Skip to contentSkip to conversation

Move fscrypt keyring destruction to after ->put_super

PATCH Applied Eric Biggers 3 participants 5 messages Dec 27, 2023 – Jan 16, 2024

Subsystems (3)
lore
Share & tools
lore

Eric Biggers

author maintainer
#0 lore raw

v2 cover Move fscrypt keyring destruction to after ->put_super

This series moves the fscrypt keyring destruction to after ->put_super, as this will be needed by the btrfs fscrypt support. To make this possible, it also changes f2fs to release its block devices after generic_shutdown_super() rather than before.

This supersedes "[PATCH] fscrypt: move the call to fscrypt_destroy_keyring() into ->put_super()"
(https://lore.kernel.org/linux-fscrypt/[email protected]/T/#u)

Changed in v2:
- Added a comment to f2fs patch.
- Dropped btrfs patch from series; it will go in separately.
- Added some Reviewed-bys.

Eric Biggers (1):
f2fs: move release of block devices to after kill_block_super()

Josef Bacik (1):
fs: move fscrypt keyring destruction to after ->put_super

 fs/f2fs/super.c | 13 ++++++++-----
 fs/super.c      | 12 ++++++------
 2 files changed, 14 insertions(+), 11 deletions(-)

base-commit: fbafc3e621c3f4ded43720fdb1d6ce1728ec664e

Signature · 1 line
2.43.0
author maintainer
#1 lore raw

PATCH v2 1/2 f2fs: move release of block devices to after kill_block_super()

From: Eric Biggers <[email protected]>

Call destroy_device_list() and free the f2fs_sb_info from kill_f2fs_super(), after the call to kill_block_super(). This is necessary to order it after the call to fscrypt_destroy_keyring() once generic_shutdown_super() starts calling fscrypt_destroy_keyring() just after calling ->put_super. This is because fscrypt_destroy_keyring() may call into f2fs_get_devices() via the fscrypt_operations.

---
 fs/f2fs/super.c | 13 ++++++++-----
 1 file changed, 8 insertions(+), 5 deletions(-)
fs/f2fs/super.c +8 -5
index 033af907c3b1d..d66e0692ac02e 100644
@@ -1710,42 +1710,39 @@ static void f2fs_put_super(struct super_block *sb)
 	f2fs_destroy_node_manager(sbi);
 	f2fs_destroy_segment_manager(sbi);
 
 	/* flush s_error_work before sbi destroy */
 	flush_work(&sbi->s_error_work);
 
 	f2fs_destroy_post_read_wq(sbi);
 
 	kvfree(sbi->ckpt);
 
-	sb->s_fs_info = NULL;
 	if (sbi->s_chksum_driver)
 		crypto_free_shash(sbi->s_chksum_driver);
 	kfree(sbi->raw_super);
 
-	destroy_device_list(sbi);
 	f2fs_destroy_page_array_cache(sbi);
 	f2fs_destroy_xattr_caches(sbi);
 	mempool_destroy(sbi->write_io_dummy);
 #ifdef CONFIG_QUOTA
 	for (i = 0; i < MAXQUOTAS; i++)
 		kfree(F2FS_OPTION(sbi).s_qf_names[i]);
 #endif
 	fscrypt_free_dummy_policy(&F2FS_OPTION(sbi).dummy_enc_policy);
 	destroy_percpu_info(sbi);
 	f2fs_destroy_iostat(sbi);
 	for (i = 0; i < NR_PAGE_TYPE; i++)
 		kvfree(sbi->write_io[i]);
 #if IS_ENABLED(CONFIG_UNICODE)
 	utf8_unload(sb->s_encoding);
 #endif
-	kfree(sbi);
 }
 
 int f2fs_sync_fs(struct super_block *sb, int sync)
 {
 	struct f2fs_sb_info *sbi = F2FS_SB(sb);
 	int err = 0;
 
 	if (unlikely(f2fs_cp_error(sbi)))
 		return 0;
 	if (unlikely(is_sbi_flag_set(sbi, SBI_CP_DISABLED)))
@@ -4895,23 +4892,23 @@ static int f2fs_fill_super(struct super_block *sb, void *data, int silent)
 }
 
 static struct dentry *f2fs_mount(struct file_system_type *fs_type, int flags,
 			const char *dev_name, void *data)
 {
 	return mount_bdev(fs_type, flags, dev_name, data, f2fs_fill_super);
 }
 
 static void kill_f2fs_super(struct super_block *sb)
 {
-	if (sb->s_root) {
-		struct f2fs_sb_info *sbi = F2FS_SB(sb);
+	struct f2fs_sb_info *sbi = F2FS_SB(sb);
 
+	if (sb->s_root) {
 		set_sbi_flag(sbi, SBI_IS_CLOSE);
 		f2fs_stop_gc_thread(sbi);
 		f2fs_stop_discard_thread(sbi);
 
