From: Hasso Tepper <hasso@dragonflybsd.org>
Date: Fri, 5 Sep 2008 10:28:35 +0000 (+0000)
Subject: acpi_cpu(4) update. It's now possible to use higher (lower power usage) C
X-Git-Tag: v2.1.1~492
X-Git-Url: https://gitweb.dragonflybsd.org/dragonfly.git/commitdiff_plain/5ebadb2c919e68287bf639bd3ad330c74a3247b3

acpi_cpu(4) update. It's now possible to use higher (lower power usage) C
states than C1 in modern (multicore) CPUs.

Obtained-from: FreeBSD
---

diff --git a/sys/dev/acpica5/acpi.c b/sys/dev/acpica5/acpi.c
index 5bec581090..92893c17f2 100644
--- a/sys/dev/acpica5/acpi.c
+++ b/sys/dev/acpica5/acpi.c
@@ -27,7 +27,7 @@
  * SUCH DAMAGE.
  *
  *	$FreeBSD: src/sys/dev/acpica/acpi.c,v 1.160 2004/06/14 03:52:19 njl Exp $
- *	$DragonFly: src/sys/dev/acpica5/acpi.c,v 1.35 2008/08/27 16:35:19 hasso Exp $
+ *	$DragonFly: src/sys/dev/acpica5/acpi.c,v 1.36 2008/09/05 10:28:35 hasso Exp $
  */
 
 #include "opt_acpi.h"
@@ -1010,7 +1010,8 @@ acpi_release_resource(device_t bus, device_t child, int type, int rid,
 
 /* Allocate an IO port or memory resource, given its GAS. */
 struct resource *
-acpi_bus_alloc_gas(device_t dev, int *rid, ACPI_GENERIC_ADDRESS *gas)
+acpi_bus_alloc_gas(device_t dev, int *rid, ACPI_GENERIC_ADDRESS *gas,
+    u_int flags)
 {
     int type;
 
@@ -1029,7 +1030,7 @@ acpi_bus_alloc_gas(device_t dev, int *rid, ACPI_GENERIC_ADDRESS *gas)
     }
 
     bus_set_resource(dev, type, *rid, gas->Address, gas->BitWidth / 8);
-    return (bus_alloc_resource_any(dev, type, rid, RF_ACTIVE));
+    return (bus_alloc_resource_any(dev, type, rid, RF_ACTIVE | flags));
 }
 
 /*
diff --git a/sys/dev/acpica5/acpi_cpu.c b/sys/dev/acpica5/acpi_cpu.c
index 8290dc9432..992f7263f5 100644
--- a/sys/dev/acpica5/acpi_cpu.c
+++ b/sys/dev/acpica5/acpi_cpu.c
@@ -1,5 +1,5 @@
 /*-
- * Copyright (c) 2003 Nate Lawson (SDG)
+ * Copyright (c) 2003-2005 Nate Lawson (SDG)
  * Copyright (c) 2001 Michael Smith
  * All rights reserved.
  *
@@ -24,8 +24,8 @@
  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  * SUCH DAMAGE.
  *
- * $FreeBSD: src/sys/dev/acpica/acpi_cpu.c,v 1.41 2004/06/24 00:38:51 njl Exp $
- * $DragonFly: src/sys/dev/acpica5/acpi_cpu.c,v 1.20 2007/03/21 22:19:00 y0netan1 Exp $
+ * $FreeBSD: src/sys/dev/acpica/acpi_cpu.c,v 1.72 2008/04/12 12:06:00 rpaulo Exp $
+ * $DragonFly: src/sys/dev/acpica5/acpi_cpu.c,v 1.21 2008/09/05 10:28:35 hasso Exp $
  */
 
 #include "opt_acpi.h"
@@ -50,11 +50,7 @@
 #include "acpivar.h"
 
 /*
- * Support for ACPI Processor devices, including ACPI 2.0 throttling
- * and C[1-3] sleep states.
- *
- * TODO: implement scans of all CPUs to be sure all Cx states are
- * equivalent.
+ * Support for ACPI Processor devices, including C[1-3] sleep states.
  */
 
 /* Hooks for the ACPI CA debugging infrastructure */
@@ -66,19 +62,35 @@ struct acpi_cx {
     uint32_t		 type;		/* C1-3 (C4 and up treated as C3). */
     uint32_t		 trans_lat;	/* Transition latency (usec). */
     uint32_t		 power;		/* Power consumed (mW). */
+    int			 res_type;	/* Resource type for p_lvlx. */
 };
 #define MAX_CX_STATES	 8
 
 struct acpi_cpu_softc {
     device_t		 cpu_dev;
     ACPI_HANDLE		 cpu_handle;
-    uint32_t		 acpi_id;	/* ACPI processor id */
+    struct mdglobaldata *md;
+    uint32_t		 cpu_acpi_id;	/* ACPI processor id */
     uint32_t		 cpu_p_blk;	/* ACPI P_BLK location */
     uint32_t		 cpu_p_blk_len;	/* P_BLK length (must be 6). */
-    struct resource	*cpu_p_cnt;	/* Throttling control register */
     struct acpi_cx	 cpu_cx_states[MAX_CX_STATES];
     int			 cpu_cx_count;	/* Number of valid Cx states. */
     int			 cpu_prev_sleep;/* Last idle sleep duration. */
+    int			 cpu_features;	/* Child driver supported features. */
+    /* Runtime state. */
+    int			 cpu_non_c3;	/* Index of lowest non-C3 state. */
+    int			 cpu_short_slp;	/* Count of < 1us sleeps. */
+    u_int		 cpu_cx_stats[MAX_CX_STATES];/* Cx usage history. */
+    /* Values for sysctl. */
+    struct sysctl_ctx_list cpu_sysctl_ctx;
+    struct sysctl_oid	*cpu_sysctl_tree;
+    int			 cpu_cx_lowest;
+    char 		 cpu_cx_supported[64];
+    int			 cpu_rid;
+};
+
+struct acpi_cpu_device {
+    struct resource_list	ad_rl;
 };
 
 #define CPU_GET_REG(reg, width) 					\
@@ -88,24 +100,11 @@ struct acpi_cpu_softc {
     (bus_space_write_ ## width(rman_get_bustag((reg)), 			\
 		       rman_get_bushandle((reg)), 0, (val)))
 
-/*
- * Speeds are stored in counts, from 1 to CPU_MAX_SPEED, and
- * reported to the user in tenths of a percent.
- */
-static uint32_t		 cpu_duty_offset;
-static uint32_t		 cpu_duty_width;
-#define CPU_MAX_SPEED		(1 << cpu_duty_width)
-#define CPU_SPEED_PERCENT(x)	((1000 * (x)) / CPU_MAX_SPEED)
-#define CPU_SPEED_PRINTABLE(x)	(CPU_SPEED_PERCENT(x) / 10),	\
-				(CPU_SPEED_PERCENT(x) % 10)
-#define CPU_P_CNT_THT_EN (1<<4)
 #define PM_USEC(x)	 ((x) >> 2)	/* ~4 clocks per usec (3.57955 Mhz) */
 
-#define ACPI_CPU_NOTIFY_PERF_STATES	0x80	/* _PSS changed. */
-#define ACPI_CPU_NOTIFY_CX_STATES	0x81	/* _CST changed. */
+#define ACPI_NOTIFY_CX_STATES	0x81	/* _CST changed. */
 
 #define CPU_QUIRK_NO_C3		(1<<0)	/* C3-type states are not usable. */
-#define CPU_QUIRK_NO_THROTTLE	(1<<1)	/* Throttling is not usable. */
 #define CPU_QUIRK_NO_BM_CTRL	(1<<2)	/* No bus mastering control. */
 
 #define PCI_VENDOR_INTEL	0x8086
@@ -114,60 +113,82 @@ static uint32_t		 cpu_duty_width;
 #define PCI_REVISION_B_STEP	1
 #define PCI_REVISION_4E		2
 #define PCI_REVISION_4M		3
+#define PIIX4_DEVACTB_REG	0x58
+#define PIIX4_BRLD_EN_IRQ0	(1<<0)
+#define PIIX4_BRLD_EN_IRQ	(1<<1)
+#define PIIX4_BRLD_EN_IRQ8	(1<<5)
+#define PIIX4_STOP_BREAK_MASK	(PIIX4_BRLD_EN_IRQ0 | PIIX4_BRLD_EN_IRQ | PIIX4_BRLD_EN_IRQ8)
+#define PIIX4_PCNTRL_BST_EN	(1<<10)
 
 /* Platform hardware resource information. */
 static uint32_t		 cpu_smi_cmd;	/* Value to write to SMI_CMD. */
 static uint8_t		 cpu_cst_cnt;	/* Indicate we are _CST aware. */
-static int		 cpu_rid;	/* Driver-wide resource id. */
 static int		 cpu_quirks;	/* Indicate any hardware bugs. */
 
 /* Runtime state. */
-static int		 cpu_cx_count;	/* Number of valid states */
-static int		 cpu_non_c3;	/* Index of lowest non-C3 state. */
-static u_int		 cpu_cx_stats[MAX_CX_STATES];/* Cx usage history. */
+static int		 cpu_disable_idle; /* Disable entry to idle function */
+static int		 cpu_cx_count;	/* Number of valid Cx states */
 
 /* Values for sysctl. */
-static uint32_t		 cpu_throttle_state;
-static uint32_t		 cpu_throttle_max;
-static uint32_t		 cpu_throttle_performance;
-static uint32_t		 cpu_throttle_economy;
+static struct sysctl_ctx_list cpu_sysctl_ctx;
+static struct sysctl_oid *cpu_sysctl_tree;
+static int		 cpu_cx_generic;
 static int		 cpu_cx_lowest;
-static char 		 cpu_cx_supported[64];
 
 static device_t		*cpu_devices;
 static int		 cpu_ndevices;
 static struct acpi_cpu_softc **cpu_softc;
 
