All of lore.kernel.org
 help / color / mirror / Atom feed
From: Peter Zijlstra <peterz@infradead.org>
To: mingo@kernel.org, will@kernel.org, tglx@linutronix.de
Cc: x86@kernel.org, linux-kernel@vger.kernel.org,
	a.darwish@linutronix.de, rostedt@goodmis.org,
	bigeasy@linutronix.de, peterz@infradead.org, davem@davemloft.net,
	sparclinux@vger.kernel.org, mpe@ellerman.id.au,
	linuxppc-dev@lists.ozlabs.org, heiko.carstens@de.ibm.com,
	linux-s390@vger.kernel.org, linux@armlinux.org.uk
Subject: [PATCH v4 1/8] lockdep: Prepare for NMI IRQ state tracking
Date: Tue, 23 Jun 2020 10:36:46 +0200	[thread overview]
Message-ID: <20200623083721.155449112@infradead.org> (raw)
In-Reply-To: 20200623083645.277342609@infradead.org

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@infradead.org>
---
 kernel/locking/lockdep.c |   44 +++++++++++++++++++++++++++++++++++++++++---
 1 file changed, 41 insertions(+), 3 deletions(-)

--- a/kernel/locking/lockdep.c
+++ b/kernel/locking/lockdep.c
@@ -3646,7 +3646,16 @@ static void __trace_hardirqs_on_caller(v
  */
 void lockdep_hardirqs_on_prepare(unsigned long ip)
 {
-	if (unlikely(!debug_locks || current->lockdep_recursion))
+	if (unlikely(!debug_locks))
+		return;
+
+	/*
+	 * NMIs do not (and cannot) track lock dependencies, nothing to do.
+	 */
+	if (unlikely(in_nmi()))
+		return;
+
+	if (unlikely(current->lockdep_recursion & LOCKDEP_RECURSION_MASK))
 		return;
 
 	if (unlikely(current->hardirqs_enabled)) {
@@ -3692,7 +3701,27 @@ void noinstr lockdep_hardirqs_on(unsigne
 {
 	struct task_struct *curr = current;
 
-	if (unlikely(!debug_locks || curr->lockdep_recursion))
+	if (unlikely(!debug_locks))
+		return;
+
+	/*
+	 * NMIs can happen in the middle of local_irq_{en,dis}able() where the
+	 * tracking state and hardware state are out of sync.
+	 *
+	 * NMIs must save lockdep_hardirqs_enabled() to restore IRQ state from,
+	 * and not rely on hardware state like normal interrupts.
+	 */
+	if (unlikely(in_nmi())) {
+		/*
+		 * Skip:
+		 *  - recursion check, because NMI can hit lockdep;
+		 *  - hardware state check, because above;
+		 *  - chain_key check, see lockdep_hardirqs_on_prepare().
+		 */
+		goto skip_checks;
+	}
+
+	if (unlikely(current->lockdep_recursion & LOCKDEP_RECURSION_MASK))
 		return;
 
 	if (curr->hardirqs_enabled) {
@@ -3720,6 +3749,7 @@ void noinstr lockdep_hardirqs_on(unsigne
 	DEBUG_LOCKS_WARN_ON(current->hardirq_chain_key !=
 			    current->curr_chain_key);
 
+skip_checks:
 	/* we'll do an OFF -> ON transition: */
 	curr->hardirqs_enabled = 1;
 	curr->hardirq_enable_ip = ip;
@@ -3735,7 +3765,15 @@ void noinstr lockdep_hardirqs_off(unsign
 {
 	struct task_struct *curr = current;
 
-	if (unlikely(!debug_locks || curr->lockdep_recursion))
+	if (unlikely(!debug_locks))
+		return;
+
+	/*
+	 * Matching lockdep_hardirqs_on(), allow NMIs in the middle of lockdep;
+	 * they will restore the software state. This ensures the software
+	 * state is consistent inside NMIs as well.
+	 */
+	if (unlikely(!in_nmi() && (current->lockdep_recursion & LOCKDEP_RECURSION_MASK)))
 		return;
 
