s390/mm: Convert bootmem to memblock

The original bootmem allocator is getting replaced by memblock. To
cover the needs of the s390 kdump implementation the physical memory
list is used.
With this patch the bootmem allocator and its bitmaps are completely
removed from s390.

Signed-off-by: Philipp Hachtmann <phacht@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>

5 files changed, 208 insertions, 337 deletions

diff --git a/arch/s390/kernel/crash_dump.c b/arch/s390/kernel/crash_dump.c
index d7658c4b2ed5..a3b9150e6802 100644
--- a/arch/s390/kernel/crash_dump.c
+++ b/arch/s390/kernel/crash_dump.c
@@ -13,6 +13,7 @@
 #include <linux/slab.h>
 #include <linux/bootmem.h>
 #include <linux/elf.h>
+#include <linux/memblock.h>
 #include <asm/os_info.h>
 #include <asm/elf.h>
 #include <asm/ipl.h>
@@ -22,6 +23,24 @@
 #define PTR_SUB(x, y) (((char *) (x)) - ((unsigned long) (y)))
 #define PTR_DIFF(x, y) ((unsigned long)(((char *) (x)) - ((unsigned long) (y))))
 
+static struct memblock_region oldmem_region;
+
+static struct memblock_type oldmem_type = {
+	.cnt = 1,
+	.max = 1,
+	.total_size = 0,
+	.regions = &oldmem_region,
+};
+
+#define for_each_dump_mem_range(i, nid, p_start, p_end, p_nid)		\
+	for (i = 0, __next_mem_range(&i, nid, &memblock.physmem,	\
+				     &oldmem_type, p_start,		\
+				     p_end, p_nid);			\
+	     i != (u64)ULLONG_MAX;					\
+	     __next_mem_range(&i, nid, &memblock.physmem,		\
+			      &oldmem_type,				\
+			      p_start, p_end, p_nid))
+
 struct dump_save_areas dump_save_areas;
 
 /*
@@ -264,19 +283,6 @@ static void *kzalloc_panic(int len)
 }
 
 /*
- * Get memory layout and create hole for oldmem
- */
-static struct mem_chunk *get_memory_layout(void)
-{
-	struct mem_chunk *chunk_array;
-
-	chunk_array = kzalloc_panic(MEMORY_CHUNKS * sizeof(struct mem_chunk));
-	detect_memory_layout(chunk_array, 0);
-	create_mem_hole(chunk_array, OLDMEM_BASE, OLDMEM_SIZE);
-	return chunk_array;
-}
-
-/*
  * Initialize ELF note
  */
 static void *nt_init(void *buf, Elf64_Word type, void *desc, int d_len,
@@ -490,52 +496,33 @@ static int get_cpu_cnt(void)
  */
 static int get_mem_chunk_cnt(void)
 {
-	struct mem_chunk *chunk_array, *mem_chunk;
-	int i, cnt = 0;
+	int cnt = 0;
+	u64 idx;
 
-	chunk_array = get_memory_layout();
-	for (i = 0; i < MEMORY_CHUNKS; i++) {
-		mem_chunk = &chunk_array[i];
-		if (chunk_array[i].type != CHUNK_READ_WRITE &&
-		    chunk_array[i].type != CHUNK_READ_ONLY)
-			continue;
-		if (mem_chunk->size == 0)
-			continue;
+	for_each_dump_mem_range(idx, NUMA_NO_NODE, NULL, NULL, NULL)
 		cnt++;
-	}
-	kfree(chunk_array);
 	return cnt;
 }
 
 /*
  * Initialize ELF loads (new kernel)
  */
-static int loads_init(Elf64_Phdr *phdr, u64 loads_offset)
+static void loads_init(Elf64_Phdr *phdr, u64 loads_offset)
 {
-	struct mem_chunk *chunk_array, *mem_chunk;
-	int i;
+	phys_addr_t start, end;
+	u64 idx;
 
