Skip to contentSkip to conversation

Assorted overlayfs fixes for 4.19-rc5

PATCH In review Amir Goldstein 3 participants 9 messages Sep 18, 2018 – Sep 24, 2018

Subsystems (5)
lore
Share & tools
lore

Amir Goldstein

author maintainer
#0 lore raw

cover Assorted overlayfs fixes for 4.19-rc5

Miklos,

Found one memory leak that needs a fix for stable and two missing freeze protections in stacked f_ops.

Last patch is not a bugfix, but it fixes a trap that we set for developers back in v4.10 - and you happened to step into.

The series passed xfstests -g quick with overlayfs over xfs and ext4 and the usual unionmount testsuite torture.

The memory leak was detected with tests overlay/051,053 after I enabled kmemleak on my test setup.

Thanks,
Amir.

Amir Goldstein (4):
ovl: fix memory leak on unlink of indexed file
ovl: fix freeze protection bypass in ovl_write_iter()
ovl: fix freeze protection bypass in ovl_clone_file_range()
vfs: swap names of {do,vfs}_clone_file_range()

 fs/ioctl.c             |  2 +-
 fs/nfsd/vfs.c          |  3 ++-
 fs/overlayfs/copy_up.c |  2 +-
 fs/overlayfs/file.c    |  2 ++
 fs/overlayfs/util.c    |  3 ++-
 fs/read_write.c        | 17 +++++++++++++++--
 include/linux/fs.h     | 17 +++--------------
 7 files changed, 26 insertions(+), 20 deletions(-)
Signature · 1 line
2.17.1
author maintainer
#1 lore raw

PATCH 1/4 ovl: fix memory leak on unlink of indexed file

Fixes caf70cb2ba5d ("ovl: cleanup orphan index entries")
Cc <[email protected]> # v4.13
---
 fs/overlayfs/util.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
fs/overlayfs/util.c +2 -1
index 8cfb62cc8672..ace4fe4c39a9 100644
@@ -683,7 +683,7 @@ static void ovl_cleanup_index(struct dentry *dentry)
 	struct dentry *upperdentry = ovl_dentry_upper(dentry);
 	struct dentry *index = NULL;
 	struct inode *inode;
-	struct qstr name;
+	struct qstr name = { };
 	int err;
 
 	err = ovl_get_index_name(lowerdentry, &name);
@@ -726,6 +726,7 @@ static void ovl_cleanup_index(struct dentry *dentry)
 		goto fail;
 
 out:
+	kfree(name.name);
 	dput(index);
 	return;
 
Signature · 1 line
2.17.1
maintainer
#2 lore raw

Re: [PATCH 1/4] ovl: fix memory leak on unlink of indexed file

Quoted 6 lines · Amir Goldstein
Fixes caf70cb2ba5d ("ovl: cleanup orphan index entries")
Cc <[email protected]> # v4.13
---
 fs/overlayfs/util.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

I know I reject patches without any changelog text present. Maybe Miklos is more forgiving than me... :)

greg k-h

author maintainer
#3 lore raw

Re: [PATCH 1/4] ovl: fix memory leak on unlink of indexed file

Quoted 9 lines · Greg KH
Quoted 6 lines · Amir Goldstein
Fixes caf70cb2ba5d ("ovl: cleanup orphan index entries")
Cc <[email protected]> # v4.13
---
 fs/overlayfs/util.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
Quoted 3 lines

I know I reject patches without any changelog text present. Maybe Miklos is more forgiving than me... :)

Well, I did not intend to CC stable at this point, but since I did and since you chimed in, Miklos, feel free to add to commit message:

The memory leak was detected by kmemleak when running xfstests overlay/051,053

Thanks,
Amir.

author maintainer
#4 lore raw

PATCH 2/4 ovl: fix freeze protection bypass in ovl_write_iter()

In reply to Amir Goldstein

Tested by re-writing to an open overlayfs file while upper ext4 is frozen:

xfs_io -f /ovl/x
xfs_io> pwrite 0 4096
fsfreeze -f /ext4
xfs_io> pwrite 0 4096

  WARNING: CPU: 0 PID: 1492 at fs/ext4/ext4_jbd2.c:53 \
           ext4_journal_check_start+0x48/0x82

After the fix, the second write blocks in ovl_write_iter() and avoids hitting WARN_ON(sb->s_writers.frozen == SB_FREEZE_COMPLETE) in ext4_journal_check_start().

