/* * CDDL HEADER START * * The contents of this file are subject to the terms of the * Common Development and Distribution License (the "License"). * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or http://www.opensolaris.org/os/licensing. * See the License for the specific language governing permissions * and limitations under the License. * * When distributing Covered Code, include this CDDL HEADER in each * file and include the License file at usr/src/OPENSOLARIS.LICENSE. * If applicable, add the following below this CDDL HEADER, with the * fields enclosed by brackets "[]" replaced with your own identifying * information: Portions Copyright [yyyy] [name of copyright owner] * * CDDL HEADER END */ /* Copyright (c) 1984, 1986, 1987, 1988, 1989 AT&T */ /* All Rights Reserved */ /* * Copyright 2009 Sun Microsystems, Inc. All rights reserved. * Use is subject to license terms. */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include /* * for NTP support */ #include #include #include #include /* * clock() is called straight from the clock cyclic; see clock_init(). * * Functions: * reprime clock * maintain date * jab the scheduler */ extern kcondvar_t fsflush_cv; extern sysinfo_t sysinfo; extern vminfo_t vminfo; extern int idleswtch; /* flag set while idle in pswtch() */ extern hrtime_t volatile devinfo_freeze; /* * high-precision avenrun values. These are needed to make the * regular avenrun values accurate. */ static uint64_t hp_avenrun[3]; int avenrun[3]; /* FSCALED average run queue lengths */ time_t time; /* time in seconds since 1970 - for compatibility only */ static struct loadavg_s loadavg; /* * Phase/frequency-lock loop (PLL/FLL) definitions * * The following variables are read and set by the ntp_adjtime() system * call. * * time_state shows the state of the system clock, with values defined * in the timex.h header file. * * time_status shows the status of the system clock, with bits defined * in the timex.h header file. * * time_offset is used by the PLL/FLL to adjust the system time in small * increments. * * time_constant determines the bandwidth or "stiffness" of the PLL. * * time_tolerance determines maximum frequency error or tolerance of the * CPU clock oscillator and is a property of the architecture; however, * in principle it could change as result of the presence of external * discipline signals, for instance. * * time_precision is usually equal to the kernel tick variable; however, * in cases where a precision clock counter or external clock is * available, the resolution can be much less than this and depend on * whether the external clock is working or not. * * time_maxerror is initialized by a ntp_adjtime() call and increased by * the kernel once each second to reflect the maximum error bound * growth. * * time_esterror is set and read by the ntp_adjtime() call, but * otherwise not used by the kernel. */ int32_t time_state = TIME_OK; /* clock state */ int32_t time_status = STA_UNSYNC; /* clock status bits */ int32_t time_offset = 0; /* time offset (us) */ int32_t time_constant = 0; /* pll time constant */ int32_t time_tolerance = MAXFREQ; /* frequency tolerance (scaled ppm) */ int32_t time_precision = 1; /* clock precision (us) */ int32_t time_maxerror = MAXPHASE; /* maximum error (us) */ int32_t time_esterror = MAXPHASE; /* estimated error (us) */ /* * The following variables establish the state of the PLL/FLL and the * residual time and frequency offset of the local clock. The scale * factors are defined in the timex.h header file. * * time_phase and time_freq are the phase increment and the frequency * increment, respectively, of the kernel time variable. * * time_freq is set via ntp_adjtime() from a value stored in a file when * the synchronization daemon is first started. Its value is retrieved * via ntp_adjtime() and written to the file about once per hour by the * daemon. * * time_adj is the adjustment added to the value of tick at each timer * interrupt and is recomputed from time_phase and time_freq at each * seconds rollover. * * time_reftime is the second's portion of the system time at the last * call to ntp_adjtime(). It is used to adjust the time_freq variable * and to increase the time_maxerror as the time since last update * increases. */ int32_t time_phase = 0; /* phase offset (scaled us) */ int32_t time_freq = 0; /* frequency offset (scaled ppm) */ int32_t time_adj = 0; /* tick adjust (scaled 1 / hz) */ int32_t time_reftime = 0; /* time at last adjustment (s) */ /* * The scale factors of the following variables are defined in the * timex.h header file. * * pps_time contains the time at each calibration interval, as read by * microtime(). pps_count counts the seconds of the calibration * interval, the duration of which is nominally pps_shift in powers of * two. * * pps_offset is the time offset produced by the time median filter * pps_tf[], while pps_jitter is the dispersion (jitter) measured by * this filter. * * pps_freq is the frequency offset produced by the frequency median * filter pps_ff[], while pps_stabil is the dispersion (wander) measured * by this filter. * * pps_usec is latched from a high resolution counter or external clock * at pps_time. Here we want the hardware counter contents only, not the * contents plus the time_tv.usec as usual. * * pps_valid counts the number of seconds since the last PPS update. It * is used as a watchdog timer to disable the PPS discipline should the * PPS signal be lost. * * pps_glitch counts the number of seconds since the beginning of an * offset burst more than tick/2 from current nominal offset. It is used * mainly to suppress error bursts due to priority conflicts between the * PPS interrupt and timer interrupt. * * pps_intcnt counts the calibration intervals for use in the interval- * adaptation algorithm. It's just too complicated for words. */ struct timeval pps_time; /* kernel time at last interval */ int32_t pps_tf[] = {0, 0, 0}; /* pps time offset median filter (us) */ int32_t pps_offset = 0; /* pps time offset (us) */ int32_t pps_jitter = MAXTIME; /* time dispersion (jitter) (us) */ int32_t pps_ff[] = {0, 0, 0}; /* pps frequency offset median filter */ int32_t pps_freq = 0; /* frequency offset (scaled ppm) */ int32_t pps_stabil = MAXFREQ; /* frequency dispersion (scaled ppm) */ int32_t pps_usec = 0; /* microsec counter at last interval */ int32_t pps_valid = PPS_VALID; /* pps signal watchdog counter */ int32_t pps_glitch = 0; /* pps signal glitch counter */ int32_t pps_count = 0; /* calibration interval counter (s) */ int32_t pps_shift = PPS_SHIFT; /* interval duration (s) (shift) */ int32_t pps_intcnt = 0; /* intervals at current duration */ /* * PPS signal quality monitors * * pps_jitcnt counts the seconds that have been discarded because the * jitter measured by the time median filter exceeds the limit MAXTIME * (100 us). * * pps_calcnt counts the frequency calibration intervals, which are * variable from 4 s to 256 s. * * pps_errcnt counts the calibration intervals which have been discarded * because the wander exceeds the limit MAXFREQ (100 ppm) or where the * calibration interval jitter exceeds two ticks. * * pps_stbcnt counts the calibration intervals that have been discarded * because the frequency wander exceeds the limit MAXFREQ / 4 (25 us). */ int32_t pps_jitcnt = 0; /* jitter limit exceeded */ int32_t pps_calcnt = 0; /* calibration intervals */ int32_t pps_errcnt = 0; /* calibration errors */ int32_t pps_stbcnt = 0; /* stability limit exceeded */ kcondvar_t lbolt_cv; /* * Hybrid lbolt implementation: * * The service historically provided by the lbolt and lbolt64 variables has * been replaced by the ddi_get_lbolt() and ddi_get_lbolt64() routines, and the * original symbols removed from the system. The once clock driven variables are * now implemented in an event driven fashion, backed by gethrtime() coarsed to * the appropriate clock resolution. The default event driven implementation is * complemented by a cyclic driven one, active only during periods of intense * activity around the DDI lbolt routines, when a lbolt specific cyclic is * reprogramed to fire at a clock tick interval to serve consumers of lbolt who * rely on the original low cost of consulting a memory position. * * The implementation uses the number of calls to these routines and the * frequency of these to determine when to transition from event to cyclic * driven and vice-versa. These values are kept on a per CPU basis for * scalability reasons and to prevent CPUs from constantly invalidating a single * cache line when modifying a global variable. The transition from event to * cyclic mode happens once the thresholds are crossed, and activity on any CPU * can cause such transition. * * The lbolt_hybrid function pointer is called by ddi_get_lbolt() and * ddi_get_lbolt64(), and will point to lbolt_event_driven() or * lbolt_cyclic_driven() according to the current mode. When the thresholds * are exceeded, lbolt_event_driven() will reprogram the lbolt cyclic to * fire at a nsec_per_tick interval and increment an internal variable at * each firing. lbolt_hybrid will then point to lbolt_cyclic_driven(), which * will simply return the value of such variable. lbolt_cyclic() will attempt * to shut itself off at each threshold interval (sampling period for calls * to the DDI lbolt routines), and return to the event driven mode, but will * be prevented from doing so if lbolt_cyclic_driven() is being heavily used. * * lbolt_bootstrap is used during boot to serve lbolt consumers who don't wait * for the cyclic subsystem to be intialized. * */ static int64_t lbolt_bootstrap(void); int64_t lbolt_event_driven(void); int64_t lbolt_cyclic_driven(void); int64_t (*lbolt_hybrid)(void) = lbolt_bootstrap; uint_t lbolt_ev_to_cyclic(caddr_t, caddr_t); /* * lbolt's cyclic, installed by clock_init(). */ static void lbolt_cyclic(void); /* * Tunable to keep lbolt in cyclic driven mode. This will prevent the system * from switching back to event driven, once it reaches cyclic mode. */ static boolean_t lbolt_cyc_only = B_FALSE; /* * Cache aligned, per CPU structure with lbolt usage statistics. */ static lbolt_cpu_t *lb_cpu; /* * Single, cache aligned, structure with all the information required by * the lbolt implementation. */ lbolt_info_t *lb_info; int one_sec = 1; /* turned on once every second */ static int fsflushcnt; /* counter for t_fsflushr */ int dosynctodr = 1; /* patchable; enable/disable sync to TOD chip */ int tod_needsync = 0; /* need to sync tod chip with software time */ static int tod_broken = 0; /* clock chip doesn't work */ time_t boot_time = 0; /* Boot time in seconds since 1970 */ cyclic_id_t clock_cyclic; /* clock()'s cyclic_id */ cyclic_id_t deadman_cyclic; /* deadman()'s cyclic_id */ cyclic_id_t ddi_timer_cyclic; /* cyclic_timer()'s cyclic_id */ extern void clock_tick_schedule(int); static int lgrp_ticks; /* counter to schedule lgrp load calcs */ /* * for tod fault detection */ #define TOD_REF_FREQ ((longlong_t)(NANOSEC)) #define TOD_STALL_THRESHOLD (TOD_REF_FREQ * 3 / 2) #define TOD_JUMP_THRESHOLD (TOD_REF_FREQ / 2) #define TOD_FILTER_N 4 #define TOD_FILTER_SETTLE (4 * TOD_FILTER_N) static int tod_faulted = TOD_NOFAULT; static int tod_fault_reset_flag = 0; /* patchable via /etc/system */ int tod_validate_enable = 1; /* Diagnose/Limit messages about delay(9F) called from interrupt context */ int delay_from_interrupt_diagnose = 0; volatile uint32_t delay_from_interrupt_msg = 20; /* * On non-SPARC systems, TOD validation must be deferred until gethrtime * returns non-zero values (after mach_clkinit's execution). * On SPARC systems, it must be deferred until after hrtime_base * and hres_last_tick are set (in the first invocation of hres_tick). * Since in both cases the prerequisites occur before the invocation of * tod_get() in clock(), the deferment is lifted there. */ static boolean_t tod_validate_deferred = B_TRUE; /* * tod_fault_table[] must be aligned with * enum tod_fault_type in systm.h */ static char *tod_fault_table[] = { "Reversed", /* TOD_REVERSED */ "Stalled", /* TOD_STALLED */ "Jumped", /* TOD_JUMPED */ "Changed in Clock Rate", /* TOD_RATECHANGED */ "Is Read-Only" /* TOD_RDONLY */ /* * no strings needed for TOD_NOFAULT */ }; /* * test hook for tod broken detection in tod_validate */ int tod_unit_test = 0; time_t tod_test_injector; #define CLOCK_ADJ_HIST_SIZE 4 static int adj_hist_entry; int64_t clock_adj_hist[CLOCK_ADJ_HIST_SIZE]; static void calcloadavg(int, uint64_t *); static int genloadavg(struct loadavg_s *); static void loadavg_update(); void (*cmm_clock_callout)() = NULL; void (*cpucaps_clock_callout)() = NULL; extern clock_t clock_tick_proc_max; static int64_t deadman_counter = 0; static void clock(void) { kthread_t *t; uint_t nrunnable; uint_t w_io; cpu_t *cp; cpupart_t *cpupart; extern void set_anoninfo(); extern void set_freemem(); void (*funcp)(); int32_t ltemp; int64_t lltemp; int s; int do_lgrp_load; int i; clock_t now = LBOLT_NO_ACCOUNT; /* current tick */ if (panicstr) return; set_anoninfo(); /* * Make sure that 'freemem' do not drift too far from the truth */ set_freemem(); /* * Before the section which is repeated is executed, we do * the time delta processing which occurs every clock tick * * There is additional processing which happens every time * the nanosecond counter rolls over which is described * below - see the section which begins with : if (one_sec) * * This section marks the beginning of the precision-kernel * code fragment. * * First, compute the phase adjustment. If the low-order bits * (time_phase) of the update overflow, bump the higher order * bits (time_update). */ time_phase += time_adj; if (time_phase <= -FINEUSEC) { ltemp = -time_phase / SCALE_PHASE; time_phase += ltemp * SCALE_PHASE; s = hr_clock_lock(); timedelta -= ltemp * (NANOSEC/MICROSEC); hr_clock_unlock(s); } else if (time_phase >= FINEUSEC) { ltemp = time_phase / SCALE_PHASE; time_phase -= ltemp * SCALE_PHASE; s = hr_clock_lock(); timedelta += ltemp * (NANOSEC/MICROSEC); hr_clock_unlock(s); } /* * End of precision-kernel code fragment which is processed * every timer interrupt. * * Continue with the interrupt processing as scheduled. */ /* * Count the number of runnable threads and the number waiting * for some form of I/O to complete -- gets added to * sysinfo.waiting. To know the state of the system, must add * wait counts from all CPUs. Also add up the per-partition * statistics. */ w_io = 0; nrunnable = 0; /* * keep track of when to update lgrp/part loads */ do_lgrp_load = 0; if (lgrp_ticks++ >= hz / 10) { lgrp_ticks = 0; do_lgrp_load = 1; } if (one_sec) { loadavg_update(); deadman_counter++; } /* * First count the threads waiting on kpreempt queues in each * CPU partition. */ cpupart = cp_list_head; do { uint_t cpupart_nrunnable = cpupart->cp_kp_queue.disp_nrunnable; cpupart->cp_updates++; nrunnable += cpupart_nrunnable; cpupart->cp_nrunnable_cum += cpupart_nrunnable; if (one_sec) { cpupart->cp_nrunning = 0; cpupart->cp_nrunnable = cpupart_nrunnable; } } while ((cpupart = cpupart->cp_next) != cp_list_head); /* Now count the per-CPU statistics. */ cp = cpu_list; do { uint_t cpu_nrunnable = cp->cpu_disp->disp_nrunnable; nrunnable += cpu_nrunnable; cpupart = cp->cpu_part; cpupart->cp_nrunnable_cum += cpu_nrunnable; if (one_sec) { cpupart->cp_nrunnable += cpu_nrunnable; /* * Update user, system, and idle cpu times. */ cpupart->cp_nrunning++; /* * w_io is used to update sysinfo.waiting during * one_second processing below. Only gather w_io * information when we walk the list of cpus if we're * going to perform one_second processing. */ w_io += CPU_STATS(cp, sys.iowait); } if (one_sec && (cp->cpu_flags & CPU_EXISTS)) { int i, load, change; hrtime_t intracct, intrused; const hrtime_t maxnsec = 1000000000; const int precision = 100; /* * Estimate interrupt load on this cpu each second. * Computes cpu_intrload as %utilization (0-99). */ /* add up interrupt time from all micro states */ for (intracct = 0, i = 0; i < NCMSTATES; i++) intracct += cp->cpu_intracct[i]; scalehrtime(&intracct); /* compute nsec used in the past second */ intrused = intracct - cp->cpu_intrlast; cp->cpu_intrlast = intracct; /* limit the value for safety (and the first pass) */ if (intrused >= maxnsec) intrused = maxnsec - 1; /* calculate %time in interrupt */ load = (precision * intrused) / maxnsec; ASSERT(load >= 0 && load < precision); change = cp->cpu_intrload - load; /* jump to new max, or decay the old max */ if (change < 0) cp->cpu_intrload = load; else if (change > 0) cp->cpu_intrload -= (change + 3) / 4; DTRACE_PROBE3(cpu_intrload, cpu_t *, cp, hrtime_t, intracct, hrtime_t, intrused); } if (do_lgrp_load && (cp->cpu_flags & CPU_EXISTS)) { /* * When updating the lgroup's load average, * account for the thread running on the CPU. * If the CPU is the current one, then we need * to account for the underlying thread which * got the clock interrupt not the thread that is * handling the interrupt and caculating the load * average */ t = cp->cpu_thread; if (CPU == cp) t = t->t_intr; /* * Account for the load average for this thread if * it isn't the idle thread or it is on the interrupt * stack and not the current CPU handling the clock * interrupt */ if ((t && t != cp->cpu_idle_thread) || (CPU != cp && CPU_ON_INTR(cp))) { if (t->t_lpl == cp->cpu_lpl) { /* local thread */ cpu_nrunnable++; } else { /* * This is a remote thread, charge it * against its home lgroup. Note that * we notice that a thread is remote * only if it's currently executing. * This is a reasonable approximation, * since queued remote threads are rare. * Note also that if we didn't charge * it to its home lgroup, remote * execution would often make a system * appear balanced even though it was * not, and thread placement/migration * would often not be done correctly. */ lgrp_loadavg(t->t_lpl, LGRP_LOADAVG_IN_THREAD_MAX, 0); } } lgrp_loadavg(cp->cpu_lpl, cpu_nrunnable * LGRP_LOADAVG_IN_THREAD_MAX, 1); } } while ((cp = cp->cpu_next) != cpu_list); clock_tick_schedule(one_sec); /* * Check for a callout that needs be called from the clock * thread to support the membership protocol in a clustered * system. Copy the function pointer so that we can reset * this to NULL if needed. */ if ((funcp = cmm_clock_callout) != NULL) (*funcp)(); if ((funcp = cpucaps_clock_callout) != NULL) (*funcp)(); /* * Wakeup the cageout thread waiters once per second. */ if (one_sec) kcage_tick(); if (one_sec) { int drift, absdrift; timestruc_t tod; int s; /* * Beginning of precision-kernel code fragment executed * every second. * * On rollover of the second the phase adjustment to be * used for the next second is calculated. Also, the * maximum error is increased by the tolerance. If the * PPS frequency discipline code is present, the phase is * increased to compensate for the CPU clock oscillator * frequency error. * * On a 32-bit machine and given parameters in the timex.h * header file, the maximum phase adjustment is +-512 ms * and maximum frequency offset is (a tad less than) * +-512 ppm. On a 64-bit machine, you shouldn't need to ask. */ time_maxerror += time_tolerance / SCALE_USEC; /* * Leap second processing. If in leap-insert state at * the end of the day, the system clock is set back one * second; if in leap-delete state, the system clock is * set ahead one second. The microtime() routine or * external clock driver will insure that reported time * is always monotonic. The ugly divides should be * replaced. */ switch (time_state) { case TIME_OK: if (time_status & STA_INS) time_state = TIME_INS; else if (time_status & STA_DEL) time_state = TIME_DEL; break; case TIME_INS: if (hrestime.tv_sec % 86400 == 0) { s = hr_clock_lock(); hrestime.tv_sec--; hr_clock_unlock(s); time_state = TIME_OOP; } break; case TIME_DEL: if ((hrestime.tv_sec + 1) % 86400 == 0) { s = hr_clock_lock(); hrestime.tv_sec++; hr_clock_unlock(s); time_state = TIME_WAIT; } break; case TIME_OOP: time_state = TIME_WAIT; break; case TIME_WAIT: if (!(time_status & (STA_INS | STA_DEL))) time_state = TIME_OK; default: break; } /* * Compute the phase adjustment for the next second. In * PLL mode, the offset is reduced by a fixed factor * times the time constant. In FLL mode the offset is * used directly. In either mode, the maximum phase * adjustment for each second is clamped so as to spread * the adjustment over not more than the number of * seconds between updates. */ if (time_offset == 0) time_adj = 0; else if (time_offset < 0) { lltemp = -time_offset; if (!(time_status & STA_FLL)) { if ((1 << time_constant) >= SCALE_KG) lltemp *= (1 << time_constant) / SCALE_KG; else lltemp = (lltemp / SCALE_KG) >> time_constant; } if (lltemp > (MAXPHASE / MINSEC) * SCALE_UPDATE) lltemp = (MAXPHASE / MINSEC) * SCALE_UPDATE; time_offset += lltemp; time_adj = -(lltemp * SCALE_PHASE) / hz / SCALE_UPDATE; } else { lltemp = time_offset; if (!(time_status & STA_FLL)) { if ((1 << time_constant) >= SCALE_KG) lltemp *= (1 << time_constant) / SCALE_KG; else lltemp = (lltemp / SCALE_KG) >> time_constant; } if (lltemp > (MAXPHASE / MINSEC) * SCALE_UPDATE) lltemp = (MAXPHASE / MINSEC) * SCALE_UPDATE; time_offset -= lltemp; time_adj = (lltemp * SCALE_PHASE) / hz / SCALE_UPDATE; } /* * Compute the frequency estimate and additional phase * adjustment due to frequency error for the next * second. When the PPS signal is engaged, gnaw on the * watchdog counter and update the frequency computed by * the pll and the PPS signal. */ pps_valid++; if (pps_valid == PPS_VALID) { pps_jitter = MAXTIME; pps_stabil = MAXFREQ; time_status &= ~(STA_PPSSIGNAL | STA_PPSJITTER | STA_PPSWANDER | STA_PPSERROR); } lltemp = time_freq + pps_freq; if (lltemp) time_adj += (lltemp * SCALE_PHASE) / (SCALE_USEC * hz); /* * End of precision kernel-code fragment * * The section below should be modified if we are planning * to use NTP for synchronization. * * Note: the clock synchronization code now assumes * the following: * - if dosynctodr is 1, then compute the drift between * the tod chip and software time and adjust one or * the other depending on the circumstances * * - if dosynctodr is 0, then the tod chip is independent * of the software clock and should not be adjusted, * but allowed to free run. this allows NTP to sync. * hrestime without any interference from the tod chip. */ tod_validate_deferred = B_FALSE; mutex_enter(&tod_lock); tod = tod_get(); drift = tod.tv_sec - hrestime.tv_sec; absdrift = (drift >= 0) ? drift : -drift; if (tod_needsync || absdrift > 1) { int s; if (absdrift > 2) { if (!tod_broken && tod_faulted == TOD_NOFAULT) { s = hr_clock_lock(); hrestime = tod; membar_enter(); /* hrestime visible */ timedelta = 0; timechanged++; tod_needsync = 0; hr_clock_unlock(s); callout_hrestime(); } } else { if (tod_needsync || !dosynctodr) { gethrestime(&tod); tod_set(tod); s = hr_clock_lock(); if (timedelta == 0) tod_needsync = 0; hr_clock_unlock(s); } else { /* * If the drift is 2 seconds on the * money, then the TOD is adjusting * the clock; record that. */ clock_adj_hist[adj_hist_entry++ % CLOCK_ADJ_HIST_SIZE] = now; s = hr_clock_lock(); timedelta = (int64_t)drift*NANOSEC; hr_clock_unlock(s); } } } one_sec = 0; time = gethrestime_sec(); /* for crusty old kmem readers */ mutex_exit(&tod_lock); /* * Some drivers still depend on this... XXX */ cv_broadcast(&lbolt_cv); sysinfo.updates++; vminfo.freemem += freemem; { pgcnt_t maxswap, resv, free; pgcnt_t avail = MAX((spgcnt_t)(availrmem - swapfs_minfree), 0); maxswap = k_anoninfo.ani_mem_resv + k_anoninfo.ani_max +avail; free = k_anoninfo.ani_free + avail; resv = k_anoninfo.ani_phys_resv + k_anoninfo.ani_mem_resv; vminfo.swap_resv += resv; /* number of reserved and allocated pages */ #ifdef DEBUG if (maxswap < free) cmn_err(CE_WARN, "clock: maxswap < free"); if (maxswap < resv) cmn_err(CE_WARN, "clock: maxswap < resv"); #endif vminfo.swap_alloc += maxswap - free; vminfo.swap_avail += maxswap - resv; vminfo.swap_free += free; } if (nrunnable) { sysinfo.runque += nrunnable; sysinfo.runocc++; } if (nswapped) { sysinfo.swpque += nswapped; sysinfo.swpocc++; } sysinfo.waiting += w_io; /* * Wake up fsflush to write out DELWRI * buffers, dirty pages and other cached * administrative data, e.g. inodes. */ if (--fsflushcnt <= 0) { fsflushcnt = tune.t_fsflushr; cv_signal(&fsflush_cv); } vmmeter(); calcloadavg(genloadavg(&loadavg), hp_avenrun); for (i = 0; i < 3; i++) /* * At the moment avenrun[] can only hold 31 * bits of load average as it is a signed * int in the API. We need to ensure that * hp_avenrun[i] >> (16 - FSHIFT) will not be * too large. If it is, we put the largest value * that we can use into avenrun[i]. This is * kludgey, but about all we can do until we * avenrun[] is declared as an array of uint64[] */ if (hp_avenrun[i] < ((uint64_t)1<<(31+16-FSHIFT))) avenrun[i] = (int32_t)(hp_avenrun[i] >> (16 - FSHIFT)); else avenrun[i] = 0x7fffffff; cpupart = cp_list_head; do { calcloadavg(genloadavg(&cpupart->cp_loadavg), cpupart->cp_hp_avenrun); } while ((cpupart = cpupart->cp_next) != cp_list_head); /* * Wake up the swapper thread if necessary. */ if (runin || (runout && (avefree < desfree || wake_sched_sec))) { t = &t0; thread_lock(t); if (t->t_state == TS_STOPPED) { runin = runout = 0; wake_sched_sec = 0; t->t_whystop = 0; t->t_whatstop = 0; t->t_schedflag &= ~TS_ALLSTART; THREAD_TRANSITION(t); setfrontdq(t); } thread_unlock(t); } } /* * Wake up the swapper if any high priority swapped-out threads * became runable during the last tick. */ if (wake_sched) { t = &t0; thread_lock(t); if (t->t_state == TS_STOPPED) { runin = runout = 0; wake_sched = 0; t->t_whystop = 0; t->t_whatstop = 0; t->t_schedflag &= ~TS_ALLSTART; THREAD_TRANSITION(t); setfrontdq(t); } thread_unlock(t); } } void clock_init(void) { cyc_handler_t clk_hdlr, timer_hdlr, lbolt_hdlr; cyc_time_t clk_when, lbolt_when; int i, sz; intptr_t buf; /* * Setup handler and timer for the clock cyclic. */ clk_hdlr.cyh_func = (cyc_func_t)clock; clk_hdlr.cyh_level = CY_LOCK_LEVEL; clk_hdlr.cyh_arg = NULL; clk_when.cyt_when = 0; clk_when.cyt_interval = nsec_per_tick; /* * cyclic_timer is dedicated to the ddi interface, which * uses the same clock resolution as the system one. */ timer_hdlr.cyh_func = (cyc_func_t)cyclic_timer; timer_hdlr.cyh_level = CY_LOCK_LEVEL; timer_hdlr.cyh_arg = NULL; /* * Setup the necessary structures for the lbolt cyclic and add the * soft interrupt which will switch from event to cyclic mode when * under high pil. */ lbolt_hdlr.cyh_func = (cyc_func_t)lbolt_cyclic; lbolt_hdlr.cyh_level = CY_LOCK_LEVEL; lbolt_hdlr.cyh_arg = NULL; lbolt_when.cyt_interval = nsec_per_tick; if (lbolt_cyc_only) { lbolt_when.cyt_when = 0; lbolt_hybrid = lbolt_cyclic_driven; } else { lbolt_when.cyt_when = CY_INFINITY; lbolt_hybrid = lbolt_event_driven; } /* * Allocate cache line aligned space for the per CPU lbolt data and * lb_info structure. We also initialize these structures with their * default values and install the softint to change from event to * cyclic driven mode. */ sz = sizeof (lbolt_info_t) + CPU_CACHE_COHERENCE_SIZE; buf = (intptr_t)kmem_zalloc(sz, KM_SLEEP); lb_info = (lbolt_info_t *)P2ROUNDUP(buf, CPU_CACHE_COHERENCE_SIZE); if (hz != HZ_DEFAULT) lb_info->lbi_thresh_interval = LBOLT_THRESH_INTERVAL * hz/HZ_DEFAULT; else lb_info->lbi_thresh_interval = LBOLT_THRESH_INTERVAL; lb_info->lbi_thresh_calls = LBOLT_THRESH_CALLS; sz = (sizeof (lbolt_info_t) * max_ncpus) + CPU_CACHE_COHERENCE_SIZE; buf = (intptr_t)kmem_zalloc(sz, KM_SLEEP); lb_cpu = (lbolt_cpu_t *)P2ROUNDUP(buf, CPU_CACHE_COHERENCE_SIZE); for (i = 0; i < max_ncpus; i++) lb_cpu[i].lbc_counter = lb_info->lbi_thresh_calls; lbolt_softint_add(); /* * Grab cpu_lock and install all three cyclics. */ mutex_enter(&cpu_lock); clock_cyclic = cyclic_add(&clk_hdlr, &clk_when); ddi_timer_cyclic = cyclic_add(&timer_hdlr, &clk_when); lb_info->lbi_cyclic_id = cyclic_add(&lbolt_hdlr, &lbolt_when); mutex_exit(&cpu_lock); } /* * Called before calcloadavg to get 10-sec moving loadavg together */ static int genloadavg(struct loadavg_s *avgs) { int avg; int spos; /* starting position */ int cpos; /* moving current position */ int i; int slen; hrtime_t hr_avg; /* 10-second snapshot, calculate first positon */ if (avgs->lg_len == 0) { return (0); } slen = avgs->lg_len < S_MOVAVG_SZ ? avgs->lg_len : S_MOVAVG_SZ; spos = (avgs->lg_cur - 1) >= 0 ? avgs->lg_cur - 1 : S_LOADAVG_SZ + (avgs->lg_cur - 1); for (i = hr_avg = 0; i < slen; i++) { cpos = (spos - i) >= 0 ? spos - i : S_LOADAVG_SZ + (spos - i); hr_avg += avgs->lg_loads[cpos]; } hr_avg = hr_avg / slen; avg = hr_avg / (NANOSEC / LGRP_LOADAVG_IN_THREAD_MAX); return (avg); } /* * Run every second from clock () to update the loadavg count available to the * system and cpu-partitions. * * This works by sampling the previous usr, sys, wait time elapsed, * computing a delta, and adding that delta to the elapsed usr, sys, * wait increase. */ static void loadavg_update() { cpu_t *cp; cpupart_t *cpupart; hrtime_t cpu_total; int prev; cp = cpu_list; loadavg.lg_total = 0; /* * first pass totals up per-cpu statistics for system and cpu * partitions */ do { struct loadavg_s *lavg; lavg = &cp->cpu_loadavg; cpu_total = cp->cpu_acct[CMS_USER] + cp->cpu_acct[CMS_SYSTEM] + cp->cpu_waitrq; /* compute delta against last total */ scalehrtime(&cpu_total); prev = (lavg->lg_cur - 1) >= 0 ? lavg->lg_cur - 1 : S_LOADAVG_SZ + (lavg->lg_cur - 1); if (lavg->lg_loads[prev] <= 0) { lavg->lg_loads[lavg->lg_cur] = cpu_total; cpu_total = 0; } else { lavg->lg_loads[lavg->lg_cur] = cpu_total; cpu_total = cpu_total - lavg->lg_loads[prev]; if (cpu_total < 0) cpu_total = 0; } lavg->lg_cur = (lavg->lg_cur + 1) % S_LOADAVG_SZ; lavg->lg_len = (lavg->lg_len + 1) < S_LOADAVG_SZ ? lavg->lg_len + 1 : S_LOADAVG_SZ; loadavg.lg_total += cpu_total; cp->cpu_part->cp_loadavg.lg_total += cpu_total; } while ((cp = cp->cpu_next) != cpu_list); loadavg.lg_loads[loadavg.lg_cur] = loadavg.lg_total; loadavg.lg_cur = (loadavg.lg_cur + 1) % S_LOADAVG_SZ; loadavg.lg_len = (loadavg.lg_len + 1) < S_LOADAVG_SZ ? loadavg.lg_len + 1 : S_LOADAVG_SZ; /* * Second pass updates counts */ cpupart = cp_list_head; do { struct loadavg_s *lavg; lavg = &cpupart->cp_loadavg; lavg->lg_loads[lavg->lg_cur] = lavg->lg_total; lavg->lg_total = 0; lavg->lg_cur = (lavg->lg_cur + 1) % S_LOADAVG_SZ; lavg->lg_len = (lavg->lg_len + 1) < S_LOADAVG_SZ ? lavg->lg_len + 1 : S_LOADAVG_SZ; } while ((cpupart = cpupart->cp_next) != cp_list_head); } /* * clock_update() - local clock update * * This routine is called by ntp_adjtime() to update the local clock * phase and frequency. The implementation is of an * adaptive-parameter, hybrid phase/frequency-lock loop (PLL/FLL). The * routine computes new time and frequency offset estimates for each * call. The PPS signal itself determines the new time offset, * instead of the calling argument. Presumably, calls to * ntp_adjtime() occur only when the caller believes the local clock * is valid within some bound (+-128 ms with NTP). If the caller's * time is far different than the PPS time, an argument will ensue, * and it's not clear who will lose. * * For uncompensated quartz crystal oscillatores and nominal update * intervals less than 1024 s, operation should be in phase-lock mode * (STA_FLL = 0), where the loop is disciplined to phase. For update * intervals greater than this, operation should be in frequency-lock * mode (STA_FLL = 1), where the loop is disciplined to frequency. * * Note: mutex(&tod_lock) is in effect. */ void clock_update(int offset) { int ltemp, mtemp, s; ASSERT(MUTEX_HELD(&tod_lock)); if (!(time_status & STA_PLL) && !