x86/crash: Fix double NMI shootdown bug
Subsystems (2)
v4 cover x86/crash: Fix double NMI shootdown bug
Fix a double NMI shootdown bug found and debugged by Guilherme, who did all the hard work. NMI shootdown is a one-time thing; the handler leaves NMIs blocked and enters halt. At best, a second (or third...) shootdown is an expensive nop, at worst it can hang the kernel and prevent kexec'ing into a new kernel, e.g. prior to the hardening of register_nmi_handler(), a double shootdown resulted in a double list_add(), which is fatal when running with CONFIG_BUG_ON_DATA_CORRUPTION=y.
With the "right" kexec/kdump configuration, emergency_vmx_disable_all() can be reached after kdump_nmi_shootdown_cpus() (currently the only two users of nmi_shootdown_cpus()).
To fix, move the disabling of virtualization into crash_nmi_callback(), remove emergency_vmx_disable_all()'s callback, and do a shootdown for emergency_vmx_disable_all() if and only if a shootdown hasn't yet occurred. The only thing emergency_vmx_disable_all() cares about is disabling VMX/SVM (obviously), and since I can't envision a use case for an NMI shootdown that doesn't want to disable virtualization, doing that in the core handler means emergency_vmx_disable_all() only needs to ensure _a_ shootdown occurs, it doesn't care when that shootdown happened or what callback may have run.
Patches 2-4 bring SVM on par with VMX.
v4:
- Set GIF=1 prior to disabling SVM. [Andrew, Tom]
- Name new helper cpu_emergency_disable_virtualization() instead of cpu_crash_disable_virtualization().
- Add patch to handle SVM in the stop/reboot case.
- Drop patch to fold __cpu_emergency_vmxoff() into cpu_emergency_vmxoff(), will be sent as part of a larger cleanup.
v3:
- https://lore.kernel.org/all/[email protected]
- Re-collect Guilherme's Tested-by.
- Tweak comment in patch 1 to reference STGI instead of CLGI.
- Celebrate this series' half-birthday.
v2:
- Use a NULL handler and crash_ipi_issued instead of a magic nop handler. [tglx]
- Add comments to call out that modifying the existing handler once the NMI is sent may cause explosions.
- Add a patch to cleanup cpu_emergency_vmxoff().
- https://lore.kernel.org/all/[email protected]
v1: https://lore.kernel.org/all/[email protected]
Sean Christopherson (4):
x86/crash: Disable virt in core NMI crash handler to avoid double
shootdown
x86/virt: Force GIF=1 prior to disabling SVM (for reboot flows) x86/reboot: Disable virtualization in an emergency if SVM is supported x86/reboot: Disable SVM, not just VMX, when stopping CPUs
arch/x86/include/asm/reboot.h | 2 + arch/x86/include/asm/virtext.h | 13 ++++- arch/x86/kernel/crash.c | 17 +------ arch/x86/kernel/reboot.c | 88 ++++++++++++++++++++++++---------- arch/x86/kernel/smp.c | 6 +-- 5 files changed, 82 insertions(+), 44 deletions(-)
base-commit: d800169041c0e035160c8b81f30d4b7e8f8ef777
Signature · 1 line
2.38.1.584.g0f3c55d4c2-goog
PATCH v4 1/4 x86/crash: Disable virt in core NMI crash handler to avoid double shootdown
Disable virtualization in crash_nmi_callback() and rework the emergency_vmx_disable_all() path to do an NMI shootdown if and only if a
shootdown has not already occurred. NMI crash shootdown fundamentally
can't support multiple invocations as responding CPUs are deliberately put into halt state without unblocking NMIs. But, the emergency reboot path doesn't have any work of its own, it simply cares about disabling virtualization, i.e. so long as a shootdown occurred, emergency reboot doesn't care who initiated the shootdown, or when.
If "crash_kexec_post_notifiers" is specified on the kernel command line, panic() will invoke crash_smp_send_stop() and result in a second call to nmi_shootdown_cpus() during native_machine_emergency_restart().
