svn commit: r212541 - in head/sys: amd64/amd64 amd64/include dev/acpica i386/i386 i386/include kern mips/include mips/mips pc98/pc98 powerpc/aim powerpc/booke powerpc/include powerpc/powerpc sparc6...

Alexander Motin mav at FreeBSD.org
Mon Sep 13 07:25:36 UTC 2010


Author: mav
Date: Mon Sep 13 07:25:35 2010
New Revision: 212541
URL: http://svn.freebsd.org/changeset/base/212541

Log:
  Refactor timer management code with priority to one-shot operation mode.
  The main goal of this is to generate timer interrupts only when there is
  some work to do. When CPU is busy interrupts are generating at full rate
  of hz + stathz to fullfill scheduler and timekeeping requirements. But
  when CPU is idle, only minimum set of interrupts (down to 8 interrupts per
  second per CPU now), needed to handle scheduled callouts is executed.
  This allows significantly increase idle CPU sleep time, increasing effect
  of static power-saving technologies. Also it should reduce host CPU load
  on virtualized systems, when guest system is idle.
  
  There is set of tunables, also available as writable sysctls, allowing to
  control wanted event timer subsystem behavior:
    kern.eventtimer.timer - allows to choose event timer hardware to use.
  On x86 there is up to 4 different kinds of timers. Depending on whether
  chosen timer is per-CPU, behavior of other options slightly differs.
    kern.eventtimer.periodic - allows to choose periodic and one-shot
  operation mode. In periodic mode, current timer hardware taken as the only
  source of time for time events. This mode is quite alike to previous kernel
  behavior. One-shot mode instead uses currently selected time counter
  hardware to schedule all needed events one by one and program timer to
  generate interrupt exactly in specified time. Default value depends of
  chosen timer capabilities, but one-shot mode is preferred, until other is
  forced by user or hardware.
    kern.eventtimer.singlemul - in periodic mode specifies how much times
  higher timer frequency should be, to not strictly alias hardclock() and
  statclock() events. Default values are 2 and 4, but could be reduced to 1
  if extra interrupts are unwanted.
    kern.eventtimer.idletick - makes each CPU to receive every timer interrupt
  independently of whether they busy or not. By default this options is
  disabled. If chosen timer is per-CPU and runs in periodic mode, this option
  has no effect - all interrupts are generating.
  
  As soon as this patch modifies cpu_idle() on some platforms, I have also
  refactored one on x86. Now it makes use of MONITOR/MWAIT instrunctions
  (if supported) under high sleep/wakeup rate, as fast alternative to other
  methods. It allows SMP scheduler to wake up sleeping CPUs much faster
  without using IPI, significantly increasing performance on some highly
  task-switching loads.
  
  Tested by:	many (on i386, amd64, sparc64 and powerc)
  H/W donated by:	Gheorghe Ardelean
  Sponsored by:	iXsystems, Inc.

Modified:
  head/sys/amd64/amd64/machdep.c
  head/sys/amd64/amd64/mp_machdep.c
  head/sys/amd64/include/apicvar.h
  head/sys/dev/acpica/acpi_cpu.c
  head/sys/dev/acpica/acpi_hpet.c
  head/sys/i386/i386/machdep.c
  head/sys/i386/i386/mp_machdep.c
  head/sys/i386/include/apicvar.h
  head/sys/kern/kern_clock.c
  head/sys/kern/kern_clocksource.c
  head/sys/kern/kern_et.c
  head/sys/kern/kern_tc.c
  head/sys/kern/kern_timeout.c
  head/sys/kern/sched_4bsd.c
  head/sys/kern/sched_ule.c
  head/sys/mips/include/smp.h
  head/sys/mips/mips/mp_machdep.c
  head/sys/pc98/pc98/machdep.c
  head/sys/powerpc/aim/machdep.c
  head/sys/powerpc/booke/machdep.c
  head/sys/powerpc/include/smp.h
  head/sys/powerpc/powerpc/mp_machdep.c
  head/sys/sparc64/include/intr_machdep.h
  head/sys/sparc64/include/smp.h
  head/sys/sparc64/sparc64/intr_machdep.c
  head/sys/sparc64/sparc64/mp_machdep.c
  head/sys/sun4v/include/intr_machdep.h
  head/sys/sun4v/include/smp.h
  head/sys/sun4v/sun4v/intr_machdep.c
  head/sys/sun4v/sun4v/mp_machdep.c
  head/sys/sys/callout.h
  head/sys/sys/sched.h
  head/sys/sys/systm.h
  head/sys/sys/timeet.h
  head/sys/sys/timetc.h
  head/sys/x86/x86/local_apic.c

Modified: head/sys/amd64/amd64/machdep.c
==============================================================================
--- head/sys/amd64/amd64/machdep.c	Mon Sep 13 07:22:14 2010	(r212540)
+++ head/sys/amd64/amd64/machdep.c	Mon Sep 13 07:25:35 2010	(r212541)
@@ -585,59 +585,89 @@ cpu_halt(void)
 }
 
 void (*cpu_idle_hook)(void) = NULL;	/* ACPI idle hook. */
+static int	cpu_ident_amdc1e = 0;	/* AMD C1E supported. */
+static int	idle_mwait = 1;		/* Use MONITOR/MWAIT for short idle. */
+TUNABLE_INT("machdep.idle_mwait", &idle_mwait);
+SYSCTL_INT(_machdep, OID_AUTO, idle_mwait, CTLFLAG_RW, &idle_mwait,
+    0, "Use MONITOR/MWAIT for short idle");
+
+#define	STATE_RUNNING	0x0
+#define	STATE_MWAIT	0x1
+#define	STATE_SLEEPING	0x2
 
