Merge tag 's390-6.11-1' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux

Pull s390 updates from Vasily Gorbik:

 - Remove restrictions on PAI NNPA and crypto counters, enabling
   concurrent per-task and system-wide sampling and counting events

 - Switch to GENERIC_CPU_DEVICES by setting up the CPU present mask in
   the architecture code and letting the generic code handle CPU
   bring-up

 - Add support for the diag204 busy indication facility to prevent
   undesirable blocking during hypervisor logical CPU utilization
   queries. Implement results caching

 - Improve the handling of Store Data SCLP events by suppressing
   unnecessary warning, preventing buffer release in I/O during
   failures, and adding timeout handling for Store Data requests to
   address potential firmware issues

 - Provide optimized __arch_hweight*() implementations

 - Remove the unnecessary CPU KOBJ_CHANGE uevents generated during
   topology updates, as they are unused and also not present on other
   architectures

 - Cleanup atomic_ops, optimize __atomic_set() for small values and
   __atomic_cmpxchg_bool() for compilers supporting flag output
   constraint

 - Couple of cleanups for KVM:
     - Move and improve KVM struct definitions for DAT tables from
       gaccess.c to a new header
     - Pass the asce as parameter to sie64a()

 - Make the crdte() and cspg() page table handling wrappers return a
   boolean to indicate success, like the other existing "compare and
   swap" wrappers

 - Add documentation for HWCAP flags

 - Switch to obtaining total RAM pages from memblock instead of
   totalram_pages() during mm init, to ensure correct calculation of
   zero page size, when defer_init is enabled

 - Refactor lowcore access and switch to using the get_lowcore()
   function instead of the S390_lowcore macro

 - Cleanups for PG_arch_1 and folio handling in UV and hugetlb code

 - Add missing MODULE_DESCRIPTION() macros

 - Fix VM_FAULT_HWPOISON handling in do_exception()

* tag 's390-6.11-1' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux: (54 commits)
  s390/mm: Fix VM_FAULT_HWPOISON handling in do_exception()
  s390/kvm: Move bitfields for dat tables
  s390/entry: Pass the asce as parameter to sie64a()
  s390/sthyi: Use cached data when diag is busy
  s390/sthyi: Move diag operations
  s390/hypfs_diag: Diag204 busy loop
  s390/diag: Add busy-indication-facility requirements
  s390/diag: Diag204 add busy return errno
  s390/diag: Return errno's from diag204
  s390/sclp: Diag204 busy indication facility detection
  s390/atomic_ops: Make use of flag output constraint
  s390/atomic_ops: Improve __atomic_set() for small values
  s390/atomic_ops: Use symbolic names
  s390/smp: Switch to GENERIC_CPU_DEVICES
  s390/hwcaps: Add documentation for HWCAP flags
  s390/pgtable: Make crdte() and cspg() return a value
  s390/topology: Remove CPU KOBJ_CHANGE uevents
  s390/sclp: Add timeout to Store Data requests
  s390/sclp: Prevent release of buffer in I/O
  s390/sclp: Suppress unnecessary Store Data warning
  ...

22 files changed, 610 insertions, 414 deletions

diff --git a/arch/s390/kernel/asm-offsets.c b/arch/s390/kernel/asm-offsets.c
index f55979f64d49..26bb45d0e6f1 100644
--- a/arch/s390/kernel/asm-offsets.c
+++ b/arch/s390/kernel/asm-offsets.c
@@ -63,6 +63,7 @@ int main(void)
 	OFFSET(__SF_SIE_REASON, stack_frame, sie_reason);
 	OFFSET(__SF_SIE_FLAGS, stack_frame, sie_flags);
 	OFFSET(__SF_SIE_CONTROL_PHYS, stack_frame, sie_control_block_phys);
+	OFFSET(__SF_SIE_GUEST_ASCE, stack_frame, sie_guest_asce);
 	DEFINE(STACK_FRAME_OVERHEAD, sizeof(struct stack_frame));
 	BLANK();
 	OFFSET(__SFUSER_BACKCHAIN, stack_frame_user, back_chain);
diff --git a/arch/s390/kernel/diag.c b/arch/s390/kernel/diag.c
index 8dee9aa0ec95..9b65f04c83de 100644
--- a/arch/s390/kernel/diag.c
+++ b/arch/s390/kernel/diag.c
@@ -185,6 +185,8 @@ int diag14(unsigned long rx, unsigned long ry1, unsigned long subcode)
 }
 EXPORT_SYMBOL(diag14);
 
+#define DIAG204_BUSY_RC 8
+
 static inline int __diag204(unsigned long *subcode, unsigned long size, void *addr)
 {
 	union register_pair rp = { .even = *subcode, .odd = size };
@@ -215,16 +217,18 @@ int diag204(unsigned long subcode, unsigned long size, void *addr)
 {
 	if (addr) {
 		if (WARN_ON_ONCE(!is_vmalloc_addr(addr)))
-			return -1;
+			return -EINVAL;
 		if (WARN_ON_ONCE(!IS_ALIGNED((unsigned long)addr, PAGE_SIZE)))
-			return -1;
+			return -EINVAL;
 	}
 	if ((subcode & DIAG204_SUBCODE_MASK) == DIAG204_SUBC_STIB4)
 		addr = (void *)pfn_to_phys(vmalloc_to_pfn(addr));
 	diag_stat_inc(DIAG_STAT_X204);
 	size = __diag204(&subcode, size, addr);
-	if (subcode)
-		return -1;
+	if (subcode == DIAG204_BUSY_RC)
+		return -EBUSY;
+	else if (subcode)
+		return -EOPNOTSUPP;
 	return size;
 }
 EXPORT_SYMBOL(diag204);
diff --git a/arch/s390/kernel/dumpstack.c b/arch/s390/kernel/dumpstack.c
index d2012635b093..1ecd0580561f 100644
--- a/arch/s390/kernel/dumpstack.c
+++ b/arch/s390/kernel/dumpstack.c
@@ -61,28 +61,28 @@ static bool in_task_stack(unsigned long sp, struct task_struct *task,
 
 static bool in_irq_stack(unsigned long sp, struct stack_info *info)
 {
-	unsigned long stack = S390_lowcore.async_stack - STACK_INIT_OFFSET;
+	unsigned long stack = get_lowcore()->async_stack - STACK_INIT_OFFSET;
 
 	return in_stack(sp, info, STACK_TYPE_IRQ, stack);
 }
 
 static bool in_nodat_stack(unsigned long sp, struct stack_info *info)
 {
-	unsigned long stack = S390_lowcore.nodat_stack - STACK_INIT_OFFSET;
+	unsigned long stack = get_lowcore()->nodat_stack - STACK_INIT_OFFSET;
 
 	return in_stack(sp, info, STACK_TYPE_NODAT, stack);
 }
 
 static bool in_mcck_stack(unsigned long sp, struct stack_info *info)
 {
-	unsigned long stack = S390_lowcore.mcck_stack - STACK_INIT_OFFSET;
+	unsigned long stack = get_lowcore()->mcck_stack - STACK_INIT_OFFSET;
 
 	return in_stack(sp, info, STACK_TYPE_MCCK, stack);
 }
 
 static bool in_restart_stack(unsigned long sp, struct stack_info *info)
 {
-	unsigned long stack = S390_lowcore.restart_stack - STACK_INIT_OFFSET;
+	unsigned long stack = get_lowcore()->restart_stack - STACK_INIT_OFFSET;
 
 	return in_stack(sp, info, STACK_TYPE_RESTART, stack);
 }
diff --git a/arch/s390/kernel/early.c b/arch/s390/kernel/early.c
index c666271433fb..467ed4dba817 100644
--- a/arch/s390/kernel/early.c
+++ b/arch/s390/kernel/early.c
@@ -72,7 +72,7 @@ static void __init reset_tod_clock(void)
 
 	memset(&tod_clock_base, 0, sizeof(tod_clock_base));
 	tod_clock_base.tod = TOD_UNIX_EPOCH;
-	S390_lowcore.last_update_clock = TOD_UNIX_EPOCH;
+	get_lowcore()->last_update_clock = TOD_UNIX_EPOCH;
 }
 
 /*
@@ -99,7 +99,7 @@ static noinline __init void detect_machine_type(void)
 
 	/* Check current-configuration-level */
 	if (stsi(NULL, 0, 0, 0) <= 2) {
-		S390_lowcore.machine_flags |= MACHINE_FLAG_LPAR;
+		get_lowcore()->machine_flags |= MACHINE_FLAG_LPAR;
 		return;
 	}
 	/* Get virtual-machine cpu information. */
@@ -108,9 +108,9 @@ static noinline __init void detect_machine_type(void)
 
 	/* Detect known hypervisors */
 	if (!memcmp(vmms->vm[0].cpi, "\xd2\xe5\xd4", 3))
-		S390_lowcore.machine_flags |= MACHINE_FLAG_KVM;
+		get_lowcore()->machine_flags |= MACHINE_FLAG_KVM;
 	else if (!memcmp(vmms->vm[0].cpi, "\xa9\x61\xe5\xd4", 4))
-		S390_lowcore.machine_flags |= MACHINE_FLAG_VM;
+		get_lowcore()->machine_flags |= MACHINE_FLAG_VM;
 }
 
 /* Remove leading, trailing and double whitespace. */
@@ -166,7 +166,7 @@ static __init void setup_topology(void)
 
 	if (!test_facility(11))
 		return;
-	S390_lowcore.machine_flags |= MACHINE_FLAG_TOPOLOGY;
+	get_lowcore()->machine_flags |= MACHINE_FLAG_TOPOLOGY;
 	for (max_mnest = 6; max_mnest > 1; max_mnest--) {
 		if (stsi(&sysinfo_page, 15, 1, max_mnest) == 0)
 			break;
@@ -186,8 +186,8 @@ static noinline __init void setup_lowcore_early(void)
 
 	psw.addr = (unsigned long)early_pgm_check_handler;
 	psw.mask = PSW_KERNEL_BITS;
-	S390_lowcore.program_new_psw = psw;
-	S390_lowcore.preempt_count = INIT_PREEMPT_COUNT;
+	get_lowcore()->program_new_psw = psw;
+	get_lowcore()->preempt_count = INIT_PREEMPT_COUNT;
 }
 
 static noinline __init void setup_facility_list(void)
@@ -211,43 +211,43 @@ static __init void detect_diag9c(void)
 		EX_TABLE(0b,1b)
 		: "=d" (rc) : "0" (-EOPNOTSUPP), "d" (cpu_address) : "cc");
 	if (!rc)
-		S390_lowcore.machine_flags |= MACHINE_FLAG_DIAG9C;
+		get_lowcore()->machine_flags |= MACHINE_FLAG_DIAG9C;
 }
 
