Skip to content

cloud-kernel
cloud-kernel

openanolis / cloud-kernel
接近 2 年 前同步成功

通知 170
Star 36
Fork 7
cloud-kernel
cloud-kernel
切换分支/标签
  1. 16 12月, 2016 3 次提交
  3. 29 11月, 2016 1 次提交
    • M
      ovl: fix d_real() for stacked fs · c4fcfc16
      由 Miklos Szeredi 提交于 
      Handling of recursion in d_real() is completely broken.  Recursion is only
done in the 'inode != NULL' case.  But when opening the file we have
'inode == NULL' hence d_real() will return an overlay dentry.  This won't
work since overlayfs doesn't define its own file operations, so all file
ops will fail.

Fix by doing the recursion first and the check against the inode second.

Bash script to reproduce the issue written by Quentin:

 - 8< - - - - - 8< - - - - - 8< - - - - - 8< - - - -
tmpdir=$(mktemp -d)
pushd ${tmpdir}

mkdir -p {upper,lower,work}
echo -n 'rocks' > lower/ksplice
mount -t overlay level_zero upper -o lowerdir=lower,upperdir=upper,workdir=work
cat upper/ksplice

tmpdir2=$(mktemp -d)
pushd ${tmpdir2}

mkdir -p {upper,work}
mount -t overlay level_one upper -o lowerdir=${tmpdir}/upper,upperdir=upper,workdir=work
ls -l upper/ksplice
cat upper/ksplice
 - 8< - - - - - 8< - - - - - 8< - - - - - 8< - - - - 
Reported-by: NQuentin Casasnovas <quentin.casasnovas@oracle.com>
Signed-off-by: NMiklos Szeredi <mszeredi@redhat.com>
Fixes: 2d902671 ("vfs: merge .d_select_inode() into .d_real()")
Cc: <stable@vger.kernel.org> # v4.8+
      c4fcfc16
  5. 31 10月, 2016 3 次提交
  7. 15 10月, 2016 1 次提交
  9. 14 10月, 2016 3 次提交
  11. 08 10月, 2016 2 次提交
  13. 27 9月, 2016 1 次提交
  15. 22 9月, 2016 1 次提交
  17. 21 9月, 2016 1 次提交
  19. 19 9月, 2016 1 次提交
    • V
      ovl: during copy up, switch to mounter's creds early · 8eac98b8
      由 Vivek Goyal 提交于 
      Now, we have the notion that copy up of a file is done with the creds
of mounter of overlay filesystem (as opposed to task). Right now before
we switch creds, we do some vfs_getattr() operations in the context of
task and that itself can fail. We should do that getattr() using the
creds of mounter instead.

So this patch switches to mounter's creds early during copy up process so
that even vfs_getattr() is done with mounter's creds.

Do not call revert_creds() unless we have already called
ovl_override_creds(). [Reported by Arnd Bergmann]
Signed-off-by: NVivek Goyal <vgoyal@redhat.com>
Signed-off-by: NMiklos Szeredi <mszeredi@redhat.com>
      8eac98b8
  21. 16 9月, 2016 3 次提交
    • M
      ovl: lookup: do getxattr with mounter's permission · 2b6bc7f4
      由 Miklos Szeredi 提交于 
      The getxattr() in ovl_is_opaquedir() was missed when converting all
operations on underlying fs to be done under mounter's permission.

This patch fixes this by moving the ovl_override_creds()/revert_creds() out
from ovl_lookup_real() to ovl_lookup().

Also convert to using vfs_getxattr() instead of directly calling
i_op->getxattr().
Signed-off-by: NMiklos Szeredi <mszeredi@redhat.com>
      2b6bc7f4
    • M
      ovl: copy_up_xattr(): use strnlen · 8b326c61
      由 Miklos Szeredi 提交于 
      Be defensive about what underlying fs provides us in the returned xattr
list buffer.  strlen() may overrun the buffer, so use strnlen() and WARN if
the contents are not properly null terminated.
Signed-off-by: NMiklos Szeredi <mszeredi@redhat.com>
Cc: <stable@vger.kernel.org>
      8b326c61
    • M
      locks: fix file locking on overlayfs · c568d683
      由 Miklos Szeredi 提交于 
      This patch allows flock, posix locks, ofd locks and leases to work
correctly on overlayfs.

Instead of using the underlying inode for storing lock context use the
overlay inode.  This allows locks to be persistent across copy-up.

This is done by introducing locks_inode() helper and using it instead of
file_inode() to get the inode in locking code.  For non-overlayfs the two
are equivalent, except for an extra pointer dereference in locks_inode().

Since lock operations are in "struct file_operations" we must also make
sure not to call underlying filesystem's lock operations.  Introcude a
super block flag MS_NOREMOTELOCK to this effect.
Signed-off-by: NMiklos Szeredi <mszeredi@redhat.com>
Acked-by: NJeff Layton <jlayton@poochiereds.net>
Cc: "J. Bruce Fields" <bfields@fieldses.org>
      c568d683
  23. 05 9月, 2016 1 次提交
  25. 01 9月, 2016 12 次提交
  27. 09 8月, 2016 3 次提交
  29. 08 8月, 2016 1 次提交
    • M
      ovl: don't copy up opaqueness · 0956254a
      由 Miklos Szeredi 提交于 
      When a copy up of a directory occurs which has the opaque xattr set, the
xattr remains in the upper directory. The immediate behavior with overlayfs
is that the upper directory is not treated as opaque, however after a
remount the opaque flag is used and upper directory is treated as opaque.
This causes files created in the lower layer to be hidden when using
multiple lower directories.

Fix by not copying up the opaque flag.

To reproduce:

 ----8<---------8<---------8<---------8<---------8<---------8<----
mkdir -p l/d/s u v w mnt
mount -t overlay overlay -olowerdir=l,upperdir=u,workdir=w mnt
rm -rf mnt/d/
mkdir -p mnt/d/n
umount mnt
mount -t overlay overlay -olowerdir=u:l,upperdir=v,workdir=w mnt
touch mnt/d/foo
umount mnt
mount -t overlay overlay -olowerdir=u:l,upperdir=v,workdir=w mnt
ls mnt/d
 ----8<---------8<---------8<---------8<---------8<---------8<----
 
output should be:  "foo  n"
Reported-by: NDerek McGowan <dmcg@drizz.net>
Link: https://bugzilla.kernel.org/show_bug.cgi?id=151291Signed-off-by: NMiklos Szeredi <mszeredi@redhat.com>
Cc: <stable@vger.kernel.org>
      0956254a
  31. 29 7月, 2016 3 次提交