* ARM/ARM64 locking fixes

* x86 fixes: PCID, UMIP, locking
 * Improved support for recent Windows version that have a 2048 Hz
 APIC timer.
 * Rename KVM_HINTS_DEDICATED CPUID bit to KVM_HINTS_REALTIME
 * Better behaved selftests.
 -----BEGIN PGP SIGNATURE-----
 Version: GnuPG v2.0.22 (GNU/Linux)
 
 iQEcBAABAgAGBQJa/bkTAAoJEL/70l94x66Dzf8IAJ1GqtXi0CNbq8MvU4QIqw0L
 HLIRoe/QgkTeTUa2fwirEuu5I+/wUyPvy5sAIsn/F5eiZM7nciLm+fYzw6F2uPIm
 lSCqKpVwmh8dPl1SBaqPnTcB1HPVwcCgc2SF9Ph7yZCUwFUtoeUuPj8v6Qy6y21g
 jfobHFZa3MrFgi7kPxOXSrC1qxuNJL9yLB5mwCvCK/K7jj2nrGJkLLDuzgReCqvz
 isOdpof3hz8whXDQG5cTtybBgE9veym4YqJY8R5ANXBKqbFlhaNF1T3xXrdPMISZ
 7bsGgkhYEOqeQsPrFwzAIiFxe2DogFwkn1BcvJ1B+duXrayt5CBnDPRB6Yxg00M=
 =H0d0
 -----END PGP SIGNATURE-----

Merge tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm

Pull kvm fixes from Paolo Bonzini:

 - ARM/ARM64 locking fixes

 - x86 fixes: PCID, UMIP, locking

 - improved support for recent Windows version that have a 2048 Hz APIC
   timer

 - rename KVM_HINTS_DEDICATED CPUID bit to KVM_HINTS_REALTIME

 - better behaved selftests

* tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm:
  kvm: rename KVM_HINTS_DEDICATED to KVM_HINTS_REALTIME
  KVM: arm/arm64: VGIC/ITS save/restore: protect kvm_read_guest() calls
  KVM: arm/arm64: VGIC/ITS: protect kvm_read_guest() calls with SRCU lock
  KVM: arm/arm64: VGIC/ITS: Promote irq_lock() in update_affinity
  KVM: arm/arm64: Properly protect VGIC locks from IRQs
  KVM: X86: Lower the default timer frequency limit to 200us
  KVM: vmx: update sec exec controls for UMIP iff emulating UMIP
  kvm: x86: Suppress CR3_PCID_INVD bit only when PCIDs are enabled
  KVM: selftests: exit with 0 status code when tests cannot be run
  KVM: hyperv: idr_find needs RCU protection
  x86: Delay skip of emulated hypercall instruction
  KVM: Extend MAX_IRQ_ROUTES to 4096 for all archs
This commit is contained in:
Linus Torvalds 2018-05-17 10:23:36 -07:00
commit 58ddfe6c3a
18 changed files with 157 additions and 87 deletions

View File

@ -72,8 +72,8 @@ KVM_FEATURE_CLOCKSOURCE_STABLE_BIT || 24 || host will warn if no guest-side
flag || value || meaning
==================================================================================
KVM_HINTS_DEDICATED || 0 || guest checks this feature bit to
|| || determine if there is vCPU pinning
|| || and there is no vCPU over-commitment,
KVM_HINTS_REALTIME || 0 || guest checks this feature bit to
|| || determine that vCPUs are never
|| || preempted for an unlimited time,
|| || allowing optimizations
----------------------------------------------------------------------------------

View File

@ -309,6 +309,22 @@ static inline unsigned int kvm_get_vmid_bits(void)
return 8;
}
/*
* We are not in the kvm->srcu critical section most of the time, so we take
* the SRCU read lock here. Since we copy the data from the user page, we
* can immediately drop the lock again.
*/
static inline int kvm_read_guest_lock(struct kvm *kvm,
gpa_t gpa, void *data, unsigned long len)
{
int srcu_idx = srcu_read_lock(&kvm->srcu);
int ret = kvm_read_guest(kvm, gpa, data, len);
srcu_read_unlock(&kvm->srcu, srcu_idx);
return ret;
}
static inline void *kvm_get_hyp_vector(void)
{
return kvm_ksym_ref(__kvm_hyp_vector);

