sched_ext: Maintain an online cid mask in the scheduler arenagrafted

Schedulers on the default cid mapping treat [0, nr_online_cids) as the
online set and restart on hotplug. Schedulers that install their own mapping
with scx_bpf_cid_override() have no way to learn which cids are online: the
count no longer identifies members and the CPU-form cpumask is unusable from
cid programs. This is an obvious hole in the cid API.

Add scx_bpf_online_cmask(), a kernel-maintained cmask in the scheduler's
arena, allocated alongside the per-CPU scratch masks and populated after the
cid mapping is finalized and before ops.init(), for child schedulers too.
The pointer stays valid through ops.exit() with no reference to take. It is
the arena offset as a void pointer, the same form struct_ops arena arguments
arrive in. The verifier types the void return as a scalar for the program's
arena cast.

The mask follows the SCX hotplug notifications: seeded from cpu_active_mask
and updated before ops.cid_online/offline() runs, so it lags cpu_online_mask
only inside a hotplug transition. Updates walk the scheduler list under the
lock that also serializes unlinking. Reads are live, not atomic snapshots.
Root initialization excludes hotplug.

v2: Reworded the getter kerneldoc (Andrea Righi).

Signed-off-by: Tejun Heo <tj@kernel.org>
Reviewed-by: Andrea Righi <arighi@nvidia.com>

