x86: merge setup_32/64.c

Index: linux-2.6/arch/x86/kernel/setup_32.c
===================================================================
@@ -186,43 +186,18 @@ static inline void copy_edd(void)
 
 #ifdef CONFIG_BLK_DEV_INITRD
 
-static void __init relocate_initrd(void);
+#ifdef CONFIG_X86_32
 
-static void __init reserve_initrd(void)
+#define MAX_MAP_CHUNK	(NR_FIX_BTMAPS << PAGE_SHIFT)
+static void __init relocate_initrd(void)
 {
+
 	u64 ramdisk_image = boot_params.hdr.ramdisk_image;
 	u64 ramdisk_size  = boot_params.hdr.ramdisk_size;
-	u64 ramdisk_end   = ramdisk_image + ramdisk_size;
 	u64 end_of_lowmem = max_low_pfn << PAGE_SHIFT;
 	u64 ramdisk_here;
-
-	if (!boot_params.hdr.type_of_loader ||
-	    !ramdisk_image || !ramdisk_size)
-		return;		/* No initrd provided by bootloader */
-
-	initrd_start = 0;
-
-	if (ramdisk_size >= (end_of_lowmem>>1)) {
-		free_early(ramdisk_image, ramdisk_end);
-		printk(KERN_ERR "initrd too large to handle, "
-		       "disabling initrd\n");
-		return;
-	}
-
-	printk(KERN_INFO "old RAMDISK: %08llx - %08llx\n", ramdisk_image,
-			ramdisk_end);
-
-
-	if (ramdisk_end <= end_of_lowmem) {
-		/* All in lowmem, easy case */
-		/*
-		 * don't need to reserve again, already reserved early
-		 * in i386_start_kernel
-		 */
-		initrd_start = ramdisk_image + PAGE_OFFSET;
-		initrd_end = initrd_start + ramdisk_size;
-		return;
-	}
+	unsigned long slop, clen, mapaddr;
+	char *p, *q;
 
 	/* We need to move the initrd down into lowmem */
 	ramdisk_here = find_e820_area(0, end_of_lowmem, ramdisk_size,

@@ -241,22 +216,6 @@ static void __init reserve_initrd(void)
 	printk(KERN_INFO "Allocated new RAMDISK: %08llx - %08llx\n",
 			 ramdisk_here, ramdisk_here + ramdisk_size);
 
-	relocate_initrd();
-}
-
-#define MAX_MAP_CHUNK	(NR_FIX_BTMAPS << PAGE_SHIFT)
-
-static void __init relocate_initrd(void)
-{
-	u64 ramdisk_image = boot_params.hdr.ramdisk_image;
-	u64 ramdisk_size  = boot_params.hdr.ramdisk_size;
-	u64 end_of_lowmem = max_low_pfn << PAGE_SHIFT;
-	u64 ramdisk_here;
-	unsigned long slop, clen, mapaddr;
-	char *p, *q;
-
-	ramdisk_here = initrd_start - PAGE_OFFSET;
-
 	q = (char *)initrd_start;
 
 	/* Copy any lowmem portion of the initrd */

@@ -286,18 +245,60 @@ static void __init relocate_initrd(void)
 	/* high pages is not converted by early_res_to_bootmem */
 	ramdisk_image = boot_params.hdr.ramdisk_image;
 	ramdisk_size  = boot_params.hdr.ramdisk_size;
-	printk(KERN_INFO "Copied RAMDISK from %016llx - %016llx to %08llx - %08llx\n",
+	printk(KERN_INFO "Move RAMDISK from %016llx - %016llx to"
+		" %08llx - %08llx\n",
 		ramdisk_image, ramdisk_image + ramdisk_size - 1,
 		ramdisk_here, ramdisk_here + ramdisk_size - 1);
+}
+#endif
 
-	/*
-	 * need to free old one, otherwise init cross max_low_pfn could be
-	 * converted to bootmem
-	 */
-	free_early(ramdisk_image, ramdisk_image+ramdisk_size);
+static void __init reserve_initrd(void)
+{
+	u64 ramdisk_image = boot_params.hdr.ramdisk_image;
+	u64 ramdisk_size  = boot_params.hdr.ramdisk_size;
+	u64 ramdisk_end   = ramdisk_image + ramdisk_size;
+	u64 end_of_lowmem = max_low_pfn << PAGE_SHIFT;
+
+	if (!boot_params.hdr.type_of_loader ||
+	    !ramdisk_image || !ramdisk_size)
+		return;		/* No initrd provided by bootloader */
+
+	initrd_start = 0;
+
+	if (ramdisk_size >= (end_of_lowmem>>1)) {
+		free_early(ramdisk_image, ramdisk_end);
+		printk(KERN_ERR "initrd too large to handle, "
+		       "disabling initrd\n");
+		return;
+	}
+
+	printk(KERN_INFO "RAMDISK: %08llx - %08llx\n", ramdisk_image,
+			ramdisk_end);
+
+
+	if (ramdisk_end <= end_of_lowmem) {
+		/* All in lowmem, easy case */
+		/*
+		 * don't need to reserve again, already reserved early
+		 * in i386_start_kernel
+		 */
+		initrd_start = ramdisk_image + PAGE_OFFSET;
+		initrd_end = initrd_start + ramdisk_size;
+		return;
+	}
+
+#ifdef CONFIG_X86_32
+	relocate_initrd();
+#else
+	printk(KERN_ERR "initrd extends beyond end of memory "
+	       "(0x%08llx > 0x%08llx)\ndisabling initrd\n",
+	       ramdisk_end, end_of_lowmem);
+	initrd_start = 0;
+#endif
+	free_early(ramdisk_image, ramdisk_end);
 }
 #else
-void __init reserve_initrd(void)
+static void __init reserve_initrd(void)
 {
 }
 #endif /* CONFIG_BLK_DEV_INITRD */

Index: linux-2.6/arch/x86/kernel/setup_32.c
===================================================================
@@ -100,6 +100,8 @@ static struct resource bss_resource = {
 	.flags	= IORESOURCE_BUSY | IORESOURCE_MEM
 };
 
+
+#ifdef CONFIG_X86_32
 static struct resource video_ram_resource = {
 	.name	= "Video RAM area",
 	.start	= 0xa0000,

@@ -108,24 +110,47 @@ static struct resource video_ram_resourc
 };
 
 /* cpu data as detected by the assembly code in head.S */
-struct cpuinfo_x86 new_cpu_data __cpuinitdata = { 0, 0, 0, 0, -1, 1, 0, 0, -1 };
+struct cpuinfo_x86 new_cpu_data __cpuinitdata = {0, 0, 0, 0, -1, 1, 0, 0, -1};
 /* common cpu data for all cpus */
-struct cpuinfo_x86 boot_cpu_data __read_mostly = { 0, 0, 0, 0, -1, 1, 0, 0, -1 };
+struct cpuinfo_x86 boot_cpu_data __read_mostly = {0, 0, 0, 0, -1, 1, 0, 0, -1};
 EXPORT_SYMBOL(boot_cpu_data);
+static void set_mca_bus(int x)
+{
+#ifdef CONFIG_MCA
+	MCA_bus = x;
+#endif
+}
 
 unsigned int def_to_bigsmp;
 
-#ifndef CONFIG_X86_PAE
-unsigned long mmu_cr4_features;
-#else
-unsigned long mmu_cr4_features = X86_CR4_PAE;
-#endif
-
 /* for MCA, but anyone else can use it if they want */
 unsigned int machine_id;
 unsigned int machine_submodel_id;
 unsigned int BIOS_revision;
 
+struct apm_info apm_info;
+EXPORT_SYMBOL(apm_info);
+
+#if defined(CONFIG_X86_SPEEDSTEP_SMI) || \
+	defined(CONFIG_X86_SPEEDSTEP_SMI_MODULE)
+struct ist_info ist_info;
+EXPORT_SYMBOL(ist_info);
+#else
+struct ist_info ist_info;
+#endif
+
+#else
+struct cpuinfo_x86 boot_cpu_data __read_mostly;
+EXPORT_SYMBOL(boot_cpu_data);
+#endif
+
+
+#if !defined(CONFIG_X86_PAE) || defined(CONFIG_X86_64)
+unsigned long mmu_cr4_features;
+#else
+unsigned long mmu_cr4_features = X86_CR4_PAE;
+#endif
+
 /* Boot loader ID as an integer, for the benefit of proc_dointvec */
 int bootloader_type;
 

@@ -140,15 +165,8 @@ char dmi_alloc_data[DMI_MAX_DATA];
  */
 struct screen_info screen_info;
 EXPORT_SYMBOL(screen_info);
-struct apm_info apm_info;
-EXPORT_SYMBOL(apm_info);
 struct edid_info edid_info;
 EXPORT_SYMBOL_GPL(edid_info);
-struct ist_info ist_info;
-#if defined(CONFIG_X86_SPEEDSTEP_SMI) || \
-	defined(CONFIG_X86_SPEEDSTEP_SMI_MODULE)
-EXPORT_SYMBOL(ist_info);
-#endif
 
 extern int root_mountflags;
 

@@ -303,15 +321,6 @@ static void __init reserve_initrd(void)
 }
 #endif /* CONFIG_BLK_DEV_INITRD */
 
-#ifdef CONFIG_MCA
-static void set_mca_bus(int x)
-{
-	MCA_bus = x;
-}
-#else
-static void set_mca_bus(int x) { }
-#endif
-
 /*
  * Determine if we were loaded by an EFI loader.  If so, then we have also been
  * passed the efi memmap, systab, etc., so we should use these data structures

Index: linux-2.6/arch/x86/kernel/setup_32.c
===================================================================
@@ -45,31 +45,75 @@
 #include <linux/dmi.h>
 #include <linux/pfn.h>
 #include <linux/pci.h>
+#include <asm/pci-direct.h>
 #include <linux/init_ohci1394_dma.h>
 #include <linux/kvm_para.h>
 
