diff options
Diffstat (limited to 'drivers/gpu/drm/i915/intel_pm.c')
| -rw-r--r-- | drivers/gpu/drm/i915/intel_pm.c | 882 |
1 files changed, 410 insertions, 472 deletions
diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c index 2d2481392824..a2f751cd187a 100644 --- a/drivers/gpu/drm/i915/intel_pm.c +++ b/drivers/gpu/drm/i915/intel_pm.c @@ -340,6 +340,11 @@ void intel_set_memory_cxsr(struct drm_i915_private *dev_priv, bool enable) I915_WRITE(FW_BLC_SELF, val); POSTING_READ(FW_BLC_SELF); } else if (IS_I915GM(dev)) { + /* + * FIXME can't find a bit like this for 915G, and + * and yet it does have the related watermark in + * FW_BLC_SELF. What's going on? + */ val = enable ? _MASKED_BIT_ENABLE(INSTPM_SELF_EN) : _MASKED_BIT_DISABLE(INSTPM_SELF_EN); I915_WRITE(INSTPM, val); @@ -960,7 +965,7 @@ static uint16_t vlv_compute_wm_level(struct intel_plane *plane, if (dev_priv->wm.pri_latency[level] == 0) return USHRT_MAX; - if (!state->visible) + if (!state->base.visible) return 0; cpp = drm_format_plane_cpp(state->base.fb->pixel_format, 0); @@ -1002,7 +1007,7 @@ static void vlv_compute_fifo(struct intel_crtc *crtc) if (plane->base.type == DRM_PLANE_TYPE_CURSOR) continue; - if (state->visible) { + if (state->base.visible) { wm_state->num_active_planes++; total_rate += drm_format_plane_cpp(state->base.fb->pixel_format, 0); } @@ -1018,7 +1023,7 @@ static void vlv_compute_fifo(struct intel_crtc *crtc) continue; } - if (!state->visible) { + if (!state->base.visible) { plane->wm.fifo_size = 0; continue; } @@ -1118,7 +1123,7 @@ static void vlv_compute_wm(struct intel_crtc *crtc) struct intel_plane_state *state = to_intel_plane_state(plane->base.state); - if (!state->visible) + if (!state->base.visible) continue; /* normal watermarks */ @@ -1580,7 +1585,7 @@ static void i9xx_update_wm(struct drm_crtc *unused_crtc) obj = intel_fb_obj(enabled->primary->state->fb); /* self-refresh seems busted with untiled */ - if (obj->tiling_mode == I915_TILING_NONE) + if (!i915_gem_object_is_tiled(obj)) enabled = NULL; } @@ -1604,6 +1609,9 @@ static void i9xx_update_wm(struct drm_crtc *unused_crtc) unsigned long line_time_us; int entries; + if (IS_I915GM(dev) || IS_I945GM(dev)) + cpp = 4; + line_time_us = max(htotal * 1000 / clock, 1); /* Use ns/us then divide to preserve precision */ @@ -1618,7 +1626,7 @@ static void i9xx_update_wm(struct drm_crtc *unused_crtc) if (IS_I945G(dev) || IS_I945GM(dev)) I915_WRITE(FW_BLC_SELF, FW_BLC_SELF_FIFO_MASK | (srwm & 0xff)); - else if (IS_I915GM(dev)) + else I915_WRITE(FW_BLC_SELF, srwm & 0x3f); } @@ -1767,7 +1775,7 @@ static uint32_t ilk_compute_pri_wm(const struct intel_crtc_state *cstate, drm_format_plane_cpp(pstate->base.fb->pixel_format, 0) : 0; uint32_t method1, method2; - if (!cstate->base.active || !pstate->visible) + if (!cstate->base.active || !pstate->base.visible) return 0; method1 = ilk_wm_method1(ilk_pipe_pixel_rate(cstate), cpp, mem_value); @@ -1777,7 +1785,7 @@ static uint32_t ilk_compute_pri_wm(const struct intel_crtc_state *cstate, method2 = ilk_wm_method2(ilk_pipe_pixel_rate(cstate), cstate->base.adjusted_mode.crtc_htotal, - drm_rect_width(&pstate->dst), + drm_rect_width(&pstate->base.dst), cpp, mem_value); return min(method1, method2); @@ -1795,13 +1803,13 @@ static uint32_t ilk_compute_spr_wm(const struct intel_crtc_state *cstate, drm_format_plane_cpp(pstate->base.fb->pixel_format, 0) : 0; uint32_t method1, method2; - if (!cstate->base.active || !pstate->visible) + if (!cstate->base.active || !pstate->base.visible) return 0; method1 = ilk_wm_method1(ilk_pipe_pixel_rate(cstate), cpp, mem_value); method2 = ilk_wm_method2(ilk_pipe_pixel_rate(cstate), cstate->base.adjusted_mode.crtc_htotal, - drm_rect_width(&pstate->dst), + drm_rect_width(&pstate->base.dst), cpp, mem_value); return min(method1, method2); } @@ -1820,7 +1828,7 @@ static uint32_t ilk_compute_cur_wm(const struct intel_crtc_state *cstate, * this is necessary to avoid flickering. */ int cpp = 4; - int width = pstate->visible ? pstate->base.crtc_w : 64; + int width = pstate->base.visible ? pstate->base.crtc_w : 64; if (!cstate->base.active) return 0; @@ -1838,10 +1846,10 @@ static uint32_t ilk_compute_fbc_wm(const struct intel_crtc_state *cstate, int cpp = pstate->base.fb ? drm_format_plane_cpp(pstate->base.fb->pixel_format, 0) : 0; - if (!cstate->base.active || !pstate->visible) + if (!cstate->base.active || !pstate->base.visible) return 0; - return ilk_wm_fbc(pri_val, drm_rect_width(&pstate->dst), cpp); + return ilk_wm_fbc(pri_val, drm_rect_width(&pstate->base.dst), cpp); } static unsigned int ilk_display_fifo_size(const struct drm_device *dev) @@ -2119,32 +2127,34 @@ static void intel_read_wm_latency(struct drm_device *dev, uint16_t wm[8]) GEN9_MEM_LATENCY_LEVEL_MASK; /* + * If a level n (n > 1) has a 0us latency, all levels m (m >= n) + * need to be disabled. We make sure to sanitize the values out + * of the punit to satisfy this requirement. + */ + for (level = 1; level <= max_level; level++) { + if (wm[level] == 0) { + for (i = level + 1; i <= max_level; i++) + wm[i] = 0; + break; + } + } + + /* * WaWmMemoryReadLatency:skl * * punit doesn't take into account the read latency so we need - * to add 2us to the various latency levels we retrieve from - * the punit. - * - W0 is a bit special in that it's the only level that - * can't be disabled if we want to have display working, so - * we always add 2us there. - * - For levels >=1, punit returns 0us latency when they are - * disabled, so we respect that and don't add 2us then - * - * Additionally, if a level n (n > 1) has a 0us latency, all - * levels m (m >= n) need to be disabled. We make sure to - * sanitize the values out of the punit to satisfy this - * requirement. + * to add 2us to the various latency levels we retrieve from the + * punit when level 0 response data us 0us. */ - wm[0] += 2; - for (level = 1; level <= max_level; level++) - if (wm[level] != 0) + if (wm[0] == 0) { + wm[0] += 2; + for (level = 1; level <= max_level; level++) { + if (wm[level] == 0) + break; wm[level] += 2; - else { - for (i = level + 1; i <= max_level; i++) - wm[i] = 0; - - break; } + } + } else if (IS_HASWELL(dev) || IS_BROADWELL(dev)) { uint64_t sskpd = I915_READ64(MCH_SSKPD); @@ -2358,10 +2368,10 @@ static int ilk_compute_pipe_wm(struct intel_crtc_state *cstate) pipe_wm->pipe_enabled = cstate->base.active; if (sprstate) { - pipe_wm->sprites_enabled = sprstate->visible; - pipe_wm->sprites_scaled = sprstate->visible && - (drm_rect_width(&sprstate->dst) != drm_rect_width(&sprstate->src) >> 16 || - drm_rect_height(&sprstate->dst) != drm_rect_height(&sprstate->src) >> 16); + pipe_wm->sprites_enabled = sprstate->base.visible; + pipe_wm->sprites_scaled = sprstate->base.visible && + (drm_rect_width(&sprstate->base.dst) != drm_rect_width(&sprstate->base.src) >> 16 || + drm_rect_height(&sprstate->base.dst) != drm_rect_height(&sprstate->base.src) >> 16); } usable_level = max_level; @@ -2845,13 +2855,6 @@ bool ilk_disable_lp_wm(struct drm_device *dev) return _ilk_disable_lp_wm(dev_priv, WM_DIRTY_LP_ALL); } -/* - * On gen9, we need to allocate Display Data Buffer (DDB) portions to the - * different active planes. - */ - -#define SKL_DDB_SIZE 896 /* in blocks */ -#define BXT_DDB_SIZE 512 #define SKL_SAGV_BLOCK_TIME 30 /* µs */ /* @@ -2876,6 +2879,19 @@ skl_wm_plane_id(const struct intel_plane *plane) } } +static bool +intel_has_sagv(struct drm_i915_private *dev_priv) +{ + if (IS_KABYLAKE(dev_priv)) + return true; + + if (IS_SKYLAKE(dev_priv) && + dev_priv->sagv_status != I915_SAGV_NOT_CONTROLLED) + return true; + + return false; +} + /* * SAGV dynamically adjusts the system agent voltage and clock frequencies * depending on power and performance requirements. The display engine access @@ -2888,12 +2904,14 @@ skl_wm_plane_id(const struct intel_plane *plane) * - We're not using an interlaced display configuration */ int -skl_enable_sagv(struct drm_i915_private *dev_priv) +intel_enable_sagv(struct drm_i915_private *dev_priv) { int ret; - if (dev_priv->skl_sagv_status == I915_SKL_SAGV_NOT_CONTROLLED || - dev_priv->skl_sagv_status == I915_SKL_SAGV_ENABLED) + if (!intel_has_sagv(dev_priv)) + return 0; + + if (dev_priv->sagv_status == I915_SAGV_ENABLED) return 0; DRM_DEBUG_KMS("Enabling the SAGV\n"); @@ -2909,21 +2927,21 @@ skl_enable_sagv(struct drm_i915_private *dev_priv) * Some skl systems, pre-release machines in particular, * don't actually have an SAGV. */ - if (ret == -ENXIO) { + if (IS_SKYLAKE(dev_priv) && ret == -ENXIO) { DRM_DEBUG_DRIVER("No SAGV found on system, ignoring\n"); - dev_priv->skl_sagv_status = I915_SKL_SAGV_NOT_CONTROLLED; + dev_priv->sagv_status = I915_SAGV_NOT_CONTROLLED; return 0; } else if (ret < 0) { DRM_ERROR("Failed to enable the SAGV\n"); return ret; } - dev_priv->skl_sagv_status = I915_SKL_SAGV_ENABLED; + dev_priv->sagv_status = I915_SAGV_ENABLED; return 0; } static int -skl_do_sagv_disable(struct drm_i915_private *dev_priv) +intel_do_sagv_disable(struct drm_i915_private *dev_priv) { int ret; uint32_t temp = GEN9_SAGV_DISABLE; @@ -2937,19 +2955,21 @@ skl_do_sagv_disable(struct drm_i915_private *dev_priv) } int -skl_disable_sagv(struct drm_i915_private *dev_priv) +intel_disable_sagv(struct drm_i915_private *dev_priv) { int ret, result; - if (dev_priv->skl_sagv_status == I915_SKL_SAGV_NOT_CONTROLLED || - dev_priv->skl_sagv_status == I915_SKL_SAGV_DISABLED) + if (!intel_has_sagv(dev_priv)) + return 0; + + if (dev_priv->sagv_status == I915_SAGV_DISABLED) return 0; DRM_DEBUG_KMS("Disabling the SAGV\n"); mutex_lock(&dev_priv->rps.hw_lock); /* bspec says to keep retrying for at least 1 ms */ - ret = wait_for(result = skl_do_sagv_disable(dev_priv), 1); + ret = wait_for(result = intel_do_sagv_disable(dev_priv), 1); mutex_unlock(&dev_priv->rps.hw_lock); if (ret == -ETIMEDOUT) { @@ -2961,20 +2981,20 @@ skl_disable_sagv(struct drm_i915_private *dev_priv) * Some skl systems, pre-release machines in particular, * don't actually have an SAGV. */ - if (result == -ENXIO) { + if (IS_SKYLAKE(dev_priv) && result == -ENXIO) { DRM_DEBUG_DRIVER("No SAGV found on system, ignoring\n"); - dev_priv->skl_sagv_status = I915_SKL_SAGV_NOT_CONTROLLED; + dev_priv->sagv_status = I915_SAGV_NOT_CONTROLLED; return 0; } else if (result < 0) { DRM_ERROR("Failed to disable the SAGV\n"); return result; } - dev_priv->skl_sagv_status = I915_SKL_SAGV_DISABLED; + dev_priv->sagv_status = I915_SAGV_DISABLED; return 0; } -bool skl_can_enable_sagv(struct drm_atomic_state *state) +bool intel_can_enable_sagv(struct drm_atomic_state *state) { struct drm_device *dev = state->dev; struct drm_i915_private *dev_priv = to_i915(dev); @@ -2983,6 +3003,9 @@ bool skl_can_enable_sagv(struct drm_atomic_state *state) enum pipe pipe; int level, plane; + if (!intel_has_sagv(dev_priv)) + return false; + /* * SKL workaround: bspec recommends we disable the SAGV when we have * more then one pipe enabled @@ -3049,10 +3072,8 @@ skl_ddb_get_pipe_allocation_limits(struct drm_device *dev, else *num_active = hweight32(dev_priv->active_crtcs); - if (IS_BROXTON(dev)) - ddb_size = BXT_DDB_SIZE; - else - ddb_size = SKL_DDB_SIZE; + ddb_size = INTEL_INFO(dev_priv)->ddb_size; + WARN_ON(ddb_size == 0); ddb_size -= 4; /* 4 blocks for bypass path allocation */ @@ -3144,14 +3165,14 @@ skl_plane_downscale_amount(const struct intel_plane_state *pstate) uint32_t downscale_h, downscale_w; uint32_t src_w, src_h, dst_w, dst_h; - if (WARN_ON(!pstate->visible)) + if (WARN_ON(!pstate->base.visible)) return DRM_PLANE_HELPER_NO_SCALING; /* n.b., src is 16.16 fixed point, dst is whole integer */ - src_w = drm_rect_width(&pstate->src); - src_h = drm_rect_height(&pstate->src); - dst_w = drm_rect_width(&pstate->dst); - dst_h = drm_rect_height(&pstate->dst); + src_w = drm_rect_width(&pstate->base.src); + src_h = drm_rect_height(&pstate->base.src); + dst_w = drm_rect_width(&pstate->base.dst); + dst_h = drm_rect_height(&pstate->base.dst); if (intel_rotation_90_or_270(pstate->base.rotation)) swap(dst_w, dst_h); @@ -3173,15 +3194,15 @@ skl_plane_relative_data_rate(const struct intel_crtc_state *cstate, uint32_t width = 0, height = 0; unsigned format = fb ? fb->pixel_format : DRM_FORMAT_XRGB8888; - if (!intel_pstate->visible) + if (!intel_pstate->base.visible) return 0; if (pstate->plane->type == DRM_PLANE_TYPE_CURSOR) return 0; if (y && format != DRM_FORMAT_NV12) return 0; - width = drm_rect_width(&intel_pstate->src) >> 16; - height = drm_rect_height(&intel_pstate->src) >> 16; + width = drm_rect_width(&intel_pstate->base.src) >> 16; + height = drm_rect_height(&intel_pstate->base.src) >> 16; if (intel_rotation_90_or_270(pstate->rotation)) swap(width, height); @@ -3280,8 +3301,8 @@ skl_ddb_min_alloc(const struct drm_plane_state *pstate, fb->modifier[0] != I915_FORMAT_MOD_Yf_TILED) return 8; - src_w = drm_rect_width(&intel_pstate->src) >> 16; - src_h = drm_rect_height(&intel_pstate->src) >> 16; + src_w = drm_rect_width(&intel_pstate->base.src) >> 16; + src_h = drm_rect_height(&intel_pstate->base.src) >> 16; if (intel_rotation_90_or_270(pstate->rotation)) swap(src_w, src_h); @@ -3372,7 +3393,7 @@ skl_allocate_pipe_ddb(struct intel_crtc_state *cstate, if (intel_plane->pipe != pipe) continue; - if (!to_intel_plane_state(pstate)->visible) { + if (!to_intel_plane_state(pstate)->base.visible) { minimum[id] = 0; y_minimum[id] = 0; continue; @@ -3473,29 +3494,14 @@ static uint32_t skl_wm_method1(uint32_t pixel_rate, uint8_t cpp, uint32_t latenc } static uint32_t skl_wm_method2(uint32_t pixel_rate, uint32_t pipe_htotal, - uint32_t horiz_pixels, uint8_t cpp, - uint64_t tiling, uint32_t latency) + uint32_t latency, uint32_t plane_blocks_per_line) { uint32_t ret; - uint32_t plane_bytes_per_line, plane_blocks_per_line; uint32_t wm_intermediate_val; if (latency == 0) return UINT_MAX; - plane_bytes_per_line = horiz_pixels * cpp; - - if (tiling == I915_FORMAT_MOD_Y_TILED || - tiling == I915_FORMAT_MOD_Yf_TILED) { - plane_bytes_per_line *= 4; - plane_blocks_per_line = DIV_ROUND_UP(plane_bytes_per_line, 512); - plane_blocks_per_line /= 4; - } else if (tiling == DRM_FORMAT_MOD_NONE) { - plane_blocks_per_line = DIV_ROUND_UP(plane_bytes_per_line, 512) + 1; - } else { - plane_blocks_per_line = DIV_ROUND_UP(plane_bytes_per_line, 512); - } - wm_intermediate_val = latency * pixel_rate; ret = DIV_ROUND_UP(wm_intermediate_val, pipe_htotal * 1000) * plane_blocks_per_line; @@ -3511,7 +3517,7 @@ static uint32_t skl_adjusted_plane_pixel_rate(const struct intel_crtc_state *cst uint64_t pixel_rate; /* Shouldn't reach here on disabled planes... */ - if (WARN_ON(!pstate->visible)) + if (WARN_ON(!pstate->base.visible)) return 0; /* @@ -3546,14 +3552,15 @@ static int skl_compute_plane_wm(const struct drm_i915_private *dev_priv, uint8_t cpp; uint32_t width = 0, height = 0; uint32_t plane_pixel_rate; + uint32_t y_tile_minimum, y_min_scanlines; - if (latency == 0 || !cstate->base.active || !intel_pstate->visible) { + if (latency == 0 || !cstate->base.active || !intel_pstate->base.visible) { *enabled = false; return 0; } - width = drm_rect_width(&intel_pstate->src) >> 16; - height = drm_rect_height(&intel_pstate->src) >> 16; + width = drm_rect_width(&intel_pstate->base.src) >> 16; + height = drm_rect_height(&intel_pstate->base.src) >> 16; if (intel_rotation_90_or_270(pstate->rotation)) swap(width, height); @@ -3561,38 +3568,51 @@ static int skl_compute_plane_wm(const struct drm_i915_private *dev_priv, cpp = drm_format_plane_cpp(fb->pixel_format, 0); plane_pixel_rate = skl_adjusted_plane_pixel_rate(cstate, intel_pstate); + if (intel_rotation_90_or_270(pstate->rotation)) { + int cpp = (fb->pixel_format == DRM_FORMAT_NV12) ? + drm_format_plane_cpp(fb->pixel_format, 1) : + drm_format_plane_cpp(fb->pixel_format, 0); + + switch (cpp) { + case 1: + y_min_scanlines = 16; + break; + case 2: + y_min_scanlines = 8; + break; + default: + WARN(1, "Unsupported pixel depth for rotation"); + case 4: + y_min_scanlines = 4; + break; + } + } else { + y_min_scanlines = 4; + } + + plane_bytes_per_line = width * cpp; + if (fb->modifier[0] == I915_FORMAT_MOD_Y_TILED || + fb->modifier[0] == I915_FORMAT_MOD_Yf_TILED) { + plane_blocks_per_line = + DIV_ROUND_UP(plane_bytes_per_line * y_min_scanlines, 512); + plane_blocks_per_line /= y_min_scanlines; + } else if (fb->modifier[0] == DRM_FORMAT_MOD_NONE) { + plane_blocks_per_line = DIV_ROUND_UP(plane_bytes_per_line, 512) + + 1; + } else { + plane_blocks_per_line = DIV_ROUND_UP(plane_bytes_per_line, 512); + } + method1 = skl_wm_method1(plane_pixel_rate, cpp, latency); method2 = skl_wm_method2(plane_pixel_rate, cstate->base.adjusted_mode.crtc_htotal, - width, - cpp, - fb->modifier[0], - latency); + latency, + plane_blocks_per_line); - plane_bytes_per_line = width * cpp; - plane_blocks_per_line = DIV_ROUND_UP(plane_bytes_per_line, 512); + y_tile_minimum = plane_blocks_per_line * y_min_scanlines; if (fb->modifier[0] == I915_FORMAT_MOD_Y_TILED || fb->modifier[0] == I915_FORMAT_MOD_Yf_TILED) { - uint32_t min_scanlines = 4; - uint32_t y_tile_minimum; - if (intel_rotation_90_or_270(pstate->rotation)) { - int cpp = (fb->pixel_format == DRM_FORMAT_NV12) ? - drm_format_plane_cpp(fb->pixel_format, 1) : - drm_format_plane_cpp(fb->pixel_format, 0); - - switch (cpp) { - case 1: - min_scanlines = 16; - break; - case 2: - min_scanlines = 8; - break; - case 8: - WARN(1, "Unsupported pixel depth for rotation"); - } - } - y_tile_minimum = plane_blocks_per_line * min_scanlines; selected_result = max(method2, y_tile_minimum); } else { if ((ddb_allocation / plane_blocks_per_line) >= 1) @@ -3606,10 +3626,12 @@ static int skl_compute_plane_wm(const struct drm_i915_private *dev_priv, if (level >= 1 && level <= 7) { if (fb->modifier[0] == I915_FORMAT_MOD_Y_TILED || - fb->modifier[0] == I915_FORMAT_MOD_Yf_TILED) - res_lines += 4; - else + fb->modifier[0] == I915_FORMAT_MOD_Yf_TILED) { + res_blocks += y_tile_minimum; + res_lines += y_min_scanlines; + } else { res_blocks++; + } } if (res_blocks >= ddb_allocation || res_lines > 31) { @@ -3828,183 +3850,82 @@ static void skl_ddb_entry_write(struct drm_i915_private *dev_priv, I915_WRITE(reg, 0); } -static void skl_write_wm_values(struct drm_i915_private *dev_priv, - const struct skl_wm_values *new) +void skl_write_plane_wm(struct