Merge branch 'next' into for-linus

Prepare first round of input updates for 3.18.

66 files changed, 3452 insertions, 1675 deletions

diff --git a/drivers/gpu/drm/radeon/Makefile b/drivers/gpu/drm/radeon/Makefile
index dbcbfe80aac0..f77b7135ee4c 100644
--- a/drivers/gpu/drm/radeon/Makefile
+++ b/drivers/gpu/drm/radeon/Makefile
@@ -76,11 +76,11 @@ radeon-y += radeon_device.o radeon_asic.o radeon_kms.o \
 	evergreen.o evergreen_cs.o evergreen_blit_shaders.o \
 	evergreen_hdmi.o radeon_trace_points.o ni.o cayman_blit_shaders.o \
 	atombios_encoders.o radeon_semaphore.o radeon_sa.o atombios_i2c.o si.o \
-	si_blit_shaders.o radeon_prime.o radeon_uvd.o cik.o cik_blit_shaders.o \
+	si_blit_shaders.o radeon_prime.o cik.o cik_blit_shaders.o \
 	r600_dpm.o rs780_dpm.o rv6xx_dpm.o rv770_dpm.o rv730_dpm.o rv740_dpm.o \
 	rv770_smc.o cypress_dpm.o btc_dpm.o sumo_dpm.o sumo_smc.o trinity_dpm.o \
 	trinity_smc.o ni_dpm.o si_smc.o si_dpm.o kv_smc.o kv_dpm.o ci_smc.o \
-	ci_dpm.o dce6_afmt.o radeon_vm.o
+	ci_dpm.o dce6_afmt.o radeon_vm.o radeon_ucode.o radeon_ib.o
 
 # add async DMA block
 radeon-y += \
diff --git a/drivers/gpu/drm/radeon/atombios_encoders.c b/drivers/gpu/drm/radeon/atombios_encoders.c
index 7d68203a3737..a7f2ddf09a9d 100644
--- a/drivers/gpu/drm/radeon/atombios_encoders.c
+++ b/drivers/gpu/drm/radeon/atombios_encoders.c
@@ -331,12 +331,10 @@ static bool radeon_atom_mode_fixup(struct drm_encoder *encoder,
 	    && (mode->crtc_vsync_start < (mode->crtc_vdisplay + 2)))
 		adjusted_mode->crtc_vsync_start = adjusted_mode->crtc_vdisplay + 2;
 
-	/* get the native mode for LVDS */
-	if (radeon_encoder->active_device & (ATOM_DEVICE_LCD_SUPPORT))
+	/* get the native mode for scaling */
+	if (radeon_encoder->active_device & (ATOM_DEVICE_LCD_SUPPORT)) {
 		radeon_panel_mode_fixup(encoder, adjusted_mode);
-
-	/* get the native mode for TV */
-	if (radeon_encoder->active_device & (ATOM_DEVICE_TV_SUPPORT)) {
+	} else if (radeon_encoder->active_device & (ATOM_DEVICE_TV_SUPPORT)) {
 		struct radeon_encoder_atom_dac *tv_dac = radeon_encoder->enc_priv;
 		if (tv_dac) {
 			if (tv_dac->tv_std == TV_STD_NTSC ||
@@ -346,6 +344,8 @@ static bool radeon_atom_mode_fixup(struct drm_encoder *encoder,
 			else
 				radeon_atom_get_tv_timings(rdev, 1, adjusted_mode);
 		}
+	} else if (radeon_encoder->rmx_type != RMX_OFF) {
+		radeon_panel_mode_fixup(encoder, adjusted_mode);
 	}
 
 	if (ASIC_IS_DCE3(rdev) &&
@@ -716,7 +716,7 @@ atombios_get_encoder_mode(struct drm_encoder *encoder)
 			if (radeon_connector->use_digital &&
 			    (radeon_connector->audio == RADEON_AUDIO_ENABLE))
 				return ATOM_ENCODER_MODE_HDMI;
-			else if (drm_detect_hdmi_monitor(radeon_connector->edid) &&
+			else if (drm_detect_hdmi_monitor(radeon_connector_edid(connector)) &&
 				 (radeon_connector->audio == RADEON_AUDIO_AUTO))
 				return ATOM_ENCODER_MODE_HDMI;
 			else if (radeon_connector->use_digital)
@@ -735,7 +735,7 @@ atombios_get_encoder_mode(struct drm_encoder *encoder)
 		if (radeon_audio != 0) {
 			if (radeon_connector->audio == RADEON_AUDIO_ENABLE)
 				return ATOM_ENCODER_MODE_HDMI;
-			else if (drm_detect_hdmi_monitor(radeon_connector->edid) &&
+			else if (drm_detect_hdmi_monitor(radeon_connector_edid(connector)) &&
 				 (radeon_connector->audio == RADEON_AUDIO_AUTO))
 				return ATOM_ENCODER_MODE_HDMI;
 			else
@@ -755,7 +755,7 @@ atombios_get_encoder_mode(struct drm_encoder *encoder)
 		} else if (radeon_audio != 0) {
 			if (radeon_connector->audio == RADEON_AUDIO_ENABLE)
 				return ATOM_ENCODER_MODE_HDMI;
-			else if (drm_detect_hdmi_monitor(radeon_connector->edid) &&
+			else if (drm_detect_hdmi_monitor(radeon_connector_edid(connector)) &&
 				 (radeon_connector->audio == RADEON_AUDIO_AUTO))
 				return ATOM_ENCODER_MODE_HDMI;
 			else
diff --git a/drivers/gpu/drm/radeon/ci_dpm.c b/drivers/gpu/drm/radeon/ci_dpm.c
index 584090ac3eb9..d416bb2ff48d 100644
--- a/drivers/gpu/drm/radeon/ci_dpm.c
+++ b/drivers/gpu/drm/radeon/ci_dpm.c
@@ -869,6 +869,9 @@ static int ci_set_thermal_temperature_range(struct radeon_device *rdev,
 	WREG32_SMC(CG_THERMAL_CTRL, tmp);
 #endif
 
+	rdev->pm.dpm.thermal.min_temp = low_temp;
+	rdev->pm.dpm.thermal.max_temp = high_temp;
+
 	return 0;
 }
 
@@ -940,7 +943,18 @@ static void ci_get_leakage_voltages(struct radeon_device *rdev)
 	pi->vddc_leakage.count = 0;
 	pi->vddci_leakage.count = 0;
 
-	if (radeon_atom_get_leakage_id_from_vbios(rdev, &leakage_id) == 0) {
+	if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_EVV) {
+		for (i = 0; i < CISLANDS_MAX_LEAKAGE_COUNT; i++) {
+			virtual_voltage_id = ATOM_VIRTUAL_VOLTAGE_ID0 + i;
+			if (radeon_atom_get_voltage_evv(rdev, virtual_voltage_id, &vddc) != 0)
+				continue;
+			if (vddc != 0 && vddc != virtual_voltage_id) {
+				pi->vddc_leakage.actual_voltage[pi->vddc_leakage.count] = vddc;
+				pi->vddc_leakage.leakage_id[pi->vddc_leakage.count] = virtual_voltage_id;
+				pi->vddc_leakage.count++;
+			}
+		}
+	} else if (radeon_atom_get_leakage_id_from_vbios(rdev, &leakage_id) == 0) {
 		for (i = 0; i < CISLANDS_MAX_LEAKAGE_COUNT; i++) {
 			virtual_voltage_id = ATOM_VIRTUAL_VOLTAGE_ID0 + i;
 			if (radeon_atom_get_leakage_vddc_based_on_leakage_params(rdev, &vddc, &vddci,
diff --git a/drivers/gpu/drm/radeon/ci_smc.c b/drivers/gpu/drm/radeon/ci_smc.c
index 8debc9d47362..b630edc2fd0c 100644
--- a/drivers/gpu/drm/radeon/ci_smc.c
+++ b/drivers/gpu/drm/radeon/ci_smc.c
@@ -213,24 +213,37 @@ int ci_load_smc_ucode(struct radeon_device *rdev, u32 limit)
 	if (!rdev->smc_fw)
 		return -EINVAL;
 
-	switch (rdev->family) {
-	case CHIP_BONAIRE:
-		ucode_start_address = BONAIRE_SMC_UCODE_START;
-		ucode_size = BONAIRE_SMC_UCODE_SIZE;
-		break;
-	case CHIP_HAWAII:
-		ucode_start_address = HAWAII_SMC_UCODE_START;
-		ucode_size = HAWAII_SMC_UCODE_SIZE;
-		break;
-	default:
-		DRM_ERROR("unknown asic in smc ucode loader\n");
-		BUG();
+	if (rdev->new_fw) {
+		const struct smc_firmware_header_v1_0 *hdr =
+			(const struct smc_firmware_header_v1_0 *)rdev->smc_fw->data;
+
+		radeon_ucode_print_smc_hdr(&hdr->header);
+
+		ucode_start_address = le32_to_cpu(hdr->ucode_start_addr);
+		ucode_size = le32_to_cpu(hdr->header.ucode_size_bytes);
+		src = (const u8 *)
+			(rdev->smc_fw->data + le32_to_cpu(hdr->header.ucode_array_offset_bytes));
+	} else {
+		switch (rdev->family) {
+		case CHIP_BONAIRE:
+			ucode_start_address = BONAIRE_SMC_UCODE_START;
+			ucode_size = BONAIRE_SMC_UCODE_SIZE;
+			break;
+		case CHIP_HAWAII:
+			ucode_start_address = HAWAII_SMC_UCODE_START;
+			ucode_size = HAWAII_SMC_UCODE_SIZE;
+			break;
+		default:
+			DRM_ERROR("unknown asic in smc ucode loader\n");
+			BUG();
+		}
+
+		src = (const u8 *)rdev->smc_fw->data;
 	}
 
 	if (ucode_size & 3)
 		return -EINVAL;
 
-	src = (const u8 *)rdev->smc_fw->data;
 	spin_lock_irqsave(&rdev->smc_idx_lock, flags);
 	WREG32(SMC_IND_INDEX_0, ucode_start_address);
 	WREG32_P(SMC_IND_ACCESS_CNTL, AUTO_INCREMENT_IND_0, ~AUTO_INCREMENT_IND_0);
diff --git a/drivers/gpu/drm/radeon/cik.c b/drivers/gpu/drm/radeon/cik.c
index 0b2471107137..fa9565957f9d 100644
--- a/drivers/gpu/drm/radeon/cik.c
+++ b/drivers/gpu/drm/radeon/cik.c
@@ -42,6 +42,16 @@ MODULE_FIRMWARE("radeon/BONAIRE_mc2.bin");
 MODULE_FIRMWARE("radeon/BONAIRE_rlc.bin");
 MODULE_FIRMWARE("radeon/BONAIRE_sdma.bin");
 MODULE_FIRMWARE("radeon/BONAIRE_smc.bin");
+
+MODULE_FIRMWARE("radeon/bonaire_pfp.bin");
+MODULE_FIRMWARE("radeon/bonaire_me.bin");
+MODULE_FIRMWARE("radeon/bonaire_ce.bin");
+MODULE_FIRMWARE("radeon/bonaire_mec.bin");
+MODULE_FIRMWARE("radeon/bonaire_mc.bin");
+MODULE_FIRMWARE("radeon/bonaire_rlc.bin");
+MODULE_FIRMWARE("radeon/bonaire_sdma.bin");
+MODULE_FIRMWARE("radeon/bonaire_smc.bin");
+
 MODULE_FIRMWARE("radeon/HAWAII_pfp.bin");
 MODULE_FIRMWARE("radeon/HAWAII_me.bin");
 MODULE_FIRMWARE("radeon/HAWAII_ce.bin");
@@ -51,18 +61,45 @@ MODULE_FIRMWARE("radeon/HAWAII_mc2.bin");
 MODULE_FIRMWARE("radeon/HAWAII_rlc.bin");
 MODULE_FIRMWARE("radeon/HAWAII_sdma.bin");
 MODULE_FIRMWARE("radeon/HAWAII_smc.bin");
+
+MODULE_FIRMWARE("radeon/hawaii_pfp.bin");
+MODULE_FIRMWARE("radeon/hawaii_me.bin");
+MODULE_FIRMWARE("radeon/hawaii_ce.bin");
+MODULE_FIRMWARE("radeon/hawaii_mec.bin");
+MODULE_FIRMWARE("radeon/hawaii_mc.bin");
+MODULE_FIRMWARE("radeon/hawaii_rlc.bin");
+MODULE_FIRMWARE("radeon/hawaii_sdma.bin");
+MODULE_FIRMWARE("radeon/hawaii_smc.bin");
+
 MODULE_FIRMWARE("radeon/KAVERI_pfp.bin");
 MODULE_FIRMWARE("radeon/KAVERI_me.bin");
 MODULE_FIRMWARE("radeon/KAVERI_ce.bin");
 MODULE_FIRMWARE("radeon/KAVERI_mec.bin");
 MODULE_FIRMWARE("radeon/KAVERI_rlc.bin");
 MODULE_FIRMWARE("radeon/KAVERI_sdma.bin");
+
+MODULE_FIRMWARE("radeon/kaveri_pfp.bin");
+MODULE_FIRMWARE("radeon/kaveri_me.bin");
+MODULE_FIRMWARE("radeon/kaveri_ce.bin");
+MODULE_FIRMWARE("radeon/kaveri_mec.bin");
+MODULE_FIRMWARE("radeon/kaveri_mec2.bin");
+MODULE_FIRMWARE("radeon/kaveri_rlc.bin");
+MODULE_FIRMWARE("radeon/kaveri_sdma.bin");
+
 MODULE_FIRMWARE("radeon/KABINI_pfp.bin");
 MODULE_FIRMWARE("radeon/KABINI_me.bin");
 MODULE_FIRMWARE("radeon/KABINI_ce.bin");
 MODULE_FIRMWARE("radeon/KABINI_mec.bin");
 MODULE_FIRMWARE("radeon/KABINI_rlc.bin");
 MODULE_FIRMWARE("radeon/KABINI_sdma.bin");
+
+MODULE_FIRMWARE("radeon/kabini_pfp.bin");
+MODULE_FIRMWARE("radeon/kabini_me.bin");
+MODULE_FIRMWARE("radeon/kabini_ce.bin");
+MODULE_FIRMWARE("radeon/kabini_mec.bin");
+MODULE_FIRMWARE("radeon/kabini_rlc.bin");
+MODULE_FIRMWARE("radeon/kabini_sdma.bin");
+
 MODULE_FIRMWARE("radeon/MULLINS_pfp.bin");
 MODULE_FIRMWARE("radeon/MULLINS_me.bin");
 MODULE_FIRMWARE("radeon/MULLINS_ce.bin");
@@ -70,6 +107,13 @@ MODULE_FIRMWARE("radeon/MULLINS_mec.bin");
 MODULE_FIRMWARE("radeon/MULLINS_rlc.bin");
 MODULE_FIRMWARE("radeon/MULLINS_sdma.bin");
 
+MODULE_FIRMWARE("radeon/mullins_pfp.bin");
+MODULE_FIRMWARE("radeon/mullins_me.bin");
+MODULE_FIRMWARE("radeon/mullins_ce.bin");
+MODULE_FIRMWARE("radeon/mullins_mec.bin");
+MODULE_FIRMWARE("radeon/mullins_rlc.bin");
+MODULE_FIRMWARE("radeon/mullins_sdma.bin");
+
 extern int r600_ih_ring_alloc(struct radeon_device *rdev);
 extern void r600_ih_ring_fini(struct radeon_device *rdev);
 extern void evergreen_mc_stop(struct radeon_device *rdev, struct evergreen_mc_save *save);
@@ -1760,27 +1804,44 @@ static void cik_srbm_select(struct radeon_device *rdev,
  */
 int ci_mc_load_microcode(struct radeon_device *rdev)
 {
-	const __be32 *fw_data;
+	const __be32 *fw_data = NULL;
+	const __le32 *new_fw_data = NULL;
 	u32 running, blackout = 0;
-	u32 *io_mc_regs;
+	u32 *io_mc_regs = NULL;
+	const __le32 *new_io_mc_regs = NULL;
 	int i, regs_size, ucode_size;
 
 	if (!rdev->mc_fw)
 		return -EINVAL;
 
-	ucode_size = rdev->mc_fw->size / 4;
+	if (rdev->new_fw) {
+		const struct mc_firmware_header_v1_0 *hdr =
+			(const struct mc_firmware_header_v1_0 *)rdev->mc_fw->data;
 
-	switch (rdev->family) {
-	case CHIP_BONAIRE:
-		io_mc_regs = (u32 *)&bonaire_io_mc_regs;
-		regs_size = BONAIRE_IO_MC_REGS_SIZE;
-		break;
-	case CHIP_HAWAII:
-		io_mc_regs = (u32 *)&hawaii_io_mc_regs;
-		regs_size = HAWAII_IO_MC_REGS_SIZE;
-		break;
-	default:
-		return -EINVAL;
+		radeon_ucode_print_mc_hdr(&hdr->header);
+
+		regs_size = le32_to_cpu(hdr->io_debug_size_bytes) / (4 * 2);
+		new_io_mc_regs = (const __le32 *)
+			(rdev->mc_fw->data + le32_to_cpu(hdr->io_debug_array_offset_bytes));
+		ucode_size = le32_to_cpu(hdr->header.ucode_size_bytes) / 4;
+		new_fw_data = (const __le32 *)
+			(rdev->mc_fw->data + le32_to_cpu(hdr->header.ucode_array_offset_bytes));
+	} else {
+		ucode_size = rdev->mc_fw->size / 4;
+
+		switch (rdev->family) {
+		case CHIP_BONAIRE:
+			io_mc_regs = (u32 *)&bonaire_io_mc_regs;
+			regs_size = BONAIRE_IO_MC_REGS_SIZE;
+			break;
+		case CHIP_HAWAII:
+			io_mc_regs = (u32 *)&hawaii_io_mc_regs;
+			regs_size = HAWAII_IO_MC_REGS_SIZE;
+			break;
+		default:
+			return -EINVAL;
+		}
+		fw_data = (const __be32 *)rdev->mc_fw->data;
 	}
 
 	running = RREG32(MC_SEQ_SUP_CNTL) & RUN_MASK;
@@ -1797,13 +1858,21 @@ int ci_mc_load_microcode(struct radeon_device *rdev)
 
 		/* load mc io regs */
 		for (i = 0; i < regs_size; i++) {
-			WREG32(MC_SEQ_IO_DEBUG_INDEX, io_mc_regs[(i << 1)]);
-			WREG32(MC_SEQ_IO_DEBUG_DATA, io_mc_regs[(i << 1) + 1]);
+			if (rdev->new_fw) {
+				WREG32(MC_SEQ_IO_DEBUG_INDEX, le32_to_cpup(new_io_mc_regs++));
+				WREG32(MC_SEQ_IO_DEBUG_DATA, le32_to_cpup(new_io_mc_regs++));
+			} else {
+				WREG32(MC_SEQ_IO_DEBUG_INDEX, io_mc_regs[(i << 1)]);
+				WREG32(MC_SEQ_IO_DEBUG_DATA, io_mc_regs[(i << 1) + 1]);
+			}
 		}
 		/* load the MC ucode */
-		fw_data = (const __be32 *)rdev->mc_fw->data;
-		for (i = 0; i < ucode_size; i++)
-			WREG32(MC_SEQ_SUP_PGM, be32_to_cpup(fw_data++));
+		for (i = 0; i < ucode_size; i++) {
+			if (rdev->new_fw)
+				WREG32(MC_SEQ_SUP_PGM, le32_to_cpup(new_fw_data++));
+			else
+				WREG32(MC_SEQ_SUP_PGM, be32_to_cpup(fw_data++));
+		}
 
 		/* put the engine back into the active state */
 		WREG32(MC_SEQ_SUP_CNTL, 0x00000008);
@@ -1841,17 +1910,21 @@ int ci_mc_load_microcode(struct radeon_device *rdev)
 static int cik_init_microcode(struct radeon_device *rdev)
 {
 	const char *chip_name;
+	const char *new_chip_name;
 	size_t pfp_req_size, me_req_size, ce_req_size,
 		mec_req_size, rlc_req_size, mc_req_size = 0,
 		sdma_req_size, smc_req_size = 0, mc2_req_size = 0;
 	char fw_name[30];
+	int new_fw = 0;
 	int err;
+	int num_fw;
 
 	DRM_DEBUG("\n");
 
 	switch (rdev->family) {
 	case CHIP_BONAIRE:
 		chip_name = "BONAIRE";
+		new_chip_name = "bonaire";
 		pfp_req_size = CIK_PFP_UCODE_SIZE * 4;
 		me_req_size = CIK_ME_UCODE_SIZE * 4;
 		ce_req_size = CIK_CE_UCODE_SIZE * 4;
@@ -1861,9 +1934,11 @@ static int cik_init_microcode(struct radeon_device *rdev)
 		mc2_req_size = BONAIRE_MC2_UCODE_SIZE * 4;
 		sdma_req_size = CIK_SDMA_UCODE_SIZE * 4;
 		smc_req_size = ALIGN(BONAIRE_SMC_UCODE_SIZE, 4);
+		num_fw = 8;
 		break;
 	case CHIP_HAWAII:
 		chip_name = "HAWAII";
+		new_chip_name = "hawaii";
 		pfp_req_size = CIK_PFP_UCODE_SIZE * 4;
 		me_req_size = CIK_ME_UCODE_SIZE * 4;
 		ce_req_size = CIK_CE_UCODE_SIZE * 4;
@@ -1873,142 +1948,285 @@ static int cik_init_microcode(struct radeon_device *rdev)
 		mc2_req_size = HAWAII_MC2_UCODE_SIZE * 4;
 		sdma_req_size = CIK_SDMA_UCODE_SIZE * 4;
 		smc_req_size = ALIGN(HAWAII_SMC_UCODE_SIZE, 4);
+		num_fw = 8;
 		break;
 	case CHIP_KAVERI:
 		chip_name = "KAVERI";
+		new_chip_name = "kaveri";
 		pfp_req_size = CIK_PFP_UCODE_SIZE * 4;
 		me_req_size = CIK_ME_UCODE_SIZE * 4;
 		ce_req_size = CIK_CE_UCODE_SIZE * 4;
 		mec_req_size = CIK_MEC_UCODE_SIZE * 4;
 		rlc_req_size = KV_RLC_UCODE_SIZE * 4;
 		sdma_req_size = CIK_SDMA_UCODE_SIZE * 4;
+		num_fw = 7;
 		break;
 	case CHIP_KABINI:
 		chip_name = "KABINI";
+		new_chip_name = "kabini";
 		pfp_req_size = CIK_PFP_UCODE_SIZE * 4;
 		me_req_size = CIK_ME_UCODE_SIZE * 4;
 		ce_req_size = CIK_CE_UCODE_SIZE * 4;
 		mec_req_size = CIK_MEC_UCODE_SIZE * 4;
 		rlc_req_size = KB_RLC_UCODE_SIZE * 4;
 		sdma_req_size = CIK_SDMA_UCODE_SIZE * 4;
+		num_fw = 6;
 		break;
 	case CHIP_MULLINS:
 		chip_name = "MULLINS";
+		new_chip_name = "mullins";
 		pfp_req_size = CIK_PFP_UCODE_SIZE * 4;
 		me_req_size = CIK_ME_UCODE_SIZE * 4;
 		ce_req_size = CIK_CE_UCODE_SIZE * 4;
 		mec_req_size = CIK_MEC_UCODE_SIZE * 4;
 		rlc_req_size = ML_RLC_UCODE_SIZE * 4;
 		sdma_req_size = CIK_SDMA_UCODE_SIZE * 4;
+		num_fw = 6;
 		break;
 	default: BUG();
 	}
 
-	DRM_INFO("Loading %s Microcode\n", chip_name);
+	DRM_INFO("Loading %s Microcode\n", new_chip_name);
 
-	snprintf(fw_name, sizeof(fw_name), "radeon/%s_pfp.bin", chip_name);
+	snprintf(fw_name, sizeof(fw_name), "radeon/%s_pfp.bin", new_chip_name);
 	err = request_firmware(&rdev->pfp_fw, fw_name, rdev->dev);
-	if (err)
-		goto out;
-	if (rdev->pfp_fw->size != pfp_req_size) {
-		printk(KERN_ERR
-		       "cik_cp: Bogus length %zu in firmware \"%s\"\n",
-		       rdev->pfp_fw->size, fw_name);
-		err = -EINVAL;
-		goto out;
+	if (err) {
+		snprintf(fw_name, sizeof(fw_name), "radeon/%s_pfp.bin", chip_name);
+		err = request_firmware(&rdev->pfp_fw, fw_name, rdev->dev);
+		if (err)
+			goto out;
+		if (rdev->pfp_fw->size != pfp_req_size) {
+			printk(KERN_ERR
+			       "cik_cp: Bogus length %zu in firmware \"%s\"\n",
+			       rdev->pfp_fw->size, fw_name);
+			err = -EINVAL;
+			goto out;
+		}
+	} else {
+		err = radeon_ucode_validate(rdev->pfp_fw);
+		if (err) {
+			printk(KERN_ERR
+			       "cik_fw: validation failed for firmware \"%s\"\n",
+			       fw_name);
+			goto out;
+		} else {
+			new_fw++;
+		}
 	}
 
-	snprintf(fw_name, sizeof(fw_name), "radeon/%s_me.bin", chip_name);
+	snprintf(fw_name, sizeof(fw_name), "radeon/%s_me.bin", new_chip_name);
 	err = request_firmware(&rdev->me_fw, fw_name, rdev->dev);
-	if (err)
-		goto out;
-	if (rdev->me_fw->size != me_req_size) {
-		printk(KERN_ERR
-		       "cik_cp: Bogus length %zu in firmware \"%s\"\n",
-		       rdev->me_fw->size, fw_name);
-		err = -EINVAL;
+	if (err) {
+		snprintf(fw_name, sizeof(fw_name), "radeon/%s_me.bin", chip_name);
+		err = request_firmware(&rdev->me_fw, fw_name, rdev->dev);
+		if (err)
+			goto out;
+		if (rdev->me_fw->size != me_req_size) {
+			printk(KERN_ERR
+			       "cik_cp: Bogus length %zu in firmware \"%s\"\n",
+			       rdev->me_fw->size, fw_name);
+			err = -EINVAL;
+		}
+	} else {
+		err = radeon_ucode_validate(rdev->me_fw);
+		if (err) {
+			printk(KERN_ERR
+			       "cik_fw: validation failed for firmware \"%s\"\n",
+			       fw_name);
+			goto out;
+		} else {
+			new_fw++;
+		}
 	}
 
-	snprintf(fw_name, sizeof(fw_name), "radeon/%s_ce.bin", chip_name);
+	snprintf(fw_name, sizeof(fw_name), "radeon/%s_ce.bin", new_chip_name);
 	err = request_firmware(&rdev->ce_fw, fw_name, rdev->dev);
-	if (err)
-		goto out;
-	if (rdev->ce_fw->size != ce_req_size) {
-		printk(KERN_ERR
-		       "cik_cp: Bogus length %zu in firmware \"%s\"\n",
-		       rdev->ce_fw->size, fw_name);
-		err = -EINVAL;
+	if (err) {
+		snprintf(fw_name, sizeof(fw_name), "radeon/%s_ce.bin", chip_name);
+		err = request_firmware(&rdev->ce_fw, fw_name, rdev->dev);
+		if (err)
+			goto out;
+		if (rdev->ce_fw->size != ce_req_size) {
+			printk(KERN_ERR
+			       "cik_cp: Bogus length %zu in firmware \"%s\"\n",
+			       rdev->ce_fw->size, fw_name);
+			err = -EINVAL;
+		}
+	} else {
+		err = radeon_ucode_validate(rdev->ce_fw);
+		if (err) {
+			printk(KERN_ERR
+			       "cik_fw: validation failed for firmware \"%s\"\n",
+			       fw_name);
+			goto out;
+		} else {
+			new_fw++;
+		}
 	}
 
-	snprintf(fw_name, sizeof(fw_name), "radeon/%s_mec.bin", chip_name);
+	snprintf(fw_name, sizeof(fw_name), "radeon/%s_mec.bin", new_chip_name);
 	err = request_firmware(&rdev->mec_fw, fw_name, rdev->dev);
-	if (err)
-		goto out;
-	if (rdev->mec_fw->size != mec_req_size) {
-		printk(KERN_ERR
-		       "cik_cp: Bogus length %zu in firmware \"%s\"\n",
-		       rdev->mec_fw->size, fw_name);
-		err = -EINVAL;
+	if (err) {
+		snprintf(fw_name, sizeof(fw_name), "radeon/%s_mec.bin", chip_name);
+		err = request_firmware(&rdev->mec_fw, fw_name, rdev->dev);
+		if (err)
+			goto out;
+		if (rdev->mec_fw->size != mec_req_size) {
+			printk(KERN_ERR
+			       "cik_cp: Bogus length %zu in firmware \"%s\"\n",
+			       rdev->mec_fw->size, fw_name);
+			err = -EINVAL;
+		}
+	} else {
+		err = radeon_ucode_validate(rdev->mec_fw);
+		if (err) {
+			printk(KERN_ERR
+			       "cik_fw: validation failed for firmware \"%s\"\n",
+			       fw_name);
+			goto out;
+		} else {
+			new_fw++;
+		}
 	}
 
-	snprintf(fw_name, sizeof(fw_name), "radeon/%s_rlc.bin", chip_name);
+	if (rdev->family == CHIP_KAVERI) {
+		snprintf(fw_name, sizeof(fw_name), "radeon/%s_mec2.bin", new_chip_name);
+		err = request_firmware(&rdev->mec2_fw, fw_name, rdev->dev);
+		if (err) {
+			goto out;
+		} else {
+			err = radeon_ucode_validate(rdev->mec2_fw);
+			if (err) {
+				goto out;
+			} else {
+				new_fw++;
+			}
+		}
+	}
+
+	snprintf(fw_name, sizeof(fw_name), "radeon/%s_rlc.bin", new_chip_name);
 	err = request_firmware(&rdev->rlc_fw, fw_name, rdev->dev);
-	if (err)
-		goto out;
-	if (rdev->rlc_fw->size != rlc_req_size) {
-		printk(KERN_ERR
-		       "cik_rlc: Bogus length %zu in firmware \"%s\"\n",
-		       rdev->rlc_fw->size, fw_name);
-		err = -EINVAL;
+	if (err) {
+		snprintf(fw_name, sizeof(fw_name), "radeon/%s_rlc.bin", chip_name);
+		err = request_firmware(&rdev->rlc_fw, fw_name, rdev->dev);
+		if (err)
+			goto out;
+		if (rdev->rlc_fw->size != rlc_req_size) {
+			printk(KERN_ERR
+			       "cik_rlc: Bogus length %zu in firmware \"%s\"\n",
+			       rdev->rlc_fw->size, fw_name);
+			err = -EINVAL;
+		}
+	} else {
+		err = radeon_ucode_validate(rdev->rlc_fw);
+		if (err) {
+			printk(KERN_ERR
+			       "cik_fw: validation failed for firmware \"%s\"\n",
+			       fw_name);
+			goto out;
+		} else {
+			new_fw++;
+		}
 	}
 
-	snprintf(fw_name, sizeof(fw_name), "radeon/%s_sdma.bin", chip_name);
+	snprintf(fw_name, sizeof(fw_name), "radeon/%s_sdma.bin", new_chip_name);
 	err = request_firmware(&rdev->sdma_fw, fw_name, rdev->dev);
-	if (err)
-		goto out;
-	if (rdev->sdma_fw->size != sdma_req_size) {
-		printk(KERN_ERR
-		       "cik_sdma: Bogus length %zu in firmware \"%s\"\n",
-		       rdev->sdma_fw->size, fw_name);
-		err = -EINVAL;
+	if (err) {
+		snprintf(fw_name, sizeof(fw_name), "radeon/%s_sdma.bin", chip_name);
+		err = request_firmware(&rdev->sdma_fw, fw_name, rdev->dev);
+		if (err)
+			goto out;
+		if (rdev->sdma_fw->size != sdma_req_size) {
+			printk(KERN_ERR
+			       "cik_sdma: Bogus length %zu in firmware \"%s\"\n",
+			       rdev->sdma_fw->size, fw_name);
+			err = -EINVAL;
+		}
+	} else {
+		err = radeon_ucode_validate(rdev->sdma_fw);
+		if (err) {
+			printk(KERN_ERR
+			       "cik_fw: validation failed for firmware \"%s\"\n",
+			       fw_name);
+			goto out;
+		} else {
+			new_fw++;
+		}
 	}
 
 	/* No SMC, MC ucode on APUs */
 	if (!(rdev->flags & RADEON_IS_IGP)) {
-		snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc2.bin", chip_name);
+		snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc.bin", new_chip_name);
 		err = request_firmware(&rdev->mc_fw, fw_name, rdev->dev);
 		if (err) {
-			snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc.bin", chip_name);
+			snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc2.bin", chip_name);
 			err = request_firmware(&rdev->mc_fw, fw_name, rdev->dev);
-			if (err)
+			if (err) {
+				snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc.bin", chip_name);
+				err = request_firmware(&rdev->mc_fw, fw_name, rdev->dev);
+				if (err)
+					goto out;
+			}
+			if ((rdev->mc_fw->size != mc_req_size) &&
+			    (rdev->mc_fw->size != mc2_req_size)){
+				printk(KERN_ERR
+				       "cik_mc: Bogus length %zu in firmware \"%s\"\n",
+				       rdev->mc_fw->size, fw_name);
+				err = -EINVAL;
+			}
+			DRM_INFO("%s: %zu bytes\n", fw_name, rdev->mc_fw->size);
+		} else {
+			err = radeon_ucode_validate(rdev->mc_fw);
+			if (err) {
+				printk(KERN_ERR
+				       "cik_fw: validation failed for firmware \"%s\"\n",
+				       fw_name);
 				goto out;
+			} else {
+				new_fw++;
+			}
 		}
-		if ((rdev->mc_fw->size != mc_req_size) &&
-		    (rdev->mc_fw->size != mc2_req_size)){
-			printk(KERN_ERR
-			       "cik_mc: Bogus length %zu in firmware \"%s\"\n",
-			       rdev->mc_fw->size, fw_name);
-			err = -EINVAL;
-		}
-		DRM_INFO("%s: %zu bytes\n", fw_name, rdev->mc_fw->size);
 
-		snprintf(fw_name, sizeof(fw_name), "radeon/%s_smc.bin", chip_name);
+		snprintf(fw_name, sizeof(fw_name), "radeon/%s_smc.bin", new_chip_name);
 		err = request_firmware(&rdev->smc_fw, fw_name, rdev->dev);
 		if (err) {
-			printk(KERN_ERR
-			       "smc: error loading firmware \"%s\"\n",
-			       fw_name);
-			release_firmware(rdev->smc_fw);
-			rdev->smc_fw = NULL;
-			err = 0;
-		} else if (rdev->smc_fw->size != smc_req_size) {
-			printk(KERN_ERR
-			       "cik_smc: Bogus length %zu in firmware \"%s\"\n",
-			       rdev->smc_fw->size, fw_name);
-			err = -EINVAL;
+			snprintf(fw_name, sizeof(fw_name), "radeon/%s_smc.bin", chip_name);
+			err = request_firmware(&rdev->smc_fw, fw_name, rdev->dev);
+			if (err) {
+				printk(KERN_ERR
+				       "smc: error loading firmware \"%s\"\n",
+				       fw_name);
+				release_firmware(rdev->smc_fw);
+				rdev->smc_fw = NULL;
+				err = 0;
+			} else if (rdev->smc_fw->size != smc_req_size) {
+				printk(KERN_ERR
+				       "cik_smc: Bogus length %zu in firmware \"%s\"\n",
+				       rdev->smc_fw->size, fw_name);
+				err = -EINVAL;
+			}
+		} else {
+			err = radeon_ucode_validate(rdev->smc_fw);
+			if (err) {
+				printk(KERN_ERR
+				       "cik_fw: validation failed for firmware \"%s\"\n",
+				       fw_name);
+				goto out;
+			} else {
+				new_fw++;
+			}
 		}
 	}
 
+	if (new_fw == 0) {
+		rdev->new_fw = false;
+	} else if (new_fw < num_fw) {
+		printk(KERN_ERR "ci_fw: mixing new and old firmware!\n");
+		err = -EINVAL;
+	} else {
+		rdev->new_fw = true;
+	}
+
 out:
 	if (err) {
 		if (err != -EINVAL)
@@ -2021,8 +2239,14 @@ out:
 		rdev->me_fw = NULL;
 		release_firmware(rdev->ce_fw);
 		rdev->ce_fw = NULL;
+		release_firmware(rdev->mec_fw);
+		rdev->mec_fw = NULL;
+		release_firmware(rdev->mec2_fw);
+		rdev->mec2_fw = NULL;
 		release_firmware(rdev->rlc_fw);
 		rdev->rlc_fw = NULL;
+		release_firmware(rdev->sdma_fw);
+		rdev->sdma_fw = NULL;
 		release_firmware(rdev->mc_fw);
 		rdev->mc_fw = NULL;
 		release_firmware(rdev->smc_fw);
@@ -2291,6 +2515,7 @@ static void cik_tiling_mode_table_init(struct radeon_device *rdev)
 				gb_tile_moden = 0;
 				break;
 			}
+			rdev->config.cik.macrotile_mode_array[reg_offset] = gb_tile_moden;
 			WREG32(GB_MACROTILE_MODE0 + (reg_offset * 4), gb_tile_moden);
 		}
 	} else if (num_pipe_configs == 8) {
@@ -3258,7 +3483,7 @@ static void cik_gpu_init(struct radeon_device *rdev)
 	u32 mc_shared_chmap, mc_arb_ramcfg;
 	u32 hdp_host_path_cntl;
 	u32 tmp;
-	int i, j, k;
+	int i, j;
 
 	switch (rdev->family) {
 	case CHIP_BONAIRE:
@@ -3319,6 +3544,7 @@ static void cik_gpu_init(struct radeon_device *rdev)
 			   (rdev->pdev->device == 0x130B) ||
 			   (rdev->pdev->device == 0x130E) ||
 			   (rdev->pdev->device == 0x1315) ||
+			   (rdev->pdev->device == 0x1318) ||
 			   (rdev->pdev->device == 0x131B)) {
 			rdev->config.cik.max_cu_per_sh = 4;
 			rdev->config.cik.max_backends_per_se = 1;
@@ -3447,12 +3673,11 @@ static void cik_gpu_init(struct radeon_device *rdev)
 		     rdev->config.cik.max_sh_per_se,
 		     rdev->config.cik.max_backends_per_se);
 
+	rdev->config.cik.active_cus = 0;
 	for (i = 0; i < rdev->config.cik.max_shader_engines; i++) {
 		for (j = 0; j < rdev->config.cik.max_sh_per_se; j++) {
-			for (k = 0; k < rdev->config.cik.max_cu_per_sh; k++) {
-				rdev->config.cik.active_cus +=
-					hweight32(cik_get_cu_active_bitmap(rdev, i, j));
-			}
+			rdev->config.cik.active_cus +=
+				hweight32(cik_get_cu_active_bitmap(rdev, i, j));
 		}
 	}
 
@@ -3576,7 +3801,7 @@ int cik_ring_test(struct radeon_device *rdev, struct radeon_ring *ring)
 	radeon_ring_write(ring, PACKET3(PACKET3_SET_UCONFIG_REG, 1));
 	radeon_ring_write(ring, ((scratch - PACKET3_SET_UCONFIG_REG_START) >> 2));
 	radeon_ring_write(ring, 0xDEADBEEF);
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev, ring, false);
 
 	for (i = 0; i < rdev->usec_timeout; i++) {
 		tmp = RREG32(scratch);
@@ -3665,8 +3890,6 @@ void cik_fence_gfx_ring_emit(struct radeon_device *rdev,
 	radeon_ring_write(ring, (upper_32_bits(addr) & 0xffff) | DATA_SEL(1) | INT_SEL(2));
 	radeon_ring_write(ring, fence->seq);
 	radeon_ring_write(ring, 0);
-	/* HDP flush */
-	cik_hdp_flush_cp_ring_emit(rdev, fence->ring);
 }
 
 /**
@@ -3695,10 +3918,19 @@ void cik_fence_compute_ring_emit(struct radeon_device *rdev,
 	radeon_ring_write(ring, upper_32_bits(addr));
 	radeon_ring_write(ring, fence->seq);
 	radeon_ring_write(ring, 0);
-	/* HDP flush */
-	cik_hdp_flush_cp_ring_emit(rdev, fence->ring);
 }
 
+/**
+ * cik_semaphore_ring_emit - emit a semaphore on the CP ring
+ *
+ * @rdev: radeon_device pointer
+ * @ring: radeon ring buffer object
+ * @semaphore: radeon semaphore object
+ * @emit_wait: Is this a sempahore wait?
+ *
+ * Emits a semaphore signal/wait packet to the CP ring and prevents the PFP
+ * from running ahead of semaphore waits.
+ */
 bool cik_semaphore_ring_emit(struct radeon_device *rdev,
 			     struct radeon_ring *ring,
 			     struct radeon_semaphore *semaphore,
@@ -3711,6 +3943,12 @@ bool cik_semaphore_ring_emit(struct radeon_device *rdev,
 	radeon_ring_write(ring, lower_32_bits(addr));
 	radeon_ring_write(ring, (upper_32_bits(addr) & 0xffff) | sel);
 
+	if (emit_wait && ring->idx == RADEON_RING_TYPE_GFX_INDEX) {
+		/* Prevent the PFP from running ahead of the semaphore wait */
+		radeon_ring_write(ring, PACKET3(PACKET3_PFP_SYNC_ME, 0));
+		radeon_ring_write(ring, 0x0);
+	}
+
 	return true;
 }
 
@@ -3783,7 +4021,7 @@ int cik_copy_cpdma(struct radeon_device *rdev,
 		return r;
 	}
 
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev, ring, false);
 	radeon_semaphore_free(rdev, &sem, *fence);
 
 	return r;
@@ -3882,7 +4120,7 @@ int cik_ib_test(struct radeon_device *rdev, struct radeon_ring *ring)
 	ib.ptr[1] = ((scratch - PACKET3_SET_UCONFIG_REG_START) >> 2);
 	ib.ptr[2] = 0xDEADBEEF;
 	ib.length_dw = 3;
-	r = radeon_ib_schedule(rdev, &ib, NULL);
+	r = radeon_ib_schedule(rdev, &ib, NULL, false);
 	if (r) {
 		radeon_scratch_free(rdev, scratch);
 		radeon_ib_free(rdev, &ib);
@@ -3968,7 +4206,6 @@ static void cik_cp_gfx_enable(struct radeon_device *rdev, bool enable)
  */
 static int cik_cp_gfx_load_microcode(struct radeon_device *rdev)
 {
-	const __be32 *fw_data;
 	int i;
 
 	if (!rdev->me_fw || !rdev->pfp_fw || !rdev->ce_fw)
@@ -3976,26 +4213,70 @@ static int cik_cp_gfx_load_microcode(struct radeon_device *rdev)
 
 	cik_cp_gfx_enable(rdev, false);
 
-	/* PFP */
-	fw_data = (const __be32 *)rdev->pfp_fw->data;
-	WREG32(CP_PFP_UCODE_ADDR, 0);
-	for (i = 0; i < CIK_PFP_UCODE_SIZE; i++)
-		WREG32(CP_PFP_UCODE_DATA, be32_to_cpup(fw_data++));
-	WREG32(CP_PFP_UCODE_ADDR, 0);
-
-	/* CE */
-	fw_data = (const __be32 *)rdev->ce_fw->data;
-	WREG32(CP_CE_UCODE_ADDR, 0);
-	for (i = 0; i < CIK_CE_UCODE_SIZE; i++)
-		WREG32(CP_CE_UCODE_DATA, be32_to_cpup(fw_data++));
-	WREG32(CP_CE_UCODE_ADDR, 0);
-
-	/* ME */
-	fw_data = (const __be32 *)rdev->me_fw->data;
-	WREG32(CP_ME_RAM_WADDR, 0);
-	for (i = 0; i < CIK_ME_UCODE_SIZE; i++)
-		WREG32(CP_ME_RAM_DATA, be32_to_cpup(fw_data++));
-	WREG32(CP_ME_RAM_WADDR, 0);
+	if (rdev->new_fw) {
+		const struct gfx_firmware_header_v1_0 *pfp_hdr =
+			(const struct gfx_firmware_header_v1_0 *)rdev->pfp_fw->data;
+		const struct gfx_firmware_header_v1_0 *ce_hdr =
+			(const struct gfx_firmware_header_v1_0 *)rdev->ce_fw->data;
+		const struct gfx_firmware_header_v1_0 *me_hdr =
+			(const struct gfx_firmware_header_v1_0 *)rdev->me_fw->data;
+		const __le32 *fw_data;
+		u32 fw_size;
+
+		radeon_ucode_print_gfx_hdr(&pfp_hdr->header);
+		radeon_ucode_print_gfx_hdr(&ce_hdr->header);
+		radeon_ucode_print_gfx_hdr(&me_hdr->header);
+
+		/* PFP */
+		fw_data = (const __le32 *)
+			(rdev->pfp_fw->data + le32_to_cpu(pfp_hdr->header.ucode_array_offset_bytes));
+		fw_size = le32_to_cpu(pfp_hdr->header.ucode_size_bytes) / 4;
+		WREG32(CP_PFP_UCODE_ADDR, 0);
+		for (i = 0; i < fw_size; i++)
+			WREG32(CP_PFP_UCODE_DATA, le32_to_cpup(fw_data++));
+		WREG32(CP_PFP_UCODE_ADDR, 0);
+
+		/* CE */
+		fw_data = (const __le32 *)
+			(rdev->ce_fw->data + le32_to_cpu(ce_hdr->header.ucode_array_offset_bytes));
+		fw_size = le32_to_cpu(ce_hdr->header.ucode_size_bytes) / 4;
+		WREG32(CP_CE_UCODE_ADDR, 0);
+		for (i = 0; i < fw_size; i++)
+			WREG32(CP_CE_UCODE_DATA, le32_to_cpup(fw_data++));
+		WREG32(CP_CE_UCODE_ADDR, 0);
+
+		/* ME */
+		fw_data = (const __be32 *)
+			(rdev->me_fw->data + le32_to_cpu(me_hdr->header.ucode_array_offset_bytes));
+		fw_size = le32_to_cpu(me_hdr->header.ucode_size_bytes) / 4;
+		WREG32(CP_ME_RAM_WADDR, 0);
+		for (i = 0; i < fw_size; i++)
+			WREG32(CP_ME_RAM_DATA, le32_to_cpup(fw_data++));
+		WREG32(CP_ME_RAM_WADDR, 0);
+	} else {
+		const __be32 *fw_data;
+
+		/* PFP */
+		fw_data = (const __be32 *)rdev->pfp_fw->data;
+		WREG32(CP_PFP_UCODE_ADDR, 0);
+		for (i = 0; i < CIK_PFP_UCODE_SIZE; i++)
+			WREG32(CP_PFP_UCODE_DATA, be32_to_cpup(fw_data++));
+		WREG32(CP_PFP_UCODE_ADDR, 0);
+
+		/* CE */
+		fw_data = (const __be32 *)rdev->ce_fw->data;
+		WREG32(CP_CE_UCODE_ADDR, 0);
+		for (i = 0; i < CIK_CE_UCODE_SIZE; i++)
+			WREG32(CP_CE_UCODE_DATA, be32_to_cpup(fw_data++));
+		WREG32(CP_CE_UCODE_ADDR, 0);
+
+		/* ME */
+		fw_data = (const __be32 *)rdev->me_fw->data;
+		WREG32(CP_ME_RAM_WADDR, 0);
+		for (i = 0; i < CIK_ME_UCODE_SIZE; i++)
+			WREG32(CP_ME_RAM_DATA, be32_to_cpup(fw_data++));
+		WREG32(CP_ME_RAM_WADDR, 0);
+	}
 
 	WREG32(CP_PFP_UCODE_ADDR, 0);
 	WREG32(CP_CE_UCODE_ADDR, 0);
@@ -4060,7 +4341,7 @@ static int cik_cp_gfx_start(struct radeon_device *rdev)
 	radeon_ring_write(ring, 0x0000000e); /* VGT_VERTEX_REUSE_BLOCK_CNTL */
 	radeon_ring_write(ring, 0x00000010); /* VGT_OUT_DEALLOC_CNTL */
 
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev, ring, false);
 
 	return 0;
 }
@@ -4260,7 +4541,6 @@ static void cik_cp_compute_enable(struct radeon_device *rdev, bool enable)
  */
 static int cik_cp_compute_load_microcode(struct radeon_device *rdev)
 {
-	const __be32 *fw_data;
 	int i;
 
 	if (!rdev->mec_fw)
@@ -4268,20 +4548,55 @@ static int cik_cp_compute_load_microcode(struct radeon_device *rdev)
 
 	cik_cp_compute_enable(rdev, false);
 
-	/* MEC1 */
-	fw_data = (const __be32 *)rdev->mec_fw->data;
-	WREG32(CP_MEC_ME1_UCODE_ADDR, 0);
-	for (i = 0; i < CIK_MEC_UCODE_SIZE; i++)
-		WREG32(CP_MEC_ME1_UCODE_DATA, be32_to_cpup(fw_data++));
-	WREG32(CP_MEC_ME1_UCODE_ADDR, 0);
+	if (rdev->new_fw) {
+		const struct gfx_firmware_header_v1_0 *mec_hdr =
+			(const struct gfx_firmware_header_v1_0 *)rdev->mec_fw->data;
+		const __le32 *fw_data;
+		u32 fw_size;
+
+		radeon_ucode_print_gfx_hdr(&mec_hdr->header);
+
+		/* MEC1 */
+		fw_data = (const __le32 *)
+			(rdev->mec_fw->data + le32_to_cpu(mec_hdr->header.ucode_array_offset_bytes));
+		fw_size = le32_to_cpu(mec_hdr->header.ucode_size_bytes) / 4;
+		WREG32(CP_MEC_ME1_UCODE_ADDR, 0);
+		for (i = 0; i < fw_size; i++)
+			WREG32(CP_MEC_ME1_UCODE_DATA, le32_to_cpup(fw_data++));
+		WREG32(CP_MEC_ME1_UCODE_ADDR, 0);
 
-	if (rdev->family == CHIP_KAVERI) {
 		/* MEC2 */
+		if (rdev->family == CHIP_KAVERI) {
+			const struct gfx_firmware_header_v1_0 *mec2_hdr =
+				(const struct gfx_firmware_header_v1_0 *)rdev->mec2_fw->data;
+
+			fw_data = (const __le32 *)
+				(rdev->mec2_fw->data +
+				 le32_to_cpu(mec2_hdr->header.ucode_array_offset_bytes));
+			fw_size = le32_to_cpu(mec2_hdr->header.ucode_size_bytes) / 4;
+			WREG32(CP_MEC_ME2_UCODE_ADDR, 0);
+			for (i = 0; i < fw_size; i++)
+				WREG32(CP_MEC_ME2_UCODE_DATA, le32_to_cpup(fw_data++));
+			WREG32(CP_MEC_ME2_UCODE_ADDR, 0);
+		}
+	} else {
+		const __be32 *fw_data;
+
+		/* MEC1 */
 		fw_data = (const __be32 *)rdev->mec_fw->data;
-		WREG32(CP_MEC_ME2_UCODE_ADDR, 0);
+		WREG32(CP_MEC_ME1_UCODE_ADDR, 0);
 		for (i = 0; i < CIK_MEC_UCODE_SIZE; i++)
-			WREG32(CP_MEC_ME2_UCODE_DATA, be32_to_cpup(fw_data++));
-		WREG32(CP_MEC_ME2_UCODE_ADDR, 0);
+			WREG32(CP_MEC_ME1_UCODE_DATA, be32_to_cpup(fw_data++));
+		WREG32(CP_MEC_ME1_UCODE_ADDR, 0);
+
+		if (rdev->family == CHIP_KAVERI) {
+			/* MEC2 */
+			fw_data = (const __be32 *)rdev->mec_fw->data;
+			WREG32(CP_MEC_ME2_UCODE_ADDR, 0);
+			for (i = 0; i < CIK_MEC_UCODE_SIZE; i++)
+				WREG32(CP_MEC_ME2_UCODE_DATA, be32_to_cpup(fw_data++));
+			WREG32(CP_MEC_ME2_UCODE_ADDR, 0);
+		}
 	}
 
 	return 0;
@@ -4374,7 +4689,7 @@ static int cik_mec_init(struct radeon_device *rdev)
 		r = radeon_bo_create(rdev,
 				     rdev->mec.num_mec *rdev->mec.num_pipe * MEC_HPD_SIZE * 2,
 				     PAGE_SIZE, true,
-				     RADEON_GEM_DOMAIN_GTT, NULL,
+				     RADEON_GEM_DOMAIN_GTT, 0, NULL,
 				     &rdev->mec.hpd_eop_obj);
 		if (r) {
 			dev_warn(rdev->dev, "(%d) create HDP EOP bo failed\n", r);
@@ -4544,7 +4859,7 @@ static int cik_cp_compute_resume(struct radeon_device *rdev)
 			r = radeon_bo_create(rdev,
 					     sizeof(struct bonaire_mqd),
 					     PAGE_SIZE, true,
-					     RADEON_GEM_DOMAIN_GTT, NULL,
+					     RADEON_GEM_DOMAIN_GTT, 0, NULL,
 					     &rdev->ring[idx].mqd_obj);
 			if (r) {
 				dev_warn(rdev->dev, "(%d) create MQD bo failed\n", r);
@@ -5401,7 +5716,6 @@ static int cik_pcie_gart_enable(struct radeon_device *rdev)
 	r = radeon_gart_table_vram_pin(rdev);
 	if (r)
 		return r;
-	radeon_gart_restore(rdev);
 	/* Setup TLB control */
 	WREG32(MC_VM_MX_L1_TLB_CNTL,
 	       (0xA << 7) |
@@ -5435,20 +5749,17 @@ static int cik_pcie_gart_enable(struct radeon_device *rdev)
 	WREG32(0x15D8, 0);
 	WREG32(0x15DC, 0);
 
