All of lore.kernel.org
 help / color / mirror / Atom feed
From: Steven Rostedt <rostedt@goodmis.org>
To: linux-kernel@vger.kernel.org, linux-trace-kernel@vger.kernel.org,
	bpf@vger.kernel.org, x86@kernel.org
Cc: Masami Hiramatsu <mhiramat@kernel.org>,
	Mathieu Desnoyers <mathieu.desnoyers@efficios.com>,
	Josh Poimboeuf <jpoimboe@kernel.org>,
	Peter Zijlstra <peterz@infradead.org>,
	Ingo Molnar <mingo@kernel.org>, Jiri Olsa <jolsa@kernel.org>,
	Namhyung Kim <namhyung@kernel.org>,
	Thomas Gleixner <tglx@linutronix.de>,
	Andrii Nakryiko <andrii@kernel.org>,
	Indu Bhagat <indu.bhagat@oracle.com>,
	"Jose E. Marchesi" <jemarch@gnu.org>,
	Beau Belgrave <beaub@linux.microsoft.com>,
	Jens Remus <jremus@linux.ibm.com>,
	Linus Torvalds <torvalds@linux-foundation.org>,
	Andrew Morton <akpm@linux-foundation.org>
Subject: [PATCH v10 09/14] unwind deferred: Use SRCU unwind_deferred_task_work()
Date: Tue, 10 Jun 2025 20:54:30 -0400	[thread overview]
Message-ID: <20250611010429.274682576@goodmis.org> (raw)
In-Reply-To: 20250611005421.144238328@goodmis.org

From: Steven Rostedt <rostedt@goodmis.org>

Instead of using the callback_mutex to protect the link list of callbacks
in unwind_deferred_task_work(), use SRCU instead. This gets called every
time a task exits that has to record a stack trace that was requested.
This can happen for many tasks on several CPUs at the same time. A mutex
is a bottleneck and can cause a bit of contention and slow down performance.

As the callbacks themselves are allowed to sleep, regular RCU cannot be
used to protect the list. Instead use SRCU, as that still allows the
callbacks to sleep and the list can be read without needing to hold the
callback_mutex.

Link: https://lore.kernel.org/all/ca9bd83a-6c80-4ee0-a83c-224b9d60b755@efficios.com/

Suggested-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
---
 kernel/unwind/deferred.c | 33 +++++++++++++++++++++++++--------
 1 file changed, 25 insertions(+), 8 deletions(-)

diff --git a/kernel/unwind/deferred.c b/kernel/unwind/deferred.c
index 268afae31ba4..e44538a2be6c 100644
--- a/kernel/unwind/deferred.c
+++ b/kernel/unwind/deferred.c
@@ -13,10 +13,11 @@
 
 #define UNWIND_MAX_ENTRIES 512
 
-/* Guards adding to and reading the list of callbacks */
+/* Guards adding to or removing from the list of callbacks */
 static DEFINE_MUTEX(callback_mutex);
 static LIST_HEAD(callbacks);
 static unsigned long unwind_mask;
+DEFINE_STATIC_SRCU(unwind_srcu);
 
 /*
  * Read the task context timestamp, if this is the first caller then
@@ -107,6 +108,7 @@ static void unwind_deferred_task_work(struct callback_head *head)
 	struct unwind_stacktrace trace;
 	struct unwind_work *work;
 	u64 timestamp;
+	int idx;
 
 	if (WARN_ON_ONCE(!info->pending))
 		return;
@@ -132,13 +134,15 @@ static void unwind_deferred_task_work(struct callback_head *head)
 
 	timestamp = info->timestamp;
 
-	guard(mutex)(&callback_mutex);
-	list_for_each_entry(work, &callbacks, list) {
+	idx = srcu_read_lock(&unwind_srcu);
+	list_for_each_entry_srcu(work, &callbacks, list,
+				 srcu_read_lock_held(&unwind_srcu)) {
 		if (info->unwind_mask & BIT(work->bit)) {
 			work->func(work, &trace, timestamp);
 			clear_bit(work->bit, &info->unwind_mask);
 		}
 	}
+	srcu_read_unlock(&unwind_srcu, idx);
 }
 
 static int unwind_deferred_request_nmi(struct unwind_work *work, u64 *timestamp)
@@ -215,6 +219,7 @@ int unwind_deferred_request(struct unwind_work *work, u64 *timestamp)
 {
 	struct unwind_task_info *info = &current->unwind_info;
 	int pending;
+	int bit;
 	int ret;
 
