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, 87 insertions, 0 deletions

diff --git a/Documentation/admin-guide/binderfs.rst b/Documentation/admin-guide/binderfs.rst
new file mode 100644
index 000000000..41a4db00d
--- /dev/null
+++ b/Documentation/admin-guide/binderfs.rst
@@ -0,0 +1,87 @@
+.. SPDX-License-Identifier: GPL-2.0
+
+The Android binderfs Filesystem
+===============================
+
+Android binderfs is a filesystem for the Android binder IPC mechanism.  It
+allows to dynamically add and remove binder devices at runtime.  Binder devices
+located in a new binderfs instance are independent of binder devices located in
+other binderfs instances.  Mounting a new binderfs instance makes it possible
+to get a set of private binder devices.
+
+Mounting binderfs
+-----------------
+
+Android binderfs can be mounted with::
+
+  mkdir /dev/binderfs
+  mount -t binder binder /dev/binderfs
+
+at which point a new instance of binderfs will show up at ``/dev/binderfs``.
+In a fresh instance of binderfs no binder devices will be present.  There will
+only be a ``binder-control`` device which serves as the request handler for
+binderfs. Mounting another binderfs instance at a different location will
+create a new and separate instance from all other binderfs mounts.  This is
+identical to the behavior of e.g. ``devpts`` and ``tmpfs``. The Android
+binderfs filesystem can be mounted in user namespaces.
+
+Options
+-------
+max
+  binderfs instances can be mounted with a limit on the number of binder
+  devices that can be allocated. The ``max=<count>`` mount option serves as
+  a per-instance limit. If ``max=<count>`` is set then only ``<count>`` number
+  of binder devices can be allocated in this binderfs instance.
+
+stats
+  Using ``stats=global`` enables global binder statistics.
+  ``stats=global`` is only available for a binderfs instance mounted in the
+  initial user namespace. An attempt to use the option to mount a binderfs
+  instance in another user namespace will return a permission error.
+
+Allocating binder Devices
+-------------------------
+
+.. _ioctl: http://man7.org/linux/man-pages/man2/ioctl.2.html
+
+To allocate a new binder device in a binderfs instance a request needs to be
+sent through the ``binder-control`` device node.  A request is sent in the form
+of an `ioctl() <ioctl_>`_.
+
+What a program needs to do is to open the ``binder-control`` device node and
+send a ``BINDER_CTL_ADD`` request to the kernel.  Users of binderfs need to
+tell the kernel which name the new binder device should get.  By default a name
+can only contain up to ``BINDERFS_MAX_NAME`` chars including the terminating
+zero byte.
+
+Once the request is made via an `ioctl() <ioctl_>`_ passing a ``struct
+binder_device`` with the name to the kernel it will allocate a new binder
+device and return the major and minor number of the new device in the struct
+(This is necessary because binderfs allocates a major device number
+dynamically.).  After the `ioctl() <ioctl_>`_ returns there will be a new
+binder device located under /dev/binderfs with the chosen name.
+
+Deleting binder Devices
+-----------------------
+
+.. _unlink: http://man7.org/linux/man-pages/man2/unlink.2.html
+.. _rm: http://man7.org/linux/man-pages/man1/rm.1.html
+
+Binderfs binder devices can be deleted via `unlink() <unlink_>`_.  This means
+that the `rm() <rm_>`_ tool can be used to delete them. Note that the
+``binder-control`` device cannot be deleted since this would make the binderfs
+instance unusable.  The ``binder-control`` device will be deleted when the
+binderfs instance is unmounted and all references to it have been dropped.
+
+Binder features
+---------------
+
+Assuming an instance of binderfs has been mounted at ``/dev/binderfs``, the
+features supported by the binder driver can be located under
+``/dev/binderfs/features/``. The presence of individual files can be tested
+to determine whether a particular feature is supported by the driver.
+
+Example::
+
+        cat /dev/binderfs/features/oneway_spam_detection
+        1