 static void
-cpu_idle_hlt(int busy)
+cpu_idle_acpi(int busy)
 {
-	/*
-	 * we must absolutely guarentee that hlt is the next instruction
-	 * after sti or we introduce a timing window.
-	 */
+	int *state;
+
+	state = (int *)PCPU_PTR(monitorbuf);
+	*state = STATE_SLEEPING;
 	disable_intr();
-  	if (sched_runnable())
+	if (sched_runnable())
 		enable_intr();
+	else if (cpu_idle_hook)
+		cpu_idle_hook();
 	else
 		__asm __volatile("sti; hlt");
+	*state = STATE_RUNNING;
 }
 
 static void
-cpu_idle_acpi(int busy)
+cpu_idle_hlt(int busy)
 {
+	int *state;
+
+	state = (int *)PCPU_PTR(monitorbuf);
+	*state = STATE_SLEEPING;
+	/*
+	 * We must absolutely guarentee that hlt is the next instruction
+	 * after sti or we introduce a timing window.
+	 */
 	disable_intr();
-  	if (sched_runnable())
+	if (sched_runnable())
 		enable_intr();
-	else if (cpu_idle_hook)
-		cpu_idle_hook();
 	else
 		__asm __volatile("sti; hlt");
+	*state = STATE_RUNNING;
 }
 
-static int cpu_ident_amdc1e = 0;
+/*
+ * MWAIT cpu power states.  Lower 4 bits are sub-states.
+ */
+#define	MWAIT_C0	0xf0
+#define	MWAIT_C1	0x00
+#define	MWAIT_C2	0x10
+#define	MWAIT_C3	0x20
+#define	MWAIT_C4	0x30
 
-static int
-cpu_probe_amdc1e(void)
+static void
+cpu_idle_mwait(int busy)
 {
-	int i;
+	int *state;
 
-	/*
-	 * Forget it, if we're not using local APIC timer.
-	 */
-	if (resource_disabled("apic", 0) ||
-	    (resource_int_value("apic", 0, "clock", &i) == 0 && i == 0))
-		return (0);
-
-	/*
-	 * Detect the presence of C1E capability mostly on latest
-	 * dual-cores (or future) k8 family.
-	 */
-	if (cpu_vendor_id == CPU_VENDOR_AMD &&
-	    (cpu_id & 0x00000f00) == 0x00000f00 &&
-	    (cpu_id & 0x0fff0000) >=  0x00040000) {
-		cpu_ident_amdc1e = 1;
-		return (1);
+	state = (int *)PCPU_PTR(monitorbuf);
+	*state = STATE_MWAIT;
+	if (!sched_runnable()) {
+		cpu_monitor(state, 0, 0);
+		if (*state == STATE_MWAIT)
+			cpu_mwait(0, MWAIT_C1);
 	}
+	*state = STATE_RUNNING;
+}
 
-	return (0);
+static void
+cpu_idle_spin(int busy)
+{
+	int *state;
+	int i;
+
+	state = (int *)PCPU_PTR(monitorbuf);
+	*state = STATE_RUNNING;
+	for (i = 0; i < 1000; i++) {
+		if (sched_runnable())
+			return;
+		cpu_spinwait();
+	}
 }
 
 /*
@@ -655,110 +685,83 @@ cpu_probe_amdc1e(void)
 #define	AMDK8_CMPHALT		(AMDK8_SMIONCMPHALT | AMDK8_C1EONCMPHALT)
 
 static void
-cpu_idle_amdc1e(int busy)
+cpu_probe_amdc1e(void)
 {
 
-	disable_intr();
-	if (sched_runnable())
-		enable_intr();
-	else {
-		uint64_t msr;
-
-		msr = rdmsr(MSR_AMDK8_IPM);
-		if (msr & AMDK8_CMPHALT)
-			wrmsr(MSR_AMDK8_IPM, msr & ~AMDK8_CMPHALT);
-
-		if (cpu_idle_hook)
-			cpu_idle_hook();
-		else
-			__asm __volatile("sti; hlt");
+	/*
+	 * Detect the presence of C1E capability mostly on latest
+	 * dual-cores (or future) k8 family.
+	 */
+	if (cpu_vendor_id == CPU_VENDOR_AMD &&
+	    (cpu_id & 0x00000f00) == 0x00000f00 &&
+	    (cpu_id & 0x0fff0000) >=  0x00040000) {
+		cpu_ident_amdc1e = 1;
 	}
 }
 
-static void
-cpu_idle_spin(int busy)
-{
-	return;
-}
-
 void (*cpu_idle_fn)(int) = cpu_idle_acpi;
 
 void
 cpu_idle(int busy)
 {
+	uint64_t msr;
+
+	CTR2(KTR_SPARE2, "cpu_idle(%d) at %d",
+	    busy, curcpu);
 #ifdef SMP
 	if (mp_grab_cpu_hlt())
 		return;
 #endif
-	cpu_idle_fn(busy);
-}
-
-/*
- * mwait cpu power states.  Lower 4 bits are sub-states.
- */
-#define	MWAIT_C0	0xf0
-#define	MWAIT_C1	0x00
-#define	MWAIT_C2	0x10
-#define	MWAIT_C3	0x20
-#define	MWAIT_C4	0x30
-
-#define	MWAIT_DISABLED	0x0
-#define	MWAIT_WOKEN	0x1
-#define	MWAIT_WAITING	0x2
+	/* If we are busy - try to use fast methods. */
+	if (busy) {
+		if ((cpu_feature2 & CPUID2_MON) && idle_mwait) {
+			cpu_idle_mwait(busy);
+			goto out;
+		}
+	}
 
