Merge branch 'core-smp-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

Pull core SMP updates from Ingo Molnar:
 "Two main change is generic vCPU pinning and physical CPU SMP-call
  support, for Xen to be able to perform certain calls on specific
  physical CPUs - by Juergen Gross"

* 'core-smp-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  smp: Allocate smp_call_on_cpu() workqueue on stack too
  hwmon: Use smp_call_on_cpu() for dell-smm i8k
  dcdbas: Make use of smp_call_on_cpu()
  xen: Add xen_pin_vcpu() to support calling functions on a dedicated pCPU
  smp: Add function to execute a function synchronously on a CPU
  virt, sched: Add generic vCPU pinning support
  xen: Sync xen header

11 files changed, 274 insertions, 59 deletions

diff --git a/MAINTAINERS b/MAINTAINERS
index c8926c1acd22..f2ae3a4bc5c2 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -8860,6 +8860,7 @@ S:	Supported
 F:	Documentation/virtual/paravirt_ops.txt
 F:	arch/*/kernel/paravirt*
 F:	arch/*/include/asm/paravirt.h
+F:	include/linux/hypervisor.h
 
 PARIDE DRIVERS FOR PARALLEL PORT IDE DEVICES
 M:	Tim Waugh <tim@cyberelk.net>
diff --git a/arch/x86/include/asm/hypervisor.h b/arch/x86/include/asm/hypervisor.h
index 055ea9941dd5..67942b6ad4b7 100644
--- a/arch/x86/include/asm/hypervisor.h
+++ b/arch/x86/include/asm/hypervisor.h
@@ -43,6 +43,9 @@ struct hypervisor_x86 {
 
 	/* X2APIC detection (run once per boot) */
 	bool		(*x2apic_available)(void);
+
+	/* pin current vcpu to specified physical cpu (run rarely) */
+	void		(*pin_vcpu)(int);
 };
 
 extern const struct hypervisor_x86 *x86_hyper;
@@ -56,6 +59,7 @@ extern const struct hypervisor_x86 x86_hyper_kvm;
 extern void init_hypervisor(struct cpuinfo_x86 *c);
 extern void init_hypervisor_platform(void);
 extern bool hypervisor_x2apic_available(void);
+extern void hypervisor_pin_vcpu(int cpu);
 #else
 static inline void init_hypervisor(struct cpuinfo_x86 *c) { }
 static inline void init_hypervisor_platform(void) { }
diff --git a/arch/x86/kernel/cpu/hypervisor.c b/arch/x86/kernel/cpu/hypervisor.c
index 27e46658ebe3..35691a6b0d32 100644
--- a/arch/x86/kernel/cpu/hypervisor.c
+++ b/arch/x86/kernel/cpu/hypervisor.c
@@ -86,3 +86,14 @@ bool __init hypervisor_x2apic_available(void)
 	       x86_hyper->x2apic_available &&
 	       x86_hyper->x2apic_available();
 }
+
+void hypervisor_pin_vcpu(int cpu)
+{
+	if (!x86_hyper)
+		return;
+
+	if (x86_hyper->pin_vcpu)
+		x86_hyper->pin_vcpu(cpu);
+	else
+		WARN_ONCE(1, "vcpu pinning requested but not supported!\n");
+}
diff --git a/arch/x86/xen/enlighten.c b/arch/x86/xen/enlighten.c
index b86ebb1a9a7f..bc9aaba01a22 100644
--- a/arch/x86/xen/enlighten.c
+++ b/arch/x86/xen/enlighten.c
@@ -1925,6 +1925,45 @@ static void xen_set_cpu_features(struct cpuinfo_x86 *c)
 	}
 }
 
+static void xen_pin_vcpu(int cpu)
+{
+	static bool disable_pinning;
+	struct sched_pin_override pin_override;
+	int ret;
+
+	if (disable_pinning)
+		return;
+
+	pin_override.pcpu = cpu;
+	ret = HYPERVISOR_sched_op(SCHEDOP_pin_override, &pin_override);
+
+	/* Ignore errors when removing override. */
+	if (cpu < 0)
+		return;
+
+	switch (ret) {
+	case -ENOSYS:
+		pr_warn("Unable to pin on physical cpu %d. In case of problems consider vcpu pinning.\n",
+			cpu);
+		disable_pinning = true;
+		break;
+	case -EPERM:
+		WARN(1, "Trying to pin vcpu without having privilege to do so\n");
+		disable_pinning = true;
+		break;
+	case -EINVAL:
+	case -EBUSY:
+		pr_warn("Physical cpu %d not available for pinning. Check Xen cpu configuration.\n",
+			cpu);
+		break;
+	case 0:
+		break;
+	default:
+		WARN(1, "rc %d while trying to pin vcpu\n", ret);
+		disable_pinning = true;
+	}
+}
+
 const struct hypervisor_x86 x86_hyper_xen = {
 	.name			= "Xen",
 	.detect			= xen_platform,
@@ -1933,6 +1972,7 @@ const struct hypervisor_x86 x86_hyper_xen = {
 #endif
 	.x2apic_available	= xen_x2apic_para_available,
 	.set_cpu_features       = xen_set_cpu_features,
+	.pin_vcpu               = xen_pin_vcpu,
 };
 EXPORT_SYMBOL(x86_hyper_xen);
 
