From 33def1ff7b09645f3631059ad9ce23e2c65e9016 Mon Sep 17 00:00:00 2001 From: Tvrtko Ursulin Date: Fri, 16 Jun 2017 14:03:38 +0100 Subject: drm/i915: Simplify intel_engines_init We do not want to carry on over missing constructors and don't need a duplicated engine mask checking which is already done in the setup phase. Signed-off-by: Tvrtko Ursulin Reviewed-by: Chris Wilson --- drivers/gpu/drm/i915/intel_engine_cs.c | 36 ++++++++++++---------------------- 1 file changed, 12 insertions(+), 24 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_engine_cs.c') diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c index a4487c5b7e37..3b46c1f7b88b 100644 --- a/drivers/gpu/drm/i915/intel_engine_cs.c +++ b/drivers/gpu/drm/i915/intel_engine_cs.c @@ -291,11 +291,9 @@ cleanup: */ int intel_engines_init(struct drm_i915_private *dev_priv) { - struct intel_device_info *device_info = mkwrite_device_info(dev_priv); struct intel_engine_cs *engine; enum intel_engine_id id, err_id; - unsigned int mask = 0; - int err = 0; + int err; for_each_engine(engine, dev_priv, id) { const struct engine_class_info *class_info = @@ -306,40 +304,30 @@ int intel_engines_init(struct drm_i915_private *dev_priv) init = class_info->init_execlists; else init = class_info->init_legacy; - if (!init) { - kfree(engine); - dev_priv->engine[id] = NULL; - continue; - } + + err = -EINVAL; + err_id = id; + + if (GEM_WARN_ON(!init)) + goto cleanup; err = init(engine); - if (err) { - err_id = id; + if (err) goto cleanup; - } GEM_BUG_ON(!engine->submit_request); - mask |= ENGINE_MASK(id); } - /* - * Catch failures to update intel_engines table when the new engines - * are added to the driver by a warning and disabling the forgotten - * engines. - */ - if (WARN_ON(mask != INTEL_INFO(dev_priv)->ring_mask)) - device_info->ring_mask = mask; - - device_info->num_rings = hweight32(mask); - return 0; cleanup: for_each_engine(engine, dev_priv, id) { - if (id >= err_id) + if (id >= err_id) { kfree(engine); - else + dev_priv->engine[id] = NULL; + } else { dev_priv->gt.cleanup_engine(engine); + } } return err; } -- cgit v1.2.3 From 9cd90018ebd5b4f484300f2a5af804317d3428a1 Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Tue, 27 Jun 2017 16:25:10 +0100 Subject: drm/i915: Cancel pending execlists irq handler upon idling Due to the slight asynchronicity in handling the execlists interrupts (i.e. we defer the work to a handler that may consume more than one interrupt event), when the engine is idle we may still have an irq tasklet queued (especially when it has been deferred to a ksoftirqd). At the beginning of the tasklet, we assert that we do hold a device wakeref for the access we are about to perform. This assumes that when we idle and release the GT wakeref, all execlists work has been completed (since the elsp tracking says the hw is idle). However, there may still be a tasklet queued, so as we mark the engine idle, also cancel any pending tasklet. Signed-off-by: Chris Wilson Cc: Tvrtko Ursulin Link: http://patchwork.freedesktop.org/patch/msgid/20170627152510.28589-1-chris@chris-wilson.co.uk Reviewed-by: Tvrtko Ursulin --- drivers/gpu/drm/i915/intel_engine_cs.c | 1 + 1 file changed, 1 insertion(+) (limited to 'drivers/gpu/drm/i915/intel_engine_cs.c') diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c index 3b46c1f7b88b..49e875c46c96 100644 --- a/drivers/gpu/drm/i915/intel_engine_cs.c +++ b/drivers/gpu/drm/i915/intel_engine_cs.c @@ -1328,6 +1328,7 @@ void intel_engines_mark_idle(struct drm_i915_private *i915) for_each_engine(engine, i915, id) { intel_engine_disarm_breadcrumbs(engine); i915_gem_batch_pool_fini(&engine->batch_pool); + tasklet_kill(&engine->irq_tasklet); engine->no_priolist = false; } } -- cgit v1.2.3 From 98eed3d1ade53596e1c8785e049f03da4480a820 Mon Sep 17 00:00:00 2001 From: Rodrigo Vivi Date: Mon, 19 Jun 2017 14:21:47 -0700 Subject: drm/i915/cfl: Fix Workarounds. MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit During the review of Coffee Lake workarounds Mika pointed out that WaDisableKillLogic and GEN9_DISABLE_OCL_OOB_SUPPRESS_LOGIC should be removed from CFL and with that I should carry the rv-b. However when doing the v2 I removed another Workaround that should remain because although not mentioned by spec the history of hangs around it advocates on its favor. On some follow-up patches I continued operating on the wrong workardound, but Ville noticed that, so here is the fix for the current CFL code that is upstream already. Fixes: 46c26662d2f ("drm/i915/cfl: Introduce Coffee Lake workarounds.") Cc: Ville Syrjälä Cc: Dhinakaran Pandiyan Cc: Mika Kuoppala Signed-off-by: Rodrigo Vivi Reviewed-by: Dhinakaran Pandiyan --- drivers/gpu/drm/i915/intel_engine_cs.c | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_engine_cs.c') diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c index 49e875c46c96..a55cd72aeeff 100644 --- a/drivers/gpu/drm/i915/intel_engine_cs.c +++ b/drivers/gpu/drm/i915/intel_engine_cs.c @@ -809,9 +809,10 @@ static int gen9_init_workarounds(struct intel_engine_cs *engine) I915_WRITE(BDW_SCRATCH1, I915_READ(BDW_SCRATCH1) | GEN9_LBS_SLA_RETRY_TIMER_DECREMENT_ENABLE); - /* WaDisableKillLogic:bxt,skl,kbl,cfl */ - I915_WRITE(GAM_ECOCHK, I915_READ(GAM_ECOCHK) | - ECOCHK_DIS_TLB); + /* WaDisableKillLogic:bxt,skl,kbl */ + if (!IS_COFFEELAKE(dev_priv)) + I915_WRITE(GAM_ECOCHK, I915_READ(GAM_ECOCHK) | + ECOCHK_DIS_TLB); /* WaClearFlowControlGpgpuContextSave:skl,bxt,kbl,glk,cfl */ /* WaDisablePartialInstShootdown:skl,bxt,kbl,glk,cfl */ @@ -882,10 +883,9 @@ static int gen9_init_workarounds(struct intel_engine_cs *engine) WA_SET_BIT_MASKED(HDC_CHICKEN0, HDC_FORCE_NON_COHERENT); - /* WaDisableHDCInvalidation:skl,bxt,kbl */ - if (!IS_COFFEELAKE(dev_priv)) - I915_WRITE(GAM_ECOCHK, I915_READ(GAM_ECOCHK) | - BDW_DISABLE_HDC_INVALIDATION); + /* WaDisableHDCInvalidation:skl,bxt,kbl,cfl */ + I915_WRITE(GAM_ECOCHK, I915_READ(GAM_ECOCHK) | + BDW_DISABLE_HDC_INVALIDATION); /* WaDisableSamplerPowerBypassForSOPingPong:skl,bxt,kbl,cfl */ if (IS_SKYLAKE(dev_priv) || -- cgit v1.2.3 From f65f84178999c2c0227f9ae7dafd8af62c0ce5ad Mon Sep 17 00:00:00 2001 From: Rodrigo Vivi Date: Thu, 6 Jul 2017 14:06:24 -0700 Subject: drm/i915/cnl: Gen10 render context size. No change on render context size is required for Gen10. So this patch doesn't change the default behaviour, but only avoid the missing_case message. Cc: Ben Widawsky Signed-off-by: Rodrigo Vivi Reviewed-by: Ben Widawsky Link: http://patchwork.freedesktop.org/patch/msgid/1499375184-5725-1-git-send-email-rodrigo.vivi@intel.com --- drivers/gpu/drm/i915/intel_engine_cs.c | 1 + 1 file changed, 1 insertion(+) (limited to 'drivers/gpu/drm/i915/intel_engine_cs.c') diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c index a55cd72aeeff..24db316e0fd1 100644 --- a/drivers/gpu/drm/i915/intel_engine_cs.c +++ b/drivers/gpu/drm/i915/intel_engine_cs.c @@ -149,6 +149,7 @@ __intel_engine_context_size(struct drm_i915_private *dev_priv, u8 class) switch (INTEL_GEN(dev_priv)) { default: MISSING_CASE(INTEL_GEN(dev_priv)); + case 10: case 9: return GEN9_LR_CONTEXT_RENDER_SIZE; case 8: -- cgit v1.2.3 From d6edb6e3b63faae5cd96993e0977d3807bd87be6 Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Fri, 21 Jul 2017 13:32:24 +0100 Subject: drm/i915: Check the execlist queue for pending requests before declaring idle Including a check against the execlist queue before calling the engine idle and passing hangcheck. Signed-off-by: Chris Wilson Reviewed-by: Mika Kuoppala Link: https://patchwork.freedesktop.org/patch/msgid/20170721123238.16428-6-chris@chris-wilson.co.uk Signed-off-by: Daniel Vetter --- drivers/gpu/drm/i915/intel_engine_cs.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'drivers/gpu/drm/i915/intel_engine_cs.c') diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c index 24db316e0fd1..c6ebfe7fb4f1 100644 --- a/drivers/gpu/drm/i915/intel_engine_cs.c +++ b/drivers/gpu/drm/i915/intel_engine_cs.c @@ -1283,6 +1283,10 @@ bool intel_engine_is_idle(struct intel_engine_cs *engine) if (port_request(&engine->execlist_port[0])) return false; + /* ELSP is empty, but there are ready requests? */ + if (READ_ONCE(engine->execlist_first)) + return false; + /* Ring stopped? */ if (!ring_is_idle(engine)) return false; -- cgit v1.2.3 From 4d53568cca96015d7ad44353dfd83c10ef14c6cf Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Fri, 21 Jul 2017 13:32:26 +0100 Subject: drm/i915: Move idle checks before intel_engine_init_global_seqno() intel_engine_init_globa_seqno() may be called from an uncontrolled set-wedged path where we have given up waiting for broken hw and declare it defunct. Along that path, any sanity checks that the hw is idle before we adjust its state will expectedly fail, so we simply cannot. Instead of asserting inside init_global_seqno, we move them to the normal caller reset_all_global_seqno() as it handles runtime seqno wraparound. Signed-off-by: Chris Wilson Reviewed-by: Mika Kuoppala Link: https://patchwork.freedesktop.org/patch/msgid/20170721123238.16428-8-chris@chris-wilson.co.uk Signed-off-by: Daniel Vetter --- drivers/gpu/drm/i915/i915_gem_request.c | 4 ++++ drivers/gpu/drm/i915/intel_engine_cs.c | 3 --- 2 files changed, 4 insertions(+), 3 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_engine_cs.c') diff --git a/drivers/gpu/drm/i915/i915_gem_request.c b/drivers/gpu/drm/i915/i915_gem_request.c index 483af8921060..d93a185c0f0a 100644 --- a/drivers/gpu/drm/i915/i915_gem_request.c +++ b/drivers/gpu/drm/i915/i915_gem_request.c @@ -213,6 +213,10 @@ static int reset_all_global_seqno(struct drm_i915_private *i915, u32 seqno) cond_resched(); } + /* Check we are idle before we fiddle with hw state! */ + GEM_BUG_ON(!intel_engine_is_idle(engine)); + GEM_BUG_ON(i915_gem_active_isset(&engine->timeline->last_request)); + /* Finally reset hw state */ intel_engine_init_global_seqno(engine, seqno); tl->seqno = seqno; diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c index c6ebfe7fb4f1..9ab596941372 100644 --- a/drivers/gpu/drm/i915/intel_engine_cs.c +++ b/drivers/gpu/drm/i915/intel_engine_cs.c @@ -337,9 +337,6 @@ void intel_engine_init_global_seqno(struct intel_engine_cs *engine, u32 seqno) { struct drm_i915_private *dev_priv = engine->i915; - GEM_BUG_ON(!intel_engine_is_idle(engine)); - GEM_BUG_ON(i915_gem_active_isset(&engine->timeline->last_request)); - /* Our semaphore implementation is strictly monotonic (i.e. we proceed * so long as the semaphore value in the register/page is greater * than the sync value), so whenever we reset the seqno, -- cgit v1.2.3 From 90007bca61627bb2fbd63bc3da0277abe4a43550 Mon Sep 17 00:00:00 2001 From: Rodrigo Vivi Date: Tue, 15 Aug 2017 16:16:48 -0700 Subject: drm/i915/cnl: Introduce initial Cannonlake Workarounds. Let's inherit workarounds from previous platforms that according to wa_database and BSpec are still valid for Cannonlake. v2: Add missed workarounds. v3: Rebase v4: Remove bad chunk that was added to rc6 disable. (Ander) Also remove A0 W/a that are not needed anymore. v5: Rebase on top of CFL. v6: Remove empty gen9_init_perctx_bb and gen9_init_indirectctx_bb since they don't carry any gen10 related W/a. (by Oscar). Also Remove A0 exclusive workaround. v7: Remove more A0 exclusive workarounds. As pointed out by Oscar many workarounds were changed to be A0 only so let's remove them. Cc: Oscar Mateo Cc: Mika Kuoppala Signed-off-by: Rodrigo Vivi Reviewed-by: Oscar Mateo Link: https://patchwork.freedesktop.org/patch/msgid/20170815231651.975-1-rodrigo.vivi@intel.com --- drivers/gpu/drm/i915/i915_gem_gtt.c | 4 ++-- drivers/gpu/drm/i915/i915_reg.h | 6 ++++++ drivers/gpu/drm/i915/intel_engine_cs.c | 19 +++++++++++++++++++ drivers/gpu/drm/i915/intel_lrc.c | 2 ++ drivers/gpu/drm/i915/intel_pm.c | 21 ++++++++++++++++++++- 5 files changed, 49 insertions(+), 3 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_engine_cs.c') diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c index d60f38adc4c4..0f7399801398 100644 --- a/drivers/gpu/drm/i915/i915_gem_gtt.c +++ b/drivers/gpu/drm/i915/i915_gem_gtt.c @@ -1885,12 +1885,12 @@ 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,cfl */ + /* WaIncreaseDefaultTLBEntries:chv,bdw,skl,bxt,kbl,glk,cfl,cnl */ 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)) I915_WRITE(GEN8_L3_LRA_1_GPGPU, GEN8_L3_LRA_1_GPGPU_DEFAULT_VALUE_CHV); - else if (IS_GEN9_BC(dev_priv)) + else if (IS_GEN9_BC(dev_priv) || IS_GEN10(dev_priv)) I915_WRITE(GEN8_L3_LRA_1_GPGPU, GEN9_L3_LRA_1_GPGPU_DEFAULT_VALUE_SKL); else if (IS_GEN9_LP(dev_priv)) I915_WRITE(GEN8_L3_LRA_1_GPGPU, GEN9_L3_LRA_1_GPGPU_DEFAULT_VALUE_BXT); diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index ed7cd9ee2c2a..2dcae9f24a85 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -3806,6 +3806,12 @@ enum { #define PWM2_GATING_DIS (1 << 14) #define PWM1_GATING_DIS (1 << 13) +/* + * GEN10 clock gating regs + */ +#define SLICE_UNIT_LEVEL_CLKGATE _MMIO(0x94d4) +#define SARBUNIT_CLKGATE_DIS (1 << 5) + /* * Display engine regs */ diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c index 9ab596941372..58a235316f92 100644 --- a/drivers/gpu/drm/i915/intel_engine_cs.c +++ b/drivers/gpu/drm/i915/intel_engine_cs.c @@ -1065,6 +1065,23 @@ static int bxt_init_workarounds(struct intel_engine_cs *engine) return 0; } +static int cnl_init_workarounds(struct intel_engine_cs *engine) +{ + struct drm_i915_private *dev_priv = engine->i915; + int ret; + + /* WaInPlaceDecompressionHang:cnl */ + WA_SET_BIT(GEN9_GAMT_ECO_REG_RW_IA, + GAMT_ECO_ENABLE_IN_PLACE_DECOMPRESS); + + /* WaEnablePreemptionGranularityControlByUMD:cnl */ + ret= wa_ring_whitelist_reg(engine, GEN8_CS_CHICKEN1); + if (ret) + return ret; + + return 0; +} + static int kbl_init_workarounds(struct intel_engine_cs *engine) { struct drm_i915_private *dev_priv = engine->i915; @@ -1185,6 +1202,8 @@ int init_workarounds_ring(struct intel_engine_cs *engine) err = glk_init_workarounds(engine); else if (IS_COFFEELAKE(dev_priv)) err = cfl_init_workarounds(engine); + else if (IS_CANNONLAKE(dev_priv)) + err = cnl_init_workarounds(engine); else err = 0; if (err) diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c index 6f972e6ec663..d89e1b8e1cc5 100644 --- a/drivers/gpu/drm/i915/intel_lrc.c +++ b/drivers/gpu/drm/i915/intel_lrc.c @@ -1175,6 +1175,8 @@ static int intel_init_workaround_bb(struct intel_engine_cs *engine) return -EINVAL; switch (INTEL_GEN(engine->i915)) { + case 10: + return 0; case 9: wa_bb_fn[0] = gen9_init_indirectctx_bb; wa_bb_fn[1] = gen9_init_perctx_bb; diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c index ed662937ec3c..48db4b54f586 100644 --- a/drivers/gpu/drm/i915/intel_pm.c +++ b/drivers/gpu/drm/i915/intel_pm.c @@ -8263,6 +8263,23 @@ static void gen8_set_l3sqc_credits(struct drm_i915_private *dev_priv, I915_WRITE(GEN7_MISCCPCTL, misccpctl); } +static void cannonlake_init_clock_gating(struct drm_i915_private *dev_priv) +{ + /* WaEnableChickenDCPR:cnl */ + I915_WRITE(GEN8_CHICKEN_DCPR_1, + I915_READ(GEN8_CHICKEN_DCPR_1) | MASK_WAKEMEM); + + /* WaFbcWakeMemOn:cnl */ + I915_WRITE(DISP_ARB_CTL, I915_READ(DISP_ARB_CTL) | + DISP_FBC_MEMORY_WAKE); + + /* WaSarbUnitClockGatingDisable:cnl (pre-prod) */ + if (IS_CNL_REVID(dev_priv, CNL_REVID_A0, CNL_REVID_B0)) + I915_WRITE(SLICE_UNIT_LEVEL_CLKGATE, + I915_READ(SLICE_UNIT_LEVEL_CLKGATE) | + SARBUNIT_CLKGATE_DIS); +} + static void kabylake_init_clock_gating(struct drm_i915_private *dev_priv) { gen9_init_clock_gating(dev_priv); @@ -8743,7 +8760,9 @@ static void nop_init_clock_gating(struct drm_i915_private *dev_priv) */ void intel_init_clock_gating_hooks(struct drm_i915_private *dev_priv) { - if (IS_SKYLAKE(dev_priv)) + if (IS_CANNONLAKE(dev_priv)) + dev_priv->display.init_clock_gating = cannonlake_init_clock_gating; + else if (IS_SKYLAKE(dev_priv)) dev_priv->display.init_clock_gating = skylake_init_clock_gating; else if (IS_KABYLAKE(dev_priv) || IS_COFFEELAKE(dev_priv)) dev_priv->display.init_clock_gating = kabylake_init_clock_gating; -- cgit v1.2.3 From e6d1a4f6b267933f81a9299c80b56cf1a36ca79e Mon Sep 17 00:00:00 2001 From: Rodrigo Vivi Date: Tue, 15 Aug 2017 16:16:49 -0700 Subject: drm/i915/cnl: Add WaDisableReplayBufferBankArbitrationOptimization WA to disable replay buffer destination buffer arbitration optimization. Same Wa on previous platforms has a different name: WaToEnableHwFixForPushConstHWBug Signed-off-by: Rodrigo Vivi Reviewed-by: Mika Kuoppala Link: https://patchwork.freedesktop.org/patch/msgid/20170815231651.975-2-rodrigo.vivi@intel.com --- drivers/gpu/drm/i915/intel_engine_cs.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'drivers/gpu/drm/i915/intel_engine_cs.c') diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c index 58a235316f92..e66817d5d4ad 100644 --- a/drivers/gpu/drm/i915/intel_engine_cs.c +++ b/drivers/gpu/drm/i915/intel_engine_cs.c @@ -1070,6 +1070,10 @@ static int cnl_init_workarounds(struct intel_engine_cs *engine) struct drm_i915_private *dev_priv = engine->i915; int ret; + /* WaDisableReplayBufferBankArbitrationOptimization:cnl */ + WA_SET_BIT_MASKED(COMMON_SLICE_CHICKEN2, + GEN8_SBE_DISABLE_REPLAY_BUF_OPTIMIZATION); + /* WaInPlaceDecompressionHang:cnl */ WA_SET_BIT(GEN9_GAMT_ECO_REG_RW_IA, GAMT_ECO_ENABLE_IN_PLACE_DECOMPRESS); -- cgit v1.2.3 From d1d247543c5626eac1064aafd0a35c231828c211 Mon Sep 17 00:00:00 2001 From: Rodrigo Vivi Date: Tue, 15 Aug 2017 16:16:50 -0700 Subject: drm/i915/cnl: WaDisableEnhancedSBEVertexCaching WA forTDS handle reallocation getting dropped by SDE, which may result in PS attribute corruption. Disable enhanced SBE vertex caching in COMMON_SLICE_CHICKEN2 offset. v2: Make it until B0 as spec tells. (by Mika). Signed-off-by: Rodrigo Vivi Reviewed-by: Mika Kuoppala Reviewed-by: Oscar Mateo Link: https://patchwork.freedesktop.org/patch/msgid/20170815231651.975-3-rodrigo.vivi@intel.com --- drivers/gpu/drm/i915/intel_engine_cs.c | 5 +++++ 1 file changed, 5 insertions(+) (limited to 'drivers/gpu/drm/i915/intel_engine_cs.c') diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c index e66817d5d4ad..d23f18874309 100644 --- a/drivers/gpu/drm/i915/intel_engine_cs.c +++ b/drivers/gpu/drm/i915/intel_engine_cs.c @@ -1074,6 +1074,11 @@ static int cnl_init_workarounds(struct intel_engine_cs *engine) WA_SET_BIT_MASKED(COMMON_SLICE_CHICKEN2, GEN8_SBE_DISABLE_REPLAY_BUF_OPTIMIZATION); + /* WaDisableEnhancedSBEVertexCaching:cnl (pre-prod) */ + if (IS_CNL_REVID(dev_priv, 0, CNL_REVID_B0)) + WA_SET_BIT_MASKED(COMMON_SLICE_CHICKEN2, + GEN8_CSC2_SBE_VUE_CACHE_CONSERVATIVE); + /* WaInPlaceDecompressionHang:cnl */ WA_SET_BIT(GEN9_GAMT_ECO_REG_RW_IA, GAMT_ECO_ENABLE_IN_PLACE_DECOMPRESS); -- cgit v1.2.3 From 2cbecff4122cedff329e3efa32c7f2266125c4a1 Mon Sep 17 00:00:00 2001 From: Oscar Mateo Date: Wed, 23 Aug 2017 12:56:31 -0700 Subject: drm/i915/cnl: WaPushConstantDereferenceHoldDisable CS sometimes hangs on 3D Push Constant dispatches with the new deref enhancement logic in CNL. v2: Improve the commit message (Rodrigo) Cc: Mika Kuoppala Signed-off-by: Oscar Mateo Reviewed-by: Rodrigo Vivi Signed-off-by: Rodrigo Vivi Link: https://patchwork.freedesktop.org/patch/msgid/1503518191-19116-1-git-send-email-oscar.mateo@intel.com --- drivers/gpu/drm/i915/i915_reg.h | 1 + drivers/gpu/drm/i915/intel_engine_cs.c | 3 +++ 2 files changed, 4 insertions(+) (limited to 'drivers/gpu/drm/i915/intel_engine_cs.c') diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index d4ecb1905ad8..d9b0249fe5a1 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -8055,6 +8055,7 @@ enum { #define GEN7_ROW_CHICKEN2 _MMIO(0xe4f4) #define GEN7_ROW_CHICKEN2_GT2 _MMIO(0xf4f4) #define DOP_CLOCK_GATING_DISABLE (1<<0) +#define PUSH_CONSTANT_DEREF_DISABLE (1<<8) #define HSW_ROW_CHICKEN3 _MMIO(0xe49c) #define HSW_ROW_CHICKEN3_L3_GLOBAL_ATOMICS_DISABLE (1 << 6) diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c index d23f18874309..d7e1ccf778a2 100644 --- a/drivers/gpu/drm/i915/intel_engine_cs.c +++ b/drivers/gpu/drm/i915/intel_engine_cs.c @@ -1083,6 +1083,9 @@ static int cnl_init_workarounds(struct intel_engine_cs *engine) WA_SET_BIT(GEN9_GAMT_ECO_REG_RW_IA, GAMT_ECO_ENABLE_IN_PLACE_DECOMPRESS); + /* WaPushConstantDereferenceHoldDisable:cnl */ + WA_SET_BIT(GEN7_ROW_CHICKEN2, PUSH_CONSTANT_DEREF_DISABLE); + /* WaEnablePreemptionGranularityControlByUMD:cnl */ ret= wa_ring_whitelist_reg(engine, GEN8_CS_CHICKEN1); if (ret) -- cgit v1.2.3 From acfb5554c769ad7e09b9e4e42b572cc297a728e9 Mon Sep 17 00:00:00 2001 From: Rodrigo Vivi Date: Wed, 23 Aug 2017 13:35:04 -0700 Subject: drm/i915/cnl: WaForceContextSaveRestoreNonCoherent To avoid a potential hang condition with TLB invalidation we need to enable masked bit 5 of MMIO 0xE5F0 at boot. Same workaround was in place for previous platforms, but the register offset has changed for CNL. But also BSpec doesn't mention the bit 15 as set on gen9 platforms and mark bit as reserved on CNL. v2: Improve commit message accepting Oscar's suggestion. Cc: Mika Kuoppala Cc: Oscar Mateo Signed-off-by: Rodrigo Vivi Reviewed-by: Oscar Mateo Link: https://patchwork.freedesktop.org/patch/msgid/20170823203504.10012-1-rodrigo.vivi@intel.com --- drivers/gpu/drm/i915/i915_reg.h | 1 + drivers/gpu/drm/i915/intel_engine_cs.c | 4 ++++ 2 files changed, 5 insertions(+) (limited to 'drivers/gpu/drm/i915/intel_engine_cs.c') diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index d9b0249fe5a1..c59c590e45c4 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -7024,6 +7024,7 @@ enum { /* GEN8 chicken */ #define HDC_CHICKEN0 _MMIO(0x7300) +#define CNL_HDC_CHICKEN0 _MMIO(0xE5F0) #define HDC_FORCE_CSR_NON_COHERENT_OVR_DISABLE (1<<15) #define HDC_FENCE_DEST_SLM_DISABLE (1<<14) #define HDC_DONOT_FETCH_MEM_WHEN_MASKED (1<<11) diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c index d7e1ccf778a2..a6ac9d0a4156 100644 --- a/drivers/gpu/drm/i915/intel_engine_cs.c +++ b/drivers/gpu/drm/i915/intel_engine_cs.c @@ -1070,6 +1070,10 @@ static int cnl_init_workarounds(struct intel_engine_cs *engine) struct drm_i915_private *dev_priv = engine->i915; int ret; + /* WaForceContextSaveRestoreNonCoherent:cnl */ + WA_SET_BIT_MASKED(CNL_HDC_CHICKEN0, + HDC_FORCE_CONTEXT_SAVE_RESTORE_NON_COHERENT); + /* WaDisableReplayBufferBankArbitrationOptimization:cnl */ WA_SET_BIT_MASKED(COMMON_SLICE_CHICKEN2, GEN8_SBE_DISABLE_REPLAY_BUF_OPTIMIZATION); -- cgit v1.2.3 From 392572feb01c03c9db2f73993bdbff2b5ed45c38 Mon Sep 17 00:00:00 2001 From: Rodrigo Vivi Date: Tue, 29 Aug 2017 16:07:23 -0700 Subject: drm/i915/cnl: WA FtrEnableFastAnisoL1BankingFix WA to enable HW L1 Banking fix that allows aniso to operate at full sample rate. References: HSD#1937670 Cc: Mika Kuoppala Cc: Oscar Mateo Cc: Ben Widawsky Cc: Anuj Phogat Signed-off-by: Rodrigo Vivi Reviewed-by: Mika Kuoppala Link: https://patchwork.freedesktop.org/patch/msgid/20170829230723.20898-1-rodrigo.vivi@intel.com --- drivers/gpu/drm/i915/i915_reg.h | 1 + drivers/gpu/drm/i915/intel_engine_cs.c | 3 +++ 2 files changed, 4 insertions(+) (limited to 'drivers/gpu/drm/i915/intel_engine_cs.c') diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index e2908ae34004..1ad22a824921 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -8072,6 +8072,7 @@ enum { #define HSW_SAMPLE_C_PERFORMANCE (1<<9) #define GEN8_CENTROID_PIXEL_OPT_DIS (1<<8) #define GEN9_DISABLE_OCL_OOB_SUPPRESS_LOGIC (1<<5) +#define CNL_FAST_ANISO_L1_BANKING_FIX (1<<4) #define GEN8_SAMPLER_POWER_BYPASS_DIS (1<<1) #define GEN9_HALF_SLICE_CHICKEN7 _MMIO(0xe194) diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c index a6ac9d0a4156..4b9b7828802d 100644 --- a/drivers/gpu/drm/i915/intel_engine_cs.c +++ b/drivers/gpu/drm/i915/intel_engine_cs.c @@ -1090,6 +1090,9 @@ static int cnl_init_workarounds(struct intel_engine_cs *engine) /* WaPushConstantDereferenceHoldDisable:cnl */ WA_SET_BIT(GEN7_ROW_CHICKEN2, PUSH_CONSTANT_DEREF_DISABLE); + /* FtrEnableFastAnisoL1BankingFix: cnl */ + WA_SET_BIT_MASKED(HALF_SLICE_CHICKEN3, CNL_FAST_ANISO_L1_BANKING_FIX); + /* WaEnablePreemptionGranularityControlByUMD:cnl */ ret= wa_ring_whitelist_reg(engine, GEN8_CS_CHICKEN1); if (ret) -- cgit v1.2.3 From 86ebb015fa744dd1e265c9b45ade870ac859a4d5 Mon Sep 17 00:00:00 2001 From: Rodrigo Vivi Date: Tue, 29 Aug 2017 16:07:51 -0700 Subject: drm/i915/cnl: WaDisableI2mCycleOnWRPort On CNL B0 stepping GAM is not able to detect some deadlock condition and then rise the rise the gam_coh_flush. WA database and spec both mentions to set 4AB8[24]=1 as workaround. Although register offset 0x4AB8 is not documented for any platform. References: HSD#1945815, BSID#1112 Cc: Mika Kuoppala Signed-off-by: Rodrigo Vivi Reviewed-by: Mika Kuoppala Link: https://patchwork.freedesktop.org/patch/msgid/20170829230751.21047-1-rodrigo.vivi@intel.com --- drivers/gpu/drm/i915/i915_reg.h | 1 + drivers/gpu/drm/i915/intel_engine_cs.c | 5 +++++ 2 files changed, 6 insertions(+) (limited to 'drivers/gpu/drm/i915/intel_engine_cs.c') diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index 1ad22a824921..c718c2f2eaeb 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -2373,6 +2373,7 @@ enum i915_power_well_id { #define GAMT_CHKN_BIT_REG _MMIO(0x4ab8) #define GAMT_CHKN_DISABLE_DYNAMIC_CREDIT_SHARING (1<<28) +#define GAMT_CHKN_DISABLE_I2M_CYCLE_ON_WR_PORT (1<<24) #if 0 #define PRB0_TAIL _MMIO(0x2030) diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c index 4b9b7828802d..ae668340620e 100644 --- a/drivers/gpu/drm/i915/intel_engine_cs.c +++ b/drivers/gpu/drm/i915/intel_engine_cs.c @@ -1070,6 +1070,11 @@ static int cnl_init_workarounds(struct intel_engine_cs *engine) struct drm_i915_private *dev_priv = engine->i915; int ret; + /* WaDisableI2mCycleOnWRPort: cnl (pre-prod) */ + if (IS_CNL_REVID(dev_priv, CNL_REVID_B0, CNL_REVID_B0)) + WA_SET_BIT(GAMT_CHKN_BIT_REG, + GAMT_CHKN_DISABLE_I2M_CYCLE_ON_WR_PORT); + /* WaForceContextSaveRestoreNonCoherent:cnl */ WA_SET_BIT_MASKED(CNL_HDC_CHICKEN0, HDC_FORCE_CONTEXT_SAVE_RESTORE_NON_COHERENT); -- cgit v1.2.3 From 90cad095eeaa43711d60066096cddaf0d3b58bf6 Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Wed, 6 Sep 2017 16:28:59 +0100 Subject: drm/i915: Disable MI_STORE_DATA_IMM for i915g/i915gm MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The early gen3 machines (i915g/Grantsdale and i915gm/Alviso) share a lot of characteristics in their MI/GTT blocks with gen2, and in particular can only use physical addresses in MI_STORE_DATA_IMM. This makes it incompatible with our usage, so include those two machines in the blacklist to prevent usage. v2: Make it easy for gcc and rewrite it as a switch to save some space. Signed-off-by: Chris Wilson Reviewed-by: Ville Syrjälä #v1 Link: https://patchwork.freedesktop.org/patch/msgid/20170906152859.5304-1-chris@chris-wilson.co.uk --- drivers/gpu/drm/i915/i915_drv.h | 7 ------- drivers/gpu/drm/i915/i915_gem_execbuffer.c | 7 ++++--- drivers/gpu/drm/i915/intel_engine_cs.c | 15 +++++++++++++++ drivers/gpu/drm/i915/intel_ringbuffer.h | 12 +----------- 4 files changed, 20 insertions(+), 21 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_engine_cs.c') diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h index 789f7502cd1f..6020a94daf81 100644 --- a/drivers/gpu/drm/i915/i915_drv.h +++ b/drivers/gpu/drm/i915/i915_drv.h @@ -4360,11 +4360,4 @@ int remap_io_mapping(struct vm_area_struct *vma, unsigned long addr, unsigned long pfn, unsigned long size, struct io_mapping *iomap); -static inline bool -intel_engine_can_store_dword(struct intel_engine_cs *engine) -{ - return __intel_engine_can_store_dword(INTEL_GEN(engine->i915), - engine->class); -} - #endif diff --git a/drivers/gpu/drm/i915/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/i915_gem_execbuffer.c index 53718a245ded..7687483ff218 100644 --- a/drivers/gpu/drm/i915/i915_gem_execbuffer.c +++ b/drivers/gpu/drm/i915/i915_gem_execbuffer.c @@ -1168,6 +1168,9 @@ static u32 *reloc_gpu(struct i915_execbuffer *eb, if (eb_use_cmdparser(eb)) return ERR_PTR(-EWOULDBLOCK); + if (!intel_engine_can_store_dword(eb->engine)) + return ERR_PTR(-ENODEV); + err = __reloc_gpu_alloc(eb, vma, len); if (unlikely(err)) return ERR_PTR(err); @@ -1192,9 +1195,7 @@ relocate_entry(struct i915_vma *vma, if (!eb->reloc_cache.vaddr && (DBG_FORCE_RELOC == FORCE_GPU_RELOC || - !reservation_object_test_signaled_rcu(vma->resv, true)) && - __intel_engine_can_store_dword(eb->reloc_cache.gen, - eb->engine->class)) { + !reservation_object_test_signaled_rcu(vma->resv, true))) { const unsigned int gen = eb->reloc_cache.gen; unsigned int len; u32 *batch; diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c index ae668340620e..23812ec23969 100644 --- a/drivers/gpu/drm/i915/intel_engine_cs.c +++ b/drivers/gpu/drm/i915/intel_engine_cs.c @@ -1378,6 +1378,21 @@ void intel_engines_mark_idle(struct drm_i915_private *i915) } } +bool intel_engine_can_store_dword(struct intel_engine_cs *engine) +{ + switch (INTEL_GEN(engine->i915)) { + case 2: + return false; /* uses physical not virtual addresses */ + case 3: + /* maybe only uses physical not virtual addresses */ + return !