-static void
-cpu_idle_mwait(int busy)
-{
-	int *mwait;
+	/* If we have time - switch timers into idle mode. */
+	if (!busy) {
+		critical_enter();
+		cpu_idleclock();
+	}
 
-	mwait = (int *)PCPU_PTR(monitorbuf);
-	*mwait = MWAIT_WAITING;
-	if (sched_runnable())
-		return;
-	cpu_monitor(mwait, 0, 0);
-	if (*mwait == MWAIT_WAITING)
-		cpu_mwait(0, MWAIT_C1);
-}
+	/* Apply AMD APIC timer C1E workaround. */
+	if (cpu_ident_amdc1e && cpu_disable_deep_sleep) {
+		msr = rdmsr(MSR_AMDK8_IPM);
+		if (msr & AMDK8_CMPHALT)
+			wrmsr(MSR_AMDK8_IPM, msr & ~AMDK8_CMPHALT);
+	}
 
-static void
-cpu_idle_mwait_hlt(int busy)
-{
-	int *mwait;
+	/* Call main idle method. */
+	cpu_idle_fn(busy);
 
-	mwait = (int *)PCPU_PTR(monitorbuf);
-	if (busy == 0) {
-		*mwait = MWAIT_DISABLED;
-		cpu_idle_hlt(busy);
-		return;
-	}
-	*mwait = MWAIT_WAITING;
-	if (sched_runnable())
-		return;
-	cpu_monitor(mwait, 0, 0);
-	if (*mwait == MWAIT_WAITING)
-		cpu_mwait(0, MWAIT_C1);
+	/* Switch timers mack into active mode. */
+	if (!busy) {
+		cpu_activeclock();
+		critical_exit();
+	}
+out:
+	CTR2(KTR_SPARE2, "cpu_idle(%d) at %d done",
+	    busy, curcpu);
 }
 
 int
 cpu_idle_wakeup(int cpu)
 {
 	struct pcpu *pcpu;
-	int *mwait;
+	int *state;
 
-	if (cpu_idle_fn == cpu_idle_spin)
-		return (1);
-	if (cpu_idle_fn != cpu_idle_mwait && cpu_idle_fn != cpu_idle_mwait_hlt)
-		return (0);
 	pcpu = pcpu_find(cpu);
-	mwait = (int *)pcpu->pc_monitorbuf;
+	state = (int *)pcpu->pc_monitorbuf;
 	/*
 	 * This doesn't need to be atomic since missing the race will
 	 * simply result in unnecessary IPIs.
 	 */
-	if (cpu_idle_fn == cpu_idle_mwait_hlt && *mwait == MWAIT_DISABLED)
+	if (*state == STATE_SLEEPING)
 		return (0);
-	*mwait = MWAIT_WOKEN;
-
+	if (*state == STATE_MWAIT)
+		*state = STATE_RUNNING;
 	return (1);
 }
 
@@ -771,8 +774,6 @@ struct {
 } idle_tbl[] = {
 	{ cpu_idle_spin, "spin" },
 	{ cpu_idle_mwait, "mwait" },
-	{ cpu_idle_mwait_hlt, "mwait_hlt" },
-	{ cpu_idle_amdc1e, "amdc1e" },
 	{ cpu_idle_hlt, "hlt" },
 	{ cpu_idle_acpi, "acpi" },
 	{ NULL, NULL }
@@ -791,8 +792,8 @@ idle_sysctl_available(SYSCTL_HANDLER_ARG
 		if (strstr(idle_tbl[i].id_name, "mwait") &&
 		    (cpu_feature2 & CPUID2_MON) == 0)
 			continue;
-		if (strcmp(idle_tbl[i].id_name, "amdc1e") == 0 &&
-		    cpu_ident_amdc1e == 0)
+		if (strcmp(idle_tbl[i].id_name, "acpi") == 0 &&
+		    cpu_idle_hook == NULL)
 			continue;
 		p += sprintf(p, "%s, ", idle_tbl[i].id_name);
 	}
@@ -801,6 +802,9 @@ idle_sysctl_available(SYSCTL_HANDLER_ARG
 	return (error);
 }
 
+SYSCTL_PROC(_machdep, OID_AUTO, idle_available, CTLTYPE_STRING | CTLFLAG_RD,
+    0, 0, idle_sysctl_available, "A", "list of available idle functions");
+
 static int
 idle_sysctl(SYSCTL_HANDLER_ARGS)
 {
@@ -824,8 +828,8 @@ idle_sysctl(SYSCTL_HANDLER_ARGS)
 		if (strstr(idle_tbl[i].id_name, "mwait") &&
 		    (cpu_feature2 & CPUID2_MON) == 0)
 			continue;
-		if (strcmp(idle_tbl[i].id_name, "amdc1e") == 0 &&
-		    cpu_ident_amdc1e == 0)
+		if (strcmp(idle_tbl[i].id_name, "acpi") == 0 &&
+		    cpu_idle_hook == NULL)
 			continue;
 		if (strcmp(idle_tbl[i].id_name, buf))
 			continue;
@@ -835,9 +839,6 @@ idle_sysctl(SYSCTL_HANDLER_ARGS)
 	return (EINVAL);
 }
 
-SYSCTL_PROC(_machdep, OID_AUTO, idle_available, CTLTYPE_STRING | CTLFLAG_RD,
-    0, 0, idle_sysctl_available, "A", "list of available idle functions");
-
 SYSCTL_PROC(_machdep, OID_AUTO, idle, CTLTYPE_STRING | CTLFLAG_RW, 0, 0,
     idle_sysctl, "A", "currently selected idle function");
 
