| /linux/tools/testing/selftests/cgroup/ |
| H A D | memcg_protection.m | 6 % hierarchy to illustrate how overcommitted protection spreads among siblings
9 % Simulation assumes siblings consumed the initial amount of memory (w/out
16 % n vector nominal protection of siblings set at the given level (memory.low)
48 siblings = sum(u); variable
52 e = protected * min(1, E / siblings); % normalize overcommit
55 unclaimed = max(0, E - siblings);
56 parent_overuse = sum(c) - siblings;
79 % XXX here I do parallel reclaim of all siblings
|
| /linux/Documentation/admin-guide/hw-vuln/ |
| H A D | core-scheduling.rst | 100 siblings of a core such that all the selected tasks running on a core are
107 the sibling has the task enqueued. For rest of the siblings in the core,
112 Once a task has been selected for all the siblings in the core, an IPI is sent to
113 siblings for whom a new task was selected. Siblings on receiving the IPI will
125 siblings could be forced to select a lower priority task if the highest
157 and are considered system-wide trusted. The forced-idling of siblings running
174 the siblings to switch to the new task. But there could be hardware delays in
176 cause an attacker task to start running on a CPU before its siblings receive the
177 IPI. Even though cache is flushed on entry to user mode, victim tasks on siblings
185 Core scheduling cannot protect against MDS attacks between the siblings
[all …]
|
| /linux/drivers/infiniband/hw/irdma/ |
| H A D | ws.c | 133 list_for_each_entry(node, &parent->child_list_head, siblings) { in ws_find_node()
139 list_for_each_entry(node, &parent->child_list_head, siblings) { in ws_find_node()
214 list_del(&tc_node->siblings); in irdma_remove_leaf()
219 list_del(&vsi_node->siblings); in irdma_remove_leaf()
295 list_add(&vsi_node->siblings, &ws_tree_root->child_list_head); in irdma_ws_add()
322 list_add(&tc_node->siblings, &vsi_node->child_list_head); in irdma_ws_add()
356 list_del(&tc_node->siblings); in irdma_ws_add()
362 list_del(&vsi_node->siblings); in irdma_ws_add()
|
| H A D | ws.h | 19 struct list_head siblings; member
|
| /linux/drivers/gpu/drm/i915/gt/uc/ |
| H A D | selftest_guc_multi_lrc.c | 34 struct intel_engine_cs *siblings[MAX_ENGINE_INSTANCE + 1]; in multi_lrc_create_parent() local
43 siblings[i++] = engine; in multi_lrc_create_parent()
49 logical_sort(siblings, i); in multi_lrc_create_parent()
51 return intel_engine_create_parallel(siblings, 1, i); in multi_lrc_create_parent()
|
| /linux/arch/x86/kernel/cpu/microcode/ |
| H A D | core.c | 247 * - HT siblings must be idle and not execute other code while the other sibling
402 * If the update was successful, let the siblings run the apply() in __load_primary()
451 /* Kick soft-offlined SMT siblings if required */ in load_primary()
459 /* Unconditionally release soft-offlined SMT siblings if required */ in load_primary()
534 unsigned int cpu, updated = 0, failed = 0, timedout = 0, siblings = 0; in load_late_stop_cpus() local
567 case UCODE_OK: siblings++; break; in load_late_stop_cpus()
583 pr_warn("%u offline siblings did not respond.\n", in load_late_stop_cpus()
595 pr_info("load: updated on %u primary CPUs with %u siblings\n", updated, siblings); in load_late_stop_cpus()
598 num_online_cpus() - (updated + siblings)); in load_late_stop_cpus()
603 return updated + siblings == num_online_cpus() ? 0 : -EIO; in load_late_stop_cpus()
|
| /linux/drivers/gpu/drm/i915/gem/ |
| H A D | i915_gem_context.c | 206 kfree(pc->user_engines[i].siblings); in proto_context_close()
407 struct intel_engine_cs **siblings; in set_proto_ctx_engines_balance() local
445 siblings = kmalloc_array(num_siblings, sizeof(*siblings), GFP_KERNEL); in set_proto_ctx_engines_balance()
446 if (!siblings) in set_proto_ctx_engines_balance()
457 siblings[n] = intel_engine_lookup_user(i915, in set_proto_ctx_engines_balance()
460 if (!siblings[n]) { in set_proto_ctx_engines_balance()
471 set->engines[idx].engine = siblings[0]; in set_proto_ctx_engines_balance()
472 kfree(siblings); in set_proto_ctx_engines_balance()
476 set->engines[idx].siblings = siblings; in set_proto_ctx_engines_balance()
482 kfree(siblings); in set_proto_ctx_engines_balance()
[all …]
|
| H A D | i915_gem_context_types.h | 102 * i915_gem_proto_engine::num_siblings and i915_gem_proto_engine::siblings.
