linux/arch/x86/kernel/reboot.c

835 lines
21 KiB
C
Raw Normal View History

#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
#include <linux/module.h>
#include <linux/reboot.h>
#include <linux/init.h>
#include <linux/pm.h>
#include <linux/efi.h>
#include <linux/dmi.h>
#include <linux/sched.h>
#include <linux/tboot.h>
#include <linux/delay.h>
#include <acpi/reboot.h>
#include <asm/io.h>
#include <asm/apic.h>
#include <asm/io_apic.h>
#include <asm/desc.h>
#include <asm/hpet.h>
#include <asm/pgtable.h>
#include <asm/proto.h>
#include <asm/reboot_fixups.h>
#include <asm/reboot.h>
#include <asm/pci_x86.h>
#include <asm/virtext.h>
#include <asm/cpu.h>
#include <asm/nmi.h>
#include <asm/smp.h>
#include <linux/ctype.h>
#include <linux/mc146818rtc.h>
#include <asm/realmode.h>
#include <asm/x86_init.h>
#include <asm/efi.h>
/*
* Power off function, if any
*/
void (*pm_power_off)(void);
EXPORT_SYMBOL(pm_power_off);
static const struct desc_ptr no_idt = {};
/*
* This is set if we need to go through the 'emergency' path.
* When machine_emergency_restart() is called, we may be on
* an inconsistent state and won't be able to do a clean cleanup
*/
static int reboot_emergency;
/* This is set by the PCI code if either type 1 or type 2 PCI is detected */
bool port_cf9_safe = false;
/*
* Reboot options and system auto-detection code provided by
* Dell Inc. so their systems "just work". :-)
*/
/*
* Some machines require the "reboot=b" or "reboot=k" commandline options,
* this quirk makes that automatic.
*/
static int __init set_bios_reboot(const struct dmi_system_id *d)
{
if (reboot_type != BOOT_BIOS) {
reboot_type = BOOT_BIOS;
pr_info("%s series board detected. Selecting %s-method for reboots.\n",
d->ident, "BIOS");
}
return 0;
}
void __noreturn machine_real_restart(unsigned int type)
{
local_irq_disable();
/*
* Write zero to CMOS register number 0x0f, which the BIOS POST
* routine will recognize as telling it to do a proper reboot. (Well
* that's what this book in front of me says -- it may only apply to
* the Phoenix BIOS though, it's not clear). At the same time,
* disable NMIs by setting the top bit in the CMOS address register,
* as we're about to do peculiar things to the CPU. I'm not sure if
* `outb_p' is needed instead of just `outb'. Use it to be on the
* safe side. (Yes, CMOS_WRITE does outb_p's. - Paul G.)
*/
spin_lock(&rtc_lock);
CMOS_WRITE(0x00, 0x8f);
spin_unlock(&rtc_lock);
/*
* Switch back to the initial page table.
*/
#ifdef CONFIG_X86_32
load_cr3(initial_page_table);
#else
write_cr3(real_mode_header->trampoline_pgd);
#endif
/* Jump to the identity-mapped low memory code */
#ifdef CONFIG_X86_32
asm volatile("jmpl *%0" : :
"rm" (real_mode_header->machine_real_restart_asm),
"a" (type));
#else
asm volatile("ljmpl *%0" : :
"m" (real_mode_header->machine_real_restart_asm),
"D" (type));
#endif
unreachable();
}
#ifdef CONFIG_APM_MODULE
EXPORT_SYMBOL(machine_real_restart);
#endif
/*
* Some Apple MacBook and MacBookPro's needs reboot=p to be able to reboot
*/
static int __init set_pci_reboot(const struct dmi_system_id *d)
{
x86: Remove the PCI reboot method from the default chain Steve reported a reboot hang and bisected it back to this commit: a4f1987e4c54 x86, reboot: Add EFI and CF9 reboot methods into the default list He heroically tested all reboot methods and found the following: reboot=t # triple fault ok reboot=k # keyboard ctrl FAIL reboot=b # BIOS ok reboot=a # ACPI FAIL reboot=e # EFI FAIL [system has no EFI] reboot=p # PCI 0xcf9 FAIL And I think it's pretty obvious that we should only try PCI 0xcf9 as a last resort - if at all. The other observation is that (on this box) we should never try the PCI reboot method, but close with either the 'triple fault' or the 'BIOS' (terminal!) reboot methods. Thirdly, CF9_COND is a total misnomer - it should be something like CF9_SAFE or CF9_CAREFUL, and 'CF9' should be 'CF9_FORCE' ... So this patch fixes the worst problems: - it orders the actual reboot logic to follow the reboot ordering pattern - it was in a pretty random order before for no good reason. - it fixes the CF9 misnomers and uses BOOT_CF9_FORCE and BOOT_CF9_SAFE flags to make the code more obvious. - it tries the BIOS reboot method before the PCI reboot method. (Since 'BIOS' is a terminal reboot method resulting in a hang if it does not work, this is essentially equivalent to removing the PCI reboot method from the default reboot chain.) - just for the miraculous possibility of terminal (resulting in hang) reboot methods of triple fault or BIOS returning without having done their job, there's an ordering between them as well. Reported-and-bisected-and-tested-by: Steven Rostedt <rostedt@goodmis.org> Cc: Li Aubrey <aubrey.li@linux.intel.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Matthew Garrett <mjg59@srcf.ucam.org> Link: http://lkml.kernel.org/r/20140404064120.GB11877@gmail.com Signed-off-by: Ingo Molnar <mingo@kernel.org>
2014-04-04 14:41:26 +08:00
if (reboot_type != BOOT_CF9_FORCE) {
reboot_type = BOOT_CF9_FORCE;
pr_info("%s series board detected. Selecting %s-method for reboots.\n",
d->ident, "PCI");
}
return 0;
}
static int __init set_kbd_reboot(const struct dmi_system_id *d)
{
if (reboot_type != BOOT_KBD) {
reboot_type = BOOT_KBD;
pr_info("%s series board detected. Selecting %s-method for reboot.\n",
d->ident, "KBD");
}
return 0;
}
/*
* This is a single dmi_table handling all reboot quirks.
*/
static struct dmi_system_id __initdata reboot_dmi_table[] = {
/* Acer */
{ /* Handle reboot issue on Acer Aspire one */
.callback = set_kbd_reboot,
.ident = "Acer Aspire One A110",
.matches = {
DMI_MATCH(DMI_SYS_VENDOR, "Acer"),
DMI_MATCH(DMI_PRODUCT_NAME, "AOA110"),
},
},
/* Apple */
{ /* Handle problems with rebooting on Apple MacBook5 */
.callback = set_pci_reboot,
.ident = "Apple MacBook5",
.matches = {
DMI_MATCH(DMI_SYS_VENDOR, "Apple Inc."),
DMI_MATCH(DMI_PRODUCT_NAME, "MacBook5"),
},
},
{ /* Handle problems with rebooting on Apple MacBookPro5 */
.callback = set_pci_reboot,
.ident = "Apple MacBookPro5",
.matches = {
DMI_MATCH(DMI_SYS_VENDOR, "Apple Inc."),
DMI_MATCH(DMI_PRODUCT_NAME, "MacBookPro5"),
},
},
{ /* Handle problems with rebooting on Apple Macmini3,1 */
.callback = set_pci_reboot,
.ident = "Apple Macmini3,1",
.matches = {
DMI_MATCH(DMI_SYS_VENDOR, "Apple Inc."),
DMI_MATCH(DMI_PRODUCT_NAME, "Macmini3,1"),
},
},
{ /* Handle problems with rebooting on the iMac9,1. */
.callback = set_pci_reboot,
.ident = "Apple iMac9,1",
.matches = {
DMI_MATCH(DMI_SYS_VENDOR, "Apple Inc."),
DMI_MATCH(DMI_PRODUCT_NAME, "iMac9,1"),
},
},
{ /* Handle problems with rebooting on the iMac10,1. */
.callback = set_pci_reboot,
.ident = "Apple iMac10,1",
.matches = {
DMI_MATCH(DMI_SYS_VENDOR, "Apple Inc."),
DMI_MATCH(DMI_PRODUCT_NAME, "iMac10,1"),
},
},
/* ASRock */
{ /* Handle problems with rebooting on ASRock Q1900DC-ITX */
.callback = set_pci_reboot,
.ident = "ASRock Q1900DC-ITX",
.matches = {
DMI_MATCH(DMI_BOARD_VENDOR, "ASRock"),
DMI_MATCH(DMI_BOARD_NAME, "Q1900DC-ITX"),
},
},
/* ASUS */
{ /* Handle problems with rebooting on ASUS P4S800 */
.callback = set_bios_reboot,
.ident = "ASUS P4S800",
.matches = {
DMI_MATCH(DMI_BOARD_VENDOR, "ASUSTeK Computer INC."),
DMI_MATCH(DMI_BOARD_NAME, "P4S800"),
},
},
/* Certec */
{ /* Handle problems with rebooting on Certec BPC600 */
.callback = set_pci_reboot,
.ident = "Certec BPC600",
.matches = {
DMI_MATCH(DMI_SYS_VENDOR, "Certec"),
DMI_MATCH(DMI_PRODUCT_NAME, "BPC600"),
},
},
/* Dell */
{ /* Handle problems with rebooting on Dell DXP061 */
.callback = set_bios_reboot,
.ident = "Dell DXP061",
.matches = {
DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc."),
DMI_MATCH(DMI_PRODUCT_NAME, "Dell DXP061"),
},
},
{ /* Handle problems with rebooting on Dell E520's */
.callback = set_bios_reboot,
.ident = "Dell E520",
.matches = {
DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc."),
DMI_MATCH(DMI_PRODUCT_NAME, "Dell DM061"),
},
},
{ /* Handle problems with rebooting on the Latitude E5410. */
.callback = set_pci_reboot,
.ident = "Dell Latitude E5410",
.matches = {
DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc."),
DMI_MATCH(DMI_PRODUCT_NAME, "Latitude E5410"),
},
},
{ /* Handle problems with rebooting on the Latitude E5420. */
.callback = set_pci_reboot,
.ident = "Dell Latitude E5420",
.matches = {
DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc."),
DMI_MATCH(DMI_PRODUCT_NAME, "Latitude E5420"),
},
},
{ /* Handle problems with rebooting on the Latitude E6320. */
.callback = set_pci_reboot,
.ident = "Dell Latitude E6320",
.matches = {
DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc."),
DMI_MATCH(DMI_PRODUCT_NAME, "Latitude E6320"),
},
},
{ /* Handle problems with rebooting on the Latitude E6420. */
.callback = set_pci_reboot,
.ident = "Dell Latitude E6420",
.matches = {
DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc."),
DMI_MATCH(DMI_PRODUCT_NAME, "Latitude E6420"),
},
},
{ /* Handle problems with rebooting on Dell Optiplex 330 with 0KP561 */
.callback = set_bios_reboot,
.ident = "Dell OptiPlex 330",
.matches = {
DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc."),
DMI_MATCH(DMI_PRODUCT_NAME, "OptiPlex 330"),
DMI_MATCH(DMI_BOARD_NAME, "0KP561"),
},
},
{ /* Handle problems with rebooting on Dell Optiplex 360 with 0T656F */