-static struct sysctl_ctx_list	acpi_cpu_sysctl_ctx;
-static struct sysctl_oid	*acpi_cpu_sysctl_tree;
-
 static int	acpi_cpu_probe(device_t dev);
 static int	acpi_cpu_attach(device_t dev);
+static int	acpi_cpu_suspend(device_t dev);
+static int	acpi_cpu_resume(device_t dev);
 static int	acpi_pcpu_get_id(uint32_t idx, uint32_t *acpi_id,
-				 uint32_t *cpu_id);
+		    uint32_t *cpu_id);
+static struct resource_list *acpi_cpu_get_rlist(device_t dev, device_t child);
+static device_t	acpi_cpu_add_child(device_t dev, int order, const char *name,
+		    int unit);
+static int	acpi_cpu_read_ivar(device_t dev, device_t child, int index,
+		    uintptr_t *result);
 static int	acpi_cpu_shutdown(device_t dev);
-static int	acpi_cpu_throttle_probe(struct acpi_cpu_softc *sc);
-static void	acpi_cpu_power_profile(void *arg);
-static int	acpi_cpu_cx_probe(struct acpi_cpu_softc *sc);
+static void	acpi_cpu_cx_probe(struct acpi_cpu_softc *sc);
+static void	acpi_cpu_generic_cx_probe(struct acpi_cpu_softc *sc);
 static int	acpi_cpu_cx_cst(struct acpi_cpu_softc *sc);
 static void	acpi_cpu_startup(void *arg);
-static void	acpi_cpu_startup_throttling(void);
-static void	acpi_cpu_startup_cx(void);
-static void	acpi_cpu_throttle_set(uint32_t speed);
+static void	acpi_cpu_startup_cx(struct acpi_cpu_softc *sc);
+static void	acpi_cpu_cx_list(struct acpi_cpu_softc *sc);
 static void	acpi_cpu_idle(void);
-static void	acpi_cpu_c1(void);
 static void	acpi_cpu_notify(ACPI_HANDLE h, UINT32 notify, void *context);
-static int	acpi_cpu_quirks(struct acpi_cpu_softc *sc);
-static int	acpi_cpu_throttle_sysctl(SYSCTL_HANDLER_ARGS);
+static int	acpi_cpu_quirks(void);
 static int	acpi_cpu_usage_sysctl(SYSCTL_HANDLER_ARGS);
+static int	acpi_cpu_set_cx_lowest(struct acpi_cpu_softc *sc, int val);
 static int	acpi_cpu_cx_lowest_sysctl(SYSCTL_HANDLER_ARGS);
+static int	acpi_cpu_global_cx_lowest_sysctl(SYSCTL_HANDLER_ARGS);
+
+static void	acpi_cpu_c1(void);	/* XXX */
 
 static device_method_t acpi_cpu_methods[] = {
     /* Device interface */
     DEVMETHOD(device_probe,	acpi_cpu_probe),
     DEVMETHOD(device_attach,	acpi_cpu_attach),
+    DEVMETHOD(device_detach,	bus_generic_detach),
     DEVMETHOD(device_shutdown,	acpi_cpu_shutdown),
-
+    DEVMETHOD(device_suspend,	acpi_cpu_suspend),
+    DEVMETHOD(device_resume,	acpi_cpu_resume),
+
+    /* Bus interface */
+    DEVMETHOD(bus_add_child,	acpi_cpu_add_child),
+    DEVMETHOD(bus_read_ivar,	acpi_cpu_read_ivar),
+    DEVMETHOD(bus_get_resource_list, acpi_cpu_get_rlist),
+    DEVMETHOD(bus_get_resource,	bus_generic_rl_get_resource),
+    DEVMETHOD(bus_set_resource,	bus_generic_rl_set_resource),
+    DEVMETHOD(bus_alloc_resource, bus_generic_rl_alloc_resource),
+    DEVMETHOD(bus_release_resource, bus_generic_rl_release_resource),
+    DEVMETHOD(bus_driver_added,	bus_generic_driver_added),
+    DEVMETHOD(bus_activate_resource, bus_generic_activate_resource),
+    DEVMETHOD(bus_deactivate_resource, bus_generic_deactivate_resource),
+    DEVMETHOD(bus_setup_intr,	bus_generic_setup_intr),
+    DEVMETHOD(bus_teardown_intr, bus_generic_teardown_intr),
     {0, 0}
 };
 
@@ -184,10 +205,9 @@ MODULE_DEPEND(cpu, acpi, 1, 1, 1);
 static int
 acpi_cpu_probe(device_t dev)
 {
-    int			   acpi_id, cpu_id, cx_count;
+    int			   acpi_id, cpu_id;
     ACPI_BUFFER		   buf;
     ACPI_HANDLE		   handle;
-    char		   msg[32];
     ACPI_OBJECT		   *obj;
     ACPI_STATUS		   status;
 
@@ -232,33 +252,10 @@ acpi_cpu_probe(device_t dev)
     if (cpu_softc[cpu_id] != NULL)
 	return (ENXIO);
 
-    /* Get a count of Cx states for our device string. */
-    cx_count = 0;
-    buf.Pointer = NULL;
-    buf.Length = ACPI_ALLOCATE_BUFFER;
-    status = AcpiEvaluateObject(handle, "_CST", NULL, &buf);
-    if (ACPI_SUCCESS(status)) {
-	obj = (ACPI_OBJECT *)buf.Pointer;
-	if (ACPI_PKG_VALID(obj, 2))
-	    acpi_PkgInt32(obj, 0, &cx_count);
-	AcpiOsFree(obj);
-    } else {
-	if (AcpiGbl_FADT.C2Latency <= 100)
-	    cx_count++;
-	if (AcpiGbl_FADT.C3Latency <= 1000)
-	    cx_count++;
-	if (cx_count > 0)
-	    cx_count++;
-    }
-    if (cx_count > 0)
-	ksnprintf(msg, sizeof(msg), "ACPI CPU (%d Cx states)", cx_count);
-    else
-	strlcpy(msg, "ACPI CPU", sizeof(msg));
-    device_set_desc_copy(dev, msg);
-
     /* Mark this processor as in-use and save our derived id for attach. */
     cpu_softc[cpu_id] = (void *)1;
     acpi_set_magic(dev, cpu_id);
+    device_set_desc(dev, "ACPI CPU");
 
     return (0);
 }
@@ -267,20 +264,33 @@ static int
 acpi_cpu_attach(device_t dev)
 {
     ACPI_BUFFER		   buf;
-    ACPI_OBJECT		   *obj;
+    ACPI_OBJECT		   arg[4], *obj;
+    ACPI_OBJECT_LIST	   arglist;
+    struct mdglobaldata	  *md;
     struct acpi_cpu_softc *sc;
     struct acpi_softc	  *acpi_sc;
     ACPI_STATUS		   status;
-    int			   thr_ret, cx_ret;
+    u_int		   features;
+    int			   cpu_id, drv_count, i;
+    driver_t 		  **drivers;
+    uint32_t		   cap_set[3];
 
-    ACPI_FUNCTION_TRACE((char *)(uintptr_t)__func__);
+    /* UUID needed by _OSC evaluation */
+    static uint8_t cpu_oscuuid[16] = { 0x16, 0xA6, 0x77, 0x40, 0x0C, 0x29,
+				       0xBE, 0x47, 0x9E, 0xBD, 0xD8, 0x70,
+				       0x58, 0x71, 0x39, 0x53 };
 
-    ACPI_ASSERTLOCK;
+    ACPI_FUNCTION_TRACE((char *)(uintptr_t)__func__);
 
     sc = device_get_softc(dev);
     sc->cpu_dev = dev;
     sc->cpu_handle = acpi_get_handle(dev);
-    cpu_softc[acpi_get_magic(dev)] = sc;
+    cpu_id = acpi_get_magic(dev);
+    cpu_softc[cpu_id] = sc;
+    md = (struct mdglobaldata *)globaldata_find(device_get_unit(dev));
+    sc->md = md;
+    cpu_smi_cmd = AcpiGbl_FADT.SmiCommand;
+    cpu_cst_cnt = AcpiGbl_FADT.CstControl;
 
     buf.Pointer = NULL;
     buf.Length = ACPI_ALLOCATE_BUFFER;
@@ -293,37 +303,121 @@ acpi_cpu_attach(device_t dev)
     obj = (ACPI_OBJECT *)buf.Pointer;
     sc->cpu_p_blk = obj->Processor.PblkAddress;
     sc->cpu_p_blk_len = obj->Processor.PblkLength;
-    sc->acpi_id = obj->Processor.ProcId;
+    sc->cpu_acpi_id = obj->Processor.ProcId;
     AcpiOsFree(obj);
     ACPI_DEBUG_PRINT((ACPI_DB_INFO, "acpi_cpu%d: P_BLK at %#x/%d\n",
 		     device_get_unit(dev), sc->cpu_p_blk, sc->cpu_p_blk_len));
 
-    acpi_sc = acpi_device_get_parent_softc(dev);
-    sysctl_ctx_init(&acpi_cpu_sysctl_ctx);
-    acpi_cpu_sysctl_tree = SYSCTL_ADD_NODE(&acpi_cpu_sysctl_ctx,
-				SYSCTL_CHILDREN(acpi_sc->acpi_sysctl_tree),
-				OID_AUTO, "cpu", CTLFLAG_RD, 0, "");
+    /*
+     * If this is the first cpu we attach, create and initialize the generic
+     * resources that will be used by all acpi cpu devices.
+     */
+    if (device_get_unit(dev) == 0) {
+	/* Assume we won't be using generic Cx mode by default */
+	cpu_cx_generic = FALSE;
+
+	/* Install hw.acpi.cpu sysctl tree */
+	acpi_sc = acpi_device_get_parent_softc(dev);
+	sysctl_ctx_init(&cpu_sysctl_ctx);
+	cpu_sysctl_tree = SYSCTL_ADD_NODE(&cpu_sysctl_ctx,
+	    SYSCTL_CHILDREN(acpi_sc->acpi_sysctl_tree), OID_AUTO, "cpu",
+	    CTLFLAG_RD, 0, "node for CPU children");
+
+	/* Queue post cpu-probing task handler */
+	AcpiOsExecute(OSL_NOTIFY_HANDLER, acpi_cpu_startup, NULL);
+    }
 
-    /* If this is the first device probed, check for quirks. */
-    if (device_get_unit(dev) == 0)
-	acpi_cpu_quirks(sc);
+    sysctl_ctx_init(&sc->cpu_sysctl_ctx);
+    sc->cpu_sysctl_tree = SYSCTL_ADD_NODE(&sc->cpu_sysctl_ctx,
+				      SYSCTL_STATIC_CHILDREN(_hw), OID_AUTO,
+				      device_get_nameunit(dev), CTLFLAG_RD,
+				      0, "");
 