@@ -1743,8 +1744,7 @@ hammer_time(u_int64_t modulep, u_int64_t
 	}
 #endif
 
-	if (cpu_probe_amdc1e())
-		cpu_idle_fn = cpu_idle_amdc1e;
+	cpu_probe_amdc1e();
 
 	/* Location of kernel stack for locore */
 	return ((u_int64_t)thread0.td_pcb);

Modified: head/sys/amd64/amd64/mp_machdep.c
==============================================================================
--- head/sys/amd64/amd64/mp_machdep.c	Mon Sep 13 07:22:14 2010	(r212540)
+++ head/sys/amd64/amd64/mp_machdep.c	Mon Sep 13 07:25:35 2010	(r212541)
@@ -118,7 +118,6 @@ u_long *ipi_invlcache_counts[MAXCPU];
 u_long *ipi_rendezvous_counts[MAXCPU];
 u_long *ipi_lazypmap_counts[MAXCPU];
 static u_long *ipi_hardclock_counts[MAXCPU];
-static u_long *ipi_statclock_counts[MAXCPU];
 #endif
 
 extern inthand_t IDTVEC(fast_syscall), IDTVEC(fast_syscall32);
@@ -1196,16 +1195,22 @@ smp_masked_invlpg_range(cpumask_t mask, 
 void
 ipi_bitmap_handler(struct trapframe frame)
 {
+	struct trapframe *oldframe;
+	struct thread *td;
 	int cpu = PCPU_GET(cpuid);
 	u_int ipi_bitmap;
 
+	critical_enter();
+	td = curthread;
+	td->td_intr_nesting_level++;
+	oldframe = td->td_intr_frame;
+	td->td_intr_frame = &frame;
 	ipi_bitmap = atomic_readandclear_int(&cpu_ipi_pending[cpu]);
-
 	if (ipi_bitmap & (1 << IPI_PREEMPT)) {
 #ifdef COUNT_IPIS
 		(*ipi_preempt_counts[cpu])++;
 #endif
-		sched_preempt(curthread);
+		sched_preempt(td);
 	}
 	if (ipi_bitmap & (1 << IPI_AST)) {
 #ifdef COUNT_IPIS
@@ -1217,14 +1222,11 @@ ipi_bitmap_handler(struct trapframe fram
 #ifdef COUNT_IPIS
 		(*ipi_hardclock_counts[cpu])++;
 #endif
-		hardclockintr(&frame);
-	}
-	if (ipi_bitmap & (1 << IPI_STATCLOCK)) {
-#ifdef COUNT_IPIS
-		(*ipi_statclock_counts[cpu])++;
-#endif
-		statclockintr(&frame);
+		hardclockintr();
 	}
+	td->td_intr_frame = oldframe;
+	td->td_intr_nesting_level--;
+	critical_exit();
 }
 
 /*
@@ -1579,8 +1581,6 @@ mp_ipi_intrcnt(void *dummy)
 		intrcnt_add(buf, &ipi_lazypmap_counts[i]);
 		snprintf(buf, sizeof(buf), "cpu%d:hardclock", i);
 		intrcnt_add(buf, &ipi_hardclock_counts[i]);
-		snprintf(buf, sizeof(buf), "cpu%d:statclock", i);
-		intrcnt_add(buf, &ipi_statclock_counts[i]);
 	}
 }
 SYSINIT(mp_ipi_intrcnt, SI_SUB_INTR, SI_ORDER_MIDDLE, mp_ipi_intrcnt, NULL);

Modified: head/sys/amd64/include/apicvar.h
==============================================================================
--- head/sys/amd64/include/apicvar.h	Mon Sep 13 07:22:14 2010	(r212540)
+++ head/sys/amd64/include/apicvar.h	Mon Sep 13 07:25:35 2010	(r212541)
@@ -123,8 +123,7 @@
 #define	IPI_AST		0 	/* Generate software trap. */
 #define IPI_PREEMPT     1
 #define IPI_HARDCLOCK   2
-#define IPI_STATCLOCK   3
-#define IPI_BITMAP_LAST IPI_STATCLOCK
+#define IPI_BITMAP_LAST IPI_HARDCLOCK
 #define IPI_IS_BITMAPED(x) ((x) <= IPI_BITMAP_LAST)
 
 #define	IPI_STOP	(APIC_IPI_INTS + 7)	/* Stop CPU until restarted. */

Modified: head/sys/dev/acpica/acpi_cpu.c
==============================================================================
--- head/sys/dev/acpica/acpi_cpu.c	Mon Sep 13 07:22:14 2010	(r212540)
+++ head/sys/dev/acpica/acpi_cpu.c	Mon Sep 13 07:25:35 2010	(r212541)
@@ -900,7 +900,13 @@ acpi_cpu_idle()
 
