linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Lance Yang <ioworker0@gmail.com>
To: akpm@linux-foundation.org
Cc: zi.li@linux.dev, anna.schumaker@oracle.com, boqun.feng@gmail.com,
	joel.granados@kernel.org, jstultz@google.com,
	kent.overstreet@linux.dev, leonylgao@tencent.com,
	linux-kernel@vger.kernel.org, longman@redhat.com,
	mhiramat@kernel.org, mingo@redhat.com, mingzhe.yang@ly.com,
	peterz@infradead.org, rostedt@goodmis.org,
	senozhatsky@chromium.org, tfiga@chromium.org, will@kernel.org
Subject: [PATCH 3/3] samples: enhance hung_task detector test with read-write semaphore support
Date: Fri, 27 Jun 2025 15:29:24 +0800	[thread overview]
Message-ID: <20250627072924.36567-4-lance.yang@linux.dev> (raw)
In-Reply-To: <20250627072924.36567-1-lance.yang@linux.dev>

From: Zi Li <zi.li@linux.dev>

Extend the hung_task detector test module to include read-write semaphore
support alongside existing mutex and semaphore tests. This module now
creates additional debugfs files under <debugfs>/hung_task, namely
'rw_semaphore_read' and 'rw_semaphore_write', in addition to 'mutex' and
'semaphore'. Reading these files with multiple processes triggers a
prolonged sleep (256 seconds) while holding the respective lock, enabling
hung_task detector testing for various locking mechanisms.

This change builds on the extensible hung_task_tests module, adding
read-write semaphore functionality to improve test coverage for kernel
locking primitives. The implementation ensures proper lock handling and
includes checks to prevent redundant data reads.

Usage is:

        > cd /sys/kernel/debug/hung_task
        > cat mutex & cat mutex           # Test mutex blocking
        > cat semaphore & cat semaphore   # Test semaphore blocking
        > cat rw_semaphore_write \
            & cat rw_semaphore_read       # Test rwsem blocking
        > cat rw_semaphore_write \
            & cat rw_semaphore_write      # Test rwsem blocking

Update the Kconfig description to reflect the addition of read-write
semaphore debugfs files.

Suggested-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
Signed-off-by: Zi Li <zi.li@linux.dev>
---
 samples/Kconfig                     |  7 ++-
 samples/hung_task/hung_task_tests.c | 81 ++++++++++++++++++++++++++---
 2 files changed, 77 insertions(+), 11 deletions(-)

diff --git a/samples/Kconfig b/samples/Kconfig
index ffef99950206..a8880c62d4c8 100644
--- a/samples/Kconfig
+++ b/samples/Kconfig
@@ -316,10 +316,9 @@ config SAMPLE_HUNG_TASK
 	depends on DETECT_HUNG_TASK && DEBUG_FS
 	help
 	  Build a module that provides debugfs files (e.g., mutex, semaphore,
-	  etc.) under <debugfs>/hung_task. If user reads one of these files,
-	  it will sleep long time (256 seconds) with holding a lock. Thus,
-	  if 2 or more processes read the same file concurrently, it will
-	  be detected by the hung_task watchdog.
+	  rw_semaphore_read, rw_semaphore_write) under <debugfs>/hung_task.
+	  Reading these files with multiple processes triggers hung task
+	  detection by holding locks for a long time (256 seconds).
 
 source "samples/rust/Kconfig"
 
diff --git a/samples/hung_task/hung_task_tests.c b/samples/hung_task/hung_task_tests.c
index a5c09bd3a47d..0360ec916890 100644
--- a/samples/hung_task/hung_task_tests.c
+++ b/samples/hung_task/hung_task_tests.c
@@ -4,11 +4,12 @@
  * semaphore, etc.
  *
  * Usage: Load this module and read `<debugfs>/hung_task/mutex`,
- *        `<debugfs>/hung_task/semaphore`, etc., with 2 or more processes.
+ *        `<debugfs>/hung_task/semaphore`, `<debugfs>/hung_task/rw_semaphore_read`,
+ *        `<debugfs>/hung_task/rw_semaphore_write`, etc., with 2 or more processes.
  *
  * This is for testing kernel hung_task error messages with various locking
- * mechanisms (e.g., mutex, semaphore, etc.). Note that this may freeze
- * your system or cause a panic. Use only for testing purposes.
+ * mechanisms (e.g., mutex, semaphore, rw_semaphore_read, rw_semaphore_write, etc.).
+ * Note that this may freeze your system or cause a panic. Use only for testing purposes.
  */
 
 #include <linux/debugfs.h>
@@ -17,21 +18,29 @@
 #include <linux/module.h>
 #include <linux/mutex.h>
 #include <linux/semaphore.h>
+#include <linux/rwsem.h>
 
-#define HUNG_TASK_DIR		"hung_task"
-#define HUNG_TASK_MUTEX_FILE	"mutex"
-#define HUNG_TASK_SEM_FILE	"semaphore"
-#define SLEEP_SECOND		256
+#define HUNG_TASK_DIR			"hung_task"
+#define HUNG_TASK_MUTEX_FILE		"mutex"
+#define HUNG_TASK_SEM_FILE		"semaphore"
+#define HUNG_TASK_RWSEM_READ_FILE	"rw_semaphore_read"
+#define HUNG_TASK_RWSEM_WRITE_FILE	"rw_semaphore_write"
+#define SLEEP_SECOND			256
 
 static const char dummy_string[] = "This is a dummy string.";
 static DEFINE_MUTEX(dummy_mutex);
 static DEFINE_SEMAPHORE(dummy_sem, 1);
+static DECLARE_RWSEM(dummy_rwsem);
 static struct dentry *hung_task_dir;
 
