Upgrade to 4.4.50-rt62

The current kernel is based on rt kernel v4.4.6-rt14. We will upgrade it
to 4.4.50-rt62.
The command to achieve it is:
a) Clone a git repo from
git://git.kernel.org/pub/scm/linux/kernel/git/rt/linux-stable-rt.git

b) Get the diff between this two changesets:
git diff 640eca2901f3435e616157b11379d3223a44b391
705619beeea1b0b48219a683fd1a901a86fdaf5e

where the two commits are:
[yjiang5@jnakajim-build linux-stable-rt]$ git show --oneline --name-only
640eca2901f3435e616157b11379d3223a44b391
640eca2901f3 v4.4.6-rt14
localversion-rt
[yjiang5@jnakajim-build linux-stable-rt]$ git show --oneline --name-only
705619beeea1b0b48219a683fd1a901a86fdaf5e
705619beeea1 Linux 4.4.50-rt62
localversion-rt

c) One patch has been backported thus revert the patch before applying.

 filterdiff -p1 -x scripts/package/Makefile
~/tmp/v4.4.6-rt14-4.4.50-rt62.diff  |patch -p1 --dry-run

Upstream status: backport

Change-Id: I244d57a32f6066e5a5b9915f9fbf99e7bbca6e01
Signed-off-by: Yunhong Jiang <yunhong.jiang@linux.intel.com>

23 files changed, 119 insertions, 215 deletions

diff --git a/kernel/arch/s390/crypto/prng.c b/kernel/arch/s390/crypto/prng.c
index b8045b97f..683a966b5 100644
--- a/kernel/arch/s390/crypto/prng.c
+++ b/kernel/arch/s390/crypto/prng.c
@@ -565,8 +565,10 @@ static ssize_t prng_tdes_read(struct file *file, char __user *ubuf,
 		prng_data->prngws.byte_counter += n;
 		prng_data->prngws.reseed_counter += n;
 
-		if (copy_to_user(ubuf, prng_data->buf, chunk))
-			return -EFAULT;
+		if (copy_to_user(ubuf, prng_data->buf, chunk)) {
+			ret = -EFAULT;
+			break;
+		}
 
 		nbytes -= chunk;
 		ret += chunk;
@@ -669,11 +671,13 @@ static const struct file_operations prng_tdes_fops = {
 static struct miscdevice prng_sha512_dev = {
 	.name	= "prandom",
 	.minor	= MISC_DYNAMIC_MINOR,
+	.mode	= 0644,
 	.fops	= &prng_sha512_fops,
 };
 static struct miscdevice prng_tdes_dev = {
 	.name	= "prandom",
 	.minor	= MISC_DYNAMIC_MINOR,
+	.mode	= 0644,
 	.fops	= &prng_tdes_fops,
 };
 
diff --git a/kernel/arch/s390/hypfs/hypfs_diag.c b/kernel/arch/s390/hypfs/hypfs_diag.c
index 045035796..b63b9a42a 100644
--- a/kernel/arch/s390/hypfs/hypfs_diag.c
+++ b/kernel/arch/s390/hypfs/hypfs_diag.c
@@ -525,11 +525,11 @@ static int diag224(void *ptr)
 static int diag224_get_name_table(void)
 {
 	/* memory must be below 2GB */
-	diag224_cpu_names = kmalloc(PAGE_SIZE, GFP_KERNEL | GFP_DMA);
+	diag224_cpu_names = (char *) __get_free_page(GFP_KERNEL | GFP_DMA);
 	if (!diag224_cpu_names)
 		return -ENOMEM;
 	if (diag224(diag224_cpu_names)) {
-		kfree(diag224_cpu_names);
+		free_page((unsigned long) diag224_cpu_names);
 		return -EOPNOTSUPP;
 	}
 	EBCASC(diag224_cpu_names + 16, (*diag224_cpu_names + 1) * 16);
@@ -538,7 +538,7 @@ static int diag224_get_name_table(void)
 
 static void diag224_delete_name_table(void)
 {
-	kfree(diag224_cpu_names);
+	free_page((unsigned long) diag224_cpu_names);
 }
 
