Message ID | 20220823202449.83727-1-matthew.d.roper@intel.com (mailing list archive) |
---|---|
State | New, archived |
Headers | show |
Series | drm/i915/dg2: Incorporate Wa_16014892111 into DRAW_WATERMARK tuning | expand |
On 23.08.2022 13:24, Matt Roper wrote: > Although register tuning settings are generally implemented via the > workaround infrastructure, it turns out that the DRAW_WATERMARK register > is not properly saved/restored by hardware around power events (i.e., > RC6 entry) so updates to the value cannot be applied in the usual > manner. New workaround Wa_16014892111 informs us that any tuning > updates to this register must instead be applied via an INDIRECT_CTX > batch buffer. This will ensure that the necessary value is re-applied > when a context begins running, even if an RC6 entry had wiped the > register back to hardware defaults since the last context ran. > > Fixes: 6dc85721df74 ("drm/i915/dg2: Add additional tuning settings") > Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/6642 > Signed-off-by: Matt Roper <matthew.d.roper@intel.com> Reviewed-by: Balasubramani Vivekanandan <balasubramani.vivekanandan@intel.com> > --- > drivers/gpu/drm/i915/gt/intel_lrc.c | 21 +++++++++++++++++++++ > drivers/gpu/drm/i915/gt/intel_workarounds.c | 2 -- > 2 files changed, 21 insertions(+), 2 deletions(-) > > diff --git a/drivers/gpu/drm/i915/gt/intel_lrc.c b/drivers/gpu/drm/i915/gt/intel_lrc.c > index eec73c66406c..070cec4ff8a4 100644 > --- a/drivers/gpu/drm/i915/gt/intel_lrc.c > +++ b/drivers/gpu/drm/i915/gt/intel_lrc.c > @@ -1242,6 +1242,23 @@ dg2_emit_rcs_hang_wabb(const struct intel_context *ce, u32 *cs) > return cs; > } > > +/* > + * The bspec's tuning guide asks us to program a vertical watermark value of > + * 0x3FF. However this register is not saved/restored properly by the > + * hardware, so we're required to apply the desired value via INDIRECT_CTX > + * batch buffer to ensure the value takes effect properly. All other bits > + * in this register should remain at 0 (the hardware default). > + */ > +static u32 * > +dg2_emit_draw_watermark_setting(u32 *cs) > +{ > + *cs++ = MI_LOAD_REGISTER_IMM(1); > + *cs++ = i915_mmio_reg_offset(DRAW_WATERMARK); > + *cs++ = REG_FIELD_PREP(VERT_WM_VAL, 0x3FF); > + > + return cs; > +} > + > static u32 * > gen12_emit_indirect_ctx_rcs(const struct intel_context *ce, u32 *cs) > { > @@ -1263,6 +1280,10 @@ gen12_emit_indirect_ctx_rcs(const struct intel_context *ce, u32 *cs) > if (!HAS_FLAT_CCS(ce->engine->i915)) > cs = gen12_emit_aux_table_inv(cs, GEN12_GFX_CCS_AUX_NV); > > + /* Wa_16014892111 */ > + if (IS_DG2(ce->engine->i915)) > + cs = dg2_emit_draw_watermark_setting(cs); > + > return cs; > } > > diff --git a/drivers/gpu/drm/i915/gt/intel_workarounds.c b/drivers/gpu/drm/i915/gt/intel_workarounds.c > index 31e129329fb0..3cdb8294e13f 100644 > --- a/drivers/gpu/drm/i915/gt/intel_workarounds.c > +++ b/drivers/gpu/drm/i915/gt/intel_workarounds.c > @@ -2685,8 +2685,6 @@ add_render_compute_tuning_settings(struct drm_i915_private *i915, > if (IS_DG2(i915)) { > wa_write_or(wal, XEHP_L3SCQREG7, BLEND_FILL_CACHING_OPT_DIS); > wa_write_clr_set(wal, RT_CTRL, STACKID_CTRL, STACKID_CTRL_512); > - wa_write_clr_set(wal, DRAW_WATERMARK, VERT_WM_VAL, > - REG_FIELD_PREP(VERT_WM_VAL, 0x3FF)); > > /* > * This is also listed as Wa_22012654132 for certain DG2 > -- > 2.37.2 >
