[Intel-gfx] [PATCH 13/17] drm/i915/icl: Enable 2nd DBuf slice only when needed
James Ausmus
james.ausmus at intel.com
Thu Jan 25 22:56:29 UTC 2018
On Tue, Jan 23, 2018 at 05:05:32PM -0200, Paulo Zanoni wrote:
> From: Mahesh Kumar <mahesh1.kumar at intel.com>
>
> ICL has two slices of DBuf, each slice of size 1024 blocks.
> We should not always enable slice-2. It should be enabled only if
> display total required BW is > 12GBps OR more than 1 pipes are enabled.
>
> Changes since V1:
> - typecast total_data_rate to u64 before multiplication to solve any
> possible overflow (Rodrigo)
> - fix where skl_wm_get_hw_state was memsetting ddb, resulting
> enabled_slices to become zero
> - Fix the logic of calculating ddb_size
> Changes since V2:
> - If no-crtc is part of commit required_slices will have value "0",
> don't try to disable DBuf slice.
>
> Signed-off-by: Mahesh Kumar <mahesh1.kumar at intel.com>
> ---
> drivers/gpu/drm/i915/intel_display.c | 12 +++++++
> drivers/gpu/drm/i915/intel_drv.h | 6 ++++
> drivers/gpu/drm/i915/intel_pm.c | 64 +++++++++++++++++++++++++--------
> drivers/gpu/drm/i915/intel_runtime_pm.c | 47 +++++++++++++++++++++---
> 4 files changed, 110 insertions(+), 19 deletions(-)
>
> diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
> index bad3b112ac3e..3eb2359c221b 100644
> --- a/drivers/gpu/drm/i915/intel_display.c
> +++ b/drivers/gpu/drm/i915/intel_display.c
> @@ -12117,6 +12117,8 @@ static void skl_update_crtcs(struct drm_atomic_state *state)
> bool progress;
> enum pipe pipe;
> int i;
> + uint8_t hw_enabled_slices = dev_priv->wm.skl_hw.ddb.enabled_slices;
> + uint8_t required_slices = intel_state->wm_results.ddb.enabled_slices;
>
> const struct skl_ddb_entry *entries[I915_MAX_PIPES] = {};
>
> @@ -12125,6 +12127,11 @@ static void skl_update_crtcs(struct drm_atomic_state *state)
> if (new_crtc_state->active)
> entries[i] = &to_intel_crtc_state(old_crtc_state)->wm.skl.ddb;
>
> + /* If 2nd DBuf slice required, enable it here */
> + if (INTEL_GEN(dev_priv) >= 11 && required_slices &&
> + required_slices > hw_enabled_slices)
> + icl_dbuf_slices_update(dev_priv, required_slices);
> +
> /*
> * Whenever the number of active pipes changes, we need to make sure we
> * update the pipes in the right order so that their ddb allocations
> @@ -12175,6 +12182,11 @@ static void skl_update_crtcs(struct drm_atomic_state *state)
> progress = true;
> }
> } while (progress);
> +
> + /* If 2nd DBuf slice is no more required disable it */
> + if (INTEL_GEN(dev_priv) >= 11 && required_slices &&
> + required_slices < hw_enabled_slices)
> + icl_dbuf_slices_update(dev_priv, required_slices);
> }
>
> static void intel_atomic_helper_free_state(struct drm_i915_private *dev_priv)
> diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h
> index c5d6092aca41..d4639a161fe3 100644
> --- a/drivers/gpu/drm/i915/intel_drv.h
> +++ b/drivers/gpu/drm/i915/intel_drv.h
> @@ -140,6 +140,10 @@
> #define KHz(x) (1000 * (x))
> #define MHz(x) KHz(1000 * (x))
>
> +#define KBps(x) (1000 * (x))
> +#define MBps(x) KBps(1000 * (x))
