[Intel-gfx] [PATCH v18 7/8] drm/i915: Restrict qgv points which don't have enough bandwidth.
Ville Syrjälä
ville.syrjala at linux.intel.com
Thu Feb 27 16:20:14 UTC 2020
On Tue, Feb 25, 2020 at 05:00:43PM +0200, Stanislav Lisovskiy wrote:
> According to BSpec 53998, we should try to
> restrict qgv points, which can't provide
> enough bandwidth for desired display configuration.
>
> Currently we are just comparing against all of
> those and take minimum(worst case).
>
> v2: Fixed wrong PCode reply mask, removed hardcoded
> values.
>
> v3: Forbid simultaneous legacy SAGV PCode requests and
> restricting qgv points. Put the actual restriction
> to commit function, added serialization(thanks to Ville)
> to prevent commit being applied out of order in case of
> nonblocking and/or nomodeset commits.
>
> v4:
> - Minor code refactoring, fixed few typos(thanks to James Ausmus)
> - Change the naming of qgv point
> masking/unmasking functions(James Ausmus).
> - Simplify the masking/unmasking operation itself,
> as we don't need to mask only single point per request(James Ausmus)
> - Reject and stick to highest bandwidth point if SAGV
> can't be enabled(BSpec)
>
> v5:
> - Add new mailbox reply codes, which seems to happen during boot
> time for TGL and indicate that QGV setting is not yet available.
>
> v6:
> - Increase number of supported QGV points to be in sync with BSpec.
>
> v7: - Rebased and resolved conflict to fix build failure.
> - Fix NUM_QGV_POINTS to 8 and moved that to header file(James Ausmus)
>
> v8: - Don't report an error if we can't restrict qgv points, as SAGV
> can be disabled by BIOS, which is completely legal. So don't
> make CI panic. Instead if we detect that there is only 1 QGV
> point accessible just analyze if we can fit the required bandwidth
> requirements, but no need in restricting.
>
> v9: - Fix wrong QGV transition if we have 0 planes and no SAGV
> simultaneously.
>
> v10: - Fix CDCLK corruption, because of global state getting serialized
> without modeset, which caused copying of non-calculated cdclk
> to be copied to dev_priv(thanks to Ville for the hint).
>
> v11: - Remove unneeded headers and spaces(Matthew Roper)
> - Remove unneeded intel_qgv_info qi struct from bw check and zero
> out the needed one(Matthew Roper)
> - Changed QGV error message to have more clear meaning(Matthew Roper)
> - Use state->modeset_set instead of any_ms(Matthew Roper)
> - Moved NUM_SAGV_POINTS from i915_reg.h to i915_drv.h where it's used
> - Keep using crtc_state->hw.active instead of .enable(Matthew Roper)
> - Moved unrelated changes to other patch(using latency as parameter
> for plane wm calculation, moved to SAGV refactoring patch)
>
> v12: - Fix rebase conflict with own temporary SAGV/QGV fix.
> - Remove unnecessary mask being zero check when unmasking
> qgv points as this is completely legal(Matt Roper)
> - Check if we are setting the same mask as already being set
> in hardware to prevent error from PCode.
> - Fix error message when restricting/unrestricting qgv points
> to "mask/unmask" which sounds more accurate(Matt Roper)
> - Move sagv status setting to icl_get_bw_info from atomic check
> as this should be calculated only once.(Matt Roper)
> - Edited comments for the case when we can't enable SAGV and
> use only 1 QGV point with highest bandwidth to be more
> understandable.(Matt Roper)
>
> v13: - Moved max_data_rate in bw check to closer scope(Ville Syrjälä)
> - Changed comment for zero new_mask in qgv points masking function
> to better reflect reality(Ville Syrjälä)
> - Simplified bit mask operation in qgv points masking function
> (Ville Syrjälä)
> - Moved intel_qgv_points_mask closer to gen11 SAGV disabling,
> however this still can't be under modeset condition(Ville Syrjälä)
> - Packed qgv_points_mask as u8 and moved closer to pipe_sagv_mask
> (Ville Syrjälä)
> - Extracted PCode changes to separate patch.(Ville Syrjälä)
> - Now treat num_planes 0 same as 1 to avoid confusion and
> returning max_bw as 0, which would prevent choosing QGV
> point having max bandwidth in case if SAGV is not allowed,
> as per BSpec(Ville Syrjälä)
> - Do the actual qgv_points_mask swap in the same place as
> all other global state parts like cdclk are swapped.
