[PATCH] drm/i915: Fix AUX power domain toggling across TypeC mode resets

[PATCH] drm/i915: Fix AUX power domain toggling across TypeC mode resets

[Imre Deak]

Make sure to select the port's AUX power domain while holding the TC
port lock. The domain depends on the port's current TC mode, which may
get changed under us if we're not holding the lock.

This was left out from
commit 8c10e2262663 ("drm/i915: Keep the TypeC port mode fixed for detect/AUX transfers")

Cc: <stable@vger.kernel.org> # v5.4+
Signed-off-by: Imre Deak <imre.deak@intel.com>
---
 drivers/gpu/drm/i915/display/intel_dp.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/i915/display/intel_dp.c b/drivers/gpu/drm/i915/display/intel_dp.c
index 408c3c1c5e81..40d42dcff0b7 100644
--- a/drivers/gpu/drm/i915/display/intel_dp.c
+++ b/drivers/gpu/drm/i915/display/intel_dp.c
@@ -1359,8 +1359,7 @@ intel_dp_aux_xfer(struct intel_dp *intel_dp,
 	bool is_tc_port = intel_phy_is_tc(i915, phy);
 	i915_reg_t ch_ctl, ch_data[5];
 	u32 aux_clock_divider;
-	enum intel_display_power_domain aux_domain =
-		intel_aux_power_domain(intel_dig_port);
+	enum intel_display_power_domain aux_domain;
 	intel_wakeref_t aux_wakeref;
 	intel_wakeref_t pps_wakeref;
 	int i, ret, recv_bytes;
@@ -1375,6 +1374,8 @@ intel_dp_aux_xfer(struct intel_dp *intel_dp,
 	if (is_tc_port)
 		intel_tc_port_lock(intel_dig_port);
 
+	aux_domain = intel_aux_power_domain(intel_dig_port);
+
 	aux_wakeref = intel_display_power_get(i915, aux_domain);
 	pps_wakeref = pps_lock(intel_dp);
 
-- 
2.23.1

Re: [Intel-gfx] [PATCH] drm/i915: Fix AUX power domain toggling across TypeC mode resets

[Souza, Jose]

On Thu, 2020-05-14 at 23:45 +0300, Imre Deak wrote:
> Make sure to select the port's AUX power domain while holding the TC
> port lock. The domain depends on the port's current TC mode, which may
> get changed under us if we're not holding the lock.
> 
> This was left out from
> commit 8c10e2262663 ("drm/i915: Keep the TypeC port mode fixed for detect/AUX transfers")
> 

Reviewed-by: José Roberto de Souza <jose.souza@intel.com>

> Cc: <stable@vger.kernel.org> # v5.4+
> Signed-off-by: Imre Deak <imre.deak@intel.com>
> ---
>  drivers/gpu/drm/i915/display/intel_dp.c | 5 +++--
>  1 file changed, 3 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/gpu/drm/i915/display/intel_dp.c b/drivers/gpu/drm/i915/display/intel_dp.c
> index 408c3c1c5e81..40d42dcff0b7 100644
> --- a/drivers/gpu/drm/i915/display/intel_dp.c
> +++ b/drivers/gpu/drm/i915/display/intel_dp.c
> @@ -1359,8 +1359,7 @@ intel_dp_aux_xfer(struct intel_dp *intel_dp,
>  	bool is_tc_port = intel_phy_is_tc(i915, phy);
>  	i915_reg_t ch_ctl, ch_data[5];
>  	u32 aux_clock_divider;
> -	enum intel_display_power_domain aux_domain =
> -		intel_aux_power_domain(intel_dig_port);
> +	enum intel_display_power_domain aux_domain;
>  	intel_wakeref_t aux_wakeref;
>  	intel_wakeref_t pps_wakeref;
>  	int i, ret, recv_bytes;
> @@ -1375,6 +1374,8 @@ intel_dp_aux_xfer(struct intel_dp *intel_dp,
>  	if (is_tc_port)
>  		intel_tc_port_lock(intel_dig_port);
>  
> +	aux_domain = intel_aux_power_domain(intel_dig_port);
> +
>  	aux_wakeref = intel_display_power_get(i915, aux_domain);
>  	pps_wakeref = pps_lock(intel_dp);
>  

Re: [Intel-gfx] [PATCH] drm/i915: Fix AUX power domain toggling across TypeC mode resets

[Ville Syrjälä]

On Thu, May 14, 2020 at 11:45:53PM +0300, Imre Deak wrote:
> Make sure to select the port's AUX power domain while holding the TC
> port lock. The domain depends on the port's current TC mode, which may
> get changed under us if we're not holding the lock.