 	/*



WARNING: multiple messages have this Message-ID (diff)
From: Peter Zijlstra <peterz@infradead.org>
To: mingo@kernel.org, will@kernel.org, tglx@linutronix.de
Cc: x86@kernel.org, linux-kernel@vger.kernel.org,
	a.darwish@linutronix.de, rostedt@goodmis.org,
	bigeasy@linutronix.de, peterz@infradead.org, davem@davemloft.net,
	sparclinux@vger.kernel.org, mpe@ellerman.id.au,
	linuxppc-dev@lists.ozlabs.org, heiko.carstens@de.ibm.com,
	linux-s390@vger.kernel.org, linux@armlinux.org.uk
Subject: [PATCH v4 1/8] lockdep: Prepare for NMI IRQ state tracking
Date: Tue, 23 Jun 2020 08:36:46 +0000	[thread overview]
Message-ID: <20200623083721.155449112@infradead.org> (raw)
In-Reply-To: 20200623083645.277342609@infradead.org

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@infradead.org>
---
 kernel/locking/lockdep.c |   44 +++++++++++++++++++++++++++++++++++++++++---
 1 file changed, 41 insertions(+), 3 deletions(-)

--- a/kernel/locking/lockdep.c
+++ b/kernel/locking/lockdep.c
@@ -3646,7 +3646,16 @@ static void __trace_hardirqs_on_caller(v
  */
 void lockdep_hardirqs_on_prepare(unsigned long ip)
 {
-	if (unlikely(!debug_locks || current->lockdep_recursion))
+	if (unlikely(!debug_locks))
+		return;
+
+	/*
+	 * NMIs do not (and cannot) track lock dependencies, nothing to do.
+	 */
+	if (unlikely(in_nmi()))
+		return;
+
+	if (unlikely(current->lockdep_recursion & LOCKDEP_RECURSION_MASK))
 		return;
 
 	if (unlikely(current->hardirqs_enabled)) {
@@ -3692,7 +3701,27 @@ void noinstr lockdep_hardirqs_on(unsigne
 {
 	struct task_struct *curr = current;
 
-	if (unlikely(!debug_locks || curr->lockdep_recursion))
+	if (unlikely(!debug_locks))
+		return;
+
+	/*
+	 * NMIs can happen in the middle of local_irq_{en,dis}able() where the
+	 * tracking state and hardware state are out of sync.
+	 *
+	 * NMIs must save lockdep_hardirqs_enabled() to restore IRQ state from,
+	 * and not rely on hardware state like normal interrupts.
+	 */
+	if (unlikely(in_nmi())) {
+		/*
+		 * Skip:
+		 *  - recursion check, because NMI can hit lockdep;
+		 *  - hardware state check, because above;
+		 *  - chain_key check, see lockdep_hardirqs_on_prepare().
+		 */
+		goto skip_checks;
+	}
+
+	if (unlikely(current->lockdep_recursion & LOCKDEP_RECURSION_MASK))
 		return;
 
 	if (curr->hardirqs_enabled) {
@@ -3720,6 +3749,7 @@ void noinstr lockdep_hardirqs_on(unsigne
 	DEBUG_LOCKS_WARN_ON(current->hardirq_chain_key ! 			    current->curr_chain_key);
 
+skip_checks:
 	/* we'll do an OFF -> ON transition: */
 	curr->hardirqs_enabled = 1;
 	curr->hardirq_enable_ip = ip;
@@ -3735,7 +3765,15 @@ void noinstr lockdep_hardirqs_off(unsign
 {
 	struct task_struct *curr = current;
 
-	if (unlikely(!debug_locks || curr->lockdep_recursion))
+	if (unlikely(!debug_locks))
+		return;
+
+	/*
+	 * Matching lockdep_hardirqs_on(), allow NMIs in the middle of lockdep;
+	 * they will restore the software state. This ensures the software
+	 * state is consistent inside NMIs as well.
+	 */
+	if (unlikely(!in_nmi() && (current->lockdep_recursion & LOCKDEP_RECURSION_MASK)))
 		return;
 
