From nobody Wed Oct 8 07:35:04 2025 Received: from us-smtp-delivery-124.mimecast.com (us-smtp-delivery-124.mimecast.com [170.10.129.124]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 2958D2857F8 for ; Mon, 30 Jun 2025 19:53:31 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=170.10.129.124 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1751313213; cv=none; b=Ir4mhrF8ngL76JH/wWtp+gMOHFLXtJ9pflud7+YpYjaR39+r0eGCfsbs++iA6JuQILhguqPr+DR6AUM3vlKtPF18kaP69ICShsYGKDXKrzf3Jyp2UKiLrRrRtt+j0tg9/2QkJQrNC0n3TMGGHDDlnVk8Sr+EvENZX/zf/5qCLYY= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1751313213; c=relaxed/simple; bh=rysp77v0m2sVwjmkmSgFE5NF/p+PFjSMmBieBHk2lsU=; h=From:To:Cc:Subject:Date:Message-ID:In-Reply-To:References: MIME-Version:Content-Type; b=ks6o8Q7zvVUuFhqMEgYFYZp5ltY4hycN179HxvHkOs26D4YfeINJfSG4MVs6G4pWdtulcxobUBqb2p7Y+nRv0ZIBFHFOXF/p1BanrBab9A2l+z6na5uvGR7dYH0JpK7vyXAsdrzEAL9p9p9eD0ByY1m1s/N8EbVeaW+StviVmC0= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=pass (p=quarantine dis=none) header.from=redhat.com; spf=pass smtp.mailfrom=redhat.com; dkim=pass (1024-bit key) header.d=redhat.com header.i=@redhat.com header.b=alxL+S+F; arc=none smtp.client-ip=170.10.129.124 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=quarantine dis=none) header.from=redhat.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=redhat.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (1024-bit key) header.d=redhat.com header.i=@redhat.com header.b="alxL+S+F" DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=redhat.com; s=mimecast20190719; t=1751313211; h=from:from:reply-to:subject:subject:date:date:message-id:message-id: to:to:cc:cc:mime-version:mime-version:content-type:content-type: content-transfer-encoding:content-transfer-encoding: in-reply-to:in-reply-to:references:references; bh=CXZLUBjW746zrWrHi02YgbbLPfAn2nFzYFn+w9qge5E=; b=alxL+S+FesKVInxMnZ1EgQ1CduBjvcFcAVtXy6I1S5dCjQCjhsI9PQZmlEBDxQxaMXl9i3 u48nyKQHrbPkFN/d19M/RSiAcT79roC2zS1S3fGFr1fPZ9wha2yzXD9KXIn+N/+oQFVma4 1XUwd6b9/Ugf6JLG12NLDufutQa8Zjo= Received: from mx-prod-mc-03.mail-002.prod.us-west-2.aws.redhat.com (ec2-54-186-198-63.us-west-2.compute.amazonaws.com [54.186.198.63]) by relay.mimecast.com with ESMTP with STARTTLS (version=TLSv1.3, cipher=TLS_AES_256_GCM_SHA384) id us-mta-315-lfhuDkGWOYOpv_KEXkk15g-1; Mon, 30 Jun 2025 15:53:27 -0400 X-MC-Unique: lfhuDkGWOYOpv_KEXkk15g-1 X-Mimecast-MFC-AGG-ID: lfhuDkGWOYOpv_KEXkk15g_1751313205 Received: from mx-prod-int-03.mail-002.prod.us-west-2.aws.redhat.com (mx-prod-int-03.mail-002.prod.us-west-2.aws.redhat.com [10.30.177.12]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature RSA-PSS (2048 bits) server-digest SHA256) (No client certificate requested) by mx-prod-mc-03.mail-002.prod.us-west-2.aws.redhat.com (Postfix) with ESMTPS id 3B08318DA5C2; Mon, 30 Jun 2025 19:53:25 +0000 (UTC) Received: from wcosta-thinkpadt14gen4.rmtbr.csb (unknown [10.22.65.29]) by mx-prod-int-03.mail-002.prod.us-west-2.aws.redhat.com (Postfix) with ESMTP id 6D2BA19560A7; Mon, 30 Jun 2025 19:53:18 +0000 (UTC) From: Wander Lairson Costa To: Ingo Molnar , Peter Zijlstra , Juri Lelli , Vincent Guittot , Dietmar Eggemann , Steven Rostedt , Ben Segall , Mel Gorman , Valentin Schneider , Masami Hiramatsu , Mathieu Desnoyers , Wander Lairson Costa , Thomas Gleixner , Boqun Feng , David Woodhouse , linux-kernel@vger.kernel.org (open list), linux-trace-kernel@vger.kernel.org (open list:TRACING) Cc: Arnaldo Carvalho de Melo , Clark Williams , Gabriele Monaco Subject: [PATCH v2 1/2] trace/preemptirq: reduce overhead of irq_enable/disable tracepoints Date: Mon, 30 Jun 2025 16:52:38 -0300 Message-ID: <20250630195243.701516-2-wander@redhat.com> In-Reply-To: <20250630195243.701516-1-wander@redhat.com> References: <20250630195243.701516-1-wander@redhat.com> Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: quoted-printable X-Scanned-By: MIMEDefang 3.0 on 10.30.177.12 The irqsoff tracer is rarely enabled in production systems due to the non-negligible overhead it introduces=E2=80=94even when unused. This is cau= sed by how trace_hardirqs_on/off() are always invoked in local_irq_enable/disable(), evaluate the tracepoint static key. This patch reduces the overhead in the common case where the tracepoint is disabled by performing the static key check earlier, avoiding the call to trace_hardirqs_on/off() entirely. This makes the impact of disabled preemptirq IRQ tracing negligible in performance-sensitive environments. Signed-off-by: Wander Lairson Costa Suggested-by: Steven Rostedt Cc: Arnaldo Carvalho de Melo Cc: Clark Williams Cc: Gabriele Monaco Cc: Juri Lelli --- include/linux/irqflags.h | 30 +++++++++++++++++++++--------- kernel/trace/trace_preemptirq.c | 3 +++ 2 files changed, 24 insertions(+), 9 deletions(-) diff --git a/include/linux/irqflags.h b/include/linux/irqflags.h index 57b074e0cfbb..40e456fa3d10 100644 --- a/include/linux/irqflags.h +++ b/include/linux/irqflags.h @@ -17,6 +17,7 @@ #include #include #include +#include =20 struct task_struct; =20 @@ -197,9 +198,17 @@ extern void warn_bogus_irq_restore(void); */ #ifdef CONFIG_TRACE_IRQFLAGS =20 +DECLARE_TRACEPOINT(irq_enable); +DECLARE_TRACEPOINT(irq_disable); + +#define __trace_enabled(tp) \ + (IS_ENABLED(CONFIG_PROVE_LOCKING) || \ + tracepoint_enabled(tp)) + #define local_irq_enable() \ do { \ - trace_hardirqs_on(); \ + if (__trace_enabled(irq_enable)) \ + trace_hardirqs_on(); \ raw_local_irq_enable(); \ } while (0) =20 @@ -207,31 +216,34 @@ extern void warn_bogus_irq_restore(void); do { \ bool was_disabled =3D raw_irqs_disabled();\ raw_local_irq_disable(); \ - if (!was_disabled) \ + if (__trace_enabled(irq_disable) && \ + !was_disabled) \ trace_hardirqs_off(); \ } while (0) =20 #define local_irq_save(flags) \ do { \ raw_local_irq_save(flags); \ - if (!raw_irqs_disabled_flags(flags)) \ + if (__trace_enabled(irq_disable) && \ + !raw_irqs_disabled_flags(flags)) \ trace_hardirqs_off(); \ } while (0) =20 #define local_irq_restore(flags) \ do { \ - if (!raw_irqs_disabled_flags(flags)) \ + if (__trace_enabled(irq_enable) && \ + !raw_irqs_disabled_flags(flags)) \ trace_hardirqs_on(); \ raw_local_irq_restore(flags); \ } while (0) =20 -#define safe_halt() \ - do { \ - trace_hardirqs_on(); \ - raw_safe_halt(); \ +#define safe_halt() \ + do { \ + if (__trace_enabled(irq_enable)) \ + trace_hardirqs_on(); \ + raw_safe_halt(); \ } while (0) =20 - #else /* !CONFIG_TRACE_IRQFLAGS */ =20 #define local_irq_enable() do { raw_local_irq_enable(); } while (0) diff --git a/kernel/trace/trace_preemptirq.c b/kernel/trace/trace_preemptir= q.c index 0c42b15c3800..90ee65db4516 100644 --- a/kernel/trace/trace_preemptirq.c +++ b/kernel/trace/trace_preemptirq.c @@ -111,6 +111,9 @@ void trace_hardirqs_off(void) } EXPORT_SYMBOL(trace_hardirqs_off); NOKPROBE_SYMBOL(trace_hardirqs_off); + +EXPORT_TRACEPOINT_SYMBOL(irq_disable); +EXPORT_TRACEPOINT_SYMBOL(irq_enable); #endif /* CONFIG_TRACE_IRQFLAGS */ =20 #ifdef CONFIG_TRACE_PREEMPT_TOGGLE --=20 2.50.0 From nobody Wed Oct 8 07:35:04 2025 Received: from us-smtp-delivery-124.mimecast.com (us-smtp-delivery-124.mimecast.com [170.10.129.124]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id D77362857F8 for ; Mon, 30 Jun 2025 19:53:40 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=170.10.129.124 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1751313223; cv=none; b=pCnLLbgkBnH2hHdjCXBrqsxDgEhFDQC+1Np6VZRXp08aFcOQvdFzVYcOxr3tEpTY9B3PKmuFwGUgT62eP+U4uwVLZ+wwczo+QdrGFT2vDxFUnJN8hs53VtB15znIv0Hi8jiRGwWJZtLsaQUTubMPslJ4Iqc1CkxndOlM12S1bBk= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1751313223; c=relaxed/simple; bh=pb209bXaugEvQjg2MuIrZK2EF+7xdDuFhOeXfAta4ww=; h=From:To:Cc:Subject:Date:Message-ID:In-Reply-To:References: MIME-Version; b=IOZS2cpYb5yKcQ81g0LPPdzxUEN8KEnciivxXHE/lgGprXpnRa3XK/qGbtH7Ws4QY8kgBezJPvWzaT1+alUIKVhzvX/em5Fxj5uO+oJ+k1zvCGIEU62l8UOpgdXJ2O9XlE/n8MqxCoSkWS3Aa9YguzNGXYnwILdFmaMRgrZ2p9k= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=pass (p=quarantine dis=none) header.from=redhat.com; spf=pass smtp.mailfrom=redhat.com; dkim=pass (1024-bit key) header.d=redhat.com header.i=@redhat.com header.b=GKlekG/6; arc=none smtp.client-ip=170.10.129.124 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=quarantine dis=none) header.from=redhat.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=redhat.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (1024-bit key) header.d=redhat.com header.i=@redhat.com header.b="GKlekG/6" DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=redhat.com; s=mimecast20190719; t=1751313219; h=from:from:reply-to:subject:subject:date:date:message-id:message-id: to:to:cc:cc:mime-version:mime-version: content-transfer-encoding:content-transfer-encoding: in-reply-to:in-reply-to:references:references; bh=4Fb7vbgp9yWJFAfr2LZ2H/YK07/bv8dqvyUraYZjnSI=; b=GKlekG/6s7tJTyT5k2SAp0cP3Oxj9f7U96iPkJQgLVrjnKbz+msbI5pkiALmpxwt/c5/0+ sMLkIgkvdePu6EP95faDGQsMTVE30/9mucWOEw5Wdzp16VHil87r7vvV/UvWZ+1T68GXJY kE8aimnmVOL4PtlGD8kBgRxobiSvMlg= Received: from mx-prod-mc-06.mail-002.prod.us-west-2.aws.redhat.com (ec2-35-165-154-97.us-west-2.compute.amazonaws.com [35.165.154.97]) by relay.mimecast.com with ESMTP with STARTTLS (version=TLSv1.3, cipher=TLS_AES_256_GCM_SHA384) id us-mta-635-vfYpM8YzPHGCdd--bkl8OQ-1; Mon, 30 Jun 2025 15:53:36 -0400 X-MC-Unique: vfYpM8YzPHGCdd--bkl8OQ-1 X-Mimecast-MFC-AGG-ID: vfYpM8YzPHGCdd--bkl8OQ_1751313214 Received: from mx-prod-int-03.mail-002.prod.us-west-2.aws.redhat.com (mx-prod-int-03.mail-002.prod.us-west-2.aws.redhat.com [10.30.177.12]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature RSA-PSS (2048 bits) server-digest SHA256) (No client certificate requested) by mx-prod-mc-06.mail-002.prod.us-west-2.aws.redhat.com (Postfix) with ESMTPS id 03B6A180029F; Mon, 30 Jun 2025 19:53:34 +0000 (UTC) Received: from wcosta-thinkpadt14gen4.rmtbr.csb (unknown [10.22.65.29]) by mx-prod-int-03.mail-002.prod.us-west-2.aws.redhat.com (Postfix) with ESMTP id BF03719560A7; Mon, 30 