| Message ID | 20221220094618.207126-1-maarten.lankhorst@linux.intel.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | drm/i915/display: Fix a use-after-free when intel_edp_init_connector fails | expand |
On Tue, 20 Dec 2022, Maarten Lankhorst <maarten.lankhorst@linux.intel.com> wrote: > We enable the DP aux channel during probe, but may free the connector > soon afterwards. Ensure the DP aux display power put is completed before > everything is freed, to prevent a use-after-free in icl_aux_pw_to_phy(), > called from icl_combo_phy_aux_power_well_disable. Feels like the placement of the intel_display_power_flush_work_sync() call in intel_dp_aux_fini() is a bit arbitrary. If we add it in intel_dp_aux_fini(), the async and sync waits will both be called on the regular encoder destroy path. Maybe both intel_ddi_encoder_destroy() and intel_dp_encoder_destroy() should call intel_display_power_flush_work_sync(), instead of async, and maybe the error paths should call those functions instead of just drm_encoder_cleanup()? Imre? BR, Jani. > > Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> > --- > drivers/gpu/drm/i915/display/intel_display_power.c | 2 +- > drivers/gpu/drm/i915/display/intel_display_power.h | 1 + > drivers/gpu/drm/i915/display/intel_dp_aux.c | 2 ++ > 3 files changed, 4 insertions(+), 1 deletion(-) > > diff --git a/drivers/gpu/drm/i915/display/intel_display_power.c b/drivers/gpu/drm/i915/display/intel_display_power.c > index 04915f85a0df..0edb5532461f 100644 > --- a/drivers/gpu/drm/i915/display/intel_display_power.c > +++ b/drivers/gpu/drm/i915/display/intel_display_power.c > @@ -776,7 +776,7 @@ void intel_display_power_flush_work(struct drm_i915_private *i915) > * Like intel_display_power_flush_work(), but also ensure that the work > * handler function is not running any more when this function returns. > */ > -static void > +void > intel_display_power_flush_work_sync(struct drm_i915_private *i915) > { > struct i915_power_domains *power_domains = &i915->display.power.domains; > diff --git a/drivers/gpu/drm/i915/display/intel_display_power.h b/drivers/gpu/drm/i915/display/intel_display_power.h > index 7136ea3f233e..dc10ee0519e6 100644 > --- a/drivers/gpu/drm/i915/display/intel_display_power.h > +++ b/drivers/gpu/drm/i915/display/intel_display_power.h > @@ -188,6 +188,7 @@ void __intel_display_power_put_async(struct drm_i915_private *i915, > enum intel_display_power_domain domain, > intel_wakeref_t wakeref); > void intel_display_power_flush_work(struct drm_i915_private *i915); > +void intel_display_power_flush_work_sync(struct drm_i915_private *i915); > #if IS_ENABLED(CONFIG_DRM_I915_DEBUG_RUNTIME_PM) > void intel_display_power_put(struct drm_i915_private *dev_priv, > enum intel_display_power_domain domain, > diff --git a/drivers/gpu/drm/i915/display/intel_dp_aux.c b/drivers/gpu/drm/i915/display/intel_dp_aux.c > index f1835c74bff0..1006dddad2d5 100644 > --- a/drivers/gpu/drm/i915/display/intel_dp_aux.c > +++ b/drivers/gpu/drm/i915/display/intel_dp_aux.c > @@ -680,6 +680,8 @@ void intel_dp_aux_fini(struct intel_dp *intel_dp) > if (cpu_latency_qos_request_active(&intel_dp->pm_qos)) > cpu_latency_qos_remove_request(&intel_dp->pm_qos); > > + /* Ensure async work from intel_dp_aux_xfer() is flushed before we clean up */ > + intel_display_power_flush_work_sync(dp_to_i915(intel_dp)); > kfree(intel_dp->aux.name); > }