intel_crtc *intel_crtc, + const struct skl_wm_values *wm, + int plane) { - struct drm_device *dev = &dev_priv->drm; - struct intel_crtc *crtc; - - for_each_intel_crtc(dev, crtc) { - int i, level, max_level = ilk_wm_max_level(dev); - enum pipe pipe = crtc->pipe; - - if ((new->dirty_pipes & drm_crtc_mask(&crtc->base)) == 0) - continue; - if (!crtc->active) - continue; - - I915_WRITE(PIPE_WM_LINETIME(pipe), new->wm_linetime[pipe]); - - for (level = 0; level <= max_level; level++) { - for (i = 0; i < intel_num_planes(crtc); i++) - I915_WRITE(PLANE_WM(pipe, i, level), - new->plane[pipe][i][level]); - I915_WRITE(CUR_WM(pipe, level), - new->plane[pipe][PLANE_CURSOR][level]); - } - for (i = 0; i < intel_num_planes(crtc); i++) - I915_WRITE(PLANE_WM_TRANS(pipe, i), - new->plane_trans[pipe][i]); - I915_WRITE(CUR_WM_TRANS(pipe), - new->plane_trans[pipe][PLANE_CURSOR]); - - for (i = 0; i < intel_num_planes(crtc); i++) { - skl_ddb_entry_write(dev_priv, - PLANE_BUF_CFG(pipe, i), - &new->ddb.plane[pipe][i]); - skl_ddb_entry_write(dev_priv, - PLANE_NV12_BUF_CFG(pipe, i), - &new->ddb.y_plane[pipe][i]); - } + struct drm_crtc *crtc = &intel_crtc->base; + struct drm_device *dev = crtc->dev; + struct drm_i915_private *dev_priv = to_i915(dev); + int level, max_level = ilk_wm_max_level(dev); + enum pipe pipe = intel_crtc->pipe; - skl_ddb_entry_write(dev_priv, CUR_BUF_CFG(pipe), - &new->ddb.plane[pipe][PLANE_CURSOR]); + for (level = 0; level <= max_level; level++) { + I915_WRITE(PLANE_WM(pipe, plane, level), + wm->plane[pipe][plane][level]); } -} + I915_WRITE(PLANE_WM_TRANS(pipe, plane), wm->plane_trans[pipe][plane]); -/* - * When setting up a new DDB allocation arrangement, we need to correctly - * sequence the times at which the new allocations for the pipes are taken into - * account or we'll have pipes fetching from space previously allocated to - * another pipe. - * - * Roughly the sequence looks like: - * 1. re-allocate the pipe(s) with the allocation being reduced and not - * overlapping with a previous light-up pipe (another way to put it is: - * pipes with their new allocation strickly included into their old ones). - * 2. re-allocate the other pipes that get their allocation reduced - * 3. allocate the pipes having their allocation increased - * - * Steps 1. and 2. are here to take care of the following case: - * - Initially DDB looks like this: - * | B | C | - * - enable pipe A. - * - pipe B has a reduced DDB allocation that overlaps with the old pipe C - * allocation - * | A | B | C | - * - * We need to sequence the re-allocation: C, B, A (and not B, C, A). - */ + skl_ddb_entry_write(dev_priv, PLANE_BUF_CFG(pipe, plane), + &wm->ddb.plane[pipe][plane]); + skl_ddb_entry_write(dev_priv, PLANE_NV12_BUF_CFG(pipe, plane), + &wm->ddb.y_plane[pipe][plane]); +} -static void -skl_wm_flush_pipe(struct drm_i915_private *dev_priv, enum pipe pipe, int pass) +void skl_write_cursor_wm(struct intel_crtc *intel_crtc, + const struct skl_wm_values *wm) { - int plane; - - DRM_DEBUG_KMS("flush pipe %c (pass %d)\n", pipe_name(pipe), pass); + struct drm_crtc *crtc = &intel_crtc->base; + struct drm_device *dev = crtc->dev; + struct drm_i915_private *dev_priv = to_i915(dev); + int level, max_level = ilk_wm_max_level(dev); + enum pipe pipe = intel_crtc->pipe; - for_each_plane(dev_priv, pipe, plane) { - I915_WRITE(PLANE_SURF(pipe, plane), - I915_READ(PLANE_SURF(pipe, plane))); + for (level = 0; level <= max_level; level++) { + I915_WRITE(CUR_WM(pipe, level), + wm->plane[pipe][PLANE_CURSOR][level]); } - I915_WRITE(CURBASE(pipe), I915_READ(CURBASE(pipe))); + I915_WRITE(CUR_WM_TRANS(pipe), wm->plane_trans[pipe][PLANE_CURSOR]); + + skl_ddb_entry_write(dev_priv, CUR_BUF_CFG(pipe), + &wm->ddb.plane[pipe][PLANE_CURSOR]); } -static bool -skl_ddb_allocation_included(const struct skl_ddb_allocation *old, - const struct skl_ddb_allocation *new, - enum pipe pipe) +bool skl_ddb_allocation_equals(const struct skl_ddb_allocation *old, + const struct skl_ddb_allocation *new, + enum pipe pipe) { - uint16_t old_size, new_size; - - old_size = skl_ddb_entry_size(&old->pipe[pipe]); - new_size = skl_ddb_entry_size(&new->pipe[pipe]); - - return old_size != new_size && - new->pipe[pipe].start >= old->pipe[pipe].start && - new->pipe[pipe].end <= old->pipe[pipe].end; + return new->pipe[pipe].start == old->pipe[pipe].start && + new->pipe[pipe].end == old->pipe[pipe].end; } -static void skl_flush_wm_values(struct drm_i915_private *dev_priv, - struct skl_wm_values *new_values) +static inline bool skl_ddb_entries_overlap(const struct skl_ddb_entry *a, + const struct skl_ddb_entry *b) { - struct drm_device *dev = &dev_priv->drm; - struct skl_ddb_allocation *cur_ddb, *new_ddb; - bool reallocated[I915_MAX_PIPES] = {}; - struct intel_crtc *crtc; - enum pipe pipe; - - new_ddb = &new_values->ddb; - cur_ddb = &dev_priv->wm.skl_hw.ddb; - - /* - * First pass: flush the pipes with the new allocation contained into - * the old space. - * - * We'll wait for the vblank on those pipes to ensure we can safely - * re-allocate the freed space without this pipe fetching from it. - */ - for_each_intel_crtc(dev, crtc) { - if (!crtc->active) - continue; - - pipe = crtc->pipe; - - if (!skl_ddb_allocation_included(cur_ddb, new_ddb, pipe)) - continue; - - skl_wm_flush_pipe(dev_priv, pipe, 1); - intel_wait_for_vblank(dev, pipe); - - reallocated[pipe] = true; - } - + return a->start < b->end && b->start < a->end; +} - /* - * Second pass: flush the pipes that are having their allocation - * reduced, but overlapping with a previous allocation. - * - * Here as well we need to wait for the vblank to make sure the freed - * space is not used anymore. - */ - for_each_intel_crtc(dev, crtc) { - if (!crtc->active) - continue; +bool skl_ddb_allocation_overlaps(struct drm_atomic_state *state, + const struct skl_ddb_allocation *old, + const struct skl_ddb_allocation *new, + enum pipe pipe) +{ + struct drm_device *dev = state->dev; + struct intel_crtc *intel_crtc; + enum pipe otherp; - pipe = crtc->pipe; + for_each_intel_crtc(dev, intel_crtc) { + otherp = intel_crtc->pipe; - if (reallocated[pipe]) + if (otherp == pipe) continue; - if (skl_ddb_entry_size(&new_ddb->pipe[pipe]) < - skl_ddb_entry_size(&cur_ddb->pipe[pipe])) { - skl_wm_flush_pipe(dev_priv, pipe, 2); - intel_wait_for_vblank(dev, pipe); - reallocated[pipe] = true; - } + if (skl_ddb_entries_overlap(&new->pipe[pipe], + &old->pipe[otherp])) + return true; } - /* - * Third pass: flush the pipes that got more space allocated. - * - * We don't need to actively wait for the update here, next vblank - * will just get more DDB space with the correct WM values. - */ - for_each_intel_crtc(dev, crtc) { - if (!crtc->active) - continue; - - pipe = crtc->pipe; - - /* - * At this point, only the pipes more space than before are - * left to re-allocate. - */ - if (reallocated[pipe]) - continue; - - skl_wm_flush_pipe(dev_priv, pipe, 3); - } + return false; } static int skl_update_pipe_wm(struct drm_crtc_state *cstate, @@ -4041,6 +3962,41 @@ pipes_modified(struct drm_atomic_state *state) return ret; } +int +skl_ddb_add_affected_planes(struct intel_crtc_state *cstate) +{ + struct drm_atomic_state *state = cstate->base.state; + struct drm_device *dev = state->dev; + struct drm_crtc *crtc = cstate->base.crtc; + struct intel_crtc *intel_crtc = to_intel_crtc(crtc); + struct drm_i915_private *dev_priv = to_i915(dev); + struct intel_atomic_state *intel_state = to_intel_atomic_state(state); + struct skl_ddb_allocation *new_ddb = &intel_state->wm_results.ddb; + struct skl_ddb_allocation *cur_ddb = &dev_priv->wm.skl_hw.ddb; + struct drm_plane_state *plane_state; + struct drm_plane *plane; + enum pipe pipe = intel_crtc->pipe; + int id; + + WARN_ON(!drm_atomic_get_existing_crtc_state(state, crtc)); + + drm_for_each_plane_mask(plane, dev, crtc->state->plane_mask) { + id = skl_wm_plane_id(to_intel_plane(plane)); + + if (skl_ddb_entry_equal(&cur_ddb->plane[pipe][id], + &new_ddb->plane[pipe][id]) && + skl_ddb_entry_equal(&cur_ddb->y_plane[pipe][id], + &new_ddb->y_plane[pipe][id])) + continue; + + plane_state = drm_atomic_get_plane_state(state, plane); + if (IS_ERR(plane_state)) + return PTR_ERR(plane_state); + } + + return 0; +} + static int skl_compute_ddb(struct drm_atomic_state *state) { @@ -4105,7 +4061,7 @@ skl_compute_ddb(struct drm_atomic_state *state) if (ret) return ret; - ret = drm_atomic_add_affected_planes(state, &intel_crtc->base); + ret = skl_ddb_add_affected_planes(cstate); if (ret) return ret; } @@ -4206,7 +4162,7 @@ static void skl_update_wm(struct drm_crtc *crtc) struct skl_wm_values *hw_vals = &dev_priv->wm.skl_hw; struct intel_crtc_state *cstate = to_intel_crtc_state(crtc->state); struct skl_pipe_wm *pipe_wm = &cstate->wm.skl.optimal; - int pipe; + enum pipe pipe = intel_crtc->pipe; if ((results->dirty_pipes & drm_crtc_mask(crtc)) == 0) return; @@ -4215,15 +4171,22 @@ static void skl_update_wm(struct drm_crtc *crtc) mutex_lock(&dev_priv->wm.wm_mutex); - skl_write_wm_values(dev_priv, results); - skl_flush_wm_values(dev_priv, results); - /* - * Store the new configuration (but only for the pipes that have - * changed; the other values weren't recomputed). + * If this pipe isn't active already, we're going to be enabling it + * very soon. Since it's safe to update a pipe's ddb allocation while + * the pipe's shut off, just do so here. Already active pipes will have + * their watermarks updated once we update their planes. */ - for_each_pipe_masked(dev_priv, pipe, results->dirty_pipes) - skl_copy_wm_for_pipe(hw_vals, results, pipe); + if (crtc->state->active_changed) { + int plane; + + for (plane = 0; plane < intel_num_planes(intel_crtc); plane++) + skl_write_plane_wm(intel_crtc, results, plane); + + skl_write_cursor_wm(intel_crtc, results); + } + + skl_copy_wm_for_pipe(hw_vals, results, pipe); mutex_unlock(&dev_priv->wm.wm_mutex); } @@ -5103,7 +5066,7 @@ void gen6_rps_boost(struct drm_i915_private *dev_priv, */ if (!(dev_priv->gt.awake && dev_priv->rps.enabled && - dev_priv->rps.cur_freq < dev_priv->rps.max_freq_softlimit)) + dev_priv->rps.cur_freq < dev_priv->rps.boost_freq)) return; /* Force a RPS boost (and don't count it against the client) if @@ -5294,35 +5257,31 @@ int sanitize_rc6_option(struct drm_i915_private *dev_priv, int enable_rc6) static void gen6_init_rps_frequencies(struct drm_i915_private *dev_priv) { - uint32_t rp_state_cap; - u32 ddcc_status = 0; - int ret; - /* All of these values are in units of 50MHz */ - dev_priv->rps.cur_freq = 0; + /* static values from HW: RP0 > RP1 > RPn (min_freq) */ if (IS_BROXTON(dev_priv)) { - rp_state_cap = I915_READ(BXT_RP_STATE_CAP); + u32 rp_state_cap = I915_READ(BXT_RP_STATE_CAP); dev_priv->rps.rp0_freq = (rp_state_cap >> 16) & 0xff; dev_priv->rps.rp1_freq = (rp_state_cap >> 8) & 0xff; dev_priv->rps.min_freq = (rp_state_cap >> 0) & 0xff; } else { - rp_state_cap = I915_READ(GEN6_RP_STATE_CAP); + u32 rp_state_cap = I915_READ(GEN6_RP_STATE_CAP); dev_priv->rps.rp0_freq = (rp_state_cap >> 0) & 0xff; dev_priv->rps.rp1_freq = (rp_state_cap >> 8) & 0xff; dev_priv->rps.min_freq = (rp_state_cap >> 16) & 0xff; } - /* hw_max = RP0 until we check for overclocking */ - dev_priv->rps.max_freq = dev_priv->rps.rp0_freq; + dev_priv->rps.max_freq = dev_priv->rps.rp0_freq; dev_priv->rps.efficient_freq = dev_priv->rps.rp1_freq; if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv) || IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv)) { - ret = sandybridge_pcode_read(dev_priv, - HSW_PCODE_DYNAMIC_DUTY_CYCLE_CONTROL, - &ddcc_status); - if (0 == ret) + u32 ddcc_status = 0; + + if (sandybridge_pcode_read(dev_priv, + HSW_PCODE_DYNAMIC_DUTY_CYCLE_CONTROL, + &ddcc_status) == 0) dev_priv->rps.efficient_freq = clamp_t(u8, ((ddcc_status >> 8) & 0xff), @@ -5332,29 +5291,26 @@ static void gen6_init_rps_frequencies(struct drm_i915_private *dev_priv) if (IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv)) { /* Store the frequency values in 16.66 MHZ units, which is - the natural hardware unit for SKL */ + * the natural hardware unit for SKL + */ dev_priv->rps.rp0_freq *= GEN9_FREQ_SCALER; dev_priv->rps.rp1_freq *= GEN9_FREQ_SCALER; dev_priv->rps.min_freq *= GEN9_FREQ_SCALER; dev_priv->rps.max_freq *= GEN9_FREQ_SCALER; dev_priv->rps.efficient_freq *= GEN9_FREQ_SCALER; } +} - dev_priv->rps.idle_freq = dev_priv->rps.min_freq; +static void reset_rps(struct drm_i915_private *dev_priv, + void (*set)(struct drm_i915_private *, u8)) +{ + u8 freq = dev_priv->rps.cur_freq; - /* Preserve min/max settings in case of re-init */ - if (dev_priv->rps.max_freq_softlimit == 0) - dev_priv->rps.max_freq_softlimit = dev_priv->rps.max_freq; + /* force a reset */ + dev_priv->rps.power = -1; + dev_priv->rps.cur_freq = -1; - if (dev_priv->rps.min_freq_softlimit == 0) { - if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv)) - dev_priv->rps.min_freq_softlimit = - max_t(int, dev_priv->rps.efficient_freq, - intel_freq_opcode(dev_priv, 450)); - else - dev_priv->rps.min_freq_softlimit = - dev_priv->rps.min_freq; - } + set(dev_priv, freq); } /* See the Gen9_GT_PM_Programming_Guide doc for the below */ @@ -5362,8 +5318,6 @@ static void gen9_enable_rps(struct drm_i915_private *dev_priv) { intel_uncore_forcewake_get(dev_priv, FORCEWAKE_ALL); - gen6_init_rps_frequencies(dev_priv); - /* WaGsvDisableTurbo: Workaround to disable turbo on BXT A* */ if (IS_BXT_REVID(dev_priv, 0, BXT_REVID_A1)) { /* @@ -5393,8 +5347,7 @@ static void gen9_enable_rps(struct drm_i915_private *dev_priv) /* Leaning on the below call to gen6_set_rps to program/setup the * Up/Down EI & threshold registers, as well as the RP_CONTROL, * RP_INTERRUPT_LIMITS & RPNSWREQ registers */ - dev_priv->rps.power = HIGH_POWER; /* force a reset */ - gen6_set_rps(dev_priv, dev_priv->rps.idle_freq); + reset_rps(dev_priv, gen6_set_rps); intel_uncore_forcewake_put(dev_priv, FORCEWAKE_ALL); } @@ -5481,9 +5434,6 @@ static void gen8_enable_rps(struct drm_i915_private *dev_priv) /* 2a: Disable RC states. */ I915_WRITE(GEN6_RC_CONTROL, 0); - /* Initialize rps frequencies */ - gen6_init_rps_frequencies(dev_priv); - /* 2b: Program RC6 thresholds.*/ I915_WRITE(GEN6_RC6_WAKE_RATE_LIMIT, 40 << 16); I915_WRITE(GEN6_RC_EVALUATION_INTERVAL, 125000); /* 12500 * 1280ns */ @@ -5540,8 +5490,7 @@ static void gen8_enable_rps(struct drm_i915_private *dev_priv) /* 6: Ring frequency + overclocking (our driver does this later */ - dev_priv->rps.power = HIGH_POWER; /* force a reset */ - gen6_set_rps(dev_priv, dev_priv->rps.idle_freq); + reset_rps(dev_priv, gen6_set_rps); intel_uncore_forcewake_put(dev_priv, FORCEWAKE_ALL); } @@ -5549,7 +5498,7 @@ static void gen8_enable_rps(struct drm_i915_private *dev_priv) static void gen6_enable_rps(struct drm_i915_private *dev_priv) { struct intel_engine_cs *engine; - u32 rc6vids, pcu_mbox = 0, rc6_mask = 0; + u32 rc6vids, rc6_mask = 0; u32 gtfifodbg; int rc6_mode; int ret; @@ -5573,9 +5522,6 @@ static void gen6_enable_rps(struct drm_i915_private *dev_priv) intel_uncore_forcewake_get(dev_priv, FORCEWAKE_ALL); - /* Initialize rps frequencies */ - gen6_init_rps_frequencies(dev_priv); - /* disable the counters and set deterministic thresholds */ I915_WRITE(GEN6_RC_CONTROL, 0); @@ -5626,16 +5572,7 @@ static void gen6_enable_rps(struct drm_i915_private *dev_priv) if (ret) DRM_DEBUG_DRIVER("Failed to set the min frequency\n"); - ret = sandybridge_pcode_read(dev_priv, GEN6_READ_OC_PARAMS, &pcu_mbox); - if (!ret && (pcu_mbox & (1<<31))) { /* OC supported */ - DRM_DEBUG_DRIVER("Overclocking supported. Max: %dMHz, Overclock max: %dMHz\n", - (dev_priv->rps.max_freq_softlimit & 0xff) * 50, - (pcu_mbox & 0xff) * 50); - dev_priv->rps.max_freq = pcu_mbox & 0xff; - } - - dev_priv->rps.power = HIGH_POWER; /* force a reset */ - gen6_set_rps(dev_priv, dev_priv->rps.idle_freq); + reset_rps(dev_priv, gen6_set_rps); rc6vids = 0; ret = sandybridge_pcode_read(dev_priv, GEN6_PCODE_READ_RC6VIDS, &rc6vids); @@ -5654,7 +5591,7 @@ static void gen6_enable_rps(struct drm_i915_private *dev_priv) intel_uncore_forcewake_put(dev_priv, FORCEWAKE_ALL); } -static void __gen6_update_ring_freq(struct drm_i915_private *dev_priv) +static void gen6_update_ring_freq(struct drm_i915_private *dev_priv) { int min_freq = 15; unsigned int gpu_freq; @@ -5738,23 +5675,13 @@ static void __gen6_update_ring_freq(struct drm_i915_private *dev_priv) } } -void gen6_update_ring_freq(struct drm_i915_private *dev_priv) -{ - if (!HAS_CORE_RING_FREQ(dev_priv)) - return; - - mutex_lock(&dev_priv->rps.hw_lock); - __gen6_update_ring_freq(dev_priv); - mutex_unlock(&dev_priv->rps.hw_lock); -} - static int cherryview_rps_max_freq(struct drm_i915_private *dev_priv) { u32 val, rp0; val = vlv_punit_read(dev_priv, FB_GFX_FMAX_AT_VMAX_FUSE); - switch (INTEL_INFO(dev_priv)->eu_total) { + switch (INTEL_INFO(dev_priv)->sseu.eu_total) { case 8: /* (2 * 4) config */ rp0 = (val >> FB_GFX_FMAX_AT_VMAX_2SS4EU_FUSE_SHIFT); @@ -5892,8 +5819,6 @@ static void valleyview_setup_pctx(struct drm_i915_private *dev_priv) u32 pcbr; int pctx_size = 24*1024; - mutex_lock(&dev_priv->drm.struct_mutex); - pcbr = I915_READ(VLV_PCBR); if (pcbr) { /* BIOS set it up already, grab the pre-alloc'd space */ @@ -5929,7 +5854,6 @@ static void valleyview_setup_pctx(struct drm_i915_private *dev_priv) out: DRM_DEBUG_DRIVER("PCBR: 0x%08x\n", I915_READ(VLV_PCBR)); dev_priv->vlv_pctx = pctx; - mutex_unlock(&dev_priv->drm.struct_mutex); } static void valleyview_cleanup_pctx(struct drm_i915_private *dev_priv) @@ -5937,7 +5861,7 @@ static void valleyview_cleanup_pctx(struct drm_i915_private *dev_priv) if (WARN_ON(!dev_priv->vlv_pctx)) return; - drm_gem_object_unreference_unlocked(&dev_priv->vlv_pctx->base); + i915_gem_object_put_unlocked(dev_priv->vlv_pctx); dev_priv->vlv_pctx = NULL; } @@ -5960,8 +5884,6 @@ static void valleyview_init_gt_powersave(struct drm_i915_private *dev_priv) vlv_init_gpll_ref_freq(dev_priv); - mutex_lock(&dev_priv->rps.hw_lock); - val = vlv_punit_read(dev_priv, PUNIT_REG_GPU_FREQ_STS); switch ((val >> 6) & 3) { case 0: @@ -5997,17 +5919,6 @@ static void valleyview_init_gt_powersave(struct drm_i915_private *dev_priv) DRM_DEBUG_DRIVER("min GPU freq: %d MHz (%u)\n", intel_gpu_freq(dev_priv, dev_priv->rps.min_freq), dev_priv->rps.min_freq); - - dev_priv->rps.idle_freq = dev_priv->rps.min_freq; - - /* Preserve min/max settings in case of re-init */ - if (dev_priv->rps.max_freq_softlimit == 0) - dev_priv->rps.max_freq_softlimit = dev_priv->rps.max_freq; - - if (dev_priv->rps.min_freq_softlimit == 0) - dev_priv->rps.min_freq_softlimit = dev_priv->rps.min_freq; - - mutex_unlock(&dev_priv->rps.hw_lock); } static void cherryview_init_gt_powersave(struct drm_i915_private *dev_priv) @@ -6018,8 +5929,6 @@ static void cherryview_init_gt_powersave(struct drm_i915_private *dev_priv) vlv_init_gpll_ref_freq(dev_priv); - mutex_lock(&dev_priv->rps.hw_lock); - mutex_lock(&dev_priv->sb_lock); val = vlv_cck_read(dev_priv, CCK_FUSE_REG); mutex_unlock(&dev_priv->sb_lock); @@ -6061,17 +5970,6 @@ static void cherryview_init_gt_powersave(struct drm_i915_private *dev_priv) dev_priv->rps.rp1_freq | dev_priv->rps.min_freq) & 1, "Odd GPU freq values\n"); - - dev_priv->rps.idle_freq = dev_priv->rps.min_freq; - - /* Preserve min/max settings in case of re-init */ - if (dev_priv->rps.max_freq_softlimit == 0) - dev_priv->rps.max_freq_softlimit = dev_priv->rps.max_freq; - - if (dev_priv->rps.min_freq_softlimit == 0) - dev_priv->rps.min_freq_softlimit = dev_priv->rps.min_freq; - - mutex_unlock(&dev_priv->rps.hw_lock); } static void valleyview_cleanup_gt_powersave(struct drm_i915_private *dev_priv) @@ -6162,16 +6060,7 @@ static void cherryview_enable_rps(struct drm_i915_private *dev_priv) DRM_DEBUG_DRIVER("GPLL enabled? %s\n", yesno(val & GPLLENABLE)); DRM_DEBUG_DRIVER("GPU status: 0x%08x\n", val); - dev_priv->rps.cur_freq = (val >> 8) & 0xff; - DRM_DEBUG_DRIVER("current GPU freq: %d MHz (%u)\n", - intel_gpu_freq(dev_priv, dev_priv->rps.cur_freq), - dev_priv->rps.cur_freq); - - DRM_DEBUG_DRIVER("setting GPU freq to %d MHz (%u)\n", - intel_gpu_freq(dev_priv, dev_priv->rps.idle_freq), - dev_priv->rps.idle_freq); - - valleyview_set_rps(dev_priv, dev_priv->rps.idle_freq); + reset_rps(dev_priv, valleyview_set_rps); intel_uncore_forcewake_put(dev_priv, FORCEWAKE_ALL); } @@ -6251,16 +6140,7 @@ static void valleyview_enable_rps(struct drm_i915_private *dev_priv) DRM_DEBUG_DRIVER("GPLL enabled? %s\n", yesno(val & GPLLENABLE)); DRM_DEBUG_DRIVER("GPU status: 0x%08x\n", val); - dev_priv->rps.cur_freq = (val >> 8) & 0xff; - DRM_DEBUG_DRIVER("current GPU freq: %d MHz (%u)\n", - intel_gpu_freq(dev_priv, dev_priv->rps.cur_freq), - dev_priv->rps.cur_freq); - - DRM_DEBUG_DRIVER("setting GPU freq to %d MHz (%u)\n", - intel_gpu_freq(dev_priv, dev_priv->rps.idle_freq), - dev_priv->rps.idle_freq); - - valleyview_set_rps(dev_priv, dev_priv->rps.idle_freq); + reset_rps(dev_priv, valleyview_set_rps); intel_uncore_forcewake_put(dev_priv, FORCEWAKE_ALL); } @@ -6589,19 +6469,11 @@ EXPORT_SYMBOL_GPL(i915_gpu_lower); */ bool i915_gpu_busy(void) { - struct drm_i915_private *dev_priv; - struct intel_engine_cs *engine; bool ret = false; spin_lock_irq(&mchdev_lock); - if (!i915_mch_dev) - goto out_unlock; - dev_priv = i915_mch_dev; - - for_each_engine(engine, dev_priv) - ret |= !list_empty(&engine->request_list); - -out_unlock: + if (i915_mch_dev) + ret = i915_mch_dev->gt.awake; spin_unlock_irq(&mchdev_lock); return ret; @@ -6757,10 +6629,51 @@ void intel_init_gt_powersave(struct drm_i915_private *dev_priv) intel_runtime_pm_get(dev_priv); } + mutex_lock(&dev_priv->drm.struct_mutex); + mutex_lock(&dev_priv->rps.hw_lock); + + /* Initialize RPS limits (for userspace) */ if (IS_CHERRYVIEW(dev_priv)) cherryview_init_gt_powersave(dev_priv); else if (IS_VALLEYVIEW(dev_priv)) valleyview_init_gt_powersave(dev_priv); + else if (INTEL_GEN(dev_priv) >= 6) + gen6_init_rps_frequencies(dev_priv); + + /* Derive initial user preferences/limits from the hardware limits */ + dev_priv->rps.idle_freq = dev_priv->rps.min_freq; + dev_priv->rps.cur_freq = dev_priv->rps.idle_freq; + + dev_priv->rps.max_freq_softlimit = dev_priv->rps.max_freq; + dev_priv->rps.min_freq_softlimit = dev_priv->rps.min_freq; + + if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv)) + dev_priv->rps.min_freq_softlimit = + max_t(int, + dev_priv->rps.efficient_freq, + intel_freq_opcode(dev_priv, 450)); + + /* After setting max-softlimit, find the overclock max freq */ + if (IS_GEN6(dev_priv) || + IS_IVYBRIDGE(dev_priv) || IS_HASWELL(dev_priv)) { + u32 params = 0; + + sandybridge_pcode_read(dev_priv, GEN6_READ_OC_PARAMS, ¶ms); + if (params & BIT(31)) { /* OC supported */ + DRM_DEBUG_DRIVER("Overclocking supported, max: %dMHz, overclock: %dMHz\n", + (dev_priv->rps.max_freq & 0xff) * 50, + (params & 0xff) * 50); + dev_priv->rps.max_freq = params & 0xff; + } + } + + /* Finally allow us to boost to max by default */ + dev_priv->rps.boost_freq = dev_priv->rps.max_freq; + + mutex_unlock(&dev_priv->rps.hw_lock); + mutex_unlock(&dev_priv->drm.struct_mutex); + + intel_autoenable_gt_powersave(dev_priv); } void intel_cleanup_gt_powersave(struct drm_i915_private *dev_priv) @@ -6772,13 +6685,6 @@ void intel_cleanup_gt_powersave(struct drm_i915_private *dev_priv) intel_runtime_pm_put(dev_priv); } -static void gen6_suspend_rps(struct drm_i915_private *dev_priv) -{ - flush_delayed_work(&dev_priv->rps.delayed_resume_work); - - gen6_disable_rps_interrupts(dev_priv); -} - /** * intel_suspend_gt_powersave - suspend PM work and helper threads * @dev_priv: i915 device @@ -6792,60 +6698,76 @@ void intel_suspend_gt_powersave(struct drm_i915_private *dev_priv) if (INTEL_GEN(dev_priv) < 6) return; - gen6_suspend_rps(dev_priv); + if (cancel_delayed_work_sync(&dev_priv->rps.autoenable_work)) + intel_runtime_pm_put(dev_priv); + + /* gen6_rps_idle() will be called later to disable interrupts */ +} + +void intel_sanitize_gt_powersave(struct drm_i915_private *dev_priv) +{ + dev_priv->rps.enabled = true; /* force disabling */ + intel_disable_gt_powersave(dev_priv); - /* Force GPU to min freq during suspend */ - gen6_rps_idle(dev_priv); + gen6_reset_rps_interrupts(dev_priv); } void intel_disable_gt_powersave(struct drm_i915_private *dev_priv) { - if (IS_IRONLAKE_M(dev_priv)) { - ironlake_disable_drps(dev_priv); - } else