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, 50 insertions, 0 deletions

diff --git a/Documentation/driver-api/gpio/using-gpio.rst b/Documentation/driver-api/gpio/using-gpio.rst
new file mode 100644
index 000000000..894d88855
--- /dev/null
+++ b/Documentation/driver-api/gpio/using-gpio.rst
@@ -0,0 +1,50 @@
+=========================
+Using GPIO Lines in Linux
+=========================
+
+The Linux kernel exists to abstract and present hardware to users. GPIO lines
+as such are normally not user facing abstractions. The most obvious, natural
+and preferred way to use GPIO lines is to let kernel hardware drivers deal
+with them.
+
+For examples of already existing generic drivers that will also be good
+examples for any other kernel drivers you want to author, refer to
+Documentation/driver-api/gpio/drivers-on-gpio.rst
+
+For any kind of mass produced system you want to support, such as servers,
+laptops, phones, tablets, routers, and any consumer or office or business goods
+using appropriate kernel drivers is paramount. Submit your code for inclusion
+in the upstream Linux kernel when you feel it is mature enough and you will get
+help to refine it, see Documentation/process/submitting-patches.rst.
+
+In Linux GPIO lines also have a userspace ABI.
+
+The userspace ABI is intended for one-off deployments. Examples are prototypes,
+factory lines, maker community projects, workshop specimen, production tools,
+industrial automation, PLC-type use cases, door controllers, in short a piece
+of specialized equipment that is not produced by the numbers, requiring
+operators to have a deep knowledge of the equipment and knows about the
+software-hardware interface to be set up. They should not have a natural fit
+to any existing kernel subsystem and not be a good fit for an operating system,
+because of not being reusable or abstract enough, or involving a lot of non
+computer hardware related policy.
+
+Applications that have a good reason to use the industrial I/O (IIO) subsystem
+from userspace will likely be a good fit for using GPIO lines from userspace as
+well.
+
+Do not under any circumstances abuse the GPIO userspace ABI to cut corners in
+any product development projects. If you use it for prototyping, then do not
+productify the prototype: rewrite it using proper kernel drivers. Do not under
+any circumstances deploy any uniform products using GPIO from userspace.
+
+The userspace ABI is a character device for each GPIO hardware unit (GPIO chip).
+These devices will appear on the system as ``/dev/gpiochip0`` thru
+``/dev/gpiochipN``. Examples of how to directly use the userspace ABI can be
+found in the kernel tree ``tools/gpio`` subdirectory.
+
+For structured and managed applications, we recommend that you make use of the
+libgpiod_ library. This provides helper abstractions, command line utilities
+and arbitration for multiple simultaneous consumers on the same GPIO chip.
+
+.. _libgpiod: https://git.kernel.org/pub/scm/libs/libgpiod/libgpiod.git/