From 6c08ec1216b7b2f84b9755f339d6009768599256 Mon Sep 17 00:00:00 2001 From: Michael Roth Date: Thu, 8 Nov 2018 21:27:23 -0600 Subject: [PATCH 01/15] KVM: PPC: Book3S HV: Fix handling for interrupted H_ENTER_NESTED While running a nested guest VCPU on L0 via H_ENTER_NESTED hcall, a pending signal in the L0 QEMU process can generate the following sequence: ret0 = kvmppc_pseries_do_hcall() ret1 = kvmhv_enter_nested_guest() ret2 = kvmhv_run_single_vcpu() if (ret2 == -EINTR) return H_INTERRUPT if (ret1 == H_INTERRUPT) kvmppc_set_gpr(vcpu, 3, 0) return -EINTR /* skipped: */ kvmppc_set_gpr(vcpu, 3, ret) vcpu->arch.hcall_needed = 0 return RESUME_GUEST which causes an exit to L0 userspace with ret0 == -EINTR. The intention seems to be to set the hcall return value to 0 (via VCPU r3) so that L1 will see a successful return from H_ENTER_NESTED once we resume executing the VCPU. However, because we don't set vcpu->arch.hcall_needed = 0, we do the following once userspace resumes execution via kvm_arch_vcpu_ioctl_run(): ... } else if (vcpu->arch.hcall_needed) { int i kvmppc_set_gpr(vcpu, 3, run->papr_hcall.ret); for (i = 0; i < 9; ++i) kvmppc_set_gpr(vcpu, 4 + i, run->papr_hcall.args[i]); vcpu->arch.hcall_needed = 0; since vcpu->arch.hcall_needed == 1 indicates that userspace should have handled the hcall and stored the return value in run->papr_hcall.ret. Since that's not the case here, we can get an unexpected value in VCPU r3, which can result in kvmhv_p9_guest_entry() reporting an unexpected trap value when it returns from H_ENTER_NESTED, causing the following register dump to console via subsequent call to kvmppc_handle_exit_hv() in L1: [ 350.612854] vcpu 00000000f9564cf8 (0): [ 350.612915] pc = c00000000013eb98 msr = 8000000000009033 trap = 1 [ 350.613020] r 0 = c0000000004b9044 r16 = 0000000000000000 [ 350.613075] r 1 = c00000007cffba30 r17 = 0000000000000000 [ 350.613120] r 2 = c00000000178c100 r18 = 00007fffc24f3b50 [ 350.613166] r 3 = c00000007ef52480 r19 = 00007fffc24fff58 [ 350.613212] r 4 = 0000000000000000 r20 = 00000a1e96ece9d0 [ 350.613253] r 5 = 70616d00746f6f72 r21 = 00000a1ea117c9b0 [ 350.613295] r 6 = 0000000000000020 r22 = 00000a1ea1184360 [ 350.613338] r 7 = c0000000783be440 r23 = 0000000000000003 [ 350.613380] r 8 = fffffffffffffffc r24 = 00000a1e96e9e124 [ 350.613423] r 9 = c00000007ef52490 r25 = 00000000000007ff [ 350.613469] r10 = 0000000000000004 r26 = c00000007eb2f7a0 [ 350.613513] r11 = b0616d0009eccdb2 r27 = c00000007cffbb10 [ 350.613556] r12 = c0000000004b9000 r28 = c00000007d83a2c0 [ 350.613597] r13 = c000000001b00000 r29 = c0000000783cdf68 [ 350.613639] r14 = 0000000000000000 r30 = 0000000000000000 [ 350.613681] r15 = 0000000000000000 r31 = c00000007cffbbf0 [ 350.613723] ctr = c0000000004b9000 lr = c0000000004b9044 [ 350.613765] srr0 = 0000772f954dd48c srr1 = 800000000280f033 [ 350.613808] sprg0 = 0000000000000000 sprg1 = c000000001b00000 [ 350.613859] sprg2 = 0000772f9565a280 sprg3 = 0000000000000000 [ 350.613911] cr = 88002848 xer = 0000000020040000 dsisr = 42000000 [ 350.613962] dar = 0000772f95390000 [ 350.614031] fault dar = c000000244b278c0 dsisr = 00000000 [ 350.614073] SLB (0 entries): [ 350.614157] lpcr = 0040000003d40413 sdr1 = 0000000000000000 last_inst = ffffffff [ 350.614252] trap=0x1 | pc=0xc00000000013eb98 | msr=0x8000000000009033 followed by L1's QEMU reporting the following before stopping execution of the nested guest: KVM: unknown exit, hardware reason 1 NIP c00000000013eb98 LR c0000000004b9044 CTR c0000000004b9000 XER 0000000020040000 CPU#0 MSR 8000000000009033 HID0 0000000000000000 HF 8000000000000000 iidx 3 didx 3 TB 00000000 00000000 DECR 00000000 GPR00 c0000000004b9044 c00000007cffba30 c00000000178c100 c00000007ef52480 GPR04 0000000000000000 70616d00746f6f72 0000000000000020 c0000000783be440 GPR08 fffffffffffffffc c00000007ef52490 0000000000000004 b0616d0009eccdb2 GPR12 c0000000004b9000 c000000001b00000 0000000000000000 0000000000000000 GPR16 0000000000000000 0000000000000000 00007fffc24f3b50 00007fffc24fff58 GPR20 00000a1e96ece9d0 00000a1ea117c9b0 00000a1ea1184360 0000000000000003 GPR24 00000a1e96e9e124 00000000000007ff c00000007eb2f7a0 c00000007cffbb10 GPR28 c00000007d83a2c0 c0000000783cdf68 0000000000000000 c00000007cffbbf0 CR 88002848 [ L L - - E L G L ] RES ffffffffffffffff SRR0 0000772f954dd48c SRR1 800000000280f033 PVR 00000000004e1202 VRSAVE 0000000000000000 SPRG0 0000000000000000 SPRG1 c000000001b00000 SPRG2 0000772f9565a280 SPRG3 0000000000000000 SPRG4 0000000000000000 SPRG5 0000000000000000 SPRG6 0000000000000000 SPRG7 0000000000000000 HSRR0 0000000000000000 HSRR1 0000000000000000 CFAR 0000000000000000 LPCR 0000000003d40413 PTCR 0000000000000000 DAR 0000772f95390000 DSISR 0000000042000000 Fix this by setting vcpu->arch.hcall_needed = 0 to indicate completion of H_ENTER_NESTED before we exit to L0 userspace. Fixes: 360cae313702 ("KVM: PPC: Book3S HV: Nested guest entry via hypercall") Cc: linuxppc-dev@ozlabs.org Cc: David Gibson Signed-off-by: Michael Roth