From 7102404cb471973cdcc2d3bfe5755153b3137bca Mon Sep 17 00:00:00 2001 From: Ville Syrjälä Date: Fri, 25 Sep 2020 15:17:48 +0300 Subject: drm/i915: Make intel_{enable,disable}_sagv() static MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit intel_{enable,disable}_sagv() are no longer needed outside the compilation unit. Make them static. Signed-off-by: Ville Syrjälä Link: https://patchwork.freedesktop.org/patch/msgid/20200925121749.708-1-ville.syrjala@linux.intel.com Reviewed-by: José Roberto de Souza --- drivers/gpu/drm/i915/intel_pm.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_pm.c') diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c index 34e0d22d456b..8cd62402d597 100644 --- a/drivers/gpu/drm/i915/intel_pm.c +++ b/drivers/gpu/drm/i915/intel_pm.c @@ -3706,7 +3706,7 @@ skl_setup_sagv_block_time(struct drm_i915_private *dev_priv) * - All planes can enable watermarks for latencies >= SAGV engine block time * - We're not using an interlaced display configuration */ -int +static int intel_enable_sagv(struct drm_i915_private *dev_priv) { int ret; @@ -3740,7 +3740,7 @@ intel_enable_sagv(struct drm_i915_private *dev_priv) return 0; } -int +static int intel_disable_sagv(struct drm_i915_private *dev_priv) { int ret; -- cgit From da942750928a037d6285b64826a637918f3fe3cc Mon Sep 17 00:00:00 2001 From: Stuart Summers Date: Wed, 14 Oct 2020 12:19:34 -0700 Subject: drm/i915/dg1: Add initial DG1 workarounds MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit DG1 shares some workarounds with TGL and RKL and also has some additional workarounds of its own. v2: Correct location of Wa_1408615072 (JohnH). v3: Apply WAs 1606700617, 18011464164 and 22010931296 to DG1 (José) v4 (Anusha) - Add Wa_22010271021 - s/Wa_14010096844/Wa_1409836686 v5: - Extend Wa_14010919138 to all revs (Matt Atwood) - Power gate media is global gen12 design. (Rodrigo) - Rebase (Lucas) v6: use REG_BIT() to fix checkpatch warning (Lucas) BSpec: 53508 Cc: Matt Atwood Cc: Matt Roper Cc: Radhakrishna Sripada Cc: José Roberto de Souza Signed-off-by: Stuart Summers Signed-off-by: Anusha Srivatsa Signed-off-by: Rodrigo Vivi Reviewed-by: Lucas De Marchi Signed-off-by: Lucas De Marchi Link: https://patchwork.freedesktop.org/patch/msgid/20201014191937.1266226-8-lucas.demarchi@intel.com --- drivers/gpu/drm/i915/display/intel_display_power.c | 5 +- drivers/gpu/drm/i915/display/intel_sprite.c | 4 +- drivers/gpu/drm/i915/gt/intel_workarounds.c | 111 +++++++++++++++++---- drivers/gpu/drm/i915/i915_pci.c | 2 + drivers/gpu/drm/i915/i915_reg.h | 14 ++- drivers/gpu/drm/i915/intel_pm.c | 39 +++++--- 6 files changed, 131 insertions(+), 44 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_pm.c') diff --git a/drivers/gpu/drm/i915/display/intel_display_power.c b/drivers/gpu/drm/i915/display/intel_display_power.c index 4934c89882b1..18af078c208b 100644 --- a/drivers/gpu/drm/i915/display/intel_display_power.c +++ b/drivers/gpu/drm/i915/display/intel_display_power.c @@ -5273,8 +5273,9 @@ static void tgl_bw_buddy_init(struct drm_i915_private *dev_priv) unsigned long abox_mask = INTEL_INFO(dev_priv)->abox_mask; int config, i; - if (IS_TGL_DISP_REVID(dev_priv, TGL_REVID_A0, TGL_REVID_B0)) - /* Wa_1409767108: tgl */ + if (IS_DG1_REVID(dev_priv, DG1_REVID_A0, DG1_REVID_A0) || + IS_TGL_DISP_REVID(dev_priv, TGL_REVID_A0, TGL_REVID_B0)) + /* Wa_1409767108:tgl,dg1 */ table = wa_1409767108_buddy_page_masks; else table = tgl_buddy_page_masks; diff --git a/drivers/gpu/drm/i915/display/intel_sprite.c b/drivers/gpu/drm/i915/display/intel_sprite.c index 3ae7470c1b8b..88bfebdf9228 100644 --- a/drivers/gpu/drm/i915/display/intel_sprite.c +++ b/drivers/gpu/drm/i915/display/intel_sprite.c @@ -2886,8 +2886,8 @@ static bool skl_plane_format_mod_supported(struct drm_plane *_plane, static bool gen12_plane_supports_mc_ccs(struct drm_i915_private *dev_priv, enum plane_id plane_id) { - /* Wa_14010477008:tgl[a0..c0],rkl[all] */ - if (IS_ROCKETLAKE(dev_priv) || + /* Wa_14010477008:tgl[a0..c0],rkl[all],dg1[all] */ + if (IS_DG1(dev_priv) || IS_ROCKETLAKE(dev_priv) || IS_TGL_DISP_REVID(dev_priv, TGL_REVID_A0, TGL_REVID_C0)) return false; diff --git a/drivers/gpu/drm/i915/gt/intel_workarounds.c b/drivers/gpu/drm/i915/gt/intel_workarounds.c index c6433b72f5e9..fed9503a7c4e 100644 --- a/drivers/gpu/drm/i915/gt/intel_workarounds.c +++ b/drivers/gpu/drm/i915/gt/intel_workarounds.c @@ -672,6 +672,20 @@ static void tgl_ctx_workarounds_init(struct intel_engine_cs *engine, 0); } +static void dg1_ctx_workarounds_init(struct intel_engine_cs *engine, + struct i915_wa_list *wal) +{ + gen12_ctx_workarounds_init(engine, wal); + + /* Wa_1409044764 */ + WA_CLR_BIT_MASKED(GEN11_COMMON_SLICE_CHICKEN3, + DG1_FLOAT_POINT_BLEND_OPT_STRICT_MODE_EN); + + /* Wa_22010493298 */ + WA_SET_BIT_MASKED(HIZ_CHICKEN, + DG1_HZ_READ_SUPPRESSION_OPTIMIZATION_DISABLE); +} + static void __intel_engine_init_ctx_wa(struct intel_engine_cs *engine, struct i915_wa_list *wal, @@ -684,7 +698,9 @@ __intel_engine_init_ctx_wa(struct intel_engine_cs *engine, wa_init_start(wal, name, engine->name); - if (IS_ROCKETLAKE(i915) || IS_TIGERLAKE(i915)) + if (IS_DG1(i915)) + dg1_ctx_workarounds_init(engine, wal); + else if (IS_ROCKETLAKE(i915) || IS_TIGERLAKE(i915)) tgl_ctx_workarounds_init(engine, wal); else if (IS_GEN(i915, 12)) gen12_ctx_workarounds_init(engine, wal); @@ -1244,10 +1260,36 @@ tgl_gt_workarounds_init(struct drm_i915_private *i915, struct i915_wa_list *wal) L3_CLKGATE_DIS | L3_CR2X_CLKGATE_DIS); } +static void +dg1_gt_workarounds_init(struct drm_i915_private *i915, struct i915_wa_list *wal) +{ + gen12_gt_workarounds_init(i915, wal); + + /* Wa_1607087056:dg1 */ + if (IS_DG1_REVID(i915, DG1_REVID_A0, DG1_REVID_A0)) + wa_write_or(wal, + SLICE_UNIT_LEVEL_CLKGATE, + L3_CLKGATE_DIS | L3_CR2X_CLKGATE_DIS); + + /* Wa_1409420604:dg1 */ + if (IS_DG1(i915)) + wa_write_or(wal, + SUBSLICE_UNIT_LEVEL_CLKGATE2, + CPSSUNIT_CLKGATE_DIS); + + /* Wa_1408615072:dg1 */ + /* Empirical testing shows this register is unaffected by engine reset. */ + if (IS_DG1(i915)) + wa_write_or(wal, UNSLICE_UNIT_LEVEL_CLKGATE2, + VSUNIT_CLKGATE_DIS_TGL); +} + static void gt_init_workarounds(struct drm_i915_private *i915, struct i915_wa_list *wal) { - if (IS_TIGERLAKE(i915)) + if (IS_DG1(i915)) + dg1_gt_workarounds_init(i915, wal); + else if (IS_TIGERLAKE(i915)) tgl_gt_workarounds_init(i915, wal); else if (IS_GEN(i915, 12)) gen12_gt_workarounds_init(i915, wal); @@ -1612,6 +1654,20 @@ static void tgl_whitelist_build(struct intel_engine_cs *engine) } } +static void dg1_whitelist_build(struct intel_engine_cs *engine) +{ + struct i915_wa_list *w = &engine->whitelist; + + tgl_whitelist_build(engine); + + /* GEN:BUG:1409280441:dg1 */ + if (IS_DG1_REVID(engine->i915, DG1_REVID_A0, DG1_REVID_A0) && + (engine->class == RENDER_CLASS || + engine->class == COPY_ENGINE_CLASS)) + whitelist_reg_ext(w, RING_ID(engine->mmio_base), + RING_FORCE_TO_NONPRIV_ACCESS_RD); +} + void intel_engine_init_whitelist(struct intel_engine_cs *engine) { struct drm_i915_private *i915 = engine->i915; @@ -1619,7 +1675,9 @@ void intel_engine_init_whitelist(struct intel_engine_cs *engine) wa_init_start(w, "whitelist", engine->name); - if (IS_GEN(i915, 12)) + if (IS_DG1(i915)) + dg1_whitelist_build(engine); + else if (IS_GEN(i915, 12)) tgl_whitelist_build(engine); else if (IS_GEN(i915, 11)) icl_whitelist_build(engine); @@ -1673,15 +1731,18 @@ rcs_engine_wa_init(struct intel_engine_cs *engine, struct i915_wa_list *wal) { struct drm_i915_private *i915 = engine->i915; - if (IS_TGL_UY_GT_REVID(i915, TGL_REVID_A0, TGL_REVID_A0)) { + if (IS_DG1_REVID(i915, DG1_REVID_A0, DG1_REVID_A0) || + IS_TGL_UY_GT_REVID(i915, TGL_REVID_A0, TGL_REVID_A0)) { /* - * Wa_1607138336:tgl - * Wa_1607063988:tgl + * Wa_1607138336:tgl[a0],dg1[a0] + * Wa_1607063988:tgl[a0],dg1[a0] */ wa_write_or(wal, GEN9_CTX_PREEMPT_REG, GEN12_DISABLE_POSH_BUSY_FF_DOP_CG); + } + if (IS_TGL_UY_GT_REVID(i915, TGL_REVID_A0, TGL_REVID_A0)) { /* * Wa_1606679103:tgl * (see also Wa_1606682166:icl) @@ -1695,35 +1756,41 @@ rcs_engine_wa_init(struct intel_engine_cs *engine, struct i915_wa_list *wal) VSUNIT_CLKGATE_DIS_TGL); } - if (IS_ROCKETLAKE(i915) || IS_TIGERLAKE(i915)) { - /* Wa_1606931601:tgl,rkl */ + if (IS_DG1(i915) || IS_ROCKETLAKE(i915) || IS_TIGERLAKE(i915)) { + /* Wa_1606931601:tgl,rkl,dg1 */ wa_masked_en(wal, GEN7_ROW_CHICKEN2, GEN12_DISABLE_EARLY_READ); - /* Wa_1409804808:tgl,rkl */ + /* + * Wa_1407928979:tgl A* + * Wa_18011464164:tgl[B0+],dg1[B0+] + * Wa_22010931296:tgl[B0+],dg1[B0+] + * Wa_14010919138:rkl, dg1 + */ + wa_write_or(wal, GEN7_FF_THREAD_MODE, + GEN12_FF_TESSELATION_DOP_GATE_DISABLE); + } + + if (IS_DG1_REVID(i915, DG1_REVID_A0, DG1_REVID_A0) || + IS_ROCKETLAKE(i915) || IS_TIGERLAKE(i915)) { + /* Wa_1409804808:tgl,rkl,dg1[a0] */ wa_masked_en(wal, GEN7_ROW_CHICKEN2, GEN12_PUSH_CONST_DEREF_HOLD_DIS); /* * Wa_1409085225:tgl - * Wa_14010229206:tgl,rkl + * Wa_14010229206:tgl,rkl,dg1[a0] */ wa_masked_en(wal, GEN9_ROW_CHICKEN4, GEN12_DISABLE_TDL_PUSH); - /* - * Wa_1407928979:tgl A* - * Wa_18011464164:tgl B0+ - * Wa_22010931296:tgl B0+ - * Wa_14010919138:rkl,tgl - */ - wa_write_or(wal, GEN7_FF_THREAD_MODE, - GEN12_FF_TESSELATION_DOP_GATE_DISABLE); - /* * Wa_1607030317:tgl * Wa_1607186500:tgl - * Wa_1607297627:tgl,rkl there are multiple entries for this - * WA in the BSpec; some indicate this is an A0-only WA, - * others indicate it applies to all steppings. + * Wa_1607297627:tgl,rkl,dg1[a0] + * + * On TGL and RKL there are multiple entries for this WA in the + * BSpec; some indicate this is an A0-only WA, others indicate + * it applies to all steppings so we trust the "all steppings." + * For DG1 this only applies to A0. */ wa_masked_en(wal, GEN6_RC_SLEEP_PSMI_CONTROL, diff --git a/drivers/gpu/drm/i915/i915_pci.c b/drivers/gpu/drm/i915/i915_pci.c index 16d4e72bed09..d0eeb2181e74 100644 --- a/drivers/gpu/drm/i915/i915_pci.c +++ b/drivers/gpu/drm/i915/i915_pci.c @@ -918,6 +918,8 @@ static const struct intel_device_info dg1_info __maybe_unused = { .platform_engine_mask = BIT(RCS0) | BIT(BCS0) | BIT(VECS0) | BIT(VCS0) | BIT(VCS2), + /* Wa_16011227922 */ + .ppgtt_size = 47, }; #undef GEN diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index 49945e33f573..d33d0057aef4 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -2528,6 +2528,7 @@ static inline bool i915_mmio_reg_valid(i915_reg_t reg) #define RING_PSMI_CTL(base) _MMIO((base) + 0x50) #define RING_MAX_IDLE(base) _MMIO((base) + 0x54) #define RING_HWS_PGA(base) _MMIO((base) + 0x80) +#define RING_ID(base) _MMIO((base) + 0x8c) #define RING_HWS_PGA_GEN6(base) _MMIO((base) + 0x2080) #define RING_RESET_CTL(base) _MMIO((base) + 0xd0) #define RESET_CTL_CAT_ERROR REG_BIT(2) @@ -4147,6 +4148,7 @@ enum { #define GEN9_CLKGATE_DIS_3 _MMIO(0x46538) #define TGL_VRH_GATING_DIS REG_BIT(31) +#define DPT_GATING_DIS REG_BIT(22) #define GEN9_CLKGATE_DIS_4 _MMIO(0x4653C) #define BXT_GMBUS_GATING_DIS (1 << 14) @@ -8019,13 +8021,15 @@ enum { #define GEN8_L3CNTLREG _MMIO(0x7034) #define GEN8_ERRDETBCTRL (1 << 9) -#define GEN11_COMMON_SLICE_CHICKEN3 _MMIO(0x7304) - #define GEN11_BLEND_EMB_FIX_DISABLE_IN_RCC (1 << 11) - #define GEN12_DISABLE_CPS_AWARE_COLOR_PIPE (1 << 9) +#define GEN11_COMMON_SLICE_CHICKEN3 _MMIO(0x7304) + #define DG1_FLOAT_POINT_BLEND_OPT_STRICT_MODE_EN REG_BIT(12) + #define GEN11_BLEND_EMB_FIX_DISABLE_IN_RCC REG_BIT(11) + #define GEN12_DISABLE_CPS_AWARE_COLOR_PIPE REG_BIT(9) #define HIZ_CHICKEN _MMIO(0x7018) -# define CHV_HZ_8X8_MODE_IN_1X (1 << 15) -# define BDW_HIZ_POWER_COMPILER_CLOCK_GATING_DISABLE (1 << 3) +# define CHV_HZ_8X8_MODE_IN_1X REG_BIT(15) +# define DG1_HZ_READ_SUPPRESSION_OPTIMIZATION_DISABLE REG_BIT(14) +# define BDW_HIZ_POWER_COMPILER_CLOCK_GATING_DISABLE REG_BIT(3) #define GEN9_SLICE_COMMON_ECO_CHICKEN0 _MMIO(0x7308) #define DISABLE_PIXEL_MASK_CAMMING (1 << 14) diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c index 8cd62402d597..ae6b367e63cb 100644 --- a/drivers/gpu/drm/i915/intel_pm.c +++ b/drivers/gpu/drm/i915/intel_pm.c @@ -7116,25 +7116,26 @@ static void icl_init_clock_gating(struct drm_i915_private *dev_priv) 0, CNL_DELAY_PMRSP); } -static void tgl_init_clock_gating(struct drm_i915_private *dev_priv) +static void gen12_init_clock_gating(struct drm_i915_private *i915) { - u32 vd_pg_enable = 0; unsigned int i; + /* This is not a WA. Enable VD HCP & MFX_ENC powergate */ + for (i = 0; i < I915_MAX_VCS; i++) + if (HAS_ENGINE(&i915->gt, _VCS(i))) + intel_uncore_rmw(&i915->uncore, POWERGATE_ENABLE, 0, + VDN_HCP_POWERGATE_ENABLE(i) | + VDN_MFX_POWERGATE_ENABLE(i)); +} + +static void tgl_init_clock_gating(struct drm_i915_private *dev_priv) +{ + gen12_init_clock_gating(dev_priv); + /* Wa_1409120013:tgl */ I915_WRITE(ILK_DPFC_CHICKEN, ILK_DPFC_CHICKEN_COMP_DUMMY_PIXEL); - /* This is not a WA. Enable VD HCP & MFX_ENC powergate */ - for (i = 0; i < I915_MAX_VCS; i++) { - if (HAS_ENGINE(&dev_priv->gt, _VCS(i))) - vd_pg_enable |= VDN_HCP_POWERGATE_ENABLE(i) | - VDN_MFX_POWERGATE_ENABLE(i); - } - - I915_WRITE(POWERGATE_ENABLE, - I915_READ(POWERGATE_ENABLE) | vd_pg_enable); - /* Wa_1409825376:tgl (pre-prod)*/ if (IS_TGL_DISP_REVID(dev_priv, TGL_REVID_A0, TGL_REVID_B1)) I915_WRITE(GEN9_CLKGATE_DIS_3, I915_READ(GEN9_CLKGATE_DIS_3) | @@ -7145,6 +7146,16 @@ static void tgl_init_clock_gating(struct drm_i915_private *dev_priv) 0, DFR_DISABLE); } +static void dg1_init_clock_gating(struct drm_i915_private *dev_priv) +{ + gen12_init_clock_gating(dev_priv); + + /* Wa_1409836686:dg1[a0] */ + if (IS_DG1_REVID(dev_priv, DG1_REVID_A0, DG1_REVID_A0)) + I915_WRITE(GEN9_CLKGATE_DIS_3, I915_READ(GEN9_CLKGATE_DIS_3) | + DPT_GATING_DIS); +} + static void cnp_init_clock_gating(struct drm_i915_private *dev_priv) { if (!HAS_PCH_CNP(dev_priv)) @@ -7590,7 +7601,9 @@ 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_GEN(dev_priv, 12)) + if (IS_DG1(dev_priv)) + dev_priv->display.init_clock_gating = dg1_init_clock_gating; + else if (IS_GEN(dev_priv, 12)) dev_priv->display.init_clock_gating = tgl_init_clock_gating; else if (IS_GEN(dev_priv, 11)) dev_priv->display.init_clock_gating = icl_init_clock_gating; -- cgit From 4d6bde58a026c5cac263081c572aebe89bf32167 Mon Sep 17 00:00:00 2001 From: Ville Syrjälä Date: Thu, 16 Jul 2020 22:04:26 +0300 Subject: drm/i915: Apply WAC6entrylatency to kbl/cfl MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit WAC6entrylatency is trying to fix excessive rc6 entry latency caused by the extra delay from FBC_LLC_READ_CTRL, which is there for some extra sync with uncore for frame buffer caching in LLC. Reading through the hsd the recommendation was to set the FBC_LLC_FULLY_OPEN bit to disable this extra delay entirely. This can be done whenever fb LLC caching is not used. The alternative suggestion was to reduce the delay to eg. 0x5 via updated BIOS programming instructions. But all the kbl/cfl machines I've seen still have the default 0xff programmed. As we never use fb LLC caching let's just apply the w/a to all skl derivatives to get consistent rc6 latencies. I was able to measure the effect of FBC_LLC_READ_CTRL to rc6 latency via forcewake. Here's a graph of some of the results: sleep;fw_req=1;wait fw_ack==1;sleep;fw_req=0;wait fw_ack==0 fw_ack==1 duration 160us +----------------------------------------------------------------+ | + + $$+ + + | | $$ $ $ ******$$ ** $ $**$* #########$$######| 140us |-$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$*$$$$$$$$$$$$$$$$ $$$$$$| | $ * # | | $ * # | 120us |$+ * # +-| |$ * # | |$ * # # | 100us |$+ ************######################## +-| |$ * *# | |$ ***** ######### | 80us |$+ * # #### ## +-| |$ **** ### # # | | ** #### FBC_LLC_READ_CTRL: 0x8000 ******* | 60us |-###### FBC_LLC_READ_CTRL: 0xffff #######-| |## + + FBC_LLC_READ_CTRL: 0x400000ff $$$$$$$ | +----------------------------------------------------------------+ 0ms 10ms 20ms 30ms 40ms 50ms 60ms sleep duration The default FBC_LLC_READ_CTRL value of 0xff is documented to give us a 170usec delay. That tracks well with the knees at 0xffff->~44msec and 0x8000->~22msec we see in the graph. We can see that if we sleep longer than the FBC_LLC_READ_CTRL delay we always observe the full (~145usec) rc6 wakeup latency. But if we sleep for less than the FBC_LLC_READ_CTRL delay we see a quicker fw wakeup, presumably due the hardware not having yet entered rc6 fully. The other plateaus in the graph I suspect correspond to some shallower internal rc states. v2: s/usec/msec/ typo in commit msg Signed-off-by: Ville Syrjälä Link: https://patchwork.freedesktop.org/patch/msgid/20200716190426.17047-2-ville.syrjala@linux.intel.com Reviewed-by: Chris Wilson --- drivers/gpu/drm/i915/intel_pm.c | 8 ++++++++ 1 file changed, 8 insertions(+) (limited to 'drivers/gpu/drm/i915/intel_pm.c') diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c index ae6b367e63cb..0ef01a01ef8d 