Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux

Pull s390 updates from Martin Schwidefsky:
 "The new features and main improvements in this merge for v4.9

   - Support for the UBSAN sanitizer

   - Set HAVE_EFFICIENT_UNALIGNED_ACCESS, it improves the code in some
     places

   - Improvements for the in-kernel fpu code, in particular the overhead
     for multiple consecutive in kernel fpu users is recuded

   - Add a SIMD implementation for the RAID6 gen and xor operations

   - Add RAID6 recovery based on the XC instruction

   - The PCI DMA flush logic has been improved to increase the speed of
     the map / unmap operations

   - The time synchronization code has seen some updates

  And bug fixes all over the place"

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux: (48 commits)
  s390/con3270: fix insufficient space padding
  s390/con3270: fix use of uninitialised data
  MAINTAINERS: update DASD maintainer
  s390/cio: fix accidental interrupt enabling during resume
  s390/dasd: add missing \n to end of dev_err messages
  s390/config: Enable config options for Docker
  s390/dasd: make query host access interruptible
  s390/dasd: fix panic during offline processing
  s390/dasd: fix hanging offline processing
  s390/pci_dma: improve lazy flush for unmap
  s390/pci_dma: split dma_update_trans
  s390/pci_dma: improve map_sg
  s390/pci_dma: simplify dma address calculation
  s390/pci_dma: remove dma address range check
  iommu/s390: simplify registration of I/O address translation parameters
  s390: migrate exception table users off module.h and onto extable.h
  s390: export header for CLP ioctl
  s390/vmur: fix irq pointer dereference in int handler
  s390/dasd: add missing KOBJ_CHANGE event for unformatted devices
  s390: enable UBSAN
  ...

11 files changed, 189 insertions, 303 deletions

diff --git a/arch/s390/kernel/Makefile b/arch/s390/kernel/Makefile
index 3234817..72ccc41 100644
--- a/arch/s390/kernel/Makefile
+++ b/arch/s390/kernel/Makefile
@@ -48,6 +48,9 @@ AFLAGS_head.o		+= -march=z900
 endif
 GCOV_PROFILE_sclp.o := n
 GCOV_PROFILE_als.o := n
+UBSAN_SANITIZE_als.o := n
+UBSAN_SANITIZE_early.o := n
+UBSAN_SANITIZE_sclp.o := n
 
 obj-y	:= traps.o time.o process.o base.o early.o setup.o idle.o vtime.o
 obj-y	+= processor.o sys_s390.o ptrace.o signal.o cpcmd.o ebcdic.o nmi.o
diff --git a/arch/s390/kernel/crash_dump.c b/arch/s390/kernel/crash_dump.c
index 29df8484..f9293bf 100644
--- a/arch/s390/kernel/crash_dump.c
+++ b/arch/s390/kernel/crash_dump.c
@@ -71,9 +71,7 @@ struct save_area * __init save_area_alloc(bool is_boot_cpu)
  */
 struct save_area * __init save_area_boot_cpu(void)
 {
-	if (list_empty(&dump_save_areas))
-		return NULL;
-	return list_first_entry(&dump_save_areas, struct save_area, list);
+	return list_first_entry_or_null(&dump_save_areas, struct save_area, list);
 }
 
 /*
diff --git a/arch/s390/kernel/early.c b/arch/s390/kernel/early.c
index 717b03a..2374c5b 100644
--- a/arch/s390/kernel/early.c
+++ b/arch/s390/kernel/early.c
@@ -13,7 +13,7 @@
 #include <linux/string.h>
 #include <linux/ctype.h>
 #include <linux/lockdep.h>
-#include <linux/module.h>
+#include <linux/extable.h>
 #include <linux/pfn.h>
 #include <linux/uaccess.h>
 #include <linux/kernel.h>
diff --git a/arch/s390/kernel/fpu.c b/arch/s390/kernel/fpu.c
index 81d1d18..1235b94 100644
--- a/arch/s390/kernel/fpu.c
+++ b/arch/s390/kernel/fpu.c
@@ -10,240 +10,167 @@
 #include <asm/fpu/types.h>
 #include <asm/fpu/api.h>
 