diff --git a/drivers/firmware/dcdbas.c b/drivers/firmware/dcdbas.c
index 829eec8959f2..2fe1a130189f 100644
--- a/drivers/firmware/dcdbas.c
+++ b/drivers/firmware/dcdbas.c
@@ -23,6 +23,7 @@
 #include <linux/platform_device.h>
 #include <linux/dma-mapping.h>
 #include <linux/errno.h>
+#include <linux/cpu.h>
 #include <linux/gfp.h>
 #include <linux/init.h>
 #include <linux/kernel.h>
@@ -238,33 +239,14 @@ static ssize_t host_control_on_shutdown_store(struct device *dev,
 	return count;
 }
 
-/**
- * dcdbas_smi_request: generate SMI request
- *
- * Called with smi_data_lock.
- */
-int dcdbas_smi_request(struct smi_cmd *smi_cmd)
+static int raise_smi(void *par)
 {
-	cpumask_var_t old_mask;
-	int ret = 0;
-
-	if (smi_cmd->magic != SMI_CMD_MAGIC) {
-		dev_info(&dcdbas_pdev->dev, "%s: invalid magic value\n",
-			 __func__);
-		return -EBADR;
-	}
+	struct smi_cmd *smi_cmd = par;
 
-	/* SMI requires CPU 0 */
-	if (!alloc_cpumask_var(&old_mask, GFP_KERNEL))
-		return -ENOMEM;
-
-	cpumask_copy(old_mask, &current->cpus_allowed);
-	set_cpus_allowed_ptr(current, cpumask_of(0));
 	if (smp_processor_id() != 0) {
 		dev_dbg(&dcdbas_pdev->dev, "%s: failed to get CPU 0\n",
 			__func__);
-		ret = -EBUSY;
-		goto out;
+		return -EBUSY;
 	}
 
 	/* generate SMI */
@@ -280,9 +262,28 @@ int dcdbas_smi_request(struct smi_cmd *smi_cmd)
 		: "memory"
 	);
 
-out:
-	set_cpus_allowed_ptr(current, old_mask);
-	free_cpumask_var(old_mask);
+	return 0;
+}
+/**
+ * dcdbas_smi_request: generate SMI request
+ *
+ * Called with smi_data_lock.
+ */
+int dcdbas_smi_request(struct smi_cmd *smi_cmd)
+{
+	int ret;
+
+	if (smi_cmd->magic != SMI_CMD_MAGIC) {
+		dev_info(&dcdbas_pdev->dev, "%s: invalid magic value\n",
+			 __func__);
+		return -EBADR;
+	}
+
+	/* SMI requires CPU 0 */
+	get_online_cpus();
+	ret = smp_call_on_cpu(0, raise_smi, smi_cmd, true);
+	put_online_cpus();
+
 	return ret;
 }
 
diff --git a/drivers/hwmon/dell-smm-hwmon.c b/drivers/hwmon/dell-smm-hwmon.c
index acf9c0361d9f..34704b0451b4 100644
--- a/drivers/hwmon/dell-smm-hwmon.c
+++ b/drivers/hwmon/dell-smm-hwmon.c
@@ -21,6 +21,7 @@
 
 #define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
 
+#include <linux/cpu.h>
 #include <linux/delay.h>
 #include <linux/module.h>
 #include <linux/types.h>
@@ -36,6 +37,7 @@
 #include <linux/io.h>
 #include <linux/sched.h>
 #include <linux/ctype.h>
+#include <linux/smp.h>
 
 #include <linux/i8k.h>
 
@@ -134,11 +136,11 @@ static inline const char *i8k_get_dmi_data(int field)
 /*
  * Call the System Management Mode BIOS. Code provided by Jonathan Buzzard.
  */
-static int i8k_smm(struct smm_regs *regs)
+static int i8k_smm_func(void *par)
 {
 	int rc;
+	struct smm_regs *regs = par;
 	int eax = regs->eax;
-	cpumask_var_t old_mask;
 
