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 --- Documentation/driver-api/hte/hte.rst | 79 ++++++++++++++++++++++++++++++ Documentation/driver-api/hte/index.rst | 22 +++++++++ Documentation/driver-api/hte/tegra-hte.rst | 47 ++++++++++++++++++ 3 files changed, 148 insertions(+) create mode 100644 Documentation/driver-api/hte/hte.rst create mode 100644 Documentation/driver-api/hte/index.rst create mode 100644 Documentation/driver-api/hte/tegra-hte.rst (limited to 'Documentation/driver-api/hte') diff --git a/Documentation/driver-api/hte/hte.rst b/Documentation/driver-api/hte/hte.rst new file mode 100644 index 000000000..153f3233c --- /dev/null +++ b/Documentation/driver-api/hte/hte.rst @@ -0,0 +1,79 @@ +.. SPDX-License-Identifier: GPL-2.0+ + +============================================ +The Linux Hardware Timestamping Engine (HTE) +============================================ + +:Author: Dipen Patel + +Introduction +------------ + +Certain devices have built in hardware timestamping engines which can +monitor sets of system signals, lines, buses etc... in realtime for state +change; upon detecting the change they can automatically store the timestamp at +the moment of occurrence. Such functionality may help achieve better accuracy +in obtaining timestamps than using software counterparts i.e. ktime and +friends. + +This document describes the API that can be used by hardware timestamping +engine provider and consumer drivers that want to use the hardware timestamping +engine (HTE) framework. Both consumers and providers must include +``#include ``. + +The HTE framework APIs for the providers +---------------------------------------- + +.. kernel-doc:: drivers/hte/hte.c + :functions: devm_hte_register_chip hte_push_ts_ns + +The HTE framework APIs for the consumers +---------------------------------------- + +.. kernel-doc:: drivers/hte/hte.c + :functions: hte_init_line_attr hte_ts_get hte_ts_put devm_hte_request_ts_ns hte_request_ts_ns hte_enable_ts hte_disable_ts of_hte_req_count hte_get_clk_src_info + +The HTE framework public structures +----------------------------------- +.. kernel-doc:: include/linux/hte.h + +More on the HTE timestamp data +------------------------------ +The ``struct hte_ts_data`` is used to pass timestamp details between the +consumers and the providers. It expresses timestamp data in nanoseconds in +u64. An example of the typical timestamp data life cycle, for the GPIO line is +as follows:: + + - Monitors GPIO line change. + - Detects the state change on GPIO line. + - Converts timestamps in nanoseconds. + - Stores GPIO raw level in raw_level variable if the provider has that + hardware capability. + - Pushes this hte_ts_data object to HTE subsystem. + - HTE subsystem increments seq counter and invokes consumer provided callback. + Based on callback return value, the HTE core invokes secondary callback in + the thread context. + +HTE subsystem debugfs attributes +-------------------------------- +HTE subsystem creates debugfs attributes at ``/sys/kernel/debug/hte/``. +It also creates line/signal-related debugfs attributes at +``/sys/kernel/debug/hte//