> In the next patch, this all will be moved to bw state as
> global state, once new global state patch series from Ville
> lands
>
> v14: - Now using global state to serialize access to qgv points
> - Added global state locking back, otherwise we seem to read
> bw state in a wrong way.
>
> v15: - Added TODO comment for near atomic global state locking in
> bw code.
>
> v16: - Fixed intel_atomic_bw_* functions to be intel_bw_* as discussed
> with Jani Nikula.
> - Take bw_state_changed flag into use.
>
> Signed-off-by: Stanislav Lisovskiy <stanislav.lisovskiy at intel.com>
> Cc: Ville Syrjälä <ville.syrjala at intel.com>
> Cc: James Ausmus <james.ausmus at intel.com>
> ---
> drivers/gpu/drm/i915/display/intel_bw.c | 184 +++++++++++++------
> drivers/gpu/drm/i915/display/intel_bw.h | 9 +
> drivers/gpu/drm/i915/display/intel_display.c | 125 ++++++++++++-
> drivers/gpu/drm/i915/i915_drv.h | 3 +
> 4 files changed, 255 insertions(+), 66 deletions(-)
>
> diff --git a/drivers/gpu/drm/i915/display/intel_bw.c b/drivers/gpu/drm/i915/display/intel_bw.c
> index d5be603b8b03..4986a5464700 100644
> --- a/drivers/gpu/drm/i915/display/intel_bw.c
> +++ b/drivers/gpu/drm/i915/display/intel_bw.c
> @@ -8,6 +8,9 @@
> #include "intel_bw.h"
> #include "intel_display_types.h"
> #include "intel_sideband.h"
> +#include "intel_atomic.h"
> +#include "intel_pm.h"
> +
>
> /* Parameters for Qclk Geyserville (QGV) */
> struct intel_qgv_point {
> @@ -113,6 +116,26 @@ static int icl_pcode_read_qgv_point_info(struct drm_i915_private *dev_priv,
> return 0;
> }
>
> +int icl_pcode_restrict_qgv_points(struct drm_i915_private *dev_priv,
> + u32 points_mask)
> +{
> + int ret;
> +
> + /* bspec says to keep retrying for at least 1 ms */
> + ret = skl_pcode_request(dev_priv, ICL_PCODE_SAGV_DE_MEM_SS_CONFIG,
> + points_mask,
> + GEN11_PCODE_POINTS_RESTRICTED_MASK,
> + GEN11_PCODE_POINTS_RESTRICTED,
> + 1);
> +
> + if (ret < 0) {
> + DRM_ERROR("Failed to disable qgv points (%d)\n", ret);
> + return ret;
> + }
> +
> + return 0;
> +}
> +
> static int icl_get_qgv_points(struct drm_i915_private *dev_priv,
> struct intel_qgv_info *qi)
> {
> @@ -240,6 +263,16 @@ static int icl_get_bw_info(struct drm_i915_private *dev_priv, const struct intel
> break;
> }
>
> + /*
> + * In case if SAGV is disabled in BIOS, we always get 1
> + * SAGV point, but we can't send PCode commands to restrict it
> + * as it will fail and pointless anyway.
> + */
> + if (qi.num_points == 1)
> + dev_priv->sagv_status = I915_SAGV_NOT_CONTROLLED;
> + else
> + dev_priv->sagv_status = I915_SAGV_ENABLED;
> +
> return 0;
> }
>
> @@ -259,7 +292,7 @@ static unsigned int icl_max_bw(struct drm_i915_private *dev_priv,
> if (qgv_point >= bi->num_qgv_points)
> return UINT_MAX;
>
> - if (num_planes >= bi->num_planes)
> + if (num_planes >= bi->num_planes || !num_planes)
> return bi->deratedbw[qgv_point];
> }
>
> @@ -277,34 +310,6 @@ void intel_bw_init_hw(struct drm_i915_private *dev_priv)
> icl_get_bw_info(dev_priv, &icl_sa_info);
> }
>
> -static unsigned int intel_max_data_rate(struct drm_i915_private *dev_priv,
> - int num_planes)
> -{
> - if (INTEL_GEN(dev_priv) >= 11) {
> - /*
> - * Any bw group has same amount of QGV points
> - */
> - const struct intel_bw_info *bi =
> - &dev_priv->max_bw[0];
> - unsigned int min_bw = UINT_MAX;
> - int i;
> -
> - /*
> - * FIXME with SAGV disabled maybe we can assume
> - * point 1 will always be used? Seems to match
> - * the behaviour observed in the wild.