+#include <linux/errno.h>
+#include <linux/kernel.h>
+#include <linux/stddef.h>
+#include <linux/unistd.h>
+#include <linux/ptrace.h>
+#include <linux/slab.h>
+#include <linux/user.h>
+#include <linux/delay.h>
+#include <linux/highmem.h>
+
+#include <linux/kallsyms.h>
+#include <linux/edd.h>
+#include <linux/iscsi_ibft.h>
+#include <linux/kexec.h>
+#include <linux/cpufreq.h>
+#include <linux/dma-mapping.h>
+#include <linux/ctype.h>
+#include <linux/uaccess.h>
+
+#include <linux/percpu.h>
+#include <linux/crash_dump.h>
+
 #include <video/edid.h>
 
 #include <asm/mtrr.h>
 #include <asm/apic.h>
 #include <asm/e820.h>
 #include <asm/mpspec.h>
-#include <asm/mmzone.h>
 #include <asm/setup.h>
 #include <asm/arch_hooks.h>
 #include <asm/sections.h>
 #include <asm/dmi.h>
 #include <asm/io_apic.h>
 #include <asm/ist.h>
-#include <asm/io.h>
 #include <asm/vmi.h>
 #include <setup_arch.h>
 #include <asm/bios_ebda.h>
 #include <asm/cacheflush.h>
 #include <asm/processor.h>
-#include <asm/efi.h>
 #include <asm/bugs.h>
 
+#include <asm/system.h>
+#include <asm/vsyscall.h>
+#include <asm/smp.h>
+#include <asm/desc.h>
+#include <asm/dma.h>
+#include <asm/gart.h>
+#include <asm/mmu_context.h>
+#include <asm/proto.h>
+
+#include <mach_apic.h>
+#ifdef CONFIG_PARAVIRT
+#include <asm/paravirt.h>
+#else
+#define ARCH_SETUP
+#endif
+
+#include <asm/percpu.h>
+#include <asm/sections.h>
+#include <asm/topology.h>
+#include <asm/apicdef.h>
+#ifdef CONFIG_X86_32
+#include <asm/highmem.h>
+#endif
+
 /* This value is set up by the early boot code to point to the value
    immediately after the boot time page tables.  It contains a *physical*
    address, and must not be in the .bss segment! */

Index: linux-2.6/arch/x86/kernel/setup_32.c
===================================================================
@@ -372,26 +372,38 @@ static void __init reserve_initrd(void)
  * global efi_enabled. This allows the same kernel image to be used on existing
  * systems (with a traditional BIOS) as well as on EFI systems.
  */
+/*
+ * setup_arch - architecture-specific boot-time initializations
+ *
+ * Note: On x86_64, fixmaps are ready for use even before this is called.
+ */
+
 void __init setup_arch(char **cmdline_p)
 {
+#ifdef CONFIG_X86_32
 	memcpy(&boot_cpu_data, &new_cpu_data, sizeof(new_cpu_data));
 	pre_setup_arch_hook();
 	early_cpu_init();
 	early_ioremap_init();
 	reserve_setup_data();
+#else
+	printk(KERN_INFO "Command line: %s\n", boot_command_line);
+#endif
 
 	ROOT_DEV = old_decode_dev(boot_params.hdr.root_dev);
 	screen_info = boot_params.screen_info;
 	edid_info = boot_params.edid_info;
+#ifdef CONFIG_X86_32
 	apm_info.bios = boot_params.apm_bios_info;
 	ist_info = boot_params.ist_info;
-	saved_video_mode = boot_params.hdr.vid_mode;
-	if( boot_params.sys_desc_table.length != 0 ) {
+	if (boot_params.sys_desc_table.length != 0) {
 		set_mca_bus(boot_params.sys_desc_table.table[3] & 0x2);
 		machine_id = boot_params.sys_desc_table.table[0];
 		machine_submodel_id = boot_params.sys_desc_table.table[1];
 		BIOS_revision = boot_params.sys_desc_table.table[2];
 	}
+#endif
+	saved_video_mode = boot_params.hdr.vid_mode;
 	bootloader_type = boot_params.hdr.type_of_loader;
 
 #ifdef CONFIG_BLK_DEV_RAM

@@ -401,7 +413,12 @@ void __init setup_arch(char **cmdline_p)
 #endif
 #ifdef CONFIG_EFI
 	if (!strncmp((char *)&boot_params.efi_info.efi_loader_signature,
-		     "EL32", 4)) {
+#ifdef CONFIG_X86_32
+		     "EL32",
+#else
+		     "EL64",
+#endif
+	 4)) {
 		efi_enabled = 1;
 		efi_reserve_early();
 	}

@@ -417,7 +434,11 @@ void __init setup_arch(char **cmdline_p)
 	init_mm.start_code = (unsigned long) _text;
 	init_mm.end_code = (unsigned long) _etext;
 	init_mm.end_data = (unsigned long) _edata;
+#ifdef CONFIG_X86_32
 	init_mm.brk = init_pg_tables_end + PAGE_OFFSET;
+#else
+	init_mm.brk = (unsigned long) &_end;
+#endif
 
 	code_resource.start = virt_to_phys(_text);
 	code_resource.end = virt_to_phys(_etext)-1;

@@ -426,6 +447,9 @@ void __init setup_arch(char **cmdline_p)
 	bss_resource.start = virt_to_phys(&__bss_start);
 	bss_resource.end = virt_to_phys(&__bss_stop)-1;
 
+#ifdef CONFIG_X86_64
+	early_cpu_init();
+#endif
 	strlcpy(command_line, boot_command_line, COMMAND_LINE_SIZE);
 	*cmdline_p = command_line;
 

@@ -433,16 +457,27 @@ void __init setup_arch(char **cmdline_p)
 
 	parse_early_param();
 
-	if (acpi_mps_check()){
+	if (acpi_mps_check()) {
 #ifdef CONFIG_X86_LOCAL_APIC
+#ifdef CONFIG_X86_32
 		enable_local_apic = -1;
+#else
+		disable_apic = 1;
+#endif
 #endif
 		clear_cpu_cap(&boot_cpu_data, X86_FEATURE_APIC);
 	}
 
 	finish_e820_parsing();
 
+#ifdef CONFIG_X86_32
 	probe_roms();
+#else
+# ifdef CONFIG_PROVIDE_OHCI1394_DMA_INIT
+	if (init_ohci1394_dma_early)
+		init_ohci1394_dma_on_all_controllers();
+# endif
+#endif
 
 	/* after parse_early_param, so could debug it */
 	insert_resource(&iomem_resource, &code_resource);

@@ -452,6 +487,7 @@ void __init setup_arch(char **cmdline_p)
 	if (efi_enabled)
 		efi_init();
 
+#ifdef CONFIG_X86_32
 	if (ppro_with_ram_bug()) {
 		e820_update_range(0x70000000ULL, 0x40000ULL, E820_RAM,
 				  E820_RESERVED);

@@ -459,6 +495,9 @@ void __init setup_arch(char **cmdline_p)
 		printk(KERN_INFO "fixed physical RAM map:\n");
 		e820_print_map("bad_ppro");
 	}
+#else
+	early_gart_iommu_check();
+#endif
 
 	e820_register_active_regions(0, 0, -1UL);
 	/*

@@ -477,14 +516,32 @@ void __init setup_arch(char **cmdline_p)
 		max_pfn = e820_end_of_ram();
 	}
 
+#ifdef CONFIG_X86_32
 	/* max_low_pfn get updated here */
 	find_low_pfn_range();
+#else
+	num_physpages = max_pfn;
+
+	check_efer();
+
+	/* How many end-of-memory variables you have, grandma! */
+	/* need this before calling reserve_initrd */
+	max_low_pfn = max_pfn;
+	high_memory = (void *)__va(max_pfn * PAGE_SIZE - 1) + 1;
+#endif
 
 	/* max_pfn_mapped is updated here */
+#ifdef CONFIG_X86_64
+	max_pfn_mapped =
+#endif
 	init_memory_mapping(0, (max_low_pfn << PAGE_SHIFT));
 
 	reserve_initrd();
 
+#ifdef CONFIG_X86_64
+	vsmp_init();
+#endif
+
 	dmi_scan_machine();
 
 	io_delay_init();

@@ -494,6 +551,11 @@ void __init setup_arch(char **cmdline_p)
 	 */
 	acpi_boot_table_init();
 
+#ifdef CONFIG_X86_64
+	/* Remove active ranges so rediscovery with NUMA-awareness happens */
+	remove_all_active_ranges();
+#endif
+
 #ifdef CONFIG_ACPI_NUMA
         /*
          * Parse SRAT to discover nodes.

@@ -503,13 +565,18 @@ void __init setup_arch(char **cmdline_p)
 
 	initmem_init(0, max_pfn);
 
+#ifdef CONFIG_X86_64
+	dma32_reserve_bootmem();
+#endif
+
 #ifdef CONFIG_ACPI_SLEEP
 	/*
 	 * Reserve low memory region for sleep support.
 	 */
 	acpi_reserve_bootmem();
 #endif
-#ifdef CONFIG_X86_FIND_SMP_CONFIG
+#if defined(CONFIG_X86_FIND_SMP_CONFIG) && defined(CONFIG_X86_32) || \
+    defined(CONFIG_X86_MPPARSE) && defined(CONFIG_X86_64)
 	/*
 	 * Find and reserve possible boot-time SMP configuration:
 	 */

@@ -523,7 +590,7 @@ void __init setup_arch(char **cmdline_p)
 	kvmclock_init();
 #endif
 
-#ifdef CONFIG_VMI
+#if defined(CONFIG_VMI) && defined(CONFIG_X86_32)
 	/*
 	 * Must be after max_low_pfn is determined, and before kernel
 	 * pagetables are setup.

