From 0d6526f82c3cdefcca47f73f5fc08dc6f335eac6 Mon Sep 17 00:00:00 2001 From: Tim Chen Date: Mon, 21 Sep 2026 17:37:22 -0700 Subject: sched/cache: Keep nr_pref_llc_running in the runnable domain, to fix LLC mis-scheduling bug alb_break_llc() decides whether to break LLC preference during active load balance. It does so by testing that every runnable fair task on the source rq prefers its LLC: env->src_rq->nr_pref_llc_running == env->src_rq->cfs.h_nr_runnable But the two counters cover different sets. nr_pref_llc_running is updated in account_llc_enqueue()/account_llc_dequeue(), next to cfs_rq->nr_queued, so it follows queued tasks. h_nr_runnable is updated in set_delayed()/ clear_delayed() and drops delay-dequeued tasks. So under DELAY_DEQUEUE, a preferring task that goes to sleep stays counted in nr_pref_llc_running while h_nr_runnable falls. The equality then breaks, alb_break_llc() returns false, and active balance is free to pull a task off its preferred LLC. Active balance only moves runnable tasks, and this is the only LLC check it consults: once the stopper runs, LBF_ACTIVE_LB skips the per-task test in can_migrate_task(). The runnable set is the one we want. Fix it on the counter side. A task should be counted in nr_pref_llc_running exactly while it is both queued on its preferred LLC (pref_llc_queued) and runnable (!sched_delayed). Define that membership once in task_pref_llc_runnable(), and adjust the counter only through pref_llc_running_inc()/pref_llc_running_dec() from the four sites that change either input: account_llc_enqueue(), account_llc_dequeue(), set_delayed() and clear_delayed(). Gating every update on the same predicate keeps the delay, wake and dequeue paths from double-counting or underflowing; see the comments at those sites for the ordering. nr_llc_running and sd->llc_counts are not touched and stay on queued semantics. Fixes: 714059f79ff0 ("sched/cache: Handle moving single tasks to/from their preferred LLC") Closes: https://lore.kernel.org/lkml/20260827135000.735138-1-zhanxusheng@xiaomi.com/ Reported-by: Zhan Xusheng Suggested-by: Chen Yu Signed-off-by: Tim Chen Signed-off-by: Peter Zijlstra (Intel) Signed-off-by: Ingo Molnar Reviewed-by: Kayra Cizmeci Cc: # v7.2.x Link: https://patch.msgid.link/06af61afedac32e6477f57feb4d658f6c411c3af.1790035273.git.tim.c.chen@linux.intel.com --- net/rds/info.h | 30 ++++++++++++++++++++++++++++++ 1 file changed, 30 insertions(+) create mode 100644 net/rds/info.h (limited to 'net/rds/info.h') diff --git a/net/rds/info.h b/net/rds/info.h new file mode 100644 index 000000000..1aab62ab6 --- /dev/null +++ b/net/rds/info.h @@ -0,0 +1,30 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +#ifndef _RDS_INFO_H +#define _RDS_INFO_H + +struct rds_info_lengths { + unsigned int nr; + unsigned int each; +}; + +struct rds_info_iterator; + +/* + * These functions must fill in the fields of @lens to reflect the size + * of the available info source. If the snapshot fits in @len then it + * should be copied using @iter. The caller will deduce if it was copied + * or not by comparing the lengths. + */ +typedef void (*rds_info_func)(struct socket *sock, unsigned int len, + struct rds_info_iterator *iter, + struct rds_info_lengths *lens); + +void rds_info_register_func(int optname, rds_info_func func); +void rds_info_deregister_func(int optname, rds_info_func func); +int rds_info_getsockopt(struct socket *sock, int optname, sockopt_t *opt); +void rds_info_copy(struct rds_info_iterator *iter, void *data, + unsigned long bytes); +void rds_info_iter_unmap(struct rds_info_iterator *iter); + + +#endif -- cgit v1.3.1