Reviewed-by: Suraj Jitindar Singh Signed-off-by: Paul Mackerras --- arch/powerpc/kvm/book3s_hv.c | 1 + 1 file changed, 1 insertion(+) diff --git a/arch/powerpc/kvm/book3s_hv.c b/arch/powerpc/kvm/book3s_hv.c index d65b961661fb..a56f8413758a 100644 --- a/arch/powerpc/kvm/book3s_hv.c +++ b/arch/powerpc/kvm/book3s_hv.c @@ -983,6 +983,7 @@ int kvmppc_pseries_do_hcall(struct kvm_vcpu *vcpu) ret = kvmhv_enter_nested_guest(vcpu); if (ret == H_INTERRUPT) { kvmppc_set_gpr(vcpu, 3, 0); + vcpu->arch.hcall_needed = 0; return -EINTR; } break; From a87c99e61236ba8ca962ce97a19fab5ebd588d35 Mon Sep 17 00:00:00 2001 From: Luiz Capitulino Date: Fri, 23 Nov 2018 12:02:14 -0500 Subject: [PATCH 02/15] KVM: VMX: re-add ple_gap module parameter Apparently, the ple_gap parameter was accidentally removed by commit c8e88717cfc6b36bedea22368d97667446318291. Add it back. Signed-off-by: Luiz Capitulino Cc: stable@vger.kernel.org Fixes: c8e88717cfc6b36bedea22368d97667446318291 Signed-off-by: Paolo Bonzini --- arch/x86/kvm/vmx.c | 1 + 1 file changed, 1 insertion(+) diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c index 4555077d69ce..be6f13f1c25f 100644 --- a/arch/x86/kvm/vmx.c +++ b/arch/x86/kvm/vmx.c @@ -174,6 +174,7 @@ module_param_named(preemption_timer, enable_preemption_timer, bool, S_IRUGO); * refer SDM volume 3b section 21.6.13 & 22.1.3. */ static unsigned int ple_gap = KVM_DEFAULT_PLE_GAP; +module_param(ple_gap, uint, 0444); static unsigned int ple_window = KVM_VMX_DEFAULT_PLE_WINDOW; module_param(ple_window, uint, 0444); From 38ab012f109caf10f471db1adf284e620dd8d701 Mon Sep 17 00:00:00 2001 From: Wanpeng Li Date: Tue, 20 Nov 2018 09:39:30 +0800 Subject: [PATCH 03/15] KVM: LAPIC: Fix pv ipis use-before-initialization MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Reported by syzkaller: BUG: unable to handle kernel NULL pointer dereference at 0000000000000014 PGD 800000040410c067 P4D 800000040410c067 PUD 40410d067 PMD 0 Oops: 0000 [#1] PREEMPT SMP PTI CPU: 3 PID: 2567 Comm: poc Tainted: G OE 4.19.0-rc5 #16 RIP: 0010:kvm_pv_send_ipi+0x94/0x350 [kvm] Call Trace: kvm_emulate_hypercall+0x3cc/0x700 [kvm] handle_vmcall+0xe/0x10 [kvm_intel] vmx_handle_exit+0xc1/0x11b0 [kvm_intel] vcpu_enter_guest+0x9fb/0x1910 [kvm] kvm_arch_vcpu_ioctl_run+0x35c/0x610 [kvm] kvm_vcpu_ioctl+0x3e9/0x6d0 [kvm] do_vfs_ioctl+0xa5/0x690 ksys_ioctl+0x6d/0x80 __x64_sys_ioctl+0x1a/0x20 do_syscall_64+0x83/0x6e0 entry_SYSCALL_64_after_hwframe+0x49/0xbe The reason is that the apic map has not yet been initialized, the testcase triggers pv_send_ipi interface by vmcall which results in kvm->arch.apic_map is dereferenced. This patch fixes it by checking whether or not apic map is NULL and bailing out immediately if that is the case. Fixes: 4180bf1b65 (KVM: X86: Implement "send IPI" hypercall) Reported-by: Wei Wu Cc: Paolo Bonzini Cc: Radim Krčmář Cc: Wei Wu Signed-off-by: Wanpeng Li Cc: stable@vger.kernel.org Signed-off-by: Paolo Bonzini --- arch/x86/kvm/lapic.c | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/arch/x86/kvm/lapic.c b/arch/x86/kvm/lapic.c index 89db20f8cb70..02f2291dcf7e 100644 --- a/arch/x86/kvm/lapic.c +++ b/arch/x86/kvm/lapic.c @@ -576,6 +576,11 @@ int kvm_pv_send_ipi(struct kvm *kvm, unsigned long ipi_bitmap_low, rcu_read_lock(); map = rcu_dereference(kvm->arch.apic_map); + if (unlikely(!map)) { + count = -EOPNOTSUPP; + goto out; + } + if (min > map->max_apic_id) goto out; /* Bits above cluster_size are masked in the caller. */ From e97f852fd4561e77721bb9a4e0ea9d98305b1e93 Mon Sep 17 00:00:00 2001 From: Wanpeng Li Date: Tue, 20 Nov 2018 16:34:18 +0800 Subject: [PATCH 04/15] KVM: X86: Fix scan ioapic use-before-initialization MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Reported by syzkaller: BUG: unable to handle kernel NULL pointer dereference at 00000000000001c8 PGD 80000003ec4da067 P4D 80000003ec4da067 PUD 3f7bfa067 PMD 0 Oops: 0000 [#1] PREEMPT SMP PTI CPU: 7 PID: 5059 Comm: debug Tainted: G OE 4.19.0-rc5 #16 RIP: 0010:__lock_acquire+0x1a6/0x1990 Call Trace: lock_acquire+0xdb/0x210 _raw_spin_lock+0x38/0x70 kvm_ioapic_scan_entry+0x3e/0x110 [kvm] vcpu_enter_guest+0x167e/0x1910 [kvm] kvm_arch_vcpu_ioctl_run+0x35c/0x610 [kvm] kvm_vcpu_ioctl+0x3e9/0x6d0 [kvm] do_vfs_ioctl+0xa5/0x690 ksys_ioctl+0x6d/0x80 __x64_sys_ioctl+0x1a/0x20 do_syscall_64+0x83/0x6e0 entry_SYSCALL_64_after_hwframe+0x49/0xbe The reason is that the testcase writes hyperv synic HV_X64_MSR_SINT6 msr and triggers scan ioapic logic to load synic vectors into EOI exit bitmap. However, irqchip is not initialized by this simple testcase, ioapic/apic objects should not be accessed. This can be triggered by the following program: #define _GNU_SOURCE #include #include #include #include #include #include #include #include uint64_t r[3] = {0xffffffffffffffff, 0xffffffffffffffff, 0xffffffffffffffff}; int main(void) { syscall(__NR_mmap, 0x20000000, 0x1000000, 3, 0x32, -1, 0); long res = 0; memcpy((void*)0x20000040, "/dev/kvm", 9); res = syscall(__NR_openat, 0xffffffffffffff9c, 