linux-kernel - [PATCH 5/6] lockdep: Prepare for NMI IRQ state tracking

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20200527155003.407192056@infradead.org>
Date:   Wed, 27 May 2020 17:45:32 +0200
From:   Peter Zijlstra <peterz@...radead.org>
To:     mingo@...nel.org, will@...nel.org, tglx@...utronix.de
Cc:     x86@...nel.org, linux-kernel@...r.kernel.org,
        a.darwish@...utronix.de, rostedt@...dmis.org,
        bigeasy@...utronix.de, peterz@...radead.org
Subject: [PATCH 5/6] lockdep: Prepare for NMI IRQ state tracking

There is no reason not to always, accurately, track IRQ state.

This change also makes IRQ state tracking ignore lockdep_off().

Signed-off-by: Peter Zijlstra (Intel) <peterz@...radead.org>
---
 kernel/locking/lockdep.c |   23 ++++++++++++++++-------
 1 file changed, 16 insertions(+), 7 deletions(-)

--- a/kernel/locking/lockdep.c
+++ b/kernel/locking/lockdep.c
@@ -3646,7 +3646,12 @@ static void __trace_hardirqs_on_caller(v
  */
 void lockdep_hardirqs_on_prepare(unsigned long ip)
 {
-	if (unlikely(!debug_locks || current->lockdep_recursion))
+	/*
+	 * Even though NMIs can happen while in lockdep; we only call this
+	 * then the NMI returns to an IRQs enabled context, which guarantees
+	 * we're not in lockdep.
+	 */
+	if (DEBUG_LOCKS_WARN_ON(current->lockdep_recursion & LOCKDEP_RECURSION_MASK))
 		return;
 
 	if (unlikely(lockdep_hardirqs_enabled())) {
@@ -3682,9 +3687,11 @@ void lockdep_hardirqs_on_prepare(unsigne
 
 	current->hardirq_chain_key = current->curr_chain_key;
 
-	current->lockdep_recursion++;
-	__trace_hardirqs_on_caller();
-	lockdep_recursion_finish();
+	if (!in_nmi()) {
+		current->lockdep_recursion++;
+		__trace_hardirqs_on_caller();
+		lockdep_recursion_finish();
+	}
 }
 EXPORT_SYMBOL_GPL(lockdep_hardirqs_on_prepare);
 
@@ -3692,7 +3699,7 @@ void noinstr lockdep_hardirqs_on(unsigne
 {
 	struct task_struct *curr = current;
 
-	if (unlikely(!debug_locks || curr->lockdep_recursion))
+	if (DEBUG_LOCKS_WARN_ON(curr->lockdep_recursion & LOCKDEP_RECURSION_MASK))
 		return;
 
 	if (lockdep_hardirqs_enabled()) {
@@ -3735,8 +3742,10 @@ void noinstr lockdep_hardirqs_off(unsign
 {
 	struct task_struct *curr = current;
 
-	if (unlikely(!debug_locks || curr->lockdep_recursion))
-		return;
+	/*
+	 * Can't test recursion; NMIs can happen while in lockdep.
+	 * Harmless though; all we do is clear hardirqs_enabled.
+	 */
 
 	/*
 	 * So we're supposed to get called after you mask local IRQs, but for