.callback = set_bios_reboot,
.ident = "Dell OptiPlex 360",
.matches = {
DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc."),
DMI_MATCH(DMI_PRODUCT_NAME, "OptiPlex 360"),
DMI_MATCH(DMI_BOARD_NAME, "0T656F"),
},
},
{ /* Handle problems with rebooting on Dell Optiplex 745's SFF */
.callback = set_bios_reboot,
.ident = "Dell OptiPlex 745",
.matches = {
DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc."),
DMI_MATCH(DMI_PRODUCT_NAME, "OptiPlex 745"),
},
},
{ /* Handle problems with rebooting on Dell Optiplex 745's DFF */
.callback = set_bios_reboot,
.ident = "Dell OptiPlex 745",
.matches = {
DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc."),
DMI_MATCH(DMI_PRODUCT_NAME, "OptiPlex 745"),
DMI_MATCH(DMI_BOARD_NAME, "0MM599"),
},
},
{ /* Handle problems with rebooting on Dell Optiplex 745 with 0KW626 */
.callback = set_bios_reboot,
.ident = "Dell OptiPlex 745",
.matches = {
DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc."),
DMI_MATCH(DMI_PRODUCT_NAME, "OptiPlex 745"),
DMI_MATCH(DMI_BOARD_NAME, "0KW626"),
},
},
{ /* Handle problems with rebooting on Dell OptiPlex 760 with 0G919G */
.callback = set_bios_reboot,
.ident = "Dell OptiPlex 760",
.matches = {
DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc."),
DMI_MATCH(DMI_PRODUCT_NAME, "OptiPlex 760"),
DMI_MATCH(DMI_BOARD_NAME, "0G919G"),
},
},
{ /* Handle problems with rebooting on the OptiPlex 990. */
.callback = set_pci_reboot,
.ident = "Dell OptiPlex 990",
.matches = {
DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc."),
DMI_MATCH(DMI_PRODUCT_NAME, "OptiPlex 990"),
},
},
{ /* Handle problems with rebooting on Dell 300's */
.callback = set_bios_reboot,
.ident = "Dell PowerEdge 300",
.matches = {
DMI_MATCH(DMI_SYS_VENDOR, "Dell Computer Corporation"),
DMI_MATCH(DMI_PRODUCT_NAME, "PowerEdge 300/"),
},
},
{ /* Handle problems with rebooting on Dell 1300's */
.callback = set_bios_reboot,
.ident = "Dell PowerEdge 1300",
.matches = {
DMI_MATCH(DMI_SYS_VENDOR, "Dell Computer Corporation"),
DMI_MATCH(DMI_PRODUCT_NAME, "PowerEdge 1300/"),
},
},
{ /* Handle problems with rebooting on Dell 2400's */
.callback = set_bios_reboot,
.ident = "Dell PowerEdge 2400",
.matches = {
DMI_MATCH(DMI_SYS_VENDOR, "Dell Computer Corporation"),
DMI_MATCH(DMI_PRODUCT_NAME, "PowerEdge 2400"),
},
},
{ /* Handle problems with rebooting on the Dell PowerEdge C6100. */
.callback = set_pci_reboot,
.ident = "Dell PowerEdge C6100",
.matches = {
DMI_MATCH(DMI_SYS_VENDOR, "Dell"),
DMI_MATCH(DMI_PRODUCT_NAME, "C6100"),
},
},
{ /* Handle problems with rebooting on the Precision M6600. */
.callback = set_pci_reboot,
.ident = "Dell Precision M6600",
.matches = {
DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc."),
DMI_MATCH(DMI_PRODUCT_NAME, "Precision M6600"),
},
},
{ /* Handle problems with rebooting on Dell T5400's */
.callback = set_bios_reboot,
.ident = "Dell Precision T5400",
.matches = {
DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc."),
DMI_MATCH(DMI_PRODUCT_NAME, "Precision WorkStation T5400"),
},
},
{ /* Handle problems with rebooting on Dell T7400's */
.callback = set_bios_reboot,
.ident = "Dell Precision T7400",
.matches = {
DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc."),
DMI_MATCH(DMI_PRODUCT_NAME, "Precision WorkStation T7400"),
},
},
{ /* Handle problems with rebooting on Dell XPS710 */
.callback = set_bios_reboot,
.ident = "Dell XPS710",
.matches = {
DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc."),
DMI_MATCH(DMI_PRODUCT_NAME, "Dell XPS710"),
},
},
/* Hewlett-Packard */
{ /* Handle problems with rebooting on HP laptops */
.callback = set_bios_reboot,
.ident = "HP Compaq Laptop",
.matches = {
DMI_MATCH(DMI_SYS_VENDOR, "Hewlett-Packard"),
DMI_MATCH(DMI_PRODUCT_NAME, "HP Compaq"),
},
},
/* Sony */
{ /* Handle problems with rebooting on Sony VGN-Z540N */
.callback = set_bios_reboot,
.ident = "Sony VGN-Z540N",
.matches = {
DMI_MATCH(DMI_SYS_VENDOR, "Sony Corporation"),
DMI_MATCH(DMI_PRODUCT_NAME, "VGN-Z540N"),
},
},
{ }
};
static int __init reboot_init(void)
{
int rv;
/*
* Only do the DMI check if reboot_type hasn't been overridden
* on the command line
*/
if (!reboot_default)
return 0;
/*
* The DMI quirks table takes precedence. If no quirks entry
* matches and the ACPI Hardware Reduced bit is set, force EFI
* reboot.
*/
rv = dmi_check_system(reboot_dmi_table);
if (!rv && efi_reboot_required())
reboot_type = BOOT_EFI;
return 0;
}
core_initcall(reboot_init);
static inline void kb_wait(void)
{
int i;
for (i = 0; i < 0x10000; i++) {
if ((inb(0x64) & 0x02) == 0)
break;
udelay(2);
}
}
static void vmxoff_nmi(int cpu, struct pt_regs *regs)
{
cpu_emergency_vmxoff();
}
/* Use NMIs as IPIs to tell all CPUs to disable virtualization */
static void emergency_vmx_disable_all(void)
{
/* Just make sure we won't change CPUs while doing this */
local_irq_disable();
/*
* We need to disable VMX on all CPUs before rebooting, otherwise
* we risk hanging up the machine, because the CPU ignore INIT
* signals when VMX is enabled.
*
* We can't take any locks and we may be on an inconsistent
* state, so we use NMIs as IPIs to tell the other CPUs to disable
* VMX and halt.
*
* For safety, we will avoid running the nmi_shootdown_cpus()
* stuff unnecessarily, but we don't have a way to check
* if other CPUs have VMX enabled. So we will call it only if the
* CPU we are running on has VMX enabled.
*
* We will miss cases where VMX is not enabled on all CPUs. This
* shouldn't do much harm because KVM always enable VMX on all
* CPUs anyway. But we can miss it on the small window where KVM
* is still enabling VMX.
*/
if (cpu_has_vmx() && cpu_vmx_enabled()) {
/* Disable VMX on this CPU. */
cpu_vmxoff();
/* Halt and disable VMX on the other CPUs */
nmi_shootdown_cpus(vmxoff_nmi);
}
}
void __attribute__((weak)) mach_reboot_fixups(void)
{
}
x86: Reorder reboot method preferences We have a never ending stream of 'reboot quirks' for new boxes that will not reboot properly under Linux (they will hang on reboot). The reason is widespread 'Windows compatible' assumption of modern x86 hardware, which expects the following reboot sequence: - hitting the ACPI reboot vector (if available) - trying the keyboard controller - hitting the ACPI reboot vector again - then giving the keyboard controller one last go This sequence expectation gets more and more embedded in modern hardware, which often lacks a keyboard controller and may even lock up if the legacy io ports are hit - and which hardware is often not tested with Linux during development. The end result is that reboot works under Windows-alike OSs but not under Linux. Rework our reboot process to meet this hardware externality a little better and match this assumption of newer x86 hardware. In addition to the ACPI,kbd,ACPI,kbd sequence we'll still fall through to attempting a legacy triple fault if nothing else works - and keep trying that and the kbd reset. Signed-off-by: Matthew Garrett <mjg@redhat.com> [ this commit will also save special casing Oaktrail boards ] Acked-by: Alan Cox <alan@linux.intel.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Leann Ogasawara <leann.ogasawara@canonical.com> Cc: Dave Jones <davej@redhat.com> Cc: Len Brown <len.brown@intel.com> LKML-Reference: <1301939705-2404-1-git-send-email-mjg@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2011-04-05 01:55:05 +08:00
/*
x86: Remove the PCI reboot method from the default chain Steve reported a reboot hang and bisected it back to this commit: a4f1987e4c54 x86, reboot: Add EFI and CF9 reboot methods into the default list He heroically tested all reboot methods and found the following: reboot=t # triple fault ok reboot=k # keyboard ctrl FAIL reboot=b # BIOS ok reboot=a # ACPI FAIL reboot=e # EFI FAIL [system has no EFI] reboot=p # PCI 0xcf9 FAIL And I think it's pretty obvious that we should only try PCI 0xcf9 as a last resort - if at all. The other observation is that (on this box) we should never try the PCI reboot method, but close with either the 'triple fault' or the 'BIOS' (terminal!) reboot methods. Thirdly, CF9_COND is a total misnomer - it should be something like CF9_SAFE or CF9_CAREFUL, and 'CF9' should be 'CF9_FORCE' ... So this patch fixes the worst problems: - it orders the actual reboot logic to follow the reboot ordering pattern - it was in a pretty random order before for no good reason. - it fixes the CF9 misnomers and uses BOOT_CF9_FORCE and BOOT_CF9_SAFE flags to make the code more obvious. - it tries the BIOS reboot method before the PCI reboot method. (Since 'BIOS' is a terminal reboot method resulting in a hang if it does not work, this is essentially equivalent to removing the PCI reboot method from the default reboot chain.) - just for the miraculous possibility of terminal (resulting in hang) reboot methods of triple fault or BIOS returning without having done their job, there's an ordering between them as well. Reported-and-bisected-and-tested-by: Steven Rostedt <rostedt@goodmis.org> Cc: Li Aubrey <aubrey.li@linux.intel.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Matthew Garrett <mjg59@srcf.ucam.org> Link: http://lkml.kernel.org/r/20140404064120.GB11877@gmail.com Signed-off-by: Ingo Molnar <mingo@kernel.org>
2014-04-04 14:41:26 +08:00
