Merge branch 'drm-fixes' of git://people.freedesktop.org/~airlied/linux

Pull drm fixes from Dave Airlie:
 "Mostly radeon, more fixes for dynamic power management which is is off
  by default for this release anyways, but there are a large number of
  testers, so I'd like to keep merging the fixes.

  Otherwise, radeon UVD fixes affecting suspend/resume regressions, i915
  regression fixes, one for your mac mini, ast, mgag200, cirrus ttm fix
  and one regression fix in the core"

* 'drm-fixes' of git://people.freedesktop.org/~airlied/linux: (25 commits)
  drm: Don't pass negative delta to ktime_sub_ns()
  drm/radeon: make missing smc ucode non-fatal
  drm/radeon/dpm: require rlc for dpm
  drm/radeon/cik: use a mutex to properly lock srbm instanced registers
  drm/radeon: remove unnecessary unpin
  drm/radeon: add more UVD CS checking
  drm/radeon: stop sending invalid UVD destroy msg
  drm/radeon: only save UVD bo when we have open handles
  drm/radeon: always program the MC on startup
  drm/radeon: fix audio dto calculation on DCE3+ (v3)
  drm/radeon/dpm: disable sclk ss on rv6xx
  drm/radeon: fix halting UVD
  drm/radeon/dpm: adjust power state properly for UVD on SI
  drm/radeon/dpm: fix spread spectrum setup (v2)
  drm/radeon/dpm: adjust thermal protection requirements
  drm/radeon: select audio dto based on encoder id for DCE3
  drm/radeon: properly handle pm on gpu reset
  drm/i915: do not disable backlight on vgaswitcheroo switch off
  drm/i915: Don't call encoder's get_config unless encoder is active
  drm/i915: avoid brightness overflow when doing scale
  ...

32 files changed, 353 insertions, 165 deletions

diff --git a/drivers/gpu/drm/ast/ast_ttm.c b/drivers/gpu/drm/ast/ast_ttm.c
index 98d670825a1..6e8887fe6c1 100644
--- a/drivers/gpu/drm/ast/ast_ttm.c
+++ b/drivers/gpu/drm/ast/ast_ttm.c
@@ -323,6 +323,7 @@ int ast_bo_create(struct drm_device *dev, int size, int align,
 
 	astbo->gem.driver_private = NULL;
 	astbo->bo.bdev = &ast->ttm.bdev;
+	astbo->bo.bdev->dev_mapping = dev->dev_mapping;
 
 	ast_ttm_placement(astbo, TTM_PL_FLAG_VRAM | TTM_PL_FLAG_SYSTEM);
 
diff --git a/drivers/gpu/drm/cirrus/cirrus_ttm.c b/drivers/gpu/drm/cirrus/cirrus_ttm.c
index 0047012045c..69fd8f1ac8d 100644
--- a/drivers/gpu/drm/cirrus/cirrus_ttm.c
+++ b/drivers/gpu/drm/cirrus/cirrus_ttm.c
@@ -328,6 +328,7 @@ int cirrus_bo_create(struct drm_device *dev, int size, int align,
 
 	cirrusbo->gem.driver_private = NULL;
 	cirrusbo->bo.bdev = &cirrus->ttm.bdev;
+	cirrusbo->bo.bdev->dev_mapping = dev->dev_mapping;
 
 	cirrus_ttm_placement(cirrusbo, TTM_PL_FLAG_VRAM | TTM_PL_FLAG_SYSTEM);
 
diff --git a/drivers/gpu/drm/drm_irq.c b/drivers/gpu/drm/drm_irq.c
index 8bcce7866d3..f92da0a32f0 100644
--- a/drivers/gpu/drm/drm_irq.c
+++ b/drivers/gpu/drm/drm_irq.c
@@ -708,7 +708,10 @@ int drm_calc_vbltimestamp_from_scanoutpos(struct drm_device *dev, int crtc,
 	/* Subtract time delta from raw timestamp to get final
 	 * vblank_time timestamp for end of vblank.
 	 */
-	etime = ktime_sub_ns(etime, delta_ns);
+	if (delta_ns < 0)
+		etime = ktime_add_ns(etime, -delta_ns);
+	else
+		etime = ktime_sub_ns(etime, delta_ns);
 	*vblank_time = ktime_to_timeval(etime);
 
 	DRM_DEBUG("crtc %d : v %d p(%d,%d)@ %ld.%ld -> %ld.%ld [e %d us, %d rep]\n",
diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index f2326fc60ac..6f514297c48 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -1856,10 +1856,16 @@
 #define CRT_HOTPLUG_DETECT_VOLTAGE_475MV	(1 << 2)
 
 #define PORT_HOTPLUG_STAT	(dev_priv->info->display_mmio_offset + 0x61114)
-/* HDMI/DP bits are gen4+ */
-#define   PORTB_HOTPLUG_LIVE_STATUS               (1 << 29)
+/*
+ * HDMI/DP bits are gen4+
+ *
+ * WARNING: Bspec for hpd status bits on gen4 seems to be completely confused.
+ * Please check the detailed lore in the commit message for for experimental
+ * evidence.
+ */
+#define   PORTD_HOTPLUG_LIVE_STATUS               (1 << 29)
 #define   PORTC_HOTPLUG_LIVE_STATUS               (1 << 28)
-#define   PORTD_HOTPLUG_LIVE_STATUS               (1 << 27)
+#define   PORTB_HOTPLUG_LIVE_STATUS               (1 << 27)
 #define   PORTD_HOTPLUG_INT_STATUS		(3 << 21)
 #define   PORTC_HOTPLUG_INT_STATUS		(3 << 19)
 #define   PORTB_HOTPLUG_INT_STATUS		(3 << 17)
diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 5fb305840db..e38b4578665 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -8269,9 +8269,11 @@ check_crtc_state(struct drm_device *dev)
 
 		list_for_each_entry(encoder, &dev->mode_config.encoder_list,
 				    base.head) {
+			enum pipe pipe;
 			if (encoder->base.crtc != &crtc->base)
 				continue;
-			if (encoder->get_config)
+			if (encoder->get_config &&
+			    encoder->get_hw_state(encoder, &pipe))
 				encoder->get_config(encoder, &pipe_config);
 		}
 
diff --git a/drivers/gpu/drm/i915/intel_panel.c b/drivers/gpu/drm/i915/intel_panel.c
index 67e2c1f1c9a..5950888ae1d 100644
--- a/drivers/gpu/drm/i915/intel_panel.c
+++ b/drivers/gpu/drm/i915/intel_panel.c
@@ -497,8 +497,11 @@ void intel_panel_set_backlight(struct drm_device *dev, u32 level, u32 max)
 		goto out;
 	}
 
-	/* scale to hardware */
-	level = level * freq / max;
+	/* scale to hardware, but be careful to not overflow */
+	if (freq < max)
+		level = level * freq / max;
+	else
+		level = freq / max * level;
 
 	dev_priv->backlight.level = level;
 	if (dev_priv->backlight.device)
@@ -515,6 +518,17 @@ void intel_panel_disable_backlight(struct drm_device *dev)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	unsigned long flags;
 
+	/*
+	 * Do not disable backlight on the vgaswitcheroo path. When switching
+	 * away from i915, the other client may depend on i915 to handle the
+	 * backlight. This will leave the backlight on unnecessarily when
+	 * another client is not activated.
+	 */
+	if (dev->switch_power_state == DRM_SWITCH_POWER_CHANGING) {
+		DRM_DEBUG_DRIVER("Skipping backlight disable on vga switch\n");
+		return;
+	}
+
 	spin_lock_irqsave(&dev_priv->backlight.lock, flags);
 
 	dev_priv->backlight.enabled = false;
diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c
index f895d1508df..b0e4a0bd131 100644
--- a/drivers/gpu/drm/i915/intel_pm.c
+++ b/drivers/gpu/drm/i915/intel_pm.c
@@ -5063,8 +5063,26 @@ static void __intel_set_power_well(struct drm_device *dev, bool enable)
 		}
 	} else {
 		if (enable_requested) {
+			unsigned long irqflags;
+			enum pipe p;
+
 			I915_WRITE(HSW_PWR_WELL_DRIVER, 0);
+			POSTING_READ(HSW_PWR_WELL_DRIVER);
 			DRM_DEBUG_KMS("Requesting to disable the power well\n");
+
+			/*
+			 * After this, the registers on the pipes that are part
+			 * of the power well will become zero, so we have to
+			 * adjust our counters according to that.
+			 *
+			 * FIXME: Should we do this in general in
+			 * drm_vblank_post_modeset?
+			 */
+			spin_lock_irqsave(&dev->vbl_lock, irqflags);
+			for_each_pipe(p)
+				if (p != PIPE_A)
+					dev->last_vblank[p] = 0;
+			spin_unlock_irqrestore(&dev->vbl_lock, irqflags);
 		}
 	}
 }
diff --git a/drivers/gpu/drm/mgag200/mgag200_ttm.c b/drivers/gpu/drm/mgag200/mgag200_ttm.c
index 13878d5de06..d70e4a92773 100644
--- a/drivers/gpu/drm/mgag200/mgag200_ttm.c
+++ b/drivers/gpu/drm/mgag200/mgag200_ttm.c
@@ -323,6 +323,7 @@ int mgag200_bo_create(struct drm_device *dev, int size, int align,
 
 	mgabo->gem.driver_private = NULL;
 	mgabo->bo.bdev = &mdev->ttm.bdev;
+	mgabo->bo.bdev->dev_mapping = dev->dev_mapping;
 
 	mgag200_ttm_placement(mgabo, TTM_PL_FLAG_VRAM | TTM_PL_FLAG_SYSTEM);
 
diff --git a/drivers/gpu/drm/radeon/btc_dpm.c b/drivers/gpu/drm/radeon/btc_dpm.c
index 0bfd55e0882..9953e1fbc46 100644
--- a/drivers/gpu/drm/radeon/btc_dpm.c
+++ b/drivers/gpu/drm/radeon/btc_dpm.c
@@ -2548,9 +2548,6 @@ int btc_dpm_init(struct radeon_device *rdev)
 {
 	struct rv7xx_power_info *pi;
 	struct evergreen_power_info *eg_pi;
-	int index = GetIndexIntoMasterTable(DATA, ASIC_InternalSS_Info);
-	u16 data_offset, size;
-	u8 frev, crev;
 	struct atom_clock_dividers dividers;
 	int ret;
 