 static int diag224_idx2name(int index, char *name)
diff --git a/kernel/arch/s390/include/asm/fpu/api.h b/kernel/arch/s390/include/asm/fpu/api.h
index 5e04f3cbd..8ae236b0f 100644
--- a/kernel/arch/s390/include/asm/fpu/api.h
+++ b/kernel/arch/s390/include/asm/fpu/api.h
@@ -22,7 +22,7 @@ static inline int test_fp_ctl(u32 fpc)
 		"	la	%0,0\n"
 		"1:\n"
 		EX_TABLE(0b,1b)
-		: "=d" (rc), "=d" (orig_fpc)
+		: "=d" (rc), "=&d" (orig_fpc)
 		: "d" (fpc), "0" (-EINVAL));
 	return rc;
 }
diff --git a/kernel/arch/s390/include/asm/mmu.h b/kernel/arch/s390/include/asm/mmu.h
index d29ad9545..081b2ad99 100644
--- a/kernel/arch/s390/include/asm/mmu.h
+++ b/kernel/arch/s390/include/asm/mmu.h
@@ -11,7 +11,7 @@ typedef struct {
 	spinlock_t list_lock;
 	struct list_head pgtable_list;
 	struct list_head gmap_list;
-	unsigned long asce_bits;
+	unsigned long asce;
 	unsigned long asce_limit;
 	unsigned long vdso_base;
 	/* The mmu context allocates 4K page tables. */
diff --git a/kernel/arch/s390/include/asm/mmu_context.h b/kernel/arch/s390/include/asm/mmu_context.h
index e485817f7..22877c944 100644
--- a/kernel/arch/s390/include/asm/mmu_context.h
+++ b/kernel/arch/s390/include/asm/mmu_context.h
@@ -26,12 +26,28 @@ static inline int init_new_context(struct task_struct *tsk,
 	mm->context.has_pgste = 0;
 	mm->context.use_skey = 0;
 #endif
-	if (mm->context.asce_limit == 0) {
+	switch (mm->context.asce_limit) {
+	case 1UL << 42:
+		/*
+		 * forked 3-level task, fall through to set new asce with new
+		 * mm->pgd
+		 */
+	case 0:
 		/* context created by exec, set asce limit to 4TB */
-		mm->context.asce_bits = _ASCE_TABLE_LENGTH |
-			_ASCE_USER_BITS | _ASCE_TYPE_REGION3;
 		mm->context.asce_limit = STACK_TOP_MAX;
-	} else if (mm->context.asce_limit == (1UL << 31)) {
+		mm->context.asce = __pa(mm->pgd) | _ASCE_TABLE_LENGTH |
+				   _ASCE_USER_BITS | _ASCE_TYPE_REGION3;
+		break;
+	case 1UL << 53:
+		/* forked 4-level task, set new asce with new mm->pgd */
+		mm->context.asce = __pa(mm->pgd) | _ASCE_TABLE_LENGTH |
+				   _ASCE_USER_BITS | _ASCE_TYPE_REGION2;
+		break;
+	case 1UL << 31:
+		/* forked 2-level compat task, set new asce with new mm->pgd */
+		mm->context.asce = __pa(mm->pgd) | _ASCE_TABLE_LENGTH |
+				   _ASCE_USER_BITS | _ASCE_TYPE_SEGMENT;
+		/* pgd_alloc() did not increase mm->nr_pmds */
 		mm_inc_nr_pmds(mm);
 	}
 	crst_table_init((unsigned long *) mm->pgd, pgd_entry_type(mm));
@@ -42,7 +58,7 @@ static inline int init_new_context(struct task_struct *tsk,
 
 static inline void set_user_asce(struct mm_struct *mm)
 {
-	S390_lowcore.user_asce = mm->context.asce_bits | __pa(mm->pgd);
+	S390_lowcore.user_asce = mm->context.asce;
 	if (current->thread.mm_segment.ar4)
 		__ctl_load(S390_lowcore.user_asce, 7, 7);
 	set_cpu_flag(CIF_ASCE);
@@ -71,7 +87,7 @@ static inline void switch_mm(struct mm_struct *prev, struct mm_struct *next,
 {
 	int cpu = smp_processor_id();
 
-	S390_lowcore.user_asce = next->context.asce_bits | __pa(next->pgd);
+	S390_lowcore.user_asce = next->context.asce;
 	if (prev == next)
 		return;
 	if (MACHINE_HAS_TLB_LC)
diff --git a/kernel/arch/s390/include/asm/pci.h b/kernel/arch/s390/include/asm/pci.h
index c873e682b..6dafabb6a 100644
--- a/kernel/arch/s390/include/asm/pci.h
+++ b/kernel/arch/s390/include/asm/pci.h
@@ -45,7 +45,8 @@ struct zpci_fmb {
 	u64 rpcit_ops;
 	u64 dma_rbytes;
 	u64 dma_wbytes;
-} __packed __aligned(16);
+	u64 pad[2];
+} __packed __aligned(128);
 
 enum zpci_state {
 	ZPCI_FN_STATE_RESERVED,
diff --git a/kernel/arch/s390/include/asm/pci_dma.h b/kernel/arch/s390/include/asm/pci_dma.h
index 1aac41e83..92df3eb8d 100644
--- a/kernel/arch/s390/include/asm/pci_dma.h
+++ b/kernel/arch/s390/include/asm/pci_dma.h
@@ -23,6 +23,8 @@ enum zpci_ioat_dtype {
 #define ZPCI_IOTA_FS_2G			2
 #define ZPCI_KEY			(PAGE_DEFAULT_KEY << 5)
 
