Merge tag 'drm-next-2020-06-08' of git://anongit.freedesktop.org/drm/drm

Pull drm fixes from Dave Airlie:
 "These are the fixes from last week for the stuff merged in the merge
  window. It got a bunch of nouveau fixes for HDA audio on some new
  GPUs, some i915 and some amdpgu fixes.

  i915:
   - gvt: Fix one clang warning on debug only function
   - Use ARRAY_SIZE for coccicheck warning
   - Use after free fix for display global state.
   - Whitelisting context-local timestamp on Gen9 and two scheduler
     fixes with deps (Cc: stable)
   - Removal of write flag from sysfs files where ineffective

  nouveau:
   - HDMI/DP audio HDA fixes
   - display hang fix for Volta/Turing
   - GK20A regression fix.

  amdgpu:
   - Prevent hwmon accesses while GPU is in reset
   - CTF interrupt fix
   - Backlight fix for renoir
   - Fix for display sync groups
   - Display bandwidth validation workaround"

* tag 'drm-next-2020-06-08' of git://anongit.freedesktop.org/drm/drm: (28 commits)
  drm/nouveau/kms/nv50-: clear SW state of disabled windows harder
  drm/nouveau: gr/gk20a: Use firmware version 0
  drm/nouveau/disp/gm200-: detect and potentially disable HDA support on some SORs
  drm/nouveau/disp/gp100: split SOR implementation from gm200
  drm/nouveau/disp: modify OR allocation policy to account for HDA requirements
  drm/nouveau/disp: split part of OR allocation logic into a function
  drm/nouveau/disp: provide hint to OR allocation about HDA requirements
  drm/amd/display: Revalidate bandwidth before commiting DC updates
  drm/amdgpu/display: use blanked rather than plane state for sync groups
  drm/i915/params: fix i915.fake_lmem_start module param sysfs permissions
  drm/i915/params: don't expose inject_probe_failure in debugfs
  drm/i915: Whitelist context-local timestamp in the gen9 cmdparser
  drm/i915: Fix global state use-after-frees with a refcount
  drm/i915: Check for awaits on still currently executing requests
  drm/i915/gt: Do not schedule normal requests immediately along virtual
  drm/i915: Reorder await_execution before await_request
  drm/nouveau/kms/gt215-: fix race with audio driver runpm
  drm/nouveau/disp/gm200-: fix NV_PDISP_SOR_HDMI2_CTRL(n) selection
  Revert "drm/amd/display: disable dcn20 abm feature for bring up"
  drm/amd/powerplay: ack the SMUToHost interrupt on receive V2
  ...

32 files changed, 778 insertions, 211 deletions

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_pm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_pm.c
index d7646cbce346e0..775e389c9a13e4 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_pm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_pm.c
@@ -163,6 +163,9 @@ static ssize_t amdgpu_get_power_dpm_state(struct device *dev,
 	enum amd_pm_state_type pm;
 	int ret;
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	ret = pm_runtime_get_sync(ddev->dev);
 	if (ret < 0)
 		return ret;
@@ -196,6 +199,9 @@ static ssize_t amdgpu_set_power_dpm_state(struct device *dev,
 	enum amd_pm_state_type  state;
 	int ret;
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	if (strncmp("battery", buf, strlen("battery")) == 0)
 		state = POWER_STATE_TYPE_BATTERY;
 	else if (strncmp("balanced", buf, strlen("balanced")) == 0)
@@ -297,6 +303,9 @@ static ssize_t amdgpu_get_power_dpm_force_performance_level(struct device *dev,
 	enum amd_dpm_forced_level level = 0xff;
 	int ret;
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	ret = pm_runtime_get_sync(ddev->dev);
 	if (ret < 0)
 		return ret;
@@ -334,6 +343,9 @@ static ssize_t amdgpu_set_power_dpm_force_performance_level(struct device *dev,
 	enum amd_dpm_forced_level current_level = 0xff;
 	int ret = 0;
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	if (strncmp("low", buf, strlen("low")) == 0) {
 		level = AMD_DPM_FORCED_LEVEL_LOW;
 	} else if (strncmp("high", buf, strlen("high")) == 0) {
@@ -433,6 +445,9 @@ static ssize_t amdgpu_get_pp_num_states(struct device *dev,
 	struct pp_states_info data;
 	int i, buf_len, ret;
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	ret = pm_runtime_get_sync(ddev->dev);
 	if (ret < 0)
 		return ret;
@@ -472,6 +487,9 @@ static ssize_t amdgpu_get_pp_cur_state(struct device *dev,
 	enum amd_pm_state_type pm = 0;
 	int i = 0, ret = 0;
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	ret = pm_runtime_get_sync(ddev->dev);
 	if (ret < 0)
 		return ret;
@@ -508,6 +526,9 @@ static ssize_t amdgpu_get_pp_force_state(struct device *dev,
 	struct drm_device *ddev = dev_get_drvdata(dev);
 	struct amdgpu_device *adev = ddev->dev_private;
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	if (adev->pp_force_state_enabled)
 		return amdgpu_get_pp_cur_state(dev, attr, buf);
 	else
@@ -525,6 +546,9 @@ static ssize_t amdgpu_set_pp_force_state(struct device *dev,
 	unsigned long idx;
 	int ret;
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	if (strlen(buf) == 1)
 		adev->pp_force_state_enabled = false;
 	else if (is_support_sw_smu(adev))
@@ -580,6 +604,9 @@ static ssize_t amdgpu_get_pp_table(struct device *dev,
 	char *table = NULL;
 	int size, ret;
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	ret = pm_runtime_get_sync(ddev->dev);
 	if (ret < 0)
 		return ret;
@@ -619,6 +646,9 @@ static ssize_t amdgpu_set_pp_table(struct device *dev,
 	struct amdgpu_device *adev = ddev->dev_private;
 	int ret = 0;
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	ret = pm_runtime_get_sync(ddev->dev);
 	if (ret < 0)
 		return ret;
@@ -721,6 +751,9 @@ static ssize_t amdgpu_set_pp_od_clk_voltage(struct device *dev,
 	const char delimiter[3] = {' ', '\n', '\0'};
 	uint32_t type;
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	if (count > 127)
 		return -EINVAL;
 
@@ -810,6 +843,9 @@ static ssize_t amdgpu_get_pp_od_clk_voltage(struct device *dev,
 	ssize_t size;
 	int ret;
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	ret = pm_runtime_get_sync(ddev->dev);
 	if (ret < 0)
 		return ret;
@@ -859,6 +895,9 @@ static ssize_t amdgpu_set_pp_features(struct device *dev,
 	uint64_t featuremask;
 	int ret;
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	ret = kstrtou64(buf, 0, &featuremask);
 	if (ret)
 		return -EINVAL;
@@ -899,6 +938,9 @@ static ssize_t amdgpu_get_pp_features(struct device *dev,
 	ssize_t size;
 	int ret;
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	ret = pm_runtime_get_sync(ddev->dev);
 	if (ret < 0)
 		return ret;
@@ -955,6 +997,9 @@ static ssize_t amdgpu_get_pp_dpm_sclk(struct device *dev,
 	ssize_t size;
 	int ret;
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	ret = pm_runtime_get_sync(ddev->dev);
 	if (ret < 0)
 		return ret;
@@ -1018,6 +1063,9 @@ static ssize_t amdgpu_set_pp_dpm_sclk(struct device *dev,
 	int ret;
 	uint32_t mask = 0;
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	ret = amdgpu_read_mask(buf, count, &mask);
 	if (ret)
 		return ret;
@@ -1049,6 +1097,9 @@ static ssize_t amdgpu_get_pp_dpm_mclk(struct device *dev,
 	ssize_t size;
 	int ret;
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	ret = pm_runtime_get_sync(ddev->dev);
 	if (ret < 0)
 		return ret;
@@ -1076,6 +1127,9 @@ static ssize_t amdgpu_set_pp_dpm_mclk(struct device *dev,
 	uint32_t mask = 0;
 	int ret;
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	ret = amdgpu_read_mask(buf, count, &mask);
 	if (ret)
 		return ret;
@@ -1107,6 +1161,9 @@ static ssize_t amdgpu_get_pp_dpm_socclk(struct device *dev,
 	ssize_t size;
 	int ret;
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	ret = pm_runtime_get_sync(ddev->dev);
 	if (ret < 0)
 		return ret;
@@ -1134,6 +1191,9 @@ static ssize_t amdgpu_set_pp_dpm_socclk(struct device *dev,
 	int ret;
 	uint32_t mask = 0;
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	ret = amdgpu_read_mask(buf, count, &mask);
 	if (ret)
 		return ret;
@@ -1167,6 +1227,9 @@ static ssize_t amdgpu_get_pp_dpm_fclk(struct device *dev,
 	ssize_t size;
 	int ret;
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	ret = pm_runtime_get_sync(ddev->dev);
 	if (ret < 0)
 		return ret;
@@ -1194,6 +1257,9 @@ static ssize_t amdgpu_set_pp_dpm_fclk(struct device *dev,
 	int ret;
 	uint32_t mask = 0;
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	ret = amdgpu_read_mask(buf, count, &mask);
 	if (ret)
 		return ret;
@@ -1227,6 +1293,9 @@ static ssize_t amdgpu_get_pp_dpm_dcefclk(struct device *dev,
 	ssize_t size;
 	int ret;
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	ret = pm_runtime_get_sync(ddev->dev);
 	if (ret < 0)
 		return ret;
@@ -1254,6 +1323,9 @@ static ssize_t amdgpu_set_pp_dpm_dcefclk(struct device *dev,
 	int ret;
 	uint32_t mask = 0;
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	ret = amdgpu_read_mask(buf, count, &mask);
 	if (ret)
 		return ret;
@@ -1287,6 +1359,9 @@ static ssize_t amdgpu_get_pp_dpm_pcie(struct device *dev,
 	ssize_t size;
 	int ret;
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	ret = pm_runtime_get_sync(ddev->dev);
 	if (ret < 0)
 		return ret;
@@ -1314,6 +1389,9 @@ static ssize_t amdgpu_set_pp_dpm_pcie(struct device *dev,
 	int ret;
 	uint32_t mask = 0;
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	ret = amdgpu_read_mask(buf, count, &mask);
 	if (ret)
 		return ret;
@@ -1347,6 +1425,9 @@ static ssize_t amdgpu_get_pp_sclk_od(struct device *dev,
 	uint32_t value = 0;
 	int ret;
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	ret = pm_runtime_get_sync(ddev->dev);
 	if (ret < 0)
 		return ret;
@@ -1372,6 +1453,9 @@ static ssize_t amdgpu_set_pp_sclk_od(struct device *dev,
 	int ret;
 	long int value;
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	ret = kstrtol(buf, 0, &value);
 
