All of lore.kernel.org
 help / color / mirror / Atom feed
From: Chao Yu via Linux-f2fs-devel <linux-f2fs-devel@lists.sourceforge.net>
To: jaegeuk@kernel.org
Cc: linux-kernel@vger.kernel.org, linux-f2fs-devel@lists.sourceforge.net
Subject: [f2fs-dev] [PATCH] f2fs: fix to avoid invalid wait context issue
Date: Wed, 11 Jun 2025 16:42:18 +0800	[thread overview]
Message-ID: <20250611084218.62774-1-chao@kernel.org> (raw)

=============================
[ BUG: Invalid wait context ]
6.13.0-rc1 #84 Tainted: G           O
-----------------------------
cat/56160 is trying to lock:
ffff888105c86648 (&cprc->stat_lock){+.+.}-{3:3}, at: update_general_status+0x32a/0x8c0 [f2fs]
other info that might help us debug this:
context-{5:5}
2 locks held by cat/56160:
 #0: ffff88810a002a98 (&p->lock){+.+.}-{4:4}, at: seq_read_iter+0x56/0x4c0
 #1: ffffffffa0462638 (f2fs_stat_lock){....}-{2:2}, at: stat_show+0x29/0x1020 [f2fs]
stack backtrace:
CPU: 0 UID: 0 PID: 56160 Comm: cat Tainted: G           O       6.13.0-rc1 #84
Tainted: [O]=OOT_MODULE
Hardware name: innotek GmbH VirtualBox/VirtualBox, BIOS VirtualBox 12/01/2006
Call Trace:
 <TASK>
 dump_stack_lvl+0x88/0xd0
 dump_stack+0x14/0x20
 __lock_acquire+0x8d4/0xbb0
 lock_acquire+0xd6/0x300
 _raw_spin_lock+0x38/0x50
 update_general_status+0x32a/0x8c0 [f2fs]
 stat_show+0x50/0x1020 [f2fs]
 seq_read_iter+0x116/0x4c0
 seq_read+0xfa/0x130
 full_proxy_read+0x66/0x90
 vfs_read+0xc4/0x350
 ksys_read+0x74/0xf0
 __x64_sys_read+0x1d/0x20
 x64_sys_call+0x17d9/0x1b80
 do_syscall_64+0x68/0x130
 entry_SYSCALL_64_after_hwframe+0x67/0x6f
RIP: 0033:0x7f2ca53147e2

- seq_read
 - stat_show
  - raw_spin_lock_irqsave(&f2fs_stat_lock, flags)
  : f2fs_stat_lock is raw_spinlock_t type variable
  - update_general_status
   - spin_lock(&sbi->cprc_info.stat_lock);
   : stat_lock is spinlock_t type variable

The root cause is the lock order is incorrect [1], we should not acquire
spinlock_t lock after raw_spinlock_t lock, as if CONFIG_PREEMPT_LOCK is
on, spinlock_t is implemented based on rtmutex, which can sleep after
holding the lock.

To fix this issue, let's use change f2fs_stat_lock lock type from
raw_spinlock_t to spinlock_t, it's safe due to:
- we don't need to use raw version of spinlock as the path is not
performance sensitive.
- we don't need to use irqsave version of spinlock as it won't be
used in irq context.

Quoted from [1]:

"Extend lockdep to validate lock wait-type context.

The current wait-types are:

	LD_WAIT_FREE,		/* wait free, rcu etc.. */
	LD_WAIT_SPIN,		/* spin loops, raw_spinlock_t etc.. */
	LD_WAIT_CONFIG,		/* CONFIG_PREEMPT_LOCK, spinlock_t etc.. */
	LD_WAIT_SLEEP,		/* sleeping locks, mutex_t etc.. */

Where lockdep validates that the current lock (the one being acquired)
fits in the current wait-context (as generated by the held stack).

This ensures that there is no attempt to acquire mutexes while holding
spinlocks, to acquire spinlocks while holding raw_spinlocks and so on. In
other words, its a more fancy might_sleep()."

