Merge tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm
Pull KVM updates from Paolo Bonzini:
"One NULL pointer dereference, and two fixes for regressions introduced
during the merge window.
The rest are fixes for MIPS, s390 and nested VMX"
* tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm:
kvm: x86: Check memopp before dereference (CVE-2016-8630)
kvm: nVMX: VMCLEAR an active shadow VMCS after last use
KVM: x86: drop TSC offsetting kvm_x86_ops to fix KVM_GET/SET_CLOCK
KVM: x86: fix wbinvd_dirty_mask use-after-free
kvm/x86: Show WRMSR data is in hex
kvm: nVMX: Fix kernel panics induced by illegal INVEPT/INVVPID types
KVM: document lock orders
KVM: fix OOPS on flush_work
KVM: s390: Fix STHYI buffer alignment for diag224
KVM: MIPS: Precalculate MMIO load resume PC
KVM: MIPS: Make ERET handle ERL before EXL
KVM: MIPS: Fix lazy user ASID regenerate for SMP
diff --git a/Documentation/virtual/kvm/locking.txt b/Documentation/virtual/kvm/locking.txt
index f2491a8..e5dd9f4 100644
--- a/Documentation/virtual/kvm/locking.txt
+++ b/Documentation/virtual/kvm/locking.txt
@@ -4,7 +4,17 @@
1. Acquisition Orders
---------------------
-(to be written)
+The acquisition orders for mutexes are as follows:
+
+- kvm->lock is taken outside vcpu->mutex
+
+- kvm->lock is taken outside kvm->slots_lock and kvm->irq_lock
+
+- kvm->slots_lock is taken outside kvm->irq_lock, though acquiring
+ them together is quite rare.
+
+For spinlocks, kvm_lock is taken outside kvm->mmu_lock. Everything
+else is a leaf: no other lock is taken inside the critical sections.
2: Exception
------------
diff --git a/arch/mips/include/asm/kvm_host.h b/arch/mips/include/asm/kvm_host.h
index 07f58cf..bebec37 100644
--- a/arch/mips/include/asm/kvm_host.h
+++ b/arch/mips/include/asm/kvm_host.h
@@ -293,7 +293,10 @@
/* Host KSEG0 address of the EI/DI offset */
void *kseg0_commpage;
- u32 io_gpr; /* GPR used as IO source/target */
+ /* Resume PC after MMIO completion */
+ unsigned long io_pc;
+ /* GPR used as IO source/target */
+ u32 io_gpr;
struct hrtimer comparecount_timer;
/* Count timer control KVM register */
@@ -315,8 +318,6 @@
/* Bitmask of pending exceptions to be cleared */
unsigned long pending_exceptions_clr;
- u32 pending_load_cause;
-
/* Save/Restore the entryhi register when are are preempted/scheduled back in */
unsigned long preempt_entryhi;
diff --git a/arch/mips/kvm/emulate.c b/arch/mips/kvm/emulate.c
index 8770f32..aa09374 100644
--- a/arch/mips/kvm/emulate.c
+++ b/arch/mips/kvm/emulate.c
@@ -790,15 +790,15 @@
struct mips_coproc *cop0 = vcpu->arch.cop0;
enum emulation_result er = EMULATE_DONE;
- if (kvm_read_c0_guest_status(cop0) & ST0_EXL) {
+ if (kvm_read_c0_guest_status(cop0) & ST0_ERL) {
+ kvm_clear_c0_guest_status(cop0, ST0_ERL);
+ vcpu->arch.pc = kvm_read_c0_guest_errorepc(cop0);
+ } else if (kvm_read_c0_guest_status(cop0) & ST0_EXL) {
kvm_debug("[%#lx] ERET to %#lx\n", vcpu->arch.pc,
kvm_read_c0_guest_epc(cop0));
kvm_clear_c0_guest_status(cop0, ST0_EXL);
vcpu->arch.pc = kvm_read_c0_guest_epc(cop0);
- } else if (kvm_read_c0_guest_status(cop0) & ST0_ERL) {
- kvm_clear_c0_guest_status(cop0, ST0_ERL);
- vcpu->arch.pc = kvm_read_c0_guest_errorepc(cop0);
} else {
kvm_err("[%#lx] ERET when MIPS_SR_EXL|MIPS_SR_ERL == 0\n",
vcpu->arch.pc);
@@ -1528,13 +1528,25 @@
struct kvm_vcpu *vcpu)
{
enum emulation_result er = EMULATE_DO_MMIO;
+ unsigned long curr_pc;
u32 op, rt;
u32 bytes;
rt = inst.i_format.rt;
op = inst.i_format.opcode;
- vcpu->arch.pending_load_cause = cause;
+ /*
+ * Find the resume PC now while we have safe and easy access to the
+ * prior branch instruction, and save it for
+ * kvm_mips_complete_mmio_load() to restore later.
