netdev - Re: [net-next] seg6: add support for optional attributes during behavior construction

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Mon, 3 Feb 2020 17:08:00 +0200
From:   Leon Romanovsky <leon@...nel.org>
To:     Andrea Mayer <andrea.mayer@...roma2.it>
Cc:     "David S. Miller" <davem@...emloft.net>,
        Alexey Kuznetsov <kuznet@....inr.ac.ru>,
        Hideaki YOSHIFUJI <yoshfuji@...ux-ipv6.org>,
        David Lebrun <dav.lebrun@...il.com>, netdev@...r.kernel.org,
        linux-kernel@...r.kernel.org, Alexei Starovoitov <ast@...nel.org>,
        Daniel Borkmann <daniel@...earbox.net>,
        Martin KaFai Lau <kafai@...com>,
        Song Liu <songliubraving@...com>, Yonghong Song <yhs@...com>,
        Andrii Nakryiko <andriin@...com>, bpf@...r.kernel.org,
        Paolo Lungaroni <paolo.lungaroni@...t.it>
Subject: Re: [net-next] seg6: add support for optional attributes during
 behavior construction

On Mon, Feb 03, 2020 at 03:36:58PM +0100, Andrea Mayer wrote:
> before this patch, each SRv6 behavior specifies a set of required
> attributes that must be provided by the userspace application when the
> behavior is created. If an attribute is not supplied, the creation
> operation fails.
> As a workaround, if an attribute is not needed by a behavior, it requires
> to be set by the userspace application to a conventional skip-value. The
> kernel side, that processes the creation request of a behavior, reads the
> supplied attribute values and checks if it has been set to the
> conventional skip-value or not. Hence, each optional attribute must have a
> conventional skip-value which is known a priori and shared between
> userspace applications and kernel.
>
> Messy code and complicated tricks may arise from this approach.
> On the other hand, this patch explicitly differentiates the required
> mandatory attributes from the optional ones. Now, each behavior can declare
> a set of required attributes and a set of optional ones. The behavior
> creation fails in case a required attribute is missing, while it goes on
> without generating any issue if an optional attribute is not supplied by
> the userspace application.
>
> To properly combine the required and optional attributes, a new callback
> function called destroy() is used for releasing resources that have been
> acquired, during the parse() operation, by a given attribute.
> However, the destroy() function is optional and if an attribute does not
> require resources that have to be later released, the callback can be
> omitted.
>
> Signed-off-by: Andrea Mayer <andrea.mayer@...roma2.it>
> ---
>  net/ipv6/seg6_local.c | 226 ++++++++++++++++++++++++++++++++++++------
>  1 file changed, 198 insertions(+), 28 deletions(-)
>
> diff --git a/net/ipv6/seg6_local.c b/net/ipv6/seg6_local.c
> index 85a5447a3e8d..480f1ab35221 100644
> --- a/net/ipv6/seg6_local.c
> +++ b/net/ipv6/seg6_local.c
> @@ -7,6 +7,13 @@
>   *  eBPF support: Mathieu Xhonneux <m.xhonneux@...il.com>
>   */
>
> +/* Changes:
> + *
> + * Andrea Mayer <andrea.mayer@...roma2.it>
> + *	add support for optional attributes during behavior construction
> + *
> + */

The lines above look strange in 2020 when all of us are using git.

Thanks