diff options
Diffstat (limited to 'drivers/gpu/drm/i915/intel_pm.c')
| -rw-r--r-- | drivers/gpu/drm/i915/intel_pm.c | 235 |
1 files changed, 159 insertions, 76 deletions
diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c index 3907c60a72de..008ec0bb017f 100644 --- a/drivers/gpu/drm/i915/intel_pm.c +++ b/drivers/gpu/drm/i915/intel_pm.c @@ -370,7 +370,7 @@ static void intel_enable_fbc(struct drm_crtc *crtc, unsigned long interval) intel_cancel_fbc_work(dev_priv); - work = kzalloc(sizeof *work, GFP_KERNEL); + work = kzalloc(sizeof(*work), GFP_KERNEL); if (work == NULL) { DRM_ERROR("Failed to allocate FBC work structure\n"); dev_priv->display.enable_fbc(crtc, interval); @@ -1100,8 +1100,12 @@ static void pineview_update_wm(struct drm_crtc *unused_crtc) crtc = single_enabled_crtc(dev); if (crtc) { - int clock = to_intel_crtc(crtc)->config.adjusted_mode.clock; + const struct drm_display_mode *adjusted_mode; int pixel_size = crtc->fb->bits_per_pixel / 8; + int clock; + + adjusted_mode = &to_intel_crtc(crtc)->config.adjusted_mode; + clock = adjusted_mode->crtc_clock; /* Display SR */ wm = intel_calculate_wm(clock, &pineview_display_wm, @@ -1174,7 +1178,7 @@ static bool g4x_compute_wm0(struct drm_device *dev, } adjusted_mode = &to_intel_crtc(crtc)->config.adjusted_mode; - clock = adjusted_mode->clock; + clock = adjusted_mode->crtc_clock; htotal = adjusted_mode->htotal; hdisplay = to_intel_crtc(crtc)->config.pipe_src_w; pixel_size = crtc->fb->bits_per_pixel / 8; @@ -1261,7 +1265,7 @@ static bool g4x_compute_srwm(struct drm_device *dev, crtc = intel_get_crtc_for_plane(dev, plane); adjusted_mode = &to_intel_crtc(crtc)->config.adjusted_mode; - clock = adjusted_mode->clock; + clock = adjusted_mode->crtc_clock; htotal = adjusted_mode->htotal; hdisplay = to_intel_crtc(crtc)->config.pipe_src_w; pixel_size = crtc->fb->bits_per_pixel / 8; @@ -1302,7 +1306,7 @@ static bool vlv_compute_drain_latency(struct drm_device *dev, if (!intel_crtc_active(crtc)) return false; - clock = to_intel_crtc(crtc)->config.adjusted_mode.clock; + clock = to_intel_crtc(crtc)->config.adjusted_mode.crtc_clock; pixel_size = crtc->fb->bits_per_pixel / 8; /* BPP */ entries = (clock / 1000) * pixel_size; @@ -1492,7 +1496,7 @@ static void i965_update_wm(struct drm_crtc *unused_crtc) static const int sr_latency_ns = 12000; const struct drm_display_mode *adjusted_mode = &to_intel_crtc(crtc)->config.adjusted_mode; - int clock = adjusted_mode->clock; + int clock = adjusted_mode->crtc_clock; int htotal = adjusted_mode->htotal; int hdisplay = to_intel_crtc(crtc)->config.pipe_src_w; int pixel_size = crtc->fb->bits_per_pixel / 8; @@ -1567,11 +1571,13 @@ static void i9xx_update_wm(struct drm_crtc *unused_crtc) fifo_size = dev_priv->display.get_fifo_size(dev, 0); crtc = intel_get_crtc_for_plane(dev, 0); if (intel_crtc_active(crtc)) { + const struct drm_display_mode *adjusted_mode; int cpp = crtc->fb->bits_per_pixel / 8; if (IS_GEN2(dev)) cpp = 4; - planea_wm = intel_calculate_wm(to_intel_crtc(crtc)->config.adjusted_mode.clock, + adjusted_mode = &to_intel_crtc(crtc)->config.adjusted_mode; + planea_wm = intel_calculate_wm(adjusted_mode->crtc_clock, wm_info, fifo_size, cpp, latency_ns); enabled = crtc; @@ -1581,11 +1587,13 @@ static void i9xx_update_wm(struct drm_crtc *unused_crtc) fifo_size = dev_priv->display.get_fifo_size(dev, 1); crtc = intel_get_crtc_for_plane(dev, 1); if (intel_crtc_active(crtc)) { + const struct drm_display_mode *adjusted_mode; int cpp = crtc->fb->bits_per_pixel / 8; if (IS_GEN2(dev)) cpp = 4; - planeb_wm = intel_calculate_wm(to_intel_crtc(crtc)->config.adjusted_mode.clock, + adjusted_mode = &to_intel_crtc(crtc)->config.adjusted_mode; + planeb_wm = intel_calculate_wm(adjusted_mode->crtc_clock, wm_info, fifo_size, cpp, latency_ns); if (enabled == NULL) @@ -1614,7 +1622,7 @@ static void i9xx_update_wm(struct drm_crtc *unused_crtc) static const int sr_latency_ns = 6000; const struct drm_display_mode *adjusted_mode = &to_intel_crtc(enabled)->config.adjusted_mode; - int clock = adjusted_mode->clock; + int clock = adjusted_mode->crtc_clock; int htotal = adjusted_mode->htotal; int hdisplay = to_intel_crtc(crtc)->config.pipe_src_w; int pixel_size = enabled->fb->bits_per_pixel / 8; @@ -1670,6 +1678,7 @@ static void i830_update_wm(struct drm_crtc *unused_crtc) struct drm_device *dev = unused_crtc->dev; struct drm_i915_private *dev_priv = dev->dev_private; struct drm_crtc *crtc; + const struct drm_display_mode *adjusted_mode; uint32_t fwater_lo; int planea_wm; @@ -1677,7 +1686,8 @@ static void i830_update_wm(struct drm_crtc *unused_crtc) if (crtc == NULL) return; - planea_wm = intel_calculate_wm(to_intel_crtc(crtc)->config.adjusted_mode.clock, + adjusted_mode = &to_intel_crtc(crtc)->config.adjusted_mode; + planea_wm = intel_calculate_wm(adjusted_mode->crtc_clock, &i830_wm_info, dev_priv->display.get_fifo_size(dev, 0), 4, latency_ns); @@ -1764,7 +1774,7 @@ static bool ironlake_compute_srwm(struct drm_device *dev, int level, int plane, crtc = intel_get_crtc_for_plane(dev, plane); adjusted_mode = &to_intel_crtc(crtc)->config.adjusted_mode; - clock = adjusted_mode->clock; + clock = adjusted_mode->crtc_clock; htotal = adjusted_mode->htotal; hdisplay = to_intel_crtc(crtc)->config.pipe_src_w; pixel_size = crtc->fb->bits_per_pixel / 8; @@ -2112,7 +2122,7 @@ static uint32_t ilk_pipe_pixel_rate(struct drm_device *dev, struct intel_crtc *intel_crtc = to_intel_crtc(crtc); uint32_t pixel_rate; - pixel_rate = intel_crtc->config.adjusted_mode.clock; + pixel_rate = intel_crtc->config.adjusted_mode.crtc_clock; /* We only use IF-ID interlacing. If we ever use PF-ID we'll need to * adjust the pixel_rate here. */ @@ -2913,7 +2923,7 @@ sandybridge_compute_sprite_wm(struct drm_device *dev, int plane, return false; } - clock = to_intel_crtc(crtc)->config.adjusted_mode.clock; + clock = to_intel_crtc(crtc)->config.adjusted_mode.crtc_clock; /* Use the small buffer method to calculate the sprite watermark */ entries = ((clock * pixel_size / 1000) * display_latency_ns) / 1000; @@ -2948,7 +2958,7 @@ sandybridge_compute_sprite_srwm(struct drm_device *dev, int plane, } crtc = intel_get_crtc_for_plane(dev, plane); - clock = to_intel_crtc(crtc)->config.adjusted_mode.clock; + clock = to_intel_crtc(crtc)->config.adjusted_mode.crtc_clock; if (!clock) { *sprite_wm = 0; return