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, 109 insertions, 0 deletions

diff --git a/Documentation/dev-tools/kunit/faq.rst b/Documentation/dev-tools/kunit/faq.rst
new file mode 100644
index 000000000..b1341c1a6
--- /dev/null
+++ b/Documentation/dev-tools/kunit/faq.rst
@@ -0,0 +1,109 @@
+.. SPDX-License-Identifier: GPL-2.0
+
+==========================
+Frequently Asked Questions
+==========================
+
+How is this different from Autotest, kselftest, and so on?
+==========================================================
+KUnit is a unit testing framework. Autotest, kselftest (and some others) are
+not.
+
+A `unit test <https://martinfowler.com/bliki/UnitTest.html>`_ is supposed to
+test a single unit of code in isolation and hence the name *unit test*. A unit
+test should be the finest granularity of testing and should allow all possible
+code paths to be tested in the code under test. This is only possible if the
+code under test is small and does not have any external dependencies outside of
+the test's control like hardware.
+
+There are no testing frameworks currently available for the kernel that do not
+require installing the kernel on a test machine or in a virtual machine. All
+testing frameworks require tests to be written in userspace and run on the
+kernel under test. This is true for Autotest, kselftest, and some others,
+disqualifying any of them from being considered unit testing frameworks.
+
+Does KUnit support running on architectures other than UML?
+===========================================================
+
+Yes. KUnit can run on any architecture, though the kunit.py tool can only
+build and run kernels for some architectures (of which UML is the default).
+
+You can build and run tests without kunit.py at all on any architecture by
+enabling ``CONFIG_KUNIT=y`` and booting the kernel.
+See Documentation/dev-tools/kunit/run_manual.rst for more details.
+
+Alternatively, kunit.py supports many common architectures using
+cross-compilers and the qemu emulator. This can be done using the ``--arch``
+parameter when running the tests, and the ``--cross_compile`` parameter
+when building (if the architecture is not supported by the host compiler).
+See :ref:`kunit-on-qemu` for more details.
+
+When writing tests targeting other architectures, it's worth keeping the tips
+on the :ref:`kunit-on-non-uml` page in mind.
+
+.. _kinds-of-tests:
+
+What is the difference between a unit test and other kinds of tests?
+====================================================================
+Most existing tests for the Linux kernel would be categorized as an integration
+test, or an end-to-end test.
+
+- A unit test is supposed to test a single unit of code in isolation. A unit
+  test should be the finest granularity of testing and, as such, allows all
+  possible code paths to be tested in the code under test. This is only possible
+  if the code under test is small and does not have any external dependencies
+  outside of the test's control like hardware.
+- An integration test tests the interaction between a minimal set of components,
+  usually just two or three. For example, someone might write an integration
+  test to test the interaction between a driver and a piece of hardware, or to
+  test the interaction between the userspace libraries the kernel provides and
+  the kernel itself. However, one of these tests would probably not test the
+  entire kernel along with hardware interactions and interactions with the
+  userspace.
+- An end-to-end test usually tests the entire system from the perspective of the
+  code under test. For example, someone might write an end-to-end test for the
+  kernel by installing a production configuration of the kernel on production
+  hardware with a production userspace and then trying to exercise some behavior
+  that depends on interactions between the hardware, the kernel, and userspace.
+
+KUnit is not working, what should I do?
+=======================================
+
+Unfortunately, there are a number of things which can break, but here are some
+things to try.
+
+1. Run ``./tools/testing/kunit/kunit.py run`` with the ``--raw_output``
+   parameter. This might show details or error messages hidden by the kunit_tool
+   parser.
+2. Instead of running ``kunit.py run``, try running ``kunit.py config``,
+   ``kunit.py build``, and ``kunit.py exec`` independently. This can help track
+   down where an issue is occurring. (If you think the parser is at fault, you
+   can run it manually against ``stdin`` or a file with ``kunit.py parse``.)
+3. Running the UML kernel directly can often reveal issues or error messages,
+   ``kunit_tool`` ignores. This should be as simple as runningi the ``vmlinux``
+   binary in the output directory (by default ``./.kunit/vmlinux``) after
+   building the UML kernel (for example, by using ``kunit.py build``).
+   Note that UML has some unusual requirements (such as the host having a tmpfs
+   filesystem mounted), and has had issues in the past when built statically and
+   the host has KASLR enabled. (On older host kernels, you may need to run
+   ``setarch `uname -m` -R ./vmlinux`` to disable KASLR.)
+4. Try running KUnit on a different architecture by using the ``--arch``
+   option. On an x86_64 host, using ``--arch=x86_64`` is a good first step.
+5. Make sure the kernel .config has ``CONFIG_KUNIT=y`` and at least one test
+   (e.g. ``CONFIG_KUNIT_EXAMPLE_TEST=y``). kunit_tool will keep its .config
+   around, so you can see what config was used after running ``kunit.py run``.
+   It also preserves any config changes you might make, so you can
+   enable/disable things with ``make ARCH=um menuconfig`` or similar, and then
+   re-run kunit_tool.
+6. Try to run ``make ARCH=um defconfig`` before running ``kunit.py run``. This
+   may help clean up any residual config items which could be causing problems.
+7. Finally, try running KUnit manually, instead of via ``kunit.py``. KUnit can
+   be built into any kernel, or can be built as a module and loaded at runtime.
+   When tests are built-in, they will execute when the kernel boots, and
+   modules will automatically execute associated tests when loaded. Test results
+   can be collected from ``/sys/kernel/debug/kunit/<test suite>/results``, and
+   can be parsed with ``kunit.py parse``. For more details, see
+   Documentation/dev-tools/kunit/run_manual.rst
+
+If none of the above tricks help, you are always welcome to email any issues to
+kunit-dev@googlegroups.com.