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, 248 insertions, 0 deletions

diff --git a/arch/alpha/mm/fault.c b/arch/alpha/mm/fault.c
new file mode 100644
index 000000000..a9816bbc9
--- /dev/null
+++ b/arch/alpha/mm/fault.c
@@ -0,0 +1,248 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ *  linux/arch/alpha/mm/fault.c
+ *
+ *  Copyright (C) 1995  Linus Torvalds
+ */
+
+#include <linux/sched/signal.h>
+#include <linux/kernel.h>
+#include <linux/mm.h>
+#include <asm/io.h>
+
+#define __EXTERN_INLINE inline
+#include <asm/mmu_context.h>
+#include <asm/tlbflush.h>
+#undef  __EXTERN_INLINE
+
+#include <linux/signal.h>
+#include <linux/errno.h>
+#include <linux/string.h>
+#include <linux/types.h>
+#include <linux/ptrace.h>
+#include <linux/mman.h>
+#include <linux/smp.h>
+#include <linux/interrupt.h>
+#include <linux/extable.h>
+#include <linux/uaccess.h>
+#include <linux/perf_event.h>
+
+extern void die_if_kernel(char *,struct pt_regs *,long, unsigned long *);
+
+
+/*
+ * Force a new ASN for a task.
+ */
+
+#ifndef CONFIG_SMP
+unsigned long last_asn = ASN_FIRST_VERSION;
+#endif
+
+void
+__load_new_mm_context(struct mm_struct *next_mm)
+{
+	unsigned long mmc;
+	struct pcb_struct *pcb;
+
+	mmc = __get_new_mm_context(next_mm, smp_processor_id());
+	next_mm->context[smp_processor_id()] = mmc;
+
+	pcb = &current_thread_info()->pcb;
+	pcb->asn = mmc & HARDWARE_ASN_MASK;
+	pcb->ptbr = ((unsigned long) next_mm->pgd - IDENT_ADDR) >> PAGE_SHIFT;
+
+	__reload_thread(pcb);
+}
+
+
+/*
+ * This routine handles page faults.  It determines the address,
+ * and the problem, and then passes it off to handle_mm_fault().
+ *
+ * mmcsr:
+ *	0 = translation not valid
+ *	1 = access violation
+ *	2 = fault-on-read
+ *	3 = fault-on-execute
+ *	4 = fault-on-write
+ *
+ * cause:
+ *	-1 = instruction fetch
+ *	0 = load
+ *	1 = store
+ *
+ * Registers $9 through $15 are saved in a block just prior to `regs' and
+ * are saved and restored around the call to allow exception code to
+ * modify them.
+ */
+
+/* Macro for exception fixup code to access integer registers.  */
+#define dpf_reg(r)							\
+	(((unsigned long *)regs)[(r) <= 8 ? (r) : (r) <= 15 ? (r)-17 :	\
+				 (r) <= 18 ? (r)+11 : (r)-10])
+
+asmlinkage void
+do_page_fault(unsigned long address, unsigned long mmcsr,
+	      long cause, struct pt_regs *regs)
+{
+	struct vm_area_struct * vma;
+	struct mm_struct *mm = current->mm;
+	const struct exception_table_entry *fixup;
+	int si_code = SEGV_MAPERR;
+	vm_fault_t fault;
+	unsigned int flags = FAULT_FLAG_DEFAULT;
+
+	/* As of EV6, a load into $31/$f31 is a prefetch, and never faults
+	   (or is suppressed by the PALcode).  Support that for older CPUs
+	   by ignoring such an instruction.  */
+	if (cause == 0) {
+		unsigned int insn;
+		__get_user(insn, (unsigned int __user *)regs->pc);
+		if ((insn >> 21 & 0x1f) == 0x1f &&
+		    /* ldq ldl ldt lds ldg ldf ldwu ldbu */
+		    (1ul << (insn >> 26) & 0x30f00001400ul)) {
+			regs->pc += 4;
+			return;
+		}
+	}
+
+	/* If we're in an interrupt context, or have no user context,
+	   we must not take the fault.  */
+	if (!mm || faulthandler_disabled())
+		goto no_context;
+
+#ifdef CONFIG_ALPHA_LARGE_VMALLOC
+	if (address >= TASK_SIZE)
+		goto vmalloc_fault;
+#endif
+	if (user_mode(regs))
+		flags |= FAULT_FLAG_USER;
