Re: [PATCH v2] x86: Preserve iopl on fork and execve

From: Arjan van de Ven <arjanvandeven@gmail.com>
To: Ingo Molnar <mingo@kernel.org>
Cc: Alex Henrie <alexhenrie24@gmail.com>,
	One Thousand Gnomes <gnomes@lxorguk.ukuu.org.uk>,
	Kees Cook <keescook@chromium.org>,
	"H . Peter Anvin" <hpa@zytor.com>,
	Doug Johnson <dougvj@gmail.com>,
	Thomas Gleixner <tglx@linutronix.de>,
	Ingo Molnar <mingo@redhat.com>,
	Tyler Hicks <tyhicks@canonical.com>,
	Al Viro <viro@zeniv.linux.org.uk>,
	LKML <linux-kernel@vger.kernel.org>,
	Andy Lutomirski <luto@kernel.org>,
	Linus Torvalds <torvalds@linux-foundation.org>,
	Andrew Morton <akpm@linux-foundation.org>,
	Borislav Petkov <bp@alien8.de>,
	Peter Zijlstra <a.p.zijlstra@chello.nl>,
	Arjan van de Ven <arjan@infradead.org>,
	Denys Vlasenko <dvlasenk@redhat.com>,
	Brian Gerst <brgerst@gmail.com>
Subject: Re: [PATCH v2] x86: Preserve iopl on fork and execve
Date: Tue, 12 May 2015 08:24:38 -0700	[thread overview]
Message-ID: <CADyApD3Kom234GhvNdxTiF8xpUdqp9cLp8TuGLTkrQ=Z3ESkfg@mail.gmail.com> (raw)
In-Reply-To: <20150512064032.GA25097@gmail.com>

On Mon, May 11, 2015 at 11:40 PM, Ingo Molnar <mingo@kernel.org> wrote:
>  - Nothing actually broke that people cared about in the last 2.5
>    years, thus this might be one of the (very very rare) cases where
>    preserving a breakage is the right thing to do.

>  - These syscalls are rarely used, and we could as well insist that
>    every new context should have the permissions to (re-)acquire them
>    and should actively seek them - instead of inheriting it to shells
>    via system(), etc. The best strategy with dangerous APIs is to make
>    it really, really explicit when they are used.

since nothing really broke and its a "nasty either way" regression
wise, picking the more secure path looks the most sane.

the most likely impact path is in the X world, where X normally gets
iopl type permissions (even thought it doesn't need
them anymore nowadays).. reverting this behavior would give all the
processes X spawns off those perms as well...

also the interesting question is:
can a process give up these perms?
otherwise it becomes a "once given, never gotten rid of" hell hole.