[PATCH 00/16] x86: merge setup_32/64.c

[PATCH 00/16] x86: merge setup_32/64.c

[Yinghai Lu]

please check the patches

it will rename setup.c to setup_percpu.c and make setup_32.c to support 64bit, and then change setup_32.c to setup.c 
and delete setup_64.c.

Thanks

Yinghai Lu

[PATCH 03/16] x86: update reserve_initrd to support 64bit

[Yinghai Lu]

Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>

Index: linux-2.6/arch/x86/kernel/setup_32.c
===================================================================
--- linux-2.6.orig/arch/x86/kernel/setup_32.c
+++ 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 */

[PATCH 04/16] x86: put global variable for 32bit all together

[Yinghai Lu]

those variables are not needed by 64 bit.

Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>

Index: linux-2.6/arch/x86/kernel/setup_32.c
===================================================================
--- linux-2.6.orig/arch/x86/kernel/setup_32.c
+++ 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

[PATCH 05/16] x86: add extra includes for 64bit support

[Yinghai Lu]

Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>

Index: linux-2.6/arch/x86/kernel/setup_32.c
===================================================================
--- linux-2.6.orig/arch/x86/kernel/setup_32.c
+++ 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! */

[PATCH 06/16] x86: merge 64bit setup_arch into setup_32

[Yinghai Lu]

Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>

Index: linux-2.6/arch/x86/kernel/setup_32.c
===================================================================
--- linux-2.6.orig/arch/x86/kernel/setup_32.c
+++ 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();

[PATCH 07/16] x86: space to tab in setup_arch

[Yinghai Lu]

Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>

Index: linux-2.6/arch/x86/kernel/setup_32.c
===================================================================
--- linux-2.6.orig/arch/x86/kernel/setup_32.c
+++ 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);

[PATCH 08/16] x86: rename setup_32.c to setup.c

[Yinghai Lu]

and let 64 bit use that instead of setup_64.c

Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>

---
 arch/x86/kernel/Makefile   |    2 
 arch/x86/kernel/setup.c    |  671 +++++++++++++++++++++++++++++++++++++++++++++
 arch/x86/kernel/setup_32.c |  671 ---------------------------------------------
 arch/x86/kernel/setup_64.c |  380 -------------------------
 4 files changed, 672 insertions(+), 1052 deletions(-)