 	if (ret)
@@ -1410,6 +1494,9 @@ static ssize_t amdgpu_get_pp_mclk_od(struct device *dev,
 	uint32_t value = 0;
 	int ret;
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	ret = pm_runtime_get_sync(ddev->dev);
 	if (ret < 0)
 		return ret;
@@ -1435,6 +1522,9 @@ static ssize_t amdgpu_set_pp_mclk_od(struct device *dev,
 	int ret;
 	long int value;
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	ret = kstrtol(buf, 0, &value);
 
 	if (ret)
@@ -1493,6 +1583,9 @@ static ssize_t amdgpu_get_pp_power_profile_mode(struct device *dev,
 	ssize_t size;
 	int ret;
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	ret = pm_runtime_get_sync(ddev->dev);
 	if (ret < 0)
 		return ret;
@@ -1528,6 +1621,9 @@ static ssize_t amdgpu_set_pp_power_profile_mode(struct device *dev,
 	long int profile_mode = 0;
 	const char delimiter[3] = {' ', '\n', '\0'};
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	tmp[0] = *(buf);
 	tmp[1] = '\0';
 	ret = kstrtol(tmp, 0, &profile_mode);
@@ -1587,6 +1683,9 @@ static ssize_t amdgpu_get_gpu_busy_percent(struct device *dev,
 	struct amdgpu_device *adev = ddev->dev_private;
 	int r, value, size = sizeof(value);
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	r = pm_runtime_get_sync(ddev->dev);
 	if (r < 0)
 		return r;
@@ -1620,6 +1719,9 @@ static ssize_t amdgpu_get_mem_busy_percent(struct device *dev,
 	struct amdgpu_device *adev = ddev->dev_private;
 	int r, value, size = sizeof(value);
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	r = pm_runtime_get_sync(ddev->dev);
 	if (r < 0)
 		return r;
@@ -1658,6 +1760,9 @@ static ssize_t amdgpu_get_pcie_bw(struct device *dev,
 	uint64_t count0 = 0, count1 = 0;
 	int ret;
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	if (adev->flags & AMD_IS_APU)
 		return -ENODATA;
 
@@ -1694,6 +1799,9 @@ static ssize_t amdgpu_get_unique_id(struct device *dev,
 	struct drm_device *ddev = dev_get_drvdata(dev);
 	struct amdgpu_device *adev = ddev->dev_private;
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	if (adev->unique_id)
 		return snprintf(buf, PAGE_SIZE, "%016llx\n", adev->unique_id);
 
@@ -1888,6 +1996,9 @@ static ssize_t amdgpu_hwmon_show_temp(struct device *dev,
 	int channel = to_sensor_dev_attr(attr)->index;
 	int r, temp = 0, size = sizeof(temp);
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	if (channel >= PP_TEMP_MAX)
 		return -EINVAL;
 
@@ -2019,6 +2130,9 @@ static ssize_t amdgpu_hwmon_get_pwm1_enable(struct device *dev,
 	u32 pwm_mode = 0;
 	int ret;
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	ret = pm_runtime_get_sync(adev->ddev->dev);
 	if (ret < 0)
 		return ret;
@@ -2050,6 +2164,9 @@ static ssize_t amdgpu_hwmon_set_pwm1_enable(struct device *dev,
 	int err, ret;
 	int value;
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	err = kstrtoint(buf, 10, &value);
 	if (err)
 		return err;
@@ -2099,6 +2216,9 @@ static ssize_t amdgpu_hwmon_set_pwm1(struct device *dev,
 	u32 value;
 	u32 pwm_mode;
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	err = pm_runtime_get_sync(adev->ddev->dev);
 	if (err < 0)
 		return err;
@@ -2148,6 +2268,9 @@ static ssize_t amdgpu_hwmon_get_pwm1(struct device *dev,
 	int err;
 	u32 speed = 0;
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	err = pm_runtime_get_sync(adev->ddev->dev);
 	if (err < 0)
 		return err;
@@ -2178,6 +2301,9 @@ static ssize_t amdgpu_hwmon_get_fan1_input(struct device *dev,
 	int err;
 	u32 speed = 0;
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	err = pm_runtime_get_sync(adev->ddev->dev);
 	if (err < 0)
 		return err;
@@ -2207,6 +2333,9 @@ static ssize_t amdgpu_hwmon_get_fan1_min(struct device *dev,
 	u32 size = sizeof(min_rpm);
 	int r;
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	r = pm_runtime_get_sync(adev->ddev->dev);
 	if (r < 0)
 		return r;
@@ -2232,6 +2361,9 @@ static ssize_t amdgpu_hwmon_get_fan1_max(struct device *dev,
 	u32 size = sizeof(max_rpm);
 	int r;
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	r = pm_runtime_get_sync(adev->ddev->dev);
 	if (r < 0)
 		return r;
@@ -2256,6 +2388,9 @@ static ssize_t amdgpu_hwmon_get_fan1_target(struct device *dev,
 	int err;
 	u32 rpm = 0;
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	err = pm_runtime_get_sync(adev->ddev->dev);
 	if (err < 0)
 		return err;
@@ -2285,6 +2420,9 @@ static ssize_t amdgpu_hwmon_set_fan1_target(struct device *dev,
 	u32 value;
 	u32 pwm_mode;
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	err = pm_runtime_get_sync(adev->ddev->dev);
 	if (err < 0)
 		return err;
@@ -2331,6 +2469,9 @@ static ssize_t amdgpu_hwmon_get_fan1_enable(struct device *dev,
 	u32 pwm_mode = 0;
 	int ret;
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	ret = pm_runtime_get_sync(adev->ddev->dev);
 	if (ret < 0)
 		return ret;
@@ -2363,6 +2504,9 @@ static ssize_t amdgpu_hwmon_set_fan1_enable(struct device *dev,
 	int value;
 	u32 pwm_mode;
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	err = kstrtoint(buf, 10, &value);
 	if (err)
 		return err;
@@ -2403,6 +2547,9 @@ static ssize_t amdgpu_hwmon_show_vddgfx(struct device *dev,
 	u32 vddgfx;
 	int r, size = sizeof(vddgfx);
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	r = pm_runtime_get_sync(adev->ddev->dev);
 	if (r < 0)
 		return r;
@@ -2435,6 +2582,9 @@ static ssize_t amdgpu_hwmon_show_vddnb(struct device *dev,
 	u32 vddnb;
 	int r, size = sizeof(vddnb);
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	/* only APUs have vddnb */
 	if  (!(adev->flags & AMD_IS_APU))
 		return -EINVAL;
@@ -2472,6 +2622,9 @@ static ssize_t amdgpu_hwmon_show_power_avg(struct device *dev,
 	int r, size = sizeof(u32);
 	unsigned uw;
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	r = pm_runtime_get_sync(adev->ddev->dev);
 	if (r < 0)
 		return r;
@@ -2508,6 +2661,9 @@ static ssize_t amdgpu_hwmon_show_power_cap_max(struct device *dev,
 	ssize_t size;
 	int r;
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	r = pm_runtime_get_sync(adev->ddev->dev);
 	if (r < 0)
 		return r;
@@ -2537,6 +2693,9 @@ static ssize_t amdgpu_hwmon_show_power_cap(struct device *dev,
 	ssize_t size;
 	int r;
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	r = pm_runtime_get_sync(adev->ddev->dev);
 	if (r < 0)
 		return r;
@@ -2567,6 +2726,9 @@ static ssize_t amdgpu_hwmon_set_power_cap(struct device *dev,
 	int err;
 	u32 value;
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	if (amdgpu_sriov_vf(adev))
 		return -EINVAL;
 
@@ -2605,6 +2767,9 @@ static ssize_t amdgpu_hwmon_show_sclk(struct device *dev,
 	uint32_t sclk;
 	int r, size = sizeof(sclk);
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	r = pm_runtime_get_sync(adev->ddev->dev);
 	if (r < 0)
 		return r;
@@ -2637,6 +2802,9 @@ static ssize_t amdgpu_hwmon_show_mclk(struct device *dev,
 	uint32_t mclk;
 	int r, size = sizeof(mclk);
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	r = pm_runtime_get_sync(adev->ddev->dev);
 	if (r < 0)
 		return r;
@@ -3497,6 +3665,9 @@ static int amdgpu_debugfs_pm_info(struct seq_file *m, void *data)
 	u32 flags = 0;
 	int r;
 
+	if (adev->in_gpu_reset)
+		return -EPERM;
+
 	r = pm_runtime_get_sync(dev->dev);
 	if (r < 0)
 		return r;
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
index bdba0bfd6df1be..7ced9f87be9793 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
@@ -1356,7 +1356,7 @@ static int dm_late_init(void *handle)
 	unsigned int linear_lut[16];
 	int i;
 	struct dmcu *dmcu = NULL;
-	bool ret = false;
+	bool ret;
 
 	if (!adev->dm.fw_dmcu)
 		return detect_mst_link_for_all_connectors(adev->ddev);
@@ -1377,13 +1377,10 @@ static int dm_late_init(void *handle)
 	 */
 	params.min_abm_backlight = 0x28F;
 
-	/* todo will enable for navi10 */
-	if (adev->asic_type <= CHIP_RAVEN) {
-		ret = dmcu_load_iram(dmcu, params);
+	ret = dmcu_load_iram(dmcu, params);
 
-		if (!ret)
-			return -EINVAL;
-	}
+	if (!ret)
+		return -EINVAL;
 
 	return detect_mst_link_for_all_connectors(adev->ddev);
 }
diff --git a/drivers/gpu/drm/amd/display/dc/core/dc.c b/drivers/gpu/drm/amd/display/dc/core/dc.c
index 45cfb7c45566af..6f93a6ca4cf0c8 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc.c
@@ -1016,9 +1016,17 @@ static void program_timing_sync(
 			}
 		}
 
