4 files changed, 286 insertions, 47 deletions

diff --git a/arch/microblaze/mm/Makefile b/arch/microblaze/mm/Makefile
index 09c49ed87235..7313bd8acbb7 100644
--- a/arch/microblaze/mm/Makefile
+++ b/arch/microblaze/mm/Makefile
@@ -5,3 +5,4 @@
 obj-y := consistent.o init.o
 
 obj-$(CONFIG_MMU) += pgtable.o mmu_context.o fault.o
+obj-$(CONFIG_HIGHMEM) += highmem.o
diff --git a/arch/microblaze/mm/highmem.c b/arch/microblaze/mm/highmem.c
new file mode 100644
index 000000000000..7d78838e8bfa
--- /dev/null
+++ b/arch/microblaze/mm/highmem.c
@@ -0,0 +1,88 @@
+/*
+ * highmem.c: virtual kernel memory mappings for high memory
+ *
+ * PowerPC version, stolen from the i386 version.
+ *
+ * Used in CONFIG_HIGHMEM systems for memory pages which
+ * are not addressable by direct kernel virtual addresses.
+ *
+ * Copyright (C) 1999 Gerhard Wichert, Siemens AG
+ *		      Gerhard.Wichert@pdb.siemens.de
+ *
+ *
+ * Redesigned the x86 32-bit VM architecture to deal with
+ * up to 16 Terrabyte physical memory. With current x86 CPUs
+ * we now support up to 64 Gigabytes physical RAM.
+ *
+ * Copyright (C) 1999 Ingo Molnar <mingo@redhat.com>
+ *
+ * Reworked for PowerPC by various contributors. Moved from
+ * highmem.h by Benjamin Herrenschmidt (c) 2009 IBM Corp.
+ */
+
+#include <linux/highmem.h>
+#include <linux/module.h>
+
+/*
+ * The use of kmap_atomic/kunmap_atomic is discouraged - kmap/kunmap
+ * gives a more generic (and caching) interface. But kmap_atomic can
+ * be used in IRQ contexts, so in some (very limited) cases we need
+ * it.
+ */
+#include <asm/tlbflush.h>
+
+void *kmap_atomic_prot(struct page *page, pgprot_t prot)
+{
+
+	unsigned long vaddr;
+	int idx, type;
+
+	/* even !CONFIG_PREEMPT needs this, for in_atomic in do_page_fault */
+	pagefault_disable();
+	if (!PageHighMem(page))
+		return page_address(page);
+
+
+	type = kmap_atomic_idx_push();
+	idx = type + KM_TYPE_NR*smp_processor_id();
+	vaddr = __fix_to_virt(FIX_KMAP_BEGIN + idx);
+#ifdef CONFIG_DEBUG_HIGHMEM
+	BUG_ON(!pte_none(*(kmap_pte-idx)));
+#endif
+	set_pte_at(&init_mm, vaddr, kmap_pte-idx, mk_pte(page, prot));
+	local_flush_tlb_page(NULL, vaddr);
+
+	return (void *) vaddr;
+}
+EXPORT_SYMBOL(kmap_atomic_prot);
+
+void __kunmap_atomic(void *kvaddr)
+{
+	unsigned long vaddr = (unsigned long) kvaddr & PAGE_MASK;
+	int type;
+
+	if (vaddr < __fix_to_virt(FIX_KMAP_END)) {
+		pagefault_enable();
+		return;
+	}
+
+	type = kmap_atomic_idx();
+#ifdef CONFIG_DEBUG_HIGHMEM
+	{
+		unsigned int idx;
+
+		idx = type + KM_TYPE_NR * smp_processor_id();
+		BUG_ON(vaddr != __fix_to_virt(FIX_KMAP_BEGIN + idx));
+
+		/*
+		 * force other mappings to Oops if they'll try to access
+		 * this pte without first remap it
+		 */
+		pte_clear(&init_mm, vaddr, kmap_pte-idx);
+		local_flush_tlb_page(NULL, vaddr);
+	}
+#endif
+	kmap_atomic_idx_pop();
+	pagefault_enable();
+}
+EXPORT_SYMBOL(__kunmap_atomic);
diff --git a/arch/microblaze/mm/init.c b/arch/microblaze/mm/init.c
index 565d193c7ebf..ce80823051ba 100644
--- a/arch/microblaze/mm/init.c
+++ b/arch/microblaze/mm/init.c
@@ -24,6 +24,7 @@
 #include <asm/pgalloc.h>
 #include <asm/sections.h>
 #include <asm/tlb.h>
+#include <asm/fixmap.h>
 