Index: linux-2.6/arch/x86/kernel/Makefile
===================================================================
--- linux-2.6.orig/arch/x86/kernel/Makefile
+++ linux-2.6/arch/x86/kernel/Makefile
@@ -26,7 +26,7 @@ CFLAGS_paravirt.o	:= $(nostackp)
 obj-y			:= process_$(BITS).o signal_$(BITS).o entry_$(BITS).o
 obj-y			+= traps_$(BITS).o irq_$(BITS).o
 obj-y			+= time_$(BITS).o ioport.o ldt.o
-obj-y			+= setup_$(BITS).o i8259.o irqinit_$(BITS).o setup_percpu.o
+obj-y			+= setup.o i8259.o irqinit_$(BITS).o setup_percpu.o
 obj-$(CONFIG_X86_32)	+= probe_roms_32.o
 obj-$(CONFIG_X86_32)	+= sys_i386_32.o i386_ksyms_32.o
 obj-$(CONFIG_X86_64)	+= sys_x86_64.o x8664_ksyms_64.o
Index: linux-2.6/arch/x86/kernel/setup.c
===================================================================
--- /dev/null
+++ linux-2.6/arch/x86/kernel/setup.c
@@ -0,0 +1,671 @@
+/*
+ *  Copyright (C) 1995  Linus Torvalds
+ *
+ *  Support of BIGMEM added by Gerhard Wichert, Siemens AG, July 1999
+ *
+ *  Memory region support
+ *	David Parsons <orc@pell.chi.il.us>, July-August 1999
+ *
+ *  Added E820 sanitization routine (removes overlapping memory regions);
+ *  Brian Moyle <bmoyle@mvista.com>, February 2001
+ *
+ * Moved CPU detection code to cpu/${cpu}.c
+ *    Patrick Mochel <mochel@osdl.org>, March 2002
+ *
+ *  Provisions for empty E820 memory regions (reported by certain BIOSes).
+ *  Alex Achenbach <xela@slit.de>, December 2002.
+ *
+ */
+
+/*
+ * This file handles the architecture-dependent parts of initialization
+ */
+
+#include <linux/sched.h>
+#include <linux/mm.h>
+#include <linux/mmzone.h>
+#include <linux/screen_info.h>
+#include <linux/ioport.h>
+#include <linux/acpi.h>
+#include <linux/apm_bios.h>
+#include <linux/initrd.h>
+#include <linux/bootmem.h>
+#include <linux/seq_file.h>
+#include <linux/console.h>
+#include <linux/mca.h>
+#include <linux/root_dev.h>
+#include <linux/highmem.h>
+#include <linux/module.h>
+#include <linux/efi.h>
+#include <linux/init.h>
+#include <linux/edd.h>
+#include <linux/iscsi_ibft.h>
+#include <linux/nodemask.h>
+#include <linux/kexec.h>
+#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/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/vmi.h>
+#include <setup_arch.h>
+#include <asm/bios_ebda.h>
+#include <asm/cacheflush.h>
+#include <asm/processor.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! */
+unsigned long init_pg_tables_start __initdata = ~0UL;
+unsigned long init_pg_tables_end __initdata = ~0UL;
+
+/*
+ * Machine setup..
+ */
+static struct resource data_resource = {
+	.name	= "Kernel data",
+	.start	= 0,
+	.end	= 0,
+	.flags	= IORESOURCE_BUSY | IORESOURCE_MEM
+};
+
+static struct resource code_resource = {
+	.name	= "Kernel code",
+	.start	= 0,
+	.end	= 0,
+	.flags	= IORESOURCE_BUSY | IORESOURCE_MEM
+};
+
+static struct resource bss_resource = {
+	.name	= "Kernel bss",
+	.start	= 0,
+	.end	= 0,
+	.flags	= IORESOURCE_BUSY | IORESOURCE_MEM
+};
+
+
+#ifdef CONFIG_X86_32
+static struct resource video_ram_resource = {
+	.name	= "Video RAM area",
+	.start	= 0xa0000,
+	.end	= 0xbffff,
+	.flags	= IORESOURCE_BUSY | IORESOURCE_MEM
+};
+
+/* 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};
+/* common cpu data for all cpus */
+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;
+
+/* 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;
+
+/*
+ * Early DMI memory
+ */
+int dmi_alloc_index;
+char dmi_alloc_data[DMI_MAX_DATA];
+
+/*
+ * Setup options
+ */
+struct screen_info screen_info;
+EXPORT_SYMBOL(screen_info);
+struct edid_info edid_info;
+EXPORT_SYMBOL_GPL(edid_info);
+
+extern int root_mountflags;
+
+unsigned long saved_video_mode;
+
+#define RAMDISK_IMAGE_START_MASK	0x07FF
+#define RAMDISK_PROMPT_FLAG		0x8000
+#define RAMDISK_LOAD_FLAG		0x4000
+
+static char __initdata command_line[COMMAND_LINE_SIZE];
+
+#if defined(CONFIG_EDD) || defined(CONFIG_EDD_MODULE)
+struct edd edd;
+#ifdef CONFIG_EDD_MODULE
+EXPORT_SYMBOL(edd);
+#endif
+/**
+ * copy_edd() - Copy the BIOS EDD information
+ *              from boot_params into a safe place.
+ *
+ */
+static inline void copy_edd(void)
+{
+     memcpy(edd.mbr_signature, boot_params.edd_mbr_sig_buffer,
+	    sizeof(edd.mbr_signature));
+     memcpy(edd.edd_info, boot_params.eddbuf, sizeof(edd.edd_info));
+     edd.mbr_signature_nr = boot_params.edd_mbr_sig_buf_entries;
+     edd.edd_info_nr = boot_params.eddbuf_entries;
+}
+#else
+static inline void copy_edd(void)
+{
+}
+#endif
+
+#ifdef CONFIG_BLK_DEV_INITRD
+
+#ifdef CONFIG_X86_32
+
+#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;
+
+	/* We need to move the initrd down into lowmem */
+	ramdisk_here = find_e820_area(0, end_of_lowmem, ramdisk_size,
+					 PAGE_SIZE);
+
+	if (ramdisk_here == -1ULL)
+		panic("Cannot find place for new RAMDISK of size %lld\n",
+			 ramdisk_size);
+
+	/* Note: this includes all the lowmem currently occupied by
+	   the initrd, we rely on that fact to keep the data intact. */
+	reserve_early(ramdisk_here, ramdisk_here + ramdisk_size,
+			 "NEW RAMDISK");
+	initrd_start = ramdisk_here + PAGE_OFFSET;
+	initrd_end   = initrd_start + ramdisk_size;
+	printk(KERN_INFO "Allocated new RAMDISK: %08llx - %08llx\n",
+			 ramdisk_here, ramdisk_here + ramdisk_size);
+
+	q = (char *)initrd_start;
+
+	/* Copy any lowmem portion of the initrd */
+	if (ramdisk_image < end_of_lowmem) {
+		clen = end_of_lowmem - ramdisk_image;
+		p = (char *)__va(ramdisk_image);
+		memcpy(q, p, clen);
+		q += clen;
+		ramdisk_image += clen;
+		ramdisk_size  -= clen;
+	}
+
+	/* Copy the highmem portion of the initrd */
+	while (ramdisk_size) {
+		slop = ramdisk_image & ~PAGE_MASK;
+		clen = ramdisk_size;
+		if (clen > MAX_MAP_CHUNK-slop)
+			clen = MAX_MAP_CHUNK-slop;
+		mapaddr = ramdisk_image & PAGE_MASK;
+		p = early_ioremap(mapaddr, clen+slop);
+		memcpy(q, p+slop, clen);
+		early_iounmap(p, clen+slop);
+		q += clen;
+		ramdisk_image += clen;
+		ramdisk_size  -= clen;
+	}
+	/* 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 "Move RAMDISK from %016llx - %016llx to"
+		" %08llx - %08llx\n",
+		ramdisk_image, ramdisk_image + ramdisk_size - 1,
+		ramdisk_here, ramdisk_here + ramdisk_size - 1);
+}
+#endif
+
+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
+static void __init reserve_initrd(void)
+{
+}
+#endif /* CONFIG_BLK_DEV_INITRD */
+
+/*
+ * 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
+ * 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;
+	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
+	rd_image_start = boot_params.hdr.ram_size & RAMDISK_IMAGE_START_MASK;
+	rd_prompt = ((boot_params.hdr.ram_size & RAMDISK_PROMPT_FLAG) != 0);
+	rd_doload = ((boot_params.hdr.ram_size & RAMDISK_LOAD_FLAG) != 0);
+#endif
+#ifdef CONFIG_EFI
+	if (!strncmp((char *)&boot_params.efi_info.efi_loader_signature,
+#ifdef CONFIG_X86_32
+		     "EL32",
+#else
+		     "EL64",
+#endif
+	 4)) {
+		efi_enabled = 1;
+		efi_reserve_early();
+	}
+#endif
+
+	ARCH_SETUP
+
+	setup_memory_map();
+	copy_edd();
+
+	if (!boot_params.hdr.root_flags)
+		root_mountflags &= ~MS_RDONLY;
+	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;
+	data_resource.start = virt_to_phys(_etext);
+	data_resource.end = virt_to_phys(_edata)-1;
+	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;
+
+	parse_setup_data();
+
+	parse_early_param();
+
+	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);
+	insert_resource(&iomem_resource, &data_resource);
+	insert_resource(&iomem_resource, &bss_resource);
+
+	if (efi_enabled)
+		efi_init();
+
+#ifdef CONFIG_X86_32
+	if (ppro_with_ram_bug()) {
+		e820_update_range(0x70000000ULL, 0x40000ULL, E820_RAM,
+				  E820_RESERVED);
+		sanitize_e820_map(e820.map, ARRAY_SIZE(e820.map), &e820.nr_map);
+		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);
+	/*
+	 * partially used pages are not usable - thus
+	 * we are rounding upwards:
+	 */
+	max_pfn = e820_end_of_ram();
+
+	/* preallocate 4k for mptable mpc */
+	early_reserve_e820_mpc_new();
+	/* update e820 for memory not covered by WB MTRRs */
+	mtrr_bp_init();
+	if (mtrr_trim_uncached_memory(max_pfn)) {
+		remove_all_active_ranges();
+		e820_register_active_regions(0, 0, -1UL);
+		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();
+
+	/*
+	 * Parse the ACPI tables for possible boot-time SMP configuration.
+	 */
+	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.
+	 */
+	acpi_numa_init();
+#endif
+
+	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
+#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:
+	 */
+	find_smp_config();
+#endif
+	reserve_crashkernel();
+
+	reserve_ibft_region();
+
+#ifdef CONFIG_KVM_CLOCK
+	kvmclock_init();
+#endif
+
+#if defined(CONFIG_VMI) && defined(CONFIG_X86_32)
+	/*
+	 * Must be after max_low_pfn is determined, and before kernel
+	 * pagetables are setup.
+	 */
+	vmi_init();
+#endif
+
+	paging_init();
+
+#ifdef CONFIG_X86_64
+	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
+
+	early_quirks();
+
+	/*
+	 * Read APIC and some other early information from ACPI tables.
+	 */
+	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:
+	 */
+	if (smp_found_config)
+		get_smp_config();
+#endif
+
+#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();
+
+#ifdef CONFIG_VT
+#if defined(CONFIG_VGA_CONSOLE)
+	if (!efi_enabled || (efi_mem_type(0xa0000) != EFI_CONVENTIONAL_MEMORY))
+		conswitchp = &vga_con;
+#elif defined(CONFIG_DUMMY_CONSOLE)
+	conswitchp = &dummy_con;
+#endif
+#endif
+}
+
Index: linux-2.6/arch/x86/kernel/setup_32.c
===================================================================
--- linux-2.6.orig/arch/x86/kernel/setup_32.c
+++ /dev/null
@@ -1,671 +0,0 @@
-/*
- *  Copyright (C) 1995  Linus Torvalds
- *
- *  Support of BIGMEM added by Gerhard Wichert, Siemens AG, July 1999
- *
- *  Memory region support
- *	David Parsons <orc@pell.chi.il.us>, July-August 1999
- *
- *  Added E820 sanitization routine (removes overlapping memory regions);
- *  Brian Moyle <bmoyle@mvista.com>, February 2001
- *
- * Moved CPU detection code to cpu/${cpu}.c
- *    Patrick Mochel <mochel@osdl.org>, March 2002
- *
- *  Provisions for empty E820 memory regions (reported by certain BIOSes).
- *  Alex Achenbach <xela@slit.de>, December 2002.
- *
- */
-
-/*
- * This file handles the architecture-dependent parts of initialization
- */
-
-#include <linux/sched.h>
-#include <linux/mm.h>
-#include <linux/mmzone.h>
-#include <linux/screen_info.h>
-#include <linux/ioport.h>
-#include <linux/acpi.h>
-#include <linux/apm_bios.h>
-#include <linux/initrd.h>
-#include <linux/bootmem.h>
-#include <linux/seq_file.h>
-#include <linux/console.h>
-#include <linux/mca.h>
-#include <linux/root_dev.h>
-#include <linux/highmem.h>
-#include <linux/module.h>
-#include <linux/efi.h>
-#include <linux/init.h>
-#include <linux/edd.h>
-#include <linux/iscsi_ibft.h>
-#include <linux/nodemask.h>
-#include <linux/kexec.h>
-#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/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/vmi.h>
-#include <setup_arch.h>
-#include <asm/bios_ebda.h>
-#include <asm/cacheflush.h>
-#include <asm/processor.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! */
-unsigned long init_pg_tables_start __initdata = ~0UL;
-unsigned long init_pg_tables_end __initdata = ~0UL;
-
-/*
- * Machine setup..
- */
-static struct resource data_resource = {
-	.name	= "Kernel data",
-	.start	= 0,
-	.end	= 0,
-	.flags	= IORESOURCE_BUSY | IORESOURCE_MEM
-};
-
-static struct resource code_resource = {
-	.name	= "Kernel code",
-	.start	= 0,
-	.end	= 0,
-	.flags	= IORESOURCE_BUSY | IORESOURCE_MEM
-};
-
-static struct resource bss_resource = {
-	.name	= "Kernel bss",
-	.start	= 0,
-	.end	= 0,
-	.flags	= IORESOURCE_BUSY | IORESOURCE_MEM
-};
-
-
-#ifdef CONFIG_X86_32
-static struct resource video_ram_resource = {
-	.name	= "Video RAM area",
-	.start	= 0xa0000,
-	.end	= 0xbffff,
-	.flags	= IORESOURCE_BUSY | IORESOURCE_MEM
-};
-
-/* 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};
-/* common cpu data for all cpus */
-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;
-
-/* 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;
-
-/*
- * Early DMI memory
- */
-int dmi_alloc_index;
-char dmi_alloc_data[DMI_MAX_DATA];
-
-/*
- * Setup options
- */
-struct screen_info screen_info;
-EXPORT_SYMBOL(screen_info);
-struct edid_info edid_info;
-EXPORT_SYMBOL_GPL(edid_info);
-
-extern int root_mountflags;
-
-unsigned long saved_video_mode;
-
-#define RAMDISK_IMAGE_START_MASK	0x07FF
-#define RAMDISK_PROMPT_FLAG		0x8000
-#define RAMDISK_LOAD_FLAG		0x4000
-
-static char __initdata command_line[COMMAND_LINE_SIZE];
-
-#if defined(CONFIG_EDD) || defined(CONFIG_EDD_MODULE)
-struct edd edd;
-#ifdef CONFIG_EDD_MODULE
-EXPORT_SYMBOL(edd);
-#endif
-/**
- * copy_edd() - Copy the BIOS EDD information
- *              from boot_params into a safe place.
- *
- */
-static inline void copy_edd(void)
-{
-     memcpy(edd.mbr_signature, boot_params.edd_mbr_sig_buffer,
-	    sizeof(edd.mbr_signature));
-     memcpy(edd.edd_info, boot_params.eddbuf, sizeof(edd.edd_info));
-     edd.mbr_signature_nr = boot_params.edd_mbr_sig_buf_entries;
-     edd.edd_info_nr = boot_params.eddbuf_entries;
-}
-#else
-static inline void copy_edd(void)
-{
-}
-#endif
-
-#ifdef CONFIG_BLK_DEV_INITRD
-
-#ifdef CONFIG_X86_32
-
-#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;
-
-	/* We need to move the initrd down into lowmem */
-	ramdisk_here = find_e820_area(0, end_of_lowmem, ramdisk_size,
-					 PAGE_SIZE);
-
-	if (ramdisk_here == -1ULL)
-		panic("Cannot find place for new RAMDISK of size %lld\n",
-			 ramdisk_size);
-
-	/* Note: this includes all the lowmem currently occupied by
