[PATCH v2.1 8/8] x86/debug: Remove the historical junk

[peterz@infradead.org]

On Fri, Aug 21, 2020 at 11:39:20AM +0200, Peter Zijlstra wrote:
> Remove the historical junk and replace it with a WARN and a comment.
> 
> The problem is that even though the kernel only uses TF single-step in
> kprobes and KGDB, both of which consume the event before this,
> QEMU/KVM has bugs in this area that can trigger this state so we have
> to deal with it.
> 
> Suggested-by: Brian Gerst <brgerst@gmail.com>
> Suggested-by: Andy Lutomirski <luto@amacapital.net>
> Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>

Damn, I lost a refresh.

---
 arch/x86/kernel/traps.c |   23 ++++++++++++-----------
 1 file changed, 12 insertions(+), 11 deletions(-)

--- a/arch/x86/kernel/traps.c
+++ b/arch/x86/kernel/traps.c
@@ -843,18 +843,19 @@ static __always_inline void exc_debug_ke
 	if (notify_debug(regs, &dr6))
 		goto out;
 
-	if (WARN_ON_ONCE(dr6 & DR_STEP)) {
-		/*
-		 * Historical junk that used to handle SYSENTER single-stepping.
-		 * This should be unreachable now.  If we survive for a while
-		 * without anyone hitting this warning, we'll turn this into
-		 * an oops.
-		 */
-		dr6 &= ~DR_STEP;
-		set_thread_flag(TIF_SINGLESTEP);
+	/*
+	 * The kernel doesn't use TF single-step outside of:
+	 *
+	 *  - Kprobes, consumed through kprobe_debug_handler()
+	 *  - KGDB, consumed through notify_debug()
+	 *
+	 * So if we get here with DR_STEP set, something is wonky.
+	 *
+	 * A known way to trigger this is through QEMU's GDB stub,
+	 * which leaks #DB into the guest and causes IST recursion.
+	 */
+	if (WARN_ON_ONCE(current->thread.debugreg6 & DR_STEP))
 		regs->flags &= ~X86_EFLAGS_TF;
-	}
-
 out:
 	instrumentation_end();
 	idtentry_exit_nmi(regs, irq_state);