 static __init void detect_machine_facilities(void)
 {
 	if (test_facility(8)) {
-		S390_lowcore.machine_flags |= MACHINE_FLAG_EDAT1;
+		get_lowcore()->machine_flags |= MACHINE_FLAG_EDAT1;
 		system_ctl_set_bit(0, CR0_EDAT_BIT);
 	}
 	if (test_facility(78))
-		S390_lowcore.machine_flags |= MACHINE_FLAG_EDAT2;
+		get_lowcore()->machine_flags |= MACHINE_FLAG_EDAT2;
 	if (test_facility(3))
-		S390_lowcore.machine_flags |= MACHINE_FLAG_IDTE;
+		get_lowcore()->machine_flags |= MACHINE_FLAG_IDTE;
 	if (test_facility(50) && test_facility(73)) {
-		S390_lowcore.machine_flags |= MACHINE_FLAG_TE;
+		get_lowcore()->machine_flags |= MACHINE_FLAG_TE;
 		system_ctl_set_bit(0, CR0_TRANSACTIONAL_EXECUTION_BIT);
 	}
 	if (test_facility(51))
-		S390_lowcore.machine_flags |= MACHINE_FLAG_TLB_LC;
+		get_lowcore()->machine_flags |= MACHINE_FLAG_TLB_LC;
 	if (test_facility(129))
 		system_ctl_set_bit(0, CR0_VECTOR_BIT);
 	if (test_facility(130))
-		S390_lowcore.machine_flags |= MACHINE_FLAG_NX;
+		get_lowcore()->machine_flags |= MACHINE_FLAG_NX;
 	if (test_facility(133))
-		S390_lowcore.machine_flags |= MACHINE_FLAG_GS;
+		get_lowcore()->machine_flags |= MACHINE_FLAG_GS;
 	if (test_facility(139) && (tod_clock_base.tod >> 63)) {
 		/* Enabled signed clock comparator comparisons */
-		S390_lowcore.machine_flags |= MACHINE_FLAG_SCC;
+		get_lowcore()->machine_flags |= MACHINE_FLAG_SCC;
 		clock_comparator_max = -1ULL >> 1;
 		system_ctl_set_bit(0, CR0_CLOCK_COMPARATOR_SIGN_BIT);
 	}
 	if (IS_ENABLED(CONFIG_PCI) && test_facility(153)) {
-		S390_lowcore.machine_flags |= MACHINE_FLAG_PCI_MIO;
+		get_lowcore()->machine_flags |= MACHINE_FLAG_PCI_MIO;
 		/* the control bit is set during PCI initialization */
 	}
 	if (test_facility(194))
-		S390_lowcore.machine_flags |= MACHINE_FLAG_RDP;
+		get_lowcore()->machine_flags |= MACHINE_FLAG_RDP;
 }
 
 static inline void save_vector_registers(void)
diff --git a/arch/s390/kernel/entry.S b/arch/s390/kernel/entry.S
index 60cf917a7122..454b6b92c7f8 100644
--- a/arch/s390/kernel/entry.S
+++ b/arch/s390/kernel/entry.S
@@ -179,6 +179,7 @@ SYM_FUNC_END(__switch_to_asm)
  * %r2 pointer to sie control block phys
  * %r3 pointer to sie control block virt
  * %r4 guest register save area
+ * %r5 guest asce
  */
 SYM_FUNC_START(__sie64a)
 	stmg	%r6,%r14,__SF_GPRS(%r15)	# save kernel registers
@@ -186,15 +187,12 @@ SYM_FUNC_START(__sie64a)
 	stg	%r2,__SF_SIE_CONTROL_PHYS(%r15)	# save sie block physical..
 	stg	%r3,__SF_SIE_CONTROL(%r15)	# ...and virtual addresses
 	stg	%r4,__SF_SIE_SAVEAREA(%r15)	# save guest register save area
+	stg	%r5,__SF_SIE_GUEST_ASCE(%r15)	# save guest asce
 	xc	__SF_SIE_REASON(8,%r15),__SF_SIE_REASON(%r15) # reason code = 0
 	mvc	__SF_SIE_FLAGS(8,%r15),__TI_flags(%r12) # copy thread flags
 	lmg	%r0,%r13,0(%r4)			# load guest gprs 0-13
-	lg	%r14,__LC_GMAP			# get gmap pointer
-	ltgr	%r14,%r14
-	jz	.Lsie_gmap
 	oi	__LC_CPU_FLAGS+7,_CIF_SIE
-	lctlg	%c1,%c1,__GMAP_ASCE(%r14)	# load primary asce
-.Lsie_gmap:
+	lctlg	%c1,%c1,__SF_SIE_GUEST_ASCE(%r15) # load primary asce
 	lg	%r14,__SF_SIE_CONTROL(%r15)	# get control block pointer
 	oi	__SIE_PROG0C+3(%r14),1		# we are going into SIE now
 	tm	__SIE_PROG20+3(%r14),3		# last exit...
diff --git a/arch/s390/kernel/idle.c b/arch/s390/kernel/idle.c
index af9c97c0ad73..39cb8d0ae348 100644
--- a/arch/s390/kernel/idle.c
+++ b/arch/s390/kernel/idle.c
@@ -24,6 +24,7 @@ static DEFINE_PER_CPU(struct s390_idle_data, s390_idle);
 void account_idle_time_irq(void)
 {
 	struct s390_idle_data *idle = this_cpu_ptr(&s390_idle);
+	struct lowcore *lc = get_lowcore();
 	unsigned long idle_time;
 	u64 cycles_new[8];
 	int i;
@@ -34,13 +35,13 @@ void account_idle_time_irq(void)
 			this_cpu_add(mt_cycles[i], cycles_new[i] - idle->mt_cycles_enter[i]);
 	}
 
-	idle_time = S390_lowcore.int_clock - idle->clock_idle_enter;
+	idle_time = lc->int_clock - idle->clock_idle_enter;
 
-	S390_lowcore.steal_timer += idle->clock_idle_enter - S390_lowcore.last_update_clock;
-	S390_lowcore.last_update_clock = S390_lowcore.int_clock;
+	lc->steal_timer += idle->clock_idle_enter - lc->last_update_clock;
+	lc->last_update_clock = lc->int_clock;
 
-	S390_lowcore.system_timer += S390_lowcore.last_update_timer - idle->timer_idle_enter;
-	S390_lowcore.last_update_timer = S390_lowcore.sys_enter_timer;
+	lc->system_timer += lc->last_update_timer - idle->timer_idle_enter;
+	lc->last_update_timer = lc->sys_enter_timer;
 
 	/* Account time spent with enabled wait psw loaded as idle time. */
 	WRITE_ONCE(idle->idle_time, READ_ONCE(idle->idle_time) + idle_time);
diff --git a/arch/s390/kernel/irq.c b/arch/s390/kernel/irq.c
index 9acc6630abd3..1af5a08d72ab 100644
--- a/arch/s390/kernel/irq.c
+++ b/arch/s390/kernel/irq.c
@@ -100,8 +100,8 @@ static const struct irq_class irqclass_sub_desc[] = {
 
 static void do_IRQ(struct pt_regs *regs, int irq)
 {
-	if (tod_after_eq(S390_lowcore.int_clock,
-			 S390_lowcore.clock_comparator))
+	if (tod_after_eq(get_lowcore()->int_clock,
+			 get_lowcore()->clock_comparator))
 		/* Serve timer interrupts first. */
 		clock_comparator_work();
 	generic_handle_irq(irq);
@@ -111,7 +111,7 @@ static int on_async_stack(void)
 {
 	unsigned long frame = current_frame_address();
 
-	return ((S390_lowcore.async_stack ^ frame) & ~(THREAD_SIZE - 1)) == 0;
+	return ((get_lowcore()->async_stack ^ frame) & ~(THREAD_SIZE - 1)) == 0;
 }
 
 static void do_irq_async(struct pt_regs *regs, int irq)
@@ -119,7 +119,7 @@ static void do_irq_async(struct pt_regs *regs, int irq)
 	if (on_async_stack()) {
 		do_IRQ(regs, irq);
 	} else {
-		call_on_stack(2, S390_lowcore.async_stack, void, do_IRQ,
+		call_on_stack(2, get_lowcore()->async_stack, void, do_IRQ,
 			      struct pt_regs *, regs, int, irq);
 	}
 }
@@ -153,8 +153,8 @@ void noinstr do_io_irq(struct pt_regs *regs)
 
 	set_cpu_flag(CIF_NOHZ_DELAY);
 	do {
-		regs->tpi_info = S390_lowcore.tpi_info;
-		if (S390_lowcore.tpi_info.adapter_IO)
+		regs->tpi_info = get_lowcore()->tpi_info;
+		if (get_lowcore()->tpi_info.adapter_IO)
 			do_irq_async(regs, THIN_INTERRUPT);
 		else
 			do_irq_async(regs, IO_INTERRUPT);
@@ -183,9 +183,9 @@ void noinstr do_ext_irq(struct pt_regs *regs)
 			current->thread.last_break = regs->last_break;
 	}
 
-	regs->int_code = S390_lowcore.ext_int_code_addr;
-	regs->int_parm = S390_lowcore.ext_params;
-	regs->int_parm_long = S390_lowcore.ext_params2;
+	regs->int_code = get_lowcore()->ext_int_code_addr;
+	regs->int_parm = get_lowcore()->ext_params;
+	regs->int_parm_long = get_lowcore()->ext_params2;
 
 	from_idle = test_and_clear_cpu_flag(CIF_ENABLED_WAIT);
 	if (from_idle)
diff --git a/arch/s390/kernel/machine_kexec.c b/arch/s390/kernel/machine_kexec.c
index 3aee98efc374..f4cf65da6d49 100644
--- a/arch/s390/kernel/machine_kexec.c
+++ b/arch/s390/kernel/machine_kexec.c
@@ -52,7 +52,7 @@ static void __do_machine_kdump(void *data)
 	purgatory = (purgatory_t)image->start;
 
 	/* store_status() saved the prefix register to lowcore */
-	prefix = (unsigned long) S390_lowcore.prefixreg_save_area;
+	prefix = (unsigned long)get_lowcore()->prefixreg_save_area;
 
 	/* Now do the reset  */
 	s390_reset_system();
@@ -91,7 +91,7 @@ static noinline void __machine_kdump(void *image)
 			continue;
 	}
 	/* Store status of the boot CPU */
-	mcesa = __va(S390_lowcore.mcesad & MCESA_ORIGIN_MASK);
+	mcesa = __va(get_lowcore()->mcesad & MCESA_ORIGIN_MASK);
 	if (cpu_has_vx())
 		save_vx_regs((__vector128 *) mcesa->vector_save_area);
 	if (MACHINE_HAS_GS) {
diff --git a/arch/s390/kernel/nmi.c b/arch/s390/kernel/nmi.c
index 230d010bac9b..fbd218b6fc8e 100644
--- a/arch/s390/kernel/nmi.c
+++ b/arch/s390/kernel/nmi.c
@@ -117,6 +117,7 @@ static __always_inline char *u64_to_hex(char *dest, u64 val)
 
 static notrace void s390_handle_damage(void)
 {
+	struct lowcore *lc = get_lowcore();
 	union ctlreg0 cr0, cr0_new;
 	char message[100];
 	psw_t psw_save;
@@ -125,7 +126,7 @@ static notrace void s390_handle_damage(void)
 	smp_emergency_stop();
 	diag_amode31_ops.diag308_reset();
 	ptr = nmi_puts(message, "System stopped due to unrecoverable machine check, code: 0x");
-	u64_to_hex(ptr, S390_lowcore.mcck_interruption_code);
+	u64_to_hex(ptr, lc->mcck_interruption_code);
 
 	/*
 	 * Disable low address protection and make machine check new PSW a
@@ -135,17 +136,17 @@ static notrace void s390_handle_damage(void)
 	cr0_new = cr0;
 	cr0_new.lap = 0;
 	local_ctl_load(0, &cr0_new.reg);
-	psw_save = S390_lowcore.mcck_new_psw;
-	psw_bits(S390_lowcore.mcck_new_psw).io = 0;
-	psw_bits(S390_lowcore.mcck_new_psw).ext = 0;
-	psw_bits(S390_lowcore.mcck_new_psw).wait = 1;
+	psw_save = lc->mcck_new_psw;
+	psw_bits(lc->mcck_new_psw).io = 0;
+	psw_bits(lc->mcck_new_psw).ext = 0;
+	psw_bits(lc->mcck_new_psw).wait = 1;
 	sclp_emergency_printk(message);
 
