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, 60 insertions, 0 deletions

diff --git a/scripts/stackdelta b/scripts/stackdelta
new file mode 100755
index 000000000..44d2dfd62
--- /dev/null
+++ b/scripts/stackdelta
@@ -0,0 +1,60 @@
+#!/usr/bin/env perl
+# SPDX-License-Identifier: GPL-2.0
+
+# Read two files produced by the stackusage script, and show the
+# delta between them.
+#
+# Currently, only shows changes for functions listed in both files. We
+# could add an option to show also functions which have vanished or
+# appeared (which would often be due to gcc making other inlining
+# decisions).
+#
+# Another possible option would be a minimum absolute value for the
+# delta.
+#
+# A third possibility is for sorting by delta, but that can be
+# achieved by piping to sort -k5,5g.
+
+sub read_stack_usage_file {
+    my %su;
+    my $f = shift;
+    open(my $fh, '<', $f)
+	or die "cannot open $f: $!";
+    while (<$fh>) {
+	chomp;
+	my ($file, $func, $size, $type) = split;
+	# Old versions of gcc (at least 4.7) have an annoying quirk in
+	# that a (static) function whose name has been changed into
+	# for example ext4_find_unwritten_pgoff.isra.11 will show up
+	# in the .su file with a name of just "11". Since such a
+	# numeric suffix is likely to change across different
+	# commits/compilers/.configs or whatever else we're trying to
+	# tweak, we can't really track those functions, so we just
+	# silently skip them.
+	#
+	# Newer gcc (at least 5.0) report the full name, so again,
+	# since the suffix is likely to change, we strip it.
+	next if $func =~ m/^[0-9]+$/;
+	$func =~ s/\..*$//;
+	# Line numbers are likely to change; strip those.
+	$file =~ s/:[0-9]+$//;
+	$su{"${file}\t${func}"} = {size => $size, type => $type};
+    }
+    close($fh);
+    return \%su;
+}
+
+@ARGV == 2
+    or die "usage: $0 <old> <new>";
+
+my $old = read_stack_usage_file($ARGV[0]);
+my $new = read_stack_usage_file($ARGV[1]);
+my @common = sort grep {exists $new->{$_}} keys %$old;
+for (@common) {
+    my $x = $old->{$_}{size};
+    my $y = $new->{$_}{size};
+    my $delta = $y - $x;
+    if ($delta) {
+	printf "%s\t%d\t%d\t%+d\n", $_, $x, $y, $delta;
+    }
+}