AI/rocm-systems
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

libhsakmt: add FamilyID info into HsaNodeProperties

Browse Source 
Query family id info from drm render node, then
ROCr can query this info directly from Thunk
instead of parsing the info by itself.

Signed-off-by: Lang Yu <Lang.Yu@amd.com>
Change-Id: I030bd27ab2379fbf87f3d787302c3b8613456278
This commit is contained in:
Lang Yu
2022-08-17 11:24:16 +08:00
parent 4267c4b524
commit 66e9e97e0d
2 changed files with 35 additions and 19 deletions
Download Patch File Download Diff File Expand all files Collapse all files
include/hsakmttypes.h
+3 -1
View File
@@ -330,7 +330,9 @@ typedef struct _HsaNodeProperties
HSAuint32 SGPRSizePerCU; // SGPR size in bytes per CU
HSAuint32 KFDGpuID; // GPU Hash ID generated by KFD
HSAuint8 Reserved[8];
HSAuint32 FamilyID; // GPU family id
HSAuint8 Reserved[4];
} HsaNodeProperties;
src/topology.c
+32 -18
View File
@@ -1004,34 +1004,48 @@ exit:
return ret;
}
static int topology_get_marketing_name(int minor, uint16_t *marketing_name)
static int topology_get_node_props_from_drm(HsaNodeProperties *props)
{
int drm_fd;
uint32_t major_version;
uint32_t minor_version;
amdgpu_device_handle device_handle;
struct amdgpu_gpu_info gpu_info;
const char *name;
int i;
int i, ret = 0;
if (marketing_name == NULL)
if (props->MarketingName == NULL)
return -1;
drm_fd = drmOpenRender(minor);
drm_fd = drmOpenRender(props->DrmRenderMinor);
if (drm_fd < 0)
return -1;
if (amdgpu_device_initialize(drm_fd,
&major_version, &minor_version, &device_handle) < 0) {
drmClose(drm_fd);
return -1;
ret = -1;
goto err_device_initialize;
}
name = amdgpu_get_marketing_name(device_handle);
if (name != NULL) {
for (i = 0; name[i] != 0 && i < HSA_PUBLIC_NAME_SIZE - 1; i++)
marketing_name[i] = name[i];
marketing_name[i] = '\0';
props->MarketingName[i] = name[i];
props->MarketingName[i] = '\0';
}
if (amdgpu_query_gpu_info(device_handle, &gpu_info)) {
ret = -1;
goto err_query_gpu_info;
}
props->FamilyID = gpu_info.family_id;
err_query_gpu_info:
amdgpu_device_deinitialize(device_handle);
err_device_initialize:
drmClose(drm_fd);
return 0;
return ret;
}
static HSAKMT_STATUS topology_sysfs_get_node_props(uint32_t node_id,
@@ -1170,6 +1184,10 @@ static HSAKMT_STATUS topology_sysfs_get_node_props(uint32_t node_id,
gfxv = (uint32_t)prop_val;
}
/* Bail out early, if a CPU node */
if (props->NumCPUCores)
goto err;
gfxv_major = HSA_GET_GFX_VERSION_MAJOR(gfxv);
gfxv_minor = HSA_GET_GFX_VERSION_MINOR(gfxv);
gfxv_stepping = HSA_GET_GFX_VERSION_STEP(gfxv);
@@ -1210,15 +1228,11 @@ static HSAKMT_STATUS topology_sysfs_get_node_props(uint32_t node_id,
snprintf((char *)props->AMDName, sizeof(props->AMDName)-1, "GFX%06x",
HSA_GET_GFX_VERSION_FULL(props->EngineId.ui32));
if (!props->NumCPUCores) {
/* Is dGPU Node, not APU
* Retrieve the marketing name of the node.
*/
if (topology_get_marketing_name(props->DrmRenderMinor,
props->MarketingName) != 0)
pr_info("failed to get marketing name for device ID 0x%x\n",
props->DeviceId);
}
/* Is dGPU Node, not APU
* Retrieve the marketing name of the node.
*/
if (topology_get_node_props_from_drm(props))
pr_info("failed to get marketing name for device ID 0x%x\n", props->DeviceId);
/* Get VGPR/SGPR size in byte per CU */
props->SGPRSizePerCU = SGPR_SIZE_PER_CU;