 /* Use for MMU and noMMU because of PCI generic code */
 int mem_init_done;
@@ -44,9 +45,56 @@ char *klimit = _end;
  */
 unsigned long memory_start;
 EXPORT_SYMBOL(memory_start);
-unsigned long memory_end; /* due to mm/nommu.c */
 unsigned long memory_size;
 EXPORT_SYMBOL(memory_size);
+unsigned long lowmem_size;
+
+#ifdef CONFIG_HIGHMEM
+pte_t *kmap_pte;
+EXPORT_SYMBOL(kmap_pte);
+pgprot_t kmap_prot;
+EXPORT_SYMBOL(kmap_prot);
+
+static inline pte_t *virt_to_kpte(unsigned long vaddr)
+{
+	return pte_offset_kernel(pmd_offset(pgd_offset_k(vaddr),
+			vaddr), vaddr);
+}
+
+static void __init highmem_init(void)
+{
+	pr_debug("%x\n", (u32)PKMAP_BASE);
+	map_page(PKMAP_BASE, 0, 0);	/* XXX gross */
+	pkmap_page_table = virt_to_kpte(PKMAP_BASE);
+
+	kmap_pte = virt_to_kpte(__fix_to_virt(FIX_KMAP_BEGIN));
+	kmap_prot = PAGE_KERNEL;
+}
+
+static unsigned long highmem_setup(void)
+{
+	unsigned long pfn;
+	unsigned long reservedpages = 0;
+
+	for (pfn = max_low_pfn; pfn < max_pfn; ++pfn) {
+		struct page *page = pfn_to_page(pfn);
+
+		/* FIXME not sure about */
+		if (memblock_is_reserved(pfn << PAGE_SHIFT))
+			continue;
+		ClearPageReserved(page);
+		init_page_count(page);
+		__free_page(page);
+		totalhigh_pages++;
+		reservedpages++;
+	}
+	totalram_pages += totalhigh_pages;
+	printk(KERN_INFO "High memory: %luk\n",
+					totalhigh_pages << (PAGE_SHIFT-10));
+
+	return reservedpages;
+}
+#endif /* CONFIG_HIGHMEM */
 
 /*
  * paging_init() sets up the page tables - in fact we've already done this.
@@ -54,17 +102,28 @@ EXPORT_SYMBOL(memory_size);
 static void __init paging_init(void)
 {
 	unsigned long zones_size[MAX_NR_ZONES];
+#ifdef CONFIG_MMU
+	int idx;
+
+	/* Setup fixmaps */
+	for (idx = 0; idx < __end_of_fixed_addresses; idx++)
+		clear_fixmap(idx);
+#endif
 
 	/* Clean every zones */
 	memset(zones_size, 0, sizeof(zones_size));
 
-	/*
-	 * old: we can DMA to/from any address.put all page into ZONE_DMA
-	 * We use only ZONE_NORMAL
-	 */
-	zones_size[ZONE_NORMAL] = max_mapnr;
+#ifdef CONFIG_HIGHMEM
+	highmem_init();
 
-	free_area_init(zones_size);
+	zones_size[ZONE_DMA] = max_low_pfn;
+	zones_size[ZONE_HIGHMEM] = max_pfn;
+#else
+	zones_size[ZONE_DMA] = max_pfn;
+#endif
+
+	/* We don't have holes in memory map */
+	free_area_init_nodes(zones_size);
 }
 
 void __init setup_memory(void)
@@ -78,32 +137,31 @@ void __init setup_memory(void)
 	/* Find main memory where is the kernel */
 	for_each_memblock(memory, reg) {
 		memory_start = (u32)reg->base;
-		memory_end = (u32) reg->base + reg->size;
+		lowmem_size = reg->size;
 		if ((memory_start <= (u32)_text) &&
-					((u32)_text <= memory_end)) {
-			memory_size = memory_end - memory_start;
+			((u32)_text <= (memory_start + lowmem_size - 1))) {
+			memory_size = lowmem_size;
 			PAGE_OFFSET = memory_start;
-			printk(KERN_INFO "%s: Main mem: 0x%x-0x%x, "
+			printk(KERN_INFO "%s: Main mem: 0x%x, "
 				"size 0x%08x\n", __func__, (u32) memory_start,
-					(u32) memory_end, (u32) memory_size);
+					(u32) memory_size);
 			break;
 		}
 	}
 