@@ -2633,16 +2630,7 @@ int btc_dpm_init(struct radeon_device *rdev)
 	eg_pi->vddci_control =
 		radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_VDDCI, 0);
 
-	if (atom_parse_data_header(rdev->mode_info.atom_context, index, &size,
-                                   &frev, &crev, &data_offset)) {
-		pi->sclk_ss = true;
-		pi->mclk_ss = true;
-		pi->dynamic_ss = true;
-	} else {
-		pi->sclk_ss = false;
-		pi->mclk_ss = false;
-		pi->dynamic_ss = true;
-	}
+	rv770_get_engine_memory_ss(rdev);
 
 	pi->asi = RV770_ASI_DFLT;
 	pi->pasi = CYPRESS_HASI_DFLT;
@@ -2659,8 +2647,7 @@ int btc_dpm_init(struct radeon_device *rdev)
 
 	pi->dynamic_pcie_gen2 = true;
 
-	if (pi->gfx_clock_gating &&
-	    (rdev->pm.int_thermal_type != THERMAL_TYPE_NONE))
+	if (rdev->pm.int_thermal_type != THERMAL_TYPE_NONE)
 		pi->thermal_protection = true;
 	else
 		pi->thermal_protection = false;
diff --git a/drivers/gpu/drm/radeon/cik.c b/drivers/gpu/drm/radeon/cik.c
index 6dacec4e209..8928bd109c1 100644
--- a/drivers/gpu/drm/radeon/cik.c
+++ b/drivers/gpu/drm/radeon/cik.c
@@ -2587,9 +2587,11 @@ u32 cik_compute_ring_get_rptr(struct radeon_device *rdev,
 	if (rdev->wb.enabled) {
 		rptr = le32_to_cpu(rdev->wb.wb[ring->rptr_offs/4]);
 	} else {
+		mutex_lock(&rdev->srbm_mutex);
 		cik_srbm_select(rdev, ring->me, ring->pipe, ring->queue, 0);
 		rptr = RREG32(CP_HQD_PQ_RPTR);
 		cik_srbm_select(rdev, 0, 0, 0, 0);
+		mutex_unlock(&rdev->srbm_mutex);
 	}
 	rptr = (rptr & ring->ptr_reg_mask) >> ring->ptr_reg_shift;
 
@@ -2604,9 +2606,11 @@ u32 cik_compute_ring_get_wptr(struct radeon_device *rdev,
 	if (rdev->wb.enabled) {
 		wptr = le32_to_cpu(rdev->wb.wb[ring->wptr_offs/4]);
 	} else {
+		mutex_lock(&rdev->srbm_mutex);
 		cik_srbm_select(rdev, ring->me, ring->pipe, ring->queue, 0);
 		wptr = RREG32(CP_HQD_PQ_WPTR);
 		cik_srbm_select(rdev, 0, 0, 0, 0);
+		mutex_unlock(&rdev->srbm_mutex);
 	}
 	wptr = (wptr & ring->ptr_reg_mask) >> ring->ptr_reg_shift;
 
@@ -2897,6 +2901,7 @@ static int cik_cp_compute_resume(struct radeon_device *rdev)
 	WREG32(CP_CPF_DEBUG, tmp);
 
 	/* init the pipes */
+	mutex_lock(&rdev->srbm_mutex);
 	for (i = 0; i < (rdev->mec.num_pipe * rdev->mec.num_mec); i++) {
 		int me = (i < 4) ? 1 : 2;
 		int pipe = (i < 4) ? i : (i - 4);
@@ -2919,6 +2924,7 @@ static int cik_cp_compute_resume(struct radeon_device *rdev)
 		WREG32(CP_HPD_EOP_CONTROL, tmp);
 	}
 	cik_srbm_select(rdev, 0, 0, 0, 0);
+	mutex_unlock(&rdev->srbm_mutex);
 
 	/* init the queues.  Just two for now. */
 	for (i = 0; i < 2; i++) {
@@ -2972,6 +2978,7 @@ static int cik_cp_compute_resume(struct radeon_device *rdev)
 		mqd->static_thread_mgmt23[0] = 0xffffffff;
 		mqd->static_thread_mgmt23[1] = 0xffffffff;
 
+		mutex_lock(&rdev->srbm_mutex);
 		cik_srbm_select(rdev, rdev->ring[idx].me,
 				rdev->ring[idx].pipe,
 				rdev->ring[idx].queue, 0);
@@ -3099,6 +3106,7 @@ static int cik_cp_compute_resume(struct radeon_device *rdev)
 		WREG32(CP_HQD_ACTIVE, mqd->queue_state.cp_hqd_active);
 
 		cik_srbm_select(rdev, 0, 0, 0, 0);
+		mutex_unlock(&rdev->srbm_mutex);
 
 		radeon_bo_kunmap(rdev->ring[idx].mqd_obj);
 		radeon_bo_unreserve(rdev->ring[idx].mqd_obj);
@@ -4320,6 +4328,7 @@ static int cik_pcie_gart_enable(struct radeon_device *rdev)
 
 	/* XXX SH_MEM regs */
 	/* where to put LDS, scratch, GPUVM in FSA64 space */
+	mutex_lock(&rdev->srbm_mutex);
 	for (i = 0; i < 16; i++) {
 		cik_srbm_select(rdev, 0, 0, 0, i);
 		/* CP and shaders */
@@ -4335,6 +4344,7 @@ static int cik_pcie_gart_enable(struct radeon_device *rdev)
 		/* XXX SDMA RLC - todo */
 	}
 	cik_srbm_select(rdev, 0, 0, 0, 0);
+	mutex_unlock(&rdev->srbm_mutex);
 
 	cik_pcie_gart_tlb_flush(rdev);
 	DRM_INFO("PCIE GART of %uM enabled (table at 0x%016llX).\n",
@@ -5954,6 +5964,8 @@ static int cik_startup(struct radeon_device *rdev)
 	struct radeon_ring *ring;
 	int r;
 
+	cik_mc_program(rdev);
+
 	if (rdev->flags & RADEON_IS_IGP) {
 		if (!rdev->me_fw || !rdev->pfp_fw || !rdev->ce_fw ||
 		    !rdev->mec_fw || !rdev->sdma_fw || !rdev->rlc_fw) {
@@ -5985,7 +5997,6 @@ static int cik_startup(struct radeon_device *rdev)
 	if (r)
 		return r;
 
-	cik_mc_program(rdev);
 	r = cik_pcie_gart_enable(rdev);
 	if (r)
 		return r;
@@ -6194,7 +6205,7 @@ int cik_suspend(struct radeon_device *rdev)
 	radeon_vm_manager_fini(rdev);
 	cik_cp_enable(rdev, false);
 	cik_sdma_enable(rdev, false);
-	r600_uvd_rbc_stop(rdev);
+	r600_uvd_stop(rdev);
 	radeon_uvd_suspend(rdev);
 	cik_irq_suspend(rdev);
 	radeon_wb_disable(rdev);
@@ -6358,6 +6369,7 @@ void cik_fini(struct radeon_device *rdev)
 	radeon_vm_manager_fini(rdev);
 	radeon_ib_pool_fini(rdev);
 	radeon_irq_kms_fini(rdev);
+	r600_uvd_stop(rdev);
 	radeon_uvd_fini(rdev);
 	cik_pcie_gart_fini(rdev);
 	r600_vram_scratch_fini(rdev);
@@ -6978,7 +6990,7 @@ int cik_uvd_resume(struct radeon_device *rdev)
 