-	chunk_array = get_memory_layout();
-	for (i = 0; i < MEMORY_CHUNKS; i++) {
-		mem_chunk = &chunk_array[i];
-		if (mem_chunk->size == 0)
-			continue;
-		if (chunk_array[i].type != CHUNK_READ_WRITE &&
-		    chunk_array[i].type != CHUNK_READ_ONLY)
-			continue;
-		else
-			phdr->p_filesz = mem_chunk->size;
+	for_each_dump_mem_range(idx, NUMA_NO_NODE, &start, &end, NULL) {
+		phdr->p_filesz = end - start;
 		phdr->p_type = PT_LOAD;
-		phdr->p_offset = mem_chunk->addr;
-		phdr->p_vaddr = mem_chunk->addr;
-		phdr->p_paddr = mem_chunk->addr;
-		phdr->p_memsz = mem_chunk->size;
+		phdr->p_offset = start;
+		phdr->p_vaddr = start;
+		phdr->p_paddr = start;
+		phdr->p_memsz = end - start;
 		phdr->p_flags = PF_R | PF_W | PF_X;
 		phdr->p_align = PAGE_SIZE;
 		phdr++;
 	}
-	kfree(chunk_array);
-	return i;
 }
 
 /*
@@ -584,6 +571,14 @@ int elfcorehdr_alloc(unsigned long long *addr, unsigned long long *size)
 	/* If we cannot get HSA size for zfcpdump return error */
 	if (ipl_info.type == IPL_TYPE_FCP_DUMP && !sclp_get_hsa_size())
 		return -ENODEV;
+
+	/* For kdump, exclude previous crashkernel memory */
+	if (OLDMEM_BASE) {
+		oldmem_region.base = OLDMEM_BASE;
+		oldmem_region.size = OLDMEM_SIZE;
+		oldmem_type.total_size = OLDMEM_SIZE;
+	}
+
 	mem_chunk_cnt = get_mem_chunk_cnt();
 
 	alloc_size = 0x1000 + get_cpu_cnt() * 0x300 +
diff --git a/arch/s390/kernel/early.c b/arch/s390/kernel/early.c
index a734f3585ceb..0dff972a169c 100644
--- a/arch/s390/kernel/early.c
+++ b/arch/s390/kernel/early.c
@@ -258,13 +258,19 @@ static __init void setup_topology(void)
 static void early_pgm_check_handler(void)
 {
 	const struct exception_table_entry *fixup;
+	unsigned long cr0, cr0_new;
 	unsigned long addr;
 
 	addr = S390_lowcore.program_old_psw.addr;
 	fixup = search_exception_tables(addr & PSW_ADDR_INSN);
 	if (!fixup)
 		disabled_wait(0);
+	/* Disable low address protection before storing into lowcore. */
+	__ctl_store(cr0, 0, 0);
+	cr0_new = cr0 & ~(1UL << 28);
+	__ctl_load(cr0_new, 0, 0);
 	S390_lowcore.program_old_psw.addr = extable_fixup(fixup)|PSW_ADDR_AMODE;
+	__ctl_load(cr0, 0, 0);
 }
 
 static noinline __init void setup_lowcore_early(void)
diff --git a/arch/s390/kernel/head31.S b/arch/s390/kernel/head31.S
index 9a99856df1c9..6dbe80983a24 100644
--- a/arch/s390/kernel/head31.S
+++ b/arch/s390/kernel/head31.S
@@ -59,7 +59,6 @@ ENTRY(startup_continue)
 	.long	0			# cr13: home space segment table
 	.long	0xc0000000		# cr14: machine check handling off
 	.long	0			# cr15: linkage stack operations
-.Lmchunk:.long	memory_chunk
 .Lbss_bgn:  .long __bss_start
 .Lbss_end:  .long _end
 .Lparmaddr: .long PARMAREA
diff --git a/arch/s390/kernel/setup.c b/arch/s390/kernel/setup.c
index 88d1ca81e2dd..1f5536c2fd02 100644
--- a/arch/s390/kernel/setup.c
+++ b/arch/s390/kernel/setup.c
@@ -78,10 +78,9 @@ EXPORT_SYMBOL(console_irq);
 unsigned long elf_hwcap = 0;
 char elf_platform[ELF_PLATFORM_SIZE];
 
