diff mbox series

[2/2] security.capability: fix conversions on getxattr

Message ID 20210119162204.2081137-3-mszeredi@redhat.com (mailing list archive)
State New
Headers show
Series capability conversion fixes | expand

Commit Message

Miklos Szeredi Jan. 19, 2021, 4:22 p.m. UTC
If a capability is stored on disk in v2 format cap_inode_getsecurity() will
currently return in v2 format unconditionally.

This is wrong: v2 cap should be equivalent to a v3 cap with zero rootid,
and so the same conversions performed on it.

If the rootid cannot be mapped v3 is returned unconverted.  Fix this so
that both v2 and v3 return -EOVERFLOW if the rootid (or the owner of the fs
user namespace in case of v2) cannot be mapped in the current user
namespace.

Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>
---
 security/commoncap.c | 67 ++++++++++++++++++++++++++++----------------
 1 file changed, 43 insertions(+), 24 deletions(-)

Comments

Eric W. Biederman Jan. 20, 2021, 1:34 a.m. UTC | #1
Miklos Szeredi <mszeredi@redhat.com> writes:

> If a capability is stored on disk in v2 format cap_inode_getsecurity() will
> currently return in v2 format unconditionally.
>
> This is wrong: v2 cap should be equivalent to a v3 cap with zero rootid,
> and so the same conversions performed on it.
>
> If the rootid cannot be mapped v3 is returned unconverted.  Fix this so
> that both v2 and v3 return -EOVERFLOW if the rootid (or the owner of the fs
> user namespace in case of v2) cannot be mapped in the current user
> namespace.

This looks like a good cleanup.

I do wonder how well this works with stacking.  In particular
ovl_xattr_set appears to call vfs_getxattr without overriding the creds.
What the purpose of that is I haven't quite figured out.  It looks like
it is just a probe to see if an xattr is present so maybe it is ok.

Acked-by: "Eric W. Biederman" <ebiederm@xmission.com>

>
> Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>
> ---
>  security/commoncap.c | 67 ++++++++++++++++++++++++++++----------------
>  1 file changed, 43 insertions(+), 24 deletions(-)
>
> diff --git a/security/commoncap.c b/security/commoncap.c
> index bacc1111d871..c9d99f8f4c82 100644
> --- a/security/commoncap.c
> +++ b/security/commoncap.c
> @@ -371,10 +371,11 @@ int cap_inode_getsecurity(struct inode *inode, const char *name, void **buffer,
>  {
>  	int size, ret;
>  	kuid_t kroot;
> +	__le32 nsmagic, magic;
>  	uid_t root, mappedroot;
>  	char *tmpbuf = NULL;
>  	struct vfs_cap_data *cap;
> -	struct vfs_ns_cap_data *nscap;
> +	struct vfs_ns_cap_data *nscap = NULL;
>  	struct dentry *dentry;
>  	struct user_namespace *fs_ns;
>  
> @@ -396,46 +397,61 @@ int cap_inode_getsecurity(struct inode *inode, const char *name, void **buffer,
>  	fs_ns = inode->i_sb->s_user_ns;
>  	cap = (struct vfs_cap_data *) tmpbuf;
>  	if (is_v2header((size_t) ret, cap)) {
> -		/* If this is sizeof(vfs_cap_data) then we're ok with the
> -		 * on-disk value, so return that.  */
> -		if (alloc)
> -			*buffer = tmpbuf;
> -		else
> -			kfree(tmpbuf);
> -		return ret;
> -	} else if (!is_v3header((size_t) ret, cap)) {
> -		kfree(tmpbuf);
> -		return -EINVAL;
> +		root = 0;
> +	} else if (is_v3header((size_t) ret, cap)) {
> +		nscap = (struct vfs_ns_cap_data *) tmpbuf;
> +		root = le32_to_cpu(nscap->rootid);
> +	} else {
> +		size = -EINVAL;
> +		goto out_free;
>  	}
>  
> -	nscap = (struct vfs_ns_cap_data *) tmpbuf;
> -	root = le32_to_cpu(nscap->rootid);
>  	kroot = make_kuid(fs_ns, root);
>  
>  	/* If the root kuid maps to a valid uid in current ns, then return
>  	 * this as a nscap. */
>  	mappedroot = from_kuid(current_user_ns(), kroot);
>  	if (mappedroot != (uid_t)-1 && mappedroot != (uid_t)0) {
> +		size = sizeof(struct vfs_ns_cap_data);
>  		if (alloc) {
> -			*buffer = tmpbuf;
> +			if (!nscap) {
> +				/* v2 -> v3 conversion */
> +				nscap = kzalloc(size, GFP_ATOMIC);
> +				if (!nscap) {
> +					size = -ENOMEM;
> +					goto out_free;
> +				}
> +				nsmagic = VFS_CAP_REVISION_3;
> +				magic = le32_to_cpu(cap->magic_etc);
> +				if (magic & VFS_CAP_FLAGS_EFFECTIVE)
> +					nsmagic |= VFS_CAP_FLAGS_EFFECTIVE;
> +				memcpy(&nscap->data, &cap->data, sizeof(__le32) * 2 * VFS_CAP_U32);
> +				nscap->magic_etc = cpu_to_le32(nsmagic);
> +			} else {
> +				/* use allocated v3 buffer */
> +				tmpbuf = NULL;
> +			}
>  			nscap->rootid = cpu_to_le32(mappedroot);
> -		} else
> -			kfree(tmpbuf);
> -		return size;
> +			*buffer = nscap;
> +		}
> +		goto out_free;
>  	}
>  
>  	if (!rootid_owns_currentns(kroot)) {
> -		kfree(tmpbuf);
> -		return -EOPNOTSUPP;
> +		size = -EOVERFLOW;
> +		goto out_free;
>  	}
>  
>  	/* This comes from a parent namespace.  Return as a v2 capability */
>  	size = sizeof(struct vfs_cap_data);
>  	if (alloc) {
> -		*buffer = kmalloc(size, GFP_ATOMIC);
> -		if (*buffer) {
> -			struct vfs_cap_data *cap = *buffer;
> -			__le32 nsmagic, magic;
> +		if (nscap) {
> +			/* v3 -> v2 conversion */
> +			cap = kzalloc(size, GFP_ATOMIC);
> +			if (!cap) {
> +				size = -ENOMEM;
> +				goto out_free;
> +			}
>  			magic = VFS_CAP_REVISION_2;
>  			nsmagic = le32_to_cpu(nscap->magic_etc);
>  			if (nsmagic & VFS_CAP_FLAGS_EFFECTIVE)
> @@ -443,9 +459,12 @@ int cap_inode_getsecurity(struct inode *inode, const char *name, void **buffer,
>  			memcpy(&cap->data, &nscap->data, sizeof(__le32) * 2 * VFS_CAP_U32);
>  			cap->magic_etc = cpu_to_le32(magic);
>  		} else {
> -			size = -ENOMEM;
> +			/* use unconverted v2 */
> +			tmpbuf = NULL;
>  		}
> +		*buffer = cap;
>  	}
> +out_free:
>  	kfree(tmpbuf);
>  	return size;
>  }
Miklos Szeredi Jan. 20, 2021, 7:58 a.m. UTC | #2
On Wed, Jan 20, 2021 at 2:39 AM Eric W. Biederman <ebiederm@xmission.com> wrote:
>
> Miklos Szeredi <mszeredi@redhat.com> writes:
>
> > If a capability is stored on disk in v2 format cap_inode_getsecurity() will
> > currently return in v2 format unconditionally.
> >
> > This is wrong: v2 cap should be equivalent to a v3 cap with zero rootid,
> > and so the same conversions performed on it.
> >
> > If the rootid cannot be mapped v3 is returned unconverted.  Fix this so
> > that both v2 and v3 return -EOVERFLOW if the rootid (or the owner of the fs
> > user namespace in case of v2) cannot be mapped in the current user
> > namespace.
>
> This looks like a good cleanup.
>
> I do wonder how well this works with stacking.  In particular
> ovl_xattr_set appears to call vfs_getxattr without overriding the creds.
> What the purpose of that is I haven't quite figured out.  It looks like
> it is just a probe to see if an xattr is present so maybe it is ok.