 	/* programm the VCPU memory controller bits 0-27 */
 	addr = rdev->uvd.gpu_addr >> 3;
-	size = RADEON_GPU_PAGE_ALIGN(rdev->uvd.fw_size + 4) >> 3;
+	size = RADEON_GPU_PAGE_ALIGN(rdev->uvd_fw->size + 4) >> 3;
 	WREG32(UVD_VCPU_CACHE_OFFSET0, addr);
 	WREG32(UVD_VCPU_CACHE_SIZE0, size);
 
diff --git a/drivers/gpu/drm/radeon/cypress_dpm.c b/drivers/gpu/drm/radeon/cypress_dpm.c
index 9bcdd174780..7e5d0b570a3 100644
--- a/drivers/gpu/drm/radeon/cypress_dpm.c
+++ b/drivers/gpu/drm/radeon/cypress_dpm.c
@@ -2038,9 +2038,6 @@ int cypress_dpm_init(struct radeon_device *rdev)
 {
 	struct rv7xx_power_info *pi;
 	struct evergreen_power_info *eg_pi;
-	int index = GetIndexIntoMasterTable(DATA, ASIC_InternalSS_Info);
-	uint16_t data_offset, size;
-	uint8_t frev, crev;
 	struct atom_clock_dividers dividers;
 	int ret;
 
@@ -2092,16 +2089,7 @@ int cypress_dpm_init(struct radeon_device *rdev)
 	eg_pi->vddci_control =
 		radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_VDDCI, 0);
 
-	if (atom_parse_data_header(rdev->mode_info.atom_context, index, &size,
-                                   &frev, &crev, &data_offset)) {
-		pi->sclk_ss = true;
-		pi->mclk_ss = true;
-		pi->dynamic_ss = true;
-	} else {
-		pi->sclk_ss = false;
-		pi->mclk_ss = false;
-		pi->dynamic_ss = true;
-	}
+	rv770_get_engine_memory_ss(rdev);
 
 	pi->asi = RV770_ASI_DFLT;
 	pi->pasi = CYPRESS_HASI_DFLT;
@@ -2122,8 +2110,7 @@ int cypress_dpm_init(struct radeon_device *rdev)
 
 	pi->dynamic_pcie_gen2 = true;
 
-	if (pi->gfx_clock_gating &&
-	    (rdev->pm.int_thermal_type != THERMAL_TYPE_NONE))
+	if (rdev->pm.int_thermal_type != THERMAL_TYPE_NONE)
 		pi->thermal_protection = true;
 	else
 		pi->thermal_protection = false;
diff --git a/drivers/gpu/drm/radeon/evergreen.c b/drivers/gpu/drm/radeon/evergreen.c
index 038dcac7670..d5b49e33315 100644
--- a/drivers/gpu/drm/radeon/evergreen.c
+++ b/drivers/gpu/drm/radeon/evergreen.c
@@ -5106,6 +5106,8 @@ static int evergreen_startup(struct radeon_device *rdev)
 	/* enable aspm */
 	evergreen_program_aspm(rdev);
 
+	evergreen_mc_program(rdev);
+
 	if (ASIC_IS_DCE5(rdev)) {
 		if (!rdev->me_fw || !rdev->pfp_fw || !rdev->rlc_fw || !rdev->mc_fw) {
 			r = ni_init_microcode(rdev);
@@ -5133,7 +5135,6 @@ static int evergreen_startup(struct radeon_device *rdev)
 	if (r)
 		return r;
 
-	evergreen_mc_program(rdev);
 	if (rdev->flags & RADEON_IS_AGP) {
 		evergreen_agp_enable(rdev);
 	} else {
@@ -5291,10 +5292,10 @@ int evergreen_resume(struct radeon_device *rdev)
 int evergreen_suspend(struct radeon_device *rdev)
 {
 	r600_audio_fini(rdev);
+	r600_uvd_stop(rdev);
 	radeon_uvd_suspend(rdev);
 	r700_cp_stop(rdev);
 	r600_dma_stop(rdev);
-	r600_uvd_rbc_stop(rdev);
 	evergreen_irq_suspend(rdev);
 	radeon_wb_disable(rdev);
 	evergreen_pcie_gart_disable(rdev);
@@ -5429,6 +5430,7 @@ void evergreen_fini(struct radeon_device *rdev)
 	radeon_ib_pool_fini(rdev);
 	radeon_irq_kms_fini(rdev);
 	evergreen_pcie_gart_fini(rdev);
+	r600_uvd_stop(rdev);
 	radeon_uvd_fini(rdev);
 	r600_vram_scratch_fini(rdev);
 	radeon_gem_fini(rdev);
diff --git a/drivers/gpu/drm/radeon/evergreen_hdmi.c b/drivers/gpu/drm/radeon/evergreen_hdmi.c
index bb9ea364131..b0e280058b9 100644
--- a/drivers/gpu/drm/radeon/evergreen_hdmi.c
+++ b/drivers/gpu/drm/radeon/evergreen_hdmi.c
@@ -148,18 +148,40 @@ static void evergreen_audio_set_dto(struct drm_encoder *encoder, u32 clock)
 	struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv;
 	struct radeon_crtc *radeon_crtc = to_radeon_crtc(encoder->crtc);
 	u32 base_rate = 24000;
+	u32 max_ratio = clock / base_rate;
+	u32 dto_phase;
+	u32 dto_modulo = clock;
+	u32 wallclock_ratio;
+	u32 dto_cntl;
 
 	if (!dig || !dig->afmt)
 		return;
 
+	if (max_ratio >= 8) {
+		dto_phase = 192 * 1000;
+		wallclock_ratio = 3;
+	} else if (max_ratio >= 4) {
+		dto_phase = 96 * 1000;
+		wallclock_ratio = 2;
+	} else if (max_ratio >= 2) {
+		dto_phase = 48 * 1000;
+		wallclock_ratio = 1;
+	} else {
+		dto_phase = 24 * 1000;
+		wallclock_ratio = 0;
+	}
+	dto_cntl = RREG32(DCCG_AUDIO_DTO0_CNTL) & ~DCCG_AUDIO_DTO_WALLCLOCK_RATIO_MASK;
+	dto_cntl |= DCCG_AUDIO_DTO_WALLCLOCK_RATIO(wallclock_ratio);
+	WREG32(DCCG_AUDIO_DTO0_CNTL, dto_cntl);
+
 	/* XXX two dtos; generally use dto0 for hdmi */
 	/* Express [24MHz / target pixel clock] as an exact rational
 	 * number (coefficient of two integer numbers.  DCCG_AUDIO_DTOx_PHASE
 	 * is the numerator, DCCG_AUDIO_DTOx_MODULE is the denominator
 	 */
 	WREG32(DCCG_AUDIO_DTO_SOURCE, DCCG_AUDIO_DTO0_SOURCE_SEL(radeon_crtc->crtc_id));
-	WREG32(DCCG_AUDIO_DTO0_PHASE, base_rate * 100);
-	WREG32(DCCG_AUDIO_DTO0_MODULE, clock * 100);
+	WREG32(DCCG_AUDIO_DTO0_PHASE, dto_phase);
+	WREG32(DCCG_AUDIO_DTO0_MODULE, dto_modulo);
 }
 
 
diff --git a/drivers/gpu/drm/radeon/evergreend.h b/drivers/gpu/drm/radeon/evergreend.h
index a7baf67aef6..0d582ac1dc3 100644
--- a/drivers/gpu/drm/radeon/evergreend.h
+++ b/drivers/gpu/drm/radeon/evergreend.h
@@ -497,6 +497,9 @@
 #define DCCG_AUDIO_DTO0_MODULE            0x05b4
 #define DCCG_AUDIO_DTO0_LOAD              0x05b8
 #define DCCG_AUDIO_DTO0_CNTL              0x05bc
+#       define DCCG_AUDIO_DTO_WALLCLOCK_RATIO(x) (((x) & 7) << 0)
+#       define DCCG_AUDIO_DTO_WALLCLOCK_RATIO_MASK 7
+#       define DCCG_AUDIO_DTO_WALLCLOCK_RATIO_SHIFT 0
 
 #define DCCG_AUDIO_DTO1_PHASE             0x05c0
 #define DCCG_AUDIO_DTO1_MODULE            0x05c4
diff --git a/drivers/gpu/drm/radeon/ni.c b/drivers/gpu/drm/radeon/ni.c
index 56bd4f3be4f..ccb4f8b5485 100644
--- a/drivers/gpu/drm/radeon/ni.c
+++ b/drivers/gpu/drm/radeon/ni.c
@@ -794,9 +794,13 @@ int ni_init_microcode(struct radeon_device *rdev)
 	if ((rdev->family >= CHIP_BARTS) && (rdev->family <= CHIP_CAYMAN)) {
 		snprintf(fw_name, sizeof(fw_name), "radeon/%s_smc.bin", chip_name);
 		err = request_firmware(&rdev->smc_fw, fw_name, rdev->dev);
-		if (err)
-			goto out;
-		if (rdev->smc_fw->size != smc_req_size) {
+		if (err) {
+			printk(KERN_ERR
+			       "smc: error loading firmware \"%s\"\n",
+			       fw_name);
+			release_firmware(rdev->smc_fw);
+			rdev->smc_fw = NULL;
+		} else if (rdev->smc_fw->size != smc_req_size) {
 			printk(KERN_ERR
 			       "ni_mc: Bogus length %zu in firmware \"%s\"\n",
 			       rdev->mc_fw->size, fw_name);
@@ -2079,6 +2083,8 @@ static int cayman_startup(struct radeon_device *rdev)
 	/* enable aspm */
 	evergreen_program_aspm(rdev);
 