View File

@ -360,6 +360,22 @@ static inline unsigned int kvm_get_vmid_bits(void)
return (cpuid_feature_extract_unsigned_field(reg, ID_AA64MMFR1_VMIDBITS_SHIFT) == 2) ? 16 : 8;
}
/*
* We are not in the kvm->srcu critical section most of the time, so we take
* the SRCU read lock here. Since we copy the data from the user page, we
* can immediately drop the lock again.
*/
static inline int kvm_read_guest_lock(struct kvm *kvm,
gpa_t gpa, void *data, unsigned long len)
{
int srcu_idx = srcu_read_lock(&kvm->srcu);
int ret = kvm_read_guest(kvm, gpa, data, len);
srcu_read_unlock(&kvm->srcu, srcu_idx);
return ret;
}
#ifdef CONFIG_KVM_INDIRECT_VECTORS
/*
* EL2 vectors can be mapped and rerouted in a number of ways,

View File

@ -29,7 +29,7 @@
#define KVM_FEATURE_PV_TLB_FLUSH 9
#define KVM_FEATURE_ASYNC_PF_VMEXIT 10
#define KVM_HINTS_DEDICATED 0
#define KVM_HINTS_REALTIME 0
/* The last 8 bits are used to indicate how to interpret the flags field
* in pvclock structure. If no bits are set, all flags are ignored.

View File

@ -457,7 +457,7 @@ static void __init sev_map_percpu_data(void)
static void __init kvm_smp_prepare_cpus(unsigned int max_cpus)
{
native_smp_prepare_cpus(max_cpus);
if (kvm_para_has_hint(KVM_HINTS_DEDICATED))
if (kvm_para_has_hint(KVM_HINTS_REALTIME))
static_branch_disable(&virt_spin_lock_key);
}
@ -553,7 +553,7 @@ static void __init kvm_guest_init(void)
}
if (kvm_para_has_feature(KVM_FEATURE_PV_TLB_FLUSH) &&
!kvm_para_has_hint(KVM_HINTS_DEDICATED) &&
!kvm_para_has_hint(KVM_HINTS_REALTIME) &&
kvm_para_has_feature(KVM_FEATURE_STEAL_TIME))
pv_mmu_ops.flush_tlb_others = kvm_flush_tlb_others;
@ -649,7 +649,7 @@ static __init int kvm_setup_pv_tlb_flush(void)
int cpu;
if (kvm_para_has_feature(KVM_FEATURE_PV_TLB_FLUSH) &&
!kvm_para_has_hint(KVM_HINTS_DEDICATED) &&
!kvm_para_has_hint(KVM_HINTS_REALTIME) &&
kvm_para_has_feature(KVM_FEATURE_STEAL_TIME)) {
for_each_possible_cpu(cpu) {
zalloc_cpumask_var_node(per_cpu_ptr(&__pv_tlb_mask, cpu),
@ -745,7 +745,7 @@ void __init kvm_spinlock_init(void)
if (!kvm_para_has_feature(KVM_FEATURE_PV_UNHALT))
return;
if (kvm_para_has_hint(KVM_HINTS_DEDICATED))
if (kvm_para_has_hint(KVM_HINTS_REALTIME))
return;
__pv_init_lock_hash();

View File

@ -1265,7 +1265,7 @@ static int kvm_hv_hypercall_complete_userspace(struct kvm_vcpu *vcpu)
struct kvm_run *run = vcpu->run;
kvm_hv_hypercall_set_result(vcpu, run->hyperv.u.hcall.result);
return 1;
return kvm_skip_emulated_instruction(vcpu);
}
static u16 kvm_hvcall_signal_event(struct kvm_vcpu *vcpu, bool fast, u64 param)
@ -1296,8 +1296,10 @@ static u16 kvm_hvcall_signal_event(struct kvm_vcpu *vcpu, bool fast, u64 param)
if (param & ~KVM_HYPERV_CONN_ID_MASK)
return HV_STATUS_INVALID_HYPERCALL_INPUT;
/* conn_to_evt is protected by vcpu->kvm->srcu */
/* the eventfd is protected by vcpu->kvm->srcu, but conn_to_evt isn't */
rcu_read_lock();
eventfd = idr_find(&vcpu->kvm->arch.hyperv.conn_to_evt, param);
rcu_read_unlock();
if (!eventfd)
return HV_STATUS_INVALID_PORT_ID;