-		/* set first pipe with plane as master */
+		/* set first unblanked pipe as master */
 		for (j = 0; j < group_size; j++) {
-			if (pipe_set[j]->plane_state) {
+			bool is_blanked;
+
+			if (pipe_set[j]->stream_res.opp->funcs->dpg_is_blanked)
+				is_blanked =
+					pipe_set[j]->stream_res.opp->funcs->dpg_is_blanked(pipe_set[j]->stream_res.opp);
+			else
+				is_blanked =
+					pipe_set[j]->stream_res.tg->funcs->is_blanked(pipe_set[j]->stream_res.tg);
+			if (!is_blanked) {
 				if (j == 0)
 					break;
 
@@ -1039,9 +1047,17 @@ static void program_timing_sync(
 				status->timing_sync_info.master = false;
 
 		}
-		/* remove any other pipes with plane as they have already been synced */
+		/* remove any other unblanked pipes as they have already been synced */
 		for (j = j + 1; j < group_size; j++) {
-			if (pipe_set[j]->plane_state) {
+			bool is_blanked;
+
+			if (pipe_set[j]->stream_res.opp->funcs->dpg_is_blanked)
+				is_blanked =
+					pipe_set[j]->stream_res.opp->funcs->dpg_is_blanked(pipe_set[j]->stream_res.opp);
+			else
+				is_blanked =
+					pipe_set[j]->stream_res.tg->funcs->is_blanked(pipe_set[j]->stream_res.tg);
+			if (!is_blanked) {
 				group_size--;
 				pipe_set[j] = pipe_set[group_size];
 				j--;
@@ -2522,6 +2538,12 @@ void dc_commit_updates_for_stream(struct dc *dc,
 
 	copy_stream_update_to_stream(dc, context, stream, stream_update);
 
+	if (!dc->res_pool->funcs->validate_bandwidth(dc, context, false)) {
+		DC_ERROR("Mode validation failed for stream update!\n");
+		dc_release_state(context);
+		return;
+	}
+
 	commit_planes_for_stream(
 				dc,
 				srf_updates,
diff --git a/drivers/gpu/drm/amd/powerplay/smu_v11_0.c b/drivers/gpu/drm/amd/powerplay/smu_v11_0.c
index ae0361e225bb27..aa76c2cea7471b 100644
--- a/drivers/gpu/drm/amd/powerplay/smu_v11_0.c
+++ b/drivers/gpu/drm/amd/powerplay/smu_v11_0.c
@@ -1561,6 +1561,7 @@ static int smu_v11_0_irq_process(struct amdgpu_device *adev,
 	 * events for SMCToHost interrupt.
 	 */
 	uint32_t ctxid = entry->src_data[0];
+	uint32_t data;
 
 	if (client_id == SOC15_IH_CLIENTID_THM) {
 		switch (src_id) {
@@ -1590,6 +1591,11 @@ static int smu_v11_0_irq_process(struct amdgpu_device *adev,
 		orderly_poweroff(true);
 	} else if (client_id == SOC15_IH_CLIENTID_MP1) {
 		if (src_id == 0xfe) {
+			/* ACK SMUToHost interrupt */
+			data = RREG32_SOC15(MP1, 0, mmMP1_SMN_IH_SW_INT_CTRL);
+			data = REG_SET_FIELD(data, MP1_SMN_IH_SW_INT_CTRL, INT_ACK, 1);
+			WREG32_SOC15(MP1, 0, mmMP1_SMN_IH_SW_INT_CTRL, data);
+
 			switch (ctxid) {
 			case 0x3:
 				dev_dbg(adev->dev, "Switched to AC mode!\n");
diff --git a/drivers/gpu/drm/i915/display/intel_global_state.c b/drivers/gpu/drm/i915/display/intel_global_state.c
index 212d4ee682059e..7a19215ad8441c 100644
--- a/drivers/gpu/drm/i915/display/intel_global_state.c
+++ b/drivers/gpu/drm/i915/display/intel_global_state.c
@@ -10,6 +10,28 @@
 #include "intel_display_types.h"
 #include "intel_global_state.h"
 
+static void __intel_atomic_global_state_free(struct kref *kref)
+{
+	struct intel_global_state *obj_state =
+		container_of(kref, struct intel_global_state, ref);
+	struct intel_global_obj *obj = obj_state->obj;
+
+	obj->funcs->atomic_destroy_state(obj, obj_state);
+}
+
+static void intel_atomic_global_state_put(struct intel_global_state *obj_state)
+{
+	kref_put(&obj_state->ref, __intel_atomic_global_state_free);
+}
+
+static struct intel_global_state *
+intel_atomic_global_state_get(struct intel_global_state *obj_state)
+{
+	kref_get(&obj_state->ref);
+
+	return obj_state;
+}
+
 void intel_atomic_global_obj_init(struct drm_i915_private *dev_priv,
 				  struct intel_global_obj *obj,
 				  struct intel_global_state *state,
@@ -17,6 +39,10 @@ void intel_atomic_global_obj_init(struct drm_i915_private *dev_priv,
 {
 	memset(obj, 0, sizeof(*obj));
 
+	state->obj = obj;
+
+	kref_init(&state->ref);
+
 	obj->state = state;
 	obj->funcs = funcs;
 	list_add_tail(&obj->head, &dev_priv->global_obj_list);
@@ -28,7 +54,9 @@ void intel_atomic_global_obj_cleanup(struct drm_i915_private *dev_priv)
 
 	list_for_each_entry_safe(obj, next, &dev_priv->global_obj_list, head) {
 		list_del(&obj->head);
-		obj->funcs->atomic_destroy_state(obj, obj->state);
+
+		drm_WARN_ON(&dev_priv->drm, kref_read(&obj->state->ref) != 1);
+		intel_atomic_global_state_put(obj->state);
 	}
 }
 
@@ -97,10 +125,14 @@ intel_atomic_get_global_obj_state(struct intel_atomic_state *state,
 	if (!obj_state)
 		return ERR_PTR(-ENOMEM);
 
+	obj_state->obj = obj;
 	obj_state->changed = false;
 
+	kref_init(&obj_state->ref);
+
 	state->global_objs[index].state = obj_state;
-	state->global_objs[index].old_state = obj->state;
+	state->global_objs[index].old_state =
+		intel_atomic_global_state_get(obj->state);
 	state->global_objs[index].new_state = obj_state;
 	state->global_objs[index].ptr = obj;
 	obj_state->state = state;
@@ -163,7 +195,9 @@ void intel_atomic_swap_global_state(struct intel_atomic_state *state)
 		new_obj_state->state = NULL;
 
 		state->global_objs[i].state = old_obj_state;
-		obj->state = new_obj_state;
+
+		intel_atomic_global_state_put(obj->state);
+		obj->state = intel_atomic_global_state_get(new_obj_state);
 	}
 }
 
@@ -172,10 +206,9 @@ void intel_atomic_clear_global_state(struct intel_atomic_state *state)
 	int i;
 
 	for (i = 0; i < state->num_global_objs; i++) {
-		struct intel_global_obj *obj = state->global_objs[i].ptr;
+		intel_atomic_global_state_put(state->global_objs[i].old_state);
+		intel_atomic_global_state_put(state->global_objs[i].new_state);
 
-		obj->funcs->atomic_destroy_state(obj,
-						 state->global_objs[i].state);
 		state->global_objs[i].ptr = NULL;
 		state->global_objs[i].state = NULL;
 		state->global_objs[i].old_state = NULL;
diff --git a/drivers/gpu/drm/i915/display/intel_global_state.h b/drivers/gpu/drm/i915/display/intel_global_state.h
index e6163a46902966..1f16fa3073c9fc 100644
--- a/drivers/gpu/drm/i915/display/intel_global_state.h
+++ b/drivers/gpu/drm/i915/display/intel_global_state.h
@@ -6,6 +6,7 @@
 #ifndef __INTEL_GLOBAL_STATE_H__
 #define __INTEL_GLOBAL_STATE_H__
 
+#include <linux/kref.h>
 #include <linux/list.h>
 
 struct drm_i915_private;
@@ -54,7 +55,9 @@ struct intel_global_obj {
 		for_each_if(obj)
 
 struct intel_global_state {
+	struct intel_global_obj *obj;
 	struct intel_atomic_state *state;
+	struct kref ref;
 	bool changed;
 };
 
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_context.c b/drivers/gpu/drm/i915/gem/i915_gem_context.c
index 900ea8b7fc8fb5..f5d59d18cd5ba3 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_context.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_context.c
@@ -230,7 +230,7 @@ static void intel_context_set_gem(struct intel_context *ce,
 		ce->timeline = intel_timeline_get(ctx->timeline);
 
 	if (ctx->sched.priority >= I915_PRIORITY_NORMAL &&
-	    intel_engine_has_semaphores(ce->engine))
+	    intel_engine_has_timeslices(ce->engine))
 		__set_bit(CONTEXT_USE_SEMAPHORES, &ce->flags);
 }
 
@@ -1969,7 +1969,7 @@ static int __apply_priority(struct intel_context *ce, void *arg)
 {
 	struct i915_gem_context *ctx = arg;
 
-	if (!intel_engine_has_semaphores(ce->engine))
+	if (!intel_engine_has_timeslices(ce->engine))
 		return 0;
 
 	if (ctx->sched.priority >= I915_PRIORITY_NORMAL)
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_shmem.c b/drivers/gpu/drm/i915/gem/i915_gem_shmem.c
index 5d5d7eef3f43af..7aff3514d97a98 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_shmem.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_shmem.c
@@ -39,7 +39,6 @@ static int shmem_get_pages(struct drm_i915_gem_object *obj)
 	unsigned long last_pfn = 0;	/* suppress gcc warning */
 	unsigned int max_segment = i915_sg_segment_size();
 	unsigned int sg_page_sizes;
-	struct pagevec pvec;
 	gfp_t noreclaim;
 	int ret;
 
