Merge tag 'powerpc-4.12-1' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux

Pull powerpc updates from Michael Ellerman:
 "Highlights include:

   - Larger virtual address space on 64-bit server CPUs. By default we
     use a 128TB virtual address space, but a process can request access
     to the full 512TB by passing a hint to mmap().

   - Support for the new Power9 "XIVE" interrupt controller.

   - TLB flushing optimisations for the radix MMU on Power9.

   - Support for CAPI cards on Power9, using the "Coherent Accelerator
     Interface Architecture 2.0".

   - The ability to configure the mmap randomisation limits at build and
     runtime.

   - Several small fixes and cleanups to the kprobes code, as well as
     support for KPROBES_ON_FTRACE.

   - Major improvements to handling of system reset interrupts,
     correctly treating them as NMIs, giving them a dedicated stack and
     using a new hypervisor call to trigger them, all of which should
     aid debugging and robustness.

   - Many fixes and other minor enhancements.

  Thanks to: Alastair D'Silva, Alexey Kardashevskiy, Alistair Popple,
  Andrew Donnellan, Aneesh Kumar K.V, Anshuman Khandual, Anton
  Blanchard, Balbir Singh, Ben Hutchings, Benjamin Herrenschmidt,
  Bhupesh Sharma, Chris Packham, Christian Zigotzky, Christophe Leroy,
  Christophe Lombard, Daniel Axtens, David Gibson, Gautham R. Shenoy,
  Gavin Shan, Geert Uytterhoeven, Guilherme G. Piccoli, Hamish Martin,
  Hari Bathini, Kees Cook, Laurent Dufour, Madhavan Srinivasan, Mahesh J
  Salgaonkar, Mahesh Salgaonkar, Masami Hiramatsu, Matt Brown, Matthew
  R. Ochs, Michael Neuling, Naveen N. Rao, Nicholas Piggin, Oliver
  O'Halloran, Pan Xinhui, Paul Mackerras, Rashmica Gupta, Russell
  Currey, Sukadev Bhattiprolu, Thadeu Lima de Souza Cascardo, Tobin C.
  Harding, Tyrel Datwyler, Uma Krishnan, Vaibhav Jain, Vipin K Parashar,
  Yang Shi"

* tag 'powerpc-4.12-1' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux: (214 commits)
  powerpc/64s: Power9 has no LPCR[VRMASD] field so don't set it
  powerpc/powernv: Fix TCE kill on NVLink2
  powerpc/mm/radix: Drop support for CPUs without lockless tlbie
  powerpc/book3s/mce: Move add_taint() later in virtual mode
  powerpc/sysfs: Move #ifdef CONFIG_HOTPLUG_CPU out of the function body
  powerpc/smp: Document irq enable/disable after migrating IRQs
  powerpc/mpc52xx: Don't select user-visible RTAS_PROC
  powerpc/powernv: Document cxl dependency on special case in pnv_eeh_reset()
  powerpc/eeh: Clean up and document event handling functions
  powerpc/eeh: Avoid use after free in eeh_handle_special_event()
  cxl: Mask slice error interrupts after first occurrence
  cxl: Route eeh events to all drivers in cxl_pci_error_detected()
  cxl: Force context lock during EEH flow
  powerpc/64: Allow CONFIG_RELOCATABLE if COMPILE_TEST
  powerpc/xmon: Teach xmon oops about radix vectors
  powerpc/mm/hash: Fix off-by-one in comment about kernel contexts ids
  powerpc/pseries: Enable VFIO
  powerpc/powernv: Fix iommu table size calculation hook for small tables
  powerpc/powernv: Check kzalloc() return value in pnv_pci_table_alloc
  powerpc: Add arch/powerpc/tools directory
  ...