100644 --- a/drivers/gpu/drm/i915/intel_pm.c +++ b/drivers/gpu/drm/i915/intel_pm.c @@ -7208,6 +7208,10 @@ static void cfl_init_clock_gating(struct drm_i915_private *dev_priv) cnp_init_clock_gating(dev_priv); gen9_init_clock_gating(dev_priv); + /* WAC6entrylatency:cfl */ + I915_WRITE(FBC_LLC_READ_CTRL, I915_READ(FBC_LLC_READ_CTRL) | + FBC_LLC_FULLY_OPEN); + /* * WaFbcTurnOffFbcWatermark:cfl * Display WA #0562: cfl @@ -7227,6 +7231,10 @@ static void kbl_init_clock_gating(struct drm_i915_private *dev_priv) { gen9_init_clock_gating(dev_priv); + /* WAC6entrylatency:kbl */ + I915_WRITE(FBC_LLC_READ_CTRL, I915_READ(FBC_LLC_READ_CTRL) | + FBC_LLC_FULLY_OPEN); + /* WaDisableSDEUnitClockGating:kbl */ if (IS_KBL_GT_REVID(dev_priv, 0, KBL_REVID_B0)) I915_WRITE(GEN8_UCGCTL6, I915_READ(GEN8_UCGCTL6) | -- cgit From 96eaeb3dfa40576a7aa195303c538202311cffbc Mon Sep 17 00:00:00 2001 From: Ville Syrjälä Date: Wed, 12 Dec 2018 23:17:38 +0200 Subject: drm/i915: Use _MMIO_PIPE3() for ilk+ WM0_PIPE registers MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Remove the hand rolled array of WM0_PIPE register offsets and use the standard _MMIO_PIPE3() instead. v2: Take care of gvt too Signed-off-by: Ville Syrjälä Link: https://patchwork.freedesktop.org/patch/msgid/20181212211738.27770-1-ville.syrjala@linux.intel.com Reviewed-by: Lucas De Marchi --- drivers/gpu/drm/i915/gvt/handlers.c | 6 +++--- drivers/gpu/drm/i915/i915_reg.h | 9 +++++---- drivers/gpu/drm/i915/intel_pm.c | 13 ++++--------- 3 files changed, 12 insertions(+), 16 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_pm.c') diff --git a/drivers/gpu/drm/i915/gvt/handlers.c b/drivers/gpu/drm/i915/gvt/handlers.c index e0edc9d1f357..c7cf15fe9ef6 100644 --- a/drivers/gpu/drm/i915/gvt/handlers.c +++ b/drivers/gpu/drm/i915/gvt/handlers.c @@ -2209,9 +2209,9 @@ static int init_generic_mmio_info(struct intel_gvt *gvt) MMIO_D(PF_VSCALE(PIPE_C), D_ALL); MMIO_D(PF_HSCALE(PIPE_C), D_ALL); - MMIO_D(WM0_PIPEA_ILK, D_ALL); - MMIO_D(WM0_PIPEB_ILK, D_ALL); - MMIO_D(WM0_PIPEC_IVB, D_ALL); + MMIO_D(WM0_PIPE_ILK(PIPE_A), D_ALL); + MMIO_D(WM0_PIPE_ILK(PIPE_B), D_ALL); + MMIO_D(WM0_PIPE_ILK(PIPE_C), D_ALL); MMIO_D(WM1_LP_ILK, D_ALL); MMIO_D(WM2_LP_ILK, D_ALL); MMIO_D(WM3_LP_ILK, D_ALL); diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index 8b021f77cb1f..993d0b4c6cf3 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -6434,15 +6434,16 @@ enum { _MMIO(_PLANE(plane, _PLANE_WM_TRANS_1(pipe), _PLANE_WM_TRANS_2(pipe))) /* define the Watermark register on Ironlake */ -#define WM0_PIPEA_ILK _MMIO(0x45100) +#define _WM0_PIPEA_ILK 0x45100 +#define _WM0_PIPEB_ILK 0x45104 +#define _WM0_PIPEC_IVB 0x45200 +#define WM0_PIPE_ILK(pipe) _MMIO_PIPE3((pipe), _WM0_PIPEA_ILK, \ + _WM0_PIPEB_ILK, _WM0_PIPEC_IVB) #define WM0_PIPE_PLANE_MASK (0xffff << 16) #define WM0_PIPE_PLANE_SHIFT 16 #define WM0_PIPE_SPRITE_MASK (0xff << 8) #define WM0_PIPE_SPRITE_SHIFT 8 #define WM0_PIPE_CURSOR_MASK (0xff) - -#define WM0_PIPEB_ILK _MMIO(0x45104) -#define WM0_PIPEC_IVB _MMIO(0x45200) #define WM1_LP_ILK _MMIO(0x45108) #define WM1_LP_SR_EN (1 << 31) #define WM1_LP_LATENCY_SHIFT 24 diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c index 0ef01a01ef8d..f54375b11964 100644 --- a/drivers/gpu/drm/i915/intel_pm.c +++ b/drivers/gpu/drm/i915/intel_pm.c @@ -3573,11 +3573,11 @@ static void ilk_write_wm_values(struct drm_i915_private *dev_priv, _ilk_disable_lp_wm(dev_priv, dirty); if (dirty & WM_DIRTY_PIPE(PIPE_A)) - I915_WRITE(WM0_PIPEA_ILK, results->wm_pipe[0]); + I915_WRITE(WM0_PIPE_ILK(PIPE_A), results->wm_pipe[0]); if (dirty & WM_DIRTY_PIPE(PIPE_B)) - I915_WRITE(WM0_PIPEB_ILK, results->wm_pipe[1]); + I915_WRITE(WM0_PIPE_ILK(PIPE_B), results->wm_pipe[1]); if (dirty & WM_DIRTY_PIPE(PIPE_C)) - I915_WRITE(WM0_PIPEC_IVB, results->wm_pipe[2]); + I915_WRITE(WM0_PIPE_ILK(PIPE_C), results->wm_pipe[2]); if (dirty & WM_DIRTY_DDB) { if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv)) { @@ -6287,13 +6287,8 @@ static void ilk_pipe_wm_get_hw_state(struct intel_crtc *crtc) struct intel_crtc_state *crtc_state = to_intel_crtc_state(crtc->base.state); struct intel_pipe_wm *active = &crtc_state->wm.ilk.optimal; enum pipe pipe = crtc->pipe; - static const i915_reg_t wm0_pipe_reg[] = { - [PIPE_A] = WM0_PIPEA_ILK, - [PIPE_B] = WM0_PIPEB_ILK, - [PIPE_C] = WM0_PIPEC_IVB, - }; - hw->wm_pipe[pipe] = I915_READ(wm0_pipe_reg[pipe]); + hw->wm_pipe[pipe] = I915_READ(WM0_PIPE_ILK(pipe)); memset(active, 0, sizeof(*active)); -- cgit From 695dc55b573985569259e18f8e6261a77924342b Mon Sep 17 00:00:00 2001 From: Rodrigo Vivi Date: Wed, 11 Nov 2020 09:09:36 -0500 Subject: drm/i915/tgl: Fix Media power gate sequence. Some media power gates are disabled by default. commit 5d86923060fc ("drm/i915/tgl: Enable VD HCP/MFX sub-pipe power gating") tried to enable it, but it duplicated an existent register. So, the main PG setup sequences ended up overwriting it. So, let's now merge this to the main PG setup sequence. v2: (Chris): s/BIT/REG_BIT, remove useless comment, remove useless =0, use the right gt, remove rc6 sequence doubt from commit message. Fixes: 5d86923060fc ("drm/i915/tgl: Enable VD HCP/MFX sub-pipe power gating") Cc: Lucas De Marchi Cc: stable@vger.kernel.org#v5.5+ Cc: Dale B Stimson Signed-off-by: Rodrigo Vivi Cc: Chris Wilson Reviewed-by: Chris Wilson Signed-off-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20201111072859.1186070-1-rodrigo.vivi@intel.com --- drivers/gpu/drm/i915/gt/intel_rc6.c | 22 +++++++++++++++++----- drivers/gpu/drm/i915/i915_reg.h | 12 +++++------- drivers/gpu/drm/i915/intel_pm.c | 13 ------------- 3 files changed, 22 insertions(+), 25 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_pm.c') diff --git a/drivers/gpu/drm/i915/gt/intel_rc6.c b/drivers/gpu/drm/i915/gt/intel_rc6.c index ab675d35030d..d7b8e4457fc2 100644 --- a/drivers/gpu/drm/i915/gt/intel_rc6.c +++ b/drivers/gpu/drm/i915/gt/intel_rc6.c @@ -56,9 +56,12 @@ static inline void set(struct intel_uncore *uncore, i915_reg_t reg, u32 val) static void gen11_rc6_enable(struct intel_rc6 *rc6) { - struct intel_uncore *uncore = rc6_to_uncore(rc6); + struct intel_gt *gt = rc6_to_gt(rc6); + struct intel_uncore *uncore = gt->uncore; struct intel_engine_cs *engine; enum intel_engine_id id; + u32 pg_enable; + int i; /* 2b: Program RC6 thresholds.*/ set(uncore, GEN6_RC6_WAKE_RATE_LIMIT, 54 << 16 | 85); @@ -102,10 +105,19 @@ static void gen11_rc6_enable(struct intel_rc6 *rc6) GEN6_RC_CTL_RC6_ENABLE | GEN6_RC_CTL_EI_MODE(1); - set(uncore, GEN9_PG_ENABLE, - GEN9_RENDER_PG_ENABLE | - GEN9_MEDIA_PG_ENABLE | - GEN11_MEDIA_SAMPLER_PG_ENABLE); + pg_enable = + GEN9_RENDER_PG_ENABLE | + GEN9_MEDIA_PG_ENABLE | + GEN11_MEDIA_SAMPLER_PG_ENABLE; + + if (INTEL_GEN(gt->i915) >= 12) { + for (i = 0; i < I915_MAX_VCS; i++) + if (HAS_ENGINE(gt, _VCS(i))) + pg_enable |= (VDN_HCP_POWERGATE_ENABLE(i) | + VDN_MFX_POWERGATE_ENABLE(i)); + } + + set(uncore, GEN9_PG_ENABLE, pg_enable); } static void gen9_rc6_enable(struct intel_rc6 *rc6) diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index ac691927e29d..6fb1746e2125 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -8984,10 +8984,6 @@ enum { #define GEN9_PWRGT_MEDIA_STATUS_MASK (1 << 0) #define GEN9_PWRGT_RENDER_STATUS_MASK (1 << 1) -#define POWERGATE_ENABLE _MMIO(0xa210) -#define VDN_HCP_POWERGATE_ENABLE(n) BIT(((n) * 2) + 3) -#define VDN_MFX_POWERGATE_ENABLE(n) BIT(((n) * 2) + 4) - #define GTFIFODBG _MMIO(0x120000) #define GT_FIFO_SBDEDICATE_FREE_ENTRY_CHV (0x1f << 20) #define GT_FIFO_FREE_ENTRIES_CHV (0x7f << 13) @@ -9127,9 +9123,11 @@ enum { #define GEN9_MEDIA_PG_IDLE_HYSTERESIS _MMIO(0xA0C4) #define GEN9_RENDER_PG_IDLE_HYSTERESIS _MMIO(0xA0C8) #define GEN9_PG_ENABLE _MMIO(0xA210) -#define GEN9_RENDER_PG_ENABLE REG_BIT(0) -#define GEN9_MEDIA_PG_ENABLE REG_BIT(1) -#define GEN11_MEDIA_SAMPLER_PG_ENABLE REG_BIT(2) +#define GEN9_RENDER_PG_ENABLE REG_BIT(0) +#define GEN9_MEDIA_PG_ENABLE REG_BIT(1) +#define GEN11_MEDIA_SAMPLER_PG_ENABLE REG_BIT(2) +#define VDN_HCP_POWERGATE_ENABLE(n) REG_BIT(3 + 2 * (n)) +#define VDN_MFX_POWERGATE_ENABLE(n) REG_BIT(4 + 2 * (n)) #define GEN8_PUSHBUS_CONTROL _MMIO(0xA248) #define GEN8_PUSHBUS_ENABLE _MMIO(0xA250) #define GEN8_PUSHBUS_SHIFT _MMIO(0xA25C) diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c index b4bd19266b8c..2000d3991d12 100644 --- a/drivers/gpu/drm/i915/intel_pm.c +++ b/drivers/gpu/drm/i915/intel_pm.c @@ -7118,23 +7118,10 @@ static void icl_init_clock_gating(struct drm_i915_private *dev_priv) static void tgl_init_clock_gating(struct drm_i915_private *dev_priv) { - u32 vd_pg_enable = 0; - unsigned int i; - /* Wa_1409120013:tgl */ I915_WRITE(ILK_DPFC_CHICKEN, ILK_DPFC_CHICKEN_COMP_DUMMY_PIXEL); - /* This is not a WA. Enable VD HCP & MFX_ENC powergate */ - for (i = 0; i < I915_MAX_VCS; i++) { - if (HAS_ENGINE(&dev_priv->gt, _VCS(i))) - vd_pg_enable |= VDN_HCP_POWERGATE_ENABLE(i) | - VDN_MFX_POWERGATE_ENABLE(i); - } - - I915_WRITE(POWERGATE_ENABLE, - I915_READ(POWERGATE_ENABLE) | vd_pg_enable); - /* Wa_1409825376:tgl (pre-prod)*/ if (IS_TGL_REVID(dev_priv, TGL_REVID_A0, TGL_REVID_A0)) I915_WRITE(GEN9_CLKGATE_DIS_3, I915_READ(GEN9_CLKGATE_DIS_3) | -- cgit From bafcdad643462471628275a120a9e2b52262559c Mon Sep 17 00:00:00 2001 From: Maarten Lankhorst Date: Thu, 12 Nov 2020 21:17:18 +0200 Subject: drm/i915: Add hw.pipe_mode to allow bigjoiner pipe/transcoder split MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit With bigjoiner, there will be 2 pipes driving 2 halves of 1 transcoder, because of this, we need a pipe_mode for various calculations, including for example watermarks, plane clipping, etc. v10: * remove redundant pipe_mode assignment (Ville) v9: * pipe_mode in state dump nd state check (Ville) v8: * Add pipe_mode in readout in verify_crtc_state (Ville) v7: * Remove redundant comment (Ville) * Just keep mode instead of pipe_mode (Ville) v6: * renaming in separate function, only pipe_mode here (Ville) * Add description (Maarten) v5: * Rebase (Manasi) v4: * Manual rebase (Manasi) v3: * Change state to crtc_state, fix rebase err (Manasi) v2: * Manual Rebase (Manasi) Signed-off-by: Maarten Lankhorst Signed-off-by: Manasi Navare Reviewed-by: Animesh Manna [vsyrjala: * Fix state checker * Fix state dump * Use pipe_mode for linetime watermarks * Make sure pipe_mode normal timings are correct since the silly ddb code uses them * Drop the redundant pipe_mode copies from intel_modeset_pipe_config() and intel_crtc_copy_uapi_to_hw_state() * Use drm_mode_copy() all over] Signed-off-by: Ville Syrjälä Link: https://patchwork.freedesktop.org/patch/msgid/20201112191718.16683-7-ville.syrjala@linux.intel.com --- drivers/gpu/drm/i915/display/intel_display.c | 67 ++++++++++++------- drivers/gpu/drm/i915/display/intel_display_types.h | 11 +++- drivers/gpu/drm/i915/intel_pm.c | 76 +++++++++++----------- 3 files changed, 92 insertions(+), 62 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_pm.c') diff --git a/drivers/gpu/drm/i915/display/intel_display.c b/drivers/gpu/drm/i915/display/intel_display.c index 07b8d945c41f..03577ee5d9b7 100644 --- a/drivers/gpu/drm/i915/display/intel_display.c +++ b/drivers/gpu/drm/i915/display/intel_display.c @@ -6073,18 +6073,16 @@ skl_update_scaler(struct intel_crtc_state *crtc_state, bool force_detach, static int skl_update_scaler_crtc(struct intel_crtc_state *crtc_state) { - const struct drm_display_mode *adjusted_mode = - &crtc_state->hw.adjusted_mode; + const struct drm_display_mode *pipe_mode = &crtc_state->hw.pipe_mode; int width, height; if (crtc_state->pch_pfit.enabled) { width = drm_rect_width(&crtc_state->pch_pfit.dst); height = drm_rect_height(&crtc_state->pch_pfit.dst); } else { - width = adjusted_mode->crtc_hdisplay; - height = adjusted_mode->crtc_vdisplay; + width = pipe_mode->crtc_hdisplay; + height = pipe_mode->crtc_vdisplay; } - return skl_update_scaler(crtc_state, !crtc_state->hw.active, SKL_CRTC_INDEX, &crtc_state->scaler_state.scaler_id, @@ -8098,7 +8096,7 @@ static bool intel_crtc_supports_double_wide(const struct intel_crtc *crtc) static u32 ilk_pipe_pixel_rate(const struct intel_crtc_state *crtc_state) { - u32 pixel_rate = crtc_state->hw.adjusted_mode.crtc_clock; + u32 pixel_rate = crtc_state->hw.pipe_mode.crtc_clock; unsigned int pipe_w, pipe_h, pfit_w, pfit_h; /* @@ -8156,7 +8154,7 @@ static void intel_crtc_compute_pixel_rate(struct intel_crtc_state *crtc_state) if (HAS_GMCH(dev_priv)) /* FIXME calculate proper pipe pixel rate for GMCH pfit */ crtc_state->pixel_rate = - crtc_state->hw.adjusted_mode.crtc_clock; + crtc_state->hw.pipe_mode.crtc_clock; else crtc_state->pixel_rate = ilk_pipe_pixel_rate(crtc_state); @@ -8165,8 +8163,12 @@ static void intel_crtc_compute_pixel_rate(struct intel_crtc_state *crtc_state) static void intel_crtc_readout_derived_state(struct intel_crtc_state *crtc_state) { struct drm_display_mode *mode = &crtc_state->hw.mode; + struct drm_display_mode *pipe_mode = &crtc_state->hw.pipe_mode; struct drm_display_mode *adjusted_mode = &crtc_state->hw.adjusted_mode; + drm_mode_copy(pipe_mode, adjusted_mode); + + intel_mode_from_crtc_timings(pipe_mode, pipe_mode); intel_mode_from_crtc_timings(adjusted_mode, adjusted_mode); intel_crtc_compute_pixel_rate(crtc_state); @@ -8188,9 +8190,12 @@ static int intel_crtc_compute_config(struct intel_crtc *crtc, struct intel_crtc_state *pipe_config) { struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); - const struct drm_display_mode *adjusted_mode = &pipe_config->hw.adjusted_mode; + struct drm_display_mode *pipe_mode = &pipe_config->hw.pipe_mode; int clock_limit = dev_priv->max_dotclk_freq; + drm_mode_copy(pipe_mode, &pipe_config->hw.adjusted_mode); + intel_mode_from_crtc_timings(pipe_mode, pipe_mode); + if (INTEL_GEN(dev_priv) < 4) { clock_limit = dev_priv->max_cdclk_freq * 9 / 10; @@ -8199,16 +8204,16 @@ static int intel_crtc_compute_config(struct intel_crtc *crtc, * is > 90% of the (display) core speed. */ if (intel_crtc_supports_double_wide(crtc) && - adjusted_mode->crtc_clock > clock_limit) { + pipe_mode->crtc_clock > clock_limit) { clock_limit = dev_priv->max_dotclk_freq; pipe_config->double_wide = true; } } - if (adjusted_mode->crtc_clock > clock_limit) { + if (pipe_mode->crtc_clock > clock_limit) { drm_dbg_kms(&dev_priv->drm, "requested pixel clock (%d kHz) too high (max: %d kHz, double wide: %s)\n", - adjusted_mode->crtc_clock, clock_limit, + pipe_mode->crtc_clock, clock_limit, yesno(pipe_config->double_wide)); return -EINVAL; } @@ -8251,7 +8256,7 @@ static int intel_crtc_compute_config(struct intel_crtc *crtc, * WaPruneModeWithIncorrectHsyncOffset:ctg,elk,ilk,snb,ivb,vlv,hsw. */ if ((INTEL_GEN(dev_priv) > 4 || IS_G4X(dev_priv)) && - adjusted_mode->crtc_hsync_start == adjusted_mode->crtc_hdisplay) + pipe_mode->crtc_hsync_start == pipe_mode->crtc_hdisplay) return -EINVAL; intel_crtc_compute_pixel_rate(pipe_config); @@ -12783,15 +12788,15 @@ static bool c8_planes_changed(const struct intel_crtc_state *new_crtc_state) static u16 hsw_linetime_wm(const struct intel_crtc_state *crtc_state) { - const struct drm_display_mode *adjusted_mode = - &crtc_state->hw.adjusted_mode; + const struct drm_display_mode *pipe_mode = + &crtc_state->hw.pipe_mode; int linetime_wm; if (!crtc_state->hw.enable) return 0; - linetime_wm = DIV_ROUND_CLOSEST(adjusted_mode->crtc_htotal * 1000 * 8, - adjusted_mode->crtc_clock); + linetime_wm = DIV_ROUND_CLOSEST(pipe_mode->crtc_htotal * 1000 * 8, + pipe_mode->crtc_clock); return min(linetime_wm, 0x1ff); } @@ -12799,14 +12804,14 @@ static u16 hsw_linetime_wm(const struct intel_crtc_state *crtc_state) static u16 hsw_ips_linetime_wm(const struct intel_crtc_state *crtc_state, const struct intel_cdclk_state *cdclk_state) { - const struct drm_display_mode *adjusted_mode = - &crtc_state->hw.adjusted_mode; + const struct drm_display_mode *pipe_mode = + &crtc_state->hw.pipe_mode; int linetime_wm; if (!crtc_state->hw.enable) return 0; - linetime_wm = DIV_ROUND_CLOSEST(adjusted_mode->crtc_htotal * 1000 * 8, + linetime_wm = DIV_ROUND_CLOSEST(pipe_mode->crtc_htotal * 1000 * 8, cdclk_state->logical.cdclk); return min(linetime_wm, 0x1ff); @@ -12816,14 +12821,14 @@ static u16 skl_linetime_wm(const struct intel_crtc_state *crtc_state) { struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc); struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); - const struct drm_display_mode *adjusted_mode = - &crtc_state->hw.adjusted_mode; + const struct drm_display_mode *pipe_mode = + &crtc_state->hw.pipe_mode; int linetime_wm; if (!crtc_state->hw.enable) return 0; - linetime_wm = DIV_ROUND_UP(adjusted_mode->crtc_htotal * 1000 * 8, + linetime_wm = DIV_ROUND_UP(pipe_mode->crtc_htotal * 1000 * 8, crtc_state->pixel_rate); /* Display WA #1135: BXT:ALL GLK:ALL */ @@ -13279,6 +13284,9 @@ static void intel_dump_pipe_config(const struct intel_crtc_state *pipe_config, drm_dbg_kms(&dev_priv->drm, "adjusted mode:\n"); drm_mode_debug_printmodeline(&pipe_config->hw.adjusted_mode); intel_dump_crtc_timings(dev_priv, &pipe_config->hw.adjusted_mode); + drm_dbg_kms(&dev_priv->drm, "pipe mode:\n"); + drm_mode_debug_printmodeline(&pipe_config->hw.pipe_mode); + intel_dump_crtc_timings(dev_priv, &pipe_config->hw.pipe_mode); drm_dbg_kms(&dev_priv->drm, "port clock: %d, pipe src size: %dx%d, pixel rate %d\n", pipe_config->port_clock, @@ -14027,6 +14035,20 @@ intel_pipe_config_compare(const struct intel_crtc_state *current_config, PIPE_CONF_CHECK_X(output_types); + PIPE_CONF_CHECK_I(hw.pipe_mode.crtc_hdisplay); + PIPE_CONF_CHECK_I(hw.pipe_mode.crtc_htotal); + PIPE_CONF_CHECK_I(hw.pipe_mode.crtc_hblank_start); + PIPE_CONF_CHECK_I(hw.pipe_mode.crtc_hblank_end); + PIPE_CONF_CHECK_I(hw.pipe_mode.crtc_hsync_start); + PIPE_CONF_CHECK_I(hw.pipe_mode.crtc_hsync_end); + + PIPE_CONF_CHECK_I(hw.pipe_mode.crtc_vdisplay); + PIPE_CONF_CHECK_I(hw.pipe_mode.crtc_vtotal); + PIPE_CONF_CHECK_I(hw.pipe_mode.crtc_vblank_start); + PIPE_CONF_CHECK_I(hw.pipe_mode.crtc_vblank_end); + PIPE_CONF_CHECK_I(hw.pipe_mode.crtc_vsync_start); + PIPE_CONF_CHECK_I(hw.pipe_mode.crtc_vsync_end); + PIPE_CONF_CHECK_I(hw.adjusted_mode.crtc_hdisplay); PIPE_CONF_CHECK_I(hw.adjusted_mode.crtc_htotal); PIPE_CONF_CHECK_I(hw.adjusted_mode.crtc_hblank_start); @@ -14153,6 +14175,7 @@ intel_pipe_config_compare(const struct intel_crtc_state *current_config, if (IS_G4X(dev_priv) || INTEL_GEN(dev_priv) >= 5) PIPE_CONF_CHECK_I(pipe_bpp); + PIPE_CONF_CHECK_CLOCK_FUZZY(hw.pipe_mode.crtc_clock); PIPE_CONF_CHECK_CLOCK_FUZZY(hw.adjusted_mode.crtc_clock); PIPE_CONF_CHECK_CLOCK_FUZZY(port_clock); diff --git a/drivers/gpu/drm/i915/display/intel_display_types.h b/drivers/gpu/drm/i915/display/intel_display_types.h index 6e72c9d52843..35ab5944a3f7 100644 --- a/drivers/gpu/drm/i915/display/intel_display_types.h +++ b/drivers/gpu/drm/i915/display/intel_display_types.h @@ -817,15 +817,22 @@ struct intel_crtc_state { * The following members are used to verify the hardware state: * - enable * - active - * - mode / adjusted_mode + * - mode / pipe_mode / adjusted_mode * - color property blobs. * * During initial hw readout, they need to be copied to uapi. + * + * Bigjoiner will allow a transcoder mode that spans 2 pipes; + * Use the pipe_mode for calculations like watermarks, pipe + * scaler, and bandwidth. + * + * Use adjusted_mode for things that need to know the full + * mode on the transcoder, which spans all pipes. */ struct { bool active, enable; struct drm_property_blob *degamma_lut, *gamma_lut, *ctm; - struct drm_display_mode mode, adjusted_mode; + struct drm_display_mode mode, pipe_mode, adjusted_mode; enum drm_scaling_filter scaling_filter; } hw; diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c index f54375b11964..9898c257d3e0 100644 --- a/drivers/gpu/drm/i915/intel_pm.c +++ b/drivers/gpu/drm/i915/intel_pm.c @@ -899,12 +899,12 @@ static void pnv_update_wm(struct intel_crtc *unused_crtc) crtc = single_enabled_crtc(dev_priv); if (crtc) { - const struct drm_display_mode *adjusted_mode = - &crtc->config->hw.adjusted_mode; + const struct drm_display_mode *pipe_mode = + &crtc->config->hw.pipe_mode; const struct drm_framebuffer *fb = crtc->base.primary->state->fb; int cpp = fb->format->cpp[0]; - int clock = adjusted_mode->crtc_clock; + int clock = pipe_mode->crtc_clock; /* Display SR */ wm = intel_calculate_wm(clock, &pnv_display_wm, @@ -1135,8 +1135,8 @@ static u16 g4x_compute_wm(const struct intel_crtc_state *crtc_state, { struct intel_plane *plane = to_intel_plane(plane_state->uapi.plane); struct drm_i915_private *dev_priv = to_i915(plane->base.dev); - const struct drm_display_mode *adjusted_mode = - &crtc_state->hw.adjusted_mode; + const struct drm_display_mode *pipe_mode = + &crtc_state->hw.pipe_mode; unsigned int latency = dev_priv->wm.pri_latency[level] * 10; unsigned int clock, htotal, cpp, width, wm; @@ -1163,8 +1163,8 @@ static u16 g4x_compute_wm(const struct intel_crtc_state *crtc_state, level != G4X_WM_LEVEL_NORMAL) cpp = max(cpp, 4u); - clock = adjusted_mode->crtc_clock; - htotal = adjusted_mode->crtc_htotal; + clock = pipe_mode->crtc_clock; + htotal = pipe_mode->crtc_htotal; width = drm_rect_width(&plane_state->uapi.dst); @@ -1660,8 +1660,8 @@ static u16 vlv_compute_wm_level(const struct intel_crtc_state *crtc_state, { struct intel_plane *plane = to_intel_plane(plane_state->uapi.plane); struct drm_i915_private *dev_priv = to_i915(plane->base.dev); - const struct drm_display_mode *adjusted_mode = - &crtc_state->hw.adjusted_mode; + const struct drm_display_mode *pipe_mode = + &crtc_state->hw.pipe_mode; unsigned int clock, htotal, cpp, width, wm; if (dev_priv->wm.pri_latency[level] == 0) @@ -1671,8 +1671,8 @@ static u16 vlv_compute_wm_level(const struct intel_crtc_state *crtc_state, return 0; cpp = plane_state->hw.fb->format->cpp[0]; - clock = adjusted_mode->crtc_clock; - htotal = adjusted_mode->crtc_htotal; + clock = pipe_mode->crtc_clock; + htotal = pipe_mode->crtc_htotal; width = crtc_state->pipe_src_w; if (plane->id == PLANE_CURSOR) { @@ -2261,12 +2261,12 @@ static void i965_update_wm(struct intel_crtc *unused_crtc) if (crtc) { /* self-refresh has much higher latency */ static const int sr_latency_ns = 12000; - const struct drm_display_mode *adjusted_mode = - &crtc->config->hw.adjusted_mode; + const struct drm_display_mode *pipe_mode = + &crtc->config->hw.pipe_mode; const struct drm_framebuffer *fb = crtc->base.primary->state->fb; - int clock = adjusted_mode->crtc_clock; - int htotal = adjusted_mode->crtc_htotal; + int clock = pipe_mode->crtc_clock; + int htotal = pipe_mode->crtc_htotal; int hdisplay = crtc->config->pipe_src_w; int cpp = fb->format->cpp[0]; int entries; @@ -2345,8 +2345,8 @@ static void i9xx_update_wm(struct intel_crtc *unused_crtc) fifo_size = dev_priv->display.get_fifo_size(dev_priv, PLANE_A); crtc = intel_get_crtc_for_plane(dev_priv, PLANE_A); if (intel_crtc_active(crtc)) { - const struct drm_display_mode *adjusted_mode = - &crtc->config->hw.adjusted_mode; + const struct drm_display_mode *pipe_mode = + &crtc->config->hw.pipe_mode; const struct drm_framebuffer *fb = crtc->base.primary->state->fb; int cpp; @@ -2356,7 +2356,7 @@ static void i9xx_update_wm(struct intel_crtc *unused_crtc) else cpp = fb->format->cpp[0]; - planea_wm = intel_calculate_wm(adjusted_mode->crtc_clock, + planea_wm = intel_calculate_wm(pipe_mode->crtc_clock, wm_info, fifo_size, cpp, pessimal_latency_ns); enabled = crtc; @@ -2372,8 +2372,8 @@ static void i9xx_update_wm(struct intel_crtc *unused_crtc) fifo_size = dev_priv->display.get_fifo_size(dev_priv, PLANE_B); crtc = intel_get_crtc_for_plane(dev_priv, PLANE_B); if (intel_crtc_active(crtc)) { - const struct drm_display_mode *adjusted_mode = - &crtc->config->hw.adjusted_mode; + const struct drm_display_mode *pipe_mode = + &crtc->config->hw.pipe_mode; const struct drm_framebuffer *fb = crtc->base.primary->state->fb; int cpp; @@ -2383,7 +2383,7 @@ static void i9xx_update_wm(struct intel_crtc *unused_crtc) else cpp = fb->format->cpp[0]; - planeb_wm = intel_calculate_wm(adjusted_mode->crtc_clock, + planeb_wm = intel_calculate_wm(pipe_mode->crtc_clock, wm_info, fifo_size, cpp, pessimal_latency_ns); if (enabled == NULL) @@ -2421,12 +2421,12 @@ static void i9xx_update_wm(struct intel_crtc *unused_crtc) if (HAS_FW_BLC(dev_priv) && enabled) { /* self-refresh has much higher latency */ static const int sr_latency_ns = 6000; - const struct drm_display_mode *adjusted_mode = - &enabled->config->hw.adjusted_mode; + const struct drm_display_mode *pipe_mode = + &enabled->config->hw.pipe_mode; const struct drm_framebuffer *fb = enabled->base.primary->state->fb; - int clock = adjusted_mode->crtc_clock; - int htotal = adjusted_mode->crtc_htotal; + int clock = pipe_mode->crtc_clock; + int htotal = pipe_mode->crtc_htotal; int hdisplay = enabled->config->pipe_src_w; int cpp; int entries; @@ -2474,7 +2474,7 @@ static void i845_update_wm(struct intel_crtc *unused_crtc) { struct drm_i915_private *dev_priv = to_i915(unused_crtc->base.dev); struct intel_crtc *crtc; - const struct drm_display_mode *adjusted_mode; + const struct drm_display_mode *pipe_mode; u32 fwater_lo; int planea_wm; @@ -2482,8 +2482,8 @@ static void i845_update_wm(struct intel_crtc *unused_crtc) if (crtc == NULL) return; - adjusted_mode = &crtc->config->hw.adjusted_mode; - planea_wm = intel_calculate_wm(adjusted_mode->crtc_clock, + pipe_mode = &crtc->config->hw.pipe_mode; + planea_wm = intel_calculate_wm(pipe_mode->crtc_clock, &i845_wm_info, dev_priv->display.get_fifo_size(dev_priv, PLANE_A), 4, pessimal_latency_ns); @@ -2573,7 +2573,7 @@ static u32 ilk_compute_pri_wm(const struct intel_crtc_state *crtc_state, return method1; method2 = ilk_wm_method2(crtc_state->pixel_rate, - crtc_state->hw.adjusted_mode.crtc_htotal, + crtc_state->hw.pipe_mode.crtc_htotal, drm_rect_width(&plane_state->uapi.dst), cpp, mem_value); @@ -2601,7 +2601,7 @@ static u32 ilk_compute_spr_wm(const struct intel_crtc_state *crtc_state, method1 = ilk_wm_method1(crtc_state->pixel_rate, cpp, mem_value); method2 = ilk_wm_method2(crtc_state->pixel_rate, - crtc_state->hw.adjusted_mode.crtc_htotal, + crtc_state->hw.pipe_mode.crtc_htotal, drm_rect_width(&plane_state->uapi.dst), cpp, mem_value); return min(method1, method2); @@ -2626,7 +2626,7 @@ static u32 ilk_compute_cur_wm(const struct intel_crtc_state *crtc_state, cpp = plane_state->hw.fb->format->cpp[0]; return ilk_wm_method2(crtc_state->pixel_rate, - crtc_state->hw.adjusted_mode.crtc_htotal, + crtc_state->hw.pipe_mode.crtc_htotal, drm_rect_width(&plane_state->uapi.dst), cpp, mem_value); } @@ -3883,7 +3883,7 @@ static bool skl_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state) if (!crtc_state->hw.active) return true; - if (crtc_state->hw.adjusted_mode.flags & DRM_MODE_FLAG_INTERLACE) + if (crtc_state->hw.pipe_mode.flags & DRM_MODE_FLAG_INTERLACE) return false; intel_atomic_crtc_state_for_each_plane_state(plane, plane_state, crtc_state) { @@ -4174,8 +4174,8 @@ skl_ddb_get_pipe_allocation_limits(struct drm_i915_private *dev_priv, */ total_slice_mask = dbuf_slice_mask; for_each_new_intel_crtc_in_state(intel_state, crtc, crtc_state, i) { - const struct drm_display_mode *adjusted_mode = - &crtc_state->hw.adjusted_mode; + const struct drm_display_mode *pipe_mode = + &crtc_state->hw.pipe_mode; enum pipe pipe = crtc->pipe; int hdisplay, vdisplay; u32 pipe_dbuf_slice_mask; @@ -4205,7 +4205,7 @@ skl_ddb_get_pipe_allocation_limits(struct drm_i915_private *dev_priv, if (dbuf_slice_mask != pipe_dbuf_slice_mask) continue; - drm_mode_get_hv_timing(adjusted_mode, &hdisplay, &vdisplay); + drm_mode_get_hv_timing(pipe_mode, &hdisplay, &vdisplay); total_width_in_range += hdisplay; @@ -5093,7 +5093,7 @@ intel_get_linetime_us(const struct intel_crtc_state *crtc_state) if (drm_WARN_ON(&dev_priv->drm, pixel_rate == 0)) return u32_to_fixed16(0); - crtc_htotal = crtc_state->hw.adjusted_mode.crtc_htotal; + crtc_htotal = crtc_state->hw.pipe_mode.crtc_htotal; linetime_us = div_fixed16(crtc_htotal * 1000, pixel_rate); return linetime_us; @@ -5282,14 +5282,14 @@ static void skl_compute_plane_wm(const struct intel_crtc_state *crtc_state, method1 = skl_wm_method1(dev_priv, wp->plane_pixel_rate, wp->cpp, latency, wp->dbuf_block_size); method2 = skl_wm_method2(wp->plane_pixel_rate, - crtc_state->hw.adjusted_mode.crtc_htotal, + crtc_state->hw.pipe_mode.crtc_htotal, latency, wp->plane_blocks_per_line); if (wp->y_tiled) { selected_result = max_fixed16(method2, wp->y_tile_minimum); } else { - if ((wp->cpp * crtc_state->hw.adjusted_mode.crtc_htotal / + if ((wp->cpp * crtc_state->hw.pipe_mode.crtc_htotal / wp->dbuf_block_size < 1) && (wp->plane_bytes_per_line / wp->dbuf_block_size < 1)) { selected_result = method2; -- cgit From ffc90033dfc1f5a857ac96ca648d1250b3fcccf9 Mon Sep 17 00:00:00 2001 From: Ville Syrjälä Date: Fri, 6 Nov 2020 19:30:37 +0200 Subject: drm/i915: Pass intel_atomic_state around MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Pass the whole intel_atomic_state to skl_build_pipe_wm() and skl_allocate_pipe_ddb() so we can start to iterate stuff containerd in the commit. Signed-off-by: Ville Syrjälä Reviewed-by: Manasi Navare Signed-off-by: Manasi Navare Link: https://patchwork.freedesktop.org/patch/msgid/20201106173042.7534-2-ville.syrjala@linux.intel.com --- drivers/gpu/drm/i915/intel_pm.c | 24 +++++++++++++----------- 1 file changed, 13 insertions(+), 11 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_pm.c') diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c index 9898c257d3e0..a413e4b10e45 100644 --- a/drivers/gpu/drm/i915/intel_pm.c +++ b/drivers/gpu/drm/i915/intel_pm.c @@ -4791,9 +4791,11 @@ skl_plane_wm_level(const struct intel_crtc_state *crtc_state, } static int -skl_allocate_pipe_ddb(struct intel_crtc_state *crtc_state) +skl_allocate_pipe_ddb(struct intel_atomic_state *state, + struct intel_crtc *crtc) { - struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc); + struct intel_crtc_state *crtc_state = + intel_atomic_get_new_crtc_state(state, crtc); struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); struct skl_ddb_entry *alloc = &crtc_state->wm.skl.ddb; u16 alloc_size, start = 0; @@ -5583,9 +5585,12 @@ static int icl_build_plane_wm(struct intel_crtc_state *crtc_state, return 0; } -static int skl_build_pipe_wm(struct intel_crtc_state *crtc_state) +static int skl_build_pipe_wm(struct intel_atomic_state *state, + struct intel_crtc *crtc) { - struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev); + struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); + struct intel_crtc_state *crtc_state = + intel_atomic_get_new_crtc_state(state, crtc); struct skl_pipe_wm *pipe_wm = &crtc_state->wm.skl.optimal; struct intel_plane *plane; const struct intel_plane_state *plane_state; @@ -5794,7 +5799,7 @@ skl_compute_ddb(struct intel_atomic_state *state) for_each_oldnew_intel_crtc_in_state(state, crtc, old_crtc_state, new_crtc_state, i) { - ret = skl_allocate_pipe_ddb(new_crtc_state); + ret = skl_allocate_pipe_ddb(state, crtc); if (ret) return ret; @@ -6092,7 +6097,6 @@ skl_compute_wm(struct intel_atomic_state *state) { struct intel_crtc *crtc; struct intel_crtc_state *new_crtc_state; - struct intel_crtc_state *old_crtc_state; int ret, i; ret = skl_ddb_add_affected_pipes(state); @@ -6104,9 +6108,8 @@ skl_compute_wm(struct intel_atomic_state *state) * Note that skl_ddb_add_affected_pipes may have added more CRTC's that * weren't otherwise being modified if pipe allocations had to change. */ - for_each_oldnew_intel_crtc_in_state(state, crtc, old_crtc_state, - new_crtc_state, i) { - ret = skl_build_pipe_wm(new_crtc_state); + for_each_new_intel_crtc_in_state(state, crtc, new_crtc_state, i) { + ret = skl_build_pipe_wm(state, crtc); if (ret) return ret; } @@ -6124,8 +6127,7 @@ skl_compute_wm(struct intel_atomic_state *state) * based on how much ddb is available. Now we can actually * check if the final watermarks changed. */ - for_each_oldnew_intel_crtc_in_state(state, crtc, old_crtc_state, - new_crtc_state, i) { + for_each_new_intel_crtc_in_state(state, crtc, new_crtc_state, i) { ret = skl_wm_add_affected_planes(state, crtc); if (ret) return ret; -- cgit From dbf71381d73390d1e2cedd84eb04ea9b32fd96b3 Mon Sep 17 00:00:00 2001 From: Ville Syrjälä Date: Fri, 6 Nov 2020 19:30:38 +0200 Subject: drm/i915: Nuke intel_atomic_crtc_state_for_each_plane_state() from skl+ wm code MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit intel_atomic_crtc_state_for_each_plane_state() peeks at the plane's current state without holding the plane's mutex, trusting that the crtc's mutex will protect it. In practice that does work since our planes can't move between pipes, but it sets a bad example. intel_atomic_crtc_state_for_each_plane_state() also relies on crtc_state.uapi.plane_mask which may be full of lies when it comes to the bigjoiner stuff, so soon we can't use it as is anyway. So best to just get rid of it entirely. Which we can easily do by switching to the g4x/vlv "raw" watermark approach. Later on we should even be able to move the "raw" watermark computation into the normal .plane_check() code, leaving only the merging/clamping of the final watermarks to the later stages. But that will require adjusting the ilk+ wm code similarly as well. Signed-off-by: Ville Syrjälä Reviewed-by: Stanislav Lisovskiy Signed-off-by: Manasi Navare Link: https://patchwork.freedesktop.org/patch/msgid/20201106173042.7534-3-ville.syrjala@linux.intel.com --- drivers/gpu/drm/i915/display/intel_display_types.h | 2 ++ drivers/gpu/drm/i915/intel_pm.c | 41 +++++++++++++--------- 2 files changed, 27 insertions(+), 16 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_pm.c') diff --git a/drivers/gpu/drm/i915/display/intel_display_types.h