[1] https://lore.kernel.org/all/20200321113242.427089655@linutronix.de

Fixes: 98237fcda4a2 ("f2fs: use spin_lock to avoid hang")
Signed-off-by: Chao Yu <chao@kernel.org>
---
 fs/f2fs/debug.c | 17 +++++++----------
 1 file changed, 7 insertions(+), 10 deletions(-)

diff --git a/fs/f2fs/debug.c b/fs/f2fs/debug.c
index 16c2dfb4f595..3417e7e550b2 100644
--- a/fs/f2fs/debug.c
+++ b/fs/f2fs/debug.c
@@ -21,7 +21,7 @@
 #include "gc.h"
 
 static LIST_HEAD(f2fs_stat_list);
-static DEFINE_RAW_SPINLOCK(f2fs_stat_lock);
+static DEFINE_SPINLOCK(f2fs_stat_lock);
 #ifdef CONFIG_DEBUG_FS
 static struct dentry *f2fs_debugfs_root;
 #endif
@@ -439,9 +439,8 @@ static int stat_show(struct seq_file *s, void *v)
 {
 	struct f2fs_stat_info *si;
 	int i = 0, j = 0;
-	unsigned long flags;
 
-	raw_spin_lock_irqsave(&f2fs_stat_lock, flags);
+	spin_lock(&f2fs_stat_lock);
 	list_for_each_entry(si, &f2fs_stat_list, stat_list) {
 		struct f2fs_sb_info *sbi = si->sbi;
 
@@ -753,7 +752,7 @@ static int stat_show(struct seq_file *s, void *v)
 		seq_printf(s, "  - paged : %llu KB\n",
 				si->page_mem >> 10);
 	}
-	raw_spin_unlock_irqrestore(&f2fs_stat_lock, flags);
+	spin_unlock(&f2fs_stat_lock);
 	return 0;
 }
 
@@ -765,7 +764,6 @@ int f2fs_build_stats(struct f2fs_sb_info *sbi)
 	struct f2fs_super_block *raw_super = F2FS_RAW_SUPER(sbi);
 	struct f2fs_stat_info *si;
 	struct f2fs_dev_stats *dev_stats;
-	unsigned long flags;
 	int i;
 
 	si = f2fs_kzalloc(sbi, sizeof(struct f2fs_stat_info), GFP_KERNEL);
@@ -817,9 +815,9 @@ int f2fs_build_stats(struct f2fs_sb_info *sbi)
 
 	atomic_set(&sbi->max_aw_cnt, 0);
 
-	raw_spin_lock_irqsave(&f2fs_stat_lock, flags);
+	spin_lock(&f2fs_stat_lock);
 	list_add_tail(&si->stat_list, &f2fs_stat_list);
-	raw_spin_unlock_irqrestore(&f2fs_stat_lock, flags);
+	spin_unlock(&f2fs_stat_lock);
 
 	return 0;
 }
@@ -827,11 +825,10 @@ int f2fs_build_stats(struct f2fs_sb_info *sbi)
 void f2fs_destroy_stats(struct f2fs_sb_info *sbi)
 {
 	struct f2fs_stat_info *si = F2FS_STAT(sbi);
-	unsigned long flags;
 
-	raw_spin_lock_irqsave(&f2fs_stat_lock, flags);
+	spin_lock(&f2fs_stat_lock);
 	list_del(&si->stat_list);
-	raw_spin_unlock_irqrestore(&f2fs_stat_lock, flags);
+	spin_unlock(&f2fs_stat_lock);
 
 	kfree(si->dev_stats);
 	kfree(si);
-- 
2.49.0



_______________________________________________
Linux-f2fs-devel mailing list
Linux-f2fs-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel

WARNING: multiple messages have this Message-ID (diff)
From: Chao Yu <chao@kernel.org>
To: jaegeuk@kernel.org
Cc: linux-f2fs-devel@lists.sourceforge.net,
	linux-kernel@vger.kernel.org, Chao Yu <chao@kernel.org>
Subject: [PATCH] f2fs: fix to avoid invalid wait context issue
Date: Wed, 11 Jun 2025 16:42:18 +0800	[thread overview]
Message-ID: <20250611084218.62774-1-chao@kernel.org> (raw)