 	*timestamp = 0;
@@ -226,12 +231,17 @@ int unwind_deferred_request(struct unwind_work *work, u64 *timestamp)
 	if (in_nmi())
 		return unwind_deferred_request_nmi(work, timestamp);
 
+	/* Do not allow cancelled works to request again */
+	bit = READ_ONCE(work->bit);
+	if (WARN_ON_ONCE(bit < 0))
+		return -EINVAL;
+
 	guard(irqsave)();
 
 	*timestamp = get_timestamp(info);
 
 	/* This is already queued */
-	if (info->unwind_mask & BIT(work->bit))
+	if (info->unwind_mask & BIT(bit))
 		return 1;
 
 	/* callback already pending? */
@@ -257,19 +267,26 @@ int unwind_deferred_request(struct unwind_work *work, u64 *timestamp)
 void unwind_deferred_cancel(struct unwind_work *work)
 {
 	struct task_struct *g, *t;
+	int bit;
 
 	if (!work)
 		return;
 
 	guard(mutex)(&callback_mutex);
-	list_del(&work->list);
+	list_del_rcu(&work->list);
+	bit = work->bit;
+
+	/* Do not allow any more requests and prevent callbacks */
+	work->bit = -1;
+
+	clear_bit(bit, &unwind_mask);
 
-	clear_bit(work->bit, &unwind_mask);
+	synchronize_srcu(&unwind_srcu);
 
 	guard(rcu)();
 	/* Clear this bit from all threads */
 	for_each_process_thread(g, t) {
-		clear_bit(work->bit, &t->unwind_info.unwind_mask);
+		clear_bit(bit, &t->unwind_info.unwind_mask);
 	}
 }
 
@@ -286,7 +303,7 @@ int unwind_deferred_init(struct unwind_work *work, unwind_callback_t func)
 	work->bit = ffz(unwind_mask);
 	unwind_mask |= BIT(work->bit);
 
-	list_add(&work->list, &callbacks);
+	list_add_rcu(&work->list, &callbacks);
 	work->func = func;
 	return 0;
 }
-- 
2.47.2



  parent reply	other threads:[~2025-06-11  1:03 UTC|newest]

