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, 74 insertions, 0 deletions

diff --git a/Documentation/userspace-api/dma-buf-heaps.rst b/Documentation/userspace-api/dma-buf-heaps.rst
new file mode 100644
index 000000000..f56b743cd
--- /dev/null
+++ b/Documentation/userspace-api/dma-buf-heaps.rst
@@ -0,0 +1,74 @@
+.. SPDX-License-Identifier: GPL-2.0
+
+==============================
+Allocating dma-buf using heaps
+==============================
+
+Dma-buf Heaps are a way for userspace to allocate dma-buf objects. They are
+typically used to allocate buffers from a specific allocation pool, or to share
+buffers across frameworks.
+
+Heaps
+=====
+
+A heap represents a specific allocator. The Linux kernel currently supports the
+following heaps:
+
+ - The ``system`` heap allocates virtually contiguous, cacheable, buffers.
+
+ - The ``system_cc_shared`` heap allocates virtually contiguous, cacheable,
+   buffers using shared (decrypted) memory. It is only present on
+   confidential computing (CoCo) VMs where memory encryption is active
+   (e.g., AMD SEV, Intel TDX). The allocated pages have the encryption
+   bit cleared, making them accessible for device DMA without TDISP
+   support. On non-CoCo VM configurations, this heap is not registered.
+
+ - The ``default_cma_region`` heap allocates physically contiguous,
+   cacheable, buffers. Only present if a CMA region is present. Such a
+   region is usually created either through the kernel commandline
+   through the ``cma`` parameter, a memory region Device-Tree node with
+   the ``linux,cma-default`` property set, or through the
+   ``CMA_SIZE_MBYTES`` or ``CMA_SIZE_PERCENTAGE`` Kconfig options. Prior
+   to Linux 6.17, its name wasn't stable and could be called
+   ``reserved``, ``linux,cma``, or ``default-pool``, depending on the
+   platform.
+
+ - A heap will be created for each reusable region in the device tree
+   with the ``shared-dma-pool`` compatible, using the full device tree
+   node name as its name. The buffer semantics are identical to
+   ``default-cma-region``.
+
+Naming Convention
+=================
+
+``dma-buf`` heaps name should meet a number of constraints:
+
+- The name must be stable, and must not change from one version to the other.
+  Userspace identifies heaps by their name, so if the names ever change, we
+  would be likely to introduce regressions.
+
+- The name must describe the memory region the heap will allocate from, and
+  must uniquely identify it in a given platform. Since userspace applications
+  use the heap name as the discriminant, it must be able to tell which heap it
+  wants to use reliably if there's multiple heaps.
+
+- The name must not mention implementation details, such as the allocator. The
+  heap driver will change over time, and implementation details when it was
+  introduced might not be relevant in the future.
+
+- The name should describe properties of the buffers that would be allocated.
+  Doing so will make heap identification easier for userspace. Such properties
+  are:
+
+  - ``contiguous`` for physically contiguous buffers;
+
+  - ``protected`` for encrypted buffers not accessible the OS;
+
+- The name may describe intended usage. Doing so will make heap identification
+  easier for userspace applications and users.
+
+For example, assuming a platform with a reserved memory region located
+at the RAM address 0x42000000, intended to allocate video framebuffers,
+physically contiguous, and backed by the CMA kernel allocator, good
+names would be ``memory@42000000-contiguous`` or ``video@42000000``, but
+``cma-video`` wouldn't.