-/*
- * Per-CPU variable to maintain FPU register ranges that are in use
- * by the kernel.
- */
-static DEFINE_PER_CPU(u32, kernel_fpu_state);
-
-#define KERNEL_FPU_STATE_MASK	(KERNEL_FPU_MASK|KERNEL_FPC)
-
+asm(".include \"asm/vx-insn.h\"\n");
 
 void __kernel_fpu_begin(struct kernel_fpu *state, u32 flags)
 {
-	if (!__this_cpu_read(kernel_fpu_state)) {
-		/*
-		 * Save user space FPU state and register contents.  Multiple
-		 * calls because of interruptions do not matter and return
-		 * immediately.  This also sets CIF_FPU to lazy restore FP/VX
-		 * register contents when returning to user space.
-		 */
-		save_fpu_regs();
-	}
-
-	/* Update flags to use the vector facility for KERNEL_FPR */
-	if (MACHINE_HAS_VX && (state->mask & KERNEL_FPR)) {
-		flags |= KERNEL_VXR_LOW | KERNEL_FPC;
-		flags &= ~KERNEL_FPR;
-	}
-
-	/* Save and update current kernel VX state */
-	state->mask = __this_cpu_read(kernel_fpu_state);
-	__this_cpu_or(kernel_fpu_state, flags & KERNEL_FPU_STATE_MASK);
-
 	/*
-	 * If this is the first call to __kernel_fpu_begin(), no additional
-	 * work is required.
+	 * Limit the save to the FPU/vector registers already
+	 * in use by the previous context
 	 */
-	if (!(state->mask & KERNEL_FPU_STATE_MASK))
-		return;
+	flags &= state->mask;
 
-	/*
-	 * If KERNEL_FPR is still set, the vector facility is not available
-	 * and, thus, save floating-point control and registers only.
-	 */
-	if (state->mask & KERNEL_FPR) {
-		asm volatile("stfpc %0" : "=Q" (state->fpc));
-		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_FPC)
+		/* Save floating point control */
+		asm volatile("stfpc %0" : "=m" (state->fpc));
+
+	if (!MACHINE_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]));
+		}
 		return;
 	}
 