* To the best of our knowledge Windows compatible x86 hardware expects
* the following on reboot:
x86: Reorder reboot method preferences We have a never ending stream of 'reboot quirks' for new boxes that will not reboot properly under Linux (they will hang on reboot). The reason is widespread 'Windows compatible' assumption of modern x86 hardware, which expects the following reboot sequence: - hitting the ACPI reboot vector (if available) - trying the keyboard controller - hitting the ACPI reboot vector again - then giving the keyboard controller one last go This sequence expectation gets more and more embedded in modern hardware, which often lacks a keyboard controller and may even lock up if the legacy io ports are hit - and which hardware is often not tested with Linux during development. The end result is that reboot works under Windows-alike OSs but not under Linux. Rework our reboot process to meet this hardware externality a little better and match this assumption of newer x86 hardware. In addition to the ACPI,kbd,ACPI,kbd sequence we'll still fall through to attempting a legacy triple fault if nothing else works - and keep trying that and the kbd reset. Signed-off-by: Matthew Garrett <mjg@redhat.com> [ this commit will also save special casing Oaktrail boards ] Acked-by: Alan Cox <alan@linux.intel.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Leann Ogasawara <leann.ogasawara@canonical.com> Cc: Dave Jones <davej@redhat.com> Cc: Len Brown <len.brown@intel.com> LKML-Reference: <1301939705-2404-1-git-send-email-mjg@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2011-04-05 01:55:05 +08:00
*
* 1) If the FADT has the ACPI reboot register flag set, try it
* 2) If still alive, write to the keyboard controller
* 3) If still alive, write to the ACPI reboot register again
* 4) If still alive, write to the keyboard controller again
x86, reboot: Add EFI and CF9 reboot methods into the default list Reboot is the last service linux OS provides to the end user. We are supposed to make this function more robust than today. This patch adds all of the known reboot methods into the default attempt list. The machines requiring reboot=efi or reboot=p or reboot=bios get a chance to reboot automatically now. If there is a new reboot method emerged, we are supposed to add it to the default list as well, instead of adding the endless dmidecode entry. If one method required is in the default list in this patch but the machine reboot still hangs, that means some methods ahead of the required method cause the system hangs, then reboot the machine by passing reboot= arguments and submit the reboot dmidecode table quirk. We are supposed to remove the reboot dmidecode table from the kernel, but to be safe, we keep it. This patch prevents us from adding more. If you happened to have a machine listed in the reboot dmidecode table and this patch makes reboot work on your machine, please submit a patch to remove the quirk. The default reboot order with this patch is now: ACPI > KBD > ACPI > KBD > EFI > CF9_COND > BIOS Because BIOS and TRIPLE are mutually exclusive (either will either work or hang the machine) that method is not included. [ hpa: as with any changes to the reboot order, this patch will have to be monitored carefully for regressions. ] Signed-off-by: Aubrey Li <aubrey.li@intel.com> Acked-by: Matthew Garrett <mjg59@srcf.ucam.org> Link: http://lkml.kernel.org/r/53130A46.1010801@linux.intel.com Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
2014-03-02 18:39:02 +08:00
* 5) If still alive, call the EFI runtime service to reboot
x86: Remove the PCI reboot method from the default chain Steve reported a reboot hang and bisected it back to this commit: a4f1987e4c54 x86, reboot: Add EFI and CF9 reboot methods into the default list He heroically tested all reboot methods and found the following: reboot=t # triple fault ok reboot=k # keyboard ctrl FAIL reboot=b # BIOS ok reboot=a # ACPI FAIL reboot=e # EFI FAIL [system has no EFI] reboot=p # PCI 0xcf9 FAIL And I think it's pretty obvious that we should only try PCI 0xcf9 as a last resort - if at all. The other observation is that (on this box) we should never try the PCI reboot method, but close with either the 'triple fault' or the 'BIOS' (terminal!) reboot methods. Thirdly, CF9_COND is a total misnomer - it should be something like CF9_SAFE or CF9_CAREFUL, and 'CF9' should be 'CF9_FORCE' ... So this patch fixes the worst problems: - it orders the actual reboot logic to follow the reboot ordering pattern - it was in a pretty random order before for no good reason. - it fixes the CF9 misnomers and uses BOOT_CF9_FORCE and BOOT_CF9_SAFE flags to make the code more obvious. - it tries the BIOS reboot method before the PCI reboot method. (Since 'BIOS' is a terminal reboot method resulting in a hang if it does not work, this is essentially equivalent to removing the PCI reboot method from the default reboot chain.) - just for the miraculous possibility of terminal (resulting in hang) reboot methods of triple fault or BIOS returning without having done their job, there's an ordering between them as well. Reported-and-bisected-and-tested-by: Steven Rostedt <rostedt@goodmis.org> Cc: Li Aubrey <aubrey.li@linux.intel.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Matthew Garrett <mjg59@srcf.ucam.org> Link: http://lkml.kernel.org/r/20140404064120.GB11877@gmail.com Signed-off-by: Ingo Molnar <mingo@kernel.org>
2014-04-04 14:41:26 +08:00
* 6) If no EFI runtime service, call the BIOS to do a reboot
x86: Reorder reboot method preferences We have a never ending stream of 'reboot quirks' for new boxes that will not reboot properly under Linux (they will hang on reboot). The reason is widespread 'Windows compatible' assumption of modern x86 hardware, which expects the following reboot sequence: - hitting the ACPI reboot vector (if available) - trying the keyboard controller - hitting the ACPI reboot vector again - then giving the keyboard controller one last go This sequence expectation gets more and more embedded in modern hardware, which often lacks a keyboard controller and may even lock up if the legacy io ports are hit - and which hardware is often not tested with Linux during development. The end result is that reboot works under Windows-alike OSs but not under Linux. Rework our reboot process to meet this hardware externality a little better and match this assumption of newer x86 hardware. In addition to the ACPI,kbd,ACPI,kbd sequence we'll still fall through to attempting a legacy triple fault if nothing else works - and keep trying that and the kbd reset. Signed-off-by: Matthew Garrett <mjg@redhat.com> [ this commit will also save special casing Oaktrail boards ] Acked-by: Alan Cox <alan@linux.intel.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Leann Ogasawara <leann.ogasawara@canonical.com> Cc: Dave Jones <davej@redhat.com> Cc: Len Brown <len.brown@intel.com> LKML-Reference: <1301939705-2404-1-git-send-email-mjg@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2011-04-05 01:55:05 +08:00
*
x86: Remove the PCI reboot method from the default chain Steve reported a reboot hang and bisected it back to this commit: a4f1987e4c54 x86, reboot: Add EFI and CF9 reboot methods into the default list He heroically tested all reboot methods and found the following: reboot=t # triple fault ok reboot=k # keyboard ctrl FAIL reboot=b # BIOS ok reboot=a # ACPI FAIL reboot=e # EFI FAIL [system has no EFI] reboot=p # PCI 0xcf9 FAIL And I think it's pretty obvious that we should only try PCI 0xcf9 as a last resort - if at all. The other observation is that (on this box) we should never try the PCI reboot method, but close with either the 'triple fault' or the 'BIOS' (terminal!) reboot methods. Thirdly, CF9_COND is a total misnomer - it should be something like CF9_SAFE or CF9_CAREFUL, and 'CF9' should be 'CF9_FORCE' ... So this patch fixes the worst problems: - it orders the actual reboot logic to follow the reboot ordering pattern - it was in a pretty random order before for no good reason. - it fixes the CF9 misnomers and uses BOOT_CF9_FORCE and BOOT_CF9_SAFE flags to make the code more obvious. - it tries the BIOS reboot method before the PCI reboot method. (Since 'BIOS' is a terminal reboot method resulting in a hang if it does not work, this is essentially equivalent to removing the PCI reboot method from the default reboot chain.) - just for the miraculous possibility of terminal (resulting in hang) reboot methods of triple fault or BIOS returning without having done their job, there's an ordering between them as well. Reported-and-bisected-and-tested-by: Steven Rostedt <rostedt@goodmis.org> Cc: Li Aubrey <aubrey.li@linux.intel.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Matthew Garrett <mjg59@srcf.ucam.org> Link: http://lkml.kernel.org/r/20140404064120.GB11877@gmail.com Signed-off-by: Ingo Molnar <mingo@kernel.org>
2014-04-04 14:41:26 +08:00
* We default to following the same pattern. We also have
* two other reboot methods: 'triple fault' and 'PCI', which
* can be triggered via the reboot= kernel boot option or
* via quirks.
*
* This means that this function can never return, it can misbehave
* by not rebooting properly and hanging.