     /*
-     * Probe for throttling and Cx state support.
-     * If none of these is present, free up unused resources.
+     * Before calling any CPU methods, collect child driver feature hints
+     * and notify ACPI of them.  We support unified SMP power control
+     * so advertise this ourselves.  Note this is not the same as independent
+     * SMP control where each CPU can have different settings.
      */
-    thr_ret = acpi_cpu_throttle_probe(sc);
-    cx_ret = acpi_cpu_cx_probe(sc);
-    if (thr_ret == 0 || cx_ret == 0) {
-	status = AcpiInstallNotifyHandler(sc->cpu_handle, ACPI_DEVICE_NOTIFY,
-					  acpi_cpu_notify, sc);
-	if (device_get_unit(dev) == 0)
-	    AcpiOsExecute(OSL_NOTIFY_HANDLER, acpi_cpu_startup, NULL);
-    } else {
-	sysctl_ctx_free(&acpi_cpu_sysctl_ctx);
+    sc->cpu_features = ACPI_CAP_SMP_SAME | ACPI_CAP_SMP_SAME_C3;
+    if (devclass_get_drivers(acpi_cpu_devclass, &drivers, &drv_count) == 0) {
+	for (i = 0; i < drv_count; i++) {
+	    if (ACPI_GET_FEATURES(drivers[i], &features) == 0)
+		sc->cpu_features |= features;
+	}
+	kfree(drivers, M_TEMP);
     }
 
-    return_VALUE (0);
+    /*
+     * CPU capabilities are specified as a buffer of 32-bit integers:
+     * revision, count, and one or more capabilities.  The revision of
+     * "1" is not specified anywhere but seems to match Linux.
+     */
+    if (sc->cpu_features) {
+	arglist.Pointer = arg;
+	arglist.Count = 1;
+	arg[0].Type = ACPI_TYPE_BUFFER;
+	arg[0].Buffer.Length = sizeof(cap_set);
+	arg[0].Buffer.Pointer = (uint8_t *)cap_set;
+	cap_set[0] = 1; /* revision */
+	cap_set[1] = 1; /* number of capabilities integers */
+	cap_set[2] = sc->cpu_features;
+	AcpiEvaluateObject(sc->cpu_handle, "_PDC", &arglist, NULL);
+
+	/*
+	 * On some systems we need to evaluate _OSC so that the ASL
+	 * loads the _PSS and/or _PDC methods at runtime.
+	 *
+	 * TODO: evaluate failure of _OSC.
+	 */
+	arglist.Pointer = arg;
+	arglist.Count = 4;
+	arg[0].Type = ACPI_TYPE_BUFFER;
+	arg[0].Buffer.Length = sizeof(cpu_oscuuid);
+	arg[0].Buffer.Pointer = cpu_oscuuid;	/* UUID */
+	arg[1].Type = ACPI_TYPE_INTEGER;
+	arg[1].Integer.Value = 1;		/* revision */
+	arg[2].Type = ACPI_TYPE_INTEGER;
+	arg[2].Integer.Value = 1;		/* count */
+	arg[3].Type = ACPI_TYPE_BUFFER;
+	arg[3].Buffer.Length = sizeof(cap_set);	/* Capabilities buffer */
+	arg[3].Buffer.Pointer = (uint8_t *)cap_set;
+	cap_set[0] = 0;
+	AcpiEvaluateObject(sc->cpu_handle, "_OSC", &arglist, NULL);
+    }
+
+    /* Probe for Cx state support. */
+    acpi_cpu_cx_probe(sc);
+
+    /* Finally,  call identify and probe/attach for child devices. */
+    bus_generic_probe(dev);
+    bus_generic_attach(dev);
+
+    return (0);
+}
+
+/*
+ * Disable any entry to the idle function during suspend and re-enable it
+ * during resume.
+ */
+static int
+acpi_cpu_suspend(device_t dev)
+{
+    int error;
+
+    error = bus_generic_suspend(dev);
+    if (error)
+	return (error);
+    cpu_disable_idle = TRUE;
+    return (0);
+}
+
+static int
+acpi_cpu_resume(device_t dev)
+{
+
+    cpu_disable_idle = FALSE;
+    return (bus_generic_resume(dev));
 }
 
 /*
@@ -351,7 +445,7 @@ acpi_pcpu_get_id(uint32_t idx, uint32_t *acpi_id, uint32_t *cpu_id)
 	     * return the pc_cpuid to reference this processor.
 	     */
 	    if (md->gd_acpi_id == 0xffffffff)
-		 md->gd_acpi_id = *acpi_id;
+		md->gd_acpi_id = *acpi_id;
 	    else if (md->gd_acpi_id != *acpi_id)
 		*acpi_id = md->gd_acpi_id;
 	    *cpu_id = md->mi.gd_cpuid;
@@ -362,206 +456,170 @@ acpi_pcpu_get_id(uint32_t idx, uint32_t *acpi_id, uint32_t *cpu_id)
     return (ESRCH);
 }
 
-static int
-acpi_cpu_shutdown(device_t dev)
+static struct resource_list *
+acpi_cpu_get_rlist(device_t dev, device_t child)
 {
-    ACPI_FUNCTION_TRACE((char *)(uintptr_t)__func__);
+    struct acpi_cpu_device *ad;
 
-    /* Disable any entry to the idle function. */
-    cpu_cx_count = 0;
+    ad = device_get_ivars(child);
+    if (ad == NULL)
+	return (NULL);
+    return (&ad->ad_rl);
+}
 
-    /* Signal and wait for all processors to exit acpi_cpu_idle(). */
-#ifdef SMP
-    if (mycpu->gd_cpuid == 0)
-	lwkt_cpusync_simple(0, NULL, NULL);
-#endif
-    DELAY(1);
+static device_t
+acpi_cpu_add_child(device_t dev, int order, const char *name, int unit)
+{
+    struct acpi_cpu_device *ad;
+    device_t child;
 
-    return_VALUE (0);
+    if ((ad = kmalloc(sizeof(*ad), M_TEMP, M_NOWAIT | M_ZERO)) == NULL)
+	return (NULL);
+
+    resource_list_init(&ad->ad_rl);
+    
+    child = device_add_child_ordered(dev, order, name, unit);
+    if (child != NULL)
+	device_set_ivars(child, ad);
+    else
+	kfree(ad, M_TEMP);
+    return (child);
 }
 
 static int
-acpi_cpu_throttle_probe(struct acpi_cpu_softc *sc)
+acpi_cpu_read_ivar(device_t dev, device_t child, int index, uintptr_t *result)
 {
-    uint32_t		 duty_end;
-    ACPI_BUFFER		 buf;
-    ACPI_OBJECT		 obj;
-    ACPI_GENERIC_ADDRESS gas;
-    ACPI_STATUS		 status;
-
-    ACPI_FUNCTION_TRACE((char *)(uintptr_t)__func__);
-
-    ACPI_ASSERTLOCK;
+    struct acpi_cpu_softc *sc;
 
-    /* Get throttling parameters from the FADT.  0 means not supported. */
-    if (device_get_unit(sc->cpu_dev) == 0) {
-	cpu_smi_cmd = AcpiGbl_FADT.SmiCommand;
-	cpu_cst_cnt = AcpiGbl_FADT.CstControl;
-	cpu_duty_offset = AcpiGbl_FADT.DutyOffset;
-	cpu_duty_width = AcpiGbl_FADT.DutyWidth;
+    sc = device_get_softc(dev);
+    switch (index) {
+    case ACPI_IVAR_HANDLE:
+	*result = (uintptr_t)sc->cpu_handle;
+	break;
+#if 0
+    case CPU_IVAR_PCPU:
+	*result = (uintptr_t)sc->cpu_pcpu;
+	break;
+#endif
+    default:
+	return (ENOENT);
     }
-    if (cpu_duty_width == 0 || (cpu_quirks & CPU_QUIRK_NO_THROTTLE) != 0)
-	return (ENXIO);
+    return (0);
+}
 
-    /* Validate the duty offset/width. */
-    duty_end = cpu_duty_offset + cpu_duty_width - 1;
-    if (duty_end > 31) {
-	device_printf(sc->cpu_dev, "CLK_VAL field overflows P_CNT register\n");
-	return (ENXIO);
-    }
-    if (cpu_duty_offset <= 4 && duty_end >= 4) {
-	device_printf(sc->cpu_dev, "CLK_VAL field overlaps THT_EN bit\n");
-	return (ENXIO);
-    }
+static int
+acpi_cpu_shutdown(device_t dev)
+{
+    ACPI_FUNCTION_TRACE((char *)(uintptr_t)__func__);
+
+    /* Allow children to shutdown first. */
+    bus_generic_shutdown(dev);
 
     /*
-     * If not present, fall back to using the processor's P_BLK to find
-     * the P_CNT register.
-     *
-     * Note that some systems seem to duplicate the P_BLK pointer
-     * across multiple CPUs, so not getting the resource is not fatal.
+     * Disable any entry to the idle function.  There is a small race where
+     * an idle thread have passed this check but not gone to sleep.  This
+     * is ok since device_shutdown() does not free the softc, otherwise
+     * we'd have to be sure all threads were evicted before returning.
      */
-    buf.Pointer = &obj;
-    buf.Length = sizeof(obj);
-    status = AcpiEvaluateObject(sc->cpu_handle, "_PTC", NULL, &buf);
-    if (ACPI_SUCCESS(status)) {
-	if (obj.Buffer.Length < sizeof(ACPI_GENERIC_ADDRESS) + 3) {
-	    device_printf(sc->cpu_dev, "_PTC buffer too small\n");
-	    return (ENXIO);
-	}
-	memcpy(&gas, obj.Buffer.Pointer + 3, sizeof(gas));
-	sc->cpu_p_cnt = acpi_bus_alloc_gas(sc->cpu_dev, &cpu_rid, &gas);
-	if (sc->cpu_p_cnt != NULL) {
-	    ACPI_DEBUG_PRINT((ACPI_DB_INFO, "acpi_cpu%d: P_CNT from _PTC\n",
-			     device_get_unit(sc->cpu_dev)));
-	}
-    }
+    cpu_disable_idle = TRUE;
 
-    /* If _PTC not present or other failure, try the P_BLK. */
-    if (sc->cpu_p_cnt == NULL) {
-	/* 
-	 * The spec says P_BLK must be 6 bytes long.  However, some
-	 * systems use it to indicate a fractional set of features
-	 * present so we take anything >= 4.
-	 */
-	if (sc->cpu_p_blk_len < 4)
-	    return (ENXIO);
-	gas.Address = sc->cpu_p_blk;
-	gas.SpaceId = ACPI_ADR_SPACE_SYSTEM_IO;
-	gas.BitWidth = 32;
-	sc->cpu_p_cnt = acpi_bus_alloc_gas(sc->cpu_dev, &cpu_rid, &gas);
-	if (sc->cpu_p_cnt != NULL) {
-	    ACPI_DEBUG_PRINT((ACPI_DB_INFO, "acpi_cpu%d: P_CNT from P_BLK\n",
-			     device_get_unit(sc->cpu_dev)));
-	} else {
-	    device_printf(sc->cpu_dev, "Failed to attach throttling P_CNT\n");
-	    return (ENXIO);
-	}
-    }
-    cpu_rid++;
-
-    return (0);
+    return_VALUE (0);
 }
 