-	   the initrd, we rely on that fact to keep the data intact. */
-	reserve_early(ramdisk_here, ramdisk_here + ramdisk_size,
-			 "NEW RAMDISK");
-	initrd_start = ramdisk_here + PAGE_OFFSET;
-	initrd_end   = initrd_start + ramdisk_size;
-	printk(KERN_INFO "Allocated new RAMDISK: %08llx - %08llx\n",
-			 ramdisk_here, ramdisk_here + ramdisk_size);
-
-	q = (char *)initrd_start;
-
-	/* Copy any lowmem portion of the initrd */
-	if (ramdisk_image < end_of_lowmem) {
-		clen = end_of_lowmem - ramdisk_image;
-		p = (char *)__va(ramdisk_image);
-		memcpy(q, p, clen);
-		q += clen;
-		ramdisk_image += clen;
-		ramdisk_size  -= clen;
-	}
-
-	/* Copy the highmem portion of the initrd */
-	while (ramdisk_size) {
-		slop = ramdisk_image & ~PAGE_MASK;
-		clen = ramdisk_size;
-		if (clen > MAX_MAP_CHUNK-slop)
-			clen = MAX_MAP_CHUNK-slop;
-		mapaddr = ramdisk_image & PAGE_MASK;
-		p = early_ioremap(mapaddr, clen+slop);
-		memcpy(q, p+slop, clen);
-		early_iounmap(p, clen+slop);
-		q += clen;
-		ramdisk_image += clen;
-		ramdisk_size  -= clen;
-	}
-	/* 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 "Move RAMDISK from %016llx - %016llx to"
-		" %08llx - %08llx\n",
-		ramdisk_image, ramdisk_image + ramdisk_size - 1,
-		ramdisk_here, ramdisk_here + ramdisk_size - 1);
-}
-#endif
-
-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
-static void __init reserve_initrd(void)
-{
-}
-#endif /* CONFIG_BLK_DEV_INITRD */
-
-/*
- * 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
- * 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;
-	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
-	rd_image_start = boot_params.hdr.ram_size & RAMDISK_IMAGE_START_MASK;
-	rd_prompt = ((boot_params.hdr.ram_size & RAMDISK_PROMPT_FLAG) != 0);
-	rd_doload = ((boot_params.hdr.ram_size & RAMDISK_LOAD_FLAG) != 0);
-#endif
-#ifdef CONFIG_EFI
-	if (!strncmp((char *)&boot_params.efi_info.efi_loader_signature,
-#ifdef CONFIG_X86_32
-		     "EL32",
-#else
-		     "EL64",
-#endif
-	 4)) {
-		efi_enabled = 1;
-		efi_reserve_early();
-	}
-#endif
-
-	ARCH_SETUP
-
-	setup_memory_map();
-	copy_edd();
-
-	if (!boot_params.hdr.root_flags)
-		root_mountflags &= ~MS_RDONLY;
-	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;
-	data_resource.start = virt_to_phys(_etext);
-	data_resource.end = virt_to_phys(_edata)-1;
-	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;
-
-	parse_setup_data();
-
-	parse_early_param();
-
-	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);
-	insert_resource(&iomem_resource, &data_resource);
-	insert_resource(&iomem_resource, &bss_resource);
-
-	if (efi_enabled)
-		efi_init();
-
-#ifdef CONFIG_X86_32
-	if (ppro_with_ram_bug()) {
-		e820_update_range(0x70000000ULL, 0x40000ULL, E820_RAM,
-				  E820_RESERVED);
-		sanitize_e820_map(e820.map, ARRAY_SIZE(e820.map), &e820.nr_map);
-		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);
-	/*
-	 * partially used pages are not usable - thus
-	 * we are rounding upwards:
-	 */
-	max_pfn = e820_end_of_ram();
-
-	/* preallocate 4k for mptable mpc */
-	early_reserve_e820_mpc_new();
-	/* update e820 for memory not covered by WB MTRRs */
-	mtrr_bp_init();
-	if (mtrr_trim_uncached_memory(max_pfn)) {
-		remove_all_active_ranges();
-		e820_register_active_regions(0, 0, -1UL);
-		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();
-
-	/*
-	 * Parse the ACPI tables for possible boot-time SMP configuration.
-	 */
-	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.
-	 */
-	acpi_numa_init();
-#endif
-
-	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
-#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:
-	 */
-	find_smp_config();
-#endif
-	reserve_crashkernel();
-
-	reserve_ibft_region();
-
-#ifdef CONFIG_KVM_CLOCK
-	kvmclock_init();
-#endif
-
-#if defined(CONFIG_VMI) && defined(CONFIG_X86_32)
-	/*
-	 * Must be after max_low_pfn is determined, and before kernel
-	 * pagetables are setup.
-	 */
-	vmi_init();
-#endif
-
-	paging_init();
-
-#ifdef CONFIG_X86_64
-	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
-
-	early_quirks();
-
-	/*
-	 * Read APIC and some other early information from ACPI tables.
-	 */
-	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:
-	 */
-	if (smp_found_config)
-		get_smp_config();
-#endif
-
-#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();
-
-#ifdef CONFIG_VT
-#if defined(CONFIG_VGA_CONSOLE)
-	if (!efi_enabled || (efi_mem_type(0xa0000) != EFI_CONVENTIONAL_MEMORY))
-		conswitchp = &vga_con;
-#elif defined(CONFIG_DUMMY_CONSOLE)
-	conswitchp = &dummy_con;
-#endif
-#endif
-}
-
Index: linux-2.6/arch/x86/kernel/setup_64.c
===================================================================
--- linux-2.6.orig/arch/x86/kernel/setup_64.c
+++ /dev/null
@@ -1,380 +0,0 @@
-/*
- *  Copyright (C) 1995  Linus Torvalds
- */
-
-/*
- * This file handles the architecture-dependent parts of initialization
- */
-
-#include <linux/errno.h>
-#include <linux/sched.h>
-#include <linux/kernel.h>
-#include <linux/mm.h>
-#include <linux/stddef.h>
-#include <linux/unistd.h>
-#include <linux/ptrace.h>
-#include <linux/slab.h>
-#include <linux/user.h>
-#include <linux/screen_info.h>
-#include <linux/ioport.h>
-#include <linux/delay.h>
-#include <linux/init.h>
-#include <linux/initrd.h>
-#include <linux/highmem.h>
-#include <linux/bootmem.h>
-#include <linux/module.h>
-#include <asm/processor.h>
-#include <linux/console.h>
-#include <linux/seq_file.h>
-#include <linux/root_dev.h>
-#include <linux/pci.h>
-#include <asm/pci-direct.h>
-#include <linux/efi.h>
-#include <linux/acpi.h>
-#include <linux/kallsyms.h>
-#include <linux/edd.h>
-#include <linux/iscsi_ibft.h>
-#include <linux/mmzone.h>
-#include <linux/kexec.h>
-#include <linux/cpufreq.h>
-#include <linux/dmi.h>
-#include <linux/dma-mapping.h>
-#include <linux/ctype.h>
-#include <linux/uaccess.h>
-#include <linux/init_ohci1394_dma.h>
-#include <linux/kvm_para.h>
-
-#include <asm/mtrr.h>
-#include <asm/uaccess.h>
-#include <asm/system.h>
-#include <asm/vsyscall.h>
-#include <asm/io.h>
-#include <asm/smp.h>
-#include <asm/desc.h>
-#include <video/edid.h>
-#include <asm/e820.h>
-#include <asm/mpspec.h>
-#include <asm/dma.h>
-#include <asm/gart.h>
-#include <asm/mpspec.h>
-#include <asm/mmu_context.h>
-#include <asm/proto.h>
-#include <asm/setup.h>
-#include <asm/numa.h>
-#include <asm/sections.h>
-#include <asm/dmi.h>
-#include <asm/cacheflush.h>
-
-#include <mach_apic.h>
-#ifdef CONFIG_PARAVIRT
-#include <asm/paravirt.h>
-#else
-#define ARCH_SETUP
-#endif
-
-/*
- * Machine setup..
- */
-
-struct cpuinfo_x86 boot_cpu_data __read_mostly;
-EXPORT_SYMBOL(boot_cpu_data);
-
-unsigned long mmu_cr4_features;
-
-/* Boot loader ID as an integer, for the benefit of proc_dointvec */
-int bootloader_type;
-
-unsigned long saved_video_mode;
-
-/*
- * Early DMI memory
- */
-int dmi_alloc_index;
-char dmi_alloc_data[DMI_MAX_DATA];
-
-/*
- * Setup options
- */
-struct screen_info screen_info;
-EXPORT_SYMBOL(screen_info);
-struct sys_desc_table_struct {
-	unsigned short length;
-	unsigned char table[0];
-};
-
-struct edid_info edid_info;
-EXPORT_SYMBOL_GPL(edid_info);
-
-extern int root_mountflags;
-
-static char __initdata command_line[COMMAND_LINE_SIZE];
-
-#define IORESOURCE_RAM (IORESOURCE_BUSY | IORESOURCE_MEM)
-
-static struct resource data_resource = {
-	.name = "Kernel data",
-	.start = 0,
-	.end = 0,
-	.flags = IORESOURCE_RAM,
-};
-static struct resource code_resource = {
-	.name = "Kernel code",
-	.start = 0,
-	.end = 0,
-	.flags = IORESOURCE_RAM,
-};
-static struct resource bss_resource = {
-	.name = "Kernel bss",
-	.start = 0,
-	.end = 0,
-	.flags = IORESOURCE_RAM,
-};
-
-#if defined(CONFIG_EDD) || defined(CONFIG_EDD_MODULE)
-struct edd edd;
-#ifdef CONFIG_EDD_MODULE
-EXPORT_SYMBOL(edd);
-#endif
-/**
- * copy_edd() - Copy the BIOS EDD information
- *              from boot_params into a safe place.
- *
- */
-static inline void copy_edd(void)
-{
-     memcpy(edd.mbr_signature, boot_params.edd_mbr_sig_buffer,
-	    sizeof(edd.mbr_signature));
-     memcpy(edd.edd_info, boot_params.eddbuf, sizeof(edd.edd_info));
-     edd.mbr_signature_nr = boot_params.edd_mbr_sig_buf_entries;
-     edd.edd_info_nr = boot_params.eddbuf_entries;
-}
-#else
-static inline void copy_edd(void)
-{
-}
-#endif
-
-static void __init reserve_initrd(void)
-{
-#ifdef CONFIG_BLK_DEV_INITRD
-	if (boot_params.hdr.type_of_loader && boot_params.hdr.ramdisk_image) {
-		unsigned long ramdisk_image = boot_params.hdr.ramdisk_image;
-		unsigned long ramdisk_size  = boot_params.hdr.ramdisk_size;
-		unsigned long ramdisk_end   = ramdisk_image + ramdisk_size;
-		unsigned long end_of_mem    = max_pfn << PAGE_SHIFT;
-
-		if (ramdisk_end <= end_of_mem) {
-			/*
-			 * don't need to reserve again, already reserved early
-			 * in x86_64_start_kernel, and early_res_to_bootmem
-			 * will convert that to reserved in bootmem
-			 */
-			initrd_start = ramdisk_image + PAGE_OFFSET;
-			initrd_end = initrd_start+ramdisk_size;
-		} else {
-			free_early(ramdisk_image, ramdisk_end);
-			printk(KERN_ERR "initrd extends beyond end of memory "
-			       "(0x%08lx > 0x%08lx)\ndisabling initrd\n",
-			       ramdisk_end, end_of_mem);
-			initrd_start = 0;
-		}
-	}
-#endif
-}
-
-/*
- * 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)
-{
-	printk(KERN_INFO "Command line: %s\n", boot_command_line);
-
-	ROOT_DEV = old_decode_dev(boot_params.hdr.root_dev);
-	screen_info = boot_params.screen_info;
-	edid_info = boot_params.edid_info;
-	saved_video_mode = boot_params.hdr.vid_mode;
-	bootloader_type = boot_params.hdr.type_of_loader;
-
-#ifdef CONFIG_BLK_DEV_RAM
-	rd_image_start = boot_params.hdr.ram_size & RAMDISK_IMAGE_START_MASK;
-	rd_prompt = ((boot_params.hdr.ram_size & RAMDISK_PROMPT_FLAG) != 0);
-	rd_doload = ((boot_params.hdr.ram_size & RAMDISK_LOAD_FLAG) != 0);
-#endif
-#ifdef CONFIG_EFI
-	if (!strncmp((char *)&boot_params.efi_info.efi_loader_signature,
-		     "EL64", 4)) {
-		efi_enabled = 1;
-		efi_reserve_early();
-	}
-#endif
-
-	ARCH_SETUP
-
-	setup_memory_map();
-	copy_edd();
-
-	if (!boot_params.hdr.root_flags)
-		root_mountflags &= ~MS_RDONLY;
-	init_mm.start_code = (unsigned long) &_text;
-	init_mm.end_code = (unsigned long) &_etext;
-	init_mm.end_data = (unsigned long) &_edata;
-	init_mm.brk = (unsigned long) &_end;
-
-	code_resource.start = virt_to_phys(&_text);
-	code_resource.end = virt_to_phys(&_etext)-1;
-	data_resource.start = virt_to_phys(&_etext);
-	data_resource.end = virt_to_phys(&_edata)-1;
-	bss_resource.start = virt_to_phys(&__bss_start);
-	bss_resource.end = virt_to_phys(&__bss_stop)-1;
-
-	early_cpu_init();
-
-	strlcpy(command_line, boot_command_line, COMMAND_LINE_SIZE);
-	*cmdline_p = command_line;
-
-	parse_setup_data();
-
-	parse_early_param();
-
-	if (acpi_mps_check()) {
-		disable_apic = 1;
-		clear_cpu_cap(&boot_cpu_data, X86_FEATURE_APIC);
-	}
-
-#ifdef CONFIG_PROVIDE_OHCI1394_DMA_INIT
-	if (init_ohci1394_dma_early)
-		init_ohci1394_dma_on_all_controllers();
-#endif
-
-	finish_e820_parsing();
-
-	/* after parse_early_param, so could debug it */
-	insert_resource(&iomem_resource, &code_resource);
-	insert_resource(&iomem_resource, &data_resource);
-	insert_resource(&iomem_resource, &bss_resource);
-
-	if (efi_enabled)
-		efi_init();
-
-	early_gart_iommu_check();
-
-	e820_register_active_regions(0, 0, -1UL);
-	/*
-	 * partially used pages are not usable - thus
-	 * we are rounding upwards:
-	 */
-	max_pfn = e820_end_of_ram();
-
-	/* pre allocte 4k for mptable mpc */
-	early_reserve_e820_mpc_new();
-	/* update e820 for memory not covered by WB MTRRs */
-	mtrr_bp_init();
-	if (mtrr_trim_uncached_memory(max_pfn)) {
-		remove_all_active_ranges();
-		e820_register_active_regions(0, 0, -1UL);
-		max_pfn = e820_end_of_ram();
-	}
-
-	reserve_initrd();
-
-	num_physpages = max_pfn;
-
-	check_efer();
-
-	max_pfn_mapped = init_memory_mapping(0, (max_pfn << PAGE_SHIFT));
-
-	vsmp_init();
-
-	dmi_scan_machine();
-
-	io_delay_init();
-
-	/*
-	 * Initialize the ACPI boot-time table parser (gets the RSDP and SDT).
-	 * Call this early for SRAT node setup.
-	 */
-	acpi_boot_table_init();
-
-	/* How many end-of-memory variables you have, grandma! */
-	max_low_pfn = max_pfn;
-	high_memory = (void *)__va(max_pfn * PAGE_SIZE - 1) + 1;
-
-	/* Remove active ranges so rediscovery with NUMA-awareness happens */
-	remove_all_active_ranges();
-
-#ifdef CONFIG_ACPI_NUMA
-	/*
-	 * Parse SRAT to discover nodes.
-	 */
-	acpi_numa_init();
-#endif
-
-	initmem_init(0, max_pfn);
-
-	dma32_reserve_bootmem();
-
-#ifdef CONFIG_ACPI_SLEEP
-	/*
-	 * Reserve low memory region for sleep support.
-	 */
-       acpi_reserve_bootmem();
-#endif
-
-#ifdef CONFIG_X86_MPPARSE
-       /*
-	* Find and reserve possible boot-time SMP configuration:
-	*/
-	find_smp_config();
-#endif
-	reserve_crashkernel();
-
-	reserve_ibft_region();
-
-#ifdef CONFIG_KVM_CLOCK
-	kvmclock_init();
-#endif
-	paging_init();
-	map_vsyscall();
-
-	early_quirks();
-
-	/*
-	 * Read APIC and some other early information from ACPI tables.
-	 */
-	acpi_boot_init();
-
-	init_cpu_to_node();
-
-#ifdef CONFIG_X86_MPPARSE
-	/*
-	 * get boot-time SMP configuration:
-	 */
-	if (smp_found_config)
-		get_smp_config();
-#endif
-	init_apic_mappings();
-	ioapic_init_mappings();
-
-	kvm_guest_init();
-
-	/*
-	 * We trust e820 completely. No explicit ROM probing in memory.
-	 */
-	e820_reserve_resources();
-	e820_mark_nosave_regions(max_pfn);
-
-	reserve_standard_io_resources();
-
-	e820_setup_gap();
-
-#ifdef CONFIG_VT
-#if defined(CONFIG_VGA_CONSOLE)
-	if (!efi_enabled || (efi_mem_type(0xa0000) != EFI_CONVENTIONAL_MEMORY))
-		conswitchp = &vga_con;
-#elif defined(CONFIG_DUMMY_CONSOLE)
-	conswitchp = &dummy_con;
-#endif
-#endif
-}

