1 files changed, 181 insertions, 139 deletions

diff --git a/arch/s390/mm/vmem.c b/arch/s390/mm/vmem.c
index fd594d5fe14..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>
  */
@@ -10,15 +8,15 @@
 #include <linux/mm.h>
 #include <linux/module.h>
 #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>
 #include <asm/tlbflush.h>
+#include <asm/sections.h>
 
-unsigned long vmalloc_end;
-EXPORT_SYMBOL(vmalloc_end);
-
-static struct page *vmem_map;
 static DEFINE_MUTEX(vmem_mutex);
 
 struct memory_segment {
@@ -29,114 +27,119 @@ struct memory_segment {
 
 static LIST_HEAD(mem_segs);
 
-void __meminit memmap_init(unsigned long size, int nid, unsigned long zone,
-			   unsigned long start_pfn)
-{
-	struct page *start, *end;
-	struct page *map_start, *map_end;
-	int i;
-
-	start = pfn_to_page(start_pfn);
-	end = start + size;
-
-	for (i = 0; i < MEMORY_CHUNKS && memory_chunk[i].size > 0; i++) {
-		unsigned long cstart, cend;
-
-		cstart = PFN_DOWN(memory_chunk[i].addr);
-		cend = cstart + PFN_DOWN(memory_chunk[i].size);
-
-		map_start = mem_map + cstart;
-		map_end = mem_map + cend;
-
-		if (map_start < start)
-			map_start = start;
-		if (map_end > end)
-			map_end = end;
-
-		map_start -= ((unsigned long) map_start & (PAGE_SIZE - 1))
-			/ sizeof(struct page);
-		map_end += ((PFN_ALIGN((unsigned long) map_end)
-			     - (unsigned long) map_end)
-			    / sizeof(struct page));
-
-		if (map_start < map_end)
-			memmap_init_zone((unsigned long)(map_end - map_start),
-					 nid, zone, page_to_pfn(map_start),
-					 MEMMAP_EARLY);
-	}
-}
-
-static void __init_refok *vmem_alloc_pages(unsigned int order)
+static void __ref *vmem_alloc_pages(unsigned int order)
 {
 	if (slab_is_available())
 		return (void *)__get_free_pages(GFP_KERNEL, order);
 	return alloc_bootmem_pages((1 << order) * PAGE_SIZE);
 }
 
+static inline pud_t *vmem_pud_alloc(void)
+{
+	pud_t *pud = NULL;
+
+#ifdef CONFIG_64BIT
+	pud = vmem_alloc_pages(2);
+	if (!pud)
+		return NULL;
+	clear_table((unsigned long *) pud, _REGION3_ENTRY_EMPTY, PAGE_SIZE * 4);
+#endif
+	return pud;
+}
+
 static inline pmd_t *vmem_pmd_alloc(void)
 {
-	pmd_t *pmd;
-	int i;
+	pmd_t *pmd = NULL;
 
-	pmd = vmem_alloc_pages(PMD_ALLOC_ORDER);
+#ifdef CONFIG_64BIT
+	pmd = vmem_alloc_pages(2);
 	if (!pmd)
 		return NULL;
-	for (i = 0; i < PTRS_PER_PMD; i++)
-		pmd_clear_kernel(pmd + i);
+	clear_table((unsigned long *) pmd, _SEGMENT_ENTRY_EMPTY, PAGE_SIZE * 4);
+#endif
 	return pmd;
 }
 
-static inline pte_t *vmem_pte_alloc(void)
+static pte_t __ref *vmem_pte_alloc(unsigned long address)
 {
 	pte_t *pte;
-	pte_t empty_pte;
-	int i;
 
-	pte = vmem_alloc_pages(PTE_ALLOC_ORDER);
+	if (slab_is_available())
+		pte = (pte_t *) page_table_alloc(&init_mm, address);
+	else
+		pte = alloc_bootmem_align(PTRS_PER_PTE * sizeof(pte_t),
+					  PTRS_PER_PTE * sizeof(pte_t));
 	if (!pte)
 		return NULL;
-	pte_val(empty_pte) = _PAGE_TYPE_EMPTY;
-	for (i = 0; i < PTRS_PER_PTE; i++)
-		pte[i] = empty_pte;
+	clear_table((unsigned long *) pte, _PAGE_INVALID,
+		    PTRS_PER_PTE * sizeof(pte_t));
 	return pte;
 }
 
 /*
  * Add a physical memory range to the 1:1 mapping.
  */
-static int vmem_add_range(unsigned long start, unsigned long size)
+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(&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;
-			pgd_populate_kernel(&init_mm, pg_dir, pm_dir);
+			pud_populate(&init_mm, pu_dir, pm_dir);
 		}
-
-		pm_dir = pmd_offset(pg_dir, address);
+		pm_dir = pmd_offset(pu_dir, address);
+#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);
-		pte = pfn_pte(address >> PAGE_SHIFT, PAGE_KERNEL);
-		*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;
 }
 
@@ -146,60 +149,106 @@ 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))
+		if (pgd_none(*pg_dir)) {
+			address += PGDIR_SIZE;
 			continue;
-		pm_dir = pmd_offset(pg_dir, address);
-		if (pmd_none(*pm_dir))
+		}
+		pu_dir = pud_offset(pg_dir, address);
+		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)) {
+			address += PMD_SIZE;
+			continue;
+		}
+		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.
  */
-static int vmem_add_mem_map(unsigned long start, unsigned long size)
+int __meminit vmemmap_populate(unsigned long start, unsigned long end, int node)
 {
-	unsigned long address, start_addr, end_addr;
-	struct page *map_start, *map_end;
+	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;
 
