1 files changed, 212 insertions, 314 deletions

diff --git a/arch/s390/kernel/smp.c b/arch/s390/kernel/smp.c
index 30c91d565933..81f12bb77f62 100644
--- a/arch/s390/kernel/smp.c
+++ b/arch/s390/kernel/smp.c
@@ -18,6 +18,7 @@
 #define KMSG_COMPONENT "cpu"
 #define pr_fmt(fmt) KMSG_COMPONENT ": " fmt
 
+#include <linux/cpufeature.h>
 #include <linux/workqueue.h>
 #include <linux/memblock.h>
 #include <linux/export.h>
@@ -36,16 +37,20 @@
 #include <linux/sched/task_stack.h>
 #include <linux/crash_dump.h>
 #include <linux/kprobes.h>
+#include <asm/access-regs.h>
 #include <asm/asm-offsets.h>
+#include <asm/machine.h>
+#include <asm/ctlreg.h>
+#include <asm/pfault.h>
 #include <asm/diag.h>
-#include <asm/switch_to.h>
 #include <asm/facility.h>
+#include <asm/fpu.h>
 #include <asm/ipl.h>
 #include <asm/setup.h>
 #include <asm/irq.h>
 #include <asm/tlbflush.h>
 #include <asm/vtimer.h>
-#include <asm/lowcore.h>
+#include <asm/abs_lowcore.h>
 #include <asm/sclp.h>
 #include <asm/debug.h>
 #include <asm/os_info.h>
@@ -55,6 +60,7 @@
 #include <asm/stacktrace.h>
 #include <asm/topology.h>
 #include <asm/vdso.h>
+#include <asm/maccess.h>
 #include "entry.h"
 
 enum {
@@ -70,18 +76,15 @@ enum {
 	CPU_STATE_CONFIGURED,
 };
 
-static DEFINE_PER_CPU(struct cpu *, cpu_device);
-
-struct pcpu {
-	unsigned long ec_mask;		/* bit mask for ec_xxx functions */
-	unsigned long ec_clk;		/* sigp timestamp for ec_xxx */
-	signed char state;		/* physical cpu state */
-	signed char polarization;	/* physical polarization */
-	u16 address;			/* physical cpu address */
-};
-
 static u8 boot_core_type;
-static struct pcpu pcpu_devices[NR_CPUS];
+DEFINE_PER_CPU(struct pcpu, pcpu_devices);
+/*
+ * Pointer to the pcpu area of the boot CPU. This is required when a restart
+ * interrupt is triggered on an offline CPU. For that case accessing percpu
+ * data with the common primitives does not work, since the percpu offset is
+ * stored in a non existent lowcore.
+ */
+static struct pcpu *ipl_pcpu;
 
 unsigned int smp_cpu_mt_shift;
 EXPORT_SYMBOL(smp_cpu_mt_shift);
@@ -96,13 +99,6 @@ __vector128 __initdata boot_cpu_vector_save_area[__NUM_VXRS];
 static unsigned int smp_max_threads __initdata = -1U;
 cpumask_t cpu_setup_mask;
 
-static int __init early_nosmt(char *s)
-{
-	smp_max_threads = 1;
-	return 0;
-}
-early_param("nosmt", early_nosmt);
-
 static int __init early_smt(char *s)
 {
 	get_option(&s, &smp_max_threads);
@@ -112,7 +108,7 @@ early_param("smt", early_smt);
 
 /*
  * The smp_cpu_state_mutex must be held when changing the state or polarization
- * member of a pcpu data structure within the pcpu_devices arreay.
+ * member of a pcpu data structure within the pcpu_devices array.
  */
 DEFINE_MUTEX(smp_cpu_state_mutex);
 
@@ -172,8 +168,8 @@ static struct pcpu *pcpu_find_address(const struct cpumask *mask, u16 address)
 	int cpu;
 
 	for_each_cpu(cpu, mask)
-		if (pcpu_devices[cpu].address == address)
-			return pcpu_devices + cpu;
+		if (per_cpu(pcpu_devices, cpu).address == address)
+			return &per_cpu(pcpu_devices, cpu);
 	return NULL;
 }
 
@@ -199,7 +195,7 @@ static int pcpu_alloc_lowcore(struct pcpu *pcpu, int cpu)
 	mcck_stack = stack_alloc();
 	if (!lc || !nodat_stack || !async_stack || !mcck_stack)
 		goto out;
-	memcpy(lc, &S390_lowcore, 512);
+	memcpy(lc, get_lowcore(), 512);
 	memset((char *) lc + 512, 0, sizeof(*lc) - 512);
 	lc->async_stack = async_stack + STACK_INIT_OFFSET;
 	lc->nodat_stack = nodat_stack + STACK_INIT_OFFSET;
@@ -212,10 +208,14 @@ static int pcpu_alloc_lowcore(struct pcpu *pcpu, int cpu)
 	lc->preempt_count = PREEMPT_DISABLED;
 	if (nmi_alloc_mcesa(&lc->mcesad))
 		goto out;
+	if (abs_lowcore_map(cpu, lc, true))
+		goto out_mcesa;
 	lowcore_ptr[cpu] = lc;
 	pcpu_sigp_retry(pcpu, SIGP_SET_PREFIX, __pa(lc));
 	return 0;
 
+out_mcesa:
+	nmi_free_mcesa(&lc->mcesad);
 out:
 	stack_free(mcck_stack);
 	stack_free(async_stack);
@@ -224,19 +224,18 @@ out:
 	return -ENOMEM;
 }
 
-static void pcpu_free_lowcore(struct pcpu *pcpu)
+static void pcpu_free_lowcore(struct pcpu *pcpu, int cpu)
 {
 	unsigned long async_stack, nodat_stack, mcck_stack;
 	struct lowcore *lc;
-	int cpu;
 
-	cpu = pcpu - pcpu_devices;
 	lc = lowcore_ptr[cpu];
 	nodat_stack = lc->nodat_stack - STACK_INIT_OFFSET;
 	async_stack = lc->async_stack - STACK_INIT_OFFSET;
 	mcck_stack = lc->mcck_stack - STACK_INIT_OFFSET;
 	pcpu_sigp_retry(pcpu, SIGP_SET_PREFIX, 0);
 	lowcore_ptr[cpu] = NULL;
+	abs_lowcore_unmap(cpu);
 	nmi_free_mcesa(&lc->mcesad);
 	stack_free(async_stack);
 	stack_free(mcck_stack);
@@ -246,37 +245,37 @@ static void pcpu_free_lowcore(struct pcpu *pcpu)
 
 static void pcpu_prepare_secondary(struct pcpu *pcpu, int cpu)
 {
-	struct lowcore *lc = lowcore_ptr[cpu];
+	struct lowcore *lc, *abs_lc;
 