@@ -192,13 +191,17 @@ err_sg:
 	sg_mark_end(sg);
 err_pages:
 	mapping_clear_unevictable(mapping);
-	pagevec_init(&pvec);
-	for_each_sgt_page(page, sgt_iter, st) {
-		if (!pagevec_add(&pvec, page))
+	if (sg != st->sgl) {
+		struct pagevec pvec;
+
+		pagevec_init(&pvec);
+		for_each_sgt_page(page, sgt_iter, st) {
+			if (!pagevec_add(&pvec, page))
+				check_release_pagevec(&pvec);
+		}
+		if (pagevec_count(&pvec))
 			check_release_pagevec(&pvec);
 	}
-	if (pagevec_count(&pvec))
-		check_release_pagevec(&pvec);
 	sg_free_table(st);
 	kfree(st);
 
diff --git a/drivers/gpu/drm/i915/gt/intel_context.c b/drivers/gpu/drm/i915/gt/intel_context.c
index 74ddb49b294155..e4aece20bc8089 100644
--- a/drivers/gpu/drm/i915/gt/intel_context.c
+++ b/drivers/gpu/drm/i915/gt/intel_context.c
@@ -97,8 +97,6 @@ int __intel_context_do_pin(struct intel_context *ce)
 {
 	int err;
 
-	GEM_BUG_ON(intel_context_is_closed(ce));
-
 	if (unlikely(!test_bit(CONTEXT_ALLOC_BIT, &ce->flags))) {
 		err = intel_context_alloc_state(ce);
 		if (err)
diff --git a/drivers/gpu/drm/i915/gvt/vgpu.c b/drivers/gpu/drm/i915/gvt/vgpu.c
index 1d5ff88078bd12..7d361623ff6792 100644
--- a/drivers/gpu/drm/i915/gvt/vgpu.c
+++ b/drivers/gpu/drm/i915/gvt/vgpu.c
@@ -124,7 +124,7 @@ int intel_gvt_init_vgpu_types(struct intel_gvt *gvt)
 	 */
 	low_avail = gvt_aperture_sz(gvt) - HOST_LOW_GM_SIZE;
 	high_avail = gvt_hidden_sz(gvt) - HOST_HIGH_GM_SIZE;
-	num_types = sizeof(vgpu_types) / sizeof(vgpu_types[0]);
+	num_types = ARRAY_SIZE(vgpu_types);
 
 	gvt->types = kcalloc(num_types, sizeof(struct intel_vgpu_type),
 			     GFP_KERNEL);
diff --git a/drivers/gpu/drm/i915/i915_cmd_parser.c b/drivers/gpu/drm/i915/i915_cmd_parser.c
index 189b573d02beb9..372354d33f5521 100644
--- a/drivers/gpu/drm/i915/i915_cmd_parser.c
+++ b/drivers/gpu/drm/i915/i915_cmd_parser.c
@@ -572,6 +572,9 @@ struct drm_i915_reg_descriptor {
 #define REG32(_reg, ...) \
 	{ .addr = (_reg), __VA_ARGS__ }
 
+#define REG32_IDX(_reg, idx) \
+	{ .addr = _reg(idx) }
+
 /*
  * Convenience macro for adding 64-bit registers.
  *
@@ -669,6 +672,7 @@ static const struct drm_i915_reg_descriptor gen9_blt_regs[] = {
 	REG64_IDX(RING_TIMESTAMP, BSD_RING_BASE),
 	REG32(BCS_SWCTRL),
 	REG64_IDX(RING_TIMESTAMP, BLT_RING_BASE),
+	REG32_IDX(RING_CTX_TIMESTAMP, BLT_RING_BASE),
 	REG64_IDX(BCS_GPR, 0),
 	REG64_IDX(BCS_GPR, 1),
 	REG64_IDX(BCS_GPR, 2),
diff --git a/drivers/gpu/drm/i915/i915_params.c b/drivers/gpu/drm/i915/i915_params.c
index add00ec1f7878b..a3dde770226dec 100644
--- a/drivers/gpu/drm/i915/i915_params.c
+++ b/drivers/gpu/drm/i915/i915_params.c
@@ -173,7 +173,7 @@ i915_param_named(enable_gvt, bool, 0400,
 #endif
 
 #if IS_ENABLED(CONFIG_DRM_I915_UNSTABLE_FAKE_LMEM)
-i915_param_named_unsafe(fake_lmem_start, ulong, 0600,
+i915_param_named_unsafe(fake_lmem_start, ulong, 0400,
 	"Fake LMEM start offset (default: 0)");
 #endif
 
diff --git a/drivers/gpu/drm/i915/i915_params.h b/drivers/gpu/drm/i915/i915_params.h
index 45323732f0997e..4f21bfffbf0ed0 100644
--- a/drivers/gpu/drm/i915/i915_params.h
+++ b/drivers/gpu/drm/i915/i915_params.h
@@ -64,7 +64,7 @@ struct drm_printer;
 	param(int, mmio_debug, -IS_ENABLED(CONFIG_DRM_I915_DEBUG_MMIO), 0600) \
 	param(int, edp_vswing, 0, 0400) \
 	param(unsigned int, reset, 3, 0600) \
-	param(unsigned int, inject_probe_failure, 0, 0600) \
+	param(unsigned int, inject_probe_failure, 0, 0) \
 	param(int, fastboot, -1, 0600) \
 	param(int, enable_dpcd_backlight, -1, 0600) \
 	param(char *, force_probe, CONFIG_DRM_I915_FORCE_PROBE, 0400) \
diff --git a/drivers/gpu/drm/i915/i915_request.c b/drivers/gpu/drm/i915/i915_request.c
index 526c1e9acbd587..def62100e666d0 100644
--- a/drivers/gpu/drm/i915/i915_request.c
+++ b/drivers/gpu/drm/i915/i915_request.c
@@ -121,8 +121,39 @@ static void i915_fence_release(struct dma_fence *fence)
 	i915_sw_fence_fini(&rq->submit);
 	i915_sw_fence_fini(&rq->semaphore);
 
-	/* Keep one request on each engine for reserved use under mempressure */
-	if (!cmpxchg(&rq->engine->request_pool, NULL, rq))
+	/*
+	 * Keep one request on each engine for reserved use under mempressure
+	 *
+	 * We do not hold a reference to the engine here and so have to be
+	 * very careful in what rq->engine we poke. The virtual engine is
+	 * referenced via the rq->context and we released that ref during
+	 * i915_request_retire(), ergo we must not dereference a virtual
+	 * engine here. Not that we would want to, as the only consumer of
+	 * the reserved engine->request_pool is the power management parking,
+	 * which must-not-fail, and that is only run on the physical engines.
+	 *
+	 * Since the request must have been executed to be have completed,
+	 * we know that it will have been processed by the HW and will
+	 * not be unsubmitted again, so rq->engine and rq->execution_mask
+	 * at this point is stable. rq->execution_mask will be a single
+	 * bit if the last and _only_ engine it could execution on was a
+	 * physical engine, if it's multiple bits then it started on and
+	 * could still be on a virtual engine. Thus if the mask is not a
+	 * power-of-two we assume that rq->engine may still be a virtual
+	 * engine and so a dangling invalid pointer that we cannot dereference
+	 *
+	 * For example, consider the flow of a bonded request through a virtual
+	 * engine. The request is created with a wide engine mask (all engines
+	 * that we might execute on). On processing the bond, the request mask
+	 * is reduced to one or more engines. If the request is subsequently
+	 * bound to a single engine, it will then be constrained to only
+	 * execute on that engine and never returned to the virtual engine
+	 * after timeslicing away, see __unwind_incomplete_requests(). Thus we
+	 * know that if the rq->execution_mask is a single bit, rq->engine
+	 * can be a physical engine with the exact corresponding mask.
+	 */
+	if (is_power_of_2(rq->execution_mask) &&
+	    !cmpxchg(&rq->engine->request_pool, NULL, rq))
 		return;
 
 	kmem_cache_free(global.slab_requests, rq);
@@ -326,6 +357,53 @@ void i915_request_retire_upto(struct i915_request *rq)
 	} while (i915_request_retire(tmp) && tmp != rq);
 }
 
+static struct i915_request * const *
+__engine_active(struct intel_engine_cs *engine)
+{
+	return READ_ONCE(engine->execlists.active);
+}
+
+static bool __request_in_flight(const struct i915_request *signal)
+{
+	struct i915_request * const *port, *rq;
+	bool inflight = false;
+
+	if (!i915_request_is_ready(signal))
+		return false;
+
+	/*
+	 * Even if we have unwound the request, it may still be on
+	 * the GPU (preempt-to-busy). If that request is inside an
+	 * unpreemptible critical section, it will not be removed. Some
+	 * GPU functions may even be stuck waiting for the paired request
+	 * (__await_execution) to be submitted and cannot be preempted
+	 * until the bond is executing.
+	 *
+	 * As we know that there are always preemption points between
+	 * requests, we know that only the currently executing request
+	 * may be still active even though we have cleared the flag.
+	 * However, we can't rely on our tracking of ELSP[0] to known
+	 * which request is currently active and so maybe stuck, as
+	 * the tracking maybe an event behind. Instead assume that
+	 * if the context is still inflight, then it is still active
+	 * even if the active flag has been cleared.
+	 */
+	if (!intel_context_inflight(signal->context))
+		return false;
+
+	rcu_read_lock();
+	for (port = __engine_active(signal->engine); (rq = *port); port++) {
+		if (rq->context == signal->context) {
+			inflight = i915_seqno_passed(rq->fence.seqno,
+						     signal->fence.seqno);
+			break;
+		}
+	}
+	rcu_read_unlock();
+
+	return inflight;
+}
+
 static int
 __await_execution(struct i915_request *rq,
 		  struct i915_request *signal,
@@ -356,7 +434,7 @@ __await_execution(struct i915_request *rq,
 	}
 
 	spin_lock_irq(&signal->lock);
-	if (i915_request_is_active(signal)) {
+	if (i915_request_is_active(signal) || __request_in_flight(signal)) {
 		if (hook) {
 			hook(rq, &signal->fence);
 			i915_request_put(signal);
@@ -1022,37 +1100,91 @@ await_fence:
 					     I915_FENCE_GFP);
 }
 
+static bool intel_timeline_sync_has_start(struct intel_timeline *tl,
+					  struct dma_fence *fence)
+{
+	return __intel_timeline_sync_is_later(tl,
+					      fence->context,
+					      fence->seqno - 1);
+}
+
+static int intel_timeline_sync_set_start(struct intel_timeline *tl,
+					 const struct dma_fence *fence)
+{
+	return __intel_timeline_sync_set(tl, fence->context, fence->seqno - 1);
+}
+
 static int
-i915_request_await_request(struct i915_request *to, struct i915_request *from)
+__i915_request_await_execution(struct i915_request *to,
+			       struct i915_request *from,
+			       void (*hook)(struct i915_request *rq,
+					    struct dma_fence *signal))
 {
-	int ret;
+	int err;
 
