netdev - Re: [PATCH iproute2 1/2] man: add ip-netns(8) as generation target

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: <CA+ASDXM540KLNXjRh0swrp=ATGfxWS-VUcZcqYT1Udm4QLPaVQ@mail.gmail.com>
Date:   Tue, 7 Apr 2020 11:38:59 -0700
From:   Brian Norris <briannorris@...omium.org>
To:     Stephen Hemminger <stephen@...workplumber.org>
Cc:     "<netdev@...r.kernel.org>" <netdev@...r.kernel.org>
Subject: Re: [PATCH iproute2 1/2] man: add ip-netns(8) as generation target

On Tue, Apr 7, 2020 at 11:24 AM Stephen Hemminger
<stephen@...workplumber.org> wrote:
>
> On Tue,  7 Apr 2020 10:43:05 -0700
> Brian Norris <briannorris@...omium.org> wrote:
>
> > Prepare for adding new variable substitutions. Unify the sed rules while
> > we're at it, since there's no need to write this out 4 times.
> >
> > Signed-off-by: Brian Norris <briannorris@...omium.org>
>
> Why is this needed?

For patch 1: it's only for the sake of patch 2.
If you're implying that patch 2 doesn't describe the "why?" well
enough: I'll try again:

> man: replace $(NETNS_ETC_DIR) and $(NETNS_RUN_DIR) in ip-netns(8)
>
> These can be configured to different paths. Reflect that in the
> generated documentation.

This is needed because Chrom{ium,e} OS patches iproute2 to use /run
directly instead of /var/run [1]. We also build the man pages, so we'd
like the man-pages to match.

Incidentally, we were already manually patching this out (in both
source and man-page) before this upstream patch existed:
e2f5ceccdab5 Allow to configure /var/run/netns directory
It would be nice if we could just use the Makefile variable instead.

I can resubmit if you'd like a more verbose description in the patch
submission itself.

Brian

[1] Longer answer: because the latter traverses a symlink on a
read/write partition, whereas the former is a direct-mounted tmpfs. We
can provide better guarantees for programs that avoid symlinks like
this.