Merge tag 's390-6.9-1' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux

Pull s390 updates from Heiko Carstens:

 - Various virtual vs physical address usage fixes

 - Fix error handling in Processor Activity Instrumentation device
   driver, and export number of counters with a sysfs file

 - Allow for multiple events when Processor Activity Instrumentation
   counters are monitored in system wide sampling

 - Change multiplier and shift values of the Time-of-Day clock source to
   improve steering precision

 - Remove a couple of unneeded GFP_DMA flags from allocations

 - Disable mmap alignment if randomize_va_space is also disabled, to
   avoid a too small heap

 - Various changes to allow s390 to be compiled with LLVM=1, since
   ld.lld and llvm-objcopy will have proper s390 support witch clang 19

 - Add __uninitialized macro to Compiler Attributes. This is helpful
   with s390's FPU code where some users have up to 520 byte stack
   frames. Clearing such stack frames (if INIT_STACK_ALL_PATTERN or
   INIT_STACK_ALL_ZERO is enabled) before they are used contradicts the
   intention (performance improvement) of such code sections.

 - Convert switch_to() to an out-of-line function, and use the generic
   switch_to header file

 - Replace the usage of s390's debug feature with pr_debug() calls
   within the zcrypt device driver

 - Improve hotplug support of the Adjunct Processor device driver

 - Improve retry handling in the zcrypt device driver

 - Various changes to the in-kernel FPU code:

     - Make in-kernel FPU sections preemptible

     - Convert various larger inline assemblies and assembler files to
       C, mainly by using singe instruction inline assemblies. This
       increases readability, but also allows makes it easier to add
       proper instrumentation hooks

     - Cleanup of the header files

 - Provide fast variants of csum_partial() and
   csum_partial_copy_nocheck() based on vector instructions

 - Introduce and use a lock to synchronize accesses to zpci device data
   structures to avoid inconsistent states caused by concurrent accesses

 - Compile the kernel without -fPIE. This addresses the following
   problems if the kernel is compiled with -fPIE:

     - It uses dynamic symbols (.dynsym), for which the linker refuses
       to allow more than 64k sections. This can break features which
       use '-ffunction-sections' and '-fdata-sections', including
       kpatch-build and function granular KASLR

     - It unnecessarily uses GOT relocations, adding an extra layer of
       indirection for many memory accesses

 - Fix shared_cpu_list for CPU private L2 caches, which incorrectly were
   reported as globally shared

* tag 's390-6.9-1' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux: (117 commits)
  s390/tools: handle rela R_390_GOTPCDBL/R_390_GOTOFF64
  s390/cache: prevent rebuild of shared_cpu_list
  s390/crypto: remove retry loop with sleep from PAES pkey invocation
  s390/pkey: improve pkey retry behavior
  s390/zcrypt: improve zcrypt retry behavior
  s390/zcrypt: introduce retries on in-kernel send CPRB functions
  s390/ap: introduce mutex to lock the AP bus scan
  s390/ap: rework ap_scan_bus() to return true on config change
  s390/ap: clarify AP scan bus related functions and variables
  s390/ap: rearm APQNs bindings complete completion
  s390/configs: increase number of LOCKDEP_BITS
  s390/vfio-ap: handle hardware checkstop state on queue reset operation
  s390/pai: change sampling event assignment for PMU device driver
  s390/boot: fix minor comment style damages
  s390/boot: do not check for zero-termination relocation entry
  s390/boot: make type of __vmlinux_relocs_64_start|end consistent
  s390/boot: sanitize kaslr_adjust_relocs() function prototype
  s390/boot: simplify GOT handling
  s390: vmlinux.lds.S: fix .got.plt assertion
  s390/boot: workaround current 'llvm-objdump -t -j ...' behavior
  ...

30 files changed, 503 insertions, 549 deletions

diff --git a/arch/s390/kernel/cache.c b/arch/s390/kernel/cache.c
index 56254fa06f99..4f2669030220 100644
--- a/arch/s390/kernel/cache.c
+++ b/arch/s390/kernel/cache.c
@@ -166,5 +166,6 @@ int populate_cache_leaves(unsigned int cpu)
 			ci_leaf_init(this_leaf++, pvt, ctype, level, cpu);
 		}
 	}
+	this_cpu_ci->cpu_map_populated = true;
 	return 0;
 }
diff --git a/arch/s390/kernel/compat_signal.c b/arch/s390/kernel/compat_signal.c
index f8fc6c25d051..1942e2a9f8db 100644
--- a/arch/s390/kernel/compat_signal.c
+++ b/arch/s390/kernel/compat_signal.c
@@ -24,12 +24,12 @@
 #include <linux/tty.h>
 #include <linux/personality.h>
 #include <linux/binfmts.h>
+#include <asm/access-regs.h>
 #include <asm/ucontext.h>
 #include <linux/uaccess.h>
 #include <asm/lowcore.h>
-#include <asm/switch_to.h>
 #include <asm/vdso.h>
-#include <asm/fpu/api.h>
+#include <asm/fpu.h>
 #include "compat_linux.h"
 #include "compat_ptrace.h"
 #include "entry.h"
@@ -56,7 +56,7 @@ typedef struct
 static void store_sigregs(void)
 {
 	save_access_regs(current->thread.acrs);
-	save_fpu_regs();
+	save_user_fpu_regs();
 }
 
 /* Load registers after signal return */
@@ -79,7 +79,7 @@ static int save_sigregs32(struct pt_regs *regs, _sigregs32 __user *sregs)
 		user_sregs.regs.gprs[i] = (__u32) regs->gprs[i];
 	memcpy(&user_sregs.regs.acrs, current->thread.acrs,
 	       sizeof(user_sregs.regs.acrs));
-	fpregs_store((_s390_fp_regs *) &user_sregs.fpregs, &current->thread.fpu);
+	fpregs_store((_s390_fp_regs *) &user_sregs.fpregs, &current->thread.ufpu);
 	if (__copy_to_user(sregs, &user_sregs, sizeof(_sigregs32)))
 		return -EFAULT;
 	return 0;
@@ -113,7 +113,7 @@ static int restore_sigregs32(struct pt_regs *regs,_sigregs32 __user *sregs)
 		regs->gprs[i] = (__u64) user_sregs.regs.gprs[i];
 	memcpy(&current->thread.acrs, &user_sregs.regs.acrs,
 	       sizeof(current->thread.acrs));
-	fpregs_load((_s390_fp_regs *) &user_sregs.fpregs, &current->thread.fpu);
+	fpregs_load((_s390_fp_regs *)&user_sregs.fpregs, &current->thread.ufpu);
 
 	clear_pt_regs_flag(regs, PIF_SYSCALL); /* No longer in a system call */
 	return 0;
@@ -136,11 +136,11 @@ static int save_sigregs_ext32(struct pt_regs *regs,
 	/* Save vector registers to signal stack */
 	if (cpu_has_vx()) {
 		for (i = 0; i < __NUM_VXRS_LOW; i++)
-			vxrs[i] = current->thread.fpu.vxrs[i].low;
+			vxrs[i] = current->thread.ufpu.vxrs[i].low;
 		if (__copy_to_user(&sregs_ext->vxrs_low, vxrs,
 				   sizeof(sregs_ext->vxrs_low)) ||
 		    __copy_to_user(&sregs_ext->vxrs_high,
-				   current->thread.fpu.vxrs + __NUM_VXRS_LOW,
+				   current->thread.ufpu.vxrs + __NUM_VXRS_LOW,
 				   sizeof(sregs_ext->vxrs_high)))
 			return -EFAULT;
 	}
@@ -165,12 +165,12 @@ static int restore_sigregs_ext32(struct pt_regs *regs,
 	if (cpu_has_vx()) {
 		if (__copy_from_user(vxrs, &sregs_ext->vxrs_low,
 				     sizeof(sregs_ext->vxrs_low)) ||
-		    __copy_from_user(current->thread.fpu.vxrs + __NUM_VXRS_LOW,
+		    __copy_from_user(current->thread.ufpu.vxrs + __NUM_VXRS_LOW,
 				     &sregs_ext->vxrs_high,
 				     sizeof(sregs_ext->vxrs_high)))
 			return -EFAULT;
 		for (i = 0; i < __NUM_VXRS_LOW; i++)
-			current->thread.fpu.vxrs[i].low = vxrs[i];
+			current->thread.ufpu.vxrs[i].low = vxrs[i];
 	}
 	return 0;
 }
@@ -184,7 +184,7 @@ COMPAT_SYSCALL_DEFINE0(sigreturn)
 	if (get_compat_sigset(&set, (compat_sigset_t __user *)frame->sc.oldmask))
 		goto badframe;
 	set_current_blocked(&set);
-	save_fpu_regs();
+	save_user_fpu_regs();
 	if (restore_sigregs32(regs, &frame->sregs))
 		goto badframe;
 	if (restore_sigregs_ext32(regs, &frame->sregs_ext))
@@ -207,7 +207,7 @@ COMPAT_SYSCALL_DEFINE0(rt_sigreturn)
 	set_current_blocked(&set);
 	if (compat_restore_altstack(&frame->uc.uc_stack))
 		goto badframe;
-	save_fpu_regs();
+	save_user_fpu_regs();
 	if (restore_sigregs32(regs, &frame->uc.uc_mcontext))
 		goto badframe;
 	if (restore_sigregs_ext32(regs, &frame->uc.uc_mcontext_ext))
diff --git a/arch/s390/kernel/crash_dump.c b/arch/s390/kernel/crash_dump.c
index 5c46c2659305..d09ebb6f5262 100644
--- a/arch/s390/kernel/crash_dump.c
+++ b/arch/s390/kernel/crash_dump.c
@@ -22,7 +22,7 @@
 #include <asm/ipl.h>
 #include <asm/sclp.h>
 #include <asm/maccess.h>
-#include <asm/fpu/api.h>
+#include <asm/fpu.h>
 
 #define PTR_ADD(x, y) (((char *) (x)) + ((unsigned long) (y)))
 #define PTR_SUB(x, y) (((char *) (x)) - ((unsigned long) (y)))
diff --git a/arch/s390/kernel/diag.c b/arch/s390/kernel/diag.c
index 92fdc35f028c..8dee9aa0ec95 100644
--- a/arch/s390/kernel/diag.c
+++ b/arch/s390/kernel/diag.c
@@ -147,11 +147,40 @@ void notrace diag_stat_inc_norecursion(enum diag_stat_enum nr)
 EXPORT_SYMBOL(diag_stat_inc_norecursion);
 
 /*
+ * Diagnose 0c: Pseudo Timer
+ */
+void diag0c(struct hypfs_diag0c_entry *data)
+{
+	diag_stat_inc(DIAG_STAT_X00C);
+	diag_amode31_ops.diag0c(virt_to_phys(data));
+}
+
+/*
  * Diagnose 14: Input spool file manipulation
+ *
+ * The subcode parameter determines the type of the first parameter rx.
+ * Currently used are the following 3 subcommands:
+ * 0x0:   Read the Next Spool File Buffer (Data Record)
+ * 0x28:  Position a Spool File to the Designated Record
+ * 0xfff: Retrieve Next File Descriptor
+ *
+ * For subcommands 0x0 and 0xfff, the value of the first parameter is
+ * a virtual address of a memory buffer and needs virtual to physical
+ * address translation. For other subcommands the rx parameter is not
+ * a virtual address.
  */
 int diag14(unsigned long rx, unsigned long ry1, unsigned long subcode)
 {
 	diag_stat_inc(DIAG_STAT_X014);
+	switch (subcode) {
+	case 0x0:
+	case 0xfff:
+		rx = virt_to_phys((void *)rx);
+		break;
+	default:
+		/* Do nothing */
+		break;
+	}
 	return diag_amode31_ops.diag14(rx, ry1, subcode);
 }
 EXPORT_SYMBOL(diag14);
@@ -265,6 +294,6 @@ EXPORT_SYMBOL(diag224);
 int diag26c(void *req, void *resp, enum diag26c_sc subcode)
 {
 	diag_stat_inc(DIAG_STAT_X26C);
-	return diag_amode31_ops.diag26c(req, resp, subcode);
+	return diag_amode31_ops.diag26c(virt_to_phys(req), virt_to_phys(resp), subcode);
 }
 EXPORT_SYMBOL(diag26c);
diff --git a/arch/s390/kernel/early.c b/arch/s390/kernel/early.c
index 2345ea332b97..c666271433fb 100644
--- a/arch/s390/kernel/early.c
+++ b/arch/s390/kernel/early.c
@@ -19,8 +19,10 @@
 #include <linux/kernel.h>
 #include <asm/asm-extable.h>
 #include <linux/memblock.h>
+#include <asm/access-regs.h>
 #include <asm/diag.h>
 #include <asm/ebcdic.h>
+#include <asm/fpu.h>
 #include <asm/ipl.h>
 #include <asm/lowcore.h>
 #include <asm/processor.h>
@@ -31,7 +33,6 @@
 #include <asm/sclp.h>
 #include <asm/facility.h>
 #include <asm/boot_data.h>
-#include <asm/switch_to.h>
 #include "entry.h"
 