 	/*

WARNING: multiple messages have this Message-ID (diff)
From: Peter Zijlstra <peterz@infradead.org>
To: mingo@kernel.org, will@kernel.org, tglx@linutronix.de
Cc: linux-s390@vger.kernel.org, peterz@infradead.org,
	bigeasy@linutronix.de, x86@kernel.org, heiko.carstens@de.ibm.com,
	linux-kernel@vger.kernel.org, rostedt@goodmis.org,
	linux@armlinux.org.uk, a.darwish@linutronix.de,
	sparclinux@vger.kernel.org, linuxppc-dev@lists.ozlabs.org,
	davem@davemloft.net
Subject: [PATCH v4 1/8] lockdep: Prepare for NMI IRQ state tracking
Date: Tue, 23 Jun 2020 10:36:46 +0200	[thread overview]
Message-ID: <20200623083721.155449112@infradead.org> (raw)
In-Reply-To: 20200623083645.277342609@infradead.org

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@infradead.org>
---
 kernel/locking/lockdep.c |   44 +++++++++++++++++++++++++++++++++++++++++---
 1 file changed, 41 insertions(+), 3 deletions(-)

--- a/kernel/locking/lockdep.c
+++ b/kernel/locking/lockdep.c
@@ -3646,7 +3646,16 @@ static void __trace_hardirqs_on_caller(v
  */
 void lockdep_hardirqs_on_prepare(unsigned long ip)
 {
-	if (unlikely(!debug_locks || current->lockdep_recursion))
+	if (unlikely(!debug_locks))
+		return;
+
+	/*
+	 * NMIs do not (and cannot) track lock dependencies, nothing to do.
+	 */
+	if (unlikely(in_nmi()))
+		return;
+
+	if (unlikely(current->lockdep_recursion & LOCKDEP_RECURSION_MASK))
 		return;
 
 	if (unlikely(current->hardirqs_enabled)) {
@@ -3692,7 +3701,27 @@ void noinstr lockdep_hardirqs_on(unsigne
 {
 	struct task_struct *curr = current;
 
-	if (unlikely(!debug_locks || curr->lockdep_recursion))
+	if (unlikely(!debug_locks))
+		return;
+
+	/*
+	 * NMIs can happen in the middle of local_irq_{en,dis}able() where the
+	 * tracking state and hardware state are out of sync.
+	 *
+	 * NMIs must save lockdep_hardirqs_enabled() to restore IRQ state from,
+	 * and not rely on hardware state like normal interrupts.
+	 */
+	if (unlikely(in_nmi())) {
+		/*
+		 * Skip:
+		 *  - recursion check, because NMI can hit lockdep;
+		 *  - hardware state check, because above;
+		 *  - chain_key check, see lockdep_hardirqs_on_prepare().
+		 */
+		goto skip_checks;
+	}
+
+	if (unlikely(current->lockdep_recursion & LOCKDEP_RECURSION_MASK))
 		return;
 
 	if (curr->hardirqs_enabled) {
@@ -3720,6 +3749,7 @@ void noinstr lockdep_hardirqs_on(unsigne
 	DEBUG_LOCKS_WARN_ON(current->hardirq_chain_key !=
 			    current->curr_chain_key);
 
+skip_checks:
 	/* we'll do an OFF -> ON transition: */
 	curr->hardirqs_enabled = 1;
 	curr->hardirq_enable_ip = ip;
@@ -3735,7 +3765,15 @@ void noinstr lockdep_hardirqs_off(unsign
 {
 	struct task_struct *curr = current;
 
-	if (unlikely(!debug_locks || curr->lockdep_recursion))
+	if (unlikely(!debug_locks))
+		return;
+
+	/*
+	 * Matching lockdep_hardirqs_on(), allow NMIs in the middle of lockdep;
+	 * they will restore the software state. This ensures the software
+	 * state is consistent inside NMIs as well.
+	 */
+	if (unlikely(!in_nmi() && (current->lockdep_recursion & LOCKDEP_RECURSION_MASK)))
 		return;
 
 	/*



  reply	other threads:[~2020-06-23  8:39 UTC|newest]