On Tue, Dec 20, 2022 at 02:40:47PM +0200, Jani Nikula wrote: > On Tue, 20 Dec 2022, Maarten Lankhorst <maarten.lankhorst@linux.intel.com> wrote: > > We enable the DP aux channel during probe, but may free the connector > > soon afterwards. Ensure the DP aux display power put is completed before > > everything is freed, to prevent a use-after-free in icl_aux_pw_to_phy(), > > called from icl_combo_phy_aux_power_well_disable. > > Feels like the placement of the intel_display_power_flush_work_sync() > call in intel_dp_aux_fini() is a bit arbitrary. > > If we add it in intel_dp_aux_fini(), the async and sync waits will both > be called on the regular encoder destroy path. Yes, calling intel_display_power_flush_work() from the error handler at the end of intel_dp_init_connector() would be better. > Maybe both intel_ddi_encoder_destroy() and intel_dp_encoder_destroy() > should call intel_display_power_flush_work_sync(), instead of async, intel_display_power_flush_work() ensures that power wells without a reference held are disabled when it returns, so no need to call the _sync() version for encoders (the _sync() version ensures in addition during driver unloading that the work function is not running). > and maybe the error paths should call those functions instead of just > drm_encoder_cleanup()? Yes, the cleanup in those functions could be shared with the error handling in g4x_dp_init() and intel_ddi_init(), except kfree(dig_port) which also happens if drm_encoder_init() fails. For this intel_pps_vdd_off_sync() / intel_dp_aux_fini() would also happen later at the end of g4x_dp_init()/intel_ddi_init(), I guess that's ok. I wonder if not handling drm_encoder_init() error in intel_ddi_init() was on purpose, can't see a reason for it. > Imre? > > > BR, > Jani. > > > > > > Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> > > --- > > drivers/gpu/drm/i915/display/intel_display_power.c | 2 +- > > drivers/gpu/drm/i915/display/intel_display_power.h | 1 + > > drivers/gpu/drm/i915/display/intel_dp_aux.c | 2 ++ > > 3 files changed, 4 insertions(+), 1 deletion(-) > > > > diff --git a/drivers/gpu/drm/i915/display/intel_display_power.c b/drivers/gpu/drm/i915/display/intel_display_power.c > > index 04915f85a0df..0edb5532461f 100644 > > --- a/drivers/gpu/drm/i915/display/intel_display_power.c > > +++ b/drivers/gpu/drm/i915/display/intel_display_power.c > > @@ -776,7 +776,7 @@ void intel_display_power_flush_work(struct drm_i915_private *i915) > > * Like intel_display_power_flush_work(), but also ensure that the work > > * handler function is not running any more when this function returns. > > */ > > -static void > > +void > > intel_display_power_flush_work_sync(struct drm_i915_private *i915) > > { > > struct i915_power_domains *power_domains = &i915->display.power.domains; > > diff --git a/drivers/gpu/drm/i915/display/intel_display_power.h b/drivers/gpu/drm/i915/display/intel_display_power.h > > index 7136ea3f233e..dc10ee0519e6 100644 > > --- a/drivers/gpu/drm/i915/display/intel_display_power.h > > +++ b/drivers/gpu/drm/i915/display/intel_display_power.h > > @@ -188,6 +188,7 @@ void __intel_display_power_put_async(struct drm_i915_private *i915, > > enum intel_display_power_domain domain, > > intel_wakeref_t wakeref); > > void intel_display_power_flush_work(struct drm_i915_private *i915); > > +void intel_display_power_flush_work_sync(struct drm_i915_private *i915); > > #if IS_ENABLED(CONFIG_DRM_I915_DEBUG_RUNTIME_PM) > > void intel_display_power_put(struct drm_i915_private *dev_priv, > > enum intel_display_power_domain domain, > > diff --git a/drivers/gpu/drm/i915/display/intel_dp_aux.c b/drivers/gpu/drm/i915/display/intel_dp_aux.c > > index f1835c74bff0..1006dddad2d5 100644 > > --- a/drivers/gpu/drm/i915/display/intel_dp_aux.c > > +++ b/drivers/gpu/drm/i915/display/intel_dp_aux.c > > @@ -680,6 +680,8 @@ void intel_dp_aux_fini(struct intel_dp *intel_dp) > > if (cpu_latency_qos_request_active(&intel_dp->pm_qos)) > > cpu_latency_qos_remove_request(&intel_dp->pm_qos); > > > > + /* Ensure async work from intel_dp_aux_xfer() is flushed before we clean up */ > > + intel_display_power_flush_work_sync(dp_to_i915(intel_dp)); > > kfree(intel_dp->aux.name); > > } > > -- > Jani Nikula, Intel Open Source Graphics Center