 #define decompressor_handled_param(param)			\
diff --git a/arch/s390/kernel/entry.S b/arch/s390/kernel/entry.S
index 49a11f6dd7ae..fc5277eab554 100644
--- a/arch/s390/kernel/entry.S
+++ b/arch/s390/kernel/entry.S
@@ -24,7 +24,7 @@
 #include <asm/page.h>
 #include <asm/sigp.h>
 #include <asm/irq.h>
-#include <asm/vx-insn.h>
+#include <asm/fpu-insn.h>
 #include <asm/setup.h>
 #include <asm/nmi.h>
 #include <asm/nospec-insn.h>
@@ -171,13 +171,13 @@ _LPP_OFFSET	= __LC_LPP
 	nop	0
 
 /*
- * Scheduler resume function, called by switch_to
- *  gpr2 = (task_struct *) prev
- *  gpr3 = (task_struct *) next
+ * Scheduler resume function, called by __switch_to
+ *  gpr2 = (task_struct *)prev
+ *  gpr3 = (task_struct *)next
  * Returns:
  *  gpr2 = prev
  */
-SYM_FUNC_START(__switch_to)
+SYM_FUNC_START(__switch_to_asm)
 	stmg	%r6,%r15,__SF_GPRS(%r15)	# store gprs of prev task
 	lghi	%r4,__TASK_stack
 	lghi	%r1,__TASK_thread
@@ -193,7 +193,7 @@ SYM_FUNC_START(__switch_to)
 	lmg	%r6,%r15,__SF_GPRS(%r15)	# load gprs of next task
 	ALTERNATIVE "nop", "lpp _LPP_OFFSET", 40
 	BR_EX	%r14
-SYM_FUNC_END(__switch_to)
+SYM_FUNC_END(__switch_to_asm)
 
 #if IS_ENABLED(CONFIG_KVM)
 /*
@@ -220,8 +220,6 @@ SYM_FUNC_START(__sie64a)
 	oi	__SIE_PROG0C+3(%r14),1		# we are going into SIE now
 	tm	__SIE_PROG20+3(%r14),3		# last exit...
 	jnz	.Lsie_skip
-	TSTMSK	__LC_CPU_FLAGS,_CIF_FPU
-	jo	.Lsie_skip			# exit if fp/vx regs changed
 	lg	%r14,__SF_SIE_CONTROL_PHYS(%r15)	# get sie block phys addr
 	BPEXIT	__SF_SIE_FLAGS(%r15),_TIF_ISOLATE_BP_GUEST
 .Lsie_entry:
@@ -489,16 +487,11 @@ SYM_FUNC_END(psw_idle)
  */
 SYM_CODE_START(mcck_int_handler)
 	BPOFF
-	la	%r1,4095		# validate r1
-	spt	__LC_CPU_TIMER_SAVE_AREA-4095(%r1)	# validate cpu timer
-	LBEAR	__LC_LAST_BREAK_SAVE_AREA-4095(%r1)		# validate bear
-	lmg	%r0,%r15,__LC_GPREGS_SAVE_AREA		# validate gprs
 	lmg	%r8,%r9,__LC_MCK_OLD_PSW
 	TSTMSK	__LC_MCCK_CODE,MCCK_CODE_SYSTEM_DAMAGE
 	jo	.Lmcck_panic		# yes -> rest of mcck code invalid
 	TSTMSK	__LC_MCCK_CODE,MCCK_CODE_CR_VALID
 	jno	.Lmcck_panic		# control registers invalid -> panic
-	lctlg	%c0,%c15,__LC_CREGS_SAVE_AREA		# validate ctl regs
 	ptlb
 	lghi	%r14,__LC_CPU_TIMER_SAVE_AREA
 	mvc	__LC_MCCK_ENTER_TIMER(8),0(%r14)
diff --git a/arch/s390/kernel/entry.h b/arch/s390/kernel/entry.h
index 9f41853f36b9..21969520f947 100644
--- a/arch/s390/kernel/entry.h
+++ b/arch/s390/kernel/entry.h
@@ -19,6 +19,7 @@ void mcck_int_handler(void);
 void restart_int_handler(void);
 void early_pgm_check_handler(void);
 
+struct task_struct *__switch_to_asm(struct task_struct *prev, struct task_struct *next);
 void __ret_from_fork(struct task_struct *prev, struct pt_regs *regs);
 void __do_pgm_check(struct pt_regs *regs);
 void __do_syscall(struct pt_regs *regs, int per_trap);
diff --git a/arch/s390/kernel/fpu.c b/arch/s390/kernel/fpu.c
index a4f3449cc814..fa90bbdc5ef9 100644
--- a/arch/s390/kernel/fpu.c
+++ b/arch/s390/kernel/fpu.c
@@ -8,256 +8,186 @@
 #include <linux/kernel.h>
 #include <linux/cpu.h>
 #include <linux/sched.h>
-#include <asm/fpu/types.h>
-#include <asm/fpu/api.h>
-#include <asm/vx-insn.h>
+#include <asm/fpu.h>
 
-void __kernel_fpu_begin(struct kernel_fpu *state, u32 flags)
+void __kernel_fpu_begin(struct kernel_fpu *state, int flags)
 {
+	__vector128 *vxrs = state->vxrs;
+	int mask;
+
 	/*
 	 * Limit the save to the FPU/vector registers already
-	 * in use by the previous context
+	 * in use by the previous context.
 	 */
-	flags &= state->mask;
-
+	flags &= state->hdr.mask;
 	if (flags & KERNEL_FPC)
-		/* Save floating point control */
-		asm volatile("stfpc %0" : "=Q" (state->fpc));
-
+		fpu_stfpc(&state->hdr.fpc);
 	if (!cpu_has_vx()) {
-		if (flags & KERNEL_VXR_V0V7) {
-			/* Save floating-point registers */
-			asm volatile("std 0,%0" : "=Q" (state->fprs[0]));
-			asm volatile("std 1,%0" : "=Q" (state->fprs[1]));
-			asm volatile("std 2,%0" : "=Q" (state->fprs[2]));
-			asm volatile("std 3,%0" : "=Q" (state->fprs[3]));
-			asm volatile("std 4,%0" : "=Q" (state->fprs[4]));
-			asm volatile("std 5,%0" : "=Q" (state->fprs[5]));
-			asm volatile("std 6,%0" : "=Q" (state->fprs[6]));
-			asm volatile("std 7,%0" : "=Q" (state->fprs[7]));
-			asm volatile("std 8,%0" : "=Q" (state->fprs[8]));
-			asm volatile("std 9,%0" : "=Q" (state->fprs[9]));
-			asm volatile("std 10,%0" : "=Q" (state->fprs[10]));
-			asm volatile("std 11,%0" : "=Q" (state->fprs[11]));
-			asm volatile("std 12,%0" : "=Q" (state->fprs[12]));
-			asm volatile("std 13,%0" : "=Q" (state->fprs[13]));
-			asm volatile("std 14,%0" : "=Q" (state->fprs[14]));
-			asm volatile("std 15,%0" : "=Q" (state->fprs[15]));
-		}
+		if (flags & KERNEL_VXR_LOW)
+			save_fp_regs_vx(vxrs);
 		return;
 	}
-
-	/* Test and save vector registers */
-	asm volatile (
-		/*
-		 * Test if any vector register must be saved and, if so,
-		 * test if all register can be saved.
-		 */
-		"	la	1,%[vxrs]\n"	/* load save area */
-		"	tmll	%[m],30\n"	/* KERNEL_VXR */
-		"	jz	7f\n"		/* no work -> done */
-		"	jo	5f\n"		/* -> save V0..V31 */
-		/*
-		 * Test for special case KERNEL_FPU_MID only. In this
-		 * case a vstm V8..V23 is the best instruction
-		 */
-		"	chi	%[m],12\n"	/* KERNEL_VXR_MID */
-		"	jne	0f\n"		/* -> save V8..V23 */
-		"	VSTM	8,23,128,1\n"	/* vstm %v8,%v23,128(%r1) */
-		"	j	7f\n"
-		/* Test and save the first half of 16 vector registers */
-		"0:	tmll	%[m],6\n"	/* KERNEL_VXR_LOW */
-		"	jz	3f\n"		/* -> KERNEL_VXR_HIGH */
-		"	jo	2f\n"		/* 11 -> save V0..V15 */
-		"	brc	2,1f\n"		/* 10 -> save V8..V15 */
-		"	VSTM	0,7,0,1\n"	/* vstm %v0,%v7,0(%r1) */
-		"	j	3f\n"
-		"1:	VSTM	8,15,128,1\n"	/* vstm %v8,%v15,128(%r1) */
-		"	j	3f\n"
-		"2:	VSTM	0,15,0,1\n"	/* vstm %v0,%v15,0(%r1) */
-		/* Test and save the second half of 16 vector registers */
-		"3:	tmll	%[m],24\n"	/* KERNEL_VXR_HIGH */
-		"	jz	7f\n"
-		"	jo	6f\n"		/* 11 -> save V16..V31 */
-		"	brc	2,4f\n"		/* 10 -> save V24..V31 */
-		"	VSTM	16,23,256,1\n"	/* vstm %v16,%v23,256(%r1) */
-		"	j	7f\n"
-		"4:	VSTM	24,31,384,1\n"	/* vstm %v24,%v31,384(%r1) */
-		"	j	7f\n"
-		"5:	VSTM	0,15,0,1\n"	/* vstm %v0,%v15,0(%r1) */
-		"6:	VSTM	16,31,256,1\n"	/* vstm %v16,%v31,256(%r1) */
-		"7:"
-		: [vxrs] "=Q" (*(struct vx_array *) &state->vxrs)
-		: [m] "d" (flags)
-		: "1", "cc");
+	mask = flags & KERNEL_VXR;
+	if (mask == KERNEL_VXR) {
+		vxrs += fpu_vstm(0, 15, vxrs);
+		vxrs += fpu_vstm(16, 31, vxrs);
+		return;
+	}
+	if (mask == KERNEL_VXR_MID) {
+		vxrs += fpu_vstm(8, 23, vxrs);
+		return;
+	}
+	mask = flags & KERNEL_VXR_LOW;
+	if (mask) {
+		if (mask == KERNEL_VXR_LOW)
+			vxrs += fpu_vstm(0, 15, vxrs);
+		else if (mask == KERNEL_VXR_V0V7)
+			vxrs += fpu_vstm(0, 7, vxrs);
+		else
+			vxrs += fpu_vstm(8, 15, vxrs);
+	}
+	mask = flags & KERNEL_VXR_HIGH;
+	if (mask) {
+		if (mask == KERNEL_VXR_HIGH)
+			vxrs += fpu_vstm(16, 31, vxrs);
+		else if (mask == KERNEL_VXR_V16V23)
+			vxrs += fpu_vstm(16, 23, vxrs);
+		else
+			vxrs += fpu_vstm(24, 31, vxrs);
+	}
 }
 EXPORT_SYMBOL(__kernel_fpu_begin);
 