1 files changed, 248 insertions, 51 deletions

diff --git a/arch/powerpc/kernel/smp.c b/arch/powerpc/kernel/smp.c
index d68ed1f004a306..df2a41647d8ed3 100644
--- a/arch/powerpc/kernel/smp.c
+++ b/arch/powerpc/kernel/smp.c
@@ -39,6 +39,7 @@
 #include <asm/irq.h>
 #include <asm/hw_irq.h>
 #include <asm/kvm_ppc.h>
+#include <asm/dbell.h>
 #include <asm/page.h>
 #include <asm/pgtable.h>
 #include <asm/prom.h>
@@ -86,8 +87,6 @@ volatile unsigned int cpu_callin_map[NR_CPUS];
 
 int smt_enabled_at_boot = 1;
 
-static void (*crash_ipi_function_ptr)(struct pt_regs *) = NULL;
-
 /*
  * Returns 1 if the specified cpu should be brought up during boot.
  * Used to inhibit booting threads if they've been disabled or
@@ -158,32 +157,33 @@ static irqreturn_t tick_broadcast_ipi_action(int irq, void *data)
 	return IRQ_HANDLED;
 }
 
-static irqreturn_t debug_ipi_action(int irq, void *data)
+#ifdef CONFIG_NMI_IPI
+static irqreturn_t nmi_ipi_action(int irq, void *data)
 {
-	if (crash_ipi_function_ptr) {
-		crash_ipi_function_ptr(get_irq_regs());
-		return IRQ_HANDLED;
-	}
-
-#ifdef CONFIG_DEBUGGER
-	debugger_ipi(get_irq_regs());
-#endif /* CONFIG_DEBUGGER */
-
+	smp_handle_nmi_ipi(get_irq_regs());
 	return IRQ_HANDLED;
 }
+#endif
 
 static irq_handler_t smp_ipi_action[] = {
 	[PPC_MSG_CALL_FUNCTION] =  call_function_action,
 	[PPC_MSG_RESCHEDULE] = reschedule_action,
 	[PPC_MSG_TICK_BROADCAST] = tick_broadcast_ipi_action,
-	[PPC_MSG_DEBUGGER_BREAK] = debug_ipi_action,
+#ifdef CONFIG_NMI_IPI
+	[PPC_MSG_NMI_IPI] = nmi_ipi_action,
+#endif
 };
 
+/*
+ * The NMI IPI is a fallback and not truly non-maskable. It is simpler
+ * than going through the call function infrastructure, and strongly
+ * serialized, so it is more appropriate for debugging.
+ */
 const char *smp_ipi_name[] = {
 	[PPC_MSG_CALL_FUNCTION] =  "ipi call function",
 	[PPC_MSG_RESCHEDULE] = "ipi reschedule",
 	[PPC_MSG_TICK_BROADCAST] = "ipi tick-broadcast",
-	[PPC_MSG_DEBUGGER_BREAK] = "ipi debugger",
+	[PPC_MSG_NMI_IPI] = "nmi ipi",
 };
 
 /* optional function to request ipi, for controllers with >= 4 ipis */
@@ -191,14 +191,13 @@ int smp_request_message_ipi(int virq, int msg)
 {
 	int err;
 
-	if (msg < 0 || msg > PPC_MSG_DEBUGGER_BREAK) {
+	if (msg < 0 || msg > PPC_MSG_NMI_IPI)
 		return -EINVAL;
-	}
-#if !defined(CONFIG_DEBUGGER) && !defined(CONFIG_KEXEC_CORE)
-	if (msg == PPC_MSG_DEBUGGER_BREAK) {
+#ifndef CONFIG_NMI_IPI
+	if (msg == PPC_MSG_NMI_IPI)
 		return 1;
-	}
 #endif
+
 	err = request_irq(virq, smp_ipi_action[msg],
 			  IRQF_PERCPU | IRQF_NO_THREAD | IRQF_NO_SUSPEND,
 			  smp_ipi_name[msg], NULL);
@@ -211,17 +210,9 @@ int smp_request_message_ipi(int virq, int msg)
 #ifdef CONFIG_PPC_SMP_MUXED_IPI
 struct cpu_messages {
 	long messages;			/* current messages */
-	unsigned long data;		/* data for cause ipi */
 };
 static DEFINE_PER_CPU_SHARED_ALIGNED(struct cpu_messages, ipi_message);
 
