7 files changed, 168 insertions, 34 deletions

diff --git a/arch/x86_64/ia32/ia32_signal.c b/arch/x86_64/ia32/ia32_signal.c
index 66e2821533db..0903cc1faef2 100644
--- a/arch/x86_64/ia32/ia32_signal.c
+++ b/arch/x86_64/ia32/ia32_signal.c
@@ -425,7 +425,11 @@ get_sigframe(struct k_sigaction *ka, struct pt_regs * regs, size_t frame_size)
 		rsp = (unsigned long) ka->sa.sa_restorer;
 	}
 
-	return (void __user *)((rsp - frame_size) & -8UL);
+	rsp -= frame_size;
+	/* Align the stack pointer according to the i386 ABI,
+	 * i.e. so that on function entry ((sp + 4) & 15) == 0. */
+	rsp = ((rsp + 4) & -16ul) - 4;
+	return (void __user *) rsp;
 }
 
 int ia32_setup_frame(int sig, struct k_sigaction *ka,
diff --git a/arch/x86_64/kernel/head.S b/arch/x86_64/kernel/head.S
index 4592bf21fcaf..b92e5f45ed46 100644
--- a/arch/x86_64/kernel/head.S
+++ b/arch/x86_64/kernel/head.S
@@ -270,26 +270,26 @@ ENTRY(level3_kernel_pgt)
 .org 0x4000
 ENTRY(level2_ident_pgt)
 	/* 40MB for bootup. 	*/
-	.quad	0x0000000000000183
-	.quad	0x0000000000200183
-	.quad	0x0000000000400183
-	.quad	0x0000000000600183
-	.quad	0x0000000000800183
-	.quad	0x0000000000A00183
-	.quad	0x0000000000C00183
-	.quad	0x0000000000E00183
-	.quad	0x0000000001000183
-	.quad	0x0000000001200183
-	.quad	0x0000000001400183
-	.quad	0x0000000001600183
-	.quad	0x0000000001800183
-	.quad	0x0000000001A00183
-	.quad	0x0000000001C00183
-	.quad	0x0000000001E00183
-	.quad	0x0000000002000183
-	.quad	0x0000000002200183
-	.quad	0x0000000002400183
-	.quad	0x0000000002600183
+	.quad	0x0000000000000083
+	.quad	0x0000000000200083
+	.quad	0x0000000000400083
+	.quad	0x0000000000600083
+	.quad	0x0000000000800083
+	.quad	0x0000000000A00083
+	.quad	0x0000000000C00083
+	.quad	0x0000000000E00083
+	.quad	0x0000000001000083
+	.quad	0x0000000001200083
+	.quad	0x0000000001400083
+	.quad	0x0000000001600083
+	.quad	0x0000000001800083
+	.quad	0x0000000001A00083
+	.quad	0x0000000001C00083
+	.quad	0x0000000001E00083
+	.quad	0x0000000002000083
+	.quad	0x0000000002200083
+	.quad	0x0000000002400083
+	.quad	0x0000000002600083
 	/* Temporary mappings for the super early allocator in arch/x86_64/mm/init.c */
 	.globl temp_boot_pmds
 temp_boot_pmds:
diff --git a/arch/x86_64/kernel/setup64.c b/arch/x86_64/kernel/setup64.c
index bd33be24a386..79190891fbc5 100644
--- a/arch/x86_64/kernel/setup64.c
+++ b/arch/x86_64/kernel/setup64.c
@@ -87,6 +87,10 @@ void __init setup_per_cpu_areas(void)
 	int i;
 	unsigned long size;
 