> - */
> - for (i = 0; i < bi->num_qgv_points; i++) {
> - unsigned int bw = icl_max_bw(dev_priv, num_planes, i);
> -
> - min_bw = min(bw, min_bw);
> - }
> - return min_bw;
> - } else {
> - return UINT_MAX;
> - }
> -}
> -
> static unsigned int intel_bw_crtc_num_active_planes(const struct intel_crtc_state *crtc_state)
> {
> /*
> @@ -418,11 +423,16 @@ int intel_bw_atomic_check(struct intel_atomic_state *state)
> {
> struct drm_i915_private *dev_priv = to_i915(state->base.dev);
> struct intel_crtc_state *new_crtc_state, *old_crtc_state;
> - struct intel_bw_state *bw_state = NULL;
> - unsigned int data_rate, max_data_rate;
> + struct intel_bw_state *new_bw_state = NULL;
> + struct intel_bw_state *old_bw_state = NULL;
> + unsigned int data_rate;
> unsigned int num_active_planes;
> struct intel_crtc *crtc;
> int i, ret;
> + u32 allowed_points = 0;
> + unsigned int max_bw_point = 0, max_bw = 0;
> + unsigned int num_qgv_points = dev_priv->max_bw[0].num_qgv_points;
> + u32 mask = (1 << num_qgv_points) - 1;
>
> /*
> * For earlier Gens let's consider bandwidth changed if ddb requirements,
> @@ -430,11 +440,11 @@ int intel_bw_atomic_check(struct intel_atomic_state *state)
> */
> if (INTEL_GEN(dev_priv) < 11) {
> if (state->ddb_state_changed) {
> - bw_state = intel_bw_get_state(state);
> - if (IS_ERR(bw_state))
> - return PTR_ERR(bw_state);
> + new_bw_state = intel_bw_get_state(state);
> + if (IS_ERR(new_bw_state))
> + return PTR_ERR(new_bw_state);
Lot's of rename noise in the patch. Can't really see what's going on
unless those are split out.
>
> - ret = intel_atomic_lock_global_state(&bw_state->base);
> + ret = intel_atomic_lock_global_state(&new_bw_state->base);
> if (ret)
> return ret;
>
> @@ -465,45 +475,107 @@ int intel_bw_atomic_check(struct intel_atomic_state *state)
> old_active_planes == new_active_planes)
> continue;
>
> - bw_state = intel_bw_get_state(state);
> - if (IS_ERR(bw_state))
> - return PTR_ERR(bw_state);
> + new_bw_state = intel_bw_get_state(state);
> + if (IS_ERR(new_bw_state))
> + return PTR_ERR(new_bw_state);
>
> - bw_state->data_rate[crtc->pipe] = new_data_rate;
> - bw_state->num_active_planes[crtc->pipe] = new_active_planes;
> + new_bw_state->data_rate[crtc->pipe] = new_data_rate;
> + new_bw_state->num_active_planes[crtc->pipe] = new_active_planes;
>
> drm_dbg_kms(&dev_priv->drm,
> "pipe %c data rate %u num active planes %u\n",
> pipe_name(crtc->pipe),
> - bw_state->data_rate[crtc->pipe],
> - bw_state->num_active_planes[crtc->pipe]);
> + new_bw_state->data_rate[crtc->pipe],
> + new_bw_state->num_active_planes[crtc->pipe]);
> }
>
> - if (!bw_state)
> + if (!new_bw_state)
> return 0;
>
> - ret = intel_atomic_lock_global_state(&bw_state->base);
> - if (ret)
> + ret = intel_atomic_lock_global_state(&new_bw_state->base);
> + if (ret) {
> + DRM_DEBUG_KMS("Could not lock global state\n");
> return ret;
> + }
>
> - DRM_DEBUG_KMS("Marking bw state changed for atomic state %p\n", state);
> -
> - state->bw_state_changed = true;
> + data_rate = intel_bw_data_rate(dev_priv, new_bw_state);
> + data_rate = DIV_ROUND_UP(data_rate, 1000);
>
> - data_rate = intel_bw_data_rate(dev_priv, bw_state);
> - num_active_planes = intel_bw_num_active_planes(dev_priv, bw_state);
