[PATCH -mm 0/4] Hibernation: Arbitrary boot kernel support on x86_64 (rev. 2)

[PATCH -mm 0/4] Hibernation: Arbitrary boot kernel support on x86_64 (rev. 2)

[Rafael J. Wysocki]

Hi,

The first two patches in this series make it possible to restore the memory
state from a hibernation image with the help of a kernel different from the
image one.

The first patch adds the generic, platform independent code needed for that.
The second patch implements the idea on x86_64.

The remaining two patches remove some implicit assumptions about the boot and
image kernels from the x86_64 restore code.

Greetings,
Rafael


-- 
"Premature optimization is the root of all evil." - Donald Knuth

[PATCH -mm 1/4] Hibernation: Arbitrary boot kernel support - generic code (rev. 2)

[Rafael J. Wysocki]

From: Rafael J. Wysocki <rjw@sisk.pl>

Add the bits needed for supporting arbitrary boot kernels to the common
hibernation code.

To support arbitrary boot kernels, make it possible to replace the 'struct
new_utsname' and the kernel version in the hibernation image header by some
architecture specific data that will be used to verify if the image is valid
and to restore the image.

Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
---
 kernel/power/power.h    |   20 +++++++++++++++++-
 kernel/power/snapshot.c |   53 +++++++++++++++++++++++++++++++-----------------
 2 files changed, 54 insertions(+), 19 deletions(-)

Index: linux-2.6.23-rc3/kernel/power/power.h
===================================================================
--- linux-2.6.23-rc3.orig/kernel/power/power.h	2007-08-23 23:13:34.000000000 +0200
+++ linux-2.6.23-rc3/kernel/power/power.h	2007-08-25 21:18:59.000000000 +0200
@@ -11,14 +11,32 @@ struct swsusp_info {
 	unsigned long		size;
 } __attribute__((aligned(PAGE_SIZE)));
 
+#ifdef CONFIG_HIBERNATION
+#ifdef CONFIG_ARCH_HIBERNATION_HEADER
+/* Maximum size of architecture specific data in a hibernation header */
+#define MAX_ARCH_HEADER_SIZE	(sizeof(struct new_utsname) + 4)
 
+extern int arch_hibernation_header_save(void *addr, unsigned int max_size);
+extern int arch_hibernation_header_restore(void *addr);
+
+static inline int init_header_complete(struct swsusp_info *info)
+{
+	return arch_hibernation_header_save(info, MAX_ARCH_HEADER_SIZE);
+}
+
+static inline char *check_image_kernel(struct swsusp_info *info)
+{
+	return arch_hibernation_header_restore(info) ?
+			"architecture specific data" : NULL;
+}
+#endif /* CONFIG_ARCH_HIBERNATION_HEADER */
 
-#ifdef CONFIG_HIBERNATION
 /*
  * Keep some memory free so that I/O operations can succeed without paging
  * [Might this be more than 4 MB?]
  */
 #define PAGES_FOR_IO	((4096 * 1024) >> PAGE_SHIFT)
+
 /*
  * Keep 1 MB of memory free so that device drivers can allocate some pages in
  * their .suspend() routines without breaking the suspend to disk.
Index: linux-2.6.23-rc3/kernel/power/snapshot.c
===================================================================
--- linux-2.6.23-rc3.orig/kernel/power/snapshot.c	2007-08-23 23:13:34.000000000 +0200
+++ linux-2.6.23-rc3/kernel/power/snapshot.c	2007-08-25 21:21:55.000000000 +0200
@@ -1239,17 +1239,39 @@ asmlinkage int swsusp_save(void)
 	return 0;
 }
 