 	/*
 	 * Restore machine check new PSW and control register 0 to original
 	 * values. This makes possible system dump analysis easier.
 	 */
-	S390_lowcore.mcck_new_psw = psw_save;
+	lc->mcck_new_psw = psw_save;
 	local_ctl_load(0, &cr0.reg);
 	disabled_wait();
 	while (1);
@@ -226,7 +227,7 @@ static bool notrace nmi_registers_valid(union mci mci)
 	/*
 	 * Set the clock comparator register to the next expected value.
 	 */
-	set_clock_comparator(S390_lowcore.clock_comparator);
+	set_clock_comparator(get_lowcore()->clock_comparator);
 	if (!mci.gr || !mci.fp || !mci.fc)
 		return false;
 	/*
@@ -252,7 +253,7 @@ static bool notrace nmi_registers_valid(union mci mci)
 	 * check handling must take care of this. The host values are saved by
 	 * KVM and are not affected.
 	 */
-	cr2.reg = S390_lowcore.cregs_save_area[2];
+	cr2.reg = get_lowcore()->cregs_save_area[2];
 	if (cr2.gse && !mci.gs && !test_cpu_flag(CIF_MCCK_GUEST))
 		return false;
 	if (!mci.ms || !mci.pm || !mci.ia)
@@ -278,11 +279,10 @@ static void notrace s390_backup_mcck_info(struct pt_regs *regs)
 
 	sie_page = container_of(sie_block, struct sie_page, sie_block);
 	mcck_backup = &sie_page->mcck_info;
-	mcck_backup->mcic = S390_lowcore.mcck_interruption_code &
+	mcck_backup->mcic = get_lowcore()->mcck_interruption_code &
 				~(MCCK_CODE_CP | MCCK_CODE_EXT_DAMAGE);
-	mcck_backup->ext_damage_code = S390_lowcore.external_damage_code;
-	mcck_backup->failing_storage_address
-			= S390_lowcore.failing_storage_address;
+	mcck_backup->ext_damage_code = get_lowcore()->external_damage_code;
+	mcck_backup->failing_storage_address = get_lowcore()->failing_storage_address;
 }
 NOKPROBE_SYMBOL(s390_backup_mcck_info);
 
@@ -302,6 +302,7 @@ void notrace s390_do_machine_check(struct pt_regs *regs)
 	static int ipd_count;
 	static DEFINE_SPINLOCK(ipd_lock);
 	static unsigned long long last_ipd;
+	struct lowcore *lc = get_lowcore();
 	struct mcck_struct *mcck;
 	unsigned long long tmp;
 	irqentry_state_t irq_state;
@@ -314,7 +315,7 @@ void notrace s390_do_machine_check(struct pt_regs *regs)
 	if (user_mode(regs))
 		update_timer_mcck();
 	inc_irq_stat(NMI_NMI);
-	mci.val = S390_lowcore.mcck_interruption_code;
+	mci.val = lc->mcck_interruption_code;
 	mcck = this_cpu_ptr(&cpu_mcck);
 
 	/*
@@ -382,9 +383,9 @@ void notrace s390_do_machine_check(struct pt_regs *regs)
 	}
 	if (mci.ed && mci.ec) {
 		/* External damage */
-		if (S390_lowcore.external_damage_code & (1U << ED_STP_SYNC))
+		if (lc->external_damage_code & (1U << ED_STP_SYNC))
 			mcck->stp_queue |= stp_sync_check();
-		if (S390_lowcore.external_damage_code & (1U << ED_STP_ISLAND))
+		if (lc->external_damage_code & (1U << ED_STP_ISLAND))
 			mcck->stp_queue |= stp_island_check();
 		mcck_pending = 1;
 	}
diff --git a/arch/s390/kernel/perf_cpum_sf.c b/arch/s390/kernel/perf_cpum_sf.c
index 06efad5b4f93..736c1d9632dd 100644
--- a/arch/s390/kernel/perf_cpum_sf.c
+++ b/arch/s390/kernel/perf_cpum_sf.c
@@ -1022,7 +1022,7 @@ static void cpumsf_pmu_enable(struct pmu *pmu)
 	}
 
 	/* Load current program parameter */
-	lpp(&S390_lowcore.lpp);
+	lpp(&get_lowcore()->lpp);
 
 	debug_sprintf_event(sfdbg, 6, "%s: es %i cs %i ed %i cd %i "
 			    "interval %#lx tear %#lx dear %#lx\n", __func__,
diff --git a/arch/s390/kernel/perf_pai_crypto.c b/arch/s390/kernel/perf_pai_crypto.c
index 4ad472d130a3..2f5a20e300f6 100644
--- a/arch/s390/kernel/perf_pai_crypto.c
+++ b/arch/s390/kernel/perf_pai_crypto.c
@@ -36,8 +36,8 @@ struct paicrypt_map {
 	struct pai_userdata *save;	/* Page to store no-zero counters */
 	unsigned int active_events;	/* # of PAI crypto users */
 	refcount_t refcnt;		/* Reference count mapped buffers */
-	enum paievt_mode mode;		/* Type of event */
 	struct perf_event *event;	/* Perf event for sampling */
+	struct list_head syswide_list;	/* List system-wide sampling events */
 };
 
 struct paicrypt_mapptr {
@@ -84,20 +84,16 @@ static DEFINE_MUTEX(pai_reserve_mutex);
 /* Adjust usage counters and remove allocated memory when all users are
  * gone.
  */
-static void paicrypt_event_destroy(struct perf_event *event)
+static void paicrypt_event_destroy_cpu(struct perf_event *event, int cpu)
 {
-	struct paicrypt_mapptr *mp = per_cpu_ptr(paicrypt_root.mapptr,
-						 event->cpu);
+	struct paicrypt_mapptr *mp = per_cpu_ptr(paicrypt_root.mapptr, cpu);
 	struct paicrypt_map *cpump = mp->mapptr;
 
-	static_branch_dec(&pai_key);
 	mutex_lock(&pai_reserve_mutex);
-	debug_sprintf_event(cfm_dbg, 5, "%s event %#llx cpu %d users %d"
-			    " mode %d refcnt %u\n", __func__,
-			    event->attr.config, event->cpu,
-			    cpump->active_events, cpump->mode,
+	debug_sprintf_event(cfm_dbg, 5, "%s event %#llx cpu %d users %d "
+			    "refcnt %u\n", __func__, event->attr.config,
+			    event->cpu, cpump->active_events,
 			    refcount_read(&cpump->refcnt));
-	free_page(PAI_SAVE_AREA(event));
 	if (refcount_dec_and_test(&cpump->refcnt)) {
 		debug_sprintf_event(cfm_dbg, 4, "%s page %#lx save %p\n",
 				    __func__, (unsigned long)cpump->page,
@@ -111,6 +107,23 @@ static void paicrypt_event_destroy(struct perf_event *event)
 	mutex_unlock(&pai_reserve_mutex);
 }
 
+static void paicrypt_event_destroy(struct perf_event *event)
+{
+	int cpu;
+
+	static_branch_dec(&pai_key);
+	free_page(PAI_SAVE_AREA(event));
+	if (event->cpu == -1) {
+		struct cpumask *mask = PAI_CPU_MASK(event);
+
+		for_each_cpu(cpu, mask)
+			paicrypt_event_destroy_cpu(event, cpu);
+		kfree(mask);
+	} else {
+		paicrypt_event_destroy_cpu(event, event->cpu);
+	}
+}
+
 static u64 paicrypt_getctr(unsigned long *page, int nr, bool kernel)
 {
 	if (kernel)
@@ -156,23 +169,15 @@ static u64 paicrypt_getall(struct perf_event *event)
 	return sum;
 }
 
-/* Used to avoid races in checking concurrent access of counting and
- * sampling for crypto events
- *
- * Only one instance of event pai_crypto/CRYPTO_ALL/ for sampling is
- * allowed and when this event is running, no counting event is allowed.
- * Several counting events are allowed in parallel, but no sampling event
- * is allowed while one (or more) counting events are running.
- *
+/* Check concurrent access of counting and sampling for crypto events.
  * This function is called in process context and it is save to block.
  * When the event initialization functions fails, no other call back will
  * be invoked.
  *
  * Allocate the memory for the event.
  */
-static struct paicrypt_map *paicrypt_busy(struct perf_event *event)
+static struct paicrypt_map *paicrypt_busy(struct perf_event *event, int cpu)
 {
-	struct perf_event_attr *a = &event->attr;
 	struct paicrypt_map *cpump = NULL;
 	struct paicrypt_mapptr *mp;
 	int rc;
@@ -185,7 +190,7 @@ static struct paicrypt_map *paicrypt_busy(struct perf_event *event)
 		goto unlock;
 
 	/* Allocate node for this event */
-	mp = per_cpu_ptr(paicrypt_root.mapptr, event->cpu);
+	mp = per_cpu_ptr(paicrypt_root.mapptr, cpu);
 	cpump = mp->mapptr;
 	if (!cpump) {			/* Paicrypt_map allocated? */
 		cpump = kzalloc(sizeof(*cpump), GFP_KERNEL);
@@ -193,25 +198,9 @@ static struct paicrypt_map *paicrypt_busy(struct perf_event *event)
 			rc = -ENOMEM;
 			goto free_root;
 		}
+		INIT_LIST_HEAD(&cpump->syswide_list);
 	}
 
-	if (a->sample_period) {		/* Sampling requested */
-		if (cpump->mode != PAI_MODE_NONE)
-			rc = -EBUSY;	/* ... sampling/counting active */
-	} else {			/* Counting requested */
-		if (cpump->mode == PAI_MODE_SAMPLING)
-			rc = -EBUSY;	/* ... and sampling active */
-	}
-	/*
-	 * This error case triggers when there is a conflict:
-	 * Either sampling requested and counting already active, or visa
-	 * versa. Therefore the struct paicrypto_map for this CPU is
-	 * needed or the error could not have occurred. Only adjust root
-	 * node refcount.
-	 */
-	if (rc)
-		goto free_root;
-
 	/* Allocate memory for counter page and counter extraction.
 	 * Only the first counting event has to allocate a page.
 	 */
@@ -235,26 +224,58 @@ static struct paicrypt_map *paicrypt_busy(struct perf_event *event)
 	/* Set mode and reference count */
 	rc = 0;
 	refcount_set(&cpump->refcnt, 1);
-	cpump->mode = a->sample_period ? PAI_MODE_SAMPLING : PAI_MODE_COUNTING;
 	mp->mapptr = cpump;
-	debug_sprintf_event(cfm_dbg, 5, "%s sample_period %#llx users %d"
-			    " mode %d refcnt %u page %#lx save %p rc %d\n",
-			    __func__, a->sample_period, cpump->active_events,
-			    cpump->mode, refcount_read(&cpump->refcnt),
+	debug_sprintf_event(cfm_dbg, 5, "%s users %d refcnt %u page %#lx "
+			    "save %p rc %d\n", __func__, cpump->active_events,
+			    refcount_read(&cpump->refcnt),
 			    (unsigned long)cpump->page, cpump->save, rc);
 	goto unlock;
 
 free_paicrypt_map:
+	/* Undo memory allocation */
 	kfree(cpump);
 	mp->mapptr = NULL;
 free_root:
 	paicrypt_root_free();
-
 unlock:
 	mutex_unlock(&pai_reserve_mutex);
 	return rc ? ERR_PTR(rc) : cpump;
 }
 
+static int paicrypt_event_init_all(struct perf_event *event)
+{
+	struct paicrypt_map *cpump;
+	struct cpumask *maskptr;
+	int cpu, rc = -ENOMEM;
+
+	maskptr = kzalloc(sizeof(*maskptr), GFP_KERNEL);
+	if (!maskptr)
+		goto out;
+
+	for_each_online_cpu(cpu) {
+		cpump = paicrypt_busy(event, cpu);
+		if (IS_ERR(cpump)) {
+			for_each_cpu(cpu, maskptr)
+				paicrypt_event_destroy_cpu(event, cpu);
+			kfree(maskptr);
+			rc = PTR_ERR(cpump);
+			goto out;
+		}
+		cpumask_set_cpu(cpu, maskptr);
+	}
+
+	/*
+	 * On error all cpumask are freed and all events have been destroyed.
+	 * Save of which CPUs data structures have been allocated for.
+	 * Release them in paicrypt_event_destroy call back function
+	 * for this event.
+	 */
+	PAI_CPU_MASK(event) = maskptr;
+	rc = 0;
+out:
+	return rc;
+}
+
 /* Might be called on different CPU than the one the event is intended for. */
 static int paicrypt_event_init(struct perf_event *event)
 {
@@ -269,10 +290,7 @@ static int paicrypt_event_init(struct perf_event *event)
 	if (a->config < PAI_CRYPTO_BASE ||
 	    a->config > PAI_CRYPTO_BASE + paicrypt_cnt)
 		return -EINVAL;
-	/* Allow only CPU wide operation, no process context for now. */
-	if ((event->attach_state & PERF_ATTACH_TASK) || event->cpu == -1)
-		return -ENOENT;
-	/* Allow only CRYPTO_ALL for sampling. */
+	/* Allow only CRYPTO_ALL for sampling */
 	if (a->sample_period && a->config != PAI_CRYPTO_BASE)
 		return -EINVAL;
 	/* Get a page to store last counter values for sampling */
@@ -284,13 +302,17 @@ static int paicrypt_event_init(struct perf_event *event)
 		}
 	}
 
