8 files changed, 116 insertions, 205 deletions

diff --git a/arch/x86/kernel/acpi/boot.c b/arch/x86/kernel/acpi/boot.c
index bba162c81d5b..956c1dee6fbe 100644
--- a/arch/x86/kernel/acpi/boot.c
+++ b/arch/x86/kernel/acpi/boot.c
@@ -108,35 +108,18 @@ enum acpi_irq_model_id acpi_irq_model = ACPI_IRQ_MODEL_PIC;
  */
 char *__init __acpi_map_table(unsigned long phys, unsigned long size)
 {
-        unsigned long base, offset, mapped_size;
-        int idx;
 
         if (!phys || !size)
                 return NULL;
 
-        if (phys+size <= (max_low_pfn_mapped << PAGE_SHIFT))
-                return __va(phys);
-
-        offset = phys & (PAGE_SIZE - 1);
-        mapped_size = PAGE_SIZE - offset;
-        clear_fixmap(FIX_ACPI_END);
-        set_fixmap(FIX_ACPI_END, phys);
-        base = fix_to_virt(FIX_ACPI_END);
-
-        /*
-         * Most cases can be covered by the below.
-         */
-        idx = FIX_ACPI_END;
-        while (mapped_size < size) {
-                if (--idx < FIX_ACPI_BEGIN)
-                        return NULL;    /* cannot handle this */
-                phys += PAGE_SIZE;
-                clear_fixmap(idx);
-                set_fixmap(idx, phys);
-                mapped_size += PAGE_SIZE;
-        }
+        return early_ioremap(phys, size);
+}
+void __init __acpi_unmap_table(char *map, unsigned long size)
+{
+        if (!map || !size)
+                return;
 
-        return ((unsigned char *)base + offset);
+        early_iounmap(map, size);
 }
 
 #ifdef CONFIG_PCI_MMCONFIG
diff --git a/arch/x86/kernel/apic.c b/arch/x86/kernel/apic.c
index 8bd801db24d9..cf2ca19e62da 100644
--- a/arch/x86/kernel/apic.c
+++ b/arch/x86/kernel/apic.c
@@ -555,7 +555,8 @@ static void __init lapic_cal_handler(struct clock_event_device *dev)
         }
 }
 
-static int __init calibrate_by_pmtimer(long deltapm, long *delta)
+static int __init
+calibrate_by_pmtimer(long deltapm, long *delta, long *deltatsc)
 {
         const long pm_100ms = PMTMR_TICKS_PER_SEC / 10;
         const long pm_thresh = pm_100ms / 100;
@@ -566,7 +567,7 @@ static int __init calibrate_by_pmtimer(long deltapm, long *delta)
         return -1;
 #endif
 
-        apic_printk(APIC_VERBOSE, "... PM timer delta = %ld\n", deltapm);
+        apic_printk(APIC_VERBOSE, "... PM-Timer delta = %ld\n", deltapm);
 
         /* Check, if the PM timer is available */
         if (!deltapm)
@@ -576,19 +577,30 @@ static int __init calibrate_by_pmtimer(long deltapm, long *delta)
 
         if (deltapm > (pm_100ms - pm_thresh) &&
             deltapm < (pm_100ms + pm_thresh)) {
-                apic_printk(APIC_VERBOSE, "... PM timer result ok\n");
-        } else {
-                res = (((u64)deltapm) *  mult) >> 22;
-                do_div(res, 1000000);
-                pr_warning("APIC calibration not consistent "
-                        "with PM Timer: %ldms instead of 100ms\n",
-                        (long)res);
-                /* Correct the lapic counter value */
-                res = (((u64)(*delta)) * pm_100ms);
+                apic_printk(APIC_VERBOSE, "... PM-Timer result ok\n");
+                return 0;
+        }
+
+        res = (((u64)deltapm) *  mult) >> 22;
+        do_div(res, 1000000);
+        pr_warning("APIC calibration not consistent "
+                   "with PM-Timer: %ldms instead of 100ms\n",(long)res);
+
+        /* Correct the lapic counter value */
+        res = (((u64)(*delta)) * pm_100ms);
+        do_div(res, deltapm);
+        pr_info("APIC delta adjusted to PM-Timer: "
+                "%lu (%ld)\n", (unsigned long)res, *delta);
+        *delta = (long)res;
+
+        /* Correct the tsc counter value */
+        if (cpu_has_tsc) {
+                res = (((u64)(*deltatsc)) * pm_100ms);
                 do_div(res, deltapm);
-                pr_info("APIC delta adjusted to PM-Timer: "
-                        "%lu (%ld)\n", (unsigned long)res, *delta);
-                *delta = (long)res;
+                apic_printk(APIC_VERBOSE, "TSC delta adjusted to "
+                                          "PM-Timer: %lu (%ld) \n",
+                                        (unsigned long)res, *deltatsc);
+                *deltatsc = (long)res;
         }
 