Yeah, it's checking in the removexattr case whether copy-up is needed
or not (i.e. if trying to remove a non-existent xattr, then no need to
copy up).

But for consistency it should also be wrapped in override creds.
Adding fix to this series.

I'll also audit for any remaining omissions.  One known and documented
case is vfs_ioctl(FS_IOC_{[SG]ETFLAGS,FS[SG]ETXATTR}), but that
shouldn't be affected by user namespaces.

Thanks,
Miklos
kernel test robot Jan. 20, 2021, 7:37 p.m. UTC | #3
Hi Miklos,

I love your patch! Perhaps something to improve:

[auto build test WARNING on security/next-testing]
[also build test WARNING on linux/master linus/master v5.11-rc4 next-20210120]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch]

url:    https://github.com/0day-ci/linux/commits/Miklos-Szeredi/capability-conversion-fixes/20210120-152933
base:   https://git.kernel.org/pub/scm/linux/kernel/git/jmorris/linux-security.git next-testing
config: xtensa-randconfig-s032-20210120 (attached as .config)
compiler: xtensa-linux-gcc (GCC) 9.3.0
reproduce:
        wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross
        chmod +x ~/bin/make.cross
        # apt-get install sparse
        # sparse version: v0.6.3-208-g46a52ca4-dirty
        # https://github.com/0day-ci/linux/commit/bcf70adf8bcc3e52cb1b262ae2e1d9154da75097
        git remote add linux-review https://github.com/0day-ci/linux
        git fetch --no-tags linux-review Miklos-Szeredi/capability-conversion-fixes/20210120-152933
        git checkout bcf70adf8bcc3e52cb1b262ae2e1d9154da75097
        # save the attached .config to linux build tree
        COMPILER_INSTALL_PATH=$HOME/0day COMPILER=gcc-9.3.0 make.cross C=1 CF='-fdiagnostic-prefix -D__CHECK_ENDIAN__' ARCH=xtensa 

If you fix the issue, kindly add following tag as appropriate
Reported-by: kernel test robot <lkp@intel.com>


"sparse warnings: (new ones prefixed by >>)"
   security/commoncap.c:424:41: sparse: sparse: incorrect type in assignment (different base types) @@     expected restricted __le32 [usertype] nsmagic @@     got int @@
   security/commoncap.c:424:41: sparse:     expected restricted __le32 [usertype] nsmagic
   security/commoncap.c:424:41: sparse:     got int
>> security/commoncap.c:425:39: sparse: sparse: incorrect type in assignment (different base types) @@     expected restricted __le32 [usertype] magic @@     got unsigned int @@
   security/commoncap.c:425:39: sparse:     expected restricted __le32 [usertype] magic
   security/commoncap.c:425:39: sparse:     got unsigned int
   security/commoncap.c:426:37: sparse: sparse: restricted __le32 degrades to integer
   security/commoncap.c:427:49: sparse: sparse: invalid assignment: |=
   security/commoncap.c:427:49: sparse:    left side has type restricted __le32
   security/commoncap.c:427:49: sparse:    right side has type int
   security/commoncap.c:429:52: sparse: sparse: cast from restricted __le32
>> security/commoncap.c:429:52: sparse: sparse: incorrect type in argument 1 (different base types) @@     expected unsigned int [usertype] val @@     got restricted __le32 [usertype] nsmagic @@
   security/commoncap.c:429:52: sparse:     expected unsigned int [usertype] val
   security/commoncap.c:429:52: sparse:     got restricted __le32 [usertype] nsmagic
   security/commoncap.c:429:52: sparse: sparse: cast from restricted __le32
   security/commoncap.c:429:52: sparse: sparse: cast from restricted __le32
   security/commoncap.c:429:52: sparse: sparse: cast from restricted __le32
   security/commoncap.c:429:52: sparse: sparse: cast from restricted __le32
   security/commoncap.c:455:31: sparse: sparse: incorrect type in assignment (different base types) @@     expected restricted __le32 [usertype] magic @@     got int @@
   security/commoncap.c:455:31: sparse:     expected restricted __le32 [usertype] magic
   security/commoncap.c:455:31: sparse:     got int
   security/commoncap.c:456:33: sparse: sparse: incorrect type in assignment (different base types) @@     expected restricted __le32 [usertype] nsmagic @@     got unsigned int @@
   security/commoncap.c:456:33: sparse:     expected restricted __le32 [usertype] nsmagic
   security/commoncap.c:456:33: sparse:     got unsigned int
   security/commoncap.c:457:29: sparse: sparse: restricted __le32 degrades to integer
   security/commoncap.c:458:39: sparse: sparse: invalid assignment: |=
   security/commoncap.c:458:39: sparse:    left side has type restricted __le32
   security/commoncap.c:458:39: sparse:    right side has type int
   security/commoncap.c:460:42: sparse: sparse: cast from restricted __le32
   security/commoncap.c:460:42: sparse: sparse: incorrect type in argument 1 (different base types) @@     expected unsigned int [usertype] val @@     got restricted __le32 [usertype] magic @@
   security/commoncap.c:460:42: sparse:     expected unsigned int [usertype] val
   security/commoncap.c:460:42: sparse:     got restricted __le32 [usertype] magic
   security/commoncap.c:460:42: sparse: sparse: cast from restricted __le32
   security/commoncap.c:460:42: sparse: sparse: cast from restricted __le32
   security/commoncap.c:460:42: sparse: sparse: cast from restricted __le32
   security/commoncap.c:460:42: sparse: sparse: cast from restricted __le32
   security/commoncap.c:1281:41: sparse: sparse: dubious: !x | y