-	/*
-	 * If this is a nested call to __kernel_fpu_begin(), check the saved
-	 * state mask to save and later restore the vector registers that
-	 * are already in use.	Let's start with checking floating-point
-	 * controls.
-	 */
-	if (state->mask & KERNEL_FPC)
-		asm volatile("stfpc %0" : "=m" (state->fpc));
-
 	/* 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.
 		 */
-		"	tmll	%[m],15\n"	/* KERNEL_VXR_MASK */
-		"	jz	20f\n"		/* no work -> done */
 		"	la	1,%[vxrs]\n"	/* load save area */
-		"	jo	18f\n"		/* -> save V0..V31 */
-
+		"	tmll	%[m],30\n"	/* KERNEL_VXR */
+		"	jz	7f\n"		/* no work -> done */
+		"	jo	5f\n"		/* -> save V0..V31 */
 		/*
-		 * Test if V8..V23 can be saved at once... this speeds up
-		 * for KERNEL_fpu_MID only. Otherwise continue to split the
-		 * range of vector registers into two halves and test them
-		 * separately.
+		 * Test for special case KERNEL_FPU_MID only. In this
+		 * case a vstm V8..V23 is the best instruction
 		 */
-		"	tmll	%[m],6\n"	/* KERNEL_VXR_MID */
-		"	jo	17f\n"		/* -> save V8..V23 */
-
+		"	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 */
-		"1:	tmll	%[m],3\n"	/* KERNEL_VXR_LOW */
-		"	jz	10f\n"		/* -> KERNEL_VXR_HIGH */
+		"0:	tmll	%[m],6\n"	/* KERNEL_VXR_LOW */
+		"	jz	3f\n"		/* -> KERNEL_VXR_HIGH */
 		"	jo	2f\n"		/* 11 -> save V0..V15 */
-		"	brc	4,3f\n"		/* 01 -> save V0..V7  */
-		"	brc	2,4f\n"		/* 10 -> save V8..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 */
-		"10:	tmll	%[m],12\n"	/* KERNEL_VXR_HIGH */
-		"	jo	19f\n"		/* 11 -> save V16..V31 */
-		"	brc	4,11f\n"	/* 01 -> save V16..V23	*/
-		"	brc	2,12f\n"	/* 10 -> save V24..V31 */
-		"	j	20f\n"		/* 00 -> done */
-
-		/*
-		 * Below are the vstm combinations to save multiple vector
-		 * registers at once.
-		 */
-		"2:	.word	0xe70f,0x1000,0x003e\n"	/* vstm 0,15,0(1) */
-		"	j	10b\n"			/* -> VXR_HIGH */
-		"3:	.word	0xe707,0x1000,0x003e\n" /* vstm 0,7,0(1) */
-		"	j	10b\n"			/* -> VXR_HIGH */
-		"4:	.word	0xe78f,0x1080,0x003e\n" /* vstm 8,15,128(1) */
-		"	j	10b\n"			/* -> VXR_HIGH */
-		"\n"
-		"11:	.word	0xe707,0x1100,0x0c3e\n"	/* vstm 16,23,256(1) */
-		"	j	20f\n"			/* -> done */
-		"12:	.word	0xe78f,0x1180,0x0c3e\n" /* vstm 24,31,384(1) */
-		"	j	20f\n"			/* -> done */
-		"\n"
-		"17:	.word	0xe787,0x1080,0x043e\n"	/* vstm 8,23,128(1) */
-		"	nill	%[m],249\n"		/* m &= ~VXR_MID    */
-		"	j	1b\n"			/* -> VXR_LOW */
-		"\n"
-		"18:	.word	0xe70f,0x1000,0x003e\n"	/* vstm 0,15,0(1) */
-		"19:	.word	0xe70f,0x1100,0x0c3e\n"	/* vstm 16,31,256(1) */
-		"20:"
+		"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" (state->mask)
+		: [m] "d" (flags)
 		: "1", "cc");
 }
 EXPORT_SYMBOL(__kernel_fpu_begin);
 
-void __kernel_fpu_end(struct kernel_fpu *state)
+void __kernel_fpu_end(struct kernel_fpu *state, u32 flags)
 {
-	/* Just update the per-CPU state if there is nothing to restore */
-	if (!(state->mask & KERNEL_FPU_STATE_MASK))
-		goto update_fpu_state;
-
 	/*
-	 * If KERNEL_FPR is specified, the vector facility is not available
-	 * and, thus, restore floating-point control and registers only.
+	 * Limit the restore to the FPU/vector registers of the
+	 * previous context that have been overwritte by the
+	 * current context
 	 */
-	if (state->mask & KERNEL_FPR) {
-		asm volatile("lfpc %0" : : "Q" (state->fpc));
-		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]));
-		goto update_fpu_state;
-	}
+	flags &= state->mask;
 
-	/* Test and restore floating-point controls */
-	if (state->mask & KERNEL_FPC)
+	if (flags & KERNEL_FPC)
+		/* Restore floating-point controls */
 		asm volatile("lfpc %0" : : "Q" (state->fpc));
 