+	lc = lowcore_ptr[cpu];
 	cpumask_set_cpu(cpu, &init_mm.context.cpu_attach_mask);
 	cpumask_set_cpu(cpu, mm_cpumask(&init_mm));
 	lc->cpu_nr = cpu;
+	lc->pcpu = (unsigned long)pcpu;
 	lc->restart_flags = RESTART_FLAG_CTLREGS;
 	lc->spinlock_lockval = arch_spin_lockval(cpu);
 	lc->spinlock_index = 0;
 	lc->percpu_offset = __per_cpu_offset[cpu];
-	lc->kernel_asce = S390_lowcore.kernel_asce;
+	lc->kernel_asce = get_lowcore()->kernel_asce;
 	lc->user_asce = s390_invalid_asce;
-	lc->machine_flags = S390_lowcore.machine_flags;
 	lc->user_timer = lc->system_timer =
 		lc->steal_timer = lc->avg_steal_timer = 0;
-	__ctl_store(lc->cregs_save_area, 0, 15);
-	lc->cregs_save_area[1] = lc->kernel_asce;
+	abs_lc = get_abs_lowcore();
+	memcpy(lc->cregs_save_area, abs_lc->cregs_save_area, sizeof(lc->cregs_save_area));
+	put_abs_lowcore(abs_lc);
+	lc->cregs_save_area[1] = lc->user_asce;
 	lc->cregs_save_area[7] = lc->user_asce;
 	save_access_regs((unsigned int *) lc->access_regs_save_area);
 	arch_spin_lock_setup(cpu);
 }
 
-static void pcpu_attach_task(struct pcpu *pcpu, struct task_struct *tsk)
+static void pcpu_attach_task(int cpu, struct task_struct *tsk)
 {
 	struct lowcore *lc;
-	int cpu;
 
-	cpu = pcpu - pcpu_devices;
 	lc = lowcore_ptr[cpu];
-	lc->kernel_stack = (unsigned long) task_stack_page(tsk)
-		+ THREAD_SIZE - STACK_FRAME_OVERHEAD - sizeof(struct pt_regs);
-	lc->current_task = (unsigned long) tsk;
+	lc->kernel_stack = (unsigned long)task_stack_page(tsk) + STACK_INIT_OFFSET;
+	lc->current_task = (unsigned long)tsk;
 	lc->lpp = LPP_MAGIC;
 	lc->current_pid = tsk->pid;
 	lc->user_timer = tsk->thread.user_timer;
@@ -287,18 +286,16 @@ static void pcpu_attach_task(struct pcpu *pcpu, struct task_struct *tsk)
 	lc->steal_timer = 0;
 }
 
-static void pcpu_start_fn(struct pcpu *pcpu, void (*func)(void *), void *data)
+static void pcpu_start_fn(int cpu, void (*func)(void *), void *data)
 {
 	struct lowcore *lc;
-	int cpu;
 
-	cpu = pcpu - pcpu_devices;
 	lc = lowcore_ptr[cpu];
 	lc->restart_stack = lc->kernel_stack;
 	lc->restart_fn = (unsigned long) func;
 	lc->restart_data = (unsigned long) data;
 	lc->restart_source = -1U;
-	pcpu_sigp_retry(pcpu, SIGP_RESTART, 0);
+	pcpu_sigp_retry(per_cpu_ptr(&pcpu_devices, cpu), SIGP_RESTART, 0);
 }
 
 typedef void (pcpu_delegate_fn)(void *);
@@ -311,20 +308,23 @@ static void __pcpu_delegate(pcpu_delegate_fn *func, void *data)
 	func(data);	/* should not return */
 }
 
-static void pcpu_delegate(struct pcpu *pcpu,
+static void pcpu_delegate(struct pcpu *pcpu, int cpu,
 			  pcpu_delegate_fn *func,
 			  void *data, unsigned long stack)
 {
-	struct lowcore *lc = lowcore_ptr[pcpu - pcpu_devices];
-	unsigned int source_cpu = stap();
+	struct lowcore *lc, *abs_lc;
+	unsigned int source_cpu;
+
+	lc = lowcore_ptr[cpu];
+	source_cpu = stap();
 
-	__load_psw_mask(PSW_KERNEL_BITS | PSW_MASK_DAT);
 	if (pcpu->address == source_cpu) {
 		call_on_stack(2, stack, void, __pcpu_delegate,
 			      pcpu_delegate_fn *, func, void *, data);
 	}
 	/* Stop target cpu (if func returns this stops the current cpu). */
 	pcpu_sigp_retry(pcpu, SIGP_STOP, 0);
+	pcpu_sigp_retry(pcpu, SIGP_CPU_RESET, 0);
 	/* Restart func on the target cpu and stop the current cpu. */
 	if (lc) {
 		lc->restart_stack = stack;
@@ -332,12 +332,13 @@ static void pcpu_delegate(struct pcpu *pcpu,
 		lc->restart_data = (unsigned long)data;
 		lc->restart_source = source_cpu;
 	} else {
-		put_abs_lowcore(restart_stack, stack);
-		put_abs_lowcore(restart_fn, (unsigned long)func);
-		put_abs_lowcore(restart_data, (unsigned long)data);
-		put_abs_lowcore(restart_source, source_cpu);
+		abs_lc = get_abs_lowcore();
+		abs_lc->restart_stack = stack;
+		abs_lc->restart_fn = (unsigned long)func;
+		abs_lc->restart_data = (unsigned long)data;
+		abs_lc->restart_source = source_cpu;
+		put_abs_lowcore(abs_lc);
 	}
-	__bpon();
 	asm volatile(
 		"0:	sigp	0,%0,%2	# sigp restart to target cpu\n"
 		"	brc	2,0b	# busy, try again\n"
@@ -364,38 +365,22 @@ static int pcpu_set_smt(unsigned int mtid)
 		smp_cpu_mt_shift = 0;
 		while (smp_cpu_mtid >= (1U << smp_cpu_mt_shift))
 			smp_cpu_mt_shift++;
-		pcpu_devices[0].address = stap();
+		per_cpu(pcpu_devices, 0).address = stap();
 	}
 	return cc;
 }
 
