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] [thread-next>] [day] [month] [year] [list]
Message-ID: <CADnq5_NWX7u=S+jrC8YA6fJxN7GXpSN+kqsQieqphdOz2HT6EA@mail.gmail.com>
Date: Fri, 6 Sep 2024 11:02:32 -0400
From: Alex Deucher <alexdeucher@...il.com>
To: Mukul Sikka <mukul.sikka@...adcom.com>
Cc: stable@...r.kernel.org, gregkh@...uxfoundation.org, evan.quan@....com, 
	alexander.deucher@....com, christian.koenig@....com, airlied@...ux.ie, 
	daniel@...ll.ch, Jun.Ma2@....com, kevinyang.wang@....com, sashal@...nel.org, 
	amd-gfx@...ts.freedesktop.org, dri-devel@...ts.freedesktop.org, 
	linux-kernel@...r.kernel.org, ajay.kaher@...adcom.com, 
	alexey.makhalov@...adcom.com, vasavi.sirnapalli@...adcom.com, 
	Bob Zhou <bob.zhou@....com>, Tim Huang <Tim.Huang@....com>
Subject: Re: [PATCH v5.15-v5.10] drm/amd/pm: Fix the null pointer dereference
 for vega10_hwmgr

On Fri, Sep 6, 2024 at 4:50 AM Mukul Sikka <mukul.sikka@...adcom.com> wrote:
>
> On Fri, Sep 6, 2024 at 12:05 AM Alex Deucher <alexdeucher@...il.com> wrote:
> >
> > On Tue, Sep 3, 2024 at 5:53 AM sikkamukul <mukul.sikka@...adcom.com> wrote:
> > >
> > > From: Bob Zhou <bob.zhou@....com>
> > >
> > > [ Upstream commit 50151b7f1c79a09117837eb95b76c2de76841dab ]
> > >
> > > Check return value and conduct null pointer handling to avoid null pointer dereference.
> > >
> > > Signed-off-by: Bob Zhou <bob.zhou@....com>
> > > Reviewed-by: Tim Huang <Tim.Huang@....com>
> > > Signed-off-by: Alex Deucher <alexander.deucher@....com>
> > > Signed-off-by: Sasha Levin <sashal@...nel.org>
> > > Signed-off-by: Mukul Sikka <mukul.sikka@...adcom.com>
> >
> > Just out of curiosity, are you actually seeing an issue?  This and a
> > lot of the other recent NULL check patches are just static checker
> > fixes.  They don't actually fix a known issue.
> >
> No, according to the description of this patch and CVE-2024-43905.
> It seems to be applicable to LTS.

I don't know that this is really CVE material, but oh well.  I'm not
sure if it's actually possible to hit this in practice.

