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, 127 insertions, 0 deletions

diff --git a/include/asm-generic/cacheflush.h b/include/asm-generic/cacheflush.h
new file mode 100644
index 000000000..7ee8a179d
--- /dev/null
+++ b/include/asm-generic/cacheflush.h
@@ -0,0 +1,127 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef _ASM_GENERIC_CACHEFLUSH_H
+#define _ASM_GENERIC_CACHEFLUSH_H
+
+#include <linux/instrumented.h>
+
+struct mm_struct;
+struct vm_area_struct;
+struct page;
+struct address_space;
+
+/*
+ * The cache doesn't need to be flushed when TLB entries change when
+ * the cache is mapped to physical memory, not virtual memory
+ */
+#ifndef flush_cache_all
+static inline void flush_cache_all(void)
+{
+}
+#endif
+
+#ifndef flush_cache_mm
+static inline void flush_cache_mm(struct mm_struct *mm)
+{
+}
+#endif
+
+#ifndef flush_cache_dup_mm
+static inline void flush_cache_dup_mm(struct mm_struct *mm)
+{
+}
+#endif
+
+#ifndef flush_cache_range
+static inline void flush_cache_range(struct vm_area_struct *vma,
+				     unsigned long start,
+				     unsigned long end)
+{
+}
+#endif
+
+#ifndef flush_cache_page
+static inline void flush_cache_page(struct vm_area_struct *vma,
+				    unsigned long vmaddr,
+				    unsigned long pfn)
+{
+}
+#endif
+
+#ifndef ARCH_IMPLEMENTS_FLUSH_DCACHE_PAGE
+static inline void flush_dcache_page(struct page *page)
+{
+}
+
+#define ARCH_IMPLEMENTS_FLUSH_DCACHE_PAGE 0
+#endif
+
+#ifndef flush_dcache_mmap_lock
+static inline void flush_dcache_mmap_lock(struct address_space *mapping)
+{
+}
+#endif
+
+#ifndef flush_dcache_mmap_unlock
+static inline void flush_dcache_mmap_unlock(struct address_space *mapping)
+{
+}
+#endif
+
+#ifndef flush_icache_range
+static inline void flush_icache_range(unsigned long start, unsigned long end)
+{
+}
+#endif
+
+#ifndef flush_icache_user_range
+#define flush_icache_user_range flush_icache_range
+#endif
+
+#ifndef flush_icache_user_page
+static inline void flush_icache_user_page(struct vm_area_struct *vma,
+					   struct page *page,
+					   unsigned long addr, int len)
+{
+}
+#endif
+
+#ifndef flush_cache_vmap
+static inline void flush_cache_vmap(unsigned long start, unsigned long end)
+{
+}
+#endif
+
+#ifndef flush_cache_vmap_early
+static inline void flush_cache_vmap_early(unsigned long start, unsigned long end)
+{
+}
+#endif
+
+#ifndef flush_cache_vunmap
+static inline void flush_cache_vunmap(unsigned long start, unsigned long end)
+{
+}
+#endif
+
+#ifndef copy_to_user_page
+#define copy_to_user_page(vma, page, vaddr, dst, src, len)	\
+	do { \
+		instrument_copy_to_user((void __user *)dst, src, len); \
+		memcpy(dst, src, len); \
+		flush_icache_user_page(vma, page, vaddr, len); \
+	} while (0)
+#endif
+
+
+#ifndef copy_from_user_page
+#define copy_from_user_page(vma, page, vaddr, dst, src, len)		  \
+	do {								  \
+		instrument_copy_from_user_before(dst, (void __user *)src, \
+						 len);			  \
+		memcpy(dst, src, len);					  \
+		instrument_copy_from_user_after(dst, (void __user *)src, len, \
+						0);			  \
+	} while (0)
+#endif
+
+#endif /* _ASM_GENERIC_CACHEFLUSH_H */