1 files changed, 215 insertions, 0 deletions

diff --git a/drivers/mtd/tests/readtest.c b/drivers/mtd/tests/readtest.c
new file mode 100644
index 000000000..99670ef91
--- /dev/null
+++ b/drivers/mtd/tests/readtest.c
@@ -0,0 +1,215 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (C) 2006-2008 Nokia Corporation
+ *
+ * Check MTD device read.
+ *
+ * Author: Adrian Hunter <ext-adrian.hunter@nokia.com>
+ */
+
+#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
+
+#include <linux/init.h>
+#include <linux/module.h>
+#include <linux/moduleparam.h>
+#include <linux/err.h>
+#include <linux/mtd/mtd.h>
+#include <linux/slab.h>
+#include <linux/sched.h>
+
+#include "mtd_test.h"
+
+static int dev = -EINVAL;
+module_param(dev, int, S_IRUGO);
+MODULE_PARM_DESC(dev, "MTD device number to use");
+
+static struct mtd_info *mtd;
+static unsigned char *iobuf;
+static unsigned char *iobuf1;
+static unsigned char *bbt;
+
+static int pgsize;
+static int ebcnt;
+static int pgcnt;
+
+static int read_eraseblock_by_page(int ebnum)
+{
+	int i, ret, err = 0;
+	loff_t addr = (loff_t)ebnum * mtd->erasesize;
+	void *buf = iobuf;
+	void *oobbuf = iobuf1;
+
+	for (i = 0; i < pgcnt; i++) {
+		memset(buf, 0 , pgsize);
+		ret = mtdtest_read(mtd, addr, pgsize, buf);
+		if (ret) {
+			if (!err)
+				err = ret;
+		}
+		if (mtd->oobsize) {
+			struct mtd_oob_ops ops = { };
+
+			ops.mode      = MTD_OPS_PLACE_OOB;
+			ops.len       = 0;
+			ops.retlen    = 0;
+			ops.ooblen    = mtd->oobsize;
+			ops.oobretlen = 0;
+			ops.ooboffs   = 0;
+			ops.datbuf    = NULL;
+			ops.oobbuf    = oobbuf;
+			ret = mtd_read_oob(mtd, addr, &ops);
+			if ((ret && !mtd_is_bitflip(ret)) ||
+					ops.oobretlen != mtd->oobsize) {
+				pr_err("error: read oob failed at "
+						  "%#llx\n", (long long)addr);
+				if (!err)
+					err = ret;
+				if (!err)
+					err = -EINVAL;
+			}
+			oobbuf += mtd->oobsize;
+		}
+		addr += pgsize;
+		buf += pgsize;
+	}
+
+	return err;
+}
+
+static void dump_eraseblock(int ebnum)
+{
+	int i, j, n;
+	char line[128];
+	int pg, oob;
+
+	pr_info("dumping eraseblock %d\n", ebnum);
+	n = mtd->erasesize;
+	for (i = 0; i < n;) {
+		char *p = line;
+
+		p += sprintf(p, "%05x: ", i);
+		for (j = 0; j < 32 && i < n; j++, i++)
+			p += sprintf(p, "%02x", (unsigned int)iobuf[i]);
+		printk(KERN_CRIT "%s\n", line);
+		cond_resched();
+	}
+	if (!mtd->oobsize)
+		return;
+	pr_info("dumping oob from eraseblock %d\n", ebnum);
+	n = mtd->oobsize;
+	for (pg = 0, i = 0; pg < pgcnt; pg++)
+		for (oob = 0; oob < n;) {
+			char *p = line;
+
+			p += sprintf(p, "%05x: ", i);
+			for (j = 0; j < 32 && oob < n; j++, oob++, i++)
+				p += sprintf(p, "%02x",
+					     (unsigned int)iobuf1[i]);
+			printk(KERN_CRIT "%s\n", line);
+			cond_resched();
+		}
+}
+
+static int __init mtd_readtest_init(void)
+{
+	uint64_t tmp;
+	int err, i;
+
+	printk(KERN_INFO "\n");
+	printk(KERN_INFO "=================================================\n");
+
+	if (dev < 0) {
+		pr_info("Please specify a valid mtd-device via module parameter\n");
+		return -EINVAL;
+	}
+
+	pr_info("MTD device: %d\n", dev);
+
+	mtd = get_mtd_device(NULL, dev);
+	if (IS_ERR(mtd)) {
+		err = PTR_ERR(mtd);
+		pr_err("error: Cannot get MTD device\n");
+		return err;
+	}
+
+	if (mtd->writesize == 1) {
+		pr_info("not NAND flash, assume page size is 512 "
+		       "bytes.\n");
+		pgsize = 512;
+	} else
+		pgsize = mtd->writesize;
+
+	tmp = mtd->size;
+	do_div(tmp, mtd->erasesize);
+	ebcnt = tmp;
+	pgcnt = mtd->erasesize / pgsize;
+
+	pr_info("MTD device size %llu, eraseblock size %u, "
+	       "page size %u, count of eraseblocks %u, pages per "
+	       "eraseblock %u, OOB size %u\n",
+	       (unsigned long long)mtd->size, mtd->erasesize,
+	       pgsize, ebcnt, pgcnt, mtd->oobsize);
+
+	err = -ENOMEM;
+	iobuf = kmalloc(mtd->erasesize, GFP_KERNEL);
+	if (!iobuf)
+		goto out;
+	iobuf1 = kmalloc(mtd->erasesize, GFP_KERNEL);
+	if (!iobuf1)
+		goto out;
+
+	bbt = kzalloc(ebcnt, GFP_KERNEL);
+	if (!bbt)
+		goto out;
+	err = mtdtest_scan_for_bad_eraseblocks(mtd, bbt, 0, ebcnt);
+	if (err)
+		goto out;
+
+	/* Read all eraseblocks 1 page at a time */
+	pr_info("testing page read\n");
+	for (i = 0; i < ebcnt; ++i) {
+		int ret;
+
+		if (bbt[i])
+			continue;
+		ret = read_eraseblock_by_page(i);
+		if (ret) {
+			dump_eraseblock(i);
+			if (!err)
+				err = ret;
+		}
+
+		ret = mtdtest_relax();
+		if (ret) {
+			err = ret;
+			goto out;
+		}
+	}
+
+	if (err)
+		pr_info("finished with errors\n");
+	else
+		pr_info("finished\n");
+
+out:
+
+	kfree(iobuf);
+	kfree(iobuf1);
+	kfree(bbt);
+	put_mtd_device(mtd);
+	if (err)
+		pr_info("error %d occurred\n", err);
+	printk(KERN_INFO "=================================================\n");
+	return err;
+}
+module_init(mtd_readtest_init);
+
+static void __exit mtd_readtest_exit(void)
+{
+	return;
+}
+module_exit(mtd_readtest_exit);
+
+MODULE_DESCRIPTION("Read test module");
+MODULE_AUTHOR("Adrian Hunter");
+MODULE_LICENSE("GPL");