Re: [PATCH 00/12] x86: disable virt on kdump and emergency_restart(v4)
From: Ingo Molnar
Date: Tue Nov 18 2008 - 03:15:51 EST
* Eduardo Habkost <ehabkost@xxxxxxxxxx> wrote:
> Hi, Ingo,
>
> This is yet another spin of the series to disable vmx on kdump and
> on emergency_restart, after some feedback from Avi.
this is going to interact with the KVM tree, wont it?
i think the best way forward would be to keep your changes in the KVM
tree.
Lets try a Git trick for that. Avi could do that by pulling your other
x86 changes from the x86 topic tree into the kvm tree. They are
reviewed, acked and well-tested now, and kept in a separate tree so no
other x86 change will be pulled in via them.
We can do this if Avi can guarantee that these commits wont ever be
rebased within KVM - then the two trees will merge up just fine in
linux-next (and later in v2.6.29 as well), without any awkward merge
dependencies or merge conflicts.
See the pull request below.
Ingo
------------>
Please pull the latest x86/crashdump git tree from:
git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip.git x86/crashdump
out-of-topic modifications in x86/crashdump:
--------------------------------------------
include/linux/kexec.h # 92be3d6: kexec/i386: allocate page table p
Thanks,
Ingo
------------------>
Eduardo Habkost (8):
x86 kdump: extract kdump-specific code from crash_nmi_callback()
x86 kdump: move crashing_cpu assignment to nmi_shootdown_cpus()
x86 kdump: create kdump_nmi_shootdown_cpus()
x86 kdump: make kdump_nmi_callback() a function ptr on crash_nmi_callback()
x86 kdump: make nmi_shootdown_cpus() non-static
x86: move nmi_shootdown_cpus() to reboot.c
x86: make nmi_shootdown_cpus() available on !SMP and !X86_LOCAL_APIC
x86: disable IRQs before doing anything on nmi_shootdown_cpus()
Huang Ying (3):
kexec/i386: remove PAGE_SIZE alignment from relocate_kernel
kexec/i386: allocate page table pages dynamically
kexec/i386: setup kexec page table in C
arch/x86/include/asm/kexec.h | 31 +++++-----
arch/x86/include/asm/reboot.h | 5 ++
arch/x86/kernel/crash.c | 70 +++------------------
arch/x86/kernel/machine_kexec_32.c | 104 ++++++++++++++++++++++++-------
arch/x86/kernel/reboot.c | 95 ++++++++++++++++++++++++++++
arch/x86/kernel/relocate_kernel_32.S | 115 ----------------------------------
include/linux/kexec.h | 4 +
7 files changed, 210 insertions(+), 214 deletions(-)
diff --git a/arch/x86/include/asm/kexec.h b/arch/x86/include/asm/kexec.h
index a1f2277..c61d8b2 100644
--- a/arch/x86/include/asm/kexec.h
+++ b/arch/x86/include/asm/kexec.h
@@ -5,21 +5,8 @@
# define PA_CONTROL_PAGE 0
# define VA_CONTROL_PAGE 1
# define PA_PGD 2
-# define VA_PGD 3
-# define PA_PTE_0 4
-# define VA_PTE_0 5
-# define PA_PTE_1 6
-# define VA_PTE_1 7
-# define PA_SWAP_PAGE 8
-# ifdef CONFIG_X86_PAE
-# define PA_PMD_0 9
-# define VA_PMD_0 10
-# define PA_PMD_1 11
-# define VA_PMD_1 12
-# define PAGES_NR 13
-# else
-# define PAGES_NR 9
-# endif
+# define PA_SWAP_PAGE 3
+# define PAGES_NR 4
#else
# define PA_CONTROL_PAGE 0
# define VA_CONTROL_PAGE 1
@@ -170,6 +157,20 @@ relocate_kernel(unsigned long indirection_page,
unsigned long start_address) ATTRIB_NORET;
#endif
+#ifdef CONFIG_X86_32
+#define ARCH_HAS_KIMAGE_ARCH
+
+struct kimage_arch {
+ pgd_t *pgd;
+#ifdef CONFIG_X86_PAE
+ pmd_t *pmd0;
+ pmd_t *pmd1;
+#endif
+ pte_t *pte0;
+ pte_t *pte1;
+};