-	GEM_BUG_ON(to == from);
-	GEM_BUG_ON(to->timeline == from->timeline);
+	GEM_BUG_ON(intel_context_is_barrier(from->context));
 
-	if (i915_request_completed(from)) {
-		i915_sw_fence_set_error_once(&to->submit, from->fence.error);
+	/* Submit both requests at the same time */
+	err = __await_execution(to, from, hook, I915_FENCE_GFP);
+	if (err)
+		return err;
+
+	/* Squash repeated depenendices to the same timelines */
+	if (intel_timeline_sync_has_start(i915_request_timeline(to),
+					  &from->fence))
 		return 0;
+
+	/*
+	 * Wait until the start of this request.
+	 *
+	 * The execution cb fires when we submit the request to HW. But in
+	 * many cases this may be long before the request itself is ready to
+	 * run (consider that we submit 2 requests for the same context, where
+	 * the request of interest is behind an indefinite spinner). So we hook
+	 * up to both to reduce our queues and keep the execution lag minimised
+	 * in the worst case, though we hope that the await_start is elided.
+	 */
+	err = i915_request_await_start(to, from);
+	if (err < 0)
+		return err;
+
+	/*
+	 * Ensure both start together [after all semaphores in signal]
+	 *
+	 * Now that we are queued to the HW at roughly the same time (thanks
+	 * to the execute cb) and are ready to run at roughly the same time
+	 * (thanks to the await start), our signaler may still be indefinitely
+	 * delayed by waiting on a semaphore from a remote engine. If our
+	 * signaler depends on a semaphore, so indirectly do we, and we do not
+	 * want to start our payload until our signaler also starts theirs.
+	 * So we wait.
+	 *
+	 * However, there is also a second condition for which we need to wait
+	 * for the precise start of the signaler. Consider that the signaler
+	 * was submitted in a chain of requests following another context
+	 * (with just an ordinary intra-engine fence dependency between the
+	 * two). In this case the signaler is queued to HW, but not for
+	 * immediate execution, and so we must wait until it reaches the
+	 * active slot.
+	 */
+	if (intel_engine_has_semaphores(to->engine) &&
+	    !i915_request_has_initial_breadcrumb(to)) {
+		err = __emit_semaphore_wait(to, from, from->fence.seqno - 1);
+		if (err < 0)
+			return err;
 	}
 
+	/* Couple the dependency tree for PI on this exposed to->fence */
 	if (to->engine->schedule) {
-		ret = i915_sched_node_add_dependency(&to->sched,
+		err = i915_sched_node_add_dependency(&to->sched,
 						     &from->sched,
-						     I915_DEPENDENCY_EXTERNAL);
-		if (ret < 0)
-			return ret;
+						     I915_DEPENDENCY_WEAK);
+		if (err < 0)
+			return err;
 	}
 
-	if (to->engine == from->engine)
-		ret = i915_sw_fence_await_sw_fence_gfp(&to->submit,
-						       &from->submit,
-						       I915_FENCE_GFP);
-	else
-		ret = emit_semaphore_wait(to, from, I915_FENCE_GFP);
-	if (ret < 0)
-		return ret;
-
-	return 0;
+	return intel_timeline_sync_set_start(i915_request_timeline(to),
+					     &from->fence);
 }
 
 static void mark_external(struct i915_request *rq)
@@ -1105,23 +1237,20 @@ i915_request_await_external(struct i915_request *rq, struct dma_fence *fence)
 }
 
 int
-i915_request_await_dma_fence(struct i915_request *rq, struct dma_fence *fence)
+i915_request_await_execution(struct i915_request *rq,
+			     struct dma_fence *fence,
+			     void (*hook)(struct i915_request *rq,
+					  struct dma_fence *signal))
 {
 	struct dma_fence **child = &fence;
 	unsigned int nchild = 1;
 	int ret;
 
-	/*
-	 * Note that if the fence-array was created in signal-on-any mode,
-	 * we should *not* decompose it into its individual fences. However,
-	 * we don't currently store which mode the fence-array is operating
-	 * in. Fortunately, the only user of signal-on-any is private to
-	 * amdgpu and we should not see any incoming fence-array from
-	 * sync-file being in signal-on-any mode.
-	 */
 	if (dma_fence_is_array(fence)) {
 		struct dma_fence_array *array = to_dma_fence_array(fence);
 
+		/* XXX Error for signal-on-any fence arrays */
+
 		child = array->fences;
 		nchild = array->num_fences;
 		GEM_BUG_ON(!nchild);
@@ -1134,138 +1263,95 @@ i915_request_await_dma_fence(struct i915_request *rq, struct dma_fence *fence)
 			continue;
 		}
 
-		/*
-		 * Requests on the same timeline are explicitly ordered, along
-		 * with their dependencies, by i915_request_add() which ensures
-		 * that requests are submitted in-order through each ring.
-		 */
 		if (fence->context == rq->fence.context)
 			continue;
 
-		/* Squash repeated waits to the same timelines */
-		if (fence->context &&
-		    intel_timeline_sync_is_later(i915_request_timeline(rq),
-						 fence))
-			continue;
+		/*
+		 * We don't squash repeated fence dependencies here as we
+		 * want to run our callback in all cases.
+		 */
 
 		if (dma_fence_is_i915(fence))
-			ret = i915_request_await_request(rq, to_request(fence));
+			ret = __i915_request_await_execution(rq,
+							     to_request(fence),
+							     hook);
 		else
 			ret = i915_request_await_external(rq, fence);
 		if (ret < 0)
 			return ret;
-
-		/* Record the latest fence used against each timeline */
-		if (fence->context)
-			intel_timeline_sync_set(i915_request_timeline(rq),
-						fence);
 	} while (--nchild);
 
 	return 0;
 }
 
-static bool intel_timeline_sync_has_start(struct intel_timeline *tl,
-					  struct dma_fence *fence)
-{
-	return __intel_timeline_sync_is_later(tl,
-					      fence->context,
-					      fence->seqno - 1);
-}
-
-static int intel_timeline_sync_set_start(struct intel_timeline *tl,
-					 const struct dma_fence *fence)
+static int
+await_request_submit(struct i915_request *to, struct i915_request *from)
 {
-	return __intel_timeline_sync_set(tl, fence->context, fence->seqno - 1);
+	/*
+	 * If we are waiting on a virtual engine, then it may be
+	 * constrained to execute on a single engine *prior* to submission.
+	 * When it is submitted, it will be first submitted to the virtual
+	 * engine and then passed to the physical engine. We cannot allow
+	 * the waiter to be submitted immediately to the physical engine
+	 * as it may then bypass the virtual request.
+	 */
+	if (to->engine == READ_ONCE(from->engine))
+		return i915_sw_fence_await_sw_fence_gfp(&to->submit,
+							&from->submit,
+							I915_FENCE_GFP);
+	else
+		return __i915_request_await_execution(to, from, NULL);
 }
 
 static int
-__i915_request_await_execution(struct i915_request *to,
-			       struct i915_request *from,
-			       void (*hook)(struct i915_request *rq,
-					    struct dma_fence *signal))
+i915_request_await_request(struct i915_request *to, struct i915_request *from)
 {
-	int err;
-
-	GEM_BUG_ON(intel_context_is_barrier(from->context));
+	int ret;
 
-	/* Submit both requests at the same time */
-	err = __await_execution(to, from, hook, I915_FENCE_GFP);
-	if (err)
-		return err;
+	GEM_BUG_ON(to == from);
+	GEM_BUG_ON(to->timeline == from->timeline);
 
-	/* Squash repeated depenendices to the same timelines */
-	if (intel_timeline_sync_has_start(i915_request_timeline(to),
-					  &from->fence))
+	if (i915_request_completed(from)) {
+		i915_sw_fence_set_error_once(&to->submit, from->fence.error);
 		return 0;
-
-	/*
-	 * Wait until the start of this request.
-	 *
-	 * The execution cb fires when we submit the request to HW. But in
-	 * many cases this may be long before the request itself is ready to
-	 * run (consider that we submit 2 requests for the same context, where
-	 * the request of interest is behind an indefinite spinner). So we hook
-	 * up to both to reduce our queues and keep the execution lag minimised
-	 * in the worst case, though we hope that the await_start is elided.
-	 */
-	err = i915_request_await_start(to, from);
-	if (err < 0)
-		return err;
-
-	/*
-	 * Ensure both start together [after all semaphores in signal]
-	 *
-	 * Now that we are queued to the HW at roughly the same time (thanks
-	 * to the execute cb) and are ready to run at roughly the same time
-	 * (thanks to the await start), our signaler may still be indefinitely
-	 * delayed by waiting on a semaphore from a remote engine. If our
-	 * signaler depends on a semaphore, so indirectly do we, and we do not
-	 * want to start our payload until our signaler also starts theirs.
-	 * So we wait.
-	 *
-	 * However, there is also a second condition for which we need to wait
-	 * for the precise start of the signaler. Consider that the signaler
-	 * was submitted in a chain of requests following another context
-	 * (with just an ordinary intra-engine fence dependency between the
-	 * two). In this case the signaler is queued to HW, but not for
-	 * immediate execution, and so we must wait until it reaches the
-	 * active slot.
-	 */
-	if (intel_engine_has_semaphores(to->engine) &&
-	    !i915_request_has_initial_breadcrumb(to)) {
-		err = __emit_semaphore_wait(to, from, from->fence.seqno - 1);
-		if (err < 0)
-			return err;
 	}
 
-	/* Couple the dependency tree for PI on this exposed to->fence */
 	if (to->engine->schedule) {
-		err = i915_sched_node_add_dependency(&to->sched,
+		ret = i915_sched_node_add_dependency(&to->sched,
 						     &from->sched,
-						     I915_DEPENDENCY_WEAK);
-		if (err < 0)
-			return err;
+						     I915_DEPENDENCY_EXTERNAL);
+		if (ret < 0)
+			return ret;
 	}
 