 /*
- * Call function on an online CPU.
- */
-void smp_call_online_cpu(void (*func)(void *), void *data)
-{
-	struct pcpu *pcpu;
-
-	/* Use the current cpu if it is online. */
-	pcpu = pcpu_find_address(cpu_online_mask, stap());
-	if (!pcpu)
-		/* Use the first online cpu. */
-		pcpu = pcpu_devices + cpumask_first(cpu_online_mask);
-	pcpu_delegate(pcpu, func, data, (unsigned long) restart_stack);
-}
-
-/*
  * Call function on the ipl CPU.
  */
 void smp_call_ipl_cpu(void (*func)(void *), void *data)
 {
 	struct lowcore *lc = lowcore_ptr[0];
 
-	if (pcpu_devices[0].address == stap())
-		lc = &S390_lowcore;
+	if (ipl_pcpu->address == stap())
+		lc = get_lowcore();
 
-	pcpu_delegate(&pcpu_devices[0], func, data,
-		      lc->nodat_stack);
+	pcpu_delegate(ipl_pcpu, 0, func, data, lc->nodat_stack);
 }
 
 int smp_find_processor_id(u16 address)
@@ -403,21 +388,21 @@ int smp_find_processor_id(u16 address)
 	int cpu;
 
 	for_each_present_cpu(cpu)
-		if (pcpu_devices[cpu].address == address)
+		if (per_cpu(pcpu_devices, cpu).address == address)
 			return cpu;
 	return -1;
 }
 
 void schedule_mcck_handler(void)
 {
-	pcpu_ec_call(pcpu_devices + smp_processor_id(), ec_mcck_pending);
+	pcpu_ec_call(this_cpu_ptr(&pcpu_devices), ec_mcck_pending);
 }
 
 bool notrace arch_vcpu_is_preempted(int cpu)
 {
 	if (test_cpu_flag_of(CIF_ENABLED_WAIT, cpu))
 		return false;
-	if (pcpu_running(pcpu_devices + cpu))
+	if (pcpu_running(per_cpu_ptr(&pcpu_devices, cpu)))
 		return false;
 	return true;
 }
@@ -425,11 +410,11 @@ EXPORT_SYMBOL(arch_vcpu_is_preempted);
 
 void notrace smp_yield_cpu(int cpu)
 {
-	if (!MACHINE_HAS_DIAG9C)
+	if (!machine_has_diag9c())
 		return;
 	diag_stat_inc_norecursion(DIAG_STAT_X09C);
 	asm volatile("diag %0,0,0x9c"
-		     : : "d" (pcpu_devices[cpu].address));
+		     : : "d" (per_cpu(pcpu_devices, cpu).address));
 }
 EXPORT_SYMBOL_GPL(smp_yield_cpu);
 
@@ -450,7 +435,7 @@ void notrace smp_emergency_stop(void)
 
 	end = get_tod_clock() + (1000000UL << 12);
 	for_each_cpu(cpu, &cpumask) {
-		struct pcpu *pcpu = pcpu_devices + cpu;
+		struct pcpu *pcpu = per_cpu_ptr(&pcpu_devices, cpu);
 		set_bit(ec_stop_cpu, &pcpu->ec_mask);
 		while (__pcpu_sigp(pcpu->address, SIGP_EMERGENCY_SIGNAL,
 				   0, NULL) == SIGP_CC_BUSY &&
@@ -459,7 +444,7 @@ void notrace smp_emergency_stop(void)
 	}
 	while (get_tod_clock() < end) {
 		for_each_cpu(cpu, &cpumask)
-			if (pcpu_stopped(pcpu_devices + cpu))
+			if (pcpu_stopped(per_cpu_ptr(&pcpu_devices, cpu)))
 				cpumask_clear_cpu(cpu, &cpumask);
 		if (cpumask_empty(&cpumask))
 			break;
@@ -474,10 +459,11 @@ NOKPROBE_SYMBOL(smp_emergency_stop);
  */
 void smp_send_stop(void)
 {
+	struct pcpu *pcpu;
 	int cpu;
 
 	/* Disable all interrupts/machine checks */
-	__load_psw_mask(PSW_KERNEL_BITS | PSW_MASK_DAT);
+	__load_psw_mask(PSW_KERNEL_BITS);
 	trace_hardirqs_off();
 
 	debug_set_critical();
@@ -489,8 +475,9 @@ void smp_send_stop(void)
 	for_each_online_cpu(cpu) {
 		if (cpu == smp_processor_id())
 			continue;
-		pcpu_sigp_retry(pcpu_devices + cpu, SIGP_STOP, 0);
-		while (!pcpu_stopped(pcpu_devices + cpu))
+		pcpu = per_cpu_ptr(&pcpu_devices, cpu);
+		pcpu_sigp_retry(pcpu, SIGP_STOP, 0);
+		while (!pcpu_stopped(pcpu))
 			cpu_relax();
 	}
 }
@@ -504,7 +491,7 @@ static void smp_handle_ext_call(void)
 	unsigned long bits;
 
 	/* handle bit signal external calls */
-	bits = xchg(&pcpu_devices[smp_processor_id()].ec_mask, 0);
+	bits = this_cpu_xchg(pcpu_devices.ec_mask, 0);
 	if (test_bit(ec_stop_cpu, &bits))
 		smp_stop_cpu();
 	if (test_bit(ec_schedule, &bits))
@@ -512,7 +499,7 @@ static void smp_handle_ext_call(void)
 	if (test_bit(ec_call_function_single, &bits))
 		generic_smp_call_function_single_interrupt();
 	if (test_bit(ec_mcck_pending, &bits))
-		__s390_handle_mcck();
+		s390_handle_mcck();
 	if (test_bit(ec_irq_work, &bits))
 		irq_work_run();
 }
@@ -529,12 +516,12 @@ void arch_send_call_function_ipi_mask(const struct cpumask *mask)
 	int cpu;
 
 	for_each_cpu(cpu, mask)
-		pcpu_ec_call(pcpu_devices + cpu, ec_call_function_single);
+		pcpu_ec_call(per_cpu_ptr(&pcpu_devices, cpu), ec_call_function_single);
 }
 
 void arch_send_call_function_single_ipi(int cpu)
 {
-	pcpu_ec_call(pcpu_devices + cpu, ec_call_function_single);
+	pcpu_ec_call(per_cpu_ptr(&pcpu_devices, cpu), ec_call_function_single);
 }
 