vim +425 security/commoncap.c

   357	
   358	/*
   359	 * getsecurity: We are called for security.* before any attempt to read the
   360	 * xattr from the inode itself.
   361	 *
   362	 * This gives us a chance to read the on-disk value and convert it.  If we
   363	 * return -EOPNOTSUPP, then vfs_getxattr() will call the i_op handler.
   364	 *
   365	 * Note we are not called by vfs_getxattr_alloc(), but that is only called
   366	 * by the integrity subsystem, which really wants the unconverted values -
   367	 * so that's good.
   368	 */
   369	int cap_inode_getsecurity(struct inode *inode, const char *name, void **buffer,
   370				  bool alloc)
   371	{
   372		int size, ret;
   373		kuid_t kroot;
   374		__le32 nsmagic, magic;
   375		uid_t root, mappedroot;
   376		char *tmpbuf = NULL;
   377		struct vfs_cap_data *cap;
   378		struct vfs_ns_cap_data *nscap = NULL;
   379		struct dentry *dentry;
   380		struct user_namespace *fs_ns;
   381	
   382		if (strcmp(name, "capability") != 0)
   383			return -EOPNOTSUPP;
   384	
   385		dentry = d_find_any_alias(inode);
   386		if (!dentry)
   387			return -EINVAL;
   388	
   389		size = sizeof(struct vfs_ns_cap_data);
   390		ret = (int) vfs_getxattr_alloc(dentry, XATTR_NAME_CAPS,
   391					 &tmpbuf, size, GFP_NOFS);
   392		dput(dentry);
   393	
   394		if (ret < 0)
   395			return ret;
   396	
   397		fs_ns = inode->i_sb->s_user_ns;
   398		cap = (struct vfs_cap_data *) tmpbuf;
   399		if (is_v2header((size_t) ret, cap)) {
   400			root = 0;
   401		} else if (is_v3header((size_t) ret, cap)) {
   402			nscap = (struct vfs_ns_cap_data *) tmpbuf;
   403			root = le32_to_cpu(nscap->rootid);
   404		} else {
   405			size = -EINVAL;
   406			goto out_free;
   407		}
   408	
   409		kroot = make_kuid(fs_ns, root);
   410	
   411		/* If the root kuid maps to a valid uid in current ns, then return
   412		 * this as a nscap. */
   413		mappedroot = from_kuid(current_user_ns(), kroot);
   414		if (mappedroot != (uid_t)-1 && mappedroot != (uid_t)0) {
   415			size = sizeof(struct vfs_ns_cap_data);
   416			if (alloc) {
   417				if (!nscap) {
   418					/* v2 -> v3 conversion */
   419					nscap = kzalloc(size, GFP_ATOMIC);
   420					if (!nscap) {
   421						size = -ENOMEM;
   422						goto out_free;
   423					}
   424					nsmagic = VFS_CAP_REVISION_3;
 > 425					magic = le32_to_cpu(cap->magic_etc);
   426					if (magic & VFS_CAP_FLAGS_EFFECTIVE)
   427						nsmagic |= VFS_CAP_FLAGS_EFFECTIVE;
   428					memcpy(&nscap->data, &cap->data, sizeof(__le32) * 2 * VFS_CAP_U32);
 > 429					nscap->magic_etc = cpu_to_le32(nsmagic);
   430				} else {
   431					/* use allocated v3 buffer */
   432					tmpbuf = NULL;
   433				}
   434				nscap->rootid = cpu_to_le32(mappedroot);
   435				*buffer = nscap;
   436			}
   437			goto out_free;
   438		}
   439	
   440		if (!rootid_owns_currentns(kroot)) {
   441			size = -EOVERFLOW;
   442			goto out_free;
   443		}
   444	
   445		/* This comes from a parent namespace.  Return as a v2 capability */
   446		size = sizeof(struct vfs_cap_data);
   447		if (alloc) {
   448			if (nscap) {
   449				/* v3 -> v2 conversion */
   450				cap = kzalloc(size, GFP_ATOMIC);
   451				if (!cap) {
   452					size = -ENOMEM;
   453					goto out_free;
   454				}
   455				magic = VFS_CAP_REVISION_2;
   456				nsmagic = le32_to_cpu(nscap->magic_etc);
   457				if (nsmagic & VFS_CAP_FLAGS_EFFECTIVE)
   458					magic |= VFS_CAP_FLAGS_EFFECTIVE;
   459				memcpy(&cap->data, &nscap->data, sizeof(__le32) * 2 * VFS_CAP_U32);
   460				cap->magic_etc = cpu_to_le32(magic);
   461			} else {
   462				/* use unconverted v2 */
   463				tmpbuf = NULL;
   464			}
   465			*buffer = cap;
   466		}
   467	out_free:
   468		kfree(tmpbuf);
   469		return size;
   470	}
   471	

---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/kbuild-all@lists.01.org
kernel test robot Jan. 20, 2021, 9:08 p.m. UTC | #4
Hi Miklos,

I love your patch! Perhaps something to improve:

[auto build test WARNING on security/next-testing]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch]

url:    https://github.com/0day-ci/linux/commits/Miklos-Szeredi/capability-conversion-fixes/20210120-152933
base:   https://git.kernel.org/pub/scm/linux/kernel/git/jmorris/linux-security.git next-testing
config: x86_64-randconfig-s022-20210120 (attached as .config)
compiler: gcc-9 (Debian 9.3.0-15) 9.3.0
reproduce:
        # apt-get install sparse
        # sparse version: v0.6.3-208-g46a52ca4-dirty
        # https://github.com/0day-ci/linux/commit/bcf70adf8bcc3e52cb1b262ae2e1d9154da75097
        git remote add linux-review https://github.com/0day-ci/linux
        git fetch --no-tags linux-review Miklos-Szeredi/capability-conversion-fixes/20210120-152933
        git checkout bcf70adf8bcc3e52cb1b262ae2e1d9154da75097
        # save the attached .config to linux build tree
        make W=1 C=1 CF='-fdiagnostic-prefix -D__CHECK_ENDIAN__' ARCH=x86_64 

If you fix the issue, kindly add following tag as appropriate
Reported-by: kernel test robot <lkp@intel.com>


"sparse warnings: (new ones prefixed by >>)"
>> security/commoncap.c:424:41: sparse: sparse: incorrect type in assignment (different base types) @@     expected restricted __le32 [usertype] nsmagic @@     got int @@
   security/commoncap.c:424:41: sparse:     expected restricted __le32 [usertype] nsmagic
   security/commoncap.c:424:41: sparse:     got int
>> security/commoncap.c:425:39: sparse: sparse: incorrect type in assignment (different base types) @@     expected restricted __le32 [usertype] magic @@     got unsigned int [usertype] @@
   security/commoncap.c:425:39: sparse:     expected restricted __le32 [usertype] magic
   security/commoncap.c:425:39: sparse:     got unsigned int [usertype]
   security/commoncap.c:426:37: sparse: sparse: restricted __le32 degrades to integer
   security/commoncap.c:427:49: sparse: sparse: invalid assignment: |=
   security/commoncap.c:427:49: sparse:    left side has type restricted __le32
   security/commoncap.c:427:49: sparse:    right side has type int
   security/commoncap.c:429:52: sparse: sparse: cast from restricted __le32
   security/commoncap.c:455:31: sparse: sparse: incorrect type in assignment (different base types) @@     expected restricted __le32 [usertype] magic @@     got int @@
   security/commoncap.c:455:31: sparse:     expected restricted __le32 [usertype] magic
   security/commoncap.c:455:31: sparse:     got int
   security/commoncap.c:456:33: sparse: sparse: incorrect type in assignment (different base types) @@     expected restricted __le32 [usertype] nsmagic @@     got unsigned int [usertype] @@
   security/commoncap.c:456:33: sparse:     expected restricted __le32 [usertype] nsmagic
   security/commoncap.c:456:33: sparse:     got unsigned int [usertype]
   security/commoncap.c:457:29: sparse: sparse: restricted __le32 degrades to integer
   security/commoncap.c:458:39: sparse: sparse: invalid assignment: |=
   security/commoncap.c:458:39: sparse:    left side has type restricted __le32
   security/commoncap.c:458:39: sparse:    right side has type int
   security/commoncap.c:460:42: sparse: sparse: cast from restricted __le32
   security/commoncap.c:1281:41: sparse: sparse: dubious: !x | y