On Tue, 20 Dec 2022, Imre Deak <imre.deak@intel.com> wrote: > On Tue, Dec 20, 2022 at 02:40:47PM +0200, Jani Nikula wrote: >> On Tue, 20 Dec 2022, Maarten Lankhorst <maarten.lankhorst@linux.intel.com> wrote: >> > We enable the DP aux channel during probe, but may free the connector >> > soon afterwards. Ensure the DP aux display power put is completed before >> > everything is freed, to prevent a use-after-free in icl_aux_pw_to_phy(), >> > called from icl_combo_phy_aux_power_well_disable. >> >> Feels like the placement of the intel_display_power_flush_work_sync() >> call in intel_dp_aux_fini() is a bit arbitrary. >> >> If we add it in intel_dp_aux_fini(), the async and sync waits will both >> be called on the regular encoder destroy path. > > Yes, calling intel_display_power_flush_work() from the error handler at > the end of intel_dp_init_connector() would be better. > >> Maybe both intel_ddi_encoder_destroy() and intel_dp_encoder_destroy() >> should call intel_display_power_flush_work_sync(), instead of async, > > intel_display_power_flush_work() ensures that power wells without a > reference held are disabled when it returns, so no need to call the > _sync() version for encoders (the _sync() version ensures in addition > during driver unloading that the work function is not running). > >> and maybe the error paths should call those functions instead of just >> drm_encoder_cleanup()? > > Yes, the cleanup in those functions could be shared with the error > handling in g4x_dp_init() and intel_ddi_init(), except kfree(dig_port) > which also happens if drm_encoder_init() fails. > > For this intel_pps_vdd_off_sync() / intel_dp_aux_fini() would also > happen later at the end of g4x_dp_init()/intel_ddi_init(), I guess > that's ok. > > I wonder if not handling drm_encoder_init() error in intel_ddi_init() > was on purpose, can't see a reason for it. This patch has been forgotten... Maarten, can you follow up on it please? BR, Jani. > >> Imre? >> >> >> BR, >> Jani. >> >> >> > >> > Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> >> > --- >> > drivers/gpu/drm/i915/display/intel_display_power.c | 2 +- >> > drivers/gpu/drm/i915/display/intel_display_power.h | 1 + >> > drivers/gpu/drm/i915/display/intel_dp_aux.c | 2 ++ >> > 3 files changed, 4 insertions(+), 1 deletion(-) >> > >> > diff --git a/drivers/gpu/drm/i915/display/intel_display_power.c b/drivers/gpu/drm/i915/display/intel_display_power.c >> > index 04915f85a0df..0edb5532461f 100644 >> > --- a/drivers/gpu/drm/i915/display/intel_display_power.c >> > +++ b/drivers/gpu/drm/i915/display/intel_display_power.c >> > @@ -776,7 +776,7 @@ void intel_display_power_flush_work(struct drm_i915_private *i915) >> > * Like intel_display_power_flush_work(), but also ensure that the work >> > * handler function is not running any more when this function returns. >> > */ >> > -static void >> > +void >> > intel_display_power_flush_work_sync(struct drm_i915_private *i915) >> > { >> > struct i915_power_domains *power_domains = &i915->display.power.domains; >> > diff --git a/drivers/gpu/drm/i915/display/intel_display_power.h b/drivers/gpu/drm/i915/display/intel_display_power.h >> > index 7136ea3f233e..dc10ee0519e6 100644 >> > --- a/drivers/gpu/drm/i915/display/intel_display_power.h >> > +++ b/drivers/gpu/drm/i915/display/intel_display_power.h >> > @@ -188,6 +188,7 @@ void __intel_display_power_put_async(struct drm_i915_private *i915, >> > enum intel_display_power_domain domain, >> > intel_wakeref_t wakeref); >> > void intel_display_power_flush_work(struct drm_i915_private *i915); >> > +void intel_display_power_flush_work_sync(struct drm_i915_private *i915); >> > #if IS_ENABLED(CONFIG_DRM_I915_DEBUG_RUNTIME_PM) >> > void intel_display_power_put(struct drm_i915_private *dev_priv, >> > enum intel_display_power_domain domain, >> > diff --git a/drivers/gpu/drm/i915/display/intel_dp_aux.c b/drivers/gpu/drm/i915/display/intel_dp_aux.c >> > index f1835c74bff0..1006dddad2d5 100644 >> > --- a/drivers/gpu/drm/i915/display/intel_dp_aux.c >> > +++ b/drivers/gpu/drm/i915/display/intel_dp_aux.c >> > @@ -680,6 +680,8 @@ void intel_dp_aux_fini(struct intel_dp *intel_dp) >> > if (cpu_latency_qos_request_active(&intel_dp->pm_qos)) >> > cpu_latency_qos_remove_request(&intel_dp->pm_qos); >> > >> > + /* Ensure async work from intel_dp_aux_xfer() is flushed before we clean up */ >> > + intel_display_power_flush_work_sync(dp_to_i915(intel_dp)); >> > kfree(intel_dp->aux.name); >> > } >> >> -- >> Jani Nikula, Intel Open Source Graphics Center
diff --git a/drivers/gpu/drm/i915/display/intel_display_power.c b/drivers/gpu/drm/i915/display/intel_display_power.c index 04915f85a0df..0edb5532461f 100644 --- a/drivers/gpu/drm/i915/display/intel_display_power.c +++ b/drivers/gpu/drm/i915/display/intel_display_power.c @@ -776,7 +776,7 @@ void intel_display_power_flush_work(struct drm_i915_private *i915) * Like intel_display_power_flush_work(), but also ensure that the work * handler function is not running any more when this function returns. */ -static void +void intel_display_power_flush_work_sync(struct drm_i915_private *i915) { struct i915_power_domains *power_domains = &i915->display.power.domains; diff --git a/drivers/gpu/drm/i915/display/intel_display_power.h b/drivers/gpu/drm/i915/display/intel_display_power.h index 7136ea3f233e..dc10ee0519e6 100644 --- a/drivers/gpu/drm/i915/display/intel_display_power.h +++ b/drivers/gpu/drm/i915/display/intel_display_power.h @@ -188,6 +188,7 @@ void __intel_display_power_put_async(struct drm_i915_private *i915, enum intel_display_power_domain domain, intel_wakeref_t wakeref); void intel_display_power_flush_work(struct drm_i915_private *i915); +void intel_display_power_flush_work_sync(struct drm_i915_private *i915); #if IS_ENABLED(CONFIG_DRM_I915_DEBUG_RUNTIME_PM) void intel_display_power_put(struct drm_i915_private *dev_priv, enum intel_display_power_domain domain, diff --git a/drivers/gpu/drm/i915/display/intel_dp_aux.c b/drivers/gpu/drm/i915/display/intel_dp_aux.c index f1835c74bff0..1006dddad2d5 100644 --- a/drivers/gpu/drm/i915/display/intel_dp_aux.c +++ b/drivers/gpu/drm/i915/display/intel_dp_aux.c @@ -680,6 +680,8 @@ void intel_dp_aux_fini(struct intel_dp *intel_dp) if (cpu_latency_qos_request_active(&intel_dp->pm_qos)) cpu_latency_qos_remove_request(&intel_dp->pm_qos); + /* Ensure async work from intel_dp_aux_xfer() is flushed before we clean up */ + intel_display_power_flush_work_sync(dp_to_i915(intel_dp)); kfree(intel_dp->aux.name); }
We enable the DP aux channel during probe, but may free the connector soon afterwards. Ensure the DP aux display power put is completed before everything is freed, to prevent a use-after-free in icl_aux_pw_to_phy(), called from icl_combo_phy_aux_power_well_disable. Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> --- drivers/gpu/drm/i915/display/intel_display_power.c | 2 +- drivers/gpu/drm/i915/display/intel_display_power.h | 1 + drivers/gpu/drm/i915/display/intel_dp_aux.c | 2 ++ 3 files changed, 4 insertions(+), 1 deletion(-)