 #ifdef DEBUG
 	int ebx = regs->ebx;
@@ -149,16 +151,8 @@ static int i8k_smm(struct smm_regs *regs)
 #endif
 
 	/* SMM requires CPU 0 */
-	if (!alloc_cpumask_var(&old_mask, GFP_KERNEL))
-		return -ENOMEM;
-	cpumask_copy(old_mask, &current->cpus_allowed);
-	rc = set_cpus_allowed_ptr(current, cpumask_of(0));
-	if (rc)
-		goto out;
-	if (smp_processor_id() != 0) {
-		rc = -EBUSY;
-		goto out;
-	}
+	if (smp_processor_id() != 0)
+		return -EBUSY;
 
 #if defined(CONFIG_X86_64)
 	asm volatile("pushq %%rax\n\t"
@@ -216,10 +210,6 @@ static int i8k_smm(struct smm_regs *regs)
 	if (rc != 0 || (regs->eax & 0xffff) == 0xffff || regs->eax == eax)
 		rc = -EINVAL;
 
-out:
-	set_cpus_allowed_ptr(current, old_mask);
-	free_cpumask_var(old_mask);
-
 #ifdef DEBUG
 	rettime = ktime_get();
 	delta = ktime_sub(rettime, calltime);
@@ -232,6 +222,20 @@ out:
 }
 
 /*
+ * Call the System Management Mode BIOS.
+ */
+static int i8k_smm(struct smm_regs *regs)
+{
+	int ret;
+
+	get_online_cpus();
+	ret = smp_call_on_cpu(0, i8k_smm_func, regs, true);
+	put_online_cpus();
+
+	return ret;
+}
+
+/*
  * Read the fan status.
  */
 static int i8k_get_fan_status(int fan)
diff --git a/include/linux/hypervisor.h b/include/linux/hypervisor.h
new file mode 100644
index 000000000000..3fa5ef2b3759
--- /dev/null
+++ b/include/linux/hypervisor.h
@@ -0,0 +1,17 @@
+#ifndef __LINUX_HYPEVISOR_H
+#define __LINUX_HYPEVISOR_H
+
+/*
+ *	Generic Hypervisor support
+ *		Juergen Gross <jgross@suse.com>
+ */
+
+#ifdef CONFIG_HYPERVISOR_GUEST
+#include <asm/hypervisor.h>
+#else
+static inline void hypervisor_pin_vcpu(int cpu)
+{
+}
+#endif
+
+#endif /* __LINUX_HYPEVISOR_H */
diff --git a/include/linux/smp.h b/include/linux/smp.h
index eccae4690f41..8e0cb7a0f836 100644
--- a/include/linux/smp.h
+++ b/include/linux/smp.h
@@ -196,6 +196,9 @@ extern void arch_enable_nonboot_cpus_end(void);
 
 void smp_setup_processor_id(void);
 
+int smp_call_on_cpu(unsigned int cpu, int (*func)(void *), void *par,
+		    bool phys);
+
 /* SMP core functions */
 int smpcfd_prepare_cpu(unsigned int cpu);
 int smpcfd_dead_cpu(unsigned int cpu);
diff --git a/include/xen/interface/sched.h b/include/xen/interface/sched.h
index f18490985fc8..a4c4d735d781 100644
--- a/include/xen/interface/sched.h
+++ b/include/xen/interface/sched.h
@@ -3,6 +3,24 @@
  *
  * Scheduler state interactions
  *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
  * Copyright (c) 2005, Keir Fraser <keir@xensource.com>
  */
 
@@ -12,18 +30,30 @@
 #include <xen/interface/event_channel.h>
 
 /*
+ * Guest Scheduler Operations
+ *
+ * The SCHEDOP interface provides mechanisms for a guest to interact
+ * with the scheduler, including yield, blocking and shutting itself
+ * down.
+ */
+
+/*
  * The prototype for this hypercall is:
- *  long sched_op_new(int cmd, void *arg)
+ * long HYPERVISOR_sched_op(enum sched_op cmd, void *arg, ...)
+ *
  * @cmd == SCHEDOP_??? (scheduler operation).
  * @arg == Operation-specific extra argument(s), as described below.
+ * ...  == Additional Operation-specific extra arguments, described below.
  *
- * **NOTE**:
- * Versions of Xen prior to 3.0.2 provide only the following legacy version
+ * Versions of Xen prior to 3.0.2 provided only the following legacy version
  * of this hypercall, supporting only the commands yield, block and shutdown:
  *  long sched_op(int cmd, unsigned long arg)
  * @cmd == SCHEDOP_??? (scheduler operation).
  * @arg == 0               (SCHEDOP_yield and SCHEDOP_block)
  *      == SHUTDOWN_* code (SCHEDOP_shutdown)
+ *
+ * This legacy version is available to new guests as:
+ * long HYPERVISOR_sched_op_compat(enum sched_op cmd, unsigned long arg)
  */
 