-void smp_muxed_ipi_set_data(int cpu, unsigned long data)
-{
-	struct cpu_messages *info = &per_cpu(ipi_message, cpu);
-
-	info->data = data;
-}
-
 void smp_muxed_ipi_set_message(int cpu, int msg)
 {
 	struct cpu_messages *info = &per_cpu(ipi_message, cpu);
@@ -236,14 +227,13 @@ void smp_muxed_ipi_set_message(int cpu, int msg)
 
 void smp_muxed_ipi_message_pass(int cpu, int msg)
 {
-	struct cpu_messages *info = &per_cpu(ipi_message, cpu);
-
 	smp_muxed_ipi_set_message(cpu, msg);
+
 	/*
 	 * cause_ipi functions are required to include a full barrier
 	 * before doing whatever causes the IPI.
 	 */
-	smp_ops->cause_ipi(cpu, info->data);
+	smp_ops->cause_ipi(cpu);
 }
 
 #ifdef __BIG_ENDIAN__
@@ -254,11 +244,18 @@ void smp_muxed_ipi_message_pass(int cpu, int msg)
 
 irqreturn_t smp_ipi_demux(void)
 {
-	struct cpu_messages *info = this_cpu_ptr(&ipi_message);
-	unsigned long all;
-
 	mb();	/* order any irq clear */
 
+	return smp_ipi_demux_relaxed();
+}
+
+/* sync-free variant. Callers should ensure synchronization */
+irqreturn_t smp_ipi_demux_relaxed(void)
+{
+	struct cpu_messages *info;
+	unsigned long all;
+
+	info = this_cpu_ptr(&ipi_message);
 	do {
 		all = xchg(&info->messages, 0);
 #if defined(CONFIG_KVM_XICS) && defined(CONFIG_KVM_BOOK3S_HV_POSSIBLE)
@@ -278,8 +275,10 @@ irqreturn_t smp_ipi_demux(void)
 			scheduler_ipi();
 		if (all & IPI_MESSAGE(PPC_MSG_TICK_BROADCAST))
 			tick_broadcast_ipi_handler();
-		if (all & IPI_MESSAGE(PPC_MSG_DEBUGGER_BREAK))
-			debug_ipi_action(0, NULL);
+#ifdef CONFIG_NMI_IPI
+		if (all & IPI_MESSAGE(PPC_MSG_NMI_IPI))
+			nmi_ipi_action(0, NULL);
+#endif
 	} while (info->messages);
 
 	return IRQ_HANDLED;
@@ -316,6 +315,187 @@ void arch_send_call_function_ipi_mask(const struct cpumask *mask)
 		do_message_pass(cpu, PPC_MSG_CALL_FUNCTION);
 }
 