+#define ZPCI_TABLE_SIZE_RT	(1UL << 42)
+
 #define ZPCI_IOTA_STO_FLAG	(ZPCI_IOTA_IOT_ENABLED | ZPCI_KEY | ZPCI_IOTA_DT_ST)
 #define ZPCI_IOTA_RTTO_FLAG	(ZPCI_IOTA_IOT_ENABLED | ZPCI_KEY | ZPCI_IOTA_DT_RT)
 #define ZPCI_IOTA_RSTO_FLAG	(ZPCI_IOTA_IOT_ENABLED | ZPCI_KEY | ZPCI_IOTA_DT_RS)
diff --git a/kernel/arch/s390/include/asm/pgalloc.h b/kernel/arch/s390/include/asm/pgalloc.h
index d7cc79fb6..5991cdcb5 100644
--- a/kernel/arch/s390/include/asm/pgalloc.h
+++ b/kernel/arch/s390/include/asm/pgalloc.h
@@ -56,8 +56,8 @@ static inline unsigned long pgd_entry_type(struct mm_struct *mm)
 	return _REGION2_ENTRY_EMPTY;
 }
 
-int crst_table_upgrade(struct mm_struct *, unsigned long limit);
-void crst_table_downgrade(struct mm_struct *, unsigned long limit);
+int crst_table_upgrade(struct mm_struct *);
+void crst_table_downgrade(struct mm_struct *);
 
 static inline pud_t *pud_alloc_one(struct mm_struct *mm, unsigned long address)
 {
diff --git a/kernel/arch/s390/include/asm/processor.h b/kernel/arch/s390/include/asm/processor.h
index b16c3d0a1..c1ea67db8 100644
--- a/kernel/arch/s390/include/asm/processor.h
+++ b/kernel/arch/s390/include/asm/processor.h
@@ -163,7 +163,7 @@ extern __vector128 init_task_fpu_regs[__NUM_VXRS];
 	regs->psw.mask	= PSW_USER_BITS | PSW_MASK_BA;			\
 	regs->psw.addr	= new_psw | PSW_ADDR_AMODE;			\
 	regs->gprs[15]	= new_stackp;					\
-	crst_table_downgrade(current->mm, 1UL << 31);			\
+	crst_table_downgrade(current->mm);				\
 	execve_tail();							\
 } while (0)
 
diff --git a/kernel/arch/s390/include/asm/tlbflush.h b/kernel/arch/s390/include/asm/tlbflush.h
index ca148f7c3..0a2031618 100644
--- a/kernel/arch/s390/include/asm/tlbflush.h
+++ b/kernel/arch/s390/include/asm/tlbflush.h
@@ -81,7 +81,8 @@ static inline void __tlb_flush_full(struct mm_struct *mm)
 }
 
 /*
- * Flush TLB entries for a specific ASCE on all CPUs.
+ * Flush TLB entries for a specific ASCE on all CPUs. Should never be used
+ * when more than one asce (e.g. gmap) ran on this mm.
  */
 static inline void __tlb_flush_asce(struct mm_struct *mm, unsigned long asce)
 {
@@ -110,8 +111,7 @@ static inline void __tlb_flush_asce(struct mm_struct *mm, unsigned long asce)
 static inline void __tlb_flush_kernel(void)
 {
 	if (MACHINE_HAS_IDTE)
-		__tlb_flush_idte((unsigned long) init_mm.pgd |
-				 init_mm.context.asce_bits);
+		__tlb_flush_idte(init_mm.context.asce);
 	else
 		__tlb_flush_global();
 }
@@ -133,8 +133,7 @@ static inline void __tlb_flush_asce(struct mm_struct *mm, unsigned long asce)
 static inline void __tlb_flush_kernel(void)
 {
 	if (MACHINE_HAS_TLB_LC)
-		__tlb_flush_idte_local((unsigned long) init_mm.pgd |
-				       init_mm.context.asce_bits);
+		__tlb_flush_idte_local(init_mm.context.asce);
 	else
 		__tlb_flush_local();
 }
@@ -148,8 +147,7 @@ static inline void __tlb_flush_mm(struct mm_struct * mm)
 	 * only ran on the local cpu.
 	 */
 	if (MACHINE_HAS_IDTE && list_empty(&mm->context.gmap_list))
-		__tlb_flush_asce(mm, (unsigned long) mm->pgd |
-				 mm->context.asce_bits);
+		__tlb_flush_asce(mm, mm->context.asce);
 	else
 		__tlb_flush_full(mm);
 }
diff --git a/kernel/arch/s390/include/asm/uaccess.h b/kernel/arch/s390/include/asm/uaccess.h
index 9dd4cc47d..5c7381c5a 100644
--- a/kernel/arch/s390/include/asm/uaccess.h
+++ b/kernel/arch/s390/include/asm/uaccess.h
@@ -215,28 +215,28 @@ int __put_user_bad(void) __attribute__((noreturn));
 	__chk_user_ptr(ptr);					\
 	switch (sizeof(*(ptr))) {				\
 	case 1: {						\
-		unsigned char __x;				\
+		unsigned char __x = 0;				\
 		__gu_err = __get_user_fn(&__x, ptr,		\
 					 sizeof(*(ptr)));	\
 		(x) = *(__force __typeof__(*(ptr)) *) &__x;	\
 		break;						\
 	};							\
 	case 2: {						\
-		unsigned short __x;				\
+		unsigned short __x = 0;				\
 		__gu_err = __get_user_fn(&__x, ptr,		\
 					 sizeof(*(ptr)));	\
 		(x) = *(__force __typeof__(*(ptr)) *) &__x;	\
 		break;						\
 	};							\
 	case 4: {						\
-		unsigned int __x;				\
+		unsigned int __x = 0;				\
 		__gu_err = __get_user_fn(&__x, ptr,		\
 					 sizeof(*(ptr)));	\
 		(x) = *(__force __typeof__(*(ptr)) *) &__x;	\
 		break;						\
 	};							\
 	case 8: {						\
-		unsigned long long __x;				\
+		unsigned long long __x = 0;			\
 		__gu_err = __get_user_fn(&__x, ptr,		\
 					 sizeof(*(ptr)));	\
 		(x) = *(__force __typeof__(*(ptr)) *) &__x;	\
diff --git a/kernel/arch/s390/kernel/entry.S b/kernel/arch/s390/kernel/entry.S
index 857b6526d..424e6809a 100644
--- a/kernel/arch/s390/kernel/entry.S
+++ b/kernel/arch/s390/kernel/entry.S
@@ -1197,114 +1197,12 @@ cleanup_critical:
 	.quad	.Lpsw_idle_lpsw
 
