Linux kernel mirror (for testing)
git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
kernel
os
linux
sched_ext: idle: clarify comments
Add a comments to clarify about the usage of cpumask_intersects().
Moreover, update scx_select_cpu_dfl() description clarifying that the
final step of the idle selection logic involves searching for any idle
CPU in the system that the task can use.
Reviewed-by: Yury Norov <yury.norov@gmail.com>
Signed-off-by: Andrea Righi <arighi@nvidia.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
authored by
Andrea Righi
and committed by
Tejun Heo
02f034dc
9cf9acee
+6
+6
kernel/sched/ext.c
+6
kernel/sched/ext.c
···
3180
3180
* scx_pick_idle_cpu() can get caught in an infinite loop as
3181
3181
* @cpu is never cleared from idle_masks.smt. Ensure that @cpu
3182
3182
* is eventually cleared.
3183
+
*
3184
+
* NOTE: Use cpumask_intersects() and cpumask_test_cpu() to
3185
+
* reduce memory writes, which may help alleviate cache
3186
+
* coherence pressure.
3183
3187
*/
3184
3188
if (cpumask_intersects(smt, idle_masks.smt))
3185
3189
cpumask_andnot(idle_masks.smt, idle_masks.smt, smt);
···
3411
3407
*
3412
3408
* 4. Pick a CPU within the same NUMA node, if enabled:
3413
3409
* - choose a CPU from the same NUMA node to reduce memory access latency.
3410
+
*
3411
+
* 5. Pick any idle CPU usable by the task.
3414
3412
*
3415
3413
* Step 3 and 4 are performed only if the system has, respectively, multiple
3416
3414
* LLC domains / multiple NUMA nodes (see scx_selcpu_topo_llc and