+	if (!MACHINE_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]));
+		}
+		return;
+	}
+
 	/* Test and restore (load) vector registers */
 	asm volatile (
 		/*
-		 * Test if any vector registers must be loaded and, if so,
+		 * Test if any vector register must be loaded and, if so,
 		 * test if all registers can be loaded at once.
 		 */
-		"	tmll	%[m],15\n"	/* KERNEL_VXR_MASK */
-		"	jz	20f\n"		/* no work -> done */
-		"	la	1,%[vxrs]\n"	/* load load area */
-		"	jo	18f\n"		/* -> load V0..V31 */
-
-		/*
-		 * Test if V8..V23 can be restored at once... this speeds up
-		 * for KERNEL_VXR_MID only. Otherwise continue to split the
-		 * range of vector registers into two halves and test them
-		 * separately.
-		 */
-		"	tmll	%[m],6\n"	/* KERNEL_VXR_MID */
-		"	jo	17f\n"		/* -> load V8..V23 */
-
-		/* Test and load the first half of 16 vector registers */
-		"1:	tmll	%[m],3\n"	/* KERNEL_VXR_LOW */
-		"	jz	10f\n"		/* -> KERNEL_VXR_HIGH */
-		"	jo	2f\n"		/* 11 -> load V0..V15 */
-		"	brc	4,3f\n"		/* 01 -> load V0..V7  */
-		"	brc	2,4f\n"		/* 10 -> load V8..V15 */
-
-		/* Test and load the second half of 16 vector registers */
-		"10:	tmll	%[m],12\n"	/* KERNEL_VXR_HIGH */
-		"	jo	19f\n"		/* 11 -> load V16..V31 */
-		"	brc	4,11f\n"	/* 01 -> load V16..V23	*/
-		"	brc	2,12f\n"	/* 10 -> load V24..V31 */
-		"	j	20f\n"		/* 00 -> done */
-
+		"	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 */
 		/*
-		 * Below are the vstm combinations to load multiple vector
-		 * registers at once.
+		 * Test for special case KERNEL_FPU_MID only. In this
+		 * case a vlm V8..V23 is the best instruction
 		 */
-		"2:	.word	0xe70f,0x1000,0x0036\n"	/* vlm 0,15,0(1) */
-		"	j	10b\n"			/* -> VXR_HIGH */
-		"3:	.word	0xe707,0x1000,0x0036\n" /* vlm 0,7,0(1) */
-		"	j	10b\n"			/* -> VXR_HIGH */
-		"4:	.word	0xe78f,0x1080,0x0036\n" /* vlm 8,15,128(1) */
-		"	j	10b\n"			/* -> VXR_HIGH */
-		"\n"
-		"11:	.word	0xe707,0x1100,0x0c36\n"	/* vlm 16,23,256(1) */
-		"	j	20f\n"			/* -> done */
-		"12:	.word	0xe78f,0x1180,0x0c36\n" /* vlm 24,31,384(1) */
-		"	j	20f\n"			/* -> done */
-		"\n"
-		"17:	.word	0xe787,0x1080,0x0436\n"	/* vlm 8,23,128(1) */
-		"	nill	%[m],249\n"		/* m &= ~VXR_MID    */
-		"	j	1b\n"			/* -> VXR_LOW */
-		"\n"
-		"18:	.word	0xe70f,0x1000,0x0036\n"	/* vlm 0,15,0(1) */
-		"19:	.word	0xe70f,0x1100,0x0c36\n"	/* vlm 16,31,256(1) */
-		"20:"
-		:
-		: [vxrs] "Q" (*(struct vx_array *) &state->vxrs),
-		  [m] "d" (state->mask)
+		"	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");
-
-update_fpu_state:
-	/* Update current kernel VX state */
-	__this_cpu_write(kernel_fpu_state, state->mask);
 }
 EXPORT_SYMBOL(__kernel_fpu_end);
diff --git a/arch/s390/kernel/kprobes.c b/arch/s390/kernel/kprobes.c
index dd6306c..fdb4042 100644
--- a/arch/s390/kernel/kprobes.c
+++ b/arch/s390/kernel/kprobes.c
@@ -26,12 +26,14 @@
 #include <linux/stop_machine.h>
 #include <linux/kdebug.h>
 #include <linux/uaccess.h>
+#include <linux/extable.h>
 #include <linux/module.h>
 #include <linux/slab.h>
 #include <linux/hardirq.h>
 #include <linux/ftrace.h>
 #include <asm/cacheflush.h>
 #include <asm/sections.h>
+#include <asm/uaccess.h>
 #include <asm/dis.h>
 
