Skip to contentSkip to conversation

6.1.189-rc1 build failure: implicit declaration of 'is_core_idle' when CONFIG_NUMA_BALANCING=n

linux-6.1.y Francesco Dolcini 2 participants 2 messages Sep 21 – Sep 22 Stable branch

lore
Share & tools
lore
#0 lore raw

Hello Linux stable maintainers,

Building linux-6.1.y at commit 7c5e16cff0f2 ("Linux 6.1.189-rc1") fails on configs with CONFIG_NUMA_BALANCING disabled

kernel/sched/fair.c: In function 'update_sd_lb_stats':
kernel/sched/fair.c:9912:53: error: implicit declaration of function 'is_core_idle' [-Werror=implicit-function-declaration]

          env->dst_core_idle = !sched_smt_active() || is_core_idle(env->dst_cpu);

This was introduced by:

commit fa94761a8f440dae2a2eddb36da1f017e332aa5c

Author: Andrea Righi <[email protected]>
Date:   Sat May 9 20:07:28 2026 +0200

sched/fair: Reject misfit pulls onto busy SMT siblings on asym-capacity

[ Upstream commit bf6aa722198d3c06e4236e8c5a480f30a64e1513 ]

Thanks,
Francesco

maintainer
#1 lore raw
Quoted 4 lines

kernel/sched/fair.c: In function 'update_sd_lb_stats':
kernel/sched/fair.c:9912:53: error: implicit declaration of function 'is_core_idle' [-Werror=implicit-function-declaration]

          env->dst_core_idle = !sched_smt_active() || is_core_idle(env->dst_cpu);

Confirmed, and 6.1 is the only tree that needs a prerequisite here. is_core_idle() still lives inside the #ifdef CONFIG_NUMA_BALANCING block there, while update_sd_lb_stats() calls it unconditionally. 6.6 and newer already carry 8b36d07f1d63 ("sched/fair: Move is_core_idle() out of CONFIG_NUMA") in their base release, which is why only 6.1 breaks.

I have queued that move ahead of bf6aa722198d ("sched/fair: Reject misfit pulls onto busy SMT siblings on asym-capacity") for 6.1, in that order.

Signature · 2 lines
Thanks,
Sasha