Re: [PATCH v7] vfs, security: Fix automount superblock LSM init problem, preventing NFS sb sharing

[Christian Brauner]

On Mon, Aug 07, 2023 at 08:57:03AM -0400, Jeff Layton wrote:
> On Sat, 2023-08-05 at 14:43 +0200, Christian Brauner wrote:
> > On Fri, Aug 04, 2023 at 12:09:34PM -0400, Jeff Layton wrote:
> > > From: David Howells <dhowells@xxxxxxxxxx>
> > > 
> > > When NFS superblocks are created by automounting, their LSM parameters
> > > aren't set in the fs_context struct prior to sget_fc() being called,
> > > leading to failure to match existing superblocks.
> > > 
> > > This bug leads to messages like the following appearing in dmesg when
> > > fscache is enabled:
> > > 
> > >     NFS: Cache volume key already in use (nfs,4.2,2,108,106a8c0,1,,,,100000,100000,2ee,3a98,1d4c,3a98,1)
> > > 
> > > Fix this by adding a new LSM hook to load fc->security for submount
> > > creation when alloc_fs_context() is creating the fs_context for it.
> > > 
> > > Signed-off-by: David Howells <dhowells@xxxxxxxxxx>
> > > Signed-off-by: Jeff Layton <jlayton@xxxxxxxxxx>
> > > Fixes: 9bc61ab18b1d ("vfs: Introduce fs_context, switch vfs_kern_mount() to it.")
> > > Fixes: 779df6a5480f ("NFS: Ensure security label is set for root inode)
> > > Tested-by: Jeff Layton <jlayton@xxxxxxxxxx>
> > > Reviewed-by: Jeff Layton <jlayton@xxxxxxxxxx>
> > > Acked-by: Casey Schaufler <casey@xxxxxxxxxxxxxxxx>
> > > Acked-by: "Christian Brauner (Microsoft)" <brauner@xxxxxxxxxx>
> > > Link: https://lore.kernel.org/r/165962680944.3334508.6610023900349142034.stgit@xxxxxxxxxxxxxxxxxxxxxx/ # v1
> > > Link: https://lore.kernel.org/r/165962729225.3357250.14350728846471527137.stgit@xxxxxxxxxxxxxxxxxxxxxx/ # v2
> > > Link: https://lore.kernel.org/r/165970659095.2812394.6868894171102318796.stgit@xxxxxxxxxxxxxxxxxxxxxx/ # v3
> > > Link: https://lore.kernel.org/r/166133579016.3678898.6283195019480567275.stgit@xxxxxxxxxxxxxxxxxxxxxx/ # v4
> > > Link: https://lore.kernel.org/r/217595.1662033775@xxxxxxxxxxxxxxxxxxxxxx/ # v5
> > > ---
> > > ver #7)
> > >  - Drop lsm_set boolean
> > >  - Link to v6: https://lore.kernel.org/r/20230802-master-v6-1-45d48299168b@xxxxxxxxxx
> > > 
> > > ver #6)
> > >  - Rebase onto v6.5.0-rc4
> > > 
> > > ver #5)
> > >  - Removed unused variable.
> > >  - Only allocate smack_mnt_opts if we're dealing with a submount.
> > > 
> > > ver #4)
> > >  - When doing a FOR_SUBMOUNT mount, don't set the root label in SELinux or
> > >    Smack.
> > > 
> > > ver #3)
> > >  - Made LSM parameter extraction dependent on fc->purpose ==
> > >    FS_CONTEXT_FOR_SUBMOUNT.  Shouldn't happen on FOR_RECONFIGURE.
> > > 
> > > ver #2)
> > >  - Added Smack support
> > >  - Made LSM parameter extraction dependent on reference != NULL.
> > > ---
> > >  fs/fs_context.c               |  4 ++++
> > >  include/linux/lsm_hook_defs.h |  1 +
> > >  include/linux/security.h      |  6 +++++
> > >  security/security.c           | 14 +++++++++++
> > >  security/selinux/hooks.c      | 25 ++++++++++++++++++++
> > >  security/smack/smack_lsm.c    | 54 +++++++++++++++++++++++++++++++++++++++++++
> > >  6 files changed, 104 insertions(+)
> > > 
> > > diff --git a/fs/fs_context.c b/fs/fs_context.c
> > > index 851214d1d013..a523aea956c4 100644
> > > --- a/fs/fs_context.c
> > > +++ b/fs/fs_context.c
> > > @@ -282,6 +282,10 @@ static struct fs_context *alloc_fs_context(struct file_system_type *fs_type,
> > >  		break;
> > >  	}
> > >  
> > > +	ret = security_fs_context_init(fc, reference);
> > > +	if (ret < 0)
> > > +		goto err_fc;
> > > +
> > >  	/* TODO: Make all filesystems support this unconditionally */
> > >  	init_fs_context = fc->fs_type->init_fs_context;
> > >  	if (!init_fs_context)
> > > diff --git a/include/linux/lsm_hook_defs.h b/include/linux/lsm_hook_defs.h
