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, 133 insertions, 0 deletions

diff --git a/drivers/firmware/efi/fdtparams.c b/drivers/firmware/efi/fdtparams.c
new file mode 100644
index 000000000..b815d2a75
--- /dev/null
+++ b/drivers/firmware/efi/fdtparams.c
@@ -0,0 +1,133 @@
+// SPDX-License-Identifier: GPL-2.0-only
+
+#define pr_fmt(fmt) "efi: " fmt
+
+#include <linux/module.h>
+#include <linux/init.h>
+#include <linux/efi.h>
+#include <linux/libfdt.h>
+#include <linux/of_fdt.h>
+
+#include <linux/unaligned.h>
+
+enum {
+	SYSTAB,
+	MMBASE,
+	MMSIZE,
+	DCSIZE,
+	DCVERS,
+
+	PARAMCOUNT
+};
+
+static __initconst const char name[][22] = {
+	[SYSTAB] = "System Table         ",
+	[MMBASE] = "MemMap Address       ",
+	[MMSIZE] = "MemMap Size          ",
+	[DCSIZE] = "MemMap Desc. Size    ",
+	[DCVERS] = "MemMap Desc. Version ",
+};
+
+static __initconst const struct {
+	const char	path[17];
+	u8		paravirt;
+	const char	params[PARAMCOUNT][26];
+} dt_params[] = {
+	{
+#ifdef CONFIG_XEN    //  <-------17------>
+		.path = "/hypervisor/uefi",
+		.paravirt = 1,
+		.params = {
+			[SYSTAB] = "xen,uefi-system-table",
+			[MMBASE] = "xen,uefi-mmap-start",
+			[MMSIZE] = "xen,uefi-mmap-size",
+			[DCSIZE] = "xen,uefi-mmap-desc-size",
+			[DCVERS] = "xen,uefi-mmap-desc-ver",
+		}
+	}, {
+#endif
+		.path = "/chosen",
+		.params = {	//  <-----------26----------->
+			[SYSTAB] = "linux,uefi-system-table",
+			[MMBASE] = "linux,uefi-mmap-start",
+			[MMSIZE] = "linux,uefi-mmap-size",
+			[DCSIZE] = "linux,uefi-mmap-desc-size",
+			[DCVERS] = "linux,uefi-mmap-desc-ver",
+		}
+	}
+};
+
+static int __init efi_get_fdt_prop(const void *fdt, int node, const char *pname,
+				   const char *rname, void *var, int size)
+{
+	const void *prop;
+	int len;
+	u64 val;
+
+	prop = fdt_getprop(fdt, node, pname, &len);
+	if (!prop)
+		return 1;
+
+	val = (len == 4) ? (u64)be32_to_cpup(prop) : get_unaligned_be64(prop);
+
+	if (size == 8)
+		*(u64 *)var = val;
+	else
+		*(u32 *)var = (val < U32_MAX) ? val : U32_MAX; // saturate
+
+	if (efi_enabled(EFI_DBG))
+		pr_info("  %s: 0x%0*llx\n", rname, size * 2, val);
+
+	return 0;
+}
+
+u64 __init efi_get_fdt_params(struct efi_memory_map_data *mm)
+{
+	const void *fdt = initial_boot_params;
+	unsigned long systab;
+	int i, j, node;
+	struct {
+		void	*var;
+		int	size;
+	} target[] = {
+		[SYSTAB] = { &systab,		sizeof(systab) },
+		[MMBASE] = { &mm->phys_map,	sizeof(mm->phys_map) },
+		[MMSIZE] = { &mm->size,		sizeof(mm->size) },
+		[DCSIZE] = { &mm->desc_size,	sizeof(mm->desc_size) },
+		[DCVERS] = { &mm->desc_version,	sizeof(mm->desc_version) },
+	};
+
+	BUILD_BUG_ON(ARRAY_SIZE(target) != ARRAY_SIZE(name));
+	BUILD_BUG_ON(ARRAY_SIZE(target) != ARRAY_SIZE(dt_params[0].params));
+
+	if (!fdt)
+		return 0;
+
+	for (i = 0; i < ARRAY_SIZE(dt_params); i++) {
+		node = fdt_path_offset(fdt, dt_params[i].path);
+		if (node < 0)
+			continue;
+
+		if (efi_enabled(EFI_DBG))
+			pr_info("Getting UEFI parameters from %s in DT:\n",
+				dt_params[i].path);
+
+		for (j = 0; j < ARRAY_SIZE(target); j++) {
+			const char *pname = dt_params[i].params[j];
+
+			if (!efi_get_fdt_prop(fdt, node, pname, name[j],
+					      target[j].var, target[j].size))
+				continue;
+			if (!j)
+				goto notfound;
+			pr_err("Can't find property '%s' in DT!\n", pname);
+			return 0;
+		}
+		if (dt_params[i].paravirt)
+			set_bit(EFI_PARAVIRT, &efi.flags);
+		return systab;
+	}
+notfound:
+	pr_info("UEFI not found.\n");
+	return 0;
+}