x86: Reorder reboot method preferences We have a never ending stream of 'reboot quirks' for new boxes that will not reboot properly under Linux (they will hang on reboot). The reason is widespread 'Windows compatible' assumption of modern x86 hardware, which expects the following reboot sequence: - hitting the ACPI reboot vector (if available) - trying the keyboard controller - hitting the ACPI reboot vector again - then giving the keyboard controller one last go This sequence expectation gets more and more embedded in modern hardware, which often lacks a keyboard controller and may even lock up if the legacy io ports are hit - and which hardware is often not tested with Linux during development. The end result is that reboot works under Windows-alike OSs but not under Linux. Rework our reboot process to meet this hardware externality a little better and match this assumption of newer x86 hardware. In addition to the ACPI,kbd,ACPI,kbd sequence we'll still fall through to attempting a legacy triple fault if nothing else works - and keep trying that and the kbd reset. Signed-off-by: Matthew Garrett <mjg@redhat.com> [ this commit will also save special casing Oaktrail boards ] Acked-by: Alan Cox <alan@linux.intel.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Leann Ogasawara <leann.ogasawara@canonical.com> Cc: Dave Jones <davej@redhat.com> Cc: Len Brown <len.brown@intel.com> LKML-Reference: <1301939705-2404-1-git-send-email-mjg@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2011-04-05 01:55:05 +08:00
*/
static void native_machine_emergency_restart(void)
{
int i;
x86: Reorder reboot method preferences We have a never ending stream of 'reboot quirks' for new boxes that will not reboot properly under Linux (they will hang on reboot). The reason is widespread 'Windows compatible' assumption of modern x86 hardware, which expects the following reboot sequence: - hitting the ACPI reboot vector (if available) - trying the keyboard controller - hitting the ACPI reboot vector again - then giving the keyboard controller one last go This sequence expectation gets more and more embedded in modern hardware, which often lacks a keyboard controller and may even lock up if the legacy io ports are hit - and which hardware is often not tested with Linux during development. The end result is that reboot works under Windows-alike OSs but not under Linux. Rework our reboot process to meet this hardware externality a little better and match this assumption of newer x86 hardware. In addition to the ACPI,kbd,ACPI,kbd sequence we'll still fall through to attempting a legacy triple fault if nothing else works - and keep trying that and the kbd reset. Signed-off-by: Matthew Garrett <mjg@redhat.com> [ this commit will also save special casing Oaktrail boards ] Acked-by: Alan Cox <alan@linux.intel.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Leann Ogasawara <leann.ogasawara@canonical.com> Cc: Dave Jones <davej@redhat.com> Cc: Len Brown <len.brown@intel.com> LKML-Reference: <1301939705-2404-1-git-send-email-mjg@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2011-04-05 01:55:05 +08:00
int attempt = 0;
int orig_reboot_type = reboot_type;
unsigned short mode;
if (reboot_emergency)
emergency_vmx_disable_all();
tboot_shutdown(TB_SHUTDOWN_REBOOT);
/* Tell the BIOS if we want cold or warm reboot */
mode = reboot_mode == REBOOT_WARM ? 0x1234 : 0;
*((unsigned short *)__va(0x472)) = mode;
for (;;) {
/* Could also try the reset bit in the Hammer NB */
switch (reboot_type) {
x86: Remove the PCI reboot method from the default chain Steve reported a reboot hang and bisected it back to this commit: a4f1987e4c54 x86, reboot: Add EFI and CF9 reboot methods into the default list He heroically tested all reboot methods and found the following: reboot=t # triple fault ok reboot=k # keyboard ctrl FAIL reboot=b # BIOS ok reboot=a # ACPI FAIL reboot=e # EFI FAIL [system has no EFI] reboot=p # PCI 0xcf9 FAIL And I think it's pretty obvious that we should only try PCI 0xcf9 as a last resort - if at all. The other observation is that (on this box) we should never try the PCI reboot method, but close with either the 'triple fault' or the 'BIOS' (terminal!) reboot methods. Thirdly, CF9_COND is a total misnomer - it should be something like CF9_SAFE or CF9_CAREFUL, and 'CF9' should be 'CF9_FORCE' ... So this patch fixes the worst problems: - it orders the actual reboot logic to follow the reboot ordering pattern - it was in a pretty random order before for no good reason. - it fixes the CF9 misnomers and uses BOOT_CF9_FORCE and BOOT_CF9_SAFE flags to make the code more obvious. - it tries the BIOS reboot method before the PCI reboot method. (Since 'BIOS' is a terminal reboot method resulting in a hang if it does not work, this is essentially equivalent to removing the PCI reboot method from the default reboot chain.) - just for the miraculous possibility of terminal (resulting in hang) reboot methods of triple fault or BIOS returning without having done their job, there's an ordering between them as well. Reported-and-bisected-and-tested-by: Steven Rostedt <rostedt@goodmis.org> Cc: Li Aubrey <aubrey.li@linux.intel.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Matthew Garrett <mjg59@srcf.ucam.org> Link: http://lkml.kernel.org/r/20140404064120.GB11877@gmail.com Signed-off-by: Ingo Molnar <mingo@kernel.org>
2014-04-04 14:41:26 +08:00
case BOOT_ACPI:
acpi_reboot();
reboot_type = BOOT_KBD;
break;
case BOOT_KBD:
mach_reboot_fixups(); /* For board specific fixups */
for (i = 0; i < 10; i++) {
kb_wait();
udelay(50);
outb(0xfe, 0x64); /* Pulse reset low */
udelay(50);
}
x86: Reorder reboot method preferences We have a never ending stream of 'reboot quirks' for new boxes that will not reboot properly under Linux (they will hang on reboot). The reason is widespread 'Windows compatible' assumption of modern x86 hardware, which expects the following reboot sequence: - hitting the ACPI reboot vector (if available) - trying the keyboard controller - hitting the ACPI reboot vector again - then giving the keyboard controller one last go This sequence expectation gets more and more embedded in modern hardware, which often lacks a keyboard controller and may even lock up if the legacy io ports are hit - and which hardware is often not tested with Linux during development. The end result is that reboot works under Windows-alike OSs but not under Linux. Rework our reboot process to meet this hardware externality a little better and match this assumption of newer x86 hardware. In addition to the ACPI,kbd,ACPI,kbd sequence we'll still fall through to attempting a legacy triple fault if nothing else works - and keep trying that and the kbd reset. Signed-off-by: Matthew Garrett <mjg@redhat.com> [ this commit will also save special casing Oaktrail boards ] Acked-by: Alan Cox <alan@linux.intel.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Leann Ogasawara <leann.ogasawara@canonical.com> Cc: Dave Jones <davej@redhat.com> Cc: Len Brown <len.brown@intel.com> LKML-Reference: <1301939705-2404-1-git-send-email-mjg@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2011-04-05 01:55:05 +08:00
if (attempt == 0 && orig_reboot_type == BOOT_ACPI) {
attempt = 1;
reboot_type = BOOT_ACPI;
} else {
x86, reboot: Add EFI and CF9 reboot methods into the default list Reboot is the last service linux OS provides to the end user. We are supposed to make this function more robust than today. This patch adds all of the known reboot methods into the default attempt list. The machines requiring reboot=efi or reboot=p or reboot=bios get a chance to reboot automatically now. If there is a new reboot method emerged, we are supposed to add it to the default list as well, instead of adding the endless dmidecode entry. If one method required is in the default list in this patch but the machine reboot still hangs, that means some methods ahead of the required method cause the system hangs, then reboot the machine by passing reboot= arguments and submit the reboot dmidecode table quirk. We are supposed to remove the reboot dmidecode table from the kernel, but to be safe, we keep it. This patch prevents us from adding more. If you happened to have a machine listed in the reboot dmidecode table and this patch makes reboot work on your machine, please submit a patch to remove the quirk. The default reboot order with this patch is now: ACPI > KBD > ACPI > KBD > EFI > CF9_COND > BIOS Because BIOS and TRIPLE are mutually exclusive (either will either work or hang the machine) that method is not included. [ hpa: as with any changes to the reboot order, this patch will have to be monitored carefully for regressions. ] Signed-off-by: Aubrey Li <aubrey.li@intel.com> Acked-by: Matthew Garrett <mjg59@srcf.ucam.org> Link: http://lkml.kernel.org/r/53130A46.1010801@linux.intel.com Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
2014-03-02 18:39:02 +08:00
reboot_type = BOOT_EFI;
x86: Reorder reboot method preferences We have a never ending stream of 'reboot quirks' for new boxes that will not reboot properly under Linux (they will hang on reboot). The reason is widespread 'Windows compatible' assumption of modern x86 hardware, which expects the following reboot sequence: - hitting the ACPI reboot vector (if available) - trying the keyboard controller - hitting the ACPI reboot vector again - then giving the keyboard controller one last go This sequence expectation gets more and more embedded in modern hardware, which often lacks a keyboard controller and may even lock up if the legacy io ports are hit - and which hardware is often not tested with Linux during development. The end result is that reboot works under Windows-alike OSs but not under Linux. Rework our reboot process to meet this hardware externality a little better and match this assumption of newer x86 hardware. In addition to the ACPI,kbd,ACPI,kbd sequence we'll still fall through to attempting a legacy triple fault if nothing else works - and keep trying that and the kbd reset. Signed-off-by: Matthew Garrett <mjg@redhat.com> [ this commit will also save special casing Oaktrail boards ] Acked-by: Alan Cox <alan@linux.intel.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Leann Ogasawara <leann.ogasawara@canonical.com> Cc: Dave Jones <davej@redhat.com> Cc: Len Brown <len.brown@intel.com> LKML-Reference: <1301939705-2404-1-git-send-email-mjg@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2011-04-05 01:55:05 +08:00
}
break;
case BOOT_EFI:
efi_reboot(reboot_mode, NULL);