+#ifdef CONFIG_HOTPLUG_CPU
+	prefill_possible_map();
+#endif
+
 	/* Copy section for each CPU (we discard the original) */
 	size = ALIGN(__per_cpu_end - __per_cpu_start, SMP_CACHE_BYTES);
 #ifdef CONFIG_MODULES
diff --git a/arch/x86_64/kernel/smpboot.c b/arch/x86_64/kernel/smpboot.c
index e12d7baeb33e..658a81b33f3b 100644
--- a/arch/x86_64/kernel/smpboot.c
+++ b/arch/x86_64/kernel/smpboot.c
@@ -892,7 +892,7 @@ static __init void disable_smp(void)
  * those NR_CPUS, hence cpu_possible_map represents entire NR_CPUS range.
  * - Ashok Raj
  */
-static void prefill_possible_map(void)
+__init void prefill_possible_map(void)
 {
 	int i;
 	for (i = 0; i < NR_CPUS; i++)
@@ -967,10 +967,6 @@ void __init smp_prepare_cpus(unsigned int max_cpus)
 	current_cpu_data = boot_cpu_data;
 	current_thread_info()->cpu = 0;  /* needed? */
 
-#ifdef CONFIG_HOTPLUG_CPU
-	prefill_possible_map();
-#endif
-
 	if (smp_sanity_check(max_cpus) < 0) {
 		printk(KERN_INFO "SMP disabled\n");
 		disable_smp();
diff --git a/arch/x86_64/kernel/suspend.c b/arch/x86_64/kernel/suspend.c
index ebb9abf3ce6d..f066c6ab3618 100644
--- a/arch/x86_64/kernel/suspend.c
+++ b/arch/x86_64/kernel/suspend.c
@@ -11,6 +11,8 @@
 #include <linux/smp.h>
 #include <linux/suspend.h>
 #include <asm/proto.h>
+#include <asm/page.h>
+#include <asm/pgtable.h>
 
 struct saved_context saved_context;
 
@@ -140,4 +142,129 @@ void fix_processor_context(void)
 
 }
 
+#ifdef CONFIG_SOFTWARE_SUSPEND
+/* Defined in arch/x86_64/kernel/suspend_asm.S */
+extern int restore_image(void);
 