+ */
+ curr_pc = vcpu->arch.pc;
+ er = update_pc(vcpu, cause);
+ if (er == EMULATE_FAIL)
+ return er;
+ vcpu->arch.io_pc = vcpu->arch.pc;
+ vcpu->arch.pc = curr_pc;
+
vcpu->arch.io_gpr = rt;
switch (op) {
@@ -2494,9 +2506,8 @@
goto done;
}
- er = update_pc(vcpu, vcpu->arch.pending_load_cause);
- if (er == EMULATE_FAIL)
- return er;
+ /* Restore saved resume PC */
+ vcpu->arch.pc = vcpu->arch.io_pc;
switch (run->mmio.len) {
case 4:
@@ -2518,11 +2529,6 @@
break;
}
- if (vcpu->arch.pending_load_cause & CAUSEF_BD)
- kvm_debug("[%#lx] Completing %d byte BD Load to gpr %d (0x%08lx) type %d\n",
- vcpu->arch.pc, run->mmio.len, vcpu->arch.io_gpr, *gpr,
- vcpu->mmio_needed);
-
done:
return er;
}
diff --git a/arch/mips/kvm/mips.c b/arch/mips/kvm/mips.c
index 622037d..06a60b1 100644
--- a/arch/mips/kvm/mips.c
+++ b/arch/mips/kvm/mips.c
@@ -426,7 +426,7 @@
static void kvm_mips_check_asids(struct kvm_vcpu *vcpu)
{
struct mips_coproc *cop0 = vcpu->arch.cop0;
- int cpu = smp_processor_id();
+ int i, cpu = smp_processor_id();
unsigned int gasid;
/*
@@ -442,6 +442,9 @@
vcpu);
vcpu->arch.guest_user_asid[cpu] =
vcpu->arch.guest_user_mm.context.asid[cpu];
+ for_each_possible_cpu(i)
+ if (i != cpu)
+ vcpu->arch.guest_user_asid[cpu] = 0;
vcpu->arch.last_user_gasid = gasid;
}
}
diff --git a/arch/mips/kvm/mmu.c b/arch/mips/kvm/mmu.c
index 03883ba..3b677c8 100644
--- a/arch/mips/kvm/mmu.c
+++ b/arch/mips/kvm/mmu.c
@@ -260,13 +260,9 @@
if ((vcpu->arch.guest_user_asid[cpu] ^ asid_cache(cpu)) &
asid_version_mask(cpu)) {
- u32 gasid = kvm_read_c0_guest_entryhi(vcpu->arch.cop0) &
- KVM_ENTRYHI_ASID;
-
kvm_get_new_mmu_context(&vcpu->arch.guest_user_mm, cpu, vcpu);
vcpu->arch.guest_user_asid[cpu] =
vcpu->arch.guest_user_mm.context.asid[cpu];
- vcpu->arch.last_user_gasid = gasid;
newasid++;
kvm_debug("[%d]: cpu_context: %#lx\n", cpu,
diff --git a/arch/s390/kvm/sthyi.c b/arch/s390/kvm/sthyi.c
index bd98b7d..05c98bb 100644
--- a/arch/s390/kvm/sthyi.c
+++ b/arch/s390/kvm/sthyi.c
@@ -315,7 +315,7 @@
if (r < 0)
goto out;
- diag224_buf = kmalloc(PAGE_SIZE, GFP_KERNEL | GFP_DMA);
+ diag224_buf = (void *)__get_free_page(GFP_KERNEL | GFP_DMA);
if (!diag224_buf || diag224(diag224_buf))
goto out;
@@ -378,7 +378,7 @@
sctns->par.infpval1 |= PAR_WGHT_VLD;
out:
- kfree(diag224_buf);
+ free_page((unsigned long)diag224_buf);
vfree(diag204_buf);
}
diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h
index 4b20f73..bdde807 100644
--- a/arch/x86/include/asm/kvm_host.h
+++ b/arch/x86/include/asm/kvm_host.h
@@ -948,7 +948,6 @@
int (*get_lpage_level)(void);