0x20000040, 0, 0); if (res != -1) r[0] = res; res = syscall(__NR_ioctl, r[0], 0xae01, 0); if (res != -1) r[1] = res; res = syscall(__NR_ioctl, r[1], 0xae41, 0); if (res != -1) r[2] = res; memcpy( (void*)0x20000080, "\x01\x00\x00\x00\x00\x5b\x61\xbb\x96\x00\x00\x40\x00\x00\x00\x00\x01\x00" "\x08\x00\x00\x00\x00\x00\x0b\x77\xd1\x78\x4d\xd8\x3a\xed\xb1\x5c\x2e\x43" "\xaa\x43\x39\xd6\xff\xf5\xf0\xa8\x98\xf2\x3e\x37\x29\x89\xde\x88\xc6\x33" "\xfc\x2a\xdb\xb7\xe1\x4c\xac\x28\x61\x7b\x9c\xa9\xbc\x0d\xa0\x63\xfe\xfe" "\xe8\x75\xde\xdd\x19\x38\xdc\x34\xf5\xec\x05\xfd\xeb\x5d\xed\x2e\xaf\x22" "\xfa\xab\xb7\xe4\x42\x67\xd0\xaf\x06\x1c\x6a\x35\x67\x10\x55\xcb", 106); syscall(__NR_ioctl, r[2], 0x4008ae89, 0x20000080); syscall(__NR_ioctl, r[2], 0xae80, 0); return 0; } This patch fixes it by bailing out scan ioapic if ioapic is not initialized in kernel. Reported-by: Wei Wu Cc: Paolo Bonzini Cc: Radim Krčmář Cc: Wei Wu Signed-off-by: Wanpeng Li Cc: stable@vger.kernel.org Signed-off-by: Paolo Bonzini --- arch/x86/kvm/x86.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c index 5cd5647120f2..64cae03b2c20 100644 --- a/arch/x86/kvm/x86.c +++ b/arch/x86/kvm/x86.c @@ -7455,7 +7455,8 @@ static void vcpu_scan_ioapic(struct kvm_vcpu *vcpu) else { if (vcpu->arch.apicv_active) kvm_x86_ops->sync_pir_to_irr(vcpu); - kvm_ioapic_scan_entry(vcpu, vcpu->arch.ioapic_handled_vectors); + if (ioapic_in_kernel(vcpu->kvm)) + kvm_ioapic_scan_entry(vcpu, vcpu->arch.ioapic_handled_vectors); } if (is_guest_mode(vcpu)) From 30510387a5e45bfcf8190e03ec7aa15b295828e2 Mon Sep 17 00:00:00 2001 From: Wei Wang Date: Mon, 12 Nov 2018 12:23:14 +0000 Subject: [PATCH 05/15] svm: Add mutex_lock to protect apic_access_page_done on AMD systems There is a race condition when accessing kvm->arch.apic_access_page_done. Due to it, x86_set_memory_region will fail when creating the second vcpu for a svm guest. Add a mutex_lock to serialize the accesses to apic_access_page_done. This lock is also used by vmx for the same purpose. Signed-off-by: Wei Wang Signed-off-by: Amadeusz Juskowiak Signed-off-by: Julian Stecklina Signed-off-by: Suravee Suthikulpanit Reviewed-by: Joerg Roedel Cc: stable@vger.kernel.org Signed-off-by: Paolo Bonzini --- arch/x86/kvm/svm.c | 19 +++++++++++-------- 1 file changed, 11 insertions(+), 8 deletions(-) diff --git a/arch/x86/kvm/svm.c b/arch/x86/kvm/svm.c index 0e21ccc46792..033ec01661b9 100644 --- a/arch/x86/kvm/svm.c +++ b/arch/x86/kvm/svm.c @@ -1664,20 +1664,23 @@ static u64 *avic_get_physical_id_entry(struct kvm_vcpu *vcpu, static int avic_init_access_page(struct kvm_vcpu *vcpu) { struct kvm *kvm = vcpu->kvm; - int ret; + int ret = 0; + mutex_lock(&kvm->slots_lock); if (kvm->arch.apic_access_page_done) - return 0; + goto out; - ret = x86_set_memory_region(kvm, - APIC_ACCESS_PAGE_PRIVATE_MEMSLOT, - APIC_DEFAULT_PHYS_BASE, - PAGE_SIZE); + ret = __x86_set_memory_region(kvm, + APIC_ACCESS_PAGE_PRIVATE_MEMSLOT, + APIC_DEFAULT_PHYS_BASE, + PAGE_SIZE); if (ret) - return ret; + goto out; kvm->arch.apic_access_page_done = true; - return 0; +out: + mutex_unlock(&kvm->slots_lock); + return ret; } static int avic_init_backing_page(struct kvm_vcpu *vcpu) From 7f9ad1dfa3c768d1116c2dbacd7a09f9a871534e Mon Sep 17 00:00:00 2001 From: Liran Alon Date: Sat, 17 Nov 2018 14:05:06 +0200 Subject: [PATCH 06/15] KVM: nVMX: Fix kernel info-leak when enabling KVM_CAP_HYPERV_ENLIGHTENED_VMCS more than once Consider the case that userspace enables KVM_CAP_HYPERV_ENLIGHTENED_VMCS twice: 1) kvm_vcpu_ioctl_enable_cap() is called to enable KVM_CAP_HYPERV_ENLIGHTENED_VMCS which calls nested_enable_evmcs(). 2) nested_enable_evmcs() sets enlightened_vmcs_enabled to true and fills vmcs_version which is then copied to userspace. 3) kvm_vcpu_ioctl_enable_cap() is called again to enable KVM_CAP_HYPERV_ENLIGHTENED_VMCS which calls nested_enable_evmcs(). 4) This time nested_enable_evmcs() just returns 0 as enlightened_vmcs_enabled is already true. *Without filling vmcs_version*. 5) kvm_vcpu_ioctl_enable_cap() continues as usual and copies *uninitialized* vmcs_version to userspace which leads to kernel info-leak. Fix this issue by simply changing nested_enable_evmcs() to always fill vmcs_version output argument. Even when enlightened_vmcs_enabled is already set to true. Note that SVM's nested_enable_evmcs() should not be modified because it always returns a non-zero value (-ENODEV) which results in kvm_vcpu_ioctl_enable_cap() skipping the copy of vmcs_version to userspace (as it should). Fixes: 57b119da3594 ("KVM: nVMX: add KVM_CAP_HYPERV_ENLIGHTENED_VMCS capability") Reported-by: syzbot+cfbc368e283d381f8cef@syzkaller.appspotmail.com Reviewed-by: Krish Sadhukhan Reviewed-by: Vitaly Kuznetsov Signed-off-by: Liran Alon Signed-off-by: Paolo Bonzini --- arch/x86/kvm/vmx.c | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c index be6f13f1c25f..065f1df74000 100644 --- a/arch/x86/kvm/vmx.c +++ b/arch/x86/kvm/vmx.c @@ -1611,12 +1611,6 @@ static int nested_enable_evmcs(struct kvm_vcpu *vcpu, { struct vcpu_vmx *vmx = to_vmx(vcpu); - /* We don't support disabling the feature for simplicity. */ - if (vmx->nested.enlightened_vmcs_enabled) - return 0; - - vmx->nested.enlightened_vmcs_enabled = true; - /* * vmcs_version represents the range of supported Enlightened VMCS * versions: lower 8 bits is the minimal version, higher 8 bits is the @@ -1626,6 +1620,12 @@ static int nested_enable_evmcs(struct kvm_vcpu *vcpu, if (vmcs_version) *vmcs_version = (KVM_EVMCS_VERSION << 8) | 1; + /* We don't support disabling the feature for simplicity. */ + if (vmx->nested.enlightened_vmcs_enabled) + return 0; + + vmx->nested.enlightened_vmcs_enabled = true; + vmx->nested.msrs.pinbased_ctls_high &= ~EVMCS1_UNSUPPORTED_PINCTRL; vmx->nested.msrs.entry_ctls_high &= ~EVMCS1_UNSUPPORTED_VMENTRY_CTRL; vmx->nested.msrs.exit_ctls_high &= ~EVMCS1_UNSUPPORTED_VMEXIT_CTRL; From bcbfbd8ec21096027f1ee13ce6c185e8175166f6 Mon Sep 17 00:00:00 2001 From: Liran Alon Date: Thu, 8 Nov 2018 00:43:06 +0200 Subject: [PATCH 07/15] KVM: x86: Fix kernel info-leak in KVM_HC_CLOCK_PAIRING hypercall kvm_pv_clock_pairing() allocates local var "struct kvm_clock_pairing clock_pairing" on stack and initializes all it's fields besides padding (clock_pairing.pad[]). Because clock_pairing var is written completely (including padding) to guest memory, failure to init struct padding results in kernel info-leak. Fix the issue by making sure to also init the padding with zeroes. Fixes: 55dd00a73a51 ("KVM: x86: add KVM_HC_CLOCK_PAIRING hypercall") Reported-by: syzbot+a8ef68d71211ba264f56@syzkaller.appspotmail.com Reviewed-by: Mark Kanda Signed-off-by: Liran Alon Cc: stable@vger.kernel.org Signed-off-by: Paolo Bonzini --- arch/x86/kvm/x86.c | 1 + 1 file changed, 1 insertion(+) diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c index 64cae03b2c20..7e4be1f2f253 100644 --- a/arch/x86/kvm/x86.c +++ b/arch/x86/kvm/x86.c @@ -6918,6 +6918,7 @@ static int kvm_pv_clock_pairing(struct kvm_vcpu *vcpu, gpa_t paddr, clock_pairing.nsec = ts.tv_nsec; clock_pairing.tsc = kvm_read_l1_tsc(vcpu, cycle); clock_pairing.flags = 0; + memset(&clock_pairing.pad, 0, sizeof(clock_pairing.pad)); ret = 0; if (kvm_write_guest(vcpu->kvm, paddr, &clock_pairing, From f48b4711dd6e1cf282f9dfd159c14a305909c97c Mon Sep 17 00:00:00 2001 From: Liran Alon Date: Tue, 20 Nov 2018 18:03:25 +0200 Subject: [PATCH 08/15] KVM: VMX: Update shared MSRs to be saved/restored on MSR_EFER.LMA changes When guest transitions from/to long-mode by modifying MSR_EFER.LMA, the list of shared MSRs to be saved/restored on guest<->host transitions is updated (See vmx_set_efer() call to setup_msrs()). On every entry to guest, vcpu_enter_guest() calls vmx_prepare_switch_to_guest(). This function should also take care of setting the shared MSRs to be saved/restored. However, the function does nothing in case we are already running with loaded guest state (vmx->loaded_cpu_state != NULL). This means that even when guest modifies MSR_EFER.LMA which results in updating the list of shared MSRs, it isn't being taken into account by vmx_prepare_switch_to_guest() because it happens while we are running with loaded guest state. To fix above mentioned issue, add a flag to mark that the list of shared MSRs has been updated and modify vmx_prepare_switch_to_guest() to set shared MSRs when running with host state *OR* list of shared MSRs has been updated. Note that this issue was mistakenly introduced by commit 678e315e78a7 ("KVM: vmx: add dedicated utility to access guest's kernel_gs_base") because previously vmx_set_efer() always called vmx_load_host_state() which resulted in vmx_prepare_switch_to_guest() to set shared MSRs. Fixes: 678e315e78a7 ("KVM: vmx: add dedicated utility to access guest's kernel_gs_base") Reported-by: Eyal Moscovici Reviewed-by: Mihai Carabas Reviewed-by: Liam Merwick Reviewed-by: Jim Mattson Signed-off-by: Liran Alon Signed-off-by: Paolo Bonzini --- arch/x86/kvm/vmx.c | 21 ++++++++++++++++----- 1 file changed, 16 insertions(+), 5 deletions(-) diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c index 065f1df74000..d09d67310012 100644 --- a/arch/x86/kvm/vmx.c +++ b/arch/x86/kvm/vmx.c @@ -985,6 +985,7 @@ struct vcpu_vmx { struct shared_msr_entry *guest_msrs; int nmsrs; int save_nmsrs; + bool guest_msrs_dirty; unsigned long host_idt_base; #ifdef CONFIG_X86_64 u64 msr_host_kernel_gs_base; @@ -2898,6 +2899,20 @@ static void vmx_prepare_switch_to_guest(struct kvm_vcpu *vcpu) vmx->req_immediate_exit = false; + /* + * Note that guest MSRs to be saved/restored can also be changed + * when guest state is loaded. This happens when guest transitions + * to/from long-mode by setting MSR_EFER.LMA. + */ + if (!vmx->loaded_cpu_state || vmx->guest_msrs_dirty) { + vmx->guest_msrs_dirty = false; + for (i = 0; i < vmx->save_nmsrs; ++i) + kvm_set_shared_msr(vmx->guest_msrs[i].index, + vmx->guest_msrs[i].data, + vmx->guest_msrs[i].mask); + + } + if (vmx->loaded_cpu_state) return; @@ -2958,11 +2973,6 @@ static void