-static void init_header(struct swsusp_info *info)
+#ifndef CONFIG_ARCH_HIBERNATION_HEADER
+static int init_header_complete(struct swsusp_info *info)
 {
-	memset(info, 0, sizeof(struct swsusp_info));
+	memcpy(&info->uts, init_utsname(), sizeof(struct new_utsname));
 	info->version_code = LINUX_VERSION_CODE;
+	return 0;
+}
+
+static char *check_image_kernel(struct swsusp_info *info)
+{
+	if (info->version_code != LINUX_VERSION_CODE)
+		return "kernel version";
+	if (strcmp(info->uts.sysname,init_utsname()->sysname))
+		return "system type";
+	if (strcmp(info->uts.release,init_utsname()->release))
+		return "kernel release";
+	if (strcmp(info->uts.version,init_utsname()->version))
+		return "version";
+	if (strcmp(info->uts.machine,init_utsname()->machine))
+		return "machine";
+	return NULL;
+}
+#endif /* CONFIG_ARCH_HIBERNATION_HEADER */
+
+static int init_header(struct swsusp_info *info)
+{
+	memset(info, 0, sizeof(struct swsusp_info));
 	info->num_physpages = num_physpages;
-	memcpy(&info->uts, init_utsname(), sizeof(struct new_utsname));
-	info->cpus = num_online_cpus();
 	info->image_pages = nr_copy_pages;
 	info->pages = nr_copy_pages + nr_meta_pages + 1;
 	info->size = info->pages;
 	info->size <<= PAGE_SHIFT;
+	return init_header_complete(info);
 }
 
 /**
@@ -1303,7 +1325,11 @@ int snapshot_read_next(struct snapshot_h
 			return -ENOMEM;
 	}
 	if (!handle->offset) {
-		init_header((struct swsusp_info *)buffer);
+		int error;
+
+		error = init_header((struct swsusp_info *)buffer);
+		if (error)
+			return error;
 		handle->buffer = buffer;
 		memory_bm_position_reset(&orig_bm);
 		memory_bm_position_reset(&copy_bm);
@@ -1394,22 +1420,13 @@ duplicate_memory_bitmap(struct memory_bi
 	}
 }
 
-static inline int check_header(struct swsusp_info *info)
+static int check_header(struct swsusp_info *info)
 {
-	char *reason = NULL;
+	char *reason;
 
-	if (info->version_code != LINUX_VERSION_CODE)
-		reason = "kernel version";
-	if (info->num_physpages != num_physpages)
+	reason = check_image_kernel(info);
+	if (!reason && info->num_physpages != num_physpages)
 		reason = "memory size";
-	if (strcmp(info->uts.sysname,init_utsname()->sysname))
-		reason = "system type";
-	if (strcmp(info->uts.release,init_utsname()->release))
-		reason = "kernel release";
-	if (strcmp(info->uts.version,init_utsname()->version))
-		reason = "version";
-	if (strcmp(info->uts.machine,init_utsname()->machine))
-		reason = "machine";
 	if (reason) {
 		printk(KERN_ERR "swsusp: Resume mismatch: %s\n", reason);
 		return -EPERM;

[PATCH -mm 2/4] Hibernation: Arbitrary boot kernel support on x86_64 (rev. 2)

[Rafael J. Wysocki]

From: Rafael J. Wysocki <rjw@sisk.pl>

Make it possible to restore a hibernation image on x86_64 with the help of a
kernel different from the one in the image.

The idea is to split the core restoration code into two separate parts and to
place each of them in a different page.  The first part belongs to the boot
kernel and is executed as the last step of the image kernel's memory restoration
procedure.  Before being executed, it is relocated to a safe page that won't be
overwritten while copying the image kernel pages.

The final operation performed by it is a jump to the second part of the core
restoration code that belongs to the image kernel and has just been restored.
This code makes the CPU switch to the image kernel's page tables and
restores the state of general purpose registers (including the stack pointer)
from before the hibernation.

The main issue with this idea is that in order to jump to the second part of the
core restoration code the boot kernel needs to know its address.  However, this
address may be passed to it in the image header.  Namely, the part of the image
header previously used for checking if the version of the image kernel is
correct can be replaced with some architecture specific data that will allow
the boot kernel to jump to the right address within the image kernel.  These
data should also be used for checking if the image kernel is compatible with
the boot kernel (as far as the memory restroration procedure is concerned).
It can be done, for example, with the help of a "magic" value that has to be
equal in both kernels, so that they can be regarded as compatible.

Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Acked-by: Pavel Machek <pavel@ucw.cz>
---
 arch/x86_64/Kconfig              |    5 +++
 arch/x86_64/kernel/suspend.c     |   54 ++++++++++++++++++++++++++++++++++++++-
 arch/x86_64/kernel/suspend_asm.S |   41 ++++++++++++++++++++++++-----
 include/asm-x86_64/suspend.h     |    3 ++
 4 files changed, 95 insertions(+), 8 deletions(-)

Index: linux-2.6.23-rc3/arch/x86_64/kernel/suspend_asm.S
===================================================================
--- linux-2.6.23-rc3.orig/arch/x86_64/kernel/suspend_asm.S	2007-08-25 22:09:53.000000000 +0200
+++ linux-2.6.23-rc3/arch/x86_64/kernel/suspend_asm.S	2007-08-25 22:10:25.000000000 +0200
@@ -2,8 +2,8 @@
  *
  * Distribute under GPLv2.
  *
- * swsusp_arch_resume may not use any stack, nor any variable that is
- * not "NoSave" during copying pages:
+ * swsusp_arch_resume must not use any stack or any nonlocal variables while
+ * copying pages:
  *
  * Its rewriting one kernel image with another. What is stack in "old"
  * image could very well be data page in "new" image, and overwriting
@@ -36,6 +36,10 @@ ENTRY(swsusp_arch_suspend)
 	pushfq
 	popq	pt_regs_eflags(%rax)
 
+	/* save the address of restore_registers */
+	movq	$restore_registers, %rax
+	movq	%rax, restore_jump_address(%rip)
+
 	call swsusp_save
 	ret
 
@@ -54,7 +58,16 @@ ENTRY(restore_image)
 	movq	%rcx, %cr3;
 	movq	%rax, %cr4;  # turn PGE back on
 
+	/* prepare to jump to the image kernel */
+	movq	restore_jump_address(%rip), %rax
+
+	/* prepare to copy image data to their original locations */
 	movq	restore_pblist(%rip), %rdx
+	movq	relocated_restore_code(%rip), %rcx
+	jmpq	*%rcx
+
+	/* code below has been relocated to a safe page */
+ENTRY(core_restore_code)
 loop:
 	testq	%rdx, %rdx
 	jz	done
@@ -62,7 +75,7 @@ loop:
 	/* get addresses from the pbe and copy the page */
 	movq	pbe_address(%rdx), %rsi
 	movq	pbe_orig_address(%rdx), %rdi
-	movq	$512, %rcx
+	movq	$(PAGE_SIZE >> 3), %rcx
 	rep
 	movsq
 
@@ -70,6 +83,20 @@ loop:
 	movq	pbe_next(%rdx), %rdx
 	jmp	loop
 done:
+	/* jump to the restore_registers address from the image header */
+	jmpq	*%rax
+	/*
+	 * NOTE: This assumes that the boot kernel's text mapping covers the
+	 * image kernel's page containing restore_registers and the address of
+	 * this page is the same as in the image kernel's text mapping (it
+	 * should always be true, because the text mapping is linear, starting
+	 * from 0, and is supposed to cover the entire kernel text for every
+	 * kernel).
+	 *
+	 * code below belongs to the image kernel
+	 */
+
+ENTRY(restore_registers)
 	/* go back to the original page tables */
 	movq    $(init_level4_pgt - __START_KERNEL_map), %rax
 	addq    phys_base(%rip), %rax
@@ -84,10 +111,7 @@ done:
 	movq	%rcx, %cr3
 	movq	%rax, %cr4;  # turn PGE back on
 
-	movl	$24, %eax
-	movl	%eax, %ds
-
-	/* We don't restore %rax, it must be 0 anyway */
+	/* restore GPRs (we don't restore %rax, it must be 0 anyway) */
 	movq	$saved_context, %rax
 	movq	pt_regs_rsp(%rax), %rsp
 	movq	pt_regs_rbp(%rax), %rbp
@@ -109,4 +133,7 @@ done:
 
 	xorq	%rax, %rax
 
+	/* tell the hibernation core that we've just restored the memory */
+	movq	%rax, in_suspend(%rip)
+
 	ret
Index: linux-2.6.23-rc3/arch/x86_64/kernel/suspend.c
===================================================================
--- linux-2.6.23-rc3.orig/arch/x86_64/kernel/suspend.c	2007-08-25 22:09:53.000000000 +0200
+++ linux-2.6.23-rc3/arch/x86_64/kernel/suspend.c	2007-08-25 22:11:43.000000000 +0200
@@ -144,8 +144,16 @@ void fix_processor_context(void)
 /* Defined in arch/x86_64/kernel/suspend_asm.S */
 extern int restore_image(void);
 