vim +424 security/commoncap.c

   357	
   358	/*
   359	 * getsecurity: We are called for security.* before any attempt to read the
   360	 * xattr from the inode itself.
   361	 *
   362	 * This gives us a chance to read the on-disk value and convert it.  If we
   363	 * return -EOPNOTSUPP, then vfs_getxattr() will call the i_op handler.
   364	 *
   365	 * Note we are not called by vfs_getxattr_alloc(), but that is only called
   366	 * by the integrity subsystem, which really wants the unconverted values -
   367	 * so that's good.
   368	 */
   369	int cap_inode_getsecurity(struct inode *inode, const char *name, void **buffer,
   370				  bool alloc)
   371	{
   372		int size, ret;
   373		kuid_t kroot;
   374		__le32 nsmagic, magic;
   375		uid_t root, mappedroot;
   376		char *tmpbuf = NULL;
   377		struct vfs_cap_data *cap;
   378		struct vfs_ns_cap_data *nscap = NULL;
   379		struct dentry *dentry;
   380		struct user_namespace *fs_ns;
   381	
   382		if (strcmp(name, "capability") != 0)
   383			return -EOPNOTSUPP;
   384	
   385		dentry = d_find_any_alias(inode);
   386		if (!dentry)
   387			return -EINVAL;
   388	
   389		size = sizeof(struct vfs_ns_cap_data);
   390		ret = (int) vfs_getxattr_alloc(dentry, XATTR_NAME_CAPS,
   391					 &tmpbuf, size, GFP_NOFS);
   392		dput(dentry);
   393	
   394		if (ret < 0)
   395			return ret;
   396	
   397		fs_ns = inode->i_sb->s_user_ns;
   398		cap = (struct vfs_cap_data *) tmpbuf;
   399		if (is_v2header((size_t) ret, cap)) {
   400			root = 0;
   401		} else if (is_v3header((size_t) ret, cap)) {
   402			nscap = (struct vfs_ns_cap_data *) tmpbuf;
   403			root = le32_to_cpu(nscap->rootid);
   404		} else {
   405			size = -EINVAL;
   406			goto out_free;
   407		}
   408	
   409		kroot = make_kuid(fs_ns, root);
   410	
   411		/* If the root kuid maps to a valid uid in current ns, then return
   412		 * this as a nscap. */
   413		mappedroot = from_kuid(current_user_ns(), kroot);
   414		if (mappedroot != (uid_t)-1 && mappedroot != (uid_t)0) {
   415			size = sizeof(struct vfs_ns_cap_data);
   416			if (alloc) {
   417				if (!nscap) {
   418					/* v2 -> v3 conversion */
   419					nscap = kzalloc(size, GFP_ATOMIC);
   420					if (!nscap) {
   421						size = -ENOMEM;
   422						goto out_free;
   423					}
 > 424					nsmagic = VFS_CAP_REVISION_3;
 > 425					magic = le32_to_cpu(cap->magic_etc);
   426					if (magic & VFS_CAP_FLAGS_EFFECTIVE)
   427						nsmagic |= VFS_CAP_FLAGS_EFFECTIVE;
   428					memcpy(&nscap->data, &cap->data, sizeof(__le32) * 2 * VFS_CAP_U32);
   429					nscap->magic_etc = cpu_to_le32(nsmagic);
   430				} else {
   431					/* use allocated v3 buffer */
   432					tmpbuf = NULL;
   433				}
   434				nscap->rootid = cpu_to_le32(mappedroot);
   435				*buffer = nscap;
   436			}
   437			goto out_free;
   438		}
   439	
   440		if (!rootid_owns_currentns(kroot)) {
   441			size = -EOVERFLOW;
   442			goto out_free;
   443		}
   444	
   445		/* This comes from a parent namespace.  Return as a v2 capability */
   446		size = sizeof(struct vfs_cap_data);
   447		if (alloc) {
   448			if (nscap) {
   449				/* v3 -> v2 conversion */
   450				cap = kzalloc(size, GFP_ATOMIC);
   451				if (!cap) {
   452					size = -ENOMEM;
   453					goto out_free;
   454				}
   455				magic = VFS_CAP_REVISION_2;
   456				nsmagic = le32_to_cpu(nscap->magic_etc);
   457				if (nsmagic & VFS_CAP_FLAGS_EFFECTIVE)
   458					magic |= VFS_CAP_FLAGS_EFFECTIVE;
   459				memcpy(&cap->data, &nscap->data, sizeof(__le32) * 2 * VFS_CAP_U32);
   460				cap->magic_etc = cpu_to_le32(magic);
   461			} else {
   462				/* use unconverted v2 */
   463				tmpbuf = NULL;
   464			}
   465			*buffer = cap;
   466		}
   467	out_free:
   468		kfree(tmpbuf);
   469		return size;
   470	}
   471	

---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/kbuild-all@lists.01.org
Serge E. Hallyn Jan. 28, 2021, 4:58 p.m. UTC | #5
On Tue, Jan 19, 2021 at 07:34:49PM -0600, Eric W. Biederman wrote:
> Miklos Szeredi <mszeredi@redhat.com> writes:
> 
> > If a capability is stored on disk in v2 format cap_inode_getsecurity() will
> > currently return in v2 format unconditionally.
> >
> > This is wrong: v2 cap should be equivalent to a v3 cap with zero rootid,
> > and so the same conversions performed on it.
> >
> > If the rootid cannot be mapped v3 is returned unconverted.  Fix this so
> > that both v2 and v3 return -EOVERFLOW if the rootid (or the owner of the fs
> > user namespace in case of v2) cannot be mapped in the current user
> > namespace.
> 
> This looks like a good cleanup.

Sorry, I'm not following.  Why is this a good cleanup?  Why should
the xattr be shown as faked v3 in this case?

A separate question below.

> I do wonder how well this works with stacking.  In particular
> ovl_xattr_set appears to call vfs_getxattr without overriding the creds.
> What the purpose of that is I haven't quite figured out.  It looks like
> it is just a probe to see if an xattr is present so maybe it is ok.
> 
> Acked-by: "Eric W. Biederman" <ebiederm@xmission.com>
> 
> >
> > Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>
> > ---
> >  security/commoncap.c | 67 ++++++++++++++++++++++++++++----------------
> >  1 file changed, 43 insertions(+), 24 deletions(-)
> >
> > diff --git a/security/commoncap.c b/security/commoncap.c
> > index bacc1111d871..c9d99f8f4c82 100644
> > --- a/security/commoncap.c
> > +++ b/security/commoncap.c
> > @@ -371,10 +371,11 @@ int cap_inode_getsecurity(struct inode *inode, const char *name, void **buffer,
> >  {
> >  	int size, ret;
> >  	kuid_t kroot;
> > +	__le32 nsmagic, magic;
> >  	uid_t root, mappedroot;
> >  	char *tmpbuf = NULL;
> >  	struct vfs_cap_data *cap;
> > -	struct vfs_ns_cap_data *nscap;
> > +	struct vfs_ns_cap_data *nscap = NULL;
> >  	struct dentry *dentry;
> >  	struct user_namespace *fs_ns;
> >  
> > @@ -396,46 +397,61 @@ int cap_inode_getsecurity(struct inode *inode, const char *name, void **buffer,
> >  	fs_ns = inode->i_sb->s_user_ns;
> >  	cap = (struct vfs_cap_data *) tmpbuf;
> >  	if (is_v2header((size_t) ret, cap)) {
> > -		/* If this is sizeof(vfs_cap_data) then we're ok with the
> > -		 * on-disk value, so return that.  */
> > -		if (alloc)
> > -			*buffer = tmpbuf;
> > -		else
> > -			kfree(tmpbuf);
> > -		return ret;
> > -	} else if (!is_v3header((size_t) ret, cap)) {
> > -		kfree(tmpbuf);
> > -		return -EINVAL;
> > +		root = 0;
> > +	} else if (is_v3header((size_t) ret, cap)) {
> > +		nscap = (struct vfs_ns_cap_data *) tmpbuf;
> > +		root = le32_to_cpu(nscap->rootid);
> > +	} else {
> > +		size = -EINVAL;
> > +		goto out_free;
> >  	}
> >  
> > -	nscap = (struct vfs_ns_cap_data *) tmpbuf;
> > -	root = le32_to_cpu(nscap->rootid);
> >  	kroot = make_kuid(fs_ns, root);
> >  
> >  	/* If the root kuid maps to a valid uid in current ns, then return
> >  	 * this as a nscap. */
> >  	mappedroot = from_kuid(current_user_ns(), kroot);
> >  	if (mappedroot != (uid_t)-1 && mappedroot != (uid_t)0) {
> > +		size = sizeof(struct vfs_ns_cap_data);
> >  		if (alloc) {
> > -			*buffer = tmpbuf;
> > +			if (!nscap) {
> > +				/* v2 -> v3 conversion */
> > +				nscap = kzalloc(size, GFP_ATOMIC);
> > +				if (!nscap) {
> > +					size = -ENOMEM;
> > +					goto out_free;
> > +				}
> > +				nsmagic = VFS_CAP_REVISION_3;
> > +				magic = le32_to_cpu(cap->magic_etc);
> > +				if (magic & VFS_CAP_FLAGS_EFFECTIVE)
> > +					nsmagic |= VFS_CAP_FLAGS_EFFECTIVE;
> > +				memcpy(&nscap->data, &cap->data, sizeof(__le32) * 2 * VFS_CAP_U32);
> > +				nscap->magic_etc = cpu_to_le32(nsmagic);
> > +			} else {
> > +				/* use allocated v3 buffer */
> > +				tmpbuf = NULL;
> > +			}
> >  			nscap->rootid = cpu_to_le32(mappedroot);
> > -		} else
> > -			kfree(tmpbuf);
> > -		return size;
> > +			*buffer = nscap;
> > +		}
> > +		goto out_free;
> >  	}
> >  
> >  	if (!rootid_owns_currentns(kroot)) {
> > -		kfree(tmpbuf);
> > -		return -EOPNOTSUPP;
> > +		size = -EOVERFLOW;

Why this change?  Christian (cc:d) noticed that this is a user visible change.
Without this change, if you are in a userns which has different rootid, the
EOVERFLOW tells vfs_getxattr to vall back to __vfs_getxattr() and so you can
see the v3 capability with its rootid.