-	/* empty context1-15 */
-	/* FIXME start with 4G, once using 2 level pt switch to full
-	 * vm size space
-	 */
+	/* restore context1-15 */
 	/* set vm size, must be a multiple of 4 */
 	WREG32(VM_CONTEXT1_PAGE_TABLE_START_ADDR, 0);
 	WREG32(VM_CONTEXT1_PAGE_TABLE_END_ADDR, rdev->vm_manager.max_pfn);
 	for (i = 1; i < 16; i++) {
 		if (i < 8)
 			WREG32(VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (i << 2),
-			       rdev->gart.table_addr >> 12);
+			       rdev->vm_manager.saved_table_addr[i]);
 		else
 			WREG32(VM_CONTEXT8_PAGE_TABLE_BASE_ADDR + ((i - 8) << 2),
-			       rdev->gart.table_addr >> 12);
+			       rdev->vm_manager.saved_table_addr[i]);
 	}
 
 	/* enable context1-15 */
@@ -5513,6 +5824,17 @@ static int cik_pcie_gart_enable(struct radeon_device *rdev)
  */
 static void cik_pcie_gart_disable(struct radeon_device *rdev)
 {
+	unsigned i;
+
+	for (i = 1; i < 16; ++i) {
+		uint32_t reg;
+		if (i < 8)
+			reg = VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (i << 2);
+		else
+			reg = VM_CONTEXT8_PAGE_TABLE_BASE_ADDR + ((i - 8) << 2);
+		rdev->vm_manager.saved_table_addr[i] = RREG32(reg);
+	}
+
 	/* Disable all tables */
 	WREG32(VM_CONTEXT0_CNTL, 0);
 	WREG32(VM_CONTEXT1_CNTL, 0);
@@ -5641,12 +5963,13 @@ static void cik_vm_decode_fault(struct radeon_device *rdev,
 void cik_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm)
 {
 	struct radeon_ring *ring = &rdev->ring[ridx];
+	int usepfp = (ridx == RADEON_RING_TYPE_GFX_INDEX);
 
 	if (vm == NULL)
 		return;
 
 	radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
-	radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(0) |
+	radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(usepfp) |
 				 WRITE_DATA_DST_SEL(0)));
 	if (vm->id < 8) {
 		radeon_ring_write(ring,
@@ -5660,14 +5983,14 @@ void cik_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm)
 
 	/* update SH_MEM_* regs */
 	radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
-	radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(0) |
+	radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(usepfp) |
 				 WRITE_DATA_DST_SEL(0)));
 	radeon_ring_write(ring, SRBM_GFX_CNTL >> 2);
 	radeon_ring_write(ring, 0);
 	radeon_ring_write(ring, VMID(vm->id));
 
 	radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 6));
-	radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(0) |
+	radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(usepfp) |
 				 WRITE_DATA_DST_SEL(0)));
 	radeon_ring_write(ring, SH_MEM_BASES >> 2);
 	radeon_ring_write(ring, 0);
@@ -5678,7 +6001,7 @@ void cik_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm)
 	radeon_ring_write(ring, 0); /* SH_MEM_APE1_LIMIT */
 
 	radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
-	radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(0) |
+	radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(usepfp) |
 				 WRITE_DATA_DST_SEL(0)));
 	radeon_ring_write(ring, SRBM_GFX_CNTL >> 2);
 	radeon_ring_write(ring, 0);
@@ -5689,14 +6012,14 @@ void cik_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm)
 
 	/* bits 0-15 are the VM contexts0-15 */
 	radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
-	radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(0) |
+	radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(usepfp) |
 				 WRITE_DATA_DST_SEL(0)));
 	radeon_ring_write(ring, VM_INVALIDATE_REQUEST >> 2);
 	radeon_ring_write(ring, 0);
 	radeon_ring_write(ring, 1 << vm->id);
 
 	/* compute doesn't have PFP */
-	if (ridx == RADEON_RING_TYPE_GFX_INDEX) {
+	if (usepfp) {
 		/* sync PFP to ME, otherwise we might get invalid PFP reads */
 		radeon_ring_write(ring, PACKET3(PACKET3_PFP_SYNC_ME, 0));
 		radeon_ring_write(ring, 0x0);
@@ -5864,28 +6187,10 @@ static void cik_rlc_start(struct radeon_device *rdev)
 static int cik_rlc_resume(struct radeon_device *rdev)
 {
 	u32 i, size, tmp;
-	const __be32 *fw_data;
 
 	if (!rdev->rlc_fw)
 		return -EINVAL;
 
-	switch (rdev->family) {
-	case CHIP_BONAIRE:
-	case CHIP_HAWAII:
-	default:
-		size = BONAIRE_RLC_UCODE_SIZE;
-		break;
-	case CHIP_KAVERI:
-		size = KV_RLC_UCODE_SIZE;
-		break;
-	case CHIP_KABINI:
-		size = KB_RLC_UCODE_SIZE;
-		break;
-	case CHIP_MULLINS:
-		size = ML_RLC_UCODE_SIZE;
-		break;
-	}
-
 	cik_rlc_stop(rdev);
 
 	/* disable CG */
@@ -5909,11 +6214,45 @@ static int cik_rlc_resume(struct radeon_device *rdev)
 	WREG32(RLC_MC_CNTL, 0);
 	WREG32(RLC_UCODE_CNTL, 0);
 
-	fw_data = (const __be32 *)rdev->rlc_fw->data;
+	if (rdev->new_fw) {
+		const struct rlc_firmware_header_v1_0 *hdr =
+			(const struct rlc_firmware_header_v1_0 *)rdev->rlc_fw->data;
+		const __le32 *fw_data = (const __le32 *)
+			(rdev->rlc_fw->data + le32_to_cpu(hdr->header.ucode_array_offset_bytes));
+
+		radeon_ucode_print_rlc_hdr(&hdr->header);
+
+		size = le32_to_cpu(hdr->header.ucode_size_bytes) / 4;
+		WREG32(RLC_GPM_UCODE_ADDR, 0);
+		for (i = 0; i < size; i++)
+			WREG32(RLC_GPM_UCODE_DATA, le32_to_cpup(fw_data++));
+		WREG32(RLC_GPM_UCODE_ADDR, 0);
+	} else {
+		const __be32 *fw_data;
+
+		switch (rdev->family) {
+		case CHIP_BONAIRE:
+		case CHIP_HAWAII:
+		default:
+			size = BONAIRE_RLC_UCODE_SIZE;
+			break;
+		case CHIP_KAVERI:
+			size = KV_RLC_UCODE_SIZE;
+			break;
+		case CHIP_KABINI:
+			size = KB_RLC_UCODE_SIZE;
+			break;
+		case CHIP_MULLINS:
+			size = ML_RLC_UCODE_SIZE;
+			break;
+		}
+
+		fw_data = (const __be32 *)rdev->rlc_fw->data;
+		WREG32(RLC_GPM_UCODE_ADDR, 0);
+		for (i = 0; i < size; i++)
+			WREG32(RLC_GPM_UCODE_DATA, be32_to_cpup(fw_data++));
 		WREG32(RLC_GPM_UCODE_ADDR, 0);
-	for (i = 0; i < size; i++)
-		WREG32(RLC_GPM_UCODE_DATA, be32_to_cpup(fw_data++));
-	WREG32(RLC_GPM_UCODE_ADDR, 0);
+	}
 
 	/* XXX - find out what chips support lbpw */
 	cik_enable_lbpw(rdev, false);
@@ -6347,11 +6686,10 @@ static void cik_enable_gds_pg(struct radeon_device *rdev, bool enable)
 
 void cik_init_cp_pg_table(struct radeon_device *rdev)
 {
-	const __be32 *fw_data;
 	volatile u32 *dst_ptr;
 	int me, i, max_me = 4;
 	u32 bo_offset = 0;
-	u32 table_offset;
+	u32 table_offset, table_size;
 
 	if (rdev->family == CHIP_KAVERI)
 		max_me = 5;
@@ -6362,24 +6700,71 @@ void cik_init_cp_pg_table(struct radeon_device *rdev)
 	/* write the cp table buffer */
 	dst_ptr = rdev->rlc.cp_table_ptr;
 	for (me = 0; me < max_me; me++) {
-		if (me == 0) {
-			fw_data = (const __be32 *)rdev->ce_fw->data;
-			table_offset = CP_ME_TABLE_OFFSET;
-		} else if (me == 1) {
-			fw_data = (const __be32 *)rdev->pfp_fw->data;
-			table_offset = CP_ME_TABLE_OFFSET;
-		} else if (me == 2) {
-			fw_data = (const __be32 *)rdev->me_fw->data;
-			table_offset = CP_ME_TABLE_OFFSET;
+		if (rdev->new_fw) {
+			const __le32 *fw_data;
+			const struct gfx_firmware_header_v1_0 *hdr;
+
+			if (me == 0) {
+				hdr = (const struct gfx_firmware_header_v1_0 *)rdev->ce_fw->data;
+				fw_data = (const __le32 *)
+					(rdev->ce_fw->data + le32_to_cpu(hdr->header.ucode_array_offset_bytes));
+				table_offset = le32_to_cpu(hdr->jt_offset);
+				table_size = le32_to_cpu(hdr->jt_size);
+			} else if (me == 1) {
+				hdr = (const struct gfx_firmware_header_v1_0 *)rdev->pfp_fw->data;
+				fw_data = (const __le32 *)
+					(rdev->pfp_fw->data + le32_to_cpu(hdr->header.ucode_array_offset_bytes));
+				table_offset = le32_to_cpu(hdr->jt_offset);
+				table_size = le32_to_cpu(hdr->jt_size);
+			} else if (me == 2) {
+				hdr = (const struct gfx_firmware_header_v1_0 *)rdev->me_fw->data;
+				fw_data = (const __le32 *)
+					(rdev->me_fw->data + le32_to_cpu(hdr->header.ucode_array_offset_bytes));
+				table_offset = le32_to_cpu(hdr->jt_offset);
+				table_size = le32_to_cpu(hdr->jt_size);
+			} else if (me == 3) {
+				hdr = (const struct gfx_firmware_header_v1_0 *)rdev->mec_fw->data;
+				fw_data = (const __le32 *)
+					(rdev->mec_fw->data + le32_to_cpu(hdr->header.ucode_array_offset_bytes));
+				table_offset = le32_to_cpu(hdr->jt_offset);
+				table_size = le32_to_cpu(hdr->jt_size);
+			} else {
+				hdr = (const struct gfx_firmware_header_v1_0 *)rdev->mec2_fw->data;
+				fw_data = (const __le32 *)
+					(rdev->mec2_fw->data + le32_to_cpu(hdr->header.ucode_array_offset_bytes));
+				table_offset = le32_to_cpu(hdr->jt_offset);
+				table_size = le32_to_cpu(hdr->jt_size);
+			}
+
+			for (i = 0; i < table_size; i ++) {
+				dst_ptr[bo_offset + i] =
+					cpu_to_le32(le32_to_cpu(fw_data[table_offset + i]));
+			}
+			bo_offset += table_size;
 		} else {
-			fw_data = (const __be32 *)rdev->mec_fw->data;
-			table_offset = CP_MEC_TABLE_OFFSET;
-		}
+			const __be32 *fw_data;
+			table_size = CP_ME_TABLE_SIZE;
+
+			if (me == 0) {
+				fw_data = (const __be32 *)rdev->ce_fw->data;
+				table_offset = CP_ME_TABLE_OFFSET;
+			} else if (me == 1) {
+				fw_data = (const __be32 *)rdev->pfp_fw->data;
+				table_offset = CP_ME_TABLE_OFFSET;
+			} else if (me == 2) {
+				fw_data = (const __be32 *)rdev->me_fw->data;
+				table_offset = CP_ME_TABLE_OFFSET;
+			} else {
+				fw_data = (const __be32 *)rdev->mec_fw->data;
+				table_offset = CP_MEC_TABLE_OFFSET;
+			}
 
-		for (i = 0; i < CP_ME_TABLE_SIZE; i ++) {
-			dst_ptr[bo_offset + i] = cpu_to_le32(be32_to_cpu(fw_data[table_offset + i]));
+			for (i = 0; i < table_size; i ++) {
+				dst_ptr[bo_offset + i] =
+					cpu_to_le32(be32_to_cpu(fw_data[table_offset + i]));
+			}
+			bo_offset += table_size;
 		}
-		bo_offset += CP_ME_TABLE_SIZE;
 	}
 }
 
@@ -7376,6 +7761,7 @@ static inline u32 cik_get_ih_wptr(struct radeon_device *rdev)
 		tmp = RREG32(IH_RB_CNTL);
 		tmp |= IH_WPTR_OVERFLOW_CLEAR;
 		WREG32(IH_RB_CNTL, tmp);
+		wptr &= ~RB_OVERFLOW;
 	}
 	return (wptr & rdev->ih.ptr_mask);
 }
@@ -7616,7 +8002,8 @@ restart_ih:
 		case 16: /* D5 page flip */
 		case 18: /* D6 page flip */
 			DRM_DEBUG("IH: D%d flip\n", ((src_id - 8) >> 1) + 1);
-			radeon_crtc_handle_flip(rdev, (src_id - 8) >> 1);
+			if (radeon_use_pflipirq > 0)
+				radeon_crtc_handle_flip(rdev, (src_id - 8) >> 1);
 			break;
 		case 42: /* HPD hotplug */
 			switch (src_data) {
@@ -7898,6 +8285,7 @@ restart_ih:
 static int cik_startup(struct radeon_device *rdev)
 {
 	struct radeon_ring *ring;
+	u32 nop;
 	int r;
 
 	/* enable pcie gen2/3 link */
@@ -8031,9 +8419,18 @@ static int cik_startup(struct radeon_device *rdev)
 	}
 	cik_irq_set(rdev);
 
+	if (rdev->family == CHIP_HAWAII) {
+		if (rdev->new_fw)
+			nop = PACKET3(PACKET3_NOP, 0x3FFF);
+		else
+			nop = RADEON_CP_PACKET2;
+	} else {
+		nop = PACKET3(PACKET3_NOP, 0x3FFF);
+	}
+
 	ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
 	r = radeon_ring_init(rdev, ring, ring->ring_size, RADEON_WB_CP_RPTR_OFFSET,
-			     PACKET3(PACKET3_NOP, 0x3FFF));
+			     nop);
 	if (r)
 		return r;
 
@@ -8041,7 +8438,7 @@ static int cik_startup(struct radeon_device *rdev)
 	/* type-2 packets are deprecated on MEC, use type-3 instead */
 	ring = &rdev->ring[CAYMAN_RING_TYPE_CP1_INDEX];
 	r = radeon_ring_init(rdev, ring, ring->ring_size, RADEON_WB_CP1_RPTR_OFFSET,
-			     PACKET3(PACKET3_NOP, 0x3FFF));
+			     nop);
 	if (r)
 		return r;
 	ring->me = 1; /* first MEC */
@@ -8052,7 +8449,7 @@ static int cik_startup(struct radeon_device *rdev)
 	/* type-2 packets are deprecated on MEC, use type-3 instead */
 	ring = &rdev->ring[CAYMAN_RING_TYPE_CP2_INDEX];
 	r = radeon_ring_init(rdev, ring, ring->ring_size, RADEON_WB_CP2_RPTR_OFFSET,
-			     PACKET3(PACKET3_NOP, 0x3FFF));
+			     nop);
 	if (r)
 		return r;
 	/* dGPU only have 1 MEC */
@@ -9166,6 +9563,9 @@ static void cik_pcie_gen3_enable(struct radeon_device *rdev)
 	int ret, i;
 	u16 tmp16;
 
+	if (pci_is_root_bus(rdev->pdev->bus))
+		return;
+
 	if (radeon_pcie_gen2 == 0)
 		return;
 
@@ -9392,7 +9792,8 @@ static void cik_program_aspm(struct radeon_device *rdev)
 			if (orig != data)
 				WREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL, data);
 
-			if (!disable_clkreq) {
+			if (!disable_clkreq &&
+			    !pci_is_root_bus(rdev->pdev->bus)) {
 				struct pci_dev *root = rdev->pdev->bus->self;
 				u32 lnkcap;
 
diff --git a/drivers/gpu/drm/radeon/cik_sdma.c b/drivers/gpu/drm/radeon/cik_sdma.c
index 8e9d0f1d858e..192278bc993c 100644
--- a/drivers/gpu/drm/radeon/cik_sdma.c
+++ b/drivers/gpu/drm/radeon/cik_sdma.c
@@ -24,6 +24,7 @@
 #include <linux/firmware.h>
 #include <drm/drmP.h>
 #include "radeon.h"
+#include "radeon_ucode.h"
 #include "radeon_asic.h"
 #include "radeon_trace.h"
 #include "cikd.h"
@@ -118,6 +119,7 @@ void cik_sdma_set_wptr(struct radeon_device *rdev,
 		reg = SDMA0_GFX_RB_WPTR + SDMA1_REGISTER_OFFSET;
 
 	WREG32(reg, (ring->wptr << 2) & 0x3fffc);
+	(void)RREG32(reg);
 }
 
 /**
@@ -419,7 +421,6 @@ static int cik_sdma_rlc_resume(struct radeon_device *rdev)
  */
 static int cik_sdma_load_microcode(struct radeon_device *rdev)
 {
-	const __be32 *fw_data;
 	int i;
 
 	if (!rdev->sdma_fw)
@@ -428,19 +429,48 @@ static int cik_sdma_load_microcode(struct radeon_device *rdev)
 	/* halt the MEs */
 	cik_sdma_enable(rdev, false);
 
-	/* sdma0 */
-	fw_data = (const __be32 *)rdev->sdma_fw->data;
-	WREG32(SDMA0_UCODE_ADDR + SDMA0_REGISTER_OFFSET, 0);
-	for (i = 0; i < CIK_SDMA_UCODE_SIZE; i++)
-		WREG32(SDMA0_UCODE_DATA + SDMA0_REGISTER_OFFSET, be32_to_cpup(fw_data++));
-	WREG32(SDMA0_UCODE_DATA + SDMA0_REGISTER_OFFSET, CIK_SDMA_UCODE_VERSION);
-
-	/* sdma1 */
-	fw_data = (const __be32 *)rdev->sdma_fw->data;
-	WREG32(SDMA0_UCODE_ADDR + SDMA1_REGISTER_OFFSET, 0);
-	for (i = 0; i < CIK_SDMA_UCODE_SIZE; i++)
-		WREG32(SDMA0_UCODE_DATA + SDMA1_REGISTER_OFFSET, be32_to_cpup(fw_data++));
-	WREG32(SDMA0_UCODE_DATA + SDMA1_REGISTER_OFFSET, CIK_SDMA_UCODE_VERSION);
+	if (rdev->new_fw) {
+		const struct sdma_firmware_header_v1_0 *hdr =
+			(const struct sdma_firmware_header_v1_0 *)rdev->sdma_fw->data;
+		const __le32 *fw_data;
+		u32 fw_size;
+
+		radeon_ucode_print_sdma_hdr(&hdr->header);
+
+		/* sdma0 */
+		fw_data = (const __le32 *)
+			(rdev->sdma_fw->data + le32_to_cpu(hdr->header.ucode_array_offset_bytes));
+		fw_size = le32_to_cpu(hdr->header.ucode_size_bytes) / 4;
+		WREG32(SDMA0_UCODE_ADDR + SDMA0_REGISTER_OFFSET, 0);
+		for (i = 0; i < fw_size; i++)
+			WREG32(SDMA0_UCODE_DATA + SDMA0_REGISTER_OFFSET, le32_to_cpup(fw_data++));
+		WREG32(SDMA0_UCODE_DATA + SDMA0_REGISTER_OFFSET, CIK_SDMA_UCODE_VERSION);
+
+		/* sdma1 */
+		fw_data = (const __le32 *)
+			(rdev->sdma_fw->data + le32_to_cpu(hdr->header.ucode_array_offset_bytes));
+		fw_size = le32_to_cpu(hdr->header.ucode_size_bytes) / 4;
+		WREG32(SDMA0_UCODE_ADDR + SDMA1_REGISTER_OFFSET, 0);
+		for (i = 0; i < fw_size; i++)
+			WREG32(SDMA0_UCODE_DATA + SDMA1_REGISTER_OFFSET, le32_to_cpup(fw_data++));
+		WREG32(SDMA0_UCODE_DATA + SDMA1_REGISTER_OFFSET, CIK_SDMA_UCODE_VERSION);
+	} else {
+		const __be32 *fw_data;
+
+		/* sdma0 */
+		fw_data = (const __be32 *)rdev->sdma_fw->data;
+		WREG32(SDMA0_UCODE_ADDR + SDMA0_REGISTER_OFFSET, 0);
+		for (i = 0; i < CIK_SDMA_UCODE_SIZE; i++)
+			WREG32(SDMA0_UCODE_DATA + SDMA0_REGISTER_OFFSET, be32_to_cpup(fw_data++));
+		WREG32(SDMA0_UCODE_DATA + SDMA0_REGISTER_OFFSET, CIK_SDMA_UCODE_VERSION);
+
+		/* sdma1 */
+		fw_data = (const __be32 *)rdev->sdma_fw->data;
+		WREG32(SDMA0_UCODE_ADDR + SDMA1_REGISTER_OFFSET, 0);
+		for (i = 0; i < CIK_SDMA_UCODE_SIZE; i++)
+			WREG32(SDMA0_UCODE_DATA + SDMA1_REGISTER_OFFSET, be32_to_cpup(fw_data++));
+		WREG32(SDMA0_UCODE_DATA + SDMA1_REGISTER_OFFSET, CIK_SDMA_UCODE_VERSION);
+	}
 
 	WREG32(SDMA0_UCODE_ADDR + SDMA0_REGISTER_OFFSET, 0);
 	WREG32(SDMA0_UCODE_ADDR + SDMA1_REGISTER_OFFSET, 0);
@@ -566,7 +596,7 @@ int cik_copy_dma(struct radeon_device *rdev,
 		return r;
 	}
 
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev, ring, false);
 	radeon_semaphore_free(rdev, &sem, *fence);
 
 	return r;
@@ -608,7 +638,7 @@ int cik_sdma_ring_test(struct radeon_device *rdev,
 	radeon_ring_write(ring, upper_32_bits(rdev->vram_scratch.gpu_addr));
 	radeon_ring_write(ring, 1); /* number of DWs to follow */
 	radeon_ring_write(ring, 0xDEADBEEF);
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev, ring, false);
 
 	for (i = 0; i < rdev->usec_timeout; i++) {
 		tmp = readl(ptr);
@@ -665,7 +695,7 @@ int cik_sdma_ib_test(struct radeon_device *rdev, struct radeon_ring *ring)
 	ib.ptr[4] = 0xDEADBEEF;
 	ib.length_dw = 5;
 
-	r = radeon_ib_schedule(rdev, &ib, NULL);
+	r = radeon_ib_schedule(rdev, &ib, NULL, false);
 	if (r) {
 		radeon_ib_free(rdev, &ib);
 		DRM_ERROR("radeon: failed to schedule ib (%d).\n", r);
@@ -719,7 +749,43 @@ bool cik_sdma_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring)
 }
 
 /**
- * cik_sdma_vm_set_page - update the page tables using sDMA
+ * cik_sdma_vm_copy_pages - update PTEs by copying them from the GART
+ *
+ * @rdev: radeon_device pointer
+ * @ib: indirect buffer to fill with commands
+ * @pe: addr of the page entry
+ * @src: src addr to copy from
+ * @count: number of page entries to update
+ *
+ * Update PTEs by copying them from the GART using sDMA (CIK).
+ */
+void cik_sdma_vm_copy_pages(struct radeon_device *rdev,
+			    struct radeon_ib *ib,
+			    uint64_t pe, uint64_t src,
+			    unsigned count)
+{
+	while (count) {
+		unsigned bytes = count * 8;
+		if (bytes > 0x1FFFF8)
+			bytes = 0x1FFFF8;
+
+		ib->ptr[ib->length_dw++] = SDMA_PACKET(SDMA_OPCODE_COPY,
+			SDMA_WRITE_SUB_OPCODE_LINEAR, 0);
+		ib->ptr[ib->length_dw++] = bytes;
+		ib->ptr[ib->length_dw++] = 0; /* src/dst endian swap */
+		ib->ptr[ib->length_dw++] = lower_32_bits(src);
+		ib->ptr[ib->length_dw++] = upper_32_bits(src);
+		ib->ptr[ib->length_dw++] = lower_32_bits(pe);
+		ib->ptr[ib->length_dw++] = upper_32_bits(pe);
+
+		pe += bytes;
+		src += bytes;
+		count -= bytes / 8;
+	}
+}
+
+/**
+ * cik_sdma_vm_write_pages - update PTEs by writing them manually
  *
  * @rdev: radeon_device pointer
  * @ib: indirect buffer to fill with commands
@@ -729,84 +795,103 @@ bool cik_sdma_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring)
  * @incr: increase next addr by incr bytes
  * @flags: access flags
  *
- * Update the page tables using sDMA (CIK).
+ * Update PTEs by writing them manually using sDMA (CIK).
  */
-void cik_sdma_vm_set_page(struct radeon_device *rdev,
-			  struct radeon_ib *ib,
-			  uint64_t pe,
-			  uint64_t addr, unsigned count,
-			  uint32_t incr, uint32_t flags)
+void cik_sdma_vm_write_pages(struct radeon_device *rdev,
+			     struct radeon_ib *ib,
+			     uint64_t pe,
+			     uint64_t addr, unsigned count,
+			     uint32_t incr, uint32_t flags)
 {
 	uint64_t value;
 	unsigned ndw;
 
-	trace_radeon_vm_set_page(pe, addr, count, incr, flags);
-
-	if (flags == R600_PTE_GART) {
-		uint64_t src = rdev->gart.table_addr + (addr >> 12) * 8;
-		while (count) {
-			unsigned bytes = count * 8;
-			if (bytes > 0x1FFFF8)
-				bytes = 0x1FFFF8;
-
-			ib->ptr[ib->length_dw++] = SDMA_PACKET(SDMA_OPCODE_COPY, SDMA_WRITE_SUB_OPCODE_LINEAR, 0);
-			ib->ptr[ib->length_dw++] = bytes;
-			ib->ptr[ib->length_dw++] = 0; /* src/dst endian swap */
-			ib->ptr[ib->length_dw++] = lower_32_bits(src);
-			ib->ptr[ib->length_dw++] = upper_32_bits(src);
-			ib->ptr[ib->length_dw++] = lower_32_bits(pe);
-			ib->ptr[ib->length_dw++] = upper_32_bits(pe);
-
-			pe += bytes;
-			src += bytes;
-			count -= bytes / 8;
-		}
-	} else if (flags & R600_PTE_SYSTEM) {
-		while (count) {
-			ndw = count * 2;
-			if (ndw > 0xFFFFE)
-				ndw = 0xFFFFE;
-
-			/* for non-physically contiguous pages (system) */
-			ib->ptr[ib->length_dw++] = SDMA_PACKET(SDMA_OPCODE_WRITE, SDMA_WRITE_SUB_OPCODE_LINEAR, 0);
-			ib->ptr[ib->length_dw++] = pe;
-			ib->ptr[ib->length_dw++] = upper_32_bits(pe);
-			ib->ptr[ib->length_dw++] = ndw;
-			for (; ndw > 0; ndw -= 2, --count, pe += 8) {
+	while (count) {
+		ndw = count * 2;
+		if (ndw > 0xFFFFE)
+			ndw = 0xFFFFE;
+
+		/* for non-physically contiguous pages (system) */
+		ib->ptr[ib->length_dw++] = SDMA_PACKET(SDMA_OPCODE_WRITE,
+			SDMA_WRITE_SUB_OPCODE_LINEAR, 0);
+		ib->ptr[ib->length_dw++] = pe;
+		ib->ptr[ib->length_dw++] = upper_32_bits(pe);
+		ib->ptr[ib->length_dw++] = ndw;
+		for (; ndw > 0; ndw -= 2, --count, pe += 8) {
+			if (flags & R600_PTE_SYSTEM) {
 				value = radeon_vm_map_gart(rdev, addr);
 				value &= 0xFFFFFFFFFFFFF000ULL;
-				addr += incr;
-				value |= flags;
-				ib->ptr[ib->length_dw++] = value;
-				ib->ptr[ib->length_dw++] = upper_32_bits(value);
-			}
-		}
-	} else {
-		while (count) {
-			ndw = count;
-			if (ndw > 0x7FFFF)
-				ndw = 0x7FFFF;
-
-			if (flags & R600_PTE_VALID)
+			} else if (flags & R600_PTE_VALID) {
 				value = addr;
-			else
+			} else {
 				value = 0;
-			/* for physically contiguous pages (vram) */
-			ib->ptr[ib->length_dw++] = SDMA_PACKET(SDMA_OPCODE_GENERATE_PTE_PDE, 0, 0);
-			ib->ptr[ib->length_dw++] = pe; /* dst addr */
-			ib->ptr[ib->length_dw++] = upper_32_bits(pe);
-			ib->ptr[ib->length_dw++] = flags; /* mask */
-			ib->ptr[ib->length_dw++] = 0;
-			ib->ptr[ib->length_dw++] = value; /* value */
+			}
+			addr += incr;
+			value |= flags;
+			ib->ptr[ib->length_dw++] = value;
 			ib->ptr[ib->length_dw++] = upper_32_bits(value);
-			ib->ptr[ib->length_dw++] = incr; /* increment size */
-			ib->ptr[ib->length_dw++] = 0;
-			ib->ptr[ib->length_dw++] = ndw; /* number of entries */
-			pe += ndw * 8;
-			addr += ndw * incr;
-			count -= ndw;
 		}
 	}
+}
+
+/**
+ * cik_sdma_vm_set_pages - update the page tables using sDMA
+ *
+ * @rdev: radeon_device pointer
+ * @ib: indirect buffer to fill with commands
+ * @pe: addr of the page entry
+ * @addr: dst addr to write into pe
+ * @count: number of page entries to update
+ * @incr: increase next addr by incr bytes
+ * @flags: access flags
+ *
+ * Update the page tables using sDMA (CIK).
+ */
+void cik_sdma_vm_set_pages(struct radeon_device *rdev,
+			   struct radeon_ib *ib,
+			   uint64_t pe,
+			   uint64_t addr, unsigned count,
+			   uint32_t incr, uint32_t flags)
+{
+	uint64_t value;
+	unsigned ndw;
+
+	while (count) {
+		ndw = count;
+		if (ndw > 0x7FFFF)
+			ndw = 0x7FFFF;
+
+		if (flags & R600_PTE_VALID)
+			value = addr;
+		else
+			value = 0;
+
+		/* for physically contiguous pages (vram) */
+		ib->ptr[ib->length_dw++] = SDMA_PACKET(SDMA_OPCODE_GENERATE_PTE_PDE, 0, 0);
+		ib->ptr[ib->length_dw++] = pe; /* dst addr */
+		ib->ptr[ib->length_dw++] = upper_32_bits(pe);
+		ib->ptr[ib->length_dw++] = flags; /* mask */
+		ib->ptr[ib->length_dw++] = 0;
+		ib->ptr[ib->length_dw++] = value; /* value */
+		ib->ptr[ib->length_dw++] = upper_32_bits(value);
+		ib->ptr[ib->length_dw++] = incr; /* increment size */
+		ib->ptr[ib->length_dw++] = 0;
+		ib->ptr[ib->length_dw++] = ndw; /* number of entries */
+
+		pe += ndw * 8;
+		addr += ndw * incr;
+		count -= ndw;
+	}
+}
+
+/**
+ * cik_sdma_vm_pad_ib - pad the IB to the required number of dw
+ *
+ * @ib: indirect buffer to fill with padding
+ *
+ */
+void cik_sdma_vm_pad_ib(struct radeon_ib *ib)
+{
 	while (ib->length_dw & 0x7)
 		ib->ptr[ib->length_dw++] = SDMA_PACKET(SDMA_OPCODE_NOP, 0, 0);
 }
diff --git a/drivers/gpu/drm/radeon/dce6_afmt.c b/drivers/gpu/drm/radeon/dce6_afmt.c
index 0a65dc7e93e7..ab29f953a767 100644
--- a/drivers/gpu/drm/radeon/dce6_afmt.c
+++ b/drivers/gpu/drm/radeon/dce6_afmt.c
@@ -136,13 +136,13 @@ void dce6_afmt_write_latency_fields(struct drm_encoder *encoder,
 			tmp = VIDEO_LIPSYNC(connector->video_latency[1]) |
 				AUDIO_LIPSYNC(connector->audio_latency[1]);
 		else
-			tmp = VIDEO_LIPSYNC(255) | AUDIO_LIPSYNC(255);
+			tmp = VIDEO_LIPSYNC(0) | AUDIO_LIPSYNC(0);
 	} else {
 		if (connector->latency_present[0])
 			tmp = VIDEO_LIPSYNC(connector->video_latency[0]) |
 				AUDIO_LIPSYNC(connector->audio_latency[0]);
 		else
-			tmp = VIDEO_LIPSYNC(255) | AUDIO_LIPSYNC(255);
+			tmp = VIDEO_LIPSYNC(0) | AUDIO_LIPSYNC(0);
 	}
 	WREG32_ENDPOINT(offset, AZ_F0_CODEC_PIN_CONTROL_RESPONSE_LIPSYNC, tmp);
 }
@@ -164,8 +164,10 @@ void dce6_afmt_write_speaker_allocation(struct drm_encoder *encoder)
 	offset = dig->afmt->pin->offset;
 
 	list_for_each_entry(connector, &encoder->dev->mode_config.connector_list, head) {
-		if (connector->encoder == encoder)
+		if (connector->encoder == encoder) {
 			radeon_connector = to_radeon_connector(connector);
+			break;
+		}
 	}
 
 	if (!radeon_connector) {
@@ -173,7 +175,7 @@ void dce6_afmt_write_speaker_allocation(struct drm_encoder *encoder)
 		return;
 	}
 
-	sad_count = drm_edid_to_speaker_allocation(radeon_connector->edid, &sadb);
+	sad_count = drm_edid_to_speaker_allocation(radeon_connector_edid(connector), &sadb);
 	if (sad_count <= 0) {
 		DRM_ERROR("Couldn't read Speaker Allocation Data Block: %d\n", sad_count);
 		return;
@@ -225,8 +227,10 @@ void dce6_afmt_write_sad_regs(struct drm_encoder *encoder)
 	offset = dig->afmt->pin->offset;
 
 	list_for_each_entry(connector, &encoder->dev->mode_config.connector_list, head) {
-		if (connector->encoder == encoder)
+		if (connector->encoder == encoder) {
 			radeon_connector = to_radeon_connector(connector);
+			break;
+		}
 	}
 
 	if (!radeon_connector) {
@@ -234,7 +238,7 @@ void dce6_afmt_write_sad_regs(struct drm_encoder *encoder)
 		return;
 	}
 
-	sad_count = drm_edid_to_sad(radeon_connector->edid, &sads);
+	sad_count = drm_edid_to_sad(radeon_connector_edid(connector), &sads);
 	if (sad_count <= 0) {
 		DRM_ERROR("Couldn't read SADs: %d\n", sad_count);
 		return;
diff --git a/drivers/gpu/drm/radeon/evergreen.c b/drivers/gpu/drm/radeon/evergreen.c
index 250bac3935a4..dbca60c7d097 100644
--- a/drivers/gpu/drm/radeon/evergreen.c
+++ b/drivers/gpu/drm/radeon/evergreen.c
@@ -2424,7 +2424,6 @@ static int evergreen_pcie_gart_enable(struct radeon_device *rdev)
 	r = radeon_gart_table_vram_pin(rdev);
 	if (r)
 		return r;
-	radeon_gart_restore(rdev);
 	/* Setup L2 cache */
 	WREG32(VM_L2_CNTL, ENABLE_L2_CACHE | ENABLE_L2_FRAGMENT_PROCESSING |
 				ENABLE_L2_PTE_CACHE_LRU_UPDATE_BY_WRITE |
@@ -2677,7 +2676,7 @@ void evergreen_mc_resume(struct radeon_device *rdev, struct evergreen_mc_save *s
 		if (save->crtc_enabled[i]) {
 			if (ASIC_IS_DCE6(rdev)) {
 				tmp = RREG32(EVERGREEN_CRTC_BLANK_CONTROL + crtc_offsets[i]);
-				tmp |= EVERGREEN_CRTC_BLANK_DATA_EN;
+				tmp &= ~EVERGREEN_CRTC_BLANK_DATA_EN;
 				WREG32(EVERGREEN_CRTC_UPDATE_LOCK + crtc_offsets[i], 1);
 				WREG32(EVERGREEN_CRTC_BLANK_CONTROL + crtc_offsets[i], tmp);
 				WREG32(EVERGREEN_CRTC_UPDATE_LOCK + crtc_offsets[i], 0);
@@ -2870,7 +2869,7 @@ static int evergreen_cp_start(struct radeon_device *rdev)
 	radeon_ring_write(ring, PACKET3_ME_INITIALIZE_DEVICE_ID(1));
 	radeon_ring_write(ring, 0);
 	radeon_ring_write(ring, 0);
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev, ring, false);
 
 	cp_me = 0xff;
 	WREG32(CP_ME_CNTL, cp_me);
@@ -2913,7 +2912,7 @@ static int evergreen_cp_start(struct radeon_device *rdev)
 	radeon_ring_write(ring, 0x0000000e); /* VGT_VERTEX_REUSE_BLOCK_CNTL */
 	radeon_ring_write(ring, 0x00000010); /*  */
 
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev, ring, false);
 
 	return 0;
 }
@@ -4023,7 +4022,8 @@ int sumo_rlc_init(struct radeon_device *rdev)
 		/* save restore block */
 		if (rdev->rlc.save_restore_obj == NULL) {
 			r = radeon_bo_create(rdev, dws * 4, PAGE_SIZE, true,
-					     RADEON_GEM_DOMAIN_VRAM, NULL, &rdev->rlc.save_restore_obj);
+					     RADEON_GEM_DOMAIN_VRAM, 0, NULL,
+					     &rdev->rlc.save_restore_obj);
 			if (r) {
 				dev_warn(rdev->dev, "(%d) create RLC sr bo failed\n", r);
 				return r;
@@ -4101,7 +4101,8 @@ int sumo_rlc_init(struct radeon_device *rdev)
 
 		if (rdev->rlc.clear_state_obj == NULL) {
 			r = radeon_bo_create(rdev, dws * 4, PAGE_SIZE, true,
-					     RADEON_GEM_DOMAIN_VRAM, NULL, &rdev->rlc.clear_state_obj);
+					     RADEON_GEM_DOMAIN_VRAM, 0, NULL,
+					     &rdev->rlc.clear_state_obj);
 			if (r) {
 				dev_warn(rdev->dev, "(%d) create RLC c bo failed\n", r);
 				sumo_rlc_fini(rdev);
@@ -4175,8 +4176,10 @@ int sumo_rlc_init(struct radeon_device *rdev)
 
 	if (rdev->rlc.cp_table_size) {
 		if (rdev->rlc.cp_table_obj == NULL) {
-			r = radeon_bo_create(rdev, rdev->rlc.cp_table_size, PAGE_SIZE, true,
-					     RADEON_GEM_DOMAIN_VRAM, NULL, &rdev->rlc.cp_table_obj);
+			r = radeon_bo_create(rdev, rdev->rlc.cp_table_size,
+					     PAGE_SIZE, true,
+					     RADEON_GEM_DOMAIN_VRAM, 0, NULL,
+					     &rdev->rlc.cp_table_obj);
 			if (r) {
 				dev_warn(rdev->dev, "(%d) create RLC cp table bo failed\n", r);
 				sumo_rlc_fini(rdev);
@@ -4756,6 +4759,7 @@ static u32 evergreen_get_ih_wptr(struct radeon_device *rdev)
 		tmp = RREG32(IH_RB_CNTL);
 		tmp |= IH_WPTR_OVERFLOW_CLEAR;
 		WREG32(IH_RB_CNTL, tmp);
+		wptr &= ~RB_OVERFLOW;
 	}
 	return (wptr & rdev->ih.ptr_mask);
 }
@@ -4960,7 +4964,8 @@ restart_ih:
 		case 16: /* D5 page flip */
 		case 18: /* D6 page flip */
 			DRM_DEBUG("IH: D%d flip\n", ((src_id - 8) >> 1) + 1);
-			radeon_crtc_handle_flip(rdev, (src_id - 8) >> 1);
+			if (radeon_use_pflipirq > 0)
+				radeon_crtc_handle_flip(rdev, (src_id - 8) >> 1);
 			break;
 		case 42: /* HPD hotplug */
 			switch (src_data) {
diff --git a/drivers/gpu/drm/radeon/evergreen_dma.c b/drivers/gpu/drm/radeon/evergreen_dma.c
index 478caefe0fef..afaba388c36d 100644
--- a/drivers/gpu/drm/radeon/evergreen_dma.c
+++ b/drivers/gpu/drm/radeon/evergreen_dma.c
@@ -155,7 +155,7 @@ int evergreen_copy_dma(struct radeon_device *rdev,
 		return r;
 	}
 
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev, ring, false);
 	radeon_semaphore_free(rdev, &sem, *fence);
 
 	return r;
diff --git a/drivers/gpu/drm/radeon/evergreen_hdmi.c b/drivers/gpu/drm/radeon/evergreen_hdmi.c
index 1ec0e6e83f9f..278c7a139d74 100644
--- a/drivers/gpu/drm/radeon/evergreen_hdmi.c
+++ b/drivers/gpu/drm/radeon/evergreen_hdmi.c
@@ -117,7 +117,7 @@ static void dce4_afmt_write_speaker_allocation(struct drm_encoder *encoder)
 		return;
 	}
 
-	sad_count = drm_edid_to_speaker_allocation(radeon_connector->edid, &sadb);
+	sad_count = drm_edid_to_speaker_allocation(radeon_connector_edid(connector), &sadb);
 	if (sad_count <= 0) {
 		DRM_ERROR("Couldn't read Speaker Allocation Data Block: %d\n", sad_count);
 		return;
@@ -172,7 +172,7 @@ static void evergreen_hdmi_write_sad_regs(struct drm_encoder *encoder)
 		return;
 	}
 
-	sad_count = drm_edid_to_sad(radeon_connector->edid, &sads);
+	sad_count = drm_edid_to_sad(radeon_connector_edid(connector), &sads);
 	if (sad_count <= 0) {
 		DRM_ERROR("Couldn't read SADs: %d\n", sad_count);
 		return;
diff --git a/drivers/gpu/drm/radeon/kv_dpm.c b/drivers/gpu/drm/radeon/kv_dpm.c
index 9ef8c38f2d66..8b58e11b64fa 100644
--- a/drivers/gpu/drm/radeon/kv_dpm.c
+++ b/drivers/gpu/drm/radeon/kv_dpm.c
@@ -1438,14 +1438,14 @@ static int kv_update_uvd_dpm(struct radeon_device *rdev, bool gate)
 	return kv_enable_uvd_dpm(rdev, !gate);
 }
 
-static u8 kv_get_vce_boot_level(struct radeon_device *rdev)
+static u8 kv_get_vce_boot_level(struct radeon_device *rdev, u32 evclk)
 {
 	u8 i;
 	struct radeon_vce_clock_voltage_dependency_table *table =
 		&rdev->pm.dpm.dyn_state.vce_clock_voltage_dependency_table;
 
 	for (i = 0; i < table->count; i++) {
-		if (table->entries[i].evclk >= 0) /* XXX */
+		if (table->entries[i].evclk >= evclk)
 			break;
 	}
 
@@ -1468,7 +1468,7 @@ static int kv_update_vce_dpm(struct radeon_device *rdev,
 		if (pi->caps_stable_p_state)
 			pi->vce_boot_level = table->count - 1;
 		else
-			pi->vce_boot_level = kv_get_vce_boot_level(rdev);
+			pi->vce_boot_level = kv_get_vce_boot_level(rdev, radeon_new_state->evclk);
 
 		ret = kv_copy_bytes_to_smc(rdev,
 					   pi->dpm_table_start +
@@ -2726,7 +2726,10 @@ int kv_dpm_init(struct radeon_device *rdev)
 	pi->caps_sclk_ds = true;
 	pi->enable_auto_thermal_throttling = true;
 	pi->disable_nb_ps3_in_battery = false;
-	pi->bapm_enable = true;
+	if (radeon_bapm == 0)
+		pi->bapm_enable = false;
+	else
+		pi->bapm_enable = true;
 	pi->voltage_drop_t = 0;
 	pi->caps_sclk_throttle_low_notification = false;
 	pi->caps_fps = false; /* true? */
diff --git a/drivers/gpu/drm/radeon/ni.c b/drivers/gpu/drm/radeon/ni.c
index 5a33ca681867..3faee58946dd 100644
--- a/drivers/gpu/drm/radeon/ni.c
+++ b/drivers/gpu/drm/radeon/ni.c
@@ -1229,7 +1229,6 @@ static int cayman_pcie_gart_enable(struct radeon_device *rdev)
 	r = radeon_gart_table_vram_pin(rdev);
 	if (r)
 		return r;
-	radeon_gart_restore(rdev);
 	/* Setup TLB control */
 	WREG32(MC_VM_MX_L1_TLB_CNTL,
 	       (0xA << 7) |
@@ -1272,7 +1271,7 @@ static int cayman_pcie_gart_enable(struct radeon_device *rdev)
 		WREG32(VM_CONTEXT0_PAGE_TABLE_START_ADDR + (i << 2), 0);
 		WREG32(VM_CONTEXT0_PAGE_TABLE_END_ADDR + (i << 2), rdev->vm_manager.max_pfn);
 		WREG32(VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (i << 2),
-			rdev->gart.table_addr >> 12);
+		       rdev->vm_manager.saved_table_addr[i]);
 	}
 
 	/* enable context1-7 */
@@ -1304,6 +1303,13 @@ static int cayman_pcie_gart_enable(struct radeon_device *rdev)
 
 static void cayman_pcie_gart_disable(struct radeon_device *rdev)
 {
+	unsigned i;
+
+	for (i = 1; i < 8; ++i) {
+		rdev->vm_manager.saved_table_addr[i] = RREG32(
+			VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (i << 2));
+	}
+
 	/* Disable all tables */
 	WREG32(VM_CONTEXT0_CNTL, 0);
 	WREG32(VM_CONTEXT1_CNTL, 0);
@@ -1506,7 +1512,7 @@ static int cayman_cp_start(struct radeon_device *rdev)
 	radeon_ring_write(ring, PACKET3_ME_INITIALIZE_DEVICE_ID(1));
 	radeon_ring_write(ring, 0);
 	radeon_ring_write(ring, 0);
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev, ring, false);
 
 	cayman_cp_enable(rdev, true);
 
@@ -1548,7 +1554,7 @@ static int cayman_cp_start(struct radeon_device *rdev)
 	radeon_ring_write(ring, 0x0000000e); /* VGT_VERTEX_REUSE_BLOCK_CNTL */
 	radeon_ring_write(ring, 0x00000010); /*  */
 
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev, ring, false);
 
 	/* XXX init other rings */
 
diff --git a/drivers/gpu/drm/radeon/ni_dma.c b/drivers/gpu/drm/radeon/ni_dma.c
index 6378e0276691..8a3e6221cece 100644
--- a/drivers/gpu/drm/radeon/ni_dma.c
+++ b/drivers/gpu/drm/radeon/ni_dma.c
@@ -307,7 +307,43 @@ bool cayman_dma_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring)
 }
 
 /**
- * cayman_dma_vm_set_page - update the page tables using the DMA
+ * cayman_dma_vm_copy_pages - update PTEs by copying them from the GART
+ *
+ * @rdev: radeon_device pointer
+ * @ib: indirect buffer to fill with commands
+ * @pe: addr of the page entry
+ * @src: src addr where to copy from
+ * @count: number of page entries to update
+ *
+ * Update PTEs by copying them from the GART using the DMA (cayman/TN).
+ */
+void cayman_dma_vm_copy_pages(struct radeon_device *rdev,
+			      struct radeon_ib *ib,
+			      uint64_t pe, uint64_t src,
+			      unsigned count)
+{
+	unsigned ndw;
+
+	while (count) {
+		ndw = count * 2;
+		if (ndw > 0xFFFFE)
+			ndw = 0xFFFFE;
+
+		ib->ptr[ib->length_dw++] = DMA_PACKET(DMA_PACKET_COPY,
+						      0, 0, ndw);
+		ib->ptr[ib->length_dw++] = lower_32_bits(pe);
+		ib->ptr[ib->length_dw++] = lower_32_bits(src);
+		ib->ptr[ib->length_dw++] = upper_32_bits(pe) & 0xff;
+		ib->ptr[ib->length_dw++] = upper_32_bits(src) & 0xff;
+
+		pe += ndw * 4;
+		src += ndw * 4;
+		count -= ndw / 2;
+	}
+}
+
+/**
+ * cayman_dma_vm_write_pages - update PTEs by writing them manually
  *
  * @rdev: radeon_device pointer
  * @ib: indirect buffer to fill with commands
@@ -315,71 +351,103 @@ bool cayman_dma_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring)
  * @addr: dst addr to write into pe
  * @count: number of page entries to update
  * @incr: increase next addr by incr bytes
- * @flags: hw access flags 
+ * @flags: hw access flags
  *
- * Update the page tables using the DMA (cayman/TN).
+ * Update PTEs by writing them manually using the DMA (cayman/TN).
  */
-void cayman_dma_vm_set_page(struct radeon_device *rdev,
-			    struct radeon_ib *ib,
-			    uint64_t pe,
-			    uint64_t addr, unsigned count,
-			    uint32_t incr, uint32_t flags)
+void cayman_dma_vm_write_pages(struct radeon_device *rdev,
+			       struct radeon_ib *ib,
+			       uint64_t pe,
+			       uint64_t addr, unsigned count,
+			       uint32_t incr, uint32_t flags)
 {
 	uint64_t value;
 	unsigned ndw;
 
-	trace_radeon_vm_set_page(pe, addr, count, incr, flags);
-
-	if ((flags & R600_PTE_SYSTEM) || (count == 1)) {
-		while (count) {
-			ndw = count * 2;
-			if (ndw > 0xFFFFE)
-				ndw = 0xFFFFE;
-
-			/* for non-physically contiguous pages (system) */
-			ib->ptr[ib->length_dw++] = DMA_PACKET(DMA_PACKET_WRITE, 0, 0, ndw);
-			ib->ptr[ib->length_dw++] = pe;
-			ib->ptr[ib->length_dw++] = upper_32_bits(pe) & 0xff;
-			for (; ndw > 0; ndw -= 2, --count, pe += 8) {
-				if (flags & R600_PTE_SYSTEM) {
-					value = radeon_vm_map_gart(rdev, addr);
-					value &= 0xFFFFFFFFFFFFF000ULL;
-				} else if (flags & R600_PTE_VALID) {
-					value = addr;
-				} else {
-					value = 0;
-				}
-				addr += incr;
-				value |= flags;
-				ib->ptr[ib->length_dw++] = value;
-				ib->ptr[ib->length_dw++] = upper_32_bits(value);
-			}
-		}
-	} else {
-		while (count) {
-			ndw = count * 2;
-			if (ndw > 0xFFFFE)
-				ndw = 0xFFFFE;
-
-			if (flags & R600_PTE_VALID)
+	while (count) {
+		ndw = count * 2;
+		if (ndw > 0xFFFFE)
+			ndw = 0xFFFFE;
+
+		/* for non-physically contiguous pages (system) */
+		ib->ptr[ib->length_dw++] = DMA_PACKET(DMA_PACKET_WRITE,
+						      0, 0, ndw);
+		ib->ptr[ib->length_dw++] = pe;
+		ib->ptr[ib->length_dw++] = upper_32_bits(pe) & 0xff;
+		for (; ndw > 0; ndw -= 2, --count, pe += 8) {
+			if (flags & R600_PTE_SYSTEM) {
+				value = radeon_vm_map_gart(rdev, addr);
+				value &= 0xFFFFFFFFFFFFF000ULL;
+			} else if (flags & R600_PTE_VALID) {
 				value = addr;
-			else
+			} else {
 				value = 0;
-			/* for physically contiguous pages (vram) */
-			ib->ptr[ib->length_dw++] = DMA_PTE_PDE_PACKET(ndw);
-			ib->ptr[ib->length_dw++] = pe; /* dst addr */
-			ib->ptr[ib->length_dw++] = upper_32_bits(pe) & 0xff;
-			ib->ptr[ib->length_dw++] = flags; /* mask */
-			ib->ptr[ib->length_dw++] = 0;
-			ib->ptr[ib->length_dw++] = value; /* value */
+			}
+			addr += incr;
+			value |= flags;
+			ib->ptr[ib->length_dw++] = value;
 			ib->ptr[ib->length_dw++] = upper_32_bits(value);
-			ib->ptr[ib->length_dw++] = incr; /* increment size */
-			ib->ptr[ib->length_dw++] = 0;
-			pe += ndw * 4;
-			addr += (ndw / 2) * incr;
-			count -= ndw / 2;
 		}
 	}
+}
+
+/**
+ * cayman_dma_vm_set_pages - update the page tables using the DMA
+ *
+ * @rdev: radeon_device pointer
+ * @ib: indirect buffer to fill with commands
+ * @pe: addr of the page entry
+ * @addr: dst addr to write into pe
+ * @count: number of page entries to update
+ * @incr: increase next addr by incr bytes
+ * @flags: hw access flags
+ *
+ * Update the page tables using the DMA (cayman/TN).
+ */
+void cayman_dma_vm_set_pages(struct radeon_device *rdev,
+			     struct radeon_ib *ib,
+			     uint64_t pe,
+			     uint64_t addr, unsigned count,
+			     uint32_t incr, uint32_t flags)
+{
+	uint64_t value;
+	unsigned ndw;
+
+	while (count) {
+		ndw = count * 2;
+		if (ndw > 0xFFFFE)
+			ndw = 0xFFFFE;
+
+		if (flags & R600_PTE_VALID)
+			value = addr;
+		else
+			value = 0;
+
+		/* for physically contiguous pages (vram) */
+		ib->ptr[ib->length_dw++] = DMA_PTE_PDE_PACKET(ndw);
+		ib->ptr[ib->length_dw++] = pe; /* dst addr */
+		ib->ptr[ib->length_dw++] = upper_32_bits(pe) & 0xff;
+		ib->ptr[ib->length_dw++] = flags; /* mask */
+		ib->ptr[ib->length_dw++] = 0;
+		ib->ptr[ib->length_dw++] = value; /* value */
+		ib->ptr[ib->length_dw++] = upper_32_bits(value);
+		ib->ptr[ib->length_dw++] = incr; /* increment size */
+		ib->ptr[ib->length_dw++] = 0;
+
+		pe += ndw * 4;
+		addr += (ndw / 2) * incr;
+		count -= ndw / 2;
+	}
+}
+
+/**
+ * cayman_dma_vm_pad_ib - pad the IB to the required number of dw
+ *
+ * @ib: indirect buffer to fill with padding
+ *
+ */
+void cayman_dma_vm_pad_ib(struct radeon_ib *ib)
+{
 	while (ib->length_dw & 0x7)
 		ib->ptr[ib->length_dw++] = DMA_PACKET(DMA_PACKET_NOP, 0, 0, 0);
 }
diff --git a/drivers/gpu/drm/radeon/r100.c b/drivers/gpu/drm/radeon/r100.c
index 1544efcf1c3a..4c5ec44ff328 100644
--- a/drivers/gpu/drm/radeon/r100.c
+++ b/drivers/gpu/drm/radeon/r100.c
@@ -652,7 +652,6 @@ int r100_pci_gart_enable(struct radeon_device *rdev)
 {
 	uint32_t tmp;
 