+/*
+ * Address to jump to in the last phase of restore in order to get to the image
+ * kernel's text (this value is passed in the image header).
+ */
+unsigned long restore_jump_address;
+
 pgd_t *temp_level4_pgt;
 
+void *relocated_restore_code;
+
 static int res_phys_pud_init(pud_t *pud, unsigned long address, unsigned long end)
 {
 	long i, j;
@@ -169,7 +177,7 @@ static int res_phys_pud_init(pud_t *pud,
 
 			if (paddr >= end)
 				break;
-			pe = _PAGE_NX | _PAGE_PSE | _KERNPG_TABLE | paddr;
+			pe = __PAGE_KERNEL_LARGE_EXEC | paddr;
 			pe &= __supported_pte_mask;
 			set_pmd(pmd, __pmd(pe));
 		}
@@ -216,6 +224,13 @@ int swsusp_arch_resume(void)
 	/* We have got enough memory and from now on we cannot recover */
 	if ((error = set_up_temporary_mappings()))
 		return error;
+
+	relocated_restore_code = (void *)get_safe_page(GFP_ATOMIC);
+	if (!relocated_restore_code)
+		return -ENOMEM;
+	memcpy(relocated_restore_code, &core_restore_code,
+	       &restore_registers - &core_restore_code);
+
 	restore_image();
 	return 0;
 }
@@ -230,4 +245,41 @@ int pfn_is_nosave(unsigned long pfn)
 	unsigned long nosave_end_pfn = PAGE_ALIGN(__pa_symbol(&__nosave_end)) >> PAGE_SHIFT;
 	return (pfn >= nosave_begin_pfn) && (pfn < nosave_end_pfn);
 }
+
+struct restore_data_record {
+	unsigned long jump_address;
+	unsigned long control;
+};
+
+#define RESTORE_MAGIC	0x0123456789ABCDEFUL
+
+/**
+ *	arch_hibernation_header_save - populate the architecture specific part
+ *		of a hibernation image header
+ *	@addr: address to save the data at
+ */
+int arch_hibernation_header_save(void *addr, unsigned int max_size)
+{
+	struct restore_data_record *rdr = addr;
+
+	if (max_size < sizeof(struct restore_data_record))
+		return -EOVERFLOW;
+	rdr->jump_address = restore_jump_address;
+	rdr->control = (restore_jump_address ^ RESTORE_MAGIC);
+	return 0;
+}
+
+/**
+ *	arch_hibernation_header_restore - read the architecture specific data
+ *		from the hibernation image header
+ *	@addr: address to read the data from
+ */
+int arch_hibernation_header_restore(void *addr)
+{
+	struct restore_data_record *rdr = addr;
+
+	restore_jump_address = rdr->jump_address;
+	return (rdr->control == (restore_jump_address ^ RESTORE_MAGIC)) ?
+			0 : -EINVAL;
+}
 #endif /* CONFIG_HIBERNATION */
Index: linux-2.6.23-rc3/arch/x86_64/Kconfig
===================================================================
--- linux-2.6.23-rc3.orig/arch/x86_64/Kconfig	2007-08-25 22:09:53.000000000 +0200
+++ linux-2.6.23-rc3/arch/x86_64/Kconfig	2007-08-25 22:10:25.000000000 +0200
@@ -710,6 +710,11 @@ menu "Power management options"
 
 source kernel/power/Kconfig
 
+config ARCH_HIBERNATION_HEADER
+	bool
+	depends on HIBERNATION
+	default y
+
 source "drivers/acpi/Kconfig"
 
 source "arch/x86_64/kernel/cpufreq/Kconfig"
Index: linux-2.6.23-rc3/include/asm-x86_64/suspend.h
===================================================================
--- linux-2.6.23-rc3.orig/include/asm-x86_64/suspend.h	2007-08-25 22:09:53.000000000 +0200
+++ linux-2.6.23-rc3/include/asm-x86_64/suspend.h	2007-08-25 22:10:25.000000000 +0200
@@ -43,4 +43,7 @@ extern void fix_processor_context(void);
 /* routines for saving/restoring kernel state */
 extern int acpi_save_state_mem(void);
 
