Merge tag 'amd-drm-next-6.13-2024-11-06' of https://gitlab.freedesktop.org/agd5f/linux into drm-next

amd-drm-next-6.13-2024-11-06:

amdgpu:
- Misc cleanups
- OLED fixes
- DCN 4.x fixes
- DCN 3.5 fixes
- 8K fixes
- IPS fixes
- DSC fixes
- S3 fix
- KASAN fix
- SMU13 fixes
- fdinfo fixes
- USB-C fixes
- ACPI fix
- Fix dummy page overlapping mappings
- Fix workload profile handling
- Add user control for zero RPM on SMU13
- Cleaner shader updates
- Stop syncing PRT map operations
- Debugfs permissions fixes
- Debugfs bounds check fix
- RAS cleanups
- Enforce isolation updates

amdkfd:
- Add topology cap flag for per queue reset
- Add an interface to query whether KFD queues are present
- Use dynamic allocation for get_cu_occupancy

From: Alex Deucher <alexander.deucher@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20241106163904.189108-1-alexander.deucher@amd.com
Signed-off-by: Dave Airlie <airlied@redhat.com>

Author: airlied

Documentation/gpu/amdgpu/thermal.rst

@@ -100,6 +100,18 @@ fan_minimum_pwm
 .. kernel-doc:: drivers/gpu/drm/amd/pm/amdgpu_pm.c
    :doc: fan_minimum_pwm
 
+fan_zero_rpm_enable
+----------------------
+
+.. kernel-doc:: drivers/gpu/drm/amd/pm/amdgpu_pm.c
+   :doc: fan_zero_rpm_enable
+
+fan_zero_rpm_stop_temperature
+-----------------------------
+
+.. kernel-doc:: drivers/gpu/drm/amd/pm/amdgpu_pm.c
+   :doc: fan_zero_rpm_stop_temperature
+
 GFXOFF
 ======
 

drivers/gpu/drm/amd/amdgpu/amdgpu.h

@@ -118,7 +118,7 @@
 
 #define MAX_GPU_INSTANCE		64
 
-#define GFX_SLICE_PERIOD		msecs_to_jiffies(250)
+#define GFX_SLICE_PERIOD_MS		250
 
 struct amdgpu_gpu_instance {
 	struct amdgpu_device		*adev;
@@ -1111,8 +1111,6 @@ struct amdgpu_device {
 	bool				in_s3;
 	bool				in_s4;
 	bool				in_s0ix;
-	/* indicate amdgpu suspension status */
-	bool				suspend_complete;
 
 	enum pp_mp1_state               mp1_state;
 	struct amdgpu_doorbell_index doorbell_index;

drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c

@@ -172,8 +172,8 @@ static union acpi_object *amdgpu_atif_call(struct amdgpu_atif *atif,
 				      &buffer);
 	obj = (union acpi_object *)buffer.pointer;
 
-	/* Fail if calling the method fails and ATIF is supported */
-	if (ACPI_FAILURE(status) && status != AE_NOT_FOUND) {
+	/* Fail if calling the method fails */
+	if (ACPI_FAILURE(status)) {
 		DRM_DEBUG_DRIVER("failed to evaluate ATIF got %s\n",
 				 acpi_format_exception(status));
 		kfree(obj);

drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c

@@ -890,6 +890,15 @@ int amdgpu_amdkfd_start_sched(struct amdgpu_device *adev, uint32_t node_id)
 	return kgd2kfd_start_sched(adev->kfd.dev, node_id);
 }
 
+/* check if there are KFD queues active */
+bool amdgpu_amdkfd_compute_active(struct amdgpu_device *adev, uint32_t node_id)
+{
+	if (!adev->kfd.init_complete)
+		return false;
+
+	return kgd2kfd_compute_active(adev->kfd.dev, node_id);
+}
+
 /* Config CGTT_SQ_CLK_CTRL */
 int amdgpu_amdkfd_config_sq_perfmon(struct amdgpu_device *adev, uint32_t xcp_id,
 	bool core_override_enable, bool reg_override_enable, bool perfmon_override_enable)

drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h

@@ -268,6 +268,7 @@ int amdgpu_amdkfd_start_sched(struct amdgpu_device *adev, uint32_t node_id);
 int amdgpu_amdkfd_stop_sched(struct amdgpu_device *adev, uint32_t node_id);
 int amdgpu_amdkfd_config_sq_perfmon(struct amdgpu_device *adev, uint32_t xcp_id,
 	bool core_override_enable, bool reg_override_enable, bool perfmon_override_enable);
+bool amdgpu_amdkfd_compute_active(struct amdgpu_device *adev, uint32_t node_id);
 
 
 /* Read user wptr from a specified user address space with page fault
@@ -431,6 +432,7 @@ int kgd2kfd_check_and_lock_kfd(void);
 void kgd2kfd_unlock_kfd(void);
 int kgd2kfd_start_sched(struct kfd_dev *kfd, uint32_t node_id);
 int kgd2kfd_stop_sched(struct kfd_dev *kfd, uint32_t node_id);
+bool kgd2kfd_compute_active(struct kfd_dev *kfd, uint32_t node_id);
 #else
 static inline int kgd2kfd_init(void)
 {
@@ -511,5 +513,10 @@ static inline int kgd2kfd_stop_sched(struct kfd_dev *kfd, uint32_t node_id)
 {
 	return 0;
 }
+
+static inline bool kgd2kfd_compute_active(struct kfd_dev *kfd, uint32_t node_id)
+{
+	return false;
+}
 #endif
 #endif /* AMDGPU_AMDKFD_H_INCLUDED */

drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c

@@ -402,7 +402,7 @@ static ssize_t amdgpu_debugfs_gprwave_read(struct file *f, char __user *buf, siz
 	int r;
 	uint32_t *data, x;
 