-static int
+static void
 acpi_cpu_cx_probe(struct acpi_cpu_softc *sc)
 {
-    ACPI_GENERIC_ADDRESS gas;
-    struct acpi_cx	*cx_ptr;
-    int			 error;
-
     ACPI_FUNCTION_TRACE((char *)(uintptr_t)__func__);
 
+    /* Use initial sleep value of 1 sec. to start with lowest idle state. */
+    sc->cpu_prev_sleep = 1000000;
+    sc->cpu_cx_lowest = 0;
+
     /*
-     * Bus mastering arbitration control is needed to keep caches coherent
-     * while sleeping in C3.  If it's not present but a working flush cache
-     * instruction is present, flush the caches before entering C3 instead.
-     * Otherwise, just disable C3 completely.
+     * Check for the ACPI 2.0 _CST sleep states object. If we can't find
+     * any, we'll revert to generic FADT/P_BLK Cx control method which will
+     * be handled by acpi_cpu_startup. We need to defer to after having
+     * probed all the cpus in the system before probing for generic Cx
+     * states as we may already have found cpus with valid _CST packages
      */
-    if (AcpiGbl_FADT.Pm2ControlBlock == 0 || AcpiGbl_FADT.Pm2ControlLength == 0) {
-	if ((AcpiGbl_FADT.Flags & ACPI_FADT_WBINVD) != 0 &&
-	    (AcpiGbl_FADT.Flags & ACPI_FADT_WBINVD_FLUSH) == 0) {
-	    cpu_quirks |= CPU_QUIRK_NO_BM_CTRL;
-	    ACPI_DEBUG_PRINT((ACPI_DB_INFO,
-		"acpi_cpu%d: no BM control, using flush cache method\n",
-		device_get_unit(sc->cpu_dev)));
-	} else {
-	    cpu_quirks |= CPU_QUIRK_NO_C3;
-	    ACPI_DEBUG_PRINT((ACPI_DB_INFO,
-		"acpi_cpu%d: no BM control, C3 not available\n",
-		device_get_unit(sc->cpu_dev)));
-	}
+    if (!cpu_cx_generic && acpi_cpu_cx_cst(sc) != 0) {
+	/*
+	 * We were unable to find a _CST package for this cpu or there
+	 * was an error parsing it. Switch back to generic mode.
+	 */
+	cpu_cx_generic = TRUE;
+	if (bootverbose)
+	    device_printf(sc->cpu_dev, "switching to generic Cx mode\n");
     }
 
     /*
-     * First, check for the ACPI 2.0 _CST sleep states object.
-     * If not usable, fall back to the P_BLK's P_LVL2 and P_LVL3.
+     * TODO: _CSD Package should be checked here.
      */
-    sc->cpu_cx_count = 0;
-    error = acpi_cpu_cx_cst(sc);
-    if (error != 0) {
-	cx_ptr = sc->cpu_cx_states;
-
-	/* C1 has been required since just after ACPI 1.0 */
-	cx_ptr->type = ACPI_STATE_C1;
-	cx_ptr->trans_lat = 0;
-	cpu_non_c3 = 0;
-	cx_ptr++;
-	sc->cpu_cx_count++;
-
-	/* 
-	 * The spec says P_BLK must be 6 bytes long.  However, some systems
-	 * use it to indicate a fractional set of features present so we
-	 * take 5 as C2.  Some may also have a value of 7 to indicate
-	 * another C3 but most use _CST for this (as required) and having
-	 * "only" C1-C3 is not a hardship.
-	 */
-	if (sc->cpu_p_blk_len < 5)
-	    goto done;
-
-	/* Validate and allocate resources for C2 (P_LVL2). */
-	gas.SpaceId = ACPI_ADR_SPACE_SYSTEM_IO;
-	gas.BitWidth = 8;
-	if (AcpiGbl_FADT.C2Latency <= 100) {
-	    gas.Address = sc->cpu_p_blk + 4;
-	    cx_ptr->p_lvlx = acpi_bus_alloc_gas(sc->cpu_dev, &cpu_rid, &gas);
-	    if (cx_ptr->p_lvlx != NULL) {
-		cpu_rid++;
-		cx_ptr->type = ACPI_STATE_C2;
-		cx_ptr->trans_lat = AcpiGbl_FADT.C2Latency;
-		cpu_non_c3 = 1;
-		cx_ptr++;
-		sc->cpu_cx_count++;
-	    }
-	}
-	if (sc->cpu_p_blk_len < 6)
-	    goto done;
-
-	/* Validate and allocate resources for C3 (P_LVL3). */
-	if (AcpiGbl_FADT.C3Latency <= 1000 &&
-	    (cpu_quirks & CPU_QUIRK_NO_C3) == 0) {
-
-	    gas.Address = sc->cpu_p_blk + 5;
-	    cx_ptr->p_lvlx = acpi_bus_alloc_gas(sc->cpu_dev, &cpu_rid, &gas);
-	    if (cx_ptr->p_lvlx != NULL) {
-		cpu_rid++;
-		cx_ptr->type = ACPI_STATE_C3;
-		cx_ptr->trans_lat = AcpiGbl_FADT.C3Latency;
-		cx_ptr++;
-		sc->cpu_cx_count++;
-	    }
-	}
-    }
+}
 
-done:
-    /* If no valid registers were found, don't attach. */
-    if (sc->cpu_cx_count == 0)
-	return (ENXIO);
+static void
+acpi_cpu_generic_cx_probe(struct acpi_cpu_softc *sc)
+{
+    ACPI_GENERIC_ADDRESS	 gas;
+    struct acpi_cx		*cx_ptr;
+
+    sc->cpu_cx_count = 0;
+    cx_ptr = sc->cpu_cx_states;
 
     /* Use initial sleep value of 1 sec. to start with lowest idle state. */
     sc->cpu_prev_sleep = 1000000;
 
-    return (0);
+    /* C1 has been required since just after ACPI 1.0 */
+    cx_ptr->type = ACPI_STATE_C1;
+    cx_ptr->trans_lat = 0;
+    cx_ptr++;
+    sc->cpu_cx_count++;
+
+    /* 
+     * The spec says P_BLK must be 6 bytes long.  However, some systems
+     * use it to indicate a fractional set of features present so we
+     * take 5 as C2.  Some may also have a value of 7 to indicate
+     * another C3 but most use _CST for this (as required) and having
+     * "only" C1-C3 is not a hardship.
+     */
+    if (sc->cpu_p_blk_len < 5)
+	return; 
+
+    /* Validate and allocate resources for C2 (P_LVL2). */
+    gas.SpaceId = ACPI_ADR_SPACE_SYSTEM_IO;
+    gas.BitWidth = 8;
+    if (AcpiGbl_FADT.C2Latency <= 100) {
+	gas.Address = sc->cpu_p_blk + 4;
+	cx_ptr->p_lvlx = acpi_bus_alloc_gas(sc->cpu_dev, &sc->cpu_rid, &gas,
+					    RF_SHAREABLE);
+	if (cx_ptr->p_lvlx != NULL) {
+	    sc->cpu_rid++;
+	    cx_ptr->type = ACPI_STATE_C2;
+	    cx_ptr->trans_lat = AcpiGbl_FADT.C2Latency;
+	    cx_ptr++;
+	    sc->cpu_cx_count++;
+	}
+    }
+    if (sc->cpu_p_blk_len < 6)
+	return;
+
+    /* Validate and allocate resources for C3 (P_LVL3). */
+    if (AcpiGbl_FADT.C3Latency <= 1000 && !(cpu_quirks & CPU_QUIRK_NO_C3)) {
+	gas.Address = sc->cpu_p_blk + 5;
+	cx_ptr->p_lvlx = acpi_bus_alloc_gas(sc->cpu_dev, &sc->cpu_rid, &gas,
+					    RF_SHAREABLE);
+	if (cx_ptr->p_lvlx != NULL) {
+	    sc->cpu_rid++;
+	    cx_ptr->type = ACPI_STATE_C3;
+	    cx_ptr->trans_lat = AcpiGbl_FADT.C3Latency;
+	    cx_ptr++;
+	    sc->cpu_cx_count++;
+	}
+    }
+
+    /* Update the largest cx_count seen so far */
+    if (sc->cpu_cx_count > cpu_cx_count)
+	cpu_cx_count = sc->cpu_cx_count;
 }
 
 /*
@@ -591,12 +649,12 @@ acpi_cpu_cx_cst(struct acpi_cpu_softc *sc)
     /* _CST is a package with a count and at least one Cx package. */
     top = (ACPI_OBJECT *)buf.Pointer;
     if (!ACPI_PKG_VALID(top, 2) || acpi_PkgInt32(top, 0, &count) != 0) {
-	device_printf(sc->cpu_dev, "Invalid _CST package\n");
+	device_printf(sc->cpu_dev, "invalid _CST package\n");
 	AcpiOsFree(buf.Pointer);
 	return (ENXIO);
     }
     if (count != top->Package.Count - 1) {
-	device_printf(sc->cpu_dev, "Invalid _CST state count (%d != %d)\n",
+	device_printf(sc->cpu_dev, "invalid _CST state count (%d != %d)\n",
 	       count, top->Package.Count - 1);
 	count = top->Package.Count - 1;
     }
@@ -615,14 +673,14 @@ acpi_cpu_cx_cst(struct acpi_cpu_softc *sc)
 	    acpi_PkgInt32(pkg, 2, &cx_ptr->trans_lat) != 0 ||
 	    acpi_PkgInt32(pkg, 3, &cx_ptr->power) != 0) {
 
-	    device_printf(sc->cpu_dev, "Skipping invalid Cx state package\n");
+	    device_printf(sc->cpu_dev, "skipping invalid Cx state package\n");
 	    continue;
 	}
 