 /* Mutex-based read function */
 static ssize_t read_dummy_mutex(struct file *file, char __user *user_buf,
 				size_t count, loff_t *ppos)
 {
+	/* Check if data is already read */
+	if (*ppos >= sizeof(dummy_string))
+		return 0;
+
 	/* Second task waits on mutex, entering uninterruptible sleep */
 	guard(mutex)(&dummy_mutex);
 
@@ -46,6 +55,10 @@ static ssize_t read_dummy_mutex(struct file *file, char __user *user_buf,
 static ssize_t read_dummy_semaphore(struct file *file, char __user *user_buf,
 				    size_t count, loff_t *ppos)
 {
+	/* Check if data is already read */
+	if (*ppos >= sizeof(dummy_string))
+		return 0;
+
 	/* Second task waits on semaphore, entering uninterruptible sleep */
 	down(&dummy_sem);
 
@@ -58,6 +71,46 @@ static ssize_t read_dummy_semaphore(struct file *file, char __user *user_buf,
 				       sizeof(dummy_string));
 }
 
+/* Read-write semaphore read function */
+static ssize_t read_dummy_rwsem_read(struct file *file, char __user *user_buf,
+				     size_t count, loff_t *ppos)
+{
+	/* Check if data is already read */
+	if (*ppos >= sizeof(dummy_string))
+		return 0;
+
+	/* Acquires read lock, allowing concurrent readers but blocks if write lock is held */
+	down_read(&dummy_rwsem);
+
+	/* Sleeps here, potentially triggering hung task detection if lock is held too long */
+	msleep_interruptible(SLEEP_SECOND * 1000);
+
+	up_read(&dummy_rwsem);
+
+	return simple_read_from_buffer(user_buf, count, ppos, dummy_string,
+				       sizeof(dummy_string));
+}
+
+/* Read-write semaphore write function */
+static ssize_t read_dummy_rwsem_write(struct file *file, char __user *user_buf,
+				      size_t count, loff_t *ppos)
+{
+	/* Check if data is already read */
+	if (*ppos >= sizeof(dummy_string))
+		return 0;
+
+	/* Acquires exclusive write lock, blocking all other readers and writers */
+	down_write(&dummy_rwsem);
+
+	/* Sleeps here, potentially triggering hung task detection if lock is held too long */
+	msleep_interruptible(SLEEP_SECOND * 1000);
+
+	up_write(&dummy_rwsem);
+
+	return simple_read_from_buffer(user_buf, count, ppos, dummy_string,
+				       sizeof(dummy_string));
+}
+
 /* File operations for mutex */
 static const struct file_operations hung_task_mutex_fops = {
 	.read = read_dummy_mutex,
@@ -68,6 +121,16 @@ static const struct file_operations hung_task_sem_fops = {
 	.read = read_dummy_semaphore,
 };
 
+/* File operations for rw_semaphore read */
+static const struct file_operations hung_task_rwsem_read_fops = {
+	.read = read_dummy_rwsem_read,
+};
+
+/* File operations for rw_semaphore write */
+static const struct file_operations hung_task_rwsem_write_fops = {
+	.read = read_dummy_rwsem_write,
+};
+
 static int __init hung_task_tests_init(void)
 {
 	hung_task_dir = debugfs_create_dir(HUNG_TASK_DIR, NULL);
@@ -79,6 +142,10 @@ static int __init hung_task_tests_init(void)
 			    &hung_task_mutex_fops);
 	debugfs_create_file(HUNG_TASK_SEM_FILE, 0400, hung_task_dir, NULL,
 			    &hung_task_sem_fops);
+	debugfs_create_file(HUNG_TASK_RWSEM_READ_FILE, 0400, hung_task_dir, NULL,
+			    &hung_task_rwsem_read_fops);
+	debugfs_create_file(HUNG_TASK_RWSEM_WRITE_FILE, 0400, hung_task_dir, NULL,
+			    &hung_task_rwsem_write_fops);
 
 	return 0;
 }
-- 
2.49.0


      parent reply	other threads:[~2025-06-27  7:30 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-06-27  7:29 [PATCH 0/3] extend hung task blocker tracking to rwsems Lance Yang
2025-06-27  7:29 ` [PATCH 1/3] locking/rwsem: make owner helpers globally available Lance Yang
2025-06-27  7:29 ` [PATCH 2/3] hung_task: extend hung task blocker tracking to rwsems Lance Yang
2025-08-22  7:34   ` Geert Uytterhoeven
2025-08-22  7:37     ` Geert Uytterhoeven
2025-08-22  8:25     ` Sergey Senozhatsky
2025-06-27  7:29 ` Lance Yang [this message]

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=20250627072924.36567-4-lance.yang@linux.dev \
    --to=ioworker0@gmail.com \
    --cc=akpm@linux-foundation.org \
    --cc=anna.schumaker@oracle.com \
    --cc=boqun.feng@gmail.com \
    --cc=joel.granados@kernel.org \
    --cc=jstultz@google.com \
    --cc=kent.overstreet@linux.dev \
    --cc=leonylgao@tencent.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=longman@redhat.com \
    --cc=mhiramat@kernel.org \
    --cc=mingo@redhat.com \
    --cc=mingzhe.yang@ly.com \
    --cc=peterz@infradead.org \
    --cc=rostedt@goodmis.org \
    --cc=senozhatsky@chromium.org \
    --cc=tfiga@chromium.org \
    --cc=will@kernel.org \
    --cc=zi.li@linux.dev \
    /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;
as well as URLs for NNTP newsgroup(s).