+extern char core_restore_code;
+extern char restore_registers;
+
 #endif /* __ASM_X86_64_SUSPEND_H */

[PATCH -mm 3/4] Hibernation: Pass CR3 in the image header on x86_64 (rev. 2)

[Rafael J. Wysocki]

From: Rafael J. Wysocki <rjw@sisk.pl>

Since we already pass the address of restore_registers() in the image header,
we can also pass the value of the CR3 register from before the hibernation in
the same way.  This will allow us to avoid using init_level4_pgt page tables
during the restore.

Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Acked-by: Pavel Machek <pavel@ucw.cz>
---
 arch/x86_64/kernel/suspend.c     |   16 ++++++++++++----
 arch/x86_64/kernel/suspend_asm.S |    8 +++++---
 2 files changed, 17 insertions(+), 7 deletions(-)

Index: linux-2.6.23-rc3/arch/x86_64/kernel/suspend.c
===================================================================
--- linux-2.6.23-rc3.orig/arch/x86_64/kernel/suspend.c	2007-08-25 22:11:43.000000000 +0200
+++ linux-2.6.23-rc3/arch/x86_64/kernel/suspend.c	2007-08-25 22:51:56.000000000 +0200
@@ -150,6 +150,12 @@ extern int restore_image(void);
  */
 unsigned long restore_jump_address;
 
+/*
+ * Value of the cr3 register from before the hibernation (this value is passed
+ * in the image header).
+ */
+unsigned long restore_cr3;
+
 pgd_t *temp_level4_pgt;
 
 void *relocated_restore_code;
@@ -248,7 +254,8 @@ int pfn_is_nosave(unsigned long pfn)
 
 struct restore_data_record {
 	unsigned long jump_address;
-	unsigned long control;
+	unsigned long cr3;
+	unsigned long magic;
 };
 
 #define RESTORE_MAGIC	0x0123456789ABCDEFUL
@@ -265,7 +272,8 @@ int arch_hibernation_header_save(void *a
 	if (max_size < sizeof(struct restore_data_record))
 		return -EOVERFLOW;
 	rdr->jump_address = restore_jump_address;
-	rdr->control = (restore_jump_address ^ RESTORE_MAGIC);
+	rdr->cr3 = restore_cr3;
+	rdr->magic = RESTORE_MAGIC;
 	return 0;
 }
 
@@ -279,7 +287,7 @@ int arch_hibernation_header_restore(void
 	struct restore_data_record *rdr = addr;
 
 	restore_jump_address = rdr->jump_address;
-	return (rdr->control == (restore_jump_address ^ RESTORE_MAGIC)) ?
-			0 : -EINVAL;
+	restore_cr3 = rdr->cr3;
+	return (rdr->magic == RESTORE_MAGIC) ? 0 : -EINVAL;
 }
 #endif /* CONFIG_HIBERNATION */
Index: linux-2.6.23-rc3/arch/x86_64/kernel/suspend_asm.S
===================================================================
--- linux-2.6.23-rc3.orig/arch/x86_64/kernel/suspend_asm.S	2007-08-25 22:10:25.000000000 +0200
+++ linux-2.6.23-rc3/arch/x86_64/kernel/suspend_asm.S	2007-08-25 22:51:56.000000000 +0200
@@ -39,6 +39,9 @@ ENTRY(swsusp_arch_suspend)
 	/* save the address of restore_registers */
 	movq	$restore_registers, %rax
 	movq	%rax, restore_jump_address(%rip)
+	/* save cr3 */
+	movq	%cr3, %rax
+	movq	%rax, restore_cr3(%rip)
 
 	call swsusp_save
 	ret
@@ -60,6 +63,7 @@ ENTRY(restore_image)
 