 .Lcleanup_save_fpu_regs:
-	TSTMSK	__LC_CPU_FLAGS,_CIF_FPU
-	bor	%r14
-	clg	%r9,BASED(.Lcleanup_save_fpu_regs_done)
-	jhe	5f
-	clg	%r9,BASED(.Lcleanup_save_fpu_regs_fp)
-	jhe	4f
-	clg	%r9,BASED(.Lcleanup_save_fpu_regs_vx_high)
-	jhe	3f
-	clg	%r9,BASED(.Lcleanup_save_fpu_regs_vx_low)
-	jhe	2f
-	clg	%r9,BASED(.Lcleanup_save_fpu_fpc_end)
-	jhe	1f
-	lg	%r2,__LC_CURRENT
-	aghi	%r2,__TASK_thread
-0:	# Store floating-point controls
-	stfpc	__THREAD_FPU_fpc(%r2)
-1:	# Load register save area and check if VX is active
-	lg	%r3,__THREAD_FPU_regs(%r2)
-	TSTMSK	__LC_MACHINE_FLAGS,MACHINE_FLAG_VX
-	jz	4f			  # no VX -> store FP regs
-2:	# Store vector registers (V0-V15)
-	VSTM	%v0,%v15,0,%r3		  # vstm 0,15,0(3)
-3:	# Store vector registers (V16-V31)
-	VSTM	%v16,%v31,256,%r3	  # vstm 16,31,256(3)
-	j	5f			  # -> done, set CIF_FPU flag
-4:	# Store floating-point registers
-	std	0,0(%r3)
-	std	1,8(%r3)
-	std	2,16(%r3)
-	std	3,24(%r3)
-	std	4,32(%r3)
-	std	5,40(%r3)
-	std	6,48(%r3)
-	std	7,56(%r3)
-	std	8,64(%r3)
-	std	9,72(%r3)
-	std	10,80(%r3)
-	std	11,88(%r3)
-	std	12,96(%r3)
-	std	13,104(%r3)
-	std	14,112(%r3)
-	std	15,120(%r3)
-5:	# Set CIF_FPU flag
-	oi	__LC_CPU_FLAGS+7,_CIF_FPU
-	lg	%r9,48(%r11)		# return from save_fpu_regs
+	larl	%r9,save_fpu_regs
 	br	%r14
-.Lcleanup_save_fpu_fpc_end:
-	.quad	.Lsave_fpu_regs_fpc_end
-.Lcleanup_save_fpu_regs_vx_low:
-	.quad	.Lsave_fpu_regs_vx_low
-.Lcleanup_save_fpu_regs_vx_high:
-	.quad	.Lsave_fpu_regs_vx_high
-.Lcleanup_save_fpu_regs_fp:
-	.quad	.Lsave_fpu_regs_fp
-.Lcleanup_save_fpu_regs_done:
-	.quad	.Lsave_fpu_regs_done
 