 DEFINE_PER_CPU(struct kprobe *, current_kprobe);
diff --git a/arch/s390/kernel/nmi.c b/arch/s390/kernel/nmi.c
index 29376f0..9a32f74 100644
--- a/arch/s390/kernel/nmi.c
+++ b/arch/s390/kernel/nmi.c
@@ -98,7 +98,7 @@ EXPORT_SYMBOL_GPL(s390_handle_mcck);
  * returns 0 if all registers could be validated
  * returns 1 otherwise
  */
-static int notrace s390_validate_registers(union mci mci)
+static int notrace s390_validate_registers(union mci mci, int umode)
 {
 	int kill_task;
 	u64 zero;
@@ -110,26 +110,41 @@ static int notrace s390_validate_registers(union mci mci)
 	if (!mci.gr) {
 		/*
 		 * General purpose registers couldn't be restored and have
-		 * unknown contents. Process needs to be terminated.
+		 * unknown contents. Stop system or terminate process.
 		 */
+		if (!umode)
+			s390_handle_damage();
 		kill_task = 1;
 	}
 	if (!mci.fp) {
 		/*
-		 * Floating point registers can't be restored and
-		 * therefore the process needs to be terminated.
+		 * Floating point registers can't be restored. If the
+		 * kernel currently uses floating point registers the
+		 * system is stopped. If the process has its floating
+		 * pointer registers loaded it is terminated.
+		 * Otherwise just revalidate the registers.
 		 */
-		kill_task = 1;
+		if (S390_lowcore.fpu_flags & KERNEL_VXR_V0V7)
+			s390_handle_damage();
+		if (!test_cpu_flag(CIF_FPU))
+			kill_task = 1;
 	}
 	fpt_save_area = &S390_lowcore.floating_pt_save_area;
 	fpt_creg_save_area = &S390_lowcore.fpt_creg_save_area;
 	if (!mci.fc) {
 		/*
 		 * Floating point control register can't be restored.
-		 * Task will be terminated.
+		 * If the kernel currently uses the floating pointer
+		 * registers and needs the FPC register the system is
+		 * stopped. If the process has its floating pointer
+		 * registers loaded it is terminated. Otherwiese the
+		 * FPC is just revalidated.
 		 */
+		if (S390_lowcore.fpu_flags & KERNEL_FPC)
+			s390_handle_damage();
 		asm volatile("lfpc 0(%0)" : : "a" (&zero), "m" (zero));
-		kill_task = 1;
+		if (!test_cpu_flag(CIF_FPU))
+			kill_task = 1;
 	} else
 		asm volatile("lfpc 0(%0)" : : "a" (fpt_creg_save_area));
 
@@ -159,10 +174,16 @@ static int notrace s390_validate_registers(union mci mci)
 
 		if (!mci.vr) {
 			/*
-			 * Vector registers can't be restored and therefore
-			 * the process needs to be terminated.
+			 * Vector registers can't be restored. If the kernel
+			 * currently uses vector registers the system is
+			 * stopped. If the process has its vector registers
+			 * loaded it is terminated. Otherwise just revalidate
+			 * the registers.
 			 */
-			kill_task = 1;
+			if (S390_lowcore.fpu_flags & KERNEL_VXR)
+				s390_handle_damage();
+			if (!test_cpu_flag(CIF_FPU))
+				kill_task = 1;
 		}
 		cr0.val = S390_lowcore.cregs_save_area[0];
 		cr0.afp = cr0.vx = 1;
@@ -250,13 +271,11 @@ void notrace s390_do_machine_check(struct pt_regs *regs)
 	struct mcck_struct *mcck;
 	unsigned long long tmp;
 	union mci mci;
-	int umode;
 
 	nmi_enter();
 	inc_irq_stat(NMI_NMI);
 	mci.val = S390_lowcore.mcck_interruption_code;
 	mcck = this_cpu_ptr(&cpu_mcck);
-	umode = user_mode(regs);
 
