8 files changed, 185 insertions, 30 deletions

diff --git a/arch/x86/kernel/alternative.c b/arch/x86/kernel/alternative.c
index 4adbe65afe23..2400ad62f330 100644
--- a/arch/x86/kernel/alternative.c
+++ b/arch/x86/kernel/alternative.c
@@ -807,6 +807,15 @@ static inline temp_mm_state_t use_temporary_mm(struct mm_struct *mm)
 	temp_mm_state_t temp_state;
 
 	lockdep_assert_irqs_disabled();
+
+	/*
+	 * Make sure not to be in TLB lazy mode, as otherwise we'll end up
+	 * with a stale address space WITHOUT being in lazy mode after
+	 * restoring the previous mm.
+	 */
+	if (this_cpu_read(cpu_tlbstate.is_lazy))
+		leave_mm(smp_processor_id());
+
 	temp_state.mm = this_cpu_read(cpu_tlbstate.loaded_mm);
 	switch_mm_irqs_off(NULL, mm, current);
 
diff --git a/arch/x86/kernel/head_64.S b/arch/x86/kernel/head_64.S
index 7eb2a1c87969..3c417734790f 100644
--- a/arch/x86/kernel/head_64.S
+++ b/arch/x86/kernel/head_64.S
@@ -161,6 +161,21 @@ SYM_INNER_LABEL(secondary_startup_64_no_verify, SYM_L_GLOBAL)
 
 	/* Setup early boot stage 4-/5-level pagetables. */
 	addq	phys_base(%rip), %rax
+
+	/*
+	 * For SEV guests: Verify that the C-bit is correct. A malicious
+	 * hypervisor could lie about the C-bit position to perform a ROP
+	 * attack on the guest by writing to the unencrypted stack and wait for
+	 * the next RET instruction.
+	 * %rsi carries pointer to realmode data and is callee-clobbered. Save
+	 * and restore it.
+	 */
+	pushq	%rsi
+	movq	%rax, %rdi
+	call	sev_verify_cbit
+	popq	%rsi
+
+	/* Switch to new page-table */
 	movq	%rax, %cr3
 
 	/* Ensure I am executing from virtual addresses */
@@ -279,6 +294,7 @@ SYM_INNER_LABEL(secondary_startup_64_no_verify, SYM_L_GLOBAL)
 SYM_CODE_END(secondary_startup_64)
 
 #include "verify_cpu.S"
+#include "sev_verify_cbit.S"
 
 #ifdef CONFIG_HOTPLUG_CPU
 /*
diff --git a/arch/x86/kernel/kexec-bzimage64.c b/arch/x86/kernel/kexec-bzimage64.c
index 57c2ecf43134..ce831f9448e7 100644
--- a/arch/x86/kernel/kexec-bzimage64.c
+++ b/arch/x86/kernel/kexec-bzimage64.c
@@ -200,8 +200,7 @@ setup_boot_parameters(struct kimage *image, struct boot_params *params,
 	params->hdr.hardware_subarch = boot_params.hdr.hardware_subarch;
 
 	/* Copying screen_info will do? */
-	memcpy(&params->screen_info, &boot_params.screen_info,
-				sizeof(struct screen_info));
+	memcpy(&params->screen_info, &screen_info, sizeof(struct screen_info));
 
 	/* Fill in memsize later */
 	params->screen_info.ext_mem_k = 0;
diff --git a/arch/x86/kernel/sev-es-shared.c b/arch/x86/kernel/sev-es-shared.c
index 5f83ccaab877..7d04b356d44d 100644
--- a/arch/x86/kernel/sev-es-shared.c
+++ b/arch/x86/kernel/sev-es-shared.c
@@ -178,6 +178,32 @@ void __init do_vc_no_ghcb(struct pt_regs *regs, unsigned long exit_code)
 		goto fail;
 	regs->dx = val >> 32;
 