Invoke the callback _before_ disabling virtualization, as the current VMCS needs to be cleared before doing VMXOFF. Note, this results in a subtle change in ordering between disabling virtualization and stopping Intel PT on the responding CPUs. While VMX and Intel PT do interact, VMXOFF and writes to MSR_IA32_RTIT_CTL do not induce faults between one another, which is all that matters when panicking.
Harden nmi_shootdown_cpus() against multiple invocations to try and capture any such kernel bugs via a WARN instead of hanging the system during a crash/dump, e.g. prior to the recent hardening of register_nmi_handler(), re-registering the NMI handler would trigger a double list_add() and hang the system if CONFIG_BUG_ON_DATA_CORRUPTION=y.
list_add double add: new=ffffffff82220800, prev=ffffffff8221cfe8, next=ffffffff82220800.
WARNING: CPU: 2 PID: 1319 at lib/list_debug.c:29 __list_add_valid+0x67/0x70 Call Trace: __register_nmi_handler+0xcf/0x130 nmi_shootdown_cpus+0x39/0x90 native_machine_emergency_restart+0x1c9/0x1d0 panic+0x237/0x29b
Extract the disabling logic to a common helper to deduplicate code, and to prepare for doing the shootdown in the emergency reboot path if SVM is supported.
Note, prior to commit ed72736183c4 ("x86/reboot: Force all cpus to exit VMX root if VMX is supported"), nmi_shootdown_cpus() was subtly protected against a second invocation by a cpu_vmx_enabled() check as the kdump handler would disable VMX if it ran first.
--- arch/x86/include/asm/reboot.h | 2 ++ arch/x86/kernel/crash.c | 17 +-------- arch/x86/kernel/reboot.c | 65 ++++++++++++++++++++++++++++------- 3 files changed, 56 insertions(+), 28 deletions(-)
arch/x86/include/asm/reboot.h +2 -0
@@ -25,6 +25,8 @@ void __noreturn machine_real_restart(unsigned int type);#define MRR_BIOS 0#define MRR_APM 1 +void cpu_emergency_disable_virtualization (void ); +typedef void (*nmi_shootdown_cb )(int ,struct pt_regs *);void nmi_panic_self_stop (struct pt_regs *regs);void nmi_shootdown_cpus (nmi_shootdown_cb callback);
arch/x86/kernel/crash.c +1 -16
@@ -37,7 +37,6 @@#include <linux/kdebug.h> #include <asm/cpu.h> #include <asm/reboot.h> -#include <asm/virtext.h> #include <asm/intel_pt.h> #include <asm/crash.h> #include <asm/cmdline.h>
@@ -81,15 +80,6 @@ static void kdump_nmi_callback(int cpu, struct pt_regs *regs)*/ cpu_crash_vmclear_loaded_vmcss (); -/* Disable VMX or SVM if needed. -* -* We need to disable virtualization on all CPUs. -* Having VMX or SVM enabled on any CPU may break rebooting -* after the kdump kernel has finished its task. -*/ -cpu_emergency_vmxoff (); -cpu_emergency_svm_disable (); -/* * Disable Intel PT to stop its logging */
@@ -148,12 +138,7 @@ void native_machine_crash_shutdown(struct pt_regs *regs)*/ cpu_crash_vmclear_loaded_vmcss (); -/* Booting kdump kernel with VMX or SVM enabled won't work, -* because (among other limitations) we can't disable paging -* with the virt flags. -*/ -cpu_emergency_vmxoff (); -cpu_emergency_svm_disable (); +cpu_emergency_disable_virtualization ();/* * Disable Intel PT to stop its logging
arch/x86/kernel/reboot.c +53 -12
114 lines not on this page · raw message
Signature · 1 line
2.38.1.584.g0f3c55d4c2-goog
PATCH v4 2/4 x86/virt: Force GIF=1 prior to disabling SVM (for reboot flows)
In reply to Sean ChristophersonSet GIF=1 prior to disabling SVM to ensure that INIT is recognized if the kernel is disabling SVM in an emergency, e.g. if the kernel is about to jump into a crash kernel or may reboot without doing a full CPU RESET. If GIF is left cleared, the new kernel (or firmware) will be unabled to awaken APs. Eat faults on STGI (due to EFER.SVME=0) as it's possible that SVM could be disabled via NMI shootdown between reading EFER.SVME and executing STGI.
