Rodrigo Vivi <rodrigo.v...@intel.com> writes:
> Coffee Lake inherit most of Kabylake production
> workardounds.
>
> Only difference identified so far is:
> - WaDisableLSQCROPERFforOCL is marked as SIWA_NEVER
>
> Cc: Dhinakaran Pandiyan <dhinakaran.pandi...@intel.com>
> Signed-off-by: Rodrigo Vivi <rodrigo.v...@intel.com>
> ---
> drivers/gpu/drm/i915/i915_gem_gtt.c | 2 +-
> drivers/gpu/drm/i915/intel_engine_cs.c | 75
> +++++++++++++++++++++++++---------
> drivers/gpu/drm/i915/intel_pm.c | 10 ++---
> 3 files changed, 61 insertions(+), 26 deletions(-)
>
> diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c
> b/drivers/gpu/drm/i915/i915_gem_gtt.c
> index 4ff854e..8e055b1 100644
> --- a/drivers/gpu/drm/i915/i915_gem_gtt.c
> +++ b/drivers/gpu/drm/i915/i915_gem_gtt.c
> @@ -1884,7 +1884,7 @@ static void gtt_write_workarounds(struct
> drm_i915_private *dev_priv)
> * called on driver load and after a GPU reset, so you can place
> * workarounds here even if they get overwritten by GPU reset.
> */
> - /* WaIncreaseDefaultTLBEntries:chv,bdw,skl,bxt,kbl,glk */
> + /* WaIncreaseDefaultTLBEntries:chv,bdw,skl,bxt,kbl,glk,cfl */
> if (IS_BROADWELL(dev_priv))
> I915_WRITE(GEN8_L3_LRA_1_GPGPU,
> GEN8_L3_LRA_1_GPGPU_DEFAULT_VALUE_BDW);
> else if (IS_CHERRYVIEW(dev_priv))
> diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c
> b/drivers/gpu/drm/i915/intel_engine_cs.c
> index bc38bd1..630ff6e 100644
> --- a/drivers/gpu/drm/i915/intel_engine_cs.c
> +++ b/drivers/gpu/drm/i915/intel_engine_cs.c
> @@ -814,24 +814,24 @@ static int gen9_init_workarounds(struct intel_engine_cs
> *engine)
> struct drm_i915_private *dev_priv = engine->i915;
> int ret;
>
> - /* WaConextSwitchWithConcurrentTLBInvalidate:skl,bxt,kbl,glk */
> + /* WaConextSwitchWithConcurrentTLBInvalidate:skl,bxt,kbl,glk,cfl */
> I915_WRITE(GEN9_CSFE_CHICKEN1_RCS,
> _MASKED_BIT_ENABLE(GEN9_PREEMPT_GPGPU_SYNC_SWITCH_DISABLE));
>
> - /* WaEnableLbsSlaRetryTimerDecrement:skl,bxt,kbl,glk */
> + /* WaEnableLbsSlaRetryTimerDecrement:skl,bxt,kbl,glk,cfl */
> I915_WRITE(BDW_SCRATCH1, I915_READ(BDW_SCRATCH1) |
> GEN9_LBS_SLA_RETRY_TIMER_DECREMENT_ENABLE);
>
> - /* WaDisableKillLogic:bxt,skl,kbl */
> + /* WaDisableKillLogic:bxt,skl,kbl,cfl */
> I915_WRITE(GAM_ECOCHK, I915_READ(GAM_ECOCHK) |
> ECOCHK_DIS_TLB);
>
I couldn't find reference to this in anywhere. So I
would suggest excluding it, unless you have better sources :)
> - /* WaClearFlowControlGpgpuContextSave:skl,bxt,kbl,glk */
> - /* WaDisablePartialInstShootdown:skl,bxt,kbl,glk */
> + /* WaClearFlowControlGpgpuContextSave:skl,bxt,kbl,glk,cfl */
> + /* WaDisablePartialInstShootdown:skl,bxt,kbl,glk,cfl */
> WA_SET_BIT_MASKED(GEN8_ROW_CHICKEN,
> FLOW_CONTROL_ENABLE |
> PARTIAL_INSTRUCTION_SHOOTDOWN_DISABLE);
>
> - /* Syncing dependencies between camera and graphics:skl,bxt,kbl */
> + /* Syncing dependencies between camera and graphics:skl,bxt,kbl,cfl */
> WA_SET_BIT_MASKED(HALF_SLICE_CHICKEN3,
> GEN9_DISABLE_OCL_OOB_SUPPRESS_LOGIC);
>
Bspec says that should be disabled for skl+. Bspec is outdated?
