Improve simple directory offset wrap behavior

EBUSY O_EXCL was specified in flags and pathname refers
to a block device that is in use by the system (e.g., it is mounted).

From: Chuck Lever <[email protected]>

Testing shows that the EBUSY error return from mtree_alloc_cyclic() leaks into user space. The ERRORS section of "man creat(2)" says:

Quoted 3 lines

EBUSY O_EXCL was specified in flags and pathname refers
to a block device that is in use by the system (e.g., it is mounted).

ENOSPC is closer to what applications expect in this situation.

Note that the normal range of simple directory offset values is 2..2^63, so hitting this error is going to be rare to impossible.

Fixes 6faddda69f62 ("libfs: Add directory operations for stable offsets")

Cc <[email protected]> # v6.9+

Reviewed-by Jeff Layton <[email protected]>

Reviewed-by Yang Erkun <[email protected]>

Signed-off-by Chuck Lever <[email protected]>

---
 fs/libfs.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

fs/libfs.c +2 -2

index 748ac5923154..3da58a92f48f 100644
@@ -292,8 +292,8 @@ int simple_offset_add(struct offset_ctx *octx, struct dentry *dentry)
 
 	ret = mtree_alloc_cyclic(&octx->mt, &offset, dentry, DIR_OFFSET_MIN,
 				 LONG_MAX, &octx->next_offset, GFP_KERNEL);
-	if (ret < 0)
-		return ret;
+	if (unlikely(ret < 0))
+		return ret == -EBUSY ? -ENOSPC : ret;
 
 	offset_set(dentry, offset);
 	return 0;

Signature · 1 line

2.47.0

Quoted 5 lines · * [email protected] <[email protected]> [241220 10:33]:

From: Chuck Lever <[email protected]>

Testing shows that the EBUSY error return from mtree_alloc_cyclic() leaks into user space. The ERRORS section of "man creat(2)" says:

Quoted 5 lines

Quoted 3 lines

EBUSY O_EXCL was specified in flags and pathname refers
to a block device that is in use by the system (e.g., it is mounted).

ENOSPC is closer to what applications expect in this situation.

Should the tree be returning ENOSPC in this case as apposed to translating it here?

Quoted 31 lines

Note that the normal range of simple directory offset values is 2..2^63, so hitting this error is going to be rare to impossible.

Fixes 6faddda69f62 ("libfs: Add directory operations for stable offsets")

Cc <[email protected]> # v6.9+

Reviewed-by Jeff Layton <[email protected]>

Reviewed-by Yang Erkun <[email protected]>

Signed-off-by Chuck Lever <[email protected]>

---
  fs/libfs.c | 4 ++--
  1 file changed, 2 insertions(+), 2 deletions(-)

fs/libfs.c +2 -2

index 748ac5923154..3da58a92f48f 100644
@@ -292,8 +292,8 @@ int simple_offset_add(struct offset_ctx *octx, struct dentry *dentry)
 
  	ret = mtree_alloc_cyclic(&octx->mt, &offset, dentry, DIR_OFFSET_MIN,
  				 LONG_MAX, &octx->next_offset, GFP_KERNEL);
-	if (ret < 0)
-		return ret;
+	if (unlikely(ret < 0))
+		return ret == -EBUSY ? -ENOSPC : ret;
 
  	offset_set(dentry, offset);
  	return 0;

Signature · 1 line

2.47.0

From: Chuck Lever <[email protected]>

simple_empty() and simple_offset_empty() perform the same task. The latter's use as a canary to find bugs has not found any new issues. A subsequent patch will remove the use of the mtree for iterating directory contents, so revert back to using a similar mechanism for determining whether a directory is indeed empty.

Only one such mechanism is ever needed.

Signed-off-by Chuck Lever <[email protected]>

---
  fs/libfs.c         | 32 --------------------------------
  include/linux/fs.h |  1 -
  mm/shmem.c         |  4 ++--
  3 files changed, 2 insertions(+), 35 deletions(-)

fs/libfs.c +0 -32

index 3da58a92f48f..8380d9314ebd 100644
@@ -329,38 +329,6 @@ void simple_offset_remove(struct offset_ctx *octx, struct dentry *dentry)
  	offset_set(dentry, 0);
  }
 
-/**
- * simple_offset_empty - Check if a dentry can be unlinked
- * @dentry: dentry to be tested
- *
- * Returns 0 if @dentry is a non-empty directory; otherwise returns 1.
- */
-int simple_offset_empty(struct dentry *dentry)
-{
-	struct inode *inode = d_inode(dentry);
-	struct offset_ctx *octx;
-	struct dentry *child;
-	unsigned long index;
-	int ret = 1;
-
-	if (!inode || !S_ISDIR(inode->i_mode))
-		return ret;
-
-	index = DIR_OFFSET_MIN;
-	octx = inode->i_op->get_offset_ctx(inode);
-	mt_for_each(&octx->mt, child, index, LONG_MAX) {
-		spin_lock(&child->d_lock);
-		if (simple_positive(child)) {
-			spin_unlock(&child->d_lock);
-			ret = 0;
-			break;
-		}
-		spin_unlock(&child->d_lock);
-	}
-
-	return ret;
-}
-
  /**
   * simple_offset_rename - handle directory offsets for rename
   * @old_dir: parent directory of source entry

include/linux/fs.h +0 -1

index 7e29433c5ecc..f7efc6866ebc 100644
@@ -3468,7 +3468,6 @@ struct offset_ctx {
  void simple_offset_init(struct offset_ctx *octx);
  int simple_offset_add(struct offset_ctx *octx, struct dentry *dentry);
  void simple_offset_remove(struct offset_ctx *octx, struct dentry *dentry);
-int simple_offset_empty(struct dentry *dentry);
  int simple_offset_rename(struct inode *old_dir, struct dentry *old_dentry,
  			 struct inode *new_dir, struct dentry *new_dentry);
  int simple_offset_rename_exchange(struct inode *old_dir,

mm/shmem.c +2 -2

index ccb9629a0f70..274c2666f457 100644
@@ -3818,7 +3818,7 @@ static int shmem_unlink(struct inode *dir, struct dentry *dentry)
 
  static int shmem_rmdir(struct inode *dir, struct dentry *dentry)
  {
-	if (!simple_offset_empty(dentry))
+	if (!simple_empty(dentry))
  		return -ENOTEMPTY;
 
  	drop_nlink(d_inode(dentry));

@@ -3875,7 +3875,7 @@ static int shmem_rename2(struct mnt_idmap *idmap,
  		return simple_offset_rename_exchange(old_dir, old_dentry,
  						     new_dir, new_dentry);
 
-	if (!simple_offset_empty(new_dentry))
+	if (!simple_empty(new_dentry))
  		return -ENOTEMPTY;
 
  	if (flags & RENAME_WHITEOUT) {

From: Chuck Lever <[email protected]>

The current directory offset allocator (based on mtree_alloc_cyclic) stores the next offset value to return in octx->next_offset. This mechanism typically returns values that increase monotonically over time. Eventually, though, the newly allocated offset value wraps back to a low number (say, 2) which is smaller than other already- allocated offset values.

Yu Kuai <[email protected]> reports that, after commit 64a7ce76fb90 ("libfs: fix infinite directory reads for offset dir"), if a directory's offset allocator wraps, existing entries are no longer visible via readdir/getdents because offset_readdir() stops listing entries once an entry's offset is larger than octx->next_offset. These entries vanish persistently -- they can be looked up, but will never again appear in readdir(3) output.

The reason for this is that the commit treats directory offsets as monotonically increasing integer values rather than opaque cookies, and introduces this comparison:

	if (dentry2offset(dentry) >= last_index) {

On 64-bit platforms, the directory offset value upper bound is 2^63 - 1. Directory offsets will monotonically increase for millions of years without wrapping.

On 32-bit platforms, however, LONG_MAX is 2^31 - 1. The allocator can wrap after only a few weeks (at worst).

Revert commit 64a7ce76fb90 ("libfs: fix infinite directory reads for offset dir") to prepare for a fix that can work properly on 32-bit systems and might apply to recent LTS kernels where shmem employs the simple_offset mechanism.

Reported-by Yu Kuai <[email protected]>

Signed-off-by Chuck Lever <[email protected]>

---
  fs/libfs.c | 35 +++++++++++------------------------
  1 file changed, 11 insertions(+), 24 deletions(-)

fs/libfs.c +11 -24

index 8380d9314ebd..8c9364a0174c 100644
@@ -422,14 +422,6 @@ void simple_offset_destroy(struct offset_ctx *octx)
  	mtree_destroy(&octx->mt);
  }
 
-static int offset_dir_open(struct inode *inode, struct file *file)
-{
-	struct offset_ctx *ctx = inode->i_op->get_offset_ctx(inode);
-
-	file->private_data = (void *)ctx->next_offset;
-	return 0;
-}
-
  /**
   * offset_dir_llseek - Advance the read position of a directory descriptor
   * @file: an open directory whose position is to be updated

@@ -443,9 +435,6 @@ static int offset_dir_open(struct inode *inode, struct file *file)
   */
  static loff_t offset_dir_llseek(struct file *file, loff_t offset, int whence)
  {
-	struct inode *inode = file->f_inode;
-	struct offset_ctx *ctx = inode->i_op->get_offset_ctx(inode);
-
  	switch (whence) {
  	case SEEK_CUR:
  		offset += file->f_pos;

@@ -459,8 +448,7 @@ static loff_t offset_dir_llseek(struct file *file, loff_t offset, int whence)
  	}
 
  	/* In this case, ->private_data is protected by f_pos_lock */
-	if (!offset)
-		file->private_data = (void *)ctx->next_offset;
+	file->private_data = NULL;
  	return vfs_setpos(file, offset, LONG_MAX);
  }
 

@@ -491,7 +479,7 @@ static bool offset_dir_emit(struct dir_context *ctx, struct dentry *dentry)
  			  inode->i_ino, fs_umode_to_dtype(inode->i_mode));
  }
 
-static void offset_iterate_dir(struct inode *inode, struct dir_context *ctx, long last_index)
+static void *offset_iterate_dir(struct inode *inode, struct dir_context *ctx)
  {
  	struct offset_ctx *octx = inode->i_op->get_offset_ctx(inode);
  	struct dentry *dentry;