-struct mem_chunk __initdata memory_chunk[MEMORY_CHUNKS];
-
 int __initdata memory_end_set;
 unsigned long __initdata memory_end;
+unsigned long __initdata max_physmem_end;
 
 unsigned long VMALLOC_START;
 EXPORT_SYMBOL(VMALLOC_START);
@@ -273,6 +272,7 @@ EXPORT_SYMBOL_GPL(pm_power_off);
 static int __init early_parse_mem(char *p)
 {
 	memory_end = memparse(p, &p);
+	memory_end &= PAGE_MASK;
 	memory_end_set = 1;
 	return 0;
 }
@@ -401,7 +401,8 @@ static struct resource __initdata *standard_resources[] = {
 static void __init setup_resources(void)
 {
 	struct resource *res, *std_res, *sub_res;
-	int i, j;
+	struct memblock_region *reg;
+	int j;
 
 	code_resource.start = (unsigned long) &_text;
 	code_resource.end = (unsigned long) &_etext - 1;
@@ -410,24 +411,13 @@ static void __init setup_resources(void)
 	bss_resource.start = (unsigned long) &__bss_start;
 	bss_resource.end = (unsigned long) &__bss_stop - 1;
 
-	for (i = 0; i < MEMORY_CHUNKS; i++) {
-		if (!memory_chunk[i].size)
-			continue;
+	for_each_memblock(memory, reg) {
 		res = alloc_bootmem_low(sizeof(*res));
 		res->flags = IORESOURCE_BUSY | IORESOURCE_MEM;
-		switch (memory_chunk[i].type) {
-		case CHUNK_READ_WRITE:
-			res->name = "System RAM";
-			break;
-		case CHUNK_READ_ONLY:
-			res->name = "System ROM";
-			res->flags |= IORESOURCE_READONLY;
-			break;
-		default:
-			res->name = "reserved";
-		}
-		res->start = memory_chunk[i].addr;
-		res->end = res->start + memory_chunk[i].size - 1;
+
+		res->name = "System RAM";
+		res->start = reg->base;
+		res->end = reg->base + reg->size - 1;
 		request_resource(&iomem_resource, res);
 
 		for (j = 0; j < ARRAY_SIZE(standard_resources); j++) {
@@ -451,48 +441,11 @@ static void __init setup_resources(void)
 static void __init setup_memory_end(void)
 {
 	unsigned long vmax, vmalloc_size, tmp;
-	unsigned long real_memory_size = 0;
-	int i;
-
-
-#ifdef CONFIG_ZFCPDUMP
-	if (ipl_info.type == IPL_TYPE_FCP_DUMP &&
-	    !OLDMEM_BASE && sclp_get_hsa_size()) {
-		memory_end = sclp_get_hsa_size();
-		memory_end_set = 1;
-	}
-#endif
-	memory_end &= PAGE_MASK;
-
-	/*
-	 * Make sure all chunks are MAX_ORDER aligned so we don't need the
-	 * extra checks that HOLES_IN_ZONE would require.
-	 */
-	for (i = 0; i < MEMORY_CHUNKS; i++) {
-		unsigned long start, end;
-		struct mem_chunk *chunk;
-		unsigned long align;
-
-		chunk = &memory_chunk[i];
-		if (!chunk->size)
-			continue;
-		align = 1UL << (MAX_ORDER + PAGE_SHIFT - 1);
-		start = (chunk->addr + align - 1) & ~(align - 1);
-		end = (chunk->addr + chunk->size) & ~(align - 1);
-		if (start >= end)
-			memset(chunk, 0, sizeof(*chunk));
-		else {
-			chunk->addr = start;
-			chunk->size = end - start;
-		}
-		real_memory_size = max(real_memory_size,
-				       chunk->addr + chunk->size);
-	}
 