+	perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, regs, address);
+retry:
+	vma = lock_mm_and_find_vma(mm, address, regs);
+	if (!vma)
+		goto bad_area_nosemaphore;
+
+	/* Ok, we have a good vm_area for this memory access, so
+	   we can handle it.  */
+	si_code = SEGV_ACCERR;
+	if (cause < 0) {
+		if (!(vma->vm_flags & VM_EXEC))
+			goto bad_area;
+	} else if (!cause) {
+		/* Allow reads even for write-only mappings */
+		if (!(vma->vm_flags & (VM_READ | VM_WRITE)))
+			goto bad_area;
+	} else {
+		if (!(vma->vm_flags & VM_WRITE))
+			goto bad_area;
+		flags |= FAULT_FLAG_WRITE;
+	}
+
+	/* If for any reason at all we couldn't handle the fault,
+	   make sure we exit gracefully rather than endlessly redo
+	   the fault.  */
+	fault = handle_mm_fault(vma, address, flags, regs);
+
+	if (fault_signal_pending(fault, regs)) {
+		if (!user_mode(regs))
+			goto no_context;
+		return;
+	}
+
+	/* The fault is fully completed (including releasing mmap lock) */
+	if (fault & VM_FAULT_COMPLETED)
+		return;
+
+	if (unlikely(fault & VM_FAULT_ERROR)) {
+		if (fault & VM_FAULT_OOM)
+			goto out_of_memory;
+		else if (fault & VM_FAULT_SIGSEGV)
+			goto bad_area;
+		else if (fault & VM_FAULT_SIGBUS)
+			goto do_sigbus;
+		BUG();
+	}
+
+	if (fault & VM_FAULT_RETRY) {
+		flags |= FAULT_FLAG_TRIED;
+
+		/* No need to mmap_read_unlock(mm) as we would
+		 * have already released it in __lock_page_or_retry
+		 * in mm/filemap.c.
+		 */
+
+		goto retry;
+	}
+
+	mmap_read_unlock(mm);
+
+	return;
+
+	/* Something tried to access memory that isn't in our memory map.
+	   Fix it, but check if it's kernel or user first.  */
+ bad_area:
+	mmap_read_unlock(mm);
+
+ bad_area_nosemaphore:
+	if (user_mode(regs))
+		goto do_sigsegv;
+
+ no_context:
+	/* Are we prepared to handle this fault as an exception?  */
+	if ((fixup = search_exception_tables(regs->pc)) != 0) {
+		unsigned long newpc;
+		newpc = fixup_exception(dpf_reg, fixup, regs->pc);
+		regs->pc = newpc;
+		return;
+	}
+
+	/* Oops. The kernel tried to access some bad page. We'll have to
+	   terminate things with extreme prejudice.  */
+	printk(KERN_ALERT "Unable to handle kernel paging request at "
+	       "virtual address %016lx\n", address);
+	die_if_kernel("Oops", regs, cause, (unsigned long*)regs - 16);
+	make_task_dead(SIGKILL);
+
+	/* We ran out of memory, or some other thing happened to us that
+	   made us unable to handle the page fault gracefully.  */
+ out_of_memory:
+	mmap_read_unlock(mm);
+	if (!user_mode(regs))
+		goto no_context;
+	pagefault_out_of_memory();
+	return;
+
+ do_sigbus:
+	mmap_read_unlock(mm);
+	/* Send a sigbus, regardless of whether we were in kernel
+	   or user mode.  */
+	force_sig_fault(SIGBUS, BUS_ADRERR, (void __user *) address);
+	if (!user_mode(regs))
+		goto no_context;
+	return;
+
+ do_sigsegv:
+	force_sig_fault(SIGSEGV, si_code, (void __user *) address);
+	return;
+
+#ifdef CONFIG_ALPHA_LARGE_VMALLOC
+ vmalloc_fault:
+	if (user_mode(regs))
+		goto do_sigsegv;
+	else {
+		/* Synchronize this task's top level page-table
+		   with the "reference" page table from init.  */
+		long index = pgd_index(address);
+		pgd_t *pgd, *pgd_k;
+
+		pgd = current->active_mm->pgd + index;
+		pgd_k = swapper_pg_dir + index;
+		if (!pgd_present(*pgd) && pgd_present(*pgd_k)) {
+			pgd_val(*pgd) = pgd_val(*pgd_k);
+			return;
+		}
+		goto no_context;
+	}
+#endif
+}