View File

@ -1494,6 +1494,12 @@ static inline bool cpu_has_vmx_vmfunc(void)
SECONDARY_EXEC_ENABLE_VMFUNC;
}
static bool vmx_umip_emulated(void)
{
return vmcs_config.cpu_based_2nd_exec_ctrl &
SECONDARY_EXEC_DESC;
}
static inline bool report_flexpriority(void)
{
return flexpriority_enabled;
@ -4761,14 +4767,16 @@ static int vmx_set_cr4(struct kvm_vcpu *vcpu, unsigned long cr4)
else
hw_cr4 |= KVM_PMODE_VM_CR4_ALWAYS_ON;
if ((cr4 & X86_CR4_UMIP) && !boot_cpu_has(X86_FEATURE_UMIP)) {
vmcs_set_bits(SECONDARY_VM_EXEC_CONTROL,
SECONDARY_EXEC_DESC);
hw_cr4 &= ~X86_CR4_UMIP;
} else if (!is_guest_mode(vcpu) ||
!nested_cpu_has2(get_vmcs12(vcpu), SECONDARY_EXEC_DESC))
vmcs_clear_bits(SECONDARY_VM_EXEC_CONTROL,
if (!boot_cpu_has(X86_FEATURE_UMIP) && vmx_umip_emulated()) {
if (cr4 & X86_CR4_UMIP) {
vmcs_set_bits(SECONDARY_VM_EXEC_CONTROL,
SECONDARY_EXEC_DESC);
hw_cr4 &= ~X86_CR4_UMIP;
} else if (!is_guest_mode(vcpu) ||
!nested_cpu_has2(get_vmcs12(vcpu), SECONDARY_EXEC_DESC))
vmcs_clear_bits(SECONDARY_VM_EXEC_CONTROL,
SECONDARY_EXEC_DESC);
}
if (cr4 & X86_CR4_VMXE) {
/*
@ -9497,12 +9505,6 @@ static bool vmx_xsaves_supported(void)
SECONDARY_EXEC_XSAVES;
}
static bool vmx_umip_emulated(void)
{
return vmcs_config.cpu_based_2nd_exec_ctrl &
SECONDARY_EXEC_DESC;
}
static void vmx_recover_nmi_blocking(struct vcpu_vmx *vmx)
{
u32 exit_intr_info;

View File

@ -114,7 +114,7 @@ module_param(ignore_msrs, bool, S_IRUGO | S_IWUSR);
static bool __read_mostly report_ignored_msrs = true;
module_param(report_ignored_msrs, bool, S_IRUGO | S_IWUSR);
unsigned int min_timer_period_us = 500;
unsigned int min_timer_period_us = 200;
module_param(min_timer_period_us, uint, S_IRUGO | S_IWUSR);
static bool __read_mostly kvmclock_periodic_sync = true;
@ -843,7 +843,10 @@ EXPORT_SYMBOL_GPL(kvm_set_cr4);
int kvm_set_cr3(struct kvm_vcpu *vcpu, unsigned long cr3)
{
#ifdef CONFIG_X86_64
cr3 &= ~CR3_PCID_INVD;
bool pcid_enabled = kvm_read_cr4_bits(vcpu, X86_CR4_PCIDE);
if (pcid_enabled)
cr3 &= ~CR3_PCID_INVD;
#endif
if (cr3 == kvm_read_cr3(vcpu) && !pdptrs_changed(vcpu)) {
@ -6671,12 +6674,13 @@ void kvm_vcpu_deactivate_apicv(struct kvm_vcpu *vcpu)
int kvm_emulate_hypercall(struct kvm_vcpu *vcpu)
{
unsigned long nr, a0, a1, a2, a3, ret;
int op_64_bit, r;
int op_64_bit;
r = kvm_skip_emulated_instruction(vcpu);
if (kvm_hv_hypercall_enabled(vcpu->kvm))
return kvm_hv_hypercall(vcpu);
if (kvm_hv_hypercall_enabled(vcpu->kvm)) {
if (!kvm_hv_hypercall(vcpu))
return 0;
goto out;
}
nr = kvm_register_read(vcpu, VCPU_REGS_RAX);
a0 = kvm_register_read(vcpu, VCPU_REGS_RBX);
@ -6697,7 +6701,7 @@ int kvm_emulate_hypercall(struct kvm_vcpu *vcpu)
if (kvm_x86_ops->get_cpl(vcpu) != 0) {
ret = -KVM_EPERM;
goto out;
goto out_error;
}
switch (nr) {
@ -6717,12 +6721,14 @@ int kvm_emulate_hypercall(struct kvm_vcpu *vcpu)
ret = -KVM_ENOSYS;
break;
}
out:
out_error:
if (!op_64_bit)
ret = (u32)ret;
kvm_register_write(vcpu, VCPU_REGS_RAX, ret);
out:
++vcpu->stat.hypercalls;
return r;
return kvm_skip_emulated_instruction(vcpu);
}
EXPORT_SYMBOL_GPL(kvm_emulate_hypercall);