vmx_prepare_switch_to_guest(struct kvm_vcpu *vcpu) vmcs_writel(HOST_GS_BASE, gs_base); host_state->gs_base = gs_base; } - - for (i = 0; i < vmx->save_nmsrs; ++i) - kvm_set_shared_msr(vmx->guest_msrs[i].index, - vmx->guest_msrs[i].data, - vmx->guest_msrs[i].mask); } static void vmx_prepare_switch_to_host(struct vcpu_vmx *vmx) @@ -3437,6 +3447,7 @@ static void setup_msrs(struct vcpu_vmx *vmx) move_msr_up(vmx, index, save_nmsrs++); vmx->save_nmsrs = save_nmsrs; + vmx->guest_msrs_dirty = true; if (cpu_has_vmx_msr_bitmap()) vmx_update_msr_bitmap(&vmx->vcpu); From 354cb410d87314e2eda344feea84809e4261570a Mon Sep 17 00:00:00 2001 From: Yi Wang Date: Thu, 8 Nov 2018 16:48:36 +0800 Subject: [PATCH 09/15] KVM: x86: fix empty-body warnings MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit We get the following warnings about empty statements when building with 'W=1': arch/x86/kvm/lapic.c:632:53: warning: suggest braces around empty body in an ‘if’ statement [-Wempty-body] arch/x86/kvm/lapic.c:1907:42: warning: suggest braces around empty body in an ‘if’ statement [-Wempty-body] arch/x86/kvm/lapic.c:1936:65: warning: suggest braces around empty body in an ‘if’ statement [-Wempty-body] arch/x86/kvm/lapic.c:1975:44: warning: suggest braces around empty body in an ‘if’ statement [-Wempty-body] Rework the debug helper macro to get rid of these warnings. Signed-off-by: Yi Wang Signed-off-by: Paolo Bonzini --- arch/x86/kvm/lapic.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/arch/x86/kvm/lapic.c b/arch/x86/kvm/lapic.c index 02f2291dcf7e..c4533d05c214 100644 --- a/arch/x86/kvm/lapic.c +++ b/arch/x86/kvm/lapic.c @@ -55,7 +55,7 @@ #define PRIo64 "o" /* #define apic_debug(fmt,arg...) printk(KERN_WARNING fmt,##arg) */ -#define apic_debug(fmt, arg...) +#define apic_debug(fmt, arg...) do {} while (0) /* 14 is the version for Xeon and Pentium 8.4.8*/ #define APIC_VERSION (0x14UL | ((KVM_APIC_LVT_NUM - 1) << 16)) From 1e4329ee2c52692ea42cc677fb2133519718b34a Mon Sep 17 00:00:00 2001 From: Yi Wang Date: Thu, 8 Nov 2018 11:22:21 +0800 Subject: [PATCH 10/15] x86/kvm/vmx: fix old-style function declaration MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The inline keyword which is not at the beginning of the function declaration may trigger the following build warnings, so let's fix it: arch/x86/kvm/vmx.c:1309:1: warning: ‘inline’ is not at beginning of declaration [-Wold-style-declaration] arch/x86/kvm/vmx.c:5947:1: warning: ‘inline’ is not at beginning of declaration [-Wold-style-declaration] arch/x86/kvm/vmx.c:5985:1: warning: ‘inline’ is not at beginning of declaration [-Wold-style-declaration] arch/x86/kvm/vmx.c:6023:1: warning: ‘inline’ is not at beginning of declaration [-Wold-style-declaration] Signed-off-by: Yi Wang Signed-off-by: Paolo Bonzini --- arch/x86/kvm/vmx.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c index d09d67310012..5f43fcfc225b 100644 --- a/arch/x86/kvm/vmx.c +++ b/arch/x86/kvm/vmx.c @@ -1308,7 +1308,7 @@ static void vmx_set_nmi_mask(struct kvm_vcpu *vcpu, bool masked); static bool nested_vmx_is_page_fault_vmexit(struct vmcs12 *vmcs12, u16 error_code); static void vmx_update_msr_bitmap(struct kvm_vcpu *vcpu); -static void __always_inline vmx_disable_intercept_for_msr(unsigned long *msr_bitmap, +static __always_inline void vmx_disable_intercept_for_msr(unsigned long *msr_bitmap, u32 msr, int type); static DEFINE_PER_CPU(struct vmcs *, vmxarea); @@ -5956,7 +5956,7 @@ static void free_vpid(int vpid) spin_unlock(&vmx_vpid_lock); } -static void __always_inline vmx_disable_intercept_for_msr(unsigned long *msr_bitmap, +static __always_inline void vmx_disable_intercept_for_msr(unsigned long *msr_bitmap, u32 msr, int type) { int f = sizeof(unsigned long); @@ -5994,7 +5994,7 @@ static void __always_inline vmx_disable_intercept_for_msr(unsigned long *msr_bit } } -static void __always_inline vmx_enable_intercept_for_msr(unsigned long *msr_bitmap, +static __always_inline void vmx_enable_intercept_for_msr(unsigned long *msr_bitmap, u32 msr, int type) { int f = sizeof(unsigned long); @@ -6032,7 +6032,7 @@ static void __always_inline vmx_enable_intercept_for_msr(unsigned long *msr_bitm } } -static void __always_inline vmx_set_intercept_for_msr(unsigned long *msr_bitmap, +static __always_inline void vmx_set_intercept_for_msr(unsigned long *msr_bitmap, u32 msr, int type, bool value) { if (value) From 326e742533bf0a23f0127d8ea62fb558ba665f08 Mon Sep 17 00:00:00 2001 From: Leonid Shatz Date: Tue, 6 Nov 2018 12:14:25 +0200 Subject: [PATCH 11/15] KVM: nVMX/nSVM: Fix bug which sets vcpu->arch.tsc_offset to L1 tsc_offset Since commit e79f245ddec1 ("X86/KVM: Properly update 'tsc_offset' to represent the running guest"), vcpu->arch.tsc_offset meaning was changed to always reflect the tsc_offset value set on active VMCS. Regardless if vCPU is currently running L1 or L2. However, above mentioned commit failed to also change kvm_vcpu_write_tsc_offset() to set vcpu->arch.tsc_offset correctly. This is because vmx_write_tsc_offset() could