[PATCH 09/16] x86: move boot_params back to setup.c

[Yinghai Lu]

Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>

---
 arch/x86/kernel/setup.c        |    6 ++++++
 arch/x86/kernel/setup_percpu.c |    6 ------
 2 files changed, 6 insertions(+), 6 deletions(-)

Index: linux-2.6/arch/x86/kernel/setup.c
===================================================================
--- linux-2.6.orig/arch/x86/kernel/setup.c
+++ 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
===================================================================
--- linux-2.6.orig/arch/x86/kernel/setup_percpu.c
+++ 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;

[PATCH 10/16] x86: move parse_setup_data back to setup.c

[Yinghai Lu]

Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>

---
 arch/x86/kernel/setup.c        |   25 +++++++++++++++++++++++++
 arch/x86/kernel/setup_percpu.c |   25 -------------------------
 2 files changed, 25 insertions(+), 25 deletions(-)

Index: linux-2.6/arch/x86/kernel/setup.c
===================================================================
--- linux-2.6.orig/arch/x86/kernel/setup.c
+++ 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
===================================================================
--- linux-2.6.orig/arch/x86/kernel/setup_percpu.c
+++ 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
 
 /*

[PATCH 11/16] x86: move back crashkernel back to setup.c

[Yinghai Lu]

Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>

---
 arch/x86/kernel/setup.c        |   58 +++++++++++++++++++++++++++++++++++++++++
 arch/x86/kernel/setup_percpu.c |   57 ----------------------------------------
 2 files changed, 58 insertions(+), 57 deletions(-)

Index: linux-2.6/arch/x86/kernel/setup.c
===================================================================
--- linux-2.6.orig/arch/x86/kernel/setup.c
+++ 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
Index: linux-2.6/arch/x86/kernel/setup_percpu.c
===================================================================
--- linux-2.6.orig/arch/x86/kernel/setup_percpu.c
+++ linux-2.6/arch/x86/kernel/setup_percpu.c
@@ -387,63 +387,6 @@ EXPORT_SYMBOL(node_to_cpumask);
 
 #endif /* X86_64_NUMA */
 
-/*
- * --------- Crashkernel reservation ------------------------------
- */
-
-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;
-}
-
-#ifdef CONFIG_KEXEC
-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
-
 static struct resource standard_io_resources[] = {
 	{ .name = "dma1", .start = 0x00, .end = 0x1f,
 		.flags = IORESOURCE_BUSY | IORESOURCE_IO },

[PATCH 12/16] x86: move reserve_standard_io_resources back to setup.c

[Yinghai Lu]

Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>

---
 arch/x86/kernel/setup.c        |   33 +++++++++++++++++++++++++++++++++
 arch/x86/kernel/setup_percpu.c |   32 --------------------------------
 2 files changed, 33 insertions(+), 32 deletions(-)

Index: linux-2.6/arch/x86/kernel/setup.c
===================================================================
--- linux-2.6.orig/arch/x86/kernel/setup.c
+++ 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
===================================================================
--- linux-2.6.orig/arch/x86/kernel/setup_percpu.c
+++ 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

[PATCH 13/16] x86: move parse elfvorehdr back to setup.c

[Yinghai Lu]

Signed-off-by: Yinghai <yhlu.kernel@gmail.com>

---
 arch/x86/kernel/setup.c        |   16 ++++++++++++++++
 arch/x86/kernel/setup_percpu.c |   16 ----------------
 2 files changed, 16 insertions(+), 16 deletions(-)

Index: linux-2.6/arch/x86/kernel/setup.c
===================================================================
--- linux-2.6.orig/arch/x86/kernel/setup.c
+++ 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
===================================================================
--- linux-2.6.orig/arch/x86/kernel/setup_percpu.c
+++ 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

[PATCH 14/16] x86: make x86_find_smp_config depends on 64 bit too

[Yinghai Lu]

Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>

---
 arch/x86/Kconfig        |    1 -
 arch/x86/kernel/setup.c |    3 +--
 2 files changed, 1 insertion(+), 3 deletions(-)

Index: linux-2.6/arch/x86/Kconfig
===================================================================
--- linux-2.6.orig/arch/x86/Kconfig
+++ linux-2.6/arch/x86/Kconfig
@@ -236,7 +236,6 @@ config SMP
 config X86_FIND_SMP_CONFIG
 	def_bool y
 	depends on X86_MPPARSE || X86_VOYAGER || X86_VISWS
-	depends on X86_32
 
 if ACPI
 config X86_MPPARSE
Index: linux-2.6/arch/x86/kernel/setup.c
===================================================================
--- linux-2.6.orig/arch/x86/kernel/setup.c
+++ 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:
 	 */

[PATCH 15/16] x86: change some functions in setup.c to static

[Yinghai Lu]

Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>

---
 arch/x86/kernel/setup.c     |    8 ++++----
 include/asm-x86/bootparam.h |    1 -
 include/asm-x86/setup.h     |    5 -----
 3 files changed, 4 insertions(+), 10 deletions(-)

Index: linux-2.6/arch/x86/kernel/setup.c
===================================================================
--- linux-2.6.orig/arch/x86/kernel/setup.c
+++ 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
===================================================================
--- linux-2.6.orig/include/asm-x86/bootparam.h
+++ 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
===================================================================
--- linux-2.6.orig/include/asm-x86/setup.h
+++ 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
 
 /*

[PATCH 16/16] x86: we only have init_pg_tables_end for 32bit

[Yinghai Lu]

Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>


---
 arch/x86/kernel/setup.c |   12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

Index: linux-2.6/arch/x86/kernel/setup.c
===================================================================
--- linux-2.6.orig/arch/x86/kernel/setup.c
+++ 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,

[PATCH] x86: clean up ARCH_SETUP

[Yinghai Lu]

asm-x86/paravirt.h already have protection with CONFIG_PARAVIRT inside

Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>

---
 arch/x86/kernel/setup.c                   |    8 ++++----
 include/asm-x86/mach-default/setup_arch.h |    4 ----
 include/asm-x86/mach-visws/setup_arch.h   |    2 --
 3 files changed, 4 insertions(+), 10 deletions(-)

Index: linux-2.6/arch/x86/kernel/setup.c
===================================================================
--- linux-2.6.orig/arch/x86/kernel/setup.c
+++ 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
===================================================================
--- linux-2.6.orig/include/asm-x86/mach-default/setup_arch.h
+++ 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
===================================================================
--- linux-2.6.orig/include/asm-x86/mach-visws/setup_arch.h
+++ 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

[PATCH] x86: move fix mapping page table range early

[Yinghai Lu]

do that in init_memory_mapping

also remove one init_ohci1394_dma_on_all_controllers

Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>

---
 arch/x86/kernel/setup.c |   23 +++++++++--------------
 arch/x86/mm/init_32.c   |   15 +++++++++++----
 2 files changed, 20 insertions(+), 18 deletions(-)

Index: linux-2.6/arch/x86/kernel/setup.c
===================================================================
--- linux-2.6.orig/arch/x86/kernel/setup.c
+++ 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
===================================================================
--- linux-2.6.orig/arch/x86/mm/init_32.c
+++ 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.