+	evergreen_mc_program(rdev);
+
 	if (rdev->flags & RADEON_IS_IGP) {
 		if (!rdev->me_fw || !rdev->pfp_fw || !rdev->rlc_fw) {
 			r = ni_init_microcode(rdev);
@@ -2107,7 +2113,6 @@ static int cayman_startup(struct radeon_device *rdev)
 	if (r)
 		return r;
 
-	evergreen_mc_program(rdev);
 	r = cayman_pcie_gart_enable(rdev);
 	if (r)
 		return r;
@@ -2286,7 +2291,7 @@ int cayman_suspend(struct radeon_device *rdev)
 	radeon_vm_manager_fini(rdev);
 	cayman_cp_enable(rdev, false);
 	cayman_dma_stop(rdev);
-	r600_uvd_rbc_stop(rdev);
+	r600_uvd_stop(rdev);
 	radeon_uvd_suspend(rdev);
 	evergreen_irq_suspend(rdev);
 	radeon_wb_disable(rdev);
@@ -2418,6 +2423,7 @@ void cayman_fini(struct radeon_device *rdev)
 	radeon_vm_manager_fini(rdev);
 	radeon_ib_pool_fini(rdev);
 	radeon_irq_kms_fini(rdev);
+	r600_uvd_stop(rdev);
 	radeon_uvd_fini(rdev);
 	cayman_pcie_gart_fini(rdev);
 	r600_vram_scratch_fini(rdev);
diff --git a/drivers/gpu/drm/radeon/ni_dpm.c b/drivers/gpu/drm/radeon/ni_dpm.c
index 4f9b9bc20da..f0f5f748938 100644
--- a/drivers/gpu/drm/radeon/ni_dpm.c
+++ b/drivers/gpu/drm/radeon/ni_dpm.c
@@ -4067,9 +4067,6 @@ int ni_dpm_init(struct radeon_device *rdev)
 	struct rv7xx_power_info *pi;
 	struct evergreen_power_info *eg_pi;
 	struct ni_power_info *ni_pi;
-	int index = GetIndexIntoMasterTable(DATA, ASIC_InternalSS_Info);
-	u16 data_offset, size;
-	u8 frev, crev;
 	struct atom_clock_dividers dividers;
 	int ret;
 
@@ -4162,16 +4159,7 @@ int ni_dpm_init(struct radeon_device *rdev)
 	eg_pi->vddci_control =
 		radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_VDDCI, 0);
 
-	if (atom_parse_data_header(rdev->mode_info.atom_context, index, &size,
-                                   &frev, &crev, &data_offset)) {
-		pi->sclk_ss = true;
-		pi->mclk_ss = true;
-		pi->dynamic_ss = true;
-	} else {
-		pi->sclk_ss = false;
-		pi->mclk_ss = false;
-		pi->dynamic_ss = true;
-	}
+	rv770_get_engine_memory_ss(rdev);
 
 	pi->asi = RV770_ASI_DFLT;
 	pi->pasi = CYPRESS_HASI_DFLT;
@@ -4188,8 +4176,7 @@ int ni_dpm_init(struct radeon_device *rdev)
 
 	pi->dynamic_pcie_gen2 = true;
 
-	if (pi->gfx_clock_gating &&
-	    (rdev->pm.int_thermal_type != THERMAL_TYPE_NONE))
+	if (rdev->pm.int_thermal_type != THERMAL_TYPE_NONE)
 		pi->thermal_protection = true;
 	else
 		pi->thermal_protection = false;
diff --git a/drivers/gpu/drm/radeon/r600.c b/drivers/gpu/drm/radeon/r600.c
index 10f712e3700..e66e7207735 100644
--- a/drivers/gpu/drm/radeon/r600.c
+++ b/drivers/gpu/drm/radeon/r600.c
@@ -2299,9 +2299,13 @@ int r600_init_microcode(struct radeon_device *rdev)
 	if ((rdev->family >= CHIP_RV770) && (rdev->family <= CHIP_HEMLOCK)) {
 		snprintf(fw_name, sizeof(fw_name), "radeon/%s_smc.bin", smc_chip_name);
 		err = request_firmware(&rdev->smc_fw, fw_name, rdev->dev);
-		if (err)
-			goto out;
-		if (rdev->smc_fw->size != smc_req_size) {
+		if (err) {
+			printk(KERN_ERR
+			       "smc: error loading firmware \"%s\"\n",
+			       fw_name);
+			release_firmware(rdev->smc_fw);
+			rdev->smc_fw = NULL;
+		} else if (rdev->smc_fw->size != smc_req_size) {
 			printk(KERN_ERR
 			       "smc: Bogus length %zu in firmware \"%s\"\n",
 			       rdev->smc_fw->size, fw_name);
@@ -2697,12 +2701,29 @@ int r600_uvd_rbc_start(struct radeon_device *rdev)
 	return 0;
 }
 
-void r600_uvd_rbc_stop(struct radeon_device *rdev)
+void r600_uvd_stop(struct radeon_device *rdev)
 {
 	struct radeon_ring *ring = &rdev->ring[R600_RING_TYPE_UVD_INDEX];
 
 	/* force RBC into idle state */
 	WREG32(UVD_RBC_RB_CNTL, 0x11010101);
+
+	/* Stall UMC and register bus before resetting VCPU */
+	WREG32_P(UVD_LMI_CTRL2, 1 << 8, ~(1 << 8));
+	WREG32_P(UVD_RB_ARB_CTRL, 1 << 3, ~(1 << 3));
+	mdelay(1);
+
+	/* put VCPU into reset */
+	WREG32(UVD_SOFT_RESET, VCPU_SOFT_RESET);
+	mdelay(5);
+
+	/* disable VCPU clock */
+	WREG32(UVD_VCPU_CNTL, 0x0);
+
+	/* Unstall UMC and register bus */
+	WREG32_P(UVD_LMI_CTRL2, 0, ~(1 << 8));
+	WREG32_P(UVD_RB_ARB_CTRL, 0, ~(1 << 3));
+
 	ring->ready = false;
 }
 
@@ -2722,6 +2743,11 @@ int r600_uvd_init(struct radeon_device *rdev)
 	/* disable interupt */
 	WREG32_P(UVD_MASTINT_EN, 0, ~(1 << 1));
 
+	/* Stall UMC and register bus before resetting VCPU */
+	WREG32_P(UVD_LMI_CTRL2, 1 << 8, ~(1 << 8));
+	WREG32_P(UVD_RB_ARB_CTRL, 1 << 3, ~(1 << 3));
+	mdelay(1);
+
 	/* put LMI, VCPU, RBC etc... into reset */
 	WREG32(UVD_SOFT_RESET, LMI_SOFT_RESET | VCPU_SOFT_RESET |
 	       LBSI_SOFT_RESET | RBC_SOFT_RESET | CSM_SOFT_RESET |
@@ -2751,10 +2777,6 @@ int r600_uvd_init(struct radeon_device *rdev)
 	WREG32(UVD_MPC_SET_ALU, 0);
 	WREG32(UVD_MPC_SET_MUX, 0x88);
 
-	/* Stall UMC */
-	WREG32_P(UVD_LMI_CTRL2, 1 << 8, ~(1 << 8));
-	WREG32_P(UVD_RB_ARB_CTRL, 1 << 3, ~(1 << 3));
-
 	/* take all subblocks out of reset, except VCPU */
 	WREG32(UVD_SOFT_RESET, VCPU_SOFT_RESET);
 	mdelay(5);
@@ -3312,6 +3334,8 @@ static int r600_startup(struct radeon_device *rdev)
 	/* enable pcie gen2 link */
 	r600_pcie_gen2_enable(rdev);
 
+	r600_mc_program(rdev);
+
 	if (!rdev->me_fw || !rdev->pfp_fw || !rdev->rlc_fw) {
 		r = r600_init_microcode(rdev);
 		if (r) {
@@ -3324,7 +3348,6 @@ static int r600_startup(struct radeon_device *rdev)
 	if (r)
 		return r;
 
-	r600_mc_program(rdev);
 	if (rdev->flags & RADEON_IS_AGP) {
 		r600_agp_enable(rdev);
 	} else {
diff --git a/drivers/gpu/drm/radeon/r600_hdmi.c b/drivers/gpu/drm/radeon/r600_hdmi.c
index f48240bb8c5..f264df5470f 100644
--- a/drivers/gpu/drm/radeon/r600_hdmi.c
+++ b/drivers/gpu/drm/radeon/r600_hdmi.c
@@ -226,10 +226,29 @@ void r600_audio_set_dto(struct drm_encoder *encoder, u32 clock)
 	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
 	struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv;
 	u32 base_rate = 24000;
+	u32 max_ratio = clock / base_rate;
+	u32 dto_phase;
+	u32 dto_modulo = clock;
+	u32 wallclock_ratio;
+	u32 dto_cntl;
 
 	if (!dig || !dig->afmt)
 		return;
 