With this change, you instead just get EOVERFLOW.

> > +		goto out_free;
> >  	}
> >  
> >  	/* This comes from a parent namespace.  Return as a v2 capability */
> >  	size = sizeof(struct vfs_cap_data);
> >  	if (alloc) {
> > -		*buffer = kmalloc(size, GFP_ATOMIC);
> > -		if (*buffer) {
> > -			struct vfs_cap_data *cap = *buffer;
> > -			__le32 nsmagic, magic;
> > +		if (nscap) {
> > +			/* v3 -> v2 conversion */
> > +			cap = kzalloc(size, GFP_ATOMIC);
> > +			if (!cap) {
> > +				size = -ENOMEM;
> > +				goto out_free;
> > +			}
> >  			magic = VFS_CAP_REVISION_2;
> >  			nsmagic = le32_to_cpu(nscap->magic_etc);
> >  			if (nsmagic & VFS_CAP_FLAGS_EFFECTIVE)
> > @@ -443,9 +459,12 @@ int cap_inode_getsecurity(struct inode *inode, const char *name, void **buffer,
> >  			memcpy(&cap->data, &nscap->data, sizeof(__le32) * 2 * VFS_CAP_U32);
> >  			cap->magic_etc = cpu_to_le32(magic);
> >  		} else {
> > -			size = -ENOMEM;
> > +			/* use unconverted v2 */
> > +			tmpbuf = NULL;
> >  		}
> > +		*buffer = cap;
> >  	}
> > +out_free:
> >  	kfree(tmpbuf);
> >  	return size;
> >  }
Eric W. Biederman Jan. 28, 2021, 8:19 p.m. UTC | #6
"Serge E. Hallyn" <serge@hallyn.com> writes:

> On Tue, Jan 19, 2021 at 07:34:49PM -0600, Eric W. Biederman wrote:
>> Miklos Szeredi <mszeredi@redhat.com> writes:
>> 
>> > If a capability is stored on disk in v2 format cap_inode_getsecurity() will
>> > currently return in v2 format unconditionally.
>> >
>> > This is wrong: v2 cap should be equivalent to a v3 cap with zero rootid,
>> > and so the same conversions performed on it.
>> >
>> > If the rootid cannot be mapped v3 is returned unconverted.  Fix this so
>> > that both v2 and v3 return -EOVERFLOW if the rootid (or the owner of the fs
>> > user namespace in case of v2) cannot be mapped in the current user
>> > namespace.
>> 
>> This looks like a good cleanup.
>
> Sorry, I'm not following.  Why is this a good cleanup?  Why should
> the xattr be shown as faked v3 in this case?

If the reader is in &init_user_ns.  If the filesystem was mounted in a
user namespace.   Then the reader looses the information that the
capability xattr only applies to a subset of user namespaces.

A trivial place where this would be important is if userspace was to
copy the file and the associated  capability xattr to another
filesystem, that is mounted differently.


<aside>
From our previous discussions I would also argue it would be good
if there was a bypass that skipped all conversions if the reader
and the filesystem are in the same user namespace.
</aside>


> A separate question below.
>
>> I do wonder how well this works with stacking.  In particular
>> ovl_xattr_set appears to call vfs_getxattr without overriding the creds.
>> What the purpose of that is I haven't quite figured out.  It looks like
>> it is just a probe to see if an xattr is present so maybe it is ok.
>> 
>> Acked-by: "Eric W. Biederman" <ebiederm@xmission.com>
>> 
>> >
>> > Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>
>> > ---
>> >  security/commoncap.c | 67 ++++++++++++++++++++++++++++----------------
>> >  1 file changed, 43 insertions(+), 24 deletions(-)
>> >
>> > diff --git a/security/commoncap.c b/security/commoncap.c
>> > index bacc1111d871..c9d99f8f4c82 100644
>> > --- a/security/commoncap.c
>> > +++ b/security/commoncap.c
>> > @@ -371,10 +371,11 @@ int cap_inode_getsecurity(struct inode *inode, const char *name, void **buffer,
>> >  {
>> >  	int size, ret;
>> >  	kuid_t kroot;
>> > +	__le32 nsmagic, magic;
>> >  	uid_t root, mappedroot;
>> >  	char *tmpbuf = NULL;
>> >  	struct vfs_cap_data *cap;
>> > -	struct vfs_ns_cap_data *nscap;
>> > +	struct vfs_ns_cap_data *nscap = NULL;
>> >  	struct dentry *dentry;
>> >  	struct user_namespace *fs_ns;
>> >  
>> > @@ -396,46 +397,61 @@ int cap_inode_getsecurity(struct inode *inode, const char *name, void **buffer,
>> >  	fs_ns = inode->i_sb->s_user_ns;
>> >  	cap = (struct vfs_cap_data *) tmpbuf;
>> >  	if (is_v2header((size_t) ret, cap)) {
>> > -		/* If this is sizeof(vfs_cap_data) then we're ok with the
>> > -		 * on-disk value, so return that.  */
>> > -		if (alloc)
>> > -			*buffer = tmpbuf;
>> > -		else
>> > -			kfree(tmpbuf);
>> > -		return ret;
>> > -	} else if (!is_v3header((size_t) ret, cap)) {
>> > -		kfree(tmpbuf);
>> > -		return -EINVAL;
>> > +		root = 0;
>> > +	} else if (is_v3header((size_t) ret, cap)) {
>> > +		nscap = (struct vfs_ns_cap_data *) tmpbuf;
>> > +		root = le32_to_cpu(nscap->rootid);
>> > +	} else {
>> > +		size = -EINVAL;
>> > +		goto out_free;
>> >  	}
>> >  
>> > -	nscap = (struct vfs_ns_cap_data *) tmpbuf;
>> > -	root = le32_to_cpu(nscap->rootid);
>> >  	kroot = make_kuid(fs_ns, root);
>> >  
>> >  	/* If the root kuid maps to a valid uid in current ns, then return
>> >  	 * this as a nscap. */
>> >  	mappedroot = from_kuid(current_user_ns(), kroot);
>> >  	if (mappedroot != (uid_t)-1 && mappedroot != (uid_t)0) {
>> > +		size = sizeof(struct vfs_ns_cap_data);
>> >  		if (alloc) {
>> > -			*buffer = tmpbuf;
>> > +			if (!nscap) {
>> > +				/* v2 -> v3 conversion */
>> > +				nscap = kzalloc(size, GFP_ATOMIC);
>> > +				if (!nscap) {
>> > +					size = -ENOMEM;
>> > +					goto out_free;
>> > +				}
>> > +				nsmagic = VFS_CAP_REVISION_3;
>> > +				magic = le32_to_cpu(cap->magic_etc);
>> > +				if (magic & VFS_CAP_FLAGS_EFFECTIVE)
>> > +					nsmagic |= VFS_CAP_FLAGS_EFFECTIVE;
>> > +				memcpy(&nscap->data, &cap->data, sizeof(__le32) * 2 * VFS_CAP_U32);
>> > +				nscap->magic_etc = cpu_to_le32(nsmagic);
>> > +			} else {
>> > +				/* use allocated v3 buffer */
>> > +				tmpbuf = NULL;
>> > +			}
>> >  			nscap->rootid = cpu_to_le32(mappedroot);
>> > -		} else
>> > -			kfree(tmpbuf);
>> > -		return size;
>> > +			*buffer = nscap;
>> > +		}
>> > +		goto out_free;
>> >  	}
>> >  
>> >  	if (!rootid_owns_currentns(kroot)) {
>> > -		kfree(tmpbuf);
>> > -		return -EOPNOTSUPP;
>> > +		size = -EOVERFLOW;
>
> Why this change?  Christian (cc:d) noticed that this is a user visible change.
> Without this change, if you are in a userns which has different rootid, the
> EOVERFLOW tells vfs_getxattr to vall back to __vfs_getxattr() and so you can
> see the v3 capability with its rootid.
>
> With this change, you instead just get EOVERFLOW.