Jun 2025 19:53:26 +0000 (UTC) From: Wander Lairson Costa To: Ingo Molnar , Peter Zijlstra , Juri Lelli , Vincent Guittot , Dietmar Eggemann , Steven Rostedt , Ben Segall , Mel Gorman , Valentin Schneider , Masami Hiramatsu , Mathieu Desnoyers , Thomas Gleixner , Boqun Feng , David Woodhouse , Wander Lairson Costa , linux-kernel@vger.kernel.org (open list), linux-trace-kernel@vger.kernel.org (open list:TRACING) Cc: Arnaldo Carvalho de Melo , Clark Williams , Gabriele Monaco Subject: [PATCH v2 2/2] tracing/preemptirq: Optimize preempt_disable/enable() tracepoint overhead Date: Mon, 30 Jun 2025 16:52:39 -0300 Message-ID: <20250630195243.701516-3-wander@redhat.com> In-Reply-To: <20250630195243.701516-1-wander@redhat.com> References: <20250630195243.701516-1-wander@redhat.com> Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable X-Scanned-By: MIMEDefang 3.0 on 10.30.177.12 Content-Type: text/plain; charset="utf-8" Similar to the IRQ tracepoint, the preempt tracepoints are typically disabled in production systems due to the significant overhead they introduce even when not in use. The overhead primarily comes from two sources: First, when tracepoints are compiled into the kernel, preempt_count_add() and preempt_count_sub() become external function calls rather than inlined operations. Second, these functions perform unnecessary preempt_count() checks even when the tracepoint itself is disabled. This optimization introduces an early check of the tracepoint static key, which allows us to skip both the function call overhead and the redundant preempt_count() checks when tracing is disabled. The change maintains all existing functionality when tracing is active while significantly reducing overhead for the common case where tracing is inactive. Signed-off-by: Wander Lairson Costa Suggested-by: Steven Rostedt Cc: Arnaldo Carvalho de Melo Cc: Clark Williams Cc: Gabriele Monaco Cc: Juri Lelli --- include/linux/preempt.h | 35 ++++++++++++++++++++++++++++++--- kernel/sched/core.c | 12 +---------- kernel/trace/trace_preemptirq.c | 19 ++++++++++++++++++ 3 files changed, 52 insertions(+), 14 deletions(-) diff --git a/include/linux/preempt.h b/include/linux/preempt.h index b0af8d4ef6e6..d13c755cd934 100644 --- a/include/linux/preempt.h +++ b/include/linux/preempt.h @@ -10,6 +10,7 @@ #include #include #include +#include =20 /* * We put the hardirq and softirq counter into the preemption @@ -191,17 +192,45 @@ static __always_inline unsigned char interrupt_contex= t_level(void) */ #define in_atomic_preempt_off() (preempt_count() !=3D PREEMPT_DISABLE_OFFS= ET) =20 -#if defined(CONFIG_DEBUG_PREEMPT) || defined(CONFIG_TRACE_PREEMPT_TOGGLE) +#if defined(CONFIG_DEBUG_PREEMPT) extern void preempt_count_add(int val); extern void preempt_count_sub(int val); -#define preempt_count_dec_and_test() \ - ({ preempt_count_sub(1); should_resched(0); }) +#elif defined(CONFIG_TRACE_PREEMPT_TOGGLE) +extern void __trace_preempt_on(void); +extern void __trace_preempt_off(void); + +DECLARE_TRACEPOINT(preempt_enable); +DECLARE_TRACEPOINT(preempt_disable); + +#define __preempt_trace_enabled(type) \ + (tracepoint_enabled(preempt_##type) && preempt_count() =3D=3D val) + +static inline void preempt_count_add(int val) +{ + __preempt_count_add(val); + + if (__preempt_trace_enabled(disable)) + __trace_preempt_off(); +} + +static