x86: Remove the PCI reboot method from the default chain Steve reported a reboot hang and bisected it back to this commit: a4f1987e4c54 x86, reboot: Add EFI and CF9 reboot methods into the default list He heroically tested all reboot methods and found the following: reboot=t # triple fault ok reboot=k # keyboard ctrl FAIL reboot=b # BIOS ok reboot=a # ACPI FAIL reboot=e # EFI FAIL [system has no EFI] reboot=p # PCI 0xcf9 FAIL And I think it's pretty obvious that we should only try PCI 0xcf9 as a last resort - if at all. The other observation is that (on this box) we should never try the PCI reboot method, but close with either the 'triple fault' or the 'BIOS' (terminal!) reboot methods. Thirdly, CF9_COND is a total misnomer - it should be something like CF9_SAFE or CF9_CAREFUL, and 'CF9' should be 'CF9_FORCE' ... So this patch fixes the worst problems: - it orders the actual reboot logic to follow the reboot ordering pattern - it was in a pretty random order before for no good reason. - it fixes the CF9 misnomers and uses BOOT_CF9_FORCE and BOOT_CF9_SAFE flags to make the code more obvious. - it tries the BIOS reboot method before the PCI reboot method. (Since 'BIOS' is a terminal reboot method resulting in a hang if it does not work, this is essentially equivalent to removing the PCI reboot method from the default reboot chain.) - just for the miraculous possibility of terminal (resulting in hang) reboot methods of triple fault or BIOS returning without having done their job, there's an ordering between them as well. Reported-and-bisected-and-tested-by: Steven Rostedt <rostedt@goodmis.org> Cc: Li Aubrey <aubrey.li@linux.intel.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Matthew Garrett <mjg59@srcf.ucam.org> Link: http://lkml.kernel.org/r/20140404064120.GB11877@gmail.com Signed-off-by: Ingo Molnar <mingo@kernel.org>
2014-04-04 14:41:26 +08:00
reboot_type = BOOT_BIOS;
break;
case BOOT_BIOS:
machine_real_restart(MRR_BIOS);
/* We're probably dead after this, but... */
reboot_type = BOOT_CF9_SAFE;
break;
x86: Remove the PCI reboot method from the default chain Steve reported a reboot hang and bisected it back to this commit: a4f1987e4c54 x86, reboot: Add EFI and CF9 reboot methods into the default list He heroically tested all reboot methods and found the following: reboot=t # triple fault ok reboot=k # keyboard ctrl FAIL reboot=b # BIOS ok reboot=a # ACPI FAIL reboot=e # EFI FAIL [system has no EFI] reboot=p # PCI 0xcf9 FAIL And I think it's pretty obvious that we should only try PCI 0xcf9 as a last resort - if at all. The other observation is that (on this box) we should never try the PCI reboot method, but close with either the 'triple fault' or the 'BIOS' (terminal!) reboot methods. Thirdly, CF9_COND is a total misnomer - it should be something like CF9_SAFE or CF9_CAREFUL, and 'CF9' should be 'CF9_FORCE' ... So this patch fixes the worst problems: - it orders the actual reboot logic to follow the reboot ordering pattern - it was in a pretty random order before for no good reason. - it fixes the CF9 misnomers and uses BOOT_CF9_FORCE and BOOT_CF9_SAFE flags to make the code more obvious. - it tries the BIOS reboot method before the PCI reboot method. (Since 'BIOS' is a terminal reboot method resulting in a hang if it does not work, this is essentially equivalent to removing the PCI reboot method from the default reboot chain.) - just for the miraculous possibility of terminal (resulting in hang) reboot methods of triple fault or BIOS returning without having done their job, there's an ordering between them as well. Reported-and-bisected-and-tested-by: Steven Rostedt <rostedt@goodmis.org> Cc: Li Aubrey <aubrey.li@linux.intel.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Matthew Garrett <mjg59@srcf.ucam.org> Link: http://lkml.kernel.org/r/20140404064120.GB11877@gmail.com Signed-off-by: Ingo Molnar <mingo@kernel.org>
2014-04-04 14:41:26 +08:00
case BOOT_CF9_FORCE:
port_cf9_safe = true;
/* Fall through */
x86: Remove the PCI reboot method from the default chain Steve reported a reboot hang and bisected it back to this commit: a4f1987e4c54 x86, reboot: Add EFI and CF9 reboot methods into the default list He heroically tested all reboot methods and found the following: reboot=t # triple fault ok reboot=k # keyboard ctrl FAIL reboot=b # BIOS ok reboot=a # ACPI FAIL reboot=e # EFI FAIL [system has no EFI] reboot=p # PCI 0xcf9 FAIL And I think it's pretty obvious that we should only try PCI 0xcf9 as a last resort - if at all. The other observation is that (on this box) we should never try the PCI reboot method, but close with either the 'triple fault' or the 'BIOS' (terminal!) reboot methods. Thirdly, CF9_COND is a total misnomer - it should be something like CF9_SAFE or CF9_CAREFUL, and 'CF9' should be 'CF9_FORCE' ... So this patch fixes the worst problems: - it orders the actual reboot logic to follow the reboot ordering pattern - it was in a pretty random order before for no good reason. - it fixes the CF9 misnomers and uses BOOT_CF9_FORCE and BOOT_CF9_SAFE flags to make the code more obvious. - it tries the BIOS reboot method before the PCI reboot method. (Since 'BIOS' is a terminal reboot method resulting in a hang if it does not work, this is essentially equivalent to removing the PCI reboot method from the default reboot chain.) - just for the miraculous possibility of terminal (resulting in hang) reboot methods of triple fault or BIOS returning without having done their job, there's an ordering between them as well. Reported-and-bisected-and-tested-by: Steven Rostedt <rostedt@goodmis.org> Cc: Li Aubrey <aubrey.li@linux.intel.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Matthew Garrett <mjg59@srcf.ucam.org> Link: http://lkml.kernel.org/r/20140404064120.GB11877@gmail.com Signed-off-by: Ingo Molnar <mingo@kernel.org>
2014-04-04 14:41:26 +08:00
case BOOT_CF9_SAFE:
if (port_cf9_safe) {
x86: Remove the PCI reboot method from the default chain Steve reported a reboot hang and bisected it back to this commit: a4f1987e4c54 x86, reboot: Add EFI and CF9 reboot methods into the default list He heroically tested all reboot methods and found the following: reboot=t # triple fault ok reboot=k # keyboard ctrl FAIL reboot=b # BIOS ok reboot=a # ACPI FAIL reboot=e # EFI FAIL [system has no EFI] reboot=p # PCI 0xcf9 FAIL And I think it's pretty obvious that we should only try PCI 0xcf9 as a last resort - if at all. The other observation is that (on this box) we should never try the PCI reboot method, but close with either the 'triple fault' or the 'BIOS' (terminal!) reboot methods. Thirdly, CF9_COND is a total misnomer - it should be something like CF9_SAFE or CF9_CAREFUL, and 'CF9' should be 'CF9_FORCE' ... So this patch fixes the worst problems: - it orders the actual reboot logic to follow the reboot ordering pattern - it was in a pretty random order before for no good reason. - it fixes the CF9 misnomers and uses BOOT_CF9_FORCE and BOOT_CF9_SAFE flags to make the code more obvious. - it tries the BIOS reboot method before the PCI reboot method. (Since 'BIOS' is a terminal reboot method resulting in a hang if it does not work, this is essentially equivalent to removing the PCI reboot method from the default reboot chain.) - just for the miraculous possibility of terminal (resulting in hang) reboot methods of triple fault or BIOS returning without having done their job, there's an ordering between them as well. Reported-and-bisected-and-tested-by: Steven Rostedt <rostedt@goodmis.org> Cc: Li Aubrey <aubrey.li@linux.intel.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Matthew Garrett <mjg59@srcf.ucam.org> Link: http://lkml.kernel.org/r/20140404064120.GB11877@gmail.com Signed-off-by: Ingo Molnar <mingo@kernel.org>
2014-04-04 14:41:26 +08:00
u8 reboot_code = reboot_mode == REBOOT_WARM ? 0x06 : 0x0E;
u8 cf9 = inb(0xcf9) & ~reboot_code;
outb(cf9|2, 0xcf9); /* Request hard reset */
udelay(50);
/* Actually do the reset */
outb(cf9|reboot_code, 0xcf9);
udelay(50);
}
x86: Remove the PCI reboot method from the default chain Steve reported a reboot hang and bisected it back to this commit: a4f1987e4c54 x86, reboot: Add EFI and CF9 reboot methods into the default list He heroically tested all reboot methods and found the following: reboot=t # triple fault ok reboot=k # keyboard ctrl FAIL reboot=b # BIOS ok reboot=a # ACPI FAIL reboot=e # EFI FAIL [system has no EFI] reboot=p # PCI 0xcf9 FAIL And I think it's pretty obvious that we should only try PCI 0xcf9 as a last resort - if at all. The other observation is that (on this box) we should never try the PCI reboot method, but close with either the 'triple fault' or the 'BIOS' (terminal!) reboot methods. Thirdly, CF9_COND is a total misnomer - it should be something like CF9_SAFE or CF9_CAREFUL, and 'CF9' should be 'CF9_FORCE' ... So this patch fixes the worst problems: - it orders the actual reboot logic to follow the reboot ordering pattern - it was in a pretty random order before for no good reason. - it fixes the CF9 misnomers and uses BOOT_CF9_FORCE and BOOT_CF9_SAFE flags to make the code more obvious. - it tries the BIOS reboot method before the PCI reboot method. (Since 'BIOS' is a terminal reboot method resulting in a hang if it does not work, this is essentially equivalent to removing the PCI reboot method from the default reboot chain.) - just for the miraculous possibility of terminal (resulting in hang) reboot methods of triple fault or BIOS returning without having done their job, there's an ordering between them as well. Reported-and-bisected-and-tested-by: Steven Rostedt <rostedt@goodmis.org> Cc: Li Aubrey <aubrey.li@linux.intel.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Matthew Garrett <mjg59@srcf.ucam.org> Link: http://lkml.kernel.org/r/20140404064120.GB11877@gmail.com Signed-off-by: Ingo Molnar <mingo@kernel.org>
2014-04-04 14:41:26 +08:00
reboot_type = BOOT_TRIPLE;
break;
case BOOT_TRIPLE:
load_idt(&no_idt);
__asm__ __volatile__("int3");
/* We're probably dead after this, but... */
reboot_type = BOOT_KBD;
break;
}
}
}
void native_machine_shutdown(void)
{
/* Stop the cpus and apics */
#ifdef CONFIG_X86_IO_APIC
/*
* Disabling IO APIC before local APIC is a workaround for
* erratum AVR31 in "Intel Atom Processor C2000 Product Family
* Specification Update". In this situation, interrupts that target
* a Logical Processor whose Local APIC is either in the process of
* being hardware disabled or software disabled are neither delivered
* nor discarded. When this erratum occurs, the processor may hang.
*
* Even without the erratum, it still makes sense to quiet IO APIC
* before disabling Local APIC.
*/
disable_IO_APIC();
#endif
#ifdef CONFIG_SMP
/*
* Stop all of the others. Also disable the local irq to
* not receive the per-cpu timer interrupt which may trigger
* scheduler's load balance.
