All of lore.kernel.org
 help / color / mirror / Atom feed
From: Imre Deak <imre.deak@intel.com>
To: "Ville Syrjälä" <ville.syrjala@linux.intel.com>
Cc: intel-gfx@lists.freedesktop.org
Subject: Re: [Intel-gfx] [PATCH v4 13/14] drm/i915/tc: Call TypeC port flush_work/cleanup without modeset locks held
Date: Fri, 12 May 2023 16:38:07 +0300	[thread overview]
Message-ID: <ZF5BP68S41e26o8x@ideak-desk.fi.intel.com> (raw)
In-Reply-To: <ZF0oqq8BD62Hl7Xo@intel.com>

On Thu, May 11, 2023 at 08:40:58PM +0300, Ville Syrjälä wrote:
> On Wed, May 10, 2023 at 05:10:22PM +0300, Imre Deak wrote:
> > On Wed, May 10, 2023 at 05:03:17PM +0300, Ville Syrjälä wrote:
> > > On Wed, May 10, 2023 at 01:31:30PM +0300, Imre Deak wrote:
> > > > Call the TypeC port flush_work and cleanup handlers without the modeset
> > > > locks held. These don't require the locks, as the work takes - as it
> > > > should be able to at any point in time - any locks it needs and by the
> > > > time cleanup is called and after cleanup returns the encoder is not in
> > > > use.
> > > > 
> > > > This is required by the next patch canceling a TypeC port work
> > > > synchronously during encoder suspend and shutdown, where the work can
> > > > take modeset locks as well, hence the canceling must be done without
> > > > holding the locks.
> > > > 
> > > > I also considered moving the modeset locking down to each encoder
> > > > suspend()/shutdown() hook instead, however locking the full modeset
> > > > state for each encoder separately would be odd, and the bigger change -
> > > > affecting all encoders - is beyond the scope of this patchset.
> > > 
> > > Hmm. What is it in the encoder shutdown/suspend hooks that
> > > actually needs the modeset locks?
> > 
> > In the case of intel_dp_encoder_suspend() for instance, I assume
> > nothing, since the VDD work and intel_pps_vdd_off_sync() should take
> > whatever locks they require.
> > 
> > So presumably most (all) of those could be made lockless. However, I
> > would like to leave that kind of change for a follow-up if possible, not
> > to affect in this patchset any other encoder types (again because of
> > possible need for CC stable).
> 
> Yeah sure.

Ok, will add a TODO comment about this.

> A bit irksome to have to add vfuncs and whatnot for
> it though, but it's not the end of the world.

Direct intel_tc calls from intel_suspend/shutdown_encoders didn't seem
ideal either, but can do that if you think that's better.

> Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
> 
> > 
> > > 
> > > > 
> > > > Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
> > > > Signed-off-by: Imre Deak <imre.deak@intel.com>
> > > > ---
> > > >  drivers/gpu/drm/i915/display/intel_ddi.c      | 27 +++++++++----------
> > > >  .../drm/i915/display/intel_display_types.h    | 12 +++++++++
> > > >  drivers/gpu/drm/i915/i915_driver.c            |  8 ++++++
> > > >  3 files changed, 33 insertions(+), 14 deletions(-)
> > > > 
> > > > diff --git a/drivers/gpu/drm/i915/display/intel_ddi.c b/drivers/gpu/drm/i915/display/intel_ddi.c
> > > > index 813be957ed11b..7d09bd2412352 100644
> > > > --- a/drivers/gpu/drm/i915/display/intel_ddi.c
> > > > +++ b/drivers/gpu/drm/i915/display/intel_ddi.c
> > > > @@ -4617,31 +4617,27 @@ static bool intel_ddi_is_tc(struct drm_i915_private *i915, enum port port)
> > > >  
> > > >  static void intel_ddi_encoder_suspend(struct intel_encoder *encoder)
> > > >  {
> > > > -	struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
> > > > -	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
> > > > -	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
> > > > -	enum phy phy = intel_port_to_phy(i915, encoder->port);
> > > > -
> > > >  	intel_dp_encoder_suspend(encoder);
> > > > +}
> > > >  
> > > > -	if (!intel_phy_is_tc(i915, phy))
> > > > -		return;
> > > > +static void intel_ddi_tc_encoder_suspend_complete(struct intel_encoder *encoder)
> > > > +{
> > > > +	struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
> > > > +	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
> > > >  
> > > >  	intel_tc_port_flush_work(dig_port);
> > > >  }
> > > >  
> > > >  static void intel_ddi_encoder_shutdown(struct intel_encoder *encoder)
> > > >  {
> > > > -	struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
> > > > -	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
> > > > -	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
> > > > -	enum phy phy = intel_port_to_phy(i915, encoder->port);
> > > > -
> > > >  	intel_dp_encoder_shutdown(encoder);
> > > >  	intel_hdmi_encoder_shutdown(encoder);
> > > > +}
> > > >  
> > > > -	if (!intel_phy_is_tc(i915, phy))
> > > > -		return;
> > > > +static void intel_ddi_tc_encoder_shutdown_complete(struct intel_encoder *encoder)
> > > > +{
> > > > +	struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
> > > > +	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
> > > >  
> > > >  	intel_tc_port_cleanup(dig_port);
> > > >  }
> > > > @@ -4908,6 +4904,9 @@ void intel_ddi_init(struct drm_i915_private *dev_priv, enum port port)
> > > >  				    is_legacy ? "legacy" : "non-legacy");
> > > >  		}
> > > >  
> > > > +		encoder->suspend_complete = intel_ddi_tc_encoder_suspend_complete;
> > > > +		encoder->shutdown_complete = intel_ddi_tc_encoder_shutdown_complete;
> > > > +
> > > >  		if (intel_tc_port_init(dig_port, is_legacy) < 0)
> > > >  			goto err;
> > > >  	}
> > > > diff --git a/drivers/gpu/drm/i915/display/intel_display_types.h b/drivers/gpu/drm/i915/display/intel_display_types.h
> > > > index 270c4c84a2920..88b2a55d19f21 100644
> > > > --- a/drivers/gpu/drm/i915/display/intel_display_types.h
> > > > +++ b/drivers/gpu/drm/i915/display/intel_display_types.h
> > > > @@ -233,13 +233,25 @@ struct intel_encoder {
> > > >  	 * Called during system suspend after all pending requests for the
> > > >  	 * encoder are flushed (for example for DP AUX transactions) and
> > > >  	 * device interrupts are disabled.
> > > > +	 * All modeset locks are held while the hook is called.
> > > >  	 */
> > > >  	void (*suspend)(struct intel_encoder *);
> > > > +	/*
> > > > +	 * Called without the modeset locks held after the suspend() hook for
> > > > +	 * all encoders have been called.
> > > > +	 */
> > > > +	void (*suspend_complete)(struct intel_encoder *encoder);
> > > >  	/*
> > > >  	 * Called during system reboot/shutdown after all the
> > > >  	 * encoders have been disabled and suspended.
> > > > +	 * All modeset locks are held while the hook is called.
> > > >  	 */
> > > >  	void (*shutdown)(struct intel_encoder *encoder);
> > > > +	/*
> > > > +	 * Called without the modeset locks held after the shutdown() hook for
> > > > +	 * all encoders have been called.
> > > > +	 */
> > > > +	void (*shutdown_complete)(struct intel_encoder *encoder);
> > > >  	/*
> > > >  	 * Enable/disable the clock to the port.
> > > >  	 */
> > > > diff --git a/drivers/gpu/drm/i915/i915_driver.c b/drivers/gpu/drm/i915/i915_driver.c
> > > > index fd198700272b1..705ba65f2ff9a 100644
> > > > --- a/drivers/gpu/drm/i915/i915_driver.c
> > > > +++ b/drivers/gpu/drm/i915/i915_driver.c
> > > > @@ -964,6 +964,10 @@ static void intel_suspend_encoders(struct drm_i915_private *dev_priv)
> > > >  		if (encoder->suspend)
> > > >  			encoder->suspend(encoder);
> > > >  	drm_modeset_unlock_all(&dev_priv->drm);
> > > > +
> > > > +	for_each_intel_encoder(&dev_priv->drm, encoder)
> > > > +		if (encoder->suspend_complete)
> > > > +			encoder->suspend_complete(encoder);
> > > >  }
> > > >  
> > > >  static void intel_shutdown_encoders(struct drm_i915_private *dev_priv)
> > > > @@ -978,6 +982,10 @@ static void intel_shutdown_encoders(struct drm_i915_private *dev_priv)
> > > >  		if (encoder->shutdown)
> > > >  			encoder->shutdown(encoder);
> > > >  	drm_modeset_unlock_all(&dev_priv->drm);
> > > > +
> > > > +	for_each_intel_encoder(&dev_priv->drm, encoder)
> > > > +		if (encoder->shutdown_complete)
> > > > +			encoder->shutdown_complete(encoder);
> > > >  }
> > > >  
> > > >  void i915_driver_shutdown(struct drm_i915_private *i915)
> > > > -- 
> > > > 2.37.2
> > > 
> > > -- 
> > > Ville Syrjälä
> > > Intel
> 
> -- 
> Ville Syrjälä
> Intel

  reply	other threads:[~2023-05-12 13:37 UTC|newest]

