]> git.proxmox.com Git - mirror_ubuntu-eoan-kernel.git/commitdiff
drm/amd/powerplay: implement smu_init(fini)_fb_allocations function
authorKevin Wang <Kevin1.Wang@amd.com>
Fri, 14 Dec 2018 10:34:20 +0000 (18:34 +0800)
committerAlex Deucher <alexander.deucher@amd.com>
Tue, 19 Mar 2019 20:03:55 +0000 (15:03 -0500)
This patch implements smu_init_fb_allocations/smu_fini_fb_allocations function
for smu to reserve the BOs for smc tables.

Signed-off-by: Kevin Wang <Kevin1.Wang@amd.com>
Reviewed-by: Huang Rui <ray.huang@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
drivers/gpu/drm/amd/powerplay/amdgpu_smu.c

index b275139ae96b411f4f9ae49950f5fef16bd6848b..ef377ef0e3896d48dec481931567ba20355f211a 100644 (file)
@@ -178,10 +178,63 @@ static int smu_sw_fini(void *handle)
 
 static int smu_init_fb_allocations(struct smu_context *smu)
 {
-       /* TODO */
+       struct amdgpu_device *adev = smu->adev;
+       struct smu_table_context *smu_table = &smu->smu_table;
+       struct smu_table *tables = smu_table->tables;
+       uint32_t table_count = smu_table->table_count;
+       uint32_t i = 0;
+       int32_t ret = 0;
+
+       if (table_count <= 0)
+               return -EINVAL;
+
+       for (i = 0 ; i < table_count; i++) {
+               if (tables[i].size == 0)
+                       continue;
+               ret = amdgpu_bo_create_kernel(adev,
+                                             tables[i].size,
+                                             tables[i].align,
+                                             tables[i].domain,
+                                             &tables[i].bo,
+                                             &tables[i].mc_address,
+                                             &tables[i].cpu_addr);
+               if (ret)
+                       goto failed;
+       }
+
        return 0;
+failed:
+       for (; i > 0; i--) {
+               if (tables[i].size == 0)
+                       continue;
+               amdgpu_bo_free_kernel(&tables[i].bo,
+                                     &tables[i].mc_address,
+                                     &tables[i].cpu_addr);
+
+       }
+       return ret;
 }
 
+static int smu_fini_fb_allocations(struct smu_context *smu)
+{
+       struct smu_table_context *smu_table = &smu->smu_table;
+       struct smu_table *tables = smu_table->tables;
+       uint32_t table_count = smu_table->table_count;
+       uint32_t i = 0;
+
+       if (table_count == 0 || tables == NULL)
+               return -EINVAL;
+
+       for (i = 0 ; i < table_count; i++) {
+               if (tables[i].size == 0)
+                       continue;
+               amdgpu_bo_free_kernel(&tables[i].bo,
+                                     &tables[i].mc_address,
+                                     &tables[i].cpu_addr);
+       }
+
+       return 0;
+}
 static int smu_smc_table_hw_init(struct smu_context *smu)
 {
        int ret;
@@ -329,10 +382,15 @@ static int smu_hw_fini(void *handle)
 {
        struct amdgpu_device *adev = (struct amdgpu_device *)handle;
        struct smu_context *smu = &adev->smu;
+       int ret = 0;
 
        if (adev->asic_type < CHIP_VEGA20)
                return -EINVAL;
 
+       ret = smu_fini_fb_allocations(smu);
+       if (ret)
+               return ret;
+
        return 0;
 }