 /*
@@ -542,63 +529,18 @@ void arch_send_call_function_single_ipi(int cpu)
  * it goes straight through and wastes no time serializing
  * anything. Worst case is that we lose a reschedule ...
  */
-void smp_send_reschedule(int cpu)
+void arch_smp_send_reschedule(int cpu)
 {
-	pcpu_ec_call(pcpu_devices + cpu, ec_schedule);
+	pcpu_ec_call(per_cpu_ptr(&pcpu_devices, cpu), ec_schedule);
 }
 
 #ifdef CONFIG_IRQ_WORK
 void arch_irq_work_raise(void)
 {
-	pcpu_ec_call(pcpu_devices + smp_processor_id(), ec_irq_work);
+	pcpu_ec_call(this_cpu_ptr(&pcpu_devices), ec_irq_work);
 }
 #endif
 
-/*
- * parameter area for the set/clear control bit callbacks
- */
-struct ec_creg_mask_parms {
-	unsigned long orval;
-	unsigned long andval;
-	int cr;
-};
-
-/*
- * callback for setting/clearing control bits
- */
-static void smp_ctl_bit_callback(void *info)
-{
-	struct ec_creg_mask_parms *pp = info;
-	unsigned long cregs[16];
-
-	__ctl_store(cregs, 0, 15);
-	cregs[pp->cr] = (cregs[pp->cr] & pp->andval) | pp->orval;
-	__ctl_load(cregs, 0, 15);
-}
-
-static DEFINE_SPINLOCK(ctl_lock);
-
-void smp_ctl_set_clear_bit(int cr, int bit, bool set)
-{
-	struct ec_creg_mask_parms parms = { .cr = cr, };
-	u64 ctlreg;
-
-	if (set) {
-		parms.orval = 1UL << bit;
-		parms.andval = -1UL;
-	} else {
-		parms.orval = 0;
-		parms.andval = ~(1UL << bit);
-	}
-	spin_lock(&ctl_lock);
-	get_abs_lowcore(ctlreg, cregs_save_area[cr]);
-	ctlreg = (ctlreg & parms.andval) | parms.orval;
-	put_abs_lowcore(cregs_save_area[cr], ctlreg);
-	spin_unlock(&ctl_lock);
-	on_each_cpu(smp_ctl_bit_callback, &parms, 1);
-}
-EXPORT_SYMBOL(smp_ctl_set_clear_bit);
-
 #ifdef CONFIG_CRASH_DUMP
 
 int smp_store_status(int cpu)
@@ -607,16 +549,16 @@ int smp_store_status(int cpu)
 	struct pcpu *pcpu;
 	unsigned long pa;
 
-	pcpu = pcpu_devices + cpu;
+	pcpu = per_cpu_ptr(&pcpu_devices, cpu);
 	lc = lowcore_ptr[cpu];
 	pa = __pa(&lc->floating_pt_save_area);
 	if (__pcpu_sigp_relax(pcpu->address, SIGP_STORE_STATUS_AT_ADDRESS,
 			      pa) != SIGP_CC_ORDER_CODE_ACCEPTED)
 		return -EIO;
-	if (!MACHINE_HAS_VX && !MACHINE_HAS_GS)
+	if (!cpu_has_vx() && !cpu_has_gs())
 		return 0;
 	pa = lc->mcesad & MCESA_ORIGIN_MASK;
-	if (MACHINE_HAS_GS)
+	if (cpu_has_gs())
 		pa |= lc->mcesad & MCESA_LC_MASK;
 	if (__pcpu_sigp_relax(pcpu->address, SIGP_STORE_ADDITIONAL_STATUS,
 			      pa) != SIGP_CC_ORDER_CODE_ACCEPTED)
@@ -626,7 +568,7 @@ int smp_store_status(int cpu)
 
 /*
  * Collect CPU state of the previous, crashed system.
- * There are four cases:
+ * There are three cases:
  * 1) standard zfcp/nvme dump
  *    condition: OLDMEM_BASE == NULL && is_ipl_type_dump() == true
  *    The state for all CPUs except the boot CPU needs to be collected
@@ -639,46 +581,45 @@ int smp_store_status(int cpu)
  *    with sigp stop-and-store-status. The firmware or the boot-loader
  *    stored the registers of the boot CPU in the absolute lowcore in the
  *    memory of the old system.
- * 3) kdump and the old kernel did not store the CPU state,
- *    or stand-alone kdump for DASD
- *    condition: OLDMEM_BASE != NULL && !is_kdump_kernel()
+ * 3) kdump or stand-alone kdump for DASD
+ *    condition: OLDMEM_BASE != NULL && is_ipl_type_dump() == false
  *    The state for all CPUs except the boot CPU needs to be collected
  *    with sigp stop-and-store-status. The kexec code or the boot-loader
  *    stored the registers of the boot CPU in the memory of the old system.
- * 4) kdump and the old kernel stored the CPU state
- *    condition: OLDMEM_BASE != NULL && is_kdump_kernel()
- *    This case does not exist for s390 anymore, setup_arch explicitly
- *    deactivates the elfcorehdr= kernel parameter
+ *
+ * Note that the legacy kdump mode where the old kernel stored the CPU states
+ * does no longer exist: setup_arch() explicitly deactivates the elfcorehdr=
+ * kernel parameter. The is_kdump_kernel() implementation on s390 is independent
+ * of the elfcorehdr= parameter.
  */
-static __init void smp_save_cpu_vxrs(struct save_area *sa, u16 addr,
-				     bool is_boot_cpu, __vector128 *vxrs)
+static bool dump_available(void)
 {
-	if (is_boot_cpu)
-		vxrs = boot_cpu_vector_save_area;
-	else
-		__pcpu_sigp_relax(addr, SIGP_STORE_ADDITIONAL_STATUS, __pa(vxrs));
-	save_area_add_vxrs(sa, vxrs);
+	return oldmem_data.start || is_ipl_type_dump();
 }
 