 /*
@@ -44,12 +74,17 @@
 /*
  * Halt execution of this domain (all VCPUs) and notify the system controller.
  * @arg == pointer to sched_shutdown structure.
+ *
+ * If the sched_shutdown_t reason is SHUTDOWN_suspend then
+ * x86 PV guests must also set RDX (EDX for 32-bit guests) to the MFN
+ * of the guest's start info page.  RDX/EDX is the third hypercall
+ * argument.
+ *
+ * In addition, which reason is SHUTDOWN_suspend this hypercall
+ * returns 1 if suspend was cancelled or the domain was merely
+ * checkpointed, and 0 if it is resuming in a new domain.
  */
 #define SCHEDOP_shutdown    2
-struct sched_shutdown {
-    unsigned int reason; /* SHUTDOWN_* */
-};
-DEFINE_GUEST_HANDLE_STRUCT(sched_shutdown);
 
 /*
  * Poll a set of event-channel ports. Return when one or more are pending. An
@@ -57,12 +92,6 @@ DEFINE_GUEST_HANDLE_STRUCT(sched_shutdown);
  * @arg == pointer to sched_poll structure.
  */
 #define SCHEDOP_poll        3
-struct sched_poll {
-    GUEST_HANDLE(evtchn_port_t) ports;
-    unsigned int nr_ports;
-    uint64_t timeout;
-};
-DEFINE_GUEST_HANDLE_STRUCT(sched_poll);
 
 /*
  * Declare a shutdown for another domain. The main use of this function is
@@ -71,15 +100,11 @@ DEFINE_GUEST_HANDLE_STRUCT(sched_poll);
  * @arg == pointer to sched_remote_shutdown structure.
  */
 #define SCHEDOP_remote_shutdown        4
-struct sched_remote_shutdown {
-    domid_t domain_id;         /* Remote domain ID */
-    unsigned int reason;       /* SHUTDOWN_xxx reason */
-};
 
 /*
  * Latch a shutdown code, so that when the domain later shuts down it
  * reports this code to the control tools.
- * @arg == as for SCHEDOP_shutdown.
+ * @arg == sched_shutdown, as for SCHEDOP_shutdown.
  */
 #define SCHEDOP_shutdown_code 5
 
@@ -92,10 +117,47 @@ struct sched_remote_shutdown {
  * With id != 0 and timeout != 0, poke watchdog timer and set new timeout.
  */
 #define SCHEDOP_watchdog    6
+
+/*
+ * Override the current vcpu affinity by pinning it to one physical cpu or
+ * undo this override restoring the previous affinity.
+ * @arg == pointer to sched_pin_override structure.
+ *
+ * A negative pcpu value will undo a previous pin override and restore the
+ * previous cpu affinity.
+ * This call is allowed for the hardware domain only and requires the cpu
+ * to be part of the domain's cpupool.
+ */
+#define SCHEDOP_pin_override 7
+
+struct sched_shutdown {
+    unsigned int reason; /* SHUTDOWN_* => shutdown reason */
+};
+DEFINE_GUEST_HANDLE_STRUCT(sched_shutdown);
+
+struct sched_poll {
+    GUEST_HANDLE(evtchn_port_t) ports;
+    unsigned int nr_ports;
+    uint64_t timeout;
+};
+DEFINE_GUEST_HANDLE_STRUCT(sched_poll);
+
+struct sched_remote_shutdown {
+    domid_t domain_id;         /* Remote domain ID */
+    unsigned int reason;       /* SHUTDOWN_* => shutdown reason */
+};
+DEFINE_GUEST_HANDLE_STRUCT(sched_remote_shutdown);
+
 struct sched_watchdog {
     uint32_t id;                /* watchdog ID */
     uint32_t timeout;           /* timeout */
 };
+DEFINE_GUEST_HANDLE_STRUCT(sched_watchdog);
+
+struct sched_pin_override {
+    int32_t pcpu;
+};
+DEFINE_GUEST_HANDLE_STRUCT(sched_pin_override);
 