-	cpump = paicrypt_busy(event);
-	if (IS_ERR(cpump)) {
+	if (event->cpu >= 0) {
+		cpump = paicrypt_busy(event, event->cpu);
+		if (IS_ERR(cpump))
+			rc = PTR_ERR(cpump);
+	} else {
+		rc = paicrypt_event_init_all(event);
+	}
+	if (rc) {
 		free_page(PAI_SAVE_AREA(event));
-		rc = PTR_ERR(cpump);
 		goto out;
 	}
-
 	event->destroy = paicrypt_event_destroy;
 
 	if (a->sample_period) {
@@ -331,8 +353,14 @@ static void paicrypt_start(struct perf_event *event, int flags)
 		sum = paicrypt_getall(event);	/* Get current value */
 		local64_set(&event->hw.prev_count, sum);
 	} else {				/* Sampling */
-		cpump->event = event;
-		perf_sched_cb_inc(event->pmu);
+		memcpy((void *)PAI_SAVE_AREA(event), cpump->page, PAGE_SIZE);
+		/* Enable context switch callback for system-wide sampling */
+		if (!(event->attach_state & PERF_ATTACH_TASK)) {
+			list_add_tail(PAI_SWLIST(event), &cpump->syswide_list);
+			perf_sched_cb_inc(event->pmu);
+		} else {
+			cpump->event = event;
+		}
 	}
 }
 
@@ -344,7 +372,7 @@ static int paicrypt_add(struct perf_event *event, int flags)
 
 	if (++cpump->active_events == 1) {
 		ccd = virt_to_phys(cpump->page) | PAI_CRYPTO_KERNEL_OFFSET;
-		WRITE_ONCE(S390_lowcore.ccd, ccd);
+		WRITE_ONCE(get_lowcore()->ccd, ccd);
 		local_ctl_set_bit(0, CR0_CRYPTOGRAPHY_COUNTER_BIT);
 	}
 	if (flags & PERF_EF_START)
@@ -353,6 +381,7 @@ static int paicrypt_add(struct perf_event *event, int flags)
 	return 0;
 }
 
+static void paicrypt_have_sample(struct perf_event *, struct paicrypt_map *);
 static void paicrypt_stop(struct perf_event *event, int flags)
 {
 	struct paicrypt_mapptr *mp = this_cpu_ptr(paicrypt_root.mapptr);
@@ -361,8 +390,13 @@ static void paicrypt_stop(struct perf_event *event, int flags)
 	if (!event->attr.sample_period) {	/* Counting */
 		paicrypt_read(event);
 	} else {				/* Sampling */
-		perf_sched_cb_dec(event->pmu);
-		cpump->event = NULL;
+		if (!(event->attach_state & PERF_ATTACH_TASK)) {
+			perf_sched_cb_dec(event->pmu);
+			list_del(PAI_SWLIST(event));
+		} else {
+			paicrypt_have_sample(event, cpump);
+			cpump->event = NULL;
+		}
 	}
 	event->hw.state = PERF_HES_STOPPED;
 }
@@ -375,7 +409,7 @@ static void paicrypt_del(struct perf_event *event, int flags)
 	paicrypt_stop(event, PERF_EF_UPDATE);
 	if (--cpump->active_events == 0) {
 		local_ctl_clear_bit(0, CR0_CRYPTOGRAPHY_COUNTER_BIT);
-		WRITE_ONCE(S390_lowcore.ccd, 0);
+		WRITE_ONCE(get_lowcore()->ccd, 0);
 	}
 }
 
@@ -455,23 +489,30 @@ static int paicrypt_push_sample(size_t rawsize, struct paicrypt_map *cpump,
 }
 
 /* Check if there is data to be saved on schedule out of a task. */
-static int paicrypt_have_sample(void)
+static void paicrypt_have_sample(struct perf_event *event,
+				 struct paicrypt_map *cpump)
 {
-	struct paicrypt_mapptr *mp = this_cpu_ptr(paicrypt_root.mapptr);
-	struct paicrypt_map *cpump = mp->mapptr;
-	struct perf_event *event = cpump->event;
 	size_t rawsize;
-	int rc = 0;
 
 	if (!event)		/* No event active */
-		return 0;
+		return;
 	rawsize = paicrypt_copy(cpump->save, cpump->page,
 				(unsigned long *)PAI_SAVE_AREA(event),
-				cpump->event->attr.exclude_user,
-				cpump->event->attr.exclude_kernel);
+				event->attr.exclude_user,
+				event->attr.exclude_kernel);
 	if (rawsize)			/* No incremented counters */
-		rc = paicrypt_push_sample(rawsize, cpump, event);
-	return rc;
+		paicrypt_push_sample(rawsize, cpump, event);
+}
+
+/* Check if there is data to be saved on schedule out of a task. */
+static void paicrypt_have_samples(void)
+{
+	struct paicrypt_mapptr *mp = this_cpu_ptr(paicrypt_root.mapptr);
+	struct paicrypt_map *cpump = mp->mapptr;
+	struct perf_event *event;
+
+	list_for_each_entry(event, &cpump->syswide_list, hw.tp_list)
+		paicrypt_have_sample(event, cpump);
 }
 
 /* Called on schedule-in and schedule-out. No access to event structure,
@@ -480,10 +521,10 @@ static int paicrypt_have_sample(void)
 static void paicrypt_sched_task(struct perf_event_pmu_context *pmu_ctx, bool sched_in)
 {
 	/* We started with a clean page on event installation. So read out
-	 * results on schedule_out and if page was dirty, clear values.
+	 * results on schedule_out and if page was dirty, save old values.
 	 */
 	if (!sched_in)
-		paicrypt_have_sample();
+		paicrypt_have_samples();
 }
 
 /* Attribute definitions for paicrypt interface. As with other CPU
@@ -527,7 +568,7 @@ static const struct attribute_group *paicrypt_attr_groups[] = {
 
 /* Performance monitoring unit for mapped counters */
 static struct pmu paicrypt = {
-	.task_ctx_nr  = perf_invalid_context,
+	.task_ctx_nr  = perf_hw_context,
 	.event_init   = paicrypt_event_init,
 	.add	      = paicrypt_add,
 	.del	      = paicrypt_del,
diff --git a/arch/s390/kernel/perf_pai_ext.c b/arch/s390/kernel/perf_pai_ext.c
index a6da7e0cc7a6..6295531b39a2 100644
--- a/arch/s390/kernel/perf_pai_ext.c
+++ b/arch/s390/kernel/perf_pai_ext.c
@@ -47,11 +47,11 @@ struct paiext_cb {		/* PAI extension 1 control block */
 struct paiext_map {
 	unsigned long *area;		/* Area for CPU to store counters */
 	struct pai_userdata *save;	/* Area to store non-zero counters */
-	enum paievt_mode mode;		/* Type of event */
 	unsigned int active_events;	/* # of PAI Extension users */
 	refcount_t refcnt;
 	struct perf_event *event;	/* Perf event for sampling */
 	struct paiext_cb *paiext_cb;	/* PAI extension control block area */
+	struct list_head syswide_list;	/* List system-wide sampling events */
 };
 
 struct paiext_mapptr {
@@ -70,6 +70,8 @@ static void paiext_root_free(void)
 		free_percpu(paiext_root.mapptr);
 		paiext_root.mapptr = NULL;
 	}
+	debug_sprintf_event(paiext_dbg, 5, "%s root.refcount %d\n", __func__,
+			    refcount_read(&paiext_root.refcnt));
 }
 
 /* On initialization of first event also allocate per CPU data dynamically.
@@ -115,20 +117,34 @@ static void paiext_free(struct paiext_mapptr *mp)
 }
 
 /* Release the PMU if event is the last perf event */
-static void paiext_event_destroy(struct perf_event *event)
+static void paiext_event_destroy_cpu(struct perf_event *event, int cpu)
 {
-	struct paiext_mapptr *mp = per_cpu_ptr(paiext_root.mapptr, event->cpu);
+	struct paiext_mapptr *mp = per_cpu_ptr(paiext_root.mapptr, cpu);
 	struct paiext_map *cpump = mp->mapptr;
 
-	free_page(PAI_SAVE_AREA(event));
 	mutex_lock(&paiext_reserve_mutex);
 	if (refcount_dec_and_test(&cpump->refcnt))	/* Last reference gone */
 		paiext_free(mp);
 	paiext_root_free();
 	mutex_unlock(&paiext_reserve_mutex);
-	debug_sprintf_event(paiext_dbg, 4, "%s cpu %d mapptr %p\n", __func__,
-			    event->cpu, mp->mapptr);
+}
+
+static void paiext_event_destroy(struct perf_event *event)
+{
+	int cpu;
+
+	free_page(PAI_SAVE_AREA(event));
+	if (event->cpu == -1) {
+		struct cpumask *mask = PAI_CPU_MASK(event);
 
+		for_each_cpu(cpu, mask)
+			paiext_event_destroy_cpu(event, cpu);
+		kfree(mask);
+	} else {
+		paiext_event_destroy_cpu(event, event->cpu);
+	}
+	debug_sprintf_event(paiext_dbg, 4, "%s cpu %d\n", __func__,
+			    event->cpu);
 }
 
 /* Used to avoid races in checking concurrent access of counting and
@@ -145,19 +161,18 @@ static void paiext_event_destroy(struct perf_event *event)
  *
  * Allocate the memory for the event.
  */
-static int paiext_alloc(struct perf_event_attr *a, struct perf_event *event)
+static int paiext_alloc_cpu(struct perf_event *event, int cpu)
 {
 	struct paiext_mapptr *mp;
 	struct paiext_map *cpump;
 	int rc;
 
 	mutex_lock(&paiext_reserve_mutex);
-
 	rc = paiext_root_alloc();
 	if (rc)
 		goto unlock;
 
-	mp = per_cpu_ptr(paiext_root.mapptr, event->cpu);
+	mp = per_cpu_ptr(paiext_root.mapptr, cpu);
 	cpump = mp->mapptr;
 	if (!cpump) {			/* Paiext_map allocated? */
 		rc = -ENOMEM;
@@ -185,24 +200,13 @@ static int paiext_alloc(struct perf_event_attr *a, struct perf_event *event)
 			paiext_free(mp);
 			goto undo;
 		}
+		INIT_LIST_HEAD(&cpump->syswide_list);
 		refcount_set(&cpump->refcnt, 1);
-		cpump->mode = a->sample_period ? PAI_MODE_SAMPLING
-					       : PAI_MODE_COUNTING;
+		rc = 0;
 	} else {
-		/* Multiple invocation, check what is active.
-		 * Supported are multiple counter events or only one sampling
-		 * event concurrently at any one time.
-		 */
-		if (cpump->mode == PAI_MODE_SAMPLING ||
-		    (cpump->mode == PAI_MODE_COUNTING && a->sample_period)) {
-			rc = -EBUSY;
-			goto undo;
-		}
 		refcount_inc(&cpump->refcnt);
 	}
 
-	rc = 0;
-
 undo:
 	if (rc) {
 		/* Error in allocation of event, decrement anchor. Since
@@ -217,6 +221,38 @@ unlock:
 	return rc;
 }
 
