netdev - Re: iproute2: ip address add prefer keyword confusion

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] [day] [month] [year] [list]

Date:   Fri, 28 May 2021 11:02:01 -0700
From:   Stephen Hemminger <stephen@...workplumber.org>
To:     David Ahern <dsahern@...il.com>
Cc:     Norman Rasmussen <norman@...mussen.co.za>, netdev@...r.kernel.org
Subject: Re: iproute2: ip address add prefer keyword confusion

On Thu, 27 May 2021 22:31:24 -0600
David Ahern <dsahern@...il.com> wrote:

> On 5/27/21 2:58 PM, Norman Rasmussen wrote:
> > commit 78d04c7b27cf ("ipaddress: Add support for address metric")
> > added "priority" and "preference" as aliases for the "metric" keyword,
> > but they are entirely undocumented.
> > 
> > I only noticed because I was adding addresses with a preferred
> > lifetime, but I was using "pref" as the keyword. The metric code was
> > added _above_ the lifetime code, so after the change "pref" matches
> > "preference", instead of "preferred_lft".
> > 
> > Is there an existing way to deal with conflicts between keyword
> > prefixes? Should "prefer" (or shorter) fail with a clear error
> > instead? Should the metric code have been added below the lifetime
> > code? Should it be moved or is it too late?
> >   
> 
> It is in general a known problem with iproute2's use of "matches" to
> allow shorthand commands.
> 
> The change where "pref" goes to metric vs "preferred_lft" was
> unintentional. At this point (3 years after the commit) it would be hard
> to revert the change.

Agreed, matches() is a real bug trap.