 	/* prepare to jump to the image kernel */
 	movq	restore_jump_address(%rip), %rax
+	movq	restore_cr3(%rip), %rbx
 
 	/* prepare to copy image data to their original locations */
 	movq	restore_pblist(%rip), %rdx
@@ -98,9 +102,7 @@ done:
 
 ENTRY(restore_registers)
 	/* go back to the original page tables */
-	movq    $(init_level4_pgt - __START_KERNEL_map), %rax
-	addq    phys_base(%rip), %rax
-	movq    %rax, %cr3
+	movq    %rbx, %cr3
 
 	/* Flush TLB, including "global" things (vmalloc) */
 	movq	mmu_cr4_features(%rip), %rax

[PATCH -mm 4/4] Hibernation: Use temporary page tables for kernel text mapping on x86_64

[Rafael J. Wysocki]

From: Rafael J. Wysocki <rjw@sisk.pl>

Use temporary page tables for the kernel text mapping during hibernation restore
on x86_64.

Without the patch, the original boot kernel's page tables that represent the
kernel text mapping are used while the core of the image kernel is being
restored.  However, in principle, if the boot kernel is not identical to the
image kernel, the location of these page tables in the image kernel need not be
the same, so we should create a safe copy of the kernel text mapping prior to
restoring the core of the image kernel.

Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Acked-by: Pavel Machek <pavel@ucw.cz>
---
 arch/x86_64/kernel/suspend.c |   39 +++++++++++++++++++++++++++++++++------
 1 file changed, 33 insertions(+), 6 deletions(-)

Index: linux-2.6.23-rc3/arch/x86_64/kernel/suspend.c
===================================================================
--- linux-2.6.23-rc3.orig/arch/x86_64/kernel/suspend.c	2007-08-25 22:51:56.000000000 +0200
+++ linux-2.6.23-rc3/arch/x86_64/kernel/suspend.c	2007-08-25 22:53:01.000000000 +0200
@@ -191,25 +191,42 @@ static int res_phys_pud_init(pud_t *pud,
 	return 0;
 }
 
+static int res_kernel_text_pud_init(pud_t *pud, unsigned long start)
+{
+	pmd_t *pmd;
+	unsigned long paddr;
+
+	pmd = (pmd_t *)get_safe_page(GFP_ATOMIC);
+	if (!pmd)
+		return -ENOMEM;
+	set_pud(pud + pud_index(start), __pud(__pa(pmd) | _KERNPG_TABLE));
+	for (paddr = 0; paddr < KERNEL_TEXT_SIZE; pmd++, paddr += PMD_SIZE) {
+		unsigned long pe;
+
+		pe = __PAGE_KERNEL_LARGE_EXEC | _PAGE_GLOBAL | paddr;
+		pe &= __supported_pte_mask;
+		set_pmd(pmd, __pmd(pe));
+	}
+
+	return 0;
+}
+
 static int set_up_temporary_mappings(void)
 {
 	unsigned long start, end, next;
+	pud_t *pud;
 	int error;
 
 	temp_level4_pgt = (pgd_t *)get_safe_page(GFP_ATOMIC);
 	if (!temp_level4_pgt)
 		return -ENOMEM;
 
-	/* 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 *)get_safe_page(GFP_ATOMIC);
+		pud = (pud_t *)get_safe_page(GFP_ATOMIC);
 		if (!pud)
 			return -ENOMEM;
 		next = start + PGDIR_SIZE;
@@ -220,7 +237,17 @@ static int set_up_temporary_mappings(voi
 		set_pgd(temp_level4_pgt + pgd_index(start),
 			mk_kernel_pgd(__pa(pud)));
 	}
-	return 0;
+
+	/* Set up the kernel text mapping from scratch */
+	pud = (pud_t *)get_safe_page(GFP_ATOMIC);
+	if (!pud)
+		return -ENOMEM;
+	error = res_kernel_text_pud_init(pud, __START_KERNEL_map);
+	if (!error)
+		set_pgd(temp_level4_pgt + pgd_index(__START_KERNEL_map),
+			__pgd(__pa(pud) | _PAGE_TABLE));
+
+	return error;
 }
 
 int swsusp_arch_resume(void)