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 --- drivers/block/rnull/rnull.rs | 100 +++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 100 insertions(+) create mode 100644 drivers/block/rnull/rnull.rs (limited to 'drivers/block/rnull/rnull.rs') diff --git a/drivers/block/rnull/rnull.rs b/drivers/block/rnull/rnull.rs new file mode 100644 index 000000000..0ca8715fe --- /dev/null +++ b/drivers/block/rnull/rnull.rs @@ -0,0 +1,100 @@ +// SPDX-License-Identifier: GPL-2.0 + +//! This is a Rust implementation of the C null block driver. + +mod configfs; + +use configfs::IRQMode; +use kernel::{ + block::{ + self, + mq::{ + self, + gen_disk::{self, GenDisk}, + Operations, TagSet, + }, + }, + prelude::*, + sync::{aref::ARef, Arc}, +}; + +module! { + type: NullBlkModule, + name: "rnull_mod", + authors: ["Andreas Hindborg"], + description: "Rust implementation of the C null block driver", + license: "GPL v2", +} + +#[pin_data] +struct NullBlkModule { + #[pin] + configfs_subsystem: kernel::configfs::Subsystem, +} + +impl kernel::InPlaceModule for NullBlkModule { + fn init(_module: &'static ThisModule) -> impl PinInit { + pr_info!("Rust null_blk loaded\n"); + + try_pin_init!(Self { + configfs_subsystem <- configfs::subsystem(), + }) + } +} + +struct NullBlkDevice; + +impl NullBlkDevice { + fn new( + name: &CStr, + block_size: u32, + rotational: bool, + capacity_mib: u64, + irq_mode: IRQMode, + ) -> Result> { + let tagset = Arc::pin_init(TagSet::new(1, 256, 1), GFP_KERNEL)?; + + let queue_data = Box::new(QueueData { irq_mode }, GFP_KERNEL)?; + + gen_disk::GenDiskBuilder::new() + .capacity_sectors(capacity_mib << (20 - block::SECTOR_SHIFT)) + .logical_block_size(block_size)? + .physical_block_size(block_size)? + .rotational(rotational) + .build(fmt!("{}", name.to_str()?), tagset, queue_data) + } +} + +struct QueueData { + irq_mode: IRQMode, +} + +#[vtable] +impl Operations for NullBlkDevice { + type QueueData = KBox; + + #[inline(always)] + fn queue_rq(queue_data: &QueueData, rq: ARef>, _is_last: bool) -> Result { + match queue_data.irq_mode { + IRQMode::None => mq::Request::end_ok(rq) + .map_err(|_e| kernel::error::code::EIO) + // We take no refcounts on the request, so we expect to be able to + // end the request. The request reference must be unique at this + // point, and so `end_ok` cannot fail. + .expect("Fatal error - expected to be able to end request"), + IRQMode::Soft => mq::Request::complete(rq), + } + Ok(()) + } + + fn commit_rqs(_queue_data: &QueueData) {} + + fn complete(rq: ARef>) { + mq::Request::end_ok(rq) + .map_err(|_e| kernel::error::code::EIO) + // We take no refcounts on the request, so we expect to be able to + // end the request. The request reference must be unique at this + // point, and so `end_ok` cannot fail. + .expect("Fatal error - expected to be able to end request"); + } +} -- cgit v1.3.1