 	/* Validate the state to see if we should use it. */
 	switch (cx_ptr->type) {
 	case ACPI_STATE_C1:
-	    cpu_non_c3 = i;
+	    sc->cpu_non_c3 = i;
 	    cx_ptr++;
 	    sc->cpu_cx_count++;
 	    continue;
@@ -633,7 +691,7 @@ acpi_cpu_cx_cst(struct acpi_cpu_softc *sc)
 				 device_get_unit(sc->cpu_dev), i));
 		continue;
 	    }
-	    cpu_non_c3 = i;
+	    sc->cpu_non_c3 = i;
 	    break;
 	case ACPI_STATE_C3:
 	default:
@@ -657,9 +715,10 @@ acpi_cpu_cx_cst(struct acpi_cpu_softc *sc)
 #endif
 
 	/* Allocate the control register for C2 or C3. */
-	acpi_PkgGas(sc->cpu_dev, pkg, 0, &cpu_rid, &cx_ptr->p_lvlx);
-	if (cx_ptr->p_lvlx != NULL) {
-	    cpu_rid++;
+	acpi_PkgGas(sc->cpu_dev, pkg, 0, &sc->cpu_rid, &cx_ptr->p_lvlx,
+		    RF_SHAREABLE);
+	if (cx_ptr->p_lvlx) {
+	    sc->cpu_rid++;
 	    ACPI_DEBUG_PRINT((ACPI_DB_INFO,
 			     "acpi_cpu%d: Got C%d - %d latency\n",
 			     device_get_unit(sc->cpu_dev), cx_ptr->type,
@@ -680,203 +739,125 @@ static void
 acpi_cpu_startup(void *arg)
 {
     struct acpi_cpu_softc *sc;
-    int count, i;
+    int i;
 
     /* Get set of CPU devices */
     devclass_get_devices(acpi_cpu_devclass, &cpu_devices, &cpu_ndevices);
 
     /*
-     * Make sure all the processors' Cx counts match.  We should probably
-     * also check the contents of each.  However, no known systems have
-     * non-matching Cx counts so we'll deal with this later.
+     * Setup any quirks that might necessary now that we have probed
+     * all the CPUs
      */
-    count = MAX_CX_STATES;
-    for (i = 0; i < cpu_ndevices; i++) {
-	sc = device_get_softc(cpu_devices[i]);
-	count = min(sc->cpu_cx_count, count);
-    }
-    cpu_cx_count = count;
+    acpi_cpu_quirks();
 
-    /* Perform throttling and Cx final initialization. */
-    sc = device_get_softc(cpu_devices[0]);
-    if (sc->cpu_p_cnt != NULL)
-	acpi_cpu_startup_throttling();
-    if (cpu_cx_count > 0)
-	acpi_cpu_startup_cx();
-
-    /* register performance profile change handler */
-    EVENTHANDLER_REGISTER(power_profile_change, acpi_cpu_power_profile, NULL, 0);
-}
-
-/*
- * Power profile change hook.
- *
- * Uses the ACPI lock to avoid reentrancy.
- */
-static void
-acpi_cpu_power_profile(void *arg)
-{
-    int	state;
-    int speed;
-    ACPI_LOCK_DECL;
+    cpu_cx_count = 0;
+    if (cpu_cx_generic) {
+	/*
+	 * We are using generic Cx mode, probe for available Cx states
+	 * for all processors.
+	 */
+	for (i = 0; i < cpu_ndevices; i++) {
+	    sc = device_get_softc(cpu_devices[i]);
+	    acpi_cpu_generic_cx_probe(sc);
+	}
 
-    state = power_profile_get_state();
-    if (state != POWER_PROFILE_PERFORMANCE &&
-	state != POWER_PROFILE_ECONOMY) {
-	return;
+	/*
+	 * Find the highest Cx state common to all CPUs
+	 * in the system, taking quirks into account.
+	 */
+	for (i = 0; i < cpu_ndevices; i++) {
+	    sc = device_get_softc(cpu_devices[i]);
+	    if (sc->cpu_cx_count < cpu_cx_count)
+		cpu_cx_count = sc->cpu_cx_count;
+	}
+    } else {
+	/*
+	 * We are using _CST mode, remove C3 state if necessary.
+	 * Update the largest Cx state supported in the global cpu_cx_count.
+	 * It will be used in the global Cx sysctl handler.
+	 * As we now know for sure that we will be using _CST mode
+	 * install our notify handler.
+	 */
+	for (i = 0; i < cpu_ndevices; i++) {
+	    sc = device_get_softc(cpu_devices[i]);
+	    if (cpu_quirks & CPU_QUIRK_NO_C3) {
+		sc->cpu_cx_count = sc->cpu_non_c3 + 1;
+	    }
+	    if (sc->cpu_cx_count > cpu_cx_count)
+		cpu_cx_count = sc->cpu_cx_count;
+	    AcpiInstallNotifyHandler(sc->cpu_handle, ACPI_DEVICE_NOTIFY,
+		acpi_cpu_notify, sc);
+	}
     }
 
-    ACPI_LOCK;
-    switch(state) {
-    case POWER_PROFILE_PERFORMANCE:
-	speed = cpu_throttle_performance;
-	break;
-    case POWER_PROFILE_ECONOMY:
-	speed = cpu_throttle_economy;
-	break;
-    default:
-	speed = cpu_throttle_state;
-	break;
+    /* Perform Cx final initialization. */
+    for (i = 0; i < cpu_ndevices; i++) {
+	sc = device_get_softc(cpu_devices[i]);
+	acpi_cpu_startup_cx(sc);
     }
-    if (speed != cpu_throttle_state)
-	acpi_cpu_throttle_set(speed);
-    ACPI_UNLOCK;
-}
 
-/*
- * Takes the ACPI lock to avoid fighting anyone over the SMI command
- * port.
- */
-static void
-acpi_cpu_startup_throttling(void)
-{
-    ACPI_LOCK_DECL;
-
-    /* Initialise throttling states */
-    cpu_throttle_max = CPU_MAX_SPEED;
-    cpu_throttle_state = CPU_MAX_SPEED;
-    cpu_throttle_performance = cpu_throttle_max;
-    cpu_throttle_economy = cpu_throttle_performance / 2;
-
-    SYSCTL_ADD_INT(&acpi_cpu_sysctl_ctx,
-		   SYSCTL_CHILDREN(acpi_cpu_sysctl_tree),
-		   OID_AUTO, "throttle_max", CTLFLAG_RD,
-		   &cpu_throttle_max, 0, "maximum CPU speed");
-    SYSCTL_ADD_PROC(&acpi_cpu_sysctl_ctx,
-		    SYSCTL_CHILDREN(acpi_cpu_sysctl_tree),
-		    OID_AUTO, "throttle_state",
-		    CTLTYPE_INT | CTLFLAG_RW, &cpu_throttle_state,
-		    0, acpi_cpu_throttle_sysctl, "I", "current CPU speed");
+    /* Add a sysctl handler to handle global Cx lowest setting */
+    SYSCTL_ADD_PROC(&cpu_sysctl_ctx, SYSCTL_CHILDREN(cpu_sysctl_tree),
+	OID_AUTO, "cx_lowest", CTLTYPE_STRING | CTLFLAG_RW,
+	NULL, 0, acpi_cpu_global_cx_lowest_sysctl, "A",
+	"Global lowest Cx sleep state to use");
 
-    /*
-     * Performance/Economy throttle settings
-     */
-    SYSCTL_ADD_PROC(&acpi_cpu_sysctl_ctx, 
-		    SYSCTL_CHILDREN(acpi_cpu_sysctl_tree),
-		    OID_AUTO, "performance_speed",
-		    CTLTYPE_INT | CTLFLAG_RW, &cpu_throttle_performance,
-		    0, acpi_cpu_throttle_sysctl, "I", "performance CPU speed");
-    SYSCTL_ADD_PROC(&acpi_cpu_sysctl_ctx,
-		    SYSCTL_CHILDREN(acpi_cpu_sysctl_tree),
-		    OID_AUTO, "economy_speed",
-		    CTLTYPE_INT | CTLFLAG_RW, &cpu_throttle_economy,
-		    0, acpi_cpu_throttle_sysctl, "I", "economy CPU speed");
-
-    /* Set initial speed to maximum. */
-    ACPI_LOCK;
-    acpi_cpu_throttle_set(cpu_throttle_max);
-    ACPI_UNLOCK;
-
-    kprintf("acpi_cpu: throttling enabled, %d steps (100%% to %d.%d%%), "
-	   "currently %d.%d%%\n", CPU_MAX_SPEED, CPU_SPEED_PRINTABLE(1),
-	   CPU_SPEED_PRINTABLE(cpu_throttle_state));
+    /* Take over idling from cpu_idle_default(). */
+    cpu_cx_lowest = 0;
+    cpu_disable_idle = FALSE;
+    cpu_idle_hook = acpi_cpu_idle;
 }
 
 static void
-acpi_cpu_startup_cx(void)
+acpi_cpu_cx_list(struct acpi_cpu_softc *sc)
 {
-    struct acpi_cpu_softc *sc;
-    struct sbuf		 sb;
+    struct sbuf sb;
     int i;
 
