netdev - Re: [PATCH ethtool] uapi: if.h: fix linux/libc-compat.h include on Linux

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]

Message-ID: <CACXRmJj8hkni1NdKHvutCQw3An-uwu0MJkHFDS14d+OiwzDHZA@mail.gmail.com>
Date:   Fri, 24 Feb 2023 11:05:43 +0100
From:   Thomas Devoogdt <thomas@...oogdt.com>
To:     Michal Kubecek <mkubecek@...e.cz>
Cc:     netdev@...r.kernel.org, Thomas Devoogdt <thomas.devoogdt@...co.com>
Subject: Re: [PATCH ethtool] uapi: if.h: fix linux/libc-compat.h include on
 Linux < 3.12

Hi,

I now remember (while looking at the other patches I had to add) that
I'm also missing __kernel_sa_family_t from /uapi/linux/socket.h (for
Linux < 3.7). So it's indeed not just libc-compat.h which is causing
problems. So perhaps take that one along while at it.

Thx in advance,

Thomas Devoogdt

Op do 23 feb. 2023 om 22:17 schreef Michal Kubecek <mkubecek@...e.cz>:
>
> On Thu, Feb 23, 2023 at 09:38:41PM +0100, Thomas Devoogdt wrote:
> > Hi,
> >
> > I now see that these headers are simply synced with (and even
> > committed to) the upstream kernel. So having a kernel version check
> > there is probably not something we want to do. Better would be to
> > incorporate the "libc-compat.h" header as well to fix compilation on
> > Linux < 3.12. This is similar to the added "if.h" header itself in
> > commit https://git.kernel.org/pub/scm/network/ethtool/ethtool.git/commit/uapi/linux/if.h?id=0b09751eb84178d19e4f92543b7fb1e4f162c236,
> > which added support for Linux < 4.11.
> >
> > Let me know what you think, and if further action is needed from my
>
> Yes, adding libc-compat.h would be a cleaner solution than having
> a modified version of one header file. The easiest way should be
> creating a bogus header file (e.g. "touch uapi/linux/libc-compat.h") and
> running the ethtool-import-uapi script.
>
> Seeing that this is not the first problem of this type - and likely not the
> last either - I'm considering if we shouldn't go all the way and prevent
> mixing potentially incompatible kernel header versions by pulling every
> kernel header included in the source (and every kernel header included
> by those etc.). That's something that could be scripted easily so I'm
> going to try it and see how big the full set would be.
>
> Michal