Can we toss in a lockdep assert?

Did this by any chance help with the MST issues you were seeing?

> 
> This was left out from
> commit 8c10e2262663 ("drm/i915: Keep the TypeC port mode fixed for detect/AUX transfers")
> 
> Cc: <stable@vger.kernel.org> # v5.4+
> Signed-off-by: Imre Deak <imre.deak@intel.com>
> ---
>  drivers/gpu/drm/i915/display/intel_dp.c | 5 +++--
>  1 file changed, 3 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/gpu/drm/i915/display/intel_dp.c b/drivers/gpu/drm/i915/display/intel_dp.c
> index 408c3c1c5e81..40d42dcff0b7 100644
> --- a/drivers/gpu/drm/i915/display/intel_dp.c
> +++ b/drivers/gpu/drm/i915/display/intel_dp.c
> @@ -1359,8 +1359,7 @@ intel_dp_aux_xfer(struct intel_dp *intel_dp,
>  	bool is_tc_port = intel_phy_is_tc(i915, phy);
>  	i915_reg_t ch_ctl, ch_data[5];
>  	u32 aux_clock_divider;
> -	enum intel_display_power_domain aux_domain =
> -		intel_aux_power_domain(intel_dig_port);
> +	enum intel_display_power_domain aux_domain;
>  	intel_wakeref_t aux_wakeref;
>  	intel_wakeref_t pps_wakeref;
>  	int i, ret, recv_bytes;
> @@ -1375,6 +1374,8 @@ intel_dp_aux_xfer(struct intel_dp *intel_dp,
>  	if (is_tc_port)
>  		intel_tc_port_lock(intel_dig_port);
>  
> +	aux_domain = intel_aux_power_domain(intel_dig_port);
> +
>  	aux_wakeref = intel_display_power_get(i915, aux_domain);
>  	pps_wakeref = pps_lock(intel_dp);
>  
> -- 
> 2.23.1
> 
> _______________________________________________
> Intel-gfx mailing list
> Intel-gfx@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/intel-gfx

-- 
Ville Syrjälä
Intel

Re: [Intel-gfx] [PATCH] drm/i915: Fix AUX power domain toggling across TypeC mode resets

[Imre Deak]

On Fri, May 15, 2020 at 07:21:06PM +0300, Ville Syrjälä wrote:
> On Thu, May 14, 2020 at 11:45:53PM +0300, Imre Deak wrote:
> > Make sure to select the port's AUX power domain while holding the TC
> > port lock. The domain depends on the port's current TC mode, which may
> > get changed under us if we're not holding the lock.
> 
> Can we toss in a lockdep assert?

Yes, can do that. However, now I realize, there is one more place to fix
in intel_dp_force(), where I think we could just drop the aux get/put.

So are you ok to do that as a follow-up?

> Did this by any chance help with the MST issues you were seeing?

No, at least it didn't resolve it. Still some missing responses to MST
down messages. More precisely the sink does send interrupts as a
response, but the interrupt handler won't see the DP_DOWN_REP_MSG_RDY
being set in ESI.

> > This was left out from
> > commit 8c10e2262663 ("drm/i915: Keep the TypeC port mode fixed for detect/AUX transfers")
> > 
> > Cc: <stable@vger.kernel.org> # v5.4+
> > Signed-off-by: Imre Deak <imre.deak@intel.com>
> > ---
> >  drivers/gpu/drm/i915/display/intel_dp.c | 5 +++--
> >  1 file changed, 3 insertions(+), 2 deletions(-)
> > 
> > diff --git a/drivers/gpu/drm/i915/display/intel_dp.c b/drivers/gpu/drm/i915/display/intel_dp.c
> > index 408c3c1c5e81..40d42dcff0b7 100644
> > --- a/drivers/gpu/drm/i915/display/intel_dp.c
> > +++ b/drivers/gpu/drm/i915/display/intel_dp.c
> > @@ -1359,8 +1359,7 @@ intel_dp_aux_xfer(struct intel_dp *intel_dp,
> >  	bool is_tc_port = intel_phy_is_tc(i915, phy);
> >  	i915_reg_t ch_ctl, ch_data[5];
> >  	u32 aux_clock_divider;
> > -	enum intel_display_power_domain aux_domain =
> > -		intel_aux_power_domain(intel_dig_port);
> > +	enum intel_display_power_domain aux_domain;
> >  	intel_wakeref_t aux_wakeref;
> >  	intel_wakeref_t pps_wakeref;
> >  	int i, ret, recv_bytes;
> > @@ -1375,6 +1374,8 @@ intel_dp_aux_xfer(struct intel_dp *intel_dp,
> >  	if (is_tc_port)
> >  		intel_tc_port_lock(intel_dig_port);
> >  
> > +	aux_domain = intel_aux_power_domain(intel_dig_port);
> > +
> >  	aux_wakeref = intel_display_power_get(i915, aux_domain);
> >  	pps_wakeref = pps_lock(intel_dp);
> >  
> > -- 
> > 2.23.1
> > 
> > _______________________________________________
> > Intel-gfx mailing list
> > Intel-gfx@lists.freedesktop.org
> > https://lists.freedesktop.org/mailman/listinfo/intel-gfx
> 
> -- 
> Ville Syrjälä
> Intel