> + num_active_planes = intel_bw_num_active_planes(dev_priv, new_bw_state);
>
> - max_data_rate = intel_max_data_rate(dev_priv, num_active_planes);
> + for (i = 0; i < num_qgv_points; i++) {
> + unsigned int max_data_rate;
>
> - data_rate = DIV_ROUND_UP(data_rate, 1000);
> + max_data_rate = icl_max_bw(dev_priv, num_active_planes, i);
> + /*
> + * We need to know which qgv point gives us
> + * maximum bandwidth in order to disable SAGV
> + * if we find that we exceed SAGV block time
> + * with watermarks. By that moment we already
> + * have those, as it is calculated earlier in
> + * intel_atomic_check,
> + */
> + if (max_data_rate > max_bw) {
> + max_bw_point = i;
> + max_bw = max_data_rate;
> + }
> + if (max_data_rate >= data_rate)
> + allowed_points |= BIT(i);
> + DRM_DEBUG_KMS("QGV point %d: max bw %d required %d\n",
> + i, max_data_rate, data_rate);
> + }
>
> - if (data_rate > max_data_rate) {
> - drm_dbg_kms(&dev_priv->drm,
> - "Bandwidth %u MB/s exceeds max available %d MB/s (%d active planes)\n",
> - data_rate, max_data_rate, num_active_planes);
> + /*
> + * BSpec states that we always should have at least one allowed point
> + * left, so if we couldn't - simply reject the configuration for obvious
> + * reasons.
> + */
> + if (allowed_points == 0) {
> + DRM_DEBUG_KMS("No QGV points provide sufficient memory"
> + " bandwidth for display configuration.\n");
> return -EINVAL;
> }
>
> + /*
> + * Leave only single point with highest bandwidth, if
> + * we can't enable SAGV due to the increased memory latency it may
> + * cause.
> + */
> + if (!intel_can_enable_sagv_for_state(state)) {
> + allowed_points = 1 << max_bw_point;
> + DRM_DEBUG_KMS("No SAGV, using single QGV point %d\n",
> + max_bw_point);
> + }
> + /*
> + * We store the ones which need to be masked as that is what PCode
> + * actually accepts as a parameter.
> + */
> + new_bw_state->qgv_points_mask = (~allowed_points) & mask;
> +
> + DRM_DEBUG_KMS("New state %p qgv mask %x\n",
> + state, new_bw_state->qgv_points_mask);
> +
> + old_bw_state = intel_bw_get_old_state(state);
> + if (IS_ERR(old_bw_state)) {
> + DRM_DEBUG_KMS("Could not get old bw state!\n");
> + return PTR_ERR(old_bw_state);
> + }
> +
> + /*
> + * If the actual mask had changed we need to make sure that
> + * the commits are serialized(in case this is a nomodeset, nonblocking)
> + */
> + if (new_bw_state->qgv_points_mask != old_bw_state->qgv_points_mask) {
> + ret = intel_atomic_serialize_global_state(&new_bw_state->base);
> + if (ret) {
> + DRM_DEBUG_KMS("Could not serialize global state\n");
> + return ret;
> + }
> +
> + DRM_DEBUG_KMS("Marking bw state changed for atomic state %p\n", state);
> + state->bw_state_changed = true;
> + }
> +
> return 0;
> }
>
> diff --git a/drivers/gpu/drm/i915/display/intel_bw.h b/drivers/gpu/drm/i915/display/intel_bw.h
> index c32b5285c12f..b3522389a181 100644
> --- a/drivers/gpu/drm/i915/display/intel_bw.h
> +++ b/drivers/gpu/drm/i915/display/intel_bw.h
> @@ -36,6 +36,13 @@ struct intel_bw_state {
> */
> bool can_sagv;
>
> + /*
> + * Current QGV points mask, which restricts
> + * some particular SAGV states, not to confuse
> + * with pipe_sagv_mask.
