2007-10-10 22:16:19 +07:00
|
|
|
/*
|
|
|
|
* Kernel-based Virtual Machine driver for Linux
|
|
|
|
*
|
|
|
|
* derived from drivers/kvm/kvm_main.c
|
|
|
|
*
|
|
|
|
* Copyright (C) 2006 Qumranet, Inc.
|
2008-07-28 23:26:26 +07:00
|
|
|
* Copyright (C) 2008 Qumranet, Inc.
|
|
|
|
* Copyright IBM Corporation, 2008
|
2010-10-06 19:23:22 +07:00
|
|
|
* Copyright 2010 Red Hat, Inc. and/or its affiliates.
|
2007-10-10 22:16:19 +07:00
|
|
|
*
|
|
|
|
* Authors:
|
|
|
|
* Avi Kivity <avi@qumranet.com>
|
|
|
|
* Yaniv Kamay <yaniv@qumranet.com>
|
2008-07-28 23:26:26 +07:00
|
|
|
* Amit Shah <amit.shah@qumranet.com>
|
|
|
|
* Ben-Ami Yassour <benami@il.ibm.com>
|
2007-10-10 22:16:19 +07:00
|
|
|
*
|
|
|
|
* This work is licensed under the terms of the GNU GPL, version 2. See
|
|
|
|
* the COPYING file in the top-level directory.
|
|
|
|
*
|
|
|
|
*/
|
|
|
|
|
2007-12-16 16:02:48 +07:00
|
|
|
#include <linux/kvm_host.h>
|
KVM: Portability: split kvm_vcpu_ioctl
This patch splits kvm_vcpu_ioctl into archtecture independent parts, and
x86 specific parts which go to kvm_arch_vcpu_ioctl in x86.c.
Common ioctls for all architectures are:
KVM_RUN, KVM_GET/SET_(S-)REGS, KVM_TRANSLATE, KVM_INTERRUPT,
KVM_DEBUG_GUEST, KVM_SET_SIGNAL_MASK, KVM_GET/SET_FPU
Note that some PPC chips don't have an FPU, so we might need an #ifdef
around KVM_GET/SET_FPU one day.
x86 specific ioctls are:
KVM_GET/SET_LAPIC, KVM_SET_CPUID, KVM_GET/SET_MSRS
An interresting aspect is vcpu_load/vcpu_put. We now have a common
vcpu_load/put which does the preemption stuff, and an architecture
specific kvm_arch_vcpu_load/put. In the x86 case, this one calls the
vmx/svm function defined in kvm_x86_ops.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Reviewed-by: Christian Ehrhardt <ehrhardt@linux.vnet.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-12 00:16:52 +07:00
|
|
|
#include "irq.h"
|
2007-12-14 08:35:10 +07:00
|
|
|
#include "mmu.h"
|
2008-01-28 04:10:22 +07:00
|
|
|
#include "i8254.h"
|
2008-03-25 04:14:53 +07:00
|
|
|
#include "tss.h"
|
2008-06-28 00:58:02 +07:00
|
|
|
#include "kvm_cache_regs.h"
|
2008-07-03 18:59:22 +07:00
|
|
|
#include "x86.h"
|
2011-11-23 21:30:32 +07:00
|
|
|
#include "cpuid.h"
|
2014-11-24 21:27:17 +07:00
|
|
|
#include "assigned-dev.h"
|
2015-06-19 18:54:23 +07:00
|
|
|
#include "pmu.h"
|
2015-07-03 19:01:34 +07:00
|
|
|
#include "hyperv.h"
|
KVM: Portability: split kvm_vcpu_ioctl
This patch splits kvm_vcpu_ioctl into archtecture independent parts, and
x86 specific parts which go to kvm_arch_vcpu_ioctl in x86.c.
Common ioctls for all architectures are:
KVM_RUN, KVM_GET/SET_(S-)REGS, KVM_TRANSLATE, KVM_INTERRUPT,
KVM_DEBUG_GUEST, KVM_SET_SIGNAL_MASK, KVM_GET/SET_FPU
Note that some PPC chips don't have an FPU, so we might need an #ifdef
around KVM_GET/SET_FPU one day.
x86 specific ioctls are:
KVM_GET/SET_LAPIC, KVM_SET_CPUID, KVM_GET/SET_MSRS
An interresting aspect is vcpu_load/vcpu_put. We now have a common
vcpu_load/put which does the preemption stuff, and an architecture
specific kvm_arch_vcpu_load/put. In the x86 case, this one calls the
vmx/svm function defined in kvm_x86_ops.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Reviewed-by: Christian Ehrhardt <ehrhardt@linux.vnet.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-12 00:16:52 +07:00
|
|
|
|
2008-02-16 02:52:47 +07:00
|
|
|
#include <linux/clocksource.h>
|
2008-07-28 23:26:26 +07:00
|
|
|
#include <linux/interrupt.h>
|
KVM: Portability: split kvm_vcpu_ioctl
This patch splits kvm_vcpu_ioctl into archtecture independent parts, and
x86 specific parts which go to kvm_arch_vcpu_ioctl in x86.c.
Common ioctls for all architectures are:
KVM_RUN, KVM_GET/SET_(S-)REGS, KVM_TRANSLATE, KVM_INTERRUPT,
KVM_DEBUG_GUEST, KVM_SET_SIGNAL_MASK, KVM_GET/SET_FPU
Note that some PPC chips don't have an FPU, so we might need an #ifdef
around KVM_GET/SET_FPU one day.
x86 specific ioctls are:
KVM_GET/SET_LAPIC, KVM_SET_CPUID, KVM_GET/SET_MSRS
An interresting aspect is vcpu_load/vcpu_put. We now have a common
vcpu_load/put which does the preemption stuff, and an architecture
specific kvm_arch_vcpu_load/put. In the x86 case, this one calls the
vmx/svm function defined in kvm_x86_ops.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Reviewed-by: Christian Ehrhardt <ehrhardt@linux.vnet.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-12 00:16:52 +07:00
|
|
|
#include <linux/kvm.h>
|
|
|
|
#include <linux/fs.h>
|
|
|
|
#include <linux/vmalloc.h>
|
2016-07-14 07:19:00 +07:00
|
|
|
#include <linux/export.h>
|
|
|
|
#include <linux/moduleparam.h>
|
2007-11-20 15:25:04 +07:00
|
|
|
#include <linux/mman.h>
|
2007-12-12 22:46:12 +07:00
|
|
|
#include <linux/highmem.h>
|
2008-12-03 20:43:34 +07:00
|
|
|
#include <linux/iommu.h>
|
2008-09-14 07:48:28 +07:00
|
|
|
#include <linux/intel-iommu.h>
|
2009-02-04 23:52:04 +07:00
|
|
|
#include <linux/cpufreq.h>
|
2009-09-07 15:12:18 +07:00
|
|
|
#include <linux/user-return-notifier.h>
|
2009-12-23 23:35:23 +07:00
|
|
|
#include <linux/srcu.h>
|
include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h
percpu.h is included by sched.h and module.h and thus ends up being
included when building most .c files. percpu.h includes slab.h which
in turn includes gfp.h making everything defined by the two files
universally available and complicating inclusion dependencies.
percpu.h -> slab.h dependency is about to be removed. Prepare for
this change by updating users of gfp and slab facilities include those
headers directly instead of assuming availability. As this conversion
needs to touch large number of source files, the following script is
used as the basis of conversion.
http://userweb.kernel.org/~tj/misc/slabh-sweep.py
The script does the followings.
* Scan files for gfp and slab usages and update includes such that
only the necessary includes are there. ie. if only gfp is used,
gfp.h, if slab is used, slab.h.
* When the script inserts a new include, it looks at the include
blocks and try to put the new include such that its order conforms
to its surrounding. It's put in the include block which contains
core kernel includes, in the same order that the rest are ordered -
alphabetical, Christmas tree, rev-Xmas-tree or at the end if there
doesn't seem to be any matching order.
* If the script can't find a place to put a new include (mostly
because the file doesn't have fitting include block), it prints out
an error message indicating which .h file needs to be added to the
file.
The conversion was done in the following steps.
1. The initial automatic conversion of all .c files updated slightly
over 4000 files, deleting around 700 includes and adding ~480 gfp.h
and ~3000 slab.h inclusions. The script emitted errors for ~400
files.
2. Each error was manually checked. Some didn't need the inclusion,
some needed manual addition while adding it to implementation .h or
embedding .c file was more appropriate for others. This step added
inclusions to around 150 files.
3. The script was run again and the output was compared to the edits
from #2 to make sure no file was left behind.
4. Several build tests were done and a couple of problems were fixed.
e.g. lib/decompress_*.c used malloc/free() wrappers around slab
APIs requiring slab.h to be added manually.
5. The script was run on all .h files but without automatically
editing them as sprinkling gfp.h and slab.h inclusions around .h
files could easily lead to inclusion dependency hell. Most gfp.h
inclusion directives were ignored as stuff from gfp.h was usually
wildly available and often used in preprocessor macros. Each
slab.h inclusion directive was examined and added manually as
necessary.
6. percpu.h was updated not to include slab.h.
7. Build test were done on the following configurations and failures
were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my
distributed build env didn't work with gcov compiles) and a few
more options had to be turned off depending on archs to make things
build (like ipr on powerpc/64 which failed due to missing writeq).
* x86 and x86_64 UP and SMP allmodconfig and a custom test config.
* powerpc and powerpc64 SMP allmodconfig
* sparc and sparc64 SMP allmodconfig
* ia64 SMP allmodconfig
* s390 SMP allmodconfig
* alpha SMP allmodconfig
* um on x86_64 SMP allmodconfig
8. percpu.h modifications were reverted so that it could be applied as
a separate patch and serve as bisection point.
Given the fact that I had only a couple of failures from tests on step
6, I'm fairly confident about the coverage of this conversion patch.
If there is a breakage, it's likely to be something in one of the arch
headers which should be easily discoverable easily on most builds of
the specific arch.
Signed-off-by: Tejun Heo <tj@kernel.org>
Guess-its-ok-by: Christoph Lameter <cl@linux-foundation.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
2010-03-24 15:04:11 +07:00
|
|
|
#include <linux/slab.h>
|
2010-04-19 12:32:45 +07:00
|
|
|
#include <linux/perf_event.h>
|
2010-06-02 16:06:03 +07:00
|
|
|
#include <linux/uaccess.h>
|
2010-10-14 16:22:46 +07:00
|
|
|
#include <linux/hash.h>
|
2011-09-06 23:46:34 +07:00
|
|
|
#include <linux/pci.h>
|
2012-11-28 08:29:00 +07:00
|
|
|
#include <linux/timekeeper_internal.h>
|
|
|
|
#include <linux/pvclock_gtod.h>
|
2015-09-18 21:29:40 +07:00
|
|
|
#include <linux/kvm_irqfd.h>
|
|
|
|
#include <linux/irqbypass.h>
|
2009-07-01 20:01:02 +07:00
|
|
|
#include <trace/events/kvm.h>
|
2010-03-10 18:00:43 +07:00
|
|
|
|
2009-09-10 00:22:48 +07:00
|
|
|
#include <asm/debugreg.h>
|
2007-11-14 19:08:51 +07:00
|
|
|
#include <asm/msr.h>
|
2008-02-20 22:57:21 +07:00
|
|
|
#include <asm/desc.h>
|
2009-05-11 15:48:15 +07:00
|
|
|
#include <asm/mce.h>
|
2015-04-22 15:58:10 +07:00
|
|
|
#include <linux/kernel_stat.h>
|
2015-04-24 07:54:44 +07:00
|
|
|
#include <asm/fpu/internal.h> /* Ugh! */
|
2010-08-20 15:07:30 +07:00
|
|
|
#include <asm/pvclock.h>
|
2010-08-26 17:38:03 +07:00
|
|
|
#include <asm/div64.h>
|
2015-09-18 21:29:51 +07:00
|
|
|
#include <asm/irq_remapping.h>
|
2007-10-10 22:16:19 +07:00
|
|
|
|
2016-06-02 00:42:20 +07:00
|
|
|
#define CREATE_TRACE_POINTS
|
|
|
|
#include "trace.h"
|
|
|
|
|
KVM: Portability: split kvm_vcpu_ioctl
This patch splits kvm_vcpu_ioctl into archtecture independent parts, and
x86 specific parts which go to kvm_arch_vcpu_ioctl in x86.c.
Common ioctls for all architectures are:
KVM_RUN, KVM_GET/SET_(S-)REGS, KVM_TRANSLATE, KVM_INTERRUPT,
KVM_DEBUG_GUEST, KVM_SET_SIGNAL_MASK, KVM_GET/SET_FPU
Note that some PPC chips don't have an FPU, so we might need an #ifdef
around KVM_GET/SET_FPU one day.
x86 specific ioctls are:
KVM_GET/SET_LAPIC, KVM_SET_CPUID, KVM_GET/SET_MSRS
An interresting aspect is vcpu_load/vcpu_put. We now have a common
vcpu_load/put which does the preemption stuff, and an architecture
specific kvm_arch_vcpu_load/put. In the x86 case, this one calls the
vmx/svm function defined in kvm_x86_ops.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Reviewed-by: Christian Ehrhardt <ehrhardt@linux.vnet.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-12 00:16:52 +07:00
|
|
|
#define MAX_IO_MSRS 256
|
2009-05-11 15:48:15 +07:00
|
|
|
#define KVM_MAX_MCE_BANKS 32
|
2016-06-22 13:59:56 +07:00
|
|
|
u64 __read_mostly kvm_mce_cap_supported = MCG_CTL_P | MCG_SER_P;
|
|
|
|
EXPORT_SYMBOL_GPL(kvm_mce_cap_supported);
|
2009-05-11 15:48:15 +07:00
|
|
|
|
2011-04-20 17:37:53 +07:00
|
|
|
#define emul_to_vcpu(ctxt) \
|
|
|
|
container_of(ctxt, struct kvm_vcpu, arch.emulate_ctxt)
|
|
|
|
|
2008-01-31 20:57:38 +07:00
|
|
|
/* EFER defaults:
|
|
|
|
* - enable syscall per default because its emulated by KVM
|
|
|
|
* - enable LME and LMA per default on 64 bit KVM
|
|
|
|
*/
|
|
|
|
#ifdef CONFIG_X86_64
|
2011-02-21 10:51:35 +07:00
|
|
|
static
|
|
|
|
u64 __read_mostly efer_reserved_bits = ~((u64)(EFER_SCE | EFER_LME | EFER_LMA));
|
2008-01-31 20:57:38 +07:00
|
|
|
#else
|
2011-02-21 10:51:35 +07:00
|
|
|
static u64 __read_mostly efer_reserved_bits = ~((u64)EFER_SCE);
|
2008-01-31 20:57:38 +07:00
|
|
|
#endif
|
KVM: Portability: split kvm_vcpu_ioctl
This patch splits kvm_vcpu_ioctl into archtecture independent parts, and
x86 specific parts which go to kvm_arch_vcpu_ioctl in x86.c.
Common ioctls for all architectures are:
KVM_RUN, KVM_GET/SET_(S-)REGS, KVM_TRANSLATE, KVM_INTERRUPT,
KVM_DEBUG_GUEST, KVM_SET_SIGNAL_MASK, KVM_GET/SET_FPU
Note that some PPC chips don't have an FPU, so we might need an #ifdef
around KVM_GET/SET_FPU one day.
x86 specific ioctls are:
KVM_GET/SET_LAPIC, KVM_SET_CPUID, KVM_GET/SET_MSRS
An interresting aspect is vcpu_load/vcpu_put. We now have a common
vcpu_load/put which does the preemption stuff, and an architecture
specific kvm_arch_vcpu_load/put. In the x86 case, this one calls the
vmx/svm function defined in kvm_x86_ops.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Reviewed-by: Christian Ehrhardt <ehrhardt@linux.vnet.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-12 00:16:52 +07:00
|
|
|
|
2007-11-18 21:24:12 +07:00
|
|
|
#define VM_STAT(x) offsetof(struct kvm, stat.x), KVM_STAT_VM
|
|
|
|
#define VCPU_STAT(x) offsetof(struct kvm_vcpu, stat.x), KVM_STAT_VCPU
|
2007-11-01 05:24:23 +07:00
|
|
|
|
2016-07-13 03:09:28 +07:00
|
|
|
#define KVM_X2APIC_API_VALID_FLAGS (KVM_X2APIC_API_USE_32BIT_IDS | \
|
|
|
|
KVM_X2APIC_API_DISABLE_BROADCAST_QUIRK)
|
2016-07-13 03:09:27 +07:00
|
|
|
|
2009-08-09 19:17:40 +07:00
|
|
|
static void update_cr8_intercept(struct kvm_vcpu *vcpu);
|
2011-09-20 17:43:14 +07:00
|
|
|
static void process_nmi(struct kvm_vcpu *vcpu);
|
2016-06-02 03:26:01 +07:00
|
|
|
static void enter_smm(struct kvm_vcpu *vcpu);
|
2014-03-27 17:29:28 +07:00
|
|
|
static void __kvm_set_rflags(struct kvm_vcpu *vcpu, unsigned long rflags);
|
2008-02-11 23:37:23 +07:00
|
|
|
|
2015-11-06 17:46:24 +07:00
|
|
|
struct kvm_x86_ops *kvm_x86_ops __read_mostly;
|
2008-06-28 00:58:02 +07:00
|
|
|
EXPORT_SYMBOL_GPL(kvm_x86_ops);
|
2007-11-14 19:09:30 +07:00
|
|
|
|
2015-11-06 17:46:24 +07:00
|
|
|
static bool __read_mostly ignore_msrs = 0;
|
2012-01-13 06:02:18 +07:00
|
|
|
module_param(ignore_msrs, bool, S_IRUGO | S_IWUSR);
|
2009-06-25 17:36:49 +07:00
|
|
|
|
2014-01-06 21:00:02 +07:00
|
|
|
unsigned int min_timer_period_us = 500;
|
|
|
|
module_param(min_timer_period_us, uint, S_IRUGO | S_IWUSR);
|
|
|
|
|
2015-05-13 08:42:04 +07:00
|
|
|
static bool __read_mostly kvmclock_periodic_sync = true;
|
|
|
|
module_param(kvmclock_periodic_sync, bool, S_IRUGO);
|
|
|
|
|
2015-11-06 17:46:24 +07:00
|
|
|
bool __read_mostly kvm_has_tsc_control;
|
2011-03-25 15:44:51 +07:00
|
|
|
EXPORT_SYMBOL_GPL(kvm_has_tsc_control);
|
2015-11-06 17:46:24 +07:00
|
|
|
u32 __read_mostly kvm_max_guest_tsc_khz;
|
2011-03-25 15:44:51 +07:00
|
|
|
EXPORT_SYMBOL_GPL(kvm_max_guest_tsc_khz);
|
2015-10-20 14:39:01 +07:00
|
|
|
u8 __read_mostly kvm_tsc_scaling_ratio_frac_bits;
|
|
|
|
EXPORT_SYMBOL_GPL(kvm_tsc_scaling_ratio_frac_bits);
|
|
|
|
u64 __read_mostly kvm_max_tsc_scaling_ratio;
|
|
|
|
EXPORT_SYMBOL_GPL(kvm_max_tsc_scaling_ratio);
|
2016-06-14 04:19:59 +07:00
|
|
|
u64 __read_mostly kvm_default_tsc_scaling_ratio;
|
|
|
|
EXPORT_SYMBOL_GPL(kvm_default_tsc_scaling_ratio);
|
2011-03-25 15:44:51 +07:00
|
|
|
|
KVM: Infrastructure for software and hardware based TSC rate scaling
This requires some restructuring; rather than use 'virtual_tsc_khz'
to indicate whether hardware rate scaling is in effect, we consider
each VCPU to always have a virtual TSC rate. Instead, there is new
logic above the vendor-specific hardware scaling that decides whether
it is even necessary to use and updates all rate variables used by
common code. This means we can simply query the virtual rate at
any point, which is needed for software rate scaling.
There is also now a threshold added to the TSC rate scaling; minor
differences and variations of measured TSC rate can accidentally
provoke rate scaling to be used when it is not needed. Instead,
we have a tolerance variable called tsc_tolerance_ppm, which is
the maximum variation from user requested rate at which scaling
will be used. The default is 250ppm, which is the half the
threshold for NTP adjustment, allowing for some hardware variation.
In the event that hardware rate scaling is not available, we can
kludge a bit by forcing TSC catchup to turn on when a faster than
hardware speed has been requested, but there is nothing available
yet for the reverse case; this requires a trap and emulate software
implementation for RDTSC, which is still forthcoming.
[avi: fix 64-bit division on i386]
Signed-off-by: Zachary Amsden <zamsden@gmail.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
2012-02-04 00:43:50 +07:00
|
|
|
/* tsc tolerance in parts per million - default to 1/2 of the NTP threshold */
|
2015-11-06 17:46:24 +07:00
|
|
|
static u32 __read_mostly tsc_tolerance_ppm = 250;
|
KVM: Infrastructure for software and hardware based TSC rate scaling
This requires some restructuring; rather than use 'virtual_tsc_khz'
to indicate whether hardware rate scaling is in effect, we consider
each VCPU to always have a virtual TSC rate. Instead, there is new
logic above the vendor-specific hardware scaling that decides whether
it is even necessary to use and updates all rate variables used by
common code. This means we can simply query the virtual rate at
any point, which is needed for software rate scaling.
There is also now a threshold added to the TSC rate scaling; minor
differences and variations of measured TSC rate can accidentally
provoke rate scaling to be used when it is not needed. Instead,
we have a tolerance variable called tsc_tolerance_ppm, which is
the maximum variation from user requested rate at which scaling
will be used. The default is 250ppm, which is the half the
threshold for NTP adjustment, allowing for some hardware variation.
In the event that hardware rate scaling is not available, we can
kludge a bit by forcing TSC catchup to turn on when a faster than
hardware speed has been requested, but there is nothing available
yet for the reverse case; this requires a trap and emulate software
implementation for RDTSC, which is still forthcoming.
[avi: fix 64-bit division on i386]
Signed-off-by: Zachary Amsden <zamsden@gmail.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
2012-02-04 00:43:50 +07:00
|
|
|
module_param(tsc_tolerance_ppm, uint, S_IRUGO | S_IWUSR);
|
|
|
|
|
2014-12-16 21:08:15 +07:00
|
|
|
/* lapic timer advance (tscdeadline mode only) in nanoseconds */
|
2015-11-06 17:46:24 +07:00
|
|
|
unsigned int __read_mostly lapic_timer_advance_ns = 0;
|
2014-12-16 21:08:15 +07:00
|
|
|
module_param(lapic_timer_advance_ns, uint, S_IRUGO | S_IWUSR);
|
|
|
|
|
2016-01-25 15:53:33 +07:00
|
|
|
static bool __read_mostly vector_hashing = true;
|
|
|
|
module_param(vector_hashing, bool, S_IRUGO);
|
|
|
|
|
2015-11-06 17:46:24 +07:00
|
|
|
static bool __read_mostly backwards_tsc_observed = false;
|
2014-05-14 22:43:24 +07:00
|
|
|
|
2009-09-07 15:12:18 +07:00
|
|
|
#define KVM_NR_SHARED_MSRS 16
|
|
|
|
|
|
|
|
struct kvm_shared_msrs_global {
|
|
|
|
int nr;
|
2009-12-18 15:48:44 +07:00
|
|
|
u32 msrs[KVM_NR_SHARED_MSRS];
|
2009-09-07 15:12:18 +07:00
|
|
|
};
|
|
|
|
|
|
|
|
struct kvm_shared_msrs {
|
|
|
|
struct user_return_notifier urn;
|
|
|
|
bool registered;
|
2009-12-18 15:48:44 +07:00
|
|
|
struct kvm_shared_msr_values {
|
|
|
|
u64 host;
|
|
|
|
u64 curr;
|
|
|
|
} values[KVM_NR_SHARED_MSRS];
|
2009-09-07 15:12:18 +07:00
|
|
|
};
|
|
|
|
|
|
|
|
static struct kvm_shared_msrs_global __read_mostly shared_msrs_global;
|
2013-01-03 20:41:39 +07:00
|
|
|
static struct kvm_shared_msrs __percpu *shared_msrs;
|
2009-09-07 15:12:18 +07:00
|
|
|
|
2007-11-01 05:24:23 +07:00
|
|
|
struct kvm_stats_debugfs_item debugfs_entries[] = {
|
2007-11-18 21:24:12 +07:00
|
|
|
{ "pf_fixed", VCPU_STAT(pf_fixed) },
|
|
|
|
{ "pf_guest", VCPU_STAT(pf_guest) },
|
|
|
|
{ "tlb_flush", VCPU_STAT(tlb_flush) },
|
|
|
|
{ "invlpg", VCPU_STAT(invlpg) },
|
|
|
|
{ "exits", VCPU_STAT(exits) },
|
|
|
|
{ "io_exits", VCPU_STAT(io_exits) },
|
|
|
|
{ "mmio_exits", VCPU_STAT(mmio_exits) },
|
|
|
|
{ "signal_exits", VCPU_STAT(signal_exits) },
|
|
|
|
{ "irq_window", VCPU_STAT(irq_window_exits) },
|
2008-05-15 17:23:25 +07:00
|
|
|
{ "nmi_window", VCPU_STAT(nmi_window_exits) },
|
2007-11-18 21:24:12 +07:00
|
|
|
{ "halt_exits", VCPU_STAT(halt_exits) },
|
kvm: add halt_poll_ns module parameter
This patch introduces a new module parameter for the KVM module; when it
is present, KVM attempts a bit of polling on every HLT before scheduling
itself out via kvm_vcpu_block.
This parameter helps a lot for latency-bound workloads---in particular
I tested it with O_DSYNC writes with a battery-backed disk in the host.
In this case, writes are fast (because the data doesn't have to go all
the way to the platters) but they cannot be merged by either the host or
the guest. KVM's performance here is usually around 30% of bare metal,
or 50% if you use cache=directsync or cache=writethrough (these
parameters avoid that the guest sends pointless flush requests, and
at the same time they are not slow because of the battery-backed cache).
The bad performance happens because on every halt the host CPU decides
to halt itself too. When the interrupt comes, the vCPU thread is then
migrated to a new physical CPU, and in general the latency is horrible
because the vCPU thread has to be scheduled back in.
With this patch performance reaches 60-65% of bare metal and, more
important, 99% of what you get if you use idle=poll in the guest. This
means that the tunable gets rid of this particular bottleneck, and more
work can be done to improve performance in the kernel or QEMU.
Of course there is some price to pay; every time an otherwise idle vCPUs
is interrupted by an interrupt, it will poll unnecessarily and thus
impose a little load on the host. The above results were obtained with
a mostly random value of the parameter (500000), and the load was around
1.5-2.5% CPU usage on one of the host's core for each idle guest vCPU.
The patch also adds a new stat, /sys/kernel/debug/kvm/halt_successful_poll,
that can be used to tune the parameter. It counts how many HLT
instructions received an interrupt during the polling period; each
successful poll avoids that Linux schedules the VCPU thread out and back
in, and may also avoid a likely trip to C1 and back for the physical CPU.
While the VM is idle, a Linux 4 VCPU VM halts around 10 times per second.
Of these halts, almost all are failed polls. During the benchmark,
instead, basically all halts end within the polling period, except a more
or less constant stream of 50 per second coming from vCPUs that are not
running the benchmark. The wasted time is thus very low. Things may
be slightly different for Windows VMs, which have a ~10 ms timer tick.
The effect is also visible on Marcelo's recently-introduced latency
test for the TSC deadline timer. Though of course a non-RT kernel has
awful latency bounds, the latency of the timer is around 8000-10000 clock
cycles compared to 20000-120000 without setting halt_poll_ns. For the TSC
deadline timer, thus, the effect is both a smaller average latency and
a smaller variance.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2015-02-05 00:20:58 +07:00
|
|
|
{ "halt_successful_poll", VCPU_STAT(halt_successful_poll) },
|
2015-09-15 23:27:57 +07:00
|
|
|
{ "halt_attempted_poll", VCPU_STAT(halt_attempted_poll) },
|
2016-05-13 17:16:35 +07:00
|
|
|
{ "halt_poll_invalid", VCPU_STAT(halt_poll_invalid) },
|
2007-11-18 21:24:12 +07:00
|
|
|
{ "halt_wakeup", VCPU_STAT(halt_wakeup) },
|
2008-02-21 02:30:30 +07:00
|
|
|
{ "hypercalls", VCPU_STAT(hypercalls) },
|
2007-11-18 21:24:12 +07:00
|
|
|
{ "request_irq", VCPU_STAT(request_irq_exits) },
|
|
|
|
{ "irq_exits", VCPU_STAT(irq_exits) },
|
|
|
|
{ "host_state_reload", VCPU_STAT(host_state_reload) },
|
|
|
|
{ "efer_reload", VCPU_STAT(efer_reload) },
|
|
|
|
{ "fpu_reload", VCPU_STAT(fpu_reload) },
|
|
|
|
{ "insn_emulation", VCPU_STAT(insn_emulation) },
|
|
|
|
{ "insn_emulation_fail", VCPU_STAT(insn_emulation_fail) },
|
2008-09-01 19:57:51 +07:00
|
|
|
{ "irq_injections", VCPU_STAT(irq_injections) },
|
2008-09-26 14:30:55 +07:00
|
|
|
{ "nmi_injections", VCPU_STAT(nmi_injections) },
|
2007-11-18 21:37:07 +07:00
|
|
|
{ "mmu_shadow_zapped", VM_STAT(mmu_shadow_zapped) },
|
|
|
|
{ "mmu_pte_write", VM_STAT(mmu_pte_write) },
|
|
|
|
{ "mmu_pte_updated", VM_STAT(mmu_pte_updated) },
|
|
|
|
{ "mmu_pde_zapped", VM_STAT(mmu_pde_zapped) },
|
|
|
|
{ "mmu_flooded", VM_STAT(mmu_flooded) },
|
|
|
|
{ "mmu_recycled", VM_STAT(mmu_recycled) },
|
2007-12-19 00:47:18 +07:00
|
|
|
{ "mmu_cache_miss", VM_STAT(mmu_cache_miss) },
|
2008-09-23 23:18:39 +07:00
|
|
|
{ "mmu_unsync", VM_STAT(mmu_unsync) },
|
2007-11-21 04:01:14 +07:00
|
|
|
{ "remote_tlb_flush", VM_STAT(remote_tlb_flush) },
|
2008-02-23 21:44:30 +07:00
|
|
|
{ "largepages", VM_STAT(lpages) },
|
2007-11-01 05:24:23 +07:00
|
|
|
{ NULL }
|
|
|
|
};
|
|
|
|
|
2010-06-10 10:27:12 +07:00
|
|
|
u64 __read_mostly host_xcr0;
|
|
|
|
|
2012-09-20 12:43:17 +07:00
|
|
|
static int emulator_fix_hypercall(struct x86_emulate_ctxt *ctxt);
|
2011-04-20 19:47:13 +07:00
|
|
|
|
2010-10-14 16:22:46 +07:00
|
|
|
static inline void kvm_async_pf_hash_reset(struct kvm_vcpu *vcpu)
|
|
|
|
{
|
|
|
|
int i;
|
|
|
|
for (i = 0; i < roundup_pow_of_two(ASYNC_PF_PER_VCPU); i++)
|
|
|
|
vcpu->arch.apf.gfns[i] = ~0;
|
|
|
|
}
|
|
|
|
|
2009-09-07 15:12:18 +07:00
|
|
|
static void kvm_on_user_return(struct user_return_notifier *urn)
|
|
|
|
{
|
|
|
|
unsigned slot;
|
|
|
|
struct kvm_shared_msrs *locals
|
|
|
|
= container_of(urn, struct kvm_shared_msrs, urn);
|
2009-12-18 15:48:44 +07:00
|
|
|
struct kvm_shared_msr_values *values;
|
2009-09-07 15:12:18 +07:00
|
|
|
|
|
|
|
for (slot = 0; slot < shared_msrs_global.nr; ++slot) {
|
2009-12-18 15:48:44 +07:00
|
|
|
values = &locals->values[slot];
|
|
|
|
if (values->host != values->curr) {
|
|
|
|
wrmsrl(shared_msrs_global.msrs[slot], values->host);
|
|
|
|
values->curr = values->host;
|
2009-09-07 15:12:18 +07:00
|
|
|
}
|
|
|
|
}
|
|
|
|
locals->registered = false;
|
|
|
|
user_return_notifier_unregister(urn);
|
|
|
|
}
|
|
|
|
|
2009-12-18 15:48:44 +07:00
|
|
|
static void shared_msr_update(unsigned slot, u32 msr)
|
2009-09-07 15:12:18 +07:00
|
|
|
{
|
|
|
|
u64 value;
|
2013-01-03 20:41:39 +07:00
|
|
|
unsigned int cpu = smp_processor_id();
|
|
|
|
struct kvm_shared_msrs *smsr = per_cpu_ptr(shared_msrs, cpu);
|
2009-09-07 15:12:18 +07:00
|
|
|
|
2009-12-18 15:48:44 +07:00
|
|
|
/* only read, and nobody should modify it at this time,
|
|
|
|
* so don't need lock */
|
|
|
|
if (slot >= shared_msrs_global.nr) {
|
|
|
|
printk(KERN_ERR "kvm: invalid MSR slot!");
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
rdmsrl_safe(msr, &value);
|
|
|
|
smsr->values[slot].host = value;
|
|
|
|
smsr->values[slot].curr = value;
|
|
|
|
}
|
|
|
|
|
|
|
|
void kvm_define_shared_msr(unsigned slot, u32 msr)
|
|
|
|
{
|
2014-07-24 19:06:56 +07:00
|
|
|
BUG_ON(slot >= KVM_NR_SHARED_MSRS);
|
2015-07-29 16:06:34 +07:00
|
|
|
shared_msrs_global.msrs[slot] = msr;
|
2009-09-07 15:12:18 +07:00
|
|
|
if (slot >= shared_msrs_global.nr)
|
|
|
|
shared_msrs_global.nr = slot + 1;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(kvm_define_shared_msr);
|
|
|
|
|
|
|
|
static void kvm_shared_msr_cpu_online(void)
|
|
|
|
{
|
|
|
|
unsigned i;
|
|
|
|
|
|
|
|
for (i = 0; i < shared_msrs_global.nr; ++i)
|
2009-12-18 15:48:44 +07:00
|
|
|
shared_msr_update(i, shared_msrs_global.msrs[i]);
|
2009-09-07 15:12:18 +07:00
|
|
|
}
|
|
|
|
|
2014-08-28 01:16:44 +07:00
|
|
|
int kvm_set_shared_msr(unsigned slot, u64 value, u64 mask)
|
2009-09-07 15:12:18 +07:00
|
|
|
{
|
2013-01-03 20:41:39 +07:00
|
|
|
unsigned int cpu = smp_processor_id();
|
|
|
|
struct kvm_shared_msrs *smsr = per_cpu_ptr(shared_msrs, cpu);
|
2014-08-28 01:16:44 +07:00
|
|
|
int err;
|
2009-09-07 15:12:18 +07:00
|
|
|
|
2009-12-18 15:48:44 +07:00
|
|
|
if (((value ^ smsr->values[slot].curr) & mask) == 0)
|
2014-08-28 01:16:44 +07:00
|
|
|
return 0;
|
2009-12-18 15:48:44 +07:00
|
|
|
smsr->values[slot].curr = value;
|
2014-08-28 01:16:44 +07:00
|
|
|
err = wrmsrl_safe(shared_msrs_global.msrs[slot], value);
|
|
|
|
if (err)
|
|
|
|
return 1;
|
|
|
|
|
2009-09-07 15:12:18 +07:00
|
|
|
if (!smsr->registered) {
|
|
|
|
smsr->urn.on_user_return = kvm_on_user_return;
|
|
|
|
user_return_notifier_register(&smsr->urn);
|
|
|
|
smsr->registered = true;
|
|
|
|
}
|
2014-08-28 01:16:44 +07:00
|
|
|
return 0;
|
2009-09-07 15:12:18 +07:00
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(kvm_set_shared_msr);
|
|
|
|
|
2014-08-28 20:13:03 +07:00
|
|
|
static void drop_user_return_notifiers(void)
|
2009-11-28 19:18:47 +07:00
|
|
|
{
|
2013-01-03 20:41:39 +07:00
|
|
|
unsigned int cpu = smp_processor_id();
|
|
|
|
struct kvm_shared_msrs *smsr = per_cpu_ptr(shared_msrs, cpu);
|
2009-11-28 19:18:47 +07:00
|
|
|
|
|
|
|
if (smsr->registered)
|
|
|
|
kvm_on_user_return(&smsr->urn);
|
|
|
|
}
|
|
|
|
|
2007-10-29 22:09:10 +07:00
|
|
|
u64 kvm_get_apic_base(struct kvm_vcpu *vcpu)
|
|
|
|
{
|
2012-08-05 19:58:26 +07:00
|
|
|
return vcpu->arch.apic_base;
|
2007-10-29 22:09:10 +07:00
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(kvm_get_apic_base);
|
|
|
|
|
2014-01-24 22:48:44 +07:00
|
|
|
int kvm_set_apic_base(struct kvm_vcpu *vcpu, struct msr_data *msr_info)
|
|
|
|
{
|
|
|
|
u64 old_state = vcpu->arch.apic_base &
|
|
|
|
(MSR_IA32_APICBASE_ENABLE | X2APIC_ENABLE);
|
|
|
|
u64 new_state = msr_info->data &
|
|
|
|
(MSR_IA32_APICBASE_ENABLE | X2APIC_ENABLE);
|
|
|
|
u64 reserved_bits = ((~0ULL) << cpuid_maxphyaddr(vcpu)) |
|
|
|
|
0x2ff | (guest_cpuid_has_x2apic(vcpu) ? 0 : X2APIC_ENABLE);
|
|
|
|
|
|
|
|
if (!msr_info->host_initiated &&
|
|
|
|
((msr_info->data & reserved_bits) != 0 ||
|
|
|
|
new_state == X2APIC_ENABLE ||
|
|
|
|
(new_state == MSR_IA32_APICBASE_ENABLE &&
|
|
|
|
old_state == (MSR_IA32_APICBASE_ENABLE | X2APIC_ENABLE)) ||
|
|
|
|
(new_state == (MSR_IA32_APICBASE_ENABLE | X2APIC_ENABLE) &&
|
|
|
|
old_state == 0)))
|
|
|
|
return 1;
|
|
|
|
|
|
|
|
kvm_lapic_set_base(vcpu, msr_info->data);
|
|
|
|
return 0;
|
2007-10-29 22:09:10 +07:00
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(kvm_set_apic_base);
|
|
|
|
|
2014-05-02 05:44:37 +07:00
|
|
|
asmlinkage __visible void kvm_spurious_fault(void)
|
2013-04-06 02:20:30 +07:00
|
|
|
{
|
|
|
|
/* Fault while not rebooting. We want the trace. */
|
|
|
|
BUG();
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(kvm_spurious_fault);
|
|
|
|
|
2009-11-19 22:54:07 +07:00
|
|
|
#define EXCPT_BENIGN 0
|
|
|
|
#define EXCPT_CONTRIBUTORY 1
|
|
|
|
#define EXCPT_PF 2
|
|
|
|
|
|
|
|
static int exception_class(int vector)
|
|
|
|
{
|
|
|
|
switch (vector) {
|
|
|
|
case PF_VECTOR:
|
|
|
|
return EXCPT_PF;
|
|
|
|
case DE_VECTOR:
|
|
|
|
case TS_VECTOR:
|
|
|
|
case NP_VECTOR:
|
|
|
|
case SS_VECTOR:
|
|
|
|
case GP_VECTOR:
|
|
|
|
return EXCPT_CONTRIBUTORY;
|
|
|
|
default:
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
return EXCPT_BENIGN;
|
|
|
|
}
|
|
|
|
|
2014-07-24 18:51:24 +07:00
|
|
|
#define EXCPT_FAULT 0
|
|
|
|
#define EXCPT_TRAP 1
|
|
|
|
#define EXCPT_ABORT 2
|
|
|
|
#define EXCPT_INTERRUPT 3
|
|
|
|
|
|
|
|
static int exception_type(int vector)
|
|
|
|
{
|
|
|
|
unsigned int mask;
|
|
|
|
|
|
|
|
if (WARN_ON(vector > 31 || vector == NMI_VECTOR))
|
|
|
|
return EXCPT_INTERRUPT;
|
|
|
|
|
|
|
|
mask = 1 << vector;
|
|
|
|
|
|
|
|
/* #DB is trap, as instruction watchpoints are handled elsewhere */
|
|
|
|
if (mask & ((1 << DB_VECTOR) | (1 << BP_VECTOR) | (1 << OF_VECTOR)))
|
|
|
|
return EXCPT_TRAP;
|
|
|
|
|
|
|
|
if (mask & ((1 << DF_VECTOR) | (1 << MC_VECTOR)))
|
|
|
|
return EXCPT_ABORT;
|
|
|
|
|
|
|
|
/* Reserved exceptions will result in fault */
|
|
|
|
return EXCPT_FAULT;
|
|
|
|
}
|
|
|
|
|
2009-11-19 22:54:07 +07:00
|
|
|
static void kvm_multiple_exception(struct kvm_vcpu *vcpu,
|
2010-04-22 17:33:13 +07:00
|
|
|
unsigned nr, bool has_error, u32 error_code,
|
|
|
|
bool reinject)
|
2009-11-19 22:54:07 +07:00
|
|
|
{
|
|
|
|
u32 prev_nr;
|
|
|
|
int class1, class2;
|
|
|
|
|
2010-07-27 16:30:24 +07:00
|
|
|
kvm_make_request(KVM_REQ_EVENT, vcpu);
|
|
|
|
|
2009-11-19 22:54:07 +07:00
|
|
|
if (!vcpu->arch.exception.pending) {
|
|
|
|
queue:
|
2014-11-02 16:54:42 +07:00
|
|
|
if (has_error && !is_protmode(vcpu))
|
|
|
|
has_error = false;
|
2009-11-19 22:54:07 +07:00
|
|
|
vcpu->arch.exception.pending = true;
|
|
|
|
vcpu->arch.exception.has_error_code = has_error;
|
|
|
|
vcpu->arch.exception.nr = nr;
|
|
|
|
vcpu->arch.exception.error_code = error_code;
|
2010-05-05 21:04:41 +07:00
|
|
|
vcpu->arch.exception.reinject = reinject;
|
2009-11-19 22:54:07 +07:00
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* to check exception */
|
|
|
|
prev_nr = vcpu->arch.exception.nr;
|
|
|
|
if (prev_nr == DF_VECTOR) {
|
|
|
|
/* triple fault -> shutdown */
|
2010-05-10 16:34:53 +07:00
|
|
|
kvm_make_request(KVM_REQ_TRIPLE_FAULT, vcpu);
|
2009-11-19 22:54:07 +07:00
|
|
|
return;
|
|
|
|
}
|
|
|
|
class1 = exception_class(prev_nr);
|
|
|
|
class2 = exception_class(nr);
|
|
|
|
if ((class1 == EXCPT_CONTRIBUTORY && class2 == EXCPT_CONTRIBUTORY)
|
|
|
|
|| (class1 == EXCPT_PF && class2 != EXCPT_BENIGN)) {
|
|
|
|
/* generate double fault per SDM Table 5-5 */
|
|
|
|
vcpu->arch.exception.pending = true;
|
|
|
|
vcpu->arch.exception.has_error_code = true;
|
|
|
|
vcpu->arch.exception.nr = DF_VECTOR;
|
|
|
|
vcpu->arch.exception.error_code = 0;
|
|
|
|
} else
|
|
|
|
/* replace previous exception with a new one in a hope
|
|
|
|
that instruction re-execution will regenerate lost
|
|
|
|
exception */
|
|
|
|
goto queue;
|
|
|
|
}
|
|
|
|
|
2007-11-25 18:41:11 +07:00
|
|
|
void kvm_queue_exception(struct kvm_vcpu *vcpu, unsigned nr)
|
|
|
|
{
|
2010-04-22 17:33:13 +07:00
|
|
|
kvm_multiple_exception(vcpu, nr, false, 0, false);
|
2007-11-25 18:41:11 +07:00
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(kvm_queue_exception);
|
|
|
|
|
2010-04-22 17:33:13 +07:00
|
|
|
void kvm_requeue_exception(struct kvm_vcpu *vcpu, unsigned nr)
|
|
|
|
{
|
|
|
|
kvm_multiple_exception(vcpu, nr, false, 0, true);
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(kvm_requeue_exception);
|
|
|
|
|
2010-12-21 17:12:01 +07:00
|
|
|
void kvm_complete_insn_gp(struct kvm_vcpu *vcpu, int err)
|
2007-11-25 19:04:58 +07:00
|
|
|
{
|
2010-12-21 17:12:01 +07:00
|
|
|
if (err)
|
|
|
|
kvm_inject_gp(vcpu, 0);
|
|
|
|
else
|
|
|
|
kvm_x86_ops->skip_emulated_instruction(vcpu);
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(kvm_complete_insn_gp);
|
2010-09-10 22:30:46 +07:00
|
|
|
|
2010-11-29 21:12:30 +07:00
|
|
|
void kvm_inject_page_fault(struct kvm_vcpu *vcpu, struct x86_exception *fault)
|
2007-11-25 19:04:58 +07:00
|
|
|
{
|
|
|
|
++vcpu->stat.pf_guest;
|
2010-11-29 21:12:30 +07:00
|
|
|
vcpu->arch.cr2 = fault->address;
|
|
|
|
kvm_queue_exception_e(vcpu, PF_VECTOR, fault->error_code);
|
2007-11-25 19:04:58 +07:00
|
|
|
}
|
2011-05-26 03:06:59 +07:00
|
|
|
EXPORT_SYMBOL_GPL(kvm_inject_page_fault);
|
2007-11-25 19:04:58 +07:00
|
|
|
|
2014-09-05 00:46:15 +07:00
|
|
|
static bool kvm_propagate_fault(struct kvm_vcpu *vcpu, struct x86_exception *fault)
|
2010-09-10 22:30:55 +07:00
|
|
|
{
|
2010-11-29 21:12:30 +07:00
|
|
|
if (mmu_is_nested(vcpu) && !fault->nested_page_fault)
|
|
|
|
vcpu->arch.nested_mmu.inject_page_fault(vcpu, fault);
|
2010-09-10 22:30:55 +07:00
|
|
|
else
|
2010-11-29 21:12:30 +07:00
|
|
|
vcpu->arch.mmu.inject_page_fault(vcpu, fault);
|
2014-09-05 00:46:15 +07:00
|
|
|
|
|
|
|
return fault->nested_page_fault;
|
2010-09-10 22:30:55 +07:00
|
|
|
}
|
|
|
|
|
2008-05-15 08:52:48 +07:00
|
|
|
void kvm_inject_nmi(struct kvm_vcpu *vcpu)
|
|
|
|
{
|
2011-09-20 17:43:14 +07:00
|
|
|
atomic_inc(&vcpu->arch.nmi_queued);
|
|
|
|
kvm_make_request(KVM_REQ_NMI, vcpu);
|
2008-05-15 08:52:48 +07:00
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(kvm_inject_nmi);
|
|
|
|
|
2007-11-25 18:41:11 +07:00
|
|
|
void kvm_queue_exception_e(struct kvm_vcpu *vcpu, unsigned nr, u32 error_code)
|
|
|
|
{
|
2010-04-22 17:33:13 +07:00
|
|
|
kvm_multiple_exception(vcpu, nr, true, error_code, false);
|
2007-11-25 18:41:11 +07:00
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(kvm_queue_exception_e);
|
|
|
|
|
2010-04-22 17:33:13 +07:00
|
|
|
void kvm_requeue_exception_e(struct kvm_vcpu *vcpu, unsigned nr, u32 error_code)
|
|
|
|
{
|
|
|
|
kvm_multiple_exception(vcpu, nr, true, error_code, true);
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(kvm_requeue_exception_e);
|
|
|
|
|
2009-09-01 16:03:25 +07:00
|
|
|
/*
|
|
|
|
* Checks if cpl <= required_cpl; if true, return true. Otherwise queue
|
|
|
|
* a #GP and return false.
|
|
|
|
*/
|
|
|
|
bool kvm_require_cpl(struct kvm_vcpu *vcpu, int required_cpl)
|
2007-11-25 18:41:11 +07:00
|
|
|
{
|
2009-09-01 16:03:25 +07:00
|
|
|
if (kvm_x86_ops->get_cpl(vcpu) <= required_cpl)
|
|
|
|
return true;
|
|
|
|
kvm_queue_exception_e(vcpu, GP_VECTOR, 0);
|
|
|
|
return false;
|
2007-11-25 18:41:11 +07:00
|
|
|
}
|
2009-09-01 16:03:25 +07:00
|
|
|
EXPORT_SYMBOL_GPL(kvm_require_cpl);
|
2007-11-25 18:41:11 +07:00
|
|
|
|
2014-10-03 05:10:05 +07:00
|
|
|
bool kvm_require_dr(struct kvm_vcpu *vcpu, int dr)
|
|
|
|
{
|
|
|
|
if ((dr != 4 && dr != 5) || !kvm_read_cr4_bits(vcpu, X86_CR4_DE))
|
|
|
|
return true;
|
|
|
|
|
|
|
|
kvm_queue_exception(vcpu, UD_VECTOR);
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(kvm_require_dr);
|
|
|
|
|
2010-09-10 22:30:51 +07:00
|
|
|
/*
|
|
|
|
* This function will be used to read from the physical memory of the currently
|
2015-04-08 20:39:23 +07:00
|
|
|
* running guest. The difference to kvm_vcpu_read_guest_page is that this function
|
2010-09-10 22:30:51 +07:00
|
|
|
* can read from guest physical or from the guest's guest physical memory.
|
|
|
|
*/
|
|
|
|
int kvm_read_guest_page_mmu(struct kvm_vcpu *vcpu, struct kvm_mmu *mmu,
|
|
|
|
gfn_t ngfn, void *data, int offset, int len,
|
|
|
|
u32 access)
|
|
|
|
{
|
2014-09-02 18:23:06 +07:00
|
|
|
struct x86_exception exception;
|
2010-09-10 22:30:51 +07:00
|
|
|
gfn_t real_gfn;
|
|
|
|
gpa_t ngpa;
|
|
|
|
|
|
|
|
ngpa = gfn_to_gpa(ngfn);
|
2014-09-02 18:23:06 +07:00
|
|
|
real_gfn = mmu->translate_gpa(vcpu, ngpa, access, &exception);
|
2010-09-10 22:30:51 +07:00
|
|
|
if (real_gfn == UNMAPPED_GVA)
|
|
|
|
return -EFAULT;
|
|
|
|
|
|
|
|
real_gfn = gpa_to_gfn(real_gfn);
|
|
|
|
|
2015-04-08 20:39:23 +07:00
|
|
|
return kvm_vcpu_read_guest_page(vcpu, real_gfn, data, offset, len);
|
2010-09-10 22:30:51 +07:00
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(kvm_read_guest_page_mmu);
|
|
|
|
|
2015-01-19 21:33:39 +07:00
|
|
|
static int kvm_read_nested_guest_page(struct kvm_vcpu *vcpu, gfn_t gfn,
|
2010-09-10 22:30:53 +07:00
|
|
|
void *data, int offset, int len, u32 access)
|
|
|
|
{
|
|
|
|
return kvm_read_guest_page_mmu(vcpu, vcpu->arch.walk_mmu, gfn,
|
|
|
|
data, offset, len, access);
|
|
|
|
}
|
|
|
|
|
KVM: Portability: Move control register helper functions to x86.c
This patch moves the definitions of CR0_RESERVED_BITS,
CR4_RESERVED_BITS, and CR8_RESERVED_BITS along with the following
functions from kvm_main.c to x86.c:
set_cr0(), set_cr3(), set_cr4(), set_cr8(), get_cr8(), lmsw(),
load_pdptrs()
The static function wrapper inject_gp is duplicated in kvm_main.c and
x86.c for now, the version in kvm_main.c should disappear once the last
user of it is gone too.
The function load_pdptrs is no longer static, and now defined in x86.h
for the time being, until the last user of it is gone from kvm_main.c.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-29 22:09:35 +07:00
|
|
|
/*
|
|
|
|
* Load the pae pdptrs. Return true is they are all valid.
|
|
|
|
*/
|
2010-09-10 22:30:57 +07:00
|
|
|
int load_pdptrs(struct kvm_vcpu *vcpu, struct kvm_mmu *mmu, unsigned long cr3)
|
KVM: Portability: Move control register helper functions to x86.c
This patch moves the definitions of CR0_RESERVED_BITS,
CR4_RESERVED_BITS, and CR8_RESERVED_BITS along with the following
functions from kvm_main.c to x86.c:
set_cr0(), set_cr3(), set_cr4(), set_cr8(), get_cr8(), lmsw(),
load_pdptrs()
The static function wrapper inject_gp is duplicated in kvm_main.c and
x86.c for now, the version in kvm_main.c should disappear once the last
user of it is gone too.
The function load_pdptrs is no longer static, and now defined in x86.h
for the time being, until the last user of it is gone from kvm_main.c.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-29 22:09:35 +07:00
|
|
|
{
|
|
|
|
gfn_t pdpt_gfn = cr3 >> PAGE_SHIFT;
|
|
|
|
unsigned offset = ((cr3 & (PAGE_SIZE-1)) >> 5) << 2;
|
|
|
|
int i;
|
|
|
|
int ret;
|
2010-09-10 22:30:57 +07:00
|
|
|
u64 pdpte[ARRAY_SIZE(mmu->pdptrs)];
|
KVM: Portability: Move control register helper functions to x86.c
This patch moves the definitions of CR0_RESERVED_BITS,
CR4_RESERVED_BITS, and CR8_RESERVED_BITS along with the following
functions from kvm_main.c to x86.c:
set_cr0(), set_cr3(), set_cr4(), set_cr8(), get_cr8(), lmsw(),
load_pdptrs()
The static function wrapper inject_gp is duplicated in kvm_main.c and
x86.c for now, the version in kvm_main.c should disappear once the last
user of it is gone too.
The function load_pdptrs is no longer static, and now defined in x86.h
for the time being, until the last user of it is gone from kvm_main.c.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-29 22:09:35 +07:00
|
|
|
|
2010-09-10 22:30:57 +07:00
|
|
|
ret = kvm_read_guest_page_mmu(vcpu, mmu, pdpt_gfn, pdpte,
|
|
|
|
offset * sizeof(u64), sizeof(pdpte),
|
|
|
|
PFERR_USER_MASK|PFERR_WRITE_MASK);
|
KVM: Portability: Move control register helper functions to x86.c
This patch moves the definitions of CR0_RESERVED_BITS,
CR4_RESERVED_BITS, and CR8_RESERVED_BITS along with the following
functions from kvm_main.c to x86.c:
set_cr0(), set_cr3(), set_cr4(), set_cr8(), get_cr8(), lmsw(),
load_pdptrs()
The static function wrapper inject_gp is duplicated in kvm_main.c and
x86.c for now, the version in kvm_main.c should disappear once the last
user of it is gone too.
The function load_pdptrs is no longer static, and now defined in x86.h
for the time being, until the last user of it is gone from kvm_main.c.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-29 22:09:35 +07:00
|
|
|
if (ret < 0) {
|
|
|
|
ret = 0;
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
for (i = 0; i < ARRAY_SIZE(pdpte); ++i) {
|
2016-07-13 05:18:50 +07:00
|
|
|
if ((pdpte[i] & PT_PRESENT_MASK) &&
|
2015-08-05 11:04:21 +07:00
|
|
|
(pdpte[i] &
|
|
|
|
vcpu->arch.mmu.guest_rsvd_check.rsvd_bits_mask[0][2])) {
|
KVM: Portability: Move control register helper functions to x86.c
This patch moves the definitions of CR0_RESERVED_BITS,
CR4_RESERVED_BITS, and CR8_RESERVED_BITS along with the following
functions from kvm_main.c to x86.c:
set_cr0(), set_cr3(), set_cr4(), set_cr8(), get_cr8(), lmsw(),
load_pdptrs()
The static function wrapper inject_gp is duplicated in kvm_main.c and
x86.c for now, the version in kvm_main.c should disappear once the last
user of it is gone too.
The function load_pdptrs is no longer static, and now defined in x86.h
for the time being, until the last user of it is gone from kvm_main.c.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-29 22:09:35 +07:00
|
|
|
ret = 0;
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
ret = 1;
|
|
|
|
|
2010-09-10 22:30:57 +07:00
|
|
|
memcpy(mmu->pdptrs, pdpte, sizeof(mmu->pdptrs));
|
2009-06-01 02:58:47 +07:00
|
|
|
__set_bit(VCPU_EXREG_PDPTR,
|
|
|
|
(unsigned long *)&vcpu->arch.regs_avail);
|
|
|
|
__set_bit(VCPU_EXREG_PDPTR,
|
|
|
|
(unsigned long *)&vcpu->arch.regs_dirty);
|
KVM: Portability: Move control register helper functions to x86.c
This patch moves the definitions of CR0_RESERVED_BITS,
CR4_RESERVED_BITS, and CR8_RESERVED_BITS along with the following
functions from kvm_main.c to x86.c:
set_cr0(), set_cr3(), set_cr4(), set_cr8(), get_cr8(), lmsw(),
load_pdptrs()
The static function wrapper inject_gp is duplicated in kvm_main.c and
x86.c for now, the version in kvm_main.c should disappear once the last
user of it is gone too.
The function load_pdptrs is no longer static, and now defined in x86.h
for the time being, until the last user of it is gone from kvm_main.c.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-29 22:09:35 +07:00
|
|
|
out:
|
|
|
|
|
|
|
|
return ret;
|
|
|
|
}
|
2008-02-07 19:47:43 +07:00
|
|
|
EXPORT_SYMBOL_GPL(load_pdptrs);
|
KVM: Portability: Move control register helper functions to x86.c
This patch moves the definitions of CR0_RESERVED_BITS,
CR4_RESERVED_BITS, and CR8_RESERVED_BITS along with the following
functions from kvm_main.c to x86.c:
set_cr0(), set_cr3(), set_cr4(), set_cr8(), get_cr8(), lmsw(),
load_pdptrs()
The static function wrapper inject_gp is duplicated in kvm_main.c and
x86.c for now, the version in kvm_main.c should disappear once the last
user of it is gone too.
The function load_pdptrs is no longer static, and now defined in x86.h
for the time being, until the last user of it is gone from kvm_main.c.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-29 22:09:35 +07:00
|
|
|
|
2007-11-21 07:57:59 +07:00
|
|
|
static bool pdptrs_changed(struct kvm_vcpu *vcpu)
|
|
|
|
{
|
2010-09-10 22:30:57 +07:00
|
|
|
u64 pdpte[ARRAY_SIZE(vcpu->arch.walk_mmu->pdptrs)];
|
2007-11-21 07:57:59 +07:00
|
|
|
bool changed = true;
|
2010-09-10 22:30:53 +07:00
|
|
|
int offset;
|
|
|
|
gfn_t gfn;
|
2007-11-21 07:57:59 +07:00
|
|
|
int r;
|
|
|
|
|
|
|
|
if (is_long_mode(vcpu) || !is_pae(vcpu))
|
|
|
|
return false;
|
|
|
|
|
2009-06-01 02:58:47 +07:00
|
|
|
if (!test_bit(VCPU_EXREG_PDPTR,
|
|
|
|
(unsigned long *)&vcpu->arch.regs_avail))
|
|
|
|
return true;
|
|
|
|
|
2010-12-05 22:30:00 +07:00
|
|
|
gfn = (kvm_read_cr3(vcpu) & ~31u) >> PAGE_SHIFT;
|
|
|
|
offset = (kvm_read_cr3(vcpu) & ~31u) & (PAGE_SIZE - 1);
|
2010-09-10 22:30:53 +07:00
|
|
|
r = kvm_read_nested_guest_page(vcpu, gfn, pdpte, offset, sizeof(pdpte),
|
|
|
|
PFERR_USER_MASK | PFERR_WRITE_MASK);
|
2007-11-21 07:57:59 +07:00
|
|
|
if (r < 0)
|
|
|
|
goto out;
|
2010-09-10 22:30:57 +07:00
|
|
|
changed = memcmp(pdpte, vcpu->arch.walk_mmu->pdptrs, sizeof(pdpte)) != 0;
|
2007-11-21 07:57:59 +07:00
|
|
|
out:
|
|
|
|
|
|
|
|
return changed;
|
|
|
|
}
|
|
|
|
|
2010-06-10 21:02:14 +07:00
|
|
|
int kvm_set_cr0(struct kvm_vcpu *vcpu, unsigned long cr0)
|
KVM: Portability: Move control register helper functions to x86.c
This patch moves the definitions of CR0_RESERVED_BITS,
CR4_RESERVED_BITS, and CR8_RESERVED_BITS along with the following
functions from kvm_main.c to x86.c:
set_cr0(), set_cr3(), set_cr4(), set_cr8(), get_cr8(), lmsw(),
load_pdptrs()
The static function wrapper inject_gp is duplicated in kvm_main.c and
x86.c for now, the version in kvm_main.c should disappear once the last
user of it is gone too.
The function load_pdptrs is no longer static, and now defined in x86.h
for the time being, until the last user of it is gone from kvm_main.c.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-29 22:09:35 +07:00
|
|
|
{
|
2010-05-12 15:40:42 +07:00
|
|
|
unsigned long old_cr0 = kvm_read_cr0(vcpu);
|
2015-05-13 13:42:28 +07:00
|
|
|
unsigned long update_bits = X86_CR0_PG | X86_CR0_WP;
|
2010-05-12 15:40:42 +07:00
|
|
|
|
2010-01-07 00:10:22 +07:00
|
|
|
cr0 |= X86_CR0_ET;
|
|
|
|
|
2010-01-21 20:28:46 +07:00
|
|
|
#ifdef CONFIG_X86_64
|
2010-04-28 23:15:31 +07:00
|
|
|
if (cr0 & 0xffffffff00000000UL)
|
|
|
|
return 1;
|
2010-01-21 20:28:46 +07:00
|
|
|
#endif
|
|
|
|
|
|
|
|
cr0 &= ~CR0_RESERVED_BITS;
|
KVM: Portability: Move control register helper functions to x86.c
This patch moves the definitions of CR0_RESERVED_BITS,
CR4_RESERVED_BITS, and CR8_RESERVED_BITS along with the following
functions from kvm_main.c to x86.c:
set_cr0(), set_cr3(), set_cr4(), set_cr8(), get_cr8(), lmsw(),
load_pdptrs()
The static function wrapper inject_gp is duplicated in kvm_main.c and
x86.c for now, the version in kvm_main.c should disappear once the last
user of it is gone too.
The function load_pdptrs is no longer static, and now defined in x86.h
for the time being, until the last user of it is gone from kvm_main.c.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-29 22:09:35 +07:00
|
|
|
|
2010-04-28 23:15:31 +07:00
|
|
|
if ((cr0 & X86_CR0_NW) && !(cr0 & X86_CR0_CD))
|
|
|
|
return 1;
|
KVM: Portability: Move control register helper functions to x86.c
This patch moves the definitions of CR0_RESERVED_BITS,
CR4_RESERVED_BITS, and CR8_RESERVED_BITS along with the following
functions from kvm_main.c to x86.c:
set_cr0(), set_cr3(), set_cr4(), set_cr8(), get_cr8(), lmsw(),
load_pdptrs()
The static function wrapper inject_gp is duplicated in kvm_main.c and
x86.c for now, the version in kvm_main.c should disappear once the last
user of it is gone too.
The function load_pdptrs is no longer static, and now defined in x86.h
for the time being, until the last user of it is gone from kvm_main.c.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-29 22:09:35 +07:00
|
|
|
|
2010-04-28 23:15:31 +07:00
|
|
|
if ((cr0 & X86_CR0_PG) && !(cr0 & X86_CR0_PE))
|
|
|
|
return 1;
|
KVM: Portability: Move control register helper functions to x86.c
This patch moves the definitions of CR0_RESERVED_BITS,
CR4_RESERVED_BITS, and CR8_RESERVED_BITS along with the following
functions from kvm_main.c to x86.c:
set_cr0(), set_cr3(), set_cr4(), set_cr8(), get_cr8(), lmsw(),
load_pdptrs()
The static function wrapper inject_gp is duplicated in kvm_main.c and
x86.c for now, the version in kvm_main.c should disappear once the last
user of it is gone too.
The function load_pdptrs is no longer static, and now defined in x86.h
for the time being, until the last user of it is gone from kvm_main.c.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-29 22:09:35 +07:00
|
|
|
|
|
|
|
if (!is_paging(vcpu) && (cr0 & X86_CR0_PG)) {
|
|
|
|
#ifdef CONFIG_X86_64
|
2010-01-21 20:31:50 +07:00
|
|
|
if ((vcpu->arch.efer & EFER_LME)) {
|
KVM: Portability: Move control register helper functions to x86.c
This patch moves the definitions of CR0_RESERVED_BITS,
CR4_RESERVED_BITS, and CR8_RESERVED_BITS along with the following
functions from kvm_main.c to x86.c:
set_cr0(), set_cr3(), set_cr4(), set_cr8(), get_cr8(), lmsw(),
load_pdptrs()
The static function wrapper inject_gp is duplicated in kvm_main.c and
x86.c for now, the version in kvm_main.c should disappear once the last
user of it is gone too.
The function load_pdptrs is no longer static, and now defined in x86.h
for the time being, until the last user of it is gone from kvm_main.c.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-29 22:09:35 +07:00
|
|
|
int cs_db, cs_l;
|
|
|
|
|
2010-04-28 23:15:31 +07:00
|
|
|
if (!is_pae(vcpu))
|
|
|
|
return 1;
|
KVM: Portability: Move control register helper functions to x86.c
This patch moves the definitions of CR0_RESERVED_BITS,
CR4_RESERVED_BITS, and CR8_RESERVED_BITS along with the following
functions from kvm_main.c to x86.c:
set_cr0(), set_cr3(), set_cr4(), set_cr8(), get_cr8(), lmsw(),
load_pdptrs()
The static function wrapper inject_gp is duplicated in kvm_main.c and
x86.c for now, the version in kvm_main.c should disappear once the last
user of it is gone too.
The function load_pdptrs is no longer static, and now defined in x86.h
for the time being, until the last user of it is gone from kvm_main.c.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-29 22:09:35 +07:00
|
|
|
kvm_x86_ops->get_cs_db_l_bits(vcpu, &cs_db, &cs_l);
|
2010-04-28 23:15:31 +07:00
|
|
|
if (cs_l)
|
|
|
|
return 1;
|
KVM: Portability: Move control register helper functions to x86.c
This patch moves the definitions of CR0_RESERVED_BITS,
CR4_RESERVED_BITS, and CR8_RESERVED_BITS along with the following
functions from kvm_main.c to x86.c:
set_cr0(), set_cr3(), set_cr4(), set_cr8(), get_cr8(), lmsw(),
load_pdptrs()
The static function wrapper inject_gp is duplicated in kvm_main.c and
x86.c for now, the version in kvm_main.c should disappear once the last
user of it is gone too.
The function load_pdptrs is no longer static, and now defined in x86.h
for the time being, until the last user of it is gone from kvm_main.c.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-29 22:09:35 +07:00
|
|
|
} else
|
|
|
|
#endif
|
2010-09-10 22:30:57 +07:00
|
|
|
if (is_pae(vcpu) && !load_pdptrs(vcpu, vcpu->arch.walk_mmu,
|
2010-12-05 22:30:00 +07:00
|
|
|
kvm_read_cr3(vcpu)))
|
2010-04-28 23:15:31 +07:00
|
|
|
return 1;
|
KVM: Portability: Move control register helper functions to x86.c
This patch moves the definitions of CR0_RESERVED_BITS,
CR4_RESERVED_BITS, and CR8_RESERVED_BITS along with the following
functions from kvm_main.c to x86.c:
set_cr0(), set_cr3(), set_cr4(), set_cr8(), get_cr8(), lmsw(),
load_pdptrs()
The static function wrapper inject_gp is duplicated in kvm_main.c and
x86.c for now, the version in kvm_main.c should disappear once the last
user of it is gone too.
The function load_pdptrs is no longer static, and now defined in x86.h
for the time being, until the last user of it is gone from kvm_main.c.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-29 22:09:35 +07:00
|
|
|
}
|
|
|
|
|
2012-07-02 08:18:48 +07:00
|
|
|
if (!(cr0 & X86_CR0_PG) && kvm_read_cr4_bits(vcpu, X86_CR4_PCIDE))
|
|
|
|
return 1;
|
|
|
|
|
KVM: Portability: Move control register helper functions to x86.c
This patch moves the definitions of CR0_RESERVED_BITS,
CR4_RESERVED_BITS, and CR8_RESERVED_BITS along with the following
functions from kvm_main.c to x86.c:
set_cr0(), set_cr3(), set_cr4(), set_cr8(), get_cr8(), lmsw(),
load_pdptrs()
The static function wrapper inject_gp is duplicated in kvm_main.c and
x86.c for now, the version in kvm_main.c should disappear once the last
user of it is gone too.
The function load_pdptrs is no longer static, and now defined in x86.h
for the time being, until the last user of it is gone from kvm_main.c.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-29 22:09:35 +07:00
|
|
|
kvm_x86_ops->set_cr0(vcpu, cr0);
|
|
|
|
|
2011-02-21 10:21:30 +07:00
|
|
|
if ((cr0 ^ old_cr0) & X86_CR0_PG) {
|
2010-11-12 13:47:01 +07:00
|
|
|
kvm_clear_async_pf_completion_queue(vcpu);
|
2011-02-21 10:21:30 +07:00
|
|
|
kvm_async_pf_hash_reset(vcpu);
|
|
|
|
}
|
2010-11-12 13:47:01 +07:00
|
|
|
|
2010-05-12 15:40:42 +07:00
|
|
|
if ((cr0 ^ old_cr0) & update_bits)
|
|
|
|
kvm_mmu_reset_context(vcpu);
|
2015-06-15 15:55:21 +07:00
|
|
|
|
2015-11-04 18:54:41 +07:00
|
|
|
if (((cr0 ^ old_cr0) & X86_CR0_CD) &&
|
|
|
|
kvm_arch_has_noncoherent_dma(vcpu->kvm) &&
|
|
|
|
!kvm_check_has_quirk(vcpu->kvm, KVM_X86_QUIRK_CD_NW_CLEARED))
|
2015-06-15 15:55:21 +07:00
|
|
|
kvm_zap_gfn_range(vcpu->kvm, 0, ~0ULL);
|
|
|
|
|
2010-04-28 23:15:31 +07:00
|
|
|
return 0;
|
|
|
|
}
|
2008-02-24 16:20:43 +07:00
|
|
|
EXPORT_SYMBOL_GPL(kvm_set_cr0);
|
KVM: Portability: Move control register helper functions to x86.c
This patch moves the definitions of CR0_RESERVED_BITS,
CR4_RESERVED_BITS, and CR8_RESERVED_BITS along with the following
functions from kvm_main.c to x86.c:
set_cr0(), set_cr3(), set_cr4(), set_cr8(), get_cr8(), lmsw(),
load_pdptrs()
The static function wrapper inject_gp is duplicated in kvm_main.c and
x86.c for now, the version in kvm_main.c should disappear once the last
user of it is gone too.
The function load_pdptrs is no longer static, and now defined in x86.h
for the time being, until the last user of it is gone from kvm_main.c.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-29 22:09:35 +07:00
|
|
|
|
2008-02-24 16:20:43 +07:00
|
|
|
void kvm_lmsw(struct kvm_vcpu *vcpu, unsigned long msw)
|
KVM: Portability: Move control register helper functions to x86.c
This patch moves the definitions of CR0_RESERVED_BITS,
CR4_RESERVED_BITS, and CR8_RESERVED_BITS along with the following
functions from kvm_main.c to x86.c:
set_cr0(), set_cr3(), set_cr4(), set_cr8(), get_cr8(), lmsw(),
load_pdptrs()
The static function wrapper inject_gp is duplicated in kvm_main.c and
x86.c for now, the version in kvm_main.c should disappear once the last
user of it is gone too.
The function load_pdptrs is no longer static, and now defined in x86.h
for the time being, until the last user of it is gone from kvm_main.c.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-29 22:09:35 +07:00
|
|
|
{
|
2010-06-10 21:02:14 +07:00
|
|
|
(void)kvm_set_cr0(vcpu, kvm_read_cr0_bits(vcpu, ~0x0eul) | (msw & 0x0f));
|
KVM: Portability: Move control register helper functions to x86.c
This patch moves the definitions of CR0_RESERVED_BITS,
CR4_RESERVED_BITS, and CR8_RESERVED_BITS along with the following
functions from kvm_main.c to x86.c:
set_cr0(), set_cr3(), set_cr4(), set_cr8(), get_cr8(), lmsw(),
load_pdptrs()
The static function wrapper inject_gp is duplicated in kvm_main.c and
x86.c for now, the version in kvm_main.c should disappear once the last
user of it is gone too.
The function load_pdptrs is no longer static, and now defined in x86.h
for the time being, until the last user of it is gone from kvm_main.c.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-29 22:09:35 +07:00
|
|
|
}
|
2008-02-24 16:20:43 +07:00
|
|
|
EXPORT_SYMBOL_GPL(kvm_lmsw);
|
KVM: Portability: Move control register helper functions to x86.c
This patch moves the definitions of CR0_RESERVED_BITS,
CR4_RESERVED_BITS, and CR8_RESERVED_BITS along with the following
functions from kvm_main.c to x86.c:
set_cr0(), set_cr3(), set_cr4(), set_cr8(), get_cr8(), lmsw(),
load_pdptrs()
The static function wrapper inject_gp is duplicated in kvm_main.c and
x86.c for now, the version in kvm_main.c should disappear once the last
user of it is gone too.
The function load_pdptrs is no longer static, and now defined in x86.h
for the time being, until the last user of it is gone from kvm_main.c.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-29 22:09:35 +07:00
|
|
|
|
2013-04-16 09:30:13 +07:00
|
|
|
static void kvm_load_guest_xcr0(struct kvm_vcpu *vcpu)
|
|
|
|
{
|
|
|
|
if (kvm_read_cr4_bits(vcpu, X86_CR4_OSXSAVE) &&
|
|
|
|
!vcpu->guest_xcr0_loaded) {
|
|
|
|
/* kvm_set_xcr() also depends on this */
|
|
|
|
xsetbv(XCR_XFEATURE_ENABLED_MASK, vcpu->arch.xcr0);
|
|
|
|
vcpu->guest_xcr0_loaded = 1;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static void kvm_put_guest_xcr0(struct kvm_vcpu *vcpu)
|
|
|
|
{
|
|
|
|
if (vcpu->guest_xcr0_loaded) {
|
|
|
|
if (vcpu->arch.xcr0 != host_xcr0)
|
|
|
|
xsetbv(XCR_XFEATURE_ENABLED_MASK, host_xcr0);
|
|
|
|
vcpu->guest_xcr0_loaded = 0;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2015-01-19 21:33:39 +07:00
|
|
|
static int __kvm_set_xcr(struct kvm_vcpu *vcpu, u32 index, u64 xcr)
|
2010-06-10 10:27:12 +07:00
|
|
|
{
|
2014-02-22 00:39:02 +07:00
|
|
|
u64 xcr0 = xcr;
|
|
|
|
u64 old_xcr0 = vcpu->arch.xcr0;
|
2013-10-17 21:50:46 +07:00
|
|
|
u64 valid_bits;
|
2010-06-10 10:27:12 +07:00
|
|
|
|
|
|
|
/* Only support XCR_XFEATURE_ENABLED_MASK(xcr0) now */
|
|
|
|
if (index != XCR_XFEATURE_ENABLED_MASK)
|
|
|
|
return 1;
|
2015-09-03 06:31:26 +07:00
|
|
|
if (!(xcr0 & XFEATURE_MASK_FP))
|
2010-06-10 10:27:12 +07:00
|
|
|
return 1;
|
2015-09-03 06:31:26 +07:00
|
|
|
if ((xcr0 & XFEATURE_MASK_YMM) && !(xcr0 & XFEATURE_MASK_SSE))
|
2010-06-10 10:27:12 +07:00
|
|
|
return 1;
|
2013-10-17 21:50:46 +07:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Do not allow the guest to set bits that we do not support
|
|
|
|
* saving. However, xcr0 bit 0 is always set, even if the
|
|
|
|
* emulated CPU does not support XSAVE (see fx_init).
|
|
|
|
*/
|
2015-09-03 06:31:26 +07:00
|
|
|
valid_bits = vcpu->arch.guest_supported_xcr0 | XFEATURE_MASK_FP;
|
2013-10-17 21:50:46 +07:00
|
|
|
if (xcr0 & ~valid_bits)
|
2010-06-10 10:27:12 +07:00
|
|
|
return 1;
|
2013-10-17 21:50:46 +07:00
|
|
|
|
2015-09-03 06:31:26 +07:00
|
|
|
if ((!(xcr0 & XFEATURE_MASK_BNDREGS)) !=
|
|
|
|
(!(xcr0 & XFEATURE_MASK_BNDCSR)))
|
2014-02-24 17:58:09 +07:00
|
|
|
return 1;
|
|
|
|
|
2015-09-03 06:31:26 +07:00
|
|
|
if (xcr0 & XFEATURE_MASK_AVX512) {
|
|
|
|
if (!(xcr0 & XFEATURE_MASK_YMM))
|
2014-10-22 16:35:24 +07:00
|
|
|
return 1;
|
2015-09-03 06:31:26 +07:00
|
|
|
if ((xcr0 & XFEATURE_MASK_AVX512) != XFEATURE_MASK_AVX512)
|
2014-10-22 16:35:24 +07:00
|
|
|
return 1;
|
|
|
|
}
|
2010-06-10 10:27:12 +07:00
|
|
|
vcpu->arch.xcr0 = xcr0;
|
2014-02-22 00:39:02 +07:00
|
|
|
|
2015-09-03 06:31:26 +07:00
|
|
|
if ((xcr0 ^ old_xcr0) & XFEATURE_MASK_EXTEND)
|
2014-02-22 00:39:02 +07:00
|
|
|
kvm_update_cpuid(vcpu);
|
2010-06-10 10:27:12 +07:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
int kvm_set_xcr(struct kvm_vcpu *vcpu, u32 index, u64 xcr)
|
|
|
|
{
|
2013-06-14 14:36:13 +07:00
|
|
|
if (kvm_x86_ops->get_cpl(vcpu) != 0 ||
|
|
|
|
__kvm_set_xcr(vcpu, index, xcr)) {
|
2010-06-10 10:27:12 +07:00
|
|
|
kvm_inject_gp(vcpu, 0);
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(kvm_set_xcr);
|
|
|
|
|
2010-06-10 21:02:15 +07:00
|
|
|
int kvm_set_cr4(struct kvm_vcpu *vcpu, unsigned long cr4)
|
KVM: Portability: Move control register helper functions to x86.c
This patch moves the definitions of CR0_RESERVED_BITS,
CR4_RESERVED_BITS, and CR8_RESERVED_BITS along with the following
functions from kvm_main.c to x86.c:
set_cr0(), set_cr3(), set_cr4(), set_cr8(), get_cr8(), lmsw(),
load_pdptrs()
The static function wrapper inject_gp is duplicated in kvm_main.c and
x86.c for now, the version in kvm_main.c should disappear once the last
user of it is gone too.
The function load_pdptrs is no longer static, and now defined in x86.h
for the time being, until the last user of it is gone from kvm_main.c.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-29 22:09:35 +07:00
|
|
|
{
|
2009-12-07 17:16:48 +07:00
|
|
|
unsigned long old_cr4 = kvm_read_cr4(vcpu);
|
2015-05-11 21:55:21 +07:00
|
|
|
unsigned long pdptr_bits = X86_CR4_PGE | X86_CR4_PSE | X86_CR4_PAE |
|
2016-03-22 15:51:21 +07:00
|
|
|
X86_CR4_SMEP | X86_CR4_SMAP | X86_CR4_PKE;
|
2015-05-11 21:55:21 +07:00
|
|
|
|
2010-04-28 23:15:31 +07:00
|
|
|
if (cr4 & CR4_RESERVED_BITS)
|
|
|
|
return 1;
|
KVM: Portability: Move control register helper functions to x86.c
This patch moves the definitions of CR0_RESERVED_BITS,
CR4_RESERVED_BITS, and CR8_RESERVED_BITS along with the following
functions from kvm_main.c to x86.c:
set_cr0(), set_cr3(), set_cr4(), set_cr8(), get_cr8(), lmsw(),
load_pdptrs()
The static function wrapper inject_gp is duplicated in kvm_main.c and
x86.c for now, the version in kvm_main.c should disappear once the last
user of it is gone too.
The function load_pdptrs is no longer static, and now defined in x86.h
for the time being, until the last user of it is gone from kvm_main.c.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-29 22:09:35 +07:00
|
|
|
|
2010-06-10 10:27:12 +07:00
|
|
|
if (!guest_cpuid_has_xsave(vcpu) && (cr4 & X86_CR4_OSXSAVE))
|
|
|
|
return 1;
|
|
|
|
|
2011-06-03 10:13:42 +07:00
|
|
|
if (!guest_cpuid_has_smep(vcpu) && (cr4 & X86_CR4_SMEP))
|
|
|
|
return 1;
|
|
|
|
|
2014-04-01 16:46:34 +07:00
|
|
|
if (!guest_cpuid_has_smap(vcpu) && (cr4 & X86_CR4_SMAP))
|
|
|
|
return 1;
|
|
|
|
|
2013-04-28 06:37:47 +07:00
|
|
|
if (!guest_cpuid_has_fsgsbase(vcpu) && (cr4 & X86_CR4_FSGSBASE))
|
2011-06-14 19:10:18 +07:00
|
|
|
return 1;
|
|
|
|
|
2016-03-22 15:51:21 +07:00
|
|
|
if (!guest_cpuid_has_pku(vcpu) && (cr4 & X86_CR4_PKE))
|
|
|
|
return 1;
|
|
|
|
|
KVM: Portability: Move control register helper functions to x86.c
This patch moves the definitions of CR0_RESERVED_BITS,
CR4_RESERVED_BITS, and CR8_RESERVED_BITS along with the following
functions from kvm_main.c to x86.c:
set_cr0(), set_cr3(), set_cr4(), set_cr8(), get_cr8(), lmsw(),
load_pdptrs()
The static function wrapper inject_gp is duplicated in kvm_main.c and
x86.c for now, the version in kvm_main.c should disappear once the last
user of it is gone too.
The function load_pdptrs is no longer static, and now defined in x86.h
for the time being, until the last user of it is gone from kvm_main.c.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-29 22:09:35 +07:00
|
|
|
if (is_long_mode(vcpu)) {
|
2010-04-28 23:15:31 +07:00
|
|
|
if (!(cr4 & X86_CR4_PAE))
|
|
|
|
return 1;
|
2009-05-25 02:19:00 +07:00
|
|
|
} else if (is_paging(vcpu) && (cr4 & X86_CR4_PAE)
|
|
|
|
&& ((cr4 ^ old_cr4) & pdptr_bits)
|
2010-12-05 22:30:00 +07:00
|
|
|
&& !load_pdptrs(vcpu, vcpu->arch.walk_mmu,
|
|
|
|
kvm_read_cr3(vcpu)))
|
2010-04-28 23:15:31 +07:00
|
|
|
return 1;
|
|
|
|
|
2012-07-02 08:18:48 +07:00
|
|
|
if ((cr4 & X86_CR4_PCIDE) && !(old_cr4 & X86_CR4_PCIDE)) {
|
|
|
|
if (!guest_cpuid_has_pcid(vcpu))
|
|
|
|
return 1;
|
|
|
|
|
|
|
|
/* PCID can not be enabled when cr3[11:0]!=000H or EFER.LMA=0 */
|
|
|
|
if ((kvm_read_cr3(vcpu) & X86_CR3_PCID_MASK) || !is_long_mode(vcpu))
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
|
2011-05-26 03:03:24 +07:00
|
|
|
if (kvm_x86_ops->set_cr4(vcpu, cr4))
|
2010-04-28 23:15:31 +07:00
|
|
|
return 1;
|
KVM: Portability: Move control register helper functions to x86.c
This patch moves the definitions of CR0_RESERVED_BITS,
CR4_RESERVED_BITS, and CR8_RESERVED_BITS along with the following
functions from kvm_main.c to x86.c:
set_cr0(), set_cr3(), set_cr4(), set_cr8(), get_cr8(), lmsw(),
load_pdptrs()
The static function wrapper inject_gp is duplicated in kvm_main.c and
x86.c for now, the version in kvm_main.c should disappear once the last
user of it is gone too.
The function load_pdptrs is no longer static, and now defined in x86.h
for the time being, until the last user of it is gone from kvm_main.c.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-29 22:09:35 +07:00
|
|
|
|
2012-07-02 08:18:48 +07:00
|
|
|
if (((cr4 ^ old_cr4) & pdptr_bits) ||
|
|
|
|
(!(cr4 & X86_CR4_PCIDE) && (old_cr4 & X86_CR4_PCIDE)))
|
2010-05-12 15:40:42 +07:00
|
|
|
kvm_mmu_reset_context(vcpu);
|
2010-04-28 23:15:31 +07:00
|
|
|
|
2016-03-22 15:51:21 +07:00
|
|
|
if ((cr4 ^ old_cr4) & (X86_CR4_OSXSAVE | X86_CR4_PKE))
|
2011-11-23 21:30:32 +07:00
|
|
|
kvm_update_cpuid(vcpu);
|
2010-06-10 10:27:12 +07:00
|
|
|
|
2010-04-28 23:15:31 +07:00
|
|
|
return 0;
|
|
|
|
}
|
2008-02-24 16:20:43 +07:00
|
|
|
EXPORT_SYMBOL_GPL(kvm_set_cr4);
|
KVM: Portability: Move control register helper functions to x86.c
This patch moves the definitions of CR0_RESERVED_BITS,
CR4_RESERVED_BITS, and CR8_RESERVED_BITS along with the following
functions from kvm_main.c to x86.c:
set_cr0(), set_cr3(), set_cr4(), set_cr8(), get_cr8(), lmsw(),
load_pdptrs()
The static function wrapper inject_gp is duplicated in kvm_main.c and
x86.c for now, the version in kvm_main.c should disappear once the last
user of it is gone too.
The function load_pdptrs is no longer static, and now defined in x86.h
for the time being, until the last user of it is gone from kvm_main.c.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-29 22:09:35 +07:00
|
|
|
|
2010-06-10 21:02:16 +07:00
|
|
|
int kvm_set_cr3(struct kvm_vcpu *vcpu, unsigned long cr3)
|
KVM: Portability: Move control register helper functions to x86.c
This patch moves the definitions of CR0_RESERVED_BITS,
CR4_RESERVED_BITS, and CR8_RESERVED_BITS along with the following
functions from kvm_main.c to x86.c:
set_cr0(), set_cr3(), set_cr4(), set_cr8(), get_cr8(), lmsw(),
load_pdptrs()
The static function wrapper inject_gp is duplicated in kvm_main.c and
x86.c for now, the version in kvm_main.c should disappear once the last
user of it is gone too.
The function load_pdptrs is no longer static, and now defined in x86.h
for the time being, until the last user of it is gone from kvm_main.c.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-29 22:09:35 +07:00
|
|
|
{
|
2014-11-10 19:53:25 +07:00
|
|
|
#ifdef CONFIG_X86_64
|
2014-11-02 16:54:52 +07:00
|
|
|
cr3 &= ~CR3_PCID_INVD;
|
2014-11-10 19:53:25 +07:00
|
|
|
#endif
|
2014-11-02 16:54:52 +07:00
|
|
|
|
2010-12-05 22:30:00 +07:00
|
|
|
if (cr3 == kvm_read_cr3(vcpu) && !pdptrs_changed(vcpu)) {
|
2008-09-23 23:18:34 +07:00
|
|
|
kvm_mmu_sync_roots(vcpu);
|
2014-09-18 23:38:37 +07:00
|
|
|
kvm_make_request(KVM_REQ_TLB_FLUSH, vcpu);
|
2010-04-28 23:15:31 +07:00
|
|
|
return 0;
|
2007-11-21 07:57:59 +07:00
|
|
|
}
|
|
|
|
|
KVM: Portability: Move control register helper functions to x86.c
This patch moves the definitions of CR0_RESERVED_BITS,
CR4_RESERVED_BITS, and CR8_RESERVED_BITS along with the following
functions from kvm_main.c to x86.c:
set_cr0(), set_cr3(), set_cr4(), set_cr8(), get_cr8(), lmsw(),
load_pdptrs()
The static function wrapper inject_gp is duplicated in kvm_main.c and
x86.c for now, the version in kvm_main.c should disappear once the last
user of it is gone too.
The function load_pdptrs is no longer static, and now defined in x86.h
for the time being, until the last user of it is gone from kvm_main.c.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-29 22:09:35 +07:00
|
|
|
if (is_long_mode(vcpu)) {
|
2014-05-10 14:24:34 +07:00
|
|
|
if (cr3 & CR3_L_MODE_RESERVED_BITS)
|
|
|
|
return 1;
|
|
|
|
} else if (is_pae(vcpu) && is_paging(vcpu) &&
|
|
|
|
!load_pdptrs(vcpu, vcpu->arch.walk_mmu, cr3))
|
2014-04-18 07:35:09 +07:00
|
|
|
return 1;
|
KVM: Portability: Move control register helper functions to x86.c
This patch moves the definitions of CR0_RESERVED_BITS,
CR4_RESERVED_BITS, and CR8_RESERVED_BITS along with the following
functions from kvm_main.c to x86.c:
set_cr0(), set_cr3(), set_cr4(), set_cr8(), get_cr8(), lmsw(),
load_pdptrs()
The static function wrapper inject_gp is duplicated in kvm_main.c and
x86.c for now, the version in kvm_main.c should disappear once the last
user of it is gone too.
The function load_pdptrs is no longer static, and now defined in x86.h
for the time being, until the last user of it is gone from kvm_main.c.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-29 22:09:35 +07:00
|
|
|
|
2010-04-28 23:15:31 +07:00
|
|
|
vcpu->arch.cr3 = cr3;
|
2010-12-05 23:56:11 +07:00
|
|
|
__set_bit(VCPU_EXREG_CR3, (ulong *)&vcpu->arch.regs_avail);
|
2013-10-02 21:56:11 +07:00
|
|
|
kvm_mmu_new_cr3(vcpu);
|
2010-04-28 23:15:31 +07:00
|
|
|
return 0;
|
|
|
|
}
|
2008-02-24 16:20:43 +07:00
|
|
|
EXPORT_SYMBOL_GPL(kvm_set_cr3);
|
KVM: Portability: Move control register helper functions to x86.c
This patch moves the definitions of CR0_RESERVED_BITS,
CR4_RESERVED_BITS, and CR8_RESERVED_BITS along with the following
functions from kvm_main.c to x86.c:
set_cr0(), set_cr3(), set_cr4(), set_cr8(), get_cr8(), lmsw(),
load_pdptrs()
The static function wrapper inject_gp is duplicated in kvm_main.c and
x86.c for now, the version in kvm_main.c should disappear once the last
user of it is gone too.
The function load_pdptrs is no longer static, and now defined in x86.h
for the time being, until the last user of it is gone from kvm_main.c.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-29 22:09:35 +07:00
|
|
|
|
2010-12-21 17:12:00 +07:00
|
|
|
int kvm_set_cr8(struct kvm_vcpu *vcpu, unsigned long cr8)
|
KVM: Portability: Move control register helper functions to x86.c
This patch moves the definitions of CR0_RESERVED_BITS,
CR4_RESERVED_BITS, and CR8_RESERVED_BITS along with the following
functions from kvm_main.c to x86.c:
set_cr0(), set_cr3(), set_cr4(), set_cr8(), get_cr8(), lmsw(),
load_pdptrs()
The static function wrapper inject_gp is duplicated in kvm_main.c and
x86.c for now, the version in kvm_main.c should disappear once the last
user of it is gone too.
The function load_pdptrs is no longer static, and now defined in x86.h
for the time being, until the last user of it is gone from kvm_main.c.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-29 22:09:35 +07:00
|
|
|
{
|
2010-04-28 23:15:31 +07:00
|
|
|
if (cr8 & CR8_RESERVED_BITS)
|
|
|
|
return 1;
|
2015-07-29 17:05:37 +07:00
|
|
|
if (lapic_in_kernel(vcpu))
|
KVM: Portability: Move control register helper functions to x86.c
This patch moves the definitions of CR0_RESERVED_BITS,
CR4_RESERVED_BITS, and CR8_RESERVED_BITS along with the following
functions from kvm_main.c to x86.c:
set_cr0(), set_cr3(), set_cr4(), set_cr8(), get_cr8(), lmsw(),
load_pdptrs()
The static function wrapper inject_gp is duplicated in kvm_main.c and
x86.c for now, the version in kvm_main.c should disappear once the last
user of it is gone too.
The function load_pdptrs is no longer static, and now defined in x86.h
for the time being, until the last user of it is gone from kvm_main.c.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-29 22:09:35 +07:00
|
|
|
kvm_lapic_set_tpr(vcpu, cr8);
|
|
|
|
else
|
2007-12-13 22:50:52 +07:00
|
|
|
vcpu->arch.cr8 = cr8;
|
2010-04-28 23:15:31 +07:00
|
|
|
return 0;
|
|
|
|
}
|
2008-02-24 16:20:43 +07:00
|
|
|
EXPORT_SYMBOL_GPL(kvm_set_cr8);
|
KVM: Portability: Move control register helper functions to x86.c
This patch moves the definitions of CR0_RESERVED_BITS,
CR4_RESERVED_BITS, and CR8_RESERVED_BITS along with the following
functions from kvm_main.c to x86.c:
set_cr0(), set_cr3(), set_cr4(), set_cr8(), get_cr8(), lmsw(),
load_pdptrs()
The static function wrapper inject_gp is duplicated in kvm_main.c and
x86.c for now, the version in kvm_main.c should disappear once the last
user of it is gone too.