--- arch/x86/include/asm/virtext.h | 13 ++++++++++++- 1 file changed, 12 insertions(+), 1 deletion(-)
arch/x86/include/asm/virtext.h +12 -1
@@ -126,7 +126,18 @@ static inline void cpu_svm_disable(void)wrmsrl (MSR_VM_HSAVE_PA ,0 );rdmsrl (MSR_EFER , efer); -wrmsrl (MSR_EFER , efer & ~EFER_SVME ); +if (efer &EFER_SVME ) { +/* +* Force GIF=1 prior to disabling SVM, e.g. to ensure INIT and +* NMI aren't blocked. Eat faults on STGI, as it #UDs if SVM +* isn't enabled and SVM can be disabled by an NMI callback. +*/ + asm_volatile_goto("1: stgi\n\t" + _ASM_EXTABLE(1b, %l[fault]) + ::: "memory" : fault); +fault: + wrmsrl(MSR_EFER, efer & ~EFER_SVME); + } }/** Makes sure SVM is disabled, if it is supported on the CPU
Signature · 1 line
2.38.1.584.g0f3c55d4c2-goog
Re: [PATCH v4 2/4] x86/virt: Force GIF=1 prior to disabling SVM (for reboot flows)
Quoted 5 lines · Sean Christopherson
arch/x86/include/asm/virtext.h +0
-0
arch/x86/include/asm/virtext.h static inline void cpu_svm_disable(void) · line 133
Quoted 9 lines
arch/x86/include/asm/virtext.h +5
-1@@ -126,7 +126,18 @@ static inline void cpu_svm_disable(void)wrmsrl (MSR_VM_HSAVE_PA ,0 );rdmsrl (MSR_EFER , efer); -wrmsrl (MSR_EFER , efer & ~EFER_SVME ); +if (efer &EFER_SVME ) { +/* +* Force GIF=1 prior to disabling SVM, e.g. to ensure INIT and +* NMI aren't blocked. Eat faults on STGI, as it #UDs if SVM +* isn't enabled and SVM can be disabled by an NMI callback.
I'd be tempted to tweak this for clarity.
How about "We don't know the state of GIF, and if NMIs are enabled, there is a race condition where EFER.SVME can be cleared behind our back. Ignore #UD, and force GIF=1 in case INIT/NMI are currently blocked." ?
The STGI can't actually #UD on real hardware, because SKINIT and SVM exist in identical sets of parts, but it can #UD in principle in a VM which doesn't offer emulate SKINIT.
Given that we are in cpu_svm_disable(), there's also MSR_VM_CR.INIT_REDIRECTION to consider, but perhaps that's better left to the series which adds SKINIT support.
~Andrew
Re: [PATCH v4 2/4] x86/virt: Force GIF=1 prior to disabling SVM (for reboot flows)
Quoted 18 lines · Andrew Cooper
Quoted 5 lines · Sean Christopherson
arch/x86/include/asm/virtext.h +0
-0arch/x86/include/asm/virtext.h static inline void cpu_svm_disable(void) · line 133
Quoted 9 lines
arch/x86/include/asm/virtext.h +5
-1@@ -126,7 +126,18 @@ static inline void cpu_svm_disable(void)wrmsrl (MSR_VM_HSAVE_PA ,0 );rdmsrl (MSR_EFER , efer); -wrmsrl (MSR_EFER , efer & ~EFER_SVME ); +if (efer &EFER_SVME ) { +/* +* Force GIF=1 prior to disabling SVM, e.g. to ensure INIT and +* NMI aren't blocked. Eat faults on STGI, as it #UDs if SVM +* isn't enabled and SVM can be disabled by an NMI callback. I'd be tempted to tweak this for clarity.