> + */
> + u8 qgv_points_mask;
> +
> unsigned int data_rate[I915_MAX_PIPES];
> u8 num_active_planes[I915_MAX_PIPES];
> };
> @@ -56,5 +63,7 @@ int intel_bw_init(struct drm_i915_private *dev_priv);
> int intel_bw_atomic_check(struct intel_atomic_state *state);
> void intel_bw_crtc_update(struct intel_bw_state *bw_state,
> const struct intel_crtc_state *crtc_state);
> +int icl_pcode_restrict_qgv_points(struct drm_i915_private *dev_priv,
> + u32 points_mask);
>
> #endif /* __INTEL_BW_H__ */
> diff --git a/drivers/gpu/drm/i915/display/intel_display.c b/drivers/gpu/drm/i915/display/intel_display.c
> index 6df836cbe0cd..cb1d10af88ce 100644
> --- a/drivers/gpu/drm/i915/display/intel_display.c
> +++ b/drivers/gpu/drm/i915/display/intel_display.c
> @@ -15504,6 +15504,105 @@ static void intel_atomic_cleanup_work(struct work_struct *work)
> intel_atomic_helper_free_state(i915);
> }
>
> +static void intel_qgv_points_mask(struct intel_atomic_state *state)
I think we should move these next to the other sagv code and just
call them somehting like intel_sagv_{pre,post}_plane_update() to
a) avoid the caller having to know anything about them, b) match
the pattern used for other similar things.
And I think the internals should just be something like:
intel_sagv_pre/post_plane_update()
{
new_bw_state = get_new_bw_state();
old_bw_state = get_old_bw_state();
if (!new_bw_state ||
new_bw_state->sagv_mask == old_bw_state->sagv_mask)
return;
do the pcode dance
}
I don't see a need for any more checks than than because that
would just mean that bw_check() is broken and needs to be fixed.
> +{
> + struct drm_device *dev = state->base.dev;
> + struct drm_i915_private *dev_priv = to_i915(dev);
> + int ret;
> + struct intel_bw_state *new_bw_state = NULL;
> + struct intel_bw_state *old_bw_state = NULL;
> + u32 new_mask = 0;
> + unsigned int num_qgv_points = dev_priv->max_bw[0].num_qgv_points;
> + unsigned int mask = (1 << num_qgv_points) - 1;
> +
> + new_bw_state = intel_bw_get_state(state);
> + if (IS_ERR(new_bw_state)) {
> + WARN(1, "Could not get new bw_state!\n");
> + return;
> + }
> +
> + old_bw_state = intel_bw_get_old_state(state);
> + if (IS_ERR(old_bw_state)) {
> + WARN(1, "Could not get old bw_state!\n");
> + return;
> + }
> +
> + new_mask = old_bw_state->qgv_points_mask | new_bw_state->qgv_points_mask;
> +
> + /*
> + * If new mask is zero - means there is nothing to mask,
> + * we can only unmask, which should be done in unmask.
> + */
> + if (!new_mask)
> + return;
> +
> + WARN_ON(new_mask == mask);
> +
> + /*
> + * Just return if we can't control SAGV or don't have it.
> + */
> + if (!intel_has_sagv(dev_priv))
> + return;
> +
> + /*
> + * Restrict required qgv points before updating the configuration.
> + * According to BSpec we can't mask and unmask qgv points at the same
> + * time. Also masking should be done before updating the configuration
> + * and unmasking afterwards.