inline void preempt_count_sub(int val) +{ + if (__preempt_trace_enabled(enable)) + __trace_preempt_on(); + + __preempt_count_sub(val); +} #else #define preempt_count_add(val) __preempt_count_add(val) #define preempt_count_sub(val) __preempt_count_sub(val) #define preempt_count_dec_and_test() __preempt_count_dec_and_test() #endif =20 +#if defined(CONFIG_DEBUG_PREEMPT) || defined(CONFIG_TRACE_PREEMPT_TOGGLE) +#define preempt_count_dec_and_test() \ + ({ preempt_count_sub(1); should_resched(0); }) +#endif + #define __preempt_count_inc() __preempt_count_add(1) #define __preempt_count_dec() __preempt_count_sub(1) =20 diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 8988d38d46a3..4feba4738d79 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -5840,8 +5840,7 @@ static inline void sched_tick_start(int cpu) { } static inline void sched_tick_stop(int cpu) { } #endif =20 -#if defined(CONFIG_PREEMPTION) && (defined(CONFIG_DEBUG_PREEMPT) || \ - defined(CONFIG_TRACE_PREEMPT_TOGGLE)) +#if defined(CONFIG_PREEMPTION) && defined(CONFIG_DEBUG_PREEMPT) /* * If the value passed in is equal to the current preempt count * then we just disabled preemption. Start timing the latency. @@ -5850,30 +5849,24 @@ static inline void preempt_latency_start(int val) { if (preempt_count() =3D=3D val) { unsigned long ip =3D get_lock_parent_ip(); -#ifdef CONFIG_DEBUG_PREEMPT current->preempt_disable_ip =3D ip; -#endif trace_preempt_off(CALLER_ADDR0, ip); } } =20 void preempt_count_add(int val) { -#ifdef CONFIG_DEBUG_PREEMPT /* * Underflow? */ if (DEBUG_LOCKS_WARN_ON((preempt_count() < 0))) return; -#endif __preempt_count_add(val); -#ifdef CONFIG_DEBUG_PREEMPT /* * Spinlock count overflowing soon? */ DEBUG_LOCKS_WARN_ON((preempt_count() & PREEMPT_MASK) >=3D PREEMPT_MASK - 10); -#endif preempt_latency_start(val); } EXPORT_SYMBOL(preempt_count_add); @@ -5891,7 +5884,6 @@ static inline void preempt_latency_stop(int val) =20 void preempt_count_sub(int val) { -#ifdef CONFIG_DEBUG_PREEMPT /* * Underflow? */ @@ -5903,14 +5895,12 @@ void preempt_count_sub(int val) if (DEBUG_LOCKS_WARN_ON((val < PREEMPT_MASK) && !(preempt_count() & PREEMPT_MASK))) return; -#endif =20 preempt_latency_stop(val); __preempt_count_sub(val); } EXPORT_SYMBOL(preempt_count_sub); NOKPROBE_SYMBOL(preempt_count_sub); - #else static inline void preempt_latency_start(int val) { } static inline void preempt_latency_stop(int val) { } diff --git a/kernel/trace/trace_preemptirq.c b/kernel/trace/trace_preemptir= q.c index 90ee65db4516..deb2428b34a2 100644 --- a/kernel/trace/trace_preemptirq.c +++ b/kernel/trace/trace_preemptirq.c @@ -118,6 +118,25 @@ EXPORT_TRACEPOINT_SYMBOL(irq_enable); =20 #ifdef CONFIG_TRACE_PREEMPT_TOGGLE =20 +#if !defined(CONFIG_DEBUG_PREEMPT) +EXPORT_SYMBOL(__tracepoint_preempt_disable); +EXPORT_SYMBOL(__tracepoint_preempt_enable); + +void __trace_preempt_on(void) +{ + trace_preempt_on(CALLER_ADDR0, get_lock_parent_ip()); +} +EXPORT_SYMBOL(__trace_preempt_on); +NOKPROBE_SYMBOL(__trace_preempt_on); + +void __trace_preempt_off(void) +{ + trace_preempt_off(CALLER_ADDR0, get_lock_parent_ip()); +} +EXPORT_SYMBOL(__trace_preempt_off); +NOKPROBE_SYMBOL(__trace_preempt_off); +#endif /* !CONFIG_DEBUG_PREEMPT */ + void trace_preempt_on(unsigned long a0, unsigned long a1) { trace(preempt_enable, TP_ARGS(a0, a1)); --=20 2.50.0