-	return intel_timeline_sync_set_start(i915_request_timeline(to),
-					     &from->fence);
+	if (is_power_of_2(to->execution_mask | READ_ONCE(from->execution_mask)))
+		ret = await_request_submit(to, from);
+	else
+		ret = emit_semaphore_wait(to, from, I915_FENCE_GFP);
+	if (ret < 0)
+		return ret;
+
+	return 0;
 }
 
 int
-i915_request_await_execution(struct i915_request *rq,
-			     struct dma_fence *fence,
-			     void (*hook)(struct i915_request *rq,
-					  struct dma_fence *signal))
+i915_request_await_dma_fence(struct i915_request *rq, struct dma_fence *fence)
 {
 	struct dma_fence **child = &fence;
 	unsigned int nchild = 1;
 	int ret;
 
+	/*
+	 * Note that if the fence-array was created in signal-on-any mode,
+	 * we should *not* decompose it into its individual fences. However,
+	 * we don't currently store which mode the fence-array is operating
+	 * in. Fortunately, the only user of signal-on-any is private to
+	 * amdgpu and we should not see any incoming fence-array from
+	 * sync-file being in signal-on-any mode.
+	 */
 	if (dma_fence_is_array(fence)) {
 		struct dma_fence_array *array = to_dma_fence_array(fence);
 
-		/* XXX Error for signal-on-any fence arrays */
-
 		child = array->fences;
 		nchild = array->num_fences;
 		GEM_BUG_ON(!nchild);
@@ -1278,22 +1364,31 @@ i915_request_await_execution(struct i915_request *rq,
 			continue;
 		}
 
+		/*
+		 * Requests on the same timeline are explicitly ordered, along
+		 * with their dependencies, by i915_request_add() which ensures
+		 * that requests are submitted in-order through each ring.
+		 */
 		if (fence->context == rq->fence.context)
 			continue;
 
-		/*
-		 * We don't squash repeated fence dependencies here as we
-		 * want to run our callback in all cases.
-		 */
+		/* Squash repeated waits to the same timelines */
+		if (fence->context &&
+		    intel_timeline_sync_is_later(i915_request_timeline(rq),
+						 fence))
+			continue;
 
 		if (dma_fence_is_i915(fence))
-			ret = __i915_request_await_execution(rq,
-							     to_request(fence),
-							     hook);
+			ret = i915_request_await_request(rq, to_request(fence));
 		else
 			ret = i915_request_await_external(rq, fence);
 		if (ret < 0)
 			return ret;
+
+		/* Record the latest fence used against each timeline */
+		if (fence->context)
+			intel_timeline_sync_set(i915_request_timeline(rq),
+						fence);
 	} while (--nchild);
 
 	return 0;
diff --git a/drivers/gpu/drm/i915/i915_scheduler.c b/drivers/gpu/drm/i915/i915_scheduler.c
index f4ea318781f0f5..cbb880b10c650d 100644
--- a/drivers/gpu/drm/i915/i915_scheduler.c
+++ b/drivers/gpu/drm/i915/i915_scheduler.c
@@ -209,14 +209,6 @@ static void kick_submission(struct intel_engine_cs *engine,
 	if (!inflight)
 		goto unlock;
 
-	ENGINE_TRACE(engine,
-		     "bumping queue-priority-hint:%d for rq:%llx:%lld, inflight:%llx:%lld prio %d\n",
-		     prio,
-		     rq->fence.context, rq->fence.seqno,
-		     inflight->fence.context, inflight->fence.seqno,
-		     inflight->sched.attr.priority);
-	engine->execlists.queue_priority_hint = prio;
-
 	/*
 	 * If we are already the currently executing context, don't
 	 * bother evaluating if we should preempt ourselves.
@@ -224,6 +216,14 @@ static void kick_submission(struct intel_engine_cs *engine,
 	if (inflight->context == rq->context)
 		goto unlock;
 
+	ENGINE_TRACE(engine,
+		     "bumping queue-priority-hint:%d for rq:%llx:%lld, inflight:%llx:%lld prio %d\n",
+		     prio,
+		     rq->fence.context, rq->fence.seqno,
+		     inflight->fence.context, inflight->fence.seqno,
+		     inflight->sched.attr.priority);
+
+	engine->execlists.queue_priority_hint = prio;
 	if (need_preempt(prio, rq_prio(inflight)))
 		tasklet_hi_schedule(&engine->execlists.tasklet);
 
diff --git a/drivers/gpu/drm/nouveau/dispnv50/disp.c b/drivers/gpu/drm/nouveau/dispnv50/disp.c
index 7622490d86024c..d472942102f50d 100644
--- a/drivers/gpu/drm/nouveau/dispnv50/disp.c
+++ b/drivers/gpu/drm/nouveau/dispnv50/disp.c
@@ -277,7 +277,7 @@ nv50_outp_release(struct nouveau_encoder *nv_encoder)
 }
 
 static int
-nv50_outp_acquire(struct nouveau_encoder *nv_encoder)
+nv50_outp_acquire(struct nouveau_encoder *nv_encoder, bool hda)
 {
 	struct nouveau_drm *drm = nouveau_drm(nv_encoder->base.base.dev);
 	struct nv50_disp *disp = nv50_disp(drm->dev);
@@ -289,6 +289,7 @@ nv50_outp_acquire(struct nouveau_encoder *nv_encoder)
 		.base.method = NV50_DISP_MTHD_V1_ACQUIRE,
 		.base.hasht  = nv_encoder->dcb->hasht,
 		.base.hashm  = nv_encoder->dcb->hashm,
+		.info.hda = hda,
 	};
 	int ret;
 
@@ -393,7 +394,7 @@ nv50_dac_enable(struct drm_encoder *encoder)
 	struct nv50_head_atom *asyh = nv50_head_atom(nv_crtc->base.state);
 	struct nv50_core *core = nv50_disp(encoder->dev)->core;
 
-	nv50_outp_acquire(nv_encoder);
+	nv50_outp_acquire(nv_encoder, false);
 
 	core->func->dac->ctrl(core, nv_encoder->or, 1 << nv_crtc->index, asyh);
 	asyh->or.depth = 0;
@@ -510,7 +511,7 @@ nv50_audio_component_get_eld(struct device *kdev, int port, int dev_id,
 		if (!nv_connector || !nv_crtc || nv_encoder->or != port ||
 		    nv_crtc->index != dev_id)
 			continue;
-		*enabled = drm_detect_monitor_audio(nv_connector->edid);
+		*enabled = nv_encoder->audio;
 		if (*enabled) {
 			ret = drm_eld_size(nv_connector->base.eld);
 			memcpy(buf, nv_connector->base.eld,
@@ -600,6 +601,7 @@ nv50_audio_disable(struct drm_encoder *encoder, struct nouveau_crtc *nv_crtc)
 				(0x0100 << nv_crtc->index),
 	};
 
+	nv_encoder->audio = false;
 	nvif_mthd(&disp->disp->object, 0, &args, sizeof(args));
 
 	nv50_audio_component_eld_notify(drm->audio.component, nv_encoder->or,
@@ -636,6 +638,7 @@ nv50_audio_enable(struct drm_encoder *encoder, struct drm_display_mode *mode)
 
 	nvif_mthd(&disp->disp->object, 0, &args,
 		  sizeof(args.base) + drm_eld_size(args.data));
+	nv_encoder->audio = true;
 
 	nv50_audio_component_eld_notify(drm->audio.component, nv_encoder->or,
 					nv_crtc->index);
@@ -966,7 +969,7 @@ nv50_msto_enable(struct drm_encoder *encoder)
 		DRM_DEBUG_KMS("Failed to allocate VCPI\n");
 
 	if (!mstm->links++)
-		nv50_outp_acquire(mstm->outp);
+		nv50_outp_acquire(mstm->outp, false /*XXX: MST audio.*/);
 
 	if (mstm->outp->link & 1)
 		proto = 0x8;
@@ -1560,12 +1563,18 @@ nv50_sor_enable(struct drm_encoder *encoder)
 	struct nouveau_drm *drm = nouveau_drm(dev);
 	struct nouveau_connector *nv_connector;
 	struct nvbios *bios = &drm->vbios;
+	bool hda = false;
 	u8 proto = 0xf;
 	u8 depth = 0x0;
 
 	nv_connector = nouveau_encoder_connector_get(nv_encoder);
 	nv_encoder->crtc = encoder->crtc;
-	nv50_outp_acquire(nv_encoder);
+
+	if ((disp->disp->object.oclass == GT214_DISP ||
+	     disp->disp->object.oclass >= GF110_DISP) &&
+	    drm_detect_monitor_audio(nv_connector->edid))
+		hda = true;
+	nv50_outp_acquire(nv_encoder, hda);
 
 	switch (nv_encoder->dcb->type) {
 	case DCB_OUTPUT_TMDS:
@@ -1775,7 +1784,7 @@ nv50_pior_enable(struct drm_encoder *encoder)
 	u8 owner = 1 << nv_crtc->index;
 	u8 proto;
 
-	nv50_outp_acquire(nv_encoder);
+	nv50_outp_acquire(nv_encoder, false);
 
 	switch (asyh->or.bpc) {
 	case 10: asyh->or.depth = 0x6; break;
diff --git a/drivers/gpu/drm/nouveau/dispnv50/wndw.c b/drivers/gpu/drm/nouveau/dispnv50/wndw.c
index e25ead56052cd8..99b9b681736da2 100644
--- a/drivers/gpu/drm/nouveau/dispnv50/wndw.c
+++ b/drivers/gpu/drm/nouveau/dispnv50/wndw.c
@@ -192,6 +192,8 @@ nv50_wndw_atomic_check_release(struct nv50_wndw *wndw,
 	wndw->func->release(wndw, asyw, asyh);
 	asyw->ntfy.handle = 0;
 	asyw->sema.handle = 0;
+	asyw->xlut.handle = 0;
+	memset(asyw->image.handle, 0x00, sizeof(asyw->image.handle));
 }
 
 static int
@@ -519,7 +521,8 @@ nv50_wndw_prepare_fb(struct drm_plane *plane, struct drm_plane_state *state)
 			return PTR_ERR(ctxdma);
 		}
 
