1 files changed, 110 insertions, 68 deletions

diff --git a/arch/s390/mm/vmem.c b/arch/s390/mm/vmem.c
index 34c43f23b28..fe9012a49aa 100644
--- a/arch/s390/mm/vmem.c
+++ b/arch/s390/mm/vmem.c
@@ -1,6 +1,4 @@
 /*
- *  arch/s390/mm/vmem.c
- *
  *    Copyright IBM Corp. 2006
  *    Author(s): Heiko Carstens <heiko.carstens@de.ibm.com>
  */
@@ -12,6 +10,7 @@
 #include <linux/list.h>
 #include <linux/hugetlb.h>
 #include <linux/slab.h>
+#include <linux/memblock.h>
 #include <asm/pgalloc.h>
 #include <asm/pgtable.h>
 #include <asm/setup.h>
@@ -61,17 +60,18 @@ static inline pmd_t *vmem_pmd_alloc(void)
 	return pmd;
 }
 
-static pte_t __ref *vmem_pte_alloc(void)
+static pte_t __ref *vmem_pte_alloc(unsigned long address)
 {
 	pte_t *pte;
 
 	if (slab_is_available())
-		pte = (pte_t *) page_table_alloc(&init_mm);
+		pte = (pte_t *) page_table_alloc(&init_mm, address);
 	else
-		pte = alloc_bootmem(PTRS_PER_PTE * sizeof(pte_t));
+		pte = alloc_bootmem_align(PTRS_PER_PTE * sizeof(pte_t),
+					  PTRS_PER_PTE * sizeof(pte_t));
 	if (!pte)
 		return NULL;
-	clear_table((unsigned long *) pte, _PAGE_TYPE_EMPTY,
+	clear_table((unsigned long *) pte, _PAGE_INVALID,
 		    PTRS_PER_PTE * sizeof(pte_t));
 	return pte;
 }
@@ -81,57 +81,65 @@ static pte_t __ref *vmem_pte_alloc(void)
  */
 static int vmem_add_mem(unsigned long start, unsigned long size, int ro)
 {
-	unsigned long address;
+	unsigned long end = start + size;
+	unsigned long address = start;
 	pgd_t *pg_dir;
 	pud_t *pu_dir;
 	pmd_t *pm_dir;
 	pte_t *pt_dir;
-	pte_t  pte;
 	int ret = -ENOMEM;
 
-	for (address = start; address < start + size; address += PAGE_SIZE) {
+	while (address < end) {
 		pg_dir = pgd_offset_k(address);
 		if (pgd_none(*pg_dir)) {
 			pu_dir = vmem_pud_alloc();
 			if (!pu_dir)
 				goto out;
-			pgd_populate_kernel(&init_mm, pg_dir, pu_dir);
+			pgd_populate(&init_mm, pg_dir, pu_dir);
 		}
-
 		pu_dir = pud_offset(pg_dir, address);
+#if defined(CONFIG_64BIT) && !defined(CONFIG_DEBUG_PAGEALLOC)
+		if (MACHINE_HAS_EDAT2 && pud_none(*pu_dir) && address &&
+		    !(address & ~PUD_MASK) && (address + PUD_SIZE <= end)) {
+			pud_val(*pu_dir) = __pa(address) |
+				_REGION_ENTRY_TYPE_R3 | _REGION3_ENTRY_LARGE |
+				(ro ? _REGION_ENTRY_PROTECT : 0);
+			address += PUD_SIZE;
+			continue;
+		}
+#endif
 		if (pud_none(*pu_dir)) {
 			pm_dir = vmem_pmd_alloc();
 			if (!pm_dir)
 				goto out;
-			pud_populate_kernel(&init_mm, pu_dir, pm_dir);
+			pud_populate(&init_mm, pu_dir, pm_dir);
 		}
-
-		pte = mk_pte_phys(address, __pgprot(ro ? _PAGE_RO : 0));
 		pm_dir = pmd_offset(pu_dir, address);
-
-#ifdef __s390x__
-		if (MACHINE_HAS_HPAGE && !(address & ~HPAGE_MASK) &&
-		    (address + HPAGE_SIZE <= start + size) &&
-		    (address >= HPAGE_SIZE)) {
-			pte_val(pte) |= _SEGMENT_ENTRY_LARGE;
-			pmd_val(*pm_dir) = pte_val(pte);
-			address += HPAGE_SIZE - PAGE_SIZE;
+#if defined(CONFIG_64BIT) && !defined(CONFIG_DEBUG_PAGEALLOC)
+		if (MACHINE_HAS_EDAT1 && pmd_none(*pm_dir) && address &&
+		    !(address & ~PMD_MASK) && (address + PMD_SIZE <= end)) {
+			pmd_val(*pm_dir) = __pa(address) |
+				_SEGMENT_ENTRY | _SEGMENT_ENTRY_LARGE |
+				_SEGMENT_ENTRY_YOUNG |
+				(ro ? _SEGMENT_ENTRY_PROTECT : 0);
+			address += PMD_SIZE;
 			continue;
 		}
 #endif
 		if (pmd_none(*pm_dir)) {
-			pt_dir = vmem_pte_alloc();
+			pt_dir = vmem_pte_alloc(address);
 			if (!pt_dir)
 				goto out;
-			pmd_populate_kernel(&init_mm, pm_dir, pt_dir);
+			pmd_populate(&init_mm, pm_dir, pt_dir);
 		}
 
