Message-ID: <20180417175755.tlhb7iza2rgu4nyo@treble>
Date:   Tue, 17 Apr 2018 12:57:55 -0500
From:   Josh Poimboeuf <jpoimboe@...hat.com>
To:     Petr Mladek <pmladek@...e.com>
Cc:     Miroslav Benes <mbenes@...e.cz>, Jiri Kosina <jikos@...nel.org>,
        Jason Baron <jbaron@...mai.com>,
        Joe Lawrence <joe.lawrence@...hat.com>,
        Jessica Yu <jeyu@...nel.org>,
        Evgenii Shatokhin <eshatokhin@...tuozzo.com>,
        live-patching@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH v10 05/10] livepatch: Support separate list for replaced
 patches.

On Tue, Apr 17, 2018 at 05:37:46PM +0200, Petr Mladek wrote:
> On Mon 2018-04-16 14:04:25, Josh Poimboeuf wrote:
> > On Mon, Apr 16, 2018 at 04:58:11PM +0200, Petr Mladek wrote:
> > > On Wed 2018-04-11 10:48:52, Josh Poimboeuf wrote:
> > > > On Wed, Apr 11, 2018 at 04:17:11PM +0200, Petr Mladek wrote:
> > > > > Second, unrelated patches must never patch the same functions.
> > > > > Otherwise we would not be able to define which implementation
> > > > > should be used. This is especially important when a patch is
> > > > > removed and we need to fallback either to another patch or
> > > > > original code. Yes, it makes perfect sense. But it needs code
> > > > > that will check it, refuse loading the patch, ... It is not
> > > > > complicated. But it is rather additional code than
> > > > > simplification. I might make livepatching more safe
> > > > > but probably not simplify the code.
> > > > 
> > > > We don't need to enforce that.  The func stack can stay.  If somebody
> > > > wants to patch the same function multiple times (without using
> > > > 'replace'), that's inadvisable, but it's also their business.  They're
> > > > responsible for the tooling to ensure the patch stack order is sane.
> > > 
> > > 
> > > While it might make sense to ignore the patch stack (ordering) for
> > > the enable operation. Do we really want to ignore it when disabling
> > > a patch.
> > > 
> > > By other words, do we want to allow disabling a patch that is in
> > > the middle of the stack, only partly in use? Does not this allow
> > > some other crazy scenarios? Is it really the user business?
> > > Will it make our life easier?
> > 
> > If there's no longer a patch stack, then there's no concept of a middle.
> > We would expect the patches to be independent of one another, and so
> > disabling any of them independently would be harmless.
> >
> > If any of the patches share a func, and the user disables one in the
> > "middle", it's not our job to support that.  The vendor / patch author
> > should prevent such cases from occurring with tooling, packaging,
> > documentation, etc.  Or they can just use 'replace'.
> > 
> > We can already have similar unexpected situations today.  For example,
> > what if patch B is a cumulative superset of patch A, but the user
> > mistakenly loads patch A (without replace) *after* loading patch B?
> > Then some unforeseen craziness could ensue.
> > 
> > We can't control all such scenarios (and that's ok), but we shouldn't
> > pretend that we support them.
> 
> I recall some earlier discussion. The aim was to make livepatching
> more safe. AFAIK, it was more related to the helper tooling around,
> like "automatic" generation of the special relocation entries, ...
> Anyway, I feel that the above gets somehow against it.
> 
> The patch stack does not solve everything. But it makes it harder
> to do wrong things.

It also makes it harder to do _right_ things.  For example, disabling a
patch in the "middle" should be allowable if there are no patch
dependencies.