Fixes 2a92e07edc5e ("ovl: add ovl_write_iter()")
---
 fs/overlayfs/file.c | 2 ++
 1 file changed, 2 insertions(+)
fs/overlayfs/file.c +2 -0
index aeaefd2a551b..986313da0c88 100644
@@ -240,8 +240,10 @@ static ssize_t ovl_write_iter(struct kiocb *iocb, struct iov_iter *iter)
 		goto out_unlock;
 
 	old_cred = ovl_override_creds(file_inode(file)->i_sb);
+	file_start_write(real.file);
 	ret = vfs_iter_write(real.file, iter, &iocb->ki_pos,
 			     ovl_iocb_to_rwf(iocb));
+	file_end_write(real.file);
 	revert_creds(old_cred);
 
 	/* Update size */
Signature · 1 line
2.17.1
author maintainer
#5 lore raw

PATCH 3/4 ovl: fix freeze protection bypass in ovl_clone_file_range()

In reply to Amir Goldstein

Tested by doing clone on overlayfs while upper xfs+reflink is frozen:

xfs_io -f /ovl/y
fsfreeze -f /xfs
xfs_io> reflink /ovl/x

Before the fix xfs_io enters xfs_reflink_remap_range() and blocks in xfs_trans_alloc(). After the fix, xfs_io blocks outside xfs code in ovl_clone_file_range().

Fixes 8ede205541ff ("ovl: add reflink/copyfile/dedup support")
---
 fs/overlayfs/file.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
fs/overlayfs/file.c +1 -1
index 986313da0c88..5d1b4b38f743 100644
@@ -461,7 +461,7 @@ static ssize_t ovl_copyfile(struct file *file_in, loff_t pos_in,
 		break;
 
 	case OVL_CLONE:
-		ret = vfs_clone_file_range(real_in.file, pos_in,
+		ret = do_clone_file_range(real_in.file, pos_in,
 					   real_out.file, pos_out, len);
 		break;
 
Signature · 1 line
2.17.1
author maintainer
#6 lore raw

PATCH 4/4 vfs: swap names of {do,vfs}_clone_file_range()

In reply to Amir Goldstein

Commit 031a072a0b8a ("vfs: call vfs_clone_file_range() under freeze protection") created a wrapper do_clone_file_range() around vfs_clone_file_range() moving the freeze protection to former, so overlayfs could call the latter.

The more common vfs practice is to call do_xxx helpers from vfs_xxx helpers, where freeze protecction is taken in the vfs_xxx helper, so this anomality could be a source of confusion.

It seems that commit 8ede205541ff ("ovl: add reflink/copyfile/dedup support") may have fallen a victim to this confusion - ovl_clone_file_range() calls the vfs_clone_file_range() helper in the hope of getting freeze protection on upper fs, but in fact results in overlayfs allowing to bypass upper fs freeze protection.

Swap the names of the two helpers to conform to common vfs practice and call the correct helpers from overlayfs and nfsd.

---
 fs/ioctl.c             |  2 +-
 fs/nfsd/vfs.c          |  3 ++-
 fs/overlayfs/copy_up.c |  2 +-
 fs/overlayfs/file.c    |  2 +-
 fs/read_write.c        | 17 +++++++++++++++--
 include/linux/fs.h     | 17 +++--------------
 6 files changed, 23 insertions(+), 20 deletions(-)
fs/ioctl.c +1 -1
index 3212c29235ce..2005529af560 100644
@@ -230,7 +230,7 @@ static long ioctl_file_clone(struct file *dst_file, unsigned long srcfd,
 	ret = -EXDEV;
 	if (src_file.file->f_path.mnt != dst_file->f_path.mnt)
 		goto fdput;
-	ret = do_clone_file_range(src_file.file, off, dst_file, destoff, olen);
+	ret = vfs_clone_file_range(src_file.file, off, dst_file, destoff, olen);
 fdput:
 	fdput(src_file);
 	return ret;
fs/nfsd/vfs.c +2 -1
index 55a099e47ba2..b53e76391e52 100644
@@ -541,7 +541,8 @@ __be32 nfsd4_set_nfs4_label(struct svc_rqst *rqstp, struct svc_fh *fhp,
 __be32 nfsd4_clone_file_range(struct file *src, u64 src_pos, struct file *dst,
 		u64 dst_pos, u64 count)
 {
-	return nfserrno(do_clone_file_range(src, src_pos, dst, dst_pos, count));
+	return nfserrno(vfs_clone_file_range(src, src_pos, dst, dst_pos,
+					     count));
 }
 
 ssize_t nfsd_copy_file_range(struct file *src, u64 src_pos, struct file *dst,
fs/overlayfs/copy_up.c +1 -1
index 296037afecdb..1cc797a08a5b 100644
@@ -141,7 +141,7 @@ static int ovl_copy_up_data(struct path *old, struct path *new, loff_t len)
 	}
 
 	/* Try to use clone_file_range to clone up within the same fs */
-	error = vfs_clone_file_range(old_file, 0, new_file, 0, len);
+	error = do_clone_file_range(old_file, 0, new_file, 0, len);
 	if (!error)
 		goto out;
 	/* Couldn't clone, so now we try to copy the data */
fs/overlayfs/file.c +1 -1
index 5d1b4b38f743..986313da0c88 100644
@@ -461,7 +461,7 @@ static ssize_t ovl_copyfile(struct file *file_in, loff_t pos_in,
 		break;
 
 	case OVL_CLONE:
-		ret = do_clone_file_range(real_in.file, pos_in,
+		ret = vfs_clone_file_range(real_in.file, pos_in,
 					   real_out.file, pos_out, len);
 		break;
 
fs/read_write.c +15 -2
index 39b4a21dd933..8a2737f0d61d 100644
@@ -1818,8 +1818,8 @@ int vfs_clone_file_prep_inodes(struct inode *inode_in, loff_t pos_in,
 }
 EXPORT_SYMBOL(vfs_clone_file_prep_inodes);
 
