
|
#include <linux/init.h>
|
|
#include <linux/kernel.h>
|
|
#include <linux/module.h>
|
|
#include <linux/profile.h>
|
|
#include <linux/sched.h>
|
|
#include <linux/time.h>
|
|
#include <linux/interrupt.h>
|
|
#include <linux/efi.h>
|
|
#include <linux/profile.h>
|
|
#include <linux/timex.h>
|
|
|
|
#include <asm/machvec.h>
|
|
#include <asm/delay.h>
|
|
#include <asm/hw_irq.h>
|
|
#include <asm/ptrace.h>
|
|
#include <asm/sal.h>
|
|
#include <asm/sections.h>
|
|
#include <asm/system.h>
|
|
|
|
extern unsigned long wall_jiffies;
|
|
|
|
#define TIME_KEEPER_ID 0 /* smp_processor_id() of time-keeper */
|
|
|
|
#ifdef CONFIG_IA64_DEBUG_IRQ
|
|
|
|
unsigned long last_cli_ip;
|
|
EXPORT_SYMBOL(last_cli_ip);
|
|
|
|
#endif
|
|
|
|
static struct time_interpolator itc_interpolator = {
|
|
.shift = 16,
|
|
.mask = 0xffffffffffffffffLL,
|
|
.source = TIME_SOURCE_CPU
|
|
};
|
|
|
|
static irqreturn_t
|
|
timer_interrupt (int irq, void *dev_id, struct pt_regs *regs)
|
|
{
|
|
unsigned long new_itm;
|
|
|
|
if (unlikely(cpu_is_offline(smp_processor_id()))) {
|
|
return IRQ_HANDLED;
|
|
}
|
|
|
|
platform_timer_interrupt(irq, dev_id, regs);
|
|
|
|
new_itm = local_cpu_data->itm_next;
|
|
|
|
if (!time_after(ia64_get_itc(), new_itm))
|
|
printk(KERN_ERR "Oops: timer tick before it's due (itc=%lx,itm=%lx)\n",
|
|
ia64_get_itc(), new_itm);
|
|
|
|
profile_tick(CPU_PROFILING, regs);
|
|
|
|
while (1) {
|
|
update_process_times(user_mode(regs));
|
|
|
|
new_itm += local_cpu_data->itm_delta;
|
|
|
|
if (smp_processor_id() == TIME_KEEPER_ID) {
|
|
/*
|
|
* Here we are in the timer irq handler. We have irqs locally
|
|
* disabled, but we don't know if the timer_bh is running on
|
|
* another CPU. We need to avoid to SMP race by acquiring the
|
|
* xtime_lock.
|
|
*/
|
|
write_seqlock(&xtime_lock);
|
|
do_timer(regs);
|
|
local_cpu_data->itm_next = new_itm;
|
|
write_sequnlock(&xtime_lock);
|
|
} else
|
|
local_cpu_data->itm_next = new_itm;
|
|
|
|
if (time_after(new_itm, ia64_get_itc()))
|
|
break;
|
|
}
|
|
|
|
do {
|
|
/*
|
|
* If we're too close to the next clock tick for
|
|
* comfort, we increase the safety margin by
|
|
* intentionally dropping the next tick(s). We do NOT
|
|
* update itm.next because that would force us to call
|
|
* do_timer() which in turn would let our clock run
|
|
* too fast (with the potentially devastating effect
|
|
* of losing monotony of time).
|
|
*/
|
|
while (!time_after(new_itm, ia64_get_itc() + local_cpu_data->itm_delta/2))
|
|
new_itm += local_cpu_data->itm_delta;
|
|
ia64_set_itm(new_itm);
|
|
/* double check, in case we got hit by a (slow) PMI: */
|
|
} while (time_after_eq(ia64_get_itc(), new_itm));
|
|
return IRQ_HANDLED;
|
|
}
|
|
|
|
/*
|
|
* Encapsulate access to the itm structure for SMP.