 		pt_dir = pte_offset_kernel(pm_dir, address);
-		*pt_dir = pte;
+		pte_val(*pt_dir) = __pa(address) |
+			pgprot_val(ro ? PAGE_KERNEL_RO : PAGE_KERNEL);
+		address += PAGE_SIZE;
 	}
 	ret = 0;
 out:
-	flush_tlb_kernel_range(start, start + size);
 	return ret;
 }
 
@@ -141,58 +149,67 @@ out:
  */
 static void vmem_remove_range(unsigned long start, unsigned long size)
 {
-	unsigned long address;
+	unsigned long end = start + size;
+	unsigned long address = start;
 	pgd_t *pg_dir;
 	pud_t *pu_dir;
 	pmd_t *pm_dir;
 	pte_t *pt_dir;
 	pte_t  pte;
 
-	pte_val(pte) = _PAGE_TYPE_EMPTY;
-	for (address = start; address < start + size; address += PAGE_SIZE) {
+	pte_val(pte) = _PAGE_INVALID;
+	while (address < end) {
 		pg_dir = pgd_offset_k(address);
+		if (pgd_none(*pg_dir)) {
+			address += PGDIR_SIZE;
+			continue;
+		}
 		pu_dir = pud_offset(pg_dir, address);
-		if (pud_none(*pu_dir))
+		if (pud_none(*pu_dir)) {
+			address += PUD_SIZE;
+			continue;
+		}
+		if (pud_large(*pu_dir)) {
+			pud_clear(pu_dir);
+			address += PUD_SIZE;
 			continue;
+		}
 		pm_dir = pmd_offset(pu_dir, address);
-		if (pmd_none(*pm_dir))
+		if (pmd_none(*pm_dir)) {
+			address += PMD_SIZE;
 			continue;
-
-		if (pmd_huge(*pm_dir)) {
-			pmd_clear_kernel(pm_dir);
-			address += HPAGE_SIZE - PAGE_SIZE;
+		}
+		if (pmd_large(*pm_dir)) {
+			pmd_clear(pm_dir);
+			address += PMD_SIZE;
 			continue;
 		}
-
 		pt_dir = pte_offset_kernel(pm_dir, address);
 		*pt_dir = pte;
+		address += PAGE_SIZE;
 	}
-	flush_tlb_kernel_range(start, start + size);
+	flush_tlb_kernel_range(start, end);
 }
 
 /*
  * Add a backed mem_map array to the virtual mem_map array.
  */
-int __meminit vmemmap_populate(struct page *start, unsigned long nr, int node)
+int __meminit vmemmap_populate(unsigned long start, unsigned long end, int node)
 {
-	unsigned long address, start_addr, end_addr;
+	unsigned long address = start;
 	pgd_t *pg_dir;
 	pud_t *pu_dir;
 	pmd_t *pm_dir;
 	pte_t *pt_dir;
-	pte_t  pte;
 	int ret = -ENOMEM;
 
-	start_addr = (unsigned long) start;
-	end_addr = (unsigned long) (start + nr);
-
-	for (address = start_addr; address < end_addr; address += PAGE_SIZE) {
+	for (address = start; address < end;) {
 		pg_dir = pgd_offset_k(address);
 		if (pgd_none(*pg_dir)) {
 			pu_dir = vmem_pud_alloc();
 			if (!pu_dir)
 				goto out;
-			pgd_populate_kernel(&init_mm, pg_dir, pu_dir);
+			pgd_populate(&init_mm, pg_dir, pu_dir);
 		}
 
 		pu_dir = pud_offset(pg_dir, address);
@@ -200,15 +217,38 @@ int __meminit vmemmap_populate(struct page *start, unsigned long nr, int node)
 			pm_dir = vmem_pmd_alloc();
 			if (!pm_dir)
 				goto out;
-			pud_populate_kernel(&init_mm, pu_dir, pm_dir);
+			pud_populate(&init_mm, pu_dir, pm_dir);
 		}
 
