x86/debug: Fix BTF handling

The SDM states that #DB clears DEBUGCTLMSR_BTF, this means that when the
bit is set for userspace (TIF_BLOCKSTEP) and a kernel #DB happens first,
the BTF bit meant for userspace execution is lost.

Have the kernel #DB handler restore the BTF bit when it was requested
for userspace.

Signed-off-by: Peter Zijlstra (Intel) <[email protected]>
Signed-off-by: Thomas Gleixner <[email protected]>
Tested-by: Kyle Huey <[email protected]> 
Link: https://lore.kernel.org/r/[email protected]

Author: Peter Zijlstra

arch/x86/kernel/traps.c

@@ -799,13 +799,6 @@ static __always_inline unsigned long debug_read_clear_dr6(void)
 	 */
 	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 +866,20 @@ 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.
@@ -935,6 +942,13 @@ static __always_inline void exc_debug_user(struct pt_regs *regs,
 	irqentry_enter_from_user_mode(regs);
 	instrumentation_begin();
 
+	/*
+	 * 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.