linux-kernel - Re: [PATCH v1 3/7] powercap/dtpm: Fixup kfree for virtual node

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 for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Thu, 17 Feb 2022 16:45:43 +0100
From:   Ulf Hansson <ulf.hansson@...aro.org>
To:     Daniel Lezcano <daniel.lezcano@...aro.org>
Cc:     rjw@...ysocki.net, heiko@...ech.de, lukasz.luba@....com,
        linux-kernel@...r.kernel.org, linux-pm@...r.kernel.org,
        Daniel Lezcano <daniel.lezcano@...nel.org>,
        "Rafael J. Wysocki" <rafael@...nel.org>
Subject: Re: [PATCH v1 3/7] powercap/dtpm: Fixup kfree for virtual node

On Thu, 17 Feb 2022 at 14:54, Daniel Lezcano <daniel.lezcano@...aro.org> wrote:
>
> On 17/02/2022 14:17, Ulf Hansson wrote:
> > On Wed, 16 Feb 2022 at 19:10, Daniel Lezcano <daniel.lezcano@...aro.org> wrote:
> >>
> >> On 16/02/2022 17:22, Ulf Hansson wrote:
> >>> On Sun, 30 Jan 2022 at 22:02, Daniel Lezcano <daniel.lezcano@...aro.org> wrote:
> >>>>
> >>>> When the node is virtual there is no release function associated which
> >>>> can free the memory.
> >>>>
> >>>> Free the memory when no 'ops' exists.
> >>>>
> >>>> Signed-off-by: Daniel Lezcano <daniel.lezcano@...aro.org>
> >>>> ---
> >>>>    drivers/powercap/dtpm.c | 4 ++--
> >>>>    1 file changed, 2 insertions(+), 2 deletions(-)
> >>>>
> >>>> diff --git a/drivers/powercap/dtpm.c b/drivers/powercap/dtpm.c
> >>>> index 0b0121c37a1b..7bddd25a6767 100644
> >>>> --- a/drivers/powercap/dtpm.c
> >>>> +++ b/drivers/powercap/dtpm.c
> >>>> @@ -181,12 +181,12 @@ int dtpm_release_zone(struct powercap_zone *pcz)
> >>>>
> >>>>           if (dtpm->ops)
> >>>>                   dtpm->ops->release(dtpm);
> >>>> +       else
> >>>> +               kfree(dtpm);
> >>>>
> >>>
> >>> This doesn't look correct. Below you check dtpm against "root", which
> >>> may be after its memory has been freed.
> >>>
> >>> If the ->release() function should be responsible for freeing the
> >>> dtpm, it needs to be called after the check below.
> >>
> >> It is harmless, 'root' is not dereferenced but used as an ID
> >>
> >> Moreover, in the patch 5/7 it is moved out this function.
> >
> > Right. It just looks a bit odd here.
> >
> >>
> >>
> >>>>           if (root == dtpm)
> >>>>                   root = NULL;
> >>>>
> >>>> -       kfree(dtpm);
> >
> > So then why doesn't this kfree do the job already?
> >
> > kfree(NULL) works fine, if dtpm->ops->release(dtpm) already freed the data.
>
> The description is confusing.
>
> Actually, there is a double kfree. When there is a ops->release, the
> kfree is done there and again a few lines after.
>
> The issue was introduced with the change where dtpm had a private data
> field to store the backend specific structure and was converted to a
> backend specific structure containing a dtpm node [1].
>
> So this function was calling release from the dtpm backend which was
> freeing the specific data in the dtpm->private and then here was freeing
> the dtpm. Now, the backend frees the structure which contains the dtpm
> structure, so when returning from ops->release(), dtpm is already free.
>
> I should change the description and add a Fixes tag to the change
> described above.

Does ops->release() also resets the "dtpm" pointer to NULL? If not,
it's good practice that it should, right?

In that case, we would be calling "kfree(NULL);" the second time,
which is perfectly fine.

>
> [1]
> https://lore.kernel.org/r/20210312130411.29833-4-daniel.lezcano@linaro.org
>

Kind regards
Uffe