bool (*rdtscp_supported)(void);
bool (*invpcid_supported)(void);
- void (*adjust_tsc_offset_guest)(struct kvm_vcpu *vcpu, s64 adjustment);
void (*set_tdp_cr3)(struct kvm_vcpu *vcpu, unsigned long cr3);
@@ -958,8 +957,6 @@
void (*write_tsc_offset)(struct kvm_vcpu *vcpu, u64 offset);
- u64 (*read_l1_tsc)(struct kvm_vcpu *vcpu, u64 host_tsc);
-
void (*get_exit_info)(struct kvm_vcpu *vcpu, u64 *info1, u64 *info2);
int (*check_intercept)(struct kvm_vcpu *vcpu,
diff --git a/arch/x86/kvm/emulate.c b/arch/x86/kvm/emulate.c
index 4e95d3e..cbd7b92 100644
--- a/arch/x86/kvm/emulate.c
+++ b/arch/x86/kvm/emulate.c
@@ -5045,7 +5045,7 @@
/* Decode and fetch the destination operand: register or memory. */
rc = decode_operand(ctxt, &ctxt->dst, (ctxt->d >> DstShift) & OpMask);
- if (ctxt->rip_relative)
+ if (ctxt->rip_relative && likely(ctxt->memopp))
ctxt->memopp->addr.mem.ea = address_mask(ctxt,
ctxt->memopp->addr.mem.ea + ctxt->_eip);
diff --git a/arch/x86/kvm/svm.c b/arch/x86/kvm/svm.c
index f8157a3..8ca1eca 100644
--- a/arch/x86/kvm/svm.c
+++ b/arch/x86/kvm/svm.c
@@ -1138,21 +1138,6 @@
mark_dirty(svm->vmcb, VMCB_INTERCEPTS);
}
-static void svm_adjust_tsc_offset_guest(struct kvm_vcpu *vcpu, s64 adjustment)
-{
- struct vcpu_svm *svm = to_svm(vcpu);
-
- svm->vmcb->control.tsc_offset += adjustment;
- if (is_guest_mode(vcpu))
- svm->nested.hsave->control.tsc_offset += adjustment;
- else
- trace_kvm_write_tsc_offset(vcpu->vcpu_id,
- svm->vmcb->control.tsc_offset - adjustment,
- svm->vmcb->control.tsc_offset);
-
- mark_dirty(svm->vmcb, VMCB_INTERCEPTS);
-}
-
static void avic_init_vmcb(struct vcpu_svm *svm)
{
struct vmcb *vmcb = svm->vmcb;
@@ -3449,12 +3434,6 @@
return 0;
}
-static u64 svm_read_l1_tsc(struct kvm_vcpu *vcpu, u64 host_tsc)
-{
- struct vmcb *vmcb = get_host_vmcb(to_svm(vcpu));
- return vmcb->control.tsc_offset + host_tsc;
-}
-
static int svm_get_msr(struct kvm_vcpu *vcpu, struct msr_data *msr_info)
{
struct vcpu_svm *svm = to_svm(vcpu);
@@ -5422,8 +5401,6 @@
.has_wbinvd_exit = svm_has_wbinvd_exit,
.write_tsc_offset = svm_write_tsc_offset,
- .adjust_tsc_offset_guest = svm_adjust_tsc_offset_guest,
- .read_l1_tsc = svm_read_l1_tsc,
.set_tdp_cr3 = set_tdp_cr3,
diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
index cf1b16d..5382b82 100644
--- a/arch/x86/kvm/vmx.c
+++ b/arch/x86/kvm/vmx.c
@@ -187,6 +187,7 @@
*/
struct loaded_vmcs {
struct vmcs *vmcs;
+ struct vmcs *shadow_vmcs;
int cpu;
int launched;
struct list_head loaded_vmcss_on_cpu_link;
@@ -411,7 +412,6 @@
* memory during VMXOFF, VMCLEAR, VMPTRLD.
