[GIT PULL] x86/trampoline for 2.6.37

[GIT PULL] x86/trampoline for 2.6.37

[H. Peter Anvin]

Hi Linus,

This is a set of changes which completes the already-begun work of
replacing the zap_low_mappings() with a 1:1 set of page tables which
are kept around permanently, and then used for things like the SMP
trampoline, suspend to RAM, and reboot.

Part of it has already been integrated in 2.6.36, because it also
worked around an AMD hardware erratum; this patch completes the
cleanup.

	-hpa


The following changes since commit b5153163ed580e00c67bdfecb02b2e3843817b3e:

  Merge branch 'devel' of master.kernel.org:/home/rmk/linux-2.6-arm (2010-10-21 16:42:32 -0700)

are available in the git repository at:

  git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip.git x86-trampoline-for-linus

Borislav Petkov (1):
      x86-32, mm: Add an initial page table for core bootstrapping

H. Peter Anvin (2):
      Merge branch 'x86/vmware' into x86/trampoline
      Merge branch 'x86/cleanups' into x86/trampoline

 arch/x86/include/asm/pgtable_32.h |    2 +-
 arch/x86/include/asm/tlbflush.h   |    2 -
 arch/x86/include/asm/trampoline.h |    3 --
 arch/x86/kernel/acpi/sleep.c      |    7 ++++-
 arch/x86/kernel/head32.c          |    1 +
 arch/x86/kernel/head_32.S         |   55 +++++++++++++++++--------------------
 arch/x86/kernel/reboot.c          |   10 +-----
 arch/x86/kernel/setup.c           |   18 +++++++++++-
 arch/x86/kernel/smpboot.c         |   16 +++--------
 arch/x86/kernel/trampoline.c      |   16 -----------
 arch/x86/mm/init_32.c             |   45 ------------------------------
 11 files changed, 56 insertions(+), 119 deletions(-)

diff --git a/arch/x86/include/asm/pgtable_32.h b/arch/x86/include/asm/pgtable_32.h
index f686f49..8abde9e 100644
--- a/arch/x86/include/asm/pgtable_32.h
+++ b/arch/x86/include/asm/pgtable_32.h
@@ -26,7 +26,7 @@ struct mm_struct;
 struct vm_area_struct;
 
 extern pgd_t swapper_pg_dir[1024];
-extern pgd_t trampoline_pg_dir[1024];
+extern pgd_t initial_page_table[1024];
 
 static inline void pgtable_cache_init(void) { }
 static inline void check_pgt_cache(void) { }
diff --git a/arch/x86/include/asm/tlbflush.h b/arch/x86/include/asm/tlbflush.h
index 7f3eba0..169be89 100644
--- a/arch/x86/include/asm/tlbflush.h
+++ b/arch/x86/include/asm/tlbflush.h
@@ -172,6 +172,4 @@ static inline void flush_tlb_kernel_range(unsigned long start,
 	flush_tlb_all();
 }
 
-extern void zap_low_mappings(bool early);
-
 #endif /* _ASM_X86_TLBFLUSH_H */
diff --git a/arch/x86/include/asm/trampoline.h b/arch/x86/include/asm/trampoline.h
index 4dde797..f4500fb 100644
--- a/arch/x86/include/asm/trampoline.h
+++ b/arch/x86/include/asm/trampoline.h
@@ -13,16 +13,13 @@ extern unsigned char *trampoline_base;
 
 extern unsigned long init_rsp;
 extern unsigned long initial_code;
-extern unsigned long initial_page_table;
 extern unsigned long initial_gs;
 
 #define TRAMPOLINE_SIZE roundup(trampoline_end - trampoline_data, PAGE_SIZE)
 
 extern unsigned long setup_trampoline(void);
-extern void __init setup_trampoline_page_table(void);
 extern void __init reserve_trampoline_memory(void);
 #else
-static inline void setup_trampoline_page_table(void) {}
 static inline void reserve_trampoline_memory(void) {}
 #endif /* CONFIG_X86_TRAMPOLINE */
 
diff --git a/arch/x86/kernel/acpi/sleep.c b/arch/x86/kernel/acpi/sleep.c
index 33cec15..b35e1ab 100644
--- a/arch/x86/kernel/acpi/sleep.c
+++ b/arch/x86/kernel/acpi/sleep.c
@@ -12,6 +12,11 @@
 #include <asm/segment.h>
 #include <asm/desc.h>
 
+#ifdef CONFIG_X86_32
+#include <asm/pgtable.h>
+#include <asm/pgtable_32.h>
+#endif
+
 #include "realmode/wakeup.h"
 #include "sleep.h"
 
@@ -90,7 +95,7 @@ int acpi_save_state_mem(void)
 
 #ifndef CONFIG_64BIT
 	header->pmode_entry = (u32)&wakeup_pmode_return;
-	header->pmode_cr3 = (u32)(swsusp_pg_dir - __PAGE_OFFSET);
+	header->pmode_cr3 = (u32)__pa(&initial_page_table);
 	saved_magic = 0x12345678;
 #else /* CONFIG_64BIT */
 	header->trampoline_segment = setup_trampoline() >> 4;
diff --git a/arch/x86/kernel/head32.c b/arch/x86/kernel/head32.c
index 784360c..8b9c201 100644
--- a/arch/x86/kernel/head32.c
+++ b/arch/x86/kernel/head32.c
@@ -17,6 +17,7 @@
 #include <asm/apic.h>
 #include <asm/io_apic.h>
 #include <asm/bios_ebda.h>
+#include <asm/tlbflush.h>
 
 static void __init i386_default_early_setup(void)
 {
diff --git a/arch/x86/kernel/head_32.S b/arch/x86/kernel/head_32.S
index fa8c1b8..bcece91 100644
--- a/arch/x86/kernel/head_32.S
+++ b/arch/x86/kernel/head_32.S
@@ -183,13 +183,12 @@ default_entry:
 #ifdef CONFIG_X86_PAE
 