@@ -499,21 +487,17 @@ static void offset_iterate_dir(struct inode *inode, struct dir_context *ctx, lon
  	while (true) {
  		dentry = offset_find_next(octx, ctx->pos);
  		if (!dentry)
-			return;
-
-		if (dentry2offset(dentry) >= last_index) {
-			dput(dentry);
-			return;
-		}
+			return ERR_PTR(-ENOENT);
 
  		if (!offset_dir_emit(ctx, dentry)) {
  			dput(dentry);
-			return;
+			break;
  		}
 
  		ctx->pos = dentry2offset(dentry) + 1;
  		dput(dentry);
  	}
+	return NULL;
  }
 
  /**

@@ -540,19 +524,22 @@ static void offset_iterate_dir(struct inode *inode, struct dir_context *ctx, lon
  static int offset_readdir(struct file *file, struct dir_context *ctx)
  {
  	struct dentry *dir = file->f_path.dentry;
-	long last_index = (long)file->private_data;
 
  	lockdep_assert_held(&d_inode(dir)->i_rwsem);
 
  	if (!dir_emit_dots(file, ctx))
  		return 0;
 
-	offset_iterate_dir(d_inode(dir), ctx, last_index);
+	/* In this case, ->private_data is protected by f_pos_lock */
+	if (ctx->pos == DIR_OFFSET_MIN)
+		file->private_data = NULL;
+	else if (file->private_data == ERR_PTR(-ENOENT))
+		return 0;
+	file->private_data = offset_iterate_dir(d_inode(dir), ctx);
  	return 0;
  }
 
  const struct file_operations simple_offset_dir_operations = {
-	.open		= offset_dir_open,
  	.llseek		= offset_dir_llseek,
  	.iterate_shared	= offset_readdir,
  	.read		= generic_read_dir,

From: Chuck Lever <[email protected]>

According to getdents(3), the d_off field in each returned directory entry points to the next entry in the directory. The d_off field in the last returned entry in the readdir buffer must contain a valid offset value, but if it points to an actual directory entry, then readdir/getdents can loop.

This patch introduces a specific fixed offset value that is placed in the d_off field of the last entry in a directory. Some user space applications assume that the EOD offset value is larger than the offsets of real directory entries, so the largest possible offset value is reserved for this purpose. This new value is never allocated by simple_offset_add().

When ->iterate_dir() returns, getdents{64} inserts the ctx->pos value into the d_off field of the last valid entry in the readdir buffer. When it hits EOD, offset_readdir() sets ctx->pos to the EOD offset value so the last entry is updated to point to the EOD marker.

When trying to read the entry at the EOD offset, offset_readdir() terminates immediately.

It is worth noting that using a Maple tree for directory offset value allocation does not guarantee a 63-bit range of values -- on platforms where "long" is a 32-bit type, the directory offset value range is still 0..(2^31 - 1).

Fixes 796432efab1e ("libfs: getdents() should return 0 after reaching EOD")

Signed-off-by Chuck Lever <[email protected]>

---
  fs/libfs.c | 38 ++++++++++++++++++++++----------------
  1 file changed, 22 insertions(+), 16 deletions(-)

fs/libfs.c +22 -16

index 8c9364a0174c..5c56783c03a5 100644
@@ -245,9 +245,16 @@ const struct inode_operations simple_dir_inode_operations = {
  };
  EXPORT_SYMBOL(simple_dir_inode_operations);
 
-/* 0 is '.', 1 is '..', so always start with offset 2 or more */
+/* simple_offset_add() allocation range */
  enum {
-	DIR_OFFSET_MIN	= 2,
+	DIR_OFFSET_MIN		= 2,
+	DIR_OFFSET_MAX		= LONG_MAX - 1,
+};
+
+/* simple_offset_add() never assigns these to a dentry */
+enum {
+	DIR_OFFSET_EOD		= LONG_MAX,	/* Marks EOD */
+
  };
 
  static void offset_set(struct dentry *dentry, long offset)

@@ -291,7 +298,8 @@ int simple_offset_add(struct offset_ctx *octx, struct dentry *dentry)
  		return -EBUSY;
 
  	ret = mtree_alloc_cyclic(&octx->mt, &offset, dentry, DIR_OFFSET_MIN,
-				 LONG_MAX, &octx->next_offset, GFP_KERNEL);
+				 DIR_OFFSET_MAX, &octx->next_offset,
+				 GFP_KERNEL);
  	if (unlikely(ret < 0))
  		return ret == -EBUSY ? -ENOSPC : ret;
 

@@ -447,8 +455,6 @@ static loff_t offset_dir_llseek(struct file *file, loff_t offset, int whence)
  		return -EINVAL;
  	}
 
-	/* In this case, ->private_data is protected by f_pos_lock */
-	file->private_data = NULL;
  	return vfs_setpos(file, offset, LONG_MAX);
  }
 

@@ -458,7 +464,7 @@ static struct dentry *offset_find_next(struct offset_ctx *octx, loff_t offset)
  	struct dentry *child, *found = NULL;
 
  	rcu_read_lock();
-	child = mas_find(&mas, LONG_MAX);
+	child = mas_find(&mas, DIR_OFFSET_MAX);
  	if (!child)
  		goto out;
  	spin_lock(&child->d_lock);

@@ -479,7 +485,7 @@ static bool offset_dir_emit(struct dir_context *ctx, struct dentry *dentry)
  			  inode->i_ino, fs_umode_to_dtype(inode->i_mode));
  }
 
-static void *offset_iterate_dir(struct inode *inode, struct dir_context *ctx)
+static void offset_iterate_dir(struct inode *inode, struct dir_context *ctx)
  {
  	struct offset_ctx *octx = inode->i_op->get_offset_ctx(inode);
  	struct dentry *dentry;

@@ -487,7 +493,7 @@ static void *offset_iterate_dir(struct inode *inode, struct dir_context *ctx)
  	while (true) {
  		dentry = offset_find_next(octx, ctx->pos);
  		if (!dentry)
-			return ERR_PTR(-ENOENT);
+			goto out_eod;
 
  		if (!offset_dir_emit(ctx, dentry)) {
  			dput(dentry);

@@ -497,7 +503,10 @@ static void *offset_iterate_dir(struct inode *inode, struct dir_context *ctx)
  		ctx->pos = dentry2offset(dentry) + 1;
  		dput(dentry);
  	}
-	return NULL;
+	return;
+
+out_eod:
+	ctx->pos = DIR_OFFSET_EOD;
  }
 
  /**

@@ -517,6 +526,8 @@ static void *offset_iterate_dir(struct inode *inode, struct dir_context *ctx)
   *
   * On return, @ctx->pos contains an offset that will read the next entry
   * in this directory when offset_readdir() is called again with @ctx.
+ * Caller places this value in the d_off field of the last entry in the
+ * user's buffer.
   *
   * Return values:
   *   %0 - Complete

@@ -529,13 +540,8 @@ static int offset_readdir(struct file *file, struct dir_context *ctx)
 
  	if (!dir_emit_dots(file, ctx))
  		return 0;
-
-	/* In this case, ->private_data is protected by f_pos_lock */
-	if (ctx->pos == DIR_OFFSET_MIN)
-		file->private_data = NULL;
-	else if (file->private_data == ERR_PTR(-ENOENT))
-		return 0;
-	file->private_data = offset_iterate_dir(d_inode(dir), ctx);
+	if (ctx->pos != DIR_OFFSET_EOD)
+		offset_iterate_dir(d_inode(dir), ctx);
  	return 0;
  }
 

From: Chuck Lever <[email protected]>

According to getdents(3), the d_off field in each returned directory entry points to the next entry in the directory. The d_off field in the last returned entry in the readdir buffer must contain a valid offset value, but if it points to an actual directory entry, then readdir/getdents can loop.

This patch introduces a specific fixed offset value that is placed in the d_off field of the last entry in a directory. Some user space applications assume that the EOD offset value is larger than the offsets of real directory entries, so the largest possible offset value is reserved for this purpose. This new value is never allocated by simple_offset_add().

When ->iterate_dir() returns, getdents{64} inserts the ctx->pos value into the d_off field of the last valid entry in the readdir buffer. When it hits EOD, offset_readdir() sets ctx->pos to the EOD offset value so the last entry is updated to point to the EOD marker.

When trying to read the entry at the EOD offset, offset_readdir() terminates immediately.

It is worth noting that using a Maple tree for directory offset value allocation does not guarantee a 63-bit range of values -- on platforms where "long" is a 32-bit type, the directory offset value range is still 0..(2^31 - 1).

Fixes 796432efab1e ("libfs: getdents() should return 0 after reaching EOD")

Signed-off-by Chuck Lever <[email protected]>

---
 fs/libfs.c | 38 ++++++++++++++++++++++----------------
 1 file changed, 22 insertions(+), 16 deletions(-)

fs/libfs.c +22 -16

index 8c9364a0174c..5c56783c03a5 100644
@@ -245,9 +245,16 @@ const struct inode_operations simple_dir_inode_operations = {
 };
 EXPORT_SYMBOL(simple_dir_inode_operations);
 
-/* 0 is '.', 1 is '..', so always start with offset 2 or more */
+/* simple_offset_add() allocation range */
 enum {
-	DIR_OFFSET_MIN	= 2,
+	DIR_OFFSET_MIN		= 2,
+	DIR_OFFSET_MAX		= LONG_MAX - 1,
+};
+
+/* simple_offset_add() never assigns these to a dentry */
+enum {
+	DIR_OFFSET_EOD		= LONG_MAX,	/* Marks EOD */
+
 };
 
 static void offset_set(struct dentry *dentry, long offset)

@@ -291,7 +298,8 @@ int simple_offset_add(struct offset_ctx *octx, struct dentry *dentry)
 		return -EBUSY;
 
 	ret = mtree_alloc_cyclic(&octx->mt, &offset, dentry, DIR_OFFSET_MIN,
-				 LONG_MAX, &octx->next_offset, GFP_KERNEL);
+				 DIR_OFFSET_MAX, &octx->next_offset,
+				 GFP_KERNEL);
 	if (unlikely(ret < 0))
 		return ret == -EBUSY ? -ENOSPC : ret;
 

@@ -447,8 +455,6 @@ static loff_t offset_dir_llseek(struct file *file, loff_t offset, int whence)
 		return -EINVAL;
 	}
 
-	/* In this case, ->private_data is protected by f_pos_lock */
-	file->private_data = NULL;
 	return vfs_setpos(file, offset, LONG_MAX);
 }
 