The function load_pdptrs is no longer static, and now defined in x86.h
for the time being, until the last user of it is gone from kvm_main.c.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-29 22:09:35 +07:00
|
|
|
|
2008-02-24 16:20:43 +07:00
|
|
|
unsigned long kvm_get_cr8(struct kvm_vcpu *vcpu)
|
KVM: Portability: Move control register helper functions to x86.c
This patch moves the definitions of CR0_RESERVED_BITS,
CR4_RESERVED_BITS, and CR8_RESERVED_BITS along with the following
functions from kvm_main.c to x86.c:
set_cr0(), set_cr3(), set_cr4(), set_cr8(), get_cr8(), lmsw(),
load_pdptrs()
The static function wrapper inject_gp is duplicated in kvm_main.c and
x86.c for now, the version in kvm_main.c should disappear once the last
user of it is gone too.
The function load_pdptrs is no longer static, and now defined in x86.h
for the time being, until the last user of it is gone from kvm_main.c.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-29 22:09:35 +07:00
|
|
|
{
|
2015-07-29 17:05:37 +07:00
|
|
|
if (lapic_in_kernel(vcpu))
|
KVM: Portability: Move control register helper functions to x86.c
This patch moves the definitions of CR0_RESERVED_BITS,
CR4_RESERVED_BITS, and CR8_RESERVED_BITS along with the following
functions from kvm_main.c to x86.c:
set_cr0(), set_cr3(), set_cr4(), set_cr8(), get_cr8(), lmsw(),
load_pdptrs()
The static function wrapper inject_gp is duplicated in kvm_main.c and
x86.c for now, the version in kvm_main.c should disappear once the last
user of it is gone too.
The function load_pdptrs is no longer static, and now defined in x86.h
for the time being, until the last user of it is gone from kvm_main.c.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-29 22:09:35 +07:00
|
|
|
return kvm_lapic_get_cr8(vcpu);
|
|
|
|
else
|
2007-12-13 22:50:52 +07:00
|
|
|
return vcpu->arch.cr8;
|
KVM: Portability: Move control register helper functions to x86.c
This patch moves the definitions of CR0_RESERVED_BITS,
CR4_RESERVED_BITS, and CR8_RESERVED_BITS along with the following
functions from kvm_main.c to x86.c:
set_cr0(), set_cr3(), set_cr4(), set_cr8(), get_cr8(), lmsw(),
load_pdptrs()
The static function wrapper inject_gp is duplicated in kvm_main.c and
x86.c for now, the version in kvm_main.c should disappear once the last
user of it is gone too.
The function load_pdptrs is no longer static, and now defined in x86.h
for the time being, until the last user of it is gone from kvm_main.c.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-29 22:09:35 +07:00
|
|
|
}
|
2008-02-24 16:20:43 +07:00
|
|
|
EXPORT_SYMBOL_GPL(kvm_get_cr8);
|
KVM: Portability: Move control register helper functions to x86.c
This patch moves the definitions of CR0_RESERVED_BITS,
CR4_RESERVED_BITS, and CR8_RESERVED_BITS along with the following
functions from kvm_main.c to x86.c:
set_cr0(), set_cr3(), set_cr4(), set_cr8(), get_cr8(), lmsw(),
load_pdptrs()
The static function wrapper inject_gp is duplicated in kvm_main.c and
x86.c for now, the version in kvm_main.c should disappear once the last
user of it is gone too.
The function load_pdptrs is no longer static, and now defined in x86.h
for the time being, until the last user of it is gone from kvm_main.c.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-29 22:09:35 +07:00
|
|
|
|
2015-04-02 07:10:37 +07:00
|
|
|
static void kvm_update_dr0123(struct kvm_vcpu *vcpu)
|
|
|
|
{
|
|
|
|
int i;
|
|
|
|
|
|
|
|
if (!(vcpu->guest_debug & KVM_GUESTDBG_USE_HW_BP)) {
|
|
|
|
for (i = 0; i < KVM_NR_DB_REGS; i++)
|
|
|
|
vcpu->arch.eff_db[i] = vcpu->arch.db[i];
|
|
|
|
vcpu->arch.switch_db_regs |= KVM_DEBUGREG_RELOAD;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2014-01-05 00:47:16 +07:00
|
|
|
static void kvm_update_dr6(struct kvm_vcpu *vcpu)
|
|
|
|
{
|
|
|
|
if (!(vcpu->guest_debug & KVM_GUESTDBG_USE_HW_BP))
|
|
|
|
kvm_x86_ops->set_dr6(vcpu, vcpu->arch.dr6);
|
|
|
|
}
|
|
|
|
|
2012-09-21 10:42:55 +07:00
|
|
|
static void kvm_update_dr7(struct kvm_vcpu *vcpu)
|
|
|
|
{
|
|
|
|
unsigned long dr7;
|
|
|
|
|
|
|
|
if (vcpu->guest_debug & KVM_GUESTDBG_USE_HW_BP)
|
|
|
|
dr7 = vcpu->arch.guest_debug_dr7;
|
|
|
|
else
|
|
|
|
dr7 = vcpu->arch.dr7;
|
|
|
|
kvm_x86_ops->set_dr7(vcpu, dr7);
|
2014-02-21 15:55:56 +07:00
|
|
|
vcpu->arch.switch_db_regs &= ~KVM_DEBUGREG_BP_ENABLED;
|
|
|
|
if (dr7 & DR7_BP_EN_MASK)
|
|
|
|
vcpu->arch.switch_db_regs |= KVM_DEBUGREG_BP_ENABLED;
|
2012-09-21 10:42:55 +07:00
|
|
|
}
|
|
|
|
|
2014-07-15 21:37:46 +07:00
|
|
|
static u64 kvm_dr6_fixed(struct kvm_vcpu *vcpu)
|
|
|
|
{
|
|
|
|
u64 fixed = DR6_FIXED_1;
|
|
|
|
|
|
|
|
if (!guest_cpuid_has_rtm(vcpu))
|
|
|
|
fixed |= DR6_RTM;
|
|
|
|
return fixed;
|
|
|
|
}
|
|
|
|
|
2010-04-28 23:15:32 +07:00
|
|
|
static int __kvm_set_dr(struct kvm_vcpu *vcpu, int dr, unsigned long val)
|
2010-04-13 14:05:23 +07:00
|
|
|
{
|
|
|
|
switch (dr) {
|
|
|
|
case 0 ... 3:
|
|
|
|
vcpu->arch.db[dr] = val;
|
|
|
|
if (!(vcpu->guest_debug & KVM_GUESTDBG_USE_HW_BP))
|
|
|
|
vcpu->arch.eff_db[dr] = val;
|
|
|
|
break;
|
|
|
|
case 4:
|
|
|
|
/* fall through */
|
|
|
|
case 6:
|
2010-04-28 23:15:32 +07:00
|
|
|
if (val & 0xffffffff00000000ULL)
|
|
|
|
return -1; /* #GP */
|
2014-07-15 21:37:46 +07:00
|
|
|
vcpu->arch.dr6 = (val & DR6_VOLATILE) | kvm_dr6_fixed(vcpu);
|
2014-01-05 00:47:16 +07:00
|
|
|
kvm_update_dr6(vcpu);
|
2010-04-13 14:05:23 +07:00
|
|
|
break;
|
|
|
|
case 5:
|
|
|
|
/* fall through */
|
|
|
|
default: /* 7 */
|
2010-04-28 23:15:32 +07:00
|
|
|
if (val & 0xffffffff00000000ULL)
|
|
|
|
return -1; /* #GP */
|
2010-04-13 14:05:23 +07:00
|
|
|
vcpu->arch.dr7 = (val & DR7_VOLATILE) | DR7_FIXED_1;
|
2012-09-21 10:42:55 +07:00
|
|
|
kvm_update_dr7(vcpu);
|
2010-04-13 14:05:23 +07:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
2010-04-28 23:15:32 +07:00
|
|
|
|
|
|
|
int kvm_set_dr(struct kvm_vcpu *vcpu, int dr, unsigned long val)
|
|
|
|
{
|
2014-10-03 05:10:05 +07:00
|
|
|
if (__kvm_set_dr(vcpu, dr, val)) {
|
2010-04-28 23:15:32 +07:00
|
|
|
kvm_inject_gp(vcpu, 0);
|
2014-10-03 05:10:05 +07:00
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
return 0;
|
2010-04-28 23:15:32 +07:00
|
|
|
}
|
2010-04-13 14:05:23 +07:00
|
|
|
EXPORT_SYMBOL_GPL(kvm_set_dr);
|
|
|
|
|
2014-10-03 05:10:05 +07:00
|
|
|
int kvm_get_dr(struct kvm_vcpu *vcpu, int dr, unsigned long *val)
|
2010-04-13 14:05:23 +07:00
|
|
|
{
|
|
|
|
switch (dr) {
|
|
|
|
case 0 ... 3:
|
|
|
|
*val = vcpu->arch.db[dr];
|
|
|
|
break;
|
|
|
|
case 4:
|
|
|
|
/* fall through */
|
|
|
|
case 6:
|
2014-01-05 00:47:16 +07:00
|
|
|
if (vcpu->guest_debug & KVM_GUESTDBG_USE_HW_BP)
|
|
|
|
*val = vcpu->arch.dr6;
|
|
|
|
else
|
|
|
|
*val = kvm_x86_ops->get_dr6(vcpu);
|
2010-04-13 14:05:23 +07:00
|
|
|
break;
|
|
|
|
case 5:
|
|
|
|
/* fall through */
|
|
|
|
default: /* 7 */
|
|
|
|
*val = vcpu->arch.dr7;
|
|
|
|
break;
|
|
|
|
}
|
2010-04-28 23:15:32 +07:00
|
|
|
return 0;
|
|
|
|
}
|
2010-04-13 14:05:23 +07:00
|
|
|
EXPORT_SYMBOL_GPL(kvm_get_dr);
|
|
|
|
|
2011-11-10 19:57:23 +07:00
|
|
|
bool kvm_rdpmc(struct kvm_vcpu *vcpu)
|
|
|
|
{
|
|
|
|
u32 ecx = kvm_register_read(vcpu, VCPU_REGS_RCX);
|
|
|
|
u64 data;
|
|
|
|
int err;
|
|
|
|
|
2015-06-19 18:44:45 +07:00
|
|
|
err = kvm_pmu_rdpmc(vcpu, ecx, &data);
|
2011-11-10 19:57:23 +07:00
|
|
|
if (err)
|
|
|
|
return err;
|
|
|
|
kvm_register_write(vcpu, VCPU_REGS_RAX, (u32)data);
|
|
|
|
kvm_register_write(vcpu, VCPU_REGS_RDX, data >> 32);
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(kvm_rdpmc);
|
|
|
|
|
2007-10-10 22:16:19 +07:00
|
|
|
/*
|
|
|
|
* List of msr numbers which we expose to userspace through KVM_GET_MSRS
|
|
|
|
* and KVM_SET_MSRS, and KVM_GET_MSR_INDEX_LIST.
|
|
|
|
*
|
|
|
|
* This list is modified at module load time to reflect the
|
2009-10-07 00:24:50 +07:00
|
|
|
* capabilities of the host cpu. This capabilities test skips MSRs that are
|
2015-05-05 17:08:55 +07:00
|
|
|
* kvm-specific. Those are put in emulated_msrs; filtering of emulated_msrs
|
|
|
|
* may depend on host virtualization features rather than host cpu features.
|
2007-10-10 22:16:19 +07:00
|
|
|
*/
|
2009-10-07 00:24:50 +07:00
|
|
|
|
2007-10-10 22:16:19 +07:00
|
|
|
static u32 msrs_to_save[] = {
|
|
|
|
MSR_IA32_SYSENTER_CS, MSR_IA32_SYSENTER_ESP, MSR_IA32_SYSENTER_EIP,
|
2010-07-17 20:03:26 +07:00
|
|
|
MSR_STAR,
|
2007-10-10 22:16:19 +07:00
|
|
|
#ifdef CONFIG_X86_64
|
|
|
|
MSR_CSTAR, MSR_KERNEL_GS_BASE, MSR_SYSCALL_MASK, MSR_LSTAR,
|
|
|
|
#endif
|
2013-07-08 18:12:35 +07:00
|
|
|
MSR_IA32_TSC, MSR_IA32_CR_PAT, MSR_VM_HSAVE_PA,
|
2015-11-12 20:49:17 +07:00
|
|
|
MSR_IA32_FEATURE_CONTROL, MSR_IA32_BNDCFGS, MSR_TSC_AUX,
|
2007-10-10 22:16:19 +07:00
|
|
|
};
|
|
|
|
|
|
|
|
static unsigned num_msrs_to_save;
|
|
|
|
|
2015-05-05 17:08:55 +07:00
|
|
|
static u32 emulated_msrs[] = {
|
|
|
|
MSR_KVM_SYSTEM_TIME, MSR_KVM_WALL_CLOCK,
|
|
|
|
MSR_KVM_SYSTEM_TIME_NEW, MSR_KVM_WALL_CLOCK_NEW,
|
|
|
|
HV_X64_MSR_GUEST_OS_ID, HV_X64_MSR_HYPERCALL,
|
|
|
|
HV_X64_MSR_TIME_REF_COUNT, HV_X64_MSR_REFERENCE_TSC,
|
2015-07-03 19:01:37 +07:00
|
|
|
HV_X64_MSR_CRASH_P0, HV_X64_MSR_CRASH_P1, HV_X64_MSR_CRASH_P2,
|
|
|
|
HV_X64_MSR_CRASH_P3, HV_X64_MSR_CRASH_P4, HV_X64_MSR_CRASH_CTL,
|
2015-09-16 16:29:48 +07:00
|
|
|
HV_X64_MSR_RESET,
|
2015-09-16 16:29:49 +07:00
|
|
|
HV_X64_MSR_VP_INDEX,
|
2015-09-16 16:29:50 +07:00
|
|
|
HV_X64_MSR_VP_RUNTIME,
|
2015-11-10 19:36:34 +07:00
|
|
|
HV_X64_MSR_SCONTROL,
|
2015-11-30 23:22:21 +07:00
|
|
|
HV_X64_MSR_STIMER0_CONFIG,
|
2015-05-05 17:08:55 +07:00
|
|
|
HV_X64_MSR_APIC_ASSIST_PAGE, MSR_KVM_ASYNC_PF_EN, MSR_KVM_STEAL_TIME,
|
|
|
|
MSR_KVM_PV_EOI_EN,
|
|
|
|
|
2012-11-30 03:42:50 +07:00
|
|
|
MSR_IA32_TSC_ADJUST,
|
2011-09-22 15:55:52 +07:00
|
|
|
MSR_IA32_TSCDEADLINE,
|
2007-10-10 22:16:19 +07:00
|
|
|
MSR_IA32_MISC_ENABLE,
|
2010-07-07 18:09:38 +07:00
|
|
|
MSR_IA32_MCG_STATUS,
|
|
|
|
MSR_IA32_MCG_CTL,
|
2016-06-22 13:59:56 +07:00
|
|
|
MSR_IA32_MCG_EXT_CTL,
|
2015-05-07 16:36:11 +07:00
|
|
|
MSR_IA32_SMBASE,
|
2007-10-10 22:16:19 +07:00
|
|
|
};
|
|
|
|
|
2015-05-05 17:08:55 +07:00
|
|
|
static unsigned num_emulated_msrs;
|
|
|
|
|
2013-04-20 15:52:36 +07:00
|
|
|
bool kvm_valid_efer(struct kvm_vcpu *vcpu, u64 efer)
|
2007-10-31 00:44:17 +07:00
|
|
|
{
|
2010-05-06 16:38:43 +07:00
|
|
|
if (efer & efer_reserved_bits)
|
2013-04-20 15:52:36 +07:00
|
|
|
return false;
|
2007-10-31 00:44:17 +07:00
|
|
|
|
2009-02-02 22:23:51 +07:00
|
|
|
if (efer & EFER_FFXSR) {
|
|
|
|
struct kvm_cpuid_entry2 *feat;
|
|
|
|
|
|
|
|
feat = kvm_find_cpuid_entry(vcpu, 0x80000001, 0);
|
2010-05-06 16:38:43 +07:00
|
|
|
if (!feat || !(feat->edx & bit(X86_FEATURE_FXSR_OPT)))
|
2013-04-20 15:52:36 +07:00
|
|
|
return false;
|
2009-02-02 22:23:51 +07:00
|
|
|
}
|
|
|
|
|
2008-11-26 02:17:11 +07:00
|
|
|
if (efer & EFER_SVME) {
|
|
|
|
struct kvm_cpuid_entry2 *feat;
|
|
|
|
|
|
|
|
feat = kvm_find_cpuid_entry(vcpu, 0x80000001, 0);
|
2010-05-06 16:38:43 +07:00
|
|
|
if (!feat || !(feat->ecx & bit(X86_FEATURE_SVM)))
|
2013-04-20 15:52:36 +07:00
|
|
|
return false;
|
2008-11-26 02:17:11 +07:00
|
|
|
}
|
|
|
|
|
2013-04-20 15:52:36 +07:00
|
|
|
return true;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(kvm_valid_efer);
|
|
|
|
|
|
|
|
static int set_efer(struct kvm_vcpu *vcpu, u64 efer)
|
|
|
|
{
|
|
|
|
u64 old_efer = vcpu->arch.efer;
|
|
|
|
|
|
|
|
if (!kvm_valid_efer(vcpu, efer))
|
|
|
|
return 1;
|
|
|
|
|
|
|
|
if (is_paging(vcpu)
|
|
|
|
&& (vcpu->arch.efer & EFER_LME) != (efer & EFER_LME))
|
|
|
|
return 1;
|
|
|
|
|
2007-10-31 00:44:17 +07:00
|
|
|
efer &= ~EFER_LMA;
|
2010-01-21 20:31:50 +07:00
|
|
|
efer |= vcpu->arch.efer & EFER_LMA;
|
2007-10-31 00:44:17 +07:00
|
|
|
|
2010-05-12 15:40:40 +07:00
|
|
|
kvm_x86_ops->set_efer(vcpu, efer);
|
|
|
|
|
2010-05-12 15:40:42 +07:00
|
|
|
/* Update reserved bits */
|
|
|
|
if ((efer ^ old_efer) & EFER_NX)
|
|
|
|
kvm_mmu_reset_context(vcpu);
|
|
|
|
|
2010-05-06 16:38:43 +07:00
|
|
|
return 0;
|
2007-10-31 00:44:17 +07:00
|
|
|
}
|
|
|
|
|
2008-01-31 20:57:37 +07:00
|
|
|
void kvm_enable_efer_bits(u64 mask)
|
|
|
|
{
|
|
|
|
efer_reserved_bits &= ~mask;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(kvm_enable_efer_bits);
|
|
|
|
|
2007-10-31 00:44:17 +07:00
|
|
|
/*
|
|
|
|
* Writes msr value into into the appropriate "register".
|
|
|
|
* Returns 0 on success, non-0 otherwise.
|
|
|
|
* Assumes vcpu_load() was already called.
|
|
|
|
*/
|
2012-11-30 03:42:12 +07:00
|
|
|
int kvm_set_msr(struct kvm_vcpu *vcpu, struct msr_data *msr)
|
2007-10-31 00:44:17 +07:00
|
|
|
{
|
KVM: x86: Check non-canonical addresses upon WRMSR
Upon WRMSR, the CPU should inject #GP if a non-canonical value (address) is
written to certain MSRs. The behavior is "almost" identical for AMD and Intel
(ignoring MSRs that are not implemented in either architecture since they would
anyhow #GP). However, IA32_SYSENTER_ESP and IA32_SYSENTER_EIP cause #GP if
non-canonical address is written on Intel but not on AMD (which ignores the top
32-bits).
Accordingly, this patch injects a #GP on the MSRs which behave identically on
Intel and AMD. To eliminate the differences between the architecutres, the
value which is written to IA32_SYSENTER_ESP and IA32_SYSENTER_EIP is turned to
canonical value before writing instead of injecting a #GP.
Some references from Intel and AMD manuals:
According to Intel SDM description of WRMSR instruction #GP is expected on
WRMSR "If the source register contains a non-canonical address and ECX
specifies one of the following MSRs: IA32_DS_AREA, IA32_FS_BASE, IA32_GS_BASE,
IA32_KERNEL_GS_BASE, IA32_LSTAR, IA32_SYSENTER_EIP, IA32_SYSENTER_ESP."
According to AMD manual instruction manual:
LSTAR/CSTAR (SYSCALL): "The WRMSR instruction loads the target RIP into the
LSTAR and CSTAR registers. If an RIP written by WRMSR is not in canonical
form, a general-protection exception (#GP) occurs."
IA32_GS_BASE and IA32_FS_BASE (WRFSBASE/WRGSBASE): "The address written to the
base field must be in canonical form or a #GP fault will occur."
IA32_KERNEL_GS_BASE (SWAPGS): "The address stored in the KernelGSbase MSR must
be in canonical form."
This patch fixes CVE-2014-3610.
Cc: stable@vger.kernel.org
Signed-off-by: Nadav Amit <namit@cs.technion.ac.il>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2014-09-16 07:24:05 +07:00
|
|
|
switch (msr->index) {
|
|
|
|
case MSR_FS_BASE:
|
|
|
|
case MSR_GS_BASE:
|
|
|
|
case MSR_KERNEL_GS_BASE:
|
|
|
|
case MSR_CSTAR:
|
|
|
|
case MSR_LSTAR:
|
|
|
|
if (is_noncanonical_address(msr->data))
|
|
|
|
return 1;
|
|
|
|
break;
|
|
|
|
case MSR_IA32_SYSENTER_EIP:
|
|
|
|
case MSR_IA32_SYSENTER_ESP:
|
|
|
|
/*
|
|
|
|
* IA32_SYSENTER_ESP and IA32_SYSENTER_EIP cause #GP if
|
|
|
|
* non-canonical address is written on Intel but not on
|
|
|
|
* AMD (which ignores the top 32-bits, because it does
|
|
|
|
* not implement 64-bit SYSENTER).
|
|
|
|
*
|
|
|
|
* 64-bit code should hence be able to write a non-canonical
|
|
|
|
* value on AMD. Making the address canonical ensures that
|
|
|
|
* vmentry does not fail on Intel after writing a non-canonical
|
|
|
|
* value, and that something deterministic happens if the guest
|
|
|
|
* invokes 64-bit SYSENTER.
|
|
|
|
*/
|
|
|
|
msr->data = get_canonical(msr->data);
|
|
|
|
}
|
2012-11-30 03:42:12 +07:00
|
|
|
return kvm_x86_ops->set_msr(vcpu, msr);
|
2007-10-31 00:44:17 +07:00
|
|
|
}
|
KVM: x86: Check non-canonical addresses upon WRMSR
Upon WRMSR, the CPU should inject #GP if a non-canonical value (address) is
written to certain MSRs. The behavior is "almost" identical for AMD and Intel
(ignoring MSRs that are not implemented in either architecture since they would
anyhow #GP). However, IA32_SYSENTER_ESP and IA32_SYSENTER_EIP cause #GP if
non-canonical address is written on Intel but not on AMD (which ignores the top
32-bits).
Accordingly, this patch injects a #GP on the MSRs which behave identically on
Intel and AMD. To eliminate the differences between the architecutres, the
value which is written to IA32_SYSENTER_ESP and IA32_SYSENTER_EIP is turned to
canonical value before writing instead of injecting a #GP.
Some references from Intel and AMD manuals:
According to Intel SDM description of WRMSR instruction #GP is expected on
WRMSR "If the source register contains a non-canonical address and ECX
specifies one of the following MSRs: IA32_DS_AREA, IA32_FS_BASE, IA32_GS_BASE,
IA32_KERNEL_GS_BASE, IA32_LSTAR, IA32_SYSENTER_EIP, IA32_SYSENTER_ESP."
According to AMD manual instruction manual:
LSTAR/CSTAR (SYSCALL): "The WRMSR instruction loads the target RIP into the
LSTAR and CSTAR registers. If an RIP written by WRMSR is not in canonical
form, a general-protection exception (#GP) occurs."
IA32_GS_BASE and IA32_FS_BASE (WRFSBASE/WRGSBASE): "The address written to the
base field must be in canonical form or a #GP fault will occur."
IA32_KERNEL_GS_BASE (SWAPGS): "The address stored in the KernelGSbase MSR must
be in canonical form."
This patch fixes CVE-2014-3610.
Cc: stable@vger.kernel.org
Signed-off-by: Nadav Amit <namit@cs.technion.ac.il>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2014-09-16 07:24:05 +07:00
|
|
|
EXPORT_SYMBOL_GPL(kvm_set_msr);
|
2007-10-31 00:44:17 +07:00
|
|
|
|
KVM: Portability: split kvm_vcpu_ioctl
This patch splits kvm_vcpu_ioctl into archtecture independent parts, and
x86 specific parts which go to kvm_arch_vcpu_ioctl in x86.c.
Common ioctls for all architectures are:
KVM_RUN, KVM_GET/SET_(S-)REGS, KVM_TRANSLATE, KVM_INTERRUPT,
KVM_DEBUG_GUEST, KVM_SET_SIGNAL_MASK, KVM_GET/SET_FPU
Note that some PPC chips don't have an FPU, so we might need an #ifdef
around KVM_GET/SET_FPU one day.
x86 specific ioctls are:
KVM_GET/SET_LAPIC, KVM_SET_CPUID, KVM_GET/SET_MSRS
An interresting aspect is vcpu_load/vcpu_put. We now have a common
vcpu_load/put which does the preemption stuff, and an architecture
specific kvm_arch_vcpu_load/put. In the x86 case, this one calls the
vmx/svm function defined in kvm_x86_ops.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Reviewed-by: Christian Ehrhardt <ehrhardt@linux.vnet.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-12 00:16:52 +07:00
|
|
|
/*
|
|
|
|
* Adapt set_msr() to msr_io()'s calling convention
|
|
|
|
*/
|
2015-04-08 20:30:38 +07:00
|
|
|
static int do_get_msr(struct kvm_vcpu *vcpu, unsigned index, u64 *data)
|
|
|
|
{
|
|
|
|
struct msr_data msr;
|
|
|
|
int r;
|
|
|
|
|
|
|
|
msr.index = index;
|
|
|
|
msr.host_initiated = true;
|
|
|
|
r = kvm_get_msr(vcpu, &msr);
|
|
|
|
if (r)
|
|
|
|
return r;
|
|
|
|
|
|
|
|
*data = msr.data;
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
KVM: Portability: split kvm_vcpu_ioctl
This patch splits kvm_vcpu_ioctl into archtecture independent parts, and
x86 specific parts which go to kvm_arch_vcpu_ioctl in x86.c.
Common ioctls for all architectures are:
KVM_RUN, KVM_GET/SET_(S-)REGS, KVM_TRANSLATE, KVM_INTERRUPT,
KVM_DEBUG_GUEST, KVM_SET_SIGNAL_MASK, KVM_GET/SET_FPU
Note that some PPC chips don't have an FPU, so we might need an #ifdef
around KVM_GET/SET_FPU one day.
x86 specific ioctls are:
KVM_GET/SET_LAPIC, KVM_SET_CPUID, KVM_GET/SET_MSRS
An interresting aspect is vcpu_load/vcpu_put. We now have a common
vcpu_load/put which does the preemption stuff, and an architecture
specific kvm_arch_vcpu_load/put. In the x86 case, this one calls the
vmx/svm function defined in kvm_x86_ops.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Reviewed-by: Christian Ehrhardt <ehrhardt@linux.vnet.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-12 00:16:52 +07:00
|
|
|
static int do_set_msr(struct kvm_vcpu *vcpu, unsigned index, u64 *data)
|
|
|
|
{
|
2012-11-30 03:42:12 +07:00
|
|
|
struct msr_data msr;
|
|
|
|
|
|
|
|
msr.data = *data;
|
|
|
|
msr.index = index;
|
|
|
|
msr.host_initiated = true;
|
|
|
|
return kvm_set_msr(vcpu, &msr);
|
KVM: Portability: split kvm_vcpu_ioctl
This patch splits kvm_vcpu_ioctl into archtecture independent parts, and
x86 specific parts which go to kvm_arch_vcpu_ioctl in x86.c.
Common ioctls for all architectures are:
KVM_RUN, KVM_GET/SET_(S-)REGS, KVM_TRANSLATE, KVM_INTERRUPT,
KVM_DEBUG_GUEST, KVM_SET_SIGNAL_MASK, KVM_GET/SET_FPU
Note that some PPC chips don't have an FPU, so we might need an #ifdef
around KVM_GET/SET_FPU one day.
x86 specific ioctls are:
KVM_GET/SET_LAPIC, KVM_SET_CPUID, KVM_GET/SET_MSRS
An interresting aspect is vcpu_load/vcpu_put. We now have a common
vcpu_load/put which does the preemption stuff, and an architecture
specific kvm_arch_vcpu_load/put. In the x86 case, this one calls the
vmx/svm function defined in kvm_x86_ops.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Reviewed-by: Christian Ehrhardt <ehrhardt@linux.vnet.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-12 00:16:52 +07:00
|
|
|
}
|
|
|
|
|
2012-11-28 08:29:00 +07:00
|
|
|
#ifdef CONFIG_X86_64
|
|
|
|
struct pvclock_gtod_data {
|
|
|
|
seqcount_t seq;
|
|
|
|
|
|
|
|
struct { /* extract of a clocksource struct */
|
|
|
|
int vclock_mode;
|
|
|
|
cycle_t cycle_last;
|
|
|
|
cycle_t mask;
|
|
|
|
u32 mult;
|
|
|
|
u32 shift;
|
|
|
|
} clock;
|
|
|
|
|
2014-07-17 04:04:54 +07:00
|
|
|
u64 boot_ns;
|
|
|
|
u64 nsec_base;
|
2012-11-28 08:29:00 +07:00
|
|
|
};
|
|
|
|
|
|
|
|
static struct pvclock_gtod_data pvclock_gtod_data;
|
|
|
|
|
|
|
|
static void update_pvclock_gtod(struct timekeeper *tk)
|
|
|
|
{
|
|
|
|
struct pvclock_gtod_data *vdata = &pvclock_gtod_data;
|
2014-07-17 04:04:54 +07:00
|
|
|
u64 boot_ns;
|
|
|
|
|
2015-03-19 16:09:06 +07:00
|
|
|
boot_ns = ktime_to_ns(ktime_add(tk->tkr_mono.base, tk->offs_boot));
|
2012-11-28 08:29:00 +07:00
|
|
|
|
|
|
|
write_seqcount_begin(&vdata->seq);
|
|
|
|
|
|
|
|
/* copy pvclock gtod data */
|
2015-03-19 16:09:06 +07:00
|
|
|
vdata->clock.vclock_mode = tk->tkr_mono.clock->archdata.vclock_mode;
|
|
|
|
vdata->clock.cycle_last = tk->tkr_mono.cycle_last;
|
|
|
|
vdata->clock.mask = tk->tkr_mono.mask;
|
|
|
|
vdata->clock.mult = tk->tkr_mono.mult;
|
|
|
|
vdata->clock.shift = tk->tkr_mono.shift;
|
2012-11-28 08:29:00 +07:00
|
|
|
|
2014-07-17 04:04:54 +07:00
|
|
|
vdata->boot_ns = boot_ns;
|
2015-03-19 16:09:06 +07:00
|
|
|
vdata->nsec_base = tk->tkr_mono.xtime_nsec;
|
2012-11-28 08:29:00 +07:00
|
|
|
|
|
|
|
write_seqcount_end(&vdata->seq);
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
|
2015-01-02 10:05:18 +07:00
|
|
|
void kvm_set_pending_timer(struct kvm_vcpu *vcpu)
|
|
|
|
{
|
|
|
|
/*
|
|
|
|
* Note: KVM_REQ_PENDING_TIMER is implicitly checked in
|
|
|
|
* vcpu_enter_guest. This function is only called from
|
|
|
|
* the physical CPU that is running vcpu.
|
|
|
|
*/
|
|
|
|
kvm_make_request(KVM_REQ_PENDING_TIMER, vcpu);
|
|
|
|
}
|
2012-11-28 08:29:00 +07:00
|
|
|
|
2008-02-16 02:52:47 +07:00
|
|
|
static void kvm_write_wall_clock(struct kvm *kvm, gpa_t wall_clock)
|
|
|
|
{
|
2010-05-04 19:00:37 +07:00
|
|
|
int version;
|
|
|
|
int r;
|
2008-06-03 21:17:31 +07:00
|
|
|
struct pvclock_wall_clock wc;
|
2016-06-17 22:48:56 +07:00
|
|
|
struct timespec64 boot;
|
2008-02-16 02:52:47 +07:00
|
|
|
|
|
|
|
if (!wall_clock)
|
|
|
|
return;
|
|
|
|
|
2010-05-04 19:00:37 +07:00
|
|
|
r = kvm_read_guest(kvm, wall_clock, &version, sizeof(version));
|
|
|
|
if (r)
|
|
|
|
return;
|
|
|
|
|
|
|
|
if (version & 1)
|
|
|
|
++version; /* first time write, random junk */
|
|
|
|
|
|
|
|
++version;
|
2008-02-16 02:52:47 +07:00
|
|
|
|
2015-12-31 01:08:46 +07:00
|
|
|
if (kvm_write_guest(kvm, wall_clock, &version, sizeof(version)))
|
|
|
|
return;
|
2008-02-16 02:52:47 +07:00
|
|
|
|
2008-06-03 21:17:31 +07:00
|
|
|
/*
|
|
|
|
* The guest calculates current wall clock time by adding
|
2010-09-19 07:38:14 +07:00
|
|
|
* system time (updated by kvm_guest_time_update below) to the
|
2008-06-03 21:17:31 +07:00
|
|
|
* wall clock specified here. guest system time equals host
|
|
|
|
* system time for us, thus we must fill in host boot time here.
|
|
|
|
*/
|
2016-06-17 22:48:56 +07:00
|
|
|
getboottime64(&boot);
|
2008-06-03 21:17:31 +07:00
|
|
|
|
2012-07-20 23:44:24 +07:00
|
|
|
if (kvm->arch.kvmclock_offset) {
|
2016-06-17 22:48:56 +07:00
|
|
|
struct timespec64 ts = ns_to_timespec64(kvm->arch.kvmclock_offset);
|
|
|
|
boot = timespec64_sub(boot, ts);
|
2012-07-20 23:44:24 +07:00
|
|
|
}
|
2016-06-17 22:48:56 +07:00
|
|
|
wc.sec = (u32)boot.tv_sec; /* overflow in 2106 guest time */
|
2008-06-03 21:17:31 +07:00
|
|
|
wc.nsec = boot.tv_nsec;
|
|
|
|
wc.version = version;
|
2008-02-16 02:52:47 +07:00
|
|
|
|
|
|
|
kvm_write_guest(kvm, wall_clock, &wc, sizeof(wc));
|
|
|
|
|
|
|
|
version++;
|
|
|
|
kvm_write_guest(kvm, wall_clock, &version, sizeof(version));
|
|
|
|
}
|
|
|
|
|
2008-06-03 21:17:31 +07:00
|
|
|
static uint32_t div_frac(uint32_t dividend, uint32_t divisor)
|
|
|
|
{
|
2016-01-22 17:39:22 +07:00
|
|
|
do_shl32_div32(dividend, divisor);
|
|
|
|
return dividend;
|
2008-06-03 21:17:31 +07:00
|
|
|
}
|
|
|
|
|
2016-02-08 21:11:15 +07:00
|
|
|
static void kvm_get_time_scale(uint64_t scaled_hz, uint64_t base_hz,
|
2010-09-19 07:38:13 +07:00
|
|
|
s8 *pshift, u32 *pmultiplier)
|
2008-06-03 21:17:31 +07:00
|
|
|
{
|
2010-09-19 07:38:13 +07:00
|
|
|
uint64_t scaled64;
|
2008-06-03 21:17:31 +07:00
|
|
|
int32_t shift = 0;
|
|
|
|
uint64_t tps64;
|
|
|
|
uint32_t tps32;
|
|
|
|
|
2016-02-08 21:11:15 +07:00
|
|
|
tps64 = base_hz;
|
|
|
|
scaled64 = scaled_hz;
|
2010-09-26 18:00:53 +07:00
|
|
|
while (tps64 > scaled64*2 || tps64 & 0xffffffff00000000ULL) {
|
2008-06-03 21:17:31 +07:00
|
|
|
tps64 >>= 1;
|
|
|
|
shift--;
|
|
|
|
}
|
|
|
|
|
|
|
|
tps32 = (uint32_t)tps64;
|
2010-09-26 18:00:53 +07:00
|
|
|
while (tps32 <= scaled64 || scaled64 & 0xffffffff00000000ULL) {
|
|
|
|
if (scaled64 & 0xffffffff00000000ULL || tps32 & 0x80000000)
|
2010-09-19 07:38:13 +07:00
|
|
|
scaled64 >>= 1;
|
|
|
|
else
|
|
|
|
tps32 <<= 1;
|
2008-06-03 21:17:31 +07:00
|
|
|
shift++;
|
|
|
|
}
|
|
|
|
|
2010-09-19 07:38:13 +07:00
|
|
|
*pshift = shift;
|
|
|
|
*pmultiplier = div_frac(scaled64, tps32);
|
2008-06-03 21:17:31 +07:00
|
|
|
|
2016-02-08 21:11:15 +07:00
|
|
|
pr_debug("%s: base_hz %llu => %llu, shift %d, mul %u\n",
|
|
|
|
__func__, base_hz, scaled_hz, shift, *pmultiplier);
|
2008-06-03 21:17:31 +07:00
|
|
|
}
|
|
|
|
|
2012-11-28 08:29:01 +07:00
|
|
|
#ifdef CONFIG_X86_64
|
2012-11-28 08:29:00 +07:00
|
|
|
static atomic_t kvm_guest_has_master_clock = ATOMIC_INIT(0);
|
2012-11-28 08:29:01 +07:00
|
|
|
#endif
|
2012-11-28 08:29:00 +07:00
|
|
|
|
2009-02-04 23:52:04 +07:00
|
|
|
static DEFINE_PER_CPU(unsigned long, cpu_tsc_khz);
|
2015-01-19 21:33:39 +07:00
|
|
|
static unsigned long max_tsc_khz;
|
2009-02-04 23:52:04 +07:00
|
|
|
|
KVM: Infrastructure for software and hardware based TSC rate scaling
This requires some restructuring; rather than use 'virtual_tsc_khz'
to indicate whether hardware rate scaling is in effect, we consider
each VCPU to always have a virtual TSC rate. Instead, there is new
logic above the vendor-specific hardware scaling that decides whether
it is even necessary to use and updates all rate variables used by
common code. This means we can simply query the virtual rate at
any point, which is needed for software rate scaling.
There is also now a threshold added to the TSC rate scaling; minor
differences and variations of measured TSC rate can accidentally
provoke rate scaling to be used when it is not needed. Instead,
we have a tolerance variable called tsc_tolerance_ppm, which is
the maximum variation from user requested rate at which scaling
will be used. The default is 250ppm, which is the half the
threshold for NTP adjustment, allowing for some hardware variation.
In the event that hardware rate scaling is not available, we can
kludge a bit by forcing TSC catchup to turn on when a faster than
hardware speed has been requested, but there is nothing available
yet for the reverse case; this requires a trap and emulate software
implementation for RDTSC, which is still forthcoming.
[avi: fix 64-bit division on i386]
Signed-off-by: Zachary Amsden <zamsden@gmail.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
2012-02-04 00:43:50 +07:00
|
|
|
static u32 adjust_tsc_khz(u32 khz, s32 ppm)
|
2011-03-25 15:44:47 +07:00
|
|
|
{
|
KVM: Infrastructure for software and hardware based TSC rate scaling
This requires some restructuring; rather than use 'virtual_tsc_khz'
to indicate whether hardware rate scaling is in effect, we consider
each VCPU to always have a virtual TSC rate. Instead, there is new
logic above the vendor-specific hardware scaling that decides whether
it is even necessary to use and updates all rate variables used by
common code. This means we can simply query the virtual rate at
any point, which is needed for software rate scaling.
There is also now a threshold added to the TSC rate scaling; minor
differences and variations of measured TSC rate can accidentally
provoke rate scaling to be used when it is not needed. Instead,
we have a tolerance variable called tsc_tolerance_ppm, which is
the maximum variation from user requested rate at which scaling
will be used. The default is 250ppm, which is the half the
threshold for NTP adjustment, allowing for some hardware variation.
In the event that hardware rate scaling is not available, we can
kludge a bit by forcing TSC catchup to turn on when a faster than
hardware speed has been requested, but there is nothing available
yet for the reverse case; this requires a trap and emulate software
implementation for RDTSC, which is still forthcoming.
[avi: fix 64-bit division on i386]
Signed-off-by: Zachary Amsden <zamsden@gmail.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
2012-02-04 00:43:50 +07:00
|
|
|
u64 v = (u64)khz * (1000000 + ppm);
|
|
|
|
do_div(v, 1000000);
|
|
|
|
return v;
|
2011-03-25 15:44:47 +07:00
|
|
|
}
|
|
|
|
|
2015-10-20 14:39:04 +07:00
|
|
|
static int set_tsc_khz(struct kvm_vcpu *vcpu, u32 user_tsc_khz, bool scale)
|
|
|
|
{
|
|
|
|
u64 ratio;
|
|
|
|
|
|
|
|
/* Guest TSC same frequency as host TSC? */
|
|
|
|
if (!scale) {
|
|
|
|
vcpu->arch.tsc_scaling_ratio = kvm_default_tsc_scaling_ratio;
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* TSC scaling supported? */
|
|
|
|
if (!kvm_has_tsc_control) {
|
|
|
|
if (user_tsc_khz > tsc_khz) {
|
|
|
|
vcpu->arch.tsc_catchup = 1;
|
|
|
|
vcpu->arch.tsc_always_catchup = 1;
|
|
|
|
return 0;
|
|
|
|
} else {
|
|
|
|
WARN(1, "user requested TSC rate below hardware speed\n");
|
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/* TSC scaling required - calculate ratio */
|
|
|
|
ratio = mul_u64_u32_div(1ULL << kvm_tsc_scaling_ratio_frac_bits,
|
|
|
|
user_tsc_khz, tsc_khz);
|
|
|
|
|
|
|
|
if (ratio == 0 || ratio >= kvm_max_tsc_scaling_ratio) {
|
|
|
|
WARN_ONCE(1, "Invalid TSC scaling ratio - virtual-tsc-khz=%u\n",
|
|
|
|
user_tsc_khz);
|
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
|
|
|
|
vcpu->arch.tsc_scaling_ratio = ratio;
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2016-02-08 20:51:12 +07:00
|
|
|
static int kvm_set_tsc_khz(struct kvm_vcpu *vcpu, u32 user_tsc_khz)
|
2010-08-20 15:07:25 +07:00
|
|
|
{
|
KVM: Infrastructure for software and hardware based TSC rate scaling
This requires some restructuring; rather than use 'virtual_tsc_khz'
to indicate whether hardware rate scaling is in effect, we consider
each VCPU to always have a virtual TSC rate. Instead, there is new
logic above the vendor-specific hardware scaling that decides whether
it is even necessary to use and updates all rate variables used by
common code. This means we can simply query the virtual rate at
any point, which is needed for software rate scaling.
There is also now a threshold added to the TSC rate scaling; minor
differences and variations of measured TSC rate can accidentally
provoke rate scaling to be used when it is not needed. Instead,
we have a tolerance variable called tsc_tolerance_ppm, which is
the maximum variation from user requested rate at which scaling
will be used. The default is 250ppm, which is the half the
threshold for NTP adjustment, allowing for some hardware variation.
In the event that hardware rate scaling is not available, we can
kludge a bit by forcing TSC catchup to turn on when a faster than
hardware speed has been requested, but there is nothing available
yet for the reverse case; this requires a trap and emulate software
implementation for RDTSC, which is still forthcoming.
[avi: fix 64-bit division on i386]
Signed-off-by: Zachary Amsden <zamsden@gmail.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
2012-02-04 00:43:50 +07:00
|
|
|
u32 thresh_lo, thresh_hi;
|
|
|
|
int use_scaling = 0;
|
2010-08-26 17:38:03 +07:00
|
|
|
|
2013-03-12 09:10:24 +07:00
|
|
|
/* tsc_khz can be zero if TSC calibration fails */
|
2016-02-08 20:51:12 +07:00
|
|
|
if (user_tsc_khz == 0) {
|
2015-10-20 14:39:02 +07:00
|
|
|
/* set tsc_scaling_ratio to a safe value */
|
|
|
|
vcpu->arch.tsc_scaling_ratio = kvm_default_tsc_scaling_ratio;
|
2015-10-20 14:39:04 +07:00
|
|
|
return -1;
|
2015-10-20 14:39:02 +07:00
|
|
|
}
|
2013-03-12 09:10:24 +07:00
|
|
|
|
2010-09-19 07:38:15 +07:00
|
|
|
/* Compute a scale to convert nanoseconds in TSC cycles */
|
2016-02-08 21:11:15 +07:00
|
|
|
kvm_get_time_scale(user_tsc_khz * 1000LL, NSEC_PER_SEC,
|
KVM: Infrastructure for software and hardware based TSC rate scaling
This requires some restructuring; rather than use 'virtual_tsc_khz'
to indicate whether hardware rate scaling is in effect, we consider
each VCPU to always have a virtual TSC rate. Instead, there is new
logic above the vendor-specific hardware scaling that decides whether
it is even necessary to use and updates all rate variables used by
common code. This means we can simply query the virtual rate at
any point, which is needed for software rate scaling.
There is also now a threshold added to the TSC rate scaling; minor
differences and variations of measured TSC rate can accidentally
provoke rate scaling to be used when it is not needed. Instead,
we have a tolerance variable called tsc_tolerance_ppm, which is
the maximum variation from user requested rate at which scaling
will be used. The default is 250ppm, which is the half the
threshold for NTP adjustment, allowing for some hardware variation.
In the event that hardware rate scaling is not available, we can
kludge a bit by forcing TSC catchup to turn on when a faster than
hardware speed has been requested, but there is nothing available
yet for the reverse case; this requires a trap and emulate software
implementation for RDTSC, which is still forthcoming.
[avi: fix 64-bit division on i386]
Signed-off-by: Zachary Amsden <zamsden@gmail.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
2012-02-04 00:43:50 +07:00
|
|
|
&vcpu->arch.virtual_tsc_shift,
|
|
|
|
&vcpu->arch.virtual_tsc_mult);
|
2016-02-08 20:51:12 +07:00
|
|
|
vcpu->arch.virtual_tsc_khz = user_tsc_khz;
|
KVM: Infrastructure for software and hardware based TSC rate scaling
This requires some restructuring; rather than use 'virtual_tsc_khz'
to indicate whether hardware rate scaling is in effect, we consider
each VCPU to always have a virtual TSC rate. Instead, there is new
logic above the vendor-specific hardware scaling that decides whether
it is even necessary to use and updates all rate variables used by
common code. This means we can simply query the virtual rate at
any point, which is needed for software rate scaling.
There is also now a threshold added to the TSC rate scaling; minor
differences and variations of measured TSC rate can accidentally
provoke rate scaling to be used when it is not needed. Instead,
we have a tolerance variable called tsc_tolerance_ppm, which is
the maximum variation from user requested rate at which scaling
will be used. The default is 250ppm, which is the half the
threshold for NTP adjustment, allowing for some hardware variation.
In the event that hardware rate scaling is not available, we can
kludge a bit by forcing TSC catchup to turn on when a faster than
hardware speed has been requested, but there is nothing available
yet for the reverse case; this requires a trap and emulate software
implementation for RDTSC, which is still forthcoming.
[avi: fix 64-bit division on i386]
Signed-off-by: Zachary Amsden <zamsden@gmail.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
2012-02-04 00:43:50 +07:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Compute the variation in TSC rate which is acceptable
|
|
|
|
* within the range of tolerance and decide if the
|
|
|
|
* rate being applied is within that bounds of the hardware
|
|
|
|
* rate. If so, no scaling or compensation need be done.
|
|
|
|
*/
|
|
|
|
thresh_lo = adjust_tsc_khz(tsc_khz, -tsc_tolerance_ppm);
|
|
|
|
thresh_hi = adjust_tsc_khz(tsc_khz, tsc_tolerance_ppm);
|
2016-02-08 20:51:12 +07:00
|
|
|
if (user_tsc_khz < thresh_lo || user_tsc_khz > thresh_hi) {
|
|
|
|
pr_debug("kvm: requested TSC rate %u falls outside tolerance [%u,%u]\n", user_tsc_khz, thresh_lo, thresh_hi);
|
KVM: Infrastructure for software and hardware based TSC rate scaling
This requires some restructuring; rather than use 'virtual_tsc_khz'
to indicate whether hardware rate scaling is in effect, we consider
each VCPU to always have a virtual TSC rate. Instead, there is new
logic above the vendor-specific hardware scaling that decides whether
it is even necessary to use and updates all rate variables used by
common code. This means we can simply query the virtual rate at
any point, which is needed for software rate scaling.
There is also now a threshold added to the TSC rate scaling; minor
differences and variations of measured TSC rate can accidentally
provoke rate scaling to be used when it is not needed. Instead,
we have a tolerance variable called tsc_tolerance_ppm, which is
the maximum variation from user requested rate at which scaling
will be used. The default is 250ppm, which is the half the
threshold for NTP adjustment, allowing for some hardware variation.
In the event that hardware rate scaling is not available, we can
kludge a bit by forcing TSC catchup to turn on when a faster than
hardware speed has been requested, but there is nothing available
yet for the reverse case; this requires a trap and emulate software
implementation for RDTSC, which is still forthcoming.
[avi: fix 64-bit division on i386]
Signed-off-by: Zachary Amsden <zamsden@gmail.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
2012-02-04 00:43:50 +07:00
|
|
|
use_scaling = 1;
|
|
|
|
}
|
2016-02-08 20:51:12 +07:00
|
|
|
return set_tsc_khz(vcpu, user_tsc_khz, use_scaling);
|
2010-09-19 07:38:15 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
static u64 compute_guest_tsc(struct kvm_vcpu *vcpu, s64 kernel_ns)
|
|
|
|
{
|
2012-02-04 00:43:57 +07:00
|
|
|
u64 tsc = pvclock_scale_delta(kernel_ns-vcpu->arch.this_tsc_nsec,
|
KVM: Infrastructure for software and hardware based TSC rate scaling
This requires some restructuring; rather than use 'virtual_tsc_khz'
to indicate whether hardware rate scaling is in effect, we consider
each VCPU to always have a virtual TSC rate. Instead, there is new
logic above the vendor-specific hardware scaling that decides whether
it is even necessary to use and updates all rate variables used by
common code. This means we can simply query the virtual rate at
any point, which is needed for software rate scaling.
There is also now a threshold added to the TSC rate scaling; minor
differences and variations of measured TSC rate can accidentally
provoke rate scaling to be used when it is not needed. Instead,
we have a tolerance variable called tsc_tolerance_ppm, which is
the maximum variation from user requested rate at which scaling
will be used. The default is 250ppm, which is the half the
threshold for NTP adjustment, allowing for some hardware variation.
In the event that hardware rate scaling is not available, we can
kludge a bit by forcing TSC catchup to turn on when a faster than
hardware speed has been requested, but there is nothing available
yet for the reverse case; this requires a trap and emulate software
implementation for RDTSC, which is still forthcoming.
[avi: fix 64-bit division on i386]
Signed-off-by: Zachary Amsden <zamsden@gmail.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
2012-02-04 00:43:50 +07:00
|
|
|
vcpu->arch.virtual_tsc_mult,
|
|
|
|
vcpu->arch.virtual_tsc_shift);
|
2012-02-04 00:43:57 +07:00
|
|
|
tsc += vcpu->arch.this_tsc_write;
|
2010-09-19 07:38:15 +07:00
|
|
|
return tsc;
|
|
|
|
}
|
|
|
|
|
2015-01-19 21:33:39 +07:00
|
|
|
static void kvm_track_tsc_matching(struct kvm_vcpu *vcpu)
|
2012-11-28 08:29:03 +07:00
|
|
|
{
|
|
|
|
#ifdef CONFIG_X86_64
|
|
|
|
bool vcpus_matched;
|
|
|
|
struct kvm_arch *ka = &vcpu->kvm->arch;
|
|
|
|
struct pvclock_gtod_data *gtod = &pvclock_gtod_data;
|
|
|
|
|
|
|
|
vcpus_matched = (ka->nr_vcpus_matched_tsc + 1 ==
|
|
|
|
atomic_read(&vcpu->kvm->online_vcpus));
|
|
|
|
|
2014-11-05 06:30:44 +07:00
|
|
|
/*
|
|
|
|
* Once the masterclock is enabled, always perform request in
|
|
|
|
* order to update it.
|
|
|
|
*
|
|
|
|
* In order to enable masterclock, the host clocksource must be TSC
|
|
|
|
* and the vcpus need to have matched TSCs. When that happens,
|
|
|
|
* perform request to enable masterclock.
|
|
|
|
*/
|
|
|
|
if (ka->use_master_clock ||
|
|
|
|
(gtod->clock.vclock_mode == VCLOCK_TSC && vcpus_matched))
|
2012-11-28 08:29:03 +07:00
|
|
|
kvm_make_request(KVM_REQ_MASTERCLOCK_UPDATE, vcpu);
|
|
|
|
|
|
|
|
trace_kvm_track_tsc(vcpu->vcpu_id, ka->nr_vcpus_matched_tsc,
|
|
|
|
atomic_read(&vcpu->kvm->online_vcpus),
|
|
|
|
ka->use_master_clock, gtod->clock.vclock_mode);
|
|
|
|
#endif
|
|
|
|
}
|
|
|
|
|
2012-11-30 03:42:50 +07:00
|
|
|
static void update_ia32_tsc_adjust_msr(struct kvm_vcpu *vcpu, s64 offset)
|
|
|
|
{
|
|
|
|
u64 curr_offset = kvm_x86_ops->read_tsc_offset(vcpu);
|
|
|
|
vcpu->arch.ia32_tsc_adjust_msr += offset - curr_offset;
|
|
|
|
}
|
|
|
|
|
2015-10-20 14:39:03 +07:00
|
|
|
/*
|
|
|
|
* Multiply tsc by a fixed point number represented by ratio.
|
|
|
|
*
|
|
|
|
* The most significant 64-N bits (mult) of ratio represent the
|
|
|
|
* integral part of the fixed point number; the remaining N bits
|
|
|
|
* (frac) represent the fractional part, ie. ratio represents a fixed
|
|
|
|
* point number (mult + frac * 2^(-N)).
|
|
|
|
*
|
|
|
|
* N equals to kvm_tsc_scaling_ratio_frac_bits.
|
|
|
|
*/
|
|
|
|
static inline u64 __scale_tsc(u64 ratio, u64 tsc)
|
|
|
|
{
|
|
|
|
return mul_u64_u64_shr(tsc, ratio, kvm_tsc_scaling_ratio_frac_bits);
|
|
|
|
}
|
|
|
|
|
|
|
|
u64 kvm_scale_tsc(struct kvm_vcpu *vcpu, u64 tsc)
|
|
|
|
{
|
|
|
|
u64 _tsc = tsc;
|
|
|
|
u64 ratio = vcpu->arch.tsc_scaling_ratio;
|
|
|
|
|
|
|
|
if (ratio != kvm_default_tsc_scaling_ratio)
|
|
|
|
_tsc = __scale_tsc(ratio, tsc);
|
|
|
|
|
|
|
|
return _tsc;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(kvm_scale_tsc);
|
|
|
|
|
2015-10-20 14:39:05 +07:00
|
|
|
static u64 kvm_compute_tsc_offset(struct kvm_vcpu *vcpu, u64 target_tsc)
|
|
|
|
{
|
|
|
|
u64 tsc;
|
|
|
|
|
|
|
|
tsc = kvm_scale_tsc(vcpu, rdtsc());
|
|
|
|
|
|
|
|
return target_tsc - tsc;
|
|
|
|
}
|
|
|
|
|
2015-10-20 14:39:07 +07:00
|
|
|
u64 kvm_read_l1_tsc(struct kvm_vcpu *vcpu, u64 host_tsc)
|
|
|
|
{
|
|
|
|
return kvm_x86_ops->read_l1_tsc(vcpu, kvm_scale_tsc(vcpu, host_tsc));
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(kvm_read_l1_tsc);
|
|
|
|
|
2012-11-30 03:42:12 +07:00
|
|
|
void kvm_write_tsc(struct kvm_vcpu *vcpu, struct msr_data *msr)
|
2010-08-20 15:07:17 +07:00
|
|
|
{
|
|
|
|
struct kvm *kvm = vcpu->kvm;
|
2010-08-20 15:07:20 +07:00
|
|
|
u64 offset, ns, elapsed;
|
2010-08-20 15:07:17 +07:00
|
|
|
unsigned long flags;
|
2012-03-09 04:46:57 +07:00
|
|
|
s64 usdiff;
|
2012-11-28 08:29:03 +07:00
|
|
|
bool matched;
|
2014-06-24 14:42:43 +07:00
|
|
|
bool already_matched;
|
2012-11-30 03:42:12 +07:00
|
|
|
u64 data = msr->data;
|
2010-08-20 15:07:17 +07:00
|
|
|
|
2011-02-04 16:49:11 +07:00
|
|
|
raw_spin_lock_irqsave(&kvm->arch.tsc_write_lock, flags);
|
2015-10-20 14:39:05 +07:00
|
|
|
offset = kvm_compute_tsc_offset(vcpu, data);
|
2010-08-20 15:07:25 +07:00
|
|
|
ns = get_kernel_ns();
|
2010-08-20 15:07:20 +07:00
|
|
|
elapsed = ns - kvm->arch.last_tsc_nsec;
|
KVM: Improve TSC offset matching
There are a few improvements that can be made to the TSC offset
matching code. First, we don't need to call the 128-bit multiply
(especially on a constant number), the code works much nicer to
do computation in nanosecond units.
Second, the way everything is setup with software TSC rate scaling,
we currently have per-cpu rates. Obviously this isn't too desirable
to use in practice, but if for some reason we do change the rate of
all VCPUs at runtime, then reset the TSCs, we will only want to
match offsets for VCPUs running at the same rate.
Finally, for the case where we have an unstable host TSC, but
rate scaling is being done in hardware, we should call the platform
code to compute the TSC offset, so the math is reorganized to recompute
the base instead, then transform the base into an offset using the
existing API.
[avi: fix 64-bit division on i386]
Signed-off-by: Zachary Amsden <zamsden@gmail.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
KVM: Fix 64-bit division in kvm_write_tsc()
Breaks i386 build.
Signed-off-by: Avi Kivity <avi@redhat.com>
2012-02-04 00:43:51 +07:00
|
|
|
|
2013-03-12 09:10:24 +07:00
|
|
|
if (vcpu->arch.virtual_tsc_khz) {
|
2013-06-12 09:31:12 +07:00
|
|
|
int faulted = 0;
|
|
|
|
|
2013-03-12 09:10:24 +07:00
|
|
|
/* n.b - signed multiplication and division required */
|
|
|
|
usdiff = data - kvm->arch.last_tsc_write;
|
KVM: Improve TSC offset matching
There are a few improvements that can be made to the TSC offset
matching code. First, we don't need to call the 128-bit multiply
(especially on a constant number), the code works much nicer to
do computation in nanosecond units.
Second, the way everything is setup with software TSC rate scaling,
we currently have per-cpu rates. Obviously this isn't too desirable
to use in practice, but if for some reason we do change the rate of
all VCPUs at runtime, then reset the TSCs, we will only want to
match offsets for VCPUs running at the same rate.
Finally, for the case where we have an unstable host TSC, but
rate scaling is being done in hardware, we should call the platform
code to compute the TSC offset, so the math is reorganized to recompute
the base instead, then transform the base into an offset using the
existing API.
[avi: fix 64-bit division on i386]
Signed-off-by: Zachary Amsden <zamsden@gmail.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
KVM: Fix 64-bit division in kvm_write_tsc()
Breaks i386 build.
Signed-off-by: Avi Kivity <avi@redhat.com>
2012-02-04 00:43:51 +07:00
|
|
|
#ifdef CONFIG_X86_64
|
2013-03-12 09:10:24 +07:00
|
|
|
usdiff = (usdiff * 1000) / vcpu->arch.virtual_tsc_khz;
|
KVM: Improve TSC offset matching
There are a few improvements that can be made to the TSC offset
matching code. First, we don't need to call the 128-bit multiply
(especially on a constant number), the code works much nicer to
do computation in nanosecond units.
Second, the way everything is setup with software TSC rate scaling,
we currently have per-cpu rates. Obviously this isn't too desirable
to use in practice, but if for some reason we do change the rate of
all VCPUs at runtime, then reset the TSCs, we will only want to
match offsets for VCPUs running at the same rate.
Finally, for the case where we have an unstable host TSC, but
rate scaling is being done in hardware, we should call the platform
code to compute the TSC offset, so the math is reorganized to recompute
the base instead, then transform the base into an offset using the
existing API.
[avi: fix 64-bit division on i386]
Signed-off-by: Zachary Amsden <zamsden@gmail.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
KVM: Fix 64-bit division in kvm_write_tsc()
Breaks i386 build.
Signed-off-by: Avi Kivity <avi@redhat.com>
2012-02-04 00:43:51 +07:00
|
|
|
#else
|
2013-03-12 09:10:24 +07:00
|
|
|
/* do_div() only does unsigned */
|
2013-06-12 09:31:12 +07:00
|
|
|
asm("1: idivl %[divisor]\n"
|
|
|
|
"2: xor %%edx, %%edx\n"
|
|
|
|
" movl $0, %[faulted]\n"
|
|
|
|
"3:\n"
|
|
|
|
".section .fixup,\"ax\"\n"
|
|
|
|
"4: movl $1, %[faulted]\n"
|
|
|
|
" jmp 3b\n"
|
|
|
|
".previous\n"
|
|
|
|
|
|
|
|
_ASM_EXTABLE(1b, 4b)
|
|
|
|
|
|
|
|
: "=A"(usdiff), [faulted] "=r" (faulted)
|
|
|
|
: "A"(usdiff * 1000), [divisor] "rm"(vcpu->arch.virtual_tsc_khz));
|
|
|
|
|
KVM: Improve TSC offset matching
There are a few improvements that can be made to the TSC offset
matching code. First, we don't need to call the 128-bit multiply
(especially on a constant number), the code works much nicer to
do computation in nanosecond units.
Second, the way everything is setup with software TSC rate scaling,
we currently have per-cpu rates. Obviously this isn't too desirable
to use in practice, but if for some reason we do change the rate of
all VCPUs at runtime, then reset the TSCs, we will only want to
match offsets for VCPUs running at the same rate.
Finally, for the case where we have an unstable host TSC, but
rate scaling is being done in hardware, we should call the platform
code to compute the TSC offset, so the math is reorganized to recompute
the base instead, then transform the base into an offset using the
existing API.
[avi: fix 64-bit division on i386]
Signed-off-by: Zachary Amsden <zamsden@gmail.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
KVM: Fix 64-bit division in kvm_write_tsc()
Breaks i386 build.
Signed-off-by: Avi Kivity <avi@redhat.com>
2012-02-04 00:43:51 +07:00
|
|
|
#endif
|
2013-03-12 09:10:24 +07:00
|
|
|
do_div(elapsed, 1000);
|
|
|
|
usdiff -= elapsed;
|
|
|
|
if (usdiff < 0)
|
|
|
|
usdiff = -usdiff;
|
2013-06-12 09:31:12 +07:00
|
|
|
|
|
|
|
/* idivl overflow => difference is larger than USEC_PER_SEC */
|
|
|
|
if (faulted)
|
|
|
|
usdiff = USEC_PER_SEC;
|
2013-03-12 09:10:24 +07:00
|
|
|
} else
|
|
|
|
usdiff = USEC_PER_SEC; /* disable TSC match window below */
|
2010-08-20 15:07:20 +07:00
|
|
|
|
|
|
|
/*
|
KVM: Improve TSC offset matching
There are a few improvements that can be made to the TSC offset
matching code. First, we don't need to call the 128-bit multiply
(especially on a constant number), the code works much nicer to
do computation in nanosecond units.
Second, the way everything is setup with software TSC rate scaling,
we currently have per-cpu rates. Obviously this isn't too desirable
to use in practice, but if for some reason we do change the rate of
all VCPUs at runtime, then reset the TSCs, we will only want to
match offsets for VCPUs running at the same rate.
Finally, for the case where we have an unstable host TSC, but
rate scaling is being done in hardware, we should call the platform
code to compute the TSC offset, so the math is reorganized to recompute
the base instead, then transform the base into an offset using the
existing API.
[avi: fix 64-bit division on i386]
Signed-off-by: Zachary Amsden <zamsden@gmail.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
KVM: Fix 64-bit division in kvm_write_tsc()
Breaks i386 build.
Signed-off-by: Avi Kivity <avi@redhat.com>
2012-02-04 00:43:51 +07:00
|
|
|
* Special case: TSC write with a small delta (1 second) of virtual
|
|
|
|
* cycle time against real time is interpreted as an attempt to
|
|
|
|
* synchronize the CPU.
|
|
|
|
*
|
|
|
|
* For a reliable TSC, we can match TSC offsets, and for an unstable
|
|
|
|
* TSC, we add elapsed time in this computation. We could let the
|
|
|
|
* compensation code attempt to catch up if we fall behind, but
|
|
|
|
* it's better to try to match offsets from the beginning.
|
|
|
|
*/
|
2012-03-09 04:46:57 +07:00
|
|
|
if (usdiff < USEC_PER_SEC &&
|
KVM: Improve TSC offset matching
There are a few improvements that can be made to the TSC offset
matching code. First, we don't need to call the 128-bit multiply
(especially on a constant number), the code works much nicer to
do computation in nanosecond units.
Second, the way everything is setup with software TSC rate scaling,
we currently have per-cpu rates. Obviously this isn't too desirable
to use in practice, but if for some reason we do change the rate of
all VCPUs at runtime, then reset the TSCs, we will only want to
match offsets for VCPUs running at the same rate.
Finally, for the case where we have an unstable host TSC, but
rate scaling is being done in hardware, we should call the platform
code to compute the TSC offset, so the math is reorganized to recompute
the base instead, then transform the base into an offset using the
existing API.
[avi: fix 64-bit division on i386]
Signed-off-by: Zachary Amsden <zamsden@gmail.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
KVM: Fix 64-bit division in kvm_write_tsc()
Breaks i386 build.
Signed-off-by: Avi Kivity <avi@redhat.com>
2012-02-04 00:43:51 +07:00
|
|
|
vcpu->arch.virtual_tsc_khz == kvm->arch.last_tsc_khz) {
|
2010-08-20 15:07:20 +07:00
|
|
|
if (!check_tsc_unstable()) {
|
2012-02-04 00:43:57 +07:00
|
|
|
offset = kvm->arch.cur_tsc_offset;
|
2010-08-20 15:07:20 +07:00
|
|
|
pr_debug("kvm: matched tsc offset for %llu\n", data);
|
|
|
|
} else {
|
2011-03-25 15:44:50 +07:00
|
|
|
u64 delta = nsec_to_cycles(vcpu, elapsed);
|
KVM: Improve TSC offset matching
There are a few improvements that can be made to the TSC offset
matching code. First, we don't need to call the 128-bit multiply
(especially on a constant number), the code works much nicer to
do computation in nanosecond units.
Second, the way everything is setup with software TSC rate scaling,
we currently have per-cpu rates. Obviously this isn't too desirable
to use in practice, but if for some reason we do change the rate of
all VCPUs at runtime, then reset the TSCs, we will only want to
match offsets for VCPUs running at the same rate.
Finally, for the case where we have an unstable host TSC, but
rate scaling is being done in hardware, we should call the platform
code to compute the TSC offset, so the math is reorganized to recompute
the base instead, then transform the base into an offset using the
existing API.
[avi: fix 64-bit division on i386]
Signed-off-by: Zachary Amsden <zamsden@gmail.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
KVM: Fix 64-bit division in kvm_write_tsc()
Breaks i386 build.
Signed-off-by: Avi Kivity <avi@redhat.com>
2012-02-04 00:43:51 +07:00
|
|
|
data += delta;
|
2015-10-20 14:39:05 +07:00
|
|
|
offset = kvm_compute_tsc_offset(vcpu, data);
|
2010-08-20 15:07:25 +07:00
|
|
|
pr_debug("kvm: adjusted tsc offset by %llu\n", delta);
|
2010-08-20 15:07:20 +07:00
|
|
|
}
|
2012-11-28 08:29:03 +07:00
|
|
|
matched = true;
|
2014-06-24 14:42:43 +07:00
|
|
|
already_matched = (vcpu->arch.this_tsc_generation == kvm->arch.cur_tsc_generation);
|
2012-02-04 00:43:57 +07:00
|
|
|
} else {
|
|
|
|
/*
|
|
|
|
* We split periods of matched TSC writes into generations.
|
|
|
|
* For each generation, we track the original measured
|
|
|
|
* nanosecond time, offset, and write, so if TSCs are in
|
|
|
|
* sync, we can match exact offset, and if not, we can match
|
2012-06-28 14:17:27 +07:00
|
|
|
* exact software computation in compute_guest_tsc()
|
2012-02-04 00:43:57 +07:00
|
|
|
*
|
|
|
|
* These values are tracked in kvm->arch.cur_xxx variables.
|
|
|
|
*/
|
|
|
|
kvm->arch.cur_tsc_generation++;
|
|
|
|
kvm->arch.cur_tsc_nsec = ns;
|
|
|
|
kvm->arch.cur_tsc_write = data;
|
|
|
|
kvm->arch.cur_tsc_offset = offset;
|
2012-11-28 08:29:03 +07:00
|
|
|
matched = false;
|
2014-06-24 14:42:43 +07:00
|
|
|
pr_debug("kvm: new tsc generation %llu, clock %llu\n",
|
2012-02-04 00:43:57 +07:00
|
|
|
kvm->arch.cur_tsc_generation, data);
|
2010-08-20 15:07:20 +07:00
|
|
|
}
|
2012-02-04 00:43:57 +07:00
|
|
|
|
|
|
|
/*
|
|
|
|
* We also track th most recent recorded KHZ, write and time to
|
|
|
|
* allow the matching interval to be extended at each write.
|
|
|
|
*/
|
2010-08-20 15:07:20 +07:00
|
|
|
kvm->arch.last_tsc_nsec = ns;
|
|
|
|
kvm->arch.last_tsc_write = data;
|
KVM: Improve TSC offset matching
There are a few improvements that can be made to the TSC offset
matching code. First, we don't need to call the 128-bit multiply
(especially on a constant number), the code works much nicer to
do computation in nanosecond units.
Second, the way everything is setup with software TSC rate scaling,
we currently have per-cpu rates. Obviously this isn't too desirable
to use in practice, but if for some reason we do change the rate of
all VCPUs at runtime, then reset the TSCs, we will only want to
match offsets for VCPUs running at the same rate.
Finally, for the case where we have an unstable host TSC, but
rate scaling is being done in hardware, we should call the platform
code to compute the TSC offset, so the math is reorganized to recompute
the base instead, then transform the base into an offset using the
existing API.
[avi: fix 64-bit division on i386]
Signed-off-by: Zachary Amsden <zamsden@gmail.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
KVM: Fix 64-bit division in kvm_write_tsc()
Breaks i386 build.
Signed-off-by: Avi Kivity <avi@redhat.com>
2012-02-04 00:43:51 +07:00
|
|
|
kvm->arch.last_tsc_khz = vcpu->arch.virtual_tsc_khz;
|
2010-08-20 15:07:17 +07:00
|
|
|
|
KVM: Fix last_guest_tsc / tsc_offset semantics
The variable last_guest_tsc was being used as an ad-hoc indicator
that guest TSC has been initialized and recorded correctly. However,
it may not have been, it could be that guest TSC has been set to some
large value, the back to a small value (by, say, a software reboot).
This defeats the logic and causes KVM to falsely assume that the
guest TSC has gone backwards, marking the host TSC unstable, which
is undesirable behavior.
In addition, rather than try to compute an offset adjustment for the
TSC on unstable platforms, just recompute the whole offset. This
allows us to get rid of one callsite for adjust_tsc_offset, which
is problematic because the units it takes are in guest units, but
here, the computation was originally being done in host units.
Doing this, and also recording last_guest_tsc when the TSC is written
allow us to remove the tricky logic which depended on last_guest_tsc
being zero to indicate a reset of uninitialized value.
Instead, we now have the guarantee that the guest TSC offset is
always at least something which will get us last_guest_tsc.
Signed-off-by: Zachary Amsden <zamsden@gmail.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
2012-02-04 00:43:53 +07:00
|
|
|
vcpu->arch.last_guest_tsc = data;
|
2012-02-04 00:43:57 +07:00
|
|
|
|
|
|
|
/* Keep track of which generation this VCPU has synchronized to */
|
|
|
|
vcpu->arch.this_tsc_generation = kvm->arch.cur_tsc_generation;
|
|
|
|
vcpu->arch.this_tsc_nsec = kvm->arch.cur_tsc_nsec;
|
|
|
|
vcpu->arch.this_tsc_write = kvm->arch.cur_tsc_write;
|
|
|
|
|
2012-11-30 03:42:50 +07:00
|
|
|
if (guest_cpuid_has_tsc_adjust(vcpu) && !msr->host_initiated)
|
|
|
|
update_ia32_tsc_adjust_msr(vcpu, offset);
|
2012-02-04 00:43:57 +07:00
|
|
|
kvm_x86_ops->write_tsc_offset(vcpu, offset);
|
|
|
|
raw_spin_unlock_irqrestore(&kvm->arch.tsc_write_lock, flags);
|
2012-11-28 08:29:03 +07:00
|
|
|
|
|
|
|
spin_lock(&kvm->arch.pvclock_gtod_sync_lock);
|
2014-06-24 14:42:43 +07:00
|
|
|
if (!matched) {
|
2012-11-28 08:29:03 +07:00
|
|
|
kvm->arch.nr_vcpus_matched_tsc = 0;
|
2014-06-24 14:42:43 +07:00
|
|
|
} else if (!already_matched) {
|
|
|
|
kvm->arch.nr_vcpus_matched_tsc++;
|
|
|
|
}
|
2012-11-28 08:29:03 +07:00
|
|
|
|
|
|
|
kvm_track_tsc_matching(vcpu);
|
|
|
|
spin_unlock(&kvm->arch.pvclock_gtod_sync_lock);
|
2010-08-20 15:07:17 +07:00
|
|
|
}
|
2012-02-04 00:43:57 +07:00
|
|
|
|
2010-08-20 15:07:17 +07:00
|
|
|
EXPORT_SYMBOL_GPL(kvm_write_tsc);
|
|
|
|
|
2015-10-20 14:39:06 +07:00
|
|
|
static inline void adjust_tsc_offset_guest(struct kvm_vcpu *vcpu,
|
|
|
|
s64 adjustment)
|
|
|
|
{
|
|
|
|
kvm_x86_ops->adjust_tsc_offset_guest(vcpu, adjustment);
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline void adjust_tsc_offset_host(struct kvm_vcpu *vcpu, s64 adjustment)
|
|
|
|
{
|
|
|
|
if (vcpu->arch.tsc_scaling_ratio != kvm_default_tsc_scaling_ratio)
|
|
|
|
WARN_ON(adjustment < 0);
|
|
|
|
adjustment = kvm_scale_tsc(vcpu, (u64) adjustment);
|
|
|
|
kvm_x86_ops->adjust_tsc_offset_guest(vcpu, adjustment);
|
|
|
|
}
|
|
|
|
|
2012-11-28 08:29:01 +07:00
|
|
|
#ifdef CONFIG_X86_64
|
|
|
|
|
|
|
|
static cycle_t read_tsc(void)
|
|
|
|
{
|
2015-06-25 23:44:08 +07:00
|
|
|
cycle_t ret = (cycle_t)rdtsc_ordered();
|
|
|
|
u64 last = pvclock_gtod_data.clock.cycle_last;
|
2012-11-28 08:29:01 +07:00
|
|
|
|
|
|
|
if (likely(ret >= last))
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* GCC likes to generate cmov here, but this branch is extremely
|
2016-02-24 06:34:30 +07:00
|
|
|
* predictable (it's just a function of time and the likely is
|
2012-11-28 08:29:01 +07:00
|
|
|
* very likely) and there's a data dependence, so force GCC
|
|
|
|
* to generate a branch instead. I don't barrier() because
|
|
|
|
* we don't actually need a barrier, and if this function
|
|
|
|
* ever gets inlined it will generate worse code.
|
|
|
|
*/
|
|
|
|
asm volatile ("");
|
|
|
|
return last;
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline u64 vgettsc(cycle_t *cycle_now)
|
|
|
|
{
|
|
|
|
long v;
|
|
|
|
struct pvclock_gtod_data *gtod = &pvclock_gtod_data;
|
|
|
|
|
|
|
|
*cycle_now = read_tsc();
|
|
|
|
|
|
|
|
v = (*cycle_now - gtod->clock.cycle_last) & gtod->clock.mask;
|
|
|
|
return v * gtod->clock.mult;
|
|
|
|
}
|
|
|
|
|
2014-07-17 04:04:54 +07:00
|
|
|
static int do_monotonic_boot(s64 *t, cycle_t *cycle_now)
|
2012-11-28 08:29:01 +07:00
|
|
|
{
|
2014-07-17 04:04:54 +07:00
|
|
|
struct pvclock_gtod_data *gtod = &pvclock_gtod_data;
|
2012-11-28 08:29:01 +07:00
|
|
|
unsigned long seq;
|
|
|
|
int mode;
|
2014-07-17 04:04:54 +07:00
|
|
|
u64 ns;
|
2012-11-28 08:29:01 +07:00
|
|
|
|
|
|
|
do {
|
|
|
|
seq = read_seqcount_begin(>od->seq);
|
|
|
|
mode = gtod->clock.vclock_mode;
|
2014-07-17 04:04:54 +07:00
|
|
|
ns = gtod->nsec_base;
|
2012-11-28 08:29:01 +07:00
|
|
|
ns += vgettsc(cycle_now);
|
|
|
|
ns >>= gtod->clock.shift;
|
2014-07-17 04:04:54 +07:00
|
|
|
ns += gtod->boot_ns;
|
2012-11-28 08:29:01 +07:00
|
|
|
} while (unlikely(read_seqcount_retry(>od->seq, seq)));
|
2014-07-17 04:04:54 +07:00
|
|
|
*t = ns;
|
2012-11-28 08:29:01 +07:00
|
|
|
|
|
|
|
return mode;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* returns true if host is using tsc clocksource */
|
|
|
|
static bool kvm_get_time_and_clockread(s64 *kernel_ns, cycle_t *cycle_now)
|
|
|
|
{
|
|
|
|
/* checked again under seqlock below */
|
|
|
|
if (pvclock_gtod_data.clock.vclock_mode != VCLOCK_TSC)
|
|
|
|
return false;
|
|
|
|
|
2014-07-17 04:04:54 +07:00
|
|
|
return do_monotonic_boot(kernel_ns, cycle_now) == VCLOCK_TSC;
|
2012-11-28 08:29:01 +07:00
|
|
|
}
|
|
|
|
#endif
|
|
|
|
|
|
|
|
/*
|
|
|
|
*
|
2012-11-28 08:29:03 +07:00
|
|
|
* Assuming a stable TSC across physical CPUS, and a stable TSC
|
|
|
|
* across virtual CPUs, the following condition is possible.
|
|
|
|
* Each numbered line represents an event visible to both
|
2012-11-28 08:29:01 +07:00
|
|
|
* CPUs at the next numbered event.
|
|
|
|
*
|
|
|
|
* "timespecX" represents host monotonic time. "tscX" represents
|
|
|
|
* RDTSC value.
|
|
|
|
*
|
|
|
|
* VCPU0 on CPU0 | VCPU1 on CPU1
|
|
|
|
*
|
|
|
|
* 1. read timespec0,tsc0
|
|
|
|
* 2. | timespec1 = timespec0 + N
|
|
|
|
* | tsc1 = tsc0 + M
|
|
|
|
* 3. transition to guest | transition to guest
|
|
|
|
* 4. ret0 = timespec0 + (rdtsc - tsc0) |
|
|
|
|
* 5. | ret1 = timespec1 + (rdtsc - tsc1)
|
|
|
|
* | ret1 = timespec0 + N + (rdtsc - (tsc0 + M))
|
|
|
|
*
|
|
|
|
* Since ret0 update is visible to VCPU1 at time 5, to obey monotonicity:
|
|
|
|
*
|
|
|
|
* - ret0 < ret1
|
|
|
|
* - timespec0 + (rdtsc - tsc0) < timespec0 + N + (rdtsc - (tsc0 + M))
|
|
|
|
* ...
|
|
|
|
* - 0 < N - M => M < N
|
|
|
|
*
|
|
|
|
* That is, when timespec0 != timespec1, M < N. Unfortunately that is not
|
|
|
|
* always the case (the difference between two distinct xtime instances
|
|
|
|
* might be smaller then the difference between corresponding TSC reads,
|
|
|
|
* when updating guest vcpus pvclock areas).
|
|
|
|
*
|
|
|
|
* To avoid that problem, do not allow visibility of distinct
|
|
|
|
* system_timestamp/tsc_timestamp values simultaneously: use a master
|
|
|
|
* copy of host monotonic time values. Update that master copy
|
|
|
|
* in lockstep.
|
|
|
|
*
|
2012-11-28 08:29:03 +07:00
|
|
|
* Rely on synchronization of host TSCs and guest TSCs for monotonicity.
|
2012-11-28 08:29:01 +07:00
|
|
|
*
|
|
|
|
*/
|
|
|
|
|
|
|
|
static void pvclock_update_vm_gtod_copy(struct kvm *kvm)
|
|
|
|
{
|
|
|
|
#ifdef CONFIG_X86_64
|
|
|
|
struct kvm_arch *ka = &kvm->arch;
|
|
|
|
int vclock_mode;
|
2012-11-28 08:29:03 +07:00
|
|
|
bool host_tsc_clocksource, vcpus_matched;
|
|
|
|
|
|
|
|
vcpus_matched = (ka->nr_vcpus_matched_tsc + 1 ==
|
|
|
|
atomic_read(&kvm->online_vcpus));
|
2012-11-28 08:29:01 +07:00
|
|
|
|
|
|
|
/*
|
|
|
|
* If the host uses TSC clock, then passthrough TSC as stable
|
|
|
|
* to the guest.
|
|
|
|
*/
|
2012-11-28 08:29:03 +07:00
|
|
|
host_tsc_clocksource = kvm_get_time_and_clockread(
|
2012-11-28 08:29:01 +07:00
|
|
|
&ka->master_kernel_ns,
|
|
|
|
&ka->master_cycle_now);
|
|
|
|
|
2014-05-14 22:43:24 +07:00
|
|
|
ka->use_master_clock = host_tsc_clocksource && vcpus_matched
|
2015-01-21 00:54:52 +07:00
|
|
|
&& !backwards_tsc_observed
|
|
|
|
&& !ka->boot_vcpu_runs_old_kvmclock;
|
2012-11-28 08:29:03 +07:00
|
|
|
|
2012-11-28 08:29:01 +07:00
|
|
|
if (ka->use_master_clock)
|
|
|
|
atomic_set(&kvm_guest_has_master_clock, 1);
|
|
|
|
|
|
|
|
vclock_mode = pvclock_gtod_data.clock.vclock_mode;
|
2012-11-28 08:29:03 +07:00
|
|
|
trace_kvm_update_master_clock(ka->use_master_clock, vclock_mode,
|
|
|
|
vcpus_matched);
|
2012-11-28 08:29:01 +07:00
|
|
|
#endif
|
|
|
|
}
|
|
|
|
|
2016-01-07 21:05:10 +07:00
|
|
|
void kvm_make_mclock_inprogress_request(struct kvm *kvm)
|
|
|
|
{
|
|
|
|
kvm_make_all_cpus_request(kvm, KVM_REQ_MCLOCK_INPROGRESS);
|
|
|
|
}
|
|
|
|
|
2013-08-28 09:55:29 +07:00
|
|
|
static void kvm_gen_update_masterclock(struct kvm *kvm)
|
|
|
|
{
|
|
|
|
#ifdef CONFIG_X86_64
|
|
|
|
int i;
|
|
|
|
struct kvm_vcpu *vcpu;
|
|
|
|
struct kvm_arch *ka = &kvm->arch;
|
|
|
|
|
|
|
|
spin_lock(&ka->pvclock_gtod_sync_lock);
|
|
|
|
kvm_make_mclock_inprogress_request(kvm);
|
|
|
|
/* no guest entries from this point */
|
|
|
|
pvclock_update_vm_gtod_copy(kvm);
|
|
|
|
|
|
|
|
kvm_for_each_vcpu(i, vcpu, kvm)
|
2014-09-12 12:43:19 +07:00
|
|
|
kvm_make_request(KVM_REQ_CLOCK_UPDATE, vcpu);
|
2013-08-28 09:55:29 +07:00
|
|
|
|
|
|
|
/* guest entries allowed */
|
|
|
|
kvm_for_each_vcpu(i, vcpu, kvm)
|
|
|
|
clear_bit(KVM_REQ_MCLOCK_INPROGRESS, &vcpu->requests);
|
|
|
|
|
|
|
|
spin_unlock(&ka->pvclock_gtod_sync_lock);
|
|
|
|
#endif
|
|
|
|
}
|
|
|
|
|
2010-09-19 07:38:14 +07:00
|
|
|
static int kvm_guest_time_update(struct kvm_vcpu *v)
|
2008-02-16 02:52:47 +07:00
|
|
|
{
|
2016-02-08 20:51:40 +07:00
|
|
|
unsigned long flags, tgt_tsc_khz;
|
2008-02-16 02:52:47 +07:00
|
|
|
struct kvm_vcpu_arch *vcpu = &v->arch;
|
2012-11-28 08:29:01 +07:00
|
|
|
struct kvm_arch *ka = &v->kvm->arch;
|
2014-01-06 21:18:59 +07:00
|
|
|
s64 kernel_ns;
|
2012-11-28 08:29:01 +07:00
|
|
|
u64 tsc_timestamp, host_tsc;
|
2013-02-21 05:48:10 +07:00
|
|
|
struct pvclock_vcpu_time_info guest_hv_clock;
|
2012-08-04 01:57:49 +07:00
|
|
|
u8 pvclock_flags;
|
2012-11-28 08:29:01 +07:00
|
|
|
bool use_master_clock;
|
|
|
|
|
|
|
|
kernel_ns = 0;
|
|
|
|
host_tsc = 0;
|
2008-02-16 02:52:47 +07:00
|
|
|
|
2012-11-28 08:29:01 +07:00
|
|
|
/*
|
|
|
|
* If the host uses TSC clock, then passthrough TSC as stable
|
|
|
|
* to the guest.
|
|
|
|
*/
|
|
|
|
spin_lock(&ka->pvclock_gtod_sync_lock);
|
|
|
|
use_master_clock = ka->use_master_clock;
|
|
|
|
if (use_master_clock) {
|
|
|
|
host_tsc = ka->master_cycle_now;
|
|
|
|
kernel_ns = ka->master_kernel_ns;
|
|
|
|
}
|
|
|
|
spin_unlock(&ka->pvclock_gtod_sync_lock);
|
2013-03-18 23:54:32 +07:00
|
|
|
|
|
|
|
/* Keep irq disabled to prevent changes to the clock */
|
|
|
|
local_irq_save(flags);
|
2016-02-08 20:51:40 +07:00
|
|
|
tgt_tsc_khz = __this_cpu_read(cpu_tsc_khz);
|
|
|
|
if (unlikely(tgt_tsc_khz == 0)) {
|
2013-03-18 23:54:32 +07:00
|
|
|
local_irq_restore(flags);
|
|
|
|
kvm_make_request(KVM_REQ_CLOCK_UPDATE, v);
|
|
|
|
return 1;
|
|
|
|
}
|
2012-11-28 08:29:01 +07:00
|
|
|
if (!use_master_clock) {
|
2015-06-25 23:44:07 +07:00
|
|
|
host_tsc = rdtsc();
|
2012-11-28 08:29:01 +07:00
|
|
|
kernel_ns = get_kernel_ns();
|
|
|
|
}
|
|
|
|
|
2015-10-20 14:39:07 +07:00
|
|
|
tsc_timestamp = kvm_read_l1_tsc(v, host_tsc);
|
2012-11-28 08:29:01 +07:00
|
|
|
|
2010-09-19 07:38:15 +07:00
|
|
|
/*
|
|
|
|
* We may have to catch up the TSC to match elapsed wall clock
|
|
|
|
* time for two reasons, even if kvmclock is used.
|
|
|
|
* 1) CPU could have been running below the maximum TSC rate
|
|
|
|
* 2) Broken TSC compensation resets the base at each VCPU
|
|
|
|
* entry to avoid unknown leaps of TSC even when running
|
|
|
|
* again on the same CPU. This may cause apparent elapsed
|
|
|
|
* time to disappear, and the guest to stand still or run
|
|
|
|
* very slowly.
|
|
|
|
*/
|
|
|
|
if (vcpu->tsc_catchup) {
|
|
|
|
u64 tsc = compute_guest_tsc(v, kernel_ns);
|
|
|
|
if (tsc > tsc_timestamp) {
|
2012-02-04 00:43:55 +07:00
|
|
|
adjust_tsc_offset_guest(v, tsc - tsc_timestamp);
|
2010-09-19 07:38:15 +07:00
|
|
|
tsc_timestamp = tsc;
|
|
|
|
}
|
2008-06-03 21:17:31 +07:00
|
|
|
}
|
|
|
|
|
2008-02-16 02:52:47 +07:00
|
|
|
local_irq_restore(flags);
|
|
|
|
|
2013-02-21 05:48:10 +07:00
|
|
|
if (!vcpu->pv_time_enabled)
|
2010-09-19 07:38:15 +07:00
|
|
|
return 0;
|
2008-02-16 02:52:47 +07:00
|
|
|
|
2016-02-08 20:51:40 +07:00
|
|
|
if (kvm_has_tsc_control)
|
|
|
|
tgt_tsc_khz = kvm_scale_tsc(v, tgt_tsc_khz);
|
|
|
|
|
|
|
|
if (unlikely(vcpu->hw_tsc_khz != tgt_tsc_khz)) {
|
2016-02-08 21:11:15 +07:00
|
|
|
kvm_get_time_scale(NSEC_PER_SEC, tgt_tsc_khz * 1000LL,
|
2010-09-19 07:38:13 +07:00
|
|
|
&vcpu->hv_clock.tsc_shift,
|
|
|
|
&vcpu->hv_clock.tsc_to_system_mul);
|
2016-02-08 20:51:40 +07:00
|
|
|
vcpu->hw_tsc_khz = tgt_tsc_khz;
|
2010-08-20 15:07:21 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
/* With all the info we got, fill in the values */
|
2010-08-20 15:07:30 +07:00
|
|
|
vcpu->hv_clock.tsc_timestamp = tsc_timestamp;
|
2010-08-20 15:07:25 +07:00
|
|
|
vcpu->hv_clock.system_time = kernel_ns + v->kvm->arch.kvmclock_offset;
|
2010-09-19 07:38:12 +07:00
|
|
|
vcpu->last_guest_tsc = tsc_timestamp;
|
2012-08-04 01:57:49 +07:00
|
|
|
|
2014-11-04 07:57:18 +07:00
|
|
|
if (unlikely(kvm_read_guest_cached(v->kvm, &vcpu->pv_time,
|
|
|
|
&guest_hv_clock, sizeof(guest_hv_clock))))
|
|
|
|
return 0;
|
|
|
|
|
2015-03-25 18:08:14 +07:00
|
|
|
/* This VCPU is paused, but it's legal for a guest to read another
|
|
|
|
* VCPU's kvmclock, so we really have to follow the specification where
|
|
|
|
* it says that version is odd if data is being modified, and even after
|
|
|
|
* it is consistent.
|
|
|
|
*
|
|
|
|
* Version field updates must be kept separate. This is because
|
|
|
|
* kvm_write_guest_cached might use a "rep movs" instruction, and
|
|
|
|
* writes within a string instruction are weakly ordered. So there
|
|
|
|
* are three writes overall.
|
|
|
|
*
|
|
|
|
* As a small optimization, only write the version field in the first
|
|
|
|
* and third write. The vcpu->pv_time cache is still valid, because the
|
|
|
|
* version field is the first in the struct.
|
2008-02-16 02:52:47 +07:00
|
|
|
*/
|
2015-03-25 18:08:14 +07:00
|
|
|
BUILD_BUG_ON(offsetof(struct pvclock_vcpu_time_info, version) != 0);
|
|
|
|
|
|
|
|
vcpu->hv_clock.version = guest_hv_clock.version + 1;
|
|
|
|
kvm_write_guest_cached(v->kvm, &vcpu->pv_time,
|
|
|
|
&vcpu->hv_clock,
|
|
|
|
sizeof(vcpu->hv_clock.version));
|
|
|
|
|
|
|
|
smp_wmb();
|
2012-11-28 08:28:47 +07:00
|
|
|
|
|
|
|
/* retain PVCLOCK_GUEST_STOPPED if set in guest copy */
|
2013-02-21 05:48:10 +07:00
|
|
|
pvclock_flags = (guest_hv_clock.flags & PVCLOCK_GUEST_STOPPED);
|
2012-11-28 08:28:47 +07:00
|
|
|
|
|
|
|
if (vcpu->pvclock_set_guest_stopped_request) {
|
|
|
|
pvclock_flags |= PVCLOCK_GUEST_STOPPED;
|
|
|
|
vcpu->pvclock_set_guest_stopped_request = false;
|
|
|
|
}
|
|
|
|
|
2012-11-28 08:29:01 +07:00
|
|
|
/* If the host uses TSC clocksource, then it is stable */
|
|
|
|
if (use_master_clock)
|
|
|
|
pvclock_flags |= PVCLOCK_TSC_STABLE_BIT;
|
|
|
|
|
2012-11-28 08:28:47 +07:00
|
|
|
vcpu->hv_clock.flags = pvclock_flags;
|
|
|
|
|
2014-11-06 02:46:42 +07:00
|
|
|
trace_kvm_pvclock_update(v->vcpu_id, &vcpu->hv_clock);
|
|
|
|
|
2013-02-21 05:48:10 +07:00
|
|
|
kvm_write_guest_cached(v->kvm, &vcpu->pv_time,
|
|
|
|
&vcpu->hv_clock,
|
|
|
|
sizeof(vcpu->hv_clock));
|
2015-03-25 18:08:14 +07:00
|
|
|
|
|
|
|
smp_wmb();
|
|
|
|
|
|
|
|
vcpu->hv_clock.version++;
|
|
|
|
kvm_write_guest_cached(v->kvm, &vcpu->pv_time,
|
|
|
|
&vcpu->hv_clock,
|
|
|
|
sizeof(vcpu->hv_clock.version));
|
2010-08-20 15:07:21 +07:00
|
|
|
return 0;
|
2009-02-04 23:52:04 +07:00
|
|
|
}
|
|
|
|
|
2013-05-10 06:21:41 +07:00
|
|
|
/*
|
|
|
|
* kvmclock updates which are isolated to a given vcpu, such as
|
|
|
|
* vcpu->cpu migration, should not allow system_timestamp from
|
|
|
|
* the rest of the vcpus to remain static. Otherwise ntp frequency
|
|
|
|
* correction applies to one vcpu's system_timestamp but not
|
|
|
|
* the others.