-		asyw->image.handle[0] = ctxdma->object.handle;
+		if (asyw->visible)
+			asyw->image.handle[0] = ctxdma->object.handle;
 	}
 
 	asyw->state.fence = dma_resv_get_excl_rcu(nvbo->bo.base.resv);
diff --git a/drivers/gpu/drm/nouveau/include/nvif/cl5070.h b/drivers/gpu/drm/nouveau/include/nvif/cl5070.h
index 38bf4f38e8692a..53800fb4658292 100644
--- a/drivers/gpu/drm/nouveau/include/nvif/cl5070.h
+++ b/drivers/gpu/drm/nouveau/include/nvif/cl5070.h
@@ -46,7 +46,8 @@ struct nv50_disp_acquire_v0 {
 	__u8  version;
 	__u8  or;
 	__u8  link;
-	__u8  pad03[5];
+	__u8  hda;
+	__u8  pad04[4];
 };
 
 struct nv50_disp_dac_load_v0 {
diff --git a/drivers/gpu/drm/nouveau/nouveau_encoder.h b/drivers/gpu/drm/nouveau/nouveau_encoder.h
index de51733b047615..a72c412ac8b148 100644
--- a/drivers/gpu/drm/nouveau/nouveau_encoder.h
+++ b/drivers/gpu/drm/nouveau/nouveau_encoder.h
@@ -52,6 +52,7 @@ struct nouveau_encoder {
 	 * actually programmed on the hw, not the proposed crtc */
 	struct drm_crtc *crtc;
 	u32 ctrl;
+	bool audio;
 
 	struct drm_display_mode mode;
 	int last_dpms;
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/Kbuild b/drivers/gpu/drm/nouveau/nvkm/engine/disp/Kbuild
index 571687ba85b8f1..cf075311cdd27a 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/Kbuild
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/Kbuild
@@ -39,6 +39,7 @@ nvkm-y += nvkm/engine/disp/sorgf119.o
 nvkm-y += nvkm/engine/disp/sorgk104.o
 nvkm-y += nvkm/engine/disp/sorgm107.o
 nvkm-y += nvkm/engine/disp/sorgm200.o
+nvkm-y += nvkm/engine/disp/sorgp100.o
 nvkm-y += nvkm/engine/disp/sorgv100.o
 nvkm-y += nvkm/engine/disp/sortu102.o
 
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/gp100.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/gp100.c
index fd6216684f6d2e..8471de3f3b616c 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/gp100.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/gp100.c
@@ -36,7 +36,7 @@ gp100_disp = {
 	.super = gf119_disp_super,
 	.root = &gp100_disp_root_oclass,
 	.head = { .cnt = gf119_head_cnt, .new = gf119_head_new },
-	.sor = { .cnt = gf119_sor_cnt, .new = gm200_sor_new },
+	.sor = { .cnt = gf119_sor_cnt, .new = gp100_sor_new },
 };
 
 int
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/gp102.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/gp102.c
index 3468ddec1270b6..a3779c5046ead1 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/gp102.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/gp102.c
@@ -63,7 +63,7 @@ gp102_disp = {
 	.super = gf119_disp_super,
 	.root = &gp102_disp_root_oclass,
 	.head = { .cnt = gf119_head_cnt, .new = gf119_head_new },
-	.sor = { .cnt = gf119_sor_cnt, .new = gm200_sor_new },
+	.sor = { .cnt = gf119_sor_cnt, .new = gp100_sor_new },
 };
 
 int
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/hdmigm200.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/hdmigm200.c
index 9b16a08eb4d9f4..bf6d41fb0c9fe5 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/hdmigm200.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/hdmigm200.c
@@ -27,10 +27,10 @@ void
 gm200_hdmi_scdc(struct nvkm_ior *ior, int head, u8 scdc)
 {
 	struct nvkm_device *device = ior->disp->engine.subdev.device;
-	const u32 hoff = head * 0x800;
+	const u32 soff = nv50_ior_base(ior);
 	const u32 ctrl = scdc & 0x3;
 
-	nvkm_mask(device, 0x61c5bc + hoff, 0x00000003, ctrl);
+	nvkm_mask(device, 0x61c5bc + soff, 0x00000003, ctrl);
 
 	ior->tmds.high_speed = !!(scdc & 0x2);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/ior.h b/drivers/gpu/drm/nouveau/nvkm/engine/disp/ior.h
index c1d7a36e4d3c9c..1a200a9ba4e467 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/ior.h
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/ior.h
@@ -201,6 +201,7 @@ int gf119_sor_new(struct nvkm_disp *, int);
 int gk104_sor_new(struct nvkm_disp *, int);
 int gm107_sor_new(struct nvkm_disp *, int);
 int gm200_sor_new(struct nvkm_disp *, int);
+int gp100_sor_new(struct nvkm_disp *, int);
 
 int gv100_sor_cnt(struct nvkm_disp *, unsigned long *);
 int gv100_sor_new(struct nvkm_disp *, int);
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/outp.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/outp.c
index c62030c96fba09..dcf08249374a35 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/outp.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/outp.c
@@ -111,8 +111,44 @@ nvkm_outp_acquire_ior(struct nvkm_outp *outp, u8 user, struct nvkm_ior *ior)
 	return 0;
 }
 
+static inline int
+nvkm_outp_acquire_hda(struct nvkm_outp *outp, enum nvkm_ior_type type,
+		      u8 user, bool hda)
+{
+	struct nvkm_ior *ior;
+
+	/* First preference is to reuse the OR that is currently armed
+	 * on HW, if any, in order to prevent unnecessary switching.
+	 */
+	list_for_each_entry(ior, &outp->disp->ior, head) {
+		if (!ior->identity && !!ior->func->hda.hpd == hda &&
+		    !ior->asy.outp && ior->arm.outp == outp)
+			return nvkm_outp_acquire_ior(outp, user, ior);
+	}
+
+	/* Failing that, a completely unused OR is the next best thing. */
+	list_for_each_entry(ior, &outp->disp->ior, head) {
+		if (!ior->identity && !!ior->func->hda.hpd == hda &&
+		    !ior->asy.outp && ior->type == type && !ior->arm.outp &&
+		    (ior->func->route.set || ior->id == __ffs(outp->info.or)))
+			return nvkm_outp_acquire_ior(outp, user, ior);
+	}
+
+	/* Last resort is to assign an OR that's already active on HW,
+	 * but will be released during the next modeset.
+	 */
+	list_for_each_entry(ior, &outp->disp->ior, head) {
+		if (!ior->identity && !!ior->func->hda.hpd == hda &&
+		    !ior->asy.outp && ior->type == type &&
+		    (ior->func->route.set || ior->id == __ffs(outp->info.or)))
+			return nvkm_outp_acquire_ior(outp, user, ior);
+	}
+
+	return -ENOSPC;
+}
+
 int
-nvkm_outp_acquire(struct nvkm_outp *outp, u8 user)
+nvkm_outp_acquire(struct nvkm_outp *outp, u8 user, bool hda)
 {
 	struct nvkm_ior *ior = outp->ior;
 	enum nvkm_ior_proto proto;
@@ -137,32 +173,25 @@ nvkm_outp_acquire(struct nvkm_outp *outp, u8 user)
 		return nvkm_outp_acquire_ior(outp, user, ior);
 	}
 
-	/* First preference is to reuse the OR that is currently armed
-	 * on HW, if any, in order to prevent unnecessary switching.
+	/* If we don't need HDA, first try to acquire an OR that doesn't
+	 * support it to leave free the ones that do.
 	 */
-	list_for_each_entry(ior, &outp->disp->ior, head) {
-		if (!ior->identity && !ior->asy.outp && ior->arm.outp == outp)
-			return nvkm_outp_acquire_ior(outp, user, ior);
-	}
+	if (!hda) {
+		if (!nvkm_outp_acquire_hda(outp, type, user, false))
+			return 0;
 
-	/* Failing that, a completely unused OR is the next best thing. */
-	list_for_each_entry(ior, &outp->disp->ior, head) {
-		if (!ior->identity &&
-		    !ior->asy.outp && ior->type == type && !ior->arm.outp &&
-		    (ior->func->route.set || ior->id == __ffs(outp->info.or)))
-			return nvkm_outp_acquire_ior(outp, user, ior);
+		/* Use a HDA-supporting SOR anyway. */
+		return nvkm_outp_acquire_hda(outp, type, user, true);
 	}
 
-	/* Last resort is to assign an OR that's already active on HW,
-	 * but will be released during the next modeset.
-	 */
-	list_for_each_entry(ior, &outp->disp->ior, head) {
-		if (!ior->identity && !ior->asy.outp && ior->type == type &&
-		    (ior->func->route.set || ior->id == __ffs(outp->info.or)))
-			return nvkm_outp_acquire_ior(outp, user, ior);
-	}
+	/* We want HDA, try to acquire an OR that supports it. */
+	if (!nvkm_outp_acquire_hda(outp, type, user, true))
+		return 0;
 
-	return -ENOSPC;
+	/* There weren't any free ORs that support HDA, grab one that
+	 * doesn't and at least allow display to work still.
+	 */
+	return nvkm_outp_acquire_hda(outp, type, user, false);
 }
 
 void
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/outp.h b/drivers/gpu/drm/nouveau/nvkm/engine/disp/outp.h
index 721b068b87efa7..ee028d30cfe705 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/outp.h
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/outp.h
@@ -32,7 +32,7 @@ int nvkm_outp_new(struct nvkm_disp *, int index, struct dcb_output *,
 void nvkm_outp_del(struct nvkm_outp **);
 void nvkm_outp_init(struct nvkm_outp *);
 void nvkm_outp_fini(struct nvkm_outp *);
-int nvkm_outp_acquire(struct nvkm_outp *, u8 user);
+int nvkm_outp_acquire(struct nvkm_outp *, u8 user, bool hda);
 void nvkm_outp_release(struct nvkm_outp *, u8 user);
 void nvkm_outp_route(struct nvkm_disp *);
 