Thread overview: 79+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-06-11  0:54 [PATCH v10 00/14] unwind_user: x86: Deferred unwinding infrastructure Steven Rostedt
2025-06-11  0:54 ` [PATCH v10 01/14] unwind_user: Add user space unwinding API Steven Rostedt
2025-06-18 13:49   ` Peter Zijlstra
2025-06-11  0:54 ` [PATCH v10 02/14] unwind_user: Add frame pointer support Steven Rostedt
2025-06-18 13:52   ` Peter Zijlstra
2025-06-18 15:09     ` Steven Rostedt
2025-06-23 16:31       ` Indu Bhagat
2025-06-24 20:30         ` Steven Rostedt
2025-06-11  0:54 ` [PATCH v10 03/14] unwind_user: Add compat mode " Steven Rostedt
2025-06-18 13:46   ` Peter Zijlstra
2025-06-18 15:10     ` Steven Rostedt
2025-06-18 17:52       ` Linus Torvalds
2025-06-18 18:37         ` Steven Rostedt
2025-06-18 13:47   ` Peter Zijlstra
2025-06-18 15:18     ` Steven Rostedt
2025-06-19  7:51       ` Peter Zijlstra
2025-06-19  8:39         ` Steven Rostedt
2025-06-11  0:54 ` [PATCH v10 04/14] unwind_user/deferred: Add unwind_deferred_trace() Steven Rostedt
2025-06-18 13:54   ` Peter Zijlstra
2025-06-18 13:59   ` Peter Zijlstra
2025-06-18 15:20     ` Steven Rostedt
2025-06-18 16:26       ` Steven Rostedt
2025-06-18 14:01   ` Peter Zijlstra
2025-06-18 15:23     ` Steven Rostedt
2025-06-18 14:02   ` Peter Zijlstra
2025-06-18 15:29     ` Steven Rostedt
2025-06-19  7:54       ` Peter Zijlstra
2025-06-19  8:44         ` Steven Rostedt
2025-06-11  0:54 ` [PATCH v10 05/14] unwind_user/deferred: Add unwind cache Steven Rostedt
2025-06-18 14:13   ` Peter Zijlstra
2025-06-18 15:33     ` Steven Rostedt
2025-06-19  7:56       ` Peter Zijlstra
2025-06-19  8:47         ` Steven Rostedt
2025-06-19  9:04           ` Peter Zijlstra
2025-06-19  9:12             ` Steven Rostedt
2025-06-11  0:54 ` [PATCH v10 06/14] unwind_user/deferred: Add deferred unwinding interface Steven Rostedt
2025-06-18 14:20   ` Peter Zijlstra
2025-06-18 15:37     ` Steven Rostedt
2025-06-18 15:38       ` Steven Rostedt
2025-06-19  8:01       ` Peter Zijlstra
2025-06-19  8:49         ` Steven Rostedt
2025-06-18 18:46   ` Peter Zijlstra
2025-06-18 19:09     ` Steven Rostedt
2025-06-18 19:36       ` Steven Rostedt
2025-06-19  7:50       ` Peter Zijlstra
2025-06-19  8:56         ` Steven Rostedt
2025-06-19  9:11           ` Peter Zijlstra
2025-06-24 14:03             ` Steven Rostedt
2025-06-24 22:36     ` Steven Rostedt
2025-06-11  0:54 ` [PATCH v10 07/14] unwind_user/deferred: Make unwind deferral requests NMI-safe Steven Rostedt
2025-06-19  8:34   ` Peter Zijlstra
2025-06-19  8:37     ` Steven Rostedt
2025-06-19  8:44       ` Peter Zijlstra
2025-06-19  8:48         ` Peter Zijlstra
2025-06-19  9:10           ` Steven Rostedt
2025-06-19  9:24             ` Peter Zijlstra
2025-06-19  8:57   ` Peter Zijlstra
2025-06-19  9:07     ` Steven Rostedt
2025-06-19  9:32       ` Peter Zijlstra
2025-06-19  9:34         ` Peter Zijlstra
2025-06-19  9:42         ` Steven Rostedt
2025-06-19  9:45           ` Peter Zijlstra
2025-06-19 10:19             ` Steven Rostedt
2025-06-19 10:39               ` Peter Zijlstra
2025-06-19 13:04                 ` Steven Rostedt
2025-06-11  0:54 ` [PATCH v10 08/14] unwind deferred: Use bitmask to determine which callbacks to call Steven Rostedt
2025-06-20  8:15   ` Peter Zijlstra
2025-06-24 14:55     ` Steven Rostedt
2025-06-24 15:00       ` Peter Zijlstra
2025-06-24 16:36         ` Steven Rostedt
2025-06-11  0:54 ` Steven Rostedt [this message]
2025-06-11  0:54 ` [PATCH v10 10/14] unwind: Clear unwind_mask on exit back to user space Steven Rostedt
2025-06-11  0:54 ` [PATCH v10 11/14] unwind: Finish up unwind when a task exits Steven Rostedt
2025-06-11  0:54 ` [PATCH v10 12/14] unwind_user/x86: Enable frame pointer unwinding on x86 Steven Rostedt
2025-06-11  0:54 ` [PATCH v10 13/14] perf/x86: Rename and move get_segment_base() and make it global Steven Rostedt
2025-06-11  0:54 ` [PATCH v10 14/14] unwind_user/x86: Enable compat mode frame pointer unwinding on x86 Steven Rostedt
2025-06-12 21:44 ` [PATCH v10 00/14] unwind_user: x86: Deferred unwinding infrastructure Andrii Nakryiko
2025-06-12 22:02   ` Josh Poimboeuf
2025-06-12 23:30     ` Andrii Nakryiko

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=20250611010429.274682576@goodmis.org \
    --to=rostedt@goodmis.org \
    --cc=akpm@linux-foundation.org \
    --cc=andrii@kernel.org \
    --cc=beaub@linux.microsoft.com \
    --cc=bpf@vger.kernel.org \
    --cc=indu.bhagat@oracle.com \
    --cc=jemarch@gnu.org \
    --cc=jolsa@kernel.org \
    --cc=jpoimboe@kernel.org \
    --cc=jremus@linux.ibm.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-trace-kernel@vger.kernel.org \
    --cc=mathieu.desnoyers@efficios.com \
    --cc=mhiramat@kernel.org \
    --cc=mingo@kernel.org \
    --cc=namhyung@kernel.org \
    --cc=peterz@infradead.org \
    --cc=tglx@linutronix.de \
    --cc=torvalds@linux-foundation.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.