 	if (mci.sd) {
 		/* System damage -> stopping machine */
@@ -297,22 +316,14 @@ void notrace s390_do_machine_check(struct pt_regs *regs)
 			s390_handle_damage();
 		}
 	}
-	if (s390_validate_registers(mci)) {
-		if (umode) {
-			/*
-			 * Couldn't restore all register contents while in
-			 * user mode -> mark task for termination.
-			 */
-			mcck->kill_task = 1;
-			mcck->mcck_code = mci.val;
-			set_cpu_flag(CIF_MCCK_PENDING);
-		} else {
-			/*
-			 * Couldn't restore all register contents while in
-			 * kernel mode -> stopping machine.
-			 */
-			s390_handle_damage();
-		}
+	if (s390_validate_registers(mci, user_mode(regs))) {
+		/*
+		 * Couldn't restore all register contents for the
+		 * user space process -> mark task for termination.
+		 */
+		mcck->kill_task = 1;
+		mcck->mcck_code = mci.val;
+		set_cpu_flag(CIF_MCCK_PENDING);
 	}
 	if (mci.cd) {
 		/* Timing facility damage */
diff --git a/arch/s390/kernel/sysinfo.c b/arch/s390/kernel/sysinfo.c
index 050b8d0..bfda6aa 100644
--- a/arch/s390/kernel/sysinfo.c
+++ b/arch/s390/kernel/sysinfo.c
@@ -454,7 +454,7 @@ void s390_adjust_jiffies(void)
 			: "Q" (info->capability), "d" (10000000), "d" (0)
 			: "cc"
 			);
-		kernel_fpu_end(&fpu);
+		kernel_fpu_end(&fpu, KERNEL_FPR);
 	} else
 		/*
 		 * Really old machine without stsi block for basic
diff --git a/arch/s390/kernel/time.c b/arch/s390/kernel/time.c
index 4e99498..0bfcc49 100644
--- a/arch/s390/kernel/time.c
+++ b/arch/s390/kernel/time.c
@@ -50,10 +50,6 @@
 #include <asm/cio.h>
 #include "entry.h"
 
-/* change this if you have some constant time drift */
-#define USECS_PER_JIFFY     ((unsigned long) 1000000/HZ)
-#define CLK_TICKS_PER_JIFFY ((unsigned long) USECS_PER_JIFFY << 12)
-
 u64 sched_clock_base_cc = -1;	/* Force to data section. */
 EXPORT_SYMBOL_GPL(sched_clock_base_cc);
 
@@ -282,13 +278,8 @@ extern struct timezone sys_tz;
 
 void update_vsyscall_tz(void)
 {
-	/* Make userspace gettimeofday spin until we're done. */
-	++vdso_data->tb_update_count;
-	smp_wmb();
 	vdso_data->tz_minuteswest = sys_tz.tz_minuteswest;
 	vdso_data->tz_dsttime = sys_tz.tz_dsttime;
-	smp_wmb();
-	++vdso_data->tb_update_count;
 }
 
 /*
@@ -318,51 +309,12 @@ void __init time_init(void)
 	vtime_init();
 }
 
-/*
- * The time is "clock". old is what we think the time is.
- * Adjust the value by a multiple of jiffies and add the delta to ntp.
- * "delay" is an approximation how long the synchronization took. If
- * the time correction is positive, then "delay" is subtracted from
- * the time difference and only the remaining part is passed to ntp.
- */
-static unsigned long long adjust_time(unsigned long long old,
-				      unsigned long long clock,
-				      unsigned long long delay)
-{
-	unsigned long long delta, ticks;
-	struct timex adjust;
-
-	if (clock > old) {
-		/* It is later than we thought. */
-		delta = ticks = clock - old;
-		delta = ticks = (delta < delay) ? 0 : delta - delay;
-		delta -= do_div(ticks, CLK_TICKS_PER_JIFFY);
-		adjust.offset = ticks * (1000000 / HZ);
-	} else {
-		/* It is earlier than we thought. */
-		delta = ticks = old - clock;
-		delta -= do_div(ticks, CLK_TICKS_PER_JIFFY);
-		delta = -delta;
-		adjust.offset = -ticks * (1000000 / HZ);
-	}
-	sched_clock_base_cc += delta;
-	if (adjust.offset != 0) {
-		pr_notice("The ETR interface has adjusted the clock "
-			  "by %li microseconds\n", adjust.offset);
-		adjust.modes = ADJ_OFFSET_SINGLESHOT;
-		do_adjtimex(&adjust);
-	}
-	return delta;
-}
-
 static DEFINE_PER_CPU(atomic_t, clock_sync_word);
 static DEFINE_MUTEX(clock_sync_mutex);
 static unsigned long clock_sync_flags;
 