|
|
*/
|
|
void
|
|
ia64_cpu_local_tick (void)
|
|
{
|
|
int cpu = smp_processor_id();
|
|
unsigned long shift = 0, delta;
|
|
|
|
/* arrange for the cycle counter to generate a timer interrupt: */
|
|
ia64_set_itv(IA64_TIMER_VECTOR);
|
|
|
|
delta = local_cpu_data->itm_delta;
|
|
/*
|
|
* Stagger the timer tick for each CPU so they don't occur all at (almost) the
|
|
* same time:
|
|
*/
|
|
if (cpu) {
|
|
unsigned long hi = 1UL << ia64_fls(cpu);
|
|
shift = (2*(cpu - hi) + 1) * delta/hi/2;
|
|
}
|
|
local_cpu_data->itm_next = ia64_get_itc() + delta + shift;
|
|
ia64_set_itm(local_cpu_data->itm_next);
|
|
}
|
|
|
|
static int nojitter;
|
|
|
|
static int __init nojitter_setup(char *str)
|
|
{
|
|
nojitter = 1;
|
|
printk("Jitter checking for ITC timers disabled\n");
|
|
return 1;
|
|
}
|
|
|
|
__setup("nojitter", nojitter_setup);
|
|
|
|
|
|
void __devinit
|
|
ia64_init_itm (void)
|
|
{
|
|
unsigned long platform_base_freq, itc_freq;
|
|
struct pal_freq_ratio itc_ratio, proc_ratio;
|
|
long status, platform_base_drift, itc_drift;
|
|
|
|
/*
|
|
* According to SAL v2.6, we need to use a SAL call to determine the platform base
|
|
* frequency and then a PAL call to determine the frequency ratio between the ITC
|
|
* and the base frequency.
|
|
*/
|
|
status = ia64_sal_freq_base(SAL_FREQ_BASE_PLATFORM,
|
|
&platform_base_freq, &platform_base_drift);
|
|
if (status != 0) {
|
|
printk(KERN_ERR "SAL_FREQ_BASE_PLATFORM failed: %s\n", ia64_sal_strerror(status));
|
|
} else {
|
|
status = ia64_pal_freq_ratios(&proc_ratio, NULL, &itc_ratio);
|
|
if (status != 0)
|
|
printk(KERN_ERR "PAL_FREQ_RATIOS failed with status=%ld\n", status);
|
|
}
|
|
if (status != 0) {
|
|
/* invent "random" values */
|
|
printk(KERN_ERR
|
|
"SAL/PAL failed to obtain frequency info---inventing reasonable values\n");
|
|
platform_base_freq = 100000000;
|
|
platform_base_drift = -1; /* no drift info */
|
|
itc_ratio.num = 3;
|
|
itc_ratio.den = 1;
|
|
}
|
|
if (platform_base_freq < 40000000) {
|
|
printk(KERN_ERR "Platform base frequency %lu bogus---resetting to 75MHz!\n",
|
|
platform_base_freq);
|
|
platform_base_freq = 75000000;
|
|
platform_base_drift = -1;
|
|
}
|
|
if (!proc_ratio.den)
|
|
proc_ratio.den = 1; /* avoid division by zero */
|
|
if (!itc_ratio.den)
|
|
itc_ratio.den = 1; /* avoid division by zero */
|
|
|
|
itc_freq = (platform_base_freq*itc_ratio.num)/itc_ratio.den;
|
|
|
|
local_cpu_data->itm_delta = (itc_freq + HZ/2) / HZ;
|
|
printk(KERN_DEBUG "CPU %d: base freq=%lu.%03luMHz, ITC ratio=%lu/%lu, "
|
|
"ITC freq=%lu.%03luMHz", smp_processor_id(),
|
|
platform_base_freq / 1000000, (platform_base_freq / 1000) % 1000,
|
|
itc_ratio.num, itc_ratio.den, itc_freq / 1000000, (itc_freq / 1000) % 1000);
|
|
|
|