View File

@ -1045,13 +1045,7 @@ static inline int mmu_notifier_retry(struct kvm *kvm, unsigned long mmu_seq)
#ifdef CONFIG_HAVE_KVM_IRQ_ROUTING
#ifdef CONFIG_S390
#define KVM_MAX_IRQ_ROUTES 4096 //FIXME: we can have more than that...
#elif defined(CONFIG_ARM64)
#define KVM_MAX_IRQ_ROUTES 4096
#else
#define KVM_MAX_IRQ_ROUTES 1024
#endif
#define KVM_MAX_IRQ_ROUTES 4096 /* might need extension/rework in the future */
bool kvm_arch_can_set_irq_routing(struct kvm *kvm);
int kvm_set_irq_routing(struct kvm *kvm,

View File

@ -15,7 +15,7 @@ LIBKVM += $(LIBKVM_$(UNAME_M))
INSTALL_HDR_PATH = $(top_srcdir)/usr
LINUX_HDR_PATH = $(INSTALL_HDR_PATH)/include/
CFLAGS += -O2 -g -std=gnu99 -I$(LINUX_HDR_PATH) -Iinclude -I$(<D)
CFLAGS += -O2 -g -std=gnu99 -I$(LINUX_HDR_PATH) -Iinclude -I$(<D) -I..
# After inclusion, $(OUTPUT) is defined and
# $(TEST_GEN_PROGS) starts with $(OUTPUT)/

View File

@ -19,6 +19,7 @@
#include <errno.h>
#include <unistd.h>
#include <fcntl.h>
#include "kselftest.h"
ssize_t test_write(int fd, const void *buf, size_t count);
ssize_t test_read(int fd, void *buf, size_t count);

View File

@ -50,8 +50,8 @@ int kvm_check_cap(long cap)
int kvm_fd;
kvm_fd = open(KVM_DEV_PATH, O_RDONLY);
TEST_ASSERT(kvm_fd >= 0, "open %s failed, rc: %i errno: %i",
KVM_DEV_PATH, kvm_fd, errno);
if (kvm_fd < 0)
exit(KSFT_SKIP);
ret = ioctl(kvm_fd, KVM_CHECK_EXTENSION, cap);
TEST_ASSERT(ret != -1, "KVM_CHECK_EXTENSION IOCTL failed,\n"
@ -91,8 +91,8 @@ struct kvm_vm *vm_create(enum vm_guest_mode mode, uint64_t phy_pages, int perm)
vm->mode = mode;
kvm_fd = open(KVM_DEV_PATH, perm);
TEST_ASSERT(kvm_fd >= 0, "open %s failed, rc: %i errno: %i",
KVM_DEV_PATH, kvm_fd, errno);
if (kvm_fd < 0)
exit(KSFT_SKIP);
/* Create VM. */
vm->fd = ioctl(kvm_fd, KVM_CREATE_VM, NULL);
@ -418,8 +418,8 @@ struct kvm_cpuid2 *kvm_get_supported_cpuid(void)
cpuid = allocate_kvm_cpuid2();
kvm_fd = open(KVM_DEV_PATH, O_RDONLY);
TEST_ASSERT(kvm_fd >= 0, "open %s failed, rc: %i errno: %i",
KVM_DEV_PATH, kvm_fd, errno);
if (kvm_fd < 0)
exit(KSFT_SKIP);
ret = ioctl(kvm_fd, KVM_GET_SUPPORTED_CPUID, cpuid);
TEST_ASSERT(ret == 0, "KVM_GET_SUPPORTED_CPUID failed %d %d\n",
@ -675,8 +675,8 @@ static int vcpu_mmap_sz(void)
int dev_fd, ret;
dev_fd = open(KVM_DEV_PATH, O_RDONLY);
TEST_ASSERT(dev_fd >= 0, "%s open %s failed, rc: %i errno: %i",
__func__, KVM_DEV_PATH, dev_fd, errno);
if (dev_fd < 0)
exit(KSFT_SKIP);
ret = ioctl(dev_fd, KVM_GET_VCPU_MMAP_SIZE, NULL);
TEST_ASSERT(ret >= sizeof(struct kvm_run),

