Linux Trace Kernel
 help / color / mirror / Atom feed
From: Ran Xiaokai <ranxiaokai627@163.com>
To: rostedt@goodmis.org, mhiramat@kernel.org, mathieu.desnoyers@efficios.com
Cc: linux-kernel@vger.kernel.org, linux-trace-kernel@vger.kernel.org,
	ran.xiaokai@zte.com.cn, yang.guang5@zte.com.cn,
	Wang Yong <wang.yong12@zte.com.cn>
Subject: [PATCH] tracing/osnoise: Fix possible recursive locking for cpus_read_lock()
Date: Tue, 25 Feb 2025 12:31:32 +0000	[thread overview]
Message-ID: <20250225123132.2583820-1-ranxiaokai627@163.com> (raw)

From: Ran Xiaokai <ran.xiaokai@zte.com.cn>

Lockdep reports this deadlock log:
============================================
WARNING: possible recursive locking detected
--------------------------------------------
sh/31444 is trying to acquire lock:
ffffffff82c51af0 (cpu_hotplug_lock){++++}-{0:0}, at:
stop_per_cpu_kthreads+0x7/0x60

but task is already holding lock:
ffffffff82c51af0 (cpu_hotplug_lock){++++}-{0:0}, at:
start_per_cpu_kthreads+0x28/0x140

other info that might help us debug this:
 Possible unsafe locking scenario:

       CPU0
       ----
  lock(cpu_hotplug_lock);
  lock(cpu_hotplug_lock);

Call Trace:
 <TASK>
 __lock_acquire+0x1612/0x29b0
 lock_acquire+0xd0/0x2e0
 cpus_read_lock+0x49/0x120
 stop_per_cpu_kthreads+0x7/0x60
 start_kthread+0x105/0x120
 start_per_cpu_kthreads+0xdd/0x140
 osnoise_workload_start+0x261/0x2f0
 osnoise_tracer_start+0x18/0x4

In start_kthread(), when kthread_run_on_cpu() fails,
cpus_read_unlock() should be called before stop_per_cpu_kthreads(),
but both start_per_cpu_kthreads() and start_kthread() call the error
handling routine stop_per_cpu_kthreads(),
which is redundant. Only one call is necessary.
To fix this, move stop_per_cpu_kthreads() outside of start_kthread(),
use the return value of start_kthread() to determine kthread creation
error.
The same issue exists in osnoise_hotplug_workfn() too.

Reviewed-by: Yang Guang <yang.guang5@zte.com.cn>
Reviewed-by: Wang Yong <wang.yong12@zte.com.cn>
Signed-off-by: Ran Xiaokai <ran.xiaokai@zte.com.cn>
---
 kernel/trace/trace_osnoise.c | 10 +++++++---
 1 file changed, 7 insertions(+), 3 deletions(-)

diff --git a/kernel/trace/trace_osnoise.c b/kernel/trace/trace_osnoise.c
index 92e16f03fa4e..38fb0c655f5b 100644
--- a/kernel/trace/trace_osnoise.c
+++ b/kernel/trace/trace_osnoise.c
@@ -2029,7 +2029,6 @@ static int start_kthread(unsigned int cpu)
 
 	if (IS_ERR(kthread)) {
 		pr_err(BANNER "could not start sampling thread\n");
-		stop_per_cpu_kthreads();
 		return -ENOMEM;
 	}
 
@@ -2097,7 +2096,7 @@ static void osnoise_hotplug_workfn(struct
work_struct *dummy)
 		return;
 
 	guard(mutex)(&interface_lock);
-	guard(cpus_read_lock)();
+	cpus_read_lock();
 
 	if (!cpu_online(cpu))
 		return;
@@ -2105,7 +2104,12 @@ static void osnoise_hotplug_workfn(struct
work_struct *dummy)
 	if (!cpumask_test_cpu(cpu, &osnoise_cpumask))
 		return;
 
-	start_kthread(cpu);
+	if (start_kthread(cpu)) {
+		cpus_read_unlock();
+		stop_per_cpu_kthreads();
+		return;
+	}
+	cpus_read_unlock();
 }
 
 static DECLARE_WORK(osnoise_hotplug_work, osnoise_hotplug_workfn);
-- 
2.15.2


             reply	other threads:[~2025-02-25 12:32 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-02-25 12:31 Ran Xiaokai [this message]
2025-02-25 16:30 ` [PATCH] tracing/osnoise: Fix possible recursive locking for cpus_read_lock() Steven Rostedt
2025-02-26  3:42   ` Ran Xiaokai
2025-02-26 15:05     ` Steven Rostedt
2025-03-17 12:28       ` Ran Xiaokai
2025-02-27  7:40 ` Vishal Chourasia
2025-03-17 12:52   ` Ran Xiaokai

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=20250225123132.2583820-1-ranxiaokai627@163.com \
    --to=ranxiaokai627@163.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-trace-kernel@vger.kernel.org \
    --cc=mathieu.desnoyers@efficios.com \
    --cc=mhiramat@kernel.org \
    --cc=ran.xiaokai@zte.com.cn \
    --cc=rostedt@goodmis.org \
    --cc=wang.yong12@zte.com.cn \
    --cc=yang.guang5@zte.com.cn \
    /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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox