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/sched/em_nbyte.c | 79 ++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 79 insertions(+) create mode 100644 net/sched/em_nbyte.c (limited to 'net/sched/em_nbyte.c') diff --git a/net/sched/em_nbyte.c b/net/sched/em_nbyte.c new file mode 100644 index 000000000..c65ffa5ff --- /dev/null +++ b/net/sched/em_nbyte.c @@ -0,0 +1,79 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* + * net/sched/em_nbyte.c N-Byte ematch + * + * Authors: Thomas Graf + */ + +#include +#include +#include +#include +#include +#include +#include +#include + +struct nbyte_data { + struct tcf_em_nbyte hdr; + char pattern[]; +}; + +static int em_nbyte_change(struct net *net, void *data, int data_len, + struct tcf_ematch *em) +{ + struct tcf_em_nbyte *nbyte = data; + + if (data_len < sizeof(*nbyte) || + data_len < (sizeof(*nbyte) + nbyte->len)) + return -EINVAL; + + em->datalen = sizeof(*nbyte) + nbyte->len; + em->data = (unsigned long)kmemdup(data, em->datalen, GFP_KERNEL); + if (em->data == 0UL) + return -ENOMEM; + + return 0; +} + +static int em_nbyte_match(struct sk_buff *skb, struct tcf_ematch *em, + struct tcf_pkt_info *info) +{ + struct nbyte_data *nbyte = (struct nbyte_data *) em->data; + unsigned char *ptr = tcf_get_base_ptr(skb, nbyte->hdr.layer); + + if (!ptr) + return 0; + ptr += nbyte->hdr.off; + + if (!tcf_valid_offset(skb, ptr, nbyte->hdr.len)) + return 0; + + return !memcmp(ptr, nbyte->pattern, nbyte->hdr.len); +} + +static struct tcf_ematch_ops em_nbyte_ops = { + .kind = TCF_EM_NBYTE, + .change = em_nbyte_change, + .match = em_nbyte_match, + .owner = THIS_MODULE, + .link = LIST_HEAD_INIT(em_nbyte_ops.link) +}; + +static int __init init_em_nbyte(void) +{ + return tcf_em_register(&em_nbyte_ops); +} + +static void __exit exit_em_nbyte(void) +{ + tcf_em_unregister(&em_nbyte_ops); +} + +MODULE_DESCRIPTION("ematch classifier for arbitrary skb multi-bytes"); +MODULE_LICENSE("GPL"); + +module_init(init_em_nbyte); +module_exit(exit_em_nbyte); + +MODULE_ALIAS_TCF_EMATCH(TCF_EM_NBYTE); -- cgit v1.3.1