-static __init void smp_save_cpu_regs(struct save_area *sa, u16 addr,
-				     bool is_boot_cpu, void *regs)
+void __init smp_save_dump_ipl_cpu(void)
 {
-	if (is_boot_cpu)
-		copy_oldmem_kernel(regs, __LC_FPREGS_SAVE_AREA, 512);
-	else
-		__pcpu_sigp_relax(addr, SIGP_STORE_STATUS_AT_ADDRESS, __pa(regs));
+	struct save_area *sa;
+	void *regs;
+
+	if (!dump_available())
+		return;
+	sa = save_area_alloc(true);
+	regs = memblock_alloc_or_panic(512, 8);
+	copy_oldmem_kernel(regs, __LC_FPREGS_SAVE_AREA, 512);
 	save_area_add_regs(sa, regs);
+	memblock_free(regs, 512);
+	if (cpu_has_vx())
+		save_area_add_vxrs(sa, boot_cpu_vector_save_area);
 }
 
-void __init smp_save_dump_cpus(void)
+void __init smp_save_dump_secondary_cpus(void)
 {
 	int addr, boot_cpu_addr, max_cpu_addr;
 	struct save_area *sa;
-	bool is_boot_cpu;
 	void *page;
 
-	if (!(oldmem_data.start || is_ipl_type_dump()))
-		/* No previous system present, normal boot. */
+	if (!dump_available())
 		return;
 	/* Allocate a page as dumping area for the store status sigps */
 	page = memblock_alloc_low(PAGE_SIZE, PAGE_SIZE);
@@ -691,26 +632,18 @@ void __init smp_save_dump_cpus(void)
 	boot_cpu_addr = stap();
 	max_cpu_addr = SCLP_MAX_CORES << sclp.mtid_prev;
 	for (addr = 0; addr <= max_cpu_addr; addr++) {
+		if (addr == boot_cpu_addr)
+			continue;
 		if (__pcpu_sigp_relax(addr, SIGP_SENSE, 0) ==
 		    SIGP_CC_NOT_OPERATIONAL)
 			continue;
-		is_boot_cpu = (addr == boot_cpu_addr);
-		/* Allocate save area */
-		sa = save_area_alloc(is_boot_cpu);
-		if (!sa)
-			panic("could not allocate memory for save area\n");
-		if (MACHINE_HAS_VX)
-			/* Get the vector registers */
-			smp_save_cpu_vxrs(sa, addr, is_boot_cpu, page);
-		/*
-		 * For a zfcp/nvme dump OLDMEM_BASE == NULL and the registers
-		 * of the boot CPU are stored in the HSA. To retrieve
-		 * these registers an SCLP request is required which is
-		 * done by drivers/s390/char/zcore.c:init_cpu_info()
-		 */
-		if (!is_boot_cpu || oldmem_data.start)
-			/* Get the CPU registers */
-			smp_save_cpu_regs(sa, addr, is_boot_cpu, page);
+		sa = save_area_alloc(false);
+		__pcpu_sigp_relax(addr, SIGP_STORE_STATUS_AT_ADDRESS, __pa(page));
+		save_area_add_regs(sa, page);
+		if (cpu_has_vx()) {
+			__pcpu_sigp_relax(addr, SIGP_STORE_ADDITIONAL_STATUS, __pa(page));
+			save_area_add_vxrs(sa, page);
+		}
 	}
 	memblock_free(page, PAGE_SIZE);
 	diag_amode31_ops.diag308_reset();
@@ -720,17 +653,36 @@ void __init smp_save_dump_cpus(void)
 
 void smp_cpu_set_polarization(int cpu, int val)
 {
-	pcpu_devices[cpu].polarization = val;
+	per_cpu(pcpu_devices, cpu).polarization = val;
 }
 
 int smp_cpu_get_polarization(int cpu)
 {
-	return pcpu_devices[cpu].polarization;
+	return per_cpu(pcpu_devices, cpu).polarization;
+}
+
+void smp_cpu_set_capacity(int cpu, unsigned long val)
+{
+	per_cpu(pcpu_devices, cpu).capacity = val;
+}
+
+unsigned long smp_cpu_get_capacity(int cpu)
+{
+	return per_cpu(pcpu_devices, cpu).capacity;
+}
+
+void smp_set_core_capacity(int cpu, unsigned long val)
+{
+	int i;
+
+	cpu = smp_get_base_cpu(cpu);
+	for (i = cpu; (i <= cpu + smp_cpu_mtid) && (i < nr_cpu_ids); i++)
+		smp_cpu_set_capacity(i, val);
 }
 
 int smp_cpu_get_cpu_address(int cpu)
 {
-	return pcpu_devices[cpu].address;
+	return per_cpu(pcpu_devices, cpu).address;
 }
 
 static void __ref smp_get_core_info(struct sclp_core_info *info, int early)
@@ -754,8 +706,6 @@ static void __ref smp_get_core_info(struct sclp_core_info *info, int early)
 	}
 }
 
-static int smp_add_present_cpu(int cpu);
-
 static int smp_add_core(struct sclp_core_entry *core, cpumask_t *avail,
 			bool configured, bool early)
 {
@@ -771,15 +721,16 @@ static int smp_add_core(struct sclp_core_entry *core, cpumask_t *avail,
 	for (i = 0; (i <= smp_cpu_mtid) && (cpu < nr_cpu_ids); i++) {
 		if (pcpu_find_address(cpu_present_mask, address + i))
 			continue;
-		pcpu = pcpu_devices + cpu;
+		pcpu = per_cpu_ptr(&pcpu_devices, cpu);
 		pcpu->address = address + i;
 		if (configured)
 			pcpu->state = CPU_STATE_CONFIGURED;
 		else
 			pcpu->state = CPU_STATE_STANDBY;
 		smp_cpu_set_polarization(cpu, POLARIZATION_UNKNOWN);
+		smp_cpu_set_capacity(cpu, CPU_CAPACITY_HIGH);
 		set_cpu_present(cpu, true);
-		if (!early && smp_add_present_cpu(cpu) != 0)
+		if (!early && arch_register_cpu(cpu))
 			set_cpu_present(cpu, false);
 		else
 			nr++;
@@ -806,7 +757,7 @@ static int __smp_rescan_cpus(struct sclp_core_info *info, bool early)
 	 * that all SMT threads get subsequent logical CPU numbers.
 	 */
 	if (early) {
-		core_id = pcpu_devices[0].address >> smp_cpu_mt_shift;
+		core_id = per_cpu(pcpu_devices, 0).address >> smp_cpu_mt_shift;
 		for (i = 0; i < info->configured; i++) {
 			core = &info->core[i];
 			if (core->core_id == core_id) {
@@ -831,10 +782,7 @@ void __init smp_detect_cpus(void)
 	u16 address;
 