Returning EOVERFLOW is the desired behavior when the rootid can not be
represented by the calling userspace.

Today when you execute such a file from such a namespace the file will
run without any file capabilities because get_vfs_caps_from_disk
returns -ENODATA.

However today if you copy the file will all of it's xattrs onto another
filesystem the new file will have a v3 cap that will grant capabilities
in some contexts.  That mismatch is potentially a security problem.

Which means the only sane thing to do is to fail so userspace does not
think it can safely copy or comprehend all of the xattrs of the file.

>> > +		goto out_free;
>> >  	}
>> >  
>> >  	/* This comes from a parent namespace.  Return as a v2 capability */
>> >  	size = sizeof(struct vfs_cap_data);
>> >  	if (alloc) {
>> > -		*buffer = kmalloc(size, GFP_ATOMIC);
>> > -		if (*buffer) {
>> > -			struct vfs_cap_data *cap = *buffer;
>> > -			__le32 nsmagic, magic;
>> > +		if (nscap) {
>> > +			/* v3 -> v2 conversion */
>> > +			cap = kzalloc(size, GFP_ATOMIC);
>> > +			if (!cap) {
>> > +				size = -ENOMEM;
>> > +				goto out_free;
>> > +			}
>> >  			magic = VFS_CAP_REVISION_2;
>> >  			nsmagic = le32_to_cpu(nscap->magic_etc);
>> >  			if (nsmagic & VFS_CAP_FLAGS_EFFECTIVE)
>> > @@ -443,9 +459,12 @@ int cap_inode_getsecurity(struct inode *inode, const char *name, void **buffer,
>> >  			memcpy(&cap->data, &nscap->data, sizeof(__le32) * 2 * VFS_CAP_U32);
>> >  			cap->magic_etc = cpu_to_le32(magic);
>> >  		} else {
>> > -			size = -ENOMEM;
>> > +			/* use unconverted v2 */
>> > +			tmpbuf = NULL;
>> >  		}
>> > +		*buffer = cap;
>> >  	}
>> > +out_free:
>> >  	kfree(tmpbuf);
>> >  	return size;
>> >  }

Eric
Miklos Szeredi Jan. 28, 2021, 8:38 p.m. UTC | #7
On Thu, Jan 28, 2021 at 9:24 PM Eric W. Biederman <ebiederm@xmission.com> wrote:

> <aside>
> From our previous discussions I would also argue it would be good
> if there was a bypass that skipped all conversions if the reader
> and the filesystem are in the same user namespace.
> </aside>

That's however just an optimization (AFAICS) that only makes sense if
it helps a read world workload.   I'm not convinced that that's the
case.

Thanks,
Miklos
Eric W. Biederman Jan. 28, 2021, 8:49 p.m. UTC | #8
Miklos Szeredi <miklos@szeredi.hu> writes:

> On Thu, Jan 28, 2021 at 9:24 PM Eric W. Biederman <ebiederm@xmission.com> wrote:
>
>> <aside>
>> From our previous discussions I would also argue it would be good
>> if there was a bypass that skipped all conversions if the reader
>> and the filesystem are in the same user namespace.
>> </aside>
>
> That's however just an optimization (AFAICS) that only makes sense if
> it helps a read world workload.   I'm not convinced that that's the
> case.

It is definitely a different issue.

From previous conversations with Serge, there is a concern with a
sysadmin wanting to see what is actually on disk.  In case there are
bugs that care about the different layout.  Just passing everything
through when no translation is necessary will allow that kind of
diagnosis.

As your patch demonstrates we already have had bugs in this area
so being able to get at the raw data may help people if they get into a
situation where bugs matter.

Eric
Eric W. Biederman Jan. 29, 2021, 10:55 p.m. UTC | #9
"Serge E. Hallyn" <serge@hallyn.com> writes:

> On Thu, Jan 28, 2021 at 02:19:13PM -0600, Eric W. Biederman wrote:
>> "Serge E. Hallyn" <serge@hallyn.com> writes:
>> 
>> > On Tue, Jan 19, 2021 at 07:34:49PM -0600, Eric W. Biederman wrote:
>> >> Miklos Szeredi <mszeredi@redhat.com> writes:
>> >> 
>> >> > If a capability is stored on disk in v2 format cap_inode_getsecurity() will
>> >> > currently return in v2 format unconditionally.
>> >> >
>> >> > This is wrong: v2 cap should be equivalent to a v3 cap with zero rootid,
>> >> > and so the same conversions performed on it.
>> >> >
>> >> > If the rootid cannot be mapped v3 is returned unconverted.  Fix this so
>> >> > that both v2 and v3 return -EOVERFLOW if the rootid (or the owner of the fs
>> >> > user namespace in case of v2) cannot be mapped in the current user
>> >> > namespace.
>> >> 
>> >> This looks like a good cleanup.
>> >
>> > Sorry, I'm not following.  Why is this a good cleanup?  Why should
>> > the xattr be shown as faked v3 in this case?
>> 
>> If the reader is in &init_user_ns.  If the filesystem was mounted in a
>> user namespace.   Then the reader looses the information that the
>
> Can you be more precise about "filesystem was mounted in a user namespace"?
> Is this a FUSE thing, the fs is marked as being mounted in a non-init userns?
> If that's a possible case, then yes that must be represented as v3.  Using
> is_v2header() may be the simpler way to check for that, but the more accurate
> check would be "is it v2 header and mounted by init_user_ns".

I think the filesystems current relevant are fuse,overlayfs,ramfs,tmpfs.

> Basically yes, in as many cases as possible we want to just give a v2
> cap because more userspace knows what to do with that, but a non-init-userns
> mounted fs which provides a v2 fscap should have it represented as v3 cap
> with rootid being the kuid that owns the userns.

That is the case we that is being fixed in the patch.

> Or am I still thinking wrongly?  Wouldn't be entirely surprised :)

No you got it.

Eric
Eric W. Biederman Jan. 29, 2021, 11:11 p.m. UTC | #10
"Serge E. Hallyn" <serge@hallyn.com> writes:

> On Thu, Jan 28, 2021 at 08:44:26PM +0100, Miklos Szeredi wrote:
>> On Thu, Jan 28, 2021 at 6:09 PM Serge E. Hallyn <serge@hallyn.com> wrote:
>> >
>> > On Tue, Jan 19, 2021 at 07:34:49PM -0600, Eric W. Biederman wrote:
>> > > Miklos Szeredi <mszeredi@redhat.com> writes:
>> > >
>> > > >     if (!rootid_owns_currentns(kroot)) {
>> > > > -           kfree(tmpbuf);
>> > > > -           return -EOPNOTSUPP;
>> > > > +           size = -EOVERFLOW;
>> >
>> > Why this change?  Christian (cc:d) noticed that this is a user visible change.
>> > Without this change, if you are in a userns which has different rootid, the
>> > EOVERFLOW tells vfs_getxattr to vall back to __vfs_getxattr() and so you can
>> > see the v3 capability with its rootid.
>> >
>> > With this change, you instead just get EOVERFLOW.
>> 
>> Why would the user want to see nonsense (in its own userns) rootid and
>> what would it do with it?
>
> They would know that the data is there.