-	map_start = vmem_map + PFN_DOWN(start);
-	map_end	= vmem_map + PFN_DOWN(start + size);
-
-	start_addr = (unsigned long) map_start & PAGE_MASK;
-	end_addr = PFN_ALIGN((unsigned long) map_end);
-
-	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(&init_mm, pg_dir, pu_dir);
+		}
+
+		pu_dir = pud_offset(pg_dir, address);
+		if (pud_none(*pu_dir)) {
 			pm_dir = vmem_pmd_alloc();
 			if (!pm_dir)
 				goto out;
-			pgd_populate_kernel(&init_mm, pg_dir, pm_dir);
+			pud_populate(&init_mm, pu_dir, pm_dir);
 		}
 
-		pm_dir = pmd_offset(pg_dir, address);
+		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);
@@ -209,24 +258,19 @@ static int vmem_add_mem_map(unsigned long start, unsigned long size)
 			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((void *)start, 0, end - start);
 	ret = 0;
 out:
-	flush_tlb_kernel_range(start_addr, end_addr);
 	return ret;
 }
 
-static int vmem_add_mem(unsigned long start, unsigned long size)
+void vmemmap_free(unsigned long start, unsigned long end)
 {
-	int ret;
-
-	ret = vmem_add_range(start, size);
-	if (ret)
-		return ret;
-	return vmem_add_mem_map(start, size);
 }
 
 /*
@@ -237,7 +281,7 @@ static int insert_memory_segment(struct memory_segment *seg)
 {
 	struct memory_segment *tmp;
 
-	if (PFN_DOWN(seg->start + seg->size) > max_pfn ||
+	if (seg->start + seg->size > VMEM_MAX_PHYS ||
 	    seg->start + seg->size < seg->start)
 		return -ERANGE;
 
@@ -266,7 +310,7 @@ static void __remove_shared_memory(struct memory_segment *seg)
 	vmem_remove_range(seg->start, seg->size);
 }
 
-int remove_shared_memory(unsigned long start, unsigned long size)
+int vmem_remove_mapping(unsigned long start, unsigned long size)
 {
 	struct memory_segment *seg;
 	int ret;
@@ -290,11 +334,9 @@ out:
 	return ret;
 }
 
-int add_shared_memory(unsigned long start, unsigned long size)
+int vmem_add_mapping(unsigned long start, unsigned long size)
 {
 	struct memory_segment *seg;
-	struct page *page;
-	unsigned long pfn, num_pfn, end_pfn;
 	int ret;
 
 	mutex_lock(&vmem_mutex);
@@ -309,24 +351,9 @@ int add_shared_memory(unsigned long start, unsigned long size)
 	if (ret)
 		goto out_free;
 
-	ret = vmem_add_mem(start, size);
+	ret = vmem_add_mem(start, size, 0);
 	if (ret)
 		goto out_remove;
-
-	pfn = PFN_DOWN(start);
-	num_pfn = PFN_DOWN(size);
-	end_pfn = pfn + num_pfn;
-
-	page = pfn_to_page(pfn);
-	memset(page, 0, num_pfn * sizeof(struct page));
-
-	for (; pfn < end_pfn; pfn++) {
-		page = pfn_to_page(pfn);
-		init_page_count(page);
-		reset_page_mapcount(page);
-		SetPageReserved(page);
-		INIT_LIST_HEAD(&page->lru);
-	}
 	goto out;
 
 out_remove:
@@ -340,39 +367,54 @@ out:
 
 /*
  * map whole physical memory to virtual memory (identity mapping)
+ * we reserve enough space in the vmalloc area for vmemmap to hotplug
+ * additional memory segments.
  */
 void __init vmem_map_init(void)
 {
-	unsigned long map_size;
-	int i;
-
-	map_size = ALIGN(max_low_pfn, MAX_ORDER_NR_PAGES) * sizeof(struct page);
-	vmalloc_end = PFN_ALIGN(VMALLOC_END_INIT) - PFN_ALIGN(map_size);
-	vmem_map = (struct page *) vmalloc_end;
-	NODE_DATA(0)->node_mem_map = vmem_map;
-
-	for (i = 0; i < MEMORY_CHUNKS && memory_chunk[i].size > 0; i++)
-		vmem_add_mem(memory_chunk[i].addr, memory_chunk[i].size);
+	unsigned long ro_start, ro_end;
+	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)
+			vmem_add_mem(start, end - start, 1);
+		else if (start >= ro_start) {
+			vmem_add_mem(start, ro_end - start, 1);
+			vmem_add_mem(ro_end, end - ro_end, 0);
+		} else if (end < ro_end) {
+			vmem_add_mem(start, ro_start - start, 0);
+			vmem_add_mem(ro_start, end - ro_start, 1);
+		} else {
+			vmem_add_mem(start, ro_start - start, 0);
+			vmem_add_mem(ro_start, ro_end - ro_start, 1);
+			vmem_add_mem(ro_end, end - ro_end, 0);
+		}
+	}
 }
 
 /*
- * 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 && memory_chunk[i].size > 0; 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);