From 0cb5b30698fdc8f6b4646012e3acb4ddce430788 Mon Sep 17 00:00:00 2001 From: Jim Mattson Date: Wed, 3 Jan 2018 14:31:38 -0800 Subject: kvm: vmx: Scrub hardware GPRs at VM-exit Guest GPR values are live in the hardware GPRs at VM-exit. Do not leave any guest values in hardware GPRs after the guest GPR values are saved to the vcpu_vmx structure. This is a partial mitigation for CVE 2017-5715 and CVE 2017-5753. Specifically, it defeats the Project Zero PoC for CVE 2017-5715. Suggested-by: Eric Northup Signed-off-by: Jim Mattson Reviewed-by: Eric Northup Reviewed-by: Benjamin Serebrin Reviewed-by: Andrew Honig [Paolo: Add AMD bits, Signed-off-by: Tom Lendacky ] Signed-off-by: Paolo Bonzini --- arch/x86/kvm/vmx.c | 14 +++++++++++++- 1 file changed, 13 insertions(+), 1 deletion(-) (limited to 'arch/x86/kvm/vmx.c') diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c index 8eba631c4dbd..c1e7ed371259 100644 --- a/arch/x86/kvm/vmx.c +++ b/arch/x86/kvm/vmx.c @@ -9415,6 +9415,7 @@ static void __noclone vmx_vcpu_run(struct kvm_vcpu *vcpu) /* Save guest registers, load host registers, keep flags */ "mov %0, %c[wordsize](%%" _ASM_SP ") \n\t" "pop %0 \n\t" + "setbe %c[fail](%0)\n\t" "mov %%" _ASM_AX ", %c[rax](%0) \n\t" "mov %%" _ASM_BX ", %c[rbx](%0) \n\t" __ASM_SIZE(pop) " %c[rcx](%0) \n\t" @@ -9431,12 +9432,23 @@ static void __noclone vmx_vcpu_run(struct kvm_vcpu *vcpu) "mov %%r13, %c[r13](%0) \n\t" "mov %%r14, %c[r14](%0) \n\t" "mov %%r15, %c[r15](%0) \n\t" + "xor %%r8d, %%r8d \n\t" + "xor %%r9d, %%r9d \n\t" + "xor %%r10d, %%r10d \n\t" + "xor %%r11d, %%r11d \n\t" + "xor %%r12d, %%r12d \n\t" + "xor %%r13d, %%r13d \n\t" + "xor %%r14d, %%r14d \n\t" + "xor %%r15d, %%r15d \n\t" #endif "mov %%cr2, %%" _ASM_AX " \n\t" "mov %%" _ASM_AX ", %c[cr2](%0) \n\t" + "xor %%eax, %%eax \n\t" + "xor %%ebx, %%ebx \n\t" + "xor %%esi, %%esi \n\t" + "xor %%edi, %%edi \n\t" "pop %%" _ASM_BP "; pop %%" _ASM_DX " \n\t" - "setbe %c[fail](%0) \n\t" ".pushsection .rodata \n\t" ".global vmx_return \n\t" "vmx_return: " _ASM_PTR " 2b \n\t" -- cgit v1.2.3-59-g8ed1b From bd89525a823ce6edddcedbe9aed79faa1b9cf544 Mon Sep 17 00:00:00 2001 From: Paolo Bonzini Date: Thu, 11 Jan 2018 16:55:24 +0100 Subject: KVM: x86: emulate #UD while in guest mode This reverts commits ae1f57670703656cc9f293722c3b8b6782f8ab3f and ac9b305caa0df6f5b75d294e4b86c1027648991e. If the hardware doesn't support MOVBE, but L0 sets CPUID.01H:ECX.MOVBE in L1's emulated CPUID information, then L1 is likely to pass that CPUID bit through to L2. L2 will expect MOVBE to work, but if L1 doesn't intercept #UD, then any MOVBE instruction executed in L2 will raise #UD, and the exception will be delivered in L2. Commit ac9b305caa0df6f5b75d294e4b86c1027648991e is a better and more complete version of ae1f57670703 ("KVM: nVMX: Do not emulate #UD while in guest mode"); however, neither considers the above case. Suggested-by: Jim Mattson Signed-off-by: Paolo Bonzini --- arch/x86/kvm/svm.c | 9 +-------- arch/x86/kvm/vmx.c | 5 +---- 2 files changed, 2 insertions(+), 12 deletions(-) (limited to 'arch/x86/kvm/vmx.c') diff --git a/arch/x86/kvm/svm.c b/arch/x86/kvm/svm.c index bb31c801f1fc..3158dac87f82 100644 --- a/arch/x86/kvm/svm.c +++ b/arch/x86/kvm/svm.c @@ -361,7 +361,6 @@ static void recalc_intercepts(struct vcpu_svm *svm) { struct vmcb_control_area *c, *h; struct nested_state *g; - u32 h_intercept_exceptions; mark_dirty(svm->vmcb, VMCB_INTERCEPTS); @@ -372,14 +371,9 @@ static void recalc_intercepts(struct vcpu_svm *svm) h = &svm->nested.hsave->control; g = &svm->nested; - /* No need to intercept #UD if L1 doesn't intercept it */ - h_intercept_exceptions = - h->intercept_exceptions & ~(1U << UD_VECTOR); - c->intercept_cr = h->intercept_cr | g->intercept_cr; c->intercept_dr = h->intercept_dr | g->intercept_dr; - c->intercept_exceptions = - h_intercept_exceptions | g->intercept_exceptions; + c->intercept_exceptions = h->intercept_exceptions | g->intercept_exceptions; c->intercept = h->intercept | g->intercept; } @@ -2202,7 +2196,6 @@ static int ud_interception(struct vcpu_svm *svm) { int er; - WARN_ON_ONCE(is_guest_mode(&svm->vcpu)); er = emulate_instruction(&svm->vcpu, EMULTYPE_TRAP_UD); if (er == EMULATE_USER_EXIT) return 0; diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c index 5c14d65f676a..427fd3200dd8 100644 --- a/arch/x86/kvm/vmx.c +++ b/arch/x86/kvm/vmx.c @@ -1887,7 +1887,7 @@ static void update_exception_bitmap(struct kvm_vcpu *vcpu) { u32 eb; - eb = (1u << PF_VECTOR) | (1u << MC_VECTOR) | + eb = (1u << PF_VECTOR) | (1u << UD_VECTOR) | (1u << MC_VECTOR) | (1u << DB_VECTOR) | (1u << AC_VECTOR); if ((vcpu->guest_debug & (KVM_GUESTDBG_ENABLE | KVM_GUESTDBG_USE_SW_BP)) == @@ -1905,8 +1905,6 @@ static void update_exception_bitmap(struct kvm_vcpu *vcpu) */ if (is_guest_mode(vcpu)) eb |= get_vmcs12(vcpu)->exception_bitmap; - else - eb |= 1u << UD_VECTOR; vmcs_write32(EXCEPTION_BITMAP, eb); } @@ -5917,7 +5915,6 @@ static int handle_exception(struct kvm_vcpu *vcpu) return 1; /* already handled by vmx_vcpu_run() */ if (is_invalid_opcode(intr_info)) { - WARN_ON_ONCE(is_guest_mode(vcpu)); er = emulate_instruction(vcpu, EMULTYPE_TRAP_UD); if (er == EMULATE_USER_EXIT) return 0; -- cgit v1.2.3-59-g8ed1b From 75f139aaf896d6fdeec2e468ddfa4b2fe469bf40 Mon Sep 17 00:00:00 2001 From: Andrew Honig Date: Wed, 10 Jan 2018 10:12:03 -0800 Subject: KVM: x86: Add memory barrier on vmcs field lookup This adds a memory barrier when performing a lookup into the vmcs_field_to_offset_table. This is related to CVE-2017-5753. Signed-off-by: Andrew Honig Reviewed-by: Jim Mattson Cc: stable@vger.kernel.org Signed-off-by: Paolo Bonzini --- arch/x86/kvm/vmx.c | 12 ++++++++++-- 1 file changed, 10 insertions(+), 2 deletions(-) (limited to 'arch/x86/kvm/vmx.c') diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c index a6f4f095f8f4..7f8fcc5ce664 100644 --- a/arch/x86/kvm/vmx.c +++ b/arch/x86/kvm/vmx.c @@ -884,8 +884,16 @@ static inline short vmcs_field_to_offset(unsigned long field) { BUILD_BUG_ON(ARRAY_SIZE(vmcs_field_to_offset_table) > SHRT_MAX); - if (field >= ARRAY_SIZE(vmcs_field_to_offset_table) || - vmcs_field_to_offset_table[field] == 0) + if (field >= ARRAY_SIZE(vmcs_field_to_offset_table)) + return -ENOENT; + + /* + * FIXME: Mitigation for CVE-2017-5753. To be replaced with a + * generic mechanism. + */ + asm("lfence"); + + if (vmcs_field_to_offset_table[field] == 0) return -ENOENT; return vmcs_field_to_offset_table[field]; -- cgit v1.2.3-59-g8ed1b From 117cc7a908c83697b0b737d15ae1eb5943afe35b Mon Sep 17 00:00:00 2001 From: David Woodhouse Date: Fri, 12 Jan 2018 11:11:27 +0000 Subject: x86/retpoline: Fill return stack buffer on vmexit In accordance with the Intel and AMD documentation, we need to overwrite all entries in the RSB on exiting a guest, to prevent malicious branch target predictions from affecting the host kernel. This is needed both for retpoline and for IBRS. [ak: numbers again for the RSB stuffing labels] Signed-off-by: David Woodhouse Signed-off-by: Thomas Gleixner Tested-by: Peter Zijlstra (Intel) Cc: gnomes@lxorguk.ukuu.org.uk Cc: Rik van Riel Cc: Andi Kleen Cc: Josh Poimboeuf Cc: thomas.lendacky@amd.com Cc: Linus Torvalds Cc: Jiri Kosina Cc: Andy Lutomirski Cc: Dave Hansen Cc: Kees Cook Cc: Tim Chen Cc: Greg Kroah-Hartman Cc: Paul Turner Link: https://lkml.kernel.org/r/1515755487-8524-1-git-send-email-dwmw@amazon.co.uk --- arch/x86/include/asm/nospec-branch.h | 78 +++++++++++++++++++++++++++++++++++- arch/x86/kvm/svm.c | 4 ++ arch/x86/kvm/vmx.c | 4 ++ 3 files changed, 85 insertions(+), 1 deletion(-) (limited to 'arch/x86/kvm/vmx.c') diff --git a/arch/x86/include/asm/nospec-branch.h b/arch/x86/include/asm/nospec-branch.h index ea034fa6e261..402a11c803c3 100644 --- a/arch/x86/include/asm/nospec-branch.h +++ b/arch/x86/include/asm/nospec-branch.h @@ -7,6 +7,48 @@ #include #include +/* + * Fill the CPU return stack buffer. + * + * Each entry in the RSB, if used for a speculative 'ret', contains an + * infinite 'pause; jmp' loop to capture speculative execution. + * + * This is required in various cases for retpoline and IBRS-based + * mitigations for the Spectre variant 2 vulnerability. Sometimes to + * eliminate potentially bogus entries from the RSB, and sometimes + * purely to ensure that it doesn't get empty, which on some CPUs would + * allow predictions from other (unwanted!) sources to be used. + * + * We define a CPP macro such that it can be used from both .S files and + * inline assembly. It's possible to do a .macro and then include that + * from C via asm(".include ") but let's not go there. + */ + +#define RSB_CLEAR_LOOPS 32 /* To forcibly overwrite all entries */ +#define RSB_FILL_LOOPS 16 /* To avoid underflow */ + +/* + * Google experimented with loop-unrolling and this turned out to be + * the optimal