sched/cache: Keep nr_pref_llc_running in the runnable domain, to fix LLC mis-scheduling buggrafted

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 <zhanxusheng@xiaomi.com>
Suggested-by: Chen Yu <yu.c.chen@intel.com>
Signed-off-by: Tim Chen <tim.c.chen@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Kayra Cizmeci <kayracizmeci@gmail.com>
Cc: <stable@kernel.org> # v7.2.x
Link: https://patch.msgid.link/06af61afedac32e6477f57feb4d658f6c411c3af.1790035273.git.tim.c.chen@linux.intel.com

1 files changed, 179 insertions, 0 deletions

diff --git a/samples/rust/rust_debugfs.rs b/samples/rust/rust_debugfs.rs
new file mode 100644
index 000000000..77c19a437
--- /dev/null
+++ b/samples/rust/rust_debugfs.rs
@@ -0,0 +1,179 @@
+// SPDX-License-Identifier: GPL-2.0
+
+// Copyright (C) 2025 Google LLC.
+
+//! Sample DebugFS exporting platform driver
+//!
+//! To successfully probe this driver with ACPI, use an ssdt that looks like
+//!
+//! ```dsl
+//! DefinitionBlock ("", "SSDT", 2, "TEST", "VIRTACPI", 0x00000001)
+//! {
+//!    Scope (\_SB)
+//!    {
+//!        Device (T432)
+//!        {
+//!            Name (_HID, "LNUXBEEF")  // ACPI hardware ID to match
+//!            Name (_UID, 1)
+//!            Name (_STA, 0x0F)        // Device present, enabled
+//!            Name (_DSD, Package () { // Sample attribute
+//!                ToUUID("daffd814-6eba-4d8c-8a91-bc9bbf4aa301"),
+//!                Package() {
+//!                    Package(2) {"compatible", "sample-debugfs"}
+//!                }
+//!            })
+//!            Name (_CRS, ResourceTemplate ()
+//!            {
+//!                Memory32Fixed (ReadWrite, 0xFED00000, 0x1000)
+//!            })
+//!        }
+//!    }
+//! }
+//! ```
+
+use core::str::FromStr;
+use kernel::{
+    acpi,
+    debugfs::{
+        Dir,
+        File, //
+    },
+    device::Core,
+    new_mutex,
+    of,
+    platform,
+    prelude::*,
+    sizes::*,
+    str::CString,
+    sync::{
+        aref::ARef,
+        atomic::{
+            Atomic,
+            Relaxed, //
+        },
+        Mutex,
+    }, //
+};
+
+kernel::module_platform_driver! {
+    type: RustDebugFs,
+    name: "rust_debugfs",
+    authors: ["Matthew Maurer"],
+    description: "Rust DebugFS usage sample",
+    license: "GPL",
+}
+
+#[pin_data]
+struct RustDebugFs {
+    pdev: ARef<platform::Device>,
+    // As we only hold these for drop effect (to remove the directory/files) we have a leading
+    // underscore to indicate to the compiler that we don't expect to use this field directly.
+    _debugfs: Dir,
+    #[pin]
+    _compatible: File<CString>,
+    #[pin]
+    counter: File<Atomic<usize>>,
+    #[pin]
+    inner: File<Mutex<Inner>>,
+    #[pin]
+    array_blob: File<Mutex<[u8; 4]>>,
+    #[pin]
+    vector_blob: File<Mutex<KVec<u8>>>,
+}
+
+#[derive(Debug)]
+struct Inner {
+    x: u32,
+    y: u32,
+}
+
+impl FromStr for Inner {
+    type Err = Error;
+    fn from_str(s: &str) -> Result<Self> {
+        let mut parts = s.split_whitespace();
+        let x = parts
+            .next()
+            .ok_or(EINVAL)?
+            .parse::<u32>()
+            .map_err(|_| EINVAL)?;
+        let y = parts
+            .next()
+            .ok_or(EINVAL)?
+            .parse::<u32>()
+            .map_err(|_| EINVAL)?;
+        if parts.next().is_some() {
+            return Err(EINVAL);
+        }
+        Ok(Inner { x, y })
+    }
+}
+
+kernel::acpi_device_table!(
+    ACPI_TABLE,
+    <RustDebugFs as platform::Driver>::IdInfo,
+    [(acpi::DeviceId::new(c"LNUXBEEF"), ())]
+);
+
+impl platform::Driver for RustDebugFs {
+    type IdInfo = ();
+    type Data<'bound> = Self;
+    const OF_ID_TABLE: Option<of::IdTable<Self::IdInfo>> = None;
+    const ACPI_ID_TABLE: Option<acpi::IdTable<Self::IdInfo>> = Some(&ACPI_TABLE);
+
+    fn probe<'bound>(
+        pdev: &'bound platform::Device<Core<'_>>,
+        _info: Option<&'bound Self::IdInfo>,
+    ) -> impl PinInit<Self, Error> + 'bound {
+        RustDebugFs::new(pdev).pin_chain(|this| {
+            this.counter.store(91, Relaxed);
+            {
+                let mut guard = this.inner.lock();
+                guard.x = guard.y;
+                guard.y = 42;
+            }
+
+            Ok(())
+        })
+    }
+}
+
+impl RustDebugFs {
+    fn build_counter(dir: &Dir) -> impl PinInit<File<Atomic<usize>>> + '_ {
+        dir.read_write_file(c"counter", Atomic::<usize>::new(0))
+    }
+
+    fn build_inner(dir: &Dir) -> impl PinInit<File<Mutex<Inner>>> + '_ {
+        dir.read_write_file(c"pair", new_mutex!(Inner { x: 3, y: 10 }))
+    }
+
+    fn new<'a, 'b>(
+        pdev: &'a platform::Device<Core<'b>>,
+    ) -> impl PinInit<Self, Error> + use<'a, 'b> {
+        let debugfs = Dir::new(c"sample_debugfs");
+        let dev = pdev.as_ref();
+
+        try_pin_init! {
+            Self {
+                _compatible <- debugfs.read_only_file(
+                    c"compatible",
+                    dev.fwnode()
+                        .ok_or(ENOENT)?
+                        .property_read::<CString>(c"compatible")
+                        .required_by(dev)?,
+                ),
+                counter <- Self::build_counter(&debugfs),
+                inner <- Self::build_inner(&debugfs),
+                array_blob <- debugfs.read_write_binary_file(
+                    c"array_blob",
+                    new_mutex!([0x62, 0x6c, 0x6f, 0x62]),
+                ),
+                vector_blob <- debugfs.read_write_binary_file(
+                    c"vector_blob",
+                    new_mutex!(kernel::kvec!(0x42; SZ_4K)?),
+                ),
+                _debugfs: debugfs,
+                pdev: pdev.into(),
+            }
+        }
+    }
+}