|
|
|
|
*
|
|
|
|
* So in those cases, request a kvmclock update for all vcpus.
|
2014-02-28 18:52:54 +07:00
|
|
|
* We need to rate-limit these requests though, as they can
|
|
|
|
* considerably slow guests that have a large number of vcpus.
|
|
|
|
* The time for a remote vcpu to update its kvmclock is bound
|
|
|
|
* by the delay we use to rate-limit the updates.
|
2013-05-10 06:21:41 +07:00
|
|
|
*/
|
|
|
|
|
2014-02-28 18:52:54 +07:00
|
|
|
#define KVMCLOCK_UPDATE_DELAY msecs_to_jiffies(100)
|
|
|
|
|
|
|
|
static void kvmclock_update_fn(struct work_struct *work)
|
2013-05-10 06:21:41 +07:00
|
|
|
{
|
|
|
|
int i;
|
2014-02-28 18:52:54 +07:00
|
|
|
struct delayed_work *dwork = to_delayed_work(work);
|
|
|
|
struct kvm_arch *ka = container_of(dwork, struct kvm_arch,
|
|
|
|
kvmclock_update_work);
|
|
|
|
struct kvm *kvm = container_of(ka, struct kvm, arch);
|
2013-05-10 06:21:41 +07:00
|
|
|
struct kvm_vcpu *vcpu;
|
|
|
|
|
|
|
|
kvm_for_each_vcpu(i, vcpu, kvm) {
|
2014-09-12 12:43:19 +07:00
|
|
|
kvm_make_request(KVM_REQ_CLOCK_UPDATE, vcpu);
|
2013-05-10 06:21:41 +07:00
|
|
|
kvm_vcpu_kick(vcpu);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2014-02-28 18:52:54 +07:00
|
|
|
static void kvm_gen_kvmclock_update(struct kvm_vcpu *v)
|
|
|
|
{
|
|
|
|
struct kvm *kvm = v->kvm;
|
|
|
|
|
2014-09-12 12:43:19 +07:00
|
|
|
kvm_make_request(KVM_REQ_CLOCK_UPDATE, v);
|
2014-02-28 18:52:54 +07:00
|
|
|
schedule_delayed_work(&kvm->arch.kvmclock_update_work,
|
|
|
|
KVMCLOCK_UPDATE_DELAY);
|
|
|
|
}
|
|
|
|
|
2014-02-28 18:52:55 +07:00
|
|
|
#define KVMCLOCK_SYNC_PERIOD (300 * HZ)
|
|
|
|
|
|
|
|
static void kvmclock_sync_fn(struct work_struct *work)
|
|
|
|
{
|
|
|
|
struct delayed_work *dwork = to_delayed_work(work);
|
|
|
|
struct kvm_arch *ka = container_of(dwork, struct kvm_arch,
|
|
|
|
kvmclock_sync_work);
|
|
|
|
struct kvm *kvm = container_of(ka, struct kvm, arch);
|
|
|
|
|
2015-05-13 08:42:04 +07:00
|
|
|
if (!kvmclock_periodic_sync)
|
|
|
|
return;
|
|
|
|
|
2014-02-28 18:52:55 +07:00
|
|
|
schedule_delayed_work(&kvm->arch.kvmclock_update_work, 0);
|
|
|
|
schedule_delayed_work(&kvm->arch.kvmclock_sync_work,
|
|
|
|
KVMCLOCK_SYNC_PERIOD);
|
|
|
|
}
|
|
|
|
|
2009-05-11 15:48:15 +07:00
|
|
|
static int set_msr_mce(struct kvm_vcpu *vcpu, u32 msr, u64 data)
|
2007-10-31 00:44:17 +07:00
|
|
|
{
|
2009-05-11 15:48:15 +07:00
|
|
|
u64 mcg_cap = vcpu->arch.mcg_cap;
|
|
|
|
unsigned bank_num = mcg_cap & 0xff;
|
|
|
|
|
2007-10-31 00:44:17 +07:00
|
|
|
switch (msr) {
|
|
|
|
case MSR_IA32_MCG_STATUS:
|
2009-05-11 15:48:15 +07:00
|
|
|
vcpu->arch.mcg_status = data;
|
2007-10-31 00:44:17 +07:00
|
|
|
break;
|
2008-02-12 02:28:27 +07:00
|
|
|
case MSR_IA32_MCG_CTL:
|
2009-05-11 15:48:15 +07:00
|
|
|
if (!(mcg_cap & MCG_CTL_P))
|
|
|
|
return 1;
|
|
|
|
if (data != 0 && data != ~(u64)0)
|
|
|
|
return -1;
|
|
|
|
vcpu->arch.mcg_ctl = data;
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
if (msr >= MSR_IA32_MC0_CTL &&
|
2014-09-23 09:44:35 +07:00
|
|
|
msr < MSR_IA32_MCx_CTL(bank_num)) {
|
2009-05-11 15:48:15 +07:00
|
|
|
u32 offset = msr - MSR_IA32_MC0_CTL;
|
2010-03-24 23:46:42 +07:00
|
|
|
/* only 0 or all 1s can be written to IA32_MCi_CTL
|
|
|
|
* some Linux kernels though clear bit 10 in bank 4 to
|
|
|
|
* workaround a BIOS/GART TBL issue on AMD K8s, ignore
|
|
|
|
* this to avoid an uncatched #GP in the guest
|
|
|
|
*/
|
2009-05-11 15:48:15 +07:00
|
|
|
if ((offset & 0x3) == 0 &&
|
2010-03-24 23:46:42 +07:00
|
|
|
data != 0 && (data | (1 << 10)) != ~(u64)0)
|
2009-05-11 15:48:15 +07:00
|
|
|
return -1;
|
|
|
|
vcpu->arch.mce_banks[offset] = data;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2009-10-16 05:21:43 +07:00
|
|
|
static int xen_hvm_config(struct kvm_vcpu *vcpu, u64 data)
|
|
|
|
{
|
|
|
|
struct kvm *kvm = vcpu->kvm;
|
|
|
|
int lm = is_long_mode(vcpu);
|
|
|
|
u8 *blob_addr = lm ? (u8 *)(long)kvm->arch.xen_hvm_config.blob_addr_64
|
|
|
|
: (u8 *)(long)kvm->arch.xen_hvm_config.blob_addr_32;
|
|
|
|
u8 blob_size = lm ? kvm->arch.xen_hvm_config.blob_size_64
|
|
|
|
: kvm->arch.xen_hvm_config.blob_size_32;
|
|
|
|
u32 page_num = data & ~PAGE_MASK;
|
|
|
|
u64 page_addr = data & PAGE_MASK;
|
|
|
|
u8 *page;
|
|
|
|
int r;
|
|
|
|
|
|
|
|
r = -E2BIG;
|
|
|
|
if (page_num >= blob_size)
|
|
|
|
goto out;
|
|
|
|
r = -ENOMEM;
|
2011-12-05 00:36:29 +07:00
|
|
|
page = memdup_user(blob_addr + (page_num * PAGE_SIZE), PAGE_SIZE);
|
|
|
|
if (IS_ERR(page)) {
|
|
|
|
r = PTR_ERR(page);
|
2009-10-16 05:21:43 +07:00
|
|
|
goto out;
|
2011-12-05 00:36:29 +07:00
|
|
|
}
|
2015-04-08 20:39:23 +07:00
|
|
|
if (kvm_vcpu_write_guest(vcpu, page_addr, page, PAGE_SIZE))
|
2009-10-16 05:21:43 +07:00
|
|
|
goto out_free;
|
|
|
|
r = 0;
|
|
|
|
out_free:
|
|
|
|
kfree(page);
|
|
|
|
out:
|
|
|
|
return r;
|
|
|
|
}
|
|
|
|
|
2010-10-14 16:22:50 +07:00
|
|
|
static int kvm_pv_enable_async_pf(struct kvm_vcpu *vcpu, u64 data)
|
|
|
|
{
|
|
|
|
gpa_t gpa = data & ~0x3f;
|
|
|
|
|
2012-06-28 14:17:27 +07:00
|
|
|
/* Bits 2:5 are reserved, Should be zero */
|
2010-10-14 16:22:55 +07:00
|
|
|
if (data & 0x3c)
|
2010-10-14 16:22:50 +07:00
|
|
|
return 1;
|
|
|
|
|
|
|
|
vcpu->arch.apf.msr_val = data;
|
|
|
|
|
|
|
|
if (!(data & KVM_ASYNC_PF_ENABLED)) {
|
|
|
|
kvm_clear_async_pf_completion_queue(vcpu);
|
|
|
|
kvm_async_pf_hash_reset(vcpu);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2013-03-29 23:35:21 +07:00
|
|
|
if (kvm_gfn_to_hva_cache_init(vcpu->kvm, &vcpu->arch.apf.data, gpa,
|
|
|
|
sizeof(u32)))
|
2010-10-14 16:22:50 +07:00
|
|
|
return 1;
|
|
|
|
|
2010-10-14 16:22:55 +07:00
|
|
|
vcpu->arch.apf.send_user_only = !(data & KVM_ASYNC_PF_SEND_ALWAYS);
|
2010-10-14 16:22:50 +07:00
|
|
|
kvm_async_pf_wakeup_all(vcpu);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2011-02-02 02:16:40 +07:00
|
|
|
static void kvmclock_reset(struct kvm_vcpu *vcpu)
|
|
|
|
{
|
2013-02-21 05:48:10 +07:00
|
|
|
vcpu->arch.pv_time_enabled = false;
|
2011-02-02 02:16:40 +07:00
|
|
|
}
|
|
|
|
|
2011-07-12 02:28:14 +07:00
|
|
|
static void record_steal_time(struct kvm_vcpu *vcpu)
|
|
|
|
{
|
|
|
|
if (!(vcpu->arch.st.msr_val & KVM_MSR_ENABLED))
|
|
|
|
return;
|
|
|
|
|
|
|
|
if (unlikely(kvm_read_guest_cached(vcpu->kvm, &vcpu->arch.st.stime,
|
|
|
|
&vcpu->arch.st.steal, sizeof(struct kvm_steal_time))))
|
|
|
|
return;
|
|
|
|
|
2016-05-03 10:43:10 +07:00
|
|
|
if (vcpu->arch.st.steal.version & 1)
|
|
|
|
vcpu->arch.st.steal.version += 1; /* first time write, random junk */
|
|
|
|
|
|
|
|
vcpu->arch.st.steal.version += 1;
|
|
|
|
|
|
|
|
kvm_write_guest_cached(vcpu->kvm, &vcpu->arch.st.stime,
|
|
|
|
&vcpu->arch.st.steal, sizeof(struct kvm_steal_time));
|
|
|
|
|
|
|
|
smp_wmb();
|
|
|
|
|
2016-03-16 18:33:16 +07:00
|
|
|
vcpu->arch.st.steal.steal += current->sched_info.run_delay -
|
|
|
|
vcpu->arch.st.last_steal;
|
|
|
|
vcpu->arch.st.last_steal = current->sched_info.run_delay;
|
2016-05-03 10:43:10 +07:00
|
|
|
|
|
|
|
kvm_write_guest_cached(vcpu->kvm, &vcpu->arch.st.stime,
|
|
|
|
&vcpu->arch.st.steal, sizeof(struct kvm_steal_time));
|
|
|
|
|
|
|
|
smp_wmb();
|
|
|
|
|
|
|
|
vcpu->arch.st.steal.version += 1;
|
2011-07-12 02:28:14 +07:00
|
|
|
|
|
|
|
kvm_write_guest_cached(vcpu->kvm, &vcpu->arch.st.stime,
|
|
|
|
&vcpu->arch.st.steal, sizeof(struct kvm_steal_time));
|
|
|
|
}
|
|
|
|
|
2012-11-30 03:42:12 +07:00
|
|
|
int kvm_set_msr_common(struct kvm_vcpu *vcpu, struct msr_data *msr_info)
|
2007-10-31 00:44:17 +07:00
|
|
|
{
|
2012-01-15 19:17:22 +07:00
|
|
|
bool pr = false;
|
2012-11-30 03:42:12 +07:00
|
|
|
u32 msr = msr_info->index;
|
|
|
|
u64 data = msr_info->data;
|
2012-01-15 19:17:22 +07:00
|
|
|
|
2007-10-31 00:44:17 +07:00
|
|
|
switch (msr) {
|
2013-02-20 01:33:13 +07:00
|
|
|
case MSR_AMD64_NB_CFG:
|
|
|
|
case MSR_IA32_UCODE_REV:
|
|
|
|
case MSR_IA32_UCODE_WRITE:
|
|
|
|
case MSR_VM_HSAVE_PA:
|
|
|
|
case MSR_AMD64_PATCH_LOADER:
|
|
|
|
case MSR_AMD64_BU_CFG2:
|
|
|
|
break;
|
|
|
|
|
2007-10-31 00:44:17 +07:00
|
|
|
case MSR_EFER:
|
2010-05-06 16:38:43 +07:00
|
|
|
return set_efer(vcpu, data);
|
2009-06-24 17:44:33 +07:00
|
|
|
case MSR_K7_HWCR:
|
|
|
|
data &= ~(u64)0x40; /* ignore flush filter disable */
|
2010-02-25 00:59:16 +07:00
|
|
|
data &= ~(u64)0x100; /* ignore ignne emulation enable */
|
2012-02-22 04:44:21 +07:00
|
|
|
data &= ~(u64)0x8; /* ignore TLB cache disable */
|
2014-06-26 18:50:15 +07:00
|
|
|
data &= ~(u64)0x40000; /* ignore Mc status write enable */
|
2009-06-24 17:44:33 +07:00
|
|
|
if (data != 0) {
|
KVM: Cleanup the kvm_print functions and introduce pr_XX wrappers
Introduces a couple of print functions, which are essentially wrappers
around standard printk functions, with a KVM: prefix.
Functions introduced or modified are:
- kvm_err(fmt, ...)
- kvm_info(fmt, ...)
- kvm_debug(fmt, ...)
- kvm_pr_unimpl(fmt, ...)
- pr_unimpl(vcpu, fmt, ...) -> vcpu_unimpl(vcpu, fmt, ...)
Signed-off-by: Christoffer Dall <c.dall@virtualopensystems.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
2012-06-04 01:17:48 +07:00
|
|
|
vcpu_unimpl(vcpu, "unimplemented HWCR wrmsr: 0x%llx\n",
|
|
|
|
data);
|
2009-06-24 17:44:33 +07:00
|
|
|
return 1;
|
|
|
|
}
|
2007-10-31 00:44:17 +07:00
|
|
|
break;
|
2009-07-02 20:04:14 +07:00
|
|
|
case MSR_FAM10H_MMIO_CONF_BASE:
|
|
|
|
if (data != 0) {
|
KVM: Cleanup the kvm_print functions and introduce pr_XX wrappers
Introduces a couple of print functions, which are essentially wrappers
around standard printk functions, with a KVM: prefix.
Functions introduced or modified are:
- kvm_err(fmt, ...)
- kvm_info(fmt, ...)
- kvm_debug(fmt, ...)
- kvm_pr_unimpl(fmt, ...)
- pr_unimpl(vcpu, fmt, ...) -> vcpu_unimpl(vcpu, fmt, ...)
Signed-off-by: Christoffer Dall <c.dall@virtualopensystems.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
2012-06-04 01:17:48 +07:00
|
|
|
vcpu_unimpl(vcpu, "unimplemented MMIO_CONF_BASE wrmsr: "
|
|
|
|
"0x%llx\n", data);
|
2009-07-02 20:04:14 +07:00
|
|
|
return 1;
|
|
|
|
}
|
2007-10-31 00:44:17 +07:00
|
|
|
break;
|
2008-07-22 13:00:45 +07:00
|
|
|
case MSR_IA32_DEBUGCTLMSR:
|
|
|
|
if (!data) {
|
|
|
|
/* We support the non-activated case already */
|
|
|
|
break;
|
|
|
|
} else if (data & ~(DEBUGCTLMSR_LBR | DEBUGCTLMSR_BTF)) {
|
|
|
|
/* Values other than LBR and BTF are vendor-specific,
|
|
|
|
thus reserved and should throw a #GP */
|
|
|
|
return 1;
|
|
|
|
}
|
KVM: Cleanup the kvm_print functions and introduce pr_XX wrappers
Introduces a couple of print functions, which are essentially wrappers
around standard printk functions, with a KVM: prefix.
Functions introduced or modified are:
- kvm_err(fmt, ...)
- kvm_info(fmt, ...)
- kvm_debug(fmt, ...)
- kvm_pr_unimpl(fmt, ...)
- pr_unimpl(vcpu, fmt, ...) -> vcpu_unimpl(vcpu, fmt, ...)
Signed-off-by: Christoffer Dall <c.dall@virtualopensystems.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
2012-06-04 01:17:48 +07:00
|
|
|
vcpu_unimpl(vcpu, "%s: MSR_IA32_DEBUGCTLMSR 0x%llx, nop\n",
|
|
|
|
__func__, data);
|
2008-07-22 13:00:45 +07:00
|
|
|
break;
|
2008-05-27 00:06:35 +07:00
|
|
|
case 0x200 ... 0x2ff:
|
2015-06-15 15:55:22 +07:00
|
|
|
return kvm_mtrr_set_msr(vcpu, msr, data);
|
2007-10-31 00:44:17 +07:00
|
|
|
case MSR_IA32_APICBASE:
|
2014-01-24 22:48:44 +07:00
|
|
|
return kvm_set_apic_base(vcpu, msr_info);
|
2009-07-05 21:39:36 +07:00
|
|
|
case APIC_BASE_MSR ... APIC_BASE_MSR + 0x3ff:
|
|
|
|
return kvm_x2apic_msr_write(vcpu, msr, data);
|
2011-09-22 15:55:52 +07:00
|
|
|
case MSR_IA32_TSCDEADLINE:
|
|
|
|
kvm_set_lapic_tscdeadline_msr(vcpu, data);
|
|
|
|
break;
|
2012-11-30 03:42:50 +07:00
|
|
|
case MSR_IA32_TSC_ADJUST:
|
|
|
|
if (guest_cpuid_has_tsc_adjust(vcpu)) {
|
|
|
|
if (!msr_info->host_initiated) {
|
2014-11-13 10:00:39 +07:00
|
|
|
s64 adj = data - vcpu->arch.ia32_tsc_adjust_msr;
|
2015-08-07 10:24:32 +07:00
|
|
|
adjust_tsc_offset_guest(vcpu, adj);
|
2012-11-30 03:42:50 +07:00
|
|
|
}
|
|
|
|
vcpu->arch.ia32_tsc_adjust_msr = data;
|
|
|
|
}
|
|
|
|
break;
|
2007-10-31 00:44:17 +07:00
|
|
|
case MSR_IA32_MISC_ENABLE:
|
2007-12-13 22:50:52 +07:00
|
|
|
vcpu->arch.ia32_misc_enable_msr = data;
|
2007-10-31 00:44:17 +07:00
|
|
|
break;
|
2015-05-07 16:36:11 +07:00
|
|
|
case MSR_IA32_SMBASE:
|
|
|
|
if (!msr_info->host_initiated)
|
|
|
|
return 1;
|
|
|
|
vcpu->arch.smbase = data;
|
|
|
|
break;
|
2010-05-11 23:17:41 +07:00
|
|
|
case MSR_KVM_WALL_CLOCK_NEW:
|
2008-02-16 02:52:47 +07:00
|
|
|
case MSR_KVM_WALL_CLOCK:
|
|
|
|
vcpu->kvm->arch.wall_clock = data;
|
|
|
|
kvm_write_wall_clock(vcpu->kvm, data);
|
|
|
|
break;
|
2010-05-11 23:17:41 +07:00
|
|
|
case MSR_KVM_SYSTEM_TIME_NEW:
|
2008-02-16 02:52:47 +07:00
|
|
|
case MSR_KVM_SYSTEM_TIME: {
|
2013-02-21 05:48:10 +07:00
|
|
|
u64 gpa_offset;
|
2015-01-21 00:54:52 +07:00
|
|
|
struct kvm_arch *ka = &vcpu->kvm->arch;
|
|
|
|
|
2011-02-02 02:16:40 +07:00
|
|
|
kvmclock_reset(vcpu);
|
2008-02-16 02:52:47 +07:00
|
|
|
|
2015-01-21 00:54:52 +07:00
|
|
|
if (vcpu->vcpu_id == 0 && !msr_info->host_initiated) {
|
|
|
|
bool tmp = (msr == MSR_KVM_SYSTEM_TIME);
|
|
|
|
|
|
|
|
if (ka->boot_vcpu_runs_old_kvmclock != tmp)
|
|
|
|
set_bit(KVM_REQ_MASTERCLOCK_UPDATE,
|
|
|
|
&vcpu->requests);
|
|
|
|
|
|
|
|
ka->boot_vcpu_runs_old_kvmclock = tmp;
|
|
|
|
}
|
|
|
|
|
2008-02-16 02:52:47 +07:00
|
|
|
vcpu->arch.time = data;
|
2013-05-10 06:21:41 +07:00
|
|
|
kvm_make_request(KVM_REQ_GLOBAL_CLOCK_UPDATE, vcpu);
|
2008-02-16 02:52:47 +07:00
|
|
|
|
|
|
|
/* we verify if the enable bit is set... */
|
|
|
|
if (!(data & 1))
|
|
|
|
break;
|
|
|
|
|
2013-02-21 05:48:10 +07:00
|
|
|
gpa_offset = data & ~(PAGE_MASK | 1);
|
2008-02-16 02:52:47 +07:00
|
|
|
|
2013-02-21 05:48:10 +07:00
|
|
|
if (kvm_gfn_to_hva_cache_init(vcpu->kvm,
|
2013-03-29 23:35:21 +07:00
|
|
|
&vcpu->arch.pv_time, data & ~1ULL,
|
|
|
|
sizeof(struct pvclock_vcpu_time_info)))
|
2013-02-21 05:48:10 +07:00
|
|
|
vcpu->arch.pv_time_enabled = false;
|
|
|
|
else
|
|
|
|
vcpu->arch.pv_time_enabled = true;
|
2012-08-03 14:42:52 +07:00
|
|
|
|
2008-02-16 02:52:47 +07:00
|
|
|
break;
|
|
|
|
}
|
2010-10-14 16:22:50 +07:00
|
|
|
case MSR_KVM_ASYNC_PF_EN:
|
|
|
|
if (kvm_pv_enable_async_pf(vcpu, data))
|
|
|
|
return 1;
|
|
|
|
break;
|
2011-07-12 02:28:14 +07:00
|
|
|
case MSR_KVM_STEAL_TIME:
|
|
|
|
|
|
|
|
if (unlikely(!sched_info_on()))
|
|
|
|
return 1;
|
|
|
|
|
|
|
|
if (data & KVM_STEAL_RESERVED_MASK)
|
|
|
|
return 1;
|
|
|
|
|
|
|
|
if (kvm_gfn_to_hva_cache_init(vcpu->kvm, &vcpu->arch.st.stime,
|
2013-03-29 23:35:21 +07:00
|
|
|
data & KVM_STEAL_VALID_BITS,
|
|
|
|
sizeof(struct kvm_steal_time)))
|
2011-07-12 02:28:14 +07:00
|
|
|
return 1;
|
|
|
|
|
|
|
|
vcpu->arch.st.msr_val = data;
|
|
|
|
|
|
|
|
if (!(data & KVM_MSR_ENABLED))
|
|
|
|
break;
|
|
|
|
|
|
|
|
kvm_make_request(KVM_REQ_STEAL_UPDATE, vcpu);
|
|
|
|
|
|
|
|
break;
|
2012-06-24 23:25:07 +07:00
|
|
|
case MSR_KVM_PV_EOI_EN:
|
|
|
|
if (kvm_lapic_enable_pv_eoi(vcpu, data))
|
|
|
|
return 1;
|
|
|
|
break;
|
2011-07-12 02:28:14 +07:00
|
|
|
|
2009-05-11 15:48:15 +07:00
|
|
|
case MSR_IA32_MCG_CTL:
|
|
|
|
case MSR_IA32_MCG_STATUS:
|
2014-09-23 09:44:35 +07:00
|
|
|
case MSR_IA32_MC0_CTL ... MSR_IA32_MCx_CTL(KVM_MAX_MCE_BANKS) - 1:
|
2009-05-11 15:48:15 +07:00
|
|
|
return set_msr_mce(vcpu, msr, data);
|
2009-06-13 03:01:29 +07:00
|
|
|
|
2015-06-12 12:34:56 +07:00
|
|
|
case MSR_K7_PERFCTR0 ... MSR_K7_PERFCTR3:
|
|
|
|
case MSR_P6_PERFCTR0 ... MSR_P6_PERFCTR1:
|
|
|
|
pr = true; /* fall through */
|
|
|
|
case MSR_K7_EVNTSEL0 ... MSR_K7_EVNTSEL3:
|
|
|
|
case MSR_P6_EVNTSEL0 ... MSR_P6_EVNTSEL1:
|
2015-06-19 18:44:45 +07:00
|
|
|
if (kvm_pmu_is_valid_msr(vcpu, msr))
|
2013-03-28 23:18:35 +07:00
|
|
|
return kvm_pmu_set_msr(vcpu, msr_info);
|
2012-01-15 19:17:22 +07:00
|
|
|
|
|
|
|
if (pr || data != 0)
|
KVM: Cleanup the kvm_print functions and introduce pr_XX wrappers
Introduces a couple of print functions, which are essentially wrappers
around standard printk functions, with a KVM: prefix.
Functions introduced or modified are:
- kvm_err(fmt, ...)
- kvm_info(fmt, ...)
- kvm_debug(fmt, ...)
- kvm_pr_unimpl(fmt, ...)
- pr_unimpl(vcpu, fmt, ...) -> vcpu_unimpl(vcpu, fmt, ...)
Signed-off-by: Christoffer Dall <c.dall@virtualopensystems.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
2012-06-04 01:17:48 +07:00
|
|
|
vcpu_unimpl(vcpu, "disabled perfctr wrmsr: "
|
|
|
|
"0x%x data 0x%llx\n", msr, data);
|
2012-01-15 19:17:22 +07:00
|
|
|
break;
|
2010-09-01 16:42:04 +07:00
|
|
|
case MSR_K7_CLK_CTL:
|
|
|
|
/*
|
|
|
|
* Ignore all writes to this no longer documented MSR.
|
|
|
|
* Writes are only relevant for old K7 processors,
|
|
|
|
* all pre-dating SVM, but a recommended workaround from
|
2012-06-28 14:17:27 +07:00
|
|
|
* AMD for these chips. It is possible to specify the
|
2010-09-01 16:42:04 +07:00
|
|
|
* affected processor models on the command line, hence
|
|
|
|
* the need to ignore the workaround.
|
|
|
|
*/
|
|
|
|
break;
|
2010-01-17 20:51:22 +07:00
|
|
|
case HV_X64_MSR_GUEST_OS_ID ... HV_X64_MSR_SINT15:
|
2015-07-03 19:01:37 +07:00
|
|
|
case HV_X64_MSR_CRASH_P0 ... HV_X64_MSR_CRASH_P4:
|
|
|
|
case HV_X64_MSR_CRASH_CTL:
|
2015-11-30 23:22:21 +07:00
|
|
|
case HV_X64_MSR_STIMER0_CONFIG ... HV_X64_MSR_STIMER3_COUNT:
|
2015-07-03 19:01:37 +07:00
|
|
|
return kvm_hv_set_msr_common(vcpu, msr, data,
|
|
|
|
msr_info->host_initiated);
|
2011-01-21 12:21:00 +07:00
|
|
|
case MSR_IA32_BBL_CR_CTL3:
|
|
|
|
/* Drop writes to this legacy MSR -- see rdmsr
|
|
|
|
* counterpart for further detail.
|
|
|
|
*/
|
KVM: Cleanup the kvm_print functions and introduce pr_XX wrappers
Introduces a couple of print functions, which are essentially wrappers
around standard printk functions, with a KVM: prefix.
Functions introduced or modified are:
- kvm_err(fmt, ...)
- kvm_info(fmt, ...)
- kvm_debug(fmt, ...)
- kvm_pr_unimpl(fmt, ...)
- pr_unimpl(vcpu, fmt, ...) -> vcpu_unimpl(vcpu, fmt, ...)
Signed-off-by: Christoffer Dall <c.dall@virtualopensystems.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
2012-06-04 01:17:48 +07:00
|
|
|
vcpu_unimpl(vcpu, "ignored wrmsr: 0x%x data %llx\n", msr, data);
|
2011-01-21 12:21:00 +07:00
|
|
|
break;
|
2012-01-10 02:00:35 +07:00
|
|
|
case MSR_AMD64_OSVW_ID_LENGTH:
|
|
|
|
if (!guest_cpuid_has_osvw(vcpu))
|
|
|
|
return 1;
|
|
|
|
vcpu->arch.osvw.length = data;
|
|
|
|
break;
|
|
|
|
case MSR_AMD64_OSVW_STATUS:
|
|
|
|
if (!guest_cpuid_has_osvw(vcpu))
|
|
|
|
return 1;
|
|
|
|
vcpu->arch.osvw.status = data;
|
|
|
|
break;
|
2007-10-31 00:44:17 +07:00
|
|
|
default:
|
2009-10-16 05:21:43 +07:00
|
|
|
if (msr && (msr == vcpu->kvm->arch.xen_hvm_config.msr))
|
|
|
|
return xen_hvm_config(vcpu, data);
|
2015-06-19 18:44:45 +07:00
|
|
|
if (kvm_pmu_is_valid_msr(vcpu, msr))
|
2013-03-28 23:18:35 +07:00
|
|
|
return kvm_pmu_set_msr(vcpu, msr_info);
|
2009-06-25 17:36:49 +07:00
|
|
|
if (!ignore_msrs) {
|
KVM: Cleanup the kvm_print functions and introduce pr_XX wrappers
Introduces a couple of print functions, which are essentially wrappers
around standard printk functions, with a KVM: prefix.
Functions introduced or modified are:
- kvm_err(fmt, ...)
- kvm_info(fmt, ...)
- kvm_debug(fmt, ...)
- kvm_pr_unimpl(fmt, ...)
- pr_unimpl(vcpu, fmt, ...) -> vcpu_unimpl(vcpu, fmt, ...)
Signed-off-by: Christoffer Dall <c.dall@virtualopensystems.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
2012-06-04 01:17:48 +07:00
|
|
|
vcpu_unimpl(vcpu, "unhandled wrmsr: 0x%x data %llx\n",
|
|
|
|
msr, data);
|
2009-06-25 17:36:49 +07:00
|
|
|
return 1;
|
|
|
|
} else {
|
KVM: Cleanup the kvm_print functions and introduce pr_XX wrappers
Introduces a couple of print functions, which are essentially wrappers
around standard printk functions, with a KVM: prefix.
Functions introduced or modified are:
- kvm_err(fmt, ...)
- kvm_info(fmt, ...)
- kvm_debug(fmt, ...)
- kvm_pr_unimpl(fmt, ...)
- pr_unimpl(vcpu, fmt, ...) -> vcpu_unimpl(vcpu, fmt, ...)
Signed-off-by: Christoffer Dall <c.dall@virtualopensystems.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
2012-06-04 01:17:48 +07:00
|
|
|
vcpu_unimpl(vcpu, "ignored wrmsr: 0x%x data %llx\n",
|
|
|
|
msr, data);
|
2009-06-25 17:36:49 +07:00
|
|
|
break;
|
|
|
|
}
|
2007-10-31 00:44:17 +07:00
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(kvm_set_msr_common);
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Reads an msr value (of 'msr_index') into 'pdata'.
|
|
|
|
* Returns 0 on success, non-0 otherwise.
|
|
|
|
* Assumes vcpu_load() was already called.
|
|
|
|
*/
|
2015-04-08 20:30:38 +07:00
|
|
|
int kvm_get_msr(struct kvm_vcpu *vcpu, struct msr_data *msr)
|
2007-10-31 00:44:17 +07:00
|
|
|
{
|
2015-04-08 20:30:38 +07:00
|
|
|
return kvm_x86_ops->get_msr(vcpu, msr);
|
2007-10-31 00:44:17 +07:00
|
|
|
}
|
2014-12-11 12:52:58 +07:00
|
|
|
EXPORT_SYMBOL_GPL(kvm_get_msr);
|
2007-10-31 00:44:17 +07:00
|
|
|
|
2009-05-11 15:48:15 +07:00
|
|
|
static int get_msr_mce(struct kvm_vcpu *vcpu, u32 msr, u64 *pdata)
|
2007-10-31 00:44:17 +07:00
|
|
|
{
|
|
|
|
u64 data;
|
2009-05-11 15:48:15 +07:00
|
|
|
u64 mcg_cap = vcpu->arch.mcg_cap;
|
|
|
|
unsigned bank_num = mcg_cap & 0xff;
|
2007-10-31 00:44:17 +07:00
|
|
|
|
|
|
|
switch (msr) {
|
|
|
|
case MSR_IA32_P5_MC_ADDR:
|
|
|
|
case MSR_IA32_P5_MC_TYPE:
|
2009-05-11 15:48:15 +07:00
|
|
|
data = 0;
|
|
|
|
break;
|
2007-10-31 00:44:17 +07:00
|
|
|
case MSR_IA32_MCG_CAP:
|
2009-05-11 15:48:15 +07:00
|
|
|
data = vcpu->arch.mcg_cap;
|
|
|
|
break;
|
2008-02-12 02:28:27 +07:00
|
|
|
case MSR_IA32_MCG_CTL:
|
2009-05-11 15:48:15 +07:00
|
|
|
if (!(mcg_cap & MCG_CTL_P))
|
|
|
|
return 1;
|
|
|
|
data = vcpu->arch.mcg_ctl;
|
|
|
|
break;
|
|
|
|
case MSR_IA32_MCG_STATUS:
|
|
|
|
data = vcpu->arch.mcg_status;
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
if (msr >= MSR_IA32_MC0_CTL &&
|
2014-09-23 09:44:35 +07:00
|
|
|
msr < MSR_IA32_MCx_CTL(bank_num)) {
|
2009-05-11 15:48:15 +07:00
|
|
|
u32 offset = msr - MSR_IA32_MC0_CTL;
|
|
|
|
data = vcpu->arch.mce_banks[offset];
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
*pdata = data;
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2015-04-08 20:30:38 +07:00
|
|
|
int kvm_get_msr_common(struct kvm_vcpu *vcpu, struct msr_data *msr_info)
|
2009-05-11 15:48:15 +07:00
|
|
|
{
|
2015-04-08 20:30:38 +07:00
|
|
|
switch (msr_info->index) {
|
2009-05-11 15:48:15 +07:00
|
|
|
case MSR_IA32_PLATFORM_ID:
|
2007-10-31 00:44:17 +07:00
|
|
|
case MSR_IA32_EBL_CR_POWERON:
|
2008-07-22 13:00:45 +07:00
|
|
|
case MSR_IA32_DEBUGCTLMSR:
|
|
|
|
case MSR_IA32_LASTBRANCHFROMIP:
|
|
|
|
case MSR_IA32_LASTBRANCHTOIP:
|
|
|
|
case MSR_IA32_LASTINTFROMIP:
|
|
|
|
case MSR_IA32_LASTINTTOIP:
|
2009-05-14 12:30:10 +07:00
|
|
|
case MSR_K8_SYSCFG:
|
2015-09-18 22:33:04 +07:00
|
|
|
case MSR_K8_TSEG_ADDR:
|
|
|
|
case MSR_K8_TSEG_MASK:
|
2009-05-14 12:30:10 +07:00
|
|
|
case MSR_K7_HWCR:
|
2008-12-29 22:32:28 +07:00
|
|
|
case MSR_VM_HSAVE_PA:
|
2009-06-24 17:44:34 +07:00
|
|
|
case MSR_K8_INT_PENDING_MSG:
|
2009-06-24 20:37:05 +07:00
|
|
|
case MSR_AMD64_NB_CFG:
|
2009-07-02 20:04:14 +07:00
|
|
|
case MSR_FAM10H_MMIO_CONF_BASE:
|
2013-02-20 01:33:13 +07:00
|
|
|
case MSR_AMD64_BU_CFG2:
|
2016-05-31 21:38:24 +07:00
|
|
|
case MSR_IA32_PERF_CTL:
|
2015-04-08 20:30:38 +07:00
|
|
|
msr_info->data = 0;
|
2007-10-31 00:44:17 +07:00
|
|
|
break;
|
2015-06-12 12:34:56 +07:00
|
|
|
case MSR_K7_EVNTSEL0 ... MSR_K7_EVNTSEL3:
|
|
|
|
case MSR_K7_PERFCTR0 ... MSR_K7_PERFCTR3:
|
|
|
|
case MSR_P6_PERFCTR0 ... MSR_P6_PERFCTR1:
|
|
|
|
case MSR_P6_EVNTSEL0 ... MSR_P6_EVNTSEL1:
|
2015-06-19 18:44:45 +07:00
|
|
|
if (kvm_pmu_is_valid_msr(vcpu, msr_info->index))
|
2015-04-08 20:30:38 +07:00
|
|
|
return kvm_pmu_get_msr(vcpu, msr_info->index, &msr_info->data);
|
|
|
|
msr_info->data = 0;
|
2012-01-15 19:17:22 +07:00
|
|
|
break;
|
2011-07-30 05:44:21 +07:00
|
|
|
case MSR_IA32_UCODE_REV:
|
2015-04-08 20:30:38 +07:00
|
|
|
msr_info->data = 0x100000000ULL;
|
2011-07-30 05:44:21 +07:00
|
|
|
break;
|
2008-05-27 00:06:35 +07:00
|
|
|
case MSR_MTRRcap:
|
|
|
|
case 0x200 ... 0x2ff:
|
2015-06-15 15:55:22 +07:00
|
|
|
return kvm_mtrr_get_msr(vcpu, msr_info->index, &msr_info->data);
|
2007-10-31 00:44:17 +07:00
|
|
|
case 0xcd: /* fsb frequency */
|
2015-04-08 20:30:38 +07:00
|
|
|
msr_info->data = 3;
|
2007-10-31 00:44:17 +07:00
|
|
|
break;
|
2010-09-09 17:06:46 +07:00
|
|
|
/*
|
|
|
|
* MSR_EBC_FREQUENCY_ID
|
|
|
|
* Conservative value valid for even the basic CPU models.
|
|
|
|
* Models 0,1: 000 in bits 23:21 indicating a bus speed of
|
|
|
|
* 100MHz, model 2 000 in bits 18:16 indicating 100MHz,
|
|
|
|
* and 266MHz for model 3, or 4. Set Core Clock
|
|
|
|
* Frequency to System Bus Frequency Ratio to 1 (bits
|
|
|
|
* 31:24) even though these are only valid for CPU
|
|
|
|
* models > 2, however guests may end up dividing or
|
|
|
|
* multiplying by zero otherwise.
|
|
|
|
*/
|
|
|
|
case MSR_EBC_FREQUENCY_ID:
|
2015-04-08 20:30:38 +07:00
|
|
|
msr_info->data = 1 << 24;
|
2010-09-09 17:06:46 +07:00
|
|
|
break;
|
2007-10-31 00:44:17 +07:00
|
|
|
case MSR_IA32_APICBASE:
|
2015-04-08 20:30:38 +07:00
|
|
|
msr_info->data = kvm_get_apic_base(vcpu);
|
2007-10-31 00:44:17 +07:00
|
|
|
break;
|
2009-07-05 21:39:36 +07:00
|
|
|
case APIC_BASE_MSR ... APIC_BASE_MSR + 0x3ff:
|
2015-04-08 20:30:38 +07:00
|
|
|
return kvm_x2apic_msr_read(vcpu, msr_info->index, &msr_info->data);
|
2009-07-05 21:39:36 +07:00
|
|
|
break;
|
2011-09-22 15:55:52 +07:00
|
|
|
case MSR_IA32_TSCDEADLINE:
|
2015-04-08 20:30:38 +07:00
|
|
|
msr_info->data = kvm_get_lapic_tscdeadline_msr(vcpu);
|
2011-09-22 15:55:52 +07:00
|
|
|
break;
|
2012-11-30 03:42:50 +07:00
|
|
|
case MSR_IA32_TSC_ADJUST:
|
2015-04-08 20:30:38 +07:00
|
|
|
msr_info->data = (u64)vcpu->arch.ia32_tsc_adjust_msr;
|
2012-11-30 03:42:50 +07:00
|
|
|
break;
|
2007-10-31 00:44:17 +07:00
|
|
|
case MSR_IA32_MISC_ENABLE:
|
2015-04-08 20:30:38 +07:00
|
|
|
msr_info->data = vcpu->arch.ia32_misc_enable_msr;
|
2007-10-31 00:44:17 +07:00
|
|
|
break;
|
2015-05-07 16:36:11 +07:00
|
|
|
case MSR_IA32_SMBASE:
|
|
|
|
if (!msr_info->host_initiated)
|
|
|
|
return 1;
|
|
|
|
msr_info->data = vcpu->arch.smbase;
|
2007-10-31 00:44:17 +07:00
|
|
|
break;
|
2008-02-21 18:11:01 +07:00
|
|
|
case MSR_IA32_PERF_STATUS:
|
|
|
|
/* TSC increment by tick */
|
2015-04-08 20:30:38 +07:00
|
|
|
msr_info->data = 1000ULL;
|
2008-02-21 18:11:01 +07:00
|
|
|
/* CPU multiplier */
|
2015-06-29 17:39:23 +07:00
|
|
|
msr_info->data |= (((uint64_t)4ULL) << 40);
|
2008-02-21 18:11:01 +07:00
|
|
|
break;
|
2007-10-31 00:44:17 +07:00
|
|
|
case MSR_EFER:
|
2015-04-08 20:30:38 +07:00
|
|
|
msr_info->data = vcpu->arch.efer;
|
2007-10-31 00:44:17 +07:00
|
|
|
break;
|
2008-02-16 02:52:47 +07:00
|
|
|
case MSR_KVM_WALL_CLOCK:
|
2010-05-11 23:17:41 +07:00
|
|
|
case MSR_KVM_WALL_CLOCK_NEW:
|
2015-04-08 20:30:38 +07:00
|
|
|
msr_info->data = vcpu->kvm->arch.wall_clock;
|
2008-02-16 02:52:47 +07:00
|
|
|
break;
|
|
|
|
case MSR_KVM_SYSTEM_TIME:
|
2010-05-11 23:17:41 +07:00
|
|
|
case MSR_KVM_SYSTEM_TIME_NEW:
|
2015-04-08 20:30:38 +07:00
|
|
|
msr_info->data = vcpu->arch.time;
|
2008-02-16 02:52:47 +07:00
|
|
|
break;
|
2010-10-14 16:22:50 +07:00
|
|
|
case MSR_KVM_ASYNC_PF_EN:
|
2015-04-08 20:30:38 +07:00
|
|
|
msr_info->data = vcpu->arch.apf.msr_val;
|
2010-10-14 16:22:50 +07:00
|
|
|
break;
|
2011-07-12 02:28:14 +07:00
|
|
|
case MSR_KVM_STEAL_TIME:
|
2015-04-08 20:30:38 +07:00
|
|
|
msr_info->data = vcpu->arch.st.msr_val;
|
2011-07-12 02:28:14 +07:00
|
|
|
break;
|
2012-08-26 22:00:29 +07:00
|
|
|
case MSR_KVM_PV_EOI_EN:
|
2015-04-08 20:30:38 +07:00
|
|
|
msr_info->data = vcpu->arch.pv_eoi.msr_val;
|
2012-08-26 22:00:29 +07:00
|
|
|
break;
|
2009-05-11 15:48:15 +07:00
|
|
|
case MSR_IA32_P5_MC_ADDR:
|
|
|
|
case MSR_IA32_P5_MC_TYPE:
|
|
|
|
case MSR_IA32_MCG_CAP:
|
|
|
|
case MSR_IA32_MCG_CTL:
|
|
|
|
case MSR_IA32_MCG_STATUS:
|
2014-09-23 09:44:35 +07:00
|
|
|
case MSR_IA32_MC0_CTL ... MSR_IA32_MCx_CTL(KVM_MAX_MCE_BANKS) - 1:
|
2015-04-08 20:30:38 +07:00
|
|
|
return get_msr_mce(vcpu, msr_info->index, &msr_info->data);
|
2010-09-01 16:42:04 +07:00
|
|
|
case MSR_K7_CLK_CTL:
|
|
|
|
/*
|
|
|
|
* Provide expected ramp-up count for K7. All other
|
|
|
|
* are set to zero, indicating minimum divisors for
|
|
|
|
* every field.
|
|
|
|
*
|
|
|
|
* This prevents guest kernels on AMD host with CPU
|
|
|
|
* type 6, model 8 and higher from exploding due to
|
|
|
|
* the rdmsr failing.
|
|
|
|
*/
|
2015-04-08 20:30:38 +07:00
|
|
|
msr_info->data = 0x20000000;
|
2010-09-01 16:42:04 +07:00
|
|
|
break;
|
2010-01-17 20:51:22 +07:00
|
|
|
case HV_X64_MSR_GUEST_OS_ID ... HV_X64_MSR_SINT15:
|
2015-07-03 19:01:37 +07:00
|
|
|
case HV_X64_MSR_CRASH_P0 ... HV_X64_MSR_CRASH_P4:
|
|
|
|
case HV_X64_MSR_CRASH_CTL:
|
2015-11-30 23:22:21 +07:00
|
|
|
case HV_X64_MSR_STIMER0_CONFIG ... HV_X64_MSR_STIMER3_COUNT:
|
2015-07-03 19:01:34 +07:00
|
|
|
return kvm_hv_get_msr_common(vcpu,
|
|
|
|
msr_info->index, &msr_info->data);
|
2010-01-17 20:51:22 +07:00
|
|
|
break;
|
2011-01-21 12:21:00 +07:00
|
|
|
case MSR_IA32_BBL_CR_CTL3:
|
|
|
|
/* This legacy MSR exists but isn't fully documented in current
|
|
|
|
* silicon. It is however accessed by winxp in very narrow
|
|
|
|
* scenarios where it sets bit #19, itself documented as
|
|
|
|
* a "reserved" bit. Best effort attempt to source coherent
|
|
|
|
* read data here should the balance of the register be
|
|
|
|
* interpreted by the guest:
|
|
|
|
*
|
|
|
|
* L2 cache control register 3: 64GB range, 256KB size,
|
|
|
|
* enabled, latency 0x1, configured
|
|
|
|
*/
|
2015-04-08 20:30:38 +07:00
|
|
|
msr_info->data = 0xbe702111;
|
2011-01-21 12:21:00 +07:00
|
|
|
break;
|
2012-01-10 02:00:35 +07:00
|
|
|
case MSR_AMD64_OSVW_ID_LENGTH:
|
|
|
|
if (!guest_cpuid_has_osvw(vcpu))
|
|
|
|
return 1;
|
2015-04-08 20:30:38 +07:00
|
|
|
msr_info->data = vcpu->arch.osvw.length;
|
2012-01-10 02:00:35 +07:00
|
|
|
break;
|
|
|
|
case MSR_AMD64_OSVW_STATUS:
|
|
|
|
if (!guest_cpuid_has_osvw(vcpu))
|
|
|
|
return 1;
|
2015-04-08 20:30:38 +07:00
|
|
|
msr_info->data = vcpu->arch.osvw.status;
|
2012-01-10 02:00:35 +07:00
|
|
|
break;
|
2007-10-31 00:44:17 +07:00
|
|
|
default:
|
2015-06-19 18:44:45 +07:00
|
|
|
if (kvm_pmu_is_valid_msr(vcpu, msr_info->index))
|
2015-04-08 20:30:38 +07:00
|
|
|
return kvm_pmu_get_msr(vcpu, msr_info->index, &msr_info->data);
|
2009-06-25 17:36:49 +07:00
|
|
|
if (!ignore_msrs) {
|
2015-04-08 20:30:38 +07:00
|
|
|
vcpu_unimpl(vcpu, "unhandled rdmsr: 0x%x\n", msr_info->index);
|
2009-06-25 17:36:49 +07:00
|
|
|
return 1;
|
|
|
|
} else {
|
2015-04-08 20:30:38 +07:00
|
|
|
vcpu_unimpl(vcpu, "ignored rdmsr: 0x%x\n", msr_info->index);
|
|
|
|
msr_info->data = 0;
|
2009-06-25 17:36:49 +07:00
|
|
|
}
|
|
|
|
break;
|
2007-10-31 00:44:17 +07:00
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(kvm_get_msr_common);
|
|
|
|
|
KVM: Portability: split kvm_vcpu_ioctl
This patch splits kvm_vcpu_ioctl into archtecture independent parts, and
x86 specific parts which go to kvm_arch_vcpu_ioctl in x86.c.
Common ioctls for all architectures are:
KVM_RUN, KVM_GET/SET_(S-)REGS, KVM_TRANSLATE, KVM_INTERRUPT,
KVM_DEBUG_GUEST, KVM_SET_SIGNAL_MASK, KVM_GET/SET_FPU
Note that some PPC chips don't have an FPU, so we might need an #ifdef
around KVM_GET/SET_FPU one day.
x86 specific ioctls are:
KVM_GET/SET_LAPIC, KVM_SET_CPUID, KVM_GET/SET_MSRS
An interresting aspect is vcpu_load/vcpu_put. We now have a common
vcpu_load/put which does the preemption stuff, and an architecture
specific kvm_arch_vcpu_load/put. In the x86 case, this one calls the
vmx/svm function defined in kvm_x86_ops.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Reviewed-by: Christian Ehrhardt <ehrhardt@linux.vnet.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-12 00:16:52 +07:00
|
|
|
/*
|
|
|
|
* Read or write a bunch of msrs. All parameters are kernel addresses.
|
|
|
|
*
|
|
|
|
* @return number of msrs set successfully.
|
|
|
|
*/
|
|
|
|
static int __msr_io(struct kvm_vcpu *vcpu, struct kvm_msrs *msrs,
|
|
|
|
struct kvm_msr_entry *entries,
|
|
|
|
int (*do_msr)(struct kvm_vcpu *vcpu,
|
|
|
|
unsigned index, u64 *data))
|
|
|
|
{
|
2009-12-23 23:35:25 +07:00
|
|
|
int i, idx;
|
KVM: Portability: split kvm_vcpu_ioctl
This patch splits kvm_vcpu_ioctl into archtecture independent parts, and
x86 specific parts which go to kvm_arch_vcpu_ioctl in x86.c.
Common ioctls for all architectures are:
KVM_RUN, KVM_GET/SET_(S-)REGS, KVM_TRANSLATE, KVM_INTERRUPT,
KVM_DEBUG_GUEST, KVM_SET_SIGNAL_MASK, KVM_GET/SET_FPU
Note that some PPC chips don't have an FPU, so we might need an #ifdef
around KVM_GET/SET_FPU one day.
x86 specific ioctls are:
KVM_GET/SET_LAPIC, KVM_SET_CPUID, KVM_GET/SET_MSRS
An interresting aspect is vcpu_load/vcpu_put. We now have a common
vcpu_load/put which does the preemption stuff, and an architecture
specific kvm_arch_vcpu_load/put. In the x86 case, this one calls the
vmx/svm function defined in kvm_x86_ops.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Reviewed-by: Christian Ehrhardt <ehrhardt@linux.vnet.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-12 00:16:52 +07:00
|
|
|
|
2009-12-23 23:35:25 +07:00
|
|
|
idx = srcu_read_lock(&vcpu->kvm->srcu);
|
KVM: Portability: split kvm_vcpu_ioctl
This patch splits kvm_vcpu_ioctl into archtecture independent parts, and
x86 specific parts which go to kvm_arch_vcpu_ioctl in x86.c.
Common ioctls for all architectures are:
KVM_RUN, KVM_GET/SET_(S-)REGS, KVM_TRANSLATE, KVM_INTERRUPT,
KVM_DEBUG_GUEST, KVM_SET_SIGNAL_MASK, KVM_GET/SET_FPU
Note that some PPC chips don't have an FPU, so we might need an #ifdef
around KVM_GET/SET_FPU one day.
x86 specific ioctls are:
KVM_GET/SET_LAPIC, KVM_SET_CPUID, KVM_GET/SET_MSRS
An interresting aspect is vcpu_load/vcpu_put. We now have a common
vcpu_load/put which does the preemption stuff, and an architecture
specific kvm_arch_vcpu_load/put. In the x86 case, this one calls the
vmx/svm function defined in kvm_x86_ops.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Reviewed-by: Christian Ehrhardt <ehrhardt@linux.vnet.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-12 00:16:52 +07:00
|
|
|
for (i = 0; i < msrs->nmsrs; ++i)
|
|
|
|
if (do_msr(vcpu, entries[i].index, &entries[i].data))
|
|
|
|
break;
|
2009-12-23 23:35:25 +07:00
|
|
|
srcu_read_unlock(&vcpu->kvm->srcu, idx);
|
KVM: Portability: split kvm_vcpu_ioctl
This patch splits kvm_vcpu_ioctl into archtecture independent parts, and
x86 specific parts which go to kvm_arch_vcpu_ioctl in x86.c.
Common ioctls for all architectures are:
KVM_RUN, KVM_GET/SET_(S-)REGS, KVM_TRANSLATE, KVM_INTERRUPT,
KVM_DEBUG_GUEST, KVM_SET_SIGNAL_MASK, KVM_GET/SET_FPU
Note that some PPC chips don't have an FPU, so we might need an #ifdef
around KVM_GET/SET_FPU one day.
x86 specific ioctls are:
KVM_GET/SET_LAPIC, KVM_SET_CPUID, KVM_GET/SET_MSRS
An interresting aspect is vcpu_load/vcpu_put. We now have a common
vcpu_load/put which does the preemption stuff, and an architecture
specific kvm_arch_vcpu_load/put. In the x86 case, this one calls the
vmx/svm function defined in kvm_x86_ops.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Reviewed-by: Christian Ehrhardt <ehrhardt@linux.vnet.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-12 00:16:52 +07:00
|
|
|
|
|
|
|
return i;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Read or write a bunch of msrs. Parameters are user addresses.
|
|
|
|
*
|
|
|
|
* @return number of msrs set successfully.
|
|
|
|
*/
|
|
|
|
static int msr_io(struct kvm_vcpu *vcpu, struct kvm_msrs __user *user_msrs,
|
|
|
|
int (*do_msr)(struct kvm_vcpu *vcpu,
|
|
|
|
unsigned index, u64 *data),
|
|
|
|
int writeback)
|
|
|
|
{
|
|
|
|
struct kvm_msrs msrs;
|
|
|
|
struct kvm_msr_entry *entries;
|
|
|
|
int r, n;
|
|
|
|
unsigned size;
|
|
|
|
|
|
|
|
r = -EFAULT;
|
|
|
|
if (copy_from_user(&msrs, user_msrs, sizeof msrs))
|
|
|
|
goto out;
|
|
|
|
|
|
|
|
r = -E2BIG;
|
|
|
|
if (msrs.nmsrs >= MAX_IO_MSRS)
|
|
|
|
goto out;
|
|
|
|
|
|
|
|
size = sizeof(struct kvm_msr_entry) * msrs.nmsrs;
|
2011-12-05 00:36:29 +07:00
|
|
|
entries = memdup_user(user_msrs->entries, size);
|
|
|
|
if (IS_ERR(entries)) {
|
|
|
|
r = PTR_ERR(entries);
|
KVM: Portability: split kvm_vcpu_ioctl
This patch splits kvm_vcpu_ioctl into archtecture independent parts, and
x86 specific parts which go to kvm_arch_vcpu_ioctl in x86.c.
Common ioctls for all architectures are:
KVM_RUN, KVM_GET/SET_(S-)REGS, KVM_TRANSLATE, KVM_INTERRUPT,
KVM_DEBUG_GUEST, KVM_SET_SIGNAL_MASK, KVM_GET/SET_FPU
Note that some PPC chips don't have an FPU, so we might need an #ifdef
around KVM_GET/SET_FPU one day.
x86 specific ioctls are:
KVM_GET/SET_LAPIC, KVM_SET_CPUID, KVM_GET/SET_MSRS
An interresting aspect is vcpu_load/vcpu_put. We now have a common
vcpu_load/put which does the preemption stuff, and an architecture
specific kvm_arch_vcpu_load/put. In the x86 case, this one calls the
vmx/svm function defined in kvm_x86_ops.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Reviewed-by: Christian Ehrhardt <ehrhardt@linux.vnet.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-12 00:16:52 +07:00
|
|
|
goto out;
|
2011-12-05 00:36:29 +07:00
|
|
|
}
|
KVM: Portability: split kvm_vcpu_ioctl
This patch splits kvm_vcpu_ioctl into archtecture independent parts, and
x86 specific parts which go to kvm_arch_vcpu_ioctl in x86.c.
Common ioctls for all architectures are:
KVM_RUN, KVM_GET/SET_(S-)REGS, KVM_TRANSLATE, KVM_INTERRUPT,
KVM_DEBUG_GUEST, KVM_SET_SIGNAL_MASK, KVM_GET/SET_FPU
Note that some PPC chips don't have an FPU, so we might need an #ifdef
around KVM_GET/SET_FPU one day.
x86 specific ioctls are:
KVM_GET/SET_LAPIC, KVM_SET_CPUID, KVM_GET/SET_MSRS
An interresting aspect is vcpu_load/vcpu_put. We now have a common
vcpu_load/put which does the preemption stuff, and an architecture
specific kvm_arch_vcpu_load/put. In the x86 case, this one calls the
vmx/svm function defined in kvm_x86_ops.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Reviewed-by: Christian Ehrhardt <ehrhardt@linux.vnet.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-12 00:16:52 +07:00
|
|
|
|
|
|
|
r = n = __msr_io(vcpu, &msrs, entries, do_msr);
|
|
|
|
if (r < 0)
|
|
|
|
goto out_free;
|
|
|
|
|
|
|
|
r = -EFAULT;
|
|
|
|
if (writeback && copy_to_user(user_msrs->entries, entries, size))
|
|
|
|
goto out_free;
|
|
|
|
|
|
|
|
r = n;
|
|
|
|
|
|
|
|
out_free:
|
2010-07-23 03:24:52 +07:00
|
|
|
kfree(entries);
|
KVM: Portability: split kvm_vcpu_ioctl
This patch splits kvm_vcpu_ioctl into archtecture independent parts, and
x86 specific parts which go to kvm_arch_vcpu_ioctl in x86.c.
Common ioctls for all architectures are:
KVM_RUN, KVM_GET/SET_(S-)REGS, KVM_TRANSLATE, KVM_INTERRUPT,
KVM_DEBUG_GUEST, KVM_SET_SIGNAL_MASK, KVM_GET/SET_FPU
Note that some PPC chips don't have an FPU, so we might need an #ifdef
around KVM_GET/SET_FPU one day.
x86 specific ioctls are:
KVM_GET/SET_LAPIC, KVM_SET_CPUID, KVM_GET/SET_MSRS
An interresting aspect is vcpu_load/vcpu_put. We now have a common
vcpu_load/put which does the preemption stuff, and an architecture
specific kvm_arch_vcpu_load/put. In the x86 case, this one calls the
vmx/svm function defined in kvm_x86_ops.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Reviewed-by: Christian Ehrhardt <ehrhardt@linux.vnet.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-12 00:16:52 +07:00
|
|
|
out:
|
|
|
|
return r;
|
|
|
|
}
|
|
|
|
|
2014-07-14 23:27:35 +07:00
|
|
|
int kvm_vm_ioctl_check_extension(struct kvm *kvm, long ext)
|
2007-11-15 22:07:47 +07:00
|
|
|
{
|
|
|
|
int r;
|
|
|
|
|
|
|
|
switch (ext) {
|
|
|
|
case KVM_CAP_IRQCHIP:
|
|
|
|
case KVM_CAP_HLT:
|
|
|
|
case KVM_CAP_MMU_SHADOW_CACHE_CONTROL:
|
|
|
|
case KVM_CAP_SET_TSS_ADDR:
|
2007-11-21 22:10:04 +07:00
|
|
|
case KVM_CAP_EXT_CPUID:
|
2013-09-22 21:44:50 +07:00
|
|
|
case KVM_CAP_EXT_EMUL_CPUID:
|
2009-02-04 23:52:04 +07:00
|
|
|
case KVM_CAP_CLOCKSOURCE:
|
2008-01-28 04:10:22 +07:00
|
|
|
case KVM_CAP_PIT:
|
2008-02-23 00:21:36 +07:00
|
|
|
case KVM_CAP_NOP_IO_DELAY:
|
2008-04-11 23:24:45 +07:00
|
|
|
case KVM_CAP_MP_STATE:
|
2008-07-29 15:30:57 +07:00
|
|
|
case KVM_CAP_SYNC_MMU:
|
2010-12-14 16:57:47 +07:00
|
|
|
case KVM_CAP_USER_NMI:
|
2008-12-31 00:55:06 +07:00
|
|
|
case KVM_CAP_REINJECT_CONTROL:
|
2009-02-04 22:28:14 +07:00
|
|
|
case KVM_CAP_IRQ_INJECT_STATUS:
|
KVM: add ioeventfd support
ioeventfd is a mechanism to register PIO/MMIO regions to trigger an eventfd
signal when written to by a guest. Host userspace can register any
arbitrary IO address with a corresponding eventfd and then pass the eventfd
to a specific end-point of interest for handling.
Normal IO requires a blocking round-trip since the operation may cause
side-effects in the emulated model or may return data to the caller.
Therefore, an IO in KVM traps from the guest to the host, causes a VMX/SVM
"heavy-weight" exit back to userspace, and is ultimately serviced by qemu's
device model synchronously before returning control back to the vcpu.
However, there is a subclass of IO which acts purely as a trigger for
other IO (such as to kick off an out-of-band DMA request, etc). For these
patterns, the synchronous call is particularly expensive since we really
only want to simply get our notification transmitted asychronously and
return as quickly as possible. All the sychronous infrastructure to ensure
proper data-dependencies are met in the normal IO case are just unecessary
overhead for signalling. This adds additional computational load on the
system, as well as latency to the signalling path.
Therefore, we provide a mechanism for registration of an in-kernel trigger
point that allows the VCPU to only require a very brief, lightweight
exit just long enough to signal an eventfd. This also means that any
clients compatible with the eventfd interface (which includes userspace
and kernelspace equally well) can now register to be notified. The end
result should be a more flexible and higher performance notification API
for the backend KVM hypervisor and perhipheral components.
To test this theory, we built a test-harness called "doorbell". This
module has a function called "doorbell_ring()" which simply increments a
counter for each time the doorbell is signaled. It supports signalling
from either an eventfd, or an ioctl().
We then wired up two paths to the doorbell: One via QEMU via a registered
io region and through the doorbell ioctl(). The other is direct via
ioeventfd.
You can download this test harness here:
ftp://ftp.novell.com/dev/ghaskins/doorbell.tar.bz2
The measured results are as follows:
qemu-mmio: 110000 iops, 9.09us rtt
ioeventfd-mmio: 200100 iops, 5.00us rtt
ioeventfd-pio: 367300 iops, 2.72us rtt
I didn't measure qemu-pio, because I have to figure out how to register a
PIO region with qemu's device model, and I got lazy. However, for now we
can extrapolate based on the data from the NULLIO runs of +2.56us for MMIO,
and -350ns for HC, we get:
qemu-pio: 153139 iops, 6.53us rtt
ioeventfd-hc: 412585 iops, 2.37us rtt
these are just for fun, for now, until I can gather more data.
Here is a graph for your convenience:
http://developer.novell.com/wiki/images/7/76/Iofd-chart.png
The conclusion to draw is that we save about 4us by skipping the userspace
hop.
--------------------
Signed-off-by: Gregory Haskins <ghaskins@novell.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
2009-07-08 04:08:49 +07:00
|
|
|
case KVM_CAP_IOEVENTFD:
|
2014-04-01 01:50:38 +07:00
|
|
|
case KVM_CAP_IOEVENTFD_NO_LENGTH:
|
2009-05-15 03:42:53 +07:00
|
|
|
case KVM_CAP_PIT2:
|
2009-07-07 22:50:38 +07:00
|
|
|
case KVM_CAP_PIT_STATE2:
|
2009-07-21 09:42:48 +07:00
|
|
|
case KVM_CAP_SET_IDENTITY_MAP_ADDR:
|
2009-10-16 05:21:43 +07:00
|
|
|
case KVM_CAP_XEN_HVM:
|
2009-10-17 02:28:36 +07:00
|
|
|
case KVM_CAP_ADJUST_CLOCK:
|
2009-11-12 07:04:25 +07:00
|
|
|
case KVM_CAP_VCPU_EVENTS:
|
2010-01-17 20:51:22 +07:00
|
|
|
case KVM_CAP_HYPERV:
|
2010-01-17 20:51:23 +07:00
|
|
|
case KVM_CAP_HYPERV_VAPIC:
|
2010-01-17 20:51:24 +07:00
|
|
|
case KVM_CAP_HYPERV_SPIN:
|
2015-11-10 19:36:34 +07:00
|
|
|
case KVM_CAP_HYPERV_SYNIC:
|
2010-01-29 13:38:44 +07:00
|
|
|
case KVM_CAP_PCI_SEGMENT:
|
2010-02-15 16:45:43 +07:00
|
|
|
case KVM_CAP_DEBUGREGS:
|
2010-02-23 23:47:57 +07:00
|
|
|
case KVM_CAP_X86_ROBUST_SINGLESTEP:
|
2010-06-13 16:29:39 +07:00
|
|
|
case KVM_CAP_XSAVE:
|
2010-10-14 16:22:50 +07:00
|
|
|
case KVM_CAP_ASYNC_PF:
|
2011-03-25 15:44:51 +07:00
|
|
|
case KVM_CAP_GET_TSC_KHZ:
|
2012-03-11 02:37:27 +07:00
|
|
|
case KVM_CAP_KVMCLOCK_CTRL:
|
2012-08-21 10:02:51 +07:00
|
|
|
case KVM_CAP_READONLY_MEM:
|
2014-01-30 00:10:45 +07:00
|
|
|
case KVM_CAP_HYPERV_TIME:
|
2014-02-28 11:06:17 +07:00
|
|
|
case KVM_CAP_IOAPIC_POLARITY_IGNORED:
|
2015-01-08 21:59:30 +07:00
|
|
|
case KVM_CAP_TSC_DEADLINE_TIMER:
|
2015-04-13 05:53:41 +07:00
|
|
|
case KVM_CAP_ENABLE_CAP_VM:
|
|
|
|
case KVM_CAP_DISABLE_QUIRKS:
|
2015-07-29 16:56:48 +07:00
|
|
|
case KVM_CAP_SET_BOOT_CPU_ID:
|
2015-07-30 13:21:40 +07:00
|
|
|
case KVM_CAP_SPLIT_IRQCHIP:
|
2013-04-17 02:49:18 +07:00
|
|
|
#ifdef CONFIG_KVM_DEVICE_ASSIGNMENT
|
|
|
|
case KVM_CAP_ASSIGN_DEV_IRQ:
|
|
|
|
case KVM_CAP_PCI_2_3:
|
|
|
|
#endif
|
2007-11-15 22:07:47 +07:00
|
|
|
r = 1;
|
|
|
|
break;
|
2015-04-01 19:25:33 +07:00
|
|
|
case KVM_CAP_X86_SMM:
|
|
|
|
/* SMBASE is usually relocated above 1M on modern chipsets,
|
|
|
|
* and SMM handlers might indeed rely on 4G segment limits,
|
|
|
|
* so do not report SMM to be available if real mode is
|
|
|
|
* emulated via vm86 mode. Still, do not go to great lengths
|
|
|
|
* to avoid userspace's usage of the feature, because it is a
|
|
|
|
* fringe case that is not enabled except via specific settings
|
|
|
|
* of the module parameters.
|
|
|
|
*/
|
|
|
|
r = kvm_x86_ops->cpu_has_high_real_mode_segbase();
|
|
|
|
break;
|
2008-05-30 21:05:55 +07:00
|
|
|
case KVM_CAP_COALESCED_MMIO:
|
|
|
|
r = KVM_COALESCED_MMIO_PAGE_OFFSET;
|
|
|
|
break;
|
2007-12-26 18:57:04 +07:00
|
|
|
case KVM_CAP_VAPIC:
|
|
|
|
r = !kvm_x86_ops->cpu_has_accelerated_tpr();
|
|
|
|
break;
|
2008-02-20 16:53:16 +07:00
|
|
|
case KVM_CAP_NR_VCPUS:
|
2011-07-18 21:17:15 +07:00
|
|
|
r = KVM_SOFT_MAX_VCPUS;
|
|
|
|
break;
|
|
|
|
case KVM_CAP_MAX_VCPUS:
|
2008-02-20 16:53:16 +07:00
|
|
|
r = KVM_MAX_VCPUS;
|
|
|
|
break;
|
2008-02-20 16:59:20 +07:00
|
|
|
case KVM_CAP_NR_MEMSLOTS:
|
2012-12-11 00:33:09 +07:00
|
|
|
r = KVM_USER_MEM_SLOTS;
|
2008-02-20 16:59:20 +07:00
|
|
|
break;
|
2009-10-02 05:28:39 +07:00
|
|
|
case KVM_CAP_PV_MMU: /* obsolete */
|
|
|
|
r = 0;
|
2008-02-23 00:21:37 +07:00
|
|
|
break;
|
2013-04-29 23:54:08 +07:00
|
|
|
#ifdef CONFIG_KVM_DEVICE_ASSIGNMENT
|
2008-09-14 07:48:28 +07:00
|
|
|
case KVM_CAP_IOMMU:
|
2011-09-06 23:46:34 +07:00
|
|
|
r = iommu_present(&pci_bus_type);
|
2008-09-14 07:48:28 +07:00
|
|
|
break;
|
2013-04-29 23:54:08 +07:00
|
|
|
#endif
|
2009-05-11 15:48:15 +07:00
|
|
|
case KVM_CAP_MCE:
|
|
|
|
r = KVM_MAX_MCE_BANKS;
|
|
|
|
break;
|
2010-06-13 16:29:39 +07:00
|
|
|
case KVM_CAP_XCRS:
|
2016-04-05 03:25:02 +07:00
|
|
|
r = boot_cpu_has(X86_FEATURE_XSAVE);
|
2010-06-13 16:29:39 +07:00
|
|
|
break;
|
2011-03-25 15:44:51 +07:00
|
|
|
case KVM_CAP_TSC_CONTROL:
|
|
|
|
r = kvm_has_tsc_control;
|
|
|
|
break;
|
2016-07-13 03:09:27 +07:00
|
|
|
case KVM_CAP_X2APIC_API:
|
|
|
|
r = KVM_X2APIC_API_VALID_FLAGS;
|
|
|
|
break;
|
2007-11-15 22:07:47 +07:00
|
|
|
default:
|
|
|
|
r = 0;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
return r;
|
|
|
|
|
|
|
|
}
|
|
|
|
|
2007-10-10 22:16:19 +07:00
|
|
|
long kvm_arch_dev_ioctl(struct file *filp,
|
|
|
|
unsigned int ioctl, unsigned long arg)
|
|
|
|
{
|
|
|
|
void __user *argp = (void __user *)arg;
|
|
|
|
long r;
|
|
|
|
|
|
|
|
switch (ioctl) {
|
|
|
|
case KVM_GET_MSR_INDEX_LIST: {
|
|
|
|
struct kvm_msr_list __user *user_msr_list = argp;
|
|
|
|
struct kvm_msr_list msr_list;
|
|
|
|
unsigned n;
|
|
|
|
|
|
|
|
r = -EFAULT;
|
|
|
|
if (copy_from_user(&msr_list, user_msr_list, sizeof msr_list))
|
|
|
|
goto out;
|
|
|
|
n = msr_list.nmsrs;
|
2015-05-05 17:08:55 +07:00
|
|
|
msr_list.nmsrs = num_msrs_to_save + num_emulated_msrs;
|
2007-10-10 22:16:19 +07:00
|
|
|
if (copy_to_user(user_msr_list, &msr_list, sizeof msr_list))
|
|
|
|
goto out;
|
|
|
|
r = -E2BIG;
|
2009-07-03 02:45:47 +07:00
|
|
|
if (n < msr_list.nmsrs)
|
2007-10-10 22:16:19 +07:00
|
|
|
goto out;
|
|
|
|
r = -EFAULT;
|
|
|
|
if (copy_to_user(user_msr_list->indices, &msrs_to_save,
|
|
|
|
num_msrs_to_save * sizeof(u32)))
|
|
|
|
goto out;
|
2009-07-03 02:45:47 +07:00
|
|
|
if (copy_to_user(user_msr_list->indices + num_msrs_to_save,
|
2007-10-10 22:16:19 +07:00
|
|
|
&emulated_msrs,
|
2015-05-05 17:08:55 +07:00
|
|
|
num_emulated_msrs * sizeof(u32)))
|
2007-10-10 22:16:19 +07:00
|
|
|
goto out;
|
|
|
|
r = 0;
|
|
|
|
break;
|
|
|
|
}
|
2013-09-22 21:44:50 +07:00
|
|
|
case KVM_GET_SUPPORTED_CPUID:
|
|
|
|
case KVM_GET_EMULATED_CPUID: {
|
2008-02-11 23:37:23 +07:00
|
|
|
struct kvm_cpuid2 __user *cpuid_arg = argp;
|
|
|
|
struct kvm_cpuid2 cpuid;
|
|
|
|
|
|
|
|
r = -EFAULT;
|
|
|
|
if (copy_from_user(&cpuid, cpuid_arg, sizeof cpuid))
|
|
|
|
goto out;
|
2013-09-22 21:44:50 +07:00
|
|
|
|
|
|
|
r = kvm_dev_ioctl_get_cpuid(&cpuid, cpuid_arg->entries,
|
|
|
|
ioctl);
|
2008-02-11 23:37:23 +07:00
|
|
|
if (r)
|
|
|
|
goto out;
|
|
|
|
|
|
|
|
r = -EFAULT;
|
|
|
|
if (copy_to_user(cpuid_arg, &cpuid, sizeof cpuid))
|
|
|
|
goto out;
|
|
|
|
r = 0;
|
|
|
|
break;
|
|
|
|
}
|
2009-05-11 15:48:15 +07:00
|
|
|
case KVM_X86_GET_MCE_CAP_SUPPORTED: {
|
|
|
|
r = -EFAULT;
|
2016-06-22 13:59:56 +07:00
|
|
|
if (copy_to_user(argp, &kvm_mce_cap_supported,
|
|
|
|
sizeof(kvm_mce_cap_supported)))
|
2009-05-11 15:48:15 +07:00
|
|
|
goto out;
|
|
|
|
r = 0;
|
|
|
|
break;
|
|
|
|
}
|
2007-10-10 22:16:19 +07:00
|
|
|
default:
|
|
|
|
r = -EINVAL;
|
|
|
|
}
|
|
|
|
out:
|
|
|
|
return r;
|
|
|
|
}
|
|
|
|
|
2010-06-30 11:25:15 +07:00
|
|
|
static void wbinvd_ipi(void *garbage)
|
|
|
|
{
|
|
|
|
wbinvd();
|
|
|
|
}
|
|
|
|
|
|
|
|
static bool need_emulate_wbinvd(struct kvm_vcpu *vcpu)
|
|
|
|
{
|
2013-10-31 00:02:30 +07:00
|
|
|
return kvm_arch_has_noncoherent_dma(vcpu->kvm);
|
2010-06-30 11:25:15 +07:00
|
|
|
}
|
|
|
|
|
2016-01-07 21:05:10 +07:00
|
|
|
static inline void kvm_migrate_timers(struct kvm_vcpu *vcpu)
|
|
|
|
{
|
|
|
|
set_bit(KVM_REQ_MIGRATE_TIMER, &vcpu->requests);
|
|
|
|
}
|
|
|
|
|
KVM: Portability: split kvm_vcpu_ioctl
This patch splits kvm_vcpu_ioctl into archtecture independent parts, and
x86 specific parts which go to kvm_arch_vcpu_ioctl in x86.c.
Common ioctls for all architectures are:
KVM_RUN, KVM_GET/SET_(S-)REGS, KVM_TRANSLATE, KVM_INTERRUPT,
KVM_DEBUG_GUEST, KVM_SET_SIGNAL_MASK, KVM_GET/SET_FPU
Note that some PPC chips don't have an FPU, so we might need an #ifdef
around KVM_GET/SET_FPU one day.
x86 specific ioctls are:
KVM_GET/SET_LAPIC, KVM_SET_CPUID, KVM_GET/SET_MSRS
An interresting aspect is vcpu_load/vcpu_put. We now have a common
vcpu_load/put which does the preemption stuff, and an architecture
specific kvm_arch_vcpu_load/put. In the x86 case, this one calls the
vmx/svm function defined in kvm_x86_ops.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Reviewed-by: Christian Ehrhardt <ehrhardt@linux.vnet.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-12 00:16:52 +07:00
|
|
|
void kvm_arch_vcpu_load(struct kvm_vcpu *vcpu, int cpu)
|
|
|
|
{
|
2010-06-30 11:25:15 +07:00
|
|
|
/* Address WBINVD may be executed by guest */
|
|
|
|
if (need_emulate_wbinvd(vcpu)) {
|
|
|
|
if (kvm_x86_ops->has_wbinvd_exit())
|
|
|
|
cpumask_set_cpu(cpu, vcpu->arch.wbinvd_dirty_mask);
|
|
|
|
else if (vcpu->cpu != -1 && vcpu->cpu != cpu)
|
|
|
|
smp_call_function_single(vcpu->cpu,
|
|
|
|
wbinvd_ipi, NULL, 1);
|
|
|
|
}
|
|
|
|
|
KVM: Portability: split kvm_vcpu_ioctl
This patch splits kvm_vcpu_ioctl into archtecture independent parts, and
x86 specific parts which go to kvm_arch_vcpu_ioctl in x86.c.
Common ioctls for all architectures are:
KVM_RUN, KVM_GET/SET_(S-)REGS, KVM_TRANSLATE, KVM_INTERRUPT,
KVM_DEBUG_GUEST, KVM_SET_SIGNAL_MASK, KVM_GET/SET_FPU
Note that some PPC chips don't have an FPU, so we might need an #ifdef
around KVM_GET/SET_FPU one day.
x86 specific ioctls are:
KVM_GET/SET_LAPIC, KVM_SET_CPUID, KVM_GET/SET_MSRS
An interresting aspect is vcpu_load/vcpu_put. We now have a common
vcpu_load/put which does the preemption stuff, and an architecture
specific kvm_arch_vcpu_load/put. In the x86 case, this one calls the
vmx/svm function defined in kvm_x86_ops.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Reviewed-by: Christian Ehrhardt <ehrhardt@linux.vnet.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-12 00:16:52 +07:00
|
|
|
kvm_x86_ops->vcpu_load(vcpu, cpu);
|
2011-03-25 15:44:48 +07:00
|
|
|
|
2012-02-04 00:43:56 +07:00
|
|
|
/* Apply any externally detected TSC adjustments (due to suspend) */
|
|
|
|
if (unlikely(vcpu->arch.tsc_offset_adjustment)) {
|
|
|
|
adjust_tsc_offset_host(vcpu, vcpu->arch.tsc_offset_adjustment);
|
|
|
|
vcpu->arch.tsc_offset_adjustment = 0;
|
2014-09-12 12:43:19 +07:00
|
|
|
kvm_make_request(KVM_REQ_CLOCK_UPDATE, vcpu);
|
2012-02-04 00:43:56 +07:00
|
|
|
}
|
2011-03-25 15:44:48 +07:00
|
|
|
|
2010-08-20 15:07:24 +07:00
|
|
|
if (unlikely(vcpu->cpu != cpu) || check_tsc_unstable()) {
|
2012-02-04 00:43:54 +07:00
|
|
|
s64 tsc_delta = !vcpu->arch.last_host_tsc ? 0 :
|
2015-06-25 23:44:07 +07:00
|
|
|
rdtsc() - vcpu->arch.last_host_tsc;
|
2010-08-20 15:07:23 +07:00
|
|
|
if (tsc_delta < 0)
|
|
|
|
mark_tsc_unstable("KVM discovered backwards TSC");
|
2016-06-14 04:20:01 +07:00
|
|
|
|
|
|
|
if (kvm_lapic_hv_timer_in_use(vcpu) &&
|
|
|
|
kvm_x86_ops->set_hv_timer(vcpu,
|
|
|
|
kvm_get_lapic_tscdeadline_msr(vcpu)))
|
|
|
|
kvm_lapic_switch_to_sw_timer(vcpu);
|
2010-09-19 07:38:15 +07:00
|
|
|
if (check_tsc_unstable()) {
|
2015-10-20 14:39:05 +07:00
|
|
|
u64 offset = kvm_compute_tsc_offset(vcpu,
|
KVM: Fix last_guest_tsc / tsc_offset semantics
The variable last_guest_tsc was being used as an ad-hoc indicator
that guest TSC has been initialized and recorded correctly. However,
it may not have been, it could be that guest TSC has been set to some
large value, the back to a small value (by, say, a software reboot).
This defeats the logic and causes KVM to falsely assume that the
guest TSC has gone backwards, marking the host TSC unstable, which
is undesirable behavior.
In addition, rather than try to compute an offset adjustment for the
TSC on unstable platforms, just recompute the whole offset. This
allows us to get rid of one callsite for adjust_tsc_offset, which
is problematic because the units it takes are in guest units, but
here, the computation was originally being done in host units.
Doing this, and also recording last_guest_tsc when the TSC is written
allow us to remove the tricky logic which depended on last_guest_tsc
being zero to indicate a reset of uninitialized value.
Instead, we now have the guarantee that the guest TSC offset is
always at least something which will get us last_guest_tsc.
Signed-off-by: Zachary Amsden <zamsden@gmail.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
2012-02-04 00:43:53 +07:00
|
|
|
vcpu->arch.last_guest_tsc);
|
|
|
|
kvm_x86_ops->write_tsc_offset(vcpu, offset);
|
2010-09-19 07:38:15 +07:00
|
|
|
vcpu->arch.tsc_catchup = 1;
|
|
|
|
}
|
2012-11-28 08:29:04 +07:00
|
|
|
/*
|
|
|
|
* On a host with synchronized TSC, there is no need to update
|
|
|
|
* kvmclock on vcpu->cpu migration
|
|
|
|
*/
|
|
|
|
if (!vcpu->kvm->arch.use_master_clock || vcpu->cpu == -1)
|
2013-05-10 06:21:41 +07:00
|
|
|
kvm_make_request(KVM_REQ_GLOBAL_CLOCK_UPDATE, vcpu);
|
2010-09-19 07:38:15 +07:00
|
|
|
if (vcpu->cpu != cpu)
|
|
|
|
kvm_migrate_timers(vcpu);
|
2010-08-20 15:07:23 +07:00
|
|
|
vcpu->cpu = cpu;
|
2009-10-10 09:26:08 +07:00
|
|
|
}
|
2011-07-12 02:28:14 +07:00
|
|
|
|
|
|
|
kvm_make_request(KVM_REQ_STEAL_UPDATE, vcpu);
|
KVM: Portability: split kvm_vcpu_ioctl
This patch splits kvm_vcpu_ioctl into archtecture independent parts, and
x86 specific parts which go to kvm_arch_vcpu_ioctl in x86.c.
Common ioctls for all architectures are:
KVM_RUN, KVM_GET/SET_(S-)REGS, KVM_TRANSLATE, KVM_INTERRUPT,
KVM_DEBUG_GUEST, KVM_SET_SIGNAL_MASK, KVM_GET/SET_FPU
Note that some PPC chips don't have an FPU, so we might need an #ifdef
around KVM_GET/SET_FPU one day.
x86 specific ioctls are:
KVM_GET/SET_LAPIC, KVM_SET_CPUID, KVM_GET/SET_MSRS
An interresting aspect is vcpu_load/vcpu_put. We now have a common
vcpu_load/put which does the preemption stuff, and an architecture
specific kvm_arch_vcpu_load/put. In the x86 case, this one calls the
vmx/svm function defined in kvm_x86_ops.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Reviewed-by: Christian Ehrhardt <ehrhardt@linux.vnet.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-12 00:16:52 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
void kvm_arch_vcpu_put(struct kvm_vcpu *vcpu)
|
|
|
|
{
|
2009-12-30 17:40:26 +07:00
|
|
|
kvm_x86_ops->vcpu_put(vcpu);
|
2010-05-03 20:05:44 +07:00
|
|
|
kvm_put_guest_fpu(vcpu);
|
2015-06-25 23:44:07 +07:00
|
|
|
vcpu->arch.last_host_tsc = rdtsc();
|
KVM: Portability: split kvm_vcpu_ioctl
This patch splits kvm_vcpu_ioctl into archtecture independent parts, and
x86 specific parts which go to kvm_arch_vcpu_ioctl in x86.c.
Common ioctls for all architectures are:
KVM_RUN, KVM_GET/SET_(S-)REGS, KVM_TRANSLATE, KVM_INTERRUPT,
KVM_DEBUG_GUEST, KVM_SET_SIGNAL_MASK, KVM_GET/SET_FPU
Note that some PPC chips don't have an FPU, so we might need an #ifdef
around KVM_GET/SET_FPU one day.
x86 specific ioctls are:
KVM_GET/SET_LAPIC, KVM_SET_CPUID, KVM_GET/SET_MSRS
An interresting aspect is vcpu_load/vcpu_put. We now have a common
vcpu_load/put which does the preemption stuff, and an architecture
specific kvm_arch_vcpu_load/put. In the x86 case, this one calls the
vmx/svm function defined in kvm_x86_ops.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Reviewed-by: Christian Ehrhardt <ehrhardt@linux.vnet.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-12 00:16:52 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
static int kvm_vcpu_ioctl_get_lapic(struct kvm_vcpu *vcpu,
|
|
|
|
struct kvm_lapic_state *s)
|
|
|
|
{
|
2015-11-10 19:36:33 +07:00
|
|
|
if (vcpu->arch.apicv_active)
|
|
|
|
kvm_x86_ops->sync_pir_to_irr(vcpu);
|
|
|
|
|
2016-07-13 03:09:22 +07:00
|
|
|
return kvm_apic_get_state(vcpu, s);
|
KVM: Portability: split kvm_vcpu_ioctl
This patch splits kvm_vcpu_ioctl into archtecture independent parts, and
x86 specific parts which go to kvm_arch_vcpu_ioctl in x86.c.
Common ioctls for all architectures are:
KVM_RUN, KVM_GET/SET_(S-)REGS, KVM_TRANSLATE, KVM_INTERRUPT,
KVM_DEBUG_GUEST, KVM_SET_SIGNAL_MASK, KVM_GET/SET_FPU
Note that some PPC chips don't have an FPU, so we might need an #ifdef
around KVM_GET/SET_FPU one day.
x86 specific ioctls are:
KVM_GET/SET_LAPIC, KVM_SET_CPUID, KVM_GET/SET_MSRS
An interresting aspect is vcpu_load/vcpu_put. We now have a common
vcpu_load/put which does the preemption stuff, and an architecture
specific kvm_arch_vcpu_load/put. In the x86 case, this one calls the
vmx/svm function defined in kvm_x86_ops.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Reviewed-by: Christian Ehrhardt <ehrhardt@linux.vnet.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-12 00:16:52 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
static int kvm_vcpu_ioctl_set_lapic(struct kvm_vcpu *vcpu,
|
|
|
|
struct kvm_lapic_state *s)
|
|
|
|
{
|
2016-07-13 03:09:22 +07:00
|
|
|
int r;
|
|
|
|
|
|
|
|
r = kvm_apic_set_state(vcpu, s);
|
|
|
|
if (r)
|
|
|
|
return r;
|
2009-08-09 19:17:40 +07:00
|
|
|
update_cr8_intercept(vcpu);
|
KVM: Portability: split kvm_vcpu_ioctl
This patch splits kvm_vcpu_ioctl into archtecture independent parts, and
x86 specific parts which go to kvm_arch_vcpu_ioctl in x86.c.
Common ioctls for all architectures are:
KVM_RUN, KVM_GET/SET_(S-)REGS, KVM_TRANSLATE, KVM_INTERRUPT,
KVM_DEBUG_GUEST, KVM_SET_SIGNAL_MASK, KVM_GET/SET_FPU
Note that some PPC chips don't have an FPU, so we might need an #ifdef
around KVM_GET/SET_FPU one day.
x86 specific ioctls are:
KVM_GET/SET_LAPIC, KVM_SET_CPUID, KVM_GET/SET_MSRS
An interresting aspect is vcpu_load/vcpu_put. We now have a common
vcpu_load/put which does the preemption stuff, and an architecture
specific kvm_arch_vcpu_load/put. In the x86 case, this one calls the
vmx/svm function defined in kvm_x86_ops.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Reviewed-by: Christian Ehrhardt <ehrhardt@linux.vnet.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-12 00:16:52 +07:00
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2015-11-17 23:32:05 +07:00
|
|
|
static int kvm_cpu_accept_dm_intr(struct kvm_vcpu *vcpu)
|
|
|
|
{
|
|
|
|
return (!lapic_in_kernel(vcpu) ||
|
|
|
|
kvm_apic_accept_pic_intr(vcpu));
|
|
|
|
}
|
|
|
|
|
2015-11-17 06:26:00 +07:00
|
|
|
/*
|
|
|
|
* if userspace requested an interrupt window, check that the
|
|
|
|
* interrupt window is open.
|
|
|
|
*
|
|
|
|
* No need to exit to userspace if we already have an interrupt queued.
|
|
|
|
*/
|
|
|
|
static int kvm_vcpu_ready_for_interrupt_injection(struct kvm_vcpu *vcpu)
|
|
|
|
{
|
|
|
|
return kvm_arch_interrupt_allowed(vcpu) &&
|
|
|
|
!kvm_cpu_has_interrupt(vcpu) &&
|
|
|
|
!kvm_event_needs_reinjection(vcpu) &&
|
|
|
|
kvm_cpu_accept_dm_intr(vcpu);
|
|
|
|
}
|
|
|
|
|
2007-11-21 03:36:41 +07:00
|
|
|
static int kvm_vcpu_ioctl_interrupt(struct kvm_vcpu *vcpu,
|
|
|
|
struct kvm_interrupt *irq)
|
|
|
|
{
|
2013-02-27 10:33:25 +07:00
|
|
|
if (irq->irq >= KVM_NR_INTERRUPTS)
|
2007-11-21 03:36:41 +07:00
|
|
|
return -EINVAL;
|
KVM: x86: Add support for local interrupt requests from userspace
In order to enable userspace PIC support, the userspace PIC needs to
be able to inject local interrupts even when the APICs are in the
kernel.
KVM_INTERRUPT now supports sending local interrupts to an APIC when
APICs are in the kernel.
The ready_for_interrupt_request flag is now only set when the CPU/APIC
will immediately accept and inject an interrupt (i.e. APIC has not
masked the PIC).
When the PIC wishes to initiate an INTA cycle with, say, CPU0, it
kicks CPU0 out of the guest, and renedezvous with CPU0 once it arrives
in userspace.
When the CPU/APIC unmasks the PIC, a KVM_EXIT_IRQ_WINDOW_OPEN is
triggered, so that userspace has a chance to inject a PIC interrupt
if it had been pending.
Overall, this design can lead to a small number of spurious userspace
renedezvous. In particular, whenever the PIC transistions from low to
high while it is masked and whenever the PIC becomes unmasked while
it is low.
Note: this does not buffer more than one local interrupt in the
kernel, so the VMM needs to enter the guest in order to complete
interrupt injection before injecting an additional interrupt.
Compiles for x86.
Can pass the KVM Unit Tests.
Signed-off-by: Steve Rutherford <srutherford@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2015-07-30 16:27:16 +07:00
|
|
|
|
|
|
|
if (!irqchip_in_kernel(vcpu->kvm)) {
|
|
|
|
kvm_queue_interrupt(vcpu, irq->irq, false);
|
|
|
|
kvm_make_request(KVM_REQ_EVENT, vcpu);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* With in-kernel LAPIC, we only use this to inject EXTINT, so
|
|
|
|
* fail for in-kernel 8259.
|
|
|
|
*/
|
|
|
|
if (pic_in_kernel(vcpu->kvm))
|
2007-11-21 03:36:41 +07:00
|
|
|
return -ENXIO;
|
|
|
|
|
KVM: x86: Add support for local interrupt requests from userspace
In order to enable userspace PIC support, the userspace PIC needs to
be able to inject local interrupts even when the APICs are in the
kernel.
KVM_INTERRUPT now supports sending local interrupts to an APIC when
APICs are in the kernel.
The ready_for_interrupt_request flag is now only set when the CPU/APIC
will immediately accept and inject an interrupt (i.e. APIC has not
masked the PIC).
When the PIC wishes to initiate an INTA cycle with, say, CPU0, it
kicks CPU0 out of the guest, and renedezvous with CPU0 once it arrives
in userspace.
When the CPU/APIC unmasks the PIC, a KVM_EXIT_IRQ_WINDOW_OPEN is
triggered, so that userspace has a chance to inject a PIC interrupt
if it had been pending.