-    sc = device_get_softc(cpu_devices[0]);
-    sbuf_new(&sb, cpu_cx_supported, sizeof(cpu_cx_supported), SBUF_FIXEDLEN);
-    for (i = 0; i < cpu_cx_count; i++)
+    /*
+     * Set up the list of Cx states
+     */
+    sc->cpu_non_c3 = 0;
+    sbuf_new(&sb, sc->cpu_cx_supported, sizeof(sc->cpu_cx_supported),
+	SBUF_FIXEDLEN);
+    for (i = 0; i < sc->cpu_cx_count; i++) {
 	sbuf_printf(&sb, "C%d/%d ", i + 1, sc->cpu_cx_states[i].trans_lat);
+	if (sc->cpu_cx_states[i].type < ACPI_STATE_C3)
+	    sc->cpu_non_c3 = i;
+    }
     sbuf_trim(&sb);
     sbuf_finish(&sb);
-    SYSCTL_ADD_STRING(&acpi_cpu_sysctl_ctx,
-		      SYSCTL_CHILDREN(acpi_cpu_sysctl_tree),
-		      OID_AUTO, "cx_supported", CTLFLAG_RD, cpu_cx_supported,
-		      0, "Cx/microsecond values for supported Cx states");
-    SYSCTL_ADD_PROC(&acpi_cpu_sysctl_ctx,
-		    SYSCTL_CHILDREN(acpi_cpu_sysctl_tree),
+}	
+
+static void
+acpi_cpu_startup_cx(struct acpi_cpu_softc *sc)
+{
+    acpi_cpu_cx_list(sc);
+    
+    SYSCTL_ADD_STRING(&sc->cpu_sysctl_ctx,
+		      SYSCTL_CHILDREN(sc->cpu_sysctl_tree),
+		      OID_AUTO, "cx_supported", CTLFLAG_RD,
+		      sc->cpu_cx_supported, 0,
+		      "Cx/microsecond values for supported Cx states");
+    SYSCTL_ADD_PROC(&sc->cpu_sysctl_ctx,
+		    SYSCTL_CHILDREN(sc->cpu_sysctl_tree),
 		    OID_AUTO, "cx_lowest", CTLTYPE_STRING | CTLFLAG_RW,
-		    NULL, 0, acpi_cpu_cx_lowest_sysctl, "A",
+		    (void *)sc, 0, acpi_cpu_cx_lowest_sysctl, "A",
 		    "lowest Cx sleep state to use");
-    SYSCTL_ADD_PROC(&acpi_cpu_sysctl_ctx,
-		    SYSCTL_CHILDREN(acpi_cpu_sysctl_tree),
+    SYSCTL_ADD_PROC(&sc->cpu_sysctl_ctx,
+		    SYSCTL_CHILDREN(sc->cpu_sysctl_tree),
 		    OID_AUTO, "cx_usage", CTLTYPE_STRING | CTLFLAG_RD,
-		    NULL, 0, acpi_cpu_usage_sysctl, "A",
+		    (void *)sc, 0, acpi_cpu_usage_sysctl, "A",
 		    "percent usage for each Cx state");
 
 #ifdef notyet
     /* Signal platform that we can handle _CST notification. */
-    if (cpu_cst_cnt != 0) {
-	ACPI_LOCK;
+    if (!cpu_cx_generic && cpu_cst_cnt != 0) {
+	ACPI_LOCK(acpi);
 	AcpiOsWritePort(cpu_smi_cmd, cpu_cst_cnt, 8);
-	ACPI_UNLOCK;
+	ACPI_UNLOCK(acpi);
     }
 #endif
-
-    /* Take over idling from cpu_idle_default_hook(). */
-    if (ncpus == 1)
-	cpu_idle_hook = acpi_cpu_idle;
-    else
-	kprintf("Warning: ACPI idle hook not yet supported for SMP\n");
-}
-
-/*
- * Set CPUs to the new state.
- *
- * Must be called with the ACPI lock held.
- */
-static void
-acpi_cpu_throttle_set(uint32_t speed)
-{
-    struct acpi_cpu_softc	*sc;
-    int				i;
-    uint32_t			p_cnt, clk_val;
-
-    ACPI_ASSERTLOCK;
-
-    /* Iterate over processors */
-    for (i = 0; i < cpu_ndevices; i++) {
-	sc = device_get_softc(cpu_devices[i]);
-	if (sc->cpu_p_cnt == NULL)
-	    continue;
-
-	/* Get the current P_CNT value and disable throttling */
-	p_cnt = CPU_GET_REG(sc->cpu_p_cnt, 4);
-	p_cnt &= ~CPU_P_CNT_THT_EN;
-	CPU_SET_REG(sc->cpu_p_cnt, 4, p_cnt);
-
-	/* If we're at maximum speed, that's all */
-	if (speed < CPU_MAX_SPEED) {
-	    /* Mask the old CLK_VAL off and or-in the new value */
-	    clk_val = (CPU_MAX_SPEED - 1) << cpu_duty_offset;
-	    p_cnt &= ~clk_val;
-	    p_cnt |= (speed << cpu_duty_offset);
-
-	    /* Write the new P_CNT value and then enable throttling */
-	    CPU_SET_REG(sc->cpu_p_cnt, 4, p_cnt);
-	    p_cnt |= CPU_P_CNT_THT_EN;
-	    CPU_SET_REG(sc->cpu_p_cnt, 4, p_cnt);
-	}
-	ACPI_VPRINT(sc->cpu_dev, acpi_device_get_parent_softc(sc->cpu_dev),
-		    "set speed to %d.%d%%\n", CPU_SPEED_PRINTABLE(speed));
-    }
-    cpu_throttle_state = speed;
 }
 
 /*
@@ -886,7 +867,7 @@ acpi_cpu_throttle_set(uint32_t speed)
  * interrupts are re-enabled.
  */
 static void
-acpi_cpu_idle(void)
+acpi_cpu_idle()
 {
     struct	acpi_cpu_softc *sc;
     struct	acpi_cx *cx_next;
@@ -894,7 +875,7 @@ acpi_cpu_idle(void)
     int		bm_active, cx_next_idx, i;
 
     /* If disabled, return immediately. */
-    if (cpu_cx_count == 0) {
+    if (cpu_disable_idle) {
 	ACPI_ENABLE_IRQS();
 	return;
     }
@@ -915,13 +896,39 @@ acpi_cpu_idle(void)
      * find the lowest state that has a latency less than or equal to
      * the length of our last sleep.
      */
-    cx_next_idx = cpu_cx_lowest;
-    if (sc->cpu_prev_sleep < 100)
-	for (i = cpu_cx_lowest; i >= 0; i--)
+    cx_next_idx = sc->cpu_cx_lowest;
+    if (sc->cpu_prev_sleep < 100) {
+	/*
+	 * If we sleep too short all the time, this system may not implement
+	 * C2/3 correctly (i.e. reads return immediately).  In this case,
+	 * back off and use the next higher level.
+	 * It seems that when you have a dual core cpu (like the Intel Core Duo)
+	 * that both cores will get out of C3 state as soon as one of them
+	 * requires it. This breaks the sleep detection logic as the sleep
+	 * counter is local to each cpu. Disable the sleep logic for now as a
+	 * workaround if there's more than one CPU. The right fix would probably
+	 * be to add quirks for system that don't really support C3 state.
+	 */
+	if (ncpus < 2 && sc->cpu_prev_sleep <= 1) {
+	    sc->cpu_short_slp++;
+	    if (sc->cpu_short_slp == 1000 && sc->cpu_cx_lowest != 0) {
+		if (sc->cpu_non_c3 == sc->cpu_cx_lowest && sc->cpu_non_c3 != 0)
+		    sc->cpu_non_c3--;
+		sc->cpu_cx_lowest--;
+		sc->cpu_short_slp = 0;
+		device_printf(sc->cpu_dev,
+		    "too many short sleeps, backing off to C%d\n",
+		    sc->cpu_cx_lowest + 1);
+	    }
+	} else
+	    sc->cpu_short_slp = 0;
+
+	for (i = sc->cpu_cx_lowest; i >= 0; i--)
 	    if (sc->cpu_cx_states[i].trans_lat <= sc->cpu_prev_sleep) {
 		cx_next_idx = i;
 		break;
 	    }
+    }
 
     /*
      * Check for bus master activity.  If there was activity, clear
@@ -933,13 +940,13 @@ acpi_cpu_idle(void)
 	AcpiGetRegister(ACPI_BITREG_BUS_MASTER_STATUS, &bm_active);
 	if (bm_active != 0) {
 	    AcpiSetRegister(ACPI_BITREG_BUS_MASTER_STATUS, 1);
-	    cx_next_idx = min(cx_next_idx, cpu_non_c3);
+	    cx_next_idx = min(cx_next_idx, sc->cpu_non_c3);
 	}
     }
 
     /* Select the next state and update statistics. */
     cx_next = &sc->cpu_cx_states[cx_next_idx];
-    cpu_cx_stats[cx_next_idx]++;
+    sc->cpu_cx_stats[cx_next_idx]++;
     KASSERT(cx_next->type != ACPI_STATE_C0, ("acpi_cpu_idle: C0 sleep"));
 
     /*
@@ -989,38 +996,15 @@ acpi_cpu_idle(void)
 	AcpiSetRegister(ACPI_BITREG_ARB_DISABLE, 0);
 	AcpiSetRegister(ACPI_BITREG_BUS_MASTER_RLD, 0);
     }
+    ACPI_ENABLE_IRQS();
 
     /* Find the actual time asleep in microseconds, minus overhead. */
     end_time = acpi_TimerDelta(end_time, start_time);
     sc->cpu_prev_sleep = PM_USEC(end_time) - cx_next->trans_lat;
-    ACPI_ENABLE_IRQS();
-}
-
-/* Put the CPU in C1 in a machine-dependant way. */
-static void
-acpi_cpu_c1(void)
-{
-#ifdef __ia64__
-    ia64_call_pal_static(PAL_HALT_LIGHT, 0, 0, 0);
-#else
-    splz();
-#ifdef SMP
-    if (!lwkt_runnable())
-	__asm __volatile("sti; hlt");
-    else
-	__asm __volatile("sti; pause");
-#else
-    if (!lwkt_runnable())
-	__asm __volatile("sti; hlt");
-    else
-	__asm __volatile("sti");
-#endif
-#endif /* !__ia64__ */
 }
 
 /*
- * Re-evaluate the _PSS and _CST objects when we are notified that they
- * have changed.
+ * Re-evaluate the _CST object when we are notified that it changed.
  *
  * XXX Re-evaluation disabled until locking is done.
  */
@@ -1028,50 +1012,69 @@ static void
 acpi_cpu_notify(ACPI_HANDLE h, UINT32 notify, void *context)
 {
     struct acpi_cpu_softc *sc = (struct acpi_cpu_softc *)context;
+    struct acpi_cpu_softc *isc;
+    int i;
+    
+    if (notify != ACPI_NOTIFY_CX_STATES)
+	return;
 
-    switch (notify) {
-    case ACPI_CPU_NOTIFY_PERF_STATES:
-	device_printf(sc->cpu_dev, "Performance states changed\n");
-	/* acpi_cpu_px_available(sc); */
-	break;
-    case ACPI_CPU_NOTIFY_CX_STATES:
-	device_printf(sc->cpu_dev, "Cx states changed\n");
-	/* acpi_cpu_cx_cst(sc); */
-	break;
-    default:
-	device_printf(sc->cpu_dev, "Unknown notify %#x\n", notify);
-	break;
+    /* Update the list of Cx states. */
+    acpi_cpu_cx_cst(sc);
+    acpi_cpu_cx_list(sc);
+
+    /* Update the new lowest useable Cx state for all CPUs. */
+    crit_enter();
+    cpu_cx_count = 0;
+    for (i = 0; i < cpu_ndevices; i++) {
+	isc = device_get_softc(cpu_devices[i]);
+	if (isc->cpu_cx_count > cpu_cx_count)
+	    cpu_cx_count = isc->cpu_cx_count;
     }
+    crit_exit();
 }
 
 static int
-acpi_cpu_quirks(struct acpi_cpu_softc *sc)
+acpi_cpu_quirks(void)
 {
+    device_t acpi_dev;
+    uint32_t val;
+
+    ACPI_FUNCTION_TRACE((char *)(uintptr_t)__func__);
 