Thread overview: 104+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-06-23  8:36 [PATCH v4 0/8] lockdep: Change IRQ state tracking to use per-cpu variables Peter Zijlstra
2020-06-23  8:36 ` Peter Zijlstra
2020-06-23  8:36 ` Peter Zijlstra
2020-06-23  8:36 ` Peter Zijlstra [this message]
2020-06-23  8:36   ` [PATCH v4 1/8] lockdep: Prepare for NMI IRQ state tracking Peter Zijlstra
2020-06-23  8:36   ` Peter Zijlstra
2020-07-11 10:09   ` [tip: locking/core] " tip-bot2 for Peter Zijlstra
2020-06-23  8:36 ` [PATCH v4 2/8] x86/entry: Fix NMI vs " Peter Zijlstra
2020-06-23  8:36   ` Peter Zijlstra
2020-06-23  8:36   ` Peter Zijlstra
2020-07-11 10:09   ` [tip: locking/core] " tip-bot2 for Peter Zijlstra
2020-06-23  8:36 ` [PATCH v4 3/8] sparc64: Fix asm/percpu.h build error Peter Zijlstra
2020-06-23  8:36   ` Peter Zijlstra
2020-06-23  8:36   ` Peter Zijlstra
2020-06-23 21:35   ` David Miller
2020-06-23 21:35     ` David Miller
2020-06-23 21:35     ` David Miller
2020-07-11 10:09   ` [tip: locking/core] " tip-bot2 for Peter Zijlstra
2020-06-23  8:36 ` [PATCH v4 4/8] powerpc64: Break asm/percpu.h vs spinlock_types.h dependency Peter Zijlstra
2020-06-23  8:36   ` Peter Zijlstra
2020-06-23  8:36   ` Peter Zijlstra
2020-06-23  8:36 ` [PATCH v4 5/8] s390: Break cyclic percpu include Peter Zijlstra
2020-06-23  8:36   ` Peter Zijlstra
2020-06-23  8:36   ` Peter Zijlstra
2020-07-11 10:09   ` [tip: locking/core] " tip-bot2 for Peter Zijlstra
2020-06-23  8:36 ` [PATCH v4 6/8] arm: " Peter Zijlstra
2020-06-23  8:36   ` Peter Zijlstra
2020-06-23  8:36   ` Peter Zijlstra
2020-06-23  9:02   ` Will Deacon
2020-06-23  9:02     ` Will Deacon
2020-06-23  9:02     ` Will Deacon
2020-06-24 17:53     ` Peter Zijlstra
2020-06-24 17:53       ` Peter Zijlstra
2020-06-24 17:53       ` Peter Zijlstra
2020-06-25  7:31       ` Will Deacon
2020-06-25  7:31         ` Will Deacon
2020-06-25  7:31         ` Will Deacon
2020-07-11 10:09   ` [tip: locking/core] " tip-bot2 for Peter Zijlstra
2020-06-23  8:36 ` [PATCH v4 7/8] lockdep: Change hardirq{s_enabled,_context} to per-cpu variables Peter Zijlstra
2020-06-23  8:36   ` [PATCH v4 7/8] lockdep: Change hardirq{s_enabled, _context} " Peter Zijlstra
2020-06-23  8:36   ` [PATCH v4 7/8] lockdep: Change hardirq{s_enabled,_context} " Peter Zijlstra
2020-06-23 15:00   ` Ahmed S. Darwish
2020-06-23 15:00     ` Ahmed S. Darwish
2020-06-23 15:00     ` Ahmed S. Darwish
2020-06-23 15:24     ` Peter Zijlstra
2020-06-23 15:24       ` Peter Zijlstra
2020-06-23 15:24       ` Peter Zijlstra
2020-06-23 16:13       ` Ahmed S. Darwish
2020-06-23 16:13         ` Ahmed S. Darwish
2020-06-23 16:13         ` Ahmed S. Darwish
2020-06-23 16:37         ` Peter Zijlstra
2020-06-23 16:37           ` Peter Zijlstra
2020-06-23 16:37           ` Peter Zijlstra
2020-06-23 17:59           ` Marco Elver
2020-06-23 17:59             ` Marco Elver
2020-06-23 17:59             ` Marco Elver
2020-06-23 18:12             ` Peter Zijlstra
2020-06-23 18:12               ` Peter Zijlstra
2020-06-23 18:12               ` Peter Zijlstra
2020-06-23 18:39               ` Marco Elver
2020-06-23 18:39                 ` Marco Elver