Re: [PATCH 00/16] x86: merge setup_32/64.c

[Ingo Molnar]

* Yinghai Lu <yhlu.kernel@gmail.com> wrote:

> please check the patches
> 
> it will rename setup.c to setup_percpu.c and make setup_32.c to 
> support 64bit, and then change setup_32.c to setup.c and delete 
> setup_64.c.

applied them to tip/x86/unify-setup:

 Yinghai Lu (17):
      x86: fix memory setup bug
      x86: rename setup.c to setup_percpu.c
      x86: we can use full bootmem after have init_memory_mapping
      x86: update reserve_initrd to support 64bit
      x86: put global variable for 32bit all together
      x86: add extra includes for 64bit support
      x86: merge 64bit setup_arch into setup_32
      x86: space to tab in setup_arch
      x86: rename setup_32.c to setup.c
      x86: move boot_params back to setup.c
      x86: move parse_setup_data back to setup.c
      x86: move back crashkernel back to setup.c
      x86: move reserve_standard_io_resources back to setup.c
      x86: move parse elfvorehdr back to setup.c
      x86: make x86_find_smp_config depends on 64 bit too
      x86: change some functions in setup.c to static
      x86: we only have init_pg_tables_end for 32bit

thanks Yinghai.

	Ingo

Re: [PATCH 00/16] x86: merge setup_32/64.c

[Ingo Molnar]

* Ingo Molnar <mingo@elte.hu> wrote:

>       x86: rename setup.c to setup_percpu.c

this brought a build failure:

 arch/x86/kernel/setup.c: In function ‘setup_arch':
 arch/x86/kernel/setup.c:561: error: implicit declaration of function ‘efi_reserve_early'

09865f3049050458a5c236e5b4630f137d9b179d is first bad commit
commit 09865f3049050458a5c236e5b4630f137d9b179d
Author: Yinghai Lu <yhlu.kernel@gmail.com>
Date:   Wed Jun 25 17:48:14 2008 -0700

    x86: rename setup.c to setup_percpu.c

we lost an asm/efi.h include. (which, a bit illogically, is not implicit 
via the existing linux/efi.h include.) Fixed below.

	Ingo

-------------->
commit 9faea5db92f5ff3b39e7b66656f847dcb9618caa
Author: Ingo Molnar <mingo@elte.hu>
Date:   Thu Jun 26 12:02:01 2008 +0200

    x86: build fix
    
    fix:
    
    arch/x86/kernel/setup.c: In function ‘setup_arch':
    arch/x86/kernel/setup.c:561: error: implicit declaration of function ‘efi_reserve_early'
    
    Signed-off-by: Ingo Molnar <mingo@elte.hu>

diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c
index a3f2263..29f04da 100644
--- a/arch/x86/kernel/setup.c
+++ b/arch/x86/kernel/setup.c
@@ -79,6 +79,7 @@
 #include <asm/mpspec.h>
 #include <asm/setup.h>
 #include <asm/arch_hooks.h>
+#include <asm/efi.h>
 #include <asm/sections.h>
 #include <asm/dmi.h>
 #include <asm/io_apic.h>

Re: [PATCH 00/16] x86: merge setup_32/64.c

[Yinghai Lu]

On Thu, Jun 26, 2008 at 3:02 AM, Ingo Molnar <mingo@elte.hu> wrote:
>
> * Ingo Molnar <mingo@elte.hu> wrote:
>
>>       x86: rename setup.c to setup_percpu.c
>
> this brought a build failure:
>
>  arch/x86/kernel/setup.c: In function 'setup_arch':
>  arch/x86/kernel/setup.c:561: error: implicit declaration of function 'efi_reserve_early'
>
> 09865f3049050458a5c236e5b4630f137d9b179d is first bad commit
> commit 09865f3049050458a5c236e5b4630f137d9b179d
> Author: Yinghai Lu <yhlu.kernel@gmail.com>
> Date:   Wed Jun 25 17:48:14 2008 -0700
>
>    x86: rename setup.c to setup_percpu.c
>
> we lost an asm/efi.h include. (which, a bit illogically, is not implicit
> via the existing linux/efi.h include.) Fixed below.
>
>        Ingo
>
> -------------->
> commit 9faea5db92f5ff3b39e7b66656f847dcb9618caa
> Author: Ingo Molnar <mingo@elte.hu>
> Date:   Thu Jun 26 12:02:01 2008 +0200
>
>    x86: build fix
>
>    fix:
>
>    arch/x86/kernel/setup.c: In function 'setup_arch':
>    arch/x86/kernel/setup.c:561: error: implicit declaration of function 'efi_reserve_early'
>
>    Signed-off-by: Ingo Molnar <mingo@elte.hu>
>
> diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c
> index a3f2263..29f04da 100644
> --- a/arch/x86/kernel/setup.c
> +++ b/arch/x86/kernel/setup.c
> @@ -79,6 +79,7 @@
>  #include <asm/mpspec.h>
>  #include <asm/setup.h>
>  #include <asm/arch_hooks.h>
> +#include <asm/efi.h>
>  #include <asm/sections.h>
>  #include <asm/dmi.h>
>  #include <asm/io_apic.h>

checkpatch.pl want me to remove that.

YH

Re: [PATCH 00/16] x86: merge setup_32/64.c

[Ingo Molnar]

* Yinghai Lu <yhlu.kernel@gmail.com> wrote:

> >  #include <asm/mpspec.h>
> >  #include <asm/setup.h>
> >  #include <asm/arch_hooks.h>
> > +#include <asm/efi.h>
> >  #include <asm/sections.h>
> >  #include <asm/dmi.h>
> >  #include <asm/io_apic.h>
> 
> checkpatch.pl want me to remove that.

checkpatch.pl is right about that, but the proper fix would be to decode 
this dependency - why isnt it enough to include linux/efi.h?

	Ingo

Re: [PATCH] x86: clean up ARCH_SETUP

[Ingo Molnar]

* Yinghai Lu <yhlu.kernel@gmail.com> wrote:

> asm-x86/paravirt.h already have protection with CONFIG_PARAVIRT inside

applied to tip/x86/unify-setup - thanks Yinghai.

	Ingo

Re: [PATCH] x86: move fix mapping page table range early

[Ingo Molnar]

* Yinghai Lu <yhlu.kernel@gmail.com> wrote:

> do that in init_memory_mapping
> 
> also remove one init_ohci1394_dma_on_all_controllers

applied to tip/x86/unify-setup - thanks Yinghai.

	Ingo

Re: [PATCH 00/16] x86: merge setup_32/64.c

[Huang, Ying]

On Thu, 2008-06-26 at 13:01 +0200, Ingo Molnar wrote:
> * Yinghai Lu <yhlu.kernel@gmail.com> wrote:
> 
> > >  #include <asm/mpspec.h>
> > >  #include <asm/setup.h>
> > >  #include <asm/arch_hooks.h>
> > > +#include <asm/efi.h>
> > >  #include <asm/sections.h>
> > >  #include <asm/dmi.h>
> > >  #include <asm/io_apic.h>
> > 
> > checkpatch.pl want me to remove that.
> 
> checkpatch.pl is right about that, but the proper fix would be to decode 
> this dependency - why isnt it enough to include linux/efi.h?

There is one dependency for asm/efi.h: declaration of
efi_reserve_early(). This declaration can be moved to linux/efi.h, maybe
after being renamed to efi_reserve_memmap().

Best Regards,
Huang Ying

[PATCH] x86: early res print out alignment

[Yinghai Lu]

Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>

---
 arch/x86/kernel/e820.c |   20 +++++++++++++++-----
 1 file changed, 15 insertions(+), 5 deletions(-)

Index: linux-2.6/arch/x86/kernel/e820.c
===================================================================
--- linux-2.6.orig/arch/x86/kernel/e820.c
+++ 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);
 	}

[PATCH] x86: let setup_arch call init_apic_mappings for 32bit

[Yinghai Lu]

instead of calling it from trap_init()

also move init ioapic mapping out of apic_32.c

so 32 bit do same as 64 bit

Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>

---
 arch/x86/kernel/apic_32.c    |   30 ------------------------------
 arch/x86/kernel/io_apic_32.c |   32 ++++++++++++++++++++++++++++++++
 arch/x86/kernel/setup.c      |    6 ++----
 arch/x86/kernel/traps_32.c   |    4 ----
 include/asm-x86/apic.h       |    1 +
 include/asm-x86/io_apic.h    |    1 +
 6 files changed, 36 insertions(+), 38 deletions(-)

Index: linux-2.6/arch/x86/kernel/apic_32.c
===================================================================
--- linux-2.6.orig/arch/x86/kernel/apic_32.c
+++ 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
===================================================================
--- linux-2.6.orig/arch/x86/kernel/io_apic_32.c
+++ 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
===================================================================
--- linux-2.6.orig/arch/x86/kernel/setup.c
+++ 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
===================================================================
--- linux-2.6.orig/arch/x86/kernel/traps_32.c
+++ 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
===================================================================
--- linux-2.6.orig/include/asm-x86/apic.h
+++ 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
===================================================================
--- linux-2.6.orig/include/asm-x86/io_apic.h
+++ 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