@@ -533,11 +600,15 @@ void __init setup_arch(char **cmdline_p)
 
 	paging_init();
 
+#ifdef CONFIG_X86_64
+	map_vsyscall();
+#endif
+
 	/*
 	 * NOTE: On x86-32, only from this point on, fixmaps are ready for use.
 	 */
 
-#ifdef CONFIG_PROVIDE_OHCI1394_DMA_INIT
+#if defined(CONFIG_PROVIDE_OHCI1394_DMA_INIT) && defined(CONFIG_X86_32)
 	if (init_ohci1394_dma_early)
 		init_ohci1394_dma_on_all_controllers();
 #endif

@@ -553,6 +624,10 @@ void __init setup_arch(char **cmdline_p)
 	 */
 	acpi_boot_init();
 
+#ifdef CONFIG_X86_64
+	init_cpu_to_node();
+#endif
+
 #if defined(CONFIG_X86_MPPARSE) || defined(CONFIG_X86_VISWS)
 	/*
 	 * get boot-time SMP configuration:

@@ -560,18 +635,26 @@ void __init setup_arch(char **cmdline_p)
 	if (smp_found_config)
 		get_smp_config();
 #endif
-#if defined(CONFIG_SMP) && defined(CONFIG_X86_PC)
+
+#ifdef CONFIG_X86_64
+	init_apic_mappings();
+	ioapic_init_mappings();
+#else
+# if defined(CONFIG_SMP) && defined(CONFIG_X86_PC)
 	if (def_to_bigsmp)
 		printk(KERN_WARNING "More than 8 CPUs detected and "
 			"CONFIG_X86_PC cannot handle it.\nUse "
 			"CONFIG_X86_GENERICARCH or CONFIG_X86_BIGSMP.\n");
+# endif
 #endif
 	kvm_guest_init();
 
 	e820_reserve_resources();
 	e820_mark_nosave_regions(max_low_pfn);
 
+#ifdef CONFIG_X86_32
 	request_resource(&iomem_resource, &video_ram_resource);
+#endif
 	reserve_standard_io_resources();
 
 	e820_setup_gap();

Index: linux-2.6/arch/x86/kernel/setup_32.c
===================================================================
@@ -557,10 +557,10 @@ void __init setup_arch(char **cmdline_p)
 #endif
 
 #ifdef CONFIG_ACPI_NUMA
-        /*
-         * Parse SRAT to discover nodes.
-         */
-        acpi_numa_init();
+	/*
+	 * Parse SRAT to discover nodes.
+	 */
+	acpi_numa_init();
 #endif
 
 	initmem_init(0, max_pfn);

671 lines not on this page · raw message

671 lines not on this page · raw message

380 lines not on this page · raw message

Index: linux-2.6/arch/x86/kernel/setup.c
===================================================================
@@ -114,6 +114,12 @@
 #include <asm/highmem.h>
 #endif
 
+#ifndef CONFIG_DEBUG_BOOT_PARAMS
+struct boot_params __initdata boot_params;
+#else
+struct boot_params boot_params;
+#endif
+
 /* This value is set up by the early boot code to point to the value
    immediately after the boot time page tables.  It contains a *physical*
    address, and must not be in the .bss segment! */

Index: linux-2.6/arch/x86/kernel/setup_percpu.c
===================================================================
@@ -15,12 +15,6 @@
 #include <asm/apicdef.h>
 #include <asm/highmem.h>
 
-#ifndef CONFIG_DEBUG_BOOT_PARAMS
-struct boot_params __initdata boot_params;
-#else
-struct boot_params boot_params;
-#endif
-
 #ifdef CONFIG_X86_LOCAL_APIC
 unsigned int num_processors;
 unsigned disabled_cpus __cpuinitdata;

Index: linux-2.6/arch/x86/kernel/setup.c
===================================================================
@@ -371,6 +371,31 @@ static void __init reserve_initrd(void)
 }
 #endif /* CONFIG_BLK_DEV_INITRD */
 
+void __init parse_setup_data(void)
+{
+	struct setup_data *data;
+	u64 pa_data;
+
+	if (boot_params.hdr.version < 0x0209)
+		return;
+	pa_data = boot_params.hdr.setup_data;
+	while (pa_data) {
+		data = early_ioremap(pa_data, PAGE_SIZE);
+		switch (data->type) {
+		case SETUP_E820_EXT:
+			parse_e820_ext(data, pa_data);
+			break;
+		default:
+			break;
+		}
+#ifndef CONFIG_DEBUG_BOOT_PARAMS
+		free_early(pa_data, pa_data+sizeof(*data)+data->len);
+#endif
+		pa_data = data->next;
+		early_iounmap(data, PAGE_SIZE);
+	}
+}
+
 /*
  * Determine if we were loaded by an EFI loader.  If so, then we have also been
  * passed the efi memmap, systab, etc., so we should use these data structures

Index: linux-2.6/arch/x86/kernel/setup_percpu.c
===================================================================
@@ -214,31 +214,6 @@ void __init setup_per_cpu_areas(void)
 
 #endif
 
-void __init parse_setup_data(void)
-{
-	struct setup_data *data;
-	u64 pa_data;
-
-	if (boot_params.hdr.version < 0x0209)
-		return;
-	pa_data = boot_params.hdr.setup_data;
-	while (pa_data) {
-		data = early_ioremap(pa_data, PAGE_SIZE);
-		switch (data->type) {
-		case SETUP_E820_EXT:
-			parse_e820_ext(data, pa_data);
-			break;
-		default:
-			break;
-		}
-#ifndef CONFIG_DEBUG_BOOT_PARAMS
-		free_early(pa_data, pa_data+sizeof(*data)+data->len);
-#endif
-		pa_data = data->next;
-		early_iounmap(data, PAGE_SIZE);
-	}
-}
-
 #ifdef X86_64_NUMA
 
 /*

Index: linux-2.6/arch/x86/kernel/setup.c
===================================================================
@@ -397,6 +397,64 @@ void __init parse_setup_data(void)
 }
 
 /*
+ * --------- Crashkernel reservation ------------------------------
+ */
+
+#ifdef CONFIG_KEXEC
+static inline unsigned long long get_total_mem(void)
+{
+	unsigned long long total;
+
+	total = max_low_pfn - min_low_pfn;
+#ifdef CONFIG_HIGHMEM
+	total += highend_pfn - highstart_pfn;
+#endif
+
+	return total << PAGE_SHIFT;
+}
+
+void __init reserve_crashkernel(void)
+{
+	unsigned long long total_mem;
+	unsigned long long crash_size, crash_base;
+	int ret;
+
+	total_mem = get_total_mem();
+
+	ret = parse_crashkernel(boot_command_line, total_mem,
+			&crash_size, &crash_base);
+	if (ret == 0 && crash_size > 0) {
+		if (crash_base <= 0) {
+			printk(KERN_INFO "crashkernel reservation failed - "
+					"you have to specify a base address\n");
+			return;
+		}
+
+		if (reserve_bootmem_generic(crash_base, crash_size,
+					BOOTMEM_EXCLUSIVE) < 0) {
+			printk(KERN_INFO "crashkernel reservation failed - "
+					"memory is in use\n");
+			return;
+		}
+
+		printk(KERN_INFO "Reserving %ldMB of memory at %ldMB "
+				"for crashkernel (System RAM: %ldMB)\n",
+				(unsigned long)(crash_size >> 20),
+				(unsigned long)(crash_base >> 20),
+				(unsigned long)(total_mem >> 20));
+
+		crashk_res.start = crash_base;
+		crashk_res.end   = crash_base + crash_size - 1;
+		insert_resource(&iomem_resource, &crashk_res);
+	}
+}
+#else
+void __init reserve_crashkernel(void)
+{
+}
+#endif
+
+/*
  * Determine if we were loaded by an EFI loader.  If so, then we have also been
  * passed the efi memmap, systab, etc., so we should use these data structures
  * for initialization.  Note, the efi init code path is determined by the

63 lines not on this page · raw message

Index: linux-2.6/arch/x86/kernel/setup.c
===================================================================
@@ -454,6 +454,39 @@ void __init reserve_crashkernel(void)
 }
 #endif
 
+static struct resource standard_io_resources[] = {
+	{ .name = "dma1", .start = 0x00, .end = 0x1f,
+		.flags = IORESOURCE_BUSY | IORESOURCE_IO },
+	{ .name = "pic1", .start = 0x20, .end = 0x21,
+		.flags = IORESOURCE_BUSY | IORESOURCE_IO },
+	{ .name = "timer0", .start = 0x40, .end = 0x43,
+		.flags = IORESOURCE_BUSY | IORESOURCE_IO },
+	{ .name = "timer1", .start = 0x50, .end = 0x53,
+		.flags = IORESOURCE_BUSY | IORESOURCE_IO },
+	{ .name = "keyboard", .start = 0x60, .end = 0x60,
+		.flags = IORESOURCE_BUSY | IORESOURCE_IO },
+	{ .name = "keyboard", .start = 0x64, .end = 0x64,
+		.flags = IORESOURCE_BUSY | IORESOURCE_IO },
+	{ .name = "dma page reg", .start = 0x80, .end = 0x8f,
+		.flags = IORESOURCE_BUSY | IORESOURCE_IO },
+	{ .name = "pic2", .start = 0xa0, .end = 0xa1,
+		.flags = IORESOURCE_BUSY | IORESOURCE_IO },
+	{ .name = "dma2", .start = 0xc0, .end = 0xdf,
+		.flags = IORESOURCE_BUSY | IORESOURCE_IO },
+	{ .name = "fpu", .start = 0xf0, .end = 0xff,
+		.flags = IORESOURCE_BUSY | IORESOURCE_IO }
+};
+
+void __init reserve_standard_io_resources(void)
+{
+	int i;
+
+	/* request I/O space for devices used on all i[345]86 PCs */
+	for (i = 0; i < ARRAY_SIZE(standard_io_resources); i++)
+		request_resource(&ioport_resource, &standard_io_resources[i]);
+
+}
+
 /*
  * Determine if we were loaded by an EFI loader.  If so, then we have also been
  * passed the efi memmap, systab, etc., so we should use these data structures