-#define CLOCK_SYNC_HAS_ETR	0
-#define CLOCK_SYNC_HAS_STP	1
-#define CLOCK_SYNC_ETR		2
-#define CLOCK_SYNC_STP		3
+#define CLOCK_SYNC_HAS_STP	0
+#define CLOCK_SYNC_STP		1
 
 /*
  * The get_clock function for the physical clock. It will get the current
@@ -384,34 +336,32 @@ int get_phys_clock(unsigned long long *clock)
 	if (sw0 == sw1 && (sw0 & 0x80000000U))
 		/* Success: time is in sync. */
 		return 0;
-	if (!test_bit(CLOCK_SYNC_HAS_ETR, &clock_sync_flags) &&
-	    !test_bit(CLOCK_SYNC_HAS_STP, &clock_sync_flags))
+	if (!test_bit(CLOCK_SYNC_HAS_STP, &clock_sync_flags))
 		return -EOPNOTSUPP;
-	if (!test_bit(CLOCK_SYNC_ETR, &clock_sync_flags) &&
-	    !test_bit(CLOCK_SYNC_STP, &clock_sync_flags))
+	if (!test_bit(CLOCK_SYNC_STP, &clock_sync_flags))
 		return -EACCES;
 	return -EAGAIN;
 }
 EXPORT_SYMBOL(get_phys_clock);
 
 /*
- * Make get_sync_clock return -EAGAIN.
+ * Make get_phys_clock() return -EAGAIN.
  */
 static void disable_sync_clock(void *dummy)
 {
 	atomic_t *sw_ptr = this_cpu_ptr(&clock_sync_word);
 	/*
-	 * Clear the in-sync bit 2^31. All get_sync_clock calls will
+	 * Clear the in-sync bit 2^31. All get_phys_clock calls will
 	 * fail until the sync bit is turned back on. In addition
 	 * increase the "sequence" counter to avoid the race of an
-	 * etr event and the complete recovery against get_sync_clock.
+	 * stp event and the complete recovery against get_phys_clock.
 	 */
 	atomic_andnot(0x80000000, sw_ptr);
 	atomic_inc(sw_ptr);
 }
 
 /*
- * Make get_sync_clock return 0 again.
+ * Make get_phys_clock() return 0 again.
  * Needs to be called from a context disabled for preemption.
  */
 static void enable_sync_clock(void)
@@ -434,7 +384,7 @@ static inline int check_sync_clock(void)
 	return rc;
 }
 
-/* Single threaded workqueue used for etr and stp sync events */
+/* Single threaded workqueue used for stp sync events */
 static struct workqueue_struct *time_sync_wq;
 
 static void __init time_init_wq(void)
@@ -448,20 +398,12 @@ struct clock_sync_data {
 	atomic_t cpus;
 	int in_sync;
 	unsigned long long fixup_cc;
-	int etr_port;
-	struct etr_aib *etr_aib;
 };
 