-	radeon_gart_restore(rdev);
 	/* discard memory request outside of configured range */
 	tmp = RREG32(RADEON_AIC_CNTL) | RADEON_DIS_OUT_OF_PCI_GART_ACCESS;
 	WREG32(RADEON_AIC_CNTL, tmp);
@@ -683,7 +682,7 @@ void r100_pci_gart_disable(struct radeon_device *rdev)
 }
 
 void r100_pci_gart_set_page(struct radeon_device *rdev, unsigned i,
-			    uint64_t addr)
+			    uint64_t addr, uint32_t flags)
 {
 	u32 *gtt = rdev->gart.ptr;
 	gtt[i] = cpu_to_le32(lower_32_bits(addr));
@@ -838,11 +837,7 @@ void r100_fence_ring_emit(struct radeon_device *rdev,
 	/* Wait until IDLE & CLEAN */
 	radeon_ring_write(ring, PACKET0(RADEON_WAIT_UNTIL, 0));
 	radeon_ring_write(ring, RADEON_WAIT_2D_IDLECLEAN | RADEON_WAIT_3D_IDLECLEAN);
-	radeon_ring_write(ring, PACKET0(RADEON_HOST_PATH_CNTL, 0));
-	radeon_ring_write(ring, rdev->config.r100.hdp_cntl |
-				RADEON_HDP_READ_BUFFER_INVALIDATE);
-	radeon_ring_write(ring, PACKET0(RADEON_HOST_PATH_CNTL, 0));
-	radeon_ring_write(ring, rdev->config.r100.hdp_cntl);
+	r100_ring_hdp_flush(rdev, ring);
 	/* Emit fence sequence & fire IRQ */
 	radeon_ring_write(ring, PACKET0(rdev->fence_drv[fence->ring].scratch_reg, 0));
 	radeon_ring_write(ring, fence->seq);
@@ -930,7 +925,7 @@ int r100_copy_blit(struct radeon_device *rdev,
 	if (fence) {
 		r = radeon_fence_emit(rdev, fence, RADEON_RING_TYPE_GFX_INDEX);
 	}
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev, ring, false);
 	return r;
 }
 
@@ -963,7 +958,7 @@ void r100_ring_start(struct radeon_device *rdev, struct radeon_ring *ring)
 			  RADEON_ISYNC_ANY3D_IDLE2D |
 			  RADEON_ISYNC_WAIT_IDLEGUI |
 			  RADEON_ISYNC_CPSCRATCH_IDLEGUI);
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev, ring, false);
 }
 
 
@@ -1061,6 +1056,20 @@ void r100_gfx_set_wptr(struct radeon_device *rdev,
 	(void)RREG32(RADEON_CP_RB_WPTR);
 }
 
+/**
+ * r100_ring_hdp_flush - flush Host Data Path via the ring buffer
+ * rdev: radeon device structure
+ * ring: ring buffer struct for emitting packets
+ */
+void r100_ring_hdp_flush(struct radeon_device *rdev, struct radeon_ring *ring)
+{
+	radeon_ring_write(ring, PACKET0(RADEON_HOST_PATH_CNTL, 0));
+	radeon_ring_write(ring, rdev->config.r100.hdp_cntl |
+				RADEON_HDP_READ_BUFFER_INVALIDATE);
+	radeon_ring_write(ring, PACKET0(RADEON_HOST_PATH_CNTL, 0));
+	radeon_ring_write(ring, rdev->config.r100.hdp_cntl);
+}
+
 static void r100_cp_load_microcode(struct radeon_device *rdev)
 {
 	const __be32 *fw_data;
@@ -1401,7 +1410,6 @@ int r100_cs_parse_packet0(struct radeon_cs_parser *p,
  */
 int r100_cs_packet_parse_vline(struct radeon_cs_parser *p)
 {
-	struct drm_mode_object *obj;
 	struct drm_crtc *crtc;
 	struct radeon_crtc *radeon_crtc;
 	struct radeon_cs_packet p3reloc, waitreloc;
@@ -1441,12 +1449,11 @@ int r100_cs_packet_parse_vline(struct radeon_cs_parser *p)
 	header = radeon_get_ib_value(p, h_idx);
 	crtc_id = radeon_get_ib_value(p, h_idx + 5);
 	reg = R100_CP_PACKET0_GET_REG(header);
-	obj = drm_mode_object_find(p->rdev->ddev, crtc_id, DRM_MODE_OBJECT_CRTC);
-	if (!obj) {
+	crtc = drm_crtc_find(p->rdev->ddev, crtc_id);
+	if (!crtc) {
 		DRM_ERROR("cannot find crtc %d\n", crtc_id);
 		return -ENOENT;
 	}
-	crtc = obj_to_crtc(obj);
 	radeon_crtc = to_radeon_crtc(crtc);
 	crtc_id = radeon_crtc->crtc_id;
 
@@ -3631,7 +3638,7 @@ int r100_ring_test(struct radeon_device *rdev, struct radeon_ring *ring)
 	}
 	radeon_ring_write(ring, PACKET0(scratch, 0));
 	radeon_ring_write(ring, 0xDEADBEEF);
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev, ring, false);
 	for (i = 0; i < rdev->usec_timeout; i++) {
 		tmp = RREG32(scratch);
 		if (tmp == 0xDEADBEEF) {
@@ -3693,7 +3700,7 @@ int r100_ib_test(struct radeon_device *rdev, struct radeon_ring *ring)
 	ib.ptr[6] = PACKET2(0);
 	ib.ptr[7] = PACKET2(0);
 	ib.length_dw = 8;
-	r = radeon_ib_schedule(rdev, &ib, NULL);
+	r = radeon_ib_schedule(rdev, &ib, NULL, false);
 	if (r) {
 		DRM_ERROR("radeon: failed to schedule ib (%d).\n", r);
 		goto free_ib;
@@ -4067,39 +4074,6 @@ int r100_init(struct radeon_device *rdev)
 	return 0;
 }
 
-uint32_t r100_mm_rreg(struct radeon_device *rdev, uint32_t reg,
-		      bool always_indirect)
-{
-	if (reg < rdev->rmmio_size && !always_indirect)
-		return readl(((void __iomem *)rdev->rmmio) + reg);
-	else {
-		unsigned long flags;
-		uint32_t ret;
-
-		spin_lock_irqsave(&rdev->mmio_idx_lock, flags);
-		writel(reg, ((void __iomem *)rdev->rmmio) + RADEON_MM_INDEX);
-		ret = readl(((void __iomem *)rdev->rmmio) + RADEON_MM_DATA);
-		spin_unlock_irqrestore(&rdev->mmio_idx_lock, flags);
-
-		return ret;
-	}
-}
-
-void r100_mm_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v,
-		  bool always_indirect)
-{
-	if (reg < rdev->rmmio_size && !always_indirect)
-		writel(v, ((void __iomem *)rdev->rmmio) + reg);
-	else {
-		unsigned long flags;
-
-		spin_lock_irqsave(&rdev->mmio_idx_lock, flags);
-		writel(reg, ((void __iomem *)rdev->rmmio) + RADEON_MM_INDEX);
-		writel(v, ((void __iomem *)rdev->rmmio) + RADEON_MM_DATA);
-		spin_unlock_irqrestore(&rdev->mmio_idx_lock, flags);
-	}
-}
-
 u32 r100_io_rreg(struct radeon_device *rdev, u32 reg)
 {
 	if (reg < rdev->rio_mem_size)
diff --git a/drivers/gpu/drm/radeon/r200.c b/drivers/gpu/drm/radeon/r200.c
index 58f0473aa73f..67780374a652 100644
--- a/drivers/gpu/drm/radeon/r200.c
+++ b/drivers/gpu/drm/radeon/r200.c
@@ -121,7 +121,7 @@ int r200_copy_dma(struct radeon_device *rdev,
 	if (fence) {
 		r = radeon_fence_emit(rdev, fence, RADEON_RING_TYPE_GFX_INDEX);
 	}
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev, ring, false);
 	return r;
 }
 
diff --git a/drivers/gpu/drm/radeon/r300.c b/drivers/gpu/drm/radeon/r300.c
index 3c21d77a483d..1bc4704034ce 100644
--- a/drivers/gpu/drm/radeon/r300.c
+++ b/drivers/gpu/drm/radeon/r300.c
@@ -69,17 +69,23 @@ void rv370_pcie_gart_tlb_flush(struct radeon_device *rdev)
 	mb();
 }
 
+#define R300_PTE_UNSNOOPED (1 << 0)
 #define R300_PTE_WRITEABLE (1 << 2)
 #define R300_PTE_READABLE  (1 << 3)
 
 void rv370_pcie_gart_set_page(struct radeon_device *rdev, unsigned i,
-			      uint64_t addr)
+			      uint64_t addr, uint32_t flags)
 {
 	void __iomem *ptr = rdev->gart.ptr;
 
 	addr = (lower_32_bits(addr) >> 8) |
-	       ((upper_32_bits(addr) & 0xff) << 24) |
-	       R300_PTE_WRITEABLE | R300_PTE_READABLE;
+		((upper_32_bits(addr) & 0xff) << 24);
+	if (flags & RADEON_GART_PAGE_READ)
+		addr |= R300_PTE_READABLE;
+	if (flags & RADEON_GART_PAGE_WRITE)
+		addr |= R300_PTE_WRITEABLE;
+	if (!(flags & RADEON_GART_PAGE_SNOOP))
+		addr |= R300_PTE_UNSNOOPED;
 	/* on x86 we want this to be CPU endian, on powerpc
 	 * on powerpc without HW swappers, it'll get swapped on way
 	 * into VRAM - so no need for cpu_to_le32 on VRAM tables */
@@ -120,7 +126,6 @@ int rv370_pcie_gart_enable(struct radeon_device *rdev)
 	r = radeon_gart_table_vram_pin(rdev);
 	if (r)
 		return r;
-	radeon_gart_restore(rdev);
 	/* discard memory request outside of configured range */
 	tmp = RADEON_PCIE_TX_GART_UNMAPPED_ACCESS_DISCARD;
 	WREG32_PCIE(RADEON_PCIE_TX_GART_CNTL, tmp);
@@ -290,7 +295,7 @@ void r300_ring_start(struct radeon_device *rdev, struct radeon_ring *ring)
 	radeon_ring_write(ring,
 			  R300_GEOMETRY_ROUND_NEAREST |
 			  R300_COLOR_ROUND_NEAREST);
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev, ring, false);
 }
 
 static void r300_errata(struct radeon_device *rdev)
diff --git a/drivers/gpu/drm/radeon/r420.c b/drivers/gpu/drm/radeon/r420.c
index 802b19220a21..2828605aef3f 100644
--- a/drivers/gpu/drm/radeon/r420.c
+++ b/drivers/gpu/drm/radeon/r420.c
@@ -219,7 +219,7 @@ static void r420_cp_errata_init(struct radeon_device *rdev)
 	radeon_ring_write(ring, PACKET0(R300_CP_RESYNC_ADDR, 1));
 	radeon_ring_write(ring, rdev->config.r300.resync_scratch);
 	radeon_ring_write(ring, 0xDEADBEEF);
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev, ring, false);
 }
 
 static void r420_cp_errata_fini(struct radeon_device *rdev)
@@ -232,7 +232,7 @@ static void r420_cp_errata_fini(struct radeon_device *rdev)
 	radeon_ring_lock(rdev, ring, 8);
 	radeon_ring_write(ring, PACKET0(R300_RB3D_DSTCACHE_CTLSTAT, 0));
 	radeon_ring_write(ring, R300_RB3D_DC_FINISH);
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev, ring, false);
 	radeon_scratch_free(rdev, rdev->config.r300.resync_scratch);
 }
 
diff --git a/drivers/gpu/drm/radeon/r600.c b/drivers/gpu/drm/radeon/r600.c
index c66952d4b00c..e616eb5f6e7a 100644
--- a/drivers/gpu/drm/radeon/r600.c
+++ b/drivers/gpu/drm/radeon/r600.c
@@ -968,7 +968,6 @@ static int r600_pcie_gart_enable(struct radeon_device *rdev)
 	r = radeon_gart_table_vram_pin(rdev);
 	if (r)
 		return r;
-	radeon_gart_restore(rdev);
 
 	/* Setup L2 cache */
 	WREG32(VM_L2_CNTL, ENABLE_L2_CACHE | ENABLE_L2_FRAGMENT_PROCESSING |
@@ -1339,7 +1338,7 @@ int r600_vram_scratch_init(struct radeon_device *rdev)
 	if (rdev->vram_scratch.robj == NULL) {
 		r = radeon_bo_create(rdev, RADEON_GPU_PAGE_SIZE,
 				     PAGE_SIZE, true, RADEON_GEM_DOMAIN_VRAM,
-				     NULL, &rdev->vram_scratch.robj);
+				     0, NULL, &rdev->vram_scratch.robj);
 		if (r) {
 			return r;
 		}
@@ -1813,7 +1812,6 @@ static void r600_gpu_init(struct radeon_device *rdev)
 {
 	u32 tiling_config;
 	u32 ramcfg;
-	u32 cc_rb_backend_disable;
 	u32 cc_gc_shader_pipe_config;
 	u32 tmp;
 	int i, j;
@@ -1940,29 +1938,20 @@ static void r600_gpu_init(struct radeon_device *rdev)
 	}
 	tiling_config |= BANK_SWAPS(1);
 
-	cc_rb_backend_disable = RREG32(CC_RB_BACKEND_DISABLE) & 0x00ff0000;
-	tmp = R6XX_MAX_BACKENDS -
-		r600_count_pipe_bits((cc_rb_backend_disable >> 16) & R6XX_MAX_BACKENDS_MASK);
-	if (tmp < rdev->config.r600.max_backends) {
-		rdev->config.r600.max_backends = tmp;
-	}
-
 	cc_gc_shader_pipe_config = RREG32(CC_GC_SHADER_PIPE_CONFIG) & 0x00ffff00;
-	tmp = R6XX_MAX_PIPES -
-		r600_count_pipe_bits((cc_gc_shader_pipe_config >> 8) & R6XX_MAX_PIPES_MASK);
-	if (tmp < rdev->config.r600.max_pipes) {
-		rdev->config.r600.max_pipes = tmp;
-	}
-	tmp = R6XX_MAX_SIMDS -
-		r600_count_pipe_bits((cc_gc_shader_pipe_config >> 16) & R6XX_MAX_SIMDS_MASK);
-	if (tmp < rdev->config.r600.max_simds) {
-		rdev->config.r600.max_simds = tmp;
-	}
 	tmp = rdev->config.r600.max_simds -
 		r600_count_pipe_bits((cc_gc_shader_pipe_config >> 16) & R6XX_MAX_SIMDS_MASK);
 	rdev->config.r600.active_simds = tmp;
 
 	disabled_rb_mask = (RREG32(CC_RB_BACKEND_DISABLE) >> 16) & R6XX_MAX_BACKENDS_MASK;
+	tmp = 0;
+	for (i = 0; i < rdev->config.r600.max_backends; i++)
+		tmp |= (1 << i);
+	/* if all the backends are disabled, fix it up here */
+	if ((disabled_rb_mask & tmp) == tmp) {
+		for (i = 0; i < rdev->config.r600.max_backends; i++)
+			disabled_rb_mask &= ~(1 << i);
+	}
 	tmp = (tiling_config & PIPE_TILING__MASK) >> PIPE_TILING__SHIFT;
 	tmp = r6xx_remap_render_backend(rdev, tmp, rdev->config.r600.max_backends,
 					R6XX_MAX_BACKENDS, disabled_rb_mask);
@@ -2548,7 +2537,7 @@ int r600_cp_start(struct radeon_device *rdev)
 	radeon_ring_write(ring, PACKET3_ME_INITIALIZE_DEVICE_ID(1));
 	radeon_ring_write(ring, 0);
 	radeon_ring_write(ring, 0);
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev, ring, false);
 
 	cp_me = 0xff;
 	WREG32(R_0086D8_CP_ME_CNTL, cp_me);
@@ -2684,7 +2673,7 @@ int r600_ring_test(struct radeon_device *rdev, struct radeon_ring *ring)
 	radeon_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 1));
 	radeon_ring_write(ring, ((scratch - PACKET3_SET_CONFIG_REG_OFFSET) >> 2));
 	radeon_ring_write(ring, 0xDEADBEEF);
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev, ring, false);
 	for (i = 0; i < rdev->usec_timeout; i++) {
 		tmp = RREG32(scratch);
 		if (tmp == 0xDEADBEEF)
@@ -2754,6 +2743,17 @@ void r600_fence_ring_emit(struct radeon_device *rdev,
 	}
 }
 
+/**
+ * r600_semaphore_ring_emit - emit a semaphore on the CP ring
+ *
+ * @rdev: radeon_device pointer
+ * @ring: radeon ring buffer object
+ * @semaphore: radeon semaphore object
+ * @emit_wait: Is this a sempahore wait?
+ *
+ * Emits a semaphore signal/wait packet to the CP ring and prevents the PFP
+ * from running ahead of semaphore waits.
+ */
 bool r600_semaphore_ring_emit(struct radeon_device *rdev,
 			      struct radeon_ring *ring,
 			      struct radeon_semaphore *semaphore,
@@ -2769,6 +2769,13 @@ bool r600_semaphore_ring_emit(struct radeon_device *rdev,
 	radeon_ring_write(ring, lower_32_bits(addr));
 	radeon_ring_write(ring, (upper_32_bits(addr) & 0xff) | sel);
 
+	/* PFP_SYNC_ME packet only exists on 7xx+ */
+	if (emit_wait && (rdev->family >= CHIP_RV770)) {
+		/* Prevent the PFP from running ahead of the semaphore wait */
+		radeon_ring_write(ring, PACKET3(PACKET3_PFP_SYNC_ME, 0));
+		radeon_ring_write(ring, 0x0);
+	}
+
 	return true;
 }
 
@@ -2846,7 +2853,7 @@ int r600_copy_cpdma(struct radeon_device *rdev,
 		return r;
 	}
 
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev, ring, false);
 	radeon_semaphore_free(rdev, &sem, *fence);
 
 	return r;
@@ -3166,7 +3173,7 @@ int r600_ib_test(struct radeon_device *rdev, struct radeon_ring *ring)
 	ib.ptr[1] = ((scratch - PACKET3_SET_CONFIG_REG_OFFSET) >> 2);
 	ib.ptr[2] = 0xDEADBEEF;
 	ib.length_dw = 3;
-	r = radeon_ib_schedule(rdev, &ib, NULL);
+	r = radeon_ib_schedule(rdev, &ib, NULL, false);
 	if (r) {
 		DRM_ERROR("radeon: failed to schedule ib (%d).\n", r);
 		goto free_ib;
@@ -3227,7 +3234,7 @@ int r600_ih_ring_alloc(struct radeon_device *rdev)
 	if (rdev->ih.ring_obj == NULL) {
 		r = radeon_bo_create(rdev, rdev->ih.ring_size,
 				     PAGE_SIZE, true,
-				     RADEON_GEM_DOMAIN_GTT,
+				     RADEON_GEM_DOMAIN_GTT, 0,
 				     NULL, &rdev->ih.ring_obj);
 		if (r) {
 			DRM_ERROR("radeon: failed to create ih ring buffer (%d).\n", r);
@@ -3795,6 +3802,7 @@ static u32 r600_get_ih_wptr(struct radeon_device *rdev)
 		tmp = RREG32(IH_RB_CNTL);
 		tmp |= IH_WPTR_OVERFLOW_CLEAR;
 		WREG32(IH_RB_CNTL, tmp);
+		wptr &= ~RB_OVERFLOW;
 	}
 	return (wptr & rdev->ih.ptr_mask);
 }
@@ -3923,11 +3931,13 @@ restart_ih:
 			break;
 		case 9: /* D1 pflip */
 			DRM_DEBUG("IH: D1 flip\n");
-			radeon_crtc_handle_flip(rdev, 0);
+			if (radeon_use_pflipirq > 0)
+				radeon_crtc_handle_flip(rdev, 0);
 			break;
 		case 11: /* D2 pflip */
 			DRM_DEBUG("IH: D2 flip\n");
-			radeon_crtc_handle_flip(rdev, 1);
+			if (radeon_use_pflipirq > 0)
+				radeon_crtc_handle_flip(rdev, 1);
 			break;
 		case 19: /* HPD/DAC hotplug */
 			switch (src_data) {
@@ -4088,16 +4098,15 @@ int r600_debugfs_mc_info_init(struct radeon_device *rdev)
 }
 
 /**
- * r600_ioctl_wait_idle - flush host path cache on wait idle ioctl
+ * r600_mmio_hdp_flush - flush Host Data Path cache via MMIO
  * rdev: radeon device structure
- * bo: buffer object struct which userspace is waiting for idle
  *
- * Some R6XX/R7XX doesn't seems to take into account HDP flush performed
- * through ring buffer, this leads to corruption in rendering, see
- * http://bugzilla.kernel.org/show_bug.cgi?id=15186 to avoid this we
- * directly perform HDP flush by writing register through MMIO.
+ * Some R6XX/R7XX don't seem to take into account HDP flushes performed
+ * through the ring buffer. This leads to corruption in rendering, see
+ * http://bugzilla.kernel.org/show_bug.cgi?id=15186 . To avoid this, we
+ * directly perform the HDP flush by writing the register through MMIO.
  */
-void r600_ioctl_wait_idle(struct radeon_device *rdev, struct radeon_bo *bo)
+void r600_mmio_hdp_flush(struct radeon_device *rdev)
 {
 	/* r7xx hw bug.  write to HDP_DEBUG1 followed by fb read
 	 * rather than write to HDP_REG_COHERENCY_FLUSH_CNTL.
diff --git a/drivers/gpu/drm/radeon/r600_cs.c b/drivers/gpu/drm/radeon/r600_cs.c
index 12511bb5fd6f..c47537a1ddba 100644
--- a/drivers/gpu/drm/radeon/r600_cs.c
+++ b/drivers/gpu/drm/radeon/r600_cs.c
@@ -825,7 +825,6 @@ int r600_cs_common_vline_parse(struct radeon_cs_parser *p,
 			       uint32_t *vline_start_end,
 			       uint32_t *vline_status)
 {
-	struct drm_mode_object *obj;
 	struct drm_crtc *crtc;
 	struct radeon_crtc *radeon_crtc;
 	struct radeon_cs_packet p3reloc, wait_reg_mem;
@@ -887,12 +886,11 @@ int r600_cs_common_vline_parse(struct radeon_cs_parser *p,
 	crtc_id = radeon_get_ib_value(p, h_idx + 2 + 7 + 1);
 	reg = R600_CP_PACKET0_GET_REG(header);
 
-	obj = drm_mode_object_find(p->rdev->ddev, crtc_id, DRM_MODE_OBJECT_CRTC);
-	if (!obj) {
+	crtc = drm_crtc_find(p->rdev->ddev, crtc_id);
+	if (!crtc) {
 		DRM_ERROR("cannot find crtc %d\n", crtc_id);
 		return -ENOENT;
 	}
-	crtc = obj_to_crtc(obj);
 	radeon_crtc = to_radeon_crtc(crtc);
 	crtc_id = radeon_crtc->crtc_id;
 
diff --git a/drivers/gpu/drm/radeon/r600_dma.c b/drivers/gpu/drm/radeon/r600_dma.c
index 4969cef44a19..51fd98553eaf 100644
--- a/drivers/gpu/drm/radeon/r600_dma.c
+++ b/drivers/gpu/drm/radeon/r600_dma.c
@@ -261,7 +261,7 @@ int r600_dma_ring_test(struct radeon_device *rdev,
 	radeon_ring_write(ring, rdev->vram_scratch.gpu_addr & 0xfffffffc);
 	radeon_ring_write(ring, upper_32_bits(rdev->vram_scratch.gpu_addr) & 0xff);
 	radeon_ring_write(ring, 0xDEADBEEF);
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev, ring, false);
 
 	for (i = 0; i < rdev->usec_timeout; i++) {
 		tmp = readl(ptr);
@@ -368,7 +368,7 @@ int r600_dma_ib_test(struct radeon_device *rdev, struct radeon_ring *ring)
 	ib.ptr[3] = 0xDEADBEEF;
 	ib.length_dw = 4;
 
-	r = radeon_ib_schedule(rdev, &ib, NULL);
+	r = radeon_ib_schedule(rdev, &ib, NULL, false);
 	if (r) {
 		radeon_ib_free(rdev, &ib);
 		DRM_ERROR("radeon: failed to schedule ib (%d).\n", r);
@@ -493,7 +493,7 @@ int r600_copy_dma(struct radeon_device *rdev,
 		return r;
 	}
 
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev, ring, false);
 	radeon_semaphore_free(rdev, &sem, *fence);
 
 	return r;
diff --git a/drivers/gpu/drm/radeon/r600d.h b/drivers/gpu/drm/radeon/r600d.h
index f94e7a9afe75..0c4a7d8d93e0 100644
--- a/drivers/gpu/drm/radeon/r600d.h
+++ b/drivers/gpu/drm/radeon/r600d.h
@@ -1597,6 +1597,7 @@
 		 */
 #              define PACKET3_CP_DMA_CMD_SAIC      (1 << 28)
 #              define PACKET3_CP_DMA_CMD_DAIC      (1 << 29)
+#define	PACKET3_PFP_SYNC_ME				0x42 /* r7xx+ only */
 #define	PACKET3_SURFACE_SYNC				0x43
 #              define PACKET3_CB0_DEST_BASE_ENA    (1 << 6)
 #              define PACKET3_FULL_CACHE_ENA       (1 << 20) /* r7xx+ only */
diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h
index b7204500a9a6..5f05b4c84338 100644
--- a/drivers/gpu/drm/radeon/radeon.h
+++ b/drivers/gpu/drm/radeon/radeon.h
@@ -64,6 +64,7 @@
 #include <linux/wait.h>
 #include <linux/list.h>
 #include <linux/kref.h>
+#include <linux/interval_tree.h>
 
 #include <ttm/ttm_bo_api.h>
 #include <ttm/ttm_bo_driver.h>
@@ -103,6 +104,8 @@ extern int radeon_hard_reset;
 extern int radeon_vm_size;
 extern int radeon_vm_block_size;
 extern int radeon_deep_color;
+extern int radeon_use_pflipirq;
+extern int radeon_bapm;
 
 /*
  * Copy from radeon_drv.h so we don't have to include both and have conflicting
@@ -304,6 +307,9 @@ int radeon_atom_get_leakage_vddc_based_on_leakage_params(struct radeon_device *r
 							 u16 *vddc, u16 *vddci,
 							 u16 virtual_voltage_id,
 							 u16 vbios_voltage_id);
+int radeon_atom_get_voltage_evv(struct radeon_device *rdev,
+				u16 virtual_voltage_id,
+				u16 *voltage);
 int radeon_atom_round_to_true_voltage(struct radeon_device *rdev,
 				      u8 voltage_type,
 				      u16 nominal_voltage,
@@ -317,6 +323,9 @@ int radeon_atom_get_voltage_table(struct radeon_device *rdev,
 				  struct atom_voltage_table *voltage_table);
 bool radeon_atom_is_voltage_gpio(struct radeon_device *rdev,
 				 u8 voltage_type, u8 voltage_mode);
+int radeon_atom_get_svi2_info(struct radeon_device *rdev,
+			      u8 voltage_type,
+			      u8 *svd_gpio_id, u8 *svc_gpio_id);
 void radeon_atom_update_memory_dll(struct radeon_device *rdev,
 				   u32 mem_clock);
 void radeon_atom_set_ac_timing(struct radeon_device *rdev,
@@ -441,14 +450,13 @@ struct radeon_mman {
 struct radeon_bo_va {
 	/* protected by bo being reserved */
 	struct list_head		bo_list;
-	uint64_t			soffset;
-	uint64_t			eoffset;
 	uint32_t			flags;
-	bool				valid;
+	uint64_t			addr;
 	unsigned			ref_count;
 
 	/* protected by vm mutex */
-	struct list_head		vm_list;
+	struct interval_tree_node	it;
+	struct list_head		vm_status;
 
 	/* constant after initialization */
 	struct radeon_vm		*vm;
@@ -464,6 +472,7 @@ struct radeon_bo {
 	struct ttm_placement		placement;
 	struct ttm_buffer_object	tbo;
 	struct ttm_bo_kmap_obj		kmap;
+	u32				flags;
 	unsigned			pin_count;
 	void				*kptr;
 	u32				tiling_flags;
@@ -542,9 +551,9 @@ struct radeon_gem {
 
 int radeon_gem_init(struct radeon_device *rdev);
 void radeon_gem_fini(struct radeon_device *rdev);
-int radeon_gem_object_create(struct radeon_device *rdev, int size,
+int radeon_gem_object_create(struct radeon_device *rdev, unsigned long size,
 				int alignment, int initial_domain,
-				bool discardable, bool kernel,
+				u32 flags, bool kernel,
 				struct drm_gem_object **obj);
 
 int radeon_mode_dumb_create(struct drm_file *file_priv,
@@ -589,6 +598,12 @@ struct radeon_mc;
 #define RADEON_GPU_PAGE_SHIFT 12
 #define RADEON_GPU_PAGE_ALIGN(a) (((a) + RADEON_GPU_PAGE_MASK) & ~RADEON_GPU_PAGE_MASK)
 
+#define RADEON_GART_PAGE_DUMMY  0
+#define RADEON_GART_PAGE_VALID	(1 << 0)
+#define RADEON_GART_PAGE_READ	(1 << 1)
+#define RADEON_GART_PAGE_WRITE	(1 << 2)
+#define RADEON_GART_PAGE_SNOOP	(1 << 3)
+
 struct radeon_gart {
 	dma_addr_t			table_addr;
 	struct radeon_bo		*robj;
@@ -613,8 +628,7 @@ void radeon_gart_unbind(struct radeon_device *rdev, unsigned offset,
 			int pages);
 int radeon_gart_bind(struct radeon_device *rdev, unsigned offset,
 		     int pages, struct page **pagelist,
-		     dma_addr_t *dma_addr);
-void radeon_gart_restore(struct radeon_device *rdev);
+		     dma_addr_t *dma_addr, uint32_t flags);
 
 
 /*
@@ -854,9 +868,9 @@ struct radeon_mec {
 #define R600_PTE_FRAG_64KB	(4 << 7)
 #define R600_PTE_FRAG_256KB	(6 << 7)
 
-/* flags used for GART page table entries on R600+ */
-#define R600_PTE_GART	( R600_PTE_VALID | R600_PTE_SYSTEM | R600_PTE_SNOOPED \
-			| R600_PTE_READABLE | R600_PTE_WRITEABLE)
+/* flags needed to be set so we can copy directly from the GART table */
+#define R600_PTE_GART_MASK	( R600_PTE_READABLE | R600_PTE_WRITEABLE | \
+				  R600_PTE_SYSTEM | R600_PTE_VALID )
 
 struct radeon_vm_pt {
 	struct radeon_bo		*bo;
@@ -864,9 +878,15 @@ struct radeon_vm_pt {
 };
 
 struct radeon_vm {
-	struct list_head		va;
+	struct rb_root			va;
 	unsigned			id;
 
+	/* BOs moved, but not yet updated in the PT */
+	struct list_head		invalidated;
+
+	/* BOs freed, but not yet updated in the PT */
+	struct list_head		freed;
+
 	/* contains the page directory */
 	struct radeon_bo		*page_directory;
 	uint64_t			pd_gpu_addr;
@@ -875,6 +895,8 @@ struct radeon_vm {
 	/* array of page tables, one for each page directory entry */
 	struct radeon_vm_pt		*page_tables;
 
+	struct radeon_bo_va		*ib_bo_va;
+
 	struct mutex			mutex;
 	/* last fence for cs using this vm */
 	struct radeon_fence		*fence;
@@ -893,6 +915,8 @@ struct radeon_vm_manager {
 	u64				vram_base_offset;
 	/* is vm enabled? */
 	bool				enabled;
+	/* for hw to save the PD addr on suspend/resume */
+	uint32_t			saved_table_addr[RADEON_NUM_VM];
 };
 
 /*
@@ -946,7 +970,7 @@ int radeon_ib_get(struct radeon_device *rdev, int ring,
 		  unsigned size);
 void radeon_ib_free(struct radeon_device *rdev, struct radeon_ib *ib);
 int radeon_ib_schedule(struct radeon_device *rdev, struct radeon_ib *ib,
-		       struct radeon_ib *const_ib);
+		       struct radeon_ib *const_ib, bool hdp_flush);
 int radeon_ib_pool_init(struct radeon_device *rdev);
 void radeon_ib_pool_fini(struct radeon_device *rdev);
 int radeon_ib_ring_tests(struct radeon_device *rdev);
@@ -956,8 +980,10 @@ bool radeon_ring_supports_scratch_reg(struct radeon_device *rdev,
 void radeon_ring_free_size(struct radeon_device *rdev, struct radeon_ring *cp);
 int radeon_ring_alloc(struct radeon_device *rdev, struct radeon_ring *cp, unsigned ndw);
 int radeon_ring_lock(struct radeon_device *rdev, struct radeon_ring *cp, unsigned ndw);
-void radeon_ring_commit(struct radeon_device *rdev, struct radeon_ring *cp);
-void radeon_ring_unlock_commit(struct radeon_device *rdev, struct radeon_ring *cp);
+void radeon_ring_commit(struct radeon_device *rdev, struct radeon_ring *cp,
+			bool hdp_flush);
+void radeon_ring_unlock_commit(struct radeon_device *rdev, struct radeon_ring *cp,
+			       bool hdp_flush);
 void radeon_ring_undo(struct radeon_ring *ring);
 void radeon_ring_unlock_undo(struct radeon_device *rdev, struct radeon_ring *cp);
 int radeon_ring_test(struct radeon_device *rdev, struct radeon_ring *cp);
@@ -1734,6 +1760,7 @@ struct radeon_asic_ring {
 	/* command emmit functions */
 	void (*ib_execute)(struct radeon_device *rdev, struct radeon_ib *ib);
 	void (*emit_fence)(struct radeon_device *rdev, struct radeon_fence *fence);
+	void (*hdp_flush)(struct radeon_device *rdev, struct radeon_ring *ring);
 	bool (*emit_semaphore)(struct radeon_device *rdev, struct radeon_ring *cp,
 			       struct radeon_semaphore *semaphore, bool emit_wait);
 	void (*vm_flush)(struct radeon_device *rdev, int ridx, struct radeon_vm *vm);
@@ -1757,13 +1784,8 @@ struct radeon_asic {
 	int (*suspend)(struct radeon_device *rdev);
 	void (*vga_set_state)(struct radeon_device *rdev, bool state);
 	int (*asic_reset)(struct radeon_device *rdev);
-	/* ioctl hw specific callback. Some hw might want to perform special
-	 * operation on specific ioctl. For instance on wait idle some hw
-	 * might want to perform and HDP flush through MMIO as it seems that
-	 * some R6XX/R7XX hw doesn't take HDP flush into account if programmed
-	 * through ring.
-	 */
-	void (*ioctl_wait_idle)(struct radeon_device *rdev, struct radeon_bo *bo);
+	/* Flush the HDP cache via MMIO */
+	void (*mmio_hdp_flush)(struct radeon_device *rdev);
 	/* check if 3D engine is idle */
 	bool (*gui_idle)(struct radeon_device *rdev);
 	/* wait for mc_idle */
@@ -1776,16 +1798,26 @@ struct radeon_asic {
 	struct {
 		void (*tlb_flush)(struct radeon_device *rdev);
 		void (*set_page)(struct radeon_device *rdev, unsigned i,
-				 uint64_t addr);
+				 uint64_t addr, uint32_t flags);
 	} gart;
 	struct {
 		int (*init)(struct radeon_device *rdev);
 		void (*fini)(struct radeon_device *rdev);
-		void (*set_page)(struct radeon_device *rdev,
-				 struct radeon_ib *ib,
-				 uint64_t pe,
-				 uint64_t addr, unsigned count,
-				 uint32_t incr, uint32_t flags);
+		void (*copy_pages)(struct radeon_device *rdev,
+				   struct radeon_ib *ib,
+				   uint64_t pe, uint64_t src,
+				   unsigned count);
+		void (*write_pages)(struct radeon_device *rdev,
+				    struct radeon_ib *ib,
+				    uint64_t pe,
+				    uint64_t addr, unsigned count,
+				    uint32_t incr, uint32_t flags);
+		void (*set_pages)(struct radeon_device *rdev,
+				  struct radeon_ib *ib,
+				  uint64_t pe,
+				  uint64_t addr, unsigned count,
+				  uint32_t incr, uint32_t flags);
+		void (*pad_ib)(struct radeon_ib *ib);
 	} vm;
 	/* ring specific callbacks */
 	struct radeon_asic_ring *ring[RADEON_NUM_RINGS];
@@ -2293,10 +2325,12 @@ struct radeon_device {
 	const struct firmware *mc_fw;	/* NI MC firmware */
 	const struct firmware *ce_fw;	/* SI CE firmware */
 	const struct firmware *mec_fw;	/* CIK MEC firmware */
+	const struct firmware *mec2_fw;	/* KV MEC2 firmware */
 	const struct firmware *sdma_fw;	/* CIK SDMA firmware */
 	const struct firmware *smc_fw;	/* SMC firmware */
 	const struct firmware *uvd_fw;	/* UVD firmware */
 	const struct firmware *vce_fw;	/* VCE firmware */
+	bool new_fw;
 	struct r600_vram_scratch vram_scratch;
 	int msi_enabled; /* msi enabled */
 	struct r600_ih ih; /* r6/700 interrupt ring */
@@ -2336,6 +2370,11 @@ struct radeon_device {
 
 	struct dev_pm_domain vga_pm_domain;
 	bool have_disp_power_ref;
+	u32 px_quirk_flags;
+
+	/* tracking pinned memory */
+	u64 vram_pin_size;
+	u64 gart_pin_size;
 };
 
 bool radeon_is_px(struct drm_device *dev);
@@ -2346,10 +2385,42 @@ int radeon_device_init(struct radeon_device *rdev,
 void radeon_device_fini(struct radeon_device *rdev);
 int radeon_gpu_wait_for_idle(struct radeon_device *rdev);
 
-uint32_t r100_mm_rreg(struct radeon_device *rdev, uint32_t reg,
-		      bool always_indirect);
-void r100_mm_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v,
-		  bool always_indirect);
+#define RADEON_MIN_MMIO_SIZE 0x10000
+
+static inline uint32_t r100_mm_rreg(struct radeon_device *rdev, uint32_t reg,
+				    bool always_indirect)
+{
+	/* The mmio size is 64kb at minimum. Allows the if to be optimized out. */
+	if ((reg < rdev->rmmio_size || reg < RADEON_MIN_MMIO_SIZE) && !always_indirect)
+		return readl(((void __iomem *)rdev->rmmio) + reg);
+	else {
+		unsigned long flags;
+		uint32_t ret;
+
+		spin_lock_irqsave(&rdev->mmio_idx_lock, flags);
+		writel(reg, ((void __iomem *)rdev->rmmio) + RADEON_MM_INDEX);
+		ret = readl(((void __iomem *)rdev->rmmio) + RADEON_MM_DATA);
+		spin_unlock_irqrestore(&rdev->mmio_idx_lock, flags);
+
+		return ret;
+	}
+}
+
+static inline void r100_mm_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v,
+				bool always_indirect)
+{
+	if ((reg < rdev->rmmio_size || reg < RADEON_MIN_MMIO_SIZE) && !always_indirect)
+		writel(v, ((void __iomem *)rdev->rmmio) + reg);
+	else {
+		unsigned long flags;
+
+		spin_lock_irqsave(&rdev->mmio_idx_lock, flags);
+		writel(reg, ((void __iomem *)rdev->rmmio) + RADEON_MM_INDEX);
+		writel(v, ((void __iomem *)rdev->rmmio) + RADEON_MM_DATA);
+		spin_unlock_irqrestore(&rdev->mmio_idx_lock, flags);
+	}
+}
+
 u32 r100_io_rreg(struct radeon_device *rdev, u32 reg);
 void r100_io_wreg(struct radeon_device *rdev, u32 reg, u32 v);
 
@@ -2703,10 +2774,13 @@ void radeon_ring_write(struct radeon_ring *ring, uint32_t v);
 #define radeon_vga_set_state(rdev, state) (rdev)->asic->vga_set_state((rdev), (state))
 #define radeon_asic_reset(rdev) (rdev)->asic->asic_reset((rdev))
 #define radeon_gart_tlb_flush(rdev) (rdev)->asic->gart.tlb_flush((rdev))
-#define radeon_gart_set_page(rdev, i, p) (rdev)->asic->gart.set_page((rdev), (i), (p))
+#define radeon_gart_set_page(rdev, i, p, f) (rdev)->asic->gart.set_page((rdev), (i), (p), (f))
 #define radeon_asic_vm_init(rdev) (rdev)->asic->vm.init((rdev))
 #define radeon_asic_vm_fini(rdev) (rdev)->asic->vm.fini((rdev))
-#define radeon_asic_vm_set_page(rdev, ib, pe, addr, count, incr, flags) ((rdev)->asic->vm.set_page((rdev), (ib), (pe), (addr), (count), (incr), (flags)))
+#define radeon_asic_vm_copy_pages(rdev, ib, pe, src, count) ((rdev)->asic->vm.copy_pages((rdev), (ib), (pe), (src), (count)))
+#define radeon_asic_vm_write_pages(rdev, ib, pe, addr, count, incr, flags) ((rdev)->asic->vm.write_pages((rdev), (ib), (pe), (addr), (count), (incr), (flags)))
+#define radeon_asic_vm_set_pages(rdev, ib, pe, addr, count, incr, flags) ((rdev)->asic->vm.set_pages((rdev), (ib), (pe), (addr), (count), (incr), (flags)))
+#define radeon_asic_vm_pad_ib(rdev, ib) ((rdev)->asic->vm.pad_ib((ib)))
 #define radeon_ring_start(rdev, r, cp) (rdev)->asic->ring[(r)]->ring_start((rdev), (cp))
 #define radeon_ring_test(rdev, r, cp) (rdev)->asic->ring[(r)]->ring_test((rdev), (cp))
 #define radeon_ib_test(rdev, r, cp) (rdev)->asic->ring[(r)]->ib_test((rdev), (cp))
@@ -2832,9 +2906,12 @@ void radeon_vm_fence(struct radeon_device *rdev,
 uint64_t radeon_vm_map_gart(struct radeon_device *rdev, uint64_t addr);
 int radeon_vm_update_page_directory(struct radeon_device *rdev,
 				    struct radeon_vm *vm);
+int radeon_vm_clear_freed(struct radeon_device *rdev,
+			  struct radeon_vm *vm);
+int radeon_vm_clear_invalids(struct radeon_device *rdev,
+			     struct radeon_vm *vm);
 int radeon_vm_bo_update(struct radeon_device *rdev,
-			struct radeon_vm *vm,
-			struct radeon_bo *bo,
+			struct radeon_bo_va *bo_va,
 			struct ttm_mem_reg *mem);
 void radeon_vm_bo_invalidate(struct radeon_device *rdev,
 			     struct radeon_bo *bo);
@@ -2847,8 +2924,8 @@ int radeon_vm_bo_set_addr(struct radeon_device *rdev,
 			  struct radeon_bo_va *bo_va,
 			  uint64_t offset,
 			  uint32_t flags);
-int radeon_vm_bo_rmv(struct radeon_device *rdev,
-		     struct radeon_bo_va *bo_va);
+void radeon_vm_bo_rmv(struct radeon_device *rdev,
+		      struct radeon_bo_va *bo_va);
 