But an error of -EOVERFLOW still indicates data is there.
You just don't get the data because it can not be represented.

>> Please give an example where an untranslatable rootid would make any
>> sense at all to the user.
>
> I may have accidentally, from init_user_ns, as uid 1000, set an
> fscap with rootid 100001 instead of 100000, and wonder why the
> cap is not working in the container where 100000 is root.

Getting -EOVERFLOW when attempting to read the cap from inside
the user namespace will immediately tell you what is wrong. The rootid
does not map.

That is how all the non-mapping situations are handled.  Either
-EOVERFLOW or returning INVALID_UID/the unmapped user id aka nobody.

The existing code is wrong because it returns a completely untranslated
uid, which is completely non-sense.

An argument could be made for returning a rootid of 0xffffffff aka
INVALID_UID in a v3 cap xattr when the rootid can not be mapped.  I
think that is what we do with posix_acls that contain ids that don't
map.  My sense is returning -EOVERFLOW inside the container and
returning the v3 cap xattr outside the container will most quickly get
the problem diagnosed, and will be the most likely to not cause
problems.

If there is a good case for returning a v3 cap with rootid of 0xffffffff
instead of -EOVERFLOW we can do that.  Right now I don't see anything
that would be compelling in either direction.

Eric
Serge E. Hallyn Jan. 30, 2021, 2:04 a.m. UTC | #11
On Fri, Jan 29, 2021 at 05:11:53PM -0600, Eric W. Biederman wrote:
> "Serge E. Hallyn" <serge@hallyn.com> writes:
> 
> > On Thu, Jan 28, 2021 at 08:44:26PM +0100, Miklos Szeredi wrote:
> >> On Thu, Jan 28, 2021 at 6:09 PM Serge E. Hallyn <serge@hallyn.com> wrote:
> >> >
> >> > On Tue, Jan 19, 2021 at 07:34:49PM -0600, Eric W. Biederman wrote:
> >> > > Miklos Szeredi <mszeredi@redhat.com> writes:
> >> > >
> >> > > >     if (!rootid_owns_currentns(kroot)) {
> >> > > > -           kfree(tmpbuf);
> >> > > > -           return -EOPNOTSUPP;
> >> > > > +           size = -EOVERFLOW;
> >> >
> >> > Why this change?  Christian (cc:d) noticed that this is a user visible change.
> >> > Without this change, if you are in a userns which has different rootid, the
> >> > EOVERFLOW tells vfs_getxattr to vall back to __vfs_getxattr() and so you can
> >> > see the v3 capability with its rootid.
> >> >
> >> > With this change, you instead just get EOVERFLOW.
> >> 
> >> Why would the user want to see nonsense (in its own userns) rootid and
> >> what would it do with it?
> >
> > They would know that the data is there.
> 
> But an error of -EOVERFLOW still indicates data is there.
> You just don't get the data because it can not be represented.

Ok - and this happens *after* the check for whether the rootid to maps
into the current ns.

That sounds reasonable, thanks.

> >> Please give an example where an untranslatable rootid would make any
> >> sense at all to the user.
> >
> > I may have accidentally, from init_user_ns, as uid 1000, set an
> > fscap with rootid 100001 instead of 100000, and wonder why the
> > cap is not working in the container where 100000 is root.
> 
> Getting -EOVERFLOW when attempting to read the cap from inside
> the user namespace will immediately tell you what is wrong. The rootid
> does not map.
> 
> That is how all the non-mapping situations are handled.  Either
> -EOVERFLOW or returning INVALID_UID/the unmapped user id aka nobody.
> 
> The existing code is wrong because it returns a completely untranslated
> uid, which is completely non-sense.
> 
> An argument could be made for returning a rootid of 0xffffffff aka
> INVALID_UID in a v3 cap xattr when the rootid can not be mapped.  I
> think that is what we do with posix_acls that contain ids that don't
> map.  My sense is returning -EOVERFLOW inside the container and
> returning the v3 cap xattr outside the container will most quickly get
> the problem diagnosed, and will be the most likely to not cause
> problems.
> 
> If there is a good case for returning a v3 cap with rootid of 0xffffffff
> instead of -EOVERFLOW we can do that.  Right now I don't see anything
> that would be compelling in either direction.
> 
> Eric
> 
> 
>
Serge E. Hallyn Jan. 30, 2021, 2:06 a.m. UTC | #12
On Fri, Jan 29, 2021 at 04:55:29PM -0600, Eric W. Biederman wrote:
> "Serge E. Hallyn" <serge@hallyn.com> writes:
> 
> > On Thu, Jan 28, 2021 at 02:19:13PM -0600, Eric W. Biederman wrote:
> >> "Serge E. Hallyn" <serge@hallyn.com> writes:
> >> 
> >> > On Tue, Jan 19, 2021 at 07:34:49PM -0600, Eric W. Biederman wrote:
> >> >> Miklos Szeredi <mszeredi@redhat.com> writes:
> >> >> 
> >> >> > If a capability is stored on disk in v2 format cap_inode_getsecurity() will
> >> >> > currently return in v2 format unconditionally.
> >> >> >
> >> >> > This is wrong: v2 cap should be equivalent to a v3 cap with zero rootid,
> >> >> > and so the same conversions performed on it.
> >> >> >
> >> >> > If the rootid cannot be mapped v3 is returned unconverted.  Fix this so
> >> >> > that both v2 and v3 return -EOVERFLOW if the rootid (or the owner of the fs
> >> >> > user namespace in case of v2) cannot be mapped in the current user
> >> >> > namespace.
> >> >> 
> >> >> This looks like a good cleanup.
> >> >
> >> > Sorry, I'm not following.  Why is this a good cleanup?  Why should
> >> > the xattr be shown as faked v3 in this case?
> >> 
> >> If the reader is in &init_user_ns.  If the filesystem was mounted in a
> >> user namespace.   Then the reader looses the information that the
> >
> > Can you be more precise about "filesystem was mounted in a user namespace"?
> > Is this a FUSE thing, the fs is marked as being mounted in a non-init userns?
> > If that's a possible case, then yes that must be represented as v3.  Using
> > is_v2header() may be the simpler way to check for that, but the more accurate
> > check would be "is it v2 header and mounted by init_user_ns".
> 
> I think the filesystems current relevant are fuse,overlayfs,ramfs,tmpfs.
> 
> > Basically yes, in as many cases as possible we want to just give a v2
> > cap because more userspace knows what to do with that, but a non-init-userns
> > mounted fs which provides a v2 fscap should have it represented as v3 cap
> > with rootid being the kuid that owns the userns.
> 
> That is the case we that is being fixed in the patch.
> 
> > Or am I still thinking wrongly?  Wouldn't be entirely surprised :)
> 
> No you got it.