Quoted 8 lines
How about "We don't know the state of GIF, and if NMIs are enabled, there is a race condition where EFER.SVME can be cleared behind our back. Ignore #UD, and force GIF=1 in case INIT/NMI are currently blocked." ?
The STGI can't actually #UD on real hardware, because SKINIT and SVM exist in identical sets of parts, but it can #UD in principle in a VM which doesn't offer emulate SKINIT.
Ah, right, "may #UD", not "will #UD". And despite writing this, I also keep forgetting why forcing GIF is even necessary. How about?
/*
* Force GIF=1 prior to disabling SVM to ensure INIT and NMI
* aren't blocked, e.g. if a fatal error occurred between CLGI
* and STGI. Note, STGI may #UD if SVM is disabled from NMI
* context between reading EFER and executing STGI. In that
* case, GIF must already be set, otherwise the NMI would have
* been blocked, so just eat the fault.
*/
Quoted 5 lines
Given that we are in
cpu_svm_disable(), there's also MSR_VM_CR.INIT_REDIRECTION to consider, but perhaps that's better left to the series which adds SKINIT support.~Andrew
Re: [PATCH v4 2/4] x86/virt: Force GIF=1 prior to disabling SVM (for reboot flows)
Quoted 29 lines · Sean Christopherson
Quoted 17 lines · Andrew Cooper
Quoted 5 lines · Sean Christopherson
arch/x86/include/asm/virtext.h +0
-0arch/x86/include/asm/virtext.h static inline void cpu_svm_disable(void) · line 133
Quoted 9 lines
arch/x86/include/asm/virtext.h +5
-1@@ -126,7 +126,18 @@ static inline void cpu_svm_disable(void)wrmsrl (MSR_VM_HSAVE_PA ,0 );rdmsrl (MSR_EFER , efer); -wrmsrl (MSR_EFER , efer & ~EFER_SVME ); +if (efer &EFER_SVME ) { +/* +* Force GIF=1 prior to disabling SVM, e.g. to ensure INIT and +* NMI aren't blocked. Eat faults on STGI, as it #UDs if SVM +* isn't enabled and SVM can be disabled by an NMI callback. I'd be tempted to tweak this for clarity.
Quoted 8 lines
How about "We don't know the state of GIF, and if NMIs are enabled, there is a race condition where EFER.SVME can be cleared behind our back. Ignore #UD, and force GIF=1 in case INIT/NMI are currently blocked." ?
The STGI can't actually #UD on real hardware, because SKINIT and SVM exist in identical sets of parts, but it can #UD in principle in a VM which doesn't offer emulate SKINIT.
Ah, right, "may #UD", not "will #UD". And despite writing this, I also keep forgetting why forcing GIF is even necessary. How about?
Quoted 8 lines
/*
* Force GIF=1 prior to disabling SVM to ensure INIT and NMI
* aren't blocked, e.g. if a fatal error occurred between CLGI
* and STGI. Note, STGI may #UD if SVM is disabled from NMI
* context between reading EFER and executing STGI. In that
* case, GIF must already be set, otherwise the NMI would have
* been blocked, so just eat the fault.
*/
LGTM.
~Andrew
PATCH v4 3/4 x86/reboot: Disable virtualization in an emergency if SVM is supported
In reply to Sean ChristophersonDisable SVM on all CPUs via NMI shootdown during an emergency reboot. Like VMX, SVM can block INIT, e.g. if the emergency reboot is triggered between CLGI and STGI, and thus can prevent bringing up other CPUs via INIT-SIPI-SIPI.