View File

@ -85,6 +85,9 @@ static void compare_vcpu_events(struct kvm_vcpu_events *left,
{
}
#define TEST_SYNC_FIELDS (KVM_SYNC_X86_REGS|KVM_SYNC_X86_SREGS|KVM_SYNC_X86_EVENTS)
#define INVALID_SYNC_FIELD 0x80000000
int main(int argc, char *argv[])
{
struct kvm_vm *vm;
@ -98,9 +101,14 @@ int main(int argc, char *argv[])
setbuf(stdout, NULL);
cap = kvm_check_cap(KVM_CAP_SYNC_REGS);
TEST_ASSERT((unsigned long)cap == KVM_SYNC_X86_VALID_FIELDS,
"KVM_CAP_SYNC_REGS (0x%x) != KVM_SYNC_X86_VALID_FIELDS (0x%lx)\n",
cap, KVM_SYNC_X86_VALID_FIELDS);
if ((cap & TEST_SYNC_FIELDS) != TEST_SYNC_FIELDS) {
fprintf(stderr, "KVM_CAP_SYNC_REGS not supported, skipping test\n");
exit(KSFT_SKIP);
}
if ((cap & INVALID_SYNC_FIELD) != 0) {
fprintf(stderr, "The \"invalid\" field is not invalid, skipping test\n");
exit(KSFT_SKIP);
}
/* Create VM */
vm = vm_create_default(VCPU_ID, guest_code);
@ -108,7 +116,14 @@ int main(int argc, char *argv[])
run = vcpu_state(vm, VCPU_ID);
/* Request reading invalid register set from VCPU. */
run->kvm_valid_regs = KVM_SYNC_X86_VALID_FIELDS << 1;
run->kvm_valid_regs = INVALID_SYNC_FIELD;
rv = _vcpu_run(vm, VCPU_ID);
TEST_ASSERT(rv < 0 && errno == EINVAL,
"Invalid kvm_valid_regs did not cause expected KVM_RUN error: %d\n",
rv);
vcpu_state(vm, VCPU_ID)->kvm_valid_regs = 0;
run->kvm_valid_regs = INVALID_SYNC_FIELD | TEST_SYNC_FIELDS;
rv = _vcpu_run(vm, VCPU_ID);
TEST_ASSERT(rv < 0 && errno == EINVAL,
"Invalid kvm_valid_regs did not cause expected KVM_RUN error: %d\n",
@ -116,7 +131,14 @@ int main(int argc, char *argv[])
vcpu_state(vm, VCPU_ID)->kvm_valid_regs = 0;
/* Request setting invalid register set into VCPU. */
run->kvm_dirty_regs = KVM_SYNC_X86_VALID_FIELDS << 1;
run->kvm_dirty_regs = INVALID_SYNC_FIELD;
rv = _vcpu_run(vm, VCPU_ID);
TEST_ASSERT(rv < 0 && errno == EINVAL,
"Invalid kvm_dirty_regs did not cause expected KVM_RUN error: %d\n",
rv);
vcpu_state(vm, VCPU_ID)->kvm_dirty_regs = 0;
run->kvm_dirty_regs = INVALID_SYNC_FIELD | TEST_SYNC_FIELDS;
rv = _vcpu_run(vm, VCPU_ID);
TEST_ASSERT(rv < 0 && errno == EINVAL,
"Invalid kvm_dirty_regs did not cause expected KVM_RUN error: %d\n",
@ -125,7 +147,7 @@ int main(int argc, char *argv[])
/* Request and verify all valid register sets. */
/* TODO: BUILD TIME CHECK: TEST_ASSERT(KVM_SYNC_X86_NUM_FIELDS != 3); */
run->kvm_valid_regs = KVM_SYNC_X86_VALID_FIELDS;
run->kvm_valid_regs = TEST_SYNC_FIELDS;
rv = _vcpu_run(vm, VCPU_ID);
TEST_ASSERT(run->exit_reason == KVM_EXIT_IO,
"Unexpected exit reason: %u (%s),\n",
@ -146,7 +168,7 @@ int main(int argc, char *argv[])
run->s.regs.sregs.apic_base = 1 << 11;
/* TODO run->s.regs.events.XYZ = ABC; */
run->kvm_valid_regs = KVM_SYNC_X86_VALID_FIELDS;
run->kvm_valid_regs = TEST_SYNC_FIELDS;
run->kvm_dirty_regs = KVM_SYNC_X86_REGS | KVM_SYNC_X86_SREGS;
rv = _vcpu_run(vm, VCPU_ID);
TEST_ASSERT(run->exit_reason == KVM_EXIT_IO,
@ -172,7 +194,7 @@ int main(int argc, char *argv[])
/* Clear kvm_dirty_regs bits, verify new s.regs values are
* overwritten with existing guest values.
*/
run->kvm_valid_regs = KVM_SYNC_X86_VALID_FIELDS;
run->kvm_valid_regs = TEST_SYNC_FIELDS;
run->kvm_dirty_regs = 0;
run->s.regs.regs.r11 = 0xDEADBEEF;
rv = _vcpu_run(vm, VCPU_ID);
@ -211,7 +233,7 @@ int main(int argc, char *argv[])
* with kvm_sync_regs values.
*/
run->kvm_valid_regs = 0;
run->kvm_dirty_regs = KVM_SYNC_X86_VALID_FIELDS;
run->kvm_dirty_regs = TEST_SYNC_FIELDS;
run->s.regs.regs.r11 = 0xBBBB;
rv = _vcpu_run(vm, VCPU_ID);
TEST_ASSERT(run->exit_reason == KVM_EXIT_IO,

