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, 142 insertions, 0 deletions

diff --git a/drivers/nvmem/rcar-efuse.c b/drivers/nvmem/rcar-efuse.c
new file mode 100644
index 000000000..b94ff83b7
--- /dev/null
+++ b/drivers/nvmem/rcar-efuse.c
@@ -0,0 +1,142 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Renesas R-Car E-FUSE/OTP Driver
+ *
+ * Copyright (C) 2024 Glider bv
+ */
+
+#include <linux/device.h>
+#include <linux/export.h>
+#include <linux/io.h>
+#include <linux/nvmem-provider.h>
+#include <linux/platform_device.h>
+#include <linux/pm_runtime.h>
+#include <linux/property.h>
+
+struct rcar_fuse {
+	struct nvmem_keepout keepouts[2];
+	struct nvmem_device *nvmem;
+	struct device *dev;
+	void __iomem *base;
+};
+
+struct rcar_fuse_data {
+	unsigned int bank;	/* 0: PFC + E-FUSE, 1: OPT_MEM + E-FUSE */
+	unsigned int start;	/* inclusive */
+	unsigned int end;	/* exclusive */
+};
+
+static int rcar_fuse_reg_read(void *priv, unsigned int offset, void *val,
+			      size_t bytes)
+{
+	struct rcar_fuse *fuse = priv;
+	int ret;
+
+	ret = pm_runtime_resume_and_get(fuse->dev);
+	if (ret < 0)
+		return ret;
+
+	__ioread32_copy(val, fuse->base + offset, bytes / 4);
+
+	pm_runtime_put(fuse->dev);
+
+	return 0;
+}
+
+static int rcar_fuse_probe(struct platform_device *pdev)
+{
+	struct device *dev = &pdev->dev;
+	const struct rcar_fuse_data *data = device_get_match_data(dev);
+	struct nvmem_config config = {
+		.dev = dev,
+		.name = "rcar-fuse",
+		.id = NVMEM_DEVID_NONE,
+		.owner = THIS_MODULE,
+		.type = NVMEM_TYPE_OTP,
+		.read_only = true,
+		.root_only = true,
+		.reg_read = rcar_fuse_reg_read,
+		.word_size = 4,
+		.stride = 4,
+	};
+	struct rcar_fuse *fuse;
+	struct resource *res;
+	int ret;
+
+	ret = devm_pm_runtime_enable(dev);
+	if (ret < 0)
+		return ret;
+
+	fuse = devm_kzalloc(dev, sizeof(*fuse), GFP_KERNEL);
+	if (!fuse)
+		return -ENOMEM;
+
+	fuse->base = devm_platform_get_and_ioremap_resource(pdev, data->bank,
+							    &res);
+	if (IS_ERR(fuse->base))
+		return PTR_ERR(fuse->base);
+
+	fuse->dev = dev;
+	fuse->keepouts[0].start = 0;
+	fuse->keepouts[0].end = data->start;
+	fuse->keepouts[1].start = data->end;
+	fuse->keepouts[1].end = resource_size(res);
+
+	config.keepout = fuse->keepouts;
+	config.nkeepout = ARRAY_SIZE(fuse->keepouts);
+	config.size = resource_size(res);
+	config.priv = fuse;
+
+	fuse->nvmem = devm_nvmem_register(dev, &config);
+	if (IS_ERR(fuse->nvmem))
+		return dev_err_probe(dev, PTR_ERR(fuse->nvmem),
+				     "Failed to register NVMEM device\n");
+
+	return 0;
+}
+
+static const struct rcar_fuse_data rcar_fuse_v3u = {
+	.bank = 0,
+	.start = 0x0c0,
+	.end = 0x0e8,
+};
+
+static const struct rcar_fuse_data rcar_fuse_s4 = {
+	.bank = 0,
+	.start = 0x0c0,
+	.end = 0x14c,
+};
+
+static const struct rcar_fuse_data rcar_fuse_v4h = {
+	.bank = 1,
+	.start = 0x100,
+	.end = 0x1a0,
+};
+
+static const struct rcar_fuse_data rcar_fuse_v4m = {
+	.bank = 1,
+	.start = 0x100,
+	.end = 0x110,
+};
+
+static const struct of_device_id rcar_fuse_match[] = {
+	{ .compatible = "renesas,r8a779a0-efuse", .data = &rcar_fuse_v3u },
+	{ .compatible = "renesas,r8a779f0-efuse", .data = &rcar_fuse_s4 },
+	{ .compatible = "renesas,r8a779g0-otp", .data = &rcar_fuse_v4h },
+	{ .compatible = "renesas,r8a779h0-otp", .data = &rcar_fuse_v4m },
+	{ /* sentinel */ }
+};
+MODULE_DEVICE_TABLE(of, rcar_fuse_match);
+
+static struct platform_driver rcar_fuse_driver = {
+	.probe = rcar_fuse_probe,
+	.driver = {
+		.name = "rcar_fuse",
+		.of_match_table = rcar_fuse_match,
+	},
+};
+module_platform_driver(rcar_fuse_driver);
+
+MODULE_DESCRIPTION("Renesas R-Car E-FUSE/OTP driver");
+MODULE_LICENSE("GPL");
+MODULE_AUTHOR("Geert Uytterhoeven");