     /* Find the lowest state that has small enough latency. */
     cx_next_idx = 0;
-    for (i = sc->cpu_cx_lowest; i >= 0; i--) {
+#ifndef __ia64__
+    if (cpu_disable_deep_sleep)
+	i = sc->cpu_non_c3;
+    else
+#endif
+	i = sc->cpu_cx_lowest;
+    for (; i >= 0; i--) {
 	if (sc->cpu_cx_states[i].trans_lat * 3 <= sc->cpu_prev_sleep) {
 	    cx_next_idx = i;
 	    break;
@@ -929,15 +935,17 @@ acpi_cpu_idle()
     /*
      * Execute HLT (or equivalent) and wait for an interrupt.  We can't
      * precisely calculate the time spent in C1 since the place we wake up
-     * is an ISR.  Assume we slept no more then half of quantum.
+     * is an ISR.  Assume we slept no more then half of quantum, unless
+     * we are called inside critical section, delaying context switch.
      */
     if (cx_next->type == ACPI_STATE_C1) {
 	AcpiHwRead(&start_time, &AcpiGbl_FADT.XPmTimerBlock);
 	acpi_cpu_c1();
 	AcpiHwRead(&end_time, &AcpiGbl_FADT.XPmTimerBlock);
-        end_time = acpi_TimerDelta(end_time, start_time);
-	sc->cpu_prev_sleep = (sc->cpu_prev_sleep * 3 +
-	    min(PM_USEC(end_time), 500000 / hz)) / 4;
+        end_time = PM_USEC(acpi_TimerDelta(end_time, start_time));
+        if (curthread->td_critnest == 0)
+		end_time = min(end_time, 500000 / hz);
+	sc->cpu_prev_sleep = (sc->cpu_prev_sleep * 3 + end_time) / 4;
 	return;
     }
 

Modified: head/sys/dev/acpica/acpi_hpet.c
==============================================================================
--- head/sys/dev/acpica/acpi_hpet.c	Mon Sep 13 07:22:14 2010	(r212540)
+++ head/sys/dev/acpica/acpi_hpet.c	Mon Sep 13 07:25:35 2010	(r212541)
@@ -683,15 +683,15 @@ hpet_detach(device_t dev)
 static int
 hpet_suspend(device_t dev)
 {
-	struct hpet_softc *sc;
+//	struct hpet_softc *sc;
 
 	/*
 	 * Disable the timer during suspend.  The timer will not lose
 	 * its state in S1 or S2, but we are required to disable
 	 * it.
 	 */
-	sc = device_get_softc(dev);
-	hpet_disable(sc);
+//	sc = device_get_softc(dev);
+//	hpet_disable(sc);
 
 	return (0);
 }

Modified: head/sys/i386/i386/machdep.c
==============================================================================
--- head/sys/i386/i386/machdep.c	Mon Sep 13 07:22:14 2010	(r212540)
+++ head/sys/i386/i386/machdep.c	Mon Sep 13 07:25:35 2010	(r212541)
@@ -1175,9 +1175,6 @@ cpu_est_clockrate(int cpu_id, uint64_t *
 	return (0);
 }
 
-
-void (*cpu_idle_hook)(void) = NULL;	/* ACPI idle hook. */
-
 #ifdef XEN
 
 void
@@ -1208,60 +1205,94 @@ cpu_halt(void)
 		__asm__ ("hlt");
 }
 
+#endif
+
+void (*cpu_idle_hook)(void) = NULL;	/* ACPI idle hook. */
+static int	cpu_ident_amdc1e = 0;	/* AMD C1E supported. */
+static int	idle_mwait = 1;		/* Use MONITOR/MWAIT for short idle. */
+TUNABLE_INT("machdep.idle_mwait", &idle_mwait);
+SYSCTL_INT(_machdep, OID_AUTO, idle_mwait, CTLFLAG_RW, &idle_mwait,
+    0, "Use MONITOR/MWAIT for short idle");
+
+#define	STATE_RUNNING	0x0
+#define	STATE_MWAIT	0x1
+#define	STATE_SLEEPING	0x2
+
 static void
-cpu_idle_hlt(int busy)
+cpu_idle_acpi(int busy)
 {
-	/*
-	 * we must absolutely guarentee that hlt is the next instruction
-	 * after sti or we introduce a timing window.
-	 */
+	int *state;
+
+	state = (int *)PCPU_PTR(monitorbuf);
+	*state = STATE_SLEEPING;
 	disable_intr();
-  	if (sched_runnable())
+	if (sched_runnable())
 		enable_intr();
+	else if (cpu_idle_hook)
+		cpu_idle_hook();
 	else
 		__asm __volatile("sti; hlt");
+	*state = STATE_RUNNING;
 }
-#endif
 
+#ifndef XEN
 static void
-cpu_idle_acpi(int busy)
+cpu_idle_hlt(int busy)
 {
+	int *state;
+
+	state = (int *)PCPU_PTR(monitorbuf);
+	*state = STATE_SLEEPING;
+	/*
+	 * We must absolutely guarentee that hlt is the next instruction
+	 * after sti or we introduce a timing window.
+	 */
 	disable_intr();
-  	if (sched_runnable())
+	if (sched_runnable())
 		enable_intr();
-	else if (cpu_idle_hook)
-		cpu_idle_hook();
 	else
 		__asm __volatile("sti; hlt");
+	*state = STATE_RUNNING;
 }
+#endif
 
-static int cpu_ident_amdc1e = 0;
+/*
+ * MWAIT cpu power states.  Lower 4 bits are sub-states.
+ */
+#define	MWAIT_C0	0xf0
+#define	MWAIT_C1	0x00
+#define	MWAIT_C2	0x10
+#define	MWAIT_C3	0x20
+#define	MWAIT_C4	0x30
 
-static int
-cpu_probe_amdc1e(void)
-{ 
-#ifdef DEV_APIC
-	int i;
+static void
+cpu_idle_mwait(int busy)
+{
+	int *state;
 
-	/*
-	 * Forget it, if we're not using local APIC timer.
-	 */
-	if (resource_disabled("apic", 0) ||
-	    (resource_int_value("apic", 0, "clock", &i) == 0 && i == 0))
-		return (0);
+	state = (int *)PCPU_PTR(monitorbuf);
+	*state = STATE_MWAIT;
+	if (!sched_runnable()) {
+		cpu_monitor(state, 0, 0);
+		if (*state == STATE_MWAIT)
+			cpu_mwait(0, MWAIT_C1);
+	}
+	*state = STATE_RUNNING;
+}
 