-	if (!memory_start || !memory_end) {
-		panic("%s: Missing memory setting 0x%08x-0x%08x\n",
-			__func__, (u32) memory_start, (u32) memory_end);
+	if (!memory_start || !memory_size) {
+		panic("%s: Missing memory setting 0x%08x, size=0x%08x\n",
+			__func__, (u32) memory_start, (u32) memory_size);
 	}
 
 	/* reservation of region where is the kernel */
 	kernel_align_start = PAGE_DOWN((u32)_text);
 	/* ALIGN can be remove because _end in vmlinux.lds.S is align */
 	kernel_align_size = PAGE_UP((u32)klimit) - kernel_align_start;
-	memblock_reserve(kernel_align_start, kernel_align_size);
-	printk(KERN_INFO "%s: kernel addr=0x%08x-0x%08x size=0x%08x\n",
+	printk(KERN_INFO "%s: kernel addr:0x%08x-0x%08x size=0x%08x\n",
 		__func__, kernel_align_start, kernel_align_start
 			+ kernel_align_size, kernel_align_size);
-
+	memblock_reserve(kernel_align_start, kernel_align_size);
 #endif
 	/*
 	 * Kernel:
@@ -120,11 +178,13 @@ void __init setup_memory(void)
 	min_low_pfn = memory_start >> PAGE_SHIFT; /* minimum for allocation */
 	/* RAM is assumed contiguous */
 	num_physpages = max_mapnr = memory_size >> PAGE_SHIFT;
-	max_pfn = max_low_pfn = memory_end >> PAGE_SHIFT;
+	max_low_pfn = ((u64)memory_start + (u64)lowmem_size) >> PAGE_SHIFT;
+	max_pfn = ((u64)memory_start + (u64)memory_size) >> PAGE_SHIFT;
 
 	printk(KERN_INFO "%s: max_mapnr: %#lx\n", __func__, max_mapnr);
 	printk(KERN_INFO "%s: min_low_pfn: %#lx\n", __func__, min_low_pfn);
 	printk(KERN_INFO "%s: max_low_pfn: %#lx\n", __func__, max_low_pfn);
+	printk(KERN_INFO "%s: max_pfn: %#lx\n", __func__, max_pfn);
 
 	/*
 	 * Find an area to use for the bootmem bitmap.
@@ -137,15 +197,39 @@ void __init setup_memory(void)
 		PFN_UP(TOPHYS((u32)klimit)), min_low_pfn, max_low_pfn);
 	memblock_reserve(PFN_UP(TOPHYS((u32)klimit)) << PAGE_SHIFT, map_size);
 
+	/* Add active regions with valid PFNs */
+	for_each_memblock(memory, reg) {
+		unsigned long start_pfn, end_pfn;
+
+		start_pfn = memblock_region_memory_base_pfn(reg);
+		end_pfn = memblock_region_memory_end_pfn(reg);
+		memblock_set_node(start_pfn << PAGE_SHIFT,
+					(end_pfn - start_pfn) << PAGE_SHIFT, 0);
+	}
+
 	/* free bootmem is whole main memory */
-	free_bootmem(memory_start, memory_size);
+	free_bootmem_with_active_regions(0, max_low_pfn);
 
 	/* reserve allocate blocks */
 	for_each_memblock(reserved, reg) {
-		pr_debug("reserved - 0x%08x-0x%08x\n",
-			 (u32) reg->base, (u32) reg->size);
-		reserve_bootmem(reg->base, reg->size, BOOTMEM_DEFAULT);
+		unsigned long top = reg->base + reg->size - 1;
+
+		pr_debug("reserved - 0x%08x-0x%08x, %lx, %lx\n",
+			 (u32) reg->base, (u32) reg->size, top,
+						memory_start + lowmem_size - 1);
+
+		if (top <= (memory_start + lowmem_size - 1)) {
+			reserve_bootmem(reg->base, reg->size, BOOTMEM_DEFAULT);
+		} else if (reg->base < (memory_start + lowmem_size - 1)) {
+			unsigned long trunc_size = memory_start + lowmem_size -
+								reg->base;
+			reserve_bootmem(reg->base, trunc_size, BOOTMEM_DEFAULT);
+		}
 	}
+
+	/* XXX need to clip this if using highmem? */
+	sparse_memory_present_with_active_regions(0);
+
 #ifdef CONFIG_MMU
 	init_bootmem_done = 1;
 #endif
@@ -190,13 +274,58 @@ void free_initmem(void)
 
 void __init mem_init(void)
 {
-	high_memory = (void *)__va(memory_end);
+	pg_data_t *pgdat;
+	unsigned long reservedpages = 0, codesize, initsize, datasize, bsssize;
+
+	high_memory = (void *)__va(memory_start + lowmem_size - 1);
+
 	/* this will put all memory onto the freelists */
 	totalram_pages += free_all_bootmem();
 
