[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <20241129044845.GJ3387508@ZenIV>
Date: Fri, 29 Nov 2024 04:48:45 +0000
From: Al Viro <viro@...iv.linux.org.uk>
To: "Eric W. Biederman" <ebiederm@...ssion.com>
Cc: Casey Schaufler <casey@...aufler-ca.com>,
Linus Torvalds <torvalds@...ux-foundation.org>,
Kees Cook <kees@...nel.org>, linux-kernel@...r.kernel.org,
Christophe JAILLET <christophe.jaillet@...adoo.fr>,
Nir Lichtman <nir@...htman.org>,
Tycho Andersen <tandersen@...flix.com>,
Vegard Nossum <vegard.nossum@...cle.com>,
linux-security-module@...r.kernel.org
Subject: Re: [GIT PULL] execve updates for v6.13-rc1 (take 2)
On Thu, Nov 28, 2024 at 10:23:18PM -0600, Eric W. Biederman wrote:
> > I agree that for fexecve() the only sane approach is to go by whatever
> > that opened file refers to; I'm not sold on the _usefulness_ of
> > fexecve() to start with, but if we want that thing, that's the way
> > to go.
>
> The craziness is that apparently systemd wants to implement execve in
> terms of fexecve, not execveat.
... presumably because the pathname might have changed its meaning
just as we called execve(). Which is why we want it to show up in
comm, got it.
</sarcasm>
Powered by blists - more mailing lists