[Intel-gfx] [PATCH v7 3/7] drm/i915/tgl: Enable DC3CO state in "DC Off" power well
Anshuman Gupta
anshuman.gupta at intel.com
Mon Sep 9 16:19:17 UTC 2019
On 2019-09-08 at 19:44:35 +0300, Imre Deak wrote:
> On Sat, Sep 07, 2019 at 10:44:39PM +0530, Anshuman Gupta wrote:
Hi Imre,
Thanks for reviewing the pacthes i will rework the patches.
There are few comments from my side which will help to rework.
> > Add max_dc_state and tgl_set_target_dc_state() API
> > in order to enable DC3CO state with existing DC states.
> > max_dc_state will enable/disable the desired DC state in
> > DC_STATE_EN reg when "DC Off" power well gets disable/enable.
> >
> > v2: commit log improvement.
> > v3: Used intel_wait_for_register to wait for DC3CO exit. [Imre]
> > Used gen9_set_dc_state() to allow/disallow DC3CO. [Imre]
> > Moved transcoder psr2 exit line enablement from tgl_allow_dc3co()
> > to a appropriate place haswell_crtc_enable(). [Imre]
> > Changed the DC3CO power well enabled call back logic as
> > recommended in review comments. [Imre]
> > v4: Used wait_for_us() instead of intel_wait_for_reg(). [Imre (IRC)]
> > v5: using udelay() instead of waiting for DC3CO exit status.
> > v6: Fixed minor unwanted change.
> > v7: Removed DC3CO powerwell and POWER_DOMAIN_VIDEO.
> >
> > Cc: Jani Nikula <jani.nikula at intel.com>
> > Cc: Imre Deak <imre.deak at intel.com>
> > Cc: Animesh Manna <animesh.manna at intel.com>
> > Signed-off-by: Anshuman Gupta <anshuman.gupta at intel.com>
> > ---
> > .../drm/i915/display/intel_display_power.c | 111 ++++++++++++++----
> > .../drm/i915/display/intel_display_power.h | 3 +
> > drivers/gpu/drm/i915/i915_drv.h | 1 +
> > 3 files changed, 95 insertions(+), 20 deletions(-)
> >
> > diff --git a/drivers/gpu/drm/i915/display/intel_display_power.c b/drivers/gpu/drm/i915/display/intel_display_power.c
> > index 496fa1b53ffb..83b10f61ee42 100644
> > --- a/drivers/gpu/drm/i915/display/intel_display_power.c
> > +++ b/drivers/gpu/drm/i915/display/intel_display_power.c
> > @@ -772,6 +772,29 @@ static void gen9_set_dc_state(struct drm_i915_private *dev_priv, u32 state)
> > dev_priv->csr.dc_state = val & mask;
> > }
> >
> > +static void tgl_allow_dc3co(struct drm_i915_private *dev_priv)
>
> Should be tgl_enable_dc3co(), to match the rest of DC state helpers.
>
> > +{
> > + if (!dev_priv->psr.sink_psr2_support)
> > + return;
>
> PSR knows when to enable DC3co, so no need to double-check that here.
>
> > +
> > + if (dev_priv->csr.allowed_dc_mask & DC_STATE_EN_DC3CO)
>
> This check is out-of-place wrt. the same checks for other DC states.
>
> > + gen9_set_dc_state(dev_priv, DC_STATE_EN_DC3CO);
> > +}
> > +
> > +static void tgl_disallow_dc3co(struct drm_i915_private *dev_priv)
> > +{
> > + u32 val;
> > +
> > + val = I915_READ(DC_STATE_EN);
> > + val &= ~DC_STATE_DC3CO_STATUS;
> > + I915_WRITE(DC_STATE_EN, val);
> > + gen9_set_dc_state(dev_priv, DC_STATE_DISABLE);
> > + /*
> > + * Delay of 200us DC3CO Exit time B.Spec 49196
> > + */
> > + udelay(200);
> > +}
> > +
> > static void bxt_enable_dc9(struct drm_i915_private *dev_priv)
> > {
> > assert_can_enable_dc9(dev_priv);
> > @@ -939,7 +962,8 @@ static void bxt_verify_ddi_phy_power_wells(struct drm_i915_private *dev_priv)
> > static bool gen9_dc_off_power_well_enabled(struct drm_i915_private *dev_priv,
> > struct i915_power_well *power_well)
> > {
> > - return (I915_READ(DC_STATE_EN) & DC_STATE_EN_UPTO_DC5_DC6_MASK) == 0;
> > + return ((I915_READ(DC_STATE_EN) & DC_STATE_EN_DC3CO) == 0 &&
> > + (I915_READ(DC_STATE_EN) & DC_STATE_EN_UPTO_DC5_DC6_MASK) == 0);
> > }
> >
> > static void gen9_assert_dbuf_enabled(struct drm_i915_private *dev_priv)
> > @@ -955,24 +979,32 @@ static void gen9_disable_dc_states(struct drm_i915_private *dev_priv)
> > {
> > struct intel_cdclk_state cdclk_state = {};
> >
> > - gen9_set_dc_state(dev_priv, DC_STATE_DISABLE);
> > + if (dev_priv->csr.max_dc_state & DC_STATE_EN_DC3CO) {
> > + tgl_disallow_dc3co(dev_priv);
> > + } else {
>
> With an early return you can avoid the extra diff and make reviewing
> easier.