+	if (max_ratio >= 8) {
+		dto_phase = 192 * 1000;
+		wallclock_ratio = 3;
+	} else if (max_ratio >= 4) {
+		dto_phase = 96 * 1000;
+		wallclock_ratio = 2;
+	} else if (max_ratio >= 2) {
+		dto_phase = 48 * 1000;
+		wallclock_ratio = 1;
+	} else {
+		dto_phase = 24 * 1000;
+		wallclock_ratio = 0;
+	}
+
 	/* there are two DTOs selected by DCCG_AUDIO_DTO_SELECT.
 	 * doesn't matter which one you use.  Just use the first one.
 	 */
@@ -242,9 +261,21 @@ void r600_audio_set_dto(struct drm_encoder *encoder, u32 clock)
 		/* according to the reg specs, this should DCE3.2 only, but in
 		 * practice it seems to cover DCE3.0 as well.
 		 */
-		WREG32(DCCG_AUDIO_DTO0_PHASE, base_rate * 100);
-		WREG32(DCCG_AUDIO_DTO0_MODULE, clock * 100);
-		WREG32(DCCG_AUDIO_DTO_SELECT, 0); /* select DTO0 */
+		if (dig->dig_encoder == 0) {
+			dto_cntl = RREG32(DCCG_AUDIO_DTO0_CNTL) & ~DCCG_AUDIO_DTO_WALLCLOCK_RATIO_MASK;
+			dto_cntl |= DCCG_AUDIO_DTO_WALLCLOCK_RATIO(wallclock_ratio);
+			WREG32(DCCG_AUDIO_DTO0_CNTL, dto_cntl);
+			WREG32(DCCG_AUDIO_DTO0_PHASE, dto_phase);
+			WREG32(DCCG_AUDIO_DTO0_MODULE, dto_modulo);
+			WREG32(DCCG_AUDIO_DTO_SELECT, 0); /* select DTO0 */
+		} else {
+			dto_cntl = RREG32(DCCG_AUDIO_DTO1_CNTL) & ~DCCG_AUDIO_DTO_WALLCLOCK_RATIO_MASK;
+			dto_cntl |= DCCG_AUDIO_DTO_WALLCLOCK_RATIO(wallclock_ratio);
+			WREG32(DCCG_AUDIO_DTO1_CNTL, dto_cntl);
+			WREG32(DCCG_AUDIO_DTO1_PHASE, dto_phase);
+			WREG32(DCCG_AUDIO_DTO1_MODULE, dto_modulo);
+			WREG32(DCCG_AUDIO_DTO_SELECT, 1); /* select DTO1 */
+		}
 	} else {
 		/* according to the reg specs, this should be DCE2.0 and DCE3.0 */
 		WREG32(AUDIO_DTO, AUDIO_DTO_PHASE(base_rate / 10) |
diff --git a/drivers/gpu/drm/radeon/r600d.h b/drivers/gpu/drm/radeon/r600d.h
index 8e3fe815eda..7c780839a7f 100644
--- a/drivers/gpu/drm/radeon/r600d.h
+++ b/drivers/gpu/drm/radeon/r600d.h
@@ -933,6 +933,9 @@
 #define DCCG_AUDIO_DTO0_LOAD              0x051c
 #       define DTO_LOAD                   (1 << 31)
 #define DCCG_AUDIO_DTO0_CNTL              0x0520
+#       define DCCG_AUDIO_DTO_WALLCLOCK_RATIO(x) (((x) & 7) << 0)
+#       define DCCG_AUDIO_DTO_WALLCLOCK_RATIO_MASK 7
+#       define DCCG_AUDIO_DTO_WALLCLOCK_RATIO_SHIFT 0
 
 #define DCCG_AUDIO_DTO1_PHASE             0x0524
 #define DCCG_AUDIO_DTO1_MODULE            0x0528
diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h
index 2f08219c39b..274b8e1b889 100644
--- a/drivers/gpu/drm/radeon/radeon.h
+++ b/drivers/gpu/drm/radeon/radeon.h
@@ -1468,7 +1468,6 @@ struct radeon_uvd {
 	void			*cpu_addr;
 	uint64_t		gpu_addr;
 	void			*saved_bo;
-	unsigned		fw_size;
 	atomic_t		handles[RADEON_MAX_UVD_HANDLES];
 	struct drm_file		*filp[RADEON_MAX_UVD_HANDLES];
 	struct delayed_work	idle_work;
@@ -2066,6 +2065,7 @@ struct radeon_device {
 	const struct firmware *mec_fw;	/* CIK MEC firmware */
 	const struct firmware *sdma_fw;	/* CIK SDMA firmware */
 	const struct firmware *smc_fw;	/* SMC firmware */
+	const struct firmware *uvd_fw;	/* UVD firmware */
 	struct r600_blit r600_blit;
 	struct r600_vram_scratch vram_scratch;
 	int msi_enabled; /* msi enabled */
@@ -2095,6 +2095,8 @@ struct radeon_device {
 	/* ACPI interface */
 	struct radeon_atif		atif;
 	struct radeon_atcs		atcs;
+	/* srbm instance registers */
+	struct mutex			srbm_mutex;
 };
 
 int radeon_device_init(struct radeon_device *rdev,
diff --git a/drivers/gpu/drm/radeon/radeon_asic.h b/drivers/gpu/drm/radeon/radeon_asic.h
index 902479fa737..3d61d5aac18 100644
--- a/drivers/gpu/drm/radeon/radeon_asic.h
+++ b/drivers/gpu/drm/radeon/radeon_asic.h
@@ -441,7 +441,7 @@ void rs780_dpm_debugfs_print_current_performance_level(struct radeon_device *rde
 /* uvd */
 int r600_uvd_init(struct radeon_device *rdev);
 int r600_uvd_rbc_start(struct radeon_device *rdev);
-void r600_uvd_rbc_stop(struct radeon_device *rdev);
+void r600_uvd_stop(struct radeon_device *rdev);
 int r600_uvd_ib_test(struct radeon_device *rdev, struct radeon_ring *ring);
 void r600_uvd_fence_emit(struct radeon_device *rdev,
 			 struct radeon_fence *fence);
diff --git a/drivers/gpu/drm/radeon/radeon_device.c b/drivers/gpu/drm/radeon/radeon_device.c
index 82335e38ec4..63398ae1dbf 100644
--- a/drivers/gpu/drm/radeon/radeon_device.c
+++ b/drivers/gpu/drm/radeon/radeon_device.c
@@ -1163,6 +1163,7 @@ int radeon_device_init(struct radeon_device *rdev,
 	mutex_init(&rdev->gem.mutex);
 	mutex_init(&rdev->pm.mutex);
 	mutex_init(&rdev->gpu_clock_mutex);
+	mutex_init(&rdev->srbm_mutex);
 	init_rwsem(&rdev->pm.mclk_lock);
 	init_rwsem(&rdev->exclusive_lock);
 	init_waitqueue_head(&rdev->irq.vblank_queue);
@@ -1519,6 +1520,7 @@ int radeon_gpu_reset(struct radeon_device *rdev)
 	radeon_save_bios_scratch_regs(rdev);
 	/* block TTM */
 	resched = ttm_bo_lock_delayed_workqueue(&rdev->mman.bdev);
+	radeon_pm_suspend(rdev);
 	radeon_suspend(rdev);
 
 	for (i = 0; i < RADEON_NUM_RINGS; ++i) {
@@ -1564,6 +1566,7 @@ retry:
 		}
 	}
 
+	radeon_pm_resume(rdev);
 	drm_helper_resume_force_mode(rdev->ddev);
 
 	ttm_bo_unlock_delayed_workqueue(&rdev->mman.bdev, resched);
diff --git a/drivers/gpu/drm/radeon/radeon_fence.c b/drivers/gpu/drm/radeon/radeon_fence.c
index 7ddb0efe240..ddb8f8e04eb 100644
--- a/drivers/gpu/drm/radeon/radeon_fence.c
+++ b/drivers/gpu/drm/radeon/radeon_fence.c
@@ -782,7 +782,7 @@ int radeon_fence_driver_start_ring(struct radeon_device *rdev, int ring)
 