--- arch/x86/kernel/reboot.c | 23 +++++++++++------------ 1 file changed, 11 insertions(+), 12 deletions(-)
arch/x86/kernel/reboot.c +11 -12
@@ -530,27 +530,26 @@ static inline void kb_wait(void)static inline void nmi_shootdown_cpus_on_restart (void ); -/* Use NMIs as IPIs to tell all CPUs to disable virtualization */ -static void emergency_vmx_disable_all (void ) +static void emergency_reboot_disable_virtualization (void ) {/* Just make sure we won't change CPUs while doing this */ local_irq_disable ();/* -* Disable VMX on all CPUs before rebooting, otherwise we risk hanging -* the machine, because the CPU blocks INIT when it's in VMX root. +* Disable virtualization on all CPUs before rebooting to avoid hanging +* the system, as VMX and SVM block INIT when running in the host. * * We can't take any locks and we may be on an inconsistent state, so -* use NMIs as IPIs to tell the other CPUs to exit VMX root and halt. +* use NMIs as IPIs to tell the other CPUs to disable VMX/SVM and halt. * -* Do the NMI shootdown even if VMX if off on _this_ CPU, as that -* doesn't prevent a different CPU from being in VMX root operation. +* Do the NMI shootdown even if virtualization is off on _this_ CPU, as +* other CPUs may have virtualization enabled. */ -if (cpu_has_vmx ()) { -/* Safely force _this_ CPU out of VMX root operation. */ -__cpu_emergency_vmxoff (); +if (cpu_has_vmx () ||cpu_has_svm (NULL )) { +/* Safely force _this_ CPU out of VMX/SVM operation. */ +cpu_emergency_disable_virtualization (); -/* Halt and exit VMX root operation on the other CPUs. */ +/* Disable VMX/SVM and halt on other CPUs. */ nmi_shootdown_cpus_on_restart (); } }
@@ -587,7 +586,7 @@ static void native_machine_emergency_restart(void)unsigned short mode;if (reboot_emergency) -emergency_vmx_disable_all (); +emergency_reboot_disable_virtualization ();tboot_shutdown (TB_SHUTDOWN_REBOOT );
Signature · 1 line
2.38.1.584.g0f3c55d4c2-goog
PATCH v4 4/4 x86/reboot: Disable SVM, not just VMX, when stopping CPUs
In reply to Sean ChristophersonDisable SVM and more importantly force GIF=1 when halting a CPU or rebooting the machine. Similar to VMX, SVM allows software to block INITs via CLGI, and thus can be problematic for a crash/reboot. The window for failure is smaller with SVM as INIT is only blocked while GIF=0, i.e. between CLGI and STGI, but the window does exist.
--- arch/x86/kernel/smp.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-)
arch/x86/kernel/smp.c +3 -3
@@ -32,7 +32,7 @@#include <asm/mce.h> #include <asm/trace/irq_vectors.h> #include <asm/kexec.h> -#include <asm/virtext.h> +#include <asm/reboot.h> /* * Some notes on x86 processor bugs affecting SMP operation:
@@ -122,7 +122,7 @@ static int smp_stop_nmi_callback(unsigned int val, struct pt_regs *regs)if (raw_smp_processor_id () ==atomic_read (&stopping_cpu))return NMI_HANDLED ; -cpu_emergency_vmxoff (); +cpu_emergency_disable_virtualization ();stop_this_cpu (NULL );return NMI_HANDLED ;
@@ -134,7 +134,7 @@ static int smp_stop_nmi_callback(unsigned int val, struct pt_regs *regs)DEFINE_IDTENTRY_SYSVEC (sysvec_reboot) {ack_APIC_irq (); -cpu_emergency_vmxoff (); +cpu_emergency_disable_virtualization ();stop_this_cpu (NULL ); }
Signature · 1 line
2.38.1.584.g0f3c55d4c2-goog
Re: [PATCH v4 4/4] x86/reboot: Disable SVM, not just VMX, when stopping CPUs
Quoted 8 lines · Sean Christopherson
Disable SVM and more importantly force GIF=1 when halting a CPU or rebooting the machine. Similar to VMX, SVM allows software to block INITs via CLGI, and thus can be problematic for a crash/reboot. The window for failure is smaller with SVM as INIT is only blocked while GIF=0, i.e. between CLGI and STGI, but the window does exist.
Fixes fba4f472b33a ("x86/reboot: Turn off KVM when halting a CPU")
Argh, forgot the .org, and of course my scripts then failed to filter out the address.