-	printk(KERN_INFO "Memory: %luk/%luk available\n",
-	       nr_free_pages() << (PAGE_SHIFT-10),
-	       num_physpages << (PAGE_SHIFT-10));
+	for_each_online_pgdat(pgdat) {
+		unsigned long i;
+		struct page *page;
+
+		for (i = 0; i < pgdat->node_spanned_pages; i++) {
+			if (!pfn_valid(pgdat->node_start_pfn + i))
+				continue;
+			page = pgdat_page_nr(pgdat, i);
+			if (PageReserved(page))
+				reservedpages++;
+		}
+	}
+
+#ifdef CONFIG_HIGHMEM
+	reservedpages -= highmem_setup();
+#endif
+
+	codesize = (unsigned long)&_sdata - (unsigned long)&_stext;
+	datasize = (unsigned long)&_edata - (unsigned long)&_sdata;
+	initsize = (unsigned long)&__init_end - (unsigned long)&__init_begin;
+	bsssize = (unsigned long)&__bss_stop - (unsigned long)&__bss_start;
+
+	pr_info("Memory: %luk/%luk available (%luk kernel code, "
+		"%luk reserved, %luk data, %luk bss, %luk init)\n",
+		nr_free_pages() << (PAGE_SHIFT-10),
+		num_physpages << (PAGE_SHIFT-10),
+		codesize >> 10,
+		reservedpages << (PAGE_SHIFT-10),
+		datasize >> 10,
+		bsssize >> 10,
+		initsize >> 10);
+
+#ifdef CONFIG_MMU
+	pr_info("Kernel virtual memory layout:\n");
+	pr_info("  * 0x%08lx..0x%08lx  : fixmap\n", FIXADDR_START, FIXADDR_TOP);
+#ifdef CONFIG_HIGHMEM
+	pr_info("  * 0x%08lx..0x%08lx  : highmem PTEs\n",
+		PKMAP_BASE, PKMAP_ADDR(LAST_PKMAP));
+#endif /* CONFIG_HIGHMEM */
+	pr_info("  * 0x%08lx..0x%08lx  : early ioremap\n",
+		ioremap_bot, ioremap_base);
+	pr_info("  * 0x%08lx..0x%08lx  : vmalloc & ioremap\n",
+		(unsigned long)VMALLOC_START, VMALLOC_END);
+#endif
 	mem_init_done = 1;
 }
 
@@ -226,7 +355,6 @@ static void mm_cmdline_setup(void)
 		maxmem = memparse(p, &p);
 		if (maxmem && memory_size > maxmem) {
 			memory_size = maxmem;
-			memory_end = memory_start + memory_size;
 			memblock.memory.regions[0].size = memory_size;
 		}
 	}
@@ -270,15 +398,26 @@ asmlinkage void __init mmu_init(void)
 		machine_restart(NULL);
 	}
 
-	if ((u32) memblock.memory.regions[0].size < 0x1000000) {
-		printk(KERN_EMERG "Memory must be greater than 16MB\n");
+	if ((u32) memblock.memory.regions[0].size < 0x400000) {
+		printk(KERN_EMERG "Memory must be greater than 4MB\n");
+		machine_restart(NULL);
+	}
+
+	if ((u32) memblock.memory.regions[0].size < kernel_tlb) {
+		printk(KERN_EMERG "Kernel size is greater than memory node\n");
 		machine_restart(NULL);
 	}
+
 	/* Find main memory where the kernel is */
 	memory_start = (u32) memblock.memory.regions[0].base;
-	memory_end = (u32) memblock.memory.regions[0].base +
-				(u32) memblock.memory.regions[0].size;
-	memory_size = memory_end - memory_start;
+	lowmem_size = memory_size = (u32) memblock.memory.regions[0].size;
+
+	if (lowmem_size > CONFIG_LOWMEM_SIZE) {
+		lowmem_size = CONFIG_LOWMEM_SIZE;
+#ifndef CONFIG_HIGHMEM
+		memory_size = lowmem_size;
+#endif
+	}
 
 	mm_cmdline_setup(); /* FIXME parse args from command line - not used */
 
@@ -305,15 +444,20 @@ asmlinkage void __init mmu_init(void)
 	/* Map in all of RAM starting at CONFIG_KERNEL_START */
 	mapin_ram();
 