@@ -458,7 +464,7 @@ static struct dentry *offset_find_next(struct offset_ctx *octx, loff_t offset)
 	struct dentry *child, *found = NULL;
 
 	rcu_read_lock();
-	child = mas_find(&mas, LONG_MAX);
+	child = mas_find(&mas, DIR_OFFSET_MAX);
 	if (!child)
 		goto out;
 	spin_lock(&child->d_lock);

@@ -479,7 +485,7 @@ static bool offset_dir_emit(struct dir_context *ctx, struct dentry *dentry)
 			  inode->i_ino, fs_umode_to_dtype(inode->i_mode));
 }
 
-static void *offset_iterate_dir(struct inode *inode, struct dir_context *ctx)
+static void offset_iterate_dir(struct inode *inode, struct dir_context *ctx)
 {
 	struct offset_ctx *octx = inode->i_op->get_offset_ctx(inode);
 	struct dentry *dentry;

@@ -487,7 +493,7 @@ static void *offset_iterate_dir(struct inode *inode, struct dir_context *ctx)
 	while (true) {
 		dentry = offset_find_next(octx, ctx->pos);
 		if (!dentry)
-			return ERR_PTR(-ENOENT);
+			goto out_eod;
 
 		if (!offset_dir_emit(ctx, dentry)) {
 			dput(dentry);

@@ -497,7 +503,10 @@ static void *offset_iterate_dir(struct inode *inode, struct dir_context *ctx)
 		ctx->pos = dentry2offset(dentry) + 1;
 		dput(dentry);
 	}
-	return NULL;
+	return;
+
+out_eod:
+	ctx->pos = DIR_OFFSET_EOD;
 }
 
 /**

@@ -517,6 +526,8 @@ static void *offset_iterate_dir(struct inode *inode, struct dir_context *ctx)
  *
  * On return, @ctx->pos contains an offset that will read the next entry
  * in this directory when offset_readdir() is called again with @ctx.
+ * Caller places this value in the d_off field of the last entry in the
+ * user's buffer.
  *
  * Return values:
  *   %0 - Complete

@@ -529,13 +540,8 @@ static int offset_readdir(struct file *file, struct dir_context *ctx)
 
 	if (!dir_emit_dots(file, ctx))
 		return 0;
-
-	/* In this case, ->private_data is protected by f_pos_lock */
-	if (ctx->pos == DIR_OFFSET_MIN)
-		file->private_data = NULL;
-	else if (file->private_data == ERR_PTR(-ENOENT))
-		return 0;
-	file->private_data = offset_iterate_dir(d_inode(dir), ctx);
+	if (ctx->pos != DIR_OFFSET_EOD)
+		offset_iterate_dir(d_inode(dir), ctx);
 	return 0;
 }
 

Signature · 1 line

2.47.0

Quoted 23 lines · * [email protected] <[email protected]> [241220 10:33]:

From: Chuck Lever <[email protected]>

According to getdents(3), the d_off field in each returned directory entry points to the next entry in the directory. The d_off field in the last returned entry in the readdir buffer must contain a valid offset value, but if it points to an actual directory entry, then readdir/getdents can loop.

This patch introduces a specific fixed offset value that is placed in the d_off field of the last entry in a directory. Some user space applications assume that the EOD offset value is larger than the offsets of real directory entries, so the largest possible offset value is reserved for this purpose. This new value is never allocated by simple_offset_add().

When ->iterate_dir() returns, getdents{64} inserts the ctx->pos value into the d_off field of the last valid entry in the readdir buffer. When it hits EOD, offset_readdir() sets ctx->pos to the EOD offset value so the last entry is updated to point to the EOD marker.

When trying to read the entry at the EOD offset, offset_readdir() terminates immediately.

Quoted 4 lines

It is worth noting that using a Maple tree for directory offset value allocation does not guarantee a 63-bit range of values -- on platforms where "long" is a 32-bit type, the directory offset value range is still 0..(2^31 - 1).

I have a standing request to have 32-bit archs return 64-bit values. Is this another 'nice to have' 64 bit values on 32 bit archs?

Quoted 116 lines

Fixes 796432efab1e ("libfs: getdents() should return 0 after reaching EOD")

Signed-off-by Chuck Lever <[email protected]>

---
  fs/libfs.c | 38 ++++++++++++++++++++++----------------
  1 file changed, 22 insertions(+), 16 deletions(-)

fs/libfs.c +22 -16

index 8c9364a0174c..5c56783c03a5 100644
@@ -245,9 +245,16 @@ const struct inode_operations simple_dir_inode_operations = {
  };
  EXPORT_SYMBOL(simple_dir_inode_operations);
 
-/* 0 is '.', 1 is '..', so always start with offset 2 or more */
+/* simple_offset_add() allocation range */
  enum {
-	DIR_OFFSET_MIN	= 2,
+	DIR_OFFSET_MIN		= 2,
+	DIR_OFFSET_MAX		= LONG_MAX - 1,
+};
+
+/* simple_offset_add() never assigns these to a dentry */
+enum {
+	DIR_OFFSET_EOD		= LONG_MAX,	/* Marks EOD */
+
  };
 
  static void offset_set(struct dentry *dentry, long offset)

@@ -291,7 +298,8 @@ int simple_offset_add(struct offset_ctx *octx, struct dentry *dentry)
  		return -EBUSY;
 
  	ret = mtree_alloc_cyclic(&octx->mt, &offset, dentry, DIR_OFFSET_MIN,
-				 LONG_MAX, &octx->next_offset, GFP_KERNEL);
+				 DIR_OFFSET_MAX, &octx->next_offset,
+				 GFP_KERNEL);
  	if (unlikely(ret < 0))
  		return ret == -EBUSY ? -ENOSPC : ret;
 

@@ -447,8 +455,6 @@ static loff_t offset_dir_llseek(struct file *file, loff_t offset, int whence)
  		return -EINVAL;
  	}
 
-	/* In this case, ->private_data is protected by f_pos_lock */
-	file->private_data = NULL;
  	return vfs_setpos(file, offset, LONG_MAX);
  }
 

@@ -458,7 +464,7 @@ static struct dentry *offset_find_next(struct offset_ctx *octx, loff_t offset)
  	struct dentry *child, *found = NULL;
 
  	rcu_read_lock();
-	child = mas_find(&mas, LONG_MAX);
+	child = mas_find(&mas, DIR_OFFSET_MAX);
  	if (!child)
  		goto out;
  	spin_lock(&child->d_lock);

@@ -479,7 +485,7 @@ static bool offset_dir_emit(struct dir_context *ctx, struct dentry *dentry)
  			  inode->i_ino, fs_umode_to_dtype(inode->i_mode));
  }
 
-static void *offset_iterate_dir(struct inode *inode, struct dir_context *ctx)
+static void offset_iterate_dir(struct inode *inode, struct dir_context *ctx)
  {
  	struct offset_ctx *octx = inode->i_op->get_offset_ctx(inode);
  	struct dentry *dentry;

@@ -487,7 +493,7 @@ static void *offset_iterate_dir(struct inode *inode, struct dir_context *ctx)
  	while (true) {
  		dentry = offset_find_next(octx, ctx->pos);
  		if (!dentry)
-			return ERR_PTR(-ENOENT);
+			goto out_eod;
 
  		if (!offset_dir_emit(ctx, dentry)) {
  			dput(dentry);

@@ -497,7 +503,10 @@ static void *offset_iterate_dir(struct inode *inode, struct dir_context *ctx)
  		ctx->pos = dentry2offset(dentry) + 1;
  		dput(dentry);
  	}
-	return NULL;
+	return;
+
+out_eod:
+	ctx->pos = DIR_OFFSET_EOD;
  }
 
  /**

@@ -517,6 +526,8 @@ static void *offset_iterate_dir(struct inode *inode, struct dir_context *ctx)
   *
   * On return, @ctx->pos contains an offset that will read the next entry
   * in this directory when offset_readdir() is called again with @ctx.
+ * Caller places this value in the d_off field of the last entry in the
+ * user's buffer.
   *
   * Return values:
   *   %0 - Complete

@@ -529,13 +540,8 @@ static int offset_readdir(struct file *file, struct dir_context *ctx)
 
  	if (!dir_emit_dots(file, ctx))
  		return 0;
-
-	/* In this case, ->private_data is protected by f_pos_lock */
-	if (ctx->pos == DIR_OFFSET_MIN)
-		file->private_data = NULL;
-	else if (file->private_data == ERR_PTR(-ENOENT))
-		return 0;
-	file->private_data = offset_iterate_dir(d_inode(dir), ctx);
+	if (ctx->pos != DIR_OFFSET_EOD)
+		offset_iterate_dir(d_inode(dir), ctx);
  	return 0;
  }
 

Signature · 1 line

2.47.0

Quoted 23 lines · * [email protected] <[email protected]> [241220 10:33]:

From: Chuck Lever <[email protected]>

According to getdents(3), the d_off field in each returned directory entry points to the next entry in the directory. The d_off field in the last returned entry in the readdir buffer must contain a valid offset value, but if it points to an actual directory entry, then readdir/getdents can loop.

This patch introduces a specific fixed offset value that is placed in the d_off field of the last entry in a directory. Some user space applications assume that the EOD offset value is larger than the offsets of real directory entries, so the largest possible offset value is reserved for this purpose. This new value is never allocated by simple_offset_add().

When ->iterate_dir() returns, getdents{64} inserts the ctx->pos value into the d_off field of the last valid entry in the readdir buffer. When it hits EOD, offset_readdir() sets ctx->pos to the EOD offset value so the last entry is updated to point to the EOD marker.

When trying to read the entry at the EOD offset, offset_readdir() terminates immediately.

Quoted 4 lines

It is worth noting that using a Maple tree for directory offset value allocation does not guarantee a 63-bit range of values -- on platforms where "long" is a 32-bit type, the directory offset value range is still 0..(2^31 - 1).

I have a standing request to have 32-bit archs return 64-bit values. Is

From: Chuck Lever <[email protected]>

The mtree mechanism has been effective at creating directory offsets that are stable over multiple opendir instances. However, it has not been able to handle the subtleties of renames that are concurrent with readdir.

Instead of using the mtree to emit entries in the order of their offset values, use it only to map incoming ctx->pos to a starting entry. Then use the directory's d_children list, which is already maintained properly by the dcache, to find the next child to emit.