Index: linux-2.6/arch/x86/kernel/setup_percpu.c
===================================================================
@@ -387,38 +387,6 @@ EXPORT_SYMBOL(node_to_cpumask);
 
 #endif /* X86_64_NUMA */
 
-static struct resource standard_io_resources[] = {
-	{ .name = "dma1", .start = 0x00, .end = 0x1f,
-		.flags = IORESOURCE_BUSY | IORESOURCE_IO },
-	{ .name = "pic1", .start = 0x20, .end = 0x21,
-		.flags = IORESOURCE_BUSY | IORESOURCE_IO },
-	{ .name = "timer0", .start = 0x40, .end = 0x43,
-		.flags = IORESOURCE_BUSY | IORESOURCE_IO },
-	{ .name = "timer1", .start = 0x50, .end = 0x53,
-		.flags = IORESOURCE_BUSY | IORESOURCE_IO },
-	{ .name = "keyboard", .start = 0x60, .end = 0x60,
-		.flags = IORESOURCE_BUSY | IORESOURCE_IO },
-	{ .name = "keyboard", .start = 0x64, .end = 0x64,
-		.flags = IORESOURCE_BUSY | IORESOURCE_IO },
-	{ .name = "dma page reg", .start = 0x80, .end = 0x8f,
-		.flags = IORESOURCE_BUSY | IORESOURCE_IO },
-	{ .name = "pic2", .start = 0xa0, .end = 0xa1,
-		.flags = IORESOURCE_BUSY | IORESOURCE_IO },
-	{ .name = "dma2", .start = 0xc0, .end = 0xdf,
-		.flags = IORESOURCE_BUSY | IORESOURCE_IO },
-	{ .name = "fpu", .start = 0xf0, .end = 0xff,
-		.flags = IORESOURCE_BUSY | IORESOURCE_IO }
-};
-
-void __init reserve_standard_io_resources(void)
-{
-	int i;
-
-	/* request I/O space for devices used on all i[345]86 PCs */
-	for (i = 0; i < ARRAY_SIZE(standard_io_resources); i++)
-		request_resource(&ioport_resource, &standard_io_resources[i]);
-
-}
 
 #ifdef CONFIG_PROC_VMCORE
 /* elfcorehdr= specifies the location of elf core header

Index: linux-2.6/arch/x86/kernel/setup.c
===================================================================
@@ -487,6 +487,22 @@ void __init reserve_standard_io_resource
 
 }
 
+#ifdef CONFIG_PROC_VMCORE
+/* elfcorehdr= specifies the location of elf core header
+ * stored by the crashed kernel. This option will be passed
+ * by kexec loader to the capture kernel.
+ */
+static int __init setup_elfcorehdr(char *arg)
+{
+	char *end;
+	if (!arg)
+		return -EINVAL;
+	elfcorehdr_addr = memparse(arg, &end);
+	return end > arg ? 0 : -EINVAL;
+}
+early_param("elfcorehdr", setup_elfcorehdr);
+#endif
+
 /*
  * Determine if we were loaded by an EFI loader.  If so, then we have also been
  * passed the efi memmap, systab, etc., so we should use these data structures

Index: linux-2.6/arch/x86/kernel/setup_percpu.c
===================================================================
@@ -387,19 +387,3 @@ EXPORT_SYMBOL(node_to_cpumask);
 
 #endif /* X86_64_NUMA */
 
-
-#ifdef CONFIG_PROC_VMCORE
-/* elfcorehdr= specifies the location of elf core header
- * stored by the crashed kernel. This option will be passed
- * by kexec loader to the capture kernel.
- */
-static int __init setup_elfcorehdr(char *arg)
-{
-	char *end;
-	if (!arg)
-		return -EINVAL;
-	elfcorehdr_addr = memparse(arg, &end);
-	return end > arg ? 0 : -EINVAL;
-}
-early_param("elfcorehdr", setup_elfcorehdr);
-#endif

Index: linux-2.6/arch/x86/kernel/setup.c
===================================================================
@@ -713,8 +713,7 @@ void __init setup_arch(char **cmdline_p)
 	 */
 	acpi_reserve_bootmem();
 #endif
-#if defined(CONFIG_X86_FIND_SMP_CONFIG) && defined(CONFIG_X86_32) || \
-    defined(CONFIG_X86_MPPARSE) && defined(CONFIG_X86_64)
+#ifdef CONFIG_X86_FIND_SMP_CONFIG
 	/*
 	 * Find and reserve possible boot-time SMP configuration:
 	 */

Index: linux-2.6/arch/x86/kernel/setup.c
===================================================================
@@ -371,7 +371,7 @@ static void __init reserve_initrd(void)
 }
 #endif /* CONFIG_BLK_DEV_INITRD */
 
-void __init parse_setup_data(void)
+static void __init parse_setup_data(void)
 {
 	struct setup_data *data;
 	u64 pa_data;

@@ -413,7 +413,7 @@ static inline unsigned long long get_tot
 	return total << PAGE_SHIFT;
 }
 
-void __init reserve_crashkernel(void)
+static void __init reserve_crashkernel(void)
 {
 	unsigned long long total_mem;
 	unsigned long long crash_size, crash_base;

@@ -449,7 +449,7 @@ void __init reserve_crashkernel(void)
 	}
 }
 #else
-void __init reserve_crashkernel(void)
+static void __init reserve_crashkernel(void)
 {
 }
 #endif

@@ -477,7 +477,7 @@ static struct resource standard_io_resou
 		.flags = IORESOURCE_BUSY | IORESOURCE_IO }
 };
 
-void __init reserve_standard_io_resources(void)
+static void __init reserve_standard_io_resources(void)
 {
 	int i;
 

Index: linux-2.6/include/asm-x86/bootparam.h
===================================================================
@@ -109,6 +109,5 @@ struct boot_params {
 } __attribute__((packed));
 
 void reserve_setup_data(void);
-void parse_setup_data(void);
 
 #endif /* _ASM_BOOTPARAM_H */

Index: linux-2.6/include/asm-x86/setup.h
===================================================================
@@ -8,9 +8,6 @@
 /* Interrupt control for vSMPowered x86_64 systems */
 void vsmp_init(void);
 
-/* Crashkernel reservation */
-void reserve_crashkernel(void);
-
 #ifndef CONFIG_PARAVIRT
 #define paravirt_post_allocator_init()	do {} while (0)
 #endif

@@ -38,8 +35,6 @@ void reserve_crashkernel(void);
 #ifndef __ASSEMBLY__
 #include <asm/bootparam.h>
 
-void reserve_standard_io_resources(void);
-
 #ifndef _SETUP
 
 /*

Index: linux-2.6/arch/x86/kernel/setup.c
===================================================================
@@ -120,12 +120,6 @@ struct boot_params __initdata boot_param
 struct boot_params boot_params;
 #endif
 
-/* This value is set up by the early boot code to point to the value
-   immediately after the boot time page tables.  It contains a *physical*
-   address, and must not be in the .bss segment! */
-unsigned long init_pg_tables_start __initdata = ~0UL;
-unsigned long init_pg_tables_end __initdata = ~0UL;
-
 /*
  * Machine setup..
  */

@@ -152,6 +146,12 @@ static struct resource bss_resource = {
 
 
 #ifdef CONFIG_X86_32
+/* This value is set up by the early boot code to point to the value
+   immediately after the boot time page tables.  It contains a *physical*
+   address, and must not be in the .bss segment! */
+unsigned long init_pg_tables_start __initdata = ~0UL;
+unsigned long init_pg_tables_end __initdata = ~0UL;
+
 static struct resource video_ram_resource = {
 	.name	= "Video RAM area",
 	.start	= 0xa0000,

Index: linux-2.6/arch/x86/kernel/setup.c
===================================================================
@@ -100,11 +100,7 @@
 #include <asm/proto.h>
 
 #include <mach_apic.h>
-#ifdef CONFIG_PARAVIRT
 #include <asm/paravirt.h>
-#else
-#define ARCH_SETUP
-#endif
 
 #include <asm/percpu.h>
 #include <asm/sections.h>

@@ -114,6 +110,10 @@
 #include <asm/highmem.h>
 #endif
 
+#ifndef ARCH_SETUP
+#define ARCH_SETUP
+#endif
+
 #ifndef CONFIG_DEBUG_BOOT_PARAMS
 struct boot_params __initdata boot_params;
 #else

Index: linux-2.6/include/asm-x86/mach-default/setup_arch.h
===================================================================
@@ -1,7 +1,3 @@
 /* Hook to call BIOS initialisation function */
 
 /* no action for generic */
-
-#ifndef ARCH_SETUP
-#define ARCH_SETUP
-#endif

Index: linux-2.6/include/asm-x86/mach-visws/setup_arch.h
===================================================================
@@ -4,5 +4,3 @@ extern unsigned long sgivwfb_mem_phys;
 extern unsigned long sgivwfb_mem_size;
 
 /* no action for visws */
-
-#define ARCH_SETUP

Index: linux-2.6/arch/x86/kernel/setup.c
===================================================================
@@ -615,11 +615,6 @@ void __init setup_arch(char **cmdline_p)
 
 #ifdef CONFIG_X86_32
 	probe_roms();
-#else
-# ifdef CONFIG_PROVIDE_OHCI1394_DMA_INIT
-	if (init_ohci1394_dma_early)
-		init_ohci1394_dma_on_all_controllers();
-# endif
 #endif
 
 	/* after parse_early_param, so could debug it */

@@ -679,6 +674,15 @@ void __init setup_arch(char **cmdline_p)
 #endif
 	init_memory_mapping(0, (max_low_pfn << PAGE_SHIFT));
 
+	/*
+	 * NOTE: On x86-32, only from this point on, fixmaps are ready for use.
+	 */
+
+#ifdef CONFIG_PROVIDE_OHCI1394_DMA_INIT
+	if (init_ohci1394_dma_early)
+		init_ohci1394_dma_on_all_controllers();
+#endif
+
 	reserve_initrd();
 