Overall, this design can lead to a small number of spurious userspace
renedezvous. In particular, whenever the PIC transistions from low to
high while it is masked and whenever the PIC becomes unmasked while
it is low.
Note: this does not buffer more than one local interrupt in the
kernel, so the VMM needs to enter the guest in order to complete
interrupt injection before injecting an additional interrupt.
Compiles for x86.
Can pass the KVM Unit Tests.
Signed-off-by: Steve Rutherford <srutherford@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2015-07-30 16:27:16 +07:00
|
|
|
if (vcpu->arch.pending_external_vector != -1)
|
|
|
|
return -EEXIST;
|
2007-11-21 03:36:41 +07:00
|
|
|
|
KVM: x86: Add support for local interrupt requests from userspace
In order to enable userspace PIC support, the userspace PIC needs to
be able to inject local interrupts even when the APICs are in the
kernel.
KVM_INTERRUPT now supports sending local interrupts to an APIC when
APICs are in the kernel.
The ready_for_interrupt_request flag is now only set when the CPU/APIC
will immediately accept and inject an interrupt (i.e. APIC has not
masked the PIC).
When the PIC wishes to initiate an INTA cycle with, say, CPU0, it
kicks CPU0 out of the guest, and renedezvous with CPU0 once it arrives
in userspace.
When the CPU/APIC unmasks the PIC, a KVM_EXIT_IRQ_WINDOW_OPEN is
triggered, so that userspace has a chance to inject a PIC interrupt
if it had been pending.
Overall, this design can lead to a small number of spurious userspace
renedezvous. In particular, whenever the PIC transistions from low to
high while it is masked and whenever the PIC becomes unmasked while
it is low.
Note: this does not buffer more than one local interrupt in the
kernel, so the VMM needs to enter the guest in order to complete
interrupt injection before injecting an additional interrupt.
Compiles for x86.
Can pass the KVM Unit Tests.
Signed-off-by: Steve Rutherford <srutherford@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2015-07-30 16:27:16 +07:00
|
|
|
vcpu->arch.pending_external_vector = irq->irq;
|
2015-11-17 06:26:05 +07:00
|
|
|
kvm_make_request(KVM_REQ_EVENT, vcpu);
|
2007-11-21 03:36:41 +07:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2008-09-26 14:30:55 +07:00
|
|
|
static int kvm_vcpu_ioctl_nmi(struct kvm_vcpu *vcpu)
|
|
|
|
{
|
|
|
|
kvm_inject_nmi(vcpu);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2015-04-01 20:06:40 +07:00
|
|
|
static int kvm_vcpu_ioctl_smi(struct kvm_vcpu *vcpu)
|
|
|
|
{
|
2015-05-07 16:36:11 +07:00
|
|
|
kvm_make_request(KVM_REQ_SMI, vcpu);
|
|
|
|
|
2015-04-01 20:06:40 +07:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2007-10-22 21:50:39 +07:00
|
|
|
static int vcpu_ioctl_tpr_access_reporting(struct kvm_vcpu *vcpu,
|
|
|
|
struct kvm_tpr_access_ctl *tac)
|
|
|
|
{
|
|
|
|
if (tac->flags)
|
|
|
|
return -EINVAL;
|
|
|
|
vcpu->arch.tpr_access_reporting = !!tac->enabled;
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2009-05-11 15:48:15 +07:00
|
|
|
static int kvm_vcpu_ioctl_x86_setup_mce(struct kvm_vcpu *vcpu,
|
|
|
|
u64 mcg_cap)
|
|
|
|
{
|
|
|
|
int r;
|
|
|
|
unsigned bank_num = mcg_cap & 0xff, bank;
|
|
|
|
|
|
|
|
r = -EINVAL;
|
2009-10-23 14:37:00 +07:00
|
|
|
if (!bank_num || bank_num >= KVM_MAX_MCE_BANKS)
|
2009-05-11 15:48:15 +07:00
|
|
|
goto out;
|
2016-06-22 13:59:56 +07:00
|
|
|
if (mcg_cap & ~(kvm_mce_cap_supported | 0xff | 0xff0000))
|
2009-05-11 15:48:15 +07:00
|
|
|
goto out;
|
|
|
|
r = 0;
|
|
|
|
vcpu->arch.mcg_cap = mcg_cap;
|
|
|
|
/* Init IA32_MCG_CTL to all 1s */
|
|
|
|
if (mcg_cap & MCG_CTL_P)
|
|
|
|
vcpu->arch.mcg_ctl = ~(u64)0;
|
|
|
|
/* Init IA32_MCi_CTL to all 1s */
|
|
|
|
for (bank = 0; bank < bank_num; bank++)
|
|
|
|
vcpu->arch.mce_banks[bank*4] = ~(u64)0;
|
2016-06-22 13:59:56 +07:00
|
|
|
|
|
|
|
if (kvm_x86_ops->setup_mce)
|
|
|
|
kvm_x86_ops->setup_mce(vcpu);
|
2009-05-11 15:48:15 +07:00
|
|
|
out:
|
|
|
|
return r;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int kvm_vcpu_ioctl_x86_set_mce(struct kvm_vcpu *vcpu,
|
|
|
|
struct kvm_x86_mce *mce)
|
|
|
|
{
|
|
|
|
u64 mcg_cap = vcpu->arch.mcg_cap;
|
|
|
|
unsigned bank_num = mcg_cap & 0xff;
|
|
|
|
u64 *banks = vcpu->arch.mce_banks;
|
|
|
|
|
|
|
|
if (mce->bank >= bank_num || !(mce->status & MCI_STATUS_VAL))
|
|
|
|
return -EINVAL;
|
|
|
|
/*
|
|
|
|
* if IA32_MCG_CTL is not all 1s, the uncorrected error
|
|
|
|
* reporting is disabled
|
|
|
|
*/
|
|
|
|
if ((mce->status & MCI_STATUS_UC) && (mcg_cap & MCG_CTL_P) &&
|
|
|
|
vcpu->arch.mcg_ctl != ~(u64)0)
|
|
|
|
return 0;
|
|
|
|
banks += 4 * mce->bank;
|
|
|
|
/*
|
|
|
|
* if IA32_MCi_CTL is not all 1s, the uncorrected error
|
|
|
|
* reporting is disabled for the bank
|
|
|
|
*/
|
|
|
|
if ((mce->status & MCI_STATUS_UC) && banks[0] != ~(u64)0)
|
|
|
|
return 0;
|
|
|
|
if (mce->status & MCI_STATUS_UC) {
|
|
|
|
if ((vcpu->arch.mcg_status & MCG_STATUS_MCIP) ||
|
2009-12-07 17:16:48 +07:00
|
|
|
!kvm_read_cr4_bits(vcpu, X86_CR4_MCE)) {
|
2010-05-10 16:34:53 +07:00
|
|
|
kvm_make_request(KVM_REQ_TRIPLE_FAULT, vcpu);
|
2009-05-11 15:48:15 +07:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
if (banks[1] & MCI_STATUS_VAL)
|
|
|
|
mce->status |= MCI_STATUS_OVER;
|
|
|
|
banks[2] = mce->addr;
|
|
|
|
banks[3] = mce->misc;
|
|
|
|
vcpu->arch.mcg_status = mce->mcg_status;
|
|
|
|
banks[1] = mce->status;
|
|
|
|
kvm_queue_exception(vcpu, MC_VECTOR);
|
|
|
|
} else if (!(banks[1] & MCI_STATUS_VAL)
|
|
|
|
|| !(banks[1] & MCI_STATUS_UC)) {
|
|
|
|
if (banks[1] & MCI_STATUS_VAL)
|
|
|
|
mce->status |= MCI_STATUS_OVER;
|
|
|
|
banks[2] = mce->addr;
|
|
|
|
banks[3] = mce->misc;
|
|
|
|
banks[1] = mce->status;
|
|
|
|
} else
|
|
|
|
banks[1] |= MCI_STATUS_OVER;
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2009-11-12 07:04:25 +07:00
|
|
|
static void kvm_vcpu_ioctl_x86_get_vcpu_events(struct kvm_vcpu *vcpu,
|
|
|
|
struct kvm_vcpu_events *events)
|
|
|
|
{
|
2011-09-20 17:43:14 +07:00
|
|
|
process_nmi(vcpu);
|
2010-02-15 16:45:41 +07:00
|
|
|
events->exception.injected =
|
|
|
|
vcpu->arch.exception.pending &&
|
|
|
|
!kvm_exception_is_soft(vcpu->arch.exception.nr);
|
2009-11-12 07:04:25 +07:00
|
|
|
events->exception.nr = vcpu->arch.exception.nr;
|
|
|
|
events->exception.has_error_code = vcpu->arch.exception.has_error_code;
|
2010-10-31 01:54:47 +07:00
|
|
|
events->exception.pad = 0;
|
2009-11-12 07:04:25 +07:00
|
|
|
events->exception.error_code = vcpu->arch.exception.error_code;
|
|
|
|
|
2010-02-15 16:45:41 +07:00
|
|
|
events->interrupt.injected =
|
|
|
|
vcpu->arch.interrupt.pending && !vcpu->arch.interrupt.soft;
|
2009-11-12 07:04:25 +07:00
|
|
|
events->interrupt.nr = vcpu->arch.interrupt.nr;
|
2010-02-15 16:45:41 +07:00
|
|
|
events->interrupt.soft = 0;
|
2014-05-20 19:29:47 +07:00
|
|
|
events->interrupt.shadow = kvm_x86_ops->get_interrupt_shadow(vcpu);
|
2009-11-12 07:04:25 +07:00
|
|
|
|
|
|
|
events->nmi.injected = vcpu->arch.nmi_injected;
|
2011-09-20 17:43:14 +07:00
|
|
|
events->nmi.pending = vcpu->arch.nmi_pending != 0;
|
2009-11-12 07:04:25 +07:00
|
|
|
events->nmi.masked = kvm_x86_ops->get_nmi_mask(vcpu);
|
2010-10-31 01:54:47 +07:00
|
|
|
events->nmi.pad = 0;
|
2009-11-12 07:04:25 +07:00
|
|
|
|
2013-03-13 18:42:34 +07:00
|
|
|
events->sipi_vector = 0; /* never valid when reporting to user space */
|
2009-11-12 07:04:25 +07:00
|
|
|
|
2015-04-01 20:06:40 +07:00
|
|
|
events->smi.smm = is_smm(vcpu);
|
|
|
|
events->smi.pending = vcpu->arch.smi_pending;
|
|
|
|
events->smi.smm_inside_nmi =
|
|
|
|
!!(vcpu->arch.hflags & HF_SMM_INSIDE_NMI_MASK);
|
|
|
|
events->smi.latched_init = kvm_lapic_latched_init(vcpu);
|
|
|
|
|
2009-12-07 00:24:15 +07:00
|
|
|
events->flags = (KVM_VCPUEVENT_VALID_NMI_PENDING
|
2015-04-01 20:06:40 +07:00
|
|
|
| KVM_VCPUEVENT_VALID_SHADOW
|
|
|
|
| KVM_VCPUEVENT_VALID_SMM);
|
2010-10-31 01:54:47 +07:00
|
|
|
memset(&events->reserved, 0, sizeof(events->reserved));
|
2009-11-12 07:04:25 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
static int kvm_vcpu_ioctl_x86_set_vcpu_events(struct kvm_vcpu *vcpu,
|
|
|
|
struct kvm_vcpu_events *events)
|
|
|
|
{
|
2009-12-07 00:24:15 +07:00
|
|
|
if (events->flags & ~(KVM_VCPUEVENT_VALID_NMI_PENDING
|
2010-02-20 01:38:07 +07:00
|
|
|
| KVM_VCPUEVENT_VALID_SIPI_VECTOR
|
2015-04-01 20:06:40 +07:00
|
|
|
| KVM_VCPUEVENT_VALID_SHADOW
|
|
|
|
| KVM_VCPUEVENT_VALID_SMM))
|
2009-11-12 07:04:25 +07:00
|
|
|
return -EINVAL;
|
|
|
|
|
KVM: fail KVM_SET_VCPU_EVENTS with invalid exception number
This cannot be returned by KVM_GET_VCPU_EVENTS, so it is okay to return
EINVAL. It causes a WARN from exception_type:
WARNING: CPU: 3 PID: 16732 at arch/x86/kvm/x86.c:345 exception_type+0x49/0x50 [kvm]()
CPU: 3 PID: 16732 Comm: a.out Tainted: G W 4.4.6-300.fc23.x86_64 #1
Hardware name: LENOVO 2325F51/2325F51, BIOS G2ET32WW (1.12 ) 05/30/2012
0000000000000286 000000006308a48b ffff8800bec7fcf8 ffffffff813b542e
0000000000000000 ffffffffa0966496 ffff8800bec7fd30 ffffffff810a40f2
ffff8800552a8000 0000000000000000 00000000002c267c 0000000000000001
Call Trace:
[<ffffffff813b542e>] dump_stack+0x63/0x85
[<ffffffff810a40f2>] warn_slowpath_common+0x82/0xc0
[<ffffffff810a423a>] warn_slowpath_null+0x1a/0x20
[<ffffffffa0924809>] exception_type+0x49/0x50 [kvm]
[<ffffffffa0934622>] kvm_arch_vcpu_ioctl_run+0x10a2/0x14e0 [kvm]
[<ffffffffa091c04d>] kvm_vcpu_ioctl+0x33d/0x620 [kvm]
[<ffffffff81241248>] do_vfs_ioctl+0x298/0x480
[<ffffffff812414a9>] SyS_ioctl+0x79/0x90
[<ffffffff817a04ee>] entry_SYSCALL_64_fastpath+0x12/0x71
---[ end trace b1a0391266848f50 ]---
Testcase (beautified/reduced from syzkaller output):
#include <unistd.h>
#include <sys/syscall.h>
#include <string.h>
#include <stdint.h>
#include <fcntl.h>
#include <sys/ioctl.h>
#include <linux/kvm.h>
long r[31];
int main()
{
memset(r, -1, sizeof(r));
r[2] = open("/dev/kvm", O_RDONLY);
r[3] = ioctl(r[2], KVM_CREATE_VM, 0);
r[7] = ioctl(r[3], KVM_CREATE_VCPU, 0);
struct kvm_vcpu_events ve = {
.exception.injected = 1,
.exception.nr = 0xd4
};
r[27] = ioctl(r[7], KVM_SET_VCPU_EVENTS, &ve);
r[30] = ioctl(r[7], KVM_RUN, 0);
return 0;
}
Reported-by: Dmitry Vyukov <dvyukov@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Radim Krčmář <rkrcmar@redhat.com>
2016-06-01 19:09:20 +07:00
|
|
|
if (events->exception.injected &&
|
|
|
|
(events->exception.nr > 31 || events->exception.nr == NMI_VECTOR))
|
|
|
|
return -EINVAL;
|
|
|
|
|
2011-09-20 17:43:14 +07:00
|
|
|
process_nmi(vcpu);
|
2009-11-12 07:04:25 +07:00
|
|
|
vcpu->arch.exception.pending = events->exception.injected;
|
|
|
|
vcpu->arch.exception.nr = events->exception.nr;
|
|
|
|
vcpu->arch.exception.has_error_code = events->exception.has_error_code;
|
|
|
|
vcpu->arch.exception.error_code = events->exception.error_code;
|
|
|
|
|
|
|
|
vcpu->arch.interrupt.pending = events->interrupt.injected;
|
|
|
|
vcpu->arch.interrupt.nr = events->interrupt.nr;
|
|
|
|
vcpu->arch.interrupt.soft = events->interrupt.soft;
|
2010-02-20 01:38:07 +07:00
|
|
|
if (events->flags & KVM_VCPUEVENT_VALID_SHADOW)
|
|
|
|
kvm_x86_ops->set_interrupt_shadow(vcpu,
|
|
|
|
events->interrupt.shadow);
|
2009-11-12 07:04:25 +07:00
|
|
|
|
|
|
|
vcpu->arch.nmi_injected = events->nmi.injected;
|
2009-12-07 00:24:15 +07:00
|
|
|
if (events->flags & KVM_VCPUEVENT_VALID_NMI_PENDING)
|
|
|
|
vcpu->arch.nmi_pending = events->nmi.pending;
|
2009-11-12 07:04:25 +07:00
|
|
|
kvm_x86_ops->set_nmi_mask(vcpu, events->nmi.masked);
|
|
|
|
|
2013-03-13 18:42:34 +07:00
|
|
|
if (events->flags & KVM_VCPUEVENT_VALID_SIPI_VECTOR &&
|
2016-01-08 19:48:51 +07:00
|
|
|
lapic_in_kernel(vcpu))
|
2013-03-13 18:42:34 +07:00
|
|
|
vcpu->arch.apic->sipi_vector = events->sipi_vector;
|
2009-11-12 07:04:25 +07:00
|
|
|
|
2015-04-01 20:06:40 +07:00
|
|
|
if (events->flags & KVM_VCPUEVENT_VALID_SMM) {
|
|
|
|
if (events->smi.smm)
|
|
|
|
vcpu->arch.hflags |= HF_SMM_MASK;
|
|
|
|
else
|
|
|
|
vcpu->arch.hflags &= ~HF_SMM_MASK;
|
|
|
|
vcpu->arch.smi_pending = events->smi.pending;
|
|
|
|
if (events->smi.smm_inside_nmi)
|
|
|
|
vcpu->arch.hflags |= HF_SMM_INSIDE_NMI_MASK;
|
|
|
|
else
|
|
|
|
vcpu->arch.hflags &= ~HF_SMM_INSIDE_NMI_MASK;
|
2016-01-08 19:48:51 +07:00
|
|
|
if (lapic_in_kernel(vcpu)) {
|
2015-04-01 20:06:40 +07:00
|
|
|
if (events->smi.latched_init)
|
|
|
|
set_bit(KVM_APIC_INIT, &vcpu->arch.apic->pending_events);
|
|
|
|
else
|
|
|
|
clear_bit(KVM_APIC_INIT, &vcpu->arch.apic->pending_events);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2010-07-27 16:30:24 +07:00
|
|
|
kvm_make_request(KVM_REQ_EVENT, vcpu);
|
|
|
|
|
2009-11-12 07:04:25 +07:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2010-02-15 16:45:43 +07:00
|
|
|
static void kvm_vcpu_ioctl_x86_get_debugregs(struct kvm_vcpu *vcpu,
|
|
|
|
struct kvm_debugregs *dbgregs)
|
|
|
|
{
|
2014-01-05 00:47:16 +07:00
|
|
|
unsigned long val;
|
|
|
|
|
2010-02-15 16:45:43 +07:00
|
|
|
memcpy(dbgregs->db, vcpu->arch.db, sizeof(vcpu->arch.db));
|
2014-10-03 05:10:05 +07:00
|
|
|
kvm_get_dr(vcpu, 6, &val);
|
2014-01-05 00:47:16 +07:00
|
|
|
dbgregs->dr6 = val;
|
2010-02-15 16:45:43 +07:00
|
|
|
dbgregs->dr7 = vcpu->arch.dr7;
|
|
|
|
dbgregs->flags = 0;
|
2010-10-31 01:54:47 +07:00
|
|
|
memset(&dbgregs->reserved, 0, sizeof(dbgregs->reserved));
|
2010-02-15 16:45:43 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
static int kvm_vcpu_ioctl_x86_set_debugregs(struct kvm_vcpu *vcpu,
|
|
|
|
struct kvm_debugregs *dbgregs)
|
|
|
|
{
|
|
|
|
if (dbgregs->flags)
|
|
|
|
return -EINVAL;
|
|
|
|
|
KVM: x86: fix OOPS after invalid KVM_SET_DEBUGREGS
MOV to DR6 or DR7 causes a #GP if an attempt is made to write a 1 to
any of bits 63:32. However, this is not detected at KVM_SET_DEBUGREGS
time, and the next KVM_RUN oopses:
general protection fault: 0000 [#1] SMP
CPU: 2 PID: 14987 Comm: a.out Not tainted 4.4.9-300.fc23.x86_64 #1
Hardware name: LENOVO 2325F51/2325F51, BIOS G2ET32WW (1.12 ) 05/30/2012
[...]
Call Trace:
[<ffffffffa072c93d>] kvm_arch_vcpu_ioctl_run+0x141d/0x14e0 [kvm]
[<ffffffffa071405d>] kvm_vcpu_ioctl+0x33d/0x620 [kvm]
[<ffffffff81241648>] do_vfs_ioctl+0x298/0x480
[<ffffffff812418a9>] SyS_ioctl+0x79/0x90
[<ffffffff817a0f2e>] entry_SYSCALL_64_fastpath+0x12/0x71
Code: 55 83 ff 07 48 89 e5 77 27 89 ff ff 24 fd 90 87 80 81 0f 23 fe 5d c3 0f 23 c6 5d c3 0f 23 ce 5d c3 0f 23 d6 5d c3 0f 23 de 5d c3 <0f> 23 f6 5d c3 0f 0b 66 66 66 66 66 2e 0f 1f 84 00 00 00 00 00
RIP [<ffffffff810639eb>] native_set_debugreg+0x2b/0x40
RSP <ffff88005836bd50>
Testcase (beautified/reduced from syzkaller output):
#include <unistd.h>
#include <sys/syscall.h>
#include <string.h>
#include <stdint.h>
#include <linux/kvm.h>
#include <fcntl.h>
#include <sys/ioctl.h>
long r[8];
int main()
{
struct kvm_debugregs dr = { 0 };
r[2] = open("/dev/kvm", O_RDONLY);
r[3] = ioctl(r[2], KVM_CREATE_VM, 0);
r[4] = ioctl(r[3], KVM_CREATE_VCPU, 7);
memcpy(&dr,
"\x5d\x6a\x6b\xe8\x57\x3b\x4b\x7e\xcf\x0d\xa1\x72"
"\xa3\x4a\x29\x0c\xfc\x6d\x44\x00\xa7\x52\xc7\xd8"
"\x00\xdb\x89\x9d\x78\xb5\x54\x6b\x6b\x13\x1c\xe9"
"\x5e\xd3\x0e\x40\x6f\xb4\x66\xf7\x5b\xe3\x36\xcb",
48);
r[7] = ioctl(r[4], KVM_SET_DEBUGREGS, &dr);
r[6] = ioctl(r[4], KVM_RUN, 0);
}
Reported-by: Dmitry Vyukov <dvyukov@google.com>
Cc: stable@vger.kernel.org
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Radim Krčmář <rkrcmar@redhat.com>
2016-06-01 19:09:23 +07:00
|
|
|
if (dbgregs->dr6 & ~0xffffffffull)
|
|
|
|
return -EINVAL;
|
|
|
|
if (dbgregs->dr7 & ~0xffffffffull)
|
|
|
|
return -EINVAL;
|
|
|
|
|
2010-02-15 16:45:43 +07:00
|
|
|
memcpy(vcpu->arch.db, dbgregs->db, sizeof(vcpu->arch.db));
|
2015-04-02 07:10:37 +07:00
|
|
|
kvm_update_dr0123(vcpu);
|
2010-02-15 16:45:43 +07:00
|
|
|
vcpu->arch.dr6 = dbgregs->dr6;
|
2014-01-05 00:47:16 +07:00
|
|
|
kvm_update_dr6(vcpu);
|
2010-02-15 16:45:43 +07:00
|
|
|
vcpu->arch.dr7 = dbgregs->dr7;
|
2014-01-05 00:47:15 +07:00
|
|
|
kvm_update_dr7(vcpu);
|
2010-02-15 16:45:43 +07:00
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2014-11-22 01:05:07 +07:00
|
|
|
#define XSTATE_COMPACTION_ENABLED (1ULL << 63)
|
|
|
|
|
|
|
|
static void fill_xsave(u8 *dest, struct kvm_vcpu *vcpu)
|
|
|
|
{
|
2015-04-30 22:15:32 +07:00
|
|
|
struct xregs_state *xsave = &vcpu->arch.guest_fpu.state.xsave;
|
2015-04-24 15:19:47 +07:00
|
|
|
u64 xstate_bv = xsave->header.xfeatures;
|
2014-11-22 01:05:07 +07:00
|
|
|
u64 valid;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Copy legacy XSAVE area, to avoid complications with CPUID
|
|
|
|
* leaves 0 and 1 in the loop below.
|
|
|
|
*/
|
|
|
|
memcpy(dest, xsave, XSAVE_HDR_OFFSET);
|
|
|
|
|
|
|
|
/* Set XSTATE_BV */
|
|
|
|
*(u64 *)(dest + XSAVE_HDR_OFFSET) = xstate_bv;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Copy each region from the possibly compacted offset to the
|
|
|
|
* non-compacted offset.
|
|
|
|
*/
|
2015-09-03 06:31:26 +07:00
|
|
|
valid = xstate_bv & ~XFEATURE_MASK_FPSSE;
|
2014-11-22 01:05:07 +07:00
|
|
|
while (valid) {
|
|
|
|
u64 feature = valid & -valid;
|
|
|
|
int index = fls64(feature) - 1;
|
|
|
|
void *src = get_xsave_addr(xsave, feature);
|
|
|
|
|
|
|
|
if (src) {
|
|
|
|
u32 size, offset, ecx, edx;
|
|
|
|
cpuid_count(XSTATE_CPUID, index,
|
|
|
|
&size, &offset, &ecx, &edx);
|
|
|
|
memcpy(dest + offset, src, size);
|
|
|
|
}
|
|
|
|
|
|
|
|
valid -= feature;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static void load_xsave(struct kvm_vcpu *vcpu, u8 *src)
|
|
|
|
{
|
2015-04-30 22:15:32 +07:00
|
|
|
struct xregs_state *xsave = &vcpu->arch.guest_fpu.state.xsave;
|
2014-11-22 01:05:07 +07:00
|
|
|
u64 xstate_bv = *(u64 *)(src + XSAVE_HDR_OFFSET);
|
|
|
|
u64 valid;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Copy legacy XSAVE area, to avoid complications with CPUID
|
|
|
|
* leaves 0 and 1 in the loop below.
|
|
|
|
*/
|
|
|
|
memcpy(xsave, src, XSAVE_HDR_OFFSET);
|
|
|
|
|
|
|
|
/* Set XSTATE_BV and possibly XCOMP_BV. */
|
2015-04-24 15:19:47 +07:00
|
|
|
xsave->header.xfeatures = xstate_bv;
|
2016-04-05 03:25:03 +07:00
|
|
|
if (boot_cpu_has(X86_FEATURE_XSAVES))
|
2015-04-24 15:14:36 +07:00
|
|
|
xsave->header.xcomp_bv = host_xcr0 | XSTATE_COMPACTION_ENABLED;
|
2014-11-22 01:05:07 +07:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Copy each region from the non-compacted offset to the
|
|
|
|
* possibly compacted offset.
|
|
|
|
*/
|
2015-09-03 06:31:26 +07:00
|
|
|
valid = xstate_bv & ~XFEATURE_MASK_FPSSE;
|
2014-11-22 01:05:07 +07:00
|
|
|
while (valid) {
|
|
|
|
u64 feature = valid & -valid;
|
|
|
|
int index = fls64(feature) - 1;
|
|
|
|
void *dest = get_xsave_addr(xsave, feature);
|
|
|
|
|
|
|
|
if (dest) {
|
|
|
|
u32 size, offset, ecx, edx;
|
|
|
|
cpuid_count(XSTATE_CPUID, index,
|
|
|
|
&size, &offset, &ecx, &edx);
|
|
|
|
memcpy(dest, src + offset, size);
|
2015-07-09 14:44:52 +07:00
|
|
|
}
|
2014-11-22 01:05:07 +07:00
|
|
|
|
|
|
|
valid -= feature;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2010-06-13 16:29:39 +07:00
|
|
|
static void kvm_vcpu_ioctl_x86_get_xsave(struct kvm_vcpu *vcpu,
|
|
|
|
struct kvm_xsave *guest_xsave)
|
|
|
|
{
|
2016-04-05 03:25:02 +07:00
|
|
|
if (boot_cpu_has(X86_FEATURE_XSAVE)) {
|
2014-11-22 01:05:07 +07:00
|
|
|
memset(guest_xsave, 0, sizeof(struct kvm_xsave));
|
|
|
|
fill_xsave((u8 *) guest_xsave->region, vcpu);
|
2013-10-02 21:06:16 +07:00
|
|
|
} else {
|
2010-06-13 16:29:39 +07:00
|
|
|
memcpy(guest_xsave->region,
|
x86/fpu: Simplify FPU handling by embedding the fpstate in task_struct (again)
So 6 years ago we made the FPU fpstate dynamically allocated:
aa283f49276e ("x86, fpu: lazy allocation of FPU area - v5")
61c4628b5386 ("x86, fpu: split FPU state from task struct - v5")
In hindsight this was a mistake:
- it complicated context allocation failure handling, such as:
/* kthread execs. TODO: cleanup this horror. */
if (WARN_ON(fpstate_alloc_init(fpu)))
force_sig(SIGKILL, tsk);
- it caused us to enable irqs in fpu__restore():
local_irq_enable();
/*
* does a slab alloc which can sleep
*/
if (fpstate_alloc_init(fpu)) {
/*
* ran out of memory!
*/
do_group_exit(SIGKILL);
return;
}
local_irq_disable();
- it (slightly) slowed down task creation/destruction by adding
slab allocation/free pattens.
- it made access to context contents (slightly) slower by adding
one more pointer dereference.
The motivation for the dynamic allocation was two-fold:
- reduce memory consumption by non-FPU tasks
- allocate and handle only the necessary amount of context for
various XSAVE processors that have varying hardware frame
sizes.
These days, with glibc using SSE memcpy by default and GCC optimizing
for SSE/AVX by default, the scope of FPU using apps on an x86 system is
much larger than it was 6 years ago.
For example on a freshly installed Fedora 21 desktop system, with a
recent kernel, all non-kthread tasks have used the FPU shortly after
bootup.
Also, even modern embedded x86 CPUs try to support the latest vector
instruction set - so they'll too often use the larger xstate frame
sizes.
So remove the dynamic allocation complication by embedding the FPU
fpstate in task_struct again. This should make the FPU a lot more
accessible to all sorts of atomic contexts.
We could still optimize for the xstate frame size in the future,
by moving the state structure to the last element of task_struct,
and allocating only a part of that.
This change is kept minimal by still keeping the ctx_alloc()/free()
routines (that now do nothing substantial) - we'll remove them in
the following patches.
Reviewed-by: Borislav Petkov <bp@alien8.de>
Cc: Andy Lutomirski <luto@amacapital.net>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
2015-04-27 09:19:39 +07:00
|
|
|
&vcpu->arch.guest_fpu.state.fxsave,
|
2015-04-30 22:15:32 +07:00
|
|
|
sizeof(struct fxregs_state));
|
2010-06-13 16:29:39 +07:00
|
|
|
*(u64 *)&guest_xsave->region[XSAVE_HDR_OFFSET / sizeof(u32)] =
|
2015-09-03 06:31:26 +07:00
|
|
|
XFEATURE_MASK_FPSSE;
|
2010-06-13 16:29:39 +07:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static int kvm_vcpu_ioctl_x86_set_xsave(struct kvm_vcpu *vcpu,
|
|
|
|
struct kvm_xsave *guest_xsave)
|
|
|
|
{
|
|
|
|
u64 xstate_bv =
|
|
|
|
*(u64 *)&guest_xsave->region[XSAVE_HDR_OFFSET / sizeof(u32)];
|
|
|
|
|
2016-04-05 03:25:02 +07:00
|
|
|
if (boot_cpu_has(X86_FEATURE_XSAVE)) {
|
2013-10-02 21:06:15 +07:00
|
|
|
/*
|
|
|
|
* Here we allow setting states that are not present in
|
|
|
|
* CPUID leaf 0xD, index 0, EDX:EAX. This is for compatibility
|
|
|
|
* with old userspace.
|
|
|
|
*/
|
2014-02-24 18:15:16 +07:00
|
|
|
if (xstate_bv & ~kvm_supported_xcr0())
|
2013-10-02 21:06:15 +07:00
|
|
|
return -EINVAL;
|
2014-11-22 01:05:07 +07:00
|
|
|
load_xsave(vcpu, (u8 *)guest_xsave->region);
|
2013-10-02 21:06:15 +07:00
|
|
|
} else {
|
2015-09-03 06:31:26 +07:00
|
|
|
if (xstate_bv & ~XFEATURE_MASK_FPSSE)
|
2010-06-13 16:29:39 +07:00
|
|
|
return -EINVAL;
|
x86/fpu: Simplify FPU handling by embedding the fpstate in task_struct (again)
So 6 years ago we made the FPU fpstate dynamically allocated:
aa283f49276e ("x86, fpu: lazy allocation of FPU area - v5")
61c4628b5386 ("x86, fpu: split FPU state from task struct - v5")
In hindsight this was a mistake:
- it complicated context allocation failure handling, such as:
/* kthread execs. TODO: cleanup this horror. */
if (WARN_ON(fpstate_alloc_init(fpu)))
force_sig(SIGKILL, tsk);
- it caused us to enable irqs in fpu__restore():
local_irq_enable();
/*
* does a slab alloc which can sleep
*/
if (fpstate_alloc_init(fpu)) {
/*
* ran out of memory!
*/
do_group_exit(SIGKILL);
return;
}
local_irq_disable();
- it (slightly) slowed down task creation/destruction by adding
slab allocation/free pattens.
- it made access to context contents (slightly) slower by adding
one more pointer dereference.
The motivation for the dynamic allocation was two-fold:
- reduce memory consumption by non-FPU tasks
- allocate and handle only the necessary amount of context for
various XSAVE processors that have varying hardware frame
sizes.
These days, with glibc using SSE memcpy by default and GCC optimizing
for SSE/AVX by default, the scope of FPU using apps on an x86 system is
much larger than it was 6 years ago.
For example on a freshly installed Fedora 21 desktop system, with a
recent kernel, all non-kthread tasks have used the FPU shortly after
bootup.
Also, even modern embedded x86 CPUs try to support the latest vector
instruction set - so they'll too often use the larger xstate frame
sizes.
So remove the dynamic allocation complication by embedding the FPU
fpstate in task_struct again. This should make the FPU a lot more
accessible to all sorts of atomic contexts.
We could still optimize for the xstate frame size in the future,
by moving the state structure to the last element of task_struct,
and allocating only a part of that.
This change is kept minimal by still keeping the ctx_alloc()/free()
routines (that now do nothing substantial) - we'll remove them in
the following patches.
Reviewed-by: Borislav Petkov <bp@alien8.de>
Cc: Andy Lutomirski <luto@amacapital.net>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
2015-04-27 09:19:39 +07:00
|
|
|
memcpy(&vcpu->arch.guest_fpu.state.fxsave,
|
2015-04-30 22:15:32 +07:00
|
|
|
guest_xsave->region, sizeof(struct fxregs_state));
|
2010-06-13 16:29:39 +07:00
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void kvm_vcpu_ioctl_x86_get_xcrs(struct kvm_vcpu *vcpu,
|
|
|
|
struct kvm_xcrs *guest_xcrs)
|
|
|
|
{
|
2016-04-05 03:25:02 +07:00
|
|
|
if (!boot_cpu_has(X86_FEATURE_XSAVE)) {
|
2010-06-13 16:29:39 +07:00
|
|
|
guest_xcrs->nr_xcrs = 0;
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
guest_xcrs->nr_xcrs = 1;
|
|
|
|
guest_xcrs->flags = 0;
|
|
|
|
guest_xcrs->xcrs[0].xcr = XCR_XFEATURE_ENABLED_MASK;
|
|
|
|
guest_xcrs->xcrs[0].value = vcpu->arch.xcr0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int kvm_vcpu_ioctl_x86_set_xcrs(struct kvm_vcpu *vcpu,
|
|
|
|
struct kvm_xcrs *guest_xcrs)
|
|
|
|
{
|
|
|
|
int i, r = 0;
|
|
|
|
|
2016-04-05 03:25:02 +07:00
|
|
|
if (!boot_cpu_has(X86_FEATURE_XSAVE))
|
2010-06-13 16:29:39 +07:00
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
if (guest_xcrs->nr_xcrs > KVM_MAX_XCRS || guest_xcrs->flags)
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
for (i = 0; i < guest_xcrs->nr_xcrs; i++)
|
|
|
|
/* Only support XCR0 currently */
|
2013-10-17 21:50:47 +07:00
|
|
|
if (guest_xcrs->xcrs[i].xcr == XCR_XFEATURE_ENABLED_MASK) {
|
2010-06-13 16:29:39 +07:00
|
|
|
r = __kvm_set_xcr(vcpu, XCR_XFEATURE_ENABLED_MASK,
|
2013-10-17 21:50:47 +07:00
|
|
|
guest_xcrs->xcrs[i].value);
|
2010-06-13 16:29:39 +07:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
if (r)
|
|
|
|
r = -EINVAL;
|
|
|
|
return r;
|
|
|
|
}
|
|
|
|
|
2012-03-11 02:37:27 +07:00
|
|
|
/*
|
|
|
|
* kvm_set_guest_paused() indicates to the guest kernel that it has been
|
|
|
|
* stopped by the hypervisor. This function will be called from the host only.
|
|
|
|
* EINVAL is returned when the host attempts to set the flag for a guest that
|
|
|
|
* does not support pv clocks.
|
|
|
|
*/
|
|
|
|
static int kvm_set_guest_paused(struct kvm_vcpu *vcpu)
|
|
|
|
{
|
2013-02-21 05:48:10 +07:00
|
|
|
if (!vcpu->arch.pv_time_enabled)
|
2012-03-11 02:37:27 +07:00
|
|
|
return -EINVAL;
|
2012-08-04 01:57:49 +07:00
|
|
|
vcpu->arch.pvclock_set_guest_stopped_request = true;
|
2012-03-11 02:37:27 +07:00
|
|
|
kvm_make_request(KVM_REQ_CLOCK_UPDATE, vcpu);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2015-11-10 19:36:34 +07:00
|
|
|
static int kvm_vcpu_ioctl_enable_cap(struct kvm_vcpu *vcpu,
|
|
|
|
struct kvm_enable_cap *cap)
|
|
|
|
{
|
|
|
|
if (cap->flags)
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
switch (cap->cap) {
|
|
|
|
case KVM_CAP_HYPERV_SYNIC:
|
|
|
|
return kvm_hv_activate_synic(vcpu);
|
|
|
|
default:
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
KVM: Portability: split kvm_vcpu_ioctl
This patch splits kvm_vcpu_ioctl into archtecture independent parts, and
x86 specific parts which go to kvm_arch_vcpu_ioctl in x86.c.
Common ioctls for all architectures are:
KVM_RUN, KVM_GET/SET_(S-)REGS, KVM_TRANSLATE, KVM_INTERRUPT,
KVM_DEBUG_GUEST, KVM_SET_SIGNAL_MASK, KVM_GET/SET_FPU
Note that some PPC chips don't have an FPU, so we might need an #ifdef
around KVM_GET/SET_FPU one day.
x86 specific ioctls are:
KVM_GET/SET_LAPIC, KVM_SET_CPUID, KVM_GET/SET_MSRS
An interresting aspect is vcpu_load/vcpu_put. We now have a common
vcpu_load/put which does the preemption stuff, and an architecture
specific kvm_arch_vcpu_load/put. In the x86 case, this one calls the
vmx/svm function defined in kvm_x86_ops.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Reviewed-by: Christian Ehrhardt <ehrhardt@linux.vnet.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-12 00:16:52 +07:00
|
|
|
long kvm_arch_vcpu_ioctl(struct file *filp,
|
|
|
|
unsigned int ioctl, unsigned long arg)
|
|
|
|
{
|
|
|
|
struct kvm_vcpu *vcpu = filp->private_data;
|
|
|
|
void __user *argp = (void __user *)arg;
|
|
|
|
int r;
|
2010-06-20 19:54:43 +07:00
|
|
|
union {
|
|
|
|
struct kvm_lapic_state *lapic;
|
|
|
|
struct kvm_xsave *xsave;
|
|
|
|
struct kvm_xcrs *xcrs;
|
|
|
|
void *buffer;
|
|
|
|
} u;
|
|
|
|
|
|
|
|
u.buffer = NULL;
|
KVM: Portability: split kvm_vcpu_ioctl
This patch splits kvm_vcpu_ioctl into archtecture independent parts, and
x86 specific parts which go to kvm_arch_vcpu_ioctl in x86.c.
Common ioctls for all architectures are:
KVM_RUN, KVM_GET/SET_(S-)REGS, KVM_TRANSLATE, KVM_INTERRUPT,
KVM_DEBUG_GUEST, KVM_SET_SIGNAL_MASK, KVM_GET/SET_FPU
Note that some PPC chips don't have an FPU, so we might need an #ifdef
around KVM_GET/SET_FPU one day.
x86 specific ioctls are:
KVM_GET/SET_LAPIC, KVM_SET_CPUID, KVM_GET/SET_MSRS
An interresting aspect is vcpu_load/vcpu_put. We now have a common
vcpu_load/put which does the preemption stuff, and an architecture
specific kvm_arch_vcpu_load/put. In the x86 case, this one calls the
vmx/svm function defined in kvm_x86_ops.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Reviewed-by: Christian Ehrhardt <ehrhardt@linux.vnet.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-12 00:16:52 +07:00
|
|
|
switch (ioctl) {
|
|
|
|
case KVM_GET_LAPIC: {
|
2009-10-29 22:44:16 +07:00
|
|
|
r = -EINVAL;
|
2016-01-08 19:48:51 +07:00
|
|
|
if (!lapic_in_kernel(vcpu))
|
2009-10-29 22:44:16 +07:00
|
|
|
goto out;
|
2010-06-20 19:54:43 +07:00
|
|
|
u.lapic = kzalloc(sizeof(struct kvm_lapic_state), GFP_KERNEL);
|
KVM: Portability: split kvm_vcpu_ioctl
This patch splits kvm_vcpu_ioctl into archtecture independent parts, and
x86 specific parts which go to kvm_arch_vcpu_ioctl in x86.c.
Common ioctls for all architectures are:
KVM_RUN, KVM_GET/SET_(S-)REGS, KVM_TRANSLATE, KVM_INTERRUPT,
KVM_DEBUG_GUEST, KVM_SET_SIGNAL_MASK, KVM_GET/SET_FPU
Note that some PPC chips don't have an FPU, so we might need an #ifdef
around KVM_GET/SET_FPU one day.
x86 specific ioctls are:
KVM_GET/SET_LAPIC, KVM_SET_CPUID, KVM_GET/SET_MSRS
An interresting aspect is vcpu_load/vcpu_put. We now have a common
vcpu_load/put which does the preemption stuff, and an architecture
specific kvm_arch_vcpu_load/put. In the x86 case, this one calls the
vmx/svm function defined in kvm_x86_ops.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Reviewed-by: Christian Ehrhardt <ehrhardt@linux.vnet.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-12 00:16:52 +07:00
|
|
|
|
2008-08-12 00:01:47 +07:00
|
|
|
r = -ENOMEM;
|
2010-06-20 19:54:43 +07:00
|
|
|
if (!u.lapic)
|
2008-08-12 00:01:47 +07:00
|
|
|
goto out;
|
2010-06-20 19:54:43 +07:00
|
|
|
r = kvm_vcpu_ioctl_get_lapic(vcpu, u.lapic);
|
KVM: Portability: split kvm_vcpu_ioctl
This patch splits kvm_vcpu_ioctl into archtecture independent parts, and
x86 specific parts which go to kvm_arch_vcpu_ioctl in x86.c.
Common ioctls for all architectures are:
KVM_RUN, KVM_GET/SET_(S-)REGS, KVM_TRANSLATE, KVM_INTERRUPT,
KVM_DEBUG_GUEST, KVM_SET_SIGNAL_MASK, KVM_GET/SET_FPU
Note that some PPC chips don't have an FPU, so we might need an #ifdef
around KVM_GET/SET_FPU one day.
x86 specific ioctls are:
KVM_GET/SET_LAPIC, KVM_SET_CPUID, KVM_GET/SET_MSRS
An interresting aspect is vcpu_load/vcpu_put. We now have a common
vcpu_load/put which does the preemption stuff, and an architecture
specific kvm_arch_vcpu_load/put. In the x86 case, this one calls the
vmx/svm function defined in kvm_x86_ops.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Reviewed-by: Christian Ehrhardt <ehrhardt@linux.vnet.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-12 00:16:52 +07:00
|
|
|
if (r)
|
|
|
|
goto out;
|
|
|
|
r = -EFAULT;
|
2010-06-20 19:54:43 +07:00
|
|
|
if (copy_to_user(argp, u.lapic, sizeof(struct kvm_lapic_state)))
|
KVM: Portability: split kvm_vcpu_ioctl
This patch splits kvm_vcpu_ioctl into archtecture independent parts, and
x86 specific parts which go to kvm_arch_vcpu_ioctl in x86.c.
Common ioctls for all architectures are:
KVM_RUN, KVM_GET/SET_(S-)REGS, KVM_TRANSLATE, KVM_INTERRUPT,
KVM_DEBUG_GUEST, KVM_SET_SIGNAL_MASK, KVM_GET/SET_FPU
Note that some PPC chips don't have an FPU, so we might need an #ifdef
around KVM_GET/SET_FPU one day.
x86 specific ioctls are:
KVM_GET/SET_LAPIC, KVM_SET_CPUID, KVM_GET/SET_MSRS
An interresting aspect is vcpu_load/vcpu_put. We now have a common
vcpu_load/put which does the preemption stuff, and an architecture
specific kvm_arch_vcpu_load/put. In the x86 case, this one calls the
vmx/svm function defined in kvm_x86_ops.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Reviewed-by: Christian Ehrhardt <ehrhardt@linux.vnet.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-12 00:16:52 +07:00
|
|
|
goto out;
|
|
|
|
r = 0;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
case KVM_SET_LAPIC: {
|
2009-10-29 22:44:16 +07:00
|
|
|
r = -EINVAL;
|
2016-01-08 19:48:51 +07:00
|
|
|
if (!lapic_in_kernel(vcpu))
|
2009-10-29 22:44:16 +07:00
|
|
|
goto out;
|
2011-12-05 00:36:29 +07:00
|
|
|
u.lapic = memdup_user(argp, sizeof(*u.lapic));
|
2012-11-02 17:33:21 +07:00
|
|
|
if (IS_ERR(u.lapic))
|
|
|
|
return PTR_ERR(u.lapic);
|
2011-12-05 00:36:29 +07:00
|
|
|
|
2010-06-20 19:54:43 +07:00
|
|
|
r = kvm_vcpu_ioctl_set_lapic(vcpu, u.lapic);
|
KVM: Portability: split kvm_vcpu_ioctl
This patch splits kvm_vcpu_ioctl into archtecture independent parts, and
x86 specific parts which go to kvm_arch_vcpu_ioctl in x86.c.
Common ioctls for all architectures are:
KVM_RUN, KVM_GET/SET_(S-)REGS, KVM_TRANSLATE, KVM_INTERRUPT,
KVM_DEBUG_GUEST, KVM_SET_SIGNAL_MASK, KVM_GET/SET_FPU
Note that some PPC chips don't have an FPU, so we might need an #ifdef
around KVM_GET/SET_FPU one day.
x86 specific ioctls are:
KVM_GET/SET_LAPIC, KVM_SET_CPUID, KVM_GET/SET_MSRS
An interresting aspect is vcpu_load/vcpu_put. We now have a common
vcpu_load/put which does the preemption stuff, and an architecture
specific kvm_arch_vcpu_load/put. In the x86 case, this one calls the
vmx/svm function defined in kvm_x86_ops.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Reviewed-by: Christian Ehrhardt <ehrhardt@linux.vnet.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-12 00:16:52 +07:00
|
|
|
break;
|
|
|
|
}
|
2007-11-21 03:36:41 +07:00
|
|
|
case KVM_INTERRUPT: {
|
|
|
|
struct kvm_interrupt irq;
|
|
|
|
|
|
|
|
r = -EFAULT;
|
|
|
|
if (copy_from_user(&irq, argp, sizeof irq))
|
|
|
|
goto out;
|
|
|
|
r = kvm_vcpu_ioctl_interrupt(vcpu, &irq);
|
|
|
|
break;
|
|
|
|
}
|
2008-09-26 14:30:55 +07:00
|
|
|
case KVM_NMI: {
|
|
|
|
r = kvm_vcpu_ioctl_nmi(vcpu);
|
|
|
|
break;
|
|
|
|
}
|
2015-04-01 20:06:40 +07:00
|
|
|
case KVM_SMI: {
|
|
|
|
r = kvm_vcpu_ioctl_smi(vcpu);
|
|
|
|
break;
|
|
|
|
}
|
KVM: Portability: split kvm_vcpu_ioctl
This patch splits kvm_vcpu_ioctl into archtecture independent parts, and
x86 specific parts which go to kvm_arch_vcpu_ioctl in x86.c.
Common ioctls for all architectures are:
KVM_RUN, KVM_GET/SET_(S-)REGS, KVM_TRANSLATE, KVM_INTERRUPT,
KVM_DEBUG_GUEST, KVM_SET_SIGNAL_MASK, KVM_GET/SET_FPU
Note that some PPC chips don't have an FPU, so we might need an #ifdef
around KVM_GET/SET_FPU one day.
x86 specific ioctls are:
KVM_GET/SET_LAPIC, KVM_SET_CPUID, KVM_GET/SET_MSRS
An interresting aspect is vcpu_load/vcpu_put. We now have a common
vcpu_load/put which does the preemption stuff, and an architecture
specific kvm_arch_vcpu_load/put. In the x86 case, this one calls the
vmx/svm function defined in kvm_x86_ops.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Reviewed-by: Christian Ehrhardt <ehrhardt@linux.vnet.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-12 00:16:52 +07:00
|
|
|
case KVM_SET_CPUID: {
|
|
|
|
struct kvm_cpuid __user *cpuid_arg = argp;
|
|
|
|
struct kvm_cpuid cpuid;
|
|
|
|
|
|
|
|
r = -EFAULT;
|
|
|
|
if (copy_from_user(&cpuid, cpuid_arg, sizeof cpuid))
|
|
|
|
goto out;
|
|
|
|
r = kvm_vcpu_ioctl_set_cpuid(vcpu, &cpuid, cpuid_arg->entries);
|
|
|
|
break;
|
|
|
|
}
|
2007-11-21 22:10:04 +07:00
|
|
|
case KVM_SET_CPUID2: {
|
|
|
|
struct kvm_cpuid2 __user *cpuid_arg = argp;
|
|
|
|
struct kvm_cpuid2 cpuid;
|
|
|
|
|
|
|
|
r = -EFAULT;
|
|
|
|
if (copy_from_user(&cpuid, cpuid_arg, sizeof cpuid))
|
|
|
|
goto out;
|
|
|
|
r = kvm_vcpu_ioctl_set_cpuid2(vcpu, &cpuid,
|
2009-01-14 23:56:00 +07:00
|
|
|
cpuid_arg->entries);
|
2007-11-21 22:10:04 +07:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
case KVM_GET_CPUID2: {
|
|
|
|
struct kvm_cpuid2 __user *cpuid_arg = argp;
|
|
|
|
struct kvm_cpuid2 cpuid;
|
|
|
|
|
|
|
|
r = -EFAULT;
|
|
|
|
if (copy_from_user(&cpuid, cpuid_arg, sizeof cpuid))
|
|
|
|
goto out;
|
|
|
|
r = kvm_vcpu_ioctl_get_cpuid2(vcpu, &cpuid,
|
2009-01-14 23:56:00 +07:00
|
|
|
cpuid_arg->entries);
|
2007-11-21 22:10:04 +07:00
|
|
|
if (r)
|
|
|
|
goto out;
|
|
|
|
r = -EFAULT;
|
|
|
|
if (copy_to_user(cpuid_arg, &cpuid, sizeof cpuid))
|
|
|
|
goto out;
|
|
|
|
r = 0;
|
|
|
|
break;
|
|
|
|
}
|
KVM: Portability: split kvm_vcpu_ioctl
This patch splits kvm_vcpu_ioctl into archtecture independent parts, and
x86 specific parts which go to kvm_arch_vcpu_ioctl in x86.c.
Common ioctls for all architectures are:
KVM_RUN, KVM_GET/SET_(S-)REGS, KVM_TRANSLATE, KVM_INTERRUPT,
KVM_DEBUG_GUEST, KVM_SET_SIGNAL_MASK, KVM_GET/SET_FPU
Note that some PPC chips don't have an FPU, so we might need an #ifdef
around KVM_GET/SET_FPU one day.
x86 specific ioctls are:
KVM_GET/SET_LAPIC, KVM_SET_CPUID, KVM_GET/SET_MSRS
An interresting aspect is vcpu_load/vcpu_put. We now have a common
vcpu_load/put which does the preemption stuff, and an architecture
specific kvm_arch_vcpu_load/put. In the x86 case, this one calls the
vmx/svm function defined in kvm_x86_ops.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Reviewed-by: Christian Ehrhardt <ehrhardt@linux.vnet.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-12 00:16:52 +07:00
|
|
|
case KVM_GET_MSRS:
|
2015-04-08 20:30:38 +07:00
|
|
|
r = msr_io(vcpu, argp, do_get_msr, 1);
|
KVM: Portability: split kvm_vcpu_ioctl
This patch splits kvm_vcpu_ioctl into archtecture independent parts, and
x86 specific parts which go to kvm_arch_vcpu_ioctl in x86.c.
Common ioctls for all architectures are:
KVM_RUN, KVM_GET/SET_(S-)REGS, KVM_TRANSLATE, KVM_INTERRUPT,
KVM_DEBUG_GUEST, KVM_SET_SIGNAL_MASK, KVM_GET/SET_FPU
Note that some PPC chips don't have an FPU, so we might need an #ifdef
around KVM_GET/SET_FPU one day.
x86 specific ioctls are:
KVM_GET/SET_LAPIC, KVM_SET_CPUID, KVM_GET/SET_MSRS
An interresting aspect is vcpu_load/vcpu_put. We now have a common
vcpu_load/put which does the preemption stuff, and an architecture
specific kvm_arch_vcpu_load/put. In the x86 case, this one calls the
vmx/svm function defined in kvm_x86_ops.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Reviewed-by: Christian Ehrhardt <ehrhardt@linux.vnet.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-12 00:16:52 +07:00
|
|
|
break;
|
|
|
|
case KVM_SET_MSRS:
|
|
|
|
r = msr_io(vcpu, argp, do_set_msr, 0);
|
|
|
|
break;
|
2007-10-22 21:50:39 +07:00
|
|
|
case KVM_TPR_ACCESS_REPORTING: {
|
|
|
|
struct kvm_tpr_access_ctl tac;
|
|
|
|
|
|
|
|
r = -EFAULT;
|
|
|
|
if (copy_from_user(&tac, argp, sizeof tac))
|
|
|
|
goto out;
|
|
|
|
r = vcpu_ioctl_tpr_access_reporting(vcpu, &tac);
|
|
|
|
if (r)
|
|
|
|
goto out;
|
|
|
|
r = -EFAULT;
|
|
|
|
if (copy_to_user(argp, &tac, sizeof tac))
|
|
|
|
goto out;
|
|
|
|
r = 0;
|
|
|
|
break;
|
|
|
|
};
|
2007-10-25 21:52:32 +07:00
|
|
|
case KVM_SET_VAPIC_ADDR: {
|
|
|
|
struct kvm_vapic_addr va;
|
|
|
|
|
|
|
|
r = -EINVAL;
|
2015-07-29 17:05:37 +07:00
|
|
|
if (!lapic_in_kernel(vcpu))
|
2007-10-25 21:52:32 +07:00
|
|
|
goto out;
|
|
|
|
r = -EFAULT;
|
|
|
|
if (copy_from_user(&va, argp, sizeof va))
|
|
|
|
goto out;
|
2013-11-21 01:23:22 +07:00
|
|
|
r = kvm_lapic_set_vapic_addr(vcpu, va.vapic_addr);
|
2007-10-25 21:52:32 +07:00
|
|
|
break;
|
|
|
|
}
|
2009-05-11 15:48:15 +07:00
|
|
|
case KVM_X86_SETUP_MCE: {
|
|
|
|
u64 mcg_cap;
|
|
|
|
|
|
|
|
r = -EFAULT;
|
|
|
|
if (copy_from_user(&mcg_cap, argp, sizeof mcg_cap))
|
|
|
|
goto out;
|
|
|
|
r = kvm_vcpu_ioctl_x86_setup_mce(vcpu, mcg_cap);
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
case KVM_X86_SET_MCE: {
|
|
|
|
struct kvm_x86_mce mce;
|
|
|
|
|
|
|
|
r = -EFAULT;
|
|
|
|
if (copy_from_user(&mce, argp, sizeof mce))
|
|
|
|
goto out;
|
|
|
|
r = kvm_vcpu_ioctl_x86_set_mce(vcpu, &mce);
|
|
|
|
break;
|
|
|
|
}
|
2009-11-12 07:04:25 +07:00
|
|
|
case KVM_GET_VCPU_EVENTS: {
|
|
|
|
struct kvm_vcpu_events events;
|
|
|
|
|
|
|
|
kvm_vcpu_ioctl_x86_get_vcpu_events(vcpu, &events);
|
|
|
|
|
|
|
|
r = -EFAULT;
|
|
|
|
if (copy_to_user(argp, &events, sizeof(struct kvm_vcpu_events)))
|
|
|
|
break;
|
|
|
|
r = 0;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
case KVM_SET_VCPU_EVENTS: {
|
|
|
|
struct kvm_vcpu_events events;
|
|
|
|
|
|
|
|
r = -EFAULT;
|
|
|
|
if (copy_from_user(&events, argp, sizeof(struct kvm_vcpu_events)))
|
|
|
|
break;
|
|
|
|
|
|
|
|
r = kvm_vcpu_ioctl_x86_set_vcpu_events(vcpu, &events);
|
|
|
|
break;
|
|
|
|
}
|
2010-02-15 16:45:43 +07:00
|
|
|
case KVM_GET_DEBUGREGS: {
|
|
|
|
struct kvm_debugregs dbgregs;
|
|
|
|
|
|
|
|
kvm_vcpu_ioctl_x86_get_debugregs(vcpu, &dbgregs);
|
|
|
|
|
|
|
|
r = -EFAULT;
|
|
|
|
if (copy_to_user(argp, &dbgregs,
|
|
|
|
sizeof(struct kvm_debugregs)))
|
|
|
|
break;
|
|
|
|
r = 0;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
case KVM_SET_DEBUGREGS: {
|
|
|
|
struct kvm_debugregs dbgregs;
|
|
|
|
|
|
|
|
r = -EFAULT;
|
|
|
|
if (copy_from_user(&dbgregs, argp,
|
|
|
|
sizeof(struct kvm_debugregs)))
|
|
|
|
break;
|
|
|
|
|
|
|
|
r = kvm_vcpu_ioctl_x86_set_debugregs(vcpu, &dbgregs);
|
|
|
|
break;
|
|
|
|
}
|
2010-06-13 16:29:39 +07:00
|
|
|
case KVM_GET_XSAVE: {
|
2010-06-20 19:54:43 +07:00
|
|
|
u.xsave = kzalloc(sizeof(struct kvm_xsave), GFP_KERNEL);
|
2010-06-13 16:29:39 +07:00
|
|
|
r = -ENOMEM;
|
2010-06-20 19:54:43 +07:00
|
|
|
if (!u.xsave)
|
2010-06-13 16:29:39 +07:00
|
|
|
break;
|
|
|
|
|
2010-06-20 19:54:43 +07:00
|
|
|
kvm_vcpu_ioctl_x86_get_xsave(vcpu, u.xsave);
|
2010-06-13 16:29:39 +07:00
|
|
|
|
|
|
|
r = -EFAULT;
|
2010-06-20 19:54:43 +07:00
|
|
|
if (copy_to_user(argp, u.xsave, sizeof(struct kvm_xsave)))
|
2010-06-13 16:29:39 +07:00
|
|
|
break;
|
|
|
|
r = 0;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
case KVM_SET_XSAVE: {
|
2011-12-05 00:36:29 +07:00
|
|
|
u.xsave = memdup_user(argp, sizeof(*u.xsave));
|
2012-11-02 17:33:21 +07:00
|
|
|
if (IS_ERR(u.xsave))
|
|
|
|
return PTR_ERR(u.xsave);
|
2010-06-13 16:29:39 +07:00
|
|
|
|
2010-06-20 19:54:43 +07:00
|
|
|
r = kvm_vcpu_ioctl_x86_set_xsave(vcpu, u.xsave);
|
2010-06-13 16:29:39 +07:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
case KVM_GET_XCRS: {
|
2010-06-20 19:54:43 +07:00
|
|
|
u.xcrs = kzalloc(sizeof(struct kvm_xcrs), GFP_KERNEL);
|
2010-06-13 16:29:39 +07:00
|
|
|
r = -ENOMEM;
|
2010-06-20 19:54:43 +07:00
|
|
|
if (!u.xcrs)
|
2010-06-13 16:29:39 +07:00
|
|
|
break;
|
|
|
|
|
2010-06-20 19:54:43 +07:00
|
|
|
kvm_vcpu_ioctl_x86_get_xcrs(vcpu, u.xcrs);
|
2010-06-13 16:29:39 +07:00
|
|
|
|
|
|
|
r = -EFAULT;
|
2010-06-20 19:54:43 +07:00
|
|
|
if (copy_to_user(argp, u.xcrs,
|
2010-06-13 16:29:39 +07:00
|
|
|
sizeof(struct kvm_xcrs)))
|
|
|
|
break;
|
|
|
|
r = 0;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
case KVM_SET_XCRS: {
|
2011-12-05 00:36:29 +07:00
|
|
|
u.xcrs = memdup_user(argp, sizeof(*u.xcrs));
|
2012-11-02 17:33:21 +07:00
|
|
|
if (IS_ERR(u.xcrs))
|
|
|
|
return PTR_ERR(u.xcrs);
|
2010-06-13 16:29:39 +07:00
|
|
|
|
2010-06-20 19:54:43 +07:00
|
|
|
r = kvm_vcpu_ioctl_x86_set_xcrs(vcpu, u.xcrs);
|
2010-06-13 16:29:39 +07:00
|
|
|
break;
|
|
|
|
}
|
2011-03-25 15:44:51 +07:00
|
|
|
case KVM_SET_TSC_KHZ: {
|
|
|
|
u32 user_tsc_khz;
|
|
|
|
|
|
|
|
r = -EINVAL;
|
|
|
|
user_tsc_khz = (u32)arg;
|
|
|
|
|
|
|
|
if (user_tsc_khz >= kvm_max_guest_tsc_khz)
|
|
|
|
goto out;
|
|
|
|
|
KVM: Infrastructure for software and hardware based TSC rate scaling
This requires some restructuring; rather than use 'virtual_tsc_khz'
to indicate whether hardware rate scaling is in effect, we consider
each VCPU to always have a virtual TSC rate. Instead, there is new
logic above the vendor-specific hardware scaling that decides whether
it is even necessary to use and updates all rate variables used by
common code. This means we can simply query the virtual rate at
any point, which is needed for software rate scaling.
There is also now a threshold added to the TSC rate scaling; minor
differences and variations of measured TSC rate can accidentally
provoke rate scaling to be used when it is not needed. Instead,
we have a tolerance variable called tsc_tolerance_ppm, which is
the maximum variation from user requested rate at which scaling
will be used. The default is 250ppm, which is the half the
threshold for NTP adjustment, allowing for some hardware variation.
In the event that hardware rate scaling is not available, we can
kludge a bit by forcing TSC catchup to turn on when a faster than
hardware speed has been requested, but there is nothing available
yet for the reverse case; this requires a trap and emulate software
implementation for RDTSC, which is still forthcoming.
[avi: fix 64-bit division on i386]
Signed-off-by: Zachary Amsden <zamsden@gmail.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
2012-02-04 00:43:50 +07:00
|
|
|
if (user_tsc_khz == 0)
|
|
|
|
user_tsc_khz = tsc_khz;
|
|
|
|
|
2015-10-20 14:39:04 +07:00
|
|
|
if (!kvm_set_tsc_khz(vcpu, user_tsc_khz))
|
|
|
|
r = 0;
|
2011-03-25 15:44:51 +07:00
|
|
|
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
case KVM_GET_TSC_KHZ: {
|
KVM: Infrastructure for software and hardware based TSC rate scaling
This requires some restructuring; rather than use 'virtual_tsc_khz'
to indicate whether hardware rate scaling is in effect, we consider
each VCPU to always have a virtual TSC rate. Instead, there is new
logic above the vendor-specific hardware scaling that decides whether
it is even necessary to use and updates all rate variables used by
common code. This means we can simply query the virtual rate at
any point, which is needed for software rate scaling.
There is also now a threshold added to the TSC rate scaling; minor
differences and variations of measured TSC rate can accidentally
provoke rate scaling to be used when it is not needed. Instead,
we have a tolerance variable called tsc_tolerance_ppm, which is
the maximum variation from user requested rate at which scaling
will be used. The default is 250ppm, which is the half the
threshold for NTP adjustment, allowing for some hardware variation.
In the event that hardware rate scaling is not available, we can
kludge a bit by forcing TSC catchup to turn on when a faster than
hardware speed has been requested, but there is nothing available
yet for the reverse case; this requires a trap and emulate software
implementation for RDTSC, which is still forthcoming.
[avi: fix 64-bit division on i386]
Signed-off-by: Zachary Amsden <zamsden@gmail.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
2012-02-04 00:43:50 +07:00
|
|
|
r = vcpu->arch.virtual_tsc_khz;
|
2011-03-25 15:44:51 +07:00
|
|
|
goto out;
|
|
|
|
}
|
2012-03-11 02:37:27 +07:00
|
|
|
case KVM_KVMCLOCK_CTRL: {
|
|
|
|
r = kvm_set_guest_paused(vcpu);
|
|
|
|
goto out;
|
|
|
|
}
|
2015-11-10 19:36:34 +07:00
|
|
|
case KVM_ENABLE_CAP: {
|
|
|
|
struct kvm_enable_cap cap;
|
|
|
|
|
|
|
|
r = -EFAULT;
|
|
|
|
if (copy_from_user(&cap, argp, sizeof(cap)))
|
|
|
|
goto out;
|
|
|
|
r = kvm_vcpu_ioctl_enable_cap(vcpu, &cap);
|
|
|
|
break;
|
|
|
|
}
|
KVM: Portability: split kvm_vcpu_ioctl
This patch splits kvm_vcpu_ioctl into archtecture independent parts, and
x86 specific parts which go to kvm_arch_vcpu_ioctl in x86.c.
Common ioctls for all architectures are:
KVM_RUN, KVM_GET/SET_(S-)REGS, KVM_TRANSLATE, KVM_INTERRUPT,
KVM_DEBUG_GUEST, KVM_SET_SIGNAL_MASK, KVM_GET/SET_FPU
Note that some PPC chips don't have an FPU, so we might need an #ifdef
around KVM_GET/SET_FPU one day.
x86 specific ioctls are:
KVM_GET/SET_LAPIC, KVM_SET_CPUID, KVM_GET/SET_MSRS
An interresting aspect is vcpu_load/vcpu_put. We now have a common
vcpu_load/put which does the preemption stuff, and an architecture
specific kvm_arch_vcpu_load/put. In the x86 case, this one calls the
vmx/svm function defined in kvm_x86_ops.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Reviewed-by: Christian Ehrhardt <ehrhardt@linux.vnet.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-12 00:16:52 +07:00
|
|
|
default:
|
|
|
|
r = -EINVAL;
|
|
|
|
}
|
|
|
|
out:
|
2010-06-20 19:54:43 +07:00
|
|
|
kfree(u.buffer);
|
KVM: Portability: split kvm_vcpu_ioctl
This patch splits kvm_vcpu_ioctl into archtecture independent parts, and
x86 specific parts which go to kvm_arch_vcpu_ioctl in x86.c.
Common ioctls for all architectures are:
KVM_RUN, KVM_GET/SET_(S-)REGS, KVM_TRANSLATE, KVM_INTERRUPT,
KVM_DEBUG_GUEST, KVM_SET_SIGNAL_MASK, KVM_GET/SET_FPU
Note that some PPC chips don't have an FPU, so we might need an #ifdef
around KVM_GET/SET_FPU one day.
x86 specific ioctls are:
KVM_GET/SET_LAPIC, KVM_SET_CPUID, KVM_GET/SET_MSRS
An interresting aspect is vcpu_load/vcpu_put. We now have a common
vcpu_load/put which does the preemption stuff, and an architecture
specific kvm_arch_vcpu_load/put. In the x86 case, this one calls the
vmx/svm function defined in kvm_x86_ops.
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Reviewed-by: Christian Ehrhardt <ehrhardt@linux.vnet.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-12 00:16:52 +07:00
|
|
|
return r;
|
|
|
|
}
|
|
|
|
|
2012-01-04 16:25:23 +07:00
|
|
|
int kvm_arch_vcpu_fault(struct kvm_vcpu *vcpu, struct vm_fault *vmf)
|
|
|
|
{
|
|
|
|
return VM_FAULT_SIGBUS;
|
|
|
|
}
|
|
|
|
|
2007-10-29 22:08:35 +07:00
|
|
|
static int kvm_vm_ioctl_set_tss_addr(struct kvm *kvm, unsigned long addr)
|
|
|
|
{
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
if (addr > (unsigned int)(-3 * PAGE_SIZE))
|
2012-11-02 17:33:22 +07:00
|
|
|
return -EINVAL;
|
2007-10-29 22:08:35 +07:00
|
|
|
ret = kvm_x86_ops->set_tss_addr(kvm, addr);
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
2009-07-21 09:42:48 +07:00
|
|
|
static int kvm_vm_ioctl_set_identity_map_addr(struct kvm *kvm,
|
|
|
|
u64 ident_addr)
|
|
|
|
{
|
|
|
|
kvm->arch.ept_identity_map_addr = ident_addr;
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2007-10-29 22:08:35 +07:00
|
|
|
static int kvm_vm_ioctl_set_nr_mmu_pages(struct kvm *kvm,
|
|
|
|
u32 kvm_nr_mmu_pages)
|
|
|
|
{
|
|
|
|
if (kvm_nr_mmu_pages < KVM_MIN_ALLOC_MMU_PAGES)
|
|
|
|
return -EINVAL;
|
|
|
|
|
2009-12-23 23:35:26 +07:00
|
|
|
mutex_lock(&kvm->slots_lock);
|
2007-10-29 22:08:35 +07:00
|
|
|
|
|
|
|
kvm_mmu_change_mmu_pages(kvm, kvm_nr_mmu_pages);
|
2007-12-14 09:01:48 +07:00
|
|
|
kvm->arch.n_requested_mmu_pages = kvm_nr_mmu_pages;
|
2007-10-29 22:08:35 +07:00
|
|
|
|
2009-12-23 23:35:26 +07:00
|
|
|
mutex_unlock(&kvm->slots_lock);
|
2007-10-29 22:08:35 +07:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int kvm_vm_ioctl_get_nr_mmu_pages(struct kvm *kvm)
|
|
|
|
{
|
2010-08-20 08:11:14 +07:00
|
|
|
return kvm->arch.n_max_mmu_pages;
|
2007-10-29 22:08:35 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
static int kvm_vm_ioctl_get_irqchip(struct kvm *kvm, struct kvm_irqchip *chip)
|
|
|
|
{
|
|
|
|
int r;
|
|
|
|
|
|
|
|
r = 0;
|
|
|
|
switch (chip->chip_id) {
|
|
|
|
case KVM_IRQCHIP_PIC_MASTER:
|
|
|
|
memcpy(&chip->chip.pic,
|
|
|
|
&pic_irqchip(kvm)->pics[0],
|
|
|
|
sizeof(struct kvm_pic_state));
|
|
|
|
break;
|
|
|
|
case KVM_IRQCHIP_PIC_SLAVE:
|
|
|
|
memcpy(&chip->chip.pic,
|
|
|
|
&pic_irqchip(kvm)->pics[1],
|
|
|
|
sizeof(struct kvm_pic_state));
|
|
|
|
break;
|
|
|
|
case KVM_IRQCHIP_IOAPIC:
|
2009-08-24 15:54:25 +07:00
|
|
|
r = kvm_get_ioapic(kvm, &chip->chip.ioapic);
|
2007-10-29 22:08:35 +07:00
|
|
|
break;
|
|
|
|
default:
|
|
|
|
r = -EINVAL;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
return r;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int kvm_vm_ioctl_set_irqchip(struct kvm *kvm, struct kvm_irqchip *chip)
|
|
|
|
{
|
|
|
|
int r;
|
|
|
|
|
|
|
|
r = 0;
|
|
|
|
switch (chip->chip_id) {
|
|
|
|
case KVM_IRQCHIP_PIC_MASTER:
|
2010-09-19 23:44:07 +07:00
|
|
|
spin_lock(&pic_irqchip(kvm)->lock);
|
2007-10-29 22:08:35 +07:00
|
|
|
memcpy(&pic_irqchip(kvm)->pics[0],
|
|
|
|
&chip->chip.pic,
|
|
|
|
sizeof(struct kvm_pic_state));
|
2010-09-19 23:44:07 +07:00
|
|
|
spin_unlock(&pic_irqchip(kvm)->lock);
|
2007-10-29 22:08:35 +07:00
|
|
|
break;
|
|
|
|
case KVM_IRQCHIP_PIC_SLAVE:
|
2010-09-19 23:44:07 +07:00
|
|
|
spin_lock(&pic_irqchip(kvm)->lock);
|
2007-10-29 22:08:35 +07:00
|
|
|
memcpy(&pic_irqchip(kvm)->pics[1],
|
|
|
|
&chip->chip.pic,
|
|
|
|
sizeof(struct kvm_pic_state));
|
2010-09-19 23:44:07 +07:00
|
|
|
spin_unlock(&pic_irqchip(kvm)->lock);
|
2007-10-29 22:08:35 +07:00
|
|
|
break;
|
|
|
|
case KVM_IRQCHIP_IOAPIC:
|
2009-08-24 15:54:25 +07:00
|
|
|
r = kvm_set_ioapic(kvm, &chip->chip.ioapic);
|
2007-10-29 22:08:35 +07:00
|
|
|
break;
|
|
|
|
default:
|
|
|
|
r = -EINVAL;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
kvm_pic_update_irq(pic_irqchip(kvm));
|
|
|
|
return r;
|
|
|
|
}
|
|
|
|
|
2008-03-03 23:50:59 +07:00
|
|
|
static int kvm_vm_ioctl_get_pit(struct kvm *kvm, struct kvm_pit_state *ps)
|
|
|
|
{
|
2016-03-03 04:56:50 +07:00
|
|
|
struct kvm_kpit_state *kps = &kvm->arch.vpit->pit_state;
|
|
|
|
|
|
|
|
BUILD_BUG_ON(sizeof(*ps) != sizeof(kps->channels));
|
|
|
|
|
|
|
|
mutex_lock(&kps->lock);
|
|
|
|
memcpy(ps, &kps->channels, sizeof(*ps));
|
|
|
|
mutex_unlock(&kps->lock);
|
2015-10-30 14:26:11 +07:00
|
|
|
return 0;
|
2008-03-03 23:50:59 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
static int kvm_vm_ioctl_set_pit(struct kvm *kvm, struct kvm_pit_state *ps)
|
|
|
|
{
|
2015-11-19 05:50:23 +07:00
|
|
|
int i;
|
2016-03-03 04:56:43 +07:00
|
|
|
struct kvm_pit *pit = kvm->arch.vpit;
|
|
|
|
|
|
|
|
mutex_lock(&pit->pit_state.lock);
|
2016-03-03 04:56:50 +07:00
|
|
|
memcpy(&pit->pit_state.channels, ps, sizeof(*ps));
|
2015-11-19 05:50:23 +07:00
|
|
|
for (i = 0; i < 3; i++)
|
2016-03-03 04:56:43 +07:00
|
|
|
kvm_pit_load_count(pit, i, ps->channels[i].count, 0);
|
|
|
|
mutex_unlock(&pit->pit_state.lock);
|
2015-10-30 14:26:11 +07:00
|
|
|
return 0;
|
2009-07-07 22:50:38 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
static int kvm_vm_ioctl_get_pit2(struct kvm *kvm, struct kvm_pit_state2 *ps)
|
|
|
|
{
|
|
|
|
mutex_lock(&kvm->arch.vpit->pit_state.lock);
|
|
|
|
memcpy(ps->channels, &kvm->arch.vpit->pit_state.channels,
|
|
|
|
sizeof(ps->channels));
|
|
|
|
ps->flags = kvm->arch.vpit->pit_state.flags;
|
|
|
|
mutex_unlock(&kvm->arch.vpit->pit_state.lock);
|
2010-10-31 01:54:47 +07:00
|
|
|
memset(&ps->reserved, 0, sizeof(ps->reserved));
|
2015-10-30 14:26:11 +07:00
|
|
|
return 0;
|
2009-07-07 22:50:38 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
static int kvm_vm_ioctl_set_pit2(struct kvm *kvm, struct kvm_pit_state2 *ps)
|
|
|
|
{
|
2015-10-30 14:26:11 +07:00
|
|
|
int start = 0;
|
2015-11-19 05:50:23 +07:00
|
|
|
int i;
|
2009-07-07 22:50:38 +07:00
|
|
|
u32 prev_legacy, cur_legacy;
|
2016-03-03 04:56:43 +07:00
|
|
|
struct kvm_pit *pit = kvm->arch.vpit;
|
|
|
|
|
|
|
|
mutex_lock(&pit->pit_state.lock);
|
|
|
|
prev_legacy = pit->pit_state.flags & KVM_PIT_FLAGS_HPET_LEGACY;
|
2009-07-07 22:50:38 +07:00
|
|
|
cur_legacy = ps->flags & KVM_PIT_FLAGS_HPET_LEGACY;
|
|
|
|
if (!prev_legacy && cur_legacy)
|
|
|
|
start = 1;
|
2016-03-03 04:56:43 +07:00
|
|
|
memcpy(&pit->pit_state.channels, &ps->channels,
|
|
|
|
sizeof(pit->pit_state.channels));
|
|
|
|
pit->pit_state.flags = ps->flags;
|
2015-11-19 05:50:23 +07:00
|
|
|
for (i = 0; i < 3; i++)
|
2016-03-03 04:56:43 +07:00
|
|
|
kvm_pit_load_count(pit, i, pit->pit_state.channels[i].count,
|
2016-01-07 19:50:38 +07:00
|
|
|
start && i == 0);
|
2016-03-03 04:56:43 +07:00
|
|
|
mutex_unlock(&pit->pit_state.lock);
|
2015-10-30 14:26:11 +07:00
|
|
|
return 0;
|
2008-03-03 23:50:59 +07:00
|
|
|
}
|
|
|
|
|
2008-12-31 00:55:06 +07:00
|
|
|
static int kvm_vm_ioctl_reinject(struct kvm *kvm,
|
|
|
|
struct kvm_reinject_control *control)
|
|
|
|
{
|
2016-03-03 04:56:45 +07:00
|
|
|
struct kvm_pit *pit = kvm->arch.vpit;
|
|
|
|
|
|
|
|
if (!pit)
|
2008-12-31 00:55:06 +07:00
|
|
|
return -ENXIO;
|
2016-03-03 04:56:44 +07:00
|
|
|
|
2016-03-03 04:56:45 +07:00
|
|
|
/* pit->pit_state.lock was overloaded to prevent userspace from getting
|
|
|
|
* an inconsistent state after running multiple KVM_REINJECT_CONTROL
|
|
|
|
* ioctls in parallel. Use a separate lock if that ioctl isn't rare.
|
|
|
|
*/
|
|
|
|
mutex_lock(&pit->pit_state.lock);
|
|
|
|
kvm_pit_set_reinject(pit, control->pit_reinject);
|
|
|
|
mutex_unlock(&pit->pit_state.lock);
|
2016-03-03 04:56:44 +07:00
|
|
|
|
2008-12-31 00:55:06 +07:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2011-11-14 16:24:50 +07:00
|
|
|
/**
|
2012-03-03 12:21:48 +07:00
|
|
|
* kvm_vm_ioctl_get_dirty_log - get and clear the log of dirty pages in a slot
|
|
|
|
* @kvm: kvm instance
|
|
|
|
* @log: slot id and address to which we copy the log
|
2011-11-14 16:24:50 +07:00
|
|
|
*
|
2015-01-16 06:58:54 +07:00
|
|
|
* Steps 1-4 below provide general overview of dirty page logging. See
|
|
|
|
* kvm_get_dirty_log_protect() function description for additional details.
|
|
|
|
*
|
|
|
|
* We call kvm_get_dirty_log_protect() to handle steps 1-3, upon return we
|
|
|
|
* always flush the TLB (step 4) even if previous step failed and the dirty
|
|
|
|
* bitmap may be corrupt. Regardless of previous outcome the KVM logging API
|
|
|
|
* does not preclude user space subsequent dirty log read. Flushing TLB ensures
|
|
|
|
* writes will be marked dirty for next log read.
|
2011-11-14 16:24:50 +07:00
|
|
|
*
|
2012-03-03 12:21:48 +07:00
|
|
|
* 1. Take a snapshot of the bit and clear it if needed.
|
|
|
|
* 2. Write protect the corresponding page.
|
2015-01-16 06:58:54 +07:00
|
|
|
* 3. Copy the snapshot to the userspace.
|
|
|
|
* 4. Flush TLB's if needed.
|
2007-11-18 19:29:43 +07:00
|
|
|
*/
|
2012-03-03 12:21:48 +07:00
|
|
|
int kvm_vm_ioctl_get_dirty_log(struct kvm *kvm, struct kvm_dirty_log *log)
|
2007-11-18 19:29:43 +07:00
|
|
|
{
|
2012-03-03 12:21:48 +07:00
|
|
|
bool is_dirty = false;
|
2015-01-16 06:58:54 +07:00
|
|
|
int r;
|
2007-11-18 19:29:43 +07:00
|
|
|
|
2009-12-23 23:35:26 +07:00
|
|
|
mutex_lock(&kvm->slots_lock);
|
2007-11-18 19:29:43 +07:00
|
|
|
|
2015-01-28 09:54:27 +07:00
|
|
|
/*
|
|
|
|
* Flush potentially hardware-cached dirty pages to dirty_bitmap.
|
|
|
|
*/
|
|
|
|
if (kvm_x86_ops->flush_log_dirty)
|
|
|
|
kvm_x86_ops->flush_log_dirty(kvm);
|
|
|
|
|
2015-01-16 06:58:54 +07:00
|
|
|
r = kvm_get_dirty_log_protect(kvm, log, &is_dirty);
|
2014-04-17 16:06:16 +07:00
|
|
|
|
|
|
|
/*
|
|
|
|
* All the TLBs can be flushed out of mmu lock, see the comments in
|
|
|
|
* kvm_mmu_slot_remove_write_access().
|
|
|
|
*/
|
2015-01-16 06:58:54 +07:00
|
|
|
lockdep_assert_held(&kvm->slots_lock);
|
2014-04-17 16:06:16 +07:00
|
|
|
if (is_dirty)
|
|
|
|
kvm_flush_remote_tlbs(kvm);
|
|
|
|
|
2009-12-23 23:35:26 +07:00
|
|
|
mutex_unlock(&kvm->slots_lock);
|
2007-11-18 19:29:43 +07:00
|
|
|
return r;
|
|
|
|
}
|
|
|
|
|
2013-04-11 18:21:40 +07:00
|
|
|
int kvm_vm_ioctl_irq_line(struct kvm *kvm, struct kvm_irq_level *irq_event,
|
|
|
|
bool line_status)
|
2012-07-24 19:51:20 +07:00
|
|
|
{
|
|
|
|
if (!irqchip_in_kernel(kvm))
|
|
|
|
return -ENXIO;
|
|
|
|
|
|
|
|
irq_event->status = kvm_set_irq(kvm, KVM_USERSPACE_IRQ_SOURCE_ID,
|
2013-04-11 18:21:40 +07:00
|
|
|
irq_event->irq, irq_event->level,
|
|
|
|
line_status);
|
2012-07-24 19:51:20 +07:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2015-04-13 05:53:41 +07:00
|
|
|
static int kvm_vm_ioctl_enable_cap(struct kvm *kvm,
|
|
|
|
struct kvm_enable_cap *cap)
|
|
|
|
{
|
|
|
|
int r;
|
|
|
|
|
|
|
|
if (cap->flags)
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
switch (cap->cap) {
|
|
|
|
case KVM_CAP_DISABLE_QUIRKS:
|
|
|
|
kvm->arch.disabled_quirks = cap->args[0];
|
|
|
|
r = 0;
|
|
|
|
break;
|
2015-07-30 13:21:40 +07:00
|
|
|
case KVM_CAP_SPLIT_IRQCHIP: {
|
|
|
|
mutex_lock(&kvm->lock);
|
2015-07-30 13:32:35 +07:00
|
|
|
r = -EINVAL;
|
|
|
|
if (cap->args[0] > MAX_NR_RESERVED_IOAPIC_PINS)
|
|
|
|
goto split_irqchip_unlock;
|
2015-07-30 13:21:40 +07:00
|
|
|
r = -EEXIST;
|
|
|
|
if (irqchip_in_kernel(kvm))
|
|
|
|
goto split_irqchip_unlock;
|
2016-06-13 19:50:04 +07:00
|
|
|
if (kvm->created_vcpus)
|
2015-07-30 13:21:40 +07:00
|
|
|
goto split_irqchip_unlock;
|
|
|
|
r = kvm_setup_empty_irq_routing(kvm);
|
|
|
|
if (r)
|
|
|
|
goto split_irqchip_unlock;
|
|
|
|
/* Pairs with irqchip_in_kernel. */
|
|
|
|
smp_wmb();
|
|
|
|
kvm->arch.irqchip_split = true;
|
2015-07-30 13:32:35 +07:00
|
|
|
kvm->arch.nr_reserved_ioapic_pins = cap->args[0];
|
2015-07-30 13:21:40 +07:00
|
|
|
r = 0;
|
|
|
|
split_irqchip_unlock:
|
|
|
|
mutex_unlock(&kvm->lock);
|
|
|
|
break;
|
|
|
|
}
|
2016-07-13 03:09:27 +07:00
|
|
|
case KVM_CAP_X2APIC_API:
|
|
|
|
r = -EINVAL;
|
|
|
|
if (cap->args[0] & ~KVM_X2APIC_API_VALID_FLAGS)
|
|
|
|
break;
|
|
|
|
|
|
|
|
if (cap->args[0] & KVM_X2APIC_API_USE_32BIT_IDS)
|
|
|
|
kvm->arch.x2apic_format = true;
|
2016-07-13 03:09:28 +07:00
|
|
|
if (cap->args[0] & KVM_X2APIC_API_DISABLE_BROADCAST_QUIRK)
|
|
|
|
kvm->arch.x2apic_broadcast_quirk_disabled = true;
|
2016-07-13 03:09:27 +07:00
|
|
|
|
|
|
|
r = 0;
|
|
|
|
break;
|
2015-04-13 05:53:41 +07:00
|
|
|
default:
|
|
|
|
r = -EINVAL;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
return r;
|
|
|
|
}
|
|
|
|
|
2007-10-29 22:08:35 +07:00
|
|
|
long kvm_arch_vm_ioctl(struct file *filp,
|
|
|
|
unsigned int ioctl, unsigned long arg)
|
|
|
|
{
|
|
|
|
struct kvm *kvm = filp->private_data;
|
|
|
|
void __user *argp = (void __user *)arg;
|
2009-08-26 18:57:07 +07:00
|
|
|
int r = -ENOTTY;
|
2008-08-12 00:01:45 +07:00
|
|
|
/*
|
|
|
|
* This union makes it completely explicit to gcc-3.x
|
|
|
|
* that these two variables' stack usage should be
|
|
|
|
* combined, not added together.
|
|
|
|
*/
|
|
|
|
union {
|
|
|
|
struct kvm_pit_state ps;
|
2009-07-07 22:50:38 +07:00
|
|
|
struct kvm_pit_state2 ps2;
|
2009-05-15 03:42:53 +07:00
|
|
|
struct kvm_pit_config pit_config;
|
2008-08-12 00:01:45 +07:00
|
|
|
} u;
|
2007-10-29 22:08:35 +07:00
|
|
|
|
|
|
|
switch (ioctl) {
|
|
|
|
case KVM_SET_TSS_ADDR:
|
|
|
|
r = kvm_vm_ioctl_set_tss_addr(kvm, arg);
|
|
|
|
break;
|
2009-07-21 09:42:48 +07:00
|
|
|
case KVM_SET_IDENTITY_MAP_ADDR: {
|
|
|
|
u64 ident_addr;
|
|
|
|
|
|
|
|
r = -EFAULT;
|
|
|
|
if (copy_from_user(&ident_addr, argp, sizeof ident_addr))
|
|
|
|
goto out;
|
|
|
|
r = kvm_vm_ioctl_set_identity_map_addr(kvm, ident_addr);
|
|
|
|
break;
|
|
|
|
}
|
2007-10-29 22:08:35 +07:00
|
|
|
case KVM_SET_NR_MMU_PAGES:
|
|
|
|
r = kvm_vm_ioctl_set_nr_mmu_pages(kvm, arg);
|
|
|
|
break;
|
|
|
|
case KVM_GET_NR_MMU_PAGES:
|
|
|
|
r = kvm_vm_ioctl_get_nr_mmu_pages(kvm);
|
|
|
|
break;
|
2009-10-29 22:44:15 +07:00
|
|
|
case KVM_CREATE_IRQCHIP: {
|
|
|
|
struct kvm_pic *vpic;
|
|
|
|
|
|
|
|
mutex_lock(&kvm->lock);
|
|
|
|
r = -EEXIST;
|
|
|
|
if (kvm->arch.vpic)
|
|
|
|
goto create_irqchip_unlock;
|
2012-03-05 19:23:29 +07:00
|
|
|
r = -EINVAL;
|
2016-06-13 19:50:04 +07:00
|
|
|
if (kvm->created_vcpus)
|
2012-03-05 19:23:29 +07:00
|
|
|
goto create_irqchip_unlock;
|
2007-10-29 22:08:35 +07:00
|
|
|
r = -ENOMEM;
|
2009-10-29 22:44:15 +07:00
|
|
|
vpic = kvm_create_pic(kvm);
|
|
|
|
if (vpic) {
|
2007-10-29 22:08:35 +07:00
|
|
|
r = kvm_ioapic_init(kvm);
|
|
|
|
if (r) {
|
2010-12-15 23:41:37 +07:00
|
|
|
mutex_lock(&kvm->slots_lock);
|
2015-07-29 17:31:15 +07:00
|
|
|
kvm_destroy_pic(vpic);
|
2010-12-15 23:41:37 +07:00
|
|
|
mutex_unlock(&kvm->slots_lock);
|
2009-10-29 22:44:15 +07:00
|
|
|
goto create_irqchip_unlock;
|
2007-10-29 22:08:35 +07:00
|
|
|
}
|
|
|
|
} else
|
2009-10-29 22:44:15 +07:00
|
|
|
goto create_irqchip_unlock;
|
2008-11-19 18:58:46 +07:00
|
|
|
r = kvm_setup_default_irq_routing(kvm);
|
|
|
|
if (r) {
|
2010-12-15 23:41:37 +07:00
|
|
|
mutex_lock(&kvm->slots_lock);
|
2009-10-29 22:44:15 +07:00
|
|
|
mutex_lock(&kvm->irq_lock);
|
2010-02-09 09:33:03 +07:00
|
|
|
kvm_ioapic_destroy(kvm);
|
2015-07-29 17:31:15 +07:00
|
|
|
kvm_destroy_pic(vpic);
|
2009-10-29 22:44:15 +07:00
|
|
|
mutex_unlock(&kvm->irq_lock);
|
2010-12-15 23:41:37 +07:00
|
|
|
mutex_unlock(&kvm->slots_lock);
|
2015-07-29 17:31:15 +07:00
|
|
|
goto create_irqchip_unlock;
|
2008-11-19 18:58:46 +07:00
|
|
|
}
|
2015-07-29 17:31:15 +07:00
|
|
|
/* Write kvm->irq_routing before kvm->arch.vpic. */
|
|
|
|
smp_wmb();
|
|
|
|
kvm->arch.vpic = vpic;
|
2009-10-29 22:44:15 +07:00
|
|
|
create_irqchip_unlock:
|
|
|
|
mutex_unlock(&kvm->lock);
|
2007-10-29 22:08:35 +07:00
|
|
|
break;
|
2009-10-29 22:44:15 +07:00
|
|
|
}
|
2008-01-28 04:10:22 +07:00
|
|
|
case KVM_CREATE_PIT:
|
2009-05-15 03:42:53 +07:00
|
|
|
u.pit_config.flags = KVM_PIT_SPEAKER_DUMMY;
|
|
|
|
goto create_pit;
|
|
|
|
case KVM_CREATE_PIT2:
|
|
|
|
r = -EFAULT;
|
|
|
|
if (copy_from_user(&u.pit_config, argp,
|
|
|
|
sizeof(struct kvm_pit_config)))
|
|
|
|
goto out;
|
|
|
|
create_pit:
|
KVM: x86: protect KVM_CREATE_PIT/KVM_CREATE_PIT2 with kvm->lock
The syzkaller folks reported a NULL pointer dereference that seems
to be cause by a race between KVM_CREATE_IRQCHIP and KVM_CREATE_PIT2.
The former takes kvm->lock (except when registering the devices,
which needs kvm->slots_lock); the latter takes kvm->slots_lock only.
Change KVM_CREATE_PIT2 to follow the same model as KVM_CREATE_IRQCHIP.
