Merge tag 'amd-drm-next-6.5-2023-06-09' of https://gitlab.freedesktop.org/agd5f/linux into drm-next

amd-drm-next-6.5-2023-06-02:

amdgpu:
- SR-IOV fixes
- Warning fixes
- Misc code cleanups and spelling fixes
- DCN 3.2 updates
- Improved DC FAMS support for better power management
- Improved DC SubVP support for better power management
- DCN 3.1.x fixes
- Max IB size query
- DC GPU reset fixes
- RAS updates
- DCN 3.0.x fixes
- S/G display fixes
- CP shadow buffer support
- Implement connector force callback
- Z8 power improvements
- PSP 13.0.10 vbflash support
- Mode2 reset fixes
- Store MQDs in VRAM to improve queue switch latency
- VCN 3.x fixes
- JPEG 3.x fixes
- Enable DC_FP on LoongArch
- GFXOFF fixes
- GC 9.4.3 partition support
- SDMA 4.4.2 partition support
- VCN/JPEG 4.0.3 partition support
- VCN 4.0.3 updates
- NBIO 7.9 updates
- GC 9.4.3 updates
- Take NUMA into account when allocating memory
- Handle NUMA for partitions
- SMU 13.0.6 updates
- GC 9.4.3 RAS updates
- Stop including unused swiotlb.h
- SMU 13.0.7 fixes
- Fix clock output ordering on some APUs
- Clean up DC FPGA code
- GFX9 preemption fixes
- Misc irq fixes
- S0ix fixes
- Add new DRM_AMDGPU_WERROR config parameter to help with CI
- PCIe fix for RDNA2
- kdoc fixes
- Documentation updates

amdkfd:
- Query TTM mem limit rather than hardcoding it
- GC 9.4.3 partition support
- Handle NUMA for partitions

radeon:
- Fix possible double free
- Stop including unused swiotlb.h
- Fix possible division by zero

ttm:
- Add query for TTM mem limit
- Add NUMA awareness to pools
- Export ttm_pool_fini()

UAPI:
- Add new ctx query flag to better handle GPU resets
  Mesa MR: https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/22290
- Add new interface to query and set shadow buffer for RDNA3
  Mesa MR: https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21986
- Add new INFO query for max IB size
  Proposed userspace: https://gitlab.freedesktop.org/bnieuwenhuizen/mesa/-/commits/ib-rejection-v3

amd-drm-next-6.5-2023-06-09:

amdgpu:
- S0ix fixes
- Initial SMU13 Overdrive support
- kdoc fixes
- Misc clode cleanups
- Flexible array fixes
- Display OTG fixes
- SMU 13.0.6 updates
- Revert some broken clock counter updates
- Misc display fixes
- GFX9 preemption fixes
- Add support for newer EEPROM bad page table format
- Add missing radeon secondary id
- Add support for new colorspace KMS API
- CSA fix
- Stable pstate fixes for APUs
- make vbl interface admin only
- Handle PCI accelerator class

amdkfd:
- Add debugger support for gdb

radeon:
- Fix possible UAF

drm:
- Add Colorspace functionality

UAPI:
- Add debugger interface for enabling gdb
  Proposed userspace: https://github.com/ROCm-Developer-Tools/ROCdbgapi/tree/wip-dbgapi
- Add KMS colorspace API
  Discussion: https://lists.freedesktop.org/archives/dri-devel/2023-June/408128.html

From: Alex Deucher <alexander.deucher@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20230609174817.7764-1-alexander.deucher@amd.com

1 files changed, 83 insertions, 15 deletions

diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_process_queue_manager.c b/drivers/gpu/drm/amd/amdkfd/kfd_process_queue_manager.c
index 4236539d9f93..9ad1a2186a24 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_process_queue_manager.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_process_queue_manager.c
@@ -81,7 +81,7 @@ static int find_available_queue_slot(struct process_queue_manager *pqm,
 
 void kfd_process_dequeue_from_device(struct kfd_process_device *pdd)
 {
-	struct kfd_dev *dev = pdd->dev;
+	struct kfd_node *dev = pdd->dev;
 
 	if (pdd->already_dequeued)
 		return;
@@ -93,7 +93,7 @@ void kfd_process_dequeue_from_device(struct kfd_process_device *pdd)
 int pqm_set_gws(struct process_queue_manager *pqm, unsigned int qid,
 			void *gws)
 {
-	struct kfd_dev *dev = NULL;
+	struct kfd_node *dev = NULL;
 	struct process_queue_node *pqn;
 	struct kfd_process_device *pdd;
 	struct kgd_mem *mem = NULL;
@@ -178,7 +178,7 @@ void pqm_uninit(struct process_queue_manager *pqm)
 }
 
 static int init_user_queue(struct process_queue_manager *pqm,
-				struct kfd_dev *dev, struct queue **q,
+				struct kfd_node *dev, struct queue **q,
 				struct queue_properties *q_properties,
 				struct file *f, struct amdgpu_bo *wptr_bo,
 				unsigned int qid)
@@ -187,6 +187,7 @@ static int init_user_queue(struct process_queue_manager *pqm,
 
