6 files changed, 68 insertions, 132 deletions

diff --git a/arch/s390/mm/cmm.c b/arch/s390/mm/cmm.c
index 9d84a1f..79ddd58 100644
--- a/arch/s390/mm/cmm.c
+++ b/arch/s390/mm/cmm.c
@@ -253,12 +253,12 @@ static int cmm_skip_blanks(char *cp, char **endp)
 
 static struct ctl_table cmm_table[];
 
-static int cmm_pages_handler(ctl_table *ctl, int write, void __user *buffer,
-			     size_t *lenp, loff_t *ppos)
+static int cmm_pages_handler(struct ctl_table *ctl, int write,
+			     void __user *buffer, size_t *lenp, loff_t *ppos)
 {
 	char buf[16], *p;
+	unsigned int len;
 	long nr;
-	int len;
 
 	if (!*lenp || (*ppos && !write)) {
 		*lenp = 0;
@@ -293,12 +293,12 @@ static int cmm_pages_handler(ctl_table *ctl, int write, void __user *buffer,
 	return 0;
 }
 
-static int cmm_timeout_handler(ctl_table *ctl, int write,  void __user *buffer,
-			       size_t *lenp, loff_t *ppos)
+static int cmm_timeout_handler(struct ctl_table *ctl, int write,
+			       void __user *buffer, size_t *lenp, loff_t *ppos)
 {
 	char buf[64], *p;
 	long nr, seconds;
-	int len;
+	unsigned int len;
 
 	if (!*lenp || (*ppos && !write)) {
 		*lenp = 0;
diff --git a/arch/s390/mm/fault.c b/arch/s390/mm/fault.c
index fc66792..d95265b2 100644
--- a/arch/s390/mm/fault.c
+++ b/arch/s390/mm/fault.c
@@ -115,13 +115,8 @@ static inline int user_space_fault(unsigned long trans_exc_code)
 	if (trans_exc_code == 2)
 		/* Access via secondary space, set_fs setting decides */
 		return current->thread.mm_segment.ar4;
-	if (s390_user_mode == HOME_SPACE_MODE)
-		/* User space if the access has been done via home space. */
-		return trans_exc_code == 3;
 	/*
-	 * If the user space is not the home space the kernel runs in home
-	 * space. Access via secondary space has already been covered,
-	 * access via primary space or access register is from user space
+	 * Access via primary space or access register is from user space
 	 * and access via home space is from the kernel.
 	 */
 	return trans_exc_code != 3;
@@ -428,50 +423,13 @@ void __kprobes do_dat_exception(struct pt_regs *regs)
 		do_fault_error(regs, fault);
 }
 
-#ifdef CONFIG_64BIT
-void __kprobes do_asce_exception(struct pt_regs *regs)
-{
-	struct mm_struct *mm = current->mm;
-	struct vm_area_struct *vma;
-	unsigned long trans_exc_code;
-
-	/*
-	 * The instruction that caused the program check has
-	 * been nullified. Don't signal single step via SIGTRAP.
-	 */
-	clear_tsk_thread_flag(current, TIF_PER_TRAP);
-
-	trans_exc_code = regs->int_parm_long;
-	if (unlikely(!user_space_fault(trans_exc_code) || in_atomic() || !mm))
-		goto no_context;
-
-	down_read(&mm->mmap_sem);
-	vma = find_vma(mm, trans_exc_code & __FAIL_ADDR_MASK);
-	up_read(&mm->mmap_sem);
-
-	if (vma) {
-		update_mm(mm, current);
-		return;
-	}
-
-	/* User mode accesses just cause a SIGSEGV */
-	if (user_mode(regs)) {
-		do_sigsegv(regs, SEGV_MAPERR);
-		return;
-	}
-
-no_context:
-	do_no_context(regs);
-}
-#endif
-
 int __handle_fault(unsigned long uaddr, unsigned long pgm_int_code, int write)
 {
 	struct pt_regs regs;
 	int access, fault;
 