 	/* Choose kernel address space layout: 2, 3, or 4 levels. */
 #ifdef CONFIG_64BIT
 	vmalloc_size = VMALLOC_END ?: (128UL << 30) - MODULES_LEN;
-	tmp = (memory_end ?: real_memory_size) / PAGE_SIZE;
+	tmp = (memory_end ?: max_physmem_end) / PAGE_SIZE;
 	tmp = tmp * (sizeof(struct page) + PAGE_SIZE) + vmalloc_size;
 	if (tmp <= (1UL << 42))
 		vmax = 1UL << 42;	/* 3-level kernel page table */
@@ -520,21 +473,11 @@ static void __init setup_memory_end(void)
 	vmemmap = (struct page *) tmp;
 
 	/* Take care that memory_end is set and <= vmemmap */
-	memory_end = min(memory_end ?: real_memory_size, tmp);
-
-	/* Fixup memory chunk array to fit into 0..memory_end */
-	for (i = 0; i < MEMORY_CHUNKS; i++) {
-		struct mem_chunk *chunk = &memory_chunk[i];
+	memory_end = min(memory_end ?: max_physmem_end, tmp);
+	max_pfn = max_low_pfn = PFN_DOWN(memory_end);
+	memblock_remove(memory_end, ULONG_MAX);
 
-		if (!chunk->size)
-			continue;
-		if (chunk->addr >= memory_end) {
-			memset(chunk, 0, sizeof(*chunk));
-			continue;
-		}
-		if (chunk->addr + chunk->size > memory_end)
-			chunk->size = memory_end - chunk->addr;
-	}
+	pr_notice("Max memory size: %luMB\n", memory_end >> 20);
 }
 
 static void __init setup_vmcoreinfo(void)
@@ -545,89 +488,6 @@ static void __init setup_vmcoreinfo(void)
 #ifdef CONFIG_CRASH_DUMP
 
 /*
- * Find suitable location for crashkernel memory
- */
-static unsigned long __init find_crash_base(unsigned long crash_size,
-					    char **msg)
-{
-	unsigned long crash_base;
-	struct mem_chunk *chunk;
-	int i;
-
-	if (memory_chunk[0].size < crash_size) {
-		*msg = "first memory chunk must be at least crashkernel size";
-		return 0;
-	}
-	if (OLDMEM_BASE && crash_size == OLDMEM_SIZE)
-		return OLDMEM_BASE;
-
-	for (i = MEMORY_CHUNKS - 1; i >= 0; i--) {
-		chunk = &memory_chunk[i];
-		if (chunk->size == 0)
-			continue;
-		if (chunk->type != CHUNK_READ_WRITE)
-			continue;
-		if (chunk->size < crash_size)
-			continue;
-		crash_base = (chunk->addr + chunk->size) - crash_size;
-		if (crash_base < crash_size)
-			continue;
-		if (crash_base < sclp_get_hsa_size())
-			continue;
-		if (crash_base < (unsigned long) INITRD_START + INITRD_SIZE)
-			continue;
-		return crash_base;
-	}
-	*msg = "no suitable area found";
-	return 0;
-}
-
-/*
- * Check if crash_base and crash_size is valid
- */
-static int __init verify_crash_base(unsigned long crash_base,
-				    unsigned long crash_size,
-				    char **msg)
-{
-	struct mem_chunk *chunk;
-	int i;
-
-	/*
-	 * Because we do the swap to zero, we must have at least 'crash_size'
-	 * bytes free space before crash_base
-	 */
-	if (crash_size > crash_base) {
-		*msg = "crashkernel offset must be greater than size";
-		return -EINVAL;
-	}
-
-	/* First memory chunk must be at least crash_size */
-	if (memory_chunk[0].size < crash_size) {
-		*msg = "first memory chunk must be at least crashkernel size";
-		return -EINVAL;
-	}
-	/* Check if we fit into the respective memory chunk */
-	for (i = 0; i < MEMORY_CHUNKS; i++) {
-		chunk = &memory_chunk[i];
-		if (chunk->size == 0)
-			continue;
-		if (crash_base < chunk->addr)
-			continue;
-		if (crash_base >= chunk->addr + chunk->size)
-			continue;
-		/* we have found the memory chunk */
-		if (crash_base + crash_size > chunk->addr + chunk->size) {
-			*msg = "selected memory chunk is too small for "
-				"crashkernel memory";
-			return -EINVAL;
-		}
-		return 0;
-	}
-	*msg = "invalid memory range specified";
-	return -EINVAL;
-}
-
-/*
  * When kdump is enabled, we have to ensure that no memory from
  * the area [0 - crashkernel memory size] and
  * [crashk_res.start - crashk_res.end] is set offline.
@@ -653,23 +513,44 @@ static struct notifier_block kdump_mem_nb = {
 #endif
 
