Re: [PATCH] selinux: replace strlcat() with seq_buf in selinux_ima_collect_state()

[David Laight <david.laight.linux@gmail.com>]

On Mon, 22 Jun 2026 10:00:07 -0500
Ian Bridges <icb@fastmail.org> wrote:

> In preparation for removing the deprecated strlcat() API[1], replace the
> strscpy()/strlcat() chain in selinux_ima_collect_state() with a struct
> seq_buf, which tracks the write position and remaining space internally.
> 
> The seven open-coded WARN_ON(rc >= buf_len) truncation checks become a
> single seq_buf_has_overflowed() check after the string is built. The
> kzalloc() and its exact-size computation are unchanged, so the
> measurement string passed to IMA is unchanged.
> 
> Link: https://github.com/KSPP/linux/issues/370 [1]
> Signed-off-by: Ian Bridges <icb@fastmail.org>
> ---
>  security/selinux/ima.c | 35 ++++++++++++++---------------------
>  1 file changed, 14 insertions(+), 21 deletions(-)
> 
> diff --git a/security/selinux/ima.c b/security/selinux/ima.c
> index aa34da9b0aeb..3d81093d16aa 100644
> --- a/security/selinux/ima.c
> +++ b/security/selinux/ima.c
> @@ -9,6 +9,7 @@
>   */
>  #include <linux/vmalloc.h>
>  #include <linux/ima.h>
> +#include <linux/seq_buf.h>
>  #include "security.h"
>  #include "ima.h"
>  
> @@ -21,8 +22,9 @@
>  static char *selinux_ima_collect_state(void)
>  {
>  	const char *on = "=1;", *off = "=0;";
> +	struct seq_buf s;
>  	char *buf;
> -	int buf_len, len, i, rc;
> +	int buf_len, len, i;
>  
>  	buf_len = strlen("initialized=0;enforcing=0;checkreqprot=0;") + 1;
>  
> @@ -34,33 +36,24 @@ static char *selinux_ima_collect_state(void)
>  	if (!buf)
>  		return NULL;
>  
> -	rc = strscpy(buf, "initialized", buf_len);
> -	WARN_ON(rc < 0);
> +	seq_buf_init(&s, buf, buf_len);

That is silly, you need the length of the buffer not the length of a string
that is the expected length of the output.

>  
> -	rc = strlcat(buf, selinux_initialized() ? on : off, buf_len);
> -	WARN_ON(rc >= buf_len);
> +	seq_buf_puts(&s, "initialized");
> +	seq_buf_puts(&s, selinux_initialized() ? on : off);
>  
> -	rc = strlcat(buf, "enforcing", buf_len);
> -	WARN_ON(rc >= buf_len);
> +	seq_buf_puts(&s, "enforcing");
> +	seq_buf_puts(&s, enforcing_enabled() ? on : off);
>  
> -	rc = strlcat(buf, enforcing_enabled() ? on : off, buf_len);
> -	WARN_ON(rc >= buf_len);
> -
> -	rc = strlcat(buf, "checkreqprot", buf_len);
> -	WARN_ON(rc >= buf_len);
> -
> -	rc = strlcat(buf, checkreqprot_get() ? on : off, buf_len);
> -	WARN_ON(rc >= buf_len);
> +	seq_buf_puts(&s, "checkreqprot");
> +	seq_buf_puts(&s, checkreqprot_get() ? on : off);

That lot would be easier to read as a seq_printf() - with %d and
kill 'on' and 'off'.
Why does 'security' code so often look like c**p.

	David

>  
>  	for (i = 0; i < __POLICYDB_CAP_MAX; i++) {
> -		rc = strlcat(buf, selinux_policycap_names[i], buf_len);
> -		WARN_ON(rc >= buf_len);
> -
> -		rc = strlcat(buf, selinux_state.policycap[i] ? on : off,
> -			buf_len);
> -		WARN_ON(rc >= buf_len);
> +		seq_buf_puts(&s, selinux_policycap_names[i]);
> +		seq_buf_puts(&s, selinux_state.policycap[i] ? on : off);
>  	}
>  
> +	WARN_ON(seq_buf_has_overflowed(&s));
> +
>  	return buf;
>  }
>