 	/* Emulate a uaccess fault from kernel mode. */
-	regs.psw.mask = psw_kernel_bits | PSW_MASK_DAT | PSW_MASK_MCHECK;
+	regs.psw.mask = PSW_KERNEL_BITS | PSW_MASK_DAT | PSW_MASK_MCHECK;
 	if (!irqs_disabled())
 		regs.psw.mask |= PSW_MASK_IO | PSW_MASK_EXT;
 	regs.psw.addr = (unsigned long) __builtin_return_address(0);
diff --git a/arch/s390/mm/gup.c b/arch/s390/mm/gup.c
index 5d758db..639fce46 100644
--- a/arch/s390/mm/gup.c
+++ b/arch/s390/mm/gup.c
@@ -180,9 +180,15 @@ int __get_user_pages_fast(unsigned long start, int nr_pages, int write,
 	addr = start;
 	len = (unsigned long) nr_pages << PAGE_SHIFT;
 	end = start + len;
-	if ((end < start) || (end > TASK_SIZE))
+	if ((end <= start) || (end > TASK_SIZE))
 		return 0;
-
+	/*
+	 * local_irq_save() doesn't prevent pagetable teardown, but does
+	 * prevent the pagetables from being freed on s390.
+	 *
+	 * So long as we atomically load page table pointers versus teardown,
+	 * we can follow the address down to the the page and take a ref on it.
+	 */
 	local_irq_save(flags);
 	pgdp = pgd_offset(mm, addr);
 	do {
@@ -219,63 +225,22 @@ int get_user_pages_fast(unsigned long start, int nr_pages, int write,
 			struct page **pages)
 {
 	struct mm_struct *mm = current->mm;
-	unsigned long addr, len, end;
-	unsigned long next;
-	pgd_t *pgdp, pgd;
-	int nr = 0;
+	int nr, ret;
 
 	start &= PAGE_MASK;
-	addr = start;
-	len = (unsigned long) nr_pages << PAGE_SHIFT;
-	end = start + len;
-	if ((end < start) || (end > TASK_SIZE))
-		goto slow_irqon;
-
-	/*
-	 * local_irq_disable() doesn't prevent pagetable teardown, but does
-	 * prevent the pagetables from being freed on s390.
-	 *
-	 * So long as we atomically load page table pointers versus teardown,
-	 * we can follow the address down to the the page and take a ref on it.
-	 */
-	local_irq_disable();
-	pgdp = pgd_offset(mm, addr);
-	do {
-		pgd = *pgdp;
-		barrier();
-		next = pgd_addr_end(addr, end);
-		if (pgd_none(pgd))
-			goto slow;
-		if (!gup_pud_range(pgdp, pgd, addr, next, write, pages, &nr))
-			goto slow;
-	} while (pgdp++, addr = next, addr != end);
-	local_irq_enable();
-
-	VM_BUG_ON(nr != (end - start) >> PAGE_SHIFT);
-	return nr;
-
-	{
-		int ret;
-slow:
-		local_irq_enable();
-slow_irqon:
-		/* Try to get the remaining pages with get_user_pages */
-		start += nr << PAGE_SHIFT;
-		pages += nr;
-
-		down_read(&mm->mmap_sem);
-		ret = get_user_pages(current, mm, start,
-			(end - start) >> PAGE_SHIFT, write, 0, pages, NULL);
-		up_read(&mm->mmap_sem);
-
-		/* Have to be a bit careful with return values */
-		if (nr > 0) {
-			if (ret < 0)
-				ret = nr;
-			else
-				ret += nr;
-		}
-
-		return ret;
-	}
+	nr = __get_user_pages_fast(start, nr_pages, write, pages);
+	if (nr == nr_pages)
+		return nr;
+
+	/* Try to get the remaining pages with get_user_pages */
+	start += nr << PAGE_SHIFT;
+	pages += nr;
+	down_read(&mm->mmap_sem);
+	ret = get_user_pages(current, mm, start,
+			     nr_pages - nr, write, 0, pages, NULL);
+	up_read(&mm->mmap_sem);
+	/* Have to be a bit careful with return values */
+	if (nr > 0)
+		ret = (ret < 0) ? nr : ret + nr;
+	return ret;
 }
diff --git a/arch/s390/mm/mmap.c b/arch/s390/mm/mmap.c
index 4002329..6bcb045 100644
--- a/arch/s390/mm/mmap.c
+++ b/arch/s390/mm/mmap.c
@@ -101,18 +101,12 @@ void arch_pick_mmap_layout(struct mm_struct *mm)
 
