On Mon, Apr 8, 2024 at 11:31 AM Christian Göttsche
<cgoettsche@xxxxxxxxxxxxx> wrote:
>
> From: Christian Göttsche <cgzones@xxxxxxxxxxxxxx>
>
> Perform the password hash comparison in a time constant way to avoid
> leaking the length of the identical prefix via a side-channel.
> Since only hashes are compared leaking the total length is non critical.
>
> Signed-off-by: Christian Göttsche <cgzones@xxxxxxxxxxxxxx>
Should we just require PAM for newrole and be done with it?
> ---
> policycoreutils/newrole/newrole.c | 20 +++++++++++++++++++-
> 1 file changed, 19 insertions(+), 1 deletion(-)
>
> diff --git a/policycoreutils/newrole/newrole.c b/policycoreutils/newrole/newrole.c
> index 5a1a1129..1e01d2ef 100644
> --- a/policycoreutils/newrole/newrole.c
> +++ b/policycoreutils/newrole/newrole.c
> @@ -338,6 +338,24 @@ static void memzero(void *ptr, size_t size)
> }
> }
>
> +static int streq_constant(const char *userinput, const char *secret)
> +{
> + const volatile char *x = userinput, *y = secret;
> + size_t i, u_len, s_len;
> + int ret = 0;
> +
> + u_len = strlen(userinput);
> + s_len = strlen(secret);
> +
> + if (u_len != s_len)
> + return 0;
> +
> + for (i = 0; i < u_len; i++)
> + ret |= x[i] ^ y[i];
> +
> + return ret == 0;
> +}
> +
> /* authenticate_via_shadow_passwd()
> *
> * in: uname - the calling user's user name
> @@ -383,7 +401,7 @@ static int authenticate_via_shadow_passwd(const char *uname)
> return 0;
> }
>
> - ret = !strcmp(encrypted_password_s, p_shadow_line->sp_pwdp);
> + ret = streq_constant(encrypted_password_s, p_shadow_line->sp_pwdp);
> memzero(encrypted_password_s, strlen(encrypted_password_s));
> return ret;
> }
> --
> 2.43.0
>
>