b/drivers/gpu/drm/i915/display/intel_display_types.h index 35ab5944a3f7..3d91b116aadc 100644 --- a/drivers/gpu/drm/i915/display/intel_display_types.h +++ b/drivers/gpu/drm/i915/display/intel_display_types.h @@ -755,6 +755,8 @@ struct intel_crtc_wm_state { } ilk; struct { + /* "raw" watermarks */ + struct skl_pipe_wm raw; /* gen9+ only needs 1-step wm programming */ struct skl_pipe_wm optimal; struct skl_ddb_entry ddb; diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c index a413e4b10e45..98eec96d5c23 100644 --- a/drivers/gpu/drm/i915/intel_pm.c +++ b/drivers/gpu/drm/i915/intel_pm.c @@ -5480,7 +5480,7 @@ static int skl_build_plane_wm_single(struct intel_crtc_state *crtc_state, { struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc); struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); - struct skl_plane_wm *wm = &crtc_state->wm.skl.optimal.planes[plane_id]; + struct skl_plane_wm *wm = &crtc_state->wm.skl.raw.planes[plane_id]; struct skl_wm_params wm_params; int ret; @@ -5503,7 +5503,7 @@ static int skl_build_plane_wm_uv(struct intel_crtc_state *crtc_state, const struct intel_plane_state *plane_state, enum plane_id plane_id) { - struct skl_plane_wm *wm = &crtc_state->wm.skl.optimal.planes[plane_id]; + struct skl_plane_wm *wm = &crtc_state->wm.skl.raw.planes[plane_id]; struct skl_wm_params wm_params; int ret; @@ -5524,10 +5524,13 @@ static int skl_build_plane_wm(struct intel_crtc_state *crtc_state, const struct intel_plane_state *plane_state) { struct intel_plane *plane = to_intel_plane(plane_state->uapi.plane); - const struct drm_framebuffer *fb = plane_state->hw.fb; enum plane_id plane_id = plane->id; + struct skl_plane_wm *wm = &crtc_state->wm.skl.raw.planes[plane_id]; + const struct drm_framebuffer *fb = plane_state->hw.fb; int ret; + memset(wm, 0, sizeof(*wm)); + if (!intel_wm_plane_visible(crtc_state, plane_state)) return 0; @@ -5549,10 +5552,14 @@ static int skl_build_plane_wm(struct intel_crtc_state *crtc_state, static int icl_build_plane_wm(struct intel_crtc_state *crtc_state, const struct intel_plane_state *plane_state) { - struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev); - enum plane_id plane_id = to_intel_plane(plane_state->uapi.plane)->id; + struct intel_plane *plane = to_intel_plane(plane_state->uapi.plane); + struct drm_i915_private *dev_priv = to_i915(plane->base.dev); + enum plane_id plane_id = plane->id; + struct skl_plane_wm *wm = &crtc_state->wm.skl.raw.planes[plane_id]; int ret; + memset(wm, 0, sizeof(*wm)); + /* Watermarks calculated in master */ if (plane_state->planar_slave) return 0; @@ -5591,19 +5598,18 @@ static int skl_build_pipe_wm(struct intel_atomic_state *state, struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); struct intel_crtc_state *crtc_state = intel_atomic_get_new_crtc_state(state, crtc); - struct skl_pipe_wm *pipe_wm = &crtc_state->wm.skl.optimal; - struct intel_plane *plane; const struct intel_plane_state *plane_state; - int ret; - - /* - * We'll only calculate watermarks for planes that are actually - * enabled, so make sure all other planes are set as disabled. - */ - memset(pipe_wm->planes, 0, sizeof(pipe_wm->planes)); + struct intel_plane *plane; + int ret, i; - intel_atomic_crtc_state_for_each_plane_state(plane, plane_state, - crtc_state) { + for_each_new_intel_plane_in_state(state, plane, plane_state, i) { + /* + * FIXME should perhaps check {old,new}_plane_crtc->hw.crtc + * instead but we don't populate that correctly for NV12 Y + * planes so for now hack this. + */ + if (plane->pipe != crtc->pipe) + continue; if (INTEL_GEN(dev_priv) >= 11) ret = icl_build_plane_wm(crtc_state, plane_state); @@ -5613,6 +5619,8 @@ static int skl_build_pipe_wm(struct intel_atomic_state *state, return ret; } + crtc_state->wm.skl.optimal = crtc_state->wm.skl.raw; + return 0; } @@ -6273,6 +6281,7 @@ void skl_wm_get_hw_state(struct drm_i915_private *dev_priv) crtc_state = to_intel_crtc_state(crtc->base.state); skl_pipe_wm_get_hw_state(crtc, &crtc_state->wm.skl.optimal); + crtc_state->wm.skl.raw = crtc_state->wm.skl.optimal; } if (dev_priv->active_pipes) { -- cgit From 9c31212b247832d71882a9942f37025544264892 Mon Sep 17 00:00:00 2001 From: Ville Syrjälä Date: Fri, 6 Nov 2020 19:30:40 +0200 Subject: drm/i915: Precompute can_sagv for each wm level MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit In order to remove intel_atomic_crtc_state_for_each_plane_state() from skl_crtc_can_enable_sagv() we can simply precompute whether each wm level can tolerate the SAGV block time latency or not. This has the nice side benefit that we remove the duplicated wm level latency calculation. In fact the copy of that code we had in skl_crtc_can_enable_sagv() didn't even handle WaIncreaseLatencyIPCEnabled/Display WA #1141 whereas the copy in skl_compute_plane_wm() did. So now we just have the one copy which handles all the w/as. Signed-off-by: Ville Syrjälä Link: https://patchwork.freedesktop.org/patch/msgid/20201106173042.7534-5-ville.syrjala@linux.intel.com Reviewed-by: Stanislav Lisovskiy --- drivers/gpu/drm/i915/display/intel_display_types.h | 1 + drivers/gpu/drm/i915/intel_pm.c | 21 ++++++++------------- 2 files changed, 9 insertions(+), 13 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_pm.c') diff --git a/drivers/gpu/drm/i915/display/intel_display_types.h b/drivers/gpu/drm/i915/display/intel_display_types.h index 0c1df34a0cce..86de89d621d8 100644 --- a/drivers/gpu/drm/i915/display/intel_display_types.h +++ b/drivers/gpu/drm/i915/display/intel_display_types.h @@ -686,6 +686,7 @@ struct skl_wm_level { u8 plane_res_l; bool plane_en; bool ignore_lines; + bool can_sagv; }; struct skl_plane_wm { diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c index 98eec96d5c23..7b747088d3da 100644 --- a/drivers/gpu/drm/i915/intel_pm.c +++ b/drivers/gpu/drm/i915/intel_pm.c @@ -3873,9 +3873,7 @@ static bool skl_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state) { struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc); struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); - struct intel_plane *plane; - const struct intel_plane_state *plane_state; - int level, latency; + enum plane_id plane_id; if (!intel_has_sagv(dev_priv)) return false; @@ -3886,9 +3884,10 @@ static bool skl_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state) if (crtc_state->hw.pipe_mode.flags & DRM_MODE_FLAG_INTERLACE) return false; - intel_atomic_crtc_state_for_each_plane_state(plane, plane_state, crtc_state) { + for_each_plane_id_on_crtc(crtc, plane_id) { const struct skl_plane_wm *wm = - &crtc_state->wm.skl.optimal.planes[plane->id]; + &crtc_state->wm.skl.optimal.planes[plane_id]; + int level; /* Skip this plane if it's not enabled */ if (!wm->wm[0].plane_en) @@ -3899,19 +3898,12 @@ static bool skl_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state) !wm->wm[level].plane_en; --level) { } - latency = dev_priv->wm.skl_latency[level]; - - if (skl_needs_memory_bw_wa(dev_priv) && - plane_state->uapi.fb->modifier == - I915_FORMAT_MOD_X_TILED) - latency += 15; - /* * If any of the planes on this pipe don't enable wm levels that * incur memory latencies higher than sagv_block_time_us we * can't enable SAGV. */ - if (latency < dev_priv->sagv_block_time_us) + if (!wm->wm[level].can_sagv) return false; } @@ -5375,6 +5367,9 @@ static void skl_compute_plane_wm(const struct intel_crtc_state *crtc_state, /* Bspec says: value >= plane ddb allocation -> invalid, hence the +1 here */ result->min_ddb_alloc = max(min_ddb_alloc, res_blocks) + 1; result->plane_en = true; + + if (INTEL_GEN(dev_priv) < 12) + result->can_sagv = latency >= dev_priv->sagv_block_time_us; } static void -- cgit From ab01630b64ce19e1b845a4d83879eb9727306aa1 Mon Sep 17 00:00:00 2001 From: Ville Syrjälä Date: Fri, 6 Nov 2020 19:30:41 +0200 Subject: drm/i915: Store plane relative data rate in crtc_state MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Store the relative data rate for planes in the crtc state so that we don't have to use intel_atomic_crtc_state_for_each_plane_state() to compute it even for the planes that are no part of the current state. Should probably just nuke this stuff entirely an use the normal plane data rate instead. The two are slightly different since this relative data rate doesn't factor in the actual pixel clock, so it's a bit odd thing to even call a "data rate". And since the watermarks are computed based on the actual data rate