View File

@ -189,8 +189,8 @@ int main(int argc, char *argv[])
struct kvm_cpuid_entry2 *entry = kvm_get_supported_cpuid_entry(1);
if (!(entry->ecx & CPUID_VMX)) {
printf("nested VMX not enabled, skipping test");
return 0;
fprintf(stderr, "nested VMX not enabled, skipping test\n");
exit(KSFT_SKIP);
}
vm = vm_create_default_vmx(VCPU_ID, (void *) l1_guest_code);

View File

@ -211,6 +211,7 @@ static int vgic_debug_show(struct seq_file *s, void *v)
struct vgic_state_iter *iter = (struct vgic_state_iter *)v;
struct vgic_irq *irq;
struct kvm_vcpu *vcpu = NULL;
unsigned long flags;
if (iter->dist_id == 0) {
print_dist_state(s, &kvm->arch.vgic);
@ -227,9 +228,9 @@ static int vgic_debug_show(struct seq_file *s, void *v)
irq = &kvm->arch.vgic.spis[iter->intid - VGIC_NR_PRIVATE_IRQS];
}
spin_lock(&irq->irq_lock);
spin_lock_irqsave(&irq->irq_lock, flags);
print_irq_state(s, irq, vcpu);
spin_unlock(&irq->irq_lock);
spin_unlock_irqrestore(&irq->irq_lock, flags);
return 0;
}

