linux-kernel - Re: [PATCH] thermal: intel_powerclamp: Check for a possible array out of bounds access.

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: Windows password security audit tool. GUI, reports in PDF.

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

Message-ID: <9a4660438e0466c04880fad4450215707921521e.camel@linux.intel.com>
Date:   Mon, 05 Jun 2023 09:36:54 -0700
From:   srinivas pandruvada <srinivas.pandruvada@...ux.intel.com>
To:     "Rafael J. Wysocki" <rafael@...nel.org>,
        Daniil Dulov <d.dulov@...ddin.ru>,
        Zhang Rui <rui.zhang@...el.com>
Cc:     Daniel Lezcano <daniel.lezcano@...aro.org>,
        Amit Kucheria <amitk@...nel.org>,
        Chen Yu <yu.c.chen@...el.com>,
        Jacob Pan <jacob.jun.pan@...ux.intel.com>,
        linux-pm@...r.kernel.org, linux-kernel@...r.kernel.org,
        lvc-project@...uxtesting.org
Subject: Re: [PATCH] thermal: intel_powerclamp: Check for a possible array
 out of bounds access.

On Sun, 2023-06-04 at 18:13 +0200, Rafael J. Wysocki wrote:
> Cc list trimmed.
> 
> On Fri, Jun 2, 2023 at 11:12 AM Daniil Dulov <d.dulov@...ddin.ru>
> wrote:
> > 
> > ratio may be equal to MAX_TARGET_RATIO - 1 that will result in
> > out of bound access.
> > 
The description was not clear to me.

May be something like this:

"
The max value of "ratio" parameter passed to the function
get_compensation() is MAX_TARGET_RATIO - 1. The size of cal_data array
is MAX_TARGET_RATIO. Hence, accessing cal_data[ratio + 1], will result
in out of bound access.

Add a condition to check ratio < MAX_TARGET_RATIO - 1, before accsssing
cal_data[ratio + 1].
"

The change is correct. But for actual code change,
the ratio < MAX_TARGET_RATIO - 1 is also checked in else if() before,
which can be merged to check only once for this condition.

Thanks,
Srinivas






> > Found by Linux Verification Center (linuxtesting.org) with SVACE.
> > 
> > Fixes: d6d71ee4a14a ("PM: Introduce Intel PowerClamp Driver")
> > Signed-off-by: Daniil Dulov <d.dulov@...ddin.ru>
> > ---
> >  drivers/thermal/intel/intel_powerclamp.c | 3 ++-
> >  1 file changed, 2 insertions(+), 1 deletion(-)
> > 
> > diff --git a/drivers/thermal/intel/intel_powerclamp.c
> > b/drivers/thermal/intel/intel_powerclamp.c
> > index fb04470d7d4b..9deaf2b8ccf6 100644
> > --- a/drivers/thermal/intel/intel_powerclamp.c
> > +++ b/drivers/thermal/intel/intel_powerclamp.c
> > @@ -277,7 +277,8 @@ static unsigned int get_compensation(int ratio)
> >                 comp = (cal_data[ratio].steady_comp +
> >                         cal_data[ratio - 1].steady_comp +
> >                         cal_data[ratio - 2].steady_comp) / 3;
> > -       } else if (cal_data[ratio].confidence >= CONFIDENCE_OK &&

I think the concern is that cal_data[ratio + 1] is going out of bound
for the "ratio" passed to this function, when ratio ==
ARRAY_SIZE(cal_data) - 1;

Here size of cal_data array is 50. The max possible "ratio" passed can
be 49.




> > +       } else if (ratio < MAX_TARGET_RATIO - 1 &&
> > +               cal_data[ratio].confidence >= CONFIDENCE_OK &&
> >                 cal_data[ratio - 1].confidence >= CONFIDENCE_OK &&
> >                 cal_data[ratio + 1].confidence >= CONFIDENCE_OK) {
> >                 comp = (cal_data[ratio].steady_comp +
> > --
> 
> Rui, Srinivas, can you have a look at this, please?