anyway I don't really see what the point of this relative data rate is. But that's for the future... Signed-off-by: Ville Syrjälä Link: https://patchwork.freedesktop.org/patch/msgid/20201106173042.7534-6-ville.syrjala@linux.intel.com Reviewed-by: Stanislav Lisovskiy --- drivers/gpu/drm/i915/display/intel_display_types.h | 4 ++ drivers/gpu/drm/i915/intel_pm.c | 83 ++++++++++++---------- 2 files changed, 50 insertions(+), 37 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_pm.c') diff --git a/drivers/gpu/drm/i915/display/intel_display_types.h b/drivers/gpu/drm/i915/display/intel_display_types.h index 86de89d621d8..5bb7adc1ff3e 100644 --- a/drivers/gpu/drm/i915/display/intel_display_types.h +++ b/drivers/gpu/drm/i915/display/intel_display_types.h @@ -1031,6 +1031,10 @@ struct intel_crtc_state { u32 data_rate[I915_MAX_PLANES]; + /* FIXME unify with data_rate[] */ + u64 plane_data_rate[I915_MAX_PLANES]; + u64 uv_plane_data_rate[I915_MAX_PLANES]; + /* Gamma mode programmed on the pipe */ u32 gamma_mode; diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c index 7b747088d3da..e5ebaec434f4 100644 --- a/drivers/gpu/drm/i915/intel_pm.c +++ b/drivers/gpu/drm/i915/intel_pm.c @@ -4696,50 +4696,63 @@ skl_plane_relative_data_rate(const struct intel_crtc_state *crtc_state, } static u64 -skl_get_total_relative_data_rate(struct intel_crtc_state *crtc_state, - u64 *plane_data_rate, - u64 *uv_plane_data_rate) +skl_get_total_relative_data_rate(struct intel_atomic_state *state, + struct intel_crtc *crtc) { - struct intel_plane *plane; + struct intel_crtc_state *crtc_state = + intel_atomic_get_new_crtc_state(state, crtc); const struct intel_plane_state *plane_state; + struct intel_plane *plane; u64 total_data_rate = 0; + enum plane_id plane_id; + int i; /* Calculate and cache data rate for each plane */ - intel_atomic_crtc_state_for_each_plane_state(plane, plane_state, crtc_state) { - enum plane_id plane_id = plane->id; - u64 rate; + for_each_new_intel_plane_in_state(state, plane, plane_state, i) { + if (plane->pipe != crtc->pipe) + continue; + + plane_id = plane->id; /* packed/y */ - rate = skl_plane_relative_data_rate(crtc_state, plane_state, 0); - plane_data_rate[plane_id] = rate; - total_data_rate += rate; + crtc_state->plane_data_rate[plane_id] = + skl_plane_relative_data_rate(crtc_state, plane_state, 0); /* uv-plane */ - rate = skl_plane_relative_data_rate(crtc_state, plane_state, 1); - uv_plane_data_rate[plane_id] = rate; - total_data_rate += rate; + crtc_state->uv_plane_data_rate[plane_id] = + skl_plane_relative_data_rate(crtc_state, plane_state, 1); + } + + for_each_plane_id_on_crtc(crtc, plane_id) { + total_data_rate += crtc_state->plane_data_rate[plane_id]; + total_data_rate += crtc_state->uv_plane_data_rate[plane_id]; } return total_data_rate; } static u64 -icl_get_total_relative_data_rate(struct intel_crtc_state *crtc_state, - u64 *plane_data_rate) +icl_get_total_relative_data_rate(struct intel_atomic_state *state, + struct intel_crtc *crtc) { - struct intel_plane *plane; + struct intel_crtc_state *crtc_state = + intel_atomic_get_new_crtc_state(state, crtc); const struct intel_plane_state *plane_state; + struct intel_plane *plane; u64 total_data_rate = 0; + enum plane_id plane_id; + int i; /* Calculate and cache data rate for each plane */ - intel_atomic_crtc_state_for_each_plane_state(plane, plane_state, crtc_state) { - enum plane_id plane_id = plane->id; - u64 rate; + for_each_new_intel_plane_in_state(state, plane, plane_state, i) { + if (plane->pipe != crtc->pipe) + continue; + + plane_id = plane->id; if (!plane_state->planar_linked_plane) { - rate = skl_plane_relative_data_rate(crtc_state, plane_state, 0); - plane_data_rate[plane_id] = rate; - total_data_rate += rate; + crtc_state->plane_data_rate[plane_id] = + skl_plane_relative_data_rate(crtc_state, plane_state, 0); } else { enum plane_id y_plane_id; @@ -4754,17 +4767,18 @@ icl_get_total_relative_data_rate(struct intel_crtc_state *crtc_state, continue; /* Y plane rate is calculated on the slave */ - rate = skl_plane_relative_data_rate(crtc_state, plane_state, 0); y_plane_id = plane_state->planar_linked_plane->id; - plane_data_rate[y_plane_id] = rate; - total_data_rate += rate; + crtc_state->plane_data_rate[y_plane_id] = + skl_plane_relative_data_rate(crtc_state, plane_state, 0); - rate = skl_plane_relative_data_rate(crtc_state, plane_state, 1); - plane_data_rate[plane_id] = rate; - total_data_rate += rate; + crtc_state->plane_data_rate[plane_id] = + skl_plane_relative_data_rate(crtc_state, plane_state, 1); } } + for_each_plane_id_on_crtc(crtc, plane_id) + total_data_rate += crtc_state->plane_data_rate[plane_id]; + return total_data_rate; } @@ -4796,8 +4810,6 @@ skl_allocate_pipe_ddb(struct intel_atomic_state *state, u64 total_data_rate; enum plane_id plane_id; int num_active; - u64 plane_data_rate[I915_MAX_PLANES] = {}; - u64 uv_plane_data_rate[I915_MAX_PLANES] = {}; u32 blocks; int level; int ret; @@ -4837,13 +4849,10 @@ skl_allocate_pipe_ddb(struct intel_atomic_state *state, if (INTEL_GEN(dev_priv) >= 11) total_data_rate = - icl_get_total_relative_data_rate(crtc_state, - plane_data_rate); + icl_get_total_relative_data_rate(state, crtc); else total_data_rate = - skl_get_total_relative_data_rate(crtc_state, - plane_data_rate, - uv_plane_data_rate); + skl_get_total_relative_data_rate(state, crtc); ret = skl_ddb_get_pipe_allocation_limits(dev_priv, crtc_state, total_data_rate, @@ -4924,7 +4933,7 @@ skl_allocate_pipe_ddb(struct intel_atomic_state *state, if (total_data_rate == 0) break; - rate = plane_data_rate[plane_id]; + rate = crtc_state->plane_data_rate[plane_id]; extra = min_t(u16, alloc_size, DIV64_U64_ROUND_UP(alloc_size * rate, total_data_rate)); @@ -4935,7 +4944,7 @@ skl_allocate_pipe_ddb(struct intel_atomic_state *state, if (total_data_rate == 0) break; - rate = uv_plane_data_rate[plane_id]; + rate = crtc_state->uv_plane_data_rate[plane_id]; extra = min_t(u16, alloc_size, DIV64_U64_ROUND_UP(alloc_size * rate, total_data_rate)); -- cgit From 3df3fe2412000fec34f9185f67f7c63d37b615ad Mon Sep 17 00:00:00 2001 From: Ville Syrjälä Date: Fri, 6 Nov 2020 19:30:42 +0200 Subject: drm/i915: Remove skl_adjusted_plane_pixel_rate() MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Replace skl_adjusted_plane_pixel_rate() with the generic intel_plane_pixel_rate(). The two should produce identical results. Signed-off-by: Ville Syrjälä Link: https://patchwork.freedesktop.org/patch/msgid/20201106173042.7534-7-ville.syrjala@linux.intel.com Reviewed-by: Stanislav Lisovskiy --- drivers/gpu/drm/i915/intel_pm.c | 27 ++------------------------- 1 file changed, 2 insertions(+), 25 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_pm.c') diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c index e5ebaec434f4..ad76e0a835d4 100644 --- a/drivers/gpu/drm/i915/intel_pm.c +++ b/drivers/gpu/drm/i915/intel_pm.c @@ -33,6 +33,7 @@ #include #include "display/intel_atomic.h" +#include "display/intel_atomic_plane.h" #include "display/intel_bw.h" #include "display/intel_display_types.h" #include "display/intel_fbc.h" @@ -5102,30 +5103,6 @@ intel_get_linetime_us(const struct intel_crtc_state *crtc_state) return linetime_us; } -static u32 -skl_adjusted_plane_pixel_rate(const struct intel_crtc_state *crtc_state, - const struct intel_plane_state *plane_state) -{ - struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev); - u64 adjusted_pixel_rate; - uint_fixed_16_16_t downscale_amount; - - /* Shouldn't reach here on disabled planes... */ - if (drm_WARN_ON(&dev_priv->drm, - !intel_wm_plane_visible(crtc_state, plane_state))) - return 0; - - /* - * Adjusted plane pixel rate is just the pipe's adjusted pixel rate - * with additional adjustments for plane-specific scaling. - */ - adjusted_pixel_rate = crtc_state->pixel_rate; - downscale_amount = skl_plane_downscale_amount(crtc_state, plane_state); - - return mul_round_up_u32_fixed16(adjusted_pixel_rate, - downscale_amount); -} - static int skl_compute_wm_params(const struct intel_crtc_state *crtc_state, int width, const struct drm_format_info *format, @@ -5238,7 +5215,7 @@ skl_compute_plane_wm_params(const struct intel_crtc_state *crtc_state, return skl_compute_wm_params(crtc_state, width, fb->format, fb->modifier, plane_state->hw.rotation, - skl_adjusted_plane_pixel_rate(crtc_state, plane_state), + intel_plane_pixel_rate(crtc_state, plane_state), wp, color_plane); } -- cgit