*/
struct vmcs12 *cached_vmcs12;
- struct vmcs *current_shadow_vmcs;
/*
* Indicates if the shadow vmcs must be updated with the
* data hold by vmcs12
@@ -421,7 +421,6 @@
/* vmcs02_list cache of VMCSs recently used to run L2 guests */
struct list_head vmcs02_pool;
int vmcs02_num;
- u64 vmcs01_tsc_offset;
bool change_vmcs01_virtual_x2apic_mode;
/* L2 must run next, and mustn't decide to exit to L1. */
bool nested_run_pending;
@@ -1419,6 +1418,8 @@
static inline void loaded_vmcs_init(struct loaded_vmcs *loaded_vmcs)
{
vmcs_clear(loaded_vmcs->vmcs);
+ if (loaded_vmcs->shadow_vmcs && loaded_vmcs->launched)
+ vmcs_clear(loaded_vmcs->shadow_vmcs);
loaded_vmcs->cpu = -1;
loaded_vmcs->launched = 0;
}
@@ -2605,20 +2606,6 @@
}
/*
- * Like guest_read_tsc, but always returns L1's notion of the timestamp
- * counter, even if a nested guest (L2) is currently running.
- */
-static u64 vmx_read_l1_tsc(struct kvm_vcpu *vcpu, u64 host_tsc)
-{
- u64 tsc_offset;
-
- tsc_offset = is_guest_mode(vcpu) ?
- to_vmx(vcpu)->nested.vmcs01_tsc_offset :
- vmcs_read64(TSC_OFFSET);
- return host_tsc + tsc_offset;
-}
-
-/*
* writes 'offset' into guest's timestamp counter offset register
*/
static void vmx_write_tsc_offset(struct kvm_vcpu *vcpu, u64 offset)
@@ -2631,7 +2618,6 @@
* to the newly set TSC to get L2's TSC.