Re: [Intel-gfx] [PATCH] drm/i915: Fix AUX power domain toggling across TypeC mode resets

[Ville Syrjälä]

On Fri, May 15, 2020 at 08:24:45PM +0300, Imre Deak wrote:
> On Fri, May 15, 2020 at 07:21:06PM +0300, Ville Syrjälä wrote:
> > On Thu, May 14, 2020 at 11:45:53PM +0300, Imre Deak wrote:
> > > Make sure to select the port's AUX power domain while holding the TC
> > > port lock. The domain depends on the port's current TC mode, which may
> > > get changed under us if we're not holding the lock.
> > 
> > Can we toss in a lockdep assert?
> 
> Yes, can do that. However, now I realize, there is one more place to fix
> in intel_dp_force(), where I think we could just drop the aux get/put.
> 
> So are you ok to do that as a follow-up?

Sure.

> 
> > Did this by any chance help with the MST issues you were seeing?
> 
> No, at least it didn't resolve it. Still some missing responses to MST
> down messages. More precisely the sink does send interrupts as a
> response, but the interrupt handler won't see the DP_DOWN_REP_MSG_RDY
> being set in ESI.

:(

> 
> > > This was left out from
> > > commit 8c10e2262663 ("drm/i915: Keep the TypeC port mode fixed for detect/AUX transfers")
> > > 
> > > Cc: <stable@vger.kernel.org> # v5.4+
> > > Signed-off-by: Imre Deak <imre.deak@intel.com>
> > > ---
> > >  drivers/gpu/drm/i915/display/intel_dp.c | 5 +++--
> > >  1 file changed, 3 insertions(+), 2 deletions(-)
> > > 
> > > diff --git a/drivers/gpu/drm/i915/display/intel_dp.c b/drivers/gpu/drm/i915/display/intel_dp.c
> > > index 408c3c1c5e81..40d42dcff0b7 100644
> > > --- a/drivers/gpu/drm/i915/display/intel_dp.c
> > > +++ b/drivers/gpu/drm/i915/display/intel_dp.c
> > > @@ -1359,8 +1359,7 @@ intel_dp_aux_xfer(struct intel_dp *intel_dp,
> > >  	bool is_tc_port = intel_phy_is_tc(i915, phy);
> > >  	i915_reg_t ch_ctl, ch_data[5];
> > >  	u32 aux_clock_divider;
> > > -	enum intel_display_power_domain aux_domain =
> > > -		intel_aux_power_domain(intel_dig_port);
> > > +	enum intel_display_power_domain aux_domain;
> > >  	intel_wakeref_t aux_wakeref;
> > >  	intel_wakeref_t pps_wakeref;
> > >  	int i, ret, recv_bytes;
> > > @@ -1375,6 +1374,8 @@ intel_dp_aux_xfer(struct intel_dp *intel_dp,
> > >  	if (is_tc_port)
> > >  		intel_tc_port_lock(intel_dig_port);
> > >  
> > > +	aux_domain = intel_aux_power_domain(intel_dig_port);
> > > +
> > >  	aux_wakeref = intel_display_power_get(i915, aux_domain);
> > >  	pps_wakeref = pps_lock(intel_dp);
> > >  
> > > -- 
> > > 2.23.1
> > > 
> > > _______________________________________________
> > > Intel-gfx mailing list
> > > Intel-gfx@lists.freedesktop.org
> > > https://lists.freedesktop.org/mailman/listinfo/intel-gfx
> > 
> > -- 
> > Ville Syrjälä
> > Intel

-- 
Ville Syrjälä
Intel