linux-kernel - Re: [PATCH] syscalls: define and explain goal to not call syscalls in the kernel

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]

Message-ID: <20180330183117.GA11444@light.dominikbrodowski.net>
Date:   Fri, 30 Mar 2018 20:31:17 +0200
From:   Dominik Brodowski <linux@...inikbrodowski.net>
To:     Jonathan Corbet <corbet@....net>
Cc:     linux-kernel@...r.kernel.org, linux-doc@...r.kernel.org,
        viro@...iv.linux.org.uk, x86@...nel.org,
        torvalds@...ux-foundation.org, mingo@...nel.org,
        tglx@...utronix.de, luto@...capital.net
Subject: Re: [PATCH] syscalls: define and explain goal to not call syscalls
 in the kernel

Jon,

On Fri, Mar 30, 2018 at 09:35:18AM -0600, Jonathan Corbet wrote:
> On Sun, 25 Mar 2018 18:25:27 +0200
> Dominik Brodowski <linux@...inikbrodowski.net> wrote:
> 
> > As there have been multiple inquiries on the rationale of my patchsets
> > removing in-kernel calls to sys_xyzzy(), here is an updated patch 01/NN
> > which I will push upstream for v4.17-rc1. I will also include a reference
> > to this mail (and therefore to the explanation below) in all related
> > patches of the series. Any improvements, hints, suggestions, spelling
> > fixes, and/or objections?
> 
> I have no objections to the text, but I do wonder about the placement.
> The "adding syscalls" document isn't about *invoking* them; I suspect that
> few people will see it there.  The coding-style document isn't quite right
> either, but I wonder if it might not be a better place in the short term?

Well, most of the existing instances where syscalls were called in the
kernel were common codepaths for old and new syscalls or native and compat
syscalls, and syscall multiplexers like sys_ipc() which got replaced or
superseded by many new syscalls. That's what lead me to 
Documentation/process/adding-syscalls.rst . I'm happy to move this text to
Documentation/process/coding-style.rst (as new section 21?), or even to
Documentation/process/do-not-call-syscalls.rst . Just let me know what you
prefer me to push upstream.

Thanks,
	Dominik