-#ifdef CONFIG_HIGHMEM_START_BOOL
-	ioremap_base = CONFIG_HIGHMEM_START;
+	/* Extend vmalloc and ioremap area as big as possible */
+#ifdef CONFIG_HIGHMEM
+	ioremap_base = ioremap_bot = PKMAP_BASE;
 #else
-	ioremap_base = 0xfe000000UL;	/* for now, could be 0xfffff000 */
-#endif /* CONFIG_HIGHMEM_START_BOOL */
-	ioremap_bot = ioremap_base;
+	ioremap_base = ioremap_bot = FIXADDR_START;
+#endif
 
 	/* Initialize the context management stuff */
 	mmu_context_init();
+
+	/* Shortly after that, the entire linear mapping will be available */
+	/* This will also cause that unflatten device tree will be allocated
+	 * inside 768MB limit */
+	memblock_set_current_limit(memory_start + lowmem_size - 1);
 }
 
 /* This is only called until mem_init is done. */
@@ -324,11 +468,11 @@ void __init *early_get_page(void)
 		p = alloc_bootmem_pages(PAGE_SIZE);
 	} else {
 		/*
-		 * Mem start + 32MB -> here is limit
+		 * Mem start + kernel_tlb -> here is limit
 		 * because of mem mapping from head.S
 		 */
 		p = __va(memblock_alloc_base(PAGE_SIZE, PAGE_SIZE,
-					memory_start + 0x2000000));
+					memory_start + kernel_tlb));
 	}
 	return p;
 }
diff --git a/arch/microblaze/mm/pgtable.c b/arch/microblaze/mm/pgtable.c
index 59bf2335a4ce..d1c06d07fed8 100644
--- a/arch/microblaze/mm/pgtable.c
+++ b/arch/microblaze/mm/pgtable.c
@@ -37,6 +37,7 @@
 #include <linux/io.h>
 #include <asm/mmu.h>
 #include <asm/sections.h>
+#include <asm/fixmap.h>
 
 #define flush_HPTE(X, va, pg)	_tlbie(va)
 
@@ -44,11 +45,6 @@ unsigned long ioremap_base;
 unsigned long ioremap_bot;
 EXPORT_SYMBOL(ioremap_bot);
 
-/* The maximum lowmem defaults to 768Mb, but this can be configured to
- * another value.
- */
-#define MAX_LOW_MEM	CONFIG_LOWMEM_SIZE
-
 #ifndef CONFIG_SMP
 struct pgtable_cache_struct quicklists;
 #endif
@@ -80,7 +76,7 @@ static void __iomem *__ioremap(phys_addr_t addr, unsigned long size,
 		!(p >= virt_to_phys((unsigned long)&__bss_stop) &&
 		p < virt_to_phys((unsigned long)__bss_stop))) {
 		printk(KERN_WARNING "__ioremap(): phys addr "PTE_FMT
-			" is RAM lr %p\n", (unsigned long)p,
+			" is RAM lr %pf\n", (unsigned long)p,
 			__builtin_return_address(0));
 		return NULL;
 	}
@@ -171,7 +167,7 @@ void __init mapin_ram(void)
 
 	v = CONFIG_KERNEL_START;
 	p = memory_start;
-	for (s = 0; s < memory_size; s += PAGE_SIZE) {
+	for (s = 0; s < lowmem_size; s += PAGE_SIZE) {
 		f = _PAGE_PRESENT | _PAGE_ACCESSED |
 				_PAGE_SHARED | _PAGE_HWEXEC;
 		if ((char *) v < _stext || (char *) v >= _etext)
@@ -254,3 +250,13 @@ __init_refok pte_t *pte_alloc_one_kernel(struct mm_struct *mm,
 	}
 	return pte;
 }
+
+void __set_fixmap(enum fixed_addresses idx, phys_addr_t phys, pgprot_t flags)
+{
+	unsigned long address = __fix_to_virt(idx);
+
+	if (idx >= __end_of_fixed_addresses)
+		BUG();
+
+	map_page(address, phys, pgprot_val(flags));
+}