| /xnu-12377.41.6/doc/observability/ |
| H A D | recount.md | 8 It supports attributing counts to a specific level of the CPU topology (per-CPU and per-CPU kind).
11 These two metrics are abbreviated to cycles-per-instruction, or CPI, for brevity.
17 …ault, Recount tracks its counters per-CPU kind (e.g. performance or efficiency) for threads, per-C…
52 …t each metric as a column and then uses rows for the groupings, like per-CPU or per-CPU kind value…
56 - `recount thread <thread-ptr> [...]` prints a table of per-CPU kind counts for threads.
58 - `recount task <task-ptr> [...]` prints a table of per-CPU counts for tasks.
62 - `recount coalition <coalition-ptr>` prints a table of per-CPU kind counts for each coalition, not…
89 The CPU's previous snapshot, stored in the `_snaps_percpu` per-CPU variable, is subtracted from the…
90 The per-CPU variable is then updated with the current snapshot for the next switch.
114 …ng the lock and iterating the constituent tasks, grouping their per-CPU counters into per-CPU kind…
|
| H A D | mt_stackshot.md | 68 Stackshot uses a basic per-cluster bump allocator to allocate space within the
73 Memory that is freed is put into a per-cluster freelist. Even if the data was
121 execution. There are circular per-cpu buffers containing a list of tracepoints,
|
| H A D | coalitions.md | 19 …nce empty coalitions are valid, `launchd` tries to keep a single coalition per loaded service and …
77 Note that these fields are summed per-task, so in the case where `x` amount of `cpu_time` is billed…
85 …d: [rdar://122812962](rdar://122812962) (Capture ANE energy billed to me and others per coalition).
|
| /xnu-12377.41.6/tools/tests/MPMMTest/ |
| H A D | README | 8 1 server, 4 clients per server (4 total) 400000 messages... in 4.820 seconds
|
| /xnu-12377.41.6/doc/debugging/ |
| H A D | task_ref.md | 61 one available). At the bottom level, there's a per-task (local) ref group under
63 The exact hierarchy of task references (specifically what per-task reference
|
| /xnu-12377.41.6/doc/scheduler/ |
| H A D | sched_clutch_edge.md | 12 … efficiency charecterstics. The Edge scheduler uses the clutch timesharing per cluster and adds ot…
24 …e picked for execution. The kernel maintains a notion of scheduling bucket per thread which are de…
74 The clutch bucket group has an entry to represent runnable threads for the thread group per cluster…
112 The per-bucket thread quantum allows the scheduler to bound the worst case execution latency for a …
186 The Edge scheduler uses **per-thread group recommendations** from the performance controller and **…
190 …e performance controller to specify a preferred cluster per QoS within the thread group i.e. per s…
216 …e and destination nodes (i.e. clusters) for the thread to be migrated. The per-cluster scheduling …
222 The Edge scheduler maintains a per-cluster scheduling latency metric which indicates the latency fo…
230 …ke sure they capture the recent behavior of the threads on the system. The per-cluster scheduling …
273 …ent implementation for cluster shared resource load balancing looks at the per-cluster load at thr…
|
| /xnu-12377.41.6/san/memory/ |
| H A D | kasan-denylist | 26 # Exclude per-cpu data structures the assembly accesses
|
| /xnu-12377.41.6/tools/tests/perf_index/ |
| H A D | README | 50 file_write - writes n bytes to files on disk. There is one file per each thread.
51 file_read - initializes by creating one large file on disk per each thread.
|
| /xnu-12377.41.6/doc/building/ |
| H A D | xnu_build_consolidation.md | 69 2. Splitting the tunable-application code across different files, one per supported architecture (e…
116 element per core or cluster. Whilst this information is not known precisely at compile time anymore…
133 Currently, a family of per-SoC macros are defined at build time to enable XNU to conditionally comp…
|
| H A D | bound_checks.md | 35 We can ask clang to give us one trap instruction per function, which can have
42 On the other hand, we can ask clang to give us one trap instruction per bounds
|
| /xnu-12377.41.6/doc/vm/ |
| H A D | memorystatus_kills.md | 69 … soft limit. In that case the callback does a simulated crash instead of a per process limit kill.…
82 … `memstat_kill_idle_process()` to kill 1 idle daemon per second, up to a maximum of 100 daemons pe…
|
| H A D | memorystatus.md | 20 for killing processes that go over their per-process memory limits.
|
| H A D | freezer.md | 80 …file limit), so we don't have to keep processes around forever. Thus, once per day, we demote froz…
|
| /xnu-12377.41.6/doc/mach_ipc/ |
| H A D | ipc_security_concepts.md | 19 system policies into a single enum per IPC space. This policy is an inherent
|
| /xnu-12377.41.6/tests/unit/ |
| H A D | README.md | 143 High-level per-file textual report:
147 Low-level per-line html pages in a directory structure:
|
| /xnu-12377.41.6/osfmk/mach/ |
| H A D | task.defs | 336 * per-task basis (i.e. the task owns them).
619 * You can only call this once per task.
|
| H A D | thread_act.defs | 417 * per task.
|
| /xnu-12377.41.6/doc/lifecycle/ |
| H A D | startup.md | 128 - `cpu_data_startup_init`: Allocate per-CPU memory that needs to be accessible with MMU disabled
|
| /xnu-12377.41.6/doc/arm/ |
| H A D | apple_speculative_hardening.md | 93 by, on boot, binding each type signature to one of a number of per-size class
104 whether it uses the left or right front depends on its per-boot random type
271 signature groups are randomly assigned to one a handful per-size-class zones.
316 a median and a mean of 9 zones per kalloc_type size class, and so most attacks
628 requires no more than one extra instruction per hardening site. At a machine
|
| /xnu-12377.41.6/config/ |
| H A D | MASTER | 289 # maximum size of the per-process Mach IPC table
369 # enable per-process memory priority tracking
|
| /xnu-12377.41.6/doc/allocators/ |
| H A D | guard-objects.md | 261 of types are randomized per boot.
|
| H A D | api-basics.md | 86 - the allocation must be per-cpu,
|
| /xnu-12377.41.6/osfmk/kern/ |
| H A D | zalloc.h | 1208 #pragma mark XNU only: per-cpu allocations
|
| /xnu-12377.41.6/tests/ |
| H A D | stackshot_tests.m | 297 scenario->nocheck_recordfile = true; // don't do this more than once per scenario
2622 dt_stat_t duration = dt_stat_create("nanoseconds per thread", "%s_duration", flagname);
2623 dt_stat_t size = dt_stat_create("bytes per thread", "%s_size", flagname);
3330 "in per-task shared_cache_dyld_load_info, "
3334 "in per-task shared_cache_dyld_load_info, "
3737 …/* If the task counts below check out, we can be sure that the per-cpu reported thread counts are …
|
| /xnu-12377.41.6/makedefs/ |
| H A D | MakeInc.def | 742 # setup phase. The denylist file itself is divided per-sanitizer, so won't
1252 # CFLAGS_NOLTO_FLAG is needed on a per-file basis (for files
|