 	/* Get CPU information */
-	info = memblock_alloc(sizeof(*info), 8);
-	if (!info)
-		panic("%s: Failed to allocate %zu bytes align=0x%x\n",
-		      __func__, sizeof(*info), 8);
+	info = memblock_alloc_or_panic(sizeof(*info), 8);
 	smp_get_core_info(info, 1);
 	/* Find boot CPU type */
 	if (sclp.has_core_type) {
@@ -853,6 +801,7 @@ void __init smp_detect_cpus(void)
 	mtid = boot_core_type ? sclp.mtid : sclp.mtid_cp;
 	mtid = (mtid < smp_max_threads) ? mtid : smp_max_threads - 1;
 	pcpu_set_smt(mtid);
+	cpu_smt_set_num_threads(smp_cpu_mtid + 1, smp_cpu_mtid + 1);
 
 	/* Print number of CPUs */
 	c_cpus = s_cpus = 0;
@@ -866,9 +815,6 @@ void __init smp_detect_cpus(void)
 			s_cpus += smp_cpu_mtid + 1;
 	}
 	pr_info("%d configured CPUs, %d standby CPUs\n", c_cpus, s_cpus);
-
-	/* Add CPUs present at boot */
-	__smp_rescan_cpus(info, true);
 	memblock_free(info, sizeof(*info));
 }
 
@@ -877,17 +823,18 @@ void __init smp_detect_cpus(void)
  */
 static void smp_start_secondary(void *cpuvoid)
 {
+	struct lowcore *lc = get_lowcore();
 	int cpu = raw_smp_processor_id();
 
-	S390_lowcore.last_update_clock = get_tod_clock();
-	S390_lowcore.restart_stack = (unsigned long)restart_stack;
-	S390_lowcore.restart_fn = (unsigned long)do_restart;
-	S390_lowcore.restart_data = 0;
-	S390_lowcore.restart_source = -1U;
-	S390_lowcore.restart_flags = 0;
-	restore_access_regs(S390_lowcore.access_regs_save_area);
+	lc->last_update_clock = get_tod_clock();
+	lc->restart_stack = (unsigned long)restart_stack;
+	lc->restart_fn = (unsigned long)do_restart;
+	lc->restart_data = 0;
+	lc->restart_source = -1U;
+	lc->restart_flags = 0;
+	restore_access_regs(lc->access_regs_save_area);
 	cpu_init();
-	rcu_cpu_starting(cpu);
+	rcutree_report_cpu_starting(cpu);
 	init_cpu_timer();
 	vtime_init();
 	vdso_getcpu_init();
@@ -908,7 +855,7 @@ static void smp_start_secondary(void *cpuvoid)
 /* Upping and downing of CPUs */
 int __cpu_up(unsigned int cpu, struct task_struct *tidle)
 {
-	struct pcpu *pcpu = pcpu_devices + cpu;
+	struct pcpu *pcpu = per_cpu_ptr(&pcpu_devices, cpu);
 	int rc;
 
 	if (pcpu->state != CPU_STATE_CONFIGURED)
@@ -920,12 +867,18 @@ int __cpu_up(unsigned int cpu, struct task_struct *tidle)
 	rc = pcpu_alloc_lowcore(pcpu, cpu);
 	if (rc)
 		return rc;
+	/*
+	 * Make sure global control register contents do not change
+	 * until new CPU has initialized control registers.
+	 */
+	system_ctlreg_lock();
 	pcpu_prepare_secondary(pcpu, cpu);
-	pcpu_attach_task(pcpu, tidle);
-	pcpu_start_fn(pcpu, smp_start_secondary, NULL);
+	pcpu_attach_task(cpu, tidle);
+	pcpu_start_fn(cpu, smp_start_secondary, NULL);
 	/* Wait until cpu puts itself in the online & active maps */
 	while (!cpu_online(cpu))
 		cpu_relax();
+	system_ctlreg_unlock();
 	return 0;
 }
 
@@ -940,7 +893,7 @@ early_param("possible_cpus", _setup_possible_cpus);
 
 int __cpu_disable(void)
 {
-	unsigned long cregs[16];
+	struct ctlreg cregs[16];
 	int cpu;
 
 	/* Handle possible pending IPIs */
@@ -952,11 +905,11 @@ int __cpu_disable(void)
 	/* Disable pseudo page faults on this cpu. */
 	pfault_fini();
 	/* Disable interrupt sources via control register. */
-	__ctl_store(cregs, 0, 15);
-	cregs[0]  &= ~0x0000ee70UL;	/* disable all external interrupts */
-	cregs[6]  &= ~0xff000000UL;	/* disable all I/O interrupts */
-	cregs[14] &= ~0x1f000000UL;	/* disable most machine checks */
-	__ctl_load(cregs, 0, 15);
+	__local_ctl_store(0, 15, cregs);
+	cregs[0].val  &= ~0x0000ee70UL;	/* disable all external interrupts */
+	cregs[6].val  &= ~0xff000000UL;	/* disable all I/O interrupts */
+	cregs[14].val &= ~0x1f000000UL;	/* disable most machine checks */
+	__local_ctl_load(0, 15, cregs);
 	clear_cpu_flag(CIF_NOHZ_DELAY);
 	return 0;
 }
@@ -966,19 +919,19 @@ void __cpu_die(unsigned int cpu)
 	struct pcpu *pcpu;
 
