[PATCH 19/20] drm/amdkfd: add RAS capabilities in topology for Vega20 (v2)

[Alex Deucher]

From: Eric Huang <JinhuiEric.Huang at amd.com>

It is to collaborate with HSA_CAPABILITY in libhsakmt.

v2: squash in NULL pointer check

Signed-off-by: Eric Huang <JinhuiEric.Huang at amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling at amd.com>
Reviewed-by: Alex Deucher <alexander.deucher at amd.com>
Signed-off-by: Alex Deucher <alexander.deucher at amd.com>
---
 drivers/gpu/drm/amd/amdkfd/kfd_topology.c | 16 ++++++++++++++++
 drivers/gpu/drm/amd/amdkfd/kfd_topology.h |  4 ++++
 2 files changed, 20 insertions(+)

diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_topology.c b/drivers/gpu/drm/amd/amdkfd/kfd_topology.c
index 09da91644f9f..2cb09e088dce 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_topology.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_topology.c
@@ -37,6 +37,7 @@
 #include "kfd_device_queue_manager.h"
 #include "kfd_iommu.h"
 #include "amdgpu_amdkfd.h"
+#include "amdgpu_ras.h"
 
 /* topology_device_list - Master list of all topology devices */
 static struct list_head topology_device_list;
@@ -1197,6 +1198,7 @@ int kfd_topology_add_device(struct kfd_dev *gpu)
 	void *crat_image = NULL;
 	size_t image_size = 0;
 	int proximity_domain;
+	struct amdgpu_ras *ctx;
 
 	INIT_LIST_HEAD(&temp_topology_device_list);
 
@@ -1328,6 +1330,20 @@ int kfd_topology_add_device(struct kfd_dev *gpu)
 		dev->node_props.capability |= HSA_CAP_ATS_PRESENT;
 	}
 
+	ctx = amdgpu_ras_get_context((struct amdgpu_device *)(dev->gpu->kgd));
+	if (ctx) {
+		/* kfd only concerns sram ecc on GFX/SDMA and HBM ecc on UMC */
+		dev->node_props.capability |=
+			(((ctx->features & BIT(AMDGPU_RAS_BLOCK__SDMA)) != 0) ||
+			 ((ctx->features & BIT(AMDGPU_RAS_BLOCK__GFX)) != 0)) ?
+			HSA_CAP_SRAM_EDCSUPPORTED : 0;
+		dev->node_props.capability |= ((ctx->features & BIT(AMDGPU_RAS_BLOCK__UMC)) != 0) ?
+			HSA_CAP_MEM_EDCSUPPORTED : 0;
+
+		dev->node_props.capability |= (ctx->features != 0) ?
+			HSA_CAP_RASEVENTNOTIFY : 0;
+	}
+
 	kfd_debug_print_topology();
 
 	if (!res)
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_topology.h b/drivers/gpu/drm/amd/amdkfd/kfd_topology.h
index 92a19be07344..84710cfd23c2 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_topology.h
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_topology.h
@@ -48,6 +48,10 @@
 #define HSA_CAP_DOORBELL_TYPE_2_0		0x2
 #define HSA_CAP_AQL_QUEUE_DOUBLE_MAP		0x00004000
 
+#define HSA_CAP_SRAM_EDCSUPPORTED		0x00080000
+#define HSA_CAP_MEM_EDCSUPPORTED		0x00100000
+#define HSA_CAP_RASEVENTNOTIFY			0x00200000
+
 struct kfd_node_properties {
 	uint64_t hive_id;
 	uint32_t cpu_cores_count;
-- 
2.20.1