Testcase:
#include <pthread.h>
#include <linux/kvm.h>
#include <fcntl.h>
#include <sys/ioctl.h>
#include <stdint.h>
#include <string.h>
#include <stdlib.h>
#include <sys/syscall.h>
#include <unistd.h>
long r[23];
void* thr1(void* arg)
{
struct kvm_pit_config pitcfg = { .flags = 4 };
switch ((long)arg) {
case 0: r[2] = open("/dev/kvm", O_RDONLY|O_ASYNC); break;
case 1: r[3] = ioctl(r[2], KVM_CREATE_VM, 0); break;
case 2: r[4] = ioctl(r[3], KVM_CREATE_IRQCHIP, 0); break;
case 3: r[22] = ioctl(r[3], KVM_CREATE_PIT2, &pitcfg); break;
}
return 0;
}
int main(int argc, char **argv)
{
long i;
pthread_t th[4];
memset(r, -1, sizeof(r));
for (i = 0; i < 4; i++) {
pthread_create(&th[i], 0, thr, (void*)i);
if (argc > 1 && rand()%2) usleep(rand()%1000);
}
usleep(20000);
return 0;
}
Reported-by: Dmitry Vyukov <dvyukov@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Radim Krčmář <rkrcmar@redhat.com>
2016-06-01 19:09:24 +07:00
|
|
|
mutex_lock(&kvm->lock);
|
2009-01-05 20:21:42 +07:00
|
|
|
r = -EEXIST;
|
|
|
|
if (kvm->arch.vpit)
|
|
|
|
goto create_pit_unlock;
|
2008-01-28 04:10:22 +07:00
|
|
|
r = -ENOMEM;
|
2009-05-15 03:42:53 +07:00
|
|
|
kvm->arch.vpit = kvm_create_pit(kvm, u.pit_config.flags);
|
2008-01-28 04:10:22 +07:00
|
|
|
if (kvm->arch.vpit)
|
|
|
|
r = 0;
|
2009-01-05 20:21:42 +07:00
|
|
|
create_pit_unlock:
|
KVM: x86: protect KVM_CREATE_PIT/KVM_CREATE_PIT2 with kvm->lock
The syzkaller folks reported a NULL pointer dereference that seems
to be cause by a race between KVM_CREATE_IRQCHIP and KVM_CREATE_PIT2.
The former takes kvm->lock (except when registering the devices,
which needs kvm->slots_lock); the latter takes kvm->slots_lock only.
Change KVM_CREATE_PIT2 to follow the same model as KVM_CREATE_IRQCHIP.
Testcase:
#include <pthread.h>
#include <linux/kvm.h>
#include <fcntl.h>
#include <sys/ioctl.h>
#include <stdint.h>
#include <string.h>
#include <stdlib.h>
#include <sys/syscall.h>
#include <unistd.h>
long r[23];
void* thr1(void* arg)
{
struct kvm_pit_config pitcfg = { .flags = 4 };
switch ((long)arg) {
case 0: r[2] = open("/dev/kvm", O_RDONLY|O_ASYNC); break;
case 1: r[3] = ioctl(r[2], KVM_CREATE_VM, 0); break;
case 2: r[4] = ioctl(r[3], KVM_CREATE_IRQCHIP, 0); break;
case 3: r[22] = ioctl(r[3], KVM_CREATE_PIT2, &pitcfg); break;
}
return 0;
}
int main(int argc, char **argv)
{
long i;
pthread_t th[4];
memset(r, -1, sizeof(r));
for (i = 0; i < 4; i++) {
pthread_create(&th[i], 0, thr, (void*)i);
if (argc > 1 && rand()%2) usleep(rand()%1000);
}
usleep(20000);
return 0;
}
Reported-by: Dmitry Vyukov <dvyukov@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Radim Krčmář <rkrcmar@redhat.com>
2016-06-01 19:09:24 +07:00
|
|
|
mutex_unlock(&kvm->lock);
|
2008-01-28 04:10:22 +07:00
|
|
|
break;
|
2007-10-29 22:08:35 +07:00
|
|
|
case KVM_GET_IRQCHIP: {
|
|
|
|
/* 0: PIC master, 1: PIC slave, 2: IOAPIC */
|
2011-12-05 00:36:29 +07:00
|
|
|
struct kvm_irqchip *chip;
|
2007-10-29 22:08:35 +07:00
|
|
|
|
2011-12-05 00:36:29 +07:00
|
|
|
chip = memdup_user(argp, sizeof(*chip));
|
|
|
|
if (IS_ERR(chip)) {
|
|
|
|
r = PTR_ERR(chip);
|
2007-10-29 22:08:35 +07:00
|
|
|
goto out;
|
2011-12-05 00:36:29 +07:00
|
|
|
}
|
|
|
|
|
2007-10-29 22:08:35 +07:00
|
|
|
r = -ENXIO;
|
2015-07-30 13:21:40 +07:00
|
|
|
if (!irqchip_in_kernel(kvm) || irqchip_split(kvm))
|
2008-08-12 00:01:45 +07:00
|
|
|
goto get_irqchip_out;
|
|
|
|
r = kvm_vm_ioctl_get_irqchip(kvm, chip);
|
2007-10-29 22:08:35 +07:00
|
|
|
if (r)
|
2008-08-12 00:01:45 +07:00
|
|
|
goto get_irqchip_out;
|
2007-10-29 22:08:35 +07:00
|
|
|
r = -EFAULT;
|
2008-08-12 00:01:45 +07:00
|
|
|
if (copy_to_user(argp, chip, sizeof *chip))
|
|
|
|
goto get_irqchip_out;
|
2007-10-29 22:08:35 +07:00
|
|
|
r = 0;
|
2008-08-12 00:01:45 +07:00
|
|
|
get_irqchip_out:
|
|
|
|
kfree(chip);
|
2007-10-29 22:08:35 +07:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
case KVM_SET_IRQCHIP: {
|
|
|
|
/* 0: PIC master, 1: PIC slave, 2: IOAPIC */
|
2011-12-05 00:36:29 +07:00
|
|
|
struct kvm_irqchip *chip;
|
2007-10-29 22:08:35 +07:00
|
|
|
|
2011-12-05 00:36:29 +07:00
|
|
|
chip = memdup_user(argp, sizeof(*chip));
|
|
|
|
if (IS_ERR(chip)) {
|
|
|
|
r = PTR_ERR(chip);
|
2007-10-29 22:08:35 +07:00
|
|
|
goto out;
|
2011-12-05 00:36:29 +07:00
|
|
|
}
|
|
|
|
|
2007-10-29 22:08:35 +07:00
|
|
|
r = -ENXIO;
|
2015-07-30 13:21:40 +07:00
|
|
|
if (!irqchip_in_kernel(kvm) || irqchip_split(kvm))
|
2008-08-12 00:01:45 +07:00
|
|
|
goto set_irqchip_out;
|
|
|
|
r = kvm_vm_ioctl_set_irqchip(kvm, chip);
|
2007-10-29 22:08:35 +07:00
|
|
|
if (r)
|
2008-08-12 00:01:45 +07:00
|
|
|
goto set_irqchip_out;
|
2007-10-29 22:08:35 +07:00
|
|
|
r = 0;
|
2008-08-12 00:01:45 +07:00
|
|
|
set_irqchip_out:
|
|
|
|
kfree(chip);
|
2007-10-29 22:08:35 +07:00
|
|
|
break;
|
|
|
|
}
|
2008-03-03 23:50:59 +07:00
|
|
|
case KVM_GET_PIT: {
|
|
|
|
r = -EFAULT;
|
2008-08-12 00:01:45 +07:00
|
|
|
if (copy_from_user(&u.ps, argp, sizeof(struct kvm_pit_state)))
|
2008-03-03 23:50:59 +07:00
|
|
|
goto out;
|
|
|
|
r = -ENXIO;
|
|
|
|
if (!kvm->arch.vpit)
|
|
|
|
goto out;
|
2008-08-12 00:01:45 +07:00
|
|
|
r = kvm_vm_ioctl_get_pit(kvm, &u.ps);
|
2008-03-03 23:50:59 +07:00
|
|
|
if (r)
|
|
|
|
goto out;
|
|
|
|
r = -EFAULT;
|
2008-08-12 00:01:45 +07:00
|
|
|
if (copy_to_user(argp, &u.ps, sizeof(struct kvm_pit_state)))
|
2008-03-03 23:50:59 +07:00
|
|
|
goto out;
|
|
|
|
r = 0;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
case KVM_SET_PIT: {
|
|
|
|
r = -EFAULT;
|
2008-08-12 00:01:45 +07:00
|
|
|
if (copy_from_user(&u.ps, argp, sizeof u.ps))
|
2008-03-03 23:50:59 +07:00
|
|
|
goto out;
|
|
|
|
r = -ENXIO;
|
|
|
|
if (!kvm->arch.vpit)
|
|
|
|
goto out;
|
2008-08-12 00:01:45 +07:00
|
|
|
r = kvm_vm_ioctl_set_pit(kvm, &u.ps);
|
2008-03-03 23:50:59 +07:00
|
|
|
break;
|
|
|
|
}
|
2009-07-07 22:50:38 +07:00
|
|
|
case KVM_GET_PIT2: {
|
|
|
|
r = -ENXIO;
|
|
|
|
if (!kvm->arch.vpit)
|
|
|
|
goto out;
|
|
|
|
r = kvm_vm_ioctl_get_pit2(kvm, &u.ps2);
|
|
|
|
if (r)
|
|
|
|
goto out;
|
|
|
|
r = -EFAULT;
|
|
|
|
if (copy_to_user(argp, &u.ps2, sizeof(u.ps2)))
|
|
|
|
goto out;
|
|
|
|
r = 0;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
case KVM_SET_PIT2: {
|
|
|
|
r = -EFAULT;
|
|
|
|
if (copy_from_user(&u.ps2, argp, sizeof(u.ps2)))
|
|
|
|
goto out;
|
|
|
|
r = -ENXIO;
|
|
|
|
if (!kvm->arch.vpit)
|
|
|
|
goto out;
|
|
|
|
r = kvm_vm_ioctl_set_pit2(kvm, &u.ps2);
|
|
|
|
break;
|
|
|
|
}
|
2008-12-31 00:55:06 +07:00
|
|
|
case KVM_REINJECT_CONTROL: {
|
|
|
|
struct kvm_reinject_control control;
|
|
|
|
r = -EFAULT;
|
|
|
|
if (copy_from_user(&control, argp, sizeof(control)))
|
|
|
|
goto out;
|
|
|
|
r = kvm_vm_ioctl_reinject(kvm, &control);
|
|
|
|
break;
|
|
|
|
}
|
2015-07-29 16:56:48 +07:00
|
|
|
case KVM_SET_BOOT_CPU_ID:
|
|
|
|
r = 0;
|
|
|
|
mutex_lock(&kvm->lock);
|
2016-06-13 19:50:04 +07:00
|
|
|
if (kvm->created_vcpus)
|
2015-07-29 16:56:48 +07:00
|
|
|
r = -EBUSY;
|
|
|
|
else
|
|
|
|
kvm->arch.bsp_vcpu_id = arg;
|
|
|
|
mutex_unlock(&kvm->lock);
|
|
|
|
break;
|
2009-10-16 05:21:43 +07:00
|
|
|
case KVM_XEN_HVM_CONFIG: {
|
|
|
|
r = -EFAULT;
|
|
|
|
if (copy_from_user(&kvm->arch.xen_hvm_config, argp,
|
|
|
|
sizeof(struct kvm_xen_hvm_config)))
|
|
|
|
goto out;
|
|
|
|
r = -EINVAL;
|
|
|
|
if (kvm->arch.xen_hvm_config.flags)
|
|
|
|
goto out;
|
|
|
|
r = 0;
|
|
|
|
break;
|
|
|
|
}
|
2009-10-17 02:28:36 +07:00
|
|
|
case KVM_SET_CLOCK: {
|
|
|
|
struct kvm_clock_data user_ns;
|
|
|
|
u64 now_ns;
|
|
|
|
s64 delta;
|
|
|
|
|
|
|
|
r = -EFAULT;
|
|
|
|
if (copy_from_user(&user_ns, argp, sizeof(user_ns)))
|
|
|
|
goto out;
|
|
|
|
|
|
|
|
r = -EINVAL;
|
|
|
|
if (user_ns.flags)
|
|
|
|
goto out;
|
|
|
|
|
|
|
|
r = 0;
|
2010-10-04 17:55:49 +07:00
|
|
|
local_irq_disable();
|
2010-08-20 15:07:25 +07:00
|
|
|
now_ns = get_kernel_ns();
|
2009-10-17 02:28:36 +07:00
|
|
|
delta = user_ns.clock - now_ns;
|
2010-10-04 17:55:49 +07:00
|
|
|
local_irq_enable();
|
2009-10-17 02:28:36 +07:00
|
|
|
kvm->arch.kvmclock_offset = delta;
|
2013-08-28 09:55:29 +07:00
|
|
|
kvm_gen_update_masterclock(kvm);
|
2009-10-17 02:28:36 +07:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
case KVM_GET_CLOCK: {
|
|
|
|
struct kvm_clock_data user_ns;
|
|
|
|
u64 now_ns;
|
|
|
|
|
2010-10-04 17:55:49 +07:00
|
|
|
local_irq_disable();
|
2010-08-20 15:07:25 +07:00
|
|
|
now_ns = get_kernel_ns();
|
2009-10-17 02:28:36 +07:00
|
|
|
user_ns.clock = kvm->arch.kvmclock_offset + now_ns;
|
2010-10-04 17:55:49 +07:00
|
|
|
local_irq_enable();
|
2009-10-17 02:28:36 +07:00
|
|
|
user_ns.flags = 0;
|
2010-10-31 01:54:47 +07:00
|
|
|
memset(&user_ns.pad, 0, sizeof(user_ns.pad));
|
2009-10-17 02:28:36 +07:00
|
|
|
|
|
|
|
r = -EFAULT;
|
|
|
|
if (copy_to_user(argp, &user_ns, sizeof(user_ns)))
|
|
|
|
goto out;
|
|
|
|
r = 0;
|
|
|
|
break;
|
|
|
|
}
|
2015-04-13 05:53:41 +07:00
|
|
|
case KVM_ENABLE_CAP: {
|
|
|
|
struct kvm_enable_cap cap;
|
2009-10-17 02:28:36 +07:00
|
|
|
|
2015-04-13 05:53:41 +07:00
|
|
|
r = -EFAULT;
|
|
|
|
if (copy_from_user(&cap, argp, sizeof(cap)))
|
|
|
|
goto out;
|
|
|
|
r = kvm_vm_ioctl_enable_cap(kvm, &cap);
|
|
|
|
break;
|
|
|
|
}
|
2007-10-29 22:08:35 +07:00
|
|
|
default:
|
2014-11-22 04:21:50 +07:00
|
|
|
r = kvm_vm_ioctl_assigned_device(kvm, ioctl, arg);
|
2007-10-29 22:08:35 +07:00
|
|
|
}
|
|
|
|
out:
|
|
|
|
return r;
|
|
|
|
}
|
|
|
|
|
2007-11-16 13:38:21 +07:00
|
|
|
static void kvm_init_msr_list(void)
|
2007-10-10 22:16:19 +07:00
|
|
|
{
|
|
|
|
u32 dummy[2];
|
|
|
|
unsigned i, j;
|
|
|
|
|
2015-05-05 17:08:55 +07:00
|
|
|
for (i = j = 0; i < ARRAY_SIZE(msrs_to_save); i++) {
|
2007-10-10 22:16:19 +07:00
|
|
|
if (rdmsr_safe(msrs_to_save[i], &dummy[0], &dummy[1]) < 0)
|
|
|
|
continue;
|
2014-03-06 05:19:52 +07:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Even MSRs that are valid in the host may not be exposed
|
2015-11-12 20:49:17 +07:00
|
|
|
* to the guests in some cases.
|
2014-03-06 05:19:52 +07:00
|
|
|
*/
|
|
|
|
switch (msrs_to_save[i]) {
|
|
|
|
case MSR_IA32_BNDCFGS:
|
|
|
|
if (!kvm_x86_ops->mpx_supported())
|
|
|
|
continue;
|
|
|
|
break;
|
2015-11-12 20:49:17 +07:00
|
|
|
case MSR_TSC_AUX:
|
|
|
|
if (!kvm_x86_ops->rdtscp_supported())
|
|
|
|
continue;
|
|
|
|
break;
|
2014-03-06 05:19:52 +07:00
|
|
|
default:
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
2007-10-10 22:16:19 +07:00
|
|
|
if (j < i)
|
|
|
|
msrs_to_save[j] = msrs_to_save[i];
|
|
|
|
j++;
|
|
|
|
}
|
|
|
|
num_msrs_to_save = j;
|
2015-05-05 17:08:55 +07:00
|
|
|
|
|
|
|
for (i = j = 0; i < ARRAY_SIZE(emulated_msrs); i++) {
|
|
|
|
switch (emulated_msrs[i]) {
|
2015-04-01 19:25:33 +07:00
|
|
|
case MSR_IA32_SMBASE:
|
|
|
|
if (!kvm_x86_ops->cpu_has_high_real_mode_segbase())
|
|
|
|
continue;
|
|
|
|
break;
|
2015-05-05 17:08:55 +07:00
|
|
|
default:
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (j < i)
|
|
|
|
emulated_msrs[j] = emulated_msrs[i];
|
|
|
|
j++;
|
|
|
|
}
|
|
|
|
num_emulated_msrs = j;
|
2007-10-10 22:16:19 +07:00
|
|
|
}
|
|
|
|
|
2009-06-30 02:24:32 +07:00
|
|
|
static int vcpu_mmio_write(struct kvm_vcpu *vcpu, gpa_t addr, int len,
|
|
|
|
const void *v)
|
KVM: Portability: Move x86 emulation and mmio device hook to x86.c
This patch moves the following functions to from kvm_main.c to x86.c:
emulator_read/write_std, vcpu_find_pervcpu_dev, vcpu_find_mmio_dev,
emulator_read/write_emulated, emulator_write_phys,
emulator_write_emulated_onepage, emulator_cmpxchg_emulated,
get_setment_base, emulate_invlpg, emulate_clts, emulator_get/set_dr,
kvm_report_emulation_failure, emulate_instruction
The following data type is moved to x86.c:
struct x86_emulate_ops emulate_ops
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-31 00:44:21 +07:00
|
|
|
{
|
2010-01-19 17:51:22 +07:00
|
|
|
int handled = 0;
|
|
|
|
int n;
|
|
|
|
|
|
|
|
do {
|
|
|
|
n = min(len, 8);
|
2016-01-08 19:48:51 +07:00
|
|
|
if (!(lapic_in_kernel(vcpu) &&
|
2015-03-26 21:39:28 +07:00
|
|
|
!kvm_iodevice_write(vcpu, &vcpu->arch.apic->dev, addr, n, v))
|
|
|
|
&& kvm_io_bus_write(vcpu, KVM_MMIO_BUS, addr, n, v))
|
2010-01-19 17:51:22 +07:00
|
|
|
break;
|
|
|
|
handled += n;
|
|
|
|
addr += n;
|
|
|
|
len -= n;
|
|
|
|
v += n;
|
|
|
|
} while (len);
|
KVM: Portability: Move x86 emulation and mmio device hook to x86.c
This patch moves the following functions to from kvm_main.c to x86.c:
emulator_read/write_std, vcpu_find_pervcpu_dev, vcpu_find_mmio_dev,
emulator_read/write_emulated, emulator_write_phys,
emulator_write_emulated_onepage, emulator_cmpxchg_emulated,
get_setment_base, emulate_invlpg, emulate_clts, emulator_get/set_dr,
kvm_report_emulation_failure, emulate_instruction
The following data type is moved to x86.c:
struct x86_emulate_ops emulate_ops
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-31 00:44:21 +07:00
|
|
|
|
2010-01-19 17:51:22 +07:00
|
|
|
return handled;
|
KVM: Portability: Move x86 emulation and mmio device hook to x86.c
This patch moves the following functions to from kvm_main.c to x86.c:
emulator_read/write_std, vcpu_find_pervcpu_dev, vcpu_find_mmio_dev,
emulator_read/write_emulated, emulator_write_phys,
emulator_write_emulated_onepage, emulator_cmpxchg_emulated,
get_setment_base, emulate_invlpg, emulate_clts, emulator_get/set_dr,
kvm_report_emulation_failure, emulate_instruction
The following data type is moved to x86.c:
struct x86_emulate_ops emulate_ops
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-31 00:44:21 +07:00
|
|
|
}
|
|
|
|
|
2009-06-30 02:24:32 +07:00
|
|
|
static int vcpu_mmio_read(struct kvm_vcpu *vcpu, gpa_t addr, int len, void *v)
|
KVM: Portability: Move x86 emulation and mmio device hook to x86.c
This patch moves the following functions to from kvm_main.c to x86.c:
emulator_read/write_std, vcpu_find_pervcpu_dev, vcpu_find_mmio_dev,
emulator_read/write_emulated, emulator_write_phys,
emulator_write_emulated_onepage, emulator_cmpxchg_emulated,
get_setment_base, emulate_invlpg, emulate_clts, emulator_get/set_dr,
kvm_report_emulation_failure, emulate_instruction
The following data type is moved to x86.c:
struct x86_emulate_ops emulate_ops
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-31 00:44:21 +07:00
|
|
|
{
|
2010-01-19 17:51:22 +07:00
|
|
|
int handled = 0;
|
|
|
|
int n;
|
|
|
|
|
|
|
|
do {
|
|
|
|
n = min(len, 8);
|
2016-01-08 19:48:51 +07:00
|
|
|
if (!(lapic_in_kernel(vcpu) &&
|
2015-03-26 21:39:28 +07:00
|
|
|
!kvm_iodevice_read(vcpu, &vcpu->arch.apic->dev,
|
|
|
|
addr, n, v))
|
|
|
|
&& kvm_io_bus_read(vcpu, KVM_MMIO_BUS, addr, n, v))
|
2010-01-19 17:51:22 +07:00
|
|
|
break;
|
|
|
|
trace_kvm_mmio(KVM_TRACE_MMIO_READ, n, addr, *(u64 *)v);
|
|
|
|
handled += n;
|
|
|
|
addr += n;
|
|
|
|
len -= n;
|
|
|
|
v += n;
|
|
|
|
} while (len);
|
KVM: Portability: Move x86 emulation and mmio device hook to x86.c
This patch moves the following functions to from kvm_main.c to x86.c:
emulator_read/write_std, vcpu_find_pervcpu_dev, vcpu_find_mmio_dev,
emulator_read/write_emulated, emulator_write_phys,
emulator_write_emulated_onepage, emulator_cmpxchg_emulated,
get_setment_base, emulate_invlpg, emulate_clts, emulator_get/set_dr,
kvm_report_emulation_failure, emulate_instruction
The following data type is moved to x86.c:
struct x86_emulate_ops emulate_ops
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-31 00:44:21 +07:00
|
|
|
|
2010-01-19 17:51:22 +07:00
|
|
|
return handled;
|
KVM: Portability: Move x86 emulation and mmio device hook to x86.c
This patch moves the following functions to from kvm_main.c to x86.c:
emulator_read/write_std, vcpu_find_pervcpu_dev, vcpu_find_mmio_dev,
emulator_read/write_emulated, emulator_write_phys,
emulator_write_emulated_onepage, emulator_cmpxchg_emulated,
get_setment_base, emulate_invlpg, emulate_clts, emulator_get/set_dr,
kvm_report_emulation_failure, emulate_instruction
The following data type is moved to x86.c:
struct x86_emulate_ops emulate_ops
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-31 00:44:21 +07:00
|
|
|
}
|
|
|
|
|
2010-03-18 20:20:16 +07:00
|
|
|
static void kvm_set_segment(struct kvm_vcpu *vcpu,
|
|
|
|
struct kvm_segment *var, int seg)
|
|
|
|
{
|
|
|
|
kvm_x86_ops->set_segment(vcpu, var, seg);
|
|
|
|
}
|
|
|
|
|
|
|
|
void kvm_get_segment(struct kvm_vcpu *vcpu,
|
|
|
|
struct kvm_segment *var, int seg)
|
|
|
|
{
|
|
|
|
kvm_x86_ops->get_segment(vcpu, var, seg);
|
|
|
|
}
|
|
|
|
|
2014-09-02 18:23:06 +07:00
|
|
|
gpa_t translate_nested_gpa(struct kvm_vcpu *vcpu, gpa_t gpa, u32 access,
|
|
|
|
struct x86_exception *exception)
|
2010-09-10 22:30:54 +07:00
|
|
|
{
|
|
|
|
gpa_t t_gpa;
|
|
|
|
|
|
|
|
BUG_ON(!mmu_is_nested(vcpu));
|
|
|
|
|
|
|
|
/* NPT walks are always user-walks */
|
|
|
|
access |= PFERR_USER_MASK;
|
2014-09-02 18:23:06 +07:00
|
|
|
t_gpa = vcpu->arch.mmu.gva_to_gpa(vcpu, gpa, access, exception);
|
2010-09-10 22:30:54 +07:00
|
|
|
|
|
|
|
return t_gpa;
|
|
|
|
}
|
|
|
|
|
2010-11-22 22:53:26 +07:00
|
|
|
gpa_t kvm_mmu_gva_to_gpa_read(struct kvm_vcpu *vcpu, gva_t gva,
|
|
|
|
struct x86_exception *exception)
|
2010-02-10 19:21:32 +07:00
|
|
|
{
|
|
|
|
u32 access = (kvm_x86_ops->get_cpl(vcpu) == 3) ? PFERR_USER_MASK : 0;
|
2010-11-22 22:53:26 +07:00
|
|
|
return vcpu->arch.walk_mmu->gva_to_gpa(vcpu, gva, access, exception);
|
2010-02-10 19:21:32 +07:00
|
|
|
}
|
|
|
|
|
2010-11-22 22:53:26 +07:00
|
|
|
gpa_t kvm_mmu_gva_to_gpa_fetch(struct kvm_vcpu *vcpu, gva_t gva,
|
|
|
|
struct x86_exception *exception)
|
2010-02-10 19:21:32 +07:00
|
|
|
{
|
|
|
|
u32 access = (kvm_x86_ops->get_cpl(vcpu) == 3) ? PFERR_USER_MASK : 0;
|
|
|
|
access |= PFERR_FETCH_MASK;
|
2010-11-22 22:53:26 +07:00
|
|
|
return vcpu->arch.walk_mmu->gva_to_gpa(vcpu, gva, access, exception);
|
2010-02-10 19:21:32 +07:00
|
|
|
}
|
|
|
|
|
2010-11-22 22:53:26 +07:00
|
|
|
gpa_t kvm_mmu_gva_to_gpa_write(struct kvm_vcpu *vcpu, gva_t gva,
|
|
|
|
struct x86_exception *exception)
|
2010-02-10 19:21:32 +07:00
|
|
|
{
|
|
|
|
u32 access = (kvm_x86_ops->get_cpl(vcpu) == 3) ? PFERR_USER_MASK : 0;
|
|
|
|
access |= PFERR_WRITE_MASK;
|
2010-11-22 22:53:26 +07:00
|
|
|
return vcpu->arch.walk_mmu->gva_to_gpa(vcpu, gva, access, exception);
|
2010-02-10 19:21:32 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
/* uses this to access any guest's mapped memory without checking CPL */
|
2010-11-22 22:53:26 +07:00
|
|
|
gpa_t kvm_mmu_gva_to_gpa_system(struct kvm_vcpu *vcpu, gva_t gva,
|
|
|
|
struct x86_exception *exception)
|
2010-02-10 19:21:32 +07:00
|
|
|
{
|
2010-11-22 22:53:26 +07:00
|
|
|
return vcpu->arch.walk_mmu->gva_to_gpa(vcpu, gva, 0, exception);
|
2010-02-10 19:21:32 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
static int kvm_read_guest_virt_helper(gva_t addr, void *val, unsigned int bytes,
|
|
|
|
struct kvm_vcpu *vcpu, u32 access,
|
2010-11-22 22:53:22 +07:00
|
|
|
struct x86_exception *exception)
|
KVM: Portability: Move x86 emulation and mmio device hook to x86.c
This patch moves the following functions to from kvm_main.c to x86.c:
emulator_read/write_std, vcpu_find_pervcpu_dev, vcpu_find_mmio_dev,
emulator_read/write_emulated, emulator_write_phys,
emulator_write_emulated_onepage, emulator_cmpxchg_emulated,
get_setment_base, emulate_invlpg, emulate_clts, emulator_get/set_dr,
kvm_report_emulation_failure, emulate_instruction
The following data type is moved to x86.c:
struct x86_emulate_ops emulate_ops
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-31 00:44:21 +07:00
|
|
|
{
|
|
|
|
void *data = val;
|
2007-12-21 07:18:22 +07:00
|
|
|
int r = X86EMUL_CONTINUE;
|
KVM: Portability: Move x86 emulation and mmio device hook to x86.c
This patch moves the following functions to from kvm_main.c to x86.c:
emulator_read/write_std, vcpu_find_pervcpu_dev, vcpu_find_mmio_dev,
emulator_read/write_emulated, emulator_write_phys,
emulator_write_emulated_onepage, emulator_cmpxchg_emulated,
get_setment_base, emulate_invlpg, emulate_clts, emulator_get/set_dr,
kvm_report_emulation_failure, emulate_instruction
The following data type is moved to x86.c:
struct x86_emulate_ops emulate_ops
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-31 00:44:21 +07:00
|
|
|
|
|
|
|
while (bytes) {
|
2010-09-10 22:30:49 +07:00
|
|
|
gpa_t gpa = vcpu->arch.walk_mmu->gva_to_gpa(vcpu, addr, access,
|
2010-11-22 22:53:26 +07:00
|
|
|
exception);
|
KVM: Portability: Move x86 emulation and mmio device hook to x86.c
This patch moves the following functions to from kvm_main.c to x86.c:
emulator_read/write_std, vcpu_find_pervcpu_dev, vcpu_find_mmio_dev,
emulator_read/write_emulated, emulator_write_phys,
emulator_write_emulated_onepage, emulator_cmpxchg_emulated,
get_setment_base, emulate_invlpg, emulate_clts, emulator_get/set_dr,
kvm_report_emulation_failure, emulate_instruction
The following data type is moved to x86.c:
struct x86_emulate_ops emulate_ops
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-31 00:44:21 +07:00
|
|
|
unsigned offset = addr & (PAGE_SIZE-1);
|
2008-12-29 06:42:19 +07:00
|
|
|
unsigned toread = min(bytes, (unsigned)PAGE_SIZE - offset);
|
KVM: Portability: Move x86 emulation and mmio device hook to x86.c
This patch moves the following functions to from kvm_main.c to x86.c:
emulator_read/write_std, vcpu_find_pervcpu_dev, vcpu_find_mmio_dev,
emulator_read/write_emulated, emulator_write_phys,
emulator_write_emulated_onepage, emulator_cmpxchg_emulated,
get_setment_base, emulate_invlpg, emulate_clts, emulator_get/set_dr,
kvm_report_emulation_failure, emulate_instruction
The following data type is moved to x86.c:
struct x86_emulate_ops emulate_ops
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-31 00:44:21 +07:00
|
|
|
int ret;
|
|
|
|
|
2010-11-22 22:53:22 +07:00
|
|
|
if (gpa == UNMAPPED_GVA)
|
2010-11-22 22:53:26 +07:00
|
|
|
return X86EMUL_PROPAGATE_FAULT;
|
2015-04-08 20:39:23 +07:00
|
|
|
ret = kvm_vcpu_read_guest_page(vcpu, gpa >> PAGE_SHIFT, data,
|
|
|
|
offset, toread);
|
2007-12-21 07:18:22 +07:00
|
|
|
if (ret < 0) {
|
2010-04-28 23:15:35 +07:00
|
|
|
r = X86EMUL_IO_NEEDED;
|
2007-12-21 07:18:22 +07:00
|
|
|
goto out;
|
|
|
|
}
|
KVM: Portability: Move x86 emulation and mmio device hook to x86.c
This patch moves the following functions to from kvm_main.c to x86.c:
emulator_read/write_std, vcpu_find_pervcpu_dev, vcpu_find_mmio_dev,
emulator_read/write_emulated, emulator_write_phys,
emulator_write_emulated_onepage, emulator_cmpxchg_emulated,
get_setment_base, emulate_invlpg, emulate_clts, emulator_get/set_dr,
kvm_report_emulation_failure, emulate_instruction
The following data type is moved to x86.c:
struct x86_emulate_ops emulate_ops
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-31 00:44:21 +07:00
|
|
|
|
2008-12-29 06:42:19 +07:00
|
|
|
bytes -= toread;
|
|
|
|
data += toread;
|
|
|
|
addr += toread;
|
KVM: Portability: Move x86 emulation and mmio device hook to x86.c
This patch moves the following functions to from kvm_main.c to x86.c:
emulator_read/write_std, vcpu_find_pervcpu_dev, vcpu_find_mmio_dev,
emulator_read/write_emulated, emulator_write_phys,
emulator_write_emulated_onepage, emulator_cmpxchg_emulated,
get_setment_base, emulate_invlpg, emulate_clts, emulator_get/set_dr,
kvm_report_emulation_failure, emulate_instruction
The following data type is moved to x86.c:
struct x86_emulate_ops emulate_ops
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-31 00:44:21 +07:00
|
|
|
}
|
2007-12-21 07:18:22 +07:00
|
|
|
out:
|
|
|
|
return r;
|
KVM: Portability: Move x86 emulation and mmio device hook to x86.c
This patch moves the following functions to from kvm_main.c to x86.c:
emulator_read/write_std, vcpu_find_pervcpu_dev, vcpu_find_mmio_dev,
emulator_read/write_emulated, emulator_write_phys,
emulator_write_emulated_onepage, emulator_cmpxchg_emulated,
get_setment_base, emulate_invlpg, emulate_clts, emulator_get/set_dr,
kvm_report_emulation_failure, emulate_instruction
The following data type is moved to x86.c:
struct x86_emulate_ops emulate_ops
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-31 00:44:21 +07:00
|
|
|
}
|
2008-12-29 06:42:19 +07:00
|
|
|
|
2010-02-10 19:21:32 +07:00
|
|
|
/* used for instruction fetching */
|
2011-04-20 17:37:53 +07:00
|
|
|
static int kvm_fetch_guest_virt(struct x86_emulate_ctxt *ctxt,
|
|
|
|
gva_t addr, void *val, unsigned int bytes,
|
2010-11-22 22:53:22 +07:00
|
|
|
struct x86_exception *exception)
|
2010-02-10 19:21:32 +07:00
|
|
|
{
|
2011-04-20 17:37:53 +07:00
|
|
|
struct kvm_vcpu *vcpu = emul_to_vcpu(ctxt);
|
2010-02-10 19:21:32 +07:00
|
|
|
u32 access = (kvm_x86_ops->get_cpl(vcpu) == 3) ? PFERR_USER_MASK : 0;
|
2014-05-13 19:02:13 +07:00
|
|
|
unsigned offset;
|
|
|
|
int ret;
|
2011-04-20 17:37:53 +07:00
|
|
|
|
2014-05-13 19:02:13 +07:00
|
|
|
/* Inline kvm_read_guest_virt_helper for speed. */
|
|
|
|
gpa_t gpa = vcpu->arch.walk_mmu->gva_to_gpa(vcpu, addr, access|PFERR_FETCH_MASK,
|
|
|
|
exception);
|
|
|
|
if (unlikely(gpa == UNMAPPED_GVA))
|
|
|
|
return X86EMUL_PROPAGATE_FAULT;
|
|
|
|
|
|
|
|
offset = addr & (PAGE_SIZE-1);
|
|
|
|
if (WARN_ON(offset + bytes > PAGE_SIZE))
|
|
|
|
bytes = (unsigned)PAGE_SIZE - offset;
|
2015-04-08 20:39:23 +07:00
|
|
|
ret = kvm_vcpu_read_guest_page(vcpu, gpa >> PAGE_SHIFT, val,
|
|
|
|
offset, bytes);
|
2014-05-13 19:02:13 +07:00
|
|
|
if (unlikely(ret < 0))
|
|
|
|
return X86EMUL_IO_NEEDED;
|
|
|
|
|
|
|
|
return X86EMUL_CONTINUE;
|
2010-02-10 19:21:32 +07:00
|
|
|
}
|
|
|
|
|
2011-05-26 03:04:56 +07:00
|
|
|
int kvm_read_guest_virt(struct x86_emulate_ctxt *ctxt,
|
2011-04-20 17:37:53 +07:00
|
|
|
gva_t addr, void *val, unsigned int bytes,
|
2010-11-22 22:53:22 +07:00
|
|
|
struct x86_exception *exception)
|
2010-02-10 19:21:32 +07:00
|
|
|
{
|
2011-04-20 17:37:53 +07:00
|
|
|
struct kvm_vcpu *vcpu = emul_to_vcpu(ctxt);
|
2010-02-10 19:21:32 +07:00
|
|
|
u32 access = (kvm_x86_ops->get_cpl(vcpu) == 3) ? PFERR_USER_MASK : 0;
|
2011-04-20 17:37:53 +07:00
|
|
|
|
2010-02-10 19:21:32 +07:00
|
|
|
return kvm_read_guest_virt_helper(addr, val, bytes, vcpu, access,
|
2010-11-22 22:53:22 +07:00
|
|
|
exception);
|
2010-02-10 19:21:32 +07:00
|
|
|
}
|
2011-05-26 03:04:56 +07:00
|
|
|
EXPORT_SYMBOL_GPL(kvm_read_guest_virt);
|
2010-02-10 19:21:32 +07:00
|
|
|
|
2011-04-20 17:37:53 +07:00
|
|
|
static int kvm_read_guest_virt_system(struct x86_emulate_ctxt *ctxt,
|
|
|
|
gva_t addr, void *val, unsigned int bytes,
|
2010-11-22 22:53:22 +07:00
|
|
|
struct x86_exception *exception)
|
2010-02-10 19:21:32 +07:00
|
|
|
{
|
2011-04-20 17:37:53 +07:00
|
|
|
struct kvm_vcpu *vcpu = emul_to_vcpu(ctxt);
|
2010-11-22 22:53:22 +07:00
|
|
|
return kvm_read_guest_virt_helper(addr, val, bytes, vcpu, 0, exception);
|
2010-02-10 19:21:32 +07:00
|
|
|
}
|
|
|
|
|
2015-10-30 22:36:24 +07:00
|
|
|
static int kvm_read_guest_phys_system(struct x86_emulate_ctxt *ctxt,
|
|
|
|
unsigned long addr, void *val, unsigned int bytes)
|
|
|
|
{
|
|
|
|
struct kvm_vcpu *vcpu = emul_to_vcpu(ctxt);
|
|
|
|
int r = kvm_vcpu_read_guest(vcpu, addr, val, bytes);
|
|
|
|
|
|
|
|
return r < 0 ? X86EMUL_IO_NEEDED : X86EMUL_CONTINUE;
|
|
|
|
}
|
|
|
|
|
2011-05-26 03:08:00 +07:00
|
|
|
int kvm_write_guest_virt_system(struct x86_emulate_ctxt *ctxt,
|
2011-04-20 17:37:53 +07:00
|
|
|
gva_t addr, void *val,
|
2010-03-18 20:20:16 +07:00
|
|
|
unsigned int bytes,
|
2010-11-22 22:53:22 +07:00
|
|
|
struct x86_exception *exception)
|
2008-12-29 06:42:19 +07:00
|
|
|
{
|
2011-04-20 17:37:53 +07:00
|
|
|
struct kvm_vcpu *vcpu = emul_to_vcpu(ctxt);
|
2008-12-29 06:42:19 +07:00
|
|
|
void *data = val;
|
|
|
|
int r = X86EMUL_CONTINUE;
|
|
|
|
|
|
|
|
while (bytes) {
|
2010-09-10 22:30:49 +07:00
|
|
|
gpa_t gpa = vcpu->arch.walk_mmu->gva_to_gpa(vcpu, addr,
|
|
|
|
PFERR_WRITE_MASK,
|
2010-11-22 22:53:26 +07:00
|
|
|
exception);
|
2008-12-29 06:42:19 +07:00
|
|
|
unsigned offset = addr & (PAGE_SIZE-1);
|
|
|
|
unsigned towrite = min(bytes, (unsigned)PAGE_SIZE - offset);
|
|
|
|
int ret;
|
|
|
|
|
2010-11-22 22:53:22 +07:00
|
|
|
if (gpa == UNMAPPED_GVA)
|
2010-11-22 22:53:26 +07:00
|
|
|
return X86EMUL_PROPAGATE_FAULT;
|
2015-04-08 20:39:23 +07:00
|
|
|
ret = kvm_vcpu_write_guest(vcpu, gpa, data, towrite);
|
2008-12-29 06:42:19 +07:00
|
|
|
if (ret < 0) {
|
2010-04-28 23:15:35 +07:00
|
|
|
r = X86EMUL_IO_NEEDED;
|
2008-12-29 06:42:19 +07:00
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
|
|
|
bytes -= towrite;
|
|
|
|
data += towrite;
|
|
|
|
addr += towrite;
|
|
|
|
}
|
|
|
|
out:
|
|
|
|
return r;
|
|
|
|
}
|
2011-05-26 03:08:00 +07:00
|
|
|
EXPORT_SYMBOL_GPL(kvm_write_guest_virt_system);
|
2008-12-29 06:42:19 +07:00
|
|
|
|
2011-07-12 02:22:46 +07:00
|
|
|
static int vcpu_mmio_gva_to_gpa(struct kvm_vcpu *vcpu, unsigned long gva,
|
|
|
|
gpa_t *gpa, struct x86_exception *exception,
|
|
|
|
bool write)
|
|
|
|
{
|
KVM: MMU: Optimize pte permission checks
walk_addr_generic() permission checks are a maze of branchy code, which is
performed four times per lookup. It depends on the type of access, efer.nxe,
cr0.wp, cr4.smep, and in the near future, cr4.smap.
Optimize this away by precalculating all variants and storing them in a
bitmap. The bitmap is recalculated when rarely-changing variables change
(cr0, cr4) and is indexed by the often-changing variables (page fault error
code, pte access permissions).
The permission check is moved to the end of the loop, otherwise an SMEP
fault could be reported as a false positive, when PDE.U=1 but PTE.U=0.
Noted by Xiao Guangrong.
The result is short, branch-free code.
Reviewed-by: Xiao Guangrong <xiaoguangrong@linux.vnet.ibm.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
2012-09-12 18:52:00 +07:00
|
|
|
u32 access = ((kvm_x86_ops->get_cpl(vcpu) == 3) ? PFERR_USER_MASK : 0)
|
|
|
|
| (write ? PFERR_WRITE_MASK : 0);
|
2011-07-12 02:22:46 +07:00
|
|
|
|
2016-03-22 15:51:20 +07:00
|
|
|
/*
|
|
|
|
* currently PKRU is only applied to ept enabled guest so
|
|
|
|
* there is no pkey in EPT page table for L1 guest or EPT
|
|
|
|
* shadow page table for L2 guest.
|
|
|
|
*/
|
KVM: MMU: Optimize pte permission checks
walk_addr_generic() permission checks are a maze of branchy code, which is
performed four times per lookup. It depends on the type of access, efer.nxe,
cr0.wp, cr4.smep, and in the near future, cr4.smap.
Optimize this away by precalculating all variants and storing them in a
bitmap. The bitmap is recalculated when rarely-changing variables change
(cr0, cr4) and is indexed by the often-changing variables (page fault error
code, pte access permissions).
The permission check is moved to the end of the loop, otherwise an SMEP
fault could be reported as a false positive, when PDE.U=1 but PTE.U=0.
Noted by Xiao Guangrong.
The result is short, branch-free code.
Reviewed-by: Xiao Guangrong <xiaoguangrong@linux.vnet.ibm.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
2012-09-12 18:52:00 +07:00
|
|
|
if (vcpu_match_mmio_gva(vcpu, gva)
|
2014-04-01 16:46:34 +07:00
|
|
|
&& !permission_fault(vcpu, vcpu->arch.walk_mmu,
|
2016-03-22 15:51:20 +07:00
|
|
|
vcpu->arch.access, 0, access)) {
|
2011-07-12 02:23:20 +07:00
|
|
|
*gpa = vcpu->arch.mmio_gfn << PAGE_SHIFT |
|
|
|
|
(gva & (PAGE_SIZE - 1));
|
2011-07-12 02:34:24 +07:00
|
|
|
trace_vcpu_match_mmio(gva, *gpa, write, false);
|
2011-07-12 02:23:20 +07:00
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
|
2011-07-12 02:22:46 +07:00
|
|
|
*gpa = vcpu->arch.walk_mmu->gva_to_gpa(vcpu, gva, access, exception);
|
|
|
|
|
|
|
|
if (*gpa == UNMAPPED_GVA)
|
|
|
|
return -1;
|
|
|
|
|
|
|
|
/* For APIC access vmexit */
|
|
|
|
if ((*gpa & PAGE_MASK) == APIC_DEFAULT_PHYS_BASE)
|
|
|
|
return 1;
|
|
|
|
|
2011-07-12 02:34:24 +07:00
|
|
|
if (vcpu_match_mmio_gpa(vcpu, *gpa)) {
|
|
|
|
trace_vcpu_match_mmio(gva, *gpa, write, true);
|
2011-07-12 02:23:20 +07:00
|
|
|
return 1;
|
2011-07-12 02:34:24 +07:00
|
|
|
}
|
2011-07-12 02:23:20 +07:00
|
|
|
|
2011-07-12 02:22:46 +07:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2008-03-30 06:17:59 +07:00
|
|
|
int emulator_write_phys(struct kvm_vcpu *vcpu, gpa_t gpa,
|
2010-11-22 22:53:22 +07:00
|
|
|
const void *val, int bytes)
|
KVM: Portability: Move x86 emulation and mmio device hook to x86.c
This patch moves the following functions to from kvm_main.c to x86.c:
emulator_read/write_std, vcpu_find_pervcpu_dev, vcpu_find_mmio_dev,
emulator_read/write_emulated, emulator_write_phys,
emulator_write_emulated_onepage, emulator_cmpxchg_emulated,
get_setment_base, emulate_invlpg, emulate_clts, emulator_get/set_dr,
kvm_report_emulation_failure, emulate_instruction
The following data type is moved to x86.c:
struct x86_emulate_ops emulate_ops
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-31 00:44:21 +07:00
|
|
|
{
|
|
|
|
int ret;
|
|
|
|
|
2015-04-08 20:39:23 +07:00
|
|
|
ret = kvm_vcpu_write_guest(vcpu, gpa, val, bytes);
|
2008-03-02 19:06:05 +07:00
|
|
|
if (ret < 0)
|
KVM: Portability: Move x86 emulation and mmio device hook to x86.c
This patch moves the following functions to from kvm_main.c to x86.c:
emulator_read/write_std, vcpu_find_pervcpu_dev, vcpu_find_mmio_dev,
emulator_read/write_emulated, emulator_write_phys,
emulator_write_emulated_onepage, emulator_cmpxchg_emulated,
get_setment_base, emulate_invlpg, emulate_clts, emulator_get/set_dr,
kvm_report_emulation_failure, emulate_instruction
The following data type is moved to x86.c:
struct x86_emulate_ops emulate_ops
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-31 00:44:21 +07:00
|
|
|
return 0;
|
2016-02-24 16:51:13 +07:00
|
|
|
kvm_page_track_write(vcpu, gpa, val, bytes);
|
KVM: Portability: Move x86 emulation and mmio device hook to x86.c
This patch moves the following functions to from kvm_main.c to x86.c:
emulator_read/write_std, vcpu_find_pervcpu_dev, vcpu_find_mmio_dev,
emulator_read/write_emulated, emulator_write_phys,
emulator_write_emulated_onepage, emulator_cmpxchg_emulated,
get_setment_base, emulate_invlpg, emulate_clts, emulator_get/set_dr,
kvm_report_emulation_failure, emulate_instruction
The following data type is moved to x86.c:
struct x86_emulate_ops emulate_ops
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-31 00:44:21 +07:00
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
|
2011-07-13 13:31:50 +07:00
|
|
|
struct read_write_emulator_ops {
|
|
|
|
int (*read_write_prepare)(struct kvm_vcpu *vcpu, void *val,
|
|
|
|
int bytes);
|
|
|
|
int (*read_write_emulate)(struct kvm_vcpu *vcpu, gpa_t gpa,
|
|
|
|
void *val, int bytes);
|
|
|
|
int (*read_write_mmio)(struct kvm_vcpu *vcpu, gpa_t gpa,
|
|
|
|
int bytes, void *val);
|
|
|
|
int (*read_write_exit_mmio)(struct kvm_vcpu *vcpu, gpa_t gpa,
|
|
|
|
void *val, int bytes);
|
|
|
|
bool write;
|
|
|
|
};
|
|
|
|
|
|
|
|
static int read_prepare(struct kvm_vcpu *vcpu, void *val, int bytes)
|
|
|
|
{
|
|
|
|
if (vcpu->mmio_read_completed) {
|
|
|
|
trace_kvm_mmio(KVM_TRACE_MMIO_READ, bytes,
|
2012-04-18 23:22:47 +07:00
|
|
|
vcpu->mmio_fragments[0].gpa, *(u64 *)val);
|
2011-07-13 13:31:50 +07:00
|
|
|
vcpu->mmio_read_completed = 0;
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int read_emulate(struct kvm_vcpu *vcpu, gpa_t gpa,
|
|
|
|
void *val, int bytes)
|
|
|
|
{
|
2015-04-08 20:39:23 +07:00
|
|
|
return !kvm_vcpu_read_guest(vcpu, gpa, val, bytes);
|
2011-07-13 13:31:50 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
static int write_emulate(struct kvm_vcpu *vcpu, gpa_t gpa,
|
|
|
|
void *val, int bytes)
|
|
|
|
{
|
|
|
|
return emulator_write_phys(vcpu, gpa, val, bytes);
|
|
|
|
}
|
|
|
|
|
|
|
|
static int write_mmio(struct kvm_vcpu *vcpu, gpa_t gpa, int bytes, void *val)
|
|
|
|
{
|
|
|
|
trace_kvm_mmio(KVM_TRACE_MMIO_WRITE, bytes, gpa, *(u64 *)val);
|
|
|
|
return vcpu_mmio_write(vcpu, gpa, bytes, val);
|
|
|
|
}
|
|
|
|
|
|
|
|
static int read_exit_mmio(struct kvm_vcpu *vcpu, gpa_t gpa,
|
|
|
|
void *val, int bytes)
|
|
|
|
{
|
|
|
|
trace_kvm_mmio(KVM_TRACE_MMIO_READ_UNSATISFIED, bytes, gpa, 0);
|
|
|
|
return X86EMUL_IO_NEEDED;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int write_exit_mmio(struct kvm_vcpu *vcpu, gpa_t gpa,
|
|
|
|
void *val, int bytes)
|
|
|
|
{
|
2012-04-18 23:22:47 +07:00
|
|
|
struct kvm_mmio_fragment *frag = &vcpu->mmio_fragments[0];
|
|
|
|
|
2012-10-24 13:07:59 +07:00
|
|
|
memcpy(vcpu->run->mmio.data, frag->data, min(8u, frag->len));
|
2011-07-13 13:31:50 +07:00
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
}
|
|
|
|
|
2012-08-30 06:30:17 +07:00
|
|
|
static const struct read_write_emulator_ops read_emultor = {
|
2011-07-13 13:31:50 +07:00
|
|
|
.read_write_prepare = read_prepare,
|
|
|
|
.read_write_emulate = read_emulate,
|
|
|
|
.read_write_mmio = vcpu_mmio_read,
|
|
|
|
.read_write_exit_mmio = read_exit_mmio,
|
|
|
|
};
|
|
|
|
|
2012-08-30 06:30:17 +07:00
|
|
|
static const struct read_write_emulator_ops write_emultor = {
|
2011-07-13 13:31:50 +07:00
|
|
|
.read_write_emulate = write_emulate,
|
|
|
|
.read_write_mmio = write_mmio,
|
|
|
|
.read_write_exit_mmio = write_exit_mmio,
|
|
|
|
.write = true,
|
|
|
|
};
|
|
|
|
|
2011-07-13 13:32:31 +07:00
|
|
|
static int emulator_read_write_onepage(unsigned long addr, void *val,
|
|
|
|
unsigned int bytes,
|
|
|
|
struct x86_exception *exception,
|
|
|
|
struct kvm_vcpu *vcpu,
|
2012-08-30 06:30:17 +07:00
|
|
|
const struct read_write_emulator_ops *ops)
|
KVM: Portability: Move x86 emulation and mmio device hook to x86.c
This patch moves the following functions to from kvm_main.c to x86.c:
emulator_read/write_std, vcpu_find_pervcpu_dev, vcpu_find_mmio_dev,
emulator_read/write_emulated, emulator_write_phys,
emulator_write_emulated_onepage, emulator_cmpxchg_emulated,
get_setment_base, emulate_invlpg, emulate_clts, emulator_get/set_dr,
kvm_report_emulation_failure, emulate_instruction
The following data type is moved to x86.c:
struct x86_emulate_ops emulate_ops
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-31 00:44:21 +07:00
|
|
|
{
|
2011-07-12 02:22:46 +07:00
|
|
|
gpa_t gpa;
|
|
|
|
int handled, ret;
|
2011-07-13 13:32:31 +07:00
|
|
|
bool write = ops->write;
|
2012-04-18 23:22:47 +07:00
|
|
|
struct kvm_mmio_fragment *frag;
|
2007-12-21 07:18:22 +07:00
|
|
|
|
2011-07-13 13:32:31 +07:00
|
|
|
ret = vcpu_mmio_gva_to_gpa(vcpu, addr, &gpa, exception, write);
|
KVM: Portability: Move x86 emulation and mmio device hook to x86.c
This patch moves the following functions to from kvm_main.c to x86.c:
emulator_read/write_std, vcpu_find_pervcpu_dev, vcpu_find_mmio_dev,
emulator_read/write_emulated, emulator_write_phys,
emulator_write_emulated_onepage, emulator_cmpxchg_emulated,
get_setment_base, emulate_invlpg, emulate_clts, emulator_get/set_dr,
kvm_report_emulation_failure, emulate_instruction
The following data type is moved to x86.c:
struct x86_emulate_ops emulate_ops
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-31 00:44:21 +07:00
|
|
|
|
2011-07-12 02:22:46 +07:00
|
|
|
if (ret < 0)
|
KVM: Portability: Move x86 emulation and mmio device hook to x86.c
This patch moves the following functions to from kvm_main.c to x86.c:
emulator_read/write_std, vcpu_find_pervcpu_dev, vcpu_find_mmio_dev,
emulator_read/write_emulated, emulator_write_phys,
emulator_write_emulated_onepage, emulator_cmpxchg_emulated,
get_setment_base, emulate_invlpg, emulate_clts, emulator_get/set_dr,
kvm_report_emulation_failure, emulate_instruction
The following data type is moved to x86.c:
struct x86_emulate_ops emulate_ops
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-31 00:44:21 +07:00
|
|
|
return X86EMUL_PROPAGATE_FAULT;
|
|
|
|
|
|
|
|
/* For APIC access vmexit */
|
2011-07-12 02:22:46 +07:00
|
|
|
if (ret)
|
KVM: Portability: Move x86 emulation and mmio device hook to x86.c
This patch moves the following functions to from kvm_main.c to x86.c:
emulator_read/write_std, vcpu_find_pervcpu_dev, vcpu_find_mmio_dev,
emulator_read/write_emulated, emulator_write_phys,
emulator_write_emulated_onepage, emulator_cmpxchg_emulated,
get_setment_base, emulate_invlpg, emulate_clts, emulator_get/set_dr,
kvm_report_emulation_failure, emulate_instruction
The following data type is moved to x86.c:
struct x86_emulate_ops emulate_ops
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-31 00:44:21 +07:00
|
|
|
goto mmio;
|
|
|
|
|
2011-07-13 13:32:31 +07:00
|
|
|
if (ops->read_write_emulate(vcpu, gpa, val, bytes))
|
KVM: Portability: Move x86 emulation and mmio device hook to x86.c
This patch moves the following functions to from kvm_main.c to x86.c:
emulator_read/write_std, vcpu_find_pervcpu_dev, vcpu_find_mmio_dev,
emulator_read/write_emulated, emulator_write_phys,
emulator_write_emulated_onepage, emulator_cmpxchg_emulated,
get_setment_base, emulate_invlpg, emulate_clts, emulator_get/set_dr,
kvm_report_emulation_failure, emulate_instruction
The following data type is moved to x86.c:
struct x86_emulate_ops emulate_ops
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-31 00:44:21 +07:00
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
|
|
|
mmio:
|
|
|
|
/*
|
|
|
|
* Is this MMIO handled locally?
|
|
|
|
*/
|
2011-07-13 13:32:31 +07:00
|
|
|
handled = ops->read_write_mmio(vcpu, gpa, bytes, val);
|
2010-01-19 17:51:22 +07:00
|
|
|
if (handled == bytes)
|
KVM: Portability: Move x86 emulation and mmio device hook to x86.c
This patch moves the following functions to from kvm_main.c to x86.c:
emulator_read/write_std, vcpu_find_pervcpu_dev, vcpu_find_mmio_dev,
emulator_read/write_emulated, emulator_write_phys,
emulator_write_emulated_onepage, emulator_cmpxchg_emulated,
get_setment_base, emulate_invlpg, emulate_clts, emulator_get/set_dr,
kvm_report_emulation_failure, emulate_instruction
The following data type is moved to x86.c:
struct x86_emulate_ops emulate_ops
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-31 00:44:21 +07:00
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
2010-01-19 17:51:22 +07:00
|
|
|
gpa += handled;
|
|
|
|
bytes -= handled;
|
|
|
|
val += handled;
|
|
|
|
|
2012-10-24 13:07:59 +07:00
|
|
|
WARN_ON(vcpu->mmio_nr_fragments >= KVM_MAX_MMIO_FRAGMENTS);
|
|
|
|
frag = &vcpu->mmio_fragments[vcpu->mmio_nr_fragments++];
|
|
|
|
frag->gpa = gpa;
|
|
|
|
frag->data = val;
|
|
|
|
frag->len = bytes;
|
2012-04-18 23:22:47 +07:00
|
|
|
return X86EMUL_CONTINUE;
|
KVM: Portability: Move x86 emulation and mmio device hook to x86.c
This patch moves the following functions to from kvm_main.c to x86.c:
emulator_read/write_std, vcpu_find_pervcpu_dev, vcpu_find_mmio_dev,
emulator_read/write_emulated, emulator_write_phys,
emulator_write_emulated_onepage, emulator_cmpxchg_emulated,
get_setment_base, emulate_invlpg, emulate_clts, emulator_get/set_dr,
kvm_report_emulation_failure, emulate_instruction
The following data type is moved to x86.c:
struct x86_emulate_ops emulate_ops
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-31 00:44:21 +07:00
|
|
|
}
|
|
|
|
|
2015-03-13 16:39:45 +07:00
|
|
|
static int emulator_read_write(struct x86_emulate_ctxt *ctxt,
|
|
|
|
unsigned long addr,
|
2011-07-13 13:32:31 +07:00
|
|
|
void *val, unsigned int bytes,
|
|
|
|
struct x86_exception *exception,
|
2012-08-30 06:30:17 +07:00
|
|
|
const struct read_write_emulator_ops *ops)
|
KVM: Portability: Move x86 emulation and mmio device hook to x86.c
This patch moves the following functions to from kvm_main.c to x86.c:
emulator_read/write_std, vcpu_find_pervcpu_dev, vcpu_find_mmio_dev,
emulator_read/write_emulated, emulator_write_phys,
emulator_write_emulated_onepage, emulator_cmpxchg_emulated,
get_setment_base, emulate_invlpg, emulate_clts, emulator_get/set_dr,
kvm_report_emulation_failure, emulate_instruction
The following data type is moved to x86.c:
struct x86_emulate_ops emulate_ops
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-31 00:44:21 +07:00
|
|
|
{
|
2011-04-20 17:37:53 +07:00
|
|
|
struct kvm_vcpu *vcpu = emul_to_vcpu(ctxt);
|
2012-04-18 23:22:47 +07:00
|
|
|
gpa_t gpa;
|
|
|
|
int rc;
|
|
|
|
|
|
|
|
if (ops->read_write_prepare &&
|
|
|
|
ops->read_write_prepare(vcpu, val, bytes))
|
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
|
|
|
vcpu->mmio_nr_fragments = 0;
|
2011-04-20 17:37:53 +07:00
|
|
|
|
KVM: Portability: Move x86 emulation and mmio device hook to x86.c
This patch moves the following functions to from kvm_main.c to x86.c:
emulator_read/write_std, vcpu_find_pervcpu_dev, vcpu_find_mmio_dev,
emulator_read/write_emulated, emulator_write_phys,
emulator_write_emulated_onepage, emulator_cmpxchg_emulated,
get_setment_base, emulate_invlpg, emulate_clts, emulator_get/set_dr,
kvm_report_emulation_failure, emulate_instruction
The following data type is moved to x86.c:
struct x86_emulate_ops emulate_ops
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-31 00:44:21 +07:00
|
|
|
/* Crossing a page boundary? */
|
|
|
|
if (((addr + bytes - 1) ^ addr) & PAGE_MASK) {
|
2012-04-18 23:22:47 +07:00
|
|
|
int now;
|
KVM: Portability: Move x86 emulation and mmio device hook to x86.c
This patch moves the following functions to from kvm_main.c to x86.c:
emulator_read/write_std, vcpu_find_pervcpu_dev, vcpu_find_mmio_dev,
emulator_read/write_emulated, emulator_write_phys,
emulator_write_emulated_onepage, emulator_cmpxchg_emulated,
get_setment_base, emulate_invlpg, emulate_clts, emulator_get/set_dr,
kvm_report_emulation_failure, emulate_instruction
The following data type is moved to x86.c:
struct x86_emulate_ops emulate_ops
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-31 00:44:21 +07:00
|
|
|
|
|
|
|
now = -addr & ~PAGE_MASK;
|
2011-07-13 13:32:31 +07:00
|
|
|
rc = emulator_read_write_onepage(addr, val, now, exception,
|
|
|
|
vcpu, ops);
|
|
|
|
|
KVM: Portability: Move x86 emulation and mmio device hook to x86.c
This patch moves the following functions to from kvm_main.c to x86.c:
emulator_read/write_std, vcpu_find_pervcpu_dev, vcpu_find_mmio_dev,
emulator_read/write_emulated, emulator_write_phys,
emulator_write_emulated_onepage, emulator_cmpxchg_emulated,
get_setment_base, emulate_invlpg, emulate_clts, emulator_get/set_dr,
kvm_report_emulation_failure, emulate_instruction
The following data type is moved to x86.c:
struct x86_emulate_ops emulate_ops
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-31 00:44:21 +07:00
|
|
|
if (rc != X86EMUL_CONTINUE)
|
|
|
|
return rc;
|
|
|
|
addr += now;
|
2015-01-26 14:32:26 +07:00
|
|
|
if (ctxt->mode != X86EMUL_MODE_PROT64)
|
|
|
|
addr = (u32)addr;
|
KVM: Portability: Move x86 emulation and mmio device hook to x86.c
This patch moves the following functions to from kvm_main.c to x86.c:
emulator_read/write_std, vcpu_find_pervcpu_dev, vcpu_find_mmio_dev,
emulator_read/write_emulated, emulator_write_phys,
emulator_write_emulated_onepage, emulator_cmpxchg_emulated,
get_setment_base, emulate_invlpg, emulate_clts, emulator_get/set_dr,
kvm_report_emulation_failure, emulate_instruction
The following data type is moved to x86.c:
struct x86_emulate_ops emulate_ops
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-31 00:44:21 +07:00
|
|
|
val += now;
|
|
|
|
bytes -= now;
|
|
|
|
}
|
2011-07-13 13:32:31 +07:00
|
|
|
|
2012-04-18 23:22:47 +07:00
|
|
|
rc = emulator_read_write_onepage(addr, val, bytes, exception,
|
|
|
|
vcpu, ops);
|
|
|
|
if (rc != X86EMUL_CONTINUE)
|
|
|
|
return rc;
|
|
|
|
|
|
|
|
if (!vcpu->mmio_nr_fragments)
|
|
|
|
return rc;
|
|
|
|
|
|
|
|
gpa = vcpu->mmio_fragments[0].gpa;
|
|
|
|
|
|
|
|
vcpu->mmio_needed = 1;
|
|
|
|
vcpu->mmio_cur_fragment = 0;
|
|
|
|
|
2012-10-24 13:07:59 +07:00
|
|
|
vcpu->run->mmio.len = min(8u, vcpu->mmio_fragments[0].len);
|
2012-04-18 23:22:47 +07:00
|
|
|
vcpu->run->mmio.is_write = vcpu->mmio_is_write = ops->write;
|
|
|
|
vcpu->run->exit_reason = KVM_EXIT_MMIO;
|
|
|
|
vcpu->run->mmio.phys_addr = gpa;
|
|
|
|
|
|
|
|
return ops->read_write_exit_mmio(vcpu, gpa, val, bytes);
|
2011-07-13 13:32:31 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
static int emulator_read_emulated(struct x86_emulate_ctxt *ctxt,
|
|
|
|
unsigned long addr,
|
|
|
|
void *val,
|
|
|
|
unsigned int bytes,
|
|
|
|
struct x86_exception *exception)
|
|
|
|
{
|
|
|
|
return emulator_read_write(ctxt, addr, val, bytes,
|
|
|
|
exception, &read_emultor);
|
|
|
|
}
|
|
|
|
|
2015-03-13 16:39:45 +07:00
|
|
|
static int emulator_write_emulated(struct x86_emulate_ctxt *ctxt,
|
2011-07-13 13:32:31 +07:00
|
|
|
unsigned long addr,
|
|
|
|
const void *val,
|
|
|
|
unsigned int bytes,
|
|
|
|
struct x86_exception *exception)
|
|
|
|
{
|
|
|
|
return emulator_read_write(ctxt, addr, (void *)val, bytes,
|
|
|
|
exception, &write_emultor);
|
KVM: Portability: Move x86 emulation and mmio device hook to x86.c
This patch moves the following functions to from kvm_main.c to x86.c:
emulator_read/write_std, vcpu_find_pervcpu_dev, vcpu_find_mmio_dev,
emulator_read/write_emulated, emulator_write_phys,
emulator_write_emulated_onepage, emulator_cmpxchg_emulated,
get_setment_base, emulate_invlpg, emulate_clts, emulator_get/set_dr,
kvm_report_emulation_failure, emulate_instruction
The following data type is moved to x86.c:
struct x86_emulate_ops emulate_ops
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-31 00:44:21 +07:00
|
|
|
}
|
|
|
|
|
2010-03-15 18:59:54 +07:00
|
|
|
#define CMPXCHG_TYPE(t, ptr, old, new) \
|
|
|
|
(cmpxchg((t *)(ptr), *(t *)(old), *(t *)(new)) == *(t *)(old))
|
|
|
|
|
|
|
|
#ifdef CONFIG_X86_64
|
|
|
|
# define CMPXCHG64(ptr, old, new) CMPXCHG_TYPE(u64, ptr, old, new)
|
|
|
|
#else
|
|
|
|
# define CMPXCHG64(ptr, old, new) \
|
2010-03-20 16:14:13 +07:00
|
|
|
(cmpxchg64((u64 *)(ptr), *(u64 *)(old), *(u64 *)(new)) == *(u64 *)(old))
|
2010-03-15 18:59:54 +07:00
|
|
|
#endif
|
|
|
|
|
2011-04-20 17:37:53 +07:00
|
|
|
static int emulator_cmpxchg_emulated(struct x86_emulate_ctxt *ctxt,
|
|
|
|
unsigned long addr,
|
KVM: Portability: Move x86 emulation and mmio device hook to x86.c
This patch moves the following functions to from kvm_main.c to x86.c:
emulator_read/write_std, vcpu_find_pervcpu_dev, vcpu_find_mmio_dev,
emulator_read/write_emulated, emulator_write_phys,
emulator_write_emulated_onepage, emulator_cmpxchg_emulated,
get_setment_base, emulate_invlpg, emulate_clts, emulator_get/set_dr,
kvm_report_emulation_failure, emulate_instruction
The following data type is moved to x86.c:
struct x86_emulate_ops emulate_ops
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-31 00:44:21 +07:00
|
|
|
const void *old,
|
|
|
|
const void *new,
|
|
|
|
unsigned int bytes,
|
2011-04-20 17:37:53 +07:00
|
|
|
struct x86_exception *exception)
|
KVM: Portability: Move x86 emulation and mmio device hook to x86.c
This patch moves the following functions to from kvm_main.c to x86.c:
emulator_read/write_std, vcpu_find_pervcpu_dev, vcpu_find_mmio_dev,
emulator_read/write_emulated, emulator_write_phys,
emulator_write_emulated_onepage, emulator_cmpxchg_emulated,
get_setment_base, emulate_invlpg, emulate_clts, emulator_get/set_dr,
kvm_report_emulation_failure, emulate_instruction
The following data type is moved to x86.c:
struct x86_emulate_ops emulate_ops
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-31 00:44:21 +07:00
|
|
|
{
|
2011-04-20 17:37:53 +07:00
|
|
|
struct kvm_vcpu *vcpu = emul_to_vcpu(ctxt);
|
2010-03-15 18:59:54 +07:00
|
|
|
gpa_t gpa;
|
|
|
|
struct page *page;
|
|
|
|
char *kaddr;
|
|
|
|
bool exchanged;
|
2007-12-12 22:46:12 +07:00
|
|
|
|
2010-03-15 18:59:54 +07:00
|
|
|
/* guests cmpxchg8b have to be emulated atomically */
|
|
|
|
if (bytes > 8 || (bytes & (bytes - 1)))
|
|
|
|
goto emul_write;
|
2007-12-21 07:18:22 +07:00
|
|
|
|
2010-03-15 18:59:54 +07:00
|
|
|
gpa = kvm_mmu_gva_to_gpa_write(vcpu, addr, NULL);
|
2007-12-12 22:46:12 +07:00
|
|
|
|
2010-03-15 18:59:54 +07:00
|
|
|
if (gpa == UNMAPPED_GVA ||
|
|
|
|
(gpa & PAGE_MASK) == APIC_DEFAULT_PHYS_BASE)
|
|
|
|
goto emul_write;
|
2007-12-12 22:46:12 +07:00
|
|
|
|
2010-03-15 18:59:54 +07:00
|
|
|
if (((gpa + bytes - 1) & PAGE_MASK) != (gpa & PAGE_MASK))
|
|
|
|
goto emul_write;
|
2008-02-10 23:04:15 +07:00
|
|
|
|
2015-04-08 20:39:23 +07:00
|
|
|
page = kvm_vcpu_gfn_to_page(vcpu, gpa >> PAGE_SHIFT);
|
2012-08-03 14:42:52 +07:00
|
|
|
if (is_error_page(page))
|
2010-07-15 07:51:58 +07:00
|
|
|
goto emul_write;
|
2008-02-10 23:04:15 +07:00
|
|
|
|
2011-11-25 22:14:17 +07:00
|
|
|
kaddr = kmap_atomic(page);
|
2010-03-15 18:59:54 +07:00
|
|
|
kaddr += offset_in_page(gpa);
|
|
|
|
switch (bytes) {
|
|
|
|
case 1:
|
|
|
|
exchanged = CMPXCHG_TYPE(u8, kaddr, old, new);
|
|
|
|
break;
|
|
|
|
case 2:
|
|
|
|
exchanged = CMPXCHG_TYPE(u16, kaddr, old, new);
|
|
|
|
break;
|
|
|
|
case 4:
|
|
|
|
exchanged = CMPXCHG_TYPE(u32, kaddr, old, new);
|
|
|
|
break;
|
|
|
|
case 8:
|
|
|
|
exchanged = CMPXCHG64(kaddr, old, new);
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
BUG();
|
2007-12-12 22:46:12 +07:00
|
|
|
}
|
2011-11-25 22:14:17 +07:00
|
|
|
kunmap_atomic(kaddr);
|
2010-03-15 18:59:54 +07:00
|
|
|
kvm_release_page_dirty(page);
|
|
|
|
|
|
|
|
if (!exchanged)
|
|
|
|
return X86EMUL_CMPXCHG_FAILED;
|
|
|
|
|
2015-04-08 20:39:23 +07:00
|
|
|
kvm_vcpu_mark_page_dirty(vcpu, gpa >> PAGE_SHIFT);
|
2016-02-24 16:51:13 +07:00
|
|
|
kvm_page_track_write(vcpu, gpa, new, bytes);
|
2010-04-13 14:21:56 +07:00
|
|
|
|
|
|
|
return X86EMUL_CONTINUE;
|
2010-03-15 18:59:55 +07:00
|
|
|
|
2008-03-30 06:17:59 +07:00
|
|
|
emul_write:
|
2010-03-15 18:59:54 +07:00
|
|
|
printk_once(KERN_WARNING "kvm: emulating exchange as write\n");
|
2007-12-12 22:46:12 +07:00
|
|
|
|
2011-04-20 17:37:53 +07:00
|
|
|
return emulator_write_emulated(ctxt, addr, new, bytes, exception);
|
KVM: Portability: Move x86 emulation and mmio device hook to x86.c
This patch moves the following functions to from kvm_main.c to x86.c:
emulator_read/write_std, vcpu_find_pervcpu_dev, vcpu_find_mmio_dev,
emulator_read/write_emulated, emulator_write_phys,
emulator_write_emulated_onepage, emulator_cmpxchg_emulated,
get_setment_base, emulate_invlpg, emulate_clts, emulator_get/set_dr,
kvm_report_emulation_failure, emulate_instruction
The following data type is moved to x86.c:
struct x86_emulate_ops emulate_ops
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-31 00:44:21 +07:00
|
|
|
}
|
|
|
|
|
2010-03-18 20:20:23 +07:00
|
|
|
static int kernel_pio(struct kvm_vcpu *vcpu, void *pd)
|
|
|
|
{
|
|
|
|
/* TODO: String I/O for in kernel device */
|
|
|
|
int r;
|
|
|
|
|
|
|
|
if (vcpu->arch.pio.in)
|
2015-03-26 21:39:28 +07:00
|
|
|
r = kvm_io_bus_read(vcpu, KVM_PIO_BUS, vcpu->arch.pio.port,
|
2010-03-18 20:20:23 +07:00
|
|
|
vcpu->arch.pio.size, pd);
|
|
|
|
else
|
2015-03-26 21:39:28 +07:00
|
|
|
r = kvm_io_bus_write(vcpu, KVM_PIO_BUS,
|
2010-03-18 20:20:23 +07:00
|
|
|
vcpu->arch.pio.port, vcpu->arch.pio.size,
|
|
|
|
pd);
|
|
|
|
return r;
|
|
|
|
}
|
|
|
|
|
2011-09-22 15:55:10 +07:00
|
|
|
static int emulator_pio_in_out(struct kvm_vcpu *vcpu, int size,
|
|
|
|
unsigned short port, void *val,
|
|
|
|
unsigned int count, bool in)
|
2010-03-18 20:20:23 +07:00
|
|
|
{
|
|
|
|
vcpu->arch.pio.port = port;
|
2011-09-22 15:55:10 +07:00
|
|
|
vcpu->arch.pio.in = in;
|
2010-03-18 20:20:24 +07:00
|
|
|
vcpu->arch.pio.count = count;
|
2010-03-18 20:20:23 +07:00
|
|
|
vcpu->arch.pio.size = size;
|
|
|
|
|
|
|
|
if (!kernel_pio(vcpu, vcpu->arch.pio_data)) {
|
2010-03-18 20:20:24 +07:00
|
|
|
vcpu->arch.pio.count = 0;
|
2010-03-18 20:20:23 +07:00
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
|
|
|
|
vcpu->run->exit_reason = KVM_EXIT_IO;
|
2011-09-22 15:55:10 +07:00
|
|
|
vcpu->run->io.direction = in ? KVM_EXIT_IO_IN : KVM_EXIT_IO_OUT;
|
2010-03-18 20:20:23 +07:00
|
|
|
vcpu->run->io.size = size;
|
|
|
|
vcpu->run->io.data_offset = KVM_PIO_PAGE_OFFSET * PAGE_SIZE;
|
|
|
|
vcpu->run->io.count = count;
|
|
|
|
vcpu->run->io.port = port;
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2011-09-22 15:55:10 +07:00
|
|
|
static int emulator_pio_in_emulated(struct x86_emulate_ctxt *ctxt,
|
|
|
|
int size, unsigned short port, void *val,
|
|
|
|
unsigned int count)
|
2010-03-18 20:20:23 +07:00
|
|
|
{
|
2011-04-20 17:37:53 +07:00
|
|
|
struct kvm_vcpu *vcpu = emul_to_vcpu(ctxt);
|
2011-09-22 15:55:10 +07:00
|
|
|
int ret;
|
2011-04-20 17:37:53 +07:00
|
|
|
|
2011-09-22 15:55:10 +07:00
|
|
|
if (vcpu->arch.pio.count)
|
|
|
|
goto data_avail;
|
2010-03-18 20:20:23 +07:00
|
|
|
|
2011-09-22 15:55:10 +07:00
|
|
|
ret = emulator_pio_in_out(vcpu, size, port, val, count, true);
|
|
|
|
if (ret) {
|
|
|
|
data_avail:
|
|
|
|
memcpy(val, vcpu->arch.pio_data, size * count);
|
2014-05-02 22:57:47 +07:00
|
|
|
trace_kvm_pio(KVM_PIO_IN, port, size, count, vcpu->arch.pio_data);
|
2010-03-18 20:20:24 +07:00
|
|
|
vcpu->arch.pio.count = 0;
|
2010-03-18 20:20:23 +07:00
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2011-09-22 15:55:10 +07:00
|
|
|
static int emulator_pio_out_emulated(struct x86_emulate_ctxt *ctxt,
|
|
|
|
int size, unsigned short port,
|
|
|
|
const void *val, unsigned int count)
|
|
|
|
{
|
|
|
|
struct kvm_vcpu *vcpu = emul_to_vcpu(ctxt);
|
|
|
|
|
|
|
|
memcpy(vcpu->arch.pio_data, val, size * count);
|
2014-05-02 22:57:47 +07:00
|
|
|
trace_kvm_pio(KVM_PIO_OUT, port, size, count, vcpu->arch.pio_data);
|
2011-09-22 15:55:10 +07:00
|
|
|
return emulator_pio_in_out(vcpu, size, port, (void *)val, count, false);
|
|
|
|
}
|
|
|
|
|
KVM: Portability: Move x86 emulation and mmio device hook to x86.c
This patch moves the following functions to from kvm_main.c to x86.c:
emulator_read/write_std, vcpu_find_pervcpu_dev, vcpu_find_mmio_dev,
emulator_read/write_emulated, emulator_write_phys,
emulator_write_emulated_onepage, emulator_cmpxchg_emulated,
get_setment_base, emulate_invlpg, emulate_clts, emulator_get/set_dr,
kvm_report_emulation_failure, emulate_instruction
The following data type is moved to x86.c:
struct x86_emulate_ops emulate_ops
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-31 00:44:21 +07:00
|
|
|
static unsigned long get_segment_base(struct kvm_vcpu *vcpu, int seg)
|
|
|
|
{
|
|
|
|
return kvm_x86_ops->get_segment_base(vcpu, seg);
|
|
|
|
}
|
|
|
|
|
2011-04-20 19:38:44 +07:00
|
|
|
static void emulator_invlpg(struct x86_emulate_ctxt *ctxt, ulong address)
|
KVM: Portability: Move x86 emulation and mmio device hook to x86.c
This patch moves the following functions to from kvm_main.c to x86.c:
emulator_read/write_std, vcpu_find_pervcpu_dev, vcpu_find_mmio_dev,
emulator_read/write_emulated, emulator_write_phys,
emulator_write_emulated_onepage, emulator_cmpxchg_emulated,
get_setment_base, emulate_invlpg, emulate_clts, emulator_get/set_dr,
kvm_report_emulation_failure, emulate_instruction
The following data type is moved to x86.c:
struct x86_emulate_ops emulate_ops
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-31 00:44:21 +07:00
|
|
|
{
|
2011-04-20 19:38:44 +07:00
|
|
|
kvm_mmu_invlpg(emul_to_vcpu(ctxt), address);
|
KVM: Portability: Move x86 emulation and mmio device hook to x86.c
This patch moves the following functions to from kvm_main.c to x86.c:
emulator_read/write_std, vcpu_find_pervcpu_dev, vcpu_find_mmio_dev,
emulator_read/write_emulated, emulator_write_phys,
emulator_write_emulated_onepage, emulator_cmpxchg_emulated,
get_setment_base, emulate_invlpg, emulate_clts, emulator_get/set_dr,
kvm_report_emulation_failure, emulate_instruction
The following data type is moved to x86.c:
struct x86_emulate_ops emulate_ops
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-31 00:44:21 +07:00
|
|
|
}
|
|
|
|
|
2015-03-03 02:43:31 +07:00
|
|
|
int kvm_emulate_wbinvd_noskip(struct kvm_vcpu *vcpu)
|
2010-06-30 11:25:15 +07:00
|
|
|
{
|
|
|
|
if (!need_emulate_wbinvd(vcpu))
|
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
|
|
|
if (kvm_x86_ops->has_wbinvd_exit()) {
|
2010-11-01 20:01:29 +07:00
|
|
|
int cpu = get_cpu();
|
|
|
|
|
|
|
|
cpumask_set_cpu(cpu, vcpu->arch.wbinvd_dirty_mask);
|
2010-06-30 11:25:15 +07:00
|
|
|
smp_call_function_many(vcpu->arch.wbinvd_dirty_mask,
|
|
|
|
wbinvd_ipi, NULL, 1);
|
2010-11-01 20:01:29 +07:00
|
|
|
put_cpu();
|
2010-06-30 11:25:15 +07:00
|
|
|
cpumask_clear(vcpu->arch.wbinvd_dirty_mask);
|
2010-11-01 20:01:29 +07:00
|
|
|
} else
|
|
|
|
wbinvd();
|
2010-06-30 11:25:15 +07:00
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
}
|
2015-03-03 02:43:31 +07:00
|
|
|
|
|
|
|
int kvm_emulate_wbinvd(struct kvm_vcpu *vcpu)
|
|
|
|
{
|
|
|
|
kvm_x86_ops->skip_emulated_instruction(vcpu);
|
|
|
|
return kvm_emulate_wbinvd_noskip(vcpu);
|
|
|
|
}
|
2010-06-30 11:25:15 +07:00
|
|
|
EXPORT_SYMBOL_GPL(kvm_emulate_wbinvd);
|
|
|
|
|
2015-03-03 02:43:31 +07:00
|
|
|
|
|
|
|
|
2011-04-20 19:53:23 +07:00
|
|
|
static void emulator_wbinvd(struct x86_emulate_ctxt *ctxt)
|
|
|
|
{
|
2015-03-03 02:43:31 +07:00
|
|
|
kvm_emulate_wbinvd_noskip(emul_to_vcpu(ctxt));
|
2011-04-20 19:53:23 +07:00
|
|
|
}
|
|
|
|
|
2015-03-13 16:39:45 +07:00
|
|
|
static int emulator_get_dr(struct x86_emulate_ctxt *ctxt, int dr,
|
|
|
|
unsigned long *dest)
|
KVM: Portability: Move x86 emulation and mmio device hook to x86.c
This patch moves the following functions to from kvm_main.c to x86.c:
emulator_read/write_std, vcpu_find_pervcpu_dev, vcpu_find_mmio_dev,
emulator_read/write_emulated, emulator_write_phys,
emulator_write_emulated_onepage, emulator_cmpxchg_emulated,
get_setment_base, emulate_invlpg, emulate_clts, emulator_get/set_dr,
kvm_report_emulation_failure, emulate_instruction
The following data type is moved to x86.c:
struct x86_emulate_ops emulate_ops
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-31 00:44:21 +07:00
|
|
|
{
|
2014-10-03 05:10:05 +07:00
|
|
|
return kvm_get_dr(emul_to_vcpu(ctxt), dr, dest);
|
KVM: Portability: Move x86 emulation and mmio device hook to x86.c
This patch moves the following functions to from kvm_main.c to x86.c:
emulator_read/write_std, vcpu_find_pervcpu_dev, vcpu_find_mmio_dev,
emulator_read/write_emulated, emulator_write_phys,
emulator_write_emulated_onepage, emulator_cmpxchg_emulated,
get_setment_base, emulate_invlpg, emulate_clts, emulator_get/set_dr,
kvm_report_emulation_failure, emulate_instruction
The following data type is moved to x86.c:
struct x86_emulate_ops emulate_ops
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-31 00:44:21 +07:00
|
|
|
}
|
|
|
|
|
2015-03-13 16:39:45 +07:00
|
|
|
static int emulator_set_dr(struct x86_emulate_ctxt *ctxt, int dr,
|
|
|
|
unsigned long value)
|
KVM: Portability: Move x86 emulation and mmio device hook to x86.c
This patch moves the following functions to from kvm_main.c to x86.c:
emulator_read/write_std, vcpu_find_pervcpu_dev, vcpu_find_mmio_dev,
emulator_read/write_emulated, emulator_write_phys,
emulator_write_emulated_onepage, emulator_cmpxchg_emulated,
get_setment_base, emulate_invlpg, emulate_clts, emulator_get/set_dr,
kvm_report_emulation_failure, emulate_instruction
The following data type is moved to x86.c:
struct x86_emulate_ops emulate_ops
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-31 00:44:21 +07:00
|
|
|
{
|
2010-04-28 23:15:32 +07:00
|
|
|
|
2011-04-20 17:37:53 +07:00
|
|
|
return __kvm_set_dr(emul_to_vcpu(ctxt), dr, value);
|
KVM: Portability: Move x86 emulation and mmio device hook to x86.c
This patch moves the following functions to from kvm_main.c to x86.c:
emulator_read/write_std, vcpu_find_pervcpu_dev, vcpu_find_mmio_dev,
emulator_read/write_emulated, emulator_write_phys,
emulator_write_emulated_onepage, emulator_cmpxchg_emulated,
get_setment_base, emulate_invlpg, emulate_clts, emulator_get/set_dr,
kvm_report_emulation_failure, emulate_instruction
The following data type is moved to x86.c:
struct x86_emulate_ops emulate_ops
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-31 00:44:21 +07:00
|
|
|
}
|
|
|
|
|
2010-03-18 20:20:03 +07:00
|
|
|
static u64 mk_cr_64(u64 curr_cr, u32 new_val)
|
2008-06-28 00:58:02 +07:00
|
|
|
{
|
2010-03-18 20:20:03 +07:00
|
|
|
return (curr_cr & ~((1ULL << 32) - 1)) | new_val;
|
2008-06-28 00:58:02 +07:00
|
|
|
}
|
|
|
|
|
2011-04-20 17:37:53 +07:00
|
|
|
static unsigned long emulator_get_cr(struct x86_emulate_ctxt *ctxt, int cr)
|
KVM: Portability: Move x86 emulation and mmio device hook to x86.c
This patch moves the following functions to from kvm_main.c to x86.c:
emulator_read/write_std, vcpu_find_pervcpu_dev, vcpu_find_mmio_dev,
emulator_read/write_emulated, emulator_write_phys,
emulator_write_emulated_onepage, emulator_cmpxchg_emulated,
get_setment_base, emulate_invlpg, emulate_clts, emulator_get/set_dr,
kvm_report_emulation_failure, emulate_instruction
The following data type is moved to x86.c:
struct x86_emulate_ops emulate_ops
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-31 00:44:21 +07:00
|
|
|
{
|
2011-04-20 17:37:53 +07:00
|
|
|
struct kvm_vcpu *vcpu = emul_to_vcpu(ctxt);
|
2010-03-18 20:20:03 +07:00
|
|
|
unsigned long value;
|
|
|
|
|
|
|
|
switch (cr) {
|
|
|
|
case 0:
|
|
|
|
value = kvm_read_cr0(vcpu);
|
|
|
|
break;
|
|
|
|
case 2:
|
|
|
|
value = vcpu->arch.cr2;
|
|
|
|
break;
|
|
|
|
case 3:
|
2010-12-05 22:30:00 +07:00
|
|
|
value = kvm_read_cr3(vcpu);
|
2010-03-18 20:20:03 +07:00
|
|
|
break;
|
|
|
|
case 4:
|
|
|
|
value = kvm_read_cr4(vcpu);
|
|
|
|
break;
|
|
|
|
case 8:
|
|
|
|
value = kvm_get_cr8(vcpu);
|
|
|
|
break;
|
|
|
|
default:
|
KVM: Cleanup the kvm_print functions and introduce pr_XX wrappers
Introduces a couple of print functions, which are essentially wrappers
around standard printk functions, with a KVM: prefix.
Functions introduced or modified are:
- kvm_err(fmt, ...)
- kvm_info(fmt, ...)
- kvm_debug(fmt, ...)
- kvm_pr_unimpl(fmt, ...)
- pr_unimpl(vcpu, fmt, ...) -> vcpu_unimpl(vcpu, fmt, ...)
Signed-off-by: Christoffer Dall <c.dall@virtualopensystems.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
2012-06-04 01:17:48 +07:00
|
|
|
kvm_err("%s: unexpected cr %u\n", __func__, cr);
|
2010-03-18 20:20:03 +07:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
return value;
|
|
|
|
}
|
|
|
|
|
2011-04-20 17:37:53 +07:00
|
|
|
static int emulator_set_cr(struct x86_emulate_ctxt *ctxt, int cr, ulong val)
|
2010-03-18 20:20:03 +07:00
|
|
|
{
|
2011-04-20 17:37:53 +07:00
|
|
|
struct kvm_vcpu *vcpu = emul_to_vcpu(ctxt);
|
2010-04-28 23:15:31 +07:00
|
|
|
int res = 0;
|
|
|
|
|
2010-03-18 20:20:03 +07:00
|
|
|
switch (cr) {
|
|
|
|
case 0:
|
2010-06-10 21:02:14 +07:00
|
|
|
res = kvm_set_cr0(vcpu, mk_cr_64(kvm_read_cr0(vcpu), val));
|
2010-03-18 20:20:03 +07:00
|
|
|
break;
|
|
|
|
case 2:
|
|
|
|
vcpu->arch.cr2 = val;
|
|
|
|
break;
|
|
|
|
case 3:
|
2010-06-10 21:02:16 +07:00
|
|
|
res = kvm_set_cr3(vcpu, val);
|
2010-03-18 20:20:03 +07:00
|
|
|
break;
|
|
|
|
case 4:
|
2010-06-10 21:02:15 +07:00
|
|
|
res = kvm_set_cr4(vcpu, mk_cr_64(kvm_read_cr4(vcpu), val));
|
2010-03-18 20:20:03 +07:00
|
|
|
break;
|
|
|
|
case 8:
|
2010-12-21 17:12:00 +07:00
|
|
|
res = kvm_set_cr8(vcpu, val);
|
2010-03-18 20:20:03 +07:00
|
|
|
break;
|
|
|
|
default:
|
KVM: Cleanup the kvm_print functions and introduce pr_XX wrappers
Introduces a couple of print functions, which are essentially wrappers
around standard printk functions, with a KVM: prefix.
Functions introduced or modified are:
- kvm_err(fmt, ...)
- kvm_info(fmt, ...)
- kvm_debug(fmt, ...)
- kvm_pr_unimpl(fmt, ...)
- pr_unimpl(vcpu, fmt, ...) -> vcpu_unimpl(vcpu, fmt, ...)