-void __kernel_fpu_end(struct kernel_fpu *state, u32 flags)
+void __kernel_fpu_end(struct kernel_fpu *state, int flags)
 {
+	__vector128 *vxrs = state->vxrs;
+	int mask;
+
 	/*
 	 * Limit the restore to the FPU/vector registers of the
-	 * previous context that have been overwritte by the
-	 * current context
+	 * previous context that have been overwritten by the
+	 * current context.
 	 */
-	flags &= state->mask;
-
+	flags &= state->hdr.mask;
 	if (flags & KERNEL_FPC)
-		/* Restore floating-point controls */
-		asm volatile("lfpc %0" : : "Q" (state->fpc));
-
+		fpu_lfpc(&state->hdr.fpc);
 	if (!cpu_has_vx()) {
-		if (flags & KERNEL_VXR_V0V7) {
-			/* Restore floating-point registers */
-			asm volatile("ld 0,%0" : : "Q" (state->fprs[0]));
-			asm volatile("ld 1,%0" : : "Q" (state->fprs[1]));
-			asm volatile("ld 2,%0" : : "Q" (state->fprs[2]));
-			asm volatile("ld 3,%0" : : "Q" (state->fprs[3]));
-			asm volatile("ld 4,%0" : : "Q" (state->fprs[4]));
-			asm volatile("ld 5,%0" : : "Q" (state->fprs[5]));
-			asm volatile("ld 6,%0" : : "Q" (state->fprs[6]));
-			asm volatile("ld 7,%0" : : "Q" (state->fprs[7]));
-			asm volatile("ld 8,%0" : : "Q" (state->fprs[8]));
-			asm volatile("ld 9,%0" : : "Q" (state->fprs[9]));
-			asm volatile("ld 10,%0" : : "Q" (state->fprs[10]));
-			asm volatile("ld 11,%0" : : "Q" (state->fprs[11]));
-			asm volatile("ld 12,%0" : : "Q" (state->fprs[12]));
-			asm volatile("ld 13,%0" : : "Q" (state->fprs[13]));
-			asm volatile("ld 14,%0" : : "Q" (state->fprs[14]));
-			asm volatile("ld 15,%0" : : "Q" (state->fprs[15]));
-		}
+		if (flags & KERNEL_VXR_LOW)
+			load_fp_regs_vx(vxrs);
 		return;
 	}
-
-	/* Test and restore (load) vector registers */
-	asm volatile (
-		/*
-		 * Test if any vector register must be loaded and, if so,
-		 * test if all registers can be loaded at once.
-		 */
-		"	la	1,%[vxrs]\n"	/* load restore area */
-		"	tmll	%[m],30\n"	/* KERNEL_VXR */
-		"	jz	7f\n"		/* no work -> done */
-		"	jo	5f\n"		/* -> restore V0..V31 */
-		/*
-		 * Test for special case KERNEL_FPU_MID only. In this
-		 * case a vlm V8..V23 is the best instruction
-		 */
-		"	chi	%[m],12\n"	/* KERNEL_VXR_MID */
-		"	jne	0f\n"		/* -> restore V8..V23 */
-		"	VLM	8,23,128,1\n"	/* vlm %v8,%v23,128(%r1) */
-		"	j	7f\n"
-		/* Test and restore the first half of 16 vector registers */
-		"0:	tmll	%[m],6\n"	/* KERNEL_VXR_LOW */
-		"	jz	3f\n"		/* -> KERNEL_VXR_HIGH */
-		"	jo	2f\n"		/* 11 -> restore V0..V15 */
-		"	brc	2,1f\n"		/* 10 -> restore V8..V15 */
-		"	VLM	0,7,0,1\n"	/* vlm %v0,%v7,0(%r1) */
-		"	j	3f\n"
-		"1:	VLM	8,15,128,1\n"	/* vlm %v8,%v15,128(%r1) */
-		"	j	3f\n"
-		"2:	VLM	0,15,0,1\n"	/* vlm %v0,%v15,0(%r1) */
-		/* Test and restore the second half of 16 vector registers */
-		"3:	tmll	%[m],24\n"	/* KERNEL_VXR_HIGH */
-		"	jz	7f\n"
-		"	jo	6f\n"		/* 11 -> restore V16..V31 */
-		"	brc	2,4f\n"		/* 10 -> restore V24..V31 */
-		"	VLM	16,23,256,1\n"	/* vlm %v16,%v23,256(%r1) */
-		"	j	7f\n"
-		"4:	VLM	24,31,384,1\n"	/* vlm %v24,%v31,384(%r1) */
-		"	j	7f\n"
-		"5:	VLM	0,15,0,1\n"	/* vlm %v0,%v15,0(%r1) */
-		"6:	VLM	16,31,256,1\n"	/* vlm %v16,%v31,256(%r1) */
-		"7:"
-		: [vxrs] "=Q" (*(struct vx_array *) &state->vxrs)
-		: [m] "d" (flags)
-		: "1", "cc");
+	mask = flags & KERNEL_VXR;
+	if (mask == KERNEL_VXR) {
+		vxrs += fpu_vlm(0, 15, vxrs);
+		vxrs += fpu_vlm(16, 31, vxrs);
+		return;
+	}
+	if (mask == KERNEL_VXR_MID) {
+		vxrs += fpu_vlm(8, 23, vxrs);
+		return;
+	}
+	mask = flags & KERNEL_VXR_LOW;
+	if (mask) {
+		if (mask == KERNEL_VXR_LOW)
+			vxrs += fpu_vlm(0, 15, vxrs);
+		else if (mask == KERNEL_VXR_V0V7)
+			vxrs += fpu_vlm(0, 7, vxrs);
+		else
+			vxrs += fpu_vlm(8, 15, vxrs);
+	}
+	mask = flags & KERNEL_VXR_HIGH;
+	if (mask) {
+		if (mask == KERNEL_VXR_HIGH)
+			vxrs += fpu_vlm(16, 31, vxrs);
+		else if (mask == KERNEL_VXR_V16V23)
+			vxrs += fpu_vlm(16, 23, vxrs);
+		else
+			vxrs += fpu_vlm(24, 31, vxrs);
+	}
 }
 EXPORT_SYMBOL(__kernel_fpu_end);
 
-void __load_fpu_regs(void)
+void load_fpu_state(struct fpu *state, int flags)
 {
-	unsigned long *regs = current->thread.fpu.regs;
-	struct fpu *state = &current->thread.fpu;
+	__vector128 *vxrs = &state->vxrs[0];
+	int mask;
 
-	sfpc_safe(state->fpc);
-	if (likely(cpu_has_vx())) {
-		asm volatile("lgr	1,%0\n"
-			     "VLM	0,15,0,1\n"
-			     "VLM	16,31,256,1\n"
-			     :
-			     : "d" (regs)
-			     : "1", "cc", "memory");
-	} else {
-		asm volatile("ld 0,%0" : : "Q" (regs[0]));
-		asm volatile("ld 1,%0" : : "Q" (regs[1]));
-		asm volatile("ld 2,%0" : : "Q" (regs[2]));
-		asm volatile("ld 3,%0" : : "Q" (regs[3]));
-		asm volatile("ld 4,%0" : : "Q" (regs[4]));
-		asm volatile("ld 5,%0" : : "Q" (regs[5]));
-		asm volatile("ld 6,%0" : : "Q" (regs[6]));
-		asm volatile("ld 7,%0" : : "Q" (regs[7]));
-		asm volatile("ld 8,%0" : : "Q" (regs[8]));
-		asm volatile("ld 9,%0" : : "Q" (regs[9]));
-		asm volatile("ld 10,%0" : : "Q" (regs[10]));
-		asm volatile("ld 11,%0" : : "Q" (regs[11]));
-		asm volatile("ld 12,%0" : : "Q" (regs[12]));
-		asm volatile("ld 13,%0" : : "Q" (regs[13]));
-		asm volatile("ld 14,%0" : : "Q" (regs[14]));
-		asm volatile("ld 15,%0" : : "Q" (regs[15]));
+	if (flags & KERNEL_FPC)
+		fpu_lfpc(&state->fpc);
+	if (!cpu_has_vx()) {
+		if (flags & KERNEL_VXR_V0V7)
+			load_fp_regs_vx(state->vxrs);
+		return;
+	}
+	mask = flags & KERNEL_VXR;
+	if (mask == KERNEL_VXR) {
+		fpu_vlm(0, 15, &vxrs[0]);
+		fpu_vlm(16, 31, &vxrs[16]);
+		return;
+	}
+	if (mask == KERNEL_VXR_MID) {
+		fpu_vlm(8, 23, &vxrs[8]);
+		return;
+	}
+	mask = flags & KERNEL_VXR_LOW;
+	if (mask) {
+		if (mask == KERNEL_VXR_LOW)
+			fpu_vlm(0, 15, &vxrs[0]);
+		else if (mask == KERNEL_VXR_V0V7)
+			fpu_vlm(0, 7, &vxrs[0]);
+		else
+			fpu_vlm(8, 15, &vxrs[8]);
+	}
+	mask = flags & KERNEL_VXR_HIGH;
+	if (mask) {
+		if (mask == KERNEL_VXR_HIGH)
+			fpu_vlm(16, 31, &vxrs[16]);
+		else if (mask == KERNEL_VXR_V16V23)
+			fpu_vlm(16, 23, &vxrs[16]);
+		else
+			fpu_vlm(24, 31, &vxrs[24]);
 	}
-	clear_cpu_flag(CIF_FPU);
-}
-
-void load_fpu_regs(void)
-{
-	raw_local_irq_disable();
-	__load_fpu_regs();
-	raw_local_irq_enable();
 }
-EXPORT_SYMBOL(load_fpu_regs);
 
-void save_fpu_regs(void)
+void save_fpu_state(struct fpu *state, int flags)
 {
-	unsigned long flags, *regs;
-	struct fpu *state;
-
-	local_irq_save(flags);
+	__vector128 *vxrs = &state->vxrs[0];
+	int mask;
 