=============================
[ BUG: Invalid wait context ]
6.13.0-rc1 #84 Tainted: G           O
-----------------------------
cat/56160 is trying to lock:
ffff888105c86648 (&cprc->stat_lock){+.+.}-{3:3}, at: update_general_status+0x32a/0x8c0 [f2fs]
other info that might help us debug this:
context-{5:5}
2 locks held by cat/56160:
 #0: ffff88810a002a98 (&p->lock){+.+.}-{4:4}, at: seq_read_iter+0x56/0x4c0
 #1: ffffffffa0462638 (f2fs_stat_lock){....}-{2:2}, at: stat_show+0x29/0x1020 [f2fs]
stack backtrace:
CPU: 0 UID: 0 PID: 56160 Comm: cat Tainted: G           O       6.13.0-rc1 #84
Tainted: [O]=OOT_MODULE
Hardware name: innotek GmbH VirtualBox/VirtualBox, BIOS VirtualBox 12/01/2006
Call Trace:
 <TASK>
 dump_stack_lvl+0x88/0xd0
 dump_stack+0x14/0x20
 __lock_acquire+0x8d4/0xbb0
 lock_acquire+0xd6/0x300
 _raw_spin_lock+0x38/0x50
 update_general_status+0x32a/0x8c0 [f2fs]
 stat_show+0x50/0x1020 [f2fs]
 seq_read_iter+0x116/0x4c0
 seq_read+0xfa/0x130
 full_proxy_read+0x66/0x90
 vfs_read+0xc4/0x350
 ksys_read+0x74/0xf0
 __x64_sys_read+0x1d/0x20
 x64_sys_call+0x17d9/0x1b80
 do_syscall_64+0x68/0x130
 entry_SYSCALL_64_after_hwframe+0x67/0x6f
RIP: 0033:0x7f2ca53147e2

- seq_read
 - stat_show
  - raw_spin_lock_irqsave(&f2fs_stat_lock, flags)
  : f2fs_stat_lock is raw_spinlock_t type variable
  - update_general_status
   - spin_lock(&sbi->cprc_info.stat_lock);
   : stat_lock is spinlock_t type variable

The root cause is the lock order is incorrect [1], we should not acquire
spinlock_t lock after raw_spinlock_t lock, as if CONFIG_PREEMPT_LOCK is
on, spinlock_t is implemented based on rtmutex, which can sleep after
holding the lock.

To fix this issue, let's use change f2fs_stat_lock lock type from
raw_spinlock_t to spinlock_t, it's safe due to:
- we don't need to use raw version of spinlock as the path is not
performance sensitive.
- we don't need to use irqsave version of spinlock as it won't be
used in irq context.

Quoted from [1]:

"Extend lockdep to validate lock wait-type context.

The current wait-types are:

	LD_WAIT_FREE,		/* wait free, rcu etc.. */
	LD_WAIT_SPIN,		/* spin loops, raw_spinlock_t etc.. */
	LD_WAIT_CONFIG,		/* CONFIG_PREEMPT_LOCK, spinlock_t etc.. */
	LD_WAIT_SLEEP,		/* sleeping locks, mutex_t etc.. */

Where lockdep validates that the current lock (the one being acquired)
fits in the current wait-context (as generated by the held stack).

This ensures that there is no attempt to acquire mutexes while holding
spinlocks, to acquire spinlocks while holding raw_spinlocks and so on. In
other words, its a more fancy might_sleep()."