Signed-off-by: Christoffer Dall <c.dall@virtualopensystems.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
2012-06-04 01:17:48 +07:00
|
|
|
kvm_err("%s: unexpected cr %u\n", __func__, cr);
|
2010-04-28 23:15:31 +07:00
|
|
|
res = -1;
|
2010-03-18 20:20:03 +07:00
|
|
|
}
|
2010-04-28 23:15:31 +07:00
|
|
|
|
|
|
|
return res;
|
2010-03-18 20:20:03 +07:00
|
|
|
}
|
|
|
|
|
2011-04-20 17:37:53 +07:00
|
|
|
static int emulator_get_cpl(struct x86_emulate_ctxt *ctxt)
|
2010-03-18 20:20:05 +07:00
|
|
|
{
|
2011-04-20 17:37:53 +07:00
|
|
|
return kvm_x86_ops->get_cpl(emul_to_vcpu(ctxt));
|
2010-03-18 20:20:05 +07:00
|
|
|
}
|
|
|
|
|
2011-04-20 17:37:53 +07:00
|
|
|
static void emulator_get_gdt(struct x86_emulate_ctxt *ctxt, struct desc_ptr *dt)
|
2010-03-18 20:20:16 +07:00
|
|
|
{
|
2011-04-20 17:37:53 +07:00
|
|
|
kvm_x86_ops->get_gdt(emul_to_vcpu(ctxt), dt);
|
2010-03-18 20:20:16 +07:00
|
|
|
}
|
|
|
|
|
2011-04-20 17:37:53 +07:00
|
|
|
static void emulator_get_idt(struct x86_emulate_ctxt *ctxt, struct desc_ptr *dt)
|
2010-08-04 09:44:24 +07:00
|
|
|
{
|
2011-04-20 17:37:53 +07:00
|
|
|
kvm_x86_ops->get_idt(emul_to_vcpu(ctxt), dt);
|
2010-08-04 09:44:24 +07:00
|
|
|
}
|
|
|
|
|
2011-04-20 19:12:00 +07:00
|
|
|
static void emulator_set_gdt(struct x86_emulate_ctxt *ctxt, struct desc_ptr *dt)
|
|
|
|
{
|
|
|
|
kvm_x86_ops->set_gdt(emul_to_vcpu(ctxt), dt);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void emulator_set_idt(struct x86_emulate_ctxt *ctxt, struct desc_ptr *dt)
|
|
|
|
{
|
|
|
|
kvm_x86_ops->set_idt(emul_to_vcpu(ctxt), dt);
|
|
|
|
}
|
|
|
|
|
2011-04-20 17:37:53 +07:00
|
|
|
static unsigned long emulator_get_cached_segment_base(
|
|
|
|
struct x86_emulate_ctxt *ctxt, int seg)
|
2010-04-28 23:15:29 +07:00
|
|
|
{
|
2011-04-20 17:37:53 +07:00
|
|
|
return get_segment_base(emul_to_vcpu(ctxt), seg);
|
2010-04-28 23:15:29 +07:00
|
|
|
}
|
|
|
|
|
2011-04-27 17:20:30 +07:00
|
|
|
static bool emulator_get_segment(struct x86_emulate_ctxt *ctxt, u16 *selector,
|
|
|
|
struct desc_struct *desc, u32 *base3,
|
|
|
|
int seg)
|
2010-03-18 20:20:16 +07:00
|
|
|
{
|
|
|
|
struct kvm_segment var;
|
|
|
|
|
2011-04-20 17:37:53 +07:00
|
|
|
kvm_get_segment(emul_to_vcpu(ctxt), &var, seg);
|
2011-04-27 17:20:30 +07:00
|
|
|
*selector = var.selector;
|
2010-03-18 20:20:16 +07:00
|
|
|
|
2013-01-21 20:36:48 +07:00
|
|
|
if (var.unusable) {
|
|
|
|
memset(desc, 0, sizeof(*desc));
|
2010-03-18 20:20:16 +07:00
|
|
|
return false;
|
2013-01-21 20:36:48 +07:00
|
|
|
}
|
2010-03-18 20:20:16 +07:00
|
|
|
|
|
|
|
if (var.g)
|
|
|
|
var.limit >>= 12;
|
|
|
|
set_desc_limit(desc, var.limit);
|
|
|
|
set_desc_base(desc, (unsigned long)var.base);
|
2011-03-07 19:55:06 +07:00
|
|
|
#ifdef CONFIG_X86_64
|
|
|
|
if (base3)
|
|
|
|
*base3 = var.base >> 32;
|
|
|
|
#endif
|
2010-03-18 20:20:16 +07:00
|
|
|
desc->type = var.type;
|
|
|
|
desc->s = var.s;
|
|
|
|
desc->dpl = var.dpl;
|
|
|
|
desc->p = var.present;
|
|
|
|
desc->avl = var.avl;
|
|
|
|
desc->l = var.l;
|
|
|
|
desc->d = var.db;
|
|
|
|
desc->g = var.g;
|
|
|
|
|
|
|
|
return true;
|
|
|
|
}
|
|
|
|
|
2011-04-27 17:20:30 +07:00
|
|
|
static void emulator_set_segment(struct x86_emulate_ctxt *ctxt, u16 selector,
|
|
|
|
struct desc_struct *desc, u32 base3,
|
|
|
|
int seg)
|
2010-03-18 20:20:16 +07:00
|
|
|
{
|
2011-04-20 17:37:53 +07:00
|
|
|
struct kvm_vcpu *vcpu = emul_to_vcpu(ctxt);
|
2010-03-18 20:20:16 +07:00
|
|
|
struct kvm_segment var;
|
|
|
|
|
2011-04-27 17:20:30 +07:00
|
|
|
var.selector = selector;
|
2010-03-18 20:20:16 +07:00
|
|
|
var.base = get_desc_base(desc);
|
2011-03-07 19:55:06 +07:00
|
|
|
#ifdef CONFIG_X86_64
|
|
|
|
var.base |= ((u64)base3) << 32;
|
|
|
|
#endif
|
2010-03-18 20:20:16 +07:00
|
|
|
var.limit = get_desc_limit(desc);
|
|
|
|
if (desc->g)
|
|
|
|
var.limit = (var.limit << 12) | 0xfff;
|
|
|
|
var.type = desc->type;
|
|
|
|
var.dpl = desc->dpl;
|
|
|
|
var.db = desc->d;
|
|
|
|
var.s = desc->s;
|
|
|
|
var.l = desc->l;
|
|
|
|
var.g = desc->g;
|
|
|
|
var.avl = desc->avl;
|
|
|
|
var.present = desc->p;
|
|
|
|
var.unusable = !var.present;
|
|
|
|
var.padding = 0;
|
|
|
|
|
|
|
|
kvm_set_segment(vcpu, &var, seg);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
2011-04-20 17:37:53 +07:00
|
|
|
static int emulator_get_msr(struct x86_emulate_ctxt *ctxt,
|
|
|
|
u32 msr_index, u64 *pdata)
|
|
|
|
{
|
2015-04-08 20:30:38 +07:00
|
|
|
struct msr_data msr;
|
|
|
|
int r;
|
|
|
|
|
|
|
|
msr.index = msr_index;
|
|
|
|
msr.host_initiated = false;
|
|
|
|
r = kvm_get_msr(emul_to_vcpu(ctxt), &msr);
|
|
|
|
if (r)
|
|
|
|
return r;
|
|
|
|
|
|
|
|
*pdata = msr.data;
|
|
|
|
return 0;
|
2011-04-20 17:37:53 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
static int emulator_set_msr(struct x86_emulate_ctxt *ctxt,
|
|
|
|
u32 msr_index, u64 data)
|
|
|
|
{
|
2012-11-30 03:42:12 +07:00
|
|
|
struct msr_data msr;
|
|
|
|
|
|
|
|
msr.data = data;
|
|
|
|
msr.index = msr_index;
|
|
|
|
msr.host_initiated = false;
|
|
|
|
return kvm_set_msr(emul_to_vcpu(ctxt), &msr);
|
2011-04-20 17:37:53 +07:00
|
|
|
}
|
|
|
|
|
2015-05-07 16:36:11 +07:00
|
|
|
static u64 emulator_get_smbase(struct x86_emulate_ctxt *ctxt)
|
|
|
|
{
|
|
|
|
struct kvm_vcpu *vcpu = emul_to_vcpu(ctxt);
|
|
|
|
|
|
|
|
return vcpu->arch.smbase;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void emulator_set_smbase(struct x86_emulate_ctxt *ctxt, u64 smbase)
|
|
|
|
{
|
|
|
|
struct kvm_vcpu *vcpu = emul_to_vcpu(ctxt);
|
|
|
|
|
|
|
|
vcpu->arch.smbase = smbase;
|
|
|
|
}
|
|
|
|
|
2014-06-02 22:34:09 +07:00
|
|
|
static int emulator_check_pmc(struct x86_emulate_ctxt *ctxt,
|
|
|
|
u32 pmc)
|
|
|
|
{
|
2015-06-19 18:44:45 +07:00
|
|
|
return kvm_pmu_is_valid_msr_idx(emul_to_vcpu(ctxt), pmc);
|
2014-06-02 22:34:09 +07:00
|
|
|
}
|
|
|
|
|
2011-11-10 19:57:30 +07:00
|
|
|
static int emulator_read_pmc(struct x86_emulate_ctxt *ctxt,
|
|
|
|
u32 pmc, u64 *pdata)
|
|
|
|
{
|
2015-06-19 18:44:45 +07:00
|
|
|
return kvm_pmu_rdpmc(emul_to_vcpu(ctxt), pmc, pdata);
|
2011-11-10 19:57:30 +07:00
|
|
|
}
|
|
|
|
|
2011-04-20 19:43:05 +07:00
|
|
|
static void emulator_halt(struct x86_emulate_ctxt *ctxt)
|
|
|
|
{
|
|
|
|
emul_to_vcpu(ctxt)->arch.halt_request = 1;
|
|
|
|
}
|
|
|
|
|
2011-03-28 21:53:59 +07:00
|
|
|
static void emulator_get_fpu(struct x86_emulate_ctxt *ctxt)
|
|
|
|
{
|
|
|
|
preempt_disable();
|
2011-04-20 19:55:40 +07:00
|
|
|
kvm_load_guest_fpu(emul_to_vcpu(ctxt));
|
2011-03-28 21:53:59 +07:00
|
|
|
/*
|
|
|
|
* CR0.TS may reference the host fpu state, not the guest fpu state,
|
|
|
|
* so it may be clear at this point.
|
|
|
|
*/
|
|
|
|
clts();
|
|
|
|
}
|
|
|
|
|
|
|
|
static void emulator_put_fpu(struct x86_emulate_ctxt *ctxt)
|
|
|
|
{
|
|
|
|
preempt_enable();
|
|
|
|
}
|
|
|
|
|
2011-04-20 17:37:53 +07:00
|
|
|
static int emulator_intercept(struct x86_emulate_ctxt *ctxt,
|
2011-04-04 17:39:27 +07:00
|
|
|
struct x86_instruction_info *info,
|
2011-04-04 17:39:22 +07:00
|
|
|
enum x86_intercept_stage stage)
|
|
|
|
{
|
2011-04-20 17:37:53 +07:00
|
|
|
return kvm_x86_ops->check_intercept(emul_to_vcpu(ctxt), info, stage);
|
2011-04-04 17:39:22 +07:00
|
|
|
}
|
|
|
|
|
2012-06-07 18:10:16 +07:00
|
|
|
static void emulator_get_cpuid(struct x86_emulate_ctxt *ctxt,
|
2012-01-12 22:43:03 +07:00
|
|
|
u32 *eax, u32 *ebx, u32 *ecx, u32 *edx)
|
|
|
|
{
|
2012-06-07 18:10:16 +07:00
|
|
|
kvm_cpuid(emul_to_vcpu(ctxt), eax, ebx, ecx, edx);
|
2012-01-12 22:43:03 +07:00
|
|
|
}
|
|
|
|
|
2012-08-28 03:46:17 +07:00
|
|
|
static ulong emulator_read_gpr(struct x86_emulate_ctxt *ctxt, unsigned reg)
|
|
|
|
{
|
|
|
|
return kvm_register_read(emul_to_vcpu(ctxt), reg);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void emulator_write_gpr(struct x86_emulate_ctxt *ctxt, unsigned reg, ulong val)
|
|
|
|
{
|
|
|
|
kvm_register_write(emul_to_vcpu(ctxt), reg, val);
|
|
|
|
}
|
|
|
|
|
2015-01-26 14:32:23 +07:00
|
|
|
static void emulator_set_nmi_mask(struct x86_emulate_ctxt *ctxt, bool masked)
|
|
|
|
{
|
|
|
|
kvm_x86_ops->set_nmi_mask(emul_to_vcpu(ctxt), masked);
|
|
|
|
}
|
|
|
|
|
2012-08-30 06:30:16 +07:00
|
|
|
static const struct x86_emulate_ops emulate_ops = {
|
2012-08-28 03:46:17 +07:00
|
|
|
.read_gpr = emulator_read_gpr,
|
|
|
|
.write_gpr = emulator_write_gpr,
|
2010-02-10 19:21:32 +07:00
|
|
|
.read_std = kvm_read_guest_virt_system,
|
2010-03-18 20:20:16 +07:00
|
|
|
.write_std = kvm_write_guest_virt_system,
|
2015-10-30 22:36:24 +07:00
|
|
|
.read_phys = kvm_read_guest_phys_system,
|
2010-02-10 19:21:32 +07:00
|
|
|
.fetch = kvm_fetch_guest_virt,
|
KVM: Portability: Move x86 emulation and mmio device hook to x86.c
This patch moves the following functions to from kvm_main.c to x86.c:
emulator_read/write_std, vcpu_find_pervcpu_dev, vcpu_find_mmio_dev,
emulator_read/write_emulated, emulator_write_phys,
emulator_write_emulated_onepage, emulator_cmpxchg_emulated,
get_setment_base, emulate_invlpg, emulate_clts, emulator_get/set_dr,
kvm_report_emulation_failure, emulate_instruction
The following data type is moved to x86.c:
struct x86_emulate_ops emulate_ops
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-31 00:44:21 +07:00
|
|
|
.read_emulated = emulator_read_emulated,
|
|
|
|
.write_emulated = emulator_write_emulated,
|
|
|
|
.cmpxchg_emulated = emulator_cmpxchg_emulated,
|
2011-04-20 19:38:44 +07:00
|
|
|
.invlpg = emulator_invlpg,
|
2010-03-18 20:20:23 +07:00
|
|
|
.pio_in_emulated = emulator_pio_in_emulated,
|
|
|
|
.pio_out_emulated = emulator_pio_out_emulated,
|
2011-04-27 17:20:30 +07:00
|
|
|
.get_segment = emulator_get_segment,
|
|
|
|
.set_segment = emulator_set_segment,
|
2010-04-28 23:15:29 +07:00
|
|
|
.get_cached_segment_base = emulator_get_cached_segment_base,
|
2010-03-18 20:20:16 +07:00
|
|
|
.get_gdt = emulator_get_gdt,
|
2010-08-04 09:44:24 +07:00
|
|
|
.get_idt = emulator_get_idt,
|
2011-04-20 19:12:00 +07:00
|
|
|
.set_gdt = emulator_set_gdt,
|
|
|
|
.set_idt = emulator_set_idt,
|
2010-03-18 20:20:03 +07:00
|
|
|
.get_cr = emulator_get_cr,
|
|
|
|
.set_cr = emulator_set_cr,
|
2010-03-18 20:20:05 +07:00
|
|
|
.cpl = emulator_get_cpl,
|
2010-04-28 23:15:27 +07:00
|
|
|
.get_dr = emulator_get_dr,
|
|
|
|
.set_dr = emulator_set_dr,
|
2015-05-07 16:36:11 +07:00
|
|
|
.get_smbase = emulator_get_smbase,
|
|
|
|
.set_smbase = emulator_set_smbase,
|
2011-04-20 17:37:53 +07:00
|
|
|
.set_msr = emulator_set_msr,
|
|
|
|
.get_msr = emulator_get_msr,
|
2014-06-02 22:34:09 +07:00
|
|
|
.check_pmc = emulator_check_pmc,
|
2011-11-10 19:57:30 +07:00
|
|
|
.read_pmc = emulator_read_pmc,
|
2011-04-20 19:43:05 +07:00
|
|
|
.halt = emulator_halt,
|
2011-04-20 19:53:23 +07:00
|
|
|
.wbinvd = emulator_wbinvd,
|
2011-04-20 19:47:13 +07:00
|
|
|
.fix_hypercall = emulator_fix_hypercall,
|
2011-03-28 21:53:59 +07:00
|
|
|
.get_fpu = emulator_get_fpu,
|
|
|
|
.put_fpu = emulator_put_fpu,
|
2011-04-04 17:39:22 +07:00
|
|
|
.intercept = emulator_intercept,
|
2012-01-12 22:43:03 +07:00
|
|
|
.get_cpuid = emulator_get_cpuid,
|
2015-01-26 14:32:23 +07:00
|
|
|
.set_nmi_mask = emulator_set_nmi_mask,
|
KVM: Portability: Move x86 emulation and mmio device hook to x86.c
This patch moves the following functions to from kvm_main.c to x86.c:
emulator_read/write_std, vcpu_find_pervcpu_dev, vcpu_find_mmio_dev,
emulator_read/write_emulated, emulator_write_phys,
emulator_write_emulated_onepage, emulator_cmpxchg_emulated,
get_setment_base, emulate_invlpg, emulate_clts, emulator_get/set_dr,
kvm_report_emulation_failure, emulate_instruction
The following data type is moved to x86.c:
struct x86_emulate_ops emulate_ops
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-31 00:44:21 +07:00
|
|
|
};
|
|
|
|
|
2010-04-28 23:15:43 +07:00
|
|
|
static void toggle_interruptibility(struct kvm_vcpu *vcpu, u32 mask)
|
|
|
|
{
|
2014-05-20 19:29:47 +07:00
|
|
|
u32 int_shadow = kvm_x86_ops->get_interrupt_shadow(vcpu);
|
2010-04-28 23:15:43 +07:00
|
|
|
/*
|
|
|
|
* an sti; sti; sequence only disable interrupts for the first
|
|
|
|
* instruction. So, if the last instruction, be it emulated or
|
|
|
|
* not, left the system with the INT_STI flag enabled, it
|
|
|
|
* means that the last instruction is an sti. We should not
|
|
|
|
* leave the flag on in this case. The same goes for mov ss
|
|
|
|
*/
|
2014-05-20 19:29:47 +07:00
|
|
|
if (int_shadow & mask)
|
|
|
|
mask = 0;
|
2014-03-27 17:29:28 +07:00
|
|
|
if (unlikely(int_shadow || mask)) {
|
2010-04-28 23:15:43 +07:00
|
|
|
kvm_x86_ops->set_interrupt_shadow(vcpu, mask);
|
2014-03-27 17:29:28 +07:00
|
|
|
if (!mask)
|
|
|
|
kvm_make_request(KVM_REQ_EVENT, vcpu);
|
|
|
|
}
|
2010-04-28 23:15:43 +07:00
|
|
|
}
|
|
|
|
|
2014-09-05 00:46:15 +07:00
|
|
|
static bool inject_emulated_exception(struct kvm_vcpu *vcpu)
|
2010-04-28 23:15:44 +07:00
|
|
|
{
|
|
|
|
struct x86_emulate_ctxt *ctxt = &vcpu->arch.emulate_ctxt;
|
2010-11-22 22:53:21 +07:00
|
|
|
if (ctxt->exception.vector == PF_VECTOR)
|
2014-09-05 00:46:15 +07:00
|
|
|
return kvm_propagate_fault(vcpu, &ctxt->exception);
|
|
|
|
|
|
|
|
if (ctxt->exception.error_code_valid)
|
2010-11-22 22:53:21 +07:00
|
|
|
kvm_queue_exception_e(vcpu, ctxt->exception.vector,
|
|
|
|
ctxt->exception.error_code);
|
2010-04-28 23:15:44 +07:00
|
|
|
else
|
2010-11-22 22:53:21 +07:00
|
|
|
kvm_queue_exception(vcpu, ctxt->exception.vector);
|
2014-09-05 00:46:15 +07:00
|
|
|
return false;
|
2010-04-28 23:15:44 +07:00
|
|
|
}
|
|
|
|
|
2010-08-16 04:47:01 +07:00
|
|
|
static void init_emulate_ctxt(struct kvm_vcpu *vcpu)
|
|
|
|
{
|
2011-05-25 09:06:16 +07:00
|
|
|
struct x86_emulate_ctxt *ctxt = &vcpu->arch.emulate_ctxt;
|
2010-08-16 04:47:01 +07:00
|
|
|
int cs_db, cs_l;
|
|
|
|
|
|
|
|
kvm_x86_ops->get_cs_db_l_bits(vcpu, &cs_db, &cs_l);
|
|
|
|
|
2011-05-25 09:06:16 +07:00
|
|
|
ctxt->eflags = kvm_get_rflags(vcpu);
|
|
|
|
ctxt->eip = kvm_rip_read(vcpu);
|
|
|
|
ctxt->mode = (!is_protmode(vcpu)) ? X86EMUL_MODE_REAL :
|
|
|
|
(ctxt->eflags & X86_EFLAGS_VM) ? X86EMUL_MODE_VM86 :
|
2014-04-18 11:11:34 +07:00
|
|
|
(cs_l && is_long_mode(vcpu)) ? X86EMUL_MODE_PROT64 :
|
2011-05-25 09:06:16 +07:00
|
|
|
cs_db ? X86EMUL_MODE_PROT32 :
|
|
|
|
X86EMUL_MODE_PROT16;
|
2015-04-01 23:18:53 +07:00
|
|
|
BUILD_BUG_ON(HF_GUEST_MASK != X86EMUL_GUEST_MASK);
|
2015-05-07 16:36:11 +07:00
|
|
|
BUILD_BUG_ON(HF_SMM_MASK != X86EMUL_SMM_MASK);
|
|
|
|
BUILD_BUG_ON(HF_SMM_INSIDE_NMI_MASK != X86EMUL_SMM_INSIDE_NMI_MASK);
|
2015-04-01 23:18:53 +07:00
|
|
|
ctxt->emul_flags = vcpu->arch.hflags;
|
2011-05-25 09:06:16 +07:00
|
|
|
|
2012-08-28 03:46:17 +07:00
|
|
|
init_decode_cache(ctxt);
|
2011-03-31 17:06:41 +07:00
|
|
|
vcpu->arch.emulate_regs_need_sync_from_vcpu = false;
|
2010-08-16 04:47:01 +07:00
|
|
|
}
|
|
|
|
|
2011-04-13 21:12:54 +07:00
|
|
|
int kvm_inject_realmode_interrupt(struct kvm_vcpu *vcpu, int irq, int inc_eip)
|
2010-09-19 19:34:06 +07:00
|
|
|
{
|
2011-05-29 19:53:48 +07:00
|
|
|
struct x86_emulate_ctxt *ctxt = &vcpu->arch.emulate_ctxt;
|
2010-09-19 19:34:06 +07:00
|
|
|
int ret;
|
|
|
|
|
|
|
|
init_emulate_ctxt(vcpu);
|
|
|
|
|
2011-06-01 19:34:25 +07:00
|
|
|
ctxt->op_bytes = 2;
|
|
|
|
ctxt->ad_bytes = 2;
|
|
|
|
ctxt->_eip = ctxt->eip + inc_eip;
|
2011-05-29 19:53:48 +07:00
|
|
|
ret = emulate_int_real(ctxt, irq);
|
2010-09-19 19:34:06 +07:00
|
|
|
|
|
|
|
if (ret != X86EMUL_CONTINUE)
|
|
|
|
return EMULATE_FAIL;
|
|
|
|
|
2011-06-01 19:34:25 +07:00
|
|
|
ctxt->eip = ctxt->_eip;
|
2011-05-29 19:53:48 +07:00
|
|
|
kvm_rip_write(vcpu, ctxt->eip);
|
|
|
|
kvm_set_rflags(vcpu, ctxt->eflags);
|
2010-09-19 19:34:06 +07:00
|
|
|
|
|
|
|
if (irq == NMI_VECTOR)
|
2011-09-20 17:43:14 +07:00
|
|
|
vcpu->arch.nmi_pending = 0;
|
2010-09-19 19:34:06 +07:00
|
|
|
else
|
|
|
|
vcpu->arch.interrupt.pending = false;
|
|
|
|
|
|
|
|
return EMULATE_DONE;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(kvm_inject_realmode_interrupt);
|
|
|
|
|
2010-05-10 15:16:56 +07:00
|
|
|
static int handle_emulation_failure(struct kvm_vcpu *vcpu)
|
|
|
|
{
|
2010-11-29 23:51:49 +07:00
|
|
|
int r = EMULATE_DONE;
|
|
|
|
|
2010-05-10 15:16:56 +07:00
|
|
|
++vcpu->stat.insn_emulation_fail;
|
|
|
|
trace_kvm_emulate_insn_failed(vcpu);
|
2014-09-17 06:50:50 +07:00
|
|
|
if (!is_guest_mode(vcpu) && kvm_x86_ops->get_cpl(vcpu) == 0) {
|
2010-11-29 23:51:49 +07:00
|
|
|
vcpu->run->exit_reason = KVM_EXIT_INTERNAL_ERROR;
|
|
|
|
vcpu->run->internal.suberror = KVM_INTERNAL_ERROR_EMULATION;
|
|
|
|
vcpu->run->internal.ndata = 0;
|
|
|
|
r = EMULATE_FAIL;
|
|
|
|
}
|
2010-05-10 15:16:56 +07:00
|
|
|
kvm_queue_exception(vcpu, UD_VECTOR);
|
2010-11-29 23:51:49 +07:00
|
|
|
|
|
|
|
return r;
|
2010-05-10 15:16:56 +07:00
|
|
|
}
|
|
|
|
|
2013-01-13 22:49:07 +07:00
|
|
|
static bool reexecute_instruction(struct kvm_vcpu *vcpu, gva_t cr2,
|
2013-04-11 16:10:51 +07:00
|
|
|
bool write_fault_to_shadow_pgtable,
|
|
|
|
int emulation_type)
|
2010-07-08 16:41:12 +07:00
|
|
|
{
|
2013-01-13 22:46:52 +07:00
|
|
|
gpa_t gpa = cr2;
|
kvm: rename pfn_t to kvm_pfn_t
To date, we have implemented two I/O usage models for persistent memory,
PMEM (a persistent "ram disk") and DAX (mmap persistent memory into
userspace). This series adds a third, DAX-GUP, that allows DAX mappings
to be the target of direct-i/o. It allows userspace to coordinate
DMA/RDMA from/to persistent memory.
The implementation leverages the ZONE_DEVICE mm-zone that went into
4.3-rc1 (also discussed at kernel summit) to flag pages that are owned
and dynamically mapped by a device driver. The pmem driver, after
mapping a persistent memory range into the system memmap via
devm_memremap_pages(), arranges for DAX to distinguish pfn-only versus
page-backed pmem-pfns via flags in the new pfn_t type.
The DAX code, upon seeing a PFN_DEV+PFN_MAP flagged pfn, flags the
resulting pte(s) inserted into the process page tables with a new
_PAGE_DEVMAP flag. Later, when get_user_pages() is walking ptes it keys
off _PAGE_DEVMAP to pin the device hosting the page range active.
Finally, get_page() and put_page() are modified to take references
against the device driver established page mapping.
Finally, this need for "struct page" for persistent memory requires
memory capacity to store the memmap array. Given the memmap array for a
large pool of persistent may exhaust available DRAM introduce a
mechanism to allocate the memmap from persistent memory. The new
"struct vmem_altmap *" parameter to devm_memremap_pages() enables
arch_add_memory() to use reserved pmem capacity rather than the page
allocator.
This patch (of 18):
The core has developed a need for a "pfn_t" type [1]. Move the existing
pfn_t in KVM to kvm_pfn_t [2].
[1]: https://lists.01.org/pipermail/linux-nvdimm/2015-September/002199.html
[2]: https://lists.01.org/pipermail/linux-nvdimm/2015-September/002218.html
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Acked-by: Christoffer Dall <christoffer.dall@linaro.org>
Cc: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-16 07:56:11 +07:00
|
|
|
kvm_pfn_t pfn;
|
2010-07-08 16:41:12 +07:00
|
|
|
|
2013-04-11 16:10:51 +07:00
|
|
|
if (emulation_type & EMULTYPE_NO_REEXECUTE)
|
|
|
|
return false;
|
|
|
|
|
2013-01-13 22:46:52 +07:00
|
|
|
if (!vcpu->arch.mmu.direct_map) {
|
|
|
|
/*
|
|
|
|
* Write permission should be allowed since only
|
|
|
|
* write access need to be emulated.
|
|
|
|
*/
|
|
|
|
gpa = kvm_mmu_gva_to_gpa_write(vcpu, cr2, NULL);
|
2010-07-08 16:41:12 +07:00
|
|
|
|
2013-01-13 22:46:52 +07:00
|
|
|
/*
|
|
|
|
* If the mapping is invalid in guest, let cpu retry
|
|
|
|
* it to generate fault.
|
|
|
|
*/
|
|
|
|
if (gpa == UNMAPPED_GVA)
|
|
|
|
return true;
|
|
|
|
}
|
2010-07-08 16:41:12 +07:00
|
|
|
|
2012-08-21 09:57:42 +07:00
|
|
|
/*
|
|
|
|
* Do not retry the unhandleable instruction if it faults on the
|
|
|
|
* readonly host memory, otherwise it will goto a infinite loop:
|
|
|
|
* retry instruction -> write #PF -> emulation fail -> retry
|
|
|
|
* instruction -> ...
|
|
|
|
*/
|
|
|
|
pfn = gfn_to_pfn(vcpu->kvm, gpa_to_gfn(gpa));
|
2013-01-13 22:46:52 +07:00
|
|
|
|
|
|
|
/*
|
|
|
|
* If the instruction failed on the error pfn, it can not be fixed,
|
|
|
|
* report the error to userspace.
|
|
|
|
*/
|
|
|
|
if (is_error_noslot_pfn(pfn))
|
|
|
|
return false;
|
|
|
|
|
|
|
|
kvm_release_pfn_clean(pfn);
|
|
|
|
|
|
|
|
/* The instructions are well-emulated on direct mmu. */
|
|
|
|
if (vcpu->arch.mmu.direct_map) {
|
|
|
|
unsigned int indirect_shadow_pages;
|
|
|
|
|
|
|
|
spin_lock(&vcpu->kvm->mmu_lock);
|
|
|
|
indirect_shadow_pages = vcpu->kvm->arch.indirect_shadow_pages;
|
|
|
|
spin_unlock(&vcpu->kvm->mmu_lock);
|
|
|
|
|
|
|
|
if (indirect_shadow_pages)
|
|
|
|
kvm_mmu_unprotect_page(vcpu->kvm, gpa_to_gfn(gpa));
|
|
|
|
|
2010-07-08 16:41:12 +07:00
|
|
|
return true;
|
2012-08-21 09:57:42 +07:00
|
|
|
}
|
2010-07-08 16:41:12 +07:00
|
|
|
|
2013-01-13 22:46:52 +07:00
|
|
|
/*
|
|
|
|
* if emulation was due to access to shadowed page table
|
|
|
|
* and it failed try to unshadow page and re-enter the
|
|
|
|
* guest to let CPU execute the instruction.
|
|
|
|
*/
|
|
|
|
kvm_mmu_unprotect_page(vcpu->kvm, gpa_to_gfn(gpa));
|
2013-01-13 22:49:07 +07:00
|
|
|
|
|
|
|
/*
|
|
|
|
* If the access faults on its page table, it can not
|
|
|
|
* be fixed by unprotecting shadow page and it should
|
|
|
|
* be reported to userspace.
|
|
|
|
*/
|
|
|
|
return !write_fault_to_shadow_pgtable;
|
2010-07-08 16:41:12 +07:00
|
|
|
}
|
|
|
|
|
2011-09-22 16:02:48 +07:00
|
|
|
static bool retry_instruction(struct x86_emulate_ctxt *ctxt,
|
|
|
|
unsigned long cr2, int emulation_type)
|
|
|
|
{
|
|
|
|
struct kvm_vcpu *vcpu = emul_to_vcpu(ctxt);
|
|
|
|
unsigned long last_retry_eip, last_retry_addr, gpa = cr2;
|
|
|
|
|
|
|
|
last_retry_eip = vcpu->arch.last_retry_eip;
|
|
|
|
last_retry_addr = vcpu->arch.last_retry_addr;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* If the emulation is caused by #PF and it is non-page_table
|
|
|
|
* writing instruction, it means the VM-EXIT is caused by shadow
|
|
|
|
* page protected, we can zap the shadow page and retry this
|
|
|
|
* instruction directly.
|
|
|
|
*
|
|
|
|
* Note: if the guest uses a non-page-table modifying instruction
|
|
|
|
* on the PDE that points to the instruction, then we will unmap
|
|
|
|
* the instruction and go to an infinite loop. So, we cache the
|
|
|
|
* last retried eip and the last fault address, if we meet the eip
|
|
|
|
* and the address again, we can break out of the potential infinite
|
|
|
|
* loop.
|
|
|
|
*/
|
|
|
|
vcpu->arch.last_retry_eip = vcpu->arch.last_retry_addr = 0;
|
|
|
|
|
|
|
|
if (!(emulation_type & EMULTYPE_RETRY))
|
|
|
|
return false;
|
|
|
|
|
|
|
|
if (x86_page_table_writing_insn(ctxt))
|
|
|
|
return false;
|
|
|
|
|
|
|
|
if (ctxt->eip == last_retry_eip && last_retry_addr == cr2)
|
|
|
|
return false;
|
|
|
|
|
|
|
|
vcpu->arch.last_retry_eip = ctxt->eip;
|
|
|
|
vcpu->arch.last_retry_addr = cr2;
|
|
|
|
|
|
|
|
if (!vcpu->arch.mmu.direct_map)
|
|
|
|
gpa = kvm_mmu_gva_to_gpa_write(vcpu, cr2, NULL);
|
|
|
|
|
2013-01-13 22:44:12 +07:00
|
|
|
kvm_mmu_unprotect_page(vcpu->kvm, gpa_to_gfn(gpa));
|
2011-09-22 16:02:48 +07:00
|
|
|
|
|
|
|
return true;
|
|
|
|
}
|
|
|
|
|
2012-09-03 19:24:26 +07:00
|
|
|
static int complete_emulated_mmio(struct kvm_vcpu *vcpu);
|
|
|
|
static int complete_emulated_pio(struct kvm_vcpu *vcpu);
|
|
|
|
|
2015-05-07 16:36:11 +07:00
|
|
|
static void kvm_smm_changed(struct kvm_vcpu *vcpu)
|
2015-04-01 23:18:53 +07:00
|
|
|
{
|
2015-05-07 16:36:11 +07:00
|
|
|
if (!(vcpu->arch.hflags & HF_SMM_MASK)) {
|
2015-05-05 16:50:23 +07:00
|
|
|
/* This is a good place to trace that we are exiting SMM. */
|
|
|
|
trace_kvm_enter_smm(vcpu->vcpu_id, vcpu->arch.smbase, false);
|
|
|
|
|
2016-06-02 03:26:00 +07:00
|
|
|
/* Process a latched INIT or SMI, if any. */
|
|
|
|
kvm_make_request(KVM_REQ_EVENT, vcpu);
|
2015-05-07 16:36:11 +07:00
|
|
|
}
|
2015-05-18 20:03:39 +07:00
|
|
|
|
|
|
|
kvm_mmu_reset_context(vcpu);
|
2015-05-07 16:36:11 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
static void kvm_set_hflags(struct kvm_vcpu *vcpu, unsigned emul_flags)
|
|
|
|
{
|
|
|
|
unsigned changed = vcpu->arch.hflags ^ emul_flags;
|
|
|
|
|
2015-04-01 23:18:53 +07:00
|
|
|
vcpu->arch.hflags = emul_flags;
|
2015-05-07 16:36:11 +07:00
|
|
|
|
|
|
|
if (changed & HF_SMM_MASK)
|
|
|
|
kvm_smm_changed(vcpu);
|
2015-04-01 23:18:53 +07:00
|
|
|
}
|
|
|
|
|
2013-05-30 16:48:30 +07:00
|
|
|
static int kvm_vcpu_check_hw_bp(unsigned long addr, u32 type, u32 dr7,
|
|
|
|
unsigned long *db)
|
|
|
|
{
|
|
|
|
u32 dr6 = 0;
|
|
|
|
int i;
|
|
|
|
u32 enable, rwlen;
|
|
|
|
|
|
|
|
enable = dr7;
|
|
|
|
rwlen = dr7 >> 16;
|
|
|
|
for (i = 0; i < 4; i++, enable >>= 2, rwlen >>= 4)
|
|
|
|
if ((enable & 3) && (rwlen & 15) == type && db[i] == addr)
|
|
|
|
dr6 |= (1 << i);
|
|
|
|
return dr6;
|
|
|
|
}
|
|
|
|
|
2014-03-27 17:29:28 +07:00
|
|
|
static void kvm_vcpu_check_singlestep(struct kvm_vcpu *vcpu, unsigned long rflags, int *r)
|
2013-06-25 23:32:07 +07:00
|
|
|
{
|
|
|
|
struct kvm_run *kvm_run = vcpu->run;
|
|
|
|
|
|
|
|
/*
|
2014-03-27 17:29:28 +07:00
|
|
|
* rflags is the old, "raw" value of the flags. The new value has
|
|
|
|
* not been saved yet.
|
2013-06-25 23:32:07 +07:00
|
|
|
*
|
|
|
|
* This is correct even for TF set by the guest, because "the
|
|
|
|
* processor will not generate this exception after the instruction
|
|
|
|
* that sets the TF flag".
|
|
|
|
*/
|
|
|
|
if (unlikely(rflags & X86_EFLAGS_TF)) {
|
|
|
|
if (vcpu->guest_debug & KVM_GUESTDBG_SINGLESTEP) {
|
2014-07-15 21:37:46 +07:00
|
|
|
kvm_run->debug.arch.dr6 = DR6_BS | DR6_FIXED_1 |
|
|
|
|
DR6_RTM;
|
2013-06-25 23:32:07 +07:00
|
|
|
kvm_run->debug.arch.pc = vcpu->arch.singlestep_rip;
|
|
|
|
kvm_run->debug.arch.exception = DB_VECTOR;
|
|
|
|
kvm_run->exit_reason = KVM_EXIT_DEBUG;
|
|
|
|
*r = EMULATE_USER_EXIT;
|
|
|
|
} else {
|
|
|
|
vcpu->arch.emulate_ctxt.eflags &= ~X86_EFLAGS_TF;
|
|
|
|
/*
|
|
|
|
* "Certain debug exceptions may clear bit 0-3. The
|
|
|
|
* remaining contents of the DR6 register are never
|
|
|
|
* cleared by the processor".
|
|
|
|
*/
|
|
|
|
vcpu->arch.dr6 &= ~15;
|
2014-07-15 21:37:46 +07:00
|
|
|
vcpu->arch.dr6 |= DR6_BS | DR6_RTM;
|
2013-06-25 23:32:07 +07:00
|
|
|
kvm_queue_exception(vcpu, DB_VECTOR);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2013-05-30 16:48:30 +07:00
|
|
|
static bool kvm_vcpu_check_breakpoint(struct kvm_vcpu *vcpu, int *r)
|
|
|
|
{
|
|
|
|
if (unlikely(vcpu->guest_debug & KVM_GUESTDBG_USE_HW_BP) &&
|
|
|
|
(vcpu->arch.guest_debug_dr7 & DR7_BP_EN_MASK)) {
|
2014-11-02 16:54:45 +07:00
|
|
|
struct kvm_run *kvm_run = vcpu->run;
|
|
|
|
unsigned long eip = kvm_get_linear_rip(vcpu);
|
|
|
|
u32 dr6 = kvm_vcpu_check_hw_bp(eip, 0,
|
2013-05-30 16:48:30 +07:00
|
|
|
vcpu->arch.guest_debug_dr7,
|
|
|
|
vcpu->arch.eff_db);
|
|
|
|
|
|
|
|
if (dr6 != 0) {
|
2014-07-15 21:37:46 +07:00
|
|
|
kvm_run->debug.arch.dr6 = dr6 | DR6_FIXED_1 | DR6_RTM;
|
2014-11-02 16:54:45 +07:00
|
|
|
kvm_run->debug.arch.pc = eip;
|
2013-05-30 16:48:30 +07:00
|
|
|
kvm_run->debug.arch.exception = DB_VECTOR;
|
|
|
|
kvm_run->exit_reason = KVM_EXIT_DEBUG;
|
|
|
|
*r = EMULATE_USER_EXIT;
|
|
|
|
return true;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2014-07-17 05:19:31 +07:00
|
|
|
if (unlikely(vcpu->arch.dr7 & DR7_BP_EN_MASK) &&
|
|
|
|
!(kvm_get_rflags(vcpu) & X86_EFLAGS_RF)) {
|
2014-11-02 16:54:45 +07:00
|
|
|
unsigned long eip = kvm_get_linear_rip(vcpu);
|
|
|
|
u32 dr6 = kvm_vcpu_check_hw_bp(eip, 0,
|
2013-05-30 16:48:30 +07:00
|
|
|
vcpu->arch.dr7,
|
|
|
|
vcpu->arch.db);
|
|
|
|
|
|
|
|
if (dr6 != 0) {
|
|
|
|
vcpu->arch.dr6 &= ~15;
|
2014-07-15 21:37:46 +07:00
|
|
|
vcpu->arch.dr6 |= dr6 | DR6_RTM;
|
2013-05-30 16:48:30 +07:00
|
|
|
kvm_queue_exception(vcpu, DB_VECTOR);
|
|
|
|
*r = EMULATE_DONE;
|
|
|
|
return true;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
|
2010-12-21 17:12:02 +07:00
|
|
|
int x86_emulate_instruction(struct kvm_vcpu *vcpu,
|
|
|
|
unsigned long cr2,
|
2010-12-21 17:12:07 +07:00
|
|
|
int emulation_type,
|
|
|
|
void *insn,
|
|
|
|
int insn_len)
|
KVM: Portability: Move x86 emulation and mmio device hook to x86.c
This patch moves the following functions to from kvm_main.c to x86.c:
emulator_read/write_std, vcpu_find_pervcpu_dev, vcpu_find_mmio_dev,
emulator_read/write_emulated, emulator_write_phys,
emulator_write_emulated_onepage, emulator_cmpxchg_emulated,
get_setment_base, emulate_invlpg, emulate_clts, emulator_get/set_dr,
kvm_report_emulation_failure, emulate_instruction
The following data type is moved to x86.c:
struct x86_emulate_ops emulate_ops
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-31 00:44:21 +07:00
|
|
|
{
|
2010-04-28 23:15:43 +07:00
|
|
|
int r;
|
2011-05-29 19:53:48 +07:00
|
|
|
struct x86_emulate_ctxt *ctxt = &vcpu->arch.emulate_ctxt;
|
2011-03-31 17:06:41 +07:00
|
|
|
bool writeback = true;
|
2013-01-13 22:49:07 +07:00
|
|
|
bool write_fault_to_spt = vcpu->arch.write_fault_to_shadow_pgtable;
|
KVM: Portability: Move x86 emulation and mmio device hook to x86.c
This patch moves the following functions to from kvm_main.c to x86.c:
emulator_read/write_std, vcpu_find_pervcpu_dev, vcpu_find_mmio_dev,
emulator_read/write_emulated, emulator_write_phys,
emulator_write_emulated_onepage, emulator_cmpxchg_emulated,
get_setment_base, emulate_invlpg, emulate_clts, emulator_get/set_dr,
kvm_report_emulation_failure, emulate_instruction
The following data type is moved to x86.c:
struct x86_emulate_ops emulate_ops
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-31 00:44:21 +07:00
|
|
|
|
2013-01-13 22:49:07 +07:00
|
|
|
/*
|
|
|
|
* Clear write_fault_to_shadow_pgtable here to ensure it is
|
|
|
|
* never reused.
|
|
|
|
*/
|
|
|
|
vcpu->arch.write_fault_to_shadow_pgtable = false;
|
2008-07-03 18:59:22 +07:00
|
|
|
kvm_clear_exception_queue(vcpu);
|
2011-04-12 16:36:21 +07:00
|
|
|
|
KVM: x86 emulator: Only allow VMCALL/VMMCALL trapped by #UD
When executing a test program called "crashme", we found the KVM guest cannot
survive more than ten seconds, then encounterd kernel panic. The basic concept
of "crashme" is generating random assembly code and trying to execute it.
After some fixes on emulator insn validity judgment, we found it's hard to
get the current emulator handle the invalid instructions correctly, for the
#UD trap for hypercall patching caused troubles. The problem is, if the opcode
itself was OK, but combination of opcode and modrm_reg was invalid, and one
operand of the opcode was memory (SrcMem or DstMem), the emulator will fetch
the memory operand first rather than checking the validity, and may encounter
an error there. For example, ".byte 0xfe, 0x34, 0xcd" has this problem.
In the patch, we simply check that if the invalid opcode wasn't vmcall/vmmcall,
then return from emulate_instruction() and inject a #UD to guest. With the
patch, the guest had been running for more than 12 hours.
Signed-off-by: Feng (Eric) Liu <eric.e.liu@intel.com>
Signed-off-by: Sheng Yang <sheng.yang@intel.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2008-01-02 13:49:22 +07:00
|
|
|
if (!(emulation_type & EMULTYPE_NO_DECODE)) {
|
2010-08-16 04:47:01 +07:00
|
|
|
init_emulate_ctxt(vcpu);
|
2013-05-30 16:48:30 +07:00
|
|
|
|
|
|
|
/*
|
|
|
|
* We will reenter on the same instruction since
|
|
|
|
* we do not set complete_userspace_io. This does not
|
|
|
|
* handle watchpoints yet, those would be handled in
|
|
|
|
* the emulate_ops.
|
|
|
|
*/
|
|
|
|
if (kvm_vcpu_check_breakpoint(vcpu, &r))
|
|
|
|
return r;
|
|
|
|
|
2011-05-29 19:53:48 +07:00
|
|
|
ctxt->interruptibility = 0;
|
|
|
|
ctxt->have_exception = false;
|
2014-08-20 15:08:23 +07:00
|
|
|
ctxt->exception.vector = -1;
|
2011-05-29 19:53:48 +07:00
|
|
|
ctxt->perm_ok = false;
|
KVM: Portability: Move x86 emulation and mmio device hook to x86.c
This patch moves the following functions to from kvm_main.c to x86.c:
emulator_read/write_std, vcpu_find_pervcpu_dev, vcpu_find_mmio_dev,
emulator_read/write_emulated, emulator_write_phys,
emulator_write_emulated_onepage, emulator_cmpxchg_emulated,
get_setment_base, emulate_invlpg, emulate_clts, emulator_get/set_dr,
kvm_report_emulation_failure, emulate_instruction
The following data type is moved to x86.c:
struct x86_emulate_ops emulate_ops
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-31 00:44:21 +07:00
|
|
|
|
2013-09-22 21:44:52 +07:00
|
|
|
ctxt->ud = emulation_type & EMULTYPE_TRAP_UD;
|
2011-02-01 21:32:04 +07:00
|
|
|
|
2011-05-29 19:53:48 +07:00
|
|
|
r = x86_decode_insn(ctxt, insn, insn_len);
|
KVM: Portability: Move x86 emulation and mmio device hook to x86.c
This patch moves the following functions to from kvm_main.c to x86.c:
emulator_read/write_std, vcpu_find_pervcpu_dev, vcpu_find_mmio_dev,
emulator_read/write_emulated, emulator_write_phys,
emulator_write_emulated_onepage, emulator_cmpxchg_emulated,
get_setment_base, emulate_invlpg, emulate_clts, emulator_get/set_dr,
kvm_report_emulation_failure, emulate_instruction
The following data type is moved to x86.c:
struct x86_emulate_ops emulate_ops
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-31 00:44:21 +07:00
|
|
|
|
2010-04-11 17:05:16 +07:00
|
|
|
trace_kvm_emulate_insn_start(vcpu);
|
2007-11-18 20:17:51 +07:00
|
|
|
++vcpu->stat.insn_emulation;
|
2011-07-30 16:03:34 +07:00
|
|
|
if (r != EMULATION_OK) {
|
2011-02-01 21:32:04 +07:00
|
|
|
if (emulation_type & EMULTYPE_TRAP_UD)
|
|
|
|
return EMULATE_FAIL;
|
2013-04-11 16:10:51 +07:00
|
|
|
if (reexecute_instruction(vcpu, cr2, write_fault_to_spt,
|
|
|
|
emulation_type))
|
KVM: Portability: Move x86 emulation and mmio device hook to x86.c
This patch moves the following functions to from kvm_main.c to x86.c:
emulator_read/write_std, vcpu_find_pervcpu_dev, vcpu_find_mmio_dev,
emulator_read/write_emulated, emulator_write_phys,
emulator_write_emulated_onepage, emulator_cmpxchg_emulated,
get_setment_base, emulate_invlpg, emulate_clts, emulator_get/set_dr,
kvm_report_emulation_failure, emulate_instruction
The following data type is moved to x86.c:
struct x86_emulate_ops emulate_ops
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-31 00:44:21 +07:00
|
|
|
return EMULATE_DONE;
|
2010-05-10 15:16:56 +07:00
|
|
|
if (emulation_type & EMULTYPE_SKIP)
|
|
|
|
return EMULATE_FAIL;
|
|
|
|
return handle_emulation_failure(vcpu);
|
KVM: Portability: Move x86 emulation and mmio device hook to x86.c
This patch moves the following functions to from kvm_main.c to x86.c:
emulator_read/write_std, vcpu_find_pervcpu_dev, vcpu_find_mmio_dev,
emulator_read/write_emulated, emulator_write_phys,
emulator_write_emulated_onepage, emulator_cmpxchg_emulated,
get_setment_base, emulate_invlpg, emulate_clts, emulator_get/set_dr,
kvm_report_emulation_failure, emulate_instruction
The following data type is moved to x86.c:
struct x86_emulate_ops emulate_ops
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-31 00:44:21 +07:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2009-04-12 17:36:57 +07:00
|
|
|
if (emulation_type & EMULTYPE_SKIP) {
|
2011-06-01 19:34:25 +07:00
|
|
|
kvm_rip_write(vcpu, ctxt->_eip);
|
2014-07-21 18:37:26 +07:00
|
|
|
if (ctxt->eflags & X86_EFLAGS_RF)
|
|
|
|
kvm_set_rflags(vcpu, ctxt->eflags & ~X86_EFLAGS_RF);
|
2009-04-12 17:36:57 +07:00
|
|
|
return EMULATE_DONE;
|
|
|
|
}
|
|
|
|
|
2011-09-22 16:02:48 +07:00
|
|
|
if (retry_instruction(ctxt, cr2, emulation_type))
|
|
|
|
return EMULATE_DONE;
|
|
|
|
|
2011-03-31 17:06:41 +07:00
|
|
|
/* this is needed for vmware backdoor interface to work since it
|
2010-04-28 23:15:42 +07:00
|
|
|
changes registers values during IO operation */
|
2011-03-31 17:06:41 +07:00
|
|
|
if (vcpu->arch.emulate_regs_need_sync_from_vcpu) {
|
|
|
|
vcpu->arch.emulate_regs_need_sync_from_vcpu = false;
|
2012-08-28 03:46:17 +07:00
|
|
|
emulator_invalidate_register_cache(ctxt);
|
2011-03-31 17:06:41 +07:00
|
|
|
}
|
2010-04-28 23:15:42 +07:00
|
|
|
|
2010-03-18 20:20:26 +07:00
|
|
|
restart:
|
2011-05-29 19:53:48 +07:00
|
|
|
r = x86_emulate_insn(ctxt);
|
KVM: Portability: Move x86 emulation and mmio device hook to x86.c
This patch moves the following functions to from kvm_main.c to x86.c:
emulator_read/write_std, vcpu_find_pervcpu_dev, vcpu_find_mmio_dev,
emulator_read/write_emulated, emulator_write_phys,
emulator_write_emulated_onepage, emulator_cmpxchg_emulated,
get_setment_base, emulate_invlpg, emulate_clts, emulator_get/set_dr,
kvm_report_emulation_failure, emulate_instruction
The following data type is moved to x86.c:
struct x86_emulate_ops emulate_ops
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-31 00:44:21 +07:00
|
|
|
|
2011-04-04 17:39:24 +07:00
|
|
|
if (r == EMULATION_INTERCEPTED)
|
|
|
|
return EMULATE_DONE;
|
|
|
|
|
2010-08-25 16:47:43 +07:00
|
|
|
if (r == EMULATION_FAILED) {
|
2013-04-11 16:10:51 +07:00
|
|
|
if (reexecute_instruction(vcpu, cr2, write_fault_to_spt,
|
|
|
|
emulation_type))
|
2010-04-28 23:15:35 +07:00
|
|
|
return EMULATE_DONE;
|
|
|
|
|
2010-05-10 15:16:56 +07:00
|
|
|
return handle_emulation_failure(vcpu);
|
KVM: Portability: Move x86 emulation and mmio device hook to x86.c
This patch moves the following functions to from kvm_main.c to x86.c:
emulator_read/write_std, vcpu_find_pervcpu_dev, vcpu_find_mmio_dev,
emulator_read/write_emulated, emulator_write_phys,
emulator_write_emulated_onepage, emulator_cmpxchg_emulated,
get_setment_base, emulate_invlpg, emulate_clts, emulator_get/set_dr,
kvm_report_emulation_failure, emulate_instruction
The following data type is moved to x86.c:
struct x86_emulate_ops emulate_ops
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Acked-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
2007-10-31 00:44:21 +07:00
|
|
|
}
|
|
|
|
|
2011-05-29 19:53:48 +07:00
|
|
|
if (ctxt->have_exception) {
|
2010-08-25 16:47:43 +07:00
|
|
|
r = EMULATE_DONE;
|
2014-09-05 00:46:15 +07:00
|
|
|
if (inject_emulated_exception(vcpu))
|
|
|
|
return r;
|
2010-08-25 16:47:43 +07:00
|
|
|
} else if (vcpu->arch.pio.count) {
|
2013-08-27 20:41:43 +07:00
|
|
|
if (!vcpu->arch.pio.in) {
|
|
|
|
/* FIXME: return into emulator if single-stepping. */
|
2010-04-28 23:15:38 +07:00
|
|
|
vcpu->arch.pio.count = 0;
|
2013-08-27 20:41:43 +07:00
|
|
|
} else {
|
2011-03-31 17:06:41 +07:00
|
|
|
writeback = false;
|
2012-09-03 19:24:26 +07:00
|
|
|
vcpu->arch.complete_userspace_io = complete_emulated_pio;
|
|
|
|
}
|
2013-06-25 23:24:41 +07:00
|
|
|
r = EMULATE_USER_EXIT;
|
2011-03-31 17:06:41 +07:00
|
|
|
} else if (vcpu->mmio_needed) {
|
|
|
|
if (!vcpu->mmio_is_write)
|
|
|
|
writeback = false;
|
2013-06-25 23:24:41 +07:00
|
|
|
r = EMULATE_USER_EXIT;
|
2012-09-03 19:24:26 +07:00
|
|
|
vcpu->arch.complete_userspace_io = complete_emulated_mmio;
|
2011-03-31 17:06:41 +07:00
|
|
|
} else if (r == EMULATION_RESTART)
|
2010-03-18 20:20:26 +07:00
|
|
|
goto restart;
|
2010-08-25 16:47:43 +07:00
|
|
|
else
|
|
|
|
r = EMULATE_DONE;
|
2010-02-10 19:21:33 +07:00
|
|
|
|
2011-03-31 17:06:41 +07:00
|
|
|
if (writeback) {
|
2014-03-27 17:29:28 +07:00
|
|
|
unsigned long rflags = kvm_x86_ops->get_rflags(vcpu);
|
2011-05-29 19:53:48 +07:00
|
|
|
toggle_interruptibility(vcpu, ctxt->interruptibility);
|
2011-03-31 17:06:41 +07:00
|
|
|
vcpu->arch.emulate_regs_need_sync_to_vcpu = false;
|
2015-04-01 23:18:53 +07:00
|
|
|
if (vcpu->arch.hflags != ctxt->emul_flags)
|
|
|
|
kvm_set_hflags(vcpu, ctxt->emul_flags);
|
2011-05-29 19:53:48 +07:00
|
|
|
kvm_rip_write(vcpu, ctxt->eip);
|
2013-06-25 23:32:07 +07:00
|
|
|
if (r == EMULATE_DONE)
|
2014-03-27 17:29:28 +07:00
|
|
|
kvm_vcpu_check_singlestep(vcpu, rflags, &r);
|
2014-11-02 16:54:53 +07:00
|
|
|
if (!ctxt->have_exception ||
|
|
|
|
exception_type(ctxt->exception.vector) == EXCPT_TRAP)
|
|
|
|
__kvm_set_rflags(vcpu, ctxt->eflags);
|
2014-03-27 17:29:28 +07:00
|
|
|
|
|
|
|
/*
|
|
|
|
* For STI, interrupts are shadowed; so KVM_REQ_EVENT will
|
|
|
|
* do nothing, and it will be requested again as soon as
|
|
|
|
* the shadow expires. But we still need to check here,
|
|
|
|
* because POPF has no interrupt shadow.
|
|
|
|
*/
|
|
|
|
if (unlikely((ctxt->eflags & ~rflags) & X86_EFLAGS_IF))
|
|
|
|
kvm_make_request(KVM_REQ_EVENT, vcpu);
|
2011-03-31 17:06:41 +07:00
|
|
|
} else
|
|
|
|
vcpu->arch.emulate_regs_need_sync_to_vcpu = true;
|
2010-07-29 19:11:52 +07:00
|
|
|
|
|
|
|
return r;
|
2007-10-31 00:44:25 +07:00
|
|
|
}
|
2010-12-21 17:12:02 +07:00
|
|
|
EXPORT_SYMBOL_GPL(x86_emulate_instruction);
|
2007-10-31 00:44:25 +07:00
|
|
|
|
2010-03-18 20:20:23 +07:00
|
|
|
int kvm_fast_pio_out(struct kvm_vcpu *vcpu, int size, unsigned short port)
|
2007-10-31 00:44:25 +07:00
|
|
|
{
|
2010-03-18 20:20:23 +07:00
|
|
|
unsigned long val = kvm_register_read(vcpu, VCPU_REGS_RAX);
|
2011-04-20 17:37:53 +07:00
|
|
|
int ret = emulator_pio_out_emulated(&vcpu->arch.emulate_ctxt,
|
|
|
|
size, port, &val, 1);
|
2010-03-18 20:20:23 +07:00
|
|
|
/* do not return to emulator after return from userspace */
|
2010-03-18 20:20:24 +07:00
|
|
|
vcpu->arch.pio.count = 0;
|
2007-10-31 00:44:25 +07:00
|
|
|
return ret;
|
|
|
|
}
|
2010-03-18 20:20:23 +07:00
|
|
|
EXPORT_SYMBOL_GPL(kvm_fast_pio_out);
|
2007-10-31 00:44:25 +07:00
|
|
|
|
2016-07-14 00:16:33 +07:00
|
|
|
static int kvmclock_cpu_down_prep(unsigned int cpu)
|
2010-08-20 15:07:21 +07:00
|
|
|
{
|
2010-12-18 22:28:55 +07:00
|
|
|
__this_cpu_write(cpu_tsc_khz, 0);
|
2016-07-14 00:16:33 +07:00
|
|
|
return 0;
|
2010-08-20 15:07:21 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
static void tsc_khz_changed(void *data)
|
2009-02-04 23:52:04 +07:00
|
|
|
{
|
2010-08-20 15:07:21 +07:00
|
|
|
struct cpufreq_freqs *freq = data;
|
|
|
|
unsigned long khz = 0;
|
|
|
|
|
|
|
|
if (data)
|
|
|
|
khz = freq->new;
|
|
|
|
else if (!boot_cpu_has(X86_FEATURE_CONSTANT_TSC))
|
|
|
|
khz = cpufreq_quick_get(raw_smp_processor_id());
|
|
|
|
if (!khz)
|
|
|
|
khz = tsc_khz;
|
2010-12-18 22:28:55 +07:00
|
|
|
__this_cpu_write(cpu_tsc_khz, khz);
|
2009-02-04 23:52:04 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
static int kvmclock_cpufreq_notifier(struct notifier_block *nb, unsigned long val,
|
|
|
|
void *data)
|
|
|
|
{
|
|
|
|
struct cpufreq_freqs *freq = data;
|
|
|
|
struct kvm *kvm;
|
|
|
|
struct kvm_vcpu *vcpu;
|
|
|
|
int i, send_ipi = 0;
|
|
|
|
|
2010-08-20 15:07:21 +07:00
|
|
|
/*
|
|
|
|
* We allow guests to temporarily run on slowing clocks,
|
|
|
|
* provided we notify them after, or to run on accelerating
|
|
|
|
* clocks, provided we notify them before. Thus time never
|
|
|
|
* goes backwards.
|
|
|
|
*
|
|
|
|
* However, we have a problem. We can't atomically update
|
|
|
|
* the frequency of a given CPU from this function; it is
|
|
|
|
* merely a notifier, which can be called from any CPU.
|
|
|
|
* Changing the TSC frequency at arbitrary points in time
|
|
|
|
* requires a recomputation of local variables related to
|
|
|
|
* the TSC for each VCPU. We must flag these local variables
|
|
|
|
* to be updated and be sure the update takes place with the
|
|
|
|
* new frequency before any guests proceed.
|
|
|
|
*
|
|
|
|
* Unfortunately, the combination of hotplug CPU and frequency
|
|
|
|
* change creates an intractable locking scenario; the order
|
|
|
|
* of when these callouts happen is undefined with respect to
|
|
|
|
* CPU hotplug, and they can race with each other. As such,
|
|
|
|
* merely setting per_cpu(cpu_tsc_khz) = X during a hotadd is
|
|
|
|
* undefined; you can actually have a CPU frequency change take
|
|
|
|
* place in between the computation of X and the setting of the
|
|
|
|
* variable. To protect against this problem, all updates of
|
|
|
|
* the per_cpu tsc_khz variable are done in an interrupt
|
|
|
|
* protected IPI, and all callers wishing to update the value
|
|
|
|
* must wait for a synchronous IPI to complete (which is trivial
|
|
|
|
* if the caller is on the CPU already). This establishes the
|
|
|
|
* necessary total order on variable updates.
|
|
|
|
*
|
|
|
|
* Note that because a guest time update may take place
|
|
|
|
* anytime after the setting of the VCPU's request bit, the
|
|
|
|
* correct TSC value must be set before the request. However,
|
|
|
|
* to ensure the update actually makes it to any guest which
|
|
|
|
* starts running in hardware virtualization between the set
|
|
|
|
* and the acquisition of the spinlock, we must also ping the
|
|
|
|
* CPU after setting the request bit.
|
|
|
|
*
|
|
|
|
*/
|
|
|
|
|
2009-02-04 23:52:04 +07:00
|
|
|
if (val == CPUFREQ_PRECHANGE && freq->old > freq->new)
|
|
|
|
return 0;
|
|
|
|
if (val == CPUFREQ_POSTCHANGE && freq->old < freq->new)
|
|
|
|
return 0;
|
2010-08-20 15:07:21 +07:00
|
|
|
|
|
|
|
smp_call_function_single(freq->cpu, tsc_khz_changed, freq, 1);
|
2009-02-04 23:52:04 +07:00
|
|
|
|
2013-09-25 18:53:07 +07:00
|
|
|
spin_lock(&kvm_lock);
|
2009-02-04 23:52:04 +07:00
|
|
|
list_for_each_entry(kvm, &vm_list, vm_list) {
|
2009-06-09 19:56:29 +07:00
|
|
|
kvm_for_each_vcpu(i, vcpu, kvm) {
|
2009-02-04 23:52:04 +07:00
|
|
|
if (vcpu->cpu != freq->cpu)
|
|
|
|
continue;
|
2010-09-19 07:38:15 +07:00
|
|
|
kvm_make_request(KVM_REQ_CLOCK_UPDATE, vcpu);
|
2009-02-04 23:52:04 +07:00
|
|
|
if (vcpu->cpu != smp_processor_id())
|
2010-08-20 15:07:21 +07:00
|
|
|
send_ipi = 1;
|
2009-02-04 23:52:04 +07:00
|
|
|
}
|
|
|
|
}
|
2013-09-25 18:53:07 +07:00
|
|
|
spin_unlock(&kvm_lock);
|
2009-02-04 23:52:04 +07:00
|
|
|
|
|
|
|
if (freq->old < freq->new && send_ipi) {
|
|
|
|
/*
|
|
|
|
* We upscale the frequency. Must make the guest
|
|
|
|
* doesn't see old kvmclock values while running with
|
|
|
|
* the new frequency, otherwise we risk the guest sees
|
|
|
|
* time go backwards.
|
|
|
|
*
|
|
|
|
* In case we update the frequency for another cpu
|
|
|
|
* (which might be in guest context) send an interrupt
|
|
|
|
* to kick the cpu out of guest context. Next time
|
|
|
|
* guest context is entered kvmclock will be updated,
|
|
|
|
* so the guest will not see stale values.
|
|
|
|
*/
|
2010-08-20 15:07:21 +07:00
|
|
|
smp_call_function_single(freq->cpu, tsc_khz_changed, freq, 1);
|
2009-02-04 23:52:04 +07:00
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static struct notifier_block kvmclock_cpufreq_notifier_block = {
|
2010-08-20 15:07:21 +07:00
|
|
|
.notifier_call = kvmclock_cpufreq_notifier
|
|
|
|
};
|
|
|
|
|
2016-07-14 00:16:33 +07:00
|
|
|
static int kvmclock_cpu_online(unsigned int cpu)
|
2010-08-20 15:07:21 +07:00
|
|
|
{
|
2016-07-14 00:16:33 +07:00
|
|
|
tsc_khz_changed(NULL);
|
|
|
|
return 0;
|
2010-08-20 15:07:21 +07:00
|
|
|
}
|
|
|
|
|
2009-09-30 04:38:34 +07:00
|
|
|
static void kvm_timer_init(void)
|
|
|
|
{
|
|
|
|
int cpu;
|
|
|
|
|
2010-09-19 07:38:15 +07:00
|
|
|
max_tsc_khz = tsc_khz;
|
2014-03-11 03:39:01 +07:00
|
|
|
|
2009-09-30 04:38:34 +07:00
|
|
|
if (!boot_cpu_has(X86_FEATURE_CONSTANT_TSC)) {
|
2010-09-19 07:38:15 +07:00
|
|
|
#ifdef CONFIG_CPU_FREQ
|
|
|
|
struct cpufreq_policy policy;
|
|
|
|
memset(&policy, 0, sizeof(policy));
|
2010-12-16 17:16:34 +07:00
|
|
|
cpu = get_cpu();
|
|
|
|
cpufreq_get_policy(&policy, cpu);
|
2010-09-19 07:38:15 +07:00
|
|
|
if (policy.cpuinfo.max_freq)
|
|
|
|
max_tsc_khz = policy.cpuinfo.max_freq;
|
2010-12-16 17:16:34 +07:00
|
|
|
put_cpu();
|
2010-09-19 07:38:15 +07:00
|
|
|
#endif
|
2009-09-30 04:38:34 +07:00
|
|
|
cpufreq_register_notifier(&kvmclock_cpufreq_notifier_block,
|
|
|
|
CPUFREQ_TRANSITION_NOTIFIER);
|
|
|
|
}
|
2010-09-19 07:38:15 +07:00
|
|
|
pr_debug("kvm: max_tsc_khz = %ld\n", max_tsc_khz);
|
2014-03-11 03:39:01 +07:00
|
|
|
|
2016-07-14 00:16:33 +07:00
|
|
|
cpuhp_setup_state(CPUHP_AP_X86_KVM_CLK_ONLINE, "AP_X86_KVM_CLK_ONLINE",
|
|
|
|
kvmclock_cpu_online, kvmclock_cpu_down_prep);
|
2009-09-30 04:38:34 +07:00
|
|
|
}
|
|
|
|
|
2010-04-19 12:32:45 +07:00
|
|
|
static DEFINE_PER_CPU(struct kvm_vcpu *, current_vcpu);
|
|
|
|
|
2011-11-10 19:57:22 +07:00
|
|
|
int kvm_is_in_guest(void)
|
2010-04-19 12:32:45 +07:00
|
|
|
{
|
2011-10-20 14:34:01 +07:00
|
|
|
return __this_cpu_read(current_vcpu) != NULL;
|
2010-04-19 12:32:45 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
static int kvm_is_user_mode(void)
|
|
|
|
{
|
|
|
|
int user_mode = 3;
|
2010-04-20 09:13:58 +07:00
|
|
|
|
2011-10-20 14:34:01 +07:00
|
|
|
if (__this_cpu_read(current_vcpu))
|
|
|
|
user_mode = kvm_x86_ops->get_cpl(__this_cpu_read(current_vcpu));
|
2010-04-20 09:13:58 +07:00
|
|
|
|
2010-04-19 12:32:45 +07:00
|
|
|
return user_mode != 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static unsigned long kvm_get_guest_ip(void)
|
|
|
|
{
|
|
|
|
unsigned long ip = 0;
|
2010-04-20 09:13:58 +07:00
|
|
|
|
2011-10-20 14:34:01 +07:00
|
|
|
if (__this_cpu_read(current_vcpu))
|
|
|
|
ip = kvm_rip_read(__this_cpu_read(current_vcpu));
|
2010-04-20 09:13:58 +07:00
|
|
|
|
2010-04-19 12:32:45 +07:00
|
|
|
return ip;
|
|
|
|
}
|
|
|
|
|
|
|
|
static struct perf_guest_info_callbacks kvm_guest_cbs = {
|
|
|
|
.is_in_guest = kvm_is_in_guest,
|
|
|
|
.is_user_mode = kvm_is_user_mode,
|
|
|
|
.get_guest_ip = kvm_get_guest_ip,
|
|
|
|
};
|
|
|
|
|
|
|
|
void kvm_before_handle_nmi(struct kvm_vcpu *vcpu)
|
|
|
|
{
|
2011-10-20 14:34:01 +07:00
|
|
|
__this_cpu_write(current_vcpu, vcpu);
|
2010-04-19 12:32:45 +07:00
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(kvm_before_handle_nmi);
|
|
|
|
|
|
|
|
void kvm_after_handle_nmi(struct kvm_vcpu *vcpu)
|
|
|
|
{
|
2011-10-20 14:34:01 +07:00
|
|
|
__this_cpu_write(current_vcpu, NULL);
|
2010-04-19 12:32:45 +07:00
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(kvm_after_handle_nmi);
|
|
|
|
|
2011-07-12 02:33:44 +07:00
|
|
|
static void kvm_set_mmio_spte_mask(void)
|
|
|
|
{
|
|
|
|
u64 mask;
|
|
|
|
int maxphyaddr = boot_cpu_data.x86_phys_bits;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Set the reserved bits and the present bit of an paging-structure
|
|
|
|
* entry to generate page fault with PFER.RSV = 1.
|
|
|
|
*/
|
2013-06-07 15:51:23 +07:00
|
|
|
/* Mask the reserved physical address bits. */
|
2014-09-01 17:44:04 +07:00
|
|
|
mask = rsvd_bits(maxphyaddr, 51);
|
2013-06-07 15:51:23 +07:00
|
|
|
|
|
|
|
/* Bit 62 is always reserved for 32bit host. */
|
|
|
|
mask |= 0x3ull << 62;
|
|
|
|
|
|
|
|
/* Set the present bit. */
|
2011-07-12 02:33:44 +07:00
|
|
|
mask |= 1ull;
|
|
|
|
|
|
|
|
#ifdef CONFIG_X86_64
|
|
|
|
/*
|
|
|
|
* If reserved bit is not supported, clear the present bit to disable
|
|
|
|
* mmio page fault.
|
|
|
|
*/
|
|
|
|
if (maxphyaddr == 52)
|
|
|
|
mask &= ~1ull;
|
|
|
|
#endif
|
|
|
|
|
|
|
|
kvm_mmu_set_mmio_spte_mask(mask);
|
|
|
|
}
|
|
|
|
|
2012-11-28 08:29:00 +07:00
|
|
|
#ifdef CONFIG_X86_64
|
|
|
|
static void pvclock_gtod_update_fn(struct work_struct *work)
|
|
|
|
{
|
2012-11-28 08:29:01 +07:00
|
|
|
struct kvm *kvm;
|
|
|
|
|
|
|
|
struct kvm_vcpu *vcpu;
|
|
|
|
int i;
|
|
|
|
|
2013-09-25 18:53:07 +07:00
|
|
|
spin_lock(&kvm_lock);
|
2012-11-28 08:29:01 +07:00
|
|
|
list_for_each_entry(kvm, &vm_list, vm_list)
|
|
|
|
kvm_for_each_vcpu(i, vcpu, kvm)
|
2014-09-12 12:43:19 +07:00
|
|
|
kvm_make_request(KVM_REQ_MASTERCLOCK_UPDATE, vcpu);
|
2012-11-28 08:29:01 +07:00
|
|
|
atomic_set(&kvm_guest_has_master_clock, 0);
|
2013-09-25 18:53:07 +07:00
|
|
|
spin_unlock(&kvm_lock);
|
2012-11-28 08:29:00 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
static DECLARE_WORK(pvclock_gtod_work, pvclock_gtod_update_fn);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Notification about pvclock gtod data update.
|
|
|
|
*/
|
|
|
|
static int pvclock_gtod_notify(struct notifier_block *nb, unsigned long unused,
|
|
|
|
void *priv)
|
|
|
|
{
|
|
|
|
struct pvclock_gtod_data *gtod = &pvclock_gtod_data;
|
|
|
|
struct timekeeper *tk = priv;
|
|
|
|
|
|
|
|
update_pvclock_gtod(tk);
|
|
|
|
|
|
|
|
/* disable master clock if host does not trust, or does not
|
|
|
|
* use, TSC clocksource
|
|
|
|
*/
|
|
|
|
if (gtod->clock.vclock_mode != VCLOCK_TSC &&
|
|
|
|
atomic_read(&kvm_guest_has_master_clock) != 0)
|
|
|
|
queue_work(system_long_wq, &pvclock_gtod_work);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static struct notifier_block pvclock_gtod_notifier = {
|
|
|
|
.notifier_call = pvclock_gtod_notify,
|
|
|
|
};
|
|
|
|
#endif
|
|
|
|
|
2007-11-14 19:40:21 +07:00
|
|
|
int kvm_arch_init(void *opaque)
|
2007-10-10 22:16:19 +07:00
|
|
|
{
|
2009-09-30 04:38:34 +07:00
|
|
|
int r;
|
2013-06-27 01:36:23 +07:00
|
|
|
struct kvm_x86_ops *ops = opaque;
|
2007-11-14 19:40:21 +07:00
|
|
|
|
|
|
|
if (kvm_x86_ops) {
|
|
|
|
printk(KERN_ERR "kvm: already loaded the other module\n");
|
2007-11-18 19:43:21 +07:00
|
|
|
r = -EEXIST;
|
|
|
|
goto out;
|
2007-11-14 19:40:21 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
if (!ops->cpu_has_kvm_support()) {
|
|
|
|
printk(KERN_ERR "kvm: no hardware support\n");
|
2007-11-18 19:43:21 +07:00
|
|
|
r = -EOPNOTSUPP;
|
|
|
|
goto out;
|
2007-11-14 19:40:21 +07:00
|
|
|
}
|
|
|
|
if (ops->disabled_by_bios()) {
|
|
|
|
printk(KERN_ERR "kvm: disabled by bios\n");
|
2007-11-18 19:43:21 +07:00
|
|
|
r = -EOPNOTSUPP;
|
|
|
|
goto out;
|
2007-11-14 19:40:21 +07:00
|
|
|
}
|
|
|
|
|
2013-01-03 20:41:39 +07:00
|
|
|
r = -ENOMEM;
|
|
|
|
shared_msrs = alloc_percpu(struct kvm_shared_msrs);
|
|
|
|
if (!shared_msrs) {
|
|
|
|
printk(KERN_ERR "kvm: failed to allocate percpu kvm_shared_msrs\n");
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
2008-01-13 18:23:56 +07:00
|
|
|
r = kvm_mmu_module_init();
|
|
|
|
if (r)
|
2013-01-03 20:41:39 +07:00
|
|
|
goto out_free_percpu;
|
2008-01-13 18:23:56 +07:00
|
|
|
|
2011-07-12 02:33:44 +07:00
|
|
|
kvm_set_mmio_spte_mask();
|
2008-01-13 18:23:56 +07:00
|
|
|
|
2007-11-14 19:40:21 +07:00
|
|
|
kvm_x86_ops = ops;
|
2014-03-26 21:54:00 +07:00
|
|
|
|
2008-04-25 20:13:50 +07:00
|
|
|
kvm_mmu_set_mask_ptes(PT_USER_MASK, PT_ACCESSED_MASK,
|
2016-07-13 05:18:49 +07:00
|
|
|
PT_DIRTY_MASK, PT64_NX_MASK, 0,
|
|
|
|
PT_PRESENT_MASK);
|
2009-09-30 04:38:34 +07:00
|
|
|
kvm_timer_init();
|
2009-02-04 23:52:04 +07:00
|
|
|
|
2010-04-19 12:32:45 +07:00
|
|
|
perf_register_guest_info_callbacks(&kvm_guest_cbs);
|
|
|
|
|
2016-04-05 03:25:02 +07:00
|
|
|
if (boot_cpu_has(X86_FEATURE_XSAVE))
|
2010-06-10 10:27:12 +07:00
|
|
|
host_xcr0 = xgetbv(XCR_XFEATURE_ENABLED_MASK);
|
|
|
|
|
2012-08-05 19:58:30 +07:00
|
|
|
kvm_lapic_init();
|
2012-11-28 08:29:00 +07:00
|
|
|
#ifdef CONFIG_X86_64
|
|
|
|
pvclock_gtod_register_notifier(&pvclock_gtod_notifier);
|
|
|
|
#endif
|
|
|
|
|
2007-11-14 19:40:21 +07:00
|
|
|
return 0;
|
2007-11-18 19:43:21 +07:00
|
|
|
|
2013-01-03 20:41:39 +07:00
|
|
|
out_free_percpu:
|
|
|
|
free_percpu(shared_msrs);
|
2007-11-18 19:43:21 +07:00
|
|
|
out:
|
|
|
|
return r;
|
2007-10-10 22:16:19 +07:00
|
|
|
}
|
2007-11-01 05:24:24 +07:00
|
|
|
|
2007-11-14 19:40:21 +07:00
|
|
|
void kvm_arch_exit(void)
|
|
|
|
{
|
2010-04-19 12:32:45 +07:00
|
|
|
perf_unregister_guest_info_callbacks(&kvm_guest_cbs);
|
|
|
|
|
2009-04-18 00:24:58 +07:00
|
|
|
if (!boot_cpu_has(X86_FEATURE_CONSTANT_TSC))
|
|
|
|
cpufreq_unregister_notifier(&kvmclock_cpufreq_notifier_block,
|
|
|
|
CPUFREQ_TRANSITION_NOTIFIER);
|
2016-07-14 00:16:33 +07:00
|
|
|
cpuhp_remove_state_nocalls(CPUHP_AP_X86_KVM_CLK_ONLINE);
|
2012-11-28 08:29:00 +07:00
|
|
|
#ifdef CONFIG_X86_64
|
|
|
|
pvclock_gtod_unregister_notifier(&pvclock_gtod_notifier);
|
|
|
|
#endif
|
2007-11-14 19:40:21 +07:00
|
|
|
kvm_x86_ops = NULL;
|
2007-11-18 19:43:21 +07:00
|
|
|
kvm_mmu_module_exit();
|
2013-01-03 20:41:39 +07:00
|
|
|
free_percpu(shared_msrs);
|
2007-11-18 19:43:21 +07:00
|
|
|
}
|
2007-11-14 19:40:21 +07:00
|
|
|
|
2015-03-03 02:43:31 +07:00
|
|
|
int kvm_vcpu_halt(struct kvm_vcpu *vcpu)
|
2007-11-01 05:24:24 +07:00
|
|
|
{
|
|
|
|
++vcpu->stat.halt_exits;
|
2015-07-29 17:05:37 +07:00
|
|
|
if (lapic_in_kernel(vcpu)) {
|
2008-04-13 21:54:35 +07:00
|
|
|
vcpu->arch.mp_state = KVM_MP_STATE_HALTED;
|
2007-11-01 05:24:24 +07:00
|
|
|
return 1;
|
|
|
|
} else {
|
|
|
|
vcpu->run->exit_reason = KVM_EXIT_HLT;
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
}
|
2015-03-03 02:43:31 +07:00
|
|
|
EXPORT_SYMBOL_GPL(kvm_vcpu_halt);
|
|
|
|
|
|
|
|
int kvm_emulate_halt(struct kvm_vcpu *vcpu)
|
|
|
|
{
|
|
|
|
kvm_x86_ops->skip_emulated_instruction(vcpu);
|
|
|
|
return kvm_vcpu_halt(vcpu);
|
|
|
|
}
|
2007-11-01 05:24:24 +07:00
|
|
|
EXPORT_SYMBOL_GPL(kvm_emulate_halt);
|
|
|
|
|
2013-08-26 15:48:34 +07:00
|
|
|
/*
|
|
|
|
* kvm_pv_kick_cpu_op: Kick a vcpu.
|
|
|
|
*
|
|
|
|
* @apicid - apicid of vcpu to be kicked.
|
|
|
|
*/
|
|
|
|
static void kvm_pv_kick_cpu_op(struct kvm *kvm, unsigned long flags, int apicid)
|
|
|
|
{
|
2013-08-26 15:48:35 +07:00
|
|
|
struct kvm_lapic_irq lapic_irq;
|
2013-08-26 15:48:34 +07:00
|
|
|
|
2013-08-26 15:48:35 +07:00
|
|
|
lapic_irq.shorthand = 0;
|
|
|
|
lapic_irq.dest_mode = 0;
|
|
|
|
lapic_irq.dest_id = apicid;
|
2015-03-19 08:26:03 +07:00
|
|
|
lapic_irq.msi_redir_hint = false;
|
2013-08-26 15:48:34 +07:00
|
|
|
|
2013-08-26 15:48:35 +07:00
|
|
|
lapic_irq.delivery_mode = APIC_DM_REMRD;
|
2015-03-13 16:39:44 +07:00
|
|
|
kvm_irq_delivery_to_apic(kvm, NULL, &lapic_irq, NULL);
|
2013-08-26 15:48:34 +07:00
|
|
|
}
|
|
|
|
|
2015-11-10 19:36:33 +07:00
|
|
|
void kvm_vcpu_deactivate_apicv(struct kvm_vcpu *vcpu)
|
|
|
|
{
|
|
|
|
vcpu->arch.apicv_active = false;
|
|
|
|
kvm_x86_ops->refresh_apicv_exec_ctrl(vcpu);
|
|
|
|
}
|
|
|
|
|
2007-11-01 05:24:24 +07:00
|
|
|
int kvm_emulate_hypercall(struct kvm_vcpu *vcpu)
|
|
|
|
{
|
|
|
|
unsigned long nr, a0, a1, a2, a3, ret;
|
2014-06-18 21:19:24 +07:00
|
|
|
int op_64_bit, r = 1;
|
2007-11-01 05:24:24 +07:00
|
|
|
|
2015-03-03 02:43:31 +07:00
|
|
|
kvm_x86_ops->skip_emulated_instruction(vcpu);
|
|
|
|
|
2010-01-17 20:51:22 +07:00
|
|
|
if (kvm_hv_hypercall_enabled(vcpu->kvm))
|
|
|
|
return kvm_hv_hypercall(vcpu);
|
|
|
|
|
2008-06-28 00:58:02 +07:00
|
|
|
nr = kvm_register_read(vcpu, VCPU_REGS_RAX);
|
|
|
|
a0 = kvm_register_read(vcpu, VCPU_REGS_RBX);
|
|
|
|
a1 = kvm_register_read(vcpu, VCPU_REGS_RCX);
|
|
|
|
a2 = kvm_register_read(vcpu, VCPU_REGS_RDX);
|
|
|
|
a3 = kvm_register_read(vcpu, VCPU_REGS_RSI);
|
2007-11-01 05:24:24 +07:00
|
|
|
|
2009-06-17 19:22:14 +07:00
|
|
|
trace_kvm_hypercall(nr, a0, a1, a2, a3);
|
2008-04-11 02:31:10 +07:00
|
|
|
|
2014-06-18 21:19:24 +07:00
|
|
|
op_64_bit = is_64_bit_mode(vcpu);
|
|
|
|
if (!op_64_bit) {
|
2007-11-01 05:24:24 +07:00
|
|
|
nr &= 0xFFFFFFFF;
|
|
|
|
a0 &= 0xFFFFFFFF;
|
|
|
|
a1 &= 0xFFFFFFFF;
|
|
|
|
a2 &= 0xFFFFFFFF;
|
|
|
|
a3 &= 0xFFFFFFFF;
|
|
|
|
}
|
|
|
|
|
2009-08-03 23:43:28 +07:00
|
|
|
if (kvm_x86_ops->get_cpl(vcpu) != 0) {
|
|
|
|
ret = -KVM_EPERM;
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
2007-11-01 05:24:24 +07:00
|
|
|
switch (nr) {
|
2007-10-25 21:52:32 +07:00
|
|
|
case KVM_HC_VAPIC_POLL_IRQ:
|
|
|
|
ret = 0;
|
|
|
|
break;
|
2013-08-26 15:48:34 +07:00
|
|
|
case KVM_HC_KICK_CPU:
|
|
|
|
kvm_pv_kick_cpu_op(vcpu->kvm, a0, a1);
|
|
|
|
ret = 0;
|
|
|
|
break;
|
2007-11-01 05:24:24 +07:00
|
|
|
default:
|
|
|
|
ret = -KVM_ENOSYS;
|
|
|
|
break;
|
|
|
|
}
|
2009-08-03 23:43:28 +07:00
|
|
|
out:
|
2014-06-18 21:19:24 +07:00
|
|
|
if (!op_64_bit)
|
|
|
|
ret = (u32)ret;
|
2008-06-28 00:58:02 +07:00
|
|
|
kvm_register_write(vcpu, VCPU_REGS_RAX, ret);
|
2008-02-21 02:30:30 +07:00
|
|
|
++vcpu->stat.hypercalls;
|
2008-02-23 00:21:37 +07:00
|
|
|
return r;
|
2007-11-01 05:24:24 +07:00
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(kvm_emulate_hypercall);
|
|
|
|
|
2012-09-20 12:43:17 +07:00
|
|
|
static int emulator_fix_hypercall(struct x86_emulate_ctxt *ctxt)
|
2007-11-01 05:24:24 +07:00
|
|
|
{
|
2011-04-20 19:47:13 +07:00
|
|
|
struct kvm_vcpu *vcpu = emul_to_vcpu(ctxt);
|
2007-11-01 05:24:24 +07:00
|
|
|
char instruction[3];
|
2008-06-28 00:58:02 +07:00
|
|
|
unsigned long rip = kvm_rip_read(vcpu);
|
2007-11-01 05:24:24 +07:00
|
|
|
|
|
|
|
kvm_x86_ops->patch_hypercall(vcpu, instruction);
|
|
|
|
|
2011-05-29 19:53:48 +07:00
|
|
|
return emulator_write_emulated(ctxt, rip, instruction, 3, NULL);
|
2007-11-01 05:24:24 +07:00
|
|
|
}
|
|
|
|
|
2009-08-24 15:10:17 +07:00
|
|
|
static int dm_request_for_irq_injection(struct kvm_vcpu *vcpu)
|
2007-11-02 02:16:10 +07:00
|
|
|
{
|
2015-11-17 06:26:00 +07:00
|
|
|
return vcpu->run->request_interrupt_window &&
|
|
|
|
likely(!pic_in_kernel(vcpu->kvm));
|
2007-11-02 02:16:10 +07:00
|
|
|
}
|
|
|
|
|
2009-08-24 15:10:17 +07:00
|
|
|
static void post_kvm_run_save(struct kvm_vcpu *vcpu)
|
2007-11-02 02:16:10 +07:00
|
|
|
{
|
2009-08-24 15:10:17 +07:00
|
|
|
struct kvm_run *kvm_run = vcpu->run;
|
|
|
|
|
2009-10-05 18:07:21 +07:00
|
|
|
kvm_run->if_flag = (kvm_get_rflags(vcpu) & X86_EFLAGS_IF) != 0;
|
2015-04-01 20:06:40 +07:00
|
|
|
kvm_run->flags = is_smm(vcpu) ? KVM_RUN_X86_SMM : 0;
|
2008-02-24 16:20:43 +07:00
|
|
|
kvm_run->cr8 = kvm_get_cr8(vcpu);
|
2007-11-02 02:16:10 +07:00
|
|
|
kvm_run->apic_base = kvm_get_apic_base(vcpu);
|
2015-11-17 23:32:05 +07:00
|
|
|
kvm_run->ready_for_interrupt_injection =
|
|
|
|
pic_in_kernel(vcpu->kvm) ||
|
2015-11-17 06:26:00 +07:00
|
|
|
kvm_vcpu_ready_for_interrupt_injection(vcpu);
|
2007-11-02 02:16:10 +07:00
|
|
|
}
|
|
|
|
|
2009-04-21 21:45:08 +07:00
|
|
|
static void update_cr8_intercept(struct kvm_vcpu *vcpu)
|
|
|
|
{
|
|
|
|
int max_irr, tpr;
|
|
|
|
|
|
|
|
if (!kvm_x86_ops->update_cr8_intercept)
|
|
|
|
return;
|
|
|
|
|
2016-01-08 19:48:51 +07:00
|
|
|
if (!lapic_in_kernel(vcpu))
|
2009-08-18 02:49:40 +07:00
|
|
|
return;
|
|
|
|
|
2015-11-10 19:36:33 +07:00
|
|
|
if (vcpu->arch.apicv_active)
|
|
|
|
return;
|
|
|
|
|
2009-05-11 17:35:54 +07:00
|
|
|
if (!vcpu->arch.apic->vapic_addr)
|
|
|
|
max_irr = kvm_lapic_find_highest_irr(vcpu);
|
|
|
|
else
|
|
|
|
max_irr = -1;
|
2009-04-21 21:45:08 +07:00
|
|
|
|
|
|
|
if (max_irr != -1)
|
|
|
|
max_irr >>= 4;
|
|
|
|
|
|
|
|
tpr = kvm_lapic_get_cr8(vcpu);
|
|
|
|
|
|
|
|
kvm_x86_ops->update_cr8_intercept(vcpu, tpr, max_irr);
|
|
|
|
}
|
|
|
|
|
2014-03-08 02:03:12 +07:00
|
|
|
static int inject_pending_event(struct kvm_vcpu *vcpu, bool req_int_win)
|
2009-04-21 21:45:08 +07:00
|
|
|
{
|
2014-03-08 02:03:12 +07:00
|
|
|
int r;
|
|
|
|
|
2009-04-21 21:45:08 +07:00
|
|
|
/* try to reinject previous events if any */
|
2009-07-09 19:33:51 +07:00
|
|
|
if (vcpu->arch.exception.pending) {
|
2010-03-11 18:01:59 +07:00
|
|
|
trace_kvm_inj_exception(vcpu->arch.exception.nr,
|
|
|
|
vcpu->arch.exception.has_error_code,
|
|
|
|
vcpu->arch.exception.error_code);
|
2014-07-24 18:51:24 +07:00
|
|
|
|
|
|
|
if (exception_type(vcpu->arch.exception.nr) == EXCPT_FAULT)
|
|
|
|
__kvm_set_rflags(vcpu, kvm_get_rflags(vcpu) |
|
|
|
|
X86_EFLAGS_RF);
|
|
|
|
|
2014-10-01 00:49:14 +07:00
|
|
|
if (vcpu->arch.exception.nr == DB_VECTOR &&
|
|
|
|
(vcpu->arch.dr7 & DR7_GD)) {
|
|
|
|
vcpu->arch.dr7 &= ~DR7_GD;
|
|
|
|
kvm_update_dr7(vcpu);
|
|
|
|
}
|
|
|
|
|
2009-07-09 19:33:51 +07:00
|
|
|
kvm_x86_ops->queue_exception(vcpu, vcpu->arch.exception.nr,
|
|
|
|
vcpu->arch.exception.has_error_code,
|
2010-04-22 17:33:13 +07:00
|
|
|
vcpu->arch.exception.error_code,
|
|
|
|
vcpu->arch.exception.reinject);
|
2014-03-08 02:03:12 +07:00
|
|
|
return 0;
|
2009-07-09 19:33:51 +07:00
|
|
|
}
|
|
|
|
|
2009-04-21 21:45:08 +07:00
|
|
|
if (vcpu->arch.nmi_injected) {
|
|
|
|
kvm_x86_ops->set_nmi(vcpu);
|
2014-03-08 02:03:12 +07:00
|
|
|
return 0;
|
2009-04-21 21:45:08 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
if (vcpu->arch.interrupt.pending) {
|
2009-05-11 17:35:50 +07:00
|
|
|
kvm_x86_ops->set_irq(vcpu);
|
2014-03-08 02:03:12 +07:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (is_guest_mode(vcpu) && kvm_x86_ops->check_nested_events) {
|
|
|
|
r = kvm_x86_ops->check_nested_events(vcpu, req_int_win);
|
|
|
|
if (r != 0)
|
|
|
|
return r;
|
2009-04-21 21:45:08 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
/* try to inject new event if pending */
|
2016-06-02 03:26:00 +07:00
|
|
|
if (vcpu->arch.smi_pending && !is_smm(vcpu)) {
|
|
|
|
vcpu->arch.smi_pending = false;
|
2016-06-02 03:26:01 +07:00
|
|
|
enter_smm(vcpu);
|
2016-06-02 03:26:00 +07:00
|
|
|
} else if (vcpu->arch.nmi_pending && kvm_x86_ops->nmi_allowed(vcpu)) {
|
KVM: x86: Inject pending interrupt even if pending nmi exist
Non maskable interrupts (NMI) are preferred to interrupts in current
implementation. If a NMI is pending and NMI is blocked by the result
of nmi_allowed(), pending interrupt is not injected and
enable_irq_window() is not executed, even if interrupts injection is
allowed.
In old kernel (e.g. 2.6.32), schedule() is often called in NMI context.
In this case, interrupts are needed to execute iret that intends end
of NMI. The flag of blocking new NMI is not cleared until the guest
execute the iret, and interrupts are blocked by pending NMI. Due to
this, iret can't be invoked in the guest, and the guest is starved
until block is cleared by some events (e.g. canceling injection).
This patch injects pending interrupts, when it's allowed, even if NMI
is blocked. And, If an interrupts is pending after executing
inject_pending_event(), enable_irq_window() is executed regardless of
NMI pending counter.