 .Lcleanup_load_fpu_regs:
-	TSTMSK	__LC_CPU_FLAGS,_CIF_FPU
-	bnor	%r14
-	clg	%r9,BASED(.Lcleanup_load_fpu_regs_done)
-	jhe	1f
-	clg	%r9,BASED(.Lcleanup_load_fpu_regs_fp)
-	jhe	2f
-	clg	%r9,BASED(.Lcleanup_load_fpu_regs_vx_high)
-	jhe	3f
-	clg	%r9,BASED(.Lcleanup_load_fpu_regs_vx)
-	jhe	4f
-	lg	%r4,__LC_CURRENT
-	aghi	%r4,__TASK_thread
-	lfpc	__THREAD_FPU_fpc(%r4)
-	TSTMSK	__LC_MACHINE_FLAGS,MACHINE_FLAG_VX
-	lg	%r4,__THREAD_FPU_regs(%r4)	# %r4 <- reg save area
-	jz	2f				# -> no VX, load FP regs
-4:	# Load V0 ..V15 registers
-	VLM	%v0,%v15,0,%r4
-3:	# Load V16..V31 registers
-	VLM	%v16,%v31,256,%r4
-	j	1f
-2:	# Load floating-point registers
-	ld	0,0(%r4)
-	ld	1,8(%r4)
-	ld	2,16(%r4)
-	ld	3,24(%r4)
-	ld	4,32(%r4)
-	ld	5,40(%r4)
-	ld	6,48(%r4)
-	ld	7,56(%r4)
-	ld	8,64(%r4)
-	ld	9,72(%r4)
-	ld	10,80(%r4)
-	ld	11,88(%r4)
-	ld	12,96(%r4)
-	ld	13,104(%r4)
-	ld	14,112(%r4)
-	ld	15,120(%r4)
-1:	# Clear CIF_FPU bit
-	ni	__LC_CPU_FLAGS+7,255-_CIF_FPU
-	lg	%r9,48(%r11)		# return from load_fpu_regs
+	larl	%r9,load_fpu_regs
 	br	%r14
-.Lcleanup_load_fpu_regs_vx:
-	.quad	.Lload_fpu_regs_vx
-.Lcleanup_load_fpu_regs_vx_high:
-	.quad	.Lload_fpu_regs_vx_high
-.Lcleanup_load_fpu_regs_fp:
-	.quad	.Lload_fpu_regs_fp
-.Lcleanup_load_fpu_regs_done:
-	.quad	.Lload_fpu_regs_done
 
 /*
  * Integer constants
diff --git a/kernel/arch/s390/kernel/head64.S b/kernel/arch/s390/kernel/head64.S
index 58b719fa8..1ad2407c7 100644
--- a/kernel/arch/s390/kernel/head64.S
+++ b/kernel/arch/s390/kernel/head64.S
@@ -16,7 +16,7 @@
 
 __HEAD
 ENTRY(startup_continue)
-	tm	__LC_STFL_FAC_LIST+6,0x80	# LPP available ?
+	tm	__LC_STFL_FAC_LIST+5,0x80	# LPP available ?
 	jz	0f
 	xc	__LC_LPP+1(7,0),__LC_LPP+1	# clear lpp and current_pid
 	mvi	__LC_LPP,0x80			#   and set LPP_MAGIC
diff --git a/kernel/arch/s390/kernel/ipl.c b/kernel/arch/s390/kernel/ipl.c
index b1f0a90f9..42570d8fb 100644
--- a/kernel/arch/s390/kernel/ipl.c
+++ b/kernel/arch/s390/kernel/ipl.c
@@ -2070,13 +2070,6 @@ void s390_reset_system(void (*fn_pre)(void),
 	S390_lowcore.program_new_psw.addr =
 		PSW_ADDR_AMODE | (unsigned long) s390_base_pgm_handler;
 
-	/*
-	 * Clear subchannel ID and number to signal new kernel that no CCW or
-	 * SCSI IPL has been done (for kexec and kdump)
-	 */
-	S390_lowcore.subchannel_id = 0;
-	S390_lowcore.subchannel_nr = 0;
-
 	/* Store status at absolute zero */
 	store_status();
 
diff --git a/kernel/arch/s390/kernel/ptrace.c b/kernel/arch/s390/kernel/ptrace.c
index 01c37b36c..02bd587b6 100644
--- a/kernel/arch/s390/kernel/ptrace.c
+++ b/kernel/arch/s390/kernel/ptrace.c
@@ -963,6 +963,11 @@ static int s390_fpregs_set(struct task_struct *target,
 	if (target == current)
 		save_fpu_regs();
 
+	if (MACHINE_HAS_VX)
+		convert_vx_to_fp(fprs, target->thread.fpu.vxrs);
+	else
+		memcpy(&fprs, target->thread.fpu.fprs, sizeof(fprs));
+
 	/* If setting FPC, must validate it first. */
 	if (count > 0 && pos < offsetof(s390_fp_regs, fprs)) {
 		u32 ufpc[2] = { target->thread.fpu.fpc, 0 };
@@ -1067,6 +1072,9 @@ static int s390_vxrs_low_set(struct task_struct *target,
 	if (target == current)
 		save_fpu_regs();
 
