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

Fix wrong loop termination condition

Browse Source 
Compare with gpu_mem_count instead of deprecated NUM_OF_SUPPORTED_GPUS
to prevent overflows in case no dGPUs are present.

Change-Id: I71fcb7503ba4c20bffadbdb04cefc4e4027a7df7
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
This commit is contained in:
Felix Kuehling
2018-07-05 15:32:14 -04:00
parent 4839882fc8
commit d3228f363e
1 changed files with 2 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/fmm.c
+2 -2
View File
@@ -223,11 +223,11 @@ static void print_device_id_array(uint32_t *device_id_array, uint32_t device_id_
static int32_t find_first_dgpu(HSAuint32 *gpu_id)
{
int32_t i;
uint32_t i;
*gpu_id = NON_VALID_GPU_ID;
for (i = 0; i < NUM_OF_SUPPORTED_GPUS; i++) {
for (i = 0; i < gpu_mem_count; i++) {
if (gpu_mem[i].gpu_id == NON_VALID_GPU_ID)
continue;
if (!topology_is_dgpu(gpu_mem[i].device_id))