Merge tag 'amd-drm-next-6.7-2023-10-20' of https://gitlab.freedesktop.org/agd5f/linux into drm-next

amd-drm-next-6.7-2023-10-20:

amdgpu:
- SMU 13 updates
- UMSCH updates
- DC MPO fixes
- RAS updates
- MES 11 fixes
- Fix possible memory leaks in error pathes
- GC 11.5 fixes
- Kernel doc updates
- PSP updates
- APU IMU fixes
- Misc code cleanups
- SMU 11 fixes
- OD fix
- Frame size warning fixes
- SR-IOV fixes
- NBIO 7.11 updates
- NBIO 7.7 updates
- XGMI fixes
- devcoredump updates

amdkfd:
- Misc code cleanups
- SVM fixes

Signed-off-by: Dave Airlie <[email protected]>

From: Alex Deucher <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]

Author: airlied

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

@@ -773,6 +773,17 @@ struct amdgpu_mqd {
 struct amdgpu_reset_domain;
 struct amdgpu_fru_info;
 
+struct amdgpu_reset_info {
+	/* reset dump register */
+	u32 *reset_dump_reg_list;
+	u32 *reset_dump_reg_value;
+	int num_regs;
+
+#ifdef CONFIG_DEV_COREDUMP
+	struct amdgpu_coredump_info *coredump_info;
+#endif
+};
+
 /*
  * Non-zero (true) if the GPU has VRAM. Zero (false) otherwise.
  */
@@ -1081,10 +1092,7 @@ struct amdgpu_device {
 
 	struct mutex			benchmark_mutex;
 
-	/* reset dump register */
-	uint32_t                        *reset_dump_reg_list;
-	uint32_t			*reset_dump_reg_value;
-	int                             num_regs;
+	struct amdgpu_reset_info	reset_info;
 
 	bool                            scpm_enabled;
 	uint32_t                        scpm_status;
@@ -1111,15 +1119,6 @@ static inline uint32_t amdgpu_ip_version(const struct amdgpu_device *adev,
 	return adev->ip_versions[ip][inst] & ~0xFFU;
 }
 
-#ifdef CONFIG_DEV_COREDUMP
-struct amdgpu_coredump_info {
-	struct amdgpu_device		*adev;
-	struct amdgpu_task_info         reset_task_info;
-	struct timespec64               reset_time;
-	bool                            reset_vram_lost;
-};
-#endif
-
 static inline struct amdgpu_device *drm_to_adev(struct drm_device *ddev)
 {
 	return container_of(ddev, struct amdgpu_device, ddev);

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

@@ -2016,8 +2016,8 @@ static ssize_t amdgpu_reset_dump_register_list_read(struct file *f,
 	if (ret)
 		return ret;
 
-	for (i = 0; i < adev->num_regs; i++) {
-		sprintf(reg_offset, "0x%x\n", adev->reset_dump_reg_list[i]);
+	for (i = 0; i < adev->reset_info.num_regs; i++) {
+		sprintf(reg_offset, "0x%x\n", adev->reset_info.reset_dump_reg_list[i]);
 		up_read(&adev->reset_domain->sem);
 		if (copy_to_user(buf + len, reg_offset, strlen(reg_offset)))
 			return -EFAULT;
@@ -2074,9 +2074,9 @@ static ssize_t amdgpu_reset_dump_register_list_write(struct file *f,
 	if (ret)
 		goto error_free;
 
-	swap(adev->reset_dump_reg_list, tmp);
-	swap(adev->reset_dump_reg_value, new);
-	adev->num_regs = i;
+	swap(adev->reset_info.reset_dump_reg_list, tmp);
+	swap(adev->reset_info.reset_dump_reg_value, new);
+	adev->reset_info.num_regs = i;
 	up_write(&adev->reset_domain->sem);
 	ret = size;
 

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

@@ -32,8 +32,6 @@
 #include <linux/slab.h>
 #include <linux/iommu.h>
 #include <linux/pci.h>
-#include <linux/devcoredump.h>
-#include <generated/utsrelease.h>
 #include <linux/pci-p2pdma.h>
 #include <linux/apple-gmux.h>
 
@@ -3578,9 +3576,7 @@ static void amdgpu_device_xgmi_reset_func(struct work_struct *__work)
 		if (adev->asic_reset_res)
 			goto fail;
 
