[v1] security.bpf xattr name prefix

[PATCH bpf-next 1/2] fs/xattr: bpf: Introduce security.bpf xattr name prefix

Posted by Song Liu 1 month, 3 weeks ago

Introduct new xattr name prefix security.bpf, and enable reading these
xattrs from bpf kfuncs bpf_get_[file|dentry]_xattr(). Note that
"security.bpf" could be the name of the xattr or the prefix of the
name. For example, both "security.bpf" and "security.bpf.xxx" are
valid xattr name; while "security.bpfxxx" is not valid.

Signed-off-by: Song Liu <song@kernel.org>
---
 fs/bpf_fs_kfuncs.c         | 19 ++++++++++++++++++-
 include/uapi/linux/xattr.h |  4 ++++
 2 files changed, 22 insertions(+), 1 deletion(-)

diff --git a/fs/bpf_fs_kfuncs.c b/fs/bpf_fs_kfuncs.c
index 3fe9f59ef867..339c4fef8f6e 100644
--- a/fs/bpf_fs_kfuncs.c
+++ b/fs/bpf_fs_kfuncs.c
@@ -93,6 +93,23 @@ __bpf_kfunc int bpf_path_d_path(struct path *path, char *buf, size_t buf__sz)
 	return len;
 }
 
+static bool bpf_xattr_name_allowed(const char *name__str)
+{
+	/* Allow xattr names with user. prefix */
+	if (!strncmp(name__str, XATTR_USER_PREFIX, XATTR_USER_PREFIX_LEN))
+		return true;
+
+	/* Allow security.bpf. prefix or just security.bpf */
+	if (!strncmp(name__str, XATTR_NAME_BPF_LSM, XATTR_NAME_BPF_LSM_LEN) &&
+	    (name__str[XATTR_NAME_BPF_LSM_LEN] == '\0' ||
+	     name__str[XATTR_NAME_BPF_LSM_LEN] == '.')) {
+		return true;
+	}
+
+	/* Disallow anything else */
+	return false;
+}
+
 /**
  * bpf_get_dentry_xattr - get xattr of a dentry
  * @dentry: dentry to get xattr from
@@ -117,7 +134,7 @@ __bpf_kfunc int bpf_get_dentry_xattr(struct dentry *dentry, const char *name__st
 	if (WARN_ON(!inode))
 		return -EINVAL;
 
-	if (strncmp(name__str, XATTR_USER_PREFIX, XATTR_USER_PREFIX_LEN))
+	if (!bpf_xattr_name_allowed(name__str))
 		return -EPERM;
 
 	value_len = __bpf_dynptr_size(value_ptr);
diff --git a/include/uapi/linux/xattr.h b/include/uapi/linux/xattr.h
index 9463db2dfa9d..166ef2f1f1b3 100644
--- a/include/uapi/linux/xattr.h
+++ b/include/uapi/linux/xattr.h
@@ -76,6 +76,10 @@
 #define XATTR_CAPS_SUFFIX "capability"
 #define XATTR_NAME_CAPS XATTR_SECURITY_PREFIX XATTR_CAPS_SUFFIX
 
+#define XATTR_BPF_LSM_SUFFIX "bpf"
+#define XATTR_NAME_BPF_LSM (XATTR_SECURITY_PREFIX XATTR_BPF_LSM_SUFFIX)
+#define XATTR_NAME_BPF_LSM_LEN (sizeof(XATTR_NAME_BPF_LSM) - 1)
+
 #define XATTR_POSIX_ACL_ACCESS  "posix_acl_access"
 #define XATTR_NAME_POSIX_ACL_ACCESS XATTR_SYSTEM_PREFIX XATTR_POSIX_ACL_ACCESS
 #define XATTR_POSIX_ACL_DEFAULT  "posix_acl_default"
-- 
2.43.5