-	if (test_cpu_flag(CIF_FPU))
-		goto out;
-
-	state = &current->thread.fpu;
-	regs = current->thread.fpu.regs;
-
-	asm volatile("stfpc %0" : "=Q" (state->fpc));
-	if (likely(cpu_has_vx())) {
-		asm volatile("lgr	1,%0\n"
-			     "VSTM	0,15,0,1\n"
-			     "VSTM	16,31,256,1\n"
-			     :
-			     : "d" (regs)
-			     : "1", "cc", "memory");
-	} else {
-		asm volatile("std 0,%0" : "=Q" (regs[0]));
-		asm volatile("std 1,%0" : "=Q" (regs[1]));
-		asm volatile("std 2,%0" : "=Q" (regs[2]));
-		asm volatile("std 3,%0" : "=Q" (regs[3]));
-		asm volatile("std 4,%0" : "=Q" (regs[4]));
-		asm volatile("std 5,%0" : "=Q" (regs[5]));
-		asm volatile("std 6,%0" : "=Q" (regs[6]));
-		asm volatile("std 7,%0" : "=Q" (regs[7]));
-		asm volatile("std 8,%0" : "=Q" (regs[8]));
-		asm volatile("std 9,%0" : "=Q" (regs[9]));
-		asm volatile("std 10,%0" : "=Q" (regs[10]));
-		asm volatile("std 11,%0" : "=Q" (regs[11]));
-		asm volatile("std 12,%0" : "=Q" (regs[12]));
-		asm volatile("std 13,%0" : "=Q" (regs[13]));
-		asm volatile("std 14,%0" : "=Q" (regs[14]));
-		asm volatile("std 15,%0" : "=Q" (regs[15]));
+	if (flags & KERNEL_FPC)
+		fpu_stfpc(&state->fpc);
+	if (!cpu_has_vx()) {
+		if (flags & KERNEL_VXR_LOW)
+			save_fp_regs_vx(state->vxrs);
+		return;
+	}
+	mask = flags & KERNEL_VXR;
+	if (mask == KERNEL_VXR) {
+		fpu_vstm(0, 15, &vxrs[0]);
+		fpu_vstm(16, 31, &vxrs[16]);
+		return;
+	}
+	if (mask == KERNEL_VXR_MID) {
+		fpu_vstm(8, 23, &vxrs[8]);
+		return;
+	}
+	mask = flags & KERNEL_VXR_LOW;
+	if (mask) {
+		if (mask == KERNEL_VXR_LOW)
+			fpu_vstm(0, 15, &vxrs[0]);
+		else if (mask == KERNEL_VXR_V0V7)
+			fpu_vstm(0, 7, &vxrs[0]);
+		else
+			fpu_vstm(8, 15, &vxrs[8]);
+	}
+	mask = flags & KERNEL_VXR_HIGH;
+	if (mask) {
+		if (mask == KERNEL_VXR_HIGH)
+			fpu_vstm(16, 31, &vxrs[16]);
+		else if (mask == KERNEL_VXR_V16V23)
+			fpu_vstm(16, 23, &vxrs[16]);
+		else
+			fpu_vstm(24, 31, &vxrs[24]);
 	}
-	set_cpu_flag(CIF_FPU);
-out:
-	local_irq_restore(flags);
 }
-EXPORT_SYMBOL(save_fpu_regs);
+EXPORT_SYMBOL(save_fpu_state);
diff --git a/arch/s390/kernel/ipl.c b/arch/s390/kernel/ipl.c
index ba75f6bee774..1486350a4177 100644
--- a/arch/s390/kernel/ipl.c
+++ b/arch/s390/kernel/ipl.c
@@ -1941,8 +1941,7 @@ static void dump_reipl_run(struct shutdown_trigger *trigger)
 	    reipl_type == IPL_TYPE_UNKNOWN)
 		os_info_flags |= OS_INFO_FLAG_REIPL_CLEAR;
 	os_info_entry_add(OS_INFO_FLAGS_ENTRY, &os_info_flags, sizeof(os_info_flags));
-	csum = (__force unsigned int)
-	       csum_partial(reipl_block_actual, reipl_block_actual->hdr.len, 0);
+	csum = (__force unsigned int)cksm(reipl_block_actual, reipl_block_actual->hdr.len, 0);
 	abs_lc = get_abs_lowcore();
 	abs_lc->ipib = __pa(reipl_block_actual);
 	abs_lc->ipib_checksum = csum;
diff --git a/arch/s390/kernel/machine_kexec.c b/arch/s390/kernel/machine_kexec.c
index aa22ffc16bcd..c5d0c1cf984b 100644
--- a/arch/s390/kernel/machine_kexec.c
+++ b/arch/s390/kernel/machine_kexec.c
@@ -13,8 +13,10 @@
 #include <linux/reboot.h>
 #include <linux/ftrace.h>
 #include <linux/debug_locks.h>
+#include <asm/guarded_storage.h>
 #include <asm/pfault.h>
 #include <asm/cio.h>
+#include <asm/fpu.h>
 #include <asm/setup.h>
 #include <asm/smp.h>
 #include <asm/ipl.h>
@@ -26,7 +28,6 @@
 #include <asm/os_info.h>
 #include <asm/set_memory.h>
 #include <asm/stacktrace.h>
-#include <asm/switch_to.h>
 #include <asm/nmi.h>
 #include <asm/sclp.h>
 
diff --git a/arch/s390/kernel/nmi.c b/arch/s390/kernel/nmi.c
index 9ad44c26d1a2..c77382a67325 100644
--- a/arch/s390/kernel/nmi.c
+++ b/arch/s390/kernel/nmi.c
@@ -23,16 +23,14 @@
 #include <linux/export.h>
 #include <asm/lowcore.h>
 #include <asm/ctlreg.h>
+#include <asm/fpu.h>
 #include <asm/smp.h>
 #include <asm/stp.h>
 #include <asm/cputime.h>
 #include <asm/nmi.h>
 #include <asm/crw.h>
-#include <asm/switch_to.h>
 #include <asm/asm-offsets.h>
 #include <asm/pai.h>
-#include <asm/vx-insn.h>
-#include <asm/fpu/api.h>
 
 struct mcck_struct {
 	unsigned int kill_task : 1;
@@ -204,133 +202,63 @@ void s390_handle_mcck(void)
 	}
 }
 
-/*
- * returns 0 if register contents could be validated
- * returns 1 otherwise
+/**
+ * nmi_registers_valid - verify if registers are valid
+ * @mci: machine check interruption code
+ *
+ * Inspect a machine check interruption code and verify if all required
+ * registers are valid. For some registers the corresponding validity bit is
+ * ignored and the registers are set to the expected value.
+ * Returns true if all registers are valid, otherwise false.
  */
-static int notrace s390_validate_registers(union mci mci)
+static bool notrace nmi_registers_valid(union mci mci)
 {
-	struct mcesa *mcesa;
-	void *fpt_save_area;
 	union ctlreg2 cr2;
-	int kill_task;
-	u64 zero;
-
-	kill_task = 0;
-	zero = 0;
-
-	if (!mci.gr || !mci.fp)
-		kill_task = 1;
-	fpt_save_area = &S390_lowcore.floating_pt_save_area;
-	if (!mci.fc) {
-		kill_task = 1;
-		asm volatile(
-			"	lfpc	%0\n"
-			:
-			: "Q" (zero));
-	} else {
-		asm volatile(
-			"	lfpc	%0\n"
-			:
-			: "Q" (S390_lowcore.fpt_creg_save_area));
-	}
 
-	mcesa = __va(S390_lowcore.mcesad & MCESA_ORIGIN_MASK);
-	if (!cpu_has_vx()) {
-		/* Validate floating point registers */
-		asm volatile(
-			"	ld	0,0(%0)\n"
-			"	ld	1,8(%0)\n"
-			"	ld	2,16(%0)\n"
-			"	ld	3,24(%0)\n"
-			"	ld	4,32(%0)\n"
-			"	ld	5,40(%0)\n"
-			"	ld	6,48(%0)\n"
-			"	ld	7,56(%0)\n"
-			"	ld	8,64(%0)\n"
-			"	ld	9,72(%0)\n"
-			"	ld	10,80(%0)\n"
-			"	ld	11,88(%0)\n"
-			"	ld	12,96(%0)\n"
-			"	ld	13,104(%0)\n"
-			"	ld	14,112(%0)\n"
-			"	ld	15,120(%0)\n"
-			:
-			: "a" (fpt_save_area)
-			: "memory");
-	} else {
-		/* Validate vector registers */
-		union ctlreg0 cr0;
-
-		/*
-		 * The vector validity must only be checked if not running a
-		 * KVM guest. For KVM guests the machine check is forwarded by
-		 * KVM and it is the responsibility of the guest to take
-		 * appropriate actions. The host vector or FPU values have been
-		 * saved by KVM and will be restored by KVM.
-		 */
-		if (!mci.vr && !test_cpu_flag(CIF_MCCK_GUEST))
-			kill_task = 1;
-		cr0.reg = S390_lowcore.cregs_save_area[0];
-		cr0.afp = cr0.vx = 1;
-		local_ctl_load(0, &cr0.reg);
-		asm volatile(
-			"	la	1,%0\n"
-			"	VLM	0,15,0,1\n"
-			"	VLM	16,31,256,1\n"
-			:
-			: "Q" (*(struct vx_array *)mcesa->vector_save_area)
-			: "1");
-		local_ctl_load(0, &S390_lowcore.cregs_save_area[0]);
-	}
-	/* Validate access registers */
-	asm volatile(
-		"	lam	0,15,0(%0)\n"
-		:
-		: "a" (&S390_lowcore.access_regs_save_area)
-		: "memory");
-	if (!mci.ar)
-		kill_task = 1;
-	/* Validate guarded storage registers */
-	cr2.reg = S390_lowcore.cregs_save_area[2];
-	if (cr2.gse) {
-		if (!mci.gs) {
-			/*
-			 * 2 cases:
-			 * - machine check in kernel or userspace
-			 * - machine check while running SIE (KVM guest)
-			 * For kernel or userspace the userspace values of
-			 * guarded storage control can not be recreated, the
-			 * process must be terminated.
-			 * For SIE the guest values of guarded storage can not
-			 * be recreated. This is either due to a bug or due to
-			 * GS being disabled in the guest. The guest will be
-			 * notified by KVM code and the guests machine check
-			 * handling must take care of this.  The host values
-			 * are saved by KVM and are not affected.
-			 */
-			if (!test_cpu_flag(CIF_MCCK_GUEST))
-				kill_task = 1;
-		} else {
-			load_gs_cb((struct gs_cb *)mcesa->guarded_storage_save_area);
-		}
-	}
 	/*
-	 * The getcpu vdso syscall reads CPU number from the programmable
+	 * The getcpu vdso syscall reads the CPU number from the programmable
 	 * field of the TOD clock. Disregard the TOD programmable register
-	 * validity bit and load the CPU number into the TOD programmable
-	 * field unconditionally.
+	 * validity bit and load the CPU number into the TOD programmable field
+	 * unconditionally.
 	 */
 	set_tod_programmable_field(raw_smp_processor_id());
-	/* Validate clock comparator register */
+	/*
+	 * Set the clock comparator register to the next expected value.
+	 */
 	set_clock_comparator(S390_lowcore.clock_comparator);
-
+	if (!mci.gr || !mci.fp || !mci.fc)
+		return false;
+	/*
+	 * The vector validity must only be checked if not running a
+	 * KVM guest. For KVM guests the machine check is forwarded by
+	 * KVM and it is the responsibility of the guest to take
+	 * appropriate actions. The host vector or FPU values have been
+	 * saved by KVM and will be restored by KVM.
+	 */
+	if (!mci.vr && !test_cpu_flag(CIF_MCCK_GUEST))
+		return false;
+	if (!mci.ar)
+		return false;
+	/*
+	 * Two cases for guarded storage registers:
+	 * - machine check in kernel or userspace
+	 * - machine check while running SIE (KVM guest)
+	 * For kernel or userspace the userspace values of guarded storage
+	 * control can not be recreated, the process must be terminated.
+	 * For SIE the guest values of guarded storage can not be recreated.
+	 * This is either due to a bug or due to GS being disabled in the
+	 * guest. The guest will be notified by KVM code and the guests machine
+	 * 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];
+	if (cr2.gse && !mci.gs && !test_cpu_flag(CIF_MCCK_GUEST))
+		return false;
 	if (!mci.ms || !mci.pm || !mci.ia)
-		kill_task = 1;
-
-	return kill_task;
+		return false;
+	return true;
 }
-NOKPROBE_SYMBOL(s390_validate_registers);
+NOKPROBE_SYMBOL(nmi_registers_valid);
 
 /*
  * Backup the guest's machine check info to its description block
@@ -428,7 +356,7 @@ void notrace s390_do_machine_check(struct pt_regs *regs)
 			s390_handle_damage();
 		}
 	}
-	if (s390_validate_registers(mci)) {
+	if (!nmi_registers_valid(mci)) {
 		if (!user_mode(regs))
 			s390_handle_damage();
 		/*
diff --git a/arch/s390/kernel/os_info.c b/arch/s390/kernel/os_info.c
index 6e1824141b29..a801e6bd5341 100644
--- a/arch/s390/kernel/os_info.c
+++ b/arch/s390/kernel/os_info.c
@@ -29,7 +29,7 @@ static struct os_info os_info __page_aligned_data;
 u32 os_info_csum(struct os_info *os_info)
 {
 	int size = sizeof(*os_info) - offsetof(struct os_info, version_major);
-	return (__force u32)csum_partial(&os_info->version_major, size, 0);
+	return (__force u32)cksm(&os_info->version_major, size, 0);
 }
 
 /*
@@ -49,7 +49,7 @@ void os_info_entry_add(int nr, void *ptr, u64 size)
 {
 	os_info.entry[nr].addr = __pa(ptr);
 	os_info.entry[nr].size = size;
-	os_info.entry[nr].csum = (__force u32)csum_partial(ptr, size, 0);
+	os_info.entry[nr].csum = (__force u32)cksm(ptr, size, 0);
 	os_info.csum = os_info_csum(&os_info);
 }
 
