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 --- include/trace/syscall.h | 57 +++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 57 insertions(+) create mode 100644 include/trace/syscall.h (limited to 'include/trace/syscall.h') diff --git a/include/trace/syscall.h b/include/trace/syscall.h new file mode 100644 index 000000000..0dd7f2b33 --- /dev/null +++ b/include/trace/syscall.h @@ -0,0 +1,57 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +#ifndef _TRACE_SYSCALL_H +#define _TRACE_SYSCALL_H + +#include +#include +#include +#include + +#include + + +/* + * A syscall entry in the ftrace syscalls array. + * + * @name: name of the syscall + * @syscall_nr: number of the syscall + * @nb_args: number of parameters it takes + * @user_arg_is_str: set if the arg for @user_arg_size is a string + * @user_arg_size: holds @arg that has size of the user space to read + * @user_mask: mask of @args that will read user space + * @types: list of types as strings + * @args: list of args as strings (args[i] matches types[i]) + * @enter_fields: list of fields for syscall_enter trace event + * @enter_event: associated syscall_enter trace event + * @exit_event: associated syscall_exit trace event + */ +struct syscall_metadata { + const char *name; + int syscall_nr; + u8 nb_args:7; + u8 user_arg_is_str:1; + s8 user_arg_size; + short user_mask; + const char **types; + const char **args; + struct list_head enter_fields; + + struct trace_event_call *enter_event; + struct trace_event_call *exit_event; +}; + +#if defined(CONFIG_TRACEPOINTS) && defined(CONFIG_HAVE_SYSCALL_TRACEPOINTS) +static inline void syscall_tracepoint_update(struct task_struct *p) +{ + if (test_syscall_work(SYSCALL_TRACEPOINT)) + set_task_syscall_work(p, SYSCALL_TRACEPOINT); + else + clear_task_syscall_work(p, SYSCALL_TRACEPOINT); +} +#else +static inline void syscall_tracepoint_update(struct task_struct *p) +{ +} +#endif + +#endif /* _TRACE_SYSCALL_H */ -- cgit v1.3.1