Cc: stable@vger.kernel.org
Signed-off-by: Yuki Shibuya <shibuya.yk@ncos.nec.co.jp>
Suggested-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2016-03-24 12:17:03 +07:00
|
|
|
--vcpu->arch.nmi_pending;
|
|
|
|
vcpu->arch.nmi_injected = true;
|
|
|
|
kvm_x86_ops->set_nmi(vcpu);
|
2013-01-25 09:18:51 +07:00
|
|
|
} else if (kvm_cpu_has_injectable_intr(vcpu)) {
|
2014-07-08 11:30:23 +07:00
|
|
|
/*
|
|
|
|
* Because interrupts can be injected asynchronously, we are
|
|
|
|
* calling check_nested_events again here to avoid a race condition.
|
|
|
|
* See https://lkml.org/lkml/2014/7/2/60 for discussion about this
|
|
|
|
* proposal and current concerns. Perhaps we should be setting
|
|
|
|
* KVM_REQ_EVENT only on certain events and not unconditionally?
|
|
|
|
*/
|
|
|
|
if (is_guest_mode(vcpu) && kvm_x86_ops->check_nested_events) {
|
|
|
|
r = kvm_x86_ops->check_nested_events(vcpu, req_int_win);
|
|
|
|
if (r != 0)
|
|
|
|
return r;
|
|
|
|
}
|
2009-04-21 21:45:08 +07:00
|
|
|
if (kvm_x86_ops->interrupt_allowed(vcpu)) {
|
2009-05-11 17:35:50 +07:00
|
|
|
kvm_queue_interrupt(vcpu, kvm_cpu_get_interrupt(vcpu),
|
|
|
|
false);
|
|
|
|
kvm_x86_ops->set_irq(vcpu);
|
2009-04-21 21:45:08 +07:00
|
|
|
}
|
|
|
|
}
|
2016-06-02 03:26:01 +07:00
|
|
|
|
2014-03-08 02:03:12 +07:00
|
|
|
return 0;
|
2009-04-21 21:45:08 +07:00
|
|
|
}
|
|
|
|
|
2011-09-20 17:43:14 +07:00
|
|
|
static void process_nmi(struct kvm_vcpu *vcpu)
|
|
|
|
{
|
|
|
|
unsigned limit = 2;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* x86 is limited to one NMI running, and one NMI pending after it.
|
|
|
|
* If an NMI is already in progress, limit further NMIs to just one.
|
|
|
|
* Otherwise, allow two (and we'll inject the first one immediately).
|
|
|
|
*/
|
|
|
|
if (kvm_x86_ops->get_nmi_mask(vcpu) || vcpu->arch.nmi_injected)
|
|
|
|
limit = 1;
|
|
|
|
|
|
|
|
vcpu->arch.nmi_pending += atomic_xchg(&vcpu->arch.nmi_queued, 0);
|
|
|
|
vcpu->arch.nmi_pending = min(vcpu->arch.nmi_pending, limit);
|
|
|
|
kvm_make_request(KVM_REQ_EVENT, vcpu);
|
|
|
|
}
|
|
|
|
|
2015-05-05 16:50:23 +07:00
|
|
|
#define put_smstate(type, buf, offset, val) \
|
|
|
|
*(type *)((buf) + (offset) - 0x7e00) = val
|
|
|
|
|
2016-06-02 03:26:01 +07:00
|
|
|
static u32 enter_smm_get_segment_flags(struct kvm_segment *seg)
|
2015-05-05 16:50:23 +07:00
|
|
|
{
|
|
|
|
u32 flags = 0;
|
|
|
|
flags |= seg->g << 23;
|
|
|
|
flags |= seg->db << 22;
|
|
|
|
flags |= seg->l << 21;
|
|
|
|
flags |= seg->avl << 20;
|
|
|
|
flags |= seg->present << 15;
|
|
|
|
flags |= seg->dpl << 13;
|
|
|
|
flags |= seg->s << 12;
|
|
|
|
flags |= seg->type << 8;
|
|
|
|
return flags;
|
|
|
|
}
|
|
|
|
|
2016-06-02 03:26:01 +07:00
|
|
|
static void enter_smm_save_seg_32(struct kvm_vcpu *vcpu, char *buf, int n)
|
2015-05-05 16:50:23 +07:00
|
|
|
{
|
|
|
|
struct kvm_segment seg;
|
|
|
|
int offset;
|
|
|
|
|
|
|
|
kvm_get_segment(vcpu, &seg, n);
|
|
|
|
put_smstate(u32, buf, 0x7fa8 + n * 4, seg.selector);
|
|
|
|
|
|
|
|
if (n < 3)
|
|
|
|
offset = 0x7f84 + n * 12;
|
|
|
|
else
|
|
|
|
offset = 0x7f2c + (n - 3) * 12;
|
|
|
|
|
|
|
|
put_smstate(u32, buf, offset + 8, seg.base);
|
|
|
|
put_smstate(u32, buf, offset + 4, seg.limit);
|
2016-06-02 03:26:01 +07:00
|
|
|
put_smstate(u32, buf, offset, enter_smm_get_segment_flags(&seg));
|
2015-05-05 16:50:23 +07:00
|
|
|
}
|
|
|
|
|
2015-09-06 20:35:41 +07:00
|
|
|
#ifdef CONFIG_X86_64
|
2016-06-02 03:26:01 +07:00
|
|
|
static void enter_smm_save_seg_64(struct kvm_vcpu *vcpu, char *buf, int n)
|
2015-05-05 16:50:23 +07:00
|
|
|
{
|
|
|
|
struct kvm_segment seg;
|
|
|
|
int offset;
|
|
|
|
u16 flags;
|
|
|
|
|
|
|
|
kvm_get_segment(vcpu, &seg, n);
|
|
|
|
offset = 0x7e00 + n * 16;
|
|
|
|
|
2016-06-02 03:26:01 +07:00
|
|
|
flags = enter_smm_get_segment_flags(&seg) >> 8;
|
2015-05-05 16:50:23 +07:00
|
|
|
put_smstate(u16, buf, offset, seg.selector);
|
|
|
|
put_smstate(u16, buf, offset + 2, flags);
|
|
|
|
put_smstate(u32, buf, offset + 4, seg.limit);
|
|
|
|
put_smstate(u64, buf, offset + 8, seg.base);
|
|
|
|
}
|
2015-09-06 20:35:41 +07:00
|
|
|
#endif
|
2015-05-05 16:50:23 +07:00
|
|
|
|
2016-06-02 03:26:01 +07:00
|
|
|
static void enter_smm_save_state_32(struct kvm_vcpu *vcpu, char *buf)
|
2015-05-05 16:50:23 +07:00
|
|
|
{
|
|
|
|
struct desc_ptr dt;
|
|
|
|
struct kvm_segment seg;
|
|
|
|
unsigned long val;
|
|
|
|
int i;
|
|
|
|
|
|
|
|
put_smstate(u32, buf, 0x7ffc, kvm_read_cr0(vcpu));
|
|
|
|
put_smstate(u32, buf, 0x7ff8, kvm_read_cr3(vcpu));
|
|
|
|
put_smstate(u32, buf, 0x7ff4, kvm_get_rflags(vcpu));
|
|
|
|
put_smstate(u32, buf, 0x7ff0, kvm_rip_read(vcpu));
|
|
|
|
|
|
|
|
for (i = 0; i < 8; i++)
|
|
|
|
put_smstate(u32, buf, 0x7fd0 + i * 4, kvm_register_read(vcpu, i));
|
|
|
|
|
|
|
|
kvm_get_dr(vcpu, 6, &val);
|
|
|
|
put_smstate(u32, buf, 0x7fcc, (u32)val);
|
|
|
|
kvm_get_dr(vcpu, 7, &val);
|
|
|
|
put_smstate(u32, buf, 0x7fc8, (u32)val);
|
|
|
|
|
|
|
|
kvm_get_segment(vcpu, &seg, VCPU_SREG_TR);
|
|
|
|
put_smstate(u32, buf, 0x7fc4, seg.selector);
|
|
|
|
put_smstate(u32, buf, 0x7f64, seg.base);
|
|
|
|
put_smstate(u32, buf, 0x7f60, seg.limit);
|
2016-06-02 03:26:01 +07:00
|
|
|
put_smstate(u32, buf, 0x7f5c, enter_smm_get_segment_flags(&seg));
|
2015-05-05 16:50:23 +07:00
|
|
|
|
|
|
|
kvm_get_segment(vcpu, &seg, VCPU_SREG_LDTR);
|
|
|
|
put_smstate(u32, buf, 0x7fc0, seg.selector);
|
|
|
|
put_smstate(u32, buf, 0x7f80, seg.base);
|
|
|
|
put_smstate(u32, buf, 0x7f7c, seg.limit);
|
2016-06-02 03:26:01 +07:00
|
|
|
put_smstate(u32, buf, 0x7f78, enter_smm_get_segment_flags(&seg));
|
2015-05-05 16:50:23 +07:00
|
|
|
|
|
|
|
kvm_x86_ops->get_gdt(vcpu, &dt);
|
|
|
|
put_smstate(u32, buf, 0x7f74, dt.address);
|
|
|
|
put_smstate(u32, buf, 0x7f70, dt.size);
|
|
|
|
|
|
|
|
kvm_x86_ops->get_idt(vcpu, &dt);
|
|
|
|
put_smstate(u32, buf, 0x7f58, dt.address);
|
|
|
|
put_smstate(u32, buf, 0x7f54, dt.size);
|
|
|
|
|
|
|
|
for (i = 0; i < 6; i++)
|
2016-06-02 03:26:01 +07:00
|
|
|
enter_smm_save_seg_32(vcpu, buf, i);
|
2015-05-05 16:50:23 +07:00
|
|
|
|
|
|
|
put_smstate(u32, buf, 0x7f14, kvm_read_cr4(vcpu));
|
|
|
|
|
|
|
|
/* revision id */
|
|
|
|
put_smstate(u32, buf, 0x7efc, 0x00020000);
|
|
|
|
put_smstate(u32, buf, 0x7ef8, vcpu->arch.smbase);
|
|
|
|
}
|
|
|
|
|
2016-06-02 03:26:01 +07:00
|
|
|
static void enter_smm_save_state_64(struct kvm_vcpu *vcpu, char *buf)
|
2015-05-05 16:50:23 +07:00
|
|
|
{
|
|
|
|
#ifdef CONFIG_X86_64
|
|
|
|
struct desc_ptr dt;
|
|
|
|
struct kvm_segment seg;
|
|
|
|
unsigned long val;
|
|
|
|
int i;
|
|
|
|
|
|
|
|
for (i = 0; i < 16; i++)
|
|
|
|
put_smstate(u64, buf, 0x7ff8 - i * 8, kvm_register_read(vcpu, i));
|
|
|
|
|
|
|
|
put_smstate(u64, buf, 0x7f78, kvm_rip_read(vcpu));
|
|
|
|
put_smstate(u32, buf, 0x7f70, kvm_get_rflags(vcpu));
|
|
|
|
|
|
|
|
kvm_get_dr(vcpu, 6, &val);
|
|
|
|
put_smstate(u64, buf, 0x7f68, val);
|
|
|
|
kvm_get_dr(vcpu, 7, &val);
|
|
|
|
put_smstate(u64, buf, 0x7f60, val);
|
|
|
|
|
|
|
|
put_smstate(u64, buf, 0x7f58, kvm_read_cr0(vcpu));
|
|
|
|
put_smstate(u64, buf, 0x7f50, kvm_read_cr3(vcpu));
|
|
|
|
put_smstate(u64, buf, 0x7f48, kvm_read_cr4(vcpu));
|
|
|
|
|
|
|
|
put_smstate(u32, buf, 0x7f00, vcpu->arch.smbase);
|
|
|
|
|
|
|
|
/* revision id */
|
|
|
|
put_smstate(u32, buf, 0x7efc, 0x00020064);
|
|
|
|
|
|
|
|
put_smstate(u64, buf, 0x7ed0, vcpu->arch.efer);
|
|
|
|
|
|
|
|
kvm_get_segment(vcpu, &seg, VCPU_SREG_TR);
|
|
|
|
put_smstate(u16, buf, 0x7e90, seg.selector);
|
2016-06-02 03:26:01 +07:00
|
|
|
put_smstate(u16, buf, 0x7e92, enter_smm_get_segment_flags(&seg) >> 8);
|
2015-05-05 16:50:23 +07:00
|
|
|
put_smstate(u32, buf, 0x7e94, seg.limit);
|
|
|
|
put_smstate(u64, buf, 0x7e98, seg.base);
|
|
|
|
|
|
|
|
kvm_x86_ops->get_idt(vcpu, &dt);
|
|
|
|
put_smstate(u32, buf, 0x7e84, dt.size);
|
|
|
|
put_smstate(u64, buf, 0x7e88, dt.address);
|
|
|
|
|
|
|
|
kvm_get_segment(vcpu, &seg, VCPU_SREG_LDTR);
|
|
|
|
put_smstate(u16, buf, 0x7e70, seg.selector);
|
2016-06-02 03:26:01 +07:00
|
|
|
put_smstate(u16, buf, 0x7e72, enter_smm_get_segment_flags(&seg) >> 8);
|
2015-05-05 16:50:23 +07:00
|
|
|
put_smstate(u32, buf, 0x7e74, seg.limit);
|
|
|
|
put_smstate(u64, buf, 0x7e78, seg.base);
|
|
|
|
|
|
|
|
kvm_x86_ops->get_gdt(vcpu, &dt);
|
|
|
|
put_smstate(u32, buf, 0x7e64, dt.size);
|
|
|
|
put_smstate(u64, buf, 0x7e68, dt.address);
|
|
|
|
|
|
|
|
for (i = 0; i < 6; i++)
|
2016-06-02 03:26:01 +07:00
|
|
|
enter_smm_save_seg_64(vcpu, buf, i);
|
2015-05-05 16:50:23 +07:00
|
|
|
#else
|
|
|
|
WARN_ON_ONCE(1);
|
|
|
|
#endif
|
|
|
|
}
|
|
|
|
|
2016-06-02 03:26:01 +07:00
|
|
|
static void enter_smm(struct kvm_vcpu *vcpu)
|
2015-05-07 16:36:11 +07:00
|
|
|
{
|
2015-05-05 16:50:23 +07:00
|
|
|
struct kvm_segment cs, ds;
|
2015-08-07 17:27:54 +07:00
|
|
|
struct desc_ptr dt;
|
2015-05-05 16:50:23 +07:00
|
|
|
char buf[512];
|
|
|
|
u32 cr0;
|
|
|
|
|
|
|
|
trace_kvm_enter_smm(vcpu->vcpu_id, vcpu->arch.smbase, true);
|
|
|
|
vcpu->arch.hflags |= HF_SMM_MASK;
|
|
|
|
memset(buf, 0, 512);
|
|
|
|
if (guest_cpuid_has_longmode(vcpu))
|
2016-06-02 03:26:01 +07:00
|
|
|
enter_smm_save_state_64(vcpu, buf);
|
2015-05-05 16:50:23 +07:00
|
|
|
else
|
2016-06-02 03:26:01 +07:00
|
|
|
enter_smm_save_state_32(vcpu, buf);
|
2015-05-05 16:50:23 +07:00
|
|
|
|
2015-04-08 20:39:23 +07:00
|
|
|
kvm_vcpu_write_guest(vcpu, vcpu->arch.smbase + 0xfe00, buf, sizeof(buf));
|
2015-05-05 16:50:23 +07:00
|
|
|
|
|
|
|
if (kvm_x86_ops->get_nmi_mask(vcpu))
|
|
|
|
vcpu->arch.hflags |= HF_SMM_INSIDE_NMI_MASK;
|
|
|
|
else
|
|
|
|
kvm_x86_ops->set_nmi_mask(vcpu, true);
|
|
|
|
|
|
|
|
kvm_set_rflags(vcpu, X86_EFLAGS_FIXED);
|
|
|
|
kvm_rip_write(vcpu, 0x8000);
|
|
|
|
|
|
|
|
cr0 = vcpu->arch.cr0 & ~(X86_CR0_PE | X86_CR0_EM | X86_CR0_TS | X86_CR0_PG);
|
|
|
|
kvm_x86_ops->set_cr0(vcpu, cr0);
|
|
|
|
vcpu->arch.cr0 = cr0;
|
|
|
|
|
|
|
|
kvm_x86_ops->set_cr4(vcpu, 0);
|
|
|
|
|
2015-08-07 17:27:54 +07:00
|
|
|
/* Undocumented: IDT limit is set to zero on entry to SMM. */
|
|
|
|
dt.address = dt.size = 0;
|
|
|
|
kvm_x86_ops->set_idt(vcpu, &dt);
|
|
|
|
|
2015-05-05 16:50:23 +07:00
|
|
|
__kvm_set_dr(vcpu, 7, DR7_FIXED_1);
|
|
|
|
|
|
|
|
cs.selector = (vcpu->arch.smbase >> 4) & 0xffff;
|
|
|
|
cs.base = vcpu->arch.smbase;
|
|
|
|
|
|
|
|
ds.selector = 0;
|
|
|
|
ds.base = 0;
|
|
|
|
|
|
|
|
cs.limit = ds.limit = 0xffffffff;
|
|
|
|
cs.type = ds.type = 0x3;
|
|
|
|
cs.dpl = ds.dpl = 0;
|
|
|
|
cs.db = ds.db = 0;
|
|
|
|
cs.s = ds.s = 1;
|
|
|
|
cs.l = ds.l = 0;
|
|
|
|
cs.g = ds.g = 1;
|
|
|
|
cs.avl = ds.avl = 0;
|
|
|
|
cs.present = ds.present = 1;
|
|
|
|
cs.unusable = ds.unusable = 0;
|
|
|
|
cs.padding = ds.padding = 0;
|
|
|
|
|
|
|
|
kvm_set_segment(vcpu, &cs, VCPU_SREG_CS);
|
|
|
|
kvm_set_segment(vcpu, &ds, VCPU_SREG_DS);
|
|
|
|
kvm_set_segment(vcpu, &ds, VCPU_SREG_ES);
|
|
|
|
kvm_set_segment(vcpu, &ds, VCPU_SREG_FS);
|
|
|
|
kvm_set_segment(vcpu, &ds, VCPU_SREG_GS);
|
|
|
|
kvm_set_segment(vcpu, &ds, VCPU_SREG_SS);
|
|
|
|
|
|
|
|
if (guest_cpuid_has_longmode(vcpu))
|
|
|
|
kvm_x86_ops->set_efer(vcpu, 0);
|
|
|
|
|
|
|
|
kvm_update_cpuid(vcpu);
|
|
|
|
kvm_mmu_reset_context(vcpu);
|
2015-05-07 16:36:11 +07:00
|
|
|
}
|
|
|
|
|
2016-06-02 03:26:01 +07:00
|
|
|
static void process_smi(struct kvm_vcpu *vcpu)
|
2016-06-02 03:26:00 +07:00
|
|
|
{
|
|
|
|
vcpu->arch.smi_pending = true;
|
|
|
|
kvm_make_request(KVM_REQ_EVENT, vcpu);
|
|
|
|
}
|
|
|
|
|
2016-01-07 21:05:10 +07:00
|
|
|
void kvm_make_scan_ioapic_request(struct kvm *kvm)
|
|
|
|
{
|
|
|
|
kvm_make_all_cpus_request(kvm, KVM_REQ_SCAN_IOAPIC);
|
|
|
|
}
|
|
|
|
|
2013-04-11 18:25:13 +07:00
|
|
|
static void vcpu_scan_ioapic(struct kvm_vcpu *vcpu)
|
2013-01-25 09:18:51 +07:00
|
|
|
{
|
2015-11-10 19:36:34 +07:00
|
|
|
u64 eoi_exit_bitmap[4];
|
|
|
|
|
2013-04-11 18:25:13 +07:00
|
|
|
if (!kvm_apic_hw_enabled(vcpu->arch.apic))
|
|
|
|
return;
|
2013-01-25 09:18:51 +07:00
|
|
|
|
2015-11-10 19:36:32 +07:00
|
|
|
bitmap_zero(vcpu->arch.ioapic_handled_vectors, 256);
|
2013-01-25 09:18:51 +07:00
|
|
|
|
2015-07-30 13:32:35 +07:00
|
|
|
if (irqchip_split(vcpu->kvm))
|
2015-11-10 19:36:32 +07:00
|
|
|
kvm_scan_ioapic_routes(vcpu, vcpu->arch.ioapic_handled_vectors);
|
2015-10-09 01:23:34 +07:00
|
|
|
else {
|
2015-11-10 19:36:33 +07:00
|
|
|
if (vcpu->arch.apicv_active)
|
|
|
|
kvm_x86_ops->sync_pir_to_irr(vcpu);
|
2015-11-10 19:36:32 +07:00
|
|
|
kvm_ioapic_scan_entry(vcpu, vcpu->arch.ioapic_handled_vectors);
|
2015-10-09 01:23:34 +07:00
|
|
|
}
|
2015-11-10 19:36:34 +07:00
|
|
|
bitmap_or((ulong *)eoi_exit_bitmap, vcpu->arch.ioapic_handled_vectors,
|
|
|
|
vcpu_to_synic(vcpu)->vec_bitmap, 256);
|
|
|
|
kvm_x86_ops->load_eoi_exitmap(vcpu, eoi_exit_bitmap);
|
2013-01-25 09:18:51 +07:00
|
|
|
}
|
|
|
|
|
2014-09-18 23:38:36 +07:00
|
|
|
static void kvm_vcpu_flush_tlb(struct kvm_vcpu *vcpu)
|
|
|
|
{
|
|
|
|
++vcpu->stat.tlb_flush;
|
|
|
|
kvm_x86_ops->tlb_flush(vcpu);
|
|
|
|
}
|
|
|
|
|
2014-09-24 14:57:54 +07:00
|
|
|
void kvm_vcpu_reload_apic_access_page(struct kvm_vcpu *vcpu)
|
|
|
|
{
|
2014-09-24 14:57:58 +07:00
|
|
|
struct page *page = NULL;
|
|
|
|
|
2015-07-29 17:05:37 +07:00
|
|
|
if (!lapic_in_kernel(vcpu))
|
2014-10-02 18:53:24 +07:00
|
|
|
return;
|
|
|
|
|
2014-09-24 14:57:54 +07:00
|
|
|
if (!kvm_x86_ops->set_apic_access_page_addr)
|
|
|
|
return;
|
|
|
|
|
2014-09-24 14:57:58 +07:00
|
|
|
page = gfn_to_page(vcpu->kvm, APIC_DEFAULT_PHYS_BASE >> PAGE_SHIFT);
|
2015-05-08 19:32:56 +07:00
|
|
|
if (is_error_page(page))
|
|
|
|
return;
|
2014-09-24 14:57:58 +07:00
|
|
|
kvm_x86_ops->set_apic_access_page_addr(vcpu, page_to_phys(page));
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Do not pin apic access page in memory, the MMU notifier
|
|
|
|
* will call us again if it is migrated or swapped out.
|
|
|
|
*/
|
|
|
|
put_page(page);
|
2014-09-24 14:57:54 +07:00
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(kvm_vcpu_reload_apic_access_page);
|
|
|
|
|
2014-09-24 14:57:57 +07:00
|
|
|
void kvm_arch_mmu_notifier_invalidate_page(struct kvm *kvm,
|
|
|
|
unsigned long address)
|
|
|
|
{
|
2014-09-24 14:57:58 +07:00
|
|
|
/*
|
|
|
|
* The physical address of apic access page is stored in the VMCS.
|
|
|
|
* Update it when it becomes invalid.
|
|
|
|
*/
|
|
|
|
if (address == gfn_to_hva(kvm, APIC_DEFAULT_PHYS_BASE >> PAGE_SHIFT))
|
|
|
|
kvm_make_all_cpus_request(kvm, KVM_REQ_APIC_PAGE_RELOAD);
|
2014-09-24 14:57:57 +07:00
|
|
|
}
|
|
|
|
|
2013-12-13 13:08:38 +07:00
|
|
|
/*
|
2015-02-06 18:48:04 +07:00
|
|
|
* Returns 1 to let vcpu_run() continue the guest execution loop without
|
2013-12-13 13:08:38 +07:00
|
|
|
* exiting to the userspace. Otherwise, the value will be returned to the
|
|
|
|
* userspace.
|
|
|
|
*/
|
2009-08-24 15:10:17 +07:00
|
|
|
static int vcpu_enter_guest(struct kvm_vcpu *vcpu)
|
2007-11-02 02:16:10 +07:00
|
|
|
{
|
|
|
|
int r;
|
2015-11-17 06:26:07 +07:00
|
|
|
bool req_int_win =
|
|
|
|
dm_request_for_irq_injection(vcpu) &&
|
|
|
|
kvm_cpu_accept_dm_intr(vcpu);
|
|
|
|
|
2013-04-28 15:50:52 +07:00
|
|
|
bool req_immediate_exit = false;
|
2007-11-02 02:16:10 +07:00
|
|
|
|
2010-06-23 18:26:18 +07:00
|
|
|
if (vcpu->requests) {
|
2010-05-10 16:34:53 +07:00
|
|
|
if (kvm_check_request(KVM_REQ_MMU_RELOAD, vcpu))
|
2008-02-21 02:47:24 +07:00
|
|
|
kvm_mmu_unload(vcpu);
|
2010-05-10 16:34:53 +07:00
|
|
|
if (kvm_check_request(KVM_REQ_MIGRATE_TIMER, vcpu))
|
2008-05-27 22:10:20 +07:00
|
|
|
__kvm_migrate_timers(vcpu);
|
2012-11-28 08:29:01 +07:00
|
|
|
if (kvm_check_request(KVM_REQ_MASTERCLOCK_UPDATE, vcpu))
|
|
|
|
kvm_gen_update_masterclock(vcpu->kvm);
|
2013-05-10 06:21:41 +07:00
|
|
|
if (kvm_check_request(KVM_REQ_GLOBAL_CLOCK_UPDATE, vcpu))
|
|
|
|
kvm_gen_kvmclock_update(vcpu);
|
2010-09-19 07:38:14 +07:00
|
|
|
if (kvm_check_request(KVM_REQ_CLOCK_UPDATE, vcpu)) {
|
|
|
|
r = kvm_guest_time_update(vcpu);
|
2010-08-20 15:07:21 +07:00
|
|
|
if (unlikely(r))
|
|
|
|
goto out;
|
|
|
|
}
|
2010-05-10 16:34:53 +07:00
|
|
|
if (kvm_check_request(KVM_REQ_MMU_SYNC, vcpu))
|
2008-09-23 23:18:39 +07:00
|
|
|
kvm_mmu_sync_roots(vcpu);
|
2010-05-10 16:34:53 +07:00
|
|
|
if (kvm_check_request(KVM_REQ_TLB_FLUSH, vcpu))
|
2014-09-18 23:38:36 +07:00
|
|
|
kvm_vcpu_flush_tlb(vcpu);
|
2010-05-10 16:34:53 +07:00
|
|
|
if (kvm_check_request(KVM_REQ_REPORT_TPR_ACCESS, vcpu)) {
|
2009-08-24 15:10:17 +07:00
|
|
|
vcpu->run->exit_reason = KVM_EXIT_TPR_ACCESS;
|
2007-10-25 21:52:32 +07:00
|
|
|
r = 0;
|
|
|
|
goto out;
|
|
|
|
}
|
2010-05-10 16:34:53 +07:00
|
|
|
if (kvm_check_request(KVM_REQ_TRIPLE_FAULT, vcpu)) {
|
2009-08-24 15:10:17 +07:00
|
|
|
vcpu->run->exit_reason = KVM_EXIT_SHUTDOWN;
|
2008-02-26 22:49:16 +07:00
|
|
|
r = 0;
|
|
|
|
goto out;
|
|
|
|
}
|
2010-05-10 16:34:53 +07:00
|
|
|
if (kvm_check_request(KVM_REQ_DEACTIVATE_FPU, vcpu)) {
|
2009-12-30 17:40:26 +07:00
|
|
|
vcpu->fpu_active = 0;
|
|
|
|
kvm_x86_ops->fpu_deactivate(vcpu);
|
|
|
|
}
|
2010-10-14 16:22:46 +07:00
|
|
|
if (kvm_check_request(KVM_REQ_APF_HALT, vcpu)) {
|
|
|
|
/* Page is swapped out. Do synthetic halt */
|
|
|
|
vcpu->arch.apf.halted = true;
|
|
|
|
r = 1;
|
|
|
|
goto out;
|
|
|
|
}
|
2011-07-12 02:28:14 +07:00
|
|
|
if (kvm_check_request(KVM_REQ_STEAL_UPDATE, vcpu))
|
|
|
|
record_steal_time(vcpu);
|
2015-05-07 16:36:11 +07:00
|
|
|
if (kvm_check_request(KVM_REQ_SMI, vcpu))
|
|
|
|
process_smi(vcpu);
|
2011-09-20 17:43:14 +07:00
|
|
|
if (kvm_check_request(KVM_REQ_NMI, vcpu))
|
|
|
|
process_nmi(vcpu);
|
2011-11-10 19:57:22 +07:00
|
|
|
if (kvm_check_request(KVM_REQ_PMU, vcpu))
|
2015-06-19 18:44:45 +07:00
|
|
|
kvm_pmu_handle_event(vcpu);
|
2011-11-10 19:57:22 +07:00
|
|
|
if (kvm_check_request(KVM_REQ_PMI, vcpu))
|
2015-06-19 18:44:45 +07:00
|
|
|
kvm_pmu_deliver_pmi(vcpu);
|
2015-07-30 13:21:41 +07:00
|
|
|
if (kvm_check_request(KVM_REQ_IOAPIC_EOI_EXIT, vcpu)) {
|
|
|
|
BUG_ON(vcpu->arch.pending_ioapic_eoi > 255);
|
|
|
|
if (test_bit(vcpu->arch.pending_ioapic_eoi,
|
2015-11-10 19:36:32 +07:00
|
|
|
vcpu->arch.ioapic_handled_vectors)) {
|
2015-07-30 13:21:41 +07:00
|
|
|
vcpu->run->exit_reason = KVM_EXIT_IOAPIC_EOI;
|
|
|
|
vcpu->run->eoi.vector =
|
|
|
|
vcpu->arch.pending_ioapic_eoi;
|
|
|
|
r = 0;
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
}
|
2013-04-11 18:25:13 +07:00
|
|
|
if (kvm_check_request(KVM_REQ_SCAN_IOAPIC, vcpu))
|
|
|
|
vcpu_scan_ioapic(vcpu);
|
2014-09-24 14:57:54 +07:00
|
|
|
if (kvm_check_request(KVM_REQ_APIC_PAGE_RELOAD, vcpu))
|
|
|
|
kvm_vcpu_reload_apic_access_page(vcpu);
|
2015-07-03 19:01:41 +07:00
|
|
|
if (kvm_check_request(KVM_REQ_HV_CRASH, vcpu)) {
|
|
|
|
vcpu->run->exit_reason = KVM_EXIT_SYSTEM_EVENT;
|
|
|
|
vcpu->run->system_event.type = KVM_SYSTEM_EVENT_CRASH;
|
|
|
|
r = 0;
|
|
|
|
goto out;
|
|
|
|
}
|
2015-09-16 16:29:48 +07:00
|
|
|
if (kvm_check_request(KVM_REQ_HV_RESET, vcpu)) {
|
|
|
|
vcpu->run->exit_reason = KVM_EXIT_SYSTEM_EVENT;
|
|
|
|
vcpu->run->system_event.type = KVM_SYSTEM_EVENT_RESET;
|
|
|
|
r = 0;
|
|
|
|
goto out;
|
|
|
|
}
|
2015-11-10 19:36:35 +07:00
|
|
|
if (kvm_check_request(KVM_REQ_HV_EXIT, vcpu)) {
|
|
|
|
vcpu->run->exit_reason = KVM_EXIT_HYPERV;
|
|
|
|
vcpu->run->hyperv = vcpu->arch.hyperv.exit;
|
|
|
|
r = 0;
|
|
|
|
goto out;
|
|
|
|
}
|
2015-12-28 22:27:24 +07:00
|
|
|
|
|
|
|
/*
|
|
|
|
* KVM_REQ_HV_STIMER has to be processed after
|
|
|
|
* KVM_REQ_CLOCK_UPDATE, because Hyper-V SynIC timers
|
|
|
|
* depend on the guest clock being up-to-date
|
|
|
|
*/
|
2015-11-30 23:22:21 +07:00
|
|
|
if (kvm_check_request(KVM_REQ_HV_STIMER, vcpu))
|
|
|
|
kvm_hv_process_stimers(vcpu);
|
2008-01-16 17:49:30 +07:00
|
|
|
}
|
2007-10-25 21:52:32 +07:00
|
|
|
|
2015-09-18 21:29:55 +07:00
|
|
|
/*
|
|
|
|
* KVM_REQ_EVENT is not set when posted interrupts are set by
|
|
|
|
* VT-d hardware, so we have to update RVI unconditionally.
|
|
|
|
*/
|
|
|
|
if (kvm_lapic_enabled(vcpu)) {
|
|
|
|
/*
|
|
|
|
* Update architecture specific hints for APIC
|
|
|
|
* virtual interrupt delivery.
|
|
|
|
*/
|
2015-11-10 19:36:33 +07:00
|
|
|
if (vcpu->arch.apicv_active)
|
2015-09-18 21:29:55 +07:00
|
|
|
kvm_x86_ops->hwapic_irr_update(vcpu,
|
|
|
|
kvm_lapic_find_highest_irr(vcpu));
|
2008-01-16 17:49:30 +07:00
|
|
|
}
|
2007-10-25 21:52:32 +07:00
|
|
|
|
2010-07-20 19:06:17 +07:00
|
|
|
if (kvm_check_request(KVM_REQ_EVENT, vcpu) || req_int_win) {
|
2013-03-13 18:42:34 +07:00
|
|
|
kvm_apic_accept_events(vcpu);
|
|
|
|
if (vcpu->arch.mp_state == KVM_MP_STATE_INIT_RECEIVED) {
|
|
|
|
r = 1;
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
2014-03-08 02:03:12 +07:00
|
|
|
if (inject_pending_event(vcpu, req_int_win) != 0)
|
|
|
|
req_immediate_exit = true;
|
KVM: x86: Inject pending interrupt even if pending nmi exist
Non maskable interrupts (NMI) are preferred to interrupts in current
implementation. If a NMI is pending and NMI is blocked by the result
of nmi_allowed(), pending interrupt is not injected and
enable_irq_window() is not executed, even if interrupts injection is
allowed.
In old kernel (e.g. 2.6.32), schedule() is often called in NMI context.
In this case, interrupts are needed to execute iret that intends end
of NMI. The flag of blocking new NMI is not cleared until the guest
execute the iret, and interrupts are blocked by pending NMI. Due to
this, iret can't be invoked in the guest, and the guest is starved
until block is cleared by some events (e.g. canceling injection).
This patch injects pending interrupts, when it's allowed, even if NMI
is blocked. And, If an interrupts is pending after executing
inject_pending_event(), enable_irq_window() is executed regardless of
NMI pending counter.
Cc: stable@vger.kernel.org
Signed-off-by: Yuki Shibuya <shibuya.yk@ncos.nec.co.jp>
Suggested-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2016-03-24 12:17:03 +07:00
|
|
|
else {
|
2016-06-02 03:26:00 +07:00
|
|
|
/* Enable NMI/IRQ window open exits if needed.
|
|
|
|
*
|
|
|
|
* SMIs have two cases: 1) they can be nested, and
|
|
|
|
* then there is nothing to do here because RSM will
|
|
|
|
* cause a vmexit anyway; 2) or the SMI can be pending
|
|
|
|
* because inject_pending_event has completed the
|
|
|
|
* injection of an IRQ or NMI from the previous vmexit,
|
|
|
|
* and then we request an immediate exit to inject the SMI.
|
|
|
|
*/
|
|
|
|
if (vcpu->arch.smi_pending && !is_smm(vcpu))
|
|
|
|
req_immediate_exit = true;
|
KVM: x86: Inject pending interrupt even if pending nmi exist
Non maskable interrupts (NMI) are preferred to interrupts in current
implementation. If a NMI is pending and NMI is blocked by the result
of nmi_allowed(), pending interrupt is not injected and
enable_irq_window() is not executed, even if interrupts injection is
allowed.
In old kernel (e.g. 2.6.32), schedule() is often called in NMI context.
In this case, interrupts are needed to execute iret that intends end
of NMI. The flag of blocking new NMI is not cleared until the guest
execute the iret, and interrupts are blocked by pending NMI. Due to
this, iret can't be invoked in the guest, and the guest is starved
until block is cleared by some events (e.g. canceling injection).
This patch injects pending interrupts, when it's allowed, even if NMI
is blocked. And, If an interrupts is pending after executing
inject_pending_event(), enable_irq_window() is executed regardless of
NMI pending counter.
Cc: stable@vger.kernel.org
Signed-off-by: Yuki Shibuya <shibuya.yk@ncos.nec.co.jp>
Suggested-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2016-03-24 12:17:03 +07:00
|
|
|
if (vcpu->arch.nmi_pending)
|
|
|
|
kvm_x86_ops->enable_nmi_window(vcpu);
|
|
|
|
if (kvm_cpu_has_injectable_intr(vcpu) || req_int_win)
|
|
|
|
kvm_x86_ops->enable_irq_window(vcpu);
|
|
|
|
}
|
2010-07-20 19:06:17 +07:00
|
|
|
|
|
|
|
if (kvm_lapic_enabled(vcpu)) {
|
|
|
|
update_cr8_intercept(vcpu);
|
|
|
|
kvm_lapic_sync_to_vapic(vcpu);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2012-05-14 22:07:56 +07:00
|
|
|
r = kvm_mmu_reload(vcpu);
|
|
|
|
if (unlikely(r)) {
|
2012-06-24 23:25:00 +07:00
|
|
|
goto cancel_injection;
|
2012-05-14 22:07:56 +07:00
|
|
|
}
|
|
|
|
|
2007-11-02 02:16:10 +07:00
|
|
|
preempt_disable();
|
|
|
|
|
|
|
|
kvm_x86_ops->prepare_guest_switch(vcpu);
|
2010-01-21 20:31:45 +07:00
|
|
|
if (vcpu->fpu_active)
|
|
|
|
kvm_load_guest_fpu(vcpu);
|
2011-01-12 14:40:31 +07:00
|
|
|
vcpu->mode = IN_GUEST_MODE;
|
|
|
|
|
2013-11-05 03:36:25 +07:00
|
|
|
srcu_read_unlock(&vcpu->kvm->srcu, vcpu->srcu_idx);
|
|
|
|
|
2016-03-13 10:10:29 +07:00
|
|
|
/*
|
|
|
|
* We should set ->mode before check ->requests,
|
|
|
|
* Please see the comment in kvm_make_all_cpus_request.
|
|
|
|
* This also orders the write to mode from any reads
|
|
|
|
* to the page tables done while the VCPU is running.
|
|
|
|
* Please see the comment in kvm_flush_remote_tlbs.
|
2011-01-12 14:40:31 +07:00
|
|
|
*/
|
2013-11-05 03:36:25 +07:00
|
|
|
smp_mb__after_srcu_read_unlock();
|
2007-11-02 02:16:10 +07:00
|
|
|
|
2010-05-03 20:54:48 +07:00
|
|
|
local_irq_disable();
|
2009-05-08 03:55:12 +07:00
|
|
|
|
2011-01-12 14:40:31 +07:00
|
|
|
if (vcpu->mode == EXITING_GUEST_MODE || vcpu->requests
|
2010-05-03 20:54:48 +07:00
|
|
|
|| need_resched() || signal_pending(current)) {
|
2011-01-12 14:40:31 +07:00
|
|
|
vcpu->mode = OUTSIDE_GUEST_MODE;
|
2010-05-03 20:54:48 +07:00
|
|
|
smp_wmb();
|
2008-01-15 23:27:32 +07:00
|
|
|
local_irq_enable();
|
|
|
|
preempt_enable();
|
2013-11-05 03:36:25 +07:00
|
|
|
vcpu->srcu_idx = srcu_read_lock(&vcpu->kvm->srcu);
|
2008-01-15 23:27:32 +07:00
|
|
|
r = 1;
|
2012-06-24 23:25:00 +07:00
|
|
|
goto cancel_injection;
|
2008-01-15 23:27:32 +07:00
|
|
|
}
|
|
|
|
|
2016-03-31 02:24:47 +07:00
|
|
|
kvm_load_guest_xcr0(vcpu);
|
|
|
|
|
2016-06-02 03:26:00 +07:00
|
|
|
if (req_immediate_exit) {
|
|
|
|
kvm_make_request(KVM_REQ_EVENT, vcpu);
|
KVM: nVMX: Add KVM_REQ_IMMEDIATE_EXIT
This patch adds a new vcpu->requests bit, KVM_REQ_IMMEDIATE_EXIT.
This bit requests that when next entering the guest, we should run it only
for as little as possible, and exit again.
We use this new option in nested VMX: When L1 launches L2, but L0 wishes L1
to continue running so it can inject an event to it, we unfortunately cannot
just pretend to have run L2 for a little while - We must really launch L2,
otherwise certain one-off vmcs12 parameters (namely, L1 injection into L2)
will be lost. So the existing code runs L2 in this case.
But L2 could potentially run for a long time until it exits, and the
injection into L1 will be delayed. The new KVM_REQ_IMMEDIATE_EXIT allows us
to request that L2 will be entered, as necessary, but will exit as soon as
possible after entry.
Our implementation of this request uses smp_send_reschedule() to send a
self-IPI, with interrupts disabled. The interrupts remain disabled until the
guest is entered, and then, after the entry is complete (often including
processing an injection and jumping to the relevant handler), the physical
interrupt is noticed and causes an exit.
On recent Intel processors, we could have achieved the same goal by using
MTF instead of a self-IPI. Another technique worth considering in the future
is to use VM_EXIT_ACK_INTR_ON_EXIT and a highest-priority vector IPI - to
slightly improve performance by avoiding the useless interrupt handler
which ends up being called when smp_send_reschedule() is used.
Signed-off-by: Nadav Har'El <nyh@il.ibm.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
2011-09-22 17:52:56 +07:00
|
|
|
smp_send_reschedule(vcpu->cpu);
|
2016-06-02 03:26:00 +07:00
|
|
|
}
|
KVM: nVMX: Add KVM_REQ_IMMEDIATE_EXIT
This patch adds a new vcpu->requests bit, KVM_REQ_IMMEDIATE_EXIT.
This bit requests that when next entering the guest, we should run it only
for as little as possible, and exit again.
We use this new option in nested VMX: When L1 launches L2, but L0 wishes L1
to continue running so it can inject an event to it, we unfortunately cannot
just pretend to have run L2 for a little while - We must really launch L2,
otherwise certain one-off vmcs12 parameters (namely, L1 injection into L2)
will be lost. So the existing code runs L2 in this case.
But L2 could potentially run for a long time until it exits, and the
injection into L1 will be delayed. The new KVM_REQ_IMMEDIATE_EXIT allows us
to request that L2 will be entered, as necessary, but will exit as soon as
possible after entry.
Our implementation of this request uses smp_send_reschedule() to send a
self-IPI, with interrupts disabled. The interrupts remain disabled until the
guest is entered, and then, after the entry is complete (often including
processing an injection and jumping to the relevant handler), the physical
interrupt is noticed and causes an exit.
On recent Intel processors, we could have achieved the same goal by using
MTF instead of a self-IPI. Another technique worth considering in the future
is to use VM_EXIT_ACK_INTR_ON_EXIT and a highest-priority vector IPI - to
slightly improve performance by avoiding the useless interrupt handler
which ends up being called when smp_send_reschedule() is used.
Signed-off-by: Nadav Har'El <nyh@il.ibm.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
2011-09-22 17:52:56 +07:00
|
|
|
|
2015-12-11 00:37:32 +07:00
|
|
|
trace_kvm_entry(vcpu->vcpu_id);
|
|
|
|
wait_lapic_expire(vcpu);
|
2016-06-15 20:18:26 +07:00
|
|
|
guest_enter_irqoff();
|
2007-11-02 02:16:10 +07:00
|
|
|
|
2008-12-15 19:52:10 +07:00
|
|
|
if (unlikely(vcpu->arch.switch_db_regs)) {
|
|
|
|
set_debugreg(0, 7);
|
|
|
|
set_debugreg(vcpu->arch.eff_db[0], 0);
|
|
|
|
set_debugreg(vcpu->arch.eff_db[1], 1);
|
|
|
|
set_debugreg(vcpu->arch.eff_db[2], 2);
|
|
|
|
set_debugreg(vcpu->arch.eff_db[3], 3);
|
2014-02-21 16:17:24 +07:00
|
|
|
set_debugreg(vcpu->arch.dr6, 6);
|
2015-04-02 07:10:37 +07:00
|
|
|
vcpu->arch.switch_db_regs &= ~KVM_DEBUGREG_RELOAD;
|
2008-12-15 19:52:10 +07:00
|
|
|
}
|
2007-11-02 02:16:10 +07:00
|
|
|
|
2009-08-24 15:10:17 +07:00
|
|
|
kvm_x86_ops->run(vcpu);
|
2007-11-02 02:16:10 +07:00
|
|
|
|
2014-02-21 16:17:24 +07:00
|
|
|
/*
|
|
|
|
* Do this here before restoring debug registers on the host. And
|
|
|
|
* since we do this before handling the vmexit, a DR access vmexit
|
|
|
|
* can (a) read the correct value of the debug registers, (b) set
|
|
|
|
* KVM_DEBUGREG_WONT_EXIT again.
|
|
|
|
*/
|
|
|
|
if (unlikely(vcpu->arch.switch_db_regs & KVM_DEBUGREG_WONT_EXIT)) {
|
|
|
|
WARN_ON(vcpu->guest_debug & KVM_GUESTDBG_USE_HW_BP);
|
|
|
|
kvm_x86_ops->sync_dirty_debug_regs(vcpu);
|
2016-02-26 18:28:40 +07:00
|
|
|
kvm_update_dr0123(vcpu);
|
|
|
|
kvm_update_dr6(vcpu);
|
|
|
|
kvm_update_dr7(vcpu);
|
|
|
|
vcpu->arch.switch_db_regs &= ~KVM_DEBUGREG_RELOAD;
|
2014-02-21 16:17:24 +07:00
|
|
|
}
|
|
|
|
|
2009-09-10 00:22:48 +07:00
|
|
|
/*
|
|
|
|
* If the guest has used debug registers, at least dr7
|
|
|
|
* will be disabled while returning to the host.
|
|
|
|
* If we don't have active breakpoints in the host, we don't
|
|
|
|
* care about the messed up debug address registers. But if
|
|
|
|
* we have some of them active, restore the old state.
|
|
|
|
*/
|
2009-11-10 17:03:12 +07:00
|
|
|
if (hw_breakpoint_active())
|
2009-09-10 00:22:48 +07:00
|
|
|
hw_breakpoint_restore();
|
2008-12-15 19:52:10 +07:00
|
|
|
|
2015-10-20 14:39:07 +07:00
|
|
|
vcpu->arch.last_guest_tsc = kvm_read_l1_tsc(vcpu, rdtsc());
|
2010-08-20 15:07:30 +07:00
|
|
|
|
2011-01-12 14:40:31 +07:00
|
|
|
vcpu->mode = OUTSIDE_GUEST_MODE;
|
2010-05-03 20:54:48 +07:00
|
|
|
smp_wmb();
|
2013-04-11 18:25:10 +07:00
|
|
|
|
2016-03-31 02:24:47 +07:00
|
|
|
kvm_put_guest_xcr0(vcpu);
|
|
|
|
|
2013-04-11 18:25:10 +07:00
|
|
|
/* Interrupt is enabled by handle_external_intr() */
|
|
|
|
kvm_x86_ops->handle_external_intr(vcpu);
|
2007-11-02 02:16:10 +07:00
|
|
|
|
|
|
|
++vcpu->stat.exits;
|
|
|
|
|
2016-06-15 20:23:11 +07:00
|
|
|
guest_exit_irqoff();
|
2007-11-02 02:16:10 +07:00
|
|
|
|
2016-06-15 20:23:11 +07:00
|
|
|
local_irq_enable();
|
2007-11-02 02:16:10 +07:00
|
|
|
preempt_enable();
|
|
|
|
|
2009-12-23 23:35:25 +07:00
|
|
|
vcpu->srcu_idx = srcu_read_lock(&vcpu->kvm->srcu);
|
2008-03-30 06:17:59 +07:00
|
|
|
|
2007-11-02 02:16:10 +07:00
|
|
|
/*
|
|
|
|
* Profile KVM exit RIPs:
|
|
|
|
*/
|
|
|
|
if (unlikely(prof_on == KVM_PROFILING)) {
|
2008-06-28 00:58:02 +07:00
|
|
|
unsigned long rip = kvm_rip_read(vcpu);
|
|
|
|
profile_hit(KVM_PROFILING, (void *)rip);
|
2007-11-02 02:16:10 +07:00
|
|
|
}
|
|
|
|
|
KVM: Infrastructure for software and hardware based TSC rate scaling
This requires some restructuring; rather than use 'virtual_tsc_khz'
to indicate whether hardware rate scaling is in effect, we consider
each VCPU to always have a virtual TSC rate. Instead, there is new
logic above the vendor-specific hardware scaling that decides whether
it is even necessary to use and updates all rate variables used by
common code. This means we can simply query the virtual rate at
any point, which is needed for software rate scaling.
There is also now a threshold added to the TSC rate scaling; minor
differences and variations of measured TSC rate can accidentally
provoke rate scaling to be used when it is not needed. Instead,
we have a tolerance variable called tsc_tolerance_ppm, which is
the maximum variation from user requested rate at which scaling
will be used. The default is 250ppm, which is the half the
threshold for NTP adjustment, allowing for some hardware variation.
In the event that hardware rate scaling is not available, we can
kludge a bit by forcing TSC catchup to turn on when a faster than
hardware speed has been requested, but there is nothing available
yet for the reverse case; this requires a trap and emulate software
implementation for RDTSC, which is still forthcoming.
[avi: fix 64-bit division on i386]
Signed-off-by: Zachary Amsden <zamsden@gmail.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
2012-02-04 00:43:50 +07:00
|
|
|
if (unlikely(vcpu->arch.tsc_always_catchup))
|
|
|
|
kvm_make_request(KVM_REQ_CLOCK_UPDATE, vcpu);
|
2007-11-25 18:41:11 +07:00
|
|
|
|
2012-06-24 23:24:54 +07:00
|
|
|
if (vcpu->arch.apic_attention)
|
|
|
|
kvm_lapic_sync_from_vapic(vcpu);
|
2007-10-25 21:52:32 +07:00
|
|
|
|
2009-08-24 15:10:17 +07:00
|
|
|
r = kvm_x86_ops->handle_exit(vcpu);
|
2012-06-24 23:25:00 +07:00
|
|
|
return r;
|
|
|
|
|
|
|
|
cancel_injection:
|
|
|
|
kvm_x86_ops->cancel_injection(vcpu);
|
2012-06-24 23:25:07 +07:00
|
|
|
if (unlikely(vcpu->arch.apic_attention))
|
|
|
|
kvm_lapic_sync_from_vapic(vcpu);
|
2008-09-09 01:23:48 +07:00
|
|
|
out:
|
|
|
|
return r;
|
|
|
|
}
|
2007-11-02 02:16:10 +07:00
|
|
|
|
2015-02-06 18:48:04 +07:00
|
|
|
static inline int vcpu_block(struct kvm *kvm, struct kvm_vcpu *vcpu)
|
|
|
|
{
|
2015-09-18 21:29:55 +07:00
|
|
|
if (!kvm_arch_vcpu_runnable(vcpu) &&
|
|
|
|
(!kvm_x86_ops->pre_block || kvm_x86_ops->pre_block(vcpu) == 0)) {
|
2015-02-06 18:58:42 +07:00
|
|
|
srcu_read_unlock(&kvm->srcu, vcpu->srcu_idx);
|
|
|
|
kvm_vcpu_block(vcpu);
|
|
|
|
vcpu->srcu_idx = srcu_read_lock(&kvm->srcu);
|
2015-09-18 21:29:55 +07:00
|
|
|
|
|
|
|
if (kvm_x86_ops->post_block)
|
|
|
|
kvm_x86_ops->post_block(vcpu);
|
|
|
|
|
2015-02-06 18:58:42 +07:00
|
|
|
if (!kvm_check_request(KVM_REQ_UNHALT, vcpu))
|
|
|
|
return 1;
|
|
|
|
}
|
2015-02-06 18:48:04 +07:00
|
|
|
|
|
|
|
kvm_apic_accept_events(vcpu);
|
|
|
|
switch(vcpu->arch.mp_state) {
|
|
|
|
case KVM_MP_STATE_HALTED:
|
|
|
|
vcpu->arch.pv.pv_unhalted = false;
|
|
|
|
vcpu->arch.mp_state =
|
|
|
|
KVM_MP_STATE_RUNNABLE;
|
|
|
|
case KVM_MP_STATE_RUNNABLE:
|
|
|
|
vcpu->arch.apf.halted = false;
|
|
|
|
break;
|
|
|
|
case KVM_MP_STATE_INIT_RECEIVED:
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
return -EINTR;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
return 1;
|
|
|
|
}
|
2009-03-23 20:11:44 +07:00
|
|
|
|
2015-10-13 15:18:53 +07:00
|
|
|
static inline bool kvm_vcpu_running(struct kvm_vcpu *vcpu)
|
|
|
|
{
|
|
|
|
return (vcpu->arch.mp_state == KVM_MP_STATE_RUNNABLE &&
|
|
|
|
!vcpu->arch.apf.halted);
|
|
|
|
}
|
|
|
|
|
2015-02-06 18:48:04 +07:00
|
|
|
static int vcpu_run(struct kvm_vcpu *vcpu)
|
2008-09-09 01:23:48 +07:00
|
|
|
{
|
|
|
|
int r;
|
2009-12-23 23:35:25 +07:00
|
|
|
struct kvm *kvm = vcpu->kvm;
|
2008-09-09 01:23:48 +07:00
|
|
|
|
2009-12-23 23:35:25 +07:00
|
|
|
vcpu->srcu_idx = srcu_read_lock(&kvm->srcu);
|
2008-09-09 01:23:48 +07:00
|
|
|
|
2015-02-06 18:48:04 +07:00
|
|
|
for (;;) {
|
2015-10-14 02:32:50 +07:00
|
|
|
if (kvm_vcpu_running(vcpu)) {
|
2009-08-24 15:10:17 +07:00
|
|
|
r = vcpu_enter_guest(vcpu);
|
2015-09-18 21:29:55 +07:00
|
|
|
} else {
|
2015-02-06 18:48:04 +07:00
|
|
|
r = vcpu_block(kvm, vcpu);
|
2015-09-18 21:29:55 +07:00
|
|
|
}
|
|
|
|
|
2009-03-23 20:11:44 +07:00
|
|
|
if (r <= 0)
|
|
|
|
break;
|
|
|
|
|
|
|
|
clear_bit(KVM_REQ_PENDING_TIMER, &vcpu->requests);
|
|
|
|
if (kvm_cpu_has_pending_timer(vcpu))
|
|
|
|
kvm_inject_pending_timer_irqs(vcpu);
|
|
|
|
|
2015-11-17 06:26:00 +07:00
|
|
|
if (dm_request_for_irq_injection(vcpu) &&
|
|
|
|
kvm_vcpu_ready_for_interrupt_injection(vcpu)) {
|
2015-07-30 15:32:16 +07:00
|
|
|
r = 0;
|
|
|
|
vcpu->run->exit_reason = KVM_EXIT_IRQ_WINDOW_OPEN;
|
2009-03-23 20:11:44 +07:00
|
|
|
++vcpu->stat.request_irq_exits;
|
2015-02-06 18:48:04 +07:00
|
|
|
break;
|
2009-03-23 20:11:44 +07:00
|
|
|
}
|
2010-10-14 16:22:46 +07:00
|
|
|
|
|
|
|
kvm_check_async_pf_completion(vcpu);
|
|
|
|
|
2009-03-23 20:11:44 +07:00
|
|
|
if (signal_pending(current)) {
|
|
|
|
r = -EINTR;
|
2009-08-24 15:10:17 +07:00
|
|
|
vcpu->run->exit_reason = KVM_EXIT_INTR;
|
2009-03-23 20:11:44 +07:00
|
|
|
++vcpu->stat.signal_exits;
|
2015-02-06 18:48:04 +07:00
|
|
|
break;
|
2009-03-23 20:11:44 +07:00
|
|
|
}
|
|
|
|
if (need_resched()) {
|
2009-12-23 23:35:25 +07:00
|
|
|
srcu_read_unlock(&kvm->srcu, vcpu->srcu_idx);
|
2013-12-13 13:07:21 +07:00
|
|
|
cond_resched();
|
2009-12-23 23:35:25 +07:00
|
|
|
vcpu->srcu_idx = srcu_read_lock(&kvm->srcu);
|
2008-09-09 01:23:48 +07:00
|
|
|
}
|
2007-11-02 02:16:10 +07:00
|
|
|
}
|
|
|
|
|
2009-12-23 23:35:25 +07:00
|
|
|
srcu_read_unlock(&kvm->srcu, vcpu->srcu_idx);
|
2007-11-02 02:16:10 +07:00
|
|
|
|
|
|
|
return r;
|
|
|
|
}
|
|
|
|
|
2012-09-03 19:24:26 +07:00
|
|
|
static inline int complete_emulated_io(struct kvm_vcpu *vcpu)
|
|
|
|
{
|
|
|
|
int r;
|
|
|
|
vcpu->srcu_idx = srcu_read_lock(&vcpu->kvm->srcu);
|
|
|
|
r = emulate_instruction(vcpu, EMULTYPE_NO_DECODE);
|
|
|
|
srcu_read_unlock(&vcpu->kvm->srcu, vcpu->srcu_idx);
|
|
|
|
if (r != EMULATE_DONE)
|
|
|
|
return 0;
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int complete_emulated_pio(struct kvm_vcpu *vcpu)
|
|
|
|
{
|
|
|
|
BUG_ON(!vcpu->arch.pio.count);
|
|
|
|
|
|
|
|
return complete_emulated_io(vcpu);
|
|
|
|
}
|
|
|
|
|
2012-04-18 23:22:47 +07:00
|
|
|
/*
|
|
|
|
* Implements the following, as a state machine:
|
|
|
|
*
|
|
|
|
* read:
|
|
|
|
* for each fragment
|
2012-10-24 13:07:59 +07:00
|
|
|
* for each mmio piece in the fragment
|
|
|
|
* write gpa, len
|
|
|
|
* exit
|
|
|
|
* copy data
|
2012-04-18 23:22:47 +07:00
|
|
|
* execute insn
|
|
|
|
*
|
|
|
|
* write:
|
|
|
|
* for each fragment
|
2012-10-24 13:07:59 +07:00
|
|
|
* for each mmio piece in the fragment
|
|
|
|
* write gpa, len
|
|
|
|
* copy data
|
|
|
|
* exit
|
2012-04-18 23:22:47 +07:00
|
|
|
*/
|
2012-09-03 19:24:26 +07:00
|
|
|
static int complete_emulated_mmio(struct kvm_vcpu *vcpu)
|
2010-01-19 19:20:10 +07:00
|
|
|
{
|
|
|
|
struct kvm_run *run = vcpu->run;
|
2012-04-18 23:22:47 +07:00
|
|
|
struct kvm_mmio_fragment *frag;
|
2012-10-24 13:07:59 +07:00
|
|
|
unsigned len;
|
2010-01-19 19:20:10 +07:00
|
|
|
|
2012-09-03 19:24:26 +07:00
|
|
|
BUG_ON(!vcpu->mmio_needed);
|
2010-01-19 19:20:10 +07:00
|
|
|
|
2012-09-03 19:24:26 +07:00
|
|
|
/* Complete previous fragment */
|
2012-10-24 13:07:59 +07:00
|
|
|
frag = &vcpu->mmio_fragments[vcpu->mmio_cur_fragment];
|
|
|
|
len = min(8u, frag->len);
|
2012-09-03 19:24:26 +07:00
|
|
|
if (!vcpu->mmio_is_write)
|
2012-10-24 13:07:59 +07:00
|
|
|
memcpy(frag->data, run->mmio.data, len);
|
|
|
|
|
|
|
|
if (frag->len <= 8) {
|
|
|
|
/* Switch to the next fragment. */
|
|
|
|
frag++;
|
|
|
|
vcpu->mmio_cur_fragment++;
|
|
|
|
} else {
|
|
|
|
/* Go forward to the next mmio piece. */
|
|
|
|
frag->data += len;
|
|
|
|
frag->gpa += len;
|
|
|
|
frag->len -= len;
|
|
|
|
}
|
|
|
|
|
2014-02-28 01:35:14 +07:00
|
|
|
if (vcpu->mmio_cur_fragment >= vcpu->mmio_nr_fragments) {
|
2012-09-03 19:24:26 +07:00
|
|
|
vcpu->mmio_needed = 0;
|
2013-08-27 20:41:43 +07:00
|
|
|
|
|
|
|
/* FIXME: return into emulator if single-stepping. */
|
2010-01-20 17:01:20 +07:00
|
|
|
if (vcpu->mmio_is_write)
|
2012-09-03 19:24:26 +07:00
|
|
|
return 1;
|
|
|
|
vcpu->mmio_read_completed = 1;
|
|
|
|
return complete_emulated_io(vcpu);
|
|
|
|
}
|
2012-10-24 13:07:59 +07:00
|
|
|
|
2012-09-03 19:24:26 +07:00
|
|
|
run->exit_reason = KVM_EXIT_MMIO;
|
|
|
|
run->mmio.phys_addr = frag->gpa;
|
|
|
|
if (vcpu->mmio_is_write)
|
2012-10-24 13:07:59 +07:00
|
|
|
memcpy(run->mmio.data, frag->data, min(8u, frag->len));
|
|
|
|
run->mmio.len = min(8u, frag->len);
|
2012-09-03 19:24:26 +07:00
|
|
|
run->mmio.is_write = vcpu->mmio_is_write;
|
|
|
|
vcpu->arch.complete_userspace_io = complete_emulated_mmio;
|
|
|
|
return 0;
|
2010-01-19 19:20:10 +07:00
|
|
|
}
|
|
|
|
|
2012-09-03 19:24:26 +07:00
|
|
|
|
2007-11-02 02:16:10 +07:00
|
|
|
int kvm_arch_vcpu_ioctl_run(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run)
|
|
|
|
{
|
2015-04-23 17:49:20 +07:00
|
|
|
struct fpu *fpu = ¤t->thread.fpu;
|
2007-11-02 02:16:10 +07:00
|
|
|
int r;
|
|
|
|
sigset_t sigsaved;
|
|
|
|
|
2015-04-27 12:18:17 +07:00
|
|
|
fpu__activate_curr(fpu);
|
2011-01-11 17:15:54 +07:00
|
|
|
|
2008-07-06 19:48:31 +07:00
|
|
|
if (vcpu->sigset_active)
|
|
|
|
sigprocmask(SIG_SETMASK, &vcpu->sigset, &sigsaved);
|
|
|
|
|
2008-04-13 21:54:35 +07:00
|
|
|
if (unlikely(vcpu->arch.mp_state == KVM_MP_STATE_UNINITIALIZED)) {
|
2007-11-02 02:16:10 +07:00
|
|
|
kvm_vcpu_block(vcpu);
|
2013-03-13 18:42:34 +07:00
|
|
|
kvm_apic_accept_events(vcpu);
|
2008-09-09 01:23:48 +07:00
|
|
|
clear_bit(KVM_REQ_UNHALT, &vcpu->requests);
|
2008-07-06 19:48:31 +07:00
|
|
|
r = -EAGAIN;
|
|
|
|
goto out;
|
2007-11-02 02:16:10 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
/* re-sync apic's tpr */
|
2015-07-29 17:05:37 +07:00
|
|
|
if (!lapic_in_kernel(vcpu)) {
|
2010-12-21 17:12:00 +07:00
|
|
|
if (kvm_set_cr8(vcpu, kvm_run->cr8) != 0) {
|
|
|
|
r = -EINVAL;
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
}
|
2007-11-02 02:16:10 +07:00
|
|
|
|
2012-09-03 19:24:26 +07:00
|
|
|
if (unlikely(vcpu->arch.complete_userspace_io)) {
|
|
|
|
int (*cui)(struct kvm_vcpu *) = vcpu->arch.complete_userspace_io;
|
|
|
|
vcpu->arch.complete_userspace_io = NULL;
|
|
|
|
r = cui(vcpu);
|
|
|
|
if (r <= 0)
|
|
|
|
goto out;
|
|
|
|
} else
|
|
|
|
WARN_ON(vcpu->arch.pio.count || vcpu->mmio_needed);
|
2010-01-19 19:20:10 +07:00
|
|
|
|
2015-02-06 18:48:04 +07:00
|
|
|
r = vcpu_run(vcpu);
|
2007-11-02 02:16:10 +07:00
|
|
|
|
|
|
|
out:
|
2010-05-04 09:04:27 +07:00
|
|
|
post_kvm_run_save(vcpu);
|
2007-11-02 02:16:10 +07:00
|
|
|
if (vcpu->sigset_active)
|
|
|
|
sigprocmask(SIG_SETMASK, &sigsaved, NULL);
|
|
|
|
|
|
|
|
return r;
|
|
|
|
}
|
|
|
|
|
|
|
|
int kvm_arch_vcpu_ioctl_get_regs(struct kvm_vcpu *vcpu, struct kvm_regs *regs)
|
|
|
|
{
|
2011-03-31 17:06:41 +07:00
|
|
|
if (vcpu->arch.emulate_regs_need_sync_to_vcpu) {
|
|
|
|
/*
|
|
|
|
* We are here if userspace calls get_regs() in the middle of
|
|
|
|
* instruction emulation. Registers state needs to be copied
|
2012-06-28 14:17:27 +07:00
|
|
|
* back from emulation context to vcpu. Userspace shouldn't do
|
2011-03-31 17:06:41 +07:00
|
|
|
* that usually, but some bad designed PV devices (vmware
|
|
|
|
* backdoor interface) need this to work
|
|
|
|
*/
|
2012-08-28 03:46:17 +07:00
|
|
|
emulator_writeback_register_cache(&vcpu->arch.emulate_ctxt);
|
2011-03-31 17:06:41 +07:00
|
|
|
vcpu->arch.emulate_regs_need_sync_to_vcpu = false;
|
|
|
|
}
|
2008-06-28 00:58:02 +07:00
|
|
|
regs->rax = kvm_register_read(vcpu, VCPU_REGS_RAX);
|
|
|
|
regs->rbx = kvm_register_read(vcpu, VCPU_REGS_RBX);
|
|
|
|
regs->rcx = kvm_register_read(vcpu, VCPU_REGS_RCX);
|
|
|
|
regs->rdx = kvm_register_read(vcpu, VCPU_REGS_RDX);
|
|
|
|
regs->rsi = kvm_register_read(vcpu, VCPU_REGS_RSI);
|
|
|
|
regs->rdi = kvm_register_read(vcpu, VCPU_REGS_RDI);
|
|
|
|
regs->rsp = kvm_register_read(vcpu, VCPU_REGS_RSP);
|
|
|
|
regs->rbp = kvm_register_read(vcpu, VCPU_REGS_RBP);
|
2007-11-02 02:16:10 +07:00
|
|
|
#ifdef CONFIG_X86_64
|
2008-06-28 00:58:02 +07:00
|
|
|
regs->r8 = kvm_register_read(vcpu, VCPU_REGS_R8);
|
|
|
|
regs->r9 = kvm_register_read(vcpu, VCPU_REGS_R9);
|
|
|
|
regs->r10 = kvm_register_read(vcpu, VCPU_REGS_R10);
|
|
|
|
regs->r11 = kvm_register_read(vcpu, VCPU_REGS_R11);
|
|
|
|
regs->r12 = kvm_register_read(vcpu, VCPU_REGS_R12);
|
|
|
|
regs->r13 = kvm_register_read(vcpu, VCPU_REGS_R13);
|
|
|
|
regs->r14 = kvm_register_read(vcpu, VCPU_REGS_R14);
|
|
|
|
regs->r15 = kvm_register_read(vcpu, VCPU_REGS_R15);
|
2007-11-02 02:16:10 +07:00
|
|
|
#endif
|
|
|
|
|
2008-06-28 00:58:02 +07:00
|
|
|
regs->rip = kvm_rip_read(vcpu);
|
2009-10-05 18:07:21 +07:00
|
|
|
regs->rflags = kvm_get_rflags(vcpu);
|
2007-11-02 02:16:10 +07:00
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
int kvm_arch_vcpu_ioctl_set_regs(struct kvm_vcpu *vcpu, struct kvm_regs *regs)
|
|
|
|
{
|
2011-03-31 17:06:41 +07:00
|
|
|
vcpu->arch.emulate_regs_need_sync_from_vcpu = true;
|
|
|
|
vcpu->arch.emulate_regs_need_sync_to_vcpu = false;
|
|
|
|
|
2008-06-28 00:58:02 +07:00
|
|
|
kvm_register_write(vcpu, VCPU_REGS_RAX, regs->rax);
|
|
|
|
kvm_register_write(vcpu, VCPU_REGS_RBX, regs->rbx);
|
|
|
|
kvm_register_write(vcpu, VCPU_REGS_RCX, regs->rcx);
|
|
|
|
kvm_register_write(vcpu, VCPU_REGS_RDX, regs->rdx);
|
|
|
|
kvm_register_write(vcpu, VCPU_REGS_RSI, regs->rsi);
|
|
|
|
kvm_register_write(vcpu, VCPU_REGS_RDI, regs->rdi);
|
|
|
|
kvm_register_write(vcpu, VCPU_REGS_RSP, regs->rsp);
|
|
|
|
kvm_register_write(vcpu, VCPU_REGS_RBP, regs->rbp);
|
2007-11-02 02:16:10 +07:00
|
|
|
#ifdef CONFIG_X86_64
|
2008-06-28 00:58:02 +07:00
|
|
|
kvm_register_write(vcpu, VCPU_REGS_R8, regs->r8);
|
|
|
|
kvm_register_write(vcpu, VCPU_REGS_R9, regs->r9);
|
|
|
|
kvm_register_write(vcpu, VCPU_REGS_R10, regs->r10);
|
|
|
|
kvm_register_write(vcpu, VCPU_REGS_R11, regs->r11);
|
|
|
|
kvm_register_write(vcpu, VCPU_REGS_R12, regs->r12);
|
|
|
|
kvm_register_write(vcpu, VCPU_REGS_R13, regs->r13);
|
|
|
|
kvm_register_write(vcpu, VCPU_REGS_R14, regs->r14);
|
|
|
|
kvm_register_write(vcpu, VCPU_REGS_R15, regs->r15);
|
2007-11-02 02:16:10 +07:00
|
|
|
#endif
|
|
|
|
|
2008-06-28 00:58:02 +07:00
|
|
|
kvm_rip_write(vcpu, regs->rip);
|
2009-10-05 18:07:21 +07:00
|
|
|
kvm_set_rflags(vcpu, regs->rflags);
|
2007-11-02 02:16:10 +07:00
|
|
|
|
2008-04-30 22:59:04 +07:00
|
|
|
vcpu->arch.exception.pending = false;
|
|
|
|
|
2010-07-27 16:30:24 +07:00
|
|
|
kvm_make_request(KVM_REQ_EVENT, vcpu);
|
|
|
|
|
2007-11-02 02:16:10 +07:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
void kvm_get_cs_db_l_bits(struct kvm_vcpu *vcpu, int *db, int *l)
|
|
|
|
{
|
|
|
|
struct kvm_segment cs;
|
|
|
|
|
2008-05-27 15:18:46 +07:00
|
|
|
kvm_get_segment(vcpu, &cs, VCPU_SREG_CS);
|
2007-11-02 02:16:10 +07:00
|
|
|
*db = cs.db;
|
|
|
|
*l = cs.l;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(kvm_get_cs_db_l_bits);
|
|
|
|
|
|
|
|
int kvm_arch_vcpu_ioctl_get_sregs(struct kvm_vcpu *vcpu,
|
|
|
|
struct kvm_sregs *sregs)
|
|
|
|
{
|
2010-02-16 15:51:48 +07:00
|
|
|
struct desc_ptr dt;
|
2007-11-02 02:16:10 +07:00
|
|
|
|
2008-05-27 15:18:46 +07:00
|
|
|
kvm_get_segment(vcpu, &sregs->cs, VCPU_SREG_CS);
|
|
|
|
kvm_get_segment(vcpu, &sregs->ds, VCPU_SREG_DS);
|
|
|
|
kvm_get_segment(vcpu, &sregs->es, VCPU_SREG_ES);
|
|
|
|
kvm_get_segment(vcpu, &sregs->fs, VCPU_SREG_FS);
|
|
|
|
kvm_get_segment(vcpu, &sregs->gs, VCPU_SREG_GS);
|
|
|
|
kvm_get_segment(vcpu, &sregs->ss, VCPU_SREG_SS);
|
2007-11-02 02:16:10 +07:00
|
|
|
|
2008-05-27 15:18:46 +07:00
|
|
|
kvm_get_segment(vcpu, &sregs->tr, VCPU_SREG_TR);
|
|
|
|
kvm_get_segment(vcpu, &sregs->ldt, VCPU_SREG_LDTR);
|
2007-11-02 02:16:10 +07:00
|
|
|
|
|
|
|
kvm_x86_ops->get_idt(vcpu, &dt);
|
2010-02-16 15:51:48 +07:00
|
|
|
sregs->idt.limit = dt.size;
|
|
|
|
sregs->idt.base = dt.address;
|
2007-11-02 02:16:10 +07:00
|
|
|
kvm_x86_ops->get_gdt(vcpu, &dt);
|
2010-02-16 15:51:48 +07:00
|
|
|
sregs->gdt.limit = dt.size;
|
|
|
|
sregs->gdt.base = dt.address;
|
2007-11-02 02:16:10 +07:00
|
|
|
|
2009-12-29 23:07:30 +07:00
|
|
|
sregs->cr0 = kvm_read_cr0(vcpu);
|
2007-12-13 22:50:52 +07:00
|
|
|
sregs->cr2 = vcpu->arch.cr2;
|
2010-12-05 22:30:00 +07:00
|
|
|
sregs->cr3 = kvm_read_cr3(vcpu);
|
2009-12-07 17:16:48 +07:00
|
|
|
sregs->cr4 = kvm_read_cr4(vcpu);
|
2008-02-24 16:20:43 +07:00
|
|
|
sregs->cr8 = kvm_get_cr8(vcpu);
|
2010-01-21 20:31:50 +07:00
|
|
|
sregs->efer = vcpu->arch.efer;
|
2007-11-02 02:16:10 +07:00
|
|
|
sregs->apic_base = kvm_get_apic_base(vcpu);
|
|
|
|
|
2009-05-11 17:35:48 +07:00
|
|
|
memset(sregs->interrupt_bitmap, 0, sizeof sregs->interrupt_bitmap);
|
2007-11-02 02:16:10 +07:00
|
|
|
|
2009-05-11 17:35:53 +07:00
|
|
|
if (vcpu->arch.interrupt.pending && !vcpu->arch.interrupt.soft)
|
2009-04-21 21:45:11 +07:00
|
|
|
set_bit(vcpu->arch.interrupt.nr,
|
|
|
|
(unsigned long *)sregs->interrupt_bitmap);
|
2009-04-21 21:45:10 +07:00
|
|
|
|
2007-11-02 02:16:10 +07:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2008-04-11 23:24:45 +07:00
|
|
|
int kvm_arch_vcpu_ioctl_get_mpstate(struct kvm_vcpu *vcpu,
|
|
|
|
struct kvm_mp_state *mp_state)
|
|
|
|
{
|
2013-03-13 18:42:34 +07:00
|
|
|
kvm_apic_accept_events(vcpu);
|
2013-08-26 15:48:34 +07:00
|
|
|
if (vcpu->arch.mp_state == KVM_MP_STATE_HALTED &&
|
|
|
|
vcpu->arch.pv.pv_unhalted)
|
|
|
|
mp_state->mp_state = KVM_MP_STATE_RUNNABLE;
|
|
|
|
else
|
|
|
|
mp_state->mp_state = vcpu->arch.mp_state;
|
|
|
|
|
2008-04-11 23:24:45 +07:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
int kvm_arch_vcpu_ioctl_set_mpstate(struct kvm_vcpu *vcpu,
|
|
|
|
struct kvm_mp_state *mp_state)
|
|
|
|
{
|
2016-01-08 19:48:51 +07:00
|
|
|
if (!lapic_in_kernel(vcpu) &&
|
2013-03-13 18:42:34 +07:00
|
|
|
mp_state->mp_state != KVM_MP_STATE_RUNNABLE)
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
if (mp_state->mp_state == KVM_MP_STATE_SIPI_RECEIVED) {
|
|
|
|
vcpu->arch.mp_state = KVM_MP_STATE_INIT_RECEIVED;
|
|
|
|
set_bit(KVM_APIC_SIPI, &vcpu->arch.apic->pending_events);
|
|
|
|
} else
|
|
|
|
vcpu->arch.mp_state = mp_state->mp_state;
|
2010-07-27 16:30:24 +07:00
|
|
|
kvm_make_request(KVM_REQ_EVENT, vcpu);
|
2008-04-11 23:24:45 +07:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2012-02-08 20:34:38 +07:00
|
|
|
int kvm_task_switch(struct kvm_vcpu *vcpu, u16 tss_selector, int idt_index,
|
|
|
|
int reason, bool has_error_code, u32 error_code)
|
2007-11-02 02:16:10 +07:00
|
|
|
{
|
2011-05-29 19:53:48 +07:00
|
|
|
struct x86_emulate_ctxt *ctxt = &vcpu->arch.emulate_ctxt;
|
2010-08-16 04:47:01 +07:00
|
|
|
int ret;
|
2010-01-25 17:01:04 +07:00
|
|
|
|
2010-08-16 04:47:01 +07:00
|
|
|
init_emulate_ctxt(vcpu);
|
2010-02-18 17:15:01 +07:00
|
|
|
|
2012-02-08 20:34:38 +07:00
|
|
|
ret = emulator_task_switch(ctxt, tss_selector, idt_index, reason,
|
2011-05-29 19:53:48 +07:00
|
|
|
has_error_code, error_code);
|
2010-02-18 17:15:01 +07:00
|
|
|
|
|
|
|
if (ret)
|
2010-04-15 16:29:50 +07:00
|
|
|
return EMULATE_FAIL;
|
2008-03-25 04:14:53 +07:00
|
|
|
|
2011-05-29 19:53:48 +07:00
|
|
|
kvm_rip_write(vcpu, ctxt->eip);
|
|
|
|
kvm_set_rflags(vcpu, ctxt->eflags);
|
2010-07-27 16:30:24 +07:00
|
|
|
kvm_make_request(KVM_REQ_EVENT, vcpu);
|
2010-04-15 16:29:50 +07:00
|
|
|
return EMULATE_DONE;
|
2008-03-25 04:14:53 +07:00
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(kvm_task_switch);
|
|
|
|
|
2007-11-02 02:16:10 +07:00
|
|
|
int kvm_arch_vcpu_ioctl_set_sregs(struct kvm_vcpu *vcpu,
|
|
|
|
struct kvm_sregs *sregs)
|
|
|
|
{
|
2014-01-24 22:48:44 +07:00
|
|
|
struct msr_data apic_base_msr;
|
2007-11-02 02:16:10 +07:00
|
|
|
int mmu_reset_needed = 0;
|
2011-01-12 14:39:18 +07:00
|
|
|
int pending_vec, max_bits, idx;
|
2010-02-16 15:51:48 +07:00
|
|
|
struct desc_ptr dt;
|
2007-11-02 02:16:10 +07:00
|
|
|
|
2012-11-07 01:24:07 +07:00
|
|
|
if (!guest_cpuid_has_xsave(vcpu) && (sregs->cr4 & X86_CR4_OSXSAVE))
|
|
|
|
return -EINVAL;
|
|
|
|
|
2010-02-16 15:51:48 +07:00
|
|
|
dt.size = sregs->idt.limit;
|
|
|
|
dt.address = sregs->idt.base;
|
2007-11-02 02:16:10 +07:00
|
|
|
kvm_x86_ops->set_idt(vcpu, &dt);
|
2010-02-16 15:51:48 +07:00
|
|
|
dt.size = sregs->gdt.limit;
|
|
|
|
dt.address = sregs->gdt.base;
|
2007-11-02 02:16:10 +07:00
|
|
|
kvm_x86_ops->set_gdt(vcpu, &dt);
|
|
|
|
|
2007-12-13 22:50:52 +07:00
|
|
|
vcpu->arch.cr2 = sregs->cr2;
|
2010-12-05 22:30:00 +07:00
|
|
|
mmu_reset_needed |= kvm_read_cr3(vcpu) != sregs->cr3;
|
2009-07-02 01:52:03 +07:00
|
|
|
vcpu->arch.cr3 = sregs->cr3;
|
2010-12-05 23:56:11 +07:00
|
|
|
__set_bit(VCPU_EXREG_CR3, (ulong *)&vcpu->arch.regs_avail);
|
2007-11-02 02:16:10 +07:00
|
|
|
|
2008-02-24 16:20:43 +07:00
|
|
|
kvm_set_cr8(vcpu, sregs->cr8);
|
2007-11-02 02:16:10 +07:00
|
|
|
|
2010-01-21 20:31:50 +07:00
|
|
|
mmu_reset_needed |= vcpu->arch.efer != sregs->efer;
|
2007-11-02 02:16:10 +07:00
|
|
|
kvm_x86_ops->set_efer(vcpu, sregs->efer);
|
2014-01-24 22:48:44 +07:00
|
|
|
apic_base_msr.data = sregs->apic_base;
|
|
|
|
apic_base_msr.host_initiated = true;
|
|
|
|
kvm_set_apic_base(vcpu, &apic_base_msr);
|
2007-11-02 02:16:10 +07:00
|
|
|
|
2009-12-29 23:07:30 +07:00
|
|
|
mmu_reset_needed |= kvm_read_cr0(vcpu) != sregs->cr0;
|
2007-11-02 02:16:10 +07:00
|
|
|
kvm_x86_ops->set_cr0(vcpu, sregs->cr0);
|
2008-02-06 18:02:35 +07:00
|
|
|
vcpu->arch.cr0 = sregs->cr0;
|
2007-11-02 02:16:10 +07:00
|
|
|
|
2009-12-07 17:16:48 +07:00
|
|
|
mmu_reset_needed |= kvm_read_cr4(vcpu) != sregs->cr4;
|
2007-11-02 02:16:10 +07:00
|
|
|
kvm_x86_ops->set_cr4(vcpu, sregs->cr4);
|
2016-03-22 15:51:21 +07:00
|
|
|
if (sregs->cr4 & (X86_CR4_OSXSAVE | X86_CR4_PKE))
|
2011-11-23 21:30:32 +07:00
|
|
|
kvm_update_cpuid(vcpu);
|
2011-01-12 14:39:18 +07:00
|
|
|
|
|
|
|
idx = srcu_read_lock(&vcpu->kvm->srcu);
|
2009-10-27 01:48:33 +07:00
|
|
|
if (!is_long_mode(vcpu) && is_pae(vcpu)) {
|
2010-12-05 22:30:00 +07:00
|
|
|
load_pdptrs(vcpu, vcpu->arch.walk_mmu, kvm_read_cr3(vcpu));
|
2009-10-27 01:48:33 +07:00
|
|
|
mmu_reset_needed = 1;
|
|
|
|
}
|
2011-01-12 14:39:18 +07:00
|
|
|
srcu_read_unlock(&vcpu->kvm->srcu, idx);
|
2007-11-02 02:16:10 +07:00
|
|
|
|
|
|
|
if (mmu_reset_needed)
|
|
|
|
kvm_mmu_reset_context(vcpu);
|
|
|
|
|
2012-09-06 00:00:52 +07:00
|
|
|
max_bits = KVM_NR_INTERRUPTS;
|
2009-05-11 17:35:48 +07:00
|
|
|
pending_vec = find_first_bit(
|
|
|
|
(const unsigned long *)sregs->interrupt_bitmap, max_bits);
|
|
|
|
if (pending_vec < max_bits) {
|
2009-05-11 17:35:50 +07:00
|
|
|
kvm_queue_interrupt(vcpu, pending_vec, false);
|
2009-05-11 17:35:48 +07:00
|
|
|
pr_debug("Set back pending irq %d\n", pending_vec);
|
2007-11-02 02:16:10 +07:00
|
|
|
}
|
|
|
|
|
2008-05-27 15:18:46 +07:00
|
|
|
kvm_set_segment(vcpu, &sregs->cs, VCPU_SREG_CS);
|
|
|
|
kvm_set_segment(vcpu, &sregs->ds, VCPU_SREG_DS);
|
|
|
|
kvm_set_segment(vcpu, &sregs->es, VCPU_SREG_ES);
|
|
|
|
kvm_set_segment(vcpu, &sregs->fs, VCPU_SREG_FS);
|
|
|
|
kvm_set_segment(vcpu, &sregs->gs, VCPU_SREG_GS);
|
|
|
|
kvm_set_segment(vcpu, &sregs->ss, VCPU_SREG_SS);
|
2007-11-02 02:16:10 +07:00
|
|
|
|
2008-05-27 15:18:46 +07:00
|
|
|
kvm_set_segment(vcpu, &sregs->tr, VCPU_SREG_TR);
|
|
|
|
kvm_set_segment(vcpu, &sregs->ldt, VCPU_SREG_LDTR);
|
2007-11-02 02:16:10 +07:00
|
|
|
|
2009-08-03 18:58:25 +07:00
|
|
|
update_cr8_intercept(vcpu);
|
|
|
|
|
2008-09-11 02:40:55 +07:00
|
|
|
/* Older userspace won't unhalt the vcpu on reset. */
|
2009-06-09 19:56:26 +07:00
|
|
|
if (kvm_vcpu_is_bsp(vcpu) && kvm_rip_read(vcpu) == 0xfff0 &&
|
2008-09-11 02:40:55 +07:00
|
|
|
sregs->cs.selector == 0xf000 && sregs->cs.base == 0xffff0000 &&
|
2010-01-21 20:31:48 +07:00
|
|
|
!is_protmode(vcpu))
|
2008-09-11 02:40:55 +07:00
|
|
|
vcpu->arch.mp_state = KVM_MP_STATE_RUNNABLE;
|
|
|
|
|
2010-07-27 16:30:24 +07:00
|
|
|
kvm_make_request(KVM_REQ_EVENT, vcpu);
|
|
|
|
|
2007-11-02 02:16:10 +07:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2008-12-15 19:52:10 +07:00
|
|
|
int kvm_arch_vcpu_ioctl_set_guest_debug(struct kvm_vcpu *vcpu,
|
|
|
|
struct kvm_guest_debug *dbg)
|
2007-11-02 02:16:10 +07:00
|
|
|
{
|
2009-10-03 05:31:21 +07:00
|
|
|
unsigned long rflags;
|
2008-12-15 19:52:10 +07:00
|
|
|
int i, r;
|
2007-11-02 02:16:10 +07:00
|
|
|
|
2009-10-30 18:46:59 +07:00
|
|
|
if (dbg->control & (KVM_GUESTDBG_INJECT_DB | KVM_GUESTDBG_INJECT_BP)) {
|
|
|
|
r = -EBUSY;
|
|
|
|
if (vcpu->arch.exception.pending)
|
2010-05-13 15:25:04 +07:00
|
|
|
goto out;
|
2009-10-30 18:46:59 +07:00
|
|
|
if (dbg->control & KVM_GUESTDBG_INJECT_DB)
|
|
|
|
kvm_queue_exception(vcpu, DB_VECTOR);
|
|
|
|
else
|
|
|
|
kvm_queue_exception(vcpu, BP_VECTOR);
|
|
|
|
}
|
|
|
|
|
2009-10-05 18:07:21 +07:00
|
|
|
/*
|
|
|
|
* Read rflags as long as potentially injected trace flags are still
|
|
|
|
* filtered out.
|
|
|
|
*/
|
|
|
|
rflags = kvm_get_rflags(vcpu);
|
2009-10-03 05:31:21 +07:00
|
|
|
|
|
|
|
vcpu->guest_debug = dbg->control;
|
|
|
|
if (!(vcpu->guest_debug & KVM_GUESTDBG_ENABLE))
|
|
|
|
vcpu->guest_debug = 0;
|
|
|
|
|
|
|
|
if (vcpu->guest_debug & KVM_GUESTDBG_USE_HW_BP) {
|
2008-12-15 19:52:10 +07:00
|
|
|
for (i = 0; i < KVM_NR_DB_REGS; ++i)
|
|
|
|
vcpu->arch.eff_db[i] = dbg->arch.debugreg[i];
|
2012-09-21 10:42:55 +07:00
|
|
|
vcpu->arch.guest_debug_dr7 = dbg->arch.debugreg[7];
|
2008-12-15 19:52:10 +07:00
|
|
|
} else {
|
|
|
|
for (i = 0; i < KVM_NR_DB_REGS; i++)
|
|
|
|
vcpu->arch.eff_db[i] = vcpu->arch.db[i];
|
|
|
|
}
|
2012-09-21 10:42:55 +07:00
|
|
|
kvm_update_dr7(vcpu);
|
2008-12-15 19:52:10 +07:00
|
|
|
|
2010-02-23 23:47:55 +07:00
|
|
|
if (vcpu->guest_debug & KVM_GUESTDBG_SINGLESTEP)
|
|
|
|
vcpu->arch.singlestep_rip = kvm_rip_read(vcpu) +
|
|
|
|
get_segment_base(vcpu, VCPU_SREG_CS);
|
2009-10-18 18:24:44 +07:00
|
|
|
|
2009-10-05 18:07:21 +07:00
|
|
|
/*
|
|
|
|
* Trigger an rflags update that will inject or remove the trace
|
|
|
|
* flags.
|
|
|
|
*/
|
|
|
|
kvm_set_rflags(vcpu, rflags);
|
2007-11-02 02:16:10 +07:00
|
|
|
|
2015-11-10 17:55:36 +07:00
|
|
|
kvm_x86_ops->update_bp_intercept(vcpu);
|
2007-11-02 02:16:10 +07:00
|
|
|
|
2009-10-30 18:46:59 +07:00
|
|
|
r = 0;
|
2008-12-15 19:52:10 +07:00
|
|
|
|
2010-05-13 15:25:04 +07:00
|
|
|
out:
|
2007-11-02 02:16:10 +07:00
|
|
|
|
|
|
|
return r;
|
|
|
|
}
|
|
|
|
|
2007-11-16 12:05:55 +07:00
|
|
|
/*
|
|
|
|
* Translate a guest virtual address to a guest physical address.
|
|
|
|
*/
|
|
|
|
int kvm_arch_vcpu_ioctl_translate(struct kvm_vcpu *vcpu,
|
|
|
|
struct kvm_translation *tr)
|
|
|
|
{
|
|
|
|
unsigned long vaddr = tr->linear_address;
|
|
|
|
gpa_t gpa;
|
2009-12-23 23:35:25 +07:00
|
|
|
int idx;
|
2007-11-16 12:05:55 +07:00
|
|
|
|
2009-12-23 23:35:25 +07:00
|
|
|
idx = srcu_read_lock(&vcpu->kvm->srcu);
|
2010-02-10 19:21:32 +07:00
|
|
|
gpa = kvm_mmu_gva_to_gpa_system(vcpu, vaddr, NULL);
|
2009-12-23 23:35:25 +07:00
|
|
|
srcu_read_unlock(&vcpu->kvm->srcu, idx);
|
2007-11-16 12:05:55 +07:00
|
|
|
tr->physical_address = gpa;
|
|
|
|
tr->valid = gpa != UNMAPPED_GVA;
|
|
|
|
tr->writeable = 1;
|
|
|
|
tr->usermode = 0;
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2007-11-01 05:24:25 +07:00
|
|
|
int kvm_arch_vcpu_ioctl_get_fpu(struct kvm_vcpu *vcpu, struct kvm_fpu *fpu)
|
|
|
|
{
|
2015-04-30 22:15:32 +07:00
|
|
|
struct fxregs_state *fxsave =
|
x86/fpu: Simplify FPU handling by embedding the fpstate in task_struct (again)
So 6 years ago we made the FPU fpstate dynamically allocated:
aa283f49276e ("x86, fpu: lazy allocation of FPU area - v5")
61c4628b5386 ("x86, fpu: split FPU state from task struct - v5")
In hindsight this was a mistake:
- it complicated context allocation failure handling, such as:
/* kthread execs. TODO: cleanup this horror. */
if (WARN_ON(fpstate_alloc_init(fpu)))
force_sig(SIGKILL, tsk);
- it caused us to enable irqs in fpu__restore():
local_irq_enable();
/*
* does a slab alloc which can sleep
*/
if (fpstate_alloc_init(fpu)) {
/*
* ran out of memory!
*/
do_group_exit(SIGKILL);
return;
}
local_irq_disable();
- it (slightly) slowed down task creation/destruction by adding
slab allocation/free pattens.
- it made access to context contents (slightly) slower by adding
one more pointer dereference.
The motivation for the dynamic allocation was two-fold:
- reduce memory consumption by non-FPU tasks
- allocate and handle only the necessary amount of context for
various XSAVE processors that have varying hardware frame
sizes.
These days, with glibc using SSE memcpy by default and GCC optimizing
for SSE/AVX by default, the scope of FPU using apps on an x86 system is
much larger than it was 6 years ago.
For example on a freshly installed Fedora 21 desktop system, with a
recent kernel, all non-kthread tasks have used the FPU shortly after
bootup.
Also, even modern embedded x86 CPUs try to support the latest vector
instruction set - so they'll too often use the larger xstate frame
sizes.
So remove the dynamic allocation complication by embedding the FPU
fpstate in task_struct again. This should make the FPU a lot more
accessible to all sorts of atomic contexts.
We could still optimize for the xstate frame size in the future,
by moving the state structure to the last element of task_struct,
and allocating only a part of that.
This change is kept minimal by still keeping the ctx_alloc()/free()
routines (that now do nothing substantial) - we'll remove them in
the following patches.
Reviewed-by: Borislav Petkov <bp@alien8.de>
Cc: Andy Lutomirski <luto@amacapital.net>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
2015-04-27 09:19:39 +07:00
|
|
|
&vcpu->arch.guest_fpu.state.fxsave;
|
2007-11-01 05:24:25 +07:00
|
|
|
|
|
|
|
memcpy(fpu->fpr, fxsave->st_space, 128);
|
|
|
|
fpu->fcw = fxsave->cwd;
|
|
|
|
fpu->fsw = fxsave->swd;
|
|
|
|
fpu->ftwx = fxsave->twd;
|
|
|
|
fpu->last_opcode = fxsave->fop;
|
|
|
|
fpu->last_ip = fxsave->rip;
|
|
|
|
fpu->last_dp = fxsave->rdp;
|
|
|
|
memcpy(fpu->xmm, fxsave->xmm_space, sizeof fxsave->xmm_space);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
int kvm_arch_vcpu_ioctl_set_fpu(struct kvm_vcpu *vcpu, struct kvm_fpu *fpu)
|
|
|
|
{
|
2015-04-30 22:15:32 +07:00
|
|
|
struct fxregs_state *fxsave =
|
x86/fpu: Simplify FPU handling by embedding the fpstate in task_struct (again)
So 6 years ago we made the FPU fpstate dynamically allocated:
aa283f49276e ("x86, fpu: lazy allocation of FPU area - v5")
61c4628b5386 ("x86, fpu: split FPU state from task struct - v5")
In hindsight this was a mistake:
- it complicated context allocation failure handling, such as:
/* kthread execs. TODO: cleanup this horror. */
if (WARN_ON(fpstate_alloc_init(fpu)))
force_sig(SIGKILL, tsk);
- it caused us to enable irqs in fpu__restore():
local_irq_enable();
/*
* does a slab alloc which can sleep
*/
if (fpstate_alloc_init(fpu)) {
/*
* ran out of memory!
*/
do_group_exit(SIGKILL);
return;
}
local_irq_disable();
- it (slightly) slowed down task creation/destruction by adding
slab allocation/free pattens.
- it made access to context contents (slightly) slower by adding
one more pointer dereference.
The motivation for the dynamic allocation was two-fold:
- reduce memory consumption by non-FPU tasks
- allocate and handle only the necessary amount of context for
various XSAVE processors that have varying hardware frame
sizes.
These days, with glibc using SSE memcpy by default and GCC optimizing
for SSE/AVX by default, the scope of FPU using apps on an x86 system is
much larger than it was 6 years ago.