 int s390_mmap_check(unsigned long addr, unsigned long len, unsigned long flags)
 {
-	int rc;
-
 	if (is_compat_task() || (TASK_SIZE >= (1UL << 53)))
 		return 0;
 	if (!(flags & MAP_FIXED))
 		addr = 0;
-	if ((addr + len) >= TASK_SIZE) {
-		rc = crst_table_upgrade(current->mm, 1UL << 53);
-		if (rc)
-			return rc;
-		update_mm(current->mm, current);
-	}
+	if ((addr + len) >= TASK_SIZE)
+		return crst_table_upgrade(current->mm, 1UL << 53);
 	return 0;
 }
 
@@ -132,7 +126,6 @@ s390_get_unmapped_area(struct file *filp, unsigned long addr,
 		rc = crst_table_upgrade(mm, 1UL << 53);
 		if (rc)
 			return (unsigned long) rc;
-		update_mm(mm, current);
 		area = arch_get_unmapped_area(filp, addr, len, pgoff, flags);
 	}
 	return area;
@@ -155,7 +148,6 @@ s390_get_unmapped_area_topdown(struct file *filp, const unsigned long addr,
 		rc = crst_table_upgrade(mm, 1UL << 53);
 		if (rc)
 			return (unsigned long) rc;
-		update_mm(mm, current);
 		area = arch_get_unmapped_area_topdown(filp, addr, len,
 						      pgoff, flags);
 	}
diff --git a/arch/s390/mm/pageattr.c b/arch/s390/mm/pageattr.c
index 9903974..8400f49 100644
--- a/arch/s390/mm/pageattr.c
+++ b/arch/s390/mm/pageattr.c
@@ -9,6 +9,7 @@
 #include <asm/pgtable.h>
 #include <asm/page.h>
 
+#if PAGE_DEFAULT_KEY
 static inline unsigned long sske_frame(unsigned long addr, unsigned char skey)
 {
 	asm volatile(".insn rrf,0xb22b0000,%[skey],%[addr],9,0"
@@ -16,7 +17,7 @@ static inline unsigned long sske_frame(unsigned long addr, unsigned char skey)
 	return addr;
 }
 
-void storage_key_init_range(unsigned long start, unsigned long end)
+void __storage_key_init_range(unsigned long start, unsigned long end)
 {
 	unsigned long boundary, size;
 
@@ -36,6 +37,7 @@ void storage_key_init_range(unsigned long start, unsigned long end)
 		start += PAGE_SIZE;
 	}
 }
+#endif
 
 static pte_t *walk_page_table(unsigned long addr)
 {
diff --git a/arch/s390/mm/pgtable.c b/arch/s390/mm/pgtable.c
index de8cbc3..0a2e5e0 100644
--- a/arch/s390/mm/pgtable.c
+++ b/arch/s390/mm/pgtable.c
@@ -48,12 +48,23 @@ void crst_table_free(struct mm_struct *mm, unsigned long *table)
 }
 
 #ifdef CONFIG_64BIT
+static void __crst_table_upgrade(void *arg)
+{
+	struct mm_struct *mm = arg;
+
+	if (current->active_mm == mm)
+		update_mm(mm, current);
+	__tlb_flush_local();
+}
+
 int crst_table_upgrade(struct mm_struct *mm, unsigned long limit)
 {
 	unsigned long *table, *pgd;
 	unsigned long entry;
+	int flush;
 