(time_status & STA_PPSTIME)) return; ltemp = offset; if ((time_status & STA_PPSTIME) && (time_status & STA_PPSSIGNAL)) ltemp = pps_offset; /* * Scale the phase adjustment and clamp to the operating range. */ if (ltemp > MAXPHASE) time_offset = MAXPHASE * SCALE_UPDATE; else if (ltemp < -MAXPHASE) time_offset = -(MAXPHASE * SCALE_UPDATE); else time_offset = ltemp * SCALE_UPDATE; /* * Select whether the frequency is to be controlled and in which * mode (PLL or FLL). Clamp to the operating range. Ugly * multiply/divide should be replaced someday. */ if (time_status & STA_FREQHOLD || time_reftime == 0) time_reftime = hrestime.tv_sec; mtemp = hrestime.tv_sec - time_reftime; time_reftime = hrestime.tv_sec; if (time_status & STA_FLL) { if (mtemp >= MINSEC) { ltemp = ((time_offset / mtemp) * (SCALE_USEC / SCALE_UPDATE)); if (ltemp) time_freq += ltemp / SCALE_KH; } } else { if (mtemp < MAXSEC) { ltemp *= mtemp; if (ltemp) time_freq += (int)(((int64_t)ltemp * SCALE_USEC) / SCALE_KF) / (1 << (time_constant * 2)); } } if (time_freq > time_tolerance) time_freq = time_tolerance; else if (time_freq < -time_tolerance) time_freq = -time_tolerance; s = hr_clock_lock(); tod_needsync = 1; hr_clock_unlock(s); } /* * ddi_hardpps() - discipline CPU clock oscillator to external PPS signal * * This routine is called at each PPS interrupt in order to discipline * the CPU clock oscillator to the PPS signal. It measures the PPS phase * and leaves it in a handy spot for the clock() routine. It * integrates successive PPS phase differences and calculates the * frequency offset. This is used in clock() to discipline the CPU * clock oscillator so that intrinsic frequency error is cancelled out. * The code requires the caller to capture the time and hardware counter * value at the on-time PPS signal transition. * * Note that, on some Unix systems, this routine runs at an interrupt * priority level higher than the timer interrupt routine clock(). * Therefore, the variables used are distinct from the clock() * variables, except for certain exceptions: The PPS frequency pps_freq * and phase pps_offset variables are determined by this routine and * updated atomically. The time_tolerance variable can be considered a * constant, since it is infrequently changed, and then only when the * PPS signal is disabled. The watchdog counter pps_valid is updated * once per second by clock() and is atomically cleared in this * routine. * * tvp is the time of the last tick; usec is a microsecond count since the * last tick. * * Note: In Solaris systems, the tick value is actually given by * usec_per_tick. This is called from the serial driver cdintr(), * or equivalent, at a high PIL. Because the kernel keeps a * highresolution time, the following code can accept either * the traditional argument pair, or the current highres timestamp * in tvp and zero in usec. */ void ddi_hardpps(struct timeval *tvp, int usec) { int u_usec, v_usec, bigtick; time_t cal_sec; int cal_usec; /* * An occasional glitch can be produced when the PPS interrupt * occurs in the clock() routine before the time variable is * updated. Here the offset is discarded when the difference * between it and the last one is greater than tick/2, but not * if the interval since the first discard exceeds 30 s. */ time_status |= STA_PPSSIGNAL; time_status &= ~(STA_PPSJITTER | STA_PPSWANDER | STA_PPSERROR); pps_valid = 0; u_usec = -tvp->tv_usec; if (u_usec < -(MICROSEC/2)) u_usec += MICROSEC; v_usec = pps_offset - u_usec; if (v_usec < 0) v_usec = -v_usec; if (v_usec > (usec_per_tick >> 1)) { if (pps_glitch > MAXGLITCH) { pps_glitch = 0; pps_tf[2] = u_usec; pps_tf[1] = u_usec; } else { pps_glitch++; u_usec = pps_offset; } } else pps_glitch = 0; /* * A three-stage median filter is used to help deglitch the pps * time. The median sample becomes the time offset estimate; the * difference between the other two samples becomes the time * dispersion (jitter) estimate. */ pps_tf[2] = pps_tf[1]; pps_tf[1] = pps_tf[0]; pps_tf[0] = u_usec; if (pps_tf[0] > pps_tf[1]) { if (pps_tf[1] > pps_tf[2]) { pps_offset = pps_tf[1]; /* 0 1 2 */ v_usec = pps_tf[0] - pps_tf[2]; } else if (pps_tf[2] > pps_tf[0]) { pps_offset = pps_tf[0]; /* 2 0 1 */ v_usec = pps_tf[2] - pps_tf[1]; } else { pps_offset = pps_tf[2]; /* 0 2 1 */ v_usec = pps_tf[0] - pps_tf[1]; } } else { if (pps_tf[1] < pps_tf[2]) { pps_offset = pps_tf[1]; /* 2 1 0 */ v_usec = pps_tf[2] - pps_tf[0]; } else if (pps_tf[2] < pps_tf[0]) { pps_offset = pps_tf[0]; /* 1 0 2 */ v_usec = pps_tf[1] - pps_tf[2]; } else { pps_offset = pps_tf[2]; /* 1 2 0 */ v_usec = pps_tf[1] - pps_tf[0]; } } if (v_usec > MAXTIME) pps_jitcnt++; v_usec = (v_usec << PPS_AVG) - pps_jitter; pps_jitter += v_usec / (1 << PPS_AVG); if (pps_jitter > (MAXTIME >> 1)) time_status |= STA_PPSJITTER; /* * During the calibration interval adjust the starting time when * the tick overflows. At the end of the interval compute the * duration of the interval and the difference of the hardware * counters at the beginning and end of the interval. This code * is deliciously complicated by the fact valid differences may * exceed the value of tick when using long calibration * intervals and small ticks. Note that the counter can be * greater than tick if caught at just the wrong instant, but * the values returned and used here are correct. */ bigtick = (int)usec_per_tick * SCALE_USEC; pps_usec -= pps_freq; if (pps_usec >= bigtick) pps_usec -= bigtick; if (pps_usec < 0) pps_usec += bigtick; pps_time.tv_sec++; pps_count++; if (pps_count < (1 << pps_shift)) return; pps_count = 0; pps_calcnt++; u_usec = usec * SCALE_USEC; v_usec = pps_usec - u_usec; if (v_usec >= bigtick >> 1) v_usec -= bigtick; if (v_usec < -(bigtick >> 1)) v_usec += bigtick; if (v_usec < 0) v_usec = -(-v_usec >> pps_shift); else v_usec = v_usec >> pps_shift; pps_usec = u_usec; cal_sec = tvp->tv_sec; cal_usec = tvp->tv_usec; cal_sec -= pps_time.tv_sec; cal_usec -= pps_time.tv_usec; if (cal_usec < 0) { cal_usec += MICROSEC; cal_sec--; } pps_time = *tvp; /* * Check for lost interrupts, noise, excessive jitter and * excessive frequency error. The number of timer ticks during * the interval may vary +-1 tick. Add to this a margin of one * tick for the PPS signal jitter and maximum frequency * deviation. If the limits are exceeded, the calibration * interval is reset to the minimum and we start over. */ u_usec = (int)usec_per_tick << 1; if (!((cal_sec == -1 && cal_usec > (MICROSEC - u_usec)) || (cal_sec == 0 && cal_usec < u_usec)) || v_usec > time_tolerance || v_usec < -time_tolerance) { pps_errcnt++; pps_shift = PPS_SHIFT; pps_intcnt = 0; time_status |= STA_PPSERROR; return; } /* * A three-stage median filter is used to help deglitch the pps * frequency. The median sample becomes the frequency offset * estimate; the difference between the other two samples * becomes the frequency dispersion (stability) estimate. */ pps_ff[2] = pps_ff[1]; pps_ff[1] = pps_ff[0]; pps_ff[0] = v_usec; if (pps_ff[0] > pps_ff[1]) { if (pps_ff[1] > pps_ff[2]) { u_usec = pps_ff[1]; /* 0 1 2 */ v_usec = pps_ff[0] - pps_ff[2]; } else if (pps_ff[2] > pps_ff[0]) { u_usec = pps_ff[0]; /* 2 0 1 */ v_usec = pps_ff[2] - pps_ff[1]; } else { u_usec = pps_ff[2]; /* 0 2 1 */ v_usec = pps_ff[0] - pps_ff[1]; } } else { if (pps_ff[1] < pps_ff[2]) { u_usec = pps_ff[1]; /* 2 1 0 */ v_usec = pps_ff[2] - pps_ff[0]; } else if (pps_ff[2] < pps_ff[0]) { u_usec = pps_ff[0]; /* 1 0 2 */ v_usec = pps_ff[1] - pps_ff[2]; } else { u_usec = pps_ff[2]; /* 1 2 0 */ v_usec = pps_ff[1] - pps_ff[0]; } } /* * Here the frequency dispersion (stability) is updated. If it * is less than one-fourth the maximum (MAXFREQ), the frequency * offset is updated as well, but clamped to the tolerance. It * will be processed later by the clock() routine. */ v_usec = (v_usec >> 1) - pps_stabil; if (v_usec < 0) pps_stabil -= -v_usec >> PPS_AVG; else pps_stabil += v_usec >> PPS_AVG; if (pps_stabil > MAXFREQ >> 2) { pps_stbcnt++; time_status |= STA_PPSWANDER; return; } if (time_status & STA_PPSFREQ) { if (u_usec < 0) { pps_freq -= -u_usec >> PPS_AVG; if (pps_freq < -time_tolerance) pps_freq = -time_tolerance; u_usec = -u_usec; } else { pps_freq += u_usec >> PPS_AVG; if (pps_freq > time_tolerance) pps_freq = time_tolerance; } } /* * Here the calibration interval is adjusted. If the maximum * time difference is greater than tick / 4, reduce the interval * by half. If this is not the case for four consecutive * intervals, double the interval. */ if (u_usec << pps_shift > bigtick >> 2) { pps_intcnt = 0; if (pps_shift > PPS_SHIFT) pps_shift--; } else if (pps_intcnt >= 4) { pps_intcnt = 0; if (pps_shift < PPS_SHIFTMAX) pps_shift++; } else pps_intcnt++; /* * If recovering from kmdb, then make sure the tod chip gets resynced. * If we took an early exit above, then we don't yet have a stable * calibration signal to lock onto, so don't mark the tod for sync * until we get all the way here. */ { int s = hr_clock_lock(); tod_needsync = 1; hr_clock_unlock(s); } } /* * Handle clock tick processing for a thread. * Check for timer action, enforce CPU rlimit, do profiling etc. */ void clock_tick(kthread_t *t, int pending) { struct proc *pp; klwp_id_t lwp; struct as *as; clock_t ticks; int poke = 0; /* notify another CPU */ int user_mode; size_t rss; int i, total_usec, usec; rctl_qty_t secs; ASSERT(pending > 0); /* Must be operating on a lwp/thread */ if ((lwp = ttolwp(t)) == NULL) { panic("clock_tick: no lwp"); /*NOTREACHED*/ } for (i = 0; i < pending; i++) { CL_TICK(t); /* Class specific tick processing */ DTRACE_SCHED1(tick, kthread_t *, t); } pp = ttoproc(t); /* pp->p_lock makes sure that the thread does not exit */ ASSERT(MUTEX_HELD(&pp->p_lock)); user_mode = (lwp->lwp_state == LWP_USER); ticks = (pp->p_utime + pp->p_stime) % hz; /* * Update process times. Should use high res clock and state * changes instead of statistical sampling method. XXX */ if (user_mode) { pp->p_utime += pending; } else { pp->p_stime += pending; } pp->p_ttime += pending; as = pp->p_as; /* * Update user profiling statistics. Get the pc from the * lwp when the AST happens. */ if (pp->p_prof.pr_scale) { atomic_add_32(&lwp->lwp_oweupc, (int32_t)pending); if (user_mode) { poke = 1; aston(t); } } /* * If CPU was in user state, process lwp-virtual time * interval timer. The value passed to itimerdecr() has to be * in microseconds and has to be less than one second. Hence * this loop. */ total_usec = usec_per_tick * pending; while (total_usec > 0) { usec = MIN(total_usec, (MICROSEC - 1)); if (user_mode && timerisset(&lwp->lwp_timer[ITIMER_VIRTUAL].it_value) && itimerdecr(&lwp->lwp_timer[ITIMER_VIRTUAL], usec) == 0) { poke = 1; sigtoproc(pp, t, SIGVTALRM); } total_usec -= usec; } /* * If CPU was in user state, process lwp-profile * interval timer. */ total_usec = usec_per_tick * pending; while (total_usec > 0) { usec = MIN(total_usec, (MICROSEC - 1)); if (timerisset(&lwp->lwp_timer[ITIMER_PROF].it_value) && itimerdecr(&lwp->lwp_timer[ITIMER_PROF], usec) == 0) { poke = 1; sigtoproc(pp, t, SIGPROF); } total_usec -= usec; } /* * Enforce CPU resource controls: * (a) process.max-cpu-time resource control * * Perform the check only if we have accumulated more a second. */ if ((ticks + pending) >= hz) { (void) rctl_test(rctlproc_legacy[RLIMIT_CPU], pp->p_rctls, pp, (pp->p_utime + pp->p_stime)/hz, RCA_UNSAFE_SIGINFO); } /* * (b) task.max-cpu-time resource control * * If we have accumulated enough ticks, increment the task CPU * time usage and test for the resource limit. This minimizes the * number of calls to the rct_test(). The task CPU time mutex * is highly contentious as many processes can be sharing a task. */ if (pp->p_ttime >= clock_tick_proc_max) { secs = task_cpu_time_incr(pp->p_task, pp->p_ttime); pp->p_ttime = 0; if (secs) { (void) rctl_test(rc_task_cpu_time, pp->p_task->tk_rctls, pp, secs, RCA_UNSAFE_SIGINFO); } } /* * Update memory usage for the currently running process. */ rss = rm_asrss(as); PTOU(pp)->u_mem += rss; if (rss > PTOU(pp)->u_mem_max) PTOU(pp)->u_mem_max = rss; /* * Notify the CPU the thread is running on. */ if (poke && t->t_cpu != CPU) poke_cpu(t->t_cpu->cpu_id); } void profil_tick(uintptr_t upc) { int ticks; proc_t *p = ttoproc(curthread); klwp_t *lwp = ttolwp(curthread); struct prof *pr = &p->p_prof; do { ticks = lwp->lwp_oweupc; } while (cas32(&lwp->lwp_oweupc, ticks, 0) != ticks); mutex_enter(&p->p_pflock); if (pr->pr_scale >= 2 && upc >= pr->pr_off) { /* * Old-style profiling */ uint16_t *slot = pr->pr_base; uint16_t old, new; if (pr->pr_scale != 2) { uintptr_t delta = upc - pr->pr_off; uintptr_t byteoff = ((delta >> 16) * pr->pr_scale) + (((delta & 0xffff) * pr->pr_scale) >> 16); if (byteoff >= (uintptr_t)pr->pr_size) { mutex_exit(&p->p_pflock); return; } slot += byteoff / sizeof (uint16_t); } if (fuword16(slot, &old) < 0 || (new = old + ticks) > SHRT_MAX || suword16(slot, new) < 0) { pr->pr_scale = 0; } } else if (pr->pr_scale == 1) { /* * PC Sampling */ model_t model = lwp_getdatamodel(lwp); int result; #ifdef __lint model = model; #endif while (ticks-- > 0) { if (pr->pr_samples == pr->pr_size) { /* buffer full, turn off sampling */ pr->pr_scale = 0; break; } switch (SIZEOF_PTR(model)) { case sizeof (uint32_t): result = suword32(pr->pr_base, (uint32_t)upc); break; #ifdef _LP64 case sizeof (uint64_t): result = suword64(pr->pr_base, (uint64_t)upc); break; #endif default: cmn_err(CE_WARN, "profil_tick: unexpected " "data model"); result = -1; break; } if (result != 0) { pr->pr_scale = 0; break; } pr->pr_base = (caddr_t)pr->pr_base + SIZEOF_PTR(model); pr->pr_samples++; } } mutex_exit(&p->p_pflock); } static void delay_wakeup(void *arg) { kthread_t *t = arg; mutex_enter(&t->t_delay_lock); cv_signal(&t->t_delay_cv); mutex_exit(&t->t_delay_lock); } /* * The delay(9F) man page indicates that it can only be called from user or * kernel context - detect and diagnose bad calls. The following macro will * produce a limited number of messages identifying bad callers. This is done * in a macro so that caller() is meaningful. When a bad caller is identified, * switching to 'drv_usecwait(TICK_TO_USEC(ticks));' may be appropriate. */ #define DELAY_CONTEXT_CHECK() { \ uint32_t m; \ char *f; \ ulong_t off; \ \ m = delay_from_interrupt_msg; \ if (delay_from_interrupt_diagnose && servicing_interrupt() && \ !panicstr && !devinfo_freeze && \ atomic_cas_32(&delay_from_interrupt_msg, m ? m : 1, m-1)) { \ f = modgetsymname((uintptr_t)caller(), &off); \ cmn_err(CE_WARN, "delay(9F) called from " \ "interrupt context: %s`%s", \ mod_containing_pc(caller()), f ? f : "..."); \ } \ } /* * delay_common: common delay code. */ static void delay_common(clock_t ticks) { kthread_t *t = curthread; clock_t deadline; clock_t timeleft; callout_id_t id; /* If timeouts aren't running all we can do is spin. */ if (panicstr || devinfo_freeze) { /* Convert delay(9F) call into drv_usecwait(9F) call. */ if (ticks > 0) drv_usecwait(TICK_TO_USEC(ticks)); return; } deadline = ddi_get_lbolt() + ticks; while ((timeleft = deadline - ddi_get_lbolt()) > 0) { mutex_enter(&t->t_delay_lock); id = timeout_default(delay_wakeup, t, timeleft); cv_wait(&t->t_delay_cv, &t->t_delay_lock); mutex_exit(&t->t_delay_lock); (void) untimeout_default(id, 0); } } /* * Delay specified number of clock ticks. */ void delay(clock_t ticks) { DELAY_CONTEXT_CHECK(); delay_common(ticks); } /* * Delay a random number of clock ticks between 1 and ticks. */ void delay_random(clock_t ticks) { int r; DELAY_CONTEXT_CHECK(); (void) random_get_pseudo_bytes((void *)&r, sizeof (r)); if (ticks == 0) ticks = 1; ticks = (r % ticks) + 1; delay_common(ticks); } /* * Like delay, but interruptible by a signal. */ int delay_sig(clock_t ticks) { kthread_t *t = curthread; clock_t deadline; clock_t rc; /* If timeouts aren't running all we can do is spin. */ if (panicstr || devinfo_freeze) { if (ticks > 0) drv_usecwait(TICK_TO_USEC(ticks)); return (0); } deadline = ddi_get_lbolt() + ticks; mutex_enter(&t->t_delay_lock); do { rc = cv_timedwait_sig(&t->t_delay_cv, &t->t_delay_lock, deadline); /* loop until past deadline or signaled */ } while (rc > 0); mutex_exit(&t->t_delay_lock); if (rc == 