-		if (adev->mmhub.ras && adev->mmhub.ras->ras_block.hw_ops &&
-		    adev->mmhub.ras->ras_block.hw_ops->reset_ras_error_count)
-			adev->mmhub.ras->ras_block.hw_ops->reset_ras_error_count(adev);
+		amdgpu_ras_reset_error_count(adev, AMDGPU_RAS_BLOCK__MMHUB);
 	} else {
 
 		task_barrier_full(&hive->tb);
@@ -5050,90 +5046,16 @@ static int amdgpu_reset_reg_dumps(struct amdgpu_device *adev)
 
 	lockdep_assert_held(&adev->reset_domain->sem);
 
-	for (i = 0; i < adev->num_regs; i++) {
-		adev->reset_dump_reg_value[i] = RREG32(adev->reset_dump_reg_list[i]);
-		trace_amdgpu_reset_reg_dumps(adev->reset_dump_reg_list[i],
-					     adev->reset_dump_reg_value[i]);
-	}
-
-	return 0;
-}
-
-#ifndef CONFIG_DEV_COREDUMP
-static void amdgpu_coredump(struct amdgpu_device *adev, bool vram_lost,
-			    struct amdgpu_reset_context *reset_context)
-{
-}
-#else
-static ssize_t amdgpu_devcoredump_read(char *buffer, loff_t offset,
-		size_t count, void *data, size_t datalen)
-{
-	struct drm_printer p;
-	struct amdgpu_coredump_info *coredump = data;
-	struct drm_print_iterator iter;
-	int i;
-
-	iter.data = buffer;
-	iter.offset = 0;
-	iter.start = offset;
-	iter.remain = count;
-
-	p = drm_coredump_printer(&iter);
-
-	drm_printf(&p, "**** AMDGPU Device Coredump ****\n");
-	drm_printf(&p, "kernel: " UTS_RELEASE "\n");
-	drm_printf(&p, "module: " KBUILD_MODNAME "\n");
-	drm_printf(&p, "time: %lld.%09ld\n", coredump->reset_time.tv_sec, coredump->reset_time.tv_nsec);
-	if (coredump->reset_task_info.pid)
-		drm_printf(&p, "process_name: %s PID: %d\n",
-			   coredump->reset_task_info.process_name,
-			   coredump->reset_task_info.pid);
-
-	if (coredump->reset_vram_lost)
-		drm_printf(&p, "VRAM is lost due to GPU reset!\n");
-	if (coredump->adev->num_regs) {
-		drm_printf(&p, "AMDGPU register dumps:\nOffset:     Value:\n");
-
-		for (i = 0; i < coredump->adev->num_regs; i++)
-			drm_printf(&p, "0x%08x: 0x%08x\n",
-				   coredump->adev->reset_dump_reg_list[i],
-				   coredump->adev->reset_dump_reg_value[i]);
-	}
+	for (i = 0; i < adev->reset_info.num_regs; i++) {
+		adev->reset_info.reset_dump_reg_value[i] =
+			RREG32(adev->reset_info.reset_dump_reg_list[i]);
 
-	return count - iter.remain;
-}
-
-static void amdgpu_devcoredump_free(void *data)
-{
-	kfree(data);
-}
-
-static void amdgpu_coredump(struct amdgpu_device *adev, bool vram_lost,
-			    struct amdgpu_reset_context *reset_context)
-{
-	struct amdgpu_coredump_info *coredump;
-	struct drm_device *dev = adev_to_drm(adev);
-
-	coredump = kzalloc(sizeof(*coredump), GFP_NOWAIT);
-
-	if (!coredump) {
-		DRM_ERROR("%s: failed to allocate memory for coredump\n", __func__);
-		return;
+		trace_amdgpu_reset_reg_dumps(adev->reset_info.reset_dump_reg_list[i],
+					     adev->reset_info.reset_dump_reg_value[i]);
 	}
 