diff --git a/drivers/gpu/drm/i915/gt/intel_lrc.c b/drivers/gpu/drm/i915/gt/intel_lrc.c index eec73c66406c..070cec4ff8a4 100644 --- a/drivers/gpu/drm/i915/gt/intel_lrc.c +++ b/drivers/gpu/drm/i915/gt/intel_lrc.c @@ -1242,6 +1242,23 @@ dg2_emit_rcs_hang_wabb(const struct intel_context *ce, u32 *cs) return cs; } +/* + * The bspec's tuning guide asks us to program a vertical watermark value of + * 0x3FF. However this register is not saved/restored properly by the + * hardware, so we're required to apply the desired value via INDIRECT_CTX + * batch buffer to ensure the value takes effect properly. All other bits + * in this register should remain at 0 (the hardware default). + */ +static u32 * +dg2_emit_draw_watermark_setting(u32 *cs) +{ + *cs++ = MI_LOAD_REGISTER_IMM(1); + *cs++ = i915_mmio_reg_offset(DRAW_WATERMARK); + *cs++ = REG_FIELD_PREP(VERT_WM_VAL, 0x3FF); + + return cs; +} + static u32 * gen12_emit_indirect_ctx_rcs(const struct intel_context *ce, u32 *cs) { @@ -1263,6 +1280,10 @@ gen12_emit_indirect_ctx_rcs(const struct intel_context *ce, u32 *cs) if (!HAS_FLAT_CCS(ce->engine->i915)) cs = gen12_emit_aux_table_inv(cs, GEN12_GFX_CCS_AUX_NV); + /* Wa_16014892111 */ + if (IS_DG2(ce->engine->i915)) + cs = dg2_emit_draw_watermark_setting(cs); + return cs; } diff --git a/drivers/gpu/drm/i915/gt/intel_workarounds.c b/drivers/gpu/drm/i915/gt/intel_workarounds.c index 31e129329fb0..3cdb8294e13f 100644 --- a/drivers/gpu/drm/i915/gt/intel_workarounds.c +++ b/drivers/gpu/drm/i915/gt/intel_workarounds.c @@ -2685,8 +2685,6 @@ add_render_compute_tuning_settings(struct drm_i915_private *i915, if (IS_DG2(i915)) { wa_write_or(wal, XEHP_L3SCQREG7, BLEND_FILL_CACHING_OPT_DIS); wa_write_clr_set(wal, RT_CTRL, STACKID_CTRL, STACKID_CTRL_512); - wa_write_clr_set(wal, DRAW_WATERMARK, VERT_WM_VAL, - REG_FIELD_PREP(VERT_WM_VAL, 0x3FF)); /* * This is also listed as Wa_22012654132 for certain DG2
Although register tuning settings are generally implemented via the workaround infrastructure, it turns out that the DRAW_WATERMARK register is not properly saved/restored by hardware around power events (i.e., RC6 entry) so updates to the value cannot be applied in the usual manner. New workaround Wa_16014892111 informs us that any tuning updates to this register must instead be applied via an INDIRECT_CTX batch buffer. This will ensure that the necessary value is re-applied when a context begins running, even if an RC6 entry had wiped the register back to hardware defaults since the last context ran. Fixes: 6dc85721df74 ("drm/i915/dg2: Add additional tuning settings") Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/6642 Signed-off-by: Matt Roper <matthew.d.roper@intel.com> --- drivers/gpu/drm/i915/gt/intel_lrc.c | 21 +++++++++++++++++++++ drivers/gpu/drm/i915/gt/intel_workarounds.c | 2 -- 2 files changed, 21 insertions(+), 2 deletions(-)