So then can we make faking a v3 gated on whether
    sb->s_user_ns != &init_user_ns ?
Eric W. Biederman Jan. 31, 2021, 6:14 p.m. UTC | #13
"Serge E. Hallyn" <serge@hallyn.com> writes:

> On Fri, Jan 29, 2021 at 04:55:29PM -0600, Eric W. Biederman wrote:
>> "Serge E. Hallyn" <serge@hallyn.com> writes:
>> 
>> > On Thu, Jan 28, 2021 at 02:19:13PM -0600, Eric W. Biederman wrote:
>> >> "Serge E. Hallyn" <serge@hallyn.com> writes:
>> >> 
>> >> > On Tue, Jan 19, 2021 at 07:34:49PM -0600, Eric W. Biederman wrote:
>> >> >> Miklos Szeredi <mszeredi@redhat.com> writes:
>> >> >> 
>> >> >> > If a capability is stored on disk in v2 format cap_inode_getsecurity() will
>> >> >> > currently return in v2 format unconditionally.
>> >> >> >
>> >> >> > This is wrong: v2 cap should be equivalent to a v3 cap with zero rootid,
>> >> >> > and so the same conversions performed on it.
>> >> >> >
>> >> >> > If the rootid cannot be mapped v3 is returned unconverted.  Fix this so
>> >> >> > that both v2 and v3 return -EOVERFLOW if the rootid (or the owner of the fs
>> >> >> > user namespace in case of v2) cannot be mapped in the current user
>> >> >> > namespace.
>> >> >> 
>> >> >> This looks like a good cleanup.
>> >> >
>> >> > Sorry, I'm not following.  Why is this a good cleanup?  Why should
>> >> > the xattr be shown as faked v3 in this case?
>> >> 
>> >> If the reader is in &init_user_ns.  If the filesystem was mounted in a
>> >> user namespace.   Then the reader looses the information that the
>> >
>> > Can you be more precise about "filesystem was mounted in a user namespace"?
>> > Is this a FUSE thing, the fs is marked as being mounted in a non-init userns?
>> > If that's a possible case, then yes that must be represented as v3.  Using
>> > is_v2header() may be the simpler way to check for that, but the more accurate
>> > check would be "is it v2 header and mounted by init_user_ns".
>> 
>> I think the filesystems current relevant are fuse,overlayfs,ramfs,tmpfs.
>> 
>> > Basically yes, in as many cases as possible we want to just give a v2
>> > cap because more userspace knows what to do with that, but a non-init-userns
>> > mounted fs which provides a v2 fscap should have it represented as v3 cap
>> > with rootid being the kuid that owns the userns.
>> 
>> That is the case we that is being fixed in the patch.
>> 
>> > Or am I still thinking wrongly?  Wouldn't be entirely surprised :)
>> 
>> No you got it.
>
> So then can we make faking a v3 gated on whether
>     sb->s_user_ns != &init_user_ns ?

Sort of.

What Miklos's patch implements is always treating a v2 cap xattr on disk
as v3 internally.

>  	if (is_v2header((size_t) ret, cap)) {
>  		root = 0;
>  	} else if (is_v3header((size_t) ret, cap)) {
>  		nscap = (struct vfs_ns_cap_data *) tmpbuf;
>  		root = le32_to_cpu(nscap->rootid);
>  	} else {
>  		size = -EINVAL;
>  		goto out_free;
>  	}

Then v3 is returned if:
>  	/* If the root kuid maps to a valid uid in current ns, then return
>  	 * this as a nscap. */
>  	mappedroot = from_kuid(current_user_ns(), kroot);
>  	if (mappedroot != (uid_t)-1 && mappedroot != (uid_t)0) {

After that we verify that the fs capability can be seen by the caller
as a v2 cap xattr with:

> >  	if (!rootid_owns_currentns(kroot)) {
> > 		size = -EOVERFLOW;
> > 		goto out_free;

Anything that passes that test and does not encounter a memory
allocation error is returned as a v2.

...

Which in practice does mean that if sb->s_user_ns != &init_user_ns, 
then mappedroot != 0, and is returned as a v3.

The rest of the logic takes care of all of the other crazy silly
combinations.  Like a user namespace that identity maps uid 0,
and then mounts a filesystem.

Eric
diff mbox series

Patch

diff --git a/security/commoncap.c b/security/commoncap.c
index bacc1111d871..c9d99f8f4c82 100644
--- a/security/commoncap.c
+++ b/security/commoncap.c
@@ -371,10 +371,11 @@  int cap_inode_getsecurity(struct inode *inode, const char *name, void **buffer,
 {
 	int size, ret;
 	kuid_t kroot;
+	__le32 nsmagic, magic;
 	uid_t root, mappedroot;
 	char *tmpbuf = NULL;
 	struct vfs_cap_data *cap;
-	struct vfs_ns_cap_data *nscap;
+	struct vfs_ns_cap_data *nscap = NULL;
 	struct dentry *dentry;
 	struct user_namespace *fs_ns;
 
@@ -396,46 +397,61 @@  int cap_inode_getsecurity(struct inode *inode, const char *name, void **buffer,
 	fs_ns = inode->i_sb->s_user_ns;
 	cap = (struct vfs_cap_data *) tmpbuf;
 	if (is_v2header((size_t) ret, cap)) {
-		/* If this is sizeof(vfs_cap_data) then we're ok with the
-		 * on-disk value, so return that.  */
-		if (alloc)
-			*buffer = tmpbuf;
-		else
-			kfree(tmpbuf);
-		return ret;
-	} else if (!is_v3header((size_t) ret, cap)) {
-		kfree(tmpbuf);
-		return -EINVAL;
+		root = 0;
+	} else if (is_v3header((size_t) ret, cap)) {
+		nscap = (struct vfs_ns_cap_data *) tmpbuf;
+		root = le32_to_cpu(nscap->rootid);
+	} else {
+		size = -EINVAL;
+		goto out_free;
 	}
 
-	nscap = (struct vfs_ns_cap_data *) tmpbuf;
-	root = le32_to_cpu(nscap->rootid);
 	kroot = make_kuid(fs_ns, root);
 
 	/* If the root kuid maps to a valid uid in current ns, then return
 	 * this as a nscap. */
 	mappedroot = from_kuid(current_user_ns(), kroot);
 	if (mappedroot != (uid_t)-1 && mappedroot != (uid_t)0) {
+		size = sizeof(struct vfs_ns_cap_data);
 		if (alloc) {
-			*buffer = tmpbuf;
+			if (!nscap) {
+				/* v2 -> v3 conversion */
+				nscap = kzalloc(size, GFP_ATOMIC);
+				if (!nscap) {
+					size = -ENOMEM;
+					goto out_free;
+				}
+				nsmagic = VFS_CAP_REVISION_3;
+				magic = le32_to_cpu(cap->magic_etc);
+				if (magic & VFS_CAP_FLAGS_EFFECTIVE)
+					nsmagic |= VFS_CAP_FLAGS_EFFECTIVE;
+				memcpy(&nscap->data, &cap->data, sizeof(__le32) * 2 * VFS_CAP_U32);
+				nscap->magic_etc = cpu_to_le32(nsmagic);
+			} else {
+				/* use allocated v3 buffer */
+				tmpbuf = NULL;
+			}
 			nscap->rootid = cpu_to_le32(mappedroot);
-		} else
-			kfree(tmpbuf);
-		return size;
+			*buffer = nscap;
+		}
+		goto out_free;
 	}
 
 	if (!rootid_owns_currentns(kroot)) {
-		kfree(tmpbuf);
-		return -EOPNOTSUPP;
+		size = -EOVERFLOW;
+		goto out_free;
 	}
 
 	/* This comes from a parent namespace.  Return as a v2 capability */
 	size = sizeof(struct vfs_cap_data);
 	if (alloc) {
-		*buffer = kmalloc(size, GFP_ATOMIC);
-		if (*buffer) {
-			struct vfs_cap_data *cap = *buffer;
-			__le32 nsmagic, magic;
+		if (nscap) {
+			/* v3 -> v2 conversion */
+			cap = kzalloc(size, GFP_ATOMIC);
+			if (!cap) {
+				size = -ENOMEM;
+				goto out_free;
+			}
 			magic = VFS_CAP_REVISION_2;
 			nsmagic = le32_to_cpu(nscap->magic_etc);
 			if (nsmagic & VFS_CAP_FLAGS_EFFECTIVE)
@@ -443,9 +459,12 @@  int cap_inode_getsecurity(struct inode *inode, const char *name, void **buffer,
 			memcpy(&cap->data, &nscap->data, sizeof(__le32) * 2 * VFS_CAP_U32);
 			cap->magic_etc = cpu_to_le32(magic);
 		} else {
-			size = -ENOMEM;
+			/* use unconverted v2 */
+			tmpbuf = NULL;
 		}
+		*buffer = cap;
 	}
+out_free:
 	kfree(tmpbuf);
 	return size;
 }