 	/*
-	 * In PAE mode swapper_pg_dir is statically defined to contain enough
-	 * entries to cover the VMSPLIT option (that is the top 1, 2 or 3
-	 * entries). The identity mapping is handled by pointing two PGD
-	 * entries to the first kernel PMD.
+	 * In PAE mode initial_page_table is statically defined to contain
+	 * enough entries to cover the VMSPLIT option (that is the top 1, 2 or 3
+	 * entries). The identity mapping is handled by pointing two PGD entries
+	 * to the first kernel PMD.
 	 *
-	 * Note the upper half of each PMD or PTE are always zero at
-	 * this stage.
+	 * Note the upper half of each PMD or PTE are always zero at this stage.
 	 */
 
 #define KPMDS (((-__PAGE_OFFSET) >> 30) & 3) /* Number of kernel PMDs */
@@ -197,7 +196,7 @@ default_entry:
 	xorl %ebx,%ebx				/* %ebx is kept at zero */
 
 	movl $pa(__brk_base), %edi
-	movl $pa(swapper_pg_pmd), %edx
+	movl $pa(initial_pg_pmd), %edx
 	movl $PTE_IDENT_ATTR, %eax
 10:
 	leal PDE_IDENT_ATTR(%edi),%ecx		/* Create PMD entry */
@@ -226,14 +225,14 @@ default_entry:
 	movl %eax, pa(max_pfn_mapped)
 
 	/* Do early initialization of the fixmap area */
-	movl $pa(swapper_pg_fixmap)+PDE_IDENT_ATTR,%eax
-	movl %eax,pa(swapper_pg_pmd+0x1000*KPMDS-8)
+	movl $pa(initial_pg_fixmap)+PDE_IDENT_ATTR,%eax
+	movl %eax,pa(initial_pg_pmd+0x1000*KPMDS-8)
 #else	/* Not PAE */
 
 page_pde_offset = (__PAGE_OFFSET >> 20);
 
 	movl $pa(__brk_base), %edi
-	movl $pa(swapper_pg_dir), %edx
+	movl $pa(initial_page_table), %edx
 	movl $PTE_IDENT_ATTR, %eax
 10:
 	leal PDE_IDENT_ATTR(%edi),%ecx		/* Create PDE entry */
@@ -257,8 +256,8 @@ page_pde_offset = (__PAGE_OFFSET >> 20);
 	movl %eax, pa(max_pfn_mapped)
 
 	/* Do early initialization of the fixmap area */
-	movl $pa(swapper_pg_fixmap)+PDE_IDENT_ATTR,%eax
-	movl %eax,pa(swapper_pg_dir+0xffc)
+	movl $pa(initial_pg_fixmap)+PDE_IDENT_ATTR,%eax
+	movl %eax,pa(initial_page_table+0xffc)
 #endif
 	jmp 3f
 /*
@@ -334,7 +333,7 @@ ENTRY(startup_32_smp)
 /*
  * Enable paging
  */
-	movl pa(initial_page_table), %eax
+	movl $pa(initial_page_table), %eax
 	movl %eax,%cr3		/* set the page table pointer.. */
 	movl %cr0,%eax
 	orl  $X86_CR0_PG,%eax
@@ -614,8 +613,6 @@ ignore_int:
 .align 4
 ENTRY(initial_code)
 	.long i386_start_kernel
-ENTRY(initial_page_table)
-	.long pa(swapper_pg_dir)
 
 /*
  * BSS section
@@ -623,20 +620,18 @@ ENTRY(initial_page_table)
 __PAGE_ALIGNED_BSS
 	.align PAGE_SIZE_asm
 #ifdef CONFIG_X86_PAE
-swapper_pg_pmd:
+initial_pg_pmd:
 	.fill 1024*KPMDS,4,0
 #else
-ENTRY(swapper_pg_dir)
+ENTRY(initial_page_table)
 	.fill 1024,4,0
 #endif
-swapper_pg_fixmap:
+initial_pg_fixmap:
 	.fill 1024,4,0
-#ifdef CONFIG_X86_TRAMPOLINE
-ENTRY(trampoline_pg_dir)
-	.fill 1024,4,0
-#endif
 ENTRY(empty_zero_page)
 	.fill 4096,1,0
+ENTRY(swapper_pg_dir)
+	.fill 1024,4,0
 
 /*
  * This starts the data section.
@@ -645,20 +640,20 @@ ENTRY(empty_zero_page)
 __PAGE_ALIGNED_DATA
 	/* Page-aligned for the benefit of paravirt? */
 	.align PAGE_SIZE_asm
-ENTRY(swapper_pg_dir)
-	.long	pa(swapper_pg_pmd+PGD_IDENT_ATTR),0	/* low identity map */
+ENTRY(initial_page_table)
+	.long	pa(initial_pg_pmd+PGD_IDENT_ATTR),0	/* low identity map */
 # if KPMDS == 3
-	.long	pa(swapper_pg_pmd+PGD_IDENT_ATTR),0
-	.long	pa(swapper_pg_pmd+PGD_IDENT_ATTR+0x1000),0
-	.long	pa(swapper_pg_pmd+PGD_IDENT_ATTR+0x2000),0
+	.long	pa(initial_pg_pmd+PGD_IDENT_ATTR),0
+	.long	pa(initial_pg_pmd+PGD_IDENT_ATTR+0x1000),0
+	.long	pa(initial_pg_pmd+PGD_IDENT_ATTR+0x2000),0
 # elif KPMDS == 2
 	.long	0,0
-	.long	pa(swapper_pg_pmd+PGD_IDENT_ATTR),0
-	.long	pa(swapper_pg_pmd+PGD_IDENT_ATTR+0x1000),0
+	.long	pa(initial_pg_pmd+PGD_IDENT_ATTR),0
+	.long	pa(initial_pg_pmd+PGD_IDENT_ATTR+0x1000),0
 # elif KPMDS == 1
 	.long	0,0
 	.long	0,0
-	.long	pa(swapper_pg_pmd+PGD_IDENT_ATTR),0
+	.long	pa(initial_pg_pmd+PGD_IDENT_ATTR),0
 # else
 #  error "Kernel PMDs should be 1, 2 or 3"
 # endif
diff --git a/arch/x86/kernel/reboot.c b/arch/x86/kernel/reboot.c
index 7a4cf14..f7f53dc 100644
--- a/arch/x86/kernel/reboot.c
+++ b/arch/x86/kernel/reboot.c
@@ -371,16 +371,10 @@ void machine_real_restart(const unsigned char *code, int length)
 	CMOS_WRITE(0x00, 0x8f);
 	spin_unlock(&rtc_lock);
 
