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, 111 insertions, 0 deletions

diff --git a/drivers/firmware/efi/ovmf-debug-log.c b/drivers/firmware/efi/ovmf-debug-log.c
new file mode 100644
index 000000000..5b2471ffa
--- /dev/null
+++ b/drivers/firmware/efi/ovmf-debug-log.c
@@ -0,0 +1,111 @@
+// SPDX-License-Identifier: GPL-2.0
+
+#include <linux/efi.h>
+#include <linux/init.h>
+#include <linux/io.h>
+#include <linux/kernel.h>
+#include <linux/kobject.h>
+#include <linux/module.h>
+#include <linux/platform_device.h>
+#include <linux/sysfs.h>
+
+#define OVMF_DEBUG_LOG_MAGIC1  0x3167646d666d766f  // "ovmfmdg1"
+#define OVMF_DEBUG_LOG_MAGIC2  0x3267646d666d766f  // "ovmfmdg2"
+
+struct ovmf_debug_log_header {
+	u64    magic1;
+	u64    magic2;
+	u64    hdr_size;
+	u64    log_size;
+	u64    lock; // edk2 spinlock
+	u64    head_off;
+	u64    tail_off;
+	u64    truncated;
+	u8     fw_version[128];
+};
+
+static struct ovmf_debug_log_header *hdr;
+static u8 *logbuf;
+static u64 logbufsize;
+
+static ssize_t ovmf_log_read(struct file *filp, struct kobject *kobj,
+			     const struct bin_attribute *attr, char *buf,
+			     loff_t offset, size_t count)
+{
+	u64 start, end;
+
+	start = hdr->head_off + offset;
+	if (hdr->head_off > hdr->tail_off && start >= hdr->log_size)
+		start -= hdr->log_size;
+
+	end = start + count;
+	if (start > hdr->tail_off) {
+		if (end > hdr->log_size)
+			end = hdr->log_size;
+	} else {
+		if (end > hdr->tail_off)
+			end = hdr->tail_off;
+	}
+
+	if (start > logbufsize || end > logbufsize)
+		return 0;
+	if (start >= end)
+		return 0;
+
+	memcpy(buf, logbuf + start, end - start);
+	return end - start;
+}
+
+static struct bin_attribute ovmf_log_bin_attr = {
+	.attr = {
+		.name = "ovmf_debug_log",
+		.mode = 0444,
+	},
+	.read = ovmf_log_read,
+};
+
+int __init ovmf_log_probe(unsigned long ovmf_debug_log_table)
+{
+	int ret = -EINVAL;
+	u64 size;
+
+	/* map + verify header */
+	hdr = memremap(ovmf_debug_log_table, sizeof(*hdr), MEMREMAP_WB);
+	if (!hdr) {
+		pr_err("OVMF debug log: header map failed\n");
+		return -EINVAL;
+	}
+
+	if (hdr->magic1 != OVMF_DEBUG_LOG_MAGIC1 ||
+	    hdr->magic2 != OVMF_DEBUG_LOG_MAGIC2) {
+		printk(KERN_ERR "OVMF debug log: magic mismatch\n");
+		goto err_unmap;
+	}
+
+	size = hdr->hdr_size + hdr->log_size;
+	pr_info("OVMF debug log: firmware version: \"%s\"\n", hdr->fw_version);
+	pr_info("OVMF debug log: buffer size: %lluk\n", size / 1024);
+
+	/* map complete log buffer */
+	memunmap(hdr);
+	hdr = memremap(ovmf_debug_log_table, size, MEMREMAP_WB);
+	if (!hdr) {
+		pr_err("OVMF debug log: buffer map failed\n");
+		return -EINVAL;
+	}
+	logbuf = (void *)hdr + hdr->hdr_size;
+	logbufsize = hdr->log_size;
+
+	ovmf_log_bin_attr.size = size;
+	ret = sysfs_create_bin_file(efi_kobj, &ovmf_log_bin_attr);
+	if (ret != 0) {
+		pr_err("OVMF debug log: sysfs register failed\n");
+		goto err_unmap;
+	}
+
+	return 0;
+
+err_unmap:
+	memunmap(hdr);
+	return ret;
+}