 #ifdef CONFIG_X86_64

@@ -746,15 +750,6 @@ void __init setup_arch(char **cmdline_p)
 	map_vsyscall();
 #endif
 
-	/*
-	 * NOTE: On x86-32, only from this point on, fixmaps are ready for use.
-	 */
-
-#if defined(CONFIG_PROVIDE_OHCI1394_DMA_INIT) && defined(CONFIG_X86_32)
-	if (init_ohci1394_dma_early)
-		init_ohci1394_dma_on_all_controllers();
-#endif
-
 #ifdef CONFIG_X86_GENERICARCH
 	generic_apic_probe();
 #endif

Index: linux-2.6/arch/x86/mm/init_32.c
===================================================================
@@ -442,13 +442,10 @@ void __init native_pagetable_setup_done(
  * be partially populated, and so it avoids stomping on any existing
  * mappings.
  */
-static void __init pagetable_init(void)
+static void __init early_ioremap_page_table_range_init(pgd_t *pgd_base)
 {
-	pgd_t *pgd_base = swapper_pg_dir;
 	unsigned long vaddr, end;
 
-	paravirt_pagetable_setup_start(pgd_base);
-
 	/*
 	 * Fixed mappings, only the page table structure has to be
 	 * created - mappings will be set by set_fixmap():

@@ -458,6 +455,13 @@ static void __init pagetable_init(void)
 	end = (FIXADDR_TOP + PMD_SIZE - 1) & PMD_MASK;
 	page_table_range_init(vaddr, end, pgd_base);
 	early_ioremap_reset();
+}
+
+static void __init pagetable_init(void)
+{
+	pgd_t *pgd_base = swapper_pg_dir;
+
+	paravirt_pagetable_setup_start(pgd_base);
 
 	permanent_kmaps_init(pgd_base);
 

@@ -783,6 +787,8 @@ unsigned long __init_refok init_memory_m
 
 	kernel_physical_mapping_init(pgd_base, start, end);
 
+	early_ioremap_page_table_range_init(pgd_base);
+
 	load_cr3(swapper_pg_dir);
 
 	__flush_tlb_all();

@@ -794,6 +800,7 @@ unsigned long __init_refok init_memory_m
 	return end >> PAGE_SHIFT;
 }
 
+
 /*
  * paging_init() sets up the page tables - note that the first 8MB are
  * already mapped by head.S.

Index: linux-2.6/arch/x86/kernel/e820.c
===================================================================
@@ -828,16 +828,26 @@ void __init free_early(u64 start, u64 en
 
 void __init early_res_to_bootmem(u64 start, u64 end)
 {
-	int i;
+	int i, count;
 	u64 final_start, final_end;
-	for (i = 0; i < MAX_EARLY_RES && early_res[i].end; i++) {
+
+	count  = 0;
+	for (i = 0; i < MAX_EARLY_RES && early_res[i].end; i++)
+		count++;
+
+	printk(KERN_INFO "(%d early reservations):\n", count);
+	for (i = 0; i < count; i++) {
 		struct early_res *r = &early_res[i];
+		printk(KERN_INFO "  #%d [%08llx-%08llx] %16s", i,
+			r->start, r->end - 1, r->name);
 		final_start = max(start, r->start);
 		final_end = min(end, r->end);
-		if (final_start >= final_end)
+		if (final_start >= final_end) {
+			printk(KERN_INFO "\n");
 			continue;
-		printk(KERN_INFO "  early res: %d [%llx-%llx] %s\n", i,
-			final_start, final_end - 1, r->name);
+		}
+		printk(KERN_CONT " ==> [%08llx-%08llx] converted\n",
+			final_start, final_end - 1);
 		reserve_bootmem_generic(final_start, final_end - final_start,
 				BOOTMEM_DEFAULT);
 	}

Index: linux-2.6/arch/x86/kernel/apic_32.c
===================================================================
@@ -1197,36 +1197,6 @@ void __init init_apic_mappings(void)
 	if (boot_cpu_physical_apicid == -1U)
 		boot_cpu_physical_apicid = GET_APIC_ID(read_apic_id());
 
-#ifdef CONFIG_X86_IO_APIC
-	{
-		unsigned long ioapic_phys, idx = FIX_IO_APIC_BASE_0;
-		int i;
-
-		for (i = 0; i < nr_ioapics; i++) {
-			if (smp_found_config) {
-				ioapic_phys = mp_ioapics[i].mp_apicaddr;
-				if (!ioapic_phys) {
-					printk(KERN_ERR
-					       "WARNING: bogus zero IO-APIC "
-					       "address found in MPTABLE, "
-					       "disabling IO/APIC support!\n");
-					smp_found_config = 0;
-					skip_ioapic_setup = 1;
-					goto fake_ioapic_page;
-				}
-			} else {
-fake_ioapic_page:
-				ioapic_phys = (unsigned long)
-					      alloc_bootmem_pages(PAGE_SIZE);
-				ioapic_phys = __pa(ioapic_phys);
-			}
-			set_fixmap_nocache(idx, ioapic_phys);
-			printk(KERN_DEBUG "mapped IOAPIC to %08lx (%08lx)\n",
-			       __fix_to_virt(idx), ioapic_phys);
-			idx++;
-		}
-	}
-#endif
 }
 
 /*

Index: linux-2.6/arch/x86/kernel/io_apic_32.c
===================================================================
@@ -25,6 +25,7 @@
 #include <linux/init.h>
 #include <linux/delay.h>
 #include <linux/sched.h>
+#include <linux/bootmem.h>
 #include <linux/mc146818rtc.h>
 #include <linux/compiler.h>
 #include <linux/acpi.h>

@@ -2854,3 +2855,34 @@ static int __init parse_noapic(char *arg
 	return 0;
 }
 early_param("noapic", parse_noapic);
+
+void __init ioapic_init_mappings(void)
+{
+	unsigned long ioapic_phys, idx = FIX_IO_APIC_BASE_0;
+	int i;
+
+	for (i = 0; i < nr_ioapics; i++) {
+		if (smp_found_config) {
+			ioapic_phys = mp_ioapics[i].mp_apicaddr;
+			if (!ioapic_phys) {
+				printk(KERN_ERR
+				       "WARNING: bogus zero IO-APIC "
+				       "address found in MPTABLE, "
+				       "disabling IO/APIC support!\n");
+				smp_found_config = 0;
+				skip_ioapic_setup = 1;
+				goto fake_ioapic_page;
+			}
+		} else {
+fake_ioapic_page:
+			ioapic_phys = (unsigned long)
+				      alloc_bootmem_pages(PAGE_SIZE);
+			ioapic_phys = __pa(ioapic_phys);
+		}
+		set_fixmap_nocache(idx, ioapic_phys);
+		printk(KERN_DEBUG "mapped IOAPIC to %08lx (%08lx)\n",
+		       __fix_to_virt(idx), ioapic_phys);
+		idx++;
+	}
+}
+

Index: linux-2.6/arch/x86/kernel/setup.c
===================================================================
@@ -771,16 +771,14 @@ void __init setup_arch(char **cmdline_p)
 		get_smp_config();
 #endif
 
-#ifdef CONFIG_X86_64
 	init_apic_mappings();
 	ioapic_init_mappings();
-#else
-# if defined(CONFIG_SMP) && defined(CONFIG_X86_PC)
+
+#if defined(CONFIG_SMP) && defined(CONFIG_X86_PC) && defined(CONFIG_X86_32)
 	if (def_to_bigsmp)
 		printk(KERN_WARNING "More than 8 CPUs detected and "
 			"CONFIG_X86_PC cannot handle it.\nUse "
 			"CONFIG_X86_GENERICARCH or CONFIG_X86_BIGSMP.\n");
-# endif
 #endif
 	kvm_guest_init();
 

Index: linux-2.6/arch/x86/kernel/traps_32.c
===================================================================
@@ -1210,10 +1210,6 @@ void __init trap_init(void)
 	early_iounmap(p, 4);
 #endif
 
-#ifdef CONFIG_X86_LOCAL_APIC
-	init_apic_mappings();
-#endif
-	set_trap_gate(0,  &divide_error);
 	set_intr_gate(1,  &debug);
 	set_intr_gate(2,  &nmi);
 	set_system_intr_gate(3, &int3); /* int3/4 can be called from all */

Index: linux-2.6/include/asm-x86/apic.h
===================================================================
@@ -135,6 +135,7 @@ extern int apic_is_clustered_box(void);
 #else /* !CONFIG_X86_LOCAL_APIC */
 static inline void lapic_shutdown(void) { }
 #define local_apic_timer_c2_ok		1
+static inline void init_apic_mappings(void) { }
 
 #endif /* !CONFIG_X86_LOCAL_APIC */
 

Index: linux-2.6/include/asm-x86/io_apic.h
===================================================================
@@ -196,6 +196,7 @@ extern void ioapic_init_mappings(void);
 #else  /* !CONFIG_X86_IO_APIC */
 #define io_apic_assign_pci_irqs 0
 static const int timer_through_8259 = 0;
+static inline void ioapic_init_mappings(void) { }
 #endif
 
 #endif

Index: linux-2.6/arch/x86/kernel/e820.c
===================================================================
@@ -828,16 +828,26 @@ void __init free_early(u64 start, u64 en
 
 void __init early_res_to_bootmem(u64 start, u64 end)
 {
-	int i;
+	int i, count;
 	u64 final_start, final_end;
-	for (i = 0; i < MAX_EARLY_RES && early_res[i].end; i++) {
+
+	count  = 0;
+	for (i = 0; i < MAX_EARLY_RES && early_res[i].end; i++)
+		count++;
+
+	printk(KERN_INFO "(%d early reservations) ==> bootmem\n", count);
+	for (i = 0; i < count; i++) {
 		struct early_res *r = &early_res[i];
+		printk(KERN_INFO "  #%d %010llx - %010llx %16s", i,
+			r->start, r->end, r->name);
 		final_start = max(start, r->start);
 		final_end = min(end, r->end);
-		if (final_start >= final_end)
+		if (final_start >= final_end) {
+			printk(KERN_CONT "\n");
 			continue;
-		printk(KERN_INFO "  early res: %d [%llx-%llx] %s\n", i,
-			final_start, final_end - 1, r->name);
+		}
+		printk(KERN_CONT " ==> %010llx - %010llx\n",
+			final_start, final_end);
 		reserve_bootmem_generic(final_start, final_end - final_start,
 				BOOTMEM_DEFAULT);
 	}

