drm/amdgpu:Support retiring multiple MCA error address pages

Support retiring multiple MCA error address pages in
one in-band query for umc v12_0.

Signed-off-by: YiPeng Chai <YiPeng.Chai@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

1 files changed, 35 insertions, 8 deletions

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
index 349c810d2399..3c1106cf9d80 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
@@ -3920,8 +3920,7 @@ static int ras_err_info_cmp(void *priv, const struct list_head *a, const struct
 }
 
 static struct ras_err_info *amdgpu_ras_error_get_info(struct ras_err_data *err_data,
-				struct amdgpu_smuio_mcm_config_info *mcm_info,
-				struct ras_err_addr *err_addr)
+				struct amdgpu_smuio_mcm_config_info *mcm_info)
 {
 	struct ras_err_node *err_node;
 
@@ -3933,10 +3932,9 @@ static struct ras_err_info *amdgpu_ras_error_get_info(struct ras_err_data *err_d
 	if (!err_node)
 		return NULL;
 
-	memcpy(&err_node->err_info.mcm_info, mcm_info, sizeof(*mcm_info));
+	INIT_LIST_HEAD(&err_node->err_info.err_addr_list);
 
-	if (err_addr)
-		memcpy(&err_node->err_info.err_addr, err_addr, sizeof(*err_addr));
+	memcpy(&err_node->err_info.mcm_info, mcm_info, sizeof(*mcm_info));
 
 	err_data->err_list_count++;
 	list_add_tail(&err_node->node, &err_data->err_node_list);
@@ -3945,6 +3943,29 @@ static struct ras_err_info *amdgpu_ras_error_get_info(struct ras_err_data *err_d
 	return &err_node->err_info;
 }
 
+void amdgpu_ras_add_mca_err_addr(struct ras_err_info *err_info, struct ras_err_addr *err_addr)
+{
+	struct ras_err_addr *mca_err_addr;
+
+	mca_err_addr = kzalloc(sizeof(*mca_err_addr), GFP_KERNEL);
+	if (!mca_err_addr)
+		return;
+
+	INIT_LIST_HEAD(&mca_err_addr->node);
+
+	mca_err_addr->err_status = err_addr->err_status;
+	mca_err_addr->err_ipid = err_addr->err_ipid;
+	mca_err_addr->err_addr = err_addr->err_addr;
+
+	list_add_tail(&mca_err_addr->node, &err_info->err_addr_list);
+}
+
+void amdgpu_ras_del_mca_err_addr(struct ras_err_info *err_info, struct ras_err_addr *mca_err_addr)
+{
+	list_del(&mca_err_addr->node);
+	kfree(mca_err_addr);
+}
+
 int amdgpu_ras_error_statistic_ue_count(struct ras_err_data *err_data,
 		struct amdgpu_smuio_mcm_config_info *mcm_info,
 		struct ras_err_addr *err_addr, u64 count)
@@ -3957,10 +3978,13 @@ int amdgpu_ras_error_statistic_ue_count(struct ras_err_data *err_data,
 	if (!count)
 		return 0;
 
-	err_info = amdgpu_ras_error_get_info(err_data, mcm_info, err_addr);
+	err_info = amdgpu_ras_error_get_info(err_data, mcm_info);
 	if (!err_info)
 		return -EINVAL;
 
+	if (err_addr && err_addr->err_status)
+		amdgpu_ras_add_mca_err_addr(err_info, err_addr);
+
 	err_info->ue_count += count;
 	err_data->ue_count += count;
 
@@ -3979,7 +4003,7 @@ int amdgpu_ras_error_statistic_ce_count(struct ras_err_data *err_data,
 	if (!count)
 		return 0;
 
-	err_info = amdgpu_ras_error_get_info(err_data, mcm_info, err_addr);
+	err_info = amdgpu_ras_error_get_info(err_data, mcm_info);
 	if (!err_info)
 		return -EINVAL;
 
@@ -4001,10 +4025,13 @@ int amdgpu_ras_error_statistic_de_count(struct ras_err_data *err_data,
 	if (!count)
 		return 0;
 
-	err_info = amdgpu_ras_error_get_info(err_data, mcm_info, err_addr);
+	err_info = amdgpu_ras_error_get_info(err_data, mcm_info);
 	if (!err_info)
 		return -EINVAL;
 
+	if (err_addr && err_addr->err_status)
+		amdgpu_ras_add_mca_err_addr(err_info, err_addr);
+
 	err_info->de_count += count;
 	err_data->de_count += count;