-	/*
-	 * Detect the presence of C1E capability mostly on latest
-	 * dual-cores (or future) k8 family.
-	 */
-	if (cpu_vendor_id == CPU_VENDOR_AMD &&
-	    (cpu_id & 0x00000f00) == 0x00000f00 &&
-	    (cpu_id & 0x0fff0000) >=  0x00040000) {
-		cpu_ident_amdc1e = 1;
-		return (1);
+static void
+cpu_idle_spin(int busy)
+{
+	int *state;
+	int i;
+
+	state = (int *)PCPU_PTR(monitorbuf);
+	*state = STATE_RUNNING;
+	for (i = 0; i < 1000; i++) {
+		if (sched_runnable())
+			return;
+		cpu_spinwait();
 	}
-#endif
-	return (0);
 }
 
 /*
@@ -1279,32 +1310,20 @@ cpu_probe_amdc1e(void)
 #define	AMDK8_CMPHALT		(AMDK8_SMIONCMPHALT | AMDK8_C1EONCMPHALT)
 
 static void
-cpu_idle_amdc1e(int busy)
+cpu_probe_amdc1e(void)
 {
 
-	disable_intr();
-	if (sched_runnable())
-		enable_intr();
-	else {
-		uint64_t msr;
-
-		msr = rdmsr(MSR_AMDK8_IPM);
-		if (msr & AMDK8_CMPHALT)
-			wrmsr(MSR_AMDK8_IPM, msr & ~AMDK8_CMPHALT);
-
-		if (cpu_idle_hook)
-			cpu_idle_hook();
-		else
-			__asm __volatile("sti; hlt");
+	/*
+	 * Detect the presence of C1E capability mostly on latest
+	 * dual-cores (or future) k8 family.
+	 */
+	if (cpu_vendor_id == CPU_VENDOR_AMD &&
+	    (cpu_id & 0x00000f00) == 0x00000f00 &&
+	    (cpu_id & 0x0fff0000) >=  0x00040000) {
+		cpu_ident_amdc1e = 1;
 	}
 }
 
-static void
-cpu_idle_spin(int busy)
-{
-	return;
-}
-
 #ifdef XEN
 void (*cpu_idle_fn)(int) = cpu_idle_hlt;
 #else
@@ -1314,79 +1333,72 @@ void (*cpu_idle_fn)(int) = cpu_idle_acpi
 void
 cpu_idle(int busy)
 {
+	uint64_t msr;
+
+	CTR2(KTR_SPARE2, "cpu_idle(%d) at %d",
+	    busy, curcpu);
 #if defined(SMP) && !defined(XEN)
 	if (mp_grab_cpu_hlt())
 		return;
 #endif
-	cpu_idle_fn(busy);
-}
-
-/*
- * mwait cpu power states.  Lower 4 bits are sub-states.
- */
-#define	MWAIT_C0	0xf0
-#define	MWAIT_C1	0x00
-#define	MWAIT_C2	0x10
-#define	MWAIT_C3	0x20
-#define	MWAIT_C4	0x30
-
-#define	MWAIT_DISABLED	0x0
-#define	MWAIT_WOKEN	0x1
-#define	MWAIT_WAITING	0x2
+	/* If we are busy - try to use fast methods. */
+	if (busy) {
+		if ((cpu_feature2 & CPUID2_MON) && idle_mwait) {
+			cpu_idle_mwait(busy);
+			goto out;
+		}
+	}
 
-static void
-cpu_idle_mwait(int busy)
-{
-	int *mwait;
+#ifndef XEN
+	/* If we have time - switch timers into idle mode. */
+	if (!busy) {
+		critical_enter();
+		cpu_idleclock();
+	}
+#endif
 
-	mwait = (int *)PCPU_PTR(monitorbuf);
-	*mwait = MWAIT_WAITING;
-	if (sched_runnable())
-		return;
-	cpu_monitor(mwait, 0, 0);
-	if (*mwait == MWAIT_WAITING)
-		cpu_mwait(0, MWAIT_C1);
-}
+	/* Apply AMD APIC timer C1E workaround. */
+	if (cpu_ident_amdc1e
+#ifndef XEN
+	    && cpu_disable_deep_sleep
+#endif
+	    ) {
+		msr = rdmsr(MSR_AMDK8_IPM);
+		if (msr & AMDK8_CMPHALT)
+			wrmsr(MSR_AMDK8_IPM, msr & ~AMDK8_CMPHALT);
+	}
 
-static void
-cpu_idle_mwait_hlt(int busy)
-{
-	int *mwait;
+	/* Call main idle method. */
+	cpu_idle_fn(busy);
 
-	mwait = (int *)PCPU_PTR(monitorbuf);
-	if (busy == 0) {
-		*mwait = MWAIT_DISABLED;
-		cpu_idle_hlt(busy);
-		return;
+#ifndef XEN
+	/* Switch timers mack into active mode. */
+	if (!busy) {
+		cpu_activeclock();
+		critical_exit();
 	}
-	*mwait = MWAIT_WAITING;
-	if (sched_runnable())
-		return;
-	cpu_monitor(mwait, 0, 0);
-	if (*mwait == MWAIT_WAITING)
-		cpu_mwait(0, MWAIT_C1);
+#endif
+out:
+	CTR2(KTR_SPARE2, "cpu_idle(%d) at %d done",
+	    busy, curcpu);
 }
 
 int
 cpu_idle_wakeup(int cpu)
 {
 	struct pcpu *pcpu;
-	int *mwait;
+	int *state;
 