View File

@ -52,6 +52,7 @@ static struct vgic_irq *vgic_add_lpi(struct kvm *kvm, u32 intid,
{
struct vgic_dist *dist = &kvm->arch.vgic;
struct vgic_irq *irq = vgic_get_irq(kvm, NULL, intid), *oldirq;
unsigned long flags;
int ret;
/* In this case there is no put, since we keep the reference. */
@ -71,7 +72,7 @@ static struct vgic_irq *vgic_add_lpi(struct kvm *kvm, u32 intid,
irq->intid = intid;
irq->target_vcpu = vcpu;
spin_lock(&dist->lpi_list_lock);
spin_lock_irqsave(&dist->lpi_list_lock, flags);
/*
* There could be a race with another vgic_add_lpi(), so we need to
@ -99,7 +100,7 @@ static struct vgic_irq *vgic_add_lpi(struct kvm *kvm, u32 intid,
dist->lpi_list_count++;
out_unlock:
spin_unlock(&dist->lpi_list_lock);
spin_unlock_irqrestore(&dist->lpi_list_lock, flags);
/*
* We "cache" the configuration table entries in our struct vgic_irq's.
@ -280,8 +281,8 @@ static int update_lpi_config(struct kvm *kvm, struct vgic_irq *irq,
int ret;
unsigned long flags;
ret = kvm_read_guest(kvm, propbase + irq->intid - GIC_LPI_OFFSET,
&prop, 1);
ret = kvm_read_guest_lock(kvm, propbase + irq->intid - GIC_LPI_OFFSET,
&prop, 1);
if (ret)
return ret;
@ -315,6 +316,7 @@ static int vgic_copy_lpi_list(struct kvm_vcpu *vcpu, u32 **intid_ptr)
{
struct vgic_dist *dist = &vcpu->kvm->arch.vgic;
struct vgic_irq *irq;
unsigned long flags;
u32 *intids;
int irq_count, i = 0;
@ -330,7 +332,7 @@ static int vgic_copy_lpi_list(struct kvm_vcpu *vcpu, u32 **intid_ptr)
if (!intids)
return -ENOMEM;
spin_lock(&dist->lpi_list_lock);
spin_lock_irqsave(&dist->lpi_list_lock, flags);
list_for_each_entry(irq, &dist->lpi_list_head, lpi_list) {
if (i == irq_count)
break;
@ -339,7 +341,7 @@ static int vgic_copy_lpi_list(struct kvm_vcpu *vcpu, u32 **intid_ptr)
continue;
intids[i++] = irq->intid;
}
spin_unlock(&dist->lpi_list_lock);
spin_unlock_irqrestore(&dist->lpi_list_lock, flags);
*intid_ptr = intids;
return i;
@ -348,10 +350,11 @@ static int vgic_copy_lpi_list(struct kvm_vcpu *vcpu, u32 **intid_ptr)
static int update_affinity(struct vgic_irq *irq, struct kvm_vcpu *vcpu)
{
int ret = 0;
unsigned long flags;
spin_lock(&irq->irq_lock);
spin_lock_irqsave(&irq->irq_lock, flags);
irq->target_vcpu = vcpu;
spin_unlock(&irq->irq_lock);
spin_unlock_irqrestore(&irq->irq_lock, flags);
if (irq->hw) {
struct its_vlpi_map map;
@ -441,8 +444,9 @@ static int its_sync_lpi_pending_table(struct kvm_vcpu *vcpu)
* this very same byte in the last iteration. Reuse that.
*/
if (byte_offset != last_byte_offset) {
ret = kvm_read_guest(vcpu->kvm, pendbase + byte_offset,
&pendmask, 1);
ret = kvm_read_guest_lock(vcpu->kvm,
pendbase + byte_offset,
&pendmask, 1);
if (ret) {
kfree(intids);
return ret;
@ -786,7 +790,7 @@ static bool vgic_its_check_id(struct vgic_its *its, u64 baser, u32 id,
return false;
/* Each 1st level entry is represented by a 64-bit value. */
if (kvm_read_guest(its->dev->kvm,
if (kvm_read_guest_lock(its->dev->kvm,
BASER_ADDRESS(baser) + index * sizeof(indirect_ptr),
&indirect_ptr, sizeof(indirect_ptr)))
return false;
@ -1367,8 +1371,8 @@ static void vgic_its_process_commands(struct kvm *kvm, struct vgic_its *its)
cbaser = CBASER_ADDRESS(its->cbaser);
while (its->cwriter != its->creadr) {
int ret = kvm_read_guest(kvm, cbaser + its->creadr,
cmd_buf, ITS_CMD_SIZE);
int ret = kvm_read_guest_lock(kvm, cbaser + its->creadr,
cmd_buf, ITS_CMD_SIZE);
/*
* If kvm_read_guest() fails, this could be due to the guest
* programming a bogus value in CBASER or something else going
@ -1893,7 +1897,7 @@ static int scan_its_table(struct vgic_its *its, gpa_t base, int size, int esz,
int next_offset;
size_t byte_offset;
ret = kvm_read_guest(kvm, gpa, entry, esz);
ret = kvm_read_guest_lock(kvm, gpa, entry, esz);
if (ret)
return ret;
@ -2263,7 +2267,7 @@ static int vgic_its_restore_cte(struct vgic_its *its, gpa_t gpa, int esz)
int ret;
BUG_ON(esz > sizeof(val));
ret = kvm_read_guest(kvm, gpa, &val, esz);
ret = kvm_read_guest_lock(kvm, gpa, &val, esz);
if (ret)
return ret;
val = le64_to_cpu(val);