 	/* Doorbell initialized in user space*/
 	q_properties->doorbell_ptr = NULL;
+	q_properties->exception_status = KFD_EC_MASK(EC_QUEUE_NEW);
 
 	/* let DQM handle it*/
 	q_properties->vmid = 0;
@@ -199,7 +200,7 @@ static int init_user_queue(struct process_queue_manager *pqm,
 	(*q)->device = dev;
 	(*q)->process = pqm->process;
 
-	if (dev->shared_resources.enable_mes) {
+	if (dev->kfd->shared_resources.enable_mes) {
 		retval = amdgpu_amdkfd_alloc_gtt_mem(dev->adev,
 						AMDGPU_MES_GANG_CTX_SIZE,
 						&(*q)->gang_ctx_bo,
@@ -224,7 +225,7 @@ cleanup:
 }
 
 int pqm_create_queue(struct process_queue_manager *pqm,
-			    struct kfd_dev *dev,
+			    struct kfd_node *dev,
 			    struct file *f,
 			    struct queue_properties *properties,
 			    unsigned int *qid,
@@ -242,6 +243,13 @@ int pqm_create_queue(struct process_queue_manager *pqm,
 	enum kfd_queue_type type = properties->type;
 	unsigned int max_queues = 127; /* HWS limit */
 
+	/*
+	 * On GFX 9.4.3, increase the number of queues that
+	 * can be created to 255. No HWS limit on GFX 9.4.3.
+	 */
+	if (KFD_GC_VERSION(dev) == IP_VERSION(9, 4, 3))
+		max_queues = 255;
+
 	q = NULL;
 	kq = NULL;
 
@@ -258,7 +266,7 @@ int pqm_create_queue(struct process_queue_manager *pqm,
 	 * Hence we also check the type as well
 	 */
 	if ((pdd->qpd.is_debug) || (type == KFD_QUEUE_TYPE_DIQ))
-		max_queues = dev->device_info.max_no_of_hqd/2;
+		max_queues = dev->kfd->device_info.max_no_of_hqd/2;
 
 	if (pdd->qpd.queue_count >= max_queues)
 		return -ENOSPC;
@@ -330,6 +338,10 @@ int pqm_create_queue(struct process_queue_manager *pqm,
 		kq->queue->properties.queue_id = *qid;
 		pqn->kq = kq;
 		pqn->q = NULL;
+		retval = kfd_process_drain_interrupts(pdd);
+		if (retval)
+			break;
+
 		retval = dev->dqm->ops.create_kernel_queue(dev->dqm,
 							kq, &pdd->qpd);
 		break;
@@ -354,7 +366,7 @@ int pqm_create_queue(struct process_queue_manager *pqm,
 		 */
 		*p_doorbell_offset_in_process =
 			(q->properties.doorbell_off * sizeof(uint32_t)) &
-			(kfd_doorbell_process_slice(dev) - 1);
+			(kfd_doorbell_process_slice(dev->kfd) - 1);
 
 	pr_debug("PQM After DQM create queue\n");
 
@@ -387,7 +399,7 @@ int pqm_destroy_queue(struct process_queue_manager *pqm, unsigned int qid)
 	struct process_queue_node *pqn;
 	struct kfd_process_device *pdd;
 	struct device_queue_manager *dqm;
-	struct kfd_dev *dev;
+	struct kfd_node *dev;
 	int retval;
 
 	dqm = NULL;
@@ -439,7 +451,7 @@ int pqm_destroy_queue(struct process_queue_manager *pqm, unsigned int qid)
 			pdd->qpd.num_gws = 0;
 		}
 
-		if (dev->shared_resources.enable_mes) {
+		if (dev->kfd->shared_resources.enable_mes) {
 			amdgpu_amdkfd_free_gtt_mem(dev->adev,
 						   pqn->q->gang_ctx_bo);
 			if (pqn->q->wptr_bo)
@@ -477,6 +489,7 @@ int pqm_update_queue_properties(struct process_queue_manager *pqm,
 	pqn->q->properties.queue_size = p->queue_size;
 	pqn->q->properties.queue_percent = p->queue_percent;
 	pqn->q->properties.priority = p->priority;
+	pqn->q->properties.pm4_target_xcc = p->pm4_target_xcc;
 
 	retval = pqn->q->device->dqm->ops.update_queue(pqn->q->device->dqm,
 							pqn->q, NULL);
@@ -498,8 +511,12 @@ int pqm_update_mqd(struct process_queue_manager *pqm,
 		return -EFAULT;
 	}
 