set the tsc_offset value in active VMCS to given offset parameter *plus vmcs12->tsc_offset*. However, kvm_vcpu_write_tsc_offset() just sets vcpu->arch.tsc_offset to given offset parameter. Without taking into account the possible addition of vmcs12->tsc_offset. (Same is true for SVM case). Fix this issue by changing kvm_x86_ops->write_tsc_offset() to return actually set tsc_offset in active VMCS and modify kvm_vcpu_write_tsc_offset() to set returned value in vcpu->arch.tsc_offset. In addition, rename write_tsc_offset() callback to write_l1_tsc_offset() to make it clear that it is meant to set L1 TSC offset. Fixes: e79f245ddec1 ("X86/KVM: Properly update 'tsc_offset' to represent the running guest") Reviewed-by: Liran Alon Reviewed-by: Mihai Carabas Reviewed-by: Krish Sadhukhan Signed-off-by: Leonid Shatz Cc: stable@vger.kernel.org Signed-off-by: Paolo Bonzini --- arch/x86/include/asm/kvm_host.h | 3 ++- arch/x86/kvm/svm.c | 5 +++-- arch/x86/kvm/vmx.c | 21 +++++++++------------ arch/x86/kvm/x86.c | 6 +++--- 4 files changed, 17 insertions(+), 18 deletions(-) diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h index 55e51ff7e421..fbda5a917c5b 100644 --- a/arch/x86/include/asm/kvm_host.h +++ b/arch/x86/include/asm/kvm_host.h @@ -1094,7 +1094,8 @@ struct kvm_x86_ops { bool (*has_wbinvd_exit)(void); u64 (*read_l1_tsc_offset)(struct kvm_vcpu *vcpu); - void (*write_tsc_offset)(struct kvm_vcpu *vcpu, u64 offset); + /* Returns actual tsc_offset set in active VMCS */ + u64 (*write_l1_tsc_offset)(struct kvm_vcpu *vcpu, u64 offset); void (*get_exit_info)(struct kvm_vcpu *vcpu, u64 *info1, u64 *info2); diff --git a/arch/x86/kvm/svm.c b/arch/x86/kvm/svm.c index 033ec01661b9..a24733aade4c 100644 --- a/arch/x86/kvm/svm.c +++ b/arch/x86/kvm/svm.c @@ -1446,7 +1446,7 @@ static u64 svm_read_l1_tsc_offset(struct kvm_vcpu *vcpu) return vcpu->arch.tsc_offset; } -static void svm_write_tsc_offset(struct kvm_vcpu *vcpu, u64 offset) +static u64 svm_write_l1_tsc_offset(struct kvm_vcpu *vcpu, u64 offset) { struct vcpu_svm *svm = to_svm(vcpu); u64 g_tsc_offset = 0; @@ -1464,6 +1464,7 @@ static void svm_write_tsc_offset(struct kvm_vcpu *vcpu, u64 offset) svm->vmcb->control.tsc_offset = offset + g_tsc_offset; mark_dirty(svm->vmcb, VMCB_INTERCEPTS); + return svm->vmcb->control.tsc_offset; } static void avic_init_vmcb(struct vcpu_svm *svm) @@ -7152,7 +7153,7 @@ static struct kvm_x86_ops svm_x86_ops __ro_after_init = { .has_wbinvd_exit = svm_has_wbinvd_exit, .read_l1_tsc_offset = svm_read_l1_tsc_offset, - .write_tsc_offset = svm_write_tsc_offset, + .write_l1_tsc_offset = svm_write_l1_tsc_offset, .set_tdp_cr3 = set_tdp_cr3, diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c index 5f43fcfc225b..764c23dc444f 100644 --- a/arch/x86/kvm/vmx.c +++ b/arch/x86/kvm/vmx.c @@ -3464,11 +3464,9 @@ static u64 vmx_read_l1_tsc_offset(struct kvm_vcpu *vcpu) return vcpu->arch.tsc_offset; } -/* - * writes 'offset' into guest's timestamp counter offset register - */ -static void vmx_write_tsc_offset(struct kvm_vcpu *vcpu, u64 offset) +static u64 vmx_write_l1_tsc_offset(struct kvm_vcpu *vcpu, u64 offset) { + u64 active_offset = offset; if (is_guest_mode(vcpu)) { /* * We're here if L1 chose not to trap WRMSR to TSC. According @@ -3476,17 +3474,16 @@ static void vmx_write_tsc_offset(struct kvm_vcpu *vcpu, u64 offset) * set for L2 remains unchanged, and still needs to be added * to the newly set TSC to get L2's TSC. */ - struct vmcs12 *vmcs12; - /* recalculate vmcs02.TSC_OFFSET: */ - vmcs12 = get_vmcs12(vcpu); - vmcs_write64(TSC_OFFSET, offset + - (nested_cpu_has(vmcs12, CPU_BASED_USE_TSC_OFFSETING) ? - vmcs12->tsc_offset : 0)); + struct vmcs12 *vmcs12 = get_vmcs12(vcpu); + if (nested_cpu_has(vmcs12, CPU_BASED_USE_TSC_OFFSETING)) + active_offset += vmcs12->tsc_offset; } else { trace_kvm_write_tsc_offset(vcpu->vcpu_id, vmcs_read64(TSC_OFFSET), offset); - vmcs_write64(TSC_OFFSET, offset); } + + vmcs_write64(TSC_OFFSET, active_offset); + return active_offset; } /* @@ -15074,7 +15071,7 @@ static struct kvm_x86_ops vmx_x86_ops __ro_after_init = { .has_wbinvd_exit = cpu_has_vmx_wbinvd_exit, .read_l1_tsc_offset = vmx_read_l1_tsc_offset, - .write_tsc_offset = vmx_write_tsc_offset, + .write_l1_tsc_offset = vmx_write_l1_tsc_offset, .set_tdp_cr3 = vmx_set_cr3, diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c index 7e4be1f2f253..d02937760c3b 100644 --- a/arch/x86/kvm/x86.c +++ b/arch/x86/kvm/x86.c @@ -1665,8 +1665,7 @@ EXPORT_SYMBOL_GPL(kvm_read_l1_tsc); static void kvm_vcpu_write_tsc_offset(struct kvm_vcpu *vcpu, u64 offset) { - kvm_x86_ops->write_tsc_offset(vcpu, offset); - vcpu->arch.tsc_offset = offset; + vcpu->arch.tsc_offset = kvm_x86_ops->write_l1_tsc_offset(vcpu, offset); } static inline bool kvm_check_tsc_unstable(void) @@ -1794,7 +1793,8 @@ EXPORT_SYMBOL_GPL(kvm_write_tsc); static inline void adjust_tsc_offset_guest(struct kvm_vcpu *vcpu, s64 adjustment) { - kvm_vcpu_write_tsc_offset(vcpu, vcpu->arch.tsc_offset + adjustment); + u64 tsc_offset = kvm_x86_ops->read_l1_tsc_offset(vcpu); + kvm_vcpu_write_tsc_offset(vcpu, tsc_offset + adjustment); } static inline void adjust_tsc_offset_host(struct kvm_vcpu *vcpu, s64 adjustment) From 72aeb60c52bf74a0eeec77d6b41ce40145697d76 Mon Sep 17 00:00:00 2001 From: Liran Alon Date: Thu, 1 Nov 2018 10:57:39 +0200 Subject: [PATCH 12/15] KVM: nVMX: Verify eVMCS revision id match supported eVMCS version on eVMCS VMPTRLD According to TLFS section 16.11.2 Enlightened VMCS, the first u32 field of eVMCS should specify eVMCS VersionNumber. This version should be in the range of supported eVMCS versions exposed to guest via CPUID.0x4000000A.EAX[0:15]. The range which KVM expose to guest in this CPUID field should be the same as the value returned in vmcs_version by nested_enable_evmcs(). According to the above, eVMCS VMPTRLD should verify that version specified in given eVMCS is in the supported range. However, current code mistakenly verfies this field against VMCS12_REVISION. One can also see that when KVM use eVMCS, it makes sure that alloc_vmcs_cpu() sets allocated eVMCS revision_id to KVM_EVMCS_VERSION. Obvious fix should just change eVMCS VMPTRLD to verify first u32 field of eVMCS is equal to KVM_EVMCS_VERSION. However, it turns out that Microsoft Hyper-V fails to comply to their own invented interface: When Hyper-V use eVMCS, it just sets first u32 field of eVMCS to revision_id specified in MSR_IA32_VMX_BASIC (In our case: VMCS12_REVISION). Instead of used eVMCS version number which is one of the supported versions specified in CPUID.0x4000000A.EAX[0:15]. To overcome Hyper-V bug, we accept either a supported eVMCS version or VMCS12_REVISION as valid values for first u32 field of eVMCS. Cc: Vitaly Kuznetsov Reviewed-by: Nikita Leshenko Reviewed-by: Mark Kanda Signed-off-by: Liran Alon Reviewed-by: Vitaly Kuznetsov Signed-off-by: Paolo Bonzini --- arch/x86/kvm/vmx.c | 25 ++++++++++++++++++++++++- 1 file changed, 24 insertions(+), 1 deletion(-) diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c index 764c23dc444f..0ffd8b2dbfe2 100644 --- a/arch/x86/kvm/vmx.c +++ b/arch/x86/kvm/vmx.c @@ -9378,7 +9378,30 @@ static int nested_vmx_handle_enlightened_vmptrld(struct kvm_vcpu *vcpu, vmx->nested.hv_evmcs = kmap(vmx->nested.hv_evmcs_page); - if (vmx->nested.hv_evmcs->revision_id != VMCS12_REVISION) { + /* + * Currently, KVM only supports eVMCS version 1 + * (== KVM_EVMCS_VERSION) and thus we expect guest to set this + * value to first u32 field of eVMCS which should specify eVMCS + * VersionNumber. + * + * Guest should be aware of supported eVMCS versions by host by + * examining CPUID.0x4000000A.EAX[0:15]. Host userspace VMM is + * expected to set this CPUID leaf according to the value + * returned in vmcs_version from nested_enable_evmcs(). + * + * However, it turns out that Microsoft Hyper-V fails to comply + * to their own invented interface: When Hyper-V use eVMCS, it + * just sets first u32 field of eVMCS to revision_id specified + * in MSR_IA32_VMX_BASIC. Instead of used eVMCS version number + * which is one of the supported versions specified in + * CPUID.0x4000000A.EAX[0:15]. + * + * To overcome Hyper-V bug, we accept here either a supported + * eVMCS version or VMCS12 revision_id as valid values for first + * u32 field of eVMCS. + */ + if ((vmx->nested.hv_evmcs->revision_id != KVM_EVMCS_VERSION) && + (vmx->nested.hv_evmcs->revision_id != VMCS12_REVISION)) { nested_release_evmcs(vcpu); return 0; } From 52ad7eb3d668867f9ee2e34d715cfb4860d36a93 Mon Sep 17 00:00:00 2001 From: Liran Alon Date: Tue, 13 Nov 2018 17:44:46 +0200 Subject: [PATCH 13/15] KVM: nVMX: vmcs12 revision_id is always VMCS12_REVISION even when copied from eVMCS vmcs12 represents the per-CPU cache of L1 active vmcs12. This cache can be loaded by one of the following: 1) Guest making a vmcs12 active by exeucting VMPTRLD 2) Guest specifying eVMCS in VP assist page and executing VMLAUNCH/VMRESUME. Either way, vmcs12 should have revision_id of VMCS12_REVISION. Which is not equal to eVMCS revision_id which specifies used VersionNumber of eVMCS struct (e.g. KVM_EVMCS_VERSION). Specifically, this causes an issue in restoring a nested VM state because vmx_set_nested_state() verifies that vmcs12->revision_id is equal to VMCS12_REVISION which was not true in case vmcs12 was populated from an eVMCS by vmx_get_nested_state() which calls copy_enlightened_to_vmcs12(). Reviewed-by: Darren Kenny Signed-off-by: Liran Alon Reviewed-by: Vitaly Kuznetsov Signed-off-by: Paolo Bonzini --- arch/x86/kvm/vmx.