 /*
  * Reason codes for SCHEDOP_shutdown. These may be interpreted by control
@@ -107,6 +169,7 @@ struct sched_watchdog {
 #define SHUTDOWN_suspend    2  /* Clean up, save suspend info, kill.         */
 #define SHUTDOWN_crash      3  /* Tell controller we've crashed.             */
 #define SHUTDOWN_watchdog   4  /* Restart because watchdog time expired.     */
+
 /*
  * Domain asked to perform 'soft reset' for it. The expected behavior is to
  * reset internal Xen state for the domain returning it to the point where it
@@ -115,5 +178,6 @@ struct sched_watchdog {
  * interfaces again.
  */
 #define SHUTDOWN_soft_reset 5
+#define SHUTDOWN_MAX        5  /* Maximum valid shutdown reason.             */
 
 #endif /* __XEN_PUBLIC_SCHED_H__ */
diff --git a/kernel/smp.c b/kernel/smp.c
index 3aa642d39c03..bba3b201668d 100644
--- a/kernel/smp.c
+++ b/kernel/smp.c
@@ -14,6 +14,7 @@
 #include <linux/smp.h>
 #include <linux/cpu.h>
 #include <linux/sched.h>
+#include <linux/hypervisor.h>
 
 #include "smpboot.h"
 
@@ -724,3 +725,54 @@ void wake_up_all_idle_cpus(void)
 	preempt_enable();
 }
 EXPORT_SYMBOL_GPL(wake_up_all_idle_cpus);
+
+/**
+ * smp_call_on_cpu - Call a function on a specific cpu
+ *
+ * Used to call a function on a specific cpu and wait for it to return.
+ * Optionally make sure the call is done on a specified physical cpu via vcpu
+ * pinning in order to support virtualized environments.
+ */
+struct smp_call_on_cpu_struct {
+	struct work_struct	work;
+	struct completion	done;
+	int			(*func)(void *);
+	void			*data;
+	int			ret;
+	int			cpu;
+};
+
+static void smp_call_on_cpu_callback(struct work_struct *work)
+{
+	struct smp_call_on_cpu_struct *sscs;
+
+	sscs = container_of(work, struct smp_call_on_cpu_struct, work);
+	if (sscs->cpu >= 0)
+		hypervisor_pin_vcpu(sscs->cpu);
+	sscs->ret = sscs->func(sscs->data);
+	if (sscs->cpu >= 0)
+		hypervisor_pin_vcpu(-1);
+
+	complete(&sscs->done);
+}
+
+int smp_call_on_cpu(unsigned int cpu, int (*func)(void *), void *par, bool phys)
+{
+	struct smp_call_on_cpu_struct sscs = {
+		.done = COMPLETION_INITIALIZER_ONSTACK(sscs.done),
+		.func = func,
+		.data = par,
+		.cpu  = phys ? cpu : -1,
+	};
+
+	INIT_WORK_ONSTACK(&sscs.work, smp_call_on_cpu_callback);
+
+	if (cpu >= nr_cpu_ids || !cpu_online(cpu))
+		return -ENXIO;
+
+	queue_work_on(cpu, system_wq, &sscs.work);
+	wait_for_completion(&sscs.done);
+
+	return sscs.ret;
+}
+EXPORT_SYMBOL_GPL(smp_call_on_cpu);
diff --git a/kernel/up.c b/kernel/up.c
index 1760bf3d1463..ee81ac9af4ca 100644
--- a/kernel/up.c
+++ b/kernel/up.c
@@ -6,6 +6,7 @@
 #include <linux/kernel.h>
 #include <linux/export.h>
 #include <linux/smp.h>
+#include <linux/hypervisor.h>
 
 int smp_call_function_single(int cpu, void (*func) (void *info), void *info,
 				int wait)
@@ -82,3 +83,20 @@ void on_each_cpu_cond(bool (*cond_func)(int cpu, void *info),
 	preempt_enable();
 }
 EXPORT_SYMBOL(on_each_cpu_cond);
+
+int smp_call_on_cpu(unsigned int cpu, int (*func)(void *), void *par, bool phys)
+{
+	int ret;
+
+	if (cpu != 0)
+		return -ENXIO;
+
+	if (phys)
+		hypervisor_pin_vcpu(0);
+	ret = func(par);
+	if (phys)
+		hypervisor_pin_vcpu(-1);
+
+	return ret;
+}
+EXPORT_SYMBOL_GPL(smp_call_on_cpu);