 /* audio */
 void r600_audio_update_hdmi(struct work_struct *work);
diff --git a/drivers/gpu/drm/radeon/radeon_asic.c b/drivers/gpu/drm/radeon/radeon_asic.c
index 34b9aa9e3c06..eeeeabe09758 100644
--- a/drivers/gpu/drm/radeon/radeon_asic.c
+++ b/drivers/gpu/drm/radeon/radeon_asic.c
@@ -185,6 +185,7 @@ static struct radeon_asic_ring r100_gfx_ring = {
 	.get_rptr = &r100_gfx_get_rptr,
 	.get_wptr = &r100_gfx_get_wptr,
 	.set_wptr = &r100_gfx_set_wptr,
+	.hdp_flush = &r100_ring_hdp_flush,
 };
 
 static struct radeon_asic r100_asic = {
@@ -194,7 +195,7 @@ static struct radeon_asic r100_asic = {
 	.resume = &r100_resume,
 	.vga_set_state = &r100_vga_set_state,
 	.asic_reset = &r100_asic_reset,
-	.ioctl_wait_idle = NULL,
+	.mmio_hdp_flush = NULL,
 	.gui_idle = &r100_gui_idle,
 	.mc_wait_for_idle = &r100_mc_wait_for_idle,
 	.gart = {
@@ -260,7 +261,7 @@ static struct radeon_asic r200_asic = {
 	.resume = &r100_resume,
 	.vga_set_state = &r100_vga_set_state,
 	.asic_reset = &r100_asic_reset,
-	.ioctl_wait_idle = NULL,
+	.mmio_hdp_flush = NULL,
 	.gui_idle = &r100_gui_idle,
 	.mc_wait_for_idle = &r100_mc_wait_for_idle,
 	.gart = {
@@ -331,6 +332,7 @@ static struct radeon_asic_ring r300_gfx_ring = {
 	.get_rptr = &r100_gfx_get_rptr,
 	.get_wptr = &r100_gfx_get_wptr,
 	.set_wptr = &r100_gfx_set_wptr,
+	.hdp_flush = &r100_ring_hdp_flush,
 };
 
 static struct radeon_asic r300_asic = {
@@ -340,7 +342,7 @@ static struct radeon_asic r300_asic = {
 	.resume = &r300_resume,
 	.vga_set_state = &r100_vga_set_state,
 	.asic_reset = &r300_asic_reset,
-	.ioctl_wait_idle = NULL,
+	.mmio_hdp_flush = NULL,
 	.gui_idle = &r100_gui_idle,
 	.mc_wait_for_idle = &r300_mc_wait_for_idle,
 	.gart = {
@@ -406,7 +408,7 @@ static struct radeon_asic r300_asic_pcie = {
 	.resume = &r300_resume,
 	.vga_set_state = &r100_vga_set_state,
 	.asic_reset = &r300_asic_reset,
-	.ioctl_wait_idle = NULL,
+	.mmio_hdp_flush = NULL,
 	.gui_idle = &r100_gui_idle,
 	.mc_wait_for_idle = &r300_mc_wait_for_idle,
 	.gart = {
@@ -472,7 +474,7 @@ static struct radeon_asic r420_asic = {
 	.resume = &r420_resume,
 	.vga_set_state = &r100_vga_set_state,
 	.asic_reset = &r300_asic_reset,
-	.ioctl_wait_idle = NULL,
+	.mmio_hdp_flush = NULL,
 	.gui_idle = &r100_gui_idle,
 	.mc_wait_for_idle = &r300_mc_wait_for_idle,
 	.gart = {
@@ -538,7 +540,7 @@ static struct radeon_asic rs400_asic = {
 	.resume = &rs400_resume,
 	.vga_set_state = &r100_vga_set_state,
 	.asic_reset = &r300_asic_reset,
-	.ioctl_wait_idle = NULL,
+	.mmio_hdp_flush = NULL,
 	.gui_idle = &r100_gui_idle,
 	.mc_wait_for_idle = &rs400_mc_wait_for_idle,
 	.gart = {
@@ -604,7 +606,7 @@ static struct radeon_asic rs600_asic = {
 	.resume = &rs600_resume,
 	.vga_set_state = &r100_vga_set_state,
 	.asic_reset = &rs600_asic_reset,
-	.ioctl_wait_idle = NULL,
+	.mmio_hdp_flush = NULL,
 	.gui_idle = &r100_gui_idle,
 	.mc_wait_for_idle = &rs600_mc_wait_for_idle,
 	.gart = {
@@ -672,7 +674,7 @@ static struct radeon_asic rs690_asic = {
 	.resume = &rs690_resume,
 	.vga_set_state = &r100_vga_set_state,
 	.asic_reset = &rs600_asic_reset,
-	.ioctl_wait_idle = NULL,
+	.mmio_hdp_flush = NULL,
 	.gui_idle = &r100_gui_idle,
 	.mc_wait_for_idle = &rs690_mc_wait_for_idle,
 	.gart = {
@@ -740,7 +742,7 @@ static struct radeon_asic rv515_asic = {
 	.resume = &rv515_resume,
 	.vga_set_state = &r100_vga_set_state,
 	.asic_reset = &rs600_asic_reset,
-	.ioctl_wait_idle = NULL,
+	.mmio_hdp_flush = NULL,
 	.gui_idle = &r100_gui_idle,
 	.mc_wait_for_idle = &rv515_mc_wait_for_idle,
 	.gart = {
@@ -806,7 +808,7 @@ static struct radeon_asic r520_asic = {
 	.resume = &r520_resume,
 	.vga_set_state = &r100_vga_set_state,
 	.asic_reset = &rs600_asic_reset,
-	.ioctl_wait_idle = NULL,
+	.mmio_hdp_flush = NULL,
 	.gui_idle = &r100_gui_idle,
 	.mc_wait_for_idle = &r520_mc_wait_for_idle,
 	.gart = {
@@ -898,7 +900,7 @@ static struct radeon_asic r600_asic = {
 	.resume = &r600_resume,
 	.vga_set_state = &r600_vga_set_state,
 	.asic_reset = &r600_asic_reset,
-	.ioctl_wait_idle = r600_ioctl_wait_idle,
+	.mmio_hdp_flush = r600_mmio_hdp_flush,
 	.gui_idle = &r600_gui_idle,
 	.mc_wait_for_idle = &r600_mc_wait_for_idle,
 	.get_xclk = &r600_get_xclk,
@@ -970,7 +972,7 @@ static struct radeon_asic rv6xx_asic = {
 	.resume = &r600_resume,
 	.vga_set_state = &r600_vga_set_state,
 	.asic_reset = &r600_asic_reset,
-	.ioctl_wait_idle = r600_ioctl_wait_idle,
+	.mmio_hdp_flush = r600_mmio_hdp_flush,
 	.gui_idle = &r600_gui_idle,
 	.mc_wait_for_idle = &r600_mc_wait_for_idle,
 	.get_xclk = &r600_get_xclk,
@@ -1060,7 +1062,7 @@ static struct radeon_asic rs780_asic = {
 	.resume = &r600_resume,
 	.vga_set_state = &r600_vga_set_state,
 	.asic_reset = &r600_asic_reset,
-	.ioctl_wait_idle = r600_ioctl_wait_idle,
+	.mmio_hdp_flush = r600_mmio_hdp_flush,
 	.gui_idle = &r600_gui_idle,
 	.mc_wait_for_idle = &r600_mc_wait_for_idle,
 	.get_xclk = &r600_get_xclk,
@@ -1163,7 +1165,7 @@ static struct radeon_asic rv770_asic = {
 	.resume = &rv770_resume,
 	.asic_reset = &r600_asic_reset,
 	.vga_set_state = &r600_vga_set_state,
-	.ioctl_wait_idle = r600_ioctl_wait_idle,
+	.mmio_hdp_flush = r600_mmio_hdp_flush,
 	.gui_idle = &r600_gui_idle,
 	.mc_wait_for_idle = &r600_mc_wait_for_idle,
 	.get_xclk = &rv770_get_xclk,
@@ -1281,7 +1283,7 @@ static struct radeon_asic evergreen_asic = {
 	.resume = &evergreen_resume,
 	.asic_reset = &evergreen_asic_reset,
 	.vga_set_state = &r600_vga_set_state,
-	.ioctl_wait_idle = r600_ioctl_wait_idle,
+	.mmio_hdp_flush = r600_mmio_hdp_flush,
 	.gui_idle = &r600_gui_idle,
 	.mc_wait_for_idle = &evergreen_mc_wait_for_idle,
 	.get_xclk = &rv770_get_xclk,
@@ -1373,7 +1375,7 @@ static struct radeon_asic sumo_asic = {
 	.resume = &evergreen_resume,
 	.asic_reset = &evergreen_asic_reset,
 	.vga_set_state = &r600_vga_set_state,
-	.ioctl_wait_idle = r600_ioctl_wait_idle,
+	.mmio_hdp_flush = r600_mmio_hdp_flush,
 	.gui_idle = &r600_gui_idle,
 	.mc_wait_for_idle = &evergreen_mc_wait_for_idle,
 	.get_xclk = &r600_get_xclk,
@@ -1464,7 +1466,7 @@ static struct radeon_asic btc_asic = {
 	.resume = &evergreen_resume,
 	.asic_reset = &evergreen_asic_reset,
 	.vga_set_state = &r600_vga_set_state,
-	.ioctl_wait_idle = r600_ioctl_wait_idle,
+	.mmio_hdp_flush = r600_mmio_hdp_flush,
 	.gui_idle = &r600_gui_idle,
 	.mc_wait_for_idle = &evergreen_mc_wait_for_idle,
 	.get_xclk = &rv770_get_xclk,
@@ -1599,7 +1601,7 @@ static struct radeon_asic cayman_asic = {
 	.resume = &cayman_resume,
 	.asic_reset = &cayman_asic_reset,
 	.vga_set_state = &r600_vga_set_state,
-	.ioctl_wait_idle = r600_ioctl_wait_idle,
+	.mmio_hdp_flush = r600_mmio_hdp_flush,
 	.gui_idle = &r600_gui_idle,
 	.mc_wait_for_idle = &evergreen_mc_wait_for_idle,
 	.get_xclk = &rv770_get_xclk,
@@ -1611,7 +1613,10 @@ static struct radeon_asic cayman_asic = {
 	.vm = {
 		.init = &cayman_vm_init,
 		.fini = &cayman_vm_fini,
-		.set_page = &cayman_dma_vm_set_page,
+		.copy_pages = &cayman_dma_vm_copy_pages,
+		.write_pages = &cayman_dma_vm_write_pages,
+		.set_pages = &cayman_dma_vm_set_pages,
+		.pad_ib = &cayman_dma_vm_pad_ib,
 	},
 	.ring = {
 		[RADEON_RING_TYPE_GFX_INDEX] = &cayman_gfx_ring,
@@ -1699,7 +1704,7 @@ static struct radeon_asic trinity_asic = {
 	.resume = &cayman_resume,
 	.asic_reset = &cayman_asic_reset,
 	.vga_set_state = &r600_vga_set_state,
-	.ioctl_wait_idle = r600_ioctl_wait_idle,
+	.mmio_hdp_flush = r600_mmio_hdp_flush,
 	.gui_idle = &r600_gui_idle,
 	.mc_wait_for_idle = &evergreen_mc_wait_for_idle,
 	.get_xclk = &r600_get_xclk,
@@ -1711,7 +1716,10 @@ static struct radeon_asic trinity_asic = {
 	.vm = {
 		.init = &cayman_vm_init,
 		.fini = &cayman_vm_fini,
-		.set_page = &cayman_dma_vm_set_page,
+		.copy_pages = &cayman_dma_vm_copy_pages,
+		.write_pages = &cayman_dma_vm_write_pages,
+		.set_pages = &cayman_dma_vm_set_pages,
+		.pad_ib = &cayman_dma_vm_pad_ib,
 	},
 	.ring = {
 		[RADEON_RING_TYPE_GFX_INDEX] = &cayman_gfx_ring,
@@ -1829,7 +1837,7 @@ static struct radeon_asic si_asic = {
 	.resume = &si_resume,
 	.asic_reset = &si_asic_reset,
 	.vga_set_state = &r600_vga_set_state,
-	.ioctl_wait_idle = r600_ioctl_wait_idle,
+	.mmio_hdp_flush = r600_mmio_hdp_flush,
 	.gui_idle = &r600_gui_idle,
 	.mc_wait_for_idle = &evergreen_mc_wait_for_idle,
 	.get_xclk = &si_get_xclk,
@@ -1841,7 +1849,10 @@ static struct radeon_asic si_asic = {
 	.vm = {
 		.init = &si_vm_init,
 		.fini = &si_vm_fini,
-		.set_page = &si_dma_vm_set_page,
+		.copy_pages = &si_dma_vm_copy_pages,
+		.write_pages = &si_dma_vm_write_pages,
+		.set_pages = &si_dma_vm_set_pages,
+		.pad_ib = &cayman_dma_vm_pad_ib,
 	},
 	.ring = {
 		[RADEON_RING_TYPE_GFX_INDEX] = &si_gfx_ring,
@@ -1987,7 +1998,7 @@ static struct radeon_asic ci_asic = {
 	.resume = &cik_resume,
 	.asic_reset = &cik_asic_reset,
 	.vga_set_state = &r600_vga_set_state,
-	.ioctl_wait_idle = NULL,
+	.mmio_hdp_flush = &r600_mmio_hdp_flush,
 	.gui_idle = &r600_gui_idle,
 	.mc_wait_for_idle = &evergreen_mc_wait_for_idle,
 	.get_xclk = &cik_get_xclk,
@@ -1999,7 +2010,10 @@ static struct radeon_asic ci_asic = {
 	.vm = {
 		.init = &cik_vm_init,
 		.fini = &cik_vm_fini,
-		.set_page = &cik_sdma_vm_set_page,
+		.copy_pages = &cik_sdma_vm_copy_pages,
+		.write_pages = &cik_sdma_vm_write_pages,
+		.set_pages = &cik_sdma_vm_set_pages,
+		.pad_ib = &cik_sdma_vm_pad_ib,
 	},
 	.ring = {
 		[RADEON_RING_TYPE_GFX_INDEX] = &ci_gfx_ring,
@@ -2091,7 +2105,7 @@ static struct radeon_asic kv_asic = {
 	.resume = &cik_resume,
 	.asic_reset = &cik_asic_reset,
 	.vga_set_state = &r600_vga_set_state,
-	.ioctl_wait_idle = NULL,
+	.mmio_hdp_flush = &r600_mmio_hdp_flush,
 	.gui_idle = &r600_gui_idle,
 	.mc_wait_for_idle = &evergreen_mc_wait_for_idle,
 	.get_xclk = &cik_get_xclk,
@@ -2103,7 +2117,10 @@ static struct radeon_asic kv_asic = {
 	.vm = {
 		.init = &cik_vm_init,
 		.fini = &cik_vm_fini,
-		.set_page = &cik_sdma_vm_set_page,
+		.copy_pages = &cik_sdma_vm_copy_pages,
+		.write_pages = &cik_sdma_vm_write_pages,
+		.set_pages = &cik_sdma_vm_set_pages,
+		.pad_ib = &cik_sdma_vm_pad_ib,
 	},
 	.ring = {
 		[RADEON_RING_TYPE_GFX_INDEX] = &ci_gfx_ring,
@@ -2457,7 +2474,7 @@ int radeon_asic_init(struct radeon_device *rdev)
 			rdev->cg_flags =
 				RADEON_CG_SUPPORT_GFX_MGCG |
 				RADEON_CG_SUPPORT_GFX_MGLS |
-				RADEON_CG_SUPPORT_GFX_CGCG |
+				/*RADEON_CG_SUPPORT_GFX_CGCG |*/
 				RADEON_CG_SUPPORT_GFX_CGLS |
 				RADEON_CG_SUPPORT_GFX_CGTS |
 				RADEON_CG_SUPPORT_GFX_CGTS_LS |
@@ -2476,7 +2493,7 @@ int radeon_asic_init(struct radeon_device *rdev)
 			rdev->cg_flags =
 				RADEON_CG_SUPPORT_GFX_MGCG |
 				RADEON_CG_SUPPORT_GFX_MGLS |
-				RADEON_CG_SUPPORT_GFX_CGCG |
+				/*RADEON_CG_SUPPORT_GFX_CGCG |*/
 				RADEON_CG_SUPPORT_GFX_CGLS |
 				RADEON_CG_SUPPORT_GFX_CGTS |
 				RADEON_CG_SUPPORT_GFX_CP_LS |
@@ -2502,7 +2519,7 @@ int radeon_asic_init(struct radeon_device *rdev)
 			rdev->cg_flags =
 				RADEON_CG_SUPPORT_GFX_MGCG |
 				RADEON_CG_SUPPORT_GFX_MGLS |
-				RADEON_CG_SUPPORT_GFX_CGCG |
+				/*RADEON_CG_SUPPORT_GFX_CGCG |*/
 				RADEON_CG_SUPPORT_GFX_CGLS |
 				RADEON_CG_SUPPORT_GFX_CGTS |
 				RADEON_CG_SUPPORT_GFX_CGTS_LS |
@@ -2530,7 +2547,7 @@ int radeon_asic_init(struct radeon_device *rdev)
 			rdev->cg_flags =
 				RADEON_CG_SUPPORT_GFX_MGCG |
 				RADEON_CG_SUPPORT_GFX_MGLS |
-				RADEON_CG_SUPPORT_GFX_CGCG |
+				/*RADEON_CG_SUPPORT_GFX_CGCG |*/
 				RADEON_CG_SUPPORT_GFX_CGLS |
 				RADEON_CG_SUPPORT_GFX_CGTS |
 				RADEON_CG_SUPPORT_GFX_CGTS_LS |
diff --git a/drivers/gpu/drm/radeon/radeon_asic.h b/drivers/gpu/drm/radeon/radeon_asic.h
index 01e7c0ad8f01..275a5dc01780 100644
--- a/drivers/gpu/drm/radeon/radeon_asic.h
+++ b/drivers/gpu/drm/radeon/radeon_asic.h
@@ -68,7 +68,7 @@ int r100_asic_reset(struct radeon_device *rdev);
 u32 r100_get_vblank_counter(struct radeon_device *rdev, int crtc);
 void r100_pci_gart_tlb_flush(struct radeon_device *rdev);
 void r100_pci_gart_set_page(struct radeon_device *rdev, unsigned i,
-			    uint64_t addr);
+			    uint64_t addr, uint32_t flags);
 void r100_ring_start(struct radeon_device *rdev, struct radeon_ring *ring);
 int r100_irq_set(struct radeon_device *rdev);
 int r100_irq_process(struct radeon_device *rdev);
@@ -148,7 +148,8 @@ u32 r100_gfx_get_wptr(struct radeon_device *rdev,
 		      struct radeon_ring *ring);
 void r100_gfx_set_wptr(struct radeon_device *rdev,
 		       struct radeon_ring *ring);
-
+void r100_ring_hdp_flush(struct radeon_device *rdev,
+			 struct radeon_ring *ring);
 /*
  * r200,rv250,rs300,rv280
  */
@@ -173,7 +174,7 @@ extern void r300_fence_ring_emit(struct radeon_device *rdev,
 extern int r300_cs_parse(struct radeon_cs_parser *p);
 extern void rv370_pcie_gart_tlb_flush(struct radeon_device *rdev);
 extern void rv370_pcie_gart_set_page(struct radeon_device *rdev, unsigned i,
-				     uint64_t addr);
+				     uint64_t addr, uint32_t flags);
 extern void rv370_set_pcie_lanes(struct radeon_device *rdev, int lanes);
 extern int rv370_get_pcie_lanes(struct radeon_device *rdev);
 extern void r300_set_reg_safe(struct radeon_device *rdev);
@@ -209,7 +210,7 @@ extern int rs400_suspend(struct radeon_device *rdev);
 extern int rs400_resume(struct radeon_device *rdev);
 void rs400_gart_tlb_flush(struct radeon_device *rdev);
 void rs400_gart_set_page(struct radeon_device *rdev, unsigned i,
-			 uint64_t addr);
+			 uint64_t addr, uint32_t flags);
 uint32_t rs400_mc_rreg(struct radeon_device *rdev, uint32_t reg);
 void rs400_mc_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v);
 int rs400_gart_init(struct radeon_device *rdev);
@@ -233,7 +234,7 @@ void rs600_irq_disable(struct radeon_device *rdev);
 u32 rs600_get_vblank_counter(struct radeon_device *rdev, int crtc);
 void rs600_gart_tlb_flush(struct radeon_device *rdev);
 void rs600_gart_set_page(struct radeon_device *rdev, unsigned i,
-			 uint64_t addr);
+			 uint64_t addr, uint32_t flags);
 uint32_t rs600_mc_rreg(struct radeon_device *rdev, uint32_t reg);
 void rs600_mc_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v);
 void rs600_bandwidth_update(struct radeon_device *rdev);
@@ -351,7 +352,7 @@ void r600_hpd_fini(struct radeon_device *rdev);
 bool r600_hpd_sense(struct radeon_device *rdev, enum radeon_hpd_id hpd);
 void r600_hpd_set_polarity(struct radeon_device *rdev,
 			   enum radeon_hpd_id hpd);
-extern void r600_ioctl_wait_idle(struct radeon_device *rdev, struct radeon_bo *bo);
+extern void r600_mmio_hdp_flush(struct radeon_device *rdev);
 extern bool r600_gui_idle(struct radeon_device *rdev);
 extern void r600_pm_misc(struct radeon_device *rdev);
 extern void r600_pm_init_profile(struct radeon_device *rdev);
@@ -606,11 +607,22 @@ void cayman_dma_ring_ib_execute(struct radeon_device *rdev,
 				struct radeon_ib *ib);
 bool cayman_gfx_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring);
 bool cayman_dma_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring);
-void cayman_dma_vm_set_page(struct radeon_device *rdev,
-			    struct radeon_ib *ib,
-			    uint64_t pe,
-			    uint64_t addr, unsigned count,
-			    uint32_t incr, uint32_t flags);
+
+void cayman_dma_vm_copy_pages(struct radeon_device *rdev,
+			      struct radeon_ib *ib,
+			      uint64_t pe, uint64_t src,
+			      unsigned count);
+void cayman_dma_vm_write_pages(struct radeon_device *rdev,
+			       struct radeon_ib *ib,
+			       uint64_t pe,
+			       uint64_t addr, unsigned count,
+			       uint32_t incr, uint32_t flags);
+void cayman_dma_vm_set_pages(struct radeon_device *rdev,
+			     struct radeon_ib *ib,
+			     uint64_t pe,
+			     uint64_t addr, unsigned count,
+			     uint32_t incr, uint32_t flags);
+void cayman_dma_vm_pad_ib(struct radeon_ib *ib);
 
 void cayman_dma_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm);
 
@@ -693,11 +705,22 @@ int si_copy_dma(struct radeon_device *rdev,
 		uint64_t src_offset, uint64_t dst_offset,
 		unsigned num_gpu_pages,
 		struct radeon_fence **fence);
-void si_dma_vm_set_page(struct radeon_device *rdev,
-			struct radeon_ib *ib,
-			uint64_t pe,
-			uint64_t addr, unsigned count,
-			uint32_t incr, uint32_t flags);
+
+void si_dma_vm_copy_pages(struct radeon_device *rdev,
+			  struct radeon_ib *ib,
+			  uint64_t pe, uint64_t src,
+			  unsigned count);
+void si_dma_vm_write_pages(struct radeon_device *rdev,
+			   struct radeon_ib *ib,
+			   uint64_t pe,
+			   uint64_t addr, unsigned count,
+			   uint32_t incr, uint32_t flags);
+void si_dma_vm_set_pages(struct radeon_device *rdev,
+			 struct radeon_ib *ib,
+			 uint64_t pe,
+			 uint64_t addr, unsigned count,
+			 uint32_t incr, uint32_t flags);
+
 void si_dma_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm);
 u32 si_get_xclk(struct radeon_device *rdev);
 uint64_t si_get_gpu_clock_counter(struct radeon_device *rdev);
@@ -771,11 +794,23 @@ int cik_irq_process(struct radeon_device *rdev);
 int cik_vm_init(struct radeon_device *rdev);
 void cik_vm_fini(struct radeon_device *rdev);
 void cik_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm);
-void cik_sdma_vm_set_page(struct radeon_device *rdev,
-			  struct radeon_ib *ib,
-			  uint64_t pe,
-			  uint64_t addr, unsigned count,
-			  uint32_t incr, uint32_t flags);
+
+void cik_sdma_vm_copy_pages(struct radeon_device *rdev,
+			    struct radeon_ib *ib,
+			    uint64_t pe, uint64_t src,
+			    unsigned count);
+void cik_sdma_vm_write_pages(struct radeon_device *rdev,
+			     struct radeon_ib *ib,
+			     uint64_t pe,
+			     uint64_t addr, unsigned count,
+			     uint32_t incr, uint32_t flags);
+void cik_sdma_vm_set_pages(struct radeon_device *rdev,
+			   struct radeon_ib *ib,
+			   uint64_t pe,
+			   uint64_t addr, unsigned count,
+			   uint32_t incr, uint32_t flags);
+void cik_sdma_vm_pad_ib(struct radeon_ib *ib);
+
 void cik_dma_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm);
 int cik_ib_parse(struct radeon_device *rdev, struct radeon_ib *ib);
 u32 cik_gfx_get_rptr(struct radeon_device *rdev,
diff --git a/drivers/gpu/drm/radeon/radeon_atombios.c b/drivers/gpu/drm/radeon/radeon_atombios.c
index 173f378428a9..92b2d8dd4735 100644
--- a/drivers/gpu/drm/radeon/radeon_atombios.c
+++ b/drivers/gpu/drm/radeon/radeon_atombios.c
@@ -1963,7 +1963,7 @@ static const char *thermal_controller_names[] = {
 	"adm1032",
 	"adm1030",
 	"max6649",
-	"lm64",
+	"lm63", /* lm64 */
 	"f75375",
 	"asc7xxx",
 };
@@ -1974,7 +1974,7 @@ static const char *pp_lib_thermal_controller_names[] = {
 	"adm1032",
 	"adm1030",
 	"max6649",
-	"lm64",
+	"lm63", /* lm64 */
 	"f75375",
 	"RV6xx",
 	"RV770",
@@ -3236,6 +3236,41 @@ int radeon_atom_get_leakage_vddc_based_on_leakage_params(struct radeon_device *r
 	return 0;
 }
 
+union get_voltage_info {
+	struct  _GET_VOLTAGE_INFO_INPUT_PARAMETER_V1_2 in;
+	struct  _GET_EVV_VOLTAGE_INFO_OUTPUT_PARAMETER_V1_2 evv_out;
+};
+
+int radeon_atom_get_voltage_evv(struct radeon_device *rdev,
+				u16 virtual_voltage_id,
+				u16 *voltage)
+{
+	int index = GetIndexIntoMasterTable(COMMAND, GetVoltageInfo);
+	u32 entry_id;
+	u32 count = rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk.count;
+	union get_voltage_info args;
+
+	for (entry_id = 0; entry_id < count; entry_id++) {
+		if (rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk.entries[entry_id].v ==
+		    virtual_voltage_id)
+			break;
+	}
+
+	if (entry_id >= count)
+		return -EINVAL;
+
+	args.in.ucVoltageType = VOLTAGE_TYPE_VDDC;
+	args.in.ucVoltageMode = ATOM_GET_VOLTAGE_EVV_VOLTAGE;
+	args.in.ulSCLKFreq =
+		cpu_to_le32(rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk.entries[entry_id].clk);
+
+	atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
+
+	*voltage = le16_to_cpu(args.evv_out.usVoltageLevel);
+
+	return 0;
+}
+
 int radeon_atom_get_voltage_gpio_settings(struct radeon_device *rdev,
 					  u16 voltage_level, u8 voltage_type,
 					  u32 *gpio_value, u32 *gpio_mask)
@@ -3397,6 +3432,50 @@ radeon_atom_is_voltage_gpio(struct radeon_device *rdev,
 	return false;
 }
 
+int radeon_atom_get_svi2_info(struct radeon_device *rdev,
+			      u8 voltage_type,
+			      u8 *svd_gpio_id, u8 *svc_gpio_id)
+{
+	int index = GetIndexIntoMasterTable(DATA, VoltageObjectInfo);
+	u8 frev, crev;
+	u16 data_offset, size;
+	union voltage_object_info *voltage_info;
+	union voltage_object *voltage_object = NULL;
+
+	if (atom_parse_data_header(rdev->mode_info.atom_context, index, &size,
+				   &frev, &crev, &data_offset)) {
+		voltage_info = (union voltage_object_info *)
+			(rdev->mode_info.atom_context->bios + data_offset);
+
+		switch (frev) {
+		case 3:
+			switch (crev) {
+			case 1:
+				voltage_object = (union voltage_object *)
+					atom_lookup_voltage_object_v3(&voltage_info->v3,
+								      voltage_type,
+								      VOLTAGE_OBJ_SVID2);
+				if (voltage_object) {
+					*svd_gpio_id = voltage_object->v3.asSVID2Obj.ucSVDGpioId;
+					*svc_gpio_id = voltage_object->v3.asSVID2Obj.ucSVCGpioId;
+				} else {
+					return -EINVAL;
+				}
+				break;
+			default:
+				DRM_ERROR("unknown voltage object table\n");
+				return -EINVAL;
+			}
+			break;
+		default:
+			DRM_ERROR("unknown voltage object table\n");
+			return -EINVAL;
+		}
+
+	}
+	return 0;
+}
+
 int radeon_atom_get_max_voltage(struct radeon_device *rdev,
 				u8 voltage_type, u16 *max_voltage)
 {
diff --git a/drivers/gpu/drm/radeon/radeon_benchmark.c b/drivers/gpu/drm/radeon/radeon_benchmark.c
index 6e05a2e75a46..69f5695bdab9 100644
--- a/drivers/gpu/drm/radeon/radeon_benchmark.c
+++ b/drivers/gpu/drm/radeon/radeon_benchmark.c
@@ -97,7 +97,7 @@ static void radeon_benchmark_move(struct radeon_device *rdev, unsigned size,
 	int time;
 
 	n = RADEON_BENCHMARK_ITERATIONS;
-	r = radeon_bo_create(rdev, size, PAGE_SIZE, true, sdomain, NULL, &sobj);
+	r = radeon_bo_create(rdev, size, PAGE_SIZE, true, sdomain, 0, NULL, &sobj);
 	if (r) {
 		goto out_cleanup;
 	}
@@ -109,7 +109,7 @@ static void radeon_benchmark_move(struct radeon_device *rdev, unsigned size,
 	if (r) {
 		goto out_cleanup;
 	}
-	r = radeon_bo_create(rdev, size, PAGE_SIZE, true, ddomain, NULL, &dobj);
+	r = radeon_bo_create(rdev, size, PAGE_SIZE, true, ddomain, 0, NULL, &dobj);
 	if (r) {
 		goto out_cleanup;
 	}
diff --git a/drivers/gpu/drm/radeon/radeon_connectors.c b/drivers/gpu/drm/radeon/radeon_connectors.c
index 44831197e82e..300c4b3d4669 100644
--- a/drivers/gpu/drm/radeon/radeon_connectors.c
+++ b/drivers/gpu/drm/radeon/radeon_connectors.c
@@ -107,7 +107,7 @@ int radeon_get_monitor_bpc(struct drm_connector *connector)
 	case DRM_MODE_CONNECTOR_DVII:
 	case DRM_MODE_CONNECTOR_HDMIB:
 		if (radeon_connector->use_digital) {
-			if (drm_detect_hdmi_monitor(radeon_connector->edid)) {
+			if (drm_detect_hdmi_monitor(radeon_connector_edid(connector))) {
 				if (connector->display_info.bpc)
 					bpc = connector->display_info.bpc;
 			}
@@ -115,7 +115,7 @@ int radeon_get_monitor_bpc(struct drm_connector *connector)
 		break;
 	case DRM_MODE_CONNECTOR_DVID:
 	case DRM_MODE_CONNECTOR_HDMIA:
-		if (drm_detect_hdmi_monitor(radeon_connector->edid)) {
+		if (drm_detect_hdmi_monitor(radeon_connector_edid(connector))) {
 			if (connector->display_info.bpc)
 				bpc = connector->display_info.bpc;
 		}
@@ -124,7 +124,7 @@ int radeon_get_monitor_bpc(struct drm_connector *connector)
 		dig_connector = radeon_connector->con_priv;
 		if ((dig_connector->dp_sink_type == CONNECTOR_OBJECT_ID_DISPLAYPORT) ||
 		    (dig_connector->dp_sink_type == CONNECTOR_OBJECT_ID_eDP) ||
-		    drm_detect_hdmi_monitor(radeon_connector->edid)) {
+		    drm_detect_hdmi_monitor(radeon_connector_edid(connector))) {
 			if (connector->display_info.bpc)
 				bpc = connector->display_info.bpc;
 		}
@@ -148,7 +148,7 @@ int radeon_get_monitor_bpc(struct drm_connector *connector)
 		break;
 	}
 
-	if (drm_detect_hdmi_monitor(radeon_connector->edid)) {
+	if (drm_detect_hdmi_monitor(radeon_connector_edid(connector))) {
 		/* hdmi deep color only implemented on DCE4+ */
 		if ((bpc > 8) && !ASIC_IS_DCE4(rdev)) {
 			DRM_DEBUG("%s: HDMI deep color %d bpc unsupported. Using 8 bpc.\n",
@@ -197,10 +197,19 @@ int radeon_get_monitor_bpc(struct drm_connector *connector)
 						  connector->name, bpc);
 			}
 		}
+		else if (bpc > 8) {
+			/* max_tmds_clock missing, but hdmi spec mandates it for deep color. */
+			DRM_DEBUG("%s: Required max tmds clock for HDMI deep color missing. Using 8 bpc.\n",
+					  connector->name);
+			bpc = 8;
+		}
 	}
 
-	if ((radeon_deep_color == 0) && (bpc > 8))
+	if ((radeon_deep_color == 0) && (bpc > 8)) {
+		DRM_DEBUG("%s: Deep color disabled. Set radeon module param deep_color=1 to enable.\n",
+				  connector->name);
 		bpc = 8;
+	}
 
 	DRM_DEBUG("%s: Display bpc=%d, returned bpc=%d\n",
 			  connector->name, connector->display_info.bpc, bpc);
@@ -216,7 +225,6 @@ radeon_connector_update_scratch_regs(struct drm_connector *connector, enum drm_c
 	struct drm_encoder *best_encoder = NULL;
 	struct drm_encoder *encoder = NULL;
 	struct drm_connector_helper_funcs *connector_funcs = connector->helper_private;
-	struct drm_mode_object *obj;
 	bool connected;
 	int i;
 
@@ -226,14 +234,11 @@ radeon_connector_update_scratch_regs(struct drm_connector *connector, enum drm_c
 		if (connector->encoder_ids[i] == 0)
 			break;
 
-		obj = drm_mode_object_find(connector->dev,
-					   connector->encoder_ids[i],
-					   DRM_MODE_OBJECT_ENCODER);
-		if (!obj)
+		encoder = drm_encoder_find(connector->dev,
+					   connector->encoder_ids[i]);
+		if (!encoder)
 			continue;
 
-		encoder = obj_to_encoder(obj);
-
 		if ((encoder == best_encoder) && (status == connector_status_connected))
 			connected = true;
 		else
@@ -249,7 +254,6 @@ radeon_connector_update_scratch_regs(struct drm_connector *connector, enum drm_c
 
 static struct drm_encoder *radeon_find_encoder(struct drm_connector *connector, int encoder_type)
 {
-	struct drm_mode_object *obj;
 	struct drm_encoder *encoder;
 	int i;
 
@@ -257,34 +261,134 @@ static struct drm_encoder *radeon_find_encoder(struct drm_connector *connector,
 		if (connector->encoder_ids[i] == 0)
 			break;
 
-		obj = drm_mode_object_find(connector->dev, connector->encoder_ids[i], DRM_MODE_OBJECT_ENCODER);
-		if (!obj)
+		encoder = drm_encoder_find(connector->dev, connector->encoder_ids[i]);
+		if (!encoder)
 			continue;
 
-		encoder = obj_to_encoder(obj);
 		if (encoder->encoder_type == encoder_type)
 			return encoder;
 	}
 	return NULL;
 }
 
+struct edid *radeon_connector_edid(struct drm_connector *connector)
+{
+	struct radeon_connector *radeon_connector = to_radeon_connector(connector);
+	struct drm_property_blob *edid_blob = connector->edid_blob_ptr;
+
+	if (radeon_connector->edid) {
+		return radeon_connector->edid;
+	} else if (edid_blob) {
+		struct edid *edid = kmemdup(edid_blob->data, edid_blob->length, GFP_KERNEL);
+		if (edid)
+			radeon_connector->edid = edid;
+	}
+	return radeon_connector->edid;
+}
+
+static void radeon_connector_get_edid(struct drm_connector *connector)
+{
+	struct drm_device *dev = connector->dev;
+	struct radeon_device *rdev = dev->dev_private;
+	struct radeon_connector *radeon_connector = to_radeon_connector(connector);
+
+	if (radeon_connector->edid)
+		return;
+
+	/* on hw with routers, select right port */
+	if (radeon_connector->router.ddc_valid)
+		radeon_router_select_ddc_port(radeon_connector);
+
+	if ((radeon_connector_encoder_get_dp_bridge_encoder_id(connector) !=
+	     ENCODER_OBJECT_ID_NONE) &&
+	    radeon_connector->ddc_bus->has_aux) {
+		radeon_connector->edid = drm_get_edid(connector,
+						      &radeon_connector->ddc_bus->aux.ddc);
+	} else if ((connector->connector_type == DRM_MODE_CONNECTOR_DisplayPort) ||
+		   (connector->connector_type == DRM_MODE_CONNECTOR_eDP)) {
+		struct radeon_connector_atom_dig *dig = radeon_connector->con_priv;
+
+		if ((dig->dp_sink_type == CONNECTOR_OBJECT_ID_DISPLAYPORT ||
+		     dig->dp_sink_type == CONNECTOR_OBJECT_ID_eDP) &&
+		    radeon_connector->ddc_bus->has_aux)
+			radeon_connector->edid = drm_get_edid(&radeon_connector->base,
+							      &radeon_connector->ddc_bus->aux.ddc);
+		else if (radeon_connector->ddc_bus)
+			radeon_connector->edid = drm_get_edid(&radeon_connector->base,
+							      &radeon_connector->ddc_bus->adapter);
+	} else if (radeon_connector->ddc_bus) {
+		radeon_connector->edid = drm_get_edid(&radeon_connector->base,
+						      &radeon_connector->ddc_bus->adapter);
+	}
+
+	if (!radeon_connector->edid) {
+		if (rdev->is_atom_bios) {
+			/* some laptops provide a hardcoded edid in rom for LCDs */
+			if (((connector->connector_type == DRM_MODE_CONNECTOR_LVDS) ||
+			     (connector->connector_type == DRM_MODE_CONNECTOR_eDP)))
+				radeon_connector->edid = radeon_bios_get_hardcoded_edid(rdev);
+		} else {
+			/* some servers provide a hardcoded edid in rom for KVMs */
+			radeon_connector->edid = radeon_bios_get_hardcoded_edid(rdev);
+		}
+	}
+}
+
+static void radeon_connector_free_edid(struct drm_connector *connector)
+{
+	struct radeon_connector *radeon_connector = to_radeon_connector(connector);
+
+	if (radeon_connector->edid) {
+		kfree(radeon_connector->edid);
+		radeon_connector->edid = NULL;
+	}
+}
+
+static int radeon_ddc_get_modes(struct drm_connector *connector)
+{
+	struct radeon_connector *radeon_connector = to_radeon_connector(connector);
+	int ret;
+
+	if (radeon_connector->edid) {
+		drm_mode_connector_update_edid_property(connector, radeon_connector->edid);
+		ret = drm_add_edid_modes(connector, radeon_connector->edid);
+		drm_edid_to_eld(connector, radeon_connector->edid);
+		return ret;
+	}
+	drm_mode_connector_update_edid_property(connector, NULL);
+	return 0;
+}
+
 static struct drm_encoder *radeon_best_single_encoder(struct drm_connector *connector)
 {
 	int enc_id = connector->encoder_ids[0];
-	struct drm_mode_object *obj;
-	struct drm_encoder *encoder;
-
 	/* pick the encoder ids */
-	if (enc_id) {
-		obj = drm_mode_object_find(connector->dev, enc_id, DRM_MODE_OBJECT_ENCODER);
-		if (!obj)
-			return NULL;
-		encoder = obj_to_encoder(obj);
-		return encoder;
-	}
+	if (enc_id)
+		return drm_encoder_find(connector->dev, enc_id);
 	return NULL;
 }
 
+static void radeon_get_native_mode(struct drm_connector *connector)
+{
+	struct drm_encoder *encoder = radeon_best_single_encoder(connector);
+	struct radeon_encoder *radeon_encoder;
+
+	if (encoder == NULL)
+		return;
+
+	radeon_encoder = to_radeon_encoder(encoder);
+
+	if (!list_empty(&connector->probed_modes)) {
+		struct drm_display_mode *preferred_mode =
+			list_first_entry(&connector->probed_modes,
+					 struct drm_display_mode, head);
+
+		radeon_encoder->native_mode = *preferred_mode;
+	} else {
+		radeon_encoder->native_mode.clock = 0;
+	}
+}
+
 /*
  * radeon_connector_analog_encoder_conflict_solve
  * - search for other connectors sharing this encoder
@@ -585,6 +689,35 @@ static int radeon_connector_set_property(struct drm_connector *connector, struct
 		radeon_property_change_mode(&radeon_encoder->base);
 	}
 
+	if (property == dev->mode_config.scaling_mode_property) {
+		enum radeon_rmx_type rmx_type;
+
+		if (connector->encoder)
+			radeon_encoder = to_radeon_encoder(connector->encoder);
+		else {
+			struct drm_connector_helper_funcs *connector_funcs = connector->helper_private;
+			radeon_encoder = to_radeon_encoder(connector_funcs->best_encoder(connector));
+		}
+
+		switch (val) {
+		default:
+		case DRM_MODE_SCALE_NONE: rmx_type = RMX_OFF; break;
+		case DRM_MODE_SCALE_CENTER: rmx_type = RMX_CENTER; break;
+		case DRM_MODE_SCALE_ASPECT: rmx_type = RMX_ASPECT; break;
+		case DRM_MODE_SCALE_FULLSCREEN: rmx_type = RMX_FULL; break;
+		}
+		if (radeon_encoder->rmx_type == rmx_type)
+			return 0;
+
+		if ((rmx_type != DRM_MODE_SCALE_NONE) &&
+		    (radeon_encoder->native_mode.clock == 0))
+			return 0;
+
+		radeon_encoder->rmx_type = rmx_type;
+
+		radeon_property_change_mode(&radeon_encoder->base);
+	}
+
 	return 0;
 }
 
@@ -625,22 +758,20 @@ static void radeon_fixup_lvds_native_mode(struct drm_encoder *encoder,
 
 static int radeon_lvds_get_modes(struct drm_connector *connector)
 {
-	struct radeon_connector *radeon_connector = to_radeon_connector(connector);
 	struct drm_encoder *encoder;
 	int ret = 0;
 	struct drm_display_mode *mode;
 
-	if (radeon_connector->ddc_bus) {
-		ret = radeon_ddc_get_modes(radeon_connector);
-		if (ret > 0) {
-			encoder = radeon_best_single_encoder(connector);
-			if (encoder) {
-				radeon_fixup_lvds_native_mode(encoder, connector);
-				/* add scaled modes */
-				radeon_add_common_modes(encoder, connector);
-			}
-			return ret;
+	radeon_connector_get_edid(connector);
+	ret = radeon_ddc_get_modes(connector);
+	if (ret > 0) {
+		encoder = radeon_best_single_encoder(connector);
+		if (encoder) {
+			radeon_fixup_lvds_native_mode(encoder, connector);
+			/* add scaled modes */
+			radeon_add_common_modes(encoder, connector);
 		}
+		return ret;
 	}
 
 	encoder = radeon_best_single_encoder(connector);
@@ -715,16 +846,9 @@ radeon_lvds_detect(struct drm_connector *connector, bool force)
 	}
 
 	/* check for edid as well */
+	radeon_connector_get_edid(connector);
 	if (radeon_connector->edid)
 		ret = connector_status_connected;
-	else {
-		if (radeon_connector->ddc_bus) {
-			radeon_connector->edid = drm_get_edid(&radeon_connector->base,
-							      &radeon_connector->ddc_bus->adapter);
-			if (radeon_connector->edid)
-				ret = connector_status_connected;
-		}
-	}
 	/* check acpi lid status ??? */
 
 	radeon_connector_update_scratch_regs(connector, ret);
@@ -737,10 +861,9 @@ static void radeon_connector_destroy(struct drm_connector *connector)
 {
 	struct radeon_connector *radeon_connector = to_radeon_connector(connector);
 
-	if (radeon_connector->edid)
-		kfree(radeon_connector->edid);
+	radeon_connector_free_edid(connector);
 	kfree(radeon_connector->con_priv);
-	drm_sysfs_connector_remove(connector);
+	drm_connector_unregister(connector);
 	drm_connector_cleanup(connector);
 	kfree(connector);
 }
@@ -797,10 +920,12 @@ static const struct drm_connector_funcs radeon_lvds_connector_funcs = {
 
 static int radeon_vga_get_modes(struct drm_connector *connector)
 {
-	struct radeon_connector *radeon_connector = to_radeon_connector(connector);
 	int ret;
 
-	ret = radeon_ddc_get_modes(radeon_connector);
+	radeon_connector_get_edid(connector);
+	ret = radeon_ddc_get_modes(connector);
+
+	radeon_get_native_mode(connector);
 
 	return ret;
 }
@@ -843,28 +968,26 @@ radeon_vga_detect(struct drm_connector *connector, bool force)
 		dret = radeon_ddc_probe(radeon_connector, false);
 	if (dret) {
 		radeon_connector->detected_by_load = false;
-		if (radeon_connector->edid) {
-			kfree(radeon_connector->edid);
-			radeon_connector->edid = NULL;
-		}
-		radeon_connector->edid = drm_get_edid(&radeon_connector->base, &radeon_connector->ddc_bus->adapter);
+		radeon_connector_free_edid(connector);
+		radeon_connector_get_edid(connector);
 
 		if (!radeon_connector->edid) {
 			DRM_ERROR("%s: probed a monitor but no|invalid EDID\n",
 					connector->name);
 			ret = connector_status_connected;
 		} else {
-			radeon_connector->use_digital = !!(radeon_connector->edid->input & DRM_EDID_INPUT_DIGITAL);
+			radeon_connector->use_digital =
+				!!(radeon_connector->edid->input & DRM_EDID_INPUT_DIGITAL);
 
 			/* some oems have boards with separate digital and analog connectors
 			 * with a shared ddc line (often vga + hdmi)
 			 */
 			if (radeon_connector->use_digital && radeon_connector->shared_ddc) {
-				kfree(radeon_connector->edid);
-				radeon_connector->edid = NULL;
+				radeon_connector_free_edid(connector);
 				ret = connector_status_disconnected;
-			} else
+			} else {
 				ret = connector_status_connected;
+			}
 		}
 	} else {
 
@@ -999,15 +1122,6 @@ static const struct drm_connector_funcs radeon_tv_connector_funcs = {
 	.set_property = radeon_connector_set_property,
 };
 
-static int radeon_dvi_get_modes(struct drm_connector *connector)
-{
-	struct radeon_connector *radeon_connector = to_radeon_connector(connector);
-	int ret;
-
-	ret = radeon_ddc_get_modes(radeon_connector);
-	return ret;
-}
-
 static bool radeon_check_hpd_status_unchanged(struct drm_connector *connector)
 {
 	struct drm_device *dev = connector->dev;
@@ -1048,7 +1162,6 @@ radeon_dvi_detect(struct drm_connector *connector, bool force)
 	struct radeon_connector *radeon_connector = to_radeon_connector(connector);
 	struct drm_encoder *encoder = NULL;
 	struct drm_encoder_helper_funcs *encoder_funcs;
-	struct drm_mode_object *obj;
 	int i, r;
 	enum drm_connector_status ret = connector_status_disconnected;
 	bool dret = false, broken_edid = false;
@@ -1066,18 +1179,16 @@ radeon_dvi_detect(struct drm_connector *connector, bool force)
 		dret = radeon_ddc_probe(radeon_connector, false);
 	if (dret) {
 		radeon_connector->detected_by_load = false;
-		if (radeon_connector->edid) {
-			kfree(radeon_connector->edid);
-			radeon_connector->edid = NULL;
-		}
-		radeon_connector->edid = drm_get_edid(&radeon_connector->base, &radeon_connector->ddc_bus->adapter);
+		radeon_connector_free_edid(connector);
+		radeon_connector_get_edid(connector);
 
 		if (!radeon_connector->edid) {
 			DRM_ERROR("%s: probed a monitor but no|invalid EDID\n",
 					connector->name);
 			/* rs690 seems to have a problem with connectors not existing and always
 			 * return a block of 0's. If we see this just stop polling on this output */
-			if ((rdev->family == CHIP_RS690 || rdev->family == CHIP_RS740) && radeon_connector->base.null_edid_counter) {
+			if ((rdev->family == CHIP_RS690 || rdev->family == CHIP_RS740) &&
+			    radeon_connector->base.null_edid_counter) {
 				ret = connector_status_disconnected;
 				DRM_ERROR("%s: detected RS690 floating bus bug, stopping ddc detect\n",
 					  connector->name);
@@ -1087,18 +1198,18 @@ radeon_dvi_detect(struct drm_connector *connector, bool force)
 				broken_edid = true; /* defer use_digital to later */
 			}
 		} else {
-			radeon_connector->use_digital = !!(radeon_connector->edid->input & DRM_EDID_INPUT_DIGITAL);
+			radeon_connector->use_digital =
+				!!(radeon_connector->edid->input & DRM_EDID_INPUT_DIGITAL);
 
 			/* some oems have boards with separate digital and analog connectors
 			 * with a shared ddc line (often vga + hdmi)
 			 */
 			if ((!radeon_connector->use_digital) && radeon_connector->shared_ddc) {
-				kfree(radeon_connector->edid);
-				radeon_connector->edid = NULL;
+				radeon_connector_free_edid(connector);
 				ret = connector_status_disconnected;
-			} else
+			} else {
 				ret = connector_status_connected;
-
+			}
 			/* This gets complicated.  We have boards with VGA + HDMI with a
 			 * shared DDC line and we have boards with DVI-D + HDMI with a shared
 			 * DDC line.  The latter is more complex because with DVI<->HDMI adapters
@@ -1118,8 +1229,7 @@ radeon_dvi_detect(struct drm_connector *connector, bool force)
 						if (list_connector->connector_type != DRM_MODE_CONNECTOR_VGA) {
 							/* hpd is our only option in this case */
 							if (!radeon_hpd_sense(rdev, radeon_connector->hpd.hpd)) {
-								kfree(radeon_connector->edid);
-								radeon_connector->edid = NULL;
+								radeon_connector_free_edid(connector);
 								ret = connector_status_disconnected;
 							}
 						}
@@ -1153,14 +1263,11 @@ radeon_dvi_detect(struct drm_connector *connector, bool force)
 			if (connector->encoder_ids[i] == 0)
 				break;
 
-			obj = drm_mode_object_find(connector->dev,
-						   connector->encoder_ids[i],
-						   DRM_MODE_OBJECT_ENCODER);
-			if (!obj)
+			encoder = drm_encoder_find(connector->dev,
+						   connector->encoder_ids[i]);
+			if (!encoder)
 				continue;
 
-			encoder = obj_to_encoder(obj);
-
 			if (encoder->encoder_type != DRM_MODE_ENCODER_DAC &&
 			    encoder->encoder_type != DRM_MODE_ENCODER_TVDAC)
 				continue;
@@ -1225,19 +1332,16 @@ static struct drm_encoder *radeon_dvi_encoder(struct drm_connector *connector)
 {
 	int enc_id = connector->encoder_ids[0];
 	struct radeon_connector *radeon_connector = to_radeon_connector(connector);
-	struct drm_mode_object *obj;
 	struct drm_encoder *encoder;
 	int i;
 	for (i = 0; i < DRM_CONNECTOR_MAX_ENCODER; i++) {
 		if (connector->encoder_ids[i] == 0)
 			break;
 
-		obj = drm_mode_object_find(connector->dev, connector->encoder_ids[i], DRM_MODE_OBJECT_ENCODER);
-		if (!obj)
+		encoder = drm_encoder_find(connector->dev, connector->encoder_ids[i]);
+		if (!encoder)
 			continue;
 
-		encoder = obj_to_encoder(obj);
-
 		if (radeon_connector->use_digital == true) {
 			if (encoder->encoder_type == DRM_MODE_ENCODER_TMDS)
 				return encoder;
@@ -1252,13 +1356,8 @@ static struct drm_encoder *radeon_dvi_encoder(struct drm_connector *connector)
 
 	/* then check use digitial */
 	/* pick the first one */
-	if (enc_id) {
-		obj = drm_mode_object_find(connector->dev, enc_id, DRM_MODE_OBJECT_ENCODER);
-		if (!obj)
-			return NULL;
-		encoder = obj_to_encoder(obj);
-		return encoder;
-	}
+	if (enc_id)
+		return drm_encoder_find(connector->dev, enc_id);
 	return NULL;
 }
 
@@ -1291,7 +1390,7 @@ static int radeon_dvi_mode_valid(struct drm_connector *connector,
 		    (radeon_connector->connector_object_id == CONNECTOR_OBJECT_ID_DUAL_LINK_DVI_D) ||
 		    (radeon_connector->connector_object_id == CONNECTOR_OBJECT_ID_HDMI_TYPE_B))
 			return MODE_OK;
-		else if (ASIC_IS_DCE6(rdev) && drm_detect_hdmi_monitor(radeon_connector->edid)) {
+		else if (ASIC_IS_DCE6(rdev) && drm_detect_hdmi_monitor(radeon_connector_edid(connector))) {
 			/* HDMI 1.3+ supports max clock of 340 Mhz */
 			if (mode->clock > 340000)
 				return MODE_CLOCK_HIGH;
@@ -1310,7 +1409,7 @@ static int radeon_dvi_mode_valid(struct drm_connector *connector,
 }
 
 static const struct drm_connector_helper_funcs radeon_dvi_connector_helper_funcs = {
-	.get_modes = radeon_dvi_get_modes,
+	.get_modes = radeon_vga_get_modes,
 	.mode_valid = radeon_dvi_mode_valid,
 	.best_encoder = radeon_dvi_encoder,
 };
@@ -1339,7 +1438,8 @@ static int radeon_dp_get_modes(struct drm_connector *connector)
 			if (!radeon_dig_connector->edp_on)
 				atombios_set_edp_panel_power(connector,
 							     ATOM_TRANSMITTER_ACTION_POWER_ON);
-			ret = radeon_ddc_get_modes(radeon_connector);
+			radeon_connector_get_edid(connector);
+			ret = radeon_ddc_get_modes(connector);
 			if (!radeon_dig_connector->edp_on)
 				atombios_set_edp_panel_power(connector,
 							     ATOM_TRANSMITTER_ACTION_POWER_OFF);
@@ -1350,7 +1450,8 @@ static int radeon_dp_get_modes(struct drm_connector *connector)
 				if (encoder)
 					radeon_atom_ext_encoder_setup_ddc(encoder);
 			}
-			ret = radeon_ddc_get_modes(radeon_connector);
+			radeon_connector_get_edid(connector);
+			ret = radeon_ddc_get_modes(connector);
 		}
 
 		if (ret > 0) {
@@ -1383,7 +1484,10 @@ static int radeon_dp_get_modes(struct drm_connector *connector)
 			if (encoder)
 				radeon_atom_ext_encoder_setup_ddc(encoder);
 		}
-		ret = radeon_ddc_get_modes(radeon_connector);
+		radeon_connector_get_edid(connector);
+		ret = radeon_ddc_get_modes(connector);
+
+		radeon_get_native_mode(connector);
 	}
 
 	return ret;
@@ -1391,7 +1495,6 @@ static int radeon_dp_get_modes(struct drm_connector *connector)
 
 u16 radeon_connector_encoder_get_dp_bridge_encoder_id(struct drm_connector *connector)
 {
-	struct drm_mode_object *obj;
 	struct drm_encoder *encoder;
 	struct radeon_encoder *radeon_encoder;
 	int i;
@@ -1400,11 +1503,10 @@ u16 radeon_connector_encoder_get_dp_bridge_encoder_id(struct drm_connector *conn
 		if (connector->encoder_ids[i] == 0)
 			break;
 
-		obj = drm_mode_object_find(connector->dev, connector->encoder_ids[i], DRM_MODE_OBJECT_ENCODER);
-		if (!obj)
+		encoder = drm_encoder_find(connector->dev, connector->encoder_ids[i]);
+		if (!encoder)
 			continue;
 
-		encoder = obj_to_encoder(obj);
 		radeon_encoder = to_radeon_encoder(encoder);
 
 		switch (radeon_encoder->encoder_id) {
@@ -1419,9 +1521,8 @@ u16 radeon_connector_encoder_get_dp_bridge_encoder_id(struct drm_connector *conn
 	return ENCODER_OBJECT_ID_NONE;
 }
 
-bool radeon_connector_encoder_is_hbr2(struct drm_connector *connector)
+static bool radeon_connector_encoder_is_hbr2(struct drm_connector *connector)
 {
-	struct drm_mode_object *obj;
 	struct drm_encoder *encoder;
 	struct radeon_encoder *radeon_encoder;
 	int i;
@@ -1431,11 +1532,10 @@ bool radeon_connector_encoder_is_hbr2(struct drm_connector *connector)
 		if (connector->encoder_ids[i] == 0)
 			break;
 
-		obj = drm_mode_object_find(connector->dev, connector->encoder_ids[i], DRM_MODE_OBJECT_ENCODER);
-		if (!obj)
+		encoder = drm_encoder_find(connector->dev, connector->encoder_ids[i]);
+		if (!encoder)
 			continue;
 
-		encoder = obj_to_encoder(obj);
 		radeon_encoder = to_radeon_encoder(encoder);
 		if (radeon_encoder->caps & ATOM_ENCODER_CAP_RECORD_HBR2)
 			found = true;
@@ -1478,10 +1578,7 @@ radeon_dp_detect(struct drm_connector *connector, bool force)
 		goto out;
 	}
 
-	if (radeon_connector->edid) {
-		kfree(radeon_connector->edid);
-		radeon_connector->edid = NULL;
-	}
+	radeon_connector_free_edid(connector);
 
 	if ((connector->connector_type == DRM_MODE_CONNECTOR_eDP) ||
 	    (connector->connector_type == DRM_MODE_CONNECTOR_LVDS)) {
@@ -1587,7 +1684,7 @@ static int radeon_dp_mode_valid(struct drm_connector *connector,
 		    (radeon_dig_connector->dp_sink_type == CONNECTOR_OBJECT_ID_eDP)) {
 			return radeon_dp_mode_valid_helper(connector, mode);
 		} else {
-			if (ASIC_IS_DCE6(rdev) && drm_detect_hdmi_monitor(radeon_connector->edid)) {
+			if (ASIC_IS_DCE6(rdev) && drm_detect_hdmi_monitor(radeon_connector_edid(connector))) {
 				/* HDMI 1.3+ supports max clock of 340 Mhz */
 				if (mode->clock > 340000)
 					return MODE_CLOCK_HIGH;
@@ -1747,6 +1844,9 @@ radeon_add_atom_connector(struct drm_device *dev,
 			drm_object_attach_property(&radeon_connector->base.base,
 						      rdev->mode_info.load_detect_property,
 						      1);
+			drm_object_attach_property(&radeon_connector->base.base,
+						   dev->mode_config.scaling_mode_property,
+						   DRM_MODE_SCALE_NONE);
 			break;
 		case DRM_MODE_CONNECTOR_DVII:
 		case DRM_MODE_CONNECTOR_DVID:
@@ -1768,6 +1868,10 @@ radeon_add_atom_connector(struct drm_device *dev,
 						      0);
 
 			drm_object_attach_property(&radeon_connector->base.base,
+						      dev->mode_config.scaling_mode_property,
+						      DRM_MODE_SCALE_NONE);
+
+			drm_object_attach_property(&radeon_connector->base.base,
 						   rdev->mode_info.dither_property,
 						   RADEON_FMT_DITHER_DISABLE);
 