 		} else {
 			/* put fence directly behind firmware */
-			index = ALIGN(rdev->uvd.fw_size, 8);
+			index = ALIGN(rdev->uvd_fw->size, 8);
 			rdev->fence_drv[ring].cpu_addr = rdev->uvd.cpu_addr + index;
 			rdev->fence_drv[ring].gpu_addr = rdev->uvd.gpu_addr + index;
 		}
diff --git a/drivers/gpu/drm/radeon/radeon_gart.c b/drivers/gpu/drm/radeon/radeon_gart.c
index 6a51d943ccf..b990b1a2bd5 100644
--- a/drivers/gpu/drm/radeon/radeon_gart.c
+++ b/drivers/gpu/drm/radeon/radeon_gart.c
@@ -207,7 +207,6 @@ void radeon_gart_table_vram_free(struct radeon_device *rdev)
 	if (rdev->gart.robj == NULL) {
 		return;
 	}
-	radeon_gart_table_vram_unpin(rdev);
 	radeon_bo_unref(&rdev->gart.robj);
 }
 
diff --git a/drivers/gpu/drm/radeon/radeon_pm.c b/drivers/gpu/drm/radeon/radeon_pm.c
index f374c467aac..c557850cd34 100644
--- a/drivers/gpu/drm/radeon/radeon_pm.c
+++ b/drivers/gpu/drm/radeon/radeon_pm.c
@@ -1176,7 +1176,14 @@ int radeon_pm_init(struct radeon_device *rdev)
 	case CHIP_VERDE:
 	case CHIP_OLAND:
 	case CHIP_HAINAN:
-		if (radeon_dpm == 1)
+		/* DPM requires the RLC, RV770+ dGPU requires SMC */
+		if (!rdev->rlc_fw)
+			rdev->pm.pm_method = PM_METHOD_PROFILE;
+		else if ((rdev->family >= CHIP_RV770) &&
+			 (!(rdev->flags & RADEON_IS_IGP)) &&
+			 (!rdev->smc_fw))
+			rdev->pm.pm_method = PM_METHOD_PROFILE;
+		else if (radeon_dpm == 1)
 			rdev->pm.pm_method = PM_METHOD_DPM;
 		else
 			rdev->pm.pm_method = PM_METHOD_PROFILE;
diff --git a/drivers/gpu/drm/radeon/radeon_uvd.c b/drivers/gpu/drm/radeon/radeon_uvd.c
index 414fd145d20..f1c15754e73 100644
--- a/drivers/gpu/drm/radeon/radeon_uvd.c
+++ b/drivers/gpu/drm/radeon/radeon_uvd.c
@@ -56,7 +56,6 @@ static void radeon_uvd_idle_work_handler(struct work_struct *work);
 
 int radeon_uvd_init(struct radeon_device *rdev)
 {
-	const struct firmware *fw;
 	unsigned long bo_size;
 	const char *fw_name;
 	int i, r;
@@ -105,14 +104,14 @@ int radeon_uvd_init(struct radeon_device *rdev)
 		return -EINVAL;
 	}
 
-	r = request_firmware(&fw, fw_name, rdev->dev);
+	r = request_firmware(&rdev->uvd_fw, fw_name, rdev->dev);
 	if (r) {
 		dev_err(rdev->dev, "radeon_uvd: Can't load firmware \"%s\"\n",
 			fw_name);
 		return r;
 	}
 
-	bo_size = RADEON_GPU_PAGE_ALIGN(fw->size + 8) +
+	bo_size = RADEON_GPU_PAGE_ALIGN(rdev->uvd_fw->size + 8) +
 		  RADEON_UVD_STACK_SIZE + RADEON_UVD_HEAP_SIZE;
 	r = radeon_bo_create(rdev, bo_size, PAGE_SIZE, true,
 			     RADEON_GEM_DOMAIN_VRAM, NULL, &rdev->uvd.vcpu_bo);
@@ -145,12 +144,6 @@ int radeon_uvd_init(struct radeon_device *rdev)
 
 	radeon_bo_unreserve(rdev->uvd.vcpu_bo);
 
-	rdev->uvd.fw_size = fw->size;
-	memset(rdev->uvd.cpu_addr, 0, bo_size);
-	memcpy(rdev->uvd.cpu_addr, fw->data, fw->size);
-
-	release_firmware(fw);
-
 	for (i = 0; i < RADEON_MAX_UVD_HANDLES; ++i) {
 		atomic_set(&rdev->uvd.handles[i], 0);
 		rdev->uvd.filp[i] = NULL;
@@ -174,33 +167,60 @@ void radeon_uvd_fini(struct radeon_device *rdev)
 	}
 
 	radeon_bo_unref(&rdev->uvd.vcpu_bo);
+
+	release_firmware(rdev->uvd_fw);
 }
 
 int radeon_uvd_suspend(struct radeon_device *rdev)
 {
 	unsigned size;
+	void *ptr;
+	int i;
 
 	if (rdev->uvd.vcpu_bo == NULL)
 		return 0;
 
+	for (i = 0; i < RADEON_MAX_UVD_HANDLES; ++i)
+		if (atomic_read(&rdev->uvd.handles[i]))
+			break;
+
+	if (i == RADEON_MAX_UVD_HANDLES)
+		return 0;
+
 	size = radeon_bo_size(rdev->uvd.vcpu_bo);
+	size -= rdev->uvd_fw->size;
+
+	ptr = rdev->uvd.cpu_addr;
+	ptr += rdev->uvd_fw->size;
+
 	rdev->uvd.saved_bo = kmalloc(size, GFP_KERNEL);
-	memcpy(rdev->uvd.saved_bo, rdev->uvd.cpu_addr, size);
+	memcpy(rdev->uvd.saved_bo, ptr, size);
 
 	return 0;
 }
 
 int radeon_uvd_resume(struct radeon_device *rdev)
 {
+	unsigned size;
+	void *ptr;
+
 	if (rdev->uvd.vcpu_bo == NULL)
 		return -EINVAL;
 
+	memcpy(rdev->uvd.cpu_addr, rdev->uvd_fw->data, rdev->uvd_fw->size);
+
+	size = radeon_bo_size(rdev->uvd.vcpu_bo);
+	size -= rdev->uvd_fw->size;
+
+	ptr = rdev->uvd.cpu_addr;
+	ptr += rdev->uvd_fw->size;
+
 	if (rdev->uvd.saved_bo != NULL) {
-		unsigned size = radeon_bo_size(rdev->uvd.vcpu_bo);
-		memcpy(rdev->uvd.cpu_addr, rdev->uvd.saved_bo, size);
+		memcpy(ptr, rdev->uvd.saved_bo, size);
 		kfree(rdev->uvd.saved_bo);
 		rdev->uvd.saved_bo = NULL;
-	}
+	} else
+		memset(ptr, 0, size);
 
 	return 0;
 }
@@ -215,8 +235,8 @@ void radeon_uvd_free_handles(struct radeon_device *rdev, struct drm_file *filp)
 {
 	int i, r;
 	for (i = 0; i < RADEON_MAX_UVD_HANDLES; ++i) {
-		if (rdev->uvd.filp[i] == filp) {
-			uint32_t handle = atomic_read(&rdev->uvd.handles[i]);
+		uint32_t handle = atomic_read(&rdev->uvd.handles[i]);
+		if (handle != 0 && rdev->uvd.filp[i] == filp) {
 			struct radeon_fence *fence;
 
 			r = radeon_uvd_get_destroy_msg(rdev,
@@ -337,8 +357,10 @@ static int radeon_uvd_cs_msg(struct radeon_cs_parser *p, struct radeon_bo *bo,
 	}
 
 	r = radeon_bo_kmap(bo, &ptr);
-	if (r)
+	if (r) {
+		DRM_ERROR("Failed mapping the UVD message (%d)!\n", r);
 		return r;
+	}
 
 	msg = ptr + offset;
 
@@ -364,8 +386,14 @@ static int radeon_uvd_cs_msg(struct radeon_cs_parser *p, struct radeon_bo *bo,
 		radeon_bo_kunmap(bo);
 		return 0;
 	} else {
-		/* it's a create msg, no special handling needed */
 		radeon_bo_kunmap(bo);
+
+		if (msg_type != 0) {
+			DRM_ERROR("Illegal UVD message type (%d)!\n", msg_type);
+			return -EINVAL;
+		}
+
+		/* it's a create msg, no special handling needed */
 	}
 
 	/* create or decode, validate the handle */
@@ -388,7 +416,7 @@ static int radeon_uvd_cs_msg(struct radeon_cs_parser *p, struct radeon_bo *bo,
 
 static int radeon_uvd_cs_reloc(struct radeon_cs_parser *p,
 			       int data0, int data1,
-			       unsigned buf_sizes[])
+			       unsigned buf_sizes[], bool *has_msg_cmd)
 {
 	struct radeon_cs_chunk *relocs_chunk;
 	struct radeon_cs_reloc *reloc;
@@ -417,7 +445,7 @@ static int radeon_uvd_cs_reloc(struct radeon_cs_parser *p,
 
 	if (cmd < 0x4) {
 		if ((end - start) < buf_sizes[cmd]) {
-			DRM_ERROR("buffer to small (%d / %d)!\n",
+			DRM_ERROR("buffer (%d) to small (%d / %d)!\n", cmd,
 				  (unsigned)(end - start), buf_sizes[cmd]);
 			return -EINVAL;
 		}
@@ -442,9 +470,17 @@ static int radeon_uvd_cs_reloc(struct radeon_cs_parser *p,
 	}
 
 	if (cmd == 0) {
+		if (*has_msg_cmd) {
+			DRM_ERROR("More than one message in a UVD-IB!\n");
+			return -EINVAL;
+		}
+		*has_msg_cmd = true;
 		r = radeon_uvd_cs_msg(p, reloc->robj, offset, buf_sizes);
 		if (r)
 			return r;
+	} else if (!*has_msg_cmd) {
+		DRM_ERROR("Message needed before other commands are send!\n");
+		return -EINVAL;
 	}
 
 	return 0;
@@ -453,7 +489,8 @@ static int radeon_uvd_cs_reloc(struct radeon_cs_parser *p,
 static int radeon_uvd_cs_reg(struct radeon_cs_parser *p,
 			     struct radeon_cs_packet *pkt,
 			     int *data0, int *data1,
-			     unsigned buf_sizes[])
+			     unsigned buf_sizes[],
+			     bool *has_msg_cmd)
 {
 	int i, r;
 