-	if (cpu_idle_fn == cpu_idle_spin)
-		return (1);
-	if (cpu_idle_fn != cpu_idle_mwait && cpu_idle_fn != cpu_idle_mwait_hlt)
-		return (0);
 	pcpu = pcpu_find(cpu);
-	mwait = (int *)pcpu->pc_monitorbuf;
+	state = (int *)pcpu->pc_monitorbuf;
 	/*
 	 * This doesn't need to be atomic since missing the race will
 	 * simply result in unnecessary IPIs.
 	 */
-	if (cpu_idle_fn == cpu_idle_mwait_hlt && *mwait == MWAIT_DISABLED)
+	if (*state == STATE_SLEEPING)
 		return (0);
-	*mwait = MWAIT_WOKEN;
-
+	if (*state == STATE_MWAIT)
+		*state = STATE_RUNNING;
 	return (1);
 }
 
@@ -1399,8 +1411,6 @@ struct {
 } idle_tbl[] = {
 	{ cpu_idle_spin, "spin" },
 	{ cpu_idle_mwait, "mwait" },
-	{ cpu_idle_mwait_hlt, "mwait_hlt" },
-	{ cpu_idle_amdc1e, "amdc1e" },
 	{ cpu_idle_hlt, "hlt" },
 	{ cpu_idle_acpi, "acpi" },
 	{ NULL, NULL }
@@ -1419,8 +1429,8 @@ idle_sysctl_available(SYSCTL_HANDLER_ARG
 		if (strstr(idle_tbl[i].id_name, "mwait") &&
 		    (cpu_feature2 & CPUID2_MON) == 0)
 			continue;
-		if (strcmp(idle_tbl[i].id_name, "amdc1e") == 0 &&
-		    cpu_ident_amdc1e == 0)
+		if (strcmp(idle_tbl[i].id_name, "acpi") == 0 &&
+		    cpu_idle_hook == NULL)
 			continue;
 		p += sprintf(p, "%s, ", idle_tbl[i].id_name);
 	}
@@ -1429,6 +1439,9 @@ idle_sysctl_available(SYSCTL_HANDLER_ARG
 	return (error);
 }
 
+SYSCTL_PROC(_machdep, OID_AUTO, idle_available, CTLTYPE_STRING | CTLFLAG_RD,
+    0, 0, idle_sysctl_available, "A", "list of available idle functions");
+
 static int
 idle_sysctl(SYSCTL_HANDLER_ARGS)
 {
@@ -1452,8 +1465,8 @@ idle_sysctl(SYSCTL_HANDLER_ARGS)
 		if (strstr(idle_tbl[i].id_name, "mwait") &&
 		    (cpu_feature2 & CPUID2_MON) == 0)
 			continue;
-		if (strcmp(idle_tbl[i].id_name, "amdc1e") == 0 &&
-		    cpu_ident_amdc1e == 0)
+		if (strcmp(idle_tbl[i].id_name, "acpi") == 0 &&
+		    cpu_idle_hook == NULL)
 			continue;
 		if (strcmp(idle_tbl[i].id_name, buf))
 			continue;
@@ -1463,9 +1476,6 @@ idle_sysctl(SYSCTL_HANDLER_ARGS)
 	return (EINVAL);
 }
 
-SYSCTL_PROC(_machdep, OID_AUTO, idle_available, CTLTYPE_STRING | CTLFLAG_RD,
-    0, 0, idle_sysctl_available, "A", "list of available idle functions");
-
 SYSCTL_PROC(_machdep, OID_AUTO, idle, CTLTYPE_STRING | CTLFLAG_RW, 0, 0,
     idle_sysctl, "A", "currently selected idle function");
 
@@ -2695,8 +2705,7 @@ init386(first)
 	thread0.td_pcb->pcb_fsd = PCPU_GET(fsgs_gdt)[0];
 	thread0.td_pcb->pcb_gsd = PCPU_GET(fsgs_gdt)[1];
 
-	if (cpu_probe_amdc1e())
-		cpu_idle_fn = cpu_idle_amdc1e;
+	cpu_probe_amdc1e();
 }
 
 #else
@@ -2970,8 +2979,7 @@ init386(first)
 	thread0.td_pcb->pcb_ext = 0;
 	thread0.td_frame = &proc0_tf;
 
-	if (cpu_probe_amdc1e())
-		cpu_idle_fn = cpu_idle_amdc1e;
+	cpu_probe_amdc1e();
 }
 #endif
 

Modified: head/sys/i386/i386/mp_machdep.c
==============================================================================
--- head/sys/i386/i386/mp_machdep.c	Mon Sep 13 07:22:14 2010	(r212540)
+++ head/sys/i386/i386/mp_machdep.c	Mon Sep 13 07:25:35 2010	(r212541)
@@ -167,7 +167,6 @@ u_long *ipi_invlcache_counts[MAXCPU];
 u_long *ipi_rendezvous_counts[MAXCPU];
 u_long *ipi_lazypmap_counts[MAXCPU];
 static u_long *ipi_hardclock_counts[MAXCPU];
-static u_long *ipi_statclock_counts[MAXCPU];
 #endif
 
 /*
@@ -1284,16 +1283,22 @@ smp_masked_invlpg_range(cpumask_t mask, 
 void
 ipi_bitmap_handler(struct trapframe frame)
 {
+	struct trapframe *oldframe;
+	struct thread *td;
 	int cpu = PCPU_GET(cpuid);
 	u_int ipi_bitmap;
 