if (INTEL_INFO(dev_priv)->gen >= 6) { - intel_suspend_gt_powersave(dev_priv); + if (!READ_ONCE(dev_priv->rps.enabled)) + return; - mutex_lock(&dev_priv->rps.hw_lock); - if (INTEL_INFO(dev_priv)->gen >= 9) { - gen9_disable_rc6(dev_priv); - gen9_disable_rps(dev_priv); - } else if (IS_CHERRYVIEW(dev_priv)) - cherryview_disable_rps(dev_priv); - else if (IS_VALLEYVIEW(dev_priv)) - valleyview_disable_rps(dev_priv); - else - gen6_disable_rps(dev_priv); + mutex_lock(&dev_priv->rps.hw_lock); - dev_priv->rps.enabled = false; - mutex_unlock(&dev_priv->rps.hw_lock); + if (INTEL_GEN(dev_priv) >= 9) { + gen9_disable_rc6(dev_priv); + gen9_disable_rps(dev_priv); + } else if (IS_CHERRYVIEW(dev_priv)) { + cherryview_disable_rps(dev_priv); + } else if (IS_VALLEYVIEW(dev_priv)) { + valleyview_disable_rps(dev_priv); + } else if (INTEL_GEN(dev_priv) >= 6) { + gen6_disable_rps(dev_priv); + } else if (IS_IRONLAKE_M(dev_priv)) { + ironlake_disable_drps(dev_priv); } + + dev_priv->rps.enabled = false; + mutex_unlock(&dev_priv->rps.hw_lock); } -static void intel_gen6_powersave_work(struct work_struct *work) +void intel_enable_gt_powersave(struct drm_i915_private *dev_priv) { - struct drm_i915_private *dev_priv = - container_of(work, struct drm_i915_private, - rps.delayed_resume_work.work); + /* We shouldn't be disabling as we submit, so this should be less + * racy than it appears! + */ + if (READ_ONCE(dev_priv->rps.enabled)) + return; - mutex_lock(&dev_priv->rps.hw_lock); + /* Powersaving is controlled by the host when inside a VM */ + if (intel_vgpu_active(dev_priv)) + return; - gen6_reset_rps_interrupts(dev_priv); + mutex_lock(&dev_priv->rps.hw_lock); if (IS_CHERRYVIEW(dev_priv)) { cherryview_enable_rps(dev_priv); } else if (IS_VALLEYVIEW(dev_priv)) { valleyview_enable_rps(dev_priv); - } else if (INTEL_INFO(dev_priv)->gen >= 9) { + } else if (INTEL_GEN(dev_priv) >= 9) { gen9_enable_rc6(dev_priv); gen9_enable_rps(dev_priv); if (IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv)) - __gen6_update_ring_freq(dev_priv); + gen6_update_ring_freq(dev_priv); } else if (IS_BROADWELL(dev_priv)) { gen8_enable_rps(dev_priv); - __gen6_update_ring_freq(dev_priv); - } else { + gen6_update_ring_freq(dev_priv); + } else if (INTEL_GEN(dev_priv) >= 6) { gen6_enable_rps(dev_priv); - __gen6_update_ring_freq(dev_priv); + gen6_update_ring_freq(dev_priv); + } else if (IS_IRONLAKE_M(dev_priv)) { + ironlake_enable_drps(dev_priv); + intel_init_emon(dev_priv); } WARN_ON(dev_priv->rps.max_freq < dev_priv->rps.min_freq); @@ -6855,25 +6777,52 @@ static void intel_gen6_powersave_work(struct work_struct *work) WARN_ON(dev_priv->rps.efficient_freq > dev_priv->rps.max_freq); dev_priv->rps.enabled = true; + mutex_unlock(&dev_priv->rps.hw_lock); +} + +static void __intel_autoenable_gt_powersave(struct work_struct *work) +{ + struct drm_i915_private *dev_priv = + container_of(work, typeof(*dev_priv), rps.autoenable_work.work); + struct intel_engine_cs *rcs; + struct drm_i915_gem_request *req; - gen6_enable_rps_interrupts(dev_priv); + if (READ_ONCE(dev_priv->rps.enabled)) + goto out; - mutex_unlock(&dev_priv->rps.hw_lock); + rcs = &dev_priv->engine[RCS]; + if (rcs->last_context) + goto out; + if (!rcs->init_context) + goto out; + + mutex_lock(&dev_priv->drm.struct_mutex); + + req = i915_gem_request_alloc(rcs, dev_priv->kernel_context); + if (IS_ERR(req)) + goto unlock; + + if (!i915.enable_execlists && i915_switch_context(req) == 0) + rcs->init_context(req); + + /* Mark the device busy, calling intel_enable_gt_powersave() */ + i915_add_request_no_flush(req); + +unlock: + mutex_unlock(&dev_priv->drm.struct_mutex); +out: intel_runtime_pm_put(dev_priv); } -void intel_enable_gt_powersave(struct drm_i915_private *dev_priv) +void intel_autoenable_gt_powersave(struct drm_i915_private *dev_priv) { - /* Powersaving is controlled by the host when inside a VM */ - if (intel_vgpu_active(dev_priv)) + if (READ_ONCE(dev_priv->rps.enabled)) return; if (IS_IRONLAKE_M(dev_priv)) { ironlake_enable_drps(dev_priv); - mutex_lock(&dev_priv->drm.struct_mutex); intel_init_emon(dev_priv); - mutex_unlock(&dev_priv->drm.struct_mutex); } else if (INTEL_INFO(dev_priv)->gen >= 6) { /* * PCU communication is slow and this doesn't need to be @@ -6887,21 +6836,13 @@ void intel_enable_gt_powersave(struct drm_i915_private *dev_priv) * paths, so the _noresume version is enough (and in case of * runtime resume it's necessary). */ - if (schedule_delayed_work(&dev_priv->rps.delayed_resume_work, - round_jiffies_up_relative(HZ))) + if (queue_delayed_work(dev_priv->wq, + &dev_priv->rps.autoenable_work, + round_jiffies_up_relative(HZ))) intel_runtime_pm_get_noresume(dev_priv); } } -void intel_reset_gt_powersave(struct drm_i915_private *dev_priv) -{ - if (INTEL_INFO(dev_priv)->gen < 6) - return; - - gen6_suspend_rps(dev_priv); - dev_priv->rps.enabled = false; -} - static void ibx_init_clock_gating(struct drm_device *dev) { struct drm_i915_private *dev_priv = to_i915(dev); @@ -8046,7 +7987,7 @@ static void __intel_rps_boost_work(struct work_struct *work) if (!i915_gem_request_completed(req)) gen6_rps_boost(req->i915, NULL, req->emitted_jiffies); - i915_gem_request_unreference(req); + i915_gem_request_put(req); kfree(boost); } @@ -8064,8 +8005,7 @@ void intel_queue_rps_boost_for_request(struct drm_i915_gem_request *req) if (boost == NULL) return; - i915_gem_request_reference(req); - boost->req = req; + boost->req = i915_gem_request_get(req); INIT_WORK(&boost->work, __intel_rps_boost_work); queue_work(req->i915->wq, &boost->work); @@ -8078,11 +8018,9 @@ void intel_pm_setup(struct drm_device *dev) mutex_init(&dev_priv->rps.hw_lock); spin_lock_init(&dev_priv->rps.client_lock); - INIT_DELAYED_WORK(&dev_priv->rps.delayed_resume_work, - intel_gen6_powersave_work); + INIT_DELAYED_WORK(&dev_priv->rps.autoenable_work, + __intel_autoenable_gt_powersave); INIT_LIST_HEAD(&dev_priv->rps.clients); - INIT_LIST_HEAD(&dev_priv->rps.semaphores.link); - INIT_LIST_HEAD(&dev_priv->rps.mmioflips.link); dev_priv->pm.suspended = false; atomic_set(&dev_priv->pm.wakeref_count, 0); |