-	/* Remap the kernel at virtual address zero, as well as offset zero
-	   from the kernel segment.  This assumes the kernel segment starts at
-	   virtual address PAGE_OFFSET. */
-	memcpy(swapper_pg_dir, swapper_pg_dir + KERNEL_PGD_BOUNDARY,
-		sizeof(swapper_pg_dir [0]) * KERNEL_PGD_PTRS);
-
 	/*
-	 * Use `swapper_pg_dir' as our page directory.
+	 * Switch back to the initial page table.
 	 */
-	load_cr3(swapper_pg_dir);
+	load_cr3(initial_page_table);
 
 	/* Write 0x1234 to absolute memory location 0x472.  The BIOS reads
 	   this on booting to tell it to "Bypass memory test (also warm
diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c
index 322b24f..af6cf2b 100644
--- a/arch/x86/kernel/setup.c
+++ b/arch/x86/kernel/setup.c
@@ -728,6 +728,17 @@ void __init setup_arch(char **cmdline_p)
 #ifdef CONFIG_X86_32
 	memcpy(&boot_cpu_data, &new_cpu_data, sizeof(new_cpu_data));
 	visws_early_detect();
+
+	/*
+	 * copy kernel address range established so far and switch
+	 * to the proper swapper page table
+	 */
+	clone_pgd_range(swapper_pg_dir     + KERNEL_PGD_BOUNDARY,
+			initial_page_table + KERNEL_PGD_BOUNDARY,
+			KERNEL_PGD_PTRS);
+
+	load_cr3(swapper_pg_dir);
+	__flush_tlb_all();
 #else
 	printk(KERN_INFO "Command line: %s\n", boot_command_line);
 #endif
@@ -1009,7 +1020,12 @@ void __init setup_arch(char **cmdline_p)
 	paging_init();
 	x86_init.paging.pagetable_setup_done(swapper_pg_dir);
 
-	setup_trampoline_page_table();
+#ifdef CONFIG_X86_32
+	/* sync back kernel address range */
+	clone_pgd_range(initial_page_table + KERNEL_PGD_BOUNDARY,
+			swapper_pg_dir     + KERNEL_PGD_BOUNDARY,
+			KERNEL_PGD_PTRS);
+#endif
 
 	tboot_probe();
 
diff --git a/arch/x86/kernel/smpboot.c b/arch/x86/kernel/smpboot.c
index 63a1a55..e63bb51 100644
--- a/arch/x86/kernel/smpboot.c
+++ b/arch/x86/kernel/smpboot.c
@@ -298,22 +298,16 @@ notrace static void __cpuinit start_secondary(void *unused)
 	 * fragile that we want to limit the things done here to the
 	 * most necessary things.
 	 */
+	cpu_init();
+	preempt_disable();
+	smp_callin();
 
 #ifdef CONFIG_X86_32
-	/*
-	 * Switch away from the trampoline page-table
-	 *
-	 * Do this before cpu_init() because it needs to access per-cpu
-	 * data which may not be mapped in the trampoline page-table.
-	 */
+	/* switch away from the initial page table */
 	load_cr3(swapper_pg_dir);
 	__flush_tlb_all();
 #endif
 
-	cpu_init();
-	preempt_disable();
-	smp_callin();
-
 	/* otherwise gcc will move up smp_processor_id before the cpu_init */
 	barrier();
 	/*
@@ -772,7 +766,6 @@ do_rest:
 #ifdef CONFIG_X86_32
 	/* Stack for startup_32 can be just as for start_secondary onwards */
 	irq_ctx_init(cpu);
-	initial_page_table = __pa(&trampoline_pg_dir);
 #else
 	clear_tsk_thread_flag(c_idle.idle, TIF_FORK);
 	initial_gs = per_cpu_offset(cpu);
@@ -921,7 +914,6 @@ int __cpuinit native_cpu_up(unsigned int cpu)
 	per_cpu(cpu_state, cpu) = CPU_UP_PREPARE;
 