     /*
-     * C3 on multiple CPUs requires using the expensive flush cache
-     * instruction.
+     * Bus mastering arbitration control is needed to keep caches coherent
+     * while sleeping in C3.  If it's not present but a working flush cache
+     * instruction is present, flush the caches before entering C3 instead.
+     * Otherwise, just disable C3 completely.
      */
-    if (ncpus > 1)
+    if (AcpiGbl_FADT.Pm2ControlBlock == 0 ||
+	AcpiGbl_FADT.Pm2ControlLength == 0) {
+	if ((AcpiGbl_FADT.Flags & ACPI_FADT_WBINVD) &&
+	    (AcpiGbl_FADT.Flags & ACPI_FADT_WBINVD_FLUSH) == 0) {
+	    cpu_quirks |= CPU_QUIRK_NO_BM_CTRL;
+	    ACPI_DEBUG_PRINT((ACPI_DB_INFO,
+		"acpi_cpu: no BM control, using flush cache method\n"));
+	} else {
+	    cpu_quirks |= CPU_QUIRK_NO_C3;
+	    ACPI_DEBUG_PRINT((ACPI_DB_INFO,
+		"acpi_cpu: no BM control, C3 not available\n"));
+	}
+    }
+
+    /*
+     * If we are using generic Cx mode, C3 on multiple CPUs requires using
+     * the expensive flush cache instruction.
+     */
+    if (cpu_cx_generic && ncpus > 1) {
 	cpu_quirks |= CPU_QUIRK_NO_BM_CTRL;
+	ACPI_DEBUG_PRINT((ACPI_DB_INFO,
+	    "acpi_cpu: SMP, using flush cache mode for C3\n"));
+    }
 
-#ifdef notyet
     /* Look for various quirks of the PIIX4 part. */
     acpi_dev = pci_find_device(PCI_VENDOR_INTEL, PCI_DEVICE_82371AB_3);
     if (acpi_dev != NULL) {
 	switch (pci_get_revid(acpi_dev)) {
-	/*
-	 * Disable throttling control on PIIX4 A and B-step.
-	 * See specification changes #13 ("Manual Throttle Duty Cycle")
-	 * and #14 ("Enabling and Disabling Manual Throttle"), plus
-	 * erratum #5 ("STPCLK# Deassertion Time") from the January
-	 * 2002 PIIX4 specification update.  Note that few (if any)
-	 * mobile systems ever used this part.
-	 */
-	case PCI_REVISION_A_STEP:
-	case PCI_REVISION_B_STEP:
-	    cpu_quirks |= CPU_QUIRK_NO_THROTTLE;
-	    /* FALLTHROUGH */
 	/*
 	 * Disable C3 support for all PIIX4 chipsets.  Some of these parts
 	 * do not report the BMIDE status to the BM status register and
@@ -1083,46 +1086,30 @@ acpi_cpu_quirks(struct acpi_cpu_softc *sc)
 	 * See erratum #18 ("C3 Power State/BMIDE and Type-F DMA
 	 * Livelock") from the January 2002 PIIX4 specification update.
 	 * Applies to all PIIX4 models.
+	 *
+	 * Also, make sure that all interrupts cause a "Stop Break"
+	 * event to exit from C2 state.
 	 */
+	case PCI_REVISION_A_STEP:
+	case PCI_REVISION_B_STEP:
 	case PCI_REVISION_4E:
 	case PCI_REVISION_4M:
 	    cpu_quirks |= CPU_QUIRK_NO_C3;
+	    ACPI_DEBUG_PRINT((ACPI_DB_INFO,
+		"acpi_cpu: working around PIIX4 bug, disabling C3\n"));
+
+	    val = pci_read_config(acpi_dev, PIIX4_DEVACTB_REG, 4);
+	    if ((val & PIIX4_STOP_BREAK_MASK) != PIIX4_STOP_BREAK_MASK) {
+		ACPI_DEBUG_PRINT((ACPI_DB_INFO,
+		    "PIIX4: enabling IRQs to generate Stop Break\n"));
+	    	val |= PIIX4_STOP_BREAK_MASK;
+		pci_write_config(acpi_dev, PIIX4_DEVACTB_REG, val, 4);
+	    }
 	    break;
 	default:
 	    break;
 	}
     }
-#endif
-
-    return (0);
-}
-
-/* Handle changes in the CPU throttling setting. */
-static int
-acpi_cpu_throttle_sysctl(SYSCTL_HANDLER_ARGS)
-{
-    uint32_t	*argp;
-    uint32_t	 arg;
-    int		 error;
-    ACPI_LOCK_DECL;
-
-    argp = (uint32_t *)oidp->oid_arg1;
-    arg = *argp;
-    error = sysctl_handle_int(oidp, &arg, 0, req);
-
-    /* Error or no new value */
-    if (error != 0 || req->newptr == NULL)
-	return (error);
-    if (arg < 1 || arg > cpu_throttle_max)
-	return (EINVAL);
-
-    /* If throttling changed, notify the BIOS of the new rate. */
-    ACPI_LOCK;
-    if (*argp != arg) {
-	*argp = arg;
-	acpi_cpu_throttle_set(arg);
-    }
-    ACPI_UNLOCK;
 
     return (0);
 }
@@ -1130,18 +1117,20 @@ acpi_cpu_throttle_sysctl(SYSCTL_HANDLER_ARGS)
 static int
 acpi_cpu_usage_sysctl(SYSCTL_HANDLER_ARGS)
 {
+    struct acpi_cpu_softc *sc;
     struct sbuf	 sb;
     char	 buf[128];
     int		 i;
     uintmax_t	 fract, sum, whole;
 
+    sc = (struct acpi_cpu_softc *) arg1;
     sum = 0;
-    for (i = 0; i < cpu_cx_count; i++)
-	sum += cpu_cx_stats[i];
+    for (i = 0; i < sc->cpu_cx_count; i++)
+	sum += sc->cpu_cx_stats[i];
     sbuf_new(&sb, buf, sizeof(buf), SBUF_FIXEDLEN);
-    for (i = 0; i < cpu_cx_count; i++) {
+    for (i = 0; i < sc->cpu_cx_count; i++) {
 	if (sum > 0) {
-	    whole = (uintmax_t)cpu_cx_stats[i] * 100;
+	    whole = (uintmax_t)sc->cpu_cx_stats[i] * 100;
 	    fract = (whole % sum) * 100;
 	    sbuf_printf(&sb, "%u.%02u%% ", (u_int)(whole / sum),
 		(u_int)(fract / sum));
@@ -1156,14 +1145,59 @@ acpi_cpu_usage_sysctl(SYSCTL_HANDLER_ARGS)
     return (0);
 }
 
+static int
+acpi_cpu_set_cx_lowest(struct acpi_cpu_softc *sc, int val)
+{
+    int i;
+
+    sc->cpu_cx_lowest = val;
+
+    /* If not disabling, cache the new lowest non-C3 state. */
+    sc->cpu_non_c3 = 0;
+    for (i = sc->cpu_cx_lowest; i >= 0; i--) {
+	if (sc->cpu_cx_states[i].type < ACPI_STATE_C3) {
+	    sc->cpu_non_c3 = i;
+	    break;
+	}
+    }
+
+    /* Reset the statistics counters. */
+    bzero(sc->cpu_cx_stats, sizeof(sc->cpu_cx_stats));
+    return (0);
+}
+
 static int
 acpi_cpu_cx_lowest_sysctl(SYSCTL_HANDLER_ARGS)
 {
     struct	 acpi_cpu_softc *sc;
     char	 state[8];
-    int		 val, error, i;
+    int		 val, error;
+
+    sc = (struct acpi_cpu_softc *) arg1;
+    ksnprintf(state, sizeof(state), "C%d", sc->cpu_cx_lowest + 1);
+    error = sysctl_handle_string(oidp, state, sizeof(state), req);
+    if (error != 0 || req->newptr == NULL)
+	return (error);
+    if (strlen(state) < 2 || toupper(state[0]) != 'C')
+	return (EINVAL);
+    val = (int) strtol(state + 1, NULL, 10) - 1;
+    if (val < 0 || val > sc->cpu_cx_count - 1)
+	return (EINVAL);
+
+    crit_enter();
+    acpi_cpu_set_cx_lowest(sc, val);
+    crit_exit();
+
+    return (0);
+}
+
+static int
+acpi_cpu_global_cx_lowest_sysctl(SYSCTL_HANDLER_ARGS)
+{
+    struct	acpi_cpu_softc *sc;
+    char	state[8];
+    int		val, error, i;
 
-    sc = device_get_softc(cpu_devices[0]);
     ksnprintf(state, sizeof(state), "C%d", cpu_cx_lowest + 1);
     error = sysctl_handle_string(oidp, state, sizeof(state), req);
     if (error != 0 || req->newptr == NULL)
@@ -1173,20 +1207,41 @@ acpi_cpu_cx_lowest_sysctl(SYSCTL_HANDLER_ARGS)
     val = (int) strtol(state + 1, NULL, 10) - 1;
     if (val < 0 || val > cpu_cx_count - 1)
 	return (EINVAL);
-
     cpu_cx_lowest = val;
 