@@ -98,7 +98,7 @@ static void os_info_old_alloc(int nr, int align)
 		msg = "copy failed";
 		goto fail_free;
 	}
-	csum = (__force u32)csum_partial(buf_align, size, 0);
+	csum = (__force u32)cksm(buf_align, size, 0);
 	if (csum != os_info_old->entry[nr].csum) {
 		msg = "checksum failed";
 		goto fail_free;
diff --git a/arch/s390/kernel/perf_pai_crypto.c b/arch/s390/kernel/perf_pai_crypto.c
index bf8a672b15a4..823d652e3917 100644
--- a/arch/s390/kernel/perf_pai_crypto.c
+++ b/arch/s390/kernel/perf_pai_crypto.c
@@ -98,6 +98,7 @@ static void paicrypt_event_destroy(struct perf_event *event)
 			    event->attr.config, event->cpu,
 			    cpump->active_events, cpump->mode,
 			    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,
@@ -260,6 +261,7 @@ static int paicrypt_event_init(struct perf_event *event)
 {
 	struct perf_event_attr *a = &event->attr;
 	struct paicrypt_map *cpump;
+	int rc = 0;
 
 	/* PAI crypto PMU registered as PERF_TYPE_RAW, check event type */
 	if (a->type != PERF_TYPE_RAW && event->pmu->type != a->type)
@@ -274,10 +276,21 @@ static int paicrypt_event_init(struct perf_event *event)
 	/* 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 */
+	if (a->sample_period) {
+		PAI_SAVE_AREA(event) = get_zeroed_page(GFP_KERNEL);
+		if (!PAI_SAVE_AREA(event)) {
+			rc = -ENOMEM;
+			goto out;
+		}
+	}
 
 	cpump = paicrypt_busy(event);
-	if (IS_ERR(cpump))
-		return PTR_ERR(cpump);
+	if (IS_ERR(cpump)) {
+		free_page(PAI_SAVE_AREA(event));
+		rc = PTR_ERR(cpump);
+		goto out;
+	}
 
 	event->destroy = paicrypt_event_destroy;
 
@@ -293,7 +306,8 @@ static int paicrypt_event_init(struct perf_event *event)
 	}
 
 	static_branch_inc(&pai_key);
-	return 0;
+out:
+	return rc;
 }
 
 static void paicrypt_read(struct perf_event *event)
@@ -310,20 +324,15 @@ static void paicrypt_read(struct perf_event *event)
 
 static void paicrypt_start(struct perf_event *event, int flags)
 {
+	struct paicrypt_mapptr *mp = this_cpu_ptr(paicrypt_root.mapptr);
+	struct paicrypt_map *cpump = mp->mapptr;
 	u64 sum;
 
-	/* Event initialization sets last_tag to 0. When later on the events
-	 * are deleted and re-added, do not reset the event count value to zero.
-	 * Events are added, deleted and re-added when 2 or more events
-	 * are active at the same time.
-	 */
 	if (!event->attr.sample_period) {	/* Counting */
-		if (!event->hw.last_tag) {
-			event->hw.last_tag = 1;
-			sum = paicrypt_getall(event);	/* Get current value */
-			local64_set(&event->hw.prev_count, sum);
-		}
+		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);
 	}
 }
@@ -339,7 +348,6 @@ static int paicrypt_add(struct perf_event *event, int flags)
 		WRITE_ONCE(S390_lowcore.ccd, ccd);
 		local_ctl_set_bit(0, CR0_CRYPTOGRAPHY_COUNTER_BIT);
 	}
-	cpump->event = event;
 	if (flags & PERF_EF_START)
 		paicrypt_start(event, PERF_EF_RELOAD);
 	event->hw.state = 0;
@@ -367,23 +375,34 @@ static void paicrypt_del(struct perf_event *event, int flags)
 	}
 }
 
-/* Create raw data and save it in buffer. Returns number of bytes copied.
- * Saves only positive counter entries of the form
+/* Create raw data and save it in buffer. Calculate the delta for each
+ * counter between this invocation and the last invocation.
+ * Returns number of bytes copied.
+ * Saves only entries with positive counter difference of the form
  * 2 bytes: Number of counter
  * 8 bytes: Value of counter
  */
 static size_t paicrypt_copy(struct pai_userdata *userdata, unsigned long *page,
-			    bool exclude_user, bool exclude_kernel)
+			    unsigned long *page_old, bool exclude_user,
+			    bool exclude_kernel)
 {
 	int i, outidx = 0;
 
 	for (i = 1; i <= paicrypt_cnt; i++) {
-		u64 val = 0;
+		u64 val = 0, val_old = 0;
 
-		if (!exclude_kernel)
+		if (!exclude_kernel) {
 			val += paicrypt_getctr(page, i, true);
-		if (!exclude_user)
+			val_old += paicrypt_getctr(page_old, i, true);
+		}
+		if (!exclude_user) {
 			val += paicrypt_getctr(page, i, false);
+			val_old += paicrypt_getctr(page_old, i, false);
+		}
+		if (val >= val_old)
+			val -= val_old;
+		else
+			val = (~0ULL - val_old) + val + 1;
 		if (val) {
 			userdata[outidx].num = i;
 			userdata[outidx].value = val;
@@ -426,8 +445,8 @@ static int paicrypt_push_sample(size_t rawsize, struct paicrypt_map *cpump,
 
 	overflow = perf_event_overflow(event, &data, &regs);
 	perf_event_update_userpage(event);
-	/* Clear lowcore page after read */
-	memset(cpump->page, 0, PAGE_SIZE);
+	/* Save crypto counter lowcore page after reading event data. */
+	memcpy((void *)PAI_SAVE_AREA(event), cpump->page, PAGE_SIZE);
 	return overflow;
 }
 
@@ -443,6 +462,7 @@ static int paicrypt_have_sample(void)
 	if (!event)		/* No event active */
 		return 0;
 	rawsize = paicrypt_copy(cpump->save, cpump->page,
+				(unsigned long *)PAI_SAVE_AREA(event),
 				cpump->event->attr.exclude_user,
 				cpump->event->attr.exclude_kernel);
 	if (rawsize)			/* No incremented counters */
@@ -694,6 +714,12 @@ static int __init attr_event_init_one(struct attribute **attrs, int num)
 {
 	struct perf_pmu_events_attr *pa;
 
+	/* Index larger than array_size, no counter name available */
+	if (num >= ARRAY_SIZE(paicrypt_ctrnames)) {
+		attrs[num] = NULL;
+		return 0;
+	}
+
 	pa = kzalloc(sizeof(*pa), GFP_KERNEL);
 	if (!pa)
 		return -ENOMEM;
@@ -714,14 +740,13 @@ static int __init attr_event_init(void)
 	struct attribute **attrs;
 	int ret, i;
 
-	attrs = kmalloc_array(ARRAY_SIZE(paicrypt_ctrnames) + 1, sizeof(*attrs),
-			      GFP_KERNEL);
+	attrs = kmalloc_array(paicrypt_cnt + 2, sizeof(*attrs), GFP_KERNEL);
 	if (!attrs)
 		return -ENOMEM;
-	for (i = 0; i < ARRAY_SIZE(paicrypt_ctrnames); i++) {
+	for (i = 0; i <= paicrypt_cnt; i++) {
 		ret = attr_event_init_one(attrs, i);
 		if (ret) {
-			attr_event_free(attrs, i - 1);
+			attr_event_free(attrs, i);
 			return ret;
 		}
 	}
@@ -742,8 +767,10 @@ static int __init paicrypt_init(void)
 	paicrypt_cnt = ib.num_cc;
 	if (paicrypt_cnt == 0)
 		return 0;
-	if (paicrypt_cnt >= PAI_CRYPTO_MAXCTR)
-		paicrypt_cnt = PAI_CRYPTO_MAXCTR - 1;
+	if (paicrypt_cnt >= PAI_CRYPTO_MAXCTR) {
+		pr_err("Too many PMU pai_crypto counters %d\n", paicrypt_cnt);
+		return -E2BIG;
+	}
 
 	rc = attr_event_init();		/* Export known PAI crypto events */
 	if (rc) {
diff --git a/arch/s390/kernel/perf_pai_ext.c b/arch/s390/kernel/perf_pai_ext.c
index af7f2b538c8f..616a25606cd6 100644
--- a/arch/s390/kernel/perf_pai_ext.c
+++ b/arch/s390/kernel/perf_pai_ext.c
@@ -120,6 +120,7 @@ static void paiext_event_destroy(struct perf_event *event)
 	struct paiext_mapptr *mp = per_cpu_ptr(paiext_root.mapptr, event->cpu);
 	struct paiext_map *cpump = mp->mapptr;
 
+	free_page(PAI_SAVE_AREA(event));
 	mutex_lock(&paiext_reserve_mutex);
 	cpump->event = NULL;
 	if (refcount_dec_and_test(&cpump->refcnt))	/* Last reference gone */
@@ -202,7 +203,6 @@ static int paiext_alloc(struct perf_event_attr *a, struct perf_event *event)
 	}
 
 	rc = 0;
-	cpump->event = event;
 
 undo:
 	if (rc) {
@@ -256,10 +256,18 @@ static int paiext_event_init(struct perf_event *event)
 	/* Prohibit exclude_user event selection */
 	if (a->exclude_user)
 		return -EINVAL;
+	/* Get a page to store last counter values for sampling */
+	if (a->sample_period) {
+		PAI_SAVE_AREA(event) = get_zeroed_page(GFP_KERNEL);
+		if (!PAI_SAVE_AREA(event))
+			return -ENOMEM;
+	}
 
 	rc = paiext_alloc(a, event);
-	if (rc)
+	if (rc) {
+		free_page(PAI_SAVE_AREA(event));
 		return rc;
+	}
 	event->destroy = paiext_event_destroy;
 
 	if (a->sample_period) {
@@ -319,15 +327,15 @@ static void paiext_read(struct perf_event *event)
 
 static void paiext_start(struct perf_event *event, int flags)
 {
+	struct paiext_mapptr *mp = this_cpu_ptr(paiext_root.mapptr);
+	struct paiext_map *cpump = mp->mapptr;
 	u64 sum;
 
 	if (!event->attr.sample_period) {	/* Counting */
-		if (!event->hw.last_tag) {
-			event->hw.last_tag = 1;
-			sum = paiext_getall(event);	/* Get current value */
-			local64_set(&event->hw.prev_count, sum);
-		}
+		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);
 	}
 }
@@ -346,7 +354,6 @@ static int paiext_add(struct perf_event *event, int flags)
 		debug_sprintf_event(paiext_dbg, 4, "%s 1508 %llx acc %llx\n",
 				    __func__, S390_lowcore.aicd, pcb->acc);
 	}
-	cpump->event = event;
 	if (flags & PERF_EF_START)
 		paiext_start(event, PERF_EF_RELOAD);
 	event->hw.state = 0;
@@ -384,13 +391,19 @@ static void paiext_del(struct perf_event *event, int flags)
  * 2 bytes: Number of counter
  * 8 bytes: Value of counter
  */
-static size_t paiext_copy(struct pai_userdata *userdata, unsigned long *area)
+static size_t paiext_copy(struct pai_userdata *userdata, unsigned long *area,
+			  unsigned long *area_old)
 {
 	int i, outidx = 0;
 
 	for (i = 1; i <= paiext_cnt; i++) {
 		u64 val = paiext_getctr(area, i);
+		u64 val_old = paiext_getctr(area_old, i);
 
+		if (val >= val_old)
+			val -= val_old;
+		else
+			val = (~0ULL - val_old) + val + 1;
 		if (val) {
 			userdata[outidx].num = i;
 			userdata[outidx].value = val;
@@ -446,8 +459,9 @@ static int paiext_push_sample(size_t rawsize, struct paiext_map *cpump,
 
 	overflow = perf_event_overflow(event, &data, &regs);
 	perf_event_update_userpage(event);
-	/* Clear lowcore area after read */
-	memset(cpump->area, 0, PAIE1_CTRBLOCK_SZ);
+	/* Save NNPA lowcore area after read in event */
+	memcpy((void *)PAI_SAVE_AREA(event), cpump->area,
+	       PAIE1_CTRBLOCK_SZ);
 	return overflow;
 }
 