Quoted 12 lines · Sean Christopherson
Fix a double NMI shootdown bug found and debugged by Guilherme, who did all the hard work. NMI shootdown is a one-time thing; the handler leaves NMIs blocked and enters halt. At best, a second (or third...) shootdown is an expensive nop, at worst it can hang the kernel and prevent kexec'ing into a new kernel, e.g. prior to the hardening of
register_nmi_handler(), a double shootdown resulted in a doublelist_add(), which is fatal when running withCONFIG_BUG_ON_DATA_CORRUPTION=y.With the "right" kexec/kdump configuration,
emergency_vmx_disable_all()can be reached afterkdump_nmi_shootdown_cpus()(currently the only two users ofnmi_shootdown_cpus()).
Quoted 8 lines
To fix, move the disabling of virtualization into
crash_nmi_callback(), remove emergency_vmx_disable_all()'s callback, and do a shootdown foremergency_vmx_disable_all()if and only if a shootdown hasn't yet occurred. The only thingemergency_vmx_disable_all()cares about is disabling VMX/SVM (obviously), and since I can't envision a use case for an NMI shootdown that doesn't want to disable virtualization, doing that in the core handler meansemergency_vmx_disable_all()only needs to ensure _a_ shootdown occurs, it doesn't care when that shootdown happened or what callback may have run.
Sean Christopherson
AppliedQuoted 9 lines · Sean Christopherson
Fix a double NMI shootdown bug found and debugged by Guilherme, who did all the hard work. NMI shootdown is a one-time thing; the handler leaves NMIs blocked and enters halt. At best, a second (or third...) shootdown is an expensive nop, at worst it can hang the kernel and prevent kexec'ing into a new kernel, e.g. prior to the hardening of
register_nmi_handler(), a double shootdown resulted in a doublelist_add(), which is fatal when running withCONFIG_BUG_ON_DATA_CORRUPTION=y.[...]
Applied to kvm-x86 misc, thanks!
[1/4] x86/crash: Disable virt in core NMI crash handler to avoid double shootdown
https://github.com/kvm-x86/linux/commit/26044aff37a5 [2/4] x86/virt: Force GIF=1 prior to disabling SVM (for reboot flows) https://github.com/kvm-x86/linux/commit/6a3236580b0b [3/4] x86/reboot: Disable virtualization in an emergency if SVM is supported https://github.com/kvm-x86/linux/commit/d81f952aa657 [4/4] x86/reboot: Disable SVM, not just VMX, when stopping CPUs https://github.com/kvm-x86/linux/commit/a2b07fa7b933
--
https://github.com/kvm-x86/linux/tree/next
https://github.com/kvm-x86/linux/tree/fixes
Quoted 22 lines · Sean Christopherson
Quoted 9 lines · Sean Christopherson
Fix a double NMI shootdown bug found and debugged by Guilherme, who did all the hard work. NMI shootdown is a one-time thing; the handler leaves NMIs blocked and enters halt. At best, a second (or third...) shootdown is an expensive nop, at worst it can hang the kernel and prevent kexec'ing into a new kernel, e.g. prior to the hardening of
register_nmi_handler(), a double shootdown resulted in a doublelist_add(), which is fatal when running withCONFIG_BUG_ON_DATA_CORRUPTION=y.[...]
Applied to kvm-x86 misc, thanks!
[1/4] x86/crash: Disable virt in core NMI crash handler to avoid double shootdown
https://github.com/kvm-x86/linux/commit/26044aff37a5 [2/4] x86/virt: Force GIF=1 prior to disabling SVM (for reboot flows) https://github.com/kvm-x86/linux/commit/6a3236580b0b [3/4] x86/reboot: Disable virtualization in an emergency if SVM is supported https://github.com/kvm-x86/linux/commit/d81f952aa657 [4/4] x86/reboot: Disable SVM, not just VMX, when stopping CPUs https://github.com/kvm-x86/linux/commit/a2b07fa7b933
Quoted 3 lines
--
https://github.com/kvm-x86/linux/tree/next
https://github.com/kvm-x86/linux/tree/fixes
Thanks a bunch Sean!