-    /* If not disabling, cache the new lowest non-C3 state. */
-    cpu_non_c3 = 0;
-    for (i = cpu_cx_lowest; i >= 0; i--) {
-	if (sc->cpu_cx_states[i].type < ACPI_STATE_C3) {
-	    cpu_non_c3 = i;
-	    break;
-	}
+    /* Update the new lowest useable Cx state for all CPUs. */
+    crit_enter();
+    for (i = 0; i < cpu_ndevices; i++) {
+	sc = device_get_softc(cpu_devices[i]);
+	acpi_cpu_set_cx_lowest(sc, val);
     }
-
-    /* Reset the statistics counters. */
-    bzero(cpu_cx_stats, sizeof(cpu_cx_stats));
+    crit_exit();
 
     return (0);
 }
+
+/*
+ * Put the CPU in C1 in a machine-dependant way.
+ * XXX: shouldn't be here!
+ */
+static void
+acpi_cpu_c1(void)
+{
+#ifdef __ia64__
+    ia64_call_pal_static(PAL_HALT_LIGHT, 0, 0, 0);
+#else
+    splz();
+#ifdef SMP
+    if (!lwkt_runnable())
+        __asm __volatile("sti; hlt");
+    else
+        __asm __volatile("sti; pause");
+#else
+    if (!lwkt_runnable())
+        __asm __volatile("sti; hlt");
+    else
+        __asm __volatile("sti");
+#endif
+#endif /* !__ia64__ */
+}
+
diff --git a/sys/dev/acpica5/acpi_if.m b/sys/dev/acpica5/acpi_if.m
index 9c46815240..ab8fc0fc87 100644
--- a/sys/dev/acpica5/acpi_if.m
+++ b/sys/dev/acpica5/acpi_if.m
@@ -23,7 +23,7 @@
 # OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
 # SUCH DAMAGE.
 #
-# $DragonFly: src/sys/dev/acpica5/acpi_if.m,v 1.1 2008/08/27 16:35:19 hasso Exp $
+# $DragonFly: src/sys/dev/acpica5/acpi_if.m,v 1.2 2008/09/05 10:28:35 hasso Exp $
 #
 
 #include <sys/bus.h>
@@ -61,6 +61,19 @@ METHOD char * id_probe {
 	char		**ids;
 } DEFAULT acpi_generic_id_probe;
 
+#
+# Query a given driver for its supported feature(s).  This should be
+# called by the parent bus before the driver is probed.
+#
+# driver_t *driver:  child driver
+#
+# u_int *features:  returned bitmask of all supported features
+#
+STATICMETHOD int get_features {
+	driver_t	*driver;
+	u_int		*features;
+};
+
 #
 # Read embedded controller (EC) address space
 #
diff --git a/sys/dev/acpica5/acpi_package.c b/sys/dev/acpica5/acpi_package.c
index 30c704f741..310dba7fc7 100644
--- a/sys/dev/acpica5/acpi_package.c
+++ b/sys/dev/acpica5/acpi_package.c
@@ -24,7 +24,7 @@
  * SUCH DAMAGE.
  *
  * $FreeBSD: src/sys/dev/acpica/acpi_package.c,v 1.3 2004/04/09 06:40:03 njl Exp $
- * $DragonFly: src/sys/dev/acpica5/acpi_package.c,v 1.3 2006/10/25 20:55:52 dillon Exp $
+ * $DragonFly: src/sys/dev/acpica5/acpi_package.c,v 1.4 2008/09/05 10:28:35 hasso Exp $
  */
 
 #include <sys/param.h>
@@ -101,7 +101,7 @@ acpi_PkgStr(ACPI_OBJECT *res, int idx, void *dst, size_t size)
 
 int
 acpi_PkgGas(device_t dev, ACPI_OBJECT *res, int idx, int *rid,
-	struct resource **dst)
+	struct resource **dst, u_int flags)
 {
     ACPI_GENERIC_ADDRESS gas;
     ACPI_OBJECT		*obj;
@@ -114,7 +114,7 @@ acpi_PkgGas(device_t dev, ACPI_OBJECT *res, int idx, int *rid,
     }
 
     memcpy(&gas, obj->Buffer.Pointer + 3, sizeof(gas));
-    *dst = acpi_bus_alloc_gas(dev, rid, &gas);
+    *dst = acpi_bus_alloc_gas(dev, rid, &gas, flags);
     if (*dst == NULL)
 	return (ENXIO);
 
diff --git a/sys/dev/acpica5/acpivar.h b/sys/dev/acpica5/acpivar.h
index c43eccd3ce..a21e7975ec 100644
--- a/sys/dev/acpica5/acpivar.h
+++ b/sys/dev/acpica5/acpivar.h
@@ -26,7 +26,7 @@
  * SUCH DAMAGE.
  *
  * $FreeBSD: src/sys/dev/acpica/acpivar.h,v 1.71 2004/06/13 22:52:30 njl Exp $
- * $DragonFly: src/sys/dev/acpica5/acpivar.h,v 1.14 2008/08/28 16:46:54 hasso Exp $
+ * $DragonFly: src/sys/dev/acpica5/acpivar.h,v 1.15 2008/09/05 10:28:35 hasso Exp $
  */
 
 #include "acpi_if.h"
@@ -147,6 +147,20 @@ struct acpi_prw_data {
 #define	ACPI_INTR_APIC		1
 #define	ACPI_INTR_SAPIC		2
 
+/*
+ * Various features and capabilities for the acpi_get_features() method.
+ * In particular, these are used for the ACPI 3.0 _PDC and _OSC methods.
+ */
+#define ACPI_CAP_PERF_MSRS	(1 << 0) /* Intel SpeedStep PERF_CTL MSRs */
+#define ACPI_CAP_C1_IO_HALT	(1 << 1) /* Intel C1 "IO then halt" sequence */
+#define ACPI_CAP_THR_MSRS	(1 << 2) /* Intel OnDemand throttling MSRs */
+#define ACPI_CAP_SMP_SAME	(1 << 3) /* MP C1, Px, and Tx (all the same) */
+#define ACPI_CAP_SMP_SAME_C3	(1 << 4) /* MP C2 and C3 (all the same) */
+#define ACPI_CAP_SMP_DIFF_PX	(1 << 5) /* MP Px (different, using _PSD) */
+#define ACPI_CAP_SMP_DIFF_CX	(1 << 6) /* MP Cx (different, using _CSD) */
+#define ACPI_CAP_SMP_DIFF_TX	(1 << 7) /* MP Tx (different, using _TSD) */
+#define ACPI_CAP_SMP_C1_NATIVE	(1 << 8) /* MP C1 support other than halt */
+
 /*
  * Note that the low ivar values are reserved to provide
  * interface compatibility with ISA drivers which can also
@@ -252,7 +266,7 @@ ACPI_STATUS	acpi_Disable(struct acpi_softc *sc);
 void		acpi_UserNotify(const char *subsystem, ACPI_HANDLE h,
 		    uint8_t notify);
 struct resource *acpi_bus_alloc_gas(device_t dev, int *rid,
-		    ACPI_GENERIC_ADDRESS *gas);
+		    ACPI_GENERIC_ADDRESS *gas, u_int flags);
 
 struct acpi_parse_resource_set {
     void	(*set_init)(device_t dev, void *arg, void **context);
@@ -354,7 +368,7 @@ int		acpi_PkgInt(ACPI_OBJECT *res, int idx, ACPI_INTEGER *dst);
 int		acpi_PkgInt32(ACPI_OBJECT *res, int idx, uint32_t *dst);
 int		acpi_PkgStr(ACPI_OBJECT *res, int idx, void *dst, size_t size);
 int		acpi_PkgGas(device_t dev, ACPI_OBJECT *res, int idx, int *rid,
-			    struct resource **dst);
+			    struct resource **dst, u_int flags);
 ACPI_HANDLE	acpi_GetReference(ACPI_HANDLE scope, ACPI_OBJECT *obj);
 
 /* ACPI task kernel thread initialization. */
diff --git a/sys/kern/subr_bus.c b/sys/kern/subr_bus.c
index 3975dca634..4bda88552d 100644
--- a/sys/kern/subr_bus.c
+++ b/sys/kern/subr_bus.c
@@ -24,7 +24,7 @@
  * SUCH DAMAGE.
  *
  * $FreeBSD: src/sys/kern/subr_bus.c,v 1.54.2.9 2002/10/10 15:13:32 jhb Exp $
- * $DragonFly: src/sys/kern/subr_bus.c,v 1.42 2008/09/02 17:48:26 hasso Exp $
+ * $DragonFly: src/sys/kern/subr_bus.c,v 1.43 2008/09/05 10:28:35 hasso Exp $
  */
 
 #include "opt_bus.h"
@@ -357,6 +357,47 @@ devclass_get_devices(devclass_t dc, device_t **devlistp, int *devcountp)
 	return(0);
 }
 
+/**
+ * @brief Get a list of drivers in the devclass
+ *
+ * An array containing a list of pointers to all the drivers in the
+ * given devclass is allocated and returned in @p *listp.  The number
+ * of drivers in the array is returned in @p *countp. The caller should
+ * free the array using @c free(p, M_TEMP).
+ *
+ * @param dc            the devclass to examine
+ * @param listp         gives location for array pointer return value
+ * @param countp        gives location for number of array elements
+ *                      return value
+ *
+ * @retval 0            success
+ * @retval ENOMEM       the array allocation failed
+ */
+int
+devclass_get_drivers(devclass_t dc, driver_t ***listp, int *countp)
+{
+        driverlink_t dl;
+        driver_t **list;
+        int count;
+
+        count = 0;
+        TAILQ_FOREACH(dl, &dc->drivers, link)
+                count++;
+        list = kmalloc(count * sizeof(driver_t *), M_TEMP, M_NOWAIT);
+        if (list == NULL)
+                return (ENOMEM);
+
+        count = 0;
+        TAILQ_FOREACH(dl, &dc->drivers, link) {
+                list[count] = dl->driver;
+                count++;
+        }
+        *listp = list;
+        *countp = count;
+
+        return (0);
+}
+
 int
 devclass_get_maxunit(devclass_t dc)
 {
diff --git a/sys/sys/bus.h b/sys/sys/bus.h
index 79b4b1df5e..38402c26be 100644
--- a/sys/sys/bus.h
+++ b/sys/sys/bus.h
@@ -24,7 +24,7 @@
  * SUCH DAMAGE.
  *
  * $FreeBSD: src/sys/sys/bus.h,v 1.30.2.5 2004/03/17 17:54:25 njl Exp $
- * $DragonFly: src/sys/sys/bus.h,v 1.27 2008/08/17 04:32:32 sephe Exp $
+ * $DragonFly: src/sys/sys/bus.h,v 1.28 2008/09/05 10:28:35 hasso Exp $
  */
 
 #ifndef _SYS_BUS_H_
@@ -352,6 +352,7 @@ const char 	*devclass_get_name(devclass_t dc);
 device_t	devclass_get_device(devclass_t dc, int unit);
 void	*devclass_get_softc(devclass_t dc, int unit);
 int	devclass_get_devices(devclass_t dc, device_t **listp, int *countp);
+int	devclass_get_drivers(devclass_t dc, driver_t ***listp, int *countp);
 int	devclass_get_maxunit(devclass_t dc);
 void	devclass_set_parent(devclass_t dc, devclass_t pdc);
 devclass_t	devclass_get_parent(devclass_t dc);