[1] https://lore.kernel.org/all/20200321113242.427089655@linutronix.de

Fixes: 98237fcda4a2 ("f2fs: use spin_lock to avoid hang")
Signed-off-by: Chao Yu <chao@kernel.org>
---
 fs/f2fs/debug.c | 17 +++++++----------
 1 file changed, 7 insertions(+), 10 deletions(-)

diff --git a/fs/f2fs/debug.c b/fs/f2fs/debug.c
index 16c2dfb4f595..3417e7e550b2 100644
--- a/fs/f2fs/debug.c
+++ b/fs/f2fs/debug.c
@@ -21,7 +21,7 @@
 #include "gc.h"
 
 static LIST_HEAD(f2fs_stat_list);
-static DEFINE_RAW_SPINLOCK(f2fs_stat_lock);
+static DEFINE_SPINLOCK(f2fs_stat_lock);
 #ifdef CONFIG_DEBUG_FS
 static struct dentry *f2fs_debugfs_root;
 #endif
@@ -439,9 +439,8 @@ static int stat_show(struct seq_file *s, void *v)
 {
 	struct f2fs_stat_info *si;
 	int i = 0, j = 0;
-	unsigned long flags;
 
-	raw_spin_lock_irqsave(&f2fs_stat_lock, flags);
+	spin_lock(&f2fs_stat_lock);
 	list_for_each_entry(si, &f2fs_stat_list, stat_list) {
 		struct f2fs_sb_info *sbi = si->sbi;
 
@@ -753,7 +752,7 @@ static int stat_show(struct seq_file *s, void *v)
 		seq_printf(s, "  - paged : %llu KB\n",
 				si->page_mem >> 10);
 	}
-	raw_spin_unlock_irqrestore(&f2fs_stat_lock, flags);
+	spin_unlock(&f2fs_stat_lock);
 	return 0;
 }
 
@@ -765,7 +764,6 @@ int f2fs_build_stats(struct f2fs_sb_info *sbi)
 	struct f2fs_super_block *raw_super = F2FS_RAW_SUPER(sbi);
 	struct f2fs_stat_info *si;
 	struct f2fs_dev_stats *dev_stats;
-	unsigned long flags;
 	int i;
 
 	si = f2fs_kzalloc(sbi, sizeof(struct f2fs_stat_info), GFP_KERNEL);
@@ -817,9 +815,9 @@ int f2fs_build_stats(struct f2fs_sb_info *sbi)
 
 	atomic_set(&sbi->max_aw_cnt, 0);
 
-	raw_spin_lock_irqsave(&f2fs_stat_lock, flags);
+	spin_lock(&f2fs_stat_lock);
 	list_add_tail(&si->stat_list, &f2fs_stat_list);
-	raw_spin_unlock_irqrestore(&f2fs_stat_lock, flags);
+	spin_unlock(&f2fs_stat_lock);
 
 	return 0;
 }
@@ -827,11 +825,10 @@ int f2fs_build_stats(struct f2fs_sb_info *sbi)
 void f2fs_destroy_stats(struct f2fs_sb_info *sbi)
 {
 	struct f2fs_stat_info *si = F2FS_STAT(sbi);
-	unsigned long flags;
 
-	raw_spin_lock_irqsave(&f2fs_stat_lock, flags);
+	spin_lock(&f2fs_stat_lock);
 	list_del(&si->stat_list);
-	raw_spin_unlock_irqrestore(&f2fs_stat_lock, flags);
+	spin_unlock(&f2fs_stat_lock);
 
 	kfree(si->dev_stats);
 	kfree(si);
-- 
2.49.0


             reply	other threads:[~2025-06-11  8:42 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-06-11  8:42 Chao Yu via Linux-f2fs-devel [this message]
2025-06-11  8:42 ` [PATCH] f2fs: fix to avoid invalid wait context issue Chao Yu
2025-06-24  0:41 ` [f2fs-dev] " patchwork-bot+f2fs--- via Linux-f2fs-devel
2025-06-24  0:41   ` patchwork-bot+f2fs

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=20250611084218.62774-1-chao@kernel.org \
    --to=linux-f2fs-devel@lists.sourceforge.net \
    --cc=chao@kernel.org \
    --cc=jaegeuk@kernel.org \
    --cc=linux-kernel@vger.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.