Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Date:   Thu, 16 Apr 2020 11:45:05 +0200 (CEST)
From:   Miroslav Benes <mbenes@suse.cz>
To:     Josh Poimboeuf <jpoimboe@redhat.com>
cc:     Peter Zijlstra <peterz@infradead.org>,
        live-patching@vger.kernel.org, linux-kernel@vger.kernel.org,
        Jessica Yu <jeyu@kernel.org>
Subject: Re: [PATCH 0/7] livepatch,module: Remove .klp.arch and
 module_disable_ro()
In-Reply-To: <20200414190814.glra2gceqgy34iyx@treble>
Message-ID: <alpine.LSU.2.21.2004161136340.10475@pobox.suse.cz>
References: <cover.1586881704.git.jpoimboe@redhat.com> <20200414182726.GF2483@worktop.programming.kicks-ass.net> <20200414190814.glra2gceqgy34iyx@treble>
User-Agent: Alpine 2.21 (LSU 202 2017-01-01)
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Tue, 14 Apr 2020, Josh Poimboeuf wrote:

> On Tue, Apr 14, 2020 at 08:27:26PM +0200, Peter Zijlstra wrote:
> > On Tue, Apr 14, 2020 at 11:28:36AM -0500, Josh Poimboeuf wrote:
> > > Better late than never, these patches add simplifications and
> > > improvements for some issues Peter found six months ago, as part of his
> > > non-writable text code (W^X) cleanups.
> > 
> > Excellent stuff, thanks!!
> >
> > I'll go brush up these two patches then:
> > 
> >   https://lkml.kernel.org/r/20191018074634.801435443@infradead.org
> >   https://lkml.kernel.org/r/20191018074634.858645375@infradead.org
> 
> Ah right, I meant to bring that up.  I actually played around with those
> patches.  While it would be nice to figure out a way to converge the
> ftrace module init, I didn't really like the first patch.
> 
> It bothers me that both the notifiers and the module init() both see the
> same MODULE_STATE_COMING state, but only in the former case is the text
> writable.
> 
> I think it's cognitively simpler if MODULE_STATE_COMING always means the
> same thing, like the comments imply, "fully formed" and thus
> not-writable:
> 
> enum module_state {
> 	MODULE_STATE_LIVE,	/* Normal state. */
> 	MODULE_STATE_COMING,	/* Full formed, running module_init. */
> 	MODULE_STATE_GOING,	/* Going away. */
> 	MODULE_STATE_UNFORMED,	/* Still setting it up. */
> };
> 
> And, it keeps tighter constraints on what a notifier can do, which is a
> good thing if we can get away with it.

Agreed.

On the other hand, the first patch would remove the tiny race window when 
a module state is still UNFORMED, but the protections are (being) set up. 
Patches 4/7 and 5/7 allow to use memcpy in that case, because it is early. 
But it is in fact not already. I haven't checked yet if it really matters 
somewhere (a race with livepatch running klp_module_coming while another 
module is being loaded or anything like that).

Miroslav