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/rootnv50.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/rootnv50.c
index a7672ef17d3bc8..fb5de44e4b8d29 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/rootnv50.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/rootnv50.c
@@ -99,7 +99,7 @@ nv50_disp_root_mthd_(struct nvkm_object *object, u32 mthd, void *data, u32 size)
 		} *args = data;
 		int ret = -ENOSYS;
 		if (!(ret = nvif_unpack(ret, &data, &size, args->v0, 0, 0, false))) {
-			ret = nvkm_outp_acquire(outp, NVKM_OUTP_USER);
+			ret = nvkm_outp_acquire(outp, NVKM_OUTP_USER, args->v0.hda);
 			if (ret == 0) {
 				args->v0.or = outp->ior->id;
 				args->v0.link = outp->ior->asy.link;
@@ -119,7 +119,7 @@ nv50_disp_root_mthd_(struct nvkm_object *object, u32 mthd, void *data, u32 size)
 		if (!(ret = nvif_unpack(ret, &data, &size, args->v0, 0, 0, false))) {
 			if (args->v0.data & 0xfff00000)
 				return -EINVAL;
-			ret = nvkm_outp_acquire(outp, NVKM_OUTP_PRIV);
+			ret = nvkm_outp_acquire(outp, NVKM_OUTP_PRIV, false);
 			if (ret)
 				return ret;
 			ret = outp->ior->func->sense(outp->ior, args->v0.data);
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/sorgm200.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/sorgm200.c
index cf2075db742a21..4dd7f382968ed1 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/sorgm200.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/sorgm200.c
@@ -89,7 +89,7 @@ gm200_sor_route_get(struct nvkm_outp *outp, int *link)
 }
 
 static const struct nvkm_ior_func
-gm200_sor = {
+gm200_sor_hda = {
 	.route = {
 		.get = gm200_sor_route_get,
 		.set = gm200_sor_route_set,
@@ -119,8 +119,42 @@ gm200_sor = {
 	},
 };
 
+static const struct nvkm_ior_func
+gm200_sor = {
+	.route = {
+		.get = gm200_sor_route_get,
+		.set = gm200_sor_route_set,
+	},
+	.state = gf119_sor_state,
+	.power = nv50_sor_power,
+	.clock = gf119_sor_clock,
+	.hdmi = {
+		.ctrl = gk104_hdmi_ctrl,
+		.scdc = gm200_hdmi_scdc,
+	},
+	.dp = {
+		.lanes = { 0, 1, 2, 3 },
+		.links = gf119_sor_dp_links,
+		.power = g94_sor_dp_power,
+		.pattern = gm107_sor_dp_pattern,
+		.drive = gm200_sor_dp_drive,
+		.vcpi = gf119_sor_dp_vcpi,
+		.audio = gf119_sor_dp_audio,
+		.audio_sym = gf119_sor_dp_audio_sym,
+		.watermark = gf119_sor_dp_watermark,
+	},
+};
+
 int
 gm200_sor_new(struct nvkm_disp *disp, int id)
 {
+	struct nvkm_device *device = disp->engine.subdev.device;
+	u32 hda;
+
+	if (!((hda = nvkm_rd32(device, 0x08a15c)) & 0x40000000))
+		hda = nvkm_rd32(device, 0x101034);
+
+	if (hda & BIT(id))
+		return nvkm_ior_new_(&gm200_sor_hda, disp, SOR, id);
 	return nvkm_ior_new_(&gm200_sor, disp, SOR, id);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/sorgp100.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/sorgp100.c
new file mode 100644
index 00000000000000..c54f88317a07f5
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/sorgp100.c
@@ -0,0 +1,93 @@
+/*
+ * Copyright 2020 Red Hat 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 "ior.h"
+
+static const struct nvkm_ior_func
+gp100_sor_hda = {
+	.route = {
+		.get = gm200_sor_route_get,
+		.set = gm200_sor_route_set,
+	},
+	.state = gf119_sor_state,
+	.power = nv50_sor_power,
+	.clock = gf119_sor_clock,
+	.hdmi = {
+		.ctrl = gk104_hdmi_ctrl,
+		.scdc = gm200_hdmi_scdc,
+	},
+	.dp = {
+		.lanes = { 0, 1, 2, 3 },
+		.links = gf119_sor_dp_links,
+		.power = g94_sor_dp_power,
+		.pattern = gm107_sor_dp_pattern,
+		.drive = gm200_sor_dp_drive,
+		.vcpi = gf119_sor_dp_vcpi,
+		.audio = gf119_sor_dp_audio,
+		.audio_sym = gf119_sor_dp_audio_sym,
+		.watermark = gf119_sor_dp_watermark,
+	},
+	.hda = {
+		.hpd = gf119_hda_hpd,
+		.eld = gf119_hda_eld,
+		.device_entry = gf119_hda_device_entry,
+	},
+};
+
+static const struct nvkm_ior_func
+gp100_sor = {
+	.route = {
+		.get = gm200_sor_route_get,
+		.set = gm200_sor_route_set,
+	},
+	.state = gf119_sor_state,
+	.power = nv50_sor_power,
+	.clock = gf119_sor_clock,
+	.hdmi = {
+		.ctrl = gk104_hdmi_ctrl,
+		.scdc = gm200_hdmi_scdc,
+	},
+	.dp = {
+		.lanes = { 0, 1, 2, 3 },
+		.links = gf119_sor_dp_links,
+		.power = g94_sor_dp_power,
+		.pattern = gm107_sor_dp_pattern,
+		.drive = gm200_sor_dp_drive,
+		.vcpi = gf119_sor_dp_vcpi,
+		.audio = gf119_sor_dp_audio,
+		.audio_sym = gf119_sor_dp_audio_sym,
+		.watermark = gf119_sor_dp_watermark,
+	},
+};
+
+int
+gp100_sor_new(struct nvkm_disp *disp, int id)
+{
+	struct nvkm_device *device = disp->engine.subdev.device;
+	u32 hda;
+
+	if (!((hda = nvkm_rd32(device, 0x08a15c)) & 0x40000000))
+		hda = nvkm_rd32(device, 0x10ebb0) >> 8;
+
+	if (hda & BIT(id))
+		return nvkm_ior_new_(&gp100_sor_hda, disp, SOR, id);
+	return nvkm_ior_new_(&gp100_sor, disp, SOR, id);
+}
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/sorgv100.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/sorgv100.c
index d11a0dff10c66a..4441187e8ec905 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/sorgv100.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/sorgv100.c
@@ -78,7 +78,7 @@ gv100_sor_state(struct nvkm_ior *sor, struct nvkm_ior_state *state)
 }
 
 static const struct nvkm_ior_func
-gv100_sor = {
+gv100_sor_hda = {
 	.route = {
 		.get = gm200_sor_route_get,
 		.set = gm200_sor_route_set,
@@ -107,9 +107,42 @@ gv100_sor = {
 	},
 };
 
+static const struct nvkm_ior_func
+gv100_sor = {
+	.route = {
+		.get = gm200_sor_route_get,
+		.set = gm200_sor_route_set,
+	},
+	.state = gv100_sor_state,
+	.power = nv50_sor_power,
+	.clock = gf119_sor_clock,
+	.hdmi = {
+		.ctrl = gv100_hdmi_ctrl,
+		.scdc = gm200_hdmi_scdc,
+	},
+	.dp = {
+		.lanes = { 0, 1, 2, 3 },
+		.links = gf119_sor_dp_links,
+		.power = g94_sor_dp_power,
+		.pattern = gm107_sor_dp_pattern,
+		.drive = gm200_sor_dp_drive,
+		.audio = gv100_sor_dp_audio,
+		.audio_sym = gv100_sor_dp_audio_sym,
+		.watermark = gv100_sor_dp_watermark,
+	},
+};
+
 int
 gv100_sor_new(struct nvkm_disp *disp, int id)
 {
+	struct nvkm_device *device = disp->engine.subdev.device;
+	u32 hda;
+
+	if (!((hda = nvkm_rd32(device, 0x08a15c)) & 0x40000000))
+		hda = nvkm_rd32(device, 0x118fb0) >> 8;
+
+	if (hda & BIT(id))
+		return nvkm_ior_new_(&gv100_sor_hda, disp, SOR, id);
 	return nvkm_ior_new_(&gv100_sor, disp, SOR, id);
 }
 
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/sortu102.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/sortu102.c
index fa6d7425123767..59865a934c4b9d 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/sortu102.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/sortu102.c
@@ -62,7 +62,7 @@ tu102_sor_dp_links(struct nvkm_ior *sor, struct nvkm_i2c_aux *aux)
 }
 
 static const struct nvkm_ior_func
-tu102_sor = {
+tu102_sor_hda = {
 	.route = {
 		.get = gm200_sor_route_get,
 		.set = gm200_sor_route_set,
@@ -92,8 +92,38 @@ tu102_sor = {
 	},
 };
 
+static const struct nvkm_ior_func
+tu102_sor = {
+	.route = {
+		.get = gm200_sor_route_get,
+		.set = gm200_sor_route_set,
+	},
+	.state = gv100_sor_state,
+	.power = nv50_sor_power,
+	.clock = gf119_sor_clock,
+	.hdmi = {
+		.ctrl = gv100_hdmi_ctrl,
+		.scdc = gm200_hdmi_scdc,
+	},
+	.dp = {
+		.lanes = { 0, 1, 2, 3 },
+		.links = tu102_sor_dp_links,
+		.power = g94_sor_dp_power,
+		.pattern = gm107_sor_dp_pattern,
+		.drive = gm200_sor_dp_drive,
+		.vcpi = tu102_sor_dp_vcpi,
+		.audio = gv100_sor_dp_audio,
+		.audio_sym = gv100_sor_dp_audio_sym,
+		.watermark = gv100_sor_dp_watermark,
+	},
+};
+
 int
 tu102_sor_new(struct nvkm_disp *disp, int id)
 {
+	struct nvkm_device *device = disp->engine.subdev.device;
+	u32 hda = nvkm_rd32(device, 0x08a15c);
+	if (hda & BIT(id))
+		return nvkm_ior_new_(&tu102_sor_hda, disp, SOR, id);
 	return nvkm_ior_new_(&tu102_sor, disp, SOR, id);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gk20a.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gk20a.c
index ec330d791d15e8..e56880f3e3bd6a 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gk20a.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gk20a.c
@@ -352,7 +352,7 @@ gk20a_gr_load(struct gf100_gr *gr, int ver, const struct gf100_gr_fwif *fwif)
 
 static const struct gf100_gr_fwif
 gk20a_gr_fwif[] = {
-	{ -1, gk20a_gr_load, &gk20a_gr },
+	{ 0, gk20a_gr_load, &gk20a_gr },
 	{}
 };