@@ -462,7 +476,8 @@ static int paiext_have_sample(void)
 
 	if (!event)
 		return 0;
-	rawsize = paiext_copy(cpump->save, cpump->area);
+	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;
@@ -584,6 +599,12 @@ static int __init attr_event_init_one(struct attribute **attrs, int num)
 {
 	struct perf_pmu_events_attr *pa;
 
+	/* Index larger than array_size, no counter name available */
+	if (num >= ARRAY_SIZE(paiext_ctrnames)) {
+		attrs[num] = NULL;
+		return 0;
+	}
+
 	pa = kzalloc(sizeof(*pa), GFP_KERNEL);
 	if (!pa)
 		return -ENOMEM;
@@ -604,14 +625,13 @@ static int __init attr_event_init(void)
 	struct attribute **attrs;
 	int ret, i;
 
-	attrs = kmalloc_array(ARRAY_SIZE(paiext_ctrnames) + 1, sizeof(*attrs),
-			      GFP_KERNEL);
+	attrs = kmalloc_array(paiext_cnt + 2, sizeof(*attrs), GFP_KERNEL);
 	if (!attrs)
 		return -ENOMEM;
-	for (i = 0; i < ARRAY_SIZE(paiext_ctrnames); i++) {
+	for (i = 0; i <= paiext_cnt; i++) {
 		ret = attr_event_init_one(attrs, i);
 		if (ret) {
-			attr_event_free(attrs, i - 1);
+			attr_event_free(attrs, i);
 			return ret;
 		}
 	}
diff --git a/arch/s390/kernel/perf_regs.c b/arch/s390/kernel/perf_regs.c
index 3d93656bd948..a6b058ee4a36 100644
--- a/arch/s390/kernel/perf_regs.c
+++ b/arch/s390/kernel/perf_regs.c
@@ -5,8 +5,7 @@
 #include <linux/errno.h>
 #include <linux/bug.h>
 #include <asm/ptrace.h>
-#include <asm/fpu/api.h>
-#include <asm/fpu/types.h>
+#include <asm/fpu.h>
 
 u64 perf_reg_value(struct pt_regs *regs, int idx)
 {
@@ -20,10 +19,7 @@ u64 perf_reg_value(struct pt_regs *regs, int idx)
 			return 0;
 
 		idx -= PERF_REG_S390_FP0;
-		if (cpu_has_vx())
-			fp = *(freg_t *)(current->thread.fpu.vxrs + idx);
-		else
-			fp = current->thread.fpu.fprs[idx];
+		fp = *(freg_t *)(current->thread.ufpu.vxrs + idx);
 		return fp.ui;
 	}
 
@@ -65,6 +61,6 @@ void perf_get_regs_user(struct perf_regs *regs_user,
 	 */
 	regs_user->regs = task_pt_regs(current);
 	if (user_mode(regs_user->regs))
-		save_fpu_regs();
+		save_user_fpu_regs();
 	regs_user->abi = perf_reg_abi(current);
 }
diff --git a/arch/s390/kernel/process.c b/arch/s390/kernel/process.c
index 4e3b366589fb..dd456b475861 100644
--- a/arch/s390/kernel/process.c
+++ b/arch/s390/kernel/process.c
@@ -31,15 +31,19 @@
 #include <linux/init_task.h>
 #include <linux/entry-common.h>
 #include <linux/io.h>
+#include <asm/guarded_storage.h>
+#include <asm/access-regs.h>
+#include <asm/switch_to.h>
 #include <asm/cpu_mf.h>
 #include <asm/processor.h>
+#include <asm/ptrace.h>
 #include <asm/vtimer.h>
 #include <asm/exec.h>
+#include <asm/fpu.h>
 #include <asm/irq.h>
 #include <asm/nmi.h>
 #include <asm/smp.h>
 #include <asm/stacktrace.h>
-#include <asm/switch_to.h>
 #include <asm/runtime_instr.h>
 #include <asm/unwind.h>
 #include "entry.h"
@@ -84,13 +88,13 @@ int arch_dup_task_struct(struct task_struct *dst, struct task_struct *src)
 {
 	/*
 	 * Save the floating-point or vector register state of the current
-	 * task and set the CIF_FPU flag to lazy restore the FPU register
+	 * task and set the TIF_FPU flag to lazy restore the FPU register
 	 * state when returning to user space.
 	 */
-	save_fpu_regs();
+	save_user_fpu_regs();
 
 	*dst = *src;
-	dst->thread.fpu.regs = dst->thread.fpu.fprs;
+	dst->thread.kfpu_flags = 0;
 
 	/*
 	 * Don't transfer over the runtime instrumentation or the guarded
@@ -186,8 +190,23 @@ int copy_thread(struct task_struct *p, const struct kernel_clone_args *args)
 
 void execve_tail(void)
 {
-	current->thread.fpu.fpc = 0;
-	asm volatile("sfpc %0" : : "d" (0));
+	current->thread.ufpu.fpc = 0;
+	fpu_sfpc(0);
+}
+
+struct task_struct *__switch_to(struct task_struct *prev, struct task_struct *next)
+{
+	save_user_fpu_regs();
+	save_kernel_fpu_regs(&prev->thread);
+	save_access_regs(&prev->thread.acrs[0]);
+	save_ri_cb(prev->thread.ri_cb);
+	save_gs_cb(prev->thread.gs_cb);
+	update_cr_regs(next);
+	restore_kernel_fpu_regs(&next->thread);
+	restore_access_regs(&next->thread.acrs[0]);
+	restore_ri_cb(next->thread.ri_cb, prev->thread.ri_cb);
+	restore_gs_cb(next->thread.gs_cb);
+	return __switch_to_asm(prev, next);
 }
 
 unsigned long __get_wchan(struct task_struct *p)
diff --git a/arch/s390/kernel/ptrace.c b/arch/s390/kernel/ptrace.c
index f1897a8bb221..1cfed8b710b8 100644
--- a/arch/s390/kernel/ptrace.c
+++ b/arch/s390/kernel/ptrace.c
@@ -24,13 +24,14 @@
 #include <linux/seccomp.h>
 #include <linux/compat.h>
 #include <trace/syscall.h>
+#include <asm/guarded_storage.h>
+#include <asm/access-regs.h>
 #include <asm/page.h>
 #include <linux/uaccess.h>
 #include <asm/unistd.h>
-#include <asm/switch_to.h>
 #include <asm/runtime_instr.h>
 #include <asm/facility.h>
-#include <asm/fpu/api.h>
+#include <asm/fpu.h>
 
 #include "entry.h"
 
@@ -246,22 +247,15 @@ static unsigned long __peek_user(struct task_struct *child, addr_t addr)
 		/*
 		 * floating point control reg. is in the thread structure
 		 */
-		tmp = child->thread.fpu.fpc;
+		tmp = child->thread.ufpu.fpc;
 		tmp <<= BITS_PER_LONG - 32;
 
 	} else if (addr < offsetof(struct user, regs.fp_regs) + sizeof(s390_fp_regs)) {
 		/*
-		 * floating point regs. are either in child->thread.fpu
-		 * or the child->thread.fpu.vxrs array
+		 * floating point regs. are in the child->thread.ufpu.vxrs array
 		 */
 		offset = addr - offsetof(struct user, regs.fp_regs.fprs);
-		if (cpu_has_vx())
-			tmp = *(addr_t *)
-			       ((addr_t) child->thread.fpu.vxrs + 2*offset);
-		else
-			tmp = *(addr_t *)
-			       ((addr_t) child->thread.fpu.fprs + offset);
-
+		tmp = *(addr_t *)((addr_t)child->thread.ufpu.vxrs + 2 * offset);
 	} else if (addr < offsetof(struct user, regs.per_info) + sizeof(per_struct)) {
 		/*
 		 * Handle access to the per_info structure.
@@ -395,21 +389,14 @@ static int __poke_user(struct task_struct *child, addr_t addr, addr_t data)
 		 */
 		if ((unsigned int)data != 0)
 			return -EINVAL;
-		child->thread.fpu.fpc = data >> (BITS_PER_LONG - 32);
+		child->thread.ufpu.fpc = data >> (BITS_PER_LONG - 32);
 
 	} else if (addr < offsetof(struct user, regs.fp_regs) + sizeof(s390_fp_regs)) {
 		/*
-		 * floating point regs. are either in child->thread.fpu
-		 * or the child->thread.fpu.vxrs array
+		 * floating point regs. are in the child->thread.ufpu.vxrs array
 		 */
 		offset = addr - offsetof(struct user, regs.fp_regs.fprs);
-		if (cpu_has_vx())
-			*(addr_t *)((addr_t)
-				child->thread.fpu.vxrs + 2*offset) = data;
-		else
-			*(addr_t *)((addr_t)
-				child->thread.fpu.fprs + offset) = data;
-
+		*(addr_t *)((addr_t)child->thread.ufpu.vxrs + 2 * offset) = data;
 	} else if (addr < offsetof(struct user, regs.per_info) + sizeof(per_struct)) {
 		/*
 		 * Handle access to the per_info structure.
@@ -622,21 +609,14 @@ static u32 __peek_user_compat(struct task_struct *child, addr_t addr)
 		/*
 		 * floating point control reg. is in the thread structure
 		 */
-		tmp = child->thread.fpu.fpc;
+		tmp = child->thread.ufpu.fpc;
 
 	} else if (addr < offsetof(struct compat_user, regs.fp_regs) + sizeof(s390_fp_regs)) {
 		/*
-		 * floating point regs. are either in child->thread.fpu
-		 * or the child->thread.fpu.vxrs array
+		 * floating point regs. are in the child->thread.ufpu.vxrs array
 		 */
 		offset = addr - offsetof(struct compat_user, regs.fp_regs.fprs);
-		if (cpu_has_vx())
-			tmp = *(__u32 *)
-			       ((addr_t) child->thread.fpu.vxrs + 2*offset);
-		else
-			tmp = *(__u32 *)
-			       ((addr_t) child->thread.fpu.fprs + offset);
-
+		tmp = *(__u32 *)((addr_t)child->thread.ufpu.vxrs + 2 * offset);
 	} else if (addr < offsetof(struct compat_user, regs.per_info) + sizeof(struct compat_per_struct_kernel)) {
 		/*
 		 * Handle access to the per_info structure.
@@ -748,21 +728,14 @@ static int __poke_user_compat(struct task_struct *child,
 		/*
 		 * floating point control reg. is in the thread structure
 		 */
-		child->thread.fpu.fpc = data;
+		child->thread.ufpu.fpc = data;
 
 	} else if (addr < offsetof(struct compat_user, regs.fp_regs) + sizeof(s390_fp_regs)) {
 		/*
-		 * floating point regs. are either in child->thread.fpu
-		 * or the child->thread.fpu.vxrs array
+		 * floating point regs. are in the child->thread.ufpu.vxrs array
 		 */
 		offset = addr - offsetof(struct compat_user, regs.fp_regs.fprs);
-		if (cpu_has_vx())
-			*(__u32 *)((addr_t)
-				child->thread.fpu.vxrs + 2*offset) = tmp;
-		else
-			*(__u32 *)((addr_t)
-				child->thread.fpu.fprs + offset) = tmp;
-
+		*(__u32 *)((addr_t)child->thread.ufpu.vxrs + 2 * offset) = tmp;
 	} else if (addr < offsetof(struct compat_user, regs.per_info) + sizeof(struct compat_per_struct_kernel)) {
 		/*
 		 * Handle access to the per_info structure.
@@ -893,10 +866,10 @@ static int s390_fpregs_get(struct task_struct *target,
 	_s390_fp_regs fp_regs;
 
 	if (target == current)
-		save_fpu_regs();
+		save_user_fpu_regs();
 