Index: linux-2.6/arch/x86/mm/init_32.c
===================================================================
@@ -184,8 +184,9 @@ static inline int is_kernel_text(unsigne
  * PAGE_OFFSET:
  */
 static void __init kernel_physical_mapping_init(pgd_t *pgd_base,
-						unsigned long start,
-						unsigned long end)
+						unsigned long start_pfn,
+						unsigned long end_pfn,
+						int use_pse)
 {
 	int pgd_idx, pmd_idx, pte_ofs;
 	unsigned long pfn;

@@ -193,32 +194,30 @@ static void __init kernel_physical_mappi
 	pmd_t *pmd;
 	pte_t *pte;
 	unsigned pages_2m = 0, pages_4k = 0;
-	unsigned limit_pfn = end >> PAGE_SHIFT;
 
-	pgd_idx = pgd_index(PAGE_OFFSET);
-	pgd = pgd_base + pgd_idx;
-	pfn = start >> PAGE_SHIFT;
+	if (!cpu_has_pse)
+		use_pse = 0;
 
+	pfn = start_pfn;
+	pgd_idx = pgd_index((pfn<<PAGE_SHIFT) + PAGE_OFFSET);
+	pgd = pgd_base + pgd_idx;
 	for (; pgd_idx < PTRS_PER_PGD; pgd++, pgd_idx++) {
 		pmd = one_md_table_init(pgd);
-		if (pfn >= limit_pfn)
+
+		if (pfn >= end_pfn)
 			continue;
 
-		for (pmd_idx = 0;
-		     pmd_idx < PTRS_PER_PMD && pfn < limit_pfn;
+		pmd_idx = pmd_index((pfn<<PAGE_SHIFT) + PAGE_OFFSET);
+		pmd += pmd_idx;
+		for (; pmd_idx < PTRS_PER_PMD && pfn < end_pfn;
 		     pmd++, pmd_idx++) {
 			unsigned int addr = pfn * PAGE_SIZE + PAGE_OFFSET;
 
 			/*
 			 * Map with big pages if possible, otherwise
 			 * create normal page tables:
-			 *
-			 * Don't use a large page for the first 2/4MB of memory
-			 * because there are often fixed size MTRRs in there
-			 * and overlapping MTRRs into large pages can cause
-			 * slowdowns.
 			 */
-			if (cpu_has_pse && !(pgd_idx == 0 && pmd_idx == 0)) {
+			if (use_pse) {
 				unsigned int addr2;
 				pgprot_t prot = PAGE_KERNEL_LARGE;
 

@@ -233,13 +232,12 @@ static void __init kernel_physical_mappi
 				set_pmd(pmd, pfn_pmd(pfn, prot));
 
 				pfn += PTRS_PER_PTE;
-				max_pfn_mapped = pfn;
 				continue;
 			}
 			pte = one_page_table_init(pmd);
 
 			for (pte_ofs = 0;
-			     pte_ofs < PTRS_PER_PTE && pfn < max_low_pfn;
+			     pte_ofs < PTRS_PER_PTE && pfn < end_pfn;
 			     pte++, pfn++, pte_ofs++, addr += PAGE_SIZE) {
 				pgprot_t prot = PAGE_KERNEL;
 

@@ -249,7 +247,6 @@ static void __init kernel_physical_mappi
 				pages_4k++;
 				set_pte(pte, pfn_pte(pfn, prot));
 			}
-			max_pfn_mapped = pfn;
 		}
 	}
 	update_page_count(PG_LEVEL_2M, pages_2m);

@@ -729,7 +726,7 @@ void __init setup_bootmem_allocator(void
 
 static void __init find_early_table_space(unsigned long end)
 {
-	unsigned long puds, pmds, tables, start;
+	unsigned long puds, pmds, ptes, tables, start;
 
 	puds = (end + PUD_SIZE - 1) >> PUD_SHIFT;
 	tables = PAGE_ALIGN(puds * sizeof(pud_t));

@@ -737,10 +734,15 @@ static void __init find_early_table_spac
 	pmds = (end + PMD_SIZE - 1) >> PMD_SHIFT;
 	tables += PAGE_ALIGN(pmds * sizeof(pmd_t));
 
-	if (!cpu_has_pse) {
-		int ptes = (end + PAGE_SIZE - 1) >> PAGE_SHIFT;
-		tables += PAGE_ALIGN(ptes * sizeof(pte_t));
-	}
+	if (cpu_has_pse) {
+		int extra;
+		extra = end - ((end>>21) << 21);
+		extra += (2<<20);
+		ptes = (extra + PAGE_SIZE - 1) >> PAGE_SHIFT;
+	} else
+		ptes = (end + PAGE_SIZE - 1) >> PAGE_SHIFT;
+
+	tables += PAGE_ALIGN(ptes * sizeof(pte_t));
 
 	/*
 	 * RED-PEN putting page tables only on node 0 could

@@ -766,6 +768,7 @@ unsigned long __init_refok init_memory_m
 						unsigned long end)
 {
 	pgd_t *pgd_base = swapper_pg_dir;
+	unsigned long start_pfn, end_pfn;
 
 	/*
 	 * Find space for the kernel direct mapping tables.

@@ -790,7 +793,30 @@ unsigned long __init_refok init_memory_m
 		__PAGE_KERNEL_EXEC |= _PAGE_GLOBAL;
 	}
 
-	kernel_physical_mapping_init(pgd_base, start, end);
+	/*
+	 * Don't use a large page for the first 2/4MB of memory
+	 * because there are often fixed size MTRRs in there
+	 * and overlapping MTRRs into large pages can cause
+	 * slowdowns.
+	 */
+	if (start < (2<<20)) {
+		start_pfn = start >> PAGE_SHIFT;
+		end_pfn = min(2UL<<(20 - PAGE_SHIFT), end>>PAGE_SHIFT);
+		kernel_physical_mapping_init(pgd_base, start_pfn, end_pfn, 0);
+	}
+
+	if (end >= (2<<20)) {
+		start_pfn = 2 << (20 - PAGE_SHIFT);
+		end_pfn = (end >> 21) << (21 - PAGE_SHIFT);
+		kernel_physical_mapping_init(pgd_base, start_pfn, end_pfn,
+						cpu_has_pse);
+
+		start_pfn = end_pfn;
+		end_pfn = end >> PAGE_SHIFT;
+		if (end_pfn > start_pfn)
+			kernel_physical_mapping_init(pgd_base, start_pfn,
+							end_pfn, 0);
+	}
 
 	early_ioremap_page_table_range_init(pgd_base);
 

Index: linux-2.6/arch/x86/mm/init_32.c
===================================================================
@@ -184,8 +184,9 @@ static inline int is_kernel_text(unsigne
  * PAGE_OFFSET:
  */
 static void __init kernel_physical_mapping_init(pgd_t *pgd_base,
-						unsigned long start,
-						unsigned long end)
+						unsigned long start_pfn,
+						unsigned long end_pfn,
+						int use_pse)
 {
 	int pgd_idx, pmd_idx, pte_ofs;
 	unsigned long pfn;

@@ -193,32 +194,30 @@ static void __init kernel_physical_mappi
 	pmd_t *pmd;
 	pte_t *pte;
 	unsigned pages_2m = 0, pages_4k = 0;
-	unsigned limit_pfn = end >> PAGE_SHIFT;
 
-	pgd_idx = pgd_index(PAGE_OFFSET);
-	pgd = pgd_base + pgd_idx;
-	pfn = start >> PAGE_SHIFT;
+	if (!cpu_has_pse)
+		use_pse = 0;
 
+	pfn = start_pfn;
+	pgd_idx = pgd_index((pfn<<PAGE_SHIFT) + PAGE_OFFSET);
+	pgd = pgd_base + pgd_idx;
 	for (; pgd_idx < PTRS_PER_PGD; pgd++, pgd_idx++) {
 		pmd = one_md_table_init(pgd);
-		if (pfn >= limit_pfn)
+
+		if (pfn >= end_pfn)
 			continue;
 
-		for (pmd_idx = 0;
-		     pmd_idx < PTRS_PER_PMD && pfn < limit_pfn;
+		pmd_idx = pmd_index((pfn<<PAGE_SHIFT) + PAGE_OFFSET);
+		pmd += pmd_idx;
+		for (; pmd_idx < PTRS_PER_PMD && pfn < end_pfn;
 		     pmd++, pmd_idx++) {
 			unsigned int addr = pfn * PAGE_SIZE + PAGE_OFFSET;
 
 			/*
 			 * Map with big pages if possible, otherwise
 			 * create normal page tables:
-			 *
-			 * Don't use a large page for the first 2/4MB of memory
-			 * because there are often fixed size MTRRs in there
-			 * and overlapping MTRRs into large pages can cause
-			 * slowdowns.
 			 */
-			if (cpu_has_pse && !(pgd_idx == 0 && pmd_idx == 0)) {
+			if (use_pse) {
 				unsigned int addr2;
 				pgprot_t prot = PAGE_KERNEL_LARGE;
 

@@ -233,13 +232,12 @@ static void __init kernel_physical_mappi
 				set_pmd(pmd, pfn_pmd(pfn, prot));
 
 				pfn += PTRS_PER_PTE;
-				max_pfn_mapped = pfn;
 				continue;
 			}
 			pte = one_page_table_init(pmd);
 
 			for (pte_ofs = 0;
-			     pte_ofs < PTRS_PER_PTE && pfn < max_low_pfn;
+			     pte_ofs < PTRS_PER_PTE && pfn < end_pfn;
 			     pte++, pfn++, pte_ofs++, addr += PAGE_SIZE) {
 				pgprot_t prot = PAGE_KERNEL;
 