+	for (i = 0; i < __NUM_VXRS_LOW; i++)
+		vxrs[i] = *((__u64 *)(target->thread.fpu.vxrs + i) + 1);
+
 	rc = user_regset_copyin(&pos, &count, &kbuf, &ubuf, vxrs, 0, -1);
 	if (rc == 0)
 		for (i = 0; i < __NUM_VXRS_LOW; i++)
diff --git a/kernel/arch/s390/kernel/setup.c b/kernel/arch/s390/kernel/setup.c
index c837bcacf..1f581eb61 100644
--- a/kernel/arch/s390/kernel/setup.c
+++ b/kernel/arch/s390/kernel/setup.c
@@ -329,6 +329,7 @@ static void __init setup_lowcore(void)
 		+ PAGE_SIZE - STACK_FRAME_OVERHEAD - sizeof(struct pt_regs);
 	lc->current_task = (unsigned long) init_thread_union.thread_info.task;
 	lc->thread_info = (unsigned long) &init_thread_union;
+	lc->lpp = LPP_MAGIC;
 	lc->machine_flags = S390_lowcore.machine_flags;
 	lc->stfl_fac_list = S390_lowcore.stfl_fac_list;
 	memcpy(lc->stfle_fac_list, S390_lowcore.stfle_fac_list,
diff --git a/kernel/arch/s390/mm/init.c b/kernel/arch/s390/mm/init.c
index c722400c7..feff9caf8 100644
--- a/kernel/arch/s390/mm/init.c
+++ b/kernel/arch/s390/mm/init.c
@@ -89,7 +89,8 @@ void __init paging_init(void)
 		asce_bits = _ASCE_TYPE_REGION3 | _ASCE_TABLE_LENGTH;
 		pgd_type = _REGION3_ENTRY_EMPTY;
 	}
-	S390_lowcore.kernel_asce = (__pa(init_mm.pgd) & PAGE_MASK) | asce_bits;
+	init_mm.context.asce = (__pa(init_mm.pgd) & PAGE_MASK) | asce_bits;
+	S390_lowcore.kernel_asce = init_mm.context.asce;
 	clear_table((unsigned long *) init_mm.pgd, pgd_type,
 		    sizeof(unsigned long)*2048);
 	vmem_map_init();
diff --git a/kernel/arch/s390/mm/mmap.c b/kernel/arch/s390/mm/mmap.c
index ea01477b4..f2b6b1d9c 100644
--- a/kernel/arch/s390/mm/mmap.c
+++ b/kernel/arch/s390/mm/mmap.c
@@ -174,7 +174,7 @@ int s390_mmap_check(unsigned long addr, unsigned long len, unsigned long flags)
 	if (!(flags & MAP_FIXED))
 		addr = 0;
 	if ((addr + len) >= TASK_SIZE)
-		return crst_table_upgrade(current->mm, 1UL << 53);
+		return crst_table_upgrade(current->mm);
 	return 0;
 }
 
@@ -191,7 +191,7 @@ s390_get_unmapped_area(struct file *filp, unsigned long addr,
 		return area;
 	if (area == -ENOMEM && !is_compat_task() && TASK_SIZE < (1UL << 53)) {
 		/* Upgrade the page table to 4 levels and retry. */
-		rc = crst_table_upgrade(mm, 1UL << 53);
+		rc = crst_table_upgrade(mm);
 		if (rc)
 			return (unsigned long) rc;
 		area = arch_get_unmapped_area(filp, addr, len, pgoff, flags);
@@ -213,7 +213,7 @@ s390_get_unmapped_area_topdown(struct file *filp, const unsigned long addr,
 		return area;
 	if (area == -ENOMEM && !is_compat_task() && TASK_SIZE < (1UL << 53)) {
 		/* Upgrade the page table to 4 levels and retry. */
-		rc = crst_table_upgrade(mm, 1UL << 53);
+		rc = crst_table_upgrade(mm);
 		if (rc)
 			return (unsigned long) rc;
 		area = arch_get_unmapped_area_topdown(filp, addr, len,
diff --git a/kernel/arch/s390/mm/pgtable.c b/kernel/arch/s390/mm/pgtable.c
index 54ef3bc01..8345ae1f1 100644
--- a/kernel/arch/s390/mm/pgtable.c
+++ b/kernel/arch/s390/mm/pgtable.c
@@ -49,81 +49,52 @@ static void __crst_table_upgrade(void *arg)
 	__tlb_flush_local();
 }
 
-int crst_table_upgrade(struct mm_struct *mm, unsigned long limit)
+int crst_table_upgrade(struct mm_struct *mm)
 {
 	unsigned long *table, *pgd;
-	unsigned long entry;
-	int flush;
 
-	BUG_ON(limit > (1UL << 53));
-	flush = 0;
-repeat:
+	/* upgrade should only happen from 3 to 4 levels */
+	BUG_ON(mm->context.asce_limit != (1UL << 42));
+
 	table = crst_table_alloc(mm);
 	if (!table)
 		return -ENOMEM;
+
 	spin_lock_bh(&mm->page_table_lock);
-	if (mm->context.asce_limit < limit) {
-		pgd = (unsigned long *) mm->pgd;
-		if (mm->context.asce_limit <= (1UL << 31)) {
-			entry = _REGION3_ENTRY_EMPTY;
-			mm->context.asce_limit = 1UL << 42;
-			mm->context.asce_bits = _ASCE_TABLE_LENGTH |
-						_ASCE_USER_BITS |
-						_ASCE_TYPE_REGION3;
-		} else {
-			entry = _REGION2_ENTRY_EMPTY;
-			mm->context.asce_limit = 1UL << 53;
-			mm->context.asce_bits = _ASCE_TABLE_LENGTH |
-						_ASCE_USER_BITS |
-						_ASCE_TYPE_REGION2;
-		}
-		crst_table_init(table, entry);
-		pgd_populate(mm, (pgd_t *) table, (pud_t *) pgd);
-		mm->pgd = (pgd_t *) table;
-		mm->task_size = mm->context.asce_limit;
-		table = NULL;
-		flush = 1;
-	}
+	pgd = (unsigned long *) mm->pgd;
+	crst_table_init(table, _REGION2_ENTRY_EMPTY);
+	pgd_populate(mm, (pgd_t *) table, (pud_t *) pgd);
+	mm->pgd = (pgd_t *) table;
+	mm->context.asce_limit = 1UL << 53;
+	mm->context.asce = __pa(mm->pgd) | _ASCE_TABLE_LENGTH |
+			   _ASCE_USER_BITS | _ASCE_TYPE_REGION2;
+	mm->task_size = mm->context.asce_limit;
 	spin_unlock_bh(&mm->page_table_lock);
-	if (table)
-		crst_table_free(mm, table);
-	if (mm->context.asce_limit < limit)
-		goto repeat;
-	if (flush)
-		on_each_cpu(__crst_table_upgrade, mm, 0);
+
+	on_each_cpu(__crst_table_upgrade, mm, 0);
 	return 0;
 }
 