 	err = do_boot_cpu(apicid, cpu);
-
 	if (err) {
 		pr_debug("do_boot_cpu failed %d\n", err);
 		return -EIO;
diff --git a/arch/x86/kernel/trampoline.c b/arch/x86/kernel/trampoline.c
index e2a5952..f1488a3 100644
--- a/arch/x86/kernel/trampoline.c
+++ b/arch/x86/kernel/trampoline.c
@@ -38,19 +38,3 @@ unsigned long __trampinit setup_trampoline(void)
 	memcpy(trampoline_base, trampoline_data, TRAMPOLINE_SIZE);
 	return virt_to_phys(trampoline_base);
 }
-
-void __init setup_trampoline_page_table(void)
-{
-#ifdef CONFIG_X86_32
-	/* Copy kernel address range */
-	clone_pgd_range(trampoline_pg_dir + KERNEL_PGD_BOUNDARY,
-			swapper_pg_dir + KERNEL_PGD_BOUNDARY,
-			KERNEL_PGD_PTRS);
-
-	/* Initialize low mappings */
-	clone_pgd_range(trampoline_pg_dir,
-			swapper_pg_dir + KERNEL_PGD_BOUNDARY,
-			min_t(unsigned long, KERNEL_PGD_PTRS,
-			      KERNEL_PGD_BOUNDARY));
-#endif
-}
diff --git a/arch/x86/mm/init_32.c b/arch/x86/mm/init_32.c
index 558f2d3..1aeac2d 100644
--- a/arch/x86/mm/init_32.c
+++ b/arch/x86/mm/init_32.c
@@ -548,48 +548,6 @@ static void __init pagetable_init(void)
 	permanent_kmaps_init(pgd_base);
 }
 
-#ifdef CONFIG_ACPI_SLEEP
-/*
- * ACPI suspend needs this for resume, because things like the intel-agp
- * driver might have split up a kernel 4MB mapping.
- */
-char swsusp_pg_dir[PAGE_SIZE]
-	__attribute__ ((aligned(PAGE_SIZE)));
-
-static inline void save_pg_dir(void)
-{
-	copy_page(swsusp_pg_dir, swapper_pg_dir);
-}
-#else /* !CONFIG_ACPI_SLEEP */
-static inline void save_pg_dir(void)
-{
-}
-#endif /* !CONFIG_ACPI_SLEEP */
-
-void zap_low_mappings(bool early)
-{
-	int i;
-
-	/*
-	 * Zap initial low-memory mappings.
-	 *
-	 * Note that "pgd_clear()" doesn't do it for
-	 * us, because pgd_clear() is a no-op on i386.
-	 */
-	for (i = 0; i < KERNEL_PGD_BOUNDARY; i++) {
-#ifdef CONFIG_X86_PAE
-		set_pgd(swapper_pg_dir+i, __pgd(1 + __pa(empty_zero_page)));
-#else
-		set_pgd(swapper_pg_dir+i, __pgd(0));
-#endif
-	}
-
-	if (early)
-		__flush_tlb();
-	else
-		flush_tlb_all();
-}
-
 pteval_t __supported_pte_mask __read_mostly = ~(_PAGE_NX | _PAGE_GLOBAL | _PAGE_IOMAP);
 EXPORT_SYMBOL_GPL(__supported_pte_mask);
 
@@ -958,9 +916,6 @@ void __init mem_init(void)
 
 	if (boot_cpu_data.wp_works_ok < 0)
 		test_wp_bit();
-
-	save_pg_dir();
-	zap_low_mappings(true);
 }
 
 #ifdef CONFIG_MEMORY_HOTPLUG

Re: [GIT PULL] x86/trampoline for 2.6.37

[Jaswinder Singh Rajput]

Hello,

On Sat, Oct 23, 2010 at 6:47 AM, H. Peter Anvin <hpa@linux.intel.com> wrote:
>
> Borislav Petkov (1):
>      x86-32, mm: Add an initial page table for core bootstrapping
>
> H. Peter Anvin (2):
>      Merge branch 'x86/vmware' into x86/trampoline
>      Merge branch 'x86/cleanups' into x86/trampoline
>
> --- a/arch/x86/kernel/acpi/sleep.c
> +++ b/arch/x86/kernel/acpi/sleep.c
> @@ -12,6 +12,11 @@
>  #include <asm/segment.h>
>  #include <asm/desc.h>
>
> +#ifdef CONFIG_X86_32
> +#include <asm/pgtable.h>
> +#include <asm/pgtable_32.h>
> +#endif
> +

pgtable.h will take care of pgtable_32.h, why you are doing this ?

Thanks,
--
Jaswinder Singh.

Re: [GIT PULL] x86/trampoline for 2.6.37

[Borislav Petkov]

On Sat, Oct 23, 2010 at 11:17:42AM +0530, Jaswinder Singh Rajput wrote:
> > +#ifdef CONFIG_X86_32
> > +#include <asm/pgtable.h>
> > +#include <asm/pgtable_32.h>
> > +#endif
> > +
> 
> pgtable.h will take care of pgtable_32.h, why you are doing this ?

Frankly, I don't remember. Here's an updated patch.

Linus, you could apply the following one instead of pulling from the
branch, if hpa is ok with it. It is against your tree of the moment
v2.6.36-1431-gf5d9d24.

--
From: Borislav Petkov <bp@alien8.de>
Date: Sat, 28 Aug 2010 15:58:33 +0200
Subject: [PATCH] x86-32, mm: Add an initial page table for core bootstrapping

This patch adds an initial page table with low mappings used exclusively
for booting APs/resuming after ACPI suspend/machine restart. After this,
there's no need to add low mappings to swapper_pg_dir and zap them later
or create own swsusp PGD page solely for ACPI sleep needs - we have
initial_page_table for that.

