[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <Zr4GCYkfof1aJ-hp@pathway.suse.cz>
Date: Thu, 15 Aug 2024 15:43:37 +0200
From: Petr Mladek <pmladek@...e.com>
To: Miroslav Benes <mbenes@...e.cz>
Cc: Josh Poimboeuf <jpoimboe@...nel.org>,
Joe Lawrence <joe.lawrence@...hat.com>,
Nicolai Stange <nstange@...e.de>, live-patching@...r.kernel.org,
linux-kernel@...r.kernel.org
Subject: Re: [POC 2/7] livepatch: Allow to handle lifetime of shadow
variables using the livepatch state
On Thu 2024-07-25 13:31:40, Miroslav Benes wrote:
> > diff --git a/kernel/livepatch/state.c b/kernel/livepatch/state.c
> > index 6693d808106b..4ec65afe3a43 100644
> > --- a/kernel/livepatch/state.c
> > +++ b/kernel/livepatch/state.c
> > @@ -198,11 +198,17 @@ void klp_release_states(struct klp_patch *patch)
> > if (is_state_in_other_patches(patch, state))
> > continue;
> >
> > - if (!state->callbacks.release)
> > - continue;
> > -
> > - if (state->callbacks.setup_succeeded)
> > + if (state->callbacks.release && state->callbacks.setup_succeeded)
> > state->callbacks.release(patch, state);
> > +
> > + if (state->is_shadow)
> > + klp_shadow_free_all(state->id, state->callbacks.shadow_dtor);
>
> The following
>
> > + /*
> > + * The @release callback is supposed to restore the original
> > + * state before the @setup callback was called.
> > + */
> > + state->callbacks.setup_succeeded = 0;
>
> should go to the previous patch perhaps?
Great catch!
I am going to refactor the code in the next version so that it would
look like:
void klp_states_post_unpatch(struct klp_patch *patch)
{
struct klp_state *state;
klp_for_each_state(patch, state) {
if (is_state_in_other_patches(patch, state))
continue;
if (!state->callbacks.pre_patch_succeeded)
continue;
if (state->callbacks.post_unpatch)
state->callbacks.post_unpatch(patch, state);
+ if (state->is_shadow)
+ klp_shadow_free_all(state->id, state->callbacks.shadow_dtor);
+
state->callbacks.pre_patch_succeeded = 0;
}
}
Best Regards,
Petr
Powered by blists - more mailing lists