-	fp_regs.fpc = target->thread.fpu.fpc;
-	fpregs_store(&fp_regs, &target->thread.fpu);
+	fp_regs.fpc = target->thread.ufpu.fpc;
+	fpregs_store(&fp_regs, &target->thread.ufpu);
 
 	return membuf_write(&to, &fp_regs, sizeof(fp_regs));
 }
@@ -910,22 +883,17 @@ static int s390_fpregs_set(struct task_struct *target,
 	freg_t fprs[__NUM_FPRS];
 
 	if (target == current)
-		save_fpu_regs();
-
-	if (cpu_has_vx())
-		convert_vx_to_fp(fprs, target->thread.fpu.vxrs);
-	else
-		memcpy(&fprs, target->thread.fpu.fprs, sizeof(fprs));
-
+		save_user_fpu_regs();
+	convert_vx_to_fp(fprs, target->thread.ufpu.vxrs);
 	if (count > 0 && pos < offsetof(s390_fp_regs, fprs)) {
-		u32 ufpc[2] = { target->thread.fpu.fpc, 0 };
+		u32 ufpc[2] = { target->thread.ufpu.fpc, 0 };
 		rc = user_regset_copyin(&pos, &count, &kbuf, &ubuf, &ufpc,
 					0, offsetof(s390_fp_regs, fprs));
 		if (rc)
 			return rc;
 		if (ufpc[1] != 0)
 			return -EINVAL;
-		target->thread.fpu.fpc = ufpc[0];
+		target->thread.ufpu.fpc = ufpc[0];
 	}
 
 	if (rc == 0 && count > 0)
@@ -933,12 +901,7 @@ static int s390_fpregs_set(struct task_struct *target,
 					fprs, offsetof(s390_fp_regs, fprs), -1);
 	if (rc)
 		return rc;
-
-	if (cpu_has_vx())
-		convert_fp_to_vx(target->thread.fpu.vxrs, fprs);
-	else
-		memcpy(target->thread.fpu.fprs, &fprs, sizeof(fprs));
-
+	convert_fp_to_vx(target->thread.ufpu.vxrs, fprs);
 	return rc;
 }
 
@@ -988,9 +951,9 @@ static int s390_vxrs_low_get(struct task_struct *target,
 	if (!cpu_has_vx())
 		return -ENODEV;
 	if (target == current)
-		save_fpu_regs();
+		save_user_fpu_regs();
 	for (i = 0; i < __NUM_VXRS_LOW; i++)
-		vxrs[i] = target->thread.fpu.vxrs[i].low;
+		vxrs[i] = target->thread.ufpu.vxrs[i].low;
 	return membuf_write(&to, vxrs, sizeof(vxrs));
 }
 
@@ -1005,15 +968,15 @@ static int s390_vxrs_low_set(struct task_struct *target,
 	if (!cpu_has_vx())
 		return -ENODEV;
 	if (target == current)
-		save_fpu_regs();
+		save_user_fpu_regs();
 
 	for (i = 0; i < __NUM_VXRS_LOW; i++)
-		vxrs[i] = target->thread.fpu.vxrs[i].low;
+		vxrs[i] = target->thread.ufpu.vxrs[i].low;
 
 	rc = user_regset_copyin(&pos, &count, &kbuf, &ubuf, vxrs, 0, -1);
 	if (rc == 0)
 		for (i = 0; i < __NUM_VXRS_LOW; i++)
-			target->thread.fpu.vxrs[i].low = vxrs[i];
+			target->thread.ufpu.vxrs[i].low = vxrs[i];
 
 	return rc;
 }
@@ -1025,8 +988,8 @@ static int s390_vxrs_high_get(struct task_struct *target,
 	if (!cpu_has_vx())
 		return -ENODEV;
 	if (target == current)
-		save_fpu_regs();
-	return membuf_write(&to, target->thread.fpu.vxrs + __NUM_VXRS_LOW,
+		save_user_fpu_regs();
+	return membuf_write(&to, target->thread.ufpu.vxrs + __NUM_VXRS_LOW,
 			    __NUM_VXRS_HIGH * sizeof(__vector128));
 }
 
@@ -1040,10 +1003,10 @@ static int s390_vxrs_high_set(struct task_struct *target,
 	if (!cpu_has_vx())
 		return -ENODEV;
 	if (target == current)
-		save_fpu_regs();
+		save_user_fpu_regs();
 
 	rc = user_regset_copyin(&pos, &count, &kbuf, &ubuf,
-				target->thread.fpu.vxrs + __NUM_VXRS_LOW, 0, -1);
+				target->thread.ufpu.vxrs + __NUM_VXRS_LOW, 0, -1);
 	return rc;
 }
 
diff --git a/arch/s390/kernel/setup.c b/arch/s390/kernel/setup.c
index d1f3b56e7afc..24ed33f044ec 100644
--- a/arch/s390/kernel/setup.c
+++ b/arch/s390/kernel/setup.c
@@ -504,12 +504,12 @@ static void __init setup_resources(void)
 	int j;
 	u64 i;
 
-	code_resource.start = (unsigned long) _text;
-	code_resource.end = (unsigned long) _etext - 1;
-	data_resource.start = (unsigned long) _etext;
-	data_resource.end = (unsigned long) _edata - 1;
-	bss_resource.start = (unsigned long) __bss_start;
-	bss_resource.end = (unsigned long) __bss_stop - 1;
+	code_resource.start = __pa_symbol(_text);
+	code_resource.end = __pa_symbol(_etext) - 1;
+	data_resource.start = __pa_symbol(_etext);
+	data_resource.end = __pa_symbol(_edata) - 1;
+	bss_resource.start = __pa_symbol(__bss_start);
+	bss_resource.end = __pa_symbol(__bss_stop) - 1;
 
 	for_each_mem_range(i, &start, &end) {
 		res = memblock_alloc(sizeof(*res), 8);
diff --git a/arch/s390/kernel/signal.c b/arch/s390/kernel/signal.c
index 43e9661cd715..6c2cb345402f 100644
--- a/arch/s390/kernel/signal.c
+++ b/arch/s390/kernel/signal.c
@@ -30,8 +30,8 @@
 #include <linux/compat.h>
 #include <asm/ucontext.h>
 #include <linux/uaccess.h>
+#include <asm/access-regs.h>
 #include <asm/lowcore.h>
-#include <asm/switch_to.h>
 #include <asm/vdso.h>
 #include "entry.h"
 
@@ -109,7 +109,7 @@ struct rt_sigframe
 static void store_sigregs(void)
 {
 	save_access_regs(current->thread.acrs);
-	save_fpu_regs();
+	save_user_fpu_regs();
 }
 
 /* Load registers after signal return */
@@ -131,7 +131,7 @@ static int save_sigregs(struct pt_regs *regs, _sigregs __user *sregs)
 	memcpy(&user_sregs.regs.gprs, &regs->gprs, sizeof(sregs->regs.gprs));
 	memcpy(&user_sregs.regs.acrs, current->thread.acrs,
 	       sizeof(user_sregs.regs.acrs));
-	fpregs_store(&user_sregs.fpregs, &current->thread.fpu);
+	fpregs_store(&user_sregs.fpregs, &current->thread.ufpu);
 	if (__copy_to_user(sregs, &user_sregs, sizeof(_sigregs)))
 		return -EFAULT;
 	return 0;
@@ -165,7 +165,7 @@ static int restore_sigregs(struct pt_regs *regs, _sigregs __user *sregs)
 	memcpy(&current->thread.acrs, &user_sregs.regs.acrs,
 	       sizeof(current->thread.acrs));
 
-	fpregs_load(&user_sregs.fpregs, &current->thread.fpu);
+	fpregs_load(&user_sregs.fpregs, &current->thread.ufpu);
 
 	clear_pt_regs_flag(regs, PIF_SYSCALL); /* No longer in a system call */
 	return 0;
@@ -181,11 +181,11 @@ static int save_sigregs_ext(struct pt_regs *regs,
 	/* Save vector registers to signal stack */
 	if (cpu_has_vx()) {
 		for (i = 0; i < __NUM_VXRS_LOW; i++)
-			vxrs[i] = current->thread.fpu.vxrs[i].low;
+			vxrs[i] = current->thread.ufpu.vxrs[i].low;
 		if (__copy_to_user(&sregs_ext->vxrs_low, vxrs,
 				   sizeof(sregs_ext->vxrs_low)) ||
 		    __copy_to_user(&sregs_ext->vxrs_high,
-				   current->thread.fpu.vxrs + __NUM_VXRS_LOW,
+				   current->thread.ufpu.vxrs + __NUM_VXRS_LOW,
 				   sizeof(sregs_ext->vxrs_high)))
 			return -EFAULT;
 	}
@@ -202,12 +202,12 @@ static int restore_sigregs_ext(struct pt_regs *regs,
 	if (cpu_has_vx()) {
 		if (__copy_from_user(vxrs, &sregs_ext->vxrs_low,
 				     sizeof(sregs_ext->vxrs_low)) ||
-		    __copy_from_user(current->thread.fpu.vxrs + __NUM_VXRS_LOW,
+		    __copy_from_user(current->thread.ufpu.vxrs + __NUM_VXRS_LOW,
 				     &sregs_ext->vxrs_high,
 				     sizeof(sregs_ext->vxrs_high)))
 			return -EFAULT;
 		for (i = 0; i < __NUM_VXRS_LOW; i++)
-			current->thread.fpu.vxrs[i].low = vxrs[i];
+			current->thread.ufpu.vxrs[i].low = vxrs[i];
 	}
 	return 0;
 }
@@ -222,7 +222,7 @@ SYSCALL_DEFINE0(sigreturn)
 	if (__copy_from_user(&set.sig, &frame->sc.oldmask, _SIGMASK_COPY_SIZE))
 		goto badframe;
 	set_current_blocked(&set);
-	save_fpu_regs();
+	save_user_fpu_regs();
 	if (restore_sigregs(regs, &frame->sregs))
 		goto badframe;
 	if (restore_sigregs_ext(regs, &frame->sregs_ext))
@@ -246,7 +246,7 @@ SYSCALL_DEFINE0(rt_sigreturn)
 	set_current_blocked(&set);
 	if (restore_altstack(&frame->uc.uc_stack))
 		goto badframe;
-	save_fpu_regs();
+	save_user_fpu_regs();
 	if (restore_sigregs(regs, &frame->uc.uc_mcontext))
 		goto badframe;
 	if (restore_sigregs_ext(regs, &frame->uc.uc_mcontext_ext))
diff --git a/arch/s390/kernel/smp.c b/arch/s390/kernel/smp.c
index c39d9f0d4b1c..0324649aae0a 100644
--- a/arch/s390/kernel/smp.c
+++ b/arch/s390/kernel/smp.c
@@ -36,12 +36,13 @@
 #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/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>
diff --git a/arch/s390/kernel/sysinfo.c b/arch/s390/kernel/sysinfo.c
index f6f8f498c9be..1b1be3110cfc 100644
--- a/arch/s390/kernel/sysinfo.c
+++ b/arch/s390/kernel/sysinfo.c
@@ -20,7 +20,7 @@
 #include <asm/sysinfo.h>
 #include <asm/cpcmd.h>
 #include <asm/topology.h>
-#include <asm/fpu/api.h>
+#include <asm/fpu.h>
 
 int topology_max_mnest;
 
