timers/migration: Temporarily disable per capacity hierarchies
authorFrederic Weisbecker <frederic@kernel.org>
Tue, 9 Jun 2026 12:33:56 +0000 (14:33 +0200)
committerThomas Gleixner <tglx@kernel.org>
Tue, 9 Jun 2026 15:25:54 +0000 (17:25 +0200)
commitbb0c250e8e1132723795c1046442ceb01a5ed1b1
treee64da365afd6f5e248e9e460b55136a768375a95
parent45b49d7e3ab6490a9b957a4075344093c43d1f7e
timers/migration: Temporarily disable per capacity hierarchies

Some workloads with different CPU capacities consume more power with
timer migration than before. The recently introduced per capacity
hierarchies were supposed to alleviate this problem. However it appears
to also regress other types of workloads, especially when plenty of
capacities live together in the same machine.

Disable the feature until a reasonable solution is found.

Fixes: 098cbaad8e57 ("timers/migration: Split per-capacity hierarchies")
Reported-by: Christian Loehle <christian.loehle@arm.com>
Signed-off-by: Frederic Weisbecker <frederic@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@kernel.org>
Link: https://patch.msgid.link/20260609123356.28449-1-frederic@kernel.org
Closes: https://lore.kernel.org/all/3b79338f-6cfc-4722-8062-9103db2c8ad1@arm.com
kernel/time/timer_migration.c