From: Felix Fu <fuzhen5@huawei.com>

The return value of security_inode_set(remove)xattr should
be 1. If it return 0, cap_inode_setxattr would not be
executed when no lsm exist, which is not what we expected,
any user could set some security.* xattr for a file.

Before commit 260017f31a8c ("lsm: use default hook return
value in call_int_hook()") was approved, this issue would
still happened when lsm only include bpf, because bpf_lsm_
inode_setxattr return 0 by default which cause cap_inode_set
xattr to be not executed.

Fixes: 260017f31a8c ("lsm: use default hook return value in call_int_hook()")
Signed-off-by: Felix Fu <fuzhen5@huawei.com>
---
 include/linux/lsm_hook_defs.h | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/include/linux/lsm_hook_defs.h b/include/linux/lsm_hook_defs.h
index f804b76cde44..9c768b954264 100644
--- a/include/linux/lsm_hook_defs.h
+++ b/include/linux/lsm_hook_defs.h
@@ -144,14 +144,14 @@ LSM_HOOK(int, 0, inode_setattr, struct mnt_idmap *idmap, struct dentry *dentry,
 LSM_HOOK(void, LSM_RET_VOID, inode_post_setattr, struct mnt_idmap *idmap,
 	 struct dentry *dentry, int ia_valid)
 LSM_HOOK(int, 0, inode_getattr, const struct path *path)
-LSM_HOOK(int, 0, inode_setxattr, struct mnt_idmap *idmap,
+LSM_HOOK(int, 1, inode_setxattr, struct mnt_idmap *idmap,
 	 struct dentry *dentry, const char *name, const void *value,
 	 size_t size, int flags)
 LSM_HOOK(void, LSM_RET_VOID, inode_post_setxattr, struct dentry *dentry,
 	 const char *name, const void *value, size_t size, int flags)
 LSM_HOOK(int, 0, inode_getxattr, struct dentry *dentry, const char *name)
 LSM_HOOK(int, 0, inode_listxattr, struct dentry *dentry)
-LSM_HOOK(int, 0, inode_removexattr, struct mnt_idmap *idmap,
+LSM_HOOK(int, 1, inode_removexattr, struct mnt_idmap *idmap,
 	 struct dentry *dentry, const char *name)
 LSM_HOOK(void, LSM_RET_VOID, inode_post_removexattr, struct dentry *dentry,
 	 const char *name)
-- 
2.34.1

On Mon, Apr 29, 2024 at 7:47 AM felix <fuzhen5@huawei.com> wrote:
>
> From: Felix Fu <fuzhen5@huawei.com>
>
> The return value of security_inode_set(remove)xattr should
> be 1. If it return 0, cap_inode_setxattr would not be
> executed when no lsm exist, which is not what we expected,
> any user could set some security.* xattr for a file.
>
> Before commit 260017f31a8c ("lsm: use default hook return
> value in call_int_hook()") was approved, this issue would
> still happened when lsm only include bpf, because bpf_lsm_
> inode_setxattr return 0 by default which cause cap_inode_set
> xattr to be not executed.
>
> Fixes: 260017f31a8c ("lsm: use default hook return value in call_int_hook()")
> Signed-off-by: Felix Fu <fuzhen5@huawei.com>
> ---
>  include/linux/lsm_hook_defs.h | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)

Adding the LSM list as that is the important list for this patch.  I'm
also just realizing we don't include lsm_hook_defs.h in the LSM
MAINTAINERS entry, I'll submit a patch for that shortly.

> diff --git a/include/linux/lsm_hook_defs.h b/include/linux/lsm_hook_defs.h
> index f804b76cde44..9c768b954264 100644
> --- a/include/linux/lsm_hook_defs.h
> +++ b/include/linux/lsm_hook_defs.h
> @@ -144,14 +144,14 @@ LSM_HOOK(int, 0, inode_setattr, struct mnt_idmap *idmap, struct dentry *dentry,
>  LSM_HOOK(void, LSM_RET_VOID, inode_post_setattr, struct mnt_idmap *idmap,
>          struct dentry *dentry, int ia_valid)
>  LSM_HOOK(int, 0, inode_getattr, const struct path *path)
> -LSM_HOOK(int, 0, inode_setxattr, struct mnt_idmap *idmap,
> +LSM_HOOK(int, 1, inode_setxattr, struct mnt_idmap *idmap,
>          struct dentry *dentry, const char *name, const void *value,
>          size_t size, int flags)
>  LSM_HOOK(void, LSM_RET_VOID, inode_post_setxattr, struct dentry *dentry,
>          const char *name, const void *value, size_t size, int flags)
>  LSM_HOOK(int, 0, inode_getxattr, struct dentry *dentry, const char *name)
>  LSM_HOOK(int, 0, inode_listxattr, struct dentry *dentry)
> -LSM_HOOK(int, 0, inode_removexattr, struct mnt_idmap *idmap,
> +LSM_HOOK(int, 1, inode_removexattr, struct mnt_idmap *idmap,
>          struct dentry *dentry, const char *name)
>  LSM_HOOK(void, LSM_RET_VOID, inode_post_removexattr, struct dentry *dentry,
>          const char *name)
> --
> 2.34.1

-- 
paul-moore.com

On Wed, May 1, 2024 at 12:02 PM Paul Moore <paul@paul-moore.com> wrote:
> On Mon, Apr 29, 2024 at 7:47 AM felix <fuzhen5@huawei.com> wrote:
> >
> > From: Felix Fu <fuzhen5@huawei.com>
> >
> > The return value of security_inode_set(remove)xattr should
> > be 1. If it return 0, cap_inode_setxattr would not be
> > executed when no lsm exist, which is not what we expected,
> > any user could set some security.* xattr for a file.
> >
> > Before commit 260017f31a8c ("lsm: use default hook return
> > value in call_int_hook()") was approved, this issue would
> > still happened when lsm only include bpf, because bpf_lsm_
> > inode_setxattr return 0 by default which cause cap_inode_set
> > xattr to be not executed.
> >
> > Fixes: 260017f31a8c ("lsm: use default hook return value in call_int_hook()")
> > Signed-off-by: Felix Fu <fuzhen5@huawei.com>
> > ---
> >  include/linux/lsm_hook_defs.h | 4 ++--
> >  1 file changed, 2 insertions(+), 2 deletions(-)
>
> Adding the LSM list as that is the important list for this patch.

It's also worth noting the discussion below from earlier this year.  I
just spent a little bit of time working on a different solution which
I personally find more acceptable; I'm building a test kernel now,
assuming it works I'll post it as a RFC.

https://lore.kernel.org/linux-security-module/20240129133058.1627971-1-omosnace@redhat.com/

-- 
paul-moore.com