 	BUG_ON(limit > (1UL << 53));
+	flush = 0;
 repeat:
 	table = crst_table_alloc(mm);
 	if (!table)
@@ -79,12 +90,15 @@ repeat:
 		mm->pgd = (pgd_t *) table;
 		mm->task_size = mm->context.asce_limit;
 		table = NULL;
+		flush = 1;
 	}
 	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);
 	return 0;
 }
 
@@ -92,6 +106,8 @@ void crst_table_downgrade(struct mm_struct *mm, unsigned long limit)
 {
 	pgd_t *pgd;
 
+	if (current->active_mm == mm)
+		__tlb_flush_mm(mm);
 	while (mm->context.asce_limit > limit) {
 		pgd = mm->pgd;
 		switch (pgd_val(*pgd) & _REGION_ENTRY_TYPE_MASK) {
@@ -114,6 +130,8 @@ void crst_table_downgrade(struct mm_struct *mm, unsigned long limit)
 		mm->task_size = mm->context.asce_limit;
 		crst_table_free(mm, (unsigned long *) pgd);
 	}
+	if (current->active_mm == mm)
+		update_mm(mm, current);
 }
 #endif
 
@@ -1087,10 +1105,9 @@ again:
 			continue;
 		/* Allocate new page table with pgstes */
 		new = page_table_alloc_pgste(mm, addr);
-		if (!new) {
-			mm->context.has_pgste = 0;
-			continue;
-		}
+		if (!new)
+			return -ENOMEM;
+
 		spin_lock(&mm->page_table_lock);
 		if (likely((unsigned long *) pmd_deref(*pmd) == table)) {
 			/* Nuke pmd entry pointing to the "short" page table */
@@ -1128,13 +1145,15 @@ static unsigned long page_table_realloc_pud(struct mmu_gather *tlb,
 		if (pud_none_or_clear_bad(pud))
 			continue;
 		next = page_table_realloc_pmd(tlb, mm, pud, addr, next);
+		if (unlikely(IS_ERR_VALUE(next)))
+			return next;
 	} while (pud++, addr = next, addr != end);
 
 	return addr;
 }
 
-static void page_table_realloc(struct mmu_gather *tlb, struct mm_struct *mm,
-			       unsigned long addr, unsigned long end)
+static unsigned long page_table_realloc(struct mmu_gather *tlb, struct mm_struct *mm,
+					unsigned long addr, unsigned long end)
 {
 	unsigned long next;
 	pgd_t *pgd;
@@ -1145,7 +1164,11 @@ static void page_table_realloc(struct mmu_gather *tlb, struct mm_struct *mm,
 		if (pgd_none_or_clear_bad(pgd))
 			continue;
 		next = page_table_realloc_pud(tlb, mm, pgd, addr, next);
+		if (unlikely(IS_ERR_VALUE(next)))
+			return next;
 	} while (pgd++, addr = next, addr != end);
+
+	return 0;
 }
 
 /*
@@ -1157,10 +1180,6 @@ int s390_enable_sie(void)
 	struct mm_struct *mm = tsk->mm;
 	struct mmu_gather tlb;
 
-	/* Do we have switched amode? If no, we cannot do sie */
-	if (s390_user_mode == HOME_SPACE_MODE)
-		return -EINVAL;
-
 	/* Do we have pgstes? if yes, we are done */
 	if (mm_has_pgste(tsk->mm))
 		return 0;
@@ -1169,9 +1188,9 @@ int s390_enable_sie(void)
 	/* split thp mappings and disable thp for future mappings */
 	thp_split_mm(mm);
 	/* Reallocate the page tables with pgstes */
-	mm->context.has_pgste = 1;
 	tlb_gather_mmu(&tlb, mm, 0, TASK_SIZE);
-	page_table_realloc(&tlb, mm, 0, TASK_SIZE);
+	if (!page_table_realloc(&tlb, mm, 0, TASK_SIZE))
+		mm->context.has_pgste = 1;
 	tlb_finish_mmu(&tlb, 0, TASK_SIZE);
 	up_write(&mm->mmap_sem);
 	return mm->context.has_pgste ? 0 : -ENOMEM;