Merge tag 'vfs-6.6-merge-2' of ssh://gitolite.kernel.org/pub/scm/fs/xfs/xfs-linux

Pull filesystem freezing updates from Darrick Wong:

New code for 6.6:

 * Allow the kernel to initiate a freeze of a filesystem.  The kernel
   and userspace can both hold a freeze on a filesystem at the same
   time; the freeze is not lifted until /both/ holders lift it.  This
   will enable us to fix a longstanding bug in XFS online fsck.

Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Message-Id: <20230822182604.GB11286@frogsfrogsfrogs>
Signed-off-by: Christian Brauner <brauner@kernel.org>

1 files changed, 102 insertions, 18 deletions

diff --git a/fs/super.c b/fs/super.c
index a284052ebab2..ef87103e2a51 100644
--- a/fs/super.c
+++ b/fs/super.c
@@ -39,7 +39,7 @@
 #include <uapi/linux/mount.h>
 #include "internal.h"
 
-static int thaw_super_locked(struct super_block *sb);
+static int thaw_super_locked(struct super_block *sb, enum freeze_holder who);
 
 static LIST_HEAD(super_blocks);
 static DEFINE_SPINLOCK(sb_lock);
@@ -1182,7 +1182,7 @@ static void do_thaw_all_callback(struct super_block *sb)
 
 	if (born && sb->s_root) {
 		emergency_thaw_bdev(sb);
-		thaw_super_locked(sb);
+		thaw_super_locked(sb, FREEZE_HOLDER_USERSPACE);
 	} else {
 		super_unlock_excl(sb);
 	}
@@ -1836,14 +1836,43 @@ static void sb_freeze_unlock(struct super_block *sb, int level)
 		percpu_up_write(sb->s_writers.rw_sem + level);
 }
 
+static int wait_for_partially_frozen(struct super_block *sb)
+{
+	int ret = 0;
+
+	do {
+		unsigned short old = sb->s_writers.frozen;
+
+		up_write(&sb->s_umount);
+		ret = wait_var_event_killable(&sb->s_writers.frozen,
+					       sb->s_writers.frozen != old);
+		down_write(&sb->s_umount);
+	} while (ret == 0 &&
+		 sb->s_writers.frozen != SB_UNFROZEN &&
+		 sb->s_writers.frozen != SB_FREEZE_COMPLETE);
+
+	return ret;
+}
+
 /**
  * freeze_super - lock the filesystem and force it into a consistent state
  * @sb: the super to lock
+ * @who: context that wants to freeze
  *
  * Syncs the super to make sure the filesystem is consistent and calls the fs's
- * freeze_fs.  Subsequent calls to this without first thawing the fs will return
+ * freeze_fs.  Subsequent calls to this without first thawing the fs may return
  * -EBUSY.
  *
+ * @who should be:
+ * * %FREEZE_HOLDER_USERSPACE if userspace wants to freeze the fs;
+ * * %FREEZE_HOLDER_KERNEL if the kernel wants to freeze the fs.
+ *
+ * The @who argument distinguishes between the kernel and userspace trying to
+ * freeze the filesystem.  Although there cannot be multiple kernel freezes or
+ * multiple userspace freezes in effect at any given time, the kernel and
+ * userspace can both hold a filesystem frozen.  The filesystem remains frozen
+ * until there are no kernel or userspace freezes in effect.
+ *
  * During this function, sb->s_writers.frozen goes through these values:
  *
  * SB_UNFROZEN: File system is normal, all writes progress as usual.
@@ -1869,20 +1898,40 @@ static void sb_freeze_unlock(struct super_block *sb, int level)
  *
  * sb->s_writers.frozen is protected by sb->s_umount.
  */
-int freeze_super(struct super_block *sb)
+int freeze_super(struct super_block *sb, enum freeze_holder who)
 {
 	int ret;
 
-	/* Since the caller must already have an active reference... */
 	atomic_inc(&sb->s_active);
-
-	/* ...@sb definitely can't be dying. */
 	if (!super_lock_excl(sb))
 		WARN(1, "Dying superblock while freezing!");
 
+retry:
+	if (sb->s_writers.frozen == SB_FREEZE_COMPLETE) {
+		if (sb->s_writers.freeze_holders & who) {
+			deactivate_locked_super(sb);
+			return -EBUSY;
+		}
+
+		WARN_ON(sb->s_writers.freeze_holders == 0);
+
+		/*
+		 * Someone else already holds this type of freeze; share the
+		 * freeze and assign the active ref to the freeze.
+		 */
+		sb->s_writers.freeze_holders |= who;
+		super_unlock_excl(sb);
+		return 0;
+	}
+
 	if (sb->s_writers.frozen != SB_UNFROZEN) {
-		deactivate_locked_super(sb);
-		return -EBUSY;
+		ret = wait_for_partially_frozen(sb);
+		if (ret) {
+			deactivate_locked_super(sb);
+			return ret;
+		}
+
+		goto retry;
 	}
 
 	if (!(sb->s_flags & SB_BORN)) {
@@ -1892,7 +1941,9 @@ int freeze_super(struct super_block *sb)
 
 	if (sb_rdonly(sb)) {
 		/* Nothing to do really... */
+		sb->s_writers.freeze_holders |= who;
 		sb->s_writers.frozen = SB_FREEZE_COMPLETE;
+		wake_up_var(&sb->s_writers.frozen);
 		super_unlock_excl(sb);
 		return 0;
 	}
@@ -1901,8 +1952,6 @@ int freeze_super(struct super_block *sb)
 	/* Release s_umount to preserve sb_start_write -> s_umount ordering */
 	super_unlock_excl(sb);
 	sb_wait_write(sb, SB_FREEZE_WRITE);
-
-	/* We're still holding an active reference. */
 	if (!super_lock_excl(sb))
 		WARN(1, "Dying superblock while freezing!");
 