+static int paiext_alloc(struct perf_event *event)
+{
+	struct cpumask *maskptr;
+	int cpu, rc = -ENOMEM;
+
+	maskptr = kzalloc(sizeof(*maskptr), GFP_KERNEL);
+	if (!maskptr)
+		goto out;
+
+	for_each_online_cpu(cpu) {
+		rc = paiext_alloc_cpu(event, cpu);
+		if (rc) {
+			for_each_cpu(cpu, maskptr)
+				paiext_event_destroy_cpu(event, cpu);
+			kfree(maskptr);
+			goto out;
+		}
+		cpumask_set_cpu(cpu, maskptr);
+	}
+
+	/*
+	 * On error all cpumask are freed and all events have been destroyed.
+	 * Save of which CPUs data structures have been allocated for.
+	 * Release them in paicrypt_event_destroy call back function
+	 * for this event.
+	 */
+	PAI_CPU_MASK(event) = maskptr;
+	rc = 0;
+out:
+	return rc;
+}
+
 /* The PAI extension 1 control block supports up to 128 entries. Return
  * the index within PAIE1_CB given the event number. Also validate event
  * number.
@@ -246,9 +282,6 @@ static int paiext_event_init(struct perf_event *event)
 	rc = paiext_event_valid(event);
 	if (rc)
 		return rc;
-	/* Allow only CPU wide operation, no process context for now. */
-	if ((event->attach_state & PERF_ATTACH_TASK) || event->cpu == -1)
-		return -ENOENT;
 	/* Allow only event NNPA_ALL for sampling. */
 	if (a->sample_period && a->config != PAI_NNPA_BASE)
 		return -EINVAL;
@@ -262,7 +295,10 @@ static int paiext_event_init(struct perf_event *event)
 			return -ENOMEM;
 	}
 
-	rc = paiext_alloc(a, event);
+	if (event->cpu >= 0)
+		rc = paiext_alloc_cpu(event, event->cpu);
+	else
+		rc = paiext_alloc(event);
 	if (rc) {
 		free_page(PAI_SAVE_AREA(event));
 		return rc;
@@ -334,8 +370,15 @@ static void paiext_start(struct perf_event *event, int flags)
 		sum = paiext_getall(event);	/* Get current value */
 		local64_set(&event->hw.prev_count, sum);
 	} else {				/* Sampling */
-		cpump->event = event;
-		perf_sched_cb_inc(event->pmu);
+		memcpy((void *)PAI_SAVE_AREA(event), cpump->area,
+		       PAIE1_CTRBLOCK_SZ);
+		/* Enable context switch callback for system-wide sampling */
+		if (!(event->attach_state & PERF_ATTACH_TASK)) {
+			list_add_tail(PAI_SWLIST(event), &cpump->syswide_list);
+			perf_sched_cb_inc(event->pmu);
+		} else {
+			cpump->event = event;
+		}
 	}
 }
 
@@ -346,12 +389,10 @@ static int paiext_add(struct perf_event *event, int flags)
 	struct paiext_cb *pcb = cpump->paiext_cb;
 
 	if (++cpump->active_events == 1) {
-		S390_lowcore.aicd = virt_to_phys(cpump->paiext_cb);
+		get_lowcore()->aicd = virt_to_phys(cpump->paiext_cb);
 		pcb->acc = virt_to_phys(cpump->area) | 0x1;
 		/* Enable CPU instruction lookup for PAIE1 control block */
 		local_ctl_set_bit(0, CR0_PAI_EXTENSION_BIT);
-		debug_sprintf_event(paiext_dbg, 4, "%s 1508 %llx acc %llx\n",
-				    __func__, S390_lowcore.aicd, pcb->acc);
 	}
 	if (flags & PERF_EF_START)
 		paiext_start(event, PERF_EF_RELOAD);
@@ -359,6 +400,7 @@ static int paiext_add(struct perf_event *event, int flags)
 	return 0;
 }
 
+static void paiext_have_sample(struct perf_event *, struct paiext_map *);
 static void paiext_stop(struct perf_event *event, int flags)
 {
 	struct paiext_mapptr *mp = this_cpu_ptr(paiext_root.mapptr);
@@ -367,8 +409,13 @@ static void paiext_stop(struct perf_event *event, int flags)
 	if (!event->attr.sample_period) {	/* Counting */
 		paiext_read(event);
 	} else {				/* Sampling */
-		perf_sched_cb_dec(event->pmu);
-		cpump->event = NULL;
+		if (!(event->attach_state & PERF_ATTACH_TASK)) {
+			list_del(PAI_SWLIST(event));
+			perf_sched_cb_dec(event->pmu);
+		} else {
+			paiext_have_sample(event, cpump);
+			cpump->event = NULL;
+		}
 	}
 	event->hw.state = PERF_HES_STOPPED;
 }
@@ -384,9 +431,7 @@ static void paiext_del(struct perf_event *event, int flags)
 		/* Disable CPU instruction lookup for PAIE1 control block */
 		local_ctl_clear_bit(0, CR0_PAI_EXTENSION_BIT);
 		pcb->acc = 0;
-		S390_lowcore.aicd = 0;
-		debug_sprintf_event(paiext_dbg, 4, "%s 1508 %llx acc %llx\n",
-				    __func__, S390_lowcore.aicd, pcb->acc);
+		get_lowcore()->aicd = 0;
 	}
 }
 
@@ -470,21 +515,28 @@ static int paiext_push_sample(size_t rawsize, struct paiext_map *cpump,
 }
 
 /* Check if there is data to be saved on schedule out of a task. */
-static int paiext_have_sample(void)
+static void paiext_have_sample(struct perf_event *event,
+			       struct paiext_map *cpump)
 {
-	struct paiext_mapptr *mp = this_cpu_ptr(paiext_root.mapptr);
-	struct paiext_map *cpump = mp->mapptr;
-	struct perf_event *event = cpump->event;
 	size_t rawsize;
-	int rc = 0;
 
 	if (!event)
-		return 0;
+		return;
 	rawsize = paiext_copy(cpump->save, cpump->area,
 			      (unsigned long *)PAI_SAVE_AREA(event));
 	if (rawsize)			/* Incremented counters */
-		rc = paiext_push_sample(rawsize, cpump, event);
-	return rc;
+		paiext_push_sample(rawsize, cpump, event);
+}
+
+/* Check if there is data to be saved on schedule out of a task. */
+static void paiext_have_samples(void)
+{
+	struct paiext_mapptr *mp = this_cpu_ptr(paiext_root.mapptr);
+	struct paiext_map *cpump = mp->mapptr;
+	struct perf_event *event;
+
+	list_for_each_entry(event, &cpump->syswide_list, hw.tp_list)
+		paiext_have_sample(event, cpump);
 }
 
 /* Called on schedule-in and schedule-out. No access to event structure,
@@ -493,10 +545,10 @@ static int paiext_have_sample(void)
 static void paiext_sched_task(struct perf_event_pmu_context *pmu_ctx, bool sched_in)
 {
 	/* We started with a clean page on event installation. So read out
-	 * results on schedule_out and if page was dirty, clear values.
+	 * results on schedule_out and if page was dirty, save old values.
 	 */
 	if (!sched_in)
-		paiext_have_sample();
+		paiext_have_samples();
 }
 
 /* Attribute definitions for pai extension1 interface. As with other CPU
@@ -542,7 +594,7 @@ static const struct attribute_group *paiext_attr_groups[] = {
 
 /* Performance monitoring unit for mapped counters */
 static struct pmu paiext = {
-	.task_ctx_nr  = perf_invalid_context,
+	.task_ctx_nr  = perf_hw_context,
 	.event_init   = paiext_event_init,
 	.add	      = paiext_add,
 	.del	      = paiext_del,
diff --git a/arch/s390/kernel/process.c b/arch/s390/kernel/process.c
index d8740631df4b..9637aee43c40 100644
--- a/arch/s390/kernel/process.c
+++ b/arch/s390/kernel/process.c
@@ -71,10 +71,10 @@ void flush_thread(void)
 
 void arch_setup_new_exec(void)
 {
-	if (S390_lowcore.current_pid != current->pid) {
-		S390_lowcore.current_pid = current->pid;
+	if (get_lowcore()->current_pid != current->pid) {
+		get_lowcore()->current_pid = current->pid;
 		if (test_facility(40))
-			lpp(&S390_lowcore.lpp);
+			lpp(&get_lowcore()->lpp);
 	}
 }
 
diff --git a/arch/s390/kernel/setup.c b/arch/s390/kernel/setup.c
index 90c2c786bb35..3993f4caf224 100644
--- a/arch/s390/kernel/setup.c
+++ b/arch/s390/kernel/setup.c
@@ -421,16 +421,16 @@ static void __init setup_lowcore(void)
 	lc->clock_comparator = clock_comparator_max;
 	lc->current_task = (unsigned long)&init_task;
 	lc->lpp = LPP_MAGIC;
-	lc->machine_flags = S390_lowcore.machine_flags;
-	lc->preempt_count = S390_lowcore.preempt_count;
+	lc->machine_flags = get_lowcore()->machine_flags;
+	lc->preempt_count = get_lowcore()->preempt_count;
 	nmi_alloc_mcesa_early(&lc->mcesad);
-	lc->sys_enter_timer = S390_lowcore.sys_enter_timer;
-	lc->exit_timer = S390_lowcore.exit_timer;
-	lc->user_timer = S390_lowcore.user_timer;
-	lc->system_timer = S390_lowcore.system_timer;
-	lc->steal_timer = S390_lowcore.steal_timer;
-	lc->last_update_timer = S390_lowcore.last_update_timer;
-	lc->last_update_clock = S390_lowcore.last_update_clock;
+	lc->sys_enter_timer = get_lowcore()->sys_enter_timer;
+	lc->exit_timer = get_lowcore()->exit_timer;
+	lc->user_timer = get_lowcore()->user_timer;
+	lc->system_timer = get_lowcore()->system_timer;
+	lc->steal_timer = get_lowcore()->steal_timer;
+	lc->last_update_timer = get_lowcore()->last_update_timer;
+	lc->last_update_clock = get_lowcore()->last_update_clock;
 	/*
 	 * Allocate the global restart stack which is the same for
 	 * all CPUs in case *one* of them does a PSW restart.
@@ -439,7 +439,7 @@ static void __init setup_lowcore(void)
 	lc->mcck_stack = stack_alloc_early() + STACK_INIT_OFFSET;
 	lc->async_stack = stack_alloc_early() + STACK_INIT_OFFSET;
 	lc->nodat_stack = stack_alloc_early() + STACK_INIT_OFFSET;
-	lc->kernel_stack = S390_lowcore.kernel_stack;
+	lc->kernel_stack = get_lowcore()->kernel_stack;
 	/*
 	 * Set up PSW restart to call ipl.c:do_restart(). Copy the relevant
 	 * restart data to the absolute zero lowcore. This is necessary if
@@ -455,8 +455,8 @@ static void __init setup_lowcore(void)
 	lc->return_lpswe = gen_lpswe(__LC_RETURN_PSW);
 	lc->return_mcck_lpswe = gen_lpswe(__LC_RETURN_MCCK_PSW);
 	lc->preempt_count = PREEMPT_DISABLED;
-	lc->kernel_asce = S390_lowcore.kernel_asce;
-	lc->user_asce = S390_lowcore.user_asce;
+	lc->kernel_asce = get_lowcore()->kernel_asce;
+	lc->user_asce = get_lowcore()->user_asce;
 
 	system_ctlreg_init_save_area(lc);
 	abs_lc = get_abs_lowcore();
diff --git a/arch/s390/kernel/smp.c b/arch/s390/kernel/smp.c
index 0324649aae0a..c3c54adf67bc 100644
--- a/arch/s390/kernel/smp.c
+++ b/arch/s390/kernel/smp.c
@@ -74,8 +74,6 @@ 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 */
@@ -203,7 +201,7 @@ 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;
@@ -265,9 +263,9 @@ static void pcpu_prepare_secondary(struct pcpu *pcpu, int cpu)
 	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->machine_flags = get_lowcore()->machine_flags;
 	lc->user_timer = lc->system_timer =
 		lc->steal_timer = lc->avg_steal_timer = 0;
 	abs_lc = get_abs_lowcore();
@@ -407,7 +405,7 @@ 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;
+		lc = get_lowcore();
 
 	pcpu_delegate(&pcpu_devices[0], func, data,
 		      lc->nodat_stack);
@@ -719,8 +717,6 @@ 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)
 {
@@ -744,7 +740,7 @@ static int smp_add_core(struct sclp_core_entry *core, cpumask_t *avail,
 			pcpu->state = CPU_STATE_STANDBY;
 		smp_cpu_set_polarization(cpu, POLARIZATION_UNKNOWN);
 		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++;
@@ -831,9 +827,6 @@ 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));
 }
 