(IS_I915G(engine->i915) || IS_I915GM(engine->i915)); + case 6: + return engine->class != VIDEO_DECODE_CLASS; /* b0rked */ + default: + return true; + } +} + #if IS_ENABLED(CONFIG_DRM_I915_SELFTEST) #include "selftests/mock_engine.c" #endif diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.h b/drivers/gpu/drm/i915/intel_ringbuffer.h index 02d8974bf9ab..79c0021f3700 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.h +++ b/drivers/gpu/drm/i915/intel_ringbuffer.h @@ -735,16 +735,6 @@ bool intel_engines_are_idle(struct drm_i915_private *dev_priv); void intel_engines_mark_idle(struct drm_i915_private *i915); void intel_engines_reset_default_submission(struct drm_i915_private *i915); -static inline bool -__intel_engine_can_store_dword(unsigned int gen, unsigned int class) -{ - if (gen <= 2) - return false; /* uses physical not virtual addresses */ - - if (gen == 6 && class == VIDEO_DECODE_CLASS) - return false; /* b0rked */ - - return true; -} +bool intel_engine_can_store_dword(struct intel_engine_cs *engine); #endif /* _INTEL_RINGBUFFER_H_ */ -- cgit v1.2.3 From aa9f4c4f19131e05323d0d5a068ccb46058e8257 Mon Sep 17 00:00:00 2001 From: Rodrigo Vivi Date: Wed, 6 Sep 2017 15:03:25 -0700 Subject: drm/i915/cnl: WaThrottleEUPerfToAvoidTDBackPressure:cnl(pre-prod) Wa for B-stepping only. A for a hang issue that requires throttling EU performace to 12.5% to avoid back pressure to thread dispatch v2: Rebased. No change from v1. Signed-off-by: Rodrigo Vivi Reviewed-by: Oscar Mateo Link: https://patchwork.freedesktop.org/patch/msgid/20170906220325.24524-1-rodrigo.vivi@intel.com --- drivers/gpu/drm/i915/i915_reg.h | 1 + drivers/gpu/drm/i915/intel_engine_cs.c | 4 ++++ 2 files changed, 5 insertions(+) (limited to 'drivers/gpu/drm/i915/intel_engine_cs.c') diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index 1878c4967529..2eff98cdcfad 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -8061,6 +8061,7 @@ enum { #define FLOW_CONTROL_ENABLE (1<<15) #define PARTIAL_INSTRUCTION_SHOOTDOWN_DISABLE (1<<8) #define STALL_DOP_GATING_DISABLE (1<<5) +#define THROTTLE_12_5 (7<<2) #define GEN7_ROW_CHICKEN2 _MMIO(0xe4f4) #define GEN7_ROW_CHICKEN2_GT2 _MMIO(0xf4f4) diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c index 23812ec23969..b8e9a234af2d 100644 --- a/drivers/gpu/drm/i915/intel_engine_cs.c +++ b/drivers/gpu/drm/i915/intel_engine_cs.c @@ -1079,6 +1079,10 @@ static int cnl_init_workarounds(struct intel_engine_cs *engine) WA_SET_BIT_MASKED(CNL_HDC_CHICKEN0, HDC_FORCE_CONTEXT_SAVE_RESTORE_NON_COHERENT); + /* WaThrottleEUPerfToAvoidTDBackPressure:cnl(pre-prod) */ + if (IS_CNL_REVID(dev_priv, CNL_REVID_B0, CNL_REVID_B0)) + WA_SET_BIT_MASKED(GEN8_ROW_CHICKEN, THROTTLE_12_5); + /* WaDisableReplayBufferBankArbitrationOptimization:cnl */ WA_SET_BIT_MASKED(COMMON_SLICE_CHICKEN2, GEN8_SBE_DISABLE_REPLAY_BUF_OPTIMIZATION); -- cgit v1.2.3 From efc886cb135595c7bec481a32d000dce865b7971 Mon Sep 17 00:00:00 2001 From: Oscar Mateo Date: Thu, 7 Sep 2017 08:40:04 -0700 Subject: drm/i915: Transform WaInPlaceDecompressionHang into a simple reg write MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Afaict, GEN9_GAMT_ECO_REG_RW_IA does not live in the context, so writing it on every context creation is overkill (and wrong). v2: Missing end parenthesis Cc: Chris Wilson Cc: Mika Kuoppala Cc: Rodrigo Vivi Signed-off-by: Oscar Mateo Reviewed-by: Michał Winiarski Link: https://patchwork.freedesktop.org/patch/msgid/1504798809-5653-1-git-send-email-oscar.mateo@intel.com Signed-off-by: Chris Wilson --- drivers/gpu/drm/i915/intel_engine_cs.c | 25 +++++++++++++++---------- 1 file changed, 15 insertions(+), 10 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_engine_cs.c') diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c index b8e9a234af2d..674d686286b1 100644 --- a/drivers/gpu/drm/i915/intel_engine_cs.c +++ b/drivers/gpu/drm/i915/intel_engine_cs.c @@ -985,8 +985,9 @@ static int skl_init_workarounds(struct intel_engine_cs *engine) /* WaInPlaceDecompressionHang:skl */ if (IS_SKL_REVID(dev_priv, SKL_REVID_H0, REVID_FOREVER)) - WA_SET_BIT(GEN9_GAMT_ECO_REG_RW_IA, - GAMT_ECO_ENABLE_IN_PLACE_DECOMPRESS); + I915_WRITE(GEN9_GAMT_ECO_REG_RW_IA, + (I915_READ(GEN9_GAMT_ECO_REG_RW_IA) | + GAMT_ECO_ENABLE_IN_PLACE_DECOMPRESS)); /* WaDisableLSQCROPERFforOCL:skl */ ret = wa_ring_whitelist_reg(engine, GEN8_L3SQCREG4); @@ -1059,8 +1060,9 @@ static int bxt_init_workarounds(struct intel_engine_cs *engine) /* WaInPlaceDecompressionHang:bxt */ if (IS_BXT_REVID(dev_priv, BXT_REVID_C0, REVID_FOREVER)) - WA_SET_BIT(GEN9_GAMT_ECO_REG_RW_IA, - GAMT_ECO_ENABLE_IN_PLACE_DECOMPRESS); + I915_WRITE(GEN9_GAMT_ECO_REG_RW_IA, + (I915_READ(GEN9_GAMT_ECO_REG_RW_IA) | + GAMT_ECO_ENABLE_IN_PLACE_DECOMPRESS)); return 0; } @@ -1093,8 +1095,9 @@ static int cnl_init_workarounds(struct intel_engine_cs *engine) GEN8_CSC2_SBE_VUE_CACHE_CONSERVATIVE); /* WaInPlaceDecompressionHang:cnl */ - WA_SET_BIT(GEN9_GAMT_ECO_REG_RW_IA, - GAMT_ECO_ENABLE_IN_PLACE_DECOMPRESS); + I915_WRITE(GEN9_GAMT_ECO_REG_RW_IA, + (I915_READ(GEN9_GAMT_ECO_REG_RW_IA) | + GAMT_ECO_ENABLE_IN_PLACE_DECOMPRESS)); /* WaPushConstantDereferenceHoldDisable:cnl */ WA_SET_BIT(GEN7_ROW_CHICKEN2, PUSH_CONSTANT_DEREF_DISABLE); @@ -1147,8 +1150,9 @@ static int kbl_init_workarounds(struct intel_engine_cs *engine) GEN7_SBE_SS_CACHE_DISPATCH_PORT_SHARING_DISABLE); /* WaInPlaceDecompressionHang:kbl */ - WA_SET_BIT(GEN9_GAMT_ECO_REG_RW_IA, - GAMT_ECO_ENABLE_IN_PLACE_DECOMPRESS); + I915_WRITE(GEN9_GAMT_ECO_REG_RW_IA, + (I915_READ(GEN9_GAMT_ECO_REG_RW_IA) | + GAMT_ECO_ENABLE_IN_PLACE_DECOMPRESS)); /* WaDisableLSQCROPERFforOCL:kbl */ ret = wa_ring_whitelist_reg(engine, GEN8_L3SQCREG4); @@ -1200,8 +1204,9 @@ static int cfl_init_workarounds(struct intel_engine_cs *engine) 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); + I915_WRITE(GEN9_GAMT_ECO_REG_RW_IA, + (I915_READ(GEN9_GAMT_ECO_REG_RW_IA) | + GAMT_ECO_ENABLE_IN_PLACE_DECOMPRESS)); return 0; } -- cgit v1.2.3 From 6cf20a0128fab5368937d719bcd9ec1b233491b7 Mon Sep 17 00:00:00 2001 From: Oscar Mateo Date: Thu, 7 Sep 2017 08:40:05 -0700 Subject: drm/i915: Transform WaDisableI2mCycleOnWRPort into a simple reg write MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit GAMT_CHKN_BIT_REG does not live in the context. Cc: Chris Wilson Cc: Mika Kuoppala Cc: Rodrigo Vivi Signed-off-by: Oscar Mateo Reviewed-by: Michał Winiarski Link: https://patchwork.freedesktop.org/patch/msgid/1504798809-5653-2-git-send-email-oscar.mateo@intel.com Signed-off-by: Chris Wilson --- drivers/gpu/drm/i915/intel_engine_cs.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_engine_cs.c') diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c index 674d686286b1..0fb012369077 100644 --- a/drivers/gpu/drm/i915/intel_engine_cs.c +++ b/drivers/gpu/drm/i915/intel_engine_cs.c @@ -1072,10 +1072,11 @@ static int cnl_init_workarounds(struct intel_engine_cs *engine) struct drm_i915_private *dev_priv = engine->i915; int ret; - /* WaDisableI2mCycleOnWRPort: cnl (pre-prod) */ + /* WaDisableI2mCycleOnWRPort:cnl (pre-prod) */ if (IS_CNL_REVID(dev_priv, CNL_REVID_B0, CNL_REVID_B0)) - WA_SET_BIT(GAMT_CHKN_BIT_REG, - GAMT_CHKN_DISABLE_I2M_CYCLE_ON_WR_PORT); + I915_WRITE(GAMT_CHKN_BIT_REG, + (I915_READ(GAMT_CHKN_BIT_REG) | + GAMT_CHKN_DISABLE_I2M_CYCLE_ON_WR_PORT)); /* WaForceContextSaveRestoreNonCoherent:cnl */ WA_SET_BIT_MASKED(CNL_HDC_CHICKEN0, -- cgit v1.2.3 From b27f59010f27628f1d96e746fdf7e6e916c451d5 Mon Sep 17 00:00:00 2001 From: Oscar Mateo Date: Thu, 7 Sep 2017 08:40:06 -0700 Subject: drm/i915: WaPushConstantDereferenceHoldDisable needs to modify a masked register MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit So do it correctly. Cc: Chris Wilson Cc: Mika Kuoppala Cc: Rodrigo Vivi Signed-off-by: Oscar Mateo Reviewed-by: Michał Winiarski Link: https://patchwork.freedesktop.org/patch/msgid/1504798809-5653-3-git-send-email-oscar.mateo@intel.com Signed-off-by: Chris Wilson --- drivers/gpu/drm/i915/intel_engine_cs.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'drivers/gpu/drm/i915/intel_engine_cs.c') diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c index 0fb012369077..6e24711c106c 100644 --- a/drivers/gpu/drm/i915/intel_engine_cs.c +++ b/drivers/gpu/drm/i915/intel_engine_cs.c @@ -1101,7 +1101,7 @@ static int cnl_init_workarounds(struct intel_engine_cs *engine) GAMT_ECO_ENABLE_IN_PLACE_DECOMPRESS)); /* WaPushConstantDereferenceHoldDisable:cnl */ - WA_SET_BIT(GEN7_ROW_CHICKEN2, PUSH_CONSTANT_DEREF_DISABLE); + WA_SET_BIT_MASKED(GEN7_ROW_CHICKEN2, PUSH_CONSTANT_DEREF_DISABLE); /* FtrEnableFastAnisoL1BankingFix: cnl */ WA_SET_BIT_MASKED(HALF_SLICE_CHICKEN3, CNL_FAST_ANISO_L1_BANKING_FIX); -- cgit v1.2.3 From 4827c547c51f81579edb8f3958d53c15e0a3f662 Mon Sep 17 00:00:00 2001 From: Oscar Mateo Date: Thu, 7 Sep 2017 08:40:07 -0700 Subject: drm/i915: Transform WaDisableGafsUnitClkGating into a simple reg write MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit GEN7_UCGCTL4 does not live in the context. v2: Missing parenthesis Cc: Chris Wilson Cc: Mika Kuoppala Cc: Rodrigo Vivi Signed-off-by: Oscar Mateo Reviewed-by: Michał Winiarski Link: https://patchwork.freedesktop.org/patch/msgid/1504798809-5653-4-git-send-email-oscar.mateo@intel.com Signed-off-by: Chris Wilson --- drivers/gpu/drm/i915/intel_engine_cs.c | 9 ++++++--- 1 file changed, 6 insertions(+), 3 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_engine_cs.c') diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c index 6e24711c106c..48c7eb36d633 100644 --- a/drivers/gpu/drm/i915/intel_engine_cs.c +++ b/drivers/gpu/drm/i915/intel_engine_cs.c @@ -981,7 +981,8 @@ static int skl_init_workarounds(struct intel_engine_cs *engine) GEN9_GAPS_TSV_CREDIT_DISABLE)); /* WaDisableGafsUnitClkGating:skl */ - WA_SET_BIT(GEN7_UCGCTL4, GEN8_EU_GAUNIT_CLOCK_GATE_DISABLE); + I915_WRITE(GEN7_UCGCTL4, (I915_READ(GEN7_UCGCTL4) | + GEN8_EU_GAUNIT_CLOCK_GATE_DISABLE)); /* WaInPlaceDecompressionHang:skl */ if (IS_SKL_REVID(dev_priv, SKL_REVID_H0, REVID_FOREVER)) @@ -1143,7 +1144,8 @@ static int kbl_init_workarounds(struct intel_engine_cs *engine) GEN8_SBE_DISABLE_REPLAY_BUF_OPTIMIZATION); /* WaDisableGafsUnitClkGating:kbl */ - WA_SET_BIT(GEN7_UCGCTL4, GEN8_EU_GAUNIT_CLOCK_GATE_DISABLE); + I915_WRITE(GEN7_UCGCTL4, (I915_READ(GEN7_UCGCTL4) | + GEN8_EU_GAUNIT_CLOCK_GATE_DISABLE)); /* WaDisableSbeCacheDispatchPortSharing:kbl */ WA_SET_BIT_MASKED( @@ -1197,7 +1199,8 @@ static int cfl_init_workarounds(struct intel_engine_cs *engine) GEN8_SBE_DISABLE_REPLAY_BUF_OPTIMIZATION); /* WaDisableGafsUnitClkGating:cfl */ - WA_SET_BIT(GEN7_UCGCTL4, GEN8_EU_GAUNIT_CLOCK_GATE_DISABLE); + I915_WRITE(GEN7_UCGCTL4, (I915_READ(GEN7_UCGCTL4) | + GEN8_EU_GAUNIT_CLOCK_GATE_DISABLE)); /* WaDisableSbeCacheDispatchPortSharing:cfl */ WA_SET_BIT_MASKED( -- cgit v1.2.3 From c6ea497c40b46aaf77d9018e27b26bffdbdac6f7 Mon Sep 17 00:00:00 2001 From: Oscar Mateo Date: Thu, 7 Sep 2017 08:40:08 -0700 Subject: drm/i915: Transform WaDisableDynamicCreditSharing into a simple register write MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit GAMT_CHKN_BIT_REG does not live in the context image. Cc: Chris Wilson Cc: Mika Kuoppala Cc: Rodrigo Vivi Signed-off-by: Oscar Mateo Reviewed-by: Michał Winiarski Link: https://patchwork.freedesktop.org/patch/msgid/1504798809-5653-5-git-send-email-oscar.mateo@intel.com Signed-off-by: Chris Wilson --- drivers/gpu/drm/i915/intel_engine_cs.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_engine_cs.c') diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c index 48c7eb36d633..0701ddb68b6a 100644 --- a/drivers/gpu/drm/i915/intel_engine_cs.c +++ b/drivers/gpu/drm/i915/intel_engine_cs.c @@ -1130,8 +1130,9 @@ static int kbl_init_workarounds(struct intel_engine_cs *engine) /* WaDisableDynamicCreditSharing:kbl */ if (IS_KBL_REVID(dev_priv, 0, KBL_REVID_B0)) - WA_SET_BIT(GAMT_CHKN_BIT_REG, - GAMT_CHKN_DISABLE_DYNAMIC_CREDIT_SHARING); + I915_WRITE(GAMT_CHKN_BIT_REG, + (I915_READ(GAMT_CHKN_BIT_REG) | + GAMT_CHKN_DISABLE_DYNAMIC_CREDIT_SHARING)); /* WaDisableFenceDestinationToSLM:kbl (pre-prod) */ if (IS_KBL_REVID(dev_priv, KBL_REVID_A0, KBL_REVID_A0)) -- cgit v1.2.3 From 212154ba4d1c31ad828b8e1e2d4ecd4adcb02796 Mon Sep 17 00:00:00 2001 From: Oscar Mateo Date: Thu, 7 Sep 2017 08:40:09 -0700 Subject: drm/i915: Transform WaDisablePooledEuLoadBalancingFix into a simple register write MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit FF_SLICE_CS_CHICKEN2 does not belong to the context image. Cc: Chris Wilson Cc: Mika Kuoppala Cc: Rodrigo Vivi Signed-off-by: Oscar Mateo Reviewed-by: Michał Winiarski Link: https://patchwork.freedesktop.org/patch/msgid/1504798809-5653-6-git-send-email-oscar.mateo@intel.com Signed-off-by: Chris Wilson --- drivers/gpu/drm/i915/intel_engine_cs.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_engine_cs.c') diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c index 0701ddb68b6a..3ae89a9d6241 100644 --- a/drivers/gpu/drm/i915/intel_engine_cs.c +++ b/drivers/gpu/drm/i915/intel_engine_cs.c @@ -1024,8 +1024,8 @@ static int bxt_init_workarounds(struct intel_engine_cs *engine) /* WaDisablePooledEuLoadBalancingFix:bxt */ if (IS_BXT_REVID(dev_priv, BXT_REVID_B0, REVID_FOREVER)) { - WA_SET_BIT_MASKED(FF_SLICE_CS_CHICKEN2, - GEN9_POOLED_EU_LOAD_BALANCING_FIX_DISABLE); + I915_WRITE(FF_SLICE_CS_CHICKEN2, + _MASKED_BIT_ENABLE(GEN9_POOLED_EU_LOAD_BALANCING_FIX_DISABLE)); } /* WaDisableSbeCacheDispatchPortSharing:bxt */ -- cgit v1.2.3 From 486e93f72abd41559b740f017717c7f6b3f8bb1f Mon Sep 17 00:00:00 2001 From: Daniele Ceraolo Spurio Date: Wed, 13 Sep 2017 09:56:02 +0100 Subject: drm/i915/lrc: allocate separate page for HWSP On gen8+ we're currently using the PPHWSP of the kernel ctx as the global HWSP. However, when the kernel ctx gets submitted (e.g. from __intel_autoenable_gt_powersave) the HW will use that page as both HWSP and PPHWSP. This causes a conflict in the register arena of the HWSP, i.e. dword indices below 0x30. We don't current utilize this arena, but in the following patches we will take advantage of the cached register state for handling execlist's context status interrupt. To avoid the conflict, instead of re-using the PPHWSP of the kernel ctx we can allocate a separate page for the HWSP like what happens for pre-execlists platform. v2: Add a use-case for the register arena of the HWSP. Signed-off-by: Daniele Ceraolo Spurio Cc: Michel Thierry Link: http://patchwork.freedesktop.org/patch/msgid/1499357440-34688-1-git-send-email-daniele.ceraolospurio@intel.com Signed-off-by: Chris Wilson Reviewed-by: Michel Thierry Link: https://patchwork.freedesktop.org/patch/msgid/20170913085605.18299-3-chris@chris-wilson.co.uk --- drivers/gpu/drm/i915/intel_engine_cs.c | 126 +++++++++++++++++++++++++++++++- drivers/gpu/drm/i915/intel_lrc.c | 34 +-------- drivers/gpu/drm/i915/intel_ringbuffer.c | 125 +------------------------------ 3 files changed, 127 insertions(+), 158 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_engine_cs.c') diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c index 3ae89a9d6241..8a5535ad6552 100644 --- a/drivers/gpu/drm/i915/intel_engine_cs.c +++ b/drivers/gpu/drm/i915/intel_engine_cs.c @@ -442,6 +442,114 @@ static void intel_engine_cleanup_scratch(struct intel_engine_cs *engine) i915_vma_unpin_and_release(&engine->scratch); } +static void cleanup_phys_status_page(struct intel_engine_cs *engine) +{ + struct drm_i915_private *dev_priv = engine->i915; + + if (!dev_priv->status_page_dmah) + return; + + drm_pci_free(&dev_priv->drm, dev_priv->status_page_dmah); + engine->status_page.page_addr = NULL; +} + +static void cleanup_status_page(struct intel_engine_cs *engine) +{ + struct i915_vma *vma; + struct drm_i915_gem_object *obj; + + vma = fetch_and_zero(&engine->status_page.vma); + if (!vma) + return; + + obj = vma->obj; + + i915_vma_unpin(vma); + i915_vma_close(vma); + + i915_gem_object_unpin_map(obj); + __i915_gem_object_release_unless_active(obj); +} + +static int init_status_page(struct intel_engine_cs *engine) +{ + struct drm_i915_gem_object *obj; + struct i915_vma *vma; + unsigned int flags; + void *vaddr; + int ret; + + obj = i915_gem_object_create_internal(engine->i915, PAGE_SIZE); + if (IS_ERR(obj)) { + DRM_ERROR("Failed to allocate status page\n"); + return PTR_ERR(obj); + } + + ret = i915_gem_object_set_cache_level(obj, I915_CACHE_LLC); + if (ret) + goto err; + + vma = i915_vma_instance(obj, &engine->i915->ggtt.base, NULL); + if (IS_ERR(vma)) { + ret = PTR_ERR(vma); + goto err; + } + + flags = PIN_GLOBAL; + if (!HAS_LLC(engine->i915)) + /* On g33, we cannot place HWS above 256MiB, so + * restrict its pinning to the low mappable arena. + * Though this restriction is not documented for + * gen4, gen5, or byt, they also behave similarly + * and hang if the HWS is placed at the top of the + * GTT. To generalise, it appears that all !llc + * platforms have issues with us placing the HWS + * above the mappable region (even though we never + * actually map it). + */ + flags |= PIN_MAPPABLE; + ret = i915_vma_pin(vma, 0, 4096, flags); + if (ret) + goto err; + + vaddr = i915_gem_object_pin_map(obj, I915_MAP_WB); + if (IS_ERR(vaddr)) { + ret = PTR_ERR(vaddr); + goto err_unpin; + } + + engine->status_page.vma = vma; + engine->status_page.ggtt_offset = i915_ggtt_offset(vma); + engine->status_page.page_addr = memset(vaddr, 0, PAGE_SIZE); + + DRM_DEBUG_DRIVER("%s hws offset: 0x%08x\n", + engine->name, i915_ggtt_offset(vma)); + return 0; + +err_unpin: + i915_vma_unpin(vma); +err: + i915_gem_object_put(obj); + return ret; +} + +static int init_phys_status_page(struct intel_engine_cs *engine) +{ + struct drm_i915_private *dev_priv = engine->i915; + + GEM_BUG_ON(engine->id != RCS); + + dev_priv->status_page_dmah = + drm_pci_alloc(&dev_priv->drm, PAGE_SIZE, PAGE_SIZE); + if (!dev_priv->status_page_dmah) + return -ENOMEM; + + engine->status_page.page_addr = dev_priv->status_page_dmah->vaddr; + memset(engine->status_page.page_addr, 0, PAGE_SIZE); + + return 0; +} + /** * intel_engines_init_common - initialize cengine state which might require hw access * @engine: Engine to initialize. @@ -477,10 +585,21 @@ int intel_engine_init_common(struct intel_engine_cs *engine) ret = i915_gem_render_state_init(engine); if (ret) - goto err_unpin; + goto err_breadcrumbs; + + if (HWS_NEEDS_PHYSICAL(engine->i915)) + ret = init_phys_status_page(engine); + else + ret = init_status_page(engine); + if (ret) + goto err_rs_fini; return 0; +err_rs_fini: + i915_gem_render_state_fini(engine); +err_breadcrumbs: + intel_engine_fini_breadcrumbs(engine); err_unpin: engine->context_unpin(engine, engine->i915->kernel_context); return ret; @@ -497,6 +616,11 @@ void intel_engine_cleanup_common(struct intel_engine_cs *engine) { intel_engine_cleanup_scratch(engine); + if (HWS_NEEDS_PHYSICAL(engine->i915)) + cleanup_phys_status_page(engine); + else + cleanup_status_page(engine); + i915_gem_render_state_fini(engine); intel_engine_fini_breadcrumbs(engine); intel_engine_cleanup_cmd_parser(engine); diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c index 3b8f1a79d640..8886e3b60e82 100644 --- a/drivers/gpu/drm/i915/intel_lrc.c +++ b/drivers/gpu/drm/i915/intel_lrc.c @@ -1674,11 +1674,6 @@ void intel_logical_ring_cleanup(struct intel_engine_cs *engine) if (engine->cleanup) engine->cleanup(engine); - if (engine->status_page.vma) { - i915_gem_object_unpin_map(engine->status_page.vma->obj); - engine->status_page.vma = NULL; - } - intel_engine_cleanup_common(engine); lrc_destroy_wa_ctx(engine); @@ -1725,24 +1720,6 @@ logical_ring_default_irqs(struct intel_engine_cs *engine) engine->irq_keep_mask = GT_CONTEXT_SWITCH_INTERRUPT << shift; } -static int -lrc_setup_hws(struct intel_engine_cs *engine, struct i915_vma *vma) -{ - const int hws_offset = LRC_PPHWSP_PN * PAGE_SIZE; - void *hws; - - /* The HWSP is part of the default context object in LRC mode. */ - hws = i915_gem_object_pin_map(vma->obj, I915_MAP_WB); - if (IS_ERR(hws)) - return PTR_ERR(hws); - - engine->status_page.page_addr = hws + hws_offset; - engine->status_page.ggtt_offset = i915_ggtt_offset(vma) + hws_offset; - engine->status_page.vma = vma; - - return 0; -} - static void logical_ring_setup(struct intel_engine_cs *engine) { @@ -1775,23 +1752,14 @@ logical_ring_setup(struct intel_engine_cs *engine) logical_ring_default_irqs(engine); } -static int -logical_ring_init(struct intel_engine_cs *engine) +static int logical_ring_init(struct intel_engine_cs *engine) { - struct i915_gem_context *dctx = engine->i915->kernel_context; int ret; ret = intel_engine_init_common(engine); if (ret) goto error; - /* And setup the hardware status page. */ - ret = lrc_setup_hws(engine, dctx->engine[engine->id].state); - if (ret) { - DRM_ERROR("Failed to set up hws %s: %d\n", engine->name, ret); - goto error; - } - return 0; error: diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index 268342433a8e..8af8871a8594 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -1175,113 +1175,7 @@ i915_emit_bb_start(struct drm_i915_gem_request *req, return 0; } -static void cleanup_phys_status_page(struct intel_engine_cs *engine) -{ - struct drm_i915_private *dev_priv = engine->i915; - - if (!dev_priv->status_page_dmah) - return; - - drm_pci_free(&dev_priv->drm, dev_priv->status_page_dmah); - engine->status_page.page_addr = NULL; -} - -static void cleanup_status_page(struct intel_engine_cs *engine) -{ - struct i915_vma *vma; - struct drm_i915_gem_object *obj; - - vma = fetch_and_zero(&engine->status_page.vma); - if (!vma) - return; - - obj = vma->obj; - - i915_vma_unpin(vma); - i915_vma_close(vma); - - i915_gem_object_unpin_map(obj); - __i915_gem_object_release_unless_active(obj); -} -static int init_status_page(struct intel_engine_cs *engine) -{ - struct drm_i915_gem_object *obj; - struct i915_vma *vma; - unsigned int flags; - void *vaddr; - int ret; - - obj = i915_gem_object_create_internal(engine->i915, PAGE_SIZE); - if (IS_ERR(obj)) { - DRM_ERROR("Failed to allocate status page\n"); - return PTR_ERR(obj); - } - - ret = i915_gem_object_set_cache_level(obj, I915_CACHE_LLC); - if (ret) - goto err; - - vma = i915_vma_instance(obj, &engine->i915->ggtt.base, NULL); - if (IS_ERR(vma)) { - ret = PTR_ERR(vma); - goto err; - } - - flags = PIN_GLOBAL; - if (!HAS_LLC(engine->i915)) - /* On g33, we cannot place HWS above 256MiB, so - * restrict its pinning to the low mappable arena. - * Though this restriction is not documented for - * gen4, gen5, or byt, they also behave similarly - * and hang if the HWS is placed at the top of the - * GTT. To generalise, it appears that all !llc - * platforms have issues with us placing the HWS - * above the mappable region (even though we never - * actualy map it). - */ - flags |= PIN_MAPPABLE; - ret = i915_vma_pin(vma, 0, 4096, flags); - if (ret) - goto err; - - vaddr = i915_gem_object_pin_map(obj, I915_MAP_WB); - if (IS_ERR(vaddr)) { - ret = PTR_ERR(vaddr); - goto err_unpin; - } - - engine->status_page.vma = vma; - engine->status_page.ggtt_offset = i915_ggtt_offset(vma); - engine->status_page.page_addr = memset(vaddr, 0, PAGE_SIZE); - - DRM_DEBUG_DRIVER("%s hws offset: 0x%08x\n", - engine->name, i915_ggtt_offset(vma)); - return 0; - -err_unpin: - i915_vma_unpin(vma); -err: - i915_gem_object_put(obj); - return ret; -} - -static int init_phys_status_page(struct intel_engine_cs *engine) -{ - struct drm_i915_private *dev_priv = engine->i915; - - GEM_BUG_ON(engine->id != RCS); - - dev_priv->status_page_dmah = - drm_pci_alloc(&dev_priv->drm, PAGE_SIZE, PAGE_SIZE); - if (!dev_priv->status_page_dmah) - return -ENOMEM; - - engine->status_page.page_addr = dev_priv->status_page_dmah->vaddr; - memset(engine->status_page.page_addr, 0, PAGE_SIZE); - - return 0; -} int intel_ring_pin(struct intel_ring *ring, struct drm_i915_private *i915, @@ -1568,17 +1462,10 @@ static int intel_init_ring_buffer(struct intel_engine_cs *engine) if (err) goto err; - if (HWS_NEEDS_PHYSICAL(engine->i915)) - err = init_phys_status_page(engine); - else - err = init_status_page(engine); - if (err) - goto err; - ring = intel_engine_create_ring(engine, 32 * PAGE_SIZE); if (IS_ERR(ring)) { err = PTR_ERR(ring); - goto err_hws; + goto err; } /* Ring wraparound at offset 0 sometimes hangs. No idea why. */ @@ -1593,11 +1480,6 @@ static int intel_init_ring_buffer(struct intel_engine_cs *engine) err_ring: intel_ring_free(ring); -err_hws: - if (HWS_NEEDS_PHYSICAL(engine->i915)) - cleanup_phys_status_page(engine); - else - cleanup_status_page(engine); err: intel_engine_cleanup_common(engine); return err; @@ -1616,11 +1498,6 @@ void intel_engine_cleanup(struct intel_engine_cs *engine) if (engine->cleanup) engine->cleanup(engine); - if (HWS_NEEDS_PHYSICAL(dev_priv)) - cleanup_phys_status_page(engine); - else - cleanup_status_page(engine); - intel_engine_cleanup_common(engine); dev_priv->engine[engine->id] = NULL; -- cgit v1.2.3 From 34a04e5e46cb984a6bab336484fa856574db332f Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Wed, 13 Sep 2017 09:56:03 +0100 Subject: drm/i915: Allow HW status page to be bound high MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit At the time of commit 1f767e02d69f ("drm/i915: HWS must be in the mappable region for g33"), drm_mm insertion would often default to placing a new object high in the zone forcing us to specify that certain HWSP must be bound within the low mappable region. Since then, drm_mm has gained more finesse over its placement and exposes that to the caller, commit 4e64e5539d15 ("drm: Improve drm_mm search (and fix topdown allocation) with rbtrees"). As such where possible we want the HWSP to be outside of the mappable aperture and so need to specify that can be pinned high. Signed-off-by: Chris Wilson Cc: Joonas Lahtinen Cc: Ville Syrjälä Cc: Michel Thierry Cc: Tvrtko Ursulin Cc: Mika Kuoppala Reviewed-by: Joonas Lahtinen Link: https://patchwork.freedesktop.org/patch/msgid/20170913085605.18299-4-chris@chris-wilson.co.uk --- drivers/gpu/drm/i915/intel_engine_cs.