-int vfs_clone_file_range(struct file *file_in, loff_t pos_in,
-		struct file *file_out, loff_t pos_out, u64 len)
+int do_clone_file_range(struct file *file_in, loff_t pos_in,
+			struct file *file_out, loff_t pos_out, u64 len)
 {
 	struct inode *inode_in = file_inode(file_in);
 	struct inode *inode_out = file_inode(file_out);
@@ -1866,6 +1866,19 @@ int vfs_clone_file_range(struct file *file_in, loff_t pos_in,
 
 	return ret;
 }
+EXPORT_SYMBOL(do_clone_file_range);
+
+int vfs_clone_file_range(struct file *file_in, loff_t pos_in,
+			 struct file *file_out, loff_t pos_out, u64 len)
+{
+	int ret;
+
+	file_start_write(file_out);
+	ret = do_clone_file_range(file_in, pos_in, file_out, pos_out, len);
+	file_end_write(file_out);
+
+	return ret;
+}
 EXPORT_SYMBOL(vfs_clone_file_range);
 
 /*
include/linux/fs.h +3 -14
index 4f0a11d0d296..25a449f37bb1 100644
@@ -1833,8 +1833,10 @@ extern ssize_t vfs_copy_file_range(struct file *, loff_t , struct file *,
 extern int vfs_clone_file_prep_inodes(struct inode *inode_in, loff_t pos_in,
 				      struct inode *inode_out, loff_t pos_out,
 				      u64 *len, bool is_dedupe);
+extern int do_clone_file_range(struct file *file_in, loff_t pos_in,
+			       struct file *file_out, loff_t pos_out, u64 len);
 extern int vfs_clone_file_range(struct file *file_in, loff_t pos_in,
-		struct file *file_out, loff_t pos_out, u64 len);
+				struct file *file_out, loff_t pos_out, u64 len);
 extern int vfs_dedupe_file_range_compare(struct inode *src, loff_t srcoff,
 					 struct inode *dest, loff_t destoff,
 					 loff_t len, bool *is_same);
@@ -2778,19 +2780,6 @@ static inline void file_end_write(struct file *file)
 	__sb_end_write(file_inode(file)->i_sb, SB_FREEZE_WRITE);
 }
 
-static inline int do_clone_file_range(struct file *file_in, loff_t pos_in,
-				      struct file *file_out, loff_t pos_out,
-				      u64 len)
-{
-	int ret;
-
-	file_start_write(file_out);
-	ret = vfs_clone_file_range(file_in, pos_in, file_out, pos_out, len);
-	file_end_write(file_out);
-
-	return ret;
-}
-
 /*
  * get_write_access() gets write permission for a file.
  * put_write_access() releases this write permission.
Signature · 1 line
2.17.1
author maintainer
#7 lore raw
In reply to Amir Goldstein
Quoted 6 lines · Amir Goldstein

Miklos,

Found one memory leak that needs a fix for stable and two missing freeze protections in stacked f_ops.

Quoted 9 lines

Last patch is not a bugfix, but it fixes a trap that we set for developers back in v4.10 - and you happened to step into.

The series passed xfstests -g quick with overlayfs over xfs and ext4 and the usual unionmount testsuite torture.

The memory leak was detected with tests overlay/051,053 after I enabled kmemleak on my test setup.

Miklos,

Maybe not enough time has passed to justify a PING, but you know I get edgy towards the end of a release cycle, especially a weird release cycle as this one...

P.S. removing Al from CC because his emails have been bouncing. Al, are you there?

Thanks,
Amir.

maintainer
#8 lore raw
Quoted 17 lines · Amir Goldstein
Quoted 6 lines · Amir Goldstein

Miklos,

Found one memory leak that needs a fix for stable and two missing freeze protections in stacked f_ops.

Quoted 9 lines

Last patch is not a bugfix, but it fixes a trap that we set for developers back in v4.10 - and you happened to step into.

The series passed xfstests -g quick with overlayfs over xfs and ext4 and the usual unionmount testsuite torture.

The memory leak was detected with tests overlay/051,053 after I enabled kmemleak on my test setup.

Quoted 5 lines

Miklos,

Maybe not enough time has passed to justify a PING, but you know I get edgy towards the end of a release cycle, especially a weird release cycle as this one...

Thanks for the reminder. Pushed to overlayfs-next and will ask Greg
to pull before -rc6. If I forget, please remind again ;)

Thanks,
Miklos