+	/*
+	 * This is a VC handler and the #VC is only raised when SEV-ES is
+	 * active, which means SEV must be active too. Do sanity checks on the
+	 * CPUID results to make sure the hypervisor does not trick the kernel
+	 * into the no-sev path. This could map sensitive data unencrypted and
+	 * make it accessible to the hypervisor.
+	 *
+	 * In particular, check for:
+	 *	- Hypervisor CPUID bit
+	 *	- Availability of CPUID leaf 0x8000001f
+	 *	- SEV CPUID bit.
+	 *
+	 * The hypervisor might still report the wrong C-bit position, but this
+	 * can't be checked here.
+	 */
+
+	if ((fn == 1 && !(regs->cx & BIT(31))))
+		/* Hypervisor bit */
+		goto fail;
+	else if (fn == 0x80000000 && (regs->ax < 0x8000001f))
+		/* SEV leaf check */
+		goto fail;
+	else if ((fn == 0x8000001f && !(regs->ax & BIT(1))))
+		/* SEV bit */
+		goto fail;
+
 	/* Skip over the CPUID two-byte opcode */
 	regs->ip += 2;
 
diff --git a/arch/x86/kernel/sev-es.c b/arch/x86/kernel/sev-es.c
index 4a96726fbaf8..0bd1a0fc587e 100644
--- a/arch/x86/kernel/sev-es.c
+++ b/arch/x86/kernel/sev-es.c
@@ -374,8 +374,8 @@ fault:
 	return ES_EXCEPTION;
 }
 
-static bool vc_slow_virt_to_phys(struct ghcb *ghcb, struct es_em_ctxt *ctxt,
-				 unsigned long vaddr, phys_addr_t *paddr)
+static enum es_result vc_slow_virt_to_phys(struct ghcb *ghcb, struct es_em_ctxt *ctxt,
+					   unsigned long vaddr, phys_addr_t *paddr)
 {
 	unsigned long va = (unsigned long)vaddr;
 	unsigned int level;
@@ -394,15 +394,19 @@ static bool vc_slow_virt_to_phys(struct ghcb *ghcb, struct es_em_ctxt *ctxt,
 		if (user_mode(ctxt->regs))
 			ctxt->fi.error_code |= X86_PF_USER;
 
-		return false;
+		return ES_EXCEPTION;
 	}
 
+	if (WARN_ON_ONCE(pte_val(*pte) & _PAGE_ENC))
+		/* Emulated MMIO to/from encrypted memory not supported */
+		return ES_UNSUPPORTED;
+
 	pa = (phys_addr_t)pte_pfn(*pte) << PAGE_SHIFT;
 	pa |= va & ~page_level_mask(level);
 
 	*paddr = pa;
 
-	return true;
+	return ES_OK;
 }
 
 /* Include code shared with pre-decompression boot stage */
@@ -731,6 +735,7 @@ static enum es_result vc_do_mmio(struct ghcb *ghcb, struct es_em_ctxt *ctxt,
 {
 	u64 exit_code, exit_info_1, exit_info_2;
 	unsigned long ghcb_pa = __pa(ghcb);
+	enum es_result res;
 	phys_addr_t paddr;
 	void __user *ref;
 
@@ -740,11 +745,12 @@ static enum es_result vc_do_mmio(struct ghcb *ghcb, struct es_em_ctxt *ctxt,
 
 	exit_code = read ? SVM_VMGEXIT_MMIO_READ : SVM_VMGEXIT_MMIO_WRITE;
 
-	if (!vc_slow_virt_to_phys(ghcb, ctxt, (unsigned long)ref, &paddr)) {
-		if (!read)
+	res = vc_slow_virt_to_phys(ghcb, ctxt, (unsigned long)ref, &paddr);
+	if (res != ES_OK) {
+		if (res == ES_EXCEPTION && !read)
 			ctxt->fi.error_code |= X86_PF_WRITE;
 