One of the sneaky things about this is that when the mtree-allocated offset value wraps (which is very rare), looking up ctx->pos++ is not going to find the next entry; it will return NULL. Instead, by following the d_children list, the offset values can appear in any order but all of the entries in the directory will be visited eventually.

Note also that the readdir() is guaranteed to reach the tail of this list. Entries are added only at the head of d_children, and readdir walks from its current position in that list towards its tail.

Signed-off-by Chuck Lever <[email protected]>

---
  fs/libfs.c | 84 +++++++++++++++++++++++++++++++++++++-----------------
  1 file changed, 58 insertions(+), 26 deletions(-)

fs/libfs.c +31 -10

index 5c56783c03a5..f7ead02062ad 100644
@@ -247,12 +247,13 @@ EXPORT_SYMBOL(simple_dir_inode_operations);
 
  /* simple_offset_add() allocation range */
  enum {
-	DIR_OFFSET_MIN		= 2,
+	DIR_OFFSET_MIN		= 3,
  	DIR_OFFSET_MAX		= LONG_MAX - 1,
  };
 
  /* simple_offset_add() never assigns these to a dentry */
  enum {
+	DIR_OFFSET_FIRST	= 2,		/* Find first real entry */
  	DIR_OFFSET_EOD		= LONG_MAX,	/* Marks EOD */
 
  };

@@ -458,51 +459,82 @@ static loff_t offset_dir_llseek(struct file *file, loff_t offset, int whence)
  	return vfs_setpos(file, offset, LONG_MAX);
  }
 
-static struct dentry *offset_find_next(struct offset_ctx *octx, loff_t offset)
+static struct dentry *find_positive_dentry(struct dentry *parent,
+					   struct dentry *dentry,
+					   bool next)
  {
-	MA_STATE(mas, &octx->mt, offset, offset);
+	struct dentry *found = NULL;
+
+	spin_lock(&parent->d_lock);
+	if (next)
+		dentry = d_next_sibling(dentry);
+	else if (!dentry)
+		dentry = d_first_child(parent);
+	hlist_for_each_entry_from(dentry, d_sib) {
+		if (!simple_positive(dentry))
+			continue;
+		spin_lock_nested(&dentry->d_lock, DENTRY_D_LOCK_NESTED);
+		if (simple_positive(dentry))
+			found = dget_dlock(dentry);
+		spin_unlock(&dentry->d_lock);
+		if (likely(found))
+			break;
+	}
+	spin_unlock(&parent->d_lock);
+	return found;
+}
+
+static noinline_for_stack struct dentry *
+offset_dir_lookup(struct dentry *parent, loff_t offset)
+{
+	struct inode *inode = d_inode(parent);
+	struct offset_ctx *octx = inode->i_op->get_offset_ctx(inode);
  	struct dentry *child, *found = NULL;
 
-	rcu_read_lock();
-	child = mas_find(&mas, DIR_OFFSET_MAX);
-	if (!child)
-		goto out;
-	spin_lock(&child->d_lock);
-	if (simple_positive(child))
-		found = dget_dlock(child);

fs/libfs.c +20 -13

+		found = find_positive_dentry(parent, child, false);
+		rcu_read_unlock();
+	}
  	return found;
  }
 
  static bool offset_dir_emit(struct dir_context *ctx, struct dentry *dentry)
  {
  	struct inode *inode = d_inode(dentry);
-	long offset = dentry2offset(dentry);
 
-	return ctx->actor(ctx, dentry->d_name.name, dentry->d_name.len, offset,
-			  inode->i_ino, fs_umode_to_dtype(inode->i_mode));
+	return dir_emit(ctx, dentry->d_name.name, dentry->d_name.len,
+			inode->i_ino, fs_umode_to_dtype(inode->i_mode));
  }
 
-static void offset_iterate_dir(struct inode *inode, struct dir_context *ctx)
+static void offset_iterate_dir(struct file *file, struct dir_context *ctx)
  {
-	struct offset_ctx *octx = inode->i_op->get_offset_ctx(inode);
+	struct dentry *dir = file->f_path.dentry;
  	struct dentry *dentry;
 
+	dentry = offset_dir_lookup(dir, ctx->pos);
+	if (!dentry)
+		goto out_eod;
  	while (true) {
-		dentry = offset_find_next(octx, ctx->pos);
-		if (!dentry)
-			goto out_eod;
+		struct dentry *next;
 
-		if (!offset_dir_emit(ctx, dentry)) {
-			dput(dentry);
+		ctx->pos = dentry2offset(dentry);
+		if (!offset_dir_emit(ctx, dentry))
  			break;
-		}
 
-		ctx->pos = dentry2offset(dentry) + 1;
+		next = find_positive_dentry(dir, dentry, true);
  		dput(dentry);
+
+		if (!next)
+			goto out_eod;
+		dentry = next;
  	}
+	dput(dentry);
  	return;
 
  out_eod:

@@ -541,7 +573,7 @@ static int offset_readdir(struct file *file, struct dir_context *ctx)
  	if (!dir_emit_dots(file, ctx))
  		return 0;
  	if (ctx->pos != DIR_OFFSET_EOD)
-		offset_iterate_dir(d_inode(dir), ctx);
+		offset_iterate_dir(file, ctx);
  	return 0;
  }
 

Quoted 93 lines · 在 2024/12/20 23:33, [email protected] 写道:

From: Chuck Lever <[email protected]>

The mtree mechanism has been effective at creating directory offsets that are stable over multiple opendir instances. However, it has not been able to handle the subtleties of renames that are concurrent with readdir.

Instead of using the mtree to emit entries in the order of their offset values, use it only to map incoming ctx->pos to a starting entry. Then use the directory's d_children list, which is already maintained properly by the dcache, to find the next child to emit.

One of the sneaky things about this is that when the mtree-allocated offset value wraps (which is very rare), looking up ctx->pos++ is not going to find the next entry; it will return NULL. Instead, by following the d_children list, the offset values can appear in any order but all of the entries in the directory will be visited eventually.

Note also that the readdir() is guaranteed to reach the tail of this list. Entries are added only at the head of d_children, and readdir walks from its current position in that list towards its tail.

Signed-off-by Chuck Lever <[email protected]>

---
  fs/libfs.c | 84 +++++++++++++++++++++++++++++++++++++-----------------

  1 file changed, 58 insertions(+), 26 deletions(-)

fs/libfs.c +2 -1

index 5c56783c03a5..f7ead02062ad 100644
@@ -247,12 +247,13 @@ EXPORT_SYMBOL(simple_dir_inode_operations);
   /* simple_offset_add() allocation range */
   enum {
-    DIR_OFFSET_MIN        = 2,
+    DIR_OFFSET_MIN        = 3,
       DIR_OFFSET_MAX        = LONG_MAX - 1,
   };
   /* simple_offset_add() never assigns these to a dentry */
   enum {
+    DIR_OFFSET_FIRST    = 2,        /* Find first real entry */

      DIR_OFFSET_EOD        = LONG_MAX,    /* Marks EOD */ };

fs/libfs.c +30 -12

@@ -458,51 +459,82 @@ static loff_t offset_dir_llseek(struct file 
 *file, loff_t offset, int whence)
       return vfs_setpos(file, offset, LONG_MAX);
   }
-static struct dentry *offset_find_next(struct offset_ctx *octx, 
 loff_t offset)
+static struct dentry *find_positive_dentry(struct dentry *parent,
+                       struct dentry *dentry,
+                       bool next)
   {
-    MA_STATE(mas, &octx->mt, offset, offset);
+    struct dentry *found = NULL;
+
+    spin_lock(&parent->d_lock);
+    if (next)
+        dentry = d_next_sibling(dentry);
+    else if (!dentry)
+        dentry = d_first_child(parent);
+    hlist_for_each_entry_from(dentry, d_sib) {
+        if (!simple_positive(dentry))
+            continue;
+        spin_lock_nested(&dentry->d_lock, DENTRY_D_LOCK_NESTED);
+        if (simple_positive(dentry))
+            found = dget_dlock(dentry);
+        spin_unlock(&dentry->d_lock);
+        if (likely(found))
+            break;
+    }
+    spin_unlock(&parent->d_lock);
+    return found;
+}
+
+static noinline_for_stack struct dentry *
+offset_dir_lookup(struct dentry *parent, loff_t offset)
+{
+    struct inode *inode = d_inode(parent);
+    struct offset_ctx *octx = inode->i_op->get_offset_ctx(inode);
       struct dentry *child, *found = NULL;
-    rcu_read_lock();
-    child = mas_find(&mas, DIR_OFFSET_MAX);
-    if (!child)
-        goto out;
-    spin_lock(&child->d_lock);
-    if (simple_positive(child))
-        found = dget_dlock(child);
-    spin_unlock(&child->d_lock);
-out:
-    rcu_read_unlock();
+    MA_STATE(mas, &octx->mt, offset, offset);

fs/libfs.c static loff_t offset_dir_llseek(struct file *file, loff_t offset, int whence) · line 500

Quoted 6 lines

fs/libfs.c +6 -0

+
+    if (offset == DIR_OFFSET_FIRST)
+        found = find_positive_dentry(parent, NULL, false);
+    else {
+        rcu_read_lock();
+        child = mas_find(&mas, DIR_OFFSET_MAX);

Can this child be NULL?

Like we delete some file after first readdir, maybe we should break here, or we may rescan all dentry and return them to userspace again?