-void crst_table_downgrade(struct mm_struct *mm, unsigned long limit)
+void crst_table_downgrade(struct mm_struct *mm)
 {
 	pgd_t *pgd;
 
+	/* downgrade should only happen from 3 to 2 levels (compat only) */
+	BUG_ON(mm->context.asce_limit != (1UL << 42));
+
 	if (current->active_mm == mm) {
 		clear_user_asce();
 		__tlb_flush_mm(mm);
 	}
-	while (mm->context.asce_limit > limit) {
-		pgd = mm->pgd;
-		switch (pgd_val(*pgd) & _REGION_ENTRY_TYPE_MASK) {
-		case _REGION_ENTRY_TYPE_R2:
-			mm->context.asce_limit = 1UL << 42;
-			mm->context.asce_bits = _ASCE_TABLE_LENGTH |
-						_ASCE_USER_BITS |
-						_ASCE_TYPE_REGION3;
-			break;
-		case _REGION_ENTRY_TYPE_R3:
-			mm->context.asce_limit = 1UL << 31;
-			mm->context.asce_bits = _ASCE_TABLE_LENGTH |
-						_ASCE_USER_BITS |
-						_ASCE_TYPE_SEGMENT;
-			break;
-		default:
-			BUG();
-		}
-		mm->pgd = (pgd_t *) (pgd_val(*pgd) & _REGION_ENTRY_ORIGIN);
-		mm->task_size = mm->context.asce_limit;
-		crst_table_free(mm, (unsigned long *) pgd);
-	}
+
+	pgd = mm->pgd;
+	mm->pgd = (pgd_t *) (pgd_val(*pgd) & _REGION_ENTRY_ORIGIN);
+	mm->context.asce_limit = 1UL << 31;
+	mm->context.asce = __pa(mm->pgd) | _ASCE_TABLE_LENGTH |
+			   _ASCE_USER_BITS | _ASCE_TYPE_SEGMENT;
+	mm->task_size = mm->context.asce_limit;
+	crst_table_free(mm, (unsigned long *) pgd);
+
 	if (current->active_mm == mm)
 		set_user_asce(mm);
 }
@@ -195,7 +166,7 @@ EXPORT_SYMBOL_GPL(gmap_alloc);
 static void gmap_flush_tlb(struct gmap *gmap)
 {
 	if (MACHINE_HAS_IDTE)
-		__tlb_flush_asce(gmap->mm, gmap->asce);
+		__tlb_flush_idte(gmap->asce);
 	else
 		__tlb_flush_global();
 }
@@ -234,7 +205,7 @@ void gmap_free(struct gmap *gmap)
 
 	/* Flush tlb. */
 	if (MACHINE_HAS_IDTE)
-		__tlb_flush_asce(gmap->mm, gmap->asce);
+		__tlb_flush_idte(gmap->asce);
 	else
 		__tlb_flush_global();
 
diff --git a/kernel/arch/s390/net/bpf_jit.h b/kernel/arch/s390/net/bpf_jit.h
index f010c93a8..fda605dbc 100644
--- a/kernel/arch/s390/net/bpf_jit.h
+++ b/kernel/arch/s390/net/bpf_jit.h
@@ -37,7 +37,7 @@ extern u8 sk_load_word[], sk_load_half[], sk_load_byte[];
  *	      |		      |     |
  *	      +---------------+     |
  *	      | 8 byte skbp   |     |
- * R15+170 -> +---------------+     |
+ * R15+176 -> +---------------+     |
  *	      | 8 byte hlen   |     |
  * R15+168 -> +---------------+     |
  *	      | 4 byte align  |     |
@@ -58,7 +58,7 @@ extern u8 sk_load_word[], sk_load_half[], sk_load_byte[];
 #define STK_OFF		(STK_SPACE - STK_160_UNUSED)
 #define STK_OFF_TMP	160	/* Offset of tmp buffer on stack */
 #define STK_OFF_HLEN	168	/* Offset of SKB header length on stack */
-#define STK_OFF_SKBP	170	/* Offset of SKB pointer on stack */
+#define STK_OFF_SKBP	176	/* Offset of SKB pointer on stack */
 