 /*
+ * Make sure that the area behind memory_end is protected
+ */
+static void reserve_memory_end(void)
+{
+#ifdef CONFIG_ZFCPDUMP
+	if (ipl_info.type == IPL_TYPE_FCP_DUMP &&
+	    !OLDMEM_BASE && sclp_get_hsa_size()) {
+		memory_end = sclp_get_hsa_size();
+		memory_end &= PAGE_MASK;
+		memory_end_set = 1;
+	}
+#endif
+	if (!memory_end_set)
+		return;
+	memblock_reserve(memory_end, ULONG_MAX);
+}
+
+/*
  * Make sure that oldmem, where the dump is stored, is protected
  */
 static void reserve_oldmem(void)
 {
 #ifdef CONFIG_CRASH_DUMP
-	unsigned long real_size = 0;
-	int i;
-
-	if (!OLDMEM_BASE)
-		return;
-	for (i = 0; i < MEMORY_CHUNKS; i++) {
-		struct mem_chunk *chunk = &memory_chunk[i];
+	if (OLDMEM_BASE)
+		/* Forget all memory above the running kdump system */
+		memblock_reserve(OLDMEM_SIZE, (phys_addr_t)ULONG_MAX);
+#endif
+}
 
-		real_size = max(real_size, chunk->addr + chunk->size);
-	}
-	create_mem_hole(memory_chunk, OLDMEM_BASE, OLDMEM_SIZE);
-	create_mem_hole(memory_chunk, OLDMEM_SIZE, real_size - OLDMEM_SIZE);
+/*
+ * Make sure that oldmem, where the dump is stored, is protected
+ */
+static void remove_oldmem(void)
+{
+#ifdef CONFIG_CRASH_DUMP
+	if (OLDMEM_BASE)
+		/* Forget all memory above the running kdump system */
+		memblock_remove(OLDMEM_SIZE, (phys_addr_t)ULONG_MAX);
 #endif
 }
 
@@ -680,167 +561,132 @@ static void __init reserve_crashkernel(void)
 {
 #ifdef CONFIG_CRASH_DUMP
 	unsigned long long crash_base, crash_size;
-	char *msg = NULL;
+	phys_addr_t low, high;
 	int rc;
 
