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, 178 insertions, 0 deletions

diff --git a/drivers/firmware/efi/arm-runtime.c b/drivers/firmware/efi/arm-runtime.c
new file mode 100644
index 000000000..3167cab62
--- /dev/null
+++ b/drivers/firmware/efi/arm-runtime.c
@@ -0,0 +1,178 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Extensible Firmware Interface
+ *
+ * Based on Extensible Firmware Interface Specification version 2.4
+ *
+ * Copyright (C) 2013, 2014 Linaro Ltd.
+ */
+
+#include <linux/dmi.h>
+#include <linux/efi.h>
+#include <linux/io.h>
+#include <linux/memblock.h>
+#include <linux/mm_types.h>
+#include <linux/pgalloc.h>
+#include <linux/pgtable.h>
+#include <linux/preempt.h>
+#include <linux/rbtree.h>
+#include <linux/rwsem.h>
+#include <linux/sched.h>
+#include <linux/slab.h>
+#include <linux/spinlock.h>
+
+#include <asm/cacheflush.h>
+#include <asm/efi.h>
+#include <asm/mmu.h>
+
+#if defined(CONFIG_PTDUMP_DEBUGFS) || defined(CONFIG_ARM_PTDUMP_DEBUGFS)
+#include <asm/ptdump.h>
+
+static struct ptdump_info efi_ptdump_info = {
+	.mm		= &efi_mm,
+	.markers	= (struct addr_marker[]){
+		{ 0,				"UEFI runtime start" },
+		{ EFI_RUNTIME_MAP_END,		"UEFI runtime end" },
+		{ -1,				NULL }
+	},
+	.base_addr	= 0,
+};
+
+static int __init ptdump_init(void)
+{
+	if (efi_enabled(EFI_RUNTIME_SERVICES))
+		ptdump_debugfs_register(&efi_ptdump_info, "efi_page_tables");
+
+	return 0;
+}
+device_initcall(ptdump_init);
+
+#endif
+
+static bool __init efi_virtmap_init(void)
+{
+	efi_memory_desc_t *md;
+
+	efi_mm.pgd = pgd_alloc(&efi_mm);
+	mm_init_cpumask(&efi_mm);
+	init_new_context(NULL, &efi_mm);
+
+	for_each_efi_memory_desc(md) {
+		phys_addr_t phys = md->phys_addr;
+		int ret;
+
+		if (!(md->attribute & EFI_MEMORY_RUNTIME))
+			continue;
+		if (md->virt_addr == U64_MAX)
+			return false;
+
+		ret = efi_create_mapping(&efi_mm, md);
+		if (ret) {
+			pr_warn("  EFI remap %pa: failed to create mapping (%d)\n",
+				&phys, ret);
+			return false;
+		}
+	}
+
+	if (efi_memattr_apply_permissions(&efi_mm, efi_set_mapping_permissions))
+		return false;
+
+	return true;
+}
+
+/*
+ * Enable the UEFI Runtime Services if all prerequisites are in place, i.e.,
+ * non-early mapping of the UEFI system table and virtual mappings for all
+ * EFI_MEMORY_RUNTIME regions.
+ */
+static int __init arm_enable_runtime_services(void)
+{
+	u64 mapsize;
+
+	if (!efi_enabled(EFI_BOOT)) {
+		pr_info("EFI services will not be available.\n");
+		return 0;
+	}
+
+	efi_memmap_unmap();
+
+	mapsize = efi.memmap.desc_size * efi.memmap.nr_map;
+
+	if (efi_memmap_init_late(efi.memmap.phys_map, mapsize)) {
+		pr_err("Failed to remap EFI memory map\n");
+		return 0;
+	}
+
+	if (efi_soft_reserve_enabled()) {
+		efi_memory_desc_t *md;
+
+		for_each_efi_memory_desc(md) {
+			u64 md_size = md->num_pages << EFI_PAGE_SHIFT;
+			struct resource *res;
+
+			if (!(md->attribute & EFI_MEMORY_SP))
+				continue;
+
+			res = kzalloc_obj(*res);
+			if (WARN_ON(!res))
+				break;
+
+			res->start	= md->phys_addr;
+			res->end	= md->phys_addr + md_size - 1;
+			res->name	= "Soft Reserved";
+			res->flags	= IORESOURCE_MEM;
+			res->desc	= IORES_DESC_SOFT_RESERVED;
+
+			insert_resource(&iomem_resource, res);
+		}
+	}
+
+	if (efi_runtime_disabled()) {
+		pr_info("EFI runtime services will be disabled.\n");
+		return 0;
+	}
+
+	if (efi_enabled(EFI_RUNTIME_SERVICES)) {
+		pr_info("EFI runtime services access via paravirt.\n");
+		return 0;
+	}
+
+	pr_info("Remapping and enabling EFI services.\n");
+
+	if (!efi_virtmap_init()) {
+		pr_err("UEFI virtual mapping missing or invalid -- runtime services will not be available\n");
+		return -ENOMEM;
+	}
+
+	/* Set up runtime services function pointers */
+	efi_native_runtime_setup();
+	set_bit(EFI_RUNTIME_SERVICES, &efi.flags);
+
+	return 0;
+}
+early_initcall(arm_enable_runtime_services);
+
+void efi_virtmap_load(void)
+{
+	preempt_disable();
+	efi_set_pgd(&efi_mm);
+}
+
+void efi_virtmap_unload(void)
+{
+	efi_set_pgd(current->active_mm);
+	preempt_enable();
+}
+
+
+static int __init arm_dmi_init(void)
+{
+	/*
+	 * On arm64/ARM, DMI depends on UEFI, and dmi_setup() needs to
+	 * be called early because dmi_id_init(), which is an arch_initcall
+	 * itself, depends on dmi_scan_machine() having been called already.
+	 */
+	dmi_setup();
+	return 0;
+}
+core_initcall(arm_dmi_init);