0) return (EINTR); return (0); } #define SECONDS_PER_DAY 86400 /* * Initialize the system time based on the TOD chip. approx is used as * an approximation of time (e.g. from the filesystem) in the event that * the TOD chip has been cleared or is unresponsive. An approx of -1 * means the filesystem doesn't keep time. */ void clkset(time_t approx) { timestruc_t ts; int spl; int set_clock = 0; mutex_enter(&tod_lock); ts = tod_get(); if (ts.tv_sec > 365 * SECONDS_PER_DAY) { /* * If the TOD chip is reporting some time after 1971, * then it probably didn't lose power or become otherwise * cleared in the recent past; check to assure that * the time coming from the filesystem isn't in the future * according to the TOD chip. */ if (approx != -1 && approx > ts.tv_sec) { cmn_err(CE_WARN, "Last shutdown is later " "than time on time-of-day chip; check date."); } } else { /* * If the TOD chip isn't giving correct time, set it to the * greater of i) approx and ii) 1987. That way if approx * is negative or is earlier than 1987, we set the clock * back to a time when Oliver North, ALF and Dire Straits * were all on the collective brain: 1987. */ timestruc_t tmp; time_t diagnose_date = (1987 - 1970) * 365 * SECONDS_PER_DAY; ts.tv_sec = (approx > diagnose_date ? approx : diagnose_date); ts.tv_nsec = 0; /* * Attempt to write the new time to the TOD chip. Set spl high * to avoid getting preempted between the tod_set and tod_get. */ spl = splhi(); tod_set(ts); tmp = tod_get(); splx(spl); if (tmp.tv_sec != ts.tv_sec && tmp.tv_sec != ts.tv_sec + 1) { tod_broken = 1; dosynctodr = 0; cmn_err(CE_WARN, "Time-of-day chip unresponsive."); } else { cmn_err(CE_WARN, "Time-of-day chip had " "incorrect date; check and reset."); } set_clock = 1; } if (!boot_time) { boot_time = ts.tv_sec; set_clock = 1; } if (set_clock) set_hrestime(&ts); mutex_exit(&tod_lock); } int timechanged; /* for testing if the system time has been reset */ void set_hrestime(timestruc_t *ts) { int spl = hr_clock_lock(); hrestime = *ts; membar_enter(); /* hrestime must be visible before timechanged++ */ timedelta = 0; timechanged++; hr_clock_unlock(spl); callout_hrestime(); } static uint_t deadman_seconds; static uint32_t deadman_panics; static int deadman_enabled = 0; static int deadman_panic_timers = 1; static void deadman(void) { if (panicstr) { /* * During panic, other CPUs besides the panic * master continue to handle cyclics and some other * interrupts. The code below is intended to be * single threaded, so any CPU other than the master * must keep out. */ if (CPU->cpu_id != panic_cpu.cpu_id) return; if (!deadman_panic_timers) return; /* allow all timers to be manually disabled */ /* * If we are generating a crash dump or syncing filesystems and * the corresponding timer is set, decrement it and re-enter * the panic code to abort it and advance to the next state. * The panic states and triggers are explained in panic.c. */ if (panic_dump) { if (dump_timeleft && (--dump_timeleft == 0)) { panic("panic dump timeout"); /*NOTREACHED*/ } } else if (panic_sync) { if (sync_timeleft && (--sync_timeleft == 0)) { panic("panic sync timeout"); /*NOTREACHED*/ } } return; } if (deadman_counter != CPU->cpu_deadman_counter) { CPU->cpu_deadman_counter = deadman_counter; CPU->cpu_deadman_countdown = deadman_seconds; return; } if (--CPU->cpu_deadman_countdown > 0) return; /* * Regardless of whether or not we actually bring the system down, * bump the deadman_panics variable. * * N.B. deadman_panics is incremented once for each CPU that * passes through here. It's expected that all the CPUs will * detect this condition within one second of each other, so * when deadman_enabled is off, deadman_panics will * typically be a multiple of the total number of CPUs in * the system. */ atomic_add_32(&deadman_panics, 1); if (!deadman_enabled) { CPU->cpu_deadman_countdown = deadman_seconds; return; } /* * If we're here, we want to bring the system down. */ panic("deadman: timed out after %d seconds of clock " "inactivity", deadman_seconds); /*NOTREACHED*/ } /*ARGSUSED*/ static void deadman_online(void *arg, cpu_t *cpu, cyc_handler_t *hdlr, cyc_time_t *when) { cpu->cpu_deadman_counter = 0; cpu->cpu_deadman_countdown = deadman_seconds; hdlr->cyh_func = (cyc_func_t)deadman; hdlr->cyh_level = CY_HIGH_LEVEL; hdlr->cyh_arg = NULL; /* * Stagger the CPUs so that they don't all run deadman() at * the same time. Simplest reason to do this is to make it * more likely that only one CPU will panic in case of a * timeout. This is (strictly speaking) an aesthetic, not a * technical consideration. */ when->cyt_when = cpu->cpu_id * (NANOSEC / NCPU); when->cyt_interval = NANOSEC; } void deadman_init(void) { cyc_omni_handler_t hdlr; if (deadman_seconds == 0) deadman_seconds = snoop_interval / MICROSEC; if (snooping) deadman_enabled = 1; hdlr.cyo_online = deadman_online; hdlr.cyo_offline = NULL; hdlr.cyo_arg = NULL; mutex_enter(&cpu_lock); deadman_cyclic = cyclic_add_omni(&hdlr); mutex_exit(&cpu_lock); } /* * tod_fault() is for updating tod validate mechanism state: * (1) TOD_NOFAULT: for resetting the state to 'normal'. * currently used for debugging only * (2) The following four cases detected by tod validate mechanism: * TOD_REVERSED: current tod value is less than previous value. * TOD_STALLED: current tod value hasn't advanced. * TOD_JUMPED: current tod value advanced too far from previous value. * TOD_RATECHANGED: the ratio between average tod delta and * average tick delta has changed. * (3) TOD_RDONLY: when the TOD clock is not writeable e.g. because it is * a virtual TOD provided by a hypervisor. */ enum tod_fault_type tod_fault(enum tod_fault_type ftype, int off) { ASSERT(MUTEX_HELD(&tod_lock)); if (tod_faulted != ftype) { switch (ftype) { case TOD_NOFAULT: plat_tod_fault(TOD_NOFAULT); cmn_err(CE_NOTE, "Restarted tracking " "Time of Day clock."); tod_faulted = ftype; break; case TOD_REVERSED: case TOD_JUMPED: if (tod_faulted == TOD_NOFAULT) { plat_tod_fault(ftype); cmn_err(CE_WARN, "Time of Day clock error: " "reason [%s by 0x%x]. -- " " Stopped tracking Time Of Day clock.", tod_fault_table[ftype], off); tod_faulted = ftype; } break; case TOD_STALLED: case TOD_RATECHANGED: if (tod_faulted == TOD_NOFAULT) { plat_tod_fault(ftype); cmn_err(CE_WARN, "Time of Day clock error: " "reason [%s]. -- " " Stopped tracking Time Of Day clock.", tod_fault_table[ftype]); tod_faulted = ftype; } break; case TOD_RDONLY: if (tod_faulted == TOD_NOFAULT) { plat_tod_fault(ftype); cmn_err(CE_NOTE, "!Time of Day clock is " "Read-Only; set of Date/Time will not " "persist across reboot."); tod_faulted = ftype; } break; default: break; } } return (tod_faulted); } void tod_fault_reset() { tod_fault_reset_flag = 1; } /* * tod_validate() is used for checking values returned by tod_get(). * Four error cases can be detected by this routine: * TOD_REVERSED: current tod value is less than previous. * TOD_STALLED: current tod value hasn't advanced. * TOD_JUMPED: current tod value advanced too far from previous value. * TOD_RATECHANGED: the ratio between average tod delta and * average tick delta has changed. */ time_t tod_validate(time_t tod) { time_t diff_tod; hrtime_t diff_tick; long dtick; int dtick_delta; int off = 0; enum tod_fault_type tod_bad = TOD_NOFAULT; static int firsttime = 1; static time_t prev_tod = 0; static hrtime_t prev_tick = 0; static long dtick_avg = TOD_REF_FREQ; hrtime_t tick = gethrtime(); ASSERT(MUTEX_HELD(&tod_lock)); /* * tod_validate_enable is patchable via /etc/system. * If TOD is already faulted, or if TOD validation is deferred, * there is nothing to do. */ if ((tod_validate_enable == 0) || (tod_faulted != TOD_NOFAULT) || tod_validate_deferred) { return (tod); } /* * Update prev_tod and prev_tick values for first run */ if (firsttime) { firsttime = 0; prev_tod = tod; prev_tick = tick; return (tod); } /* * For either of these conditions, we need to reset ourself * and start validation from zero since each