-	coredump->reset_vram_lost = vram_lost;
-
-	if (reset_context->job && reset_context->job->vm)
-		coredump->reset_task_info = reset_context->job->vm->task_info;
-
-	coredump->adev = adev;
-
-	ktime_get_ts64(&coredump->reset_time);
-
-	dev_coredumpm(dev->dev, THIS_MODULE, coredump, 0, GFP_NOWAIT,
-		      amdgpu_devcoredump_read, amdgpu_devcoredump_free);
+	return 0;
 }
-#endif
 
 int amdgpu_do_asic_reset(struct list_head *device_list_handle,
 			 struct amdgpu_reset_context *reset_context)
@@ -5201,9 +5123,7 @@ int amdgpu_do_asic_reset(struct list_head *device_list_handle,
 
 	if (!r && amdgpu_ras_intr_triggered()) {
 		list_for_each_entry(tmp_adev, device_list_handle, reset_list) {
-			if (tmp_adev->mmhub.ras && tmp_adev->mmhub.ras->ras_block.hw_ops &&
-			    tmp_adev->mmhub.ras->ras_block.hw_ops->reset_ras_error_count)
-				tmp_adev->mmhub.ras->ras_block.hw_ops->reset_ras_error_count(tmp_adev);
+			amdgpu_ras_reset_error_count(tmp_adev, AMDGPU_RAS_BLOCK__MMHUB);
 		}
 
 		amdgpu_ras_intr_cleared();

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

@@ -29,6 +29,7 @@
 #include "amdgpu_rlc.h"
 #include "amdgpu_ras.h"
 #include "amdgpu_xcp.h"
+#include "amdgpu_xgmi.h"
 
 /* delay 0.1 second to enable gfx off feature */
 #define GFX_OFF_DELAY_ENABLE         msecs_to_jiffies(100)
@@ -501,6 +502,9 @@ int amdgpu_gfx_disable_kcq(struct amdgpu_device *adev, int xcc_id)
 {
 	struct amdgpu_kiq *kiq = &adev->gfx.kiq[xcc_id];
 	struct amdgpu_ring *kiq_ring = &kiq->ring;
+	struct amdgpu_hive_info *hive;
+	struct amdgpu_ras *ras;
+	int hive_ras_recovery = 0;
 	int i, r = 0;
 	int j;
 
@@ -521,6 +525,23 @@ int amdgpu_gfx_disable_kcq(struct amdgpu_device *adev, int xcc_id)
 					   RESET_QUEUES, 0, 0);
 	}
 
+	/**
+	 * This is workaround: only skip kiq_ring test
+	 * during ras recovery in suspend stage for gfx9.4.3
+	 */
+	hive = amdgpu_get_xgmi_hive(adev);
+	if (hive) {
+		hive_ras_recovery = atomic_read(&hive->ras_recovery);
+		amdgpu_put_xgmi_hive(hive);
+	}
+
+	ras = amdgpu_ras_get_context(adev);
+	if ((amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 3)) &&
+		ras && (atomic_read(&ras->in_recovery) || hive_ras_recovery)) {
+		spin_unlock(&kiq->ring_lock);
+		return 0;
+	}
+
 	if (kiq_ring->sched.ready && !adev->job_hang)
 		r = amdgpu_ring_test_helper(kiq_ring);
 	spin_unlock(&kiq->ring_lock);

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

@@ -1267,6 +1267,8 @@ int psp_xgmi_initialize(struct psp_context *psp, bool set_extended_data, bool lo
 	xgmi_cmd->cmd_id = TA_COMMAND_XGMI__INITIALIZE;
 
 	ret = psp_xgmi_invoke(psp, xgmi_cmd->cmd_id);
+	/* note down the capbility flag for XGMI TA */
+	psp->xgmi_context.xgmi_ta_caps = xgmi_cmd->caps_flag;
 
 	return ret;
 }
@@ -1388,7 +1390,7 @@ int psp_xgmi_get_topology_info(struct psp_context *psp,
 
 	/* Fill in the shared memory with topology information as input */
 	topology_info_input = &xgmi_cmd->xgmi_in_message.get_topology_info;
-	xgmi_cmd->cmd_id = TA_COMMAND_XGMI__GET_GET_TOPOLOGY_INFO;
+	xgmi_cmd->cmd_id = TA_COMMAND_XGMI__GET_TOPOLOGY_INFO;
 	topology_info_input->num_nodes = number_devices;
 
 	for (i = 0; i < topology_info_input->num_nodes; i++) {
@@ -1399,7 +1401,7 @@ int psp_xgmi_get_topology_info(struct psp_context *psp,
 	}
 