+#endif
+
#endif /* __ASSEMBLY__ */
#endif /* _ASM_X86_KEXEC_H */
diff --git a/arch/x86/include/asm/reboot.h b/arch/x86/include/asm/reboot.h
index df77103..562d4fd 100644
--- a/arch/x86/include/asm/reboot.h
+++ b/arch/x86/include/asm/reboot.h
@@ -1,6 +1,8 @@
#ifndef _ASM_X86_REBOOT_H
#define _ASM_X86_REBOOT_H
+#include <linux/kdebug.h>
+
struct pt_regs;
struct machine_ops {
@@ -18,4 +20,7 @@ void native_machine_crash_shutdown(struct pt_regs *regs);
void native_machine_shutdown(void);
void machine_real_restart(const unsigned char *code, int length);
+typedef void (*nmi_shootdown_cb)(int, struct die_args*);
+void nmi_shootdown_cpus(nmi_shootdown_cb callback);
+
#endif /* _ASM_X86_REBOOT_H */
diff --git a/arch/x86/kernel/crash.c b/arch/x86/kernel/crash.c
index 2685538..d84a852 100644
--- a/arch/x86/kernel/crash.c
+++ b/arch/x86/kernel/crash.c
@@ -29,34 +29,17 @@
#include <mach_ipi.h>
-/* This keeps a track of which one is crashing cpu. */
-static int crashing_cpu;
#if defined(CONFIG_SMP) && defined(CONFIG_X86_LOCAL_APIC)
-static atomic_t waiting_for_crash_ipi;
-static int crash_nmi_callback(struct notifier_block *self,
- unsigned long val, void *data)
+static void kdump_nmi_callback(int cpu, struct die_args *args)
{
struct pt_regs *regs;
#ifdef CONFIG_X86_32
struct pt_regs fixed_regs;
#endif
- int cpu;
- if (val != DIE_NMI_IPI)
- return NOTIFY_OK;
-
- regs = ((struct die_args *)data)->regs;
- cpu = raw_smp_processor_id();
-
- /* Don't do anything if this handler is invoked on crashing cpu.
- * Otherwise, system will completely hang. Crashing cpu can get
- * an NMI if system was initially booted with nmi_watchdog parameter.
- */
- if (cpu == crashing_cpu)
- return NOTIFY_STOP;
- local_irq_disable();
+ regs = args->regs;
#ifdef CONFIG_X86_32
if (!user_mode_vm(regs)) {
@@ -65,54 +48,19 @@ static int crash_nmi_callback(struct notifier_block *self,
}
#endif
crash_save_cpu(regs, cpu);
- disable_local_APIC();
- atomic_dec(&waiting_for_crash_ipi);
- /* Assume hlt works */
- halt();
- for (;;)
- cpu_relax();
-
- return 1;
-}
-static void smp_send_nmi_allbutself(void)
-{
- cpumask_t mask = cpu_online_map;
- cpu_clear(safe_smp_processor_id(), mask);
- if (!cpus_empty(mask))
- send_IPI_mask(mask, NMI_VECTOR);
+ disable_local_APIC();
}
-static struct notifier_block crash_nmi_nb = {
- .notifier_call = crash_nmi_callback,
-};
-
-static void nmi_shootdown_cpus(void)
+static void kdump_nmi_shootdown_cpus(void)
{
- unsigned long msecs;
-
- atomic_set(&waiting_for_crash_ipi, num_online_cpus() - 1);
- /* Would it be better to replace the trap vector here? */
- if (register_die_notifier(&crash_nmi_nb))
- return; /* return what? */
- /* Ensure the new callback function is set before sending
- * out the NMI
- */
- wmb();
+ nmi_shootdown_cpus(kdump_nmi_callback);
- smp_send_nmi_allbutself();
-
- msecs = 1000; /* Wait at most a second for the other cpus to stop */
- while ((atomic_read(&waiting_for_crash_ipi) > 0) && msecs) {
- mdelay(1);
- msecs--;
- }
-
- /* Leave the nmi callback set */
disable_local_APIC();
}
+
#else
-static void nmi_shootdown_cpus(void)
+static void kdump_nmi_shootdown_cpus(void)
{
/* There are no cpus to shootdown */
}
@@ -131,9 +79,7 @@ void native_machine_crash_shutdown(struct pt_regs *regs)
/* The kernel is broken so disable interrupts */
local_irq_disable();