> > > index 7308a1a7599b..7ce3550154b1 100644
> > > --- a/include/linux/lsm_hook_defs.h
> > > +++ b/include/linux/lsm_hook_defs.h
> > > @@ -54,6 +54,7 @@ LSM_HOOK(int, 0, bprm_creds_from_file, struct linux_binprm *bprm, struct file *f
> > >  LSM_HOOK(int, 0, bprm_check_security, struct linux_binprm *bprm)
> > >  LSM_HOOK(void, LSM_RET_VOID, bprm_committing_creds, struct linux_binprm *bprm)
> > >  LSM_HOOK(void, LSM_RET_VOID, bprm_committed_creds, struct linux_binprm *bprm)
> > > +LSM_HOOK(int, 0, fs_context_init, struct fs_context *fc, struct dentry *reference)
> > >  LSM_HOOK(int, 0, fs_context_dup, struct fs_context *fc,
> > >  	 struct fs_context *src_sc)
> > >  LSM_HOOK(int, -ENOPARAM, fs_context_parse_param, struct fs_context *fc,
> > > diff --git a/include/linux/security.h b/include/linux/security.h
> > > index 32828502f09e..61fda06fac9d 100644
> > > --- a/include/linux/security.h
> > > +++ b/include/linux/security.h
> > > @@ -293,6 +293,7 @@ int security_bprm_creds_from_file(struct linux_binprm *bprm, struct file *file);
> > >  int security_bprm_check(struct linux_binprm *bprm);
> > >  void security_bprm_committing_creds(struct linux_binprm *bprm);
> > >  void security_bprm_committed_creds(struct linux_binprm *bprm);
> > > +int security_fs_context_init(struct fs_context *fc, struct dentry *reference);
> > >  int security_fs_context_dup(struct fs_context *fc, struct fs_context *src_fc);
> > >  int security_fs_context_parse_param(struct fs_context *fc, struct fs_parameter *param);
> > >  int security_sb_alloc(struct super_block *sb);
> > > @@ -629,6 +630,11 @@ static inline void security_bprm_committed_creds(struct linux_binprm *bprm)
> > >  {
> > >  }
> > >  
> > > +static inline int security_fs_context_init(struct fs_context *fc,
> > > +					   struct dentry *reference)
> > 
> > I think that's the wrong way of doing this hook. The security hook
> > really doesn't belong into alloc_fs_context().
> > 
> > I think what we want is a dedicated helper similar to vfs_dup_context():
> > 
> > // Only pass the superblock. There's no need for the dentry. I would
> > // avoid even passing fs_context but if that's preferred then sure.
> > security_fs_context_submount(struct fs_context *fc, const struct super_block *sb)
> > 
> > vfs_submount_fs_context(struct file_system_type *fs_type, struct dentry *reference)
> > {
> >         fc = fs_context_for_submount(fs_type, reference);
> > 
> >         security_fs_context_for_submount(fc, reference->d_sb);
> > }
> > 
> > This automatically ensures it's only called for submounts, the LSM
> > doesn't need to care about fc->purpose and this isn't called
> > in a pure allocation function for all allocation calls.
> > 
> > The we should switch all callers over to that new helper and unexport
> > that fs_context_for_submount() thing completely. Yes, that's more work
> > but that's the correct thing to do. And we need to audit fuse, cifs,
> > afs, and nfs anyway that they work fine with the new security hook.*
> > 
> 
> It's the same prototype. We could just move the hook call to the end of
> fs_context_for_submount, and that would be less churn for its callers.

If you just add it into fs_context_for_submount() after the allocation
and then leave that as first class citizen it's fine ofc. It's the same
result as what I mentioned earlier. We just shouldn't put generic hooks
in an allocation function that allocates different types of filesystem
contexts. I prefer to have them as targeted as possible and also avoid
unnecessary trips into the LSM layer.

> Or were you wanting to do that to make this a more gradual changeover
> for some reason?

No, I think it's fine if you switch it all in one go. I just don't know
if fuse/virtiofs submounts expect an selinux context to be bestowed upon
them.