@@ -249,7 +247,6 @@ static void __init kernel_physical_mappi
 				pages_4k++;
 				set_pte(pte, pfn_pte(pfn, prot));
 			}
-			max_pfn_mapped = pfn;
 		}
 	}
 	update_page_count(PG_LEVEL_2M, pages_2m);

@@ -729,7 +726,7 @@ void __init setup_bootmem_allocator(void
 
 static void __init find_early_table_space(unsigned long end)
 {
-	unsigned long puds, pmds, tables, start;
+	unsigned long puds, pmds, ptes, tables, start;
 
 	puds = (end + PUD_SIZE - 1) >> PUD_SHIFT;
 	tables = PAGE_ALIGN(puds * sizeof(pud_t));

@@ -737,10 +734,15 @@ static void __init find_early_table_spac
 	pmds = (end + PMD_SIZE - 1) >> PMD_SHIFT;
 	tables += PAGE_ALIGN(pmds * sizeof(pmd_t));
 
-	if (!cpu_has_pse) {
-		int ptes = (end + PAGE_SIZE - 1) >> PAGE_SHIFT;
-		tables += PAGE_ALIGN(ptes * sizeof(pte_t));
-	}
+	if (cpu_has_pse) {
+		int extra;
+		extra = end - ((end>>21) << 21);
+		extra += (2<<20);
+		ptes = (extra + PAGE_SIZE - 1) >> PAGE_SHIFT;
+	} else
+		ptes = (end + PAGE_SIZE - 1) >> PAGE_SHIFT;
+
+	tables += PAGE_ALIGN(ptes * sizeof(pte_t));
 
 	/*
 	 * RED-PEN putting page tables only on node 0 could

@@ -766,6 +768,8 @@ unsigned long __init_refok init_memory_m
 						unsigned long end)
 {
 	pgd_t *pgd_base = swapper_pg_dir;
+	unsigned long start_pfn, end_pfn;
+	unsigned long big_page_start;
 
 	/*
 	 * Find space for the kernel direct mapping tables.

@@ -790,7 +794,41 @@ unsigned long __init_refok init_memory_m
 		__PAGE_KERNEL_EXEC |= _PAGE_GLOBAL;
 	}
 
-	kernel_physical_mapping_init(pgd_base, start, end);
+	/*
+	 * Don't use a large page for the first 2/4MB of memory
+	 * because there are often fixed size MTRRs in there
+	 * and overlapping MTRRs into large pages can cause
+	 * slowdowns.
+	 */
+	big_page_start = 2UL<<20;
+	if (start < big_page_start) {
+		start_pfn = start >> PAGE_SHIFT;
+		end_pfn = min(big_page_start>>PAGE_SHIFT, end>>PAGE_SHIFT);
+	} else {
+		/* head is not 2M alignment ? */
+		start_pfn = (start >> 21) << (21 - PAGE_SHIFT);
+		end_pfn = start >> PAGE_SHIFT;
+	}
+	if (start_pfn < end_pfn)
+		kernel_physical_mapping_init(pgd_base, start_pfn, end_pfn, 0);
+
+	/* big page range */
+	start_pfn = ((start + ((2UL<<20) - 1)) >> 21) << (21 - PAGE_SHIFT);
+	if (start_pfn < (big_page_start >> PAGE_SHIFT))
+		start_pfn =  big_page_start >> PAGE_SHIFT;
+	end_pfn = (end >> 21) << (21 - PAGE_SHIFT);
+	if (start_pfn < end_pfn)
+		kernel_physical_mapping_init(pgd_base, start_pfn, end_pfn,
+						cpu_has_pse);
+
+	/* tail is not 2M alignment ? */
+	start_pfn = end_pfn;
+	if (start_pfn > (big_page_start>>PAGE_SHIFT)) {
+		end_pfn = end >> PAGE_SHIFT;
+		if (start_pfn < end_pfn)
+			kernel_physical_mapping_init(pgd_base, start_pfn,
+							 end_pfn, 0);
+	}
 
 	early_ioremap_page_table_range_init(pgd_base);
 

Index: linux-2.6/arch/x86/mm/init_32.c
===================================================================
@@ -195,7 +195,7 @@ static void __init kernel_physical_mappi
 	unsigned pages_2m = 0, pages_4k = 0;
 	unsigned limit_pfn = end >> PAGE_SHIFT;
 
-	pgd_idx = pgd_index(PAGE_OFFSET);
+	pgd_idx = pgd_index(start + PAGE_OFFSET);
 	pgd = pgd_base + pgd_idx;
 	pfn = start >> PAGE_SHIFT;
 

@@ -218,7 +218,8 @@ static void __init kernel_physical_mappi
 			 * and overlapping MTRRs into large pages can cause
 			 * slowdowns.
 			 */
-			if (cpu_has_pse && !(pgd_idx == 0 && pmd_idx == 0)) {
+			if (cpu_has_pse && !(pgd_idx == 0 && pmd_idx == 0) &&
+				(pfn + PTRS_PER_PTE) <= limit_pfn) {
 				unsigned int addr2;
 				pgprot_t prot = PAGE_KERNEL_LARGE;
 

@@ -233,13 +234,12 @@ static void __init kernel_physical_mappi
 				set_pmd(pmd, pfn_pmd(pfn, prot));
 
 				pfn += PTRS_PER_PTE;
-				max_pfn_mapped = pfn;
 				continue;
 			}
 			pte = one_page_table_init(pmd);
 
 			for (pte_ofs = 0;
-			     pte_ofs < PTRS_PER_PTE && pfn < max_low_pfn;
+			     pte_ofs < PTRS_PER_PTE && pfn < limit_pfn;
 			     pte++, pfn++, pte_ofs++, addr += PAGE_SIZE) {
 				pgprot_t prot = PAGE_KERNEL;
 

@@ -249,7 +249,6 @@ static void __init kernel_physical_mappi
 				pages_4k++;
 				set_pte(pte, pfn_pte(pfn, prot));
 			}
-			max_pfn_mapped = pfn;
 		}
 	}
 	update_page_count(PG_LEVEL_2M, pages_2m);

@@ -729,7 +728,7 @@ void __init setup_bootmem_allocator(void
 
 static void __init find_early_table_space(unsigned long end)
 {
-	unsigned long puds, pmds, tables, start;
+	unsigned long puds, pmds, ptes, tables, start;
 
 	puds = (end + PUD_SIZE - 1) >> PUD_SHIFT;
 	tables = PAGE_ALIGN(puds * sizeof(pud_t));

@@ -737,10 +736,15 @@ static void __init find_early_table_spac
 	pmds = (end + PMD_SIZE - 1) >> PMD_SHIFT;
 	tables += PAGE_ALIGN(pmds * sizeof(pmd_t));
 
-	if (!cpu_has_pse) {
-		int ptes = (end + PAGE_SIZE - 1) >> PAGE_SHIFT;
-		tables += PAGE_ALIGN(ptes * sizeof(pte_t));
-	}
+	if (cpu_has_pse) {
+		unsigned long extra;
+		extra = end - ((end>>21) << 21);
+		extra += (2UL<<20);
+		ptes = (extra + PAGE_SIZE - 1) >> PAGE_SHIFT;
+	} else
+		ptes = (end + PAGE_SIZE - 1) >> PAGE_SHIFT;
+
+	tables += PAGE_ALIGN(ptes * sizeof(pte_t));
 
 	/*
 	 * RED-PEN putting page tables only on node 0 could

Index: linux-2.6/arch/x86/mm/init_32.c
===================================================================
@@ -184,8 +184,9 @@ static inline int is_kernel_text(unsigne
  * PAGE_OFFSET:
  */
 static void __init kernel_physical_mapping_init(pgd_t *pgd_base,
-						unsigned long start,
-						unsigned long end)
+						unsigned long start_pfn,
+						unsigned long end_pfn,
+						int use_pse)
 {
 	int pgd_idx, pmd_idx, pte_ofs;
 	unsigned long pfn;

@@ -193,32 +194,33 @@ static void __init kernel_physical_mappi
 	pmd_t *pmd;
 	pte_t *pte;
 	unsigned pages_2m = 0, pages_4k = 0;
-	unsigned limit_pfn = end >> PAGE_SHIFT;
 
-	pgd_idx = pgd_index(PAGE_OFFSET);
-	pgd = pgd_base + pgd_idx;
-	pfn = start >> PAGE_SHIFT;
+	if (!cpu_has_pse)
+		use_pse = 0;
 
+	pfn = start_pfn;
+	pgd_idx = pgd_index((pfn<<PAGE_SHIFT) + PAGE_OFFSET);
+	pgd = pgd_base + pgd_idx;
 	for (; pgd_idx < PTRS_PER_PGD; pgd++, pgd_idx++) {
 		pmd = one_md_table_init(pgd);
-		if (pfn >= limit_pfn)
-			continue;
 
-		for (pmd_idx = 0;
-		     pmd_idx < PTRS_PER_PMD && pfn < limit_pfn;
+		if (pfn >= end_pfn)
+			continue;
+#ifdef CONFIG_X86_PAE
+		pmd_idx = pmd_index((pfn<<PAGE_SHIFT) + PAGE_OFFSET);
+		pmd += pmd_idx;
+#else
+		pmd_idx = 0;
+#endif
+		for (; pmd_idx < PTRS_PER_PMD && pfn < end_pfn;
 		     pmd++, pmd_idx++) {
 			unsigned int addr = pfn * PAGE_SIZE + PAGE_OFFSET;
 