2020-06-23 18:39                 ` Marco Elver
2020-06-23 19:13                 ` Marco Elver
2020-06-23 19:13                   ` Marco Elver
2020-06-23 19:13                   ` Marco Elver
2020-06-23 19:41                   ` Peter Zijlstra
2020-06-23 19:41                     ` Peter Zijlstra
2020-06-23 19:41                     ` Peter Zijlstra
2020-06-23 20:08                   ` Peter Zijlstra
2020-06-23 20:08                     ` Peter Zijlstra
2020-06-23 20:08                     ` Peter Zijlstra
2020-06-23 20:24               ` Peter Zijlstra
2020-06-23 20:24                 ` Peter Zijlstra
2020-06-23 20:24                 ` Peter Zijlstra
2020-06-23 20:33                 ` Peter Zijlstra
2020-06-23 20:33                   ` Peter Zijlstra
2020-06-23 20:33                   ` Peter Zijlstra
2020-06-24  9:00                 ` Peter Zijlstra
2020-06-24  9:00                   ` Peter Zijlstra
2020-06-24  9:00                   ` Peter Zijlstra
2020-06-24 10:17                   ` Marco Elver
2020-06-24 10:17                     ` Marco Elver
2020-06-24 10:17                     ` Marco Elver
2020-06-24 12:31                     ` Peter Zijlstra
2020-06-24 12:31                       ` Peter Zijlstra
2020-06-24 12:31                       ` Peter Zijlstra
2020-06-24 11:32                 ` Marco Elver
2020-06-24 11:32                   ` Marco Elver
2020-06-24 11:32                   ` Marco Elver
2020-06-24 15:18                   ` Peter Zijlstra
2020-06-24 15:18                     ` Peter Zijlstra
2020-06-24 15:18                     ` Peter Zijlstra
2020-07-11 10:09                   ` [tip: locking/core] kcsan: Make KCSAN compatible with new IRQ state tracking tip-bot2 for Marco Elver
2020-06-30  5:59   ` [PATCH v4 7/8] lockdep: Change hardirq{s_enabled,_context} to per-cpu variables Ahmed S. Darwish
2020-06-30  5:59     ` Ahmed S. Darwish
2020-06-30  5:59     ` Ahmed S. Darwish
2020-06-30  9:40     ` Peter Zijlstra
2020-06-30  9:40       ` Peter Zijlstra
2020-06-30  9:40       ` Peter Zijlstra
2020-07-11 10:09   ` [tip: locking/core] " tip-bot2 for Peter Zijlstra
2020-06-23  8:36 ` [PATCH v4 8/8] lockdep: Remove lockdep_hardirq{s_enabled,_context}() argument Peter Zijlstra
2020-06-23  8:36   ` [PATCH v4 8/8] lockdep: Remove lockdep_hardirq{s_enabled, _context}() argument Peter Zijlstra
2020-06-23  8:36   ` [PATCH v4 8/8] lockdep: Remove lockdep_hardirq{s_enabled,_context}() argument Peter Zijlstra
2020-07-11 10:09   ` [tip: locking/core] " tip-bot2 for Peter Zijlstra

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20200623083721.155449112@infradead.org \
    --to=peterz@infradead.org \
    --cc=a.darwish@linutronix.de \
    --cc=bigeasy@linutronix.de \
    --cc=davem@davemloft.net \
    --cc=heiko.carstens@de.ibm.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-s390@vger.kernel.org \
    --cc=linux@armlinux.org.uk \
    --cc=linuxppc-dev@lists.ozlabs.org \
    --cc=mingo@kernel.org \
    --cc=mpe@ellerman.id.au \
    --cc=rostedt@goodmis.org \
    --cc=sparclinux@vger.kernel.org \
    --cc=tglx@linutronix.de \
    --cc=will@kernel.org \
    --cc=x86@kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.