false; @@ -3302,6 +3312,98 @@ static u32 gen6_rps_limits(struct drm_i915_private *dev_priv, u8 *val) return limits; } +static void gen6_set_rps_thresholds(struct drm_i915_private *dev_priv, u8 val) +{ + int new_power; + + new_power = dev_priv->rps.power; + switch (dev_priv->rps.power) { + case LOW_POWER: + if (val > dev_priv->rps.rpe_delay + 1 && val > dev_priv->rps.cur_delay) + new_power = BETWEEN; + break; + + case BETWEEN: + if (val <= dev_priv->rps.rpe_delay && val < dev_priv->rps.cur_delay) + new_power = LOW_POWER; + else if (val >= dev_priv->rps.rp0_delay && val > dev_priv->rps.cur_delay) + new_power = HIGH_POWER; + break; + + case HIGH_POWER: + if (val < (dev_priv->rps.rp1_delay + dev_priv->rps.rp0_delay) >> 1 && val < dev_priv->rps.cur_delay) + new_power = BETWEEN; + break; + } + /* Max/min bins are special */ + if (val == dev_priv->rps.min_delay) + new_power = LOW_POWER; + if (val == dev_priv->rps.max_delay) + new_power = HIGH_POWER; + if (new_power == dev_priv->rps.power) + return; + + /* Note the units here are not exactly 1us, but 1280ns. */ + switch (new_power) { + case LOW_POWER: + /* Upclock if more than 95% busy over 16ms */ + I915_WRITE(GEN6_RP_UP_EI, 12500); + I915_WRITE(GEN6_RP_UP_THRESHOLD, 11800); + + /* Downclock if less than 85% busy over 32ms */ + I915_WRITE(GEN6_RP_DOWN_EI, 25000); + I915_WRITE(GEN6_RP_DOWN_THRESHOLD, 21250); + + I915_WRITE(GEN6_RP_CONTROL, + GEN6_RP_MEDIA_TURBO | + GEN6_RP_MEDIA_HW_NORMAL_MODE | + GEN6_RP_MEDIA_IS_GFX | + GEN6_RP_ENABLE | + GEN6_RP_UP_BUSY_AVG | + GEN6_RP_DOWN_IDLE_AVG); + break; + + case BETWEEN: + /* Upclock if more than 90% busy over 13ms */ + I915_WRITE(GEN6_RP_UP_EI, 10250); + I915_WRITE(GEN6_RP_UP_THRESHOLD, 9225); + + /* Downclock if less than 75% busy over 32ms */ + I915_WRITE(GEN6_RP_DOWN_EI, 25000); + I915_WRITE(GEN6_RP_DOWN_THRESHOLD, 18750); + + I915_WRITE(GEN6_RP_CONTROL, + GEN6_RP_MEDIA_TURBO | + GEN6_RP_MEDIA_HW_NORMAL_MODE | + GEN6_RP_MEDIA_IS_GFX | + GEN6_RP_ENABLE | + GEN6_RP_UP_BUSY_AVG | + GEN6_RP_DOWN_IDLE_AVG); + break; + + case HIGH_POWER: + /* Upclock if more than 85% busy over 10ms */ + I915_WRITE(GEN6_RP_UP_EI, 8000); + I915_WRITE(GEN6_RP_UP_THRESHOLD, 6800); + + /* Downclock if less than 60% busy over 32ms */ + I915_WRITE(GEN6_RP_DOWN_EI, 25000); + I915_WRITE(GEN6_RP_DOWN_THRESHOLD, 15000); + + I915_WRITE(GEN6_RP_CONTROL, + GEN6_RP_MEDIA_TURBO | + GEN6_RP_MEDIA_HW_NORMAL_MODE | + GEN6_RP_MEDIA_IS_GFX | + GEN6_RP_ENABLE | + GEN6_RP_UP_BUSY_AVG | + GEN6_RP_DOWN_IDLE_AVG); + break; + } + + dev_priv->rps.power = new_power; + dev_priv->rps.last_adj = 0; +} + void gen6_set_rps(struct drm_device *dev, u8 val) { struct drm_i915_private *dev_priv = dev->dev_private; @@ -3314,6 +3416,8 @@ void gen6_set_rps(struct drm_device *dev, u8 val) if (val == dev_priv->rps.cur_delay) return; + gen6_set_rps_thresholds(dev_priv, val); + if (IS_HASWELL(dev)) I915_WRITE(GEN6_RPNSWREQ, HSW_FREQUENCY(val)); @@ -3335,6 +3439,28 @@ void gen6_set_rps(struct drm_device *dev, u8 val) trace_intel_gpu_freq_change(val * 50); } +void gen6_rps_idle(struct drm_i915_private *dev_priv) +{ + mutex_lock(&dev_priv->rps.hw_lock); + if (dev_priv->info->is_valleyview) + valleyview_set_rps(dev_priv->dev, dev_priv->rps.min_delay); + else + gen6_set_rps(dev_priv->dev, dev_priv->rps.min_delay); + dev_priv->rps.last_adj = 0; + mutex_unlock(&dev_priv->rps.hw_lock); +} + +void gen6_rps_boost(struct drm_i915_private *dev_priv) +{ + mutex_lock(&dev_priv->rps.hw_lock); + if (dev_priv->info->is_valleyview) + valleyview_set_rps(dev_priv->dev, dev_priv->rps.max_delay); + else + gen6_set_rps(dev_priv->dev, dev_priv->rps.max_delay); + dev_priv->rps.last_adj = 0; + mutex_unlock(&dev_priv->rps.hw_lock); +} + /* * Wait until the previous freq change has completed, * or the timeout elapsed, and then update our notion @@ -3516,7 +3642,10 @@ static void gen6_enable_rps(struct drm_device *dev) /* In units of 50MHz */ dev_priv->rps.hw_max = dev_priv->rps.max_delay = rp_state_cap & 0xff; - dev_priv->rps.min_delay = (rp_state_cap & 0xff0000) >> 16; + dev_priv->rps.min_delay = (rp_state_cap >> 16) & 0xff; + dev_priv->rps.rp1_delay = (rp_state_cap >> 8) & 0xff; + dev_priv->rps.rp0_delay = (rp_state_cap >> 0) & 0xff; + dev_priv->rps.rpe_delay = dev_priv->rps.rp1_delay; dev_priv->rps.cur_delay = 0; /* disable the counters and set deterministic thresholds */ @@ -3564,38 +3693,9 @@ static void gen6_enable_rps(struct drm_device *dev) GEN6_RC_CTL_EI_MODE(1) | GEN6_RC_CTL_HW_ENABLE); - if (IS_HASWELL(dev)) { - I915_WRITE(GEN6_RPNSWREQ, - HSW_FREQUENCY(10)); - I915_WRITE(GEN6_RC_VIDEO_FREQ, - HSW_FREQUENCY(12)); - } else { - I915_WRITE(GEN6_RPNSWREQ, - GEN6_FREQUENCY(10) | - GEN6_OFFSET(0) | - GEN6_AGGRESSIVE_TURBO); - I915_WRITE(GEN6_RC_VIDEO_FREQ, - GEN6_FREQUENCY(12)); - } - - I915_WRITE(GEN6_RP_DOWN_TIMEOUT, 1000000); - I915_WRITE(GEN6_RP_INTERRUPT_LIMITS, - dev_priv->rps.max_delay << 24 | - dev_priv->rps.min_delay << 16); - - I915_WRITE(GEN6_RP_UP_THRESHOLD, 59400); - I915_WRITE(GEN6_RP_DOWN_THRESHOLD, 245000); - I915_WRITE(GEN6_RP_UP_EI, 66000); - I915_WRITE(GEN6_RP_DOWN_EI, 350000); - + /* Power down if completely idle for over 50ms */ + I915_WRITE(GEN6_RP_DOWN_TIMEOUT, 50000); I915_WRITE(GEN6_RP_IDLE_HYSTERSIS, 10); - I915_WRITE(GEN6_RP_CONTROL, - GEN6_RP_MEDIA_TURBO | - GEN6_RP_MEDIA_HW_NORMAL_MODE | - GEN6_RP_MEDIA_IS_GFX | - GEN6_RP_ENABLE | - GEN6_RP_UP_BUSY_AVG | - (IS_HASWELL(dev) ? GEN7_RP_DOWN_IDLE_AVG : GEN6_RP_DOWN_IDLE_CONT)); ret = sandybridge_pcode_write(dev_priv, GEN6_PCODE_WRITE_MIN_FREQ_TABLE, 0); if (!ret) { @@ -3611,7 +3711,8 @@ static void gen6_enable_rps(struct drm_device *dev) DRM_DEBUG_DRIVER("Failed to set the min frequency\n"); } - gen6_set_rps(dev_priv->dev, (gt_perf_status & 0xff00) >> 8); + dev_priv->rps.power = HIGH_POWER; /* force a reset */ + gen6_set_rps(dev_priv->dev, dev_priv->rps.min_delay); gen6_enable_rps_interrupts(dev); @@ -3653,9 +3754,9 @@ void gen6_update_ring_freq(struct drm_device *dev) /* Convert from kHz to MHz */ max_ia_freq /= 1000; - min_ring_freq = I915_READ(MCHBAR_MIRROR_BASE_SNB + DCLK); - /* convert DDR frequency from units of 133.3MHz to bandwidth */ - min_ring_freq = (2 * 4 * min_ring_freq + 2) / 3; + min_ring_freq = I915_READ(MCHBAR_MIRROR_BASE_SNB + DCLK) & 0xf; + /* convert DDR frequency from units of 266.6MHz to bandwidth */ + min_ring_freq = mult_frac(min_ring_freq, 8, 3); /* * For each potential GPU frequency, load a ring frequency we'd like @@ -3668,7 +3769,7 @@ void gen6_update_ring_freq(struct drm_device *dev) unsigned int ia_freq = 0, ring_freq = 0; if (IS_HASWELL(dev)) { - ring_freq = (gpu_freq * 5 + 3) / 4; + ring_freq = mult_frac(gpu_freq, 5, 4); ring_freq = max(min_ring_freq, ring_freq); /* leave ia_freq as the default, chosen by cpufreq */ } else { @@ -3724,24 +3825,6 @@ int valleyview_rps_min_freq(struct drm_i915_private *dev_priv) return vlv_punit_read(dev_priv, PUNIT_REG_GPU_LFM) & 0xff; } -static void vlv_rps_timer_work(struct work_struct *work) -{ - drm_i915_private_t *dev_priv = container_of(work, drm_i915_private_t, - rps.vlv_work.work); - - /* - * Timer fired, we must be idle. Drop to min voltage state. - * Note: we use RPe here since it should match the - * Vmin we were shooting for. That should give us better - * perf when we come back out of RC6 than if we used the - * min freq available. - */ - mutex_lock(&dev_priv->rps.hw_lock); - if (dev_priv->rps.cur_delay > dev_priv->rps.rpe_delay) - valleyview_set_rps(dev_priv->dev, dev_priv->rps.rpe_delay); - mutex_unlock(&dev_priv->rps.hw_lock); -} - static void valleyview_setup_pctx(struct drm_device *dev) { struct drm_i915_private *dev_priv = dev->dev_private; @@ -3794,7 +3877,8 @@ static void valleyview_enable_rps(struct drm_device *dev) WARN_ON(!mutex_is_locked(&dev_priv->rps.hw_lock)); if ((gtfifodbg = I915_READ(GTFIFODBG))) { - DRM_ERROR("GT fifo had a previous error %x\n", gtfifodbg); + DRM_DEBUG_DRIVER("GT fifo had a previous error %x\n", + gtfifodbg); I915_WRITE(GTFIFODBG, gtfifodbg); } @@ -3827,7 +3911,10 @@ static void valleyview_enable_rps(struct drm_device *dev) I915_WRITE(GEN6_RC6_THRESHOLD, 0xc350); /* allows RC6 residency counter to work */ - I915_WRITE(0x138104, _MASKED_BIT_ENABLE(0x3)); + I915_WRITE(VLV_COUNTER_CONTROL, + _MASKED_BIT_ENABLE(VLV_COUNT_RANGE_HIGH | + VLV_MEDIA_RC6_COUNT_EN | + VLV_RENDER_RC6_COUNT_EN)); if (intel_enable_rc6(dev) & INTEL_RC6_ENABLE) rc6_mode = GEN7_RC_CTL_TO_MODE; I915_WRITE(GEN6_RC_CONTROL, rc6_mode); @@ -3880,8 +3967,6 @@ static void valleyview_enable_rps(struct drm_device *dev) dev_priv->rps.rpe_delay), dev_priv->rps.rpe_delay); - INIT_DELAYED_WORK(&dev_priv->rps.vlv_work, vlv_rps_timer_work); - valleyview_set_rps(dev_priv->dev, dev_priv->rps.rpe_delay); gen6_enable_rps_interrupts(dev); @@ -4621,8 +4706,6 @@ void intel_disable_gt_powersave(struct drm_device *dev) } else if (INTEL_INFO(dev)->gen >= 6) { cancel_delayed_work_sync(&dev_priv->rps.delayed_resume_work); cancel_work_sync(&dev_priv->rps.work); - if (IS_VALLEYVIEW(dev)) - cancel_delayed_work_sync(&dev_priv->rps.vlv_work); mutex_lock(&dev_priv->rps.hw_lock); if (IS_VALLEYVIEW(dev)) valleyview_disable_rps(dev); @@ -5498,7 +5581,7 @@ void intel_set_power_well(struct drm_device *dev, bool enable) spin_unlock_irq(&power_well->lock); } -void intel_resume_power_well(struct drm_device *dev) +static void intel_resume_power_well(struct drm_device *dev) { struct drm_i915_private *dev_priv = dev->dev_private; struct i915_power_well *power_well = &dev_priv->power_well; |