> @@ -851,18 +851,18 @@ static int gen9_init_workarounds(struct intel_engine_cs
> *engine)
> */
> }
>
> - /* WaEnableYV12BugFixInHalfSliceChicken7:skl,bxt,kbl,glk */
> - /* WaEnableSamplerGPGPUPreemptionSupport:skl,bxt,kbl */
> + /* WaEnableYV12BugFixInHalfSliceChicken7:skl,bxt,kbl,glk,cfl */
> + /* WaEnableSamplerGPGPUPreemptionSupport:skl,bxt,kbl,cfl */
> WA_SET_BIT_MASKED(GEN9_HALF_SLICE_CHICKEN7,
> GEN9_ENABLE_YV12_BUGFIX |
> GEN9_ENABLE_GPGPU_PREEMPTION);
>
> - /* Wa4x4STCOptimizationDisable:skl,bxt,kbl,glk */
> - /* WaDisablePartialResolveInVc:skl,bxt,kbl */
> + /* Wa4x4STCOptimizationDisable:skl,bxt,kbl,glk,cfl */
> + /* WaDisablePartialResolveInVc:skl,bxt,kbl,cfl */
> WA_SET_BIT_MASKED(CACHE_MODE_1, (GEN8_4x4_STC_OPTIMIZATION_DISABLE |
> GEN9_PARTIAL_RESOLVE_IN_VC_DISABLE));
>
> - /* WaCcsTlbPrefetchDisable:skl,bxt,kbl,glk */
> + /* WaCcsTlbPrefetchDisable:skl,bxt,kbl,glk,cfl */
> WA_CLR_BIT_MASKED(GEN9_HALF_SLICE_CHICKEN5,
> GEN9_CCS_TLB_PREFETCH_ENABLE);
>
> @@ -871,7 +871,7 @@ static int gen9_init_workarounds(struct intel_engine_cs
> *engine)
> WA_SET_BIT_MASKED(SLICE_ECO_CHICKEN0,
> PIXEL_MASK_CAMMING_DISABLE);
>
> - /* WaForceContextSaveRestoreNonCoherent:skl,bxt,kbl */
> + /* WaForceContextSaveRestoreNonCoherent:skl,bxt,kbl,cfl */
> WA_SET_BIT_MASKED(HDC_CHICKEN0,
> HDC_FORCE_CONTEXT_SAVE_RESTORE_NON_COHERENT |
> HDC_FORCE_CSR_NON_COHERENT_OVR_DISABLE);
> @@ -889,39 +889,40 @@ static int gen9_init_workarounds(struct intel_engine_cs
> *engine)
> * a TLB invalidation occurs during a PSD flush.
> */
>
> - /* WaForceEnableNonCoherent:skl,bxt,kbl */
> + /* WaForceEnableNonCoherent:skl,bxt,kbl,cfl */
> WA_SET_BIT_MASKED(HDC_CHICKEN0,
> HDC_FORCE_NON_COHERENT);
>
> - /* WaDisableHDCInvalidation:skl,bxt,kbl */
> + /* WaDisableHDCInvalidation:skl,bxt,kbl,cfl */
> I915_WRITE(GAM_ECOCHK, I915_READ(GAM_ECOCHK) |
> BDW_DISABLE_HDC_INVALIDATION);
>
The two above are not mentioned with cfl. But the comment and
history of these workarounds speaks volumes, so I agree with these.
> - /* WaDisableSamplerPowerBypassForSOPingPong:skl,bxt,kbl */
> + /* WaDisableSamplerPowerBypassForSOPingPong:skl,bxt,kbl,cfl */
> if (IS_SKYLAKE(dev_priv) ||
> IS_KABYLAKE(dev_priv) ||
> + IS_COFFEELAKE(dev_priv)||
> IS_BXT_REVID(dev_priv, 0, BXT_REVID_B0))
> WA_SET_BIT_MASKED(HALF_SLICE_CHICKEN3,
> GEN8_SAMPLER_POWER_BYPASS_DIS);
>
> - /* WaDisableSTUnitPowerOptimization:skl,bxt,kbl,glk */
> + /* WaDisableSTUnitPowerOptimization:skl,bxt,kbl,glk,cfl */
> WA_SET_BIT_MASKED(HALF_SLICE_CHICKEN2, GEN8_ST_PO_DISABLE);
>
> - /* WaOCLCoherentLineFlush:skl,bxt,kbl */
> + /* WaOCLCoherentLineFlush:skl,bxt,kbl,cfl */
> I915_WRITE(GEN8_L3SQCREG4, (I915_READ(GEN8_L3SQCREG4) |
> GEN8_LQSC_FLUSH_COHERENT_LINES));
>
For me it seems that the workaround name has changed:
WaPipelineFlushCoherentLines. Should be carried to cfl.
