All of lore.kernel.org
 help / color / mirror / Atom feed
From: Imre Deak <imre.deak@intel.com>
To: "José Roberto de Souza" <jose.souza@intel.com>
Cc: intel-gfx@lists.freedesktop.org
Subject: Re: [PATCH 15/16] drm/i915: Power down any power well left on by BIOS
Date: Mon, 15 Oct 2018 14:06:37 +0300	[thread overview]
Message-ID: <20181015110637.GB11298@ideak-desk.fi.intel.com> (raw)
In-Reply-To: <20181012215218.5119-15-jose.souza@intel.com>

On Fri, Oct 12, 2018 at 02:52:17PM -0700, José Roberto de Souza wrote:
> Just not enable power wells is not enough as BIOS/firmware can turn
> on some power wells during boot, so is needed disable those to save
> power and to avoid mismatch state errors in
> intel_power_domains_verify_state().
> So here disabling every non-real power well first as it could have
> some dependency in a real power well and then disabling all power
> wells in reverse(power well 2 depends on power well 1 and so on)
> other as required by spec.

You can't disable a power well while the function depending on it is
still on, that would lead to a hang. Also some power wells can only be
enabled/disabled at a specific place in the modeset sequence, so
disabling them here would lead to timeouts. The correct way to get power
wells disbled is to disable the corresponding functions by doing a
modeset disabling any active outputs.

> 
> Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
> ---
>  drivers/gpu/drm/i915/intel_runtime_pm.c | 59 +++++++++++++++++++++++++
>  1 file changed, 59 insertions(+)
> 
> diff --git a/drivers/gpu/drm/i915/intel_runtime_pm.c b/drivers/gpu/drm/i915/intel_runtime_pm.c
> index 56c65d921acd..0f5016b74228 100644
> --- a/drivers/gpu/drm/i915/intel_runtime_pm.c
> +++ b/drivers/gpu/drm/i915/intel_runtime_pm.c
> @@ -3785,6 +3785,61 @@ static void vlv_cmnlane_wa(struct drm_i915_private *dev_priv)
>  
>  static void intel_power_domains_verify_state(struct drm_i915_private *dev_priv);
>  
> +static void
> +intel_power_domains_disable_leftovers(struct drm_i915_private *dev_priv)
> +{
> +	struct i915_power_domains *power_domains = &dev_priv->power_domains;
> +	struct i915_power_well *power_well;
> +	int i;
> +
> +	mutex_lock(&power_domains->lock);
> +
> +	/* Disable everything that is enabled and is not a HW power_well */
> +	for_each_power_well(dev_priv, power_well) {
> +		WARN_ON(power_well->count);
> +
> +		/*
> +		 * Power wells not belonging to any domain (like the MISC_IO
> +		 * and PW1 power wells) are under FW control, so ignore them,
> +		 * since their state can change asynchronously.
> +		 */
> +		if (!power_well->desc->domains || power_well->desc->always_on)
> +			continue;
> +
> +		if (power_well->desc->id != DISP_PW_ID_NONE)
> +			continue;
> +
> +		if (!power_well->hw_enabled)
> +			continue;
> +
> +		intel_power_well_disable(dev_priv, power_well);
> +	}
> +
> +	/* Disabled HW power wells in reverse order, so power well 2 is
> +	 * disabled before power well 1 and so on as required by spec.
> +	 */
> +	for (i = power_domains->power_well_count - 1; i >= 0; i--) {
> +		power_well = &power_domains->power_wells[i];
> +
> +		WARN_ON(power_well->count);
> +
> +		if (!power_well->desc->domains || power_well->desc->always_on)
> +			continue;
> +
> +		if (power_well->desc->id == DISP_PW_ID_NONE)
> +			continue;
> +
> +		if (!power_well->hw_enabled)
> +			continue;
> +
> +		intel_power_well_disable(dev_priv, power_well);
> +	}
> +
> +	mutex_unlock(&power_domains->lock);
> +
> +	intel_power_domains_verify_state(dev_priv);
> +}
> +
>  /**
>   * intel_power_domains_init_hw - initialize hardware power domain state
>   * @dev_priv: i915 device instance
> @@ -3838,6 +3893,10 @@ void intel_power_domains_init_hw(struct drm_i915_private *dev_priv, bool resume)
>  		intel_display_power_get(dev_priv, POWER_DOMAIN_INIT);
>  	intel_power_domains_sync_hw(dev_priv);
>  
> +	/* Disable everything left enabled by BIOS/firmware */
> +	if (!INTEL_INFO(dev_priv)->num_pipes)
> +		intel_power_domains_disable_leftovers(dev_priv);
> +
>  	power_domains->initializing = false;
>  }
>  
> -- 
> 2.19.1
> 
> _______________________________________________
> Intel-gfx mailing list
> Intel-gfx@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/intel-gfx
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

  reply	other threads:[~2018-10-15 11:08 UTC|newest]