>
> > + gen9_set_dc_state(dev_priv, DC_STATE_DISABLE);
> >
> > - dev_priv->display.get_cdclk(dev_priv, &cdclk_state);
> > - /* Can't read out voltage_level so can't use intel_cdclk_changed() */
> > - WARN_ON(intel_cdclk_needs_modeset(&dev_priv->cdclk.hw, &cdclk_state));
> > + dev_priv->display.get_cdclk(dev_priv, &cdclk_state);
> > + /*
> > + * Can't read out voltage_level so can't use
> > + * intel_cdclk_changed()
> > + */
> > + WARN_ON(intel_cdclk_needs_modeset(&dev_priv->cdclk.hw,
> > + &cdclk_state));
> >
> > - gen9_assert_dbuf_enabled(dev_priv);
> > + gen9_assert_dbuf_enabled(dev_priv);
> >
> > - if (IS_GEN9_LP(dev_priv))
> > - bxt_verify_ddi_phy_power_wells(dev_priv);
> > + if (IS_GEN9_LP(dev_priv))
> > + bxt_verify_ddi_phy_power_wells(dev_priv);
> >
> > - if (INTEL_GEN(dev_priv) >= 11)
> > - /*
> > - * DMC retains HW context only for port A, the other combo
> > - * PHY's HW context for port B is lost after DC transitions,
> > - * so we need to restore it manually.
> > - */
> > - intel_combo_phy_init(dev_priv);
> > + if (INTEL_GEN(dev_priv) >= 11)
> > + /*
> > + * DMC retains HW context only for port A, the other
> > + * combo PHY's HW context for port B is lost after
> > + * DC transitions, so we need to restore it manually.
> > + */
> > + intel_combo_phy_init(dev_priv);
> > + }
> > }
> >
> > static void gen9_dc_off_power_well_enable(struct drm_i915_private *dev_priv,
> > @@ -987,10 +1019,48 @@ static void gen9_dc_off_power_well_disable(struct drm_i915_private *dev_priv,
> > if (!dev_priv->csr.dmc_payload)
> > return;
> >
> > - if (dev_priv->csr.allowed_dc_mask & DC_STATE_EN_UPTO_DC6)
> > - skl_enable_dc6(dev_priv);
> > - else if (dev_priv->csr.allowed_dc_mask & DC_STATE_EN_UPTO_DC5)
> > - gen9_enable_dc5(dev_priv);
> > + if (dev_priv->csr.max_dc_state & DC_STATE_EN_DC3CO) {
>
> target_dc_state would be a better name and it shold be an exact state
> instead of a mask.
>
> > + tgl_allow_dc3co(dev_priv);
> > + } else if (dev_priv->csr.max_dc_state & DC_STATE_EN_UPTO_DC5_DC6_MASK) {
> > + if (dev_priv->csr.allowed_dc_mask & DC_STATE_EN_UPTO_DC6)
>
> We should make these checks more uniform, by checking here only
> target_dc_state.
If i understand correctly this comment is about to have a uniform
condition target_dc_state in order to decide which DC state we
interested. Then we may require to check allowed_dc_mask accordingly in
tgl_set_target_dc_state(), but that would not be fitting for
older platforms.
What is your input on this? please correct me if i am wrong here.
>
> > + skl_enable_dc6(dev_priv);
> > + else if (dev_priv->csr.allowed_dc_mask & DC_STATE_EN_UPTO_DC5)
> > + gen9_enable_dc5(dev_priv);
> > + }
> > +}
> > +
> > +void tgl_set_target_dc_state(struct drm_i915_private *dev_priv, u32 state,
> > + bool psr2_deep_sleep)
>
> psr2_deep_sleep is a PSR internal stuff, while PSR is using the power
> domains framework, so let's keep the deep sleep programming in the PSR
> code.
I wanted not to enable/disable psr2 deep sleep for each flip flush request.
Though PSR2 deep sleep is PSR internal stuff but it is H/W has tied with DC3CO.