[ v1.1 Fix suspend-to-ram ]
[ v1.2 Remove duplicated include ]

Signed-off-by: Borislav Petkov <bp@alien8.de>
---
 arch/x86/include/asm/pgtable_32.h |    2 +-
 arch/x86/include/asm/tlbflush.h   |    2 -
 arch/x86/include/asm/trampoline.h |    3 --
 arch/x86/kernel/acpi/sleep.c      |    6 +++-
 arch/x86/kernel/head32.c          |    1 +
 arch/x86/kernel/head_32.S         |   55 +++++++++++++++++--------------------
 arch/x86/kernel/reboot.c          |   10 +-----
 arch/x86/kernel/setup.c           |   18 +++++++++++-
 arch/x86/kernel/smpboot.c         |   16 +++--------
 arch/x86/kernel/trampoline.c      |   16 -----------
 arch/x86/mm/init_32.c             |   45 ------------------------------
 11 files changed, 55 insertions(+), 119 deletions(-)

diff --git a/arch/x86/include/asm/pgtable_32.h b/arch/x86/include/asm/pgtable_32.h
index f686f49..8abde9e 100644
--- a/arch/x86/include/asm/pgtable_32.h
+++ b/arch/x86/include/asm/pgtable_32.h
@@ -26,7 +26,7 @@ struct mm_struct;
 struct vm_area_struct;
 
 extern pgd_t swapper_pg_dir[1024];
-extern pgd_t trampoline_pg_dir[1024];
+extern pgd_t initial_page_table[1024];
 
 static inline void pgtable_cache_init(void) { }
 static inline void check_pgt_cache(void) { }
diff --git a/arch/x86/include/asm/tlbflush.h b/arch/x86/include/asm/tlbflush.h
index 7f3eba0..169be89 100644
--- a/arch/x86/include/asm/tlbflush.h
+++ b/arch/x86/include/asm/tlbflush.h
@@ -172,6 +172,4 @@ static inline void flush_tlb_kernel_range(unsigned long start,
 	flush_tlb_all();
 }
 
-extern void zap_low_mappings(bool early);
-
 #endif /* _ASM_X86_TLBFLUSH_H */
diff --git a/arch/x86/include/asm/trampoline.h b/arch/x86/include/asm/trampoline.h
index 4dde797..f4500fb 100644
--- a/arch/x86/include/asm/trampoline.h
+++ b/arch/x86/include/asm/trampoline.h
@@ -13,16 +13,13 @@ extern unsigned char *trampoline_base;
 
 extern unsigned long init_rsp;
 extern unsigned long initial_code;
-extern unsigned long initial_page_table;
 extern unsigned long initial_gs;
 
 #define TRAMPOLINE_SIZE roundup(trampoline_end - trampoline_data, PAGE_SIZE)
 
 extern unsigned long setup_trampoline(void);
-extern void __init setup_trampoline_page_table(void);
 extern void __init reserve_trampoline_memory(void);
 #else
-static inline void setup_trampoline_page_table(void) {}
 static inline void reserve_trampoline_memory(void) {}
 #endif /* CONFIG_X86_TRAMPOLINE */
 
diff --git a/arch/x86/kernel/acpi/sleep.c b/arch/x86/kernel/acpi/sleep.c
index e125207..69fd72a 100644
--- a/arch/x86/kernel/acpi/sleep.c
+++ b/arch/x86/kernel/acpi/sleep.c
@@ -13,6 +13,10 @@
 #include <asm/segment.h>
 #include <asm/desc.h>
 
+#ifdef CONFIG_X86_32
+#include <asm/pgtable.h>
+#endif
+
 #include "realmode/wakeup.h"
 #include "sleep.h"
 
@@ -91,7 +95,7 @@ int acpi_save_state_mem(void)
 
 #ifndef CONFIG_64BIT
 	header->pmode_entry = (u32)&wakeup_pmode_return;
-	header->pmode_cr3 = (u32)(swsusp_pg_dir - __PAGE_OFFSET);
+	header->pmode_cr3 = (u32)__pa(&initial_page_table);
 	saved_magic = 0x12345678;
 #else /* CONFIG_64BIT */
 	header->trampoline_segment = setup_trampoline() >> 4;
diff --git a/arch/x86/kernel/head32.c b/arch/x86/kernel/head32.c
index 9a6ca23..7633101 100644
--- a/arch/x86/kernel/head32.c
+++ b/arch/x86/kernel/head32.c
@@ -18,6 +18,7 @@
 #include <asm/apic.h>
 #include <asm/io_apic.h>
 #include <asm/bios_ebda.h>
+#include <asm/tlbflush.h>
 
 static void __init i386_default_early_setup(void)
 {
diff --git a/arch/x86/kernel/head_32.S b/arch/x86/kernel/head_32.S
index fa8c1b8..bcece91 100644
--- a/arch/x86/kernel/head_32.S
+++ b/arch/x86/kernel/head_32.S
@@ -183,13 +183,12 @@ default_entry:
 #ifdef CONFIG_X86_PAE
 
 	/*
-	 * In PAE mode swapper_pg_dir is statically defined to contain enough
-	 * entries to cover the VMSPLIT option (that is the top 1, 2 or 3
-	 * entries). The identity mapping is handled by pointing two PGD
-	 * entries to the first kernel PMD.
+	 * In PAE mode initial_page_table is statically defined to contain
+	 * enough entries to cover the VMSPLIT option (that is the top 1, 2 or 3
+	 * entries). The identity mapping is handled by pointing two PGD entries
+	 * to the first kernel PMD.
 	 *
-	 * Note the upper half of each PMD or PTE are always zero at
-	 * this stage.
+	 * Note the upper half of each PMD or PTE are always zero at this stage.
 	 */
 
 #define KPMDS (((-__PAGE_OFFSET) >> 30) & 3) /* Number of kernel PMDs */
@@ -197,7 +196,7 @@ default_entry:
 	xorl %ebx,%ebx				/* %ebx is kept at zero */
 
 	movl $pa(__brk_base), %edi
-	movl $pa(swapper_pg_pmd), %edx
+	movl $pa(initial_pg_pmd), %edx
 	movl $PTE_IDENT_ATTR, %eax
 10:
 	leal PDE_IDENT_ATTR(%edi),%ecx		/* Create PMD entry */
@@ -226,14 +225,14 @@ default_entry:
 	movl %eax, pa(max_pfn_mapped)
 