Thread overview: 44+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-10-12 21:52 [PATCH 01/16] drm/i915: Properly set PCH as NOP when display is disabled José Roberto de Souza
2018-10-12 21:52 ` [PATCH 02/16] drm/i915: Move out non-display related calls from display/modeset init/cleanup José Roberto de Souza
2018-10-15 11:14   ` Chris Wilson
2018-10-15 23:51     ` Souza, Jose
2018-10-12 21:52 ` [PATCH 03/16] drm/i915: Move drm_vblank_init() to i915_load_modeset_init() José Roberto de Souza
2018-10-12 21:52 ` [PATCH 04/16] drm/i915: Move FBC init and cleanup calls to modeset functions José Roberto de Souza
2018-10-12 21:52 ` [PATCH 05/16] drm/i915: Move intel_init_ipc() call to i915_load_modeset_init() José Roberto de Souza
2018-10-12 21:52 ` [PATCH 06/16] drm/i915: Don't call modeset related functions when display is disabled José Roberto de Souza
2018-10-22  8:25   ` Jani Nikula
2018-10-22  8:37     ` Chris Wilson
2018-10-22  9:00       ` Jani Nikula
2018-10-25 17:38         ` Souza, Jose
2019-04-08 20:50         ` Chris Wilson
2019-08-24  9:51           ` Chris Wilson
2018-10-12 21:52 ` [PATCH 07/16] drm/i915: Remove redundant checks for num_pipes == 0 José Roberto de Souza
2018-10-22  8:31   ` Jani Nikula
2018-10-12 21:52 ` [PATCH 08/16] drm/i915: Keep overlay functions naming consistent José Roberto de Souza
2018-10-22  8:32   ` Jani Nikula
2018-10-12 21:52 ` [PATCH 09/16] drm/i915: Do not reset display when display is disabled José Roberto de Souza
2018-10-22  8:34   ` Jani Nikula
2018-10-12 21:52 ` [PATCH 10/16] drm/i915: Do not initialize display clocks " José Roberto de Souza
2018-10-22  8:37   ` Jani Nikula
2018-10-12 21:52 ` [PATCH 11/16] drm/i915: Do not initialize display core " José Roberto de Souza
2018-10-22  8:39   ` Jani Nikula
2018-10-12 21:52 ` [PATCH 12/16] drm/i915: Warn when display irq functions is executed " José Roberto de Souza
2018-10-22  8:40   ` Jani Nikula
2018-10-12 21:52 ` [PATCH 13/16] drm/i915: Do not print DC off mismatch state when DMC firmware in not loaded José Roberto de Souza
2018-10-15 10:58   ` Imre Deak
2018-10-16  1:31     ` Souza, Jose
2018-10-16  9:35       ` Imre Deak
2018-10-12 21:52 ` [PATCH 14/16] drm/i915: Do not turn power wells on or off when display is disabled José Roberto de Souza
2018-10-22  8:50   ` Jani Nikula
2018-10-12 21:52 ` [PATCH 15/16] drm/i915: Power down any power well left on by BIOS José Roberto de Souza
2018-10-15 11:06   ` Imre Deak [this message]
2018-10-16  0:05     ` Souza, Jose
2018-10-16  9:39       ` Imre Deak
2018-10-12 21:52 ` [PATCH 16/16] drm/i915: Guard debugfs against invalid access when display is disabled José Roberto de Souza
2018-10-22  8:52   ` Jani Nikula
2018-10-12 22:03 ` ✗ Fi.CI.CHECKPATCH: warning for series starting with [01/16] drm/i915: Properly set PCH as NOP " Patchwork
2018-10-12 22:07 ` ✗ Fi.CI.SPARSE: " Patchwork
2018-10-12 22:27 ` ✓ Fi.CI.BAT: success " Patchwork
2018-10-13  2:46 ` ✓ Fi.CI.IGT: " Patchwork
2018-11-03 21:41 ` [PATCH 01/16] " Jani Nikula
2018-11-30  8:29   ` Lucas De Marchi

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=20181015110637.GB11298@ideak-desk.fi.intel.com \
    --to=imre.deak@intel.com \
    --cc=intel-gfx@lists.freedesktop.org \
    --cc=jose.souza@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.