> +#define GBps(x) ((uint64_t) 1000 * MBps((x)))
> +
> /*
> * Display related stuff
> */
> @@ -1890,6 +1894,8 @@ bool intel_display_power_get_if_enabled(struct drm_i915_private *dev_priv,
> enum intel_display_power_domain domain);
> void intel_display_power_put(struct drm_i915_private *dev_priv,
> enum intel_display_power_domain domain);
> +void icl_dbuf_slices_update(struct drm_i915_private *dev_priv,
> + uint8_t req_slices);
>
> static inline void
> assert_rpm_device_not_suspended(struct drm_i915_private *dev_priv)
> diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c
> index e8d98857c208..d4cd631377da 100644
> --- a/drivers/gpu/drm/i915/intel_pm.c
> +++ b/drivers/gpu/drm/i915/intel_pm.c
> @@ -3767,9 +3767,42 @@ bool intel_can_enable_sagv(struct drm_atomic_state *state)
> return true;
> }
>
> +static unsigned int intel_get_ddb_size(struct drm_i915_private *dev_priv,
> + const struct intel_crtc_state *cstate,
> + const unsigned int total_data_rate,
> + const int num_active,
> + struct skl_ddb_allocation *ddb)
> +{
> + const struct drm_display_mode *adjusted_mode;
> + u64 total_data_bw;
> + u16 ddb_size = INTEL_INFO(dev_priv)->ddb_size;
> +
> + WARN_ON(ddb_size == 0);
> +
> + if (INTEL_GEN(dev_priv) < 11)
> + return ddb_size - 4; /* 4 blocks for bypass path allocation */
> +
> + adjusted_mode = &cstate->base.adjusted_mode;
> + total_data_bw = (u64)total_data_rate * drm_mode_vrefresh(adjusted_mode);
> +
> + /*
> + * 12GB/s is maximum BW supported by single DBuf slice.
> + */
> + if (total_data_bw >= GBps(12) || num_active > 1)
> + ddb->enabled_slices = 2;
> + else {
> + ddb->enabled_slices = 1;
> + ddb_size /= 2;
> + }
> +
> + return ddb_size;
> +}
> +
> static void
> skl_ddb_get_pipe_allocation_limits(struct drm_device *dev,
> const struct intel_crtc_state *cstate,
> + const unsigned int total_data_rate,
> + struct skl_ddb_allocation *ddb,
> struct skl_ddb_entry *alloc, /* out */
> int *num_active /* out */)
> {
> @@ -3792,11 +3825,8 @@ skl_ddb_get_pipe_allocation_limits(struct drm_device *dev,
> else
> *num_active = hweight32(dev_priv->active_crtcs);
>
> - ddb_size = INTEL_INFO(dev_priv)->ddb_size;
> - WARN_ON(ddb_size == 0);
> -
> - if (INTEL_GEN(dev_priv) < 11)
> - ddb_size -= 4; /* 4 blocks for bypass path allocation */
> + ddb_size = intel_get_ddb_size(dev_priv, cstate, total_data_rate,
> + *num_active, ddb);
>
> /*
> * If the state doesn't change the active CRTC's, then there's
> @@ -4235,7 +4265,14 @@ skl_allocate_pipe_ddb(struct intel_crtc_state *cstate,
> return 0;
> }
>
> - skl_ddb_get_pipe_allocation_limits(dev, cstate, alloc, &num_active);
> + total_data_rate = skl_get_total_relative_data_rate(cstate,
> + plane_data_rate,
> + plane_y_data_rate);
> + if (total_data_rate == 0)
> + return 0;
> +
> + skl_ddb_get_pipe_allocation_limits(dev, cstate, total_data_rate, ddb,
> + alloc, &num_active);
> alloc_size = skl_ddb_entry_size(alloc);
> if (alloc_size == 0)
> return 0;
> @@ -4270,12 +4307,6 @@ skl_allocate_pipe_ddb(struct intel_crtc_state *cstate,
> *
> * FIXME: we may not allocate every single block here.