106 * i915_gem_proto_engine::siblings.
115 /** @num_siblings: Number of balanced or parallel siblings */
121 /** @siblings: Balanced siblings or num_siblings * width for parallel */
122 struct intel_engine_cs **siblings; member
|
| /linux/drivers/gpu/drm/i915/gt/ |
| H A D | selftest_execlists.c | 3712 struct intel_engine_cs **siblings, in nop_virtual_engine() argument
3729 ve[n] = intel_engine_create_virtual(siblings, nsibling, 0); in nop_virtual_engine()
3840 struct intel_engine_cs **siblings, in __select_siblings() argument
3853 siblings[n++] = gt->engine_class[class][inst]; in __select_siblings()
3862 struct intel_engine_cs **siblings) in select_siblings() argument
3864 return __select_siblings(gt, class, siblings, NULL); in select_siblings()
3870 struct intel_engine_cs *siblings[MAX_ENGINE_INSTANCE + 1]; in live_virtual_engine() local
3891 nsibling = select_siblings(gt, class, siblings); in live_virtual_engine()
3896 err = nop_virtual_engine(gt, siblings, nsibling, in live_virtual_engine()
3902 err = nop_virtual_engine(gt, siblings, nsibling, n, CHAIN); in live_virtual_engine()
[all …]
|
| H A D | intel_execlists_submission.c | 198 struct intel_engine_cs *siblings[]; member
208 execlists_create_virtual(struct intel_engine_cs **siblings, unsigned int count,
638 * We have to kick all the siblings again in case we need to in __execlists_schedule_out()
1054 if (likely(engine == ve->siblings[0])) in virtual_xfer_context()
1068 if (ve->siblings[n] == engine) { in virtual_xfer_context()
1069 swap(ve->siblings[n], ve->siblings[0]); in virtual_xfer_context()
1443 str_yes_no(engine != ve->siblings[0])); in execlists_dequeue()
1465 * ve->siblings[] on an idle context, where in execlists_dequeue()
1466 * we may be using ve->siblings[] in in execlists_dequeue()
1470 GEM_BUG_ON(ve->siblings[0] != engine); in execlists_dequeue()
[all …]
|
| /linux/drivers/gpio/ |
| H A D | gpio-sim.c | 637 struct list_head siblings; member
671 struct list_head siblings; member
767 list_for_each_entry(line, &bank->line_list, siblings) { in gpio_sim_get_line_names_size()
782 list_for_each_entry(line, &bank->line_list, siblings) { in gpio_sim_set_line_names()
815 list_for_each_entry(bank, &dev->bank_list, siblings) { in gpio_sim_add_hogs()
816 list_for_each_entry(line, &bank->line_list, siblings) { in gpio_sim_add_hogs()
833 list_for_each_entry(bank, &dev->bank_list, siblings) { in gpio_sim_add_hogs()
834 list_for_each_entry(line, &bank->line_list, siblings) { in gpio_sim_add_hogs()
933 list_for_each_entry(this, &dev->bank_list, siblings) { in gpio_sim_bank_labels_non_unique()
934 list_for_each_entry(pos, &dev->bank_list, siblings) { in gpio_sim_bank_labels_non_unique()
[all …]
|
| H A D | gpio-virtuser.c | 1039 struct list_head siblings; member
1058 struct list_head siblings; member
1401 list_for_each_entry(lookup, &dev->lookup_list, siblings) in gpio_virtuser_get_lookup_count()
1426 list_for_each_entry(lookup, &dev->lookup_list, siblings) { in gpio_virtuser_make_lookup_table()
1427 list_for_each_entry(entry, &lookup->entry_list, siblings) { in gpio_virtuser_make_lookup_table()
1458 list_for_each_entry(lookup, &dev->lookup_list, siblings) in gpio_virtuser_make_device_swnode()
1577 list_del(&entry->siblings); in gpio_virtuser_lookup_entry_config_group_release()
1617 list_add_tail(&entry->siblings, &lookup->entry_list); in gpio_virtuser_make_lookup_entry_group()