Re: [PATCH bpf-next 1/2] fs/xattr: bpf: Introduce security.bpf xattr name prefix

Posted by Christian Brauner 1 month, 2 weeks ago

On Wed, Oct 02, 2024 at 02:46:36PM -0700, Song Liu wrote:
> Introduct new xattr name prefix security.bpf, and enable reading these
> xattrs from bpf kfuncs bpf_get_[file|dentry]_xattr(). Note that
> "security.bpf" could be the name of the xattr or the prefix of the
> name. For example, both "security.bpf" and "security.bpf.xxx" are
> valid xattr name; while "security.bpfxxx" is not valid.
> 
> Signed-off-by: Song Liu <song@kernel.org>
> ---

Acked-by: Christian Brauner <brauner@kernel.org>

Re: [PATCH bpf-next 1/2] fs/xattr: bpf: Introduce security.bpf xattr name prefix

Posted by Jiri Olsa 1 month, 3 weeks ago

On Wed, Oct 02, 2024 at 02:46:36PM -0700, Song Liu wrote:
> Introduct new xattr name prefix security.bpf, and enable reading these
> xattrs from bpf kfuncs bpf_get_[file|dentry]_xattr(). Note that
> "security.bpf" could be the name of the xattr or the prefix of the
> name. For example, both "security.bpf" and "security.bpf.xxx" are
> valid xattr name; while "security.bpfxxx" is not valid.
> 
> Signed-off-by: Song Liu <song@kernel.org>
> ---
>  fs/bpf_fs_kfuncs.c         | 19 ++++++++++++++++++-
>  include/uapi/linux/xattr.h |  4 ++++
>  2 files changed, 22 insertions(+), 1 deletion(-)
> 
> diff --git a/fs/bpf_fs_kfuncs.c b/fs/bpf_fs_kfuncs.c
> index 3fe9f59ef867..339c4fef8f6e 100644
> --- a/fs/bpf_fs_kfuncs.c
> +++ b/fs/bpf_fs_kfuncs.c
> @@ -93,6 +93,23 @@ __bpf_kfunc int bpf_path_d_path(struct path *path, char *buf, size_t buf__sz)
>  	return len;
>  }
>  
> +static bool bpf_xattr_name_allowed(const char *name__str)
> +{
> +	/* Allow xattr names with user. prefix */
> +	if (!strncmp(name__str, XATTR_USER_PREFIX, XATTR_USER_PREFIX_LEN))
> +		return true;
> +
> +	/* Allow security.bpf. prefix or just security.bpf */
> +	if (!strncmp(name__str, XATTR_NAME_BPF_LSM, XATTR_NAME_BPF_LSM_LEN) &&
> +	    (name__str[XATTR_NAME_BPF_LSM_LEN] == '\0' ||
> +	     name__str[XATTR_NAME_BPF_LSM_LEN] == '.')) {
> +		return true;
> +	}
> +
> +	/* Disallow anything else */
> +	return false;
> +}
> +
>  /**
>   * bpf_get_dentry_xattr - get xattr of a dentry
>   * @dentry: dentry to get xattr from
> @@ -117,7 +134,7 @@ __bpf_kfunc int bpf_get_dentry_xattr(struct dentry *dentry, const char *name__st

nit, I guess the comment for bpf_get_dentry_xattr function needs to be updated

 * For security reasons, only *name__str* with prefix "user." is allowed.