@@ -1817,6 +1921,10 @@ radeon_add_atom_connector(struct drm_device *dev,
 			drm_object_attach_property(&radeon_connector->base.base,
 						      rdev->mode_info.load_detect_property,
 						      1);
+			if (ASIC_IS_AVIVO(rdev))
+				drm_object_attach_property(&radeon_connector->base.base,
+							   dev->mode_config.scaling_mode_property,
+							   DRM_MODE_SCALE_NONE);
 			/* no HPD on analog connectors */
 			radeon_connector->hpd.hpd = RADEON_HPD_NONE;
 			connector->polled = DRM_CONNECTOR_POLL_CONNECT;
@@ -1835,6 +1943,10 @@ radeon_add_atom_connector(struct drm_device *dev,
 			drm_object_attach_property(&radeon_connector->base.base,
 						      rdev->mode_info.load_detect_property,
 						      1);
+			if (ASIC_IS_AVIVO(rdev))
+				drm_object_attach_property(&radeon_connector->base.base,
+							   dev->mode_config.scaling_mode_property,
+							   DRM_MODE_SCALE_NONE);
 			/* no HPD on analog connectors */
 			radeon_connector->hpd.hpd = RADEON_HPD_NONE;
 			connector->interlace_allowed = true;
@@ -1868,17 +1980,18 @@ radeon_add_atom_connector(struct drm_device *dev,
 				drm_object_attach_property(&radeon_connector->base.base,
 							      rdev->mode_info.underscan_vborder_property,
 							      0);
+				drm_object_attach_property(&radeon_connector->base.base,
+							   rdev->mode_info.dither_property,
+							   RADEON_FMT_DITHER_DISABLE);
+				drm_object_attach_property(&radeon_connector->base.base,
+							   dev->mode_config.scaling_mode_property,
+							   DRM_MODE_SCALE_NONE);
 			}
 			if (ASIC_IS_DCE2(rdev) && (radeon_audio != 0)) {
 				drm_object_attach_property(&radeon_connector->base.base,
 							   rdev->mode_info.audio_property,
 							   RADEON_AUDIO_AUTO);
 			}
-			if (ASIC_IS_AVIVO(rdev)) {
-				drm_object_attach_property(&radeon_connector->base.base,
-							   rdev->mode_info.dither_property,
-							   RADEON_FMT_DITHER_DISABLE);
-			}
 			if (connector_type == DRM_MODE_CONNECTOR_DVII) {
 				radeon_connector->dac_load_detect = true;
 				drm_object_attach_property(&radeon_connector->base.base,
@@ -1918,17 +2031,18 @@ radeon_add_atom_connector(struct drm_device *dev,
 				drm_object_attach_property(&radeon_connector->base.base,
 							      rdev->mode_info.underscan_vborder_property,
 							      0);
+				drm_object_attach_property(&radeon_connector->base.base,
+							   rdev->mode_info.dither_property,
+							   RADEON_FMT_DITHER_DISABLE);
+				drm_object_attach_property(&radeon_connector->base.base,
+							   dev->mode_config.scaling_mode_property,
+							   DRM_MODE_SCALE_NONE);
 			}
 			if (ASIC_IS_DCE2(rdev) && (radeon_audio != 0)) {
 				drm_object_attach_property(&radeon_connector->base.base,
 							   rdev->mode_info.audio_property,
 							   RADEON_AUDIO_AUTO);
 			}
-			if (ASIC_IS_AVIVO(rdev)) {
-				drm_object_attach_property(&radeon_connector->base.base,
-							   rdev->mode_info.dither_property,
-							   RADEON_FMT_DITHER_DISABLE);
-			}
 			subpixel_order = SubPixelHorizontalRGB;
 			connector->interlace_allowed = true;
 			if (connector_type == DRM_MODE_CONNECTOR_HDMIB)
@@ -1965,18 +2079,18 @@ radeon_add_atom_connector(struct drm_device *dev,
 				drm_object_attach_property(&radeon_connector->base.base,
 							      rdev->mode_info.underscan_vborder_property,
 							      0);
+				drm_object_attach_property(&radeon_connector->base.base,
+							   rdev->mode_info.dither_property,
+							   RADEON_FMT_DITHER_DISABLE);
+				drm_object_attach_property(&radeon_connector->base.base,
+							   dev->mode_config.scaling_mode_property,
+							   DRM_MODE_SCALE_NONE);
 			}
 			if (ASIC_IS_DCE2(rdev) && (radeon_audio != 0)) {
 				drm_object_attach_property(&radeon_connector->base.base,
 							   rdev->mode_info.audio_property,
 							   RADEON_AUDIO_AUTO);
 			}
-			if (ASIC_IS_AVIVO(rdev)) {
-				drm_object_attach_property(&radeon_connector->base.base,
-							   rdev->mode_info.dither_property,
-							   RADEON_FMT_DITHER_DISABLE);
-
-			}
 			connector->interlace_allowed = true;
 			/* in theory with a DP to VGA converter... */
 			connector->doublescan_allowed = false;
@@ -2050,7 +2164,7 @@ radeon_add_atom_connector(struct drm_device *dev,
 		connector->polled = DRM_CONNECTOR_POLL_HPD;
 
 	connector->display_info.subpixel_order = subpixel_order;
-	drm_sysfs_connector_add(connector);
+	drm_connector_register(connector);
 
 	if (has_aux)
 		radeon_dp_aux_init(radeon_connector);
@@ -2211,5 +2325,5 @@ radeon_add_legacy_connector(struct drm_device *dev,
 	} else
 		connector->polled = DRM_CONNECTOR_POLL_HPD;
 	connector->display_info.subpixel_order = subpixel_order;
-	drm_sysfs_connector_add(connector);
+	drm_connector_register(connector);
 }
diff --git a/drivers/gpu/drm/radeon/radeon_cs.c b/drivers/gpu/drm/radeon/radeon_cs.c
index 71a143461478..83f382e8e40e 100644
--- a/drivers/gpu/drm/radeon/radeon_cs.c
+++ b/drivers/gpu/drm/radeon/radeon_cs.c
@@ -132,7 +132,8 @@ static int radeon_cs_parser_relocs(struct radeon_cs_parser *p)
 		 * the buffers used for read only, which doubles the range
 		 * to 0 to 31. 32 is reserved for the kernel driver.
 		 */
-		priority = (r->flags & 0xf) * 2 + !!r->write_domain;
+		priority = (r->flags & RADEON_RELOC_PRIO_MASK) * 2
+			   + !!r->write_domain;
 
 		/* the first reloc of an UVD job is the msg and that must be in
 		   VRAM, also but everything into VRAM on AGP cards to avoid
@@ -450,7 +451,7 @@ static int radeon_cs_ib_chunk(struct radeon_device *rdev,
 		radeon_vce_note_usage(rdev);
 
 	radeon_cs_sync_rings(parser);
-	r = radeon_ib_schedule(rdev, &parser->ib, NULL);
+	r = radeon_ib_schedule(rdev, &parser->ib, NULL, true);
 	if (r) {
 		DRM_ERROR("Failed to schedule IB !\n");
 	}
@@ -461,13 +462,23 @@ static int radeon_bo_vm_update_pte(struct radeon_cs_parser *p,
 				   struct radeon_vm *vm)
 {
 	struct radeon_device *rdev = p->rdev;
+	struct radeon_bo_va *bo_va;
 	int i, r;
 
 	r = radeon_vm_update_page_directory(rdev, vm);
 	if (r)
 		return r;
 
-	r = radeon_vm_bo_update(rdev, vm, rdev->ring_tmp_bo.bo,
+	r = radeon_vm_clear_freed(rdev, vm);
+	if (r)
+		return r;
+
+	if (vm->ib_bo_va == NULL) {
+		DRM_ERROR("Tmp BO not in VM!\n");
+		return -EINVAL;
+	}
+
+	r = radeon_vm_bo_update(rdev, vm->ib_bo_va,
 				&rdev->ring_tmp_bo.bo->tbo.mem);
 	if (r)
 		return r;
@@ -480,11 +491,18 @@ static int radeon_bo_vm_update_pte(struct radeon_cs_parser *p,
 			continue;
 
 		bo = p->relocs[i].robj;
-		r = radeon_vm_bo_update(rdev, vm, bo, &bo->tbo.mem);
+		bo_va = radeon_vm_bo_find(vm, bo);
+		if (bo_va == NULL) {
+			dev_err(rdev->dev, "bo %p not in vm %p\n", bo, vm);
+			return -EINVAL;
+		}
+
+		r = radeon_vm_bo_update(rdev, bo_va, &bo->tbo.mem);
 		if (r)
 			return r;
 	}
-	return 0;
+
+	return radeon_vm_clear_invalids(rdev, vm);
 }
 
 static int radeon_cs_ib_vm_chunk(struct radeon_device *rdev,
@@ -524,9 +542,9 @@ static int radeon_cs_ib_vm_chunk(struct radeon_device *rdev,
 
 	if ((rdev->family >= CHIP_TAHITI) &&
 	    (parser->chunk_const_ib_idx != -1)) {
-		r = radeon_ib_schedule(rdev, &parser->ib, &parser->const_ib);
+		r = radeon_ib_schedule(rdev, &parser->ib, &parser->const_ib, true);
 	} else {
-		r = radeon_ib_schedule(rdev, &parser->ib, NULL);
+		r = radeon_ib_schedule(rdev, &parser->ib, NULL, true);
 	}
 
 out:
diff --git a/drivers/gpu/drm/radeon/radeon_device.c b/drivers/gpu/drm/radeon/radeon_device.c
index 03686fab842d..6a219bcee66d 100644
--- a/drivers/gpu/drm/radeon/radeon_device.c
+++ b/drivers/gpu/drm/radeon/radeon_device.c
@@ -103,6 +103,31 @@ static const char radeon_family_name[][16] = {
 	"LAST",
 };
 
+#define RADEON_PX_QUIRK_DISABLE_PX  (1 << 0)
+#define RADEON_PX_QUIRK_LONG_WAKEUP (1 << 1)
+
+struct radeon_px_quirk {
+	u32 chip_vendor;
+	u32 chip_device;
+	u32 subsys_vendor;
+	u32 subsys_device;
+	u32 px_quirk_flags;
+};
+
+static struct radeon_px_quirk radeon_px_quirk_list[] = {
+	/* Acer aspire 5560g (CPU: AMD A4-3305M; GPU: AMD Radeon HD 6480g + 7470m)
+	 * https://bugzilla.kernel.org/show_bug.cgi?id=74551
+	 */
+	{ PCI_VENDOR_ID_ATI, 0x6760, 0x1025, 0x0672, RADEON_PX_QUIRK_DISABLE_PX },
+	/* Asus K73TA laptop with AMD A6-3400M APU and Radeon 6550 GPU
+	 * https://bugzilla.kernel.org/show_bug.cgi?id=51381
+	 */
+	{ PCI_VENDOR_ID_ATI, 0x6741, 0x1043, 0x108c, RADEON_PX_QUIRK_DISABLE_PX },
+	/* macbook pro 8.2 */
+	{ PCI_VENDOR_ID_ATI, 0x6741, PCI_VENDOR_ID_APPLE, 0x00e2, RADEON_PX_QUIRK_LONG_WAKEUP },
+	{ 0, 0, 0, 0, 0 },
+};
+
 bool radeon_is_px(struct drm_device *dev)
 {
 	struct radeon_device *rdev = dev->dev_private;
@@ -112,6 +137,26 @@ bool radeon_is_px(struct drm_device *dev)
 	return false;
 }
 
+static void radeon_device_handle_px_quirks(struct radeon_device *rdev)
+{
+	struct radeon_px_quirk *p = radeon_px_quirk_list;
+
+	/* Apply PX quirks */
+	while (p && p->chip_device != 0) {
+		if (rdev->pdev->vendor == p->chip_vendor &&
+		    rdev->pdev->device == p->chip_device &&
+		    rdev->pdev->subsystem_vendor == p->subsys_vendor &&
+		    rdev->pdev->subsystem_device == p->subsys_device) {
+			rdev->px_quirk_flags = p->px_quirk_flags;
+			break;
+		}
+		++p;
+	}
+
+	if (rdev->px_quirk_flags & RADEON_PX_QUIRK_DISABLE_PX)
+		rdev->flags &= ~RADEON_IS_PX;
+}
+
 /**
  * radeon_program_register_sequence - program an array of registers.
  *
@@ -385,7 +430,8 @@ int radeon_wb_init(struct radeon_device *rdev)
 
 	if (rdev->wb.wb_obj == NULL) {
 		r = radeon_bo_create(rdev, RADEON_GPU_PAGE_SIZE, PAGE_SIZE, true,
-				     RADEON_GEM_DOMAIN_GTT, NULL, &rdev->wb.wb_obj);
+				     RADEON_GEM_DOMAIN_GTT, 0, NULL,
+				     &rdev->wb.wb_obj);
 		if (r) {
 			dev_warn(rdev->dev, "(%d) create WB bo failed\n", r);
 			return r;
@@ -1056,61 +1102,54 @@ static void radeon_check_arguments(struct radeon_device *rdev)
 	if (!radeon_check_pot_argument(radeon_vm_size)) {
 		dev_warn(rdev->dev, "VM size (%d) must be a power of 2\n",
 			 radeon_vm_size);
-		radeon_vm_size = 4096;
+		radeon_vm_size = 4;
 	}
 
-	if (radeon_vm_size < 4) {
-		dev_warn(rdev->dev, "VM size (%d) to small, min is 4MB\n",
+	if (radeon_vm_size < 1) {
+		dev_warn(rdev->dev, "VM size (%d) to small, min is 1GB\n",
 			 radeon_vm_size);
-		radeon_vm_size = 4096;
+		radeon_vm_size = 4;
 	}
 
        /*
         * Max GPUVM size for Cayman, SI and CI are 40 bits.
         */
-	if (radeon_vm_size > 1024*1024) {
-		dev_warn(rdev->dev, "VM size (%d) to large, max is 1TB\n",
+	if (radeon_vm_size > 1024) {
+		dev_warn(rdev->dev, "VM size (%d) too large, max is 1TB\n",
 			 radeon_vm_size);
-		radeon_vm_size = 4096;
+		radeon_vm_size = 4;
 	}
 
 	/* defines number of bits in page table versus page directory,
 	 * a page is 4KB so we have 12 bits offset, minimum 9 bits in the
 	 * page table and the remaining bits are in the page directory */
-	if (radeon_vm_block_size < 9) {
-		dev_warn(rdev->dev, "VM page table size (%d) to small\n",
+	if (radeon_vm_block_size == -1) {
+
+		/* Total bits covered by PD + PTs */
+		unsigned bits = ilog2(radeon_vm_size) + 17;
+
+		/* Make sure the PD is 4K in size up to 8GB address space.
+		   Above that split equal between PD and PTs */
+		if (radeon_vm_size <= 8)
+			radeon_vm_block_size = bits - 9;
+		else
+			radeon_vm_block_size = (bits + 3) / 2;
+
+	} else if (radeon_vm_block_size < 9) {
+		dev_warn(rdev->dev, "VM page table size (%d) too small\n",
 			 radeon_vm_block_size);
 		radeon_vm_block_size = 9;
 	}
 
 	if (radeon_vm_block_size > 24 ||
-	    radeon_vm_size < (1ull << radeon_vm_block_size)) {
-		dev_warn(rdev->dev, "VM page table size (%d) to large\n",
+	    (radeon_vm_size * 1024) < (1ull << radeon_vm_block_size)) {
+		dev_warn(rdev->dev, "VM page table size (%d) too large\n",
 			 radeon_vm_block_size);
 		radeon_vm_block_size = 9;
 	}
 }
 
 /**
- * radeon_switcheroo_quirk_long_wakeup - return true if longer d3 delay is
- * needed for waking up.
- *
- * @pdev: pci dev pointer
- */
-static bool radeon_switcheroo_quirk_long_wakeup(struct pci_dev *pdev)
-{
-
-	/* 6600m in a macbook pro */
-	if (pdev->subsystem_vendor == PCI_VENDOR_ID_APPLE &&
-	    pdev->subsystem_device == 0x00e2) {
-		printk(KERN_INFO "radeon: quirking longer d3 wakeup delay\n");
-		return true;
-	}
-
-	return false;
-}
-
-/**
  * radeon_switcheroo_set_state - set switcheroo state
  *
  * @pdev: pci dev pointer
@@ -1122,6 +1161,7 @@ static bool radeon_switcheroo_quirk_long_wakeup(struct pci_dev *pdev)
 static void radeon_switcheroo_set_state(struct pci_dev *pdev, enum vga_switcheroo_state state)
 {
 	struct drm_device *dev = pci_get_drvdata(pdev);
+	struct radeon_device *rdev = dev->dev_private;
 
 	if (radeon_is_px(dev) && state == VGA_SWITCHEROO_OFF)
 		return;
@@ -1133,7 +1173,7 @@ static void radeon_switcheroo_set_state(struct pci_dev *pdev, enum vga_switchero
 		/* don't suspend or resume card normally */
 		dev->switch_power_state = DRM_SWITCH_POWER_CHANGING;
 
-		if (d3_delay < 20 && radeon_switcheroo_quirk_long_wakeup(pdev))
+		if (d3_delay < 20 && (rdev->px_quirk_flags & RADEON_PX_QUIRK_LONG_WAKEUP))
 			dev->pdev->d3_delay = 20;
 
 		radeon_resume_kms(dev, true, true);
@@ -1238,7 +1278,7 @@ int radeon_device_init(struct radeon_device *rdev,
 	/* Adjust VM size here.
 	 * Max GPUVM size for cayman+ is 40 bits.
 	 */
-	rdev->vm_manager.max_pfn = radeon_vm_size << 8;
+	rdev->vm_manager.max_pfn = radeon_vm_size << 18;
 
 	/* Set asic functions */
 	r = radeon_asic_init(rdev);
@@ -1337,6 +1377,9 @@ int radeon_device_init(struct radeon_device *rdev,
 	if (rdev->rio_mem == NULL)
 		DRM_ERROR("Unable to find PCI I/O BAR\n");
 
+	if (rdev->flags & RADEON_IS_PX)
+		radeon_device_handle_px_quirks(rdev);
+
 	/* if we have > 1 VGA cards, then disable the radeon VGA resources */
 	/* this will fail for cards that aren't VGA class devices, just
 	 * ignore it */
@@ -1637,8 +1680,8 @@ 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);
+	radeon_hpd_fini(rdev);
 
 	for (i = 0; i < RADEON_NUM_RINGS; ++i) {
 		ring_sizes[i] = radeon_ring_backup(rdev, &rdev->ring[i],
@@ -1683,9 +1726,39 @@ retry:
 		}
 	}
 
-	radeon_pm_resume(rdev);
+	if ((rdev->pm.pm_method == PM_METHOD_DPM) && rdev->pm.dpm_enabled) {
+		/* do dpm late init */
+		r = radeon_pm_late_init(rdev);
+		if (r) {
+			rdev->pm.dpm_enabled = false;
+			DRM_ERROR("radeon_pm_late_init failed, disabling dpm\n");
+		}
+	} else {
+		/* resume old pm late */
+		radeon_pm_resume(rdev);
+	}
+
+	/* init dig PHYs, disp eng pll */
+	if (rdev->is_atom_bios) {
+		radeon_atom_encoder_init(rdev);
+		radeon_atom_disp_eng_pll_init(rdev);
+		/* turn on the BL */
+		if (rdev->mode_info.bl_encoder) {
+			u8 bl_level = radeon_get_backlight_level(rdev,
+								 rdev->mode_info.bl_encoder);
+			radeon_set_backlight_level(rdev, rdev->mode_info.bl_encoder,
+						   bl_level);
+		}
+	}
+	/* reset hpd state */
+	radeon_hpd_init(rdev);
+
 	drm_helper_resume_force_mode(rdev->ddev);
 
+	/* set the power state here in case we are a PX system or headless */
+	if ((rdev->pm.pm_method == PM_METHOD_DPM) && rdev->pm.dpm_enabled)
+		radeon_pm_compute_clocks(rdev);
+
 	ttm_bo_unlock_delayed_workqueue(&rdev->mman.bdev, resched);
 	if (r) {
 		/* bad news, how to tell it to userspace ? */
diff --git a/drivers/gpu/drm/radeon/radeon_display.c b/drivers/gpu/drm/radeon/radeon_display.c
index bf25061c8ac4..3fdf87318069 100644
--- a/drivers/gpu/drm/radeon/radeon_display.c
+++ b/drivers/gpu/drm/radeon/radeon_display.c
@@ -293,6 +293,18 @@ void radeon_crtc_handle_vblank(struct radeon_device *rdev, int crtc_id)
 	if (radeon_crtc == NULL)
 		return;
 
+	/* Skip the pageflip completion check below (based on polling) on
+	 * asics which reliably support hw pageflip completion irqs. pflip
+	 * irqs are a reliable and race-free method of handling pageflip
+	 * completion detection. A use_pflipirq module parameter < 2 allows
+	 * to override this in case of asics with faulty pflip irqs.
+	 * A module parameter of 0 would only use this polling based path,
+	 * a parameter of 1 would use pflip irq only as a backup to this
+	 * path, as in Linux 3.16.
+	 */
+	if ((radeon_use_pflipirq == 2) && ASIC_IS_DCE4(rdev))
+		return;
+
 	spin_lock_irqsave(&rdev->ddev->event_lock, flags);
 	if (radeon_crtc->flip_status != RADEON_FLIP_SUBMITTED) {
 		DRM_DEBUG_DRIVER("radeon_crtc->flip_status = %d != "
@@ -823,64 +835,6 @@ static bool radeon_setup_enc_conn(struct drm_device *dev)
 	return ret;
 }
 
-int radeon_ddc_get_modes(struct radeon_connector *radeon_connector)
-{
-	struct drm_device *dev = radeon_connector->base.dev;
-	struct radeon_device *rdev = dev->dev_private;
-	int ret = 0;
-
-	/* don't leak the edid if we already fetched it in detect() */
-	if (radeon_connector->edid)
-		goto got_edid;
-
-	/* on hw with routers, select right port */
-	if (radeon_connector->router.ddc_valid)
-		radeon_router_select_ddc_port(radeon_connector);
-
-	if (radeon_connector_encoder_get_dp_bridge_encoder_id(&radeon_connector->base) !=
-	    ENCODER_OBJECT_ID_NONE) {
-		if (radeon_connector->ddc_bus->has_aux)
-			radeon_connector->edid = drm_get_edid(&radeon_connector->base,
-							      &radeon_connector->ddc_bus->aux.ddc);
-	} else if ((radeon_connector->base.connector_type == DRM_MODE_CONNECTOR_DisplayPort) ||
-		   (radeon_connector->base.connector_type == DRM_MODE_CONNECTOR_eDP)) {
-		struct radeon_connector_atom_dig *dig = radeon_connector->con_priv;
-
-		if ((dig->dp_sink_type == CONNECTOR_OBJECT_ID_DISPLAYPORT ||
-		     dig->dp_sink_type == CONNECTOR_OBJECT_ID_eDP) &&
-		    radeon_connector->ddc_bus->has_aux)
-			radeon_connector->edid = drm_get_edid(&radeon_connector->base,
-							      &radeon_connector->ddc_bus->aux.ddc);
-		else if (radeon_connector->ddc_bus && !radeon_connector->edid)
-			radeon_connector->edid = drm_get_edid(&radeon_connector->base,
-							      &radeon_connector->ddc_bus->adapter);
-	} else {
-		if (radeon_connector->ddc_bus && !radeon_connector->edid)
-			radeon_connector->edid = drm_get_edid(&radeon_connector->base,
-							      &radeon_connector->ddc_bus->adapter);
-	}
-
-	if (!radeon_connector->edid) {
-		if (rdev->is_atom_bios) {
-			/* some laptops provide a hardcoded edid in rom for LCDs */
-			if (((radeon_connector->base.connector_type == DRM_MODE_CONNECTOR_LVDS) ||
-			     (radeon_connector->base.connector_type == DRM_MODE_CONNECTOR_eDP)))
-				radeon_connector->edid = radeon_bios_get_hardcoded_edid(rdev);
-		} else
-			/* some servers provide a hardcoded edid in rom for KVMs */
-			radeon_connector->edid = radeon_bios_get_hardcoded_edid(rdev);
-	}
-	if (radeon_connector->edid) {
-got_edid:
-		drm_mode_connector_update_edid_property(&radeon_connector->base, radeon_connector->edid);
-		ret = drm_add_edid_modes(&radeon_connector->base, radeon_connector->edid);
-		drm_edid_to_eld(&radeon_connector->base, radeon_connector->edid);
-		return ret;
-	}
-	drm_mode_connector_update_edid_property(&radeon_connector->base, NULL);
-	return 0;
-}
-
 /* avivo */
 
 /**
@@ -1749,7 +1703,7 @@ bool radeon_crtc_scaling_mode_fixup(struct drm_crtc *crtc,
 			    (!(mode->flags & DRM_MODE_FLAG_INTERLACE)) &&
 			    ((radeon_encoder->underscan_type == UNDERSCAN_ON) ||
 			     ((radeon_encoder->underscan_type == UNDERSCAN_AUTO) &&
-			      drm_detect_hdmi_monitor(radeon_connector->edid) &&
+			      drm_detect_hdmi_monitor(radeon_connector_edid(connector)) &&
 			      is_hdtv_mode(mode)))) {
 				if (radeon_encoder->underscan_hborder != 0)
 					radeon_crtc->h_border = radeon_encoder->underscan_hborder;
diff --git a/drivers/gpu/drm/radeon/radeon_drv.c b/drivers/gpu/drm/radeon/radeon_drv.c
index cb1421369e3a..8df888908833 100644
--- a/drivers/gpu/drm/radeon/radeon_drv.c
+++ b/drivers/gpu/drm/radeon/radeon_drv.c
@@ -82,9 +82,11 @@
  *   2.38.0 - RADEON_GEM_OP (GET_INITIAL_DOMAIN, SET_INITIAL_DOMAIN),
  *            CIK: 1D and linear tiling modes contain valid PIPE_CONFIG
  *   2.39.0 - Add INFO query for number of active CUs
+ *   2.40.0 - Add RADEON_GEM_GTT_WC/UC, flush HDP cache before submitting
+ *            CS to GPU
  */
 #define KMS_DRIVER_MAJOR	2
-#define KMS_DRIVER_MINOR	39
+#define KMS_DRIVER_MINOR	40
 #define KMS_DRIVER_PATCHLEVEL	0
 int radeon_driver_load_kms(struct drm_device *dev, unsigned long flags);
 int radeon_driver_unload_kms(struct drm_device *dev);
@@ -132,6 +134,7 @@ struct drm_gem_object *radeon_gem_prime_import_sg_table(struct drm_device *dev,
 							struct sg_table *sg);
 int radeon_gem_prime_pin(struct drm_gem_object *obj);
 void radeon_gem_prime_unpin(struct drm_gem_object *obj);
+struct reservation_object *radeon_gem_prime_res_obj(struct drm_gem_object *);
 void *radeon_gem_prime_vmap(struct drm_gem_object *obj);
 void radeon_gem_prime_vunmap(struct drm_gem_object *obj, void *vaddr);
 extern long radeon_kms_compat_ioctl(struct file *filp, unsigned int cmd,
@@ -173,9 +176,11 @@ int radeon_dpm = -1;
 int radeon_aspm = -1;
 int radeon_runtime_pm = -1;
 int radeon_hard_reset = 0;
-int radeon_vm_size = 4096;
-int radeon_vm_block_size = 9;
+int radeon_vm_size = 8;
+int radeon_vm_block_size = -1;
 int radeon_deep_color = 0;
+int radeon_use_pflipirq = 2;
+int radeon_bapm = -1;
 
 MODULE_PARM_DESC(no_wb, "Disable AGP writeback for scratch registers");
 module_param_named(no_wb, radeon_no_wb, int, 0444);
@@ -243,15 +248,21 @@ module_param_named(runpm, radeon_runtime_pm, int, 0444);
 MODULE_PARM_DESC(hard_reset, "PCI config reset (1 = force enable, 0 = disable (default))");
 module_param_named(hard_reset, radeon_hard_reset, int, 0444);
 
-MODULE_PARM_DESC(vm_size, "VM address space size in megabytes (default 4GB)");
+MODULE_PARM_DESC(vm_size, "VM address space size in gigabytes (default 4GB)");
 module_param_named(vm_size, radeon_vm_size, int, 0444);
 
-MODULE_PARM_DESC(vm_block_size, "VM page table size in bits (default 9)");
+MODULE_PARM_DESC(vm_block_size, "VM page table size in bits (default depending on vm_size)");
 module_param_named(vm_block_size, radeon_vm_block_size, int, 0444);
 
 MODULE_PARM_DESC(deep_color, "Deep Color support (1 = enable, 0 = disable (default))");
 module_param_named(deep_color, radeon_deep_color, int, 0444);
 
+MODULE_PARM_DESC(use_pflipirq, "Pflip irqs for pageflip completion (0 = disable, 1 = as fallback, 2 = exclusive (default))");
+module_param_named(use_pflipirq, radeon_use_pflipirq, int, 0444);
+
+MODULE_PARM_DESC(bapm, "BAPM support (1 = enable, 0 = disable, -1 = auto)");
+module_param_named(bapm, radeon_bapm, int, 0444);
+
 static struct pci_device_id pciidlist[] = {
 	radeon_PCI_IDS
 };
@@ -566,6 +577,7 @@ static struct drm_driver kms_driver = {
 	.gem_prime_import = drm_gem_prime_import,
 	.gem_prime_pin = radeon_gem_prime_pin,
 	.gem_prime_unpin = radeon_gem_prime_unpin,
+	.gem_prime_res_obj = radeon_gem_prime_res_obj,
 	.gem_prime_get_sg_table = radeon_gem_prime_get_sg_table,
 	.gem_prime_import_sg_table = radeon_gem_prime_import_sg_table,
 	.gem_prime_vmap = radeon_gem_prime_vmap,
diff --git a/drivers/gpu/drm/radeon/radeon_encoders.c b/drivers/gpu/drm/radeon/radeon_encoders.c
index bd4959ca23aa..3c2094c25b53 100644
--- a/drivers/gpu/drm/radeon/radeon_encoders.c
+++ b/drivers/gpu/drm/radeon/radeon_encoders.c
@@ -343,7 +343,7 @@ bool radeon_dig_monitor_is_duallink(struct drm_encoder *encoder,
 	case DRM_MODE_CONNECTOR_HDMIB:
 		if (radeon_connector->use_digital) {
 			/* HDMI 1.3 supports up to 340 Mhz over single link */
-			if (ASIC_IS_DCE6(rdev) && drm_detect_hdmi_monitor(radeon_connector->edid)) {
+			if (ASIC_IS_DCE6(rdev) && drm_detect_hdmi_monitor(radeon_connector_edid(connector))) {
 				if (pixel_clock > 340000)
 					return true;
 				else
@@ -365,7 +365,7 @@ bool radeon_dig_monitor_is_duallink(struct drm_encoder *encoder,
 			return false;
 		else {
 			/* HDMI 1.3 supports up to 340 Mhz over single link */
-			if (ASIC_IS_DCE6(rdev) && drm_detect_hdmi_monitor(radeon_connector->edid)) {
+			if (ASIC_IS_DCE6(rdev) && drm_detect_hdmi_monitor(radeon_connector_edid(connector))) {
 				if (pixel_clock > 340000)
 					return true;
 				else
diff --git a/drivers/gpu/drm/radeon/radeon_fb.c b/drivers/gpu/drm/radeon/radeon_fb.c
index 665ced3b7313..94b0f2aa3d7c 100644
--- a/drivers/gpu/drm/radeon/radeon_fb.c
+++ b/drivers/gpu/drm/radeon/radeon_fb.c
@@ -127,8 +127,7 @@ static int radeonfb_create_pinned_object(struct radeon_fbdev *rfbdev,
 	aligned_size = ALIGN(size, PAGE_SIZE);
 	ret = radeon_gem_object_create(rdev, aligned_size, 0,
 				       RADEON_GEM_DOMAIN_VRAM,
-				       false, true,
-				       &gobj);
+				       0, true, &gobj);
 	if (ret) {
 		printk(KERN_ERR "failed to allocate framebuffer (%d)\n",
 		       aligned_size);
@@ -331,7 +330,7 @@ static int radeon_fbdev_destroy(struct drm_device *dev, struct radeon_fbdev *rfb
 	return 0;
 }
 
-static struct drm_fb_helper_funcs radeon_fb_helper_funcs = {
+static const struct drm_fb_helper_funcs radeon_fb_helper_funcs = {
 	.gamma_set = radeon_crtc_fb_gamma_set,
 	.gamma_get = radeon_crtc_fb_gamma_get,
 	.fb_probe = radeonfb_create,
@@ -353,7 +352,9 @@ int radeon_fbdev_init(struct radeon_device *rdev)
 
 	rfbdev->rdev = rdev;
 	rdev->mode_info.rfbdev = rfbdev;
-	rfbdev->helper.funcs = &radeon_fb_helper_funcs;
+
+	drm_fb_helper_prepare(rdev->ddev, &rfbdev->helper,
+			      &radeon_fb_helper_funcs);
 
 	ret = drm_fb_helper_init(rdev->ddev, &rfbdev->helper,
 				 rdev->num_crtc,
diff --git a/drivers/gpu/drm/radeon/radeon_gart.c b/drivers/gpu/drm/radeon/radeon_gart.c
index 2e723651069b..a053a0779aac 100644
--- a/drivers/gpu/drm/radeon/radeon_gart.c
+++ b/drivers/gpu/drm/radeon/radeon_gart.c
@@ -128,7 +128,7 @@ int radeon_gart_table_vram_alloc(struct radeon_device *rdev)
 	if (rdev->gart.robj == NULL) {
 		r = radeon_bo_create(rdev, rdev->gart.table_size,
 				     PAGE_SIZE, true, RADEON_GEM_DOMAIN_VRAM,
-				     NULL, &rdev->gart.robj);
+				     0, NULL, &rdev->gart.robj);
 		if (r) {
 			return r;
 		}
@@ -243,7 +243,8 @@ void radeon_gart_unbind(struct radeon_device *rdev, unsigned offset,
 			page_base = rdev->gart.pages_addr[p];
 			for (j = 0; j < (PAGE_SIZE / RADEON_GPU_PAGE_SIZE); j++, t++) {
 				if (rdev->gart.ptr) {
-					radeon_gart_set_page(rdev, t, page_base);
+					radeon_gart_set_page(rdev, t, page_base,
+							     RADEON_GART_PAGE_DUMMY);
 				}
 				page_base += RADEON_GPU_PAGE_SIZE;
 			}
@@ -261,13 +262,15 @@ void radeon_gart_unbind(struct radeon_device *rdev, unsigned offset,
  * @pages: number of pages to bind
  * @pagelist: pages to bind
  * @dma_addr: DMA addresses of pages
+ * @flags: RADEON_GART_PAGE_* flags
  *
  * Binds the requested pages to the gart page table
  * (all asics).
  * Returns 0 for success, -EINVAL for failure.
  */
 int radeon_gart_bind(struct radeon_device *rdev, unsigned offset,
-		     int pages, struct page **pagelist, dma_addr_t *dma_addr)
+		     int pages, struct page **pagelist, dma_addr_t *dma_addr,
+		     uint32_t flags)
 {
 	unsigned t;
 	unsigned p;
@@ -287,7 +290,7 @@ int radeon_gart_bind(struct radeon_device *rdev, unsigned offset,
 		if (rdev->gart.ptr) {
 			page_base = rdev->gart.pages_addr[p];
 			for (j = 0; j < (PAGE_SIZE / RADEON_GPU_PAGE_SIZE); j++, t++) {
-				radeon_gart_set_page(rdev, t, page_base);
+				radeon_gart_set_page(rdev, t, page_base, flags);
 				page_base += RADEON_GPU_PAGE_SIZE;
 			}
 		}
@@ -298,33 +301,6 @@ int radeon_gart_bind(struct radeon_device *rdev, unsigned offset,
 }
 
 /**
- * radeon_gart_restore - bind all pages in the gart page table
- *
- * @rdev: radeon_device pointer
- *
- * Binds all pages in the gart page table (all asics).
- * Used to rebuild the gart table on device startup or resume.
- */
-void radeon_gart_restore(struct radeon_device *rdev)
-{
-	int i, j, t;
-	u64 page_base;
-
-	if (!rdev->gart.ptr) {
-		return;
-	}
-	for (i = 0, t = 0; i < rdev->gart.num_cpu_pages; i++) {
-		page_base = rdev->gart.pages_addr[i];
-		for (j = 0; j < (PAGE_SIZE / RADEON_GPU_PAGE_SIZE); j++, t++) {
-			radeon_gart_set_page(rdev, t, page_base);
-			page_base += RADEON_GPU_PAGE_SIZE;
-		}
-	}
-	mb();
-	radeon_gart_tlb_flush(rdev);
-}
-
-/**
  * radeon_gart_init - init the driver info for managing the gart
  *
  * @rdev: radeon_device pointer
diff --git a/drivers/gpu/drm/radeon/radeon_gem.c b/drivers/gpu/drm/radeon/radeon_gem.c
index d09650c1d720..bfd7e1b0ff3f 100644
--- a/drivers/gpu/drm/radeon/radeon_gem.c
+++ b/drivers/gpu/drm/radeon/radeon_gem.c
@@ -40,9 +40,9 @@ void radeon_gem_object_free(struct drm_gem_object *gobj)
 	}
 }
 
-int radeon_gem_object_create(struct radeon_device *rdev, int size,
+int radeon_gem_object_create(struct radeon_device *rdev, unsigned long size,
 				int alignment, int initial_domain,
-				bool discardable, bool kernel,
+				u32 flags, bool kernel,
 				struct drm_gem_object **obj)
 {
 	struct radeon_bo *robj;
@@ -55,23 +55,26 @@ int radeon_gem_object_create(struct radeon_device *rdev, int size,
 		alignment = PAGE_SIZE;
 	}
 
-	/* maximun bo size is the minimun btw visible vram and gtt size */
-	max_size = min(rdev->mc.visible_vram_size, rdev->mc.gtt_size);
+	/* Maximum bo size is the unpinned gtt size since we use the gtt to
+	 * handle vram to system pool migrations.
+	 */
+	max_size = rdev->mc.gtt_size - rdev->gart_pin_size;
 	if (size > max_size) {
-		printk(KERN_WARNING "%s:%d alloc size %dMb bigger than %ldMb limit\n",
-		       __func__, __LINE__, size >> 20, max_size >> 20);
+		DRM_DEBUG("Allocation size %ldMb bigger than %ldMb limit\n",
+			  size >> 20, max_size >> 20);
 		return -ENOMEM;
 	}
 
 retry:
-	r = radeon_bo_create(rdev, size, alignment, kernel, initial_domain, NULL, &robj);
+	r = radeon_bo_create(rdev, size, alignment, kernel, initial_domain,
+			     flags, NULL, &robj);
 	if (r) {
 		if (r != -ERESTARTSYS) {
 			if (initial_domain == RADEON_GEM_DOMAIN_VRAM) {
 				initial_domain |= RADEON_GEM_DOMAIN_GTT;
 				goto retry;
 			}
-			DRM_ERROR("Failed to allocate GEM object (%d, %d, %u, %d)\n",
+			DRM_ERROR("Failed to allocate GEM object (%ld, %d, %u, %d)\n",
 				  size, initial_domain, alignment, r);
 		}
 		return r;
@@ -208,18 +211,15 @@ int radeon_gem_info_ioctl(struct drm_device *dev, void *data,
 	struct radeon_device *rdev = dev->dev_private;
 	struct drm_radeon_gem_info *args = data;
 	struct ttm_mem_type_manager *man;
-	unsigned i;
 
 	man = &rdev->mman.bdev.man[TTM_PL_VRAM];
 
 	args->vram_size = rdev->mc.real_vram_size;
 	args->vram_visible = (u64)man->size << PAGE_SHIFT;
-	if (rdev->stollen_vga_memory)
-		args->vram_visible -= radeon_bo_size(rdev->stollen_vga_memory);
-	args->vram_visible -= radeon_fbdev_total_size(rdev);
-	args->gart_size = rdev->mc.gtt_size - 4096 - RADEON_IB_POOL_SIZE*64*1024;
-	for(i = 0; i < RADEON_NUM_RINGS; ++i)
-		args->gart_size -= rdev->ring[i].ring_size;
+	args->vram_visible -= rdev->vram_pin_size;
+	args->gart_size = rdev->mc.gtt_size;
+	args->gart_size -= rdev->gart_pin_size;
+
 	return 0;
 }
 
@@ -252,8 +252,8 @@ int radeon_gem_create_ioctl(struct drm_device *dev, void *data,
 	/* create a gem object to contain this object in */
 	args->size = roundup(args->size, PAGE_SIZE);
 	r = radeon_gem_object_create(rdev, args->size, args->alignment,
-					args->initial_domain, false,
-					false, &gobj);
+				     args->initial_domain, args->flags,
+				     false, &gobj);
 	if (r) {
 		up_read(&rdev->exclusive_lock);
 		r = radeon_gem_handle_lockup(rdev, r);
@@ -358,16 +358,18 @@ int radeon_gem_wait_idle_ioctl(struct drm_device *dev, void *data,
 	struct drm_gem_object *gobj;
 	struct radeon_bo *robj;
 	int r;
+	uint32_t cur_placement = 0;
 
 	gobj = drm_gem_object_lookup(dev, filp, args->handle);
 	if (gobj == NULL) {
 		return -ENOENT;
 	}
 	robj = gem_to_radeon_bo(gobj);
-	r = radeon_bo_wait(robj, NULL, false);
-	/* callback hw specific functions if any */
-	if (rdev->asic->ioctl_wait_idle)
-		robj->rdev->asic->ioctl_wait_idle(rdev, robj);
+	r = radeon_bo_wait(robj, &cur_placement, false);
+	/* Flush HDP cache via MMIO if necessary */
+	if (rdev->asic->mmio_hdp_flush &&
+	    radeon_mem_type_to_domain(cur_placement) == RADEON_GEM_DOMAIN_VRAM)
+		robj->rdev->asic->mmio_hdp_flush(rdev);
 	drm_gem_object_unreference_unlocked(gobj);
 	r = radeon_gem_handle_lockup(rdev, r);
 	return r;
@@ -461,11 +463,6 @@ int radeon_gem_va_ioctl(struct drm_device *dev, void *data,
 		args->operation = RADEON_VA_RESULT_ERROR;
 		return -EINVAL;
 	}
-	if (!(args->flags & RADEON_VM_PAGE_SNOOPED)) {
-		dev_err(&dev->pdev->dev, "only supported snooped mapping for now\n");
-		args->operation = RADEON_VA_RESULT_ERROR;
-		return -EINVAL;
-	}
 
 	switch (args->operation) {
 	case RADEON_VA_MAP:
@@ -499,9 +496,9 @@ int radeon_gem_va_ioctl(struct drm_device *dev, void *data,
 
 	switch (args->operation) {
 	case RADEON_VA_MAP:
-		if (bo_va->soffset) {
+		if (bo_va->it.start) {
 			args->operation = RADEON_VA_RESULT_VA_EXIST;
-			args->offset = bo_va->soffset;
+			args->offset = bo_va->it.start * RADEON_GPU_PAGE_SIZE;
 			goto out;
 		}
 		r = radeon_vm_bo_set_addr(rdev, bo_va, args->offset, args->flags);
@@ -572,9 +569,8 @@ int radeon_mode_dumb_create(struct drm_file *file_priv,
 	args->size = ALIGN(args->size, PAGE_SIZE);
 
 	r = radeon_gem_object_create(rdev, args->size, 0,
-				     RADEON_GEM_DOMAIN_VRAM,
-				     false, ttm_bo_type_device,
-				     &gobj);
+				     RADEON_GEM_DOMAIN_VRAM, 0,
+				     false, &gobj);
 	if (r)
 		return -ENOMEM;
 
diff --git a/drivers/gpu/drm/radeon/radeon_ib.c b/drivers/gpu/drm/radeon/radeon_ib.c
new file mode 100644
index 000000000000..5bf2c0a05827
--- /dev/null
+++ b/drivers/gpu/drm/radeon/radeon_ib.c
@@ -0,0 +1,320 @@
+/*
+ * Copyright 2008 Advanced Micro Devices, Inc.
+ * Copyright 2008 Red Hat Inc.
+ * Copyright 2009 Jerome Glisse.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Dave Airlie
+ *          Alex Deucher
+ *          Jerome Glisse
+ *          Christian König
+ */
+#include <drm/drmP.h>
+#include "radeon.h"
+
+/*
+ * IB
+ * IBs (Indirect Buffers) and areas of GPU accessible memory where
+ * commands are stored.  You can put a pointer to the IB in the
+ * command ring and the hw will fetch the commands from the IB
+ * and execute them.  Generally userspace acceleration drivers
+ * produce command buffers which are send to the kernel and
+ * put in IBs for execution by the requested ring.
+ */
+static int radeon_debugfs_sa_init(struct radeon_device *rdev);
+
+/**
+ * radeon_ib_get - request an IB (Indirect Buffer)
+ *
+ * @rdev: radeon_device pointer
+ * @ring: ring index the IB is associated with
+ * @ib: IB object returned
+ * @size: requested IB size
+ *
+ * Request an IB (all asics).  IBs are allocated using the
+ * suballocator.
+ * Returns 0 on success, error on failure.
+ */
+int radeon_ib_get(struct radeon_device *rdev, int ring,
+		  struct radeon_ib *ib, struct radeon_vm *vm,
+		  unsigned size)
+{
+	int r;
+
+	r = radeon_sa_bo_new(rdev, &rdev->ring_tmp_bo, &ib->sa_bo, size, 256);
+	if (r) {
+		dev_err(rdev->dev, "failed to get a new IB (%d)\n", r);
+		return r;
+	}
+
+	r = radeon_semaphore_create(rdev, &ib->semaphore);
+	if (r) {
+		return r;
+	}
+
+	ib->ring = ring;
+	ib->fence = NULL;
+	ib->ptr = radeon_sa_bo_cpu_addr(ib->sa_bo);
+	ib->vm = vm;
+	if (vm) {
+		/* ib pool is bound at RADEON_VA_IB_OFFSET in virtual address
+		 * space and soffset is the offset inside the pool bo
+		 */
+		ib->gpu_addr = ib->sa_bo->soffset + RADEON_VA_IB_OFFSET;
+	} else {
+		ib->gpu_addr = radeon_sa_bo_gpu_addr(ib->sa_bo);
+	}
+	ib->is_const_ib = false;
+
+	return 0;
+}
+
+/**
+ * radeon_ib_free - free an IB (Indirect Buffer)
+ *
+ * @rdev: radeon_device pointer
+ * @ib: IB object to free
+ *
+ * Free an IB (all asics).
+ */
+void radeon_ib_free(struct radeon_device *rdev, struct radeon_ib *ib)
+{
+	radeon_semaphore_free(rdev, &ib->semaphore, ib->fence);
+	radeon_sa_bo_free(rdev, &ib->sa_bo, ib->fence);
+	radeon_fence_unref(&ib->fence);
+}
+
+/**
+ * radeon_ib_schedule - schedule an IB (Indirect Buffer) on the ring
+ *
+ * @rdev: radeon_device pointer
+ * @ib: IB object to schedule
+ * @const_ib: Const IB to schedule (SI only)
+ * @hdp_flush: Whether or not to perform an HDP cache flush
+ *
+ * Schedule an IB on the associated ring (all asics).
+ * Returns 0 on success, error on failure.
+ *
+ * On SI, there are two parallel engines fed from the primary ring,
+ * the CE (Constant Engine) and the DE (Drawing Engine).  Since
+ * resource descriptors have moved to memory, the CE allows you to
+ * prime the caches while the DE is updating register state so that
+ * the resource descriptors will be already in cache when the draw is
+ * processed.  To accomplish this, the userspace driver submits two
+ * IBs, one for the CE and one for the DE.  If there is a CE IB (called
+ * a CONST_IB), it will be put on the ring prior to the DE IB.  Prior
+ * to SI there was just a DE IB.
+ */
+int radeon_ib_schedule(struct radeon_device *rdev, struct radeon_ib *ib,
+		       struct radeon_ib *const_ib, bool hdp_flush)
+{
+	struct radeon_ring *ring = &rdev->ring[ib->ring];
+	int r = 0;
+
+	if (!ib->length_dw || !ring->ready) {
+		/* TODO: Nothings in the ib we should report. */
+		dev_err(rdev->dev, "couldn't schedule ib\n");
+		return -EINVAL;
+	}
+
+	/* 64 dwords should be enough for fence too */
+	r = radeon_ring_lock(rdev, ring, 64 + RADEON_NUM_SYNCS * 8);
+	if (r) {
+		dev_err(rdev->dev, "scheduling IB failed (%d).\n", r);
+		return r;
+	}
+
+	/* grab a vm id if necessary */
+	if (ib->vm) {
+		struct radeon_fence *vm_id_fence;
+		vm_id_fence = radeon_vm_grab_id(rdev, ib->vm, ib->ring);
+        	radeon_semaphore_sync_to(ib->semaphore, vm_id_fence);
+	}
+
+	/* sync with other rings */
+	r = radeon_semaphore_sync_rings(rdev, ib->semaphore, ib->ring);
+	if (r) {
+		dev_err(rdev->dev, "failed to sync rings (%d)\n", r);
+		radeon_ring_unlock_undo(rdev, ring);
+		return r;
+	}
+
+	if (ib->vm)
+		radeon_vm_flush(rdev, ib->vm, ib->ring);
+
+	if (const_ib) {
+		radeon_ring_ib_execute(rdev, const_ib->ring, const_ib);
+		radeon_semaphore_free(rdev, &const_ib->semaphore, NULL);
+	}
+	radeon_ring_ib_execute(rdev, ib->ring, ib);
+	r = radeon_fence_emit(rdev, &ib->fence, ib->ring);
+	if (r) {
+		dev_err(rdev->dev, "failed to emit fence for new IB (%d)\n", r);
+		radeon_ring_unlock_undo(rdev, ring);
+		return r;
+	}
+	if (const_ib) {
+		const_ib->fence = radeon_fence_ref(ib->fence);
+	}
+
+	if (ib->vm)
+		radeon_vm_fence(rdev, ib->vm, ib->fence);
+
+	radeon_ring_unlock_commit(rdev, ring, hdp_flush);
+	return 0;
+}
+
+/**
+ * radeon_ib_pool_init - Init the IB (Indirect Buffer) pool
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Initialize the suballocator to manage a pool of memory
+ * for use as IBs (all asics).
+ * Returns 0 on success, error on failure.
+ */
+int radeon_ib_pool_init(struct radeon_device *rdev)
+{
+	int r;
+
+	if (rdev->ib_pool_ready) {
+		return 0;
+	}
+
+	if (rdev->family >= CHIP_BONAIRE) {
+		r = radeon_sa_bo_manager_init(rdev, &rdev->ring_tmp_bo,
+					      RADEON_IB_POOL_SIZE*64*1024,
+					      RADEON_GPU_PAGE_SIZE,
+					      RADEON_GEM_DOMAIN_GTT,
+					      RADEON_GEM_GTT_WC);
+	} else {
+		/* Before CIK, it's better to stick to cacheable GTT due
+		 * to the command stream checking
+		 */
+		r = radeon_sa_bo_manager_init(rdev, &rdev->ring_tmp_bo,
+					      RADEON_IB_POOL_SIZE*64*1024,
+					      RADEON_GPU_PAGE_SIZE,
+					      RADEON_GEM_DOMAIN_GTT, 0);
+	}
+	if (r) {
+		return r;
+	}
+
+	r = radeon_sa_bo_manager_start(rdev, &rdev->ring_tmp_bo);
+	if (r) {
+		return r;
+	}
+
+	rdev->ib_pool_ready = true;
+	if (radeon_debugfs_sa_init(rdev)) {
+		dev_err(rdev->dev, "failed to register debugfs file for SA\n");
+	}
+	return 0;
+}
+
+/**
+ * radeon_ib_pool_fini - Free the IB (Indirect Buffer) pool
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Tear down the suballocator managing the pool of memory
+ * for use as IBs (all asics).
+ */
+void radeon_ib_pool_fini(struct radeon_device *rdev)
+{
+	if (rdev->ib_pool_ready) {
+		radeon_sa_bo_manager_suspend(rdev, &rdev->ring_tmp_bo);
+		radeon_sa_bo_manager_fini(rdev, &rdev->ring_tmp_bo);
+		rdev->ib_pool_ready = false;
+	}
+}
+
+/**
+ * radeon_ib_ring_tests - test IBs on the rings
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Test an IB (Indirect Buffer) on each ring.
+ * If the test fails, disable the ring.
+ * Returns 0 on success, error if the primary GFX ring
+ * IB test fails.
+ */
+int radeon_ib_ring_tests(struct radeon_device *rdev)
+{
+	unsigned i;
+	int r;
+
+	for (i = 0; i < RADEON_NUM_RINGS; ++i) {
+		struct radeon_ring *ring = &rdev->ring[i];
+
+		if (!ring->ready)
+			continue;
+
+		r = radeon_ib_test(rdev, i, ring);
+		if (r) {
+			ring->ready = false;
+			rdev->needs_reset = false;
+
+			if (i == RADEON_RING_TYPE_GFX_INDEX) {
+				/* oh, oh, that's really bad */
+				DRM_ERROR("radeon: failed testing IB on GFX ring (%d).\n", r);
+		                rdev->accel_working = false;
+				return r;
+
+			} else {
+				/* still not good, but we can live with it */
+				DRM_ERROR("radeon: failed testing IB on ring %d (%d).\n", i, r);
+			}
+		}
+	}
+	return 0;
+}
+
+/*
+ * Debugfs info
+ */
+#if defined(CONFIG_DEBUG_FS)
+
+static int radeon_debugfs_sa_info(struct seq_file *m, void *data)
+{
+	struct drm_info_node *node = (struct drm_info_node *) m->private;
+	struct drm_device *dev = node->minor->dev;
+	struct radeon_device *rdev = dev->dev_private;
+
+	radeon_sa_bo_dump_debug_info(&rdev->ring_tmp_bo, m);
+
+	return 0;
+
+}
+
+static struct drm_info_list radeon_debugfs_sa_list[] = {
+        {"radeon_sa_info", &radeon_debugfs_sa_info, 0, NULL},
+};
+
+#endif
+
+static int radeon_debugfs_sa_init(struct radeon_device *rdev)
+{
+#if defined(CONFIG_DEBUG_FS)
+	return radeon_debugfs_add_files(rdev, radeon_debugfs_sa_list, 1);
+#else
+	return 0;
+#endif
+}
diff --git a/drivers/gpu/drm/radeon/radeon_kms.c b/drivers/gpu/drm/radeon/radeon_kms.c
index 35d931881b4b..eb7164d07985 100644
--- a/drivers/gpu/drm/radeon/radeon_kms.c
+++ b/drivers/gpu/drm/radeon/radeon_kms.c
@@ -254,7 +254,18 @@ static int radeon_info_ioctl(struct drm_device *dev, void *data, struct drm_file
 		}
 		break;
 	case RADEON_INFO_ACCEL_WORKING2:
-		*value = rdev->accel_working;
+		if (rdev->family == CHIP_HAWAII) {
+			if (rdev->accel_working) {
+				if (rdev->new_fw)
+					*value = 3;
+				else
+					*value = 2;
+			} else {
+				*value = 0;
+			}
+		} else {
+			*value = rdev->accel_working;
+		}
 		break;
 	case RADEON_INFO_TILING_CONFIG:
 		if (rdev->family >= CHIP_BONAIRE)
@@ -579,7 +590,7 @@ int radeon_driver_open_kms(struct drm_device *dev, struct drm_file *file_priv)
 	/* new gpu have virtual address space support */
 	if (rdev->family >= CHIP_CAYMAN) {
 		struct radeon_fpriv *fpriv;
-		struct radeon_bo_va *bo_va;
+		struct radeon_vm *vm;
 		int r;
 
 		fpriv = kzalloc(sizeof(*fpriv), GFP_KERNEL);
@@ -587,7 +598,8 @@ int radeon_driver_open_kms(struct drm_device *dev, struct drm_file *file_priv)
 			return -ENOMEM;
 		}
 
