Re: [PATCH v5 2/3] bpf: add bpf_init_inode_xattr kfunc for atomic inode labeling

[Paul Moore]

On Jul  7, 2026 David Windsor <dwindsor@xxxxxxxxx> wrote:
> 
> Add bpf_init_inode_xattr() kfunc for BPF LSM programs to atomically set
> xattrs via the inode_init_security hook using security_lsmxattr_add().
> The hook now passes its xattr state as a single struct lsm_xattrs
> object, which the kfunc takes directly.
> 
> This kfunc is only callable from inode_init_security; the verifier
> rejects attempts to call it elsewhere.
> 
> A previous attempt [1] required a kmalloc string output protocol for
> the xattr name. Since commit 6bcdfd2cac55 ("security: Allow all LSMs to
> provide xattrs for inode_init_security hook") [2], the xattr name is no
> longer allocated; it is a static constant.
> 
> Link: https://kernsec.org/pipermail/linux-security-module-archive/2022-October/034878.html [1]
> Link: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=6bcdfd2cac55 [2]
> Suggested-by: Song Liu <song@xxxxxxxxxx>
> Signed-off-by: David Windsor <dwindsor@xxxxxxxxx>
> ---
>  fs/bpf_fs_kfuncs.c | 36 ++++++++++++++++++++++++++++++++++++
>  1 file changed, 36 insertions(+)
> 
> diff --git a/fs/bpf_fs_kfuncs.c b/fs/bpf_fs_kfuncs.c
> index 768aca2dc0f0..ad0025f2264a 100644
> --- a/fs/bpf_fs_kfuncs.c
> +++ b/fs/bpf_fs_kfuncs.c
> @@ -11,7 +11,9 @@
>  #include <linux/file.h>
>  #include <linux/kernfs.h>
>  #include <linux/mm.h>
> +#include <linux/security.h>
>  #include <linux/xattr.h>
> +#include <uapi/linux/lsm.h>
>  
>  __bpf_kfunc_start_defs();
>  
> @@ -374,6 +376,39 @@ __bpf_kfunc struct inode *bpf_real_inode(struct dentry *dentry)
>  	return d_real_inode(dentry);
>  }
>  
> +/**
> + * bpf_init_inode_xattr - set an xattr on a new inode from inode_init_security
> + * @xattrs: inode_init_security xattr state from the hook context
> + * @name__str: xattr name (e.g., "bpf.file_label")
> + * @value_p: dynptr containing the xattr value
> + *
> + * Only callable from lsm/inode_init_security programs.
> + *
> + * Return: 0 on success, negative error on failure.
> + */
> +__bpf_kfunc int bpf_init_inode_xattr(struct lsm_xattrs *xattrs,
> +				     const char *name__str,
> +				     const struct bpf_dynptr *value_p)
> +{
> +	struct bpf_dynptr_kern *value_ptr = (struct bpf_dynptr_kern *)value_p;
> +	const void *value;
> +	u32 value_len;
> +
> +	if (!name__str)
> +		return -EINVAL;
> +	if (strncmp(name__str, XATTR_BPF_LSM_SUFFIX,
> +		    sizeof(XATTR_BPF_LSM_SUFFIX) - 1))
> +		return -EPERM;
> +
> +	value_len = __bpf_dynptr_size(value_ptr);
> +	value = __bpf_dynptr_data(value_ptr, value_len);
> +	if (!value)
> +		return -EINVAL;
> +
> +	return security_lsmxattr_add(xattrs, LSM_ID_BPF, name__str, value,
> +				     value_len);
> +}

I'm sorry David, now that I'm seeing this function again, especially
with the LSM specific bits extracted into a LSM function, this absolutely
belongs somewhere under security/.  It's only callable from within a
BPF LSM callback and all it does outside of some BPF pointer boilerplate
is call right back into a LSM helper function.

If the BPF maintainers aren't willing to accept that, then we will all
need to find another way.

>  __bpf_kfunc_end_defs();
>  
>  BTF_KFUNCS_START(bpf_fs_kfunc_set_ids)
> @@ -385,6 +420,7 @@ BTF_ID_FLAGS(func, bpf_get_file_xattr, KF_SLEEPABLE)
>  BTF_ID_FLAGS(func, bpf_set_dentry_xattr, KF_SLEEPABLE)
>  BTF_ID_FLAGS(func, bpf_remove_dentry_xattr, KF_SLEEPABLE)
>  BTF_ID_FLAGS(func, bpf_real_inode, KF_SLEEPABLE | KF_RET_NULL)
> +BTF_ID_FLAGS(func, bpf_init_inode_xattr)
>  BTF_KFUNCS_END(bpf_fs_kfunc_set_ids)
>  
>  static int bpf_fs_kfuncs_filter(const struct bpf_prog *prog, u32 kfunc_id)
> -- 
> 2.53.0

--
paul-moore.com