@@ -467,7 +504,8 @@ static int radeon_uvd_cs_reg(struct radeon_cs_parser *p,
 			*data1 = p->idx;
 			break;
 		case UVD_GPCOM_VCPU_CMD:
-			r = radeon_uvd_cs_reloc(p, *data0, *data1, buf_sizes);
+			r = radeon_uvd_cs_reloc(p, *data0, *data1,
+						buf_sizes, has_msg_cmd);
 			if (r)
 				return r;
 			break;
@@ -488,6 +526,9 @@ int radeon_uvd_cs_parse(struct radeon_cs_parser *p)
 	struct radeon_cs_packet pkt;
 	int r, data0 = 0, data1 = 0;
 
+	/* does the IB has a msg command */
+	bool has_msg_cmd = false;
+
 	/* minimum buffer sizes */
 	unsigned buf_sizes[] = {
 		[0x00000000]	=	2048,
@@ -514,8 +555,8 @@ int radeon_uvd_cs_parse(struct radeon_cs_parser *p)
 			return r;
 		switch (pkt.type) {
 		case RADEON_PACKET_TYPE0:
-			r = radeon_uvd_cs_reg(p, &pkt, &data0,
-					      &data1, buf_sizes);
+			r = radeon_uvd_cs_reg(p, &pkt, &data0, &data1,
+					      buf_sizes, &has_msg_cmd);
 			if (r)
 				return r;
 			break;
@@ -527,6 +568,12 @@ int radeon_uvd_cs_parse(struct radeon_cs_parser *p)
 			return -EINVAL;
 		}
 	} while (p->idx < p->chunks[p->chunk_ib_idx].length_dw);
+
+	if (!has_msg_cmd) {
+		DRM_ERROR("UVD-IBs need a msg command!\n");
+		return -EINVAL;
+	}
+
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/radeon/rv6xx_dpm.c b/drivers/gpu/drm/radeon/rv6xx_dpm.c
index 363018c6041..bdd888b4db2 100644
--- a/drivers/gpu/drm/radeon/rv6xx_dpm.c
+++ b/drivers/gpu/drm/radeon/rv6xx_dpm.c
@@ -1944,9 +1944,7 @@ static int rv6xx_parse_power_table(struct radeon_device *rdev)
 
 int rv6xx_dpm_init(struct radeon_device *rdev)
 {
-	int index = GetIndexIntoMasterTable(DATA, ASIC_InternalSS_Info);
-	uint16_t data_offset, size;
-	uint8_t frev, crev;
+	struct radeon_atom_ss ss;
 	struct atom_clock_dividers dividers;
 	struct rv6xx_power_info *pi;
 	int ret;
@@ -1989,16 +1987,18 @@ int rv6xx_dpm_init(struct radeon_device *rdev)
 
 	pi->gfx_clock_gating = true;
 
-	if (atom_parse_data_header(rdev->mode_info.atom_context, index, &size,
-                                   &frev, &crev, &data_offset)) {
-		pi->sclk_ss = true;
-		pi->mclk_ss = true;
+	pi->sclk_ss = radeon_atombios_get_asic_ss_info(rdev, &ss,
+						       ASIC_INTERNAL_ENGINE_SS, 0);
+	pi->mclk_ss = radeon_atombios_get_asic_ss_info(rdev, &ss,
+						       ASIC_INTERNAL_MEMORY_SS, 0);
+
+	/* Disable sclk ss, causes hangs on a lot of systems */
+	pi->sclk_ss = false;
+
+	if (pi->sclk_ss || pi->mclk_ss)
 		pi->dynamic_ss = true;
-	} else {
-		pi->sclk_ss = false;
-		pi->mclk_ss = false;
+	else
 		pi->dynamic_ss = false;
-	}
 
 	pi->dynamic_pcie_gen2 = true;
 
diff --git a/drivers/gpu/drm/radeon/rv770.c b/drivers/gpu/drm/radeon/rv770.c
index 30ea14e8854..bcc68ec204a 100644
--- a/drivers/gpu/drm/radeon/rv770.c
+++ b/drivers/gpu/drm/radeon/rv770.c
@@ -813,7 +813,7 @@ int rv770_uvd_resume(struct radeon_device *rdev)
 
 	/* programm the VCPU memory controller bits 0-27 */
 	addr = rdev->uvd.gpu_addr >> 3;
-	size = RADEON_GPU_PAGE_ALIGN(rdev->uvd.fw_size + 4) >> 3;
+	size = RADEON_GPU_PAGE_ALIGN(rdev->uvd_fw->size + 4) >> 3;
 	WREG32(UVD_VCPU_CACHE_OFFSET0, addr);
 	WREG32(UVD_VCPU_CACHE_SIZE0, size);
 
@@ -1829,6 +1829,8 @@ static int rv770_startup(struct radeon_device *rdev)
 	/* enable pcie gen2 link */
 	rv770_pcie_gen2_enable(rdev);
 
+	rv770_mc_program(rdev);
+
 	if (!rdev->me_fw || !rdev->pfp_fw || !rdev->rlc_fw) {
 		r = r600_init_microcode(rdev);
 		if (r) {
@@ -1841,7 +1843,6 @@ static int rv770_startup(struct radeon_device *rdev)
 	if (r)
 		return r;
 
-	rv770_mc_program(rdev);
 	if (rdev->flags & RADEON_IS_AGP) {
 		rv770_agp_enable(rdev);
 	} else {
@@ -1983,6 +1984,7 @@ int rv770_resume(struct radeon_device *rdev)
 int rv770_suspend(struct radeon_device *rdev)
 {
 	r600_audio_fini(rdev);
+	r600_uvd_stop(rdev);
 	radeon_uvd_suspend(rdev);
 	r700_cp_stop(rdev);
 	r600_dma_stop(rdev);
@@ -2098,6 +2100,7 @@ void rv770_fini(struct radeon_device *rdev)
 	radeon_ib_pool_fini(rdev);
 	radeon_irq_kms_fini(rdev);
 	rv770_pcie_gart_fini(rdev);
+	r600_uvd_stop(rdev);
 	radeon_uvd_fini(rdev);
 	r600_vram_scratch_fini(rdev);
 	radeon_gem_fini(rdev);
diff --git a/drivers/gpu/drm/radeon/rv770_dpm.c b/drivers/gpu/drm/radeon/rv770_dpm.c
index 2d347925f77..094c67a29d0 100644
--- a/drivers/gpu/drm/radeon/rv770_dpm.c
+++ b/drivers/gpu/drm/radeon/rv770_dpm.c
@@ -2319,12 +2319,25 @@ int rv7xx_parse_power_table(struct radeon_device *rdev)
 	return 0;
 }
 
+void rv770_get_engine_memory_ss(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct radeon_atom_ss ss;
+
+	pi->sclk_ss = radeon_atombios_get_asic_ss_info(rdev, &ss,
+						       ASIC_INTERNAL_ENGINE_SS, 0);
+	pi->mclk_ss = radeon_atombios_get_asic_ss_info(rdev, &ss,
+						       ASIC_INTERNAL_MEMORY_SS, 0);
+
+	if (pi->sclk_ss || pi->mclk_ss)
+		pi->dynamic_ss = true;
+	else
+		pi->dynamic_ss = false;
+}
+
 int rv770_dpm_init(struct radeon_device *rdev)
 {
 	struct rv7xx_power_info *pi;
-	int index = GetIndexIntoMasterTable(DATA, ASIC_InternalSS_Info);
-	uint16_t data_offset, size;
-	uint8_t frev, crev;
 	struct atom_clock_dividers dividers;
 	int ret;
 
@@ -2369,16 +2382,7 @@ int rv770_dpm_init(struct radeon_device *rdev)
 	pi->mvdd_control =
 		radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_MVDDC, 0);
 
-	if (atom_parse_data_header(rdev->mode_info.atom_context, index, &size,
-                                   &frev, &crev, &data_offset)) {
-		pi->sclk_ss = true;
-		pi->mclk_ss = true;
-		pi->dynamic_ss = true;
-	} else {
-		pi->sclk_ss = false;
-		pi->mclk_ss = false;
-		pi->dynamic_ss = false;
-	}
+	rv770_get_engine_memory_ss(rdev);
 
 	pi->asi = RV770_ASI_DFLT;
 	pi->pasi = RV770_HASI_DFLT;
@@ -2393,8 +2397,7 @@ int rv770_dpm_init(struct radeon_device *rdev)
 
 	pi->dynamic_pcie_gen2 = true;
 
-	if (pi->gfx_clock_gating &&
-	    (rdev->pm.int_thermal_type != THERMAL_TYPE_NONE))
+	if (rdev->pm.int_thermal_type != THERMAL_TYPE_NONE)
 		pi->thermal_protection = true;
 	else
 		pi->thermal_protection = false;
diff --git a/drivers/gpu/drm/radeon/rv770_dpm.h b/drivers/gpu/drm/radeon/rv770_dpm.h
index 96b1b2a62a8..9244effc6b5 100644
--- a/drivers/gpu/drm/radeon/rv770_dpm.h
+++ b/drivers/gpu/drm/radeon/rv770_dpm.h
@@ -275,6 +275,7 @@ void rv770_set_uvd_clock_before_set_eng_clock(struct radeon_device *rdev,
 void rv770_set_uvd_clock_after_set_eng_clock(struct radeon_device *rdev,
 					     struct radeon_ps *new_ps,
 					     struct radeon_ps *old_ps);
+void rv770_get_engine_memory_ss(struct radeon_device *rdev);
 