-		r = radeon_vm_init(rdev, &fpriv->vm);
+		vm = &fpriv->vm;
+		r = radeon_vm_init(rdev, vm);
 		if (r) {
 			kfree(fpriv);
 			return r;
@@ -596,22 +608,23 @@ int radeon_driver_open_kms(struct drm_device *dev, struct drm_file *file_priv)
 		if (rdev->accel_working) {
 			r = radeon_bo_reserve(rdev->ring_tmp_bo.bo, false);
 			if (r) {
-				radeon_vm_fini(rdev, &fpriv->vm);
+				radeon_vm_fini(rdev, vm);
 				kfree(fpriv);
 				return r;
 			}
 
 			/* map the ib pool buffer read only into
 			 * virtual address space */
-			bo_va = radeon_vm_bo_add(rdev, &fpriv->vm,
-						 rdev->ring_tmp_bo.bo);
-			r = radeon_vm_bo_set_addr(rdev, bo_va, RADEON_VA_IB_OFFSET,
+			vm->ib_bo_va = radeon_vm_bo_add(rdev, vm,
+							rdev->ring_tmp_bo.bo);
+			r = radeon_vm_bo_set_addr(rdev, vm->ib_bo_va,
+						  RADEON_VA_IB_OFFSET,
 						  RADEON_VM_PAGE_READABLE |
 						  RADEON_VM_PAGE_SNOOPED);
 
 			radeon_bo_unreserve(rdev->ring_tmp_bo.bo);
 			if (r) {
-				radeon_vm_fini(rdev, &fpriv->vm);
+				radeon_vm_fini(rdev, vm);
 				kfree(fpriv);
 				return r;
 			}
@@ -640,21 +653,19 @@ void radeon_driver_postclose_kms(struct drm_device *dev,
 	/* new gpu have virtual address space support */
 	if (rdev->family >= CHIP_CAYMAN && file_priv->driver_priv) {
 		struct radeon_fpriv *fpriv = file_priv->driver_priv;
-		struct radeon_bo_va *bo_va;
+		struct radeon_vm *vm = &fpriv->vm;
 		int r;
 
 		if (rdev->accel_working) {
 			r = radeon_bo_reserve(rdev->ring_tmp_bo.bo, false);
 			if (!r) {
-				bo_va = radeon_vm_bo_find(&fpriv->vm,
-							  rdev->ring_tmp_bo.bo);
-				if (bo_va)
-					radeon_vm_bo_rmv(rdev, bo_va);
+				if (vm->ib_bo_va)
+					radeon_vm_bo_rmv(rdev, vm->ib_bo_va);
 				radeon_bo_unreserve(rdev->ring_tmp_bo.bo);
 			}
 		}
 
-		radeon_vm_fini(rdev, &fpriv->vm);
+		radeon_vm_fini(rdev, vm);
 		kfree(fpriv);
 		file_priv->driver_priv = NULL;
 	}
diff --git a/drivers/gpu/drm/radeon/radeon_mode.h b/drivers/gpu/drm/radeon/radeon_mode.h
index 0592ddb0904b..e27608c29c11 100644
--- a/drivers/gpu/drm/radeon/radeon_mode.h
+++ b/drivers/gpu/drm/radeon/radeon_mode.h
@@ -685,10 +685,11 @@ extern bool radeon_dig_monitor_is_duallink(struct drm_encoder *encoder,
 
 extern u16 radeon_encoder_get_dp_bridge_encoder_id(struct drm_encoder *encoder);
 extern u16 radeon_connector_encoder_get_dp_bridge_encoder_id(struct drm_connector *connector);
-extern bool radeon_connector_encoder_is_hbr2(struct drm_connector *connector);
 extern bool radeon_connector_is_dp12_capable(struct drm_connector *connector);
 extern int radeon_get_monitor_bpc(struct drm_connector *connector);
 
+extern struct edid *radeon_connector_edid(struct drm_connector *connector);
+
 extern void radeon_connector_hotplug(struct drm_connector *connector);
 extern int radeon_dp_mode_valid_helper(struct drm_connector *connector,
 				       struct drm_display_mode *mode);
@@ -738,7 +739,6 @@ extern void radeon_i2c_put_byte(struct radeon_i2c_chan *i2c,
 extern void radeon_router_select_ddc_port(struct radeon_connector *radeon_connector);
 extern void radeon_router_select_cd_port(struct radeon_connector *radeon_connector);
 extern bool radeon_ddc_probe(struct radeon_connector *radeon_connector, bool use_aux);
-extern int radeon_ddc_get_modes(struct radeon_connector *radeon_connector);
 
 extern struct drm_encoder *radeon_best_encoder(struct drm_connector *connector);
 
diff --git a/drivers/gpu/drm/radeon/radeon_object.c b/drivers/gpu/drm/radeon/radeon_object.c
index 6c717b257d6d..480c87d8edc5 100644
--- a/drivers/gpu/drm/radeon/radeon_object.c
+++ b/drivers/gpu/drm/radeon/radeon_object.c
@@ -46,16 +46,6 @@ static void radeon_bo_clear_surface_reg(struct radeon_bo *bo);
  * function are calling it.
  */
 
-static void radeon_bo_clear_va(struct radeon_bo *bo)
-{
-	struct radeon_bo_va *bo_va, *tmp;
-
-	list_for_each_entry_safe(bo_va, tmp, &bo->va, bo_list) {
-		/* remove from all vm address space */
-		radeon_vm_bo_rmv(bo->rdev, bo_va);
-	}
-}
-
 static void radeon_update_memory_usage(struct radeon_bo *bo,
 				       unsigned mem_type, int sign)
 {
@@ -90,7 +80,7 @@ static void radeon_ttm_bo_destroy(struct ttm_buffer_object *tbo)
 	list_del_init(&bo->list);
 	mutex_unlock(&bo->rdev->gem.mutex);
 	radeon_bo_clear_surface_reg(bo);
-	radeon_bo_clear_va(bo);
+	WARN_ON(!list_empty(&bo->va));
 	drm_gem_object_release(&bo->gem_base);
 	kfree(bo);
 }
@@ -114,15 +104,23 @@ void radeon_ttm_placement_from_domain(struct radeon_bo *rbo, u32 domain)
 		rbo->placements[c++] = TTM_PL_FLAG_WC | TTM_PL_FLAG_UNCACHED |
 					TTM_PL_FLAG_VRAM;
 	if (domain & RADEON_GEM_DOMAIN_GTT) {
-		if (rbo->rdev->flags & RADEON_IS_AGP) {
-			rbo->placements[c++] = TTM_PL_FLAG_WC | TTM_PL_FLAG_TT;
+		if (rbo->flags & RADEON_GEM_GTT_UC) {
+			rbo->placements[c++] = TTM_PL_FLAG_UNCACHED | TTM_PL_FLAG_TT;
+		} else if ((rbo->flags & RADEON_GEM_GTT_WC) ||
+			   (rbo->rdev->flags & RADEON_IS_AGP)) {
+			rbo->placements[c++] = TTM_PL_FLAG_WC | TTM_PL_FLAG_UNCACHED |
+				TTM_PL_FLAG_TT;
 		} else {
 			rbo->placements[c++] = TTM_PL_FLAG_CACHED | TTM_PL_FLAG_TT;
 		}
 	}
 	if (domain & RADEON_GEM_DOMAIN_CPU) {
-		if (rbo->rdev->flags & RADEON_IS_AGP) {
-			rbo->placements[c++] = TTM_PL_FLAG_WC | TTM_PL_FLAG_SYSTEM;
+		if (rbo->flags & RADEON_GEM_GTT_UC) {
+			rbo->placements[c++] = TTM_PL_FLAG_UNCACHED | TTM_PL_FLAG_SYSTEM;
+		} else if ((rbo->flags & RADEON_GEM_GTT_WC) ||
+		    rbo->rdev->flags & RADEON_IS_AGP) {
+			rbo->placements[c++] = TTM_PL_FLAG_WC | TTM_PL_FLAG_UNCACHED |
+				TTM_PL_FLAG_SYSTEM;
 		} else {
 			rbo->placements[c++] = TTM_PL_FLAG_CACHED | TTM_PL_FLAG_SYSTEM;
 		}
@@ -146,7 +144,7 @@ void radeon_ttm_placement_from_domain(struct radeon_bo *rbo, u32 domain)
 
 int radeon_bo_create(struct radeon_device *rdev,
 		     unsigned long size, int byte_align, bool kernel, u32 domain,
-		     struct sg_table *sg, struct radeon_bo **bo_ptr)
+		     u32 flags, struct sg_table *sg, struct radeon_bo **bo_ptr)
 {
 	struct radeon_bo *bo;
 	enum ttm_bo_type type;
@@ -183,6 +181,12 @@ int radeon_bo_create(struct radeon_device *rdev,
 	bo->initial_domain = domain & (RADEON_GEM_DOMAIN_VRAM |
 	                               RADEON_GEM_DOMAIN_GTT |
 	                               RADEON_GEM_DOMAIN_CPU);
+
+	bo->flags = flags;
+	/* PCI GART is always snooped */
+	if (!(rdev->flags & RADEON_IS_PCIE))
+		bo->flags &= ~(RADEON_GEM_GTT_WC | RADEON_GEM_GTT_UC);
+
 	radeon_ttm_placement_from_domain(bo, domain);
 	/* Kernel allocation are uninterruptible */
 	down_read(&rdev->pm.mclk_lock);
@@ -232,6 +236,15 @@ void radeon_bo_kunmap(struct radeon_bo *bo)
 	ttm_bo_kunmap(&bo->kmap);
 }
 
+struct radeon_bo *radeon_bo_ref(struct radeon_bo *bo)
+{
+	if (bo == NULL)
+		return NULL;
+
+	ttm_bo_reference(&bo->tbo);
+	return bo;
+}
+
 void radeon_bo_unref(struct radeon_bo **bo)
 {
 	struct ttm_buffer_object *tbo;
@@ -241,9 +254,7 @@ void radeon_bo_unref(struct radeon_bo **bo)
 		return;
 	rdev = (*bo)->rdev;
 	tbo = &((*bo)->tbo);
-	down_read(&rdev->pm.mclk_lock);
 	ttm_bo_unref(&tbo);
-	up_read(&rdev->pm.mclk_lock);
 	if (tbo == NULL)
 		*bo = NULL;
 }
@@ -292,9 +303,13 @@ int radeon_bo_pin_restricted(struct radeon_bo *bo, u32 domain, u64 max_offset,
 		bo->pin_count = 1;
 		if (gpu_addr != NULL)
 			*gpu_addr = radeon_bo_gpu_offset(bo);
-	}
-	if (unlikely(r != 0))
+		if (domain == RADEON_GEM_DOMAIN_VRAM)
+			bo->rdev->vram_pin_size += radeon_bo_size(bo);
+		else
+			bo->rdev->gart_pin_size += radeon_bo_size(bo);
+	} else {
 		dev_err(bo->rdev->dev, "%p pin failed\n", bo);
+	}
 	return r;
 }
 
@@ -317,8 +332,14 @@ int radeon_bo_unpin(struct radeon_bo *bo)
 	for (i = 0; i < bo->placement.num_placement; i++)
 		bo->placements[i] &= ~TTM_PL_FLAG_NO_EVICT;
 	r = ttm_bo_validate(&bo->tbo, &bo->placement, false, false);
-	if (unlikely(r != 0))
+	if (likely(r == 0)) {
+		if (bo->tbo.mem.mem_type == TTM_PL_VRAM)
+			bo->rdev->vram_pin_size -= radeon_bo_size(bo);
+		else
+			bo->rdev->gart_pin_size -= radeon_bo_size(bo);
+	} else {
 		dev_err(bo->rdev->dev, "%p validate failed for unpin\n", bo);
+	}
 	return r;
 }
 
diff --git a/drivers/gpu/drm/radeon/radeon_object.h b/drivers/gpu/drm/radeon/radeon_object.h
index 5a873f31a171..98a47fdf3625 100644
--- a/drivers/gpu/drm/radeon/radeon_object.h
+++ b/drivers/gpu/drm/radeon/radeon_object.h
@@ -124,11 +124,12 @@ extern int radeon_bo_wait(struct radeon_bo *bo, u32 *mem_type,
 
 extern int radeon_bo_create(struct radeon_device *rdev,
 			    unsigned long size, int byte_align,
-			    bool kernel, u32 domain,
+			    bool kernel, u32 domain, u32 flags,
 			    struct sg_table *sg,
 			    struct radeon_bo **bo_ptr);
 extern int radeon_bo_kmap(struct radeon_bo *bo, void **ptr);
 extern void radeon_bo_kunmap(struct radeon_bo *bo);
+extern struct radeon_bo *radeon_bo_ref(struct radeon_bo *bo);
 extern void radeon_bo_unref(struct radeon_bo **bo);
 extern int radeon_bo_pin(struct radeon_bo *bo, u32 domain, u64 *gpu_addr);
 extern int radeon_bo_pin_restricted(struct radeon_bo *bo, u32 domain,
@@ -170,7 +171,8 @@ static inline void * radeon_sa_bo_cpu_addr(struct radeon_sa_bo *sa_bo)
 
 extern int radeon_sa_bo_manager_init(struct radeon_device *rdev,
 				     struct radeon_sa_manager *sa_manager,
-				     unsigned size, u32 align, u32 domain);
+				     unsigned size, u32 align, u32 domain,
+				     u32 flags);
 extern void radeon_sa_bo_manager_fini(struct radeon_device *rdev,
 				      struct radeon_sa_manager *sa_manager);
 extern int radeon_sa_bo_manager_start(struct radeon_device *rdev,
diff --git a/drivers/gpu/drm/radeon/radeon_pm.c b/drivers/gpu/drm/radeon/radeon_pm.c
index e447e390d09a..164898b0010c 100644
--- a/drivers/gpu/drm/radeon/radeon_pm.c
+++ b/drivers/gpu/drm/radeon/radeon_pm.c
@@ -460,10 +460,6 @@ static ssize_t radeon_get_dpm_state(struct device *dev,
 	struct radeon_device *rdev = ddev->dev_private;
 	enum radeon_pm_state_type pm = rdev->pm.dpm.user_state;
 
-	if  ((rdev->flags & RADEON_IS_PX) &&
-	     (ddev->switch_power_state != DRM_SWITCH_POWER_ON))
-		return snprintf(buf, PAGE_SIZE, "off\n");
-
 	return snprintf(buf, PAGE_SIZE, "%s\n",
 			(pm == POWER_STATE_TYPE_BATTERY) ? "battery" :
 			(pm == POWER_STATE_TYPE_BALANCED) ? "balanced" : "performance");
@@ -477,11 +473,6 @@ static ssize_t radeon_set_dpm_state(struct device *dev,
 	struct drm_device *ddev = dev_get_drvdata(dev);
 	struct radeon_device *rdev = ddev->dev_private;
 
-	/* Can't set dpm state when the card is off */
-	if  ((rdev->flags & RADEON_IS_PX) &&
-	     (ddev->switch_power_state != DRM_SWITCH_POWER_ON))
-		return -EINVAL;
-
 	mutex_lock(&rdev->pm.mutex);
 	if (strncmp("battery", buf, strlen("battery")) == 0)
 		rdev->pm.dpm.user_state = POWER_STATE_TYPE_BATTERY;
@@ -495,7 +486,12 @@ static ssize_t radeon_set_dpm_state(struct device *dev,
 		goto fail;
 	}
 	mutex_unlock(&rdev->pm.mutex);
-	radeon_pm_compute_clocks(rdev);
+
+	/* Can't set dpm state when the card is off */
+	if (!(rdev->flags & RADEON_IS_PX) ||
+	    (ddev->switch_power_state == DRM_SWITCH_POWER_ON))
+		radeon_pm_compute_clocks(rdev);
+
 fail:
 	return count;
 }
@@ -1303,10 +1299,6 @@ int radeon_pm_init(struct radeon_device *rdev)
 	case CHIP_RS780:
 	case CHIP_RS880:
 	case CHIP_RV770:
-	case CHIP_BARTS:
-	case CHIP_TURKS:
-	case CHIP_CAICOS:
-	case CHIP_CAYMAN:
 		/* DPM requires the RLC, RV770+ dGPU requires SMC */
 		if (!rdev->rlc_fw)
 			rdev->pm.pm_method = PM_METHOD_PROFILE;
@@ -1330,6 +1322,10 @@ int radeon_pm_init(struct radeon_device *rdev)
 	case CHIP_PALM:
 	case CHIP_SUMO:
 	case CHIP_SUMO2:
+	case CHIP_BARTS:
+	case CHIP_TURKS:
+	case CHIP_CAICOS:
+	case CHIP_CAYMAN:
 	case CHIP_ARUBA:
 	case CHIP_TAHITI:
 	case CHIP_PITCAIRN:
@@ -1400,9 +1396,7 @@ static void radeon_pm_fini_old(struct radeon_device *rdev)
 	}
 
 	radeon_hwmon_fini(rdev);
-
-	if (rdev->pm.power_state)
-		kfree(rdev->pm.power_state);
+	kfree(rdev->pm.power_state);
 }
 
 static void radeon_pm_fini_dpm(struct radeon_device *rdev)
@@ -1421,9 +1415,7 @@ static void radeon_pm_fini_dpm(struct radeon_device *rdev)
 	radeon_dpm_fini(rdev);
 
 	radeon_hwmon_fini(rdev);
-
-	if (rdev->pm.power_state)
-		kfree(rdev->pm.power_state);
+	kfree(rdev->pm.power_state);
 }
 
 void radeon_pm_fini(struct radeon_device *rdev)
diff --git a/drivers/gpu/drm/radeon/radeon_prime.c b/drivers/gpu/drm/radeon/radeon_prime.c
index 20074560fc25..0b16f2cbcf17 100644
--- a/drivers/gpu/drm/radeon/radeon_prime.c
+++ b/drivers/gpu/drm/radeon/radeon_prime.c
@@ -65,7 +65,7 @@ struct drm_gem_object *radeon_gem_prime_import_sg_table(struct drm_device *dev,
 	int ret;
 
 	ret = radeon_bo_create(rdev, size, PAGE_SIZE, false,
-			       RADEON_GEM_DOMAIN_GTT, sg, &bo);
+			       RADEON_GEM_DOMAIN_GTT, 0, sg, &bo);
 	if (ret)
 		return ERR_PTR(ret);
 
@@ -103,3 +103,11 @@ void radeon_gem_prime_unpin(struct drm_gem_object *obj)
 	radeon_bo_unpin(bo);
 	radeon_bo_unreserve(bo);
 }
+
+
+struct reservation_object *radeon_gem_prime_res_obj(struct drm_gem_object *obj)
+{
+	struct radeon_bo *bo = gem_to_radeon_bo(obj);
+
+	return bo->tbo.resv;
+}
diff --git a/drivers/gpu/drm/radeon/radeon_ring.c b/drivers/gpu/drm/radeon/radeon_ring.c
index f8050f5429e2..d65607902537 100644
--- a/drivers/gpu/drm/radeon/radeon_ring.c
+++ b/drivers/gpu/drm/radeon/radeon_ring.c
@@ -26,258 +26,8 @@
  *          Jerome Glisse
  *          Christian König
  */
-#include <linux/seq_file.h>
-#include <linux/slab.h>
 #include <drm/drmP.h>
-#include <drm/radeon_drm.h>
-#include "radeon_reg.h"
 #include "radeon.h"
-#include "atom.h"
-
-/*
- * IB
- * IBs (Indirect Buffers) and areas of GPU accessible memory where
- * commands are stored.  You can put a pointer to the IB in the
- * command ring and the hw will fetch the commands from the IB
- * and execute them.  Generally userspace acceleration drivers
- * produce command buffers which are send to the kernel and
- * put in IBs for execution by the requested ring.
- */
-static int radeon_debugfs_sa_init(struct radeon_device *rdev);
-
-/**
- * radeon_ib_get - request an IB (Indirect Buffer)
- *
- * @rdev: radeon_device pointer
- * @ring: ring index the IB is associated with
- * @ib: IB object returned
- * @size: requested IB size
- *
- * Request an IB (all asics).  IBs are allocated using the
- * suballocator.
- * Returns 0 on success, error on failure.
- */
-int radeon_ib_get(struct radeon_device *rdev, int ring,
-		  struct radeon_ib *ib, struct radeon_vm *vm,
-		  unsigned size)
-{
-	int r;
-
-	r = radeon_sa_bo_new(rdev, &rdev->ring_tmp_bo, &ib->sa_bo, size, 256);
-	if (r) {
-		dev_err(rdev->dev, "failed to get a new IB (%d)\n", r);
-		return r;
-	}
-
-	r = radeon_semaphore_create(rdev, &ib->semaphore);
-	if (r) {
-		return r;
-	}
-
-	ib->ring = ring;
-	ib->fence = NULL;
-	ib->ptr = radeon_sa_bo_cpu_addr(ib->sa_bo);
-	ib->vm = vm;
-	if (vm) {
-		/* ib pool is bound at RADEON_VA_IB_OFFSET in virtual address
-		 * space and soffset is the offset inside the pool bo
-		 */
-		ib->gpu_addr = ib->sa_bo->soffset + RADEON_VA_IB_OFFSET;
-	} else {
-		ib->gpu_addr = radeon_sa_bo_gpu_addr(ib->sa_bo);
-	}
-	ib->is_const_ib = false;
-
-	return 0;
-}
-
-/**
- * radeon_ib_free - free an IB (Indirect Buffer)
- *
- * @rdev: radeon_device pointer
- * @ib: IB object to free
- *
- * Free an IB (all asics).
- */
-void radeon_ib_free(struct radeon_device *rdev, struct radeon_ib *ib)
-{
-	radeon_semaphore_free(rdev, &ib->semaphore, ib->fence);
-	radeon_sa_bo_free(rdev, &ib->sa_bo, ib->fence);
-	radeon_fence_unref(&ib->fence);
-}
-
-/**
- * radeon_ib_schedule - schedule an IB (Indirect Buffer) on the ring
- *
- * @rdev: radeon_device pointer
- * @ib: IB object to schedule
- * @const_ib: Const IB to schedule (SI only)
- *
- * Schedule an IB on the associated ring (all asics).
- * Returns 0 on success, error on failure.
- *
- * On SI, there are two parallel engines fed from the primary ring,
- * the CE (Constant Engine) and the DE (Drawing Engine).  Since
- * resource descriptors have moved to memory, the CE allows you to
- * prime the caches while the DE is updating register state so that
- * the resource descriptors will be already in cache when the draw is
- * processed.  To accomplish this, the userspace driver submits two
- * IBs, one for the CE and one for the DE.  If there is a CE IB (called
- * a CONST_IB), it will be put on the ring prior to the DE IB.  Prior
- * to SI there was just a DE IB.
- */
-int radeon_ib_schedule(struct radeon_device *rdev, struct radeon_ib *ib,
-		       struct radeon_ib *const_ib)
-{
-	struct radeon_ring *ring = &rdev->ring[ib->ring];
-	int r = 0;
-
-	if (!ib->length_dw || !ring->ready) {
-		/* TODO: Nothings in the ib we should report. */
-		dev_err(rdev->dev, "couldn't schedule ib\n");
-		return -EINVAL;
-	}
-
-	/* 64 dwords should be enough for fence too */
-	r = radeon_ring_lock(rdev, ring, 64 + RADEON_NUM_SYNCS * 8);
-	if (r) {
-		dev_err(rdev->dev, "scheduling IB failed (%d).\n", r);
-		return r;
-	}
-
-	/* grab a vm id if necessary */
-	if (ib->vm) {
-		struct radeon_fence *vm_id_fence;
-		vm_id_fence = radeon_vm_grab_id(rdev, ib->vm, ib->ring);
-        	radeon_semaphore_sync_to(ib->semaphore, vm_id_fence);
-	}
-
-	/* sync with other rings */
-	r = radeon_semaphore_sync_rings(rdev, ib->semaphore, ib->ring);
-	if (r) {
-		dev_err(rdev->dev, "failed to sync rings (%d)\n", r);
-		radeon_ring_unlock_undo(rdev, ring);
-		return r;
-	}
-
-	if (ib->vm)
-		radeon_vm_flush(rdev, ib->vm, ib->ring);
-
-	if (const_ib) {
-		radeon_ring_ib_execute(rdev, const_ib->ring, const_ib);
-		radeon_semaphore_free(rdev, &const_ib->semaphore, NULL);
-	}
-	radeon_ring_ib_execute(rdev, ib->ring, ib);
-	r = radeon_fence_emit(rdev, &ib->fence, ib->ring);
-	if (r) {
-		dev_err(rdev->dev, "failed to emit fence for new IB (%d)\n", r);
-		radeon_ring_unlock_undo(rdev, ring);
-		return r;
-	}
-	if (const_ib) {
-		const_ib->fence = radeon_fence_ref(ib->fence);
-	}
-
-	if (ib->vm)
-		radeon_vm_fence(rdev, ib->vm, ib->fence);
-
-	radeon_ring_unlock_commit(rdev, ring);
-	return 0;
-}
-
-/**
- * radeon_ib_pool_init - Init the IB (Indirect Buffer) pool
- *
- * @rdev: radeon_device pointer
- *
- * Initialize the suballocator to manage a pool of memory
- * for use as IBs (all asics).
- * Returns 0 on success, error on failure.
- */
-int radeon_ib_pool_init(struct radeon_device *rdev)
-{
-	int r;
-
-	if (rdev->ib_pool_ready) {
-		return 0;
-	}
-	r = radeon_sa_bo_manager_init(rdev, &rdev->ring_tmp_bo,
-				      RADEON_IB_POOL_SIZE*64*1024,
-				      RADEON_GPU_PAGE_SIZE,
-				      RADEON_GEM_DOMAIN_GTT);
-	if (r) {
-		return r;
-	}
-
-	r = radeon_sa_bo_manager_start(rdev, &rdev->ring_tmp_bo);
-	if (r) {
-		return r;
-	}
-
-	rdev->ib_pool_ready = true;
-	if (radeon_debugfs_sa_init(rdev)) {
-		dev_err(rdev->dev, "failed to register debugfs file for SA\n");
-	}
-	return 0;
-}
-
-/**
- * radeon_ib_pool_fini - Free the IB (Indirect Buffer) pool
- *
- * @rdev: radeon_device pointer
- *
- * Tear down the suballocator managing the pool of memory
- * for use as IBs (all asics).
- */
-void radeon_ib_pool_fini(struct radeon_device *rdev)
-{
-	if (rdev->ib_pool_ready) {
-		radeon_sa_bo_manager_suspend(rdev, &rdev->ring_tmp_bo);
-		radeon_sa_bo_manager_fini(rdev, &rdev->ring_tmp_bo);
-		rdev->ib_pool_ready = false;
-	}
-}
-
-/**
- * radeon_ib_ring_tests - test IBs on the rings
- *
- * @rdev: radeon_device pointer
- *
- * Test an IB (Indirect Buffer) on each ring.
- * If the test fails, disable the ring.
- * Returns 0 on success, error if the primary GFX ring
- * IB test fails.
- */
-int radeon_ib_ring_tests(struct radeon_device *rdev)
-{
-	unsigned i;
-	int r;
-
-	for (i = 0; i < RADEON_NUM_RINGS; ++i) {
-		struct radeon_ring *ring = &rdev->ring[i];
-
-		if (!ring->ready)
-			continue;
-
-		r = radeon_ib_test(rdev, i, ring);
-		if (r) {
-			ring->ready = false;
-			rdev->needs_reset = false;
-
-			if (i == RADEON_RING_TYPE_GFX_INDEX) {
-				/* oh, oh, that's really bad */
-				DRM_ERROR("radeon: failed testing IB on GFX ring (%d).\n", r);
-		                rdev->accel_working = false;
-				return r;
-
-			} else {
-				/* still not good, but we can live with it */
-				DRM_ERROR("radeon: failed testing IB on ring %d (%d).\n", i, r);
-			}
-		}
-	}
-	return 0;
-}
 
 /*
  * Rings
@@ -427,17 +177,29 @@ int radeon_ring_lock(struct radeon_device *rdev, struct radeon_ring *ring, unsig
  *
  * @rdev: radeon_device pointer
  * @ring: radeon_ring structure holding ring information
+ * @hdp_flush: Whether or not to perform an HDP cache flush
  *
  * Update the wptr (write pointer) to tell the GPU to
  * execute new commands on the ring buffer (all asics).
  */
-void radeon_ring_commit(struct radeon_device *rdev, struct radeon_ring *ring)
+void radeon_ring_commit(struct radeon_device *rdev, struct radeon_ring *ring,
+			bool hdp_flush)
 {
+	/* If we are emitting the HDP flush via the ring buffer, we need to
+	 * do it before padding.
+	 */
+	if (hdp_flush && rdev->asic->ring[ring->idx]->hdp_flush)
+		rdev->asic->ring[ring->idx]->hdp_flush(rdev, ring);
 	/* We pad to match fetch size */
 	while (ring->wptr & ring->align_mask) {
 		radeon_ring_write(ring, ring->nop);
 	}
 	mb();
+	/* If we are emitting the HDP flush via MMIO, we need to do it after
+	 * all CPU writes to VRAM finished.
+	 */
+	if (hdp_flush && rdev->asic->mmio_hdp_flush)
+		rdev->asic->mmio_hdp_flush(rdev);
 	radeon_ring_set_wptr(rdev, ring);
 }
 
@@ -447,12 +209,14 @@ void radeon_ring_commit(struct radeon_device *rdev, struct radeon_ring *ring)
  *
  * @rdev: radeon_device pointer
  * @ring: radeon_ring structure holding ring information
+ * @hdp_flush: Whether or not to perform an HDP cache flush
  *
  * Call radeon_ring_commit() then unlock the ring (all asics).
  */
-void radeon_ring_unlock_commit(struct radeon_device *rdev, struct radeon_ring *ring)
+void radeon_ring_unlock_commit(struct radeon_device *rdev, struct radeon_ring *ring,
+			       bool hdp_flush)
 {
-	radeon_ring_commit(rdev, ring);
+	radeon_ring_commit(rdev, ring, hdp_flush);
 	mutex_unlock(&rdev->ring_lock);
 }
 
@@ -612,7 +376,7 @@ int radeon_ring_restore(struct radeon_device *rdev, struct radeon_ring *ring,
 		radeon_ring_write(ring, data[i]);
 	}
 
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev, ring, false);
 	kfree(data);
 	return 0;
 }
@@ -640,7 +404,7 @@ int radeon_ring_init(struct radeon_device *rdev, struct radeon_ring *ring, unsig
 	/* Allocate ring buffer */
 	if (ring->ring_obj == NULL) {
 		r = radeon_bo_create(rdev, ring->ring_size, PAGE_SIZE, true,
-				     RADEON_GEM_DOMAIN_GTT,
+				     RADEON_GEM_DOMAIN_GTT, 0,
 				     NULL, &ring->ring_obj);
 		if (r) {
 			dev_err(rdev->dev, "(%d) ring create failed\n", r);
@@ -791,22 +555,6 @@ static struct drm_info_list radeon_debugfs_ring_info_list[] = {
 	{"radeon_ring_vce2", radeon_debugfs_ring_info, 0, &si_vce2_index},
 };
 
-static int radeon_debugfs_sa_info(struct seq_file *m, void *data)
-{
-	struct drm_info_node *node = (struct drm_info_node *) m->private;
-	struct drm_device *dev = node->minor->dev;
-	struct radeon_device *rdev = dev->dev_private;
-
-	radeon_sa_bo_dump_debug_info(&rdev->ring_tmp_bo, m);
-
-	return 0;
-
-}
-
-static struct drm_info_list radeon_debugfs_sa_list[] = {
-        {"radeon_sa_info", &radeon_debugfs_sa_info, 0, NULL},
-};
-
 #endif
 
 static int radeon_debugfs_ring_init(struct radeon_device *rdev, struct radeon_ring *ring)
@@ -828,12 +576,3 @@ static int radeon_debugfs_ring_init(struct radeon_device *rdev, struct radeon_ri
 #endif
 	return 0;
 }
-
-static int radeon_debugfs_sa_init(struct radeon_device *rdev)
-{
-#if defined(CONFIG_DEBUG_FS)
-	return radeon_debugfs_add_files(rdev, radeon_debugfs_sa_list, 1);
-#else
-	return 0;
-#endif
-}
diff --git a/drivers/gpu/drm/radeon/radeon_sa.c b/drivers/gpu/drm/radeon/radeon_sa.c
index adcf3e2f07da..b84f97c8718c 100644
--- a/drivers/gpu/drm/radeon/radeon_sa.c
+++ b/drivers/gpu/drm/radeon/radeon_sa.c
@@ -49,7 +49,7 @@ static void radeon_sa_bo_try_free(struct radeon_sa_manager *sa_manager);
 
 int radeon_sa_bo_manager_init(struct radeon_device *rdev,
 			      struct radeon_sa_manager *sa_manager,
-			      unsigned size, u32 align, u32 domain)
+			      unsigned size, u32 align, u32 domain, u32 flags)
 {
 	int i, r;
 
@@ -65,7 +65,7 @@ int radeon_sa_bo_manager_init(struct radeon_device *rdev,
 	}
 
 	r = radeon_bo_create(rdev, size, align, true,
-			     domain, NULL, &sa_manager->bo);
+			     domain, flags, NULL, &sa_manager->bo);
 	if (r) {
 		dev_err(rdev->dev, "(%d) failed to allocate bo for manager\n", r);
 		return r;
diff --git a/drivers/gpu/drm/radeon/radeon_semaphore.c b/drivers/gpu/drm/radeon/radeon_semaphore.c
index dbd6bcde92de..56d9fd66d8ae 100644
--- a/drivers/gpu/drm/radeon/radeon_semaphore.c
+++ b/drivers/gpu/drm/radeon/radeon_semaphore.c
@@ -179,7 +179,7 @@ int radeon_semaphore_sync_rings(struct radeon_device *rdev,
 			continue;
 		}
 
-		radeon_ring_commit(rdev, &rdev->ring[i]);
+		radeon_ring_commit(rdev, &rdev->ring[i], false);
 		radeon_fence_note_sync(fence, ring);
 
 		semaphore->gpu_addr += 8;
diff --git a/drivers/gpu/drm/radeon/radeon_test.c b/drivers/gpu/drm/radeon/radeon_test.c
index 3a13e0d1055c..17bc3dced9f1 100644
--- a/drivers/gpu/drm/radeon/radeon_test.c
+++ b/drivers/gpu/drm/radeon/radeon_test.c
@@ -56,13 +56,7 @@ static void radeon_do_test_moves(struct radeon_device *rdev, int flag)
 	/* Number of tests =
 	 * (Total GTT - IB pool - writeback page - ring buffers) / test size
 	 */
-	n = rdev->mc.gtt_size - RADEON_IB_POOL_SIZE*64*1024;
-	for (i = 0; i < RADEON_NUM_RINGS; ++i)
-		n -= rdev->ring[i].ring_size;
-	if (rdev->wb.wb_obj)
-		n -= RADEON_GPU_PAGE_SIZE;
-	if (rdev->ih.ring_obj)
-		n -= rdev->ih.ring_size;
+	n = rdev->mc.gtt_size - rdev->gart_pin_size;
 	n /= size;
 
 	gtt_obj = kzalloc(n * sizeof(*gtt_obj), GFP_KERNEL);
@@ -73,7 +67,7 @@ static void radeon_do_test_moves(struct radeon_device *rdev, int flag)
 	}
 
 	r = radeon_bo_create(rdev, size, PAGE_SIZE, true, RADEON_GEM_DOMAIN_VRAM,
-			     NULL, &vram_obj);
+			     0, NULL, &vram_obj);
 	if (r) {
 		DRM_ERROR("Failed to create VRAM object\n");
 		goto out_cleanup;
@@ -93,7 +87,7 @@ static void radeon_do_test_moves(struct radeon_device *rdev, int flag)
 		struct radeon_fence *fence = NULL;
 
 		r = radeon_bo_create(rdev, size, PAGE_SIZE, true,
-				     RADEON_GEM_DOMAIN_GTT, NULL, gtt_obj + i);
+				     RADEON_GEM_DOMAIN_GTT, 0, NULL, gtt_obj + i);
 		if (r) {
 			DRM_ERROR("Failed to create GTT object %d\n", i);
 			goto out_lclean;
@@ -294,7 +288,7 @@ static int radeon_test_create_and_emit_fence(struct radeon_device *rdev,
 			return r;
 		}
 		radeon_fence_emit(rdev, fence, ring->idx);
-		radeon_ring_unlock_commit(rdev, ring);
+		radeon_ring_unlock_commit(rdev, ring, false);
 	}
 	return 0;
 }
@@ -319,7 +313,7 @@ void radeon_test_ring_sync(struct radeon_device *rdev,
 		goto out_cleanup;
 	}
 	radeon_semaphore_emit_wait(rdev, ringA->idx, semaphore);
-	radeon_ring_unlock_commit(rdev, ringA);
+	radeon_ring_unlock_commit(rdev, ringA, false);
 
 	r = radeon_test_create_and_emit_fence(rdev, ringA, &fence1);
 	if (r)
@@ -331,7 +325,7 @@ void radeon_test_ring_sync(struct radeon_device *rdev,
 		goto out_cleanup;
 	}
 	radeon_semaphore_emit_wait(rdev, ringA->idx, semaphore);
-	radeon_ring_unlock_commit(rdev, ringA);
+	radeon_ring_unlock_commit(rdev, ringA, false);
 
 	r = radeon_test_create_and_emit_fence(rdev, ringA, &fence2);
 	if (r)
@@ -350,7 +344,7 @@ void radeon_test_ring_sync(struct radeon_device *rdev,
 		goto out_cleanup;
 	}
 	radeon_semaphore_emit_signal(rdev, ringB->idx, semaphore);
-	radeon_ring_unlock_commit(rdev, ringB);
+	radeon_ring_unlock_commit(rdev, ringB, false);
 
 	r = radeon_fence_wait(fence1, false);
 	if (r) {
@@ -371,7 +365,7 @@ void radeon_test_ring_sync(struct radeon_device *rdev,
 		goto out_cleanup;
 	}
 	radeon_semaphore_emit_signal(rdev, ringB->idx, semaphore);
-	radeon_ring_unlock_commit(rdev, ringB);
+	radeon_ring_unlock_commit(rdev, ringB, false);
 
 	r = radeon_fence_wait(fence2, false);
 	if (r) {
@@ -414,7 +408,7 @@ static void radeon_test_ring_sync2(struct radeon_device *rdev,
 		goto out_cleanup;
 	}
 	radeon_semaphore_emit_wait(rdev, ringA->idx, semaphore);
-	radeon_ring_unlock_commit(rdev, ringA);
+	radeon_ring_unlock_commit(rdev, ringA, false);
 
 	r = radeon_test_create_and_emit_fence(rdev, ringA, &fenceA);
 	if (r)
@@ -426,7 +420,7 @@ static void radeon_test_ring_sync2(struct radeon_device *rdev,
 		goto out_cleanup;
 	}
 	radeon_semaphore_emit_wait(rdev, ringB->idx, semaphore);
-	radeon_ring_unlock_commit(rdev, ringB);
+	radeon_ring_unlock_commit(rdev, ringB, false);
 	r = radeon_test_create_and_emit_fence(rdev, ringB, &fenceB);
 	if (r)
 		goto out_cleanup;
@@ -448,7 +442,7 @@ static void radeon_test_ring_sync2(struct radeon_device *rdev,
 		goto out_cleanup;
 	}
 	radeon_semaphore_emit_signal(rdev, ringC->idx, semaphore);
-	radeon_ring_unlock_commit(rdev, ringC);
+	radeon_ring_unlock_commit(rdev, ringC, false);
 
 	for (i = 0; i < 30; ++i) {
 		mdelay(100);
@@ -474,7 +468,7 @@ static void radeon_test_ring_sync2(struct radeon_device *rdev,
 		goto out_cleanup;
 	}
 	radeon_semaphore_emit_signal(rdev, ringC->idx, semaphore);
-	radeon_ring_unlock_commit(rdev, ringC);
+	radeon_ring_unlock_commit(rdev, ringC, false);
 
 	mdelay(1000);
 
diff --git a/drivers/gpu/drm/radeon/radeon_trace.h b/drivers/gpu/drm/radeon/radeon_trace.h
index f749f2c3bbdb..9db74a96ef61 100644
--- a/drivers/gpu/drm/radeon/radeon_trace.h
+++ b/drivers/gpu/drm/radeon/radeon_trace.h
@@ -72,8 +72,8 @@ TRACE_EVENT(radeon_vm_bo_update,
 			     ),
 
 	    TP_fast_assign(
-			   __entry->soffset = bo_va->soffset;
-			   __entry->eoffset = bo_va->eoffset;
+			   __entry->soffset = bo_va->it.start;
+			   __entry->eoffset = bo_va->it.last + 1;
 			   __entry->flags = bo_va->flags;
 			   ),
 	    TP_printk("soffs=%010llx, eoffs=%010llx, flags=%08x",
@@ -104,6 +104,24 @@ TRACE_EVENT(radeon_vm_set_page,
 		      __entry->flags, __entry->count)
 );
 
+TRACE_EVENT(radeon_vm_flush,
+	    TP_PROTO(uint64_t pd_addr, unsigned ring, unsigned id),
+	    TP_ARGS(pd_addr, ring, id),
+	    TP_STRUCT__entry(
+			     __field(u64, pd_addr)
+			     __field(u32, ring)
+			     __field(u32, id)
+			     ),
+
+	    TP_fast_assign(
+			   __entry->pd_addr = pd_addr;
+			   __entry->ring = ring;
+			   __entry->id = id;
+			   ),
+	    TP_printk("pd_addr=%010Lx, ring=%u, id=%u",
+		      __entry->pd_addr, __entry->ring, __entry->id)
+);
+
 DECLARE_EVENT_CLASS(radeon_fence_request,
 
 	    TP_PROTO(struct drm_device *dev, int ring, u32 seqno),
diff --git a/drivers/gpu/drm/radeon/radeon_ttm.c b/drivers/gpu/drm/radeon/radeon_ttm.c
index c8a8a5144ec1..72afe82a95c9 100644
--- a/drivers/gpu/drm/radeon/radeon_ttm.c
+++ b/drivers/gpu/drm/radeon/radeon_ttm.c
@@ -521,6 +521,8 @@ static int radeon_ttm_backend_bind(struct ttm_tt *ttm,
 				   struct ttm_mem_reg *bo_mem)
 {
 	struct radeon_ttm_tt *gtt = (void*)ttm;
+	uint32_t flags = RADEON_GART_PAGE_VALID | RADEON_GART_PAGE_READ |
+		RADEON_GART_PAGE_WRITE;
 	int r;
 
 	gtt->offset = (unsigned long)(bo_mem->start << PAGE_SHIFT);
@@ -528,8 +530,10 @@ static int radeon_ttm_backend_bind(struct ttm_tt *ttm,
 		WARN(1, "nothing to bind %lu pages for mreg %p back %p!\n",
 		     ttm->num_pages, bo_mem, ttm);
 	}
-	r = radeon_gart_bind(gtt->rdev, gtt->offset,
-			     ttm->num_pages, ttm->pages, gtt->ttm.dma_address);
+	if (ttm->caching_state == tt_cached)
+		flags |= RADEON_GART_PAGE_SNOOP;
+	r = radeon_gart_bind(gtt->rdev, gtt->offset, ttm->num_pages,
+			     ttm->pages, gtt->ttm.dma_address, flags);
 	if (r) {
 		DRM_ERROR("failed to bind %lu pages at 0x%08X\n",
 			  ttm->num_pages, (unsigned)gtt->offset);
@@ -726,7 +730,7 @@ int radeon_ttm_init(struct radeon_device *rdev)
 	radeon_ttm_set_active_vram_size(rdev, rdev->mc.visible_vram_size);
 
 	r = radeon_bo_create(rdev, 256 * 1024, PAGE_SIZE, true,
-			     RADEON_GEM_DOMAIN_VRAM,
+			     RADEON_GEM_DOMAIN_VRAM, 0,
 			     NULL, &rdev->stollen_vga_memory);
 	if (r) {
 		return r;
diff --git a/drivers/gpu/drm/radeon/radeon_ucode.c b/drivers/gpu/drm/radeon/radeon_ucode.c
new file mode 100644
index 000000000000..6beec680390c
--- /dev/null
+++ b/drivers/gpu/drm/radeon/radeon_ucode.c
@@ -0,0 +1,167 @@
+/*
+ * Copyright 2014 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/firmware.h>
+#include <linux/slab.h>
+#include <linux/module.h>
+#include <drm/drmP.h>
+#include "radeon.h"
+#include "radeon_ucode.h"
+
+static void radeon_ucode_print_common_hdr(const struct common_firmware_header *hdr)
+{
+	DRM_DEBUG("size_bytes: %u\n", le32_to_cpu(hdr->size_bytes));
+	DRM_DEBUG("header_size_bytes: %u\n", le32_to_cpu(hdr->header_size_bytes));
+	DRM_DEBUG("header_version_major: %u\n", le16_to_cpu(hdr->header_version_major));
+	DRM_DEBUG("header_version_minor: %u\n", le16_to_cpu(hdr->header_version_minor));
+	DRM_DEBUG("ip_version_major: %u\n", le16_to_cpu(hdr->ip_version_major));
+	DRM_DEBUG("ip_version_minor: %u\n", le16_to_cpu(hdr->ip_version_minor));
+	DRM_DEBUG("ucode_version: 0x%08x\n", le32_to_cpu(hdr->ucode_version));
+	DRM_DEBUG("ucode_size_bytes: %u\n", le32_to_cpu(hdr->ucode_size_bytes));
+	DRM_DEBUG("ucode_array_offset_bytes: %u\n",
+		  le32_to_cpu(hdr->ucode_array_offset_bytes));
+	DRM_DEBUG("crc32: 0x%08x\n", le32_to_cpu(hdr->crc32));
+}
+
+void radeon_ucode_print_mc_hdr(const struct common_firmware_header *hdr)
+{
+	uint16_t version_major = le16_to_cpu(hdr->header_version_major);
+	uint16_t version_minor = le16_to_cpu(hdr->header_version_minor);
+
+	DRM_DEBUG("MC\n");
+	radeon_ucode_print_common_hdr(hdr);
+
+	if (version_major == 1) {
+		const struct mc_firmware_header_v1_0 *mc_hdr =
+			container_of(hdr, struct mc_firmware_header_v1_0, header);
+
+		DRM_DEBUG("io_debug_size_bytes: %u\n",
+			  le32_to_cpu(mc_hdr->io_debug_size_bytes));
+		DRM_DEBUG("io_debug_array_offset_bytes: %u\n",
+			  le32_to_cpu(mc_hdr->io_debug_array_offset_bytes));
+	} else {
+		DRM_ERROR("Unknown MC ucode version: %u.%u\n", version_major, version_minor);
+	}
+}
+
+void radeon_ucode_print_smc_hdr(const struct common_firmware_header *hdr)
+{
+	uint16_t version_major = le16_to_cpu(hdr->header_version_major);
+	uint16_t version_minor = le16_to_cpu(hdr->header_version_minor);
+
+	DRM_DEBUG("SMC\n");
+	radeon_ucode_print_common_hdr(hdr);
+
+	if (version_major == 1) {
+		const struct smc_firmware_header_v1_0 *smc_hdr =
+			container_of(hdr, struct smc_firmware_header_v1_0, header);
+
+		DRM_DEBUG("ucode_start_addr: %u\n", le32_to_cpu(smc_hdr->ucode_start_addr));
+	} else {
+		DRM_ERROR("Unknown SMC ucode version: %u.%u\n", version_major, version_minor);
+	}
+}
+
+void radeon_ucode_print_gfx_hdr(const struct common_firmware_header *hdr)
+{
+	uint16_t version_major = le16_to_cpu(hdr->header_version_major);
+	uint16_t version_minor = le16_to_cpu(hdr->header_version_minor);
+
+	DRM_DEBUG("GFX\n");
+	radeon_ucode_print_common_hdr(hdr);
+
+	if (version_major == 1) {
+		const struct gfx_firmware_header_v1_0 *gfx_hdr =
+			container_of(hdr, struct gfx_firmware_header_v1_0, header);
+
+		DRM_DEBUG("ucode_feature_version: %u\n",
+			  le32_to_cpu(gfx_hdr->ucode_feature_version));
+		DRM_DEBUG("jt_offset: %u\n", le32_to_cpu(gfx_hdr->jt_offset));
+		DRM_DEBUG("jt_size: %u\n", le32_to_cpu(gfx_hdr->jt_size));
+	} else {
+		DRM_ERROR("Unknown GFX ucode version: %u.%u\n", version_major, version_minor);
+	}
+}
+
+void radeon_ucode_print_rlc_hdr(const struct common_firmware_header *hdr)
+{
+	uint16_t version_major = le16_to_cpu(hdr->header_version_major);
+	uint16_t version_minor = le16_to_cpu(hdr->header_version_minor);
+
+	DRM_DEBUG("RLC\n");
+	radeon_ucode_print_common_hdr(hdr);
+
+	if (version_major == 1) {
+		const struct rlc_firmware_header_v1_0 *rlc_hdr =
+			container_of(hdr, struct rlc_firmware_header_v1_0, header);
+
+		DRM_DEBUG("ucode_feature_version: %u\n",
+			  le32_to_cpu(rlc_hdr->ucode_feature_version));
+		DRM_DEBUG("save_and_restore_offset: %u\n",
+			  le32_to_cpu(rlc_hdr->save_and_restore_offset));
+		DRM_DEBUG("clear_state_descriptor_offset: %u\n",
+			  le32_to_cpu(rlc_hdr->clear_state_descriptor_offset));
+		DRM_DEBUG("avail_scratch_ram_locations: %u\n",
+			  le32_to_cpu(rlc_hdr->avail_scratch_ram_locations));
+		DRM_DEBUG("master_pkt_description_offset: %u\n",
+			  le32_to_cpu(rlc_hdr->master_pkt_description_offset));
+	} else {
+		DRM_ERROR("Unknown RLC ucode version: %u.%u\n", version_major, version_minor);
+	}
+}
+
+void radeon_ucode_print_sdma_hdr(const struct common_firmware_header *hdr)
+{
+	uint16_t version_major = le16_to_cpu(hdr->header_version_major);
+	uint16_t version_minor = le16_to_cpu(hdr->header_version_minor);
+
+	DRM_DEBUG("SDMA\n");
+	radeon_ucode_print_common_hdr(hdr);
+
+	if (version_major == 1) {
+		const struct sdma_firmware_header_v1_0 *sdma_hdr =
+			container_of(hdr, struct sdma_firmware_header_v1_0, header);
+
+		DRM_DEBUG("ucode_feature_version: %u\n",
+			  le32_to_cpu(sdma_hdr->ucode_feature_version));
+		DRM_DEBUG("ucode_change_version: %u\n",
+			  le32_to_cpu(sdma_hdr->ucode_change_version));
+		DRM_DEBUG("jt_offset: %u\n", le32_to_cpu(sdma_hdr->jt_offset));
+		DRM_DEBUG("jt_size: %u\n", le32_to_cpu(sdma_hdr->jt_size));
+	} else {
+		DRM_ERROR("Unknown SDMA ucode version: %u.%u\n",
+			  version_major, version_minor);
+	}
+}
+
+int radeon_ucode_validate(const struct firmware *fw)
+{
+	const struct common_firmware_header *hdr =
+		(const struct common_firmware_header *)fw->data;
+
+	if (fw->size == le32_to_cpu(hdr->size_bytes))
+		return 0;
+
+	return -EINVAL;
+}
+
diff --git a/drivers/gpu/drm/radeon/radeon_ucode.h b/drivers/gpu/drm/radeon/radeon_ucode.h
index 4e7c3269b183..dc4576e4d8ad 100644
--- a/drivers/gpu/drm/radeon/radeon_ucode.h
+++ b/drivers/gpu/drm/radeon/radeon_ucode.h
@@ -153,4 +153,75 @@
 #define HAWAII_SMC_UCODE_START       0x20000
 #define HAWAII_SMC_UCODE_SIZE        0x1FDEC
 