+#ifdef CONFIG_NMI_IPI
+
+/*
+ * "NMI IPI" system.
+ *
+ * NMI IPIs may not be recoverable, so should not be used as ongoing part of
+ * a running system. They can be used for crash, debug, halt/reboot, etc.
+ *
+ * NMI IPIs are globally single threaded. No more than one in progress at
+ * any time.
+ *
+ * The IPI call waits with interrupts disabled until all targets enter the
+ * NMI handler, then the call returns.
+ *
+ * No new NMI can be initiated until targets exit the handler.
+ *
+ * The IPI call may time out without all targets entering the NMI handler.
+ * In that case, there is some logic to recover (and ignore subsequent
+ * NMI interrupts that may eventually be raised), but the platform interrupt
+ * handler may not be able to distinguish this from other exception causes,
+ * which may cause a crash.
+ */
+
+static atomic_t __nmi_ipi_lock = ATOMIC_INIT(0);
+static struct cpumask nmi_ipi_pending_mask;
+static int nmi_ipi_busy_count = 0;
+static void (*nmi_ipi_function)(struct pt_regs *) = NULL;
+
+static void nmi_ipi_lock_start(unsigned long *flags)
+{
+	raw_local_irq_save(*flags);
+	hard_irq_disable();
+	while (atomic_cmpxchg(&__nmi_ipi_lock, 0, 1) == 1) {
+		raw_local_irq_restore(*flags);
+		cpu_relax();
+		raw_local_irq_save(*flags);
+		hard_irq_disable();
+	}
+}
+
+static void nmi_ipi_lock(void)
+{
+	while (atomic_cmpxchg(&__nmi_ipi_lock, 0, 1) == 1)
+		cpu_relax();
+}
+
+static void nmi_ipi_unlock(void)
+{
+	smp_mb();
+	WARN_ON(atomic_read(&__nmi_ipi_lock) != 1);
+	atomic_set(&__nmi_ipi_lock, 0);
+}
+
+static void nmi_ipi_unlock_end(unsigned long *flags)
+{
+	nmi_ipi_unlock();
+	raw_local_irq_restore(*flags);
+}
+
+/*
+ * Platform NMI handler calls this to ack
+ */
+int smp_handle_nmi_ipi(struct pt_regs *regs)
+{
+	void (*fn)(struct pt_regs *);
+	unsigned long flags;
+	int me = raw_smp_processor_id();
+	int ret = 0;
+
+	/*
+	 * Unexpected NMIs are possible here because the interrupt may not
+	 * be able to distinguish NMI IPIs from other types of NMIs, or
+	 * because the caller may have timed out.
+	 */
+	nmi_ipi_lock_start(&flags);
+	if (!nmi_ipi_busy_count)
+		goto out;
+	if (!cpumask_test_cpu(me, &nmi_ipi_pending_mask))
+		goto out;
+
+	fn = nmi_ipi_function;
+	if (!fn)
+		goto out;
+
+	cpumask_clear_cpu(me, &nmi_ipi_pending_mask);
+	nmi_ipi_busy_count++;
+	nmi_ipi_unlock();
+
+	ret = 1;
+
+	fn(regs);
+
+	nmi_ipi_lock();
+	nmi_ipi_busy_count--;
+out:
+	nmi_ipi_unlock_end(&flags);
+
+	return ret;
+}
+
+static void do_smp_send_nmi_ipi(int cpu)
+{
+	if (smp_ops->cause_nmi_ipi && smp_ops->cause_nmi_ipi(cpu))
+		return;
+
+	if (cpu >= 0) {
+		do_message_pass(cpu, PPC_MSG_NMI_IPI);
+	} else {
+		int c;
+
+		for_each_online_cpu(c) {
+			if (c == raw_smp_processor_id())
+				continue;
+			do_message_pass(c, PPC_MSG_NMI_IPI);
+		}
+	}
+}
+
+/*
+ * - cpu is the target CPU (must not be this CPU), or NMI_IPI_ALL_OTHERS.
+ * - fn is the target callback function.
+ * - delay_us > 0 is the delay before giving up waiting for targets to
+ *   enter the handler, == 0 specifies indefinite delay.
+ */
+static int smp_send_nmi_ipi(int cpu, void (*fn)(struct pt_regs *), u64 delay_us)
+{
+	unsigned long flags;
+	int me = raw_smp_processor_id();
+	int ret = 1;
+
+	BUG_ON(cpu == me);
+	BUG_ON(cpu < 0 && cpu != NMI_IPI_ALL_OTHERS);
+
+	if (unlikely(!smp_ops))
+		return 0;
+
+	/* Take the nmi_ipi_busy count/lock with interrupts hard disabled */
+	nmi_ipi_lock_start(&flags);
+	while (nmi_ipi_busy_count) {
+		nmi_ipi_unlock_end(&flags);
+		cpu_relax();
+		nmi_ipi_lock_start(&flags);
+	}
+
+	nmi_ipi_function = fn;
+
+	if (cpu < 0) {
+		/* ALL_OTHERS */
+		cpumask_copy(&nmi_ipi_pending_mask, cpu_online_mask);
+		cpumask_clear_cpu(me, &nmi_ipi_pending_mask);
+	} else {
+		/* cpumask starts clear */
+		cpumask_set_cpu(cpu, &nmi_ipi_pending_mask);
+	}
+	nmi_ipi_busy_count++;
+	nmi_ipi_unlock();
+
+	do_smp_send_nmi_ipi(cpu);
+
+	while (!cpumask_empty(&nmi_ipi_pending_mask)) {
+		udelay(1);
+		if (delay_us) {
+			delay_us--;
+			if (!delay_us)
+				break;
+		}
+	}
+
+	nmi_ipi_lock();
+	if (!cpumask_empty(&nmi_ipi_pending_mask)) {
+		/* Could not gather all CPUs */
+		ret = 0;
+		cpumask_clear(&nmi_ipi_pending_mask);
+	}
+	nmi_ipi_busy_count--;
+	nmi_ipi_unlock_end(&flags);
+
+	return ret;
+}
+#endif /* CONFIG_NMI_IPI */
+
 #ifdef CONFIG_GENERIC_CLOCKEVENTS_BROADCAST
 void tick_broadcast(const struct cpumask *mask)
 {
@@ -326,29 +506,22 @@ void tick_broadcast(const struct cpumask *mask)
 }
 #endif
 