 	/* Wait until target cpu is down */
-	pcpu = pcpu_devices + cpu;
+	pcpu = per_cpu_ptr(&pcpu_devices, cpu);
 	while (!pcpu_stopped(pcpu))
 		cpu_relax();
-	pcpu_free_lowcore(pcpu);
+	pcpu_free_lowcore(pcpu, cpu);
 	cpumask_clear_cpu(cpu, mm_cpumask(&init_mm));
 	cpumask_clear_cpu(cpu, &init_mm.context.cpu_attach_mask);
+	pcpu->flags = 0;
 }
 
 void __noreturn cpu_die(void)
 {
 	idle_task_exit();
-	__bpon();
-	pcpu_sigp_retry(pcpu_devices + smp_processor_id(), SIGP_STOP, 0);
+	pcpu_sigp_retry(this_cpu_ptr(&pcpu_devices), SIGP_STOP, 0);
 	for (;;) ;
 }
 
@@ -997,30 +950,36 @@ void __init smp_fill_possible_mask(void)
 
 void __init smp_prepare_cpus(unsigned int max_cpus)
 {
-	/* request the 0x1201 emergency signal external interrupt */
 	if (register_external_irq(EXT_IRQ_EMERGENCY_SIG, do_ext_call_interrupt))
 		panic("Couldn't request external interrupt 0x1201");
-	/* request the 0x1202 external call external interrupt */
+	system_ctl_set_bit(0, 14);
 	if (register_external_irq(EXT_IRQ_EXTERNAL_CALL, do_ext_call_interrupt))
 		panic("Couldn't request external interrupt 0x1202");
+	system_ctl_set_bit(0, 13);
+	smp_rescan_cpus(true);
 }
 
 void __init smp_prepare_boot_cpu(void)
 {
-	struct pcpu *pcpu = pcpu_devices;
+	struct lowcore *lc = get_lowcore();
 
 	WARN_ON(!cpu_present(0) || !cpu_online(0));
-	pcpu->state = CPU_STATE_CONFIGURED;
-	S390_lowcore.percpu_offset = __per_cpu_offset[0];
+	lc->percpu_offset = __per_cpu_offset[0];
+	ipl_pcpu = per_cpu_ptr(&pcpu_devices, 0);
+	ipl_pcpu->state = CPU_STATE_CONFIGURED;
+	lc->pcpu = (unsigned long)ipl_pcpu;
 	smp_cpu_set_polarization(0, POLARIZATION_UNKNOWN);
+	smp_cpu_set_capacity(0, CPU_CAPACITY_HIGH);
 }
 
 void __init smp_setup_processor_id(void)
 {
-	pcpu_devices[0].address = stap();
-	S390_lowcore.cpu_nr = 0;
-	S390_lowcore.spinlock_lockval = arch_spin_lockval(0);
-	S390_lowcore.spinlock_index = 0;
+	struct lowcore *lc = get_lowcore();
+
+	lc->cpu_nr = 0;
+	per_cpu(pcpu_devices, 0).address = stap();
+	lc->spinlock_lockval = arch_spin_lockval(0);
+	lc->spinlock_index = 0;
 }
 
 /*
@@ -1040,7 +999,7 @@ static ssize_t cpu_configure_show(struct device *dev,
 	ssize_t count;
 
 	mutex_lock(&smp_cpu_state_mutex);
-	count = sprintf(buf, "%d\n", pcpu_devices[dev->id].state);
+	count = sysfs_emit(buf, "%d\n", per_cpu(pcpu_devices, dev->id).state);
 	mutex_unlock(&smp_cpu_state_mutex);
 	return count;
 }
@@ -1060,15 +1019,13 @@ static ssize_t cpu_configure_store(struct device *dev,
 	cpus_read_lock();
 	mutex_lock(&smp_cpu_state_mutex);
 	rc = -EBUSY;
-	/* disallow configuration changes of online cpus and cpu 0 */
+	/* disallow configuration changes of online cpus */
 	cpu = dev->id;
 	cpu = smp_get_base_cpu(cpu);
-	if (cpu == 0)
-		goto out;
 	for (i = 0; i <= smp_cpu_mtid; i++)
 		if (cpu_online(cpu + i))
 			goto out;
-	pcpu = pcpu_devices + cpu;
+	pcpu = per_cpu_ptr(&pcpu_devices, cpu);
 	rc = 0;
 	switch (val) {
 	case 0:
@@ -1080,7 +1037,7 @@ static ssize_t cpu_configure_store(struct device *dev,
 		for (i = 0; i <= smp_cpu_mtid; i++) {
 			if (cpu + i >= nr_cpu_ids || !cpu_present(cpu + i))
 				continue;
-			pcpu[i].state = CPU_STATE_STANDBY;
+			per_cpu(pcpu_devices, cpu + i).state = CPU_STATE_STANDBY;
 			smp_cpu_set_polarization(cpu + i,
 						 POLARIZATION_UNKNOWN);
 		}
@@ -1095,7 +1052,7 @@ static ssize_t cpu_configure_store(struct device *dev,
 		for (i = 0; i <= smp_cpu_mtid; i++) {
 			if (cpu + i >= nr_cpu_ids || !cpu_present(cpu + i))
 				continue;
-			pcpu[i].state = CPU_STATE_CONFIGURED;
+			per_cpu(pcpu_devices, cpu + i).state = CPU_STATE_CONFIGURED;
 			smp_cpu_set_polarization(cpu + i,
 						 POLARIZATION_UNKNOWN);
 		}
@@ -1114,7 +1071,7 @@ static DEVICE_ATTR(configure, 0644, cpu_configure_show, cpu_configure_store);
 static ssize_t show_cpu_address(struct device *dev,
 				struct device_attribute *attr, char *buf)
 {
-	return sprintf(buf, "%d\n", pcpu_devices[dev->id].address);
+	return sysfs_emit(buf, "%d\n", per_cpu(pcpu_devices, dev->id).address);
 }
 static DEVICE_ATTR(address, 0444, show_cpu_address, NULL);
 
@@ -1140,35 +1097,34 @@ static struct attribute_group cpu_online_attr_group = {
 
 static int smp_cpu_online(unsigned int cpu)
 {
-	struct device *s = &per_cpu(cpu_device, cpu)->dev;
+	struct cpu *c = per_cpu_ptr(&cpu_devices, cpu);
 
-	return sysfs_create_group(&s->kobj, &cpu_online_attr_group);
+	return sysfs_create_group(&c->dev.kobj, &cpu_online_attr_group);
 }
 
 static int smp_cpu_pre_down(unsigned int cpu)
 {
-	struct device *s = &per_cpu(cpu_device, cpu)->dev;
+	struct cpu *c = per_cpu_ptr(&cpu_devices, cpu);
 