jirka

>  	if (WARN_ON(!inode))
>  		return -EINVAL;
>  
> -	if (strncmp(name__str, XATTR_USER_PREFIX, XATTR_USER_PREFIX_LEN))
> +	if (!bpf_xattr_name_allowed(name__str))
>  		return -EPERM;
>  
>  	value_len = __bpf_dynptr_size(value_ptr);
> diff --git a/include/uapi/linux/xattr.h b/include/uapi/linux/xattr.h
> index 9463db2dfa9d..166ef2f1f1b3 100644
> --- a/include/uapi/linux/xattr.h
> +++ b/include/uapi/linux/xattr.h
> @@ -76,6 +76,10 @@
>  #define XATTR_CAPS_SUFFIX "capability"
>  #define XATTR_NAME_CAPS XATTR_SECURITY_PREFIX XATTR_CAPS_SUFFIX
>  
> +#define XATTR_BPF_LSM_SUFFIX "bpf"
> +#define XATTR_NAME_BPF_LSM (XATTR_SECURITY_PREFIX XATTR_BPF_LSM_SUFFIX)
> +#define XATTR_NAME_BPF_LSM_LEN (sizeof(XATTR_NAME_BPF_LSM) - 1)
> +
>  #define XATTR_POSIX_ACL_ACCESS  "posix_acl_access"
>  #define XATTR_NAME_POSIX_ACL_ACCESS XATTR_SYSTEM_PREFIX XATTR_POSIX_ACL_ACCESS
>  #define XATTR_POSIX_ACL_DEFAULT  "posix_acl_default"
> -- 
> 2.43.5
> 
>

Re: [PATCH bpf-next 1/2] fs/xattr: bpf: Introduce security.bpf xattr name prefix

Posted by Song Liu 1 month, 3 weeks ago

Hi Jiri,

Thanks for the review!

> On Oct 7, 2024, at 1:39 AM, Jiri Olsa <olsajiri@gmail.com> wrote:
> 
> On Wed, Oct 02, 2024 at 02:46:36PM -0700, Song Liu wrote:
>> Introduct new xattr name prefix security.bpf, and enable reading these
>> xattrs from bpf kfuncs bpf_get_[file|dentry]_xattr(). Note that
>> "security.bpf" could be the name of the xattr or the prefix of the
>> name. For example, both "security.bpf" and "security.bpf.xxx" are
>> valid xattr name; while "security.bpfxxx" is not valid.
>> 
>> Signed-off-by: Song Liu <song@kernel.org>
>> ---
>> fs/bpf_fs_kfuncs.c         | 19 ++++++++++++++++++-
>> include/uapi/linux/xattr.h |  4 ++++
>> 2 files changed, 22 insertions(+), 1 deletion(-)
>> 
>> diff --git a/fs/bpf_fs_kfuncs.c b/fs/bpf_fs_kfuncs.c
>> index 3fe9f59ef867..339c4fef8f6e 100644
>> --- a/fs/bpf_fs_kfuncs.c
>> +++ b/fs/bpf_fs_kfuncs.c
>> @@ -93,6 +93,23 @@ __bpf_kfunc int bpf_path_d_path(struct path *path, char *buf, size_t buf__sz)
>> return len;
>> }
>> 
>> +static bool bpf_xattr_name_allowed(const char *name__str)
>> +{
>> + /* Allow xattr names with user. prefix */
>> + if (!strncmp(name__str, XATTR_USER_PREFIX, XATTR_USER_PREFIX_LEN))
>> + return true;
>> +
>> + /* Allow security.bpf. prefix or just security.bpf */
>> + if (!strncmp(name__str, XATTR_NAME_BPF_LSM, XATTR_NAME_BPF_LSM_LEN) &&
>> +    (name__str[XATTR_NAME_BPF_LSM_LEN] == '\0' ||
>> +     name__str[XATTR_NAME_BPF_LSM_LEN] == '.')) {
>> + return true;
>> + }
>> +
>> + /* Disallow anything else */
>> + return false;
>> +}
>> +
>> /**
>>  * bpf_get_dentry_xattr - get xattr of a dentry
>>  * @dentry: dentry to get xattr from
>> @@ -117,7 +134,7 @@ __bpf_kfunc int bpf_get_dentry_xattr(struct dentry *dentry, const char *name__st
> 
> nit, I guess the comment for bpf_get_dentry_xattr function needs to be updated
> 
> * For security reasons, only *name__str* with prefix "user." is allowed.

Good catch! We can update it as:

 * For security reasons, only *name__str* with prefix "user." or
 * "security.bpf" is allowed.

Song