 /* smc */
 int rv770_read_smc_soft_register(struct radeon_device *rdev,
diff --git a/drivers/gpu/drm/radeon/si.c b/drivers/gpu/drm/radeon/si.c
index 6ca904673a4..daa8d2df8ec 100644
--- a/drivers/gpu/drm/radeon/si.c
+++ b/drivers/gpu/drm/radeon/si.c
@@ -1663,9 +1663,13 @@ static int si_init_microcode(struct radeon_device *rdev)
 
 	snprintf(fw_name, sizeof(fw_name), "radeon/%s_smc.bin", chip_name);
 	err = request_firmware(&rdev->smc_fw, fw_name, rdev->dev);
-	if (err)
-		goto out;
-	if (rdev->smc_fw->size != smc_req_size) {
+	if (err) {
+		printk(KERN_ERR
+		       "smc: error loading firmware \"%s\"\n",
+		       fw_name);
+		release_firmware(rdev->smc_fw);
+		rdev->smc_fw = NULL;
+	} else if (rdev->smc_fw->size != smc_req_size) {
 		printk(KERN_ERR
 		       "si_smc: Bogus length %zu in firmware \"%s\"\n",
 		       rdev->smc_fw->size, fw_name);
@@ -6418,6 +6422,8 @@ static int si_startup(struct radeon_device *rdev)
 	/* enable aspm */
 	si_program_aspm(rdev);
 
+	si_mc_program(rdev);
+
 	if (!rdev->me_fw || !rdev->pfp_fw || !rdev->ce_fw ||
 	    !rdev->rlc_fw || !rdev->mc_fw) {
 		r = si_init_microcode(rdev);
@@ -6437,7 +6443,6 @@ static int si_startup(struct radeon_device *rdev)
 	if (r)
 		return r;
 
-	si_mc_program(rdev);
 	r = si_pcie_gart_enable(rdev);
 	if (r)
 		return r;
@@ -6621,7 +6626,7 @@ int si_suspend(struct radeon_device *rdev)
 	si_cp_enable(rdev, false);
 	cayman_dma_stop(rdev);
 	if (rdev->has_uvd) {
-		r600_uvd_rbc_stop(rdev);
+		r600_uvd_stop(rdev);
 		radeon_uvd_suspend(rdev);
 	}
 	si_irq_suspend(rdev);
@@ -6763,8 +6768,10 @@ void si_fini(struct radeon_device *rdev)
 	radeon_vm_manager_fini(rdev);
 	radeon_ib_pool_fini(rdev);
 	radeon_irq_kms_fini(rdev);
-	if (rdev->has_uvd)
+	if (rdev->has_uvd) {
+		r600_uvd_stop(rdev);
 		radeon_uvd_fini(rdev);
+	}
 	si_pcie_gart_fini(rdev);
 	r600_vram_scratch_fini(rdev);
 	radeon_gem_fini(rdev);
diff --git a/drivers/gpu/drm/radeon/si_dpm.c b/drivers/gpu/drm/radeon/si_dpm.c
index 41825575b40..88699e3cd86 100644
--- a/drivers/gpu/drm/radeon/si_dpm.c
+++ b/drivers/gpu/drm/radeon/si_dpm.c
@@ -2903,7 +2903,8 @@ static void si_apply_state_adjust_rules(struct radeon_device *rdev,
 {
 	struct ni_ps *ps = ni_get_ps(rps);
 	struct radeon_clock_and_voltage_limits *max_limits;
-	bool disable_mclk_switching;
+	bool disable_mclk_switching = false;
+	bool disable_sclk_switching = false;
 	u32 mclk, sclk;
 	u16 vddc, vddci;
 	int i;
@@ -2911,8 +2912,11 @@ static void si_apply_state_adjust_rules(struct radeon_device *rdev,
 	if ((rdev->pm.dpm.new_active_crtc_count > 1) ||
 	    ni_dpm_vblank_too_short(rdev))
 		disable_mclk_switching = true;
-	else
-		disable_mclk_switching = false;
+
+	if (rps->vclk || rps->dclk) {
+		disable_mclk_switching = true;
+		disable_sclk_switching = true;
+	}
 
 	if (rdev->pm.dpm.ac_power)
 		max_limits = &rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac;
@@ -2940,27 +2944,43 @@ static void si_apply_state_adjust_rules(struct radeon_device *rdev,
 
 	if (disable_mclk_switching) {
 		mclk  = ps->performance_levels[ps->performance_level_count - 1].mclk;
-		sclk = ps->performance_levels[0].sclk;
-		vddc = ps->performance_levels[0].vddc;
 		vddci = ps->performance_levels[ps->performance_level_count - 1].vddci;
 	} else {
-		sclk = ps->performance_levels[0].sclk;
 		mclk = ps->performance_levels[0].mclk;
-		vddc = ps->performance_levels[0].vddc;
 		vddci = ps->performance_levels[0].vddci;
 	}
 
+	if (disable_sclk_switching) {
+		sclk = ps->performance_levels[ps->performance_level_count - 1].sclk;
+		vddc = ps->performance_levels[ps->performance_level_count - 1].vddc;
+	} else {
+		sclk = ps->performance_levels[0].sclk;
+		vddc = ps->performance_levels[0].vddc;
+	}
+
 	/* adjusted low state */
 	ps->performance_levels[0].sclk = sclk;
 	ps->performance_levels[0].mclk = mclk;
 	ps->performance_levels[0].vddc = vddc;
 	ps->performance_levels[0].vddci = vddci;
 
-	for (i = 1; i < ps->performance_level_count; i++) {
-		if (ps->performance_levels[i].sclk < ps->performance_levels[i - 1].sclk)
-			ps->performance_levels[i].sclk = ps->performance_levels[i - 1].sclk;
-		if (ps->performance_levels[i].vddc < ps->performance_levels[i - 1].vddc)
-			ps->performance_levels[i].vddc = ps->performance_levels[i - 1].vddc;
+	if (disable_sclk_switching) {
+		sclk = ps->performance_levels[0].sclk;
+		for (i = 1; i < ps->performance_level_count; i++) {
+			if (sclk < ps->performance_levels[i].sclk)
+				sclk = ps->performance_levels[i].sclk;
+		}
+		for (i = 0; i < ps->performance_level_count; i++) {
+			ps->performance_levels[i].sclk = sclk;
+			ps->performance_levels[i].vddc = vddc;
+		}
+	} else {
+		for (i = 1; i < ps->performance_level_count; i++) {
+			if (ps->performance_levels[i].sclk < ps->performance_levels[i - 1].sclk)
+				ps->performance_levels[i].sclk = ps->performance_levels[i - 1].sclk;
+			if (ps->performance_levels[i].vddc < ps->performance_levels[i - 1].vddc)
+				ps->performance_levels[i].vddc = ps->performance_levels[i - 1].vddc;
+		}
 	}
 
 	if (disable_mclk_switching) {
@@ -6253,9 +6273,6 @@ int si_dpm_init(struct radeon_device *rdev)
 	struct evergreen_power_info *eg_pi;
 	struct ni_power_info *ni_pi;
 	struct si_power_info *si_pi;
-	int index = GetIndexIntoMasterTable(DATA, ASIC_InternalSS_Info);
-	u16 data_offset, size;
-	u8 frev, crev;
 	struct atom_clock_dividers dividers;
 	int ret;
 	u32 mask;
@@ -6346,16 +6363,7 @@ int si_dpm_init(struct radeon_device *rdev)
 	si_pi->vddc_phase_shed_control =
 		radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_VDDC, VOLTAGE_OBJ_PHASE_LUT);
 
-	if (atom_parse_data_header(rdev->mode_info.atom_context, index, &size,
-                                   &frev, &crev, &data_offset)) {
-		pi->sclk_ss = true;
-		pi->mclk_ss = true;
-		pi->dynamic_ss = true;
-	} else {
-		pi->sclk_ss = false;
-		pi->mclk_ss = false;
-		pi->dynamic_ss = true;
-	}
+	rv770_get_engine_memory_ss(rdev);
 
 	pi->asi = RV770_ASI_DFLT;
 	pi->pasi = CYPRESS_HASI_DFLT;
@@ -6366,8 +6374,7 @@ int si_dpm_init(struct radeon_device *rdev)
 	eg_pi->sclk_deep_sleep = true;
 	si_pi->sclk_deep_sleep_above_low = false;
 
-	if (pi->gfx_clock_gating &&
-	    (rdev->pm.int_thermal_type != THERMAL_TYPE_NONE))
+	if (rdev->pm.int_thermal_type != THERMAL_TYPE_NONE)
 		pi->thermal_protection = true;
 	else
 		pi->thermal_protection = false;