 	rc = parse_crashkernel(boot_command_line, memory_end, &crash_size,
 			       &crash_base);
-	if (rc || crash_size == 0)
-		return;
+
 	crash_base = ALIGN(crash_base, KEXEC_CRASH_MEM_ALIGN);
 	crash_size = ALIGN(crash_size, KEXEC_CRASH_MEM_ALIGN);
-	if (register_memory_notifier(&kdump_mem_nb))
+	if (rc || crash_size == 0)
 		return;
-	if (!crash_base)
-		crash_base = find_crash_base(crash_size, &msg);
-	if (!crash_base) {
-		pr_info("crashkernel reservation failed: %s\n", msg);
-		unregister_memory_notifier(&kdump_mem_nb);
+
+	if (memblock.memory.regions[0].size < crash_size) {
+		pr_info("crashkernel reservation failed: %s\n",
+			"first memory chunk must be at least crashkernel size");
 		return;
 	}
-	if (verify_crash_base(crash_base, crash_size, &msg)) {
-		pr_info("crashkernel reservation failed: %s\n", msg);
-		unregister_memory_notifier(&kdump_mem_nb);
+
+	low = crash_base ?: OLDMEM_BASE;
+	high = low + crash_size;
+	if (low >= OLDMEM_BASE && high <= OLDMEM_BASE + OLDMEM_SIZE) {
+		/* The crashkernel fits into OLDMEM, reuse OLDMEM */
+		crash_base = low;
+	} else {
+		/* Find suitable area in free memory */
+		low = max_t(unsigned long, crash_size, sclp_get_hsa_size());
+		high = crash_base ? crash_base + crash_size : ULONG_MAX;
+
+		if (crash_base && crash_base < low) {
+			pr_info("crashkernel reservation failed: %s\n",
+				"crash_base too low");
+			return;
+		}
+		low = crash_base ?: low;
+		crash_base = memblock_find_in_range(low, high, crash_size,
+						    KEXEC_CRASH_MEM_ALIGN);
+	}
+
+	if (!crash_base) {
+		pr_info("crashkernel reservation failed: %s\n",
+			"no suitable area found");
 		return;
 	}
+
+	if (register_memory_notifier(&kdump_mem_nb))
+		return;
+
 	if (!OLDMEM_BASE && MACHINE_IS_VM)
 		diag10_range(PFN_DOWN(crash_base), PFN_DOWN(crash_size));
 	crashk_res.start = crash_base;
 	crashk_res.end = crash_base + crash_size - 1;
 	insert_resource(&iomem_resource, &crashk_res);
-	create_mem_hole(memory_chunk, crash_base, crash_size);
+	memblock_remove(crash_base, crash_size);
 	pr_info("Reserving %lluMB of memory at %lluMB "
 		"for crashkernel (System RAM: %luMB)\n",
-		crash_size >> 20, crash_base >> 20, memory_end >> 20);
+		crash_size >> 20, crash_base >> 20,
+		(unsigned long)memblock.memory.total_size >> 20);
 	os_info_crashkernel_add(crash_base, crash_size);
 #endif
 }
 
-static void __init setup_memory(void)
+/*
+ * Reserve the initrd from being used by memblock
+ */
+static void __init reserve_initrd(void)
 {
-        unsigned long bootmap_size;
-	unsigned long start_pfn, end_pfn;
-	int i;
+#ifdef CONFIG_BLK_DEV_INITRD
+	initrd_start = INITRD_START;
+	initrd_end = initrd_start + INITRD_SIZE;
+	memblock_reserve(INITRD_START, INITRD_SIZE);
+#endif
+}
 
-	/*
-	 * partially used pages are not usable - thus
-	 * we are rounding upwards:
-	 */
+/*
+ * Check for initrd being in usable memory
+ */
+static void __init check_initrd(void)
+{
+#ifdef CONFIG_BLK_DEV_INITRD
+	if (INITRD_START && INITRD_SIZE &&
+	    !memblock_is_region_memory(INITRD_START, INITRD_SIZE)) {
+		pr_err("initrd does not fit memory.\n");
+		memblock_free(INITRD_START, INITRD_SIZE);
+		initrd_start = initrd_end = 0;
+	}
+#endif
+}
+
+/*
+ * Reserve all kernel text
+ */
+static void __init reserve_kernel(void)
+{
+	unsigned long start_pfn;
 	start_pfn = PFN_UP(__pa(&_end));
-	end_pfn = max_pfn = PFN_DOWN(memory_end);
 
-#ifdef CONFIG_BLK_DEV_INITRD
 	/*
-	 * Move the initrd in case the bitmap of the bootmem allocater
-	 * would overwrite it.
+	 * Reserve memory used for lowcore/command line/kernel image.
 	 */
+	memblock_reserve(0, (unsigned long)_ehead);
+	memblock_reserve((unsigned long)_stext, PFN_PHYS(start_pfn)
+			 - (unsigned long)_stext);
+}
 