> + */
> + ret = icl_pcode_restrict_qgv_points(dev_priv, new_mask);
> + if (ret < 0)
> + DRM_DEBUG_KMS("Could not mask required qgv points(%d)\n",
> + ret);
> +}
> +
> +static void intel_qgv_points_unmask(struct intel_atomic_state *state)
> +{
> + struct drm_device *dev = state->base.dev;
> + struct drm_i915_private *dev_priv = to_i915(dev);
> + int ret;
> + struct intel_bw_state *new_bw_state = NULL;
> + struct intel_bw_state *old_bw_state = NULL;
> + u32 new_mask = 0;
> +
> + new_bw_state = intel_bw_get_state(state);
> + if (IS_ERR(new_bw_state)) {
> + WARN(1, "Could not get new bw_state!\n");
> + return;
> + }
> +
> + old_bw_state = intel_bw_get_old_state(state);
> + if (IS_ERR(old_bw_state)) {
> + WARN(1, "Could not get new bw_state!\n");
> + return;
> + }
> +
> + new_mask = new_bw_state->qgv_points_mask;
> +
> + /*
> + * Just return if we can't control SAGV or don't have it.
> + */
> + if (!intel_has_sagv(dev_priv))
> + return;
> +
> + /*
> + * Nothing to unmask
> + */
> + if (new_mask == old_bw_state->qgv_points_mask)
> + return;
> +
> + /*
> + * Allow required qgv points after updating the configuration.
> + * According to BSpec we can't mask and unmask qgv points at the same
> + * time. Also masking should be done before updating the configuration
> + * and unmasking afterwards.
> + */
> + ret = icl_pcode_restrict_qgv_points(dev_priv, new_mask);
> + if (ret < 0)
> + DRM_DEBUG_KMS("Could not unmask required qgv points(%d)\n",
> + ret);
> +}
> +
> static void intel_atomic_commit_tail(struct intel_atomic_state *state)
> {
> struct drm_device *dev = state->base.dev;
> @@ -15537,6 +15636,15 @@ static void intel_atomic_commit_tail(struct intel_atomic_state *state)
> for_each_new_intel_crtc_in_state(state, crtc, new_crtc_state, i)
> crtc->config = new_crtc_state;
>
> + /*
> + * Now we need to check if SAGV needs to be disabled(i.e QGV points
> + * modified even, when no modeset is done(for example plane updates
> + * can now trigger that).
> + */
> + if ((INTEL_GEN(dev_priv) >= 11))
> + if (state->bw_state_changed)
> + intel_qgv_points_mask(state);
> +
> if (state->modeset) {
> drm_atomic_helper_update_legacy_modeset_state(dev, &state->base);
>
> @@ -15546,12 +15654,9 @@ static void intel_atomic_commit_tail(struct intel_atomic_state *state)
> * SKL workaround: bspec recommends we disable the SAGV when we
> * have more then one pipe enabled
> */
> - if (INTEL_GEN(dev_priv) < 11) {
> - if (state->bw_state_changed) {
> - if (!intel_can_enable_sagv(dev_priv))
> - intel_disable_sagv(dev_priv);
> - }
> - }
> + if ((INTEL_GEN(dev_priv) < 11))
> + if (!intel_can_enable_sagv(dev_priv))
> + intel_disable_sagv(dev_priv);
>
> intel_modeset_verify_disabled(dev_priv, state);
> }
> @@ -15652,10 +15757,10 @@ static void intel_atomic_commit_tail(struct intel_atomic_state *state)
> intel_verify_planes(state);
>
> if (INTEL_GEN(dev_priv) < 11) {
> - if (state->bw_state_changed) {
> - if (intel_can_enable_sagv(dev_priv))
> - intel_enable_sagv(dev_priv);
> - }
> + if (intel_can_enable_sagv(dev_priv))
> + intel_enable_sagv(dev_priv);
> + } else if (state->bw_state_changed) {
> + intel_qgv_points_unmask(state);
> }
>
> drm_atomic_helper_commit_hw_done(&state->base);
> diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
> index 4305ccc4c683..0a589700a071 100644
> --- a/drivers/gpu/drm/i915/i915_drv.h
> +++ b/drivers/gpu/drm/i915/i915_drv.h
> @@ -841,6 +841,9 @@ enum intel_pipe_crc_source {
> INTEL_PIPE_CRC_SOURCE_MAX,
> };
>
> +/* BSpec precisely defines this */
> +#define NUM_SAGV_POINTS 8
> +
> #define INTEL_PIPE_CRC_ENTRIES_NR 128
> struct intel_pipe_crc {
> spinlock_t lock;
> --
> 2.24.1.485.gad05a3d8e5
--
Ville Syrjälä
Intel
More information about the Intel-gfx
mailing list