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, 204 insertions, 0 deletions

diff --git a/crypto/scatterwalk.c b/crypto/scatterwalk.c
new file mode 100644
index 000000000..be0e24843
--- /dev/null
+++ b/crypto/scatterwalk.c
@@ -0,0 +1,204 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Cryptographic API.
+ *
+ * Cipher operations.
+ *
+ * Copyright (c) 2002 James Morris <jmorris@intercode.com.au>
+ *               2002 Adam J. Richter <adam@yggdrasil.com>
+ *               2004 Jean-Luc Cooke <jlcooke@certainkey.com>
+ */
+
+#include <crypto/scatterwalk.h>
+#include <linux/kernel.h>
+#include <linux/mm.h>
+#include <linux/module.h>
+#include <linux/scatterlist.h>
+
+void scatterwalk_skip(struct scatter_walk *walk, unsigned int nbytes)
+{
+	struct scatterlist *sg = walk->sg;
+
+	nbytes += walk->offset - sg->offset;
+
+	while (nbytes > sg->length) {
+		nbytes -= sg->length;
+		sg = sg_next(sg);
+	}
+	walk->sg = sg;
+	walk->offset = sg->offset + nbytes;
+}
+EXPORT_SYMBOL_GPL(scatterwalk_skip);
+
+inline void memcpy_from_scatterwalk(void *buf, struct scatter_walk *walk,
+				    unsigned int nbytes)
+{
+	do {
+		unsigned int to_copy;
+
+		to_copy = scatterwalk_next(walk, nbytes);
+		memcpy(buf, walk->addr, to_copy);
+		scatterwalk_done_src(walk, to_copy);
+		buf += to_copy;
+		nbytes -= to_copy;
+	} while (nbytes);
+}
+EXPORT_SYMBOL_GPL(memcpy_from_scatterwalk);
+
+inline void memcpy_to_scatterwalk(struct scatter_walk *walk, const void *buf,
+				  unsigned int nbytes)
+{
+	do {
+		unsigned int to_copy;
+
+		to_copy = scatterwalk_next(walk, nbytes);
+		memcpy(walk->addr, buf, to_copy);
+		scatterwalk_done_dst(walk, to_copy);
+		buf += to_copy;
+		nbytes -= to_copy;
+	} while (nbytes);
+}
+EXPORT_SYMBOL_GPL(memcpy_to_scatterwalk);
+
+void memcpy_from_sglist(void *buf, struct scatterlist *sg,
+			unsigned int start, unsigned int nbytes)
+{
+	struct scatter_walk walk;
+
+	if (unlikely(nbytes == 0)) /* in case sg == NULL */
+		return;
+
+	scatterwalk_start_at_pos(&walk, sg, start);
+	memcpy_from_scatterwalk(buf, &walk, nbytes);
+}
+EXPORT_SYMBOL_GPL(memcpy_from_sglist);
+
+void memcpy_to_sglist(struct scatterlist *sg, unsigned int start,
+		      const void *buf, unsigned int nbytes)
+{
+	struct scatter_walk walk;
+
+	if (unlikely(nbytes == 0)) /* in case sg == NULL */
+		return;
+
+	scatterwalk_start_at_pos(&walk, sg, start);
+	memcpy_to_scatterwalk(&walk, buf, nbytes);
+}
+EXPORT_SYMBOL_GPL(memcpy_to_sglist);
+
+/**
+ * memcpy_sglist() - Copy data from one scatterlist to another
+ * @dst: The destination scatterlist.  Can be NULL if @nbytes == 0.
+ * @src: The source scatterlist.  Can be NULL if @nbytes == 0.
+ * @nbytes: Number of bytes to copy
+ *
+ * The scatterlists can describe exactly the same memory, in which case this
+ * function is a no-op.  No other overlaps are supported.