 		pm_dir = pmd_offset(pu_dir, address);
 		if (pmd_none(*pm_dir)) {
-			pt_dir = vmem_pte_alloc();
+#ifdef CONFIG_64BIT
+			/* Use 1MB frames for vmemmap if available. We always
+			 * use large frames even if they are only partially
+			 * used.
+			 * Otherwise we would have also page tables since
+			 * vmemmap_populate gets called for each section
+			 * separately. */
+			if (MACHINE_HAS_EDAT1) {
+				void *new_page;
+
+				new_page = vmemmap_alloc_block(PMD_SIZE, node);
+				if (!new_page)
+					goto out;
+				pmd_val(*pm_dir) = __pa(new_page) |
+					_SEGMENT_ENTRY | _SEGMENT_ENTRY_LARGE |
+					_SEGMENT_ENTRY_CO;
+				address = (address + PMD_SIZE) & PMD_MASK;
+				continue;
+			}
+#endif
+			pt_dir = vmem_pte_alloc(address);
 			if (!pt_dir)
 				goto out;
-			pmd_populate_kernel(&init_mm, pm_dir, pt_dir);
+			pmd_populate(&init_mm, pm_dir, pt_dir);
+		} else if (pmd_large(*pm_dir)) {
+			address = (address + PMD_SIZE) & PMD_MASK;
+			continue;
 		}
 
 		pt_dir = pte_offset_kernel(pm_dir, address);
@@ -218,17 +258,21 @@ int __meminit vmemmap_populate(struct page *start, unsigned long nr, int node)
 			new_page =__pa(vmem_alloc_pages(0));
 			if (!new_page)
 				goto out;
-			pte = pfn_pte(new_page >> PAGE_SHIFT, PAGE_KERNEL);
-			*pt_dir = pte;
+			pte_val(*pt_dir) =
+				__pa(new_page) | pgprot_val(PAGE_KERNEL);
 		}
+		address += PAGE_SIZE;
 	}
-	memset(start, 0, nr * sizeof(struct page));
+	memset((void *)start, 0, end - start);
 	ret = 0;
 out:
-	flush_tlb_kernel_range(start_addr, end_addr);
 	return ret;
 }
 
+void vmemmap_free(unsigned long start, unsigned long end)
+{
+}
+
 /*
  * Add memory segment to the segment list if it doesn't overlap with
  * an already present segment.
@@ -329,14 +373,14 @@ out:
 void __init vmem_map_init(void)
 {
 	unsigned long ro_start, ro_end;
-	unsigned long start, end;
-	int i;
-
-	ro_start = ((unsigned long)&_stext) & PAGE_MASK;
-	ro_end = PFN_ALIGN((unsigned long)&_eshared);
-	for (i = 0; i < MEMORY_CHUNKS && memory_chunk[i].size > 0; i++) {
-		start = memory_chunk[i].addr;
-		end = memory_chunk[i].addr + memory_chunk[i].size;
+	struct memblock_region *reg;
+	phys_addr_t start, end;
+
+	ro_start = PFN_ALIGN((unsigned long)&_stext);
+	ro_end = (unsigned long)&_eshared & PAGE_MASK;
+	for_each_memblock(memory, reg) {
+		start = reg->base;
+		end = reg->base + reg->size - 1;
 		if (start >= ro_end || end <= ro_start)
 			vmem_add_mem(start, end - start, 0);
 		else if (start >= ro_start && end <= ro_end)
@@ -356,23 +400,21 @@ void __init vmem_map_init(void)
 }
 
 /*
- * Convert memory chunk array to a memory segment list so there is a single
- * list that contains both r/w memory and shared memory segments.
+ * Convert memblock.memory  to a memory segment list so there is a single
+ * list that contains all memory segments.
  */
 static int __init vmem_convert_memory_chunk(void)
 {
+	struct memblock_region *reg;
 	struct memory_segment *seg;
-	int i;
 
 	mutex_lock(&vmem_mutex);
-	for (i = 0; i < MEMORY_CHUNKS; i++) {
-		if (!memory_chunk[i].size)
-			continue;
+	for_each_memblock(memory, reg) {
 		seg = kzalloc(sizeof(*seg), GFP_KERNEL);
 		if (!seg)
 			panic("Out of memory...\n");
-		seg->start = memory_chunk[i].addr;
-		seg->size = memory_chunk[i].size;
+		seg->start = reg->base;
+		seg->size = reg->size;
 		insert_memory_segment(seg);
 	}
 	mutex_unlock(&vmem_mutex);