That was intention to keep psr2_deep sleep enable/diable here (Doing in Patch 6).
In order to switch between DC3CO and DC5 we need to call
tgl_psr2_deep_sleep_disable/enable() from outside of PSR code.
please correct me if i am wrong here.
Thanks ,
Anshuman Gupta.
>
> > +{
> > + struct i915_power_well *power_well;
> > + bool dc_off_enabled;
> > + struct i915_power_domains *power_domains = &dev_priv->power_domains;
> > +
> > + mutex_lock(&power_domains->lock);
> > + power_well = lookup_power_well(dev_priv, TGL_DISP_DC_OFF);
> > +
> > + if (!power_well)
> > + goto unlock;
> > +
> > + dc_off_enabled = power_well->desc->ops->is_enabled(dev_priv,
> > + power_well);
> > + if (state == dev_priv->csr.max_dc_state)
> > + goto unlock;
> > +
> > + if (!dc_off_enabled) {
> > + /*
> > + * Need to disable the DC off power well to
> > + * effect target DC state.
> > + */
> > + power_well->desc->ops->enable(dev_priv, power_well);
> > + dev_priv->csr.max_dc_state = state;
> > + power_well->desc->ops->disable(dev_priv, power_well);
> > + goto unlock;
> > + }
> > + dev_priv->csr.max_dc_state = state;
> > +
> > +unlock:
> > + mutex_unlock(&power_domains->lock);
> > }
> >
> > static void i9xx_power_well_sync_hw_noop(struct drm_i915_private *dev_priv,
> > @@ -3610,7 +3680,7 @@ static const struct i915_power_well_desc tgl_power_wells[] = {
> > .name = "DC off",
> > .domains = TGL_DISPLAY_DC_OFF_POWER_DOMAINS,
> > .ops = &gen9_dc_off_power_well_ops,
> > - .id = DISP_PW_ID_NONE,
> > + .id = TGL_DISP_DC_OFF,
>
> Let's assign this ID on all platforms for consistency.
IMO that should be done in different patch, as tgl_set_target_dc_state() is
only consumer of this id ?
What is your opinion on this ?
>
> > },
> > {
> > .name = "power well 2",
> > @@ -4039,6 +4109,7 @@ int intel_power_domains_init(struct drm_i915_private *dev_priv)
> > dev_priv->csr.allowed_dc_mask =
> > get_allowed_dc_mask(dev_priv, i915_modparams.enable_dc);
> >
> > + dev_priv->csr.max_dc_state = DC_STATE_EN_UPTO_DC5_DC6_MASK;
> > BUILD_BUG_ON(POWER_DOMAIN_NUM > 64);
> >
> > mutex_init(&power_domains->lock);
> > diff --git a/drivers/gpu/drm/i915/display/intel_display_power.h b/drivers/gpu/drm/i915/display/intel_display_power.h
> > index 737b5def7fc6..69ebde992342 100644
> > --- a/drivers/gpu/drm/i915/display/intel_display_power.h
> > +++ b/drivers/gpu/drm/i915/display/intel_display_power.h
> > @@ -100,6 +100,7 @@ enum i915_power_well_id {
> > SKL_DISP_PW_MISC_IO,
> > SKL_DISP_PW_1,
> > SKL_DISP_PW_2,
> > + TGL_DISP_DC_OFF,
> > };
> >
> > #define POWER_DOMAIN_PIPE(pipe) ((pipe) + POWER_DOMAIN_PIPE_A)
> > @@ -256,6 +257,8 @@ void intel_display_power_suspend_late(struct drm_i915_private *i915);
> > void intel_display_power_resume_early(struct drm_i915_private *i915);
> > void intel_display_power_suspend(struct drm_i915_private *i915);
> > void intel_display_power_resume(struct drm_i915_private *i915);
> > +void tgl_set_target_dc_state(struct drm_i915_private *dev_priv, u32 state,
> > + bool psr2_deep_sleep);
> >
> > const char *
> > intel_display_power_domain_str(enum intel_display_power_domain domain);
> > diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
> > index db7480831e52..999da5d2da0b 100644
> > --- a/drivers/gpu/drm/i915/i915_drv.h
> > +++ b/drivers/gpu/drm/i915/i915_drv.h
> > @@ -336,6 +336,7 @@ struct intel_csr {
> > i915_reg_t mmioaddr[20];
> > u32 mmiodata[20];
> > u32 dc_state;
> > + u32 max_dc_state;
> > u32 allowed_dc_mask;
> > intel_wakeref_t wakeref;
> > };
> > --
> > 2.21.0
> >
More information about the Intel-gfx
mailing list