Quoted 57 lines

fs/libfs.c +19 -12

+        found = find_positive_dentry(parent, child, false);
+        rcu_read_unlock();
+    }
       return found;
   }
   static bool offset_dir_emit(struct dir_context *ctx, struct dentry *dentry)
   {
       struct inode *inode = d_inode(dentry);
-    long offset = dentry2offset(dentry);
-    return ctx->actor(ctx, dentry->d_name.name, dentry->d_name.len, 
 offset,
-              inode->i_ino, fs_umode_to_dtype(inode->i_mode));
+    return dir_emit(ctx, dentry->d_name.name, dentry->d_name.len,
+            inode->i_ino, fs_umode_to_dtype(inode->i_mode));
   }
-static void offset_iterate_dir(struct inode *inode, struct 
 dir_context *ctx)
+static void offset_iterate_dir(struct file *file, struct dir_context 
 *ctx)
   {
-    struct offset_ctx *octx = inode->i_op->get_offset_ctx(inode);
+    struct dentry *dir = file->f_path.dentry;
       struct dentry *dentry;
+    dentry = offset_dir_lookup(dir, ctx->pos);
+    if (!dentry)
+        goto out_eod;
       while (true) {
-        dentry = offset_find_next(octx, ctx->pos);
-        if (!dentry)
-            goto out_eod;
+        struct dentry *next;
-        if (!offset_dir_emit(ctx, dentry)) {
-            dput(dentry);
+        ctx->pos = dentry2offset(dentry);
+        if (!offset_dir_emit(ctx, dentry))
               break;
-        }
-        ctx->pos = dentry2offset(dentry) + 1;
+        next = find_positive_dentry(dir, dentry, true);
           dput(dentry);
+
+        if (!next)
+            goto out_eod;
+        dentry = next;
       }
+    dput(dentry);
       return;

  out_eod:

fs/libfs.c +1 -1

@@ -541,7 +573,7 @@ static int offset_readdir(struct file *file, 
 struct dir_context *ctx)
       if (!dir_emit_dots(file, ctx))
           return 0;
       if (ctx->pos != DIR_OFFSET_EOD)
-        offset_iterate_dir(d_inode(dir), ctx);
+        offset_iterate_dir(file, ctx);
       return 0;
   }

Quoted 96 lines · yangerkun

Quoted 93 lines · 在 2024/12/20 23:33, [email protected] 写道:

From: Chuck Lever <[email protected]>

The mtree mechanism has been effective at creating directory offsets that are stable over multiple opendir instances. However, it has not been able to handle the subtleties of renames that are concurrent with readdir.

Instead of using the mtree to emit entries in the order of their offset values, use it only to map incoming ctx->pos to a starting entry. Then use the directory's d_children list, which is already maintained properly by the dcache, to find the next child to emit.

One of the sneaky things about this is that when the mtree-allocated offset value wraps (which is very rare), looking up ctx->pos++ is not going to find the next entry; it will return NULL. Instead, by following the d_children list, the offset values can appear in any order but all of the entries in the directory will be visited eventually.

Note also that the readdir() is guaranteed to reach the tail of this list. Entries are added only at the head of d_children, and readdir walks from its current position in that list towards its tail.

Signed-off-by Chuck Lever <[email protected]>

---
  fs/libfs.c | 84 +++++++++++++++++++++++++++++++++++++-----------------

  1 file changed, 58 insertions(+), 26 deletions(-)

fs/libfs.c +2 -1

index 5c56783c03a5..f7ead02062ad 100644
@@ -247,12 +247,13 @@ EXPORT_SYMBOL(simple_dir_inode_operations);
   /* simple_offset_add() allocation range */
   enum {
-    DIR_OFFSET_MIN        = 2,
+    DIR_OFFSET_MIN        = 3,
       DIR_OFFSET_MAX        = LONG_MAX - 1,
   };
   /* simple_offset_add() never assigns these to a dentry */
   enum {
+    DIR_OFFSET_FIRST    = 2,        /* Find first real entry */

      DIR_OFFSET_EOD        = LONG_MAX,    /* Marks EOD */ };

fs/libfs.c +30 -12

@@ -458,51 +459,82 @@ static loff_t offset_dir_llseek(struct file 
 *file, loff_t offset, int whence)
       return vfs_setpos(file, offset, LONG_MAX);
   }
-static struct dentry *offset_find_next(struct offset_ctx *octx, 
 loff_t offset)
+static struct dentry *find_positive_dentry(struct dentry *parent,
+                       struct dentry *dentry,
+                       bool next)
   {
-    MA_STATE(mas, &octx->mt, offset, offset);
+    struct dentry *found = NULL;
+
+    spin_lock(&parent->d_lock);
+    if (next)
+        dentry = d_next_sibling(dentry);
+    else if (!dentry)
+        dentry = d_first_child(parent);
+    hlist_for_each_entry_from(dentry, d_sib) {
+        if (!simple_positive(dentry))
+            continue;
+        spin_lock_nested(&dentry->d_lock, DENTRY_D_LOCK_NESTED);
+        if (simple_positive(dentry))
+            found = dget_dlock(dentry);
+        spin_unlock(&dentry->d_lock);
+        if (likely(found))
+            break;
+    }
+    spin_unlock(&parent->d_lock);
+    return found;
+}
+
+static noinline_for_stack struct dentry *
+offset_dir_lookup(struct dentry *parent, loff_t offset)
+{
+    struct inode *inode = d_inode(parent);
+    struct offset_ctx *octx = inode->i_op->get_offset_ctx(inode);
       struct dentry *child, *found = NULL;
-    rcu_read_lock();
-    child = mas_find(&mas, DIR_OFFSET_MAX);
-    if (!child)
-        goto out;
-    spin_lock(&child->d_lock);
-    if (simple_positive(child))
-        found = dget_dlock(child);
-    spin_unlock(&child->d_lock);
-out:
-    rcu_read_unlock();
+    MA_STATE(mas, &octx->mt, offset, offset);

Quoted 8 lines

fs/libfs.c static loff_t offset_dir_llseek(struct file *file, loff_t offset, int whence) · line 500

Quoted 6 lines

fs/libfs.c +6 -0

+
+    if (offset == DIR_OFFSET_FIRST)
+        found = find_positive_dentry(parent, NULL, false);
+    else {
+        rcu_read_lock();
+        child = mas_find(&mas, DIR_OFFSET_MAX);

Can this child be NULL?

Yes, this mas_find() call can return NULL. find_positive_dentry() should then return NULL. Kind of subtle.

Quoted 1 line

Like we delete some file after first readdir, maybe we should break here, or we may rescan all dentry and return them to userspace again?

You mean to deal with the case where the "next" entry has an offset that is lower than @offset? mas_find() will return the entry in the tree that is "at or after" mas->index.

I'm not sure either "break" or returning repeats is safe. But, now that you point it out, this function probably does need additional logic to deal with the offset wrap case.

But since this logic already exists here, IMO it is reasonable to leave that to be addressed by a subsequent patch. So far there aren't any regression test failures that warn of a user-visible problem the way it is now.

Quoted 60 lines

Quoted 59 lines

fs/libfs.c +19 -12

+        found = find_positive_dentry(parent, child, false);
+        rcu_read_unlock();
+    }
       return found;
   }
   static bool offset_dir_emit(struct dir_context *ctx, struct dentry *dentry)
   {
       struct inode *inode = d_inode(dentry);
-    long offset = dentry2offset(dentry);
-    return ctx->actor(ctx, dentry->d_name.name, dentry->d_name.len, 
 offset,
-              inode->i_ino, fs_umode_to_dtype(inode->i_mode));
+    return dir_emit(ctx, dentry->d_name.name, dentry->d_name.len,
+            inode->i_ino, fs_umode_to_dtype(inode->i_mode));
   }
-static void offset_iterate_dir(struct inode *inode, struct 
 dir_context *ctx)
+static void offset_iterate_dir(struct file *file, struct dir_context 
 *ctx)
   {
-    struct offset_ctx *octx = inode->i_op->get_offset_ctx(inode);
+    struct dentry *dir = file->f_path.dentry;
       struct dentry *dentry;
+    dentry = offset_dir_lookup(dir, ctx->pos);
+    if (!dentry)
+        goto out_eod;
       while (true) {
-        dentry = offset_find_next(octx, ctx->pos);
-        if (!dentry)
-            goto out_eod;
+        struct dentry *next;
-        if (!offset_dir_emit(ctx, dentry)) {
-            dput(dentry);
+        ctx->pos = dentry2offset(dentry);
+        if (!offset_dir_emit(ctx, dentry))
               break;
-        }
-        ctx->pos = dentry2offset(dentry) + 1;
+        next = find_positive_dentry(dir, dentry, true);
           dput(dentry);
+
+        if (!next)
+            goto out_eod;
+        dentry = next;
       }
+    dput(dentry);
       return;

  out_eod:

fs/libfs.c +1 -1

@@ -541,7 +573,7 @@ static int offset_readdir(struct file *file, 
 struct dir_context *ctx)
       if (!dir_emit_dots(file, ctx))
           return 0;
       if (ctx->pos != DIR_OFFSET_EOD)
-        offset_iterate_dir(d_inode(dir), ctx);
+        offset_iterate_dir(file, ctx);
       return 0;
   }

Quoted 117 lines · Chuck Lever

Quoted 97 lines · yangerkun

Quoted 94 lines · 在 2024/12/20 23:33, [email protected] 写道:

From: Chuck Lever <[email protected]>

The mtree mechanism has been effective at creating directory offsets that are stable over multiple opendir instances. However, it has not been able to handle the subtleties of renames that are concurrent with readdir.

Instead of using the mtree to emit entries in the order of their offset values, use it only to map incoming ctx->pos to a starting entry. Then use the directory's d_children list, which is already maintained properly by the dcache, to find the next child to emit.

One of the sneaky things about this is that when the mtree-allocated offset value wraps (which is very rare), looking up ctx->pos++ is not going to find the next entry; it will return NULL. Instead, by following the d_children list, the offset values can appear in any order but all of the entries in the directory will be visited eventually.

Note also that the readdir() is guaranteed to reach the tail of this list. Entries are added only at the head of d_children, and readdir walks from its current position in that list towards its tail.