         return 0;
@@ -599,7 +611,7 @@ static int __init calibrate_APIC_clock(void)
         struct clock_event_device *levt = &__get_cpu_var(lapic_events);
         void (*real_handler)(struct clock_event_device *dev);
         unsigned long deltaj;
-        long delta;
+        long delta, deltatsc;
         int pm_referenced = 0;
 
         local_irq_disable();
@@ -629,9 +641,11 @@ static int __init calibrate_APIC_clock(void)
         delta = lapic_cal_t1 - lapic_cal_t2;
         apic_printk(APIC_VERBOSE, "... lapic delta = %ld\n", delta);
 
+        deltatsc = (long)(lapic_cal_tsc2 - lapic_cal_tsc1);
+
         /* we trust the PM based calibration if possible */
         pm_referenced = !calibrate_by_pmtimer(lapic_cal_pm2 - lapic_cal_pm1,
-                                        &delta);
+                                        &delta, &deltatsc);
 
         /* Calculate the scaled math multiplication factor */
         lapic_clockevent.mult = div_sc(delta, TICK_NSEC * LAPIC_CAL_LOOPS,
@@ -649,11 +663,10 @@ static int __init calibrate_APIC_clock(void)
                     calibration_result);
 
         if (cpu_has_tsc) {
-                delta = (long)(lapic_cal_tsc2 - lapic_cal_tsc1);
                 apic_printk(APIC_VERBOSE, "..... CPU clock speed is "
                             "%ld.%04ld MHz.\n",
-                            (delta / LAPIC_CAL_LOOPS) / (1000000 / HZ),
-                            (delta / LAPIC_CAL_LOOPS) % (1000000 / HZ));
+                            (deltatsc / LAPIC_CAL_LOOPS) / (1000000 / HZ),
+                            (deltatsc / LAPIC_CAL_LOOPS) % (1000000 / HZ));
         }
 
         apic_printk(APIC_VERBOSE, "..... host bus clock speed is "
diff --git a/arch/x86/kernel/dumpstack.c b/arch/x86/kernel/dumpstack.c
index 6b1f6f6f8661..87d103ded1c3 100644
--- a/arch/x86/kernel/dumpstack.c
+++ b/arch/x86/kernel/dumpstack.c
@@ -99,7 +99,7 @@ print_context_stack(struct thread_info *tinfo,
                                 frame = frame->next_frame;
                                 bp = (unsigned long) frame;
                         } else {
-                                ops->address(data, addr, bp == 0);
+                                ops->address(data, addr, 0);
                         }
                         print_ftrace_graph_addr(addr, data, ops, tinfo, graph);
                 }
diff --git a/arch/x86/kernel/early_printk.c b/arch/x86/kernel/early_printk.c
index 504ad198e4ad..639ad98238a2 100644
--- a/arch/x86/kernel/early_printk.c
+++ b/arch/x86/kernel/early_printk.c
@@ -13,8 +13,8 @@
 #include <asm/setup.h>
 #include <xen/hvc-console.h>
 #include <asm/pci-direct.h>
-#include <asm/pgtable.h>
 #include <asm/fixmap.h>
+#include <asm/pgtable.h>
 #include <linux/usb/ehci_def.h>
 
 /* Simple VGA output */
diff --git a/arch/x86/kernel/es7000_32.c b/arch/x86/kernel/es7000_32.c
index d6184c12a182..55515d73d9c2 100644
--- a/arch/x86/kernel/es7000_32.c
+++ b/arch/x86/kernel/es7000_32.c
@@ -287,24 +287,31 @@ int __init find_unisys_acpi_oem_table(unsigned long *oem_addr)
 {
         struct acpi_table_header *header = NULL;
         int i = 0;
+        acpi_size tbl_size;
 
-        while (ACPI_SUCCESS(acpi_get_table("OEM1", i++, &header))) {
+        while (ACPI_SUCCESS(acpi_get_table_with_size("OEM1", i++, &header, &tbl_size))) {
                 if (!memcmp((char *) &header->oem_id, "UNISYS", 6)) {
                         struct oem_table *t = (struct oem_table *)header;
 
                         oem_addrX = t->OEMTableAddr;
                         oem_size = t->OEMTableSize;
+                        early_acpi_os_unmap_memory(header, tbl_size);
 