-	sysfs_remove_group(&s->kobj, &cpu_online_attr_group);
+	sysfs_remove_group(&c->dev.kobj, &cpu_online_attr_group);
 	return 0;
 }
 
-static int smp_add_present_cpu(int cpu)
+bool arch_cpu_is_hotpluggable(int cpu)
+{
+	return !!cpu;
+}
+
+int arch_register_cpu(int cpu)
 {
-	struct device *s;
-	struct cpu *c;
+	struct cpu *c = per_cpu_ptr(&cpu_devices, cpu);
 	int rc;
 
-	c = kzalloc(sizeof(*c), GFP_KERNEL);
-	if (!c)
-		return -ENOMEM;
-	per_cpu(cpu_device, cpu) = c;
-	s = &c->dev;
-	c->hotpluggable = 1;
+	c->hotpluggable = arch_cpu_is_hotpluggable(cpu);
 	rc = register_cpu(c, cpu);
 	if (rc)
 		goto out;
-	rc = sysfs_create_group(&s->kobj, &cpu_common_attr_group);
+	rc = sysfs_create_group(&c->dev.kobj, &cpu_common_attr_group);
 	if (rc)
 		goto out_cpu;
 	rc = topology_cpu_init(c);
@@ -1177,14 +1133,14 @@ static int smp_add_present_cpu(int cpu)
 	return 0;
 
 out_topology:
-	sysfs_remove_group(&s->kobj, &cpu_common_attr_group);
+	sysfs_remove_group(&c->dev.kobj, &cpu_common_attr_group);
 out_cpu:
 	unregister_cpu(c);
 out:
 	return rc;
 }
 
-int __ref smp_rescan_cpus(void)
+int __ref smp_rescan_cpus(bool early)
 {
 	struct sclp_core_info *info;
 	int nr;
@@ -1193,7 +1149,7 @@ int __ref smp_rescan_cpus(void)
 	if (!info)
 		return -ENOMEM;
 	smp_get_core_info(info, 0);
-	nr = __smp_rescan_cpus(info, false);
+	nr = __smp_rescan_cpus(info, early);
 	kfree(info);
 	if (nr)
 		topology_schedule_update();
@@ -1210,7 +1166,7 @@ static ssize_t __ref rescan_store(struct device *dev,
 	rc = lock_device_hotplug_sysfs();
 	if (rc)
 		return rc;
-	rc = smp_rescan_cpus();
+	rc = smp_rescan_cpus(false);
 	unlock_device_hotplug();
 	return rc ? rc : count;
 }
@@ -1218,77 +1174,19 @@ static DEVICE_ATTR_WO(rescan);
 
 static int __init s390_smp_init(void)
 {
-	int cpu, rc = 0;
+	struct device *dev_root;
+	int rc;
 
-	rc = device_create_file(cpu_subsys.dev_root, &dev_attr_rescan);
-	if (rc)
-		return rc;
-	for_each_present_cpu(cpu) {
-		rc = smp_add_present_cpu(cpu);
+	dev_root = bus_get_dev_root(&cpu_subsys);
+	if (dev_root) {
+		rc = device_create_file(dev_root, &dev_attr_rescan);
+		put_device(dev_root);
 		if (rc)
-			goto out;
+			return rc;
 	}
-
 	rc = cpuhp_setup_state(CPUHP_AP_ONLINE_DYN, "s390/smp:online",
 			       smp_cpu_online, smp_cpu_pre_down);
 	rc = rc <= 0 ? rc : 0;
-out:
 	return rc;
 }
 subsys_initcall(s390_smp_init);
-
-static __always_inline void set_new_lowcore(struct lowcore *lc)
-{
-	union register_pair dst, src;
-	u32 pfx;
-
-	src.even = (unsigned long) &S390_lowcore;
-	src.odd  = sizeof(S390_lowcore);
-	dst.even = (unsigned long) lc;
-	dst.odd  = sizeof(*lc);
-	pfx = __pa(lc);
-
-	asm volatile(
-		"	mvcl	%[dst],%[src]\n"
-		"	spx	%[pfx]\n"
-		: [dst] "+&d" (dst.pair), [src] "+&d" (src.pair)
-		: [pfx] "Q" (pfx)
-		: "memory", "cc");
-}
-
-static int __init smp_reinit_ipl_cpu(void)
-{
-	unsigned long async_stack, nodat_stack, mcck_stack;
-	struct lowcore *lc, *lc_ipl;
-	unsigned long flags, cr0;
-	u64 mcesad;
-
-	lc_ipl = lowcore_ptr[0];
-	lc = (struct lowcore *)	__get_free_pages(GFP_KERNEL | GFP_DMA, LC_ORDER);
-	nodat_stack = __get_free_pages(GFP_KERNEL, THREAD_SIZE_ORDER);
-	async_stack = stack_alloc();
-	mcck_stack = stack_alloc();
-	if (!lc || !nodat_stack || !async_stack || !mcck_stack || nmi_alloc_mcesa(&mcesad))
-		panic("Couldn't allocate memory");
-
-	local_irq_save(flags);
-	local_mcck_disable();
-	set_new_lowcore(lc);
-	S390_lowcore.nodat_stack = nodat_stack + STACK_INIT_OFFSET;
-	S390_lowcore.async_stack = async_stack + STACK_INIT_OFFSET;
-	S390_lowcore.mcck_stack = mcck_stack + STACK_INIT_OFFSET;
-	__ctl_store(cr0, 0, 0);
-	__ctl_clear_bit(0, 28); /* disable lowcore protection */
-	S390_lowcore.mcesad = mcesad;
-	__ctl_load(cr0, 0, 0);
-	lowcore_ptr[0] = lc;
-	local_mcck_enable();
-	local_irq_restore(flags);
-
-	free_pages(lc_ipl->async_stack - STACK_INIT_OFFSET, THREAD_SIZE_ORDER);
-	memblock_free_late(__pa(lc_ipl->mcck_stack - STACK_INIT_OFFSET), THREAD_SIZE);
-	memblock_free_late(__pa(lc_ipl), sizeof(*lc_ipl));
-
-	return 0;
-}
-early_initcall(smp_reinit_ipl_cpu);