Signed-off-by Chuck Lever <[email protected]>

---
  fs/libfs.c | 84 ++++++++++++++++++++++++++++++++++++
+-----------------

  1 file changed, 58 insertions(+), 26 deletions(-)

fs/libfs.c +2 -1

index 5c56783c03a5..f7ead02062ad 100644
@@ -247,12 +247,13 @@ EXPORT_SYMBOL(simple_dir_inode_operations);
   /* simple_offset_add() allocation range */
   enum {
-    DIR_OFFSET_MIN        = 2,
+    DIR_OFFSET_MIN        = 3,
       DIR_OFFSET_MAX        = LONG_MAX - 1,
   };
   /* simple_offset_add() never assigns these to a dentry */
   enum {
+    DIR_OFFSET_FIRST    = 2,        /* Find first real entry */

      DIR_OFFSET_EOD        = LONG_MAX,    /* Marks EOD */ };

fs/libfs.c +30 -12

@@ -458,51 +459,82 @@ static loff_t offset_dir_llseek(struct file 
 *file, loff_t offset, int whence)
       return vfs_setpos(file, offset, LONG_MAX);
   }
-static struct dentry *offset_find_next(struct offset_ctx *octx, 
 loff_t offset)
+static struct dentry *find_positive_dentry(struct dentry *parent,
+                       struct dentry *dentry,
+                       bool next)
   {
-    MA_STATE(mas, &octx->mt, offset, offset);
+    struct dentry *found = NULL;
+
+    spin_lock(&parent->d_lock);
+    if (next)
+        dentry = d_next_sibling(dentry);
+    else if (!dentry)
+        dentry = d_first_child(parent);
+    hlist_for_each_entry_from(dentry, d_sib) {
+        if (!simple_positive(dentry))
+            continue;
+        spin_lock_nested(&dentry->d_lock, DENTRY_D_LOCK_NESTED);
+        if (simple_positive(dentry))
+            found = dget_dlock(dentry);
+        spin_unlock(&dentry->d_lock);
+        if (likely(found))
+            break;
+    }
+    spin_unlock(&parent->d_lock);
+    return found;
+}
+
+static noinline_for_stack struct dentry *
+offset_dir_lookup(struct dentry *parent, loff_t offset)
+{
+    struct inode *inode = d_inode(parent);
+    struct offset_ctx *octx = inode->i_op->get_offset_ctx(inode);
       struct dentry *child, *found = NULL;
-    rcu_read_lock();
-    child = mas_find(&mas, DIR_OFFSET_MAX);
-    if (!child)
-        goto out;
-    spin_lock(&child->d_lock);
-    if (simple_positive(child))
-        found = dget_dlock(child);
-    spin_unlock(&child->d_lock);
-out:
-    rcu_read_unlock();
+    MA_STATE(mas, &octx->mt, offset, offset);

Quoted 8 lines

fs/libfs.c static loff_t offset_dir_llseek(struct file *file, loff_t offset, int whence) · line 500

Quoted 6 lines

fs/libfs.c +6 -0

+
+    if (offset == DIR_OFFSET_FIRST)
+        found = find_positive_dentry(parent, NULL, false);
+    else {
+        rcu_read_lock();
+        child = mas_find(&mas, DIR_OFFSET_MAX);

Can this child be NULL?

Yes, this mas_find() call can return NULL. find_positive_dentry() should then return NULL. Kind of subtle.

Quoted 1 line

Like we delete some file after first readdir, maybe we should break here, or we may rescan all dentry and return them to userspace again?

You mean to deal with the case where the "next" entry has an offset that is lower than @offset? mas_find() will return the entry in the tree that is "at or after" mas->index.

Quoted 8 lines

I'm not sure either "break" or returning repeats is safe. But, now that you point it out, this function probably does need additional logic to deal with the offset wrap case.

But since this logic already exists here, IMO it is reasonable to leave that to be addressed by a subsequent patch. So far there aren't any regression test failures that warn of a user-visible problem the way it is now.

Sorry for the confusing, the case I am talking is something like below:

mkdir /tmp/dir && cd /tmp/dir
touch file1 # offset is 3
touch file2 # offset is 4
touch file3 # offset is 5
touch file4 # offset is 6
touch file5 # offset is 7
first readdir and get file5 file4 file3 file2 #ctx->pos is 3, which means we will get file1 for second readdir

unlink file1 # can not get entry for ctx->pos == 3

second readdir # offset_dir_lookup will use mas_find but return NULL, and we will get file5 file4 file3 file2 again?

And for the offset wrap case, I prefer it's safe with your patch if we won't unlink file between two readdir. The second readdir will use an active ctx->pos which means there is a active dentry attach to this
ctx->pos. find_positive_dentry will stop once we meet the last child.

I am not sure if I understand correctly, if not, please point out!

Thanks!

Quoted 62 lines

Quoted 60 lines

Quoted 59 lines

fs/libfs.c +19 -12

+        found = find_positive_dentry(parent, child, false);
+        rcu_read_unlock();
+    }
       return found;
   }
   static bool offset_dir_emit(struct dir_context *ctx, struct dentry *dentry)
   {
       struct inode *inode = d_inode(dentry);
-    long offset = dentry2offset(dentry);
-    return ctx->actor(ctx, dentry->d_name.name, dentry->d_name.len, 
 offset,
-              inode->i_ino, fs_umode_to_dtype(inode->i_mode));
+    return dir_emit(ctx, dentry->d_name.name, dentry->d_name.len,
+            inode->i_ino, fs_umode_to_dtype(inode->i_mode));
   }
-static void offset_iterate_dir(struct inode *inode, struct 
 dir_context *ctx)
+static void offset_iterate_dir(struct file *file, struct 
 dir_context *ctx)
   {
-    struct offset_ctx *octx = inode->i_op->get_offset_ctx(inode);
+    struct dentry *dir = file->f_path.dentry;
       struct dentry *dentry;
+    dentry = offset_dir_lookup(dir, ctx->pos);
+    if (!dentry)
+        goto out_eod;
       while (true) {
-        dentry = offset_find_next(octx, ctx->pos);
-        if (!dentry)
-            goto out_eod;
+        struct dentry *next;
-        if (!offset_dir_emit(ctx, dentry)) {
-            dput(dentry);
+        ctx->pos = dentry2offset(dentry);
+        if (!offset_dir_emit(ctx, dentry))
               break;
-        }
-        ctx->pos = dentry2offset(dentry) + 1;
+        next = find_positive_dentry(dir, dentry, true);
           dput(dentry);
+
+        if (!next)
+            goto out_eod;
+        dentry = next;
       }
+    dput(dentry);
       return;

  out_eod:

fs/libfs.c +1 -1

@@ -541,7 +573,7 @@ static int offset_readdir(struct file *file, 
 struct dir_context *ctx)
       if (!dir_emit_dots(file, ctx))
           return 0;
       if (ctx->pos != DIR_OFFSET_EOD)
-        offset_iterate_dir(d_inode(dir), ctx);
+        offset_iterate_dir(file, ctx);
       return 0;
   }

Quoted 140 lines · yangerkun

Quoted 117 lines · Chuck Lever

Quoted 97 lines · yangerkun

Quoted 94 lines · 在 2024/12/20 23:33, [email protected] 写道:

From: Chuck Lever <[email protected]>

The mtree mechanism has been effective at creating directory offsets that are stable over multiple opendir instances. However, it has not been able to handle the subtleties of renames that are concurrent with readdir.

Instead of using the mtree to emit entries in the order of their offset values, use it only to map incoming ctx->pos to a starting entry. Then use the directory's d_children list, which is already maintained properly by the dcache, to find the next child to emit.

One of the sneaky things about this is that when the mtree-allocated offset value wraps (which is very rare), looking up ctx->pos++ is not going to find the next entry; it will return NULL. Instead, by following the d_children list, the offset values can appear in any order but all of the entries in the directory will be visited eventually.

Note also that the readdir() is guaranteed to reach the tail of this list. Entries are added only at the head of d_children, and readdir walks from its current position in that list towards its tail.

Signed-off-by Chuck Lever <[email protected]>

---
  fs/libfs.c | 84 ++++++++++++++++++++++++++++++++++++
+-----------------

  1 file changed, 58 insertions(+), 26 deletions(-)

fs/libfs.c +2 -1

index 5c56783c03a5..f7ead02062ad 100644
@@ -247,12 +247,13 @@ EXPORT_SYMBOL(simple_dir_inode_operations);
   /* simple_offset_add() allocation range */
   enum {
-    DIR_OFFSET_MIN        = 2,
+    DIR_OFFSET_MIN        = 3,
       DIR_OFFSET_MAX        = LONG_MAX - 1,
   };
   /* simple_offset_add() never assigns these to a dentry */
   enum {
+    DIR_OFFSET_FIRST    = 2,        /* Find first real entry */

      DIR_OFFSET_EOD        = LONG_MAX,    /* Marks EOD */ };

fs/libfs.c +30 -12

@@ -458,51 +459,82 @@ static loff_t offset_dir_llseek(struct file 
 *file, loff_t offset, int whence)
       return vfs_setpos(file, offset, LONG_MAX);
   }
-static struct dentry *offset_find_next(struct offset_ctx *octx, 
 loff_t offset)
+static struct dentry *find_positive_dentry(struct dentry *parent,
+                       struct dentry *dentry,
+                       bool next)
   {
-    MA_STATE(mas, &octx->mt, offset, offset);
+    struct dentry *found = NULL;
+
+    spin_lock(&parent->d_lock);
+    if (next)
+        dentry = d_next_sibling(dentry);
+    else if (!dentry)
+        dentry = d_first_child(parent);
+    hlist_for_each_entry_from(dentry, d_sib) {
+        if (!simple_positive(dentry))
+            continue;
+        spin_lock_nested(&dentry->d_lock, DENTRY_D_LOCK_NESTED);
+        if (simple_positive(dentry))
+            found = dget_dlock(dentry);
+        spin_unlock(&dentry->d_lock);
+        if (likely(found))
+            break;
+    }
+    spin_unlock(&parent->d_lock);
+    return found;
+}
+
+static noinline_for_stack struct dentry *
+offset_dir_lookup(struct dentry *parent, loff_t offset)
+{
+    struct inode *inode = d_inode(parent);
+    struct offset_ctx *octx = inode->i_op->get_offset_ctx(inode);
       struct dentry *child, *found = NULL;
-    rcu_read_lock();
-    child = mas_find(&mas, DIR_OFFSET_MAX);
-    if (!child)
-        goto out;
-    spin_lock(&child->d_lock);
-    if (simple_positive(child))
-        found = dget_dlock(child);
-    spin_unlock(&child->d_lock);
-out:
-    rcu_read_unlock();
+    MA_STATE(mas, &octx->mt, offset, offset);

Quoted 8 lines

fs/libfs.c static loff_t offset_dir_llseek(struct file *file, loff_t offset, int whence) · line 500

Quoted 6 lines

fs/libfs.c +6 -0

+
+    if (offset == DIR_OFFSET_FIRST)
+        found = find_positive_dentry(parent, NULL, false);
+    else {
+        rcu_read_lock();
+        child = mas_find(&mas, DIR_OFFSET_MAX);

Can this child be NULL?

Yes, this mas_find() call can return NULL. find_positive_dentry() should then return NULL. Kind of subtle.

Quoted 1 line

Like we delete some file after first readdir, maybe we should break here, or we may rescan all dentry and return them to userspace again?

You mean to deal with the case where the "next" entry has an offset that is lower than @offset? mas_find() will return the entry in the tree that is "at or after" mas->index.