 	/* Do early initialization of the fixmap area */
-	movl $pa(swapper_pg_fixmap)+PDE_IDENT_ATTR,%eax
-	movl %eax,pa(swapper_pg_pmd+0x1000*KPMDS-8)
+	movl $pa(initial_pg_fixmap)+PDE_IDENT_ATTR,%eax
+	movl %eax,pa(initial_pg_pmd+0x1000*KPMDS-8)
 #else	/* Not PAE */
 
 page_pde_offset = (__PAGE_OFFSET >> 20);
 
 	movl $pa(__brk_base), %edi
-	movl $pa(swapper_pg_dir), %edx
+	movl $pa(initial_page_table), %edx
 	movl $PTE_IDENT_ATTR, %eax
 10:
 	leal PDE_IDENT_ATTR(%edi),%ecx		/* Create PDE entry */
@@ -257,8 +256,8 @@ page_pde_offset = (__PAGE_OFFSET >> 20);
 	movl %eax, pa(max_pfn_mapped)
 
 	/* Do early initialization of the fixmap area */
-	movl $pa(swapper_pg_fixmap)+PDE_IDENT_ATTR,%eax
-	movl %eax,pa(swapper_pg_dir+0xffc)
+	movl $pa(initial_pg_fixmap)+PDE_IDENT_ATTR,%eax
+	movl %eax,pa(initial_page_table+0xffc)
 #endif
 	jmp 3f
 /*
@@ -334,7 +333,7 @@ ENTRY(startup_32_smp)
 /*
  * Enable paging
  */
-	movl pa(initial_page_table), %eax
+	movl $pa(initial_page_table), %eax
 	movl %eax,%cr3		/* set the page table pointer.. */
 	movl %cr0,%eax
 	orl  $X86_CR0_PG,%eax
@@ -614,8 +613,6 @@ ignore_int:
 .align 4
 ENTRY(initial_code)
 	.long i386_start_kernel
-ENTRY(initial_page_table)
-	.long pa(swapper_pg_dir)
 
 /*
  * BSS section
@@ -623,20 +620,18 @@ ENTRY(initial_page_table)
 __PAGE_ALIGNED_BSS
 	.align PAGE_SIZE_asm
 #ifdef CONFIG_X86_PAE
-swapper_pg_pmd:
+initial_pg_pmd:
 	.fill 1024*KPMDS,4,0
 #else
-ENTRY(swapper_pg_dir)
+ENTRY(initial_page_table)
 	.fill 1024,4,0
 #endif
-swapper_pg_fixmap:
+initial_pg_fixmap:
 	.fill 1024,4,0
-#ifdef CONFIG_X86_TRAMPOLINE
-ENTRY(trampoline_pg_dir)
-	.fill 1024,4,0
-#endif
 ENTRY(empty_zero_page)
 	.fill 4096,1,0
+ENTRY(swapper_pg_dir)
+	.fill 1024,4,0
 
 /*
  * This starts the data section.
@@ -645,20 +640,20 @@ ENTRY(empty_zero_page)
 __PAGE_ALIGNED_DATA
 	/* Page-aligned for the benefit of paravirt? */
 	.align PAGE_SIZE_asm
-ENTRY(swapper_pg_dir)
-	.long	pa(swapper_pg_pmd+PGD_IDENT_ATTR),0	/* low identity map */
+ENTRY(initial_page_table)
+	.long	pa(initial_pg_pmd+PGD_IDENT_ATTR),0	/* low identity map */
 # if KPMDS == 3
-	.long	pa(swapper_pg_pmd+PGD_IDENT_ATTR),0
-	.long	pa(swapper_pg_pmd+PGD_IDENT_ATTR+0x1000),0
-	.long	pa(swapper_pg_pmd+PGD_IDENT_ATTR+0x2000),0
+	.long	pa(initial_pg_pmd+PGD_IDENT_ATTR),0
+	.long	pa(initial_pg_pmd+PGD_IDENT_ATTR+0x1000),0
+	.long	pa(initial_pg_pmd+PGD_IDENT_ATTR+0x2000),0
 # elif KPMDS == 2
 	.long	0,0
-	.long	pa(swapper_pg_pmd+PGD_IDENT_ATTR),0
-	.long	pa(swapper_pg_pmd+PGD_IDENT_ATTR+0x1000),0
+	.long	pa(initial_pg_pmd+PGD_IDENT_ATTR),0
+	.long	pa(initial_pg_pmd+PGD_IDENT_ATTR+0x1000),0
 # elif KPMDS == 1
 	.long	0,0
 	.long	0,0
-	.long	pa(swapper_pg_pmd+PGD_IDENT_ATTR),0
+	.long	pa(initial_pg_pmd+PGD_IDENT_ATTR),0
 # else
 #  error "Kernel PMDs should be 1, 2 or 3"
 # endif
diff --git a/arch/x86/kernel/reboot.c b/arch/x86/kernel/reboot.c
index 7a4cf14..f7f53dc 100644
--- a/arch/x86/kernel/reboot.c
+++ b/arch/x86/kernel/reboot.c
@@ -371,16 +371,10 @@ void machine_real_restart(const unsigned char *code, int length)
 	CMOS_WRITE(0x00, 0x8f);
 	spin_unlock(&rtc_lock);
 
-	/* Remap the kernel at virtual address zero, as well as offset zero
-	   from the kernel segment.  This assumes the kernel segment starts at
-	   virtual address PAGE_OFFSET. */
-	memcpy(swapper_pg_dir, swapper_pg_dir + KERNEL_PGD_BOUNDARY,
-		sizeof(swapper_pg_dir [0]) * KERNEL_PGD_PTRS);
-
 	/*
-	 * Use `swapper_pg_dir' as our page directory.
+	 * Switch back to the initial page table.
 	 */
-	load_cr3(swapper_pg_dir);
+	load_cr3(initial_page_table);
 