1629 list_del(&lookup->siblings); in gpio_virtuser_lookup_config_group_release()
1677 list_add_tail(&lookup->siblings, &dev->lookup_list); in gpio_virtuser_make_lookup_group()
|
| /linux/include/uapi/linux/ |
| H A D | membarrier.h | 70 * threads siblings have passed through a state
94 * call, that all its running threads siblings
120 * siblings have any currently running rseq
|
| /linux/arch/s390/kernel/ |
| H A D | hiperdispatch.c | 108 const struct cpumask *siblings; in hd_add_core() local
113 siblings = topology_sibling_cpumask(cpu); in hd_add_core()
120 cpumask_or(&hd_vmvl_cpumask, &hd_vmvl_cpumask, siblings); in hd_add_core()
124 cpumask_or(&hd_vmvl_cpumask, &hd_vmvl_cpumask, siblings); in hd_add_core()
|
| /linux/tools/testing/selftests/drivers/platform/x86/intel/ifs/ |
| H A D | test_ifs.sh | 32 readonly SIBLINGS="siblings"
381 "$SIBLINGS")
446 ifs_test_cpus "$SIBLINGS" "$IFS_SCAN_MODE" "$image_id"
485 ifs_test_cpus "$SIBLINGS" "$IFS_ARRAY_BIST_SCAN_MODE"
|
| /linux/mm/ |
| H A D | page_counter.c | 295 * parent's and siblings' settings, as well as the actual memory
315 * the unused remainder is available to its siblings.
332 * against immediate siblings whereas 5. is about protecting against
353 * but unused protection is available to siblings that would in effective_protection()
372 * consumed among the siblings anyway. in effective_protection()
|
| /linux/Documentation/devicetree/bindings/net/dsa/ |
| H A D | marvell,mv88e6xxx.yaml | 28 This switch and its siblings will be autodetected from
34 This switch and its siblings will be autodetected from
39 This switch and its siblings will be autodetected from
|
| /linux/kernel/ |
| H A D | Kconfig.preempt | 123 selection across SMT siblings. When enabled -- see
124 prctl(PR_SCHED_CORE) -- task selection ensures that all SMT siblings
|
| /linux/tools/testing/selftests/landlock/ |
| H A D | scoped_base_variants.h | 69 * Parent + child domain (siblings)
139 * Inherited + parent and child domain (siblings)
|
| /linux/drivers/nvme/host/ |
| H A D | multipath.c | 227 list_for_each_entry_rcu(ns, &head->list, siblings) { in nvme_mpath_revalidate_paths()
260 list_for_each_entry_rcu(ns, &head->list, siblings) { in __nvme_find_path()
298 ns = list_next_or_null_rcu(&head->list, &ns->siblings, struct nvme_ns, in nvme_next_ns()
299 siblings); in nvme_next_ns()
302 return list_first_or_null_rcu(&head->list, struct nvme_ns, siblings); in nvme_next_ns()
359 list_for_each_entry_rcu(ns, &head->list, siblings) { in nvme_queue_depth_path()
427 list_for_each_entry_rcu(ns, &head->list, siblings) { in nvme_available_path()
|
| /linux/arch/x86/power/ |
| H A D | hibernate.c | 194 * that SMT siblings are sleeping in hlt, as mwait is not safe in arch_resume_nosmt()
199 * SMT siblings out of hlt, and offline them again so that they in arch_resume_nosmt()
|
| /linux/drivers/iommu/ |
| H A D | fsl_pamu_domain.h | 22 struct list_head link; /* link to domain siblings */
|
| /linux/arch/sparc/include/asm/ |
| H A D | oplib_32.h | 118 * siblings exist.
145 /* Search all siblings starting at the passed node for "name" matching
|
| /linux/arch/powerpc/platforms/pseries/ |
| H A D | smp.c | 206 /* Doorbells can only be used for IPIs between SMT siblings */ in pSeries_smp_probe()
232 * Under PowerVM, FSCR[MSGP] is enabled as guest vCPU siblings are in pSeries_smp_probe()
|
| /linux/arch/alpha/kernel/ |
| H A D | gct.c | 39 /* Now walk the tree, siblings first. */ in gct6_find_nodes()
|