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'drivers/gpu/drm/i915/intel_engine_cs.c') diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c index 8a5535ad6552..32d35f32d289 100644 --- a/drivers/gpu/drm/i915/intel_engine_cs.c +++ b/drivers/gpu/drm/i915/intel_engine_cs.c @@ -508,6 +508,8 @@ static int init_status_page(struct intel_engine_cs *engine) * actually map it). */ flags |= PIN_MAPPABLE; + else + flags |= PIN_HIGH; ret = i915_vma_pin(vma, 0, 4096, flags); if (ret) goto err; -- cgit v1.2.3 From 93564044fb2c938e8ee4a91323157ded28072972 Mon Sep 17 00:00:00 2001 From: Ville Syrjälä Date: Thu, 24 Aug 2017 22:10:51 +0300 Subject: drm/i915: Switch over to the LLC/eLLC hotspot avoidance hash mode for CCS MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Use the LLC/eLLC hotspot avoidance mode for CCS on LLC machines. This is reported to give better performance. Testing has indicated that we don't need to enforce any massive 2 or 4 MiB alignment for all compressed resources even though there are still plenty of stale comments in the spec suggesting that we do. We do need to make sure every hardware unit that deals with the compressed data uses the same hash mode. Cc: Ben Widawsky Cc: Jason Ekstrand Cc: Daniel Stone Signed-off-by: Ville Syrjälä Link: https://patchwork.freedesktop.org/patch/msgid/20170824191100.10949-4-ville.syrjala@linux.intel.com Reviewed-by: Ben Widawsky --- drivers/gpu/drm/i915/i915_reg.h | 8 +++++++- drivers/gpu/drm/i915/intel_engine_cs.c | 13 +++++++++++++ drivers/gpu/drm/i915/intel_pm.c | 27 +++++++++++++-------------- 3 files changed, 33 insertions(+), 15 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_engine_cs.c') diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index 9f03cd063afe..0befefec7327 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -6913,7 +6913,7 @@ enum { # define CHICKEN3_DGMG_DONE_FIX_DISABLE (1 << 2) #define CHICKEN_PAR1_1 _MMIO(0x42080) -#define SKL_RC_HASH_OUTSIDE (1 << 15) +#define SKL_DE_COMPRESSED_HASH_MODE (1 << 15) #define DPA_MASK_VBLANK_SRD (1 << 15) #define FORCE_ARB_IDLE_PLANES (1 << 14) #define SKL_EDP_PSR_FIX_RDWRAP (1 << 3) @@ -6991,6 +6991,7 @@ enum { # define GEN7_CSC1_RHWO_OPT_DISABLE_IN_RCC ((1<<10) | (1<<26)) # define GEN9_RHWO_OPTIMIZATION_DISABLE (1<<14) #define COMMON_SLICE_CHICKEN2 _MMIO(0x7014) +# define GEN9_PBE_COMPRESSED_HASH_SELECTION (1<<13) # define GEN9_DISABLE_GATHER_AT_SET_SHADER_COMMON_SLICE (1<<12) # define GEN8_SBE_DISABLE_REPLAY_BUF_OPTIMIZATION (1<<8) # define GEN8_CSC2_SBE_VUE_CACHE_CONSERVATIVE (1<<0) @@ -8085,6 +8086,7 @@ enum { #define GEN8_SAMPLER_POWER_BYPASS_DIS (1<<1) #define GEN9_HALF_SLICE_CHICKEN7 _MMIO(0xe194) +#define GEN9_SAMPLER_HASH_COMPRESSED_READ_ADDR (1<<8) #define GEN9_ENABLE_YV12_BUGFIX (1<<4) #define GEN9_ENABLE_GPGPU_PREEMPTION (1<<2) @@ -9385,4 +9387,8 @@ enum skl_power_gate { #define GEN9_L3_LRA_1_GPGPU_DEFAULT_VALUE_SKL 0x67F1427F /* " " */ #define GEN9_L3_LRA_1_GPGPU_DEFAULT_VALUE_BXT 0x5FF101FF /* " " */ +#define MMCD_MISC_CTRL _MMIO(0x4ddc) /* skl+ */ +#define MMCD_PCLA (1 << 31) +#define MMCD_HOTSPOT_EN (1 << 27) + #endif /* _I915_REG_H_ */ diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c index 32d35f32d289..3d135c3cd380 100644 --- a/drivers/gpu/drm/i915/intel_engine_cs.c +++ b/drivers/gpu/drm/i915/intel_engine_cs.c @@ -938,6 +938,19 @@ static int gen9_init_workarounds(struct intel_engine_cs *engine) I915_WRITE(GAM_ECOCHK, I915_READ(GAM_ECOCHK) | ECOCHK_DIS_TLB); + if (HAS_LLC(dev_priv)) { + /* WaCompressedResourceSamplerPbeMediaNewHashMode:skl,kbl + * + * Must match Display Engine. See + * WaCompressedResourceDisplayNewHashMode. + */ + WA_SET_BIT_MASKED(COMMON_SLICE_CHICKEN2, + GEN9_PBE_COMPRESSED_HASH_SELECTION); + WA_SET_BIT_MASKED(GEN9_HALF_SLICE_CHICKEN7, + GEN9_SAMPLER_HASH_COMPRESSED_READ_ADDR); + WA_SET_BIT(MMCD_MISC_CTRL, MMCD_PCLA | MMCD_HOTSPOT_EN); + } + /* WaClearFlowControlGpgpuContextSave:skl,bxt,kbl,glk,cfl */ /* WaDisablePartialInstShootdown:skl,bxt,kbl,glk,cfl */ WA_SET_BIT_MASKED(GEN8_ROW_CHICKEN, diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c index fa9055a4f790..94624ede3479 100644 --- a/drivers/gpu/drm/i915/intel_pm.c +++ b/drivers/gpu/drm/i915/intel_pm.c @@ -58,24 +58,23 @@ static void gen9_init_clock_gating(struct drm_i915_private *dev_priv) { + if (HAS_LLC(dev_priv)) { + /* + * WaCompressedResourceDisplayNewHashMode:skl,kbl + * Display WA#0390: skl,kbl + * + * Must match Sampler, Pixel Back End, and Media. See + * WaCompressedResourceSamplerPbeMediaNewHashMode. + */ + I915_WRITE(CHICKEN_PAR1_1, + I915_READ(CHICKEN_PAR1_1) | + SKL_DE_COMPRESSED_HASH_MODE); + } + /* 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); - /* - * Display WA#0390: skl,bxt,kbl,glk - * - * Must match Sampler, Pixel Back End, and Media - * (0xE194 bit 8, 0x7014 bit 13, 0x4DDC bits 27 and 31). - * - * Including bits outside the page in the hash would - * require 2 (or 4?) MiB alignment of resources. Just - * assume the defaul hashing mode which only uses bits - * within the page. - */ - I915_WRITE(CHICKEN_PAR1_1, - I915_READ(CHICKEN_PAR1_1) & ~SKL_RC_HASH_OUTSIDE); - I915_WRITE(GEN8_CONFIG0, I915_READ(GEN8_CONFIG0) | GEN9_DEFAULT_FIXES); -- cgit v1.2.3 From 4f044a88a86adb4c8cc6cb1a7303bb9c61ea2caa Mon Sep 17 00:00:00 2001 From: Michal Wajdeczko Date: Tue, 19 Sep 2017 19:38:44 +0000 Subject: drm/i915: Rename global i915 to i915_modparams Our global struct with params is named exactly the same way as new preferred name for the drm_i915_private function parameter. To avoid such name reuse lets use different name for the global. v5: pure rename v6: fix Credits-to: Coccinelle @@ identifier n; @@ ( - i915.n + i915_modparams.n ) Signed-off-by: Michal Wajdeczko Cc: Jani Nikula Cc: Chris Wilson Cc: Tvrtko Ursulin Cc: Ville Syrjala Cc: Joonas Lahtinen Acked-by: Joonas Lahtinen Signed-off-by: Jani Nikula Link: https://patchwork.freedesktop.org/patch/msgid/20170919193846.38060-1-michal.wajdeczko@intel.com --- drivers/gpu/drm/i915/gvt/render.c | 2 +- drivers/gpu/drm/i915/i915_debugfs.c | 14 ++++---- drivers/gpu/drm/i915/i915_drv.c | 34 ++++++++++-------- drivers/gpu/drm/i915/i915_drv.h | 10 +++--- drivers/gpu/drm/i915/i915_gem.c | 4 +-- drivers/gpu/drm/i915/i915_gem_context.c | 12 +++---- drivers/gpu/drm/i915/i915_gem_execbuffer.c | 2 +- drivers/gpu/drm/i915/i915_gem_gtt.c | 6 ++-- drivers/gpu/drm/i915/i915_gpu_error.c | 6 ++-- drivers/gpu/drm/i915/i915_guc_submission.c | 2 +- drivers/gpu/drm/i915/i915_irq.c | 2 +- drivers/gpu/drm/i915/i915_params.c | 6 ++-- drivers/gpu/drm/i915/i915_params.h | 2 +- drivers/gpu/drm/i915/i915_pci.c | 6 ++-- drivers/gpu/drm/i915/i915_perf.c | 6 ++-- drivers/gpu/drm/i915/intel_bios.c | 7 ++-- drivers/gpu/drm/i915/intel_crt.c | 4 +-- drivers/gpu/drm/i915/intel_device_info.c | 2 +- drivers/gpu/drm/i915/intel_display.c | 12 +++---- drivers/gpu/drm/i915/intel_dp.c | 4 +-- drivers/gpu/drm/i915/intel_dp_aux_backlight.c | 2 +- drivers/gpu/drm/i915/intel_drv.h | 2 +- drivers/gpu/drm/i915/intel_engine_cs.c | 4 +-- drivers/gpu/drm/i915/intel_fbc.c | 11 +++--- drivers/gpu/drm/i915/intel_guc_loader.c | 13 +++---- drivers/gpu/drm/i915/intel_guc_log.c | 26 +++++++------- drivers/gpu/drm/i915/intel_gvt.c | 12 +++---- drivers/gpu/drm/i915/intel_hangcheck.c | 2 +- drivers/gpu/drm/i915/intel_huc.c | 4 +-- drivers/gpu/drm/i915/intel_lrc.c | 4 +-- drivers/gpu/drm/i915/intel_lvds.c | 4 +-- drivers/gpu/drm/i915/intel_opregion.c | 2 +- drivers/gpu/drm/i915/intel_panel.c | 8 ++--- drivers/gpu/drm/i915/intel_pm.c | 6 ++-- drivers/gpu/drm/i915/intel_psr.c | 10 +++--- drivers/gpu/drm/i915/intel_ringbuffer.c | 8 ++--- drivers/gpu/drm/i915/intel_runtime_pm.c | 17 ++++----- drivers/gpu/drm/i915/intel_uc.c | 51 ++++++++++++++------------- drivers/gpu/drm/i915/intel_uncore.c | 22 ++++++------ 39 files changed, 182 insertions(+), 169 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_engine_cs.c') diff --git a/drivers/gpu/drm/i915/gvt/render.c b/drivers/gpu/drm/i915/gvt/render.c index 2ea542257f03..6d066cf35478 100644 --- a/drivers/gpu/drm/i915/gvt/render.c +++ b/drivers/gpu/drm/i915/gvt/render.c @@ -293,7 +293,7 @@ static void switch_mmio_to_vgpu(struct intel_vgpu *vgpu, int ring_id) */ if (mmio->in_context && ((ctx_ctrl & inhibit_mask) != inhibit_mask) && - i915.enable_execlists) + i915_modparams.enable_execlists) continue; if (mmio->mask) diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c index 2518bdf95eef..b08ebed4e700 100644 --- a/drivers/gpu/drm/i915/i915_debugfs.c +++ b/drivers/gpu/drm/i915/i915_debugfs.c @@ -66,7 +66,7 @@ static int i915_capabilities(struct seq_file *m, void *data) #undef PRINT_FLAG kernel_param_lock(THIS_MODULE); -#define PRINT_PARAM(T, x) seq_print_param(m, #x, #T, &i915.x); +#define PRINT_PARAM(T, x) seq_print_param(m, #x, #T, &i915_modparams.x); I915_PARAMS_FOR_EACH(PRINT_PARAM); #undef PRINT_PARAM kernel_param_unlock(THIS_MODULE); @@ -1266,7 +1266,7 @@ static int i915_hangcheck_info(struct seq_file *m, void *unused) if (waitqueue_active(&dev_priv->gpu_error.reset_queue)) seq_puts(m, "struct_mutex blocked for reset\n"); - if (!i915.enable_hangcheck) { + if (!i915_modparams.enable_hangcheck) { seq_puts(m, "Hangcheck disabled\n"); return 0; } @@ -1701,7 +1701,7 @@ static int i915_ips_status(struct seq_file *m, void *unused) intel_runtime_pm_get(dev_priv); seq_printf(m, "Enabled by kernel parameter: %s\n", - yesno(i915.enable_ips)); + yesno(i915_modparams.enable_ips)); if (INTEL_GEN(dev_priv) >= 8) { seq_puts(m, "Currently: unknown\n"); @@ -2016,7 +2016,7 @@ static int i915_dump_lrc(struct seq_file *m, void *unused) enum intel_engine_id id; int ret; - if (!i915.enable_execlists) { + if (!i915_modparams.enable_execlists) { seq_printf(m, "Logical Ring Contexts are disabled\n"); return 0; } @@ -2592,7 +2592,7 @@ static int i915_guc_log_control_get(void *data, u64 *val) if (!dev_priv->guc.log.vma) return -EINVAL; - *val = i915.guc_log_level; + *val = i915_modparams.guc_log_level; return 0; } @@ -3310,7 +3310,7 @@ static int i915_engine_info(struct seq_file *m, void *unused) seq_printf(m, "\tBBADDR: 0x%08x_%08x\n", upper_32_bits(addr), lower_32_bits(addr)); - if (i915.enable_execlists) { + if (i915_modparams.enable_execlists) { const u32 *hws = &engine->status_page.page_addr[I915_HWS_CSB_BUF0_INDEX]; u32 ptr, read, write; unsigned int idx; @@ -3406,7 +3406,7 @@ static int i915_semaphore_status(struct seq_file *m, void *unused) enum intel_engine_id id; int j, ret; - if (!i915.semaphores) { + if (!i915_modparams.semaphores) { seq_puts(m, "Semaphores are disabled\n"); return 0; } diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c index 5c111ea96e80..7056bb299dc6 100644 --- a/drivers/gpu/drm/i915/i915_drv.c +++ b/drivers/gpu/drm/i915/i915_drv.c @@ -58,12 +58,12 @@ static unsigned int i915_load_fail_count; bool __i915_inject_load_failure(const char *func, int line) { - if (i915_load_fail_count >= i915.inject_load_failure) + if (i915_load_fail_count >= i915_modparams.inject_load_failure) return false; - if (++i915_load_fail_count == i915.inject_load_failure) { + if (++i915_load_fail_count == i915_modparams.inject_load_failure) { DRM_INFO("Injecting failure at checkpoint %u [%s:%d]\n", - i915.inject_load_failure, func, line); + i915_modparams.inject_load_failure, func, line); return true; } @@ -106,8 +106,8 @@ __i915_printk(struct drm_i915_private *dev_priv, const char *level, static bool i915_error_injected(struct drm_i915_private *dev_priv) { - return i915.inject_load_failure && - i915_load_fail_count == i915.inject_load_failure; + return i915_modparams.inject_load_failure && + i915_load_fail_count == i915_modparams.inject_load_failure; } #define i915_load_error(dev_priv, fmt, ...) \ @@ -321,7 +321,7 @@ static int i915_getparam(struct drm_device *dev, void *data, value = USES_PPGTT(dev_priv); break; case I915_PARAM_HAS_SEMAPHORES: - value = i915.semaphores; + value = i915_modparams.semaphores; break; case I915_PARAM_HAS_SECURE_BATCHES: value = capable(CAP_SYS_ADMIN); @@ -340,7 +340,8 @@ static int i915_getparam(struct drm_device *dev, void *data, return -ENODEV; break; case I915_PARAM_HAS_GPU_RESET: - value = i915.enable_hangcheck && intel_has_gpu_reset(dev_priv); + value = i915_modparams.enable_hangcheck && + intel_has_gpu_reset(dev_priv); if (value && intel_has_reset_engine(dev_priv)) value = 2; break; @@ -1031,9 +1032,9 @@ static void i915_driver_cleanup_mmio(struct drm_i915_private *dev_priv) static void intel_sanitize_options(struct drm_i915_private *dev_priv) { - i915.enable_execlists = + i915_modparams.enable_execlists = intel_sanitize_enable_execlists(dev_priv, - i915.enable_execlists); + i915_modparams.enable_execlists); /* * i915.enable_ppgtt is read-only, so do an early pass to validate the @@ -1041,12 +1042,15 @@ static void intel_sanitize_options(struct drm_i915_private *dev_priv) * do this now so that we can print out any log messages once rather * than every time we check intel_enable_ppgtt(). */ - i915.enable_ppgtt = - intel_sanitize_enable_ppgtt(dev_priv, i915.enable_ppgtt); - DRM_DEBUG_DRIVER("ppgtt mode: %i\n", i915.enable_ppgtt); + i915_modparams.enable_ppgtt = + intel_sanitize_enable_ppgtt(dev_priv, + i915_modparams.enable_ppgtt); + DRM_DEBUG_DRIVER("ppgtt mode: %i\n", i915_modparams.enable_ppgtt); - i915.semaphores = intel_sanitize_semaphores(dev_priv, i915.semaphores); - DRM_DEBUG_DRIVER("use GPU semaphores? %s\n", yesno(i915.semaphores)); + i915_modparams.semaphores = + intel_sanitize_semaphores(dev_priv, i915_modparams.semaphores); + DRM_DEBUG_DRIVER("use GPU semaphores? %s\n", + yesno(i915_modparams.semaphores)); intel_uc_sanitize_options(dev_priv); @@ -1277,7 +1281,7 @@ int i915_driver_load(struct pci_dev *pdev, const struct pci_device_id *ent) int ret; /* Enable nuclear pageflip on ILK+ */ - if (!i915.nuclear_pageflip && match_info->gen < 5) + if (!i915_modparams.nuclear_pageflip && match_info->gen < 5) driver.driver_features &= ~DRIVER_ATOMIC; ret = -ENOMEM; diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h index 08e3ae15b52e..d6babe7d362d 100644 --- a/drivers/gpu/drm/i915/i915_drv.h +++ b/drivers/gpu/drm/i915/i915_drv.h @@ -93,7 +93,7 @@ #define I915_STATE_WARN(condition, format...) ({ \ int __ret_warn_on = !!(condition); \ if (unlikely(__ret_warn_on)) \ - if (!WARN(i915.verbose_state_checks, format)) \ + if (!WARN(i915_modparams.verbose_state_checks, format)) \ DRM_ERROR(format); \ unlikely(__ret_warn_on); \ }) @@ -3076,9 +3076,9 @@ intel_info(const struct drm_i915_private *dev_priv) #define HAS_LOGICAL_RING_CONTEXTS(dev_priv) \ ((dev_priv)->info.has_logical_ring_contexts) -#define USES_PPGTT(dev_priv) (i915.enable_ppgtt) -#define USES_FULL_PPGTT(dev_priv) (i915.enable_ppgtt >= 2) -#define USES_FULL_48BIT_PPGTT(dev_priv) (i915.enable_ppgtt == 3) +#define USES_PPGTT(dev_priv) (i915_modparams.enable_ppgtt) +#define USES_FULL_PPGTT(dev_priv) (i915_modparams.enable_ppgtt >= 2) +#define USES_FULL_48BIT_PPGTT(dev_priv) (i915_modparams.enable_ppgtt == 3) #define HAS_OVERLAY(dev_priv) ((dev_priv)->info.has_overlay) #define OVERLAY_NEEDS_PHYSICAL(dev_priv) \ @@ -3279,7 +3279,7 @@ static inline void i915_queue_hangcheck(struct drm_i915_private *dev_priv) { unsigned long delay; - if (unlikely(!i915.enable_hangcheck)) + if (unlikely(!i915_modparams.enable_hangcheck)) return; /* Don't continually defer the hangcheck so that it is always run at diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c index b0bbf8729dae..12ce97d47afb 100644 --- a/drivers/gpu/drm/i915/i915_gem.c +++ b/drivers/gpu/drm/i915/i915_gem.c @@ -4738,7 +4738,7 @@ bool intel_sanitize_semaphores(struct drm_i915_private *dev_priv, int value) return false; /* TODO: make semaphores and Execlists play nicely together */ - if (i915.enable_execlists) + if (i915_modparams.enable_execlists) return false; if (value >= 0) @@ -4759,7 +4759,7 @@ int i915_gem_init(struct drm_i915_private *dev_priv) dev_priv->mm.unordered_timeline = dma_fence_context_alloc(1); - if (!i915.enable_execlists) { + if (!i915_modparams.enable_execlists) { dev_priv->gt.resume = intel_legacy_submission_resume; dev_priv->gt.cleanup_engine = intel_engine_cleanup; } else { diff --git a/drivers/gpu/drm/i915/i915_gem_context.c b/drivers/gpu/drm/i915/i915_gem_context.c index 58a2a44f88bd..921ee369c74d 100644 --- a/drivers/gpu/drm/i915/i915_gem_context.c +++ b/drivers/gpu/drm/i915/i915_gem_context.c @@ -314,7 +314,7 @@ __create_hw_context(struct drm_i915_private *dev_priv, * present or not in use we still need a small bias as ring wraparound * at offset 0 sometimes hangs. No idea why. */ - if (HAS_GUC(dev_priv) && i915.enable_guc_loading) + if (HAS_GUC(dev_priv) && i915_modparams.enable_guc_loading) ctx->ggtt_offset_bias = GUC_WOPCM_TOP; else ctx->ggtt_offset_bias = I915_GTT_PAGE_SIZE; @@ -407,7 +407,7 @@ i915_gem_context_create_gvt(struct drm_device *dev) i915_gem_context_set_closed(ctx); /* not user accessible */ i915_gem_context_clear_bannable(ctx); i915_gem_context_set_force_single_submission(ctx); - if (!i915.enable_guc_submission) + if (!i915_modparams.enable_guc_submission) ctx->ring_size = 512 * PAGE_SIZE; /* Max ring buffer size */ GEM_BUG_ON(i915_gem_context_is_kernel(ctx)); @@ -431,7 +431,7 @@ int i915_gem_contexts_init(struct drm_i915_private *dev_priv) if (intel_vgpu_active(dev_priv) && HAS_LOGICAL_RING_CONTEXTS(dev_priv)) { - if (!i915.enable_execlists) { + if (!i915_modparams.enable_execlists) { DRM_INFO("Only EXECLIST mode is supported in vgpu.\n"); return -EINVAL; } @@ -483,7 +483,7 @@ void i915_gem_contexts_lost(struct drm_i915_private *dev_priv) } /* Force the GPU state to be restored on enabling */ - if (!i915.enable_execlists) { + if (!i915_modparams.enable_execlists) { struct i915_gem_context *ctx; list_for_each_entry(ctx, &dev_priv->contexts.list, link) { @@ -568,7 +568,7 @@ mi_set_context(struct drm_i915_gem_request *req, u32 flags) enum intel_engine_id id; const int num_rings = /* Use an extended w/a on gen7 if signalling from other rings */ - (i915.semaphores && INTEL_GEN(dev_priv) == 7) ? + (i915_modparams.semaphores && INTEL_GEN(dev_priv) == 7) ? INTEL_INFO(dev_priv)->num_rings - 1 : 0; int len; @@ -837,7 +837,7 @@ int i915_switch_context(struct drm_i915_gem_request *req) struct intel_engine_cs *engine = req->engine; lockdep_assert_held(&req->i915->drm.struct_mutex); - if (i915.enable_execlists) + if (i915_modparams.enable_execlists) return 0; if (!req->ctx->engine[engine->id].state) { diff --git a/drivers/gpu/drm/i915/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/i915_gem_execbuffer.c index 61b9b079c8c8..fa18677cdb54 100644 --- a/drivers/gpu/drm/i915/i915_gem_execbuffer.c +++ b/drivers/gpu/drm/i915/i915_gem_execbuffer.c @@ -1585,7 +1585,7 @@ static int eb_prefault_relocations(const struct i915_execbuffer *eb) const unsigned int count = eb->buffer_count; unsigned int i; - if (unlikely(i915.prefault_disable)) + if (unlikely(i915_modparams.prefault_disable)) return 0; for (i = 0; i < count; i++) { diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c index 37cd0860fc29..ecb5e8cd37ba 100644 --- a/drivers/gpu/drm/i915/i915_gem_gtt.c +++ b/drivers/gpu/drm/i915/i915_gem_gtt.c @@ -180,7 +180,7 @@ int intel_sanitize_enable_ppgtt(struct drm_i915_private *dev_priv, return 0; } - if (INTEL_GEN(dev_priv) >= 8 && i915.enable_execlists) { + if (INTEL_GEN(dev_priv) >= 8 && i915_modparams.enable_execlists) { if (has_full_48bit_ppgtt) return 3; @@ -1972,7 +1972,7 @@ int i915_ppgtt_init_hw(struct drm_i915_private *dev_priv) /* In the case of execlists, PPGTT is enabled by the context descriptor * and the PDPs are contained within the context itself. We don't * need to do anything here. */ - if (i915.enable_execlists) + if (i915_modparams.enable_execlists) return 0; if (!USES_PPGTT(dev_priv)) @@ -3292,7 +3292,7 @@ int i915_ggtt_probe_hw(struct drm_i915_private *dev_priv) * currently don't have any bits spare to pass in this upper * restriction! */ - if (HAS_GUC(dev_priv) && i915.enable_guc_loading) { + if (HAS_GUC(dev_priv) && i915_modparams.enable_guc_loading) { ggtt->base.total = min_t(u64, ggtt->base.total, GUC_GGTT_TOP); ggtt->mappable_end = min(ggtt->mappable_end, ggtt->base.total); } diff --git a/drivers/gpu/drm/i915/i915_gpu_error.c b/drivers/gpu/drm/i915/i915_gpu_error.c index ed5a1eb839ad..6cd5eba643e8 100644 --- a/drivers/gpu/drm/i915/i915_gpu_error.c +++ b/drivers/gpu/drm/i915/i915_gpu_error.c @@ -1554,7 +1554,7 @@ static void i915_gem_capture_guc_log_buffer(struct drm_i915_private *dev_priv, struct i915_gpu_state *error) { /* Capturing log buf contents won't be useful if logging was disabled */ - if (!dev_priv->guc.log.vma || (i915.guc_log_level < 0)) + if (!dev_priv->guc.log.vma || (i915_modparams.guc_log_level < 0)) return; error->guc_log = i915_error_object_create(dev_priv, @@ -1696,7 +1696,7 @@ static int capture(void *data) ktime_to_timeval(ktime_sub(ktime_get(), error->i915->gt.last_init_time)); - error->params = i915; + error->params = i915_modparams; #define DUP(T, x) dup_param(#T, &error->params.x); I915_PARAMS_FOR_EACH(DUP); #undef DUP @@ -1751,7 +1751,7 @@ void i915_capture_error_state(struct drm_i915_private *dev_priv, struct i915_gpu_state *error; unsigned long flags; - if (!i915.error_capture) + if (!i915_modparams.error_capture) return; if (READ_ONCE(dev_priv->gpu_error.first_error)) diff --git a/drivers/gpu/drm/i915/i915_guc_submission.c b/drivers/gpu/drm/i915/i915_guc_submission.c index e191d56fc990..06a26c610806 100644 --- a/drivers/gpu/drm/i915/i915_guc_submission.c +++ b/drivers/gpu/drm/i915/i915_guc_submission.c @@ -1245,7 +1245,7 @@ int intel_guc_resume(struct drm_i915_private *dev_priv) if (guc->fw.load_status != INTEL_UC_FIRMWARE_SUCCESS) return 0; - if (i915.guc_log_level >= 0) + if (i915_modparams.guc_log_level >= 0) gen9_enable_guc_interrupts(dev_priv); ctx = dev_priv->kernel_context; diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c index bd38c6983eec..b1bab7605db9 100644 --- a/drivers/gpu/drm/i915/i915_irq.c +++ b/drivers/gpu/drm/i915/i915_irq.c @@ -1357,7 +1357,7 @@ gen8_cs_irq_handler(struct intel_engine_cs *engine, u32 iir, int test_shift) if (iir & (GT_RENDER_USER_INTERRUPT << test_shift)) { notify_ring(engine); - tasklet |= i915.enable_guc_submission; + tasklet |= i915_modparams.enable_guc_submission; } if (tasklet) diff --git a/drivers/gpu/drm/i915/i915_params.c b/drivers/gpu/drm/i915/i915_params.c index ddda513cc7f4..ec6534180d54 100644 --- a/drivers/gpu/drm/i915/i915_params.c +++ b/drivers/gpu/drm/i915/i915_params.c @@ -26,13 +26,13 @@ #include "i915_drv.h" #define i915_param_named(name, T, perm, desc) \ - module_param_named(name, i915.name, T, perm); \ + module_param_named(name, i915_modparams.name, T, perm); \ MODULE_PARM_DESC(name, desc) #define i915_param_named_unsafe(name, T, perm, desc) \ - module_param_named_unsafe(name, i915.name, T, perm); \ + module_param_named_unsafe(name, i915_modparams.name, T, perm); \ MODULE_PARM_DESC(name, desc) -struct i915_params i915 __read_mostly = { +struct i915_params i915_modparams __read_mostly = { .modeset = -1, .panel_ignore_lid = 1, .semaphores = -1, diff --git a/drivers/gpu/drm/i915/i915_params.h b/drivers/gpu/drm/i915/i915_params.h index ac844709c97e..a2cbb4782fcd 100644 --- a/drivers/gpu/drm/i915/i915_params.h +++ b/drivers/gpu/drm/i915/i915_params.h @@ -76,7 +76,7 @@ struct i915_params { }; #undef MEMBER -extern struct i915_params i915 __read_mostly; +extern struct i915_params i915_modparams __read_mostly; #endif diff --git a/drivers/gpu/drm/i915/i915_pci.c b/drivers/gpu/drm/i915/i915_pci.c index ce2c08eb9890..da60866b6628 100644 --- a/drivers/gpu/drm/i915/i915_pci.c +++ b/drivers/gpu/drm/i915/i915_pci.c @@ -631,7 +631,7 @@ static int i915_pci_probe(struct pci_dev *pdev, const struct pci_device_id *ent) (struct intel_device_info *) ent->driver_data; int err; - if (IS_ALPHA_SUPPORT(intel_info) && !i915.alpha_support) { + if (IS_ALPHA_SUPPORT(intel_info) && !i915_modparams.alpha_support) { DRM_INFO("The driver support for your hardware in this kernel version is alpha quality\n" "See CONFIG_DRM_I915_ALPHA_SUPPORT or i915.alpha_support module parameter\n" "to enable support in this kernel version, or check for kernel updates.