                         *oem_addr = (unsigned long)__acpi_map_table(oem_addrX,
                                                                     oem_size);
                         return 0;
                 }
+                early_acpi_os_unmap_memory(header, tbl_size);
         }
         return -1;
 }
 
 void __init unmap_unisys_acpi_oem_table(unsigned long oem_addr)
 {
+        if (!oem_addr)
+                return;
+
+        __acpi_unmap_table((char *)oem_addr, oem_size);
 }
 #endif
 
diff --git a/arch/x86/kernel/machine_kexec_64.c b/arch/x86/kernel/machine_kexec_64.c
index c43caa3a91f3..6993d51b7fd8 100644
--- a/arch/x86/kernel/machine_kexec_64.c
+++ b/arch/x86/kernel/machine_kexec_64.c
@@ -18,15 +18,6 @@
 #include <asm/mmu_context.h>
 #include <asm/io.h>
 
-#define PAGE_ALIGNED __attribute__ ((__aligned__(PAGE_SIZE)))
-static u64 kexec_pgd[512] PAGE_ALIGNED;
-static u64 kexec_pud0[512] PAGE_ALIGNED;
-static u64 kexec_pmd0[512] PAGE_ALIGNED;
-static u64 kexec_pte0[512] PAGE_ALIGNED;
-static u64 kexec_pud1[512] PAGE_ALIGNED;
-static u64 kexec_pmd1[512] PAGE_ALIGNED;
-static u64 kexec_pte1[512] PAGE_ALIGNED;
-
 static void init_level2_page(pmd_t *level2p, unsigned long addr)
 {
         unsigned long end_addr;
@@ -107,12 +98,65 @@ out:
         return result;
 }
 
+static void free_transition_pgtable(struct kimage *image)
+{
+        free_page((unsigned long)image->arch.pud);
+        free_page((unsigned long)image->arch.pmd);
+        free_page((unsigned long)image->arch.pte);
+}
+
+static int init_transition_pgtable(struct kimage *image, pgd_t *pgd)
+{
+        pud_t *pud;
+        pmd_t *pmd;
+        pte_t *pte;
+        unsigned long vaddr, paddr;
+        int result = -ENOMEM;
+
+        vaddr = (unsigned long)relocate_kernel;
+        paddr = __pa(page_address(image->control_code_page)+PAGE_SIZE);
+        pgd += pgd_index(vaddr);
+        if (!pgd_present(*pgd)) {
+                pud = (pud_t *)get_zeroed_page(GFP_KERNEL);
+                if (!pud)
+                        goto err;
+                image->arch.pud = pud;
+                set_pgd(pgd, __pgd(__pa(pud) | _KERNPG_TABLE));
+        }
+        pud = pud_offset(pgd, vaddr);
+        if (!pud_present(*pud)) {
+                pmd = (pmd_t *)get_zeroed_page(GFP_KERNEL);
+                if (!pmd)
+                        goto err;
+                image->arch.pmd = pmd;
+                set_pud(pud, __pud(__pa(pmd) | _KERNPG_TABLE));
+        }
+        pmd = pmd_offset(pud, vaddr);
+        if (!pmd_present(*pmd)) {
+                pte = (pte_t *)get_zeroed_page(GFP_KERNEL);
+                if (!pte)
+                        goto err;
+                image->arch.pte = pte;
+                set_pmd(pmd, __pmd(__pa(pte) | _KERNPG_TABLE));
+        }
+        pte = pte_offset_kernel(pmd, vaddr);
+        set_pte(pte, pfn_pte(paddr >> PAGE_SHIFT, PAGE_KERNEL_EXEC));
+        return 0;
+err:
+        free_transition_pgtable(image);
+        return result;
+}
+
 
 static int init_pgtable(struct kimage *image, unsigned long start_pgtable)
 {
         pgd_t *level4p;
+        int result;
         level4p = (pgd_t *)__va(start_pgtable);
-        return init_level4_page(image, level4p, 0, max_pfn << PAGE_SHIFT);
+        result = init_level4_page(image, level4p, 0, max_pfn << PAGE_SHIFT);
+        if (result)
+                return result;
+        return init_transition_pgtable(image, level4p);
 }
 
 static void set_idt(void *newidt, u16 limit)
@@ -174,7 +218,7 @@ int machine_kexec_prepare(struct kimage *image)
 
 void machine_kexec_cleanup(struct kimage *image)
 {
-        return;
+        free_transition_pgtable(image);
 }
 