+struct common_firmware_header {
+	uint32_t size_bytes; /* size of the entire header+image(s) in bytes */
+	uint32_t header_size_bytes; /* size of just the header in bytes */
+	uint16_t header_version_major; /* header version */
+	uint16_t header_version_minor; /* header version */
+	uint16_t ip_version_major; /* IP version */
+	uint16_t ip_version_minor; /* IP version */
+	uint32_t ucode_version;
+	uint32_t ucode_size_bytes; /* size of ucode in bytes */
+	uint32_t ucode_array_offset_bytes; /* payload offset from the start of the header */
+	uint32_t crc32;  /* crc32 checksum of the payload */
+};
+
+/* version_major=1, version_minor=0 */
+struct mc_firmware_header_v1_0 {
+	struct common_firmware_header header;
+	uint32_t io_debug_size_bytes; /* size of debug array in dwords */
+	uint32_t io_debug_array_offset_bytes; /* payload offset from the start of the header */
+};
+
+/* version_major=1, version_minor=0 */
+struct smc_firmware_header_v1_0 {
+	struct common_firmware_header header;
+	uint32_t ucode_start_addr;
+};
+
+/* version_major=1, version_minor=0 */
+struct gfx_firmware_header_v1_0 {
+	struct common_firmware_header header;
+	uint32_t ucode_feature_version;
+	uint32_t jt_offset; /* jt location */
+	uint32_t jt_size;  /* size of jt */
+};
+
+/* version_major=1, version_minor=0 */
+struct rlc_firmware_header_v1_0 {
+	struct common_firmware_header header;
+	uint32_t ucode_feature_version;
+	uint32_t save_and_restore_offset;
+	uint32_t clear_state_descriptor_offset;
+	uint32_t avail_scratch_ram_locations;
+	uint32_t master_pkt_description_offset;
+};
+
+/* version_major=1, version_minor=0 */
+struct sdma_firmware_header_v1_0 {
+	struct common_firmware_header header;
+	uint32_t ucode_feature_version;
+	uint32_t ucode_change_version;
+	uint32_t jt_offset; /* jt location */
+	uint32_t jt_size; /* size of jt */
+};
+
+/* header is fixed size */
+union radeon_firmware_header {
+	struct common_firmware_header common;
+	struct mc_firmware_header_v1_0 mc;
+	struct smc_firmware_header_v1_0 smc;
+	struct gfx_firmware_header_v1_0 gfx;
+	struct rlc_firmware_header_v1_0 rlc;
+	struct sdma_firmware_header_v1_0 sdma;
+	uint8_t raw[0x100];
+};
+
+void radeon_ucode_print_mc_hdr(const struct common_firmware_header *hdr);
+void radeon_ucode_print_smc_hdr(const struct common_firmware_header *hdr);
+void radeon_ucode_print_gfx_hdr(const struct common_firmware_header *hdr);
+void radeon_ucode_print_rlc_hdr(const struct common_firmware_header *hdr);
+void radeon_ucode_print_sdma_hdr(const struct common_firmware_header *hdr);
+int radeon_ucode_validate(const struct firmware *fw);
+
 #endif
diff --git a/drivers/gpu/drm/radeon/radeon_uvd.c b/drivers/gpu/drm/radeon/radeon_uvd.c
index a4ad270e8261..341848a14376 100644
--- a/drivers/gpu/drm/radeon/radeon_uvd.c
+++ b/drivers/gpu/drm/radeon/radeon_uvd.c
@@ -117,7 +117,7 @@ int radeon_uvd_init(struct radeon_device *rdev)
 	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);
+			     RADEON_GEM_DOMAIN_VRAM, 0, NULL, &rdev->uvd.vcpu_bo);
 	if (r) {
 		dev_err(rdev->dev, "(%d) failed to allocate UVD bo\n", r);
 		return r;
@@ -646,7 +646,7 @@ static int radeon_uvd_send_msg(struct radeon_device *rdev,
 		ib.ptr[i] = PACKET2(0);
 	ib.length_dw = 16;
 
-	r = radeon_ib_schedule(rdev, &ib, NULL);
+	r = radeon_ib_schedule(rdev, &ib, NULL, false);
 	if (r)
 		goto err;
 	ttm_eu_fence_buffer_objects(&ticket, &head, ib.fence);
@@ -674,7 +674,7 @@ int radeon_uvd_get_create_msg(struct radeon_device *rdev, int ring,
 	int r, i;
 
 	r = radeon_bo_create(rdev, 1024, PAGE_SIZE, true,
-			     RADEON_GEM_DOMAIN_VRAM, NULL, &bo);
+			     RADEON_GEM_DOMAIN_VRAM, 0, NULL, &bo);
 	if (r)
 		return r;
 
@@ -720,7 +720,7 @@ int radeon_uvd_get_destroy_msg(struct radeon_device *rdev, int ring,
 	int r, i;
 
 	r = radeon_bo_create(rdev, 1024, PAGE_SIZE, true,
-			     RADEON_GEM_DOMAIN_VRAM, NULL, &bo);
+			     RADEON_GEM_DOMAIN_VRAM, 0, NULL, &bo);
 	if (r)
 		return r;
 
diff --git a/drivers/gpu/drm/radeon/radeon_vce.c b/drivers/gpu/drm/radeon/radeon_vce.c
index aa21c31a846c..c7190aadbd89 100644
--- a/drivers/gpu/drm/radeon/radeon_vce.c
+++ b/drivers/gpu/drm/radeon/radeon_vce.c
@@ -126,7 +126,7 @@ int radeon_vce_init(struct radeon_device *rdev)
 	size = RADEON_GPU_PAGE_ALIGN(rdev->vce_fw->size) +
 	       RADEON_VCE_STACK_SIZE + RADEON_VCE_HEAP_SIZE;
 	r = radeon_bo_create(rdev, size, PAGE_SIZE, true,
-			     RADEON_GEM_DOMAIN_VRAM, NULL, &rdev->vce.vcpu_bo);
+			     RADEON_GEM_DOMAIN_VRAM, 0, NULL, &rdev->vce.vcpu_bo);
 	if (r) {
 		dev_err(rdev->dev, "(%d) failed to allocate VCE bo\n", r);
 		return r;
@@ -368,7 +368,7 @@ int radeon_vce_get_create_msg(struct radeon_device *rdev, int ring,
 	for (i = ib.length_dw; i < ib_size_dw; ++i)
 		ib.ptr[i] = 0x0;
 
-	r = radeon_ib_schedule(rdev, &ib, NULL);
+	r = radeon_ib_schedule(rdev, &ib, NULL, false);
 	if (r) {
 	        DRM_ERROR("radeon: failed to schedule ib (%d).\n", r);
 	}
@@ -425,7 +425,7 @@ int radeon_vce_get_destroy_msg(struct radeon_device *rdev, int ring,
 	for (i = ib.length_dw; i < ib_size_dw; ++i)
 		ib.ptr[i] = 0x0;
 
-	r = radeon_ib_schedule(rdev, &ib, NULL);
+	r = radeon_ib_schedule(rdev, &ib, NULL, false);
 	if (r) {
 	        DRM_ERROR("radeon: failed to schedule ib (%d).\n", r);
 	}
@@ -715,7 +715,7 @@ int radeon_vce_ring_test(struct radeon_device *rdev, struct radeon_ring *ring)
 		return r;
 	}
 	radeon_ring_write(ring, VCE_CMD_END);
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev, ring, false);
 
 	for (i = 0; i < rdev->usec_timeout; i++) {
 	        if (vce_v1_0_get_rptr(rdev, ring) != rptr)
diff --git a/drivers/gpu/drm/radeon/radeon_vm.c b/drivers/gpu/drm/radeon/radeon_vm.c
index eecff6bbd341..088ffdc2f577 100644
--- a/drivers/gpu/drm/radeon/radeon_vm.c
+++ b/drivers/gpu/drm/radeon/radeon_vm.c
@@ -238,8 +238,8 @@ void radeon_vm_flush(struct radeon_device *rdev,
 	uint64_t pd_addr = radeon_bo_gpu_offset(vm->page_directory);
 
 	/* if we can't remember our last VM flush then flush now! */
-	/* XXX figure out why we have to flush all the time */
-	if (!vm->last_flush || true || pd_addr != vm->pd_gpu_addr) {
+	if (!vm->last_flush || pd_addr != vm->pd_gpu_addr) {
+		trace_radeon_vm_flush(pd_addr, ring, vm->id);
 		vm->pd_gpu_addr = pd_addr;
 		radeon_ring_vm_flush(rdev, ring, vm);
 	}
@@ -325,16 +325,15 @@ struct radeon_bo_va *radeon_vm_bo_add(struct radeon_device *rdev,
 	}
 	bo_va->vm = vm;
 	bo_va->bo = bo;
-	bo_va->soffset = 0;
-	bo_va->eoffset = 0;
+	bo_va->it.start = 0;
+	bo_va->it.last = 0;
 	bo_va->flags = 0;
-	bo_va->valid = false;
+	bo_va->addr = 0;
 	bo_va->ref_count = 1;
 	INIT_LIST_HEAD(&bo_va->bo_list);
-	INIT_LIST_HEAD(&bo_va->vm_list);
+	INIT_LIST_HEAD(&bo_va->vm_status);
 
 	mutex_lock(&vm->mutex);
-	list_add(&bo_va->vm_list, &vm->va);
 	list_add_tail(&bo_va->bo_list, &bo->va);
 	mutex_unlock(&vm->mutex);
 
@@ -342,6 +341,42 @@ struct radeon_bo_va *radeon_vm_bo_add(struct radeon_device *rdev,
 }
 
 /**
+ * radeon_vm_set_pages - helper to call the right asic function
+ *
+ * @rdev: radeon_device pointer
+ * @ib: indirect buffer to fill with commands
+ * @pe: addr of the page entry
+ * @addr: dst addr to write into pe
+ * @count: number of page entries to update
+ * @incr: increase next addr by incr bytes
+ * @flags: hw access flags
+ *
+ * Traces the parameters and calls the right asic functions
+ * to setup the page table using the DMA.
+ */
+static void radeon_vm_set_pages(struct radeon_device *rdev,
+				struct radeon_ib *ib,
+				uint64_t pe,
+				uint64_t addr, unsigned count,
+				uint32_t incr, uint32_t flags)
+{
+	trace_radeon_vm_set_page(pe, addr, count, incr, flags);
+
+	if ((flags & R600_PTE_GART_MASK) == R600_PTE_GART_MASK) {
+		uint64_t src = rdev->gart.table_addr + (addr >> 12) * 8;
+		radeon_asic_vm_copy_pages(rdev, ib, pe, src, count);
+
+	} else if ((flags & R600_PTE_SYSTEM) || (count < 3)) {
+		radeon_asic_vm_write_pages(rdev, ib, pe, addr,
+					   count, incr, flags);
+
+	} else {
+		radeon_asic_vm_set_pages(rdev, ib, pe, addr,
+					 count, incr, flags);
+	}
+}
+
+/**
  * radeon_vm_clear_bo - initially clear the page dir/table
  *
  * @rdev: radeon_device pointer
@@ -375,16 +410,17 @@ static int radeon_vm_clear_bo(struct radeon_device *rdev,
 	addr = radeon_bo_gpu_offset(bo);
 	entries = radeon_bo_size(bo) / 8;
 
-	r = radeon_ib_get(rdev, R600_RING_TYPE_DMA_INDEX, &ib,
-			  NULL, entries * 2 + 64);
+	r = radeon_ib_get(rdev, R600_RING_TYPE_DMA_INDEX, &ib, NULL, 256);
 	if (r)
                 goto error;
 
 	ib.length_dw = 0;
 
-	radeon_asic_vm_set_page(rdev, &ib, addr, 0, entries, 0, 0);
+	radeon_vm_set_pages(rdev, &ib, addr, 0, entries, 0, 0);
+	radeon_asic_vm_pad_ib(rdev, &ib);
+	WARN_ON(ib.length_dw > 64);
 
-	r = radeon_ib_schedule(rdev, &ib, NULL);
+	r = radeon_ib_schedule(rdev, &ib, NULL, false);
 	if (r)
                 goto error;
 
@@ -418,11 +454,9 @@ int radeon_vm_bo_set_addr(struct radeon_device *rdev,
 			  uint32_t flags)
 {
 	uint64_t size = radeon_bo_size(bo_va->bo);
-	uint64_t eoffset, last_offset = 0;
 	struct radeon_vm *vm = bo_va->vm;
-	struct radeon_bo_va *tmp;
-	struct list_head *head;
 	unsigned last_pfn, pt_idx;
+	uint64_t eoffset;
 	int r;
 
 	if (soffset) {
@@ -444,38 +478,53 @@ int radeon_vm_bo_set_addr(struct radeon_device *rdev,
 	}
 
 	mutex_lock(&vm->mutex);
-	head = &vm->va;
-	last_offset = 0;
-	list_for_each_entry(tmp, &vm->va, vm_list) {
-		if (bo_va == tmp) {
-			/* skip over currently modified bo */
-			continue;
+	if (bo_va->it.start || bo_va->it.last) {
+		if (bo_va->addr) {
+			/* add a clone of the bo_va to clear the old address */
+			struct radeon_bo_va *tmp;
+			tmp = kzalloc(sizeof(struct radeon_bo_va), GFP_KERNEL);
+			if (!tmp) {
+				mutex_unlock(&vm->mutex);
+				return -ENOMEM;
+			}
+			tmp->it.start = bo_va->it.start;
+			tmp->it.last = bo_va->it.last;
+			tmp->vm = vm;
+			tmp->addr = bo_va->addr;
+			tmp->bo = radeon_bo_ref(bo_va->bo);
+			list_add(&tmp->vm_status, &vm->freed);
 		}
 
-		if (soffset >= last_offset && eoffset <= tmp->soffset) {
-			/* bo can be added before this one */
-			break;
-		}
-		if (eoffset > tmp->soffset && soffset < tmp->eoffset) {
+		interval_tree_remove(&bo_va->it, &vm->va);
+		bo_va->it.start = 0;
+		bo_va->it.last = 0;
+	}
+
+	soffset /= RADEON_GPU_PAGE_SIZE;
+	eoffset /= RADEON_GPU_PAGE_SIZE;
+	if (soffset || eoffset) {
+		struct interval_tree_node *it;
+		it = interval_tree_iter_first(&vm->va, soffset, eoffset - 1);
+		if (it) {
+			struct radeon_bo_va *tmp;
+			tmp = container_of(it, struct radeon_bo_va, it);
 			/* bo and tmp overlap, invalid offset */
-			dev_err(rdev->dev, "bo %p va 0x%08X conflict with (bo %p 0x%08X 0x%08X)\n",
-				bo_va->bo, (unsigned)bo_va->soffset, tmp->bo,
-				(unsigned)tmp->soffset, (unsigned)tmp->eoffset);
+			dev_err(rdev->dev, "bo %p va 0x%010Lx conflict with "
+				"(bo %p 0x%010lx 0x%010lx)\n", bo_va->bo,
+				soffset, tmp->bo, tmp->it.start, tmp->it.last);
 			mutex_unlock(&vm->mutex);
 			return -EINVAL;
 		}
-		last_offset = tmp->eoffset;
-		head = &tmp->vm_list;
+		bo_va->it.start = soffset;
+		bo_va->it.last = eoffset - 1;
+		interval_tree_insert(&bo_va->it, &vm->va);
 	}
 
-	bo_va->soffset = soffset;
-	bo_va->eoffset = eoffset;
 	bo_va->flags = flags;
-	bo_va->valid = false;
-	list_move(&bo_va->vm_list, head);
+	bo_va->addr = 0;
 
-	soffset = (soffset / RADEON_GPU_PAGE_SIZE) >> radeon_vm_block_size;
-	eoffset = (eoffset / RADEON_GPU_PAGE_SIZE) >> radeon_vm_block_size;
+	soffset >>= radeon_vm_block_size;
+	eoffset >>= radeon_vm_block_size;
 
 	BUG_ON(eoffset >= radeon_vm_num_pdes(rdev));
 
@@ -496,7 +545,7 @@ int radeon_vm_bo_set_addr(struct radeon_device *rdev,
 
 		r = radeon_bo_create(rdev, RADEON_VM_PTE_COUNT * 8,
 				     RADEON_GPU_PAGE_SIZE, true,
-				     RADEON_GEM_DOMAIN_VRAM, NULL, &pt);
+				     RADEON_GEM_DOMAIN_VRAM, 0, NULL, &pt);
 		if (r)
 			return r;
 
@@ -597,7 +646,7 @@ int radeon_vm_update_page_directory(struct radeon_device *rdev,
 	ndw = 64;
 
 	/* assume the worst case */
-	ndw += vm->max_pde_used * 16;
+	ndw += vm->max_pde_used * 6;
 
 	/* update too big for an IB */
 	if (ndw > 0xfffff)
@@ -626,9 +675,9 @@ int radeon_vm_update_page_directory(struct radeon_device *rdev,
 		    ((last_pt + incr * count) != pt)) {
 
 			if (count) {
-				radeon_asic_vm_set_page(rdev, &ib, last_pde,
-							last_pt, count, incr,
-							R600_PTE_VALID);
+				radeon_vm_set_pages(rdev, &ib, last_pde,
+						    last_pt, count, incr,
+						    R600_PTE_VALID);
 			}
 
 			count = 1;
@@ -640,13 +689,15 @@ int radeon_vm_update_page_directory(struct radeon_device *rdev,
 	}
 
 	if (count)
-		radeon_asic_vm_set_page(rdev, &ib, last_pde, last_pt, count,
-					incr, R600_PTE_VALID);
+		radeon_vm_set_pages(rdev, &ib, last_pde, last_pt, count,
+				    incr, R600_PTE_VALID);
 
 	if (ib.length_dw != 0) {
+		radeon_asic_vm_pad_ib(rdev, &ib);
 		radeon_semaphore_sync_to(ib.semaphore, pd->tbo.sync_obj);
 		radeon_semaphore_sync_to(ib.semaphore, vm->last_id_use);
-		r = radeon_ib_schedule(rdev, &ib, NULL);
+		WARN_ON(ib.length_dw > ndw);
+		r = radeon_ib_schedule(rdev, &ib, NULL, false);
 		if (r) {
 			radeon_ib_free(rdev, &ib);
 			return r;
@@ -711,30 +762,30 @@ static void radeon_vm_frag_ptes(struct radeon_device *rdev,
 	    (frag_start >= frag_end)) {
 
 		count = (pe_end - pe_start) / 8;
-		radeon_asic_vm_set_page(rdev, ib, pe_start, addr, count,
-					RADEON_GPU_PAGE_SIZE, flags);
+		radeon_vm_set_pages(rdev, ib, pe_start, addr, count,
+				    RADEON_GPU_PAGE_SIZE, flags);
 		return;
 	}
 
 	/* handle the 4K area at the beginning */
 	if (pe_start != frag_start) {
 		count = (frag_start - pe_start) / 8;
-		radeon_asic_vm_set_page(rdev, ib, pe_start, addr, count,
-					RADEON_GPU_PAGE_SIZE, flags);
+		radeon_vm_set_pages(rdev, ib, pe_start, addr, count,
+				    RADEON_GPU_PAGE_SIZE, flags);
 		addr += RADEON_GPU_PAGE_SIZE * count;
 	}
 
 	/* handle the area in the middle */
 	count = (frag_end - frag_start) / 8;
-	radeon_asic_vm_set_page(rdev, ib, frag_start, addr, count,
-				RADEON_GPU_PAGE_SIZE, flags | frag_flags);
+	radeon_vm_set_pages(rdev, ib, frag_start, addr, count,
+			    RADEON_GPU_PAGE_SIZE, flags | frag_flags);
 
 	/* handle the 4K area at the end */
 	if (frag_end != pe_end) {
 		addr += RADEON_GPU_PAGE_SIZE * count;
 		count = (pe_end - frag_end) / 8;
-		radeon_asic_vm_set_page(rdev, ib, frag_end, addr, count,
-					RADEON_GPU_PAGE_SIZE, flags);
+		radeon_vm_set_pages(rdev, ib, frag_end, addr, count,
+				    RADEON_GPU_PAGE_SIZE, flags);
 	}
 }
 
@@ -763,9 +814,6 @@ static void radeon_vm_update_ptes(struct radeon_device *rdev,
 	unsigned count = 0;
 	uint64_t addr;
 
-	start = start / RADEON_GPU_PAGE_SIZE;
-	end = end / RADEON_GPU_PAGE_SIZE;
-
 	/* walk over the address space and update the page tables */
 	for (addr = start; addr < end; ) {
 		uint64_t pt_idx = addr >> radeon_vm_block_size;
@@ -823,66 +871,78 @@ static void radeon_vm_update_ptes(struct radeon_device *rdev,
  * Object have to be reserved and mutex must be locked!
  */
 int radeon_vm_bo_update(struct radeon_device *rdev,
-			struct radeon_vm *vm,
-			struct radeon_bo *bo,
+			struct radeon_bo_va *bo_va,
 			struct ttm_mem_reg *mem)
 {
+	struct radeon_vm *vm = bo_va->vm;
 	struct radeon_ib ib;
-	struct radeon_bo_va *bo_va;
-	unsigned nptes, ndw;
+	unsigned nptes, ncmds, ndw;
 	uint64_t addr;
+	uint32_t flags;
 	int r;
 
-	bo_va = radeon_vm_bo_find(vm, bo);
-	if (bo_va == NULL) {
-		dev_err(rdev->dev, "bo %p not in vm %p\n", bo, vm);
-		return -EINVAL;
-	}
-
-	if (!bo_va->soffset) {
+	if (!bo_va->it.start) {
 		dev_err(rdev->dev, "bo %p don't has a mapping in vm %p\n",
-			bo, vm);
+			bo_va->bo, vm);
 		return -EINVAL;
 	}
 
-	if ((bo_va->valid && mem) || (!bo_va->valid && mem == NULL))
-		return 0;
+	list_del_init(&bo_va->vm_status);
 
 	bo_va->flags &= ~RADEON_VM_PAGE_VALID;
 	bo_va->flags &= ~RADEON_VM_PAGE_SYSTEM;
+	bo_va->flags &= ~RADEON_VM_PAGE_SNOOPED;
 	if (mem) {
 		addr = mem->start << PAGE_SHIFT;
 		if (mem->mem_type != TTM_PL_SYSTEM) {
 			bo_va->flags |= RADEON_VM_PAGE_VALID;
-			bo_va->valid = true;
 		}
 		if (mem->mem_type == TTM_PL_TT) {
 			bo_va->flags |= RADEON_VM_PAGE_SYSTEM;
+			if (!(bo_va->bo->flags & (RADEON_GEM_GTT_WC | RADEON_GEM_GTT_UC)))
+				bo_va->flags |= RADEON_VM_PAGE_SNOOPED;
+
 		} else {
 			addr += rdev->vm_manager.vram_base_offset;
 		}
 	} else {
 		addr = 0;
-		bo_va->valid = false;
 	}
 
+	if (addr == bo_va->addr)
+		return 0;
+	bo_va->addr = addr;
+
 	trace_radeon_vm_bo_update(bo_va);
 
-	nptes = radeon_bo_ngpu_pages(bo);
+	nptes = bo_va->it.last - bo_va->it.start + 1;
+
+	/* reserve space for one command every (1 << BLOCK_SIZE) entries
+	   or 2k dwords (whatever is smaller) */
+	ncmds = (nptes >> min(radeon_vm_block_size, 11)) + 1;
 
 	/* padding, etc. */
 	ndw = 64;
 
-	if (radeon_vm_block_size > 11)
-		/* reserve space for one header for every 2k dwords */
-		ndw += (nptes >> 11) * 4;
-	else
-		/* reserve space for one header for
-		    every (1 << BLOCK_SIZE) entries */
-		ndw += (nptes >> radeon_vm_block_size) * 4;
+	flags = radeon_vm_page_flags(bo_va->flags);
+	if ((flags & R600_PTE_GART_MASK) == R600_PTE_GART_MASK) {
+		/* only copy commands needed */
+		ndw += ncmds * 7;
+
+	} else if (flags & R600_PTE_SYSTEM) {
+		/* header for write data commands */
+		ndw += ncmds * 4;
+
+		/* body of write data command */
+		ndw += nptes * 2;
 
-	/* reserve space for pte addresses */
-	ndw += nptes * 2;
+	} else {
+		/* set page commands needed */
+		ndw += ncmds * 10;
+
+		/* two extra commands for begin/end of fragment */
+		ndw += 2 * 10;
+	}
 
 	/* update too big for an IB */
 	if (ndw > 0xfffff)
@@ -893,11 +953,15 @@ int radeon_vm_bo_update(struct radeon_device *rdev,
 		return r;
 	ib.length_dw = 0;
 
-	radeon_vm_update_ptes(rdev, vm, &ib, bo_va->soffset, bo_va->eoffset,
-			      addr, radeon_vm_page_flags(bo_va->flags));
+	radeon_vm_update_ptes(rdev, vm, &ib, bo_va->it.start,
+			      bo_va->it.last + 1, addr,
+			      radeon_vm_page_flags(bo_va->flags));
+
+	radeon_asic_vm_pad_ib(rdev, &ib);
+	WARN_ON(ib.length_dw > ndw);
 
 	radeon_semaphore_sync_to(ib.semaphore, vm->fence);
-	r = radeon_ib_schedule(rdev, &ib, NULL);
+	r = radeon_ib_schedule(rdev, &ib, NULL, false);
 	if (r) {
 		radeon_ib_free(rdev, &ib);
 		return r;
@@ -911,33 +975,87 @@ int radeon_vm_bo_update(struct radeon_device *rdev,
 }
 
 /**
+ * radeon_vm_clear_freed - clear freed BOs in the PT
+ *
+ * @rdev: radeon_device pointer
+ * @vm: requested vm
+ *
+ * Make sure all freed BOs are cleared in the PT.
+ * Returns 0 for success.
+ *
+ * PTs have to be reserved and mutex must be locked!
+ */
+int radeon_vm_clear_freed(struct radeon_device *rdev,
+			  struct radeon_vm *vm)
+{
+	struct radeon_bo_va *bo_va, *tmp;
+	int r;
+
+	list_for_each_entry_safe(bo_va, tmp, &vm->freed, vm_status) {
+		r = radeon_vm_bo_update(rdev, bo_va, NULL);
+		radeon_bo_unref(&bo_va->bo);
+		kfree(bo_va);
+		if (r)
+			return r;
+	}
+	return 0;
+
+}
+
+/**
+ * radeon_vm_clear_invalids - clear invalidated BOs in the PT
+ *
+ * @rdev: radeon_device pointer
+ * @vm: requested vm
+ *
+ * Make sure all invalidated BOs are cleared in the PT.
+ * Returns 0 for success.
+ *
+ * PTs have to be reserved and mutex must be locked!
+ */
+int radeon_vm_clear_invalids(struct radeon_device *rdev,
+			     struct radeon_vm *vm)
+{
+	struct radeon_bo_va *bo_va, *tmp;
+	int r;
+
+	list_for_each_entry_safe(bo_va, tmp, &vm->invalidated, vm_status) {
+		r = radeon_vm_bo_update(rdev, bo_va, NULL);
+		if (r)
+			return r;
+	}
+	return 0;
+}
+
+/**
  * radeon_vm_bo_rmv - remove a bo to a specific vm
  *
  * @rdev: radeon_device pointer
  * @bo_va: requested bo_va
  *
  * Remove @bo_va->bo from the requested vm (cayman+).
- * Remove @bo_va->bo from the list of bos associated with the bo_va->vm and
- * remove the ptes for @bo_va in the page table.
- * Returns 0 for success.
  *
  * Object have to be reserved!
  */
-int radeon_vm_bo_rmv(struct radeon_device *rdev,
-		     struct radeon_bo_va *bo_va)
+void radeon_vm_bo_rmv(struct radeon_device *rdev,
+		      struct radeon_bo_va *bo_va)
 {
-	int r = 0;
-
-	mutex_lock(&bo_va->vm->mutex);
-	if (bo_va->soffset)
-		r = radeon_vm_bo_update(rdev, bo_va->vm, bo_va->bo, NULL);
+	struct radeon_vm *vm = bo_va->vm;
 
-	list_del(&bo_va->vm_list);
-	mutex_unlock(&bo_va->vm->mutex);
 	list_del(&bo_va->bo_list);
 
-	kfree(bo_va);
-	return r;
+	mutex_lock(&vm->mutex);
+	interval_tree_remove(&bo_va->it, &vm->va);
+	list_del(&bo_va->vm_status);
+
+	if (bo_va->addr) {
+		bo_va->bo = radeon_bo_ref(bo_va->bo);
+		list_add(&bo_va->vm_status, &vm->freed);
+	} else {
+		kfree(bo_va);
+	}
+
+	mutex_unlock(&vm->mutex);
 }
 
 /**
@@ -955,7 +1073,12 @@ void radeon_vm_bo_invalidate(struct radeon_device *rdev,
 	struct radeon_bo_va *bo_va;
 
 	list_for_each_entry(bo_va, &bo->va, bo_list) {
-		bo_va->valid = false;
+		if (bo_va->addr) {
+			mutex_lock(&bo_va->vm->mutex);
+			list_del(&bo_va->vm_status);
+			list_add(&bo_va->vm_status, &bo_va->vm->invalidated);
+			mutex_unlock(&bo_va->vm->mutex);
+		}
 	}
 }
 
@@ -975,11 +1098,14 @@ int radeon_vm_init(struct radeon_device *rdev, struct radeon_vm *vm)
 	int r;
 
 	vm->id = 0;
+	vm->ib_bo_va = NULL;
 	vm->fence = NULL;
 	vm->last_flush = NULL;
 	vm->last_id_use = NULL;
 	mutex_init(&vm->mutex);
-	INIT_LIST_HEAD(&vm->va);
+	vm->va = RB_ROOT;
+	INIT_LIST_HEAD(&vm->invalidated);
+	INIT_LIST_HEAD(&vm->freed);
 
 	pd_size = radeon_vm_directory_size(rdev);
 	pd_entries = radeon_vm_num_pdes(rdev);
@@ -993,7 +1119,7 @@ int radeon_vm_init(struct radeon_device *rdev, struct radeon_vm *vm)
 	}
 
 	r = radeon_bo_create(rdev, pd_size, align, true,
-			     RADEON_GEM_DOMAIN_VRAM, NULL,
+			     RADEON_GEM_DOMAIN_VRAM, 0, NULL,
 			     &vm->page_directory);
 	if (r)
 		return r;
@@ -1022,11 +1148,11 @@ void radeon_vm_fini(struct radeon_device *rdev, struct radeon_vm *vm)
 	struct radeon_bo_va *bo_va, *tmp;
 	int i, r;
 
-	if (!list_empty(&vm->va)) {
+	if (!RB_EMPTY_ROOT(&vm->va)) {
 		dev_err(rdev->dev, "still active bo inside vm\n");
 	}
-	list_for_each_entry_safe(bo_va, tmp, &vm->va, vm_list) {
-		list_del_init(&bo_va->vm_list);
+	rbtree_postorder_for_each_entry_safe(bo_va, tmp, &vm->va, it.rb) {
+		interval_tree_remove(&bo_va->it, &vm->va);
 		r = radeon_bo_reserve(bo_va->bo, false);
 		if (!r) {
 			list_del_init(&bo_va->bo_list);
@@ -1034,7 +1160,10 @@ void radeon_vm_fini(struct radeon_device *rdev, struct radeon_vm *vm)
 			kfree(bo_va);
 		}
 	}
-
+	list_for_each_entry_safe(bo_va, tmp, &vm->freed, vm_status) {
+		radeon_bo_unref(&bo_va->bo);
+		kfree(bo_va);
+	}
 
 	for (i = 0; i < radeon_vm_num_pdes(rdev); i++)
 		radeon_bo_unref(&vm->page_tables[i].bo);
diff --git a/drivers/gpu/drm/radeon/rs400.c b/drivers/gpu/drm/radeon/rs400.c
index a0f96decece3..6c1fc339d228 100644
--- a/drivers/gpu/drm/radeon/rs400.c
+++ b/drivers/gpu/drm/radeon/rs400.c
@@ -109,7 +109,6 @@ int rs400_gart_enable(struct radeon_device *rdev)
 	uint32_t size_reg;
 	uint32_t tmp;
 
-	radeon_gart_restore(rdev);
 	tmp = RREG32_MC(RS690_AIC_CTRL_SCRATCH);
 	tmp |= RS690_DIS_OUT_OF_PCI_GART_ACCESS;
 	WREG32_MC(RS690_AIC_CTRL_SCRATCH, tmp);
@@ -209,17 +208,24 @@ void rs400_gart_fini(struct radeon_device *rdev)
 	radeon_gart_table_ram_free(rdev);
 }
 
+#define RS400_PTE_UNSNOOPED (1 << 0)
 #define RS400_PTE_WRITEABLE (1 << 2)
 #define RS400_PTE_READABLE  (1 << 3)
 
-void rs400_gart_set_page(struct radeon_device *rdev, unsigned i, uint64_t addr)
+void rs400_gart_set_page(struct radeon_device *rdev, unsigned i,
+			 uint64_t addr, uint32_t flags)
 {
 	uint32_t entry;
 	u32 *gtt = rdev->gart.ptr;
 
 	entry = (lower_32_bits(addr) & PAGE_MASK) |
-		((upper_32_bits(addr) & 0xff) << 4) |
-		RS400_PTE_WRITEABLE | RS400_PTE_READABLE;
+		((upper_32_bits(addr) & 0xff) << 4);
+	if (flags & RADEON_GART_PAGE_READ)
+		addr |= RS400_PTE_READABLE;
+	if (flags & RADEON_GART_PAGE_WRITE)
+		addr |= RS400_PTE_WRITEABLE;
+	if (!(flags & RADEON_GART_PAGE_SNOOP))
+		entry |= RS400_PTE_UNSNOOPED;
 	entry = cpu_to_le32(entry);
 	gtt[i] = entry;
 }
diff --git a/drivers/gpu/drm/radeon/rs600.c b/drivers/gpu/drm/radeon/rs600.c
index d1a35cb1c91d..5f6db4629aaa 100644
--- a/drivers/gpu/drm/radeon/rs600.c
+++ b/drivers/gpu/drm/radeon/rs600.c
@@ -555,7 +555,6 @@ static int rs600_gart_enable(struct radeon_device *rdev)
 	r = radeon_gart_table_vram_pin(rdev);
 	if (r)
 		return r;
-	radeon_gart_restore(rdev);
 	/* Enable bus master */
 	tmp = RREG32(RADEON_BUS_CNTL) & ~RS600_BUS_MASTER_DIS;
 	WREG32(RADEON_BUS_CNTL, tmp);
@@ -626,15 +625,21 @@ static void rs600_gart_fini(struct radeon_device *rdev)
 	radeon_gart_table_vram_free(rdev);
 }
 
-void rs600_gart_set_page(struct radeon_device *rdev, unsigned i, uint64_t addr)
+void rs600_gart_set_page(struct radeon_device *rdev, unsigned i,
+			 uint64_t addr, uint32_t flags)
 {
 	void __iomem *ptr = (void *)rdev->gart.ptr;
 
 	addr = addr & 0xFFFFFFFFFFFFF000ULL;
-	if (addr == rdev->dummy_page.addr)
-		addr |= R600_PTE_SYSTEM | R600_PTE_SNOOPED;
-	else
-		addr |= R600_PTE_GART;
+	addr |= R600_PTE_SYSTEM;
+	if (flags & RADEON_GART_PAGE_VALID)
+		addr |= R600_PTE_VALID;
+	if (flags & RADEON_GART_PAGE_READ)
+		addr |= R600_PTE_READABLE;
+	if (flags & RADEON_GART_PAGE_WRITE)
+		addr |= R600_PTE_WRITEABLE;
+	if (flags & RADEON_GART_PAGE_SNOOP)
+		addr |= R600_PTE_SNOOPED;
 	writeq(addr, ptr + (i * 8));
 }
 
diff --git a/drivers/gpu/drm/radeon/rv515.c b/drivers/gpu/drm/radeon/rv515.c
index 3e21e869015f..8a477bf1fdb3 100644
--- a/drivers/gpu/drm/radeon/rv515.c
+++ b/drivers/gpu/drm/radeon/rv515.c
@@ -124,7 +124,7 @@ void rv515_ring_start(struct radeon_device *rdev, struct radeon_ring *ring)
 	radeon_ring_write(ring, GEOMETRY_ROUND_NEAREST | COLOR_ROUND_NEAREST);
 	radeon_ring_write(ring, PACKET0(0x20C8, 0));
 	radeon_ring_write(ring, 0);
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev, ring, false);
 }
 
 int rv515_mc_wait_for_idle(struct radeon_device *rdev)
diff --git a/drivers/gpu/drm/radeon/rv770.c b/drivers/gpu/drm/radeon/rv770.c
index da8703d8d455..d9f5ce715c9b 100644
--- a/drivers/gpu/drm/radeon/rv770.c
+++ b/drivers/gpu/drm/radeon/rv770.c
@@ -900,7 +900,6 @@ static int rv770_pcie_gart_enable(struct radeon_device *rdev)
 	r = radeon_gart_table_vram_pin(rdev);
 	if (r)
 		return r;
-	radeon_gart_restore(rdev);
 	/* Setup L2 cache */
 	WREG32(VM_L2_CNTL, ENABLE_L2_CACHE | ENABLE_L2_FRAGMENT_PROCESSING |
 				ENABLE_L2_PTE_CACHE_LRU_UPDATE_BY_WRITE |
@@ -1178,7 +1177,6 @@ static void rv770_gpu_init(struct radeon_device *rdev)
 	u32 hdp_host_path_cntl;
 	u32 sq_dyn_gpr_size_simd_ab_0;
 	u32 gb_tiling_config = 0;
-	u32 cc_rb_backend_disable = 0;
 	u32 cc_gc_shader_pipe_config = 0;
 	u32 mc_arb_ramcfg;
 	u32 db_debug4, tmp;
@@ -1312,21 +1310,7 @@ static void rv770_gpu_init(struct radeon_device *rdev)
 		WREG32(SPI_CONFIG_CNTL, 0);
 	}
 
-	cc_rb_backend_disable = RREG32(CC_RB_BACKEND_DISABLE) & 0x00ff0000;
-	tmp = R7XX_MAX_BACKENDS - r600_count_pipe_bits(cc_rb_backend_disable >> 16);
-	if (tmp < rdev->config.rv770.max_backends) {
-		rdev->config.rv770.max_backends = tmp;
-	}
-
 	cc_gc_shader_pipe_config = RREG32(CC_GC_SHADER_PIPE_CONFIG) & 0xffffff00;
-	tmp = R7XX_MAX_PIPES - r600_count_pipe_bits((cc_gc_shader_pipe_config >> 8) & R7XX_MAX_PIPES_MASK);
-	if (tmp < rdev->config.rv770.max_pipes) {
-		rdev->config.rv770.max_pipes = tmp;
-	}
-	tmp = R7XX_MAX_SIMDS - r600_count_pipe_bits((cc_gc_shader_pipe_config >> 16) & R7XX_MAX_SIMDS_MASK);
-	if (tmp < rdev->config.rv770.max_simds) {
-		rdev->config.rv770.max_simds = tmp;
-	}
 	tmp = rdev->config.rv770.max_simds -
 		r600_count_pipe_bits((cc_gc_shader_pipe_config >> 16) & R7XX_MAX_SIMDS_MASK);
 	rdev->config.rv770.active_simds = tmp;
@@ -1349,6 +1333,14 @@ static void rv770_gpu_init(struct radeon_device *rdev)
 	rdev->config.rv770.tiling_npipes = rdev->config.rv770.max_tile_pipes;
 
 	disabled_rb_mask = (RREG32(CC_RB_BACKEND_DISABLE) >> 16) & R7XX_MAX_BACKENDS_MASK;
+	tmp = 0;
+	for (i = 0; i < rdev->config.rv770.max_backends; i++)
+		tmp |= (1 << i);
+	/* if all the backends are disabled, fix it up here */
+	if ((disabled_rb_mask & tmp) == tmp) {
+		for (i = 0; i < rdev->config.rv770.max_backends; i++)
+			disabled_rb_mask &= ~(1 << i);
+	}
 	tmp = (gb_tiling_config & PIPE_TILING__MASK) >> PIPE_TILING__SHIFT;
 	tmp = r6xx_remap_render_backend(rdev, tmp, rdev->config.rv770.max_backends,
 					R7XX_MAX_BACKENDS, disabled_rb_mask);
diff --git a/drivers/gpu/drm/radeon/rv770_dma.c b/drivers/gpu/drm/radeon/rv770_dma.c
index bbf2e076ee45..74426ac2bb5c 100644
--- a/drivers/gpu/drm/radeon/rv770_dma.c
+++ b/drivers/gpu/drm/radeon/rv770_dma.c
@@ -90,7 +90,7 @@ int rv770_copy_dma(struct radeon_device *rdev,
 		return r;
 	}
 
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev, ring, false);
 	radeon_semaphore_free(rdev, &sem, *fence);
 
 	return r;
diff --git a/drivers/gpu/drm/radeon/si.c b/drivers/gpu/drm/radeon/si.c
index eba0225259a4..6bce40847753 100644
--- a/drivers/gpu/drm/radeon/si.c
+++ b/drivers/gpu/drm/radeon/si.c
@@ -42,6 +42,14 @@ MODULE_FIRMWARE("radeon/TAHITI_mc.bin");
 MODULE_FIRMWARE("radeon/TAHITI_mc2.bin");
 MODULE_FIRMWARE("radeon/TAHITI_rlc.bin");
 MODULE_FIRMWARE("radeon/TAHITI_smc.bin");
+
+MODULE_FIRMWARE("radeon/tahiti_pfp.bin");
+MODULE_FIRMWARE("radeon/tahiti_me.bin");
+MODULE_FIRMWARE("radeon/tahiti_ce.bin");
+MODULE_FIRMWARE("radeon/tahiti_mc.bin");
+MODULE_FIRMWARE("radeon/tahiti_rlc.bin");
+MODULE_FIRMWARE("radeon/tahiti_smc.bin");
+
 MODULE_FIRMWARE("radeon/PITCAIRN_pfp.bin");
 MODULE_FIRMWARE("radeon/PITCAIRN_me.bin");
 MODULE_FIRMWARE("radeon/PITCAIRN_ce.bin");
@@ -49,6 +57,14 @@ MODULE_FIRMWARE("radeon/PITCAIRN_mc.bin");
 MODULE_FIRMWARE("radeon/PITCAIRN_mc2.bin");
 MODULE_FIRMWARE("radeon/PITCAIRN_rlc.bin");
 MODULE_FIRMWARE("radeon/PITCAIRN_smc.bin");
+
+MODULE_FIRMWARE("radeon/pitcairn_pfp.bin");
+MODULE_FIRMWARE("radeon/pitcairn_me.bin");
+MODULE_FIRMWARE("radeon/pitcairn_ce.bin");
+MODULE_FIRMWARE("radeon/pitcairn_mc.bin");
+MODULE_FIRMWARE("radeon/pitcairn_rlc.bin");
+MODULE_FIRMWARE("radeon/pitcairn_smc.bin");
+
 MODULE_FIRMWARE("radeon/VERDE_pfp.bin");
 MODULE_FIRMWARE("radeon/VERDE_me.bin");
 MODULE_FIRMWARE("radeon/VERDE_ce.bin");
@@ -56,6 +72,14 @@ MODULE_FIRMWARE("radeon/VERDE_mc.bin");
 MODULE_FIRMWARE("radeon/VERDE_mc2.bin");
 MODULE_FIRMWARE("radeon/VERDE_rlc.bin");
 MODULE_FIRMWARE("radeon/VERDE_smc.bin");
+
+MODULE_FIRMWARE("radeon/verde_pfp.bin");
+MODULE_FIRMWARE("radeon/verde_me.bin");
+MODULE_FIRMWARE("radeon/verde_ce.bin");
+MODULE_FIRMWARE("radeon/verde_mc.bin");
+MODULE_FIRMWARE("radeon/verde_rlc.bin");
+MODULE_FIRMWARE("radeon/verde_smc.bin");
+
 MODULE_FIRMWARE("radeon/OLAND_pfp.bin");
 MODULE_FIRMWARE("radeon/OLAND_me.bin");
 MODULE_FIRMWARE("radeon/OLAND_ce.bin");
@@ -63,6 +87,14 @@ MODULE_FIRMWARE("radeon/OLAND_mc.bin");
 MODULE_FIRMWARE("radeon/OLAND_mc2.bin");
 MODULE_FIRMWARE("radeon/OLAND_rlc.bin");
 MODULE_FIRMWARE("radeon/OLAND_smc.bin");
+
+MODULE_FIRMWARE("radeon/oland_pfp.bin");
+MODULE_FIRMWARE("radeon/oland_me.bin");
+MODULE_FIRMWARE("radeon/oland_ce.bin");
+MODULE_FIRMWARE("radeon/oland_mc.bin");
+MODULE_FIRMWARE("radeon/oland_rlc.bin");
+MODULE_FIRMWARE("radeon/oland_smc.bin");
+
 MODULE_FIRMWARE("radeon/HAINAN_pfp.bin");
 MODULE_FIRMWARE("radeon/HAINAN_me.bin");
 MODULE_FIRMWARE("radeon/HAINAN_ce.bin");
@@ -71,6 +103,13 @@ MODULE_FIRMWARE("radeon/HAINAN_mc2.bin");
 MODULE_FIRMWARE("radeon/HAINAN_rlc.bin");
 MODULE_FIRMWARE("radeon/HAINAN_smc.bin");
 
+MODULE_FIRMWARE("radeon/hainan_pfp.bin");
+MODULE_FIRMWARE("radeon/hainan_me.bin");
+MODULE_FIRMWARE("radeon/hainan_ce.bin");
+MODULE_FIRMWARE("radeon/hainan_mc.bin");
+MODULE_FIRMWARE("radeon/hainan_rlc.bin");
+MODULE_FIRMWARE("radeon/hainan_smc.bin");
+
 static u32 si_get_cu_active_bitmap(struct radeon_device *rdev, u32 se, u32 sh);
 static void si_pcie_gen3_enable(struct radeon_device *rdev);
 static void si_program_aspm(struct radeon_device *rdev);
@@ -1470,38 +1509,54 @@ static const u32 hainan_io_mc_regs[TAHITI_IO_MC_REGS_SIZE][2] = {
 /* ucode loading */
 int si_mc_load_microcode(struct radeon_device *rdev)
 {
-	const __be32 *fw_data;
+	const __be32 *fw_data = NULL;
+	const __le32 *new_fw_data = NULL;
 	u32 running, blackout = 0;
-	u32 *io_mc_regs;
+	u32 *io_mc_regs = NULL;
+	const __le32 *new_io_mc_regs = NULL;
 	int i, regs_size, ucode_size;
 
 	if (!rdev->mc_fw)
 		return -EINVAL;
 
-	ucode_size = rdev->mc_fw->size / 4;
+	if (rdev->new_fw) {
+		const struct mc_firmware_header_v1_0 *hdr =
+			(const struct mc_firmware_header_v1_0 *)rdev->mc_fw->data;
+
+		radeon_ucode_print_mc_hdr(&hdr->header);
+		regs_size = le32_to_cpu(hdr->io_debug_size_bytes) / (4 * 2);
+		new_io_mc_regs = (const __le32 *)
+			(rdev->mc_fw->data + le32_to_cpu(hdr->io_debug_array_offset_bytes));
+		ucode_size = le32_to_cpu(hdr->header.ucode_size_bytes) / 4;
+		new_fw_data = (const __le32 *)
+			(rdev->mc_fw->data + le32_to_cpu(hdr->header.ucode_array_offset_bytes));
+	} else {
+		ucode_size = rdev->mc_fw->size / 4;
 
-	switch (rdev->family) {
-	case CHIP_TAHITI:
-		io_mc_regs = (u32 *)&tahiti_io_mc_regs;
-		regs_size = TAHITI_IO_MC_REGS_SIZE;
-		break;
-	case CHIP_PITCAIRN:
-		io_mc_regs = (u32 *)&pitcairn_io_mc_regs;
-		regs_size = TAHITI_IO_MC_REGS_SIZE;
-		break;
-	case CHIP_VERDE:
-	default:
-		io_mc_regs = (u32 *)&verde_io_mc_regs;
-		regs_size = TAHITI_IO_MC_REGS_SIZE;
-		break;
-	case CHIP_OLAND:
-		io_mc_regs = (u32 *)&oland_io_mc_regs;
-		regs_size = TAHITI_IO_MC_REGS_SIZE;
-		break;
-	case CHIP_HAINAN:
-		io_mc_regs = (u32 *)&hainan_io_mc_regs;
-		regs_size = TAHITI_IO_MC_REGS_SIZE;
-		break;
+		switch (rdev->family) {
+		case CHIP_TAHITI:
+			io_mc_regs = (u32 *)&tahiti_io_mc_regs;
+			regs_size = TAHITI_IO_MC_REGS_SIZE;
+			break;
+		case CHIP_PITCAIRN:
+			io_mc_regs = (u32 *)&pitcairn_io_mc_regs;
+			regs_size = TAHITI_IO_MC_REGS_SIZE;
+			break;
+		case CHIP_VERDE:
+		default:
+			io_mc_regs = (u32 *)&verde_io_mc_regs;
+			regs_size = TAHITI_IO_MC_REGS_SIZE;
+			break;
+		case CHIP_OLAND:
+			io_mc_regs = (u32 *)&oland_io_mc_regs;
+			regs_size = TAHITI_IO_MC_REGS_SIZE;
+			break;
+		case CHIP_HAINAN:
+			io_mc_regs = (u32 *)&hainan_io_mc_regs;
+			regs_size = TAHITI_IO_MC_REGS_SIZE;
+			break;
+		}
+		fw_data = (const __be32 *)rdev->mc_fw->data;
 	}
 
 	running = RREG32(MC_SEQ_SUP_CNTL) & RUN_MASK;
@@ -1518,13 +1573,21 @@ int si_mc_load_microcode(struct radeon_device *rdev)
 
 		/* load mc io regs */
 		for (i = 0; i < regs_size; i++) {
-			WREG32(MC_SEQ_IO_DEBUG_INDEX, io_mc_regs[(i << 1)]);
-			WREG32(MC_SEQ_IO_DEBUG_DATA, io_mc_regs[(i << 1) + 1]);
+			if (rdev->new_fw) {
+				WREG32(MC_SEQ_IO_DEBUG_INDEX, le32_to_cpup(new_io_mc_regs++));
+				WREG32(MC_SEQ_IO_DEBUG_DATA, le32_to_cpup(new_io_mc_regs++));
+			} else {
+				WREG32(MC_SEQ_IO_DEBUG_INDEX, io_mc_regs[(i << 1)]);
+				WREG32(MC_SEQ_IO_DEBUG_DATA, io_mc_regs[(i << 1) + 1]);
+			}
 		}
 		/* load the MC ucode */
-		fw_data = (const __be32 *)rdev->mc_fw->data;
-		for (i = 0; i < ucode_size; i++)
-			WREG32(MC_SEQ_SUP_PGM, be32_to_cpup(fw_data++));
+		for (i = 0; i < ucode_size; i++) {
+			if (rdev->new_fw)
+				WREG32(MC_SEQ_SUP_PGM, le32_to_cpup(new_fw_data++));
+			else
+				WREG32(MC_SEQ_SUP_PGM, be32_to_cpup(fw_data++));
+		}
 
 		/* put the engine back into the active state */
 		WREG32(MC_SEQ_SUP_CNTL, 0x00000008);
@@ -1553,18 +1616,19 @@ int si_mc_load_microcode(struct radeon_device *rdev)
 static int si_init_microcode(struct radeon_device *rdev)
 {
 	const char *chip_name;
-	const char *rlc_chip_name;
+	const char *new_chip_name;
 	size_t pfp_req_size, me_req_size, ce_req_size, rlc_req_size, mc_req_size;
 	size_t smc_req_size, mc2_req_size;
 	char fw_name[30];
 	int err;
+	int new_fw = 0;
 