+	/* CUs are masked for debugger requirements so deny user mask  */
+	if (pqn->q->properties.is_dbg_wa && minfo && minfo->cu_mask.ptr)
+		return -EBUSY;
+
 	/* ASICs that have WGPs must enforce pairwise enabled mask checks. */
-	if (minfo && minfo->update_flag == UPDATE_FLAG_CU_MASK && minfo->cu_mask.ptr &&
+	if (minfo && minfo->cu_mask.ptr &&
 			KFD_GC_VERSION(pqn->q->device) >= IP_VERSION(10, 0, 0)) {
 		int i;
 
@@ -518,6 +535,9 @@ int pqm_update_mqd(struct process_queue_manager *pqm,
 	if (retval != 0)
 		return retval;
 
+	if (minfo && minfo->cu_mask.ptr)
+		pqn->q->properties.is_user_cu_masked = true;
+
 	return 0;
 }
 
@@ -565,6 +585,46 @@ int pqm_get_wave_state(struct process_queue_manager *pqm,
 						       save_area_used_size);
 }
 
+int pqm_get_queue_snapshot(struct process_queue_manager *pqm,
+			   uint64_t exception_clear_mask,
+			   void __user *buf,
+			   int *num_qss_entries,
+			   uint32_t *entry_size)
+{
+	struct process_queue_node *pqn;
+	struct kfd_queue_snapshot_entry src;
+	uint32_t tmp_entry_size = *entry_size, tmp_qss_entries = *num_qss_entries;
+	int r = 0;
+
+	*num_qss_entries = 0;
+	if (!(*entry_size))
+		return -EINVAL;
+
+	*entry_size = min_t(size_t, *entry_size, sizeof(struct kfd_queue_snapshot_entry));
+	mutex_lock(&pqm->process->event_mutex);
+
+	memset(&src, 0, sizeof(src));
+
+	list_for_each_entry(pqn, &pqm->queues, process_queue_list) {
+		if (!pqn->q)
+			continue;
+
+		if (*num_qss_entries < tmp_qss_entries) {
+			set_queue_snapshot_entry(pqn->q, exception_clear_mask, &src);
+
+			if (copy_to_user(buf, &src, *entry_size)) {
+				r = -EFAULT;
+				break;
+			}
+			buf += tmp_entry_size;
+		}
+		*num_qss_entries += 1;
+	}
+
+	mutex_unlock(&pqm->process->event_mutex);
+	return r;
+}
+
 static int get_queue_data_sizes(struct kfd_process_device *pdd,
 				struct queue *q,
 				uint32_t *mqd_size,
@@ -859,7 +919,7 @@ int kfd_criu_restore_queue(struct kfd_process *p,
 	}
 
 	if (!pdd->doorbell_index &&
-	    kfd_alloc_process_doorbells(pdd->dev, &pdd->doorbell_index) < 0) {
+	    kfd_alloc_process_doorbells(pdd->dev->kfd, &pdd->doorbell_index) < 0) {
 		ret = -ENOMEM;
 		goto exit;
 	}
@@ -927,7 +987,9 @@ int pqm_debugfs_mqds(struct seq_file *m, void *data)
 	struct queue *q;
 	enum KFD_MQD_TYPE mqd_type;
 	struct mqd_manager *mqd_mgr;
-	int r = 0;
+	int r = 0, xcc, num_xccs = 1;
+	void *mqd;
+	uint64_t size = 0;
 
 	list_for_each_entry(pqn, &pqm->queues, process_queue_list) {
 		if (pqn->q) {
@@ -943,6 +1005,7 @@ int pqm_debugfs_mqds(struct seq_file *m, void *data)
 				seq_printf(m, "  Compute queue on device %x\n",
 					   q->device->id);
 				mqd_type = KFD_MQD_TYPE_CP;
+				num_xccs = NUM_XCC(q->device->xcc_mask);
 				break;
 			default:
 				seq_printf(m,
@@ -951,6 +1014,8 @@ int pqm_debugfs_mqds(struct seq_file *m, void *data)
 				continue;
 			}
 			mqd_mgr = q->device->dqm->mqd_mgrs[mqd_type];
+			size = mqd_mgr->mqd_stride(mqd_mgr,
+							&q->properties);
 		} else if (pqn->kq) {
 			q = pqn->kq->queue;
 			mqd_mgr = pqn->kq->mqd_mgr;
@@ -972,9 +1037,12 @@ int pqm_debugfs_mqds(struct seq_file *m, void *data)
 			continue;
 		}
 
-		r = mqd_mgr->debugfs_show_mqd(m, q->mqd);
-		if (r != 0)
-			break;
+		for (xcc = 0; xcc < num_xccs; xcc++) {
+			mqd = q->mqd + size * xcc;
+			r = mqd_mgr->debugfs_show_mqd(m, mqd);
+			if (r != 0)
+				break;
+		}
 	}
 
 	return r;