 	/* Write 0x1234 to absolute memory location 0x472.  The BIOS reads
 	   this on booting to tell it to "Bypass memory test (also warm
diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c
index 420e641..95a3274 100644
--- a/arch/x86/kernel/setup.c
+++ b/arch/x86/kernel/setup.c
@@ -700,6 +700,17 @@ void __init setup_arch(char **cmdline_p)
 #ifdef CONFIG_X86_32
 	memcpy(&boot_cpu_data, &new_cpu_data, sizeof(new_cpu_data));
 	visws_early_detect();
+
+	/*
+	 * copy kernel address range established so far and switch
+	 * to the proper swapper page table
+	 */
+	clone_pgd_range(swapper_pg_dir     + KERNEL_PGD_BOUNDARY,
+			initial_page_table + KERNEL_PGD_BOUNDARY,
+			KERNEL_PGD_PTRS);
+
+	load_cr3(swapper_pg_dir);
+	__flush_tlb_all();
 #else
 	printk(KERN_INFO "Command line: %s\n", boot_command_line);
 #endif
@@ -985,7 +996,12 @@ void __init setup_arch(char **cmdline_p)
 	paging_init();
 	x86_init.paging.pagetable_setup_done(swapper_pg_dir);
 
-	setup_trampoline_page_table();
+#ifdef CONFIG_X86_32
+	/* sync back kernel address range */
+	clone_pgd_range(initial_page_table + KERNEL_PGD_BOUNDARY,
+			swapper_pg_dir     + KERNEL_PGD_BOUNDARY,
+			KERNEL_PGD_PTRS);
+#endif
 
 	tboot_probe();
 
diff --git a/arch/x86/kernel/smpboot.c b/arch/x86/kernel/smpboot.c
index dfb5089..6af1185 100644
--- a/arch/x86/kernel/smpboot.c
+++ b/arch/x86/kernel/smpboot.c
@@ -299,22 +299,16 @@ notrace static void __cpuinit start_secondary(void *unused)
 	 * fragile that we want to limit the things done here to the
 	 * most necessary things.
 	 */
+	cpu_init();
+	preempt_disable();
+	smp_callin();
 
 #ifdef CONFIG_X86_32
-	/*
-	 * Switch away from the trampoline page-table
-	 *
-	 * Do this before cpu_init() because it needs to access per-cpu
-	 * data which may not be mapped in the trampoline page-table.
-	 */
+	/* switch away from the initial page table */
 	load_cr3(swapper_pg_dir);
 	__flush_tlb_all();
 #endif
 
-	cpu_init();
-	preempt_disable();
-	smp_callin();
-
 	/* otherwise gcc will move up smp_processor_id before the cpu_init */
 	barrier();
 	/*
@@ -785,7 +779,6 @@ do_rest:
 #ifdef CONFIG_X86_32
 	/* Stack for startup_32 can be just as for start_secondary onwards */
 	irq_ctx_init(cpu);
-	initial_page_table = __pa(&trampoline_pg_dir);
 #else
 	clear_tsk_thread_flag(c_idle.idle, TIF_FORK);
 	initial_gs = per_cpu_offset(cpu);
@@ -934,7 +927,6 @@ int __cpuinit native_cpu_up(unsigned int cpu)
 	per_cpu(cpu_state, cpu) = CPU_UP_PREPARE;
 
 	err = do_boot_cpu(apicid, cpu);
-
 	if (err) {
 		pr_debug("do_boot_cpu failed %d\n", err);
 		return -EIO;
diff --git a/arch/x86/kernel/trampoline.c b/arch/x86/kernel/trampoline.c
index 4c3da56..a375616 100644
--- a/arch/x86/kernel/trampoline.c
+++ b/arch/x86/kernel/trampoline.c
@@ -38,19 +38,3 @@ unsigned long __trampinit setup_trampoline(void)
 	memcpy(trampoline_base, trampoline_data, TRAMPOLINE_SIZE);
 	return virt_to_phys(trampoline_base);
 }
-
-void __init setup_trampoline_page_table(void)
-{
-#ifdef CONFIG_X86_32
-	/* Copy kernel address range */
-	clone_pgd_range(trampoline_pg_dir + KERNEL_PGD_BOUNDARY,
-			swapper_pg_dir + KERNEL_PGD_BOUNDARY,
-			KERNEL_PGD_PTRS);
-
-	/* Initialize low mappings */
-	clone_pgd_range(trampoline_pg_dir,
-			swapper_pg_dir + KERNEL_PGD_BOUNDARY,
-			min_t(unsigned long, KERNEL_PGD_PTRS,
-			      KERNEL_PGD_BOUNDARY));
-#endif
-}
diff --git a/arch/x86/mm/init_32.c b/arch/x86/mm/init_32.c
index 5d0a671..0e969f9 100644
--- a/arch/x86/mm/init_32.c
+++ b/arch/x86/mm/init_32.c
@@ -528,48 +528,6 @@ static void __init pagetable_init(void)
 	permanent_kmaps_init(pgd_base);
 }
 
