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 --- scripts/livepatch/init.c | 106 +++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 106 insertions(+) create mode 100644 scripts/livepatch/init.c (limited to 'scripts/livepatch/init.c') diff --git a/scripts/livepatch/init.c b/scripts/livepatch/init.c new file mode 100644 index 000000000..16aff8f73 --- /dev/null +++ b/scripts/livepatch/init.c @@ -0,0 +1,106 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Init code for a livepatch kernel module + */ + +#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt + +#include +#include +#include + +static struct klp_patch *patch; + +static int __init livepatch_mod_init(void) +{ + struct klp_object_ext *obj_exts; + size_t obj_exts_sec_size; + struct klp_object *objs; + unsigned int nr_objs; + int ret; + + obj_exts = klp_find_section_by_name(THIS_MODULE, ".init.klp_objects", + &obj_exts_sec_size); + nr_objs = obj_exts_sec_size / sizeof(*obj_exts); + if (!nr_objs) { + pr_err("nothing to patch!\n"); + ret = -EINVAL; + goto err; + } + + patch = kzalloc_obj(*patch); + if (!patch) { + ret = -ENOMEM; + goto err; + } + + objs = kzalloc(sizeof(struct klp_object) * (nr_objs + 1), GFP_KERNEL); + if (!objs) { + ret = -ENOMEM; + goto err_free_patch; + } + + for (int i = 0; i < nr_objs; i++) { + struct klp_object_ext *obj_ext = obj_exts + i; + struct klp_func_ext *funcs_ext = obj_ext->funcs; + unsigned int nr_funcs = obj_ext->nr_funcs; + struct klp_func *funcs = objs[i].funcs; + struct klp_object *obj = objs + i; + + funcs = kzalloc(sizeof(struct klp_func) * (nr_funcs + 1), GFP_KERNEL); + if (!funcs) { + ret = -ENOMEM; + for (int j = 0; j < i; j++) + kfree(objs[j].funcs); + goto err_free_objs; + } + + for (int j = 0; j < nr_funcs; j++) { + funcs[j].old_name = funcs_ext[j].old_name; + funcs[j].new_func = funcs_ext[j].new_func; + funcs[j].old_sympos = funcs_ext[j].sympos; + } + + obj->name = obj_ext->name; + obj->funcs = funcs; + + memcpy(&obj->callbacks, &obj_ext->callbacks, sizeof(struct klp_callbacks)); + } + + patch->mod = THIS_MODULE; + patch->objs = objs; + + /* TODO patch->states */ + +#ifdef KLP_NO_REPLACE + patch->replace = false; +#else + patch->replace = true; +#endif + + return klp_enable_patch(patch); + +err_free_objs: + kfree(objs); +err_free_patch: + kfree(patch); +err: + return ret; +} + +static void __exit livepatch_mod_exit(void) +{ + struct klp_object *obj; + + klp_for_each_object_static(patch, obj) + kfree(obj->funcs); + + kfree(patch->objs); + kfree(patch); +} + +module_init(livepatch_mod_init); +module_exit(livepatch_mod_exit); +MODULE_LICENSE("GPL"); +MODULE_INFO(livepatch, "Y"); +MODULE_DESCRIPTION("Livepatch module"); -- cgit v1.3.1