Quoted 8 lines

I'm not sure either "break" or returning repeats is safe. But, now that you point it out, this function probably does need additional logic to deal with the offset wrap case.

But since this logic already exists here, IMO it is reasonable to leave that to be addressed by a subsequent patch. So far there aren't any regression test failures that warn of a user-visible problem the way it is now.

Sorry for the confusing, the case I am talking is something like below:

mkdir /tmp/dir && cd /tmp/dir
touch file1 # offset is 3
touch file2 # offset is 4
touch file3 # offset is 5
touch file4 # offset is 6
touch file5 # offset is 7
first readdir and get file5 file4 file3 file2 #ctx->pos is 3, which means we will get file1 for second readdir

Quoted 4 lines

unlink file1 # can not get entry for ctx->pos == 3

second readdir # offset_dir_lookup will use mas_find but return NULL, and we will get file5 file4 file3 file2 again?

After this patch, directory entries are reported in descending cookie order. Therefore, should this patch replace the mas_find() call with mas_find_rev() ?

Quoted 74 lines

And for the offset wrap case, I prefer it's safe with your patch if we won't unlink file between two readdir. The second readdir will use an active ctx->pos which means there is a active dentry attach to this
ctx->pos. find_positive_dentry will stop once we meet the last child.

I am not sure if I understand correctly, if not, please point out!

Thanks!

Quoted 62 lines

Quoted 60 lines

Quoted 59 lines

fs/libfs.c +19 -12

+        found = find_positive_dentry(parent, child, false);
+        rcu_read_unlock();
+    }
       return found;
   }
   static bool offset_dir_emit(struct dir_context *ctx, struct dentry *dentry)
   {
       struct inode *inode = d_inode(dentry);
-    long offset = dentry2offset(dentry);
-    return ctx->actor(ctx, dentry->d_name.name, 
 dentry->d_name.len, offset,
-              inode->i_ino, fs_umode_to_dtype(inode->i_mode));
+    return dir_emit(ctx, dentry->d_name.name, dentry->d_name.len,
+            inode->i_ino, fs_umode_to_dtype(inode->i_mode));
   }
-static void offset_iterate_dir(struct inode *inode, struct 
 dir_context *ctx)
+static void offset_iterate_dir(struct file *file, struct 
 dir_context *ctx)
   {
-    struct offset_ctx *octx = inode->i_op->get_offset_ctx(inode);
+    struct dentry *dir = file->f_path.dentry;
       struct dentry *dentry;
+    dentry = offset_dir_lookup(dir, ctx->pos);
+    if (!dentry)
+        goto out_eod;
       while (true) {
-        dentry = offset_find_next(octx, ctx->pos);
-        if (!dentry)
-            goto out_eod;
+        struct dentry *next;
-        if (!offset_dir_emit(ctx, dentry)) {
-            dput(dentry);
+        ctx->pos = dentry2offset(dentry);
+        if (!offset_dir_emit(ctx, dentry))
               break;
-        }
-        ctx->pos = dentry2offset(dentry) + 1;
+        next = find_positive_dentry(dir, dentry, true);
           dput(dentry);
+
+        if (!next)
+            goto out_eod;
+        dentry = next;
       }
+    dput(dentry);
       return;

  out_eod:

fs/libfs.c +1 -1

@@ -541,7 +573,7 @@ static int offset_readdir(struct file *file, 
 struct dir_context *ctx)
       if (!dir_emit_dots(file, ctx))
           return 0;
       if (ctx->pos != DIR_OFFSET_EOD)
-        offset_iterate_dir(d_inode(dir), ctx);
+        offset_iterate_dir(file, ctx);
       return 0;
   }

Quoted 146 lines · Chuck Lever

Quoted 144 lines · yangerkun

Quoted 117 lines · Chuck Lever

Quoted 97 lines · yangerkun

Quoted 94 lines · 在 2024/12/20 23:33, [email protected] 写道:

From: Chuck Lever <[email protected]>

The mtree mechanism has been effective at creating directory offsets that are stable over multiple opendir instances. However, it has not been able to handle the subtleties of renames that are concurrent with readdir.

Instead of using the mtree to emit entries in the order of their offset values, use it only to map incoming ctx->pos to a starting entry. Then use the directory's d_children list, which is already maintained properly by the dcache, to find the next child to emit.

One of the sneaky things about this is that when the mtree-allocated offset value wraps (which is very rare), looking up ctx->pos++ is not going to find the next entry; it will return NULL. Instead, by following the d_children list, the offset values can appear in any order but all of the entries in the directory will be visited eventually.

Note also that the readdir() is guaranteed to reach the tail of this list. Entries are added only at the head of d_children, and readdir walks from its current position in that list towards its tail.

Signed-off-by Chuck Lever <[email protected]>

---
  fs/libfs.c | 84 ++++++++++++++++++++++++++++++++++++
+-----------------

  1 file changed, 58 insertions(+), 26 deletions(-)

fs/libfs.c +2 -1

index 5c56783c03a5..f7ead02062ad 100644
@@ -247,12 +247,13 @@ EXPORT_SYMBOL(simple_dir_inode_operations);
   /* simple_offset_add() allocation range */
   enum {
-    DIR_OFFSET_MIN        = 2,
+    DIR_OFFSET_MIN        = 3,
       DIR_OFFSET_MAX        = LONG_MAX - 1,
   };
   /* simple_offset_add() never assigns these to a dentry */
   enum {
+    DIR_OFFSET_FIRST    = 2,        /* Find first real entry */

      DIR_OFFSET_EOD        = LONG_MAX,    /* Marks EOD */ };

fs/libfs.c +30 -12

@@ -458,51 +459,82 @@ static loff_t offset_dir_llseek(struct file 
 *file, loff_t offset, int whence)
       return vfs_setpos(file, offset, LONG_MAX);
   }
-static struct dentry *offset_find_next(struct offset_ctx *octx, 
 loff_t offset)
+static struct dentry *find_positive_dentry(struct dentry *parent,
+                       struct dentry *dentry,
+                       bool next)
   {
-    MA_STATE(mas, &octx->mt, offset, offset);
+    struct dentry *found = NULL;
+
+    spin_lock(&parent->d_lock);
+    if (next)
+        dentry = d_next_sibling(dentry);
+    else if (!dentry)
+        dentry = d_first_child(parent);
+    hlist_for_each_entry_from(dentry, d_sib) {
+        if (!simple_positive(dentry))
+            continue;
+        spin_lock_nested(&dentry->d_lock, DENTRY_D_LOCK_NESTED);
+        if (simple_positive(dentry))
+            found = dget_dlock(dentry);
+        spin_unlock(&dentry->d_lock);
+        if (likely(found))
+            break;
+    }
+    spin_unlock(&parent->d_lock);
+    return found;
+}
+
+static noinline_for_stack struct dentry *
+offset_dir_lookup(struct dentry *parent, loff_t offset)
+{
+    struct inode *inode = d_inode(parent);
+    struct offset_ctx *octx = inode->i_op->get_offset_ctx(inode);
       struct dentry *child, *found = NULL;
-    rcu_read_lock();
-    child = mas_find(&mas, DIR_OFFSET_MAX);
-    if (!child)
-        goto out;
-    spin_lock(&child->d_lock);
-    if (simple_positive(child))
-        found = dget_dlock(child);
-    spin_unlock(&child->d_lock);
-out:
-    rcu_read_unlock();
+    MA_STATE(mas, &octx->mt, offset, offset);

Quoted 8 lines

fs/libfs.c static loff_t offset_dir_llseek(struct file *file, loff_t offset, int whence) · line 500

Quoted 6 lines

fs/libfs.c +6 -0

+
+    if (offset == DIR_OFFSET_FIRST)
+        found = find_positive_dentry(parent, NULL, false);
+    else {
+        rcu_read_lock();
+        child = mas_find(&mas, DIR_OFFSET_MAX);

Can this child be NULL?

Yes, this mas_find() call can return NULL. find_positive_dentry() should then return NULL. Kind of subtle.

Quoted 1 line

Like we delete some file after first readdir, maybe we should break here, or we may rescan all dentry and return them to userspace again?

You mean to deal with the case where the "next" entry has an offset that is lower than @offset? mas_find() will return the entry in the tree that is "at or after" mas->index.

Quoted 8 lines

I'm not sure either "break" or returning repeats is safe. But, now that you point it out, this function probably does need additional logic to deal with the offset wrap case.

But since this logic already exists here, IMO it is reasonable to leave that to be addressed by a subsequent patch. So far there aren't any regression test failures that warn of a user-visible problem the way it is now.

Sorry for the confusing, the case I am talking is something like below:

mkdir /tmp/dir && cd /tmp/dir
touch file1 # offset is 3
touch file2 # offset is 4
touch file3 # offset is 5
touch file4 # offset is 6
touch file5 # offset is 7
first readdir and get file5 file4 file3 file2 #ctx->pos is 3, which means we will get file1 for second readdir

unlink file1 # can not get entry for ctx->pos == 3

second readdir # offset_dir_lookup will use mas_find but return NULL, and we will get file5 file4 file3 file2 again?

Quoted 3 lines

After this patch, directory entries are reported in descending cookie order. Therefore, should this patch replace the mas_find() call with mas_find_rev() ?

Emm... The reason that why readdir report file with descending cookie order is d_alloc will insert child dentry to the list head of &parent->d_subdirs, and find_positive_dentry will get child in order. So it seems this won't work?

Quoted 77 lines

Quoted 75 lines

And for the offset wrap case, I prefer it's safe with your patch if we won't unlink file between two readdir. The second readdir will use an active ctx->pos which means there is a active dentry attach to this
ctx->pos. find_positive_dentry will stop once we meet the last child.

I am not sure if I understand correctly, if not, please point out!

Thanks!