 			/*
 			 * Map with big pages if possible, otherwise
 			 * create normal page tables:
-			 *
-			 * Don't use a large page for the first 2/4MB of memory
-			 * because there are often fixed size MTRRs in there
-			 * and overlapping MTRRs into large pages can cause
-			 * slowdowns.
 			 */
-			if (cpu_has_pse && !(pgd_idx == 0 && pmd_idx == 0)) {
+			if (use_pse) {
 				unsigned int addr2;
 				pgprot_t prot = PAGE_KERNEL_LARGE;
 

@@ -233,13 +235,13 @@ static void __init kernel_physical_mappi
 				set_pmd(pmd, pfn_pmd(pfn, prot));
 
 				pfn += PTRS_PER_PTE;
-				max_pfn_mapped = pfn;
 				continue;
 			}
 			pte = one_page_table_init(pmd);
 
-			for (pte_ofs = 0;
-			     pte_ofs < PTRS_PER_PTE && pfn < max_low_pfn;
+			pte_ofs = pte_index((pfn<<PAGE_SHIFT) + PAGE_OFFSET);
+			pte += pte_ofs;
+			for (; pte_ofs < PTRS_PER_PTE && pfn < end_pfn;
 			     pte++, pfn++, pte_ofs++, addr += PAGE_SIZE) {
 				pgprot_t prot = PAGE_KERNEL;
 

@@ -249,7 +251,6 @@ static void __init kernel_physical_mappi
 				pages_4k++;
 				set_pte(pte, pfn_pte(pfn, prot));
 			}
-			max_pfn_mapped = pfn;
 		}
 	}
 	update_page_count(PG_LEVEL_2M, pages_2m);

@@ -729,7 +730,7 @@ void __init setup_bootmem_allocator(void
 
 static void __init find_early_table_space(unsigned long end)
 {
-	unsigned long puds, pmds, tables, start;
+	unsigned long puds, pmds, ptes, tables, start;
 
 	puds = (end + PUD_SIZE - 1) >> PUD_SHIFT;
 	tables = PAGE_ALIGN(puds * sizeof(pud_t));

@@ -737,10 +738,19 @@ static void __init find_early_table_spac
 	pmds = (end + PMD_SIZE - 1) >> PMD_SHIFT;
 	tables += PAGE_ALIGN(pmds * sizeof(pmd_t));
 
-	if (!cpu_has_pse) {
-		int ptes = (end + PAGE_SIZE - 1) >> PAGE_SHIFT;
-		tables += PAGE_ALIGN(ptes * sizeof(pte_t));
-	}
+	if (cpu_has_pse) {
+		unsigned long extra;
+
+		extra = end - ((end>>PMD_SHIFT) << PMD_SHIFT);
+		extra += PMD_SIZE;
+		ptes = (extra + PAGE_SIZE - 1) >> PAGE_SHIFT;
+	} else
+		ptes = (end + PAGE_SIZE - 1) >> PAGE_SHIFT;
+
+	tables += PAGE_ALIGN(ptes * sizeof(pte_t));
+
+	/* for fixmap */
+	tables += PAGE_SIZE * 2;
 
 	/*
 	 * RED-PEN putting page tables only on node 0 could

@@ -766,6 +776,8 @@ unsigned long __init_refok init_memory_m
 						unsigned long end)
 {
 	pgd_t *pgd_base = swapper_pg_dir;
+	unsigned long start_pfn, end_pfn;
+	unsigned long big_page_start;
 
 	/*
 	 * Find space for the kernel direct mapping tables.

@@ -790,7 +802,44 @@ unsigned long __init_refok init_memory_m
 		__PAGE_KERNEL_EXEC |= _PAGE_GLOBAL;
 	}
 
-	kernel_physical_mapping_init(pgd_base, start, end);
+	/*
+	 * Don't use a large page for the first 2/4MB of memory
+	 * because there are often fixed size MTRRs in there
+	 * and overlapping MTRRs into large pages can cause
+	 * slowdowns.
+	 */
+	big_page_start = PMD_SIZE;
+
+	if (start < big_page_start) {
+		start_pfn = start >> PAGE_SHIFT;
+		end_pfn = min(big_page_start>>PAGE_SHIFT, end>>PAGE_SHIFT);
+	} else {
+		/* head is not big page alignment ? */
+		start_pfn = start >> PAGE_SHIFT;
+		end_pfn = ((start + PMD_SIZE - 1)>>PMD_SHIFT)
+				 << (PMD_SHIFT - PAGE_SHIFT);
+	}
+	if (start_pfn < end_pfn)
+		kernel_physical_mapping_init(pgd_base, start_pfn, end_pfn, 0);
+
+	/* big page range */
+	start_pfn = ((start + PMD_SIZE - 1)>>PMD_SHIFT)
+			 << (PMD_SHIFT - PAGE_SHIFT);
+	if (start_pfn < (big_page_start >> PAGE_SHIFT))
+		start_pfn =  big_page_start >> PAGE_SHIFT;
+	end_pfn = (end>>PMD_SHIFT) << (PMD_SHIFT - PAGE_SHIFT);
+	if (start_pfn < end_pfn)
+		kernel_physical_mapping_init(pgd_base, start_pfn, end_pfn,
+						cpu_has_pse);
+
+	/* tail is not big page alignment ? */
+	start_pfn = end_pfn;
+	if (start_pfn > (big_page_start>>PAGE_SHIFT)) {
+		end_pfn = end >> PAGE_SHIFT;
+		if (start_pfn < end_pfn)
+			kernel_physical_mapping_init(pgd_base, start_pfn,
+							 end_pfn, 0);
+	}
 
 	early_ioremap_page_table_range_init(pgd_base);
 

Index: linux-2.6/arch/x86/mm/init_32.c
===================================================================
@@ -184,8 +184,9 @@ static inline int is_kernel_text(unsigne
  * PAGE_OFFSET:
  */
 static void __init kernel_physical_mapping_init(pgd_t *pgd_base,
-						unsigned long start,
-						unsigned long end)
+						unsigned long start_pfn,
+						unsigned long end_pfn,
+						int use_pse)
 {
 	int pgd_idx, pmd_idx, pte_ofs;
 	unsigned long pfn;

@@ -193,33 +194,33 @@ static void __init kernel_physical_mappi
 	pmd_t *pmd;
 	pte_t *pte;
 	unsigned pages_2m = 0, pages_4k = 0;
-	unsigned limit_pfn = end >> PAGE_SHIFT;
 
-	pgd_idx = pgd_index(start + PAGE_OFFSET);
-	pgd = pgd_base + pgd_idx;
-	pfn = start >> PAGE_SHIFT;
+	if (!cpu_has_pse)
+		use_pse = 0;
 
+	pfn = start_pfn;
+	pgd_idx = pgd_index((pfn<<PAGE_SHIFT) + PAGE_OFFSET);
+	pgd = pgd_base + pgd_idx;
 	for (; pgd_idx < PTRS_PER_PGD; pgd++, pgd_idx++) {
 		pmd = one_md_table_init(pgd);
-		if (pfn >= limit_pfn)
-			continue;
 
-		for (pmd_idx = 0;
-		     pmd_idx < PTRS_PER_PMD && pfn < limit_pfn;
+		if (pfn >= end_pfn)
+			continue;
+#ifdef CONFIG_X86_PAE
+		pmd_idx = pmd_index((pfn<<PAGE_SHIFT) + PAGE_OFFSET);
+		pmd += pmd_idx;
+#else
+		pmd_idx = 0;
+#endif
+		for (; pmd_idx < PTRS_PER_PMD && pfn < end_pfn;
 		     pmd++, pmd_idx++) {
 			unsigned int addr = pfn * PAGE_SIZE + PAGE_OFFSET;
 
 			/*
 			 * Map with big pages if possible, otherwise
 			 * create normal page tables:
-			 *
-			 * Don't use a large page for the first 2/4MB of memory
-			 * because there are often fixed size MTRRs in there
-			 * and overlapping MTRRs into large pages can cause
-			 * slowdowns.
 			 */
-			if (cpu_has_pse && !(pgd_idx == 0 && pmd_idx == 0) &&
-				(pfn + PTRS_PER_PTE) <= limit_pfn) {
+			if (use_pse) {
 				unsigned int addr2;
 				pgprot_t prot = PAGE_KERNEL_LARGE;
 

@@ -238,8 +239,9 @@ static void __init kernel_physical_mappi
 			}
 			pte = one_page_table_init(pmd);
 
-			for (pte_ofs = 0;
-			     pte_ofs < PTRS_PER_PTE && pfn < limit_pfn;
+			pte_ofs = pte_index((pfn<<PAGE_SHIFT) + PAGE_OFFSET);
+			pte += pte_ofs;
+			for (; pte_ofs < PTRS_PER_PTE && pfn < end_pfn;
 			     pte++, pfn++, pte_ofs++, addr += PAGE_SIZE) {
 				pgprot_t prot = PAGE_KERNEL;
 

@@ -738,14 +740,18 @@ static void __init find_early_table_spac
 
 	if (cpu_has_pse) {
 		unsigned long extra;
-		extra = end - ((end>>21) << 21);
-		extra += (2UL<<20);
+
+		extra = end - ((end>>PMD_SHIFT) << PMD_SHIFT);
+		extra += PMD_SIZE;
 		ptes = (extra + PAGE_SIZE - 1) >> PAGE_SHIFT;
 	} else
 		ptes = (end + PAGE_SIZE - 1) >> PAGE_SHIFT;
 
 	tables += PAGE_ALIGN(ptes * sizeof(pte_t));
 
+	/* for fixmap */
+	tables += PAGE_SIZE * 2;
+
 	/*
 	 * RED-PEN putting page tables only on node 0 could
 	 * cause a hotspot and fill up ZONE_DMA. The page tables

@@ -770,6 +776,8 @@ unsigned long __init_refok init_memory_m
 						unsigned long end)
 {
 	pgd_t *pgd_base = swapper_pg_dir;
+	unsigned long start_pfn, end_pfn;
+	unsigned long big_page_start;
 
 	/*
 	 * Find space for the kernel direct mapping tables.