@@ -426,9 +426,9 @@ subsys_initcall(create_proc_service_level);
  */
 void s390_adjust_jiffies(void)
 {
+	DECLARE_KERNEL_FPU_ONSTACK16(fpu);
 	struct sysinfo_1_2_2 *info;
 	unsigned long capability;
-	struct kernel_fpu fpu;
 
 	info = (void *) get_zeroed_page(GFP_KERNEL);
 	if (!info)
@@ -447,21 +447,14 @@ void s390_adjust_jiffies(void)
 		 * point division ..
 		 */
 		kernel_fpu_begin(&fpu, KERNEL_FPR);
-		asm volatile(
-			"	sfpc	%3\n"
-			"	l	%0,%1\n"
-			"	tmlh	%0,0xff80\n"
-			"	jnz	0f\n"
-			"	cefbr	%%f2,%0\n"
-			"	j	1f\n"
-			"0:	le	%%f2,%1\n"
-			"1:	cefbr	%%f0,%2\n"
-			"	debr	%%f0,%%f2\n"
-			"	cgebr	%0,5,%%f0\n"
-			: "=&d" (capability)
-			: "Q" (info->capability), "d" (10000000), "d" (0)
-			: "cc"
-			);
+		fpu_sfpc(0);
+		if (info->capability & 0xff800000)
+			fpu_ldgr(2, info->capability);
+		else
+			fpu_cefbr(2, info->capability);
+		fpu_cefbr(0, 10000000);
+		fpu_debr(0, 2);
+		capability = fpu_cgebr(0, 5);
 		kernel_fpu_end(&fpu, KERNEL_FPR);
 	} else
 		/*
diff --git a/arch/s390/kernel/text_amode31.S b/arch/s390/kernel/text_amode31.S
index 14c6d25c035f..c0a70efa2426 100644
--- a/arch/s390/kernel/text_amode31.S
+++ b/arch/s390/kernel/text_amode31.S
@@ -90,7 +90,7 @@ SYM_FUNC_START(_diag26c_amode31)
 SYM_FUNC_END(_diag26c_amode31)
 
 /*
- * void _diag0c_amode31(struct hypfs_diag0c_entry *entry)
+ * void _diag0c_amode31(unsigned long rx)
  */
 SYM_FUNC_START(_diag0c_amode31)
 	sam31
diff --git a/arch/s390/kernel/time.c b/arch/s390/kernel/time.c
index 14abad953c02..fb9f31f36628 100644
--- a/arch/s390/kernel/time.c
+++ b/arch/s390/kernel/time.c
@@ -251,8 +251,8 @@ static struct clocksource clocksource_tod = {
 	.rating		= 400,
 	.read		= read_tod_clock,
 	.mask		= CLOCKSOURCE_MASK(64),
-	.mult		= 1000,
-	.shift		= 12,
+	.mult		= 4096000,
+	.shift		= 24,
 	.flags		= CLOCK_SOURCE_IS_CONTINUOUS,
 	.vdso_clock_mode = VDSO_CLOCKMODE_TOD,
 };
@@ -716,7 +716,7 @@ out_unlock:
 /*
  * STP subsys sysfs interface functions
  */
-static struct bus_type stp_subsys = {
+static const struct bus_type stp_subsys = {
 	.name		= "stp",
 	.dev_name	= "stp",
 };
diff --git a/arch/s390/kernel/traps.c b/arch/s390/kernel/traps.c
index 46dac4540ca8..52578b5cecbd 100644
--- a/arch/s390/kernel/traps.c
+++ b/arch/s390/kernel/traps.c
@@ -28,8 +28,8 @@
 #include <linux/cpu.h>
 #include <linux/entry-common.h>
 #include <asm/asm-extable.h>
-#include <asm/fpu/api.h>
 #include <asm/vtime.h>
+#include <asm/fpu.h>
 #include "entry.h"
 
 static inline void __user *get_trap_ip(struct pt_regs *regs)
@@ -201,8 +201,8 @@ static void vector_exception(struct pt_regs *regs)
 	}
 
 	/* get vector interrupt code from fpc */
-	save_fpu_regs();
-	vic = (current->thread.fpu.fpc & 0xf00) >> 8;
+	save_user_fpu_regs();
+	vic = (current->thread.ufpu.fpc & 0xf00) >> 8;
 	switch (vic) {
 	case 1: /* invalid vector operation */
 		si_code = FPE_FLTINV;
@@ -227,9 +227,9 @@ static void vector_exception(struct pt_regs *regs)
 
 static void data_exception(struct pt_regs *regs)
 {
-	save_fpu_regs();
-	if (current->thread.fpu.fpc & FPC_DXC_MASK)
-		do_fp_trap(regs, current->thread.fpu.fpc);
+	save_user_fpu_regs();
+	if (current->thread.ufpu.fpc & FPC_DXC_MASK)
+		do_fp_trap(regs, current->thread.ufpu.fpc);
 	else
 		do_trap(regs, SIGILL, ILL_ILLOPN, "data exception");
 }
diff --git a/arch/s390/kernel/uprobes.c b/arch/s390/kernel/uprobes.c
index b88345ef8bd9..5b0633ea8d93 100644
--- a/arch/s390/kernel/uprobes.c
+++ b/arch/s390/kernel/uprobes.c
@@ -12,7 +12,6 @@
 #include <linux/kdebug.h>
 #include <linux/sched/task_stack.h>
 
-#include <asm/switch_to.h>
 #include <asm/facility.h>
 #include <asm/kprobes.h>
 #include <asm/dis.h>
diff --git a/arch/s390/kernel/vdso32/Makefile b/arch/s390/kernel/vdso32/Makefile
index caec7db6f966..b12a274cbb47 100644
--- a/arch/s390/kernel/vdso32/Makefile
+++ b/arch/s390/kernel/vdso32/Makefile
@@ -22,7 +22,7 @@ KBUILD_CFLAGS_32 := $(filter-out -m64,$(KBUILD_CFLAGS))
 KBUILD_CFLAGS_32 := $(filter-out -mno-pic-data-is-text-relative,$(KBUILD_CFLAGS_32))
 KBUILD_CFLAGS_32 += -m31 -fPIC -shared -fno-common -fno-builtin
 
-LDFLAGS_vdso32.so.dbg += -fPIC -shared -soname=linux-vdso32.so.1 \
+LDFLAGS_vdso32.so.dbg += -shared -soname=linux-vdso32.so.1 \
 	--hash-style=both --build-id=sha1 -melf_s390 -T
 
 $(targets:%=$(obj)/%.dbg): KBUILD_CFLAGS = $(KBUILD_CFLAGS_32)
diff --git a/arch/s390/kernel/vdso32/vdso32.lds.S b/arch/s390/kernel/vdso32/vdso32.lds.S
index edf5ff1debe1..65b9513a5a0e 100644
--- a/arch/s390/kernel/vdso32/vdso32.lds.S
+++ b/arch/s390/kernel/vdso32/vdso32.lds.S
@@ -9,7 +9,6 @@
 
 OUTPUT_FORMAT("elf32-s390", "elf32-s390", "elf32-s390")
 OUTPUT_ARCH(s390:31-bit)
-ENTRY(_start)
 
 SECTIONS
 {
diff --git a/arch/s390/kernel/vdso64/Makefile b/arch/s390/kernel/vdso64/Makefile
index e3c9085f8fa7..ef9832726097 100644
--- a/arch/s390/kernel/vdso64/Makefile
+++ b/arch/s390/kernel/vdso64/Makefile
@@ -25,8 +25,9 @@ KBUILD_AFLAGS_64 += -m64
 
 KBUILD_CFLAGS_64 := $(filter-out -m64,$(KBUILD_CFLAGS))
 KBUILD_CFLAGS_64 := $(filter-out -mno-pic-data-is-text-relative,$(KBUILD_CFLAGS_64))
+KBUILD_CFLAGS_64 := $(filter-out -munaligned-symbols,$(KBUILD_CFLAGS_64))
 KBUILD_CFLAGS_64 += -m64 -fPIC -fno-common -fno-builtin
-ldflags-y := -fPIC -shared -soname=linux-vdso64.so.1 \
+ldflags-y := -shared -soname=linux-vdso64.so.1 \
 	     --hash-style=both --build-id=sha1 -T
 
 $(targets:%=$(obj)/%.dbg): KBUILD_CFLAGS = $(KBUILD_CFLAGS_64)
diff --git a/arch/s390/kernel/vdso64/vdso64.lds.S b/arch/s390/kernel/vdso64/vdso64.lds.S
index 4461ea151e49..37e2a505e81d 100644
--- a/arch/s390/kernel/vdso64/vdso64.lds.S
+++ b/arch/s390/kernel/vdso64/vdso64.lds.S
@@ -9,7 +9,6 @@
 
 OUTPUT_FORMAT("elf64-s390", "elf64-s390", "elf64-s390")
 OUTPUT_ARCH(s390:64-bit)
-ENTRY(_start)
 
 SECTIONS
 {
diff --git a/arch/s390/kernel/vmlinux.lds.S b/arch/s390/kernel/vmlinux.lds.S
index e32ef446f451..48de296e8905 100644
--- a/arch/s390/kernel/vmlinux.lds.S
+++ b/arch/s390/kernel/vmlinux.lds.S
@@ -59,6 +59,14 @@ SECTIONS
 	} :text = 0x0700
 
 	RO_DATA(PAGE_SIZE)
+	.data.rel.ro : {
+		*(.data.rel.ro .data.rel.ro.*)
+	}
+	.got : {
+		__got_start = .;
+		*(.got)
+		__got_end = .;
+	}
 
 	. = ALIGN(PAGE_SIZE);
 	_sdata = .;		/* Start of data section */
@@ -73,6 +81,9 @@ SECTIONS
 	__end_ro_after_init = .;
 
 	RW_DATA(0x100, PAGE_SIZE, THREAD_SIZE)
+	.data.rel : {
+		*(.data.rel*)
+	}
 	BOOT_DATA_PRESERVED
 
 	. = ALIGN(8);
@@ -181,6 +192,7 @@ SECTIONS
 
 	PERCPU_SECTION(0x100)
 
+#ifdef CONFIG_PIE_BUILD
 	.dynsym ALIGN(8) : {
 		__dynsym_start = .;
 		*(.dynsym)
@@ -191,6 +203,19 @@ SECTIONS
 		*(.rela*)
 		__rela_dyn_end = .;
 	}
+	.dynamic ALIGN(8) : {
+		*(.dynamic)
+	}
+	.dynstr ALIGN(8) : {
+		*(.dynstr)
+	}
+#endif
+	.hash ALIGN(8) : {
+		*(.hash)
+	}
+	.gnu.hash ALIGN(8) : {
+		*(.gnu.hash)
+	}
 
 	. = ALIGN(PAGE_SIZE);
 	__init_end = .;		/* freed after init ends here */
@@ -214,9 +239,14 @@ SECTIONS
 		QUAD(__boot_data_preserved_start)		/* bootdata_preserved_off */
 		QUAD(__boot_data_preserved_end -
 		     __boot_data_preserved_start)		/* bootdata_preserved_size */
+#ifdef CONFIG_PIE_BUILD
 		QUAD(__dynsym_start)				/* dynsym_start */
 		QUAD(__rela_dyn_start)				/* rela_dyn_start */
 		QUAD(__rela_dyn_end)				/* rela_dyn_end */
+#else
+		QUAD(__got_start)				/* got_start */
+		QUAD(__got_end)					/* got_end */
+#endif
 		QUAD(_eamode31 - _samode31)			/* amode31_size */
 		QUAD(init_mm)
 		QUAD(swapper_pg_dir)
@@ -235,6 +265,30 @@ SECTIONS
 	DWARF_DEBUG
 	ELF_DETAILS
 
+	/*
+	 * Make sure that the .got.plt is either completely empty or it
+	 * contains only the three reserved double words.
+	 */
+	.got.plt : {
+		*(.got.plt)
+	}
+	ASSERT(SIZEOF(.got.plt) == 0 || SIZEOF(.got.plt) == 0x18, "Unexpected GOT/PLT entries detected!")
+
+	/*
+	 * Sections that should stay zero sized, which is safer to
+	 * explicitly check instead of blindly discarding.
+	 */
+	.plt : {
+		*(.plt) *(.plt.*) *(.iplt) *(.igot .igot.plt)
+	}
+	ASSERT(SIZEOF(.plt) == 0, "Unexpected run-time procedure linkages detected!")
+#ifndef CONFIG_PIE_BUILD
+	.rela.dyn : {
+		*(.rela.*) *(.rela_*)
+	}
+	ASSERT(SIZEOF(.rela.dyn) == 0, "Unexpected run-time relocations (.rela) detected!")
+#endif
+
 	/* Sections to be discarded */
 	DISCARDS
 	/DISCARD/ : {