-	if (size & 0x3 || *pos & 0x3)
+	if (size > 4096 || size & 0x3 || *pos & 0x3)
 		return -EINVAL;
 
 	r = pm_runtime_get_sync(adev_to_drm(adev)->dev);
@@ -1648,7 +1648,7 @@ int amdgpu_debugfs_regs_init(struct amdgpu_device *adev)
 
 	for (i = 0; i < ARRAY_SIZE(debugfs_regs); i++) {
 		ent = debugfs_create_file(debugfs_regs_names[i],
-					  S_IFREG | 0444, root,
+					  S_IFREG | 0400, root,
 					  adev, debugfs_regs[i]);
 		if (!i && !IS_ERR_OR_NULL(ent))
 			i_size_write(ent->d_inode, adev->rmmio_size);
@@ -2096,17 +2096,20 @@ int amdgpu_debugfs_init(struct amdgpu_device *adev)
 		amdgpu_debugfs_umsch_fwlog_init(adev, &adev->umsch_mm);
 
 	amdgpu_debugfs_jpeg_sched_mask_init(adev);
+	amdgpu_debugfs_gfx_sched_mask_init(adev);
+	amdgpu_debugfs_compute_sched_mask_init(adev);
+	amdgpu_debugfs_sdma_sched_mask_init(adev);
 
 	amdgpu_ras_debugfs_create_all(adev);
 	amdgpu_rap_debugfs_init(adev);
 	amdgpu_securedisplay_debugfs_init(adev);
 	amdgpu_fw_attestation_debugfs_init(adev);
 
-	debugfs_create_file("amdgpu_evict_vram", 0444, root, adev,
+	debugfs_create_file("amdgpu_evict_vram", 0400, root, adev,
 			    &amdgpu_evict_vram_fops);
-	debugfs_create_file("amdgpu_evict_gtt", 0444, root, adev,
+	debugfs_create_file("amdgpu_evict_gtt", 0400, root, adev,
 			    &amdgpu_evict_gtt_fops);
-	debugfs_create_file("amdgpu_test_ib", 0444, root, adev,
+	debugfs_create_file("amdgpu_test_ib", 0400, root, adev,
 			    &amdgpu_debugfs_test_ib_fops);
 	debugfs_create_file("amdgpu_vm_info", 0444, root, adev,
 			    &amdgpu_debugfs_vm_info_fops);

drivers/gpu/drm/amd/amdgpu/amdgpu_device.c

@@ -6452,6 +6452,9 @@ bool amdgpu_device_cache_pci_state(struct pci_dev *pdev)
 	struct amdgpu_device *adev = drm_to_adev(dev);
 	int r;
 
+	if (amdgpu_sriov_vf(adev))
+		return false;
+
 	r = pci_save_state(pdev);
 	if (!r) {
 		kfree(adev->pci_state);

drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c

@@ -1795,11 +1795,13 @@ int amdgpu_discovery_get_nps_info(struct amdgpu_device *adev,
 
 	switch (le16_to_cpu(nps_info->v1.header.version_major)) {
 	case 1:
+		mem_ranges = kvcalloc(nps_info->v1.count,
+				      sizeof(*mem_ranges),
+				      GFP_KERNEL);
+		if (!mem_ranges)
+			return -ENOMEM;
 		*nps_type = nps_info->v1.nps_type;
 		*range_cnt = nps_info->v1.count;
-		mem_ranges = kvzalloc(
-			*range_cnt * sizeof(struct amdgpu_gmc_memrange),
-			GFP_KERNEL);
 		for (i = 0; i < *range_cnt; i++) {
 			mem_ranges[i].base_address =
 				nps_info->v1.instance_info[i].base_address;

drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c

@@ -888,7 +888,7 @@ module_param_named(visualconfirm, amdgpu_dc_visual_confirm, uint, 0444);
  * the ABM algorithm, with 1 being the least reduction and 4 being the most
  * reduction.
  *
- * Defaults to -1, or disabled. Userspace can only override this level after
+ * Defaults to -1, or auto. Userspace can only override this level after
  * boot if it's set to auto.
  */
 int amdgpu_dm_abm_level = -1;
@@ -2505,7 +2505,6 @@ static int amdgpu_pmops_suspend(struct device *dev)
 	struct drm_device *drm_dev = dev_get_drvdata(dev);
 	struct amdgpu_device *adev = drm_to_adev(drm_dev);
 
-	adev->suspend_complete = false;
 	if (amdgpu_acpi_is_s0ix_active(adev))
 		adev->in_s0ix = true;
 	else if (amdgpu_acpi_is_s3_active(adev))
@@ -2520,7 +2519,6 @@ static int amdgpu_pmops_suspend_noirq(struct device *dev)
 	struct drm_device *drm_dev = dev_get_drvdata(dev);
 	struct amdgpu_device *adev = drm_to_adev(drm_dev);
 
-	adev->suspend_complete = true;
 	if (amdgpu_acpi_should_gpu_reset(adev))
 		return amdgpu_asic_reset(adev);
 

drivers/gpu/drm/amd/amdgpu/amdgpu_eeprom.c

@@ -200,7 +200,7 @@ static int amdgpu_eeprom_xfer(struct i2c_adapter *i2c_adap, u32 eeprom_addr,
 		dev_err_ratelimited(&i2c_adap->dev,
 				    "maddr:0x%04X size:0x%02X:quirk max_%s_len must be > %d",
 				    eeprom_addr, buf_size,
-				    read ? "read" : "write", EEPROM_OFFSET_SIZE);
+				    str_read_write(read), EEPROM_OFFSET_SIZE);
 		return -EINVAL;
 	}