-	if (INITRD_START && INITRD_SIZE) {
-		unsigned long bmap_size;
-		unsigned long start;
-
-		bmap_size = bootmem_bootmap_pages(end_pfn - start_pfn + 1);
-		bmap_size = PFN_PHYS(bmap_size);
-
-		if (PFN_PHYS(start_pfn) + bmap_size > INITRD_START) {
-			start = PFN_PHYS(start_pfn) + bmap_size + PAGE_SIZE;
-
+static void __init reserve_elfcorehdr(void)
+{
 #ifdef CONFIG_CRASH_DUMP
-			if (OLDMEM_BASE) {
-				/* Move initrd behind kdump oldmem */
-				if (start + INITRD_SIZE > OLDMEM_BASE &&
-				    start < OLDMEM_BASE + OLDMEM_SIZE)
-					start = OLDMEM_BASE + OLDMEM_SIZE;
-			}
-#endif
-			if (start + INITRD_SIZE > memory_end) {
-				pr_err("initrd extends beyond end of "
-				       "memory (0x%08lx > 0x%08lx) "
-				       "disabling initrd\n",
-				       start + INITRD_SIZE, memory_end);
-				INITRD_START = INITRD_SIZE = 0;
-			} else {
-				pr_info("Moving initrd (0x%08lx -> "
-					"0x%08lx, size: %ld)\n",
-					INITRD_START, start, INITRD_SIZE);
-				memmove((void *) start, (void *) INITRD_START,
-					INITRD_SIZE);
-				INITRD_START = start;
-			}
-		}
-	}
+	if (is_kdump_kernel())
+		memblock_reserve(elfcorehdr_addr - OLDMEM_BASE,
+				 PAGE_ALIGN(elfcorehdr_size));
 #endif
+}
 
-	/*
-	 * Initialize the boot-time allocator
-	 */
-	bootmap_size = init_bootmem(start_pfn, end_pfn);
+static void __init setup_memory(void)
+{
+	struct memblock_region *reg;
 
 	/*
-	 * Register RAM areas with the bootmem allocator.
+	 * Init storage key for present memory
 	 */
-
-	for (i = 0; i < MEMORY_CHUNKS; i++) {
-		unsigned long start_chunk, end_chunk, pfn;
-
-		if (!memory_chunk[i].size)
-			continue;
-		start_chunk = PFN_DOWN(memory_chunk[i].addr);
-		end_chunk = start_chunk + PFN_DOWN(memory_chunk[i].size);
-		end_chunk = min(end_chunk, end_pfn);
-		if (start_chunk >= end_chunk)
-			continue;
-		memblock_add_node(PFN_PHYS(start_chunk),
-				  PFN_PHYS(end_chunk - start_chunk), 0);
-		pfn = max(start_chunk, start_pfn);
-		storage_key_init_range(PFN_PHYS(pfn), PFN_PHYS(end_chunk));
+	for_each_memblock(memory, reg) {
+		storage_key_init_range(reg->base, reg->base + reg->size);
 	}
-
 	psw_set_key(PAGE_DEFAULT_KEY);
 