\n"); @@ -689,10 +689,10 @@ static int __init i915_init(void) * vga_text_mode_force boot option. */ - if (i915.modeset == 0) + if (i915_modparams.modeset == 0) use_kms = false; - if (vgacon_text_force() && i915.modeset == -1) + if (vgacon_text_force() && i915_modparams.modeset == -1) use_kms = false; if (!use_kms) { diff --git a/drivers/gpu/drm/i915/i915_perf.c b/drivers/gpu/drm/i915/i915_perf.c index 902722ab84c9..1383a2995a69 100644 --- a/drivers/gpu/drm/i915/i915_perf.c +++ b/drivers/gpu/drm/i915/i915_perf.c @@ -1214,7 +1214,7 @@ static int oa_get_render_ctx_id(struct i915_perf_stream *stream) { struct drm_i915_private *dev_priv = stream->dev_priv; - if (i915.enable_execlists) + if (i915_modparams.enable_execlists) dev_priv->perf.oa.specific_ctx_id = stream->ctx->hw_id; else { struct intel_engine_cs *engine = dev_priv->engine[RCS]; @@ -1260,7 +1260,7 @@ static void oa_put_render_ctx_id(struct i915_perf_stream *stream) { struct drm_i915_private *dev_priv = stream->dev_priv; - if (i915.enable_execlists) { + if (i915_modparams.enable_execlists) { dev_priv->perf.oa.specific_ctx_id = INVALID_CTX_ID; } else { struct intel_engine_cs *engine = dev_priv->engine[RCS]; @@ -3408,7 +3408,7 @@ void i915_perf_init(struct drm_i915_private *dev_priv) dev_priv->perf.oa.timestamp_frequency = 12500000; dev_priv->perf.oa.oa_formats = hsw_oa_formats; - } else if (i915.enable_execlists) { + } else if (i915_modparams.enable_execlists) { /* Note: that although we could theoretically also support the * legacy ringbuffer mode on BDW (and earlier iterations of * this driver, before upstreaming did this) it didn't seem diff --git a/drivers/gpu/drm/i915/intel_bios.c b/drivers/gpu/drm/i915/intel_bios.c index 5949750a35ee..8526da90168c 100644 --- a/drivers/gpu/drm/i915/intel_bios.c +++ b/drivers/gpu/drm/i915/intel_bios.c @@ -356,7 +356,7 @@ parse_sdvo_panel_data(struct drm_i915_private *dev_priv, struct drm_display_mode *panel_fixed_mode; int index; - index = i915.vbt_sdvo_panel_type; + index = i915_modparams.vbt_sdvo_panel_type; if (index == -2) { DRM_DEBUG_KMS("Ignore SDVO panel mode from BIOS VBT tables.\n"); return; @@ -675,8 +675,9 @@ parse_edp(struct drm_i915_private *dev_priv, const struct bdb_header *bdb) uint8_t vswing; /* Don't read from VBT if module parameter has valid value*/ - if (i915.edp_vswing) { - dev_priv->vbt.edp.low_vswing = i915.edp_vswing == 1; + if (i915_modparams.edp_vswing) { + dev_priv->vbt.edp.low_vswing = + i915_modparams.edp_vswing == 1; } else { vswing = (edp->edp_vswing_preemph >> (panel_type * 4)) & 0xF; dev_priv->vbt.edp.low_vswing = vswing == 0; diff --git a/drivers/gpu/drm/i915/intel_crt.c b/drivers/gpu/drm/i915/intel_crt.c index a77dd80a97f2..954070255b4d 100644 --- a/drivers/gpu/drm/i915/intel_crt.c +++ b/drivers/gpu/drm/i915/intel_crt.c @@ -712,7 +712,7 @@ intel_crt_detect(struct drm_connector *connector, * broken monitor (without edid) to work behind a broken kvm (that fails * to have the right resistors for HP detection) needs to fix this up. * For now just bail out. */ - if (I915_HAS_HOTPLUG(dev_priv) && !i915.load_detect_test) { + if (I915_HAS_HOTPLUG(dev_priv) && !i915_modparams.load_detect_test) { status = connector_status_disconnected; goto out; } @@ -730,7 +730,7 @@ intel_crt_detect(struct drm_connector *connector, else if (INTEL_GEN(dev_priv) < 4) status = intel_crt_load_detect(crt, to_intel_crtc(connector->state->crtc)->pipe); - else if (i915.load_detect_test) + else if (i915_modparams.load_detect_test) status = connector_status_disconnected; else status = connector_status_unknown; diff --git a/drivers/gpu/drm/i915/intel_device_info.c b/drivers/gpu/drm/i915/intel_device_info.c index 43831b09b47a..fdf9b54b71e9 100644 --- a/drivers/gpu/drm/i915/intel_device_info.c +++ b/drivers/gpu/drm/i915/intel_device_info.c @@ -343,7 +343,7 @@ void intel_device_info_runtime_init(struct drm_i915_private *dev_priv) info->num_sprites[pipe] = 1; } - if (i915.disable_display) { + if (i915_modparams.disable_display) { DRM_INFO("Display disabled (module parameter)\n"); info->num_pipes = 0; } else if (info->num_pipes > 0 && diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c index 2e407dc49c19..8a185971798b 100644 --- a/drivers/gpu/drm/i915/intel_display.c +++ b/drivers/gpu/drm/i915/intel_display.c @@ -3701,7 +3701,7 @@ void intel_prepare_reset(struct drm_i915_private *dev_priv) /* reset doesn't touch the display */ - if (!i915.force_reset_modeset_test && + if (!i915_modparams.force_reset_modeset_test && !gpu_reset_clobbers_display(dev_priv)) return; @@ -3757,7 +3757,7 @@ void intel_finish_reset(struct drm_i915_private *dev_priv) int ret; /* reset doesn't touch the display */ - if (!i915.force_reset_modeset_test && + if (!i915_modparams.force_reset_modeset_test && !gpu_reset_clobbers_display(dev_priv)) return; @@ -6313,7 +6313,7 @@ static void hsw_compute_ips_config(struct intel_crtc *crtc, struct drm_device *dev = crtc->base.dev; struct drm_i915_private *dev_priv = to_i915(dev); - pipe_config->ips_enabled = i915.enable_ips && + pipe_config->ips_enabled = i915_modparams.enable_ips && hsw_crtc_supports_ips(crtc) && pipe_config_supports_ips(dev_priv, pipe_config); } @@ -6494,8 +6494,8 @@ intel_link_compute_m_n(int bits_per_pixel, int nlanes, static inline bool intel_panel_use_ssc(struct drm_i915_private *dev_priv) { - if (i915.panel_use_ssc >= 0) - return i915.panel_use_ssc != 0; + if (i915_modparams.panel_use_ssc >= 0) + return i915_modparams.panel_use_ssc != 0; return dev_priv->vbt.lvds_use_ssc && !(dev_priv->quirks & QUIRK_LVDS_SSC_DISABLE); } @@ -12084,7 +12084,7 @@ static int intel_atomic_check(struct drm_device *dev, return ret; } - if (i915.fastboot && + if (i915_modparams.fastboot && intel_pipe_config_compare(dev_priv, to_intel_crtc_state(old_crtc_state), pipe_config, true)) { diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c index 1e0bfbe6b4f3..48ed6c1b5a76 100644 --- a/drivers/gpu/drm/i915/intel_dp.c +++ b/drivers/gpu/drm/i915/intel_dp.c @@ -3848,7 +3848,7 @@ intel_dp_can_mst(struct intel_dp *intel_dp) { u8 mstm_cap; - if (!i915.enable_dp_mst) + if (!i915_modparams.enable_dp_mst) return false; if (!intel_dp->can_mst) @@ -3866,7 +3866,7 @@ intel_dp_can_mst(struct intel_dp *intel_dp) static void intel_dp_configure_mst(struct intel_dp *intel_dp) { - if (!i915.enable_dp_mst) + if (!i915_modparams.enable_dp_mst) return; if (!intel_dp->can_mst) diff --git a/drivers/gpu/drm/i915/intel_dp_aux_backlight.c b/drivers/gpu/drm/i915/intel_dp_aux_backlight.c index d2830ba3162e..2bb2ceb9d463 100644 --- a/drivers/gpu/drm/i915/intel_dp_aux_backlight.c +++ b/drivers/gpu/drm/i915/intel_dp_aux_backlight.c @@ -264,7 +264,7 @@ int intel_dp_aux_init_backlight_funcs(struct intel_connector *intel_connector) { struct intel_panel *panel = &intel_connector->panel; - if (!i915.enable_dpcd_backlight) + if (!i915_modparams.enable_dpcd_backlight) return -ENODEV; if (!intel_dp_aux_display_control_capable(intel_connector)) diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h index 307807672896..64358d2f2422 100644 --- a/drivers/gpu/drm/i915/intel_drv.h +++ b/drivers/gpu/drm/i915/intel_drv.h @@ -1902,7 +1902,7 @@ void intel_init_ipc(struct drm_i915_private *dev_priv); void intel_enable_ipc(struct drm_i915_private *dev_priv); static inline int intel_enable_rc6(void) { - return i915.enable_rc6; + return i915_modparams.enable_rc6; } /* intel_sdvo.c */ diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c index 3d135c3cd380..d755a2ae4223 100644 --- a/drivers/gpu/drm/i915/intel_engine_cs.c +++ b/drivers/gpu/drm/i915/intel_engine_cs.c @@ -153,7 +153,7 @@ __intel_engine_context_size(struct drm_i915_private *dev_priv, u8 class) case 9: return GEN9_LR_CONTEXT_RENDER_SIZE; case 8: - return i915.enable_execlists ? + return i915_modparams.enable_execlists ? GEN8_LR_CONTEXT_RENDER_SIZE : GEN8_CXT_TOTAL_SIZE; case 7: @@ -301,7 +301,7 @@ int intel_engines_init(struct drm_i915_private *dev_priv) &intel_engine_classes[engine->class]; int (*init)(struct intel_engine_cs *engine); - if (i915.enable_execlists) + if (i915_modparams.enable_execlists) init = class_info->init_execlists; else init = class_info->init_legacy; diff --git a/drivers/gpu/drm/i915/intel_fbc.c b/drivers/gpu/drm/i915/intel_fbc.c index 58a772de6672..8e3a05505f49 100644 --- a/drivers/gpu/drm/i915/intel_fbc.c +++ b/drivers/gpu/drm/i915/intel_fbc.c @@ -859,7 +859,7 @@ static bool intel_fbc_can_enable(struct drm_i915_private *dev_priv) return false; } - if (!i915.enable_fbc) { + if (!i915_modparams.enable_fbc) { fbc->no_fbc_reason = "disabled per module param or by default"; return false; } @@ -1310,8 +1310,8 @@ void intel_fbc_init_pipe_state(struct drm_i915_private *dev_priv) */ static int intel_sanitize_fbc_option(struct drm_i915_private *dev_priv) { - if (i915.enable_fbc >= 0) - return !!i915.enable_fbc; + if (i915_modparams.enable_fbc >= 0) + return !!i915_modparams.enable_fbc; if (!HAS_FBC(dev_priv)) return 0; @@ -1355,8 +1355,9 @@ void intel_fbc_init(struct drm_i915_private *dev_priv) if (need_fbc_vtd_wa(dev_priv)) mkwrite_device_info(dev_priv)->has_fbc = false; - i915.enable_fbc = intel_sanitize_fbc_option(dev_priv); - DRM_DEBUG_KMS("Sanitized enable_fbc value: %d\n", i915.enable_fbc); + i915_modparams.enable_fbc = intel_sanitize_fbc_option(dev_priv); + DRM_DEBUG_KMS("Sanitized enable_fbc value: %d\n", + i915_modparams.enable_fbc); if (!HAS_FBC(dev_priv)) { fbc->no_fbc_reason = "unsupported by this chipset"; diff --git a/drivers/gpu/drm/i915/intel_guc_loader.c b/drivers/gpu/drm/i915/intel_guc_loader.c index 8b0ae7fce7f2..c9e25be4db40 100644 --- a/drivers/gpu/drm/i915/intel_guc_loader.c +++ b/drivers/gpu/drm/i915/intel_guc_loader.c @@ -131,14 +131,14 @@ static void guc_params_init(struct drm_i915_private *dev_priv) params[GUC_CTL_LOG_PARAMS] = guc->log.flags; - if (i915.guc_log_level >= 0) { + if (i915_modparams.guc_log_level >= 0) { params[GUC_CTL_DEBUG] = - i915.guc_log_level << GUC_LOG_VERBOSITY_SHIFT; + i915_modparams.guc_log_level << GUC_LOG_VERBOSITY_SHIFT; } else params[GUC_CTL_DEBUG] = GUC_LOG_DISABLED; /* If GuC submission is enabled, set up additional parameters here */ - if (i915.enable_guc_submission) { + if (i915_modparams.enable_guc_submission) { u32 ads = guc_ggtt_offset(guc->ads_vma) >> PAGE_SHIFT; u32 pgs = guc_ggtt_offset(dev_priv->guc.stage_desc_pool); u32 ctx_in_16 = GUC_MAX_STAGE_DESCRIPTORS / 16; @@ -368,7 +368,8 @@ int intel_guc_init_hw(struct intel_guc *guc) guc->fw.load_status = INTEL_UC_FIRMWARE_SUCCESS; DRM_INFO("GuC %s (firmware %s [version %u.%u])\n", - i915.enable_guc_submission ? "submission enabled" : "loaded", + i915_modparams.enable_guc_submission ? "submission enabled" : + "loaded", guc->fw.path, guc->fw.major_ver_found, guc->fw.minor_ver_found); @@ -390,8 +391,8 @@ int intel_guc_select_fw(struct intel_guc *guc) guc->fw.load_status = INTEL_UC_FIRMWARE_NONE; guc->fw.type = INTEL_UC_FW_TYPE_GUC; - if (i915.guc_firmware_path) { - guc->fw.path = i915.guc_firmware_path; + if (i915_modparams.guc_firmware_path) { + guc->fw.path = i915_modparams.guc_firmware_path; guc->fw.major_ver_wanted = 0; guc->fw.minor_ver_wanted = 0; } else if (IS_SKYLAKE(dev_priv)) { diff --git a/drivers/gpu/drm/i915/intel_guc_log.c b/drivers/gpu/drm/i915/intel_guc_log.c index 16d3b8719cab..6571d96704ad 100644 --- a/drivers/gpu/drm/i915/intel_guc_log.c +++ b/drivers/gpu/drm/i915/intel_guc_log.c @@ -144,7 +144,7 @@ static int guc_log_relay_file_create(struct intel_guc *guc) struct dentry *log_dir; int ret; - if (i915.guc_log_level < 0) + if (i915_modparams.guc_log_level < 0) return 0; /* For now create the log file in /sys/kernel/debug/dri/0 dir */ @@ -480,7 +480,7 @@ err_runtime: guc_log_runtime_destroy(guc); err: /* logging will remain off */ - i915.guc_log_level = -1; + i915_modparams.guc_log_level = -1; return ret; } @@ -502,7 +502,8 @@ static void guc_flush_logs(struct intel_guc *guc) { struct drm_i915_private *dev_priv = guc_to_i915(guc); - if (!i915.enable_guc_submission || (i915.guc_log_level < 0)) + if (!i915_modparams.enable_guc_submission || + (i915_modparams.guc_log_level < 0)) return; /* First disable the interrupts, will be renabled afterwards */ @@ -529,8 +530,8 @@ int intel_guc_log_create(struct intel_guc *guc) GEM_BUG_ON(guc->log.vma); - if (i915.guc_log_level > GUC_LOG_VERBOSITY_MAX) - i915.guc_log_level = GUC_LOG_VERBOSITY_MAX; + if (i915_modparams.guc_log_level > GUC_LOG_VERBOSITY_MAX) + i915_modparams.guc_log_level = GUC_LOG_VERBOSITY_MAX; /* The first page is to save log buffer state. Allocate one * extra page for others in case for overlap */ @@ -555,7 +556,7 @@ int intel_guc_log_create(struct intel_guc *guc) guc->log.vma = vma; - if (i915.guc_log_level >= 0) { + if (i915_modparams.guc_log_level >= 0) { ret = guc_log_runtime_create(guc); if (ret < 0) goto err_vma; @@ -576,7 +577,7 @@ err_vma: i915_vma_unpin_and_release(&guc->log.vma); err: /* logging will be off */ - i915.guc_log_level = -1; + i915_modparams.guc_log_level = -1; return ret; } @@ -600,7 +601,7 @@ int i915_guc_log_control(struct drm_i915_private *dev_priv, u64 control_val) return -EINVAL; /* This combination doesn't make sense & won't have any effect */ - if (!log_param.logging_enabled && (i915.guc_log_level < 0)) + if (!log_param.logging_enabled && (i915_modparams.guc_log_level < 0)) return 0; ret = guc_log_control(guc, log_param.value); @@ -610,7 +611,7 @@ int i915_guc_log_control(struct drm_i915_private *dev_priv, u64 control_val) } if (log_param.logging_enabled) { - i915.guc_log_level = log_param.verbosity; + i915_modparams.guc_log_level = log_param.verbosity; /* If log_level was set as -1 at boot time, then the relay channel file * wouldn't have been created by now and interrupts also would not have @@ -633,7 +634,7 @@ int i915_guc_log_control(struct drm_i915_private *dev_priv, u64 control_val) guc_flush_logs(guc); /* As logging is disabled, update log level to reflect that */ - i915.guc_log_level = -1; + i915_modparams.guc_log_level = -1; } return ret; @@ -641,7 +642,8 @@ int i915_guc_log_control(struct drm_i915_private *dev_priv, u64 control_val) void i915_guc_log_register(struct drm_i915_private *dev_priv) { - if (!i915.enable_guc_submission || i915.guc_log_level < 0) + if (!i915_modparams.enable_guc_submission || + (i915_modparams.guc_log_level < 0)) return; mutex_lock(&dev_priv->drm.struct_mutex); @@ -651,7 +653,7 @@ void i915_guc_log_register(struct drm_i915_private *dev_priv) void i915_guc_log_unregister(struct drm_i915_private *dev_priv) { - if (!i915.enable_guc_submission) + if (!i915_modparams.enable_guc_submission) return; mutex_lock(&dev_priv->drm.struct_mutex); diff --git a/drivers/gpu/drm/i915/intel_gvt.c b/drivers/gpu/drm/i915/intel_gvt.c index c17ed0e62b67..b4a7f31f0214 100644 --- a/drivers/gpu/drm/i915/intel_gvt.c +++ b/drivers/gpu/drm/i915/intel_gvt.c @@ -58,7 +58,7 @@ static bool is_supported_device(struct drm_i915_private *dev_priv) */ void intel_gvt_sanitize_options(struct drm_i915_private *dev_priv) { - if (!i915.enable_gvt) + if (!i915_modparams.enable_gvt) return; if (intel_vgpu_active(dev_priv)) { @@ -73,7 +73,7 @@ void intel_gvt_sanitize_options(struct drm_i915_private *dev_priv) return; bail: - i915.enable_gvt = 0; + i915_modparams.enable_gvt = 0; } /** @@ -90,17 +90,17 @@ int intel_gvt_init(struct drm_i915_private *dev_priv) { int ret; - if (!i915.enable_gvt) { + if (!i915_modparams.enable_gvt) { DRM_DEBUG_DRIVER("GVT-g is disabled by kernel params\n"); return 0; } - if (!i915.enable_execlists) { + if (!i915_modparams.enable_execlists) { DRM_ERROR("i915 GVT-g loading failed due to disabled execlists mode\n"); return -EIO; } - if (i915.enable_guc_submission) { + if (i915_modparams.enable_guc_submission) { DRM_ERROR("i915 GVT-g loading failed due to Graphics virtualization is not yet supported with GuC submission\n"); return -EIO; } @@ -123,7 +123,7 @@ int intel_gvt_init(struct drm_i915_private *dev_priv) return 0; bail: - i915.enable_gvt = 0; + i915_modparams.enable_gvt = 0; return 0; } diff --git a/drivers/gpu/drm/i915/intel_hangcheck.c b/drivers/gpu/drm/i915/intel_hangcheck.c index d9d87d96fb69..12ac270a5f93 100644 --- a/drivers/gpu/drm/i915/intel_hangcheck.c +++ b/drivers/gpu/drm/i915/intel_hangcheck.c @@ -428,7 +428,7 @@ static void i915_hangcheck_elapsed(struct work_struct *work) unsigned int hung = 0, stuck = 0; int busy_count = 0; - if (!i915.enable_hangcheck) + if (!i915_modparams.enable_hangcheck) return; if (!READ_ONCE(dev_priv->gt.awake)) diff --git a/drivers/gpu/drm/i915/intel_huc.c b/drivers/gpu/drm/i915/intel_huc.c index 6145fa0d6773..6e1779ba87a9 100644 --- a/drivers/gpu/drm/i915/intel_huc.c +++ b/drivers/gpu/drm/i915/intel_huc.c @@ -155,8 +155,8 @@ void intel_huc_select_fw(struct intel_huc *huc) huc->fw.load_status = INTEL_UC_FIRMWARE_NONE; huc->fw.type = INTEL_UC_FW_TYPE_HUC; - if (i915.huc_firmware_path) { - huc->fw.path = i915.huc_firmware_path; + if (i915_modparams.huc_firmware_path) { + huc->fw.path = i915_modparams.huc_firmware_path; huc->fw.major_ver_wanted = 0; huc->fw.minor_ver_wanted = 0; } else if (IS_SKYLAKE(dev_priv)) { diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c index 86fed9f1f1ae..955c87999280 100644 --- a/drivers/gpu/drm/i915/intel_lrc.c +++ b/drivers/gpu/drm/i915/intel_lrc.c @@ -244,7 +244,7 @@ int intel_sanitize_enable_execlists(struct drm_i915_private *dev_priv, int enabl if (HAS_LOGICAL_RING_CONTEXTS(dev_priv) && USES_PPGTT(dev_priv) && - i915.use_mmio_flip >= 0) + i915_modparams.use_mmio_flip >= 0) return 1; return 0; @@ -1324,7 +1324,7 @@ static int gen8_init_common_ring(struct intel_engine_cs *engine) engine->csb_head = -1; /* After a GPU reset, we may have requests to replay */ - if (!i915.enable_guc_submission && engine->execlist_first) + if (!i915_modparams.enable_guc_submission && engine->execlist_first) tasklet_schedule(&engine->irq_tasklet); return 0; diff --git a/drivers/gpu/drm/i915/intel_lvds.c b/drivers/gpu/drm/i915/intel_lvds.c index a9813aea89d8..a55954a89148 100644 --- a/drivers/gpu/drm/i915/intel_lvds.c +++ b/drivers/gpu/drm/i915/intel_lvds.c @@ -880,8 +880,8 @@ static bool compute_is_dual_link_lvds(struct intel_lvds_encoder *lvds_encoder) struct drm_i915_private *dev_priv = to_i915(dev); /* use the module option value if specified */ - if (i915.lvds_channel_mode > 0) - return i915.lvds_channel_mode == 2; + if (i915_modparams.lvds_channel_mode > 0) + return i915_modparams.lvds_channel_mode == 2; /* single channel LVDS is limited to 112 MHz */ if (lvds_encoder->attached_connector->base.panel.fixed_mode->clock diff --git a/drivers/gpu/drm/i915/intel_opregion.c b/drivers/gpu/drm/i915/intel_opregion.c index 98154efcb2f4..1d946240e55f 100644 --- a/drivers/gpu/drm/i915/intel_opregion.c +++ b/drivers/gpu/drm/i915/intel_opregion.c @@ -921,7 +921,7 @@ static int intel_load_vbt_firmware(struct drm_i915_private *dev_priv) { struct intel_opregion *opregion = &dev_priv->opregion; const struct firmware *fw = NULL; - const char *name = i915.vbt_firmware; + const char *name = i915_modparams.vbt_firmware; int ret; if (!name || !*name) diff --git a/drivers/gpu/drm/i915/intel_panel.c b/drivers/gpu/drm/i915/intel_panel.c index 3b1c5d783ee7..adc51e452e3e 100644 --- a/drivers/gpu/drm/i915/intel_panel.c +++ b/drivers/gpu/drm/i915/intel_panel.c @@ -379,13 +379,13 @@ enum drm_connector_status intel_panel_detect(struct drm_i915_private *dev_priv) { /* Assume that the BIOS does not lie through the OpRegion... */ - if (!i915.panel_ignore_lid && dev_priv->opregion.lid_state) { + if (!i915_modparams.panel_ignore_lid && dev_priv->opregion.lid_state) { return *dev_priv->opregion.lid_state & 0x1 ? connector_status_connected : connector_status_disconnected; } - switch (i915.panel_ignore_lid) { + switch (i915_modparams.panel_ignore_lid) { case -2: return connector_status_connected; case -1: @@ -465,10 +465,10 @@ static u32 intel_panel_compute_brightness(struct intel_connector *connector, WARN_ON(panel->backlight.max == 0); - if (i915.invert_brightness < 0) + if (i915_modparams.invert_brightness < 0) return val; - if (i915.invert_brightness > 0 || + if (i915_modparams.invert_brightness > 0 || dev_priv->quirks & QUIRK_INVERT_BRIGHTNESS) { return panel->backlight.max - val + panel->backlight.min; } diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c index adfeb7bb8874..c66af09e27a7 100644 --- a/drivers/gpu/drm/i915/intel_pm.c +++ b/drivers/gpu/drm/i915/intel_pm.c @@ -7825,7 +7825,7 @@ void intel_init_gt_powersave(struct drm_i915_private *dev_priv) * RPM depends on RC6 to save restore the GT HW context, so make RC6 a * requirement. */ - if (!i915.enable_rc6) { + if (!i915_modparams.enable_rc6) { DRM_INFO("RC6 disabled, disabling runtime PM support\n"); intel_runtime_pm_get(dev_priv); } @@ -7882,7 +7882,7 @@ void intel_cleanup_gt_powersave(struct drm_i915_private *dev_priv) if (IS_VALLEYVIEW(dev_priv)) valleyview_cleanup_gt_powersave(dev_priv); - if (!i915.enable_rc6) + if (!i915_modparams.enable_rc6) intel_runtime_pm_put(dev_priv); } @@ -8004,7 +8004,7 @@ static void __intel_autoenable_gt_powersave(struct work_struct *work) if (IS_ERR(req)) goto unlock; - if (!i915.enable_execlists && i915_switch_context(req) == 0) + if (!i915_modparams.enable_execlists && i915_switch_context(req) == 0) rcs->init_context(req); /* Mark the device busy, calling intel_enable_gt_powersave() */ diff --git a/drivers/gpu/drm/i915/intel_psr.c b/drivers/gpu/drm/i915/intel_psr.c index acb50945bfa8..0a17d1f3ca77 100644 --- a/drivers/gpu/drm/i915/intel_psr.c +++ b/drivers/gpu/drm/i915/intel_psr.c @@ -396,7 +396,7 @@ static bool intel_psr_match_conditions(struct intel_dp *intel_dp) return false; } - if (!i915.enable_psr) { + if (!i915_modparams.enable_psr) { DRM_DEBUG_KMS("PSR disable by flag\n"); return false; } @@ -943,8 +943,8 @@ void intel_psr_init(struct drm_i915_private *dev_priv) HSW_EDP_PSR_BASE : BDW_EDP_PSR_BASE; /* Per platform default: all disabled. */ - if (i915.enable_psr == -1) - i915.enable_psr = 0; + if (i915_modparams.enable_psr == -1) + i915_modparams.enable_psr = 0; /* Set link_standby x link_off defaults */ if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv)) @@ -958,11 +958,11 @@ void intel_psr_init(struct drm_i915_private *dev_priv) dev_priv->psr.link_standby = dev_priv->vbt.psr.full_link; /* Override link_standby x link_off defaults */ - if (i915.enable_psr == 2 && !dev_priv->psr.link_standby) { + if (i915_modparams.enable_psr == 2 && !dev_priv->psr.link_standby) { DRM_DEBUG_KMS("PSR: Forcing link standby\n"); dev_priv->psr.link_standby = true; } - if (i915.enable_psr == 3 && dev_priv->psr.link_standby) { + if (i915_modparams.enable_psr == 3 && dev_priv->psr.link_standby) { DRM_DEBUG_KMS("PSR: Forcing main link off\n"); dev_priv->psr.link_standby = false; } diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index 85e64a45d0bf..05c08b0bc172 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -1882,7 +1882,7 @@ static void intel_ring_init_semaphores(struct drm_i915_private *dev_priv, struct drm_i915_gem_object *obj; int ret, i; - if (!i915.semaphores) + if (!i915_modparams.semaphores) return; if (INTEL_GEN(dev_priv) >= 8 && !dev_priv->semaphore) { @@ -1982,7 +1982,7 @@ err_obj: i915_gem_object_put(obj); err: DRM_DEBUG_DRIVER("Failed to allocate space for semaphores, disabling\n"); - i915.semaphores = 0; + i915_modparams.semaphores = 0; } static void intel_ring_init_irq(struct drm_i915_private *dev_priv, @@ -2039,7 +2039,7 @@ static void intel_ring_default_vfuncs(struct drm_i915_private *dev_priv, engine->emit_breadcrumb = i9xx_emit_breadcrumb; engine->emit_breadcrumb_sz = i9xx_emit_breadcrumb_sz; - if (i915.semaphores) { + if (i915_modparams.semaphores) { int num_rings; engine->emit_breadcrumb = gen6_sema_emit_breadcrumb; @@ -2083,7 +2083,7 @@ int intel_init_render_ring_buffer(struct intel_engine_cs *engine) engine->emit_breadcrumb = gen8_render_emit_breadcrumb; engine->emit_breadcrumb_sz = gen8_render_emit_breadcrumb_sz; engine->emit_flush = gen8_render_ring_flush; - if (i915.semaphores) { + if (i915_modparams.semaphores) { int num_rings; engine->semaphore.signal = gen8_rcs_signal; diff --git a/drivers/gpu/drm/i915/intel_runtime_pm.c b/drivers/gpu/drm/i915/intel_runtime_pm.c index a3bfb9f27e7a..7933d1bc6a1c 100644 --- a/drivers/gpu/drm/i915/intel_runtime_pm.c +++ b/drivers/gpu/drm/i915/intel_runtime_pm.c @@ -2413,7 +2413,7 @@ static uint32_t get_allowed_dc_mask(const struct drm_i915_private *dev_priv, mask = 0; } - if (!i915.disable_power_well) + if (!i915_modparams.disable_power_well) max_dc = 0; if (enable_dc >= 0 && enable_dc <= max_dc) { @@ -2471,10 +2471,11 @@ int intel_power_domains_init(struct drm_i915_private *dev_priv) { struct i915_power_domains *power_domains = &dev_priv->power_domains; - i915.disable_power_well = sanitize_disable_power_well_option(dev_priv, - i915.disable_power_well); - dev_priv->csr.allowed_dc_mask = get_allowed_dc_mask(dev_priv, - i915.enable_dc); + i915_modparams.disable_power_well = + sanitize_disable_power_well_option(dev_priv, + i915_modparams.disable_power_well); + dev_priv->csr.allowed_dc_mask = + get_allowed_dc_mask(dev_priv, i915_modparams.enable_dc); BUILD_BUG_ON(POWER_DOMAIN_NUM > 64); @@ -2535,7 +2536,7 @@ void intel_power_domains_fini(struct drm_i915_private *dev_priv) intel_display_set_init_power(dev_priv, true); /* Remove the refcount we took to keep power well support disabled. */ - if (!i915.disable_power_well) + if (!i915_modparams.disable_power_well) intel_display_power_put(dev_priv, POWER_DOMAIN_INIT); /* @@ -2995,7 +2996,7 @@ void intel_power_domains_init_hw(struct drm_i915_private *dev_priv, bool resume) /* For now, we need the power well to be always enabled. */ intel_display_set_init_power(dev_priv, true); /* Disable power support if the user asked so. */ - if (!i915.disable_power_well) + if (!i915_modparams.disable_power_well) intel_display_power_get(dev_priv, POWER_DOMAIN_INIT); intel_power_domains_sync_hw(dev_priv); power_domains->initializing = false; @@ -3014,7 +3015,7 @@ void intel_power_domains_suspend(struct drm_i915_private *dev_priv) * Even if power well support was disabled we still want to disable * power wells while we are system suspended. */ - if (!i915.disable_power_well) + if (!i915_modparams.disable_power_well) intel_display_power_put(dev_priv, POWER_DOMAIN_INIT); if (IS_CANNONLAKE(dev_priv)) diff --git a/drivers/gpu/drm/i915/intel_uc.c b/drivers/gpu/drm/i915/intel_uc.c index 0178ba42a0e5..901854007664 100644 --- a/drivers/gpu/drm/i915/intel_uc.c +++ b/drivers/gpu/drm/i915/intel_uc.c @@ -63,35 +63,35 @@ static int __intel_uc_reset_hw(struct drm_i915_private *dev_priv) void intel_uc_sanitize_options(struct drm_i915_private *dev_priv) { if (!HAS_GUC(dev_priv)) { - if (i915.enable_guc_loading > 0 || - i915.enable_guc_submission > 0) + if (i915_modparams.enable_guc_loading > 0 || + i915_modparams.enable_guc_submission > 0) DRM_INFO("Ignoring GuC options, no hardware\n"); - i915.enable_guc_loading = 0; - i915.enable_guc_submission = 0; + i915_modparams.enable_guc_loading = 0; + i915_modparams.enable_guc_submission = 0; return; } /* A negative value means "use platform default" */ - if (i915.enable_guc_loading < 0) - i915.enable_guc_loading = HAS_GUC_UCODE(dev_priv); + if (i915_modparams.enable_guc_loading < 0) + i915_modparams.enable_guc_loading = HAS_GUC_UCODE(dev_priv); /* Verify firmware version */ - if (i915.enable_guc_loading) { + if (i915_modparams.enable_guc_loading) { if (HAS_HUC_UCODE(dev_priv)) intel_huc_select_fw(&dev_priv->huc); if (intel_guc_select_fw(&dev_priv->guc)) - i915.enable_guc_loading = 0; + i915_modparams.enable_guc_loading = 0; } /* Can't enable guc submission without guc loaded */ - if (!i915.enable_guc_loading) - i915.enable_guc_submission = 0; + if (!i915_modparams.enable_guc_loading) + i915_modparams.enable_guc_submission = 0; /* A negative value means "use platform default" */ - if (i915.enable_guc_submission < 0) - i915.enable_guc_submission = HAS_GUC_SCHED(dev_priv); + if (i915_modparams.enable_guc_submission < 0) + i915_modparams.enable_guc_submission = HAS_GUC_SCHED(dev_priv); } static void gen8_guc_raise_irq(struct intel_guc *guc) @@ -290,7 +290,7 @@ static void guc_init_send_regs(struct intel_guc *guc) static void guc_capture_load_err_log(struct intel_guc *guc) { - if (!guc->log.vma || i915.guc_log_level < 0) + if (!guc->log.vma || i915_modparams.guc_log_level < 0) return; if (!guc->load_err_log) @@ -333,7 +333,7 @@ int intel_uc_init_hw(struct drm_i915_private *dev_priv) struct intel_guc *guc = &dev_priv->guc; int ret, attempts; - if (!i915.enable_guc_loading) + if (!i915_modparams.enable_guc_loading) return 0; guc_disable_communication(guc); @@ -342,7 +342,7 @@ int intel_uc_init_hw(struct drm_i915_private *dev_priv) /* We need to notify the guc whenever we change the GGTT */ i915_ggtt_enable_guc(dev_priv); - if (i915.enable_guc_submission) { + if (i915_modparams.enable_guc_submission) { /* * This is stuff we need to have available at fw load time * if we are planning to enable submission later @@ -391,8 +391,8 @@ int intel_uc_init_hw(struct drm_i915_private *dev_priv) goto err_log_capture; intel_guc_auth_huc(dev_priv); - if (i915.enable_guc_submission) { - if (i915.guc_log_level >= 0) + if (i915_modparams.enable_guc_submission) { + if (i915_modparams.guc_log_level >= 0) gen9_enable_guc_interrupts(dev_priv); ret = i915_guc_submission_enable(dev_priv); @@ -417,23 +417,24 @@ err_interrupts: err_log_capture: guc_capture_load_err_log(guc); err_submission: - if (i915.enable_guc_submission) + if (i915_modparams.enable_guc_submission) i915_guc_submission_fini(dev_priv); err_guc: i915_ggtt_disable_guc(dev_priv); DRM_ERROR("GuC init failed\n"); - if (i915.enable_guc_loading > 1 || i915.enable_guc_submission > 1) + if (i915_modparams.enable_guc_loading > 1 || + i915_modparams.enable_guc_submission > 1) ret = -EIO; else ret = 0; - if (i915.enable_guc_submission) { - i915.enable_guc_submission = 0; + if (i915_modparams.enable_guc_submission) { + i915_modparams.enable_guc_submission = 0; DRM_NOTE("Falling back from GuC submission to execlist mode\n"); } - i915.enable_guc_loading = 0; + i915_modparams.enable_guc_loading = 0; DRM_NOTE("GuC firmware loading disabled\n"); return ret; @@ -443,15 +444,15 @@ void intel_uc_fini_hw(struct drm_i915_private *dev_priv) { guc_free_load_err_log(&dev_priv->guc); - if (!i915.enable_guc_loading) + if (!i915_modparams.enable_guc_loading) return; - if (i915.enable_guc_submission) + if (i915_modparams.enable_guc_submission) i915_guc_submission_disable(dev_priv); guc_disable_communication(&dev_priv->guc); - if (i915.enable_guc_submission) { + if (i915_modparams.enable_guc_submission) { gen9_disable_guc_interrupts(dev_priv); i915_guc_submission_fini(dev_priv); } diff --git a/drivers/gpu/drm/i915/intel_uncore.c b/drivers/gpu/drm/i915/intel_uncore.c index fdd7f93acb4f..b3c3f94fc7e4 100644 --- a/drivers/gpu/drm/i915/intel_uncore.c +++ b/drivers/gpu/drm/i915/intel_uncore.c @@ -436,7 +436,8 @@ void intel_uncore_resume_early(struct drm_i915_private *dev_priv) void intel_uncore_sanitize(struct drm_i915_private *dev_priv) { - i915.enable_rc6 = sanitize_rc6_option(dev_priv, i915.enable_rc6); + i915_modparams.enable_rc6 = + sanitize_rc6_option(dev_priv, i915_modparams.enable_rc6); /* BIOS often leaves RC6 enabled, but disable it for hw init */ intel_sanitize_gt_powersave(dev_priv); @@ -507,10 +508,10 @@ void intel_uncore_forcewake_user_get(struct drm_i915_private *dev_priv) dev_priv->uncore.user_forcewake.saved_mmio_check = dev_priv->uncore.unclaimed_mmio_check; dev_priv->uncore.user_forcewake.saved_mmio_debug = - i915.mmio_debug; + i915_modparams.mmio_debug; dev_priv->uncore.unclaimed_mmio_check = 0; - i915.mmio_debug = 0; + i915_modparams.mmio_debug = 0; } spin_unlock_irq(&dev_priv->uncore.lock); } @@ -532,7 +533,7 @@ void intel_uncore_forcewake_user_put(struct drm_i915_private *dev_priv) dev_priv->uncore.unclaimed_mmio_check = dev_priv->uncore.user_forcewake.saved_mmio_check; - i915.mmio_debug = + i915_modparams.mmio_debug = dev_priv->uncore.user_forcewake.saved_mmio_debug; intel_uncore_forcewake_put__locked(dev_priv, FORCEWAKE_ALL); @@ -841,7 +842,8 @@ __unclaimed_reg_debug(struct drm_i915_private *dev_priv, "Unclaimed %s register 0x%x\n", read ? "read from" : "write to", i915_mmio_reg_offset(reg))) - i915.mmio_debug--; /* Only report the first N failures */ + /* Only report the first N failures */ + i915_modparams.mmio_debug--; } static inline void @@ -850,7 +852,7 @@ unclaimed_reg_debug(struct drm_i915_private *dev_priv, const bool read, const bool before) { - if (likely(!i915.mmio_debug)) + if (likely(!i915_modparams.mmio_debug)) return; __unclaimed_reg_debug(dev_priv, reg, read, before); @@ -1703,7 +1705,7 @@ typedef int (*reset_func)(struct drm_i915_private *, unsigned engine_mask); static reset_func intel_get_gpu_reset(struct drm_i915_private *dev_priv) { - if (!i915.reset) + if (!i915_modparams.reset) return NULL; if (INTEL_INFO(dev_priv)->gen >= 8) @@ -1777,7 +1779,7 @@ bool intel_has_reset_engine(struct drm_i915_private *dev_priv) { return (dev_priv->info.has_reset_engine && !dev_priv->guc.execbuf_client && - i915.reset >= 2); + i915_modparams.reset >= 2); } int intel_guc_reset(struct drm_i915_private *dev_priv) @@ -1802,7 +1804,7 @@ bool intel_uncore_unclaimed_mmio(struct drm_i915_private *dev_priv) bool intel_uncore_arm_unclaimed_mmio_detection(struct drm_i915_private *dev_priv) { - if (unlikely(i915.mmio_debug || + if (unlikely(i915_modparams.mmio_debug || dev_priv->uncore.unclaimed_mmio_check <= 0)) return false; @@ -1810,7 +1812,7 @@ intel_uncore_arm_unclaimed_mmio_detection(struct drm_i915_private *dev_priv) DRM_DEBUG("Unclaimed register detected, " "enabling oneshot unclaimed register reporting. " "Please use i915.mmio_debug=N for more information.\n"); - i915.mmio_debug++; + i915_modparams.mmio_debug++; dev_priv->uncore.unclaimed_mmio_check--; return true; } -- cgit v1.2.3 From 7fd0b1a2593647aba4538196dd80314fd134877b Mon Sep 17 00:00:00 2001 From: Oscar Mateo Date: Thu, 21 Sep 2017 16:19:49 -0700 Subject: drm/i915/cnl: Add Gen10 LRC size The total size of the context has decreased with the removal of the URB_ATOMIC section. BSpec indicates 16750 DWORDs (17 pages), plus one page for PPHWSP, and I'm throwing an extra page for precaution. Signed-off-by: Oscar Mateo Cc: Rodrigo Vivi Cc: Daniele Ceraolo Spurio Cc: Ben Widawsky Acked-by: Rodrigo Vivi Signed-off-by: Rodrigo Vivi Link: https://patchwork.freedesktop.org/patch/msgid/1506035989-14295-1-git-send-email-oscar.mateo@intel.com --- drivers/gpu/drm/i915/intel_engine_cs.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'drivers/gpu/drm/i915/intel_engine_cs.c') diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c index d755a2ae4223..020e4c6c0192 100644 --- a/drivers/gpu/drm/i915/intel_engine_cs.c +++ b/drivers/gpu/drm/i915/intel_engine_cs.c @@ -39,6 +39,7 @@ #define GEN8_LR_CONTEXT_RENDER_SIZE (20 * PAGE_SIZE) #define GEN9_LR_CONTEXT_RENDER_SIZE (22 * PAGE_SIZE) +#define GEN10_LR_CONTEXT_RENDER_SIZE (19 * PAGE_SIZE) #define GEN8_LR_CONTEXT_OTHER_SIZE ( 2 * PAGE_SIZE) @@ -150,6 +151,7 @@ __intel_engine_context_size(struct drm_i915_private *dev_priv, u8 class) default: MISSING_CASE(INTEL_GEN(dev_priv)); case 10: + return GEN10_LR_CONTEXT_RENDER_SIZE; case 9: return GEN9_LR_CONTEXT_RENDER_SIZE; case 8: -- cgit v1.2.3 From b620e870218ebe75b8221c7596b46e36d8329c85 Mon Sep 17 00:00:00 2001 From: Mika Kuoppala Date: Fri, 22 Sep 2017 15:43:03 +0300 Subject: drm/i915: Make own struct for execlist items MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Engine's execlist related items have been increasing to a point where a separate struct is warranted. Carve execlist specific items to a dedicated struct to add clarity. v2: add kerneldoc and fix whitespace (Joonas, Chris) v3: csb_mmio changes, rebase v4: s/\b(el|execlist)\b/execlists/ (Joonas) Suggested-by: Chris Wilson Cc: Chris Wilson Cc: Joonas Lahtinen Signed-off-by: Mika Kuoppala Acked-by: Joonas Lahtinen Reviewed-by: Michał Winiarski (v3) Reviewed-by: Chris Wilson (v3) Reviewed-by: Joonas Lahtinen Link: https://patchwork.freedesktop.org/patch/msgid/20170922124307.10914-1-mika.kuoppala@intel.com --- drivers/gpu/drm/i915/i915_debugfs.c | 8 +-- drivers/gpu/drm/i915/i915_gem.c | 6 +- drivers/gpu/drm/i915/i915_gpu_error.c | 4 +- drivers/gpu/drm/i915/i915_guc_submission.c | 31 +++++---- drivers/gpu/drm/i915/i915_irq.c | 5 +- drivers/gpu/drm/i915/intel_engine_cs.c | 12 ++-- drivers/gpu/drm/i915/intel_lrc.c | 100 +++++++++++++++-------------- drivers/gpu/drm/i915/intel_ringbuffer.h | 100 +++++++++++++++++++++++------ 8 files changed, 167 insertions(+), 99 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_engine_cs.c') diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c index b08ebed4e700..87e06da55a69 100644 --- a/drivers/gpu/drm/i915/i915_debugfs.c +++ b/drivers/gpu/drm/i915/i915_debugfs.c @@ -3323,7 +3323,7 @@ static int i915_engine_info(struct seq_file *m, void *unused) read = GEN8_CSB_READ_PTR(ptr); write = GEN8_CSB_WRITE_PTR(ptr); seq_printf(m, "\tExeclist CSB read %d [%d cached], write %d [%d from hws], interrupt posted? %s\n", - read, engine->csb_head, + read, engine->execlists.csb_head, write, intel_read_status_page(engine, intel_hws_csb_write_index(engine->i915)), yesno(test_bit(ENGINE_IRQ_EXECLIST, @@ -3345,10 +3345,10 @@ static int i915_engine_info(struct seq_file *m, void *unused) } rcu_read_lock(); - for (idx = 0; idx < ARRAY_SIZE(engine->execlist_port); idx++) { + for (idx = 0; idx < ARRAY_SIZE(engine->execlists.port); idx++) { unsigned int count; - rq = port_unpack(&engine->execlist_port[idx], + rq = port_unpack(&engine->execlists.port[idx], &count); if (rq) { seq_printf(m, "\t\tELSP[%d] count=%d, ", @@ -3362,7 +3362,7 @@ static int i915_engine_info(struct seq_file *m, void *unused) rcu_read_unlock(); spin_lock_irq(&engine->timeline->lock); - for (rb = engine->execlist_first; rb; rb = rb_next(rb)){ + for (rb = engine->execlists.first; rb; rb = rb_next(rb)) { struct i915_priolist *p = rb_entry(rb, typeof(*p), node); diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c index 12ce97d47afb..49bf5ddfa7fd 100644 --- a/drivers/gpu/drm/i915/i915_gem.c +++ b/drivers/gpu/drm/i915/i915_gem.c @@ -2815,8 +2815,8 @@ i915_gem_reset_prepare_engine(struct intel_engine_cs *engine) * Turning off the engine->irq_tasklet until the reset is over * prevents the race. */ - tasklet_kill(&engine->irq_tasklet); - tasklet_disable(&engine->irq_tasklet); + tasklet_kill(&engine->execlists.irq_tasklet); + tasklet_disable(&engine->execlists.irq_tasklet); if (engine->irq_seqno_barrier) engine->irq_seqno_barrier(engine); @@ -2995,7 +2995,7 @@ void i915_gem_reset(struct drm_i915_private *dev_priv) void i915_gem_reset_finish_engine(struct intel_engine_cs *engine) { - tasklet_enable(&engine->irq_tasklet); + tasklet_enable(&engine->execlists.irq_tasklet); kthread_unpark(engine->breadcrumbs.signaler); } diff --git a/drivers/gpu/drm/i915/i915_gpu_error.c b/drivers/gpu/drm/i915/i915_gpu_error.c index 6cd5eba643e8..20a1f034bf95 100644 --- a/drivers/gpu/drm/i915/i915_gpu_error.c +++ b/drivers/gpu/drm/i915/i915_gpu_error.c @@ -1327,10 +1327,10 @@ static void engine_record_requests(struct intel_engine_cs *engine, static void error_record_engine_execlists(struct intel_engine_cs *engine, struct drm_i915_error_engine *ee) { - const struct execlist_port *port = engine->execlist_port; + const struct execlist_port *port = engine->execlists.port; unsigned int n; - for (n = 0; n < ARRAY_SIZE(engine->execlist_port); n++) { + for (n = 0; n < ARRAY_SIZE(engine->execlists.port); n++) { struct drm_i915_gem_request *rq = port_request(&port[n]); if (!rq) diff --git a/drivers/gpu/drm/i915/i915_guc_submission.c b/drivers/gpu/drm/i915/i915_guc_submission.c index 06a26c610806..bce3f1b5892b 100644 --- a/drivers/gpu/drm/i915/i915_guc_submission.c +++ b/drivers/gpu/drm/i915/i915_guc_submission.c @@ -494,11 +494,12 @@ static void i915_guc_submit(struct intel_engine_cs *engine) struct drm_i915_private *dev_priv = engine->i915; struct intel_guc *guc = &dev_priv->guc; struct i915_guc_client *client = guc->execbuf_client; - struct execlist_port *port = engine->execlist_port; - unsigned int engine_id = engine->id; + struct intel_engine_execlists * const execlists = &engine->execlists; + struct execlist_port *port = execlists->port; + const unsigned int engine_id = engine->id; unsigned int n; - for (n = 0; n < ARRAY_SIZE(engine->execlist_port); n++) { + for (n = 0; n < ARRAY_SIZE(execlists->port); n++) { struct drm_i915_gem_request *rq; unsigned int count; @@ -558,7 +559,8 @@ static void port_assign(struct execlist_port *port, static void i915_guc_dequeue(struct intel_engine_cs *engine) { - struct execlist_port *port = engine->execlist_port; + struct intel_engine_execlists * const execlists = &engine->execlists; + struct execlist_port *port = execlists->port; struct drm_i915_gem_request *last = NULL; bool submit = false; struct rb_node *rb; @@ -567,15 +569,15 @@ static void i915_guc_dequeue(struct intel_engine_cs *engine) port++; spin_lock_irq(&engine->timeline->lock); - rb = engine->execlist_first; - GEM_BUG_ON(rb_first(&engine->execlist_queue) != rb); + rb = execlists->first; + GEM_BUG_ON(rb_first(&execlists->queue) != rb); while (rb) { struct i915_priolist *p = rb_entry(rb, typeof(*p), node); struct drm_i915_gem_request *rq, *rn; list_for_each_entry_safe(rq, rn, &p->requests, priotree.link) { if (last && rq->ctx != last->ctx) { - if (port != engine->execlist_port) { + if (port != execlists->port) { __list_del_many(&p->requests, &rq->priotree.link); goto done; @@ -596,13 +598,13 @@ static void i915_guc_dequeue(struct intel_engine_cs *engine) } rb = rb_next(rb); - rb_erase(&p->node, &engine->execlist_queue); + rb_erase(&p->node, &execlists->queue); INIT_LIST_HEAD(&p->requests); if (p->priority != I915_PRIORITY_NORMAL) kmem_cache_free(engine->i915->priorities, p); } done: - engine->execlist_first = rb; + execlists->first = rb; if (submit) { port_assign(port, last); i915_guc_submit(engine); @@ -612,8 +614,8 @@ done: static void i915_guc_irq_handler(unsigned long data) { - struct intel_engine_cs *engine = (struct intel_engine_cs *)data; - struct execlist_port *port = engine->execlist_port; + struct intel_engine_cs * const engine = (struct intel_engine_cs *)data; + struct execlist_port *port = engine->execlists.port; struct drm_i915_gem_request *rq; rq = port_request(&port[0]); @@ -1144,7 +1146,7 @@ int i915_guc_submission_enable(struct drm_i915_private *dev_priv) * and it is guaranteed that it will remove the work item from the * queue before our request is completed. */ - BUILD_BUG_ON(ARRAY_SIZE(engine->execlist_port) * + BUILD_BUG_ON(ARRAY_SIZE(engine->execlists.port) * sizeof(struct guc_wq_item) * I915_NUM_ENGINES > GUC_WQ_SIZE); @@ -1175,14 +1177,15 @@ int i915_guc_submission_enable(struct drm_i915_private *dev_priv) guc_interrupts_capture(dev_priv); for_each_engine(engine, dev_priv, id) { + struct intel_engine_execlists * const execlists = &engine->execlists; /* The tasklet was initialised by execlists, and may be in * a state of flux (across a reset) and so we just want to * take over the callback without changing any other state * in the tasklet. */ - engine->irq_tasklet.func = i915_guc_irq_handler; + execlists->irq_tasklet.func = i915_guc_irq_handler; clear_bit(ENGINE_IRQ_EXECLIST, &engine->irq_posted); - tasklet_schedule(&engine->irq_tasklet); + tasklet_schedule(&execlists->irq_tasklet); } return 0; diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c index b1bab7605db9..af82bd721dbc 100644 --- a/drivers/gpu/drm/i915/i915_irq.c +++ b/drivers/gpu/drm/i915/i915_irq.c @@ -1346,10 +1346,11 @@ static void snb_gt_irq_handler(struct drm_i915_private *dev_priv, static void gen8_cs_irq_handler(struct intel_engine_cs *engine, u32 iir, int test_shift) { + struct intel_engine_execlists * const execlists = &engine->execlists; bool tasklet = false; if (iir & (GT_CONTEXT_SWITCH_INTERRUPT << test_shift)) { - if (port_count(&engine->execlist_port[0])) { + if (port_count(&execlists->port[0])) { __set_bit(ENGINE_IRQ_EXECLIST, &engine->irq_posted); tasklet = true; } @@ -1361,7 +1362,7 @@ gen8_cs_irq_handler(struct intel_engine_cs *engine, u32 iir, int test_shift) } if (tasklet) - tasklet_hi_schedule(&engine->irq_tasklet); + tasklet_hi_schedule(&execlists->irq_tasklet); } static irqreturn_t gen8_gt_irq_ack(struct drm_i915_private *dev_priv, diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c index 020e4c6c0192..bf132266a007 100644 --- a/drivers/gpu/drm/i915/intel_engine_cs.c +++ b/drivers/gpu/drm/i915/intel_engine_cs.c @@ -393,8 +393,8 @@ static void intel_engine_init_timeline(struct intel_engine_cs *engine) */ void intel_engine_setup_common(struct intel_engine_cs *engine) { - engine->execlist_queue = RB_ROOT; - engine->execlist_first = NULL; + engine->execlists.queue = RB_ROOT; + engine->execlists.first = NULL; intel_engine_init_timeline(engine); intel_engine_init_hangcheck(engine); @@ -1475,11 +1475,11 @@ bool intel_engine_is_idle(struct intel_engine_cs *engine) return false; /* Both ports drained, no more ELSP submission? */ - if (port_request(&engine->execlist_port[0])) + if (port_request(&engine->execlists.port[0])) return false; /* ELSP is empty, but there are ready requests? */ - if (READ_ONCE(engine->execlist_first)) + if (READ_ONCE(engine->execlists.first)) return false; /* Ring stopped? */ @@ -1528,8 +1528,8 @@ void intel_engines_mark_idle(struct drm_i915_private *i915) for_each_engine(engine, i915, id) { intel_engine_disarm_breadcrumbs(engine); i915_gem_batch_pool_fini(&engine->batch_pool); - tasklet_kill(&engine->irq_tasklet); - engine->no_priolist = false; + tasklet_kill(&engine->execlists.irq_tasklet); + engine->execlists.no_priolist = false; } } diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c index 955c87999280..4f202b840e3d 100644 --- a/drivers/gpu/drm/i915/intel_lrc.c +++ b/drivers/gpu/drm/i915/intel_lrc.c @@ -291,17 +291,18 @@ lookup_priolist(struct intel_engine_cs *engine, struct i915_priotree *pt, int prio) { + struct intel_engine_execlists * const execlists = &engine->execlists; struct i915_priolist *p; struct rb_node **parent, *rb; bool first = true; - if (unlikely(engine->no_priolist)) + if (unlikely(execlists->no_priolist)) prio = I915_PRIORITY_NORMAL; find_priolist: /* most positive priority is scheduled first, equal priorities fifo */ rb = NULL; - parent = &engine->execlist_queue.rb_node; + parent = &execlists->queue.rb_node; while (*parent) { rb = *parent; p = rb_entry(rb, typeof(*p), node); @@ -316,7 +317,7 @@ find_priolist: } if (prio == I915_PRIORITY_NORMAL) { - p = &engine->default_priolist; + p = &execlists->default_priolist; } else { p = kmem_cache_alloc(engine->i915->priorities, GFP_ATOMIC); /* Convert an allocation failure to a priority bump */ @@ -331,7 +332,7 @@ find_priolist: * requests, so if userspace lied about their * dependencies that reordering may be visible. */ - engine->no_priolist = true; + execlists->no_priolist = true; goto find_priolist; } } @@ -339,10 +340,10 @@ find_priolist: p->priority = prio; INIT_LIST_HEAD(&p->requests); rb_link_node(&p->node, rb, parent); - rb_insert_color(&p->node, &engine->execlist_queue); + rb_insert_color(&p->node, &execlists->queue); if (first) - engine->execlist_first = &p->node; + execlists->first = &p->node; return ptr_pack_bits(p, first, 1); } @@ -393,12 +394,12 @@ static u64 execlists_update_context(struct drm_i915_gem_request *rq) static void execlists_submit_ports(struct intel_engine_cs *engine) { - struct execlist_port *port = engine->execlist_port; + struct execlist_port *port = engine->execlists.port; u32 __iomem *elsp = engine->i915->regs + i915_mmio_reg_offset(RING_ELSP(engine)); unsigned int n; - for (n = ARRAY_SIZE(engine->execlist_port); n--; ) { + for (n = ARRAY_SIZE(engine->execlists.port); n--; ) { struct drm_i915_gem_request *rq; unsigned int count; u64 desc; @@ -453,7 +454,7 @@ static void port_assign(struct execlist_port *port, static void execlists_dequeue(struct intel_engine_cs *engine) { struct drm_i915_gem_request *last; - struct execlist_port *port = engine->execlist_port; + struct execlist_port *port = engine->execlists.port; struct rb_node *rb; bool submit = false; @@ -491,8 +492,8 @@ static void execlists_dequeue(struct intel_engine_cs *engine) */ spin_lock_irq(&engine->timeline->lock); - rb = engine->execlist_first; - GEM_BUG_ON(rb_first(&engine->execlist_queue) != rb); + rb = engine->execlists.first; + GEM_BUG_ON(rb_first(&engine->execlists.queue) != rb); while (rb) { struct i915_priolist *p = rb_entry(rb, typeof(*p), node); struct drm_i915_gem_request *rq, *rn; @@ -515,7 +516,7 @@ static void execlists_dequeue(struct intel_engine_cs *engine) * combine this request with the last, then we * are done. */ - if (port != engine->execlist_port) { + if (port != engine->execlists.port) { __list_del_many(&p->requests, &rq->priotree.link); goto done; @@ -552,13 +553,13 @@ static void execlists_dequeue(struct intel_engine_cs *engine) } rb = rb_next(rb); - rb_erase(&p->node, &engine->execlist_queue); + rb_erase(&p->node, &engine->execlists.queue); INIT_LIST_HEAD(&p->requests); if (p->priority != I915_PRIORITY_NORMAL) kmem_cache_free(engine->i915->priorities, p); } done: - engine->execlist_first = rb; + engine->execlists.first = rb; if (submit) port_assign(port, last); spin_unlock_irq(&engine->timeline->lock); @@ -569,7 +570,8 @@ done: static void execlists_cancel_requests(struct intel_engine_cs *engine) { - struct execlist_port *port = engine->execlist_port; + struct intel_engine_execlists * const execlists = &engine->execlists; + struct execlist_port *port = execlists->port; struct drm_i915_gem_request *rq, *rn; struct rb_node *rb; unsigned long flags; @@ -578,9 +580,9 @@ static void execlists_cancel_requests(struct intel_engine_cs *engine) spin_lock_irqsave(&engine->timeline->lock, flags); /* Cancel the requests on the HW and clear the ELSP tracker. */ - for (n = 0; n < ARRAY_SIZE(engine->execlist_port); n++) + for (n = 0; n < ARRAY_SIZE(execlists->port); n++) i915_gem_request_put(port_request(&port[n])); - memset(engine->execlist_port, 0, sizeof(engine->execlist_port)); + memset(execlists->port, 0, sizeof(execlists->port)); /* Mark all executing requests as skipped. */ list_for_each_entry(rq, &engine->timeline->requests, link) { @@ -590,7 +592,7 @@ static void execlists_cancel_requests(struct intel_engine_cs *engine) } /* Flush the queued requests to the timeline list (for retiring). */ - rb = engine->execlist_first; + rb = execlists->first; while (rb) { struct i915_priolist *p = rb_entry(rb, typeof(*p), node); @@ -603,7 +605,7 @@ static void execlists_cancel_requests(struct intel_engine_cs *engine) } rb = rb_next(rb); - rb_erase(&p->node, &engine->execlist_queue); + rb_erase(&p->node, &execlists->queue); INIT_LIST_HEAD(&p->requests); if (p->priority != I915_PRIORITY_NORMAL) kmem_cache_free(engine->i915->priorities, p); @@ -611,8 +613,8 @@ static void execlists_cancel_requests(struct intel_engine_cs *engine) /* Remaining _unready_ requests will be nop'ed when submitted */ - engine->execlist_queue = RB_ROOT; - engine->execlist_first = NULL; + execlists->queue = RB_ROOT; + execlists->first = NULL; GEM_BUG_ON(port_isset(&port[0])); /* @@ -628,7 +630,7 @@ static void execlists_cancel_requests(struct intel_engine_cs *engine) static bool execlists_elsp_ready(const struct intel_engine_cs *engine) { - const struct execlist_port *port = engine->execlist_port; + const struct execlist_port *port = engine->execlists.port; return port_count(&port[0]) + port_count(&port[1]) < 2; } @@ -639,8 +641,9 @@ static bool execlists_elsp_ready(const struct intel_engine_cs *engine) */ static void intel_lrc_irq_handler(unsigned long data) { - struct intel_engine_cs *engine = (struct intel_engine_cs *)data; - struct execlist_port *port = engine->execlist_port; + struct intel_engine_cs * const engine = (struct intel_engine_cs *)data; + struct intel_engine_execlists * const execlists = &engine->execlists; + struct execlist_port *port = execlists->port; struct drm_i915_private *dev_priv = engine->i915; /* We can skip acquiring intel_runtime_pm_get() here as it was taken @@ -652,7 +655,7 @@ static void intel_lrc_irq_handler(unsigned long data) */ GEM_BUG_ON(!dev_priv->gt.awake); - intel_uncore_forcewake_get(dev_priv, engine->fw_domains); + intel_uncore_forcewake_get(dev_priv, execlists->fw_domains); /* Prefer doing test_and_clear_bit() as a two stage operation to avoid * imposing the cost of a locked atomic transaction when submitting a @@ -665,10 +668,10 @@ static void intel_lrc_irq_handler(unsigned long data) unsigned int head, tail; /* However GVT emulation depends upon intercepting CSB mmio */ - if (unlikely(engine->csb_use_mmio)) { + if (unlikely(execlists->csb_use_mmio)) { buf = (u32 * __force) (dev_priv->regs + i915_mmio_reg_offset(RING_CONTEXT_STATUS_BUF_LO(engine, 0))); - engine->csb_head = -1; /* force mmio read of CSB ptrs */ + execlists->csb_head = -1; /* force mmio read of CSB ptrs */ } /* The write will be ordered by the uncached read (itself @@ -682,19 +685,20 @@ static void intel_lrc_irq_handler(unsigned long data) * is set and we do a new loop. */ __clear_bit(ENGINE_IRQ_EXECLIST, &engine->irq_posted); - if (unlikely(engine->csb_head == -1)) { /* following a reset */ + if (unlikely(execlists->csb_head == -1)) { /* following a reset */ head = readl(dev_priv->regs + i915_mmio_reg_offset(RING_CONTEXT_STATUS_PTR(engine))); tail = GEN8_CSB_WRITE_PTR(head); head = GEN8_CSB_READ_PTR(head); - engine->csb_head = head; + execlists->csb_head = head; } else { const int write_idx = intel_hws_csb_write_index(dev_priv) - I915_HWS_CSB_BUF0_INDEX; - head = engine->csb_head; + head = execlists->csb_head; tail = READ_ONCE(buf[write_idx]); } + while (head != tail) { struct drm_i915_gem_request *rq; unsigned int status; @@ -748,8 +752,8 @@ static void intel_lrc_irq_handler(unsigned long data) !