*/
struct vmcs12 *vmcs12;
- to_vmx(vcpu)->nested.vmcs01_tsc_offset = offset;
/* recalculate vmcs02.TSC_OFFSET: */
vmcs12 = get_vmcs12(vcpu);
vmcs_write64(TSC_OFFSET, offset +
@@ -2644,19 +2630,6 @@
}
}
-static void vmx_adjust_tsc_offset_guest(struct kvm_vcpu *vcpu, s64 adjustment)
-{
- u64 offset = vmcs_read64(TSC_OFFSET);
-
- vmcs_write64(TSC_OFFSET, offset + adjustment);
- if (is_guest_mode(vcpu)) {
- /* Even when running L2, the adjustment needs to apply to L1 */
- to_vmx(vcpu)->nested.vmcs01_tsc_offset += adjustment;
- } else
- trace_kvm_write_tsc_offset(vcpu->vcpu_id, offset,
- offset + adjustment);
-}
-
static bool guest_cpuid_has_vmx(struct kvm_vcpu *vcpu)
{
struct kvm_cpuid_entry2 *best = kvm_find_cpuid_entry(vcpu, 1, 0);
@@ -3562,6 +3535,7 @@
loaded_vmcs_clear(loaded_vmcs);
free_vmcs(loaded_vmcs->vmcs);
loaded_vmcs->vmcs = NULL;
+ WARN_ON(loaded_vmcs->shadow_vmcs != NULL);
}
static void free_kvm_area(void)
@@ -6696,6 +6670,7 @@
if (!item)
return NULL;
item->vmcs02.vmcs = alloc_vmcs();
+ item->vmcs02.shadow_vmcs = NULL;
if (!item->vmcs02.vmcs) {
kfree(item);
return NULL;
@@ -7072,7 +7047,7 @@
shadow_vmcs->revision_id |= (1u << 31);
/* init shadow vmcs */
vmcs_clear(shadow_vmcs);
- vmx->nested.current_shadow_vmcs = shadow_vmcs;
+ vmx->vmcs01.shadow_vmcs = shadow_vmcs;
}
INIT_LIST_HEAD(&(vmx->nested.vmcs02_pool));
@@ -7174,8 +7149,11 @@
free_page((unsigned long)vmx->nested.msr_bitmap);
vmx->nested.msr_bitmap = NULL;
}
- if (enable_shadow_vmcs)
- free_vmcs(vmx->nested.current_shadow_vmcs);
+ if (enable_shadow_vmcs) {
+ vmcs_clear(vmx->vmcs01.shadow_vmcs);
+ free_vmcs(vmx->vmcs01.shadow_vmcs);
+ vmx->vmcs01.shadow_vmcs = NULL;
+ }
kfree(vmx->nested.cached_vmcs12);
/* Unpin physical memory we referred to in current vmcs02 */
if (vmx->nested.apic_access_page) {
@@ -7352,7 +7330,7 @@
int i;
unsigned long field;
u64 field_value;
- struct vmcs *shadow_vmcs = vmx->nested.current_shadow_vmcs;
+ struct vmcs *shadow_vmcs = vmx->vmcs01.shadow_vmcs;
const unsigned long *fields = shadow_read_write_fields;
const int num_fields = max_shadow_read_write_fields;
@@ -7401,7 +7379,7 @@
int i, q;
unsigned long field;
u64 field_value = 0;
- struct vmcs *shadow_vmcs = vmx->nested.current_shadow_vmcs;
+ struct vmcs *shadow_vmcs = vmx->vmcs01.shadow_vmcs;
vmcs_load(shadow_vmcs);
@@ -7591,7 +7569,7 @@
vmcs_set_bits(SECONDARY_VM_EXEC_CONTROL,
SECONDARY_EXEC_SHADOW_VMCS);
vmcs_write64(VMCS_LINK_POINTER,
- __pa(vmx->nested.current_shadow_vmcs));
+ __pa(vmx->vmcs01.shadow_vmcs));
vmx->nested.sync_shadow_vmcs = true;
}
}
@@ -7659,7 +7637,7 @@
types = (vmx->nested.nested_vmx_ept_caps >> VMX_EPT_EXTENT_SHIFT) & 6;
- if (!(types & (1UL << type))) {