> > > This will not happen if we refuse to load non-replace patches
> > > that touch an already patches fucntion. Then the patch stack
> > > might become only implementation detail. It will not define
> > > the ordering any longer.
> > 
> > I think this would only be a partial solution.  Patches can have
> > implicit interdependencies, even if they don't patch the same function.
> > Also it doesn't solve the problem when patches are loaded in the wrong
> > order.  We have to trust vendors and admins to do the right thing.
> 
> I would still like to add this check if we remove the stack.
> Is it too restrictive? Maybe. But it would help to prevent
> creating some ugly mistakes. By other words, we will force
> people using proper cumulative patches instead of
> dangerous semi-cumulative Frankenstains.
> 
> At least, it would reduce the number of scenarios that
> we might meet and eventually need to debug.

But the non-replace cumulative model can be perfectly safe, if the
packaging/tooling supports the right workflow.

I think we're talking about theoretical situations anyway.  I'd rather
not add any code for theoretical problems.

If somebody reports a bug due to dangerous real world usage, then we can
decide how to handle it at that point -- though I suspect my response
will be "don't do that!"

> > > > > > > > > Another possibility would be to get rid of the enable/disable states.
> > > > > > > > > I mean that the patch will be automatically enabled during
> > > > > > > > > registration and removed during unregistration.
> > > 
> > > OK. What about the following solution?
> > > 
> > >    + Enable patch directly from klp_register_patch()
> > >    + Unregister the patch directly from klp_complete_transition()
> > >      when the patch is disabled.
> > >    + Put the module later when all sysfs entries are removed
> > >      in klp_unregister_patch().
> > > 
> > > As a result:
> > > 
> > >    + module_init() will need to call only klp_register_patch()
> > >    + module_exit() will do nothing
> > >    + the module can be removed only when it is not longer needed
> > 
> > Sounds good to me.
> 
> It should be consistent with sysfs interface, see below.
> 
> 
> > > Some other ideas:
> > > 
> > >    + rename /sys/kernel/livepatch/<patch>/enable -> unregister
> > >      allow to write into /sys/kernel/livepatch/<patch>/transition
> > > 
> > >      + echo 1 >unregistrer  to disable&unregister the patch
> > >      + echo 0 >transition   to revert the running transition
> > 
> > Why not keep the existing sysfs interfaces?  So
> > 
> >   echo 0 > enable
> > 
> > would disable (and eventually unregister) the patch.
> 
> First, it would be a bit inconsistent. The patch would be added by calling
> "register" function and removed by writing into "enabled" file.

True.  Although from the end user (and tooling) standpoint, it looks the
same as before, since the register/unregister and enable are done
automatically by the patch module.  The only user-visible difference is
that the sysfs directory disappears and the patch module can't be
re-enabled.

> Second, if we remove sysfs entries for disabled patches, it would
> make the meaning of "enabled" file a bit useless. I mean that
> the patch will always be enabled when the sysfs directory exists
> (when ignoring transition states).

Though as you mentioned, it does still indicate which direction the
transition is going.

> BTW: I have just today got a feedback that the user interface is not
> ideal. In particular, it is not easy to detect what process is blocking
> the transition. It is because the target value in
> /proc/<pid>/patch_state depends on the value in
> /sys/kernel/livepatch/<patch>/enable (direction of the transition).

But hopefully tooling can hide the intricacies of the interface.

> Anyway, the first problem might be solved by doing patch registration
> at the beginning of klp_enable_patch(). Then it will match the
> "enabled" sysfs file.
> 
> I would actually feel more comfortable if we do not change the
> sysfs interface.
> 
> Well, any better idea for a simplified interface is appreciated.

We can just keep the current interface, which I think will make tooling
happy since it will function basically the same as before.

> > > The question is what to do with the stack of patches. It will have
> > > no meaning for the enable operation because it will be done
> > > automatically. But what about the disable/unregistrer operation?
> > 
> > Assuming we got rid of the patch stack, would we even need to keep a
> > global list of patches anymore?
> 
> As Mirek said, a list still might be useful in some situations
> (nops generation, conflicts check). But it would become an
> implementation detail. The ordering would not be important
> any longer.

Ok.

-- 
Josh

Powered by blists - more mailing lists