(status & GEN8_CTX_STATUS_ACTIVE_IDLE)); } - if (head != engine->csb_head) { - engine->csb_head = head; + if (head != execlists->csb_head) { + execlists->csb_head = head; writel(_MASKED_FIELD(GEN8_CSB_READ_PTR_MASK, head << 8), dev_priv->regs + i915_mmio_reg_offset(RING_CONTEXT_STATUS_PTR(engine))); } @@ -758,7 +762,7 @@ static void intel_lrc_irq_handler(unsigned long data) if (execlists_elsp_ready(engine)) execlists_dequeue(engine); - intel_uncore_forcewake_put(dev_priv, engine->fw_domains); + intel_uncore_forcewake_put(dev_priv, execlists->fw_domains); } static void insert_request(struct intel_engine_cs *engine, @@ -769,7 +773,7 @@ static void insert_request(struct intel_engine_cs *engine, list_add_tail(&pt->link, &ptr_mask_bits(p, 1)->requests); if (ptr_unmask_bits(p, 1) && execlists_elsp_ready(engine)) - tasklet_hi_schedule(&engine->irq_tasklet); + tasklet_hi_schedule(&engine->execlists.irq_tasklet); } static void execlists_submit_request(struct drm_i915_gem_request *request) @@ -782,7 +786,7 @@ static void execlists_submit_request(struct drm_i915_gem_request *request) insert_request(engine, &request->priotree, request->priotree.priority); - GEM_BUG_ON(!engine->execlist_first); + GEM_BUG_ON(!engine->execlists.first); GEM_BUG_ON(list_empty(&request->priotree.link)); spin_unlock_irqrestore(&engine->timeline->lock, flags); @@ -1289,6 +1293,7 @@ static u8 gtiir[] = { static int gen8_init_common_ring(struct intel_engine_cs *engine) { struct drm_i915_private *dev_priv = engine->i915; + struct intel_engine_execlists * const execlists = &engine->execlists; int ret; ret = intel_mocs_init_engine(engine); @@ -1321,11 +1326,11 @@ static int gen8_init_common_ring(struct intel_engine_cs *engine) I915_WRITE(GEN8_GT_IIR(gtiir[engine->id]), GT_CONTEXT_SWITCH_INTERRUPT << engine->irq_shift); clear_bit(ENGINE_IRQ_EXECLIST, &engine->irq_posted); - engine->csb_head = -1; + execlists->csb_head = -1; /* After a GPU reset, we may have requests to replay */ - if (!i915_modparams.enable_guc_submission && engine->execlist_first) - tasklet_schedule(&engine->irq_tasklet); + if (!i915_modparams.enable_guc_submission && execlists->first) + tasklet_schedule(&execlists->irq_tasklet); return 0; } @@ -1366,7 +1371,8 @@ static int gen9_init_render_ring(struct intel_engine_cs *engine) static void reset_common_ring(struct intel_engine_cs *engine, struct drm_i915_gem_request *request) { - struct execlist_port *port = engine->execlist_port; + struct intel_engine_execlists * const execlists = &engine->execlists; + struct execlist_port *port = execlists->port; struct drm_i915_gem_request *rq, *rn; struct intel_context *ce; unsigned long flags; @@ -1383,9 +1389,9 @@ static void reset_common_ring(struct intel_engine_cs *engine, * guessing the missed context-switch events by looking at what * requests were completed. */ - for (n = 0; n < ARRAY_SIZE(engine->execlist_port); n++) + for (n = 0; n < ARRAY_SIZE(execlists->port); n++) i915_gem_request_put(port_request(&port[n])); - memset(engine->execlist_port, 0, sizeof(engine->execlist_port)); + memset(execlists->port, 0, sizeof(execlists->port)); /* Push back any incomplete requests for replay after the reset. */ list_for_each_entry_safe_reverse(rq, rn, @@ -1719,8 +1725,8 @@ void intel_logical_ring_cleanup(struct intel_engine_cs *engine) * Tasklet cannot be active at this point due intel_mark_active/idle * so this is just for documentation. */ - if (WARN_ON(test_bit(TASKLET_STATE_SCHED, &engine->irq_tasklet.state))) - tasklet_kill(&engine->irq_tasklet); + if (WARN_ON(test_bit(TASKLET_STATE_SCHED, &engine->execlists.irq_tasklet.state))) + tasklet_kill(&engine->execlists.irq_tasklet); dev_priv = engine->i915; @@ -1744,7 +1750,7 @@ static void execlists_set_default_submission(struct intel_engine_cs *engine) engine->submit_request = execlists_submit_request; engine->cancel_requests = execlists_cancel_requests; engine->schedule = execlists_schedule; - engine->irq_tasklet.func = intel_lrc_irq_handler; + engine->execlists.irq_tasklet.func = intel_lrc_irq_handler; } static void @@ -1806,7 +1812,7 @@ logical_ring_setup(struct intel_engine_cs *engine) /* Intentionally left blank. */ engine->buffer = NULL; - engine->csb_use_mmio = irq_handler_force_mmio(dev_priv); + engine->execlists.csb_use_mmio = irq_handler_force_mmio(dev_priv); fw_domains = intel_uncore_forcewake_for_reg(dev_priv, RING_ELSP(engine), @@ -1820,9 +1826,9 @@ logical_ring_setup(struct intel_engine_cs *engine) RING_CONTEXT_STATUS_BUF_BASE(engine), FW_REG_READ); - engine->fw_domains = fw_domains; + engine->execlists.fw_domains = fw_domains; - tasklet_init(&engine->irq_tasklet, + tasklet_init(&engine->execlists.irq_tasklet, intel_lrc_irq_handler, (unsigned long)engine); logical_ring_default_vfuncs(engine); diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.h b/drivers/gpu/drm/i915/intel_ringbuffer.h index 138116a3b537..421e769adb79 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.h +++ b/drivers/gpu/drm/i915/intel_ringbuffer.h @@ -184,6 +184,84 @@ struct i915_priolist { int priority; }; +/** + * struct intel_engine_execlists - execlist submission queue and port state + * + * The struct intel_engine_execlists represents the combined logical state of + * driver and the hardware state for execlist mode of submission. + */ +struct intel_engine_execlists { + /** + * @irq_tasklet: softirq tasklet for bottom handler + */ + struct tasklet_struct irq_tasklet; + + /** + * @default_priolist: priority list for I915_PRIORITY_NORMAL + */ + struct i915_priolist default_priolist; + + /** + * @no_priolist: priority lists disabled + */ + bool no_priolist; + + /** + * @port: execlist port states + * + * For each hardware ELSP (ExecList Submission Port) we keep + * track of the last request and the number of times we submitted + * that port to hw. We then count the number of times the hw reports + * a context completion or preemption. As only one context can + * be active on hw, we limit resubmission of context to port[0]. This + * is called Lite Restore, of the context. + */ + struct execlist_port { + /** + * @request_count: combined request and submission count + */ + struct drm_i915_gem_request *request_count; +#define EXECLIST_COUNT_BITS 2 +#define port_request(p) ptr_mask_bits((p)->request_count, EXECLIST_COUNT_BITS) +#define port_count(p) ptr_unmask_bits((p)->request_count, EXECLIST_COUNT_BITS) +#define port_pack(rq, count) ptr_pack_bits(rq, count, EXECLIST_COUNT_BITS) +#define port_unpack(p, count) ptr_unpack_bits((p)->request_count, count, EXECLIST_COUNT_BITS) +#define port_set(p, packed) ((p)->request_count = (packed)) +#define port_isset(p) ((p)->request_count) +#define port_index(p, e) ((p) - (e)->execlists.port) + + /** + * @context_id: context ID for port + */ + GEM_DEBUG_DECL(u32 context_id); + } port[2]; + + /** + * @queue: queue of requests, in priority lists + */ + struct rb_root queue; + + /** + * @first: leftmost level in priority @queue + */ + struct rb_node *first; + + /** + * @fw_domains: forcewake domains for irq tasklet + */ + unsigned int fw_domains; + + /** + * @csb_head: context status buffer head + */ + unsigned int csb_head; + + /** + * @csb_use_mmio: access csb through mmio, instead of hwsp + */ + bool csb_use_mmio; +}; + #define INTEL_ENGINE_CS_MAX_NAME 8 struct intel_engine_cs { @@ -380,27 +458,7 @@ struct intel_engine_cs { u32 *(*signal)(struct drm_i915_gem_request *req, u32 *cs); } semaphore; - /* Execlists */ - struct tasklet_struct irq_tasklet; - struct i915_priolist default_priolist; - bool no_priolist; - struct execlist_port { - struct drm_i915_gem_request *request_count; -#define EXECLIST_COUNT_BITS 2 -#define port_request(p) ptr_mask_bits((p)->request_count, EXECLIST_COUNT_BITS) -#define port_count(p) ptr_unmask_bits((p)->request_count, EXECLIST_COUNT_BITS) -#define port_pack(rq, count) ptr_pack_bits(rq, count, EXECLIST_COUNT_BITS) -#define port_unpack(p, count) ptr_unpack_bits((p)->request_count, count, EXECLIST_COUNT_BITS) -#define port_set(p, packed) ((p)->request_count = (packed)) -#define port_isset(p) ((p)->request_count) -#define port_index(p, e) ((p) - (e)->execlist_port) - GEM_DEBUG_DECL(u32 context_id); - } execlist_port[2]; - struct rb_root execlist_queue; - struct rb_node *execlist_first; - unsigned int fw_domains; - unsigned int csb_head; - bool csb_use_mmio; + struct intel_engine_execlists execlists; /* Contexts are pinned whilst they are active on the GPU. The last * context executed remains active whilst the GPU is idle - the -- cgit v1.2.3 From 19df9a5782f51c900a730dae11e4abf85a0e5ebc Mon Sep 17 00:00:00 2001 From: Mika Kuoppala Date: Fri, 22 Sep 2017 15:43:04 +0300 Subject: drm/i915: Move execlist initialization into intel_engine_cs.c Move execlist init into a common engine setup. As it is common to both guc and hw execlists. v2: rebase with csb changes v3: rebase Signed-off-by: Mika Kuoppala Reviewed-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20170922124307.10914-2-mika.kuoppala@intel.com --- drivers/gpu/drm/i915/intel_engine_cs.c | 30 ++++++++++++++++++++++++++++-- drivers/gpu/drm/i915/intel_lrc.c | 19 ------------------- 2 files changed, 28 insertions(+), 21 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_engine_cs.c') diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c index bf132266a007..30035e59a784 100644 --- a/drivers/gpu/drm/i915/intel_engine_cs.c +++ b/drivers/gpu/drm/i915/intel_engine_cs.c @@ -382,6 +382,33 @@ static void intel_engine_init_timeline(struct intel_engine_cs *engine) engine->timeline = &engine->i915->gt.global_timeline.engine[engine->id]; } +static bool csb_force_mmio(struct drm_i915_private *i915) +{ + /* GVT emulation depends upon intercepting CSB mmio */ + if (intel_vgpu_active(i915)) + return true; + + /* + * IOMMU adds unpredictable latency causing the CSB write (from the + * GPU into the HWSP) to only be visible some time after the interrupt + * (missed breadcrumb syndrome). + */ + if (intel_vtd_active()) + return true; + + return false; +} + +static void intel_engine_init_execlist(struct intel_engine_cs *engine) +{ + struct intel_engine_execlists * const execlists = &engine->execlists; + + execlists->csb_use_mmio = csb_force_mmio(engine->i915); + + execlists->queue = RB_ROOT; + execlists->first = NULL; +} + /** * intel_engines_setup_common - setup engine state not requiring hw access * @engine: Engine to setup. @@ -393,8 +420,7 @@ static void intel_engine_init_timeline(struct intel_engine_cs *engine) */ void intel_engine_setup_common(struct intel_engine_cs *engine) { - engine->execlists.queue = RB_ROOT; - engine->execlists.first = NULL; + intel_engine_init_execlist(engine); intel_engine_init_timeline(engine); intel_engine_init_hangcheck(engine); diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c index 4f202b840e3d..3186be54bbd8 100644 --- a/drivers/gpu/drm/i915/intel_lrc.c +++ b/drivers/gpu/drm/i915/intel_lrc.c @@ -1784,23 +1784,6 @@ logical_ring_default_irqs(struct intel_engine_cs *engine) engine->irq_keep_mask = GT_CONTEXT_SWITCH_INTERRUPT << shift; } -static bool irq_handler_force_mmio(struct drm_i915_private *i915) -{ - /* GVT emulation depends upon intercepting CSB mmio */ - if (intel_vgpu_active(i915)) - return true; - - /* - * IOMMU adds unpredictable latency causing the CSB write (from the - * GPU into the HWSP) to only be visible some time after the interrupt - * (missed breadcrumb syndrome). - */ - if (intel_vtd_active()) - return true; - - return false; -} - static void logical_ring_setup(struct intel_engine_cs *engine) { @@ -1812,8 +1795,6 @@ logical_ring_setup(struct intel_engine_cs *engine) /* Intentionally left blank. */ engine->buffer = NULL; - engine->execlists.csb_use_mmio = irq_handler_force_mmio(dev_priv); - fw_domains = intel_uncore_forcewake_for_reg(dev_priv, RING_ELSP(engine), FW_REG_WRITE); -- cgit v1.2.3 From 76e70087d360fdbe97f24c205d585ada04126e5f Mon Sep 17 00:00:00 2001 From: Mika Kuoppala Date: Fri, 22 Sep 2017 15:43:07 +0300 Subject: drm/i915: Make execlist port count variable As we emulate execlists on top of the GuC workqueue, it is not restricted to just 2 ports and we can increase that number arbitrarily to trade-off queue depth (i.e. scheduling latency) against pipeline bubbles. v2: rebase. better commit msg (Chris) v3: rebase Signed-off-by: Mika Kuoppala Reviewed-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20170922124307.10914-5-mika.kuoppala@intel.com --- drivers/gpu/drm/i915/i915_debugfs.c | 10 +++++----- drivers/gpu/drm/i915/i915_drv.h | 3 ++- drivers/gpu/drm/i915/i915_gpu_error.c | 17 ++++++++++++----- drivers/gpu/drm/i915/i915_guc_submission.c | 8 ++++++-- drivers/gpu/drm/i915/intel_engine_cs.c | 4 ++++ drivers/gpu/drm/i915/intel_lrc.c | 6 ++++-- drivers/gpu/drm/i915/intel_ringbuffer.h | 21 +++++++++++++++++---- 7 files changed, 50 insertions(+), 19 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_engine_cs.c') diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c index 87e06da55a69..d10d71aa4b0e 100644 --- a/drivers/gpu/drm/i915/i915_debugfs.c +++ b/drivers/gpu/drm/i915/i915_debugfs.c @@ -3312,6 +3312,7 @@ static int i915_engine_info(struct seq_file *m, void *unused) if (i915_modparams.enable_execlists) { const u32 *hws = &engine->status_page.page_addr[I915_HWS_CSB_BUF0_INDEX]; + struct intel_engine_execlists * const execlists = &engine->execlists; u32 ptr, read, write; unsigned int idx; @@ -3323,7 +3324,7 @@ static int i915_engine_info(struct seq_file *m, void *unused) read = GEN8_CSB_READ_PTR(ptr); write = GEN8_CSB_WRITE_PTR(ptr); seq_printf(m, "\tExeclist CSB read %d [%d cached], write %d [%d from hws], interrupt posted? %s\n", - read, engine->execlists.csb_head, + read, execlists->csb_head, write, intel_read_status_page(engine, intel_hws_csb_write_index(engine->i915)), yesno(test_bit(ENGINE_IRQ_EXECLIST, @@ -3345,11 +3346,10 @@ static int i915_engine_info(struct seq_file *m, void *unused) } rcu_read_lock(); - for (idx = 0; idx < ARRAY_SIZE(engine->execlists.port); idx++) { + for (idx = 0; idx < execlists_num_ports(execlists); idx++) { unsigned int count; - rq = port_unpack(&engine->execlists.port[idx], - &count); + rq = port_unpack(&execlists->port[idx], &count); if (rq) { seq_printf(m, "\t\tELSP[%d] count=%d, ", idx, count); @@ -3362,7 +3362,7 @@ static int i915_engine_info(struct seq_file *m, void *unused) rcu_read_unlock(); spin_lock_irq(&engine->timeline->lock); - for (rb = engine->execlists.first; rb; rb = rb_next(rb)) { + for (rb = execlists->first; rb; rb = rb_next(rb)) { struct i915_priolist *p = rb_entry(rb, typeof(*p), node); diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h index d6babe7d362d..0f26aa57a922 100644 --- a/drivers/gpu/drm/i915/i915_drv.h +++ b/drivers/gpu/drm/i915/i915_drv.h @@ -1001,7 +1001,8 @@ struct i915_gpu_state { u32 seqno; u32 head; u32 tail; - } *requests, execlist[2]; + } *requests, execlist[EXECLIST_MAX_PORTS]; + unsigned int num_ports; struct drm_i915_error_waiter { char comm[TASK_COMM_LEN]; diff --git a/drivers/gpu/drm/i915/i915_gpu_error.c b/drivers/gpu/drm/i915/i915_gpu_error.c index 20a1f034bf95..881fbe8d179b 100644 --- a/drivers/gpu/drm/i915/i915_gpu_error.c +++ b/drivers/gpu/drm/i915/i915_gpu_error.c @@ -396,6 +396,8 @@ static void error_print_context(struct drm_i915_error_state_buf *m, static void error_print_engine(struct drm_i915_error_state_buf *m, const struct drm_i915_error_engine *ee) { + int n; + err_printf(m, "%s command stream:\n", engine_str(ee->engine_id)); err_printf(m, " START: 0x%08x\n", ee->start); err_printf(m, " HEAD: 0x%08x [0x%08x]\n", ee->head, ee->rq_head); @@ -465,8 +467,11 @@ static void error_print_engine(struct drm_i915_error_state_buf *m, jiffies_to_msecs(jiffies - ee->hangcheck_timestamp)); err_printf(m, " engine reset count: %u\n", ee->reset_count); - error_print_request(m, " ELSP[0]: ", &ee->execlist[0]); - error_print_request(m, " ELSP[1]: ", &ee->execlist[1]); + for (n = 0; n < ee->num_ports; n++) { + err_printf(m, " ELSP[%d]:", n); + error_print_request(m, " ", &ee->execlist[n]); + } + error_print_context(m, " Active context: ", &ee->context); } @@ -1327,17 +1332,19 @@ static void engine_record_requests(struct intel_engine_cs *engine, static void error_record_engine_execlists(struct intel_engine_cs *engine, struct drm_i915_error_engine *ee) { - const struct execlist_port *port = engine->execlists.port; + const struct intel_engine_execlists * const execlists = &engine->execlists; unsigned int n; - for (n = 0; n < ARRAY_SIZE(engine->execlists.port); n++) { - struct drm_i915_gem_request *rq = port_request(&port[n]); + for (n = 0; n < execlists_num_ports(execlists); n++) { + struct drm_i915_gem_request *rq = port_request(&execlists->port[n]); if (!rq) break; record_request(rq, &ee->execlist[n]); } + + ee->num_ports = n; } static void record_context(struct drm_i915_error_context *e, diff --git a/drivers/gpu/drm/i915/i915_guc_submission.c b/drivers/gpu/drm/i915/i915_guc_submission.c index 55e15a57c3d9..04f1281d81a5 100644 --- a/drivers/gpu/drm/i915/i915_guc_submission.c +++ b/drivers/gpu/drm/i915/i915_guc_submission.c @@ -562,6 +562,8 @@ static void i915_guc_dequeue(struct intel_engine_cs *engine) struct intel_engine_execlists * const execlists = &engine->execlists; struct execlist_port *port = execlists->port; struct drm_i915_gem_request *last = NULL; + const struct execlist_port * const last_port = + &execlists->port[execlists->port_mask]; bool submit = false; struct rb_node *rb; @@ -577,7 +579,7 @@ static void i915_guc_dequeue(struct intel_engine_cs *engine) list_for_each_entry_safe(rq, rn, &p->requests, priotree.link) { if (last && rq->ctx != last->ctx) { - if (port != execlists->port) { + if (port == last_port) { __list_del_many(&p->requests, &rq->priotree.link); goto done; @@ -617,6 +619,8 @@ static void i915_guc_irq_handler(unsigned long data) struct intel_engine_cs * const engine = (struct intel_engine_cs *)data; struct intel_engine_execlists * const execlists = &engine->execlists; struct execlist_port *port = execlists->port; + const struct execlist_port * const last_port = + &execlists->port[execlists->port_mask]; struct drm_i915_gem_request *rq; rq = port_request(&port[0]); @@ -629,7 +633,7 @@ static void i915_guc_irq_handler(unsigned long data) rq = port_request(&port[0]); } - if (!port_isset(&port[1])) + if (!port_isset(last_port)) i915_guc_dequeue(engine); } diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c index 30035e59a784..a28e2a864cf1 100644 --- a/drivers/gpu/drm/i915/intel_engine_cs.c +++ b/drivers/gpu/drm/i915/intel_engine_cs.c @@ -405,6 +405,10 @@ static void intel_engine_init_execlist(struct intel_engine_cs *engine) execlists->csb_use_mmio = csb_force_mmio(engine->i915); + execlists->port_mask = 1; + BUILD_BUG_ON_NOT_POWER_OF_2(execlists_num_ports(execlists)); + GEM_BUG_ON(execlists_num_ports(execlists) > EXECLIST_MAX_PORTS); + execlists->queue = RB_ROOT; execlists->first = NULL; } diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c index 3b03f19f1395..f91e126a7f97 100644 --- a/drivers/gpu/drm/i915/intel_lrc.c +++ b/drivers/gpu/drm/i915/intel_lrc.c @@ -399,7 +399,7 @@ static void execlists_submit_ports(struct intel_engine_cs *engine) engine->i915->regs + i915_mmio_reg_offset(RING_ELSP(engine)); unsigned int n; - for (n = ARRAY_SIZE(engine->execlists.port); n--; ) { + for (n = execlists_num_ports(&engine->execlists); n--; ) { struct drm_i915_gem_request *rq; unsigned int count; u64 desc; @@ -456,6 +456,8 @@ static void execlists_dequeue(struct intel_engine_cs *engine) struct drm_i915_gem_request *last; struct intel_engine_execlists * const execlists = &engine->execlists; struct execlist_port *port = execlists->port; + const struct execlist_port * const last_port = + &execlists->port[execlists->port_mask]; struct rb_node *rb; bool submit = false; @@ -515,7 +517,7 @@ static void execlists_dequeue(struct intel_engine_cs *engine) * combine this request with the last, then we * are done. */ - if (port != execlists->port) { + if (port == last_port) { __list_del_many(&p->requests, &rq->priotree.link); goto done; diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.h b/drivers/gpu/drm/i915/intel_ringbuffer.h index 0eae5936bc3c..56d7ae9f298b 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.h +++ b/drivers/gpu/drm/i915/intel_ringbuffer.h @@ -234,7 +234,14 @@ struct intel_engine_execlists { * @context_id: context ID for port */ GEM_DEBUG_DECL(u32 context_id); - } port[2]; + +#define EXECLIST_MAX_PORTS 2 + } port[EXECLIST_MAX_PORTS]; + + /** + * @port_mask: number of execlist ports - 1 + */ + unsigned int port_mask; /** * @queue: queue of requests, in priority lists @@ -511,16 +518,22 @@ struct intel_engine_cs { u32 (*get_cmd_length_mask)(u32 cmd_header); }; +static inline unsigned int +execlists_num_ports(const struct intel_engine_execlists * const execlists) +{ + return execlists->port_mask + 1; +} + static inline void execlists_port_complete(struct intel_engine_execlists * const execlists, struct execlist_port * const port) { - struct execlist_port * const port1 = &execlists->port[1]; + const unsigned int m = execlists->port_mask; GEM_BUG_ON(port_index(port, execlists) != 0); - *port = *port1; - memset(port1, 0, sizeof(struct execlist_port)); + memmove(port, port + 1, m * sizeof(struct execlist_port)); + memset(port + m, 0, sizeof(struct execlist_port)); } static inline unsigned int -- cgit v1.2.3 From 32ced39c1b122679f829cfdac5a679b3a5aefeaf Mon Sep 17 00:00:00 2001 From: Oscar Mateo Date: Thu, 28 Sep 2017 15:40:39 -0700 Subject: drm/i915: Transform whitelisting WAs into a simple reg write RING_FORCE_TO_NONPRIV registers do not live in the logical context. They are simply global privileged MMIO registers that happen to be powercontext saved and restored (meaning only they can survive RC6). Therefore, there is absolutely no need to save them so that they can be restored everytime we create a new logical context. Suggested-by: Chris Wilson Signed-off-by: Oscar Mateo Cc: Mika Kuoppala Link: https://patchwork.freedesktop.org/patch/msgid/1506638439-6903-1-git-send-email-oscar.mateo@intel.com Acked-by: Michel Thierry Tested-by: Chris Wilson #bxt Reviewed-by: Mika Kuoppala Signed-off-by: Chris Wilson --- drivers/gpu/drm/i915/intel_engine_cs.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_engine_cs.c') diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c index a28e2a864cf1..a75f5e889927 100644 --- a/drivers/gpu/drm/i915/intel_engine_cs.c +++ b/drivers/gpu/drm/i915/intel_engine_cs.c @@ -845,8 +845,8 @@ static int wa_ring_whitelist_reg(struct intel_engine_cs *engine, if (WARN_ON(index >= RING_MAX_NONPRIV_SLOTS)) return -EINVAL; - WA_WRITE(RING_FORCE_TO_NONPRIV(engine->mmio_base, index), - i915_mmio_reg_offset(reg)); + I915_WRITE(RING_FORCE_TO_NONPRIV(engine->mmio_base, index), + i915_mmio_reg_offset(reg)); wa->hw_whitelist_count[engine->id]++; return 0; -- cgit v1.2.3 From 53221e11c7a0e85004c1a28f74e4e173f098d262 Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Wed, 4 Oct 2017 13:41:52 +0100 Subject: drm/i915: Move MMCD_MISC_CTRL from context w/a to standard Looking at gem_workarounds shows us that MMCD_MISC_CTRL is not restored following a suspend-resume cycle. This implies that MMCD_MISC_CTRL is not stored in the context, but is an ordinary register w/a that we need to restore during init_hw. Signed-off-by: Chris Wilson Cc: Mika Kuoppala Cc: Oscar Mateo Link: https://patchwork.freedesktop.org/patch/msgid/20171004124153.14142-1-chris@chris-wilson.co.uk Reviewed-by: Mika Kuoppala --- drivers/gpu/drm/i915/intel_engine_cs.