[PATCH] x86: early res print out alignment v2

[Yinghai Lu]

v2: fix print info to cont
    [ - ] change to xxx - xxx

Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>

---
 arch/x86/kernel/e820.c |   20 +++++++++++++++-----
 1 file changed, 15 insertions(+), 5 deletions(-)

Index: linux-2.6/arch/x86/kernel/e820.c
===================================================================
--- linux-2.6.orig/arch/x86/kernel/e820.c
+++ 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);
 	}

[PATCH] x86: fix init_memory_mapping over boundary

[Yinghai Lu]

some end boundary is only page alignment, instead of 2M alignment,
so call ker_phycial_mapping_init three times.
then don't overmap above the max_low_pfn

Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>

---
 arch/x86/mm/init_32.c |   74 +++++++++++++++++++++++++++++++++-----------------
 1 file changed, 50 insertions(+), 24 deletions(-)

Index: linux-2.6/arch/x86/mm/init_32.c
===================================================================
--- linux-2.6.orig/arch/x86/mm/init_32.c
+++ 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);
 

[PATCH] x86: fix init_memory_mapping over boundary v2

[Yinghai Lu]

some end boundary is only page alignment, instead of 2M alignment,
so call ker_phycial_mapping_init three times.
then don't overmap above the max_low_pfn

v2: make init_memory_mapping more solid: start could be any value other than 0

Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>

---
 arch/x86/mm/init_32.c |   86 ++++++++++++++++++++++++++++++++++++--------------
 1 file changed, 62 insertions(+), 24 deletions(-)

Index: linux-2.6/arch/x86/mm/init_32.c
===================================================================
--- linux-2.6.orig/arch/x86/mm/init_32.c
+++ 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);
 

Re: [PATCH] x86: fix init_memory_mapping over boundary v2

[Ingo Molnar]

* Yinghai Lu <yhlu.kernel@gmail.com> wrote:

> some end boundary is only page alignment, instead of 2M alignment, so 
> call ker_phycial_mapping_init three times. then don't overmap above 
> the max_low_pfn

applied to tip/x86/unify-setup, thanks Yinghai.

> v2: make init_memory_mapping more solid: start could be any value 
> other than 0

very good idea!

	Ingo

Re: [PATCH] x86: early res print out alignment v2

[Ingo Molnar]

* Yinghai Lu <yhlu.kernel@gmail.com> wrote:

> v2: fix print info to cont

applied to tip/x86/unify-setup.

>     [ - ] change to xxx - xxx

hm, i changed this back to "[ xxx - yyy ]" - i think it's slightly more 
structured hence more readable.

	Ingo

Re: [PATCH] x86: let setup_arch call init_apic_mappings for 32bit

[Ingo Molnar]

* Yinghai Lu <yhlu.kernel@gmail.com> wrote:

> instead of calling it from trap_init()
> 
> also move init ioapic mapping out of apic_32.c
> 
> so 32 bit do same as 64 bit

applied to tip/x86/unify-setup - thanks Yinghai.

	Ingo

Re: [PATCH] x86: fix init_memory_mapping over boundary v2

[Yinghai Lu]

On Fri, Jun 27, 2008 at 11:35 PM, Ingo Molnar <mingo@elte.hu> wrote:
>
> * Yinghai Lu <yhlu.kernel@gmail.com> wrote:
>
>> some end boundary is only page alignment, instead of 2M alignment, so
>> call ker_phycial_mapping_init three times. then don't overmap above
>> the max_low_pfn
>
> applied to tip/x86/unify-setup, thanks Yinghai.
>
>> v2: make init_memory_mapping more solid: start could be any value
>> other than 0
>
> very good idea!

could change
map_pfn_mapped = init_memory_mapping(0, map_low_pfn<<PAGE_SHIFT);
to
map_pfn_mapped = init_memory_mapping(map_pfn_mapped<<PAGE_SHIFT,
map_low_pfn<<PAGE_SHIFT);

only concern is
init_memory_mapping set   _EXEC in page table for kernel code.

YH

Re: [PATCH] x86: fix init_memory_mapping over boundary v2

[Ingo Molnar]

* Ingo Molnar <mingo@elte.hu> wrote:

> * Yinghai Lu <yhlu.kernel@gmail.com> wrote:
> 
> > some end boundary is only page alignment, instead of 2M alignment, so 
> > call ker_phycial_mapping_init three times. then don't overmap above 
> > the max_low_pfn
> 
> applied to tip/x86/unify-setup, thanks Yinghai.

-tip testing found a 32-bit spontaneous reboot during bootup, with this 
config:

  http://redhat.com/~mingo/misc/config-Sat_Jun_28_08_43_52_CEST_2008.bad

no serial log produced at all.

doing the reverts below fixes the problem.

	Ingo

commit 4ae56755fa1e1bd1500c97b5bf82d6519a261ffd
Author: Ingo Molnar <mingo@elte.hu>
Date:   Sat Jun 28 08:59:44 2008 +0200

    Revert "x86: let setup_arch call init_apic_mappings for 32bit"

    This reverts commit 7fce3bb20100e5b6c42ad66c495e6348522bfd37.

commit 2703b959fc588f1b77290272f60bcf0bd8f6e557
Author: Ingo Molnar <mingo@elte.hu>
Date:   Sat Jun 28 08:59:39 2008 +0200

    Revert "x86: fix init_memory_mapping over boundary v2"

    This reverts commit 7a41ed2c98f860da38a944d8de5b29e6ec39cf76.

Re: [PATCH] x86: fix init_memory_mapping over boundary v2

[Yinghai Lu]

On Sat, Jun 28, 2008 at 12:02 AM, Ingo Molnar <mingo@elte.hu> wrote:
>
> * Ingo Molnar <mingo@elte.hu> wrote:
>
>> * Yinghai Lu <yhlu.kernel@gmail.com> wrote:
>>
>> > some end boundary is only page alignment, instead of 2M alignment, so
>> > call ker_phycial_mapping_init three times. then don't overmap above
>> > the max_low_pfn
>>
>> applied to tip/x86/unify-setup, thanks Yinghai.
>
> -tip testing found a 32-bit spontaneous reboot during bootup, with this
> config:
>
>  http://redhat.com/~mingo/misc/config-Sat_Jun_28_08_43_52_CEST_2008.bad
>
> no serial log produced at all.
>
> doing the reverts below fixes the problem.
>
>        Ingo
>
> commit 4ae56755fa1e1bd1500c97b5bf82d6519a261ffd
> Author: Ingo Molnar <mingo@elte.hu>
> Date:   Sat Jun 28 08:59:44 2008 +0200
>
>    Revert "x86: let setup_arch call init_apic_mappings for 32bit"
>
>    This reverts commit 7fce3bb20100e5b6c42ad66c495e6348522bfd37.
>
> commit 2703b959fc588f1b77290272f60bcf0bd8f6e557
> Author: Ingo Molnar <mingo@elte.hu>
> Date:   Sat Jun 28 08:59:39 2008 +0200
>
>    Revert "x86: fix init_memory_mapping over boundary v2"
>
>    This reverts commit 7a41ed2c98f860da38a944d8de5b29e6ec39cf76.
>
the system with 512M ?

YH

Re: [PATCH] x86: fix init_memory_mapping over boundary v2

[Ingo Molnar]

* Yinghai Lu <yhlu.kernel@gmail.com> wrote:

> On Sat, Jun 28, 2008 at 12:02 AM, Ingo Molnar <mingo@elte.hu> wrote:
> >
> > * Ingo Molnar <mingo@elte.hu> wrote:
> >
> >> * Yinghai Lu <yhlu.kernel@gmail.com> wrote:
> >>
> >> > some end boundary is only page alignment, instead of 2M alignment, so
> >> > call ker_phycial_mapping_init three times. then don't overmap above
> >> > the max_low_pfn
> >>
> >> applied to tip/x86/unify-setup, thanks Yinghai.
> >
> > -tip testing found a 32-bit spontaneous reboot during bootup, with this
> > config:
> >
> >  http://redhat.com/~mingo/misc/config-Sat_Jun_28_08_43_52_CEST_2008.bad
> >
> > no serial log produced at all.
> >
> > doing the reverts below fixes the problem.
> >
> >        Ingo
> >
> > commit 4ae56755fa1e1bd1500c97b5bf82d6519a261ffd
> > Author: Ingo Molnar <mingo@elte.hu>
> > Date:   Sat Jun 28 08:59:44 2008 +0200
> >
> >    Revert "x86: let setup_arch call init_apic_mappings for 32bit"
> >
> >    This reverts commit 7fce3bb20100e5b6c42ad66c495e6348522bfd37.
> >
> > commit 2703b959fc588f1b77290272f60bcf0bd8f6e557
> > Author: Ingo Molnar <mingo@elte.hu>
> > Date:   Sat Jun 28 08:59:39 2008 +0200
> >
> >    Revert "x86: fix init_memory_mapping over boundary v2"
> >
> >    This reverts commit 7a41ed2c98f860da38a944d8de5b29e6ec39cf76.
> >
> the system with 512M ?

no, other 32-bit testsystems: 1GB RAM, 2GB RAM, etc.

	Ingo

Re: [PATCH] x86: fix init_memory_mapping over boundary v2

[Yinghai Lu]

On Sat, Jun 28, 2008 at 12:02 AM, Ingo Molnar <mingo@elte.hu> wrote:
>
> * Ingo Molnar <mingo@elte.hu> wrote:
>
>> * Yinghai Lu <yhlu.kernel@gmail.com> wrote:
>>
>> > some end boundary is only page alignment, instead of 2M alignment, so
>> > call ker_phycial_mapping_init three times. then don't overmap above
>> > the max_low_pfn
>>
>> applied to tip/x86/unify-setup, thanks Yinghai.
>
> -tip testing found a 32-bit spontaneous reboot during bootup, with this
> config:
>
>  http://redhat.com/~mingo/misc/config-Sat_Jun_28_08_43_52_CEST_2008.bad
>
> no serial log produced at all.
even with console=uart8250,io,0x03f8,115200n8 ?

YH

Re: [PATCH] x86: fix init_memory_mapping over boundary v2

[Yinghai Lu]

On Sat, Jun 28, 2008 at 12:21 AM, Ingo Molnar <mingo@elte.hu> wrote:
>
> * Yinghai Lu <yhlu.kernel@gmail.com> wrote:
>
>> On Sat, Jun 28, 2008 at 12:02 AM, Ingo Molnar <mingo@elte.hu> wrote:
>> >
>> > * Ingo Molnar <mingo@elte.hu> wrote:
>> >
>> >> * Yinghai Lu <yhlu.kernel@gmail.com> wrote:
>> >>
>> >> > some end boundary is only page alignment, instead of 2M alignment, so
>> >> > call ker_phycial_mapping_init three times. then don't overmap above
>> >> > the max_low_pfn
>> >>
>> >> applied to tip/x86/unify-setup, thanks Yinghai.
>> >
>> > -tip testing found a 32-bit spontaneous reboot during bootup, with this
>> > config:
>> >
>> >  http://redhat.com/~mingo/misc/config-Sat_Jun_28_08_43_52_CEST_2008.bad
>> >
>> > no serial log produced at all.
>> >
>> > doing the reverts below fixes the problem.
>> >
>> >        Ingo
>> >
>> > commit 4ae56755fa1e1bd1500c97b5bf82d6519a261ffd
>> > Author: Ingo Molnar <mingo@elte.hu>
>> > Date:   Sat Jun 28 08:59:44 2008 +0200
>> >
>> >    Revert "x86: let setup_arch call init_apic_mappings for 32bit"
>> >
>> >    This reverts commit 7fce3bb20100e5b6c42ad66c495e6348522bfd37.
>> >
>> > commit 2703b959fc588f1b77290272f60bcf0bd8f6e557
>> > Author: Ingo Molnar <mingo@elte.hu>
>> > Date:   Sat Jun 28 08:59:39 2008 +0200
>> >
>> >    Revert "x86: fix init_memory_mapping over boundary v2"
>> >
>> >    This reverts commit 7a41ed2c98f860da38a944d8de5b29e6ec39cf76.
>> >
>> the system with 512M ?
>
> no, other 32-bit testsystems: 1GB RAM, 2GB RAM, etc.

good, i duplicate that one server with your config.

YH

Re: [PATCH] x86: fix init_memory_mapping over boundary v2

[Ingo Molnar]

* Yinghai Lu <yhlu.kernel@gmail.com> wrote:

> On Sat, Jun 28, 2008 at 12:02 AM, Ingo Molnar <mingo@elte.hu> wrote:
> >
> > * Ingo Molnar <mingo@elte.hu> wrote:
> >
> >> * Yinghai Lu <yhlu.kernel@gmail.com> wrote:
> >>
> >> > some end boundary is only page alignment, instead of 2M alignment, so
> >> > call ker_phycial_mapping_init three times. then don't overmap above
> >> > the max_low_pfn
> >>
> >> applied to tip/x86/unify-setup, thanks Yinghai.
> >
> > -tip testing found a 32-bit spontaneous reboot during bootup, with this
> > config:
> >
> >  http://redhat.com/~mingo/misc/config-Sat_Jun_28_08_43_52_CEST_2008.bad
> >
> > no serial log produced at all.
> even with console=uart8250,io,0x03f8,115200n8 ?

correct, i use earlyprintk=serial,ttyS0,115200.

i just re-tested it - doing "git merge tip/x86/unify-setup" in the 
tip/master branch triggers the failure with that config on a wide range 
of 32-bit boxes.

	Ingo

Re: [PATCH] x86: fix init_memory_mapping over boundary v2

[Yinghai Lu]

[-- Attachment #1: Type: text/plain, Size: 1080 bytes --]

On Sat, Jun 28, 2008 at 12:47 AM, Ingo Molnar <mingo@elte.hu> wrote:
>
> * Yinghai Lu <yhlu.kernel@gmail.com> wrote:
>
>> On Sat, Jun 28, 2008 at 12:02 AM, Ingo Molnar <mingo@elte.hu> wrote:
>> >
>> > * Ingo Molnar <mingo@elte.hu> wrote:
>> >
>> >> * Yinghai Lu <yhlu.kernel@gmail.com> wrote:
>> >>
>> >> > some end boundary is only page alignment, instead of 2M alignment, so
>> >> > call ker_phycial_mapping_init three times. then don't overmap above
>> >> > the max_low_pfn
>> >>
>> >> applied to tip/x86/unify-setup, thanks Yinghai.
>> >
>> > -tip testing found a 32-bit spontaneous reboot during bootup, with this
>> > config:
>> >
>> >  http://redhat.com/~mingo/misc/config-Sat_Jun_28_08_43_52_CEST_2008.bad
>> >
>> > no serial log produced at all.
>> even with console=uart8250,io,0x03f8,115200n8 ?
>
> correct, i use earlyprintk=serial,ttyS0,115200.
>
> i just re-tested it - doing "git merge tip/x86/unify-setup" in the
> tip/master branch triggers the failure with that config on a wide range
> of 32-bit boxes.

please check attached v3..., that is less intrusive.

YH

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: fix_over_mapped.patch --]
[-- Type: text/x-patch; name=fix_over_mapped.patch, Size: 2810 bytes --]

[PATCH] x86: fix init_memory_mapping over boundary v3

some end boundary is only page alignment, instead of 2M alignment,

v2: make init_memory_mapping more solid: start could be any value other than 0
v3: fix NON PAE by handling left over in kernel_physical_mapping

Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>

---
 arch/x86/mm/init_32.c |   24 ++++++++++++++----------
 1 file changed, 14 insertions(+), 10 deletions(-)

Index: linux-2.6/arch/x86/mm/init_32.c
===================================================================
--- linux-2.6.orig/arch/x86/mm/init_32.c
+++ 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

Re: [PATCH] x86: fix init_memory_mapping over boundary v2

[Ingo Molnar]

* Yinghai Lu <yhlu.kernel@gmail.com> wrote:

> > i just re-tested it - doing "git merge tip/x86/unify-setup" in the 
> > tip/master branch triggers the failure with that config on a wide 
> > range of 32-bit boxes.
> 
> please check attached v3..., that is less intrusive.
> 
> some end boundary is only page alignment, instead of 2M alignment,
> 
> v2: make init_memory_mapping more solid: start could be any value other than 0
> v3: fix NON PAE by handling left over in kernel_physical_mapping

ok, this one seems to be working better.

	Ingo

Re: [PATCH] x86: fix init_memory_mapping over boundary v2

[Jeremy Fitzhardinge]

Yinghai Lu wrote:
> some end boundary is only page alignment, instead of 2M alignment,
> so call ker_phycial_mapping_init three times.
> then don't overmap above the max_low_pfn
>
> v2: make init_memory_mapping more solid: start could be any value other than 0
>
> Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>
>
> ---
>  arch/x86/mm/init_32.c |   86 ++++++++++++++++++++++++++++++++++++--------------
>  1 file changed, 62 insertions(+), 24 deletions(-)
>
> Index: linux-2.6/arch/x86/mm/init_32.c
> ===================================================================
> --- linux-2.6.orig/arch/x86/mm/init_32.c
> +++ 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)
>   

Better to use bool for 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;
>   

What are the semantics of the use_pse argument?  Is it just a hint?  If 
so, then the caller below should just pass a constant 1/true.

>  
> +	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);
>   

That's a bit magic.  Isn't that just "extra = end & ((1 << 21) - 1)"?  
And what are you actually computing here anyway?

> +		extra += (2<<20);
>   

PMD_SIZE?  The comment below says "2/4MB", which corresponds to a 
PAE/non-PAE large PMD.

> +		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;
>   

PMD_SIZE again?  Non-PAE big pages are 4MB.

> +	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);
>   

More magic.

> +	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);
> +	}

I think this code could do with a dose of comments and some symbolic 
constants.  And perhaps the tactical use of round_up/down.

But it's a nice simplification of kernel_physical_mapping_init().

    J

Re: [PATCH] x86: fix init_memory_mapping over boundary v2

[Yinghai Lu]

On Sat, Jun 28, 2008 at 10:38 AM, Jeremy Fitzhardinge <jeremy@goop.org> wrote:
> Yinghai Lu wrote:
>>
>> some end boundary is only page alignment, instead of 2M alignment,
>> so call ker_phycial_mapping_init three times.
>> then don't overmap above the max_low_pfn
>>
>> v2: make init_memory_mapping more solid: start could be any value other
>> than 0
>>
>> Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>
>>
>> ---
>>  arch/x86/mm/init_32.c |   86
>> ++++++++++++++++++++++++++++++++++++--------------
>>  1 file changed, 62 insertions(+), 24 deletions(-)
>>
>> Index: linux-2.6/arch/x86/mm/init_32.c
>> ===================================================================
>> --- linux-2.6.orig/arch/x86/mm/init_32.c
>> +++ 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)
>>
>
> Better to use bool for 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;
>>
>
> What are the semantics of the use_pse argument?  Is it just a hint?  If so,
> then the caller below should just pass a constant 1/true.
>
>>  +       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);
>>
>
> That's a bit magic.  Isn't that just "extra = end & ((1 << 21) - 1)"?  And
> what are you actually computing here anyway?
>
>> +               extra += (2<<20);
>>
>
> PMD_SIZE?  The comment below says "2/4MB", which corresponds to a
> PAE/non-PAE large PMD.
>
>> +               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;
>>
>
> PMD_SIZE again?  Non-PAE big pages are 4MB.

thanks, I will try to revive v2...

[PATCH] x86: fix init_memory_mapping over boundary v4

[Yinghai Lu]

some end boundary is only page alignment, instead of 2M alignment,
so call ker_phycial_mapping_init three times.
then don't overmap above the max_low_pfn

v2: make init_memory_mapping more solid: start could be any value other than 0
v3: fix NON PAE by handling left over in kernel_physical_mapping
v4: revert back to v2, and use PMD_SHIFT to calculate boundary
    also adjust size for pre-allocated table size

Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>

---
 arch/x86/mm/init_32.c |  101 +++++++++++++++++++++++++++++++++++++-------------
 1 file changed, 75 insertions(+), 26 deletions(-)

Index: linux-2.6/arch/x86/mm/init_32.c
===================================================================
--- linux-2.6.orig/arch/x86/mm/init_32.c
+++ 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);
 

[PATCH] x86: fix warning in e820_reserve_resources with 32bit

[Yinghai Lu]

when 64bit resource is not enabled, got:

arch/x86/kernel/e820.c: In function ‘e820_reserve_resources’:
arch/x86/kernel/e820.c:1217: warning: comparison is always false due to limited range of data type

because res->start/end is resource_t aka u32. it will overflow.