if (platform_base_drift != -1) {
|
|
itc_drift = platform_base_drift*itc_ratio.num/itc_ratio.den;
|
|
printk("+/-%ldppm\n", itc_drift);
|
|
} else {
|
|
itc_drift = -1;
|
|
printk("\n");
|
|
}
|
|
|
|
local_cpu_data->proc_freq = (platform_base_freq*proc_ratio.num)/proc_ratio.den;
|
|
local_cpu_data->itc_freq = itc_freq;
|
|
local_cpu_data->cyc_per_usec = (itc_freq + USEC_PER_SEC/2) / USEC_PER_SEC;
|
|
local_cpu_data->nsec_per_cyc = ((NSEC_PER_SEC<<IA64_NSEC_PER_CYC_SHIFT)
|
|
+ itc_freq/2)/itc_freq;
|
|
|
|
if (!(sal_platform_features & IA64_SAL_PLATFORM_FEATURE_ITC_DRIFT)) {
|
|
itc_interpolator.frequency = local_cpu_data->itc_freq;
|
|
itc_interpolator.drift = itc_drift;
|
|
#ifdef CONFIG_SMP
|
|
/* On IA64 in an SMP configuration ITCs are never accurately synchronized.
|
|
* Jitter compensation requires a cmpxchg which may limit
|
|
* the scalability of the syscalls for retrieving time.
|
|
* The ITC synchronization is usually successful to within a few
|
|
* ITC ticks but this is not a sure thing. If you need to improve
|
|
* timer performance in SMP situations then boot the kernel with the
|
|
* "nojitter" option. However, doing so may result in time fluctuating (maybe
|
|
* even going backward) if the ITC offsets between the individual CPUs
|
|
* are too large.
|
|
*/
|
|
if (!nojitter) itc_interpolator.jitter = 1;
|
|
#endif
|
|
register_time_interpolator(&itc_interpolator);
|
|
}
|
|
|
|
/* Setup the CPU local timer tick */
|
|
ia64_cpu_local_tick();
|
|
}
|
|
|
|
static struct irqaction timer_irqaction = {
|
|
.handler = timer_interrupt,
|
|
.flags = SA_INTERRUPT,
|
|
.name = "timer"
|
|
};
|
|
|
|
void __init
|
|
time_init (void)
|
|
{
|
|
register_percpu_irq(IA64_TIMER_VECTOR, &timer_irqaction);
|
|
efi_gettimeofday(&xtime);
|
|
ia64_init_itm();
|
|
|
|
/*
|
|
* Initialize wall_to_monotonic such that adding it to xtime will yield zero, the
|
|
* tv_nsec field must be normalized (i.e., 0 <= nsec < NSEC_PER_SEC).
|
|
*/
|
|
set_normalized_timespec(&wall_to_monotonic, -xtime.tv_sec, -xtime.tv_nsec);
|
|
}
|
|
|
|
#define SMALLUSECS 100
|
|
|
|
void
|
|
udelay (unsigned long usecs)
|
|
{
|
|
unsigned long start;
|
|
unsigned long cycles;
|
|
unsigned long smallusecs;
|
|
|
|
/*
|
|
* Execute the non-preemptible delay loop (because the ITC might
|
|
* not be synchronized between CPUS) in relatively short time
|
|
* chunks, allowing preemption between the chunks.
|
|
*/
|
|
while (usecs > 0) {
|
|
smallusecs = (usecs > SMALLUSECS) ? SMALLUSECS : usecs;
|
|
preempt_disable();
|
|
cycles = smallusecs*local_cpu_data->cyc_per_usec;
|
|
start = ia64_get_itc();
|
|
|
|
while (ia64_get_itc() - start < cycles)
|
|
cpu_relax();
|
|
|
|
preempt_enable();
|
|
usecs -= smallusecs;
|
|
}
|
|
}
|
|
EXPORT_SYMBOL(udelay);
|