diff --git a/arch/x86/kernel/cpu/mce/core.c b/arch/x86/kernel/cpu/mce/core.c
index 0ba24dfffdb27..a697baee9aa0b 100644
--- a/arch/x86/kernel/cpu/mce/core.c
+++ b/arch/x86/kernel/cpu/mce/core.c
@@ -373,10 +373,28 @@ static int msr_to_offset(u32 msr)
 	return -1;
 }
 
+__visible bool ex_handler_rdmsr_fault(const struct exception_table_entry *fixup,
+				      struct pt_regs *regs, int trapnr,
+				      unsigned long error_code,
+				      unsigned long fault_addr)
+{
+	pr_emerg("MSR access error: RDMSR from 0x%x at rIP: 0x%lx (%pS)\n",
+		 (unsigned int)regs->cx, regs->ip, (void *)regs->ip);
+
+	show_stack_regs(regs);
+
+	panic("MCA architectural violation!\n");
+
+	while (true)
+		cpu_relax();
+
+	return true;
+}
+
 /* MSR access wrappers used for error injection */
 static u64 mce_rdmsrl(u32 msr)
 {
-	u64 v;
+	DECLARE_ARGS(val, low, high);
 
 	if (__this_cpu_read(injectm.finished)) {
 		int offset = msr_to_offset(msr);
@@ -386,21 +404,43 @@ static u64 mce_rdmsrl(u32 msr)
 		return *(u64 *)((char *)this_cpu_ptr(&injectm) + offset);
 	}
 
-	if (rdmsrl_safe(msr, &v)) {
-		WARN_ONCE(1, "mce: Unable to read MSR 0x%x!\n", msr);
-		/*
-		 * Return zero in case the access faulted. This should
-		 * not happen normally but can happen if the CPU does
-		 * something weird, or if the code is buggy.
-		 */
-		v = 0;
-	}
+	/*
+	 * RDMSR on MCA MSRs should not fault. If they do, this is very much an
+	 * architectural violation and needs to be reported to hw vendor. Panic
+	 * the box to not allow any further progress.
+	 */
+	asm volatile("1: rdmsr\n"
+		     "2:\n"
+		     _ASM_EXTABLE_HANDLE(1b, 2b, ex_handler_rdmsr_fault)
+		     : EAX_EDX_RET(val, low, high) : "c" (msr));
 
-	return v;
+
+	return EAX_EDX_VAL(val, low, high);
+}
+
+__visible bool ex_handler_wrmsr_fault(const struct exception_table_entry *fixup,
+				      struct pt_regs *regs, int trapnr,
+				      unsigned long error_code,
+				      unsigned long fault_addr)
+{
+	pr_emerg("MSR access error: WRMSR to 0x%x (tried to write 0x%08x%08x) at rIP: 0x%lx (%pS)\n",
+		 (unsigned int)regs->cx, (unsigned int)regs->dx, (unsigned int)regs->ax,
+		  regs->ip, (void *)regs->ip);
+
+	show_stack_regs(regs);
+
+	panic("MCA architectural violation!\n");
+
+	while (true)
+		cpu_relax();
+
+	return true;
 }
 
 static void mce_wrmsrl(u32 msr, u64 v)
 {
+	u32 low, high;
+
 	if (__this_cpu_read(injectm.finished)) {
 		int offset = msr_to_offset(msr);
 
@@ -408,7 +448,15 @@ static void mce_wrmsrl(u32 msr, u64 v)
 			*(u64 *)((char *)this_cpu_ptr(&injectm) + offset) = v;
 		return;
 	}
-	wrmsrl(msr, v);
+
+	low  = (u32)v;
+	high = (u32)(v >> 32);
+
+	/* See comment in mce_rdmsrl() */
+	asm volatile("1: wrmsr\n"
+		     "2:\n"
+		     _ASM_EXTABLE_HANDLE(1b, 2b, ex_handler_wrmsr_fault)
+		     : : "c" (msr), "a"(low), "d" (high) : "memory");
 }
 
 /*
diff --git a/arch/x86/kernel/cpu/mce/internal.h b/arch/x86/kernel/cpu/mce/internal.h
index 6473070b5da49..b122610e9046a 100644
--- a/arch/x86/kernel/cpu/mce/internal.h
+++ b/arch/x86/kernel/cpu/mce/internal.h
@@ -185,4 +185,14 @@ extern bool amd_filter_mce(struct mce *m);
 static inline bool amd_filter_mce(struct mce *m)			{ return false; };
 #endif
 
+__visible bool ex_handler_rdmsr_fault(const struct exception_table_entry *fixup,
+				      struct pt_regs *regs, int trapnr,
+				      unsigned long error_code,
+				      unsigned long fault_addr);
+
+__visible bool ex_handler_wrmsr_fault(const struct exception_table_entry *fixup,
+				      struct pt_regs *regs, int trapnr,
+				      unsigned long error_code,
+				      unsigned long fault_addr);
+
 #endif /* __X86_MCE_INTERNAL_H__ */