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c index 0ffd8b2dbfe2..02edd9960e9d 100644 --- a/arch/x86/kvm/vmx.c +++ b/arch/x86/kvm/vmx.c @@ -8673,8 +8673,6 @@ static int copy_enlightened_to_vmcs12(struct vcpu_vmx *vmx) struct vmcs12 *vmcs12 = vmx->nested.cached_vmcs12; struct hv_enlightened_vmcs *evmcs = vmx->nested.hv_evmcs; - vmcs12->hdr.revision_id = evmcs->revision_id; - /* HV_VMX_ENLIGHTENED_CLEAN_FIELD_NONE */ vmcs12->tpr_threshold = evmcs->tpr_threshold; vmcs12->guest_rip = evmcs->guest_rip; @@ -9422,9 +9420,11 @@ static int nested_vmx_handle_enlightened_vmptrld(struct kvm_vcpu *vcpu, * present in struct hv_enlightened_vmcs, ...). Make sure there * are no leftovers. */ - if (from_launch) - memset(vmx->nested.cached_vmcs12, 0, - sizeof(*vmx->nested.cached_vmcs12)); + if (from_launch) { + struct vmcs12 *vmcs12 = get_vmcs12(vcpu); + memset(vmcs12, 0, sizeof(*vmcs12)); + vmcs12->hdr.revision_id = VMCS12_REVISION; + } } return 1; From 0e0fee5c539b61fdd098332e0e2cc375d9073706 Mon Sep 17 00:00:00 2001 From: Junaid Shahid Date: Wed, 31 Oct 2018 14:53:57 -0700 Subject: [PATCH 14/15] kvm: mmu: Fix race in emulated page table writes When a guest page table is updated via an emulated write, kvm_mmu_pte_write() is called to update the shadow PTE using the just written guest PTE value. But if two emulated guest PTE writes happened concurrently, it is possible that the guest PTE and the shadow PTE end up being out of sync. Emulated writes do not mark the shadow page as unsync-ed, so this inconsistency will not be resolved even by a guest TLB flush (unless the page was marked as unsync-ed at some other point). This is fixed by re-reading the current value of the guest PTE after the MMU lock has been acquired instead of just using the value that was written prior to calling kvm_mmu_pte_write(). Signed-off-by: Junaid Shahid Reviewed-by: Wanpeng Li Cc: stable@vger.kernel.org Signed-off-by: Paolo Bonzini --- arch/x86/kvm/mmu.c | 27 +++++++++------------------ 1 file changed, 9 insertions(+), 18 deletions(-) diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c index cf5f572f2305..7c03c0f35444 100644 --- a/arch/x86/kvm/mmu.c +++ b/arch/x86/kvm/mmu.c @@ -5074,9 +5074,9 @@ static bool need_remote_flush(u64 old, u64 new) } static u64 mmu_pte_write_fetch_gpte(struct kvm_vcpu *vcpu, gpa_t *gpa, - const u8 *new, int *bytes) + int *bytes) { - u64 gentry; + u64 gentry = 0; int r; /* @@ -5088,22 +5088,12 @@ static u64 mmu_pte_write_fetch_gpte(struct kvm_vcpu *vcpu, gpa_t *gpa, /* Handle a 32-bit guest writing two halves of a 64-bit gpte */ *gpa &= ~(gpa_t)7; *bytes = 8; - r = kvm_vcpu_read_guest(vcpu, *gpa, &gentry, 8); - if (r) - gentry = 0; - new = (const u8 *)&gentry; } - switch (*bytes) { - case 4: - gentry = *(const u32 *)new; - break; - case 8: - gentry = *(const u64 *)new; - break; - default: - gentry = 0; - break; + if (*bytes == 4 || *bytes == 8) { + r = kvm_vcpu_read_guest_atomic(vcpu, *gpa, &gentry, *bytes); + if (r) + gentry = 0; } return gentry; @@ -5207,8 +5197,6 @@ static void kvm_mmu_pte_write(struct kvm_vcpu *vcpu, gpa_t gpa, pgprintk("%s: gpa %llx bytes %d\n", __func__, gpa, bytes); - gentry = mmu_pte_write_fetch_gpte(vcpu, &gpa, new, &bytes); - /* * No need to care whether allocation memory is successful * or not since pte prefetch is skiped if it does not have @@ -5217,6 +5205,9 @@ static void kvm_mmu_pte_write(struct kvm_vcpu *vcpu, gpa_t gpa, mmu_topup_memory_caches(vcpu); spin_lock(&vcpu->kvm->mmu_lock); + + gentry = mmu_pte_write_fetch_gpte(vcpu, &gpa, &bytes); + ++vcpu->kvm->stat.mmu_pte_write; kvm_mmu_audit(vcpu, AUDIT_PRE_PTE_WRITE); From fd65d3142f734bc4376053c8d75670041903134d Mon Sep 17 00:00:00 2001 From: Jim Mattson Date: Tue, 22 May 2018 09:54:20 -0700 Subject: [PATCH 15/15] kvm: svm: Ensure an IBPB on all affected CPUs when freeing a vmcb Previously, we only called indirect_branch_prediction_barrier on the logical CPU that freed a vmcb. This function should be called on all logical CPUs that last loaded the vmcb in question. Fixes: 15d45071523d ("KVM/x86: Add IBPB support") Reported-by: Neel Natu Signed-off-by: Jim Mattson Reviewed-by: Konrad Rzeszutek Wilk Cc: stable@vger.kernel.org Signed-off-by: Paolo Bonzini --- arch/x86/kvm/svm.c | 20 +++++++++++++++----- 1 file changed, 15 insertions(+), 5 deletions(-) diff --git a/arch/x86/kvm/svm.c b/arch/x86/kvm/svm.c index a24733aade4c..cc6467b35a85 100644 --- a/arch/x86/kvm/svm.c +++ b/arch/x86/kvm/svm.c @@ -2193,21 +2193,31 @@ out: return ERR_PTR(err); } +static void svm_clear_current_vmcb(struct vmcb *vmcb) +{ + int i; + + for_each_online_cpu(i) + cmpxchg(&per_cpu(svm_data, i)->current_vmcb, vmcb, NULL); +} + static void svm_free_vcpu(struct kvm_vcpu *vcpu) { struct vcpu_svm *svm = to_svm(vcpu); + /* + * The vmcb page can be recycled, causing a false negative in + * svm_vcpu_load(). So, ensure that no logical CPU has this + * vmcb page recorded as its current vmcb. + */ + svm_clear_current_vmcb(svm->vmcb); + __free_page(pfn_to_page(__sme_clr(svm->vmcb_pa) >> PAGE_SHIFT)); __free_pages(virt_to_page(svm->msrpm), MSRPM_ALLOC_ORDER); __free_page(virt_to_page(svm->nested.hsave)); __free_pages(virt_to_page(svm->nested.msrpm), MSRPM_ALLOC_ORDER); kvm_vcpu_uninit(vcpu); kmem_cache_free(kvm_vcpu_cache, svm); - /* - * The vmcb page can be recycled, causing a false negative in - * svm_vcpu_load(). So do a full IBPB now. - */ - indirect_branch_prediction_barrier(); } static void svm_vcpu_load(struct kvm_vcpu *vcpu, int cpu)