- /* Make a note of crashing cpu. Will be used in NMI callback.*/
- crashing_cpu = safe_smp_processor_id();
- nmi_shootdown_cpus();
+ kdump_nmi_shootdown_cpus();
lapic_shutdown();
#if defined(CONFIG_X86_IO_APIC)
disable_IO_APIC();
diff --git a/arch/x86/kernel/machine_kexec_32.c b/arch/x86/kernel/machine_kexec_32.c
index 7a38574..37f4200 100644
--- a/arch/x86/kernel/machine_kexec_32.c
+++ b/arch/x86/kernel/machine_kexec_32.c
@@ -13,6 +13,7 @@
#include <linux/numa.h>
#include <linux/ftrace.h>
#include <linux/suspend.h>
+#include <linux/gfp.h>
#include <asm/pgtable.h>
#include <asm/pgalloc.h>
@@ -25,15 +26,6 @@
#include <asm/system.h>
#include <asm/cacheflush.h>
-#define PAGE_ALIGNED __attribute__ ((__aligned__(PAGE_SIZE)))
-static u32 kexec_pgd[1024] PAGE_ALIGNED;
-#ifdef CONFIG_X86_PAE
-static u32 kexec_pmd0[1024] PAGE_ALIGNED;
-static u32 kexec_pmd1[1024] PAGE_ALIGNED;
-#endif
-static u32 kexec_pte0[1024] PAGE_ALIGNED;
-static u32 kexec_pte1[1024] PAGE_ALIGNED;
-
static void set_idt(void *newidt, __u16 limit)
{
struct desc_ptr curidt;
@@ -76,6 +68,76 @@ static void load_segments(void)
#undef __STR
}
+static void machine_kexec_free_page_tables(struct kimage *image)
+{
+ free_page((unsigned long)image->arch.pgd);
+#ifdef CONFIG_X86_PAE
+ free_page((unsigned long)image->arch.pmd0);
+ free_page((unsigned long)image->arch.pmd1);
+#endif
+ free_page((unsigned long)image->arch.pte0);
+ free_page((unsigned long)image->arch.pte1);
+}
+
+static int machine_kexec_alloc_page_tables(struct kimage *image)
+{
+ image->arch.pgd = (pgd_t *)get_zeroed_page(GFP_KERNEL);
+#ifdef CONFIG_X86_PAE
+ image->arch.pmd0 = (pmd_t *)get_zeroed_page(GFP_KERNEL);
+ image->arch.pmd1 = (pmd_t *)get_zeroed_page(GFP_KERNEL);
+#endif
+ image->arch.pte0 = (pte_t *)get_zeroed_page(GFP_KERNEL);
+ image->arch.pte1 = (pte_t *)get_zeroed_page(GFP_KERNEL);
+ if (!image->arch.pgd ||
+#ifdef CONFIG_X86_PAE
+ !image->arch.pmd0 || !image->arch.pmd1 ||
+#endif
+ !image->arch.pte0 || !image->arch.pte1) {
+ machine_kexec_free_page_tables(image);
+ return -ENOMEM;
+ }
+ return 0;
+}
+
+static void machine_kexec_page_table_set_one(
+ pgd_t *pgd, pmd_t *pmd, pte_t *pte,
+ unsigned long vaddr, unsigned long paddr)
+{
+ pud_t *pud;
+
+ pgd += pgd_index(vaddr);
+#ifdef CONFIG_X86_PAE
+ if (!(pgd_val(*pgd) & _PAGE_PRESENT))
+ set_pgd(pgd, __pgd(__pa(pmd) | _PAGE_PRESENT));
+#endif
+ pud = pud_offset(pgd, vaddr);
+ pmd = pmd_offset(pud, vaddr);
+ if (!(pmd_val(*pmd) & _PAGE_PRESENT))
+ set_pmd(pmd, __pmd(__pa(pte) | _PAGE_TABLE));
+ pte = pte_offset_kernel(pmd, vaddr);
+ set_pte(pte, pfn_pte(paddr >> PAGE_SHIFT, PAGE_KERNEL_EXEC));
+}
+
+static void machine_kexec_prepare_page_tables(struct kimage *image)
+{
+ void *control_page;
+ pmd_t *pmd = 0;
+
+ control_page = page_address(image->control_code_page);
+#ifdef CONFIG_X86_PAE
+ pmd = image->arch.pmd0;
+#endif
+ machine_kexec_page_table_set_one(
+ image->arch.pgd, pmd, image->arch.pte0,
+ (unsigned long)control_page, __pa(control_page));
+#ifdef CONFIG_X86_PAE
+ pmd = image->arch.pmd1;
+#endif
+ machine_kexec_page_table_set_one(
+ image->arch.pgd, pmd, image->arch.pte1,
+ __pa(control_page), __pa(control_page));
+}
+
/*
* A architecture hook called to validate the
* proposed image and prepare the control pages
@@ -87,12 +149,20 @@ static void load_segments(void)
* reboot code buffer to allow us to avoid allocations
* later.