 	/* Invoke xgmi ta to get the topology information */
-	ret = psp_xgmi_invoke(psp, TA_COMMAND_XGMI__GET_GET_TOPOLOGY_INFO);
+	ret = psp_xgmi_invoke(psp, TA_COMMAND_XGMI__GET_TOPOLOGY_INFO);
 	if (ret)
 		return ret;
 
@@ -1424,28 +1426,53 @@ int psp_xgmi_get_topology_info(struct psp_context *psp,
 
 	/* Invoke xgmi ta again to get the link information */
 	if (psp_xgmi_peer_link_info_supported(psp)) {
-		struct ta_xgmi_cmd_get_peer_link_info_output *link_info_output;
+		struct ta_xgmi_cmd_get_peer_link_info *link_info_output;
+		struct ta_xgmi_cmd_get_extend_peer_link_info *link_extend_info_output;
 		bool requires_reflection =
 			(psp->xgmi_context.supports_extended_data &&
 			 get_extended_data) ||
 			amdgpu_ip_version(psp->adev, MP0_HWIP, 0) ==
 				IP_VERSION(13, 0, 6);
+		bool ta_port_num_support = psp->xgmi_context.xgmi_ta_caps &
+						EXTEND_PEER_LINK_INFO_CMD_FLAG;
 
-		xgmi_cmd->cmd_id = TA_COMMAND_XGMI__GET_PEER_LINKS;
+		/* popluate the shared output buffer rather than the cmd input buffer
+		 * with node_ids as the input for GET_PEER_LINKS command execution.
+		 * This is required for GET_PEER_LINKS per xgmi ta implementation.
+		 * The same requirement for GET_EXTEND_PEER_LINKS command.
+		 */
+		if (ta_port_num_support) {
+			link_extend_info_output = &xgmi_cmd->xgmi_out_message.get_extend_link_info;
+
+			for (i = 0; i < topology->num_nodes; i++)
+				link_extend_info_output->nodes[i].node_id = topology->nodes[i].node_id;
+
+			link_extend_info_output->num_nodes = topology->num_nodes;
+			xgmi_cmd->cmd_id = TA_COMMAND_XGMI__GET_EXTEND_PEER_LINKS;
+		} else {
+			link_info_output = &xgmi_cmd->xgmi_out_message.get_link_info;
 
-		ret = psp_xgmi_invoke(psp, TA_COMMAND_XGMI__GET_PEER_LINKS);
+			for (i = 0; i < topology->num_nodes; i++)
+				link_info_output->nodes[i].node_id = topology->nodes[i].node_id;
 
+			link_info_output->num_nodes = topology->num_nodes;
+			xgmi_cmd->cmd_id = TA_COMMAND_XGMI__GET_PEER_LINKS;
+		}
+
+		ret = psp_xgmi_invoke(psp, xgmi_cmd->cmd_id);
 		if (ret)
 			return ret;
 
-		link_info_output = &xgmi_cmd->xgmi_out_message.get_link_info;
 		for (i = 0; i < topology->num_nodes; i++) {
+			uint8_t node_num_links = ta_port_num_support ?
+				link_extend_info_output->nodes[i].num_links : link_info_output->nodes[i].num_links;
 			/* accumulate num_links on extended data */
-			topology->nodes[i].num_links = get_extended_data ?
-					topology->nodes[i].num_links +
-							link_info_output->nodes[i].num_links :
-					((requires_reflection && topology->nodes[i].num_links) ? topology->nodes[i].num_links :
-					 link_info_output->nodes[i].num_links);
+			if (get_extended_data) {
+				topology->nodes[i].num_links = topology->nodes[i].num_links + node_num_links;
+			} else {
+				topology->nodes[i].num_links = (requires_reflection && topology->nodes[i].num_links) ?
+								topology->nodes[i].num_links : node_num_links;
+			}
 
 			/* reflect the topology information for bi-directionality */
 			if (requires_reflection && topology->nodes[i].num_hops)

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

@@ -189,6 +189,7 @@ struct psp_xgmi_context {
 	struct ta_context		context;
 	struct psp_xgmi_topology_info	top_info;
 	bool				supports_extended_data;
+	uint8_t				xgmi_ta_caps;
 };
 
 struct psp_ras_context {