+ if (type >= 32 || !(types & (1 << type))) {
nested_vmx_failValid(vcpu,
VMXERR_INVALID_OPERAND_TO_INVEPT_INVVPID);
skip_emulated_instruction(vcpu);
@@ -7722,7 +7700,7 @@
types = (vmx->nested.nested_vmx_vpid_caps >> 8) & 0x7;
- if (!(types & (1UL << type))) {
+ if (type >= 32 || !(types & (1 << type))) {
nested_vmx_failValid(vcpu,
VMXERR_INVALID_OPERAND_TO_INVEPT_INVVPID);
skip_emulated_instruction(vcpu);
@@ -9156,6 +9134,7 @@
vmx->loaded_vmcs = &vmx->vmcs01;
vmx->loaded_vmcs->vmcs = alloc_vmcs();
+ vmx->loaded_vmcs->shadow_vmcs = NULL;
if (!vmx->loaded_vmcs->vmcs)
goto free_msrs;
if (!vmm_exclusive)
@@ -10061,9 +10040,9 @@
if (vmcs12->cpu_based_vm_exec_control & CPU_BASED_USE_TSC_OFFSETING)
vmcs_write64(TSC_OFFSET,
- vmx->nested.vmcs01_tsc_offset + vmcs12->tsc_offset);
+ vcpu->arch.tsc_offset + vmcs12->tsc_offset);
else
- vmcs_write64(TSC_OFFSET, vmx->nested.vmcs01_tsc_offset);
+ vmcs_write64(TSC_OFFSET, vcpu->arch.tsc_offset);
if (kvm_has_tsc_control)
decache_tsc_multiplier(vmx);
@@ -10293,8 +10272,6 @@
enter_guest_mode(vcpu);
- vmx->nested.vmcs01_tsc_offset = vmcs_read64(TSC_OFFSET);
-
if (!(vmcs12->vm_entry_controls & VM_ENTRY_LOAD_DEBUG_CONTROLS))
vmx->nested.vmcs01_debugctl = vmcs_read64(GUEST_IA32_DEBUGCTL);
@@ -10818,7 +10795,7 @@
load_vmcs12_host_state(vcpu, vmcs12);
/* Update any VMCS fields that might have changed while L2 ran */
- vmcs_write64(TSC_OFFSET, vmx->nested.vmcs01_tsc_offset);
+ vmcs_write64(TSC_OFFSET, vcpu->arch.tsc_offset);
if (vmx->hv_deadline_tsc == -1)
vmcs_clear_bits(PIN_BASED_VM_EXEC_CONTROL,
PIN_BASED_VMX_PREEMPTION_TIMER);
@@ -11339,8 +11316,6 @@
.has_wbinvd_exit = cpu_has_vmx_wbinvd_exit,
.write_tsc_offset = vmx_write_tsc_offset,
- .adjust_tsc_offset_guest = vmx_adjust_tsc_offset_guest,
- .read_l1_tsc = vmx_read_l1_tsc,
.set_tdp_cr3 = vmx_set_cr3,
diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index e375235..3017de0 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -1409,7 +1409,7 @@
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));
+ return vcpu->arch.tsc_offset + kvm_scale_tsc(vcpu, host_tsc);
}
EXPORT_SYMBOL_GPL(kvm_read_l1_tsc);
@@ -1547,7 +1547,7 @@
static inline void adjust_tsc_offset_guest(struct kvm_vcpu *vcpu,
s64 adjustment)
{
- kvm_x86_ops->adjust_tsc_offset_guest(vcpu, adjustment);
+ kvm_vcpu_write_tsc_offset(vcpu, vcpu->arch.tsc_offset + adjustment);
}
static inline void adjust_tsc_offset_host(struct kvm_vcpu *vcpu, s64 adjustment)
@@ -1555,7 +1555,7 @@
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);
+ adjust_tsc_offset_guest(vcpu, adjustment);
}
#ifdef CONFIG_X86_64
@@ -2262,7 +2262,7 @@
/* Drop writes to this legacy MSR -- see rdmsr
* counterpart for further detail.