Quoted 62 lines

Quoted 60 lines

Quoted 59 lines

fs/libfs.c +19 -12

+        found = find_positive_dentry(parent, child, false);
+        rcu_read_unlock();
+    }
       return found;
   }
   static bool offset_dir_emit(struct dir_context *ctx, struct dentry *dentry)
   {
       struct inode *inode = d_inode(dentry);
-    long offset = dentry2offset(dentry);
-    return ctx->actor(ctx, dentry->d_name.name, 
 dentry->d_name.len, offset,
-              inode->i_ino, fs_umode_to_dtype(inode->i_mode));
+    return dir_emit(ctx, dentry->d_name.name, dentry->d_name.len,
+            inode->i_ino, fs_umode_to_dtype(inode->i_mode));
   }
-static void offset_iterate_dir(struct inode *inode, struct 
 dir_context *ctx)
+static void offset_iterate_dir(struct file *file, struct 
 dir_context *ctx)
   {
-    struct offset_ctx *octx = inode->i_op->get_offset_ctx(inode);
+    struct dentry *dir = file->f_path.dentry;
       struct dentry *dentry;
+    dentry = offset_dir_lookup(dir, ctx->pos);
+    if (!dentry)
+        goto out_eod;
       while (true) {
-        dentry = offset_find_next(octx, ctx->pos);
-        if (!dentry)
-            goto out_eod;
+        struct dentry *next;
-        if (!offset_dir_emit(ctx, dentry)) {
-            dput(dentry);
+        ctx->pos = dentry2offset(dentry);
+        if (!offset_dir_emit(ctx, dentry))
               break;
-        }
-        ctx->pos = dentry2offset(dentry) + 1;
+        next = find_positive_dentry(dir, dentry, true);
           dput(dentry);
+
+        if (!next)
+            goto out_eod;
+        dentry = next;
       }
+    dput(dentry);
       return;

  out_eod:

fs/libfs.c +1 -1

@@ -541,7 +573,7 @@ static int offset_readdir(struct file *file, 
 struct dir_context *ctx)
       if (!dir_emit_dots(file, ctx))
           return 0;
       if (ctx->pos != DIR_OFFSET_EOD)
-        offset_iterate_dir(d_inode(dir), ctx);
+        offset_iterate_dir(file, ctx);
       return 0;
   }

Quoted 153 lines · yangerkun

Quoted 149 lines · Chuck Lever

Quoted 140 lines · yangerkun

Quoted 117 lines · Chuck Lever

Quoted 97 lines · yangerkun

Quoted 94 lines · 在 2024/12/20 23:33, [email protected] 写道:

From: Chuck Lever <[email protected]>

The mtree mechanism has been effective at creating directory offsets that are stable over multiple opendir instances. However, it has not been able to handle the subtleties of renames that are concurrent with readdir.

Instead of using the mtree to emit entries in the order of their offset values, use it only to map incoming ctx->pos to a starting entry. Then use the directory's d_children list, which is already maintained properly by the dcache, to find the next child to emit.

One of the sneaky things about this is that when the mtree-allocated offset value wraps (which is very rare), looking up ctx->pos++ is not going to find the next entry; it will return NULL. Instead, by following the d_children list, the offset values can appear in any order but all of the entries in the directory will be visited eventually.

Note also that the readdir() is guaranteed to reach the tail of this list. Entries are added only at the head of d_children, and readdir walks from its current position in that list towards its tail.

Signed-off-by Chuck Lever <[email protected]>

---
  fs/libfs.c | 84 ++++++++++++++++++++++++++++++++++++
+-----------------

  1 file changed, 58 insertions(+), 26 deletions(-)

fs/libfs.c +2 -1

index 5c56783c03a5..f7ead02062ad 100644
@@ -247,12 +247,13 @@ EXPORT_SYMBOL(simple_dir_inode_operations);
   /* simple_offset_add() allocation range */
   enum {
-    DIR_OFFSET_MIN        = 2,
+    DIR_OFFSET_MIN        = 3,
       DIR_OFFSET_MAX        = LONG_MAX - 1,
   };
   /* simple_offset_add() never assigns these to a dentry */
   enum {
+    DIR_OFFSET_FIRST    = 2,        /* Find first real entry */

      DIR_OFFSET_EOD        = LONG_MAX,    /* Marks EOD */ };

fs/libfs.c +30 -12

@@ -458,51 +459,82 @@ static loff_t offset_dir_llseek(struct file 
 *file, loff_t offset, int whence)
       return vfs_setpos(file, offset, LONG_MAX);
   }
-static struct dentry *offset_find_next(struct offset_ctx *octx, 
 loff_t offset)
+static struct dentry *find_positive_dentry(struct dentry *parent,
+                       struct dentry *dentry,
+                       bool next)
   {
-    MA_STATE(mas, &octx->mt, offset, offset);
+    struct dentry *found = NULL;
+
+    spin_lock(&parent->d_lock);
+    if (next)
+        dentry = d_next_sibling(dentry);
+    else if (!dentry)
+        dentry = d_first_child(parent);
+    hlist_for_each_entry_from(dentry, d_sib) {
+        if (!simple_positive(dentry))
+            continue;
+        spin_lock_nested(&dentry->d_lock, DENTRY_D_LOCK_NESTED);
+        if (simple_positive(dentry))
+            found = dget_dlock(dentry);
+        spin_unlock(&dentry->d_lock);
+        if (likely(found))
+            break;
+    }
+    spin_unlock(&parent->d_lock);
+    return found;
+}
+
+static noinline_for_stack struct dentry *
+offset_dir_lookup(struct dentry *parent, loff_t offset)
+{
+    struct inode *inode = d_inode(parent);
+    struct offset_ctx *octx = inode->i_op->get_offset_ctx(inode);
       struct dentry *child, *found = NULL;
-    rcu_read_lock();
-    child = mas_find(&mas, DIR_OFFSET_MAX);
-    if (!child)
-        goto out;
-    spin_lock(&child->d_lock);
-    if (simple_positive(child))
-        found = dget_dlock(child);
-    spin_unlock(&child->d_lock);
-out:
-    rcu_read_unlock();
+    MA_STATE(mas, &octx->mt, offset, offset);

Quoted 8 lines

fs/libfs.c static loff_t offset_dir_llseek(struct file · line 500

Quoted 6 lines

fs/libfs.c +6 -0

+
+    if (offset == DIR_OFFSET_FIRST)
+        found = find_positive_dentry(parent, NULL, false);
+    else {
+        rcu_read_lock();
+        child = mas_find(&mas, DIR_OFFSET_MAX);

Can this child be NULL?

Yes, this mas_find() call can return NULL. find_positive_dentry() should then return NULL. Kind of subtle.

Quoted 1 line

Like we delete some file after first readdir, maybe we should break here, or we may rescan all dentry and return them to userspace again?

You mean to deal with the case where the "next" entry has an offset that is lower than @offset? mas_find() will return the entry in the tree that is "at or after" mas->index.

Quoted 8 lines

I'm not sure either "break" or returning repeats is safe. But, now that you point it out, this function probably does need additional logic to deal with the offset wrap case.

But since this logic already exists here, IMO it is reasonable to leave that to be addressed by a subsequent patch. So far there aren't any regression test failures that warn of a user-visible problem the way it is now.

Sorry for the confusing, the case I am talking is something like below:

mkdir /tmp/dir && cd /tmp/dir
touch file1 # offset is 3
touch file2 # offset is 4
touch file3 # offset is 5
touch file4 # offset is 6
touch file5 # offset is 7
first readdir and get file5 file4 file3 file2 #ctx->pos is 3, which means we will get file1 for second readdir

Quoted 4 lines

unlink file1 # can not get entry for ctx->pos == 3

second readdir # offset_dir_lookup will use mas_find but return NULL, and we will get file5 file4 file3 file2 again?

After this patch, directory entries are reported in descending cookie order. Therefore, should this patch replace the mas_find() call with mas_find_rev() ?

Quoted 4 lines

Emm... The reason that why readdir report file with descending cookie order is d_alloc will insert child dentry to the list head of &parent->d_subdirs, and find_positive_dentry will get child in order. So it seems this won't work?

I prefer this is not a problem since dcache_readdir already report dir with this order.

Quoted 76 lines

Quoted 74 lines

Quoted 72 lines

And for the offset wrap case, I prefer it's safe with your patch if we won't unlink file between two readdir. The second readdir will use an active ctx->pos which means there is a active dentry attach to this
ctx->pos. find_positive_dentry will stop once we meet the last child.

I am not sure if I understand correctly, if not, please point out!

Thanks!

Quoted 62 lines

Quoted 60 lines

Quoted 59 lines

fs/libfs.c +19 -12

+        found = find_positive_dentry(parent, child, false);
+        rcu_read_unlock();
+    }
       return found;
   }
   static bool offset_dir_emit(struct dir_context *ctx, struct dentry *dentry)
   {
       struct inode *inode = d_inode(dentry);
-    long offset = dentry2offset(dentry);
-    return ctx->actor(ctx, dentry->d_name.name, dentry- 
 >d_name.len, offset,
-              inode->i_ino, fs_umode_to_dtype(inode->i_mode));
+    return dir_emit(ctx, dentry->d_name.name, dentry->d_name.len,
+            inode->i_ino, fs_umode_to_dtype(inode->i_mode));
   }
-static void offset_iterate_dir(struct inode *inode, struct 
 dir_context *ctx)
+static void offset_iterate_dir(struct file *file, struct 
 dir_context *ctx)
   {
-    struct offset_ctx *octx = inode->i_op->get_offset_ctx(inode);
+    struct dentry *dir = file->f_path.dentry;
       struct dentry *dentry;
+    dentry = offset_dir_lookup(dir, ctx->pos);
+    if (!dentry)
+        goto out_eod;
       while (true) {
-        dentry = offset_find_next(octx, ctx->pos);
-        if (!dentry)
-            goto out_eod;
+        struct dentry *next;
-        if (!offset_dir_emit(ctx, dentry)) {
-            dput(dentry);
+        ctx->pos = dentry2offset(dentry);
+        if (!offset_dir_emit(ctx, dentry))
               break;
-        }
-        ctx->pos = dentry2offset(dentry) + 1;
+        next = find_positive_dentry(dir, dentry, true);
           dput(dentry);
+
+        if (!next)
+            goto out_eod;
+        dentry = next;
       }
+    dput(dentry);
       return;

  out_eod:

fs/libfs.c +1 -1

@@ -541,7 +573,7 @@ static int offset_readdir(struct file *file, 
 struct dir_context *ctx)
       if (!dir_emit_dots(file, ctx))
           return 0;
       if (ctx->pos != DIR_OFFSET_EOD)
-        offset_iterate_dir(d_inode(dir), ctx);
+        offset_iterate_dir(file, ctx);
       return 0;
   }

From: Chuck Lever <[email protected]>

The purpose of this series is to construct a set of upstream fixes that can be backported to v6.6 to address CVE-2024-46701.

Changes since v5:
- Improve error flow in simple_offset_add()

[...]