@@ -1915,6 +1964,7 @@ int freeze_super(struct super_block *sb)
 	if (ret) {
 		sb->s_writers.frozen = SB_UNFROZEN;
 		sb_freeze_unlock(sb, SB_FREEZE_PAGEFAULT);
+		wake_up_var(&sb->s_writers.frozen);
 		deactivate_locked_super(sb);
 		return ret;
 	}
@@ -1930,6 +1980,7 @@ int freeze_super(struct super_block *sb)
 				"VFS:Filesystem freeze failed\n");
 			sb->s_writers.frozen = SB_UNFROZEN;
 			sb_freeze_unlock(sb, SB_FREEZE_FS);
+			wake_up_var(&sb->s_writers.frozen);
 			deactivate_locked_super(sb);
 			return ret;
 		}
@@ -1938,24 +1989,50 @@ int freeze_super(struct super_block *sb)
 	 * For debugging purposes so that fs can warn if it sees write activity
 	 * when frozen is set to SB_FREEZE_COMPLETE, and for thaw_super().
 	 */
+	sb->s_writers.freeze_holders |= who;
 	sb->s_writers.frozen = SB_FREEZE_COMPLETE;
+	wake_up_var(&sb->s_writers.frozen);
 	lockdep_sb_freeze_release(sb);
 	super_unlock_excl(sb);
 	return 0;
 }
 EXPORT_SYMBOL(freeze_super);
 
-static int thaw_super_locked(struct super_block *sb)
+/*
+ * Undoes the effect of a freeze_super_locked call.  If the filesystem is
+ * frozen both by userspace and the kernel, a thaw call from either source
+ * removes that state without releasing the other state or unlocking the
+ * filesystem.
+ */
+static int thaw_super_locked(struct super_block *sb, enum freeze_holder who)
 {
 	int error;
 
-	if (sb->s_writers.frozen != SB_FREEZE_COMPLETE) {
+	if (sb->s_writers.frozen == SB_FREEZE_COMPLETE) {
+		if (!(sb->s_writers.freeze_holders & who)) {
+			super_unlock_excl(sb);
+			return -EINVAL;
+		}
+
+		/*
+		 * Freeze is shared with someone else.  Release our hold and
+		 * drop the active ref that freeze_super assigned to the
+		 * freezer.
+		 */
+		if (sb->s_writers.freeze_holders & ~who) {
+			sb->s_writers.freeze_holders &= ~who;
+			deactivate_locked_super(sb);
+			return 0;
+		}
+	} else {
 		super_unlock_excl(sb);
 		return -EINVAL;
 	}
 
 	if (sb_rdonly(sb)) {
+		sb->s_writers.freeze_holders &= ~who;
 		sb->s_writers.frozen = SB_UNFROZEN;
+		wake_up_var(&sb->s_writers.frozen);
 		goto out;
 	}
 
@@ -1964,15 +2041,16 @@ static int thaw_super_locked(struct super_block *sb)
 	if (sb->s_op->unfreeze_fs) {
 		error = sb->s_op->unfreeze_fs(sb);
 		if (error) {
-			printk(KERN_ERR
-				"VFS:Filesystem thaw failed\n");
+			printk(KERN_ERR "VFS:Filesystem thaw failed\n");
 			lockdep_sb_freeze_release(sb);
 			super_unlock_excl(sb);
 			return error;
 		}
 	}
 
+	sb->s_writers.freeze_holders &= ~who;
 	sb->s_writers.frozen = SB_UNFROZEN;
+	wake_up_var(&sb->s_writers.frozen);
 	sb_freeze_unlock(sb, SB_FREEZE_FS);
 out:
 	deactivate_locked_super(sb);
@@ -1982,14 +2060,20 @@ out:
 /**
  * thaw_super -- unlock filesystem
  * @sb: the super to thaw
+ * @who: context that wants to freeze
+ *
+ * Unlocks the filesystem and marks it writeable again after freeze_super()
+ * if there are no remaining freezes on the filesystem.
  *
- * Unlocks the filesystem and marks it writeable again after freeze_super().
+ * @who should be:
+ * * %FREEZE_HOLDER_USERSPACE if userspace wants to thaw the fs;
+ * * %FREEZE_HOLDER_KERNEL if the kernel wants to thaw the fs.
  */
-int thaw_super(struct super_block *sb)
+int thaw_super(struct super_block *sb, enum freeze_holder who)
 {
 	if (!super_lock_excl(sb))
 		WARN(1, "Dying superblock while thawing!");
-	return thaw_super_locked(sb);
+	return thaw_super_locked(sb, who);
 }
 EXPORT_SYMBOL(thaw_super);