 /*
@@ -195,22 +239,6 @@ void machine_kexec(struct kimage *image)
         memcpy(control_page, relocate_kernel, PAGE_SIZE);
 
         page_list[PA_CONTROL_PAGE] = virt_to_phys(control_page);
-        page_list[VA_CONTROL_PAGE] = (unsigned long)relocate_kernel;
-        page_list[PA_PGD] = virt_to_phys(&kexec_pgd);
-        page_list[VA_PGD] = (unsigned long)kexec_pgd;
-        page_list[PA_PUD_0] = virt_to_phys(&kexec_pud0);
-        page_list[VA_PUD_0] = (unsigned long)kexec_pud0;
-        page_list[PA_PMD_0] = virt_to_phys(&kexec_pmd0);
-        page_list[VA_PMD_0] = (unsigned long)kexec_pmd0;
-        page_list[PA_PTE_0] = virt_to_phys(&kexec_pte0);
-        page_list[VA_PTE_0] = (unsigned long)kexec_pte0;
-        page_list[PA_PUD_1] = virt_to_phys(&kexec_pud1);
-        page_list[VA_PUD_1] = (unsigned long)kexec_pud1;
-        page_list[PA_PMD_1] = virt_to_phys(&kexec_pmd1);
-        page_list[VA_PMD_1] = (unsigned long)kexec_pmd1;
-        page_list[PA_PTE_1] = virt_to_phys(&kexec_pte1);
-        page_list[VA_PTE_1] = (unsigned long)kexec_pte1;
-
         page_list[PA_TABLE_PAGE] =
           (unsigned long)__pa(page_address(image->control_code_page));
 
diff --git a/arch/x86/kernel/relocate_kernel_64.S b/arch/x86/kernel/relocate_kernel_64.S
index f5afe665a82b..b0bbdd4829c9 100644
--- a/arch/x86/kernel/relocate_kernel_64.S
+++ b/arch/x86/kernel/relocate_kernel_64.S
@@ -29,122 +29,6 @@ relocate_kernel:
          * %rdx start address
          */
 
-        /* map the control page at its virtual address */
-
-        movq    $0x0000ff8000000000, %r10        /* mask */
-        mov     $(39 - 3), %cl                   /* bits to shift */
-        movq    PTR(VA_CONTROL_PAGE)(%rsi), %r11 /* address to map */
-
-        movq    %r11, %r9
-        andq    %r10, %r9
-        shrq    %cl, %r9
-
-        movq    PTR(VA_PGD)(%rsi), %r8
-        addq    %r8, %r9
-        movq    PTR(PA_PUD_0)(%rsi), %r8
-        orq     $PAGE_ATTR, %r8
-        movq    %r8, (%r9)
-
-        shrq    $9, %r10
-        sub     $9, %cl
-
-        movq    %r11, %r9
-        andq    %r10, %r9
-        shrq    %cl, %r9
-
-        movq    PTR(VA_PUD_0)(%rsi), %r8
-        addq    %r8, %r9
-        movq    PTR(PA_PMD_0)(%rsi), %r8
-        orq     $PAGE_ATTR, %r8
-        movq    %r8, (%r9)
-
-        shrq    $9, %r10
-        sub     $9, %cl
-
-        movq    %r11, %r9
-        andq    %r10, %r9
-        shrq    %cl, %r9
-
-        movq    PTR(VA_PMD_0)(%rsi), %r8
-        addq    %r8, %r9
-        movq    PTR(PA_PTE_0)(%rsi), %r8
-        orq     $PAGE_ATTR, %r8
-        movq    %r8, (%r9)
-
-        shrq    $9, %r10
-        sub     $9, %cl
-
-        movq    %r11, %r9
-        andq    %r10, %r9
-        shrq    %cl, %r9
-
-        movq    PTR(VA_PTE_0)(%rsi), %r8
-        addq    %r8, %r9
-        movq    PTR(PA_CONTROL_PAGE)(%rsi), %r8
-        orq     $PAGE_ATTR, %r8
-        movq    %r8, (%r9)
-
-        /* identity map the control page at its physical address */
-
-        movq    $0x0000ff8000000000, %r10        /* mask */
-        mov     $(39 - 3), %cl                   /* bits to shift */
-        movq    PTR(PA_CONTROL_PAGE)(%rsi), %r11 /* address to map */
-
-        movq    %r11, %r9
-        andq    %r10, %r9
-        shrq    %cl, %r9
-
-        movq    PTR(VA_PGD)(%rsi), %r8
-        addq    %r8, %r9
-        movq    PTR(PA_PUD_1)(%rsi), %r8
-        orq     $PAGE_ATTR, %r8
-        movq    %r8, (%r9)
-
-        shrq    $9, %r10
-        sub     $9, %cl
-
-        movq    %r11, %r9
-        andq    %r10, %r9
-        shrq    %cl, %r9
-
-        movq    PTR(VA_PUD_1)(%rsi), %r8
-        addq    %r8, %r9
-        movq    PTR(PA_PMD_1)(%rsi), %r8
-        orq     $PAGE_ATTR, %r8
-        movq    %r8, (%r9)
-
-        shrq    $9, %r10
-        sub     $9, %cl
-
-        movq    %r11, %r9
-        andq    %r10, %r9
-        shrq    %cl, %r9
-
-        movq    PTR(VA_PMD_1)(%rsi), %r8
-        addq    %r8, %r9
-        movq    PTR(PA_PTE_1)(%rsi), %r8
-        orq     $PAGE_ATTR, %r8
-        movq    %r8, (%r9)
-
-        shrq    $9, %r10
-        sub     $9, %cl
-
-        movq    %r11, %r9
-        andq    %r10, %r9
-        shrq    %cl, %r9
-
-        movq    PTR(VA_PTE_1)(%rsi), %r8
-        addq    %r8, %r9
-        movq    PTR(PA_CONTROL_PAGE)(%rsi), %r8
-        orq     $PAGE_ATTR, %r8
-        movq    %r8, (%r9)
-
-relocate_new_kernel:
-        /* %rdi indirection_page
-         * %rsi page_list
-         * %rdx start address
-         */
-
         /* zero out flags, and disable interrupts */
         pushq $0
         popfq
@@ -156,9 +40,8 @@ relocate_new_kernel:
         /* get physical address of page table now too */
         movq    PTR(PA_TABLE_PAGE)(%rsi), %rcx
 