Thread overview: 41+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-05-10 10:31 [Intel-gfx] [PATCH v4 00/14] drm/i915/tc: Add a workaround for an IOM/TCSS firmware hang issue Imre Deak
2023-05-10 10:31 ` [Intel-gfx] [PATCH v4 01/14] drm/i915: Fix PIPEDMC disabling for a bigjoiner configuration Imre Deak
2023-05-30 13:49   ` [v4,01/14] " Rudi Heitbaum
2023-05-30 13:49     ` [Intel-gfx] [v4, 01/14] " Rudi Heitbaum
2023-05-30 13:49     ` [v4,01/14] " Rudi Heitbaum
2023-05-31  8:47     ` Imre Deak
2023-05-31  8:47       ` Imre Deak
2023-05-31  8:47       ` [Intel-gfx] [v4, 01/14] " Imre Deak
2023-05-31 11:31       ` [v4,01/14] " Rudi Heitbaum
2023-05-31 11:31         ` [Intel-gfx] [v4, 01/14] " Rudi Heitbaum
2023-05-31 11:31         ` Rudi Heitbaum
2023-05-10 10:31 ` [Intel-gfx] [PATCH v4 02/14] drm/i915: Add helpers to reference/unreference a DPLL for a CRTC Imre Deak
2023-05-10 10:31 ` [Intel-gfx] [PATCH v4 03/14] drm/i915: Make the CRTC state consistent during sanitize-disabling Imre Deak
2023-05-10 10:31 ` [Intel-gfx] [PATCH v4 04/14] drm/i915: Update connector atomic state before crtc sanitize-disabling Imre Deak
2023-05-10 10:31 ` [Intel-gfx] [PATCH v4 05/14] drm/i915: Separate intel_crtc_disable_noatomic_begin/complete() Imre Deak
2023-05-10 10:31 ` [Intel-gfx] [PATCH v4 06/14] drm/i915: Factor out set_encoder_for_connector() Imre Deak
2023-05-10 10:31 ` [Intel-gfx] [PATCH v4 07/14] drm/i915: Add support for disabling any CRTCs during HW readout/sanitization Imre Deak
2023-05-10 10:31 ` [Intel-gfx] [PATCH v4 08/14] drm/i915/dp: Add link training debug and error printing helpers Imre Deak
2023-05-10 10:31 ` [Intel-gfx] [PATCH v4 09/14] drm/i915/dp: Convert link training error to debug message on disconnected sink Imre Deak
2023-05-10 10:31 ` [Intel-gfx] [PATCH v4 10/14] drm/i915/dp: Prevent link training fallback on disconnected port Imre Deak
2023-05-10 10:31 ` [Intel-gfx] [PATCH v4 11/14] drm/i915/dp: Factor out intel_dp_get_active_pipes() Imre Deak
2023-05-10 10:31 ` [Intel-gfx] [PATCH v4 12/14] drm/i915: Factor out a helper for handling atomic modeset locks/state Imre Deak
2023-05-10 10:31 ` [Intel-gfx] [PATCH v4 13/14] drm/i915/tc: Call TypeC port flush_work/cleanup without modeset locks held Imre Deak
2023-05-10 14:03   ` Ville Syrjälä
2023-05-10 14:10     ` Imre Deak
2023-05-11 17:40       ` Ville Syrjälä
2023-05-12 13:38         ` Imre Deak [this message]
2023-05-12 19:55   ` [Intel-gfx] [PATCH v5 " Imre Deak
2023-05-10 10:31 ` [Intel-gfx] [PATCH v4 14/14] drm/i915/tc: Reset TypeC PHYs left enabled in DP-alt mode after the sink disconnects Imre Deak
2023-05-11 19:21   ` Ville Syrjälä
2023-05-12 13:50     ` Imre Deak
2023-05-12 19:55   ` [Intel-gfx] [PATCH v5 " Imre Deak
2023-05-10 11:13 ` [Intel-gfx] ✗ Fi.CI.CHECKPATCH: warning for drm/i915/tc: Add a workaround for an IOM/TCSS firmware hang issue (rev11) Patchwork
2023-05-10 11:13 ` [Intel-gfx] ✗ Fi.CI.SPARSE: " Patchwork
2023-05-10 11:18 ` [Intel-gfx] ✓ Fi.CI.BAT: success " Patchwork
2023-05-10 13:52 ` [Intel-gfx] ✗ Fi.CI.IGT: failure " Patchwork
2023-05-12 21:17 ` [Intel-gfx] ✗ Fi.CI.CHECKPATCH: warning for drm/i915/tc: Add a workaround for an IOM/TCSS firmware hang issue (rev13) Patchwork
2023-05-12 21:17 ` [Intel-gfx] ✗ Fi.CI.SPARSE: " Patchwork
2023-05-12 21:31 ` [Intel-gfx] ✓ Fi.CI.BAT: success " Patchwork
2023-05-13  1:08 ` [Intel-gfx] ✓ Fi.CI.IGT: " Patchwork
2023-05-16 14:06   ` Imre Deak

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=ZF5BP68S41e26o8x@ideak-desk.fi.intel.com \
    --to=imre.deak@intel.com \
    --cc=intel-gfx@lists.freedesktop.org \
    --cc=ville.syrjala@linux.intel.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.