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, 219 insertions, 0 deletions

diff --git a/lib/decompress_inflate.c b/lib/decompress_inflate.c
new file mode 100644
index 000000000..e19199f4a
--- /dev/null
+++ b/lib/decompress_inflate.c
@@ -0,0 +1,219 @@
+// SPDX-License-Identifier: GPL-2.0
+#ifdef STATIC
+#define PREBOOT
+/* Pre-boot environment: included */
+
+/* prevent inclusion of _LINUX_KERNEL_H in pre-boot environment: lots
+ * errors about console_printk etc... on ARM */
+#define _LINUX_KERNEL_H
+
+#include "zlib_inflate/inftrees.c"
+#include "zlib_inflate/inffast.c"
+#include "zlib_inflate/inflate.c"
+#ifdef CONFIG_ZLIB_DFLTCC
+#include "zlib_dfltcc/dfltcc.c"
+#include "zlib_dfltcc/dfltcc_inflate.c"
+#endif
+
+#else /* STATIC */
+/* initramfs et al: linked */
+
+#include <linux/zutil.h>
+
+#include "zlib_inflate/inftrees.h"
+#include "zlib_inflate/inffast.h"
+#include "zlib_inflate/inflate.h"
+
+#include "zlib_inflate/infutil.h"
+#include <linux/decompress/inflate.h>
+
+#endif /* STATIC */
+
+#include <linux/decompress/mm.h>
+
+#define GZIP_IOBUF_SIZE (16*1024)
+
+static long INIT nofill(void *buffer, unsigned long len)
+{
+	return -1;
+}
+
+/* Included from initramfs et al code */
+static int INIT __gunzip(unsigned char *buf, long len,
+		       long (*fill)(void*, unsigned long),
+		       long (*flush)(void*, unsigned long),
+		       unsigned char *out_buf, long out_len,
+		       long *pos,
+		       void(*error)(char *x)) {
+	u8 *zbuf;
+	struct z_stream_s *strm;
+	int rc;
+
+	rc = -1;
+	if (flush) {
+		out_len = 0x8000; /* 32 K */
+		out_buf = malloc(out_len);
+	} else {
+		if (!out_len)
+			out_len = ((size_t)~0) - (size_t)out_buf; /* no limit */
+	}
+	if (!out_buf) {
+		error("Out of memory while allocating output buffer");
+		goto gunzip_nomem1;
+	}
+
+	if (buf)
+		zbuf = buf;
+	else {
+		zbuf = malloc(GZIP_IOBUF_SIZE);
+		len = 0;
+	}
+	if (!zbuf) {
+		error("Out of memory while allocating input buffer");
+		goto gunzip_nomem2;
+	}
+
+	strm = malloc(sizeof(*strm));
+	if (strm == NULL) {
+		error("Out of memory while allocating z_stream");
+		goto gunzip_nomem3;
+	}
+
+	strm->workspace = malloc(flush ? zlib_inflate_workspacesize() :
+#ifdef CONFIG_ZLIB_DFLTCC
+	/* Always allocate the full workspace for DFLTCC */
+				 zlib_inflate_workspacesize());
+#else
+				 sizeof(struct inflate_state));
+#endif
+	if (strm->workspace == NULL) {
+		error("Out of memory while allocating workspace");
+		goto gunzip_nomem4;
+	}
+
+	if (!fill)
+		fill = nofill;
+
+	if (len == 0)
+		len = fill(zbuf, GZIP_IOBUF_SIZE);
+
+	/* verify the gzip header */
+	if (len < 10 ||
+	   zbuf[0] != 0x1f || zbuf[1] != 0x8b || zbuf[2] != 0x08) {
+		if (pos)
+			*pos = 0;
+		error("Not a gzip file");
+		goto gunzip_5;
+	}
+
+	/* skip over gzip header (1f,8b,08... 10 bytes total +
+	 * possible asciz filename)
+	 */
+	strm->next_in = zbuf + 10;
+	strm->avail_in = len - 10;
+	/* skip over asciz filename */
+	if (zbuf[3] & 0x8) {
+		do {
+			/*
+			 * If the filename doesn't fit into the buffer,
+			 * the file is very probably corrupt. Don't try
+			 * to read more data.
+			 */
+			if (strm->avail_in == 0) {
+				error("header error");
+				goto gunzip_5;
+			}
+			--strm->avail_in;
+		} while (*strm->next_in++);
+	}
+
+	strm->next_out = out_buf;
+	strm->avail_out = out_len;
+
+	rc = zlib_inflateInit2(strm, -MAX_WBITS);
+
+#ifdef CONFIG_ZLIB_DFLTCC
+	/* Always keep the window for DFLTCC */
+#else
+	if (!flush) {
+		WS(strm)->inflate_state.wsize = 0;
+		WS(strm)->inflate_state.window = NULL;
+	}
+#endif
+
+	while (rc == Z_OK) {
+		if (strm->avail_in == 0) {
+			/* TODO: handle case where both pos and fill are set */
+			len = fill(zbuf, GZIP_IOBUF_SIZE);
+			if (len < 0) {
+				rc = -1;
+				error("read error");
+				break;
+			}
+			strm->next_in = zbuf;
+			strm->avail_in = len;
+		}
+		rc = zlib_inflate(strm, 0);
+
+		/* Write any data generated */
+		if (flush && strm->next_out > out_buf) {
+			long l = strm->next_out - out_buf;
+			if (l != flush(out_buf, l)) {
+				rc = -1;
+				error("write error");
+				break;
+			}
+			strm->next_out = out_buf;
+			strm->avail_out = out_len;
+		}
+
+		/* after Z_FINISH, only Z_STREAM_END is "we unpacked it all" */
+		if (rc == Z_STREAM_END) {
+			rc = 0;
+			break;
+		} else if (rc != Z_OK) {
+			error("uncompression error");
+			rc = -1;
+		}
+	}
+
+	zlib_inflateEnd(strm);
+	if (pos)
+		/* add + 8 to skip over trailer */
+		*pos = strm->next_in - zbuf+8;
+
+gunzip_5:
+	free(strm->workspace);
+gunzip_nomem4:
+	free(strm);
+gunzip_nomem3:
+	if (!buf)
+		free(zbuf);
+gunzip_nomem2:
+	if (flush)
+		free(out_buf);
+gunzip_nomem1:
+	return rc; /* returns Z_OK (0) if successful */
+}
+
+#ifndef PREBOOT
+STATIC int INIT gunzip(unsigned char *buf, long len,
+		       long (*fill)(void*, unsigned long),
+		       long (*flush)(void*, unsigned long),
+		       unsigned char *out_buf,
+		       long *pos,
+		       void (*error)(char *x))
+{
+	return __gunzip(buf, len, fill, flush, out_buf, 0, pos, error);
+}
+#else
+STATIC int INIT __decompress(unsigned char *buf, long len,
+			   long (*fill)(void*, unsigned long),
+			   long (*flush)(void*, unsigned long),
+			   unsigned char *out_buf, long out_len,
+			   long *pos,
+			   void (*error)(char *x))
+{
+	return __gunzip(buf, len, fill, flush, out_buf, out_len, pos, error);
+}
+#endif