Skip to contentSkip to conversation

fscrypt: fix lockdep warning

PATCH In review Yu Zhao 2 participants 3 messages Oct 5, 2022 fscrypt: file system level encryption support

lore
Share & tools
lore

Yu Zhao

author
#0 lore raw

PATCH 1/1 fscrypt: fix lockdep warning

fscrypt_initialize() shouldn't allocate memory without GFP_NOFS.

The problem seems to go back to 2015
commit 57e5055b0a5e ("f2fs crypto: add f2fs encryption facilities") but I have never heard of any complaints, hence not CC'ing stable.

======================================================
WARNING: possible circular locking dependency detected 6.0.0-lockdep #1 Not tainted
------------------------------------------------------
kswapd0/77 is trying to acquire lock:
71ffff808b254a18 (jbd2_handle){++++}-{0:0}, at: start_this_handle+0x76c/0x8dc

but task is already holding lock:
ffffffea26533310 (fs_reclaim){+.+.}-{0:0}, at: 0x1

which lock already depends on the new lock.

<snipped>

other info that might help us debug this:

  Chain exists of:
    jbd2_handle --> fscrypt_init_mutex --> fs_reclaim

   Possible unsafe locking scenario:
         CPU0                    CPU1
         ----                    ----
    lock(fs_reclaim);
                                 lock(fscrypt_init_mutex);
                                 lock(fs_reclaim);
    lock(jbd2_handle);

*** DEADLOCK ***

3 locks held by kswapd0/77:
#0: ffffffea26533310 (fs_reclaim){+.+.}-{0:0}, at: 0x1
#1: ffffffea26529220 (shrinker_rwsem){++++}-{3:3}, at: shrink_slab+0x54/0x464 #2: 6dffff808abe90e8 (&type->s_umount_key#47){++++}-{3:3}, at: trylock_super+0x2c/0x8c

<snipped>

---
 fs/crypto/crypto.c | 3 +++
 1 file changed, 3 insertions(+)
fs/crypto/crypto.c +3 -0
index e78be66bbf01..e10fc30142a6 100644
@@ -316,6 +316,7 @@ EXPORT_SYMBOL(fscrypt_decrypt_block_inplace);
 int fscrypt_initialize(unsigned int cop_flags)
 {
 	int err = 0;
+	unsigned int flags;
 
 	/* No need to allocate a bounce page pool if this FS won't use it. */
 	if (cop_flags & FS_CFLG_OWN_PAGES)
@@ -326,8 +327,10 @@ int fscrypt_initialize(unsigned int cop_flags)
 		goto out_unlock;
 
 	err = -ENOMEM;
+	flags = memalloc_nofs_save();
 	fscrypt_bounce_page_pool =
 		mempool_create_page_pool(num_prealloc_crypto_pages, 0);
+	memalloc_nofs_restore(flags);
 	if (!fscrypt_bounce_page_pool)
 		goto out_unlock;
 
Signature · 1 line
2.38.0.rc1.362.ged0d419d3c-goog
maintainer
#1 lore raw
Quoted 61 lines · Yu Zhao

fscrypt_initialize() shouldn't allocate memory without GFP_NOFS.

The problem seems to go back to 2015
commit 57e5055b0a5e ("f2fs crypto: add f2fs encryption facilities") but I have never heard of any complaints, hence not CC'ing stable.

======================================================
WARNING: possible circular locking dependency detected 6.0.0-lockdep #1 Not tainted
------------------------------------------------------
kswapd0/77 is trying to acquire lock:
71ffff808b254a18 (jbd2_handle){++++}-{0:0}, at: start_this_handle+0x76c/0x8dc

but task is already holding lock:
ffffffea26533310 (fs_reclaim){+.+.}-{0:0}, at: 0x1

which lock already depends on the new lock.

<snipped>

other info that might help us debug this:

  Chain exists of:
    jbd2_handle --> fscrypt_init_mutex --> fs_reclaim

   Possible unsafe locking scenario:
         CPU0                    CPU1
         ----                    ----
    lock(fs_reclaim);
                                 lock(fscrypt_init_mutex);
                                 lock(fs_reclaim);
    lock(jbd2_handle);
   *** DEADLOCK ***

3 locks held by kswapd0/77:

   #0: ffffffea26533310 (fs_reclaim){+.+.}-{0:0}, at: 0x1
   #1: ffffffea26529220 (shrinker_rwsem){++++}-{3:3}, at: shrink_slab+0x54/0x464
   #2: 6dffff808abe90e8 (&type->s_umount_key#47){++++}-{3:3}, at: trylock_super+0x2c/0x8c

<snipped>

---
 fs/crypto/crypto.c | 3 +++
 1 file changed, 3 insertions(+)
fs/crypto/crypto.c +1 -0
index e78be66bbf01..e10fc30142a6 100644
@@ -316,6 +316,7 @@ EXPORT_SYMBOL(fscrypt_decrypt_block_inplace);
 int fscrypt_initialize(unsigned int cop_flags)
 {
 	int err = 0;
+	unsigned int flags;
 