 static void clock_sync_cpu(struct clock_sync_data *sync)
 {
 	atomic_dec(&sync->cpus);
 	enable_sync_clock();
-	/*
-	 * This looks like a busy wait loop but it isn't. etr_sync_cpus
-	 * is called on all other cpus while the TOD clocks is stopped.
-	 * __udelay will stop the cpu on an enabled wait psw until the
-	 * TOD is running again.
-	 */
 	while (sync->in_sync == 0) {
 		__udelay(1);
 		/*
@@ -582,7 +524,7 @@ void stp_queue_work(void)
 static int stp_sync_clock(void *data)
 {
 	static int first;
-	unsigned long long old_clock, delta, new_clock, clock_delta;
+	unsigned long long clock_delta;
 	struct clock_sync_data *stp_sync;
 	struct ptff_qto qto;
 	int rc;
@@ -605,18 +547,18 @@ static int stp_sync_clock(void *data)
 	if (stp_info.todoff[0] || stp_info.todoff[1] ||
 	    stp_info.todoff[2] || stp_info.todoff[3] ||
 	    stp_info.tmd != 2) {
-		old_clock = get_tod_clock();
 		rc = chsc_sstpc(stp_page, STP_OP_SYNC, 0, &clock_delta);
 		if (rc == 0) {
-			new_clock = old_clock + clock_delta;
-			delta = adjust_time(old_clock, new_clock, 0);
+			/* fixup the monotonic sched clock */
+			sched_clock_base_cc += clock_delta;
 			if (ptff_query(PTFF_QTO) &&
 			    ptff(&qto, sizeof(qto), PTFF_QTO) == 0)
 				/* Update LPAR offset */
 				lpar_offset = qto.tod_epoch_difference;
 			atomic_notifier_call_chain(&s390_epoch_delta_notifier,
 						   0, &clock_delta);
-			fixup_clock_comparator(delta);
+			stp_sync->fixup_cc = clock_delta;
+			fixup_clock_comparator(clock_delta);
 			rc = chsc_sstpi(stp_page, &stp_info,
 					sizeof(struct stp_sstpi));
 			if (rc == 0 && stp_info.tmd != 2)
diff --git a/arch/s390/kernel/traps.c b/arch/s390/kernel/traps.c
index dd97a3e..d0539f7 100644
--- a/arch/s390/kernel/traps.c
+++ b/arch/s390/kernel/traps.c
@@ -14,11 +14,12 @@
  */
 #include <linux/kprobes.h>
 #include <linux/kdebug.h>
-#include <linux/module.h>
+#include <linux/extable.h>
 #include <linux/ptrace.h>
 #include <linux/sched.h>
 #include <linux/mm.h>
 #include <linux/slab.h>
+#include <asm/uaccess.h>
 #include <asm/fpu/api.h>
 #include "entry.h"
 
diff --git a/arch/s390/kernel/vdso32/Makefile b/arch/s390/kernel/vdso32/Makefile
index 6814545..6cc9478 100644
--- a/arch/s390/kernel/vdso32/Makefile
+++ b/arch/s390/kernel/vdso32/Makefile
@@ -24,8 +24,9 @@ obj-y += vdso32_wrapper.o
 extra-y += vdso32.lds
 CPPFLAGS_vdso32.lds += -P -C -U$(ARCH)
 
-# Disable gcov profiling for VDSO code
+# Disable gcov profiling and ubsan for VDSO code
 GCOV_PROFILE := n
+UBSAN_SANITIZE := n
 
 # Force dependency (incbin is bad)
 $(obj)/vdso32_wrapper.o : $(obj)/vdso32.so
diff --git a/arch/s390/kernel/vdso64/Makefile b/arch/s390/kernel/vdso64/Makefile
index 0b0fd22..2d54c18 100644
--- a/arch/s390/kernel/vdso64/Makefile
+++ b/arch/s390/kernel/vdso64/Makefile
@@ -24,8 +24,9 @@ obj-y += vdso64_wrapper.o
 extra-y += vdso64.lds
 CPPFLAGS_vdso64.lds += -P -C -U$(ARCH)
 
-# Disable gcov profiling for VDSO code
+# Disable gcov profiling and ubsan for VDSO code
 GCOV_PROFILE := n
+UBSAN_SANITIZE := n
 
 # Force dependency (incbin is bad)
 $(obj)/vdso64_wrapper.o : $(obj)/vdso64.so