*/
- vcpu_unimpl(vcpu, "ignored wrmsr: 0x%x data %llx\n", msr, data);
+ vcpu_unimpl(vcpu, "ignored wrmsr: 0x%x data 0x%llx\n", msr, data);
break;
case MSR_AMD64_OSVW_ID_LENGTH:
if (!guest_cpuid_has_osvw(vcpu))
@@ -2280,11 +2280,11 @@
if (kvm_pmu_is_valid_msr(vcpu, msr))
return kvm_pmu_set_msr(vcpu, msr_info);
if (!ignore_msrs) {
- vcpu_unimpl(vcpu, "unhandled wrmsr: 0x%x data %llx\n",
+ vcpu_unimpl(vcpu, "unhandled wrmsr: 0x%x data 0x%llx\n",
msr, data);
return 1;
} else {
- vcpu_unimpl(vcpu, "ignored wrmsr: 0x%x data %llx\n",
+ vcpu_unimpl(vcpu, "ignored wrmsr: 0x%x data 0x%llx\n",
msr, data);
break;
}
@@ -7410,10 +7410,12 @@
void kvm_arch_vcpu_free(struct kvm_vcpu *vcpu)
{
+ void *wbinvd_dirty_mask = vcpu->arch.wbinvd_dirty_mask;
+
kvmclock_reset(vcpu);
- free_cpumask_var(vcpu->arch.wbinvd_dirty_mask);
kvm_x86_ops->vcpu_free(vcpu);
+ free_cpumask_var(wbinvd_dirty_mask);
}
struct kvm_vcpu *kvm_arch_vcpu_create(struct kvm *kvm,
diff --git a/virt/kvm/eventfd.c b/virt/kvm/eventfd.c
index f397e9b..a29786d 100644
--- a/virt/kvm/eventfd.c
+++ b/virt/kvm/eventfd.c
@@ -42,6 +42,7 @@
#ifdef CONFIG_HAVE_KVM_IRQFD
+static struct workqueue_struct *irqfd_cleanup_wq;
static void
irqfd_inject(struct work_struct *work)
@@ -167,7 +168,7 @@
list_del_init(&irqfd->list);
- schedule_work(&irqfd->shutdown);
+ queue_work(irqfd_cleanup_wq, &irqfd->shutdown);
}
int __attribute__((weak)) kvm_arch_set_irq_inatomic(
@@ -554,7 +555,7 @@
* so that we guarantee there will not be any more interrupts on this
* gsi once this deassign function returns.
*/
- flush_work(&irqfd->shutdown);
+ flush_workqueue(irqfd_cleanup_wq);
return 0;
}
@@ -591,7 +592,7 @@
* Block until we know all outstanding shutdown jobs have completed
* since we do not take a kvm* reference.
*/
- flush_work(&irqfd->shutdown);
+ flush_workqueue(irqfd_cleanup_wq);
}
@@ -621,8 +622,23 @@
spin_unlock_irq(&kvm->irqfds.lock);
}
+/*
+ * create a host-wide workqueue for issuing deferred shutdown requests
+ * aggregated from all vm* instances. We need our own isolated
+ * queue to ease flushing work items when a VM exits.
+ */
+int kvm_irqfd_init(void)
+{
+ irqfd_cleanup_wq = alloc_workqueue("kvm-irqfd-cleanup", 0, 0);
+ if (!irqfd_cleanup_wq)
+ return -ENOMEM;
+
+ return 0;
+}
+
void kvm_irqfd_exit(void)
{
+ destroy_workqueue(irqfd_cleanup_wq);
}
#endif
diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c
index 2907b7b..5c36034 100644
--- a/virt/kvm/kvm_main.c
+++ b/virt/kvm/kvm_main.c
@@ -3844,7 +3844,12 @@
* kvm_arch_init makes sure there's at most one caller
* for architectures that support multiple implementations,
* like intel and amd on x86.
+ * kvm_arch_init must be called before kvm_irqfd_init to avoid creating
+ * conflicts in case kvm is already setup for another implementation.
*/
+ r = kvm_irqfd_init();
+ if (r)
+ goto out_irqfd;
if (!zalloc_cpumask_var(&cpus_hardware_enabled, GFP_KERNEL)) {
r = -ENOMEM;
@@ -3926,6 +3931,7 @@
free_cpumask_var(cpus_hardware_enabled);
out_free_0:
kvm_irqfd_exit();
+out_irqfd:
kvm_arch_exit();
out_fail:
return r;