 	/* No need to allocate a bounce page pool if this FS won't use it. */
 	if (cop_flags & FS_CFLG_OWN_PAGES)

fs/crypto/crypto.c int fscrypt_initialize(unsigned int cop_flags) · line 335

Quoted 9 lines
fs/crypto/crypto.c +2 -0
@@ -326,8 +327,10 @@ int fscrypt_initialize(unsigned int cop_flags)
 		goto out_unlock;
 
 	err = -ENOMEM;
+	flags = memalloc_nofs_save();
 	fscrypt_bounce_page_pool =
 		mempool_create_page_pool(num_prealloc_crypto_pages, 0);
+	memalloc_nofs_restore(flags);
 	if (!fscrypt_bounce_page_pool)
 		goto out_unlock;

Thanks, but this isn't the correct fix. The real problem is that ext4 is calling fscrypt_get_encryption_info() from within a jbd2 transaction, which is fundamentally unsafe. It's a known regression from commit a80f7fcf1867 ("ext4: fixup ext4_fc_track_* functions' signature"), which extended the scope of the transaction in ext4_unlink() too far.

Sorry for not getting around to fixing this earlier. Are you interested in sending a patch for it? If you do, please make sure to include "Reported-by: [email protected]", as there's a syzbot report open already
(https://lore.kernel.org/all/[email protected]/T/#u).

- Eric

author
#2 lore raw
Quoted 65 lines · Eric Biggers
Quoted 63 lines · Yu Zhao

fscrypt_initialize() shouldn't allocate memory without GFP_NOFS.

The problem seems to go back to 2015
commit 57e5055b0a5e ("f2fs crypto: add f2fs encryption facilities") but I have never heard of any complaints, hence not CC'ing stable.

======================================================
WARNING: possible circular locking dependency detected 6.0.0-lockdep #1 Not tainted
------------------------------------------------------
kswapd0/77 is trying to acquire lock:
71ffff808b254a18 (jbd2_handle){++++}-{0:0}, at: start_this_handle+0x76c/0x8dc

but task is already holding lock:
ffffffea26533310 (fs_reclaim){+.+.}-{0:0}, at: 0x1

which lock already depends on the new lock.

<snipped>

other info that might help us debug this:

  Chain exists of:
    jbd2_handle --> fscrypt_init_mutex --> fs_reclaim

   Possible unsafe locking scenario:
         CPU0                    CPU1
         ----                    ----
    lock(fs_reclaim);
                                 lock(fscrypt_init_mutex);
                                 lock(fs_reclaim);
    lock(jbd2_handle);
   *** DEADLOCK ***

3 locks held by kswapd0/77:

   #0: ffffffea26533310 (fs_reclaim){+.+.}-{0:0}, at: 0x1
   #1: ffffffea26529220 (shrinker_rwsem){++++}-{3:3}, at: shrink_slab+0x54/0x464
   #2: 6dffff808abe90e8 (&type->s_umount_key#47){++++}-{3:3}, at: trylock_super+0x2c/0x8c

<snipped>

---
 fs/crypto/crypto.c | 3 +++
 1 file changed, 3 insertions(+)
fs/crypto/crypto.c +1 -0
index e78be66bbf01..e10fc30142a6 100644
@@ -316,6 +316,7 @@ EXPORT_SYMBOL(fscrypt_decrypt_block_inplace);
 int fscrypt_initialize(unsigned int cop_flags)
 {
      int err = 0;
+     unsigned int flags;
 
      /* No need to allocate a bounce page pool if this FS won't use it. */
      if (cop_flags & FS_CFLG_OWN_PAGES)
@@ -326,8 +327,10 @@ int fscrypt_initialize(unsigned int cop_flags)
              goto out_unlock;
 
Quoted 11 lines

fs/crypto/crypto.c int fscrypt_initialize(unsigned int cop_flags) · line 335

Quoted 7 lines
fs/crypto/crypto.c +2 -0
      err = -ENOMEM;
+     flags = memalloc_nofs_save();
      fscrypt_bounce_page_pool =
              mempool_create_page_pool(num_prealloc_crypto_pages, 0);
+     memalloc_nofs_restore(flags);
      if (!fscrypt_bounce_page_pool)
              goto out_unlock;

Thanks, but this isn't the correct fix. The real problem is that ext4 is calling fscrypt_get_encryption_info() from within a jbd2 transaction, which is fundamentally unsafe.

Thanks. So the jbd2_handle -> fscrypt_init_mutex dependency is bad.

Quoted 6 lines

It's a known regression from commit a80f7fcf1867
("ext4: fixup ext4_fc_track_* functions' signature"), which extended the scope of the transaction in ext4_unlink() too far.

Sorry for not getting around to fixing this earlier. Are you interested in sending a patch for it?

I have no idea how to move fscrypt_get_encryption_info() out of jbd2 transactions. I'll just leave it to you.

Quoted 6 lines

If you do, please make sure to include
"Reported-by: [email protected]", as there's a syzbot report open already
(https://lore.kernel.org/all/[email protected]/T/#u).

- Eric