*/
x86/reboot: Fix a warning message triggered by stop_other_cpus() When rebooting our 24 CPU Westmere servers with 3.4-rc6, we always see this warning msg: Restarting system. machine restart ------------[ cut here ]------------ WARNING: at arch/x86/kernel/smp.c:125 native_smp_send_reschedule+0x74/0xa7() Hardware name: X8DTN Modules linked in: igb [last unloaded: scsi_wait_scan] Pid: 1, comm: systemd-shutdow Not tainted 3.4.0-rc6+ #22 Call Trace: <IRQ> [<ffffffff8102a41f>] warn_slowpath_common+0x7e/0x96 [<ffffffff8102a44c>] warn_slowpath_null+0x15/0x17 [<ffffffff81018cf7>] native_smp_send_reschedule+0x74/0xa7 [<ffffffff810561c1>] trigger_load_balance+0x279/0x2a6 [<ffffffff81050112>] scheduler_tick+0xe0/0xe9 [<ffffffff81036768>] update_process_times+0x60/0x70 [<ffffffff81062f2f>] tick_sched_timer+0x68/0x92 [<ffffffff81046e33>] __run_hrtimer+0xb3/0x13c [<ffffffff81062ec7>] ? tick_nohz_handler+0xd0/0xd0 [<ffffffff810474f2>] hrtimer_interrupt+0xdb/0x198 [<ffffffff81019a35>] smp_apic_timer_interrupt+0x81/0x94 [<ffffffff81655187>] apic_timer_interrupt+0x67/0x70 <EOI> [<ffffffff8101a3c4>] ? default_send_IPI_mask_allbutself_phys+0xb4/0xc4 [<ffffffff8101c680>] physflat_send_IPI_allbutself+0x12/0x14 [<ffffffff81018db4>] native_nmi_stop_other_cpus+0x8a/0xd6 [<ffffffff810188ba>] native_machine_shutdown+0x50/0x67 [<ffffffff81018926>] machine_shutdown+0xa/0xc [<ffffffff8101897e>] native_machine_restart+0x20/0x32 [<ffffffff810189b0>] machine_restart+0xa/0xc [<ffffffff8103b196>] kernel_restart+0x47/0x4c [<ffffffff8103b2e6>] sys_reboot+0x13e/0x17c [<ffffffff8164e436>] ? _raw_spin_unlock_bh+0x10/0x12 [<ffffffff810fcac9>] ? bdi_queue_work+0xcf/0xd8 [<ffffffff810fe82f>] ? __bdi_start_writeback+0xae/0xb7 [<ffffffff810e0d64>] ? iterate_supers+0xa3/0xb7 [<ffffffff816547a2>] system_call_fastpath+0x16/0x1b ---[ end trace 320af5cb1cb60c5b ]--- The root cause seems to be the default_send_IPI_mask_allbutself_phys() takes quite some time (I measured it could be several ms) to complete sending NMIs to all the other 23 CPUs, and for HZ=250/1000 system, the time is long enough for a timer interrupt to happen, which will in turn trigger to kick load balance to a stopped CPU and cause this warning in native_smp_send_reschedule(). So disabling the local irq before stop_other_cpu() can fix this problem (tested 25 times reboot ok), and it is fine as there should be nobody caring the timer interrupt in such reboot stage. The latest 3.4 kernel slightly changes this behavior by sending REBOOT_VECTOR first and only send NMI_VECTOR if the REBOOT_VCTOR fails, and this patch is still needed to prevent the problem. Signed-off-by: Feng Tang <feng.tang@intel.com> Acked-by: Don Zickus <dzickus@redhat.com> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20120530231541.4c13433a@feng-i7 Signed-off-by: Ingo Molnar <mingo@kernel.org>
2012-05-30 23:15:41 +08:00
local_irq_disable();
stop_other_cpus();
#endif
lapic_shutdown();
#ifdef CONFIG_HPET_TIMER
hpet_disable();
#endif
#ifdef CONFIG_X86_64
x86_platform.iommu_shutdown();
#endif
}
static void __machine_emergency_restart(int emergency)
{
reboot_emergency = emergency;
machine_ops.emergency_restart();
}
static void native_machine_restart(char *__unused)
{
pr_notice("machine restart\n");
if (!reboot_force)
machine_shutdown();
__machine_emergency_restart(0);
}
static void native_machine_halt(void)
{
/* Stop other cpus and apics */
machine_shutdown();
tboot_shutdown(TB_SHUTDOWN_HALT);
stop_this_cpu(NULL);
}
static void native_machine_power_off(void)
{
[PATCH] i386/x86-64: Don't IPI to offline cpus on shutdown So why are we calling smp_send_stop from machine_halt? We don't. Looking more closely at the bug report the problem here is that halt -p is called which triggers not a halt but an attempt to power off. machine_power_off calls machine_shutdown which calls smp_send_stop. If pm_power_off is set we should never make it out machine_power_off to the call of do_exit. So pm_power_off must not be set in this case. When pm_power_off is not set we expect machine_power_off to devolve into machine_halt. So how do we fix this? Playing too much with smp_send_stop is dangerous because it must also be safe to be called from panic. It looks like the obviously correct fix is to only call machine_shutdown when pm_power_off is defined. Doing that will make Andi's assumption about not scheduling true and generally simplify what must be supported. This turns machine_power_off into a noop like machine_halt when pm_power_off is not defined. If the expected behavior is that sys_reboot(LINUX_REBOOT_CMD_POWER_OFF) becomes sys_reboot(LINUX_REBOOT_CMD_HALT) if pm_power_off is NULL this is not quite a comprehensive fix as we pass a different parameter to the reboot notifier and we set system_state to a different value before calling device_shutdown(). Unfortunately any fix more comprehensive I can think of is not obviously correct. The core problem is that there is no architecture independent way to detect if machine_power will become a noop, without calling it. Signed-off-by: Andi Kleen <ak@suse.de> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-01-12 05:43:12 +08:00
if (pm_power_off) {
if (!reboot_force)
machine_shutdown();
pm_power_off();
[PATCH] i386/x86-64: Don't IPI to offline cpus on shutdown So why are we calling smp_send_stop from machine_halt? We don't. Looking more closely at the bug report the problem here is that halt -p is called which triggers not a halt but an attempt to power off. machine_power_off calls machine_shutdown which calls smp_send_stop. If pm_power_off is set we should never make it out machine_power_off to the call of do_exit. So pm_power_off must not be set in this case. When pm_power_off is not set we expect machine_power_off to devolve into machine_halt. So how do we fix this? Playing too much with smp_send_stop is dangerous because it must also be safe to be called from panic. It looks like the obviously correct fix is to only call machine_shutdown when pm_power_off is defined. Doing that will make Andi's assumption about not scheduling true and generally simplify what must be supported. This turns machine_power_off into a noop like machine_halt when pm_power_off is not defined. If the expected behavior is that sys_reboot(LINUX_REBOOT_CMD_POWER_OFF) becomes sys_reboot(LINUX_REBOOT_CMD_HALT) if pm_power_off is NULL this is not quite a comprehensive fix as we pass a different parameter to the reboot notifier and we set system_state to a different value before calling device_shutdown(). Unfortunately any fix more comprehensive I can think of is not obviously correct. The core problem is that there is no architecture independent way to detect if machine_power will become a noop, without calling it. Signed-off-by: Andi Kleen <ak@suse.de> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-01-12 05:43:12 +08:00
}
/* A fallback in case there is no PM info available */
tboot_shutdown(TB_SHUTDOWN_HALT);
}
struct machine_ops machine_ops = {
.power_off = native_machine_power_off,
.shutdown = native_machine_shutdown,
.emergency_restart = native_machine_emergency_restart,
.restart = native_machine_restart,
.halt = native_machine_halt,
2015-09-10 06:38:55 +08:00
#ifdef CONFIG_KEXEC_CORE
.crash_shutdown = native_machine_crash_shutdown,
#endif
};
void machine_power_off(void)
{
machine_ops.power_off();
}
void machine_shutdown(void)
{
machine_ops.shutdown();
}
void machine_emergency_restart(void)
{
__machine_emergency_restart(1);
}
void machine_restart(char *cmd)
{
machine_ops.restart(cmd);
}
void machine_halt(void)
{
machine_ops.halt();
}
2015-09-10 06:38:55 +08:00
#ifdef CONFIG_KEXEC_CORE
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;
panic, x86: Allow CPUs to save registers even if looping in NMI context Currently, kdump_nmi_shootdown_cpus(), a subroutine of crash_kexec(), sends an NMI IPI to CPUs which haven't called panic() to stop them, save their register information and do some cleanups for crash dumping. However, if such a CPU is infinitely looping in NMI context, we fail to save its register information into the crash dump. For example, this can happen when unknown NMIs are broadcast to all CPUs as follows: CPU 0 CPU 1 =========================== ========================== receive an unknown NMI unknown_nmi_error() panic() receive an unknown NMI spin_trylock(&panic_lock) unknown_nmi_error() crash_kexec() panic() spin_trylock(&panic_lock) panic_smp_self_stop() infinite loop kdump_nmi_shootdown_cpus() issue NMI IPI -----------> blocked until IRET infinite loop... Here, since CPU 1 is in NMI context, the second NMI from CPU 0 is blocked until CPU 1 executes IRET. However, CPU 1 never executes IRET, so the NMI is not handled and the callback function to save registers is never called. In practice, this can happen on some servers which broadcast NMIs to all CPUs when the NMI button is pushed. To save registers in this case, we need to: a) Return from NMI handler instead of looping infinitely or b) Call the callback function directly from the infinite loop Inherently, a) is risky because NMI is also used to prevent corrupted data from being propagated to devices. So, we chose b). This patch does the following: 1. Move the infinite looping of CPUs which haven't called panic() in NMI context (actually done by panic_smp_self_stop()) outside of panic() to enable us to refer pt_regs. Please note that panic_smp_self_stop() is still used for normal context. 2. Call a callback of kdump_nmi_shootdown_cpus() directly to save registers and do some cleanups after setting waiting_for_crash_ipi which is used for counting down the number of CPUs which handled the callback Signed-off-by: Hidehiro Kawai <hidehiro.kawai.ez@hitachi.com> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Aaron Tomlin <atomlin@redhat.com> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Andy Lutomirski <luto@kernel.org> Cc: Baoquan He <bhe@redhat.com> Cc: Chris Metcalf <cmetcalf@ezchip.com> Cc: Dave Young <dyoung@redhat.com> Cc: David Hildenbrand <dahi@linux.vnet.ibm.com> Cc: Don Zickus <dzickus@redhat.com> Cc: Eric Biederman <ebiederm@xmission.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Gobinda Charan Maji <gobinda.cemk07@gmail.com> Cc: HATAYAMA Daisuke <d.hatayama@jp.fujitsu.com> Cc: Hidehiro Kawai <hidehiro.kawai.ez@hitachi.com> Cc: "H. Peter Anvin" <hpa@zytor.com> Cc: Ingo Molnar <mingo@kernel.org> Cc: Javi Merino <javi.merino@arm.com> Cc: Jiang Liu <jiang.liu@linux.intel.com> Cc: Jonathan Corbet <corbet@lwn.net> Cc: kexec@lists.infradead.org Cc: linux-doc@vger.kernel.org Cc: lkml <linux-kernel@vger.kernel.org> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Michal Nazarewicz <mina86@mina86.com> Cc: Nicolas Iooss <nicolas.iooss_linux@m4x.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Prarit Bhargava <prarit@redhat.com> Cc: Rasmus Villemoes <linux@rasmusvillemoes.dk> Cc: Seth Jennings <sjenning@redhat.com> Cc: Stefan Lippers-Hollmann <s.l-h@gmx.de> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Ulrich Obergfell <uobergfe@redhat.com> Cc: Vitaly Kuznetsov <vkuznets@redhat.com> Cc: Vivek Goyal <vgoyal@redhat.com> Cc: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com> Link: http://lkml.kernel.org/r/20151210014628.25437.75256.stgit@softrs [ Cleanup comments, fixup formatting. ] Signed-off-by: Borislav Petkov <bp@suse.de> Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2015-12-14 18:19:10 +08:00