-		return ES_EXCEPTION;
+		return res;
 	}
 
 	exit_info_1 = paddr;
diff --git a/arch/x86/kernel/sev_verify_cbit.S b/arch/x86/kernel/sev_verify_cbit.S
new file mode 100644
index 000000000000..ee04941a6546
--- /dev/null
+++ b/arch/x86/kernel/sev_verify_cbit.S
@@ -0,0 +1,89 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ *	sev_verify_cbit.S - Code for verification of the C-bit position reported
+ *			    by the Hypervisor when running with SEV enabled.
+ *
+ *	Copyright (c) 2020  Joerg Roedel ([email protected])
+ *
+ * sev_verify_cbit() is called before switching to a new long-mode page-table
+ * at boot.
+ *
+ * Verify that the C-bit position is correct by writing a random value to
+ * an encrypted memory location while on the current page-table. Then it
+ * switches to the new page-table to verify the memory content is still the
+ * same. After that it switches back to the current page-table and when the
+ * check succeeded it returns. If the check failed the code invalidates the
+ * stack pointer and goes into a hlt loop. The stack-pointer is invalidated to
+ * make sure no interrupt or exception can get the CPU out of the hlt loop.
+ *
+ * New page-table pointer is expected in %rdi (first parameter)
+ *
+ */
+SYM_FUNC_START(sev_verify_cbit)
+#ifdef CONFIG_AMD_MEM_ENCRYPT
+	/* First check if a C-bit was detected */
+	movq	sme_me_mask(%rip), %rsi
+	testq	%rsi, %rsi
+	jz	3f
+
+	/* sme_me_mask != 0 could mean SME or SEV - Check also for SEV */
+	movq	sev_status(%rip), %rsi
+	testq	%rsi, %rsi
+	jz	3f
+
+	/* Save CR4 in %rsi */
+	movq	%cr4, %rsi
+
+	/* Disable Global Pages */
+	movq	%rsi, %rdx
+	andq	$(~X86_CR4_PGE), %rdx
+	movq	%rdx, %cr4
+
+	/*
+	 * Verified that running under SEV - now get a random value using
+	 * RDRAND. This instruction is mandatory when running as an SEV guest.
+	 *
+	 * Don't bail out of the loop if RDRAND returns errors. It is better to
+	 * prevent forward progress than to work with a non-random value here.
+	 */
+1:	rdrand	%rdx
+	jnc	1b
+
+	/* Store value to memory and keep it in %rdx */
+	movq	%rdx, sev_check_data(%rip)
+
+	/* Backup current %cr3 value to restore it later */
+	movq	%cr3, %rcx
+
+	/* Switch to new %cr3 - This might unmap the stack */
+	movq	%rdi, %cr3
+
+	/*
+	 * Compare value in %rdx with memory location. If C-bit is incorrect
+	 * this would read the encrypted data and make the check fail.
+	 */
+	cmpq	%rdx, sev_check_data(%rip)
+
+	/* Restore old %cr3 */
+	movq	%rcx, %cr3
+
+	/* Restore previous CR4 */
+	movq	%rsi, %cr4
+
+	/* Check CMPQ result */
+	je	3f
+
+	/*
+	 * The check failed, prevent any forward progress to prevent ROP
+	 * attacks, invalidate the stack and go into a hlt loop.
+	 */
+	xorq	%rsp, %rsp
+	subq	$0x1000, %rsp
+2:	hlt
+	jmp 2b
+3:
+#endif
+	/* Return page-table pointer */
+	movq	%rdi, %rax
+	ret
+SYM_FUNC_END(sev_verify_cbit)
diff --git a/arch/x86/kernel/traps.c b/arch/x86/kernel/traps.c
index 3c70fb34028b..e19df6cde35d 100644
--- a/arch/x86/kernel/traps.c
+++ b/arch/x86/kernel/traps.c
@@ -793,19 +793,6 @@ static __always_inline unsigned long debug_read_clear_dr6(void)
 	set_debugreg(DR6_RESERVED, 6);
 	dr6 ^= DR6_RESERVED; /* Flip to positive polarity */
 