View File

@ -344,7 +344,7 @@ retry:
bit_nr = irq->intid % BITS_PER_BYTE;
ptr = pendbase + byte_offset;
ret = kvm_read_guest(kvm, ptr, &val, 1);
ret = kvm_read_guest_lock(kvm, ptr, &val, 1);
if (ret)
return ret;
@ -397,7 +397,7 @@ int vgic_v3_save_pending_tables(struct kvm *kvm)
ptr = pendbase + byte_offset;
if (byte_offset != last_byte_offset) {
ret = kvm_read_guest(kvm, ptr, &val, 1);
ret = kvm_read_guest_lock(kvm, ptr, &val, 1);
if (ret)
return ret;
last_byte_offset = byte_offset;

View File

@ -43,9 +43,13 @@ struct vgic_global kvm_vgic_global_state __ro_after_init = {
* kvm->lock (mutex)
* its->cmd_lock (mutex)
* its->its_lock (mutex)
* vgic_cpu->ap_list_lock
* kvm->lpi_list_lock
* vgic_irq->irq_lock
* vgic_cpu->ap_list_lock must be taken with IRQs disabled
* kvm->lpi_list_lock must be taken with IRQs disabled
* vgic_irq->irq_lock must be taken with IRQs disabled
*
* As the ap_list_lock might be taken from the timer interrupt handler,
* we have to disable IRQs before taking this lock and everything lower
* than it.
*
* If you need to take multiple locks, always take the upper lock first,
* then the lower ones, e.g. first take the its_lock, then the irq_lock.
@ -72,8 +76,9 @@ static struct vgic_irq *vgic_get_lpi(struct kvm *kvm, u32 intid)
{
struct vgic_dist *dist = &kvm->arch.vgic;
struct vgic_irq *irq = NULL;
unsigned long flags;
spin_lock(&dist->lpi_list_lock);
spin_lock_irqsave(&dist->lpi_list_lock, flags);
list_for_each_entry(irq, &dist->lpi_list_head, lpi_list) {
if (irq->intid != intid)
@ -89,7 +94,7 @@ static struct vgic_irq *vgic_get_lpi(struct kvm *kvm, u32 intid)
irq = NULL;
out_unlock:
spin_unlock(&dist->lpi_list_lock);
spin_unlock_irqrestore(&dist->lpi_list_lock, flags);
return irq;
}
@ -134,19 +139,20 @@ static void vgic_irq_release(struct kref *ref)
void vgic_put_irq(struct kvm *kvm, struct vgic_irq *irq)
{
struct vgic_dist *dist = &kvm->arch.vgic;
unsigned long flags;
if (irq->intid < VGIC_MIN_LPI)
return;
spin_lock(&dist->lpi_list_lock);
spin_lock_irqsave(&dist->lpi_list_lock, flags);
if (!kref_put(&irq->refcount, vgic_irq_release)) {
spin_unlock(&dist->lpi_list_lock);
spin_unlock_irqrestore(&dist->lpi_list_lock, flags);
return;
};
list_del(&irq->lpi_list);
dist->lpi_list_count--;
spin_unlock(&dist->lpi_list_lock);
spin_unlock_irqrestore(&dist->lpi_list_lock, flags);
kfree(irq);
}