 	DRM_DEBUG("\n");
 
 	switch (rdev->family) {
 	case CHIP_TAHITI:
 		chip_name = "TAHITI";
-		rlc_chip_name = "TAHITI";
+		new_chip_name = "tahiti";
 		pfp_req_size = SI_PFP_UCODE_SIZE * 4;
 		me_req_size = SI_PM4_UCODE_SIZE * 4;
 		ce_req_size = SI_CE_UCODE_SIZE * 4;
@@ -1575,7 +1639,7 @@ static int si_init_microcode(struct radeon_device *rdev)
 		break;
 	case CHIP_PITCAIRN:
 		chip_name = "PITCAIRN";
-		rlc_chip_name = "PITCAIRN";
+		new_chip_name = "pitcairn";
 		pfp_req_size = SI_PFP_UCODE_SIZE * 4;
 		me_req_size = SI_PM4_UCODE_SIZE * 4;
 		ce_req_size = SI_CE_UCODE_SIZE * 4;
@@ -1586,7 +1650,7 @@ static int si_init_microcode(struct radeon_device *rdev)
 		break;
 	case CHIP_VERDE:
 		chip_name = "VERDE";
-		rlc_chip_name = "VERDE";
+		new_chip_name = "verde";
 		pfp_req_size = SI_PFP_UCODE_SIZE * 4;
 		me_req_size = SI_PM4_UCODE_SIZE * 4;
 		ce_req_size = SI_CE_UCODE_SIZE * 4;
@@ -1597,7 +1661,7 @@ static int si_init_microcode(struct radeon_device *rdev)
 		break;
 	case CHIP_OLAND:
 		chip_name = "OLAND";
-		rlc_chip_name = "OLAND";
+		new_chip_name = "oland";
 		pfp_req_size = SI_PFP_UCODE_SIZE * 4;
 		me_req_size = SI_PM4_UCODE_SIZE * 4;
 		ce_req_size = SI_CE_UCODE_SIZE * 4;
@@ -1607,7 +1671,7 @@ static int si_init_microcode(struct radeon_device *rdev)
 		break;
 	case CHIP_HAINAN:
 		chip_name = "HAINAN";
-		rlc_chip_name = "HAINAN";
+		new_chip_name = "hainan";
 		pfp_req_size = SI_PFP_UCODE_SIZE * 4;
 		me_req_size = SI_PM4_UCODE_SIZE * 4;
 		ce_req_size = SI_CE_UCODE_SIZE * 4;
@@ -1618,86 +1682,178 @@ static int si_init_microcode(struct radeon_device *rdev)
 	default: BUG();
 	}
 
-	DRM_INFO("Loading %s Microcode\n", chip_name);
+	DRM_INFO("Loading %s Microcode\n", new_chip_name);
 
-	snprintf(fw_name, sizeof(fw_name), "radeon/%s_pfp.bin", chip_name);
+	snprintf(fw_name, sizeof(fw_name), "radeon/%s_pfp.bin", new_chip_name);
 	err = request_firmware(&rdev->pfp_fw, fw_name, rdev->dev);
-	if (err)
-		goto out;
-	if (rdev->pfp_fw->size != pfp_req_size) {
-		printk(KERN_ERR
-		       "si_cp: Bogus length %zu in firmware \"%s\"\n",
-		       rdev->pfp_fw->size, fw_name);
-		err = -EINVAL;
-		goto out;
+	if (err) {
+		snprintf(fw_name, sizeof(fw_name), "radeon/%s_pfp.bin", chip_name);
+		err = request_firmware(&rdev->pfp_fw, fw_name, rdev->dev);
+		if (err)
+			goto out;
+		if (rdev->pfp_fw->size != pfp_req_size) {
+			printk(KERN_ERR
+			       "si_cp: Bogus length %zu in firmware \"%s\"\n",
+			       rdev->pfp_fw->size, fw_name);
+			err = -EINVAL;
+			goto out;
+		}
+	} else {
+		err = radeon_ucode_validate(rdev->pfp_fw);
+		if (err) {
+			printk(KERN_ERR
+			       "si_cp: validation failed for firmware \"%s\"\n",
+			       fw_name);
+			goto out;
+		} else {
+			new_fw++;
+		}
 	}
 
-	snprintf(fw_name, sizeof(fw_name), "radeon/%s_me.bin", chip_name);
+	snprintf(fw_name, sizeof(fw_name), "radeon/%s_me.bin", new_chip_name);
 	err = request_firmware(&rdev->me_fw, fw_name, rdev->dev);
-	if (err)
-		goto out;
-	if (rdev->me_fw->size != me_req_size) {
-		printk(KERN_ERR
-		       "si_cp: Bogus length %zu in firmware \"%s\"\n",
-		       rdev->me_fw->size, fw_name);
-		err = -EINVAL;
+	if (err) {
+		snprintf(fw_name, sizeof(fw_name), "radeon/%s_me.bin", chip_name);
+		err = request_firmware(&rdev->me_fw, fw_name, rdev->dev);
+		if (err)
+			goto out;
+		if (rdev->me_fw->size != me_req_size) {
+			printk(KERN_ERR
+			       "si_cp: Bogus length %zu in firmware \"%s\"\n",
+			       rdev->me_fw->size, fw_name);
+			err = -EINVAL;
+		}
+	} else {
+		err = radeon_ucode_validate(rdev->me_fw);
+		if (err) {
+			printk(KERN_ERR
+			       "si_cp: validation failed for firmware \"%s\"\n",
+			       fw_name);
+			goto out;
+		} else {
+			new_fw++;
+		}
 	}
 
-	snprintf(fw_name, sizeof(fw_name), "radeon/%s_ce.bin", chip_name);
+	snprintf(fw_name, sizeof(fw_name), "radeon/%s_ce.bin", new_chip_name);
 	err = request_firmware(&rdev->ce_fw, fw_name, rdev->dev);
-	if (err)
-		goto out;
-	if (rdev->ce_fw->size != ce_req_size) {
-		printk(KERN_ERR
-		       "si_cp: Bogus length %zu in firmware \"%s\"\n",
-		       rdev->ce_fw->size, fw_name);
-		err = -EINVAL;
+	if (err) {
+		snprintf(fw_name, sizeof(fw_name), "radeon/%s_ce.bin", chip_name);
+		err = request_firmware(&rdev->ce_fw, fw_name, rdev->dev);
+		if (err)
+			goto out;
+		if (rdev->ce_fw->size != ce_req_size) {
+			printk(KERN_ERR
+			       "si_cp: Bogus length %zu in firmware \"%s\"\n",
+			       rdev->ce_fw->size, fw_name);
+			err = -EINVAL;
+		}
+	} else {
+		err = radeon_ucode_validate(rdev->ce_fw);
+		if (err) {
+			printk(KERN_ERR
+			       "si_cp: validation failed for firmware \"%s\"\n",
+			       fw_name);
+			goto out;
+		} else {
+			new_fw++;
+		}
 	}
 
-	snprintf(fw_name, sizeof(fw_name), "radeon/%s_rlc.bin", rlc_chip_name);
+	snprintf(fw_name, sizeof(fw_name), "radeon/%s_rlc.bin", new_chip_name);
 	err = request_firmware(&rdev->rlc_fw, fw_name, rdev->dev);
-	if (err)
-		goto out;
-	if (rdev->rlc_fw->size != rlc_req_size) {
-		printk(KERN_ERR
-		       "si_rlc: Bogus length %zu in firmware \"%s\"\n",
-		       rdev->rlc_fw->size, fw_name);
-		err = -EINVAL;
+	if (err) {
+		snprintf(fw_name, sizeof(fw_name), "radeon/%s_rlc.bin", chip_name);
+		err = request_firmware(&rdev->rlc_fw, fw_name, rdev->dev);
+		if (err)
+			goto out;
+		if (rdev->rlc_fw->size != rlc_req_size) {
+			printk(KERN_ERR
+			       "si_rlc: Bogus length %zu in firmware \"%s\"\n",
+			       rdev->rlc_fw->size, fw_name);
+			err = -EINVAL;
+		}
+	} else {
+		err = radeon_ucode_validate(rdev->rlc_fw);
+		if (err) {
+			printk(KERN_ERR
+			       "si_cp: validation failed for firmware \"%s\"\n",
+			       fw_name);
+			goto out;
+		} else {
+			new_fw++;
+		}
 	}
 
-	snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc2.bin", chip_name);
+	snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc.bin", new_chip_name);
 	err = request_firmware(&rdev->mc_fw, fw_name, rdev->dev);
 	if (err) {
-		snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc.bin", chip_name);
+		snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc2.bin", chip_name);
 		err = request_firmware(&rdev->mc_fw, fw_name, rdev->dev);
-		if (err)
+		if (err) {
+			snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc.bin", chip_name);
+			err = request_firmware(&rdev->mc_fw, fw_name, rdev->dev);
+			if (err)
+				goto out;
+		}
+		if ((rdev->mc_fw->size != mc_req_size) &&
+		    (rdev->mc_fw->size != mc2_req_size)) {
+			printk(KERN_ERR
+			       "si_mc: Bogus length %zu in firmware \"%s\"\n",
+			       rdev->mc_fw->size, fw_name);
+			err = -EINVAL;
+		}
+		DRM_INFO("%s: %zu bytes\n", fw_name, rdev->mc_fw->size);
+	} else {
+		err = radeon_ucode_validate(rdev->mc_fw);
+		if (err) {
+			printk(KERN_ERR
+			       "si_cp: validation failed for firmware \"%s\"\n",
+			       fw_name);
 			goto out;
+		} else {
+			new_fw++;
+		}
 	}
-	if ((rdev->mc_fw->size != mc_req_size) &&
-	    (rdev->mc_fw->size != mc2_req_size)) {
-		printk(KERN_ERR
-		       "si_mc: Bogus length %zu in firmware \"%s\"\n",
-		       rdev->mc_fw->size, fw_name);
-		err = -EINVAL;
-	}
-	DRM_INFO("%s: %zu bytes\n", fw_name, rdev->mc_fw->size);
 
-	snprintf(fw_name, sizeof(fw_name), "radeon/%s_smc.bin", chip_name);
+	snprintf(fw_name, sizeof(fw_name), "radeon/%s_smc.bin", new_chip_name);
 	err = request_firmware(&rdev->smc_fw, fw_name, rdev->dev);
 	if (err) {
-		printk(KERN_ERR
-		       "smc: error loading firmware \"%s\"\n",
-		       fw_name);
-		release_firmware(rdev->smc_fw);
-		rdev->smc_fw = NULL;
-		err = 0;
-	} 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);
-		err = -EINVAL;
+		snprintf(fw_name, sizeof(fw_name), "radeon/%s_smc.bin", chip_name);
+		err = request_firmware(&rdev->smc_fw, fw_name, rdev->dev);
+		if (err) {
+			printk(KERN_ERR
+			       "smc: error loading firmware \"%s\"\n",
+			       fw_name);
+			release_firmware(rdev->smc_fw);
+			rdev->smc_fw = NULL;
+			err = 0;
+		} 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);
+			err = -EINVAL;
+		}
+	} else {
+		err = radeon_ucode_validate(rdev->smc_fw);
+		if (err) {
+			printk(KERN_ERR
+			       "si_cp: validation failed for firmware \"%s\"\n",
+			       fw_name);
+			goto out;
+		} else {
+			new_fw++;
+		}
 	}
 
+	if (new_fw == 0) {
+		rdev->new_fw = false;
+	} else if (new_fw < 6) {
+		printk(KERN_ERR "si_fw: mixing new and old firmware!\n");
+		err = -EINVAL;
+	} else {
+		rdev->new_fw = true;
+	}
 out:
 	if (err) {
 		if (err != -EINVAL)
@@ -2901,7 +3057,7 @@ static void si_gpu_init(struct radeon_device *rdev)
 	u32 sx_debug_1;
 	u32 hdp_host_path_cntl;
 	u32 tmp;
-	int i, j, k;
+	int i, j;
 
 	switch (rdev->family) {
 	case CHIP_TAHITI:
@@ -3099,12 +3255,11 @@ static void si_gpu_init(struct radeon_device *rdev)
 		     rdev->config.si.max_sh_per_se,
 		     rdev->config.si.max_cu_per_sh);
 
+	rdev->config.si.active_cus = 0;
 	for (i = 0; i < rdev->config.si.max_shader_engines; i++) {
 		for (j = 0; j < rdev->config.si.max_sh_per_se; j++) {
-			for (k = 0; k < rdev->config.si.max_cu_per_sh; k++) {
-				rdev->config.si.active_cus +=
-					hweight32(si_get_cu_active_bitmap(rdev, i, j));
-			}
+			rdev->config.si.active_cus +=
+				hweight32(si_get_cu_active_bitmap(rdev, i, j));
 		}
 	}
 
@@ -3282,34 +3437,77 @@ static void si_cp_enable(struct radeon_device *rdev, bool enable)
 
 static int si_cp_load_microcode(struct radeon_device *rdev)
 {
-	const __be32 *fw_data;
 	int i;
 
-	if (!rdev->me_fw || !rdev->pfp_fw)
+	if (!rdev->me_fw || !rdev->pfp_fw || !rdev->ce_fw)
 		return -EINVAL;
 
 	si_cp_enable(rdev, false);
 
-	/* PFP */
-	fw_data = (const __be32 *)rdev->pfp_fw->data;
-	WREG32(CP_PFP_UCODE_ADDR, 0);
-	for (i = 0; i < SI_PFP_UCODE_SIZE; i++)
-		WREG32(CP_PFP_UCODE_DATA, be32_to_cpup(fw_data++));
-	WREG32(CP_PFP_UCODE_ADDR, 0);
-
-	/* CE */
-	fw_data = (const __be32 *)rdev->ce_fw->data;
-	WREG32(CP_CE_UCODE_ADDR, 0);
-	for (i = 0; i < SI_CE_UCODE_SIZE; i++)
-		WREG32(CP_CE_UCODE_DATA, be32_to_cpup(fw_data++));
-	WREG32(CP_CE_UCODE_ADDR, 0);
-
-	/* ME */
-	fw_data = (const __be32 *)rdev->me_fw->data;
-	WREG32(CP_ME_RAM_WADDR, 0);
-	for (i = 0; i < SI_PM4_UCODE_SIZE; i++)
-		WREG32(CP_ME_RAM_DATA, be32_to_cpup(fw_data++));
-	WREG32(CP_ME_RAM_WADDR, 0);
+	if (rdev->new_fw) {
+		const struct gfx_firmware_header_v1_0 *pfp_hdr =
+			(const struct gfx_firmware_header_v1_0 *)rdev->pfp_fw->data;
+		const struct gfx_firmware_header_v1_0 *ce_hdr =
+			(const struct gfx_firmware_header_v1_0 *)rdev->ce_fw->data;
+		const struct gfx_firmware_header_v1_0 *me_hdr =
+			(const struct gfx_firmware_header_v1_0 *)rdev->me_fw->data;
+		const __le32 *fw_data;
+		u32 fw_size;
+
+		radeon_ucode_print_gfx_hdr(&pfp_hdr->header);
+		radeon_ucode_print_gfx_hdr(&ce_hdr->header);
+		radeon_ucode_print_gfx_hdr(&me_hdr->header);
+
+		/* PFP */
+		fw_data = (const __le32 *)
+			(rdev->pfp_fw->data + le32_to_cpu(pfp_hdr->header.ucode_array_offset_bytes));
+		fw_size = le32_to_cpu(pfp_hdr->header.ucode_size_bytes) / 4;
+		WREG32(CP_PFP_UCODE_ADDR, 0);
+		for (i = 0; i < fw_size; i++)
+			WREG32(CP_PFP_UCODE_DATA, le32_to_cpup(fw_data++));
+		WREG32(CP_PFP_UCODE_ADDR, 0);
+
+		/* CE */
+		fw_data = (const __le32 *)
+			(rdev->ce_fw->data + le32_to_cpu(ce_hdr->header.ucode_array_offset_bytes));
+		fw_size = le32_to_cpu(ce_hdr->header.ucode_size_bytes) / 4;
+		WREG32(CP_CE_UCODE_ADDR, 0);
+		for (i = 0; i < fw_size; i++)
+			WREG32(CP_CE_UCODE_DATA, le32_to_cpup(fw_data++));
+		WREG32(CP_CE_UCODE_ADDR, 0);
+
+		/* ME */
+		fw_data = (const __be32 *)
+			(rdev->me_fw->data + le32_to_cpu(me_hdr->header.ucode_array_offset_bytes));
+		fw_size = le32_to_cpu(me_hdr->header.ucode_size_bytes) / 4;
+		WREG32(CP_ME_RAM_WADDR, 0);
+		for (i = 0; i < fw_size; i++)
+			WREG32(CP_ME_RAM_DATA, le32_to_cpup(fw_data++));
+		WREG32(CP_ME_RAM_WADDR, 0);
+	} else {
+		const __be32 *fw_data;
+
+		/* PFP */
+		fw_data = (const __be32 *)rdev->pfp_fw->data;
+		WREG32(CP_PFP_UCODE_ADDR, 0);
+		for (i = 0; i < SI_PFP_UCODE_SIZE; i++)
+			WREG32(CP_PFP_UCODE_DATA, be32_to_cpup(fw_data++));
+		WREG32(CP_PFP_UCODE_ADDR, 0);
+
+		/* CE */
+		fw_data = (const __be32 *)rdev->ce_fw->data;
+		WREG32(CP_CE_UCODE_ADDR, 0);
+		for (i = 0; i < SI_CE_UCODE_SIZE; i++)
+			WREG32(CP_CE_UCODE_DATA, be32_to_cpup(fw_data++));
+		WREG32(CP_CE_UCODE_ADDR, 0);
+
+		/* ME */
+		fw_data = (const __be32 *)rdev->me_fw->data;
+		WREG32(CP_ME_RAM_WADDR, 0);
+		for (i = 0; i < SI_PM4_UCODE_SIZE; i++)
+			WREG32(CP_ME_RAM_DATA, be32_to_cpup(fw_data++));
+		WREG32(CP_ME_RAM_WADDR, 0);
+	}
 
 	WREG32(CP_PFP_UCODE_ADDR, 0);
 	WREG32(CP_CE_UCODE_ADDR, 0);
@@ -3342,7 +3540,7 @@ static int si_cp_start(struct radeon_device *rdev)
 	radeon_ring_write(ring, PACKET3_BASE_INDEX(CE_PARTITION_BASE));
 	radeon_ring_write(ring, 0xc000);
 	radeon_ring_write(ring, 0xe000);
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev, ring, false);
 
 	si_cp_enable(rdev, true);
 
@@ -3371,7 +3569,7 @@ static int si_cp_start(struct radeon_device *rdev)
 	radeon_ring_write(ring, 0x0000000e); /* VGT_VERTEX_REUSE_BLOCK_CNTL */
 	radeon_ring_write(ring, 0x00000010); /* VGT_OUT_DEALLOC_CNTL */
 
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev, ring, false);
 
 	for (i = RADEON_RING_TYPE_GFX_INDEX; i <= CAYMAN_RING_TYPE_CP2_INDEX; ++i) {
 		ring = &rdev->ring[i];
@@ -3381,7 +3579,7 @@ static int si_cp_start(struct radeon_device *rdev)
 		radeon_ring_write(ring, PACKET3_COMPUTE(PACKET3_CLEAR_STATE, 0));
 		radeon_ring_write(ring, 0);
 
-		radeon_ring_unlock_commit(rdev, ring);
+		radeon_ring_unlock_commit(rdev, ring, false);
 	}
 
 	return 0;
@@ -4048,7 +4246,6 @@ static int si_pcie_gart_enable(struct radeon_device *rdev)
 	r = radeon_gart_table_vram_pin(rdev);
 	if (r)
 		return r;
-	radeon_gart_restore(rdev);
 	/* Setup TLB control */
 	WREG32(MC_VM_MX_L1_TLB_CNTL,
 	       (0xA << 7) |
@@ -4093,10 +4290,10 @@ static int si_pcie_gart_enable(struct radeon_device *rdev)
 	for (i = 1; i < 16; i++) {
 		if (i < 8)
 			WREG32(VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (i << 2),
-			       rdev->gart.table_addr >> 12);
+			       rdev->vm_manager.saved_table_addr[i]);
 		else
 			WREG32(VM_CONTEXT8_PAGE_TABLE_BASE_ADDR + ((i - 8) << 2),
-			       rdev->gart.table_addr >> 12);
+			       rdev->vm_manager.saved_table_addr[i]);
 	}
 
 	/* enable context1-15 */
@@ -4128,6 +4325,17 @@ static int si_pcie_gart_enable(struct radeon_device *rdev)
 
 static void si_pcie_gart_disable(struct radeon_device *rdev)
 {
+	unsigned i;
+
+	for (i = 1; i < 16; ++i) {
+		uint32_t reg;
+		if (i < 8)
+			reg = VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (i << 2);
+		else
+			reg = VM_CONTEXT8_PAGE_TABLE_BASE_ADDR + ((i - 8) << 2);
+		rdev->vm_manager.saved_table_addr[i] = RREG32(reg);
+	}
+
 	/* Disable all tables */
 	WREG32(VM_CONTEXT0_CNTL, 0);
 	WREG32(VM_CONTEXT1_CNTL, 0);
@@ -4815,7 +5023,7 @@ void si_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm)
 
 	/* write new base address */
 	radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
-	radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(0) |
+	radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(1) |
 				 WRITE_DATA_DST_SEL(0)));
 
 	if (vm->id < 8) {
@@ -4830,7 +5038,7 @@ void si_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm)
 
 	/* flush hdp cache */
 	radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
-	radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(0) |
+	radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(1) |
 				 WRITE_DATA_DST_SEL(0)));
 	radeon_ring_write(ring, HDP_MEM_COHERENCY_FLUSH_CNTL >> 2);
 	radeon_ring_write(ring, 0);
@@ -4838,7 +5046,7 @@ void si_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm)
 
 	/* bits 0-15 are the VM contexts0-15 */
 	radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
-	radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(0) |
+	radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(1) |
 				 WRITE_DATA_DST_SEL(0)));
 	radeon_ring_write(ring, VM_INVALIDATE_REQUEST >> 2);
 	radeon_ring_write(ring, 0);
@@ -5592,7 +5800,6 @@ static void si_enable_lbpw(struct radeon_device *rdev, bool enable)
 static int si_rlc_resume(struct radeon_device *rdev)
 {
 	u32 i;
-	const __be32 *fw_data;
 
 	if (!rdev->rlc_fw)
 		return -EINVAL;
@@ -5615,10 +5822,26 @@ static int si_rlc_resume(struct radeon_device *rdev)
 	WREG32(RLC_MC_CNTL, 0);
 	WREG32(RLC_UCODE_CNTL, 0);
 
-	fw_data = (const __be32 *)rdev->rlc_fw->data;
-	for (i = 0; i < SI_RLC_UCODE_SIZE; i++) {
-		WREG32(RLC_UCODE_ADDR, i);
-		WREG32(RLC_UCODE_DATA, be32_to_cpup(fw_data++));
+	if (rdev->new_fw) {
+		const struct rlc_firmware_header_v1_0 *hdr =
+			(const struct rlc_firmware_header_v1_0 *)rdev->rlc_fw->data;
+		u32 fw_size = le32_to_cpu(hdr->header.ucode_size_bytes) / 4;
+		const __le32 *fw_data = (const __le32 *)
+			(rdev->rlc_fw->data + le32_to_cpu(hdr->header.ucode_array_offset_bytes));
+
+		radeon_ucode_print_rlc_hdr(&hdr->header);
+
+		for (i = 0; i < fw_size; i++) {
+			WREG32(RLC_UCODE_ADDR, i);
+			WREG32(RLC_UCODE_DATA, le32_to_cpup(fw_data++));
+		}
+	} else {
+		const __be32 *fw_data =
+			(const __be32 *)rdev->rlc_fw->data;
+		for (i = 0; i < SI_RLC_UCODE_SIZE; i++) {
+			WREG32(RLC_UCODE_ADDR, i);
+			WREG32(RLC_UCODE_DATA, be32_to_cpup(fw_data++));
+		}
 	}
 	WREG32(RLC_UCODE_ADDR, 0);
 
@@ -6103,6 +6326,7 @@ static inline u32 si_get_ih_wptr(struct radeon_device *rdev)
 		tmp = RREG32(IH_RB_CNTL);
 		tmp |= IH_WPTR_OVERFLOW_CLEAR;
 		WREG32(IH_RB_CNTL, tmp);
+		wptr &= ~RB_OVERFLOW;
 	}
 	return (wptr & rdev->ih.ptr_mask);
 }
@@ -6317,7 +6541,8 @@ restart_ih:
 		case 16: /* D5 page flip */
 		case 18: /* D6 page flip */
 			DRM_DEBUG("IH: D%d flip\n", ((src_id - 8) >> 1) + 1);
-			radeon_crtc_handle_flip(rdev, (src_id - 8) >> 1);
+			if (radeon_use_pflipirq > 0)
+				radeon_crtc_handle_flip(rdev, (src_id - 8) >> 1);
 			break;
 		case 42: /* HPD hotplug */
 			switch (src_data) {
@@ -6963,6 +7188,9 @@ static void si_pcie_gen3_enable(struct radeon_device *rdev)
 	int ret, i;
 	u16 tmp16;
 
+	if (pci_is_root_bus(rdev->pdev->bus))
+		return;
+
 	if (radeon_pcie_gen2 == 0)
 		return;
 
@@ -7240,7 +7468,8 @@ static void si_program_aspm(struct radeon_device *rdev)
 			if (orig != data)
 				WREG32_PIF_PHY1(PB1_PIF_CNTL, data);
 
-			if (!disable_clkreq) {
+			if (!disable_clkreq &&
+			    !pci_is_root_bus(rdev->pdev->bus)) {
 				struct pci_dev *root = rdev->pdev->bus->self;
 				u32 lnkcap;
 
diff --git a/drivers/gpu/drm/radeon/si_dma.c b/drivers/gpu/drm/radeon/si_dma.c
index e24c94b6d14d..7c22baaf94db 100644
--- a/drivers/gpu/drm/radeon/si_dma.c
+++ b/drivers/gpu/drm/radeon/si_dma.c
@@ -56,7 +56,41 @@ bool si_dma_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring)
 }
 
 /**
- * si_dma_vm_set_page - update the page tables using the DMA
+ * si_dma_vm_copy_pages - update PTEs by copying them from the GART
+ *
+ * @rdev: radeon_device pointer
+ * @ib: indirect buffer to fill with commands
+ * @pe: addr of the page entry
+ * @src: src addr where to copy from
+ * @count: number of page entries to update
+ *
+ * Update PTEs by copying them from the GART using the DMA (SI).
+ */
+void si_dma_vm_copy_pages(struct radeon_device *rdev,
+			  struct radeon_ib *ib,
+			  uint64_t pe, uint64_t src,
+			  unsigned count)
+{
+	while (count) {
+		unsigned bytes = count * 8;
+		if (bytes > 0xFFFF8)
+			bytes = 0xFFFF8;
+
+		ib->ptr[ib->length_dw++] = DMA_PACKET(DMA_PACKET_COPY,
+						      1, 0, 0, bytes);
+		ib->ptr[ib->length_dw++] = lower_32_bits(pe);
+		ib->ptr[ib->length_dw++] = lower_32_bits(src);
+		ib->ptr[ib->length_dw++] = upper_32_bits(pe) & 0xff;
+		ib->ptr[ib->length_dw++] = upper_32_bits(src) & 0xff;
+
+		pe += bytes;
+		src += bytes;
+		count -= bytes / 8;
+	}
+}
+
+/**
+ * si_dma_vm_write_pages - update PTEs by writing them manually
  *
  * @rdev: radeon_device pointer
  * @ib: indirect buffer to fill with commands
@@ -66,83 +100,89 @@ bool si_dma_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring)
  * @incr: increase next addr by incr bytes
  * @flags: access flags
  *
- * Update the page tables using the DMA (SI).
+ * Update PTEs by writing them manually using the DMA (SI).
  */
-void si_dma_vm_set_page(struct radeon_device *rdev,
-			struct radeon_ib *ib,
-			uint64_t pe,
-			uint64_t addr, unsigned count,
-			uint32_t incr, uint32_t flags)
+void si_dma_vm_write_pages(struct radeon_device *rdev,
+			   struct radeon_ib *ib,
+			   uint64_t pe,
+			   uint64_t addr, unsigned count,
+			   uint32_t incr, uint32_t flags)
 {
 	uint64_t value;
 	unsigned ndw;
 
-	trace_radeon_vm_set_page(pe, addr, count, incr, flags);
-
-	if (flags == R600_PTE_GART) {
-		uint64_t src = rdev->gart.table_addr + (addr >> 12) * 8;
-		while (count) {
-			unsigned bytes = count * 8;
-			if (bytes > 0xFFFF8)
-				bytes = 0xFFFF8;
-
-			ib->ptr[ib->length_dw++] = DMA_PACKET(DMA_PACKET_COPY,
-							      1, 0, 0, bytes);
-			ib->ptr[ib->length_dw++] = lower_32_bits(pe);
-			ib->ptr[ib->length_dw++] = lower_32_bits(src);
-			ib->ptr[ib->length_dw++] = upper_32_bits(pe) & 0xff;
-			ib->ptr[ib->length_dw++] = upper_32_bits(src) & 0xff;
-
-			pe += bytes;
-			src += bytes;
-			count -= bytes / 8;
-		}
-	} else if (flags & R600_PTE_SYSTEM) {
-		while (count) {
-			ndw = count * 2;
-			if (ndw > 0xFFFFE)
-				ndw = 0xFFFFE;
-
-			/* for non-physically contiguous pages (system) */
-			ib->ptr[ib->length_dw++] = DMA_PACKET(DMA_PACKET_WRITE, 0, 0, 0, ndw);
-			ib->ptr[ib->length_dw++] = pe;
-			ib->ptr[ib->length_dw++] = upper_32_bits(pe) & 0xff;
-			for (; ndw > 0; ndw -= 2, --count, pe += 8) {
+	while (count) {
+		ndw = count * 2;
+		if (ndw > 0xFFFFE)
+			ndw = 0xFFFFE;
+
+		/* for non-physically contiguous pages (system) */
+		ib->ptr[ib->length_dw++] = DMA_PACKET(DMA_PACKET_WRITE, 0, 0, 0, ndw);
+		ib->ptr[ib->length_dw++] = pe;
+		ib->ptr[ib->length_dw++] = upper_32_bits(pe) & 0xff;
+		for (; ndw > 0; ndw -= 2, --count, pe += 8) {
+			if (flags & R600_PTE_SYSTEM) {
 				value = radeon_vm_map_gart(rdev, addr);
 				value &= 0xFFFFFFFFFFFFF000ULL;
-				addr += incr;
-				value |= flags;
-				ib->ptr[ib->length_dw++] = value;
-				ib->ptr[ib->length_dw++] = upper_32_bits(value);
-			}
-		}
-	} else {
-		while (count) {
-			ndw = count * 2;
-			if (ndw > 0xFFFFE)
-				ndw = 0xFFFFE;
-
-			if (flags & R600_PTE_VALID)
+			} else if (flags & R600_PTE_VALID) {
 				value = addr;
-			else
+			} else {
 				value = 0;
-			/* for physically contiguous pages (vram) */
-			ib->ptr[ib->length_dw++] = DMA_PTE_PDE_PACKET(ndw);
-			ib->ptr[ib->length_dw++] = pe; /* dst addr */
-			ib->ptr[ib->length_dw++] = upper_32_bits(pe) & 0xff;
-			ib->ptr[ib->length_dw++] = flags; /* mask */
-			ib->ptr[ib->length_dw++] = 0;
-			ib->ptr[ib->length_dw++] = value; /* value */
+			}
+			addr += incr;
+			value |= flags;
+			ib->ptr[ib->length_dw++] = value;
 			ib->ptr[ib->length_dw++] = upper_32_bits(value);
-			ib->ptr[ib->length_dw++] = incr; /* increment size */
-			ib->ptr[ib->length_dw++] = 0;
-			pe += ndw * 4;
-			addr += (ndw / 2) * incr;
-			count -= ndw / 2;
 		}
 	}
-	while (ib->length_dw & 0x7)
-		ib->ptr[ib->length_dw++] = DMA_PACKET(DMA_PACKET_NOP, 0, 0, 0, 0);
+}
+
+/**
+ * si_dma_vm_set_pages - update the page tables using the DMA
+ *
+ * @rdev: radeon_device pointer
+ * @ib: indirect buffer to fill with commands
+ * @pe: addr of the page entry
+ * @addr: dst addr to write into pe
+ * @count: number of page entries to update
+ * @incr: increase next addr by incr bytes
+ * @flags: access flags
+ *
+ * Update the page tables using the DMA (SI).
+ */
+void si_dma_vm_set_pages(struct radeon_device *rdev,
+			 struct radeon_ib *ib,
+			 uint64_t pe,
+			 uint64_t addr, unsigned count,
+			 uint32_t incr, uint32_t flags)
+{
+	uint64_t value;
+	unsigned ndw;
+
+	while (count) {
+		ndw = count * 2;
+		if (ndw > 0xFFFFE)
+			ndw = 0xFFFFE;
+
+		if (flags & R600_PTE_VALID)
+			value = addr;
+		else
+			value = 0;
+
+		/* for physically contiguous pages (vram) */
+		ib->ptr[ib->length_dw++] = DMA_PTE_PDE_PACKET(ndw);
+		ib->ptr[ib->length_dw++] = pe; /* dst addr */
+		ib->ptr[ib->length_dw++] = upper_32_bits(pe) & 0xff;
+		ib->ptr[ib->length_dw++] = flags; /* mask */
+		ib->ptr[ib->length_dw++] = 0;
+		ib->ptr[ib->length_dw++] = value; /* value */
+		ib->ptr[ib->length_dw++] = upper_32_bits(value);
+		ib->ptr[ib->length_dw++] = incr; /* increment size */
+		ib->ptr[ib->length_dw++] = 0;
+		pe += ndw * 4;
+		addr += (ndw / 2) * incr;
+		count -= ndw / 2;
+	}
 }
 
 void si_dma_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm)
@@ -235,7 +275,7 @@ int si_copy_dma(struct radeon_device *rdev,
 		return r;
 	}
 
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev, ring, false);
 	radeon_semaphore_free(rdev, &sem, *fence);
 
 	return r;
diff --git a/drivers/gpu/drm/radeon/si_dpm.c b/drivers/gpu/drm/radeon/si_dpm.c
index 58918868f894..70e61ffeace2 100644
--- a/drivers/gpu/drm/radeon/si_dpm.c
+++ b/drivers/gpu/drm/radeon/si_dpm.c
@@ -3812,6 +3812,27 @@ void si_trim_voltage_table_to_fit_state_table(struct radeon_device *rdev,
 	voltage_table->count = max_voltage_steps;
 }
 
+static int si_get_svi2_voltage_table(struct radeon_device *rdev,
+				     struct radeon_clock_voltage_dependency_table *voltage_dependency_table,
+				     struct atom_voltage_table *voltage_table)
+{
+	u32 i;
+
+	if (voltage_dependency_table == NULL)
+		return -EINVAL;
+
+	voltage_table->mask_low = 0;
+	voltage_table->phase_delay = 0;
+
+	voltage_table->count = voltage_dependency_table->count;
+	for (i = 0; i < voltage_table->count; i++) {
+		voltage_table->entries[i].value = voltage_dependency_table->entries[i].v;
+		voltage_table->entries[i].smio_low = 0;
+	}
+
+	return 0;
+}
+
 static int si_construct_voltage_tables(struct radeon_device *rdev)
 {
 	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
@@ -3819,15 +3840,25 @@ static int si_construct_voltage_tables(struct radeon_device *rdev)
 	struct si_power_info *si_pi = si_get_pi(rdev);
 	int ret;
 
-	ret = radeon_atom_get_voltage_table(rdev, VOLTAGE_TYPE_VDDC,
-					    VOLTAGE_OBJ_GPIO_LUT, &eg_pi->vddc_voltage_table);
-	if (ret)
-		return ret;
+	if (pi->voltage_control) {
+		ret = radeon_atom_get_voltage_table(rdev, VOLTAGE_TYPE_VDDC,
+						    VOLTAGE_OBJ_GPIO_LUT, &eg_pi->vddc_voltage_table);
+		if (ret)
+			return ret;
 
-	if (eg_pi->vddc_voltage_table.count > SISLANDS_MAX_NO_VREG_STEPS)
-		si_trim_voltage_table_to_fit_state_table(rdev,
-							 SISLANDS_MAX_NO_VREG_STEPS,
-							 &eg_pi->vddc_voltage_table);
+		if (eg_pi->vddc_voltage_table.count > SISLANDS_MAX_NO_VREG_STEPS)
+			si_trim_voltage_table_to_fit_state_table(rdev,
+								 SISLANDS_MAX_NO_VREG_STEPS,
+								 &eg_pi->vddc_voltage_table);
+	} else if (si_pi->voltage_control_svi2) {
+		ret = si_get_svi2_voltage_table(rdev,
+						&rdev->pm.dpm.dyn_state.vddc_dependency_on_mclk,
+						&eg_pi->vddc_voltage_table);
+		if (ret)
+			return ret;
+	} else {
+		return -EINVAL;
+	}
 
 	if (eg_pi->vddci_control) {
 		ret = radeon_atom_get_voltage_table(rdev, VOLTAGE_TYPE_VDDCI,
@@ -3840,6 +3871,13 @@ static int si_construct_voltage_tables(struct radeon_device *rdev)
 								 SISLANDS_MAX_NO_VREG_STEPS,
 								 &eg_pi->vddci_voltage_table);
 	}
+	if (si_pi->vddci_control_svi2) {
+		ret = si_get_svi2_voltage_table(rdev,
+						&rdev->pm.dpm.dyn_state.vddci_dependency_on_mclk,
+						&eg_pi->vddci_voltage_table);
+		if (ret)
+			return ret;
+	}
 
 	if (pi->mvdd_control) {
 		ret = radeon_atom_get_voltage_table(rdev, VOLTAGE_TYPE_MVDDC,
@@ -3893,46 +3931,55 @@ static int si_populate_smc_voltage_tables(struct radeon_device *rdev,
 	struct si_power_info *si_pi = si_get_pi(rdev);
 	u8 i;
 
-	if (eg_pi->vddc_voltage_table.count) {
-		si_populate_smc_voltage_table(rdev, &eg_pi->vddc_voltage_table, table);
-		table->voltageMaskTable.lowMask[SISLANDS_SMC_VOLTAGEMASK_VDDC] =
-			cpu_to_be32(eg_pi->vddc_voltage_table.mask_low);
-
-		for (i = 0; i < eg_pi->vddc_voltage_table.count; i++) {
-			if (pi->max_vddc_in_table <= eg_pi->vddc_voltage_table.entries[i].value) {
-				table->maxVDDCIndexInPPTable = i;
-				break;
+	if (si_pi->voltage_control_svi2) {
+		si_write_smc_soft_register(rdev, SI_SMC_SOFT_REGISTER_svi_rework_gpio_id_svc,
+			si_pi->svc_gpio_id);
+		si_write_smc_soft_register(rdev, SI_SMC_SOFT_REGISTER_svi_rework_gpio_id_svd,
+			si_pi->svd_gpio_id);
+		si_write_smc_soft_register(rdev, SI_SMC_SOFT_REGISTER_svi_rework_plat_type,
+					   2);
+	} else {
+		if (eg_pi->vddc_voltage_table.count) {
+			si_populate_smc_voltage_table(rdev, &eg_pi->vddc_voltage_table, table);
+			table->voltageMaskTable.lowMask[SISLANDS_SMC_VOLTAGEMASK_VDDC] =
+				cpu_to_be32(eg_pi->vddc_voltage_table.mask_low);
+
+			for (i = 0; i < eg_pi->vddc_voltage_table.count; i++) {
+				if (pi->max_vddc_in_table <= eg_pi->vddc_voltage_table.entries[i].value) {
+					table->maxVDDCIndexInPPTable = i;
+					break;
+				}
 			}
 		}
-	}
 
-	if (eg_pi->vddci_voltage_table.count) {
-		si_populate_smc_voltage_table(rdev, &eg_pi->vddci_voltage_table, table);
+		if (eg_pi->vddci_voltage_table.count) {
+			si_populate_smc_voltage_table(rdev, &eg_pi->vddci_voltage_table, table);
 
-		table->voltageMaskTable.lowMask[SISLANDS_SMC_VOLTAGEMASK_VDDCI] =
-			cpu_to_be32(eg_pi->vddci_voltage_table.mask_low);
-	}
+			table->voltageMaskTable.lowMask[SISLANDS_SMC_VOLTAGEMASK_VDDCI] =
+				cpu_to_be32(eg_pi->vddci_voltage_table.mask_low);
+		}
 
 
-	if (si_pi->mvdd_voltage_table.count) {
-		si_populate_smc_voltage_table(rdev, &si_pi->mvdd_voltage_table, table);
+		if (si_pi->mvdd_voltage_table.count) {
+			si_populate_smc_voltage_table(rdev, &si_pi->mvdd_voltage_table, table);
 
-		table->voltageMaskTable.lowMask[SISLANDS_SMC_VOLTAGEMASK_MVDD] =
-			cpu_to_be32(si_pi->mvdd_voltage_table.mask_low);
-	}
+			table->voltageMaskTable.lowMask[SISLANDS_SMC_VOLTAGEMASK_MVDD] =
+				cpu_to_be32(si_pi->mvdd_voltage_table.mask_low);
+		}
 
-	if (si_pi->vddc_phase_shed_control) {
-		if (si_validate_phase_shedding_tables(rdev, &si_pi->vddc_phase_shed_table,
-						      &rdev->pm.dpm.dyn_state.phase_shedding_limits_table)) {
-			si_populate_smc_voltage_table(rdev, &si_pi->vddc_phase_shed_table, table);
+		if (si_pi->vddc_phase_shed_control) {
+			if (si_validate_phase_shedding_tables(rdev, &si_pi->vddc_phase_shed_table,
+							      &rdev->pm.dpm.dyn_state.phase_shedding_limits_table)) {
+				si_populate_smc_voltage_table(rdev, &si_pi->vddc_phase_shed_table, table);
 
-			table->phaseMaskTable.lowMask[SISLANDS_SMC_VOLTAGEMASK_VDDC] =
-				cpu_to_be32(si_pi->vddc_phase_shed_table.mask_low);
+				table->phaseMaskTable.lowMask[SISLANDS_SMC_VOLTAGEMASK_VDDC] =
+					cpu_to_be32(si_pi->vddc_phase_shed_table.mask_low);
 
-			si_write_smc_soft_register(rdev, SI_SMC_SOFT_REGISTER_phase_shedding_delay,
-						   (u32)si_pi->vddc_phase_shed_table.phase_delay);
-		} else {
-			si_pi->vddc_phase_shed_control = false;
+				si_write_smc_soft_register(rdev, SI_SMC_SOFT_REGISTER_phase_shedding_delay,
+							   (u32)si_pi->vddc_phase_shed_table.phase_delay);
+			} else {
+				si_pi->vddc_phase_shed_control = false;
+			}
 		}
 	}
 
@@ -5798,16 +5845,17 @@ int si_dpm_enable(struct radeon_device *rdev)
 {
 	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
 	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct si_power_info *si_pi = si_get_pi(rdev);
 	struct radeon_ps *boot_ps = rdev->pm.dpm.boot_ps;
 	int ret;
 
 	if (si_is_smc_running(rdev))
 		return -EINVAL;
-	if (pi->voltage_control)
+	if (pi->voltage_control || si_pi->voltage_control_svi2)
 		si_enable_voltage_control(rdev, true);
 	if (pi->mvdd_control)
 		si_get_mvdd_configuration(rdev);
-	if (pi->voltage_control) {
+	if (pi->voltage_control || si_pi->voltage_control_svi2) {
 		ret = si_construct_voltage_tables(rdev);
 		if (ret) {
 			DRM_ERROR("si_construct_voltage_tables failed\n");
@@ -6406,16 +6454,32 @@ int si_dpm_init(struct radeon_device *rdev)
 	ni_pi->mclk_rtt_mode_threshold = eg_pi->mclk_edc_wr_enable_threshold;
 
 	pi->voltage_control =
-		radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_VDDC, VOLTAGE_OBJ_GPIO_LUT);
+		radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_VDDC,
+					    VOLTAGE_OBJ_GPIO_LUT);
+	if (!pi->voltage_control) {
+		si_pi->voltage_control_svi2 =
+			radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_VDDC,
+						    VOLTAGE_OBJ_SVID2);
+		if (si_pi->voltage_control_svi2)
+			radeon_atom_get_svi2_info(rdev, SET_VOLTAGE_TYPE_ASIC_VDDC,
+						  &si_pi->svd_gpio_id, &si_pi->svc_gpio_id);
+	}
 
 	pi->mvdd_control =
-		radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_MVDDC, VOLTAGE_OBJ_GPIO_LUT);
+		radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_MVDDC,
+					    VOLTAGE_OBJ_GPIO_LUT);
 
 	eg_pi->vddci_control =
-		radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_VDDCI, VOLTAGE_OBJ_GPIO_LUT);
+		radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_VDDCI,
+					    VOLTAGE_OBJ_GPIO_LUT);
+	if (!eg_pi->vddci_control)
+		si_pi->vddci_control_svi2 =
+			radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_VDDCI,
+						    VOLTAGE_OBJ_SVID2);
 
 	si_pi->vddc_phase_shed_control =
-		radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_VDDC, VOLTAGE_OBJ_PHASE_LUT);
+		radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_VDDC,
+					    VOLTAGE_OBJ_PHASE_LUT);
 
 	rv770_get_engine_memory_ss(rdev);
 
diff --git a/drivers/gpu/drm/radeon/si_dpm.h b/drivers/gpu/drm/radeon/si_dpm.h
index 4ce5032cdf49..8b5c06a0832d 100644
--- a/drivers/gpu/drm/radeon/si_dpm.h
+++ b/drivers/gpu/drm/radeon/si_dpm.h
@@ -170,6 +170,8 @@ struct si_power_info {
 	bool vddc_phase_shed_control;
 	bool pspp_notify_required;
 	bool sclk_deep_sleep_above_low;
+	bool voltage_control_svi2;
+	bool vddci_control_svi2;
 	/* smc offsets */
 	u32 sram_end;
 	u32 state_table_start;
@@ -192,6 +194,9 @@ struct si_power_info {
 	SMC_SIslands_MCRegisters smc_mc_reg_table;
 	SISLANDS_SMC_STATETABLE smc_statetable;
 	PP_SIslands_PAPMParameters papm_parm;
+	/* SVI2 */
+	u8 svd_gpio_id;
+	u8 svc_gpio_id;
 };
 
 #define SISLANDS_INITIAL_STATE_ARB_INDEX    0
diff --git a/drivers/gpu/drm/radeon/si_smc.c b/drivers/gpu/drm/radeon/si_smc.c
index e80efcf0c230..73dbc79c959d 100644
--- a/drivers/gpu/drm/radeon/si_smc.c
+++ b/drivers/gpu/drm/radeon/si_smc.c
@@ -219,36 +219,48 @@ int si_load_smc_ucode(struct radeon_device *rdev, u32 limit)
 	if (!rdev->smc_fw)
 		return -EINVAL;
 
-	switch (rdev->family) {
-	case CHIP_TAHITI:
-		ucode_start_address = TAHITI_SMC_UCODE_START;
-		ucode_size = TAHITI_SMC_UCODE_SIZE;
-		break;
-	case CHIP_PITCAIRN:
-		ucode_start_address = PITCAIRN_SMC_UCODE_START;
-		ucode_size = PITCAIRN_SMC_UCODE_SIZE;
-		break;
-	case CHIP_VERDE:
-		ucode_start_address = VERDE_SMC_UCODE_START;
-		ucode_size = VERDE_SMC_UCODE_SIZE;
-		break;
-	case CHIP_OLAND:
-		ucode_start_address = OLAND_SMC_UCODE_START;
-		ucode_size = OLAND_SMC_UCODE_SIZE;
-		break;
-	case CHIP_HAINAN:
-		ucode_start_address = HAINAN_SMC_UCODE_START;
-		ucode_size = HAINAN_SMC_UCODE_SIZE;
-		break;
-	default:
-		DRM_ERROR("unknown asic in smc ucode loader\n");
-		BUG();
+	if (rdev->new_fw) {
+		const struct smc_firmware_header_v1_0 *hdr =
+			(const struct smc_firmware_header_v1_0 *)rdev->smc_fw->data;
+
+		radeon_ucode_print_smc_hdr(&hdr->header);
+
+		ucode_start_address = le32_to_cpu(hdr->ucode_start_addr);
+		ucode_size = le32_to_cpu(hdr->header.ucode_size_bytes);
+		src = (const u8 *)
+			(rdev->smc_fw->data + le32_to_cpu(hdr->header.ucode_array_offset_bytes));
+	} else {
+		switch (rdev->family) {
+		case CHIP_TAHITI:
+			ucode_start_address = TAHITI_SMC_UCODE_START;
+			ucode_size = TAHITI_SMC_UCODE_SIZE;
+			break;
+		case CHIP_PITCAIRN:
+			ucode_start_address = PITCAIRN_SMC_UCODE_START;
+			ucode_size = PITCAIRN_SMC_UCODE_SIZE;
+			break;
+		case CHIP_VERDE:
+			ucode_start_address = VERDE_SMC_UCODE_START;
+			ucode_size = VERDE_SMC_UCODE_SIZE;
+			break;
+		case CHIP_OLAND:
+			ucode_start_address = OLAND_SMC_UCODE_START;
+			ucode_size = OLAND_SMC_UCODE_SIZE;
+			break;
+		case CHIP_HAINAN:
+			ucode_start_address = HAINAN_SMC_UCODE_START;
+			ucode_size = HAINAN_SMC_UCODE_SIZE;
+			break;
+		default:
+			DRM_ERROR("unknown asic in smc ucode loader\n");
+			BUG();
+		}
+		src = (const u8 *)rdev->smc_fw->data;
 	}
 
 	if (ucode_size & 3)
 		return -EINVAL;
 
-	src = (const u8 *)rdev->smc_fw->data;
 	spin_lock_irqsave(&rdev->smc_idx_lock, flags);
 	WREG32(SMC_IND_INDEX_0, ucode_start_address);
 	WREG32_P(SMC_IND_ACCESS_CNTL, AUTO_INCREMENT_IND_0, ~AUTO_INCREMENT_IND_0);
diff --git a/drivers/gpu/drm/radeon/sislands_smc.h b/drivers/gpu/drm/radeon/sislands_smc.h
index 10e945a49479..623a0b1e2d9d 100644
--- a/drivers/gpu/drm/radeon/sislands_smc.h
+++ b/drivers/gpu/drm/radeon/sislands_smc.h
@@ -241,6 +241,9 @@ typedef struct SISLANDS_SMC_STATETABLE SISLANDS_SMC_STATETABLE;
 #define SI_SMC_SOFT_REGISTER_non_ulv_pcie_link_width  0xF4
 #define SI_SMC_SOFT_REGISTER_tdr_is_about_to_happen   0xFC
 #define SI_SMC_SOFT_REGISTER_vr_hot_gpio              0x100
+#define SI_SMC_SOFT_REGISTER_svi_rework_plat_type     0x118
+#define SI_SMC_SOFT_REGISTER_svi_rework_gpio_id_svd   0x11c
+#define SI_SMC_SOFT_REGISTER_svi_rework_gpio_id_svc   0x120
 
 #define SMC_SISLANDS_LKGE_LUT_NUM_OF_TEMP_ENTRIES 16
 #define SMC_SISLANDS_LKGE_LUT_NUM_OF_VOLT_ENTRIES 32
diff --git a/drivers/gpu/drm/radeon/trinity_dpm.c b/drivers/gpu/drm/radeon/trinity_dpm.c
index 20da6ff183df..57f780053b3e 100644
--- a/drivers/gpu/drm/radeon/trinity_dpm.c
+++ b/drivers/gpu/drm/radeon/trinity_dpm.c
@@ -1874,15 +1874,22 @@ int trinity_dpm_init(struct radeon_device *rdev)
 	for (i = 0; i < SUMO_MAX_HARDWARE_POWERLEVELS; i++)
 		pi->at[i] = TRINITY_AT_DFLT;
 
-	/* There are stability issues reported on latops with
-	 * bapm installed when switching between AC and battery
-	 * power.  At the same time, some desktop boards hang
-	 * if it's not enabled and dpm is enabled.
-	 */
-	if (rdev->flags & RADEON_IS_MOBILITY)
+	if (radeon_bapm == -1) {
+		/* There are stability issues reported on with
+		 * bapm enabled when switching between AC and battery
+		 * power.  At the same time, some MSI boards hang
+		 * if it's not enabled and dpm is enabled.  Just enable
+		 * it for MSI boards right now.
+		 */
+		if (rdev->pdev->subsystem_vendor == 0x1462)
+			pi->enable_bapm = true;
+		else
+			pi->enable_bapm = false;
+	} else if (radeon_bapm == 0) {
 		pi->enable_bapm = false;
-	else
+	} else {
 		pi->enable_bapm = true;
+	}
 	pi->enable_nbps_policy = true;
 	pi->enable_sclk_ds = true;
 	pi->enable_gfx_power_gating = true;
diff --git a/drivers/gpu/drm/radeon/uvd_v1_0.c b/drivers/gpu/drm/radeon/uvd_v1_0.c
index be42c8125203..cda391347286 100644
--- a/drivers/gpu/drm/radeon/uvd_v1_0.c
+++ b/drivers/gpu/drm/radeon/uvd_v1_0.c
@@ -124,7 +124,7 @@ int uvd_v1_0_init(struct radeon_device *rdev)
 	radeon_ring_write(ring, PACKET0(UVD_SEMA_CNTL, 0));
 	radeon_ring_write(ring, 3);
 
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev, ring, false);
 
 done:
 	/* lower clocks again */
@@ -331,7 +331,7 @@ int uvd_v1_0_ring_test(struct radeon_device *rdev, struct radeon_ring *ring)
 	}
 	radeon_ring_write(ring, PACKET0(UVD_CONTEXT_ID, 0));
 	radeon_ring_write(ring, 0xDEADBEEF);
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev, ring, false);
 	for (i = 0; i < rdev->usec_timeout; i++) {
 		tmp = RREG32(UVD_CONTEXT_ID);
 		if (tmp == 0xDEADBEEF)