-	/*
-	 * Clear the virtual DR6 value, ptrace routines will set bits here for
-	 * things we want signals for.
-	 */
-	current->thread.virtual_dr6 = 0;
-
-	/*
-	 * The SDM says "The processor clears the BTF flag when it
-	 * generates a debug exception."  Clear TIF_BLOCKSTEP to keep
-	 * TIF_BLOCKSTEP in sync with the hardware BTF flag.
-	 */
-	clear_thread_flag(TIF_BLOCKSTEP);
-
 	return dr6;
 }
 
@@ -873,6 +860,20 @@ static __always_inline void exc_debug_kernel(struct pt_regs *regs,
 	 */
 	WARN_ON_ONCE(user_mode(regs));
 
+	if (test_thread_flag(TIF_BLOCKSTEP)) {
+		/*
+		 * The SDM says "The processor clears the BTF flag when it
+		 * generates a debug exception." but PTRACE_BLOCKSTEP requested
+		 * it for userspace, but we just took a kernel #DB, so re-set
+		 * BTF.
+		 */
+		unsigned long debugctl;
+
+		rdmsrl(MSR_IA32_DEBUGCTLMSR, debugctl);
+		debugctl |= DEBUGCTLMSR_BTF;
+		wrmsrl(MSR_IA32_DEBUGCTLMSR, debugctl);
+	}
+
 	/*
 	 * Catch SYSENTER with TF set and clear DR_STEP. If this hit a
 	 * watchpoint at the same time then that will still be handled.
@@ -936,6 +937,22 @@ static __always_inline void exc_debug_user(struct pt_regs *regs,
 	instrumentation_begin();
 
 	/*
+	 * Start the virtual/ptrace DR6 value with just the DR_STEP mask
+	 * of the real DR6. ptrace_triggered() will set the DR_TRAPn bits.
+	 *
+	 * Userspace expects DR_STEP to be visible in ptrace_get_debugreg(6)
+	 * even if it is not the result of PTRACE_SINGLESTEP.
+	 */
+	current->thread.virtual_dr6 = (dr6 & DR_STEP);
+
+	/*
+	 * The SDM says "The processor clears the BTF flag when it
+	 * generates a debug exception."  Clear TIF_BLOCKSTEP to keep
+	 * TIF_BLOCKSTEP in sync with the hardware BTF flag.
+	 */
+	clear_thread_flag(TIF_BLOCKSTEP);
+
+	/*
 	 * If dr6 has no reason to give us about the origin of this trap,
 	 * then it's very likely the result of an icebp/int01 trap.
 	 * User wants a sigtrap for that.
diff --git a/arch/x86/kernel/unwind_orc.c b/arch/x86/kernel/unwind_orc.c
index 6a339ce328e0..73f800100066 100644
--- a/arch/x86/kernel/unwind_orc.c
+++ b/arch/x86/kernel/unwind_orc.c
@@ -321,19 +321,12 @@ EXPORT_SYMBOL_GPL(unwind_get_return_address);
 
 unsigned long *unwind_get_return_address_ptr(struct unwind_state *state)
 {
-	struct task_struct *task = state->task;
-
 	if (unwind_done(state))
 		return NULL;
 
 	if (state->regs)
 		return &state->regs->ip;
 
-	if (task != current && state->sp == task->thread.sp) {
-		struct inactive_task_frame *frame = (void *)task->thread.sp;
-		return &frame->ret_addr;
-	}
-
 	if (state->sp)
 		return (unsigned long *)state->sp - 1;
 
@@ -663,7 +656,7 @@ void __unwind_start(struct unwind_state *state, struct task_struct *task,
 	} else {
 		struct inactive_task_frame *frame = (void *)task->thread.sp;
 
-		state->sp = task->thread.sp;
+		state->sp = task->thread.sp + sizeof(*frame);
 		state->bp = READ_ONCE_NOCHECK(frame->bp);
 		state->ip = READ_ONCE_NOCHECK(frame->ret_addr);
 		state->signal = (void *)state->ip == ret_from_fork;