@@ -842,15 +835,16 @@ 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();
 	rcutree_report_cpu_starting(cpu);
 	init_cpu_timer();
@@ -973,6 +967,7 @@ void __init smp_prepare_cpus(unsigned int max_cpus)
 	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)
@@ -981,16 +976,18 @@ void __init smp_prepare_boot_cpu(void)
 
 	WARN_ON(!cpu_present(0) || !cpu_online(0));
 	pcpu->state = CPU_STATE_CONFIGURED;
-	S390_lowcore.percpu_offset = __per_cpu_offset[0];
+	get_lowcore()->percpu_offset = __per_cpu_offset[0];
 	smp_cpu_set_polarization(0, POLARIZATION_UNKNOWN);
 }
 
 void __init smp_setup_processor_id(void)
 {
+	struct lowcore *lc = get_lowcore();
+
 	pcpu_devices[0].address = stap();
-	S390_lowcore.cpu_nr = 0;
-	S390_lowcore.spinlock_lockval = arch_spin_lockval(0);
-	S390_lowcore.spinlock_index = 0;
+	lc->cpu_nr = 0;
+	lc->spinlock_lockval = arch_spin_lockval(0);
+	lc->spinlock_index = 0;
 }
 
 /*
@@ -1108,35 +1105,34 @@ 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(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(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(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 = !!cpu;
+	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);
@@ -1145,14 +1141,14 @@ 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;
@@ -1161,7 +1157,7 @@ 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();
@@ -1178,7 +1174,7 @@ 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;
 }
@@ -1187,7 +1183,7 @@ static DEVICE_ATTR_WO(rescan);
 static int __init s390_smp_init(void)
 {
 	struct device *dev_root;
-	int cpu, rc = 0;
+	int rc;
 
 	dev_root = bus_get_dev_root(&cpu_subsys);
 	if (dev_root) {
@@ -1196,17 +1192,9 @@ static int __init s390_smp_init(void)
 		if (rc)
 			return rc;
 	}
-
-	for_each_present_cpu(cpu) {
-		rc = smp_add_present_cpu(cpu);
-		if (rc)
-			goto out;
-	}
-
 	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);
diff --git a/arch/s390/kernel/sthyi.c b/arch/s390/kernel/sthyi.c
index 30bb20461db4..1cf2ad04f8e9 100644
--- a/arch/s390/kernel/sthyi.c
+++ b/arch/s390/kernel/sthyi.c
@@ -300,33 +300,56 @@ static struct diag204_x_part_block *lpar_cpu_inf(struct lpar_cpu_inf *part_inf,
 	return (struct diag204_x_part_block *)&block->cpus[i];
 }
 
-static void fill_diag(struct sthyi_sctns *sctns)
+static void *diag204_get_data(bool diag204_allow_busy)
 {
-	int i, r, pages;
-	bool this_lpar;
+	unsigned long subcode;
 	void *diag204_buf;
-	void *diag224_buf = NULL;
-	struct diag204_x_info_blk_hdr *ti_hdr;
-	struct diag204_x_part_block *part_block;
-	struct diag204_x_phys_block *phys_block;
-	struct lpar_cpu_inf lpar_inf = {};
-
-	/* Errors are handled through the validity bits in the response. */
-	pages = diag204((unsigned long)DIAG204_SUBC_RSI |
-			(unsigned long)DIAG204_INFO_EXT, 0, NULL);
-	if (pages <= 0)
-		return;
-
+	int pages, rc;
+
+	subcode = DIAG204_SUBC_RSI;
+	subcode |= DIAG204_INFO_EXT;
+	pages = diag204(subcode, 0, NULL);
+	if (pages < 0)
+		return ERR_PTR(pages);
+	if (pages == 0)
+		return ERR_PTR(-ENODATA);
 	diag204_buf = __vmalloc_node(array_size(pages, PAGE_SIZE),
 				     PAGE_SIZE, GFP_KERNEL, NUMA_NO_NODE,
 				     __builtin_return_address(0));
 	if (!diag204_buf)
-		return;
+		return ERR_PTR(-ENOMEM);
+	subcode = DIAG204_SUBC_STIB7;
+	subcode |= DIAG204_INFO_EXT;
+	if (diag204_has_bif() && diag204_allow_busy)
+		subcode |= DIAG204_BIF_BIT;
+	rc = diag204(subcode, pages, diag204_buf);
+	if (rc < 0) {
+		vfree(diag204_buf);
+		return ERR_PTR(rc);
+	}
+	return diag204_buf;
+}
 
-	r = diag204((unsigned long)DIAG204_SUBC_STIB7 |
-		    (unsigned long)DIAG204_INFO_EXT, pages, diag204_buf);
-	if (r < 0)
-		goto out;
+static bool is_diag204_cached(struct sthyi_sctns *sctns)
+{
+	/*
+	 * Check if validity bits are set when diag204 data
+	 * is gathered.
+	 */
+	if (sctns->par.infpval1)
+		return true;
+	return false;
+}
+
+static void fill_diag(struct sthyi_sctns *sctns, void *diag204_buf)
+{
+	int i;
+	bool this_lpar;
+	void *diag224_buf = NULL;
+	struct diag204_x_info_blk_hdr *ti_hdr;
+	struct diag204_x_part_block *part_block;
+	struct diag204_x_phys_block *phys_block;
+	struct lpar_cpu_inf lpar_inf = {};
 
 	diag224_buf = (void *)__get_free_page(GFP_KERNEL | GFP_DMA);
 	if (!diag224_buf || diag224(diag224_buf))
@@ -392,7 +415,6 @@ static void fill_diag(struct sthyi_sctns *sctns)
 
 out:
 	free_page((unsigned long)diag224_buf);
-	vfree(diag204_buf);
 }
 
 static int sthyi(u64 vaddr, u64 *rc)
@@ -414,19 +436,31 @@ static int sthyi(u64 vaddr, u64 *rc)
 
 static int fill_dst(void *dst, u64 *rc)
 {
+	void *diag204_buf;
+
 	struct sthyi_sctns *sctns = (struct sthyi_sctns *)dst;
 
 	/*
 	 * If the facility is on, we don't want to emulate the instruction.
 	 * We ask the hypervisor to provide the data.
 	 */
-	if (test_facility(74))
+	if (test_facility(74)) {
+		memset(dst, 0, PAGE_SIZE);
 		return sthyi((u64)dst, rc);
-
+	}
+	/*
+	 * When emulating, if diag204 returns BUSY don't reset dst buffer
+	 * and use cached data.
+	 */
+	*rc = 0;
+	diag204_buf = diag204_get_data(is_diag204_cached(sctns));
+	if (IS_ERR(diag204_buf))
+		return PTR_ERR(diag204_buf);
+	memset(dst, 0, PAGE_SIZE);
 	fill_hdr(sctns);
 	fill_stsi(sctns);
-	fill_diag(sctns);
-	*rc = 0;
+	fill_diag(sctns, diag204_buf);
+	vfree(diag204_buf);
 	return 0;
 }
 
@@ -445,11 +479,14 @@ static int sthyi_update_cache(u64 *rc)
 {
 	int r;
 
-	memset(sthyi_cache.info, 0, PAGE_SIZE);
 	r = fill_dst(sthyi_cache.info, rc);
-	if (r)
-		return r;
-	sthyi_cache.end = jiffies + CACHE_VALID_JIFFIES;
+	if (r == 0) {
+		sthyi_cache.end = jiffies + CACHE_VALID_JIFFIES;
+	} else if (r == -EBUSY) {
+		/* mark as expired and return 0 to keep using cached data */
+		sthyi_cache.end = jiffies - 1;
+		r = 0;
+	}
 	return r;
 }
 
diff --git a/arch/s390/kernel/syscall.c b/arch/s390/kernel/syscall.c
index 50cbcbbaa03d..5ec28028315b 100644
--- a/arch/s390/kernel/syscall.c
+++ b/arch/s390/kernel/syscall.c
@@ -124,8 +124,8 @@ void noinstr __do_syscall(struct pt_regs *regs, int per_trap)
 {
 	add_random_kstack_offset();
 	enter_from_user_mode(regs);
-	regs->psw = S390_lowcore.svc_old_psw;
-	regs->int_code = S390_lowcore.svc_int_code;
+	regs->psw = get_lowcore()->svc_old_psw;
+	regs->int_code = get_lowcore()->svc_int_code;
 	update_timer_sys();
 	if (static_branch_likely(&cpu_has_bear))
 		current->thread.last_break = regs->last_break;
diff --git a/arch/s390/kernel/time.c b/arch/s390/kernel/time.c
index fb9f31f36628..b713effe0579 100644
--- a/arch/s390/kernel/time.c
+++ b/arch/s390/kernel/time.c
@@ -131,7 +131,7 @@ void clock_comparator_work(void)
 {
 	struct clock_event_device *cd;
 
-	S390_lowcore.clock_comparator = clock_comparator_max;
+	get_lowcore()->clock_comparator = clock_comparator_max;
 	cd = this_cpu_ptr(&comparators);
 	cd->event_handler(cd);
 }
@@ -139,8 +139,8 @@ void clock_comparator_work(void)
 static int s390_next_event(unsigned long delta,
 			   struct clock_event_device *evt)
 {
-	S390_lowcore.clock_comparator = get_tod_clock() + delta;
-	set_clock_comparator(S390_lowcore.clock_comparator);
+	get_lowcore()->clock_comparator = get_tod_clock() + delta;
+	set_clock_comparator(get_lowcore()->clock_comparator);
 	return 0;
 }
 
@@ -153,8 +153,8 @@ void init_cpu_timer(void)
 	struct clock_event_device *cd;
 	int cpu;
 
-	S390_lowcore.clock_comparator = clock_comparator_max;
-	set_clock_comparator(S390_lowcore.clock_comparator);
+	get_lowcore()->clock_comparator = clock_comparator_max;
+	set_clock_comparator(get_lowcore()->clock_comparator);
 
 	cpu = smp_processor_id();
 	cd = &per_cpu(comparators, cpu);
@@ -184,8 +184,8 @@ static void clock_comparator_interrupt(struct ext_code ext_code,
 				       unsigned long param64)
 {
 	inc_irq_stat(IRQEXT_CLK);
-	if (S390_lowcore.clock_comparator == clock_comparator_max)
-		set_clock_comparator(S390_lowcore.clock_comparator);
+	if (get_lowcore()->clock_comparator == clock_comparator_max)
+		set_clock_comparator(get_lowcore()->clock_comparator);
 }
 
 static void stp_timing_alert(struct stp_irq_parm *);
@@ -408,12 +408,12 @@ static void clock_sync_global(long delta)
 static void clock_sync_local(long delta)
 {
 	/* Add the delta to the clock comparator. */
-	if (S390_lowcore.clock_comparator != clock_comparator_max) {
-		S390_lowcore.clock_comparator += delta;
-		set_clock_comparator(S390_lowcore.clock_comparator);
+	if (get_lowcore()->clock_comparator != clock_comparator_max) {
+		get_lowcore()->clock_comparator += delta;
+		set_clock_comparator(get_lowcore()->clock_comparator);
 	}
 	/* Adjust the last_update_clock time-stamp. */
-	S390_lowcore.last_update_clock += delta;
+	get_lowcore()->last_update_clock += delta;
 }
 