-        /* switch to new set of page tables */
-        movq    PTR(PA_PGD)(%rsi), %r9
-        movq    %r9, %cr3
+        /* Switch to the identity mapped page tables */
+        movq    %rcx, %cr3
 
         /* setup a new stack at the end of the physical control page */
         lea     PAGE_SIZE(%r8), %rsp
@@ -194,9 +77,7 @@ identity_mapped:
         jmp 1f
 1:
 
-        /* Switch to the identity mapped page tables,
-         * and flush the TLB.
-        */
+        /* Flush the TLB (needed?) */
         movq    %rcx, %cr3
 
         /* Do the copies */
diff --git a/arch/x86/kernel/trampoline_64.S b/arch/x86/kernel/trampoline_64.S
index 894293c598db..95a012a4664e 100644
--- a/arch/x86/kernel/trampoline_64.S
+++ b/arch/x86/kernel/trampoline_64.S
@@ -29,6 +29,7 @@
 #include <asm/page.h>
 #include <asm/msr.h>
 #include <asm/segment.h>
+#include <asm/processor-flags.h>
 
 .section .rodata, "a", @progbits
 
@@ -37,7 +38,7 @@
 ENTRY(trampoline_data)
 r_base = .
         cli                     # We should be safe anyway
-        wbinvd  
+        wbinvd
         mov     %cs, %ax        # Code and data in the same place
         mov     %ax, %ds
         mov     %ax, %es
@@ -73,9 +74,8 @@ r_base = .
         lidtl   tidt - r_base   # load idt with 0, 0
         lgdtl   tgdt - r_base   # load gdt with whatever is appropriate
 
-        xor     %ax, %ax
-        inc     %ax             # protected mode (PE) bit
-        lmsw    %ax             # into protected mode
+        mov     $X86_CR0_PE, %ax        # protected mode (PE) bit
+        lmsw    %ax                     # into protected mode
 
         # flush prefetch and jump to startup_32
         ljmpl   *(startup_32_vector - r_base)
@@ -86,9 +86,8 @@ startup_32:
         movl    $__KERNEL_DS, %eax      # Initialize the %ds segment register
         movl    %eax, %ds
 
-        xorl    %eax, %eax
-        btsl    $5, %eax                # Enable PAE mode
-        movl    %eax, %cr4
+        movl    $X86_CR4_PAE, %eax
+        movl    %eax, %cr4              # Enable PAE mode
 
                                         # Setup trampoline 4 level pagetables
         leal    (trampoline_level4_pgt - r_base)(%esi), %eax
@@ -99,9 +98,9 @@ startup_32:
         xorl    %edx, %edx
         wrmsr
 
-        xorl    %eax, %eax
-        btsl    $31, %eax               # Enable paging and in turn activate Long Mode
-        btsl    $0, %eax                # Enable protected mode
+        # Enable paging and in turn activate Long Mode
+        # Enable protected mode
+        movl    $(X86_CR0_PG | X86_CR0_PE), %eax
         movl    %eax, %cr0
 
         /*