 #define STK_OFF_R6	(160 - 11 * 8)	/* Offset of r6 on stack */
 #define STK_OFF_TCCNT	(160 - 12 * 8)	/* Offset of tail_call_cnt on stack */
diff --git a/kernel/arch/s390/net/bpf_jit_comp.c b/kernel/arch/s390/net/bpf_jit_comp.c
index 9a0c4c22e..0e2919dd8 100644
--- a/kernel/arch/s390/net/bpf_jit_comp.c
+++ b/kernel/arch/s390/net/bpf_jit_comp.c
@@ -45,7 +45,7 @@ struct bpf_jit {
 	int labels[1];		/* Labels for local jumps */
 };
 
-#define BPF_SIZE_MAX	0x7ffff	/* Max size for program (20 bit signed displ) */
+#define BPF_SIZE_MAX	0xffff	/* Max size for program (16 bit branches) */
 
 #define SEEN_SKB	1	/* skb access */
 #define SEEN_MEM	2	/* use mem[] for temporary storage */
@@ -446,7 +446,7 @@ static void bpf_jit_prologue(struct bpf_jit *jit, bool is_classic)
 		emit_load_skb_data_hlen(jit);
 	if (jit->seen & SEEN_SKB_CHANGE)
 		/* stg %b1,ST_OFF_SKBP(%r0,%r15) */
-		EMIT6_DISP_LH(0xe3000000, 0x0024, REG_W1, REG_0, REG_15,
+		EMIT6_DISP_LH(0xe3000000, 0x0024, BPF_REG_1, REG_0, REG_15,
 			      STK_OFF_SKBP);
 	/* Clear A (%b0) and X (%b7) registers for converted BPF programs */
 	if (is_classic) {
diff --git a/kernel/arch/s390/pci/pci.c b/kernel/arch/s390/pci/pci.c
index 7ef12a3ac..f2f6720a3 100644
--- a/kernel/arch/s390/pci/pci.c
+++ b/kernel/arch/s390/pci/pci.c
@@ -701,8 +701,7 @@ static int zpci_restore(struct device *dev)
 		goto out;
 
 	zpci_map_resources(pdev);
-	zpci_register_ioat(zdev, 0, zdev->start_dma + PAGE_OFFSET,
-			   zdev->start_dma + zdev->iommu_size - 1,
+	zpci_register_ioat(zdev, 0, zdev->start_dma, zdev->end_dma,
 			   (u64) zdev->dma_table);
 
 out:
@@ -871,8 +870,11 @@ static inline int barsize(u8 size)
 
 static int zpci_mem_init(void)
 {
+	BUILD_BUG_ON(!is_power_of_2(__alignof__(struct zpci_fmb)) ||
+		     __alignof__(struct zpci_fmb) < sizeof(struct zpci_fmb));
+
 	zdev_fmb_cache = kmem_cache_create("PCI_FMB_cache", sizeof(struct zpci_fmb),
-				16, 0, NULL);
+					   __alignof__(struct zpci_fmb), 0, NULL);
 	if (!zdev_fmb_cache)
 		goto error_zdev;
 
diff --git a/kernel/arch/s390/pci/pci_dma.c b/kernel/arch/s390/pci/pci_dma.c
index d348f2c09..3a40f718b 100644
--- a/kernel/arch/s390/pci/pci_dma.c
+++ b/kernel/arch/s390/pci/pci_dma.c
@@ -458,7 +458,19 @@ int zpci_dma_init_device(struct zpci_dev *zdev)
 		goto out_clean;
 	}
 
-	zdev->iommu_size = (unsigned long) high_memory - PAGE_OFFSET;
+	/*
+	 * Restrict the iommu bitmap size to the minimum of the following:
+	 * - main memory size
+	 * - 3-level pagetable address limit minus start_dma offset
+	 * - DMA address range allowed by the hardware (clp query pci fn)
+	 *
+	 * Also set zdev->end_dma to the actual end address of the usable
+	 * range, instead of the theoretical maximum as reported by hardware.
+	 */
+	zdev->iommu_size = min3((u64) high_memory,
+				ZPCI_TABLE_SIZE_RT - zdev->start_dma,
+				zdev->end_dma - zdev->start_dma + 1);
+	zdev->end_dma = zdev->start_dma + zdev->iommu_size - 1;
 	zdev->iommu_pages = zdev->iommu_size >> PAGE_SHIFT;
 	zdev->iommu_bitmap = vzalloc(zdev->iommu_pages / 8);
 	if (!zdev->iommu_bitmap) {
@@ -466,10 +478,7 @@ int zpci_dma_init_device(struct zpci_dev *zdev)
 		goto out_reg;
 	}
 
-	rc = zpci_register_ioat(zdev,
-				0,
-				zdev->start_dma + PAGE_OFFSET,
-				zdev->start_dma + zdev->iommu_size - 1,
+	rc = zpci_register_ioat(zdev, 0, zdev->start_dma, zdev->end_dma,
 				(u64) zdev->dma_table);
 	if (rc)
 		goto out_reg;