 #ifdef CONFIG_F2FS_FS_COMPRESSION
 		/*
 		 * latter evict_inode() can bypass checking and invalidating
 		 * compress inode cache.
 		 */
 		if (test_opt(sbi, COMPRESS_CACHE))
@@ -4924,20 +4921,26 @@ static void kill_f2fs_super(struct super_block *sb)
 				.reason = CP_UMOUNT,
 			};
 			stat_inc_cp_call_count(sbi, TOTAL_CALL);
 			f2fs_write_checkpoint(sbi, &cpc);
 		}
 
 		if (is_sbi_flag_set(sbi, SBI_IS_RECOVERED) && f2fs_readonly(sb))
 			sb->s_flags &= ~SB_RDONLY;
 	}
 	kill_block_super(sb);
+	/* Release block devices last, after fscrypt_destroy_keyring(). */
+	if (sbi) {
+		destroy_device_list(sbi);
+		kfree(sbi);
+		sb->s_fs_info = NULL;
+	}
 }
 
 static struct file_system_type f2fs_fs_type = {
 	.owner		= THIS_MODULE,
 	.name		= "f2fs",
 	.mount		= f2fs_mount,
 	.kill_sb	= kill_f2fs_super,
 	.fs_flags	= FS_REQUIRES_DEV | FS_ALLOW_IDMAP,
 };
 MODULE_ALIAS_FS("f2fs");
Signature · 1 line
2.43.0
maintainer Reviewed-by on v2 1/2
#2 lore raw

Re: [f2fs-dev] [PATCH v2 1/2] f2fs: move release of block devices to after kill_block_super()

Quoted 10 lines · Eric Biggers

From: Eric Biggers <[email protected]>

Call destroy_device_list() and free the f2fs_sb_info from kill_f2fs_super(), after the call to kill_block_super(). This is necessary to order it after the call to fscrypt_destroy_keyring() once generic_shutdown_super() starts calling fscrypt_destroy_keyring() just after calling ->put_super. This is because fscrypt_destroy_keyring() may call into f2fs_get_devices() via the fscrypt_operations.

Thanks,

author maintainer
#3 lore raw

PATCH v2 2/2 fs: move fscrypt keyring destruction to after ->put_super

In reply to Eric Biggers

From: Josef Bacik <[email protected]>

btrfs has a variety of asynchronous things we do with inodes that can potentially last until ->put_super, when we shut everything down and clean up all of our async work. Due to this we need to move fscrypt_destroy_keyring() to after ->put_super, otherwise we get warnings about still having active references on the master key.

---
 fs/super.c | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)
fs/super.c +6 -6
index 076392396e724..faf7d248145d2 100644
@@ -674,34 +674,34 @@ void generic_shutdown_super(struct super_block *sb)
 		/* Evict all inodes with zero refcount. */
 		evict_inodes(sb);
 
 		/*
 		 * Clean up and evict any inodes that still have references due
 		 * to fsnotify or the security policy.
 		 */
 		fsnotify_sb_delete(sb);
 		security_sb_delete(sb);
 
-		/*
-		 * Now that all potentially-encrypted inodes have been evicted,
-		 * the fscrypt keyring can be destroyed.
-		 */
-		fscrypt_destroy_keyring(sb);
-
 		if (sb->s_dio_done_wq) {
 			destroy_workqueue(sb->s_dio_done_wq);
 			sb->s_dio_done_wq = NULL;
 		}
 
 		if (sop->put_super)
 			sop->put_super(sb);
 
+		/*
+		 * Now that all potentially-encrypted inodes have been evicted,
+		 * the fscrypt keyring can be destroyed.
+		 */
+		fscrypt_destroy_keyring(sb);
+
 		if (CHECK_DATA_CORRUPTION(!list_empty(&sb->s_inodes),
 				"VFS: Busy inodes after unmount of %s (%s)",
 				sb->s_id, sb->s_type->name)) {
 			/*
 			 * Adding a proper bailout path here would be hard, but
 			 * we can at least make it more likely that a later
 			 * iput_final() or such crashes cleanly.
 			 */
 			struct inode *inode;
 
Signature · 1 line
2.43.0
#4 lore raw
In reply to Eric Biggers
This series was applied to jaegeuk/f2fs.git (dev) by Eric Biggers <[email protected]>:

Hello:

This series was applied to jaegeuk/f2fs.git (dev)
by Eric Biggers <[email protected]>:

Quoted 10 lines · On Wed, 27 Dec 2023 11:14:27 -0600 you wrote:

This series moves the fscrypt keyring destruction to after ->put_super, as this will be needed by the btrfs fscrypt support. To make this possible, it also changes f2fs to release its block devices after generic_shutdown_super() rather than before.

This supersedes "[PATCH] fscrypt: move the call to fscrypt_destroy_keyring() into ->put_super()"
(https://lore.kernel.org/linux-fscrypt/[email protected]/T/#u)

[...]

Here is the summary with links:
- [f2fs-dev,v2,1/2] f2fs: move release of block devices to after kill_block_super()

    https://git.kernel.org/jaegeuk/f2fs/c/275dca4630c1
  - [f2fs-dev,v2,2/2] fs: move fscrypt keyring destruction to after ->put_super
    https://git.kernel.org/jaegeuk/f2fs/c/2a0e85719892

You are awesome, thank you!

Signature · 2 lines
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html