*
- * Make control page executable.
+ * - Make control page executable.
+ * - Allocate page tables
+ * - Setup page tables
*/
int machine_kexec_prepare(struct kimage *image)
{
+ int error;
+
if (nx_enabled)
set_pages_x(image->control_code_page, 1);
+ error = machine_kexec_alloc_page_tables(image);
+ if (error)
+ return error;
+ machine_kexec_prepare_page_tables(image);
return 0;
}
@@ -104,6 +174,7 @@ void machine_kexec_cleanup(struct kimage *image)
{
if (nx_enabled)
set_pages_nx(image->control_code_page, 1);
+ machine_kexec_free_page_tables(image);
}
/*
@@ -150,18 +221,7 @@ void machine_kexec(struct kimage *image)
relocate_kernel_ptr = control_page;
page_list[PA_CONTROL_PAGE] = __pa(control_page);
page_list[VA_CONTROL_PAGE] = (unsigned long)control_page;
- page_list[PA_PGD] = __pa(kexec_pgd);
- page_list[VA_PGD] = (unsigned long)kexec_pgd;
-#ifdef CONFIG_X86_PAE
- page_list[PA_PMD_0] = __pa(kexec_pmd0);
- page_list[VA_PMD_0] = (unsigned long)kexec_pmd0;
- page_list[PA_PMD_1] = __pa(kexec_pmd1);
- page_list[VA_PMD_1] = (unsigned long)kexec_pmd1;
-#endif
- page_list[PA_PTE_0] = __pa(kexec_pte0);
- page_list[VA_PTE_0] = (unsigned long)kexec_pte0;
- page_list[PA_PTE_1] = __pa(kexec_pte1);
- page_list[VA_PTE_1] = (unsigned long)kexec_pte1;
+ page_list[PA_PGD] = __pa(image->arch.pgd);
if (image->type == KEXEC_TYPE_DEFAULT)
page_list[PA_SWAP_PAGE] = (page_to_pfn(image->swap_page)
diff --git a/arch/x86/kernel/reboot.c b/arch/x86/kernel/reboot.c
index 724adfc..c3cd512 100644
--- a/arch/x86/kernel/reboot.c
+++ b/arch/x86/kernel/reboot.c
@@ -21,6 +21,9 @@
# include <asm/iommu.h>
#endif
+#include <mach_ipi.h>
+
+
/*
* Power off function, if any
*/
@@ -514,3 +517,95 @@ void machine_crash_shutdown(struct pt_regs *regs)
machine_ops.crash_shutdown(regs);
}
#endif
+
+
+#if defined(CONFIG_SMP)
+
+/* This keeps a track of which one is crashing cpu. */
+static int crashing_cpu;
+static nmi_shootdown_cb shootdown_callback;
+
+static atomic_t waiting_for_crash_ipi;
+
+static int crash_nmi_callback(struct notifier_block *self,
+ unsigned long val, void *data)
+{
+ int cpu;
+
+ if (val != DIE_NMI_IPI)
+ return NOTIFY_OK;
+
+ cpu = raw_smp_processor_id();
+
+ /* Don't do anything if this handler is invoked on crashing cpu.
+ * Otherwise, system will completely hang. Crashing cpu can get
+ * an NMI if system was initially booted with nmi_watchdog parameter.