For example on a freshly installed Fedora 21 desktop system, with a
recent kernel, all non-kthread tasks have used the FPU shortly after
bootup.
Also, even modern embedded x86 CPUs try to support the latest vector
instruction set - so they'll too often use the larger xstate frame
sizes.
So remove the dynamic allocation complication by embedding the FPU
fpstate in task_struct again. This should make the FPU a lot more
accessible to all sorts of atomic contexts.
We could still optimize for the xstate frame size in the future,
by moving the state structure to the last element of task_struct,
and allocating only a part of that.
This change is kept minimal by still keeping the ctx_alloc()/free()
routines (that now do nothing substantial) - we'll remove them in
the following patches.
Reviewed-by: Borislav Petkov <bp@alien8.de>
Cc: Andy Lutomirski <luto@amacapital.net>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
2015-04-27 09:19:39 +07:00
|
|
|
&vcpu->arch.guest_fpu.state.fxsave;
|
2007-11-01 05:24:25 +07:00
|
|
|
|
|
|
|
memcpy(fxsave->st_space, fpu->fpr, 128);
|
|
|
|
fxsave->cwd = fpu->fcw;
|
|
|
|
fxsave->swd = fpu->fsw;
|
|
|
|
fxsave->twd = fpu->ftwx;
|
|
|
|
fxsave->fop = fpu->last_opcode;
|
|
|
|
fxsave->rip = fpu->last_ip;
|
|
|
|
fxsave->rdp = fpu->last_dp;
|
|
|
|
memcpy(fxsave->xmm_space, fpu->xmm, sizeof fxsave->xmm_space);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2015-04-27 11:58:22 +07:00
|
|
|
static void fx_init(struct kvm_vcpu *vcpu)
|
2007-11-01 05:24:25 +07:00
|
|
|
{
|
2015-04-30 15:23:42 +07:00
|
|
|
fpstate_init(&vcpu->arch.guest_fpu.state);
|
2016-04-05 03:25:03 +07:00
|
|
|
if (boot_cpu_has(X86_FEATURE_XSAVES))
|
x86/fpu: Simplify FPU handling by embedding the fpstate in task_struct (again)
So 6 years ago we made the FPU fpstate dynamically allocated:
aa283f49276e ("x86, fpu: lazy allocation of FPU area - v5")
61c4628b5386 ("x86, fpu: split FPU state from task struct - v5")
In hindsight this was a mistake:
- it complicated context allocation failure handling, such as:
/* kthread execs. TODO: cleanup this horror. */
if (WARN_ON(fpstate_alloc_init(fpu)))
force_sig(SIGKILL, tsk);
- it caused us to enable irqs in fpu__restore():
local_irq_enable();
/*
* does a slab alloc which can sleep
*/
if (fpstate_alloc_init(fpu)) {
/*
* ran out of memory!
*/
do_group_exit(SIGKILL);
return;
}
local_irq_disable();
- it (slightly) slowed down task creation/destruction by adding
slab allocation/free pattens.
- it made access to context contents (slightly) slower by adding
one more pointer dereference.
The motivation for the dynamic allocation was two-fold:
- reduce memory consumption by non-FPU tasks
- allocate and handle only the necessary amount of context for
various XSAVE processors that have varying hardware frame
sizes.
These days, with glibc using SSE memcpy by default and GCC optimizing
for SSE/AVX by default, the scope of FPU using apps on an x86 system is
much larger than it was 6 years ago.
For example on a freshly installed Fedora 21 desktop system, with a
recent kernel, all non-kthread tasks have used the FPU shortly after
bootup.
Also, even modern embedded x86 CPUs try to support the latest vector
instruction set - so they'll too often use the larger xstate frame
sizes.
So remove the dynamic allocation complication by embedding the FPU
fpstate in task_struct again. This should make the FPU a lot more
accessible to all sorts of atomic contexts.
We could still optimize for the xstate frame size in the future,
by moving the state structure to the last element of task_struct,
and allocating only a part of that.
This change is kept minimal by still keeping the ctx_alloc()/free()
routines (that now do nothing substantial) - we'll remove them in
the following patches.
Reviewed-by: Borislav Petkov <bp@alien8.de>
Cc: Andy Lutomirski <luto@amacapital.net>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
2015-04-27 09:19:39 +07:00
|
|
|
vcpu->arch.guest_fpu.state.xsave.header.xcomp_bv =
|
2014-11-22 01:05:07 +07:00
|
|
|
host_xcr0 | XSTATE_COMPACTION_ENABLED;
|
2007-11-01 05:24:25 +07:00
|
|
|
|
2010-06-10 10:27:12 +07:00
|
|
|
/*
|
|
|
|
* Ensure guest xcr0 is valid for loading
|
|
|
|
*/
|
2015-09-03 06:31:26 +07:00
|
|
|
vcpu->arch.xcr0 = XFEATURE_MASK_FP;
|
2010-06-10 10:27:12 +07:00
|
|
|
|
2007-12-13 22:50:52 +07:00
|
|
|
vcpu->arch.cr0 |= X86_CR0_ET;
|
2007-11-01 05:24:25 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
void kvm_load_guest_fpu(struct kvm_vcpu *vcpu)
|
|
|
|
{
|
2010-01-21 20:31:45 +07:00
|
|
|
if (vcpu->guest_fpu_loaded)
|
2007-11-01 05:24:25 +07:00
|
|
|
return;
|
|
|
|
|
2010-06-10 10:27:12 +07:00
|
|
|
/*
|
|
|
|
* Restore all possible states in the guest,
|
|
|
|
* and assume host would use all available bits.
|
|
|
|
* Guest xcr0 would be loaded later.
|
|
|
|
*/
|
2007-11-01 05:24:25 +07:00
|
|
|
vcpu->guest_fpu_loaded = 1;
|
2012-09-21 01:01:49 +07:00
|
|
|
__kernel_fpu_begin();
|
2015-05-25 16:59:35 +07:00
|
|
|
__copy_kernel_to_fpregs(&vcpu->arch.guest_fpu.state);
|
2010-01-21 20:31:52 +07:00
|
|
|
trace_kvm_fpu(1);
|
2007-11-01 05:24:25 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
void kvm_put_guest_fpu(struct kvm_vcpu *vcpu)
|
|
|
|
{
|
2015-04-23 22:52:37 +07:00
|
|
|
if (!vcpu->guest_fpu_loaded) {
|
|
|
|
vcpu->fpu_counter = 0;
|
2007-11-01 05:24:25 +07:00
|
|
|
return;
|
2015-04-23 22:52:37 +07:00
|
|
|
}
|
2007-11-01 05:24:25 +07:00
|
|
|
|
|
|
|
vcpu->guest_fpu_loaded = 0;
|
x86/fpu: Rename fpu_save_init() to copy_fpregs_to_fpstate()
So fpu_save_init() is a historic name that got its name when the only
way the FPU state was FNSAVE, which cleared (well, destroyed) the FPU
state after saving it.
Nowadays the name is misleading, because ever since the introduction of
FXSAVE (and more modern FPU saving instructions) the 'we need to reload
the FPU state' part is only true if there's a pending FPU exception [*],
which is almost never the case.
So rename it to copy_fpregs_to_fpstate() to make it clear what's
happening. Also add a few comments about why we cannot keep registers
in certain cases.
Also clean up the control flow a bit, to make it more apparent when
we are dropping/keeping FP registers, and to optimize the common
case (of keeping fpregs) some more.
[*] Probably not true anymore, modern instructions always leave the FPU
state intact, even if exceptions are pending: because pending FP
exceptions are posted on the next FP instruction, not asynchronously.
They were truly asynchronous back in the IRQ13 case, and we had to
synchronize with them, but that code is not working anymore: we don't
have IRQ13 mapped in the IDT anymore.
But a cleanup patch is obviously not the place to change subtle behavior.
Reviewed-by: Borislav Petkov <bp@alien8.de>
Cc: Andy Lutomirski <luto@amacapital.net>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
2015-04-27 07:53:16 +07:00
|
|
|
copy_fpregs_to_fpstate(&vcpu->arch.guest_fpu);
|
2012-09-21 01:01:49 +07:00
|
|
|
__kernel_fpu_end();
|
2007-11-18 18:54:33 +07:00
|
|
|
++vcpu->stat.fpu_reload;
|
2015-04-23 22:52:37 +07:00
|
|
|
/*
|
|
|
|
* If using eager FPU mode, or if the guest is a frequent user
|
|
|
|
* of the FPU, just leave the FPU active for next time.
|
|
|
|
* Every 255 times fpu_counter rolls over to 0; a guest that uses
|
|
|
|
* the FPU in bursts will revert to loading it on demand.
|
|
|
|
*/
|
2016-03-08 16:00:11 +07:00
|
|
|
if (!use_eager_fpu()) {
|
2015-04-23 22:52:37 +07:00
|
|
|
if (++vcpu->fpu_counter < 5)
|
|
|
|
kvm_make_request(KVM_REQ_DEACTIVATE_FPU, vcpu);
|
|
|
|
}
|
2010-01-21 20:31:52 +07:00
|
|
|
trace_kvm_fpu(0);
|
2007-11-01 05:24:25 +07:00
|
|
|
}
|
2007-11-14 19:38:21 +07:00
|
|
|
|
|
|
|
void kvm_arch_vcpu_free(struct kvm_vcpu *vcpu)
|
|
|
|
{
|
2011-02-02 02:16:40 +07:00
|
|
|
kvmclock_reset(vcpu);
|
2009-02-25 22:08:31 +07:00
|
|
|
|
2010-06-30 11:25:15 +07:00
|
|
|
free_cpumask_var(vcpu->arch.wbinvd_dirty_mask);
|
2007-11-14 19:38:21 +07:00
|
|
|
kvm_x86_ops->vcpu_free(vcpu);
|
|
|
|
}
|
|
|
|
|
|
|
|
struct kvm_vcpu *kvm_arch_vcpu_create(struct kvm *kvm,
|
|
|
|
unsigned int id)
|
|
|
|
{
|
2015-05-21 03:41:25 +07:00
|
|
|
struct kvm_vcpu *vcpu;
|
|
|
|
|
2010-08-20 15:07:22 +07:00
|
|
|
if (check_tsc_unstable() && atomic_read(&kvm->online_vcpus) != 0)
|
|
|
|
printk_once(KERN_WARNING
|
|
|
|
"kvm: SMP vm created on host with unstable TSC; "
|
|
|
|
"guest TSC will not be reliable\n");
|
2015-05-21 03:41:25 +07:00
|
|
|
|
|
|
|
vcpu = kvm_x86_ops->vcpu_create(kvm, id);
|
|
|
|
|
|
|
|
return vcpu;
|
2007-11-20 20:30:24 +07:00
|
|
|
}
|
2007-11-14 19:38:21 +07:00
|
|
|
|
2007-11-20 20:30:24 +07:00
|
|
|
int kvm_arch_vcpu_setup(struct kvm_vcpu *vcpu)
|
|
|
|
{
|
|
|
|
int r;
|
2007-11-14 19:38:21 +07:00
|
|
|
|
2015-06-15 15:55:31 +07:00
|
|
|
kvm_vcpu_mtrr_init(vcpu);
|
2012-09-16 15:50:30 +07:00
|
|
|
r = vcpu_load(vcpu);
|
|
|
|
if (r)
|
|
|
|
return r;
|
KVM: x86: INIT and reset sequences are different
x86 architecture defines differences between the reset and INIT sequences.
INIT does not initialize the FPU (including MMX, XMM, YMM, etc.), TSC, PMU,
MSRs (in general), MTRRs machine-check, APIC ID, APIC arbitration ID and BSP.
References (from Intel SDM):
"If the MP protocol has completed and a BSP is chosen, subsequent INITs (either
to a specific processor or system wide) do not cause the MP protocol to be
repeated." [8.4.2: MP Initialization Protocol Requirements and Restrictions]
[Table 9-1. IA-32 Processor States Following Power-up, Reset, or INIT]
"If the processor is reset by asserting the INIT# pin, the x87 FPU state is not
changed." [9.2: X87 FPU INITIALIZATION]
"The state of the local APIC following an INIT reset is the same as it is after
a power-up or hardware reset, except that the APIC ID and arbitration ID
registers are not affected." [10.4.7.3: Local APIC State After an INIT Reset
("Wait-for-SIPI" State)]
Signed-off-by: Nadav Amit <namit@cs.technion.ac.il>
Message-Id: <1428924848-28212-1-git-send-email-namit@cs.technion.ac.il>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2015-04-13 18:34:08 +07:00
|
|
|
kvm_vcpu_reset(vcpu, false);
|
2013-10-02 21:56:13 +07:00
|
|
|
kvm_mmu_setup(vcpu);
|
2007-11-14 19:38:21 +07:00
|
|
|
vcpu_put(vcpu);
|
2007-11-20 20:30:24 +07:00
|
|
|
return r;
|
2007-11-14 19:38:21 +07:00
|
|
|
}
|
|
|
|
|
2014-12-04 21:47:07 +07:00
|
|
|
void kvm_arch_vcpu_postcreate(struct kvm_vcpu *vcpu)
|
2012-11-28 08:29:02 +07:00
|
|
|
{
|
2012-11-30 03:42:12 +07:00
|
|
|
struct msr_data msr;
|
2014-02-28 18:52:55 +07:00
|
|
|
struct kvm *kvm = vcpu->kvm;
|
2012-11-28 08:29:02 +07:00
|
|
|
|
2014-12-04 21:47:07 +07:00
|
|
|
if (vcpu_load(vcpu))
|
|
|
|
return;
|
2012-11-30 03:42:12 +07:00
|
|
|
msr.data = 0x0;
|
|
|
|
msr.index = MSR_IA32_TSC;
|
|
|
|
msr.host_initiated = true;
|
|
|
|
kvm_write_tsc(vcpu, &msr);
|
2012-11-28 08:29:02 +07:00
|
|
|
vcpu_put(vcpu);
|
|
|
|
|
2015-05-13 08:42:04 +07:00
|
|
|
if (!kvmclock_periodic_sync)
|
|
|
|
return;
|
|
|
|
|
2014-02-28 18:52:55 +07:00
|
|
|
schedule_delayed_work(&kvm->arch.kvmclock_sync_work,
|
|
|
|
KVMCLOCK_SYNC_PERIOD);
|
2012-11-28 08:29:02 +07:00
|
|
|
}
|
|
|
|
|
2007-11-20 03:04:43 +07:00
|
|
|
void kvm_arch_vcpu_destroy(struct kvm_vcpu *vcpu)
|
2007-11-14 19:38:21 +07:00
|
|
|
{
|
2012-09-16 15:50:30 +07:00
|
|
|
int r;
|
2010-10-14 16:22:50 +07:00
|
|
|
vcpu->arch.apf.msr_val = 0;
|
|
|
|
|
2012-09-16 15:50:30 +07:00
|
|
|
r = vcpu_load(vcpu);
|
|
|
|
BUG_ON(r);
|
2007-11-14 19:38:21 +07:00
|
|
|
kvm_mmu_unload(vcpu);
|
|
|
|
vcpu_put(vcpu);
|
|
|
|
|
|
|
|
kvm_x86_ops->vcpu_free(vcpu);
|
|
|
|
}
|
|
|
|
|
KVM: x86: INIT and reset sequences are different
x86 architecture defines differences between the reset and INIT sequences.
INIT does not initialize the FPU (including MMX, XMM, YMM, etc.), TSC, PMU,
MSRs (in general), MTRRs machine-check, APIC ID, APIC arbitration ID and BSP.
References (from Intel SDM):
"If the MP protocol has completed and a BSP is chosen, subsequent INITs (either
to a specific processor or system wide) do not cause the MP protocol to be
repeated." [8.4.2: MP Initialization Protocol Requirements and Restrictions]
[Table 9-1. IA-32 Processor States Following Power-up, Reset, or INIT]
"If the processor is reset by asserting the INIT# pin, the x87 FPU state is not
changed." [9.2: X87 FPU INITIALIZATION]
"The state of the local APIC following an INIT reset is the same as it is after
a power-up or hardware reset, except that the APIC ID and arbitration ID
registers are not affected." [10.4.7.3: Local APIC State After an INIT Reset
("Wait-for-SIPI" State)]
Signed-off-by: Nadav Amit <namit@cs.technion.ac.il>
Message-Id: <1428924848-28212-1-git-send-email-namit@cs.technion.ac.il>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2015-04-13 18:34:08 +07:00
|
|
|
void kvm_vcpu_reset(struct kvm_vcpu *vcpu, bool init_event)
|
2007-11-14 19:38:21 +07:00
|
|
|
{
|
2015-06-04 15:44:44 +07:00
|
|
|
vcpu->arch.hflags = 0;
|
|
|
|
|
2016-06-02 03:26:00 +07:00
|
|
|
vcpu->arch.smi_pending = 0;
|
2011-09-20 17:43:14 +07:00
|
|
|
atomic_set(&vcpu->arch.nmi_queued, 0);
|
|
|
|
vcpu->arch.nmi_pending = 0;
|
2008-09-26 14:30:48 +07:00
|
|
|
vcpu->arch.nmi_injected = false;
|
2014-06-30 16:03:02 +07:00
|
|
|
kvm_clear_interrupt_queue(vcpu);
|
|
|
|
kvm_clear_exception_queue(vcpu);
|
2008-09-26 14:30:48 +07:00
|
|
|
|
2008-12-15 19:52:10 +07:00
|
|
|
memset(vcpu->arch.db, 0, sizeof(vcpu->arch.db));
|
2015-04-02 07:10:37 +07:00
|
|
|
kvm_update_dr0123(vcpu);
|
2014-07-15 21:37:46 +07:00
|
|
|
vcpu->arch.dr6 = DR6_INIT;
|
2014-01-05 00:47:16 +07:00
|
|
|
kvm_update_dr6(vcpu);
|
2008-12-15 19:52:10 +07:00
|
|
|
vcpu->arch.dr7 = DR7_FIXED_1;
|
2012-09-21 10:42:55 +07:00
|
|
|
kvm_update_dr7(vcpu);
|
2008-12-15 19:52:10 +07:00
|
|
|
|
2015-04-02 07:10:38 +07:00
|
|
|
vcpu->arch.cr2 = 0;
|
|
|
|
|
2010-07-27 16:30:24 +07:00
|
|
|
kvm_make_request(KVM_REQ_EVENT, vcpu);
|
2010-10-14 16:22:50 +07:00
|
|
|
vcpu->arch.apf.msr_val = 0;
|
2011-07-12 02:28:14 +07:00
|
|
|
vcpu->arch.st.msr_val = 0;
|
2010-07-27 16:30:24 +07:00
|
|
|
|
2011-02-02 02:16:40 +07:00
|
|
|
kvmclock_reset(vcpu);
|
|
|
|
|
2010-10-14 16:22:46 +07:00
|
|
|
kvm_clear_async_pf_completion_queue(vcpu);
|
|
|
|
kvm_async_pf_hash_reset(vcpu);
|
|
|
|
vcpu->arch.apf.halted = false;
|
2010-07-27 16:30:24 +07:00
|
|
|
|
2015-05-07 16:36:11 +07:00
|
|
|
if (!init_event) {
|
KVM: x86: INIT and reset sequences are different
x86 architecture defines differences between the reset and INIT sequences.
INIT does not initialize the FPU (including MMX, XMM, YMM, etc.), TSC, PMU,
MSRs (in general), MTRRs machine-check, APIC ID, APIC arbitration ID and BSP.
References (from Intel SDM):
"If the MP protocol has completed and a BSP is chosen, subsequent INITs (either
to a specific processor or system wide) do not cause the MP protocol to be
repeated." [8.4.2: MP Initialization Protocol Requirements and Restrictions]
[Table 9-1. IA-32 Processor States Following Power-up, Reset, or INIT]
"If the processor is reset by asserting the INIT# pin, the x87 FPU state is not
changed." [9.2: X87 FPU INITIALIZATION]
"The state of the local APIC following an INIT reset is the same as it is after
a power-up or hardware reset, except that the APIC ID and arbitration ID
registers are not affected." [10.4.7.3: Local APIC State After an INIT Reset
("Wait-for-SIPI" State)]
Signed-off-by: Nadav Amit <namit@cs.technion.ac.il>
Message-Id: <1428924848-28212-1-git-send-email-namit@cs.technion.ac.il>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2015-04-13 18:34:08 +07:00
|
|
|
kvm_pmu_reset(vcpu);
|
2015-05-07 16:36:11 +07:00
|
|
|
vcpu->arch.smbase = 0x30000;
|
|
|
|
}
|
2011-11-10 19:57:22 +07:00
|
|
|
|
2012-12-05 21:26:19 +07:00
|
|
|
memset(vcpu->arch.regs, 0, sizeof(vcpu->arch.regs));
|
|
|
|
vcpu->arch.regs_avail = ~0;
|
|
|
|
vcpu->arch.regs_dirty = ~0;
|
|
|
|
|
KVM: x86: INIT and reset sequences are different
x86 architecture defines differences between the reset and INIT sequences.
INIT does not initialize the FPU (including MMX, XMM, YMM, etc.), TSC, PMU,
MSRs (in general), MTRRs machine-check, APIC ID, APIC arbitration ID and BSP.
References (from Intel SDM):
"If the MP protocol has completed and a BSP is chosen, subsequent INITs (either
to a specific processor or system wide) do not cause the MP protocol to be
repeated." [8.4.2: MP Initialization Protocol Requirements and Restrictions]
[Table 9-1. IA-32 Processor States Following Power-up, Reset, or INIT]
"If the processor is reset by asserting the INIT# pin, the x87 FPU state is not
changed." [9.2: X87 FPU INITIALIZATION]
"The state of the local APIC following an INIT reset is the same as it is after
a power-up or hardware reset, except that the APIC ID and arbitration ID
registers are not affected." [10.4.7.3: Local APIC State After an INIT Reset
("Wait-for-SIPI" State)]
Signed-off-by: Nadav Amit <namit@cs.technion.ac.il>
Message-Id: <1428924848-28212-1-git-send-email-namit@cs.technion.ac.il>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2015-04-13 18:34:08 +07:00
|
|
|
kvm_x86_ops->vcpu_reset(vcpu, init_event);
|
2007-11-14 19:38:21 +07:00
|
|
|
}
|
|
|
|
|
2014-11-24 20:35:24 +07:00
|
|
|
void kvm_vcpu_deliver_sipi_vector(struct kvm_vcpu *vcpu, u8 vector)
|
2013-03-13 18:42:34 +07:00
|
|
|
{
|
|
|
|
struct kvm_segment cs;
|
|
|
|
|
|
|
|
kvm_get_segment(vcpu, &cs, VCPU_SREG_CS);
|
|
|
|
cs.selector = vector << 8;
|
|
|
|
cs.base = vector << 12;
|
|
|
|
kvm_set_segment(vcpu, &cs, VCPU_SREG_CS);
|
|
|
|
kvm_rip_write(vcpu, 0);
|
2007-11-14 19:38:21 +07:00
|
|
|
}
|
|
|
|
|
2014-08-28 20:13:03 +07:00
|
|
|
int kvm_arch_hardware_enable(void)
|
2007-11-14 19:38:21 +07:00
|
|
|
{
|
2010-08-20 15:07:28 +07:00
|
|
|
struct kvm *kvm;
|
|
|
|
struct kvm_vcpu *vcpu;
|
|
|
|
int i;
|
2012-02-04 00:43:56 +07:00
|
|
|
int ret;
|
|
|
|
u64 local_tsc;
|
|
|
|
u64 max_tsc = 0;
|
|
|
|
bool stable, backwards_tsc = false;
|
2009-09-07 15:12:18 +07:00
|
|
|
|
|
|
|
kvm_shared_msr_cpu_online();
|
2014-08-28 20:13:03 +07:00
|
|
|
ret = kvm_x86_ops->hardware_enable();
|
2012-02-04 00:43:56 +07:00
|
|
|
if (ret != 0)
|
|
|
|
return ret;
|
|
|
|
|
2015-06-25 23:44:07 +07:00
|
|
|
local_tsc = rdtsc();
|
2012-02-04 00:43:56 +07:00
|
|
|
stable = !check_tsc_unstable();
|
|
|
|
list_for_each_entry(kvm, &vm_list, vm_list) {
|
|
|
|
kvm_for_each_vcpu(i, vcpu, kvm) {
|
|
|
|
if (!stable && vcpu->cpu == smp_processor_id())
|
2014-09-12 12:43:19 +07:00
|
|
|
kvm_make_request(KVM_REQ_CLOCK_UPDATE, vcpu);
|
2012-02-04 00:43:56 +07:00
|
|
|
if (stable && vcpu->arch.last_host_tsc > local_tsc) {
|
|
|
|
backwards_tsc = true;
|
|
|
|
if (vcpu->arch.last_host_tsc > max_tsc)
|
|
|
|
max_tsc = vcpu->arch.last_host_tsc;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Sometimes, even reliable TSCs go backwards. This happens on
|
|
|
|
* platforms that reset TSC during suspend or hibernate actions, but
|
|
|
|
* maintain synchronization. We must compensate. Fortunately, we can
|
|
|
|
* detect that condition here, which happens early in CPU bringup,
|
|
|
|
* before any KVM threads can be running. Unfortunately, we can't
|
|
|
|
* bring the TSCs fully up to date with real time, as we aren't yet far
|
|
|
|
* enough into CPU bringup that we know how much real time has actually
|
|
|
|
* elapsed; our helper function, get_kernel_ns() will be using boot
|
|
|
|
* variables that haven't been updated yet.
|
|
|
|
*
|
|
|
|
* So we simply find the maximum observed TSC above, then record the
|
|
|
|
* adjustment to TSC in each VCPU. When the VCPU later gets loaded,
|
|
|
|
* the adjustment will be applied. Note that we accumulate
|
|
|
|
* adjustments, in case multiple suspend cycles happen before some VCPU
|
|
|
|
* gets a chance to run again. In the event that no KVM threads get a
|
|
|
|
* chance to run, we will miss the entire elapsed period, as we'll have
|
|
|
|
* reset last_host_tsc, so VCPUs will not have the TSC adjusted and may
|
|
|
|
* loose cycle time. This isn't too big a deal, since the loss will be
|
|
|
|
* uniform across all VCPUs (not to mention the scenario is extremely
|
|
|
|
* unlikely). It is possible that a second hibernate recovery happens
|
|
|
|
* much faster than a first, causing the observed TSC here to be
|
|
|
|
* smaller; this would require additional padding adjustment, which is
|
|
|
|
* why we set last_host_tsc to the local tsc observed here.
|
|
|
|
*
|
|
|
|
* N.B. - this code below runs only on platforms with reliable TSC,
|
|
|
|
* as that is the only way backwards_tsc is set above. Also note
|
|
|
|
* that this runs for ALL vcpus, which is not a bug; all VCPUs should
|
|
|
|
* have the same delta_cyc adjustment applied if backwards_tsc
|
|
|
|
* is detected. Note further, this adjustment is only done once,
|
|
|
|
* as we reset last_host_tsc on all VCPUs to stop this from being
|
|
|
|
* called multiple times (one for each physical CPU bringup).
|
|
|
|
*
|
2012-06-28 14:17:27 +07:00
|
|
|
* Platforms with unreliable TSCs don't have to deal with this, they
|
2012-02-04 00:43:56 +07:00
|
|
|
* will be compensated by the logic in vcpu_load, which sets the TSC to
|
|
|
|
* catchup mode. This will catchup all VCPUs to real time, but cannot
|
|
|
|
* guarantee that they stay in perfect synchronization.
|
|
|
|
*/
|
|
|
|
if (backwards_tsc) {
|
|
|
|
u64 delta_cyc = max_tsc - local_tsc;
|
2014-05-14 22:43:24 +07:00
|
|
|
backwards_tsc_observed = true;
|
2012-02-04 00:43:56 +07:00
|
|
|
list_for_each_entry(kvm, &vm_list, vm_list) {
|
|
|
|
kvm_for_each_vcpu(i, vcpu, kvm) {
|
|
|
|
vcpu->arch.tsc_offset_adjustment += delta_cyc;
|
|
|
|
vcpu->arch.last_host_tsc = local_tsc;
|
2014-09-12 12:43:19 +07:00
|
|
|
kvm_make_request(KVM_REQ_MASTERCLOCK_UPDATE, vcpu);
|
2012-02-04 00:43:56 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* We have to disable TSC offset matching.. if you were
|
|
|
|
* booting a VM while issuing an S4 host suspend....
|
|
|
|
* you may have some problem. Solving this issue is
|
|
|
|
* left as an exercise to the reader.
|
|
|
|
*/
|
|
|
|
kvm->arch.last_tsc_nsec = 0;
|
|
|
|
kvm->arch.last_tsc_write = 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
}
|
|
|
|
return 0;
|
2007-11-14 19:38:21 +07:00
|
|
|
}
|
|
|
|
|
2014-08-28 20:13:03 +07:00
|
|
|
void kvm_arch_hardware_disable(void)
|
2007-11-14 19:38:21 +07:00
|
|
|
{
|
2014-08-28 20:13:03 +07:00
|
|
|
kvm_x86_ops->hardware_disable();
|
|
|
|
drop_user_return_notifiers();
|
2007-11-14 19:38:21 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
int kvm_arch_hardware_setup(void)
|
|
|
|
{
|
2015-04-13 01:47:15 +07:00
|
|
|
int r;
|
|
|
|
|
|
|
|
r = kvm_x86_ops->hardware_setup();
|
|
|
|
if (r != 0)
|
|
|
|
return r;
|
|
|
|
|
2015-10-20 14:39:03 +07:00
|
|
|
if (kvm_has_tsc_control) {
|
|
|
|
/*
|
|
|
|
* Make sure the user can only configure tsc_khz values that
|
|
|
|
* fit into a signed integer.
|
|
|
|
* A min value is not calculated needed because it will always
|
|
|
|
* be 1 on all machines.
|
|
|
|
*/
|
|
|
|
u64 max = min(0x7fffffffULL,
|
|
|
|
__scale_tsc(kvm_max_tsc_scaling_ratio, tsc_khz));
|
|
|
|
kvm_max_guest_tsc_khz = max;
|
|
|
|
|
2015-10-20 14:39:02 +07:00
|
|
|
kvm_default_tsc_scaling_ratio = 1ULL << kvm_tsc_scaling_ratio_frac_bits;
|
2015-10-20 14:39:03 +07:00
|
|
|
}
|
2015-10-20 14:39:02 +07:00
|
|
|
|
2015-04-13 01:47:15 +07:00
|
|
|
kvm_init_msr_list();
|
|
|
|
return 0;
|
2007-11-14 19:38:21 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
void kvm_arch_hardware_unsetup(void)
|
|
|
|
{
|
|
|
|
kvm_x86_ops->hardware_unsetup();
|
|
|
|
}
|
|
|
|
|
|
|
|
void kvm_arch_check_processor_compat(void *rtn)
|
|
|
|
{
|
|
|
|
kvm_x86_ops->check_processor_compatibility(rtn);
|
2015-07-29 16:56:48 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
bool kvm_vcpu_is_reset_bsp(struct kvm_vcpu *vcpu)
|
|
|
|
{
|
|
|
|
return vcpu->kvm->arch.bsp_vcpu_id == vcpu->vcpu_id;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(kvm_vcpu_is_reset_bsp);
|
|
|
|
|
|
|
|
bool kvm_vcpu_is_bsp(struct kvm_vcpu *vcpu)
|
|
|
|
{
|
|
|
|
return (vcpu->arch.apic_base & MSR_IA32_APICBASE_BSP) != 0;
|
2007-11-14 19:38:21 +07:00
|
|
|
}
|
|
|
|
|
2012-08-05 19:58:32 +07:00
|
|
|
struct static_key kvm_no_apic_vcpu __read_mostly;
|
2016-01-08 19:48:51 +07:00
|
|
|
EXPORT_SYMBOL_GPL(kvm_no_apic_vcpu);
|
2012-08-05 19:58:32 +07:00
|
|
|
|
2007-11-14 19:38:21 +07:00
|
|
|
int kvm_arch_vcpu_init(struct kvm_vcpu *vcpu)
|
|
|
|
{
|
|
|
|
struct page *page;
|
|
|
|
struct kvm *kvm;
|
|
|
|
int r;
|
|
|
|
|
|
|
|
BUG_ON(vcpu->kvm == NULL);
|
|
|
|
kvm = vcpu->kvm;
|
|
|
|
|
2015-11-10 19:36:33 +07:00
|
|
|
vcpu->arch.apicv_active = kvm_x86_ops->get_enable_apicv();
|
2013-08-26 15:48:34 +07:00
|
|
|
vcpu->arch.pv.pv_unhalted = false;
|
2010-07-29 19:11:50 +07:00
|
|
|
vcpu->arch.emulate_ctxt.ops = &emulate_ops;
|
2015-04-02 07:10:36 +07:00
|
|
|
if (!irqchip_in_kernel(kvm) || kvm_vcpu_is_reset_bsp(vcpu))
|
2008-04-13 21:54:35 +07:00
|
|
|
vcpu->arch.mp_state = KVM_MP_STATE_RUNNABLE;
|
2007-11-14 19:38:21 +07:00
|
|
|
else
|
2008-04-13 21:54:35 +07:00
|
|
|
vcpu->arch.mp_state = KVM_MP_STATE_UNINITIALIZED;
|
2007-11-14 19:38:21 +07:00
|
|
|
|
|
|
|
page = alloc_page(GFP_KERNEL | __GFP_ZERO);
|
|
|
|
if (!page) {
|
|
|
|
r = -ENOMEM;
|
|
|
|
goto fail;
|
|
|
|
}
|
2007-12-13 22:50:52 +07:00
|
|
|
vcpu->arch.pio_data = page_address(page);
|
2007-11-14 19:38:21 +07:00
|
|
|
|
KVM: Infrastructure for software and hardware based TSC rate scaling
This requires some restructuring; rather than use 'virtual_tsc_khz'
to indicate whether hardware rate scaling is in effect, we consider
each VCPU to always have a virtual TSC rate. Instead, there is new
logic above the vendor-specific hardware scaling that decides whether
it is even necessary to use and updates all rate variables used by
common code. This means we can simply query the virtual rate at
any point, which is needed for software rate scaling.
There is also now a threshold added to the TSC rate scaling; minor
differences and variations of measured TSC rate can accidentally
provoke rate scaling to be used when it is not needed. Instead,
we have a tolerance variable called tsc_tolerance_ppm, which is
the maximum variation from user requested rate at which scaling
will be used. The default is 250ppm, which is the half the
threshold for NTP adjustment, allowing for some hardware variation.
In the event that hardware rate scaling is not available, we can
kludge a bit by forcing TSC catchup to turn on when a faster than
hardware speed has been requested, but there is nothing available
yet for the reverse case; this requires a trap and emulate software
implementation for RDTSC, which is still forthcoming.
[avi: fix 64-bit division on i386]
Signed-off-by: Zachary Amsden <zamsden@gmail.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
2012-02-04 00:43:50 +07:00
|
|
|
kvm_set_tsc_khz(vcpu, max_tsc_khz);
|
2010-09-19 07:38:15 +07:00
|
|
|
|
2007-11-14 19:38:21 +07:00
|
|
|
r = kvm_mmu_create(vcpu);
|
|
|
|
if (r < 0)
|
|
|
|
goto fail_free_pio_data;
|
|
|
|
|
|
|
|
if (irqchip_in_kernel(kvm)) {
|
|
|
|
r = kvm_create_lapic(vcpu);
|
|
|
|
if (r < 0)
|
|
|
|
goto fail_mmu_destroy;
|
2012-08-05 19:58:32 +07:00
|
|
|
} else
|
|
|
|
static_key_slow_inc(&kvm_no_apic_vcpu);
|
2007-11-14 19:38:21 +07:00
|
|
|
|
2009-05-11 15:48:15 +07:00
|
|
|
vcpu->arch.mce_banks = kzalloc(KVM_MAX_MCE_BANKS * sizeof(u64) * 4,
|
|
|
|
GFP_KERNEL);
|
|
|
|
if (!vcpu->arch.mce_banks) {
|
|
|
|
r = -ENOMEM;
|
2010-01-22 13:21:29 +07:00
|
|
|
goto fail_free_lapic;
|
2009-05-11 15:48:15 +07:00
|
|
|
}
|
|
|
|
vcpu->arch.mcg_cap = KVM_MAX_MCE_BANKS;
|
|
|
|
|
2013-04-18 06:41:00 +07:00
|
|
|
if (!zalloc_cpumask_var(&vcpu->arch.wbinvd_dirty_mask, GFP_KERNEL)) {
|
|
|
|
r = -ENOMEM;
|
2010-06-30 11:25:15 +07:00
|
|
|
goto fail_free_mce_banks;
|
2013-04-18 06:41:00 +07:00
|
|
|
}
|
2010-06-30 11:25:15 +07:00
|
|
|
|
2015-04-27 11:58:22 +07:00
|
|
|
fx_init(vcpu);
|
2012-12-05 21:26:19 +07:00
|
|
|
|
2012-11-30 03:42:50 +07:00
|
|
|
vcpu->arch.ia32_tsc_adjust_msr = 0x0;
|
2013-02-21 05:48:10 +07:00
|
|
|
vcpu->arch.pv_time_enabled = false;
|
2013-10-02 21:06:15 +07:00
|
|
|
|
|
|
|
vcpu->arch.guest_supported_xcr0 = 0;
|
2013-10-02 21:06:16 +07:00
|
|
|
vcpu->arch.guest_xstate_size = XSAVE_HDR_SIZE + XSAVE_HDR_OFFSET;
|
2013-10-02 21:06:15 +07:00
|
|
|
|
2015-03-30 03:56:12 +07:00
|
|
|
vcpu->arch.maxphyaddr = cpuid_query_maxphyaddr(vcpu);
|
|
|
|
|
2015-04-27 20:11:25 +07:00
|
|
|
vcpu->arch.pat = MSR_IA32_CR_PAT_DEFAULT;
|
|
|
|
|
2010-10-14 16:22:46 +07:00
|
|
|
kvm_async_pf_hash_reset(vcpu);
|
2011-11-10 19:57:22 +07:00
|
|
|
kvm_pmu_init(vcpu);
|
2010-10-14 16:22:46 +07:00
|
|
|
|
KVM: x86: Add support for local interrupt requests from userspace
In order to enable userspace PIC support, the userspace PIC needs to
be able to inject local interrupts even when the APICs are in the
kernel.
KVM_INTERRUPT now supports sending local interrupts to an APIC when
APICs are in the kernel.
The ready_for_interrupt_request flag is now only set when the CPU/APIC
will immediately accept and inject an interrupt (i.e. APIC has not
masked the PIC).
When the PIC wishes to initiate an INTA cycle with, say, CPU0, it
kicks CPU0 out of the guest, and renedezvous with CPU0 once it arrives
in userspace.
When the CPU/APIC unmasks the PIC, a KVM_EXIT_IRQ_WINDOW_OPEN is
triggered, so that userspace has a chance to inject a PIC interrupt
if it had been pending.
Overall, this design can lead to a small number of spurious userspace
renedezvous. In particular, whenever the PIC transistions from low to
high while it is masked and whenever the PIC becomes unmasked while
it is low.
Note: this does not buffer more than one local interrupt in the
kernel, so the VMM needs to enter the guest in order to complete
interrupt injection before injecting an additional interrupt.
Compiles for x86.
Can pass the KVM Unit Tests.
Signed-off-by: Steve Rutherford <srutherford@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2015-07-30 16:27:16 +07:00
|
|
|
vcpu->arch.pending_external_vector = -1;
|
|
|
|
|
2015-11-10 19:36:34 +07:00
|
|
|
kvm_hv_vcpu_init(vcpu);
|
|
|
|
|
2007-11-14 19:38:21 +07:00
|
|
|
return 0;
|
2015-04-27 11:58:22 +07:00
|
|
|
|
2010-06-30 11:25:15 +07:00
|
|
|
fail_free_mce_banks:
|
|
|
|
kfree(vcpu->arch.mce_banks);
|
2010-01-22 13:21:29 +07:00
|
|
|
fail_free_lapic:
|
|
|
|
kvm_free_lapic(vcpu);
|
2007-11-14 19:38:21 +07:00
|
|
|
fail_mmu_destroy:
|
|
|
|
kvm_mmu_destroy(vcpu);
|
|
|
|
fail_free_pio_data:
|
2007-12-13 22:50:52 +07:00
|
|
|
free_page((unsigned long)vcpu->arch.pio_data);
|
2007-11-14 19:38:21 +07:00
|
|
|
fail:
|
|
|
|
return r;
|
|
|
|
}
|
|
|
|
|
|
|
|
void kvm_arch_vcpu_uninit(struct kvm_vcpu *vcpu)
|
|
|
|
{
|
2009-12-23 23:35:25 +07:00
|
|
|
int idx;
|
|
|
|
|
2015-11-30 23:22:21 +07:00
|
|
|
kvm_hv_vcpu_uninit(vcpu);
|
2011-11-10 19:57:22 +07:00
|
|
|
kvm_pmu_destroy(vcpu);
|
2010-01-22 13:18:47 +07:00
|
|
|
kfree(vcpu->arch.mce_banks);
|
2007-11-14 19:38:21 +07:00
|
|
|
kvm_free_lapic(vcpu);
|
2009-12-23 23:35:25 +07:00
|
|
|
idx = srcu_read_lock(&vcpu->kvm->srcu);
|
2007-11-14 19:38:21 +07:00
|
|
|
kvm_mmu_destroy(vcpu);
|
2009-12-23 23:35:25 +07:00
|
|
|
srcu_read_unlock(&vcpu->kvm->srcu, idx);
|
2007-12-13 22:50:52 +07:00
|
|
|
free_page((unsigned long)vcpu->arch.pio_data);
|
2015-07-29 17:05:37 +07:00
|
|
|
if (!lapic_in_kernel(vcpu))
|
2012-08-05 19:58:32 +07:00
|
|
|
static_key_slow_dec(&kvm_no_apic_vcpu);
|
2007-11-14 19:38:21 +07:00
|
|
|
}
|
2007-11-18 17:43:45 +07:00
|
|
|
|
2014-08-21 23:08:05 +07:00
|
|
|
void kvm_arch_sched_in(struct kvm_vcpu *vcpu, int cpu)
|
|
|
|
{
|
2014-08-21 23:08:06 +07:00
|
|
|
kvm_x86_ops->sched_in(vcpu, cpu);
|
2014-08-21 23:08:05 +07:00
|
|
|
}
|
|
|
|
|
2012-01-04 16:25:20 +07:00
|
|
|
int kvm_arch_init_vm(struct kvm *kvm, unsigned long type)
|
2007-11-18 17:43:45 +07:00
|
|
|
{
|
2012-01-04 16:25:20 +07:00
|
|
|
if (type)
|
|
|
|
return -EINVAL;
|
|
|
|
|
2014-11-20 19:45:31 +07:00
|
|
|
INIT_HLIST_HEAD(&kvm->arch.mask_notifier_list);
|
2007-12-14 09:01:48 +07:00
|
|
|
INIT_LIST_HEAD(&kvm->arch.active_mmu_pages);
|
2013-05-31 07:36:29 +07:00
|
|
|
INIT_LIST_HEAD(&kvm->arch.zapped_obsolete_pages);
|
2008-07-28 23:26:26 +07:00
|
|
|
INIT_LIST_HEAD(&kvm->arch.assigned_dev_head);
|
2013-10-31 00:02:30 +07:00
|
|
|
atomic_set(&kvm->arch.noncoherent_dma_count, 0);
|
2007-11-18 17:43:45 +07:00
|
|
|
|
2008-10-15 19:15:06 +07:00
|
|
|
/* Reserve bit 0 of irq_sources_bitmap for userspace irq source */
|
|
|
|
set_bit(KVM_USERSPACE_IRQ_SOURCE_ID, &kvm->arch.irq_sources_bitmap);
|
2012-09-22 00:58:03 +07:00
|
|
|
/* Reserve bit 1 of irq_sources_bitmap for irqfd-resampler */
|
|
|
|
set_bit(KVM_IRQFD_RESAMPLE_IRQ_SOURCE_ID,
|
|
|
|
&kvm->arch.irq_sources_bitmap);
|
2008-10-15 19:15:06 +07:00
|
|
|
|
2011-02-04 16:49:11 +07:00
|
|
|
raw_spin_lock_init(&kvm->arch.tsc_write_lock);
|
2012-09-13 21:19:24 +07:00
|
|
|
mutex_init(&kvm->arch.apic_map_lock);
|
2012-11-28 08:29:01 +07:00
|
|
|
spin_lock_init(&kvm->arch.pvclock_gtod_sync_lock);
|
|
|
|
|
|
|
|
pvclock_update_vm_gtod_copy(kvm);
|
2008-12-12 02:45:05 +07:00
|
|
|
|
2014-02-28 18:52:54 +07:00
|
|
|
INIT_DELAYED_WORK(&kvm->arch.kvmclock_update_work, kvmclock_update_fn);
|
2014-02-28 18:52:55 +07:00
|
|
|
INIT_DELAYED_WORK(&kvm->arch.kvmclock_sync_work, kvmclock_sync_fn);
|
2014-02-28 18:52:54 +07:00
|
|
|
|
2016-02-24 16:51:13 +07:00
|
|
|
kvm_page_track_init(kvm);
|
2016-02-24 16:51:16 +07:00
|
|
|
kvm_mmu_init_vm(kvm);
|
2016-02-24 16:51:13 +07:00
|
|
|
|
2016-05-05 02:09:42 +07:00
|
|
|
if (kvm_x86_ops->vm_init)
|
|
|
|
return kvm_x86_ops->vm_init(kvm);
|
|
|
|
|
2010-11-09 23:02:49 +07:00
|
|
|
return 0;
|
2007-11-18 17:43:45 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
static void kvm_unload_vcpu_mmu(struct kvm_vcpu *vcpu)
|
|
|
|
{
|
2012-09-16 15:50:30 +07:00
|
|
|
int r;
|
|
|
|
r = vcpu_load(vcpu);
|
|
|
|
BUG_ON(r);
|
2007-11-18 17:43:45 +07:00
|
|
|
kvm_mmu_unload(vcpu);
|
|
|
|
vcpu_put(vcpu);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void kvm_free_vcpus(struct kvm *kvm)
|
|
|
|
{
|
|
|
|
unsigned int i;
|
2009-06-09 19:56:29 +07:00
|
|
|
struct kvm_vcpu *vcpu;
|
2007-11-18 17:43:45 +07:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Unpin any mmu pages first.
|
|
|
|
*/
|
2010-10-14 16:22:46 +07:00
|
|
|
kvm_for_each_vcpu(i, vcpu, kvm) {
|
|
|
|
kvm_clear_async_pf_completion_queue(vcpu);
|
2009-06-09 19:56:29 +07:00
|
|
|
kvm_unload_vcpu_mmu(vcpu);
|
2010-10-14 16:22:46 +07:00
|
|
|
}
|
2009-06-09 19:56:29 +07:00
|
|
|
kvm_for_each_vcpu(i, vcpu, kvm)
|
|
|
|
kvm_arch_vcpu_free(vcpu);
|
|
|
|
|
|
|
|
mutex_lock(&kvm->lock);
|
|
|
|
for (i = 0; i < atomic_read(&kvm->online_vcpus); i++)
|
|
|
|
kvm->vcpus[i] = NULL;
|
2007-11-18 17:43:45 +07:00
|
|
|
|
2009-06-09 19:56:29 +07:00
|
|
|
atomic_set(&kvm->online_vcpus, 0);
|
|
|
|
mutex_unlock(&kvm->lock);
|
2007-11-18 17:43:45 +07:00
|
|
|
}
|
|
|
|
|
2009-01-06 09:03:02 +07:00
|
|
|
void kvm_arch_sync_events(struct kvm *kvm)
|
|
|
|
{
|
2014-02-28 18:52:55 +07:00
|
|
|
cancel_delayed_work_sync(&kvm->arch.kvmclock_sync_work);
|
2014-02-28 18:52:54 +07:00
|
|
|
cancel_delayed_work_sync(&kvm->arch.kvmclock_update_work);
|
2009-01-06 09:03:03 +07:00
|
|
|
kvm_free_all_assigned_devices(kvm);
|
2010-07-10 16:37:56 +07:00
|
|
|
kvm_free_pit(kvm);
|
2009-01-06 09:03:02 +07:00
|
|
|
}
|
|
|
|
|
2015-10-12 18:38:32 +07:00
|
|
|
int __x86_set_memory_region(struct kvm *kvm, int id, gpa_t gpa, u32 size)
|
2015-05-18 18:33:16 +07:00
|
|
|
{
|
|
|
|
int i, r;
|
2015-10-14 20:51:08 +07:00
|
|
|
unsigned long hva;
|
2015-10-12 18:56:27 +07:00
|
|
|
struct kvm_memslots *slots = kvm_memslots(kvm);
|
|
|
|
struct kvm_memory_slot *slot, old;
|
2015-05-18 18:33:16 +07:00
|
|
|
|
|
|
|
/* Called with kvm->slots_lock held. */
|
2015-10-12 18:38:32 +07:00
|
|
|
if (WARN_ON(id >= KVM_MEM_SLOTS_NUM))
|
|
|
|
return -EINVAL;
|
2015-05-18 18:33:16 +07:00
|
|
|
|
2015-10-12 18:56:27 +07:00
|
|
|
slot = id_to_memslot(slots, id);
|
|
|
|
if (size) {
|
kvm: x86: avoid warning on repeated KVM_SET_TSS_ADDR
Found by syzkaller:
WARNING: CPU: 3 PID: 15175 at arch/x86/kvm/x86.c:7705 __x86_set_memory_region+0x1dc/0x1f0 [kvm]()
CPU: 3 PID: 15175 Comm: a.out Tainted: G W 4.4.6-300.fc23.x86_64 #1
Hardware name: LENOVO 2325F51/2325F51, BIOS G2ET32WW (1.12 ) 05/30/2012
0000000000000286 00000000950899a7 ffff88011ab3fbf0 ffffffff813b542e
0000000000000000 ffffffffa0966496 ffff88011ab3fc28 ffffffff810a40f2
00000000000001fd 0000000000003000 ffff88014fc50000 0000000000000000
Call Trace:
[<ffffffff813b542e>] dump_stack+0x63/0x85
[<ffffffff810a40f2>] warn_slowpath_common+0x82/0xc0
[<ffffffff810a423a>] warn_slowpath_null+0x1a/0x20
[<ffffffffa09251cc>] __x86_set_memory_region+0x1dc/0x1f0 [kvm]
[<ffffffffa092521b>] x86_set_memory_region+0x3b/0x60 [kvm]
[<ffffffffa09bb61c>] vmx_set_tss_addr+0x3c/0x150 [kvm_intel]
[<ffffffffa092f4d4>] kvm_arch_vm_ioctl+0x654/0xbc0 [kvm]
[<ffffffffa091d31a>] kvm_vm_ioctl+0x9a/0x6f0 [kvm]
[<ffffffff81241248>] do_vfs_ioctl+0x298/0x480
[<ffffffff812414a9>] SyS_ioctl+0x79/0x90
[<ffffffff817a04ee>] entry_SYSCALL_64_fastpath+0x12/0x71
Testcase:
#include <unistd.h>
#include <sys/ioctl.h>
#include <fcntl.h>
#include <string.h>
#include <linux/kvm.h>
long r[8];
int main()
{
memset(r, -1, sizeof(r));
r[2] = open("/dev/kvm", O_RDONLY|O_TRUNC);
r[3] = ioctl(r[2], KVM_CREATE_VM, 0x0ul);
r[5] = ioctl(r[3], KVM_SET_TSS_ADDR, 0x20000000ul);
r[7] = ioctl(r[3], KVM_SET_TSS_ADDR, 0x20000000ul);
return 0;
}
Reported-by: Dmitry Vyukov <dvyukov@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Radim Krčmář <rkrcmar@redhat.com>
2016-06-01 19:09:18 +07:00
|
|
|
if (slot->npages)
|
2015-10-12 18:56:27 +07:00
|
|
|
return -EEXIST;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* MAP_SHARED to prevent internal slot pages from being moved
|
|
|
|
* by fork()/COW.
|
|
|
|
*/
|
|
|
|
hva = vm_mmap(NULL, 0, size, PROT_READ | PROT_WRITE,
|
|
|
|
MAP_SHARED | MAP_ANONYMOUS, 0);
|
|
|
|
if (IS_ERR((void *)hva))
|
|
|
|
return PTR_ERR((void *)hva);
|
|
|
|
} else {
|
|
|
|
if (!slot->npages)
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
hva = 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
old = *slot;
|
2015-05-18 18:33:16 +07:00
|
|
|
for (i = 0; i < KVM_ADDRESS_SPACE_NUM; i++) {
|
2015-10-12 18:38:32 +07:00
|
|
|
struct kvm_userspace_memory_region m;
|
2015-05-18 18:33:16 +07:00
|
|
|
|
2015-10-12 18:38:32 +07:00
|
|
|
m.slot = id | (i << 16);
|
|
|
|
m.flags = 0;
|
|
|
|
m.guest_phys_addr = gpa;
|
2015-10-12 18:56:27 +07:00
|
|
|
m.userspace_addr = hva;
|
2015-10-12 18:38:32 +07:00
|
|
|
m.memory_size = size;
|
2015-05-18 18:33:16 +07:00
|
|
|
r = __kvm_set_memory_region(kvm, &m);
|
|
|
|
if (r < 0)
|
|
|
|
return r;
|
|
|
|
}
|
|
|
|
|
2015-10-12 18:56:27 +07:00
|
|
|
if (!size) {
|
|
|
|
r = vm_munmap(old.userspace_addr, old.npages * PAGE_SIZE);
|
|
|
|
WARN_ON(r < 0);
|
|
|
|
}
|
|
|
|
|
2015-05-18 18:33:16 +07:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(__x86_set_memory_region);
|
|
|
|
|
2015-10-12 18:38:32 +07:00
|
|
|
int x86_set_memory_region(struct kvm *kvm, int id, gpa_t gpa, u32 size)
|
2015-05-18 18:33:16 +07:00
|
|
|
{
|
|
|
|
int r;
|
|
|
|
|
|
|
|
mutex_lock(&kvm->slots_lock);
|
2015-10-12 18:38:32 +07:00
|
|
|
r = __x86_set_memory_region(kvm, id, gpa, size);
|
2015-05-18 18:33:16 +07:00
|
|
|
mutex_unlock(&kvm->slots_lock);
|
|
|
|
|
|
|
|
return r;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(x86_set_memory_region);
|
|
|
|
|
2007-11-18 17:43:45 +07:00
|
|
|
void kvm_arch_destroy_vm(struct kvm *kvm)
|
|
|
|
{
|
2013-04-18 23:38:14 +07:00
|
|
|
if (current->mm == kvm->mm) {
|
|
|
|
/*
|
|
|
|
* Free memory regions allocated on behalf of userspace,
|
|
|
|
* unless the the memory map has changed due to process exit
|
|
|
|
* or fd copying.
|
|
|
|
*/
|
2015-10-12 18:38:32 +07:00
|
|
|
x86_set_memory_region(kvm, APIC_ACCESS_PAGE_PRIVATE_MEMSLOT, 0, 0);
|
|
|
|
x86_set_memory_region(kvm, IDENTITY_PAGETABLE_PRIVATE_MEMSLOT, 0, 0);
|
|
|
|
x86_set_memory_region(kvm, TSS_PRIVATE_MEMSLOT, 0, 0);
|
2013-04-18 23:38:14 +07:00
|
|
|
}
|
2016-05-05 02:09:42 +07:00
|
|
|
if (kvm_x86_ops->vm_destroy)
|
|
|
|
kvm_x86_ops->vm_destroy(kvm);
|
2008-10-31 11:37:41 +07:00
|
|
|
kvm_iommu_unmap_guest(kvm);
|
2007-12-14 09:17:34 +07:00
|
|
|
kfree(kvm->arch.vpic);
|
|
|
|
kfree(kvm->arch.vioapic);
|
2007-11-18 17:43:45 +07:00
|
|
|
kvm_free_vcpus(kvm);
|
2016-07-13 03:09:30 +07:00
|
|
|
kvfree(rcu_dereference_check(kvm->arch.apic_map, 1));
|
2016-02-24 16:51:16 +07:00
|
|
|
kvm_mmu_uninit_vm(kvm);
|
2007-11-18 17:43:45 +07:00
|
|
|
}
|
2007-11-20 15:25:04 +07:00
|
|
|
|
2013-10-07 23:48:00 +07:00
|
|
|
void kvm_arch_free_memslot(struct kvm *kvm, struct kvm_memory_slot *free,
|
2012-02-08 11:02:18 +07:00
|
|
|
struct kvm_memory_slot *dont)
|
|
|
|
{
|
|
|
|
int i;
|
|
|
|
|
2012-08-01 16:03:28 +07:00
|
|
|
for (i = 0; i < KVM_NR_PAGE_SIZES; ++i) {
|
|
|
|
if (!dont || free->arch.rmap[i] != dont->arch.rmap[i]) {
|
2015-02-25 03:29:25 +07:00
|
|
|
kvfree(free->arch.rmap[i]);
|
2012-08-01 16:03:28 +07:00
|
|
|
free->arch.rmap[i] = NULL;
|
2012-07-02 15:57:17 +07:00
|
|
|
}
|
2012-08-01 16:03:28 +07:00
|
|
|
if (i == 0)
|
|
|
|
continue;
|
|
|
|
|
|
|
|
if (!dont || free->arch.lpage_info[i - 1] !=
|
|
|
|
dont->arch.lpage_info[i - 1]) {
|
2015-02-25 03:29:25 +07:00
|
|
|
kvfree(free->arch.lpage_info[i - 1]);
|
2012-08-01 16:03:28 +07:00
|
|
|
free->arch.lpage_info[i - 1] = NULL;
|
2012-02-08 11:02:18 +07:00
|
|
|
}
|
|
|
|
}
|
KVM: page track: add the framework of guest page tracking
The array, gfn_track[mode][gfn], is introduced in memory slot for every
guest page, this is the tracking count for the gust page on different
modes. If the page is tracked then the count is increased, the page is
not tracked after the count reaches zero
We use 'unsigned short' as the tracking count which should be enough as
shadow page table only can use 2^14 (2^3 for level, 2^1 for cr4_pae, 2^2
for quadrant, 2^3 for access, 2^1 for nxe, 2^1 for cr0_wp, 2^1 for
smep_andnot_wp, 2^1 for smap_andnot_wp, and 2^1 for smm) at most, there
is enough room for other trackers
Two callbacks, kvm_page_track_create_memslot() and
kvm_page_track_free_memslot() are implemented in this patch, they are
internally used to initialize and reclaim the memory of the array
Currently, only write track mode is supported
Signed-off-by: Xiao Guangrong <guangrong.xiao@linux.intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2016-02-24 16:51:09 +07:00
|
|
|
|
|
|
|
kvm_page_track_free_memslot(free, dont);
|
2012-02-08 11:02:18 +07:00
|
|
|
}
|
|
|
|
|
2013-10-07 23:48:00 +07:00
|
|
|
int kvm_arch_create_memslot(struct kvm *kvm, struct kvm_memory_slot *slot,
|
|
|
|
unsigned long npages)
|
2012-02-08 11:02:18 +07:00
|
|
|
{
|
|
|
|
int i;
|
|
|
|
|
2012-08-01 16:03:28 +07:00
|
|
|
for (i = 0; i < KVM_NR_PAGE_SIZES; ++i) {
|
2016-02-24 16:51:06 +07:00
|
|
|
struct kvm_lpage_info *linfo;
|
2012-02-08 11:02:18 +07:00
|
|
|
unsigned long ugfn;
|
|
|
|
int lpages;
|
2012-08-01 16:03:28 +07:00
|
|
|
int level = i + 1;
|
2012-02-08 11:02:18 +07:00
|
|
|
|
|
|
|
lpages = gfn_to_index(slot->base_gfn + npages - 1,
|
|
|
|
slot->base_gfn, level) + 1;
|
|
|
|
|
2012-08-01 16:03:28 +07:00
|
|
|
slot->arch.rmap[i] =
|
|
|
|
kvm_kvzalloc(lpages * sizeof(*slot->arch.rmap[i]));
|
|
|
|
if (!slot->arch.rmap[i])
|
2012-07-02 15:57:17 +07:00
|
|
|
goto out_free;
|
2012-08-01 16:03:28 +07:00
|
|
|
if (i == 0)
|
|
|
|
continue;
|
2012-07-02 15:57:17 +07:00
|
|
|
|
2016-02-24 16:51:06 +07:00
|
|
|
linfo = kvm_kvzalloc(lpages * sizeof(*linfo));
|
|
|
|
if (!linfo)
|
2012-02-08 11:02:18 +07:00
|
|
|
goto out_free;
|
|
|
|
|
2016-02-24 16:51:06 +07:00
|
|
|
slot->arch.lpage_info[i - 1] = linfo;
|
|
|
|
|
2012-02-08 11:02:18 +07:00
|
|
|
if (slot->base_gfn & (KVM_PAGES_PER_HPAGE(level) - 1))
|
2016-02-24 16:51:06 +07:00
|
|
|
linfo[0].disallow_lpage = 1;
|
2012-02-08 11:02:18 +07:00
|
|
|
if ((slot->base_gfn + npages) & (KVM_PAGES_PER_HPAGE(level) - 1))
|
2016-02-24 16:51:06 +07:00
|
|
|
linfo[lpages - 1].disallow_lpage = 1;
|
2012-02-08 11:02:18 +07:00
|
|
|
ugfn = slot->userspace_addr >> PAGE_SHIFT;
|
|
|
|
/*
|
|
|
|
* If the gfn and userspace address are not aligned wrt each
|
|
|
|
* other, or if explicitly asked to, disable large page
|
|
|
|
* support for this slot
|
|
|
|
*/
|
|
|
|
if ((slot->base_gfn ^ ugfn) & (KVM_PAGES_PER_HPAGE(level) - 1) ||
|
|
|
|
!kvm_largepages_enabled()) {
|
|
|
|
unsigned long j;
|
|
|
|
|
|
|
|
for (j = 0; j < lpages; ++j)
|
2016-02-24 16:51:06 +07:00
|
|
|
linfo[j].disallow_lpage = 1;
|
2012-02-08 11:02:18 +07:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
KVM: page track: add the framework of guest page tracking
The array, gfn_track[mode][gfn], is introduced in memory slot for every
guest page, this is the tracking count for the gust page on different
modes. If the page is tracked then the count is increased, the page is
not tracked after the count reaches zero
We use 'unsigned short' as the tracking count which should be enough as
shadow page table only can use 2^14 (2^3 for level, 2^1 for cr4_pae, 2^2
for quadrant, 2^3 for access, 2^1 for nxe, 2^1 for cr0_wp, 2^1 for
smep_andnot_wp, 2^1 for smap_andnot_wp, and 2^1 for smm) at most, there
is enough room for other trackers
Two callbacks, kvm_page_track_create_memslot() and
kvm_page_track_free_memslot() are implemented in this patch, they are
internally used to initialize and reclaim the memory of the array
Currently, only write track mode is supported
Signed-off-by: Xiao Guangrong <guangrong.xiao@linux.intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2016-02-24 16:51:09 +07:00
|
|
|
if (kvm_page_track_create_memslot(slot, npages))
|
|
|
|
goto out_free;
|
|
|
|
|
2012-02-08 11:02:18 +07:00
|
|
|
return 0;
|
|
|
|
|
|
|
|
out_free:
|
2012-08-01 16:03:28 +07:00
|
|
|
for (i = 0; i < KVM_NR_PAGE_SIZES; ++i) {
|
2015-02-25 03:29:25 +07:00
|
|
|
kvfree(slot->arch.rmap[i]);
|
2012-08-01 16:03:28 +07:00
|
|
|
slot->arch.rmap[i] = NULL;
|
|
|
|
if (i == 0)
|
|
|
|
continue;
|
|
|
|
|
2015-02-25 03:29:25 +07:00
|
|
|
kvfree(slot->arch.lpage_info[i - 1]);
|
2012-08-01 16:03:28 +07:00
|
|
|
slot->arch.lpage_info[i - 1] = NULL;
|
2012-02-08 11:02:18 +07:00
|
|
|
}
|
|
|
|
return -ENOMEM;
|
|
|
|
}
|
|
|
|
|
2015-05-18 02:26:08 +07:00
|
|
|
void kvm_arch_memslots_updated(struct kvm *kvm, struct kvm_memslots *slots)
|
2013-07-04 11:40:29 +07:00
|
|
|
{
|
2013-07-04 11:41:26 +07:00
|
|
|
/*
|
|
|
|
* memslots->generation has been incremented.
|
|
|
|
* mmio generation may have reached its maximum value.
|
|
|
|
*/
|
2015-04-08 20:39:23 +07:00
|
|
|
kvm_mmu_invalidate_mmio_sptes(kvm, slots);
|
2013-07-04 11:40:29 +07:00
|
|
|
}
|
|
|
|
|
2009-12-23 23:35:18 +07:00
|
|
|
int kvm_arch_prepare_memory_region(struct kvm *kvm,
|
|
|
|
struct kvm_memory_slot *memslot,
|
2015-05-18 18:59:39 +07:00
|
|
|
const struct kvm_userspace_memory_region *mem,
|
2013-02-27 17:44:34 +07:00
|
|
|
enum kvm_mr_change change)
|
2007-11-20 15:25:04 +07:00
|
|
|
{
|
2009-12-23 23:35:18 +07:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2015-01-28 09:54:27 +07:00
|
|
|
static void kvm_mmu_slot_apply_flags(struct kvm *kvm,
|
|
|
|
struct kvm_memory_slot *new)
|
|
|
|
{
|
|
|
|
/* Still write protect RO slot */
|
|
|
|
if (new->flags & KVM_MEM_READONLY) {
|
|
|
|
kvm_mmu_slot_remove_write_access(kvm, new);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Call kvm_x86_ops dirty logging hooks when they are valid.
|
|
|
|
*
|
|
|
|
* kvm_x86_ops->slot_disable_log_dirty is called when:
|
|
|
|
*
|
|
|
|
* - KVM_MR_CREATE with dirty logging is disabled
|
|
|
|
* - KVM_MR_FLAGS_ONLY with dirty logging is disabled in new flag
|
|
|
|
*
|
|
|
|
* The reason is, in case of PML, we need to set D-bit for any slots
|
|
|
|
* with dirty logging disabled in order to eliminate unnecessary GPA
|
|
|
|
* logging in PML buffer (and potential PML buffer full VMEXT). This
|
|
|
|
* guarantees leaving PML enabled during guest's lifetime won't have
|
|
|
|
* any additonal overhead from PML when guest is running with dirty
|
|
|
|
* logging disabled for memory slots.
|
|
|
|
*
|
|
|
|
* kvm_x86_ops->slot_enable_log_dirty is called when switching new slot
|
|
|
|
* to dirty logging mode.
|
|
|
|
*
|
|
|
|
* If kvm_x86_ops dirty logging hooks are invalid, use write protect.
|
|
|
|
*
|
|
|
|
* In case of write protect:
|
|
|
|
*
|
|
|
|
* Write protect all pages for dirty logging.
|
|
|
|
*
|
|
|
|
* All the sptes including the large sptes which point to this
|
|
|
|
* slot are set to readonly. We can not create any new large
|
|
|
|
* spte on this slot until the end of the logging.
|
|
|
|
*
|
|
|
|
* See the comments in fast_page_fault().
|
|
|
|
*/
|
|
|
|
if (new->flags & KVM_MEM_LOG_DIRTY_PAGES) {
|
|
|
|
if (kvm_x86_ops->slot_enable_log_dirty)
|
|
|
|
kvm_x86_ops->slot_enable_log_dirty(kvm, new);
|
|
|
|
else
|
|
|
|
kvm_mmu_slot_remove_write_access(kvm, new);
|
|
|
|
} else {
|
|
|
|
if (kvm_x86_ops->slot_disable_log_dirty)
|
|
|
|
kvm_x86_ops->slot_disable_log_dirty(kvm, new);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2009-12-23 23:35:18 +07:00
|
|
|
void kvm_arch_commit_memory_region(struct kvm *kvm,
|
2015-05-18 18:59:39 +07:00
|
|
|
const struct kvm_userspace_memory_region *mem,
|
2013-02-27 17:45:25 +07:00
|
|
|
const struct kvm_memory_slot *old,
|
2015-05-18 18:20:23 +07:00
|
|
|
const struct kvm_memory_slot *new,
|
2013-02-27 17:45:25 +07:00
|
|
|
enum kvm_mr_change change)
|
2009-12-23 23:35:18 +07:00
|
|
|
{
|
2013-02-27 17:45:25 +07:00
|
|
|
int nr_mmu_pages = 0;
|
2009-12-23 23:35:18 +07:00
|
|
|
|
2011-03-04 17:59:21 +07:00
|
|
|
if (!kvm->arch.n_requested_mmu_pages)
|
|
|
|
nr_mmu_pages = kvm_mmu_calculate_mmu_pages(kvm);
|
|
|
|
|
|
|
|
if (nr_mmu_pages)
|
2007-11-20 15:25:04 +07:00
|
|
|
kvm_mmu_change_mmu_pages(kvm, nr_mmu_pages);
|
2015-01-28 09:54:26 +07:00
|
|
|
|
2015-04-03 14:40:25 +07:00
|
|
|
/*
|
|
|
|
* Dirty logging tracks sptes in 4k granularity, meaning that large
|
|
|
|
* sptes have to be split. If live migration is successful, the guest
|
|
|
|
* in the source machine will be destroyed and large sptes will be
|
|
|
|
* created in the destination. However, if the guest continues to run
|
|
|
|
* in the source machine (for example if live migration fails), small
|
|
|
|
* sptes will remain around and cause bad performance.
|
|
|
|
*
|
|
|
|
* Scan sptes if dirty logging has been stopped, dropping those
|
|
|
|
* which can be collapsed into a single large-page spte. Later
|
|
|
|
* page faults will create the large-page sptes.
|
|
|
|
*/
|
|
|
|
if ((change != KVM_MR_DELETE) &&
|
|
|
|
(old->flags & KVM_MEM_LOG_DIRTY_PAGES) &&
|
|
|
|
!(new->flags & KVM_MEM_LOG_DIRTY_PAGES))
|
|
|
|
kvm_mmu_zap_collapsible_sptes(kvm, new);
|
|
|
|
|
2013-01-08 17:43:28 +07:00
|
|
|
/*
|
2015-01-28 09:54:27 +07:00
|
|
|
* Set up write protection and/or dirty logging for the new slot.
|
2014-04-17 16:06:14 +07:00
|
|
|
*
|
2015-01-28 09:54:27 +07:00
|
|
|
* For KVM_MR_DELETE and KVM_MR_MOVE, the shadow pages of old slot have
|
|
|
|
* been zapped so no dirty logging staff is needed for old slot. For
|
|
|
|
* KVM_MR_FLAGS_ONLY, the old slot is essentially the same one as the
|
|
|
|
* new and it's also covered when dealing with the new slot.
|
2015-05-18 18:20:23 +07:00
|
|
|
*
|
|
|
|
* FIXME: const-ify all uses of struct kvm_memory_slot.
|
2013-01-08 17:43:28 +07:00
|
|
|
*/
|
2015-01-28 09:54:27 +07:00
|
|
|
if (change != KVM_MR_DELETE)
|
2015-05-18 18:20:23 +07:00
|
|
|
kvm_mmu_slot_apply_flags(kvm, (struct kvm_memory_slot *) new);
|
2007-11-20 15:25:04 +07:00
|
|
|
}
|
2007-12-14 08:35:10 +07:00
|
|
|
|
2012-08-25 01:54:57 +07:00
|
|
|
void kvm_arch_flush_shadow_all(struct kvm *kvm)
|
2008-07-11 06:49:31 +07:00
|
|
|
{
|
2013-05-31 07:36:23 +07:00
|
|
|
kvm_mmu_invalidate_zap_all_pages(kvm);
|
2008-07-11 06:49:31 +07:00
|
|
|
}
|
|
|
|
|
2012-08-25 01:54:57 +07:00
|
|
|
void kvm_arch_flush_shadow_memslot(struct kvm *kvm,
|
|
|
|
struct kvm_memory_slot *slot)
|
|
|
|
{
|
2013-05-31 07:36:23 +07:00
|
|
|
kvm_mmu_invalidate_zap_all_pages(kvm);
|
2012-08-25 01:54:57 +07:00
|
|
|
}
|
|
|
|
|
2015-10-13 15:18:53 +07:00
|
|
|
static inline bool kvm_vcpu_has_events(struct kvm_vcpu *vcpu)
|
|
|
|
{
|
|
|
|
if (!list_empty_careful(&vcpu->async_pf.done))
|
|
|
|
return true;
|
|
|
|
|
|
|
|
if (kvm_apic_has_events(vcpu))
|
|
|
|
return true;
|
|
|
|
|
|
|
|
if (vcpu->arch.pv.pv_unhalted)
|
|
|
|
return true;
|
|
|
|
|
|
|
|
if (atomic_read(&vcpu->arch.nmi_queued))
|
|
|
|
return true;
|
|
|
|
|
2015-10-13 15:19:35 +07:00
|
|
|
if (test_bit(KVM_REQ_SMI, &vcpu->requests))
|
|
|
|
return true;
|
|
|
|
|
2015-10-13 15:18:53 +07:00
|
|
|
if (kvm_arch_interrupt_allowed(vcpu) &&
|
|
|
|
kvm_cpu_has_interrupt(vcpu))
|
|
|
|
return true;
|
|
|
|
|
2015-11-30 23:22:21 +07:00
|
|
|
if (kvm_hv_has_stimer_pending(vcpu))
|
|
|
|
return true;
|
|
|
|
|
2015-10-13 15:18:53 +07:00
|
|
|
return false;
|
|
|
|
}
|
|
|
|
|
2007-12-14 08:35:10 +07:00
|
|
|
int kvm_arch_vcpu_runnable(struct kvm_vcpu *vcpu)
|
|
|
|
{
|
2014-03-08 02:03:12 +07:00
|
|
|
if (is_guest_mode(vcpu) && kvm_x86_ops->check_nested_events)
|
|
|
|
kvm_x86_ops->check_nested_events(vcpu, false);
|
|
|
|
|
2015-10-13 15:18:53 +07:00
|
|
|
return kvm_vcpu_running(vcpu) || kvm_vcpu_has_events(vcpu);
|
2007-12-14 08:35:10 +07:00
|
|
|
}
|
2007-12-17 13:21:40 +07:00
|
|
|
|
2012-03-09 04:44:24 +07:00
|
|
|
int kvm_arch_vcpu_should_kick(struct kvm_vcpu *vcpu)
|
2007-12-17 13:21:40 +07:00
|
|
|
{
|
2012-03-09 04:44:24 +07:00
|
|
|
return kvm_vcpu_exiting_guest_mode(vcpu) == IN_GUEST_MODE;
|
2007-12-17 13:21:40 +07:00
|
|
|
}
|
2009-03-23 17:12:11 +07:00
|
|
|
|
|
|
|
int kvm_arch_interrupt_allowed(struct kvm_vcpu *vcpu)
|
|
|
|
{
|
|
|
|
return kvm_x86_ops->interrupt_allowed(vcpu);
|
|
|
|
}
|
2009-06-17 19:22:14 +07:00
|
|
|
|
2014-11-02 16:54:45 +07:00
|
|
|
unsigned long kvm_get_linear_rip(struct kvm_vcpu *vcpu)
|
2010-02-23 23:47:55 +07:00
|
|
|
{
|
2014-11-02 16:54:45 +07:00
|
|
|
if (is_64_bit_mode(vcpu))
|
|
|
|
return kvm_rip_read(vcpu);
|
|
|
|
return (u32)(get_segment_base(vcpu, VCPU_SREG_CS) +
|
|
|
|
kvm_rip_read(vcpu));
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(kvm_get_linear_rip);
|
2010-02-23 23:47:55 +07:00
|
|
|
|
2014-11-02 16:54:45 +07:00
|
|
|
bool kvm_is_linear_rip(struct kvm_vcpu *vcpu, unsigned long linear_rip)
|
|
|
|
{
|
|
|
|
return kvm_get_linear_rip(vcpu) == linear_rip;
|
2010-02-23 23:47:55 +07:00
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(kvm_is_linear_rip);
|
|
|
|
|
2009-10-18 18:24:44 +07:00
|
|
|
unsigned long kvm_get_rflags(struct kvm_vcpu *vcpu)
|
|
|
|
{
|
|
|
|
unsigned long rflags;
|
|
|
|
|
|
|
|
rflags = kvm_x86_ops->get_rflags(vcpu);
|
|
|
|
if (vcpu->guest_debug & KVM_GUESTDBG_SINGLESTEP)
|
2010-02-23 23:47:58 +07:00
|
|
|
rflags &= ~X86_EFLAGS_TF;
|
2009-10-18 18:24:44 +07:00
|
|
|
return rflags;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(kvm_get_rflags);
|
|
|
|
|
2014-03-27 17:29:28 +07:00
|
|
|
static void __kvm_set_rflags(struct kvm_vcpu *vcpu, unsigned long rflags)
|
2009-10-18 18:24:44 +07:00
|
|
|
{
|
|
|
|
if (vcpu->guest_debug & KVM_GUESTDBG_SINGLESTEP &&
|
2010-02-23 23:47:55 +07:00
|
|
|
kvm_is_linear_rip(vcpu, vcpu->arch.singlestep_rip))
|
2010-02-23 23:47:58 +07:00
|
|
|
rflags |= X86_EFLAGS_TF;
|
2009-10-18 18:24:44 +07:00
|
|
|
kvm_x86_ops->set_rflags(vcpu, rflags);
|
2014-03-27 17:29:28 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
void kvm_set_rflags(struct kvm_vcpu *vcpu, unsigned long rflags)
|
|
|
|
{
|
|
|
|
__kvm_set_rflags(vcpu, rflags);
|
2010-07-27 16:30:24 +07:00
|
|
|
kvm_make_request(KVM_REQ_EVENT, vcpu);
|
2009-10-18 18:24:44 +07:00
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(kvm_set_rflags);
|
|
|
|
|
2010-10-17 23:13:42 +07:00
|
|
|
void kvm_arch_async_page_ready(struct kvm_vcpu *vcpu, struct kvm_async_pf *work)
|
|
|
|
{
|
|
|
|
int r;
|
|
|
|
|
2010-12-07 09:35:25 +07:00
|
|
|
if ((vcpu->arch.mmu.direct_map != work->arch.direct_map) ||
|
2013-10-14 21:22:33 +07:00
|
|
|
work->wakeup_all)
|
2010-10-17 23:13:42 +07:00
|
|
|
return;
|
|
|
|
|
|
|
|
r = kvm_mmu_reload(vcpu);
|
|
|
|
if (unlikely(r))
|
|
|
|
return;
|
|
|
|
|
2010-12-07 09:35:25 +07:00
|
|
|
if (!vcpu->arch.mmu.direct_map &&
|
|
|
|
work->arch.cr3 != vcpu->arch.mmu.get_cr3(vcpu))
|
|
|
|
return;
|
|
|
|
|
2010-10-17 23:13:42 +07:00
|
|
|
vcpu->arch.mmu.page_fault(vcpu, work->gva, 0, true);
|
|
|
|
}
|
|
|
|
|
2010-10-14 16:22:46 +07:00
|
|
|
static inline u32 kvm_async_pf_hash_fn(gfn_t gfn)
|
|
|
|
{
|
|
|
|
return hash_32(gfn & 0xffffffff, order_base_2(ASYNC_PF_PER_VCPU));
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline u32 kvm_async_pf_next_probe(u32 key)
|
|
|
|
{
|
|
|
|
return (key + 1) & (roundup_pow_of_two(ASYNC_PF_PER_VCPU) - 1);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void kvm_add_async_pf_gfn(struct kvm_vcpu *vcpu, gfn_t gfn)
|
|
|
|
{
|
|
|
|
u32 key = kvm_async_pf_hash_fn(gfn);
|
|
|
|
|
|
|
|
while (vcpu->arch.apf.gfns[key] != ~0)
|
|
|
|
key = kvm_async_pf_next_probe(key);
|
|
|
|
|
|
|
|
vcpu->arch.apf.gfns[key] = gfn;
|
|
|
|
}
|
|
|
|
|
|
|
|
static u32 kvm_async_pf_gfn_slot(struct kvm_vcpu *vcpu, gfn_t gfn)
|
|
|
|
{
|
|
|
|
int i;
|
|
|
|
u32 key = kvm_async_pf_hash_fn(gfn);
|
|
|
|
|
|
|
|
for (i = 0; i < roundup_pow_of_two(ASYNC_PF_PER_VCPU) &&
|
2010-11-01 16:00:30 +07:00
|
|
|
(vcpu->arch.apf.gfns[key] != gfn &&
|
|
|
|
vcpu->arch.apf.gfns[key] != ~0); i++)
|
2010-10-14 16:22:46 +07:00
|
|
|
key = kvm_async_pf_next_probe(key);
|
|
|
|
|
|
|
|
return key;
|
|
|
|
}
|
|
|
|
|
|
|
|
bool kvm_find_async_pf_gfn(struct kvm_vcpu *vcpu, gfn_t gfn)
|
|
|
|
{
|
|
|
|
return vcpu->arch.apf.gfns[kvm_async_pf_gfn_slot(vcpu, gfn)] == gfn;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void kvm_del_async_pf_gfn(struct kvm_vcpu *vcpu, gfn_t gfn)
|
|
|
|
{
|
|
|
|
u32 i, j, k;
|
|
|
|
|
|
|
|
i = j = kvm_async_pf_gfn_slot(vcpu, gfn);
|
|
|
|
while (true) {
|
|
|
|
vcpu->arch.apf.gfns[i] = ~0;
|
|
|
|
do {
|
|
|
|
j = kvm_async_pf_next_probe(j);
|
|
|
|
if (vcpu->arch.apf.gfns[j] == ~0)
|
|
|
|
return;
|
|
|
|
k = kvm_async_pf_hash_fn(vcpu->arch.apf.gfns[j]);
|
|
|
|
/*
|
|
|
|
* k lies cyclically in ]i,j]
|
|
|
|
* | i.k.j |
|
|
|
|
* |....j i.k.| or |.k..j i...|
|
|
|
|
*/
|
|
|
|
} while ((i <= j) ? (i < k && k <= j) : (i < k || k <= j));
|
|
|
|
vcpu->arch.apf.gfns[i] = vcpu->arch.apf.gfns[j];
|
|
|
|
i = j;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2010-10-14 16:22:53 +07:00
|
|
|
static int apf_put_user(struct kvm_vcpu *vcpu, u32 val)
|
|
|
|
{
|
|
|
|
|
|
|
|
return kvm_write_guest_cached(vcpu->kvm, &vcpu->arch.apf.data, &val,
|
|
|
|
sizeof(val));
|
|
|
|
}
|
|
|
|
|
2010-10-14 16:22:46 +07:00
|
|
|
void kvm_arch_async_page_not_present(struct kvm_vcpu *vcpu,
|
|
|
|
struct kvm_async_pf *work)
|
|
|
|
{
|
2010-11-29 21:12:30 +07:00
|
|
|
struct x86_exception fault;
|
|
|
|
|
2010-10-14 16:22:53 +07:00
|
|
|
trace_kvm_async_pf_not_present(work->arch.token, work->gva);
|
2010-10-14 16:22:46 +07:00
|
|
|
kvm_add_async_pf_gfn(vcpu, work->arch.gfn);
|
2010-10-14 16:22:53 +07:00
|
|
|
|
|
|
|
if (!(vcpu->arch.apf.msr_val & KVM_ASYNC_PF_ENABLED) ||
|
2010-10-14 16:22:56 +07:00
|
|
|
(vcpu->arch.apf.send_user_only &&
|
|
|
|
kvm_x86_ops->get_cpl(vcpu) == 0))
|
2010-10-14 16:22:53 +07:00
|
|
|
kvm_make_request(KVM_REQ_APF_HALT, vcpu);
|
|
|
|
else if (!apf_put_user(vcpu, KVM_PV_REASON_PAGE_NOT_PRESENT)) {
|
2010-11-29 21:12:30 +07:00
|
|
|
fault.vector = PF_VECTOR;
|
|
|
|
fault.error_code_valid = true;
|
|
|
|
fault.error_code = 0;
|
|
|
|
fault.nested_page_fault = false;
|
|
|
|
fault.address = work->arch.token;
|
|
|
|
kvm_inject_page_fault(vcpu, &fault);
|
2010-10-14 16:22:53 +07:00
|
|
|
}
|
2010-10-14 16:22:46 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
void kvm_arch_async_page_present(struct kvm_vcpu *vcpu,
|
|
|
|
struct kvm_async_pf *work)
|
|
|
|
{
|
2010-11-29 21:12:30 +07:00
|
|
|
struct x86_exception fault;
|
|
|
|
|
2010-10-14 16:22:53 +07:00
|
|
|
trace_kvm_async_pf_ready(work->arch.token, work->gva);
|
2013-10-14 21:22:33 +07:00
|
|
|
if (work->wakeup_all)
|
2010-10-14 16:22:53 +07:00
|
|
|
work->arch.token = ~0; /* broadcast wakeup */
|
|
|
|
else
|
|
|
|
kvm_del_async_pf_gfn(vcpu, work->arch.gfn);
|
|
|
|
|
|
|
|
if ((vcpu->arch.apf.msr_val & KVM_ASYNC_PF_ENABLED) &&
|
|
|
|
!apf_put_user(vcpu, KVM_PV_REASON_PAGE_READY)) {
|
2010-11-29 21:12:30 +07:00
|
|
|
fault.vector = PF_VECTOR;
|
|
|
|
fault.error_code_valid = true;
|
|
|
|
fault.error_code = 0;
|
|
|
|
fault.nested_page_fault = false;
|
|
|
|
fault.address = work->arch.token;
|
|
|
|
kvm_inject_page_fault(vcpu, &fault);
|
2010-10-14 16:22:53 +07:00
|
|
|
}
|
2010-11-01 16:01:28 +07:00
|
|
|
vcpu->arch.apf.halted = false;
|
2012-05-03 15:36:39 +07:00
|
|
|
vcpu->arch.mp_state = KVM_MP_STATE_RUNNABLE;
|
2010-10-14 16:22:53 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
bool kvm_arch_can_inject_async_page_present(struct kvm_vcpu *vcpu)
|
|
|
|
{
|
|
|
|
if (!(vcpu->arch.apf.msr_val & KVM_ASYNC_PF_ENABLED))
|
|
|
|
return true;
|
|
|
|
else
|
|
|
|
return !kvm_event_needs_reinjection(vcpu) &&
|
|
|
|
kvm_x86_ops->interrupt_allowed(vcpu);
|
2010-10-14 16:22:46 +07:00
|
|
|
}
|
|
|
|
|
2015-07-07 20:41:58 +07:00
|
|
|
void kvm_arch_start_assignment(struct kvm *kvm)
|
|
|
|
{
|
|
|
|
atomic_inc(&kvm->arch.assigned_device_count);
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(kvm_arch_start_assignment);
|
|
|
|
|
|
|
|
void kvm_arch_end_assignment(struct kvm *kvm)
|
|
|
|
{
|
|
|
|
atomic_dec(&kvm->arch.assigned_device_count);
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(kvm_arch_end_assignment);
|
|
|
|
|
|
|
|
bool kvm_arch_has_assigned_device(struct kvm *kvm)
|
|
|
|
{
|
|
|
|
return atomic_read(&kvm->arch.assigned_device_count);
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(kvm_arch_has_assigned_device);
|
|
|
|
|
2013-10-31 00:02:30 +07:00
|
|
|
void kvm_arch_register_noncoherent_dma(struct kvm *kvm)
|
|
|
|
{
|
|
|
|
atomic_inc(&kvm->arch.noncoherent_dma_count);
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(kvm_arch_register_noncoherent_dma);
|
|
|
|
|
|
|
|
void kvm_arch_unregister_noncoherent_dma(struct kvm *kvm)
|
|
|
|
{
|
|
|
|
atomic_dec(&kvm->arch.noncoherent_dma_count);
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(kvm_arch_unregister_noncoherent_dma);
|
|
|
|
|
|
|
|
bool kvm_arch_has_noncoherent_dma(struct kvm *kvm)
|
|
|
|
{
|
|
|
|
return atomic_read(&kvm->arch.noncoherent_dma_count);
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(kvm_arch_has_noncoherent_dma);
|
|
|
|
|
2016-05-06 00:58:35 +07:00
|
|
|
bool kvm_arch_has_irq_bypass(void)
|
|
|
|
{
|
|
|
|
return kvm_x86_ops->update_pi_irte != NULL;
|
|
|
|
}
|
|
|
|
|
2015-09-18 21:29:40 +07:00
|
|
|
int kvm_arch_irq_bypass_add_producer(struct irq_bypass_consumer *cons,
|
|
|
|
struct irq_bypass_producer *prod)
|
|
|
|
{
|
|
|
|
struct kvm_kernel_irqfd *irqfd =
|
|
|
|
container_of(cons, struct kvm_kernel_irqfd, consumer);
|
|
|
|
|
2016-05-06 00:58:35 +07:00
|
|
|
irqfd->producer = prod;
|
2015-09-18 21:29:40 +07:00
|
|
|
|
2016-05-06 00:58:35 +07:00
|
|
|
return kvm_x86_ops->update_pi_irte(irqfd->kvm,
|
|
|
|
prod->irq, irqfd->gsi, 1);
|
2015-09-18 21:29:40 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
void kvm_arch_irq_bypass_del_producer(struct irq_bypass_consumer *cons,
|
|
|
|
struct irq_bypass_producer *prod)
|
|
|
|
{
|
|
|
|
int ret;
|
|
|
|
struct kvm_kernel_irqfd *irqfd =
|
|
|
|
container_of(cons, struct kvm_kernel_irqfd, consumer);
|
|
|
|
|
|
|
|
WARN_ON(irqfd->producer != prod);
|
|
|
|
irqfd->producer = NULL;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* When producer of consumer is unregistered, we change back to
|
|
|
|
* remapped mode, so we can re-use the current implementation
|
2016-05-21 19:14:44 +07:00
|
|
|
* when the irq is masked/disabled or the consumer side (KVM
|
2015-09-18 21:29:40 +07:00
|
|
|
* int this case doesn't want to receive the interrupts.
|
|
|
|
*/
|
|
|
|
ret = kvm_x86_ops->update_pi_irte(irqfd->kvm, prod->irq, irqfd->gsi, 0);
|
|
|
|
if (ret)
|
|
|
|
printk(KERN_INFO "irq bypass consumer (token %p) unregistration"
|
|
|
|
" fails: %d\n", irqfd->consumer.token, ret);
|
|
|
|
}
|
|
|
|
|
|
|
|
int kvm_arch_update_irqfd_routing(struct kvm *kvm, unsigned int host_irq,
|
|
|
|
uint32_t guest_irq, bool set)
|
|
|
|
{
|
|
|
|
if (!kvm_x86_ops->update_pi_irte)
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
return kvm_x86_ops->update_pi_irte(kvm, host_irq, guest_irq, set);
|
|
|
|
}
|
|
|
|
|
2016-01-25 15:53:33 +07:00
|
|
|
bool kvm_vector_hashing_enabled(void)
|
|
|
|
{
|
|
|
|
return vector_hashing;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(kvm_vector_hashing_enabled);
|
|
|
|
|
2009-06-17 19:22:14 +07:00
|
|
|
EXPORT_TRACEPOINT_SYMBOL_GPL(kvm_exit);
|
2015-09-15 13:41:58 +07:00
|
|
|
EXPORT_TRACEPOINT_SYMBOL_GPL(kvm_fast_mmio);
|
2009-06-17 19:22:14 +07:00
|
|
|
EXPORT_TRACEPOINT_SYMBOL_GPL(kvm_inj_virq);
|
|
|
|
EXPORT_TRACEPOINT_SYMBOL_GPL(kvm_page_fault);
|
|
|
|
EXPORT_TRACEPOINT_SYMBOL_GPL(kvm_msr);
|
|
|
|
EXPORT_TRACEPOINT_SYMBOL_GPL(kvm_cr);
|
2009-10-09 21:08:27 +07:00
|
|
|
EXPORT_TRACEPOINT_SYMBOL_GPL(kvm_nested_vmrun);
|
2009-10-09 21:08:28 +07:00
|
|
|
EXPORT_TRACEPOINT_SYMBOL_GPL(kvm_nested_vmexit);
|
2009-10-09 21:08:29 +07:00
|
|
|
EXPORT_TRACEPOINT_SYMBOL_GPL(kvm_nested_vmexit_inject);
|
2009-10-09 21:08:30 +07:00
|
|
|
EXPORT_TRACEPOINT_SYMBOL_GPL(kvm_nested_intr_vmexit);
|
2009-10-09 21:08:31 +07:00
|
|
|
EXPORT_TRACEPOINT_SYMBOL_GPL(kvm_invlpga);
|
2009-10-09 21:08:32 +07:00
|
|
|
EXPORT_TRACEPOINT_SYMBOL_GPL(kvm_skinit);
|
2010-02-25 00:59:14 +07:00
|
|
|
EXPORT_TRACEPOINT_SYMBOL_GPL(kvm_nested_intercepts);
|
2013-06-12 14:43:44 +07:00
|
|
|
EXPORT_TRACEPOINT_SYMBOL_GPL(kvm_write_tsc_offset);
|
2014-08-21 23:08:09 +07:00
|
|
|
EXPORT_TRACEPOINT_SYMBOL_GPL(kvm_ple_window);
|
2015-01-28 09:54:28 +07:00
|
|
|
EXPORT_TRACEPOINT_SYMBOL_GPL(kvm_pml_full);
|
2015-09-18 21:29:51 +07:00
|
|
|
EXPORT_TRACEPOINT_SYMBOL_GPL(kvm_pi_irte_update);
|
2016-05-05 02:09:48 +07:00
|
|
|
EXPORT_TRACEPOINT_SYMBOL_GPL(kvm_avic_unaccelerated_access);
|
|
|
|
EXPORT_TRACEPOINT_SYMBOL_GPL(kvm_avic_incomplete_ipi);
|