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) (limited to 'drivers/gpu/drm/i915/intel_engine_cs.c') diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c index a75f5e889927..8625feb0939e 100644 --- a/drivers/gpu/drm/i915/intel_engine_cs.c +++ b/drivers/gpu/drm/i915/intel_engine_cs.c @@ -980,7 +980,11 @@ static int gen9_init_workarounds(struct intel_engine_cs *engine) GEN9_PBE_COMPRESSED_HASH_SELECTION); WA_SET_BIT_MASKED(GEN9_HALF_SLICE_CHICKEN7, GEN9_SAMPLER_HASH_COMPRESSED_READ_ADDR); - WA_SET_BIT(MMCD_MISC_CTRL, MMCD_PCLA | MMCD_HOTSPOT_EN); + + I915_WRITE(MMCD_MISC_CTRL, + I915_READ(MMCD_MISC_CTRL) | + MMCD_PCLA | + MMCD_HOTSPOT_EN); } /* WaClearFlowControlGpgpuContextSave:skl,bxt,kbl,glk,cfl */ -- cgit v1.2.3 From 8d488bbec73fa36907e1b327e919bcd522b4a57f Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Wed, 4 Oct 2017 13:41:53 +0100 Subject: drm/i915: Remove WA_(SET|CLR)_BIT These macros are of dubious merit when coupled with the per-context w/a set. Instead of tweaking the value in the context, they tweak the value based on the mmio at the time of recording; they are almost by definition not per-context! Having removed the last users, remove the macros to avoid temptation in the future. v2: Kill WA_WRITE as well (now also unused). Signed-off-by: Chris Wilson Cc: Mika Kuoppala Cc: Oscar Mateo Link: https://patchwork.freedesktop.org/patch/msgid/20171004124153.14142-2-chris@chris-wilson.co.uk Reviewed-by: Mika Kuoppala --- drivers/gpu/drm/i915/intel_engine_cs.c | 5 ----- 1 file changed, 5 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_engine_cs.c') diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c index 8625feb0939e..e804a9b816f1 100644 --- a/drivers/gpu/drm/i915/intel_engine_cs.c +++ b/drivers/gpu/drm/i915/intel_engine_cs.c @@ -830,11 +830,6 @@ static int wa_add(struct drm_i915_private *dev_priv, #define WA_SET_FIELD_MASKED(addr, mask, value) \ WA_REG(addr, mask, _MASKED_FIELD(mask, value)) -#define WA_SET_BIT(addr, mask) WA_REG(addr, mask, I915_READ(addr) | (mask)) -#define WA_CLR_BIT(addr, mask) WA_REG(addr, mask, I915_READ(addr) & ~(mask)) - -#define WA_WRITE(addr, val) WA_REG(addr, 0xffffffff, val) - static int wa_ring_whitelist_reg(struct intel_engine_cs *engine, i915_reg_t reg) { -- cgit v1.2.3 From 1e998343f95b46497c56a21de1b14a302256f973 Mon Sep 17 00:00:00 2001 From: Jeff McGee Date: Tue, 3 Oct 2017 21:34:45 +0100 Subject: drm/i915/preempt: Fix WaEnablePreemptionGranularityControlByUMD MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The WA applies to all production Gen9 and requires both enabling and whitelisting of the per-context preemption control register. v2: Extend to Cannonlake. Signed-off-by: Jeff McGee Signed-off-by: Michał Winiarski Signed-off-by: Chris Wilson Reviewed-by: Joonas Lahtinen Link: https://patchwork.freedesktop.org/patch/msgid/20171003203453.15692-1-chris@chris-wilson.co.uk --- drivers/gpu/drm/i915/intel_engine_cs.c | 16 ++++++---------- 1 file changed, 6 insertions(+), 10 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_engine_cs.c') diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c index e804a9b816f1..6245970eb3ab 100644 --- a/drivers/gpu/drm/i915/intel_engine_cs.c +++ b/drivers/gpu/drm/i915/intel_engine_cs.c @@ -1075,8 +1075,10 @@ static int gen9_init_workarounds(struct intel_engine_cs *engine) if (ret) return ret; - /* WaEnablePreemptionGranularityControlByUMD:skl,bxt,kbl,cfl */ - ret= wa_ring_whitelist_reg(engine, GEN8_CS_CHICKEN1); + /* WaEnablePreemptionGranularityControlByUMD:skl,bxt,kbl,cfl,[cnl] */ + I915_WRITE(GEN7_FF_SLICE_CS_CHICKEN1, + _MASKED_BIT_ENABLE(GEN9_FFSC_PERCTX_PREEMPT_CTRL)); + ret = wa_ring_whitelist_reg(engine, GEN8_CS_CHICKEN1); if (ret) return ret; @@ -1138,14 +1140,6 @@ static int skl_init_workarounds(struct intel_engine_cs *engine) if (ret) return ret; - /* - * Actual WA is to disable percontext preemption granularity control - * until D0 which is the default case so this is equivalent to - * !WaDisablePerCtxtPreemptionGranularityControl:skl - */ - I915_WRITE(GEN7_FF_SLICE_CS_CHICKEN1, - _MASKED_BIT_ENABLE(GEN9_FFSC_PERCTX_PREEMPT_CTRL)); - /* WaEnableGapsTsvCreditFix:skl */ I915_WRITE(GEN8_GARBCNTL, (I915_READ(GEN8_GARBCNTL) | GEN9_GAPS_TSV_CREDIT_DISABLE)); @@ -1278,6 +1272,8 @@ static int cnl_init_workarounds(struct intel_engine_cs *engine) WA_SET_BIT_MASKED(HALF_SLICE_CHICKEN3, CNL_FAST_ANISO_L1_BANKING_FIX); /* WaEnablePreemptionGranularityControlByUMD:cnl */ + I915_WRITE(GEN7_FF_SLICE_CS_CHICKEN1, + _MASKED_BIT_ENABLE(GEN9_FFSC_PERCTX_PREEMPT_CTRL)); ret= wa_ring_whitelist_reg(engine, GEN8_CS_CHICKEN1); if (ret) return ret; -- cgit v1.2.3 From 5152defe4a53ad15e6d96c422440152302c8abd7 Mon Sep 17 00:00:00 2001 From: Michał Winiarski Date: Tue, 3 Oct 2017 21:34:46 +0100 Subject: drm/i915/preempt: Default to disabled mid-command preemption levels MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Supporting fine-granularity preemption levels may require changes in userspace batch buffer programming. Therefore, we need to fallback to safe default values, rather that use hardware defaults. Userspace is still able to enable fine-granularity, since we're whitelisting the register controlling it in WaEnablePreemptionGranularityControlByUMD. v2: Extend w/a to cover Cannonlake v3: Fix commentary to include both fake w/a names. Signed-off-by: Michał Winiarski Signed-off-by: Chris Wilson Reviewed-by: Joonas Lahtinen Link: https://patchwork.freedesktop.org/patch/msgid/20171003203453.15692-2-chris@chris-wilson.co.uk --- drivers/gpu/drm/i915/i915_reg.h | 6 ++++++ drivers/gpu/drm/i915/intel_engine_cs.c | 25 +++++++++++++++++++++++++ 2 files changed, 31 insertions(+) (limited to 'drivers/gpu/drm/i915/intel_engine_cs.c') diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index 39ad9327e2a0..e7dba5539b11 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -7003,6 +7003,12 @@ enum { #define GEN9_CS_DEBUG_MODE1 _MMIO(0x20ec) #define GEN9_CTX_PREEMPT_REG _MMIO(0x2248) #define GEN8_CS_CHICKEN1 _MMIO(0x2580) +#define GEN9_PREEMPT_3D_OBJECT_LEVEL (1<<0) +#define GEN9_PREEMPT_GPGPU_LEVEL(hi, lo) (((hi) << 2) | ((lo) << 1)) +#define GEN9_PREEMPT_GPGPU_MID_THREAD_LEVEL GEN9_PREEMPT_GPGPU_LEVEL(0, 0) +#define GEN9_PREEMPT_GPGPU_THREAD_GROUP_LEVEL GEN9_PREEMPT_GPGPU_LEVEL(0, 1) +#define GEN9_PREEMPT_GPGPU_COMMAND_LEVEL GEN9_PREEMPT_GPGPU_LEVEL(1, 0) +#define GEN9_PREEMPT_GPGPU_LEVEL_MASK GEN9_PREEMPT_GPGPU_LEVEL(1, 1) /* GEN7 chicken */ #define GEN7_COMMON_SLICE_CHICKEN1 _MMIO(0x7010) diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c index 6245970eb3ab..dd64e3d13aa9 100644 --- a/drivers/gpu/drm/i915/intel_engine_cs.c +++ b/drivers/gpu/drm/i915/intel_engine_cs.c @@ -1070,6 +1070,24 @@ static int gen9_init_workarounds(struct intel_engine_cs *engine) I915_WRITE(GEN8_L3SQCREG4, (I915_READ(GEN8_L3SQCREG4) | GEN8_LQSC_FLUSH_COHERENT_LINES)); + /* + * Supporting preemption with fine-granularity requires changes in the + * batch buffer programming. Since we can't break old userspace, we + * need to set our default preemption level to safe value. Userspace is + * still able to use more fine-grained preemption levels, since in + * WaEnablePreemptionGranularityControlByUMD we're whitelisting the + * per-ctx register. As such, WaDisable{3D,GPGPU}MidCmdPreemption are + * not real HW workarounds, but merely a way to start using preemption + * while maintaining old contract with userspace. + */ + + /* WaDisable3DMidCmdPreemption:skl,bxt,glk,cfl,[cnl] */ + WA_CLR_BIT_MASKED(GEN8_CS_CHICKEN1, GEN9_PREEMPT_3D_OBJECT_LEVEL); + + /* WaDisableGPGPUMidCmdPreemption:skl,bxt,blk,cfl,[cnl] */ + WA_SET_FIELD_MASKED(GEN8_CS_CHICKEN1, GEN9_PREEMPT_GPGPU_LEVEL_MASK, + GEN9_PREEMPT_GPGPU_COMMAND_LEVEL); + /* WaVFEStateAfterPipeControlwithMediaStateClear:skl,bxt,glk,cfl */ ret = wa_ring_whitelist_reg(engine, GEN9_CTX_PREEMPT_REG); if (ret) @@ -1271,6 +1289,13 @@ static int cnl_init_workarounds(struct intel_engine_cs *engine) /* FtrEnableFastAnisoL1BankingFix: cnl */ WA_SET_BIT_MASKED(HALF_SLICE_CHICKEN3, CNL_FAST_ANISO_L1_BANKING_FIX); + /* WaDisable3DMidCmdPreemption:cnl */ + WA_CLR_BIT_MASKED(GEN8_CS_CHICKEN1, GEN9_PREEMPT_3D_OBJECT_LEVEL); + + /* WaDisableGPGPUMidCmdPreemption:cnl */ + WA_SET_FIELD_MASKED(GEN8_CS_CHICKEN1, GEN9_PREEMPT_GPGPU_LEVEL_MASK, + GEN9_PREEMPT_GPGPU_COMMAND_LEVEL); + /* WaEnablePreemptionGranularityControlByUMD:cnl */ I915_WRITE(GEN7_FF_SLICE_CS_CHICKEN1, _MASKED_BIT_ENABLE(GEN9_FFSC_PERCTX_PREEMPT_CTRL)); -- cgit v1.2.3 From e7af3116836fb7feb985497f2c7776751fb27ef3 Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Tue, 3 Oct 2017 21:34:48 +0100 Subject: drm/i915: Introduce a preempt context MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Add another perma-pinned context for using for preemption at any time. We cannot just reuse the existing kernel context, as first and foremost we need to ensure that we can preempt the kernel context itself, so require a distinct context id. Similar to the kernel context, we may want to interrupt execution and switch to the preempt context at any time, and so it needs to be permanently pinned and available. To compensate for yet another permanent allocation, we shrink the existing context and the new context by reducing their ringbuffer to the minimum. v2: Assert that we never allocate a request from the preemption context. v3: Limit perma-pin to engines that may preempt. v4: Onion cleanup for early driver death v5: Onion ordering in main driver cleanup as well. Signed-off-by: Chris Wilson Reviewed-by: Michał Winiarski Reviewed-by: Joonas Lahtinen Link: https://patchwork.freedesktop.org/patch/msgid/20171003203453.15692-4-chris@chris-wilson.co.uk --- drivers/gpu/drm/i915/i915_drv.h | 6 ++- drivers/gpu/drm/i915/i915_gem_context.c | 76 ++++++++++++++++++++++++--------- drivers/gpu/drm/i915/i915_gem_request.c | 7 +++ drivers/gpu/drm/i915/intel_engine_cs.c | 22 +++++++++- 4 files changed, 87 insertions(+), 24 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_engine_cs.c') diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h index 7ca11318ac69..31292afb961d 100644 --- a/drivers/gpu/drm/i915/i915_drv.h +++ b/drivers/gpu/drm/i915/i915_drv.h @@ -783,6 +783,7 @@ struct intel_csr { func(has_l3_dpf); \ func(has_llc); \ func(has_logical_ring_contexts); \ + func(has_logical_ring_preemption); \ func(has_overlay); \ func(has_pipe_cxsr); \ func(has_pooled_eu); \ @@ -2251,8 +2252,11 @@ struct drm_i915_private { wait_queue_head_t gmbus_wait_queue; struct pci_dev *bridge_dev; - struct i915_gem_context *kernel_context; struct intel_engine_cs *engine[I915_NUM_ENGINES]; + /* Context used internally to idle the GPU and setup initial state */ + struct i915_gem_context *kernel_context; + /* Context only to be used for injecting preemption commands */ + struct i915_gem_context *preempt_context; struct i915_vma *semaphore; struct drm_dma_handle *status_page_dmah; diff --git a/drivers/gpu/drm/i915/i915_gem_context.c b/drivers/gpu/drm/i915/i915_gem_context.c index 921ee369c74d..2bb8e58706ba 100644 --- a/drivers/gpu/drm/i915/i915_gem_context.c +++ b/drivers/gpu/drm/i915/i915_gem_context.c @@ -416,14 +416,43 @@ out: return ctx; } +static struct i915_gem_context * +create_kernel_context(struct drm_i915_private *i915, int prio) +{ + struct i915_gem_context *ctx; + + ctx = i915_gem_create_context(i915, NULL); + if (IS_ERR(ctx)) + return ctx; + + i915_gem_context_clear_bannable(ctx); + ctx->priority = prio; + ctx->ring_size = PAGE_SIZE; + + GEM_BUG_ON(!i915_gem_context_is_kernel(ctx)); + + return ctx; +} + +static void +destroy_kernel_context(struct i915_gem_context **ctxp) +{ + struct i915_gem_context *ctx; + + /* Keep the context ref so that we can free it immediately ourselves */ + ctx = i915_gem_context_get(fetch_and_zero(ctxp)); + GEM_BUG_ON(!i915_gem_context_is_kernel(ctx)); + + context_close(ctx); + i915_gem_context_free(ctx); +} + int i915_gem_contexts_init(struct drm_i915_private *dev_priv) { struct i915_gem_context *ctx; + int err; - /* Init should only be called once per module load. Eventually the - * restriction on the context_disabled check can be loosened. */ - if (WARN_ON(dev_priv->kernel_context)) - return 0; + GEM_BUG_ON(dev_priv->kernel_context); INIT_LIST_HEAD(&dev_priv->contexts.list); INIT_WORK(&dev_priv->contexts.free_work, contexts_free_worker); @@ -441,28 +470,38 @@ int i915_gem_contexts_init(struct drm_i915_private *dev_priv) BUILD_BUG_ON(MAX_CONTEXT_HW_ID > INT_MAX); ida_init(&dev_priv->contexts.hw_ida); - ctx = i915_gem_create_context(dev_priv, NULL); + /* lowest priority; idle task */ + ctx = create_kernel_context(dev_priv, I915_PRIORITY_MIN); if (IS_ERR(ctx)) { - DRM_ERROR("Failed to create default global context (error %ld)\n", - PTR_ERR(ctx)); - return PTR_ERR(ctx); + DRM_ERROR("Failed to create default global context\n"); + err = PTR_ERR(ctx); + goto err; } - - /* For easy recognisablity, we want the kernel context to be 0 and then + /* + * For easy recognisablity, we want the kernel context to be 0 and then * all user contexts will have non-zero hw_id. */ GEM_BUG_ON(ctx->hw_id); - - i915_gem_context_clear_bannable(ctx); - ctx->priority = I915_PRIORITY_MIN; /* lowest priority; idle task */ dev_priv->kernel_context = ctx; - GEM_BUG_ON(!i915_gem_context_is_kernel(ctx)); + /* highest priority; preempting task */ + ctx = create_kernel_context(dev_priv, INT_MAX); + if (IS_ERR(ctx)) { + DRM_ERROR("Failed to create default preempt context\n"); + err = PTR_ERR(ctx); + goto err_kernel_context; + } + dev_priv->preempt_context = ctx; DRM_DEBUG_DRIVER("%s context support initialized\n", dev_priv->engine[RCS]->context_size ? "logical" : "fake"); return 0; + +err_kernel_context: + destroy_kernel_context(&dev_priv->kernel_context); +err: + return err; } void i915_gem_contexts_lost(struct drm_i915_private *dev_priv) @@ -507,15 +546,10 @@ void i915_gem_contexts_lost(struct drm_i915_private *dev_priv) void i915_gem_contexts_fini(struct drm_i915_private *i915) { - struct i915_gem_context *ctx; - lockdep_assert_held(&i915->drm.struct_mutex); - /* Keep the context so that we can free it immediately ourselves */ - ctx = i915_gem_context_get(fetch_and_zero(&i915->kernel_context)); - GEM_BUG_ON(!i915_gem_context_is_kernel(ctx)); - context_close(ctx); - i915_gem_context_free(ctx); + destroy_kernel_context(&i915->preempt_context); + destroy_kernel_context(&i915->kernel_context); /* Must free all deferred contexts (via flush_workqueue) first */ ida_destroy(&i915->contexts.hw_ida); diff --git a/drivers/gpu/drm/i915/i915_gem_request.c b/drivers/gpu/drm/i915/i915_gem_request.c index 14956d899911..b100b38f1dd2 100644 --- a/drivers/gpu/drm/i915/i915_gem_request.c +++ b/drivers/gpu/drm/i915/i915_gem_request.c @@ -587,6 +587,13 @@ i915_gem_request_alloc(struct intel_engine_cs *engine, lockdep_assert_held(&dev_priv->drm.struct_mutex); + /* + * Preempt contexts are reserved for exclusive use to inject a + * preemption context switch. They are never to be used for any trivial + * request! + */ + GEM_BUG_ON(ctx == dev_priv->preempt_context); + /* ABI: Before userspace accesses the GPU (e.g. execbuffer), report * EIO if the GPU is already wedged. */ diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c index dd64e3d13aa9..8b8053d8e24b 100644 --- a/drivers/gpu/drm/i915/intel_engine_cs.c +++ b/drivers/gpu/drm/i915/intel_engine_cs.c @@ -613,9 +613,22 @@ int intel_engine_init_common(struct intel_engine_cs *engine) if (IS_ERR(ring)) return PTR_ERR(ring); + /* + * Similarly the preempt context must always be available so that + * we can interrupt the engine at any time. + */ + if (INTEL_INFO(engine->i915)->has_logical_ring_preemption) { + ring = engine->context_pin(engine, + engine->i915->preempt_context); + if (IS_ERR(ring)) { + ret = PTR_ERR(ring); + goto err_unpin_kernel; + } + } + ret = intel_engine_init_breadcrumbs(engine); if (ret) - goto err_unpin; + goto err_unpin_preempt; ret = i915_gem_render_state_init(engine); if (ret) @@ -634,7 +647,10 @@ err_rs_fini: i915_gem_render_state_fini(engine); err_breadcrumbs: intel_engine_fini_breadcrumbs(engine); -err_unpin: +err_unpin_preempt: + if (INTEL_INFO(engine->i915)->has_logical_ring_preemption) + engine->context_unpin(engine, engine->i915->preempt_context); +err_unpin_kernel: engine->context_unpin(engine, engine->i915->kernel_context); return ret; } @@ -660,6 +676,8 @@ void intel_engine_cleanup_common(struct intel_engine_cs *engine) intel_engine_cleanup_cmd_parser(engine); i915_gem_batch_pool_fini(&engine->batch_pool); + if (INTEL_INFO(engine->i915)->has_logical_ring_preemption) + engine->context_unpin(engine, engine->i915->preempt_context); engine->context_unpin(engine, engine->i915->kernel_context); } -- cgit v1.2.3 From 3cf1934abe88103fcbeff88bbaee8a094502e6cb Mon Sep 17 00:00:00 2001 From: Oscar Mateo Date: Wed, 4 Oct 2017 08:39:52 -0700 Subject: drm/i915/cnl: Do not add an extra page for precaution in the Gen10 LRC size BSpec indicates exactly 16752 DWORDs (17 pages), plus one page for PPHWSP. Please notice that, when looking at the BSpec context image table, the right filter has to be applied (e.g. "CNL") as some rows are excluded for specific GENs. BSpec: 1383 v2: Update count and add BSpec tag (Joonas) v3: Warning about filters in the commit message (Joonas) Suggested-by: Joonas Lahtinen Fixes: 7fd0b1a ("drm/i915/cnl: Add Gen10 LRC size") Signed-off-by: Oscar Mateo Cc: Rodrigo Vivi Cc: Daniele Ceraolo Spurio Cc: Ben Widawsky Reviewed-by: Joonas Lahtinen Link: https://patchwork.freedesktop.org/patch/msgid/1507131592-29209-1-git-send-email-oscar.mateo@intel.com Signed-off-by: Chris Wilson --- drivers/gpu/drm/i915/intel_engine_cs.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'drivers/gpu/drm/i915/intel_engine_cs.c') diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c index 8b8053d8e24b..807a7aafc089 100644 --- a/drivers/gpu/drm/i915/intel_engine_cs.c +++ b/drivers/gpu/drm/i915/intel_engine_cs.c @@ -39,7 +39,7 @@ #define GEN8_LR_CONTEXT_RENDER_SIZE (20 * PAGE_SIZE) #define GEN9_LR_CONTEXT_RENDER_SIZE (22 * PAGE_SIZE) -#define GEN10_LR_CONTEXT_RENDER_SIZE (19 * PAGE_SIZE) +#define GEN10_LR_CONTEXT_RENDER_SIZE (18 * PAGE_SIZE) #define GEN8_LR_CONTEXT_OTHER_SIZE ( 2 * PAGE_SIZE) -- cgit v1.2.3 From f636edb214a5ffdc533d5c7198a66957322c1b40 Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Mon, 9 Oct 2017 12:02:57 +0100 Subject: drm/i915: Make i915_engine_info pretty printer to standalone We can use drm_printer to hide the differences between printk and seq_printf, and so make the i915_engine_info pretty printer able to be called from different contexts and not just debugfs. For instance, I want to use the pretty printer to debug kselftests. Signed-off-by: Chris Wilson Cc: Mika Kuoppala Link: https://patchwork.freedesktop.org/patch/msgid/20171009110301.21705-1-chris@chris-wilson.co.uk Reviewed-by: Mika Kuoppala --- drivers/gpu/drm/i915/i915_debugfs.c | 148 +---------------------------- drivers/gpu/drm/i915/intel_engine_cs.c | 160 ++++++++++++++++++++++++++++++++ drivers/gpu/drm/i915/intel_ringbuffer.h | 4 + 3 files changed, 168 insertions(+), 144 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_engine_cs.c') diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c index f7817c667958..9ec2bcd9a695 100644 --- a/drivers/gpu/drm/i915/i915_debugfs.c +++ b/drivers/gpu/drm/i915/i915_debugfs.c @@ -3292,9 +3292,9 @@ static int i915_display_info(struct seq_file *m, void *unused) static int i915_engine_info(struct seq_file *m, void *unused) { struct drm_i915_private *dev_priv = node_to_i915(m->private); - struct i915_gpu_error *error = &dev_priv->gpu_error; struct intel_engine_cs *engine; enum intel_engine_id id; + struct drm_printer p; intel_runtime_pm_get(dev_priv); @@ -3303,149 +3303,9 @@ static int i915_engine_info(struct seq_file *m, void *unused) seq_printf(m, "Global active requests: %d\n", dev_priv->gt.active_requests); - for_each_engine(engine, dev_priv, id) { - struct intel_breadcrumbs *b = &engine->breadcrumbs; - struct drm_i915_gem_request *rq; - struct rb_node *rb; - u64 addr; - - seq_printf(m, "%s\n", engine->name); - seq_printf(m, "\tcurrent seqno %x, last %x, hangcheck %x [%d ms], inflight %d\n", - intel_engine_get_seqno(engine), - intel_engine_last_submit(engine), - engine->hangcheck.seqno, - jiffies_to_msecs(jiffies - engine->hangcheck.action_timestamp), - engine->timeline->inflight_seqnos); - seq_printf(m, "\tReset count: %d\n", - i915_reset_engine_count(error, engine)); - - rcu_read_lock(); - - seq_printf(m, "\tRequests:\n"); - - rq = list_first_entry(&engine->timeline->requests, - struct drm_i915_gem_request, link); - if (&rq->link != &engine->timeline->requests) - print_request(m, rq, "\t\tfirst "); - - rq = list_last_entry(&engine->timeline->requests, - struct drm_i915_gem_request, link); - if (&rq->link != &engine->timeline->requests) - print_request(m, rq, "\t\tlast "); - - rq = i915_gem_find_active_request(engine); - if (rq) { - print_request(m, rq, "\t\tactive "); - seq_printf(m, - "\t\t[head %04x, postfix %04x, tail %04x, batch 0x%08x_%08x]\n", - rq->head, rq->postfix, rq->tail, - rq->batch ? upper_32_bits(rq->batch->node.start) : ~0u, - rq->batch ? lower_32_bits(rq->batch->node.start) : ~0u); - } - - seq_printf(m, "\tRING_START: 0x%08x [0x%08x]\n", - I915_READ(RING_START(engine->mmio_base)), - rq ? i915_ggtt_offset(rq->ring->vma) : 0); - seq_printf(m, "\tRING_HEAD: 0x%08x [0x%08x]\n", - I915_READ(RING_HEAD(engine->mmio_base)) & HEAD_ADDR, - rq ? rq->ring->head : 0); - seq_printf(m, "\tRING_TAIL: 0x%08x [0x%08x]\n", - I915_READ(RING_TAIL(engine->mmio_base)) & TAIL_ADDR, - rq ? rq->ring->tail : 0); - seq_printf(m, "\tRING_CTL: 0x%08x [%s]\n", - I915_READ(RING_CTL(engine->mmio_base)), - I915_READ(RING_CTL(engine->mmio_base)) & (RING_WAIT | RING_WAIT_SEMAPHORE) ? "waiting" : ""); - - rcu_read_unlock(); - - addr = intel_engine_get_active_head(engine); - seq_printf(m, "\tACTHD: 0x%08x_%08x\n", - upper_32_bits(addr), lower_32_bits(addr)); - addr = intel_engine_get_last_batch_head(engine); - seq_printf(m, "\tBBADDR: 0x%08x_%08x\n", - upper_32_bits(addr), lower_32_bits(addr)); - - if (i915_modparams.enable_execlists) { - const u32 *hws = &engine->status_page.page_addr[I915_HWS_CSB_BUF0_INDEX]; - struct intel_engine_execlists * const execlists = &engine->execlists; - u32 ptr, read, write; - unsigned int idx; - - seq_printf(m, "\tExeclist status: 0x%08x %08x\n", - I915_READ(RING_EXECLIST_STATUS_LO(engine)), - I915_READ(RING_EXECLIST_STATUS_HI(engine))); - - ptr = I915_READ(RING_CONTEXT_STATUS_PTR(engine)); - read = GEN8_CSB_READ_PTR(ptr); - write = GEN8_CSB_WRITE_PTR(ptr); - seq_printf(m, "\tExeclist CSB read %d [%d cached], write %d [%d from hws], interrupt posted? %s\n", - read, execlists->csb_head, - write, - intel_read_status_page(engine, intel_hws_csb_write_index(engine->i915)), - yesno(test_bit(ENGINE_IRQ_EXECLIST, - &engine->irq_posted))); - if (read >= GEN8_CSB_ENTRIES) - read = 0; - if (write >= GEN8_CSB_ENTRIES) - write = 0; - if (read > write) - write += GEN8_CSB_ENTRIES; - while (read < write) { - idx = ++read % GEN8_CSB_ENTRIES; - seq_printf(m, "\tExeclist CSB[%d]: 0x%08x [0x%08x in hwsp], context: %d [%d in hwsp]\n", - idx, - I915_READ(RING_CONTEXT_STATUS_BUF_LO(engine, idx)), - hws[idx * 2], - I915_READ(RING_CONTEXT_STATUS_BUF_HI(engine, idx)), - hws[idx * 2 + 1]); - } - - rcu_read_lock(); - for (idx = 0; idx < execlists_num_ports(execlists); idx++) { - unsigned int count; - - rq = port_unpack(&execlists->port[idx], &count); - if (rq) { - seq_printf(m, "\t\tELSP[%d] count=%d, ", - idx, count); - print_request(m, rq, "rq: "); - } else { - seq_printf(m, "\t\tELSP[%d] idle\n", - idx); - } - } - rcu_read_unlock(); - - spin_lock_irq(&engine->timeline->lock); - for (rb = execlists->first; rb; rb = rb_next(rb)) { - struct i915_priolist *p = - rb_entry(rb, typeof(*p), node); - - list_for_each_entry(rq, &p->requests, - priotree.link) - print_request(m, rq, "\t\tQ "); - } - spin_unlock_irq(&engine->timeline->lock); - } else if (INTEL_GEN(dev_priv) > 6) { - seq_printf(m, "\tPP_DIR_BASE: 0x%08x\n", - I915_READ(RING_PP_DIR_BASE(engine))); - seq_printf(m, "\tPP_DIR_BASE_READ: 0x%08x\n", - I915_READ(RING_PP_DIR_BASE_READ(engine))); - seq_printf(m, "\tPP_DIR_DCLV: 0x%08x\n", - I915_READ(RING_PP_DIR_DCLV(engine))); - } - - spin_lock_irq(&b->rb_lock); - for (rb = rb_first(&b->waiters); rb; rb = rb_next(rb)) { - struct intel_wait *w = rb_entry(rb, typeof(*w), node); - - seq_printf(m, "\t%s [%d] waiting for %x\n", - w->tsk->comm, w->tsk->pid, w->seqno); - } - spin_unlock_irq(&b->rb_lock); - - seq_puts(m, "\n"); - } + p = drm_seq_file_printer(m); + for_each_engine(engine, dev_priv, id) + intel_engine_dump(engine, &p); intel_runtime_pm_put(dev_priv); diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c index 807a7aafc089..a59b2a30ff5a 100644 --- a/drivers/gpu/drm/i915/intel_engine_cs.c +++ b/drivers/gpu/drm/i915/intel_engine_cs.c @@ -22,6 +22,8 @@ * */ +#include + #include "i915_drv.h" #include "intel_ringbuffer.h" #include "intel_lrc.h" @@ -1616,6 +1618,164 @@ bool intel_engine_can_store_dword(struct intel_engine_cs *engine) } } +static void print_request(struct drm_printer *m, + struct drm_i915_gem_request *rq, + const char *prefix) +{ + drm_printf(m, "%s%x [%x:%x] prio=%d @ %dms: %s\n", prefix, + rq->global_seqno, rq->ctx->hw_id, rq->fence.seqno, + rq->priotree.priority, + jiffies_to_msecs(jiffies - rq->emitted_jiffies), + rq->timeline->common->name); +} + +void intel_engine_dump(struct intel_engine_cs *engine, struct drm_printer *m) +{ + struct intel_breadcrumbs *b = &engine->breadcrumbs; + struct i915_gpu_error *error = &engine->i915->gpu_error; + struct drm_i915_private *dev_priv = engine->i915; + struct drm_i915_gem_request *rq; + struct rb_node *rb; + u64 addr; + + drm_printf(m, "%s\n", engine->name); + drm_printf(m, "\tcurrent seqno %x, last %x, hangcheck %x [%d ms], inflight %d\n", + intel_engine_get_seqno(engine), + intel_engine_last_submit(engine), + engine->hangcheck.seqno, + jiffies_to_msecs(jiffies - engine->hangcheck.action_timestamp), + engine->timeline->inflight_seqnos); + drm_printf(m, "\tReset count: %d\n", + i915_reset_engine_count(error, engine)); + + rcu_read_lock(); + + drm_printf(m, "\tRequests:\n"); + + rq = list_first_entry(&engine->timeline->requests, + struct drm_i915_gem_request, link); + if (&rq->link != &engine->timeline->requests) + print_request(m, rq, "\t\tfirst "); + + rq = list_last_entry(&engine->timeline->requests, + struct drm_i915_gem_request, link); + if (&rq->link != &engine->timeline->requests) + print_request(m, rq, "\t\tlast "); + + rq = i915_gem_find_active_request(engine); + if (rq) { + print_request(m, rq, "\t\tactive "); + drm_printf(m, + "\t\t[head %04x, postfix %04x, tail %04x, batch 0x%08x_%08x]\n", + rq->head, rq->postfix, rq->tail, + rq->batch ? upper_32_bits(rq->batch->node.start) : ~0u, + rq->batch ? lower_32_bits(rq->batch->node.start) : ~0u); + } + + drm_printf(m, "\tRING_START: 0x%08x [0x%08x]\n", + I915_READ(RING_START(engine->mmio_base)), + rq ? i915_ggtt_offset(rq->ring->vma) : 0); + drm_printf(m, "\tRING_HEAD: 0x%08x [0x%08x]\n", + I915_READ(RING_HEAD(engine->mmio_base)) & HEAD_ADDR, + rq ? rq->ring->head : 0); + drm_printf(m, "\tRING_TAIL: 0x%08x [0x%08x]\n", + I915_READ(RING_TAIL(engine->mmio_base)) & TAIL_ADDR, + rq ? rq->ring->tail : 0); + drm_printf(m, "\tRING_CTL: 0x%08x [%s]\n", + I915_READ(RING_CTL(engine->mmio_base)), + I915_READ(RING_CTL(engine->mmio_base)) & (RING_WAIT | RING_WAIT_SEMAPHORE) ? "waiting" : ""); + + rcu_read_unlock(); + + addr = intel_engine_get_active_head(engine); + drm_printf(m, "\tACTHD: 0x%08x_%08x\n", + upper_32_bits(addr), lower_32_bits(addr)); + addr = intel_engine_get_last_batch_head(engine); + drm_printf(m, "\tBBADDR: 0x%08x_%08x\n", + upper_32_bits(addr), lower_32_bits(addr)); + + if (i915_modparams.enable_execlists) { + const u32 *hws = &engine->status_page.page_addr[I915_HWS_CSB_BUF0_INDEX]; + struct intel_engine_execlists * const execlists = &engine->execlists; + u32 ptr, read, write; + unsigned int idx; + + drm_printf(m, "\tExeclist status: 0x%08x %08x\n", + I915_READ(RING_EXECLIST_STATUS_LO(engine)), + I915_READ(RING_EXECLIST_STATUS_HI(engine))); + + ptr = I915_READ(RING_CONTEXT_STATUS_PTR(engine)); + read = GEN8_CSB_READ_PTR(ptr); + write = GEN8_CSB_WRITE_PTR(ptr); + drm_printf(m, "\tExeclist CSB read %d [%d cached], write %d [%d from hws], interrupt posted? %s\n", + read, execlists->csb_head, + write, + intel_read_status_page(engine, intel_hws_csb_write_index(engine->i915)), + yesno(test_bit(ENGINE_IRQ_EXECLIST, + &engine->irq_posted))); + if (read >= GEN8_CSB_ENTRIES) + read = 0; + if (write >= GEN8_CSB_ENTRIES) + write = 0; + if (read > write) + write += GEN8_CSB_ENTRIES; + while (read < write) { + idx = ++read % GEN8_CSB_ENTRIES; + drm_printf(m, "\tExeclist CSB[%d]: 0x%08x [0x%08x in hwsp], context: %d [%d in hwsp]\n", + idx, + I915_READ(RING_CONTEXT_STATUS_BUF_LO(engine, idx)), + hws[idx * 2], + I915_READ(RING_CONTEXT_STATUS_BUF_HI(engine, idx)), + hws[idx * 2 + 1]); + } + + rcu_read_lock(); + for (idx = 0; idx < execlists_num_ports(execlists); idx++) { + unsigned int count; + + rq = port_unpack(&execlists->port[idx], &count); + if (rq) { + drm_printf(m, "\t\tELSP[%d] count=%d, ", + idx, count); + print_request(m, rq, "rq: "); + } else { + drm_printf(m, "\t\tELSP[%d] idle\n", + idx); + } + } + rcu_read_unlock(); + + spin_lock_irq(&engine->timeline->lock); + for (rb = execlists->first; rb; rb = rb_next(rb)) { + struct i915_priolist *p = + rb_entry(rb, typeof(*p), node); + + list_for_each_entry(rq, &p->requests, + priotree.link) + print_request(m, rq, "\t\tQ "); + } + spin_unlock_irq(&engine->timeline->lock); + } else if (INTEL_GEN(dev_priv) > 6) { + drm_printf(m, "\tPP_DIR_BASE: 0x%08x\n", + I915_READ(RING_PP_DIR_BASE(engine))); + drm_printf(m, "\tPP_DIR_BASE_READ: 0x%08x\n", + I915_READ(RING_PP_DIR_BASE_READ(engine))); + drm_printf(m, "\tPP_DIR_DCLV: 0x%08x\n", + I915_READ(RING_PP_DIR_DCLV(engine))); + } + + spin_lock_irq(&b->rb_lock); + for (rb = rb_first(&b->waiters); rb; rb = rb_next(rb)) { + struct intel_wait *w = rb_entry(rb, typeof(*w), node); + + drm_printf(m, "\t%s [%d] waiting for %x\n", + w->tsk->comm, w->tsk->pid, w->seqno); + } + spin_unlock_irq(&b->rb_lock); + + drm_printf(m, "\n"); +} + #if IS_ENABLED(CONFIG_DRM_I915_SELFTEST) #include "selftests/mock_engine.c" #endif diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.h b/drivers/gpu/drm/i915/intel_ringbuffer.h index 0fedda17488c..17186f067408 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.h +++ b/drivers/gpu/drm/i915/intel_ringbuffer.h @@ -7,6 +7,8 @@ #include "i915_gem_timeline.h" #include "i915_selftest.h" +struct drm_printer; + #define I915_CMD_HASH_ORDER 9 /* Early gen2 devices have a cacheline of just 32 bytes, using 64 is overkill, @@ -839,4 +841,6 @@ void intel_engines_reset_default_submission(struct drm_i915_private *i915); bool intel_engine_can_store_dword(struct intel_engine_cs *engine); +void intel_engine_dump(struct intel_engine_cs *engine, struct drm_printer *p); + #endif /* _INTEL_RINGBUFFER_H_ */ -- cgit v1.2.3 From 1fd51d9d97059cb7dd0bf7a3b6f7cb609d485718 Mon Sep 17 00:00:00 2001 From: Weinan Li Date: Sun, 15 Oct 2017 11:55:25 +0800 Subject: drm/i915: enable to read CSB and CSB write pointer from HWSP in GVT-g VM Let GVT-g VM read the CSB and CSB write pointer from virtual HWSP, not all the host support this feature, need to check the BIT(3) of caps in PVINFO. v3 : Remove unnecessary comments. v4 : Separate VM enable patch with GVT-g implementation patch due to code dependency. v5 : Use inline for GVT virtual HWSP caps check function. v6 : Comments refine. Signed-off-by: Weinan Li Cc: Chris Wilson Cc: Joonas Lahtinen Reviewed-by: Chris Wilson Signed-off-by: Joonas Lahtinen Link: https://patchwork.freedesktop.org/patch/msgid/1508039725-1066-1-git-send-email-weinan.z.li@intel.com --- drivers/gpu/drm/i915/i915_pvinfo.h | 1 + drivers/gpu/drm/i915/i915_vgpu.h | 6 ++++++ drivers/gpu/drm/i915/intel_engine_cs.c | 9 +++++---- drivers/gpu/drm/i915/intel_lrc.c | 1 - 4 files changed, 12 insertions(+), 5 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_engine_cs.c') diff --git a/drivers/gpu/drm/i915/i915_pvinfo.h b/drivers/gpu/drm/i915/i915_pvinfo.h index 0679a58cdbae..195203f298df 100644 --- a/drivers/gpu/drm/i915/i915_pvinfo.h +++ b/drivers/gpu/drm/i915/i915_pvinfo.h @@ -53,6 +53,7 @@ enum vgt_g2v_type { * VGT capabilities type */ #define VGT_CAPS_FULL_48BIT_PPGTT BIT(2) +#define VGT_CAPS_HWSP_EMULATION BIT(3) struct vgt_if { u64 magic; /* VGT_MAGIC */ diff --git a/drivers/gpu/drm/i915/i915_vgpu.h b/drivers/gpu/drm/i915/i915_vgpu.h index b72bd2956b70..bb8338450dc1 100644 --- a/drivers/gpu/drm/i915/i915_vgpu.h +++ b/drivers/gpu/drm/i915/i915_vgpu.h @@ -30,6 +30,12 @@ void i915_check_vgpu(struct drm_i915_private *dev_priv); bool intel_vgpu_has_full_48bit_ppgtt(struct drm_i915_private *dev_priv); +static inline bool +intel_vgpu_has_hwsp_emulation(struct drm_i915_private *dev_priv) +{ + return dev_priv->vgpu.caps & VGT_CAPS_HWSP_EMULATION; +} + int intel_vgt_balloon(struct drm_i915_private *dev_priv); void intel_vgt_deballoon(struct drm_i915_private *dev_priv); diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c index a59b2a30ff5a..83e696f67d42 100644 --- a/drivers/gpu/drm/i915/intel_engine_cs.c +++ b/drivers/gpu/drm/i915/intel_engine_cs.c @@ -25,6 +25,7 @@ #include #include "i915_drv.h" +#include "i915_vgpu.h" #include "intel_ringbuffer.h" #include "intel_lrc.h" @@ -386,10 +387,6 @@ static void intel_engine_init_timeline(struct intel_engine_cs *engine) static bool csb_force_mmio(struct drm_i915_private *i915) { - /* GVT emulation depends upon intercepting CSB mmio */ - if (intel_vgpu_active(i915)) - return true; - /* * IOMMU adds unpredictable latency causing the CSB write (from the * GPU into the HWSP) to only be visible some time after the interrupt @@ -398,6 +395,10 @@ static bool csb_force_mmio(struct drm_i915_private *i915) if (intel_vtd_active()) return true; + /* Older GVT emulation depends upon intercepting CSB mmio */ + if (intel_vgpu_active(i915) && !intel_vgpu_has_hwsp_emulation(i915)) + return true; + return false; } diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c index fbfcf88d7fe3..766552f2cfae 100644 --- a/drivers/gpu/drm/i915/intel_lrc.c +++ b/drivers/gpu/drm/i915/intel_lrc.c @@ -793,7 +793,6 @@ static void intel_lrc_irq_handler(unsigned long data) &engine->status_page.page_addr[I915_HWS_CSB_BUF0_INDEX]; unsigned int head, tail; - /* However GVT emulation depends upon intercepting CSB mmio */ if (unlikely(execlists->csb_use_mmio)) { buf = (u32 * __force) (dev_priv->regs + i915_mmio_reg_offset(RING_CONTEXT_STATUS_BUF_LO(engine, 0))); -- cgit v1.2.3 From a27d5a44ec87a019d818a82d0475b5d38856691e Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Sun, 15 Oct 2017 21:43:10 +0100 Subject: drm/i915: Add in-flight request details to intel_engine_dump() In the intel_engine_cs dumper, we were showing the request details for the request queue but not of those requests already passed to the hw (just a summary of the seqno). If we show those details, we can then eliminate the entirely redundant and forgotten debugfs/i915_gem_request Signed-off-by: Chris Wilson Cc: Jeff McGee Cc: Mika Kuoppala Link: https://patchwork.freedesktop.org/patch/msgid/20171015204310.17045-1-chris@chris-wilson.co.uk Reviewed-by: Jeff McGee --- drivers/gpu/drm/i915/i915_debugfs.c | 49 ---------------------------------- drivers/gpu/drm/i915/intel_engine_cs.c | 35 +++++++++++++----------- 2 files changed, 19 insertions(+), 65 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_engine_cs.c') diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c index 6d643642e0ce..40287e9f00d7 100644 --- a/drivers/gpu/drm/i915/i915_debugfs.c +++ b/drivers/gpu/drm/i915/i915_debugfs.c @@ -664,54 +664,6 @@ static int i915_gem_batch_pool_info(struct seq_file *m, void *data) return 0; } -static void print_request(struct seq_file *m, - struct drm_i915_gem_request *rq, - const char *prefix) -{ - seq_printf(m, "%s%x [%x:%x] prio=%d @ %dms: %s\n", prefix, - rq->global_seqno, rq->ctx->hw_id, rq->fence.seqno, - rq->priotree.priority, - jiffies_to_msecs(jiffies - rq->emitted_jiffies), - rq->timeline->common->name); -} - -static int i915_gem_request_info(struct seq_file *m, void *data) -{ - struct drm_i915_private *dev_priv = node_to_i915(m->private); - struct drm_device *dev = &dev_priv->drm; - struct drm_i915_gem_request *req; - struct intel_engine_cs *engine; - enum intel_engine_id id; - int ret, any; - - ret = mutex_lock_interruptible(&dev->struct_mutex); - if (ret) - return ret; - - any = 0; - for_each_engine(engine, dev_priv, id) { - int count; - - count = 0; - list_for_each_entry(req, &engine->timeline->requests, link) - count++; - if (count == 0) - continue; - - seq_printf(m, "%s requests: %d\n", engine->name, count); - list_for_each_entry(req, &engine->timeline->requests, link) - print_request(m, req, " "); - - any++; - } - mutex_unlock(&dev->struct_mutex); - - if (any == 0) - seq_puts(m, "No requests\n"); - - return 0; -} - static void i915_ring_seqno_info(struct seq_file *m, struct intel_engine_cs *engine) { @@ -4783,7 +4735,6 @@ static const struct drm_info_list i915_debugfs_list[] = { {"i915_gem_objects", i915_gem_object_info, 0}, {"i915_gem_gtt", i915_gem_gtt_info, 0}, {"i915_gem_stolen", i915_gem_stolen_list_info }, - {"i915_gem_request", i915_gem_request_info, 0}, {"i915_gem_seqno", i915_gem_seqno_info, 0}, {"i915_gem_fence_regs", i915_gem_fence_regs_info, 0}, {"i915_gem_interrupt", i915_interrupt_info, 0}, diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c index 83e696f67d42..9a4c4abc7450 100644 --- a/drivers/gpu/drm/i915/intel_engine_cs.c +++ b/drivers/gpu/drm/i915/intel_engine_cs.c @@ -1623,8 +1623,10 @@ static void print_request(struct drm_printer *m, struct drm_i915_gem_request *rq, const char *prefix) { - drm_printf(m, "%s%x [%x:%x] prio=%d @ %dms: %s\n", prefix, - rq->global_seqno, rq->ctx->hw_id, rq->fence.seqno, + drm_printf(m, "%s%x%s [%x:%x] prio=%d @ %dms: %s\n", prefix, + rq->global_seqno, + i915_gem_request_completed(rq) ? "!" : "", + rq->ctx->hw_id, rq->fence.seqno, rq->priotree.priority, jiffies_to_msecs(jiffies - rq->emitted_jiffies), rq->timeline->common->name); @@ -1632,8 +1634,9 @@ static void print_request(struct drm_printer *m, void intel_engine_dump(struct intel_engine_cs *engine, struct drm_printer *m) { - struct intel_breadcrumbs *b = &engine->breadcrumbs; - struct i915_gpu_error *error = &engine->i915->gpu_error; + struct intel_breadcrumbs * const b = &engine->breadcrumbs; + const struct intel_engine_execlists * const execlists = &engine->execlists; + struct i915_gpu_error * const error = &engine->i915->gpu_error; struct drm_i915_private *dev_priv = engine->i915; struct drm_i915_gem_request *rq; struct rb_node *rb; @@ -1697,7 +1700,6 @@ void intel_engine_dump(struct intel_engine_cs *engine, struct drm_printer *m) if (i915_modparams.enable_execlists) { const u32 *hws = &engine->status_page.page_addr[I915_HWS_CSB_BUF0_INDEX]; - struct intel_engine_execlists * const execlists = &engine->execlists; u32 ptr, read, write; unsigned int idx; @@ -1745,17 +1747,6 @@ void intel_engine_dump(struct intel_engine_cs *engine, struct drm_printer *m) } } rcu_read_unlock(); - - spin_lock_irq(&engine->timeline->lock); - for (rb = execlists->first; rb; rb = rb_next(rb)) { - struct i915_priolist *p = - rb_entry(rb, typeof(*p), node); - - list_for_each_entry(rq, &p->requests, - priotree.link) - print_request(m, rq, "\t\tQ "); - } - spin_unlock_irq(&engine->timeline->lock); } else if (INTEL_GEN(dev_priv) > 6) { drm_printf(m, "\tPP_DIR_BASE: 0x%08x\n", I915_READ(RING_PP_DIR_BASE(engine))); @@ -1765,6 +1756,18 @@ void intel_engine_dump(struct intel_engine_cs *engine, struct drm_printer *m) I915_READ(RING_PP_DIR_DCLV(engine))); } + spin_lock_irq(&engine->timeline->lock); + list_for_each_entry(rq, &engine->timeline->requests, link) + print_request(m, rq, "\t\tE "); + for (rb = execlists->first; rb; rb = rb_next(rb)) { + struct i915_priolist *p = + rb_entry(rb, typeof(*p), node); + + list_for_each_entry(rq, &p->requests, priotree.link) + print_request(m, rq, "\t\tQ "); + } + spin_unlock_irq(&engine->timeline->lock); + spin_lock_irq(&b->rb_lock); for (rb = rb_first(&b->waiters); rb; rb = rb_next(rb)) { struct intel_wait *w = rb_entry(rb, typeof(*w), node); -- cgit v1.2.3 From 930a784d02339be437fec07b3bb7213bde0ed53b Mon Sep 17 00:00:00 2001 From: Oscar Mateo Date: Tue, 17 Oct 2017 13:25:45 -0700 Subject: drm/i915: Use a mask when applying WaProgramL3SqcReg1Default MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Otherwise we are blasting other bits in GEN8_L3SQCREG1 that might be important (although we probably aren't at the moment because 0 seems to be the default for all the other bits). v2: Extra parentheses (Michel) Fixes: 050fc46 ("drm/i915:bxt: implement WaProgramL3SqcReg1DefaultForPerf") Fixes: 450174f ("drm/i915/chv: Tune L3 SQC credits based on actual latencies") Signed-off-by: Oscar Mateo Cc: Chris Wilson Cc: Mika Kuoppala Cc: Ville Syrjälä Cc: Imre Deak Reviewed-by: Michel Thierry Link: https://patchwork.freedesktop.org/patch/msgid/1508271945-14961-1-git-send-email-oscar.mateo@intel.com Signed-off-by: Chris Wilson --- drivers/gpu/drm/i915/i915_reg.h | 1 + drivers/gpu/drm/i915/intel_engine_cs.c | 9 ++++++--- drivers/gpu/drm/i915/intel_pm.c | 9 ++++++--- 3 files changed, 13 insertions(+), 6 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_engine_cs.c') diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index 5f99d4d6291b..68a58cce6ab1 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -7041,6 +7041,7 @@ enum { */ #define L3_GENERAL_PRIO_CREDITS(x) (((x) >> 1) << 19) #define L3_HIGH_PRIO_CREDITS(x) (((x) >> 1) << 14) +#define L3_PRIO_CREDITS_MASK ((0x1f << 19) | (0x1f << 14)) #define GEN7_L3CNTLREG1 _MMIO(0xB01C) #define GEN7_WA_FOR_GEN7_L3_CONTROL 0x3C47FF8C diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c index 9a4c4abc7450..a47a9c6bea52 100644 --- a/drivers/gpu/drm/i915/intel_engine_cs.c +++ b/drivers/gpu/drm/i915/intel_engine_cs.c @@ -1253,9 +1253,12 @@ static int bxt_init_workarounds(struct intel_engine_cs *engine) } /* WaProgramL3SqcReg1DefaultForPerf:bxt */ - if (IS_BXT_REVID(dev_priv, BXT_REVID_B0, REVID_FOREVER)) - I915_WRITE(GEN8_L3SQCREG1, L3_GENERAL_PRIO_CREDITS(62) | - L3_HIGH_PRIO_CREDITS(2)); + if (IS_BXT_REVID(dev_priv, BXT_REVID_B0, REVID_FOREVER)) { + u32 val = I915_READ(GEN8_L3SQCREG1); + val &= ~L3_PRIO_CREDITS_MASK; + val |= L3_GENERAL_PRIO_CREDITS(62) | L3_HIGH_PRIO_CREDITS(2); + I915_WRITE(GEN8_L3SQCREG1, val); + } /* WaToEnableHwFixForPushConstHWBug:bxt */ if (IS_BXT_REVID(dev_priv, BXT_REVID_C0, REVID_FOREVER)) diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c index ac4cf98030f9..c42a65a93b3a 100644 --- a/drivers/gpu/drm/i915/intel_pm.c +++ b/drivers/gpu/drm/i915/intel_pm.c @@ -8484,14 +8484,17 @@ static void gen8_set_l3sqc_credits(struct drm_i915_private *dev_priv, int high_prio_credits) { u32 misccpctl; + u32 val; /* WaTempDisableDOPClkGating:bdw */ misccpctl = I915_READ(GEN7_MISCCPCTL); I915_WRITE(GEN7_MISCCPCTL, misccpctl & ~GEN7_DOP_CLOCK_GATE_ENABLE); - I915_WRITE(GEN8_L3SQCREG1, - L3_GENERAL_PRIO_CREDITS(general_prio_credits) | - L3_HIGH_PRIO_CREDITS(high_prio_credits)); + val = I915_READ(GEN8_L3SQCREG1); + val &= ~L3_PRIO_CREDITS_MASK; + val |= L3_GENERAL_PRIO_CREDITS(general_prio_credits); + val |= L3_HIGH_PRIO_CREDITS(high_prio_credits); + I915_WRITE(GEN8_L3SQCREG1, val); /* * Wait at least 100 clocks before re-enabling clock gating. -- cgit v1.2.3 From 5d266692372dfb412150006df420185b666658f9 Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Mon, 23 Oct 2017 22:32:36 +0100 Subject: drm/i915: Filter out spurious execlists context-switch interrupts Back in commit a4b2b01523a8 ("drm/i915: Don't mark an execlists context-switch when idle") we noticed the presence of late context-switch interrupts. We were able to filter those out by looking at whether the ELSP remained active, but in commit beecec901790 ("drm/i915/execlists: Preemption!") that became problematic as we now anticipate receiving a context-switch event for preemption while ELSP may be empty. To restore the spurious interrupt suppression, add a counter for the expected number of pending context-switches and skip if we do not need to handle this interrupt to make forward progress. v2: Don't forget to switch on for preempt. v3: Reduce the counter to a on/off boolean tracker. Declare the HW as active when we first submit, and idle after the final completion event (with which we confirm the HW says it is idle), and track each source of activity separately. With a finite number of sources, it should aide us in debugging which gets stuck. Fixes: beecec901790 ("drm/i915/execlists: Preemption!") Signed-off-by: Chris Wilson Cc: Michal Winiarski Cc: Tvrtko Ursulin Cc: Arkadiusz Hiler Cc: Mika Kuoppala Cc: Joonas Lahtinen Link: https://patchwork.freedesktop.org/patch/msgid/20171023213237.26536-3-chris@chris-wilson.co.uk Reviewed-by: Mika Kuoppala (cherry picked from commit 4a118ecbe99c93cf9f9582e83a88d03f18d6cb84) Signed-off-by: Jani Nikula --- drivers/gpu/drm/i915/i915_guc_submission.c | 3 +++ drivers/gpu/drm/i915/i915_irq.c | 6 ++++-- drivers/gpu/drm/i915/intel_engine_cs.c | 5 +++-- drivers/gpu/drm/i915/intel_lrc.c | 27 ++++++++++++++++++------ drivers/gpu/drm/i915/intel_ringbuffer.h | 34 ++++++++++++++++++++++++++++-- 5 files changed, 62 insertions(+), 13 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_engine_cs.c') diff --git a/drivers/gpu/drm/i915/i915_guc_submission.c b/drivers/gpu/drm/i915/i915_guc_submission.c index a2e8114b739d..f84c267728fd 100644 --- a/drivers/gpu/drm/i915/i915_guc_submission.c +++ b/drivers/gpu/drm/i915/i915_guc_submission.c @@ -610,6 +610,7 @@ done: execlists->first = rb; if (submit) { port_assign(port, last); + execlists_set_active(execlists, EXECLISTS_ACTIVE_USER); i915_guc_submit(engine); } spin_unlock_irq(&engine->timeline->lock); @@ -633,6 +634,8 @@ static void i915_guc_irq_handler(unsigned long data) rq = port_request(&port[0]); } + if (!rq) + execlists_clear_active(execlists, EXECLISTS_ACTIVE_USER); if (!port_isset(last_port)) i915_guc_dequeue(engine); diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c index b1296a55c1e4..f8205841868b 100644 --- a/drivers/gpu/drm/i915/i915_irq.c +++ b/drivers/gpu/drm/i915/i915_irq.c @@ -1388,8 +1388,10 @@ gen8_cs_irq_handler(struct intel_engine_cs *engine, u32 iir, int test_shift) bool tasklet = false; if (iir & (GT_CONTEXT_SWITCH_INTERRUPT << test_shift)) { - __set_bit(ENGINE_IRQ_EXECLIST, &engine->irq_posted); - tasklet = true; + if (READ_ONCE(engine->execlists.active)) { + __set_bit(ENGINE_IRQ_EXECLIST, &engine->irq_posted); + tasklet = true; + } } if (iir & (GT_RENDER_USER_INTERRUPT << test_shift)) { diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c index a47a9c6bea52..ab5bf4e2e28e 100644 --- a/drivers/gpu/drm/i915/intel_engine_cs.c +++ b/drivers/gpu/drm/i915/intel_engine_cs.c @@ -1548,8 +1548,8 @@ bool intel_engine_is_idle(struct intel_engine_cs *engine) if (test_bit(ENGINE_IRQ_EXECLIST, &engine->irq_posted)) return false; - /* Both ports drained, no more ELSP submission? */ - if (port_request(&engine->execlists.port[0])) + /* Waiting to drain ELSP? */ + if (READ_ONCE(engine->execlists.active)) return false; /* ELSP is empty, but there are ready requests? */ @@ -1749,6 +1749,7 @@ void intel_engine_dump(struct intel_engine_cs *engine, struct drm_printer *m) idx); } } + drm_printf(m, "\t\tHW active? 0x%x\n", execlists->active); rcu_read_unlock(); } else if (INTEL_GEN(dev_priv) > 6) { drm_printf(m, "\tPP_DIR_BASE: 0x%08x\n", diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c index 7f45dd7dc3e5..5b87d5284a84 100644 --- a/drivers/gpu/drm/i915/intel_lrc.c +++ b/drivers/gpu/drm/i915/intel_lrc.c @@ -575,7 +575,8 @@ static void execlists_dequeue(struct intel_engine_cs *engine) * the state of the GPU is known (idle). */ inject_preempt_context(engine); - execlists->preempt = true; + execlists_set_active(execlists, + EXECLISTS_ACTIVE_PREEMPT); goto unlock; } else { /* @@ -683,8 +684,10 @@ done: unlock: spin_unlock_irq(&engine->timeline->lock); - if (submit) + if (submit) { + execlists_set_active(execlists, EXECLISTS_ACTIVE_USER); execlists_submit_ports(engine); + } } static void @@ -696,6 +699,7 @@ execlist_cancel_port_requests(struct intel_engine_execlists *execlists) while (num_ports-- && port_isset(port)) { struct drm_i915_gem_request *rq = port_request(port); + GEM_BUG_ON(!execlists->active); execlists_context_status_change(rq, INTEL_CONTEXT_SCHEDULE_PREEMPTED); i915_gem_request_put(rq); @@ -861,15 +865,21 @@ static void intel_lrc_irq_handler(unsigned long data) unwind_incomplete_requests(engine); spin_unlock_irq(&engine->timeline->lock); - GEM_BUG_ON(!execlists->preempt); - execlists->preempt = false; + GEM_BUG_ON(!execlists_is_active(execlists, + EXECLISTS_ACTIVE_PREEMPT)); + execlists_clear_active(execlists, + EXECLISTS_ACTIVE_PREEMPT); continue; } if (status & GEN8_CTX_STATUS_PREEMPTED && - execlists->preempt) + execlists_is_active(execlists, + EXECLISTS_ACTIVE_PREEMPT)) continue; + GEM_BUG_ON(!execlists_is_active(execlists, + EXECLISTS_ACTIVE_USER)); + /* Check the context/desc id for this event matches */ GEM_DEBUG_BUG_ON(buf[2 * head + 1] != port->context_id); @@ -892,6 +902,9 @@ static void intel_lrc_irq_handler(unsigned long data) /* After the final element, the hw should be idle */ GEM_BUG_ON(port_count(port) == 0 && !(status & GEN8_CTX_STATUS_ACTIVE_IDLE)); + if (port_count(port) == 0) + execlists_clear_active(execlists, + EXECLISTS_ACTIVE_USER); } if (head != execlists->csb_head) { @@ -901,7 +914,7 @@ static void intel_lrc_irq_handler(unsigned long data) } } - if (!execlists->preempt) + if (!execlists_is_active(execlists, EXECLISTS_ACTIVE_PREEMPT)) execlists_dequeue(engine); intel_uncore_forcewake_put(dev_priv, execlists->fw_domains); @@ -1460,7 +1473,7 @@ static int gen8_init_common_ring(struct intel_engine_cs *engine) GT_CONTEXT_SWITCH_INTERRUPT << engine->irq_shift); clear_bit(ENGINE_IRQ_EXECLIST, &engine->irq_posted); execlists->csb_head = -1; - execlists->preempt = false; + execlists->active = 0; /* After a GPU reset, we may have requests to replay */ if (!i915_modparams.enable_guc_submission && execlists->first) diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.h b/drivers/gpu/drm/i915/intel_ringbuffer.h index 17186f067408..6a42ed618a28 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.h +++ b/drivers/gpu/drm/i915/intel_ringbuffer.h @@ -241,9 +241,17 @@ struct intel_engine_execlists { } port[EXECLIST_MAX_PORTS]; /** - * @preempt: are we currently handling a preempting context switch? + * @active: is the HW active? We consider the HW as active after + * submitting any context for execution and until we have seen the + * last context completion event. After that, we do not expect any + * more events until we submit, and so can park the HW. + * + * As we have a small number of different sources from which we feed + * the HW, we track the state of each inside a single bitfield. */ - bool preempt; + unsigned int active; +#define EXECLISTS_ACTIVE_USER 0 +#define EXECLISTS_ACTIVE_PREEMPT 1 /** * @port_mask: number of execlist ports - 1 @@ -525,6 +533,27 @@ struct intel_engine_cs { u32 (*get_cmd_length_mask)(u32 cmd_header); }; +static inline void +execlists_set_active(struct intel_engine_execlists *execlists, + unsigned int bit) +{ + __set_bit(bit, (unsigned long *)&execlists->active); +} + +static inline void +execlists_clear_active(struct intel_engine_execlists *execlists, + unsigned int bit) +{ + __clear_bit(bit, (unsigned long *)&execlists->active); +} + +static inline bool +execlists_is_active(const struct intel_engine_execlists *execlists, + unsigned int bit) +{ + return test_bit(bit, (unsigned long *)&execlists->active); +} + static inline unsigned int execlists_num_ports(const struct intel_engine_execlists * const execlists) { @@ -538,6 +567,7 @@ execlists_port_complete(struct intel_engine_execlists * const execlists, const unsigned int m = execlists->port_mask; GEM_BUG_ON(port_index(port, execlists) != 0); + GEM_BUG_ON(!execlists_is_active(execlists, EXECLISTS_ACTIVE_USER)); memmove(port, port + 1, m * sizeof(struct execlist_port)); memset(port + m, 0, sizeof(struct execlist_port)); -- cgit v1.2.3