version — two calls, each with their own speculation + * trap should their return address end up getting used, in a loop. + */ +#define __FILL_RETURN_BUFFER(reg, nr, sp) \ + mov $(nr/2), reg; \ +771: \ + call 772f; \ +773: /* speculation trap */ \ + pause; \ + jmp 773b; \ +772: \ + call 774f; \ +775: /* speculation trap */ \ + pause; \ + jmp 775b; \ +774: \ + dec reg; \ + jnz 771b; \ + add $(BITS_PER_LONG/8) * nr, sp; + #ifdef __ASSEMBLY__ /* @@ -74,6 +116,20 @@ #else call *\reg #endif +.endm + + /* + * A simpler FILL_RETURN_BUFFER macro. Don't make people use the CPP + * monstrosity above, manually. + */ +.macro FILL_RETURN_BUFFER reg:req nr:req ftr:req +#ifdef CONFIG_RETPOLINE + ANNOTATE_NOSPEC_ALTERNATIVE + ALTERNATIVE "jmp .Lskip_rsb_\@", \ + __stringify(__FILL_RETURN_BUFFER(\reg,\nr,%_ASM_SP)) \ + \ftr +.Lskip_rsb_\@: +#endif .endm #else /* __ASSEMBLY__ */ @@ -119,7 +175,7 @@ X86_FEATURE_RETPOLINE) # define THUNK_TARGET(addr) [thunk_target] "rm" (addr) -#else /* No retpoline */ +#else /* No retpoline for C / inline asm */ # define CALL_NOSPEC "call *%[thunk_target]\n" # define THUNK_TARGET(addr) [thunk_target] "rm" (addr) #endif @@ -134,5 +190,25 @@ enum spectre_v2_mitigation { SPECTRE_V2_IBRS, }; +/* + * On VMEXIT we must ensure that no RSB predictions learned in the guest + * can be followed in the host, by overwriting the RSB completely. Both + * retpoline and IBRS mitigations for Spectre v2 need this; only on future + * CPUs with IBRS_ATT *might* it be avoided. + */ +static inline void vmexit_fill_RSB(void) +{ +#ifdef CONFIG_RETPOLINE + unsigned long loops = RSB_CLEAR_LOOPS / 2; + + asm volatile (ANNOTATE_NOSPEC_ALTERNATIVE + ALTERNATIVE("jmp 910f", + __stringify(__FILL_RETURN_BUFFER(%0, RSB_CLEAR_LOOPS, %1)), + X86_FEATURE_RETPOLINE) + "910:" + : "=&r" (loops), ASM_CALL_CONSTRAINT + : "r" (loops) : "memory" ); +#endif +} #endif /* __ASSEMBLY__ */ #endif /* __NOSPEC_BRANCH_H__ */ diff --git a/arch/x86/kvm/svm.c b/arch/x86/kvm/svm.c index 0e68f0b3cbf7..2744b97345b8 100644 --- a/arch/x86/kvm/svm.c +++ b/arch/x86/kvm/svm.c @@ -45,6 +45,7 @@ #include #include #include +#include #include #include "trace.h" @@ -4985,6 +4986,9 @@ static void svm_vcpu_run(struct kvm_vcpu *vcpu) #endif ); + /* Eliminate branch target predictions from guest mode */ + vmexit_fill_RSB(); + #ifdef CONFIG_X86_64 wrmsrl(MSR_GS_BASE, svm->host.gs_base); #else diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c index 62ee4362e1c1..d1e25dba3112 100644 --- a/arch/x86/kvm/vmx.c +++ b/arch/x86/kvm/vmx.c @@ -50,6 +50,7 @@ #include #include #include +#include #include "trace.h" #include "pmu.h" @@ -9403,6 +9404,9 @@ static void __noclone vmx_vcpu_run(struct kvm_vcpu *vcpu) #endif ); + /* Eliminate branch target predictions from guest mode */ + vmexit_fill_RSB(); + /* MSR_IA32_DEBUGCTLMSR is zeroed on vmexit. Restore it if needed */ if (debugctlmsr) update_debugctlmsr(debugctlmsr); -- cgit v1.2.3-59-g8ed1b