Alex

>
> - Mukul
>
> > > ---
> > >  .../drm/amd/pm/powerplay/hwmgr/vega10_hwmgr.c | 30 ++++++++++++++++---
> > >  1 file changed, 26 insertions(+), 4 deletions(-)
> > >
> > > diff --git a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/vega10_hwmgr.c b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/vega10_hwmgr.c
> > > index 10678b519..304874cba 100644
> > > --- a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/vega10_hwmgr.c
> > > +++ b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/vega10_hwmgr.c
> > > @@ -3391,13 +3391,17 @@ static int vega10_find_dpm_states_clocks_in_dpm_table(struct pp_hwmgr *hwmgr, co
> > >         const struct vega10_power_state *vega10_ps =
> > >                         cast_const_phw_vega10_power_state(states->pnew_state);
> > >         struct vega10_single_dpm_table *sclk_table = &(data->dpm_table.gfx_table);
> > > -       uint32_t sclk = vega10_ps->performance_levels
> > > -                       [vega10_ps->performance_level_count - 1].gfx_clock;
> > >         struct vega10_single_dpm_table *mclk_table = &(data->dpm_table.mem_table);
> > > -       uint32_t mclk = vega10_ps->performance_levels
> > > -                       [vega10_ps->performance_level_count - 1].mem_clock;
> > > +       uint32_t sclk, mclk;
> > >         uint32_t i;
> > >
> > > +       if (vega10_ps == NULL)
> > > +               return -EINVAL;
> > > +       sclk = vega10_ps->performance_levels
> > > +                       [vega10_ps->performance_level_count - 1].gfx_clock;
> > > +       mclk = vega10_ps->performance_levels
> > > +                       [vega10_ps->performance_level_count - 1].mem_clock;
> > > +
> > >         for (i = 0; i < sclk_table->count; i++) {
> > >                 if (sclk == sclk_table->dpm_levels[i].value)
> > >                         break;
> > > @@ -3704,6 +3708,9 @@ static int vega10_generate_dpm_level_enable_mask(
> > >                         cast_const_phw_vega10_power_state(states->pnew_state);
> > >         int i;
> > >
> > > +       if (vega10_ps == NULL)
> > > +               return -EINVAL;
> > > +
> > >         PP_ASSERT_WITH_CODE(!vega10_trim_dpm_states(hwmgr, vega10_ps),
> > >                         "Attempt to Trim DPM States Failed!",
> > >                         return -1);
> > > @@ -4828,6 +4835,9 @@ static int vega10_check_states_equal(struct pp_hwmgr *hwmgr,
> > >
> > >         psa = cast_const_phw_vega10_power_state(pstate1);
> > >         psb = cast_const_phw_vega10_power_state(pstate2);
> > > +       if (psa == NULL || psb == NULL)
> > > +               return -EINVAL;
> > > +
> > >         /* If the two states don't even have the same number of performance levels they cannot be the same state. */
> > >         if (psa->performance_level_count != psb->performance_level_count) {
> > >                 *equal = false;
> > > @@ -4953,6 +4963,8 @@ static int vega10_set_sclk_od(struct pp_hwmgr *hwmgr, uint32_t value)
> > >                 return -EINVAL;
> > >
> > >         vega10_ps = cast_phw_vega10_power_state(&ps->hardware);
> > > +       if (vega10_ps == NULL)
> > > +               return -EINVAL;
> > >
> > >         vega10_ps->performance_levels
> > >         [vega10_ps->performance_level_count - 1].gfx_clock =
> > > @@ -5004,6 +5016,8 @@ static int vega10_set_mclk_od(struct pp_hwmgr *hwmgr, uint32_t value)
> > >                 return -EINVAL;
> > >
> > >         vega10_ps = cast_phw_vega10_power_state(&ps->hardware);
> > > +       if (vega10_ps == NULL)
> > > +               return -EINVAL;
> > >
> > >         vega10_ps->performance_levels
> > >         [vega10_ps->performance_level_count - 1].mem_clock =
> > > @@ -5239,6 +5253,9 @@ static void vega10_odn_update_power_state(struct pp_hwmgr *hwmgr)
> > >                 return;
> > >
> > >         vega10_ps = cast_phw_vega10_power_state(&ps->hardware);
> > > +       if (vega10_ps == NULL)
> > > +               return;
> > > +
> > >         max_level = vega10_ps->performance_level_count - 1;
> > >
> > >         if (vega10_ps->performance_levels[max_level].gfx_clock !=
> > > @@ -5261,6 +5278,9 @@ static void vega10_odn_update_power_state(struct pp_hwmgr *hwmgr)
> > >
> > >         ps = (struct pp_power_state *)((unsigned long)(hwmgr->ps) + hwmgr->ps_size * (hwmgr->num_ps - 1));
> > >         vega10_ps = cast_phw_vega10_power_state(&ps->hardware);
> > > +       if (vega10_ps == NULL)
> > > +               return;
> > > +
> > >         max_level = vega10_ps->performance_level_count - 1;
> > >
> > >         if (vega10_ps->performance_levels[max_level].gfx_clock !=
> > > @@ -5451,6 +5471,8 @@ static int vega10_get_performance_level(struct pp_hwmgr *hwmgr, const struct pp_
> > >                 return -EINVAL;
> > >
> > >         ps = cast_const_phw_vega10_power_state(state);
> > > +       if (ps == NULL)
> > > +               return -EINVAL;
> > >
> > >         i = index > ps->performance_level_count - 1 ?
> > >                         ps->performance_level_count - 1 : index;
> > > --
> > > 2.39.4
> > >

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