> - /* WaVFEStateAfterPipeControlwithMediaStateClear:skl,bxt,glk */
> + /* WaVFEStateAfterPipeControlwithMediaStateClear:skl,bxt,glk,cfl */
> ret = wa_ring_whitelist_reg(engine, GEN9_CTX_PREEMPT_REG);
> if (ret)
> return ret;
>
> - /* WaEnablePreemptionGranularityControlByUMD:skl,bxt,kbl */
> + /* WaEnablePreemptionGranularityControlByUMD:skl,bxt,kbl,cfl */
> ret= wa_ring_whitelist_reg(engine, GEN8_CS_CHICKEN1);
> if (ret)
> return ret;
>
> - /* WaAllowUMDToModifyHDCChicken1:skl,bxt,kbl,glk */
> + /* WaAllowUMDToModifyHDCChicken1:skl,bxt,kbl,glk,cfl */
> ret = wa_ring_whitelist_reg(engine, GEN8_HDC_CHICKEN1);
> if (ret)
> return ret;
> @@ -1140,6 +1141,38 @@ static int glk_init_workarounds(struct intel_engine_cs
> *engine)
> return 0;
> }
>
> +static int cfl_init_workarounds(struct intel_engine_cs *engine)
> +{
> + struct drm_i915_private *dev_priv = engine->i915;
> + int ret;
> +
> + ret = gen9_init_workarounds(engine);
> + if (ret)
> + return ret;
> +
> + /* WaEnableGapsTsvCreditFix:cfl */
> + I915_WRITE(GEN8_GARBCNTL, (I915_READ(GEN8_GARBCNTL) |
> + GEN9_GAPS_TSV_CREDIT_DISABLE));
> +
> + /* WaToEnableHwFixForPushConstHWBug:cfl */
> + WA_SET_BIT_MASKED(COMMON_SLICE_CHICKEN2,
> + GEN8_SBE_DISABLE_REPLAY_BUF_OPTIMIZATION);
> +
> + /* WaDisableGafsUnitClkGating:cfl */
> + WA_SET_BIT(GEN7_UCGCTL4, GEN8_EU_GAUNIT_CLOCK_GATE_DISABLE);
> +
> + /* WaDisableSbeCacheDispatchPortSharing:cfl */
> + WA_SET_BIT_MASKED(
> + GEN7_HALF_SLICE_CHICKEN1,
> + GEN7_SBE_SS_CACHE_DISPATCH_PORT_SHARING_DISABLE);
> +
> + /* WaInPlaceDecompressionHang:cfl */
> + WA_SET_BIT(GEN9_GAMT_ECO_REG_RW_IA,
> + GAMT_ECO_ENABLE_IN_PLACE_DECOMPRESS);
> +
> + return 0;
> +}
> +
> int init_workarounds_ring(struct intel_engine_cs *engine)
> {
> struct drm_i915_private *dev_priv = engine->i915;
> @@ -1162,6 +1195,8 @@ int init_workarounds_ring(struct intel_engine_cs
> *engine)
> err = kbl_init_workarounds(engine);
> else if (IS_GEMINILAKE(dev_priv))
> err = glk_init_workarounds(engine);
> + else if (IS_COFFEELAKE(dev_priv))
> + err = cfl_init_workarounds(engine);
> else
> err = 0;
> if (err)
> diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c
> index 98aeba9..0aed13d 100644
> --- a/drivers/gpu/drm/i915/intel_pm.c
> +++ b/drivers/gpu/drm/i915/intel_pm.c
> @@ -58,24 +58,24 @@
>
> static void gen9_init_clock_gating(struct drm_i915_private *dev_priv)
> {
> - /* See Bspec note for PSR2_CTL bit 31, Wa#828:skl,bxt,kbl */
> + /* See Bspec note for PSR2_CTL bit 31, Wa#828:skl,bxt,kbl,cfl */
> I915_WRITE(CHICKEN_PAR1_1,
> I915_READ(CHICKEN_PAR1_1) | SKL_EDP_PSR_FIX_RDWRAP);
>
> I915_WRITE(GEN8_CONFIG0,
> I915_READ(GEN8_CONFIG0) | GEN9_DEFAULT_FIXES);
>
> - /* WaEnableChickenDCPR:skl,bxt,kbl,glk */
> + /* WaEnableChickenDCPR:skl,bxt,kbl,glk,cfl */
> I915_WRITE(GEN8_CHICKEN_DCPR_1,
> I915_READ(GEN8_CHICKEN_DCPR_1) | MASK_WAKEMEM);
>
> - /* WaFbcTurnOffFbcWatermark:skl,bxt,kbl */
> - /* WaFbcWakeMemOn:skl,bxt,kbl,glk */
> + /* WaFbcTurnOffFbcWatermark:skl,bxt,kbl,cfl */
> + /* WaFbcWakeMemOn:skl,bxt,kbl,glk,cfl */
> I915_WRITE(DISP_ARB_CTL, I915_READ(DISP_ARB_CTL) |
> DISP_FBC_WM_DIS |
> DISP_FBC_MEMORY_WAKE);
>
> - /* WaFbcHighMemBwCorruptionAvoidance:skl,bxt,kbl */
> + /* WaFbcHighMemBwCorruptionAvoidance:skl,bxt,kbl,cfl */
I haven't rummaged around what new we might be missing. So
we should do a second pass after month or two.
With WaDisableKillLogic and GEN9_DISABLE_OCL_OOB_SUPPRESS_LOGIC
double checked.
Reviewed-by: Mika Kuoppala <mika.kuopp...@intel.com>
> I915_WRITE(ILK_DPFC_CHICKEN, I915_READ(ILK_DPFC_CHICKEN) |
> ILK_DPFC_DISABLE_DUMMY0);
> }
> --
> 1.9.1
>
> _______________________________________________
> 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