+ *
+ * Context: Any context
+ */
+void memcpy_sglist(struct scatterlist *dst, struct scatterlist *src,
+		   unsigned int nbytes)
+{
+	unsigned int src_offset, dst_offset;
+
+	if (unlikely(nbytes == 0)) /* in case src and/or dst is NULL */
+		return;
+
+	src_offset = src->offset;
+	dst_offset = dst->offset;
+	for (;;) {
+		/* Compute the length to copy this step. */
+		unsigned int len = min3(src->offset + src->length - src_offset,
+					dst->offset + dst->length - dst_offset,
+					nbytes);
+		struct page *src_page = sg_page(src);
+		struct page *dst_page = sg_page(dst);
+		const void *src_virt;
+		void *dst_virt;
+
+		if (IS_ENABLED(CONFIG_HIGHMEM)) {
+			/* HIGHMEM: we may have to actually map the pages. */
+			const unsigned int src_oip = offset_in_page(src_offset);
+			const unsigned int dst_oip = offset_in_page(dst_offset);
+			const unsigned int limit = PAGE_SIZE;
+
+			/* Further limit len to not cross a page boundary. */
+			len = min3(len, limit - src_oip, limit - dst_oip);
+
+			/* Compute the source and destination pages. */
+			src_page += src_offset / PAGE_SIZE;
+			dst_page += dst_offset / PAGE_SIZE;
+
+			if (src_page != dst_page) {
+				/* Copy between different pages. */
+				memcpy_page(dst_page, dst_oip,
+					    src_page, src_oip, len);
+				flush_dcache_page(dst_page);
+			} else if (src_oip != dst_oip) {
+				/* Copy between different parts of same page. */
+				dst_virt = kmap_local_page(dst_page);
+				memcpy(dst_virt + dst_oip, dst_virt + src_oip,
+				       len);
+				kunmap_local(dst_virt);
+				flush_dcache_page(dst_page);
+			} /* Else, it's the same memory.  No action needed. */
+		} else {
+			/*
+			 * !HIGHMEM: no mapping needed.  Just work in the linear
+			 * buffer of each sg entry.  Note that we can cross page
+			 * boundaries, as they are not significant in this case.
+			 */
+			src_virt = page_address(src_page) + src_offset;
+			dst_virt = page_address(dst_page) + dst_offset;
+			if (src_virt != dst_virt) {
+				memcpy(dst_virt, src_virt, len);
+				if (ARCH_IMPLEMENTS_FLUSH_DCACHE_PAGE)
+					__scatterwalk_flush_dcache_pages(
+						dst_page, dst_offset, len);
+			} /* Else, it's the same memory.  No action needed. */
+		}
+		nbytes -= len;
+		if (nbytes == 0) /* No more to copy? */
+			break;
+
+		/*
+		 * There's more to copy.  Advance the offsets by the length
+		 * copied this step, and advance the sg entries as needed.
+		 */
+		src_offset += len;
+		if (src_offset >= src->offset + src->length) {
+			src = sg_next(src);
+			src_offset = src->offset;
+		}
+		dst_offset += len;
+		if (dst_offset >= dst->offset + dst->length) {
+			dst = sg_next(dst);
+			dst_offset = dst->offset;
+		}
+	}
+}
+EXPORT_SYMBOL_GPL(memcpy_sglist);
+
+struct scatterlist *scatterwalk_ffwd(struct scatterlist dst[2],
+				     struct scatterlist *src,
+				     unsigned int len)
+{
+	for (;;) {
+		if (!len)
+			return src;
+
+		if (src->length > len)
+			break;
+
+		len -= src->length;
+		src = sg_next(src);
+	}
+
+	sg_init_table(dst, 2);
+	sg_set_page(dst, sg_page(src), src->length - len, src->offset + len);
+	scatterwalk_crypto_chain(dst, sg_next(src), 2);
+
+	return dst;
+}
+EXPORT_SYMBOL_GPL(scatterwalk_ffwd);