+	critical_enter();
+	td = curthread;
+	td->td_intr_nesting_level++;
+	oldframe = td->td_intr_frame;
+	td->td_intr_frame = &frame;
 	ipi_bitmap = atomic_readandclear_int(&cpu_ipi_pending[cpu]);
-
 	if (ipi_bitmap & (1 << IPI_PREEMPT)) {
 #ifdef COUNT_IPIS
 		(*ipi_preempt_counts[cpu])++;
 #endif
-		sched_preempt(curthread);
+		sched_preempt(td);
 	}
 	if (ipi_bitmap & (1 << IPI_AST)) {
 #ifdef COUNT_IPIS
@@ -1305,14 +1310,11 @@ ipi_bitmap_handler(struct trapframe fram
 #ifdef COUNT_IPIS
 		(*ipi_hardclock_counts[cpu])++;
 #endif
-		hardclockintr(&frame); 
-	}
-	if (ipi_bitmap & (1 << IPI_STATCLOCK)) {
-#ifdef COUNT_IPIS
-		(*ipi_statclock_counts[cpu])++;
-#endif
-		statclockintr(&frame); 
+		hardclockintr();
 	}
+	td->td_intr_frame = oldframe;
+	td->td_intr_nesting_level--;
+	critical_exit();
 }
 
 /*
@@ -1627,8 +1629,6 @@ mp_ipi_intrcnt(void *dummy)
 		intrcnt_add(buf, &ipi_lazypmap_counts[i]);
 		snprintf(buf, sizeof(buf), "cpu%d:hardclock", i);
 		intrcnt_add(buf, &ipi_hardclock_counts[i]);
-		snprintf(buf, sizeof(buf), "cpu%d:statclock", i);
-		intrcnt_add(buf, &ipi_statclock_counts[i]);
 	}		
 }
 SYSINIT(mp_ipi_intrcnt, SI_SUB_INTR, SI_ORDER_MIDDLE, mp_ipi_intrcnt, NULL);

Modified: head/sys/i386/include/apicvar.h
==============================================================================
--- head/sys/i386/include/apicvar.h	Mon Sep 13 07:22:14 2010	(r212540)
+++ head/sys/i386/include/apicvar.h	Mon Sep 13 07:25:35 2010	(r212541)
@@ -124,8 +124,7 @@
 #define	IPI_AST		0 	/* Generate software trap. */
 #define IPI_PREEMPT     1
 #define IPI_HARDCLOCK   2 
-#define IPI_STATCLOCK   3 
-#define IPI_BITMAP_LAST IPI_STATCLOCK
+#define IPI_BITMAP_LAST IPI_HARDCLOCK
 #define IPI_IS_BITMAPED(x) ((x) <= IPI_BITMAP_LAST)
 
 #define	IPI_STOP	(APIC_IPI_INTS + 7)	/* Stop CPU until restarted. */
@@ -152,8 +151,7 @@
 #define	IPI_AST		0 	/* Generate software trap. */
 #define IPI_PREEMPT     1
 #define IPI_HARDCLOCK   2 
-#define IPI_STATCLOCK   3 
-#define IPI_BITMAP_LAST IPI_STATCLOCK
+#define IPI_BITMAP_LAST IPI_HARDCLOCK
 #define IPI_IS_BITMAPED(x) ((x) <= IPI_BITMAP_LAST)
 
 #define	IPI_STOP	(APIC_IPI_INTS + 7)	/* Stop CPU until restarted. */

Modified: head/sys/kern/kern_clock.c
==============================================================================
--- head/sys/kern/kern_clock.c	Mon Sep 13 07:22:14 2010	(r212540)
+++ head/sys/kern/kern_clock.c	Mon Sep 13 07:25:35 2010	(r212541)
@@ -373,11 +373,9 @@ int	profprocs;
 int	ticks;
 int	psratio;
 
-int	timer1hz;
-int	timer2hz;
-static DPCPU_DEFINE(u_int, hard_cnt);
-static DPCPU_DEFINE(u_int, stat_cnt);
-static DPCPU_DEFINE(u_int, prof_cnt);
+static DPCPU_DEFINE(int, pcputicks);	/* Per-CPU version of ticks. */
+static struct mtx	global_hardclock_mtx;
+MTX_SYSINIT(global_hardclock_mtx, &global_hardclock_mtx, "ghc_mtx", MTX_SPIN);
 
 /*
  * Initialize clock frequencies and start both clocks running.
@@ -408,52 +406,6 @@ initclocks(dummy)
 #endif
 }
 
-void
-timer1clock(int usermode, uintfptr_t pc)
-{
-	u_int *cnt;
-
-	cnt = DPCPU_PTR(hard_cnt);
-	*cnt += hz;
-	if (*cnt >= timer1hz) {
-		*cnt -= timer1hz;
-		if (*cnt >= timer1hz)
-			*cnt = 0;
-		if (PCPU_GET(cpuid) == 0)
-			hardclock(usermode, pc);
-		else
-			hardclock_cpu(usermode);
-	}
-	if (timer2hz == 0)
-		timer2clock(usermode, pc);
-}
-
-void
-timer2clock(int usermode, uintfptr_t pc)
-{
-	u_int *cnt;
-	int t2hz = timer2hz ? timer2hz : timer1hz;
-
-	cnt = DPCPU_PTR(stat_cnt);
-	*cnt += stathz;
-	if (*cnt >= t2hz) {
-		*cnt -= t2hz;
-		if (*cnt >= t2hz)
-			*cnt = 0;
-		statclock(usermode);
-	}
-	if (profprocs == 0)
-		return;
-	cnt = DPCPU_PTR(prof_cnt);
-	*cnt += profhz;
-	if (*cnt >= t2hz) {
-		*cnt -= t2hz;

*** DIFF OUTPUT TRUNCATED AT 1000 LINES ***


More information about the svn-src-all mailing list