static int crash_ipi_issued;
static int crash_nmi_callback(unsigned int val, struct pt_regs *regs)
{
int cpu;
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 NMI_HANDLED;
local_irq_disable();
shootdown_callback(cpu, regs);
atomic_dec(&waiting_for_crash_ipi);
/* Assume hlt works */
halt();
for (;;)
cpu_relax();
return NMI_HANDLED;
}
static void smp_send_nmi_allbutself(void)
{
apic->send_IPI_allbutself(NMI_VECTOR);
}
/*
* 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_nmi_handler(NMI_LOCAL, crash_nmi_callback,
NMI_FLAG_FIRST, "crash"))
return; /* Return what? */
/*
* Ensure the new callback function is set before sending
* out the NMI
*/
wmb();
smp_send_nmi_allbutself();
panic, x86: Allow CPUs to save registers even if looping in NMI context Currently, kdump_nmi_shootdown_cpus(), a subroutine of crash_kexec(), sends an NMI IPI to CPUs which haven't called panic() to stop them, save their register information and do some cleanups for crash dumping. However, if such a CPU is infinitely looping in NMI context, we fail to save its register information into the crash dump. For example, this can happen when unknown NMIs are broadcast to all CPUs as follows: CPU 0 CPU 1 =========================== ========================== receive an unknown NMI unknown_nmi_error() panic() receive an unknown NMI spin_trylock(&panic_lock) unknown_nmi_error() crash_kexec() panic() spin_trylock(&panic_lock) panic_smp_self_stop() infinite loop kdump_nmi_shootdown_cpus() issue NMI IPI -----------> blocked until IRET infinite loop... Here, since CPU 1 is in NMI context, the second NMI from CPU 0 is blocked until CPU 1 executes IRET. However, CPU 1 never executes IRET, so the NMI is not handled and the callback function to save registers is never called. In practice, this can happen on some servers which broadcast NMIs to all CPUs when the NMI button is pushed. To save registers in this case, we need to: a) Return from NMI handler instead of looping infinitely or b) Call the callback function directly from the infinite loop Inherently, a) is risky because NMI is also used to prevent corrupted data from being propagated to devices. So, we chose b). This patch does the following: 1. Move the infinite looping of CPUs which haven't called panic() in NMI context (actually done by panic_smp_self_stop()) outside of panic() to enable us to refer pt_regs. Please note that panic_smp_self_stop() is still used for normal context. 2. Call a callback of kdump_nmi_shootdown_cpus() directly to save registers and do some cleanups after setting waiting_for_crash_ipi which is used for counting down the number of CPUs which handled the callback Signed-off-by: Hidehiro Kawai <hidehiro.kawai.ez@hitachi.com> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Aaron Tomlin <atomlin@redhat.com> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Andy Lutomirski <luto@kernel.org> Cc: Baoquan He <bhe@redhat.com> Cc: Chris Metcalf <cmetcalf@ezchip.com> Cc: Dave Young <dyoung@redhat.com> Cc: David Hildenbrand <dahi@linux.vnet.ibm.com> Cc: Don Zickus <dzickus@redhat.com> Cc: Eric Biederman <ebiederm@xmission.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Gobinda Charan Maji <gobinda.cemk07@gmail.com> Cc: HATAYAMA Daisuke <d.hatayama@jp.fujitsu.com> Cc: Hidehiro Kawai <hidehiro.kawai.ez@hitachi.com> Cc: "H. Peter Anvin" <hpa@zytor.com> Cc: Ingo Molnar <mingo@kernel.org> Cc: Javi Merino <javi.merino@arm.com> Cc: Jiang Liu <jiang.liu@linux.intel.com> Cc: Jonathan Corbet <corbet@lwn.net> Cc: kexec@lists.infradead.org Cc: linux-doc@vger.kernel.org Cc: lkml <linux-kernel@vger.kernel.org> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Michal Nazarewicz <mina86@mina86.com> Cc: Nicolas Iooss <nicolas.iooss_linux@m4x.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Prarit Bhargava <prarit@redhat.com> Cc: Rasmus Villemoes <linux@rasmusvillemoes.dk> Cc: Seth Jennings <sjenning@redhat.com> Cc: Stefan Lippers-Hollmann <s.l-h@gmx.de> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Ulrich Obergfell <uobergfe@redhat.com> Cc: Vitaly Kuznetsov <vkuznets@redhat.com> Cc: Vivek Goyal <vgoyal@redhat.com> Cc: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com> Link: http://lkml.kernel.org/r/20151210014628.25437.75256.stgit@softrs [ Cleanup comments, fixup formatting. ] Signed-off-by: Borislav Petkov <bp@suse.de> Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2015-12-14 18:19:10 +08:00
/* Kick CPUs looping in NMI context. */
WRITE_ONCE(crash_ipi_issued, 1);
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 */
}
panic, x86: Allow CPUs to save registers even if looping in NMI context Currently, kdump_nmi_shootdown_cpus(), a subroutine of crash_kexec(), sends an NMI IPI to CPUs which haven't called panic() to stop them, save their register information and do some cleanups for crash dumping. However, if such a CPU is infinitely looping in NMI context, we fail to save its register information into the crash dump. For example, this can happen when unknown NMIs are broadcast to all CPUs as follows: CPU 0 CPU 1 =========================== ========================== receive an unknown NMI unknown_nmi_error() panic() receive an unknown NMI spin_trylock(&panic_lock) unknown_nmi_error() crash_kexec() panic() spin_trylock(&panic_lock) panic_smp_self_stop() infinite loop kdump_nmi_shootdown_cpus() issue NMI IPI -----------> blocked until IRET infinite loop... Here, since CPU 1 is in NMI context, the second NMI from CPU 0 is blocked until CPU 1 executes IRET. However, CPU 1 never executes IRET, so the NMI is not handled and the callback function to save registers is never called. In practice, this can happen on some servers which broadcast NMIs to all CPUs when the NMI button is pushed. To save registers in this case, we need to: a) Return from NMI handler instead of looping infinitely or b) Call the callback function directly from the infinite loop Inherently, a) is risky because NMI is also used to prevent corrupted data from being propagated to devices. So, we chose b). This patch does the following: 1. Move the infinite looping of CPUs which haven't called panic() in NMI context (actually done by panic_smp_self_stop()) outside of panic() to enable us to refer pt_regs. Please note that panic_smp_self_stop() is still used for normal context. 2. Call a callback of kdump_nmi_shootdown_cpus() directly to save registers and do some cleanups after setting waiting_for_crash_ipi which is used for counting down the number of CPUs which handled the callback Signed-off-by: Hidehiro Kawai <hidehiro.kawai.ez@hitachi.com> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Aaron Tomlin <atomlin@redhat.com> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Andy Lutomirski <luto@kernel.org> Cc: Baoquan He <bhe@redhat.com> Cc: Chris Metcalf <cmetcalf@ezchip.com> Cc: Dave Young <dyoung@redhat.com> Cc: David Hildenbrand <dahi@linux.vnet.ibm.com> Cc: Don Zickus <dzickus@redhat.com> Cc: Eric Biederman <ebiederm@xmission.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Gobinda Charan Maji <gobinda.cemk07@gmail.com> Cc: HATAYAMA Daisuke <d.hatayama@jp.fujitsu.com> Cc: Hidehiro Kawai <hidehiro.kawai.ez@hitachi.com> Cc: "H. Peter Anvin" <hpa@zytor.com> Cc: Ingo Molnar <mingo@kernel.org> Cc: Javi Merino <javi.merino@arm.com> Cc: Jiang Liu <jiang.liu@linux.intel.com> Cc: Jonathan Corbet <corbet@lwn.net> Cc: kexec@lists.infradead.org Cc: linux-doc@vger.kernel.org Cc: lkml <linux-kernel@vger.kernel.org> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Michal Nazarewicz <mina86@mina86.com> Cc: Nicolas Iooss <nicolas.iooss_linux@m4x.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Prarit Bhargava <prarit@redhat.com> Cc: Rasmus Villemoes <linux@rasmusvillemoes.dk> Cc: Seth Jennings <sjenning@redhat.com> Cc: Stefan Lippers-Hollmann <s.l-h@gmx.de> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Ulrich Obergfell <uobergfe@redhat.com> Cc: Vitaly Kuznetsov <vkuznets@redhat.com> Cc: Vivek Goyal <vgoyal@redhat.com> Cc: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com> Link: http://lkml.kernel.org/r/20151210014628.25437.75256.stgit@softrs [ Cleanup comments, fixup formatting. ] Signed-off-by: Borislav Petkov <bp@suse.de> Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2015-12-14 18:19:10 +08:00
x86/nmi: Save regs in crash dump on external NMI Now, multiple CPUs can receive an external NMI simultaneously by specifying the "apic_extnmi=all" command line parameter. When we take a crash dump by using external NMI with this option, we fail to save registers into the crash dump. This happens as follows: CPU 0 CPU 1 ================================ ============================= receive an external NMI default_do_nmi() receive an external NMI spin_lock(&nmi_reason_lock) default_do_nmi() io_check_error() spin_lock(&nmi_reason_lock) panic() busy loop ... kdump_nmi_shootdown_cpus() issue NMI IPI -----------> blocked until IRET busy loop... Here, since CPU 1 is in NMI context, an additional NMI from CPU 0 remains unhandled until CPU 1 IRETs. However, CPU 1 will never execute IRET so the NMI is not handled and the callback function to save registers is never called. To solve this issue, we check if the IPI for crash dumping was issued while waiting for nmi_reason_lock to be released, and if so, call its callback function directly. If the IPI is not issued (e.g. kdump is disabled), the actual behavior doesn't change. Signed-off-by: Hidehiro Kawai <hidehiro.kawai.ez@hitachi.com> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Andy Lutomirski <luto@kernel.org> Cc: Baoquan He <bhe@redhat.com> Cc: Dave Young <dyoung@redhat.com> Cc: "Eric W. Biederman" <ebiederm@xmission.com> Cc: "H. Peter Anvin" <hpa@zytor.com> Cc: Ingo Molnar <mingo@kernel.org> Cc: Jiang Liu <jiang.liu@linux.intel.com> Cc: Jonathan Corbet <corbet@lwn.net> Cc: kexec@lists.infradead.org Cc: linux-doc@vger.kernel.org Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Stefan Lippers-Hollmann <s.l-h@gmx.de> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Vivek Goyal <vgoyal@redhat.com> Cc: x86-ml <x86@kernel.org> Link: http://lkml.kernel.org/r/20151210065245.4587.39316.stgit@softrs Signed-off-by: Borislav Petkov <bp@suse.de> Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2015-12-14 18:19:13 +08:00
/*
* Check if the crash dumping IPI got issued and if so, call its callback
* directly. This function is used when we have already been in NMI handler.
* It doesn't return.