 /* Single threaded workqueue used for stp sync events */
diff --git a/arch/s390/kernel/topology.c b/arch/s390/kernel/topology.c
index 89e91b8ce842..98ef6dc7916b 100644
--- a/arch/s390/kernel/topology.c
+++ b/arch/s390/kernel/topology.c
@@ -320,16 +320,10 @@ static int __arch_update_cpu_topology(void)
 
 int arch_update_cpu_topology(void)
 {
-	struct device *dev;
-	int cpu, rc;
+	int rc;
 
 	rc = __arch_update_cpu_topology();
 	on_each_cpu(__arch_update_dedicated_flag, NULL, 0);
-	for_each_online_cpu(cpu) {
-		dev = get_cpu_device(cpu);
-		if (dev)
-			kobject_uevent(&dev->kobj, KOBJ_CHANGE);
-	}
 	return rc;
 }
 
diff --git a/arch/s390/kernel/traps.c b/arch/s390/kernel/traps.c
index 52578b5cecbd..a7c211a3a0c9 100644
--- a/arch/s390/kernel/traps.c
+++ b/arch/s390/kernel/traps.c
@@ -288,15 +288,16 @@ static void __init test_monitor_call(void)
 
 void __init trap_init(void)
 {
+	struct lowcore *lc = get_lowcore();
 	unsigned long flags;
 	struct ctlreg cr0;
 
 	local_irq_save(flags);
 	cr0 = local_ctl_clear_bit(0, CR0_LOW_ADDRESS_PROTECTION_BIT);
-	psw_bits(S390_lowcore.external_new_psw).mcheck = 1;
-	psw_bits(S390_lowcore.program_new_psw).mcheck = 1;
-	psw_bits(S390_lowcore.svc_new_psw).mcheck = 1;
-	psw_bits(S390_lowcore.io_new_psw).mcheck = 1;
+	psw_bits(lc->external_new_psw).mcheck = 1;
+	psw_bits(lc->program_new_psw).mcheck = 1;
+	psw_bits(lc->svc_new_psw).mcheck = 1;
+	psw_bits(lc->io_new_psw).mcheck = 1;
 	local_ctl_load(0, &cr0);
 	local_irq_restore(flags);
 	local_mcck_enable();
@@ -307,11 +308,12 @@ static void (*pgm_check_table[128])(struct pt_regs *regs);
 
 void noinstr __do_pgm_check(struct pt_regs *regs)
 {
-	unsigned int trapnr;
+	struct lowcore *lc = get_lowcore();
 	irqentry_state_t state;
+	unsigned int trapnr;
 
-	regs->int_code = S390_lowcore.pgm_int_code;
-	regs->int_parm_long = S390_lowcore.trans_exc_code;
+	regs->int_code = lc->pgm_int_code;
+	regs->int_parm_long = lc->trans_exc_code;
 
 	state = irqentry_enter(regs);
 
@@ -324,19 +326,19 @@ void noinstr __do_pgm_check(struct pt_regs *regs)
 		current->thread.last_break = regs->last_break;
 	}
 
-	if (S390_lowcore.pgm_code & 0x0200) {
+	if (lc->pgm_code & 0x0200) {
 		/* transaction abort */
-		current->thread.trap_tdb = S390_lowcore.pgm_tdb;
+		current->thread.trap_tdb = lc->pgm_tdb;
 	}
 
-	if (S390_lowcore.pgm_code & PGM_INT_CODE_PER) {
+	if (lc->pgm_code & PGM_INT_CODE_PER) {
 		if (user_mode(regs)) {
 			struct per_event *ev = &current->thread.per_event;
 
 			set_thread_flag(TIF_PER_TRAP);
-			ev->address = S390_lowcore.per_address;
-			ev->cause = S390_lowcore.per_code_combined;
-			ev->paid = S390_lowcore.per_access_id;
+			ev->address = lc->per_address;
+			ev->cause = lc->per_code_combined;
+			ev->paid = lc->per_access_id;
 		} else {
 			/* PER event in kernel is kprobes */
 			__arch_local_irq_ssm(regs->psw.mask & ~PSW_MASK_PER);
diff --git a/arch/s390/kernel/uv.c b/arch/s390/kernel/uv.c
index 265fea37e030..fa62fa0e369f 100644
--- a/arch/s390/kernel/uv.c
+++ b/arch/s390/kernel/uv.c
@@ -110,7 +110,7 @@ EXPORT_SYMBOL_GPL(uv_pin_shared);
  *
  * @paddr: Absolute host address of page to be destroyed
  */
-static int uv_destroy_page(unsigned long paddr)
+static int uv_destroy(unsigned long paddr)
 {
 	struct uv_cb_cfs uvcb = {
 		.header.cmd = UVC_CMD_DESTR_SEC_STOR,
@@ -131,28 +131,40 @@ static int uv_destroy_page(unsigned long paddr)
 }
 
 /*
- * The caller must already hold a reference to the page
+ * The caller must already hold a reference to the folio
  */
-int uv_destroy_owned_page(unsigned long paddr)
+int uv_destroy_folio(struct folio *folio)
 {
-	struct page *page = phys_to_page(paddr);
 	int rc;
 
-	get_page(page);
-	rc = uv_destroy_page(paddr);
+	/* See gmap_make_secure(): large folios cannot be secure */
+	if (unlikely(folio_test_large(folio)))
+		return 0;
+
+	folio_get(folio);
+	rc = uv_destroy(folio_to_phys(folio));
 	if (!rc)
-		clear_bit(PG_arch_1, &page->flags);
-	put_page(page);
+		clear_bit(PG_arch_1, &folio->flags);
+	folio_put(folio);
 	return rc;
 }
 
 /*
+ * The present PTE still indirectly holds a folio reference through the mapping.
+ */
+int uv_destroy_pte(pte_t pte)
+{
+	VM_WARN_ON(!pte_present(pte));
+	return uv_destroy_folio(pfn_folio(pte_pfn(pte)));
+}
+
+/*
  * Requests the Ultravisor to encrypt a guest page and make it
  * accessible to the host for paging (export).
  *
  * @paddr: Absolute host address of page to be exported
  */
-int uv_convert_from_secure(unsigned long paddr)
+static int uv_convert_from_secure(unsigned long paddr)
 {
 	struct uv_cb_cfs uvcb = {
 		.header.cmd = UVC_CMD_CONV_FROM_SEC_STOR,
@@ -166,22 +178,34 @@ int uv_convert_from_secure(unsigned long paddr)
 }
 
 /*
- * The caller must already hold a reference to the page
+ * The caller must already hold a reference to the folio.
  */
-int uv_convert_owned_from_secure(unsigned long paddr)
+static int uv_convert_from_secure_folio(struct folio *folio)
 {
-	struct page *page = phys_to_page(paddr);
 	int rc;
 
-	get_page(page);
-	rc = uv_convert_from_secure(paddr);
+	/* See gmap_make_secure(): large folios cannot be secure */
+	if (unlikely(folio_test_large(folio)))
+		return 0;
+
+	folio_get(folio);
+	rc = uv_convert_from_secure(folio_to_phys(folio));
 	if (!rc)
-		clear_bit(PG_arch_1, &page->flags);
-	put_page(page);
+		clear_bit(PG_arch_1, &folio->flags);
+	folio_put(folio);
 	return rc;
 }
 
 /*
+ * The present PTE still indirectly holds a folio reference through the mapping.
+ */
+int uv_convert_from_secure_pte(pte_t pte)
+{
+	VM_WARN_ON(!pte_present(pte));
+	return uv_convert_from_secure_folio(pfn_folio(pte_pfn(pte)));
+}
+
+/*
  * Calculate the expected ref_count for a folio that would otherwise have no
  * further pins. This was cribbed from similar functions in other places in
  * the kernel, but with some slight modifications. We know that a secure
@@ -267,6 +291,36 @@ static bool should_export_before_import(struct uv_cb_header *uvcb, struct mm_str
 }
 
 /*
+ * Drain LRU caches: the local one on first invocation and the ones of all
+ * CPUs on successive invocations. Returns "true" on the first invocation.
+ */
+static bool drain_lru(bool *drain_lru_called)
+{
+	/*
+	 * If we have tried a local drain and the folio refcount
+	 * still does not match our expected safe value, try with a
+	 * system wide drain. This is needed if the pagevecs holding
+	 * the page are on a different CPU.
+	 */
+	if (*drain_lru_called) {
+		lru_add_drain_all();
+		/* We give up here, don't retry immediately. */
+		return false;
+	}
+	/*
+	 * We are here if the folio refcount does not match the
+	 * expected safe value. The main culprits are usually
+	 * pagevecs. With lru_add_drain() we drain the pagevecs
+	 * on the local CPU so that hopefully the refcount will
+	 * reach the expected safe value.
+	 */
+	lru_add_drain();
+	*drain_lru_called = true;
+	/* The caller should try again immediately */
+	return true;
+}
+
+/*
  * Requests the Ultravisor to make a page accessible to a guest.
  * If it's brought in the first time, it will be cleared. If
  * it has been exported before, it will be decrypted and integrity
@@ -275,7 +329,7 @@ static bool should_export_before_import(struct uv_cb_header *uvcb, struct mm_str
 int gmap_make_secure(struct gmap *gmap, unsigned long gaddr, void *uvcb)
 {
 	struct vm_area_struct *vma;
-	bool local_drain = false;
+	bool drain_lru_called = false;
 	spinlock_t *ptelock;
 	unsigned long uaddr;
 	struct folio *folio;
@@ -308,52 +362,63 @@ again:
 		goto out;
 	if (pte_present(*ptep) && !(pte_val(*ptep) & _PAGE_INVALID) && pte_write(*ptep)) {
 		folio = page_folio(pte_page(*ptep));
-		rc = -EINVAL;
-		if (folio_test_large(folio))
-			goto unlock;
 		rc = -EAGAIN;
-		if (folio_trylock(folio)) {
+		if (folio_test_large(folio)) {
+			rc = -E2BIG;
+		} else if (folio_trylock(folio)) {
 			if (should_export_before_import(uvcb, gmap->mm))
 				uv_convert_from_secure(PFN_PHYS(folio_pfn(folio)));
 			rc = make_folio_secure(folio, uvcb);
 			folio_unlock(folio);
 		}
+
+		/*
+		 * Once we drop the PTL, the folio may get unmapped and
+		 * freed immediately. We need a temporary reference.
+		 */
+		if (rc == -EAGAIN || rc == -E2BIG)
+			folio_get(folio);
 	}
-unlock:
 	pte_unmap_unlock(ptep, ptelock);
 out:
 	mmap_read_unlock(gmap->mm);
 
-	if (rc == -EAGAIN) {
+	switch (rc) {
+	case -E2BIG:
+		folio_lock(folio);
+		rc = split_folio(folio);
+		folio_unlock(folio);
+		folio_put(folio);
+
+		switch (rc) {
+		case 0:
+			/* Splitting succeeded, try again immediately. */
+			goto again;
+		case -EAGAIN:
+			/* Additional folio references. */
+			if (drain_lru(&drain_lru_called))
+				goto again;
+			return -EAGAIN;
+		case -EBUSY:
+			/* Unexpected race. */
+			return -EAGAIN;
+		}
+		WARN_ON_ONCE(1);
+		return -ENXIO;
+	case -EAGAIN:
 		/*
 		 * If we are here because the UVC returned busy or partial
 		 * completion, this is just a useless check, but it is safe.
 		 */
 		folio_wait_writeback(folio);
-	} else if (rc == -EBUSY) {
-		/*
-		 * If we have tried a local drain and the folio refcount
-		 * still does not match our expected safe value, try with a
-		 * system wide drain. This is needed if the pagevecs holding
-		 * the page are on a different CPU.
-		 */
-		if (local_drain) {
-			lru_add_drain_all();
-			/* We give up here, and let the caller try again */
-			return -EAGAIN;
-		}
-		/*
-		 * We are here if the folio refcount does not match the
-		 * expected safe value. The main culprits are usually
-		 * pagevecs. With lru_add_drain() we drain the pagevecs
-		 * on the local CPU so that hopefully the refcount will
-		 * reach the expected safe value.
-		 */
-		lru_add_drain();
-		local_drain = true;
-		/* And now we try again immediately after draining */
-		goto again;
-	} else if (rc == -ENXIO) {
+		folio_put(folio);
+		return -EAGAIN;
+	case -EBUSY:
+		/* Additional folio references. */
+		if (drain_lru(&drain_lru_called))
+			goto again;
+		return -EAGAIN;
+	case -ENXIO:
 		if (gmap_fault(gmap, gaddr, FAULT_FLAG_WRITE))
 			return -EFAULT;
 		return -EAGAIN;
@@ -388,6 +453,7 @@ int gmap_destroy_page(struct gmap *gmap, unsigned long gaddr)
 {
 	struct vm_area_struct *vma;
 	unsigned long uaddr;
+	struct folio *folio;
 	struct page *page;
 	int rc;
 