-#ifdef CONFIG_ACPI_SLEEP
-/*
- * ACPI suspend needs this for resume, because things like the intel-agp
- * driver might have split up a kernel 4MB mapping.
- */
-char swsusp_pg_dir[PAGE_SIZE]
-	__attribute__ ((aligned(PAGE_SIZE)));
-
-static inline void save_pg_dir(void)
-{
-	copy_page(swsusp_pg_dir, swapper_pg_dir);
-}
-#else /* !CONFIG_ACPI_SLEEP */
-static inline void save_pg_dir(void)
-{
-}
-#endif /* !CONFIG_ACPI_SLEEP */
-
-void zap_low_mappings(bool early)
-{
-	int i;
-
-	/*
-	 * Zap initial low-memory mappings.
-	 *
-	 * Note that "pgd_clear()" doesn't do it for
-	 * us, because pgd_clear() is a no-op on i386.
-	 */
-	for (i = 0; i < KERNEL_PGD_BOUNDARY; i++) {
-#ifdef CONFIG_X86_PAE
-		set_pgd(swapper_pg_dir+i, __pgd(1 + __pa(empty_zero_page)));
-#else
-		set_pgd(swapper_pg_dir+i, __pgd(0));
-#endif
-	}
-
-	if (early)
-		__flush_tlb();
-	else
-		flush_tlb_all();
-}
-
 pteval_t __supported_pte_mask __read_mostly = ~(_PAGE_NX | _PAGE_GLOBAL | _PAGE_IOMAP);
 EXPORT_SYMBOL_GPL(__supported_pte_mask);
 
@@ -882,9 +840,6 @@ void __init mem_init(void)
 
 	if (boot_cpu_data.wp_works_ok < 0)
 		test_wp_bit();
-
-	save_pg_dir();
-	zap_low_mappings(true);
 }
 
 #ifdef CONFIG_MEMORY_HOTPLUG
-- 
1.7.2.3

-- 
Regards/Gruss,
    Boris.

Re: [GIT PULL] x86/trampoline for 2.6.37

[Ingo Molnar]

* Borislav Petkov <bp@alien8.de> wrote:

> On Sat, Oct 23, 2010 at 11:17:42AM +0530, Jaswinder Singh Rajput wrote:
> > > +#ifdef CONFIG_X86_32
> > > +#include <asm/pgtable.h>
> > > +#include <asm/pgtable_32.h>
> > > +#endif
> > > +
> > 
> > pgtable.h will take care of pgtable_32.h, why you are doing this ?
> 
> Frankly, I don't remember. Here's an updated patch.
> 
> Linus, you could apply the following one instead of pulling from the branch, if 
> hpa is ok with it. It is against your tree of the moment v2.6.36-1431-gf5d9d24.

Please send a delta patch and we can do any trivial changes via the usual channels. 
The tree hpa sent to Linus is tested and this is tricky code.

Thanks,

	Ingo

[PATCH] x86-32, mm: Remove duplicated include

[Borislav Petkov]

On Sat, Oct 23, 2010 at 04:15:34PM +0200, Ingo Molnar wrote:
> Please send a delta patch and we can do any trivial changes via the
> usual channels. The tree hpa sent to Linus is tested and this is
> tricky code.

Ok, here it is, 32-bit randconfig tested for a day.

--
From: Borislav Petkov <bp@alien8.de>
Date: Mon, 25 Oct 2010 18:15:22 +0200
Subject: [PATCH] x86-32, mm: Remove duplicated include

b40827fa7268fda8a62490728a61c2856f33830b added an include directive
which is needless and is taken care of by a previous one. Remove it.

Caught-by: Jaswinder Singh Rajput <jaswinderlinux@gmail.com>
Signed-off-by: Borislav Petkov <bp@alien8.de>
---
 arch/x86/kernel/acpi/sleep.c |    1 -
 1 files changed, 0 insertions(+), 1 deletions(-)

diff --git a/arch/x86/kernel/acpi/sleep.c b/arch/x86/kernel/acpi/sleep.c
index 74a8478..69fd72a 100644
--- a/arch/x86/kernel/acpi/sleep.c
+++ b/arch/x86/kernel/acpi/sleep.c
@@ -15,7 +15,6 @@
 
 #ifdef CONFIG_X86_32
 #include <asm/pgtable.h>
-#include <asm/pgtable_32.h>
 #endif
 
 #include "realmode/wakeup.h"
-- 
1.7.1


-- 
Regards/Gruss,
Boris.

[tip:x86/urgent] x86-32, mm: Remove duplicated #include

[tip-bot for Borislav Petkov]

Commit-ID:  610470ce804f0326ca63fbcdc5be06b750debeb1
Gitweb:     http://git.kernel.org/tip/610470ce804f0326ca63fbcdc5be06b750debeb1
Author:     Borislav Petkov <bp@alien8.de>
AuthorDate: Mon, 25 Oct 2010 18:25:23 +0200
Committer:  Ingo Molnar <mingo@elte.hu>
CommitDate: Mon, 25 Oct 2010 19:39:33 +0200

x86-32, mm: Remove duplicated #include

b40827fa7268fda8a62490728a61c2856f33830b added an include
directive which is needless and is taken care of by a previous
one. Remove it.

Caught-by: Jaswinder Singh Rajput <jaswinderlinux@gmail.com>
Signed-off-by: Borislav Petkov <bp@alien8.de>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Jaswinder Singh Rajput <jaswinderlinux@gmail.com>
Cc: H. Peter Anvin <hpa@linux.intel.com>
LKML-Reference: <20101025162523.GA4712@a1.tnic>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
---
 arch/x86/kernel/acpi/sleep.c |    1 -
 1 files changed, 0 insertions(+), 1 deletions(-)

diff --git a/arch/x86/kernel/acpi/sleep.c b/arch/x86/kernel/acpi/sleep.c
index 74a8478..69fd72a 100644
--- a/arch/x86/kernel/acpi/sleep.c
+++ b/arch/x86/kernel/acpi/sleep.c
@@ -15,7 +15,6 @@
 
 #ifdef CONFIG_X86_32
 #include <asm/pgtable.h>
-#include <asm/pgtable_32.h>
 #endif
 
 #include "realmode/wakeup.h"