*/
void run_crash_ipi_callback(struct pt_regs *regs)
{
if (crash_ipi_issued)
crash_nmi_callback(0, regs);
}
panic, x86: Allow CPUs to save registers even if looping in NMI context Currently, kdump_nmi_shootdown_cpus(), a subroutine of crash_kexec(), sends an NMI IPI to CPUs which haven't called panic() to stop them, save their register information and do some cleanups for crash dumping. However, if such a CPU is infinitely looping in NMI context, we fail to save its register information into the crash dump. For example, this can happen when unknown NMIs are broadcast to all CPUs as follows: CPU 0 CPU 1 =========================== ========================== receive an unknown NMI unknown_nmi_error() panic() receive an unknown NMI spin_trylock(&panic_lock) unknown_nmi_error() crash_kexec() panic() spin_trylock(&panic_lock) panic_smp_self_stop() infinite loop kdump_nmi_shootdown_cpus() issue NMI IPI -----------> blocked until IRET infinite loop... Here, since CPU 1 is in NMI context, the second NMI from CPU 0 is blocked until CPU 1 executes IRET. However, CPU 1 never executes IRET, so the NMI is not handled and the callback function to save registers is never called. In practice, this can happen on some servers which broadcast NMIs to all CPUs when the NMI button is pushed. To save registers in this case, we need to: a) Return from NMI handler instead of looping infinitely or b) Call the callback function directly from the infinite loop Inherently, a) is risky because NMI is also used to prevent corrupted data from being propagated to devices. So, we chose b). This patch does the following: 1. Move the infinite looping of CPUs which haven't called panic() in NMI context (actually done by panic_smp_self_stop()) outside of panic() to enable us to refer pt_regs. Please note that panic_smp_self_stop() is still used for normal context. 2. Call a callback of kdump_nmi_shootdown_cpus() directly to save registers and do some cleanups after setting waiting_for_crash_ipi which is used for counting down the number of CPUs which handled the callback Signed-off-by: Hidehiro Kawai <hidehiro.kawai.ez@hitachi.com> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Aaron Tomlin <atomlin@redhat.com> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Andy Lutomirski <luto@kernel.org> Cc: Baoquan He <bhe@redhat.com> Cc: Chris Metcalf <cmetcalf@ezchip.com> Cc: Dave Young <dyoung@redhat.com> Cc: David Hildenbrand <dahi@linux.vnet.ibm.com> Cc: Don Zickus <dzickus@redhat.com> Cc: Eric Biederman <ebiederm@xmission.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Gobinda Charan Maji <gobinda.cemk07@gmail.com> Cc: HATAYAMA Daisuke <d.hatayama@jp.fujitsu.com> Cc: Hidehiro Kawai <hidehiro.kawai.ez@hitachi.com> Cc: "H. Peter Anvin" <hpa@zytor.com> Cc: Ingo Molnar <mingo@kernel.org> Cc: Javi Merino <javi.merino@arm.com> Cc: Jiang Liu <jiang.liu@linux.intel.com> Cc: Jonathan Corbet <corbet@lwn.net> Cc: kexec@lists.infradead.org Cc: linux-doc@vger.kernel.org Cc: lkml <linux-kernel@vger.kernel.org> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Michal Nazarewicz <mina86@mina86.com> Cc: Nicolas Iooss <nicolas.iooss_linux@m4x.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Prarit Bhargava <prarit@redhat.com> Cc: Rasmus Villemoes <linux@rasmusvillemoes.dk> Cc: Seth Jennings <sjenning@redhat.com> Cc: Stefan Lippers-Hollmann <s.l-h@gmx.de> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Ulrich Obergfell <uobergfe@redhat.com> Cc: Vitaly Kuznetsov <vkuznets@redhat.com> Cc: Vivek Goyal <vgoyal@redhat.com> Cc: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com> Link: http://lkml.kernel.org/r/20151210014628.25437.75256.stgit@softrs [ Cleanup comments, fixup formatting. ] Signed-off-by: Borislav Petkov <bp@suse.de> Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2015-12-14 18:19:10 +08:00
/* Override the weak function in kernel/panic.c */
void nmi_panic_self_stop(struct pt_regs *regs)
{
while (1) {
x86/nmi: Save regs in crash dump on external NMI Now, multiple CPUs can receive an external NMI simultaneously by specifying the "apic_extnmi=all" command line parameter. When we take a crash dump by using external NMI with this option, we fail to save registers into the crash dump. This happens as follows: CPU 0 CPU 1 ================================ ============================= receive an external NMI default_do_nmi() receive an external NMI spin_lock(&nmi_reason_lock) default_do_nmi() io_check_error() spin_lock(&nmi_reason_lock) panic() busy loop ... kdump_nmi_shootdown_cpus() issue NMI IPI -----------> blocked until IRET busy loop... Here, since CPU 1 is in NMI context, an additional NMI from CPU 0 remains unhandled until CPU 1 IRETs. However, CPU 1 will never execute IRET so the NMI is not handled and the callback function to save registers is never called. To solve this issue, we check if the IPI for crash dumping was issued while waiting for nmi_reason_lock to be released, and if so, call its callback function directly. If the IPI is not issued (e.g. kdump is disabled), the actual behavior doesn't change. Signed-off-by: Hidehiro Kawai <hidehiro.kawai.ez@hitachi.com> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Andy Lutomirski <luto@kernel.org> Cc: Baoquan He <bhe@redhat.com> Cc: Dave Young <dyoung@redhat.com> Cc: "Eric W. Biederman" <ebiederm@xmission.com> Cc: "H. Peter Anvin" <hpa@zytor.com> Cc: Ingo Molnar <mingo@kernel.org> Cc: Jiang Liu <jiang.liu@linux.intel.com> Cc: Jonathan Corbet <corbet@lwn.net> Cc: kexec@lists.infradead.org Cc: linux-doc@vger.kernel.org Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Stefan Lippers-Hollmann <s.l-h@gmx.de> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Vivek Goyal <vgoyal@redhat.com> Cc: x86-ml <x86@kernel.org> Link: http://lkml.kernel.org/r/20151210065245.4587.39316.stgit@softrs Signed-off-by: Borislav Petkov <bp@suse.de> Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2015-12-14 18:19:13 +08:00
/* If no CPU is preparing crash dump, we simply loop here. */
run_crash_ipi_callback(regs);
panic, x86: Allow CPUs to save registers even if looping in NMI context Currently, kdump_nmi_shootdown_cpus(), a subroutine of crash_kexec(), sends an NMI IPI to CPUs which haven't called panic() to stop them, save their register information and do some cleanups for crash dumping. However, if such a CPU is infinitely looping in NMI context, we fail to save its register information into the crash dump. For example, this can happen when unknown NMIs are broadcast to all CPUs as follows: CPU 0 CPU 1 =========================== ========================== receive an unknown NMI unknown_nmi_error() panic() receive an unknown NMI spin_trylock(&panic_lock) unknown_nmi_error() crash_kexec() panic() spin_trylock(&panic_lock) panic_smp_self_stop() infinite loop kdump_nmi_shootdown_cpus() issue NMI IPI -----------> blocked until IRET infinite loop... Here, since CPU 1 is in NMI context, the second NMI from CPU 0 is blocked until CPU 1 executes IRET. However, CPU 1 never executes IRET, so the NMI is not handled and the callback function to save registers is never called. In practice, this can happen on some servers which broadcast NMIs to all CPUs when the NMI button is pushed. To save registers in this case, we need to: a) Return from NMI handler instead of looping infinitely or b) Call the callback function directly from the infinite loop Inherently, a) is risky because NMI is also used to prevent corrupted data from being propagated to devices. So, we chose b). This patch does the following: 1. Move the infinite looping of CPUs which haven't called panic() in NMI context (actually done by panic_smp_self_stop()) outside of panic() to enable us to refer pt_regs. Please note that panic_smp_self_stop() is still used for normal context. 2. Call a callback of kdump_nmi_shootdown_cpus() directly to save registers and do some cleanups after setting waiting_for_crash_ipi which is used for counting down the number of CPUs which handled the callback Signed-off-by: Hidehiro Kawai <hidehiro.kawai.ez@hitachi.com> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Aaron Tomlin <atomlin@redhat.com> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Andy Lutomirski <luto@kernel.org> Cc: Baoquan He <bhe@redhat.com> Cc: Chris Metcalf <cmetcalf@ezchip.com> Cc: Dave Young <dyoung@redhat.com> Cc: David Hildenbrand <dahi@linux.vnet.ibm.com> Cc: Don Zickus <dzickus@redhat.com> Cc: Eric Biederman <ebiederm@xmission.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Gobinda Charan Maji <gobinda.cemk07@gmail.com> Cc: HATAYAMA Daisuke <d.hatayama@jp.fujitsu.com> Cc: Hidehiro Kawai <hidehiro.kawai.ez@hitachi.com> Cc: "H. Peter Anvin" <hpa@zytor.com> Cc: Ingo Molnar <mingo@kernel.org> Cc: Javi Merino <javi.merino@arm.com> Cc: Jiang Liu <jiang.liu@linux.intel.com> Cc: Jonathan Corbet <corbet@lwn.net> Cc: kexec@lists.infradead.org Cc: linux-doc@vger.kernel.org Cc: lkml <linux-kernel@vger.kernel.org> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Michal Nazarewicz <mina86@mina86.com> Cc: Nicolas Iooss <nicolas.iooss_linux@m4x.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Prarit Bhargava <prarit@redhat.com> Cc: Rasmus Villemoes <linux@rasmusvillemoes.dk> Cc: Seth Jennings <sjenning@redhat.com> Cc: Stefan Lippers-Hollmann <s.l-h@gmx.de> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Ulrich Obergfell <uobergfe@redhat.com> Cc: Vitaly Kuznetsov <vkuznets@redhat.com> Cc: Vivek Goyal <vgoyal@redhat.com> Cc: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com> Link: http://lkml.kernel.org/r/20151210014628.25437.75256.stgit@softrs [ Cleanup comments, fixup formatting. ] Signed-off-by: Borislav Petkov <bp@suse.de> Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2015-12-14 18:19:10 +08:00
cpu_relax();
}
}
#else /* !CONFIG_SMP */
void nmi_shootdown_cpus(nmi_shootdown_cb callback)
{
/* No other CPUs to shoot down */
}
x86/nmi: Save regs in crash dump on external NMI Now, multiple CPUs can receive an external NMI simultaneously by specifying the "apic_extnmi=all" command line parameter. When we take a crash dump by using external NMI with this option, we fail to save registers into the crash dump. This happens as follows: CPU 0 CPU 1 ================================ ============================= receive an external NMI default_do_nmi() receive an external NMI spin_lock(&nmi_reason_lock) default_do_nmi() io_check_error() spin_lock(&nmi_reason_lock) panic() busy loop ... kdump_nmi_shootdown_cpus() issue NMI IPI -----------> blocked until IRET busy loop... Here, since CPU 1 is in NMI context, an additional NMI from CPU 0 remains unhandled until CPU 1 IRETs. However, CPU 1 will never execute IRET so the NMI is not handled and the callback function to save registers is never called. To solve this issue, we check if the IPI for crash dumping was issued while waiting for nmi_reason_lock to be released, and if so, call its callback function directly. If the IPI is not issued (e.g. kdump is disabled), the actual behavior doesn't change. Signed-off-by: Hidehiro Kawai <hidehiro.kawai.ez@hitachi.com> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Andy Lutomirski <luto@kernel.org> Cc: Baoquan He <bhe@redhat.com> Cc: Dave Young <dyoung@redhat.com> Cc: "Eric W. Biederman" <ebiederm@xmission.com> Cc: "H. Peter Anvin" <hpa@zytor.com> Cc: Ingo Molnar <mingo@kernel.org> Cc: Jiang Liu <jiang.liu@linux.intel.com> Cc: Jonathan Corbet <corbet@lwn.net> Cc: kexec@lists.infradead.org Cc: linux-doc@vger.kernel.org Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Stefan Lippers-Hollmann <s.l-h@gmx.de> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Vivek Goyal <vgoyal@redhat.com> Cc: x86-ml <x86@kernel.org> Link: http://lkml.kernel.org/r/20151210065245.4587.39316.stgit@softrs Signed-off-by: Borislav Petkov <bp@suse.de> Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2015-12-14 18:19:13 +08:00
void run_crash_ipi_callback(struct pt_regs *regs)
{
}
#endif