fix it with temp end of u64

Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>

diff --git a/arch/x86/kernel/e820.c b/arch/x86/kernel/e820.c
index fa77cb4..ba5ac88 100644
--- a/arch/x86/kernel/e820.c
+++ b/arch/x86/kernel/e820.c
@@ -1202,6 +1202,7 @@ void __init e820_reserve_resources(void)
 {
 	int i;
 	struct resource *res;
+	u64 end;
 
 	res = alloc_bootmem_low(sizeof(struct resource) * e820.nr_map);
 	for (i = 0; i < e820.nr_map; i++) {
@@ -1211,14 +1212,16 @@ void __init e820_reserve_resources(void)
 		case E820_NVS:	res->name = "ACPI Non-volatile Storage"; break;
 		default:	res->name = "reserved";
 		}
-		res->start = e820.map[i].addr;
-		res->end = res->start + e820.map[i].size - 1;
+		end = e820.map[i].addr + e820.map[i].size - 1;
 #ifndef CONFIG_RESOURCES_64BIT
-		if (res->end > 0x100000000ULL) {
+		if (end > 0x100000000ULL) {
 			res++;
 			continue;
 		}
 #endif
+		res->start = e820.map[i].addr;
+		res->end = end;
+
 		res->flags = IORESOURCE_MEM | IORESOURCE_BUSY;
 		insert_resource(&iomem_resource, res);
 		res++;

Re: [PATCH] x86: fix init_memory_mapping over boundary v4

[Jeremy Fitzhardinge]

Yinghai Lu wrote:
> some end boundary is only page alignment, instead of 2M alignment,
> so call ker_phycial_mapping_init three times.
> then don't overmap above the max_low_pfn
>
> v2: make init_memory_mapping more solid: start could be any value other than 0
> v3: fix NON PAE by handling left over in kernel_physical_mapping
> v4: revert back to v2, and use PMD_SHIFT to calculate boundary
>     also adjust size for pre-allocated table size
>
> Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>
>
> ---
>  arch/x86/mm/init_32.c |  101 +++++++++++++++++++++++++++++++++++++-------------
>  1 file changed, 75 insertions(+), 26 deletions(-)
>
> Index: linux-2.6/arch/x86/mm/init_32.c
> ===================================================================
> --- linux-2.6.orig/arch/x86/mm/init_32.c
> +++ 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
>   

We should really define X_index() to return 0 for folded pagetable level 
X.  I've had to put similar #ifdefs in other code.  Also, for the 
purposes of making this code more unifiable, it's probably better to 
test for "#if PAGETABLE_LEVELS >= 3" rather than for PAE specifically.

> +		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;
>   

Is this the same as "extra = (end + PMD_SIZE + 1) & PMD_MASK"?

> +		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;
>   

Why not correct for fixmaps by putting "ptes += 
__end_of_fixed_addresses;" above?

>  	/*
>  	 * 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);
>  
>   

    J

Re: [PATCH] x86: fix init_memory_mapping over boundary v4

[Yinghai Lu]

On Sat, Jun 28, 2008 at 7:22 PM, Jeremy Fitzhardinge <jeremy@goop.org> wrote:
> Yinghai Lu wrote:
>>
>> some end boundary is only page alignment, instead of 2M alignment,
>> so call ker_phycial_mapping_init three times.
>> then don't overmap above the max_low_pfn
>>
>> v2: make init_memory_mapping more solid: start could be any value other
>> than 0
>> v3: fix NON PAE by handling left over in kernel_physical_mapping
>> v4: revert back to v2, and use PMD_SHIFT to calculate boundary
>>    also adjust size for pre-allocated table size
>>
>> Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>
>>
>> ---
>>  arch/x86/mm/init_32.c |  101
>> +++++++++++++++++++++++++++++++++++++-------------
>>  1 file changed, 75 insertions(+), 26 deletions(-)
>>
>> Index: linux-2.6/arch/x86/mm/init_32.c
>> ===================================================================
>> --- linux-2.6.orig/arch/x86/mm/init_32.c
>> +++ 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
>>
>
> We should really define X_index() to return 0 for folded pagetable level X.
>  I've had to put similar #ifdefs in other code.  Also, for the purposes of
> making this code more unifiable, it's probably better to test for "#if
> PAGETABLE_LEVELS >= 3" rather than for PAE specifically.

does all cpu support >=3?

>
>> +               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);
                      left over at tail
>> +               extra += PMD_SIZE;
                      head...
>>
>
> Is this the same as "extra = (end + PMD_SIZE + 1) & PMD_MASK"?

extra = end & PMD_MASK;
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;
>>
>
> Why not correct for fixmaps by putting "ptes += __end_of_fixed_addresses;"
> above?

yes

YH

Re: [PATCH] x86: early res print out alignment v2

[Yinghai Lu]

On Fri, Jun 27, 2008 at 11:36 PM, Ingo Molnar <mingo@elte.hu> wrote:
>
> * Yinghai Lu <yhlu.kernel@gmail.com> wrote:
>
>> v2: fix print info to cont
>
> applied to tip/x86/unify-setup.
>
>>     [ - ] change to xxx - xxx
>
> hm, i changed this back to "[ xxx - yyy ]" - i think it's slightly more
> structured hence more readable.

should change [ xxx - yyy ] to [xxx - yyy]
otherwise print out width is bigger than 80

YH

Re: [PATCH] x86: fix init_memory_mapping over boundary v4

[Jeremy Fitzhardinge]

Yinghai Lu wrote:
>> We should really define X_index() to return 0 for folded pagetable level X.
>>  I've had to put similar #ifdefs in other code.  Also, for the purposes of
>> making this code more unifiable, it's probably better to test for "#if
>> PAGETABLE_LEVELS >= 3" rather than for PAE specifically.
>>     
>
> does all cpu support >=3?
>   

Well, x86 32-bit PAE is 3 level, and 64-bit is 4 level; both have a 
meaningful set of pmd operations.

>
>> Is this the same as "extra = (end + PMD_SIZE + 1) & PMD_MASK"?
>>     
>
> extra = end & PMD_MASK;
> extra+ = PMD_SIZE;
>   

OK.  I wasn't sure if your intent was to round up to the next PMD_SIZE 
or add a whole new PMD_SIZE.

    J

Re: [PATCH] x86: fix init_memory_mapping over boundary v4

[Ingo Molnar]

* Yinghai Lu <yhlu.kernel@gmail.com> wrote:

> some end boundary is only page alignment, instead of 2M alignment,
> so call ker_phycial_mapping_init three times.
> then don't overmap above the max_low_pfn
> 
> v2: make init_memory_mapping more solid: start could be any value other than 0
> v3: fix NON PAE by handling left over in kernel_physical_mapping
> v4: revert back to v2, and use PMD_SHIFT to calculate boundary
>     also adjust size for pre-allocated table size

i've got v3 applied to tip/master and it worked well in testing - on the 
old, 512m RAM problem laptop too. So could you send v4 as a delta patch 
instead please, against tip/master?

	Ingo

Re: [PATCH] x86: fix init_memory_mapping over boundary v4

[Yinghai Lu]

On Sun, Jun 29, 2008 at 12:08 AM, Ingo Molnar <mingo@elte.hu> wrote:
>
> * Yinghai Lu <yhlu.kernel@gmail.com> wrote:
>
>> some end boundary is only page alignment, instead of 2M alignment,
>> so call ker_phycial_mapping_init three times.
>> then don't overmap above the max_low_pfn
>>
>> v2: make init_memory_mapping more solid: start could be any value other than 0
>> v3: fix NON PAE by handling left over in kernel_physical_mapping
>> v4: revert back to v2, and use PMD_SHIFT to calculate boundary
>>     also adjust size for pre-allocated table size
>
> i've got v3 applied to tip/master and it worked well in testing - on the
> old, 512m RAM problem laptop too. So could you send v4 as a delta patch
> instead please, against tip/master?

will. before that delta patch please do test on your 512M system .

YH

[PATCH] x86: fix init_memory_mapping over boundary v4 - diff to v3

[Yinghai Lu]

some end boundary is only page alignment, instead of 2M alignment,
so call ker_phycial_mapping_init three times.
then don't overmap above the max_low_pfn

v2: make init_memory_mapping more solid: start could be any value other than 0
v3: fix NON PAE by handling left over in kernel_physical_mapping
v4: revert back to v2, and use PMD_SHIFT to calculate boundary
    also adjust size for pre-allocated table size

Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>

---
 arch/x86/mm/init_32.c |   89 +++++++++++++++++++++++++++++++++++++-------------
 1 file changed, 67 insertions(+), 22 deletions(-)

Index: linux-2.6/arch/x86/mm/init_32.c
===================================================================
--- linux-2.6.orig/arch/x86/mm/init_32.c
+++ 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.
@@ -794,7 +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);
 

Re: [PATCH] x86: fix init_memory_mapping over boundary v4 - diff to v3

[Ingo Molnar]

* Yinghai Lu <yhlu.kernel@gmail.com> wrote:

> some end boundary is only page alignment, instead of 2M alignment,
> so call ker_phycial_mapping_init three times.
> then don't overmap above the max_low_pfn
> 
> v2: make init_memory_mapping more solid: start could be any value other than 0
> v3: fix NON PAE by handling left over in kernel_physical_mapping
> v4: revert back to v2, and use PMD_SHIFT to calculate boundary
>     also adjust size for pre-allocated table size

applied, thanks Yinghai.

	Ingo

Re: [PATCH] x86: fix warning in e820_reserve_resources with 32bit

[Ingo Molnar]

* Yinghai Lu <yhlu.kernel@gmail.com> wrote:

> when 64bit resource is not enabled, got:
> 
> arch/x86/kernel/e820.c: In function ‘e820_reserve_resources’:
> arch/x86/kernel/e820.c:1217: warning: comparison is always false due to limited range of data type
> 
> because res->start/end is resource_t aka u32. it will overflow.
> 
> fix it with temp end of u64

applied, thanks.

	Ingo

Re: [PATCH] x86: fix init_memory_mapping over boundary v4

[Ingo Molnar]

* Yinghai Lu <yhlu.kernel@gmail.com> wrote:

> On Sun, Jun 29, 2008 at 12:08 AM, Ingo Molnar <mingo@elte.hu> wrote:
> >
> > * Yinghai Lu <yhlu.kernel@gmail.com> wrote:
> >
> >> some end boundary is only page alignment, instead of 2M alignment,
> >> so call ker_phycial_mapping_init three times.
> >> then don't overmap above the max_low_pfn
> >>
> >> v2: make init_memory_mapping more solid: start could be any value other than 0
> >> v3: fix NON PAE by handling left over in kernel_physical_mapping
> >> v4: revert back to v2, and use PMD_SHIFT to calculate boundary
> >>     also adjust size for pre-allocated table size
> >
> > i've got v3 applied to tip/master and it worked well in testing - on the
> > old, 512m RAM problem laptop too. So could you send v4 as a delta patch
> > instead please, against tip/master?
> 
> will. before that delta patch please do test on your 512M system .

i've tested it thoroughly on all systems over the weekend and it's all 
fine now. Thanks,

	Ingo