[Intel-gfx] [bug report] drm/i915/tc: Fix TypeC PHY connect/disconnect logic on ADL-P
Dan Carpenter
dan.carpenter at oracle.com
Fri Oct 1 14:52:10 UTC 2021
On Fri, Oct 01, 2021 at 05:20:17PM +0300, Ville Syrjälä wrote:
> On Fri, Oct 01, 2021 at 04:25:35PM +0300, Dan Carpenter wrote:
> > Hello Imre Deak,
> >
> > This is a semi-automatic email about new static checker warnings.
> >
> > The patch 3e0abc7661c8: "drm/i915/tc: Fix TypeC PHY
> > connect/disconnect logic on ADL-P" from Sep 29, 2021, leads to the
> > following Smatch complaint:
> >
> > drivers/gpu/drm/i915/display/intel_ddi.c:4028 intel_ddi_encoder_destroy()
> > warn: variable dereferenced before check 'dig_port' (see line 4020)
> >
> > drivers/gpu/drm/i915/display/intel_ddi.c
> > 4019 struct intel_digital_port *dig_port = enc_to_dig_port(to_intel_encoder(encoder));
> > 4020 enum phy phy = intel_port_to_phy(i915, dig_port->base.port);
> > ^^^^^^^^^^^^^^^^^^^
> > The patch adds a new unchecked dereference.
> >
> > 4021
> > 4022 intel_dp_encoder_flush_work(encoder);
> > 4023 if (intel_phy_is_tc(i915, phy))
> > 4024 intel_tc_port_flush_work(dig_port);
> > 4025 intel_display_power_flush_work(i915);
> > 4026
> > 4027 drm_encoder_cleanup(encoder);
> > 4028 if (dig_port)
> > ^^^^^^^^
> > But the existing code checked for NULL.
>
> That check is nonsense. Feel free to nuke it.
Thanks! Will do.
regards,
dan carpenter
More information about the Intel-gfx
mailing list