+pgd_t *temp_level4_pgt;
+
+static void **pages;
+
+static inline void *__add_page(void)
+{
+	void **c;
+
+	c = (void **)get_usable_page(GFP_ATOMIC);
+	if (c) {
+		*c = pages;
+		pages = c;
+	}
+	return c;
+}
+
+static inline void *__next_page(void)
+{
+	void **c;
+
+	c = pages;
+	if (c) {
+		pages = *c;
+		*c = NULL;
+	}
+	return c;
+}
+
+/*
+ * Try to allocate as many usable pages as needed and daisy chain them.
+ * If one allocation fails, free the pages allocated so far
+ */
+static int alloc_usable_pages(unsigned long n)
+{
+	void *p;
+
+	pages = NULL;
+	do
+		if (!__add_page())
+			break;
+	while (--n);
+	if (n) {
+		p = __next_page();
+		while (p) {
+			free_page((unsigned long)p);
+			p = __next_page();
+		}
+		return -ENOMEM;
+	}
+	return 0;
+}
+
+static void res_phys_pud_init(pud_t *pud, unsigned long address, unsigned long end)
+{
+	long i, j;
+
+	i = pud_index(address);
+	pud = pud + i;
+	for (; i < PTRS_PER_PUD; pud++, i++) {
+		unsigned long paddr;
+		pmd_t *pmd;
+
+		paddr = address + i*PUD_SIZE;
+		if (paddr >= end)
+			break;
+
+		pmd = (pmd_t *)__next_page();
+		set_pud(pud, __pud(__pa(pmd) | _KERNPG_TABLE));
+		for (j = 0; j < PTRS_PER_PMD; pmd++, j++, paddr += PMD_SIZE) {
+			unsigned long pe;
+
+			if (paddr >= end)
+				break;
+			pe = _PAGE_NX | _PAGE_PSE | _KERNPG_TABLE | paddr;
+			pe &= __supported_pte_mask;
+			set_pmd(pmd, __pmd(pe));
+		}
+	}
+}
+
+static void set_up_temporary_mappings(void)
+{
+	unsigned long start, end, next;
+
+	temp_level4_pgt = (pgd_t *)__next_page();
+
+	/* It is safe to reuse the original kernel mapping */
+	set_pgd(temp_level4_pgt + pgd_index(__START_KERNEL_map),
+		init_level4_pgt[pgd_index(__START_KERNEL_map)]);
+
+	/* Set up the direct mapping from scratch */
+	start = (unsigned long)pfn_to_kaddr(0);
+	end = (unsigned long)pfn_to_kaddr(end_pfn);
+
+	for (; start < end; start = next) {
+		pud_t *pud = (pud_t *)__next_page();
+		next = start + PGDIR_SIZE;
+		if (next > end)
+			next = end;
+		res_phys_pud_init(pud, __pa(start), __pa(next));
+		set_pgd(temp_level4_pgt + pgd_index(start),
+			mk_kernel_pgd(__pa(pud)));
+	}
+}
+
+int swsusp_arch_resume(void)
+{
+	unsigned long n;
+
+	n = ((end_pfn << PAGE_SHIFT) + PUD_SIZE - 1) >> PUD_SHIFT;
+	n += (n + PTRS_PER_PUD - 1) / PTRS_PER_PUD + 1;
+	pr_debug("swsusp_arch_resume(): pages needed = %lu\n", n);
+	if (alloc_usable_pages(n)) {
+		free_eaten_memory();
+		return -ENOMEM;
+	}
+	/* We have got enough memory and from now on we cannot recover */
+	set_up_temporary_mappings();
+	restore_image();
+	return 0;
+}
+#endif /* CONFIG_SOFTWARE_SUSPEND */
diff --git a/arch/x86_64/kernel/suspend_asm.S b/arch/x86_64/kernel/suspend_asm.S
index 4d659e97df10..320b6fb00cca 100644
--- a/arch/x86_64/kernel/suspend_asm.S
+++ b/arch/x86_64/kernel/suspend_asm.S
@@ -39,12 +39,13 @@ ENTRY(swsusp_arch_suspend)
 	call swsusp_save
 	ret
 
-ENTRY(swsusp_arch_resume)
-	/* set up cr3 */	
-	leaq	init_level4_pgt(%rip),%rax
-	subq	$__START_KERNEL_map,%rax
-	movq	%rax,%cr3
-
+ENTRY(restore_image)
+	/* switch to temporary page tables */
+	movq	$__PAGE_OFFSET, %rdx
+	movq	temp_level4_pgt(%rip), %rax
+	subq	%rdx, %rax
+	movq	%rax, %cr3
+	/* Flush TLB */
 	movq	mmu_cr4_features(%rip), %rax
 	movq	%rax, %rdx
 	andq	$~(1<<7), %rdx	# PGE
@@ -69,6 +70,10 @@ loop:
 	movq	pbe_next(%rdx), %rdx
 	jmp	loop
 done:
+	/* go back to the original page tables */
+	leaq	init_level4_pgt(%rip), %rax
+	subq	$__START_KERNEL_map, %rax
+	movq	%rax, %cr3
 	/* Flush TLB, including "global" things (vmalloc) */
 	movq	mmu_cr4_features(%rip), %rax
 	movq	%rax, %rdx
diff --git a/arch/x86_64/mm/pageattr.c b/arch/x86_64/mm/pageattr.c
index 94862e1ec032..b90e8fe9eeb0 100644
--- a/arch/x86_64/mm/pageattr.c
+++ b/arch/x86_64/mm/pageattr.c
@@ -220,8 +220,6 @@ void global_flush_tlb(void)
 	down_read(&init_mm.mmap_sem);
 	df = xchg(&df_list, NULL);
 	up_read(&init_mm.mmap_sem);
-	if (!df)
-		return;
 	flush_map((df && !df->next) ? df->address : 0);
 	for (; df; df = next_df) { 
 		next_df = df->next;