-#if defined(CONFIG_DEBUGGER) || defined(CONFIG_KEXEC_CORE)
-void smp_send_debugger_break(void)
+#ifdef CONFIG_DEBUGGER
+void debugger_ipi_callback(struct pt_regs *regs)
 {
-	int cpu;
-	int me = raw_smp_processor_id();
-
-	if (unlikely(!smp_ops))
-		return;
+	debugger_ipi(regs);
+}
 
-	for_each_online_cpu(cpu)
-		if (cpu != me)
-			do_message_pass(cpu, PPC_MSG_DEBUGGER_BREAK);
+void smp_send_debugger_break(void)
+{
+	smp_send_nmi_ipi(NMI_IPI_ALL_OTHERS, debugger_ipi_callback, 1000000);
 }
 #endif
 
 #ifdef CONFIG_KEXEC_CORE
 void crash_send_ipi(void (*crash_ipi_callback)(struct pt_regs *))
 {
-	crash_ipi_function_ptr = crash_ipi_callback;
-	if (crash_ipi_callback) {
-		mb();
-		smp_send_debugger_break();
-	}
+	smp_send_nmi_ipi(NMI_IPI_ALL_OTHERS, crash_ipi_callback, 1000000);
 }
 #endif
 
@@ -439,7 +612,21 @@ int generic_cpu_disable(void)
 #ifdef CONFIG_PPC64
 	vdso_data->processorCount--;
 #endif
-	migrate_irqs();
+	/* Update affinity of all IRQs previously aimed at this CPU */
+	irq_migrate_all_off_this_cpu();
+
+	/*
+	 * Depending on the details of the interrupt controller, it's possible
+	 * that one of the interrupts we just migrated away from this CPU is
+	 * actually already pending on this CPU. If we leave it in that state
+	 * the interrupt will never be EOI'ed, and will never fire again. So
+	 * temporarily enable interrupts here, to allow any pending interrupt to
+	 * be received (and EOI'ed), before we take this CPU offline.
+	 */
+	local_irq_enable();
+	mdelay(1);
+	local_irq_disable();
+
 	return 0;
 }
 
@@ -521,6 +708,16 @@ int __cpu_up(unsigned int cpu, struct task_struct *tidle)
 
 	cpu_idle_thread_init(cpu, tidle);
 
+	/*
+	 * The platform might need to allocate resources prior to bringing
+	 * up the CPU
+	 */
+	if (smp_ops->prepare_cpu) {
+		rc = smp_ops->prepare_cpu(cpu);
+		if (rc)
+			return rc;
+	}
+
 	/* Make sure callin-map entry is 0 (can be leftover a CPU
 	 * hotplug
 	 */