Hi Mika, kernel test robot noticed the following build warnings: [auto build test WARNING on drm-misc/drm-misc-next] [also build test WARNING on drm-tip/drm-tip v6.12 next-20241122] [If your patch is applied to the wrong git tree, kindly drop us a note. And when submitting patch, we suggest to use '--base' as documented in https://git-scm.com/docs/git-format-patch#_base_tree_information] url: https://github.com/intel-lab-lkp/linux/commits/Mika-Laitio/ammdgpu-fix-for-gfx1103-queue-evict-restore-crash/20241122-035602 base: git://anongit.freedesktop.org/drm/drm-misc drm-misc-next patch link: https://lore.kernel.org/r/20241121195233.10679-1-lamikr%40gmail.com patch subject: [PATCH] ammdgpu fix for gfx1103 queue evict/restore crash config: powerpc-allmodconfig (https://download.01.org/0day-ci/archive/20241123/202411231721.vBNjHNvr-lkp@xxxxxxxxx/config) compiler: powerpc64-linux-gcc (GCC) 14.2.0 reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20241123/202411231721.vBNjHNvr-lkp@xxxxxxxxx/reproduce) If you fix the issue in a separate patch/commit (i.e. not just a new version of the same patch/commit), kindly add following tags | Reported-by: kernel test robot <lkp@xxxxxxxxx> | Closes: https://lore.kernel.org/oe-kbuild-all/202411231721.vBNjHNvr-lkp@xxxxxxxxx/ All warnings (new ones prefixed by >>): drivers/gpu/drm/amd/amdgpu/../amdkfd/kfd_device_queue_manager.c: In function 'restore_process_queues_cpsch': >> drivers/gpu/drm/amd/amdgpu/../amdkfd/kfd_device_queue_manager.c:1354:1: warning: label 'out_unlock' defined but not used [-Wunused-label] 1354 | out_unlock: | ^~~~~~~~~~ vim +/out_unlock +1354 drivers/gpu/drm/amd/amdgpu/../amdkfd/kfd_device_queue_manager.c 1292 1293 static int restore_process_queues_cpsch(struct device_queue_manager *dqm, 1294 struct qcm_process_device *qpd) 1295 { 1296 struct queue *q; 1297 struct device *dev = dqm->dev->adev->dev; 1298 struct kfd_process_device *pdd; 1299 uint64_t eviction_duration; 1300 int retval = 0; 1301 1302 // gfx1103 APU fails to remove the queue usually after 10-50 attempts 1303 if (dqm->dev->adev->flags & AMD_IS_APU) 1304 goto out; 1305 pdd = qpd_to_pdd(qpd); 1306 1307 dqm_lock(dqm); 1308 if (WARN_ON_ONCE(!qpd->evicted)) /* already restored, do nothing */ 1309 goto out; 1310 if (qpd->evicted > 1) { /* ref count still > 0, decrement & quit */ 1311 qpd->evicted--; 1312 goto out; 1313 } 1314 1315 /* The debugger creates processes that temporarily have not acquired 1316 * all VMs for all devices and has no VMs itself. 1317 * Skip queue restore on process restore. 1318 */ 1319 if (!pdd->drm_priv) 1320 goto vm_not_acquired; 1321 1322 pr_debug_ratelimited("Restoring PASID 0x%x queues\n", 1323 pdd->process->pasid); 1324 1325 /* Update PD Base in QPD */ 1326 qpd->page_table_base = amdgpu_amdkfd_gpuvm_get_process_page_dir(pdd->drm_priv); 1327 pr_debug("Updated PD address to 0x%llx\n", qpd->page_table_base); 1328 1329 /* activate all active queues on the qpd */ 1330 list_for_each_entry(q, &qpd->queues_list, list) { 1331 q->properties.is_evicted = false; 1332 if (!QUEUE_IS_ACTIVE(q->properties)) 1333 continue; 1334 1335 q->properties.is_active = true; 1336 increment_queue_count(dqm, &pdd->qpd, q); 1337 1338 if (dqm->dev->kfd->shared_resources.enable_mes) { 1339 retval = add_queue_mes(dqm, q, qpd); 1340 if (retval) { 1341 dev_err(dev, "Failed to restore queue %d\n", 1342 q->properties.queue_id); 1343 goto out; 1344 } 1345 } 1346 } 1347 if (!dqm->dev->kfd->shared_resources.enable_mes) 1348 retval = execute_queues_cpsch(dqm, 1349 KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES, 0, USE_DEFAULT_GRACE_PERIOD); 1350 eviction_duration = get_jiffies_64() - pdd->last_evict_timestamp; 1351 atomic64_add(eviction_duration, &pdd->evict_duration_counter); 1352 vm_not_acquired: 1353 qpd->evicted = 0; > 1354 out_unlock: 1355 dqm_unlock(dqm); 1356 out: 1357 return retval; 1358 } 1359 -- 0-DAY CI Kernel Test Service https://github.com/intel/lkp-tests/wiki