condition * indicates that the TOD will be updated with new value * Also, note that tod_needsync will be reset in clock() */ if (tod_needsync || tod_fault_reset_flag) { firsttime = 1; prev_tod = 0; prev_tick = 0; dtick_avg = TOD_REF_FREQ; if (tod_fault_reset_flag) tod_fault_reset_flag = 0; return (tod); } /* test hook */ switch (tod_unit_test) { case 1: /* for testing jumping tod */ tod += tod_test_injector; tod_unit_test = 0; break; case 2: /* for testing stuck tod bit */ tod |= 1 << tod_test_injector; tod_unit_test = 0; break; case 3: /* for testing stalled tod */ tod = prev_tod; tod_unit_test = 0; break; case 4: /* reset tod fault status */ (void) tod_fault(TOD_NOFAULT, 0); tod_unit_test = 0; break; default: break; } diff_tod = tod - prev_tod; diff_tick = tick - prev_tick; ASSERT(diff_tick >= 0); if (diff_tod < 0) { /* ERROR - tod reversed */ tod_bad = TOD_REVERSED; off = (int)(prev_tod - tod); } else if (diff_tod == 0) { /* tod did not advance */ if (diff_tick > TOD_STALL_THRESHOLD) { /* ERROR - tod stalled */ tod_bad = TOD_STALLED; } else { /* * Make sure we don't update prev_tick * so that diff_tick is calculated since * the first diff_tod == 0 */ return (tod); } } else { /* calculate dtick */ dtick = diff_tick / diff_tod; /* update dtick averages */ dtick_avg += ((dtick - dtick_avg) / TOD_FILTER_N); /* * Calculate dtick_delta as * variation from reference freq in quartiles */ dtick_delta = (dtick_avg - TOD_REF_FREQ) / (TOD_REF_FREQ >> 2); /* * Even with a perfectly functioning TOD device, * when the number of elapsed seconds is low the * algorithm can calculate a rate that is beyond * tolerance, causing an error. The algorithm is * inaccurate when elapsed time is low (less than * 5 seconds). */ if (diff_tod > 4) { if (dtick < TOD_JUMP_THRESHOLD) { /* ERROR - tod jumped */ tod_bad = TOD_JUMPED; off = (int)diff_tod; } else if (dtick_delta) { /* ERROR - change in clock rate */ tod_bad = TOD_RATECHANGED; } } } if (tod_bad != TOD_NOFAULT) { (void) tod_fault(tod_bad, off); /* * Disable dosynctodr since we are going to fault * the TOD chip anyway here */ dosynctodr = 0; /* * Set tod to the correct value from hrestime */ tod = hrestime.tv_sec; } prev_tod = tod; prev_tick = tick; return (tod); } static void calcloadavg(int nrun, uint64_t *hp_ave) { static int64_t f[3] = { 135, 27, 9 }; uint_t i; int64_t q, r; /* * Compute load average over the last 1, 5, and 15 minutes * (60, 300, and 900 seconds). The constants in f[3] are for * exponential decay: * (1 - exp(-1/60)) << 13 = 135, * (1 - exp(-1/300)) << 13 = 27, * (1 - exp(-1/900)) << 13 = 9. */ /* * a little hoop-jumping to avoid integer overflow */ for (i = 0; i < 3; i++) { q = (hp_ave[i] >> 16) << 7; r = (hp_ave[i] & 0xffff) << 7; hp_ave[i] += ((nrun - q) * f[i] - ((r * f[i]) >> 16)) >> 4; } } /* * lbolt_hybrid() is used by ddi_get_lbolt() and ddi_get_lbolt64() to * calculate the value of lbolt according to the current mode. In the event * driven mode (the default), lbolt is calculated by dividing the current hires * time by the number of nanoseconds per clock tick. In the cyclic driven mode * an internal variable is incremented at each firing of the lbolt cyclic * and returned by lbolt_cyclic_driven(). * * The system will transition from event to cyclic driven mode when the number * of calls to lbolt_event_driven() exceeds the (per CPU) threshold within a * window of time. It does so by reprograming lbolt_cyclic from CY_INFINITY to * nsec_per_tick. The lbolt cyclic will remain ON while at least one CPU is * causing enough activity to cross the thresholds. */ static int64_t lbolt_bootstrap(void) { return (0); } /* ARGSUSED */ uint_t lbolt_ev_to_cyclic(caddr_t arg1, caddr_t arg2) { hrtime_t ts, exp; int ret; ASSERT(lbolt_hybrid != lbolt_cyclic_driven); kpreempt_disable(); ts = gethrtime(); lb_info->lbi_internal = (ts/nsec_per_tick); /* * Align the next expiration to a clock tick boundary. */ exp = ts + nsec_per_tick - 1; exp = (exp/nsec_per_tick) * nsec_per_tick; ret = cyclic_reprogram(lb_info->lbi_cyclic_id, exp); ASSERT(ret); lbolt_hybrid = lbolt_cyclic_driven; lb_info->lbi_cyc_deactivate = B_FALSE; lb_info->lbi_cyc_deac_start = lb_info->lbi_internal; kpreempt_enable(); ret = atomic_dec_32_nv(&lb_info->lbi_token); ASSERT(ret == 0); return (1); } int64_t lbolt_event_driven(void) { hrtime_t ts; int64_t lb; int ret, cpu = CPU->cpu_seqid; ts = gethrtime(); ASSERT(ts > 0); ASSERT(nsec_per_tick > 0); lb = (ts/nsec_per_tick); /* * Switch to cyclic mode if the number of calls to this routine * has reached the threshold within the interval. */ if ((lb - lb_cpu[cpu].lbc_cnt_start) < lb_info->lbi_thresh_interval) { if (--lb_cpu[cpu].lbc_counter == 0) { /* * Reached the threshold within the interval, reset * the usage statistics. */ lb_cpu[cpu].lbc_counter = lb_info->lbi_thresh_calls; lb_cpu[cpu].lbc_cnt_start = lb; /* * Make sure only one thread reprograms the * lbolt cyclic and changes the mode. */ if (panicstr == NULL && atomic_cas_32(&lb_info->lbi_token, 0, 1) == 0) { if (lbolt_hybrid == lbolt_cyclic_driven) { ret = atomic_dec_32_nv( &lb_info->lbi_token); ASSERT(ret == 0); return (lb); } lbolt_softint_post(); } } } else { /* * Exceeded the interval, reset the usage statistics. */ lb_cpu[cpu].lbc_counter = lb_info->lbi_thresh_calls; lb_cpu[cpu].lbc_cnt_start = lb; } ASSERT(lb >= lb_info->lbi_debug_time); return (lb - lb_info->lbi_debug_time); } int64_t lbolt_cyclic_driven(void) { int64_t lb = lb_info->lbi_internal; int cpu = CPU->cpu_seqid; if ((lb - lb_cpu[cpu].lbc_cnt_start) < lb_info->lbi_thresh_interval) { if (lb_cpu[cpu].lbc_counter == 0) /* * Reached the threshold within the interval, * prevent the lbolt cyclic from turning itself * off. */ lb_info->lbi_cyc_deactivate = B_FALSE; else lb_cpu[cpu].lbc_counter--; } else { /* * Only reset the usage statistics when the interval has * exceeded. */ lb_cpu[cpu].lbc_counter = lb_info->lbi_thresh_calls; lb_cpu[cpu].lbc_cnt_start = lb; } ASSERT(lb >= lb_info->lbi_debug_time); return (lb - lb_info->lbi_debug_time); } /* * The lbolt_cyclic() routine will fire at a nsec_per_tick rate to satisfy * performance needs of ddi_get_lbolt() and ddi_get_lbolt64() consumers. * It is inactive by default, and will be activated when switching from event * to cyclic driven lbolt. The cyclic will turn itself off unless signaled * by lbolt_cyclic_driven(). */ static void lbolt_cyclic(void) { int ret; lb_info->lbi_internal++; if (!lbolt_cyc_only) { if (lb_info->lbi_cyc_deactivate) { /* * Switching from cyclic to event driven mode. */ if (atomic_cas_32(&lb_info->lbi_token, 0, 1) == 0) { if (lbolt_hybrid == lbolt_event_driven) { ret = atomic_dec_32_nv( &lb_info->lbi_token); ASSERT(ret == 0); return; } kpreempt_disable(); lbolt_hybrid = lbolt_event_driven; ret = cyclic_reprogram(lb_info->lbi_cyclic_id, CY_INFINITY); ASSERT(ret); kpreempt_enable(); ret = atomic_dec_32_nv(&lb_info->lbi_token); ASSERT(ret == 0); } } /* * The lbolt cyclic should not try to deactivate itself before * the sampling period has elapsed. */ if (lb_info->lbi_internal - lb_info->lbi_cyc_deac_start >= lb_info->lbi_thresh_interval) { lb_info->lbi_cyc_deactivate = B_TRUE; lb_info->lbi_cyc_deac_start = lb_info->lbi_internal; } } } /* * Since the lbolt service was historically cyclic driven, it must be 'stopped' * when the system drops into the kernel debugger. lbolt_debug_entry() is * called by the KDI system claim callbacks to record a hires timestamp at * debug enter time. lbolt_debug_return() is called by the sistem release * callbacks to account for the time spent in the debugger. The value is then * accumulated in the lb_info structure and used by lbolt_event_driven() and * lbolt_cyclic_driven(), as well as the mdb_get_lbolt() routine. */ void lbolt_debug_entry(void) { lb_info->lbi_debug_ts = gethrtime(); } void lbolt_debug_return(void) { if (nsec_per_tick > 0) lb_info->lbi_debug_time += ((gethrtime() - lb_info->lbi_debug_ts)/nsec_per_tick); lb_info->lbi_debug_ts = 0; }