+ */
+ if (cpu == crashing_cpu)
+ return NOTIFY_STOP;
+ local_irq_disable();
+
+ shootdown_callback(cpu, (struct die_args *)data);
+
+ atomic_dec(&waiting_for_crash_ipi);
+ /* Assume hlt works */
+ halt();
+ for (;;)
+ cpu_relax();
+
+ return 1;
+}
+
+static void smp_send_nmi_allbutself(void)
+{
+ cpumask_t mask = cpu_online_map;
+ cpu_clear(safe_smp_processor_id(), mask);
+ if (!cpus_empty(mask))
+ send_IPI_mask(mask, NMI_VECTOR);
+}
+
+static struct notifier_block crash_nmi_nb = {
+ .notifier_call = crash_nmi_callback,
+};
+
+/* Halt all other CPUs, calling the specified function on each of them
+ *
+ * This function can be used to halt all other CPUs on crash
+ * or emergency reboot time. The function passed as parameter
+ * will be called inside a NMI handler on all CPUs.
+ */
+void nmi_shootdown_cpus(nmi_shootdown_cb callback)
+{
+ unsigned long msecs;
+ local_irq_disable();
+
+ /* Make a note of crashing cpu. Will be used in NMI callback.*/
+ crashing_cpu = safe_smp_processor_id();
+
+ shootdown_callback = callback;
+
+ atomic_set(&waiting_for_crash_ipi, num_online_cpus() - 1);
+ /* Would it be better to replace the trap vector here? */
+ if (register_die_notifier(&crash_nmi_nb))
+ return; /* return what? */
+ /* Ensure the new callback function is set before sending
+ * out the NMI
+ */
+ wmb();
+
+ smp_send_nmi_allbutself();
+
+ msecs = 1000; /* Wait at most a second for the other cpus to stop */
+ while ((atomic_read(&waiting_for_crash_ipi) > 0) && msecs) {
+ mdelay(1);
+ msecs--;
+ }
+
+ /* Leave the nmi callback set */
+}
+#else /* !CONFIG_SMP */
+void nmi_shootdown_cpus(nmi_shootdown_cb callback)
+{
+ /* No other CPUs to shoot down */
+}
+#endif
diff --git a/arch/x86/kernel/relocate_kernel_32.S b/arch/x86/kernel/relocate_kernel_32.S
index 6f50664..a160f31 100644
--- a/arch/x86/kernel/relocate_kernel_32.S
+++ b/arch/x86/kernel/relocate_kernel_32.S
@@ -10,15 +10,12 @@
#include <asm/page.h>
#include <asm/kexec.h>
#include <asm/processor-flags.h>
-#include <asm/pgtable.h>
/*
* Must be relocatable PIC code callable as a C function
*/
#define PTR(x) (x << 2)
-#define PAGE_ATTR (_PAGE_PRESENT | _PAGE_RW | _PAGE_ACCESSED | _PAGE_DIRTY)
-#define PAE_PGD_ATTR (_PAGE_PRESENT)
/* control_page + KEXEC_CONTROL_CODE_MAX_SIZE
* ~ control_page + PAGE_SIZE are used as data storage and stack for
@@ -39,7 +36,6 @@
#define CP_PA_BACKUP_PAGES_MAP DATA(0x1c)
.text
- .align PAGE_SIZE
.globl relocate_kernel
relocate_kernel:
/* Save the CPU context, used for jumping back */
@@ -60,117 +56,6 @@ relocate_kernel:
movl %cr4, %eax
movl %eax, CR4(%edi)
-#ifdef CONFIG_X86_PAE
- /* map the control page at its virtual address */
-
- movl PTR(VA_PGD)(%ebp), %edi
- movl PTR(VA_CONTROL_PAGE)(%ebp), %eax
- andl $0xc0000000, %eax
- shrl $27, %eax
- addl %edi, %eax