-	free_bootmem_with_active_regions(0, max_pfn);
-
-	/*
-	 * Reserve memory used for lowcore/command line/kernel image.
-	 */
-	reserve_bootmem(0, (unsigned long)_ehead, BOOTMEM_DEFAULT);
-	reserve_bootmem((unsigned long)_stext,
-			PFN_PHYS(start_pfn) - (unsigned long)_stext,
-			BOOTMEM_DEFAULT);
-	/*
-	 * Reserve the bootmem bitmap itself as well. We do this in two
-	 * steps (first step was init_bootmem()) because this catches
-	 * the (very unlikely) case of us accidentally initializing the
-	 * bootmem allocator with an invalid RAM area.
-	 */
-	reserve_bootmem(start_pfn << PAGE_SHIFT, bootmap_size,
-			BOOTMEM_DEFAULT);
-
-#ifdef CONFIG_CRASH_DUMP
-	if (crashk_res.start)
-		reserve_bootmem(crashk_res.start,
-				crashk_res.end - crashk_res.start + 1,
-				BOOTMEM_DEFAULT);
-	if (is_kdump_kernel())
-		reserve_bootmem(elfcorehdr_addr - OLDMEM_BASE,
-				PAGE_ALIGN(elfcorehdr_size), BOOTMEM_DEFAULT);
-#endif
-#ifdef CONFIG_BLK_DEV_INITRD
-	if (INITRD_START && INITRD_SIZE) {
-		if (INITRD_START + INITRD_SIZE <= memory_end) {
-			reserve_bootmem(INITRD_START, INITRD_SIZE,
-					BOOTMEM_DEFAULT);
-			initrd_start = INITRD_START;
-			initrd_end = initrd_start + INITRD_SIZE;
-		} else {
-			pr_err("initrd extends beyond end of "
-			       "memory (0x%08lx > 0x%08lx) "
-			       "disabling initrd\n",
-			       initrd_start + INITRD_SIZE, memory_end);
-			initrd_start = initrd_end = 0;
-		}
-	}
-#endif
+	/* Only cosmetics */
+	memblock_enforce_memory_limit(memblock_end_of_DRAM());
 }
 
 /*
@@ -989,23 +835,46 @@ void __init setup_arch(char **cmdline_p)
 
         ROOT_DEV = Root_RAM0;
 
+	/* Is init_mm really needed? */
 	init_mm.start_code = PAGE_OFFSET;
 	init_mm.end_code = (unsigned long) &_etext;
 	init_mm.end_data = (unsigned long) &_edata;
 	init_mm.brk = (unsigned long) &_end;
 
 	parse_early_param();
-	detect_memory_layout(memory_chunk, memory_end);
 	os_info_init();
 	setup_ipl();
+
+	/* Do some memory reservations *before* memory is added to memblock */
+	reserve_memory_end();
 	reserve_oldmem();
+	reserve_kernel();
+	reserve_initrd();
+	reserve_elfcorehdr();
+	memblock_allow_resize();
+
+	/* Get information about *all* installed memory */
+	detect_memory_memblock();
+
+	remove_oldmem();
+
+	/*
+	 * Make sure all chunks are MAX_ORDER aligned so we don't need the
+	 * extra checks that HOLES_IN_ZONE would require.
+	 *
+	 * Is this still required?
+	 */
+	memblock_trim_memory(1UL << (MAX_ORDER - 1 + PAGE_SHIFT));
+
 	setup_memory_end();
-	reserve_crashkernel();
 	setup_memory();
+
+	check_initrd();
+	reserve_crashkernel();
+
 	setup_resources();
 	setup_vmcoreinfo();
 	setup_lowcore();
-
 	smp_fill_possible_mask();
         cpu_init();
 	s390_init_cpu_topology();
diff --git a/arch/s390/kernel/topology.c b/arch/s390/kernel/topology.c
index 6298fed11ced..fa3b8cdaadac 100644
--- a/arch/s390/kernel/topology.c
+++ b/arch/s390/kernel/topology.c
@@ -333,7 +333,9 @@ static void __init alloc_masks(struct sysinfo_15_1_x *info,
 		nr_masks *= info->mag[TOPOLOGY_NR_MAG - offset - 1 - i];
 	nr_masks = max(nr_masks, 1);
 	for (i = 0; i < nr_masks; i++) {
-		mask->next = alloc_bootmem(sizeof(struct mask_info));
+		mask->next = alloc_bootmem_align(
+			roundup_pow_of_two(sizeof(struct mask_info)),
+			roundup_pow_of_two(sizeof(struct mask_info)));
 		mask = mask->next;
 	}
 }