> */
> - total_data_rate = skl_get_total_relative_data_rate(cstate,
> - plane_data_rate,
> - plane_y_data_rate);
> - if (total_data_rate == 0)
> - return 0;
> -
> start = alloc->start;
> for_each_plane_id_on_crtc(intel_crtc, plane_id) {
> unsigned int data_rate, y_data_rate;
> @@ -5068,7 +5099,6 @@ skl_copy_wm_for_pipe(struct skl_wm_values *dst,
> sizeof(dst->ddb.y_plane[pipe]));
> memcpy(dst->ddb.plane[pipe], src->ddb.plane[pipe],
> sizeof(dst->ddb.plane[pipe]));
> - dst->ddb.enabled_slices = src->ddb.enabled_slices;
> }
>
> static void
> @@ -5381,8 +5411,12 @@ void skl_wm_get_hw_state(struct drm_device *dev)
> /* Fully recompute DDB on first atomic commit */
> dev_priv->wm.distrust_bios_wm = true;
> } else {
> - /* Easy/common case; just sanitize DDB now if everything off */
> - memset(ddb, 0, sizeof(*ddb));
> + /*
> + * Easy/common case; just sanitize DDB now if everything off
> + * Keep dbuf slice info intact
> + */
> + memset(ddb->plane, 0, sizeof(ddb->plane));
> + memset(ddb->y_plane, 0, sizeof(ddb->y_plane));
> }
> }
>
> diff --git a/drivers/gpu/drm/i915/intel_runtime_pm.c b/drivers/gpu/drm/i915/intel_runtime_pm.c
> index 13c8dad95b84..a67a57738740 100644
> --- a/drivers/gpu/drm/i915/intel_runtime_pm.c
> +++ b/drivers/gpu/drm/i915/intel_runtime_pm.c
> @@ -2610,10 +2610,49 @@ static void gen9_dbuf_disable(struct drm_i915_private *dev_priv)
> DRM_ERROR("DBuf power disable timeout!\n");
> }
>
> -/*
> - * TODO: we shouldn't always enable DBUF_CTL_S2, we should only enable it when
> - * needed and keep it disabled as much as possible.
> - */
> +static uint8_t intel_dbuf_max_slices(struct drm_i915_private *dev_priv)
> +{
> + if (INTEL_GEN(dev_priv) < 11)
> + return 1;
> + return 2;
> +}
> +
> +void icl_dbuf_slices_update(struct drm_i915_private *dev_priv,
> + uint8_t req_slices)
> +{
> + uint8_t hw_enabled_slices = dev_priv->wm.skl_hw.ddb.enabled_slices;
> + u32 val;
> +
> + if (req_slices > intel_dbuf_max_slices(dev_priv)) {
> + DRM_ERROR("Invalid number of dbuf slices requested\n");
> + return;
> + }
> +
> + if (req_slices == hw_enabled_slices)
> + return;
> +
> + val = I915_READ(DBUF_CTL_S2);
> + if (req_slices > hw_enabled_slices) {
> + I915_WRITE(DBUF_CTL_S2, val | DBUF_POWER_REQUEST);
> + POSTING_READ(DBUF_CTL_S2);
> +
> + udelay(10);
> + if (!(I915_READ(DBUF_CTL_S2) & DBUF_POWER_STATE))
> + DRM_ERROR("DBuf power enable timeout\n");
> + else
> + dev_priv->wm.skl_hw.ddb.enabled_slices = req_slices;
> + } else {
> + I915_WRITE(DBUF_CTL_S2, val & ~DBUF_POWER_REQUEST);
> + POSTING_READ(DBUF_CTL_S2);
> +
> + udelay(10);
> + if (I915_READ(DBUF_CTL_S2) & DBUF_POWER_STATE)
> + DRM_ERROR("DBuf power disable timeout!\n");
> + else
> + dev_priv->wm.skl_hw.ddb.enabled_slices = req_slices;
It feels a little ugly to be duplicating the Write->Posting
Read->Wait->Read->Error pattern across this function and
icl_dbuf_enable/disable below - not to mention that we have the same
pattern up above in gen9_dbuf_enable/disable. Maybe we should refactor
the gen9 versions in to helpers that take a slice number, and then reuse
those for ICL as well?
> + }
> +}
> +
> static void icl_dbuf_enable(struct drm_i915_private *dev_priv)
> {
> I915_WRITE(DBUF_CTL_S1, I915_READ(DBUF_CTL_S1) | DBUF_POWER_REQUEST);
> --
> 2.14.3
>
> _______________________________________________
> Intel-gfx mailing list
> Intel-gfx at lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/intel-gfx
More information about the Intel-gfx
mailing list