mbox series

[v3,0/5] vfs: clean up SETFLAGS and FSSETXATTR option processing

Message ID 156151632209.2283456.3592379873620132456.stgit@magnolia (mailing list archive)
Headers show
Series vfs: clean up SETFLAGS and FSSETXATTR option processing | expand

Message

Darrick J. Wong June 26, 2019, 2:32 a.m. UTC 
Hi all,

The FS_IOC_SETFLAGS and FS_IOC_FSSETXATTR ioctls were promoted from ext4
and XFS, respectively, into the VFS.  However, we didn't promote any of
the parameter checking code from those filesystems, which lead to a mess
where each filesystem open-codes whatever parameter checks they want and
the behavior across filesystems is no longer consistent.

Therefore, create some generic checking functions in the VFS and remove
all the open-coded pieces in each filesystem.  This preserves the
current behavior where a filesystem can choose to ignore fields it
doesn't understand.

If you're going to start using this mess, you probably ought to just
pull from my git trees, which are linked below.

This has been lightly tested with fstests.  Enjoy!
Comments and questions are, as always, welcome.

--D

kernel git tree:
https://git.kernel.org/cgit/linux/kernel/git/djwong/xfs-linux.git/log/?h=file-ioctl-cleanups

Comments

Al Viro June 26, 2019, 4:02 a.m. UTC | #1 
On Tue, Jun 25, 2019 at 07:32:10PM -0700, Darrick J. Wong wrote:
> From: Darrick J. Wong <darrick.wong@oracle.com>
> 
> Create a generic function to check incoming FS_IOC_SETFLAGS flag values
> and later prepare the inode for updates so that we can standardize the
> implementations that follow ext4's flag values.

Change in efivarfs behaviour deserves a note.  I'm not saying it's wrong,
but behaviour is changed there - no-op FS_IOC_SETFLAGS used to fail
without CAP_LINUX_IMMUTABLE...
Darrick J. Wong June 26, 2019, 3:30 p.m. UTC | #2 
On Wed, Jun 26, 2019 at 05:02:02AM +0100, Al Viro wrote:
> On Tue, Jun 25, 2019 at 07:32:10PM -0700, Darrick J. Wong wrote:
> > From: Darrick J. Wong <darrick.wong@oracle.com>
> > 
> > Create a generic function to check incoming FS_IOC_SETFLAGS flag values
> > and later prepare the inode for updates so that we can standardize the
> > implementations that follow ext4's flag values.
> 
> Change in efivarfs behaviour deserves a note.  I'm not saying it's wrong,
> but behaviour is changed there - no-op FS_IOC_SETFLAGS used to fail
> without CAP_LINUX_IMMUTABLE...

Ok, will do.  FWIW I wondered about fixing the orangefs usage too (using
an unsigned long uval to hold the contents of an unsigned int coming
from userspace) but ... that's their decision.

--D