@@ -411,7 +477,8 @@ int gmap_destroy_page(struct gmap *gmap, unsigned long gaddr)
 	page = follow_page(vma, uaddr, FOLL_WRITE | FOLL_GET);
 	if (IS_ERR_OR_NULL(page))
 		goto out;
-	rc = uv_destroy_owned_page(page_to_phys(page));
+	folio = page_folio(page);
+	rc = uv_destroy_folio(folio);
 	/*
 	 * Fault handlers can race; it is possible that two CPUs will fault
 	 * on the same secure page. One CPU can destroy the page, reboot,
@@ -422,8 +489,8 @@ int gmap_destroy_page(struct gmap *gmap, unsigned long gaddr)
 	 * we instead try to export the page.
 	 */
 	if (rc)
-		rc = uv_convert_owned_from_secure(page_to_phys(page));
-	put_page(page);
+		rc = uv_convert_from_secure_folio(folio);
+	folio_put(folio);
 out:
 	mmap_read_unlock(gmap->mm);
 	return rc;
@@ -431,47 +498,51 @@ out:
 EXPORT_SYMBOL_GPL(gmap_destroy_page);
 
 /*
- * To be called with the page locked or with an extra reference! This will
- * prevent gmap_make_secure from touching the page concurrently. Having 2
- * parallel make_page_accessible is fine, as the UV calls will become a
- * no-op if the page is already exported.
+ * To be called with the folio locked or with an extra reference! This will
+ * prevent gmap_make_secure from touching the folio concurrently. Having 2
+ * parallel arch_make_folio_accessible is fine, as the UV calls will become a
+ * no-op if the folio is already exported.
  */
-int arch_make_page_accessible(struct page *page)
+int arch_make_folio_accessible(struct folio *folio)
 {
 	int rc = 0;
 
-	/* Hugepage cannot be protected, so nothing to do */
-	if (PageHuge(page))
+	/* See gmap_make_secure(): large folios cannot be secure */
+	if (unlikely(folio_test_large(folio)))
 		return 0;
 
 	/*
-	 * PG_arch_1 is used in 3 places:
-	 * 1. for kernel page tables during early boot
-	 * 2. for storage keys of huge pages and KVM
-	 * 3. As an indication that this page might be secure. This can
+	 * PG_arch_1 is used in 2 places:
+	 * 1. for storage keys of hugetlb folios and KVM
+	 * 2. As an indication that this small folio might be secure. This can
 	 *    overindicate, e.g. we set the bit before calling
 	 *    convert_to_secure.
-	 * As secure pages are never huge, all 3 variants can co-exists.
+	 * As secure pages are never large folios, both variants can co-exists.
 	 */
-	if (!test_bit(PG_arch_1, &page->flags))
+	if (!test_bit(PG_arch_1, &folio->flags))
 		return 0;
 
-	rc = uv_pin_shared(page_to_phys(page));
+	rc = uv_pin_shared(folio_to_phys(folio));
 	if (!rc) {
-		clear_bit(PG_arch_1, &page->flags);
+		clear_bit(PG_arch_1, &folio->flags);
 		return 0;
 	}
 
-	rc = uv_convert_from_secure(page_to_phys(page));
+	rc = uv_convert_from_secure(folio_to_phys(folio));
 	if (!rc) {
-		clear_bit(PG_arch_1, &page->flags);
+		clear_bit(PG_arch_1, &folio->flags);
 		return 0;
 	}
 
 	return rc;
 }
-EXPORT_SYMBOL_GPL(arch_make_page_accessible);
+EXPORT_SYMBOL_GPL(arch_make_folio_accessible);
 
+int arch_make_page_accessible(struct page *page)
+{
+	return arch_make_folio_accessible(page_folio(page));
+}
+EXPORT_SYMBOL_GPL(arch_make_page_accessible);
 #endif
 
 #if defined(CONFIG_PROTECTED_VIRTUALIZATION_GUEST) || IS_ENABLED(CONFIG_KVM)
diff --git a/arch/s390/kernel/vtime.c b/arch/s390/kernel/vtime.c
index ffc1db0cbf9c..234a0ba30510 100644
--- a/arch/s390/kernel/vtime.c
+++ b/arch/s390/kernel/vtime.c
@@ -35,14 +35,15 @@ static DEFINE_PER_CPU(u64, mt_scaling_jiffies);
 
 static inline void set_vtimer(u64 expires)
 {
+	struct lowcore *lc = get_lowcore();
 	u64 timer;
 
 	asm volatile(
 		"	stpt	%0\n"	/* Store current cpu timer value */
 		"	spt	%1"	/* Set new value imm. afterwards */
 		: "=Q" (timer) : "Q" (expires));
-	S390_lowcore.system_timer += S390_lowcore.last_update_timer - timer;
-	S390_lowcore.last_update_timer = expires;
+	lc->system_timer += lc->last_update_timer - timer;
+	lc->last_update_timer = expires;
 }
 
 static inline int virt_timer_forward(u64 elapsed)
@@ -117,22 +118,23 @@ static void account_system_index_scaled(struct task_struct *p, u64 cputime,
 static int do_account_vtime(struct task_struct *tsk)
 {
 	u64 timer, clock, user, guest, system, hardirq, softirq;
+	struct lowcore *lc = get_lowcore();
 
-	timer = S390_lowcore.last_update_timer;
-	clock = S390_lowcore.last_update_clock;
+	timer = lc->last_update_timer;
+	clock = lc->last_update_clock;
 	asm volatile(
 		"	stpt	%0\n"	/* Store current cpu timer value */
 		"	stckf	%1"	/* Store current tod clock value */
-		: "=Q" (S390_lowcore.last_update_timer),
-		  "=Q" (S390_lowcore.last_update_clock)
+		: "=Q" (lc->last_update_timer),
+		  "=Q" (lc->last_update_clock)
 		: : "cc");
-	clock = S390_lowcore.last_update_clock - clock;
-	timer -= S390_lowcore.last_update_timer;
+	clock = lc->last_update_clock - clock;
+	timer -= lc->last_update_timer;
 
 	if (hardirq_count())
-		S390_lowcore.hardirq_timer += timer;
+		lc->hardirq_timer += timer;
 	else
-		S390_lowcore.system_timer += timer;
+		lc->system_timer += timer;
 
 	/* Update MT utilization calculation */
 	if (smp_cpu_mtid &&
@@ -141,16 +143,16 @@ static int do_account_vtime(struct task_struct *tsk)
 
 	/* Calculate cputime delta */
 	user = update_tsk_timer(&tsk->thread.user_timer,
-				READ_ONCE(S390_lowcore.user_timer));
+				READ_ONCE(lc->user_timer));
 	guest = update_tsk_timer(&tsk->thread.guest_timer,
-				 READ_ONCE(S390_lowcore.guest_timer));
+				 READ_ONCE(lc->guest_timer));
 	system = update_tsk_timer(&tsk->thread.system_timer,
-				  READ_ONCE(S390_lowcore.system_timer));
+				  READ_ONCE(lc->system_timer));
 	hardirq = update_tsk_timer(&tsk->thread.hardirq_timer,
-				   READ_ONCE(S390_lowcore.hardirq_timer));
+				   READ_ONCE(lc->hardirq_timer));
 	softirq = update_tsk_timer(&tsk->thread.softirq_timer,
-				   READ_ONCE(S390_lowcore.softirq_timer));
-	S390_lowcore.steal_timer +=
+				   READ_ONCE(lc->softirq_timer));
+	lc->steal_timer +=
 		clock - user - guest - system - hardirq - softirq;
 
 	/* Push account value */
@@ -176,17 +178,19 @@ static int do_account_vtime(struct task_struct *tsk)
 
 void vtime_task_switch(struct task_struct *prev)
 {
+	struct lowcore *lc = get_lowcore();
+
 	do_account_vtime(prev);
-	prev->thread.user_timer = S390_lowcore.user_timer;
-	prev->thread.guest_timer = S390_lowcore.guest_timer;
-	prev->thread.system_timer = S390_lowcore.system_timer;
-	prev->thread.hardirq_timer = S390_lowcore.hardirq_timer;
-	prev->thread.softirq_timer = S390_lowcore.softirq_timer;
-	S390_lowcore.user_timer = current->thread.user_timer;
-	S390_lowcore.guest_timer = current->thread.guest_timer;
-	S390_lowcore.system_timer = current->thread.system_timer;
-	S390_lowcore.hardirq_timer = current->thread.hardirq_timer;
-	S390_lowcore.softirq_timer = current->thread.softirq_timer;
+	prev->thread.user_timer = lc->user_timer;
+	prev->thread.guest_timer = lc->guest_timer;
+	prev->thread.system_timer = lc->system_timer;
+	prev->thread.hardirq_timer = lc->hardirq_timer;
+	prev->thread.softirq_timer = lc->softirq_timer;
+	lc->user_timer = current->thread.user_timer;
+	lc->guest_timer = current->thread.guest_timer;
+	lc->system_timer = current->thread.system_timer;
+	lc->hardirq_timer = current->thread.hardirq_timer;
+	lc->softirq_timer = current->thread.softirq_timer;
 }
 
 /*
@@ -196,28 +200,29 @@ void vtime_task_switch(struct task_struct *prev)
  */
 void vtime_flush(struct task_struct *tsk)
 {
+	struct lowcore *lc = get_lowcore();
 	u64 steal, avg_steal;
 
 	if (do_account_vtime(tsk))
 		virt_timer_expire();
 
-	steal = S390_lowcore.steal_timer;
-	avg_steal = S390_lowcore.avg_steal_timer;
+	steal = lc->steal_timer;
+	avg_steal = lc->avg_steal_timer;
 	if ((s64) steal > 0) {
-		S390_lowcore.steal_timer = 0;
+		lc->steal_timer = 0;
 		account_steal_time(cputime_to_nsecs(steal));
 		avg_steal += steal;
 	}
-	S390_lowcore.avg_steal_timer = avg_steal / 2;
+	lc->avg_steal_timer = avg_steal / 2;
 }
 
 static u64 vtime_delta(void)
 {
-	u64 timer = S390_lowcore.last_update_timer;
-
-	S390_lowcore.last_update_timer = get_cpu_timer();
+	struct lowcore *lc = get_lowcore();
+	u64 timer = lc->last_update_timer;
 
-	return timer - S390_lowcore.last_update_timer;
+	lc->last_update_timer = get_cpu_timer();
+	return timer - lc->last_update_timer;
 }
 
 /*
@@ -226,12 +231,13 @@ static u64 vtime_delta(void)
  */
 void vtime_account_kernel(struct task_struct *tsk)
 {
+	struct lowcore *lc = get_lowcore();
 	u64 delta = vtime_delta();
 
 	if (tsk->flags & PF_VCPU)
-		S390_lowcore.guest_timer += delta;
+		lc->guest_timer += delta;
 	else
-		S390_lowcore.system_timer += delta;
+		lc->system_timer += delta;
 
 	virt_timer_forward(delta);
 }
@@ -241,7 +247,7 @@ void vtime_account_softirq(struct task_struct *tsk)
 {
 	u64 delta = vtime_delta();
 
-	S390_lowcore.softirq_timer += delta;
+	get_lowcore()->softirq_timer += delta;
 
 	virt_timer_forward(delta);
 }
@@ -250,7 +256,7 @@ void vtime_account_hardirq(struct task_struct *tsk)
 {
 	u64 delta = vtime_delta();
 
-	S390_lowcore.hardirq_timer += delta;
+	get_lowcore()->hardirq_timer += delta;
 
 	virt_timer_forward(delta);
 }