-
- movl PTR(PA_PMD_0)(%ebp), %edx
- orl $PAE_PGD_ATTR, %edx
- movl %edx, (%eax)
-
- movl PTR(VA_PMD_0)(%ebp), %edi
- movl PTR(VA_CONTROL_PAGE)(%ebp), %eax
- andl $0x3fe00000, %eax
- shrl $18, %eax
- addl %edi, %eax
-
- movl PTR(PA_PTE_0)(%ebp), %edx
- orl $PAGE_ATTR, %edx
- movl %edx, (%eax)
-
- movl PTR(VA_PTE_0)(%ebp), %edi
- movl PTR(VA_CONTROL_PAGE)(%ebp), %eax
- andl $0x001ff000, %eax
- shrl $9, %eax
- addl %edi, %eax
-
- movl PTR(PA_CONTROL_PAGE)(%ebp), %edx
- orl $PAGE_ATTR, %edx
- movl %edx, (%eax)
-
- /* identity map the control page at its physical address */
-
- movl PTR(VA_PGD)(%ebp), %edi
- movl PTR(PA_CONTROL_PAGE)(%ebp), %eax
- andl $0xc0000000, %eax
- shrl $27, %eax
- addl %edi, %eax
-
- movl PTR(PA_PMD_1)(%ebp), %edx
- orl $PAE_PGD_ATTR, %edx
- movl %edx, (%eax)
-
- movl PTR(VA_PMD_1)(%ebp), %edi
- movl PTR(PA_CONTROL_PAGE)(%ebp), %eax
- andl $0x3fe00000, %eax
- shrl $18, %eax
- addl %edi, %eax
-
- movl PTR(PA_PTE_1)(%ebp), %edx
- orl $PAGE_ATTR, %edx
- movl %edx, (%eax)
-
- movl PTR(VA_PTE_1)(%ebp), %edi
- movl PTR(PA_CONTROL_PAGE)(%ebp), %eax
- andl $0x001ff000, %eax
- shrl $9, %eax
- addl %edi, %eax
-
- movl PTR(PA_CONTROL_PAGE)(%ebp), %edx
- orl $PAGE_ATTR, %edx
- movl %edx, (%eax)
-#else
- /* map the control page at its virtual address */
-
- movl PTR(VA_PGD)(%ebp), %edi
- movl PTR(VA_CONTROL_PAGE)(%ebp), %eax
- andl $0xffc00000, %eax
- shrl $20, %eax
- addl %edi, %eax
-
- movl PTR(PA_PTE_0)(%ebp), %edx
- orl $PAGE_ATTR, %edx
- movl %edx, (%eax)
-
- movl PTR(VA_PTE_0)(%ebp), %edi
- movl PTR(VA_CONTROL_PAGE)(%ebp), %eax
- andl $0x003ff000, %eax
- shrl $10, %eax
- addl %edi, %eax
-
- movl PTR(PA_CONTROL_PAGE)(%ebp), %edx
- orl $PAGE_ATTR, %edx
- movl %edx, (%eax)
-
- /* identity map the control page at its physical address */
-
- movl PTR(VA_PGD)(%ebp), %edi
- movl PTR(PA_CONTROL_PAGE)(%ebp), %eax
- andl $0xffc00000, %eax
- shrl $20, %eax
- addl %edi, %eax
-
- movl PTR(PA_PTE_1)(%ebp), %edx
- orl $PAGE_ATTR, %edx
- movl %edx, (%eax)
-
- movl PTR(VA_PTE_1)(%ebp), %edi
- movl PTR(PA_CONTROL_PAGE)(%ebp), %eax
- andl $0x003ff000, %eax
- shrl $10, %eax
- addl %edi, %eax
-
- movl PTR(PA_CONTROL_PAGE)(%ebp), %edx
- orl $PAGE_ATTR, %edx
- movl %edx, (%eax)
-#endif
-
-relocate_new_kernel:
/* read the arguments and say goodbye to the stack */
movl 20+4(%esp), %ebx /* page_list */
movl 20+8(%esp), %ebp /* list of pages */
diff --git a/include/linux/kexec.h b/include/linux/kexec.h
index 17f76fc..adc34f2 100644
--- a/include/linux/kexec.h
+++ b/include/linux/kexec.h
@@ -100,6 +100,10 @@ struct kimage {
#define KEXEC_TYPE_DEFAULT 0
#define KEXEC_TYPE_CRASH 1
unsigned int preserve_context : 1;
+
+#ifdef ARCH_HAS_KIMAGE_ARCH
+ struct kimage_arch arch;
+#endif
};
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/